7/13/19

1

Introduction:MetagenomicsinBiologyandMedicine

Builtenvironment

IslandsandbeachesCharlestonharbor

Residentialand

agriculturalareas

2

Wheremicrobiomescanbefound?

7/13/19

2

CharlestonWaterkeeper:monitorbacterialevelsatfifteenofthemostfrequentlyusedrecreationallocations aroundCharlestonsoyoucansplashsafely.

3

MedicalUniversityofSouthCarolina

4

7/13/19

3

Modernprecisionmedicinedecisionsupporttools

5

Soarethese!

6

7/13/19

4

Whynotthese?

7

Whynotthese?

+Needmassivedataonmicrobiomesatclinicallyrelevanttimepointslinkedtodetailedhealthinformation. 8

7/13/19

5

Majorchallenge:Silo-ing ofclinicalpracticeandresearch

clinicalpractice

pre-clinicalresearch

clinicaltrials

outcomesresearch

9

Majorchallenge:Silo-ing ofclinicalpracticeandresearch

clinicalpractice

pre-clinicalresearch

clinicaltrials

outcomesresearch

10

7/13/19

6

Majorchallenge:Silo-ing ofclinicalpracticeandresearch

clinicalpractice

pre-clinicalresearch

clinicaltrials

outcomesresearch

11

ORDER COLLECT TEST STORE DISCARD

Microbiology specimen lifecycle

12

7/13/19

7

LivingµBiomeBank

i2b2

Definecohort

SparcRequestCreate

specimencapturerequest

SparcRequest

Accept/Rejectspecimens

Receivedeidentifiedcohortandassaydata

SparcRequest

Receivecohortcapturerequest

Epic/RDW

Createstudyrecords

HonestBroker

Identifysurplusspecimens

ResearchNexus

Storespecimens

GenomicsPerform

sequencingassays

Bioinformatics

Processassaydata

RDW

Storedata

Specimencapturepathway

Datacapturepathway

Investigatorview

13

• Everyinpatientadmittedtothehospitalissubjecttotheprogram

• SwabsforMRSAandVREareobtainedassoonaspossibleafteradmission

• Samplesareprocessedwithin24-48hourstodeterminecolonization

• BDauto-streakinginstrumentisusedtominimizebias

• 75%ofthespecimenvolumeremainsinexcessandisdiscardedwithinaweekofcollection

• Thousandsarespecimenspassthroughthisprogrameverymonth

MUSCInfectionSurveillanceCultureProgram

14

7/13/19

8

SpecimensourcesforLµBB

LµBB

µBiome

PatientResistantculture

Selectionmediawithascertainedgrowth

15

Whyinfectionusesurveillancespecimens?

• Sampling uniformity• Limited number of nurses collect the swabs, minimizing collection

biases.• Sample handling is semi-automated, minimizing handling bias.• Participation reach• All patients are subject to the program.• ~15,000/year patients swabbed for MRSA.• ~7,500/year patients swabbed for VRE (select units, e.g. surgery,

etc.)• Impact potential• Specimens are collected early in the course of providing care.• Specimens are collected throughout care provision timeline.• Outcomes available in EHR for correlative research.

16

7/13/19

9

Whatisamicrobiomeafterall?Informallydefineit

17

OntologyforHost-MicrobiomeInteractions(OHMI)terms

UniversityofMichigan• Yongqun “Oliver”He

• Haihe Wang

• HongYu

• Jiahao Wang

UPenn• Jie Zheng

• DanielP.Beiting

Duke• AnnaMariaMasci

MUSC• JihadS.Obeid

• AlexanderV.

