SOSTDC1: A BMP/WNT DUAL PATHWAY ANTAGONIST ......WAGR Wilms tumor, Aniridia, Genitourinary malformation, and mental Retardation syndrome Wg Wingless (D. melanogaster) WIF Wnt Inhibitory

SOSTDC1: A BMP/WNT DUAL PATHWAY ANTAGONIST IN BREAST AND RENAL CANCERS

BY

KIMBERLY ROSE BLISH

A Dissertation Submitted to the Graduate Faculty of WAKE FOREST UNIVERSITY GRADUATE SCHOOL OF ARTS AND

SCIENCES

in Partial Fulfillment of the Requirements

for the Degree of

DOCTOR OF PHILOSOPHY

Molecular Medicine and Translational Science

May 2009

Winston-Salem, North Carolina

Approved By:

Suzy V. Torti, Ph.D. Advisor ________________________________

Examining Committee:

Mark C. Willingham, M.D., Chairman ________________________________

Gregory A. Hawkins, Ph.D. ________________________________

Raymond Penn, Ph.D. ________________________________

William Jeffrey Petty, M.D. ________________________________

ii

ACKNOWLEDGEMENTS

At the end of a long research and writing process, I come at last to the part

where I can reflect upon the journey with gratitude. However, as I am typing

these words last, I am hampered by a diminished vocabulary and I know I shall

fall short on the most important words of the document! To all whom I salute

below: I sincerely could not have accomplished anything worthwhile without you.

These words will not do you justice for your support has meant the world to me.

This project would be nonexistent without the opportunities, guidance, and

experience offered from my advisor, Suzy Torti and co-advisor, Frank Torti. You

both challenged me to become a better scientist. I am indebted to you for giving

me and my ideas a “home” for the past 5 years.

I am grateful to my committee members (Mark Willingham, Greg Hawkins,

Anthony Atala, Ray Penn, Jeff Petty) and many other faculty collaborators who

took my ideas seriously, encouraged new directions, and gave of their time and

laboratory resources. I have acknowledged specific contributions where

appropriate throughout the text. Mark Willingham, Tim Kute, Greg Hawkins, and

members of their laboratories were colleagues with indispensible aid and

inexhaustible advice.

The entire Torti Lab contributed through advice, technical assistance, and at

times, “colleague commiseration”. Particular mention goes to Samer Kalakish,

iii

Matt Triplette, Wei Wang, Seon-Hi Jang, Lan Coffman, Zandra Pinnix, Alice

Mims, and Jonathan Storey for their direct contributions to this project. All

students rely heavily on laboratory technicians and I was no exception. My

thanks especially go to Rong Ma and Julie Brown for keeping the lab running

smoothly, yet finding time to teach and answer endless questions. I am

particularly indebted to Julie Brown for her hours of support, technical, moral, and

spiritual on this project.

I would not have arrived at Wake Forest in the first place without a lot of help. I

had not considered a PhD degree until the mentoring I received in college from

Lowell Hall and Cynthia Davis, both of whom invested in me – not only as one of

their students, but as a person. Mark Payne, David Bass, Cash McCall, Jamal

Ibdah, Paul Laurienti, Linda McPhail, Bridget Brosnihan, and Kevin High in the

Molecular Medicine and MD/PhD programs provided career and administrative

support during my transition into and throughout the dual-degree program. I

have had the honor of working with an incredibly high caliber of colleagues in the

Molecular Medicine, MD/PhD, and Cancer Biology programs and I am proud to

be part of these departments.

Personally, there was a formidable army of friends and family who would not let

me back down and were constantly rooting for me including:

• My prayer partners and cheerleaders- may you be abundantly blessed!

iv

• My extended family at Harvest Point Church, especially the youth group

who tried to keep me smiling through it all.

• The cancer survivors who shared their stories and their enthusiasm for

life: Susan Blish, Kate Hacker, and Bob Smith

• The Rose, Pomante, Smith, Blish, LaBarr, and Stotler families for

understanding my preoccupation, missed holidays, and late birthday cards

and forgiving me when finishing this thesis work became overwhelming.

At the end, I couldn’t have made it without the support network that helped

sustain me day to day (and provided excellent child care)!

• Mom and Dad- There are no better examples of self-sacrificing love and

support than you. I am the person I am, having achieved these things-

because of you. When the battle was long, you held up my arms.

• My husband, Drew. You have put up with everything you promised and

more! “Time after Time.” I can’t imagine anyone else I’d rather share this

with.

I dedicate this work to:

Evelyn Smith & Elaine Rose, because I am in awe of all that has been done,

Drew, because every day is worth it,

and Ethan, because there’s hope for the future.

