2016 09-08 Copenhagen Bioscience Llecture, Alain van Gool

Translating molecular biomarkers to novel diagnostics and impact in personalized health(care)

Prof Alain van Gool

Copenhagen Bioscience Lecture 8 September 2016

My path 1989-now

• Molecular biology

• Mechanisms of disease

• Biomarkers

• Omics / technologies

• Translational medicine

• Personalized healthcare

Senior Scientist Integrator Biomarkers

Scientific lead DTL-Technologies

Head EATRIS Biomarker Platform

Professor of Personalized Healthcare Head Radboud Center for Proteomics, Glycomics & Metabolomics Coordinator Radboud Technology Centers

2 Alain van Gool, Copenhagen Bioscience Lecture, 8 Sept 2016

https://www.google.nl/url?sa=i&rct=j&q=&esrc=s&frm=1&source=images&cd=&cad=rja&uact=8&ved=0ahUKEwiLsvL0--jMAhWKWhQKHUFEBWMQjRwIBw&url=https://en.wikipedia.org/wiki/Leiden_University&bvm=bv.122448493,d.ZGg&psig=AFQjCNG2nD-SOcwwEdKmo9jecYe4fqRj-w&ust=1463844511999998

Consider individual differences in life science research

3


Source: Chakma, Journal of Young Investigators, 16, 2009

Principle of Personalized Medicine

4

• The right drug for right patient at right dose at right time • Molecular biomarkers as key drivers of patient selection • = Precision medicine or Targeted medicine


Example: Personalized medicine in melanoma

B-RAFV600E mutation Strong growth of cell Growth of tumor

• B-RAFV600E cells always grow and become cancer cells

• RAF inhibitors will block pathway, block cell growth and inhibit cancers that have a B-RAFV600E mutation

• 60% of melanoma patients have B-RAFV600E mutation

• Basis for a personalized medicine !


Personalized medicine in melanoma

Treat patients with

B-RAFV600E mutation Inhibit growth of cell

Patients live longer Tumors disappear Cells stop growing

B-RAF inhibitor


Emerging Personalized / Precision / Targeted Medicine

2010:

5% of drugs in pipeline had companion diagnostic biomarker test

2015:

80%

50%


Moving to personalized health(care)

{Source: Barabási 2007 NEJM 357; 4}

• People are more than linear pathways • Different systems and networks • Different risk factors • Different preferences


Societal need in efficient personalized health(care)

{Source: prof Jan Kremer}

Towards cost effective care, less cure


Highest need in efficient personalized health(care)

It’s personal !

‘I want to stay healthy.’ ‘If not, how do I get healthy?’


3 key aspects of personalized health(care)

‘I want to stay healthy. If not, how do I get healthy?’

1. What to measure?

2. How much can it change?

3. What should be the follow-up for me?



Translating Personalized Health(care) in society

We need a personalized data-driven GPS for health

• Monitor on background

• Alert when you are at risk

• Advice what to do


http://www.google.nl/url?sa=i&rct=j&q=&esrc=s&frm=1&source=images&cd=&cad=rja&uact=8&ved=0CAcQjRxqFQoTCJ6kuJXIvcYCFYFcFAodO2oP3A&url=http://arstechnica.com/apple/2014/03/ios-in-the-car-becomes-carplay-coming-to-select-dashboards-this-year/&ei=TsSVVZ7FO4G5UbvUveAN&bvm=bv.96952980,d.d24&psig=AFQjCNHK-5180CQ5Wim0GGTOe8N0tIZRIQ&ust=1435964869551370



1. What to measure?





15

New data (generators, owners)

https://www.google.nl/url?sa=i&rct=j&q=&esrc=s&frm=1&source=images&cd=&cad=rja&uact=8&ved=0ahUKEwiCtoa38dTOAhXF2xoKHa-4AKIQjRwIBw&url=https://www.flickr.com/photos/philips_newscenter/albums/72157647098220647/&bvm=bv.129759880,d.d2s&psig=AFQjCNFvo2ckn1lyFQaSdHRflHAPCqnyRA&ust=1471950479842655

What does my DNA tell me?

23% chance blond hair 3.1% Neanderthaler DNA


Genetic risk lung cancer → don’t smoke !

What does my DNA tell me?

No expected adverse reaction to Warfarin


… but not all data is useful data !