Alekseyenkohttps://github.com/OHMI-ontology/OHMI

7/13/19

10

Uniqueaspectsofmicrobiomeanalysis

• ExpressionAnalysis• Unitofanalysis:Transcript(anisoformofmRNAproducedfromagene)

• Measurement:Quantitative

• VariantAnalysis• Unitofanalysis:Variant(aversionofagenomicsegment)

• Measurement:numberofcopies,usually0,1,or2;or0vs1+

• MicrobiomeAnalysis• Unitofanalysis:Abundanceofvariantsofa16SrRNA geneampliconregion

• Measurement:Countofobservedsequencevariantsinthespecimen

• Analogy:variantanalysisinanomniploid organism

19

Additionalcaveats

• Sequencingerrors• Abigproblemwhenyouneedtoget100xcoverageofageneinadiploidorganism;

• Ahugeproblem,whenthecoverageis10,000xinanomniploid microbiome.

• Functionalmultiplicity• Manymicrobesfillthesamerole

• Compositionaleffects• Inferenceaboutabsolutequantityofmicrobesishardwithamplicondata

• Causalconsiderations

20

7/13/19

11

Whatistheroleofmicrobiomeinhumanhealth?

Wearemoremicrobesthanwearehumans?

• Humanshelter10trillionmicrobes(1013)intheirgutalone,(wearemadeof10trillioncells).

• Only1in10cellsinyourbodycarries‘your’

DNA.Recentevidencesuggestsasmanybacterialcellsashuman.

• Itisestimatedthatthereare1000speciesofbacterialivinginthehumangut.

• Comparealsothenumberofhumangenes(~25,000)tothenumberofgenesandvariantsthatbacterialcommunitiesmaycarry(~4,000,000,seee.g.doi:10.1038/ncomms3151).

22

7/13/19

12

Peter Jorth et al. mBio 2014; doi:10.1128/mBio.01012-14

Differential metabolic gene expression in the diseased periodontal microbiome.

Mechanismsforhost-microbeinteractions

23

Mechanismsforhost-microbeinteractions• Witheachother

• Viaregularecologicalmechanisms(competition)

• Withthehost/environment• Produceandmetabolizehormonesandcommonnutrients

• Hostimmunesystem

24

symbiosis

competition

e.g.parasitism

nutrients

Metabolites/ho

rmones

Immune

system/inflam

mation

7/13/19

13

Aconceptualroleofthemicrobiomeinhumandisease,aninfectiousdiseaseapproach• RobertKoch’s(1843- 1910)postulates:

1. Themicroorganismmustbefoundinabundanceinallorganismssufferingfromthedisease,butshouldnotbefoundinhealthyorganisms.

2. Themicroorganismmustbeisolatedfromadiseasedorganismandgrowninpureculture.

3. Theculturedmicroorganismshouldcausediseasewhenintroducedintoahealthyorganism.

4. Themicroorganismmustbereisolated fromtheinoculated,diseasedexperimentalhostandidentifiedasbeingidenticaltotheoriginalspecificcausativeagent.

• Dotheseapplytospecificmicrobiota?

…........…........…........…........

…........

1

2

4

3

25

Causalityviagerm-freeexperiments(postulate3)

Altering the Intestinal Microbiotaduring a Critical Developmental WindowHas Lasting Metabolic ConsequencesLaura M. Cox,1 ,2 Shingo Yamanishi,2 Jiho Sohn,2 Alexander V. Alekseyenko,2 ,3 Jacqueline M. Leung,1 Ilseung Cho,2

Sungheon G. Kim,4 Huilin Li,5 Zhan Gao,2 Douglas Mahana,1 Jorge G. Zarate Rodriguez,7 Arlin B. Rogers,6

Nicolas Robine,8 P’ng Loke,1 and Martin J. Blaser1 ,2 ,9 ,*1Department of Microbiology2Department of Medicine3Center for Health Informatics and Bioinformatics4Department of Radiology5Departments of Population Health (Biostatistics) and Environmental MedicineNYU Langone Medical Center, New York, NY 10016, USA6Department of Biomedical Sciences, Cummings School of Veterinary Medicine at Tufts University, North Grafton, MA 01536, USA7Department of Biology, NYU Abu Dhabi, Abu Dhabi, United Arab Emirates8New York Genome Center, New York, NY 10013, USA9New York Harbor Department of Veterans Affairs Medical Center, New York, NY 10010, USA*Correspondence: martin.blaser@nyumc.orghttp://dx.doi.org/10.1016/j.cell.2014.05.052

