202
The Pennsylvania State University The Graduate School Intercollege Graduate Degree Program in Genetics TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL DIVISIONS IN THE NORMAL MAMMARY GLAND AND MAMMARY NEOPLASIA USING TRANSGENIC MICE A Dissertation in Genetics by Jessica L. Mathers 2011 Jessica L. Mathers Submitted in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy August 2011

TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

  • Upload
    others

  • View
    5

  • Download
    0

Embed Size (px)

Citation preview

Page 1: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

The Pennsylvania State University

The Graduate School

Intercollege Graduate Degree Program in Genetics

TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

DIVISIONS IN THE NORMAL MAMMARY GLAND AND MAMMARY

NEOPLASIA USING TRANSGENIC MICE

A Dissertation in

Genetics

by

Jessica L. Mathers

2011 Jessica L. Mathers

Submitted in Partial Fulfillment

of the Requirements

for the Degree of

Doctor of Philosophy

August 2011

Page 2: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

ii

The dissertation of Jessica L. Mathers was reviewed and approved* by the following:

Edward J. Gunther

Associate Professor of Medicine

Dissertation Advisor

Chair of Committee

Sarah Bronson

Associate Professor of Cellular and Molecular Physiology

Gary Clawson

Professor of Pathology and Biochemistry and Molecular Biology

Jiyue Zhu

Associate Professor of Cellular and Molecular Physiology

David J. Spector

Professor of Microbiology

Chair, Intercollege Graduate Degree Program in Genetics

*Signatures are on file in the Graduate School

Page 3: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

iii

Abstract

With as many as 1 in 8 women diagnosed with breast cancer in their lifetime, breast

cancer is the most commonly diagnosed cancer in women in the Western Hemisphere and the

second most common cause of cancer-related death in females. Decades of study have

uncovered causative exposures and mutations that transform normal mammary epithelial

cells (MECs) into breast cancer cells. Nonetheless, the cellular mechanisms that define the

clinical behavior of breast cancers remain incompletely defined. Lineage commitment

pathways yield diverse MEC cell types within the breast, and recent findings suggest these

MEC lineage hierarchies may persist within breast cancers, perhaps helping to explain the

cellular heterogeneity seen in tumors. To extend these studies, models are needed that permit

cell fate tracking in the discrete MEC compartments of both normal and malignant mammary

tissue.

In this work, we describe novel transgenic mouse models that permit temporally-

regulated, MEC compartment-restricted expression of a histone-fused eGFP (H2B-eGFP)

reporter in both normal and malignant mammary epithelium. Transactivator transgenes

expressed in either the luminal or basal layer of mammary ducts drove widespread H2B-

eGFP labeling of luminal or basal MECs, respectively. We tested whether the H2B-eGFP

reporter could be used to track cell divisions within labeled MEC compartments. Indeed,

when H2B-eGFP transgene expression was switched off in pulse-chase experiments, washout

of label depended on partitioning of labeled histones between daughter cells during MEC

proliferation. Moreover, the H2B-eGFP nuclear label was readily detectable in live MECs,

enabling live cell imaging of MECs propagated in culture.

H2B-eGFP labeling was used for short-term lineage tracing of MECs during

lobuloalveolar development. Hormones of pregnancy trigger the development of bi-layered

Page 4: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

iv

alveolar outgrowths that are believed to arise from luminal progenitor cells. Contrasting with

this model, we found that labeled cells from both the luminal and basal MEC compartments

contribute to alveolar outgrowths. Furthermore, both luminal and basal MECs proliferated

while contributing to alveologenesis, and did not merely migrate into alveoli or become

incorporated as “bystanders”. These findings clarify a lineage commitment pathway

operative during a key, hormonally-driven stage of mammary gland development. In

separate studies, we labeled either basal or luminal MECs residing in the secretory

epithelium of lactating mice to examine whether a subset of these MECs persist throughout

the mammary gland remodeling program triggered by weaning. Remarkably, substantial

numbers of both luminal and basal MECs survived mammary gland involution and

contributed to remodeled ducts. These findings have implications for understanding how a

lactation-involution cycle protects against breast cancer in rodents and humans.

In other studies, the H2B-eGFP labeling strategy was applied in the context of Wnt1-

driven transgenic mouse models of breast cancer. Here, we sought to use pulse-chase H2B-

eGFP labeling of tumor cells to identify a relatively slow-cycling sub-population of tumor

cells, as relatively quiescent tumor cells have been proposed to be treatment-refractory and

enriched in tumor-propagating potential. Each transactivator drove reproducible,

compartment-restricted H2B-eGFP labeling of a large fraction of MECs within Wnt1-

initiated mammary hyperplasia, as expected. In contrast, the fraction of tumor cells labeled in

Wnt1-initiated mammary cancers showed marked tumor-to-tumor variability, suggesting that

the genetic and epigenetic events that cooperate in tumorigenesis sometimes interfere with

transgene-mediated labeling. Notably, both the luminal- and basal-MEC-directed

transactivators were capable of driving H2B-eGFP labeling of a subset of tumor cells from

Page 5: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

v

MMTV-Wnt1 transgenic mice. This finding supports the contention that Wnt1 initiates

“mixed-lineage” mammary tumors comprised of both luminal and basal tumor cells,

suggesting that Wnt1 transforms a bi-potent MEC progenitor. Tumors that labeled efficiently

were studied further by quantifying H2B-eGFP washout on a per-cell basis in pulse-chase

experiments. Mammary cancers typically were comprised of tumor cell populations that were

heterogeneous with respect to rates of proliferation. Together, this work sets the stage for

prospective studies that will compare the biological potential of luminal- versus basal-type

tumor cells (i.e., those labeled using the contrasting compartment-restricted transactivators)

and rapid- versus slow-cycling tumor cells (i.e., those that have depleted versus retained

H2B-eGFP label).

Breast cancers can recur many years after eradication of all clinically-detectable

disease, but the cellular mechanisms that maintain disease dormancy remain undefined. H2B-

eGFP labeling was applied in the context of reversible Wnt1-driven mammary tumors to

investigate the cellular mechanisms that maintain tumor dormancy in a mouse breast cancer

model. Here, H2B-eGFP and Wnt1 were co-expressed such that switching off both

transgenes simultaneously initiated washout of incorporated H2B-eGFP label and regression

of Wnt1-dependent mammary cancers. Subclinical lesions comprised of dormant mammary

cancer persisted long after tumor regression, and these lesions frequently harbored tumor

cells that retained bright H2B-eGFP label, indicating they had ceased proliferating. These

label-retaining cells represent a candidate quiescent tumor cell population that may serve as a

critical link between primary mammary cancer and subsequent disease relapse.

Page 6: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

vi

Table of Contents

List of Figures ........................................................................................................................... x

List of Tables ......................................................................................................................... xiii

Abbreviations ......................................................................................................................... xiv

Acknowledgements ................................................................................................................ xvi

Chapter 1 Literature Review ..................................................................................................... 1

1.1 Structure, Function, and Development of the Mammary Gland ............................... 1

1.1.1 Development of the Mammary Gland .................................................................. 1

1.1.1.1 Prenatal Development of the Mammary Gland ........................................... 1

1.1.1.2 Post-natal Development of the Mammary Gland ........................................ 2

1.1.1.3 Terminal Differentiation Induced By Pregnancy and Lactation ................. 3

1.1.2 Cellular Composition of the Mammary Gland ..................................................... 4

1.1.3 Hierarchy of Mammary Epithelial Cells ............................................................... 5

1.1.4 Areas for Further Study ........................................................................................ 8

1.2 Breast Cancer ............................................................................................................ 9

1.2.1 Types of Breast Cancers ..................................................................................... 10

1.2.1.1 Traditional Characterization of Breast Cancers ......................................... 10

1.2.1.1.1 Hormone Receptor Positive Breast Cancers ......................................... 11

1.2.1.1.2 Hormone Receptor Negative Breast Cancers .......................................... 12

1.2.1.1.3 HER2/ErbB2/Neu Initiated Breast Cancers ............................................. 12

1.2.1.2 Transcriptional Profiling of Breast Cancers .............................................. 13

1.2.2 Hierarchy of Cells in Mammary Cancers ........................................................... 14

1.2.3 Breast Cancer and Tumor Dormancy ................................................................. 15

1.2.4 Parity-Related Protection from Breast Cancer .................................................... 16

1.3 Mouse Models of Human Breast Cancer ................................................................ 17

1.3.1 Virally-Induced Mouse Mammary Cancer and the Discovery of the Wnt1

Oncogene ........................................................................................................................ 18

1.3.1.1 Engineering Constitutive MMTV-Driven Expression of Oncogenes in the

Mammary Gland ......................................................................................................... 19

1.3.2 The Role of Wnt Signaling in Mammary Development and Breast Cancer ....... 20

1.3.2.1 Inducible Expression of Wnt1 in the Mammary Epithelium of Transgenic

Mice ............................................................................................................................ 22

1.3.2.1.1 Minimal Residual Disease Lesions Modeled in Transgenic Mice ........ 23

1.3.3 Constitutive MMTV-Neu Driven Tumorigenesis in the Mouse Mammary Gland

............................................................................................................................. 25

Page 7: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

vii

1.3.4 Compartment-Restricted Expression of Transgenes ........................................... 25

1.3.5 Inducible Expression of Fluorescently Tagged Histone H2B ............................. 26

1.3.6 Experimental Manipulation of the Mouse Mammary Gland .............................. 27

1.4 Areas Addressed in this Dissertation ...................................................................... 28

Chapter 2 Materials and Methods ........................................................................................... 34

2.1 Mouse Strains, Surgeries and Drug Treatments ........................................................... 34

2.1.1 Housing and Maintenance...................................................................................... 34

2.1.2 Genotyping and Breeding Schemes ....................................................................... 35

2.1.3 Experimental Manipulations of Mice .................................................................... 35

2.1.3.1 Surgical ........................................................................................................... 36

2.1.3.2 Chemotherapeutics .......................................................................................... 37

2.1.3.3 Pregnancy ........................................................................................................ 37

2.2 Mammary Gland and MEC Sample Processing ........................................................... 38

2.2.1 Single Cell Suspensions ......................................................................................... 38

2.2.2.1 Immunophenotyping and Fluorescent Signal Analysis of MECs and Tumor

Cells ............................................................................................................................ 40

2.2.3 Imaging of Whole Mounts and Sections ................................................................ 40

2.2.3.1 Staining ........................................................................................................... 41

2.3 Time-lapse Imaging of H2B-eGFP Labeled Cells ........................................................ 42

Chapter 3 ................................................................................................................................. 45

3.1 Developing a transgenic mouse model enabling inducible H2B-eGFP labeling of

MECs .................................................................................................................................. 45

3.2 Doxycycline-dependent labeling of MG ....................................................................... 45

3.3 Compartment-Specific Labeling in the Mammary Epithelium .................................... 47

3.3.1 MMTV-rtTA Drives Luminal Compartment Restricted Reporter Gene Expression

......................................................................................................................................... 47

3.3.2 Keratin-5 rtTA Drives Basal Compartment-Restricted Reporter Gene Expression

......................................................................................................................................... 48

3.4 Proliferation dependent washout of H2B-eGFP labeling ............................................. 49

3.4.1 Puberty Induced Proliferation Results in Dilution of Incorporated GFP Signal

Differentially in Developmentally Distinct Ductal Areas .............................................. 50

3.4.2 Ovariectomy Blocks Proliferation Dependent Washout of H2B-eGFP Label in

Mammary Epithelial Cells .............................................................................................. 51

3.5 Simultaneous Induction of H2B-eGFP and Tet-responsive Wnt Oncogene Results in

Labeling of Mammary Hyperplasia .................................................................................... 52

3.6 H2B-eGFP Transgene Labeling Allows Live Imaging of Mammary Cells ................. 53

Page 8: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

viii

3.7 Discussion ..................................................................................................................... 54

Chapter 4 H2B-eGFP Labeling and Proliferation Dynamics in Mammary Tumors Driven by

Constitutive Oncogenes .......................................................................................................... 73

4.1 Introduction ................................................................................................................... 73

4.2 Experimental Strategies Allowing the Study of Proliferation Dynamics in Mammary

Tumors Driven by Constitutive Oncogenes ........................................................................ 74

4.3 Dox-regulated GFP Labeling Independent of Tumorigenesis ...................................... 75

4.4 Identification of LRC in Explanted Mammary Tumors ............................................... 77

4.5 Adriamycin Treatment Does Not Alter GFP Labeling and Label Retention Dynamics

in MMTV-wnt1 Driven Mammary Tumors ....................................................................... 79

4.6 Keratin-5 Induced Labeling and Label Retention in MMTV-wnt1 Driven Constitutive

Mammary Tumors .............................................................................................................. 80

4.7 Constitutive MMTV-Neu Mammary Tumors can be Labeled with H2B-eGFP .......... 81

4.8 Discussion ..................................................................................................................... 82

Chapter 5 H2B-eGFP Labeling and Proliferation Dynamics in Reversible Mammary Tumors

and Minimal Residual Disease Lesions ................................................................................ 101

5.1 Introduction ................................................................................................................. 101

5.2 Experimental Strategy for the Study of Proliferation Dynamics in Reversible

Mammary Tumors and Minimal Residual Disease Lesions ............................................. 101

5.3 Doxycycline-dependent Labeling of Inducible Mammary Tumors ............................ 103

5.4 MRD Retain Bright H2B-eGFP Label Following Tumor Regression........................ 104

5.5 Discussion ................................................................................................................... 107

Chapter 6 A Transgenic Model for Short-Term Lineage Tracing of Mammary Epithelial

Cells in Pregnancy ................................................................................................................ 121

6.1 Introduction ................................................................................................................. 121

6.2 Contributions of Mammary Epithelial Cell Sub-types to Lobulo-alveolar Outgrowths

of Pregnancy ..................................................................................................................... 123

6.2.1 Tracing the Contributions of Luminal Mammary Epithelial Cells to Lobulo-

alveolar Outgrowths During Pregnancy ........................................................................ 123

6.2.2 Tracing the Contributions of Basal Mammary Epithelial Cells to Lobulo-alveolar

Outgrowths During Pregnancy ..................................................................................... 125

6.3 Contributions of Mammary Epithelial Cell Sub-types to Lobulo-alveolar Outgrowths

Induced by Hormone Stimulation ..................................................................................... 127

6.3.1 Tracing the Contributions of Luminal Mammary Epithelial Cells to Lobulo-

alveolar Outgrowths During Hormone Induced Proliferation and Differentiation ....... 128

6.3.2 Tracing the Contributions of Basal Mammary Epithelial Cells to Lobulo-alveolar

Outgrowths During Hormone Induced Proliferation and Differentiation ..................... 129

Page 9: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

ix

6.4 Persistence of Labeled Cells Through Post-Lactational Remodeling of the Mammary

Gland ................................................................................................................................. 130

6.4.1 Luminal Labeled Mammary Epithelial Cells Persist Through Involution ........... 131

6.4.2 Basal Labeled Mammary Epithelial Cells Persist Through Involution ............... 132

6.5 Discussion ................................................................................................................... 134

Chapter 7 Final Discussion ................................................................................................... 158

7.1 Labeling the Luminal and Basal MEC Compartments ......................................... 158

7.2 Using H2B-eGFP Labeling to Characterize Mammary Tumors .......................... 160

7.3 LRCs in MRD and Maintenance of MRD .................................................................. 164

7.4 Lineage Restriction and Cell of Origin in Alveologenesis ......................................... 165

7.5 Parity-related Protection from Breast cancer .............................................................. 168

7.6 Summary ..................................................................................................................... 169

References: ............................................................................................................................ 170

Appendix: Flow Cytometric Analysis Demonstrates Variability of LRCs in MMTV-Wnt1

Induced Mammary Tumor Explants, Additional Samples ................................................... 182

Page 10: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

x

List of Figures

Figure 1.1 Basic Anatomy of the Mammary Gland. ..................................................................... 30

Figure 1.2 Secretory Differentiation and Cyclical Remodeling of the Mammary Ductal Tree. .. 31

Figure 1.3 Schematic of Bilayered Mammary Ducts and Keratin Expression Patterns. .............. 32

Figure 1.4 A Proposed Hierarchy of Mammary Epithelial Cells. ................................................. 33

Figure 3.1 Strategy for Temporal Regulation of H2B-eGFP Transgene Expression in

Epithelial Cellular Compartments of the Mammary Gland. ......................................................... 58

Figure 3.2: Dox-Regulated Expression of the H2B-eGFP Transgene in Mammary

Epithelium of MMTV-rtTA/TGFP Mice. ..................................................................................... 59

Figure 3.3 Luminal-Restricted Labeling of MECs in MMTV-rtTA/TGFP Mice. ..................... 60

Figure 3.4 Immunophenotyping of Labeled Mammary Epithelial Cells Confirms

Compartment-Specific H2B-eGFP Fluorescent Labeling ............................................................ 61

Figure 3.5 Dox-Regulated Expression of H2B-eGFP in Basal Mammary Epithelium of K5-

rtTA/TGFP Mice. .......................................................................................................................... 62

Figure 3.6 Basal-Restricted Labeling of MECs in K5-rtTA/TGFP Mice. .................................... 63

Figure 3.7 Immunophenotyping of Labeled Mammary Epithelial Cells Confirms H2B-eGFP

Fluorescent Labeling. .................................................................................................................... 64

Figure 3.8 Selective Washout of H2B-eGFP Fluorescence by Developmental Proliferation

in Terminal End Buds. .................................................................................................................. 65

Figure 3.9 Proliferation is Necessary for the Washout of Induced H2B-eGFP Signal in the

Mammary Gland. .......................................................................................................................... 66

Figure 3.10: Elimination of Ovarian Hormone Induced Proliferation Limits H2B-eGFP

Signal Dilution in the Mammary Gland. ....................................................................................... 67

Figure 3.11 A Strategy for Dox-Regulated H2B-eGFP Labeling of MECs in the Context of

Inducible Wnt1 Expression. .......................................................................................................... 68

Figure 3.12 H2B-eGFP Labeling of Luminal MEC-driven, Wnt1-initiated Mammary

Hyperplasia. .................................................................................................................................. 69

Figure 3.13 H2B-eGFP Labeling of Basal MEC-driven, Wnt1-initiated Mammary

Hyperplasia. .................................................................................................................................. 70

Figure 3.14 H2B-eGFP Labeling of Mammary Epithelial Cells Allows Tracking of Cell Fates

in Real Time via Time-Lapse Confocal Microscopy. ................................................................... 71

Figure 3.15 Time-Lapse Fluorescence Microscopy of H2B-eGFP Labeled Mammary Cells in

Mammary Duct Fragments in Culture Allows Tracking of Cellular Proliferation Dynamics. .... 72

Figure 4.1 Mammary Adenocarcinomas Arising in the MMTV-wnt1 Model Show a Mixed-

Lineage Phenotype. ....................................................................................................................... 87

Figure 4.2 Strategy for the Dox-Dependent H2B-eGFP Labeling of Either Basal-Type or

Luminal-Type Tumor Cells in the MMTV-wnt1 Model. ............................................................. 88

Figure 4.3 Homogeneous Versus Heterogeneous Models of H2B-eGFP Washout in Growing

Mammary Tumors. ....................................................................................................................... 89

Page 11: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

xi

Figure 4.4 Wide Variation in the Extent of H2B-eGFP Labeling of Luminal-Type Tumor

Cells in the MMTV-Wnt1 Model. ................................................................................................ 90

Figure 4.5 Wide Variation in the Rate of H2B-eGFP Washout from Luminal-Type Tumor

cells in the MMTV-Wnt1 Model. ................................................................................................. 91

Figure 4.6 Generating Clonally-Related Tumor Outgrowths by Explanting onto Syngeneic

Host Mice. ..................................................................................................................................... 92

Figure 4.7 Variable H2B-eGFP Labeling Among Clonally-Related MMTV-Wnt1 Mammary

Tumor Explants. ............................................................................................................................ 93

Figure 4.8 Evidence for Heterogeneous H2B-eGFP Washout in a Subset of MMTV-Wnt1

Tumor Explants. ............................................................................................................................ 94

Figure 4.9 Variable H2B-eGFP Washout Among Clonally-Related MMTV-Wnt1 Mammary

Tumor Explants. ............................................................................................................................ 96

Figure 4.10 Chemotherapy with Adriamycin Minimally Impacts H2B-eGFP Labeling and

Washout in MMTV-Wnt1 Tumor Explants. ................................................................................. 97

Figure 4.11 H2B-eGFP Labeling of Basal-Type Tumor Cells in the MMTV-Wnt1 Model. ....... 98

Figure 4.12 MMTV-Neu Driven Tumors Contain a Homogeneous Cell Population Capable

of H2B-eGFP Labeling. .............................................................................................................. 100

Figure 5.1 Strategy for the Concurrent Temporal Regulation of the H2B-eGFP and Tet-op

Wnt Transgenes Expression in Mammary Tumors and Minimal Residual Disease Lesions. .... 113

Figure 5.2 Two Models for the Maintenance of Minimal Residual Disease Lesions. ................ 114

Figure 5.3 Wide Variation in the Extent of H2B-eGFP Labeling of Luminal-Type Tumor

Cells in the MMTV-rtTA/TWNT/TGFP Model. ........................................................................ 115

Figure 5.4 Minimal Residual Disease Lesions Retain H2B-eGFP Signal Following Tumor

Regression. .................................................................................................................................. 117

Figure 5.5 MRD Retain Greater Fluorescent Label than Surrounding Normal Ducts. .............. 118

Figure 5.6 Flow Cytometric Analysis of GFP Signal Intensity in Matched Tumor Biopsy,

MRD Lesions, and Regressed Hyperplasia. ............................................................................... 120

Figure 6.1 A Model of Alveolar Progenitors. ............................................................................. 137

Figure 6.2 Parity-Related Protection from Breast Cancer. ......................................................... 138

Figure 6.3 Experimental Timelines For the Investigation of Cell Fates During Pregnancy. ...... 139

Figure 6.4 Schematic Depictions of Potential Outcomes of Lineage Tracing During

Alveologenesis. ........................................................................................................................... 141

Figure 6.5 H2B-eGFP Labeling of Luminal Cells During Pregnancy in MMTV-rtTA/TGFP

Mice. ........................................................................................................................................... 142

Figure 6.6 Luminal Cells Contribute to Lobulo-alveolar Outgrowths During Pregnancy. ........ 143

Figure 6.7 H2B-eGFP Labeling of Basal Cells During Pregnancy in K5-rtTA/TGFP Mice ..... 144

Figure 6.8 Basal Cells Contribute to Lobulo-alveolar Outgrowths During Pregnancy. ............. 145

Page 12: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

xii

Figure 6.9 H2B-eGFP Labeling of Luminal Cells During Hormone Induced Lobulo-alveolar

Development. .............................................................................................................................. 146

Figure 6.10 Luminal Cells Contribute to Lobulo-alveolar Outgrowths During Hormone

Induced Lobulo-alveolar Development. ..................................................................................... 147

Figure 6.11 H2B-eGFP Labeling of Basal Cells During Hormone Induced Lobulo-alveolar

Development. .............................................................................................................................. 148

Figure 6.12 Basal Cells Contribute to Lobulo-alveolar Outgrowths During Hormone Induced

Lobulo-alveolar Development .................................................................................................... 149

Figure 6.13 Dox Regulated H2B-eGFP Labeling of Luminal Mammary Epithelial Cells

During Lactation. ........................................................................................................................ 150

Figure 6.14 Post-lactational Remodeling of the Mammary Gland Retains Luminal Labeled

Mammary Epithelial Cells .......................................................................................................... 152

Figure 6.15 Dox Regulated H2B-eGFP Labeling of Basal Mammary Epithelial Cells During

Lactation ..................................................................................................................................... 153

Figure 6.16 Post-lactational Remodeling of the Mammary Gland Retains Basal Labeled

Mammary Epithelial Cells. ......................................................................................................... 155

Figure 6.17 A Revised Model of Alveolar Progenitor Cells. ..................................................... 157

Page 13: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

xiii

List of Tables

Table 2.1 Primer Pairs Utilized in Genotyping Transgenic Mice ................................................. 44

Page 14: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

xiv

Abbreviations

3D Three-Dimensional

APC Adenomatous Polyposis Coli

BrdU 5-bromo-2-deoxyuridine

C Celsius

CSC Cancer Stem Cell

DCIS Ductal Carcinoma In Situ

DNA Deoxyribonucleic Acid

DNAse Deoxyribonuclease I

Dox Doxycycline

Dsh Dishevelled

E Estrogen

EDTA Ethylenediaminetetraacetic acid

EGF Epidermal Growth Factor

EMT Epithelial to Mesenchymal Transition

ERDs Estrogen Receptor Downregulators

ERα Estrogen Receptor α

FACS Fluorescence Assisted Cell Sorting

FISH Fluorescent In Situ Hybridization

FVB an inbred mouse strain named for their sensitivity to

Friend Virus B strain leukemia

Fz Frizzled

G0P0 Gravida 0, Parity 0; Virgin

GFP Green Fluorescent Protein

GSK3 Glycogen Synthase kinase 3β

H+E Hematoxylin and Eosin

H2B-eGFP Histone H2B Fused to enhanced Green Fluorescent

Protein

HER2/Neu/ERBB2 (HER2) Homologues of the Human ERBB2 Receptor, an EGF

Family Receptor Tyrosine Kinase

HR Hormone Receptor

i.p. intra peritoneal

IDC Invasive Ductal Carcinoma

Inv involution

K14 Keratin 14

K5 Keratin 5

Lact lactation

LRCs Label Retaining Cells

MaSC Mammary Stem Cell

MECs Mammary Epithelial Cells

MMTV Mouse Mammary Tumor Virus

MMTV-LTR Mouse Mammary Tumor Virus Long Terminal Repeat

MNU Methylnitrosourea

MRD Minimal Residual Disease

Ovx ovariectomy

P Progesterone

Page 15: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

xv

PBS Phosphate Buffered Saline

PCR Polymerase Chain Reaction

PE Phycoerythrin

PFA Paraformaldehyde

PR Progesterone Receptor

Preg pregnant

RNA Ribonucleic Acid

rtTA Reverse-Tet Transactivator

SERMs Selective Estrogen Receptor Modulators

SMA Smooth Muscle Actin

TDLU Terminal Ductal Lobular Units

TEBs Terminal End Buds

TetO Tetracycline Operator

TGFP Tet-Operator Driven H2B-eGFP Transgene

TRAS Tet Operator Driven ras Transgene

TWNT Tet-Operator Driven Wnt1 Transgene

UV Ultra-Violet Radiation

wg wingless

WIF1 Wnt Inhibitory Factor 1

Page 16: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

xvi

Acknowledgements

I am deeply indebted to my advisor, Dr. Ed Gunther, for helping me become not just a better

scientist, but also for allowing me the time and encouragement to develop into the teacher,

wife, and mother I am today. Without his patience and guidance, I would not have thrived in

the laboratory over the last several years. Thank you for all your help over my time in the

lab, I am truly grateful that you chose me to join your lab.

I would like to thank the other members of the Gunther Lab for all their assistance, both with

experiments and daily life, over the years. You have all helped me maintain my sanity even

when my mice didn’t behave and my immunos never worked. The advice I was given when

interviewing for grad school was true- don’t choose a lab based on the science- projects come

and go, but the people, they are the ones who will be with you all along. I am so very happy

that I was able to spend my grad school years in the Gunther Lab.

Shelley-Thank you for all the birthday cakes, loaning me your children, watching

mine, watching me do the procedure “just one more time” to make sure I finally can

do it as well as you, ensuring I always had everything I needed to get my work done,

and taking over all the mice I left behind on maternity leaves. I doubt I can ever

express myself sufficiently, but I admire your dedication to both your family and the

lab.

Travis- Thank you for genotyping more mice than I can count and for many late

afternoon games of “name that tune”. I am so glad you came back to Hershey as the

Gunther Lab just wasn’t the same without you there.

Kristin- I was so glad to be able to decorate the middle of the lab with you- we

certainly made some great origami! I will miss being pelted with random lab

plasticware and sharing the latest internet nonsense during a break in experiments.

Allison- I really enjoyed having you in the lab, your sense of humor helped me

survive those last months. It was so nice having you in the grad student corner with

me! Thank you.

Dan, Mike, Will- While you all were in the lab, you each helped me learn skills I will

never forget. From the rules of cricket and the function of ceiling snorkels, to the

importance of explaining the smallest detail, to the necessity of always, ALWAYS

reading the instructions you each reminded me that there is more to life than the

success or failure of any experiment.

I would like to additionally thank

Lynn Budgeon- for preparing countless sections for me, often on incredibly short notice,

and always with a smile.

Jeanette and the other handlers in the mouse room- for making each and every visit to the

mice easier. Your senses of humor and kindness brightened many of my days.

Sarah Bronson, Gary Clawson, and Jiyue Zhu, the members of my committee- for the

Page 17: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

xvii

persistence and belief that my developing project would someday grown into a thesis even

with the stumbling blocks along the way and for your encouragement to pursue my career

when I was frustrated.

Kathy Simon and Kathy Shuey- for guiding me through all the paperwork and pitfalls of

graduate school, providing much needed breaks and reminding me that this too would pass. I

am so grateful that you both helped me balance and learn to balance the demands of school

and motherhood.

To my friends in graduate school, thank you for venting lunches, potluck dinners, and nights

spent studying and laughing. I will remember you all as the people who reminded me that

there is life in graduate school and after. To my friends outside graduate school- thank you

for your patience until I “finally” graduated, I’ll try not to talk about mice so much anymore!

Kim- No words can describe how happy I am to have met you. You are more than just my

friend; you are part of our family. You know you are beloved as Aunt Kim to my children,

but I want you to know that you are very loved and appreciated by all of us. Thank you for

everything.

I owe my parents gratitude for always, always believing in me and encouraging me to never

ever quit. I would never have been able to complete this achievement without their unfailing

dedication and support.

I must end this by thanking my family, my husband Craig, and my children Katie and John.

I know the time I have spent in graduate school has been busy and chaotic, but I am grateful

for your love and patience that have supported me through all the changes of these years.

Thanks for sticking with me and not complaining even when we had to have dinner with the

microscope. I love you all very much.

Page 18: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

1

Chapter 1 Literature Review

1.1 Structure, Function, and Development of the Mammary Gland

The mammary gland evolved relatively recently, first appearing about 200 million years

ago, with the primary function being to provide newborn mammals with nutrition in the form

of milk. The mammary gland is exquisitely dependent on hormonal stimulation for

regulation of its growth and development [1]. The cellular composition of the mammary

gland has been studied for many years and putative stem cell populations for the ductal tree

have been proposed recently; however, the exact hierarchy of lineage commitment of MECs

has not been completed.

1.1.1 Development of the Mammary Gland

Mammopoiesis is distinct from the development of most organ systems in that only

rudimentary development occurs during the embryonic and fetal stages. The full elaboration

of the mammary gland requires the onset of puberty and a complex milieu of endocrine

signals. Terminal differentiation of the mammary glands is not completed until the onset of

pregnancy induces lobulo-alveolar growth and lactation [2]. The mammary gland also has

the capacity to undergo sequential rounds of proliferation, differentiation, and subsequent

remodeling in response to hormonal stimulation. This cyclical nature of mammary gland

development provides an interesting avenue for the study of lineage commitment and lineage

plasticity of mammary epithelial cells

1.1.1.1 Prenatal Development of the Mammary Gland

During embryogenesis the first sign of the developing mammary gland is found in the

appearance of the milk lines in the ectoderm of both male and female embryos [3], followed

Page 19: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

2

by the development of pairs of placodes of epithelial cells that bud into the surrounding

mesenchymal tissue forming a small number of ducts embedded in a fat-pad. Further growth

of the mammary gland is then paused until the onset of puberty, though at birth, the

rudimentary mammary gland structures are competent to produce milk [4].

1.1.1.2 Post-natal Development of the Mammary Gland

Prior to puberty, the mammary gland consists of a few short ducts extending from the

nipple surrounded by stromal tissue commonly called the fat pad (Figure 1.1). The surge in

hormone levels accompanying the onset of puberty drives the formation of club like

structures at the end of the ducts termed Terminal End Buds (TEBs) [5]. Unlike the ducts

which maintain a lumen, the TEBs are comprised of an outer layer of cap cells, which give

rise to myoepithelial cells, and a multilayered internal mass of body cells, the precursors of

luminal epithelium [6]. As circulating estrogen levels begin to rise, a marked increase in

growth of the mammary gland begins [1, 7]. This growth is characterized by: 1) increased

levels of proliferation within the TEBs that drive ductal elongation through the surrounding

fat pad, and 2) regulated apoptosis within the body cells that leads to lumen formation in the

growing duct [5, 6]. The ductal elongation, together with bifurcation of TEBs, leads to

ductal branching and the rapid filling of the fat pad with mammary duct. Morphogenesis of

the developing mammary gland, particularly the formation of lumens, is regulated by a vast

array of signals, including hormones, growth factors, extracellular matrix, morphogens, and

immune cells [6, 8]. These signals facilitate communication between the growing epithelial

cell themselves and between the ductal epithelium and the surrounding stroma [8, 9].

Elongation of the ductal tree induced by puberty ceases when the ducts reach the ends of the

fat pad. In addition to the branching that occurs during ductal morphogenesis, secondary

Page 20: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

3

side-branching occurs even after the ductal tree fully invests the fat pad [10]. Side-branching

of the ductal tree appears to be regulated by the detection of neighboring ducts and by the

arrangement of the ductal structures [11-13]. The hormonal changes of the female estrus

cycle can lead to rounds of proliferation and side-branching. Subsequent remodeling of the

mammary gland via apoptosis maintains a homeostatic level of epithelial development [2].

1.1.1.3 Terminal Differentiation Induced By Pregnancy and Lactation

Terminal differentiation of the mammary gland is dependent on the hormones of

pregnancy which induce rapid expansion of the lobulo-alveolar compartment in order to

facilitate lactation [14, 15]. Prolactin and progesterone, in cooperation with proper 3D

structure of the ducts, extracellular matrix, and growth hormones, drive increased side-

branching of the mammary ducts and ductules, and differentiation of alveoli, the structures

which eventually synthesize and secrete milk [16, 17]. As depicted in Figure 1.2, during

pregnancy, a rapid expansion of the epithelial cell population creates the lobules and alveoli

required for milk production. Pregnancy triggers secretory cell differentiation causing the

production of milk proteins and lipids [2]. The lumen of the mammary ducts serves both as

storage for milk produced during lactation and also as the egress route for that milk to the

nipple.

Continued milk production in the mammary gland is dependent on both suckling

stimulation and hormonal influences [18]. Subsequent weaning of offspring induces

involution, a remodeling of the mammary gland that eliminates milk production and restores

the normal homeostasis and virgin-like appearance of the mammary gland (Figure 1.2). This

post-lactational regression is mediated largely by apoptosis of secretory epithelial cells and a

subsequent remodeling of the mammary gland includes disruption and repair of the basement

Page 21: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

4

membrane and phagocytosis of apoptotic cells [18-21]. Remodeling of the mammary gland

during involution does not remove or reduce the capacity to undergo subsequent rounds of

pregnancy-induced growth and differentiation.

1.1.2 Cellular Composition of the Mammary Gland

Early in embryogenesis, two distinct cell types can be identified in the developing

mammary gland. These cells are the precursors to the epithelial and stromal cells that the

mammary gland contains. Ducts within the adult mammary gland are comprised of two

layers of pseuodostratified epithelium surrounded by a laminin-containing basement

membrane [22, 23]. The inner layer is composed of luminal epithelial cells surrounding a

hollow lumen and the outer layer is made up of a single cell-layer of myoepithelial (also

referred to as basal) cells (Figure 1.3). The integrity of the basement membrane is important

for regulating polarity, growth and response to hormones for both the luminal epithelial and

myoepithelial layers [24, 25].

The inner layer of luminal epithelial cells display cuboidal morphology and can be

identified by immunohistochemical staining of cells for keratins 8, 11, 14, 20 and 22 [26].

Luminal epithelial cells are also characterized by the following cell surface markers

CD29lo

/CD49f+/CD24

hi/CD61

-/Sca1

+ [27-29]. These cells line the hollow lumen of the

mammary duct and can secrete milk proteins. A small subset of luminal epithelial cells

express Estrogen Receptor α (ERα) and are thus directly hormone responsive [15, 30]. Cell-

cell interactions between luminal layer cells are proposed to maintain cell type specificity

within the ducts of the mammary gland [31].

The myoepithelial, or basal layer, consists of elongated epithelial cells. These cells

are contractile and capable of aiding milk expulsion from the alveoli and through the ducts in

Page 22: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

5

response to suckling stimulation [32, 33]. Myoepithelium is also responsible for the

elaboration of basement membrane during the development of the ductal tree [23].