Glory to the One who was, and is, and is to come (Revelations 1:8)

v

TABLE OF CONTENTS Page LIST OF ABBREVIATIONS ………………………………………………….. vi

LIST OF ILLUSTRATIONS …………………………………………………... xi

ABSTRACT…………………………………………………………………….. xiv

Chapter

I. INTRODUCTION: BMP/WNT PATHWAYS SIGNALING IN CANCER AND SOSTDC1 AS A POTENTIAL DUAL-PATHWAY INHIBITOR……………………………………………………………….. 1

II. THE HUMAN BONE MORPHOGENETIC PROTEIN ANTAGONIST

SOSTDC1 IS DOWNREGULATED IN RENAL CANCERS Published in Molecular Biology of the Cell, February 2008…………. 39

III. LOSS OF HETEROZYGOSITY AFFECTING SOSTDC1 IN RENAL

TUMORS For Submission to Cancer Epidemiology Biomarkers and Prevention ……………………………………………………………….. 77

IV. SOSTDC1, AN ANTI-PROLIFERATIVE BMP ANTAGONIST, IS

POORLY EXPRESSED IN PATIENTS WITH ADVANCED BREAST CANCER In Preparation…………………………………………………………….. 106

V. GENERAL DISCUSSION……………………………………………….. 142

SCHOLASTIC VITA……………………………………………………………. 167

vi

LIST OF ABBREVIATIONS

AHR Aryl Hydrocarbon Receptor

AKT Akt/Protein Kinase B Oncogene Family

APC Adenomatous Polyposis Coli

ALK Activin-Like Kinase

AUS Antigen Unmasking Solution

BHD Birt-Hogg-Dubé

BMP Bone Morphogenetic Protein

BMPA BMP Antagonist

BMPR BMP Receptor

CCN Family of proteins including: CTGF, cyr61, and nov

CEU Population of Utah residents with northern or western

European ancestry (HapMap designation)

CK-1,-2 Casein kinase-1, -2

CTCK C-Terminal Cystine Knot

CTGF Connective Tissue Growth Factor

Cyr61 Cysteine rich 61

DAN Differential screening-selected gene Aberrative

in Neuroblastoma

DCIS Ductal Carcinomas In Situ (Breast)

DSH/DVL Disheveled

DKK Dickkopf

vii

EGF/R Epidermal Growth Factor/Receptor

FBS Fetal Bovine Serum

FGF Fibroblast Growth Factor

FH Fumerate Hydratase

FWT Familial Wilms Tumor locus

FRZ Frizzled

FSH Follicle Stimulating Hormone

GAPDH GlycerAldehyde-3-Phosphate DeHydrogenase

GSK-3β Glycogen synthase kinase-3β

HapMap International HapMap Project- partnership to

database of frequencies of human SNPs

hCG human Chorionic Gonadotropin

HMEC Human Mammary Epithelial Cell

HMER HMECs overexpressing ras oncogene

IGF/R Insulin-like Growth Factor/Receptor

IHC ImmunoHistoChemistry

INT MMTV Integration gene (M. musculus)

LCIS Lobular Carcinoma In Situ (Breast)

LDLR Low-Density Lipoprotein Receptor

LH Luteinizing Hormone

LOH Loss-Of-Heterozygosity

LRP LDLR Related Protein

MAD Mothers against decapentaplegic (D. melanogaster)

viii

MAPK Mitogen-Activated Protein Kinase

MEOX2 MEsenchymal homeobOX-2

MMTV Mouse Mammary Tumor Virus

NDP Norrie Disease Protein

NGF Nerve Growth Factor

Nov Nephroblastoma overexpressed

PPAR Peroxisome Proliferator-Activated Receptor

PC Pearson’s Coefficient

PCP Planar Cell Polarity (Wnt signaling pathway)

PI3K PhosphoInositide-3 Kinase

P/S Penicillin/Streptomycin

PTCH Patched

PTEN Phosphatase and TENsin homolog

RPTEC Renal Proximal Tubule Epithelial Cells

RCC Renal Cell Carcinoma

RCC-Clear RCC-Clear Cell Type

rh recombinant, human

R-Smad Regulatory-Smad

RT-CES Real Time-Cell Expansion System® (ACEA

Biosciences)

sFRP secreted Frizzled-Related Protein

SBE Smad Binding Element

SDS Sodium Dodecyl Sulfate

ix

SDS-PAGE SDS-PolyAcrylamide Gel Electrophoresis

Smad Mothers against decapentaplegic and SMA family

Member (Common name for MAD and SMA

homologs)

SNP Single Nucleotide Polymorphism

SOST SclerOSTin

SOSTDC1 SclerOSTin Domain-Containing 1 (H. sapiens)

TCF/LEF T-Cell Factor/Lymphoid Enhancing Factor

TGF-β Transforming Growth Factor-β

TMA Tissue MicroArray

Tsg twisted gastrulation

UEP UnExtended Primer

USAG-1 Uterine Sensitization-Association Gene-1 (R.