Challenge: translate laboratory to society

• Heart beat • Steps / movement • Glasses water/coffee • 1.000.000 molecules per analysis


Exponential technological developments

• Next generation sequencing

• DNA, RNA • Risk analysis and therapy selection

• Mass spectrometry • Proteins, metabolites • Monitoring of disease and treatment effects

• Imaging

• Non invasive images, real time

• Spatial view of intact organs and organisms

500

1000

1500

2000

m/z

5 10 15 20 25 30 35 40 Time [min]


Principle: One genome → multiple proteomes

Body fluids

Tissues

Cells

Plasma Urine CSF

Lung Colon Adrenal gland

THP-1 Jurkat Granulosa cells



Conventional or “Bottom-Up” proteomics • Single proteins • Protein complexes • Cells • Tissues • Organelles • Membranes • Circulating vesicles • Body fluids

Example cellular proteome profiling

Sample: HEK293 whole cell proteome (1 µg tryptic digest of urea extract)

1D LC-M/MS bottom-up proteomics analysis

Retention time

m/z

400

600

800

1000

1200

1400

m/z

10 20 30 40 50 60 Time [min]

Blue: signal intensity in MS Pink dots: precursors selected for MS/MS

Detected peaks in MS spectra 1.584.599

Detected isotope patterns in MS spectra 130.172

Total number of MS/MS spectra 22.743

Av. Absolute Mass Deviation [ppm] 2,8972

Matched MS/MS spectra 5.603

Identified NR peptides 4.537

Identified proteins 1.321

False Discovery Rate 0,98%

In 1 scan:


Targeted proteomics for biomarker validation

• Select potential protein biomarkers • Select suitable peptides and fragment ions • Develop multiplex MRM assay • Measure (very complex) samples

Nature Methods: Method of the year 2012

Protein A isoform 1 Protein A isoform 2 Protein B


Innovation in protein biomarker diagnostics

Current diagnostic protein assays:

• Mostly protein abundance

• Often unknown epitope of detection

• Ignore occurence of proteoforms


Intact protein analysis

Bottom-up proteomics

Top-down proteomics



Intact protein analysis

Top-down proteomics

{Hans Wessels, Alain van Gool}

Top-Down proteomics: characterization of protein-ligand products

Measured vs simulated spectrum Chromatographic trace of Exendin-NODAGA

Annotated CID MS/MS spectrum of Exendin-NODAGA with 100% sequence coverage


Top-Down proteomics: analysis of in-vitro protein processing

HIS

-tag

re

mo

val b

y p

rote

ase


Top-Down proteomics: analysis of intact antibody

G0: GlcNAc2Man3GlcNAc2 G0F: GlcNAc2Man3GlcNAc2Fuc1 G1F: GalGlcNAc2Man3GlcNAc2Fuc1

G2F: Gal2GlcNAc2Man3GlcNAc2Fuc1

148.000 Da m/z !

Glycosylations


Top-Down proteomics: analysis of highly purified complex I

Sequence coverage of NIDM subunit

MS/MS sequence coverage: 85.7% MS sequence coverage: 100%

m/z z MASCOT Score Residues Fragmentation Modifications 841.2617 13 128 S2-K92 ETD Truncated Met, Acetyl: 1

16

29

19

7411 8860 18355 10810977 155137 17598 1406 1431411381138 13146 128 164146123 170149134 17851 120 187191

13

3923

18 70

64

61

4336

3466

27 4289

2149 93 99 173104

7231 83 1855838 91102 1111019679 161152117474 135 168121 125 15786 167158144 1627525 18815013082 18154 177103 115

CI f iltered Captive 3ul 05FA_Tray02-E1_01_1071.d

0.5

1.0

1.5

2.0

7x10

Intens.

10 20 30 40 50 Time [min]

Extracted Ion Chromatograms

NIDM subunit Mr 10.923,3198 Da Mass error: 0.0088Da (0.81 ppm)

• Unambiguous identification and quantitation of mature molecular forms of Complex I subunits • 42 subunits but >250 proteoforms detected ! • Insights into the combined dynamics of all available PTMs in a single experiment

(Complex I crystal-grade purified from Yarrowia lipolytica)


{Hans Wessels, Ulrich Brandt, Alain van Gool}

Top-Down proteomics: analysis of isolated membrane complexes

Chromatographic separation of complex V and identification of subunit ATP5J

Ragged N-terminus of ATP5J

Nat

ive

ge

l ele

ctro

ph

ore

sis


{Hans Wessels, Sergio Guerrero-Castillo, Ulrich Brandt, Alain van Gool}

Top-Down proteomics: map subunit-PTMs in specific complexes

Different proteoforms of subunit COX5A within complex IV N

ativ

e g

el e

lect

rop

ho

resi

s

Annotation Mr (Da) Relative occurrence

Proteoform ID by top down MS2?

Modification ID by bottom-up MS2?

N terminus ID by bottom-up MS2?

44-END phospho 12.509,3623 52,1 % Yes - Yes

44-END 12.429,3950 41,6 % Yes Yes

46-END phospho 12.285,2733 2,9 % - - -

43-END deamidated 12.592,4715 1,4 % - Deamidation N77 and N102

-

43-END 12.592,4715 1,1 % - -

44-END acetyl 12.471,3661 0,5 % - Acetyl N-term Acetyl K89

Yes

44-END acetyl+acetyl 12.513,3739 0,4% - Acetyl N-term Acetyl K89

Yes

(Sequence 100% proteoforms identified using top down and bottom-up MS and MS/MS) (Mass error top-down MS2 0,6-0,7 ppm)

• Accurate relative quantitation of all proteoforms of one subunit in complex IV • Much better insight as compared to:

• Bottom-up analysis of peptide fragments alone (missing protein data) • Analysis after enrichment for specific PTMs (eg phosphoproteomics)


Next: intact complexome proteins as new biomarkers?

• Native tissue biopsies

• Isolate intact membrane complexes

• Separate and isolate complexes using native gels

• LC-MS/MS analysis of intact proteins

• Data analysis

Tissue 1 (n=3)

Tissue 2 (n=3)

Subunit

Subunit – tissue 1

Subunit – tissue 2

• Identified protein sequence of subunit • Deduce simulated sequences from database • Determine fit with experimental data

35

{Hans Wessels, Sergio Guerrero-Castillo, Susanne Arnold, Ulrich Brandt, Alain van Gool}

Glycoproteomics workflow • Mass spectrometry analysis of glycoproteins in human plasma • 0,05 microliter analysis: detection of 1.000.000 signals in one scan (1,4 Gb) • ~40.000 peptides of which >80% contain sugar modification • Use to diagnose patients and identify new biomarkers

500

1000

1500

2000

m/z

5 10 15 20 25 30 35 40 Time [min]

Proof of principle study:

{Hans Wessels, Monique van Scherpenzeel, Dirk Lefeber, Alain van Gool} Biomarkers !?


New diagnostic glycoprotein biomarker

• Rare metabolic disease cases (liver disease and dilated cardiomyopathy)

• Combination glycoproteomics and exome sequencing

• Identification of deficient enzyme in glycosylation pathway

• Outcome 1: Explanation of disease

• Outcome 2: Dietary intervention as succesful personalized therapy

• Outcome 3: Glyco -transferrin profile developed as diagnostic mass spec test

{Monique van Scherpenzeel, Dirk Lefeber}




1. What to measure?



Personalized health(care) model Personalized Intervention

of patients-like-me Personal thresholds of persons-like-me

Big Biomarker Data

Molecular Non-molecular Environment …

Ho

meo

sta

sis

A

llo

sta

sis

D

isease

Time

Disease

Health

Selfmonitoring

Adapted from Jan van der Greef, TNO

Personal profile

Personalized health

Personalized medicine

39 Alain van Gool, Radboudumc Grand Rounds, 22 Aug 2016



1. What to measure?





Personal profile data

Knowledge

Understanding

Decision

Action


Biomarker innovation gaps !

Discovery Clinical

validation/confirmation

Diagnostic

test

Number of

biomarkers

Gap 1

Gap 2

Gap 3

• Too much biomarker discovery • Too little development to application


Biomarker innovation gaps: some numbers

5 biomarkers/ working day

1 biomarker/ 1-3 years

1 biomarker/ 3-10 years

?

Eg Biomarkers in time: Prostate cancer May 2011: n= 2,231 biomarkers Nov 2012: n= 6,562 biomarkers Oct 2013: n= 8,358 biomarkers Nov 2014: n= 10,350 biomarkers Oct 2015: n = 11,856 biomarkers

