The Role of Microenvironment in Susceptibility to Age-related Breast

Cancers

Mark A LaBarge, PhD

City of Hope

For the Gayle Brinkenhoff Symposium

On November 14, 2017

1

Disclosures:I own stocks in Medicustech, and am a consultant for Thrive Biosciences. Issues pertaining to neither entity are being discussed by me.

2

CD227+

K19+ K8+

K14+ : K19+

1:1

Luminal

Myoepithelial

Stem cell Progenitor

Axl+cKit+CD10+K14+

Villasen etal JCB 2007Garbe etal Cancer Res 2012LaBarge and Lorens in reveiw

Contractile &Tumor suppressive

Secretory

Stem/Progenitors

3

4

The role of aging in breast cancer is a huge problemN

ew

ca

se

s

Pe

r 1

00

,00

0 w

om

en

Age at diagnosis

American Cancer Society. Jenkins etal The Oncologist 2014.

~80% diagnosed in

women aged >50

years

LUMINAL SUBTYPES

Age group

5

The role of aging in breast cancer is a huge problem

American Cancer Society, 2017

The State of Obesity, 2017

CDC predicts

increased

incidence of

post-meno BCs

due to trend of

>BMI CDC, 2015

CDC predicts a 21% increase in all cancer cases

by 2020 due to a larger proportion of aged

individuals.

Yabroff etal, Cancer Epidemiology Biomarkers & Prev 2011

Exp

en

ditu

res in

Bill

ions o

f U

SD

The economic burden of breast cancer

treatment to the US healthcare

economy may cost as much as $24B

by 2020.

Breast cancer incidence

6

Density (

density x

100=

%)

Age at diagnosis

Adapted from Matsuno etal, Can Epidem Biomarkers & Prev 2007

USA SEER (Caucasian in HI)

Japan (Osaka)

Japanese American (in HI)

1993-1997

This portion should be

preventable

Are age-related breast cancers preventable?

7

Low mRNA expression

High mRNA expression

Breast tumors

oldyoung

Genes

Individuals

Aging impacts transcriptomes but not genomes in hormone-

dependent breast cancers, and normal breast

Yau etal BrCanRes 2007

27 77AGE

Normal mammary epithelia

Genes

Individuals

Garbe etal CanRes 2012

Discontinuities in

Extra Cellular Matrix (ECM)

co

llag

en

fat

Age (years)

Age (years)

75

75

30

30

Changes in breast microenvironment

How bad can it get when normal architecture is disrupted?

Dolberg and Bissell, Nature, 1984

+Laminin Laminin Blockade

Tissue polarity

LaBarge & Cahjar etal

Barcellos-Hoff Can Res 2000

0cGy

40cGy

Laminin Collagen

ES

A

MU

C-1

Gudjonnson JCS 2002

Overt disruptions unleash malignancies

Wound healing

Irradiated stroma

Not so bad (normal cells) Really bad (w/ pre-existing mutations)

8

9

HER2+

Tneg

HR+ HER2-

Represents majority of age-related BC

Risk of recurrence in age-related breast cancers is protracted

Esserman etal and Benz, Br Can Res Trt 2011

Dis

ease

Sp

ecif

ic S

urv

ival

Time Since Diagnosis (years)

10

Aging raises questions aplenty…

What is the relationship between normal changes that occur with age and

increased susceptibility to breast cancers?

Why are there disproportionately more Luminal subtype BCs in older women?

How might age- (and cancer)-related tissue microenvironments impact

responses to cancer therapy?

A resource of normal, pre-stasis HMEC for functional interrogation of aging.

0

20

40

60

80

0 20 40 60 80 100 120

Po

pu

lati

on

Do

ub

lin

gs

Days in culture

Commercial #1WIT-P Commercial #2

MEGM

Stasis

Cancer Research 2012, JoVE 2013, Cell Rep 2014, PNAS 2011 & 2015, Aging 20172

Long-term growth of primary cells Multi-lineage maintenance

Se

lf-org

an

iza

tion

t = 2

4 h

rt =

96

hr

lum

en

ize

d

correct

inverted

ag

aro

se

ag

aro

se

Ma

trige

l

ABC

t = 2

4 h

r

Ma

trige

l

D

0.0

1.0

frequency

ag

aro

se

Ma

trige

l

I

II

III

IV

V

EF

LEP MEP

ECM ECM

LEP

MEP

LEP MEP

HG

0

50

100

150

contact angle

0

50

100

150

contact angle

ME

P

ME

P

LE

P

LE

P

ME

P

LE

P

ME

P

EC

M

LE

P

EC

M

ME

P

LE

PL

EP

ME

P

ME

P

LE

P

t = 0

K14 K

19

CT

R C

TG

Self-organization

t = 24 hr t = 96 hr

lumenized

co

rrect

inv

ert

ed

agarose

agarose

Matrigel

A

B

C

t = 24 hr

Matrigel

D

0.0

1.0

fre

qu

en

cy

agarose Matrigel

I II III IV V

E

F

LE

PM

EP

EC

ME

CM

LE

P

ME

P

LE

PM

EP

HG

0

50

100

150

co

nta

ct a

ng

le

0

50

100

150

co

nta

ct a

ng

le

MEP

MEP

LEP

LEP

MEP

LEP

MEP

ECM

LEP

ECM

MEP

LEP LEP

MEP

MEP

LEP

t = 0

K14 K19

CTR CTG

1h

24h

96h

Self-organizing

15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 950

2

4

6

8

10

Age of donor (years)

Num

ber

of H

ME

C s

train

s

2D culture

Progenitor activity in 3-D culture

240R Diversity Index over Passage

2 4 6 8 10 120.0

0.2

0.4

0.6

0.8

1.0M87A

WIT-P

MCDB170

Passage

Sh

an

no

n D

ive

rsity In

de

x (

ln)

208 Diversity Index over Passage

2 4 6 80.0

0.2

0.4

0.6

0.8M87A

WIT-P

MCDB170

Passage

Sh

an

no

n D

ive

rsity In

de

x (

ln)

Baseline

Total Area

Total Peak Area

Number of Peaks

Peak 1

First X=

Last X=

Peak X=

Peak Y=

Area=

%Area=

M87A

0.0

2.733

2.733

1.000

2.000

7.000

4.000

0.7052

2.733

100.0

WIT-P

0.0

1.369

1.369

1.000

2.000

7.000

2.000

0.6291

1.369

100.0

MCDB170

0.0

0.9688

0.9688

1.000

2.000

7.000

4.000

0.3103

0.9688

100.0

Baseline

Total Area

Total Peak Area

Number of Peaks

Peak 1

First X=

Last X=

Peak X=

Peak Y=

Area=

%Area=

M87A

0.0

3.330

3.330

1.000

2.000

11.00

4.000

0.8528

3.330

100.0

WIT-P

0.0

2.374

2.374

1.000

2.000

7.000

2.000

0.7009

2.374

100.0

MCDB170

0.0

0.4445

0.4445

1.000

2.000

7.000

2.000

0.3031

0.4445

100.0

LBNL

media#1

#2

Large, standardized batches of normal, pre-stasis HMEC from more than 70 reduction or mastectomy patients.

Keratin 14/19

Signs of accelerated aging in luminal cells from high-risk breast tissue

12

Vatter, Bodenmiller, LaBarge, and Lorens– in revision

Tumor, Peripheral to tumor, BRCA1 or ATM mut

No history

Young and middle-aged epithelia that were incorrectly predicted by machine learning to be “old” were from high-risk women

CYTOF analysis of 57 primary strains with 29 monoclonal antibodies

27 77AGE

Laser microdissected normal human mammary epithelium

Wiring diagrams (transcriptomes) change with age:What are the functional consequences?

Why are the aging phenotypes so stable?

Cancer Research 2012

Primary cultured normal human mammary epithelial cells

<30y

>55y

Stiffness E(Pa)

Did not differentiate in responseto stiffness changes

Stiffness E(Pa)

Mammary stem cells accumulate with age

because they lose sensitivity to

microenvironment differentiation directives

Distinct stiffness-dependent differentiation responsesFo

ld c

han

geFo

ld c

han

ge

Cancer Research 2012

Epithelia composition

Cell Reports 2014

YAPP

MST2

Makes Myoeps

Aging alters the trigger points of proteins used by cells to communicate mechanical information into the nucleus

,which initiate differentiation programs

In young progenitor cells…MST2 levels

200Pa 2350Pa

Cell Reports 2014

YAPP

MST2

Makes basal-like Luminals!!

But in older progenitor cells…

MST2

MST2

MST2

MST2

MST2

200Pa 2350Pa

MST2 levels

Cell Reports 2014

Aging alters the trigger points of proteins used by cells to communicate mechanical information into the nucleus

,which initiate differentiation programs

Hypothesis: As tissue ages the new microenvironments establish a continuum of metastable epigenetic states..

Oyer… , and Turker PlosONE 2009

Len

gth

of

tim

e re

pre

sse

d

Length of time repressed = likelihood of promoter methylation

AGING

17

Luminal cell DNA genome-wide methylation patterns cluster according to chronological age

18Analysis of primary HMEC with Infinium450K arrays

Older Younger

Sayaman and Miyano

Transcriptionally and epigenetically epithelial cells lose tissue specificity with age

Miyano etal, Aging 2017

Maintenance of luminal (LEP) cells requires the correct microenvironment.

Rat

io o

f LE

P c

om

par

ed t

o d

ay 4

Miyano etal, Aging 2017

measure

Aged microenvironments drive loss of lineage fidelity that are made metastable by epigenetic regulatory states

Luminal

Myoepithelial

YOUNGMEP

YOUNGLEP

LEP specific genesPromoter

=young gene expression patterns

LEP specific genesPromoter

5mC 5mC 5mC5mC

OLDMEP

=OLD gene expression patterns

Young-LEP/Old-MEP

Young-LEP/

Young-MEP

You

ng LEP

-1o

You

ng LEP

-1o

You

ng LEP

-1o

You

ng LEP

-1o

You

ng LEP

-1o

Old

-LEP-1

o

Old

-LEP-1

o

Old

-LEP-1

o

Old

-LEP-1

o

Gen

e ex

pre

ssio

n

Miyano etal, Aging 2017

Promoter DNA methylation changes in key luminal-specific genes driven by

age of myoepithelial cells

22

Y / Y

Y / O

0

20

40

60

80

100

*

*

% o

f ce

lls w

ith

met

hyl

ated

DN

A in

exa

min

ed

regi

on

<30y vs >55y in LEPM-value Differential Methylation

* BH adj. p-val < 0.05

** BH adj. p-val < 0.01

*** BH adj. p-val < 0.001

DN

A M

eth

ylat

ion

Bet

a V

alu

es

DN

A M

eth

ylat

ion

Bet

a V

alu

es

y/yy/o

Primary LEP

Primary LEP in Co-culture with MEP

Miyano etal, Aging 2017

23

Blockade of cell-cell communication proteins with higher and more variable expression with ageprevents transmission of the aged phenotype into luminal cells by older myoeps.

Chemical blockade Cx30 knock down in MEP

Miyano, unpublished

Immortalization re-sensitized YAP (and TAZ) to the physiological elastic range

24

66 years 91 years

Log2

fo

ld c

han

geFr

om

20

0P

a

-2

-1

0

1

2

200 Pa 1500 Pa 2400 Pa

LEP

MEP

Differentiation

122LMY 805Pp16S

Cell Reports 2014

Epithelia loses specificity with age due largely to epigenetic changes, what are the consequences for transformation?

25

Cancer Res 2012, Cell Reports 2014

32y

67y

K14K19nuclei

% o

f ce

lls

34y 54y

LEP

MEP LE

P

MEP

(N)uclear

(C)ytoplasmic

(N/C) nuc & cyto

Two-hit immortalization does not introduce gross genomic errors in HMEC

26

AGE (y)

19

21

91

66

Cell Cycle 2014

Array comparative genomic hybridization

240Lp16shMY_25p

184Dp16shMY_30p

805Pp16shMY_10p

122Lp16sMY_12p

Immortalization re-sensitizes aged progenitors to a physiological range of stiffness

27

66 years 91 years

Immortal non-malignant

Primary normal

<30y

>55y

Cell Reports 2014

Immortal post-menopausal HMEC tend to maintain phenotypes of Luminal BC

28

DAPIbeta-catenin Estrogen receptor alpha

Keratin 14Keratin 19DAPI

Age=19y Age=66y

240Lp16sMy 122Lp16sMy

Front in Cell Dev Bio, 2015

Epigenetic states of chronological age and the route of stasis bypass determines cancer subtype

29Front in Cell Dev Bio, 2015

The roles of tissue microenvironments in…

30

Aging &

breast cancer

Epithelial plasticity

Drug response

Molecular

compositionsECM

Growth factor

Mechanical

propertiesSubstrata stiffness

Shear force

ArchitecturePolarity

Dimension

Cell

Cellular

Microenvironment microarray platforms for functional dissection of the microenviroment

32

MicroEnvironment MicroArray (MEMA)A highly parallel cell-based functional screening platform

Cell – ECM

Cell – Cell

Cell – Soluble Factor

Elasticity and Geometry

Thousands of spots per MEArray

Uses standard microarray robot technology

Functionally assess lineage commitment, cell proliferation, and cell death.