Norvegius)

VHL Von Hippl-Lindau

vWF von Willebrand Factor

WAGR Wilms tumor, Aniridia, Genitourinary malformation,

and mental Retardation syndrome

Wg Wingless (D. melanogaster)

WIF Wnt Inhibitory Factor

WISP Wnt Inducible Signaling Pathway

Wnt Wingless-type MMTV integration site family member

(Common name for Wg and INT homologs)

x

WT Wilms Tumor locus

YRU Population of Nigeria residents with Yoruban

ancestors (HapMap designation)

xi

LIST OF ILLUSTRATIONS

CHAPTER I

Figure Page

1. Amino acid composition of SOSTDC1 22

2. CTCK motif in various secreted signaling molecules 25

3. SOSTDC1 alignment with sclerostin 26

CHAPTER II

Figure Page

1. Downregulation of Human SOSTDC1 mRNA in 52 Kidney Tumors

2. Immunohistochemical analysis of SOSTDC1 expression In normal kidney and renal cancer subtypes 55 3. Quantification of SOSTDC1 staining 58 4. Normal human SOSTDC1 is secreted and binds to matrix or cell surface proteins around neighboring cells 60 5. rhSOSTDC1 antagonizes the production of phosphorylated Smad in response to BMP-7 signaling 62

6. SOSTDC1 antagonizes Wnt-3a signaling in renal carcinoma cells 64 7. Overexpression of SOSTDC1 suppresses proliferation of

769-P RCC-clear cell cultures 66

Table Page

I. Quantitation of SOSTDC1 Immunohistochemistry 56

xii

CHAPTER III

Figure Page

1. SOSTDC1 Locus 85

2. Genes in 2 Mbp Region of Interest on 7p 86

3. LOH Results in Wilms and Renal Cell Carcinoma 88

4. OncoMine Database Shows Downregulation of SOSTDC1 In Wilms Tumor 92 Table Page

I. SNPs in Direct Sequencing 89

SI. Primers for Direct Sequencing 96

SII. Primers for LOH Sequenom Analysis 97

CHAPTER IV

Figure Page

1. Immunohistochemistry of SOSTDC1 in Normal and Breast Tumor Tissues 118 2. Secreted SOSTDC1 Decreases in Mammary Cells with Increasing Transformation 121 3. Quantification of SOSTDC1 Message in Breast Cancer

Patients 123 4. SOSTDC1 Protein Levels by Immunohistochemistry for 81 Breast tumors 126 5. SOSTDC1 Inhibits BMP-7-initiated Phosphorylation of Intracellular R-Smads-1,-5,-8 131

6. Cells Overexpressing hSOSTDC1-FLAG Show Proliferation Suppression 133

xiii

Table Page

I. SOSTDC1 Staining of TMAs- Significant Correlations 127

II. Comparisons Between SOSTDC1 Staining Intensity and Other Tumor Markers in the TMA Tissues 129

xiv

ABSTRACT Blish, Kimberly Rose

SOSTDC1: A BMP/WNT DUAL ANTAGONIST IN BREAST AND RENAL CANCERS

Dissertation under the direction of Suzy V. Torti, Ph.D., Associate Professor of Biochemistry

The bone morphogenetic protein (BMP) pathway and the Wnt pathway are

signaling networks associated with carcinogenesis. BMPs and Wnts are powerful morphogens and downstream signaling affects critical cell functions such as proliferation, differentiation, and viability. Proper regulation of these pathways is essential to prevent of transformation in normal cells.

SOSTDC1 (SclerOSTin Domain-Containing-1) is an inhibitor of BMP/Wnt

signaling in mouse disease models; however, little is known about the expression or actions of SOSTDC1 in human tissue. This work expands upon the observation that SOSTDC1 downregulation occurs in 85-90% of breast and kidney tumors. It was hypothesized that extracellular BMP regulation by SOSTDC1 is protective against cancer; therefore, loss of SOSTDC1 may lead to tumorigenesis initiation or progression. This was investigated through studies of the SOSTDC1 gene locus on chromosome 7, examination of the expression and actions of SOSTDC1 in cell culture models, and evaluation of SOSTDC1 within clinical samples.

Human SOSTDC1 protein is expressed in normal kidney and breast tissue.

Mature SOSTDC1 is secreted from normal kidney and breast cells but is not detectable from breast cancer cell lines. The distinction is relevant as SOSTDC1 in the extracellular space effectively antagonizes BMP signaling in breast and kidney cancer cells. Additionally, SOSTDC1 antagonizes Wnt signaling in cell models of kidney cancer. When expression of SOSTDC1 is restored to kidney or breast cancer cell models, striking inhibition of culture proliferation is achieved and cannot be rescued with BMP treatment alone. This suggests that therapeutic potential of exogenous SOSTDC1 may come through concurrent dual BMP and Wnt pathway inhibition.