Discovery Clinical

validation/confirmation

Diagnostic

test

Number of

biomarkers

Gap 1

Gap 2

Gap 3


Reasons for biomarker innovation gap

• Not one integrated pipeline of biomarker R&D

• Publication pressure towards high impact papers

• Lack of interest and funding for confirmatory biomarker studies

• Hard to organize multi-lab studies

• Biology is complex on organism level

• Data cannot be reproduced

• Publishing bias towards extreme results

• Biomarker variability

• …

{Source: John Ioannidis, JAMA 2011}

{Source: Prinz, Schlange, Asadullah, Nat Rev Drug Disc 2011}

Slide from: Alain van Gool, Eur Commission advice, 11 Sept 2012


Irreproducibility of data

{Freedman et al, PLOS Biology, 2015}

{2012} {2011} {2013} {2008} {2012}


Categories of errors leading to irreproducibility



Add to this: bad Data Stewardship

{Wilkinson et al, Nature Scientific Data, 2016}

80% of data is not FAIR: Findable, Accessible, Interoperable, Reusable


Consequences downstream



Example: validation IL-8 as biomarker for melanoma For use in BRAF-MEK-ERK inhibitor drug development programs

Literature

{Yurkovetsky, et al. Clin Cancer Res, 2007}


Validation study to confirm IL-8 in melanoma

Stage 1 Stage 2 Stage 3 Stage 4

H&E staining; 20x

• 42 melanoma samples (tumor tissue + matching serum & plasma, stage I-IV, from two independent biobanks

• Genetic analysis for BRAFV600E/D mutation in genomic DNA from tissue samples

• IL-8 mRNA analysis in tissue samples by in situ hybridisation using bDNA probes (multiplexing with 12 ERK pathway response transcripts)

• IL-8 protein analysis in tissue samples by immunohistochemistry (in parallel with 4 other ERK pathway response proteins, Ki67, Tunnel)

• IL-8 protein analysis in matching plasma and serum by IL-8 immunoassay (3 formats: ELISA, Luminex, Mesoscale; singleplex and multiplex)

• Statistical data analysis

{Source: Alain van Gool, MSD, unpublished data 2010} 51

Alain van Gool, Copenhagen Bioscience Lecture, 8 Sept 2016

Example: validation IL-8 as biomarker for melanoma For use as efficacy biomarker in development BRAF inhibitor drugs

Literature

{Yurkovetsky, et al. Clin Cancer Res, 2007}

Own data

{Unpublished, 2010}

Cause?

(6 months, 4 fte, USD 1.000.000)


• Sample history ? • Tumor load?

Lessons learned?

Source: Youtube - Burn after reading ending}


Burn after reading ending.mp4

Way forward Quote Freedman paper:



Research Biomarkers Diagnostics

Department of Laboratory Medicine, Radboud univerity medical center Integrated Translational Research and Diagnostic Laboratory, 250 fte, yearly budget ~ 28M euro. Close interaction with Dept of Genetics, Pathology and Medical Microbiology

Specialities: • Proteomics, glycomics, metabolomics • Enzymatic assays • Neurochemistry • Cellulair immunotherapy • Immunomonitoring

Areas of disease: • Metabolic diseases • Mitochondrial diseases • Lysosomal /glycosylation disorders • Neuroscience • Nefrology • Iron metabolism • Pediatric oncology • Immunodeficiency • Transplantation

In development: • ~500 Biomarkers • Early and late stage • Analytical development • Clinical validation

Assay formats: • Immunoassay • Turbidicity assays • Flow cytometry • DNA sequencing • Mass spectrometry • Experimental human (-ized)

invitro and invivo models for inflammation and immunosuppression

Validated assays*: • ~ 1000 assays • 3.000.000 tests/year

Areas of application: • Personalized healthcare • Diagnosis • Prognosis • Mechanism of disease • Mechanism of drug action

Departmental community

*CCKL accreditation/RvA/EFI


www.radboudumc.nl/research/technologycenters 56

Genomics

Bioinformatics

Animal studies

Stem cells

Translational neuroscience

Image-guided treatment

Imaging

Microscopy

Biobank

Health economics

Mass Spectrometry

Radboudumc Technology

Centers

Investigational products

Clinical studies

EHR data analysis

Statistics

Human performance

Data stewardship

Molecule

Flow cytometry

3D lab

Institutional community

About 280 dedicated people working in 19 Technology Centers, ~1800 users (internal, external), ~150 consortia www.radboudumc.nl/research/technologycenters/ 57