Myoepithelial cells are characterized by expression of keratin 5, keratin 14 and smooth-

muscle actin [22, 34] and are CD29hi

/CD49fhi

/CD24+/CD61

+ [27-29]. The localization of

myoepithelial cells, interposed between the luminal layer, basement membrane and stroma

allows integration of multiple signals from the microenvironment [35-38]. Thus,

myoepithelial cells play a role in coordinating epithelial cell interactions.

The extra-epithelial or stromal compartment of the mammary gland derives from the

mesoderm. This tissue compartment is composed largely of fat cells which provide a

structural framework for the mammary ductal tree [39]. The stroma also contains the

vasculature required to support the cells of the mammary gland. While the structural role of

the fat-pad is critical, interactions between the stroma and epithelial cells are also crucial to

proper growth and function of the mammary gland. A wide assortment of candidate signal

transduction molecules have been implicated in the interaction between stromal cells and

ductal epithelial cells including matrix metalloproteinases and fibroblast growth factors [8,

39-42].

1.1.3 Hierarchy of Mammary Epithelial Cells

By using transplant studies, DeOme and colleagues demonstrated the presence of

cells throughout the mammary gland capable of giving rise to a fully developed mammary

ductal tree [43-45]. Their studies additionally provided evidence for cells capable of serially

propagating a mammary ductal tree through several rounds of transplantation suggesting the

presence of a stem cell-like population within the mammary duct fragments. Later work

Page 23: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

6

demonstrated extended but ultimately limited proliferation if serially implanted ductal

fragments [46].

To better understand mammary stem cell biology, several groups have recently

sought to prospectively isolate a candidate stem cell population from the mammary gland

based on cell surface markers. Adapting the strategy used to isolate hematopoietic stem

cells, single cells isolated from mammary glands and subsequently stained with a range of

cell surface markers can be analyzed by flow cytometry and sorted. These sorted cells can

then be transplanted or studied in culture systems in order to examine their regenerative

capacities. The downstream patterns of cell lineage during normal development are proposed

to consist of a bipotential progenitor cell with high proliferative potential that can give rise to

cells of both luminal and basal type. Further commitment of daughter cells results in ductal-

and myoepithelial-restricted progenitor populations capable of giving rise to only cells of a

single cell layer. The exact characterization and capability of the alveolar progenitor awaits

definition. It is not yet clear whether bi-layered alveolar structures derive solely from

alveolar progenitor cell populations or if each cell type arises from a distinct lineage

committed progenitor population. Figure 1.4 presents a model of lineage pathways within

the cellular hierarchy of the mammary epithelial tree.

A MaSC-enriched population is contained within the CD61+/Sca1

-/CD24

lo/CD29

hi

(β-integrin) or CD61+/Sca

-/CD24

lo /CD49f

hi (α6-integrin) population [27-29]. Cells within

these populations were capable of reconstituting a mammary ductal tree containing cells of

both the luminal and basal lineages following implantation and through sequential rounds of

implantation and gave rise to large colonies containing cells of both luminal and basal

character. Interestingly, the MaSC-enriched populations express neither the ERα or

Page 24: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

7

progesterone receptors [47, 48]. However, the absence of hormone receptors from the stem

cell population does not appear to preclude an indirect response of these cells to hormone

stimulation, as they respond to paracrine signaling from other MECs [49].

Similar to the proposed MaSC immunophenotype, basal epithelial cells are CD29hi

or

CD49fhi

and also CD24lo

/CD61+

[27-29]. The overlap in cell-surface markers between the

MaSC population and the basal epithelial cell population within the mammary gland has

supported the potential localization of the MaSC population within the basal layer of the

ducts. A myoepithelial progenitor population has not yet been isolated though it is likely to

segregate with the proposed MaSC population. The myoepithelial progenitor population is

proposed to exist based on the identification of basal restricted cell colonies in culture [50].

Cells in the luminal layer of mammary ducts derive from a luminal committed

progenitor. An immunophenotype of these cells has been suggested as

CD29lo

/CD49f+/CD24

+/CD61

+/KIT

+ and either ERα

+ or ERα

- [27-29, 51]. The variability in

ERα status of luminal progenitors recapitulates the variation in ERα expression in total ductal

cells within the mammary gland. Luminal progenitor cells isolated using this

immunophenotype (excluding ERα status) gave rise to larger colonies in 3D culture in

Matrigel than those arising from CD61- cells. Additionally, these colonies contained only

luminal cells unlike colonies derived from cells in the MaSC population [51]. Committed

cells of the luminal layer bear the CD29lo

/CD49f+/CD24

+/Sca1

+ /CD61

- immunophenotype

[27-29, 51, 52]. The proportion of committed luminal cells to luminal progenitors rises

through puberty and during pregnancy concurrent with the widespread proliferation and

differentiation occurring during those developmental periods [51].

Page 25: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

8

The precursor cells that give rise to alveolar outgrowth remain undefined though a

potential alveolar progenitor cell has been described bearing the immunophenotype

CD24+/CD49f

+/Sca1

- [53-55]. The precise contributions and origin of this population is

unknown. One model posits that a luminal progenitor may also give rise to an alveolar

progenitor. An alternative model proposes the existence of an alveolar progenitor population

that gives rise to the lobulo-alveolar differentiated cells during pregnancy from the common

progenitor cell. The first model accounts for bi-layered alveolar outgrowths by assuming

contributions from multiple lineage-committed progenitors. The second model accounts for

the bi-layer by assuming that all the cells in an alveolus are the progeny of a bi-potent

progenitor capable of begetting both luminal and basal MECs.

1.1.4 Areas for Further Study

The developmental patterns and the mechanisms of regulation of proliferation and

differentiation within the mammary gland have largely been described. However, the precise

cellular contributions and lineages that give rise to the developed mammary gland have only

recently been studied. Further characterization of the roles of stem cells and progenitors and

the precise lineage commitment pathways would provide greater insight into the complex

development of the mammary gland and possibly into other epithelial tissues. Additional

characterization of MaSCs would also allow more specific isolation and identification of

these cells with the possible localization of these cells within intact mammary ducts. As the

microenvironment plays a key role in the maintenance of stem cells in a variety of tissues it

is likely that determining the precise localization of MaSC would provide insight into the

mechanisms regulating their maintenance and differentiation.

Page 26: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

9

Current methods of isolation of MaSC rely upon varied immunophenotyping markers

to select enriched populations. These markers are not selected based on any specific

biological functions and do not necessarily provide insight into the function of the stem cells

or their regulation. While some work has focused on Hoechst dye efflux as a functional

feature of stem cells based on the proposed ability to pump out toxins, isolation of dye-

excluding cells has not proven effective for enriching stem cell populations [56]. The

prospective isolation of a candidate stem cell population based on a putative characteristic of

those cells would offer the ability to study a viable population with known stem cell

behavior.

The lineage of the cells composing lobulo-alveolar outgrowths also remains unclear.

While putative alveolar progenitors have been identified, the exact contributions of these

cells to the differentiated structures have not been defined. Identifying the source of the cells

comprising the alveoli in both cell layers would provide crucial information about the

developmental pattern of alveologenesis and pregnancy induced differentiation. By tracing

the lineage of the cells composing both the luminal and basal layers of alveoli, the hierarchy

of MECs could be more clearly defined.

1.2 Breast Cancer

Breast cancer is the most commonly diagnosed cancer in women in the Western

hemisphere and the second leading cause of cancer-related death in females. Though

malignant breast disease is very common, with a lifetime risk approaching 1 in 8 women,

breast cancer is a very heterogeneous disease with numerous possible etiologies. The

extensive heterogeneity in both causation and tumor characteristics complicates the study and

treatment of breast cancers.

Page 27: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

10

Ductal carcinoma of the breast develops through a series of steps and is characterized

by proliferation of malignant epithelial cells within lobules and ducts of the mammary gland.

The traditional understanding of progression of ductal carcinomas begins with benign

overgrowth in a focal lesion known as hyperplasia or dysplasia, then progresses to ductal

carcinoma in situ (DCIS), and finally to invasive ductal carcinoma (IDC) [57-60]. DCIS is

commonly defined as an overgrowth of epithelium that is contained within an intact

basement membrane, whereas IDC is diagnosed when the malignant growth has breached the

basement membrane. IDC is the most common type of breast cancer. Breach of the basement

membrane and subsequent invasion into the surrounding tissue allows the metastasis of

breast cancer cells to distant sites within the body.

1.2.1 Types of Breast Cancers

The heterogeneity of breast cancers renders the selection of treatment methods more

difficult due to the assortment of clinical and histological forms. Cancers of the breast are

traditionally categorized based on the presence or absence of hormone receptors, histological

appearance, and amplification of HER2/Neu/ERBB2 (hereafter, referred to as HER2). These

methods of identification are commonly used in clinical settings for the determination of

treatment protocols. Recently, transcriptional profiling methods of characterizing mammary

tumors have been developed. Though not yet commonly used in a clinical setting, these

profiles offer greater insight into the precise molecular changes occurring within the tumor.

1.2.1.1 Traditional Characterization of Breast Cancers

In the clinic, tumors are categorized based on the presence or absence of the Estrogen

Receptor (ERα + vs. ERα

-) and Progesterone Receptor (PR

+ vs. PR

-) then by histological

Page 28: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

11

appearance as luminal, basal, and HER2-like. Luminal and basal breast cancers are

diagnosed based on the appearance of tumor cells recapitulating the normal cell layers of the

mammary gland and the molecular profile of each group’s similarity to the normal cell type,

while ERBB2+/HER2 tumors are marked by the over-expression or over activation of the

ErbB2 receptor. Luminal and basal character of tumor cells are also determined by

immunohistochemical staining and HER2 status as determined by FISH.

1.2.1.1.1 Hormone Receptor Positive Breast Cancers

In many mammary tumors, the interaction of hormones and hormone receptor-

positive cells capable of responding to hormonal binding provides a critical stimulus driving

over growth of epithelial cells. Approximately 80% of human breast cancers are ERα +, with

2/3 of those also being PR+. The luminal sub-type of breast cancer is usually characterized

as expressing ERα and/or PR and is Keratin 8/18 positive.

Anti-hormonal therapies, treatments which lower the amount of estrogen or decrease

the activity of the ERα, have provided a targeted method to slow tumor growth in hormone

dependent cancers. Three categories of anti-hormonal therapies are commonly used;

aromatase inhibitors, SERMs (Selective Estrogen Receptor Modulators), and ERDs

(Estrogen Receptor Downregulators). ERDs block ERα action throughout the body by

causing its degradation [61, 62]. SERMs, such as tamoxifen, function by competitive

inhibition of the ERα, preventing binding of estrogen to the ERα and can be used in both pre

and post-menopausal women [63]. SERMs can have mixed agonist/antagonist activity. For

example, Tamoxifen can activate ERα in extramammary tissues and is not exclusively anti-

estrogenic. Aromatase inhibitors prevent the enzyme aromatase from converting androgens

into estrogen in peripheral tissues. Ovarian hormone production is largely aromatase-

Page 29: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

12

independent, therefore aromatase inhibitors are not effective in reducing estrogen levels in

pre-menopausal women. While aromatase inhibitors are effective treatments against ERα

positive breast cancers, the long-term effects of estrogen deprivation limit their use in many

patients and treated tumors often become hypersensitive to exogenous estrogen [64, 65].

1.2.1.1.2 Hormone Receptor Negative Breast Cancers

Two subsets of breast cancer that do not express high levels of ERα have also been

described. Basal-like, named because of its similarity to normal basal/myoepithelial cells of

the mammary gland and HER2/ERBB2/Neu named for the characteristic high expression of

the gene product of the same name. Basal-like breast cancers typically express keratins 5/6

and 14, EGFR are ERα -/PR

-/HER2

- , (so-called “triple-negative”) [66-68]. Not all basal-like

breast cancers are triple-negative tumors as ER expression has been seen in 5-45% of basal-

type cancers and Rouzier et al. showed HER2 expression in 14% of basal tumors. Anti-

hormonal strategies, such as tamoxifen, are not effective against basal breast cancers due to

the lack of ERα expression. In the absence of targeted therapies for these tumors, traditional

cytotoxic agents, such as chemotherapy and radiation, remain mainstays for treatment of

triple-negative tumors.

1.2.1.1.3 HER2/ErbB2/Neu Initiated Breast Cancers

HER2 is a member of the EGF family of receptor tyrosine kinases [69]. It is a

transmembrane receptor that signals through the Ras/Mek/ERK pathway and the PI3K/Akt

pathway to drive proliferation and inhibit apoptosis in the mammary gland [70]. Several

modes of HER2 dysregulation have been proposed to drive tumorigenesis. The expansion of

copy number of the gene, increased translation of HER2/Neu mRNA, increased cell surface

receptor expression, and alterations in the extracellular domain of the receptor conceivably

Page 30: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

13

all provide an oncogenic stimulus [71-75]. Of these modes of HER2 pathway activation,

gene amplification is the most clinically important as amplification detected by fluorescent in

situ hybridization (FISH) predicts response to drugs that target HER2. Alterations in the level

and activity of the HER2 gene product result in increased and deregulated signaling to the

downstream pathways [76]. A variety of therapeutic strategies have been designed that target

HER2. Most notably, the development of Herceptin (trastuzamab), a humanized monoclonal

antibody that recognizes the ErbB2 receptor, has greatly improved outcomes for patients with

HER2+ breast cancer [77]. More recently, small molecule inhibitors of HER2 kinase

activity, such as lapatinib, have shown promising activity against HER2+ cancers that have

acquired resistance to Herceptin [78].

1.2.1.2 Transcriptional Profiling of Breast Cancers

While the traditional methods of categorizing breast cancers based on histology,

immunohistochemistry, and hormone receptor status are still widely used in the clinical

setting expression profiling of tumors by microarray is providing new insight into the

molecular taxonomy of breast cancer. These expression profiles have identified 5 “intrinsic

subtypes” of breast cancer that bear characteristic gene expression patterns; luminal A,

luminal B, basal-like, HER2, and normal breast-like [79-81]. By examining breast cancers in

each subtype, potential pathways may be targeted more precisely allowing more efficient

treatment [82].

Page 31: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

14

1.2.2 Hierarchy of Cells in Mammary Cancers

Prior to the 1990s, cancers were typically thought of as homogeneous populations of

cells; however, the description of leukemia stem cells opened a window into the more

complicated hierarchical relationship of the cells composing tumors [83-87]. While “liquid”

tumors had defined hierarchies, solid tumors lacked similar understanding of cancer cell

lineage. In 2003, Clarke and colleagues described the isolation of a minority cell population

isolated from breast cancers capable of reconstituting the tumor following implantation [88].

These cells were distinct from the bulk tumor cell population in this tumor-initiating

capability. The prospective isolation of these CD44+/CD24lo/Lin- tumor-initiating cells

paired with the identification of subsets of solid tumor cells capable of giving rise to colonies

in culture provided evidence for a candidate stem cell population within the tumor.

Importantly, tumor-initiating cells gave rise to the cellular heterogeneity seen in the primary

tumor demonstrating multi-potency of the population. While the identification of these

tumor-initiating cells provides insight into the developmental patterns of many mammary

cancers, some cancer subsets have not had any tumor-initiating populations isolated and

appear to be homogeneous.

The variety of mammary cancer phenotypes may be attributable to the plasticity of

these cancer stem cells or due to cell of origin effects. Luminal type cancers do not

necessarily arise from luminal type cells affected by genetic alternations, and similarly, basal

mammary cancers do not necessarily derive from basal precursor cells [89]. It remains

unclear whether MaSC serve as the cell of origin for breast cancers or whether accumulated

genetic insults activate stem cell-like behaviors in a subset of tumor cells responsible for

tumorigenesis. It is possible that many subsets of mammary cancers derive from a single cell

Page 32: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

15

of origin and it is the specific mutations contained within any one tumor which determine the

resulting tumor type. Alternatively, identical mutations in distinct cell populations may result

in different tumor outcomes [90].

The identification of potential cancer stem cells within solid tumors, analogous to

those already described in the hematopoietic cancers provides a new target for the

development of treatments. Increased understanding of the hierarchy contributing to

developing mammary cancer would provide new avenues of attack against breast cancers and

would particularly allow targeted therapies aimed at tumor-initiating cells within the tumors.

Traditionally, therapeutic methods were evaluated on the basis of their ability to decrease

tumor volume and typically target rapidly cycling cells. However, if only a subset of tumor

cells is in fact tumor initiating (and potentially quiescent), and the majority of tumor burden

consists of non-tumorigenic proliferating cells, then targeting instead the small proportion of

cells responsible for tumor formation would prove more beneficial even if not as

immediately effective in reducing tumor bulk. Lineage tracing experiments ought to provide

insight into the cell of origin of mammary tumors and into the interactions between particular

cell of origin and mutation effects.

1.2.3 Breast Cancer and Tumor Dormancy

A relatively common feature of breast cancer progression is the recurrence of local or

metastatic disease following a long interval free of detectable tumor [91, 92]. In addition,

only a small percentage of women develop clinically detectable breast cancers even though

as much as 39% of women aged 40-50 in fact harbor DCIS [93]. These undetected lesions

are presumed to be maintained in a non-invasive state, often for extended periods of time [94,

Page 33: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

16

95]. This lack of tumor progression is inconsistent with the common model of unbridled

proliferation of tumor cells. The concept of tumor dormancy has been offered to provide an

explanation for these long periods of relative quiescence of tumor cells within the mammary

gland or metastatic lesion [91].

While tumor dormancy may help explain how breast cancer patients relapse with

metastatic disease long after eradication of their primary tumors, the mechanisms whereby

this dormancy is maintained remain unclear [96]. Immune cell mediation and the requirement

for angiogenesis have been suggested to maintain tumor latency in metastases and in dormant

primary lesions [97-99]. According to one model, tumor cells frequently disseminate to

distant organ sites, however only a small subset is capable of proliferating within the new site

and driving the formation of macrometastases. Some fraction of the remaining cells are

proposed to enter a period of quiescence. It is tempting to speculate that these dormant cells

arise from tissue-specific stem cells and function as cancer stem cells since they can enter

and exit quiescence and retain latent malignant potential. It remains unclear what causes

these cells to exit dormancy and cause disease recurrence. However, identification of the

cells within these dormant lesions and elucidation of the pathways regulating their

quiescence would provide a clear target for the treatment of dormant breast cancer [100].

1.2.4 Parity-Related Protection from Breast Cancer

Numerous epidemiologic studies have identified an array of hormonal and

reproductive factors that affect the risk of developing breast cancer. Among these, the breast

cancer risk reduction afforded by an early first full-term pregnancy is least well understood.

Younger age at first pregnancy is associated with a significant decrease in lifetime risk of

breast cancer and multiparous females also have a decreased risk of developing breast cancer

Page 34: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

17

in comparison to nulliparous females [67, 101-105]. This reduction in risk is not

immediately apparent following pregnancy but first manifests several (3-4) years later. One

proposed etiology attributes parity-induced protection to the differentiation of the mammary

ductal tree that occurs during pregnancy and lactation, although the precise cellular and

molecular mechanisms that reduce the risk are undefined [106, 107]. Another model

attributes protection to parity-induced up-regulation of the p53 tumor suppressor pathway

[108-110].

1.3 Mouse Models of Human Breast Cancer

The similarities between the human breast and murine mammary glands as well as human

breast cancers and murine mammary tumors allow animal modeling of breast cancer. Mouse

models of cancer have greatly aided the progress of our understanding of the biology of

development and tumorigenesis by allowing a variety of experimental strategies to be

explored while relying on the homology between the model system and human conditions.

The developmental and genetic homology between humans and mice have allowed the

development of transgenic, knock-out, and chimeric mice to be used to study the role of

many gene products in the mammary gland.

Embryonic development of the mammary gland is highly conserved among mammals.

The rudimentary ducts develop from the ectodermal milk line remain growth arrested until

puberty. While humans typically only develop one pair of nipples and mammary glands,

located on the chest wall, mice typically develop 5 pairs located ventro-laterally [111].

Additionally, the human mammary gland ductal structure is somewhat more complex than

that of the mouse gland. Human ducts divide into segments that feed individually into the

nipple, rather than the single ductal tree that is found in the mouse [111]. The cellular

Page 35: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

18

composition of the mammary ductal tree is similar within both human and murine ducts

consisting of an epithelial bilayer of luminal cells surrounded by myoepithelial (basal) cells.

Moreover, the cell layers exhibit similar cytokeratin expression patterns in both species [26,

112]. Unlike the cellular composition of the ducts, the cellular components of the stroma are

divergent. In the human breast, connective tissue supports the ductal tree while in the mouse

mammary gland, fat composes the majority of the stroma [113, 114].

Both the human breast and murine mammary gland respond to the cyclical hormonal

stimulation of the menstrual cycle or estrus with transient proliferation within the mammary

ductal structures [115, 116]. Prior to pregnancy in the mouse, very little lobuloalveolar

branching is seen [111]. In contrast, in the human breast, the corresponding terminal ductal

lobular units (TDLU) commonly develop prior to the onset of pregnancy, due to the

hormonal cycling induced by the menstrual cycle [111, 113]. Notwithstanding these

preexisting differences, the murine and human mammary glands respond similarly to the

hormones of pregnancy and lactation, and weaning induces similar involution via apoptosis

within the alveolar structures [117].

1.3.1 Virally-Induced Mouse Mammary Cancer and the Discovery of the Wnt1 Oncogene

The biology of the mammary gland and the study of mammary oncogenesis was

greatly advanced by the molecular characterization of the Mouse Mammary Tumor Virus

(MMTV) insertion sites in the early 1980s. Though the presence of a “milk factor”

responsible for the vertical transfer of a malignant tendency in mice had been proposed in the

1930s by Bittner, the causative agent remained unidentified [118-122]. MMTV was

identified as the viral agent responsible for the vertical transfer of mammary tumorigenesis in

mice in 1948 and the virus family was further characterized in the 1960s [123-126]. The

Page 36: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

19

wnt1 oncogene was discovered by screening of mouse mammary tumors for oncogene

activation following infection with MMTV [127]. First named int1, the gene name was later

changed to Wnt1 after the genetic homology to the previously described wingless (wg) gene

in Drosophila was noted [128]. Insertion of MMTV into the Wnt1 gene occurs in nearly 70%

of MMTV-infected mice [127]. Less commonly, other oncogenes can be activated by

MMTV insertion, including other members of the Wnt family and members of the fibroblast

growth factor family [127, 129-132].

Wnts are secreted signaling proteins that interact with cell-surface Frizzled (Fz)

receptors [133, 134]. The binding of Wnt to Fz initiates a signal transmitted to dishevelled

(Dsh) [135] . Dsh in conjunction with Axin, in turn negatively regulate GSK3 (Glycogen

synthase kinase 3β) which normally phosphorylates β-catenin and targets it for degradation

[136-138]. The inhibition of GSK3 kinase activity allows levels of β -catenin to build up

within the cytosol and subsequently translocate to the nucleus where it can act as a

transcriptional cofactor and activate transcription of Wnt target genes, leading to cell growth

and proliferation [139, 140]. Accumulation of cytosolic β-catenin is opposed by the so-

called “destruction complex” which includes well-known tumor suppressors such as APC

(adenomatous polyposis coli) and Axin [141, 142].

1.3.1.1 Engineering Constitutive MMTV-Driven Expression of Oncogenes in the Mammary

Gland

Originally constructed in 1988 by Tsukamoto and colleagues, the commonly used

transgene for the constitutive expression of Wnt1 contains a genomic fragment harboring a

portion of the Wnt1 promoter and its complete coding sequence fused to an MMTV-LTR

(Mouse Mammary Tumor Virus Long Terminal Repeat), which recapitulates a common

Page 37: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

20

insertional mutagenesis event [143]. The MMRV-LTR is responsive to hormone stimulation

and therefore drives expression that is upregulated during puberty and pregnancy [144].

Expression of Wnt1 via this construct leads to mammary gland ductal hyperplasia even prior

to puberty in female mice [145-147]. The ductal hyperplasia precludes female MMTV-Wnt1

transgenic mice from successfully nursing their offspring presumable by disrupting milk

production or delivery. The mean latency of tumors in female mice bearing the MMTV-

Wnt1 transgene is approximately 6 months with breeding mice developing tumors somewhat

earlier than virgin mice [143, 147]. Male Wnt1 transgenic mice also develop mammary

hyperplasia and, occasionally, mammary tumors [143, 148].

1.3.2 The Role of Wnt Signaling in Mammary Development and Breast Cancer

Nineteen Wnts have been identified in mammalian genomes and have been shown to

be potent regulators of proliferation and differentiation in both embryonic and adult tissues

[149-151]. In particular, Wnt signaling is required for the initiation of mammary

morphogenesis in the embryo, is localized along the mammary lines in the developing

embryo, and helps orchestrate ductal arrangement in the mammary gland [3, 149, 152, 153].

In the adult mammary gland Wnt-4, 5a, 5b, 6, and 7b are expressed [154]. Wnt-4 in

particular has been shown to be necessary for proper branching of the mammary ducts during

early pregnancy [155]. It has additionally been demonstrated that progesterone is required

for and induces the expression of Wnt-4 during pregnancy [155]. Based on these

observations, Wnt signaling appears to play a role in both initiation of mammary gland

developments and also in the subsequent maintenance of structural patterns in the epithelial

tree.

Page 38: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

21

Wnt signaling impacts both embryonic and adult development of the mammary gland,

since abrogation of the Wnt disrupts the normal developmental patterns of proliferation.

These findings suggest that Wnt family members may play a role in the regulation and

maintenance of the MaSC population [3, 155-157]. Transgenic activation of Wnt1 increases

both the population of cells carrying the stem cell immunophenotype and also the population

of cells capable of reconstituting a mammary gland following implantation [27, 158].

Additional support for Wnt-mediated regulation of stem cell behavior in the mammary gland

is derived from the characterization of Wnt1-driven mammary tumors. Transgenic tumors

from these mice have cells of both luminal and basal character and populations within these

tumors express Keratin-6 and Sca-1, both markers of stem cell-like character [159, 160].

The effects of over- and under-expression of Wnt on stem cell and progenitor populations in

the mammary epithelium support a role of Wnt signaling in the regulation and maintenance

of mammary stem cells.

Viral oncogenesis plays no known role in the development of human breast cancers

and though the importance of Wnt as an oncogene in the murine mammary gland has been

well defined, human breast cancers resulting from Wnt1 mutations have not been identified

[161]. Nonetheless, Wnt pathway mutations have been demonstrated to transform primary

human MECs in culture and autocrine Wnt signaling drives proliferation in some breast

cancer cell lines [162, 163]. Overexpression of WNTs-2, -4, -5A, -7B, -10B, and 13 have

been identified in samples from human breast cancers [164-167]. While mutation of

components of the Wnt signaling pathway does drive tumorigenesis in other human tissues

(most notably in colon cancers where 85% of tumors have lost function of APC), similar

mutations have not been identified in the human breast [141]. Overexpression of β-catenin is

Page 39: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

22

correlated with poor prognosis in breast cancers. Aberrant expression of β-catenin has been

identified in breast cancers and suggests altered Wnt signaling but mutations in Axin and β-

catenin were not found [168, 169]. The Wnt-inhibitory Factor, WIF1, has been reported to

be inactivated in cancers of the breast, prostate, and bladder by epigenetic means [170, 171].

The inactivation of inhibitory WIF may explain why Wnt signaling becomes dysregulated in

these cancers in the absence of Wnt (or downstream pathway) mutations. Though Wnt1 itself

does not appear to drive mammary tumorigenesis, other Wnt family members and the Wnt

signaling pathway in general may play critical parts in the transformation of human

mammary epithelial cells.

1.3.2.1 Inducible Expression of Wnt1 in the Mammary Epithelium of Transgenic Mice

To circumvent the difficulties inherent in constitutively activating Wnt1 oncogene

expression as seen in MMTV-Wnt1 transgenic animals, a transgenic model utilizing Tet

operator driven expression of the Wnt1 oncogene in the mammary gland allows reversible

oncogene expression. Tissue specificity and temporal regulation were achieved by the

construction and pairing of two independently incorporated transgenes [172, 173]. The

transactivator, MMTV-rtTA, contains the MMTV-LTR and drives continuous expression of

the reverse-Tet transactivator (rtTA) in the luminal epithelium of the mammary ductal tree

[172]. The second transgene construct, named TWNT, contains a minimal CMV promoter

upstream of seven Tet-operator sequences followed by the Wnt1 gene, an IRES sequence,

and a luciferase gene. This transgene allows the concurrent expression of the Wnt1

oncogene and the luciferase reporter gene in a tetracycline-dependent manner. Doxycycline

is a tetracycline analog that activates the expression Tet-R containing transgenes and

provides favorable pharmokinetics in vivo.

Page 40: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

23

In the absence of Dox in bitransgenic MMTV-rtTA/TWNT animals, TWNT is not

expressed. However, following administration of Dox via the animals’ diet, Dox binding to

rtTA causes a conformational change in its structure that allows binding to the TetO sites in

the TWNT transgene, thereby initiating transcription of the downstream sequences.

Subsequent removal of Dox stimulation abrogates TWNT transgene expression. Dox-

regulated expression of the TWNT transgene results in ductal hyperplasia in all mammary

glands [173]. Sustained Dox treatment resulted in the stochastic development of mammary

tumors, primarily adenocarcinomas, in female MMTV-rtTA/TWNT mice, suggesting a

multi-step tumorigenesis. Single transgenic mice carrying only the transactivator or the

TWNT transgene did not develop mammary tumors even during continued Dox treatment.

Similarly, Dox-naïve MMTV-rtTA/TWNT animals did not develop mammary tumors

demonstrating Dox-dependence of tumorigenesis in this model. Similar to the findings in

MMTV-Wnt1 transgenic mice, bitransgenic tumor-bearing mice induced with Dox were

sometimes found to carry lung metastases [173].

1.3.2.1.1 Minimal Residual Disease Lesions Modeled in Transgenic Mice

In addition to allowing the conditional activation of the Wnt1 transgene in order to

limit the hormonal impacts on MMTV-driven expression of the oncogene, the reversible

model of Wnt1 expression mimics the effect of a targeted therapy. Removal of Dox from

the diet terminates expression of the TWNT transgene, mimicking the effects of a

chemotherapeutic that specifically counteracts activation of wnt1. Removal of Dox

stimulation results in the nearly complete regression of mammary tumors generated in

MMTV-rtTA/TWNT animals. However, a small proportion of tumor cells remain as a

Minimal Residual Disease (MRD) lesion. These lesions remain dormant for extended

Page 41: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

24

duration and retain latent malignancy [173, 174]. Interestingly, metastatic lesions identified

in these animals were also reliant on continued TWNT expression and regressed following

Dox withdrawal. A fraction of the MRD lesions in mice maintained off Dox spontaneously

formed recurrent tumors after several months, paralleling the findings on dormant breast

cancer in women.

The MRD lesions contain cells capable of giving rise to mammary ducts with the

capacity to repopulate a normal mammary gland and respond to the hormonal signals critical

for the regulation of mammary development following implantation into the mammary fat

pad [174]. However, ductal outgrowths from MRD fragments retained latent malignancy and

rapidly redeveloped into mammary adenocarcinoma following reactivation of Wnt signaling.

Ductal outgrowths derived from the Wnt-induced, but non tumorigenic surrounding ducts did

not retain latent malignancy and did not rapidly develop tumors following readministration of

Dox. MRD-derived outgrowths, when induced to form tumors, contained cells of both

luminal and basal appearance suggesting that the cells contained within the lesion have the

capacity to give rise to cells of both epithelial lineages within the mammary ducts.

This murine model of human breast cancer derived MRD offers potential avenues for

the investigation of MRD development and maintenance. However, while the MRD

developed in this model system have been shown to retain both latent malignancy and some

multipotency in relation to mammary epithelial development, these conclusions were reached

on the basis of fragments of tissue and do not offer insight into the capabilities of single cells

within the lesion. Additionally, the study of MRD in the MMTV-rtTA/TWNT system to date

also does not examine the mechanisms of maintenance of tumor dormancy following the

abrogation of Wnt signal. Finally, the MRD studied by Gestl et al, model only those cancers

Page 42: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

25

initiated by Wnt1 expression and as human breast cancer is caused by a diverse range of

genetic and epigenetic causes may not accurately model all or even most human MRD

lesions.

1.3.3 Constitutive MMTV-Neu Driven Tumorigenesis in the Mouse Mammary Gland

Overexpression and aberrant activation of HER2 are associated with breast cancer in

humans. In order to model the effects of HER2 activation in mouse mammary cancer, a

transgenic mouse line expressing an activated rat HER2 allele driven by the MMTV

promoter was engineered. These mice rapidly develop mammary tumors, and this process is

accelerated by pregnancy [175, 176]. In contrast, mice carrying a transgene utilizing

MMTV-driven expression of wild-type (unactivated) HER2 develop mammary tumors only

after an extended latent period [75, 175]. Unlike the multiple cell types comprising MMTV-

Wnt1 driven mammary tumors, MMTV-Neu tumors have a homogeneous luminal epithelial

composition [75, 177]. MMTV-Neu tumors did not show expansion of the stem cell

population by immunophenotyping assays, suggesting a different target cell for oncogenic

transformation [27].

1.3.4 Compartment-Restricted Expression of Transgenes

The mammary gland ductal tree is composed of two epithelial layers. These layers

have distinct appearance, function, and characteristics. Both layers of the mammary ducts

contribute to the morphogenesis of the mammary tree. To elucidate the role of each cell

compartment in the development of the mammary gland, transgenes allowing specific

activation limited to the compartment of interest were developed. By using Tet-operator

Page 43: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

26

constructs driven with cell-type specific promoters, both compartment-specificity and

temporal regulation of TetO-regulated transgenes can be achieved.

Luminal-specific expression in the mammary gland is driven by MMTV-LTR. The

MMTV-LTR is hormone sensitive, and its expression increases greatly during pregnancy and

lactation but is present during the entire development of the mammary ductal structure [143-

145, 178]. While expression of reporter transgenes driven by MMTV-LTR is uniform in

younger animals, expression in the ducts of older animals is patchy but widespread [179].

Pairing the MMTV-rtTA transactivator with TetO responder transgenes results in the luminal

MEC-restricted expression of the responder gene product in a Dox-dependent manner [172].

Basal-specific expression in the mammary gland can be driven by the Keratin 5

promoter. Keratin-5 expression is common in basal epithelial tissues throughout the body

including those within the skin and digestive tract [180]. While Keratin-14 is also expressed

in the cells of the basal layer of the mammary ducts, extended expression of genes driven by

this promoter results in labeling of both luminal and basal cells [181]. As both K5 and K14

are widely expressed in a variety of tissues, expression of downstream reporters is not limited

to basal MECs but is also seen in the skin and other epithelial tissues [180]. K5-driven

expression within the mammary ductal tree however is limited to the basal layer of the

epithelial bilayer. By pairing the K5 promoter with rtTA, TetO bearing transgene constructs

can be expressed in the basal compartment in a temporally regulated manner.

1.3.5 Inducible Expression of Fluorescently Tagged Histone H2B

The TetO-H2B-eGFP transgene enables Dox-dependent fluorescent labeling of

chromatin and was developed to study stem cell localization and dynamics in the skin [182].

Page 44: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

27

The H2B-eGFP transgene was engineered by fusing human histone H2B to the modified

Aequorea Victoria gene for green fluorescent protein (eGFP) [183]. When expressed, the

fluorescently labeled histone is stably incorporated into the chromatin of living cells [183,

184]. Fuchs and colleagues adapted the H2B-eGFP transgene to allow Dox-dependent

regulation of gene expression in the skin by pairing it with a K5-driven Tet-off element

[182]. This modification allowed the temporal and cell-type specific regulation of

fluorescence expression with a K5-driven transactivator. While maintained on Dox, GFP

was expressed throughout the skin, however following removal of Dox, H2B-eGFP

expression ceased. Subsequent proliferation of labeled cells diluted chromatin-bound H2B-

eGFP in cycling cells while relatively quiescent cells retained brighter H2B-eGFP signal

allowing identification of a slow-cycling stem cell population within the skin. Reversible

expression of the H2B-eGFP transgene allows the tracking of fluorescently labeled cells in

both tissue samples and in vitro systems.

1.3.6 Experimental Manipulation of the Mouse Mammary Gland

The large number of inbred mouse strains and genetically altered lines along with the

relative ease of creating novel genetic alterations make mice an attractive model system for a

variety of developmental disease-related questions. The ventral subcutaneous location of

murine mammary gland permits a number of experimental and surgical manipulations. The

technique developed by DeOme and colleagues of clearing a fat pad and subsequent

implanting experimental tissue is of particular importance for the study of the mammary

gland [43, 44]. The relatively quiescent state of the mammary gland from birth until the

onset of puberty provides a window of time in which endogenous epithelium can be excised

from the mouse while preserving a portion of the intact stroma in the form of the mammary

Page 45: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

28

fat pad. This surgery can be paired with the implantation of small fragments of mammary

gland or injection of isolated epithelial cells of interest into the empty fat pad. The implanted

tissue or cells then can elaborate an entire ductal tree within the typical extracellular milieu

of the mammary gland. This combination of host stoma and donor epithelium, often using

donors with distinct genetic features, has provided evidence that the mammary gland

contains a population of cells that are capable of giving rise to the entire mammary ductal

tree. This capability of the proposed mammary stem cell population has been shown to

extend over several successive generations of mammary gland implants [27-29, 43, 44].

Additionally, cleared fat pad and implant surgeries have allowed the study of genetically

altered epithelium under conditions that the transgenic donor may not have supported, due to

non-specific activation of transgenes. Finally, interactions between the tissue types,

epithelial and stromal, from distinct genetic backgrounds may be studied to further elucidate

the role of compartmental interactions in the development and function of the mammary

gland. Another technique of particular benefit to mammary cancer researchers is the facility

of explanting samples of mammary tumors onto the flanks of syngeneic or

immunocomprommised host mice. This technique allows the expansion of tumor tissue and

the study of a variety of conditions and treatment on clonally related tumors.

1.4 Areas Addressed in this Dissertation

In this dissertation, new mammary gland biology techniques and reagents are described

including:

Methods for Compartment and Temporal Regulation of Transgene Expression

in the Mouse Mammary Gland

Methods for Inducible Expression of H2B-eGFP in Normal and Neoplastic

Mammary Epithelium

Page 46: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

29

We employed these strategies to measure

Proliferation Dependent Washout of H2B-eGFP Signal

Proliferation Dynamics of Constitutive Wnt-initiated Mammary Tumors

Proliferation Dynamics of Reversible Wnt-initiated Mammary Tumors and

Minimal Residual Disease Lesions

Finally, we adapted our labeling strategy to perform

Short-Term Lineage Tracing of MECs During Pregnancy and Involution

Page 47: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

30

Figure 1.1 Basic Anatomy of the Mammary Gland.

The panel depicts a carmine-stained mammary gland whole mount from a 4-week old female

imaged by wide field microscopy. The hormones of puberty drive the elongation of the

mammary ducts from the nipple to the distal end of the mammary fat pad. Club-like TEB are

indicated at the distal tips of the developing mammary ductal tree. The clearly visible lymph

node provides a convenient landmark in the orientation of 4th

inguinal mammary gland in the

mouse.

Page 48: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

31

Figure 1.2 Secretory Differentiation and Cyclical Remodeling of the Mammary Ductal Tree.

Top panels depict carmine-stained mammary gland whole-mounts collected at the

developmental stage indicated and visualized by widefield microscopy. The lower panels

depict microscopic views of H&E stained from the same mammary glands. The adult virgin

mammary gland contains an arborized epithelial tree penetrating a supporting pad of stroma

and fat. At 15 weeks virgin development, relatively few side branches present and alveolar

outgrowths are absent. The surge in hormones corresponding to the initiation of pregnancy

results in proliferation of MECs driving the formation of lobulo-alveolar outgrowths (d6

pregnancy). Secretory differentiation is indicated by filling of alveoli with lipid droplets (d18

pregnancy). Terminal differentiation with maximal milk production results in markedly

distended alveoli (d9 lactation). Cessation of suckling leads to apoptosis-mediated

elimination of alveolar outgrowths, remodeling of the ductal tree, and the restoration of the

mammary ductal appearance to one resembling that seen in virgin animals. The arrow

indicates that the mammary gland is competent to undergo multiple rounds of proliferation,

differentiation, and involution cycle in response to hormonal stimulation.

Page 49: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

32

Figure 1.3 Schematic of Bilayered Mammary Ducts and Keratin Expression Patterns.

Mammary ducts are comprised to bilayered epithelium. The inner, luminal layer, is

composed of cuboidal cells that characteristically express Keratin-8. The outer,

myoepithelial (basal) layer, is made up of epithelial cells that develop contractile capabilities

and express Keratin-5. (Adapted from Grimm et al, Development 2005 and Teuliere et al,

Breast Cancer Research 2006 [185, 186])

Page 50: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

33

Figure 1.4 A Proposed Hierarchy of Mammary Epithelial Cells.

A mammary stem cell (MaSC) has been identified and functionally characterized in

implantation studies, spurring investigation into the hierarchical organization and pathways

of lineage commitment for mammary epithelial cells. Briefly, the MaSC is proposed to give

rise to a common bipotential progenitor which in turn differentiates into luminal and

myoepithelial committed progenitors. The myoepithelial progenitor, gives rise to

differentiated myoepithelial cells in the mammary ducts, but remains unidentified. The

luminal progenitor population differentiates to produce the cells of the luminal layer in

mammary ducts. The luminal progenitor is also proposed to give rise to the alveolar

progenitor which in turn produces the cells of the alveoli during pregnancy. The alveolar

progenitor is proposed to contribute to the luminal and basal layers of the alveolar

outgrowths, potentially via the putative myoepithelial progenitor. Reported

immunophenotypes for each cell population are included to illustrate the pathways of lineage

commitment followed by each population. (Adapted from Visvader, Genes and

Development, 2009 [187].)

Page 51: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

34

Chapter 2 Materials and Methods

2.1 Mouse Strains, Surgeries and Drug Treatments

2.1.1 Housing and Maintenance

All mice utilized in this study were bred and housed in a barrier facility. Animals

were fed a standard diet of Harlan Tekland chow pellets (HT2018) and had unlimited access

to drinking water. Mouse strains received in the FVB background were maintained in this

background by crosses to other FVB animals. In order to test transgene inducibility in

MMTV-rtTA/TGFP mice, TGFP males of a mixed-strain background were crossed to FVB

females carrying the MMTV-rtTA transgene. Before generating MMTV-

rtTA/TGFP/MMTV-wnt1 tri-transgenic animals to examine the compartment specific

labeling of mammary tumors, the TGFP transgene was backcrossed for 8 generations onto an

FVB background. Wild-type nontransgenic animals used for implant and explant studies

were also FVB. Some FVB animals utilized in this work were obtained from Jackson labs

while others were bred within the Gunther Lab barrier colony. Experimental animals were

sacrificed by CO2 inhalation to allow for the collection of mammary tissue. Lungs and liver

were examined in tumor-prone and tumor-bearing animals for the presence of metastases at

necropsy.

Transgenic lines used in this study were obtained as follows:

MMTV-rtTA: gift from Lewis Chodosh, University of Pennsylvania

K5-rtTA: gift from Adam Glick, the Pennsylvania State University

TGFP: Jackson Labs; STOCK Tg(tetO-HIST1H2BJ/GFP)47Efu/J

TWNT: gift from Lewis Chodosh, University of Pennsylvania

MMTV-wnt1: Jackson Labs; FVB.Cg-Tg(Wnt1)1Hev/J

Page 52: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

35

MMTV-neu: Jackson Labs; FVB/N-Tg(MMTVneu)202Mul/J

2.1.2 Genotyping and Breeding Schemes

All transgenic animals were genotyped by PCR analysis of tail-clip DNA using

primer pairs specific for each transgene as listed in Table 2.1. Genomic DNA was prepared

from fresh or fresh-frozen tail biopsies by the Maxwell 16 instrument (Promega) or by

standard methods. Briefly, the non-mechanized method consisted of overnight digestion of

tail biopsy samples in Proteinase K followed by homogenization, extraction, and

precipitation steps. DNA pellets obtained by this method were resuspended in water. DNA

prepared by the Maxwell machine was resuspended in a proprietary buffer.

The MMTV-rtTA and TWNT transgenes were maintained as hemizygous due to

impaired fertility of homozygous animals (data not shown). Parent of origin was tracked but

did not appear to impact transgene expression or transmission to offspring. Most crosses

involving the MMTV-wnt1 transgene employed transgenic males since transgenic females

show a notable deficiency in lactation that precludes the efficient recovery of pups from these

dams. Fostering of pups from MMTV-Wnt1 dams onto wild-type lactating females shortly

following parturition was occasionally used as needed to obtain transgenic animals. MMTV-

neu transgenic animals were maintained as homozygotes in order to facilitate breeding and

genotyping and were crossed with MMTV-rtTA and TGFP bearing mice as needed.

2.1.3 Experimental Manipulations of Mice

Animals were placed on Dox chow to induce transgene expression. Dox diet was

prepared by Bio-Serv (Frenchtown, NJ) at 2.0 g Doxycycline/kg and fed ad libitum. Animals

Page 53: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

36

removed from Dox diet were isolated from Dox-naïve animals for a minimum of one week to

avoid transfer of Dox from skin or feces.

2.1.3.1 Surgical

Animals were anesthetized with inhaled isofluorane for all surgical procedures.

Surgical sites were shaved using grooming clippers in preparation of incision and all skin

incisions were closed with stainless steel wound clips. Mammary gland biopsies were

typically performed to collect the 4th

inguinal mammary gland from either the left or right

side. A skin incision was made to expose the mammary fat pad which was then separated

from the skin and peritoneal surface using both blunt and sharp dissection. Electrocautery

was used for hemostasis. Cleared fat pad surgeries were performed by excision of the

proximal, nipple-near end of the #4 mammary fat pad up to and including the lymph node of

mice prior to the onset of puberty (<4 weeks of age) removing the nascent mammary ductal

tree [45]. Fragments of mammary tissue approximately 1mm2 were inserted into the

remaining distal segment of the fat pad. Fragments were derived from freshly biopsied MRD

lesions. Following surgery, implants were allowed to mature for 10-12 weeks prior to assay.

Tumor size was monitored twice weekly with calipers. Tumor biopsies were

performed when tumors grew to approximately 1 cm in diameter. The skin was dissected

away from the tumor and approximately half the tumor volume was excised and collected.

Following biopsy, tumor-bearing mice were maintained on Dox as specified to verify

ongoing tumor growth and subsequently removed from Dox diet as indicated in each

experiment.

Tumor explants were propagated by explanting small fragments of mammary tumor

collected at biopsy or necropsy. Fresh tumor fragments approximately 3 mm in diameter

Page 54: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

37

were implanted subcutaneously on the dorsal flank of wild-type host animals. Tumor

explants dependent on Dox were explanted onto the flanks of Dox-induced host females.

Tumor-bearing hosts were subjected to Dox pulse and Dox withdrawal as indicated.

Hormone pellets were implanted subcutaneously onto the dorsal flank of adult virgin

females. 21-day release Estrogen pellets (E 0.025 mg/pellet) and Progesterone pellets (P

25.0 mg/pellet) were purchased from Innovative Research of America (Sarasota, Florida).

2.1.3.2 Chemotherapeutics

MMTV-rtTA/TWNT/TGFP tritransgenic mice were occasionally treated with MNU

(methylnitrosourea) at 6 weeks of age in order to accelerate tumorigenesis in the mammary

gland. Animals were anesthetized with inhaled isofluorane prior to administration of MNU

at a dose of 50 mg/kg body weight. Mice expressing Wnt1 transgenes typically develop

multiple synchronous tumors in the mammary glands following MNU treatment. Tumor

explants bearing animals were treated with a single 300 mg/kg dose of Adriamycin,

administered i.p..

2.1.3.3 Pregnancy

To generate timed pregnancies, adult females were bred and monitored daily for the

presence of vaginal plugs. The plug day was reported as d 0.5 of pregnancy. Embryos were

collected at necropsy and used to verify gestational age. In our early studies of pulse-chase

kinetics during pregnancy, mammary gland biopsies were performed on d 0.5 pregnant

animals to document H2B-eGFP labeling during pulse. However, plug-confirmed pregnant

mice subjected to an early biopsy frequently lacked intrauterine embryos at necropsy,

presumable because of fetal loss. Since surgical intervention early in pregnancy appeared to

Page 55: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

38

negatively impact gestation, our protocol was modified to eliminate mammary gland biopsies

on pregnant animals. Instead, pulse labeling was confirmed by analyzing mammary glands

from additional control mice that were sacrificed and necropsied and the conclusion of Dox

pulse.

2.2 Mammary Gland and MEC Sample Processing

Mammary gland whole mounts were routinely prepared on glass slides following

excision. Following a brief period to allow air-drying of the sample and securely affix the

tissue to the slide, samples were immersed in 4% PFA for 2 hours on wet ice. Whole mounts

were subsequently transferred to low PFA (0.4%) for extended storage at 4 C. Tissue

fragments from mammary tumors and normal mammary glands were snap frozen on dry ice

and then transferred to microcentrifuge tubes for long term storage at -80 C. Additional

tumor pieces were fixed in 4% PFA on ice and subsequently stored in low PFA at 4 C.

To obtain mammary gland frozen sections, mammary gland whole mounts were

removed from slides following fixation and storage, then snap frozen on dry ice and

sectioned by microtome. Mammary tumor sections were cut from tumor fragments snap

frozen on dry ice and stored at -80 C. Thick sections (~35 um) were obtained and stored at -

20 C prior to staining in order to limit desiccation of the samples.

2.2.1 Single Cell Suspensions

To prepare single cell suspensions from mammary tumors and MRD lesions, both

mechanical and enzymatic dissociation were employed as previously described [27].

Approximately 5 mm2 fragments of mammary tumor were finely minced by hand or by

McIlwain tissue chopper. This slurry was transferred to a conical tube and incubated with

Page 56: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

39

300 U/mL collagenase (Sigma) and 100 U/mL hyaluronidase (Sigma) in DMEM/F12

(Gibco) for 1 hour at 37C in an orbital shaker at 100 rpm. Samples were pelleted for 8

minutes and the supernatant discarded. The resulting pellet was resuspended in prewarmed

(37 C) 0.25% Trypsin-EDTA (Gibco) and agitated vigorously by shaking for 4 minutes and

pelleted again. The supernatant was aspirated and the pellet resuspended in prewarmed 1X

PBS + 40 uL of Deoxyribonuclease I (DNase, 2.3 U/ L, Worthington) and Dispase II

(Neutral Protease, 5 mg/mL, Roche) for 3-5 minutes. Samples were again pelleted and were

resuspended in 0.64% Ammonium Chloride in order to lyse red blood cells. Samples were

pelleted a final time and resuspended in cold 1X PBS, filtered through a single cell filter (40

um, Falcon) and placed on ice prior to further analysis.

To prepare single cell suspensions of MECs from normal mammary gland, up to 8

mammary glands were collected from a single animal (1, 3, 4, and 5 mammary glands from

both sides) and mechanically dissociated by McIlwain tissue chopper. Samples were

digested as in tumors with the following modifications. During enzymatic digestion,

Collagenase/Hyaluronidase (Stem Cell Technologies, Vancouver, British Colombia, Canada)

in DMEM/F12 digestion was extended to 6 hours and was incubated in a 37 C water bath

with agitation by moderate shaking by hand every 1-2 hours. Samples were pelleted after all

digestion steps at 550 g at 4 C for 8 minutes. Samples were passed through the single cell

filter immediately following suspension in DNAse/Dispase mixture. The final product of the

digestion steps was further purified by an immunomagnetic method. EasySep® Mouse

Mammary Stem Cell Enrichment Kits (Stem Cell Technologies) were used to remove non-

epithelial cells from the single cell suspensions. Briefly, approximately 106 resuspended

cells were incubated with the EasySep Negative Selection Epithelial Cell Enrichment

Page 57: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

40

Cocktail (25 uL), then the EasySep Biotin Selection Cocktail (50 uL) was added to the

suspension, and finally Easysep Magnetic Nanoparticles (25 ul) were incubated with the cell

suspension. This treatment labels non-epithelial cells with antibodies capable of adhering to

a magnetic column, allowing removal of these tagged cells and enrichment for MECs.

2.2.2.1 Immunophenotyping and Fluorescent Signal Analysis of MECs and Tumor Cells

Cell suspensions were divided into 5 mL Falcon tubes at 0.5-5 million cells per tube.

Unstained and single-antibody stained control cells were included in all flow cytometry runs.

Rat α-Human CD49f antibody (BD Pharmingen) and Alexafluor 647 goat α-rat (Invitrogen)

were added to each tube at 1 uL/sample and incubated on ice for a minimum of 20 minutes

and up to 1 hour. Samples were then centrifuged for 5 minutes at 550 g at 4 C and

resuspended in 1X PBS containing PE rat α-mouse CD24 (BD Pharmingen); 1 uL/sample for

tumor cells and 0.5 uL/sample for normal MECs and incubated at 4 C for a minimum of 1

hour and as long as overnight prior to FACS analysis. Single cell suspensions were

maintained on ice until assay. FSC (Forward Scatter) and SSC (Side Scatter) gates were set

to identify single cells prior to analysis of GFP fluorescence in order to collect 20,000 events.

Single cells not labeled with immunophenotyping antibodies were analyzed on FACS

Calibur in the FL-1 channel alone. FACS Calibur (BD Bioscience, USA) was used to assess

fluorescence in the FL-2 (PE), FL-4 (Alexafluor 647), and FL-1 (eGFP) channels and data

was analyzed using the Cell Quest Pro software (BD Bioscience).

2.2.3 Imaging of Whole Mounts and Sections

Whole mounted samples containing MRD lesions were photographed under both

white-light illumination and hand held long wave UV illumination prior to fixation or

Page 58: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

41

preparation of single cell suspensions. Alternatively, excitation of eGFP was achieved using

a blue LED flashlight and fluorescence visualized using a 508 nm viewing filter when

imaging fluorescent tumor samples as previously described [188]. A Nikon digital camera

was used to collect these images and the intraoperative images of fluorescing mammary

tumors.

Mammary gland ductal GFP fluorescence was detected in fixed whole mounts from

MRD bearing mammary glands and normal mammary tissues by wide-field microscopy

using a Zeiss Axiovision Observer microscope equipped with 5X, 10X, and 20X objective

lenses and an AxioCam MRm camera. Matched exposures were obtained for Dox-naïve,

pulse, and pulse-chase trios. In addition, overexposed images were routinely captured from

Dox-naïve and pulse-chase sample in order to demonstrate the range of GFP signal

detectable in mammary epithelial cells.

2.2.3.1 Staining

Nuclei within thick frozen sections of mammary gland and mammary tumors were

counterstained with Hoechst. Briefly, sections were thawed and dried on the bench top for

10 minutes prior to fixation in -20 C methanol for 10 minutes. Slides were then rinsed 3

times in 1X PBS for 2 minutes. Samples were then incubated with 1X PBS +0.1% Triton X-

100 (Shelton Scientific) for 10-30 minutes to permeabilize the tissue. Samples were then

incubated in Hoechst dye (Hoechst 33342 trihydrochloride trihydrate, Invitrogen, 8 ug/mL)

was incubated on samples at room temperature in darkness for 30 minutes. Slides were

finally washed 2X in 1XPBS +0.1% Triton-X and then 1X PBS 3X with agitation. Slides

were removed from wash, briefly shaken dry and mounted and coverslipped.

Page 59: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

42

AquaPolyMount (Polysciences). Following overnight drying time at 4 C, coverslips were

sealed to prevent slippage and dehydration.

Frozen sections, counterstained with Hoechst were imaged by Confocal microscopy.

A Leica TCS SP2 AOBS inverted stage Confocal microscope was used to capture Z-stacks

with the 20X objective in the green and blue channels using appropriate excitation

wavelengths and a 0.6 um step size. These z-stacks were processed using the associated

Leica software in order to produce projection views that compiled the stacked images into a

single image. Z-stack heights were selected to encompass the entire ductal fragment or

tumor tissue within the mounted section.

2.3 Time-lapse Imaging of H2B-eGFP Labeled Cells

Time-lapse images were obtained on the Zeiss Axiovision Observer microscope

equipped with a motorized stage and enclosed in an environmentally-controlled imaging

compartment. For time-lapse imaging, the compartment was maintained at 37C and

incubated at 5% CO2. The motorized stage and specialized software allowed the marking of

coordinates within the culture dish and repeated automated image acquisition at these

locations. Images were obtained at 15 or 20 minute intervals (depending on the experiment)

over a period of up to several days. These images were then compiled into time-lapse movie

files documenting the cellular behaviors within each imaged field using the Zeiss Axiovision

software.

Samples monitored in this manner were plated in 3D culture using Matrigel. Cells

and cell clusters were suspended in growth factor reduced Matrigel (BD Biosciences) and

overlaid with DMEM/F12 media supplemented with 10 ug/mL insulin, 5.5 ug/mL transferrin

and 2 ng/mL sodium selenite, (all from Sigma) and antibiotic/antimycotic. Matrigel plating

Page 60: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

43

was typically performed the day prior to initiation of imaging in order to allow proper

solidification of the extracellular matrix and minimize drifting of imaged cells. Refocusing

of the desired fields was necessary periodically throughout imaging (approximately once per

24 hours) and was completed concurrent with data download periods and media changes.

Dox was added to the media as needed for each experiment at 1000 ng/mL (in distilled water,

Sigma).

Page 61: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

44

Table 2.1 Primer Pairs Utilized in Genotyping Transgenic Mice

Transgenic Line Primer 1 Primer 2

MMTV-rtTA and K5-

rtTA

ACCGTACTCGTCAATTCCAAGGG TGCCGCCATTATTACGACAAGC

CGCCCAGAAGCTTGGTGTAG CGAATAAGAAGGCTGGCTCTGC

TWNT TGCGGTTCCTGATGTATTTTGC TGCATTCCTTTGGCGAGAGG

CACGAAATTGCTTCTGGTGGC TCGAAGATGTTGGGGTGTTGG

MMTV-wnt1 ATCCGCACCCTTGATGACTCCG GGCTATCAACCAACACACTGCCAC

GGACTTGCTTCTCTTCTCATAGCC CCACACAGGCATAGAGTGTCTGC

TGFP AAGTTCATCTGCACCACCG TCCTTGAAGAAGATGGTGCG

CCTTGATGCCGTTCTTCTGCTTGT

C

3 TGFP primers will make 2 bands

Page 62: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

45

Chapter 3

3.1 Developing a transgenic mouse model enabling inducible H2B-eGFP labeling of MECs

The pathways of lineage commitment and proliferation dynamics that lead to the

development of the complex structure of the mammary gland are poorly understood. Fuchs

and colleagues pioneered an experimental approach to tracking cell fates in the skin via

inducer-dependent expression of an H2B-eGFP transgene [182]. Here, we tested whether

this strategy could be adapted to permit monitoring of mammary epithelial cell lineages. We

paired the H2B-eGFP transgene with doxycycline-inducible transactivators specific to the

luminal or basal layers within the mammary ductal tree [172, 180]. By verifying restriction

of transgene expression patterns to the expected cell compartments we establish a model to

allow the tracking of cell lineage in later experiments. Additionally, in this first section, we

substantiate that the dilution of GFP signal largely results from cellular division by

contrasting the washout observed in high proliferative regions of the gland to that observed in

low-proliferative regions using fluorescent microscopic examination of tissue sections. Cell

division-dependent washout was further confirmed by eliminating hormone-induced cellular

proliferation in the mammary epithelium and documenting reduced GFP signal dilution by

fluorescence microscopy and flow cytometry. By rigorously defining the patterns of H2B-

eGFP labeling and label retention in these models, we establish a framework for our later

experiments examining proliferation dynamics in mammary tumors and during lobulo-

alveolar development and post-lactational involution.

3.2 Doxycycline-dependent labeling of MG

Our strategy for labeling of MECs is depicted schematically in Figure 3.1. Mice

carrying either the luminal or basal transactivators were engineered to permit the temporal

Page 63: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

46

regulation of mammary cell layer-specific expression of the H2B-eGFP fusion protein. As

depicted in Figure 3.1A, in the absence of Dox, H2B-eGFP should not be expressed, but the

addition of Dox to the animals’ diet should result in the TGFP transgene expression and GFP

labeling within the mammary epithelial cell layers. Removal of Dox inducer should abrogate

TGFP transgene expression whereas GFP signal from chromatin bound H2B-eGFP protein

should persist. Transactivators determine cell layer-specific activation of transgene

expression with Keratin-5-driven expression localized to the outer myoepithelial layer and

MMTV driven expression localized within the inner luminal cell layer (Figure 3.1B).

To test whether expression of the H2B-eGFP fusion protein was tightly regulated and

would permit Dox-dependent expression of the H2B-eGFP reporter, bitransgenic MMTV-

rtTA/TGFP mice and monotransgenic controls were generated. Mammary glands harvested

from adult females that were either left untreated or treated with 7 days of Dox were

examined by fluorescent microscopy (Figure 3.2, rows 1-4). Fluorescent labeling of nuclei

was undetectable in whole-mounted mammary glands from Dox-naïve mice and from Dox-

treated genetic control mice lacking either the transactivator or responder transgenes, as

expected. In contrast, bitransgenic mice exposed to a doxycycline (Dox) pulse showed

uniform labeling of mammary ducts from proximal to distal ends visualized as bright,

nuclear-localized green fluorescence in mammary gland whole-mounts.

In order to verify that activation of the H2B-eGFP transgene is reversible in the

mammary epithelium, adult virgin females were pulse labeled with Dox for 7 days and

subsequently subjected to Dox withdrawal for a washout period of 5 weeks (Figure 3.2, final

row). As depicted in Figure 3.1, cessation of Dox induction ought to result in the abrogation

of H2B-eGFP transgene expression but should not impact existing H2B-eGFP fusion protein

Page 64: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

47

already incorporated into the chromatin. The chromatin bound GFP signal intensity should

be diluted during cellular proliferation. In the setting of pulse-chase, analysis of mammary

gland whole mounts by wide-field fluorescence microscopy showed a marked decrease in

GFP fluorescence in comparison to that seen in pulse labeled mammary glands (Figure 3.2).

3.3 Compartment-Specific Labeling in the Mammary Epithelium

Next, we sought to confirm compartment-specific labeling within the mammary

epithelial cell populations at the cellular level. Toward this end, we extended our analysis of

transgenic mammary glands using two methods that offer single-cell resolution. First,

Hoechst-counterstained frozen sections of mammary gland whole mounts were imaged by

Confocal microscopy. Second, single-cell suspensions were prepared from mammary glands

and analyzed by flow cytometry. As described below, these experiments confirmed

compartment-specific and temporal regulation of H2B-eGFP expression in the mammary

gland.

3.3.1 MMTV-rtTA Drives Luminal Compartment Restricted Reporter Gene Expression

To examine the cell-layer specificity of Dox-induced H2B-eGFP labeling in MMTV-

rtTA/TGFP mice, mammary glands from Dox-naïve, pulse labeled and pulse-chased mice

were sectioned, counterstained with Hoechst and imaged by Confocal microscopy. In line

with observations from examination of whole mounts by fluorescence wide-field

microscopy, conical imaging confirmed that a large number of ductal cells labeled with GFP

fluorescence (Figure 3.3). Imaging of counterstained sections additionally demonstrated that

the GFP positive cells were localized to the inner lining of ductal structures, the anatomic

position of luminal epithelium. Hoechst-counterstained, GFP-negative nuclei exhibiting the

Page 65: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

48

typical elongated shape of myoepithelial cells were seen surrounding the uniform, round

nuclei typical of luminal epithelial cells. Likewise, the cells comprising the fat and stroma

were also devoid of GFP signal. Low-level expression of MMTV-LTR-driven transgenes in

lymphoid tissue has been described previously and rare extra-ductal GFP-positive bodies

were occasionally visualized within the lymph nodes which are readily identifiable in

sections by their distinct Hoechst-stained, tightly-packed nuclei [172].

In order to quantify GFP fluorescence on a single cell level and to further characterize the

mammary cell subtypes labeled by the MMTV-rtTA promoter, the cell surface marker profile

of MECs isolated from pulse labeled mammary glands were determined by flow cytometry.

A proportion of the single cells isolated from the mammary gland were GFP positive as

identified by flow cytometry (Figure 3.4A). GFP-positive cells from these mice were

CD24+/CD49

lo, corresponding to the luminal compartment as previously defined (Figure

3.4B). GFP-negative cells were found in the basal MEC population and non-epithelial

populations as expected.

3.3.2 Keratin-5 rtTA Drives Basal Compartment-Restricted Reporter Gene Expression

In order to verify the compartment and temporal restriction of H2B-eGFP expression

driven by the Keratin-5 transactivator whole mounts from adult K5/TGFP bitransgenic mice

were examined by wide-field and Confocal microscopy. Wide-field fluorescence

microscopy demonstrated widespread labeling of the mammary ductal structures following

transgene expression induction by Dox. GFP-positive nuclei exhibited an elongated

appearance and were located on the exterior of the mammary ducts as expected of basal

MEC-restricted expression of the GFP labeled histone (Figure 3.5). Sections obtained from

whole-mounted mammary glands counterstained with Hoechst reaffirmed the basal location

Page 66: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

49

and appearance of eGFP positive nuclei while the GFP negative cells within the ducts were

localized to the interior of the ducts and exhibit the rounded shape typical of luminal

epithelial cells (Figure 3.6). Cells with GFP-positive nuclei were not identified in the

luminal layers of mammary ducts nor within the mammary gland stroma that surrounds and

supports the ductal tree. Rare non-ductal, GFP-positive cells were occasionally visualized in

mammary glands and frozen sections contaminated with muscle or skin fragments, however

these fluorescent cells were easily distinguished from mammary epithelial cells based on

their histology and location.

Analysis of single cells isolated from pulse labeled mammary glands stained with

cell-surface markers by flow cytometry supported compartment-confined labeling induced by

K5-rtTA. A large percentage of epithelial cells isolated were GFP-positive following pulse

labeling. GFP-positive cells isolated from pulse labeled K5/TGFP bitransgenic mice were

largely restricted to the CD24lo

/CD49+ region previously described as harboring basal

mammary epithelial cells (Figure 3.7).

3.4 Proliferation dependent washout of H2B-eGFP labeling

To define the role of cellular proliferation in the dilution of H2B-eGFP signal

incorporated into the chromatin we utilized the differential growth rates that occur within

different developmental stages of mammary gland development. Mammary gland ductal

elongation begins following the onset of puberty when hormonal signals induce rapid

proliferation in club-like structures, terminal end buds, located at the distal end of the ducts.

The proliferation and migration of these distal tips results in elongation of the duct and

progression of the ductal tree through the mammary fat pad. By performing pulse-chase

Page 67: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

50

labeling experiments during puberty, we capitalize on the natural differential in proliferation

rates to examine the role of cell division in the washout of H2B-eGFP fluorescence.

Following puberty, proliferation rates generally decrease in the virgin animal. However,

some cell turnover occurs with each estrus cycle, when ovarian hormones induce short term

proliferation. Ablation of ovarian hormones via ovariectomy largely eliminates proliferation

within the adult mammary gland and we utilize this strategy as another avenue to investigate

the role of proliferation in washout of H2B-eGFP labeling. We examine pulse-chase labeling

and assess the proliferation dependent dilution of fluorescent signal in these distinct

developmental stages by Confocal microscopy and by flow cytometric analysis of GFP

intensity in single cell preparations.

3.4.1 Puberty Induced Proliferation Results in Dilution of Incorporated GFP Signal

Differentially in Developmentally Distinct Ductal Areas

In order to measure the role of proliferation in the dilution of fluorescent signal mice

were subjected to Dox pulse from 5 to 6 weeks of age, then underwent mammary gland

biopsy at the conclusion of the pulse. Post-biopsy mice were subjected to Dox withdrawal for

1 week and then necropsied for collection of remaining mammary tissue. The timing of this

experiment allowed the identification of TEB at both biopsy and necropsy and confirmed the

animal to be in mid puberty by localization of the TEB. As proliferation rates are greatest

during puberty, and particularly in the TEB, clear localization of the structures is critical to

the examination of the role of proliferation in the dilution of GFP fluorescence. In Figure

3.8, representative Confocal microscope images depicting the widespread GFP labeling of

nuclei in a pulse labeled MMTV-rtTA/TGFP bitransgenic pubertal mouse. The proximal

ducts have relatively low proliferation rates and in bitransgenic mice submitted to pulse

Page 68: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

51

labeling followed by chase, maintain a high concentration of bright GFP labeled cells.

Following off Dox chase, proximal ducts demonstrate only modest washout. In contrast, the

highly proliferative TEB have almost entirely washed out the H2B-eGFP signal as was

expected due to rapid cell cycling. Though proliferation rates vary between the regions of

the pubertal mammary gland, both regions in the chased mammary gland show lower levels

of GFP fluorescence than either region in a pulse-labeled mammary gland as proliferation

does occur in both areas.

3.4.2 Ovariectomy Blocks Proliferation Dependent Washout of H2B-eGFP Label in

Mammary Epithelial Cells

To further demonstrate that the dilution of GFP signal is largely dependent on cell

proliferation, we examined washout kinetics in adult mice subjected to ovariectomy. Ovarian

hormones play a critical role in the induction of cell cycling in the mammary gland and in the

absence of ovarian hormones, cellular proliferation is greatly reduced [189].

In order to assess the role of hormone-mediated cell cycling on GFP washout, 5

week-old bitransgenic MMTV-rtTA/TGFP mice were placed on Dox for 5 weeks and then

underwent excisional mammary gland biopsy. Dox-withdrawal was initiated post-

operatively and mice were sacrificed 4 weeks later. Experimental mice additionally

underwent ovariectomy at the time of mammary gland biopsy prior to initiating Dox

withdrawal. Figure 3.9 shows representative Hoechst counterstained frozen sections imaged

by Confocal microscopy. Ducts in pulse labeled biopsy samples show widespread GFP

fluorescence throughout the inner lining of the structures. Following chase, the mammary

ducts of ovary-intact animals showed markedly decreased GFP fluorescence by Confocal

microscopy. In contrast, mammary ducts from mice ovariectomized prior to Dox withdrawal

Page 69: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

52

retained H2B-eGFP label at a much higher level than in ovary-intact mice. This higher level

of label retention in hormone ablated animals supports the conclusion that cellular

proliferation is the cause of GFP signal washout in the mammary gland.

To quantify H2B-eGFP washout at the cellular level, flow cytometry was used to

assess GFP signal intensity in MECs isolated from MMTV-rtTA/TGFP mice under pulse and

pulse-chase conditions in the presence and absence of ovarian hormones. A large proportion

of the MECs isolated from pulse-labeled bitransgenic mice fluoresced brightly in the GFP

channel (Figure 3.10). Dilution of GFP signal by cell division ought to roughly halve the

GFP signal with each round. In the single parameter plots depicting GFP intensity in these

animals, multiple peaks of GFP intensity appeared representing successive halving of GFP

signal. Compared with ovary-intact mice, ovariectomized animals retained larger numbers of

GFP+ and GFP

bright cells following chase. Indeed, single-parameter plots from

ovariectomized animals indicated that very few cell divisions occurred as evidenced by very

modest dilution of the GFP signal. We conclude that hormone-mediated proliferation is

required for the dilution of H2B-eGFP within the mammary epithelium.

3.5 Simultaneous Induction of H2B-eGFP and Tet-responsive Wnt Oncogene Results in

Labeling of Mammary Hyperplasia

By pairing the luminal and basal transactivators with both the H2B-eGFP fusion

transgene and a Tet-responsive oncogene expressing transgene we achieved induction of

compartment specific labeling in induced mammary hyperplasia (Figure 3.11). Induction of

tri-transgenic (MMTV-rtTA/TGFP/TWNT or K5-rtTA/TGFP/TWNT) mice bearing both

responder transgenes resulted in hyperplasia appearing similar to that induced in bitransgenic

MMTV-rtTA/TWNT or K5-rtTA/TWNT mice although epithelial cells within the gland were

Page 70: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

53

additionally labeled with bright green fluorescence (Figures 3.12 and 3.13). The

combination of transgenes does not impair the incorporation of GFP labeled histone into the

chromatin of cells induced to divide by wnt oncogene stimulation nor does the GFP label

appear to impair oncogenic stimulus of growth within the mammary epithelium. The ability

to induce compartment specific labeling within oncogene induced hyperplasia and

particularly within Wnt-induced mammary tumors will be followed more in detail in

Chapters 4 and 5 of this thesis.

3.6 H2B-eGFP Transgene Labeling Allows Live Imaging of Mammary Cells

Histone labeling with GFP can be visualized in live cells without fixation allowing an

opportunity to image cells in real time and monitor cellular kinetics. We imaged GFP

labeled tissue fragments and cell clusters via time-lapsed fluorescent Confocal and traditional

microscopy. Pulse-labeled tissue fragments dissected from MMTV-rtTA/TGFP/TRAS

mammary glands. These fragments were placed in coverslip bottomed culture dishes with

Dox-supplemented media and anchored with glass coverslips. This system allowed the time-

lapse Confocal imaging of intact ductal fragments contained in the normal surrounding

stroma. Ductal fragments imaged in this way allowed monitoring of cellular movements

within the ductal fragments and of apoptosis of MECs within the ducts (Figure 3.14). We

have also suspended single cells and small clusters of MECs in Matrigel. Matrigel mimics

basement membrane allowing the maintenance of 3D orientation and development of

polarized epithelium similar to that seen in intact stroma. Time-lapse imaging of cells

suspended in Matrigel visualized cell movements, cell divisions, and apoptosis (Figure 3.15).

The capacity to monitor cellular proliferation and movement patterns within the mammary

gland or in fragments of mammary epithelium provides an interesting method of

Page 71: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

54

characterizing the behavior of labeled cells within the mammary cell compartments and of

tracking the behavior and lineage of a cell and its progeny over time.

3.7 Discussion

The data provided in this chapter support the use of H2B-eGFP for the analysis of cell

cycling dynamics within discrete populations of cells. It adapts methods used for tracking

cell lineages in other tissues to explore the developmental biology of the mammary epithelial

tree. Analysis of GFP signal dilution within the mammary gland additionally supports a

model of division driven GFP signal washout and provides a basis for the prospective

isolation of viable mammary epithelial cells based on their proliferation histories.

When compared to similar strategies used to examine the lineage and proliferative

histories within the mammary gland, the transgenic model employed here has the clear

benefit of being specific to the mammary gland and further with GFP expression being

specifically regulated in the two cell layers that make up the mammary ducts. We are able

to identify the population of cells initially labeled, and are able to compare proliferation

dynamics between the two mammary epithelial cell compartments.

The ability to specifically analyze the contributions of the epithelial cell types to the

normal development and oncogene induced neoplasia of the mammary gland provides an

opportunity to characterize the lineage of cells of the mammary gland. By pulse-chase

analysis of MECs from each compartment within a defined developmental state of

mammopoiesis or tumorigenesis, the contributions of luminal and myoepithelial cells to the

developed tissue can be teased apart. As both mature ducts and many types of mammary

tumors contain both cell types and the exact pathways contributing to these multi-cell type

Page 72: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

55

tumors are not yet understood GFP based lineage tracing may elucidate the paths to lineage

commitment.

Additionally, the use of compartment specific labeling with H2B-eGFP would allow

the prospective isolation of viable MECs with a certain labeling or label retention profile and

the further study of the capacity of those cells in culture or following implantation into

cleared mammary fat pads. A comparable model has been used to identify both the stem cell

population and the niche of those cells within the skin [182]. It is possible that long term

label retaining cells in the mammary gland have stem cell like roles and identification of

these label retaining cells by GFP fluorescence might allow verification of stem cell

capabilities via implantation and tracking of lineage of their progeny. It has been proposed

that H2B-eGFP labeling is superior to traditional BrdU labeling for the identification of

proliferation history of cells as GFP signal is detectable through more divisions than that of

BrdU labeled cells. While immunophenotypes for stem cells within the mammary gland

have been described, they lack the capacity to produce a pure population of stem cells and

instead only enrich a population of mammary epithelial cells [27-29]. Combination of the

previously defined immunophenotypes with GFP label retention properties could allow more

selective isolation of the mammary stem cell population and provide a selection criterion

based on a putative functional biological property of adult stem cells, slow cell cycling.

H2B-eGFP labeling could also be used to purify a stem cell population that is more rapidly

cycling than their surrounding cells by selection of GFP-negative cells that have diluted

H2B-eGFP signal.

GFP signal could be diluted by either cell division or by histone turnover within

labeled cells. GFP label retention over extended periods of time in retinal post-mitotic cells

Page 73: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

56

supports a model where label dilution driven almost entirely by proliferation, as non-

proliferative cells retain GFP fluorescence [190]. Photoreceptor cells in the eyes of mice

labeled with ROSA-driven GFP during embryogenesis retained label following six months of

washout supporting a model of H2B-eGFP label stability with low histone turnover.

Additional support for a division driven model of GFP label dilution is provided by the flow

cytometric analysis of cells from a variety of tissues. GFP single parameter plots of labeled

and label retaining cells from the skin show peaks of GFP corresponding to cell divisions

roughly halving the GFP signal intensity with each peak [191].

In these experiments we have labeled the mammary gland during periods of extensive

growth and proliferation and during phases of limited cell turnover providing the possibility

that GFP label incorporation may not require cellular proliferation. H2B-eGFP labeling in

the skin and in the hematopoietic system also suggests that proliferation is not required for

label incorporation by the widespread incorporation of label throughout the tissue even

following short pulse labeling periods in infrequently dividing cell populations. Though

widespread labeling within the infrequently cycling adult mammary gland suggests that

division is not required to incorporate H2B-eGFP into the chromatin further experiments

could be performed to verify this conclusion. Specifically, adult bitransgenic mice could be

ovariectomized to eliminate cell division within the mammary gland and subsequently pulse

labeled with Dox. If proliferation is required for the incorporation of GFP labeled histone,

these mice will not have detectable GFP signal within the mammary ducts.

While compartment specific labeling in the mammary gland does not appear to cross

label compartments it also does not label all of the cells within a particular cell layer. In the

luminal epithelium both hormone receptor positive and hormone receptor negative cells

Page 74: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

57

make up the inner cell layer and it is possible that MMTV driven expression reproduces this

hormone receptor status with GFP+ cells appearing in either the HR

- or HR

+ cell populations.

Similarly, the basal cell layer as defined by published immunophenotypes contains both the

myoepithelial cell population and proposed stem cell population. Identification of the precise

cell types labeled with MMTV or K5 driven expression of H2B-eGFP would provide insight

into the expression of mouse mammary tumor virus and the distinct proliferation patterns of

the subtypes of mammary epithelial cells.

Page 75: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

58

A. B.

Figure 3.1 Strategy for Temporal Regulation of H2B-eGFP Transgene Expression in

Epithelial Cellular Compartments of the Mammary Gland.

A. Schematic of Transgenes Employed for Compartment-Restricted Expression. Transgenic

mice carrying the MMTV-rtTA-pA transgene express the rtTA transcription factor in the

luminal mammary epithelium but do not express independently integrated responder

transgenes driven by minimal CMV promoters containing multimerized operator sequences.

Similarly, transgenic mice carrying the K5-rtTA-pA transgene express the rtTA transcription

factor in the basal mammary epithelium. In the presence of inducer, rtTA binds Dox and

undergoes a conformational change that allows recognition of Tet operator sequences and

responder transgene activation. Transgenic lines used in this study permit Dox regulated

expression of the reporter gene: Histone H2B-eGFP fusion protein (TGFP line). B. Depiction

of Expected Labeling Pattern of H2B-eGFP signal in Dox-Induced Mammary Epithelial

Cells

Compartment specific expression of transactivators allows compartment specific expression

of histone fused eGFP resulting in GFP positive nuclei in either the luminal (MMTV-rtTA-

driven) or basal (Keratin 5-rtTA-driven) layers of the mammary epithelium within the ductal

structures of the mammary gland.

Page 76: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

59

Figure 3.2: Dox-Regulated Expression of the H2B-eGFP Transgene in Mammary

Epithelium of MMTV-rtTA/TGFP Mice.

Panels depict widefield microscopic images of mammary gland whole mounts collected in

the brightfield and fluorescent channels (Mag. 5.1X). Mammary gland whole mounts were

prepared from adult mice of the indicated genotypes and under pulse and pulse-chase

conditions. Third column images are superimposed GFP/white light images. Final row GFP

images are 2X the exposure of other rows.

Page 77: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

60

Figure 3.3 Luminal-Restricted Labeling of MECs in MMTV-rtTA/TGFP Mice.

Image depicts a MAX projection image collected by Confocal microscopy (Mag. 20X). A

bitransgenic (MMTV-rtTA/TGFP) adult virgin female was placed on Dox for 7 days prior to

harvest of the 4th

inguinal mammary gland. Following frozen sectioning, sections were

prepared for Confocal microscopy by counterstaining with Hoechst dye (blue). Optical

sections were collected through the entire thickness of the prepared section and compiled into

MAX projections to produce a single image depicting the stack of optical sections. This view

demonstrates GFP fluorescence lining the inside of the ductal structure while single Hoechst

counterstained nuclei line the outside of the ducts, primarily composed of myoepithelial cells.

Hoechst stained fat cells are also visualized in the surrounding stromal tissue.

Page 78: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

61

Figure 3.4 Immunophenotyping of Labeled Mammary Epithelial Cells Confirms

Compartment-Specific H2B-eGFP Fluorescent Labeling

Single cells isolated from pulse labeled bitransgenic mice (MMTV-rtTA/TGFP) were stained

for CD24 and CD49 and analyzed for fluorescence by flow cytometry. Panels depict

representative flow cytometric plots demonstrating GFP fluorescence and immunophenotype

results of pulse labeled mammary glands. A. Single Parameter Plots. Histogram depicts

GFP intensity for single cell events. GFP signal intensity was assayed in single cells by flow

cytometry and cells were gated into GFP positive (>102 fluorescence) or GFP negative (<10

2

fluorescence). B. Dual-Parameter Immunophenotyping. Single cell MG preps were stained

with CD24 and CD49 to resolve luminal and basal populations using published methods.

Cells scored as GFP+ versus GFP- on the basis of the gating shown in panel A were plotted

independently as shown. GFP+ cells were primarily contained with tin the CD24 high

/CD49 lo

luminal subset.

Page 79: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

62

Figure 3.5 Dox-Regulated Expression of H2B-eGFP in Basal Mammary Epithelium of K5-

rtTA/TGFP Mice.

Panels depict fluorescent channel microscopic images collected from mammary gland whole

mounts obtained from adult K5-rtTA/TGFP bitransgenic mice after 7 days on Dox (left

image) or from a Dox-naïve animal (right image) (Mag. 10X).

Page 80: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

63

Figure 3.6 Basal-Restricted Labeling of MECs in K5-rtTA/TGFP Mice.

Image depicts a MAX projection image collected by Confocal microscopy (Mag. 20X). A

bitransgenic (K5-rtTA/TGFP) adult virgin female was placed on Dox for 7 days prior to

harvest of the 4th

inguinal mammary gland. Following frozen sectioning, sections were

prepared for Confocal microscopy by counterstaining with Hoechst dye (blue). Optical

sections were collected through the entire thickness of the prepared section and compiled into

MAX projections to produce a single image depicting the stack of optical sections. This view

demonstrates GFP fluorescence lining the outside of the ductal structure while single Hoechst

counterstained nuclei line the inside of the ducts, primarily composed of luminal epithelial

cells. Hoechst stained fat cells are also visualized in the surrounding stromal tissue.

Page 81: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

64

Figure 3.7 Immunophenotyping of Labeled Mammary Epithelial Cells Confirms H2B-

eGFP Fluorescent Labeling.

Single cells isolated from pulse labeled bitransgenic mice (K5-rtTA/TGFP) were stained for

CD24 and CD49 and analyzed for fluorescence by flow cytometry. Panels depict

representative flow cytometric plots demonstrating GFP fluorescence and immunophenotype

results of pulse labeled mammary glands. A. Single Parameter Plots. Histogram depicts

GFP intensity for single cell events. GFP signal intensity was assayed in single cells by flow

cytometry and cells were gated into GFP positive (>102 fluorescence) or GFP negative (<10

2

fluorescence). B. Dual-Parameter Immunophenotyping. Single cell MG preps were stained

with CD24 and CD49 to resolve luminal and basal populations using published methods.

Cells scored as GFP+ versus GFP- on the basis of the gating shown in panel A were plotted

independently as shown. GFP+ cells were primarily contained within the CD24+/CD49

+

epithelial cell subsets.

Page 82: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

65

Figure 3.8 Selective Washout of H2B-eGFP Fluorescence by Developmental

Proliferation in Terminal End Buds.

Panels depict representative images obtained by conventional fluorescence microscopy on

whole-mounts (Mag. 5X) and frozen tissue sections as indicated. Tissue sections were

counterstained with Hoechst, imaged by conventional fluorescence microscopy ( Mag. 20X)

and optical sections were compiled to form MAX projection images. MMTV-rtTA/TGFP

bitransgenic female mice began Dox treatment at 5 weeks of age and underwent excisional

biopsy of the right inguinal mammary gland at 6 weeks of age. Dox treatment was stopped

post-operatively, and the contralateral inguinal gland was harvested one week later at

necropsy. Proximal ducts were located between the nipple and the inguinal lymph node,

whereas terminal end buds were distal to the lymph node at the leading edge of duct

migration in pubertal mice.

Page 83: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

66

Figure 3.9 Proliferation is Necessary for the Washout of Induced H2B-eGFP Signal in

the Mammary Gland.

Images depict conventional fluorescence microscope views of the Green and Blue channels

of Hoechst counterstained frozen sections of mammary glands (Mag. ??X). MMTV-

rtTA/TGFP bitransgenic female mice began Dox treatment at 5 weeks of age and underwent

excisional biopsy of the right inguinal mammary gland at 10 weeks of age. Dox treatment

was stopped post-operatively, and the contralateral inguinal gland was harvested four weeks

later at necropsy. Ovariectomized (OVX) animals underwent excision of the ovaries at the

time of mammary gland biopsy prior to the removal of Dox.

Page 84: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

67

Figure 3.10: Elimination of Ovarian Hormone Induced Proliferation Limits H2B-eGFP

Signal Dilution in the Mammary Gland.

Panels depict representative single parameter plots of GFP signal intensity measured by flow

cytometry. Single cells isolated from pulse and pulse-labeled then chased MMTV-

rtTA/TGFP bitransgenic mice were analyzed by flow cytometry for GFP fluorescence. Top

Panel: A large proportion of the cells isolated from an on Dox mouse are GFP positive (>103

fluorescence). Middle Panel: Following 4 weeks of off Dox chase, the majority of the

brightly labeled cells have divided and have divided their fluorescence among the daughter

cells. Arrows indicate peaks of GFP fluorescence roughly half of the preceding intensity.

Lower Panel: MECs from animals pulse labeled prior to ovariectomy and Dox-withdrawal

retain bright GFP fluorescence and peaks suggesting cell division induced dilution are not

present.

Page 85: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

68

Figure 3.11 A Strategy for Dox-Regulated H2B-eGFP Labeling of MECs in the

Context of Inducible Wnt1 Expression.

Transgenic mice carrying the MMTV-rtTA transgene express the rtTA transcription factor in

the luminal mammary epithelium but do not express independently integrated responder

transgenes driven by minimal CMV promoters containing multimerized operator sequences.

Similarly, transgenic mice carrying the K5-rtTA transgene express the rtTA transcription

factor in the basal mammary epithelium. In the presence of inducer, rtTA binds Dox and

undergoes a conformational change that allows recognition of Tet operator sequences and

responder transgene activation. Transgenic lines used in this study permit Dox regulated

expression of the reporter gene: Histone H2B-eGFP fusion protein (TGFP line) and the

oncogene, wnt (TWNT line).

Page 86: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

69

Figure 3.12 H2B-eGFP Labeling of Luminal MEC-driven, Wnt1-initiated Mammary

Hyperplasia.

Images demonstrate widespread H2B-eGFP signal throughout the ducts of adult tritransgenic

(MMTV-rtTA/TWNT/TGFP) mice placed on Dox for one week. A. Mammary Gland Whole

Mount. Images depict a wide-field fluorescent micrograph (Mag. 10X) of a representative

mammary gland whole mount. B. Mammary Gland Frozen Section. A MAX projection view

of a representative duct captured by Confocal Microscopy (Mag. 20X) from a Hoechst

counterstained mammary gland showing GFP positive nuclei lining the inner surfaces of the

ducts.

Page 87: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

70

Figure 3.13 H2B-eGFP Labeling of Basal MEC-driven, Wnt1-initiated Mammary

Hyperplasia.

Images demonstrate widespread H2B-eGFP signal throughout the ducts of adult tritransgenic

(K5-rtTA/TWNT/TGFP) mice placed on Dox for one week. A. Mammary Gland Whole

Mount. Images depict a wide-field fluorescent micrograph (Mag. 10X) of a representative

mammary gland. B. Mammary Gland Frozen Section. A MAX projection view of a

representative duct captured by Confocal Microscopy (Mag. 20X) from a Hoechst

counterstained mammary gland shows GFP positive nuclei lining the outside of the duct.

.

Page 88: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

71

Figure 3.14 H2B-eGFP Labeling of Mammary Epithelial Cells Allows Tracking of Cell

Fates in Real Time via Time-Lapse Confocal Microscopy.

Panels depict frames isolated from a time lapse imaging series collected by Confocal

microscopic imaging (Mag. 20X) of mammary gland fragments in culture. Pulse labeled

MMTV-rtTA/TGFP/Tet-O-HRASG12V

mammary gland was collected, cut into small (~1mm)

pieces and anchored in culture media in glass bottom dishes. These oncogene expressing

fragments were maintained on Dox and imaged over a period of several hours by Confocal

microscopy. Optical sections were collected throughout the depth of the imaging stack at

each time interval and compiled into a series of Max projections. These Max projections

were compiled as a time-lapse movie demonstrating the cellular activity throughout the

imaging period. Of interest, imaging captured cell movements within the tissue fragment and

apoptosis of cells within the structure are indicated by the red arrows.

Page 89: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

72

J.Mathers Fig 3.15A.AVI J.Mathers Fig 3.15B.AVI

Figure 3.15 Time-Lapse Fluorescence Microscopy of H2B-eGFP Labeled Mammary

Cells in Mammary Duct Fragments in Culture Allows Tracking of Cellular

Proliferation Dynamics.

Videos show brightfield/fluorescent channel microscopic images of mammary ductal

fragments in culture collected every 15 minutes and compiled into time-lapse series (Mag.

20X). A. Brightfield/GFP fluorescence overlay. B. Fluorescent channel alone. Pulse labeled

MMTV-rtTA/TGFP/MMTV-wnt1 mammary tumors were partially dissociated and plated in

Matrigel Suspension. Organoids were maintained under Dox treatment in culture and

imaged by conventional fluorescent and brightfield microscopy. Cell movements are

apparent throughout the movies and cell division is visible in the movies in the 2 o’clock

position.

Page 90: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

73

Chapter 4 H2B-eGFP Labeling and Proliferation Dynamics in Mammary Tumors

Driven by Constitutive Oncogenes

4.1 Introduction

Breast cancers are composed of heterogeneous tumor cell populations but how this

heterogeneity contributes to tumor growth and maintenance remains poorly understood.

Mammary tumors frequently contain cells of both luminal and basal origin and the lineage

history of cells that make up these tumors is unknown. Figure 4.1 illustrates the cell types

found within mammary adenocarcinomas initiated by MMTV-wnt1 expression and

introduces the matter of tumor initiating capacity. Elucidating the role of luminal and basal

cells in the development and maintenance of mammary adenocarcinomas would assist in the

identification of cell of origin for mammary cancers and possibly in the development of

targeted therapies aimed at the specific cell-types required for tumor maintenance.

In addition to the cell-layer diversity found in mammary tumors, mammary tumors

likely contain populations with distinct patterns of activity and growth. Specifically, the

identification of cell populations from either cell type within growing mammary tumors that

have different proliferation rates would allow characterization of distinct cell populations

within a seemingly homogeneous tumor and increase understanding of the complex

developmental patterns of mammary tumors. Recently, the possibility of separate cell

populations with distinct functional characteristics has been used to propose the presence of

cancer stem cells within many kinds of malignant growths. These CSC are proposed to be

chemo-resistant and have increased longevity and proliferative capacity than that of

surrounding bulk tumor cells.

In recent years, strategies for dissecting the roles of distinct tumor cell subsets within

a tumor have emerged. Tumor cell subsets can be defined by lineage markers or by

Page 91: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

74

functional assessment such as proliferate rate. By using H2B-eGFP compartment specific

labeling we are able to track the proliferation history of cells in the luminal and basal

compartment. The use of two independent compartment restricted promoters also allows the

tracing of lineage labeled cells and analysis of cell-compartment specific behaviors within

the developing mammary tumors. Additionally, by labeling and following viable label

retaining cells in mammary tumors we establish a model that allows testing of the varying

capabilities of differentially label retaining cells from both epithelial cell compartments in

mammary tumors.

4.2 Experimental Strategies Allowing the Study of Proliferation Dynamics in Mammary

Tumors Driven by Constitutive Oncogenes

In this chapter we paired the labeling strategy employed in Chapter 3 with the

mammary epithelium-specific, constitutive expression of the Wnt1 oncogene to examine

proliferation dynamics in a mouse mammary tumor model. In addition to the MMTV-rtTA,

K5-rtTA, and TGFP transgenes previously described, we utilized an MMTV-wnt1 transgene

expressed in the luminal epithelium of the mammary gland to drive tumorigenesis (Figure

4.2A). In this model, mice develop diffuse hyperplastic growth throughout the mammary

ductal tree and stochastically develop solitary mammary tumors. These tumors are typically

mixed-lineage adenocarcinomas that include both luminal and basal epithelial cells. Tumor

growth is driven independently of labeling of mammary epithelial cells with H2B-eGFP

(Figure 4.2B).

Building on the work described in Chapter 3, we employed Dox-dependent,

compartment-restricted H2B-eGFP labeling in the context of constitutive MMTV-wnt1

driven tumors to investigate proliferation dynamics. As depicted in Figure 4.3, two possible

Page 92: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

75

models exist regarding H2B-eGFP label retention in mammary tumors: Either all the tumor

cells within the labeled cell layer proliferate at similar rates and will dilute GFP signal at a

similar pace or populations of cells within the tumor proliferate at distinct rates and will be

detectable by different levels of GFP label retention. Potentially, CSC would additionally

have distinct proliferation rates and could be identified by H2B-eGFP labeling/label-retention

studies. In addition to the differences probable between cell-compartments within the

mammary tumors, based on the descriptions of distinct tumor cell subsets common in a

diverse range of tumors, it is expected that mammary tumors also likely contain cell

populations with disparate proliferation rates.

In this chapter, we examine tumors under pulse-chase conditions to test for either

homogeneous or heterogeneous washout of GFP signal in mammary tumors and by using

both the luminal and basal driven GFP labeling strategies tested for the presence of slower

cycling label retaining cells in both cell compartments of mammary tumors. This strategy

allows the identification of cell proliferation rates in both luminal and basal populations

within growing mammary tumors and the tracing of lineage commitment within the cell

populations within mammary tumors.

4.3 Dox-regulated GFP Labeling Independent of Tumorigenesis

In order to verify the regulation of H2B-eGFP in MMTV-wnt1 driven mammary

tumors we monitored Dox-naïve tritransgenic MMTV-rtTA/MMTV-wnt1/TGFP female

mice until mammary tumors developed. Tumor bearing mice were then Dox-treated for one

week to induce H2B-eGFP expression in the luminal compartment of the mammary tumors

and subjected to tumor biopsies. Figure 4.4A shows the widespread labeling with GFP that

is apparent in gross tumor specimens with appropriate excitation of the GFP. Examination

Page 93: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

76

of frozen sections from Dox-pulsed mammary tumors showed widespread fluorescent

labeling of tumor cell nuclei (Figure 4.4B). Flow cytometric measurement of GFP signal

intensity in single cells isolated from these tumors confirmed our visual analysis of the

presence of GFP signal in mice pulse labeled with Dox.

While many tumors were brightly labeled with GFP fluorescence during Dox-pulse

there was a wide range of proportions of tumor cells labeled that varied greatly from tumor to

tumor. Figure 4.4C portrays the variability in pulse labeling between primary tumors where

GFP positive cells are those with fluorescence above 102.5

as illustrated in Figure 4.4B. A

wide range of GFP fluorescence was found with some tumors exhibiting practically no GFP

fluorescence during Dox pulse and others containing nearly 60% brightly GFP positive cells.

This range of fluorescence intensity likely results from the varied accumulated mutations

necessary for tumor initiation altering GFP signal induction patterns and the impact of those

mutations on transgene expression

To examine whether pulse-chase labeling would permit quantification of tumor cell

proliferation, mice were then subjected to Dox withdrawal to allow proliferation dependent

washout of H2B-eGFP during continues tumor growth. As depicted in Figure 4.2B, growing

tumors dilute the GFP signal and flow cytometry of single cells isolated from pulse labeled

and subsequently chased tumors quantified the H2B-eGFP signal retention. Figure 4.5

depicts pulse and post-chase analysis of 16 independent primary tritransgenic MMTV-

rtTA/MMTV-wnt1/TGFP tumors. The first column contains tumors that did not respond to

Dox-induction by producing H2B-eGFP protein. No GFP fluorescence was detectable in

these tumors following pulse eliminating the possibility of comparison to chased matched

samples. The remaining eleven tumor pairs exhibited varying levels of pulse GFP

Page 94: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

77

fluorescence and pulse-chase label washout. Washout of H2B-eGFP signal varied between

tumors possibly due to different growth rates between independent tumors and the exact

cellular composition of each tumor. The large variability in initial labeling and in subsequent

washout illustrates the array of tumor variations that occur and recapitulates the diversity

seen in mammary tumors.

4.4 Identification of LRC in Explanted Mammary Tumors

As the analysis of pulse-chase labeling in primary MMTV-wnt1 tumors showed

marked tumor to tumor variability such that only a subset of tumors yielded sufficient

labeling to allow study of proliferation dependent washout. In order to circumvent this

variability, we propagated tumors as explants to determine whether those primary tumors that

permitted efficient H2B-eGFP labeling would yield descendent tumors with similar labeling

patterns. In addition, we reasoned that explants would expand the number of specimens,

allowing a variety of pulse-chase conditions to be performed on “identical” clonally related

tumors. Accordingly, we propagated primary MMTV-rtTA/MMTV-wnt1/TGFP tumors with

bright eGFP label as described in Figure 4.6. This procedure involved taking a biopsy

fragment from a tumor of a given genetic background, in this study MMTV-wnt1/MMTV-

rtTA/TGFP, and implanting one small fragment onto each of the flanks of syngeneic but non-

transgenic mice. These fragments grow into mammary tumors over a period of weeks and

provide multiple time-matched clonally related tumors that can be used to isolate and analyze

labeled and label retaining cells. Additionally, we were able to explant some tumors through

several generations allowing analysis of clonally related tumors over these generations.

Analysis of gross mammary tumor explants pulse labeled with doxycycline displayed

bright GFP fluorescence (Figure 4.7A). Dox-naïve animals bore tumors that did not contain

Page 95: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

78

GFP positive nuclei by Confocal imaging or by flow cytometric analysis of single cells

isolated from the tumors (Figure 4.7B). Pulse labeled explanted tumors brightly labeled with

H2B-eGFP as shown in the representative Confocal image and flow cytometry plot of GFP

fluorescence. Figure 4.7C provides an analysis of GFP fluorescent labeling in sibling tumors

of subsequent generations of mammary tumor explants by quantifying GFP fluorescence

from single cells isolated from pulse labeled mammary tumor explants. Tumor generations

generally recapitulated the labeling pattern of the primary tumor with some variability in

fluorescence intensity occurring between generations. However, following rounds of explants

some tumor lines eventually lost their typical adenocarcinoma appearance and developed a

more mesenchymal appearance. This change in tumor type was accompanied by a loss of

GFP signal induction by doxycycline.

The epithelial to mesenchymal transition (EMT) of mammary tumor explant

generations results in a change in the histological appearance of the tumors. Generations of

tritransgenic tumor passaged on syngeneic hosts gradually ceased to respond to doxycycline

induction of GFP as seen in Figure 4.7C. It is most likely that the EMT altered the

expression pattern of the cells composing the mammary tumor away from the luminal

epithelial pattern required for MMTV-driven induction of a transgene. Sequencing of HRAS

mutations throughout the generations could be used to verify tumor progeny. Additionally,

EMT results in characteristic marker changes in these mammary tumors that can be assayed

by Northern analysis of mRNA expression [192].

To examine the proliferation dynamics in growing mammary tumors driven by

MMTV-wnt1 we used a pulse-chase labeling strategy. Pulse labeled mammary tumors

exhibited bright green fluorescence even when gross specimens were examined by eye under

Page 96: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

79

UV wavelength excitation (Figure 4.8A). Pulse labeled, chased tumors also had detectable

GFP labeling by this manner though contained less than on Dox tumors (Figure 4.8A).

Increasing periods of off-Dox chase resulted in decreased levels of GFP labeling retained in

the nuclei of tumor cells. Confocal microscope imaging of thick frozen sections of these

tumors showed decreasing numbers of brightly GFP labeled nuclei with increasing periods of

off Dox washout concurrent with tumor growth (Figure 4.9B). The decreased GFP signal

intensity in growing tumors can likely be attributed to dilution of H2B-eGFP by rounds of

cellular division in the rapidly proliferating cells.

We additionally quantified label retention in tumor cells by flow cytometric analysis

of single tumor cells from pulse labeled then chased mice (Figure 4.9; for additional

examples, please see the Appendix). One tumor was able to be followed over 7 sequential

explants generations and was also frozen following 4 explant generations and was

subsequently propagated and monitored for an additional 5 generations. While this tumor

was exceptional in its duration, several explants were followed for 2-4 generations of

explants. Labeling and label retention in pulse-chased tumors was generally consistent

within a generation and followed the pattern of the preceding generations but exact labeled

proportions were not inherent in tumor explants. Later generations lost inducible H2B-eGFP

as measured by flow cytometric analysis of pulse-labeled tumor samples probably due to

EMT as documented by microscopic imaging.

4.5 Adriamycin Treatment Does Not Alter GFP Labeling and Label Retention Dynamics in

MMTV-wnt1 Driven Mammary Tumors

To investigate the effects of chemotherapeutic treatment on the proliferation

dynamics in the tumors we administered a single dose of adriamycin (300mg/kg) to pulse and

Page 97: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

80

pulse-chased mammary tumors. Figure 4.10 shows the single parameter plots of GFP

intensity for single cells isolated from sibling mammary tumor explants. Tumors from a

Dox-naïve animal did not have detectable GFP positive nuclei while pulse labeling appeared

to be unaffected by the administration of adriamycin as pulse labeled treated and untreated

animals had similar GFP fluorescence profiles (Figure 4.10). Interestingly, pulse-chase

washout in tumors chases for 6 and 10 days were not distinguishable possibly due to varied

tumor growth rates (Figure 4.10C). Pulse-chase washout of GFP signal also appeared similar

in adriamycin treated and untreated animals (Figure 4.10D). Single dose treatment with

chemotherapeutic drug did not cause tumors to shrink and did not seem to halt cellular

proliferation as measured by GFP signal dilution resulting from cell divisions.

4.6 Keratin-5 Induced Labeling and Label Retention in MMTV-wnt1 Driven Constitutive

Mammary Tumors

We utilized Keratin-5 driven TGFP transgene expression in mice carrying the

MMTV-wnt1 transgene to investigate the labeling and label retention dynamics in the basal

cell components of mammary tumors. Adenocarcinomas consist of both luminal and basal

cell types and by pairing the constitutive oncogene with the basal promoter we are able to

induce labeling in the basal compartment of mammary tumors. Following induction with

doxycycline, we observed widespread H2B-eGFP labeling in tumors from K5-rtTA/MMTV-

wnt1/TGFP mice (Figure 4.11A). Examination of sections from these tumors showed

decreased GFP labeling following Dox-withdrawal. Non-neoplastic duct fragments adjacent

to the growing tumors were also brightly labeled during Dox-pulse but retained higher levels

of GFP signal following chase due to lower proliferation rates in the non-tumor portions of

the mammary gland. Flow cytometric analysis of single cells isolated from pulse-labeled

Page 98: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

81

tumor biopsy showed a significant population of GFP positive cells (Figure 4.11B). Pulse-

chased K5-rtTA/MMTV-wnt1/TGFP mammary tumors diluted H2B-eGFP showing washout

of the GFP signal by the decreased number of GFP positive cells following chase. Clonal

tumor explants of basal driven GFP in MMTV-wnt1 tumors showed bright eGFP labeling of

the myoepithelial cells and also demonstrated washout of the GFP signal in pulsed then

chased mammary tumors. The K5-driven expression of H2B-eGFP provides a counterpoint

to the labeling induced by MMTV-rtTA and an avenue for the exploration of proliferation

dynamics in the myoepithelial compartment of mammary adenocarcinoma.

4.7 Constitutive MMTV-Neu Mammary Tumors can be Labeled with H2B-eGFP

While the cellular heterogeneity of MMTV-wnt1 initiated mammary tumors mimics

some forms of human breast cancers, alterations in HER2 commonly cause breast cancer in

women and are quite distinct histological [193]. MMTV-neu driven mammary tumors are

reported to not contain an identifiable stem cell population, possibly attributable to their

homogeneous cellular composition [194]. In order to characterize the patterns of cellular

proliferation and examine the potential for distinct populations with different proliferation

histories we employed our pulse-chase labeling scheme in MMTV-neu driven mammary

tumors. Tri-transgenic MMTV-neu/MMTV-rtTA/TGFP mice were maintained off Dox and

monitored for tumor formation. Upon tumor formation, animals were placed on Dox in order

to induce H2B-eGFP labeling in the mammary tissue. Biopsy samples obtained from these

tumors were variable in efficiency of H2B-eGFP labeling. While several tumors did contain

widespread nuclear labeling with H2B-eGFP, many pulse-labeled tumor biopsy samples did

not contain detectable H2B-eGFP fluorescence (Figure 4.12C and D). The mechanisms

responsible for this loss of regulation of transgene expression are unclear and await further

Page 99: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

82

definition. Potentially, the accumulated genetic hits acquired during tumorigenesis may have

silenced or deleted the H2B-eGFP transgene or the transactivator transgene.

In order to examine the potential for the existence of multiple distinct cell populations

within the seemingly homogeneous mammary tumors, we initiated washout of H2B-eGFP

signal by the removal of Dox. MMTV-neu driven tumors that incorporated H2B-eGFP

labeling and were subsequently removed from Dox were analyzed by microscopic imaging

and by flow cytometric assessment of GFP signal intensity and immunophenotype. Frozen

sections from pulse labeled mammary tumors showed the typical uniform cellular appearance

and widespread H2B-eGFP labeling. Cells in pulse labeled then chased tumor sections

demonstrated varied washout but without an identifiable pattern distinguishing label retaining

from non-label retaining cells (Figure 4.12B). Flow cytometric analysis of single cells

isolated from pulse labeled mammary tumors showed a large subset of the single

immunophenotype typical of MMTV-neu tumors were brightly labeled with H2B-eGFP

fluorescence (Figure 4.12A). The limited availability of samples due to the unexpected

absence of H2B-eGFP labeling in a large subset of mammary tumors complicated the

analysis of label retaining cells following off Dox washout. However, MMTV-neu driven

tumors clearly label and retain label in a distinct manner than that seen in MMTV-wnt1

driven mammary tumors, potentially due to the different pathways of tumor promotion and

target cell of mutations.

4.8 Discussion

The H2B-eGFP labeling system provides an interesting manner in which to

investigate proliferation dynamics within developing mammary tumors. By pairing the

reporter transgene with two independent transactivator lines we were able to analyze cellular

Page 100: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

83

proliferation rates in both the luminal and basal compartments of MMTV-wnt1 driven

mammary tumors. This ability will allow the contributions of each cell compartment to the

developing mammary tumor to be further characterized. Pulse-chase labeling of mammary

tumors with varying time lines may uncover cell populations with distinct cell-cycle

characteristic and cell cycle characteristics may be paired with defined immunophenotyping

markers to further subdivide cell populations comprising the mammary gland.

Constitutive mammary tumors appear to frequently be composed of cells with

different cycling patterns as demonstrated by the retention of H2B-eGFP label in a small

subset of cells in both luminal and basal labeled mammary tumors. The exact

characterization of these cells would provide insight into the maintenance and growth of

mammary tumors and possibly provide a candidate population of CSC based on their

relatively infrequent cycling. While slower cycling populations within the growing tumor

were common among the samples studied, tumors did not appear to contain non-cycling cells

retaining H2B-eGFP label to the same extent as in pulse-labeled mammary tumors.

Mammary tumor cells seem to have heterogeneous proliferation patterns though the exact

mechanisms driving those patterns are at this point unclear. Isolation of LRCs from pulse-

chased mammary tumors would allow analysis of expression patterns and possible

identification of the mechanisms underlying altered proliferation rates. It is likely that the

distinct proliferation patterns of tumor cells results from specific capabilities of those cells

and plays a role in normal tumor development.

Characterization of LRCs, and comparison to washed-out cells, by a variety of means

should provide information on the heterogeneity inherent to mammary tumors. Functional

characterization of LRCs could be obtained by suspension of single isolated LRCs in

Page 101: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

84

Matrigel and monitoring of their growth capacity within that environment. If slow cycling

cells are enriched for stem cell like behaviors, increased diversity within resulting outgrowths

would likely be visible in the 3D cultures and could be imaged in real time using the

techniques described in Chapter 3. Expression profiling may provide information about the

growth regulation mechanisms used to maintain cells relatively dormant within the confines

of rapidly growing tumor. Finally, implant studies defining the capacity of LRCs to

recapitulate both mammary tumor and potentially normal mammary gland depending on the

impact of stroma on malignant growth would be useful in characterizing LRCs.

Tumor to tumor variability proved a confounding feature for the identification and

characterization of LRCs in mammary tumors. The wide range of GFP labeling proportions

seen in pulse-labeled tumors may be due to additional genetic lesions acquired during

tumorigenesis that are different in each independent tumor. Identifying these lesions and

possibly the effects they have on tumor growth and progression may help define the changes

that cause the distinct H2B-eGFP labeling patterns. Categorizing primary tumors based on

common mutation spectra may also allow further identification of commonalities in GFP

washout patterns.

As mammary tumors arise from a variety of genetic insults, the variability seen in

this work is not necessarily problematic but is a confounding factor for the isolation of a

uniform population from a series of primary tumors. By using explants we attempted to

circumvent this variability with some success in creating sequential generations that recreate

the pulse and pulse-chase patterns of the primary tumor. While general response to Dox-

pulse was carried through generations of explanted tumors, precise percentage of H2B-eGFP

labeled nuclei was variable. This variability may be due to the regional variation in tumor

Page 102: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

85

composition and the relatively small sample taken during tumor biopsy surgeries. Though

not labeled in identical proportions across generations, explants did provide a manner to

obtain pulsed and pulse-chased data from matched tumor sets. Future work may utilize a

similar strategy of serially propagating a mammary tumor but instead of tumor fragments

may use single cells of a specific label retention history to sequentially populate the explants.

A study done in this manner would not only expand the population available for different

experimental conditions but would also help eliminate the sample variability seen by using

tumor fragments consisting of many cells for explants.

Variability in H2B-eGFP labeling seen in mammary tumors while on Dox may

alternatively result from alteration of the expression of the H2B-eGFP transgene. Silencing

may result in the abrogation of fluorescent histone expression. Bisulfite PCR may provide an

avenue to examine the methylation related silencing of the transgene. The absence of a

proliferation benefit for cells expressing H2B-eGFP suggests a potential selection against the

transgene during tumor progression and sequencing of DNA from non-labeling tumors to

identify the continued maintenance of H2B-eGFP transgene in tumor tissue.

The transition of tumors across several generations from adenocarcinoma to

mesenchymal tumors eventually eliminated the utility of each explants series. This transition

varied in length from tumor to tumor, possibly due to genetic variation inherent to the tumor,

and was accompanied by a loss of response to H2B-eGFP induction as expected due to the

loss of luminal type cells. The sustained time frame of growth of the cells explanted serially

may have caused the decay of the epithelial type tumors.

While preliminary study of labeling and washout in constitutive mammary tumors

proved to be complicated by the variability encountered in primary tumors, serial explants

Page 103: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

86

provided a more consistent population of tumors for study. Additional work done on more

characterized primary tumors selected for their labeling characteristics and genetic profiles

will provide more insight into the particular patterns of proliferation within mammary

tumors. The characterization of primary tumors and the functional characterization of single

labeled and label retaining cells from those tumors should elucidate the role of distinct cell

populations that appear to comprise mammary tumors driven by MMTV-wnt1.

Page 104: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

87

Figure 4.1 Mammary Adenocarcinomas Arising in the MMTV-wnt1 Model Show a

Mixed-Lineage Phenotype.

The upper panels show immunohistochemistical identification of Keratin-8-positive, luminal-

type tumors cells (left) and Smooth Muscle Actin (SMA)-positive basal-type tumor cells

(right). Keratin-8 labeled cells make up the luminal type cells in the mammary tumor and

SMA expressing cells are from the basal cell type. Mammary adenocarcinoma contain cells

of both luminal and basal types however the tumorigenic capacities of these two populations

remains undefined with both populations potentially exhibiting tumor initiating behaviors.

Page 105: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

88

Figure 4.2 Strategy for the Dox-Dependent H2B-eGFP Labeling of Either Basal-Type

or Luminal-Type Tumor Cells in the MMTV-wnt1 Model.

A. Schematic of Transgenes Used. Transgenic mice carrying the MMTV-rtTA transgene

express the rtTA transcription factor in the luminal mammary epithelium but do not express

independently integrated responder transgenes driven by minimal CMV promoters

containing multimerized operator sequences. Similarly, transgenic mice carrying the K5-

rtTA transgene express the rtTA transcription factor in the basal mammary epithelium. In

the presence of inducer, rtTA binds Dox and undergoes a conformational change that allows

recognition of Tet operator sequences and responder transgene activation. Transgenic lines

used in this study permit Dox regulated expression of the reporter gene: Histone H2B-eGFP

fusion protein (TGFP line). Transgenic mice carrying the MMTV-wnt1 transgene express the

wnt1 oncogene in the luminal epithelium in a constitutive manner. B. Fluorescent Labeling

of Mammary Tumors. Mice carrying the MMTV-wnt1 transgene develop mammary tumors.

Following tumor development, transgenic animals also bearing the TGFP and either MMTV-

rtTA or K5-rtTA transgenes are placed on doxycycline in order to induce H2B-eGFP

expression and incorporation into the chromatin. Removal of doxycycline from the animals’

diet results in cessation of H2B-eGFP signal production and allows washout of GFP signal

from dividing cells within the tumor.

Page 106: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

89

Figure 4.3 Homogeneous Versus Heterogeneous Models of H2B-eGFP Washout in

Growing Mammary Tumors.

The schematic depicts growth of a mammary tumor driven by a constitutive oncogene. Dox

treatment triggers H2B-eGFP labeling of a tumor cell compartment. Cessation of Dox

treatment initiates H2B-eGFP washout, which can occur in at least two distinct ways; the

first, that all cells in the tumor proliferate at similar rates and evenly washout GFP signal in

all cells or alternatively, cells proliferate at different rates leading to differential washout and

the identification of bright label retaining cells. The identification of bright LRCs resulting

from either luminal or basal driven GFP labeling provides the opportunity to isolate and

characterize a distinct population within a mammary tumor.

Page 107: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

90

Figure 4.4 Wide Variation in the Extent of H2B-eGFP Labeling of Luminal-Type

Tumor Cells in the MMTV-Wnt1 Model.

Pulse labeling of tritransgenic (MMTV-rtTA/MMTV-wnt1/TGFP) with doxycycline brightly

labels mammary tumors, however variability in the intensity of staining is common. A.

H2B-eGFP Fluorescence in Gross Tumor Specimens. A photograph of a fragment of H2B-

eGFP labeled mammary tumor excised from a tritransgenic Dox-treated mouse under long

UV illumination. B. H2B-eGFP Fluorescence in Tumor Sections. A representative Confocal

microscopy image (Mag. 40X) of a Hoechst counterstained frozen section prepared from a

mammary tumor of a Dox-treated tritransgenic animal. GFP labeled cells are contained in the

luminal cell nests present within the mammary tumor. C. Flow Cytometric Detection of

H2B-eGFP Fluorescence. Single cells isolated from independent primary mammary tumor

samples collected from Dox-induced tritransgenic animals were analyzed by flow cytometry

for GFP signal intensity. The graph depicts the range of H2B-eGFP signal detected with each

bar representing a single independent tumor. The percent GFP positive cells was calculated

as the total population of cells with signal intensity greater than 102.5

.

Page 108: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

91

Figure 4.5 Wide Variation in the Rate of H2B-eGFP Washout from Luminal-Type

Tumor cells in the MMTV-Wnt1 Model.

Single cells isolated from 16 primary mammary tumors of MMTV-rtTA/MMTV-wnt1/TGFP

tritransgenic mice were analyzed for GFP content by flow cytometry following 1 week of

Dox treatment (biopsy) and following 1 week of Dox-withdrawal (necropsy). Each graph

represents an individual tumor. Blue lines are pulse-labeled biopsies (Dox treated) while red

lines are from tumors at necropsy following Dox withdrawal. Wide variability of both extent

of GFP labeling during pulse and amount of GFP signal dilution during chase can be

visualized. The right two columns contain tumors that responded to Dox while the left

column contains those tumors without detectable GFP induction in the presence of Dox.

Percent GFP positive cells during Pulse (blue) and following chase (red) are calculated as

cells with GFP signal intensity greater than 102.5

as indicated by the bars in the top of each

column.

Page 109: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

92

Figure 4.6 Generating Clonally-Related Tumor Outgrowths by Explanting onto Syngeneic

Host Mice.

Dox-independent tumors from MMTV-wnt1/MMTV-rtTA/TGFP are biopsied and divided

into small fragments. These fragments are then implanted on to the right and left flanks of

Dox naïve wild-type host mice. Following a period of weeks, tumor outgrowths develop at

the implant locations. Groups of “sibling” clonal tumor explants can then be exposed to

various pulse-chase timelines in order to analyze labeling and label retaining cells in growing

constitutively driven mammary tumors. Additional rounds of explanting achieve expansion

of the tumor pool available for study.

Page 110: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

93

Figure 4.7 Variable H2B-eGFP Labeling Among Clonally-Related MMTV-Wnt1

Mammary Tumor Explants.

A. H2B-eGFP Fluorescence in Gross Tumor Specimens. These images depict an on Dox,

wild-type FVB female bearing bilateral tritransgenic tumor explants at necropsy. Tumor

explants contain bright green H2B-eGFP signal upon examination under excitation

illumination and were photographed using XXX filter to optimize fluorescence detection and

imaging. Images: Conventional digital photography: white light (left) and UV illumination

(right). B. H2B-eGFP Fluorescence in Tumor Sections and Flow Cytometry. Representative

samples of Dox-naïve (left) and on Dox (right) mammary tumor explants collected from host

mice were prepared for Confocal microscopy imaging with Hoechst counterstain (Mag. 40X)

and flow cytometric assessment of GFP signal intensity and demonstrate the expected

absence and presence of H2B-eGFP labeling in mammary tumors dependent on Dox

induction. C. Flow Cytometric Detection of H2B-eGFP Fluorescence in Tumor Explant

Generations. The graph depicts the GFP fluorescence as quantified by flow cytometry for

mammary tumor explant generations. Single cells isolated from pulse-labeled mammary

tumors were analyzed by flow cytometry for GFP signal intensity. Primary tumors and

subsequent generations are depicted sequentially. Many generations of tumor explant have 2

samples composed of the left and right explants.

Page 111: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

94

Figure 4.8 Evidence for Heterogeneous H2B-eGFP Washout in a Subset of MMTV-Wnt1

Tumor Explants.

A. H2B-eGFP Fluorescence and Washout in Gross Explant Tumor Specimens. Photographs

depict UV illuminated gross tumor fragments from sibling tumors under conditions listed.

Dox-treated tumors are brightly green fluorescent even to the naked eye with excitation.

Previously Dox-treated tumors subjected to Dox withdrawal exhibit dimmer GFP

fluorescence but are still easily distinguishable from the unlabeled control Dox-naïve tumor.

B. H2B-eGFP Fluorescence and Washout in Explant Tumor Sections. Panels depict

Confocal microscopy images of frozen tumor sections counterstained with Hoechst (Mag.

20X). Dox-naïve, Dox-treated and post-Dox treated chase mammary tumor samples images

demonstrate labeling and label retained following pulse-chase in a growing mammary tumor

sibling group. Only a minority of labeled cells retain bright GFP fluorescence following a

one week Dox withdrawal.

Page 112: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

95

Page 113: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

96

Figure 4.9 Variable H2B-eGFP Washout Among Clonally-Related MMTV-Wnt1

Mammary Tumor Explants.

Single cells isolated from primary mammary tumors of MMTV-rtTA/MMTV-wnt1/TGFP

tritransgenic tumor explants were analyzed for GFP signal intensity by flow cytometry. Each

graph is an individual tumor generation with sequential generations appearing in the same

column and numbered by passage generation. The arrow depicts the lineage of a tumor that

was re-explanted from frozen storage of a sample from P4. Dox-naive animals appear in

Blue, Dox-treated tumors are Magenta, and Dox-treated tumors subjected to Dox withdrawal

prior to harvest appear in both Black and Yellow.

Page 114: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

97

Figure 4.10 Chemotherapy with Adriamycin Minimally Impacts H2B-eGFP Labeling

and Washout in MMTV-Wnt1 Tumor Explants.

Single parameter plots depict the GFP signal intensity of single cells isolated from sibling

tritransgenic (MMTV-rtTA/MMTV-wnt1/TGFP) mammary tumor explants on host mice.

Colors represent single mice with each mouse bearing two sibling tumors. A. Adriamycin

Treatment does Not Induce Fluorescence. Dox-naïve animals treated with adriamycin 4 days

prior to necropsy and did not have detectable GFP fluorescence. B. Adriamycin Treatment

does Not Appear to Alter GFP Labeling. The yellow and black lines represent the GFP

fluorescence detected in pulse labeled tumors (8-14 days on Dox) treated with adriamycin 4

days prior to necropsy while the pink lines depict the fluorescent signal found in an untreated

mouse on Doxycycline for 4 days. C. Proliferating MMTV-wnt1 Initiated Mammary Tumors

Dilute GFP Signal. Yellow and black lines depict GFP signal intensity of cells isolated from

4 day pulse labeled tumors from mice chased for 6 days and 10 days (Pink lines). D.

Adriamycin Treatment does not Alter H2B-eGFP Washout Pattern in Mammary Tumors.

Graph demonstrates the absence of a change in GFP washout in 4 day pulsed- 10 day chased

mammary tumor explants following adriamycin treatment. Yellow, Blue, and Black lines are

mammary tumors from mice treated with adriamycin while the pink lines are from untreated

mice.

Page 115: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

98

Figure 4.11 H2B-eGFP Labeling of Basal-Type Tumor Cells in the MMTV-Wnt1

Model.

Tritransgenic K5-rtTA/TGFP/MMTV-Wnt1 female mice were generated and monitored until

mammary tumors were detected. Mice were subsequently placed on Dox for 4 days to

induce H2B-eGFP labeling of Keratin 5 expressing cells within the mammary gland and

mammary tumor. Following Dox-treatment, animals were removed from Dox for 12 or 14

days as indicated. A. Microscopic Imaging of Primary Tumor Sections. Representative

fluorescence microscopy images (Blue and Green channels, Mag. 20X) counterstained with

Hoechst demonstrate widespread labeling with H2B-eGFP following pulse in both the

mammary tumor and surrounding normal ducts (left). Mice were subsequently removed from

Dox for 2 weeks and washout of H2B-eGFP signal occurred reducing the number of bright

GFP positive nuclei identified by fluorescence microscopic examination of sections (right).

Relatively slower proliferating normal mammary ducts surrounding the tumor retain GFP-

label at a much higher level than that seen in growing tumors. B. Confocal Imaging of

Explant Tumor Sections. Representative fluorescence microscopy images (Mag. 20X)

counterstained with Hoechst demonstrate widespread labeling with H2B-eGFP. Similarly to

what is seen in primary tumors, sibling tumor explants of K5-rtTA/MMTV-wnt1/TGFP

tumors treated with Dox contained large numbers of brightly fluorescing nuclei while

increasing periods of chase resulted in decreased GFP fluorescence. C. Flow Cytometric

Detection of H2B-eGFP Signal in Mammary Tumors. Single parameter plots depict results of

flow cytometric analysis of single cells isolated from basal labeled constitutive mammary

tumors and confirm the patterns seen in frozen sections where pulse labeled tumors exhibit

bright GFP fluorescence while pulse/chase tumors have decreased GFP intensity.

Page 116: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

99

Page 117: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

100

Figure 4.12 MMTV-Neu Driven Tumors Contain a Homogeneous Cell Population

Capable of H2B-eGFP Labeling.

Dox-treated MMTV-Neu initiated mammary tumors were collected by biopsy after 2-3

weeks of Dox treatment. Animals were subsequently removed from Dox treatment to allow

washout of H2B-eGFP signal prior to necropsy and collection of mammary tumor fragments

for analysis by fluorescent microscopy and flow cytometry. A. Preponderance of luminal-

type cells in the MMTV-Neu model. Representative immunophenotype profiles of tumor

cells derived from MMTV-Neu and MMTV-Wnt models. Nearly all MMTV-Neu tumor

cells were CD24high

, indicating luminal character. B. H2B-eGFP labeling of MMTV-Neu

tumor cells. B. Single Parameter flow of Dox treated and Dox-withdrawal tumor

fluorescence. A representative pair of MMTV-neu/MMTV-rtTA/TGFP tumors during Dox

treatment and following 1 week chase were prepared as single cells and analyzed for GFP

signal intensity. C. GFP Signal detected in MMTV-Neu initiated mammary tumors. The

graph depicts the GFP fluorescence as quantified by flow cytometry for independent MMTV-

Neu initiated mammary tumors both during Dox treatment and following 1 week of Dox

withdrawal. Single cells isolated from pulse-labeled mammary tumors were analyzed by flow

cytometry for GFP signal intensity. D. H2B-eGFP Fluorescence in Tritransgenic MMTV-

Neu-driven Tumors Sections. Hoechst counterstained frozen tumor sections from Dox

treated (1 week on Dox), and Dox-withdrawal (1 week on Dox then 4 weeks off Dox)

MMTV-rtTA/ MMTV-Neu/TGFP mammary tumors were imaged by fluorescence

microscopy demonstrating nuclear H2B-eGFP fluorescence incorporation and washout

following of Dox chase (Mag 20X).

Page 118: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

101

Chapter 5 H2B-eGFP Labeling and Proliferation Dynamics in Reversible Mammary

Tumors and Minimal Residual Disease Lesions

5.1 Introduction

While treatment of breast cancers frequently results in elimination of clinically detectable

tumor burden many mammary tumors recur months or even years following treatment.

Relapse has been attributed to a latent population of malignant cells which eventually

repopulate the tumor after a long interval of relative dormancy [174, 195, 196]. The

mechanisms by which these tumor cells are maintained are poorly understood. Here, we

used an H2B-eGFP labeling strategy in the context of a reversible breast cancer model to test

whether residual tumor cells are maintained in a state of dynamic equilibrium or in a state of

cellular dormancy. Specifically, mice were engineered to enable pulse-chase labeling of

tumor cells of Wnt1-dependent mammary cancers. In chapter 3, we established the role of

proliferation in the dilution of H2B-eGFP label in normal MECs. Here we extend this

methodology to study the cycling dynamics of cells within dormant malignant lesions of the

mammary gland. By identifying the cycling status of the cells comprising minimal residual

disease lesions (MRD) we explore the cellular mechanisms that maintain dormant

malignancies in breast cancers.

5.2 Experimental Strategy for the Study of Proliferation Dynamics in Reversible Mammary

Tumors and Minimal Residual Disease Lesions

Our strategy for H2B-eGFP labeling of luminal epithelial cells within reversible

mammary neoplasia is depicted in Figure 5.1A. The transgenes used for luminal MEC-

restricted H2B-eGFP labeling in normal mammary tissue were paired with an additional Tet-

responder transgene expressing the Wnt1 oncogene. In tri-transgenic MMTV-

rtTA/TWNT/TGFP animals, Wnt1 and H2B-eGFP should be co-expressed in a Dox-

Page 119: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

102

dependent manner. In previous work, we showed that expression of inducible Wnt1 in the

mammary epithelium of MMTV-rtTA/TWNT mice during chronic Dox treatment induced

widespread mammary hyperplasia that progressed to yield mammary tumors which arise

stochastically after several months. Notably, mammary tumors generated in this manner are

Dox-dependent since Dox-withdrawal results in the abrogation of Wnt1 expression and

tumor regression. Regressed tumors leave behind clinically undetectable MRD lesions that

retain latent malignant potential. Consistent with MRD harboring dormant mammary cancer,

these lesions serve as a source of spontaneous Dox-independent tumor relapse during

extended Dox-withdrawal [174]. Moreover, MRD lesions rapidly regrow into measureable

tumors following the readministration of Dox (Figure 5.1B) [174]. The transgene

combination MMTV-rtTA/TWNT/TGFP is designed to permit H2B-eGFP labeling of

hyperplastic mammary ducts and growing mammary tumors. As the H2B-eGFP and Wnt1

transgenes are co-regulated, Dox-dependent H2B-eGFP labeling of tumor cells should cease

coincident with oncogene withdrawal. As such, only those tumor cells residing in MRD that

cycle infrequently should retain bright H2B-eGFP label during extended Dox-withdrawal.

Two possible models for the maintenance of MRD lesions exist and are depicted in

Figure 5.2. According to the first model, MRD lesions are maintained in a dormant state

through a dynamic equilibrium. Here, the malignant cells comprising the lesion proliferate,

but total lesion size is maintained through roughly equal rates of cell death. According to the

second model, the tumor cells contained within the dormant lesion are themselves maintained

in a dormant state; that is they are relatively slow cycling. We have utilized the H2B-eGFP

labeling system to analyze the cycling status of the cells within the MRD lesions to

distinguish between these two models. If the cells of the MRD are in a state of dynamic

Page 120: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

103

equilibrium, very little GFP fluorescence should be detectable following a period of washout

as the cellular divisions will have diluted any H2B-eGFP. However, if MRD are maintained

by cellular dormancy, a large fraction of the epithelial cells composing the lesion will retain

bright H2B-eGFP signal due to their relatively low rates of proliferation.

5.3 Doxycycline-dependent Labeling of Inducible Mammary Tumors

In order to H2B-eGFP label reversible mammary tumors, a cohort of tritransgenic

MMTV-rtTA/TGFP/TWNT mice were subjected to chronic Dox treatment to inducibly co-

express the H2B-eGFP and Wnt1 transgenes. Mice developed mammary tumors that were

labeled with bright GFP fluorescence throughout the luminal compartment. In some cases,

animals were given a single dose of MNU in order to accelerate mammary tumorigenesis and

increase mammary tumor multiplicity. As shown in Figure 5.3A, GFP signal was visible in

on Dox biopsy derived gross tumor fragments and in sections imaged by Confocal

microscopy. Confocal sections demonstrate the localization of GFP-positive nuclei along

the inner surfaces of mammary ducts. In addition to imaging studies, single cells isolated

from pulse-labeled mammary tumors analyzed by flow cytometry showed GFP fluorescent

signal (Figure 5.3B). A significant proportion of cells were GFPbright

with some tumors

exhibiting nearly 90% GFP-positive cells in the analyzed sample.

Unexpectedly, though the majority of tumors contained a subset of tumor cells labeled

with eGFP, the percent of tumor cells labeled with H2B-eGFP during Dox pulse varied

widely as depicted in Figure 5.3C. The percentage of cells labeled ranged from 0 -88% GFP-

positive. Nearly half of the tumors (10/22) that displayed any GFP fluorescence by flow

cytometric analysis contained 50% or greater GFP-positive cells. GFP signal intensity of the

fluorescing GFPbright

cells was similar between labeled tumors with differences in GFP-

Page 121: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

104

positive proportions resulting from decreased numbers of nuclei containing any detectable

label. This extensive variability in tumor behavior is illustrated again in the single parameter

plots depicted below the bar graph. Representative plots of GFP signal intensity from non-

labeling, labeling, and extensively labeling mammary tumors show the range of GFP

fluorescence induced by Dox stimulation.

5.4 MRD Retain Bright H2B-eGFP Label Following Tumor Regression

To test the models of maintenance of MRD lesions in the mammary gland, mice bearing

biopsy confirmed GFP labeled mammary tumors were subjected to Dox withdrawal to cause

the regression of mammary tumors. Tritransgenic mice were maintained off Dox for a period

of 6-8 weeks in order to allow complete tumor regression. Following this Dox-withdrawal,

animals were necropsied and MRD lesions were harvested along with normal surrounding

mammary tissue from non-tumor bearing mammary glands. We analyzed these samples first

by imaging mammary gland whole-mounts by wide-field fluorescence microscopy.

Subsequently, MRD lesions were processed to generate either tissue sections (for analysis by

Confocal microscopy) or single cell suspensions (for analysis by flow cytometry). MRD

lesions typically were only several mm in diameter, and thus too small to be divided for both

types of analysis.

In order to visualize GFP label retention, mammary glands harboring dormant

malignancies were digitally imaged using appropriate illumination to excite the GFP fusion

protein. As seen in Figure 5.4A, MRD lesions are identifiable in whole mounted mammary

glands by their darkened appearance. Some MRD lesions retaining bright GFP signal

following weeks of off-Dox washout were imaged to demonstrate GFP label retention within

the lesion. A non-tumor bearing mammary gland collected from the same mouse is also

Page 122: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

105

shown to compare the lack of H2B-eGFP fluorescence visible in the normal mammary tissue.

Subsequent Confocal microscopy of frozen sections from pulse-chased MRD lesions

counterstained with Hoechst confirmed the typical appearance of MRD lesions composed of

typical looking ductal structures organized in a densely branching cluster. These ducts were

brightly labeled with H2B-eGFP within the luminal layer and were surrounded by GFP-

negative elongated nuclei typical of myoepithelial cells.

The presence of brightly GFP positive cells following the extensive remodeling of the

MECs during tumor regression is consistent with a model of cellular dormancy for the

maintenance of MRD lesions in the mammary gland and in order to quantify H2B-eGFP

label retention at the cellular level we analyzed single cells isolated from MRD lesions by

flow cytometry. A single representative example of GFP single parameter data obtained of

an MRD retaining H2B-eGFPbright

cells following regression by this analysis is depicted in

Figure 5.4B. As pulse-labeled tumors varied in the intensity of GFP label incorporation

MRD lesions also exhibit a wide range of label retained within the cells. Figure 5.4C

matches the on Dox tumor biopsy and post-washout MRD necropsy data from a panel of

tumors. Tumors generally had higher fluorescence than their regressed counterparts probably

due to the extensive remodeling and cell death associated with tumor regression. Some

MRD lesions did retain substantial levels of H2B-eGFP fluorescence with many brightly

labeled cells remaining even following periods of weeks of Dox withdrawal suggesting that

cellular dormancy may be responsible for tumor dormancy in a subset of the tumors studied.

However, many MRD lesions retain only a small population of H2B-eGFP labeled cells

following Dox withdrawal.

Page 123: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

106

The labeling retained in MRD lesions following weeks of Dox-withdrawal is

particularly notable when compared to the washout observed in Dox-independent recurrent

tumors from MMTV-rtTA/TWNT/TGFP and in Dox-withdrawn MMTV-rtTA/MMTV-

wnt1/TGFP tumors. In these settings, continued proliferation of tumor cells rapidly dilutes

H2B-eGFP fluorescence. In constitutive Wnt1 tumors, little GFP fluorescence is detectable

even after substantially shorter periods of Dox-withdrawal (see Fig 4.9 and 4.10/ Figure

5.4D). Tumors from MMTV-rtTA/TWNT/TGFP mice that, though initially labeled with

H2B-eGFP, recurred off Dox also had very little residual H2B-eGFP fluorescence detectable

at necropsy (Figure 5.4D). The maintenance of GFP fluorescence in some MRD lesions

following the extended period of Dox withdrawal, particularly in comparison to the rapid

washout of signal in proliferative cell populations, suggests the possibility of cellular

dormancy as the method of label retention in MRD and potentially explains the maintenance

of dormancy of a subset of MRD lesions.

An interesting comparison to the label retained in MRD lesions is that seen in

surrounding regressed hyperplasia, the normal appearing ducts surrounding the MRD lesion

that do not derive from the mammary tumor. These areas are subject to the normal cycles of

proliferation driven by ovarian hormones. In Figure 5.5, representative sections of tumor

biopsy, and both MRD and regressed hyperplasia harvested at necropsy show varied levels of

H2B-eGFP label retention. Cells within some MRD lesions retained higher levels of H2B-

eGFP signal than those in the surrounding mammary epithelium as seen in 3 of the 6 sample

sets examined by flow cytometry (Figure 5.5B). One additional tumor set showed very

similar levels of label retention in the MRD and surrounding normal tissue, while the

remaining two pairs had higher percentages of label retaining GFP positive cells in the

Page 124: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

107

regressed hyperplasia than in the MRD lesion. Figure 5.6 depicts the matched single

parameter GFP plots for single cells isolated from the primary tumor, regressed tumor, and

regressed hyperplasia. GFP intensity was reduced in both regressed hyperplasia and MRD

lesions when compared to pulse-labeled mammary tumor due to infrequent cell divisions,

however in contrast to the washout seen in non-oncogene induced mice (Figure 3.10) GFP

fluorescence is maintained at a higher level with a larger proportion remaining GFP positive.

The variability in MRD label retention, particularly in comparison to the surrounding normal

mammary epithelium, suggests that independent mammary tumors may have distinct

pathways and mechanisms that result in tumor dormancy. However, in a subset of sample

sets, the MRD lesions have substantially higher levels of H2B-eGFP labeling implying a

relatively lower level of cellular proliferation than in the surrounding mammary ducts.

5.5 Discussion

The use of inducible H2B-eGFP labeling of reversible mammary tumors allows the study

of proliferation dynamics within regressed mammary tumors. A population of label retaining

cells is maintained in a subset of the regressed mammary tumors and these cells possible

contribute to the recurrence of mammary tumors as MRD lesions give rise to mammary

tumors rapidly following readministration of Dox. Though it is probable that the cells

comprising the MRD lesion contribute to relapsed mammary tumors they potentially are

maintained in the regressed lesion through alternate mechanisms that would preclude their

contributing to recurrent tumor growth such as senescence or cell death. As MRD lesions

appear stable over extended periods and can be reinsured by Dox stimulation, it appears

unlikely that the cells within these lesions are in fact irreversibly impeded from division.

While specific analysis of LRCs obtained from MRD has not been completed, we have

Page 125: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

108

examined the proliferative capacity of MECs isolated from MRD in 3D culture with Matrigel

and seen growth both in the presence and absence of Dox stimulations. This suggests that

these cells are not maintained due to senescence and are in fact viable.

The maintenance of these lesions appears to be driven by cellular dormancy in a subset of

cases as long term off-Dox chased MRD lesions retain brightly labeled nuclei in different

proportions. The identification of a dormant cell population within dormant malignancies

provides a clinical target for the development of follow-up strategies to ensure duration of

effective cancer treatments. Characterization of the mechanisms driving cellular dormancy

within residual disease also provides an interesting avenue for the chronic maintenance of

remission. If dormant malignancies can be forced to remain in the dormant state by external

means it may result in extension of remission periods. By identifying a dormant population

within MRD lesions a specific population possibly responsible for the dormant malignancy

and its capacity for recurrence can be studied.

The cancer stem cell theory has been proposed as a model for the driving force behind

tumorigenesis [197]. In this model, a small population of relatively infrequently cycling cells

gives rise to all other cells within the tumor. It is tempting to propose that the cells

comprising the MRD lesions are in fact a stem cell-like pool simply awaiting reactivation to

regenerate the tumor. Previous work has shown that cells in the MRD retain the ability to

give rise to a functional mammary ductal tree but still retain the latent malignancy and can

rapidly re-establish mammary tumors and so have some normal mammary epithelial stem

cell properties [174]. If some MRD lesions are in fact composed of cancer stem cells they

additionally provide an enriched source of these cells for study. Identification of particular

susceptibilities of these populations would possibly have repercussions in the development of

Page 126: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

109

treatment strategies not just of residual disease but of growing tumors as well. Isolation of

single label retaining and non-label retaining cells from residual lesions, followed by

implantation into the cleared fat pads would provide an avenue to investigate the capabilities

of LRCs from MRDs and into the possible links between normal and cancer stem cells

within the mammary gland. Preliminary characterization of isolated LRCs could also be

achieved in culture studies with or without extracellular matrix (Matrigel) and examining the

role of hormonal regulation on the maintenance of cellular quiescence.

While this study provides evidence supporting a model of cellular dormancy for the

maintenance of residual disease it does not exclude a role for immune surveillance or other

means of maintenance of dynamic equilibrium for the continuation of tumor dormancy. It is

likely that a combination of factors preserve tumor dormancy and that the accumulation of

multiple insults is necessary to reactivate tumor growth. This study also investigated only

the contributions and patterns of proliferation in luminal cells within mammary tumors and

MRD lesions. Similar work utilizing the basal-specific promoter K5 activation of H2B-

eGFP would provide an interesting counterpoint to the data collected for the luminal layer

and would provide insight into the role of basal cells in the maintenance of tumor dormancy

and in tumor reinitiation however would be confounded by the alteration of oncogene

expression pattern necessarily resulting from the change in transactivator.

Additional complications barring the extension of this method derive from the signal

variability seen in mammary tumors. Particularly, while a subset of the samples in this study

appear to behave in a manner consistent with cellular dormancy as the mechanism for the

maintenance of tumor dormancy, several MRD lesions did not conform to this model. It is

probable that independently derived mammary tumors accumulate distinct genetic mutations

Page 127: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

110

that result in different behaviors in both the primary tumor and in the resulting MRD lesions.

Genetic diversity in MMTV-Wnt1 induced tumors is common, primarily due to the reliance

on multiple mutations to transform initiated MECs. These additional mutations confer

distinct characteristics into each mammary tumor. For example, we have previously

demonstrated that a subset of MMTV-rtTA/TWNT derived tumors will spontaneously recur

following Dox-withdrawal and it is possible that the MRD lesions in these cases are

maintained by other less efficient mechanisms [174, 192]. As variability is a key feature of

these mammary tumors, it is unsurprising that the resulting MRD lesions are also variable.

The extensive variability inherent in mammary tumors, and all cancers, limits the potential to

identify a single mechanism that affects all tumors. So, while it appears probable that some

MRD lesions are maintained by cellular dormancy, it is also possible that dynamic

equilibrium accounts for the maintenance of MRD in other cases.

The analysis of labeling and subsequent quantification of label retention was complicated

by the extensive variability seen in initial fluorescence. Variability in tumor fluorescence

may also be attributed to regional variation in tumor structure and sampling error in the

selection of fragments for flow cytometric analysis. Intratumoral variability is primarily due

to areas of necrosis and different proportions of luminal/basal tissue within in particular

samples. As basal tissue was unlabeled in these experiments but is collected identically and

indistinguishable by single-parameter flow cytometric analysis, sample fragments with

higher ratios of basal cells could alter the perception of GFP labeling and GFP label retained

following Dox-withdrawal. There was not a readily available method to select regions of

tumor for analysis based on appearance and in fact, the heterogeneity is not apparent until

H+E sections are visualized following fixation.

Page 128: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

111

The variability of independent mammary tumors greatly complicated the study of the

mechanisms of MRD maintenance. Studies utilizing explants or implant surgical techniques

to amplify tumor fragments would provide larger pools of sibling tumors however several

attempts at propagating H2B-eGFP expressing tumors and MRD in this manner were

unsuccessful. This stumbling block was unexpected due to previously published work

demonstrating the capability of tumor fragments to proliferate following explant into

syngeneic hosts. Similarly, implanted fragments of MRD from non-TGFP mice have shown

successful outgrowth [174]. There is the potential that H2B-eGFP transgene interferes with

transplant studies thus blocking this method of increasing sample size. Potentially

overcoming this hurdle would allow several avenues of investigation to be pursued on

“identical” lesions and allow further characterization of the mechanisms involved in MRD

lesion maintenance. Specifically, amplified sample availability would allow simultaneous

flow cytometric analysis of H2B-eGFP fluorescence and cell cycle characteristics by

Pyronin/7AAD staining, histological examination, and for the culturing and expression

profiling of label retaining and unlabeled cell populations.

Page 129: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

112

Page 130: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

113

Figure 5.1 Strategy for the Concurrent Temporal Regulation of the H2B-eGFP and Tet-

op Wnt Transgenes Expression in Mammary Tumors and Minimal Residual Disease

Lesions.

A. Schematic of Transgenes Employed for Compartment Specific Expression . Transgenic

mice carrying the MMTV-rtTA transgene express the rtTA transcription factor in the luminal

mammary epithelium but do not express independently integrated responder transgenes

driven by minimal CMV promoters containing multimerized operator sequences. In the

presence of inducer, rtTA binds Dox and undergoes a conformational change that allows

recognition of Tet operator sequences and responder transgene activation. Transgenic lines

used in this study permit Dox regulated expression of the oncogene TWNT and the reporter

gene Histone H2B-eGFP fusion protein (TGFP line) in the luminal epithelium. B. Reversible

Expression of the TWNT Transgene Models the Formation of Minimal Residual Disease

Lesions. Bitransgenic (MMTV-rtTA/TWNT) mice subjected to Dox treatment stochastically

develop mammary tumors. Dox withdrawal causes these tumors to regress leaving behind

clinically undetectable dormant lesions referred to as minimal residual disease (MRD).

Readministration of Dox in the animals’ diet induces regrowth of the dormant lesion and

recurrence of the tumor.

Page 131: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

114

Figure 5.2 Two Models for the Maintenance of Minimal Residual Disease Lesions.

This schematic depicts the patterns of labeling and washout of fluorescent label in a

mammary tumor during the formation and maintenance of a dormant lesion. Following Dox-

induced tumorigenesis and H2B-eGFP labeling of the mammary tumor removal of

doxycycline results in tumor regression. We have previously demonstrated that MRD can

remain dormant for extended periods of time and still retain latent malignancy. Two possible

models exist to explain the maintenance of minimal residual disease lesions following tumor

regression. The first model is one of Dynamic Equilibrium; that is that the cells comprising

the lesion are in fact cycling but equal rates of proliferation and cell death result in no net

growth of the lesion. In this model, few if any cells in the MRD lesion will retain bright

H2B-eGFP label following extended Dox withdrawal. Contrasting this model, it is possible

that dormant cells make up the lesion. In this case, non-cycling cells in the MRD would

retain bright H2B-eGFP signal even following an extended period of off-Dox chase.

Page 132: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

115

Figure 5.3 Wide Variation in the Extent of H2B-eGFP Labeling of Luminal-Type

Tumor Cells in the MMTV-rtTA/TWNT/TGFP Model.

A. H2B-eGFP Fluorescence in Gross Mammary Tumors and Tumor Sections. The first panel

depicts a fragment of on Dox mammary tumor biopsy illuminated with light in the GFP

excitation wavelength showing widespread GFP fluorescence. The second image was

obtained via Confocal microscopy of a frozen section of pulse labeled mammary tumor

counterstained with Hoechst (Mag. 40X). B. Single Parameter Plot of Flow Cytometric

Detection of H2B-eGFP Signal. A representative single parameter plot demonstrating GFP

signal detected by flow cytometric analysis of single cells isolated from mammary tumor

biopsies of tritransgenic animals treated with Dox depicts GFP signal intensity of the tumor

cells shows two populations, a large GFP positive population with fluorescence greater than

102 (40.08%) and a GFP negative population. C. Flow Cytometric Detection of H2B-eGFP

Fluorescence in Mammary Tumors. Biopsy samples from 25 independent mammary tumors

treated with Dox were prepared as single cell suspensions and analyzed for GFP signal by

flow cytometry as shown in B. The chart in panel C demonstrates the range of signal

intensities observed by these analyses with each bar depicting a single mammary tumor.

Samples marked with stars are those tumors arising in mice treated with MNU.

Page 133: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

116

Page 134: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

117

Figure 5.4 Minimal Residual Disease Lesions Retain H2B-eGFP Signal Following

Tumor Regression.

Tritransgenic MMTV-rtTA/TWNT/TGFP mice placed on Dox develop mammary tumors.

These tumors can be brightly labeled with H2B-eGFP and subsequent removal of Dox results

in tumor regression and the formation of MRD lesions. A. Detection of GFP Signal in Gross

Mammary Tumor Fragments and MRD Lesions. The top left image in Panel A depicts a

fragment of on Dox mammary tumor biopsy illuminated under UV with widespread green

signal visible throughout the tumor. The right two images depict mammary gland whole

mounts obtained from a pulse then chase labeled tritransgenic animal and photographed

under UV illumination. The left whole mount contains an MRD lesion with visible GFP

signal (see lower left image for a larger view of the MRD) while the right whole mount

image shows a tumor-free mammary gland from the same animal. The non-tumor bearing

gland from the same animal at necropsy provides a comparison for the lack of fluorescence

seen in non-malignant tissue. B. H2B-eGFP Fluorescence in Tumor Sections and Single Cell

Preparations. Representative images obtained via Confocal microscopy of frozen sections of

a pulse labeled mammary tumor and of a pulse-chased MRD lesion counterstained with

Hoechst (Mag. 40X). Single parameter plots demonstrate flow cytometric analysis of single

cells isolated from mammary tumor biopsies and MRD lesions and provide a quantitative

assessment of GFP fluorescence in tumor cells and residual lesions. Residual lesions retain

bright GFP signal in a significant portion of their cells. Shown is a representative plot of

GFP signal intensity from single cells from a pulse-labeled mammary tumor and a

pulse/chase MRD lesion. C. Flow Cytometric Detection of H2B-eGFP in Mammary Tumors

and MRD Lesions. Graph depicts thirteen independent mammary tumor biopsy-necropsy

pairs that were analyzed for GFP signal intensity by flow cytometry. Blue bars depict the %

GFP positive cells detected in on Dox tumor biopsies. Magenta bars represent the percent

GFP positive cells in matched off Dox MRD lesions. A wide range of initial GFP labeling

present in on Dox tumor samples is again demonstrated in the range of GFP signal intensity

seen in the corresponding MRD lesions. While many MRD lesions maintained a large

proportion of cells with bright green signal even after 6-8 weeks off Dox, some samples

demonstrate greater washout of the GFP signal.

Page 135: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

118

A.

B.

Figure 5.5 MRD Retain Greater Fluorescent Label than Surrounding Normal Ducts.

A. Imaging of Frozen Sections of Tumor, MRD, and Regressed Hyperplasia. Panel depicts

representative Confocal images obtained from frozen sections of Dox-treated mammary

tumor and post-Dox withdrawal MRD lesion and surrounding non-tumor ducts

counterstained with Hoechst (MAX Projections; Mag. 20X). MRD retain H2B-eGFP

labeling even following extended periods of chase and the extensive remodeling associated

with tumor regression. The relative quiescence of the cells of the MRD lesion is particularly

apparent when compared to the ducts in the surrounding normal tissue which only have

infrequent bright GFP labeled cells. B. Flow Cytometric Detection of H2B-eGFP

Fluorescence in Mammary Tumors, MRD Lesions, and Regressed Hyperplasia. The graph in

panel B summarizes GFP fluorescence results obtained by flow cytometry for 6 independent

matched sets of Dox-treated tumor biopsies and MRDs and Regressed

Hyperplasia/Mammary Gland harvested at necropsy following Dox-withdrawal.

Page 136: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

119

Page 137: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

120

Figure 5.6 Flow Cytometric Analysis of GFP Signal Intensity in Matched Tumor

Biopsy, MRD Lesions, and Regressed Hyperplasia.

Single parameter plots display the range of GFP signal detected in each sample and GFP-

positive percentages are listed in each plot and were determined using the bracketed area

depicted in the top panel of each column. Six matched sets of Dox-treated tumor biopsies,

and MRD lesions and Regressed Hyperplasia (normal ducts) following Dox-withdrawal were

prepared as single cell suspensions. These cells were analyzed by flow cytometry for H2B-

eGFP fluorescence. MRD lesions frequently retained bright GFP signal even following 6-8

weeks of chase. In particular, the samples in rows 1, 3, and 4 demonstrate greater retained

fluorescence in the MRD than in the surrounding normal mammary tissue (Regressed

Hyperplasia).

Page 138: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

121

Chapter 6 A Transgenic Model for Short-Term Lineage Tracing of Mammary

Epithelial Cells in Pregnancy

6.1 Introduction

The changes in mammary gland structure that accompany pregnancy are vast and

rapidly expand the population of cells that compose the ductal tree. While the outcome of

these hormonally-mediated changes has been clearly described and many studies have

examined the regulation of growth during this phase, the precise cellular contributions to

growth and differentiation have not been clearly defined. The most widely held model

suggests that both cell layers within alveoli develop from a single cell or group of cells of

luminal origin (Figure 6.1). Here we adapt the H2B-eGFP labeling scheme utilized in the

previous chapters in order to examine the unique proliferation patterns and development of

the mammary gland during pregnancy. By pairing the H2B-eGFP reporter gene either

luminal or basal-specific MEC transactivators we tracked the proliferation of labeled cells in

each compartment throughout the rapid morphogenesis program that creates alveoli. Using

this method of short-term lineage tracking we identified cells of both luminal and basal origin

that contribute to lobuloalveolar outgrowths. Furthermore, we use the dilution of GFP signal

in dividing cells to show that the influx of MECs from each compartment during

alveologenesis involves proliferation and not merely MEC migration.

Having examined the proliferation and lineage patterns of mammary epithelial cells

during pregnancy, we then extended our studies to examine the remodeling of the mammary

gland during involution. Early first pregnancy has been shown to protect against breast

cancer in women but the exact mechanisms for this parity-related protection are unknown

(Figure 6.2). According to one proposed model, the widespread MEC loss incurred during

involution, by eliminating a large proportion of the MECs that exist during pregnancy,

Page 139: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

122

deletes MECs containing pre-neoplastic mutations. Whether MECs from each compartment

persist through involution and contribute to the architecture of the parous gland remains

unknown, The work presented here provides a potential avenue for further study by

demonstrating the persistence of labeled cells from lactation through post-lactational

involution. The persistence of labeled cells from both cell layers following involution

suggests that MEC deletion, though widespread, is incomplete.

In order to examine the contributions of luminal and basal cells to the lobuloalveolar

outgrowths induced by pregnancy and hormonal stimulation and the persistence of cells

through involution we used the MMTV-rtTA and K5-rtTA transactivators to direct tetO-

driven H2B-eGFP labeling of luminal and basal MECs respectively as described in Chapter

3. We then performed pulse-chase labeling experiments during timed pregnancy, during

hormonal stimulation, and following pregnancy to track luminal and basal MEC fates during

MG morphogenesis and remodeling (Figure 6.3). We utilize the H2B-eGFP labeling scheme

to both trace cell layer of origin of the cells contributing to the alveolar outgrowths and to

assess the role of proliferation in the development of alveoli.

Our method of labeling cells in each compartment prior to the induction of

alveologenesis provides a unique manner of examining the contributions of each cell layer to

the completed lobuloalveoloar structure. As depicted in Figure 6.4, there are three possible

outcomes of each set of labeling experiments. Compartment labeled cells may contribute to

the same layer of the developing alveoli, to both cell layers of the alveoli, or may not

contribute at all. The current model of alveologenesis predicts a population of cells of

luminal origin that give rise to both luminal and basal layers of the lobuloalveolar outgrowths

induced during pregnancy and this outcome is highlighted in orange in Figure 6.4. We use

Page 140: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

123

H2B-eGFP labeling to explore the contribution of luminal and basal cells to alveoli and to

verify that alveolar cells of each compartment in the alveolar structure result from cell

division and not migration.

6.2 Contributions of Mammary Epithelial Cell Sub-types to Lobulo-alveolar Outgrowths of

Pregnancy

As depicted in Figure 6.4, the exact contributions of each cell type to the

lobuloalveolar outgrowths induced during pregnancy are unclear. In these studies, we

examined luminal and basal specific labeling and washout of H2B-eGFP signal at mid-

pregnancy by inducing labeling prior to pregnancy and then allowing alveologenesis to occur

during chase. We utilize fluorescent imaging of mammary gland whole-mounts, flow

cytometric assessment of GFP fluorescence of individual cells isolated from mammary

glands, and Confocal imaging of Hoechst counterstained frozen sections to track and identify

labeled and label-retaining cells within the mammary ducts and alveolar outgrowths.

6.2.1 Tracing the Contributions of Luminal Mammary Epithelial Cells to Lobulo-alveolar

Outgrowths During Pregnancy

In order to track the contribution of the luminal lineage to alveoli we pulsed adult

virgin female bitransgenic MMTV-rtTA/TGFP mice with Dox prior to pregnancy. These

mice were then paired with breeder males and examined daily for vaginal plugs. At plug,

mice were subjected to Dox withdrawal and sacrificed at day 10.5 of pregnancy (Figure 6.3).

One inguinal mammary gland was prepared for whole mount imaging and the remaining

mammary tissue was digested to generate single cell suspensions for flow cytometric analysis

of GFP signal intensity. Control mammary gland samples were analyzed from d10.5

Page 141: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

124

pregnant females that were either Dox-naïve or placed on Dox for 3 days prior to necropsy

(Pulse-only).

Examination of pulse-labeled whole-mounts by wide field microscopy showed

widespread H2B-eGFP labeling throughout the ductal structures and lobulo-alveolar

outgrowths, whereas Dox-naïve whole-mounts lacked detectable H2B-eGFP labeling as

expected (Figure 6.5). Mammary glands harvested under pulse-chase conditions contained

brightly labeled cells within the ducts of the mammary tree and additionally exhibited

dimmer GFP-positive nuclei within the lobulo-alveolar outgrowths.

To quantify washout in the mammary epithelium during pregnancy we analyzed

single cells isolated from experimental mice in all 3 groups. Pulse-labeled mice contained

40% of bright GFP positive nuclei while Dox-naïve MECs lacked detectable bright GFP

nuclei as depicted in Figure 6.5. In the setting of pulse-chase, MECs showed a wide range of

GFP signal intensity. Single parameter plots revealed populations of MECs clustered in

peaks corresponding to halving of GFP signal attributable to cell divisions.

Next, the mammary gland whole-mounts were processed to allow imaging of

Hoechst-counterstained frozen sections by Confocal fluorescent microscopy (Figure 6.6). In

these images, bright GFP signal is seen lining the ducts and along the inner surfaces of

alveolar structures of pulse-labeled mammary glands while no GFP positive nuclei were

visualize in Dox-untreated animals. Imaging of pulse-chased mammary glands showed GFP

positive nuclei lining the inner surfaces of mammary ducts and along the inner lining of

alveolar outgrowths confirming the contribution of luminal duct cells to the alveolar

outgrowths. These nuclei were generally less bright than those found in on Dox animals

however occasional bright GFP nuclei were observed. In particular, GFP positive nuclei

Page 142: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

125

within the lobulo-alveolar outgrowths were dimmer than those observed within the

surrounding ducts.

Taken together, our imaging studies and flow cytometry analysis provide compelling

evidence that the proliferating progeny of luminal MECs contribute to alveolar outgrowths.

Decreased H2B-eGFP signal intensity within the outgrowths suggests cellular proliferation of

an initially labeled luminal population drives formation of lobulo-alveolar growths. As we

have previously demonstrated the proliferation dependent washout of H2B-eGFP labeling in

the mammary gland, the presence of cells with decreased GFP signal in the alveolar

outgrowths supports a proliferative rather than migratory origin of these cells. If luminal

labeled ductal cells migrated outward to form the bulk of the alveolar cells than similar

intensity of fluorescence between ductal and alveolar cells would be visible. Strengthening

this conclusion, by pairing the finer resolution of GFP signal intensity obtained by flow

cytometric analysis with the location identifying analysis obtained by microscopic imaging

we confirm cells with decreased overall GFP intensity and can propose that they are the

dimmer labeled cells that compose the alveolar outgrowths.

6.2.2 Tracing the Contributions of Basal Mammary Epithelial Cells to Lobulo-alveolar

Outgrowths During Pregnancy

In order to examine the role of basal cells in the development of alveolar outgrowths

we placed adult virgin female bitransgenic K5-rtTA/TGFP mice on Dox prior to pregnancy.

These mice were then paired with breeder males and examined daily for vaginal plugs. At

plug, mice were subjected to Dox withdrawal and sacrificed at d10.5 of pregnancy (Figure

6.3). Following this 10-day chase period, animals were sacrificed and mammary tissue was

harvested. One inguinal mammary gland was prepared for whole mount imaging and the

Page 143: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

126

remaining mammary tissue was digested to generate single cell suspensions for flow

cytometric analysis of GFP signal intensity. In addition we prepared samples from

bitransgenic d10.5 pregnant Dox-naïve and on Dox for 3 days prior to necropsy to provide

control comparisons.

Mimicking the strategy followed in the preceding section, we imaged mammary

gland whole mounts obtained from Dox-naïve, Pulse labeled, and pulse-chased bitransgenic

mice by wide field microscopy. All whole mounts obtained from d10.5 mice demonstrated

lobulo-alveolar outgrowths typical of mid-pregnancy as expected. In Dox untreated animals

no GFP fluorescence was detectable. In animals receiving doxycycline, widespread labeling

of the ductal tree was visible with green fluorescence also detectable in the developing

alveoli (Figure 6.7). Mammary whole-mounts that had been pulse labeled and then subjected

to Dox-withdrawal had bright GFP labeling within the ducts with dimmer GFP signal

apparent in the surrounding outgrowths induced by the hormones of pregnancy. Further

analysis of mammary epithelial cells isolated from experimental animals in this group by

flow cytometry confirmed the impressions gathered from microscopic examinations (Figure

6.7). Additionally, flow cytometric analysis of GFP signal intensity of these mammary

epithelial cells showed a range of GFP brightness suggestive of cellular division mediated

dilution of H2B-eGFP signal.

In order to confirm the presence of GFP label-retaining cells in lobuloalveolar

outgrowths, whole-mounts were frozen sectioned and Hoechst counterstained and imaged by

Confocal microscopy (Figure 6.8). Sections studied in this manner reaffirmed the

impressions obtained during the examination of mammary gland whole mounts with Dox-

naïve animals lacking detectable GFP signal while pulse labeled animals demonstrated

Page 144: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

127

widespread bright GFP labeled nuclei with the typical elongated-shape nuclei. Pulse-chased

samples showed bright GFP label within a portion of the cells lining the outer surfaces of the

mammary ducts and contained dimmer GFP positive cells within the alveolar outgrowths.

Our preliminary conclusion based on these images is that the GFP label-retaining cells are

localized along the outer layer of the lobuloalveolar outgrowths but without further

characterization of the location we cannot confirm this supposition. In stark contrast to the

current model of alveologenesis, which does not define a contribution of basal cells to the

alveolar structures, we demonstrate a clear contribution of basal-derived cells within the

lobulo-alveolar outgrowths.

6.3 Contributions of Mammary Epithelial Cell Sub-types to Lobulo-alveolar Outgrowths

Induced by Hormone Stimulation

With evidence in hand that both luminal and basal MECs contribute to lobuloalveolar

outgrowths during pregnancy, we next traced MEC lineages when lobuloalveolar

development was triggered using exogenous hormones. This strategy allowed precise timing

the onset of alveologenesis and circumvented pregnancy losses encountered during timed

pregnancy experiments (see Methods, Chapter 2). Furthermore, this hormone-induction

strategy sets the stage for dissecting the contribution of individual hormones to MEC

proliferation during lobuloalveolar development. Similar to the studies involving pregnant

animals we utilized adult virgin female bitransgenic (MMTV-rtTA/TGFP or K5-rtTA/TGFP)

animals to induce compartment specific labeling of the mammary gland prior to

administration of estrogen and progesterone (E+P) pellets. Subcutaneous insertion of time-

release hormone pellets occurred simultaneous to the excisional biopsy of a single inguinal

mammary gland. Implantation of E+P pellets results in lobulo-alveolar differentiation within

Page 145: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

128

the mammary gland [189, 198]. In order to capture label-retaining cells within this

experimental strategy, mice were harvested 7 days following pellet implantation. Though

this period of off-Dox washout was shorter than previously preformed ample lobulo-alveolar

outgrowths were present.

6.3.1 Tracing the Contributions of Luminal Mammary Epithelial Cells to Lobulo-alveolar

Outgrowths During Hormone Induced Proliferation and Differentiation

In order to verify the patterns of labeling and label retention seen in pregnant mice

following luminal compartment restricted labeling, we induced MMTV-rtTA/TGFP animals

with Dox for 4 days. Animals were then subjected to excisional biopsy of a single inguinal

mammary gland and implantation of E+P pellets at the site of mammary gland biopsy, and

removed from Dox post-operatively. Pulse-only controls were placed on Dox at the time of

pellet implant and followed for 7 days prior to necropsy.

In order to confirm labeling of the mammary gland during Pulse, whole mounted

mammary glands taken at the time of biopsy were examined by wide-field fluorescence

microscopy (Figure 6.9). Pulse labeled biopsy samples taken at the time of E+P pellet

implant were brightly labeled as well but lacked the hallmarks of lobuloalveolar development

as was expected as they were not hormonally induced. Though lacking exogenous hormonal

stimulation, mammary cells were still subject to the normal hormonal cycles associated with

estrus in the mouse and the resultant transient proliferation and minimal differentiation and

therefore there was variability in the appearance of the ductal tree even during Dox pulse

prior to hormone implant.

Mammary gland whole mounts imaged by wide field fluorescence microscopy

demonstrated widespread lobulo-alveolar outgrowths that were brightly labeled in mice on

Page 146: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

129

Dox concurrent with E+P induction (Figure 6.9). In animals removed from Dox at the time

of pellet implantation widespread alveolar development is apparent in fluorescent imaged

mammary gland whole mounts with GFP signal visible both within the ducts and lobulo-

alveolar outgrowths. Flow cytometric assessment of GFP signal intensity in single cells

isolated from mammary tissue confirms both the large proportion of bright GFP labeled cells

during Dox pulse/hormone inductions but also the presence of label retaining cells following

off-Dox chase in hormone pellet implanted animals (Figure 6.9). The GFP single parameter

plots of pulse-chased MECs particularly demonstrates the saw-tooth pattern indicative of

successive rounds of cell division resulting in dilution of H2B-eGFP signal in the chromatin

among daughter cells. Further investigation into the localization of GFP label retaining cells

by Confocal imaging shows the widespread maintenance of labeling within the luminal layer

of the ducts and shows many GFP positive nuclei lining the inner surfaces of the developing

alveoli (Figure 6.10). Though exhaustive characterization of the localization of GFP label

retaining cells in these sections has not been preformed it appears that cells of the luminal

layer contribute primarily if not exclusively to the luminal layer of the lobulo-alveolar

outgrowths.

6.3.2 Tracing the Contributions of Basal Mammary Epithelial Cells to Lobulo-alveolar

Outgrowths During Hormone Induced Proliferation and Differentiation

To confirm the lineage tracing results of basal cells seen in pregnant animals we pulse

and pulse-chase labeled bitransgenic (K5-rtTA/TGFP) adult virgin females with E+P

hormone pellet implants. Mammary gland biopsies were obtained following pulse, at the

time of hormone pellet implant. Whole mount images of pulse labeled bitransgenic

mammary glands following hormone pellet induction show widespread labeling throughout

Page 147: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

130

the ducts with scattered labeling within the alveolar outgrowths, as expected in the alveoli

due to the net like arrangement of myoepithelial cells surrounding the expanding outgrowths

(Figure 6.11). In hormone induced pulse-chased mammary glands, GFP positive nuclei are

visible throughout the ductal structures. Flow cytometric plots depicting GFP signal

intensity of single cells isolated from pulse and pulse-chased hormone implanted animals

show the expected patterns of GFP signal intensity (Figure 6.11). Pulse labeled animals

show a large proportion of cells with bright H2B-eGFP fluorescence while pulse-chased

MECs have a range of GFP signal intensity but are generally dimmer than that seen in on

Dox animals. Confocal projections collected from frozen sections reaffirm the presence of

GFP label-retaining cells within the ducts and in lobulo-alveolar outgrowths (Figure 6.12).

Cells within the outlying growths are generally dimmer than those seen in the central ducts

reconfirming their derivation from initially labeled cells by cell divisions. The GFP positive

cells seen within the lobulo-alveolar outgrowths appear to be restricted to the outer surfaces

but exhaustive characterization would be required to exclude the presence of basal derived

cells within the luminal cell layer within alveoli. The repeated identification of basal-derived

MECs within hormone induced lobulo-alveolar outgrowths supports the assertion that both

luminal and basal MECs play a role in the development of pregnancy induced growth and

differentiation.

6.4 Persistence of Labeled Cells Through Post-Lactational Remodeling of the Mammary

Gland

The rapid proliferation and differentiation of MECs during pregnancy drives the

creation of secretory units specialized for lactation. However, following weaning, these

cellular structures are no longer needed and are deleted from the mammary gland via

Page 148: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

131

involution. During this post-lactational remodeling, the mammary gland sheds its extensive

lobulo-alveolar outgrowths, restoring the mammary ductal tree to a virgin-like state (Figure

6.3). The exact patterns of cellular remodeling that result in the restructuring of the ductal

tree have not been elucidated. Here, we adapt the inducible H2B-eGFP labeling strategy to

allow tracing of cells through involution. By inducing labeling during pregnancy and early

lactation and then searching for label retaining cells following an extended chase period

concurrent with involution we can begin to track cell fates through post-lactational

remodeling and perhaps provide insight into the modalities of parity-related breast cancer

protection.

6.4.1 Luminal Labeled Mammary Epithelial Cells Persist Through Involution

In order to assess the retention of luminal cells through involution, MMTV-

rtTA/TGFP bitransgenic timed-pregnant mice were subjected to timed pregnancies and

placed on Dox 4 days prior to parturition in order to induce widespread labeling of luminal

MECs. Following birth, pups were removed and dams were subjected to excisional

mammary gland biopsy. Post-operatively, mice were subjected to Dox withdrawal during a

period of involution. At the conclusion of this 4-week chase, mice were sacrificed and the

remaining mammary tissue harvested. Control mice followed an identical time-line but were

maintained Dox-naïve throughout the experiment (Figure 6.3).

Figure 6.13 demonstrates the Dox-dependence of labeling during lactation. Both

whole mount images obtained by conventional fluorescence microscopy and Confocal

imaging of frozen sections derived from pulse-labeled biopsy mammary gland samples

demonstrate widespread H2B-eGFP labeling throughout the ducts and alveoli. In the Dox-

naïve animal, while green fluorescence is visible in the whole mount, subsequent Confocal

Page 149: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

132

imaging fails to detect the presence of any GFP positive nuclei. The green color seen in the

whole mount image is likely auto-fluorescence due to milk as the non-specific illumination is

seen in other color channels.

Having confirmed Dox-dependent regulation of H2B-eGFP transgene expression

during lactation we sought to track the persistence of luminal MECs through the remodeling

of the mammary gland during lactation. Pulse-labeled mice were removed from Dox

following delivery of pups and allowed to proceed through involution for 4 weeks and then

sacrificed. Whole mount imaging of these samples was hampered by the greatly decreased

signal intensity of the remaining H2B-eGFP signal that persisted over the extended period of

washout combined with the brighter autofluorescence seen in involuted mammary glands

(Figure 6.14). Subsequent imaging of frozen sections obtained from those whole mounts

provided clear images of GFP positive label retaining cells within the inner layers of the

involuted ducts while none was seen in glands harvested from untreated animals. Flow

cytometric analysis of single cells isolated from Dox-untreated animals did not identify GFP

positive nuclei. However, analysis of MECs from pulse-chased labeled mammary glands did

identify a number of GFP positive cells within a wide range of signal intensity. Peaks

depicting the rounds of cellular proliferation that occurred during the period of chase also

suggest that maintained cells play a role in driving the development of the restructured

mammary epithelial tree.

6.4.2 Basal Labeled Mammary Epithelial Cells Persist Through Involution

In order to assess the retention of basal cells through involution, MMTV-rtTA/TGFP

bitransgenic timed-pregnant mice were subjected to timed pregnancies and placed on Dox 4

days prior to parturition in order to induce widespread labeling of basal MECs. Following

Page 150: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

133

birth, pups were removed and dams were subjected to excisional mammary gland biopsy.

Post-operatively, mice were subjected to Dox withdrawal during a period of involution. At

the conclusion of this 4-week chase, mice were sacrificed and the remaining mammary tissue

harvested. Control mice followed an identical time-line but were maintained Dox-naïve

throughout the experiment (Figure 6.3).

Figure 6.15 demonstrates the Dox-dependence of labeling during lactation. Both

whole mount images obtained by conventional fluorescence microscopy and Confocal

imaging of frozen sections derived from pulse-labeled biopsy mammary gland samples

demonstrate widespread H2B-eGFP labeling throughout the ducts and alveoli. Fluorescent

nuclei appear in a scattered pattern around the outside of alveolar structures as expected due

to the discontinuous myoepithelial layer already characterized in alveoli [10, 41, 199]. In the

Dox-naïve animal, while green fluorescence is visible in the whole mount, subsequent

Confocal imaging fails to detect the presence of any GFP positive nuclei. The green color

seen in the whole mount image is likely auto-fluorescence due to milk as the non-specific

illumination is seen in other color channels and does not have the concentrated nuclear

appearance of H2B-eGFP fluorescence.

Having confirmed Dox-dependent regulation of H2B-eGFP transgene expression

during lactation we sought to track the persistence of basal MECs through the remodeling of

the mammary gland during lactation. Pulse-labeled mice were removed from Dox following

delivery of pups and allowed to proceed through involution for 4 weeks and then sacrificed.

Repeating the situation seen in luminal-labeled glands, whole mount imaging of these

samples was hampered by the greatly decreased signal intensity of the remaining H2B-eGFP

signal that persisted over the extended period of washout combined with the brighter

Page 151: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

134

autofluorescence seen in involuted mammary glands (Figure 6.16). Subsequent imaging of

frozen sections obtained from those whole mounts provided clear images of GFP positive

label retaining cells within the involuted ducts while none was seen in glands harvested from

untreated animals. Flow cytometric analysis of single cells isolated from Dox-untreated

animals did not identify GFP positive nuclei however; analysis of MECs from labeled and

chased mammary glands did identify a number of GFP positive cells within a wide range of

signal intensity. Peaks depicting the rounds of cellular proliferation that occurred during the

period of chase also suggest that maintained cells play a role in driving the development of

the restructured mammary epithelial tree.

6.5 Discussion

By tracing the lineage of luminal and basal epithelial cells during pregnancy and

hormone-induced lobulo-alveolar differentiation we provide evidence supporting the role of

both luminal and basal type MECs in the development of alveolar outgrowths. Our findings

contrast with the existing model in which a progenitor cell of luminal origin gives rise to both

luminal and basal cells in the alveoli. This finding also helps provide a clarification in the

hierarchy of mammary epithelial cell (Figure 6.17).

The work begun on pregnant animals here also offers the possibility of study during

different phases of pregnancy and lactation to help identify the particular contributions of

each cell type to alveologenesis and lactation. The characterization of the specific cell

populations within each cell layer that give rise to the alveolar outgrowths and the regulation

of these progenitor populations would offer insight into the maintenance of a maintained

adult stem cell-like population in the mammary gland. Considering that the plasticity of

Page 152: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

135

MECs has been noted to play a role in the development of mammary cancers and the alveolar

progenitor populations may have the ability to proliferate and differentiate in a distinct

manner than the bulk cells of the mammary duct, the potential exists that alveolar progenitors

may be cells of origin for some breast cancers [89]. It is also possible that identifying the

precise capacities of luminal and basal epithelial cells following hormone stimulation may

shed light onto the capacities of normal and transformed MECs. By further elucidating the

precise hierarchy of MECs during development potential cell of origin effects of particular

mutations on ultimate tumor subtype may be better understood and more precise and

efficacious treatments for the particular lineage and mutation history can be developed [90].

Capitalizing on the use of exogenous hormonal stimulation to drive the formation of

lobulo-alveolar outgrowths in the mammary gland by providing a basis for the study of single

hormone effects may possibly elucidate the roles of individual hormones on the regulation of

proliferation or MEC subtypes during pregnancy. The ability to experimentally manipulate

the developmental state of the mammary gland and the capacity to induce labeling in viable

cells of both epithelial cell types offers a method to identify cell population subsets based on

additional distinct characteristics. This additional characterization may help further clarify

the precise signals required for the induction of alveolar progenitor populations.

This investigation also begins to shed light on the interesting phenomenon of parity-

related protection from breast cancer. While it has been recognized that a full-term

pregnancy early in life offers significant protection against breast cancer, no model has been

identified that explains the precise cause of that protection [106-110]. It has been suggested

that the overarching remodeling of the mammary ductal tree during involution results in the

elimination of the majority of cells present prior to involution and therefore any pre-

Page 153: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

136

neoplastic cells are lost [106, 107]. While the work presented here does not explain parity

related breast cancer, it provides an initial glimpse into the possible explanations by

eliminating complete reconstruction of the mammary gland as a method of removal of pre-

neoplastic cells. It is possible that involution processes somehow select particular cells for

elimination but a sizeable proportion of the cells composing the remodeled duct contained

H2B-eGFP suggesting that many cells are retained though post-lactational remodeling.

Interestingly, others have reported a cell type that is present in the mammary gland

following a round of pregnancy that is hormone responsive and preserved through involution.

They have termed these as “parity-induced” cells and they have been demonstrated to self-

renew and give rise to both ductal and alveolar outgrowths and carry an immunophenotype

consistent with the population of MECs containing the stem cell population [200, 201].

These cells have also been shown to be targets for MMTV-neu driven tumorigenesis in

parous mice [176]. Additionally, some of the more common strains of laboratory mouse are

frequently unable to nurse their first litter though subsequent litters are maintained without

issue. It is apparent that some changes in the composition of the mammary ductal tree occur

following pregnancy and perhaps altered H2B-eGFP labeling schemes would help tease out

the populations supporting these changes.

Page 154: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

137

Figure 6.1 A Model of Alveolar Progenitors.

A schematic depiction of the commonly held model of mammary alveologenesis. This

model proposes that during pregnancy a single cell (or group of cells) of luminal origin gives

rise to the complete lobulo-alveolar outgrowth consisting of both luminal and myoepithelial

cell layers.

Page 155: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

138

Figure 6.2 Parity-Related Protection from Breast Cancer.

A schematic depiction of the potential avenues of parity related protection from breast

cancer. Early first full term pregnancy offers protection from development of breast cancer in

women. The exact processes that provide this protection are unclear. It is possible that the

wholesale remodeling of the mammary ductal tree during post-lactational involution removes

pre-neoplastic cells and thereby limits the presence of transformed cells within the mammary

epithelium.

Page 156: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

139

Figure 6.3 Experimental Timelines For the Investigation of Cell Fates During

Pregnancy.

The images at the top are representative whole mount images depicting whole mount

mammary glands stained with Carmine stain and imaged by brightfield microscopy. Three

independent time courses were followed to obtain the data depicted in this chapter and to

allow lineage tracing in the mammary gland during pregnancy. Experimental protocols were

as follows: The Mid-pregnancy group was placed on Dox diet simultaneous to their breeding

and removed from Dox when pregnancy was initiated by identification of a plug. They were

maintained in Dox-withdrawal for 10 days until mid-pregnancy d10.5 and subjected to

necropsy and mammary gland harvest. In the hormone induction group, adult virgin females

were treated with Dox for 4 days. These animals were then subjected to biopsy of a single

mammary gland and subcutaneous implanting of time-released estrogen and progesterone

pellets and removed from Dox. They were maintained off Dox for 7 days prior to necropsy

and harvesting of mammary tissue. The final experimental group consisted of adult females

bred off Dox. Timed pregnant animals were placed on Dox 4 days prior to parturition. Both

Dox and pups were removed following delivery and the animals were maintained off Dox for

4 weeks in order to allow for involution to occur.

Page 157: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

140

Page 158: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

141

Figure 6.4 Schematic Depictions of Potential Outcomes of Lineage Tracing During

Alveologenesis.

This diagram depicts the possible outcomes for label tracking through lobulo-alveolar

development. The current models of lobulo-alveolar development propose that a single cell

population of luminal origin is responsible for both the luminal and myoepithelial layers of

the outgrowths with no contribution from the original basal layer. Inducible H2B-eGFP

labeling prior to pregnancy followed by chase allows the short-term tracing of lineage during

pregnancy induced growth and differentiation. Indicated in orange are the currently held

models for the contributions of luminal and basal cells to pregnancy induced lobulo-alveolar

development. Briefly, potential outcomes include; cells from each cell layer may contribute

exclusively to their layer of origin, to both cell layers of alveoli, or may not contribute to the

differentiated structure.

Page 159: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

142

Figure 6.5 H2B-eGFP Labeling of Luminal Cells During Pregnancy in MMTV-rtTA/TGFP

Mice.

Representative single parameter flow cytometry plots for GFP fluorescence obtained by

analysis of single mammary epithelial cells isolated from adult d10.5 pregnant females with

or without Dox treatment as indicated are displayed in the first column. The second column

is composed of wide field fluorescence images (Mag. 10X) of whole-mount mammary

glands from the same mice. Pulse animals (n=2) were placed on Dox from d7.5 to d10.5 of

pregnancy. Dox-naive animals (n=2) were maintained without Dox prior to and during the

experiment. Pulse-chase animals (n=5) were bred on Dox and removed from Dox when a

plug was detected.

Page 160: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

143

Figure 6.6 Luminal Cells Contribute to Lobulo-alveolar Outgrowths During Pregnancy.

A.H2B-eGFP Fluorescence in Mammary Gland Sections. Images depict whole-mounted

mammary glands that were frozen-sectioned, counterstained with Hoechst (blue) and imaged

by Confocal microscopy (MAX Projections; Mag. 20X). Dox-treated pregnant mammary

glands were brightly labeled with GFP fluorescence as expected. No GFP fluorescence was

detectable in mammary gland sections obtained from Dox naïve animals. GFP labeled cells

were present in both the ducts and outgrowths in pulse-chased mammary glands. B. H2B-

eGFP Signal is Detectable in Lobulo-alveolar Outgrowths. The panels present a more

detailed view of an imaged area of a pulse-chased mammary gland (Confocal Microscopy,

MAX Projections; Mag. 20X). Both GFP/Hoechst (left) and GFP channel alone (right)

images demonstrate the presence of GFP labeled cells in the luminal layer of the duct and in

the lobulo-alveolar outgrowth.

Page 161: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

144

Figure 6.7 H2B-eGFP Labeling of Basal Cells During Pregnancy in K5-rtTA/TGFP Mice.

Single parameter plots presenting representative flow cytometry results for GFP fluorescence

obtained by analysis of single mammary epithelial cells isolated from adult K5-rtTA/TGFP

d10.5 pregnant females with or without Dox treatment as indicated are displayed in the first

column. The second column is composed of wide field fluorescence images of whole-mount

mammary glands from the same mice. Pulse animals (n=2) were placed on Dox from d7.5 to

d10.5 of pregnancy. Dox-naive animals (n=2) were maintained without Dox both prior to and

for the duration of the experiment. Pulse-chase animals (n=4) were bred on Dox and

removed from Dox when a plug was detected.

Page 162: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

145

Figure 6.8 Basal Cells Contribute to Lobulo-alveolar Outgrowths During Pregnancy.

A.H2B-eGFP Fluorescence in Mammary Gland Sections. Images depict whole-mounted

mammary glands from K5-rtTA/TGFP mice that were frozen-sectioned, counterstained with

Hoechst (blue) and imaged by Confocal microscopy (MAX Projections; Mag. 20X). Pulse-

labeled pregnant mammary glands were brightly labeled with GFP fluorescence as expected.

No GFP fluorescence was detectable in mammary gland sections obtained from Dox naïve

animals. GFP labeled cells were present in both the ducts and outgrowths in pulse-chase

mammary glands. B. H2B-eGFP Labeling in the Lobulo-alveolar Outgrowths. The panels in

B present a more detailed view of an imaged area of a pulse-chased mammary gland

(Confocal Microscopy, MAX Projections; Mag. 20X). Both GFP/Hoechst (left) and GFP

channel alone (right) images demonstrate the presence of GFP labeled cells in the basal layer

of the duct and in the lobulo-alveolar outgrowth.

Page 163: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

146

Figure 6.9 H2B-eGFP Labeling of Luminal Cells During Hormone Induced Lobulo-alveolar

Development.

Representative single parameter flow cytometry plots for GFP fluorescence obtained by

analysis of single mammary epithelial cells isolated from adult, MMTV-rtTA/TGFP,

hormone pellet implanted females during or following Dox treatment as indicated are

displayed in the first column. The second column is composed of wide field fluorescence

images (Mag. 10X) of whole-mount mammary glands from the same mice. Pulse +pellet

animals (n=1) were implanted with estrogen and progesterone pellets, and placed on Dox for

7 days. Pulse-chase animals (n=5) were placed on Dox for 4 days then removed from Dox

and implanted with estrogen and progesterone pellets for 7 days prior to necropsy.

Page 164: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

147

Figure 6.10 Luminal Cells Contribute to Lobulo-alveolar Outgrowths During Hormone

Induced Lobulo-alveolar Development.

A. Microscopic Imaging of GFP signal in Alveolar Outgrowths. These panels depict whole-

mounted mammary glands that were frozen-sectioned, counterstained with Hoechst (blue)

and imaged by Confocal microscopy (MAX Projections; Mag. 20X). Mammary glands

collected from Dox-treated and during Dox-withdrawal demonstrated extensive lobulo-

alveolar development following hormone pellet implant. The luminal layer was brightly

labeled during on Dox pulse as expected. Bright GFP nuclei were seen in both the inner

lining of the ducts and outgrowths in pulse-chased mammary glands. B. GFP Fluorescence in

alveolar outgrowths. The images present a more detailed view of a representative imaged

area of a pulse-chase mammary gland. Both GFP/Hoechst and GFP alone images

demonstrate the presence of GFP labeled cells in the luminal layer of the duct and in the

lobulo-alveolar outgrowth (MAX Projections; Mag. 20X).

Page 165: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

148

Figure 6.11 H2B-eGFP Labeling of Basal Cells During Hormone Induced Lobulo-

alveolar Development.

Samples from adult K5-rtTA/TGFP, hormone pellet implanted females with or without Dox

treatment were collected as indicated to allow analysis of GFP fluorescent labeling. Pulse

animals (n=2) were implanted with estrogen and progesterone pellets, and placed on Dox for

7 days. Pulse/chase animals (n=4) were placed on Dox for 4 days then removed from Dox

and implanted with estrogen and progesterone pellets for 7 days prior to necropsy.

Representative single parameter flow cytometry plots for GFP fluorescence obtained by

analysis of single mammary epithelial cells isolated as indicated are displayed in the first

column. The second column is composed of wide field fluorescence images (Mag. 10X) of

whole-mount mammary glands from the same mice.

Page 166: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

149

Figure 6.12 Basal Cells Contribute to Lobulo-alveolar Outgrowths During Hormone

Induced Lobulo-alveolar Development

Mammary glands collected from mice during Dox-treatment and following Dox-withdrawal

demonstrated extensive lobulo-alveolar development following hormone pellet implant in

K5-rtTA/TGFP females. The basal layer was brightly labeled during pulse as expected.

Bright GFP nuclei were seen in both the inner lining of the ducts and outgrowths in pulse-

chased mammary glands. A. Microscopic Imaging of GFP signal in Alveolar Outgrowths.

These panels depict whole-mounted mammary glands that were frozen-sectioned,

counterstained with Hoechst (blue) and imaged by Confocal microscopy (MAX Projections;

Mag. 20X). B. GFP Fluorescence in Alveolar outgrowths. The images present a more

detailed view of a representative imaged area of a pulse-chase mammary gland. Both

GFP/Hoechst and GFP alone images demonstrate the presence of GFP labeled cells in the

basal layer of the duct and in the lobulo-alveolar outgrowths (MAX Projections; Mag. 20X).

Page 167: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

150

Figure 6.13 Dox Regulated H2B-eGFP Labeling of Luminal Mammary Epithelial Cells

During Lactation.

Bitransgenic MMTV-rtTA/TGFP females were placed on Dox for 4 days while Never Dox

samples were collected from Dox-naïve animals. Both Pulse and Never Dox samples

displayed lobulo-alveolar outgrowths as expected. Images in the top row were taken from

the mammary glands by biopsy the day of parturition and imaged as whole mounts by

fluorescent wide-field microscopy (Mag. 10X). Following sectioning, after Hoechst

counterstain, images were collected by Confocal microscopy (MAX Projections; Mag. 20X).

Widespread nuclear H2B-eGFP labeling is apparent in the Pulse labeled mammary glands

but is not observed in Dox-naïve samples. The green signal visualized in Never Dox whole-

mounts is autofluorescent and is not detected by Confocal imaging.

Page 168: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

151

Page 169: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

152

Figure 6.14 Post-lactational Remodeling of the Mammary Gland Retains Luminal

Labeled Mammary Epithelial Cells

Panels present fluorescent microscopic and Confocal microscopic imaging of whole mounts

and sections, and the single parameter depiction of GFP signal intensity obtained by flow

cytometry of single cells collected from bitransgenic MMTV-rtTA/TGFP mice that were

placed on Dox 4 days prior to parturition and removed from Dox after delivery of pups

(Pulse-Chase) or on an identical timeline from Dox-naïve animals. The mammary glands

were collected from these mice following 4 weeks of involution and analyzed by

conventional fluorescent microscopy of whole mounts (Mag. 10X). Following sectioning and

Hoechst staining, images were obtained by Confocal microscopy (MAX Projections; Mag.

20X). Single cells prepared from mammary tissue at the time of necropsy were also analyzed

by flow cytometry for GFP signal intensity. H2B-eGFP signal is visible in the remodeled

ducts of pulse labeled-chased mammary glands by both conventional and Confocal

fluorescent microscopy but is not seen in Dox-naïve samples. Involuted mammary ducts

contained globular areas of autofluorescence that did not correspond to H2B-eGFP signal.

Flow cytometric assay of GFP signal intensity replicates the observation seen in the imaging

studies. Interestingly, peaks representative of rounds of cell proliferation are visible in the

single parameter plot of Pulse-Chase group MEC populations.

Page 170: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

153

Figure 6.15 Dox Regulated H2B-eGFP Labeling of Basal Mammary Epithelial Cells

During Lactation

Panels depict photomicrographs of whole mount and sectioned mammary glands. Images in

the top row were taken from the mammary glands of transgenic K5-rtTA/TGFP mice by

biopsy the day of parturition and imaged as whole mounts by fluorescent wide-field

microscopy (Mag. 10X). Following sectioning after Hoechst counterstain, images were

collected by Confocal microscopy (MAX Projections; Mag. 20X). Samples were obtained

from mice on Dox for 4 days (Pulse) and from Dox-naïve animals. Both Pulse and Dox-naive

samples displayed lobulo-alveolar outgrowths as expected. Widespread nuclear H2B-eGFP

labeling is apparent in the Pulse labeled mammary glands but is not observed in Dox-naïve

samples. The green signal visualized in Dox-naive whole-mounts is autofluorescent and is

not detected by Confocal imaging.

Page 171: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

154

Page 172: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

155

Figure 6.16 Post-lactational Remodeling of the Mammary Gland Retains Basal Labeled

Mammary Epithelial Cells.

Panels present fluorescent microscopic and Confocal microscopic imaging of whole mounts

and sections, and the single parameter depiction of GFP signal intensity obtained by flow

cytometry of single cells collected from bitransgenic K5-rtTA/TGFP mice that were placed

on Dox 4 days prior to parturition and removed from Dox after delivery of pups (Pulse-

Chase) or on an identical timeline from Dox-naïve animals. The mammary glands were

collected from these mice following 4 weeks of involution and analyzed by conventional

fluorescent microscopy of whole mounts (Mag. 10X).Following sectioning and Hoechst

staining, images were obtained by Confocal microscopy (MAX Projections; Mag. 20X).

Single cells prepared from mammary tissue at the time of necropsy were also analyzed by

flow cytometry for GFP signal intensity. H2B-eGFP signal is visible in the remodeled ducts

of pulse-chase mammary glands by both conventional and Confocal fluorescent microscopy

but is not seen in Dox-naïve samples. Involuted mammary ducts contained globular areas of

autofluorescence that did not correspond to H2B-eGFP signal. Flow cytometric assay of GFP

signal intensity replicates the observation seen in the imaging studies. Interestingly, peaks

representative of rounds of cell proliferation are visible in the single parameter plot of Pulse-

Chase group MEC populations.

Page 173: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

156

Page 174: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

157

Figure 6.17 A Revised Model of Alveolar Progenitor Cells.

This diagram depicts our proposed model of lobuloalveolar development based on the

outcomes of our label tracking experiments. We identified label retaining cells originating

from both cell layers in the lobulo-alveolar outgrowths stimulated by pregnancy and hormone

implants. We additionally demonstrated that the cells contained within the outgrowths have

less retained H2B-eGFP fluorescence and showed flow cytometry analyses demonstrating

H2B-eGFP signal consistent with proliferation dependent washout. In light of these findings,

and in contrast to the commonly held model (boxed in orange), we propose that both luminal

and basal cell layers contribute to the development of alveoli (shaded in yellow).

Furthermore, these contributions to alveologenesis are proliferation dependent and do not

represent migration of ductal cells. We suggest that cell contributions during differentiation

may be confined to the cell layer of origin, but cannot yet confirm this model (shaded in

peach).

Page 175: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

158

Chapter 7 Final Discussion

While much is understood about the mammary gland on a structural level, the specific

cellular contributions to the developing mammary gland have not yet been fully elucidated.

By pairing compartment-specific transactivators in the mammary gland with the recently

developed H2B-eGFP transgene we developed a method that tracks proliferation history and

also lineage path for cells of the mammary epithelial tree. We additionally extended this

method to examine the proliferation histories of the cells comprising mammary tumors and

regressed MRD lesions.

7.1 Labeling the Luminal and Basal MEC Compartments

In Chapter 3, a method allowing the temporal and tissue specific regulation of

transgene expression was described. Expression of the TGFP reporter transgene results in

the labeling of chromatin within the luminal or basal cell layers of the mammary epithelial

tree induced by MMTV-driven or Keratin5 –driven transactivators. This fluorescent labeling

may subsequently be diluted by cell division allowing both the tracking of lineage and

monitoring of proliferation history. By examining reporters for both luminal and basal

specific compartments, this method allows the study of the lineage commitment in the

development of mammary ductal structures.

While MMTV is widely expressed within the luminal epithelium, this expression is

not uniform and is altered by hormone fluctuations. During pregnancy and lactation,

MMTV-driven expression is greatly increased in both intensity and proportion of cells

expressing MMTV. As an alternative method for driving transgene expression, Whey Acidic

Protein (WAP) has been used to induce transgene expression in the luminal epithelium of

adult parous mice [202]. This promoter is not expressed prior to pregnancy as it is a

Page 176: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

159

component of the secreted milk produced during lactation. Expression driven by the WAP

promoter could be useful for the study of luminal alveolar-derived epithelial cells in the adult

parous animal and during pregnancy. As this promoter is then transcribed in the cells that

have participated in lactation, the ability to monitor the fate of these cells through lactation

and involution and throughout subsequent pregnancies provides an interesting avenue to

explore the altered capacity of MECs following hormonally driven proliferation and

differentiation.

We have described the presence of slower-cycling cells in the mammary ductal tree

during development but did not identify a typical location along the branching structures that

harbors these cells. As microenvironment plays a critical role in the development and

maintenance of the mammary gland it is likely that particular subsets of cells, either slower

or rapidly cycling are maintained in a particular niche. Additional pulse-chase conditions

during puberty and adult stages may help identify locations that harbor LRCs. Some adult

stem cell populations are maintained in a quiescent state; notably the stem cells of the skin

are maintained in a quiescent state in the skin and in the corneal limbus [182, 203]. As there

does not appear to be a biological requirement for quiescence in stem cell populations the

identification of cycling stem cell populations has also occurred. The niche of the stem cell

population within the gut and the proliferation patterns of these cells has been well described

[204]. The identification of particular cycling patterns within the candidate mammary stem

cell population may also offer a pulse-chase timeline ideal for the isolation of stem cells

based on their retention or washout of H2B-eGFP labeling under specific developmental

phases.

Page 177: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

160

Niche has been demonstrated to play a regulatory role in the maintenance and growth

patterns of normal stem cells and in tumorigenic stem cells in some tissues including the skin

and blood and brain tumors [160, 182, 205]. Identification of niche locations has facilitated

the characterization of the signaling pathways and cell-extracellular matrix interactions that

regulate the division and maintenance of these cells [206, 207]. The identification of LRCs

within mammary ducts raises the question of whether these cells are stem or progenitor-like

cells; and localization of the niche for these cells will permit investigation into the

maintenance of this compartment. Fluorescent labeling studies offer the benefit of allowing

isolation of viable cells of a specific population and thus would allow isolation of a single

cell of interest based not only on H2B-eGFP labeling, or label retention, but also on location

within a tissue. H2B-eGFP label retention of cells may help identify a stem cell niche within

the mammary gland. Following identification of a structural localization of stem cells within

the mammary gland, the biochemical properties of the niche could be explored to define the

role of the niche in the maintenance of MaSCs. Cells isolated from potential niches, if found,

could subsequently be studied in culture or via implant procedures to assay for stem cell-like

capacity. While identification of cells based on incorporation of H2B-eGFP label provides a

clear manner for the selection of a desired population, non-labeled cells may also be selected

based on the absence of GFP label under desired conditions.

7.2 Using H2B-eGFP Labeling to Characterize Mammary Tumors

In Chapter 4 we adapted the inducible compartment specific labeling of MECs to

label mammary tumors. We demonstrated inducible H2B-eGFP labeling in both the luminal

and basal layers of these tumors and demonstrated proliferation dependent washout in a

subset of these tumors. The extensive variability seen in MMTV-wnt1 initiated tumors

Page 178: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

161

proved a challenge in the identification of cell populations with distinct proliferation patterns

in constitutive tumors. Though clouding the conclusions in these experiments, the variability

inherent in MMTV-Wnt1 initiated mammary tumors reflects the wide range of mammary

cancers seen in humans. Characterization of particular tumor-type variants would allow

more clear interpretation of labeling and label retaining cell populations in growing tumors.

Specifically, the potential identification of genetic causes for the loss of H2B-eGFP signal in

Dox-induced tumors would provide one manner to group transgenic tumors for study of

labeling and label retention. Additionally, the characterization of accumulated mutation

patterns accompanying MMTV-Wnt1 initiated mammary tumors may provide insight into

the variability seen in inducible H2B-eGFP labeling of those tumors.

A candidate breast cancer stem cell has been isolated from human tumors based on

animal transplant studies and 3D culture, however the in situ capabilities of this cell type

have not been defined [88]. Cancer stem cells may have different proliferative behavior than

bulk tumor cells and the isolation of distinct tumor cell populations based on proliferation

history offers an alternative method for the isolation of putative stem cells. A number of

tumors driven by constitutive Wnt1 expression did contain populations with variable

proliferation histories. Isolating these cells based on the H2B-eGFP signal intensity by

FACS, possibly paired with cell-surface markers, would allow for the characterization of

individual cells (or cell populations) with a distinct biological feature. LRCs isolated from

mammary tumors potentially harbor CSC-like capacity and may also be derived from the

normal stem cell population of the mammary epithelium.

The role of microenvironment plays a crucial part in the maintenance of adult stem

cells and there is the possibility that when implanted into normal stroma that even malignant

Page 179: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

162

cells may behave in a “normal” fashion. In particular, the mammary gland stroma and non-

tumorigenic epithelial cells have been demonstrated capable of reprogramming embyronal

cancer cells [208]. In these experiments, human embryonal cancer cells mixed with normal

MECs were implanted into the cleared mammary fat pad of athymic mice and did not

develop into new tumor lesions but instead differentiated to produce MECs. Building upon

this observation, single cells of the desired H2B-eGFP labeling phenotype from mammary

tumors could be implanted into cleared fat pads in order to examine the effects of

extracellular signaling and architecture on the regulation of proliferation and differentiation.

The implanted mammary glands can be monitored for tumor development and following a

prescribed time period animals could be induced with Dox, sacrificed, and the contributions

of the transgenic H2B-eGFP expressing cells to the developed mammary gland assessed.

This method would require isolation of MECs from mammary tumors by fluorescent assisted

cell sorting and recovery of viable cells. Furthermore, these cells would have to be

implanted into cleared fat pads via injection in suspension with wild-type MECs. While

some successes with FACS and injection implant have been described by others, in our hands

these techniques remain unpredictable. Injection of mixed cell populations would produce

mixed-lineage outgrowths and in each instance, tumor derived cells may or may not

contribute in a significant manner to the ductal outgrowths.

By isolating and implanting different cell populations from luminal and basal H2B-

eGFP labeled MMTV-wnt1 tumors the capabilities of all distinct populations within the

tumor could be elucidated. This characterization would help clarify the potential for a CSC

population within Wnt-initiated mammary cancers. The variability of these wnt-initiated

cancers could possible preclude the isolation of a single “cell of origin” for all the tumors

Page 180: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

163

studied, however, it is likely that tumors could be grouped by common genetic alterations

and similar results in the capacity of labeled or label retaining cells could be identified.

Cytokinetic resistance, the resistance to chemotherapeutic treatment by non-

proliferative cells, has complicated the development and use of many chemotherapeutics. By

isolating and characterizing the capacity of the LRCs in mammary tumors, the potential

exists to develop alternate therapies that would eliminate these treatment resistant

populations such as immunotherapies targeted to the LRCs in particular. If alternatively,

characterization of slower-cycling cells indicates that they are in fact non-tumorigenic, the

lack of response to chemotherapeutic treatments would be insignificant. Potentially, these

cells which compose a substantial portion of some tumors, may remain following treatment

that successfully removes tumor-initiating cells, but as they lack that capacity they do not

pose a high risk for either tumor relapse or for the seeding of metastases. By assaying the

response of labeled subpopulations from both the luminal and basal compartments of well

characterized mammary tumors to a range of typical cancer therapeutics, a more complete

picture of tumor response may be achieved. While the single set of clonal tumors assayed in

Chapter 4 did not display a difference in response of labeled and unlabeled luminal MECs to

a single treatment with Adriamycin, the potential exists that different time frames, dose, and

duration of treatment, as well as alternate treatments such as radiation may offer different

results. Additionally, the extensive variability seen in the mammary tumors in this study

suggest that though the example studied in Chapter 4 did not demonstrate a response to

treatment, it is by no means representative of all mammary tumors or even all Wnt-initiated

mammary tumors.

Page 181: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

164

7.3 LRCs in MRD and Maintenance of MRD

Again, though extensive variability complicated the analysis of the presence of LRCs

in MRD lesions in Chapter 5, several tumor sets were identified that provided evidence in

support of a model of cellular dormancy for the maintenance of tumor dormancy in the

mammary gland. As discussed previously, in several instances there was not extensive label

retention in the cells of the MRD when compared to the surrounding normal ducts (regressed

hyperplasia) but nearly all the MRD lesions had much higher label retention than that seen in

constitutive MMTV-Wnt1 tumors, particularly when considered in light of the substantially

longer periods of Dox-withdrawal they were subjected to.

Metastasis of cancer is a primary cause of mortality in cancer patients and an

unexamined feature of the MMTV-rtTA/TWNT/TGFP model of mammary tumors and

labeling is the identification of distant metastases by fluorescence. Previous work has

identified multiple instances of metastases, primarily to the lung and liver, in MMTV-

rtTA/TWNT mice. Though lung and liver were examined at necropsy in the tri-transgenic

tumor-bearing mice utilized in this study, no macrometastases were identified. This may be

due to the comparatively small sample size of MMTV-rtTA/TWNT/TGFP mice studied to

date in comparison to the long history of MMTV-rtTA/TWNT mice or to some unexpected

effect of the TGFP transgene on metastatic potential of the mammary tumor cells.

Disseminated tumor cells are dispersed throughout the body; however the molecular

characteristics that identify the subset of cells capable of giving rise to macrometastases at

distant sites are unknown. If metastatic lesions or disseminated fluorescent cells, can be

identified in H2B-eGFP labeled and chased animals they provide an opportunity to further

characterize the potential of a specific population of cells within the growing mammary

Page 182: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

165

tumor and of the residual disease maintained even following abrogation of oncogenic

stimulation. By examining metastases prone tissues of tumor bearing mice by fluorescence

microscopy, we offer a method to identify micrometastatic lesions in situ. The potential

exists that distant metastases are maintained in a dormant state in either identical or

dissimilar manners to those occurring within the mammary gland and by assaying for label

retention in these metastases not only the lineage of the cells composing the lesion but also

the mechanism of maintenance may be analyzed by using compartment specific transgene

expression and examining H2B-eGFP label retention when removed from Dox stimulation.

7.4 Lineage Restriction and Cell of Origin in Alveologenesis

In Chapter 6, we describe our investigation of the lineage specific contributions to the

lobuloalveolar outgrowths of pregnancy utilizing the compartment-specific induction of

H2B-eGFP expression. While this work demonstrates the contribution of both luminal and

basal cells to the lobulo-alveolar outgrowths developed during pregnancy, in contrast to the

common model suggesting a luminal cell of origin for the entire alveolar structure, it does not

define the precise contribution of each cell type to the differentiated structure nor the

pathways or signaling mechanisms required to modulate the this proscribed growth

patterning.

Common progenitor cells and dedifferentiation of lineage committed cells have both

been offered as potential sources of the development of new structures within the adult

animal. In limb amputation models in amphibians, dedifferentiation has been proposed to

occur in blastema, but may be somewhat restricted with tagged cells giving rise only to cells

of highly similar lineage and suggest the contributions of multiple populations of slightly

dedifferentiated or committed cells driven to repopulate the limb bud [209-211]. Though

Page 183: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

166

mice are not capable of the large scale regrowth of limbs as seen in amphibians, many tissues

including the mammary epithelium are capable of responding to stimuli such as wounding

and changes in the hormonal milieu by increasing proliferation rates and altering the

pathways of lineage commitment. During pregnancy, the increase in estrogen and

progesterone levels drives proliferation and formation of the alveolar structures of the

mammary gland in an analogous manner to the signals driving limb regeneration in many

amphibians. The common model for alveologenesis posits a cell of luminal origin that gives

rise to both cell layers within the developed outgrowth. As the luminal cell is proposed to be

further committed than the normal mammary stem cell and basal cell populations within the

mammary it seems contrary to predict it giving rise to cells of both the luminal and basal cell

layers. Identifying the cell-layer specific contributions to the differentiated lobulo-alveolar

outgrowths by H2B-eGFP based lineage tracing allows the characterization of the role of

dedifferentiation or transdifferentiation by identification of specific compartment labeled

cells in the other cell layer of the developed alveolar structures.

A first step to examining the possibility of cross-contributions to each cell layer

could be obtained by immunohistochemical characterization of the label-retaining cells

within the alveolar structures of mid-pregnant mice to verify the appearance of lineage

restricted contribution to the alveoli seen in Chapter 6. A variety of defined IHC markers for

both luminal and basal cell populations exist and have been clearly defined. The

complicating feature of this experiment derives from the relative instability of H2B-eGFP

when subjected to standard histological procedures such as paraffin embedding and the

difficulty in utilizing the standard cell compartment markers in frozen sections that preserve

H2B-eGFP signal. Attempts at staining paraffin embedded sections with primary antibodies

Page 184: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

167

for eGFP were partially successful but demonstrated a loss in sensitivity for the range of GFP

levels. Further optimization of the IHC procedures may allow for careful identification of

the cell-layer of LRCs in the pregnant mammary gland.

While the MMTV-rtTA/TGFP system provides a clever manner for tracing the cell-

layer of origin of differentiated mammary epithelial cells it does not provide as simple a

manner for the tracing of single-cell of origin. The lobulo-alveolar structures derive from

cells of both cell layers but the work described in Chapter 6 does not address whether all

cells within each cell population or if only a select subset contribute to alveologenesis. Cell-

culture studies might provide a method of elucidating the specific cell or cells of origin for

these structures. Specifically, isolated GFP-labeled luminal (or basal) cells could be

assembled into organoids with non-transgenic epithelial cells of both lineages and grown in

3D culture under hormone-stimulation to induce differentiation. Additional

subcategorization of candidate alveolar progenitor cells could be made on particular

characteristics, including hormone receptor status and these populations could be analyzed

for contribution to differentiated structures.

Similarly, in a manner analogous to that used to assess the role of dedifferentiation

and transdifferentiation in the regrowth of amphibian limbs, and building upon the implant

studies that identified a candidate MaSC, single cells of interest from MMTV-rtTA/TGFP (or

K5-rtTA/TGFP) could be isolated and implanted into the cleared fat pads of host animals

mixed with non-transgenic mammary epithelial cells from syngeneic donors. The host

animals could be rendered pregnant while receiving Dox and the contributions of the

transgenic cells to lobulo-alveolar outgrowths could be assessed by flow cytometry and

histological means. Several candidate populations could be examined in this manner to

Page 185: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

168

illuminate the specific contributions of each population to the terminally differentiated

mammary gland and the level lineage restriction of each population.

The identification of the cell population(s) that give rise to the alveoli may also

provide insight into the stem cell population responsible for maintenance and development of

the mammary gland as a whole. Though a putative stem cell population has been described,

the localization of this population has not yet been offered in the adult mammary gland

though it is presumed to be basal based on the similarity of the proffered immunophenotype

to that of myoepithelial cells. It is possible that alternate time frames for Dox pulse and Dox

withdrawal, and extensive examination of MECs isolated from virgin mammary glands

would identify a candidate stem cell population that retains H2B-eGFP fluorescence

indicating relative quiescence and allow localization of this population. The potential exists

that a similar or perhaps slightly more lineage committed cell population is responsible for

the development of the alveolar outgrowths of pregnancy.

7.5 Parity-related Protection from Breast cancer

The final sections of Chapter 6 offered a preliminary glimpse into a potential

mechanism for the breast cancer protection related to pregnancy. The widespread remodeling

of the mammary gland following cessation of lactation does not remove the entire population

of MECs present during pregnancy and lactation but instead pairs remodeling with copious

reductions in cell number. The identification of label-retaining cells following involution

verifies the survival of cells through this period of post-lactational remodeling.

While this work counters a model of wholesale elimination of pre-neoplastic cells as

the mechanism of parity protection from breast cancer it does not illuminate the effects on

pre-neoplastic cells at the hormonal or genetic level, nor does it eliminate the possibility that

Page 186: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

169

potentially tumorigenic cells are selectively destroyed during involution. The selective

removal of genetically altered cells could be assessed by the implantation of single cells

isolated from MRD lesions of MMTV-rtTA/TWNT/TGFP mice in conjunction with isolated

cells from wild-type female mice. The resulting mammary ductal structures should bear

mosaic ducts comprised of both normal and “pre-neoplastic” cells. Subsequent induction of

pregnancy, lactation, and involution followed by short term induction with Dox would induce

labeling in those genetically altered cells that survive involution. Microscopic and flow-

cytometric detection of GFP signal in these chimeric glands would provide an indication if

selection of tumorigenic cells for elimination during involution may be responsible for

parity-related protection from breast cancer.

7.6 Summary

The work presented in this thesis provides not only the description of a model

allowing the labeling of mammary epithelial cells in a compartment specific manner and

consequently allowing the tracking of lineage of MECs during development and

differentiation, it also provides a system that permits the tracking of cell proliferation in

normal and neoplastic mammary tissue. By building upon the strategies described in this

document, a greater understanding of the processes governing tumorigenesis and tumor

regression following treatment might be obtained. The adaptability of this strategy to the

study of disparate phases of development and differentiation will provide a wide array of

experimental avenues in the pursuit of mammary stem cells and improved treatment of

mammary cancers.

Page 187: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

170

References:

1. Halban, J., Die innere Sekretion von Ovarium und Placenta und ihre Bedeutung fuer

die Function der Milchdruese. Mschr. Geburtsh Gynaek, 1900. 12: p. 496-503.

2. Robinson, G.W., et al., Mammary epithelial cells undergo secretory differentiation in

cycling virgins but require pregnancy for the establishment of terminal

differentiation. Development, 1995. 121(7): p. 2079-90.

3. Chu, E.Y., et al., Canonical WNT signaling promotes mammary placode development

and is essential for initiation of mammary gland morphogenesis. Development, 2004.

131(19): p. 4819-29.

4. Hens, J.R. and J.J. Wysolmerski, Key stages of mammary gland development:

molecular mechanisms involved in the formation of the embryonic mammary gland.

Breast Cancer Res, 2005. 7(5): p. 220-4.

5. Daniel C.W., S.G.B., ed. Postnatal development of the rodent mammary gland. The

mammary gland: Development, regulation, and function, ed. D.C.W. Neville M.C.

1987, Plenum Press: New York, NY. 3-36.

6. Humphreys, R.C., et al., Apoptosis in the terminal endbud of the murine mammary

gland: a mechanism of ductal morphogenesis. Development, 1996. 122(12): p. 4013-

22.

7. Bern, H.A., S. Nandi, and V. Finster, Induction of lactation in precancerous

hyperplastic alveolar nodules in the mammary gland of C3H/He Crgl mice.

Experientia, 1959. 15(4): p. 155-7.

8. Sternlicht, M.D., et al., Hormonal and local control of mammary branching

morphogenesis. Differentiation, 2006. 74(7): p. 365-81.

9. Puca, G.A. and F. Bresciani, Decrease to complete loss of ability to bind estradiol by

a soluble macromolecular fraction of C3H mammary tumors. Eur J Cancer, 1968.

3(6): p. 475-9.

10. Silberstein, G.B., Postnatal mammary gland morphogenesis. Microsc Res Tech,

2001. 52(2): p. 155-62.

11. Nelson, C.M. and M.J. Bissell, Of extracellular matrix, scaffolds, and signaling:

tissue architecture regulates development, homeostasis, and cancer. Annu Rev Cell

Dev Biol, 2006. 22: p. 287-309.

12. Silberstein, G.B., et al., Regulation of mammary morphogenesis: evidence for

extracellular matrix-mediated inhibition of ductal budding by transforming growth

factor-beta 1. Dev Biol, 1992. 152(2): p. 354-62.

13. Joseph, H., et al., Overexpression of a kinase-deficient transforming growth factor-

beta type II receptor in mouse mammary stroma results in increased epithelial

branching. Mol Biol Cell, 1999. 10(4): p. 1221-34.

14. Brisken, C., et al., A paracrine role for the epithelial progesterone receptor in

mammary gland development. Proc Natl Acad Sci U S A, 1998. 95(9): p. 5076-81.

15. Bocchinfuso, W.P., et al., Induction of mammary gland development in estrogen

receptor-alpha knockout mice. Endocrinology, 2000. 141(8): p. 2982-94.

16. Lin, C.Q. and M.J. Bissell, Multi-faceted regulation of cell differentiation by

extracellular matrix. FASEB J, 1993. 7(9): p. 737-43.

17. Brisken, C., Hormonal control of alveolar development and its implications for breast

carcinogenesis. J Mammary Gland Biol Neoplasia, 2002. 7(1): p. 39-48.

Page 188: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

171

18. Li, M., et al., Mammary-derived signals activate programmed cell death during the

first stage of mammary gland involution. Proc Natl Acad Sci U S A, 1997. 94(7): p.

3425-30.

19. Strange, R., et al., Apoptotic cell death and tissue remodelling during mouse

mammary gland involution. Development, 1992. 115(1): p. 49-58.

20. Streuli, C., Extracellular matrix remodelling and cellular differentiation. Curr Opin

Cell Biol, 1999. 11(5): p. 634-40.

21. Monks, J., et al., Epithelial cells remove apoptotic epithelial cells during post-

lactation involution of the mouse mammary gland. Biol Reprod, 2008. 78(4): p. 586-

94.

22. Radice, G.L., et al., Precocious mammary gland development in P-cadherin-deficient

mice. J Cell Biol, 1997. 139(4): p. 1025-32.

23. Williams, J.M. and C.W. Daniel, Mammary ductal elongation: differentiation of

myoepithelium and basal lamina during branching morphogenesis. Dev Biol, 1983.

97(2): p. 274-90.

24. Streuli, C.H. and M.J. Bissell, Expression of extracellular matrix components is

regulated by substratum. J Cell Biol, 1990. 110(4): p. 1405-15.

25. Talhouk, R.S., M.J. Bissell, and Z. Werb, Coordinated expression of extracellular

matrix-degrading proteinases and their inhibitors regulates mammary epithelial

function during involution. J Cell Biol, 1992. 118(5): p. 1271-82.

26. Asch, H.L. and B.B. Asch, Expression of keratins and other cytoskeletal proteins in

mouse mammary epithelium during the normal developmental cycle and primary

culture. Dev Biol, 1985. 107(2): p. 470-82.

27. Shackleton, M., et al., Generation of a functional mammary gland from a single stem

cell. Nature, 2006. 439(7072): p. 84-8.

28. Sleeman, K.E., et al., CD24 staining of mouse mammary gland cells defines luminal

epithelial, myoepithelial/basal and non-epithelial cells. Breast Cancer Res, 2006.

8(1): p. R7.

29. Stingl, J., et al., Purification and unique properties of mammary epithelial stem cells.

Nature, 2006. 439(7079): p. 993-7.

30. Cheng, G., et al., Estrogen receptors ER alpha and ER beta in proliferation in the

rodent mammary gland. Proc Natl Acad Sci U S A, 2004. 101(11): p. 3739-46.

31. Smalley, M.J., et al., Differentiation of separated mouse mammary luminal epithelial

and myoepithelial cells cultured on EHS matrix analyzed by indirect

immunofluorescence of cytoskeletal antigens. J Histochem Cytochem, 1999. 47(12):

p. 1513-24.

32. Linzell, J.L., The silver staining of myoepithelial cells, particularly in the mammary

gland, and their relation to the ejection of milk. J Anat, 1952. 86(1): p. 49-57.

33. Linzell, J.L., Some observations on the contractile tissue of the mammary glands. J

Physiol, 1955. 130(2): p. 257-67.

34. Smith, G.H., T. Mehrel, and D.R. Roop, Differential keratin gene expression in

developing, differentiating, preneoplastic, and neoplastic mouse mammary

epithelium. Cell Growth Differ, 1990. 1(4): p. 161-70.

35. Glukhova, M., et al., Adhesion systems in normal breast and in invasive breast

carcinoma. Am J Pathol, 1995. 146(3): p. 706-16.

Page 189: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

172

36. Djonov, V., et al., MMP-19: cellular localization of a novel metalloproteinase within

normal breast tissue and mammary gland tumours. J Pathol, 2001. 195(2): p. 147-55.

37. Gudjonsson, T., et al., Normal and tumor-derived myoepithelial cells differ in their

ability to interact with luminal breast epithelial cells for polarity and basement

membrane deposition. J Cell Sci, 2002. 115(Pt 1): p. 39-50.

38. Runswick, S.K., et al., Desmosomal adhesion regulates epithelial morphogenesis and

cell positioning. Nat Cell Biol, 2001. 3(9): p. 823-30.

39. Neville, M.C., et al., The mammary fat pad. J Mammary Gland Biol Neoplasia, 1998.

3(2): p. 109-16.

40. Kouros-Mehr, H. and Z. Werb, Candidate regulators of mammary branching

morphogenesis identified by genome-wide transcript analysis. Dev Dyn, 2006.

235(12): p. 3404-12.

41. Richert, M.M., et al., An atlas of mouse mammary gland development. J Mammary

Gland Biol Neoplasia, 2000. 5(2): p. 227-41.

42. Stamenkovic, I., Extracellular matrix remodelling: the role of matrix

metalloproteinases. J Pathol, 2003. 200(4): p. 448-64.

43. Daniel, C.W., Regulation of cell division in aging mouse mammary epithelium. Adv

Exp Med Biol, 1975. 61: p. 1-19.

44. Daniel, C.W., et al., Unlimited division potential of precancerous mouse mammary

cells after spontaneous or carcinogen-induced transformation. Fed Proc, 1975. 34(1):

p. 64-7.

45. Deome, K.B., et al., Development of mammary tumors from hyperplastic alveolar

nodules transplanted into gland-free mammary fat pads of female C3H mice. Cancer

Res, 1959. 19(5): p. 515-20.

46. Hoshino, K. and W.U. Gardner, Transplantability and life span of mammary gland

during serial transplantation in mice. Nature, 1967. 213(5072): p. 193-4.

47. Asselin-Labat, M.L., et al., Steroid hormone receptor status of mouse mammary stem

cells. J Natl Cancer Inst, 2006. 98(14): p. 1011-4.

48. Sleeman, K.E., et al., Dissociation of estrogen receptor expression and in vivo stem

cell activity in the mammary gland. J Cell Biol, 2007. 176(1): p. 19-26.

49. Asselin-Labat, M.L., et al., Control of mammary stem cell function by steroid

hormone signalling. Nature, 2010. 465(7299): p. 798-802.

50. Stingl, J., et al., Phenotypic and functional characterization in vitro of a multipotent

epithelial cell present in the normal adult human breast. Differentiation, 1998. 63(4):

p. 201-13.

51. Asselin-Labat, M.L., et al., Gata-3 is an essential regulator of mammary-gland

morphogenesis and luminal-cell differentiation. Nat Cell Biol, 2007. 9(2): p. 201-9.

52. Asselin-Labat, M.L., et al., Delineating the epithelial hierarchy in the mouse

mammary gland. Cold Spring Harb Symp Quant Biol, 2008. 73: p. 469-78.

53. Bouras, T., et al., Notch signaling regulates mammary stem cell function and luminal

cell-fate commitment. Cell Stem Cell, 2008. 3(4): p. 429-41.

54. Oakes, S.R., et al., The Ets transcription factor Elf5 specifies mammary alveolar cell

fate. Genes Dev, 2008. 22(5): p. 581-6.

55. Harris, J., et al., Socs2 and elf5 mediate prolactin-induced mammary gland

development. Mol Endocrinol, 2006. 20(5): p. 1177-87.

Page 190: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

173

56. Vaillant, F., et al., The emerging picture of the mouse mammary stem cell. Stem Cell

Rev, 2007. 3(2): p. 114-23.

57. Wellings, S.R., A hypothesis of the origin of human breast cancer from the terminal

ductal lobular unit. Pathol Res Pract, 1980. 166(4): p. 515-35.

58. Wellings, S.R., Development of human breast cancer. Adv Cancer Res, 1980. 31: p.

287-314.

59. Cardiff, R.D., D. Moghanaki, and R.A. Jensen, Genetically engineered mouse models

of mammary intraepithelial neoplasia. J Mammary Gland Biol Neoplasia, 2000. 5(4):

p. 421-37.

60. Cardiff, R.D., et al., The mammary pathology of genetically engineered mice: the

consensus report and recommendations from the Annapolis meeting. Oncogene,

2000. 19(8): p. 968-88.

61. Wijayaratne, A.L. and D.P. McDonnell, The human estrogen receptor-alpha is a

ubiquitinated protein whose stability is affected differentially by agonists,

antagonists, and selective estrogen receptor modulators. J Biol Chem, 2001. 276(38):

p. 35684-92.

62. Wijayaratne, A.L., et al., Comparative analyses of mechanistic differences among

antiestrogens. Endocrinology, 1999. 140(12): p. 5828-40.

63. Systemic treatment of early breast cancer by hormonal, cytotoxic, or immune therapy.

133 randomised trials involving 31,000 recurrences and 24,000 deaths among 75,000

women. Early Breast Cancer Trialists' Collaborative Group. Lancet, 1992.

339(8784): p. 1-15.

64. Masamura, S., et al., Estrogen deprivation causes estradiol hypersensitivity in human

breast cancer cells. J Clin Endocrinol Metab, 1995. 80(10): p. 2918-25.

65. Masamura, S., et al., Aromatase inhibitor development for treatment of breast cancer.

Breast Cancer Res Treat, 1995. 33(1): p. 19-26.

66. Dent, R., et al., Triple-negative breast cancer: clinical features and patterns of

recurrence. Clin Cancer Res, 2007. 13(15 Pt 1): p. 4429-34.

67. Laakso, M., et al., Cytokeratin 5/14-positive breast cancer: true basal phenotype

confined to BRCA1 tumors. Mod Pathol, 2005. 18(10): p. 1321-8.

68. Li, H., et al., Nestin is expressed in the basal/myoepithelial layer of the mammary

gland and is a selective marker of basal epithelial breast tumors. Cancer Res, 2007.

67(2): p. 501-10.

69. Coussens, L., et al., Tyrosine kinase receptor with extensive homology to EGF

receptor shares chromosomal location with neu oncogene. Science, 1985. 230(4730):

p. 1132-9.

70. Craven, R.J., H. Lightfoot, and W.G. Cance, A decade of tyrosine kinases: from gene

discovery to therapeutics. Surg Oncol, 2003. 12(1): p. 39-49.

71. King, C.R., M.H. Kraus, and S.A. Aaronson, Amplification of a novel v-erbB-related

gene in a human mammary carcinoma. Science, 1985. 229(4717): p. 974-6.

72. Slamon, D.J., et al., Human breast cancer: correlation of relapse and survival with

amplification of the HER-2/neu oncogene. Science, 1987. 235(4785): p. 177-82.

73. Slamon, D.J., et al., Studies of the HER-2/neu proto-oncogene in human breast and

ovarian cancer. Science, 1989. 244(4905): p. 707-12.

Page 191: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

174

74. Hudziak, R.M., J. Schlessinger, and A. Ullrich, Increased expression of the putative

growth factor receptor p185HER2 causes transformation and tumorigenesis of NIH

3T3 cells. Proc Natl Acad Sci U S A, 1987. 84(20): p. 7159-63.

75. Guy, C.T., et al., Expression of the neu protooncogene in the mammary epithelium of

transgenic mice induces metastatic disease. Proc Natl Acad Sci U S A, 1992. 89(22):

p. 10578-82.

76. Yarden, R.I., M.A. Wilson, and S.A. Chrysogelos, Estrogen suppression of EGFR

expression in breast cancer cells: a possible mechanism to modulate growth. J Cell

Biochem Suppl, 2001. Suppl 36: p. 232-46.

77. Hudziak, R.M., et al., p185HER2 monoclonal antibody has antiproliferative effects in

vitro and sensitizes human breast tumor cells to tumor necrosis factor. Mol Cell Biol,

1989. 9(3): p. 1165-72.

78. Konecny, G.E., et al., Activity of the dual kinase inhibitor lapatinib (GW572016)

against HER-2-overexpressing and trastuzumab-treated breast cancer cells. Cancer

Res, 2006. 66(3): p. 1630-9.

79. Sorlie, T., et al., Repeated observation of breast tumor subtypes in independent gene

expression data sets. Proc Natl Acad Sci U S A, 2003. 100(14): p. 8418-23.

80. Sorlie, T., et al., Gene expression patterns of breast carcinomas distinguish tumor

subclasses with clinical implications. Proc Natl Acad Sci U S A, 2001. 98(19): p.

10869-74.

81. Perou, C.M., et al., Molecular portraits of human breast tumours. Nature, 2000.

406(6797): p. 747-52.

82. Neve, R.M., et al., A collection of breast cancer cell lines for the study of functionally

distinct cancer subtypes. Cancer Cell, 2006. 10(6): p. 515-27.

83. Weissman, I.L. and J.A. Shizuru, The origins of the identification and isolation of

hematopoietic stem cells, and their capability to induce donor-specific

transplantation tolerance and treat autoimmune diseases. Blood, 2008. 112(9): p.

3543-53.

84. Weissman, I.L., D.J. Anderson, and F. Gage, Stem and progenitor cells: origins,

phenotypes, lineage commitments, and transdifferentiations. Annu Rev Cell Dev

Biol, 2001. 17: p. 387-403.

85. Weissman, I.L., Normal and neoplastic stem cells. Novartis Found Symp, 2005. 265:

p. 35-50; discussion 50-4, 92-7.

86. Dick, J.E., Acute myeloid leukemia stem cells. Ann N Y Acad Sci, 2005. 1044: p. 1-5.

87. Bonnet, D. and J.E. Dick, Human acute myeloid leukemia is organized as a hierarchy

that originates from a primitive hematopoietic cell. Nat Med, 1997. 3(7): p. 730-7.

88. Al-Hajj, M., et al., Prospective identification of tumorigenic breast cancer cells. Proc

Natl Acad Sci U S A, 2003. 100(7): p. 3983-8.

89. Molyneux, G., et al., BRCA1 basal-like breast cancers originate from luminal

epithelial progenitors and not from basal stem cells. Cell Stem Cell, 2010. 7(3): p.

403-17.

90. Ince, T.A., et al., Transformation of different human breast epithelial cell types leads

to distinct tumor phenotypes. Cancer Cell, 2007. 12(2): p. 160-70.

91. Uhr, J.W., et al., Cancer dormancy: opportunities for new therapeutic approaches.

Nat Med, 1997. 3(5): p. 505-9.

92. Meltzer, A., Dormancy and breast cancer. J Surg Oncol, 1990. 43(3): p. 181-8.

Page 192: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

175

93. Welch, H.G. and W.C. Black, Using autopsy series to estimate the disease

"reservoir" for ductal carcinoma in situ of the breast: how much more breast cancer

can we find? Ann Intern Med, 1997. 127(11): p. 1023-8.

94. Demicheli, R., Tumour dormancy: findings and hypotheses from clinical research on

breast cancer. Semin Cancer Biol, 2001. 11(4): p. 297-306.

95. Demicheli, R., et al., Local recurrences following mastectomy: support for the

concept of tumor dormancy. J Natl Cancer Inst, 1994. 86(1): p. 45-8.

96. Aguirre-Ghiso, J.A., Models, mechanisms and clinical evidence for cancer dormancy.

Nat Rev Cancer, 2007. 7(11): p. 834-46.

97. Koebel, C.M., et al., Adaptive immunity maintains occult cancer in an equilibrium

state. Nature, 2007. 450(7171): p. 903-7.

98. Teng, M.W., et al., Immune-mediated dormancy: an equilibrium with cancer. J

Leukoc Biol, 2008. 84(4): p. 988-93.

99. Watnick, R.S., et al., Ras modulates Myc activity to repress thrombospondin-1

expression and increase tumor angiogenesis. Cancer Cell, 2003. 3(3): p. 219-31.

100. Vessella, R.L., K. Pantel, and S. Mohla, Tumor cell dormancy: an NCI workshop

report. Cancer Biol Ther, 2007. 6(9): p. 1496-504.

101. Kelsey, J.L. and M.D. Gammon, Epidemiology of breast cancer. Epidemiol Rev,

1990. 12: p. 228-40.

102. Kelsey, J.L. and M.D. Gammon, The epidemiology of breast cancer. CA Cancer J

Clin, 1991. 41(3): p. 146-65.

103. Kelsey, J.L., M.D. Gammon, and E.M. John, Reproductive factors and breast cancer.

Epidemiol Rev, 1993. 15(1): p. 36-47.

104. Lambe, M., et al., Parity, age at first and last birth, and risk of breast cancer: a

population-based study in Sweden. Breast Cancer Res Treat, 1996. 38(3): p. 305-11.

105. Layde, P.M., et al., The independent associations of parity, age at first full term

pregnancy, and duration of breastfeeding with the risk of breast cancer. Cancer and

Steroid Hormone Study Group. J Clin Epidemiol, 1989. 42(10): p. 963-73.

106. Russo, J., et al., The protective role of pregnancy in breast cancer. Breast Cancer

Res, 2005. 7(3): p. 131-42.

107. Russo, J., L.K. Tay, and I.H. Russo, Differentiation of the mammary gland and

susceptibility to carcinogenesis. Breast Cancer Res Treat, 1982. 2(1): p. 5-73.

108. Jerry, D.J., et al., Regulation of apoptosis during mammary involution by the p53

tumor suppressor gene. J Dairy Sci, 2002. 85(5): p. 1103-10.

109. Jerry, D.J., et al., Regulation of p53 and its targets during involution of the mammary

gland. J Mammary Gland Biol Neoplasia, 1999. 4(2): p. 177-81.

110. Jerry, D.J., et al., Delayed involution of the mammary epithelium in BALB/c-p53null

mice. Oncogene, 1998. 17(18): p. 2305-12.

111. Cardiff, R.D. and S.R. Wellings, The comparative pathology of human and mouse

mammary glands. J Mammary Gland Biol Neoplasia, 1999. 4(1): p. 105-22.

112. Moll, R., et al., The catalog of human cytokeratins: patterns of expression in normal

epithelia, tumors and cultured cells. Cell, 1982. 31(1): p. 11-24.

113. Haagensen, C.D., The physiology of the breast as it concerns the clinician. Am J

Obstet Gynecol, 1971. 109(2): p. 206-9.

114. Ailhaud, G., P. Grimaldi, and R. Negrel, Cellular and molecular aspects of adipose

tissue development. Annu Rev Nutr, 1992. 12: p. 207-33.

Page 193: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

176

115. Bartow, S.A., Use of the autopsy to study ontogeny and expression of the estrogen

receptor gene in human breast. J Mammary Gland Biol Neoplasia, 1998. 3(1): p. 37-

48.

116. Haslam, S.Z., et al., Progestin-regulated luminal cell and myoepithelial cell-specific

responses in mammary organoid culture. Endocrinology, 2008. 149(5): p. 2098-107.

117. Strange, R., et al., Programmed cell death during mammary gland involution.

Methods Cell Biol, 1995. 46: p. 355-68.

118. Green, R.G., M.M. Moosey, and J.J. Bittner, Serial transmission of the milk agent of

mouse mammary carcinoma. Proc Soc Exp Biol Med, 1946. 61: p. 362.

119. Barnum, C.P., et al., The Milk Agent in Spontaneous Mammary Carcinoma. Science,

1944. 100(2608): p. 575-6.

120. Bittner, J.J., The Milk-Influence of Breast Tumors in Mice. Science, 1942. 95(2470):

p. 462-3.

121. Bittner, J.J., The Preservation by Freezing and Drying in Vacuo of the Milk-Influence

for the Development of Breast Cancer in Mice. Science, 1941. 93(2422): p. 527-8.

122. Bittner, J.J., Some Possible Effects of Nursing on the Mammary Gland Tumor

Incidence in Mice. Science, 1936. 84(2172): p. 162.

123. Porter, K.R. and H.P. Thompson, A particulate body associated with epithelial cells

cultured from mammary carcinomas of mice of a milkfactor strain. J Exp Med, 1948.

88(1): p. 15-24.

124. Dux, A. and O. Muhlbock, Propagation of the mammary tumor agent (Bittner virus)

in the absence of mammary glands in mice. J Natl Cancer Inst, 1968. 40(6): p. 1309-

12.

125. Dux, A. and O. Muhlbock, Susceptibility of mammary tissues of different strains of

mice to tumor development. J Natl Cancer Inst, 1968. 40(6): p. 1259-72.

126. Hageman, P.C., J. Links, and P. Bentvelzen, Biological properties of B particles from

C3H and C3Hf mouse milk. J Natl Cancer Inst, 1968. 40(6): p. 1319-24.

127. Nusse, R. and H.E. Varmus, Many tumors induced by the mouse mammary tumor

virus contain a provirus integrated in the same region of the host genome. Cell, 1982.

31(1): p. 99-109.

128. Nusse, R., et al., A new nomenclature for int-1 and related genes: the Wnt gene

family. Cell, 1991. 64(2): p. 231.

129. Roelink, H., et al., Wnt-3, a gene activated by proviral insertion in mouse mammary

tumors, is homologous to int-1/Wnt-1 and is normally expressed in mouse embryos

and adult brain. Proc Natl Acad Sci U S A, 1990. 87(12): p. 4519-23.

130. Lee, F.S., et al., Insertional mutagenesis identifies a member of the Wnt gene family

as a candidate oncogene in the mammary epithelium of int-2/Fgf-3 transgenic mice.

Proc Natl Acad Sci U S A, 1995. 92(6): p. 2268-72.

131. Dickson, C., et al., Tumorigenesis by mouse mammary tumor virus: proviral

activation of a cellular gene in the common integration region int-2. Cell, 1984.

37(2): p. 529-36.

132. Peters, G., et al., The mouse homolog of the hst/k-FGF gene is adjacent to int-2 and is

activated by proviral insertion in some virally induced mammary tumors. Proc Natl

Acad Sci U S A, 1989. 86(15): p. 5678-82.

133. Bhanot, P., et al., A new member of the frizzled family from Drosophila functions as a

Wingless receptor. Nature, 1996. 382(6588): p. 225-30.

Page 194: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

177

134. Dann, C.E., et al., Insights into Wnt binding and signalling from the structures of two

Frizzled cysteine-rich domains. Nature, 2001. 412(6842): p. 86-90.

135. Chen, W., et al., Dishevelled 2 recruits beta-arrestin 2 to mediate Wnt5A-stimulated

endocytosis of Frizzled 4. Science, 2003. 301(5638): p. 1391-4.

136. Mao, J., et al., Low-density lipoprotein receptor-related protein-5 binds to Axin and

regulates the canonical Wnt signaling pathway. Mol Cell, 2001. 7(4): p. 801-9.

137. Zeng, X., et al., A dual-kinase mechanism for Wnt co-receptor phosphorylation and

activation. Nature, 2005. 438(7069): p. 873-7.

138. Davidson, G., et al., Casein kinase 1 gamma couples Wnt receptor activation to

cytoplasmic signal transduction. Nature, 2005. 438(7069): p. 867-72.

139. Brunner, E., et al., pangolin encodes a Lef-1 homologue that acts downstream of

Armadillo to transduce the Wingless signal in Drosophila. Nature, 1997. 385(6619):

p. 829-33.

140. van de Wetering, M., et al., Armadillo coactivates transcription driven by the product

of the Drosophila segment polarity gene dTCF. Cell, 1997. 88(6): p. 789-99.

141. Polakis, P., Wnt signaling and cancer. Genes Dev, 2000. 14(15): p. 1837-51.

142. Spink, K.E., P. Polakis, and W.I. Weis, Structural basis of the Axin-adenomatous

polyposis coli interaction. EMBO J, 2000. 19(10): p. 2270-9.

143. Tsukamoto, A.S., et al., Expression of the int-1 gene in transgenic mice is associated

with mammary gland hyperplasia and adenocarcinomas in male and female mice.

Cell, 1988. 55(4): p. 619-25.

144. Cato, A.C., D. Henderson, and H. Ponta, The hormone response element of the mouse

mammary tumour virus DNA mediates the progestin and androgen induction of

transcription in the proviral long terminal repeat region. EMBO J, 1987. 6(2): p.

363-8.

145. Lin, T.P., et al., Role of endocrine, autocrine, and paracrine interactions in the

development of mammary hyperplasia in Wnt-1 transgenic mice. Cancer Res, 1992.

52(16): p. 4413-9.

146. Cunha, G.R. and Y.K. Hom, Role of mesenchymal-epithelial interactions in

mammary gland development. J Mammary Gland Biol Neoplasia, 1996. 1(1): p. 21-

35.

147. Shackleford, G.M., et al., Mouse mammary tumor virus infection accelerates

mammary carcinogenesis in Wnt-1 transgenic mice by insertional activation of int-

2/Fgf-3 and hst/Fgf-4. Proc Natl Acad Sci U S A, 1993. 90(2): p. 740-4.

148. Kwan, H., et al., Transgenes expressing the Wnt-1 and int-2 proto-oncogenes

cooperate during mammary carcinogenesis in doubly transgenic mice. Mol Cell Biol,

1992. 12(1): p. 147-54.

149. Gavin, B.J., J.A. McMahon, and A.P. McMahon, Expression of multiple novel Wnt-

1/int-1-related genes during fetal and adult mouse development. Genes Dev, 1990.

4(12B): p. 2319-32.

150. Jakobovits, A., et al., Two proto-oncogenes implicated in mammary carcinogenesis,

int-1 and int-2, are independently regulated during mouse development. Proc Natl

Acad Sci U S A, 1986. 83(20): p. 7806-10.

151. Roelink, H. and R. Nusse, Expression of two members of the Wnt family during

mouse development--restricted temporal and spatial patterns in the developing neural

tube. Genes Dev, 1991. 5(3): p. 381-8.

Page 195: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

178

152. DasGupta, R. and E. Fuchs, Multiple roles for activated LEF/TCF transcription

complexes during hair follicle development and differentiation. Development, 1999.

126(20): p. 4557-68.

153. Weber-Hall, S.J., et al., Developmental and hormonal regulation of Wnt gene

expression in the mouse mammary gland. Differentiation, 1994. 57(3): p. 205-14.

154. Gavin, B.J. and A.P. McMahon, Differential regulation of the Wnt gene family during

pregnancy and lactation suggests a role in postnatal development of the mammary

gland. Mol Cell Biol, 1992. 12(5): p. 2418-23.

155. Brisken, C., et al., Essential function of Wnt-4 in mammary gland development

downstream of progesterone signaling. Genes Dev, 2000. 14(6): p. 650-4.

156. Hsu, W., R. Shakya, and F. Costantini, Impaired mammary gland and lymphoid

development caused by inducible expression of Axin in transgenic mice. J Cell Biol,

2001. 155(6): p. 1055-64.

157. Lindvall, C., et al., The Wnt signaling receptor Lrp5 is required for mammary ductal

stem cell activity and Wnt1-induced tumorigenesis. J Biol Chem, 2006. 281(46): p.

35081-7.

158. Liu, B.Y., et al., The transforming activity of Wnt effectors correlates with their

ability to induce the accumulation of mammary progenitor cells. Proc Natl Acad Sci

U S A, 2004. 101(12): p. 4158-63.

159. Rosner, A., et al., Pathway pathology: histological differences between ErbB/Ras and

Wnt pathway transgenic mammary tumors. Am J Pathol, 2002. 161(3): p. 1087-97.

160. Li, Y., et al., Evidence that transgenes encoding components of the Wnt signaling

pathway preferentially induce mammary cancers from progenitor cells. Proc Natl

Acad Sci U S A, 2003. 100(26): p. 15853-8.

161. Van de Vijver, M.J. and R. Nusse, The molecular biology of breast cancer. Biochim

Biophys Acta, 1991. 1072(1): p. 33-50.

162. Ayyanan, A., et al., Increased Wnt signaling triggers oncogenic conversion of human

breast epithelial cells by a Notch-dependent mechanism. Proc Natl Acad Sci U S A,

2006. 103(10): p. 3799-804.

163. Hingorani, S.R. and D.A. Tuveson, Targeting oncogene dependence and resistance.

Cancer Cell, 2003. 3(5): p. 414-7.

164. Huguet, E.L., et al., Differential expression of human Wnt genes 2, 3, 4, and 7B in

human breast cell lines and normal and disease states of human breast tissue. Cancer

Res, 1994. 54(10): p. 2615-21.

165. Dale, T.C., et al., Compartment switching of WNT-2 expression in human breast

tumors. Cancer Res, 1996. 56(19): p. 4320-3.

166. Lejeune, S., et al., Wnt5a cloning, expression, and up-regulation in human primary

breast cancers. Clin Cancer Res, 1995. 1(2): p. 215-22.

167. Bui, T.D., et al., A novel human Wnt gene, WNT10B, maps to 12q13 and is expressed

in human breast carcinomas. Oncogene, 1997. 14(10): p. 1249-53.

168. Ozaki, S., et al., Alterations and correlations of the components in the Wnt signaling

pathway and its target genes in breast cancer. Oncol Rep, 2005. 14(6): p. 1437-43.

169. Lin, S.Y., et al., Beta-catenin, a novel prognostic marker for breast cancer: its roles

in cyclin D1 expression and cancer progression. Proc Natl Acad Sci U S A, 2000.

97(8): p. 4262-6.

Page 196: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

179

170. Wissmann, C., et al., WIF1, a component of the Wnt pathway, is down-regulated in

prostate, breast, lung, and bladder cancer. J Pathol, 2003. 201(2): p. 204-12.

171. Ai, L., et al., Inactivation of Wnt inhibitory factor-1 (WIF1) expression by epigenetic

silencing is a common event in breast cancer. Carcinogenesis, 2006. 27(7): p. 1341-8.

172. Gunther, E.J., et al., A novel doxycycline-inducible system for the transgenic analysis

of mammary gland biology. FASEB J, 2002. 16(3): p. 283-92.

173. Gunther, E.J., et al., Impact of p53 loss on reversal and recurrence of conditional

Wnt-induced tumorigenesis. Genes Dev, 2003. 17(4): p. 488-501.

174. Gestl, S.A., et al., Dormant Wnt-initiated mammary cancer can participate in

reconstituting functional mammary glands. Mol Cell Biol, 2007. 27(1): p. 195-207.

175. Guy, C.T., R.D. Cardiff, and W.J. Muller, Activated neu induces rapid tumor

progression. J Biol Chem, 1996. 271(13): p. 7673-8.

176. Henry, M.D., et al., Parity-induced mammary epithelial cells facilitate tumorigenesis

in MMTV-neu transgenic mice. Oncogene, 2004. 23(41): p. 6980-5.

177. Guy, C.T., R.D. Cardiff, and W.J. Muller, Induction of mammary tumors by

expression of polyomavirus middle T oncogene: a transgenic mouse model for

metastatic disease. Mol Cell Biol, 1992. 12(3): p. 954-61.

178. Mok, E., T.V. Golovkina, and S.R. Ross, A mouse mammary tumor virus mammary

gland enhancer confers tissue-specific but not lactation-dependent expression in

transgenic mice. J Virol, 1992. 66(12): p. 7529-32.

179. Stamp, G., et al., Nonuniform expression of a mouse mammary tumor virus-driven

int-2/Fgf-3 transgene in pregnancy-responsive breast tumors. Cell Growth Differ,

1992. 3(12): p. 929-38.

180. Diamond, I., et al., Conditional gene expression in the epidermis of transgenic mice

using the tetracycline-regulated transactivators tTA and rTA linked to the keratin 5

promoter. J Invest Dermatol, 2000. 115(5): p. 788-94.

181. Vassar, R., et al., Tissue-specific and differentiation-specific expression of a human

K14 keratin gene in transgenic mice. Proc Natl Acad Sci U S A, 1989. 86(5): p. 1563-

7.

182. Tumbar, T., et al., Defining the epithelial stem cell niche in skin. Science, 2004.

303(5656): p. 359-63.

183. Kanda, T., K.F. Sullivan, and G.M. Wahl, Histone-GFP fusion protein enables

sensitive analysis of chromosome dynamics in living mammalian cells. Curr Biol,

1998. 8(7): p. 377-85.

184. Hadjantonakis, A.K. and V.E. Papaioannou, Dynamic in vivo imaging and cell

tracking using a histone fluorescent protein fusion in mice. BMC Biotechnol, 2004. 4:

p. 33.

185. Grimm, S.L., et al., Keratin 6 is not essential for mammary gland development.

Breast Cancer Res, 2006. 8(3): p. R29.

186. Teuliere, J., et al., Targeted activation of beta-catenin signaling in basal mammary

epithelial cells affects mammary development and leads to hyperplasia.

Development, 2005. 132(2): p. 267-77.

187. Visvader, J.E., Keeping abreast of the mammary epithelial hierarchy and breast

tumorigenesis. Genes Dev, 2009. 23(22): p. 2563-77.

188. Yang, M., et al., Facile whole-body imaging of internal fluorescent tumors in mice

with an LED flashlight. Biotechniques, 2005. 39(2): p. 170, 172.

Page 197: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

180

189. Daniel, C.W., G.B. Silberstein, and P. Strickland, Direct action of 17 beta-estradiol

on mouse mammary ducts analyzed by sustained release implants and steroid

autoradiography. Cancer Res, 1987. 47(22): p. 6052-7.

190. Brennand, K., D. Huangfu, and D. Melton, All beta cells contribute equally to islet

growth and maintenance. PLoS Biol, 2007. 5(7): p. e163.

191. Waghmare, S.K., et al., Quantitative proliferation dynamics and random chromosome

segregation of hair follicle stem cells. EMBO J, 2008. 27(9): p. 1309-20.

192. Debies, M.T., et al., Tumor escape in a Wnt1-dependent mouse breast cancer model

is enabled by p19Arf/p53 pathway lesions but not p16 Ink4a loss. J Clin Invest, 2008.

118(1): p. 51-63.

193. Zhou, B.P. and M.C. Hung, Dysregulation of cellular signaling by HER2/neu in

breast cancer. Semin Oncol, 2003. 30(5 Suppl 16): p. 38-48.

194. Vaillant, F., et al., The mammary progenitor marker CD61/beta3 integrin identifies

cancer stem cells in mouse models of mammary tumorigenesis. Cancer Res, 2008.

68(19): p. 7711-7.

195. Gretschel, S., et al., Markers of tumour angiogenesis and tumour cells in bone

marrow in gastric cancer patients. Eur J Surg Oncol, 2008. 34(6): p. 642-7.

196. Allgayer, H. and J.A. Aguirre-Ghiso, The urokinase receptor (u-PAR)--a link between

tumor cell dormancy and minimal residual disease in bone marrow? APMIS, 2008.

116(7-8): p. 602-14.

197. Reya, T., et al., Stem cells, cancer, and cancer stem cells. Nature, 2001. 414(6859): p.

105-11.

198. Lyons, W.R., Sako, Y., Direct action of estrone on the mammary gland. Proceedings

of the Society for Experimental Biology and Medicine, 1940. 44: p. 398-401.

199. Deugnier, M.A., et al., Myoepithelial cell differentiation in the developing mammary

gland: progressive acquisition of smooth muscle phenotype. Dev Dyn, 1995. 204(2):

p. 107-17.

200. Matulka, L.A., A.A. Triplett, and K.U. Wagner, Parity-induced mammary epithelial

cells are multipotent and express cell surface markers associated with stem cells. Dev

Biol, 2007. 303(1): p. 29-44.

201. Boulanger, C.A., K.U. Wagner, and G.H. Smith, Parity-induced mouse mammary

epithelial cells are pluripotent, self-renewing and sensitive to TGF-beta1 expression.

Oncogene, 2005. 24(4): p. 552-60.

202. Creamer, B.A., A.A. Triplett, and K.U. Wagner, Longitudinal analysis of

mammogenesis using a novel tetracycline-inducible mouse model and in vivo

imaging. Genesis, 2009. 47(4): p. 234-45.

203. Cotsarelis, G., et al., Existence of slow-cycling limbal epithelial basal cells that can

be preferentially stimulated to proliferate: implications on epithelial stem cells. Cell,

1989. 57(2): p. 201-9.

204. Barker, N., M. van de Wetering, and H. Clevers, The intestinal stem cell. Genes Dev,

2008. 22(14): p. 1856-64.

205. Calabrese, C., et al., A perivascular niche for brain tumor stem cells. Cancer Cell,

2007. 11(1): p. 69-82.

206. Li, L. and T. Xie, Stem cell niche: structure and function. Annu Rev Cell Dev Biol,

2005. 21: p. 605-31.

Page 198: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

181

207. Lin, H., The stem-cell niche theory: lessons from flies. Nat Rev Genet, 2002. 3(12): p.

931-40.

208. Bussard, K.M., et al., Reprogramming human cancer cells in the mouse mammary

gland. Cancer Res, 2010. 70(15): p. 6336-43.

209. Morrison, J.I., et al., Salamander limb regeneration involves the activation of a

multipotent skeletal muscle satellite cell population. J Cell Biol, 2006. 172(3): p. 433-

40.

210. Sobkow, L., et al., A germline GFP transgenic axolotl and its use to track cell fate:

dual origin of the fin mesenchyme during development and the fate of blood cells

during regeneration. Dev Biol, 2006. 290(2): p. 386-97.

211. Kragl, M., et al., Cells keep a memory of their tissue origin during axolotl limb

regeneration. Nature, 2009. 460(7251): p. 60-5.

Page 199: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

182

Appendix: Flow Cytometric Analysis Demonstrates Variability of LRCs in MMTV-

Wnt1 Induced Mammary Tumor Explants, Additional Samples

Single cells isolated from primary mammary tumors of MMTV-rtTA/MMTV-Wnt1/TGFP

tritransgenic tumor explants were analyzed for GFP content by flow cytometry. Each graph

is an individual tumor generation with sequential generations appearing in the same column.

The arrow depicts the lineage of a tumor that was re-explanted from frozen storage. Never

Dox animals appear in Blue, Pulsed tumors are Magenta, and pulse then chased tumors

appear in both Black and Yellow. Wide variability of both extent of GFP labeling during

pulse and amount of GFP signal dilution during off-Dox chase can be visualized.

Page 200: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

183

Page 201: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

184

Page 202: TRACKING MAMMARY EPITHELIAL CELL LINEAGE AND CELL

Vita Jessica (Biddle) Mathers

Education

Ph.D. Candidate, Intercollege Graduate Degree Program in Genetics 2003-2011

The Pennsylvania State University College of Medicine, Hershey, PA

B.S. in Biology and International Studies with a minor in Spanish 1999-2003

The Pennsylvania State University and the Schreyer Honors College, University Park, PA

Honors Thesis: Characterization of a Novel Two-Component Signaling System in

Bordetella sps Thesis Advisor: Eric T. Harvill, Ph.D.

Honors and Service

Graham Endowed Graduate Fellowship 2003

the Pennsylvania State University, Graduate School

Junior Mentor, SURIP Program 2007

Symposium Organizer 2007

5th

Annual Genetics Intercollege Graduate Degree Program Symposium

Undergraduate Research Grant 2002

the Pennsylvania State University, College of Agriculture

Academic Excellence Scholarship 1999-2003

the Pennsylvania State University, Schreyer Honors College

Posters and Publications

Jessica Mathers, Shelley Gestl, Dan Beachler, and Edward Gunther. A Transgenic Model

for Short-Term Tracking of Mammary Epithelial Cell Fates During Pregnancy in Mice

(manuscript in preparation)

Michael T. Debies, Shelley A. Gestl, Jessica L. Mathers, Oliver R. Mikse, Travis L.

Leonard, Susan E. Moody, Lewis A. Chodosh, Robert D. Cardiff and Edward J. Gunther

(2008). Tumor escape in a Wnt1-dependent mouse breast cancer model is enabled by

p19Arf

/p53 pathway lesions but not p16Ink4a

loss. Journal of Clinical Investigation

2008;118(1):51–63

Jessica Mathers, Shelley Gestl, Edward Gunther. Identification of Quiescent Cells in a

Mouse Model of Dormant Mammary Cancer. International Society for Stem Cell Research,

Philadelphia, PA July 2008

Jessica Mathers and Edward Gunther. A Strategy for Labeling Viable, Infrequently Cycling

Cells in the Mouse Mammary Gland. Keystone Symposia: Stem Cells and Cancer, Keystone,

CO March 2007

Shelley A. Gestl, Travis L. Leonard, Jessica L. Biddle, Michael T. Debies, and Edward

J. Gunther (2007). Molecular and Cellular Biology 2007;27(1):195-207

Jessica Biddle and Eric T. Harvill. Characterization of a Novel Two-Component Signaling

System in Bordetella sps. American Society for Microbiology Annual Meeting, Washington,

DC May 2003