Genetic studies revealed loss-of-heterozygosity (LOH) at the SOSTDC1 locus

in 10% of renal carcinomas and pediatric Wilms Tumors. Searches for a putative Wilms Tumor suppressor at this locus are ongoing and SOSTDC1 may be a viable candidate. Evaluation of SOSTDC1 protein in clinical renal cell carcinomas and breast tumors reveals a significant loss of protein in renal clear cell carcinomas and breast tumors of advanced stage. Loss of SOSTDC1 may be a biomarker for more advanced breast disease.

1

CHAPTER I

GENERAL INTRODUCTION

The BMP and Wnt Pathways and SOSTDC1 as a Potential Dual-Pathway

Inhibitor

K. BLISH

2

Cancer is a Complex Disease

Cancer incidence is rising, despite the exponential expansion of research into the

etiology, development, and treatment of various neoplasms. This is in part due

to a fundamental problem at the root of cancer research and treatment: the many

diseases collectively known as cancer are incredibly diverse and complex.

The first level of complexity arises from the location of the cancer. Separate

tissues are developed and maintained by many different transcriptomes; thus the

proteins and pathways susceptible to oncogenic stimuli differ from one organ to

another. The next layer of complexity to consider is the variety of causes that

initiates transformation. From heritable genetic mutations in protooncogenes or

tumor suppressor genes, to epigenetic changes, to environmental insults

chemical or viral, the list of causative agents is always expanding. Additionally,

the timing and potential additive effects of multiple exposures play a role. Next

within tissues, there are numerous interactions between cells and within their

surrounding matrix. Tumorigenesis cannot occur without disruptions in cell-cell

communication and many changes to cell-matrix interactions. Finally, there are

incredibly complex interactions within the cells themselves. Carcinogenesis even

within the cell is a non-linear, multi-step process. Despite cellular “programming”

for homeostasis, entire networks of intracellular signaling pathways are

susceptible to many classes of perturbations. In the framework of this vast and

multi-tiered complexity there is great need for models of carcinogenesis that

identify the key processes and players of cancer; models that simplify, provide

3

understanding, and ultimately lead to practical, efficacious, and safe preventions

and treatments.

Cancer Models that Simplify the Disease, Identify Key Pathways, and Lead

to More Effective Treatments

Models of carcinogenesis tend to represent two schools of thought (1): the first

stipulates that fully differentiated cells all contain some genetic predisposition to

cancer (heritable germ line mutations result in familial cancer syndromes;

random somatic mutations to sporadic disease) and then over time, stresses to

the cell cause a number of carcinogenic changes, transforming it into a self-

renewing and invasive entity. It has been proposed that the necessary and

sufficient carcinogenic changes can be categorized into six groups: evasion of

apoptosis, self-propelled growth, insensitivity to anti-growth signals, ability to

cause angiogenesis, unlimited replication potential, and the ability to invade into

normal tissues. This model suggests that between four and seven mutations

affecting these groups are necessary to cause transformation (2). The second

paradigm suggests that within all tissues reside less-differentiated stem cells,

capable of expansion as needed within the demands of tissue homeostasis.

Mutations within these populations of stem cells that prevent them from

differentiating into their proper lineage and instead cause them to proliferate

inappropriately could be responsible for tumor formation (3).

4

Common ground between these two models identifies the importance of certain

core signaling pathways of development: Wnt (wingless/int), hedgehog, notch

secreted growth factors (particularly epidermal growth factor/EGF, platelet-

derived growth factor/PDGF and fibroblast growth factor/FGF) and subsequent

PI3K/PTEN, and TGF-β/BMP. All of these pathways include the actions of

secreted morphogens functioning in both cell autonomous and paracrine

capacities to control cell cycle progression, proliferation, protein synthesis,

viability, and interactions with the surrounding matrix. Research has shown that

these pathways are not only pivotal in the maintenance or differentiation of stem-

cell populations, but also changes in these pathways in adult tissues can induce

imbalances leading to all six previously discussed properties of transformed cells

(4).

Given the importance of these pathways, it is essential that the regulatory

mechanisms governing them remain intact to prevent those transformative

imbalances. Indeed, many of the characterized tumor suppressors to date are

genes that have regulatory functions in one of these pathways (e.g. PTEN in

PI3K signaling which can influence p53 and AKT (5), APC in Wnt pathway (6),

PTCH in hedgehog signaling (7), Smad4 in BMP pathway (8). Naturally, the goal

of studying these regulatory mechanisms is to develop synthetic strategies to

restore proper regulation to cancer cells.