• Proteins • Metabolites • Drugs • PK-PD • Preclinical

• Clinical

• Behavioural • Preclinical

• Animal facility • Systematic review

• Cell analysis • Sorting

• Pediatric • Adult • Phase 1, 2, 3, 4

• Vaccines • Pharmaceutics • Cyclotron • Radio-isotopes • Malaria parasites

• Management • Analysis • Sharing • Cloud computing

• DNA • RNA

• Internal • External

• Early HTA • Evidence-

based surgery • Field lab

• Statistics • Biological • Structural

• Preclinical • Clinical

• Economic viability

• Decision analysis

• Experimental design • Biostatistical advice

• Electronic Health Records

• Transmural translation

• Best practice

• In vivo • Functional

diagnostics

• iPSC • Organoids

• 3D imaging • 3D printing • Virtual reality

Regional communities

RADBOUD RESEARCH FACILITIES


Campus Radboud University

Regional Bioscience Parks

Regional Health Innovation Network

National communities

Funding of Large scale Scientific Infrastructures (>10M euro)

Data4LifeSciences (organising biomedical data)


National Technology Infrastructure (from 40+ partners in DTL)

National Science Agenda (originated from citizens)

European communities

@Radboudumc: Peggy Manders Gerhard Zielhuis

@Radboudumc: Peter Friedl Otto Boerman

@Radboudumc: Otto Boerman (Head Imaging Platform) Alain van Gool (Head Biomarker Platform)

@Radboudumc: Arnoud van der Maas Alain van Gool

@Radboudumc: Saskia de Wildt Paul Smits


Filling the gaps - 1

GAP 1 Public increasingly funding early development, funding gap remains

Public Funding Private Funding Moves away from early development

Fundamental discovery

Early Translational F.I.M

Phase 1 Phase 2 Phase 3

GAP 1


Filling the gaps - 2

GAP 2 Outputs of publicly funded projects misaligned with privately financed requirements

for further development

Public Funding Technology push, publications

Private Funding Product and patient focus, market pull

Fundamental discovery

Early Translational

F.I.M

Phase 1 Phase 2 Phase 3

GAP 2


Collaboration & services Matching client needs to capacities

Experts

Product Platforms QA & RA

PM & Clinical

Legal & Ethical

compliance

Training & Education Com & IT

Biomarkers Group

Vaccine Group

Tracer & Imaging Group

ATMP’s Group

Small Molecules

Group

Optimise translational trajectory

Remove barriers to multi-disciplinary collaboration Disease

expertise


Ongoing independent biomarker activities

Europe

USA

{Asadullah et al, Nature Reviews Drug Discovery, Dec 2015}


Build biomarker validation pipelines

Standardisation, harmonisation, knowledge sharing in:

1. Assay development

2. Clinical validation

NL Roadmap Molecular Diagnostics (2012) NL Grant 4.3M Eur (2014)

(Netherlands)

www.biomarkerdevelopmentcenter.nl


http://www.erasmusmc.nl/index.htm

The Good Biomarker Practice initiative

Join forces among Europe’s major academic infrastructures + industry to:

1. Establish “Good Biomarker Practice” guidelines

- on translational research, biomarker technologies, biobanking, data stewardship.

2. Efficiently execute high quality biomarker projects

- work together in clinical validation and development of probable biomarkers.


Netherlands: a culture of public-private partnerships

(Centre for Translational Molecular Medicine) (Top Institute Pharma)

(BioMolecular Materials)

2006-2015


Emerging Health Research Infrastructure in Netherlands


Health RI


Acknowledgements

Hans Wessels Jolein Gloerich

Roel Tans Esther Willems

Dirk Lefeber Monique van

Scherpenzeel

Leo Kluijtmans Ron Wevers

Marcel Verbeek Lucien Engelen

Jan Kremer Bas Bloem

Nathalie Bovy Paul Smits

the Radboudumc Technology Centers

and many others

www.radboudumc.nl/personalizedhealthcare

www.radboudumc.nl/research/technologycenters

www.radboudresearchfacilities.nl

[email protected]

[email protected]

www.linkedIn.com

www.slideshare.net/alainvangool

Many collaborators and funders

Jan van der Greef Ben van Ommen

Ivana Bobeldijk Hans Princen

Lars Verschuren Marjan van Erk

Suzan Wopereis Heleen Wortelboer

Wessel Kraaij Ronald Mooi

Peter van Dijken Cyrille Krul

and many others

CarTarDis


Ruben Kok Barend Mons

Jaap Heringa Merlijn van Rijswijk

and many others

Anton Ussi Florence Bietrix

Laura Bermejo Andreas Scherer

Sulev Koks Marian Hajduch

Giovanni Migliaccio

and many others

http://www.erasmusmc.nl/index.htm

2016 09-08 Copenhagen Bioscience Llecture, Alain van Gool

Health & Medicine

Transcript of 2016 09-08 Copenhagen Bioscience Llecture, Alain van Gool