SUMMARY

Acquisition of the intestinal microbiota begins atbirth, and a stable microbial community developsfrom a succession of key organisms. Disruption ofthe microbiota during maturation by low-dose anti-biotic exposure can alter host metabolism andadiposity. We now show that low-dose penicillin(LDP), delivered from birth, induces metabolicalterations and affects ileal expression of genesinvolved in immunity. LDP that is limited to early lifetransiently perturbs the microbiota, which is suffi-cient to induce sustained effects on body composi-tion, indicating that microbiota interactions in infancymay be critical determinants of long-term host meta-bolic effects. In addition, LDP enhances the effect ofhigh-fat diet induced obesity. The growth promotionphenotype is transferrable to germ-free hosts byLDP-selected microbiota, showing that the alteredmicrobiota, not antibiotics per se, play a causalrole. These studies characterize important variablesin early-life microbe-host metabolic interaction andidentify several taxa consistently linked with meta-bolic alterations.

INTRODUCTION

Obesity, a complex disease, can increase the risk of diabetes,heart disease, and cancer (Vucenik and Stains, 2012). Alongwith dietary excess and genetic polymorphisms, the trillions ofmicrobial cells in the intestinal microbiota can contribute toobesity by increasing energy extraction (Turnbaugh et al.,

2006) or by altering metabolic signaling (Samuel et al., 2008)and inflammation (Cani et al., 2008; Henao-Mejia et al., 2012;Vijay-Kumar et al., 2010). Obese and lean humans differ inmicro-biota compositions, and these phenotypes can be transferred togerm-free mice (Ley et al., 2005; Ridaura et al., 2013), high-lighting the need for greater understanding of microbiota-hostmetabolic interactions.Because early life is a critical period for metabolic develop-

ment (Cunningham et al., 2014; Dietz, 1994; Knittle and Hirsch,1968), microbiota disruption during this window could lead tochanges in body composition. In humans, early-life microbiotadisruption, either due to delivery by Caesarian section (Domi-nguez-Bello et al., 2010) or antibiotics, is associated withincreased risk of overweight status later in childhood (Ajslevet al., 2011; Blustein et al., 2013; Huh et al., 2012; Murphyet al., 2013; Trasande et al., 2013), although in mice, antibioticexposure can increase or decrease weight, depending on thedose, diet used, or strain (Dubos et al., 1963).For decades, farmers have been exposing livestock to

low doses of antibiotics to promote growth; the earlier inlife that exposure begins, the more profound the effects (Crom-well, 2002; http://www.fda.gov/animalveterinary/products/approvedanimaldrugproducts/default.htm). We have appliedthe agricultural model of growth promotion to interrogate funda-mental host-microbe metabolic relationships. We previouslyshowed that early-life subtherapeutic antibiotic treatment in-creased fat mass, altered metabolic hormones, hepatic meta-bolism, and microbiota composition (Cho et al., 2012). Here,we extend these studies using low-dose penicillin (LDP) as amodel agent disrupting the microbiota. We examined whethertiming of LDP exposure is critical, whether synergies existbetween penicillin and dietary effects, and whether the alteredmicrobiota is sufficient to yield metabolic phenotypes. Our ex-periments answered each of these questions in the affirmative.

Cell 158, 705–721, August 14, 2014 ª2014 Elsevier Inc. 705

26

7/13/19

14

Microbiomeasamediatorinhumanhealth

Experimentalconditions• Healthyvs.disease• Exposure/Treatment

Microbiome

Changesinthehost• Geneexpression• Inflammatory

• Metabolic

• Immune,etc.