Analyzed with fluorescent mAbProgenitors on an MEArray

MyoepLuminalIntegrative Biology, 2009

Combinatorial Microenvironments Impose aContinuum of Cellular Responses to a SinglePathway-Targeted Anti-cancer Compound (Lapatinib)

tSN

E-2

tSNE-1

Inte

nsi

ty m

arke

r 1

Intensity marker 2

Single cell data

Visualization ofcell phenotypes

as functional of ME

Cell Reports 2017

ME1 ME2 ME3

Generalized linear models (GLM) identified individual ME components that had strong effects on Lapatinib responses

Cell Reports 2017

Breast Cancerw/HER2 amp

Lung Cancerw/HER2 amp

Prostate Cancerw/HER2 amp

Cell lines thatrepresent:

GLM can identify synergies between microenvioronment properties (Stiffness and ECM)

Cell Reports 2017

More resistant

More sensitive

Microenvironment-driven morphology co-organizes with drug-response phenotypes, providing clues to therapeutic combinations

Cell Reports 2017

Microenvironments modulate ratios of pHER2 to total HER2 in HER2-amplified BC cells, which correlates with Lapatinib resistance

Relevant summary from previous figures:The rigidity and ECM properties independently drove the most variation in HCC1569 responses to Lapatinib, and EGF showed potential synergy with these properties. FN was specifically implicated.

Cell Reports 2017

+DM

SO

+Lap

atin

ibRigidity ECM Soluble factors

Reports on predictability of anti-HER2 therapy efficacy based on pHER2/HER2 ratios is variable. Context matters….

Adhesion to FN confers resistance to Lapatinib via a Src-YAP related pathway:Drug-tolerance in defined contexts enables identification of work-arounds.

Adhesion to FN activates a FAK-Src-YAP pathway in mammary epithelial cell lines (Kim and Gumbiner, 2015). AZ = AZ0530, a Src inhibitor

FN drives YAP activationIndependent of rigidity

(N)uclear

(C)ytoplasmic

(N/C) nuc & cyto

Cell Reports 2017

Do the functional results that identified FN as a modulator of Lapatinib stand up in patients?

YAP/FN protein level associations

FN protein level survival associationsSUBTYPE

Cell Reports 2017

FN protein level survival associationsTumor GRADE

There are many neighborhoods in a tumor that may have the capacity to increase tolerance to a drug by a non-genetic mechanism…

e.g. This is FN staining in 4 different ductal carcinomas.

1) Cellular phenotypes of aging and high-risk, differentiation, and drug responses are significantly determined by the ME in normal and malignant mammary epithelial cells.

2) Failure to achieve durable drug responses in cancers may have as much to do with heterogeneous tumor MEs, as it does with genetic heterogeneity.

3) Cellular phenotypes of aging in breast, which are related to susceptibility, are transmissible and malleable… is aging in breast preventable?

Take aways

Endocrine changes

lum

en

Stroma Epithelia

WBC

MEP LEPFAT

FB

Stem

Collaborators

LBNL

Mina Bissell

Bahram Parvin

Hang Chang

Ben Brown

University of California Berkeley

Sanjay Kumar

Lydia Sohn

University of California San Francisco

Zev Gartner

Valerie Weaver

Catherine Park

Susan Samson (advocate)

University of California Davis

Sandy Borowsky

University of Copenhagen, Denmark

Ole Petersen

University of Bergen, Norway

James Lorens

Oregon Health Sciences University

Joe Gray

James Korkola

Laura Haiser

Notes4hope.org

Sandy Preto (advocate)

We are grateful for support from:

• National Institutes of Health (R01AG040081, R00AG033176, R01EB024989 )

• Era of Hope Scholar Award, DOD CDMRP

• NIH’s Library of Integrated Network-Based Cellular Signatures (LINCS)

• CoH Center for Cancer and Aging

• Circle 1500

• California Breast Cancer Research Program

• Anita Tarr Turk Fund for Breast Cancer Research

City of HopeMasaru MiyanoMichael TodhunterRosalyn SayamanStefan HinzTiina JokelaSundus ShalabiArrianna ZirbesJennifer Lopez

Lawrence Berkeley LabMartha StampferJames GarbeTara FresquesChunHan Lin (Graduated)