5

Studies of secreted protein antagonists, mutated receptors or monoclonal

antibodies against receptors, and small-molecule inhibitors of these pathways

have demonstrated the possibility of extracellular regulation. Regulation

occurring extracellularly holds therapeutic potential as it bypasses the need to

safely deliver the compounds to the intracellular spaces. Noteworthy examples

of therapies “targeted” directly to extracellular modulation of these pathways

include: receptor tyrosine-kinase inhibitors (EGFR inhibitors gefitinib/Iressa®,

AstraZeneca, Waltham, MA and erlotinib/Tarceva®, Genentech, South San

Francisco, CA) and monoclonal antibodies against receptors (HER2/neu

receptor and trastuzumab/Herceptin®, Genentech).

Building on the success of these targeted anti-neoplastic drugs, there has been

interest in identification of other such extracellular targets. The work presented

here grew out of collaboration between Wake Forest University and Human

Genome Sciences, Inc. (Rockville, MD) to run screens identifying secreted or

extracellular proteins that are differentially expressed in tumors. Identified

proteins associated with any of these known pathways might represent novel

extracellular regulatory mechanisms that could serve as templates for future

therapeutics. One protein identified by this screen is SOSTDC1 (SclerOSTin

Domain-Containing-1), which has associations with the Wnt and BMP (bone

morphogenetic protein) signaling pathways. Accordingly, the rest of this general

review will focus on these particular pathways and their roles in cancer.

6

The Wnt Pathway

Studies in Drosophila melanogaster identified the wingless (Wg) gene,

responsible for body patterning during fruit fly embryogenesis. The murine

ortholog, Int-1 lies near the integration site for mouse mammary tumor virus

(MMTV) and may be overactivated by the viral oncogene. After the proteins

were found to be homologous via protein comparison, the combined name of

Wnt was used to describe the members of this family of secreted growth,

differentiating, and morphogenetic proteins. There are currently 19 known

members of the vertebrate Wnt family and these proteins are able to exert their

effects through several mechanisms; including both classical and non-classical

signaling pathways (9). Traditionally, Wnt ligands are grouped into “canonical”

ligands that tend to active the classical β-catenin pathway (10) (most ubiquitous

are Wnts -1, -3, and -8) and “non-canonical” or “β-catenin-independent Wnts”

(Wnt-4, Wnt-5a, 5b, and Wnt 11) that tend to activate one of the non-classical

signaling pathways. Less is known about the non-canonical pathways as details

are still to be elucidated, but they involve one of the following mechanisms:

intracellular Ca2+ modulation , changes in cell polarity (sometimes referred to as

the planar cell polarity, PCP pathway , and co-activity of protein kinase A)

(11;12).

All Wnt family members contain an N-terminal secretion signal sequence, at least

one N-liked glycosylation site, upwards of 20 conserved cysteine residues, which

are spaced similarly in all family members and may be sites of palmitoylation.

7

Along with a presumed role in protein folding, the cysteine residues often play

crucial roles in the amount of post-translational modification, and subsequent

activity of the proteins. For example, two cysteine residues in Wnt-1 are

necessary for its ability to transform C57MG mammary epithelial cells (13). Most

of what is known about structure-function relationship in the Wnt family comes

from studies of non-functional mutants or partial-protein chimeras, as Wnt

proteins are notoriously insoluble and full-molecule crystal structures have not

been obtained. It is thought that the C-terminal region of Wnts is responsible for

signaling activity, while the N-terminal end provides the specificity for cognate

receptors (14).

Wnt proteins play a role in a variety of normal and disease processes (15);

indeed, many normal developmental roles of Wnts are mimicked in cancer

pathologies (16). In organogenesis, Wnts often cause differentiation of

mesenchymal stem cells (17); Wnt 4 expressed from the uteric bud is sufficient to

induce metanephric blastema into early kidney tubules (18). Cancer stem cell

self-renewal is controlled by aberrant Wnt signaling (19), and it has been shown

that a gene commonly mutated in renal cancers, VHL, is under the control of Wnt

signaling (20).

As research into Wnt ligands expands the knowledge base, the diversity and

complexity of Wnt signaling can be appreciated. Not only are there a rather large

number of Wnt ligands, there are also multiple receptors, co-receptors, and

8

intracellular signaling machinery that transmit the signals. A brief review of the

main signaling pathway follows here, but many variations on this basic scheme

have been reported (21).

Classical Wnt Signaling

In the absence of active Wnt ligand in the extracellular space, intracellular β-

catenin expression and localization is tightly controlled. In this state, β-catenin is

found either bound to cell surface cadherins or in a complex of proteins

responsible for the ubiquitination and subsequent degradation of β-catenin. The

degradation complex includes: axin, APC(Adenomatous polyposis coli), GSK-3β

(glycogen synthase kinase-3β), and sometimes CK-1 (casein kinase-1). The

exact role of each protein in the complex is unknown and provides material for

ongoing research. The key steps seems to be that in the presence of axin as a

scaffolding protein and APC, GSK-3β and sometimes CK-1 play the pivotal role

of “marking” β-catenin for ubiquitination by phosphorylation of critical N-terminal

serine/threonine residues (22;23).

Wnt ligands function by signaling through receptors of the frizzled (Fzd) family.

Fzd proteins are a subfamily of 7 transmembrane, G coupled-protein receptors.

There are currently 8 known members of the mammalian Fzd family and studies

thus far show differing ligand specificity amongst the Fzds. Importantly, the

Wnt/Fzd interaction is facilitated and stabilized by LRP (low-density lipoprotein

9

related-peptides) -4, -5,or -6 co-receptors, presumably forming a ternary complex

(physical association of the protein has not been proven, yet cells expressing

LRP mutants cannot transmit Wnt signals) (24). Current models suggest that

downstream Wnt signaling activity (such as whether β-catenin independent

pathways are activated) is specified by the Wnt/Fzd/LRP combination and not the

Wnt ligand alone (25).

Upon Wnt binding to a complex of frizzled with LRP co-receptors, the second-

messenger disheveled (DSH- xenopus or DVL- vertebrate) is phosphorylated.

Activated DVL causes stabilization on the intracellular face of the cell membrane

of the multi-protein complex involving DVL, Frizzled, and co-receptors. Most

importantly, activation of DSH/DVL is known to destabilize the interactions of the

APC/Axin and GSK-3β in the degradation system (26). The subsequent

dissolution of the degradation system leads to a rise in intracellular stable (non-

phosphorylated) β-catenin levels. Stable β-catenin is a viable signaling

molecule, capable of translocation to the nucleus with other signaling proteins,

and transcription of a multitude of downstream target genes through the

TCF/LEF promoter. Well-characterized target genes include: c-myc, cyclin D1,

id2, PPARd, among others (27). The downstream effects of active intracellular β-

catenin are profound: including changes in the cell cycle and the

viability/apoptosis axis. The tight regulation of β-catenin through this pathway is

a very basic cell function and is highly conserved throughout evolution. Many

10

additional molecules contribute to this regulation and are still being identified

(28).

Wnt Regulatory Molecules

There are four known types of extracellular Wnt antagonists: WIF1 (Wnt

inhibitory factor 1) , secreted Frizzled-related protein (SFRP) family (29), DKK

(dickkopf) family (30), and multi-pathway Wnt antagonists such as cerberus and

sclerostin. Each family is hypothesized to inhibit Wnt signal transduction by

some manner of disruption of the Wnt/Fzl receptor signaling complex; however,

each set of proteins has its own binding sites and mechanism of inhibition.

Current data demonstrates that WIF1, cerberus, and SFRPs bind primarily to the

Wnt ligands themselves resulting in either sequestration of the ligand away from

the receptors or disruption of the ligand-receptor interaction. DKK proteins and

sclerostin do not have high affinities for Wnt ligands, but instead bind to the co-

receptors LRP-5 and LRP-6 (12). DKKs cross link LRPs away from the Fzds and

to another type of cell-membrane proteins called Kremins (31), while sclerostin is

thought to bind to the LRP is such a way as to prevent its association with Fzd

(32). Other β-catenin pathway regulatory molecules exist including: Intracellular

antagonists such as WTX- a new tumor suppressor in Wilms Tumor (pediatric

kidney lesion), which has been cloned on the X-chromosome and has been show

to also increase the degradation of β-catenin (33) or extracellular non-Wnt

11

ligands that activate β-catenin: such as Norrie disease protein (NDP, (34), and R-

spondins .

Wnt Pathway Proteins and Antagonists in Cancer

β-catenin

The earliest connection of the Drosophila wingless genes with their vertebrate

orthologs (Wnts) came through the study of mouse mammary tumor virus

(MMTV). From this time, it has been well-established that the viral activation of

Wnts can lead to oncogenic transformation. Much effort has gone into

understanding the mechanisms the Wnt pathways and downstream effects in

tumorigenesis- branching out of the mouse model of breast cancer to almost

every type of neoplasm in all organ systems. Overexpression of active, nuclear

β-catenin has been documented in between 20-33% of various tumors (35).

However, the first connection between Wnt signaling and human cancer occurred

with the discovery of the interactions between β-catenin and APC and the

development of colorectal cancers. In intestinal epithelium, β-catenin is a proto-

oncogene and APC as a member of the β-catenin degradation complex is a

tumor suppressor.

Many types of mutations occur to activate the β-catenin proto-oncogene,

including changes in the regulatory site serine/threonine residues that render the

mutant β-catenin less susceptible to phosphorylation and subsequent

ubiquitination (36). Additionally, β-catenin has Wnt-independent roles that are

12

oncogenic when perturbed. In normal cells, β-catenin plays a role with surface

cadherins in the establishment of the adherens junctions between epithelial cells

lining organs (37). Surprisingly, mutant β-catenin is not only likely to stimulate

Wnt-like signaling responses, but it has also been shown to bind less well to

cadherins in the adherens junctions (38). The relationship between the pool of β-

catenin associated with cadherins and the pool available for nuclear signaling is

under investigation.

APC/Axin:

Along with β-catenin, many color cancers experience changes in the expression

or function of APC. Familial adenomatous polyposis (FAP) is a hereditary

condition resulting in numerous polyp formations in the large intestine and early-

onset colo-rectal cancer. Positional cloning assigned the responsible gene, APC,

on the long arm of chromosome 5 . Now, APC is a known tumor suppressor

gene (39) and the common inactivating mutations have been characterized as

ones that prematurely terminate the protein, abrogating its negative regulatory

effects on β-catenin (40). In the assembly of the β-catenin degradation complex,

the amount of axin together with APC appears to dictate the amount of

degradation complexes available. When axin is lost, spurious β-catenin signaling

can result. Thus, axin has tumor suppressive functions. Furthermore, axin has

additional roles regulating p53 and myc complexes (41;42).

13

Others:

In addition to these widely studied mechanisms, over-activity of β-catenin can be

initiated by a change in the extracellular signaling environment with the

overexpression of Wnts and/or receptors (43;44) or loss of natural antagonists

(45). In contrast, Wnt5a shows tumor suppressive capabilities (46;47). The

many mutations and dysregulation of Wnts, Fzd receptors, and antagonists

identified vary greatly from one tumor to another. After the past 20 years of

research concerning the Wnt pathway and carcinogenesis, two things are clear:

this is a key pathway with the potential to cause cancer if misappropriated and

ongoing study is needed to identify which proteins have the most critical roles

and which might be good targets for future therapies.

The Bone Morphogenetic Protein Pathway

Like the Wnt pathway, the BMP pathway is an essential pathway of development

and studies concerning it have been ongoing for the past couple of decades.

The BMPs were initially described for their ability to induce ectopic bone

formation in cartilage (48). Since then, BMPs have been found to be powerful

whole-body morphogens. The BMP family is part of the TGF-β superfamily and

contains over 20 cysteine-rich molecules extracellular signaling molecules. An

important structural fold in the cysteine rich regions is the C-terminal cystine-knot

domain (CTCK) and is found in all BMPs. BMP ligands are expressed in diverse

tissues and exert pleiotropic effects including cellular differentiation, adhesion,

viability, extracellular matrix formation, angiogenesis, and proliferation (49-51).

14

BMP Signaling

The BMP signaling pathway bears many similarities to TGF-β signaling and is

initiated when a BMP homodimer ligand binds a hetero-tetrameric receptor

complex that consists of separate type I and type II moieties (52). Specific

receptors in the BMP pathway include for type I: (activin-like kinase) ALK-3

(BMPR-IA), ALK-6 (BMPR-IB), and ALK-2 and for type II: BMP receptor-II

(BMPR-II) and ActII-R. Various proteoglycan co-receptors will stabilize this

receptor upon ligand binding. If uninhibited binding of the BMP to its receptor

tetramer occurs, the constitutively active type II receptors phosphorylate the C-

terminal serine/threonine kinase on the type I receptors. Next, the type I

receptors recruit an appropriate receptor-regulated Smad (R-Smad) and this R-

Smad will be phosphorylated and activated. R-Smads pertinent to this pathway

include Smad-1, Smad-5, and Smad-8. Activated R-Smads have a greater

affinity for other Smads.

Depending upon the availability of other Smads in the cells, either 1) the main

common partner Smad (Co-Smad), Smad-4 or -4a will associate with the R-

Smad causing further multimerization with appropriate transcription factors, co-

activators or co-repressors molecules, and DNA binding molecules. This new

multimer forms a transcription apparatus which translocates into the nucleus

through nuclear pore complexes. In the nucleus, these transcription complexes

15

can bind to a large variety of genes that regulate cell cycle and growth; or

dephosphorylation of the activated R-Smad can occur, and an inhibitory Smad (I-

Smad-6, -7) can bind to and inhibit the R-Smad from further interactions (53). If

transcription complexes are formed and enter the nucleus unimpeded,

expression of genes containing Smad Binding Elements (SBEs) in their

promoters can be altered (54).

It is important to note that signaling overlap between BMP and TGF-β signaling

does not just occur by the shared use of Smad-4. Connections between these

analogous pathways are constantly being identified, with many occurring at the

level of receptor cross-talk. Thus far, many of the activin-like (ALK) receptors

have been classified as BMP-binding or TGF-β binding, but there are an

increasing number of examples where BMP-ligands activate TGF-β receptors or

co-receptors and vice versa. This is an important source of pathway regulation

and can be confounding when it comes to understanding the contributions of

each pathway in carcinogenesis (55).

BMP Antagonists

It has been proposed that BMPs are released as free signals with paracrine and

autocrine effects on cells, depending upon the microenvironment into which they

are released. BMPs and their receptors tend to be rather ubiquitously expressed

in many tissues, so fine-tuning of the pathway must come from BMPs

16

extracellular interactions with other molecules. BMPs are fairly “sticky”

molecules and can bind heparin sulfates and other proteoglycans.

Concentrations gradients are established based on the availability of such

proteoglycans in the extracellular matrix, free BMP receptor presence on nearby

cells, and whether antagonists have been released in the area. The presence of

antagonists (BMPAs) in the local environment can affect the ability of BMPs

reach the receptor or have an available binding site, thus the antagonists act as

local “on/off” switches (56).

There are a variety of molecules with regulatory or inhibitory effects on BMP

pathway signaling including both intra- and extracellular varieties. A brief review

follows, with particular focus on the extracellular inhibitors as this group presents

the most feasible group of therapeutic targets or models for mimetics.

Intracellular Regulatory Molecules

Smads-6 and -7 function as regulatory Smads and their expression is regulated

by other Smad signaling complexes as a feedback mechanism (57). Various

reports exist on the mechanisms of Smad-6 and -7, whether its binding of

phosphorylated Smad complexes and targeting them for degradation with the

help of E3 ubiquitin ligases SMURF1 and SMURF2 or inactivation of the ALK

receptors by dephosphorylation (58;59). Other intracellular antagonists exist at

17

the transcriptional levels (such as cSki, SnoN) or kinase-dead receptors (BAMBI

(60;61).

Extracellular Antagonists

All known extracellular BMPAs contain structural similarities to the BMPs

themselves, mostly within the CTCK domain. Among the secreted antagonists,

differences in function appear to be due in part to differences in the

arrangements within the CTCK. All extracellular BMPAs are thought to bind to

BMPs directly.

The noggin and chordin family

Members of the noggin family contain a variation of the cystine knot signature

with a 9- , 10-, or 12-membered ring (see below for discussion on the CTCK).

Noggin has binding preferences for BMPs-2, and -4 and is the only BMPA to

have a solved crystal structure. It has been show that homodimers of noggin

“clip” onto BMP dimers and block receptor-binding epitopes on both parts of the

BMP dimer (62).

Chordin is a larger protein than many of the BMPs and BMPAs (120 kDa as

opposed to an average size of 35 kDa) and is a functional antagonist of BMPs

when bound to BMPs alone (63). However, chordin can partner with a unique

BMPA, twisted gastrulation (tsg) that can interestingly serve as both a BMP

18

agonist and BMP antagonist. Tsg’s antagonist actions occur in the stabilization

of the BMP-chordin interaction, guaranteeing a more successful sequestration of

BMP ligands from the receptor. Ironically, if BMP-1/Tolloid is present, the BMP-

chordin/tsg ternary complex is more susceptible than chordin/BMP alone to

cleavage by tolloid, releasing free BMP. Therefore, in the presence of tolloid,

chordin, and BMP, tsg can behave as a BMP agonist (64).

DAN Family

The DAN family, contains the BMPAs known as DAN (Differential screening-

selected gene aberrative in neuroblastoma (65)), PRDC (Protein Related to Dan

and Cerberus (66)), gremlin (67), and the dual-Wnt-pathway antagonists:

Cerberus/coco, sclerostin, and SOSTDC1. All contain 8-membered rings in the

CTCK as the BMPs themselves do; however, outside of the CTCK the members

do not display great amounts of homology. Functionally, all inhibit BMP

signaling, particularly BMP-2 and BMP-4. The dual-pathway antagonists play

critical roles in development and are remarkable for the variety of manners in

which the dual-pathway specificity is accomplished.

Roles of BMPs and Antagonists in Cancer

Smad4

Smad-4 was originally identified as DPC4 (Deleted in Pancreatic Cancer) and

mutations in Smad-4 are indeed associated with pancreatic and colon cancers

(68). Heritable germ line mutations of Smad4 cause familial juvenile polyposis

19

syndrome (69), showing that Smad4 has tumor suppressive properties in several

disease types. