A thesis submitted for the degree of Doctor of …368011/s...Publications included in this thesis...

The role of the human gut microbiome in ankylosing spondylitis

Mary-Ellen Clare Costello

Bachelor of Applied Science (Microbiology), Masters of Science (Research)

A thesis submitted for the degree of Doctor of Philosophy at

The University of Queensland in 2015

University of Queensland Diamantina Institute

2

Abstract

Intestinal microbiome dysbiosis and microbial infections have been implicated in a number

of immune mediated diseases, including multiple sclerosis, inflammatory bowel disease

(IBD), and type 1 diabetes. Bacterial infections in the gut and urogenital tract are known to

trigger episodes of reactive arthritis, a form of spondyloarthropathy (SpA), a group of

related inflammatory arthropathies of which ankylosing spondylitis (AS) is the prototypic

disease. There is a close relationship between the gut and SpA, exemplified in reactive

arthritis patients, where a typically self-limiting arthropathy follows either gastrointestinal

infection with Campylobacter, Salmonella, Shigella or Yersinia, or urogenital infection with

Chlamydia. Microbial involvement has been suggested in AS, however, no definitive link

has been established. Multiple genes associated with AS also play a role in gut immunity,

such as genes involved in the IL-23 pathway, which are important regulators of intestinal

‘health’. There is a marked over-representation of genes associated with Crohn’s disease

(CD) that are also associated with AS suggesting the two diseases my have similar

aetiopathogenic mechanisms, possibly involving gut dysbiosis. Up to 70% of AS patients

have subclinical gut inflammation with 5-10% of these patients developing clinically

defined IBD resembling CD. Therefore unravelling the relationship between underlying

host genetics, the intestinal microbiome and the immune system is imperative for

furthering out understanding of AS pathogenesis.

This thesis sought to examine the role of the gut microbiome in AS by characterising the

AS microbiome, examining how changes in host genetics influences intestinal microbial

composition and then exploring at the association between AS susceptibility loci and

microbiome composition in healthy twins.

To date, no comprehensive characterisation, using culture-independent methods, of

intestinal microbiota from terminal ileal (TI) biopsies from AS patients has been performed.

I firstly optimised the methods required for 16S rRNA characterisation of human intestinal

tissue. This included the intestinal biopsies extraction method, 16S PCR development,

next generation library construction and sequencing as well as the analysis methods

employed for characterisation and comparison of the intestinal microbiome.

3

Secondly I profiled the microbiome of TI biopsies from patients with recent-onset, TNF-

antagonist naïve AS and healthy controls (HC), using the methods detailed in Chapter 2 of

this thesis, to investigate if the AS gut carries a distinct microbial signature. Results

showed that AS patients carried a clear and distinct microbial signature when compared to

HC (P<0.001) higher abundance of five families of bacteria Lachnospiraceae (P=0.001),

Ruminococcaceae (P=0.012), Rikenellaceae (P=0.004), Porphyromonadaceae (P=0.001),

and Bacteroidaceae (P=0.001), and a decreases in abundance of two families

Veillonellaceae (P=0.01) and Prevotellaceae (P=0.004). The microbial composition was

found to correlate with disease status and greater differences were observed between

than within disease groups. Further investigation of the AS intestinal microbiome

demonstrated that an assemblage or ‘core’ of five families of bacteria were driving the

intestinal profile. These results are consistent with the hypothesis that genes associated

with AS act at least in part through effects on the gut microbiome.

I then further examined the role of host genetics in microbiome composition by examining

the effect of Endoplasmic reticulum aminopeptidase-1 (ERAP1) in a murine model.

Genome wide association studies have currently identified over 40 loci associated with

ankylosing spondylitis, with ERAP1 is one of the strongest associations, along with HLA-

B27. ERAP1 plays a critical role in overall immune system modulation with systemic

effects. In this chapter, I asked how a lack or down regulation of ERAP1 influences the

inherent gut microbiome composition, and how this impacts on the antigen repertoire and

ensuing affects the immune system.

In the final component of this thesis I further investigated the influence of host genetics,

specifically AS risk loci, on microbiome composition by examining the association between

AS SNPs and intestinal microbiome composition in a non-AS, otherwise healthy

individuals. To investigate the relationship between genes known to be associated with AS

and the gut microbiome, results of matched genetic and 16S rRNA profiling of 982 twins

from the United Kingdom Adult Twin Registry (TwinsUK) were examined to determine if

SNPs, either individually or in combination, were associated with specific microbes. We

have shown, in an otherwise healthy Twin cohort, that genes associated with developing

AS were linked to families of bacteria known to be inflammatory and key indicator species

in the AS microbiome. The allelic risk score linked an AS genetic profile with the families of

Ruminococcaceae, Lachnospiraceae and the order Clostridiales, which are keystone

species of the AS microbiome and members of the core microbiome. Further investigation

4

of individual genes involved in microbial response, peptide processing, trimming and

presentation were also correlated with these same families of bacteria. This intimates that

host genetics play a role in not only overall microbiome composition, but also structure of

core microbiome.

Collectively the work presented in this thesis highlights of the role the human gut

microbiome in AS. It demonstrates evidence for a discrete microbial signature in the TI of

patients with AS compared to HC. The microbial composition was found to correlate with

disease status and driven by a core group of five bacterial families. Genes associated with

developing AS were linked to families of bacteria known to be inflammatory and key

indicator species in the AS microbiome, and investigation of individual genes involved in

microbial response, peptide processing, trimming and presentation were also correlated

with these same families of bacteria. This indicates that underlying host genetics plays a

role in core microbiome as well as overall microbiome structure. Final discussion chapter

of this thesis discusses implications of these associations, limitations of the research as

well as future directions to be explored.

5

Declaration by author

This thesis is composed of my original work, and contains no material previously published

or written by another person except where due reference has been made in the text. I

have clearly stated the contribution by others to jointly-authored works that I have included

in my thesis.

I have clearly stated the contribution of others to my thesis as a whole, including statistical

assistance, survey design, data analysis, significant technical procedures, professional

editorial advice, and any other original research work used or reported in my thesis. The

content of my thesis is the result of work I have carried out since the commencement of

my research higher degree candidature and does not include a substantial part of work

that has been submitted to qualify for the award of any other degree or diploma in any

university or other tertiary institution. I have clearly stated which parts of my thesis, if any,

have been submitted to qualify for another award.

I acknowledge that an electronic copy of my thesis must be lodged with the University

Library and, subject to the policy and procedures of The University of Queensland, the

thesis be made available for research and study in accordance with the Copyright Act

1968 unless a period of embargo has been approved by the Dean of the Graduate School.

I acknowledge that copyright of all material contained in my thesis resides with the

copyright holder(s) of that material. Where appropriate I have obtained copyright

permission from the copyright holder to reproduce material in this thesis.

6

Publications during candidature

Publications in peer-reviewed journals

Costello M-E, Ciccia F, Willner D, Warrington N, Robinson PC, Gardiner B, Marshall M,

Kenna TJ, Triolo G, Brown MA (2014). Intestinal dysbiosis in ankylosing spondylitis.

Arthritis & Rheumatology 10.1002/art.38967 Available online

Rosenbaum J, Lin P, Asquith M, Costello M-E, Kenna T, Brown M (2014). Does the

microbiome play a causal role in spondyloarthritis? Clinical Rheumatology 33: 763-767.

Kenna TJ, Lau MC, Keith P, Ciccia F, Costello M-E, Bradbury L et al (2014). Disease-

associated polymorphisms in ERAP1 do not alter endoplasmic reticulum stress in patients

with ankylosing spondylitis. Genes Immun. 10.1038/gene.2014.62

Costello M-E, Elewaut D, Kenna T, Brown M (2013). Microbes, the gut and ankylosing

spondylitis. Arthritis Research & Therapy 15: 214.

Publications included in this thesis

Costello M-E, Ciccia F, Willner D, Warrington N, Robinson PC, Gardiner B, et al. Brief

Report: Intestinal Dysbiosis in Ankylosing Spondylitis. Arthritis & Rheumatology.

2015;67(3):686-91.– incorporated as Chapter 3 (complete article in Appendices)

Contributor Statement of contribution

Author - Costello M-E (Candidate) Conceptualised the study (20%)

Designed experiments (60%)

Designed microbial analysis strategy (40%)

Performed the experiments (95%)

Performed microbial and statistical analysis

(75%)

Wrote the paper (80%)

Author – Ciccia F Conceptualised the study (5%)

Patient recruitment and collected human

7

Tissue samples (70%)

Author – Willner D Designed microbial analysis strategy (60%)

Wrote and edited the paper (5%)

Author - Warrington N Performed statistical analysis (5%)

Author - Robinson PC Performed statistical analysis (5%)


Author – Gardiner B Designed experiments (20%)

Author – Marshall M Performed analysis (5%)

Author – Kenna TJ Conceptualised the study (10%)


Author – Triolo G Conceptualised the study (5%)

Patient recruitment and collected human tissue samples (30%)

Author – Brown MA Conceptualised the study (60%)

Designed experiments (20%)


Costello M-E, Elewaut D, Kenna T, Brown M (2013). Microbes, the gut and ankylosing

spondylitis. Arthritis Research & Therapy 15: 214 – incorporated and adapted in Chapter 1

(complete article in Appendices)

Contributor Statement of contribution

Author - Costello M-E (Candidate) Wrote and edited the paper (40%)

Author – Elewaut D Wrote and edited the paper (10%)

Author – Kenna TJ Wrote and edited the paper (25%)

Author – Brown MA Wrote and edited the paper (25%)

8

Contributions by others to the thesis

Chapter 2: Associate Professor Graham Radford-Smith, Queensland Institute of Medical

Research, provided the colon biopsies used for the methods development. Dr Francesco

Ciccia and Professor Giovanni Triolo from the University of Palermo provided the terminal

ileal biopsies for the pilot study. Dr Dana Willner provided significant guidance and

technical expertise in the planning and analysis of the experiments.

Chapter 3: Dr Francesco Ciccia and Professor Giovanni Triolo from the University of

Palermo provided all the terminal ileal biopsies and metadata.

Chapter 4: Dr Tony Kenna and Patricia Keith, The University of Queensland Diamantina

Institute, provided assistance in the collection of mouse tissue. Dr Tony Kenna also

provided significant contributions to the conception, design and analysis of the mouse

work in this chapter.

Chapter 5: The 16S rRNA sequencing was performed at Cornell University, in the Lab of

Professor Ruth Ley. 16S rRNA sequenced the faecal samples of all the Twins used in the

study. Julia Goodrich from Professor Ley’s Lab analysed the 16S rRNA sequencing,

filtered, corrected and regressed the OTU table used through out this chapter. Genotyping

was conducted in Professor Tim Spector’s Lab at St Thomas’, King’s College London, and

Michelle Beaumont extracted the genotypes. Dr Jordana Bell also from Professor Tim

Spector’s Lab at St Thomas’, King’s College London, provided technical assistance and

the approach using Merlin to assess association between genes and the microbiome.

Professor David Evans, The University of Queensland Diamantina Institute, also provided

significant technical assistance in the overall methods and design of this chapter. Dr

Adrian Cortes, The University of Queensland Diamantina Institute, provided technical

assistance with the genotype imputation. Dr Nicole Warrington, The University of

Queensland Diamantina Institute provided technical assistance with the allelic risk scores.

Significant contributions to the thesis as a whole have been made by Professor Matt

Brown and Dr Tony Kenna in their capacity as my PhD supervisors. This includes

conceptual and experimental design, writing, editing and advice regarding specific content

in this thesis.

9

Statement of parts of the thesis submitted to qualify for the award of another degree

None.

10

Acknowledgements

From beginning to end, and all the parts in the middle, this thesis has been a collective

endeavor, so I’d like to start off by thanking my supervisor Professor Matt Brown for taking

me on and giving me the opportunity to work at UQDI and Brown lab in a field and on a

project that I’ve become very passionate about. I was just a classical microbiologist with an

interest in infections diseases when I started. I’m very grateful for the space and the

latitude to think and experiment and for all the encouragement, reassurance and support

through out this thesis. I also no longer think that normal flora is boring (Thank you!). I also

owe many, many thanks to my supervisor Dr Tony “Boss man” Kenna. You taught me how

to write, how to present, and the value of practice and persistence; you challenged me to

think, really think. I know it wasn’t meant to turn out this way but I’m very glad you were my

supervisor. Thanks Boss man. Many thanks also goes to my supervisor Dr Gethin Thomas

who is always good for an honest chat and some frank advice. Thanks Gethin. Sometimes

there are people who come and go in your life and indelibly leave their mark. I know it

hasn’t always been easy, but I’m indebted to you Dana for your friendship and support.

You’re an amazing teacher who inspires me to try that little be harder and work that little

bit harder. Thank you.

I am very grateful to all members of Brown Lab past and present. Many of you have

provided crucial experimental help and support as well as friendship. My work would not

have been possible without the support of our Nursing and Admin staff Linda, Kelly and

Kim. I know my biopsies caused a few headaches and dramas, and I just want you know

how appreciative I am of your time and support over the years.

Much of the work presented here would not have been possible without the consent and

participation of the many patients who contributed to the work presented in this thesis – I

am very grateful for their contribution. I would also like to acknowledge the financial

support that I received during my PhD candidature from The University of Queensland

Diamantina Institute, and for travel through the Training fellowship. The fellowship enabled

me to further my studies, scientific education as well as collaborations. I am grateful for

this financial support and opportunity.

Having said that, special mention goes to those in my life who have kept me sane during

this PhD journey. Firstly to Helen “Prof. Brenham” Benham for the coffee, cake and chats.

11

Your support and advice has been invaluable and will continue to be. To my Brown Lab

coffee group members Adrian and Phil, I am forever grateful for your support. I’d like to

thank Adrian for all the grilled cheese and coffee that got me through my final immunology

experiments. You always made time and helped me, even when you didn’t have to. So

thank you for your patience and your friendship; it means more to me than you know. To

Phil for always being there for a whinge, a graph, to hastily read something I’ve written, or

a high-five. Not to mention telling me to suck it up princess when I really needed to hear it.

Thanks Robinson! To my climbing partners Roland, Kylie and Claudio, a massive thank

you for all the climbing and for the thesis-free space. I owe many thanks and probably

several beers to a number of my lab mates that have become very good friends over the

last few years. To Katie “KC” Cremin, Jess “J-Rab” Harris and Lawrie “Lawrence” Wheeler,

your friendship and support means the world to me (yes Lawrence, even during the rugby

season). Thanks to Rabbit for reading several chapters of this thesis, you’re fantastic! To

KC, J-Rab and Ryan, you made coming to work so much fun and our brunches and

dinners such a blast. I do believe we’ve all become ridiculous foodies over the last few

years. I give the whole thing 4.5/5 Survey Co’s and look forward to it continuing next year!

Last but not least to Nicole for all of your time, help and patience with this non-statistical

geneticist. You’ve been fabulous and I’m so thankful for your technical experience but

more importantly your friendship over the last year (especially during thesis write-up and

submitting time).

A massive thank you to my mum and dad for all their love and support, not to mention for

feeding me for the majority of this last year and looking after Bella. I am forever grateful.

To my sister who has had to live with me throughout the duration of my thesis, I know I

wasn’t always the easiest person to live with so the biggest of thanks for putting up with

me and my special kind crazy.

Thanks to my family: Mum, Dad, Katie, Sean and Patrick this wouldn’t have been possible

without you.

12

Keywords

Ankylosing spondylitis, community profiling, microbiome,

Australian and New Zealand Standard Research Classifications (ANZSRC)

110703 Autoimmunity, 40%

060408 Genomics, 30%

060504 Microbial Ecology, 30%

Fields of Research (FoR) Classification

0605 Microbiology, 60%

0604 Genetics, 20%

1107 Immunology, 20%

13

Si Vis Pacem, para bellum

If you wish for peace, prepare for war

Vegetius, 4th Centaury Roman military author

14

Table of Contents Abstract ............................................................................................................................... 2 Declaration by author ........................................................................................................ 5 Publications during candidature ...................................................................................... 6 Contributions by others to the thesis .............................................................................. 8 Statement of parts of the thesis submitted to qualify for the award of another degree ............................................................................................................................................. 9 Acknowledgements .......................................................................................................... 10 Keywords .......................................................................................................................... 12 Australian and New Zealand Standard Research Classifications (ANZSRC) ............. 12 Fields of Research (FoR) Classification ......................................................................... 12 1 Chapter 1 Introduction and Literature Review ......................................................... 23

1.1 Introduction ....................................................................................................................... 23 1.2 Ankylosing spondylitis .................................................................................................... 23

1.2.1 The gastrointestinal tract and AS ................................................................................. 24 1.3 Genetics and overlap between AS and gut disease ...................................................... 25 1.4 The gut, barrier regulation and intestinal epithelial cells ............................................. 27

1.4.1 The physical barrier ..................................................................................................... 27 1.4.1.1 Permeability and Disease .................................................................................................... 28 1.4.1.2 Innate immunity ................................................................................................................... 29

1.4.2 Adaptive Immunity ....................................................................................................... 30 1.4.2.1 IL-23-responsive cells .......................................................................................................... 30 1.4.2.2 γδ T cells ............................................................................................................................. 31 1.4.2.3 Natural killer T cells ............................................................................................................. 32 1.4.2.4 Mucosal-associated invariant T cells ................................................................................... 32 1.4.2.5 Lymphoid tissue inducer- like cells ...................................................................................... 33

1.4.3 Cytokines and gut homeostasis ................................................................................... 33 1.4.3.1 IL-17 .................................................................................................................................... 33 1.4.3.2 IL-22 .................................................................................................................................... 33

1.4.4 Interaction between intestinal microbes and the immune system ................................ 34 1.5 Interrogating the human gut microbiome ...................................................................... 35

1.5.1 Microbial amplicon sequencing .................................................................................... 35 1.5.2 The 16S rRNA .............................................................................................................. 36 1.5.3 Metagenomic sequencing ............................................................................................ 37 1.5.4 Advances in tools for metagenomic studies ................................................................. 37

1.6 Dynamics of the gut microbiome .................................................................................... 38 1.7 The normal microbiome ................................................................................................... 40

1.7.1 Enterotypes .................................................................................................................. 41 1.7.2 Homeostasis ................................................................................................................ 41

1.8 Influence of underlying host genetics on the intestinal microbiome .......................... 42 1.9 The microbiome in immune mediated disease .............................................................. 43

1.9.1 Ankylosing spondylitis .................................................................................................. 43 1.9.2 Rheumatoid arthritis ..................................................................................................... 44 1.9.3 Inflammatory bowel disease ........................................................................................ 45 1.9.4 Psoriasis ....................................................................................................................... 45

1.10 The Chicken and the egg ............................................................................................... 47 1.11 Thesis outline.................................................................................................................. 49 1.12 Aims ................................................................................................................................. 49 1.13 Significance .................................................................................................................... 49

2 Chapter 2 Community profiling the terminal ileal microbiome .............................. 50

15

2.1 Introduction ....................................................................................................................... 50 2.2 Materials and Methods ..................................................................................................... 53

2.2.1 Optimisation of 16S primers 347F 803R and 517F 803R ............................................ 53 2.2.2 16S rRNA PCR ............................................................................................................ 53 2.2.3 In silico assessment of Primer pairs ............................................................................ 53 2.2.4 Pipeline optimisation Intestinal Biopsies ...................................................................... 54 2.2.5 Sequencing of Pilot intestinal biopsies ......................................................................... 54 2.2.6 Bioinformatics .............................................................................................................. 55 2.2.7 Community profiling the terminal ileal microbiome ...................................................... 55

2.2.7.1 Ethics statement .................................................................................................................. 55 2.2.7.2 Patient clinical data .............................................................................................................. 56 2.2.7.3 DNA extraction and sequencing .......................................................................................... 56 2.2.7.4 Bacterial Mock communities ................................................................................................ 58 2.2.7.5 16S rRNA PCR .................................................................................................................... 59 2.2.7.6 Sequencing Library preparation .......................................................................................... 59 2.2.7.7 Illumina MiSeq Sequencing ................................................................................................. 59 2.2.7.8 Bioinformatics and statistics ................................................................................................ 60

2.3 Results ............................................................................................................................... 61 2.3.1 In silico assessment of Primer Pairs ............................................................................ 61 2.3.2 Sequencing of Pilot intestinal biopsies ......................................................................... 62 2.3.3 Sequence depth analysis ............................................................................................. 63 2.3.4 Community profiling the terminal ileal microbiome ...................................................... 65

2.4 Discussion ........................................................................................................................ 70 3 Chapter 3 Intestinal dysbiosis in ankylosing spondylitis ...................................... 73

3.1 Abstract ............................................................................................................................. 73 3.2 Introduction ....................................................................................................................... 74 3.3 Materials and Methods ..................................................................................................... 75

3.3.1 Patient clinical data ...................................................................................................... 75 3.3.2 DNA extraction and sequencing .................................................................................. 77 3.3.3 Mock community .......................................................................................................... 77 3.3.4 16S rRNA PCR ............................................................................................................ 77 3.3.5 Sequencing Library preparation ................................................................................... 78 3.3.6 Illumina MiSeq Sequencing ......................................................................................... 78 3.3.7 Bioinformatics and statistics ......................................................................................... 78

3.4 Results ............................................................................................................................... 81 3.5 Discussion ........................................................................................................................ 90

4 Chapter 4 Effect of ERAP1 genotype on the intestinal microbiome and the impact on the immune system in the ERAP mouse model ....................................................... 93

4.1 Introduction ....................................................................................................................... 93 4.1.1 Peptide presentation and infection ............................................................................... 94 4.1.2 Intestinal cytokines and gut homeostasis .................................................................... 96 4.1.3 Interactions between intestinal microbes and the immune system .............................. 97

4.2 Methods ............................................................................................................................. 98 4.2.1 Antibodies and Flow Cytometry ................................................................................... 98

4.2.1.1 Mice ..................................................................................................................................... 98 4.2.1.2 Antibodies ............................................................................................................................ 99 4.2.1.3 Cell preparation ................................................................................................................... 99 4.2.1.4 Statistics .............................................................................................................................. 99

4.2.2 Sequencing and analysis ........................................................................................... 100 4.2.2.1 DNA extraction .................................................................................................................. 100 4.2.2.2 Mock community ................................................................................................................ 100 4.2.2.3 16S rRNA PCR .................................................................................................................. 100 4.2.2.4 Sequencing Library preparation ........................................................................................ 100 4.2.2.5 Illumina MiSeq Sequencing ............................................................................................... 101 4.2.2.6 Bioinformatics and statistics .............................................................................................. 101

4.3 Results ............................................................................................................................. 101 4.3.1 Alterations in host genetics influences gut microbiome composition ......................... 101

16

4.3.2 Loss of ERAP does not change overall proportions of Immune cell subsets ............. 106 4.3.3 Loss of ERAP leads to differences in cytokine potential ............................................ 109

4.4 Discussion ...................................................................................................................... 113 5 Chapter 5 Effect of underlying host genetics on intestinal microbiome composition .................................................................................................................... 117

5.1 Introduction ..................................................................................................................... 117 5.2 Methods ........................................................................................................................... 121

5.2.1 TwinsUK dataset ........................................................................................................ 121 5.2.2 Sample collection ....................................................................................................... 121 5.2.3 DNA extraction and 16S rRNA amplification .............................................................. 121 5.2.4 16S rRNA filtering and analysis ................................................................................. 121 5.2.5 SNP selection, genotyping and imputation ................................................................ 122 5.2.6 Statistical analysis ...................................................................................................... 123

5.3 Results ............................................................................................................................. 123 5.3.1 Distribution of OTUs is not linked to infection or Phylum ........................................... 123 5.3.2 AS susceptibility variants are associated with overall microbiome structure ............. 125 5.3.3 Genes involved in peptide presentation and microbial sensing are associated with AS core families of bacteria ......................................................................................................... 137

5.4 Discussion ...................................................................................................................... 157 5.5 Supplementary Tables ................................................................................................... 161

6 Chapter 6 Final Discussion ..................................................................................... 183 6.1 Overview and Conclusions ............................................................................................ 183 6.2 Future directions ............................................................................................................ 185

7 References ................................................................................................................ 191 Appendices ..................................................................................................................... 205

17

Tables Table 1.1 Genes associated with ankylosing spondylitis (AS) and their association with

inflammatory bowel disease (IBD). ...................................................................................... 26 Table 2.1 Intestinal biopsies utilised for pipeline optimisation ................................................ 54 Table 2.2 Pipeline optimisation sequencing metrics ................................................................ 55 Table 2.3 Clinical details of patients used in preliminary community profiling studies. ....... 57 Table 2.4 Mock1 construction using bacteria at varying concentrations. ............................... 58 Table 2.5 Mock2 – 16S copy number normalised to 10ng/ul at varying ratios. ....................... 59 Table 2.6 Counts of 16S rRNA sequence reads obtained per biopsy. .................................... 60 Table 2.7 Taxon output by Grinder. ............................................................................................. 62 Table 3.1. Clinical details of patients supplying biopsies at time of collection. Disease

duration is from date of onset of symptoms. ............................................................................ 76 Table 3.2 Bacterial mock community composition. ................................................................... 77 Table 3.3 16S rRNA sequence read counts per sample post quality control and filtering .... 79 Table 4.1 Innate lymphoid cells including cytokine expression and immune function. ........ 98 Table 4.2 Differences in average percentage OTU counts between C57BL/6 and ERAP-/-

mice showing that changes in proportions of key bacteria contribute to the ERAP-/- microbial signature. ............................................................................................................................... 103

Table 4.3 Frequencies of cell subsets showing no significance between ERAP+/+ (WT) and ERAP-/- (KO) strains in spleen and mesenteric lymph nodes. ........................................ 107

Table 4.4 Frequencies of cytokine expression from Immune cell subsets in the spleen of ERAP-/- (KO) and ERAP+/+ (WT) mice. There was no significance in expression between ERAP WT and ERAP KO strains in spleen as determined by unpaired t test. ...................... 110

Table 4.5 Frequencies of cytokine expression from immune cell subsets in the MLN. There was no significant difference in expression between ERAP+/+ (WT) and ERAP-/- (KO) strains in spleen as determined by unpaired t test. ........................................................................... 113

Table 5.1 Reported AS associated loci including correlation of genes associated with inflammatory bowel disease. .............................................................................................. 127

Table 5.2 OTUs with taxonomy associated at P<0.05 from the weighted allelic risk score, including beta and standard error. Six of the OTUs associated belong to the family of Ruminococcaceae, three belong to the family Lachnospiraceae and the remainder belonging to Bifidobacteriaceae and the order Clostridiales. Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_). .......... 130

Table 5.3 OTUs reaching nominal significance (P<0.05) in the unweighted risk score. The major families of bacteria represented are Ruminococcaceae, Lachnospiraceae, Bifidobacteriaceae, the order Clostridiales and Veillonellaceae, which are the same families of bacteria observed in the weighted analysis and in the core microbiome in Chapter 3 of this thesis. Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_). ................................................................................ 131

Table 5.4. AS associated loci investigated for association with each OTU in the microbiome. ............................................................................................................................................... 137

Table 5.5 OTUs reaching nominal significance (P<0.05) in the ERAP locus. 53 OTUs, predominantly Ruminococcaceae and Lachnospiraceae families associated with the ERAP1-ERAP2 SNP, rs10045403; 44 OTUs belonging to the same families of bacteria associated with the ERAP1-ERAP2-LNPEP SNP, rs2910686 and 50 OTUs all belonging to Ruminococcaceae and Lachnospiraceae associated with ERAP1 SNP, rs30187. Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_). .................................................................................................................... 141

Table 5.6 OTUs reaching nominal significance (P<0.05) associated with the IL-23R SNP, rs11209026, predominantly Ruminococcaceae, Lachnospiraceae and Bacteroidaceae families. . Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_). ....................................................................... 151

Table 5.7 OTUs reaching nominal significance (P<0.05) associated with the HLA-B27 allele, predominantly Clostridiaceae, Ruminococcaceae and Lachnospiraceae families. Taxonomy is

18

shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species(s_). ..................................................................................................................... 153

Supplementary table 5.8 OTUs reaching nominal significance (P<0.05) associated with BACH2. 50 OTUs associated with rs17765610 and 81 with rs639575; the majority of which belong to Ruminococcaceae, Lachnospiraceae and Bacteroidaceae families. Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_). .................................................................................................................... 161

Supplementary table 5.9 OTUs reaching nominal significance (P<0.05) associated with NOS2. NOS2 rs2297518 showed 65 OTU associations and rs2531875 with 18 OTUs associations, all dominated by Ruminococcaceae, Lachnospiraceae and Bacteroidaceae families along with Rikenellaceae and Clostridiaceae Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_). ........................................................................................................................................ 167

Supplementary table 5.10 OTUs reaching nominal significance (P<0.05) associated with CARD9 SNP, rs1128905, predominately Lachnospiraceae, Veillonellaceae and Ruminococcaceae, including Faecalibacterium prausnitzii. Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_). ........................................................................................................................................ 171

Supplementary table 5.11 OTUs reaching nominal significance (P<0.05) for NKX2-3, ICOSLG and SH2B3. The families of Ruminococcaceae, Lachnospiraceae dominating the associations. Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_). ....................................................................... 177

19

Figures Figure 1.1 The physical barrier. Separating the intestinal lumen and its inhabiting commensal

bacteria from the underlying lamina propria is a single layer of intestinal epithelial cells (IECs). These IECs are stitched together, creating a tight junction and regulating the paracellular flux. IECs also secrete soluble factors that are crucial to intestinal homeostasis, such as mucins and anti-microbial peptides (AMPs), including lysozymes, IgA, defensins, and C-type lectins such as RegIIIγ. Release of these molecules into luminal crypts is thought to prevent microbial invasion into the crypt microenvironment as well as limit bacteria-epithelial cell contact. Toll-like receptors (TLRs) are also expressed on IECs to sense a breach of barrier or bacterial invasion. Underneath the IECs, the lamina propria contains T cells, bacteria-sampling dendritic cells (DCs), and macrophages. Figure from: Costello et al. Arthritis Research & Therapy 2013 15:214 doi:10.1186/ar4228 ............................................................................. 28

Figure 1.2 The immune barrier. Dendritic cells (DCs) and macrophages patrol the gastrointestinal tract in high numbers. They densely populate the intestinal lamina propria and form a widespread microbe-sensing network. Activated DCs can secrete a number of cytokines and chemokines, including interleukin-23 (IL-23), IL-6, and IL-1, activating IL-23-responsive cells. IL-23-, IL-17-, and IL-22-producing cells are enriched in gut mucosa, and IL-17 and IL-22 are known to be important regulators of intestinal 'health'. IL-17 plays important roles in intestinal homoeostasis in several ways, including maintenance of epithelial barrier tight junctions. LTi, lymphoid tissue inducer; MAIT, mucosa-associated invariant T; NKT, natural killer T; T reg, regulatory T; TNF, tumour necrosis factor. Figure from: Costello et al. Arthritis Research & Therapy 2013 15:214 doi:10.1186/ar4228 ............................................................................. 31

Figure 1.3 Cartoon depiction of the 16S rRNA gene. Modified 16S rRNA gene highlighting the variable regions of the gene commonly used for taxonomy. Circled in red are the two constant commonly used and circled in yellow is the V3-V5 region commonly used for amplicon sequencing in the gut. .............................................................................................................. 37

Figure 2.1 Phyla represented in gut biopsies showing percentage abundance. Comparison of primer pairs 347F 80R and 517F 803R demonstrating no significant differences when trialled in five colon biopsies for the purpose of pipeline optimisation ................................................. 63

Figure 2.2 Rank abundance graph showing number of reads (read depth) with number of species identified for the sample Q1462. ............................................................................ 64

Figure 2.3 Rarefaction curves for microbial communities in AS patients (red curve), CD patients (blue curve), healthy controls (orange curve) and the mock communities (green curve). ..................................................................................................................................... 65

Figure 2.4 Hierarchical clustering of microbial communities based on Weighted Unifrac distance clustering of terminal ileal biopsies of patients with ankylosing spondylitis (AS), Crohn’s disease (CD), healthy controls (HC) as well as the two mock community controls. ... 66

Figure 2.5 Principal coordinates analysis of technical and biological replicates (transparent points) showing the grouping of ankylosing spondylitis (AS) samples compared to Crohn’s disease (CD) and healthy controls (HC) with the grouping of the replicates shown in the red circles. ...................................................................................................................................... 67

Figure 2.6 Comparison of variation between and within samples and affection status showing greater significance between and within affection status than between biological and technical replicates. Significance was assessed using Mann-Whitney test (P<0.001 = ***) ......................................................................................................................... 68

Figure 2.7 Microbial community profiles of terminal ileal (TI) biopsies from ankylosing spondylitis (AS), Crohn’s disease (CD) patients and healthy controls (HC). Biological replicates are indicated as ‘a’ and ‘b’. Each row represents a different OTU with the abundance as a percentage of the total population as represented by colour. Phyla and selected families are noted on the left, and a subset of OTU classifications on the right. Microbial biomass of each biopsy is shown by the bar graph at the bottom of the heat map. . 69

Figure 3.1 Differences between AS and HC microbial communities. Bar chart showing the difference in taxonomy at the Phylum level between AS and HC noting the increase in Bacteroides and the change in the Bacteroides to Firmicutes ratio in the AS samples. .......... 82

Figure 4.1 Principal coordinate analysis and taxonomy showing community differences between ERAP-/- mice and control C57BL/6. (A) Unweighted Principal Coordinate Analysis

20

(PCoA) showing distinct clustering of ERAP-/- mice from control C57BL/6. (B) Bar chart showing the difference in taxonomy at the Phylum level between C57BL/6 and ERAP-/-, noting the increase in Bacteroides in C57BL/6 and the change in the Bacteroides to Firmicutes ratio between the samples. ............................................................................................................ 103

Figure 4.2 ERAP-/- Intestinal community Co-occurrence network showing the positive (blue lines) and negative (red lines) correlations between indicator species. The thickness of the line denotes the strength of the correlation (Pearson correlation). .................................. 104

Figure 4.3 Unweighted Principal coordinates analysis (PCoA) of a single litter of ERAP mice comprising three ERAP+/+, four ERAP+/- and one ERAP-/- showing distinct grouping of ERAP+/+ away from the other genotypes and separate clustering of ERAP-/- and ERAP+/-. ............................................................................................................................................... 105

Figure 4.4 Example gating strategy for isolation, identification and characterisation of T cells. Doublets were removed and dead cells were excluded using Aqua Viability dye discrimination. Cell subset specific analysis of cell phenotype and activation status was performed by analysing cell surface protein expression, with transcription factor expression determined intracellularly following cell permeabilisation. Cytokine production was determined intracellularly following in vitro stimulation. ............................................................................ 106

Figure 4.5 Frequencies of major immune cell subsets in ERAP+/+ (WT) and ERAP-/- (KO) from the Spleen (A) and mesenteric lymph nodes (C) and activation status respectfully (B and D) shown as mean ± SEM and compared by unpaired t test. ......................................... 108

Figure 4.6 Immune cell subsets and expression of cytokine from the spleen of ERAP-/- (KO) mice compared to ERAP+/+ (WT). (A) Expression of IL-17 in the spleen, noting the increase in secretion from γδ T cells, and the absence of IL-17 secretion from CD3+ T cells as a whole and CD8+ T cells. (B) Expression of IL-22 showing a slight decrease across all cell types except for NK cells and no significant differences in expression of either IL-10 (C) or TNF (D). Data shown as mean ± SEM and compared by unpaired t test. ............................................ 111

Figure 4.7 Immune cell subsets and expression of cytokine from the mesenteric lymph nodes (MLN) of ERAP-/- (KO) mice compared to ERAP+/+ (WT). (A) Expression of IL-17 in the MLN, showing an increase in secretion from macrophages. (B) Expression of IL-22 showing an increase in secretion from CD3+ T cells, but undetectable levels across other T cell subsets. A slight increase in IL-10 secretion was detected in macrophages (C) and TNF secretion in the MLN mirrored what was seen in the spleen (D). .......................................... 112

Figure 5.1 Representative OTUs prior to normalisation, demonstrating the range of distributions present within the OTU table and the same order of bacteria, showing the range of mean, skew and kurtosis of each OTU. Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_). ................... 124

Figure 5.2 Distributions of the family Lachnospiraceae demonstrating variation of distribution patterns within the same family of bacteria. Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_). ........................................................................................................................................ 125

Figure 5.3 Distributions of selected members of the family Enterobacteriaceae, which contains a number of pathogenic bacteria including Shigella, Klebsiella and E.coli, showing no clear pattern of distribution. Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_). ............................................ 125

Figure 5.4 Histogram of weighted allelic risk score demonstrating the effect of HLA-B27 genotype (GG, AG, AA). ...................................................................................................... 126

21

List of commonly used Abbreviations 16S rRNA 16S ribosomal RNA

AS ankylosing spondylitis

BLAST basic local alignment search tool

CD Crohn’s disease

CIA collagen-induced arthritis

DC dendritic cell

ER endoplasmic reticulum

ERAP1 endoplasmic reticulum aminopeptidase-1

FoxP3 forkhead box P3

γδ T cell gamma-delta T cell

GF germ free

GWAS genome wide association study

HC healthy control

HLA human leukocyte antigen

IBD Inflammatory bowel disease

IL-23 interleukin-23



IL-23R interleukin-23 receptor

ILC innate lymphoid cell

IFNγ interferon gamma

LTi cell lymphoid tissue inducer cell

MAIT mucosal-associated invariant T cells

MHC major histocompatibility complex

NSAIDs non-steroidal anti-inflammatories

NK cell natural killer cell

NKT cell natural killer T cell

PCoA principal coordinates analysis

OTU operational taxonomy unit

RA rheumatoid arthritis

RORyt retinoic acid related orphan receptor gamma

SFB segmented filamentous bacteria

SNP single nucleotide polymorphism

22

SpA spondyloarthropathy

Th17 Thelper 17

TI terminal ileal

TLR toll-like receptor

Tregs regulatory T cells

TNF tumour necrosis factor

TGF-β transforming growth factor beta

UC ulcerative colitis

23

1 Chapter 1 Introduction and Literature Review

1.1 Introduction It is increasingly clear that the interaction between host and microbiome profoundly affects

health. There are ten times more bacteria in and on our bodies than the total of our own

cells, with the human intestine containing ~100 trillion bacteria. Our microbial communities

are not passive bystanders, and we are only just beginning to appreciate the influence our

microbial residents have on our health. Until recently interrogation of our microbial

communities using classical microbiology techniques offered a very restricted view of

these communities, only allowing us to see what we can grow in isolation. However, recent

advances in sequencing technologies have greatly facilitated systematic and

comprehensive studies of the microbiome, and elucidate its role in human health and

disease.

1.2 Ankylosing spondylitis Ankylosing spondylitis (AS) is a common and highly heritable autoimmune disease

belonging to a group of arthritides called spondyloarthropathies (SpA) and affects over

22,000 Australians (1). This group of seronegative diseases comprises of AS, psoriatic

arthritis (PsA), reactive arthritis (ReA), Juvenile onset SpA, and Undifferentiated SpA

associated with inflammatory bowel diseases including Crohn’s disease (CD) and

ulcerative colitis (UC) (2). All members of SpA are characterised by a strong association

with the human leukocyte antigen (HLA) allele HLA-B*27, combined with the absence of

an association with rheumatoid factor or anti-nuclear antibodies. Additionally, they all show

a certain degree of enthesitis, which is inflammation of the site of insertion of ligaments,

tendons or joint capsules to bone, such as the Achilles tendon (3).

AS is a life long condition with first symptoms appearing at a young age. AS is

characterised by excessive bone formation (ankylosis) that gradually bridges the gap

between joints, eventually fusing and causing stiffness, inflexibility, pain, significant

morbidity and increased mortality. The mechanism that triggers AS in a genetically

susceptible person is poorly understood. There is a close relationship between the gut and

24

SpA, exemplified in reactive arthritis patients, where a typically self-limiting arthropathy

follows either gastrointestinal infection with Campylobacter, Salmonella, Shigella or

Yersinia, or urogenital infection with Chlamydia. AS has a global distribution, with no

outbreaks or clustering being reported. The gene identified as having the strongest

association with AS development is the major histocompatibility complex (MHC) gene

HLA-B27. However, only 1-5% of HLA-B27+ individuals go on to develop AS. Monozygotic

twin studies suggest that the trigger for AS in a genetically susceptible person is a

ubiquitous environmental factor that is most likely widely distributed (4, 5).

1.2.1 The gastrointestinal tract and AS

The co-existence of AS and intestinal inflammation has been known for some time (6),

with between 5-10% of AS patients developing clinically diagnosed inflammatory bowel

disease (IBD). Up to 70% of AS patients have either clinical or subclinical gut disease,

which suggests that intestinal inflammation is important in disease (7, 8). The human

gastrointestinal tract is not a complete barrier and can become permeable to especially

when there is a loss or impairment of modulation of tight junctions (9). Permeability of the

gut is increased in conditions including IBD and celiac disease (10), and also in AS (11,

12). The increased gut permeability in AS has been mostly attributed to the use of non-

steroidal anti-inflammatories (NSAID) (6) however, first-degree relatives of AS patients

have also demonstrated increased gut permeability when compared with unrelated healthy

controls (6, 11-13). This is consistent with the hypothesis that increased ‘leakiness’ of the

gut may increase the microbial dissemination and drive inflammation in the disease.

In reactive arthritis (ReA), a member of the SpA family, inflammatory arthritis develops

following urogenital infection with Chlamydia trachomatis or gastrointestinal infection with

Campylobacter, Salmonella, Shigella or Yersinia (14). Such cause and effect relationships

are not established for other SpA. Studies in the B27 transgenic rat model of AS, supports

the idea of a ubiquitous environmental trigger. The B27 transgenic rat when exposed to

normal commensal bacteria, developed disease while the rats that were maintained in a

germ-free environment remained disease free (15). Microbial involvement, especially

arising from the gut, has been associated in both diseases however in AS, no definitive

link has been established (16, 17).

25

1.3 Genetics and overlap between AS and gut disease Strong genetic overlap exists between AS and IBD (Table 1.1), particularly CD, with the

two conditions commonly occurring together in families (18, 19). Danoy et al., 2010

examined genes known to be associated with IBD in a large AS cohort (20) identifying new

loci and genes. Of particular interest were genes involved in the IL-23 pathway, such as

STAT3, IL23R, and IL12B (which encodes IL-12p40, the share subunit of IL-23 and IL-12

(20-22). How these genetic lesions influence gut homeostasis and function still remains

unclear although, many of the genes associated with disease are associated with microbial

handling, processing and response such as IL23R, NOD2, NOS2 and CARD9. Genetic

studies in CD, one of the two common forms of IBD, shows a marked over-representation

of genes including NOD2 and TNFRSF18, that correlated with an increased risk of

developing mycobacterial diseases, including leprosy and tuberculosis, (23). Seven of the

eight known genetic variants, which include SLC11A1, VDR and LGALS9, are associated

with increased risk of developing leprosy and were also found to increase and individual’s

risk of CD (23), A high proportion of genes associated with AS are genes involved in

mucosal immunity (24). Variants in transcription factors RUNX3, EOMES, and TBX21, as

well as the cytokine receptors IL7R and IL23R (24), are key regulators of the differentiation

and activation of innate lymphoid cells, which are themselves critical components of

mucosal immune defences.

Major differences also exist between the genetics of IBD and AS, with AS showing no

association to date with NOD2/CARD15 or the autophagy gene ATG16L1, which are major

susceptibility factors in IBD, whereas IBD shows no association with HLA-B or ERAP1

which are the strongest AS-susceptibility genes (18, 24).

26

Table 1.1 Genes associated with ankylosing spondylitis (AS) and their association with inflammatory bowel disease (IBD).

Reported gene SNP Locus AS Risk/Non-

risk Alleles

IBD association

RUNX3 rs6600247 1p36 C/T - IL23R rs11209026 1p31 G/A Concordant

rs12141575 1p31 A/G GPR25-KIF21B rs41299637 1q32 T/G Concordant

PTGER4 rs12186979 5p13 C/T Concordant ERAP1-ERAP2-LNPEP rs30187 5q15 T/C -

rs2910686 5q15 G/A Concordant ERAP1-ERAP2 rs10045403 5q15 A/G -

IL12B rs6871626 5q33 T/G Concordant rs6556416 5q33 G/T Concordant

CARD9 rs1128905 9q34 C/T Concordant LTBR-TNFRSF1A rs1860545 12p13 C/T -

rs7954567 12p13 A/G - NPEPPS-TBKBP1-

TBX21 rs9901869 17q21 T/C -

21q22 rs2836883 21q22 G/A Concordant IL6R rs4129267 1q21 G/A -

FCGR2A rs1801274 1q23 T/C Concordant rs2039415 1q23 C/T -

UBE2E3 rs12615545 2q31 C/T - GPR35 rs4676410 2q37 A/G - NKX2-3 rs11190133 10q24 C/T Concordant ZMIZ1 rs1250550 10q22 C/A Concordant SH2B3 rs11065898 12q24 A/G - GPR65 rs11624293 14q31 C/T Concordant

IL27-SULT1A1 rs7282490 16p11 A/G Concordant rs35448675 16p11 A/G -

TYK2 rs35164067 19p13 G/A Concordant rs6511701 19p13 A/C -

ICOSLG rs7282490 21q22 G/A Concordant EOMES rs13093489 3p24 A/C -

IL7R rs11742270 5p13 G/A - UBE2L3 rs2283790 22q11 C/T - BACH2 rs17765610 6q15 G/A Discordant

rs639575 6q15 A/T - NOS2 rs2531875 17q11 G/T -

rs2297518 17q11 A/G - HHAT rs12758027 1q32 C/T -

IL1R2-IL1R1 rs4851529 2q11 G/A Concordant

27

1.4 The gut, barrier regulation and intestinal epithelial cells Homeostasis of the normal microbial flora in the gut is essential for intestinal health. The

gastrointestinal tract is heavily populated with microbes and is the primary site for

interaction between these microorganisms and the immune system (25, 26). Furthermore,

microbes found in the gut help to shape host immune systems from an early age. The

incomplete development of the immune system in neonates and under germ-free

conditions tells us that microbiota play a role in sculpt the host immune system (27, 28).

Maintenance of intestinal and microbial homeostasis is being increasingly recognised as

playing a pivotal role in overall health (29), and dysregulation of either gut or microbial

homeostasis may play a role in autoimmunity. Physiological processes in the host that

maintain gut homeostasis and respond to perturbance in the gut microenvironment involve

both the adaptive and innate immune system, and the barrier function of the intestines

themselves.

1.4.1 The physical barrier The human gastrointestinal tract is not a complete barrier. It is comprised of a single layer

of intestinal epithelial cells (IECs), which form a physical barrier separating the intestinal

lumen from the lamina propria (Figure 1.1). IECs secrete soluble factors that are crucial to

intestinal homeostasis, such as mucins, anti-microbial peptides (AMP), including

lysozymes, defensins, cathelicidins, lipocalins and C-type lectins such as RegIIIγ (30-32).

Release of these molecules into luminal crypts is thought to prevent microbial invasion into

the crypt microenvironment as well as limit bacteria-epithelial cell contact (31, 33). CD

patients have pronounced decreases in the human α-defensins DEFA5 and DEFA6 in the

ileum compared to healthy controls, resulting in altered mucosal function and overgrowth

or dysregulation of commensal microbial flora (33, 34). Furthermore, depletion of the

mucin layer leads to an IBD-like phenotype and endoplasmic reticulum stress, potentially

driving IL-23 production (35). IL-23 excess alone is sufficient to induce spondyloarthritis in

rs2192752 2q11 C/A - 2p15 rs6759298 2p15 C/G Concordant

GPR37 rs2402752 7q31 T/C - HLA-B27 rs4349859 6p21 A/G - ANTRX2 rs4333130 4q21 A/G -

28

mice (36), and genetic evidence in particular indicates that the cytokine plays a key role in

development of spondyloarthritis in humans.

Figure 1.1 The physical barrier. Separating the intestinal lumen and its inhabiting commensal

bacteria from the underlying lamina propria is a single layer of intestinal epithelial cells (IECs).

These IECs are stitched together, creating a tight junction and regulating the paracellular flux. IECs

also secrete soluble factors that are crucial to intestinal homeostasis, such as mucins and anti-

microbial peptides (AMPs), including lysozymes, IgA, defensins, and C-type lectins such as

RegIIIγ. Release of these molecules into luminal crypts is thought to prevent microbial invasion into

the crypt microenvironment as well as limit bacteria-epithelial cell contact. Toll-like receptors

(TLRs) are also expressed on IECs to sense a breach of barrier or bacterial invasion. Underneath

the IECs, the lamina propria contains T cells, bacteria-sampling dendritic cells (DCs), and

macrophages. Figure from: Costello et al. Arthritis Research &

Therapy 2013 15:214 doi:10.1186/ar4228

1.4.1.1 Permeability and Disease The dynamic crosstalk between IECs, microbes and the local immune cells is fundamental

to intestinal homeostasis but is also suggested to play a part in disease pathogenesis (37,

38). There are a multitude of factors capable of altering the permeability of epithelial

junctions. Once considered fixed and unchanging, epithelial tight junctions are remarkably

dynamic and responsive component of the intestines. The tight junctions constantly open

and close in response to a number of stimuli including diet, neural signals, mast cell

29

productions as well as a variety of cellular pathways, all of which can be hijacked by

microbial as well as viral pathogens (39-42).

In IBD and celiac disease, epithelial tight junctions are dysregulated causing increased

permeability between IECs and resulting in a ‘leaky gut’ (10, 43). While the causes and

ramifications of a ‘leaky gut’ in the disease setting are still under investigation, It has been

suggested that an increase in intestinal permeability may lead to bacteria, either

pathogens or normal resident bacteria, entering the lumen and triggering an inflammatory

immune response (43, 44). Intestinal permeability in AS patients has always been

considered a side effect due to the long term use of NSAIDs (6). However, studies looking

at first-degree relatives of AS patients showed they also have increased gut permeability,

suggesting that there may be an underlying genetic process operating in the gut (11, 45).

1.4.1.2 Innate immunity Dendritic cells (DC) densely populate the intestinal lamina propria and form a widespread

microbial sensing network. DC recognise a broad repertoire of bacteria, sensing with

receptors such as toll-like receptors (TLRs) and monitoring the bacteria on the mucosal

surface (46). Intestinal DC orchestrate production of intestinal specific IgA to limit bacterial

contact with the intestinal epithelial cell surface (47). Activated DC can secrete a number

of cytokines and chemokines, including IL-23 and IL-6, that are involved in inflammation

and migration of DC (48).

Macrophages patrol the gastrointestinal systems in high numbers. They frequently come in

contact with ‘stray’ bacteria, including commensals that have breached the epithelial cell

barrier. Circulating macrophages phagocytose and rapidly kill such bacteria using

mechanisms that include production of antimicrobial proteins and reactive oxygen species

(49). However, intestinal macrophages have several unique characteristics including the

expression of the anti-inflammatory cytokine IL-10, both constitutively and after bacterial

stimulation (50, 51). This makes intestinal macrophages non-inflammatory cells that still

have the capacity to phagocytose. The importance of this pathway is highlighted by the

fact that loss-of-function mutations in IL10R lead to early onset IBD (52). Intestinal

macrophages also play a role in the restoration of the physical integrity of the epithelial cell

barrier following injury or bacterial insult (9). Re-establishing this barrier post injury is

30

imperative to avoid bacterial penetration and sepsis in such a microbe-laden environment

(53).

1.4.2 Adaptive Immunity

1.4.2.1 IL-23-responsive cells IL-23 is a key cytokine in the development of IL-17 and IL-22-secreting cells. IL-23 signals

through a receptor consisting of the specific IL-23 receptor (IL-23R) subunit and IL-12Rβ1,

also shared with IL-12R (54). Loss of function polymorphisms in IL23R are associated with

protection from AS (55), psoriasis (56) and IBD (57), and many other genes in the IL-23

pathway are associated with these diseases. Under physiological conditions, IL-23, IL-17

and IL-22 producing cells are enriched in gut mucosa and IL-17 and IL-22 are known to be

important regulators of intestinal ‘health’. IL-17 plays important roles in intestinal

homoeostasis in several ways, including maintenance of epithelial barrier tight junctions

(58) and induction of anti-microbial proteins such as β-defensins, S100 proteins and REG

proteins. IL-22 induces secretion of anti-microbial peptides (59). In the gut, innate-like

immune cells act as sentinels, responding very rapidly to alterations to the microbial

composition of the gut with rapid secretion of IL-17. Key among these sentinels are γδ T

cells, Natural killer T (NKT) cells, mucosal-associated invariant T (MAIT) cells and

lymphoid tissue inducer (LTi)-like cells (Figure 1.2).

31

Figure 1.2 The immune barrier. Dendritic cells (DCs) and macrophages patrol the gastrointestinal

tract in high numbers. They densely populate the intestinal lamina propria and form a widespread

microbe-sensing network. Activated DCs can secrete a number of cytokines and chemokines,

including interleukin-23 (IL-23), IL-6, and IL-1, activating IL-23-responsive cells. IL-23-, IL-17-, and

IL-22-producing cells are enriched in gut mucosa, and IL-17 and IL-22 are known to be important

regulators of intestinal 'health'. IL-17 plays important roles in intestinal homoeostasis in several

ways, including maintenance of epithelial barrier tight junctions. LTi, lymphoid tissue inducer;

MAIT, mucosa-associated invariant T; NKT, natural killer T; T reg, regulatory T; TNF, tumour

necrosis factor. Figure from: Costello et al. Arthritis Research &

Therapy 2013 15:214 doi:10.1186/ar4228

1.4.2.2 γδ T cells

γδ T cells are found in large numbers at epithelial surfaces such as the gut and skin,

where they can account for up to 50% of T cells. γδ T cells not only bear an antigen-

specific T cell receptor (TCR), but also have many properties of cells of the innate immune

system, including expression of the major innate immunity receptors and TLRs (60) and

dectin-1(61), which recognises microbial β-glucans. Expression of these receptors

supports a role for γδ T cells in early responses to microbes. Of further relevance, we and

others, have recently confirmed that CARD9, part of the dectin-1 response pathway, is a

susceptibility gene for AS and for IBD (62). γδ T cells are potent producers of inflammatory

32

cytokines such as IFN-γ, TNF-α and IL-17 (63, 64). They are pathogenic in the collagen-

induced arthritis model (65) and mouse models of colitis (66), and IL-17-secreting γδ T

cells are expanded in patients with AS (67). Intraepithelial γδ T cells also play an important

role in modulating intestinal epithelial growth through secretion of fibroblast growth factor

(68). Alterations to γδ T cell numbers or functions may, therefore, have profound effects on

intestinal health.

1.4.2.3 Natural killer T cells Natural killer T (NKT) cells are characterised by expression of an invariant T cell receptor,

Vα24Jα18 in humans and the orthologous Vα14Jα18 in mice. NKT cells recognise

glycolipd structures presented to them by the non-classical antigen presenting molecule

CD1d. Like γδ T cells, NKT cells are rapid responders to antigenic stimuli and are capable

of producing a range of immunoregulatory cytokines (69-72). Within the gut, NKT cells are

protective in Th1-mediated models of inflammatory bowel disease but pathogenic in Th2

models (73, 74). Recently, it has been shown that microbial stimulation of NKT cells in the

gut of mice affects NKT cell phenotypic and functional maturation (75). Given that NKT

cells have protective roles in models of arthritis (76) and spondyloarthropathy (77, 78) their

functional maturation in the gut provides evidence for a role for mucosal T cell priming in

inflammatory joint disease.

1.4.2.4 Mucosal-associated invariant T cells Mucosal-associated invariant T (MAIT) cells are a population of innate-like T cells that are

abundant in human gut, liver and blood and secrete a range of inflammatory cytokines

such as IL-17 and IFN-γ in response to antigenic stimulation (79-81). Like NKT cells, MAIT

cells bear an invariant T cell receptor (Vα7.2 in humans) that recognises antigen presented

by the non-classical MHC-like molecule, MR1 (82). In blood, MAIT cells display a memory

phenotype (81) and express the transcription factor ZBTB16 (83), allowing them to rapidly

secrete cytokine in response to antigenic stimuli. Furthermore, they express high levels of

IL-23R (84). MAIT cells react to a wide range of microbial stimuli, including Gram positive

and Gram negative bacteria as well as yeasts (79, 80). While the precise role of MAIT cells

in maintenance of the mucosal barrier remains unclear, the rapid postnatal expansion of

MAIT cells and their acquisition of phenotypic markers of memory likely represents

interaction with developing commensal microflora (85).

33

1.4.2.5 Lymphoid tissue inducer- like cells Lymphoid tissue inducer (LTi)-like cells are found in spleen, lymph node and gut lamina

propria. LTi-like cells constitutively express hallmarks of IL-17-secreting cells including IL-

23R, RORγt, AHR and CCR6 (86, 87). LTi cells in the intestine represent an increasingly

interesting and heterogeneous population of innate lymphoid cells (ILC) with or without

CD4 expression (88). A common characteristic of ILCs is the expression RORγt, a

transcription factor important for the development of these cells. ILC have been linked to

gut inflammation through colitis models where IL-23 responsive ILC secrete IL-17, IL-22

and IFN-γ and to promote intestinal inflammation (89). RORγt, ILCs are abundant in the

intestinal lamina propria and produce IL-17 and/or IL-22 in order to preserve mucosal

integrity against extracellular pathogens. These NKp46+ ILCs have been found to be

critical for host defence against pathogens such as Citrobacter rodentium infection through

secretion of IL-22 (87, 90).

1.4.3 Cytokines and gut homeostasis

1.4.3.1 IL-17 IL-17 is primarily described as a T cell-secreted cytokine however much of the IL-17

released during an inflammatory response is in fact produced by innate immune cells in

response to injury, stress or to pathogens (91). IL-17 has been shown to be produced 4-

8hrs after a microbial infection and is secreted by T cells including γδ T cells (63, 92, 93).

Innate IL-17 producing cells have been found to reside mainly in the skin and mucosal

tissues, especially the gastrointestinal tract, and act as a first line of defence. They also

promote mobilisation of neutrophils and cytokines that are produced by epithelial cells for

protective immunity as well as controlling physiological process such as epithelial tight

junctions and angiogenesis. The hallmark of the innate immune response is its rapid

reaction to pathogens by secreting factors that recruit large numbers of neutrophils

through increasing the activity of IL-1, IL-6 and tumour necrosis factor (TNF), which

promote tissue infiltration, which is crucial for effective and rapid control of bacterial and

fungal pathogens (91).

1.4.3.2 IL-22 IL-22 has been shown to play a critical role in host defence, tissue homeostasis and

inflammation and the IL-22-IL-22R pathway contributes to regulating immunity,

34

inflammation and tissue repair. IL-22 is expressed by innate and adaptive cells and seems

to act almost exclusively on non-haematopoietic cells, with basal IL-22R expression in the

skin, pancreas, intestine, liver, lung and kidney (94). First described in the skin by

Dumoutier et al., 2000, who showed that the stimulation of keratinocytes with IL-22

resulted in marked induction of genes encoding several proteins involved in anti-microbial

host defences including S100A7, S100A8, S100A9, β-defensin-2 and β-defensin-3 (95-

97). Further studies have shown that the induction of two intestinal antimicrobial peptides

that regulate gut homeostasis, RegIIIβ and RegIIIγ are also dependent on the production

of IL-22 (98, 99). IL-22 can act synergistically or additively with IL-17A, IL-17F or TNF to

promote the expression of many of the genes that encode molecules involved in host

defence in the skin, airway or intestine. This demonstrates the functional importance of IL-

22 in promoting barrier immunity and mucosal integrity, particularly against Gram negative

pathogens, where IL-22 is critical for limiting bacterial replication and dissemination,

possibly in party by inducing the expression of antimicrobial peptides from epithelial cells

at these barrier surfaces (98, 99). Containment of normal flora is another important arm of

homeostasis. IL-22 producing innate lymphoid cells (ILCs) found throughout the gut play a

critical role in the physical containment of resident microbes (100). Loss of epithelial

integrity leads to dissemination of intestinal commensals and systemic inflammation.

1.4.4 Interaction between intestinal microbes and the immune system There is strong evidence from murine studies to indicate that interaction between the gut

microbiome and the host determines the overall level of activation of the immune cells

producing these cytokines. Segmented filamentous bacteria (SFB) are commensal

bacteria that induce IL-17 secretion. Mice lacking SFB have low levels of intestinal IL-17

and are susceptible to infection with pathogenic Citrobacter spp. (27, 101) Restoration of

SFB in these mice increased the number of gut-resident IL-17-producing cells and

enhanced resistance to infection through induction of specific epithelia cell-specific genes,

as well as inflammatory response host genes, which were found to be up regulated by the

bacteria leading to an inflammatory environment (27). Lachnospiraceae,

Ruminococcaceae and Prevotellaceace are families of bacteria that have been observed

in IBD gut microbiomes and are strongly associated with colitis and CD (102-104), with

Prevotellaceace especially known to elicit a strong inflammatory response in the gut (104).

The demonstration that bacteria can directly influence the host response highlights the

effect commensal bacteria have on the immune response and their role in inflammation.

Conversely, members of the gut microbiota such as Clostridium have been found to play a

35

protective role by inducing T regulatory cells (Tregs) in the colonic mucosa. Tregs are an

important counter balance in the immune system and promote homeostasis. It was found

that a specific mix of cluster IV and XIVa of Clostridium was sufficient to promote Treg

accumulation in the colon (105).

1.5 Interrogating the human gut microbiome Together, the multi-factorial components of the immune system shape the microbial

population that inhabits the gut and, to an extent, vice versa, with each side pushing to

establish a stable state. Understanding the yin and yang of this relationship is at the heart

of current microbiome research.

Microbiome refers to the totality of microbes, their genetic elements and environmental

interaction in a defined environment (106). The aim of metagenomic studies is to profile

the microbes within the community, identify functions and pathways of the communities

and how the communities compare. Projects such as the Human Microbiome Project run

by the National Institute of Health, and the Metagenomics of the Human Intestinal Tract

(MetaHit) consortium, aim to determine what constitutes a ‘normal’ or healthy microbiome.

Research in this field has recently been greatly accelerated by the development of

methods for microbial profiling without the requirement to culture organisms, using high-

throughput sequencing methods.

1.5.1 Microbial amplicon sequencing Culture independent molecular investigations using the 16S rRNA have become the

foundation for identifying and characterising microbes. With the advent of next generation

sequencing, 16S rRNA profiling has taken a back seat to newer approaches such as

metagenomics. Metagenomics has been favoured as it facilitates a more comprehensive

view of the microbial community by providing the communities overall functional potential

as opposed to simply a list of microbial inhabitants. However, 16S rRNA marker genes

studies are making a return to more widespread use and are becoming the standard

method for community profiling. This is due to significant advances in sequencing

methodology and analysis programs (107).

36

1.5.2 The 16S rRNA

16S ribosomal RNA (rRNA) is a section of prokaryotic DNA found in all bacteria and

Archaea. 16S rRNA sequences are used to differentiate between organisms across all

major phyla of bacteria and to classify strains down to species level (108). 16S rRNA

sequencing has dramatically changed our understanding of phylogeny and microbial

diversity because it provides an unbiased assessment of the microbiome, and is not

restricted by the ability to culture the bacteria present. This commonly used ribosomal

RNA gene sequence analysis is a powerful method for identifying and quantifying

members of microbial communities (109). The 16S rRNA gene can be compared not only

among all bacteria but also with the 16S rRNA gene of Archaebacteria and the 18S rRNA

gene of Eukaryotes (109). The 16S rRNA sequence is ~1,550 bp long, highly conserved

and is composed of both variable and conserved regions (Figure 1.3). The gene is large

enough and has sufficient interspecific polymorphisms as to provide distinguishing and

statistically valid measurements with the variable portions of the 16S rRNA is used for

comparative taxonomy (110). There are nine diverse, species-specific variable regions

denoted as V1-9 (Figure 1.3), which can be used to infer phylogenetic relationships. This

gene also allows the analysis of microbial communities by way of high-throughput

sequencing. The 16S rRNA gene sequences allows for differentiation between organisms

across all major phyla of bacteria and has the ability to classify strains at multiple levels

including the species and subspecies level. Similar 16S sequences are then clustered

together to form operational taxonomy units (OTUs). Sequences can be clustered two

ways; the first is open reference where sequences are clustered by similarity with respect

to the other sequences clustered at the same time. The second is closed reference where

sequences are clustered at a certain threshold of sequence similarity, such as 97% (111)

against a specific reference database. Clustering at a maximum sequence identity

threshold of 97% has been widely adopted as a measure to avoid over estimating diversity

by inadvertently mapping sequence error (112). The sequence clusters are then mapped

to a database, such as Greengenes or ARB/Silva, and assigned taxonomy based on

sequence similarity (113). One drawback of the 16S rRNA technique cannot be used to

compare strains for epidemiological purposes or to detect a specific strain virulence

factors. The 16S rRNA gene does not contain enough variation for strain resolution and

does not encode virulence factors. However variations on this gene analysis using variable

regions, can infer functionality including metabolism (114, 115).

37

Figure 1.3 Cartoon depiction of the 16S rRNA gene. Modified 16S rRNA gene highlighting the

variable regions of the gene commonly used for taxonomy. Circled in red are the two constant

commonly used and circled in yellow is the V3-V5 region commonly used for amplicon sequencing

in the gut.

1.5.3 Metagenomic sequencing

Further improvements in sequencing technologies have reduced the need for targeted

studies such as 16S sequencing, and enabled high throughput shotgun sequencing. This

type of sequencing randomly samples all genes present in a habitat rather than just 16S

rRNA. Shotgun sequencing provides more information about the microbiome than just the

16S characterisation, which is of particular benefit in determining the functional capacity of

the microbial community, rather than just its phylogeny. However shotgun sequencing is

more complex to analyse, including the challenge of distinguishing between host and

bacterial genomic material, and requires far more sequencing to be performed. Therefore

most studies to date have relied on 16S rRNA sequencing approaches.

1.5.4 Advances in tools for metagenomic studies Several sequencing technologies have been developed for human genetic studies and

have since been adapted to metagenomics. Roche’s 454 sequencing platform uses large-

scale parallel pyrosequencing to produce reads of between 450bp to 1000bp length. Read

38

lengths produced by the 454 platform are well suited to 16S rRNA amplicon metagenomics

studies as well as shotgun sequencing as they are easily aligned to reference bacterial

genetic datasets. Illumina’s sequencing platforms, the HiSeq2000 and MiSeq, produce

shorter reads than the Roche 454 however, in the last 2yrs or so, the MiSeq has over

taken from Roche as the standard platform for 16S rRNA amplicon sequencing. The HiSeq

was primarily designed for human genomic sequencing and current chemistry produces

100bp paired-end reads. Illumina sequencing is best suited for shotgun sequencing or

indexed amplicon sequencing of multiple samples. The MiSeq however has rapidly

become the go to platform for amplicon sequencing due to its reduced shorter sequencing

runs, longer amplicon reads

This is a rapidly developing field. Advances in chemistry are predicted to increase both

read-lengths and output for both of these platforms over the next 12-24 months,

particularly for the Illumina platforms that are less mature than the Roche 454. New

platforms coming to the market are likely to have a major impact on metagenomics study

design. For example, the Pacific Biosystems sequencing technology provides reads of

~3000 bases, which will make it particularly suited to this field, despite its lower overall

output (~450Mb – 1GB of sequence per run). Life Technologies Ion Torrent technology is

reported to produce up to 400 base reads and up to 1Gb of sequence per run. The relative

position of these competing technologies in metagenomics has yet to be established.

1.6 Dynamics of the gut microbiome There are more microbes on the human body than cells, and about 29% of all these reside

in the gut. The human gut microbiome contains a dynamic and vast array of microbes that

are essential to health as well as provide important metabolic capabilities. Until recently

studying this community has been difficult and generally limited to classical phenotypic

techniques (109, 116). The application of molecular techniques, particularly sequencing,

has shown the remarkable amount of diversity in the human gut, exceeding all previous

held beliefs. Originally the dominant species was thought to be E. coli, pathogen detection

was the primary aim with normal flora disregarded and considered inconsequential.

Recent studies using sequencing as the investigative tool have shown that members of the

phyla Firmicutes, Actinobacteria and Bacteroidetes are abundant in healthy individuals

(114, 116). Given that the gut plays a major role in metabolism, pathogen resistance, as

well as the body’s immune response, it is important to ascertain what role the human gut

39

microbiome plays with respect to illness and disease; and what influence these

communities have on disease pathogenesis. To explore this, it must be possible to

compare microbial communities within and across individuals in different states of disease

and health (116).

Understanding how the intestinal microbiome is assembled and maintained is becoming

increasingly relevant not only to general healthy, but also potentially for the treatment of

complex chronic diseases. In recent years only a handful of studies have examined the

composition, diversity and function of the human gut microbiome.

Studies comparing the gut microbiota of lean and obese twins have shed light on the

importance of intestinal microbes and how a change in microbiome composition can affect

food metabolism in the gut (114, 117, 118). Turnbaugh et al showed that even with a

similar genetic make-up, obese twins had substantial differences in the composition and

diversity of their gut flora with a dominance of Gram Positive bacteria from the phylum

Firmicutes, compared with discordant or lean twins (114). This shift in gut flora

composition altered how food was broken down and metabolised in the gut, leading to

increased body mass index, adiposity and obesity (114, 117, 119). This demonstrates that

shifts in the intestinal microbiome have functional consequences on the health of the

individual (120). To further interrogate structure and function of the gut microbiome, and

the consequences of shifts and alterations, faecal samples from four twin pairs discordant

for obesity were transplanted into germ free mice (121). The authors found that the

Firmicute phylum dominated microbial communities from the obese twin lead to obesity in

the germ-free mice. The Bacteroides dominated microbial communities from the lean mice

kept the germ-free mice lean. When the mice were co-housed, the invasion of Bacteroides

from the lean mice lead to a decrease in adiposity and body mass in obese mice and

transformed their microbiota’s metabolic profile to a lean-like state (121).

Maslowski, Vieira et al. 2009 suggests that not only can diet induce changes in the gut

microbial community, but these associated changes may also be underpinning

inflammatory disease. The authors suggest that a lower intake of fibre from complex plant

polysaccharides adversely affects the makeup of the intestinal microbiota, which leads to

less production of immunomodulatory products in particular short-chain fatty acids (SCFA).

These SCFA are produced by the phylum Bacteroidetes, so a shift in the flora from

predominantly Bacteroidetes to Firmicutes due to a more western diet with less fibre,

40

reduces the amount of SCFA secreted. The effect SCFA has on the immune response

was investigated in a mouse strain deficient in single G protein–coupled receptor, GPR43

(122). Mice lacking GPR43 failed to suppress inflammation in models of colitis, arthritis

and asthma, as did germ-free wild-type mice also lacking SCFA. Wild type mice raised in

non-germ-free conditions were able to suppress inflammation. This suggests that diet-

induced reduction in faecal SCFA leads to a reduced ability to suppress a variety of

inflammatory conditions.

1.7 The normal microbiome

These studies have shown that shifts in microbiome can change metabolism; however,

they do not provide details of what constitutes a ‘normal’ microbiome or how it behaves.

Two large studies interrogating and cataloguing microbiomes from various regions of the

body in health and disease have been undertaken by the National Institute of Health

Human Microbiome Project in the USA and the European MetaHit project (123-125). The

Human Microbiome Project was established for the purpose of identifying and cataloguing

the human microbiome by sequencing samples from 242 individuals (129 males, 113

females) across five major body sites, including the mouth, skin, stool and vaginal tract,

multiple times with the aim of interrogating within subject variation, between subject

variation as well as microbiome variation over time. A total of 4,788 specimens were

sequenced with females sampled from 18 different habitats and males from 15 habitats

(126). Diversity and abundance of each habitat’s signature microbes was found to vary

greatly amongst the healthy subjects, with strong niche specialisation found both within

and between individuals, demonstrating the dynamic nature of the microbiome (126-128).

Metagenomic carriage of metabolic pathways was found to be stable amongst individuals,

despite the variation seen within community structure. When clinical metadata was

present, ethnic background proved to be one of the strongest predictors of both pathways

and microbes. This suggests that environmental factors such as diet, location/region and

host genetics could play a role in sculpting the microbiome. These studies further define

the range of structural and functional configurations that are present in normal microbial

communities in a healthy population (123, 126).

41

1.7.1 Enterotypes

The European MetaHit consortium combined published data sets from across the world

and added 22 newly sequenced faecal metagenomes from four different European

countries. They identified three robust clusters, termed ‘enterotypes’, that were not nation

or continent specific. These enterotypes characterised the microbial phylogenetic variation

as well as the function variation of the clusters at gene and functional class levels (123).

Each enterotype had a dominant bacterial genus: Enterotype 1 dominated by the genus

Bacteroides; enterotype 2 by Prevotella; and enterotype 3 by Ruminococcus. The three

enterotypes were also shown to be functionally different. For example enterotype 2, which

is Prevotella dominant, also contains Desulfovibrio, which may act in synergy with

Prevotella to degrade mucin glycoproteins present in the mucosal layer of the gut. It may

be that different enterotypes may be associated with diseases such as obesity and IBD,

rather than necessarily specific bacterial species, given their differing functional capacities.

However, recent work suggests that the boundaries between the enterotypes are more

vague than initially described. Multiple datasets including the Human Microbiome Project

16S rRNA gene sequence data and metagenomes with similar published data were

interrogated for the existence of enterotypes across multiple body sites. However, rather

than forming enterotypes, many of the samples examined fell into sliding scale based on

taxonomic abundances of bacteria such as Bacteroides, rather than discrete enterotypes

(129). Moreover, metatanalysis demonstrated that many of the methods used in the

analysis of enterotypes, such as clustering approaches and distance metrics affected the

likelihood of identifying enterotypes in particular body habitats. This suggests enterotypes

are not discrete but are continuous and vary widely amongst individuals (130).

1.7.2 Homeostasis

Microbes and humans have evolved over time to live in symbiosis with each other. Many

of these diverse communities carry out specific tasks that benefit the host, as well as the

microorganisms A balancing act between war and peace exists within the gut between

resident microbial members of the gut and potential pathogenic and non-pathogenic. It is

well documented that changes in the microbial composition in the gut by a pathogen, elicit

an innate immune responses (131). However, it is still unknown if a shift or dysbiosis in the

gut flora has the capacity to illicit the same response as an infection as the gut tries to

defend its self and restore homeostasis (132). Homeostasis of the normal flora in the gut

42

microbiome is essential, and both the bacteria that inhabit the gut and the human immune

system have developed strategies to regulate and protect it. Bacteria in the intestine have

several techniques including competitive exclusion, biosurfactant production and

modulation of tight junctions (9). The human body has a number of physiological

processes that respond to a disruption in gut homeostasis or bacterial insult. If a bacterial

insult is detected by antigen presenting cells (APC), such as a dendritic cell, naïve T cells

are then activated and recruited to the area. IL-17 secreting Th17 cells are sequestered to

induce epithelial cells to recruit neutrophils to the area to neutralise the disruption and in

concert, IL-22 is secreted to promote epithelial repair and antibacterial proteins (133). The

epithelial cells lining the intestine can also respond to a disturbance or insult and attempt

to restore homeostasis by secreting a range of soluble factors such as mucins and AMPs

that aim to prevent the invasion of microbes into the intestinal crypt (30-32, 134).

1.8 Influence of underlying host genetics on the intestinal microbiome

A key issue in microbiome research is determining whether host genetics directly

influences changes in microbiome composition or indirectly via the immune system. The

extent that underlying host genetics influences intestinal microbial community composition

in humans is unclear. Host gene deletions have been shown in animal models to cause

shifts in microbiota composition (135). Moreover, a recent quantitative trait locus mapping

study linked specific genetic polymorphisms with microbial abundances (136). These

studies highlight the importance of underlying host genetics in community composition in

animals. The role of host genetics is of interest in diseases such as IBD, where many of

the genes associated with disease are involved in handling, processing and response to

microbes such as IL23R, NOD2, NOS2 and CARD9. Genetic studies in CD, one of the two

common forms of IBD, shows a marked over-representation of genes, notably NOD2 and

TNFRSF18, which are associated with an increased risk of developing mycobacterial

diseases, including leprosy and tuberculosis (23). There is significant overlap between the

risk of developing leprosy and CD, with seven of the eight known genetic variants

associated with an increased risk of leprosy, including SLC11A1, VDR and LGALS9, also

associated with an increase the risk of CD; compared with only two being associated with

risk of any other immune-mediate disease. Combined with dramatic shifts in the gut

microbiota composition associated with IBD (23, 135), this strongly indicates a potential

link between host genetics and the microbiome in IBD patients. The genetic influence over

microbiome composition is further highlighted in a study of 416 monozygotic and dizygotic

43

twin pairs investigating the heritability of the gut microbiome in obesity. The authors found

the most heritable taxon was the family Christensenellaceae (137).This particular family of

bacteria is only recently described and was shown to be central to a network of co-

occurring heritable microbes associated with lean body mass index. Animal studies

showed that the introduction of any of the members of the Christensenellaceae family had

the capacity to protected the mice from weight gain (137). Variation and shifts in the gut

microbiome have until recently been associated with diet and the environment or natural

variation. Even though the exact gene-microbe association with the Christensenellaceae

family still remains unknown, this study demonstrates that microbiome composition may in

part be due to host genetics (137).

1.9 The microbiome in immune mediated disease

1.9.1 Ankylosing spondylitis

To date no study has been reported investigating the gut microbiome using sequencing

based methods in AS. One study using denaturing gradient gel electrophoresis to profile

the microbiome using faecal samples found no differences between AS cases and healthy

controls (16). . Many studies largely using antibody tests have suggested an increased

carriage of Klebsiella species in AS patients, however this has not been universally

supported (17). Several specific bacteria have been suggested to have a role in AS

pathogenesis particularly Klebsiella pneumoniae (138, 139) and Bacteroides vulgatus

(140) although current studies using antibody tests have been unable to produce

convincing evidence demonstrating increased infection with any specific triggering agent

influencing AS development. K. pneumoniae and B. vulgatus have been of considerable

focus for several reasons. Increased levels of B. vulgatus have linked to colitis in the HLA‐

B27/β2 microglobulin transgenic rats. When the HLA‐B27/β2 microglobulin transgenic rats

were raised germ free and subsequently colonised with an intestinal bacterial cocktail, the

rats developed more several colitis when increased amounts of B. vulgatus were present

(141). Substantial research effort has been directed towards the possible involvement of K.

pneumoniae. Early work examining the faeces of AS patients during active disease,

reported the presence of K. pneumoniae (142). As well as during active disease, Studies

have also reported elevated levels of serum antibodies against K. pneumoniae in patient

sera (143). Adding to the suspected involvement of K. pneumoniae, the antigenic similarity

44

noted between Klebsiella and HLA‐B27 (144) have been noted. However, recent studies

have been unable to confirm a role of Klebsiella AS pathogenesis.

As well as specific bacteria triggering disease, it has also been speculated that

antimicrobial reactivity can develop during intestinal inflammation. There have been a

number of antibodies described over the years, including anti-Saccharomyces

cerevisiae antibodies (ASCA), anti-I2 antibodies (which are associated with anti-

Pseudomonas activity), perinuclear antineutrophil cytoplasmic antibodies (pANCA), anti-

Escherichia coli outer membrane porin C (anti-OmpC) antibodies, as well as anti-flagellin

antibodies (anti-CBir1), that have shown to have clinical significance in IBD (145, 146).

Given the overlap between IBD and AS, the prevalence of several of the above

antimicrobial antibodies were examined in the serum of 80 AS patients and healthy

controls. While there was no difference in the number of antibody positive results between

AS patients and healthy controls, AS patients were more likely than healthy controls to

have elevated levels of anti-I2 and ASCA antibodies (147). A recent study has shown of

anti-CBir1 antibodies are elevated in patients with AS with IBD compared to patients who

only have AS (148). Moreover, patients with AS only had increased levels of anti-CBir1

antibodies when compared to healthy controls. This data suggests that AS patients have

an increased exposure of the immune system to commensal bacteria, causing the

antibody production. Increased levels of anti-CBir1 antibody in AS patients, compared to

healthy controls, could be accounted for, at least in part, by microscopic gut inflammation

even in the absence of clinical gastrointestinal symptoms (149). While this observation

requires further investigation, the development of anti-CBir1 antibodies in patients with AS

raises the intriguing question of their pathophysiological significance with regards to

commensal bacteria, and more specifically commensal bacteria with flagella.

1.9.2 Rheumatoid arthritis

Rheumatoid arthritis (RA) is the only inflammatory arthritis for which modern metagenomic

studies have been reported. Community profiling studies of RA patients’ gut flora reveal

differences in the composition of RA patient’s gut microbiota compared with healthy

controls, with a lower abundance of Bifidobacterium and Bacteroides bacteria observed in

RA cases (150, 151). However, these studies were undertaken using faecal samples and

not intestinal biopsies, possibly influencing the populations and the proportions observed

(152, 153). It is not just the intestinal microbiome that has an implicated in RA

pathogenesis. There is microbial profiling data that suggests that the oral microbiome,

45

specifically a periodontal infection with the common pathogen Porphyromonas gingivalis,

may be important in RA pathogenesis. P.gingivalis, which has also been identified in

synovial fluids of RA patients (154), has the ability to citrullinate host peptides like

smoking, which subsequently generates auto antigens that drive autoimmunity in RA

(155).

1.9.3 Inflammatory Bowel Disease

Several lines of evidence indicate that the gut microbiome plays an important role in IBD,

including the association of genes involved in mucosal immunity with IBD (such as

CARD15, CARD9, IL23R and ATG16L1), the therapeutic effect of antibiotics on the

condition, and the beneficial effect of faecal stream diversion in CD. Previous studies of

human IBD have been undertaken using standard culture techniques (156) or molecular

analysis (157, 158). These studies noted alterations in intestinal microbiota when

compared to non-IBD patients, a finding recently confirmed using 16S rRNA sequencing of

intestinal biopsies (159). A recent study examined the intestinal microbiome in treatment

naïve new-onset CD, and described microbial signature. 16S community profiling intestinal

and faecal samples showed that the signature comprised of an increased abundance in

bacteria including Enterobacteriaceae, Pasteurellaceae, Veillonellaceae, and

Fusobacteriaceae, and decreased in abundance of Erysipelotrichales, Bacteroidales, and

Clostridiales and correlate strongly with disease status (153).

1.9.4 Psoriasis

The microbiome is thought to play a significant role in psoriasis, another AS-related

condition. It has long been suggested that streptococcal infection, especially throat

infections, may trigger psoriasis in a genetically susceptible individual (160). Recent

studies using 16S rRNA sequencing have found significant differences between the

cutaneous microbiota of psoriasis cases and controls, and in psoriasis involved and control

skin in psoriasis cases, with less Staphylococci and Propionibacteria observed in cases

and in affected skin (161). Again, further studies will be required to determine if there is a

particular microbial profile or specific bacterial species involved in psoriasis. These studies

are thus consistent with the hypothesis that the microbiome contributes to the

aetiopathogenesis of the immune mediated arthritis or seronegative diseases like IBD and

46

psoriasis. However at this point, with the exception of reactive arthritis, there is no

definitive evidence of specific bacterial infections or changes in microbial profile that play a

causative role in these conditions.

47

1.10 The Chicken and the egg Whether the changes noted in these microbiome studies of immune mediated disease are

a consequence of disease or are involved in its development or persistence still remains

unclear. This distinction may prove impossible to dissect in human studies, but there is

considerable evidence in studies in mice to support a role for the microbiome in driving

immune mediated diseases.

In the B27 rat model of AS, rats housed under germfree (GF) conditions did not develop

disease (141) demonstrating that microbes in this model are important disease

penetrance. In contrast, in the New Zealand black (NZB) model of Systemic lupus

erythematosus, mice maintained under GF conditions produced higher levels of

antinuclear antibodies and developed worse disease (162) demonstrating a protective role

for commensal microbes.

Considering the IL-17/IL-22 cytokines, which are of relevance to AS, IBD and psoriasis in

particular, there is strong evidence from murine studies to indicate that interaction between

the gut microbiome and the host determines the overall level of activation of the immune

cells producing these cytokines. Segmented filamentous bacteria (SFB) are a commensal

bacteria that induce IL-17 secretion. Mice lacking SFB have low levels of intestinal IL-17

and are susceptible to infection with pathogenic Citrobacter spp. Restoration of SFB in

these mice increased the number of gut-resident IL-17-producing cells and enhanced

resistance to infection (27). Further, using recolonisation studies, investigators have

recently demonstrated that in neonatal mice, commensal microbes influence iNKT cell

intestinal infiltration and activation, establishing mucosal iNKT cell tolerance to later

environmental exposures (163).

The mechanism by which the microbiome influences IL-17 producing cell activation is still

being determined. Ivanov and colleagues demonstrated that serum amyloid A, produced

in the TI, can induce TH17 differentiation of CD4+ T-lymphocytes (27). It has also been

shown that development of TH17 lymphocytes in the intestine is stimulated by microbiota

induced IL-1b (but not IL-6) production (164). Colonisation with Clostridia species has

been shown to stimulate intestinal TGFb production, in turn increasing IL-10+ CTLA4high

Treg activation (105). Clostridia colonisation of neonatal mice reduced severity of induced

colitis using DSS or oxazolone, and reduced serum IgE levels in adulthood. So it is likely

48

that alterations to the gut microflora, or invasion of the gut by pathogenic bacteria,

influence the balance of IL-17- and IL-22-producing cells, and other immune cells,

influencing susceptibility to local and systemic immune mediated disease.

Dissecting out what is genetic influence on microbiome composition and what is

environmental is inherently complicated. A recent study by Hildebrand et al., 2013,

investigated exactly that across five common laboratory healthy mouse strains BALB/c,

FVB, B6, NOD and Swiss. They found that the gut microbiota differed according to

genotype, caging and inter-individual variation (165). These factors contributed on average

19%, 31.7% and 45.5% respectively to the variance in the microbial composition. Genetic

distance was also found to positively correlate to microbiota distance, indicating that

strains that are more closely related genetically have more similar microbiota than distantly

related strains. Microbiota composition was also correlated with inflammation marker

calprotectin. Mice that displayed low microbial richness had higher levels of calprotectin

but no obvious signs of inflammation suggesting low-grade inflammation. This study

demonstrates the influence of genetic background on microbiota composition as well as

the environmental influences albeit in a controlled setting. This work highlights the

complexity of host-environment-microbiota interactions (165).

The above studies highlight that changes in the microbiome can lead to inflammation

which may have far reaching effects, and demonstrate experimental approaches by which

findings of metagenomic studies in mice and humans can be explored to successfully

dissect the role of the microbiome in human immune mediated diseases.

49

1.11 Thesis outline This thesis investigates the role of the human gut microbiome in AS by characterising the

AS microbiome, examining how changes in host genetics influences intestinal microbial

composition and then exploring at the association between AS susceptibility loci and

microbiome composition in healthy twin cohort.

1.12 Aims Aim 1: To establish a pipeline for 16S community profiling the microbial communities

present in the terminal ileum of patients with AS, compared to CD and HC. Aim 2: To characterise the terminal ileum of patients with AS and determine if the AS gut

carries a distinct microbial signature in comparison to HC.

Aim 3: To investigate how loss of ERAP influences the gut microbiome composition, the

impact this has on the antigen repertoire presented and ensuing affects on the immune

system. Aim 4: To investigate the causality between AS susceptibility loci and gut microbiome

composition. The above aims will be discussed in separate results chapters that will form the body of

this thesis.

1.13 Significance The human gut microbiome is a dynamic and complex ecosystem that is only now

beginning to be understood. It is increasingly clear that the interaction between host and

microbiome profoundly affect health. It is still unclear how interactions between host

genes, microbes and environmental factors can predispose patients to the development

autoimmune diseases such as AS. We are only beginning to grasp the influence the

microbiome has on health. Improved knowledge of the composition and function of the gut

microbiome in AS patients and how the microbiome shapes the immune response and

influences inflammation, both local and systemic, will likely provide important insights into

early events in the pathogenesis of AS.

50

2 Chapter 2 Community profiling the terminal ileal microbiome

2.1 Introduction

It is increasingly clear that the interaction between host and microbiome profoundly affects

health. We are more microbe than we are human with over ten times more bacteria and

their genes living in and on our bodies, with the human intestine containing ~100 trillion

bacteria. Our ‘second genome’ of microbes are not passive bystanders and only now are

we beginning to appreciate the influence our microbial residents have on our health. Until

recently, interrogation of our microbial communities using classical microbiology

techniques offered a very restricted view of these communities, only allowing us to see

what we can grow in isolation. Advances in sequencing technologies in the past few years

have greatly facilitated systematic and comprehensive studies of the microbiome, and

elucidate its role in human health and disease.

AS has a global distribution, with no outbreaks or clustering being reported. The gene

identified as having the strongest association with AS development is the major

histocompatibility complex (MHC) gene HLA-B27. However only 1-5% of HLA-B27+

individuals go on to develop AS. Monozygotic twin studies suggest that the trigger for AS

in a genetically susceptible person is a ubiquitous environmental factor that is most likely

widely distributed (4, 5). There is a close relationship between the gut and SpA,

exemplified in reactive arthritis patients, where a typically self-limiting arthropathy follows

either gastrointestinal infection with Campylobacter, Salmonella, Shigella or Yersinia, or

urogenital infection with Chlamydia. The mechanism by which AS is triggered, in a

genetically susceptible individual, is not well understood. Studies in the B27 transgenic rat

model of AS, supports the idea of a ubiquitous environmental trigger. The B27 transgenic

rat when exposed to normal commensal bacteria, develop disease while the rats that were

maintained in a germ-free environment remain disease free (15). CD is a common form of

IBD where, again, interplay between host genetic factors and largely undefined

environmental factors are thought to drive the disease. Microbial involvement, especially

from the gut, has been associated in both diseases however in AS, no definitive link has

been established (16, 17).

51

There is a strong overlap between AS and CD, 70% of AS patients have subclinical gut

inflammation with 5-10% of these patients with more severe intestinal inflammation

developing clinically defined IBD resembling CD (7). Multiple genes associated with AS

also play a role in gut immunity, and there is a marked over-representation of genes

associated with CD also being associated with AS (20). Of these loci and genes identified,

of particular interest are genes involved in the IL-23 pathway including STAT3, IL23R, and

IL12B (which encodes IL-12p40, the share subunit of IL-23 and IL-12 (20-22). In the

intestinal mucosa IL-23, IL-17 and IL-22 producing cells are enriched with IL-17 and IL-22

are known to be important regulators of intestinal ‘health’. IL-17 plays important roles in

intestinal homoeostasis. In the gut, innate-like immune cells act as sentinels, responding

very rapidly to alterations to the microbial composition of the gut with rapid secretion of IL-

17. Loss of function polymorphisms in IL23R are associated with protection from AS (55)

and IBD (57), and many other genes in the IL-23 pathway are associated with these

diseases. This suggests that similar aetiopathogenic mechanisms may likely involve the

gut microbial communities and that dysbiosis of the gut flora is a significant contributor to

disease. There have been several studies have shown that AS patients and their first-

degree relatives have increased intestinal permeability relative to unrelated healthy

controls, again consistent with a role for the gut microbiome in the disease (11). Whilst

there are significant genetic similarities, major differences also exist between the genetics

of IBD and AS, with AS showing no association to date with NOD2/CARD15 or the

autophagy gene ATG16L1, which are major susceptibility factors in IBD, whereas IBD

shows no association with HLA-B or ERAP1 which are the strongest AS-susceptibility

genes (166).

Together, the multi-factorial components of the immune system shape the microbial

population that inhabits the gut and, to an extent, vice versa, with each side pushing to

establish a stable state. Understanding the yin and yang of this relationship is at the heart

of current microbiome research. Microbiome refers to the totality of microbes, their genetic

elements and environmental interaction in a defined environment (106). The human body

has more microbes living on it or in it, than it does human cells with approximately 29% of

all these microbes residing in the gut. Homeostasis of the intestinal microbiome and

relationship between the microbes that inhabit the gut and the immune system is critical.

The gastrointestinal tract is heavily populated with microbes and is the primary site for

interaction between these microorganisms and the immune system (25, 26). Furthermore,

microbes found in the gut help to shape host immune systems from an early age. The

52

incomplete development of the immune system in neonates and under germ-free

conditions tells us that microbiota sculpt the host immune system (27, 28). Maintenance of

intestinal and microbial homeostasis is being increasingly recognised as playing a pivotal

role in overall health (29), and dysregulation of either gut or microbial homeostasis may

play a role in autoimmunity. Physiological processes in the host that maintain gut

homeostasis and respond to perturbance in the gut micro-environment involve both the

adaptive and innate immune system, and the barrier function of the intestines themselves,

have been implicated in a number of immune mediated diseases, including multiple

sclerosis, inflammatory bowel disease, and type I diabetes (25, 102, 167, 168).

The human gut microbiome contains a dynamic and vast array of microbes that are

essential to health as well as provide important metabolic capabilities. Until recently

studying these complex communities has been difficult and generally limited to classical

phenotypic techniques (109, 116). The application of molecular techniques, particularly

sequencing, has shown the remarkable amount of diversity in the human gut, exceeding

all previous held beliefs and expectations. Development of high-throughput sequencing

methods has greatly improved our ability to profile complex microbial communities without

the requirement to culture organisms. The gene commonly used for community profiling is

the 16S ribosomal RNA (rRNA). This ~1,550bp section of prokaryotic DNA is found in all

bacteria and archaea, is highly conserved and is composed of both variable and

conserved regions. The 16S rRNA is able to differentiate between organisms across all

major phyla of bacteria and to classify strains down to species level (108). 16S rRNA

sequencing has dramatically changed our understanding of phylogeny and microbial

diversity because it provides an unbiased assessment of the microbiome, and is not

restricted by the ability to culture the bacteria present.

The aim of this chapter was to establish a pipeline to characterise the microbial community

present in the terminal ileum of patients with AS compare to CD and healthy controls (HC).

This included tissue extraction methods for optimal microbial recovery, choice of 16S

amplicon sequencing as well as analysis strategy. To validate the pipeline, a pilot study

was undertaken consisting of four AS patients, three CD patients and four HC to determine

whether the gut microbiome of AS patients is different and distinct from CD patients and

HC. This pilot study also included the use of two mock communities consisting of well

characterised bacteria of known concentrations that served as internal sequencing and

analysis controls. In addition to exploring community-level composition and differences

53

between affection states, we also evaluated technical replication by sequencing from both

forward and reverse primers and analysing them separately. Biological replication was

assessed by halving biopsies, then processing and analysing them independently. The

resulting sequence libraries were analysed using the QIIME pipeline as described in the

Material and Methods (169). To determine differences in microbial load, total microbial

biomass in biopsy samples was quantified using real-time qPCR of the 16S rRNA gene as

described in Willner et al., 2013 (170).

2.2 Materials and Methods

2.2.1 Optimisation of 16S primers 347F 803R and 517F 803R

These two primer pairs spanning the V3-V5 region were optimised using a mock

community of enteric bacteria, a human cell line control and a TI biopsy. The mock

community consists of five enteric bacteria Escherichia coli, Klebsiella pneumonia,

Klebsiella oxytoca, Enterobacter spp., Citrobacter spp. and as well as Listeria

monocytogenes. A human cell line control was used to test if the primer set amplified

human DNA as well as bacterial DNA. The primer pairs 347F 803R and 517F 803R were

taken through to sequencing as their amplicon size was compatible with the requirements

of Illumina sequencing of 300-500bp including linkers and adaptors.

2.2.2 16S rRNA PCR The master mix for the reactions in a 25ul total volume using Bioline 10x reaction buffer,

dNTP(0.625mM), mgcl2 (2.0mM), 1U of Bioline TAQ, both primers 5uM each and 2ul

DNA. Thermocycling conditions comprised 10 min 94ºC then 30 cycles at 94ºC 50 s, 54º

30 s, 72ºC 45 s and 72ºC 10 min.

2.2.3 In silico assessment of Primer pairs To compare the spectrum of diversity of our two primer sets in silico library simulation was

conducted using Grinder was used to generate mock data sets from the microbial refseq

library (NCBI) aligning to whole genomes (171). These data sets consisted of 10 patients

that contained 1000 reads per library. Reads were generated this way to best represent

what will likely happen with patient gut biopsies as well as to cache the microbial diversity.

54

Data sets were analysed using the QIIME pipeline (qiime.sourceforge.net) (172). OTU

assignment was based on 97% similarity with taxonomic assignments made using BLAST

to Greengenes (113) as implemented in QIIME.

2.2.4 Pipeline optimisation Intestinal Biopsies Resected colon Biopsies were collected from patients with diagnosed with either

Ulcerative colitis (UC) or Crohn’s disease (CD) courtesy of A/Prof Graham Radford-Smith

for the purpose of the pilot study and protocol optimisation (Table 2.1). Table 2.1 Intestinal biopsies utilised for pipeline optimisation.

Biopsy ID Sample

description

Histology Diagnosis DNA

Concentration

ng/ul

Q1462 Ileum Non-Inflamed UC 128.15

A457 Ileum Non-inflamed UC 54.57

12.001.08 Ileum Inflamed CD 48.95

53.001 Colon Inflamed UC 9.77

83.001.03 Ileum Inflamed CD 113.82

2.2.5 Sequencing of Pilot intestinal biopsies To compare the two primer sets five gut biopsies were amplified with 347F 803R and the

same six were amplified with 517F 803R, multiplex and run on the HiSeq. 10 16S rRNA

amplicon libraries were generated from five gut biopsies and converted to Illumina HiSeq

compatible prior to adaption of ligation of barcodes the amplicon libraries were QC

checked using the bioanalyzer (Agilent) looking for discrete peaks at 286bp and 456bp.

1µg amplicon library were constructed using the TruSeq DNA sample preparation kit

(Illumina) as per manufacturer’s instructions, commencing at the end repair stage and

omitting both the fragmentation and gel purification steps. The final library underwent QC

on the Agilent Bioanalyzer to confirm that indexing was successful, by looking for a shift in

the discrete peaks to 386bp and 556bp respectfully. These libraries were then pooled for

multiplexing using equimolar proportions of each library to a final concentration of 2µM.

55

Table 2.2 Pipeline optimisation sequencing metrics.

Gut biopsy Primer set Number of

sequences per

library

Q1462347F 347F 803R 2681259

Q1462517F 517F 803R 2341819

A457347F 347F 803R 2281754

A457517F 517F 803R 2263814

12.08347F 347F 803R 2739502

12.08517F 517F 803R 2198014

53.01D347F 347F 803R 2265574

53.01517F 517F 803R 2347683

83.03347F 347F 803R 1384798

83.03517F 517F 803R 1421586

2.2.6 Bioinformatics 10 16S rRNA amplicon libraries were generated from five gut biopsies using the two primer

sets and sequenced on the Illumina HiSeq (Table 2.2). The multiplexed reads were de-

multiplexed using the Illumina program CASAVA. Fastq reads were converted to QIIME

format using custom script. Samples were analysed using the QIIME pipeline

(qiime.sourceforge.net) (169). OTU assignment was based on 97% similarity with

taxonomic assignments made using BLAST to Greengenes(113) as implemented in

QIIME.

Sequence depth analysis was conducted by sub setting samples at –n 1000, 5000, 10

000, 50 000, 100 000, 500 000 and 1 000 000 reads per sample and then graphed using

the plot_rank_abundance_graph.py as implemented in QIIME.

2.2.7 Community profiling the terminal ileal microbiome

2.2.7.1 Ethics statement

Written informed consent was obtained from all subjects, and the study protocol was

approved by the local ethics committee of the University of Palermo.

56

2.2.7.2 Patient clinical data

AS patients were diagnosed according to the modified New York criteria for the disease

(173). No AS and CD patients were receiving biological agents such as TNF-inhibitors,

and only one AS patient was on nonsteroidal anti-inflammatory (NSAIDs) drugs at time of

biopsy sampling. Healthy controls were undergoing ileocolonoscopy for routine evaluation

(Table 2.3).

2.2.7.3 DNA extraction and sequencing

TI biopsies were thawed and then divided to create biological replicates. Each was then

manually disrupted using a polypropolene disposable pestle from Sigma. DNA extractions

were performed using the Qiagen DNeasy tissue extraction kit as per manufacturer’s

instructions.

57

Table 2.3 Clinical details of patients used in preliminary community profiling studies.

ID Diagnosis Biopsy

Site

Age Disease

duration

CRP mg/l

(0-10mg)

HLA-B27 Treatment

AS01 Ankylosing

spondylitis

Ileum 56 7 years 18 Pos NSAID

AS02 Ankylosing

spondylitis

Ileum 33 4 years 1 Pos None

AS03 Ankylosing

spondylitis


AS04 Ankylosing

spondylitis


CD01 Crohn’s

disease

Ileum 44 12 months 12 N/A Mesalazine

CD02 Crohn’s

disease

Ileum 32 4 months 6 N/A none

CD03 Crohn’s

disease

Ileum 47 7 months 8 N/A Mesalazine

HC01

None Ileum 58 N/A 0.3 N/A N/A

HC02


HC03


HC04


58

2.2.7.4 Bacterial Mock communities

Mock communities consisting of six different bacteria, Gram positive and Gram negative,

was employed as a control for both sequencing and bioinformatic analysis and was

processed alongside the TI biopsies. The six bacteria used were Staphylococcus aureus

(ATCC 25923), Enterococcus faecalis (ATCC 51299), Haemophilus influenzae

(ATCC49247), Klebsiella oxytoca (ATCC 700324), Escherichia coli (ATCC 35218) and

Yersinia enterocolitica (ATCC2 3715). Bacteria were extracted using the Qiagen DNeasy

tissue extraction kit as per manufacturer’s instructions. Two communities were then

constructed using differing and varying concentrations of the seven bacteria and labelled

Mock1 and Mock2.

Two mock communities (Tables 2.4 and 2.5) were created as PCR and sequencing

controls. The first mock community (Mock1) contains six bacteria at varying concentrations

and then pooled. The second mock community (Mock2) contains the same six bacteria

and was normalised for 16S copy number to 10ng/ul and then pooled at varying ratios.

Table 2.4 Mock1 construction using bacteria at varying concentrations.

Bacteria Concentration

ng/ul

Dilution Pooled

amount

ul

Relative

abundance %

Y. enterocolitica 28.39 Neat 10 89.1

E. faecalis 7.49 1:10 10 8.9

K. oxytoca 31.08 1:100 10 0.9

H. influenzae 5.47 1:100 10 0.9

S. aureus 17.89 1:1000 10 0.1

E. coli 12.47 1:1000 10 0.1

59

Table 2.5 Mock2 – 16S copy number normalised to 10ng/ul at varying ratios.

Bacteria Concentration

ng/ul

16S copy

number

Pooled

amount

ul

Relative

abundance %

Y. enterocolitica 28.39 7 3 21.4

E. faecalis 7.49 4 1 7.1

K. oxytoca 31.08 8 4 28.7

H. influenzae 5.47 6 2 14.3

S. aureus 17.89 5 1 7.1

E. coli 12.47 4 3 21.4

2.2.7.5 16S rRNA PCR

Bacterial 16S rRNA amplicon PCR was conducted using the V4 primer spanning the

region 517F 803R of the 16S gene. The master mix for the reactions in a 25ul total volume

using Bioline 10x reaction buffer, dNTP(0.625mM), mgcl2 (2.0mM), 1U of Bioline TAQ,

both primers 5uM each and 2ul DNA. Thermocycling conditions comprised 10 min 94ºC

then 30 cycles at 94ºC 50 s, 54º 30 s, 72ºC 45 s and 72ºC 10 min.

2.2.7.6 Sequencing Library preparation

16S rRNA amplicon libraries were assembled and indexed for multiplexing using the

Illumina TruSeq DNA library protocol as per manufacturer’s instructions. Samples were

then normalized using the Agilent Bioanalyzer and pooled for sequencing at a

concentration of 2nM. To create the required complexity for sequencing, PhiX was added

to the pool at 50/50 for a final pool concentration of 4pM.

2.2.7.7 Illumina MiSeq Sequencing

Barcoded 16S amplicon sequences were sequenced with 2x150bp paired-end on an

Illumina MiSeq with approximately 100 000 reads per sample (Table 2.6).

60

Table 2.6 Counts of 16S rRNA sequence reads obtained per biopsy.

Sample name Total sequences

AS01a 279,668

AS01b 170,508

AS02a 132,010

AS02b 262,060

AS03a 115,598

AS03b 178,534

AS04a 237,744

AS04b 219,418

CD01a 103,896

CD01b 124,414

CD02a 185,934

CD02b 200,200

CD03a 237,004

CD03b 272,198

HC01a 688,184

HC01b 121,098

HC02a 316,964

HC02b 327,222

HC03a 235,822

HC03b 241,690

HC04a 226,356

HC04b 242,772

Mock1 284,316

Mock2 244,840

2.2.7.8 Bioinformatics and statistics

Reads were de-multiplexed using the Illumina program CASAVA. Fastq reads were

converted to QIIME format using and amplicon orientated using custom scripts. Samples

were normalised at 10,000 reads per sample and then analysed using the QIIME pipeline

(qiime.sourceforge.net) (169). OTU assignment was based on 97% similarity with

taxonomic assignments made using BLAST to Greengenes(113) as implemented in

QIIME.

61

Alpha diversity metrics, which look at the diversity within a sample, were generated using

the QIIME workflow alpha_rarefaction.py with five repetitions at each sampling point.

Rarefaction curves for the observed number of OTUs were based on the outputs of

alpha_rarefaction.py. UniFrac distances between microbial communities were evaluated

using FastUniFrac (174). UPGMA clustering was employed both weighted and

unweighted, to detect significant differences between microbial communities between AS,

CD and HC (175). FastUniFrac was also used to generate sample distance metrics as well

as Principle Coordinates Analysis.

PERMANOVA analysis was conducted on the genus level OTU table using the R package

vegan to test the relationship between the whole microbial community and disease (176).

Indicator species analysis, using the ‘labdsv’ R package, was employed to investigate

what microbial species are driving the microbial community and driving the signature

(177).

2.3 Results

2.3.1 In silico assessment of Primer Pairs

In silico library simulation of the primer pairs 347F 803R and 517F 803R returned 14 and

21 different phyla respectively. The dominant phyla previously reported to inhabit the gut

are Bacteroidetes and Firmicutes with lower abundance of Actinobacteria, Proteobacteria,

Fusobacteria and Verrucomicrobia (152, 178). The purpose of this experiment was to

demonstrate that the primer pairs covered the range of expected phyla as well as the

capacity to discriminate other phyla. The results show that, as expected, both primer sets

were able to classify all dominant and minor phyla as previously reported to inhabit the

gastrointestinal tract (Table 2.7). This demonstrates that either primer set would

adequately profile the intestinal microbiome. However, 517F 803R identified seven more

phyla, including some phyla that have been reported as being low in abundance such as

Verrucomicrobia and Lentisphaerae (178). Based on the in silico simulation, the primer set

517F 803R would be better suited to interrogate human gut biopsies moving forward.

62

Table 2.7 Taxon output by Grinder.

Phyla 347f 803R Phyla 517F 803R

Acidobacteria Acidobacteria

Actinobacteria Actinobacteria

Bacteroidetes Bacteroidetes

Flavobacteria Flavobacteria

Sphingobacteria Sphingobacteria

Chlorobia Chlorobia

Chrysiogenetes Chloroflexi

Deferribacteres Chrysiogenetes

Fibrobacteres Deferribacteres

Firmicutes Deinococci

Fusobacteria Dictyoglomi

Nitrospira Fibrobacteres

Proteobacteria Firmicutes

Mollicutes Fusobacteria

Lentisphaerae

Nitrospira

Planctomycetacia

Proteobacteria

Spirochaetes

Mollicutes

Verrucomicrobia

2.3.2 Sequencing of Pilot intestinal biopsies

Preliminary analysis of primer pairs in a pilot study using five colon biopsies showed that

there were three major phyla observed across the gut biopsies were Bacteroidetes,

Firmicutes and Proteobacteria (Figure 2.1). This is consistent with what has been reported

in the literature (114, 152). The minor phyla present include Actinobacteria, Fusobacteria,

Cyanobacteria, and Acidobacteria with a very low unclassified ‘other’. These ‘others’ may

be members of the phylum Archea, such as Methanobacteria, which are also common in

the gut. Using an alternate reference database, such as Silva, may possibly resolve these

unclassified sequences or they may not be present in the reference databases.

63

The phylum level classification illustrates that the 517F 803R primer set seems to classify

slightly more members of the Firmicutes phylum, while the 347F possibly classified

Bacteroidetes better. The Komolgorov-Smirnov test was run at the phylum level in order to

evaluate the differences between the two primer sets and to compare the similarities and

differences between the overall sequences corresponding to the same phyla. Neither of

the primer pairs came up significantly different from each other as all p-values were

greater than 0.90. This means that using one primer compared to the other did not cause

significant differences detected in the community profiles at the phylum level.

Figure 2.1 Phyla represented in gut biopsies showing percentage abundance. Comparison of primer pairs 347F 80R and 517F 803R demonstrating no significant differences when trialled in five colon biopsies for the purpose of pipeline optimisation

2.3.3 Sequence depth analysis

Illumina high-throughput sequencing is capable of producing millions of reads, so

ascertaining the appropriate number of reads required per sample for amplicon

sequencing is important not only for taxonomy assignment but sample multiplexing as well.

12.08347F

12.08517F

53.01D347F

53.01517F

83.03347F

83.03517F

A457347F

A457517F

Q1462347F

Q1462517F

55.02347F

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Firmicutes

Bacteroidetes

Proteobacteria

Actinobacteria

Fusobacteria

Cyanobacteria

Acidobacteria

Other

64

Whilst deep sequencing of PCR amplicons enables the detection of rare or low abundant

organisms, this can also lead to over estimation of microbial diversity by the generation of

reads containing errors (112, 179). Recent work by Bokulich et al., 2012 demonstrated

that strict quality filtering of reads greatly improves taxonomy assignment and α-diversity

measures for microbial community profiling (180). The five samples from the pilot study

were sampled at varying read depths ranging from 1000 reads to 1 000 000 reads and

compared (Figure 2.2). The analysis showed that 1000 and 5000 reads per sample

detected the least number of species with 500 000 and 1 000 000 reads per sample

detecting the greatest number. The concern is that the more species being detected may

not be true and may in fact be sequence error (112). With this in mind the taxonomy

assignment from 1000 to 50 000 reads was compared, with on minor differences found.

Based on this 10 000 reads was the depth selected to continue with as it identified the

same taxa as 50 000 reads and allowed for more samples to be multiplexed.

Figure 2.2 Rank abundance graph showing number of reads (read depth) with number of species identified for the sample Q1462.

65

2.3.4 Community profiling the terminal ileal microbiome

The microbial communities present in the TI of AS patients were found to be significantly

different and more diverse than those communities associated with CD and healthy

controls (P<0.001; PERMANOVA). Alpha diversity, a measure of richness and/or

evenness of taxa within an individual community, was increased in AS samples. The CD

and HC communities were more similar to each other than AS, but were less diverse

overall than the AS microbial community (Figure 2.3).

Figure 2.3 Rarefaction curves for microbial communities in AS patients (red curve), CD patients (blue curve), healthy controls (orange curve) and the mock communities (green curve).

Communities were compared using weighted and Unweighted UniFrac distances, a

phylogenetic metric that accounts for differences in both community membership and

relative abundance (174). The Weighted UniFrac distance metric is used to detect

changes in how many organisms are present in a community. It takes into account the

relative abundance of microbes present, which is particularly important for looking at

changes in bacterial abundances between communities. The Unweighted UniFrac

distances metric informs us about community membership is used to measure the core

differences between communities in terms of presences or comparative absence of taxa.

66

Hierarchical clustering based on both Weighted and Unweighted UniFrac distances

indicated the primary factor influencing differentiation of communities is disease, and that

the microbial communities present in AS patients are distinct from those in CD and HC

(Figure 2.4).

Figure 2.4 Hierarchical clustering of microbial communities based on Weighted Unifrac

distance clustering of terminal ileal biopsies of patients with ankylosing spondylitis (AS), Crohn’s

disease (CD), healthy controls (HC) as well as the two mock community controls.

67

Figure 2.5 Principal coordinates analysis of technical and biological replicates (transparent

points) showing the grouping of ankylosing spondylitis (AS) samples compared to Crohn’s disease

(CD) and healthy controls (HC) with the grouping of the replicates shown in the red circles.

PERMANOVA analysis confirmed the presence of a significant relationship between

disease status and microbial community composition (P<0.001). This was not due to

heterogeneity in biopsy samples as the biological replicates showed no significant

difference between each other (P<0.001) (Figure 2.6). Nor was it due to a lack of technical

reproducibility, as the technical replicates, like the biological, were not statistically different

from each other (P<0.001) and cluster tightly together (Figure 2.5).

68

Figure 2.6 Comparison of variation between and within samples and affection status

showing greater significance between and within affection status than between biological and

technical replicates Significance was assessed using Mann-Whitney test (P<0.001 = ***).

Average UniFrac distances demonstrated that the differences between both biological and

technical replicates were significantly less than differences between individuals and

disease states (Figure 2.6). Total microbial load was investigated using Biomass q-PCR.

Total microbial load is of interest as this is an indicator of the overall amount of bacteria

that are present in the community. In our samples, Biomass q-PCR showed that the

apparent microbial signature in AS patients was not due to bacterial overgrowth as there

was no significant difference on average in 16S rRNA copy number between AS, CD, and

HC (Figure 2.7). One sample, AS02, displayed a higher microbial load than the other AS

samples, while AS04 exhibited a lower overall microbial mass, illustrating the variability of

microbial community load across people and the disease. The biomass of CD and HC

samples were relatively similar to each other, with no outstanding increases or decreases

in microbial load.

69

Figure 2.7 Microbial community profiles of terminal ileal (TI) biopsies from ankylosing

spondylitis (AS), Crohn’s disease (CD) patients and healthy controls (HC). Biological replicates are

indicated as ‘a’ and ‘b’. Each row represents a different OTU with the abundance as a percentage

of the total population as represented by colour. Phyla and selected families are noted on the left,

and a subset of OTU classifications on the right. Microbial biomass of each biopsy is shown by the

bar graph at the bottom of the heat map.

70

2.4 Discussion

Whether the changes in the microbiome are a consequence of disease or are involved in

its development or persistence is unclear. This distinction may prove difficult to dissect in

human studies, but there is considerable evidence in studies in mice to support a role for

the microbiome in driving immune mediated diseases such as AS. In order to understand

the role of the gut microbiome in human disease, it is important to first be able to

accurately and confidently profile the microbiome. This requires confidence in tissue

extraction techniques, 16S rRNA primer choice to ensure diversity is sufficiently captured

and that the analysis strategy employed is appropriate (181).

During 16S pipeline validation, nine different primer sets spanning the 16S rRNA gene

were evaluated to determine which set provided the best coverage and discrimination of

intestinal flora. Emerging literature coupled with in silico analysis of the primer sets, we

focused two primer sets in the V3-V5 region of the 16S rRNA gene (127, 169). Two primer

sets 347F 803R and 517F 803R demonstrated ability to classify all expected major and

minor phyla, as previously reported to inhabit the intestinal microbiome, not only in silico

but also when trialled on human colonic biopsies (127, 128, 152, 182, 183). As well as

capturing the diversity of the bacteria in the intestinal microbiome, ability to selectively

amplify bacteria over human DNA was crucial in when considering primers. 16S rRNA

shares significant homology with human mitochondrial DNA and amplification of human

DNA is an issue. The final decision to move forward with the 517F 803R primer set was

also influenced by choice of sequencing platform and chemistry. Sequencing of the TI

biopsies was 2x150bp on the Illumina MiSeq, therefore sequencing read length was an

important determinant when considering primer sets for amplicon sequencing. Two

bacterial mock communities, consisting of six well-characterised bacteria at varying

concentrations, were used as internal standards to determine reproducibility across

multiple runs. These two mock communities enabled us to accurately compare PCRs and

sequencing runs, as well as serving as internal controls to validate the pipeline and

analysis approach.

There is increasing evidence that these indigenous microbes are also driving the

development and/or function of different of immune cells in the intestine. In the intestinal

mucosa, innate lymphoid cells (ILCs) have emerged as an important innate lymphocyte

population involved in immunity to intestinal infections, especially in conjunction with IL-22

71

(184). The transcription factor T-bet, encoded by the TBX21 gene, controls the fate of ILCs

and is yet another gene in the list of genes found to be controlled by cues from the

intestinal microbiome and IL-23. Meaning that the commensal microbiota are instructing T-

bet expression and therefore ILC development. (185).

Inflammation is a critical component in autoimmune disease. In the K/BxN mouse model of

arthritis it was shown that the introduction of a single gut-residing species, segmented

filamentous bacteria (SFB), into germ free animals was sufficient to cause proliferation of

Th17 cells, which led to the production of autoantibodies and rapidly progressed to

arthritis. Neutralisation of IL-17 in specific-pathogen-free K/BxN mice prevented arthritis

development. Thus, a single commensal microbe, via its ability to promote a specific T

helper cell subset, can drive autoimmune disease (101). Conversely, members of the gut

microbiota such as Clostridium spp., have been found to play a protective role by inducing

T regulatory cells (Tregs) in the colonic mucosa. Tregs are an important counter balance in

the immune system and promote homeostasis. It was found that a specific mix of cluster

IV and XIVa of Clostridium was sufficient to promote Treg accumulation in the colon (105).

To date no comprehensive characterisation of intestinal microbiota in AS patients has

been performed. Here we described the optimisation and validation of 16S rRNA culture-

independent microbial sequencing for the purpose of profiling TI biopsies. The choice to

examine intestinal biopsies, rather than less invasive stool samples, was in order to fully

characterise the mucosal microbiome at a site of initial inflammation. Several studies have

shown that while stool is routinely used for large scale investigations of the intestinal

microbiome (123, 126, 137), examining the mucosal microbiome has tremendous value as

the mucosal microbes are in a position to influence the immune system (152, 153, 186).

Four AS patients, three individuals with CD, four HC and two microbial mock communities

were profiled using the optimised 16S rRNA pipeline to determine if the AS gut carries a

distinct microbial signature. Hierarchal clustering showed that the microbiome of AS

patients is distinct and different from CD and HC, with the PCoA confirming that the AS

cases group separately. Statically the relationship between disease status and microbial

community composition was confirmed indicating that the differences between the

microbial community compositions were not due to heterogeneity in the biopsy or lack of

technical reproducibility, but that the driving force is disease state. These results

demonstrate that the tissue extraction method coupled with choice of 16S rRNA primers,

was able to capture and profile the intestinal microbiome in human TI samples.

72

Given what we know from murine experiments, the differences we see in the microbial

composition in AS gut compared to CD and HC may have wider reaching implications. The

evidence is mounting that not just the overall composition of the intestinal microbiome, but

presences and/or absences of specific microbes that have a substantial impact on host

response, regulation of inflammation and development of intestinal cells. What is clear is

that the interactions between our genetics, immune system and resident microbiota are

complex but critical. Our intestinal microbiota are clearly not just bystanders but are in fact

orchestrators of our health and disease, so having the capacity to accurately profile the

microbiome is imperative. Further studies are needed into whether the changes in

intestinal microbial composition are due to genetics, and how this affects the overall

function of the gut microbiome in AS patients, including how the microbiome then goes on

to shape the immune response and influence inflammation. Now have the tools to

sequence and analyse intestinal microbial communities, these investigations will likely

provide new insights and help us to better understand AS disease pathogenesis.

73

3 Chapter 3 Intestinal dysbiosis in ankylosing spondylitis

Accepted and published online as a Brief Report in Arthritis and Rheumatology 2015,

2015;67(3):686-91.

Mary-Ellen Costello1, Francesco Ciccia2*, Dana Willner3*, Nicole Warrington1, Philip C.

Robinson1, Brooke Gardiner1, Mhairi Marshall1, Tony J. Kenna1*, Giovanni Triolo2*, Matthew

A. Brown1

1The University of Queensland, The University of Queensland Diamantina Institute, Translational Research Institute, Princess Alexandra Hospital, Woolloongabba, Brisbane, QLD 4102, Australia. 2University of Palermo, Palermo, Italy. 3Australian Centre for Ecogenomics, School of Chemistry and Molecular Biosciences, University of Queensland, Brisbane, QLD 4072, Australia. *These authors contributed equally to this work.

Modified article forms Chapter 3, full brief report (PDF) in appendices

3.1 Abstract

Objective. Ankylosing spondylitis (AS) is a common highly heritable immune mediated

arthropathy that occurs in genetically susceptible individuals exposed to an unknown but

likely ubiquitous environment trigger. There is a close relationship between the gut and

SpA, exemplified in reactive arthritis patients, where a typically self-limiting arthropathy

follows either gastrointestinal or urogenital infection. Microbial involvement has been

suggested in AS, however, no definitive link has been established. We sought to

determine if the AS gut carries a distinct microbial signature, in comparison to healthy

controls (HC).

Methods. Microbial profiles from terminal ileal (TI) biopsies from subjects with recent-

onset, TNF-antagonist naïve AS and HC, were generated using culture-independent 16S

rRNA gene sequencing and analysis techniques.

74

Results. Our results show the TI microbial communities of patients with AS differ

significantly (P<0.001) from HC, driven by higher abundance of five families of bacteria

Lachnospiraceae (P=0.001), Ruminococcaceae (P=0.012), Rikenellaceae (P=0.004),

Porphyromonadaceae (P=0.001), and Bacteroidaceae (P=0.001), and a decreases in

abundance of two families Veillonellaceae (P=0.01) and Prevotellaceae (P=0.004).

Conclusions. We show evidence for a discrete microbial signature in the TI of cases with

AS, compared to HC. The microbial composition was found to correlate with disease

status and greater differences were observed between than within disease groups. These

results are consistent with the hypothesis that genes associated with AS act at least in part

through effects on the gut microbiome.

3.2 Introduction Intestinal microbiome dysbiosis and microbial infections have been implicated in a number

of immune mediated diseases, including multiple sclerosis, inflammatory bowel disease

(IBD), and type 1 diabetes. Bacterial infections in the gut and urogenital tract are known to

trigger episodes of reactive arthritis, a form of spondyloarthropathy (SpA), a group of

related inflammatory arthropathies of which ankylosing spondylitis (AS) is the prototypic

disease. There is a close relationship between the gut and SpA, exemplified in reactive

arthritis patients, where a typically self-limiting arthropathy follows either gastrointestinal

infection with Campylobacter, Salmonella, Shigella or Yersinia, or urogenital infection with

Chlamydia. Microbial involvement has been suggested in AS, however, no definitive link

has been established (16, 17).

Up to 70% of AS patients have subclinical gut inflammation with 5-10% of these patients

with more severe intestinal inflammation go on to develop clinically defined IBD resembling

Crohn’s disease (CD) (7). The high heritability of AS (5), the global disease distribution

and the absence of outbreaks of the disease, suggest that AS is triggered by a common

environmental agent in genetically susceptible individuals (187). Multiple genes associated

with AS also play a role in gut immunity such as genes involved in the IL-23 pathway (22,

24), which are important regulators of intestinal ‘health’. There is a marked over-

representation of genes associated with CD that are also associated with AS (18, 20); this

suggest the two diseases my have similar aetiopathogenic mechanisms, possibly involving

gut dysbiosis (188).

75

Several studies have shown that AS patients and their first-degree relatives have

increased intestinal permeability relative to unrelated healthy controls, again consistent

with a role for the gut microbiome in the disease (6, 11, 12). To date, no comprehensive

characterization of intestinal microbiota in AS patients has been performed. Here we

performed culture-independent microbial community profiling of terminal ileal (TI) biopsies

from nine AS and HC to determine to investigate differences the gut microbiome in these

disease states in comparison with each other and with HC.

3.3 Materials and Methods

3.3.1 Patient clinical data

Terminal ileal biopsies were collected at colonoscopy from nine consecutively enrolled

TNF-inhibitor naïve, recently diagnosed AS cases (defined according to the modified New

York classification criteria for AS (173)) and nine healthy controls (Table 3.1). All patients

gave written informed consent and the study protocol was approved by the relevant

Universities of Palermo and Queensland research ethics committees. Disease duration

listed in Table 3.1 is from onset of symptoms, as reported by the patients, not duration

since diagnosis. One AS patient was on NSAIDs at the time of biopsy. AS03, AS04 and

AS10 had reported occasional although interrupted use of NSAIDs due to gastrointestinal

upset. Other AS patients did not report use of NSAIDs but were using either paracetamol

and/or tramadol.

Table 3.1. Clinical details of patients supplying biopsies at time of collection. Disease duration is from date of onset of symptoms.

ID Diagnosis Biopsy site

Age Gender Disease duration

ESR mm/Hr

CRP mg/l

BASDAI HLA-B27 status

NSAID treatment

Histological inflammation

AS01 AS Ileum 56 M 7 years 34 18 7 Pos Current Chronic AS02 AS Ileum 33 M 4 years 22 1 5.5 Pos None Acute AS03 AS Ileum 24 F 5 years 18 5 4.8 Pos None Normal AS04 AS Ileum 22 M 3 years 33 3 5.5 Pos None Normal AS05 AS Ileum 33 M 5 years 41 2.7 6 Pos None Chronic AS06 AS Ileum 28 M 6 years 28 1.5 6.2 Pos None Normal AS08 AS Ileum 36 M 8 years 34 1.5 5.6 Pos None Acute AS09 AS Ileum 38 F 6 years 45 2.2 6.7 Pos None Chronic AS10 AS Ileum 31 M 11 years 27 1.3 8 Pos None Normal HC01 HC Ileum 58 F N/A N/A 0.3 N/A N/A N/A N/A HC02 HC Ileum 43 M N/A N/A 0.5 N/A N/A N/A N/A HC03 HC Ileum 38 F N/A N/A 0.05 N/A N/A N/A N/A HC04 HC Ileum 65 M N/A N/A 0.8 N/A N/A N/A N/A HC05 HC Ileum 48 F N/A N/A 0.02 N/A N/A N/A N/A HC06 HC Ileum 34 F N/A N/A 0.03 N/A N/A N/A N/A HC07 HC Ileum 45 M N/A N/A 0.5 N/A N/A N/A N/A HC08 HC Ileum 44 F N/A N/A 0.8 N/A N/A N/A N/A HC09 HC Ileum 56 F N/A N/A 0.02 N/A N/A N/A N/A

77

3.3.2 DNA extraction and sequencing

Samples were preserved in liquid nitrogen or on dry ice till processed. TI biopsies were

thawed and divided in two to create biological replicates. Each sample was manually

disrupted using a polypropylene disposable pestle from Sigma. DNA was extracted from

tissue biopsies using Qiagen DNeasy tissue extraction kits after manual disruption as per

manufacturer’s instructions.

3.3.3 Mock community

A mock community consisting of six different bacteria, Gram positive and Gram negative,

was employed as a control for both sequencing and bioinformatic analysis and was

processed alongside the TI biopsies. The mock community contained Staphylococcus

aureus (ATCC 25923), Enterococcus faecalis (ATCC 51299), Haemophilus influenzae

(ATCC49247), Bacteroides fragilis (ATCC 25285), Corynebacterium renale (ATCC 19412)

and Yersinia enterocolitica (ATCC2 3715). The bacteria were extracted using the Qiagen

DNeasy tissue extraction kit as per manufacturer’s instructions. The bacterial mock

community was constructed using differing amounts of the six bacteria and labelled Mock

(Table 3.2).

Table 3.2 Bacterial mock community composition.

Bacteria Concentration ng/ul Pooled amount

Y. enterocolitica 28.39 1ul

E. faecalis 7.49 2ul

B. fragilis 7.3 4ul

H. influenzae 5.47 4ul

S. aureus 17.89 1ul

C. renale 2.4 2ul

3.3.4 16S rRNA PCR

Bacterial 16S rRNA amplicon PCR was conducted using the V4 primer spanning the

region 517F GCCAGCAGCCGCGGTAA and 803R CTACCRGGGTATCTAATCC of the

16S rRNA gene. All samples, including the samples presented previously in Chapter 2

(AS01-AS04 and HC01-HC04), were processed in one batch together in this experiment.

78

The master mix for the reactions in a 50ul total volume using Invitrogen Platinum PCR

SuperMix High Fidelity. Mastermix consisted of 45ul of SuperMix, 1ul each of both primers

at 10uM and 3ul of DNA. Thermocycling conditions comprised of 2min at 94ºC then 25

cycles of 94ºC for 30s, 54ºC for 30s, 68ºC for 45 s and a single 72ºC for 10 min.

3.3.5 Sequencing Library preparation

16S rRNA amplicon libraries for all samples presented in this chapter, including the

samples presented previously in Chapter 2 (AS01-AS04 and HC01-HC04), were

assembled and indexed for multiplexing using the Illumina 16S Metagenomic Sequencing

library protocol as per manufacturer’s instructions. All samples were normalised using the

Agilent Bioanalyzer and pooled for sequencing at a concentration of 2nM. To create the

required complexity for sequencing, PhiX was added to the pool at 25/75 for a final pool

concentration of 4pM.

3.3.6 Illumina MiSeq Sequencing

Barcoded 16S amplicon sequences were sequenced with 2x250bp paired-end reads on a

single Illumina MiSeq run. A total of 6.1M reads were generated with an average of

200,000 reads per sample.

3.3.7 Bioinformatics and statistics

Samples were demultiplexed and quality filtered using split_libraries_fastq.py (Table 3.3),

then analysed using the QIIME package and workflow scripts (qiime.sourceforge.net)

(169). Operation Taxonomy Unity (OTU) assignment was based on 97% similarity with

taxonomic assignments made using BLAST to Greengenes (113) as implemented in

QIIME.

79

Table 3.3 16S rRNA sequence read counts per sample post quality control and filtering

Sample names Sequence counts AS01a 82,363 AS01b 132,597 AS02a 117,012 AS02b 77,983 AS03a 123,111 AS03b 126,647 AS04a 131,851 AS04b 150,298 AS05a 127,339 AS05b 81,085 AS06a 238,920 AS06b 121,211 AS07a 105,164 AS08a 113,490 AS08b 94,585 AS09a 117,119 AS09b 139,181 AS10a 120,418 AS10b 124,143 HC01a 111,807 HC01b 103,234 HC02a 161,879 HC02b 168,001 HC03a 166,127 HC03b 172,902 HC04a 92,936 HC04b 132,558 HC05a 100,554 HC05b 99,696 HC06a 121,273 HC06b 137,985 HC07a 91,674 HC07b 119,069 HC08a 94,205 HC08b 63,347 HC09a 110,494 HC09b 112,864 Mock 103,896

80

Alpha diversity metrics were generated using the QIIME workflow alpha_rarefaction.py

with five repetitions at each sampling point. Rarefaction curves for the observed number of

OTUs were based on the outputs of alpha_rarefaction.py. UniFrac distances between

microbial communities were evaluated using FastUniFrac (174).

Hierarchical clustering was employed using Unweighted Pair Group Method with

Arithmetic mean (UPGMA) both weighted and Unweighted UniFrac distance, to detect

significant differences within microbial communities between AS and HC (175). The

Weighted UniFrac distance metric detects changes in how many organisms are present in

a community, taking into account the relative abundance of microbes present. The

Unweighted UniFrac distance metric describes community membership. UniFrac enabled

in QIIME was used to generate sample distance metrics as well as Principal Coordinate

Analysis (PCoA) (189). To determine differences in microbial load, the total microbial

biomass in biopsy samples was quantified using real-time qPCR of the 16S rRNA gene as

described in Willner et al., 2013 (170).

PERMANOVA analysis was conducted on the genus level OTU table using the R package

‘vegan’ to test the relationship between the whole microbial community and disease (176).

Indicator species analysis, using the ‘labdsv’ R package, was employed to investigate

what microbial species are driving the microbial community and driving the signature

(177). Co-occurrence network analysis, utilizing the R package ‘Spaa’, was carried out to

further investigate correlations between microbial families in the AS microbiome. Core

microbiome analysis, as well as supervised learning, was performed to further characterise

the intestinal microbial signature using the workflows compute_core_microbiome.py and

supervised_learning.py in QIIME.

To interrogate the function of bacteria and communities, Phylogenetic Investigation of

Communities by Reconstruction of Unobserved States (PICRUSt) was used to predict the

metagenome functional content from the 16S rRNA OTU table (115). Using the workflow

normalize_by_copy_number.py in PICRUSt, the OTU table was normalised by copy

number to limit copy number bias by dividing each OTU by the known or predicted 16S

copy number. The metagenome was predicted using predict_metagenomes.py, which

creates the prediction by multiplying each normalized abundance by the predicted

functional trait abundance, to produce a table of function by samples. The table produced

by the predict_metagenomes.py workflow is then analysed as a metagenome table. The

81

functions were collapsed into categories using the categorize_by_function.py script for

further analysis in R. Metagenomic pathways were compared between AS and HC using

generalized linear model glm in R.

The core microbiomes at varying thresholds were also investigated for metagenomic

pathways of interest. As described above, the Core 100, 90, 75 and 50 OTU tables were

processed through PICRUSt and then compared using Wilcox test and glm in R.

3.4 Results The TI microbial and mock communities were profiled by high-throughput barcoded

amplicon sequencing. In addition to exploring community-level differences between

disease states, we evaluated biological replication by halving biopsies, then processing

and analysing them independently

The microbial communities in the TI of AS cases were significantly different (Figure 3.1)

and more diverse (Figure 3.2C) than those communities associated with HC (P<0.001;

PERMANOVA). It was noted that three of the HC samples had higher levels of

Proteobacteria (HC01, HC08 and HC09) (Figure 3.1). The mock community profile was

consistent with control components (Table 3.2, Figure 3.1) showing 55% Proteobacteria

(Y. enterocolitica and H. influenzae), 37% Firmicutes (E. faecalis and S. aureus), 7.5%

Bacteroides (B. fragilis) and 0.5% Actinobacteria (C. renale).

82

Figure 3.1 Differences between AS and HC microbial communities. Bar chart showing the

difference in taxonomy at the Phylum level between AS and HC noting the increase in Bacteroides

and the change in the Bacteroides to Firmicutes ratio in the AS samples.

83

Figure 3.2 Shift in microbiome composition driven by disease status. (A) Weighted Principal Coordinates Analysis, which detects differences in organisms present in a community and taking into account their relative abundance, showing the distinct clustering of AS samples compared to HC, and supported by the supervised learning predicting disease status based off microbiome composition (B) Weighted Principal Coordinates Analysis, having removed the four outlying HC, still showing the distinct clustering of AS samples compared to HC (C) Alpha diversity box plot showing an increase of microbial diversity (observed species) in AS samples compared to CD and HC. (D) Comparison of variation between and within samples and affection status showing greater significance between and within affection status than between biological and technical replicates significance was assessed using Mann-Whitney test (P<0.001 = ***).

84

Figure 3.3. Weighted and Unweighted Principal Coordinates Analysis of the samples and their biological replicates (A) Weighted Principal Coordinates Analysis of the samples and their biological replicates (open symbols) showing not only distinct clustering samples by disease status, but tight clustering of the biological replicates. The bacterial (Mock) control included as a known positive control. (B) Unweighted Principle Coordinates Analysis, which detects differences in the presence and absences of microbes in a community, of samples and their biological replicates showing general but not distinct discrimination of AS samples and their biological replicates to HC, indicating there are some differences in the AS microbial community due to presence/absence of microbes.

PCoA showed AS samples group separately to HC (Figure 3.2), including biological

replicates (Figure 3.3), indicating that disease is the primary factor influencing community

differences. Four HC samples group separately and away from the other HC, and closer

to the mock community control. Three of the four HC samples contained increased

proportion of Proteobacteria (HC01, HC08 and HC09) and HC04 that had increased levels

of Actinobacteria, compared to the other HC samples. Therefore to examine if these

outlying samples were affecting the clustering seen on PCoA (Figure 3.2A), the four

outlying samples were removed and the remained samples reanalysed (Figure 3.2B). The

reanalysed PCoA showed the AS samples still clustered separately from HC, with the

exception of AS02 which also contained more Proteobacteria than the other AS samples

(Figure 3.2B).

To further demonstrate the distinct grouping of AS samples to HC, supervised learning

was conducted to test the predictive capacity of the microbiome differences, with all nine

AS cases predicted as AS (Figure 3.2A). PERMANOVA analysis further confirmed the

85

presence of a significant relationship between disease status and microbial community

composition (P<0.001). This was not due to heterogeneity in biopsy samples as biological

replicates showed no significant difference between each other (Figure 3.2D). Biomass q-

PCR showed no significant difference on average in 16S rRNA copy number between AS

and HC; indicating that the observed differences were not due to overgrowth or dominance

of bacteria driving community differences (Mann-Whitney test, P>0.05 = NS). Average

UniFrac distances demonstrated that differences between both biological and technical

replicates were significantly less than differences between individuals and disease states

(P<0.001; Mann Whitney) (Figure 3.2D).

16S community profiling, where sequencing reads were clustered against the Greengenes

reference database at 97% similarity (112, 113) showed 51 genera present across all

biopsies, with major differences observed at the phylum level between AS and HC (Figure

3.1). Indicator species analysis was performed to determine if alterations in specific

species were driving the differences observed between communities in cases with AS and

HC. Analysis showed that compared to HC, the microbial communities in AS cases were

characterized by higher abundances of Lachnospiraceae including Coprococcus spp. and

Roseburia spp. (P=0.001), Ruminococcaceae (P=0.012), Rikenellaceae (P=0.004),

Porphyromonadaceae including Parabacteroides spp. (P=0.001), and Bacteroidaceae

(P=0.001). Decreases were noted in Veillonellaceae (P=0.01) and Prevotellaceae

(P=0.004) (Figure 3.4A). Further drilling down into the AS microbiome signature, co-

occurrence analysis, which examines at the correlation between microbes, showed that

bacterial interactions further shape the AS microbial community signature with positive

correlations observed between the indicator species Lachnospiraceae and

Ruminococcaceae, and negative correlations observed between Veillonellaceae and

Prevotellaceae (Figure 3.4B).

86

Figure 3.4 Differences in bacterial family abundances and co-occurrence network. (A) Differences in relative abundances of bacterial families in AS (orange) microbiome compared to HC (red). The first seven families of bacteria are AS indicator species, including P-values, with the last three bacterial families being HC indicator species. (B) Co-occurrence network showing the positive (blue lines) correlations, between AS indicator species Lachnospiraceae and Ruminococcaceae families, and negative (red lines) correlations between families such as Veillonellaceae, Porphyromonadaceae and Prevotellaceae. The thickness of the line denotes the strength of the correlation (Pearson correlation).

Given the distinct differences seen between the AS and HC intestinal microbiome

composition, we were interested in seeing how changes in composition affect microbiome

function. Microbial function was inferred from the 16S rRNA data using the prediction

software PICRUSt (115). PICRUSt analysis indicated 31 significant pathways with

differential representation in AS compared to HC (Bonferroni corrected P<0.02) (Table

3.4). These included a decrease in Bacterial invasion of epithelial cells (P=4.33x10-5), an

increase in antimicrobial production in the Butirosin and neomycin biosynthesis pathway

87

(P=0.002), which is consistent with an increase in bacteria from Bacteroidaceae family,

and an increase in the Secondary bile acid biosynthesis pathway (P=0.004) consistent

with an increase in the order Clostridia and Ruminococcaceae species (190).

Table 3.4 Significant KEGG pathways comparing AS to HC as identified by PICRUSt analysis

P<0.02.

KEGG Pathway name AS to HC

Pvalue Bacterial invasion of epithelial cells 0.0001 Electron transfer carriers 0.0009 Fluorobenzoate degradation 0.0021 Butirosin and neomycin biosynthesis 0.0021 Germination 0.0021 Transcription related proteins 0.0030 Caprolactam degradation 0.0030 alpha-Linolenic acid metabolism 0.0030 Secondary bile acid biosynthesis 0.0041 Primary bile acid biosynthesis 0.0041 Ether lipid metabolism 0.0041 Flavone and flavonol biosynthesis 0.0057 Flavonoid biosynthesis 0.0057 Phenylpropanoid biosynthesis 0.0057 Glycan biosynthesis and metabolism 0.0076 Linoleic acid metabolism 0.0076 Adipocytokine signaling pathway 0.0101 Phosphonate and phosphinate metabolism 0.0101 Sporulation 0.0101 Sphingolipid metabolism 0.0101 Polycyclic aromatic hydrocarbon degradation 0.0101 Pathogenic Escherichia coli infection 0.0113 Carotenoid biosynthesis 0.0133 Cyanoamino acid metabolism 0.0133 Phosphotransferase system (PTS) 0.0172 Amoebiasis 0.0172 Bisphenol degradation 0.0172 Photosynthesis 0.0172 Photosynthesis proteins 0.0172 Cell cycle - Caulobacter 0.0172 Other glycan degradation 0.0172

88

Further interrogating the AS microbiome signature, which is the overall combination of

microbes which distinguishes AS from HC, by seeing if there was an assemblage or ‘core’

microbes present in all AS and HC samples at varying levels (127) and to probe their

functional capacity. Exploring the core microbiome is imperative, to better understand the

stable and consistent components across complex microbial communities, given the

distinct AS microbial signature. We began by looking at Core 100, which requires microbes

to be present in all samples 100% of the time, which included Clostridia,

Actinomycetaceae, Bacteroidaceae, Lachnospiraceae, Porphyromonadaceae,

Rikenellaceae, Ruminococcaceae and Veillonellaceae families of bacteria (Table 3.5).

These families of bacteria are also AS indicator species suggesting that the core

microbiome is driving the AS microbial signature (Figure 3.4).

Table 3.5 Significant KEGG pathways comparing in Core100 microbiome comparing AS to

HC as identified by PICRUSt analysis P<0.02.

KEGG Pathway name AS to HC

Pvalue Basal transcription factors 0.0003 Atrazine degradation 0.0004 Renal cell carcinoma 0.0015 Flavone and flavonol biosynthesis 0.0021 Germination 0.0021 African trypanosomiasis 0.0041 Bacterial invasion of epithelial cells 0.0041 Bladder cancer 0.0041 Chagas disease (American trypanosomiasis) 0.0041 Butirosin and neomycin biosynthesis 0.0076 Electron transfer carriers 0.0076 Phenylpropanoid biosynthesis 0.0079 Carotenoid biosynthesis 0.0101 Caprolactam degradation 0.0101 Flavonoid biosynthesis 0.0101 Pathogenic Escherichia coli infection 0.0126 Shigellosis 0.0126 Fluorobenzoate degradation 0.0133 Primary bile acid biosynthesis 0.0133 Secondary bile acid biosynthesis 0.0133 Alpha-Linolenic acid metabolism 0.0133 Sphingolipid metabolism 0.0172

89

Given that these microbes were present in all samples 100% of the time, we were

interested in their function, the biological pathways they were involved in and whether they

involved in known AS associated pathways. For Core 100 there were 22 significant

pathways (Table 3.5) (P< 0.02) that again included Bacterial invasion of epithelial cells (P=

0.004), antimicrobial production in the Butirosin and neomycin biosynthesis pathway

(P=0.007) and the Secondary bile acid biosynthesis pathway (P=0.013). Members of

Bacteroides, Clostridium and Ruminococcus are known to be involved in cholesterol

metabolism and secondary bile synthesis with this pathway also implicated in colorectal

cancer.

As the threshold for core membership was reduced to 90, 75 and 50% of the time, the

families of bacteria present remained the same however the number of genus and species

within these families increased, especially in the order Clostridiales, Lachnospiraceae and

Ruminococcaceae. This demonstrates the structure of the core microbiome is robust and

that decreasing the threshold only expands the number of genera and species of bacteria

within these families. However, as the core threshold was decreased the number of

significant pathways also decreased suggesting that as the threshold is relaxed, the

expansion of genera and OTUs dilutes pathway signals. This was reflected in the

metagenome prediction with PICRUSt. As the core threshold was relaxed, the number and

types of pathways that were significantly different between AS and HC was reduced.

No significant differences were noted between AS cases and controls in abundance of

bacteria known to be associated with reactive arthritis such as Campylobacter,

Salmonella, Shigella, Yersinia and Chlamydia. Also, there was no significant difference in

abundance of Klebsiella in AS cases, which has been proposed to play a role in triggering

AS (142).

90

3.5 Discussion

Here we present the first characterisation and identification of intestinal dysbiosis is the AS

microbiome using 16S rRNA community profiling of TI biopsies. We performed culture-

independent microbial community profiling of TI biopsies from nine AS patients, nine HC

and a bacterial mock community, and show evidence for a discrete microbial signature in

the TI of cases with AS. A mock community consisting of six bacteria was constructed and

used as an internal positive control for the purpose of determining if the methods used

accurately captured the microbial communities examined (112, 191). All members of the

mock community belong to bacterial families that are commonly found in the gut, with the

exception of Pasteurellaceae. Y. enterocolitica is a known intestinal pathogen belonging to

Enterobacteriaceae family, E. faecalis and B. fragilis are both known and well documented

intestinal bacteria. Corynebacteriaceae and Staphylococcaceae bacterial families, to which

C. renale and S. aureus belong, are also documented residents of the gastrointestinal

tract. H. influenzae is a member of the Pasteurellaceae family and whilst H. influenzae

specifically is not found in the gut, members of the Pasteurellaceae family are found in oral

microbiome and upper respiratory tract. There is increasing evidence of a link between

the oral and gut microbiome, particularly if patients are on proton-pump inhibitors (192),

indicating that microbial communities may not be site specific discrete environment, but

fluid communities so (129, 193, 194). Analysis of the mock community returned the

expected 16S profile dominated by Proteobacteria (Y. enterocolitica and H. influenzae),

Firmicutes (E. faecalis and S. aureus), Bacteroides (B. fragilis) and the low abundant

Actinobacteria (C. renale). This demonstrates that the resulting community profile is

consistent with the known inputs and that the sequencing and analysis methods are

producing accurate profiles. This also demonstrates that the methods used are not

introducing unwanted bias to the community profiling results (195), nor are the less

abundant bacteria being swamped by more dominate bacteria.

Taken together, this demonstrates that the observed microbial profile differences are not

due to differences in overall bacterial quantity between cases with different diseases, but

are qualitative. PCoA was able to show the clear and distinct grouping of AS cases from

HC; larger studies will be required to define the individual bacterial species involved. Four

HC samples were found to group separately and away from the other HC, closer to the

mock community control. Three of the four HC samples contained increased proportion of

Proteobacteria (HC01, HC08 and HC09) and HC04 that had increased levels of

91

Actinobacteria which has been associated with obesity (114, 196), compared to the other

HC samples, skewing their profiles. Increased levels of Proteobacteria can be associated

with issues such as poor sequencing quality, however all samples were processed and

sequenced at the same time, and thus would all have potentially been affected by any

such bias. With only these there HC samples seemingly affected, poor sequencing quality

is an unlikely explanation for the observations. There is limited metadata on the HC

patients with regards to medications, gastrointestinal issues, diet or other co-morbidities,

therefore a biological reason for the increased Proteobacteria cannot be excluded. There

is evidence to suggest that a reduction in overall microbiome diversity can lead to

increased proportions of Proteobacteria (196), with experiments in mice demonstrating

that increased dietary fat leads to the reduction of microbial diversity and the expansion of

Proteobacteria leading to a diet induced dysbiosis (197-199).

To examine if these outlying samples were affecting the clustering seen on PCoA, the four

outlying were removed and the remaining samples reanalysed. Even excluding the four

HC samples, clear clustering of AS samples away from the HC was seen with the

exception of AS02, which upon further investigation also had higher percentage

Proteobacteria compared to the other AS samples. As well as diet induced dysbiosis,

increased Proteobacteria has also been linked to colitis in animal studies (200),

demonstrating there are several possible biological reasons for an increase in this phyla.

Despite the incidence of Proteobacteria in several of the samples, statistically the

relationship between disease status and microbial community composition was confirmed

indicating that the differences between the microbial community compositions were not

due to heterogeneity in the biopsy or lack of technical reproducibility, but that the driving

force is disease state.

Of the seven families of microbes with differences in abundance within the AS microbiome,

Lachnospiraceae, Ruminococcaceae and Prevotellaceace are strongly associated with

colitis and CD (102-104) with Prevotellaceace especially known to elicit a strong

inflammatory response in the gut (104). Further investigations showed that correlations

between these families of bacteria further shape the AS microbial signature, and that these

microbes are present in all AS samples studied suggesting that they are not only driving

the microbial signature, but they are at the core of the AS microbial signature. Increases in

Prevotellaceae and decreases in Rikenellaceae have also recently been reported in the

intestinal microbiome in the HLA-B27 transgenic rat model of spondyloarthritis, suggesting

92

that underlying host genetics may play a role in sculpting the gut microbiome in this animal

model (201).

There is an increase in the diversity of the AS community without an overall change in

microbial load, showing there is not an overgrowth or dominance of a particular microbe

driving the signature. However, murine experiments demonstrate that both the overall

composition of the intestinal microbiome, and the presences and/or absence of specific

microbes can have a substantial impact on host response, regulation of inflammation and

development of intestinal cells. In the K/BxN mouse model of arthritis it was shown that the

introduction of a single gut-residing species, Segmented Filamentous Bacteria, into GF

animals was sufficient to reinstate the Th17 cells, leading to the production of

autoantibodies and arthritis. If IL-17 was neutralised in specific-pathogen-free K/BxN mice,

arthritis development was prevented. Thus, a single commensal microbe, via its ability to

promote a specific T helper cell subset, can drive immune mediated disease (101).

Further studies are needed into whether the changes in intestinal microbial composition

are due to host genetics, and how this affects the overall function of the gut microbiome in

AS cases, including how the microbiome then goes on to shape the immune response and

influence inflammation. In particular, given the strong association of HLA-B27 with AS, the

hypothesis has been raised that HLA-B27 induces AS by effects on the gut microbiome, in

turn driving spondyloarthritis-inducing immunological processes such as IL-23 production

(67, 202). Further experiments comparing the intestinal microbiome of HLA-B27 negative

and positive patients would shed light of the influence of HLA-B27 on overall intestinal

microbiome composition, particularly given the work in B27 transgenic rats showing that

HLA-B27 was associated with altered caecal microbiota (201).

Our data here, showing intestinal dysbiosis in AS cases, is consistent with this hypothesis

however further studies are clearly required to distinguish cause and effect interactions

between the host genome and immune system, and the gut microbiome. These

investigations will provide new insights and help us to better understand AS disease

pathogenesis.

93

4 Chapter 4 Effect of ERAP1 genotype on the intestinal microbiome and the impact on the immune system in the ERAP mouse model

4.1 Introduction

To date, over 40 loci have been associated with ankylosing spondylitis, and one of the

strongest associations, along with HLA-B27, is Endoplasmic reticulum aminopeptidase-1

(ERAP1) (21, 24, 55, 62). Interactions between HLA-B27 and ERAP1 have been

described in ankylosing spondylitis (AS) such that variants of ERAP1 were shown only to

influence AS disease risk in the presence of HLA-B27 (62), suggesting that ERAP1

operates through effects on peptide processing prior to MHC class I presentation. Fine

mapping studies indicate that the ERAP1 single nucleotide polymorphism (SNP) rs30187

is directly associated with disease risk and that this loss of function variant is protective in

AS (62, 187). Therefore AS-associated ERAP1 variants may change either the amount

and length of peptides presented or even influence the stability of the peptide loaded onto

the MHC class I, or the peptide MHC complex and its cell surface retention (203).

Consequently, ERAP1 could influence AS disease risk via changes in the type or amount

of volume of peptide presented. In this chapter we are asking very different questions of

ERAP’s function and role in AS pathogenesis by exploring how loss of ERAP1 impacts on

the gut microbiome and how that subsequently affects the immune system.

Early functional studies suggested two roles for ERAP1, one as a ‘molecular ruler’

trimming peptides prior to presentation on the MHC class I, and the second as a

‘sheddase’, which acts by cleaving cytokine receptors off the cell surface, including IL-6R,

IL-1R2 and tumour necrosis factor receptor (TNFR). However studies in mice showed no

difference in the levels of these receptors over time, indicating that ERAP1 does not have

a major influence on cytokine receptor trimming, at least in mice. Further, ERAP1 SNPs

were shown not to be associated with variation in serum cytokine receptor levels,

confirming in humans that this gene does not play a significant role in cytokine receptor

shedding (204). This strengthens the case that ERAP1 acts as a ‘molecular ruler’ and is

responsible for the trimming of peptides prior to presentation (205). ERAP1 is a

ubiquitously expressed, multifunctional enzyme that trims peptides and antigen for

presentation by MHC class I to the immune system. The endoplasmic reticulum (ER)

94

localised ERAP1 is critical in the MHC class I presentation pathway. Once peptides have

been process through the proteasome, the Transporter associated with Antigen

Processing (TAP) takes the subsequent peptide from the cytoplasm into the ER. ERAP1

then further trims any N-terminally extended peptides to no longer than 9 amino acids in

length which generates or destroys antigenic epitopes prior to loading onto MHC class I

molecules (187, 206-208). Interaction between ERAP and the AS associated MHC class I

HLA-B27, was the focus of a recent study by Akram et al., 2014. The authors suggest that

ERAP may play a role in stabilising the heavy chain B27 during assembly of the MHC

class I by presenting it with appropriate B27-specific peptides(209). Interestingly, this

stabilisation seems to be B27 specific. When compared to B7, presence/absence of ERAP

had no effects on host immune response, whereas the absence of ERAP significantly

altered the immune response to infection with B27 was present (209).

The establishment and maintenance of beneficial interactions between the host and their

microbiota are key requirement for overall health. The bacteria that inhabit the gut and the

human immune system have both developed a number of strategies to maintain gut

homeostasis. The gastrointestinal tract is dynamic environment heavily populated with

microbes and is the primary site for interaction between microbes and the host immune

system. Furthermore, the microbes found in the gut help to shape host immune systems

from an early age (210, 211). The incomplete development of the immune system in

neonates and mice raised under germ free conditions, tells us that microbiota sculpt the

host immune system (28, 212). The dysregulation of either gut or microbial homeostasis is

increasingly recognised as playing a pivotal role in autoimmunity. Fluctuations in microbial

composition due to changes in diet, environment, hygiene and use of antibiotics, presents

a major challenge to the immune system (30). Controlling activation of the immune system

in response to these changes in gut microbiota has to be tightly regulated as to avoid

chronic inflammation. The body has a number of physiological processes that respond to a

disruption in gut homeostasis or a bacterial insult, involving both the adaptive and innate

immune system, and the barrier function of the intestines themselves (94, 213).

4.1.1 Peptide presentation and infection Aminopeptidases play integral roles in modulating the adaptive immune responses to both

host and pathogen derived antigens. Unlike humans, mice only carry one isoform of

ERAP1, which will be referred to from here as ERAP. Recent animal studies have

95

revealed how ERAP shapes the antigenic peptide repertoire against microbial and viral

infections. This is dramatically demonstrated in ERAP-/- mice infected with the obligate

intracellular parasite Toxoplasma gondii, that causes toxoplasmosis. Mice lacking ERAP

are unable to trim and present immunodominant and appropriate length peptides derived

from T. gondii. This leads to aberrant peptide presentation, an inability to mount an

effective immune response to infection with fatal consequences (214). ERAP-/- mice

infected with lymphocytic choriomeningitis virus (LCMV) were shown to present drastically

different immunodominant epitopes derived from LCMV for presentation than wild type

(WT) mice (215). ERAP’s role in modulating the immune response to viral pathogens is

further validated by the finding that human cytomegalovirus (HCMV) has evolved an

immune evasion strategy where by a micro RNA miR-US4-1 is expressed and specifically

targets ERAP, down regulating expression and altering peptide trimming and presentation

as to prolong viral infection (206). Down regulation or the loss of ERAP not only affects

peptide presentation during infection, it also significantly alters the composition of the

endogenous MHC class I peptidome. Murine studies have shown that a lack of ERAP, in

the absence of infection, presents unstable and structurally unique peptide-MHC

complexes and that these complexes elicited a potent CD8+ T cell response (207, 208).

Antigen presenting cells (APC), such as dendritic cells (DCs), internalise antigens from the

environment, process and present these exogenous antigens to MHC class II for

presentation to CD4+ T cells. However, many viruses and bacteria have developed a

number of evasion strategies and presentation of peptide from DCs to MHC class I

mitigates this by eliciting a potent CD8+ T cell immune response. This cross presentation

is critical for immune surveillance defence against viruses, bacteria and even tumour cells

(216). Peptides presented by DCs in this manor are processed mainly via the phagosome

to cytosol pathway. There is recent evidence to suggest that peptides generated in the

cytosol may be transported by transporter associated with antigen processing (TAP) to

MHC class I in the ER via ERAP1, or may do so in phagosomes (214). One phagosome

peptidase of interest is Insulin regulated aminopeptidase (IRAP), a homolog of ERAP1,

which also trims amino acids from the N terminus and its important for cross presentation

(217). Studies have shown that DCs that lack IRAP are capable of presenting antigen

normally however, their ability to cross present antigen is substantially reduced; not unlike

what is seen with the loss of ERAP1 (217). This further highlights the vital role ERAP1

plays in modulating the adaptive immune responses to both host and pathogen derived

antigens.

96

4.1.2 Intestinal cytokines and gut homeostasis

IL-23 is a key cytokine in the development of IL-17 and IL-22-secreting cells. IL-23 signals

through a receptor consisting of the specific IL-23 receptor p19 subunit (IL-23R) and the

IL-12 shared subunit p40 (54). Loss of function polymorphisms in IL23R are associated

with protection from AS (24, 55), psoriasis (56) and IBD (23, 57), and there are many other

genes in the IL-23 pathway that are known to be associated with these diseases. IL-23, IL-

17 and IL-22 producing cells are enriched in intestinal mucosa and have been shown to be

overexpressed in the intestinal mucosa of AS patients with sub-clinical gut disease (218).

IL-17 and IL-22 work together in concert to regulate intestinal health with IL-17 playing

important roles in intestinal homoeostasis in several ways including maintenance of

epithelial barrier tight junctions (58); as well as the induction of anti-microbial proteins such

as β-defensins, S100 proteins and REG proteins. Innate IL-17 producing cells have been

found to reside mainly in the skin and mucosal tissues, especially the gastrointestinal tract,

and act as a first line of defence. IL-17 released during an inflammatory response is

produced by innate immune cells in response to injury, stress or to pathogens (91). IL-17

has been shown to be produced 4-8hrs after a microbial infection and is secreted by T

cells including T helper cells (Th17) and γδ T cells (63, 92, 219). IL-22 is known to induce

secretion of anti-microbial peptides (98) and has been shown to play a critical role in host

defence, tissue homeostasis and inflammation with the IL-22-IL-22R pathway contributing

to regulating immunity, inflammation and tissue repair. IL-22 is expressed by innate and

adaptive cells and seems to act almost exclusively on non-haematopoietic cells, with basal

IL-22R expression in the skin, pancreas, intestine, liver, lung and kidney (94). IL-22 can

act synergistically or additively with IL-17A, IL-17F or tumour necrosis factor (TNF) to

promote the expression of many of the genes that encode molecules involved in host

defence in the skin, airway or intestine. This demonstrates the functional importance of IL-

22 in promoting barrier immunity and mucosal integrity, particularly against Gram negative

pathogens, where IL-22 is critical for limiting bacterial replication and dissemination,

possibly in part by inducing the expression of antimicrobial peptides from epithelial cells at

these barrier surfaces (98, 99). Containment of normal flora is another important arm of

homeostasis. IL-22 producing innate lymphoid cells (ILCs) found throughout the gut, play

a critical role in the physical containment of resident microbes (100). Loss of epithelial

integrity leads to dissemination of intestinal commensals and systemic inflammation.

97

The initial immunological characterisation of the IL-23/Th17 axis was performed largely in

the collagen-induced arthritis (CIA) mouse model of autoimmune arthritis, which

phenotypically resembles rheumatoid arthritis (RA). In the CIA model, mice that lack IL-23

are protected against the development of disease. In IL-17 and IL-22 deficient mice CIA is

suppressed, implicating the IL-23 pathway in pathogenesis of disease (220, 221).

4.1.3 Interactions between intestinal microbes and the immune system

A number of immune cells play critical roles in inflammation, cytokine production and

homeostasis within the gut. Natural killer T (NKT) cells are protective in Th1-mediated

models of inflammatory bowel disease but pathogenic in Th2 models (73, 74). Recently, it

has been shown that microbial stimulation of NKT cells in the gut of mice affects NKT cell

phenotypic and functional maturation (75). NKT cells recognise glycolipd structures

presented to them by the non-classical antigen-presenting molecule CD1d. NKT cells are

rapid responders to antigenic stimuli and are capable of producing a range of

immunoregulatory cytokines. Given that NKT cells have protective roles in models of

arthritis (76) and spondyloarthropathy (77, 78) their functional maturation in the gut

provides evidence for a role for mucosal T cell priming in inflammatory joint disease.

Lymphoid tissue inducer (LTi)-like cells are found in spleen, lymph node and gut lamina

propria. LTi-like cells constitutively express hallmarks of IL-17-secreting cells including IL-

23R, RORγt, AHR and CCR6 (86, 87). A common characteristic of ILCs, specifically ILC3,

is the expression RORγt, a transcription factor important for the development of these

cells. ILCs were initially described as important for development of lymphoid tissues and

more recently in the initiation of inflammation at barrier surfaces in response to infection or

tissue damage. It has become apparent that the family of ILCs (Table 4.1) plays more

complex roles throughout the duration of immune responses, participating in the transition

from innate to adaptive immunity across varying insults and contributing to chronic

inflammation. ILCs have been linked to gut inflammation through colitis models where IL-

23 responsive ILCs secrete IL-17, IL-22 and IFN-γ and to promote intestinal inflammation

(89). RORγt expressing ILC3s are abundant in the intestinal lamina propria and produce

IL-17 and/or IL-22 in order to preserve mucosal integrity against extracellular pathogens

and maintain equilibrium between host and microbes. These NKp46+ ILCs have been

found to be critical for host defence against pathogens such as Citrobacter rodentium

infection through secretion of IL-22 (87, 90).

98

Table 4.1 Innate lymphoid cells including cytokine expression and immune function.

Classification Transcription

factors

Cytokine expression Immunity roles

ILC group 1 (NK cells)

IFNγ, TNF-α Bacteria, intracellular parasites, autoimmune disorders,

ILC group 2 (Natural helper cells)

RORα, GATA 3 IL-5, IL-9, IL-13 Parasitic worms, allergic disease, obesity,

ILC group 3 (Lymphoid tissue inducers (LTi)

RORγt IL-22 and/or IL-17 Bacteria, autoimmune disorders, homeostasis

γδ T cells IL-17, IFN-γ, TNF-α Bacteria, mediate innate and adaptive immunity

Taken together, this suggests ERAP1 plays a critical role in overall immune system

modulation with systemic effects. In this chapter we are asking how a lack, or down

regulation, of ERAP1 influences the inherent gut microbiome composition, and how this

impacts on the antigen repertoire and consequently the local and systemic immune

system. We hypothesise that the loss of ERAP1 leads to altered T cell selection, impairing

immune recognition of inherent gut microbiota, and leading to an overall shift in microbial

community composition.

4.2 Methods

4.2.1 Antibodies and Flow Cytometry

4.2.1.1 Mice

C57BL/6 mice were purchased from Animal Resources Centre, Perth, Australia. The

generation of ERAP-/-, which are on a C57BL/6 background, has been described

previously (207). ERAP-/- mice were a kind gift from Dr N. Shastri (University of California,

Berkeley). All mice were housed in the specific pathogen–free animal facility and fed

standard rodent chow at Translational Research Institute (TRI) Biological Research

Facility, according to the guidelines of the Australian NHMRC code for ethical use of

animals. All animal studies have been reviewed and approved by The University of

Queensland Animal Ethics Committee. F2-intercross mice were bred to generate

littermates that were ERAP-/- (KO), ERAP+/- (HET) or ERAP+/+ (WT). To minimise cage-

99

effects on the gut microbiome, littermates were raised together throughout the experiment,

and analyses were adjusted for cage.

4.2.1.2 Antibodies

Antibodies against CD4 (GK1.5), CD3 (145-2C11), CD69 (H1.2F3), CD44 (IM7), CD335

(29A1.4), CD25 (PC61), CD11b (M1/70), NK1.1 (PK136), IFNγ (XMG1.2), IL-10 (JESS-

16E3) were purchased from Biolegend (San Diego, CA). Antibodies against γδ (eBioGL3),

RORγt (AFKJS-9), FOXP3 (FJK16s) and IL-22 (IH8PWSR) were obtained from

eBiosciences (San Diego, CA). IL-17 (TC11-18H10) and TNF (MP6-XT22) were

purchased from BD Pharmagin (San Diego, CA).

4.2.1.3 Cell preparation

Spleen and mesenteric lymph node (MLN) cells were prepared from freshly harvested

ERAP-/- and C57BL/6 mice. Whole spleens and MLN were disrupted and strained using

70mL filters (BD Pharmingen) to create single cell suspensions. Splenic cells were lysed

using ACK lysis buffer for 2 minutes. Cells were washed in PBS prior and half transferred

to 48-well plate for stimulation with PMA (phorbol myristate acetate, AKA phorbol ester,

Biochimika) and Ionomycin (Ca2+ salt, Calbiochem) as previously described (222), with

the other half being used for activation marker and transcription staining.

Splenic and MLN cells were stained for extracellular markers, transcription factors and

intracellular cytokines. For detection of T cell markers cells were stained for CD3, CD4,

CD8 and γδ. For detection of non-T cells, NK1.1 and CD11b were used. Activation status

was determined with CD69, CD44, CD335 and CD25. Transcription factors were assessed

through FoxP3 and RORgt. Cells were fixed (Fixation Buffer, Biolegend), permeabilised

(Perm/Wash buffer, BD Pharmingen) and stained for intracellular cytokines TNF, IFNγ, IL-

10, IL-22 and IL-17. Cytometric data was acquired on the Beckman Coulter Gallios

cytometer.

4.2.1.4 Statistics

All data are presented as the mean ± standard error of the mean (SEM). Data were

compared by unpaired T test, using GraphPad Prism 6 (GraphPad Software). P≤0.05 were

considered significant. Power calculations were computed using an on-line power

100

calculator (http://www.statisticalsolutions.net/pss_calc.php) with effect size and Cohen’s d

computed as outlined by Olejnik and Algina, 2003 (223).

4.2.2 Sequencing and analysis

4.2.2.1 DNA extraction Faecal pellets were collected and DNA extraction was performed using the Qiagen

DNeasy tissue extraction kit as per manufacturer’s instructions.

4.2.2.2 Mock community

A mock community consisting of seven different bacteria, Gram positive and Gram

negative, was employed as a control for both sequencing and bioinformatic analysis and

was processed alongside the mouse faecal samples. The seven bacteria used were

Staphylococcus aureus (ATCC 25923), Enterococcus faecalis (ATCC 51299),

Haemophilus influenzae (ATCC49247), non-mucoid Pseudomonas aeruginosa (ATCC

27853), Klebsiella oxytoca (ATCC 700324), Escherichia coli (ATCC 35218) and Yersinia

enterocolitica (ATCC2 3715). Bacteria were extracted using the Qiagen DNeasy tissue

extraction kit as per manufacturers instructions. The community was then labelled Mock1.

4.2.2.3 16S rRNA PCR

Bacterial 16S rRNA amplicon PCR was conducted using V4 primers spanning the region

517F 803R of the 16S rRNA gene. The master mix for the reactions was conducted in a

25ul total volume using Bioline 10x reaction buffer, dNTP(0.625mM), MgCl2 (2.0mM), 1U

of Bioline TAQ, both primers at 5uM each and 2ul DNA. Thermocycling conditions

comprised 10 min at 94ºC then 30 cycles at 94ºC for 50 s, 54º for 30 s, 72ºC for 45 s and a

72ºC single step for 10 min.

4.2.2.4 Sequencing Library preparation

16S rRNA amplicon libraries were assembled and dual indexed for multiplexing using the

Nextera XT dual indexing library protocol as per manufacturer’s instructions. Samples

were then normalised and pooled for sequencing at a concentration of 2nM. To create the

required complexity for sequencing, PhiX was added to the pool at 75/25 for a final pool

concentration of 3.5pM.

101

4.2.2.5 Illumina MiSeq Sequencing

Dual barcoded 16S amplicon sequences were sequenced using 2x250 base paired-end

protocols on the Illumina MiSeq. There were 6.1million reads generated overall with an

average of 200,000 reads per sample.

4.2.2.6 Bioinformatics and statistics

Samples were demultiplexed using CASAVA. Reads were quality filtered for Q30 and

above using split_library.py and then analysed using the QIIME pipeline

(qiime.sourceforge.net). OTU assignment was based on 97% similarity with taxonomic

assignments made using BLAST to Greengenes as implemented in QIIME (Caporaso et

al., 2011).

Alpha diversity metrics were generated using the QIIME workflow alpha_rarefaction.py

with five repetitions at each sampling point. Rarefaction curves for the observed number of

OTUs were based on the outputs of alpha_rarefaction.py. UniFrac distances between

microbial communities were evaluated using FastUniFrac (Louzapone and Knight 2005).

UPGMA clustering was employed both weighted and unweighted, to detect significant

differences between microbial communities. UniFrac enabled in QIIME was used to

generate sample distance metrics as well as Principal Co-ordinate Analysis (PCoA) (189).

PERMANOVA analysis was conducted at the genus level using the R package ‘vegan’

(version 3.0.2) to test the relationship between the whole microbial community, genotype,

cage and litter effect (176). Co-occurrence network analysis was conducted using the

‘Spaa’ package in R (version 3.0.2).

4.3 Results

4.3.1 Alterations in host genetics influences gut microbiome composition

The effect of underlying host genetics on the microbiome structure and function, especially

in AS, is still poorly understood. Taking a reductionist approach, we examined the effect

that loss of ERAP1 has on the gut microbiome and subsequent effects on the local and

systemic immune system using a mouse model of ERAP-/-. The loss of ERAP1 is

protective in AS, we therefore explored if the loss of ERAP in a mouse model caused a

shift in the intestinal flora and if that shift was anti-inflammatory in nature.

102

Faecal pellets from eight ERAP-/- and 11 wild-type C57BL/6 mice were collected and

bacterial DNA was extracted, amplified for microbial 16S rRNA and sequenced. This proof

of principle study aimed to determine if the loss of a single gene in is sufficient to cause a

shift in the intestinal microbiome.

16S rRNA community profiling of the mouse gut microbiome revealed distinct clustering of

communities separating ERAP-/- from wild-type C57BL/6 mice (Figure 4.1). Microbial

communities were compared using Weighted and Unweighted UniFrac distances, a

phylogenetic metric that accounts for differences in both community membership and

relative abundance (174). These distance metrics allow us explore whether it is the

presence or absence of bacteria that is driving the overall community structure, or if it is

simply a change in the relative abundances of particular microbes in the community (189).

Unweighted UniFrac analysis indicated that overall structure, the presence or absence of

bacteria, was driving the distinct clustering between these mice. The Phylum level

differences between C57BL/6 and ERAP-/- (Figure 4.1B) included changes in the

Bacteroides and Firmicutes ratio that constitutes fundamental changes in community

structure, driving the clustering on PCoA (Figure 4.1A). Several of the species observed

driving the clustering of these mice are also indicator species driving the clustering of

microbial communities in terminal ileum of AS patients as described in Chapter 2 of this

thesis (Table 4.2). Whilst there are environmental effects such as background and cage

effect, this suggests that a single change in underlying host genetics may be sufficient to

influence microbial community structure in the gut.

103

Figure 4.1 Principal coordinate analysis and taxonomy showing community differences between ERAP-/- mice and control C57BL/6. (A) Unweighted Principal Coordinate Analysis (PCoA) showing distinct clustering of ERAP-/- mice from control C57BL/6. (B) Bar chart showing the difference in taxonomy at the Phylum level between C57BL/6 and ERAP-/-, noting the increase in Bacteroides in C57BL/6 and the change in the Bacteroides to Firmicutes ratio between the samples.

Table 4.2 Differences in average percentage OTU counts between C57BL/6 and ERAP-/- mice showing that changes in proportions of key bacteria contribute to the ERAP-/- microbial signature.

Taxonomy C57BL/6

(OTU Ave)

ERAP-/-

(OTU Ave)

Difference

Result

Bacteroidaceae

(genus Bacteroides)

0.008 0.024 0.016 Increase in ERAP-/-

Prevotellaceae

(genus Prevotella)

0.012 0.008 -0.004 Decrease in ERAP-/-

Rikenellaceae

(genus Rikenella)

0.011 0.028 0.017 Increase in ERAP-/-

S24-7

(order Bacteroidales)

0.680 0.442 -0.238 Decrease in ERAP-/-

Lactobacillaceae

(genus Lactobacillus)

0.032 0.014 -0.018 Decrease in ERAP-/-

Clostridiaceae

(genus Clostridium)

0.129 0.301 0.172 Increase in ERAP-/-

Lachnospiraceae 0.031 0.062 0.031 Increase in ERAP-/-

Ruminococcaceae 0.007 0.016 0.009 Increase in ERAP-/-

Desulfovibrionaceae

(genus Desulfovibrio)

0.024 0.012 -0.012 Decrease in ERAP-/-

104

Thus the overall structure of the gut microbiome in the ERAP-/- mice was different to

control C57BL/6. Additionally, key indicator species also seen in AS TI biopsies shown to

be driving the clustering between mouse strains differing in AS-associated genetic

makeup. Therefore I sort to explore how the microbial community members are interacting

and how these interactions shape the community structure and signature. Co-occurrence

analysis, which looks at the correlation between microbes, was undertaken to explore

correlations between keystone members of the ERAP-/- intestinal community compared to

controls. Positive correlations, shown with blue lines, were seen between the indicator

families Lachnospiraceae, Prevotellaceae, Clostridiaceae and Rikenellaceae with negative

correlations, shown with red lines observed between S24-7, Prevotellaceae,

Ruminococcaceae (Figure 4.2). The thickness of the line denotes the strength of

correlation between the bacteria and demonstrates that bacterial interactions further shape

the ERAP-/- microbial community signature (Table 4.2).

Figure 4.2 ERAP-/- Intestinal community Co-occurrence network showing the positive (blue lines) and negative (red lines) correlations between indicator species. The thickness of the line denotes the strength of the correlation (Pearson correlation).

Dissecting the effects of genetic variation, in this case comparing ERAP-/- with C57BL/6

mice, it is important to consider that these experiments can be subject to variation due to

other factors such as strain differences, cage effect and founder effect may be driving the

105

differences seen between these mice (165). To minimise such potential confounders, we

explored the gut microbiome differences within a single litter of ERAP mice comprising of

ERAP-/-, ERAP+/- and ERAP+/+. Given that these mice were the offspring of sibling

ERAP HET founder mice, were from the same litter, and were cohoused from birth,

variation due to background strain and cage/litter effects were minimized or excluded.

Faecal pellets were collected and processed as described previously.

Figure 4.3 Unweighted Principal coordinates analysis (PCoA) of a single litter of ERAP mice comprising three ERAP+/+, four ERAP+/- and one ERAP-/- showing distinct grouping of ERAP+/+ away from the other genotypes and separate clustering of ERAP-/- and ERAP+/-.

The unweighted PCoA, which shows structural differences between microbial communities

(Figure 4.3), illustrating that genotype does alter the structure of the intestinal microbiome

(P=0.001; PERMANOVA). The three ERAP+/+ mice, our controls, are seen to cluster

together and away from the ERAP+/-and the ERAP-/- mice. As expected, the ERAP+/-

and the ERAP-/- cluster away from the WT and are more similar to each other than to the

ERAP+/+ mice. It is apparent that a change in genotype, from ERAP+/+ to ERAP+/-, is

enough to influence the community structure of the microbiome. The clustering of the

ERAP+/- and the ERAP-/- together, but separately from the ERAP+/+, suggests that

change of one allele is sufficient to alter the microbiome structure. Further litters will need

to be studied to confirm this apparent dominant effect.

106

4.3.2 Loss of ERAP does not change overall proportions of Immune cell subsets

Given that we observed that the loss of a single gene influences gut microbiome

composition, and the gut is the primary interface between microbes and the immune

system, we were interested in how changes in the microbiome affect the immune system

both locally and systemically. We began by examining the major immune cell subsets in

eight ERAP-/- and eight ERAP+/+ mice to determine if an altered gut microbiome altered

the immune system (Figure 4.4). Following removal of doublets, dead cells were excluded

using Aqua Viability dye discrimination. Cell subset specific analysis of cell phenotype and

activation status was performed by analysing cell surface protein expression. Transcription

factor expression was determined intracellularly following cell permeabilisation. Cytokine

production was determined intracellularly following in vitro stimulation for 4 hours with PMA

and ionomycin in the presence of the Golgi-transport blocker monensin.

Figure 4.4 Example gating strategy for isolation, identification and characterisation of T cells. Doublets were removed and dead cells were excluded using Aqua Viability dye discrimination. Cell subset specific analysis of cell phenotype and activation status was performed by analysing cell surface protein expression, with transcription factor expression determined intracellularly following cell permeabilisation. Cytokine production was determined intracellularly following in vitro stimulation.

107

A post hoc power calculation indicated our study had 80% power to detect an effect with

Cohen’s d >0.2 α =0.05 between ERAP-/- and ERAP+/+ cells isolated from the spleen,

however our results indicated effect sizes of Cohen’s d <0.34. In cells derived from the

MLN our study had 80% power to detect an effect size Cohen’s d >1 α =0.05 and our

results indicated effect sizes with Cohen’s d >1.3. No significant differences were seen in

overall numbers of T cells (CD3+, CD4+, CD8+ or γδ), NK cells (CD3- NKp46+),

Macrophages (CD11b+) or regulatory T cells (CD4+ CD25+) either systemically in the

spleen or locally in the MLN (Figure 4.5, Table 4.3).

Table 4.3 Frequencies of cell subsets showing no significance between ERAP+/+ (WT) and ERAP-/- (KO) strains in spleen and mesenteric lymph nodes.

Spleen Mesenteric lymph nodes

Immune cell subset

PValue Mean± SEM WT

Mean± SEM KO

PValue Mean± SEM WT

Mean± SEM KO

CD3+ >0.1

15.08± 1.197

14.74± 1.514

>0.05

32.62 ± 1.129

27.83 ± 2.432

CD4+ >0.1

22.99± 2.110

22.84± 0.733

0.054

44.02 ± 2.546

37.39 ± 1.865

CD8+ >0.1

16.24± 1.314

14.97± 0.499

>0.1

22.77 ± 1.435

21.74 ± 1.314

γδ >0.1

0.66± 0.061

0.57± 0.045

>0.1

0.77 ± 0.069

0.76 ± 0.063

NK cells >0.1

3.11± 0.240

2.98± 0.343

>0.1

21.30 ± 2.603

22.27 ± 1.283

Macrophages >0.1

6.056± 0.956

4.93± 0.540

>0.1

3.34 ± 1.213

3.78 ± 1.091

Treg cells >0.1

0.96± 0.110

0.96± 0.083

>0.1

3.07 ± 0.218

3.46 ± 0.696

There is trend towards fewer CD3+, CD4+ T cells and NK cells in the MLN, this is

countered by a slight increase in regulatory T cells. A non-significant reduction in overall

numbers of Macrophages and CD3+ NKp46+ cells and increase in NK cells in the spleen

of ERAP-/- is observed. There were no other notable differences in any of the immune cell

subsets investigated suggesting that at least systemically, the loss of ERAP does not lead

to significant shifts or changes in the T cell and non T cell compartment. There were no

differences functionally in T cell activation as measured by CD69 expression (Figure 4.5).

CD69 is a marker of early activation and involved in inducing T cell proliferation and

108

lymphocyte migration, and is normally expressed at low levels on resting CD4+ or CD8+ T

cells. Increased activation in CD4+ and CD8+ T cells in either the spleen and MLN of

ERAP-/- mice, may have suggested T cell activation and mobilisation, possibly in response

to the shift in gut flora. Together this data suggests that the effect of the loss of ERAP on

the immune system is not systemic; the effect may be more subtle and localised to the

intestinal tissue at the vanguard of immune system and microbe interaction.

Figure 4.5 Frequencies of major immune cell subsets in ERAP+/+ (WT) and ERAP-/- (KO) from the Spleen (A) and mesenteric lymph nodes (C) and activation status respectfully (B and D) shown as mean ± SEM and compared by unpaired t test.

109

4.3.3 Loss of ERAP leads to differences in cytokine potential

Although the intestinal flora of ERAP-/- mice are skewed more towards containing families

of bacteria known to elicit a strong immune response, it is notable that there is minimal

evidence of an immune response to this shift in flora. There are no detectable differences

in the immune cell subsets either in the spleen or the MLN, so the lack of response is not

due to an absence of immune cell subset. Therefore, we were interested to see if there

were differences in how the immune cells responded to the intestinal microbes, and the

cytokines produced.

There was no change observed in overall numbers of immune cells between ERAP-/- and

C57BL/6 control mice. However, there was a trend observed in ERAP-/- mice away from

an inflammatory IL-17 dominated environment, towards increased levels of IL-22 creating

a less inflammatory and a more homeostatic gut environment.

In the spleens of the ERAP-/- mice there is a non-significant trend towards a reduction in

the amount of IL-17 expressed especially from CD3- and CD4+ T cells (Figure 4.6).

However, there is a statistically significant decrease in production of IL-17 from NK cells

(P=0.04) in ERAP-/- mice, suggestive of a response by the innate lymphoid system in

response to the microbiome changes (Figure 4.6, Table 4.4). With a reduction of IL-17,

there would be an expected increase in the homeostatic cytokine IL-22. However there

was no IL-22 detected from CD8+ T cells or γδ T cells. There were no differences in

expression of IL-10, TNF or IFNγ (Figure 4.6, Table 4.4).

In the MLN, IL-17 expression was mixed across the immune subsets however there was

an increase in expression in macrophages, although not significant (Figure 4.7, Table 4.5).

There was an increase of IL-22 expression from CD3+ T cells, although not significant,

and unexpectedly a decrease across all other T cell and non T cell subsets. This

demonstrates that the IL-22 is being expressed predominately from CD3+ T cells with a

slight increase in IL-10 expression from macrophages, adding to the homeostatic

environment (Figure 4.7, Table 4.5).

110

Table 4.4 Frequencies of cytokine expression from Immune cell subsets in the spleen of ERAP-/- (KO) and ERAP+/+ (WT) mice. There was no significance in expression between ERAP WT and ERAP KO strains in spleen as determined by unpaired t test.

Spleen Immune subsets

IL-17 IL-22 IL-10 TNF

PValue Mean± SEM WT, KO




CD3+ >0.1

0.052± 0.019, 0.061± 0.015

>0.1

6.220± 1.164, 5.800± 1.251

>0.1

1.043± 0.437, 0.940± 0.796

>0.1

0.262± 0.092, 0.636± 0.348

CD3- >0.05

28.55± 5.290, 15.93± 4.043

- - - - >0.1

2.615± 0.528, 5.583± 3.012

CD4+ >0.1

16.44± 1.131, 13.82± 2.155

>0.1

0.491± 0.204, 0.218± 0.068

- - >0.1

21.86± 5.255, 17.86± 3.030

CD8+ - - - - - - >0.1

9.890± 2.858, 7.486± 1.601

γδ >0.1

4.183± 2.108, 6.671± 1.671

- - - - >0.1

9.519± 2.199, 10.23± 1.239

NK cells 0.0397

0.698± 0.279, 0.061± 0.026

>0.1

8.346± 2.317, 9.755± 3.467

>0.1

1.948± 0.375, 2.105± 0.439

>0.1

17.14± 2.191, 17.30± 4.316

Macrophage >0.1

1.501± 0.645, 1.369± 0.318

>0.1

12.16± 1.545, 11.25± 2.608

>0.1

3.983± 0.807, 3.098± 0.480

>0.1

30.05± 2.961, 28.35± 4.323

111

Figure 4.6 Immune cell subsets and expression of cytokine from the spleen of ERAP-/- (KO) mice compared to ERAP+/+ (WT). (A) Expression of IL-17 in the spleen, noting the increase in secretion from γδ T cells, and the absence of IL-17 secretion from CD3+ T cells as a whole and CD8+ T cells. (B) Expression of IL-22 showing a slight decrease across all cell types except for NK cells and no significant differences in expression of either IL-10 (C) or TNF (D). Data shown as mean ± SEM and compared by unpaired t test.

112

Figure 4.7 Immune cell subsets and expression of cytokine from the mesenteric lymph nodes (MLN) of ERAP-/- (KO) mice compared to ERAP+/+ (WT). (A) Expression of IL-17 in the MLN, showing an increase in secretion from macrophages. (B) Expression of IL-22 showing an increase in secretion from CD3+ T cells, but undetectable levels across other T cell subsets. A slight increase in IL-10 secretion was detected in macrophages (C) and TNF secretion in the MLN mirrored what was seen in the spleen (D).

113

Table 4.5 Frequencies of cytokine expression from immune cell subsets in the MLN. There was no significant difference in expression between ERAP+/+ (WT) and ERAP-/- (KO) strains in spleen as determined by unpaired t test.

MLN Immune subsets

IL-17 IL-22 IL-10 TNF





CD3+ >0.1

0.008± 0.006, 1.244± 1.221

>0.1

5.150± 1.253, 5.536± 1.956

>0.1

0.7500± 0.3543, 1.619± 1.069

>0.1

3.061± 1.368, 3.068± 0.823

CD3- >0.1

20.03± 3.159, 18.96± 4.603

- - - - >0.1

8.066± 2.134, 8.850± 2.024

CD4+ >0.1

0.976± 0.656, 0.926± 0.277

>0.1

- - - - -

CD8+ - - - - - - >0.1

29.16± 4.637, 33.87± 8.483

γδ >0.1

4.083± 0.951, 3.334± 0.661

- - - - - -

NK cells 0.059

7.023± 1.794, 3.023± 0.767

0.013

2.675± 0.562, 0.921± 0.254

>0.1

11.80± 3.954, 10.59± 1.788

0.055

20.92± 4.451, 11.31± 1.186

Macrophage >0.1

21.67± 3.942, 23.71± 3.701

>0.1

4.595± 1.117, 2.455± 0.997

>0.1

66.10± 8.980, 70.75± 7.556

0.054

51.38± 9.148, 31.34± 2.772,

4.4 Discussion

The association of variants of ERAP1 is one of the strongest non-MHC associations seen

in AS. In this chapter I investigated if the loss of a single AS-associated gene was

sufficient to alter the intestinal microbiome and the immune response. Given that loss of

function genetic variants of ERAP1 in AS are protective, and that an estimated 70% of AS

patients have either clinical or sub clinical gut disease, we explored how a lack or down

regulation of ERAP1 influences the inherent gut microbiome composition, and given the

gut microbiome is the educator of the immune system, the subsequent impact this has on

the antigen repertoire and immune system.

114

To investigate this hypothesis that ERAP1 leads to altered T cell selection, impairing

immune recognition of inherent gut microbiota, and leading to an overall shift in microbial

community composition; we examined if the loss of ERAP, in a mouse model, caused a

shift in the intestinal flora. Consistent with this hypothesis, 16S community profiling

revealed that the gut microbiome in ERAP-/- did shift, however it was surprisingly towards

a more inflammatory profile, with increases of bacteria such as Rikenellaceae,

Lachnospiraceae and Ruminococcus which are not only known to elicit a strong immune

response, but are also key bacteria in the TI of AS patients (102-104). The shift in the

ERAP-/- mouse microbiome was further explored to determine if it was the overall

structure or simply the presence or absence of bacteria that was driving the signature.

Unweighted UniFrac analysis revealed that the loss of ERAP triggered a fundamental shift

of bacteria at the phylum level, causing an overall shift in community structure.

Interestingly, this also included a shift in the Bacteroides-Firmicutes ratio leading to an

increase in Firmicutes in the ERAP-/- mice, which is consistent with what we observe in

the AS gut microbiome as described in Chapter 3 of this thesis. Indicator species analysis

revealed that alterations in specific species, and not just entire Phyla, were contributing to

the microbial signature. The bacterial families that varied in ERAP+/+ and ERAP-/- mice

were Rikenellaceae, Lachnospiraceae, Ruminococcaceae, Bacteroidaceae and

Clostridiaceae which are also bacteria associated with the microbial signature in AS.

Whilst the association of particular bacterial families seen in both AS intestinal biopsies

and ERAP-/- gut microbiome is interesting, it is however incidental.

Given the loss of ERAP caused such dramatic change in gut microbiome structure an

experiment for cage and litter effects was conducted. In order to minimise such potential

confounders, the intestinal microbial composition of a single litter of ERAP mice

comprising of ERAP+/+, ERAP+/- and ERAP-/- were explored. These mice were the

offspring of sibling ERAP+/- founder mice, from the same litter, and were cohoused from

birth so that variation due to background strain and cage/litter effects were minimized or

excluded as much as possible. It is clear from the unweighted PCoA that the ERAP+/-

and ERAP-/- mice were more similar to each other and clustered away from the WT mice.

This suggests that a change in genotype, from WT to HET, is sufficient to influence the

overall structure and composition of the gut microbiome. The clustering of the ERAP+/-

and ERAP-/- together, but separately from the ERAP+/+, suggests that change of one

allele is sufficient to alter the microbiome structure, but as only one ERAP-/- mouse was

present in the litter, whether homozygous loss of ERAP has additional effects compared

115

with heterozygotes remains unclear. While only one litter was studied, and larger numbers

are needed to better characterise the impact of genotype on the intestinal microbiome,

these results are consistent with the hypothesis that AS-associated genes contribute to

disease aetiopathogenesis through effects of the underlying host genetics on intestinal

microbial composition.

The intestine is the front line between the immune system and the gut flora, and a shift in

gut flora is usually met with a strong and swift inflammatory immune response dominated

by IL-17 in an effort to return to balance and homeostasis. ERAP-/- mice demonstrated a

shift towards microbiome composition akin to what was observed in AS cases. Therefore,

we examined the affect the loss of ERAP had on the local and systemic immune system,

and how the immune system in turn sculpted the gut microbiome. There was no significant

change or alteration in the T cell or non T cell compartments. This is in keeping with

previous work that has shown that there are no differences in the profile of CD4+ T cells or

CD8+ T cells of ERAP-deficient mice when compared to ERAP competent mice (208, 209,

214, 215). There was also no significant change in the cytokine potential in ERAP-/- mice

compared to control ERAP+/+, although there was a trend towards an increase in IL-22

secretion from CD3+ T cells as well as an increase in IL-10 secretion from macrophages in

the MLN. Coupled together, this data suggests that the effect of the loss of ERAP on the

immune system is subtle and site specific. To dissect the effect of the loss of ERAP, not

only on microbiome composition but the immune system, the local intestinal tissue

environment needs to be explored. Examining the lamina propria (LP) cells from the

intestinal walls of ERAP-/- and ERAP+/+ mice would shed light on the front line

interactions between the immune system and the gut microbiota. However, this would

require significantly more mice to be studied, as LP cells are isolated in small numbers

requiring mice to be pooled to obtain sufficient cell numbers for analysis.

The impact the establishment and maintenance of beneficial interactions between the host

and their microbiota is rapidly becoming a key focus as a requirement for overall health.

The bacteria that inhabit the gut and the human immune system have developed a number

of strategies to maintain gut homeostasis in dynamic and every changing environment.

The gut also is the primary site for interaction and education between gut microbes and

the host immune system. The dysregulation of gut homeostasis is increasingly recognised

as playing a pivotal role in autoimmunity and controlling activation of the immune system in

response to these changes in gut microbiota has to be tightly regulated as to avoid chronic

116

inflammation. The role that underlying host genetics plays in gut microbiome composition

and the flow on effect to the immune system requires more interrogation. In this chapter

we described how the loss of a single AS-associated gene, ERAP1, led to dramatic shift in

the overall composition and structure of gut microbiome. Of particular interest, was the

finding that the clustering in the murine experiments was driven by bacteria that are AS

also seen in the microbiome of patients who have AS. Further investigation into the impact

of genotype on microbiome composition revealed that the change in even one allele led to

detectable shift in relative abundance of microbes. The role the immune system has in

helping shape these microbiome changes is unclear. By better understanding the

interactions locally in the intestinal tissue at the interface between gut microbes and the

immune system, we may better understand ERAP mechanism in conferring protection

from AS and possibly harnessing and modulating an anti-inflammatory immune response.

117

5 Chapter 5 Effect of underlying host genetics on intestinal microbiome composition

5.1 Introduction

The more we learn about the microbes that live in and on our body, the more we

understand about how they interact with our genes and the environment, to affect overall

health and wellbeing (211). There is an estimated 10-fold increase in the number of

bacterial cells in our body relative to human cells, essentially making us more microbe

than human (125). The majority of these microbes reside in the gut, the primary site for

interaction between microbes and the host immune system (38, 123, 126). The exact

composition of a ‘healthy’ gut microbiome differs markedly between individuals, however

microbiome composition is increasingly viewed as a risk factor in a number of chronic

diseases such as rheumatoid arthritis (224), obesity (114), IBD (102, 225), and diabetes

(226). The gut microbiota is increasingly being looked at as a target for novel therapies,

including microbiome modulation with pre- and probiotics and even faecal transplants.

These novel therapies are being increasingly used to treat patients with CD (227) and

Clostridium difficile infections (228), where conventional treatments and antibiotics have

failed. The successful use of the gut microbiome as a therapeutic target requires a deeper

understanding of factors that shape the community composition including age, diet, gender

as well as underlying host genetics.

The role of the host genetics in shaping intestinal microbial community composition in

humans is unclear. In animal models, host gene deletions have been shown to result in

shifts in microbiota composition (135). In addition, a recent quantitative trait locus mapping

study linked specific genetic polymorphisms with microbial abundances (136). These

studies highlight the importance of underlying host genetics in community composition in

animals. This is of great interest in diseases such as IBD, where many of the genes

associated with disease are associated with handling, processing and response to

microbes such as IL23R, NOD2, NOS2 and CARD9. Genetic studies in CD, one of the two

common forms of IBD, shows a marked over-representation of genes, especially NOD2

and TNFRSF18, which are associated with an increased risk of developing mycobacterial

diseases, including leprosy and tuberculosis (23). For example, seven of the eight known

genetic variants associated with increased risk of leprosy, including SLC11A1, VDR and

118

LGALS9, also increase the risk of Crohn’s disease, compared with only two being

associated with risk of any other immune-mediate disease. Coupled with the dramatic

shifts in the gut microbiota that have been associated with IBD (23, 135), this strongly

indicates a potential link between host genetics and the microbiome in IBD patients. The

genetic influence over microbiome composition is further highlighted in a study of 416

monozygotic and dizygotic twin pairs that found the most heritable taxon was the recently

described family Christensenellaceae (137).This family of bacteria was also shown to be

central to a network of co-occurring heritable microbes associated with lean body mass

index. In mice it was shown that the introduction of any of the members of the

Christensenellaceae family of protected the mice from weight gain (137). Until recently

variation and changes in the gut microbiome have been explained by diet and

environment. While the exact genes associated with the Christensenellaceae family are

unknown, this study demonstrates that microbiome composition may in part be due to host

genetics.

AS is a common inflammatory arthritis that affects over 22,000 Australians (1). Up to 70%

of AS patients have either clinical or subclinical gut disease, which suggests that intestinal

inflammation is important in the disease (7, 8). Increased gut permeability has been

demonstrated in both AS patients and their first-degree relatives compared with unrelated

healthy controls (11-13, 229). This is consistent with the hypothesis that increased

‘leakiness’ of the gut may increase the microbial dissemination and drive inflammation in

the disease.

In the SKG mouse and B27-transgenic rat models of spondyloarthritis, animals raised in

sterile conditions remain healthy and do not develop disease (230, 231). When the gut of

these mice and rats are recolonised with bacteria, arthritis often develops (15, 232).

Whilst this data supports a role for gut pathogens or the gut microbiome in AS, current

studies using antibody tests have not produced convincing evidence to demonstrate that

increased infection with any specific triggering agent influences the development of AS.

Several specific bacteria have been suggested to have a role in AS pathogenesis

particularly Klebsiella pneumoniae (138, 139), Bacteroides vulgatus (140). These two

bacteria have been of considerable focus for several reasons. When HLA‐

B27/β2 microglobulin transgenic rats, that have been raised germ free, are colonised with

an intestinal bacterial cocktail with increased amount of Bacteroides vulgatus, the rats

developed more several colitis (141). Significant research has been directed towards the

119

involvement of K. pneumoniae. Early work reported the presence of K. pneumoniae in the

faeces of AS patients during active disease (142). Studies have reported the presence of

K. pneumoniae in faeces during active inflammatory disease (142) elevated levels of

serum antibodies against K. pneumoniae in AS patient sera have been detected (143), as

well as the antigenic similarity noted between Klebsiella and HLA‐B27 (144). Recent

studies however have been unable to confirm a role of Klebsiella AS.

Understanding how the gut microbiota communities are assembled and maintained is

becoming increasingly relevant not only to general healthy, but also potentially to the

treatment of complex chronic diseases. In recent years only a few studies have examined

the composition, diversity and function of the human gut microbiome. Studies comparing

the gut microbiota of lean and obese twins have shed light on the importance of intestinal

microbes and how a change in microbiome composition can affect food metabolism in the

gut (114, 117, 118). Turnbaugh et al showed that even with a similar genetic make-up,

obese twins had substantial differences in the composition and diversity of their gut flora

with a dominance of Gram positive bacteria from the phylum Firmicutes, compared with

discordant or lean twins (114). This shift in gut flora composition altered how food was

broken down and metabolised in the gut, leading to increased body mass index, adiposity

and obesity (114, 117, 119). This demonstrates that shifts in the intestinal microbiome

have functional consequences on the health of the individual (120). To further investigate

structure and function of the gut microbiome, as well as the consequences of shifts and

alterations, faecal samples from four twin pairs discordant for obesity were transplanted

into germ free mice (121). The authors found that the Firmicute phylum dominated

microbial communities from the obese twin lead to obesity in the germ-free mice. The

Bacteroides dominated microbial communities from the lean mice kept the germ-free mice

lean. When the mice were co-housed, the invasion of Bacteroides from the lean mice lead

to a decrease in adiposity and body mass in obese mice and transformed their

microbiota’s metabolic profile to a lean-like state (121).

These studies have shown that shifts in microbiome can change metabolism; however,

they do not provide details of what constitutes a ‘normal’ microbiome or how it behaves.

Two major studies into the ‘normal’ microbiome have been undertaken by the European

MetaHit project and the National Institute of Health Human Microbiome Project in the USA

(123, 126, 233). The European MetaHIT project was developed to catalogue all the

bacteria living in the intestine. The consortium combined published data sets from across

120

the world, in addition to 22 newly sequenced faecal metagenomes from four different

European countries. They identified three clusters, or groups of bacteria, that they termed

‘enterotypes’. The authors suggested that particular enterotypes, due to their functional

capacities, might be associated with diseases such as obesity or IBD (123) Enterotypes to

date have not been confirmed in other datasets; however, the idea of groups or clusters of

bacteria being related to disease because of function is gaining popularity. The Human

Microbiome Project was developed to identify and catalogue the human microbiome by

sequencing samples from 242 individuals (129 males, 113 females) from five major body

areas. A total of 4,788 specimens were sequenced with females sampled from 18 different

habitats and males from 15 habitats (126). Metagenomic carriage of metabolic pathways

was stable among individuals, despite the variation seen in community structure. When

clinical metadata was present, ethnic background proved to be one of the strongest

predictors of both pathways and microbes. This suggests that environmental factors such

as diet, location/region and host genetics could play a role in sculpting the microbiome.

These studies further define the range of structural and functional configurations that are

present in normal microbial communities in a healthy population (123, 126).

Genes are the primary determinant of the risk of developing AS, with heritability about 97%

(4), implying that the trigger for the disease is environmental. CD is closely related to AS

with a similar prevalence and high heritability. The two commonly co-occur with an

estimated ~5% of AS patients developing CD, and 4-10% of CD patients developing AS

(234, 235). The propensity to occur in families implies shared aetiopathogenic factors, and

extensive sharing of genetic risk factors. There is strong suggestive evidence to support a

role for the gut microbiome in the aetiopathogenesis of both diseases. Understanding the

influence of host genetics on the bacterial composition of the microbiome is vital because if

the underlying genetics dictates the response of the immune system to the microbiome, or

how the body tolerates or processes microbes, then no amount of microbiota modulation

or probiotics will permanently shift the intestinal microbiome; not without doing something

to influence the gene actions. However, if host genetics favours microbiota modulation,

then the potential for the microbiome as a therapeutic target via faecal microbiota

transplants (FMT) or probiotics presents a new avenue for therapeutic developments for

AS.

We hypothesise that the major genetic predisposing factors to AS influence the gut

microbiome in healthy and affected individuals, and that genes predisposing to AS do so

121

by influencing the gut microbiome composition. We explored the association between AS

associated single nucleotide polymorphisms (SNPs) and intestinal bacteria. The intestinal

communities of 982 faecal samples were characterised in 287 DZ pairs and 204 MZ pairs,

from the TwinsUK registry as described by Goodrich et al (137).

5.2 Methods

5.2.1 TwinsUK dataset

As detailed in Goodrich et al, 982 faecal samples from as many otherwise healthy

individuals, none of which were known to be diagnosed with AS, were sequenced. Most

subjects were female aged from 23 to 86 years (average age: 60.6 ± 0.3 years).

78,938,079 quality-filtered sequences were generated and mapped to the Bacteria and

Archaea in the Greengenes database (average sequences per sample: 73,023 ± 889).

5.2.2 Sample collection All work involving human subjects was approved by the Cornell University IRB (Protocol ID

1108002388). Faecal samples were collected at home by participants in the United

Kingdom Adult Twin Registry (TwinsUK); (Moayyeri et al., 2013) in 15 ml conical tubes and

refrigerated for 1-2 days prior to the participants’ annual clinical visits at King’s College

London (KCL). Upon arrival at KCL, the samples were stored at -80ºC and shipped by

courier on dry ice to Cornell University, where they were stored at -80ºC until processing.

5.2.3 DNA extraction and 16S rRNA amplification Genomic DNA was isolated from an aliquot of ~100 mg from each sample using the

PowerSoil® - htp DNA isolation kit (MoBio Laboratories Ltd, Carlsbad, CA). 16S rRNA

genes were amplified by PCR from each of the 982 samples (287 DZ twin pairs, 204 MZ

twin pairs) using the 515F and 806R primers for the V4 hypervariable region of the 16S

rRNA as previously described (127).

5.2.4 16S rRNA filtering and analysis

As described in Goodrich et al matching paired-end sequences (mate-pairs) were merged

using the fastq-join command in the ea-utils software package and merged sequences

over 275 bp in length were filtered out of the dataset (137). The remaining merged

sequences were analysed using the open-source software package QIIME 1.7.0

(Quantitative Insights Into Microbial Ecology; (169)).

122

Quality filters were used to remove sequences containing uncorrectable barcodes,

ambiguous bases, or low quality reads (Phred quality scores ≤ 25). Closed-reference

Operational Taxonomy Unit (OTU) picking at 97% identity against the May 2013

Greengenes database (97% OTU reference sequences), which excluded 15.6% of total

sequences. The taxonomic assignment of the reference sequence was used as the

taxonomy for each OTU. α-diversity, Chao 1, Observed Species as well as β-diversity

(Unweighted and Weighted UniFrac) (189) metrics, and Bray-Curtis Dissimilarity using the

Greengenes phylogenetic tree where necessary, were calculated. β-diversity was

calculated with a rarefied OTU table containing 10,000 sequences per sample and PCoA

on the distance matrices. α-diversity metrics were calculated from 100 iterations using a

rarefaction of 10,000 sequences per sample. Summaries of the taxonomic distributions of

OTUs were generated at six levels from genus to phylum from the non-rarefied OTU table.

5.2.5 SNP selection, genotyping and imputation

Individuals were genotyped on the HumanHap 300k Duo array or the HumanHap610-

Quad array. Of the 491 twin pairs, 4 were genotyped on both chips, 252 were genotyped

only using the Illumina HumanHap 300 duo array, with the remaining 235 genotyped using

the Illumina HumanHap 610k duo array. Previous studies have identified 43 SNPs to be

associated with AS, some of which are in the MHC (24). Genotype data was extracted

1MB around the 43 confirmed susceptibility loci; SNPs with a MAF <0.05 were excluded.

For each locus we phased data with MACH (236) using the default parameters. For each

locus we imputed missing genotypes from the European panel of the 1,000 Genomes

Project (2010-11 data freeze, 2012-02-14 haplotypes as downloaded from the MACH

website*) using minimac (237). Imputed SNPs with low imputation quality score (Rsq <

0.5) were removed from all analyses. The reported index SNP for each of the susceptibility

loci was then used for subsequent analyses, unless otherwise stated. Genotyped data was

used where available otherwise the dosages from the imputed data were used (i.e.

expected number of copies of a given allele, ranging from 0 to 2).

Due to the large effect of HLA-B27, two allelic risk scores were constructed using AS

associated imputed SNPs (Table 5.1). The first score for each of the twins was derived as

a sum of AS-risk alleles across the imputed dosages for all 43 SNPs. The second score

weighted the risk allele dosage by the effect size reported in the relevant published papers

and then summed the weighted dosages. Allelic risk scores were calculated using custom

scripts in R 3.0.2 (238).

123

* http://www.sph.umich.edu/csg/abecasis/MACH/download/1000G.2012-02-14.html

5.2.6 Statistical analysis We assessed whether each of the 768 16S rRNA OTUs followed a Gaussian distribution

through plotting histograms and calculating the mean, skew and kurtosis. The mean

allowed the assessment of central tendency for each OTU. Skewness and kurtosis

describe the shape of the distribution for each OTU; skewness quantifies the asymmetry of

the distribution and kurtosis quantifies the peakedness of the distribution. Based on these

summary statistics and the histograms, it was determined that some of the OTUs did not

follow a normal distribution. The statistical tests used in the remainder of this chapter

assume the data follows a normal distribution; we therefore applied an inverse normal

transformation to each of the OTUs to ensure normality before analysis.

Linear mixed effects models were used to investigate the association between each of the

OTUs and the 43 individual AS SNPs or allelic scores. These models account for the

correlation between twins within a family. Analysis of the single SNPs was conducted in

Merlin (239) using the following command line:

merlin -d zygosity dat file,SNP dat file –p Twins zygosity pedigree file,SNP pedigree

file -m SNP map file --assoc --tabulate Results-chr22-assoc-Jan14

This uses a likelihood ratio test to compare a model including the individual SNP to a

model without the SNP. Descriptive statistics and the analysis of the allelic score was

conducted in R (version 3.0.2) using the nlme library (238).

5.3 Results

5.3.1 Distribution of OTUs is not linked to infection or Phylum

The distribution of OTUs is often ignored because currently used methods of normalisation

and quality control are deemed sufficient. However, the statistical tests used to investigate

the relationship between host genetics and the microbiome assume the OTU is normally

distributed. Figure 5.1 shows the histograms for a selection of the OTUs, and Table 5.2

shows the mean, skew and kurtosis. Histograms demonstrated a mixture of modalities

across the OTUs, with some OTUs following a normal distribution but the majority showing

124

bimodal distributions. Table 5.2 indicates that there is a large range of skewness and

kurtosis in the OTUs (Skew: -0.503 to 0.952, Kurtosis: -1.616 to 0.744), further

emphasising the departure from normality for many OTUs. This demonstrates the need for

normalisation prior to applying a statistical model.

Figure 5.1 Representative OTUs prior to normalisation, demonstrating the range of distributions present within the OTU table and the same order of bacteria, showing the range of mean, skew and kurtosis of each OTU. Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_). To establish if the variation of distributions of the OTUs was due to pathogenic bacteria,

such as E. coli, or if particular modality patterns were specific to Gram negative or Gram

positive bacteria, OTUs within a family or genus of bacteria were compared. Comparisons

showed that OTUs within the same family or genus have very different distributions (Figure

5.2). Distributions of known pathogenic bacteria, such as pathogenic E. coli, were also

compared and no trend or pattern was observed (Figure 5.3). This demonstrates that the

variability seen across the OTU table was not driven by bacterial pathogenicity or by

taxonomy. This also indicated that OTU distribution cannot be predicted through

taxonomy, and variation seen within the OTUs may be a due to temporal variance or

normal variation with the bacteria and the microbiome.

125

Figure 5.2 Distributions of the family Lachnospiraceae demonstrating variation of distribution patterns within the same family of bacteria. Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_).

Figure 5.3 Distributions of selected members of the family Enterobacteriaceae, which contains a number of pathogenic bacteria including Shigella, Klebsiella and E.coli, showing no clear pattern of distribution. Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_).

5.3.2 AS susceptibility variants are associated with overall microbiome structure

AS is a complex highly heritable disease with over 40 loci known to be associated with AS.

Due to the large effect of HLA-B27, two allelic risk scores were constructed to investigate

the combined effect of AS susceptibility variants on microbiome composition (Figure 5.4).

No OTUs were significantly associated with either the weighted or unweighted allelic risk

score, using a Bonferroni corrected experiment-wide significance level. However, 13 OTUs

126

reached nominal significance (P<0.05) on the weighted risk score (Table 5.2) and 102

OTUs reached nominal significance with the unweighted risk score (including the 13 OTUs

from the weighted score) (Table 5.3). Within the 13 OTUs that were nominally associated

with the weight risk allelic score, six OTUs belonged to the family of Ruminococcaceae,

three belong to the family Lachnospiraceae and the remaining four OTUs belonged to

Bifidobacteriaceae and the order Clostridiales. The families of Ruminococcaceae,

Lachnospiraceae and the order Clostridiales are known to be associated with AS and are

members of the core microbiome as described in Chapter 3 of this thesis. Of the 102

OTUs reaching nominal significance in the unweighted risk score, the major families of

bacteria represented were again Ruminococcaceae, Lachnospiraceae, Bifidobacteriaceae,

the order Clostridiales and Veillonellaceae. This demonstrates that regardless of whether

the risk score is weighted for effect size or not, the same AS core families of

Ruminococcaceae, Lachnospiraceae, and the order Clostridiales are associated with AS

SNPs, with only the membership and number of OTUs expanding. This is similar to what

we observed in intestinal dysbiosis in AS cases in Chapter 3. As the threshold for core

membership was reduced to 90, 75 and 50% of the time, the families of bacteria present

remained the same however the number of genus and species within these families

increased, especially in the order Clostridiales, Lachnospiraceae and Ruminococcaceae

families. This suggests the structure of the core microbiome is not only robust, but may be

influenced by underlying host genetics.

Figure 5.4 Histogram of weighted allelic risk score demonstrating the effect of HLA-B27 genotype (GG, AG, AA).

127

Table 5.1 Reported AS associated loci including correlation of genes associated with inflammatory bowel disease.

Reported gene SNP Locus AS Risk/Non-risk Alleles

Position

RUNX3 rs6600247 1p36 C/T 25,177,701 IL23R rs11209026 1p31 G/A 67,478,546 rs12141575 1p31 A/G 67,520,024 GPR25-KIF21B rs41299637 1q32 T/G 199,144,473 PTGER4 rs12186979 5p13 C/T 40,560,617 ERAP1-ERAP2-LNPEP rs30187 5q15 T/C 96,150,086 rs2910686 5q15 G/A 96,278,345 ERAP1-ERAP2 rs10045403 5q15 A/G 96, 173,489 IL12B rs6871626 5q33 T/G 158,759,370 rs6556416 5q33 G/T 158,751,323 CARD9 rs1128905 9q34 C/T 138,373,660 LTBR-TNFRSF1A rs1860545 12p13 C/T 6,317,038 rs7954567 12p13 A/G 6,361,386 NPEPPS-TBKBP1-TBX21

rs9901869 17q21 T/C 42,930,205

21q22 rs2836883 21q22 G/A 39,388,614 IL6R rs4129267 1q21 G/A 152,692,888 FCGR2A rs1801274 1q23 T/C 159,746,369 rs2039415 1q23 C/T 159,121,069 UBE2E3 rs12615545 2q31 C/T 181,756,697 GPR35 rs4676410 2q37 A/G 241,212,412 NKX2-3 rs11190133 10q24 C/T 101,268,715 ZMIZ1 rs1250550 10q22 C/A 80,730,323 SH2B3 rs11065898 12q24 A/G 110,346,958 GPR65 rs11624293 14q31 C/T 87,558,574 IL27-SULT1A1 rs7282490 16p11 A/G 28,525,386 rs35448675 16p11 A/G 28,236,248 TYK2 rs35164067 19p13 G/A 10,386,181 rs6511701 19p13 A/C 10,486,067 ICOSLG rs7282490 21q22 G/A 44,440,169 EOMES rs13093489 3p24 A/C 27,769,911 IL7R rs11742270 5p13 G/A 35,917,200 UBE2L3 rs2283790 22q11 C/T 20,286,653 BACH2 rs17765610 6q15 G/A 90,722,494 rs639575 6q15 A/T 91,047,852 NOS2 rs2531875 17q11 G/T 23,172,294 rs2297518 17q11 A/G 23,120,724 HHAT rs12758027 1q32 C/T 208,861,431

128

IL1R2-IL1R1 rs4851529 2q11 G/A 102,013,732 rs2192752 2q11 C/A 102,135,805 2p15 rs6759298 2p15 C/G 62,421,949 GPR37 rs2402752 7q31 T/C 124,226,835 HLA-B27 rs4349859 6p21 A/G 31,365,787 ANTRX2 rs4333130 4q21 A/G 81,168,853

130

Table 5.2 OTUs with taxonomy associated at P<0.05 from the weighted allelic risk score, including beta and standard error. Six of the OTUs associated belong to the family of Ruminococcaceae, three belong to the family Lachnospiraceae and the remainder belonging to Bifidobacteriaceae and the order Clostridiales. Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_). OTU Beta Standard Error P-value Taxonomy

178760 -0.0256 0.0107 0.0172 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__

179572 -0.0261 0.0112 0.0202 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__

681370 -0.0487 0.0211 0.0216 Actinobacteria; c__Actinobacteria; o__Bifidobacteriales; f__Bifidobacteriaceae; g__Bifidobacterium; s__pseudolongum

324894 -0.0392 0.0176 0.0267 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__


4347159 -0.1122 0.0505 0.0269 Actinobacteria; c__Actinobacteria; o__Bifidobacteriales; f__Bifidobacteriaceae; g__Bifidobacterium; s__adolescentis

178642 -0.0235 0.0106 0.0275 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__

2688035 -0.0313 0.0143 0.0292 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__

16054 -0.0590 0.0272 0.0310 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Ruminococcus; s__callidus




193946 -0.0339 0.0172 0.0498 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__

131

Table 5.3 OTUs reaching nominal significance (P<0.05) in the unweighted risk score. The major families of bacteria represented are Ruminococcaceae, Lachnospiraceae, Bifidobacteriaceae, the order Clostridiales and Veillonellaceae, which are the same families of bacteria observed in the weighted analysis and in the core microbiome in Chapter 3 of this thesis. Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_).

OTU Beta SE P-value Taxonomy 16054 -0.0292 0.0085 0.0006 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Ruminococcus; s__callidus 324894 -0.0190 0.0055 0.0007 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 330469 -0.0141 0.0043 0.0012 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 2617854 -0.0525 0.0164 0.0015 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Rikenellaceae; g__; s__ 189937 -0.0116 0.0039 0.0031 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Faecalibacterium;

s__prausnitzii 177207 -0.0224 0.0078 0.0045 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ 185558 -0.0123 0.0043 0.0047 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ 193477 -0.0123 0.0044 0.0051 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ 198947 -0.0141 0.0051 0.0058 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 185814 -0.0244 0.0088 0.0060 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 197698 -0.0203 0.0074 0.0061 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ 348642 -0.0289 0.0105 0.0061 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__[Ruminococcus]; s__gnavus 196513 -0.0191 0.0070 0.0066 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 319275 -0.0187 0.0069 0.0077 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Faecalibacterium;

s__prausnitzii 182036 -0.0233 0.0087 0.0080 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 193969 -0.0242 0.0091 0.0084 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Coprococcus; s__ 193968 -0.0284 0.0109 0.0094 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 197343 -0.0134 0.0051 0.0094 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 184114 -0.0325 0.0126 0.0103 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 181826 -0.0125 0.0048 0.0105 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__

132

3530697 -0.0174 0.0067 0.0105 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 190169 -0.0231 0.0091 0.0116 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Faecalibacterium;

s__prausnitzii 189092 -0.0320 0.0128 0.0127 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Faecalibacterium;

s__prausnitzii 191734 -0.0072 0.0029 0.0132 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__ 535955 -0.0210 0.0085 0.0144 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__SMB53; s__ 295258 -0.0243 0.0099 0.0145 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 180721 -0.0159 0.0065 0.0152 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 175729 -0.0233 0.0096 0.0160 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ 183970 -0.0162 0.0067 0.0165 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ 181139 -0.0222 0.0092 0.0167 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Faecalibacterium;

s__prausnitzii 185583 -0.0271 0.0113 0.0171 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 310979 -0.0167 0.0070 0.0178 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ 2943548 -0.0299 0.0125 0.0178 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Ruminococcus; s__ 44151 -0.0103 0.0043 0.0180 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 187924 -0.0203 0.0086 0.0185 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 4425669 -0.0354 0.0150 0.0186 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Coprococcus; s__ 175836 -0.0077 0.0033 0.0194 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 192127 -0.0143 0.0061 0.0197 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ 186772 -0.0181 0.0077 0.0199 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 185575 -0.0488 0.0209 0.0200 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 4466275 -0.0108 0.0047 0.0216 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 199279 -0.0089 0.0039 0.0220 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__ 192370 -0.0098 0.0043 0.0227 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 193769 -0.0169 0.0074 0.0233 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__[Ruminococcus]; s__gnavus

133

178760 -0.0076 0.0034 0.0235 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ 182577 -0.0189 0.0084 0.0246 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 189924 -0.0114 0.0050 0.0249 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 179572 -0.0079 0.0035 0.0252 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__ 179200 -0.0196 0.0087 0.0252 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 317814 -0.0206 0.0092 0.0259 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ 305318 -0.0157 0.0071 0.0268 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ 193709 -0.0226 0.0101 0.0269 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 187504 -0.0258 0.0116 0.0269 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 4347159 -0.0350 0.0158 0.0273 Actinobacteria; c__Actinobacteria; o__Bifidobacteriales; f__Bifidobacteriaceae; g__Bifidobacterium;

s__adolescentis 4425663 -0.0216 0.0098 0.0278 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ 295485 -0.0116 0.0053 0.0280 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Faecalibacterium;

s__prausnitzii 110192 -0.0117 0.0053 0.0282 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Oscillospira; s__ 183439 -0.0297 0.0135 0.0288 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 1820513 -0.0218 0.0099 0.0291 Proteobacteria; c__Betaproteobacteria; o__Burkholderiales; f__Alcaligenaceae; g__Sutterella; s__ 189877 -0.0102 0.0047 0.0310 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__ 2688035 -0.0097 0.0045 0.0312 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ 178642 -0.0071 0.0033 0.0320 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ 179486 -0.0081 0.0037 0.0320 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 3709990 0.0132 0.0061 0.0327 Firmicutes; c__Erysipelotrichi; o__Erysipelotrichales; f__Erysipelotrichaceae; g__[Eubacterium];

s__dolichum 174695 -0.0193 0.0090 0.0335 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ 174818 -0.0215 0.0100 0.0335 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 48084 -0.0311 0.0146 0.0341 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 180473 -0.0172 0.0081 0.0341 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__

134

180352 -0.0128 0.0060 0.0343 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ 357261 -0.0140 0.0066 0.0348 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Ruminococcus; s__ 191214 -0.0131 0.0062 0.0348 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ 177581 -0.0169 0.0080 0.0349 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ 179201 -0.0103 0.0049 0.0352 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 177758 -0.0136 0.0065 0.0357 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Dorea; s__ 193946 -0.0114 0.0054 0.0364 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__ 181754 -0.0146 0.0070 0.0368 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ 192461 -0.0079 0.0038 0.0371 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ 4414476 -0.0269 0.0129 0.0375 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 216710 -0.0216 0.0104 0.0377 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 181008 -0.0145 0.0069 0.0378 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ 189828 -0.0227 0.0109 0.0383 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 185584 -0.0173 0.0083 0.0384 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 681370 -0.0136 0.0066 0.0399 Actinobacteria; c__Actinobacteria; o__Bifidobacteriales; f__Bifidobacteriaceae; g__Bifidobacterium;

s__pseudolongum 178965 -0.0137 0.0067 0.0411 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ 4465124 -0.0310 0.0151 0.0412 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__Clostridium; s__ 300620 -0.0235 0.0115 0.0422 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 188676 -0.0254 0.0125 0.0424 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 198909 -0.0099 0.0049 0.0425 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Lachnospira; s__ 182116 -0.0138 0.0068 0.0431 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ 4357811 -0.0074 0.0037 0.0434 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides; s__ 288651 -0.0178 0.0088 0.0443 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 196957 -0.0229 0.0114 0.0451 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 193148 -0.0072 0.0036 0.0455 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__

135

4458959 -0.0102 0.0051 0.0462 Firmicutes; c__Clostridia; o__Clostridiales; f__Veillonellaceae; g__Veillonella; s__ 1952 -0.0356 0.0179 0.0470 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Porphyromonadaceae; g__Parabacteroides;

s__ 190772 -0.0168 0.0084 0.0474 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ 198127 -0.0112 0.0057 0.0482 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ 232828 -0.0101 0.0051 0.0489 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 187404 -0.0182 0.0092 0.0489 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ 214036 -0.0245 0.0124 0.0490 Firmicutes; c__Clostridia; o__Clostridiales; f__[Mogibacteriaceae]; g__; s__ 189176 -0.0130 0.0066 0.0499 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__[Ruminococcus]; s__gnavus

137

5.3.3 Genes involved in peptide presentation and microbial sensing are

associated with AS core families of bacteria

We were interested in examining the association between individual AS

associated SNPs and the microbiome. We began by investigating the

strongest associations to AS susceptibility HLA-B27, ERAP and IL-23R as

well as NKX2-3, ICOSLG, NOS2, CARD9, BACH2 and SH2B3. These genes

are of interest because of their strong relative effect in AS and for their roles in

processing, trimming, presenting antigens in the class I antigen presentation,

microbial sensing and immunity. A total of 13 SNPs across the 10 genes

(Table 5.4) were each tested against the each individual OTU in the

microbiome.

Table 5.4. AS associated loci investigated for association with each OTU in the microbiome.

Locus SNP Chr Risk /Non-

risk Allele

IL23R rs11209026 1p31 G/A

ERAP1-ERAP2-LNPEP

rs30187 5q15 T/C

rs2910686 5q15 G/A

ERAP1-ERAP2 rs10045403 5q15 A/G

HLA-B27 rs4349859 6

NKX2-3 rs11190133 10q24 C/T

ICOSLG rs7282490 21q22 G/A

NOS2 rs2531875 17q11 G/T

rs2297518 17q11 A/G

CARD9 rs1128905 9q34 C/T

BACH2 rs17765610 6q15 G/A

rs639575 6q15 A/T

SH2B3 rs11065898 12q24 T/C

138

Whilst none of the OTUs were associated with the SNPs involved in

processing, trimming, presenting antigens at a Bonferroni corrected

significance level, a total of 212 OTUs reached nominal significance P<0.05.

SNPs in the ERAP1 locus were more significant with OTUs than IL23R SNPs

or the HLA-B27 allele. There were 53 OTUs, predominantly

Ruminococcaceae and Lachnospiraceae families (Table 5.5), associated with

the ERAP1-ERAP2 SNP, rs10045403, and 44 OTUs belonging to the same

families of bacteria associated with the ERAP1-ERAP2-LNPEP SNP,

rs2910686. The aminopeptidase ERAP1 SNP, rs30187, which is known to

interact with HLA-B27 in AS, showed association with 50 OTUs all belonging

to Ruminococcaceae and Lachnospiraceae bacterial families. There were 25

OTUs associated with the IL-23R SNP, rs11209026, predominantly

Ruminococcus, Lachnospiraceae and Bacteroidaceae families (Table 5.6). An

additional 40 OTUs belonging to the order Clostridiales, Ruminococcaceae

and Lachnospiraceae families were associated with the HLA-B27 allele (Table

5.7).

Microbial sensing genes NKX2-3, ICOSLG, NOS2, CARD9, BACH2 and

SH2B3 were also examined to assess for association with individual microbes

within the intestinal microbiome (Table 5.4). These six loci are of interest

because they are involved in sensing and responding to microbes and

microbial antigen with four of the six loci also associated with dysregulated

microbial sensing and response in Crohn’s disease (23). Whilst none of the

OTUs reached Bonferroni corrected significance, a total of 404 OTUs reached

nominal significance P<0.05 across all six microbial sensing SNPs. BACH2

was the most associated of all the microbial sensing genes with 50 OTUs

associated with rs17765610 and 81 with rs639575; the majority of which

belong to Ruminococcaceae, Lachnospiraceae and Bacteroides families

(Supp table 5.8). NOS2 rs2297518 showed 65 OTU associations and

rs2531875 with 18 OTU associations, all dominated by Ruminococcaceae,

Lachnospiraceae and Bacteroides families along with Rikenellaceae and

Clostridiaceae (Supplementary table 5.9). There were 105 OTUs associated

with the CARD9 SNP, rs1128905, predominately Lachnospiraceae,

Veillonellaceae and Ruminococcaceae, including Faecalibacterium prausnitzii

139

(Supplementary table 5.10). The remaining genes NKX2-3, ICOSLG and

SH2B3 were associated with 36, 30 and 18 OTUs respectively, with the

families of Ruminococcaceae, Lachnospiraceae dominating the associations

(Supplementary table 5.11). This demonstrated that no one SNP, or gene, is

associated with an individual family or genes of bacteria; rather, these genes

are linked to the three major families of Ruminococcaceae, Lachnospiraceae

and Bacteroidaceae as well as members of the Rikenellaceae and

Veillonellaceae families that have been shown to dominate the core

microbiome in AS, as detailed in Chapter 3 of this thesis.

141

Table 5.5 OTUs reaching nominal significance (P<0.05) in the ERAP locus. 53 OTUs, predominantly Ruminococcaceae and Lachnospiraceae families associated with the ERAP1-ERAP2 SNP, rs10045403; 44 OTUs belonging to the same families of bacteria associated with the ERAP1-ERAP2-LNPEP SNP, rs2910686 and 50 OTUs all belonging to Ruminococcaceae and Lachnospiraceae associated with ERAP1 SNP, rs30187. Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_). SNP Gene OTU Beta Pvalue Taxonomy

rs30187 ERAP 4420408 0.108 1.87E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides;

s__

rs30187 ERAP 4256470 0.167 1.87E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides;

s__ovatus

rs30187 ERAP 182073 0.102 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__



rs30187 ERAP 2724175 0.122 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__

rs30187 ERAP 4457427 0.118 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Lachnospira; s__

rs30187 ERAP 208739 0.088 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae;

g__Faecalibacterium; s__prausnitzii

rs30187 ERAP 185420 -0.197 7.89E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides;

s__

rs30187 ERAP 352747 -0.111 7.89E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides;

s__ovatus

rs30187 ERAP 178760 -0.061 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__


142





rs30187 ERAP 178642 -0.043 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__





rs30187 ERAP 189176 -0.079 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__[Ruminococcus];

s__gnavus

rs30187 ERAP 348642 -0.119 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__[Ruminococcus];

s__gnavus

rs30187 ERAP 179557 -0.036 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__





rs30187 ERAP 181047 -0.098 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Coprococcus; s__



rs30187 ERAP 174752 -0.083 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__

143

















rs30187 ERAP 4437359 -0.212 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Oscillospira; s__

rs30187 ERAP 4428714 -0.105 7.89E-03 Tenericutes; c__Mollicutes; o__RF39; f__; g__; s__

rs10045403 ERAP1-ERAP2 12574 -0.085 6.33E-04 Actinobacteria; c__Actinobacteria; o__Actinomycetales; f__Actinomycetaceae;

g__Actinomyces; s__

rs10045403 ERAP1-ERAP2 352747 -0.091 6.33E-04 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides;

s__ovatus

144

rs10045403 ERAP1-ERAP2 4424239 -0.112 6.33E-04 Firmicutes; c__Bacilli; o__Lactobacillales; f__Streptococcaceae; g__Streptococcus; s

rs10045403 ERAP1-ERAP2 177207 -0.119 6.33E-04 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__

rs10045403 ERAP1-ERAP2 2256425 -0.1 6.33E-04 Firmicutes; c__Clostridia; o__Clostridiales; f__Christensenellaceae; g__; s__

rs10045403 ERAP1-ERAP2 194659 -0.145 6.33E-04 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__; s__

rs10045403 ERAP1-ERAP2 178018 -0.14 6.33E-04 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__


rs10045403 ERAP1-ERAP2 2210028 -0.137 6.33E-04 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__[Ruminococcus];

s__

rs10045403 ERAP1-ERAP2 148279 -0.13 6.33E-04 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__


rs10045403 ERAP1-ERAP2 181826 -0.089 6.33E-04 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__



rs10045403 ERAP1-ERAP2 110192 -0.099 6.33E-04 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Oscillospira; s__

rs10045403 ERAP1-ERAP2 186133 -0.091 6.33E-04 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Ruminococcus;

s__


s__


s__callidus

rs10045403 ERAP1-ERAP2 4374302 0.259 6.87E-04 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Dorea; s__

rs10045403 ERAP1-ERAP2 348009 0.195 6.87E-04 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Oscillospira; s__

145

rs10045403 ERAP1-ERAP2 4347159 -0.349 1.26E-03 Actinobacteria; c__Actinobacteria; o__Bifidobacteriales; f__Bifidobacteriaceae;

g__Bifidobacterium; s__adolescentis

rs10045403 ERAP1-ERAP2 535955 -0.209 1.26E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__SMB53; s__

rs10045403 ERAP1-ERAP2 681370 -0.139 1.40E-03 Actinobacteria; c__Actinobacteria; o__Bifidobacteriales; f__Bifidobacteriaceae;

g__Bifidobacterium; s__pseudolongum

rs10045403 ERAP1-ERAP2 185420 -0.309 1.40E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides;

s__

rs10045403 ERAP1-ERAP2 4476780 -0.225 1.40E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Rikenellaceae; g__; s__











rs10045403 ERAP1-ERAP2 4465124 -0.264 1.40E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__Clostridium; s__



rs10045403 ERAP1-ERAP2 348642 -0.155 1.40E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__[Ruminococcus];

146



rs10045403 ERAP1-ERAP2 187868 -0.048 1.40E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Coprococcus; s__











rs10045403 ERAP1-ERAP2 368490 -0.178 1.91E-02 Firmicutes; c__Bacilli; o__Turicibacterales; f__Turicibacteraceae; g__Turicibacter; s__

rs2910686 ERAP1-

ERAP2-LNPEP

216599 -0.22 1.40E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Rikenellaceae; g__; s__

rs2910686 ERAP1-

ERAP2-LNPEP

368490 -0.274 1.40E-03 Firmicutes; c__Bacilli; o__Turicibacterales; f__Turicibacteraceae; g__Turicibacter; s__

rs2910686 ERAP1-

ERAP2-LNPEP

174010 -0.111 1.40E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__

rs2910686 ERAP1-

ERAP2-LNPEP


147

rs2910686 ERAP1-

ERAP2-LNPEP

555547 -0.109 1.40E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Christensenellaceae; g__; s__

rs2910686 ERAP1-

ERAP2-LNPEP

177230 -0.096 1.40E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__

rs2910686 ERAP1-

ERAP2-LNPEP


rs2910686 ERAP1-

ERAP2-LNPEP


rs2910686 ERAP1-

ERAP2-LNPEP

2210028 -0.103 1.40E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__[Ruminococcus];

s__

rs2910686 ERAP1-

ERAP2-LNPEP

193666 -0.151 1.40E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__

rs2910686 ERAP1-

ERAP2-LNPEP

358798 -0.13 1.40E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__

rs2910686 ERAP1-

ERAP2-LNPEP

178238 -0.117 1.40E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Coprococcus; s__

rs2910686 ERAP1-

ERAP2-LNPEP

184770 -0.127 1.40E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Coprococcus; s__

rs2910686 ERAP1-

ERAP2-LNPEP

187924 -0.131 1.40E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__

rs2910686 ERAP1-

ERAP2-LNPEP


rs2910686 ERAP1- 688923 -0.2 1.40E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__

148

ERAP2-LNPEP

rs2910686 ERAP1-

ERAP2-LNPEP


rs2910686 ERAP1-

ERAP2-LNPEP


rs2910686 ERAP1-

ERAP2-LNPEP

146554 -0.277 1.40E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Ruminococcus;

s__

rs2910686 ERAP1-

ERAP2-LNPEP


s__

rs2910686 ERAP1-

ERAP2-LNPEP


s__

rs2910686 ERAP1-

ERAP2-LNPEP


s__bromii

rs2910686 ERAP1-

ERAP2-LNPEP

4428714 -0.108 1.40E-03 Tenericutes; c__Mollicutes; o__RF39; f__; g__; s__

rs2910686 ERAP1-

ERAP2-LNPEP

4425214 -0.168 1.91E-02 Firmicutes; c__Bacilli; o__Lactobacillales; f__Streptococcaceae; g__Streptococcus;

s__

rs2910686 ERAP1-

ERAP2-LNPEP

192684 -0.062 2.64E-02 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides;

s__

rs2910686 ERAP1-

ERAP2-LNPEP

352747 -0.087 2.64E-02 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides;

s__ovatus

rs2910686 ERAP1-

ERAP2-LNPEP

772282 -0.08 2.64E-02 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Rikenellaceae; g__; s__

149

rs2910686 ERAP1-

ERAP2-LNPEP

4424239 -0.094 2.64E-02 Firmicutes; c__Bacilli; o__Lactobacillales; f__Streptococcaceae; g__Streptococcus;

rs2910686 ERAP1-

ERAP2-LNPEP


rs2910686 ERAP1-

ERAP2-LNPEP


rs2910686 ERAP1-

ERAP2-LNPEP

178183 -0.089 2.64E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__; s__

rs2910686 ERAP1-

ERAP2-LNPEP

194659 -0.077 2.64E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__; s__

rs2910686 ERAP1-

ERAP2-LNPEP


rs2910686 ERAP1-

ERAP2-LNPEP


rs2910686 ERAP1-

ERAP2-LNPEP


rs2910686 ERAP1-

ERAP2-LNPEP


rs2910686 ERAP1-

ERAP2-LNPEP


rs2910686 ERAP1-

ERAP2-LNPEP


rs2910686 ERAP1- 179583 -0.063 2.64E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__

150

ERAP2-LNPEP

rs2910686 ERAP1-

ERAP2-LNPEP


rs2910686 ERAP1-

ERAP2-LNPEP


rs2910686 ERAP1-

ERAP2-LNPEP


rs2910686 ERAP1-

ERAP2-LNPEP


rs2910686 ERAP1-

ERAP2-LNPEP


151

Table 5.6 OTUs reaching nominal significance (P<0.05) associated with the IL-23R SNP, rs11209026, predominantly Ruminococcaceae, Lachnospiraceae and Bacteroidaceae families. . Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_). SNP Gene OTU Beta Pvalue Taxonomy rs11209026 IL23R 4453609 -0.777 7.89E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Rikenellaceae; g__; s__ rs11209026 IL23R 4424239 -0.214 7.89E-03 Firmicutes; c__Bacilli; o__Lactobacillales; f__Streptococcaceae; g__Streptococcus; s__ rs11209026 IL23R 179847 -0.117 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs11209026 IL23R 183147 -0.12 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs11209026 IL23R 186735 -0.218 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs11209026 IL23R 192735 -0.167 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs11209026 IL23R 196100 -0.113 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs11209026 IL23R 191734 -0.085 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__ rs11209026 IL23R 177663 -0.327 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs11209026 IL23R 730906 -0.133 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs11209026 IL23R 4426298 0.301 1.71E-02 Actinobacteria; c__Actinobacteria; o__Bifidobacteriales; f__Bifidobacteriaceae;

g__Bifidobacterium; s__animalis rs11209026 IL23R 336559 0.308 1.71E-02 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides;

s__ rs11209026 IL23R 3211875 0.225 1.71E-02 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides;

s__ rs11209026 IL23R 4256470 0.613 1.71E-02 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides;

s__ovatus rs11209026 IL23R 328617 0.574 1.71E-02 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides;

s__uniformis rs11209026 IL23R 157338 0.249 1.71E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ rs11209026 IL23R 173876 0.378 1.71E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ rs11209026 IL23R 173927 0.284 1.71E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ rs11209026 IL23R 181059 0.615 1.71E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ rs11209026 IL23R 192741 0.365 1.71E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ rs11209026 IL23R 192720 0.368 1.71E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs11209026 IL23R 317677 0.168 1.71E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs11209026 IL23R 571642 0.216 1.71E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__

152

rs11209026 IL23R 4408801 0.205 1.71E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Oscillospira; s__ rs11209026 IL23R 178117 0.193 1.71E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Ruminococcus;

153

Table 5.7 OTUs reaching nominal significance (P<0.05) associated with the HLA-B27 allele, predominantly Clostridiaceae, Ruminococcaceae and Lachnospiraceae families. Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species(s_). SNP Gene OTU Beta Pvalue Taxonomy

rs4349859 HLA-B27 4232045 -0.081 1.42E-02 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides; s__

rs4349859 HLA-B27 2318497 -0.135 1.42E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__; s__

rs4349859 HLA-B27 4383953 -0.157 1.42E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__; s__

rs4349859 HLA-B27 304779 -0.061 1.42E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__Clostridium;

s__perfringens

rs4349859 HLA-B27 181485 -0.049 1.42E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__[Ruminococcus]; s__

rs4349859 HLA-B27 193852 -0.054 1.42E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__

rs4349859 HLA-B27 28914 -0.054 1.42E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Roseburia; s__




rs4349859 HLA-B27 193644 -0.081 1.42E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__

rs4349859 HLA-B27 157338 0.092 1.89E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__






154

rs4349859 HLA-B27 217109 0.105 1.89E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Christensenellaceae; g__; s__



rs4349859 HLA-B27 177047 0.129 1.89E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__



rs4349859 HLA-B27 4398588 0.059 1.89E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__[Ruminococcus]; s__

rs4349859 HLA-B27 193969 0.109 1.89E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Coprococcus; s__

rs4349859 HLA-B27 175148 0.064 1.89E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__









rs4349859 HLA-B27 2307779 0.093 1.89E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Oscillospira; s__

rs4349859 HLA-B27 147969 0.149 1.89E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Ruminococcus; s__




155

rs4349859 HLA-B27 4306262 0.206 1.89E-02 Verrucomicrobia; c__Verrucomicrobiae; o__Verrucomicrobiales; f__Verrucomicrobiaceae;

157

5.4 Discussion

To better understand the possibilities of modifying the microbiome as a

therapeutic strategy we need to better understand the factors that influence

the microbiome, from diet to environment and the underlying host genetics.

Understanding the effect of host genetics in the bacterial composition of the

microbiome is important, because if the underlying genetics dictates the

response of the immune system to certain microbes, or how the body

tolerates or processes them, then no amount of microbiota modulation or

probiotics will permanently shift the intestinal microbiome. However, if host

genetics favours microbiota modulation, then the potential for the use of

microbes as a treatment through faecal transplants or probiotics, as it is

currently employed in clinical practice for antibiotic resistant Clostridium

difficile infections, presents a new avenue for therapeutic developments for

AS.

Studies comparing genotype and the microbiome, particularly in patients with

IBD, have demonstrated evidence of gene-microbiota interactions (102, 240-

242). Recently, several studies showed that the Fucosyltransferase

2 (FUT2) gene, in patients with CD, was associated both with a shift of overall

community composition as well as alterations of abundances of specific

microbes within the community (241-244).The FUT2 gene is responsible for

regulating the secretion of the blood group antigens in the digestive mucosa

and from the secretory glands (243). The suggestion has been that individulas

that do not secret blood group products in their mucosa due to a FUT2

mutation, may have a disrupted immunogenic/homeostatic equilibrium, due to

changes in carbohydrate levels in the mucos, that shifts the community of

bacteria leaving the patient more susceptible to the development of chronic

mucosal inflammation (241-244).

These studies, however, did not address the central questions of whether the

observed alterations were functional nor what their potential mechanisms of

action risk for developing IBD is. The questions of function and mechanism

are of particular relevance, because microbial composition has been shown to

158

exhibit large inter-individual variations when compared with function-based

analyses, even in otherwise healthy individuals (233). A more general

approach to this question has also linked genetic loci with alterations of

bacterial abundances in the intestines of in mice (136), however in humans

and outside of a specific disease setting, approaches using twins has failed to

reveal significant genotype effects on microbiome composition and diversity

(114, 118).

The investigation of underlying host genetics and microbiome composition

has only recently begun to be explored; therefore we sort to better understand

the role of AS susceptibility variants on microbiome composition. There are a

number of genes involved peptide presentation, immunity, microbial sensing

and response to infection associated with AS, and given the increased

incidence of clinical and sub clinical gut disease in AS patients, we asked

what impact does host genetics have on overall microbiome composition. The

demonstration of individual genes and genetic profiles associated with risk of

AS, also associated with microbes known to drive the AS microbial signature,

particularly if found in both health and disease, would be supportive of model

whereby susceptibility genes modulate the microbiome to influence disease

risk. We have shown, in an otherwise healthy Twin cohort, that genes

associated with developing AS were linked to families of bacteria known to be

inflammatory and key indicator species in the AS microbiome (102-104).

Given that AS is a complex genetic disease, an allelic risk score was

constructed to test if the combined genetic profile, rather than single genes,

was associated with changes in the intestinal microbiome (245). The allelic

risk score linked an AS genetic profile with the families of Ruminococcaceae,

Lachnospiraceae and Clostridia, which are keystone species of the AS

microbiome and members of the core microbiome as detailed in Chapter 3.

Further investigation of individual genes involved in microbial response,

peptide processing, trimming and presentation were also correlated with these

same families of bacteria. There is currently a lack of consensus with respect

to measuring and dissecting associations between underlying host genetics

and members of the intestinal microbiome. In this study, many approaches

were attempted, including false discovery rate and Bonferroni correction,

159

however none of the variants survived multiple testing corrections; therefore,

we opted for a nominal P-value of <0.05. Nevertheless, the associations seen

between AS-associated SNPs and particular bacterial families, although not

statically significant, intimates that host genetics play a role in not only overall

microbiome composition, but also structure of the core microbiome.

However, whether the underlying host genetics influences the microbiome

directly or indirectly via the immune system is still yet to be unravelled. In this

study we were only able to assess the association of AS susceptibility variants

with microbiome composition in the TwinsUK cohort and we were not able to

interrogate the effect of genetic variants known not to affect disease risk or

known to affect risk to other forms of immune-mediated diseases. A more

extensive analysis would permit the validation of our approach undertaken, by

providing the appropriate genetic loci to serve as negative controls and

ascertain possible baseline influence of genetics influence on a healthy

microbiome.

Whilst further research is required, here we show evidence that genetic

variants that affect disease susceptibility were also associated with changes

of the microbiome composition. These observations open a new avenue of

research for our understanding of AS biology, but caution needs to be taken in

the interpretations of the results, as the directionality of causation has not

been explored and warrants further investigation.

161

5.5 Supplementary Tables Supplementary table 5.8 OTUs reaching nominal significance (P<0.05) associated with BACH2. 50 OTUs associated with rs17765610 and 81 with rs639575; the majority of which belong to Ruminococcaceae, Lachnospiraceae and Bacteroidaceae families. Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_). SNP Gene OTU Beta Pvalue Taxonomy rs17765610 BACH2 681370 -0.171 1.87E-03 Actinobacteria; c__Actinobacteria; o__Bifidobacteriales; f__Bifidobacteriaceae;

g__Bifidobacterium; s__pseudolongum rs17765610 BACH2 3439403 0.191 1.87E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides; s__ rs17765610 BACH2 3563235 0.343 1.87E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides; s__ rs17765610 BACH2 3588390 0.345 1.87E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides; s__ rs17765610 BACH2 3600504 0.454 1.87E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides; s__ rs17765610 BACH2 3931537 0.282 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__Clostridium; s__ rs17765610 BACH2 4469576 0.135 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs17765610 BACH2 199761 0.224 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs17765610 BACH2 180468 0.244 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Oscillospira; s__ rs17765610 BACH2 1504042 0.309 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Oscillospira; s__ rs17765610 BACH2 4217963 -0.089 3.49E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs17765610 BACH2 189524 -0.111 3.49E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs17765610 BACH2 216599 -0.298 7.89E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Rikenellaceae; g__; s__ rs17765610 BACH2 4377149 -0.209 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ rs17765610 BACH2 182643 -0.186 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__; s__ rs17765610 BACH2 168071 -0.196 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs17765610 BACH2 183439 -0.348 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs17765610 BACH2 186416 -0.32 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs17765610 BACH2 189176 -0.165 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__[Ruminococcus];

s__gnavus

162

rs17765610 BACH2 193769 -0.199 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__[Ruminococcus]; s__gnavus

rs17765610 BACH2 193666 -0.25 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__ rs17765610 BACH2 363735 -0.209 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__ rs17765610 BACH2 1614788 -0.384 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Coprococcus; s__catus rs17765610 BACH2 176980 -0.376 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Dorea; s__ rs17765610 BACH2 181167 -0.161 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Dorea; s__ rs17765610 BACH2 184114 -0.243 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs17765610 BACH2 185814 -0.182 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs17765610 BACH2 192720 -0.269 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs17765610 BACH2 295258 -0.22 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs17765610 BACH2 312882 -0.231 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs17765610 BACH2 368236 -0.096 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs17765610 BACH2 4468466 -0.297 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs17765610 BACH2 4421273 -0.168 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Oscillospira; s__ rs17765610 BACH2 163243 -0.228 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Ruminococcus; s__ rs17765610 BACH2 16054 -0.163 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Ruminococcus;

s__callidus rs17765610 BACH2 157327 0.175 9.92E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides; s__ rs17765610 BACH2 3426658 0.215 9.92E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides; s__ rs17765610 BACH2 211706 0.077 3.17E-02 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides; s__ rs17765610 BACH2 2875735 0.097 3.17E-02 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides; s__ rs17765610 BACH2 4326080 0.085 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__Clostridium; s__ rs17765610 BACH2 184897 0.08 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Faecalibacterium;

s__prausnitzii rs17765610 BACH2 182073 -0.127 4.34E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ rs17765610 BACH2 2996838 -0.123 4.34E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__

163

rs17765610 BACH2 4419459 -0.144 4.34E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ rs17765610 BACH2 173996 -0.135 4.34E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs17765610 BACH2 306499 -0.133 4.34E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs17765610 BACH2 177758 -0.144 4.34E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Dorea; s__ rs17765610 BACH2 181826 -0.127 4.34E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs17765610 BACH2 185164 -0.144 4.34E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs17765610 BACH2 4428714 -0.161 4.34E-02 Tenericutes; c__Mollicutes; o__RF39; f__; g__; s__ rs639575 BACH2 193672 0.23 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__; s__ rs639575 BACH2 125624 0.22 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs639575 BACH2 168071 0.182 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs639575 BACH2 178018 0.161 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs639575 BACH2 190961 0.151 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs639575 BACH2 192127 0.175 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs639575 BACH2 193477 0.105 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs639575 BACH2 305318 0.205 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs639575 BACH2 3275562 0.321 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs639575 BACH2 4354235 0.341 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs639575 BACH2 4464173 0.214 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs639575 BACH2 4468384 0.179 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__ rs639575 BACH2 4374302 0.207 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Dorea; s__ rs639575 BACH2 184114 0.186 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs639575 BACH2 185583 0.196 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs639575 BACH2 185814 0.174 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs639575 BACH2 187180 0.144 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs639575 BACH2 187924 0.155 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs639575 BACH2 332185 0.131 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__

164

rs639575 BACH2 3756485 0.091 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs639575 BACH2 4437359 0.252 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Oscillospira; s__ rs639575 BACH2 216599 -0.305 3.49E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Rikenellaceae; g__; s__ rs639575 BACH2 217109 -0.088 3.49E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Christensenellaceae; g__; s__ rs639575 BACH2 231952 -0.121 3.49E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Christensenellaceae; g__; s__ rs639575 BACH2 181452 -0.072 3.49E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs639575 BACH2 195494 -0.066 3.49E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs639575 BACH2 197624 -0.104 3.49E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs639575 BACH2 718358 -0.056 3.49E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs639575 BACH2 113003 -0.062 3.49E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Ruminococcus; s__ rs639575 BACH2 3709990 -0.103 3.49E-03 Firmicutes; c__Erysipelotrichi; o__Erysipelotrichales; f__Erysipelotrichaceae;

g__[Eubacterium]; s__dolichum rs639575 BACH2 181074 -0.096 3.49E-03 Firmicutes; c__Erysipelotrichi; o__Erysipelotrichales; f__Erysipelotrichaceae; g__cc_115; s__ rs639575 BACH2 151870 -0.117 3.49E-03 Firmicutes; c__Erysipelotrichi; o__Erysipelotrichales; f__Erysipelotrichaceae; g__Coprobacillus;

s__ rs639575 BACH2 656881 -0.099 3.49E-03 Proteobacteria; c__Gammaproteobacteria; o__Enterobacteriales; f__Enterobacteriaceae; g__;

s__ rs639575 BACH2 4424239 0.136 3.76E-03 Firmicutes; c__Bacilli; o__Lactobacillales; f__Streptococcaceae; g__Streptococcus; s__ rs639575 BACH2 4347159 -0.355 7.89E-03 Actinobacteria; c__Actinobacteria; o__Bifidobacteriales; f__Bifidobacteriaceae;

g__Bifidobacterium; s__adolescentis rs639575 BACH2 72820 -0.218 7.89E-03 Actinobacteria; c__Actinobacteria; o__Bifidobacteriales; f__Bifidobacteriaceae;

g__Bifidobacterium; s__longum rs639575 BACH2 4449054 -0.165 7.89E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides; s__ rs639575 BACH2 193528 -0.266 7.89E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides;

s__caccae rs639575 BACH2 688923 -0.184 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs639575 BACH2 359872 -0.196 7.89E-03 Proteobacteria; c__Deltaproteobacteria; o__Desulfovibrionales; f__Desulfovibrionaceae;

g__Bilophila; s__ rs639575 BACH2 782953 -0.204 7.89E-03 Proteobacteria; c__Gammaproteobacteria; o__Enterobacteriales; f__Enterobacteriaceae; g__;

s__

165

rs639575 BACH2 4391262 -0.156 7.89E-03 Proteobacteria; c__Gammaproteobacteria; o__Enterobacteriales; f__Enterobacteriaceae; g__; s__



rs639575 BACH2 4425214 0.192 1.54E-02 Firmicutes; c__Bacilli; o__Lactobacillales; f__Streptococcaceae; g__Streptococcus; s__ rs639575 BACH2 157338 0.125 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ rs639575 BACH2 179744 0.063 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ rs639575 BACH2 180352 0.089 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ rs639575 BACH2 196371 0.075 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ rs639575 BACH2 4419459 0.125 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__ rs639575 BACH2 193831 0.125 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__Clostridium; s__ rs639575 BACH2 2840201 0.08 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__Clostridium; s__ rs639575 BACH2 179512 0.068 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__SMB53; s__ rs639575 BACH2 179847 0.064 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs639575 BACH2 186735 0.104 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs639575 BACH2 191214 0.108 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs639575 BACH2 192735 0.086 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs639575 BACH2 193654 0.095 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs639575 BACH2 2688035 0.099 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs639575 BACH2 4331782 0.105 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs639575 BACH2 4450010 0.082 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__ rs639575 BACH2 179557 0.05 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__ rs639575 BACH2 186022 0.069 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__ rs639575 BACH2 189067 0.057 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__ rs639575 BACH2 191779 0.077 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__

166

rs639575 BACH2 189459 0.054 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Coprococcus; s__ rs639575 BACH2 105287 0.06 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs639575 BACH2 188044 0.058 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs639575 BACH2 189924 0.084 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs639575 BACH2 192438 0.084 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs639575 BACH2 313524 0.109 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs639575 BACH2 3422630 0.09 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs639575 BACH2 174439 0.112 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Faecalibacterium;

s__prausnitzii rs639575 BACH2 357261 0.098 3.17E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Ruminococcus; s__ rs639575 BACH2 681370 -0.134 4.34E-02 Actinobacteria; c__Actinobacteria; o__Bifidobacteriales; f__Bifidobacteriaceae;

g__Bifidobacterium; s__pseudolongum rs639575 BACH2 4365130 -0.131 4.34E-02 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Porphyromonadaceae;

g__Parabacteroides; s__distasonis rs639575 BACH2 4331760 -0.155 4.34E-02 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Rikenellaceae; g__; s__ rs639575 BACH2 2442706 -0.16 4.34E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Christensenellaceae; g__; s__ rs639575 BACH2 212532 -0.151 4.34E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs639575 BACH2 4439469 -0.139 4.34E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__ rs639575 BACH2 2979308 -0.128 4.34E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Ruminococcus; s__

167

Supplementary table 5.9 OTUs reaching nominal significance (P<0.05) associated with NOS2. NOS2 rs2297518 showed 65 OTU associations and rs2531875 with 18 OTUs associations, all dominated by Ruminococcaceae, Lachnospiraceae and Bacteroidaceae families along with Rikenellaceae and Clostridiaceae Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_). SNP Gene OTU Beta Pvalue Taxonomy rs2297518 NOS2 182643 -0.159 1.67E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__; s__

rs2297518 NOS2 4383953 -0.268 1.67E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__; s__

rs2297518 NOS2 304779 -0.092 1.67E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__Clostridium; s__perfringens

rs2297518 NOS2 2123717 -0.054 1.67E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__

rs2297518 NOS2 181826 -0.068 1.67E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__

rs2297518 NOS2 4020502 0.245 3.30E-02 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides; s__

rs2297518 NOS2 107044 0.122 3.30E-02 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Rikenellaceae; g__; s__



rs2297518 NOS2 4424239 0.078 3.30E-02 Firmicutes; c__Bacilli; o__Lactobacillales; f__Streptococcaceae; g__Streptococcus; s__




rs2297518 NOS2 157338 0.122 3.30E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__






168

rs2297518 NOS2 1646171 0.093 3.30E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__; s__

rs2297518 NOS2 3931537 0.179 3.30E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__Clostridium; s__

rs2297518 NOS2 167373 0.164 3.30E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__






rs2297518 NOS2 175559 0.072 3.30E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__












rs2297518 NOS2 179267 0.248 3.30E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Coprococcus; s__




169


rs2297518 NOS2 180462 0.106 3.30E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__








rs2297518 NOS2 199145 0.29 3.30E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Faecalibacterium; s__prausnitzii

rs2297518 NOS2 199702 0.067 3.30E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Faecalibacterium; s__prausnitzii

rs2297518 NOS2 146554 0.379 3.30E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Ruminococcus; s__







rs2297518 NOS2 798581 0.212 3.30E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Ruminococcus; s__bromii

rs2297518 NOS2 191355 0.191 3.30E-02 Firmicutes; c__Erysipelotrichi; o__Erysipelotrichales; f__Erysipelotrichaceae; g__; s__



rs2531875 NOS2 161423 -0.131 2.07E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides; s__

170



rs2531875 NOS2 193528 -0.219 2.07E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides; s__caccae

rs2531875 NOS2 368490 -0.139 2.07E-03 Firmicutes; c__Bacilli; o__Turicibacterales; f__Turicibacteraceae; g__Turicibacter; s__

rs2531875 NOS2 537219 -0.077 2.07E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__

rs2531875 NOS2 4372528 -0.147 2.07E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__



rs2531875 NOS2 304779 -0.088 2.07E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__Clostridium; s__perfringens







rs2531875 NOS2 16054 -0.106 2.07E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Ruminococcus; s__callidus

rs2531875 NOS2 4453501 -0.092 2.07E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Veillonellaceae; g__Veillonella; s__dispar

rs2531875 NOS2 4477696 -0.115 2.07E-03 Proteobacteria; c__Gammaproteobacteria; o__Pasteurellales; f__Pasteurellaceae; g__Haemophilus; s__parainfluenzae

171

Supplementary table 5.10 OTUs reaching nominal significance (P<0.05) associated with CARD9 SNP, rs1128905, predominately Lachnospiraceae, Veillonellaceae and Ruminococcaceae, including Faecalibacterium prausnitzii. Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_). SNP Gene OTU Beta Pvalue Taxonomy

rs1128905 CARD9 186981 0.144 1.87E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__[Barnesiellaceae]; g__; s__

rs1128905 CARD9 157327 0.074 1.87E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides; s__




rs1128905 CARD9 107044 0.141 1.87E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Rikenellaceae; g__; s__

rs1128905 CARD9 216599 0.165 1.87E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Rikenellaceae; g__; s__

rs1128905 CARD9 4425663 0.122 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__

rs1128905 CARD9 217109 0.081 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Christensenellaceae; g__; s__

rs1128905 CARD9 181452 0.056 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__



rs1128905 CARD9 183686 0.114 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__






172



rs1128905 CARD9 4342104 0.056 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Anaerotruncus; s__

rs1128905 CARD9 2979308 0.095 1.87E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Ruminococcus; s__

rs1128905 CARD9 3709990 0.081 1.87E-03 Firmicutes; c__Erysipelotrichi; o__Erysipelotrichales; f__Erysipelotrichaceae; g__[Eubacterium];

s__dolichum

rs1128905 CARD9 4472721 0.232 7.45E-03 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__[Odoribacteraceae]; g__Odoribacter; s__


rs1128905 CARD9 176113 -0.092 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__





rs1128905 CARD9 3931537 -0.19 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__Clostridium; s__

rs1128905 CARD9 175704 -0.041 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__








173











rs1128905 CARD9 181754 -0.104 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__[Ruminococcus]; s__

rs1128905 CARD9 176381 -0.053 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__






rs1128905 CARD9 28914 -0.061 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Roseburia; s__




rs1128905 CARD9 105287 -0.037 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__

174























175










rs1128905 CARD9 174611 -0.221 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Faecalibacterium;

s__prausnitzii


s__prausnitzii


s__prausnitzii


s__prausnitzii


s__prausnitzii


s__prausnitzii

rs1128905 CARD9 180468 -0.12 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Oscillospira; s__


176


rs1128905 CARD9 4458959 -0.071 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Veillonellaceae; g__Veillonella; s__

rs1128905 CARD9 4388775 -0.106 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Veillonellaceae; g__Veillonella; s__dispar

rs1128905 CARD9 4453501 -0.105 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Veillonellaceae; g__Veillonella; s__dispar

rs1128905 CARD9 4477696 -0.112 7.89E-03 Proteobacteria; c__Gammaproteobacteria; o__Pasteurellales; f__Pasteurellaceae;

g__Haemophilus; s__parainfluenzae

177

Supplementary table 5.11 OTUs reaching nominal significance (P<0.05) for NKX2-3, ICOSLG and SH2B3. The families of Ruminococcaceae, Lachnospiraceae dominating the associations. Taxonomy is shown with full classification from phylum through class (c_), order (o_), family (f_), genus (g_) and species (s_). SNP Gene OTU Beta Pvalue Taxonomy

rs7282490 ICOSLG 191361 -0.449 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__


rs7282490 ICOSLG 195937 -0.397 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Blautia; s__

rs7282490 ICOSLG 182289 -0.415 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Coprococcus; s__

rs7282490 ICOSLG 4393532 -0.25 1.83E-02 Actinobacteria; c__Coriobacteriia; o__Coriobacteriales; f__Coriobacteriaceae; g__Eggerthella;

s__lenta

rs7282490 ICOSLG 176862 -0.139 1.83E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__

rs7282490 ICOSLG 322580 -0.12 1.83E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__

rs7282490 ICOSLG 1646171 -0.232 1.83E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__; s__

rs7282490 ICOSLG 199694 -0.136 1.83E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__Clostridium; s__









178









rs7282490 ICOSLG 174752 -0.199 1.83E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__




rs7282490 ICOSLG 2979308 -0.189 1.83E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Ruminococcus; s__

rs11190133 NKX2-3 197004 -0.393 7.89E-03 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__

rs11190133 NKX2-3 4343627 -0.302 1.67E-02 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides;

s__fragilis

rs11190133 NKX2-3 197364 -0.249 1.67E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__



rs11190133 NKX2-3 2840201 -0.118 1.67E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__Clostridium; s__



179







rs11190133 NKX2-3 2740950 -0.165 1.67E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Coprococcus; s__

rs11190133 NKX2-3 4374302 -0.333 1.67E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Dorea; s__

rs11190133 NKX2-3 162623 -0.131 1.67E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Roseburia; s__

rs11190133 NKX2-3 178485 -0.156 1.67E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__






rs11190133 NKX2-3 4458959 -0.131 1.67E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Veillonellaceae; g__Veillonella; s__

rs11190133 NKX2-3 4453501 -0.194 1.67E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Veillonellaceae; g__Veillonella; s__dispar

rs11190133 NKX2-3 4323124 0.229 3.30E-02 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__[Barnesiellaceae]; g__; s__

rs11190133 NKX2-3 185420 0.418 3.30E-02 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides; s__




180

rs11190133 NKX2-3 344525 0.16 3.30E-02 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides;

s__eggerthii

rs11190133 NKX2-3 4405146 0.124 3.30E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__

rs11190133 NKX2-3 2318497 0.247 3.30E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__; s__

rs11190133 NKX2-3 232828 0.113 3.30E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__

rs11190133 NKX2-3 4427290 0.233 3.30E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__

rs11190133 NKX2-3 147969 0.259 3.30E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__Ruminococcus; s__

rs11065898 SH2B3 332732 -0.172 3.26E-02 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides; s__

rs11065898 SH2B3 193654 -0.1 3.26E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__

rs11065898 SH2B3 189459 -0.077 3.26E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Coprococcus; s__

rs11065898 SH2B3 1740283 -0.108 3.26E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Roseburia; s__

rs11065898 SH2B3 19611 -0.152 3.26E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__




rs11065898 SH2B3 4381553 0.355 4.46E-02 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides; s__

rs11065898 SH2B3 4436552 0.137 4.46E-02 Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Prevotellaceae; g__Prevotella; s__copri

rs11065898 SH2B3 194597 0.19 4.46E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__; g__; s__

rs11065898 SH2B3 2840201 0.085 4.46E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Clostridiaceae; g__Clostridium; s__

rs11065898 SH2B3 181452 0.098 4.46E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__; s__



181

rs11065898 SH2B3 1667433 0.118 4.46E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Lachnospiraceae; g__Dorea; s__

rs11065898 SH2B3 182044 0.239 4.46E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__

rs11065898 SH2B3 350503 0.183 4.46E-02 Firmicutes; c__Clostridia; o__Clostridiales; f__Ruminococcaceae; g__; s__

183

6 Chapter 6 Final Discussion

6.1 Overview and Conclusions

The work presented in this thesis examines the role of the gut microbiome in AS by

characterising the AS microbiome. It examines how changes in host genetics influences

the composition of intestinal microbial composition as well as the association between AS

susceptibility loci and overall microbiome composition. The key findings of this work, its

limitations and future directions are summarised and discussed below.

Microbial involvement has been suggested in AS, however, a definitive link is yet to be

established (17, 138, 142, 143, 202, 246, 247). To date there has been no comprehensive

characterisation of the gut microbiome in patients with AS. I sought to determine if the TI of

AS patients carries a distinct microbial signature in comparison to healthy controls.

Microbial profiles from TI biopsies from subjects with recent-onset, TNF-antagonist naïve

AS and HC, were determined using culture-independent 16S rRNA gene sequencing and

analysis techniques developed in Chapter 2. Analysis showed evidence for a discrete

microbial signature in the TI of cases with AS compared to HC. Distinct grouping of AS

cases was discriminated from HC on PCoA with clustering shown to be driven by higher

abundances of Lachnospiraceae, Ruminococcaceae, Rikenellaceae,

Porphyromonadaceae, Bacteroidaceae and decreases in Prevotellaceae and

Veillonellaceae. Interactions between these indicator species within the microbial

community were shown to further shape the AS microbial community signature. No

difference was noted between AS cases and controls in either bacteria known to be

associated with reactive arthritis, or Klebsiella species, which have been proposed to play

a role in triggering AS. Statistically, the relationship between disease status and microbial

community composition was confirmed indicating that the differences between the

microbial community compositions was driven by disease state and not due to

heterogeneity in the biopsy or lack of technical reproducibility. These results are consistent

with the hypothesis that genes associated with AS act at least in part through effects on

the gut microbiome. While the data here shows a clear and distinct microbial signature in

AS cases, the effect of TNF treatment on the microbiome composition and its contribution

to disease pathogenesis is yet to be explored. Cause and affect interactions between the

host genome, immune system and the gut microbiome also remains to be dissected.

184

With over 40 loci currently associated with ankylosing spondylitis, one of the strongest

associations, along with HLA-B27, is ERAP1 (24, 62). Given that loss of function of

ERAP1 is protective in AS, In Chapter 4 I asked how a lack or down regulation of ERAP

influenced inherent gut microbiome composition, the antigen repertoire and the immune

system. To investigate this hypothesis I explored if the loss of ERAP, in a mouse model,

caused a shift in the intestinal flora towards an anti-inflammatory composition. Consistent

with this hypothesis, 16S community profiling revealed that the gut microbiome in ERAP-/-

mice did shift. However, it was towards a more inflammatory profile, with an increase of

bacteria such as Rikenellaceae, Lachnospiraceae and Ruminococcus which are not only

known to elicit a strong immune response, but are also key bacteria in the TI of AS

patients (103, 248, 249). To minimise such potential confounders, the intestinal microbial

composition of a single litter of ERAP mice comprising of ERAP+/+, ERAP+/- and ERAP -/-

was explored. Analysis showed that ERAP+/-and ERAP -/- mice were more similar to each

other and clustered away from the ERAP+/+ mice, suggesting that a change in genotype,

from ERAP+/+ to ERAP+/-, is sufficient to influence the overall structure and composition

of the gut microbiome. This is a small study with only one litter studied, and larger

numbers are needed to better characterise the impact of genotype on the intestinal

microbiome. Nonetheless, this is consistent with the hypothesis that AS-associated genes

contribute to disease aetiopathogenesis through effects of the underlying host genetics on

intestinal microbial composition.

The fifth chapter of this thesis further investigated the influence of host genetics,

specifically AS risk loci, on microbiome composition by examining the association between

AS SNPs and intestinal microbiome composition in a non-AS, otherwise healthy twin

cohort. Given the number of genes involved in peptide presentation, immunity, microbial

sensing and response to infection, and the increased incidence of clinical and sub clinical

gut disease, I asked what impact does host genetics have on overall microbiome

composition. Results showed that in an otherwise healthy cohort of twins, genes

associated with developing AS were associated with families of bacteria known to be

inflammatory and key indicator species in the AS microbiome. The allelic risk score linked

the AS genetic profile with the families of Ruminococcaceae, Lachnospiraceae and the

order Clostridiales, which are keystone species of the AS microbiome. Further

investigation of individual genes involved in microbial response, peptide processing,

trimming and presentation showed correlation with these same AS indicator species. This

185

suggests that host genetics plays a role in overall microbiome composition however

whether the underlying host genetics influences the microbiome directly or indirectly via

the immune system is still yet to be unravelled. In this study we were only able to assess

the association of AS susceptibility variants with microbiome composition in the TwinsUK

cohort and were unable to interrogate the effect of other genetic variants known not to

affect disease risk or known to affect risk to other forms of immune-mediated diseases. A

more extensive analysis would permit the validation of our approach undertaken by

providing the appropriate genetic loci to serve as negative controls and ascertain possible

baseline influence of genetics influence on a healthy microbiome. Caution needs to be

taken in the interpretations of these results as directionality of causation has not been

explored and requires further investigation.

6.2 Future directions

There is mounting evidence for a role of the gut microbiome in the aetiopathogenesis in

AS, although exact mechanism remains unclear. To further explore the hypothesis that

genes predisposing to AS do so by influencing the gut microbiome, either directly or

indirectly via the immune system, will require substantially larger cohorts of several

hundred patients and controls with matched genetic, clinical and microbiome data. Larger

studies may identify specific disease triggering bacteria, and further define the AS

microbiome profile and interactions, raising the possibility both the use of the microbiome

as a biomarker as well as for novel therapeutic interventions.

The gastrointestinal tract is the primary site for interaction between microbes and the host

immune system. Moreover, the microbes found in the gut help to shape host immune

systems from an early age (210, 211), although little is understood regarding how

individual and bacterial communities interact with the immune system creating an

inflammatory environment. To date only a handful of bacteria-immune interactions have

been described, such as Prevotella and the innate immune system (249). The mechanism

of interaction between other AS indicator species of bacteria such as Ruminococcus,

Rikenella and Clostridia and the immune system causing disease is yet to be investigated.

Loss of members of the Bacteroidetes and Firmicutes phyla, especially Clostridia, have

been associated with the risk alleles of CD genes NOD2 (250) and ATG16L1 (248). These

two genes are associated with abnormal Paneth cell function suggesting alteration of

secreted antimicrobials from intestinal Paneth cells perturbs the intestinal microbiome,

186

which leads to inflammation and disease. Unravelling the interactions between AS

indicator species and the immune system will help us understand to role of the gut

microbiota in AS pathogenesis and how immune system sculpts the microbiome, and in

what way microbiome counters the immune response perpetuating the inflammatory

environment.

The intestinal microbiome is a dynamic and constantly changing environment (127, 128,

251). Whist the microbiome ebbs and flows with daily living (251), overall the microbiota

remains reasonably stable for months, and possibly even years (127, 252). However, most

of what we know regarding the make-up of the human gut microbiome and its stability is

based off stool studies and not mucosal biopsies (114, 118, 126, 127, 183, 252, 253).

Several studies have shown that the microbiome profile from stool differs from the profile

obtained from mucosal biopsy suggesting that stool, while the most accessible sample, is

not optimal for detailing gut microbiome composition and may skew our understanding

(152, 153). Moreover, studies across mice and humans suggest that common aspects of a

modern Western lifestyle including antibiotic use (254, 255) and high fat and fibre poor

diets (118, 119, 256), can persistently alter commensal microbial communities. In turn,

these microbial disturbances may increase susceptibility to pathogens (257), obesity (114,

121, 197), and autoimmune disease (199, 224, 258). This makes dissecting cause and

effect challenging given the significant environmental factors which can shape the

microbiome as well as normal temporal variance; and we have only been focusing on the

bacteria, not archea, viruses or fungi which also inhabit the intestinal tract. The successful

use of the gut microbiome as a therapeutic target requires a deeper understanding of

factors that shape the community composition including age, diet, geographical location,

gender as well as underlying host genetics. Investigations to date have largely focused on

the bacterial composition of the microbiome, and so know relatively little about the viruses

and fungi that also inhabit the gut. Despite the considerable limitations in sequencing and

functional annotation of the eukaryotic viruses present in the gastrointestinal tract, recent

deep-sequencing efforts have revealed the existence of a complex enteric virome (259,

260). A recent study revealed the surprising beneficial role viruses may play a role in a

healthy microbiome. Murine norovirus (MNV) infection of germ-free or antibiotic-treated

mice rescued both intestinal morphology and lymphocyte function in the absence of overt

inflammation and disease. This study demonstrated that eukaryotic viruses have the

capacity to support intestinal homeostasis and shape mucosal immunity, similarly to

commensal bacteria (261). This work demonstrates how little we currently known and truly

187

understand about the composition of a healthy intestinal microbiome and work along these

lines will improve our knowledge of the intestinal microbiome as well as well microbiome

host interactions critical for overall health. Particularly given that ReA is triggered by a

bacterial infection and that it is looking likely that microbiome has a critical role in AS

pathogenesis.

The gut microbiota is increasingly looked at as a target for novel therapies, including

microbiome modulation with probiotics and faecal microbiota transplants (FMT). These

novel therapies are being increasingly used to treat patients with debilitating Clostridium

difficile infections, where conventional treatments and antibiotics have failed. The number

and frequency of FMT is growing exponentially (262), and in response biobanks for stool

have been created at Massachusetts General Hospital in Boston and Emory University

Hospital in Atlanta, Georgia where ‘healthy’ stool is screened and stored for medical use in

approved cases (263). A small trial comparing antibiotic treatment to FMT for C. difficile

showed that FMT was more effective at resolving patient symptoms as antibiotics alone

(264). The effectiveness of FMT caused the trial to be stopped early. Non-randomised

studies of C. difficile treated with either antibiotics only or FMT have shown a typical

success rate of about 90% (264-266). While the majority of testimonies surrounding FMT

are positive, there are still many unknowns. Screening of the donor faeces is currently not

standardised and can be very basic. We all carry different bacteria, viruses and parasites

at any given moment (126), and while they not be harmful to donor, potential pathogens as

well as the healthy flora may inadvertently be to transplanted. Little long-term safety data

exists for FMT recipients (267), however transient abdominal pain and bloated have been

observed. It may not be as simple as taking microbes from one person and giving them to

another, just as other transplants like bone marrow and solid organ, require typing and

cross matching prior to transplant. Research is underway in several laboratories, such as

Openbiome (http://www.openbiome.org/) in the US, working on patient screening

procedures as well as overall FMT regulation, similar to the model used by the Red Cross

– except for faeces (263, 266). This is to ensure the microbes transplanted behave as we

expect them to, as well as minimising unexpected consequences, with the goal of

eventually not having to use a donor for the transplant at all, but tailoring the combination

of synthetic microbes grown in the lab tailored to each individual patient (267, 268). FMT is

becoming more accepted as a valid treatment option for C. difficile infections and is

currently being trialed for patients with CD (227). Though, the hope that manipulating the

188

gut microbiome to treat diseases other than C. difficile infection is still speculative (262,

269).

The role of the host genetics in shaping intestinal microbial community composition is

unclear. Recent animal studies have shown that host gene deletions resulted in shifts in

microbiota composition as well as linking certain genetic polymorphisms with microbial

abundances (135, 136). This highlights the importance of underlying host genetics in

community composition and raises the question: if underlying host genetics influences

microbiome composition, is FMT in complex genetic diseases futile because the transplant

will never permanently shift the intestinal flora? In diseases such as AS and CD, the

expectation is that since the microbiome influences the immune system, then transplanting

the gut with healthy intestinal flora will lead to interactions with the immune system which

create a less inflammatory environment reducing gut disease and providing overall

improvement of symptoms and maybe even disease. The caveat here is that many genes

associated with AS and CD are involved in mucosal immunology and microbial processing,

so underlying host genetics may eventually override the transplant.

What is clear is the need to better understand our second genome and to do this we

require better tools for comparison and analysis. Currently there is a lack of central open

access database and repository for human microbial community data. There is the NIH

Human microbiome project database (http://www.hmpdacc.org/) with 16S rRNA and

annotated microbial genomes img/hmp where all the human microbiome project data are

deposited, the MetaHit database (http://www.metahit.eu/), National Center for

Biotechnology Information (NCBI) (http://www.ncbi.nlm.nih.gov/) as well as the European

Bioinformatics Institute (http://www.ebi.ac.uk/) where the recent TwinsUK 16S rRNA

sequence deposited (137). This demonstrates how spread out all the data is. Some of the

data is annotated, meaning the individual whole microbial genome has been sequenced

and function and biochemical pathways have been described, and some of the data is just

the 16S rRNA sequence. What is needed is the microbial equivalent of HapMap or UK10K

were there is access to samples from varying habitats (mouth, skin, gut) with well

annotated taxonomy and function so that eventually microbial communities can be imputed

in the way we currently impute human SNPs. American gut project

(http://humanfoodproject.com/americangut/) is currently one of the largest open-source

project endeavouring to understand the microbial diversity of the gut is leading the way

however, we need reasonable metadata accompanying the microbial data such as age,

189

gender, diet, lifestyle as well as geographic location and ethnicity. Much time and

development is being dedicated to improving the 16S tools we currently have as well as

creating new ones such as QIIME (169), Mothur (270) and LotuS (271) . However moving

beyond 16S rRNA taxonomy is imperative for furthering our understanding of the

interaction between host and microbes to better understand disease. Whilst using the gut

microbiome as a phenotype for association studies for the purpose of using Plink (272)

and Merlin (239) to test for association testing is a start, these programs are not designed

to be used for microbiome analysis and do not allow within microbiome interactions and

network dynamics to be taken into account, such as co-occurrence (spa). In fact no

programs currently do. The microbial network packages such as Spaa (176) and SparCC

(273), which have been borrowed from environmental microbial ecology are still rather

crude. There is of coarse OTU networks and cytoscape as enabled in QIIME however, this

program is more big picture and does not lend its self to the within clade analysis and

direction of correlation. There are currently tools for the profiling of communities and

clades in metagenomic data, including bacteria, viruses and archea using MetaPhiAn

(274); high dimensional biomarker discovery with LEfSe (275); determining the

presence/absence and abundance of microbial pathways in a community from

metagenomic data using HUMaN (276); prediction of function from 16S rRNA data using

PICRUSt (115); as well as for multivariate analysis with linear models (MaAsLin) for testing

associations between clinical metadata and the microbial community abundance or

function (277). However, none of the above packages allow for the integration or

correlation of human data, such as genotype information, with microbiome and

metagenomic data and this is currently what is lacking.

16S taxonomy does not inform us about what microbes are active and what they are

expressing (278), let alone if expression is in reaction to host gene transcription (135,

136). To better understand how host and commensal microbes interact, we also need

better tools to examine interaction at the transcriptomic level (279). Metagenome analyses

of the human intestine has previously identified genes and pathways involved in the

transport and metabolism of simple carbohydrate substrates are enriched in the intestine

microbiome (114, 278, 280) and that alterations in these pathways are linked to obesity,

suggesting their importance for the appropriate functioning of this microbial ecosystem.

However, elucidation of the actual activity and metabolic role of individual microbial

members within the intestinal microbiome is still limited. Much of the microbial

transcriptomic studies have used stool (114, 278, 280), as the use of intestinal biopsies is

190

still technically very difficult due to the shear amount of human DNA. There are microbial

enrichment kits that can be used prior to sequencing in an effort to reduce the amount of

human DNA, such as the New England Biosciences microbial enrichment kit (281) and

MoBio Powersoil isolation kit (282), however the use of these kits prior to sequencing, and

the impact on microbial community results, is still being explored. Examining host

transcriptomics in the gut in combination with microbial transcriptions will further our

understanding into how host genetics influences intestinal microbial community

composition and function (283), which contributes to disease.

Finally, the pursuit to understand the pathogenesis of AS, enabling accurate and early

diagnosis and effective treatment, requires better understanding of how underlying host

genetics influences the immune system and the intestinal microbiome. In the next few

years, it is expected that larger studies with matched genotype and microbiome data will

elucidate how it is that genetic variants influence the immune response, which in turn

shapes the gut microbiome, in such a way that causes disease. Microbial profiling of tissue

or stool samples, either alone or in combination with genetic tests, may identify individuals

at high risk of developing AS, enabling either preventative intervention or early diagnosis.

It is with high hopes that the research presented in this thesis and further studies along

these lines improves the understanding of AS pathogenesis, which in turn lead to

improved diagnosis and treatment.

191

7 References

1. Stynes G, Adams J, Tilden D, Garrett S, Wilson J. The burden of ankylosing spondylitis in Australia: An epidemiological and cost of illness model (http://www.m-tag.net/04PAR3.pdf). 2005. 2. Sieper J, Rudwaleit M, Baraliakos X, Brandt J, Braun J, Burgos-Vargas R, et al. The Assessment of SpondyloArthritis international Society (ASAS) handbook: a guide to assess spondyloarthritis. Annals of the Rheumatic Diseases. 2009;68(Suppl 2):ii1-ii44. 3. Dougados M, Linden SVD, Juhlin R, Huitfeldt B, Amor B, Calin A, et al. The European Spondylarthropathy Study Group Preliminary Criteria for the Classification of Spondylarthropathy. Arthritis & Rheumatism. 1991;34(10):1218-27. 4. Pedersen OB, Svendsen AJ, Ejstrup L, Skytthe A, Harris JR, Junker P. Ankylosing spondylitis in Danish and Norwegian twins: occurrence and the relative importance of genetic vs. environmental effectors in disease causation. Scandinavian Journal of Rheumatology. 2008;37(2):120-6. 5. Brown MA, Kennedy LG, Macgregor AJ, Darke C, Duncan E, Shatford JL, et al. Susceptibility to ankylosing spondylitis in twins the role of genes, HLA, and the environment. Arthritis & Rheumatism. 1997;40(10):1823-8. 6. Mielants H, Veys EM, Goemaere S, Goethals K, Cuvelier C, De Vos M. Gut inflammation in the spondyloarthropathies: clinical, radiologic, biologic and genetic features in relation to the type of histology. A prospective study. The Journal of Rheumatology. 1991;18(10):1542-51. 7. Mielants H, Veys EM, Cuvelier C, De Vos M, Botelberghe L. HLA-B27 related arthritis and bowel inflammation. Part 2. Ileocolonoscopy and bowel histology in patients with HLA-B27 related arthritis. The Journal of rheumatology. 1985;12(2):294-8. 8. Ciccia F, Rizzo A, Accardo-Palumbo A, Giardina A, Bombardieri M, Guggino G, et al. Increased expression of interleukin-32 in the inflamed ileum of ankylosing spondylitis patients. Rheumatology. 2012;51(11):1966-72. 9. Reid G, Younes JA, Van der Mei HC, Gloor GB, Knight R, Busscher HJ. Microbiota restoration: natural and supplemented recovery of human microbial communities. Nat Rev Micro. 2011;9(1):27-38. 10. Lubrano E, Ciacci C, Ames P, Mazzacca G, Oriente P, Scarpa R. The arthritis of coeliac disease: prevalence and pattern in 200 adult patients. Rheumatology. 1996;35:1314 - 8. 11. Martinez-Gonzalez O, Cantero-Hinojosa J, Paule-Sastre P, Gomez-Magan JC, Salvatierra-Rios D. Intestinal Permeability in Patients with Ankylosing Spondyllitis and their healthy relatives. Rheumatology. 1994;33(7):644-7. 12. Vaile J, Meddings J, Yacyshyn B, AS. R, Maksymowych W. Bowel permeability and CD45RO expression on circulating CD20+ B cells in patients with ankylosing spondylitis and their relatives. Journal of Rheumatology. 1999 ;26(1):128-35. 13. Morris AJ, Howden CW, Robertson C, Duncan A, Torley H, Sturrock RD, et al. Increased intestinal permeability in ankylosing spondylitis--primary lesion or drug effect? Gut. 1991;32(12):1470-2. 14. Uotila T, Antonen J, Laine J, Kujansuu E, Haapala A-M, Lumio J, et al. Reactive arthritis in a population exposed to an extensive waterborne gastroenteritis outbreak after sewage contamination in Pirkanmaa, Finland. Scandinavian Journal of Rheumatology. 2011;40(5):358-62. 15. Taurog JD, Richardson JA, Croft JT, Simmons WA, Zhou M, Fernández-Sueiro JL, et al. The germfree state prevents development of gut and joint inflammatory disease in HLA-B27 transgenic rats. The Journal of Experimental Medicine. 1994;180(6):2359-64. 16. Stebbings S, Munro K, Simon MA, Tannock G, Highton J, Harmsen H, et al. Comparison of the faecal microflora of patients with ankylosing spondylitis and controls using molecular methods of analysis. Rheumatology. 2002;41(12):1395-401. 17. Stone MA, Payne U, Schentag C, Rahman P, Pacheco-Tena C, Inman RD. Comparative immune responses to candidate arthritogenic bacteria do not confirm a dominant role for Klebsiella pneumonia in the pathogenesis of familial ankylosing spondylitis. Rheumatology. 2004;43(2):148-55.

192

18. Parkes M, Cortes A, van Heel DA, Brown MA. Genetic insights into common pathways and complex relationships among immune-mediated diseases. Nat Rev Genet. 2013;14(9):661-73. 19. Thjodleifsson B, Geirsson áJ, Björnsson S, Bjarnason I. A common genetic background for inflammatory bowel disease and ankylosing spondylitis: A genealogic study in Iceland. Arthritis & Rheumatism. 2007;56(8):2633-9. 20. Danoy P, Pryce K, Hadler J, Bradbury L, Farrar C, Pointon J, et al. Association of variants at 1q32 and STAT3 with ankylosing spondylitis suggests genetic overlap with Crohn's disease. PLoS Genet. 2010;6:e1001195. 21. Reveille J, Sims A, Danoy P, Evans D, Leo P, Pointon J, et al. Genome-wide association study of ankylosing spondylitis identifies non-MHC susceptibility loci. Nat Genet. 2010;42:123 - 7. 22. WTCCC. Association scan of 14,500 nonsynonymous SNPs in four diseases identifies autoimmunity variants. Nat Genet. 2007;39(11):1329-37. 23. Jostins L, Ripke S, Weersma RK, Duerr RH, McGovern DP, Hui KY, et al. Host-microbe interactions have shaped the genetic architecture of inflammatory bowel disease. Nature. 2012;491(7422):119-24. 24. Cortes A, Hadler J, Pointon JP, Robinson PC, Karaderi T, Leo P, et al. Identification of multiple risk variants for ankylosing spondylitis through high-density genotyping of immune-related loci. Nat Genet. 2013;45(7):730-8. 25. Round JL, Mazmanian SK. The gut microbiota shapes intestinal immune responses during health and disease. Nat Rev Immunol. 2009;9(5):313-23. 26. Clemente J, Ursell L, Parfrey L, Knight R. The impact of the gut microbiota on human health: an integrative view. Cell. 2012;148:1258 - 70. 27. Ivanov II, Atarashi K, Manel N, Brodie EL, Shima T, Karaoz U, et al. Induction of Intestinal Th17 Cells by Segmented Filamentous Bacteria. Cell. 2009;139(3):485-98. 28. Duan J, Chung H, Troy E, Kasper DL. Microbial Colonization Drives Expansion of IL-1 Receptor 1-Expressing and IL-17-Producing [gamma]/[delta] T Cells. Cell Host & Microbe. 2010;7(2):140-50. 29. Cho I, Blaser M. The human microbiome: at the interface of health and disease. Nat Rev Genet. 2012;13:10. 30. Maloy KJ, Powrie F. Intestinal homeostasis and its breakdown in inflammatory bowel disease. Nature. 2011;474(7351):298-306. 31. Koslowski MJ, Beisner J, Stange EF, Wehkamp J. Innate antimicrobial host defense in small intestinal Crohn's disease. International Journal of Medical Microbiology. 2010;300(1):34-40. 32. Wehkamp J, Koslowski M, Wang G, Stange EF. Barrier dysfunction due to distinct defensin deficiencies in small intestinal and colonic Crohn's disease. Mucosal Immunol. 2008;1(1s):S67-S74. 33. Simms L, Doecke J, Walsh M, Huang N, Fowler E, Radford-Smith G. Reduced alpha-defensin expression is associated with inflammation and not NOD2 mutation status in ileal Crohn's disease. Gut. 2008;57:903 - 10. 34. Wehkamp J, Salzman NH, Porter E, Nuding S, Weichenthal M, Petras RE, et al. Reduced Paneth cell α-defensins in ileal Crohn's disease. Proceedings of the National Academy of Sciences of the United States of America. 2005;102(50):18129-34. 35. Heazlewood C, Cook M, Eri R, Price G, Tauro S, Taupin D, et al. Aberrant mucin assembly in mice causes endoplasmic reticulum stress and spontaneous inflammation resembling ulcerative colitis. PLoS Med. 2008;5:e54. 36. Sherlock J, Joyce-Shaikh B, Turner S, Chao C, Sathe M, Grein J, et al. IL-23 induces spondyloarthropathy by acting on ROR-gammat+ CD3+CD4-CD8- entheseal resident T cells. Nat Med. 2012;18:1069 - 76. 37. Artis D. Epithelial-cell recognition of commensal bacteria and maintenance of immune homeostasis in the gut. Nat Rev Immunol. 2008;8(6):411-20. 38. Hooper LV, Macpherson AJ. Immune adaptations that maintain homeostasis with the intestinal microbiota. Nat Rev Immunol. 2010;10(3):159-69. 39. Fasano A, Nataro JP. Intestinal epithelial tight junctions as targets for enteric bacteria-derived toxins. Advanced Drug Delivery Reviews. 2004;56(6):795-807. 40. Collins SM. IV. Modulation of intestinal inflammation by stress: basic mechanisms and clinical relevance2001 2001-03-01 00:00:00. G315-G8 p.

193

41. Bruewer M, Luegering A, Kucharzik T, Parkos CA, Madara JL, Hopkins AM, et al. Proinflammatory Cytokines Disrupt Epithelial Barrier Function by Apoptosis-Independent Mechanisms. The Journal of Immunology. 2003;171(11):6164-72. 42. Nusrat A, Turner JR, Madara JL. IV. Regulation of tight junctions by extracellular stimuli: nutrients, cytokines, and immune cells2000 2000-11-01 00:00:00. G851-G7 p. 43. Teahon K, Smethurst P, Levi AJ, Menzies IS, Bjarnason I. Intestinal permeability in patients with Crohn's disease and their first degree relatives. Gut. 1992;33(3):320-3. 44. Podolsky DK. Inflammatory Bowel Disease. New England Journal of Medicine. 2002;347(6):417-29. 45. Vaile J, Meddings J, Yacyshyn B, Russell A, Maksymowych W. Bowel permeability and CD45RO expression on circulating CD20+ B cells in patients with ankylosing spondylitis and their relatives. J Rheumatol. 1999;26:128 - 35. 46. Chabot S, Wagner JS, Farrant S, Neutra MR. TLRs Regulate the Gatekeeping Functions of the Intestinal Follicle-Associated Epithelium. The Journal of Immunology. 2006;176(7):4275-83. 47. Suzuki K, Ha S, Tsuji M, Fagarasan S. Intestinal IgA synthes is: a primitive form of adaptive immunity that regulates microbial communities in the gut. Semin Immunol. 2007;19:127 - 35. 48. Wesa A, Galy A. Increased production of pro-inflammatory cytokines and enhanced T cell responses after activation of human dendritic cells with IL-1 and CD40 ligand. BMC Immunology. 2002;3:14. 49. Cruz CM, Rinna A, Forman HJ, Ventura ALM, Persechini PM, Ojcius DM. ATP Activates a Reactive Oxygen Species-dependent Oxidative Stress Response and Secretion of Proinflammatory Cytokines in Macrophages. Journal of Biological Chemistry. 2007;282(5):2871-9. 50. Hirotani T, Lee PY, Kuwata H, Yamamoto M, Matsumoto M, Kawase I, et al. The Nuclear IκB Protein IκBNS Selectively Inhibits Lipopolysaccharide-Induced IL-6 Production in Macrophages of the Colonic Lamina Propria. The Journal of Immunology. 2005;174(6):3650-7. 51. Denning T, Wang Y-c, Patel S, Williams I, Pulendran B. Lamina propria macrophages and dendritic cells differentially induce regulatory and interleukin 17-producing T cell responses. Nat Immunol. 2007;8:1086 - 94. 52. Glocker E, Kotlarz D, Boztug K, Gertz E, Schaffer A, Noyan F, et al. Inflammatory bowel disease and mutations affecting the interleukin-10 receptor. N Engl J Med. 2009;361:2033 - 45. 53. Pull SL, Doherty JM, Mills JC, Gordon JI, Stappenbeck TS. Activated macrophages are an adaptive element of the colonic epithelial progenitor niche necessary for regenerative responses to injury. Proceedings of the National Academy of Sciences of the United States of America. 2005;102(1):99-104. 54. Oppmann B, Lesley R, Blom B, Timans J, Xu Y, Hunte B, et al. Novel p19 protein engages IL-12p40 to form a cytokine, IL-23, with biological activities similar as well as distinct from IL-12. Immunity. 2000;13:715 - 25. 55. Burton P, Clayton D, Cardon L, Craddock N, Deloukas P, Duncanson A, et al. Association scan of 14,500 nonsynonymous SNPs in four diseases identifies autoimmunity variants. Nat Genet. 2007;39:1329 - 37. 56. Cargill M, Schrodi S, Chang M, Garcia V, Brandon R, Callis K, et al. A large-scale genetic association study confirms IL12B and leads to the identification of IL23R as psoriasis-risk genes. Am J Hum Gen. 2007;80:273 - 90. 57. Duerr R, Taylor K, Brant S, Rioux J, Silverberg M, Daly M, et al. A genome-wide association study identifies IL23R as an inflammatory bowel disease gene. Science. 2006;314:1461 - 3. 58. Ogawa A, Andoh A, Araki Y, Bamba T, Fujiyama Y. Neutralization of interleukin-17 aggravates dextran sulfate sodium-induced colitis in mice. Clin Immunol. 2004;110:55 - 62. 59. Zheng S, Wang J, Horwitz D. Cutting edge: Foxp3+CD4+CD25+ regulatory T cells induced by IL-2 and TGF-beta are resistant to Th17 conversion by IL-6. J Immunol. 2008;180:7112 - 6. 60. Mokuno Y, Matsuguchi T, Takano M, Nish imura H, Washizu J, Ogawa T, et al. Expression of Toll-like receptor 2 on gammadelta T cells bearing invariant Vgamma6/Vdelta1 induced by Escherichia coli infection in mice. J Immunol. 2000;165:931 - 40. 61. Martin B, Hirota K, Cua DJ, Stockinger B, Veldhoen M. Interleukin-17-Producing [gamma][delta] T Cells Selectively Expand in Response to Pathogen Products and Environmental Signals. Immunity. 2009;31(2):321-30.

194

62. Evans D, Spencer C, Pointon J, Su Z, Harvey D, Kochan G, et al. Interaction between ERAP1 and HLA-B27 in ankylosing spondylitis implicates peptide handling in the mechanism for HLA-B27 in disease susceptibility. Nat Genet. 2011;43:6. 63. Lockhart E, Green AM, Flynn JL. IL-17 Production Is Dominated by γδ T Cells rather than CD4 T Cells during Mycobacterium tuberculosis Infection. The Journal of Immunology. 2006;177(7):4662-9. 64. Shibata K, Yamada H, Hara H, Kishihara K, Yoshikai Y. Resident Vdelta1+ gammadelta T cells control early infiltration of neutrophils after Escherichia coli infection via IL-17 production. J Immunol. 2007;178:4466 - 72. 65. Roark C, French J, Taylor M, Bendele A, Born W, O'Brien R. Exacerbation of collagen-induced arthritis by oligoclonal, IL-17-producing gamma delta T cells. J Immunol. 2007;179:5576 - 83. 66. Nanno M, Kanari Y, Naito T, Inoue N, Hisamatsu T, Chinen H, et al. Exacerbating role of gammadelta T cells in chronic colitis of T-cell receptor alpha mutant mice. Gastroenterology. 2008;134:481 - 90. 67. Kenna T, Davidson S, Duan R, Bradbury L, McFarlane J, Smith M, et al. Enrichment of circulating interleukin-17-secreting interleukin-23 receptor-positive gamma/delta T cells in patients with active ankylosing spondylitis. Arthritis Rheum. 2012;64:1420 - 9. 68. Boismenu R, Havran W. Modulation of epithelial cell growth by intraepithelial gamma delta T cells. Science. 1994;266:1253 - 5. 69. Bendelac A, Savage P, Teyton L. The biology of NKT cells. Ann Rev Immunol. 2007;25:297 - 336. 70. Rachitskaya A, Hansen A, Horai R, Li Z, Villasmil R, Luger D, et al. Cutting edge: NKT cells constitutively express IL-23 receptor and RORgammat and rapidly produce IL-17 upon receptor ligation in an IL-6- independent fashion. J Immunol. 2008;180:5167 - 71. 71. Akbari O, Stock P, Meyer E, Kronenberg M, Sidobre S, Nakayama T, et al. Essential role of NKT cells producing IL-4 and IL-13 in the development of allergen-induced airway hyperreactivity. Nat Med. 2003;9:582 - 8. 72. Baxter A, Kinder S, Hammond K, Scollay R, Godfrey D. Association between alphabetaTCR+CD4-CD8- T-cell deficiency and IDDM in NOD/Lt mice. Diabetes. 1997;46:572 - 82. 73. Heller F, Fuss IJ, Nieuwenhuis EE, Blumberg RS, Strober W. Oxazolone Colitis, a Th2 Colitis Model Resembling Ulcerative Colitis, Is Mediated by IL-13-Producing NK-T Cells. Immunity. 2002;17(5):629-38. 74. Ueno Y, Tanaka S, Sumii M, Miyake S, Tazuma S, Taniguchi M, et al. Single dose of OCH improves mucosal T helper type 1/T helper type 2 cytokine balance and prevents experimental colitis in the presence of Vα14 natural killer T cells in mice. Inflammatory Bowel Diseases. 2005;11(1):35-41. 75. Wingender G, Stepniak D, Krebs P, Lin L, McBride S, Wei B, et al. Intestinal microbes affect phenotypes and functions of invariant natural killer T cells in mice. Gastroenterology. 2012;143:418 - 28. 76. Coppieters K, Van Beneden K, Jacques P, Dewint P, Vervloet A, Vander Cruyssen B, et al. A single early activation of invariant NK T cells confers long-term protection against collagen-induced arthritis in a ligand-specific manner. J Immunol. 2007;179:2300 - 9. 77. Jacques P, Venken K, Van Beneden K, Hammad H, Seeuws S, Drennan M, et al. Invariant natural killer T cells are natural regulators of murine spondylarthritis. Arthritis Rheum. 2010;62:988 - 99. 78. Singh A, Misra R, Aggarwal A. Th-17 associated cytokines in patients with reactive arthritis/undifferentiated spondyloarthropathy. Clinical Rheumatology. 2011;30(6):771-6. 79. Gold M, Cerri S, Smyk-Pearson S, Cansler M, Vogt T, Delepine J, et al. Human mucosal associated invariant T cells detect bacterially infected cells. PLoS Biol. 2010;8:e1000407. 80. Le Bourhis L, Martin E, Peguillet I, Guihot A, Froux N, Core M, et al. Antimicrobial activity of mucosal associated invariant T cells. Nat Immunol. 2010;11:701 - 8. 81. Dusseaux M, Martin E, Serriari N, Peguillet I, Premel V, Louis D, et al. Human MAIT cells are xenobiotic resistant, tissue-targeted, CD161hi IL-17-secreting T cells. Blood. 2011;117:1250 - 9.

195

82. Treiner E, Duban L, Bahram S, Radosavljevic M, Wanner V, Tilloy F, et al. Selection of evolutionarily conserved mucosalassociated invariant T cells by MR1. Nature. 2003;422:164 - 9. 83. Savage A, Constantinides M, Han J, Picard D, Martin E, Li B, et al. The transcription factor PLZF directs the effector program of the NKT cell lineage. Immunity. 2008;29:391 - 403. 84. Billerbeck E, Kang Y, Walker L, Lockstone H, Grafmueller S, Fleming V, et al. Analysis of CD161 expression on human CD8+ T cells defines a distinct functional subset with tissue-homing properties. Proc Natl Acad Sci USA. 2010;107:3006 - 11. 85. Le Bourhis L, Guerri L, Dusseaux M, Martin E, Soudais C, Lantz O. Mucosal associated invariant T cells: unconventional development and function. Trends Immunol. 2011;32:212 - 8. 86. Takatori H, Kanno Y, Watford W, Tato C, Weiss G, Littman D, et al. Lymphoid tissue inducer-like cells are an innate source of IL-17 and IL-22. J Exp Med. 2009;206:35 - 41. 87. Qiu J, Heller Jennifer J, Guo X, Chen Z-ming E, Fish K, Fu Y-X, et al. The Aryl Hydrocarbon Receptor Regulates Gut Immunity through Modulation of Innate Lymphoid Cells. Immunity. 2012;36(1):92-104. 88. Sawa S, Lochner M, Satoh-Takayama N, Dulauroy S, Berard M, Kleinschek M, et al. ROR[gamma]t+ innate lymphoid cells regulate intestinal homeostasis by integrating negative signals from the symbiotic microbiota. Nat Immunol. 2011;12(4):320-6. 89. Buonocore S, Ahern PP, Uhlig HH, Ivanov II, Littman DR, Maloy KJ, et al. Innate lymphoid cells drive interleukin-23-dependent innate intestinal pathology. Nature. 2010;464(7293):1371-5. 90. Cella M, Fuchs A, Vermi W, Facchetti F, Otero K, Lennerz J, et al. A human natural killer cell subset provides an innate source of IL-22 for mucosal immunity. Nature. 2009;457:722 - 5. 91. Cua DJ, Tato CM. Innate IL-17-producing cells: the sentinels of the immune system. Nat Rev Immunol. 2010;10(7):479-89. 92. Takatori H, Kanno Y, Watford WT, Tato CM, Weiss G, Ivanov II, et al. Lymphoid tissue inducer–like cells are an innate source of IL-17 and IL-22. The Journal of Experimental Medicine. 2009;206(1):35-41. 93. Kobori A, Yagi Y, Imaeda H, Ban H, Bamba S, Tsujikawa T, et al. Interleukin-33 expression is specifically enhanced in inflamed mucosa of ulcerative colitis. Journal of Gastroenterology. 2010;45(10):999-1007. 94. Sonnenberg GF, Fouser LA, Artis D. Border patrol: regulation of immunity, inflammation and tissue homeostasis at barrier surfaces by IL-22. Nat Immunol. 2011;12(5):383-90. 95. Wolk K, Kunz S, Witte E, Friedrich M, Asadullah K, Sabat R. IL-22 Increases the Innate Immunity of Tissues. Immunity. 2004;21(2):241-54. 96. Wolk K, Witte E, Wallace E, Döcke W-D, Kunz S, Asadullah K, et al. IL-22 regulates the expression of genes responsible for antimicrobial defense, cellular differentiation, and mobility in keratinocytes: a potential role in psoriasis. European Journal of Immunology. 2006;36(5):1309-23. 97. Dumoutier L, Louahed J, Renauld J-C. Cloning and Characterization of IL-10-Related T Cell-Derived Inducible Factor (IL-TIF), a Novel Cytokine Structurally Related to IL-10 and Inducible by IL-9. The Journal of Immunology. 2000;164(4):1814-9. 98. Zheng Y, Valdez PA, Danilenko DM, Hu Y, Sa SM, Gong Q, et al. Interleukin-22 mediates early host defense against attaching and effacing bacterial pathogens. Nat Med. 2008;14(3):282-9. 99. Aujla SJ, Chan YR, Zheng M, Fei M, Askew DJ, Pociask DA, et al. IL-22 mediates mucosal host defense against Gram-negative bacterial pneumonia. Nat Med. 2008;14(3):275-81. 100. Sonnenberg GF, Monticelli LA, Alenghat T, Fung TC, Hutnick NA, Kunisawa J, et al. Innate Lymphoid Cells Promote Anatomical Containment of Lymphoid-Resident Commensal Bacteria. Science. 2012;336(6086):1321-5. 101. Wu H-J, Ivanov II, Darce J, Hattori K, Shima T, Umesaki Y, et al. Gut-Residing Segmented Filamentous Bacteria Drive Autoimmune Arthritis via T Helper 17 Cells. Immunity. 2010;32(6):815-27. 102. Frank DN, St. Amand AL, Feldman RA, Boedeker EC, Harpaz N, Pace NR. Molecular-phylogenetic characterization of microbial community imbalances in human inflammatory bowel diseases. Proceedings of the National Academy of Sciences. 2007;104(34):13780-5. 103. Papa E, Docktor M, Smillie C, Weber S, Preheim SP, Gevers D, et al. Non-Invasive Mapping of the Gastrointestinal Microbiota Identifies Children with Inflammatory Bowel Disease. PLoS ONE. 2012;7(6):e39242.

196

104. Elinav E, Strowig T, Kau Andrew L, Henao-Mejia J, Thaiss Christoph A, Booth Carmen J, et al. NLRP6 Inflammasome Regulates Colonic Microbial Ecology and Risk for Colitis. Cell. 2011;145(5):745-57. 105. Atarashi K, Tanoue T, Shima T, Imaoka A, Kuwahara T, Momose Y, et al. Induction of colonic regulatory T cells by indigenous Clostridium species. Science. 2011;331:337 - 41. 106. Lederberg J, McCray AT. 'Ome Sweet 'Omics - A Genealogical Treasury of Words. The Scientist. 2001;15(8). 107. Tringe SG, Hugenholtz P. A renaissance for the pioneering 16S rRNA gene. Current Opinion in Microbiology. 2008;11(5):442-6. 108. Woese CR. Bacterial evolution. Microbiol Rev. 1987;51(2):221-71. 109. Clarridge JE, III. Impact of 16S rRNA Gene Sequence Analysis for Identification of Bacteria on Clinical Microbiology and Infectious Diseases. Clin Microbiol Rev. 2004;17(4):840-62. 110. Chen K, Neimark H, Rumore P, Steinman CR. Broad range DNA probes for detecting and amplifying eubacterial nucleic acids. FEMS Microbiology Letters. 1989;57(1):19-24. 111. Stackebrandt E, Goebel BM. Taxonomic Note: A Place for DNA-DNA Reassociation and 16S rRNA Sequence Analysis in the Present Species Definition in Bacteriology. International Journal of Systematic Bacteriology. 1994;44(4):846-9. 112. Kunin V, Engelbrektson A, Ochman H, Hugenholtz P. Wrinkles in the rare biosphere: pyrosequencing errors can lead to artificial inflation of diversity estimates. Environmental Microbiology. 2010;12(1):118-23. 113. McDonald D, Price MN, Goodrich J, Nawrocki EP, DeSantis TZ, Probst A, et al. An improved Greengenes taxonomy with explicit ranks for ecological and evolutionary analyses of bacteria and archaea. The ISME Journal. 2012;6(3):610-8. 114. Turnbaugh PJ, Hamady M, Yatsunenko T, Cantarel BL, Duncan A, Ley RE, et al. A core gut microbiome in obese and lean twins. Nature. 2009;457(7228):480-4. 115. Langille MGI, Zaneveld J, Caporaso JG, McDonald D, Knights D, Reyes JA, et al. Predictive functional profiling of microbial communities using 16S rRNA marker gene sequences. Nat Biotech. 2013;31(9):814-21. 116. Huse SM, Dethlefsen L, Huber JA, Welch DM, Relman DA, Sogin ML. Exploring Microbial Diversity and Taxonomy Using SSU rRNA Hypervariable Tag Sequencing. PLoS Genet. 2008;4(11):e1000255. 117. Turnbaugh PJ. An obesity-associated gut microbiome with increased capacity for energy harvest. Nature. 2006;444:1027-31. 118. Yatsunenko T, Rey FE, Manary MJ, Trehan I, Dominguez-Bello MG, Contreras M, et al. Human gut microbiome viewed across age and geography. Nature. 2012;486(7402):222-7. 119. Turnbaugh PJ, Backhed F, Fulton L, Gordon JI. Diet-induced obesity is linked to marked but reversible alterations in the mouse distal gut microbiome. Cell Host Microbe. 2008;3:213-23. 120. Greenblum S, Turnbaugh PJ, Borenstein E. Metagenomic systems biology of the human gut microbiome reveals topological shifts associated with obesity and inflammatory bowel disease. Proceedings of the National Academy of Sciences. 2012;109(2):594-9. 121. Ridaura VK, Faith JJ, Rey FE, Cheng J, Duncan AE, Kau AL, et al. Gut Microbiota from Twins Discordant for Obesity Modulate Metabolism in Mice. Science. 2013;341(6150). 122. Maslowski KM, Vieira AT, Ng A, Kranich J, Sierro F, Di Y, et al. Regulation of inflammatory responses by gut microbiota and chemoattractant receptor GPR43. Nature. 2009;461(7268):1282-6. 123. Arumugam M, Raes J, Pelletier E, Le Paslier D, Yamada T, Mende D, et al. Enterotypes of the human gut microbiome. Nature. 2011;473:174 - 80. 124. Group TNHW, Peterson J, Garges S, Giovanni M, McInnes P, Wang L, et al. The NIH Human Microbiome Project. Genome Research. 2009;19(12):2317-23. 125. Qin J, Li R, Raes J, Arumugam M, Burgdorf KS, Manichanh C, et al. A human gut microbial gene catalogue established by metagenomic sequencing. Nature. 2010;464(7285):59-65. 126. Consortium HMP. Structure, function and diversity of the healthy human microbiome. Nature. 2012;486(7402):207-14. 127. Caporaso JG, Lauber C, Costello E, Berg-Lyons D, Gonzalez A, Stombaugh J, et al. Moving pictures of the human microbiome. Genome Biology. 2011;12(5):R50. 128. Costello EK, Lauber CL, Hamady M, Fierer N, Gordon JI, Knight R. Bacterial Community Variation in Human Body Habitats Across Space and Time. Science. 2009;326(5960):1694-7.

197

129. Koren O, Knights D, Gonzalez A, Waldron L, Segata N, Knight R, et al. A Guide to Enterotypes across the Human Body: Meta-Analysis of Microbial Community Structures in Human Microbiome Datasets. PLoS Comput Biol. 2013;9(1):e1002863. 130. Knights D, Ward TL, McKinlay CE, Miller H, Gonzalez A, McDonald D, et al. Rethinking “Enterotypes”. Cell Host & Microbe. 2014;16(4):433-7. 131. Perdomo OJ, Cavaillon JM, Huerre M, Ohayon H, Gounon P, Sansonetti PJ. Acute inflammation causes epithelial invasion and mucosal destruction in experimental shigellosis. The Journal of Experimental Medicine. 1994;180(4):1307-19. 132. Vaishnava S, Behrendt CL, Ismail AS, Eckmann L, Hooper LV. Paneth cells directly sense gut commensals and maintain homeostasis at the intestinal host-microbial interface. Proceedings of the National Academy of Sciences. 2008;105(52):20858-63. 133. Laurence A, O'Shea JJ, Watford WT. Interleukin-22: a sheep in wolf's clothing. Nat Med. 2008;14(3):247-9. 134. Salzman NH, Hung K, Haribhai D, Chu H, Karlsson-Sjoberg J, Amir E, et al. Enteric defensins are essential regulators of intestinal microbial ecology. Nat Immunol. 2010;11(1):76-82. 135. Spor A, Koren O, Ley R. Unravelling the effects of the environment and host genotype on the gut microbiome. Nat Rev Micro. 2011;9(4):279-90. 136. Benson AK, Kelly SA, Legge R, Ma F, Low SJ, Kim J, et al. Individuality in gut microbiota composition is a complex polygenic trait shaped by multiple environmental and host genetic factors. Proceedings of the National Academy of Sciences. 2010;107(44):18933-8. 137. Goodrich Julia K, Waters Jillian L, Poole Angela C, Sutter Jessica L, Koren O, Blekhman R, et al. Human Genetics Shape the Gut Microbiome. Cell. 2014;159(4):789-99. 138. Ebringer A. The Cross-Tolerance Hypothesis, HLA-B27 and Ankylosing Spondylitis. Rheumatology. 1983;XXII(suppl 2):53-66. 139. Rashid T, Ebringer A. Ankylosing spondylitis is linked to Klebsiella—the evidence. Clinical Rheumatology. 2007;26(6):858-64. 140. Rath HC, Wilson KH, Sartor RB. Differential Induction of Colitis and Gastritis in HLA-B27 Transgenic Rats Selectively Colonized with Bacteroides vulgatus or Escherichia coli. Infection and Immunity. 1999;67(6):2969-74. 141. Rath H, Herfarth H, Ikeda J, Grenther W, Hamm T, Balish E, et al. Normal luminal bacteria, especially Bacteroides species, mediate chronic colitis, gastritis, and arthritis in HLA-B27/human beta2 microglobulin transgenic rats. J Clin Invest. 1996;98:945 - 53. 142. Ebringer RW, Cawdell DR, Cowling P, Ebringer A. Sequential studies in ankylosing spondylitis. Association of Klebsiella pneumoniae with active disease. Annals of the Rheumatic Diseases. 1978;37(2):146-51. 143. Mäki-ikola O, Nissilä M, Lehtinen K, Leirisalo-repo M, Toivanen P, Granfors K. Antibodies to Klebsiella pneumoniae, Eschericha coli and Proteus mirabilisin the sera of patients with axial and peripheral form of ankylosing spondylitis. Rheumatology. 1995;34(5):413-7. 144. Ringrose JH. HLA-B27 associated spondyloarthropathy, an autoimmune disease based on crossreactivity between bacteria and HLA-B27? Annals of the Rheumatic Diseases. 1999;58(10):598-610. 145. Quinton JF. Anti-Saccharomyces cerevisiae mannan antibodies combined with antineutrophil cytoplasmic autoantibodies in inflammatory bowel disease: prevalence and diagnostic role. Gut. 1998;42:788-91. 146. Targan SR. Antibodies to CBir1 flagellin define a unique response that is associated independently with complicated Crohn's disease. Gastroenterology. 2005;128:2020-8. 147. Mundwiler M, Mei L, Landers C, Reveille J, Targan S, Weisman M. Inflammatory bowel disease serologies in ankylosing spondylitis patients: a pilot study. Arthritis Research & Therapy. 2009;11(6):R177. 148. Wallis D. Elevated serum anti-flagellin antibodies implicate subclinical bowel inflammation in ankylosing spondylitis: an observational study. Arthritis Res Ther. 2013;15:R166. 149. van Praet L. Microscopic gut inflammation in axial spondyloarthritis: a multiparametric predictive model. Ann Rheum Dis. 2013;72:414-7. 150. Eerola E, Mottonen T, Hannonen P, Luukkainen R, Kantola I, Vuori K, et al. Intestinal flora in early rheumatoid arthritis. Rheumatology. 1994;33:1030 - 8. 151. Vaahtovuo J, Munukka E, Korkeamaki M, Luukainen R, Toivanen P. Fecal Microbiota in Early Rheumatoid Arthritis. The Journal of Rheumatology. 2008;35(8):1500-5.

198

152. Eckburg P, Bik E, Bernstein C, Purdom E, Dethlefsen L, Sargent M, et al. Diversity of the human intestinal microbial flora. Science. 2005;308:1635 - 8. 153. Gevers D, Kugathasan S, Denson LA, V·zquez-Baeza Y, Van†Treuren W, Ren B, et al. The Treatment-Naive Microbiome in New-Onset Crohn s Disease. Cell host & microbe. 2014;15(3):382-92. 154. Martinez-Martinez RE, Abud-Mendoza C, Patiño-Marin N, Rizo-Rodríguez JC, Little JW, Loyola-Rodríguez JP. Detection of periodontal bacterial DNA in serum and synovial fluid in refractory rheumatoid arthritis patients. Journal of Clinical Periodontology. 2009;36(12):1004-10. 155. Wegner N, Wait R, Sroka A, Eick S, Nguyen K-A, Lundberg K, et al. Peptidylarginine deiminase from Porphyromonas gingivalis citrullinates human fibrinogen and α-enolase: Implications for autoimmunity in rheumatoid arthritis. Arthritis & Rheumatism. 2010;62(9):2662-72. 156. Gaiffer MH, Holdsworth CD, Duerden BI. The assessment of faecal flora in patients with inflammatory bowel disease by a simplified bacteriological technique. Journal of Medical Microbiology. 1991;35(4):238-43. 157. Seksik P, Rigottier-Gois L, Gramet G, Sutren M, Pochart P, Marteau P, et al. Alterations of the dominant faecal bacterial groups in patients with Crohn's disease of the colon. Gut. 2003;52(2):237-42. 158. Ott SJ, Musfeldt M, Ullmann U, Hampe J, Schreiber S. Quantification of Intestinal Bacterial Populations by Real-Time PCR with a Universal Primer Set and Minor Groove Binder Probes: a Global Approach to the Enteric Flora. J Clin Microbiol. 2004;42(6):2566-72. 159. Walker A, Sanderson J, Churcher C, Parkes G, Hudspith B, Rayment N, et al. High-throughput clone library analysis of the mucosa-associated microbiota reveals dysbiosis and differences between inflamed and non-inflamed regions of the intestine in inflammatory bowel disease. BMC Microbiol. 2011;11:7. 160. Krueger JG. The immunologic basis for the treatment of psoriasis with new biologic agents. Journal of the American Academy of Dermatology. 2002;46(1):1-26. 161. Fahlen A, Engstrand L, Baker B, Powles A, Fry L. Comparison of bacterial microbiota in skin biopsies from normal and psoriatic skin. Arch Dermatol Res. 2012;304:15 - 22. 162. Unni K, Holley K, McDuffie F, Titus J. Comparative study of NZB mice under germf ree and conventional conditions. J Rheumatol. 1975;2:8. 163. Olszak T, An D, Zeissig S, Vera M, Richter J, Franke A, et al. Microbial exposure during early life has persistent effects on natural killer T cell function. Science. 2012;336:489 - 93. 164. Shaw M, Kamada N, Kim Y, Nunez G. Microbiota-induced IL-1beta, but not IL-6, is critical for the development of steady-state TH17 cells in the intestine. J Exp Med. 2012;209:251 - 8. 165. Hildebrand F, Nguyen A, Brinkman B, Yunta R, Cauwe B, Vandenabeele P, et al. Inflammation-associated enterotypes, host genotype, cage and inter-individual effects drive gut microbiota variation in common laboratory mice. Genome Biology. 2013;14(1):R4. 166. Robinson P, Brown M. The genetics of ankylosing spondylitis and axial spondyloarthritis. Rheum Dis Clin North Am. 2012;38:539 - 53. 167. Wen L, Ley RE, Volchkov PY, Stranges PB, Avanesyan L, Stonebraker AC, et al. Innate immunity and intestinal microbiota in the development of Type 1 diabetes. Nature. 2008;455(7216):1109-13. 168. Bhargava P, Mowry E. Gut Microbiome and Multiple Sclerosis. Curr Neurol Neurosci Rep. 2014;14(10):1-8. 169. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, et al. QIIME allows analysis of high-throughput community sequencing data. Nat Meth. 2010;7(5):335-6. 170. Willner DL, Hugenholtz P, Yerkovich ST, Tan ME, Daly JN, Lachner N, et al. Re-Establishment of Recipient-Associated Microbiota in the Lung Allograft is Linked to Reduced Risk of Bronchiolitis Obliterans Syndrome. American Journal of Respiratory and Critical Care Medicine. 2013. 171. Angly FE, Willner D, Rohwer F, Hugenholtz P, Tyson GW. Grinder: a versatile amplicon and shotgun sequence simulator. Nucleic Acids Research. 2012;40(12):e94. 172. Caporaso JG, Lauber CL, Walters WA, Berg-Lyons D, Lozupone CA, Turnbaugh PJ, et al. Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample. Proceedings of the National Academy of Sciences. 173. Van Linden SMD, Valkenburg HA, Jongh BMD, Cats A. The risk of developing ankylosing spondylitis in HLA-B27 positive individuals. Arthritis & Rheumatism. 1984;27(3):241-9.

199

174. Lozupone C, Hamady M, Knight R. UniFrac[mdash]an online tool for comparing microbial community diversity in a phylogenetic context. BMC Bioinformatics. 2006;7:371. 175. Rodriguez-Reyna TS, Martinez-Reyes C, Yamamoto-Furusho JK. Rheumatic manifestations of inflammatory bowel disease. World J Gastroenterol. 2009;15(44):5517-24. 176. Dixon P. VEGAN, a package of R functions for community ecology. Journal of Vegetation Science. 2003;14(6):927-30. 177. Team RC. R: A language and environment for statistical computing. 2012. 178. Stearns JC, Lynch MDJ, Senadheera DB, Tenenbaum HC, Goldberg MB, Cvitkovitch DG, et al. Bacterial biogeography of the human digestive tract. Sci Rep. 2011;1. 179. Huse SM, Welch DM, Morrison HG, Sogin ML. Ironing out the wrinkles in the rare biosphere through improved OTU clustering. Environmental Microbiology. 2010;12(7):1889-98. 180. Bokulich NA, Subramanian S, Faith JJ, Gevers D, Gordon JI, Knight R, et al. Quality-filtering vastly improves diversity estimates from Illumina amplicon sequencing. Nat Meth. 2013;10(1):57-9. 181. Schloss PD, Gevers D, Westcott SL. Reducing the Effects of PCR Amplification and Sequencing Artifacts on 16S rRNA-Based Studies. PLoS ONE. 2011;6(12):e27310. 182. Dethlefsen L, Eckburg PB, Bik EM, Relman DA. Assembly of the human intestinal microbiota. Trends in Ecology & Evolution. 2006;21(9):517-23. 183. Gill SR, Pop M, DeBoy RT, Eckburg PB, Turnbaugh PJ, Samuel BS, et al. Metagenomic Analysis of the Human Distal Gut Microbiome. Science. 2006;312(5778):1355-9. 184. Rankin LC, Groom JR, Chopin M, Herold MJ, Walker JA, Mielke LA, et al. The transcription factor T-bet is essential for the development of NKp46+ innate lymphocytes via the Notch pathway. Nat Immunol. 2013;14(4):389-95. 185. Klose CSN, Kiss EA, Schwierzeck V, Ebert K, Hoyler T, d/'Hargues Y, et al. A T-bet gradient controls the fate and function of CCR6-ROR[ggr]t+ innate lymphoid cells. Nature. 2013;494(7436):261-5. 186. Belkaid Y, Naik S. Compartmentalized and systemic control of tissue immunity by commensals. Nat Immunol. 2013;14(7):646-53. 187. Robinson PC, Brown MA. Genetics of ankylosing spondylitis. Molecular Immunology. 2014;57(1):2-11. 188. Costello M-E, Elewaut D, Kenna T, Brown M. Microbes, the gut and ankylosing spondylitis. Arthritis Research & Therapy. 2013;15(3):214. 189. Lozupone C, Knight R. UniFrac: a new phylogenetic method for comparing microbial communities. Appl Environ Microbiol. 2005;71:8228-35. 190. Ridlon JM, Kang D-J, Hylemon PB. Bile salt biotransformations by human intestinal bacteria. Journal of Lipid Research. 2006;47(2):241-59. 191. Caporaso JG, Lauber CL, Walters WA, Berg-Lyons D, Lozupone CA, Turnbaugh PJ, et al. Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample. Proceedings of the National Academy of Sciences. 2011;108(Supplement 1):4516-22. 192. Vesper B, Jawdi A, Altman K, Haines I, Tao L, Radosevich J. The effect of proton pump inhibitors on the human microbiota. Current drug metabolism. 2009;10(1):84-9. 193. Franzosa EA, Huang K, Meadow JF, Gevers D, Lemon KP, Bohannan BJM, et al. Identifying personal microbiomes using metagenomic codes. Proceedings of the National Academy of Sciences. 2015. 194. Koren O, Spor A, Felin J, Fåk F, Stombaugh J, Tremaroli V, et al. Human oral, gut, and plaque microbiota in patients with atherosclerosis. Proceedings of the National Academy of Sciences. 2011;108(Supplement 1):4592-8. 195. Hang J, Desai V, Zavaljevski N, Yang Y, Lin X, Satya RV, et al. 16S rRNA gene pyrosequencing of reference and clinical samples and investigation of the temperature stability of microbiome profiles. Microbiome. 2014;2:31-. 196. Le Chatelier E, Nielsen T, Qin J, Prifti E, Hildebrand F, Falony G, et al. Richness of human gut microbiome correlates with metabolic markers. Nature. 2013;500(7464):541-6. 197. Turnbaugh PJ, Bäckhed F, Fulton L, Gordon JI. Diet-Induced Obesity Is Linked to Marked but Reversible Alterations in the Mouse Distal Gut Microbiome. Cell Host & Microbe. 2008;3(4):213-23.

200

198. Hildebrandt MA, Hoffman C, Sherrill-Mix SA, Keilbaugh SA, Hamady M, Chen Y-Y, et al. High Fat Diet Determines the Composition of the Murine Gut Microbiome Independently of Obesity. Gastroenterology. 2009;137(5):1716-24.e1-2. 199. Devkota S, Wang Y, Musch MW, Leone V, Fehlner-Peach H, Nadimpalli A, et al. Dietary-fat-induced taurocholic acid promotes pathobiont expansion and colitis in Il102/2 mice. Nature. 2012;487(7405):104-8. 200. Arthur JC, Perez-Chanona E, Mühlbauer M, Tomkovich S, Uronis JM, Fan T-J, et al. Intestinal inflammation targets cancer-inducing activity of the microbiota. Science (New York, NY). 2012;338(6103):120-3. 201. Lin P, Bach M, Asquith M, Lee AY, Akileswaran L, Stauffer P, et al. HLA-B27 and Human β2-Microglobulin Affect the Gut Microbiota of Transgenic Rats. PLoS ONE. 2014;9(8):e105684. 202. Rosenbaum JT, Davey MP. Time for a gut check: Evidence for the hypothesis that HLA–B27 predisposes to ankylosing spondylitis by altering the microbiome. Arthritis & Rheumatism. 2011;63(11):3195-8. 203. García-Medel N, Sanz-Bravo A, Van Nguyen D, Galocha B, Gómez-Molina P, Martín-Esteban A, et al. Functional Interaction of the Ankylosing Spondylitis-associated Endoplasmic Reticulum Aminopeptidase 1 Polymorphism and HLA-B27 in Vivo. Molecular & Cellular Proteomics. 2012;11(11):1416-29. 204. Haroon N, Tsui FWL, Chiu B, Tsui HW, Inman RD. Serum Cytokine Receptors in Ankylosing Spondylitis: Relationship to Inflammatory Markers and Endoplasmic Reticulum Aminopeptidase Polymorphisms. The Journal of Rheumatology. 2010;37(9):1907-10. 205. Saveanu L, Carroll O, Lindo V, Del Val M, Lopez D, Lepelletier Y, et al. Concerted peptide trimming by human ERAP1 and ERAP2 aminopeptidase complexes in the endoplasmic reticulum. Nat Immunol. 2005;6(7):689-97. 206. Aldhamen YA, Seregin SS, Rastall DPW, Aylsworth CF, Pepelyayeva Y, Busuito CJ, et al. Endoplasmic Reticulum Aminopeptidase-1 Functions Regulate Key Aspects of the Innate Immune Response. PLoS ONE. 2013;8(7):e69539. 207. Hammer GE, Gonzalez F, Champsaur M, Cado D, Shastri N. The aminopeptidase ERAAP shapes the peptide repertoire displayed by major histocompatibility complex class I molecules. Nat Immunol. 2006;7(1):103-12. 208. Hammer GE, Gonzalez F, James E, Nolla H, Shastri N. In the absence of aminopeptidase ERAAP, MHC class I molecules present many unstable and highly immunogenic peptides. Nat Immunol. 2007;8(1):101-8. 209. Akram A, Lin A, Gracey E, Streutker CJ, Inman RD. HLA-B27, but Not HLA-B7, Immunodominance to Influenza Is ERAP Dependent. The Journal of Immunology. 2014;192(12):5520-8. 210. Mazmanian SK, Liu CH, Tzianabos AO, Kasper DL. An Immunomodulatory Molecule of Symbiotic Bacteria Directs Maturation of the Host Immune System. Cell. 2005;122(1):107-18. 211. Hooper LV, Littman DR, Macpherson AJ. Interactions Between the Microbiota and the Immune System. Science. 2012;336(6086):1268-73. 212. Ivanov I, Atarashi K, Manel N, Brodie E, Shima T, Karaoz U, et al. Induction of intestinal Th17 cells by segmented filamentous bacteria. Cell. 2009;139:485 - 98. 213. Asquith M, Powrie F. An innately dangerous balancing act: intestinal homeostasis, inflammation, and colitis-associated cancer. The Journal of Experimental Medicine. 2010;207(8):1573-7. 214. Blanchard N, Gonzalez F, Schaeffer M, Joncker NT, Cheng T, Shastri AJ, et al. Immunodominant, protective response to the parasite Toxoplasma gondii requires antigen processing in the endoplasmic reticulum. Nat Immunol. 2008;9(8):937-44. 215. York IA, Brehm MA, Zendzian S, Towne CF, Rock KL. Endoplasmic reticulum aminopeptidase 1 (ERAP1) trims MHC class I-presented peptides in vivo and plays an important role in immunodominance. Proceedings of the National Academy of Sciences. 2006;103(24):9202-7. 216. Rock KL, Farfán-Arribas DJ, Shen L. Proteases in MHC Class I Presentation and Cross-Presentation. The Journal of Immunology. 2010;184(1):9-15. 217. Saveanu L, Carroll O, Weimershaus M, Guermonprez P, Firat E, Lindo V, et al. IRAP Identifies an Endosomal Compartment Required for MHC Class I Cross-Presentation. Science. 2009;325(5937):213-7.

201

218. Ciccia F, Accardo-Palumbo A, Alessandro R, Rizzo A, Principe S, Peralta S, et al. Interleukin-22 and IL-22-producing NKp44+NK cells in the subclinical gut inflammation of patients with ankylosing spondylitis. Arthritis & Rheumatism. 2011:n/a-n/a. 219. Mangan PR, Harrington LE, O'Quinn DB, Helms WS, Bullard DC, Elson CO, et al. Transforming growth factor-[beta] induces development of the TH17 lineage. Nature. 2006;441(7090):231-4. 220. Geboes L, Dumoutier L, Kelchtermans H, Schurgers E, Mitera T, Renauld J-C, et al. Proinflammatory role of the Th17 cytokine interleukin-22 in collagen-induced arthritis in C57BL/6 mice. Arthritis & Rheumatism. 2009;60(2):390-5. 221. Nakae S, Nambu A, Sudo K, Iwakura Y. Suppression of Immune Induction of Collagen-Induced Arthritis in IL-17-Deficient Mice. The Journal of Immunology. 2003;171(11):6173-7. 222. Kenna TJ, Thomas R, Steptoe RJ. Steady-state dendritic cells expressing cognate antigen terminate memory CD8+ T-cell responses2008 2008-02-15 00:00:00. 2091-100 p. 223. Olejnik S, Algina J. Generalized eta and omega squared statistics: measures of effect size for some common research designs. Psychological methods. 2003;8(4):434. 224. Scher JU, Abramson SB. The microbiome and rheumatoid arthritis. Nat Rev Rheumatol. 2011;7(10):569-78. 225. Ivanov II, Frutos RdL, Manel N, Yoshinaga K, Rifkin DB, Sartor RB, et al. Specific Microbiota Direct the Differentiation of IL-17-Producing T-Helper Cells in the Mucosa of the Small Intestine. Cell Host & Microbe. 2008;4(4):337-49. 226. Larsen N, Vogensen FK, van den Berg FWJ, Nielsen DS, Andreasen AS, Pedersen BK, et al. Gut Microbiota in Human Adults with Type 2 Diabetes Differs from Non-Diabetic Adults. PLoS ONE. 2010;5(2):e9085. 227. Cui B, Feng Q, Wang H, Wang M, Peng Z, Li P, et al. Fecal microbiota transplantation through mid-gut for refractory Crohn's disease: Safety, feasibility and efficacy trial results. Journal of Gastroenterology and Hepatology. 2014:n/a-n/a. 228. Brandt LJ, Aroniadis OC, Mellow M, Kanatzar A, Kelly C, Park T, et al. Long-Term Follow-Up of Colonoscopic Fecal Microbiota Transplant for Recurrent Clostridium difficile Infection. Am J Gastroenterol. 2012;107(7):1079-87. 229. Mielants H, De Vos M, Goemaere S, Schelstraete K, Cuvelier C, Goethals K, et al. Intestinal mucosal permeability in inflammatory rheumatic diseases. II. Role of disease. The Journal of Rheumatology. 1991;18(3):394-400. 230. Benham H, Rehaume LM, Hasnain SZ, Velasco J, Baillet AC, Ruutu M, et al. Interleukin-23 Mediates the Intestinal Response to Microbial β-1,3-Glucan and the Development of Spondyloarthritis Pathology in SKG Mice. Arthritis & Rheumatology. 2014;66(7):1755-67. 231. Řeháková Z, Čapková J, Štěpánková R, Šinkora J, Loužecká A, Ivanyi P, et al. Germ-free mice do not develop ankylosing enthesopathy, a spontaneous joint disease. Human Immunology. 2000;61(6):555-8. 232. Ruutu M, Thomas G, Steck R, Degli-Esposti MA, Zinkernagel MS, Alexander K, et al. β-glucan triggers spondylarthritis and Crohn's disease–like ileitis in SKG mice. Arthritis & Rheumatism. 2012;64(7):2211-22. 233. Qin J, Li R, Raes J, Arumugam M, Burgdorf K, Manichanh C, et al. A human gut microbial gene catalogue established by metagenomic sequencing. Nature. 2010;464:59 - 65. 234. Palm O, Moum B, Ongre A, Gran JT. Prevalence of ankylosing spondylitis and other spondyloarthropathies among patients with inflammatory bowel disease: a population study (the IBSEN study). The Journal of Rheumatology. 2002;29(3):511-5. 235. Orchard TR, Holt H, Bradbury L, Hammersma J, McNally E, Jewell DP, et al. The prevalence, clinical features and association of HLA-B27 in sacroiliitis associated with established Crohn’s disease. Alimentary Pharmacology & Therapeutics. 2009;29(2):193-7. 236. Li Y, Willer CJ, Ding J, Scheet P, Abecasis GR. MaCH: using sequence and genotype data to estimate haplotypes and unobserved genotypes. Genetic Epidemiology. 2010;34(8):816-34. 237. Howie B, Fuchsberger C, Stephens M, Marchini J, Abecasis GR. Fast and accurate genotype imputation in genome-wide association studies through pre-phasing. Nat Genet. 2012;44(8):955-9. 238. Ihaka R, Gentleman R. R: A Language for Data Analysis and Graphics. Journal of Computational and Graphical Statistics. 1996;5(3):299-314.

202

239. Abecasis GR, Cherny SS, Cookson WO, Cardon LR. Merlin[mdash]rapid analysis of dense genetic maps using sparse gene flow trees. Nat Genet. 2002;30(1):97-101. 240. Khachatryan ZA, Ktsoyan ZA, Manukyan GP, Kelly D, Ghazaryan KA, Aminov RI. Predominant Role of Host Genetics in Controlling the Composition of Gut Microbiota. PLoS ONE. 2008;3(8):e3064. 241. Rausch P, Rehman A, Künzel S, Häsler R, Ott SJ, Schreiber S, et al. Colonic mucosa-associated microbiota is influenced by an interaction of Crohn disease and FUT2 (Secretor) genotype. Proceedings of the National Academy of Sciences. 2011;108(47):19030-5. 242. Wacklin P, Mäkivuokko H, Alakulppi N, Nikkilä J, Tenkanen H, Räbinä J, et al. Secretor Genotype (<italic>FUT2</italic> gene) Is Strongly Associated with the Composition of <italic>Bifidobacteria</italic> in the Human Intestine. PLoS ONE. 2011;6(5):e20113. 243. McGovern DPB, Jones MR, Taylor KD, Marciante K, Yan X, Dubinsky M, et al. Fucosyltransferase 2 (FUT2) non-secretor status is associated with Crohn's disease. Human Molecular Genetics. 2010;19(17):3468-76. 244. Shan M, Gentile M, Yeiser JR, Walland AC, Bornstein VU, Chen K, et al. Mucus Enhances Gut Homeostasis and Oral Tolerance by Delivering Immunoregulatory Signals. Science. 2013;342(6157):447-53. 245. Evans DM, Brion MJA, Paternoster L, Kemp JP, McMahon G, Munafò M, et al. Mining the Human Phenome Using Allelic Scores That Index Biological Intermediates. PLoS Genet. 2013;9(10):e1003919. 246. Stebbings S, Munro K, Simon M, Tannock G, Highton J, Harmsen H, et al. Comparison of the faecal microflora of patients with ankylosing spondylitis and controls using molecular methods of analysis. Rheumatology. 2002;41:1395 - 401. 247. O'Mahony S, Anderson N, Nuki G, Ferguson A. Systemic and mucosal antibodies to Klebsiella in patients with ankylosing spondylitis and Crohn's disease. Annals of the Rheumatic Diseases. 1992;51(12):1296-300. 248. Frank DN, Robertson CE, Hamm CM, Kpadeh Z, Zhang T, Chen H, et al. Disease phenotype and genotype are associated with shifts in intestinal-associated microbiota in inflammatory bowel diseases. Inflammatory Bowel Diseases. 2011;17(1):179-84. 249. Elinav E, Strowig T, Kau A, Henao-Mejia J, Thaiss C, Booth C, et al. NLRP6 inflammasome regulates colonic microbial ecology and risk for colitis. Cell. 2011;145:745 - 57. 250. Li E, Hamm CM, Gulati AS, Sartor RB, Chen H, Wu X, et al. Inflammatory Bowel Diseases Phenotype, <italic>C. difficile</italic> and NOD2 Genotype Are Associated with Shifts in Human Ileum Associated Microbial Composition. PLoS ONE. 2012;7(6):e26284. 251. David L, Materna A, Friedman J, Campos-Baptista M, Blackburn M, Perrotta A, et al. Host lifestyle affects human microbiota on daily timescales. Genome Biology. 2014;15(7):R89. 252. Faith JJ, Guruge JL, Charbonneau M, Subramanian S, Seedorf H, Goodman AL, et al. The long-term stability of the human gut microbiota. Science (New York, NY). 2013;341(6141):1237439-. 253. Ley RE, Turnbaugh PJ, Klein S, Gordon JI. Microbial ecology: human gut microbes associated with obesity. Nature. 2006;444:1022-3. 254. Dethlefsen L, Relman DA. Incomplete recovery and individualized responses of the human distal gut microbiota to repeated antibiotic perturbation. Proceedings of the National Academy of Sciences. 2011;108(Supplement 1):4554-61. 255. Jakobsson HE, Jernberg C, Andersson AF, Sjölund-Karlsson M, Jansson JK, Engstrand L. Short-Term Antibiotic Treatment Has Differing Long-Term Impacts on the Human Throat and Gut Microbiome. PLoS ONE. 2010;5(3):e9836. 256. Maslowski KM, Mackay CR. Diet, gut microbiota and immune responses. Nat Immunol. 2011;12(1):5-9. 257. Buffie CG, Jarchum I, Equinda M, Lipuma L, Gobourne A, Viale A, et al. Profound Alterations of Intestinal Microbiota following a Single Dose of Clindamycin Results in Sustained Susceptibility to Clostridium difficile-Induced Colitis. Infection and Immunity. 2012;80(1):62-73. 258. Scher JU, Sczesnak A, Longman RS, Segata N, Ubeda C, Bielski C, et al. Expansion of intestinal Prevotella copri correlates with enhanced susceptibility to arthritis. Mathis D, editor2013 2013-11-05 15:22:27.

203

259. Norman JM, Handley SA, Virgin HW. Kingdom-Agnostic Metagenomics and the Importance of Complete Characterization of Enteric Microbial Communities. Gastroenterology. 2014;146(6):1459-69. 260. Virgin Herbert W. The Virome in Mammalian Physiology and Disease. Cell. 2014;157(1):142-50. 261. Kernbauer E, Ding Y, Cadwell K. An enteric virus can replace the beneficial function of commensal bacteria. Nature. 2014;advance online publication. 262. van Nood E, Speelman P, Nieuwdorp M, Keller J. Fecal microbiota transplantation: facts and controversies. Current Opinion in Gastroenterology. 2014;30(1):34-9 10.1097/MOG.0000000000000024. 263. Smith M, Kassam Z, Edelstein C, Burgess J, Alm E. OpenBiome remains open to serve the medical community. Nat Biotech. 2014;32(9):867-. 264. van Nood E, Vrieze A, Nieuwdorp M, Fuentes S, Zoetendal EG, de Vos WM, et al. Duodenal Infusion of Donor Feces for Recurrent Clostridium difficile. New England Journal of Medicine. 2013;368(5):407-15. 265. Kassam Z, Lee CH, Yuan Y, Hunt RH. Fecal Microbiota Transplantation for Clostridium difficile Infection: Systematic Review and Meta-Analysis. Am J Gastroenterol. 2013;108(4):500-8. 266. Youngster I, Sauk J, Pindar C, Wilson RG, Kaplan JL, Smith MB, et al. Fecal Microbiota Transplant for Relapsing Clostridium difficile Infection Using a Frozen Inoculum From Unrelated Donors: A Randomized, Open-Label, Controlled Pilot Study. Clinical Infectious Diseases. 2014;58(11):1515-22. 267. Petrof E, Gloor G, Vanner S, Weese S, Carter D, Daigneault M, et al. Stool substitute transplant therapy for the eradication of Clostridium difficile infection: ‘RePOOPulating’ the gut. Microbiome. 2013;1(1):1-12. 268. Lawley TD, Clare S, Walker AW, Stares MD, Connor TR, Raisen C, et al. Targeted Restoration of the Intestinal Microbiota with a Simple, Defined Bacteriotherapy Resolves Relapsing <italic>Clostridium difficile</italic> Disease in Mice. PLoS Pathog. 2012;8(10):e1002995. 269. Smits LP, Bouter KEC, de Vos WM, Borody TJ, Nieuwdorp M. Therapeutic Potential of Fecal Microbiota Transplantation. Gastroenterology. 2013;145(5):946-53. 270. Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann M, Hollister EB, et al. Introducing mothur: Open-Source, Platform-Independent, Community-Supported Software for Describing and Comparing Microbial Communities. Applied and Environmental Microbiology. 2009;75(23):7537-41. 271. Hildebrand F, Tadeo R, Voigt A, Bork P, Raes J. LotuS: an efficient and user-friendly OTU processing pipeline. Microbiome. 2014;2(1):30. 272. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MAR, Bender D, et al. PLINK: A Tool Set for Whole-Genome Association and Population-Based Linkage Analyses. The American Journal of Human Genetics. 2007;81(3):559-75. 273. Friedman J, Alm EJ. Inferring Correlation Networks from Genomic Survey Data. PLoS Comput Biol. 2012;8(9):e1002687. 274. Segata N, Waldron L, Ballarini A, Narasimhan V, Jousson O, Huttenhower C. Metagenomic microbial community profiling using unique clade-specific marker genes. Nat Meth. 2012;9(8):811-4. 275. Segata N, Izard J, Waldron L, Gevers D, Miropolsky L, Garrett W, et al. Metagenomic biomarker discovery and explanation. Genome Biology. 2011;12(6):R60. 276. Abubucker S, Segata N, Goll J, Schubert AM, Izard J, Cantarel BL, et al. Metabolic Reconstruction for Metagenomic Data and Its Application to the Human Microbiome. PLoS Comput Biol. 2012;8(6):e1002358. 277. Tickle T, Waldron L, Yiren L, Huttenhower C. Multivariate association of microbial communities with rich metadata in high-dimensional studies. (In progress). 2014. 278. Booijink CCGM, Boekhorst J, Zoetendal EG, Smidt H, Kleerebezem M, de Vos WM. Metatranscriptome Analysis of the Human Fecal Microbiota Reveals Subject-Specific Expression Profiles, with Genes Encoding Proteins Involved in Carbohydrate Metabolism Being Dominantly Expressed. Applied and Environmental Microbiology. 2010;76(16):5533-40. 279. Celaj A, Markle J, Danska J, Parkinson J. Comparison of assembly algorithms for improving rate of metatranscriptomic functional annotation. Microbiome. 2014;2:39-.

204

280. Zoetendal EG, Raes J, van den Bogert B, Arumugam M, Booijink CCGM, Troost FJ, et al. The human small intestinal microbiota is driven by rapid uptake and conversion of simple carbohydrates. The ISME Journal. 2012;6(7):1415-26. 281. Feehery GR, Yigit E, Oyola SO, Langhorst BW, Schmidt VT, Stewart FJ, et al. A Method for Selectively Enriching Microbial DNA from Contaminating Vertebrate Host DNA. PLoS ONE. 2013;8(10):e76096. 282. Aleer S, Adetutu EM, Weber J, Ball AS, Juhasz AL. Potential impact of soil microbial heterogeneity on the persistence of hydrocarbons in contaminated subsurface soils. Journal of Environmental Management. 2014;136(0):27-36. 283. Leimena M, Ramiro-Garcia J, Davids M, van den Bogert B, Smidt H, Smid E, et al. A comprehensive metatranscriptome analysis pipeline and its validation using human small intestine microbiota datasets. BMC Genomics. 2013;14(1):530.

205

Appendices

Ankylosing spondylitis

Ankylosing spondylitis (AS) belongs to a common group

of arthritides called spondyloarthopathies (SpA). AS

targets primarily the spine and pelvis and is characterized

histopathologically by entheseal infl am mation. Disease

progression in AS is characterized by excessive bone

formation (ankylosis) that gradually bridges the gap

between joints, eventually fusing joints and causing stiff -

ness, pain, signifi cant morbidity, and increased mortality

[1].

Th e coexistence of AS and intestinal infl ammation has

been known for some time [2]. Between 5% and 10% of

patients with AS develop clinically diagnosed infl amma-

tory bowel disease (IBD), and a further 70% of patients

with AS develop subclinical gut infl ammation [1,2]. In

reactive arthritis, a member of the SpA family, infl am-

matory arthritis develops following urogenital infection

with Chlamydia trachomatis or gastrointestinal infection

with Campylobacter, Salmonella, Shigella, or Yersinia [3].

Such cause-and-eff ect relationships are not established

for other SpAs.

Genetic overlap between ankylosing spondylitis

and gut disease

Strong genetic overlap exists between AS and IBD, and

the two conditions commonly occur together in families

[4]. Danoy and colleagues [5] (2010) studied genes known

to be associated with IBD in a large AS cohort. New loci

and genes were identifi ed, and of particular note were

genes involved in the interleukin-23 (IL-23) pathway,

such as STAT3, IL-23 receptor (IL23R), and IL12B (which

encodes IL-12p40, the share subunit of IL-23 and IL-12)

[5-7]. How these genetic lesions infl uence gut homeo-

stasis and function remains unclear. Major diff er ences

also exist between the genetics of IBD and AS, and AS

has shown no association to date with NOD2/CARD15 or

the autophagy gene ATG16L1, which are major suscep-

tibility factors in IBD, whereas IBD shows no association

with HLA-B or ERAP1, which are the strongest AS

susceptibility genes [8]. Although no asso ciation has been

shown with NOD2/CARD15 and AS specifi cally, poly-

morphisms in NOD2/CARD15 have been associated with

an increased risk of gut disease in patients with SpA [9].

The gut, barrier regulation, and intestinal

epithelial cells

Homeostasis of the normal microbial fl ora in the gut is

essential for intestinal health. Th e gastrointestinal tract is

heavily populated with microbes and is the primary site

for interaction between these microorganisms and the

immune system [10,11]. Furthermore, microbes found in

the gut help to shape host immune systems from an early

age. Th e incomplete development of the immune system

in neonates and under germ-free (GF) conditions tells us

that microbiota sculpt the host immune system [12,13].

Maintenance of intestinal and microbial homeostasis is

increasingly recognized as playing a pivotal role in overall

health [14], and dysregulation of either gut or microbial

homeostasis may play a role in autoimmunity. Physio-

logical processes in the host that maintain gut homeo stasis

and respond to perturbance in the gut micro environ ment

involve both the adaptive and innate immune system and

the barrier function of the intestines themselves.

Abstract

It is increasingly clear that the interaction between

host and microbiome profoundly aff ects health.

There are 10 times more bacteria in and on our

bodies than the total of our own cells, and the human

intestine contains approximately 100 trillion bacteria.

Interrogation of microbial communities by using classic

microbiology techniques off ers a very restricted view

of these communities, allowing us to see only what

we can grow in isolation. However, recent advances

in sequencing technologies have greatly facilitated

systematic and comprehensive studies of the role

of the microbiome in human health and disease.

Comprehensive understanding of our microbiome

will enhance understanding of disease pathogenesis,

which in turn may lead to rationally targeted therapy

for a number of conditions, including autoimmunity.

© 2010 BioMed Central Ltd

Microbes, the gut and ankylosing spondylitisMary-Ellen Costello1, Dirk Elewaut2, Tony J Kenna1 and Matthew A Brown*1

R E V I E W

*Correspondence: [email protected] University of Queensland Diamantina Institute, Translational Research

Institute, Level 7, 37 Kent Road, Princess Alexandra Hospital, Woolloongabba,

Brisbane QLD 4102, Australia

Full list of author information is available at the end of the article

Costello et al. Arthritis Research & Therapy 2013, 15:214 http://arthritis-research.com/content/15/3/214

© 2013 BioMed Central Ltd

The physical barrier

Th e human gastrointestinal tract is not a complete

barrier. It is composed of a single layer of intestinal

epithelial cells (IECs), which form a physical barrier

separating the intestinal lumen from the lamina propria

(Figure 1). IECs secrete soluble factors that are crucial to

intestinal homeostasis, such as mucins and anti-microbial

peptides, including lysozymes, defensins, cathelicidins,

lipocalins, and C-type lectins such as RegIIIγ [15-17].

Release of these molecules into luminal crypts is thought

to prevent microbial invasion into the crypt micro-

environment as well as limit bacteria-epithelial cell

contact [16,18]. Compared with healthy controls, Crohn’s

disease patients with active disease have pronounced

decreases in the human α-defensins DEFA5 and DEFA6

in the ileum, resulting in altered mucosal function and

overgrowth or dysregulation of commensal microbial

fl ora [18,19]. Conversely, overexpression of anti-micro-

bials, including α-defensins, are reported in sub-clinically

infl amed ileum of AS patients compared with Crohn’s

disease patients and healthy controls [20]. However, it

remains unclear whether changes in innate mucosal

defenses lead to alterations in gut-resident microbial fl ora

or whether early changes in the microbiome sculpt

intestinal host responses. Furthermore, depletion of the

mucin layer leads to an IBD-like phenotype and endo-

plasmic reticu lum stress, potentially driving IL-23 pro-

duc tion [21]. IL-23 excess alone is suffi cient to induce

spondyloarthritis in mice [22], and genetic evidence, in

particular, indicates that the cytokine plays a key role in

the development of spondyloarthritis in humans.

Permeability and disease

Th e dynamic crosstalk between IECs, microbes, and the

local immune cells is fundamental to intestinal homeo-

stasis but is also suggested to play a part in disease

pathogenesis [23]. In IBD and celiac disease, tight junc-

tions are dysregulated, causing increased permeability

between IECs, resulting in a ‘leaky gut’ [24,25]. Studies

looking at fi rst-degree relatives of patients with AS

showed that they too have increased gut permeability, so

it is likely that there is an underlying genetic process

operating in the gut [26,27].

The immune barrier

Th e complexity of the intestinal immune system is the

subject of many reviews, but here we will focus on some

intestinal immune populations that are believed to be

important in rheumatic disorders.

Innate immunity

Dendritic cells (DCs) densely populate the intestinal

lamina propria and form a widespread microbe-sensing

network. DCs recognize a broad repertoire of bacteria,

sensing with receptors such as Toll-like receptors (TLRs)

and monitoring the bacteria on the mucosal surface [28].

Intestinal DCs orchestrate production of intestine-

specifi c IgA to limit bacterial contact with the intestinal

epithelial cell surface [29]. Activated DCs can secrete a

number of cytokines and chemokines, including IL-23

and IL-6, that are involved in infl am mation and migration

of DCs [30].

Macrophages patrol the gastrointestinal systems in

high numbers. Th ey frequently come in contact with

‘stray’ bacteria, including commensals that have breached

the epithelial cell barrier. Macrophages phagocytose and

rapidly kill such bacteria by using mechanisms that

include the production of antimicrobial proteins and

reactive oxygen species [31]. Intestinal macrophages have

several unique characteristics, including the expression

of the anti-infl ammatory cytokine IL-10, both constitu-

tively and after bacterial stimulation [32,33]. Th is makes

intestinal macrophages non-infl ammatory cells that still

have the capacity to phagocytose. Th e importance of this

pathway is highlighted by the fact that loss-of-function

mutations in IL10R lead to early-onset IBD [34]. How-

ever, not all intestinal macrophages are non-infl amma-

tory. In patients with Crohn’s disease, a population of

intestinal macrophages secrete infl ammatory cytokines

such as IL-23, tumor necrosis factor-alpha (TNF-α), and

IL-6. Th ese macrophages contribute to an infl ammatory

microenvironment in patients with Crohn’s disease [35].

Intestinal macrophages also play a role in the restoration

of the physical integrity of the epithelial cell barrier

following injury or bacterial insult [24]. Re-establishing

this barrier after injury is imperative to avoid bacterial

penetration and sepsis in such a microbe-laden environ-

ment [36].

Adaptive immunity

IL-23-responsive cellsIL-23 is a key cytokine in the development of IL-17- and

IL-22-secreting cells. IL-23 signals through a receptor

consisting of the specifi c IL-23R subunit and IL-12Rβ1,

also shared with IL-12R [37]. Loss-of-function polymor-

phisms in IL23R are associated with protection from AS

[7], psoriasis [38], and IBD [39], and many other genes in

the IL-23 pathway are associated with these diseases.

Under physiological conditions, IL-23-, IL-17-, and IL-

22-producing cells are enriched in gut mucosa, and IL-17

and IL-22 are known to be important regulators of

intestinal ‘health’. IL-17 plays important roles in intestinal

homoeostasis in several ways, including maintenance of

epithelial barrier tight junctions [40] and induction of

anti-microbial proteins such as β-defensins, S100

proteins, and REG proteins. IL-22 induces secretion of

anti-microbial peptides [41]. In the gut, innate-like

immune cells act as sentinels, responding very rapidly to


Page 2 of 10

alterations to the microbial composition of the gut with

rapid secretion of IL-17. Key among these sentinels are

γδ T cells, natural killer T (NKT) cells, mucosa-

associated invariant T (MAIT) cells, and lymphoid tissue

inducer (LTi)-like cells (Figure 2).

γδ T cellsγδ T cells are found in large numbers at epithelial surfaces

such as the gut and skin, where they can account for up

to 50% of T cells. γδ T cells not only bear an antigen-

specifi c T-cell receptor (TCR) but also have many pro-

per ties of cells of the innate immune system, including

expression of the major innate immunity receptors and

TLRs [42] and dectin-1 [43], which recognizes microbial

β-glucans. Expression of these receptors supports a role

for γδ T cells in early responses to microbes. Of further

relevance, we and others have recently confi rmed that

CARD9, part of the dectin-1 response pathway, is a

susceptibility gene for AS and IBD [44]. γδ T cells are

potent producers of infl ammatory cytokines such as

inter feron-gamma (IFN-γ), TNF-α, and IL-17 [45,46].

Th ey are pathogenic in the collagen-induced arthritis

model [47] and mouse models of colitis [48], and IL-17-

secreting γδ T cells are expanded in patients with AS

[49]. Intraepithelial γδ T cells also play an important role

in modulating intestinal epithelial growth through

secretion of fi broblast growth factor [50]. Alterations to

γδ T-cell numbers or functions, therefore, may have

profound eff ects on intestinal health.

Natural killer T cellsNKT cells are characterized by expression of an invariant

TCR, Vα24Jα18 in humans and the orthologous

Vα14Jα18 in mice. NKT cells recognize glycolipid struc-

tures presented to them by the non-classic antigen-

presenting molecule CD1d. Like γδ T cells, NKT cells are

rapid responders to antigenic stimuli and are capable of

producing a range of immunoregulatory cytokines

[51-54]. Within the gut, NKT cells are protective in T

helper 1 (TH1)-mediated models of IBD but are patho-

genic in TH2 models [55,56]. Recently, it has been shown

that microbial stimulation of NKT cells in the gut of mice

Figure 1. The physical barrier. Separating the intestinal lumen and its inhabiting commensal bacteria from the underlying lamina propria is a

single layer of intestinal epithelial cells (IECs). These IECs are stitched together, creating a tight junction and regulating the paracellular fl ux. IECs

also secrete soluble factors that are crucial to intestinal homeostasis, such as mucins and anti-microbial peptides (AMPs), including lysozymes, IgA,

defensins, and C-type lectins such as RegIIIγ. Release of these molecules into luminal crypts is thought to prevent microbial invasion into the crypt

microenvironment as well as limit bacteria-epithelial cell contact. Toll-like receptors (TLRs) are also expressed on IECs to sense a breach of barrier or

bacterial invasion. Underneath the IECs, the lamina propria contains T cells, bacteria-sampling dendritic cells (DCs), and macrophages.


Page 3 of 10

aff ects NKT cell phenotypic and functional maturation

[57,58]. Given that NKT cells have protective roles in

models of arthritis [59] and SpA [60], their functional

maturation in the gut provides evidence for a role for

mucosal T-cell priming in infl ammatory joint disease.

Mucosa-associated invariant T cellsMAIT cells are a population of innate-like T cells that are

abundant in human gut, liver, and blood and secrete a

range of infl ammatory cytokines such as IL-17 and IFN-γ

in response to antigenic stimulation [61-63]. Like NKT

cells, MAIT cells bear an invariant TCR (Vα7.2 in

humans) that recognizes antigen presented by the non-

classic MHC-like molecule, MR1 [64]. In blood, MAIT

cells display a memory phenotype [63] and express the

transcription factor ZBTB16 [65], allowing them to

rapidly secrete cytokine in response to antigenic stimuli.

Furthermore, they express high levels of IL-23R [66].

MAIT cells react to a wide range of microbial stimuli,

including Gram-positive and Gram-negative bacteria as

well as yeasts [61,62]. Although the precise role of MAIT

cells in maintenance of the mucosal barrier remains

unclear, the rapid postnatal expansion of MAIT cells and

their acquisition of phenotypic markers of memory likely

represent an interaction with developing commensal

microfl ora [67].

Lymphoid tissue inducer-like cells and innate lymphoid cellsLTi-like cells are found in spleen, lymph node, and gut

lamina propria. LTi-like cells constitutively express hall-

marks of IL-17-secreting cells, including IL-23R, RORγt,

AHR, and CCR6 [68]. Stimulation of LTi-like cells with

IL-23 induces IL-17 secretion [68], whereas PMA and

ionomycin stimulation invokes secretion of IL-22 [69].

LTi cells appear to be related to an increasingly interest-

ing and heterogeneous population of innate lymphoid

Figure 2. The immune barrier. Dendritic cells (DCs) and macrophages patrol the gastrointestinal tract in high numbers. They densely populate

the intestinal lamina propria and form a widespread microbe-sensing network. Activated DCs can secrete a number of cytokines and chemokines,

including interleukin-23 (IL-23), IL-6, and IL-1, activating IL-23-responsive cells. IL-23-, IL-17-, and IL-22-producing cells are enriched in gut mucosa,

and IL-17 and IL-22 are known to be important regulators of intestinal ‘health’. IL-17 plays important roles in intestinal homoeostasis in several ways,

including maintenance of epithelial barrier tight junctions. LTi, lymphoid tissue inducer; MAIT, mucosa-associated invariant T; NKT, natural killer T;

T reg, regulatory T; TNF, tumor necrosis factor.


Page 4 of 10

cells (ILCs). ILCs have been linked to gut infl ammation

through colitis models in which IL-23-responsive ILCs

secrete IL-17 and IFN-γ and promote intestinal infl am-

mation [70]. NKp46+ ILCs are involved in host defense

against Citrobacter rodentium infection through secre-

tion of IL-22 [71].

The human microbiome revolution

Together, the multi-factorial components of the immune

system shape the microbial population that inhabits the

gut and (to an extent) vice versa, with each side pushing

to establish a stable state. Understanding the yin and

yang of this relationship is at the heart of current micro-

biome research.

Microbiome refers to the totality of microbes, their

genetic elements, and environmental interaction in a

defi ned environment [72]. Projects such as the Human

Microbiome Project, run by the National Institutes of

Health, and the Metagenomics of the Human Intestinal

Tract (MetaHit) consortium aim to determine what

constitutes a ‘normal’ or healthy microbiome. Research in

this fi eld has recently been greatly accelerated by the

development of high throughput sequencing methods for

microbial profi ling, which can profi le microbial popu la-

tions whether or not the microbes present can be cultured.

16S rRNA sequencing

16S rRNA is a section of prokaryotic DNA found in all

bacteria and archaea. 16S rRNA sequences are used to

diff erentiate between organisms across all major phyla of

bacteria and to classify strains down to the species level

[73]. 16S rRNA sequencing has dramatically changed our

understanding of phylogeny and microbial diversity

because it provides an unbiased assessment of the

microbiome and is not restricted by the ability to culture

the bacteria present.

Further improvements in sequencing technologies have

reduced the need for targeted studies such as 16S sequen-

cing and enabled high-throughput shotgun sequencing.

Th is latter type of sequen cing randomly samples all genes

present in a habitat rather than just 16S rRNA. Shotgun

sequencing provides more information about the micro-

biome than just the 16S characterization, which is of

particular benefi t in determining the functional capacity of

the microbial community rather than just its phylo geny.

However, shotgun sequencing is more complex to analyze,

owing in part to the challenge of distinguishing between

host and bacterial genomic material, and requires far more

sequencing to be performed. Th erefore, most studies to

date have relied on 16S rRNA sequencing approaches.

Advances in tools for metagenomic studies

Several sequencing technologies have been developed for

human genetic studies and have since been adapted to

metagenomics. Th e Roche 454 sequencing platform

(Roche, Basel, Switzerland) uses large-scale parallel

pyrosequencing to produce reads of between 450 and

1000 base pairs (bp) in length. Read lengths produced by

the 454 platform are well suited to 16S rRNA amplicon

metagenomics studies as well as shotgun sequencing as

they are easily aligned to reference bacterial genetic data

sets. Sequencing plat forms from Illumina (San Diego,

CA, USA), the HiSeq2000 and MiSeq, produce shorter

reads than the Roche 454. Th e HiSeq was designed

primarily for human genomic sequencing, and current

chemistry produces 100-bp paired-end reads. Illumina

sequencing is best suited for shotgun sequencing or

indexed amplicon sequencing of multiple samples.

Th is is a rapidly developing fi eld. Advances in chemistry

are predicted to increase both read lengths and output

for both of these platforms over the next 12 to 24 months,

particularly for the Illumina platforms that are less

mature than the Roche 454. New platforms coming to the

market are likely to have a major impact on meta-

genomics study design. For example, the Pacifi c Bio-

sciences sequencing technology (Pacifi c Biosciences,

Menlo Park, CA, USA) provides reads of approximately

3,000 bases, which will make it particularly suited to this

fi eld, despite its lower overall output (approximately

100 Mb of sequence per run). Life Technologies Ion

Torrent technology (Life Technologies, Grand Island, NY,

USA) is reported to produce up to 400 base reads and up

to 1 Gb of sequence per run. Th e relative positions of

these competing technologies in metagenomics have yet

to be established.

The normal human gut microbiome

To date, only a handful of studies have examined in any

depth the function as well as the composition and

diversity of the human gut microbiome. Two large studies

interrogating and cataloguing microbiomes from various

regions of the body in health and disease have been

undertaken by the National Institutes of Health Human

Microbiome Project in the US and the European MetaHit

project [74-76]. Th e European MetaHit consortium

combined published data sets from around the world and

added 22 newly sequenced fecal metagenomes from four

diff erent European countries. Th ey identifi ed three robust

clusters, termed ‘enterotypes’, that were not nation- or

continent-specifi c. Th ese enterotypes characterized the

microbial phylogenetic variation as well as the function

variation of the clusters at gene and functional class levels

[74]. Each enterotype had a dominant bacterial genus:

enterotype 1 was dominated by the genus Bacteriodes,

enterotype 2 by Prevotella, and enterotype 3 by Rumino-

coccus. Th e three enterotypes were also shown to be

functionally diff erent. For example, enterotype 2, which

is Prevotella-dominant, also contains Desfulfovibrio,


Page 5 of 10

which may act in synergy with Prevotella to degrade

mucin glycoproteins present in the mucosal layer of the

gut. It may be that diff erent enterotypes may be asso-

ciated with diseases such as obesity and IBD rather than

necessarily specifi c bacterial species, given their diff ering

functional capacities. Further studies will be required to

determine whether enterotypes are consis tently found in

expanded data sets and in studies of intestinal biopsies as

well as the fecal samples used in the MetaHit study.

The microbiome in immune-mediated arthritis

Th e major fi ndings of microbial profi ling studies in major

immune-mediated diseases associated with arthritis are

summarised in Table 1. To date, no study investigating

the gut microbiome by using sequencing-based methods

in AS has been reported. Many studies largely using

antibody tests have suggested an increased carriage of

Klebsiella species in patients with AS, but this has not

been universally supported [77]. One study using

denaturing gradient gel electrophoresis to profi le the

microbiome by using fecal samples found no diff erences

between AS cases and healthy controls [78].

Rheumatoid arthritis (RA) is the only infl ammatory

arthritis for which modern metagenomic studies have

been reported. Community profi ling studies of gut fl ora

of patients with RA reveal diff erences in the composition

of gut microbiota of patients with RA compared with

those of healthy controls, and a lower abundance of

Bifi dobacterium and Bacteroides bacteria was observed

in RA cases [79,80]. However, these studies used fecal

samples and not intestinal biopsies, possibly infl uencing

the populations observed [81]. Also, microbial profi ling

data suggest that gingival infection may be important,

particularly in anti-citrullinated peptide antibody-posi-

tive RA, although the studies suggesting this have

generally used antibody tests rather than sequencing-

based approaches (reviewed in [82]).

Several lines of evidence indicate that the gut micro-

biome plays an important role in IBD, including the

association of genes involved in mucosal immunity with

IBD (such as CARD15, CARD9, IL-23R, and ATG16L1),

the therapeutic eff ect of antibiotics on the condition, and

the benefi cial eff ect of fecal stream diversion in Crohn’s

disease. Previous studies of human IBD have been under-

taken by using standard culture techniques (for example,

[83]) or molecular analysis (for example, [84,85]). Th ese

studies noted alterations in intestinal microbiota when

compared with non-IBD patients, a fi nding recently

confi rmed by using 16S rRNA sequencing of intestinal

biopsies [86]. Th is sequencing study, however, did not

identify an IBD-specifi c microbial profi le, perhaps because

of insuffi cient resolution of the sequencing performed or

small sample size (12 patients and 5 controls). Much larger

studies will be required to dissect the relationship

between the host genetic factors determining risk of IBD

and the gut microbiome.

Th e microbiome is thought to play a signifi cant role in

psoriasis, another AS-related condition. It has long been

suggested that streptococcal infection, especially throat

infections, may trigger psoriasis in a genetically suscep-

tible individual [87]. Recent studies using 16S rRNA

sequencing have found signifi cant diff erences between

the cutaneous microbiota of psoriasis cases and controls

and in involved and control skin in psoriasis cases, and

less staphylococci and propionibacteria have been ob-

served in cases and in aff ected skin [88]. Again, further

studies will be required to determine whether there is a

particular microbial profi le or specifi c bacterial species

involved in psoriasis.

Th ese studies are thus consistent with the hypothesis

that the microbiome contributes to the etiopathogenesis

of immune-mediated arthritis or seronegative diseases

like IBD and psoriasis. However, at this point, there is no

defi nitive evidence of specifi c bacterial infections or

changes in microbial profi le that play a causative role in

these conditions (with the exception of reactive arthritis).

Chicken and the egg

Whether the changes noted in these early metagenomic

studies of immune-mediated disease are a consequence

of disease or are involved in its development or persis-

tence is unclear. Th is distinction may prove impossible to

dissect in human studies, but considerable evidence in

studies in mice supports a role for the microbiome in

driving immune-mediated diseases.

In the B27 rat model of AS, rats housed under GF

conditions did not develop disease [89], demonstrating

that microbes in this model are important for disease

penetrance. In contrast, in the New Zealand black model

of systemic lupus erythematosus, mice maintained under

GF conditions produced higher levels of antinuclear

antibodies and developed worse disease [90],

demonstrating a protective role for commensal microbes.

Given the IL-17/IL-22 cytokines, which are of relevance

to AS, IBD, and psoriasis in particular, strong evidence

from murine studies indicates that interaction between

the gut microbiome and the host determines the overall

level of activation of the immune cells producing these

cytokines. Segmented fi lamentous bacteria (SFB) are

com mensal bacteria that induce IL-17 secretion. Mice

lacking SFB have low levels of intestinal IL-17 and are

susceptible to infection with pathogenic Citrobacter spp.

Restoration of SFB in these mice increased the number of

gut-resident IL-17-producing cells and enhanced resis-

tance to infection [12]. Salzman and colleagues [91]

illustrated that α-defensins modulate mucosal T-cell

response by regulating the composition, but not the total

numbers, of bacteria in the intestinal microbiome.


Page 6 of 10

Examin ing the intestinal microbiota of mice expressing

the human α-defensin gene, they demonstrated signifi -

cant α-defensin-dependent alterations in commensal

composition, leading to a loss of SFB and fewer IL-17-

producing lamina propria T cells [91]. Furthermore,

using recolonization studies, investigators recently demon-

strated that, in neonatal mice, commensal microbes

infl uence invariant NKT (iNKT) cell intestinal infi ltration

and activation, establishing mucosal iNKT cell tolerance

to later environmental exposures [92].

Th e mechanism by which the microbiome infl uences

IL-17-producing cell activation is still being determined.

Ivanov and colleagues [12] demonstrated that serum

amyloid A, produced in the terminal ileum, can induce

TH17 diff erentiation of CD4+ T lymphocytes. It has also

been shown that development of TH17 lymphocytes in

the intestine is stimulated by microbiota-induced IL-1β

(but not IL-6) production [93]. Colonization with

Clostridial species has been shown to stimulate intestinal

transforming growth factor-beta (TGF-β) production, in

turn increasing IL-10+ CTLA4high Treg (regulatory T)

activation [94]. Clostridial coloniza tion of neonatal mice

reduced severity of induced colitis by using dextran

sulphate sodium or oxazolone and reduced serum IgE

levels in adulthood. So it is likely that alterations to the

gut microfl ora or invasion of the gut by pathogenic

bacteria infl uences the balance of IL-17- and IL-22-

producing cells and other immune cells, infl uen cing

susceptibility to local and systemic immune-mediated

disease.

Th e above studies highlight that changes in the micro-

biome can lead to infl ammation which may have far-

reaching eff ects and demonstrate experimental approaches

by which fi ndings of metagenomic studies in mice and

humans can be explored to successfully dissect the role of

the microbiome in human immune-mediated diseases.

Conclusions

Th e human gut microbiome is a dynamic and complex

ecosystem that is only now beginning to be understood.

Table 1. Alterations in gut microbiota associated with immune-mediated diseases

Associated microbes Microbiota changes References

IBD – Crohn’s disease

Gut microbiome

Bacteroides ovatus Bacteroides vulgatus Bacteroides uniformis

Reduction in microbial diversity when compared with controls [95]

IBD

Gut microbiome

Bacteroidetes Lachnospiraceae Actinobacteria Proteobacteria Clostridium Firmicutes/Bacteroidetes ratio Bifi dobacteria

Associated with overall community shift and dysbiosis [96]

Celiac disease

Gut microbiome

Bacteroides vulgatus Escherichia coli Clostridium coccoides

Overall higher diversity of microbes in patients with celiac disease compared with controls [97]

Psoriasis

Skin microbiome

Firmicutes Bacteroidetes Actinobacteria Proteobacteria

Overrepresentation of Firmicutes and an underrepresentation of Actinobacteria and

Proteobacteria when compared with controls

[98,99]

Rheumatoid arthritis

Oral microbiome

Porphyromonas gingivalis

Dysbiosis and increased diversity [100,101]

IBD, infl ammatory bowel disease.


Page 7 of 10

It is increasingly clear that the interaction between host

and microbiome profoundly aff ects health. It is still

unclear how interactions between host genes, microbes,

and environmental factors can predispose patients to the

development autoimmune diseases such as AS. We are

only beginning to grasp the infl uence the microbiome has

on health. Improved knowledge of the composition and

function of the gut microbiome in patients with AS and

how the microbiome shapes the immune response and

infl uences infl ammation, both local and systemic, will

likely provide important insights into early events in the

pathogenesis of AS.

Abbreviations

AS, ankylosing spondylitis; bp, base pair(s); DC, dendritic cell; GF, germ-free;

IBD, infl ammatory bowel disease; IEC, intestinal epithelial cell; IFNγ, interferon-

gamma; IL, interleukin; IL-23R, interleukin-23 receptor; ILC, innate lymphoid

cell; iNKT, invariant natural killer T; LTi, lymphoid tissue inducer; MAIT, mucosa-

associated invariant T; MetaHit, Metagenomics of the Human Intestinal Tract;

NKT, natural killer T; RA, rheumatoid arthritis; SFB, segmented fi lamentous

bacteria; SpA, spondyloarthopathies; TCR, T-cell receptor; TH, T helper; TLR, Toll-

like receptor; TNF-α, tumor necrosis factor-alpha.

Competing interests

The authors declare that they have no competing interests.

Acknowledgments

The authors thank Murray Hargrave and Philip Robinson for their assistance

in the preparation of this review. TJK is supported by an Arthritis Australia

Heald Fellowship. MAB is suppor ted by Australian National Health and Medical

Research Council Senior Principal Research Fellowship APP1024879. DE is

supported by grants of the Fund for Scientifi c Research Flanders and the

Research Council of Ghent University.

Author details1The University of Queensland Diamantina Institute, Translational Research

Institute, Level 7, 37 Kent Road, Princess Alexandra Hospital, Woolloongabba,

Brisbane QLD 4102, Australia. 2Laboratory for Molecular Immunology and

Infl ammation, Department of Rheumatology, Ghent University Hospital, De

Pintelaan 185, 9000 Ghent, Belgium.

Published: 6 June 2013

References

1. Thomas GP, Br own MA: Genetics and genomics of ankylosing spondylitis. Immunol Rev 2010, 233:162-180.

2. Mielant s H, Veys EM, Goemaere S, Goethals K, Cuvelier C, De Vos M: Gut infl ammation in the spondyloarthropathies: clinical, radiologic, biologic and genetic features in relation to the type of histology. A prospective study. J Rheumatol 1991, 18:1542-1551.

3. Uotila T, Antonen J, Laine J, Kujansuu E, Haapala AM, Lumio J, Vuento R, Oksa

H, Herrala J, Kuusi M, Mustonen J, Korpela M, Pirkanmaa Waterborne

Outbreak Study Group: Reactive arthritis in a population exposed to an extensive waterborne gastroenteritis outbreak after sewage contamination in Pirkanmaa, Finland. Scand J Rheumatol 2011, 40:358-362.

4. Thjodle ifsson B, Geirsson AJ, Bjornsson S, Bjarnason I: A common genetic background for infl ammatory bowel disease and ankylosing spondylitis: a genealogic study in Iceland. Arthritis Rheum 2007, 56:2633-2639.

5. Danoy P , Pryce K, Hadler J, Bradbury LA, Farrar C, Pointon J; Australo-Anglo-

American Spondyloarthritis Consortium, Ward M, Weisman M, Reveille JD,

Wordsworth BP, Stone MA; Spondyloarthritis Research Consortium of Canada,

Maksymowych WP, Rahman P, Gladman D, Inman RD, Brown MA: Association of variants at 1q32 and STAT3 with ankylosing spondylitis suggests genetic overlap with Crohn’s disease. PLoS Genet 2010, 6:e1001195.

6. Austral o-Anglo-American Spondyloarthritis Consortium (TASC), Reveille JD,

Sims AM, Danoy P, Evans DM, Leo P, Pointon JJ, Jin R, Zhou X, Bradbury LA,

Appleton LH, Davis JC, Diekman L, Doan T, Dowling A, Duan R, Duncan EL,

Farrar C, Hadler J, Harvey D, Karaderi T, Mogg R, Pomeroy E, Pryce K, Taylor J,

Savage L, Deloukas P, Kumanduri V, Peltonen L, Ring SM, Whittaker P, Glazov E,

Thomas GP, Maksymowych WP, Inman RD, Ward MM, Stone MA, Weisman MH,

Wordsworth BP, Brown MA: Genome-wide association study of ankylosing spondylitis identifi es non-MHC susceptibility loci. Nat Genet 2010,

42:123-127.

7. Wellcom e Trust Case Control Consortium; Australo-Anglo-American

Spondylitis Consortium (TASC), Burton PR, Clayton DG, Cardon LR, Craddock

N, Deloukas P, Duncanson A, Kwiatkowski DP, McCarthy MI, Ouwehand WH,

Samani NJ, Todd JA, Donnelly P, Barrett JC, Davison D, Easton D, Evans DM,

Leung HT, Marchini JL, Morris AP, Spencer CC, Tobin MD, Attwood AP,

Boorman JP, Cant B, Everson U, Hussey JM, Jolley JD, Knight AS, Koch K,

Meech E, et al.: Association scan of 14,500 nonsynonymous SNPs in four diseases identifi es autoimmunity variants. Nat Genet 2007, 39:1329-1337.

8. Robinso n PC, Brown MA: The genetics of ankylosing spondylitis and axial spondyloarthritis. Rheum Dis Clin North Am 2012, 38:539-553.

9. Laukens D, Georges M, Libioulle C, Sandor C, Mni M, Vander Cruyssen B,

Peeters H, Elewaut D, De Vos M: Evidence for signifi cant overlap between common risk variants for Crohn’s disease and ankylosing spondylitis. PLoS ONE 2010, 5:e13795.

10. Round JL, Mazmanian SK: The gut microbiota shapes intestinal immune responses during health and disease. Nat Rev Immunol 2009, 9:313-323.

11. Clemen te JC, Ursell LK, Parfrey LW, Knight R: The impact of the gut microbiota on human health: an integrative view. Cell 2012, 148:1258-1270.

12. Ivanov II, Atarashi K, Manel N, Brodie EL, Shima T, Karaoz U, Wei D, Goldfarb

KC, Santee CA, Lynch SV, Tanoue T, Imaoka A, Itoh K, Takeda K, Umesaki Y,

Honda K, Littman DR: Induction of intestinal Th17 cells by segmented fi lamentous bacteria. Cell 2009, 139:485-498.

13. Duan J , Chung H, Troy E, Kasper DL: Microbial colonization drives expansion of IL-1 receptor 1-expressing and IL-17-producing gamma/delta T cells. Cell Host Microbe 2010, 7:140-150.

14. Cho I, Blaser MJ: The human microbiome: at the interface of health and disease. Nat Rev Genet 2012, 13:10.

15. Maloy KJ, Powrie F: Intestinal homeostasis and its breakdown in infl ammatory bowel disease. Nature 2011, 474:298-306.

16. Koslow ski MJ, Beisner J, Stange EF, Wehkamp J: Innate antimicrobial host defense in small intestinal Crohn’s disease. Int J Med Microbiol 2010,

300:34-40.

17. Wehkam p J, Koslowski M, Wang G, Stange EF: Barrier dysfunction due to distinct defensin defi ciencies in small intestinal and colonic Crohn’s disease. Mucosal Immunol 2008, 1:S67-S74.

18. Simms LA, Doecke JD, Walsh MD, Huang N, Fowler EV, Radford-Smith GL:

Reduced α-defensin expression is associated with infl ammation and not NOD2 mutation status in ileal Crohn’s disease. Gut 2008, 57:903-910.

19. Wehkamp J, Salzman NH, Porter E, Nuding S, Weichenthal M, Petras RE, Shen

B, Schaeff eler E, Schwab M, Linzmeier R, Feathers RW, Chu H, Lima H Jr,

Fellermann K, Ganz T, Stange EF, Bevins CL: Reduced Paneth cell α-defensins in ileal Crohn’s disease. Proc Natl Acad Sci U S A 2005, 102:18129-18134.

20. Ciccia F, Bombardieri M, Rizzo A, Principato A, Giardina AR, Raiata F, Peralta S,

Ferrante A, Drago S, Cottone M, Pitzalis C, Triolo G: Over-expression of paneth cell-derived anti-microbial peptides in the gut of patients with ankylosing spondylitis and subclinical intestinal infl ammation. Rheumatology 2010, 49:2076-2083.

21. Heazlewood CK, Cook MC, Eri R, Price GR, Tauro SB, Taupin D, Thornton DJ,

Png CW, Crockford TL, Cornall RJ, Adams R, Kato M, Nelms KA, Hong NA,

Florin TH, Goodnow CC, McGuckin MA: Aberrant mucin assembly in mice causes endoplasmic reticulum stress and spontaneous infl ammation resembling ulcerative colitis. PLoS Med 2008, 5:e54.

22. Sherlock J P, Joyce-Shaikh B, Turner SP, Chao CC, Sathe M, Grein J, Gorman DM,

Bowman EP, McClanahan TK, Yearley JH, Eberl G, Buckley CD, Kastelein RA,

Pierce RH, Laface DM, Cua DJ: IL-23 induces spondyloarthropathy by acting on ROR-gammat+ CD3+CD4-CD8- entheseal resident T cells. Nat Med

2012, 18:1069-1076.

23. Artis D: E pithelial-cell recognition of commensal bacteria and maintenance of immune homeostasis in the gut. Nat Rev Immunol 2008,

8:411-420.

24. Reid G, Yo unes JA, Van der Mei HC, Gloor GB, Knight R, Busscher HJ:

Microbiota restoration: natural and supplemented recovery of human microbial communities. Nat Rev Micro 2011, 9:27-38.

25. Lubrano E, Ciacci C, Ames PR, Mazzacca G, Oriente P, Scarpa R: The arthritis of coeliac disease: prevalence and pattern in 200 adult patients.


Page 8 of 10

Rheumatology 1996, 35:1314-1318.

26. Vaile JH, Meddings JB, Yacyshyn BR, Russell AS, Maksymowych WP: Bowel permeability and CD45RO expression on circulating CD20+ B cells in patients with ankylosing spondylitis and their relatives. J Rheumatol 1999,

26:128-135.

27. Martínez-G onzález O, Cantero-Hinojosa J, Paule-Sastre P, Gómez-Magán JC,

Salvatierra-Ríos D: Intestinal permeability in patients with ankylosing spondylitis and their healthy relatives. Rheumatology 1994, 33:644-647.

28. Chabot S, Wagner JS, Farrant S, Neutra MR: TLRs regulate the gatekeeping functions of the intestinal follicle-associated epithelium. J Immunol 2006,

176:4275-4283.

29. Suzuki K, Ha SA, Tsuji M, Fagarasan S: Intestinal IgA synthes is: a primitive form of adaptive immunity that regulates microbial communities in the gut. Semin Immunol 2007, 19:127-135.

30. Wesa A, Galy A: Increased production of pro-infl ammatory cyto kines and enhanced T cell responses after activation of human dendritic cells with IL-1 and CD40 ligand. BMC Immunology 2002, 3:14.

31. Cruz CM, Rinna A, Forman HJ, Ventura AL, Persechini PM, Ojciu s DM: ATP activates a reactive oxygen species-dependent oxidative stress response and secretion of proinfl ammatory cytokines in macrophages. J Biol Chem

2007, 282:2871-2879.

32. Hirotani T, Lee PY, Kuwata H, Yamamoto M, Matsumoto M, Kawase I, Akira S,

Takeda K: The nuclear IkappaB protein IkappaBNS selectively inhibits lipopolysaccharide-induced IL-6 production in macrophages of the colonic lamina propria. J Immunol 2005, 174:3650-3657.

33. Denning TL, Wang Y-c, Patel SR, Williams IR, Pulendran B: Lam ina propria macrophages and dendritic cells diff erentially induce regulatory and interleukin 17-producing T cell responses. Nat Immunol 2007, 8:1086-1094.

34. Glocker EO, Kotlarz D, Boztug K, Gertz EM, Schäff er AA, Noyan F, Perro M,

Diestelhorst J, Allroth A, Murugan D, Hätscher N, Pfeifer D, Sykora KW, Sauer

M, Kreipe H, Lacher M, Nustede R, Woellner C, Baumann U, Salzer U, Koletzko

S, Shah N, Segal AW, Sauerbrey A, Buderus S, Snapper SB, Grimbacher B, Klein

C: Infl ammatory bowel disease and mutations aff ecting the interleukin-10 receptor. N Engl J Med 2009, 361:2033-2045.

35. Kamada N, Hisamatsu T, Okamoto S, Chinen H, Kobayashi T, Sato T, Sakuraba

A, Kitaz ume MT, Sugita A, Koganei K, Akagawa KS, Hibi T: Unique CD14+ intestinal macrophages contribute to the pathogenesis of Crohn disease via IL-23/IFN-γ axis. J Clin Invest 2008, 118:2269-2280.

36. Pull SL, Doherty JM, Mills JC, Gordon JI, Stappenbeck TS: Activated macrophages are an adaptive element of the colonic epithelial progenitor niche necessary for regenerative responses to injury. Proc Natl Acad Sci

U S A 2005, 102:99-104.

37. Oppmann B, Lesley R, Blom B, Timans JC, Xu Y, Hunte B, Vega F, Yu N, Wang J,

Singh K , Zonin F, Vaisberg E, Churakova T, Liu M, Gorman D, Wagner J,

Zurawski S, Liu Y, Abrams JS, Moore KW, Rennick D, de Waal-Malefyt R,

Hannum C, Bazan JF, Kastelein RA: Novel p19 protein engages IL-12p40 to form a cytokine, IL-23, with biological activities similar as well as distinct from IL-12. Immunity 2000, 13:715-725.

38. Cargill M, Schrodi SJ, Chang M, Garcia VE, Brandon R, Callis KP, Matsunami N,

Ardlie KG, Civello D, Catanese JJ, Leong DU, Panko JM, McAllister LB, Hansen

CB, Papenfuss J, Prescott SM, White TJ, Leppert MF, Krueger GG, Begovich AB:

A large-scale genetic association study confi rms IL12B and leads to the identifi cation of IL23R as psoriasis-risk genes. Am J Hum Gen 2007,

80:273-290.

39. Duerr RH, Taylor KD, Brant SR, Rioux JD, Silverberg MS, Daly MJ, Steinhart AH,

Abrah am C, Regueiro M, Griffi ths A, Dassopoulos T, Bitton A, Yang H, Targan S,

Datta LW, Kistner EO, Schumm LP, Lee AT, Gregersen PK, Barmada MM, Rotter

JI, Nicolae DL, Cho JH: A genome-wide association study identifi es IL23R as an infl ammatory bowel disease gene. Science 2006, 314:1461-1463.

40. Ogawa A, Andoh A, Araki Y, Bamba T, Fu jiyama Y: Neutralization of interleukin-17 aggravates dextran sulfate sodium-induced colitis in mice. Clin Immunol 2004, 110:55-62.

41. Zheng SG, Wang J, Horwitz DA: Cutting edge: Foxp3+CD4+CD25+ regulatory T cells induced by IL-2 and TGF-beta are resistant to Th17 conversion by IL-6. J Immunol 2008, 180:7112-7116.

42. Mokuno Y, Matsuguchi T, Takano M, Nish imura H, Washizu J, Ogawa T,

Takeuchi O, Akira S, Nimura Y, Yoshikai Y: Expression of Toll-like receptor 2 on γδ T cells bearing invariant Vγ6/Vδ1 induced by Escherichia coli infection in mice. J Immunol 2000, 165:931-940.

43. Martin B, Hirota K, Cua DJ, Stockinger B, V eldhoen M: Interleukin-17-producing gammadelta T cells selectively expand in response to

pathogen products and environmental signals. Immunity 2009, 31:321-330.

44. Evans DM, Spencer CC, Pointon JJ, Su Z, Har vey D, Kochan G, Oppermann U,

Dilthey A, Pirinen M, Stone MA, Appleton L, Moutsianas L, Leslie S,

Wordsworth T, Kenna TJ, Karaderi T, Thomas GP, Ward MM, Weisman MH,

Farrar C, Bradbury LA, Danoy P, Inman RD, Maksymowych W, Gladman D,

Rahman P; Spondyloarthritis Research Consortium of Canada (SPARCC),

Morgan A, Marzo-Ortega H, Bowness P, et al.: Interaction between ERAP1 and HLA-B27 in ankylosing spondylitis implicates peptide handling in the mechanism for HLA-B27 in disease susceptibility. Nat Genet 2011, 43:6.

45. Lockhart E, Green AM, Flynn JL: IL-17 produ ction is dominated by gammadelta T cells rather than CD4 T cells during Mycobacterium tuberculosis infection. J Immunol 2006, 177:4662-4669.

46. Shibata K, Yamada H, Hara H, Kishihara K, Y oshikai Y: Resident Vdelta1+ gammadelta T cells control early infi ltration of neutrophils after Escherichia coli infection via IL-17 production. J Immunol 2007,

178:4466-4472.

47. Roark CL, French JD, Taylor MA, Bendele AM, Born WK, O’Brien RL:

Exacerbation of collagen-induced arthritis by oligoclonal, IL-17-producing gamma delta T cells. J Immunol 2007, 179:5576-5583.

48. Nanno M, Kanari Y, Naito T, Inoue N, Hisama tsu T, Chinen H, Sugimoto K,

Shimomura Y, Yamagishi H, Shiohara T, Ueha S, Matsushima K, Suematsu M,

Mizoguchi A, Hibi T, Bhan AK, Ishikawa H: Exacerbating role of gammadelta T cells in chronic colitis of T-cell receptor alpha mutant mice. Gastroenterology 2008, 134:481-490.

49. Kenna TJ, Davidson SI, Duan R, Bradbury LA, McFarlane J, Smith M, Weedon H,

Street S, Thomas R, Thomas GP, Brown MA: Enrichment of circulating interleukin-17-secreting interleukin-23 receptor-positive gamma/delta T cells in patients with active ankylosing spondylitis. Arthritis Rheum 2012,

64:1420-1429.

50. Boismenu R, Havran WL: Modulation of epithe lial cell growth by intraepithelial gamma delta T cells. Science 1994, 266:1253-1255.

51. Bendelac A, Savage PB, Teyton L: The biolog y of NKT cells. Ann Rev Immunol

2007, 25:297-336.

52. Rachitskaya AV, Hansen AM, Horai R, Li Z, V illasmil R, Luger D, Nussenblatt RB,

Caspi RR: Cutting edge: NKT cells constitutively express IL-23 receptor and RORgammat and rapidly produce IL-17 upon receptor ligation in an IL-6-independent fashion. J Immunol 2008, 180:5167-5171.

53. Akbari O, Stock P, Meyer E, Kronenberg M, S idobre S, Nakayama T, Taniguchi

M, Grusby MJ, DeKruyff RH, Umetsu DT: Essential role of NKT cells producing IL-4 and IL-13 in the development of allergen-induced airway hyperreactivity. Nat Med 2003, 9:582-588.

54. Baxter AG, Kinder SJ, Hammond KJ, Scollay R , Godfrey DI: Association between alphabetaTCR+CD4-CD8- T-cell defi ciency and IDDM in NOD/Lt mice. Diabetes 1997, 46:572-582.

55. Heller F, Fuss IJ, Nieuwenhuis EE, Blumberg RS, Strober W: Oxazolone colitis, a Th2 colitis model resembling ulcerative colitis, is mediated by IL-13-producing NK-T cells. Immunity 2002, 17:629-638.

56. Ueno Y, Tanaka S, Sumii M, Miyake S, Tazuma S, Taniguchi M, Yamamura T,

Chayama K: Single dose of OCH improves mucosal T helper type 1/T helper type 2 cytokine balance and prevents experimental colitis in the presence of Vα14 natural killer T cells in mice. Infl amm Bowel Dis 2005, 11:35-41.

57. Coppieters KT, Elewaut D: Natural killer T c ells: born in the thymus, raised in the gut. Gastroenterology 2012, 143:293-296.

58. Wingender G, Stepniak D, Krebs P, Lin L, McB ride S, Wei B, Braun J, Mazmanian

SK, Kronenberg M: Intestinal microbes aff ect phenotypes and functions of invariant natural killer T cells in mice. Gastroenterology 2012, 143:418-428.

59. Coppieters K, Van Beneden K, Jacques P, Dewi nt P, Vervloet A, Vander

Cruyssen B, Van Calenbergh S, Chen G, Franck RW, Verbruggen G, Deforce D,

Matthys P, Tsuji M, Rottiers P, Elewaut D: A single early activation of invariant NK T cells confers long-term protection against collagen-induced arthritis in a ligand-specifi c manner. J Immunol 2007, 179:2300-2309.

60. Jacques P, Venken K, Van Beneden K, Hammad H , Seeuws S, Drennan MB,

Deforce D, Verbruggen G, Apostolaki M, Kollias G, Lambrecht BN, De Vos M,

Elewaut D: Invariant natural killer T cells are natural regulators of murine spondylarthritis. Arthritis Rheum 2010, 62:988-999.

61. Gold MC, Cerri S, Smyk-Pearson S, Cansler ME , Vogt TM, Delepine J, Winata E,

Swarbrick GM, Chua WJ, Yu YY, Lantz O, Cook MS, Null MD, Jacoby DB, Harriff

MJ, Lewinsohn DA, Hansen TH, Lewinsohn DM: Human mucosal associated invariant T cells detect bacterially infected cells. PLoS Biol 2010, 8:e1000407.

62. Le Bourhis L, Martin E, Péguillet I, Guihot A, Froux N, Coré M, Lévy E, Dusseaux

M, Meyssonnier V, Premel V, Ngo C, Riteau B, Duban L, Robert D, Huang S,


Page 9 of 10

Rottman M, Soudais C, Lantz O: Antimicrobial activity of mucosal-associated invariant T cells. Nat Immunol 2010, 11:701-708.

63. Dusseaux M, Martin E, Serriari N, Péguillet I, Premel V, Louis D, M ilder M,

Le Bourhis L, Soudais C, Treiner E, Lantz O: Human MAIT cells are xenobiotic-resistant, tissue-targeted, CD161hi IL-17–secreting T cells. Blood 2011,

117:1250-1259.

64. Treiner E, Duban L, Bahram S, Radosavljevic M, Wanner V, Tilloy F, Aff aticati P,

Gilfi llan S, Lantz O: Selection of evolutionarily conserved mucosal-associated invariant T cells by MR1. Nature 2003, 422:164-169.

65. Savage AK, Constantinides MG, Han J, Picard D, Martin E, Li B, Lantz O,

Bendel ac A: The transcription factor PLZF directs the eff ector program of the NKT cell lineage. Immunity 2008, 29:391-403.

66. Billerbeck E, Kang YH, Walker L, Lockstone H, Grafmueller S, Fleming V, Flint J,

Willberg CB, Bengsch B, Seigel B, Ramamurthy N, Zitzmann N, Barnes EJ,

Thevanayagam J, Bhagwanani A, Leslie A, Oo YH, Kollnberger S, Bowness P,

Drognitz O, Adams DH, Blum HE, Thimme R, Klenerman P: Analysis of CD161 expression on human CD8+ T cells defi nes a distinct functional subset with tissue-homing properties. Proc Natl Acad Sci U S A 2010, 107:3006-3011.

67. Le Bourhis L, Guerri L, Dusseaux M, Martin E, Soudais C, Lantz O: Mucosal-asso ciated invariant T cells: unconventional development and function. Trends Immunol 2011, 32:212-218.

68. Takatori H, Kanno Y, Watford WT, Tato CM, Weiss G, Ivanov, II, Littman DR,

O’S hea JJ: Lymphoid tissue inducer-like cells are an innate source of IL-17 and IL-22. J Exp Med 2009, 206:35-41.

69. Cupedo T, Crellin NK, Papazian N, Rombouts EJ, Weijer K, Grogan JL, Fibbe WE,

Cornelissen JJ, Spits H: Human fetal lymphoid tissue-inducer cells are interleukin 17-producing precursors to RORC+ CD127+ natural killer-like cells. Nat Immunol 2009, 10:66-74.

70. Buonocore S, Ahern PP, Uhlig HH, Ivanov, II, Littman DR, Maloy KJ, Powrie F:

I nnate lymphoid cells drive interleukin-23-dependent innate intestinal pathology. Nature 2010, 464:1371-1375.

71. Cella M, Fuchs A, Vermi W, Facchetti F, Otero K, Lennerz JK, Doherty JM, Mills

JC, Colonna M: A human natural killer cell subset provides an innate source of IL-22 for mucosal immunity. Nature 2009, 457:722-725.

72. Lederberg J, McCray AT: ‘Ome sweet ‘omics - A genealogical treasury of words. Scientist 2001, 15:8.

73. Woese CR: Bacterial evolution. Microbiol Rev 1987, 51:221-271.

74. Arumugam M, Raes J, Pelletier E, Le Paslier D, Yamada T, Mende DR, Fernandes

G R, Tap J, Bruls T, Batto JM, Bertalan M, Borruel N, Casellas F, Fernandez L,

Gautier L, Hansen T, Hattori M, Hayashi T, Kleerebezem M, Kurokawa K, Leclerc

M, Levenez F, Manichanh C, Nielsen HB, Nielsen T, Pons N, Poulain J, Qin J,

Sicheritz-Ponten T, Tims S, et al.: Enterotypes of the human gut microbiome. Nature 2011, 473:174-180.

75. NIH HMP Working Group, Peterson J, Garges S, Giovanni M, McInnes P, Wang

L, Sc hloss JA, Bonazzi V, McEwen JE, Wetterstrand KA, Deal C, Baker CC, Di

Francesco V, Howcroft TK, Karp RW, Lunsford RD, Wellington CR, Belachew T,

Wright M, Giblin C, David H, Mills M, Salomon R, Mullins C, Akolkar B, Begg L,

Davis C, Grandison L, Humble M, Khalsa J, Little AR, et al.: The NIH Human Microbiome Project. Genome Res 2009, 19:2317-2323.

76. Qin J, Li R, Raes J, Arumugam M, Burgdorf KS, Manichanh C, Nielsen T, Pons N,

Levenez F, Yamada T, Mende DR, Li J, Xu J, Li S, Li D, Cao J, Wang B, Liang H,

Zheng H, Xie Y, Tap J, Lepage P, Bertalan M, Batto JM, Hansen T, Le Paslier D,

Linneberg A, Nielsen HB, Pelletier E, Renault P, et al.: A human gut microbial gene catalogue established by metagenomic sequencing. Nature 2010,

464:59-65.

77. Stone MA, Payne U, Schentag C, Rahman P, Pacheco-Tena C, Inman RD:

Comparative immune responses to candidate arthritogenic bacteria do not confi rm a dominant role for Klebsiella pneumonia in the pathogenesis of familial ankylosing spondylitis. Rheumatology 2004, 43:148-155.

78. Stebbings S, Munro K, Simon MA, Tannock G, Highton J, Harmsen H, Welling

G, Se ksik P, Dore J, Grame G, Tilsala-Timisjarvi A: Comparison of the faecal microfl ora of patients with ankylosing spondylitis and controls using molecular methods of analysis. Rheumatology 2002, 41:1395-1401.

79. Eerola E, Mottonen T, Hannonen P, Luukkainen R, Kantola I, Vuori K, Tuominen

J , Toivanen P: Intestinal fl ora in early rheumatoid arthritis. Rheumatology

1994, 33:1030-1038.

80. Vaahtovuo J, Munukka E, Korkeamaki M, Luukainen R, Toivanen P: Fecal microbiot a in early rheumatoid arthritis. J Rheumatol 2008, 35:1500-1505.

81. Eckburg PB, Bik EM, Bernstein CN, Purdom E, Dethlefsen L, Sargent M, Gill SR,

Nelson KE, Relman DA: Diversity of the human intestinal microbial fl ora. Science 2005, 308:1635-1638.

82. Mikuls TR: Help stop tooth decay...and prevent RA? J Rheumatol 2010,

37:1083-1 085.

83. Giaff er MH, Holdsworth CD, Duerden BI: The assessment of faecal fl ora in patie nts with infl ammatory bowel disease by a simplifi ed bacteriological technique. J Med Microbiol 1991, 35:238-243.

84. Seksik P, Rigottier-Gois L, Gramet G, Sutren M, Pochart P, Marteau P, Jian R,

Doré J: Alterations of the dominant faecal bacterial groups in patients with Crohn’s disease of the colon. Gut 2003, 52:237-242.

85. Ott SJ, Musfeldt M, Wenderoth DF, Hampe J, Brant O, Folsch UR, Timmis KN,

Schrei ber S: Reduction in diversity of the colonic mucosa associated bacterial microfl ora in patients with active infl ammatory bowel disease. Gut 2004, 53:685-693.

86. Walker AW, Sanderson JD, Churcher C, Parkes GC, Hudspith BN, Rayment N,

Brostoff J, Parkhill J, Dougan G, Petrovska L: High-throughput clone library analysis of the mucosa-associated microbiota reveals dysbiosis and diff erences between infl amed and non-infl amed regions of the intestine in infl ammatory bowel disease. BMC Microbiol 2011, 11:7.

87. Krueger JG: The immunologic basis for the treatment of psoriasis with new biolog ic agents. J Am Acad Dermatol 2002, 46:1-26.

88. Fahlen A, Engstrand L, Baker BS, Powles A, Fry L: Comparison of bacterial microb iota in skin biopsies from normal and psoriatic skin. Arch Dermatol

Res 2012, 304:15-22.

89. Rath HC, Herfarth HH, Ikeda JS, Grenther WB, Hamm TE, Balish E, Taurog JD,

Hamme r RE, Wilson KH, Sartor RB: Normal luminal bacteria, especially Bacteroides species, mediate chronic colitis, gastritis, and arthritis in HLA-B27/human beta2 microglobulin transgenic rats. J Clin Invest 1996,

98:945-953.

90. Unni K, Holley K, McDuffi e F, Titus J: Comparative study of NZB mice under germf ree and conventional conditions. J Rheumatol 1975, 2:8.

91. Salzman NH, Hung K, Haribhai D, Chu H, Karlsson-Sjöberg J, Amir E, Teggatz P,

Ba rman M, Hayward M, Eastwood D, Stoel M, Zhou Y, Sodergren E, Weinstock

GM, Bevins CL, Williams CB, Bos NA: Enteric defensins are essential regulators of intestinal microbial ecology. Nat Immunol 2010, 11:76-82.

92. Olszak T, An D, Zeissig S, Vera MP, Richter J, Franke A, Glickman JN, Siebert R,

Baron RM, Kasper DL, Blumberg RS: Microbial exposure during early life has persistent eff ects on natural killer T cell function. Science 2012,

336:489-493.

93. Shaw MH, Kamada N, Kim YG, Nunez G: Microbiota-induced IL-1beta, but not IL-6, is critical for the development of steady-state TH17 cells in the intestine. J Exp Med 2012, 209:251-258.

94. Atarashi K, Tanoue T, Shima T, Imaoka A, Kuwahara T, Momose Y, Cheng G,

Yamasaki S, Saito T, Ohba Y, Taniguchi T, Takeda K, Hori S, Ivanov II, Umesaki Y,

Itoh K, Honda K: Induction of colonic regulatory T cells by indigenous Clostridium species. Science 2011, 331:337-341.

95. Dicksved J, Halfvarson J, Rosenquist M, Järnerot G, Tysk C, Apajalahti J,

Engstrand L, Jansson JK: Molecular analysis of the gut microbiota of identical twins with Crohn’s disease. ISME J 2008, 2:716-727.

96. Spor A, Koren O, Ley R: Unravelling the eff ects of the environment and host genotype on the gut mic robiome. Nat Rev Micro 2011, 9:279-290.

97. Elinav E, Strowig T, Kau AL, Henao-Mejia J, Thaiss CA, Booth CJ, Peaper DR,

Bertin J, Eisenbarth SC , Gordon JI, Flavell RA: NLRP6 infl ammasome regulates colonic microbial ecology and risk for colitis. Cell 2011,

145:745-757.

98. Gao Z, Tseng C-h, Strober BE, Pei Z, Blaser MJ: Substantial alterations of the cutaneous bacterial biota in psoriatic lesions. PLoS ONE 2008, 3:e2719.

99. Grice EA, Segre JA: The skin microbiome. Nat Rev Micro 2011, 9:244-253.

100. Hitchon CA, Chandad F, Ferucci ED, Willemze A, Ioan-Facsinay A, van der

Woude D, Markland J, Robin son D, Elias B, Newkirk M, Toes RM, Huizinga TW,

El-Gabalawy HS: Antibodies to porphyromonas gingivalis are associated with anticitrullinated protein antibodies in patients with rheumatoid arthritis and their relatives. J Rheumatol 2010, 37:1105-1112.

101. Scher JU, Ubeda C, Equinda M, Khanin R, Buischi Y, Viale A, Lipuma L, Attur M,

Pillinger MH, Weiss mann G, Littman DR, Pamer EG, Bretz WA, Abramson SB:

Periodontal disease and the oral microbiota in new-onset rheumatoid arthritis. Arthritis Rheum 2012, 64:3083-3094.

doi:10.1186/ar4228Cite this article as: Costello ME, et al.: Microbes, the gut and ankylosing spondylitis. Arthritis Research & Therapy 2013, 15:214.


Page 10 of 10

Figure 1

Figure 2

ARTHRITIS & RHEUMATOLOGYVol. 67, No. 3, March 2015, pp 686–691DOI 10.1002/art.38967© 2015, American College of Rheumatology

BRIEF REPORT

Intestinal Dysbiosis in Ankylosing Spondylitis

Mary-Ellen Costello,1 Francesco Ciccia,2 Dana Willner,3 Nicole Warrington,1 Philip C. Robinson,1

Brooke Gardiner,1 Mhairi Marshall,1 Tony J. Kenna,1 Giovanni Triolo,2 and Matthew A. Brown1

Objective. Ankylosing spondylitis (AS) is a com-mon, highly heritable immune-mediated arthropathythat occurs in genetically susceptible individuals ex-posed to an unknown but likely ubiquitous environmen-tal trigger. There is a close relationship between the gutand spondyloarthritis, as exemplified in patients withreactivearthritis, inwhomatypicallyself-limitingarthro-pathy follows either a gastrointestinal or urogenitalinfection. Microbial involvement in AS has been sug-gested; however, no definitive link has been established.The aim of this study was to determine whether the gutin patients with AS carries a distinct microbial signa-ture compared with that in the gut of healthy controlsubjects.

Methods. Microbial profiles for terminal ileumbiopsy specimens obtained from patients with recent-onset tumor necrosis factor antagonist–naive AS andfrom healthy control subjects were generated usingculture-independent 16S ribosomal RNA gene sequenc-ing and analysis techniques.

Results. Our results showed that the terminalileum microbial communities in patients with AS differsignificantly (P < 0.001) from those in healthy controlsubjects, driven by a higher abundance of 5 families of

bacteria (Lachnospiraceae [P � 0.001], Ruminococcaceae[P � 0.012], Rikenellaceae [P � 0.004], Porphyromon-adaceae [P � 0.001], and Bacteroidaceae [P � 0.001])and a decrease in the abundance of 2 families ofbacteria (Veillonellaceae [P � 0.01] and Prevotellaceae[P � 0.004]).

Conclusion. We show evidence for a discretemicrobial signature in the terminal ileum of patientswith AS compared with healthy control subjects. Themicrobial composition was demonstrated to correlatewith disease status, and greater differences were ob-served between disease groups than within diseasegroups. These results are consistent with the hypothesisthat genes associated with AS act, at least in part,through effects on the gut microbiome.

Intestinal microbiome dysbiosis and microbialinfections have been implicated in several immune-mediated diseases, including multiple sclerosis, inflam-matory bowel disease (IBD), and type 1 diabetes.Bacterial infection in the gut and urogenital tract isknown to trigger episodes of reactive arthritis, a form ofspondyloarthropathy (SpA) belonging to a group ofrelated inflammatory arthropathies, of which ankylosingspondylitis (AS) is the prototypic disease. The closerelationship between the gut and SpA is exemplified inpatients with reactive arthritis, in whom a typicallyself-limiting arthropathy follows either gastrointestinalinfection with Campylobacter, Salmonella, Shigella, orYersinia, or urogenital infection with Chlamydia. Micro-bial involvement in AS has been suggested; however, nodefinitive link has been established (1).

Up to 70% of patients with AS have subclinicalgut inflammation, and 5–10% of these patients havemore severe intestinal inflammation that progresses toclinically defined IBD resembling Crohn’s disease (CD)(2). The high heritability of AS, the global diseasedistribution, and the absence of outbreaks of the diseasesuggest that AS is triggered by a common environmentalagent in genetically susceptible individuals (3). Multiple

Dr. Brown’s work was supported by a Senior PrincipalResearch Fellowship from the National Health and Medical ResearchCouncil of Australia.

1Mary-Ellen Costello, MSc, Nicole Warrington, PhD, PhilipC. Robinson, MBChB, FRACP, Brooke Gardiner, PhD, Mhairi Mar-shall, MSc, Tony J. Kenna, BSc, PhD, Matthew A. Brown, MBBS, MD,FRACP: University of Queensland Diamantina Institute, Transla-tional Research Institute, and Princess Alexandra Hospital, Woolloon-gabba, Brisbane, Queensland, Australia; 2Francesco Ciccia, MD, PhD,Giovanni Triolo, PhD: University of Palermo, Palermo, Italy; 3DanaWillner, PhD: Australian Centre for Ecogenomics and University ofQueensland, Brisbane, Queensland, Australia.

Drs. Ciccia, Willner, Kenna, and Triolo contributed equally tothis work.

Address correspondence to Matthew A. Brown, MBBS, MD,FRACP, The University of Queensland, The University of Queens-land Diamantina Institute, Translational Research Institute, PrincessAlexandra Hospital, Brisbane, Queensland 4102, Australia. E-mail:[email protected].

Submitted for publication August 28, 2014; accepted inrevised form November 18, 2014.

686

genes associated with AS also play a role in gut immu-nity, such as genes involved in the interleukin-23 (IL-23)pathway (4), which are important regulators of intestinal“health.” Marked overrepresentation of genes that areassociated with CD is also associated with AS (5); thissuggests that the 2 diseases may have similar etiologicmechanisms, possibly involving gut dysbiosis.

Several studies have shown that patients with ASand their first-degree relatives have increased intestinalpermeability relative to unrelated healthy control sub-jects, which, again, is consistent with a role for the gutmicrobiome in AS (6). To date, no comprehensivecharacterization of intestinal microbiota in patients withAS has been performed. We therefore performedculture-independent microbial community profiling ofterminal ileum biopsy specimens to characterize andinvestigate differences in the gut microbiome betweenpatients with AS and healthy control subjects.

PATIENTS AND METHODS

Biopsy specimens from the terminal ileum were ob-tained at the time of colonoscopy from 9 consecutively en-rolled tumor necrosis factor inhibitor–naive patients withrecent-onset (duration from symptom onset �48 months) AS(defined according to the modified New York classificationcriteria for AS [7]) and 9 healthy control subjects (Table 1)(see also Supplementary Table 1, available on the Arthritis &Rheumatology web site at http://onlinelibrary.wiley.com/doi/10.1002/art.38967/abstract). All patients gave written informedconsent, and the study protocol was approved by the relevant

University of Palermo and University of Queensland researchethics committees. One patient with AS was receiving non-steroidal antiinflammatory drugs (NSAIDs) at the time ofbiopsy. AS patients 3, 4, and 10 had reported occasional use ofNSAIDs, which was interrupted due to gastrointestinal upset.Other patients with AS did not report use of NSAIDs but werereceiving either acetaminophen and/or tramadol.

The terminal ileum microbial and mock communitieswere profiled by high-throughput amplicon sequencing of the16S ribosomal RNA (rRNA) gene (2� 250-bp barcode) on anIllumina MiSeq sequencer using dual-indexed v4-region for-ward primer 517F (5�-GCCAGCAGCCGCGGTAA-3�) andreverse primer 803R (5�-CTACCRGGGTATCTAATCC-3�).In addition to exploring community-level differences betweendisease states, we evaluated technical and biologic replication.Biologic replication was examined by halving the biopsy spec-imens and performing all subsequent studies in parallel toassess the reproducibility of findings from the time of biopsyforward. Technical replication was performed by analyzingforward and reverse sequencing reads separately and thencomparing them. The resulting sequence libraries were ana-lyzed using the Quantitative Insights Into Microbial Ecology(QIIME) pipeline (8).

Alpha diversity metrics, which are used to examine thediversity within a sample, were generated using the QIIMEworkflow. Hierarchical clustering was performed, using theunweighted pair group method with arithmetic mean (includ-ing both the weighted and unweighted UniFrac distance) todetect significant differences within microbial communitiesbetween patients with AS and healthy controls. The weightedUniFrac distance metric detects changes in the number oforganisms present in a community, taking into account therelative abundance of microbes present. The unweighted Uni-Frac distance metric describes community membership. Uni-Frac, enabled in QIIME, was used to generate sample distance

Table 1. Characteristics of the patients with AS and the healthy control subjects at the time of biopsy*

SubjectID

Age,years Sex

Disease duration,years†

ESR,mm/hour

CRP,mg/liter BASDAI

HLA–B27status

NSAIDtreatment

Histologicinflammation

AS01 56 M 7 34 18 7 Positive Current ChronicAS02 33 M 4 22 1 5.5 Positive None AcuteAS03 24 F 5 18 5 4.8 Positive None NormalAS04 22 M 3 33 3 5.5 Positive None NormalAS05 33 M 5 41 2.7 6 Positive None ChronicAS06 28 M 6 28 1.5 6.2 Positive None NormalAS08 36 M 8 34 1.5 5.6 Positive None AcuteAS09 38 F 6 45 2.2 6.7 Positive None ChronicAS10 31 M 11 27 1.3 8 Positive None NormalHC01 58 F – – 0.3 – – – –HC02 43 M – – 0.5 – – – –HC03 38 F – – 0.05 – – – –HC04 65 M – – 0.8 – – – –HC05 48 F – – 0.02 – – – –HC06 34 F – – 0.03 – – – –HC07 45 M – – 0.5 – – – –HC08 44 F – – 0.8 – – – –HC09 56 F – – 0.02 – – – –

* In all subjects, the ileum was the biopsy site. AS � ankylosing spondylitis; ESR � erythrocyte sedimentation rate; CRP � C-reactive protein;BASDAI � Bath AS Disease Activity Index; NSAID � nonsteroidal antiinflammatory drug; HC � healthy control.† Beginning at the onset of symptoms.

GUT MICROBIOME AND AS-RELATED GENES 687

metrics as well as principal coordinate analysis (PCoA). Coremicrobiome analysis as well as supervised learning were per-formed to further characterize the intestinal microbial signa-ture, using the workflow pipelines in QIIME (8). The functionof bacteria and communities was predicted using PICRUSt (9).

To determine differences in the microbial load, thetotal microbial biomass in biopsy samples was quantified usingreal-time quantitative polymerase chain reaction (PCR) ana-lysis of the 16S rRNA gene, as described by Willner et al (10).PERMANOVA analysis was conducted at the genus levelusing the vegan package in R to test the relationship betweenthe whole microbial community and disease. Indicator speciesanalysis was performed using the labdsv package in R. Co-occurrence network analysis was conducted using the spaapackage in R.

RESULTS

The microbial communities in the terminal ileumof patients with AS were significantly different (Figure1A) and more diverse (Figure 1C) compared with thosein healthy control subjects (P � 0.001), as determinedusing PERMANOVA. PCoA showed that AS samples,including biologic replicates, grouped separately fromcontrol samples, indicating that disease is the primaryfactor influencing the community differences (Figure

1B) (see also Supplementary Figure 1, available on theArthritis & Rheumatology web site at http://onlinelibrary.wiley.com/doi/10.1002/art.38967/abstract).

To further demonstrate the distinct groupings ofAS samples and control samples, supervised learningwas conducted to test the predictive capacity of themicrobiome differences, and all 9 samples obtained frompatients with AS had been predicted to be AS samples(Figure 1B). PERMANOVA analysis further confirmedthe presence of a significant relationship between dis-ease status and microbial community composition (P �0.001). This was not due to heterogeneity in biopsysamples, because biologic replicates showed no signifi-cant differences between each other. Quantitative PCRanalysis of biomass showed no significant differences, onaverage, in the 16S rRNA copy number between ASsamples and control samples, indicating that the ob-served differences were not attributable to overgrowthor dominance of bacteria driving community differences(see Supplementary Figure 2, available on the Arthritis &Rheumatology web site at http://onlinelibrary.wiley.com/doi/10.1002/art.38967/abstract). The average UniFracdistances demonstrated that differences between both

A

NS

******

***

0.0

0.1

0.2

0.3

0.4

Uni

Fra

c D

ista

nce

Technical replicates

Biologicalreplicates

Within disease status

Between disease status

DC

AS=9 HC=9

Obs

erve

d S

peci

es

BacteroidesAcinetobacteria

Firmicutes

Proteobacteria

HC AS AS

Fusobacteria

Supervised Learning

True/Predicted AS HCAS 9 0HC 0 9

ASHC

B

PC

2 - P

erce

nt v

aria

tion

expl

aine

d 11

.49%

PC1 - Percent variation explained 65.07%

Figure 1. A, Difference in taxonomy at the phylum level between patients with ankylosing spondylitis (AS) and healthy control subjects (HCs),showing the increase in Bacteroides and the change in the Bacteroides-to-Firmicutes ratio in the AS samples. B, Distinct clustering of AS samplescompared with control samples, as determined by weighted principal coordinate analysis (PCoA). Supervised learning showed that all 9 AS sampleswere predicted to be AS. C, Alpha diversity box plot showing increased microbial diversity (observed species) in AS samples compared with controlsamples. Data are presented as box plots, where the boxes represent the 25th to 75th percentiles, the lines within the boxes represent the median,and the lines outside the boxes represent the 10th and 90th percentiles. D, Comparison of variation within and between disease status, showinggreater significance between and within disease status than between technical replicates and biologic replicates. Values are the mean � SD.�� P � 0.001 by Mann-Whitney test. NS � not significant.

688 COSTELLO ET AL

biologic and technical replicates were significantly lessthan differences between individuals and disease states(P � 0.001, by Mann-Whitney test) (Figure 1D).

Community profiling showed 51 genera across allbiopsy specimens, with major differences observed at thephylum level between AS and control specimens (Figure1A). Indicator species analysis was performed to deter-mine whether alterations in specific species were drivingthe differences observed between communities in pa-tients with AS and healthy control subjects. This analysisshowed that, compared with the microbial communitiesin healthy control subjects, those in patients with ASwere characterized by a higher abundance of Lachno-spiraceae, including Coprococcus species and Roseburiaspecies (P � 0.001), Ruminococcaceae (P � 0.012),Rikenellaceae (P � 0.004), Porphyromonadaceae includ-ing Parabacteroides species (P � 0.001), and Bacte-roidaceae (P � 0.001). Decreases in the abundance ofVeillonellaceae (P � 0.01) and Prevotellaceae (P � 0.004)were observed (Figure 2A).

Further drilling down into the AS microbiomesignature using co-occurrence analysis, which examinesthe correlation between microbes, showed that bacterialinteractions further shape the AS microbial communitysignature, with positive correlations observed betweenthe indicator species Lachnospiraceae and Ruminococ-caceae and negative correlations observed between Veil-lonella and Prevotella (Figure 2B). Microbial functionalprediction using PICRUSt indicated 31 significant path-ways with differential representation in patients with AScompared with healthy controls (P � 0.02, with Bonfer-roni correction) (see Supplementary Table 2, availableon the Arthritis & Rheumatology web site at http://onlinelibrary.wiley.com/doi/10.1002/art.38967/abstract).These pathways included a decrease in bacterial invasionof epithelial cells (P � 4.33 � 10�5), an increase inantimicrobial production in the butirosin and neomycinbiosynthesis pathway (P � 0.002), which is consistentwith an increase in bacteria from the Bacteroidaceaefamily, and an increase in the secondary bile acidbiosynthesis pathway (P � 0.004), which is consistentwith an increase in Clostridia and Ruminococcaceaespecies.

We were interested in further interrogating theAS microbiome signature, which is the overall combina-tion of microbes that distinguishes patients with AS fromhealthy control subjects, by investigating whether anassemblage or “core” set of microbes was present in allAS and control samples at varying levels, and to probetheir functional capacity. Exploring the core microbiomeis imperative to better understand the stable and consis-

tent components across complex microbial communities,given the distinct AS microbial signature. The Core 100species (microbes present in all samples) included theClostridium, Actinomycetaceae, Bacteroidaceae, Lachno-spiraceae, Porphyromonadaceae, Rikenellaceae, Rumino-coccaceae, and Veillonellaceae families of bacteria. Thesefamilies of bacteria are also AS indicator species, sug-gesting that the core microbiome is driving the ASmicrobial signature (Figure 2A).

The Core 100 species belonged to 22 significantpathways (P � 0.02, with Bonferroni correction) (seeSupplementary Table 3, available on the Arthritis &Rheumatology web site at http://onlinelibrary.wiley.com/doi/10.1002/art.38967/abstract) that again included bac-terial invasion of epithelial cells (P � 0.004), antimicro-

Figure 2. A, Differences in the abundance of bacterial families be-tween patients with ankylosing spondylitis (AS) and healthy controlsubjects. The first 7 families of bacteria (from left to right) are ASindicator species, and the last 3 bacterial families are control indicatorspecies. B, Co-occurrence network showing positive (blue lines) andnegative (red lines) correlations between AS indicator species includ-ing members of the Lachnospiraceae family (Coprococcus and Rose-buria); Parabacteroides, which is a member of Porphyromonadaceae;and Faecalibacterium, which is a member of the Ruminococcaceaefamily. Increasing thickness of the lines indicates increasing strength of(Pearson’s) correlation.


bial production in the butirosin and neomycinbiosynthesis pathway (P � 0.007), and the secondary bileacid biosynthesis pathway (P � 0.013). Members of theBacteroides, Clostridium, and Ruminococcus groups areknown to be involved in cholesterol metabolism andsecondary bile synthesis; disordered bile acid synthesisdue to intestinal dysbiosis has previously been impli-cated in the pathogenesis of IBD (11).

As the threshold for core membership was re-duced to 90%, 75%, and 50%, the families of bacteriapresent remained the same; however, the number ofgenus and species within these families increased, espe-cially in Clostridia, Lachnospiraceae, and Ruminococcus.This demonstrates that the structure of the core micro-biome is robust, and that decreasing the thresholdexpands the number of genera and species of bacteriaonly within these families. As the core threshold de-creased, however, the number of significant pathwaysalso decreased, suggesting that as the threshold is re-laxed, the expansion of genera and operational taxo-nomic units dilutes pathway signals.

Indicator species analysis also determined thathealthy control subjects had an increased abundance ofStreptococcus and Actinomyces compared with patientswith AS. No difference was noted between patients withAS and control subjects in the presence of either bacte-ria known to be associated with reactive arthritis orKlebsiella species, which have been proposed to play arole in triggering AS (12).

DISCUSSION

Here, we present the first characterization andidentification of intestinal dysbiosis in the AS micro-biome, using 16S rRNA community profiling of terminalileum biopsy specimens. We show evidence for a discretemicrobial signature in the terminal ileum of patientswith AS compared with healthy control subjects. Themicrobial profile differences are not attributable todifferences in overall bacterial quantity between patientswith different diseases but are qualitative. PCoA wasable to show the distinct grouping of patients with ASversus healthy control subjects; larger studies will berequired to define the individual bacterial species in-volved. Statistically, the relationship between diseasestatus and microbial community composition was con-firmed, indicating that the differences in composition ofthe microbial community were not due to heterogeneityin the biopsy specimens or a lack of technical reproduc-ibility, but that the driving force is disease state.

Of the 7 families of microbes with differences in

abundance within the AS microbiome, Lachnospiraceae,Ruminococcaceae, and Prevotellaceae are strongly asso-ciated with colitis and CD, with Prevotellaceae especiallyknown to elicit a strong inflammatory response in thegut (13). Further investigations showed that correlationsbetween these families of bacteria further shape the ASmicrobial signature, and that these microbes were pres-ent in all AS samples studied, suggesting that they arenot only driving the microbial signature but also are atthe core of the AS microbial signature. Increases inPrevotellaceae and decreases in Rikenellaceae in theintestinal microbiome have also recently been observedin the HLA–B27–transgenic rat model of spondylo-arthritis, suggesting that underlying host genetics mayplay a role in sculpting the gut microbiome in this animalmodel (14).

The increased diversity of the AS communitywithout an overall change in microbial load shows thatno overgrowth or dominance of a particular microbedrives the signature. However, murine experiments dem-onstrate that both the overall composition of the intes-tinal microbiome and the presences and/or absence ofspecific microbes can have a substantial impact on hostresponse, regulation of inflammation, and developmentof intestinal cells. In the K/BxN mouse model of arthri-tis, it was shown that the introduction of a singlegut-residing species, segmented filamentous bacteria,into germ-free mice was sufficient to reinstate Th17cells, leading to the production of autoantibodies andarthritis. When IL-17 was neutralized in specificpathogen–free K/BxN mice, development of arthritiswas prevented. Thus, a single commensal microbe, via itsability to promote a specific T helper cell subset, candrive immune-mediated disease (15).

Further studies are needed to investigate whetherthe changes in intestinal microbial composition are dueto host genetics and how this affects the overall functionof the gut microbiome in AS patients, including how themicrobiome then goes on to shape the immune responseand influence inflammation. In particular, given thestrong association of HLA–B27 with AS, it has beenhypothesized that HLA–B27 induces AS by effects onthe gut microbiome, in turn driving spondyloarthritis-inducing immunologic processes such as IL-23 produc-tion (16). Our data showing intestinal dysbiosis in pa-tients with AS is consistent with this hypothesis, butfurther studies are clearly required to distinguish cause-and-effect interactions between the host genome andimmune system and the gut microbiome. These investi-gations will provide new insights and help us to betterunderstand the pathogenesis of AS.

690 COSTELLO ET AL

ACKNOWLEDGMENT

We thank the individuals who provided samples forthis study.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising itcritically for important intellectual content, and all authors approvedthe final version to be published. Dr. Brown had full access to all of thedata in the study and takes responsibility for the integrity of the dataand the accuracy of the data analysis.Study conception and design. Costello, Ciccia, Willner, Kenna, Triolo,Brown.Acquisition of data. Costello, Ciccia, Gardiner, Marshall, Triolo,Brown.Analysis and interpretation of data. Costello, Willner, Warrington,Robinson, Marshall, Kenna, Brown.

REFERENCES

1. Stebbings S, Munro K, Simon MA, Tannock G, Highton J,Harmsen H, et al. Comparison of the faecal microflora of patientswith ankylosing spondylitis and controls using molecular methodsof analysis. Rheumatology (Oxford) 2002;41:1395–401.

2. Mielants H, Veys EM, Cuvelier C, De Vos M, Botelberghe L.HLA-B27 related arthritis and bowel inflammation. Part 2: ileo-colonoscopy and bowel histology in patients with HLA-B27 relatedarthritis. J Rheumatol 1985;12:294–8.

3. Brown MA, Kennedy LG, MacGregor AJ, Darke C, Duncan E,Shatford JL, et al. Susceptibility to ankylosing spondylitis in twins:the role of genes, HLA, and the environment. Arthritis Rheum1997;40:1823–8.

4. Wellcome Trust Case Control Consortium, Australo-Anglo-American Spondylitis Consortium (TASC), Burton PR, ClaytonDG, Cardon LR, Craddock N, et al. Association scan of 14,500nonsynonymous SNPs in four diseases identifies autoimmunityvariants. Nat Genet 2007;39:1329–37.

5. Parkes M, Cortes A, van Heel DA, Brown MA. Genetic insightsinto common pathways and complex relationships among immune-mediated diseases. Nat Rev Genet 2013;14:661–73.

6. Mielants H, De Vos M, Goemaere S, Schelstraete K, Cuvelier C,Goethals K, et al. Intestinal mucosal permeability in inflammatoryrheumatic diseases. II. Role of disease. J Rheumatol 1991;18:394–400.

7. Moll JM, Wright V. New York clinical criteria for ankylosingspondylitis: a statistical evaluation. Ann Rheum Dis 1973;32:354–63.

8. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, BushmanFD, Costello EK, et al. QIIME allows analysis of high-throughputcommunity sequencing data. Nat Methods 2010;7:335–6.

9. Langille MG, Zaneveld J, Caporaso JG, McDonald D, Knights D,Reyes JA, et al. Predictive functional profiling of microbialcommunities using 16S rRNA marker gene sequences. Nat Bio-tech 2013;31:814–21.

10. Willner DL, Hugenholtz P, Yerkovich ST, Tan ME, Daly JN,Lachner N, et al. Reestablishment of recipient-associated micro-biota in the lung allograft is linked to reduced risk of bronchiolitisobliterans syndrome. Am J Respir Crit Care Med 2013;15:640–7.

11. Devkota S, Wang Y, Musch MW, Leone V, Fehlner-Peach H,Nadimpalli A, et al. Dietary-fat-induced taurocholic acid promotespathobiont expansion and colitis in Il102/2 mice. Nature 2012;487:104–8.

12. Ebringer R, Cooke D, Cawdell DR, Cowling P, Ebringer A.Ankylosing spondylitis: klebsiella and HL-A B27. RheumatolRehabil 1977;16:190–6.

13. Elinav E, Strowig T, Kau AL, Henao-Mejia J, Thaiss CA, BoothCJ, et al. NLRP6 inflammasome regulates colonic microbialecology and risk for colitis. Cell 2011;145:745–57.

14. Lin P, Bach M, Asquith M, Lee AY, Akileswaran L, Stauffer P, etal. HLA-B27 and human �2-microglobulin affect the gut micro-biota of transgenic rats. PLoS One 2014;9:e105684.

15. Wu HJ, Ivanov II, Darce J, Hattori K, Shima T, Umesaki Y, et al.Gut-residing segmented filamentous bacteria drive autoimmunearthritis via T helper 17 cells. Immunity 2010;32:815–27.

16. Kenna TJ, Brown MA. Immunopathogenesis of ankylosing spon-dylitis. Int J Clin Rheumatol 2013;8:265–74.


ORIGINAL ARTICLE

Does the microbiome play a causal role in spondyloarthritis?

James T. Rosenbaum & Phoebe Lin & Mark Asquith &

Mary-Ellen Costello & Tony J. Kenna &

Matthew A. Brown

Received: 31 January 2014 /Accepted: 24 March 2014 /Published online: 10 May 2014# Clinical Rheumatology 2014

Abstract The purpose of this study is to review the potentialcausal role of the microbiome in the pathogenesis ofspondyloarthritis. The method used for the study is literaturereview. The microbiome plays a major role in educating theimmune response. The microbiome is strongly implicated ininflammatory bowel disease which has clinical and geneticoverlap with spondyloarthritis. The microbiome also plays acausal role in bowel and joint disease in HLAB27/human beta2 microglobulin transgenic rats. The mechanism(s) by whichHLA B27 could influence the microbiome is unknown buttheories include an immune response gene selectivity, aneffect on dendritic cell function, or a mucosal immunodefi-ciency. Bacteria are strongly implicated in the pathogenesis ofspondyloarthritis. Studies to understand howHLAB27 affectsbacterial ecosystems should be encouraged.

Keywords Inflammatory bowel disease .Microbiome .

Spondyloarthritis

The microbiome is a term coined by the Nobel Laureate JoshuaLederberg to describe the collection of microorganisms that

cohabit our bodies [1]. Although this assortment of microbesincludes viruses, fungi, and parasites, bacteria are the predom-inant form of life. In fact, bacterial cells outnumber mamma-lian cells in the human body by about 10 to 1 [2]. BacterialRNA transcripts expressed in a human outnumber mammaliantranscripts by greater than 100 to 1 [3]. The National Institutesof Health has recognized the importance of characterizing themicrobiome through the initiation of the Human MicrobiomeProject [2, 3]. Europe is home to a similar initiative known atMeta Hit, the metagenomics of the human intestinal tract.Because the majority of these organisms are strictly anaerobic,they are difficult to culture. They can now be characterized bysequencing techniques which have diminished in cost dramat-ically over the past decade. These techniques exploit differ-ences in the ribosomal sequence of bacteria compared tomammals, allowing primers to be designed such that primarilybacterial DNA is amplified. Some characterization of themicrobiome has also been done bymass sequencing that relieson databases to sort the sequences into bacterial or other origin[4].

In this review, we begin with an overview of selected recentpublications that highlight the emerging recognition of themicrobiome’s role in health and disease. We conclude bydescribing observations that implicate the microbiome strong-ly in the pathogenesis of ankylosing spondylitis. The reader isalso referred to two previous reviews on the topic from ourgroups [5, 6].

One of the key functions of the microbiome is to educatethe immune response. Many authorities argue that the intesti-nal tract is the largest organ in the immune system.Whenmiceare raised in a strictly germfree environment, the lymphoidorgans do not develop normally and the immune response to aforeign antigenic challenge is blunted [7]. Our microbiomealso plays an essential role in the generation of vitamins suchas vitamin K. All food ingested is metabolized by themicrobiome which consequently has immense influence in

J. T. Rosenbaum (*) : P. Lin :M. AsquithDivision of Arthritis and Rheumatic Diseases and Department ofOphthalmology, Oregon Health & Science University, 3181 SWSamJackson Park Rd. L467Ad, Portland, OR 97239, USAe-mail: [email protected]

J. T. RosenbaumLegacy Devers Eye Institute, 1040 NW 22nd Ave, Portland,OR 97210, USA

M.<E. Costello : T. J. Kenna :M. A. BrownThe University of Queensland Diamantina Institute, PrincessAlexandra Hospital, University of Queensland, Brisbane 4102,Australia

Clin Rheumatol (2014) 33:763–767DOI 10.1007/s10067-014-2664-5

both starvation [8] and obesity [9]. Several dramatic examplesof drug metabolism or efficacy have also been shown to bedependent on the microbiome including acetaminophen [10],platinum-based chemotherapy [11], and cyclophosphamide[12].

The microbiome is now recognized to play a role in nu-merous diseases. For example, the microbiome generatesmetabolites called trimethylamine oxides which are highlyatherogenic [13, 14]. The vegetarian diet alters themicrobiome such that these metabolites are not generated[14]. Primarily based on data from rodent models, themicrobiome has been implicated in a vast array of diseaseswhich include inflammatory bowel disease [15], diabetes [16],hypertension [17], fatty liver [18], several cancers [19], ulcers[20], inflammatory arthritis [21], multiple sclerosis [22], andpsoriasis [23]. Even personality characteristics may be influ-enced by the microbiome [24, 25]. The two diseases that aremost clearly linked to the microbiome are ulcerative colitisand Crohn’s disease. A National Library of Medicine searchon the term inflammatory bowel disease and microbiome inDecember 2013 yielded 941 references. The evidence that themicrobiome is critical in inflammatory bowel disease primar-ily derives from rodents. In most but not all instances, themurinemodel of bowel inflammation ismarkedly improved ina germfree setting [15]. In some models, cohousing a mousewith bowel inflammation with a normal wild-type mouse willcause the germfree wild-type animal to develop aspects ofbowel inflammation such as in the example of the NLRP6knock-out mouse [15]. NLRP6 is expressed by intestinalepithelial cells and contributes to innate immunity. The ab-sence of NLRP6 results in colitis and an alteration of theintestinal microbiota.

Some of the evidence that ankylosing spondylitis may be amicrobiome-driven disease is based on the analogy betweeninflammatory bowel disease and ankylosing spondylitis.About 10 to 20% of patients with inflammatory bowel diseasedevelop sacroiliitis identical to ankylosing spondylitis [26].Conversely, more than 50 % of patients with ankylosingspondylitis have microscopic colonic lesions that resembleCrohn’s disease even if bowel symptoms are absent [27].Peripheral arthritis and uveitis are characteristic of both anky-losing spondylitis and inflammatory bowel disease.

A large number of genes associated with inflammatorybowel disease are also associated with ankylosing spondylitis.In nearly all cases, the genetic variant carries the same riskdirection for each disease [28]. A high proportion of thosegenes are involved in mucosal immunity [29]. For example,ankylosing spondylitis is associated with variants in the tran-scription factors RUNX3, EOMES, and TBX21, as well as thecytokine receptors IL7R and IL23R [29], all of which are keyregulators of the differentiation and activation of innate lymphoidcells, which are themselves critical components of mucosalimmune defenses.

As is the case for inflammatory bowel disease, the strongestargument that the microbiome is important in ankylosingspondylitis derives from animal data. Taurog, Hammer, andcolleagues derived transgenic rats which overexpress HLA-B27 and human beta 2 microglobulin [30]. These rats knownas 33–3 on the Fischer background develop a spontaneousillness characterized by diarrhea and subsequent axial andperipheral arthritis. When raised in a germfree environment,both the arthritis and diarrhea are largely eliminated [31].Furthermore, certain probiotics like Lactobacillus rhamnosusGG can maintain the remission [32], while colonization of thesusceptible rats with other bacteria such as Bacteroidesthetaiotamicron (B. theta) or Bacteroides vulgatuswill inducedisease. In the SKG mouse model, which has features ofspondyloarthritis and inflammatory bowel disease, and likeankylosing spondylitis shows marked involvement of IL-23dependent immunological pathways [33], mice raised ingermfree conditions are unaffected [34]. As with all animalmodels though, the relevance of these findings to healthand disease will require demonstration specifically inhumans.

A plausible hypothesis is that HLA molecules, either aloneor in combination with other ankylosing spondylitis-associated genes, determine the microbiome. No alleles withinthe genome are as polymorphic as HLA. A teleological argu-ment is that the diversity of HLA evolved to protect thesurvival of the race in case of epidemic infection. Accordingly,it is logical that the HLA type would affect the response tobacterial antigens such that different HLA types would deter-mine the differential survival of specific organisms. Someexperimental evidence to support this has been published inmice which are transgenic for the expression of the humanHLA type DR4 [35]. In HLA-B27 transgenic Lewis rats,which have a high penetrance of arthritis resembling humanankylosing spondylitis without overt colitis (the 21-3×283-2Lewis strain) [36], we have found key differences in the gutmicrobiota compared to wild-type control animals by both16S rDNA sequencing and short fragment mass sequencingtechniques (unpublished data by Phoebe Lin, Mark Asquith,Joel Taurog, Russ Van Gelder, and James Rosenbaum).

Several theories have been proposed as to how HLA-B27might predispose to disease [37]. These theories include thepossibility that B27 dictates an immune response to anautoantigen. While this theory fits with the concept thatHLA molecules are immune response gene products, noautoantigen has been identified to explain the pathogenesisof ankylosing spondylitis adequately. It has also been ob-served that B27 is unique in that it dimerizes on the cellsurface and thus can trigger a natural killer cell response.HLA-B27 within the cytosol is relatively unstable and cantrigger a series of intracellular events known as the unfoldedprotein response. None of these three observations adequatelyclarifies the clinical characteristics of the disease. In theory,

764 Clin Rheumatol (2014) 33:763–767

each hypothesis might be true and the mechanism by whicheach contributes to disease could be indirect through an alter-ation of the microbiome.

It has been proposed that HLA-B27 leads to a state ofmucosal immunodeficiency [38], and that genes associatedwith ankylosing spondylitis either contribute to that immuno-deficiency, or lead to heightened immunological reactionsdriven by bacterial invasion permitted by HLA-B27. Thismay either be direct (due to reduced ability to drive immuno-logical responses to particular bacteria), indirect due to effectson intestinal permeability, or due to alterations in the gutmicrobiome such as a loss of protective bacterial species.Given the strong interaction of HLA-B27 and ERAP1 incausing ankylosing spondylitis, and the known function ofERAP1 in preparing peptides prior to HLA Class I presenta-tion, it seems likely that these genes operate in disease throughaltered peptide handling. Two other aminopeptidases areknown to be associated with ankylosing spondylitis (ERAP2and NPEPPS). The disease-protective variants of ERAP1 andERAP2 are known to be loss-of-function variants, leading tolower HLA Class I cell surface expression. This would bemost consistent with models in which HLA-B27 in individ-uals carrying the risk, gain-of-function variants, alter themicrobiome in a manner which promotes ankylosingspondylitis-relevant immunological activation.

If HLA molecules shape the microbiome, multiple hypoth-eses could explain how this change could result in disease. Forexample, a heightened immune response to specific bacteriacould result in synovitis if these specific bacteria escape to thesynovium where they could trigger an immune response.Specific bacterial antigens have been detected in the joint inreactive arthritis [39, 40]. Another reasonable hypothesis isthat the change in microbiome changes gut permeability andallows leakage of multiple bacterial products which couldbecome arthritogenic. An increase in intestinal permeabilityhas been shown in the HLA-B27 transgenic rat [41] and inpatients with ankylosing spondylitis [42, 43]. Another expla-nation is that the presence of the HLA-B27 transgene alters thetranscriptional profiles of the bacteria present in the gut.Evidence supporting this concept is described in a study byHansen and colleagues, in which transcriptional profiles ofcecal B. theta were altered in transgenic animals resulting in abias towards bacterial persistence. They also found that bac-terial genes that encoded nutrient-binding proteins influentialin inducing adaptive immune responses were upregulated inthe transgenic rat [44].

One can surmise that an imbalance of the bacterial speciesfound in the gut can dramatically affect immune-mediateddisease since the specific components of the gut microbiomeare intricately tied to the host’s balance of immunologic ef-fector and regulatory function. Experimental data from micehave clearly shown that specific bacteria or microbial productsinduce regulatory T cells that express the transcription factor

FOXP3. Implicated bacteria include Clostridia spp. [45, 46],Helicobacter pylori [47], and Bacteroides fragilis [48]. Mu-rine studies have also implicated segmented filamentous bac-teria in the generation of T cells that synthesize interleukin 17[21]. Altered dendritic cell function has been noted inB27-related disease models [49]. Mucus from the gutalters dendritic cell function [50] so it is likely that themicrobiome will affect dendritic cell function directlyand indirectly.

Although it seems plausible that HLA-B27 shapes themicrobiome and that HLA-B27-related disease is associatedwith altered gut permeability, it is unknownwhich comes first:the change in permeability or the change in the microbiome.The alteration in the immune response could alter gut perme-ability or the alteration in the immune response could affectthe microbial population in the gut and this could result in achange in permeability. We hypothesize that altered gut per-meability allows the escape of microbial products which be-come the target of innate and adaptive immune responses.HLA-B27 might be critical in the response to the microbialproduct when it deposits in the synovium or uveal tract.Alternatively, B27’s role might be essential in creating theleaky gut but not critical in the immune response to theseproducts once they deposit in these tissues.

At the time of this presentation, several groups are activelytrying to confirm the preliminary observation that HLA-B27alters the microbiome either in rodents or in humans. Thecomplexity of the microbiome and the uncertainty as to whichmicrobiome population (for example, cecum, skin, or ileum),is most important increases the difficulty of this quest. What iscertain is that the microbiome is integral to human health andthe understanding of the microbiome will lead to novel in-sights into disease. Spondyloarthropathy seems highly likelyto be one such disease.

Acknowledgements We are grateful for the financial support from theNational Eye Institute (Grants EY 021733 and KO8EY022948), the Stanand Madelle Rosenfeld Family Trust, the William and Mary BaumanFoundation, and Research to Prevent Blindness. PL is funded by aResearch to Prevent Blindness Career development award. MAB isfunded by a National Health and Medical Research Council (Australia)Senior Principal Research Fellowship.

References

1. Lederberg J. (2001) ‘Ome Sweet’ Omics: a geneological treasury ofwords. The Scientist

2. Turnbaugh PJ, Ley RE, Hamady M, Fraser-Liggett CM, Knight R,Gordon JI (2007) The humanmicrobiome project. Nature 449(7164):804–10. doi:10.1038/nature06244, PubMed PMID: 17943116,PubMed Central PMCID: PMC3709439

3. Qin J, Li R, Raes J, Arumugam M, Burgdorf KS, Manichanh C et al(2010) A human gut microbial gene catalogue established bymetagenomic sequencing. Nature 464(7285):59–65. doi:10.1038/

Clin Rheumatol (2014) 33:763–767 765

http://dx.doi.org/10.1038/nature06244


nature08821, PubMed PMID: 20203603, PubMed Central PMCID:PMC3779803

4. Muthappan V, Lee AY, Lamprecht TL, Akileswaran L, Dintzis SM,Lee C et al (2011) Biome representational in silico karyotyping.Genome Res 21(4):626–33. doi:10.1101/gr.115758.110, PubMedPMID: 21324882; PubMed Central PMCID: PMCPMC3065710

5. Costello ME, Elewaut D, Kenna TJ, Brown MA (2013) Microbes,the gut and ankylosing spondylitis. Arthritis Res Ther 15(3):214. doi:10.1186/ar4228, PubMed PMID: 23750937

6. Rosenbaum JT, Davey MP (2011) Time for a gut check: evidence forthe hypothesis that HLA-B27 predisposes to ankylosing spondylitisby altering the microbiome. Arthritis Rheum 63(11):3195–8. doi:10.1002/art.30558, PubMed PMID: 21792826; PubMed CentralPMCID: PMCPMC3204318

7. Hill DA, Artis D (2010) Intestinal bacteria and the regulation ofimmune cell homeostasis. Annu Rev Immunol 28:623–67. doi:10.1146/annurev-immunol-030409-101330, PubMed PMID:20192812

8. Smith MI, Yatsunenko T, Manary MJ, Trehan I, Mkakosya R, ChengJ et al (2013) Gut microbiomes ofMalawian twin pairs discordant forkwashiorkor. Science 339(6119):548–54. doi:10.1126/science.1229000, PubMed PMID: 23363771, PubMed Central PMCID:PMC3667500

9. Vijay-Kumar M, Aitken JD, Carvalho FA, Cullender TC, Mwangi S,Srinivasan S et al (2010) Metabolic syndrome and altered gut micro-biota in mice lacking Toll-like receptor 5. Science 328(5975):228–31. doi:10.1126/science.1179721, PubMed PMID: 20203013

10. Clayton TA, Baker D, Lindon JC, Everett JR, Nicholson JK (2009)Pharmacometabonomic identification of a significant host-microbiome metabolic interaction affecting human drug metabolism.Proc Natl Acad Sci U S A 106(34):14728–33. doi:10.1073/pnas.0904489106, PubMed PMID: 19667173. PubMed Central PMCID:PMC2731842

11. Iida N, Dzutsev A, Stewart CA, Smith L, Bouladoux N, WeingartenRA et al (2013) Commensal bacteria control cancer response totherapy by modulating the tumor microenvironment. Science342(6161):967–70. doi:10.1126/science.1240527, PubMed PMID:24264989

12. Viaud S, Saccheri F, Mignot G, Yamazaki T, Daillere R, Hannani Det al (2013) The intestinal microbiota modulates the anticancer im-mune effects of cyclophosphamide. Science 342(6161):971–6. doi:10.1126/science.1240537, PubMed PMID: 24264990

13. Koeth RA, Wang Z, Levison BS, Buffa JA, Org E, Sheehy BT et al(2013) Intestinal microbiota metabolism of L-carnitine, a nutrient inred meat, promotes atherosclerosis. Nat Med 19(5):576–85. doi:10.1038/nm.3145, PubMed PMID: 23563705, PubMed CentralPMCID: PMC3650111

14. Tang WH, Wang Z, Levison BS, Koeth RA, Britt EB, Fu X et al(2013) Intestinal microbial metabolism of phosphatidylcholine andcardiovascular risk. N Engl J Med 368(17):1575–84. doi:10.1056/NEJMoa1109400, PubMed PMID: 23614584, PubMed CentralPMCID: PMC3701945

15. Elinav E, Strowig T, Kau AL, Henao-Mejia J, Thaiss CA, Booth CJet al (2011) NLRP6 inflammasome regulates colonic microbial ecol-ogy and risk for colitis. Cell 145(5):745–57. doi:10.1016/j.cell.2011.04.022, PubMed PMID: 21565393, PubMed Central PMCID:PMC3140910

16. Alkanani AK, Hara N, Lien E, Ir D, Kotter CV, Robertson CE et al(2013) Induction of diabetes in the RIP-B7.1 mouse model is criti-cally dependent on TLR3 and MyD88 pathways and is associatedwith alterations in the intestinal microbiome. Diabetes. doi:10.2337/db13-1007, PubMed PMID: 24353176

17. Pluznick JL, Protzko RJ, Gevorgyan H, Peterlin Z, Sipos A, Han Jet al (2013) Olfactory receptor responding to gut microbiota-derivedsignals plays a role in renin secretion and blood pressure regulation.Proc Natl Acad Sci U S A 110(11):4410–5. doi:10.1073/pnas.

1215927110, PubMed PMID: 23401498, PubMed Central PMCID:PMC3600440

18. Henao-Mejia J, Elinav E, Jin C, Hao L, Mehal WZ, Strowig T et al(2012) Inflammasome-mediated dysbiosis regulates progression ofNAFLD and obesity. Nature 482(7384):179–85. doi:10.1038/nature10809, PubMed PMID: 22297845, PubMed Central PMCID:PMC3276682

19. Kostic AD, Gevers D, Pedamallu CS, Michaud M, Duke F, Earl AMet al (2012) Genomic analysis identifies association of Fusobacteriumwith colorectal carcinoma. Genome Res 22(2):292–8. doi:10.1101/gr.126573.111, PubMed PMID: 22009990, PubMed CentralPMCID: PMC3266036

20. Enserink M (2005) Physiology or medicine. Triumph of the ulcer-bug theory. Science 310(5745):34–5. doi:10.1126/science.310.5745.34a, PubMed PMID: 16210507

21. Wu HJ, Ivanov II, Darce J, Hattori K, Shima T, Umesaki Y et al(2010) Gut-residing segmented filamentous bacteria drive autoim-mune arthritis via T helper 17 cells. Immunity 32(6):815–27. doi:10.1016/j.immuni.2010.06.001, PubMed PMID: 20620945, PubMedCentral PMCID: PMC2904693

22. Lee YK, Menezes JS, Umesaki Y, Mazmanian SK (2011)Proinflammatory T-cell responses to gut microbiota promote ex-perimental autoimmune encephalomyelitis. Proc Natl Acad Sci US A 108(Suppl 1):4615–22. doi:10.1073/pnas.1000082107,PubMed PMID: 20660719, PubMed Central PMCID:PMC3063590

23. Fry L, Baker BS, Powles AV, Fahlen A, Engstrand L (2013) Ischronic plaque psoriasis triggered by microbiota in the skin? Br JDermatol 169(1):47–52. doi:10.1111/bjd.12322, PubMed PMID:23521130

24. Douglas-Escobar M, Elliott E, Neu J (2013) Effect of intestinalmicrobial ecology on the developing brain. JAMA Pediatr 167(4):374–9. doi:10.1001/jamapediatrics.2013.497, PubMed PMID:23400224

25. Rosenbaum JT (2013). The E. Coli made me do it. The New Yorker(on-line)

26. Bluestone R, Morris RI, Metzger AL, Terasaki PI (1975) HL-AW27and the spondylitis of chronic inflammatory bowel disease and pso-riasis. Ann Rheum Dis 34:31–2

27. Mielants H, Veys EM, Cuvelier C, De Vos M, Botelberghe L (1985)HLA-B27 related arthritis and bowel inflammation. Part 2.Ileocolonoscopy and bowel histology in patients with HLA-B27related arthritis. J Rheumatol 12(2):294–8

28. Parkes M, Cortes A, van Heel DA, Brown MA (2013) Geneticinsights into common pathways and complex relationships amongimmune-mediated diseases. Nat Rev Genet 14(9):661–73. doi:10.1038/nrg3502, PubMed PMID: 23917628

29. International Genetics of Ankylosing Spondylitis C, Cortes A, HadlerJ, Pointon JP, Robinson PC, Karaderi T et al (2013) Identification ofmultiple risk variants for ankylosing spondylitis through high-densitygenotyping of immune-related loci. Nat Genet 45(7):730–8. doi:10.1038/ng.2667, PubMed PMID: 23749187; PubMed Central PMCID:PMC3757343

30. Hammer RE, Maika SD, Richardson JA, Tang JP, Taurog JD (1990)Spontaneous inflammatory disease in transgenic rats expressingHLA-B27 and human beta 2m: an animal model of HLA-B27-associated human disorders. Cell 63(5):1099–112, PubMed PMID:2257626

31. Taurog JD, Richardson JA, Croft JT, Simmons WA, Zhou M,Fernandez-Sueiro JL et al (1994) The germfree state prevents devel-opment of gut and joint inflammatory disease in HLA-B27 transgen-ic rats. J Exp Med 180(6):2359–64, PubMed PMID: 7964509,PubMed Central PMCID: PMC2191772

32. Dieleman LA, Goerres MS, Arends A, Sprengers D, Torrice C,Hoentjen F et al (2003) Lactobacillus GG prevents recurrence ofcolitis in HLA-B27 transgenic rats after antibiotic treatment. Gut

766 Clin Rheumatol (2014) 33:763–767


http://dx.doi.org/10.1101/gr.115758.110

http://dx.doi.org/10.1186/ar4228

http://dx.doi.org/10.1002/art.30558


http://dx.doi.org/10.1146/annurev-immunol-030409-101330

http://dx.doi.org/10.1126/science.1229000



http://dx.doi.org/10.1073/pnas.0904489106




http://dx.doi.org/10.1038/nm.3145

http://dx.doi.org/10.1038/nm.3145

http://dx.doi.org/10.1056/NEJMoa1109400

http://dx.doi.org/10.1056/NEJMoa1109400

http://dx.doi.org/10.1016/j.cell.2011.04.022

http://dx.doi.org/10.1016/j.cell.2011.04.022

http://dx.doi.org/10.2337/db13-1007

http://dx.doi.org/10.2337/db13-1007





http://dx.doi.org/10.1101/gr.126573.111

http://dx.doi.org/10.1101/gr.126573.111

http://dx.doi.org/10.1126/science.310.5745.34a

http://dx.doi.org/10.1126/science.310.5745.34a

http://dx.doi.org/10.1016/j.immuni.2010.06.001

http://dx.doi.org/10.1016/j.immuni.2010.06.001


http://dx.doi.org/10.1111/bjd.12322

http://dx.doi.org/10.1001/jamapediatrics.2013.497

http://dx.doi.org/10.1038/nrg3502

http://dx.doi.org/10.1038/nrg3502

http://dx.doi.org/10.1038/ng.2667

http://dx.doi.org/10.1038/ng.2667

52(3):370–6, PubMed PMID: 12584218, PubMed Central PMCID:PMC1773552

33. Ruutu M, Thomas G, Steck R, Degli-Esposti MA, Zinkernagel MS,Alexander K et al (2012) Beta-glucan triggers spondylarthritis andCrohn’s disease-like ileitis in SKG mice. Arthritis Rheum 64(7):2211–22. doi:10.1002/art.34423, PubMed PMID: 22328069

34. Hirota K, Hashimoto M, Yoshitomi H, Tanaka S, Nomura T,Yamaguchi T et al (2007) T cell self-reactivity forms a cytokinemilieu for spontaneous development of IL-17+ Th cells that causeautoimmune arthritis. J Exp Med 204(1):41–7. doi:10.1084/jem.20062259, PubMed PMID: 17227914, PubMed Central PMCID:PMC2118414

35. Gomez A, Luckey D, Yeoman CJ, Marietta EV, Berg Miller ME,Murray JA et al (2012) Loss of sex and age driven differences in thegut microbiome characterize arthritis-susceptible 0401 mice but notarthritis-resistant 0402 mice. PLoS One 7(4):e36095. doi:10.1371/journal.pone.0036095, PubMed PMID: 22553482; PubMed CentralPMCID: PMCPMC3338357

36. Fert I, Glatigny S, Poulain C, Satumtira N, Dorris ML, Taurog JDet al (2008) Correlation between dendritic cell functional defect andspondylarthritis phenotypes in HLA-B27/HUMAN beta2-microglobulin-transgenic rat lines. Arthritis Rheum 58(11):3425–9.doi:10.1002/art.24023, PubMed PMID: 18975325

37. Taurog JD (2010) The role of HLA-B27 in spondyloarthritis. JRheumatol 37(12):2606–16. doi:10.3899/jrheum.100889, PubMedPMID: 21123333

38. Kenna TJ, Brown MA (2013) Immunopathogenesis of ankylosingspondylitis. Int J Clin Rheumatol 8(2):265–74

39. Merilahti-Palo R, Soderstrom K-O, Lahesmaa-Rantala R, GranforsK, Toivanen A (1991) Bacterial antigens in synovial biopsy speci-mens in yersinia triggered reactive arthritis. Ann Rheum Dis 50:87–90

40. Granfors K, Jalkanen S, Lindberg AA, Maki-Ikola O, Von Essen R,Lahesmaa-Rantala R et al (1990) Salmonella lipopolysaccharide insynovial cells from patients with reactive arthritis. Lancet 335:685–8

41. SchepensMA, Schonewille AJ, Vink C, van Schothorst EM, KramerE, Hendriks T et al (2009) Supplemental calcium attenuates thecolitis-related increase in diarrhea, intestinal permeability, and extra-cellular matrix breakdown in HLA-B27 transgenic rats. J Nutr139(8):1525–33. doi:10.3945/jn.109.105205

42. Martinez-Gonzalez O, Cantero-Hinojosa J, Paule-Sastre P, Gomez-Magan JC, Salvatierra-Rios D (1994) Intestinal permeability in pa-tients with ankylosing spondylitis and their healthy relatives. Br JRheumatol 33(7):644–7, PubMed PMID: 8019793

43. Vaile JH, Meddings JB, Yacyshyn BR, Russell AS, MaksymowychWP (1999) Bowel permeability and CD45RO expression on circu-lating CD20+B cells in patients with ankylosing spondylitis and theirrelatives. J Rheumatol 26(1):128–35, PubMed PMID: 9918253

44. Hansen JJ, Huang Y, Peterson DA, Goeser L, Fan TJ, Chang EB et al(2012) The colitis-associated transcriptional profile of commensalBacteroides thetaiotaomicron enhances adaptive immune responsesto a bacterial antigen. PLoS ONE 7(8):e42645. doi:10.1371/journal.pone.0042645, PubMed PMID: 22880065, PubMed CentralPMCID: PMC3411805

45. Atarashi K, Tanoue T, Shima T, Imaoka A, Kuwahara T, Momose Yet al (2011) Induction of colonic regulatory T cells by indigenousClostridium species. Science 331(6015):337–41. doi:10.1126/science.1198469, PubMed PMID: 21205640

46. Arpaia N, Campbell C, Fan X, Dikiy S, van der Veeken J, Deroos Pet al (2013) Metabolites produced by commensal bacteria promoteperipheral regulatory T-cell generation. Nature. doi:10.1038/nature12726, PubMed PMID: 24226773

47. Arnold IC, Dehzad N, Reuter S, Martin H, Becher B, Taube C et al(2011) Helicobacter pylori infection prevents allergic asthma inmouse models through the induction of regulatory T cells. J ClinInvest 121(8):3088–93. doi:10.1172/JCI45041, PubMed PMID:21737881. PubMed Central PMCID: PMC3148731

48. Round JL, Mazmanian SK (2010) Inducible Foxp3+ regulatory T-cell development by a commensal bacterium of the intestinal micro-biota. Proc Natl Acad Sci U SA 107(27):12204–9. doi:10.1073/pnas.0909122107, PubMed PMID: 20566854. PubMed Central PMCID:PMC2901479

49. SangesM, Valente G, ReaM, Della Gatta R, De Franchis G, SollazzoR et al (2009) Probiotics in spondyloarthropathy associated withulcerative colitis: a pilot study. European review for medical andpharmacological. Sciences 13(3):233–4, PubMed PMID: 19673175

50. Shan M, Gentile M, Yeiser JR, Walland AC, Bornstein VU, Chen Ket al (2013) Mucus enhances gut homeostasis and oral tolerance bydelivering immunoregulatory signals. Science 342(6157):447–53.doi:10.1126/science.1237910, PubMed PMID: 24072822

Clin Rheumatol (2014) 33:763–767 767


http://dx.doi.org/10.1084/jem.20062259

http://dx.doi.org/10.1084/jem.20062259

http://dx.doi.org/10.1371/journal.pone.0036095



http://dx.doi.org/10.3899/jrheum.100889

http://dx.doi.org/10.3945/jn.109.105205







http://dx.doi.org/10.1172/JCI45041




A thesis submitted for the degree of Doctor of …368011/s...Publications included in this thesis...

Documents

Transcript of A thesis submitted for the degree of Doctor of …368011/s...Publications included in this thesis...