1

MPE Transforms Pathology, Integrating Genomics, Microbiome, and Immunology

(21 June 2018, 30-35 min)

Shuji Ogino, MD, PhD, MS

Chief, Program in MPE Molecular Pathological Epidemiology, BWH

Professor (Pathology & Epidemiology)Brigham and Women’s Hospital (BWH)

Dana-Farber Cancer InstituteHarvard Medical School

Harvard T.H. Chan School of Public HealthAssociate Member, Broad Institute of MIT and Harvard

CRC = colorectal cancer

MPE = molecular pathological epidemiology

MSI = microsatellite instability(hypermutator, high neoantigen load)(Immune checkpoint inhibitor works for MSI-high solid tumors)

Epidemiology

One Field MPE1 + 1 → 100

Big Synergism

Molecular Pathology

5th

June 2020, Boston, USA

Free registrationOpen to public

www.mpemeeting.org

2

Tumorcells

Obesity EnvironmentDiet

Genetic variation

Immune cells

Microbiome

Ogino et al. Gut 2018Ogino & Giannakis. Lancet 2018

Drug

Tumorcells

Obesity EnvironmentDiet

Genetic variation

Immune cells

Microbiome

Drug

Pages et al. Lancet 2018(International Immunoscore Project led by Jerome Galon)

Ogino et al. Gut 2018Ogino & Giannakis. Lancet 2018

Tumorcells

Obesity EnvironmentDiet

Genetic variation

Immune cells

Microbiome

Drug

Complex problems necessitateexpertise in multiple fields

Ogino et al. Gut 2018Ogino & Giannakis. Lancet 2018

Tumorcells

Obesity EnvironmentDiet

Genetic variation

Immune cells

Microbiome

Drug

It is a big challenge to recapitulate these in experimental models → let’s get data from humans

Ogino et al. Gut 2018Ogino & Giannakis. Lancet 2018

3

Tumorcells

Obesity EnvironmentDiet

Genetic variation

Immune cells

Microbiome

Drug

Exposures can easily influence tumor - immune interactions

Ogino et al. Gut 2018Ogino & Giannakis. Lancet 2018

Pathology methods

Epidemiologymethods

Integrative MPE

Bioinformatics methods(machine / deep learning)

Immune response

Mutated peptides(neoantigens)

CD274 (PD-L1)expression

Tumor - Immune Interactions

Immune response FOXP3+ Treg

Masugi et al. Gut 2016

CD274 (PD-L1)

Mutated peptides(neoantigens)

Tumor - Immune Interactions

4

2+

Sur

viva

l pro

babi

lity

CRC-specific survival (years)

P < 0.0001

3+

0

1.0

0.8

0.6

0.4

0.2

0.0

1+

Immune cells in tumor microenvironment = best prognostic biomarker (why not clinical decision making?)

Multiplex immunofluorescence assays

CD3 (brown), CD4 (yellow), CD8 (purple), CD45RO (green), FOXP3 (orange), DAPI (nuclei, blue), cytokeratin (epithelial cells, red)

Borowsky, et al.

Immunology-MPE(immuno-MPE)

studies

Colon has rich microbiota & immune tissueKnown risk (or protective) factorsBest model for immunology-MPE

5

Nurses’ Health Study (121,700 women)

Health Professionals Follow-up Study (51,500 men)

Prospective Cohort Studies

1976

1986

Nurses’ Health Study(121,700 women)

Health Professionals Follow-up Study (51,500 men)

Diet, lifestyle, environment, genetics, plasma metabolomics, etc.

Nurses’ Health Study(121,700 women)

Health Professionals Follow-up Study (51,500 men)

Diet, lifestyle, environment, genetics, plasma metabolomics, etc.

1,600 colorectal cancers

1,000 polyps with tissue

Molecular pathology (whole exome sequencing,RNA-sequencing)Tissue microbiotaImmune response

Mortality

RNF43 mutations in 20% of CRC

Giannakis, et al. Nat Genet 2014

RNF43 inactivation (in MSI CRC)Or APC inactivation (in non-MSI CRC)→ activates WNT signaling

MSI non-MSI (≈ non-hypermutators)

6

Giannakis et al. Nat Genet 2014Giannakis et al. Cell Rep 2016

Significantly mutated genes(in our dataset)

Immunogenomic Studies• Whole exome sequencing of 1000 CRCs• Neoantigen load

– Likely a predictor for response to immunotherapy• WNT activation is immunosuppressive in CRC

High

Neoantigen Low Giannakis et al. Nat Genet 2014Giannakis et al. Cell Rep 2016Grasso et al. Cancer Discov 2018

Antigen processing machinery mutations

Giannakis et al. Cell Rep 2016

(% mutated in TIL-low/ % mutated in TIL-high)

Exposure

1,600 CRC cases

Normal colon

Subtype A

Subtype B

Mor

talit

y

Exposure

MPE analyses

170,000 persons

7

Exposure

1,600 CRC cases

Immune (or microbial) subtyping can be used

Normal colon

Subtype A

Subtype BM

orta

lity

Exposure

170,000 persons

Hypothesis: (1) Changes caused by the exposure

(2) Changes confer growth advantages

Preventive effect of colonoscopy differsby microsatellite instability (MSI) status

Nishihara et al. N Engl J Med 2013

Colonoscopy screening

MSI cancer

non-MSI cancerNormal colon

MPE can advance precision prevention(smokers have higher risk for MSI cancer)

Marine ω-3 polyunsaturated fatty acids(rich in fish oil)

• (?) reduce CRC risk• Marine ω-3 fatty acids can inhibit regulatory

T cells → stimulate effector T cells

Immune response

Mutated peptides(neoantigens)

Tumor cells

FOXP3+ Treg

8

Immune response

Mutated peptides(neoantigens)

Tumor cellsMarine ω-3fatty acids

FOXP3+ Treg

ω-3 fatty acids

CRC subtypes

Normal colon

FOXP3+ cell-high

FOXP3+ cell-low

0

0.5

1

1.5

2

Ref

Pheterogeneity = 0.006

Marine ω-3 fatty acid intake

Ptrend < 0.001

Ref

Mul

tivar

iate

Rel

ativ

e R

isk

FOXP3+ cell-low CRC FOXP3+ cell-high CRC

low → high low → high

Song et al. JAMA Oncol 2016

PUFA can reduce T-cell suppressive activity of FOXP3+ cell, leading to CD4+ T cell proliferation

Song et al. JAMA Oncol 2016

CD

4+ c

ell p

rolif

erat

ion

(%)

PUFA concentration

9

Immune response

Mutated peptides(neoantigens)

Tumor cellsMarine ω-3fatty acids

FOXP3+ Treg

Exposures (or risk factors) influence risk of CRC immune subtype

• Marine ω-3 PUFAs (Song. JAMA Oncol 2016)• IBD risk SNP (Khalili. Carcinogenesis 2015)• Vitamin D (Song. Gut 2016)• Aspirin (Cao. Gastroenterology 2016)• Inflammatory diet (Liu. Gastroenterology 2018)• Smoking (Hamada. JNCI in press)• Calcium intake

Colorectal cancer

PIK3CA mutated

PIK3CAwild-type

Dea

th

Aspirin

Liao et al. N Engl J Med 2012

Survival time (years)

Pro

babi

lity

of d

eath PIK3CA mutant PIK3CA wild-type

Non-user

Aspirin user

Non-user

Aspirin user

MPE can advance precision medicine(Aspirin may be used for PIK3CA-mutated CRC)

Liao et al. N Engl J Med 2012

P < 0.001

10

PIK3CA-mutated colon cancer cells are more sensitive to aspirin

Gu, et al. Oncotarget 2017(Similar data by Zumwalt et al. Cancer Prev Res 2017)

Aspirin for PIK3CA-mutated CRC

• Possible mechanisms– Aspirin down-regulates PI3K signaling– Immunity and inflammation– Tumor thrombosis and metastasis

Aspirin• Inhibits PTGS2 (cyclooxygenase-2)

→ reduces prostaglandins → enhances adaptive anti-tumor immunity

• Synergism of aspirin and immune checkpoint blockade

• Hypothesis: Immune checkpoint activation → resistance to aspirin

T cells

+ (-) Response to aspirin

CD274 (PD-L1) expressionlow high

11

Colorectal cancer (CRC)-specific survival (years)2 4 6 8 100

1.0

0.8

0.6

0.4

P < 0.001

Regular use of aspirin

Non-use of aspirin

CD274 (PD-L1)-low CRC

2 4 6 8 100

1.0

0.6

0.8

0.4

CD274 (PD-L1)-high(≈ immune checkpoint-

activated)

Immune checkpoint activation → resistance to aspirin P(interaction) < 0.001

Hamada et al. J Clin Oncol 2017

T cells

+ (-) +?Response to aspirin

CD274 (PD-L1) expressionlow high

Checkpoint inhibitor

Microbiology-MPE

Tumorcells

Obesity EnvironmentDiet

Genetic variation

Immune cells

Microbiome

Ogino et al. Nature Rev Clin Oncol 2011Ogino et al. Gut 2018Ogino et al. Lancet 2018

Drug

12

Fusobacterium nucleatum in CRC tissue Immunosuppressive → lower T cellsHigher stage → shorter survivalMSI-highHighest in cecal cancerGo with metastasis

Mima et al. JAMA Oncol 2015Mima et al. Gut 2016Mima et al. Clin Transl Gastro 2016Bullman et al. Science 2017

Prop

ortio

n of

cas

es (%

)

Fusobacterium nucleatum (F.n.) in 1000 CRCs100

80

20

10

0

60

P(trend) < 0.0001

High

Low

F.n.Negative

Mima et al. Clin Transl Gastro 2016

Diet Gut microbiota Cancer

Prudent diet(vegetables, whole grains, beans, fiber)

Good microbiota(↓ bad bacteria)

Fusobacteriumnucleatum(+) CRC

Pheterogeneity = 0.009

Ptrend = 0.003

Prudent diet score (quartiles)

F. nucleatum (-) CRC F. nucleatum (+) CRC

RefRef

Mul

tivar

iate

Rel

ativ

e R

isk

Mehta et al. JAMA Oncol 2017

low → high low → high

13

Prudent diet (vegetables, whole grains, beans)

CRC subtypes

Normal colon

F. nucleatum (+)

F. nucleatum (-)

Mehta et al. JAMA Oncol 2017

Prudent diet → microbiota → prevention

Inflammatory diet↓

Gut inflammation↓

Impaired mucosal barrier↓ anti-tumor immunity

↓F. nucleatum (+) colorectal cancer

Liu et al. Clin Gastro Hepatol 2018

Inflammatory diets

Red and processed meatRefined grainsSugar

(anti-inflammatory)TeaCoffeeVegetables (dark yellow; green leafy)

Causal inference - MPE

Statistical methodsin MPE

Immunology-MPEMicrobiology-MPE

Network scienceMPE

Molecular Pathological

Epidemiology(MPE)

Pharmaco-MPE

Etc.

Nutritional MPE

MPE can change our thought on scientific fields (expand, but not narrow down)

Ogino et al.

14

Immuno-OncologyImmuno-prevention

Summary• Immunology-MPE (Molecular Pathological

Epidemiology) can provide new insights– Exposures can easily influence microbiota–

tumor–immune interactions– Immunoprevention & therapy (using nutrients &

lifestyle)

• MPE is making new frontiers• There are widely open opportunities!

Summary 2• Let’s do “immuno-MPE” and

“microbiology-MPE” studies!

Acknowledgements (I cannot list everyone)• Yale Cancer Center

– Charles Fuchs• Dana-Farber Cancer Institute

– Marios Giannakis– Matthew Meyerson– Gordon Freeman– Catherine Wu– Jeffrey Meyerhardt– Levi Garraway

• Mass General Hospital– Andrew Chan– Mingyang Song

• Fred Hutchinson CRC– Ulrike Peters– Amanda Phipps

• UCLA / Parker Institute– Antoni Ribas– Catherine Grasso

• Harvard T.H. Chan School of Public Health / CDNM, Brigham and Women’s Hospital– Edward Giovannucci– Walter Willett – Lorelei Mucci– Meir Stampfer – Fran Grodstein– Tyler VanderWeele– Kana Wu– Molin Wang– Xuehong Zhang– Curtis Huttenhower– Wendy Garrett

• Brigham & Women’s (Pathology)– Scott Rodig– Jonathan Nowak– Reiko Nishihara

15

Acknowledgements #2• NHS (n=121,700), NHS II (n=116,600)

and HPFS (n=51,500) prospective cohort participants

• Various US hospitals/pathology departments for proving medical records and tissue specimens

Funding• NIH/NCI• Bennett Family Fund• Dana-Farber Harvard Cancer Center• Entertainment Industry Foundation

National Colorectal Cancer Research Alliance (NCCRA)

• Japan Society for Promotion of Science

• Japanese Society for Multidisciplinary Treatment of Cancer

• Uehara Memorial Foundation• Project P Fund• DFCI Friends

• The Ogino MPE Lab – Jonathan Nowak (faculty)– Jennifer Borowsky– Kota Arima– Ana Babic– Yang Chen– Annacarolina da Silva– Andressa Dias Costa– David Drew– Chunxia Du– Carino Gurjao– Kenji Fujiyoshi– Tsuyoshi Hamada– Koichiro Haruki– Xiaosheng He– Iny Jhun– Keisuke Kosumi– Mai Chan Lau– Peilong Li– Daniel Nevo– Rinda Soong– Tyler Twombly– Melissa Zhao