?27

See related Commentary on page 229

MUSCULOSKELETAL PATHOLOGY

Antibiotic Perturbation of Gut MicrobiotaDysregulates Osteoimmune Cross Talk inPostpubertal Skeletal DevelopmentJessica D. Hathaway-Schrader,*y Heidi M. Steinkamp,*z Michael B. Chavez,*x Nicole A. Poulides,*y Joy E. Kirkpatrick,*Michael E. Chew,* Emily Huang,* Alexander V. Alekseyenko,*{ Jose I. Aguirre,k and Chad M. Novince*y

From the Department of Oral Health Sciences,* Medical University of South Carolina College of Dental Medicine, Charleston, South Carolina; theEndocrinology Division,y Department of Pediatrics, and the Department of Public Health Sciences,{ Medical University of South Carolina College ofMedicine, Charleston, South Carolina; the Divisions of Pediatric Dentistryz and Biosciences,x The Ohio State University College of Dentistry, Columbus,Ohio; and the Department of Physiological Sciences,k College of Veterinary Medicine, University of Florida, Gainesville, Florida

Accepted for publicationOctober 16, 2018.

Address correspondence toChad M. Novince, D.D.S.,M.S.D., Ph.D., Department ofOral Health Sciences, MedicalUniversity of South CarolinaCollege of Dental Medicine,173 Ashley Ave., Charleston,SC 29425. E-mail: novincec@musc.edu.

Commensal gut microbiotaehost immune responses are experimentally delineated via gnotobiotic animalmodels or alternatively by antibiotic perturbation of gut microbiota. Osteoimmunology investigations ingerm-free mice, revealing that gut microbiota immunomodulatory actions critically regulate physiologicskeletal development, highlight that antibiotic perturbation of gut microbiota may dysregulate normalosteoimmunological processes. We investigated the impact of antibiotic disruption of gut microbiota onosteoimmune response effects in postpubertal skeletal development. Sex-matched C57BL/6T mice wereadministered broad-spectrum antibiotics or vehicle-control from the age of 6 to 12 weeks. Antibioticalterations in gut bacterial composition and skeletal morphology were sex dependent. Antibiotics did notinfluence osteoblastogenesis or endochondral bone formation, but notably enhanced osteoclastogenesis.Unchanged Tnf or Ccl3 expression in marrow and elevated tumor necrosis factor-a and chemokine (C-Cmotif) ligand 3 in serum indicated that the pro-osteoclastic effects of the antibiotics are driven byincreased systemic inflammation. Antibiotic-induced broad changes in adaptive and innate immune cellsin mesenteric lymph nodes and spleen demonstrated that the perturbation of gut microbiota drives a stateof dysbiotic hyperimmune response at secondary lymphoid tissues draining local gut and systemic cir-culation. Antibiotics up-regulated the myeloid-derived suppressor cells, immature myeloid progenitorcells known for immunosuppressive properties in pathophysiologic inflammatory conditions. Myeloid-derived suppressor cellemediated immunosuppression can be antigen specific. Therefore, antibiotic-induced broad suppression of major histocompatibility complex class II antigen presentation genes inbone marrow discerns that antibiotic perturbation of gut microbiota dysregulates critical osteoimmunecross talk. (Am J Pathol 2019, 189: 370e390; https://doi.org/10.1016/j.ajpath.2018.10.017)

The gut is colonized by diverse microorganisms (ie, bacteria,fungi, and viruses) that collectively form amicrobial communityknown as the gut microbiota. Early life host-microbe in-teractions direct the development of immunity and the estab-lishment of a stable complex microbial community, commonlyreferred to as the commensal microbiota.1e5 Extensive researchhas focused on commensal gut microbiotaehost immuneresponse effects in the context of protection against pathogenicgut microbes6e8 and the pathophysiology of chronic inflam-matory/autoimmune gastrointestinal disease states.9e11 Timely

investigations have delineated that indigenous gut microbiotaimmunomodulatory effects not only influence pathologic

Supported by the American Society for Bone and Mineral ResearchRising Star Award (C.M.N.); the NIH/National Institute of Dental andCraniofacial Research grants K08DE025337 (C.M.N.), R01DE021423(K.L.K.), T32DE017551 (Medical University of South Carolina), andR25DE022677 (Medical University of South Carolina); and NIH/NationalInstitute of General Medical Sciences grant P30GM103331 (MedicalUniversity of South Carolina).Disclosures: None declared.

Copyright ª 2019 American Society for Investigative Pathology. Published by Elsevier Inc. All rights reserved.https://doi.org/10.1016/j.ajpath.2018.10.017

ajp.amjpathol.org

The American Journal of Pathology, Vol. 189, No. 2, February 2019

28

7/13/19

15

weight (Figure 3D) and tibia length (Figure 3E) in antibiotic-versus vehicle-treated mice further support the notion thatantibiotic treatment did not induce harmful effects on normalsomatic growth processes.

To validate no changes in tibia length, growth-plate chon-drocyte zone morphology and height were assessed in prox-imal tibiae to evaluate alterations in endochondral boneformation (Figure 3F). Antibiotic- versus vehicle-treated micehad similar growth plate chondrocyte zonemorphology,whichlacked differences in proliferative zone height (Figure 3G),hypertrophic zone height (Figure 3H), and total growth plateheight (Figure 3I). The findings that antibiotic- versus vehicle-treated mice had similar growth plate morphology (Figure 3,FeI) and tibia length (Figure 3E) suggest that antibioticperturbation of gut microbiota does not disrupt endochondralbone formation.

Antibiotic Perturbation of Gut Microbiota Does NotAlter Osteoblastogenesis

Static histomorphometric analysis was performed on tolui-dine blueestained undecalcified tibia sections to investigateantibiotic effects on osteoblast cellular end points (Figure 4,

AeD). Antibiotic- and vehicle-treated mice had similarosteoblast numbers (Figure 4B), no differences in osteoblastperimeter per bone perimeter (Figure 4C), and no changes inosteoid per bone perimeter (Figure 4D).

Considering investigations in the C57BL/6 germ-freemouse model have shown that the normal gut microbiotacan blunt bone modeling/remodeling,52,55 dynamic boneformation indexes were analyzed (Figure 4, EeG) toelucidate whether antibiotic disruption of gut microbiotaalters mineral apposition and formation. Mineral apposi-tion rate (Figure 4F) and bone formation rate (Figure 4G)were not different in the proximal tibia of vehicle- andantibiotic-treated mice, which demonstrates that antibioticperturbation of gut microbiota does not alter osteoblastfunction.

To further corroborate antibiotic-induced lack of changesin osteoblastogenesis, alterations in Sp7 (osterix) (Figure 4H),a transcription factor essential for osteoblastic cell differen-tiation and maturation, and Bglap (OCN) (Figure 4, I and J), anoncollagenous matrix factor produced by mature osteo-blasts, were evaluated in marrow and calvaria. The rationalefor evaluating two distinct skeletal sites was to discern sys-temic osteoblast effects. In addition, calvaria has a more

Figure 1 Antibiotic (ABX) perturbation ofindigenous gut microbiota. A: Experimental time-line of the broad-spectrum antibiotic cocktail ofvancomycin (500 mg/L) targeting Gram (þ) bac-teria, imipenem/cilastatin (500 mg/L) targetingGram (þ), Gram ("), and anaerobes, and neomycin(1000 mg/L) targeting Gram (þ) and Gram (")bacteria. BeF: 16S rDNA analysis of fecal specimensfrom 12-weekeold mice. B and C: Universal 16Sgene analysis for bacterial load in males (B) andfemales (C). DeF: Phylum-level composition inmales and females, displayed in pie chart (D) andbar graph (E and F) format. Phylum-level real-timequantitative PCR analysis was normalized to theuniversal 16S gene expression. Unpaired t-test wasused. Data are expressed as means# SEM. nZ 5 pergroup (B and C). *P < 0.05, **P < 0.01, and***P < 0.001 versus vehicle. Tx, treatment; VEH,vehicle; wk, week.

Antibiotics Alter Osteoimmune Cross Talk

The American Journal of Pathology - ajp.amjpathol.org 375

homogeneous stromal-osteoblastic composition betterreflecting osteoblast-specific gene expression. Vehicle-versus antibiotic-treated mice had similar Sp7 (osterix)(Figure 4H) and Bglap (OCN) gene expression in femoralbone marrow and calvaria (Figure 4I). To further rule outsystemic alterations in osteoblastogenesis, intact OCN pro-tein levels were evaluated in serum (Figure 4J). Importantly,the serum enzyme-linked immunosorbent assay excludeddetection of OCN fragments derived from osteoclast-mediated bone resorption. Vehicle- and antibiotic-treatedmice had similar intact OCN serum levels (Figure 4J), vali-dating that antibiotic disruption of gut microbiota does notinfluence osteoblast-mediated bone formation.

Antibiotic Alteration of Gut Microbiota EnhancesOsteoclastogenesis and Local/SystemicProinflammatory Cytokines

Recognizing that bone modeling/remodeling occurs throughdual osteoclast-osteoblast actions, studies were performed toelucidate the impact of antibiotic perturbation of gutmicrobiota on osteoclastogenesis. Histomorphometric anal-ysis of TRAP-stained proximal tibia sections was performedto investigate antibiotic effects on osteoclastogenesis(Figure 5, AeD). Antibiotic- versus vehicle-treated micehad 2.5! greater osteoclast perimeter per bone perimeter(Figure 5D), which was attributed to a 1.75! increase in

Figure 2 Antibiotic (ABX) disruption of gut microbiota effects on bone mineral density (BMD) and trabecular bone morphology. Twelve-weekeold malevehicle (VEH)e and ABX-treated mice and female VEH- and ABX-treated mice were euthanized; specimens were harvested for analysis. AeF: Microecomputedtomographic (mCT) analyses of the proximal tibiae trabecular bone in male VEH- and ABX-treated mice. A: Representative reconstructed cross-sectional images,extending 360 mm distally from where analysis was initiated, in male VEH- and ABX-treated mice. B: Male trabecular BMD. C: Male trabecular bone volumefraction (BV/TV). D: Male trabecular number (Tb.N). E: Male trabecular thickness (Tb.Th). F: Male trabecular separation (Tb.Sp). GeL: mCT analyses of theproximal tibiae trabecular bone in female VEH- and ABX-treated mice. G: Representative reconstructed cross-sectional images, extending 360 mm distally fromwhere analysis was initiated, in female VEH- and ABX-treated mice. H: Female BMD. I: Female BV/TV. J: Female Tb.N. K: Female Tb.Th. L: Female Tb.Sp.Unpaired t-test was used. Data are expressed as means " SEM. n Z 4 per group (AeL). *P < 0.05 versus vehicle. cc, cubic centimeters.

Hathaway-Schrader et al

376 ajp.amjpathol.org - The American Journal of Pathology

BMD: Bone mineral density; BV/TV: trabecular bone volume fraction; Tb.N: trabecular number;Tb.Th: trabecular thickness;Tb.Sp: trabecular separation.

29

High-dimensionalhost-microbiomecharacteristicsPsycho-physiological assessmentTissues specific gene expression

Clinical and health record dataImmune cell populations

Epigenetic variationSomatic variantsGenetic variants

Nutrition

Bacterial, viral, and fungal composition and abundanceMetagenomic compositionPhylogeographic dataPhylogenetic dataTaxonomic dataMetaproteomeMetabolomeGlycome

HEALTHY?

30

7/13/19

16

‘New’waysto‘look’atmicrobiomes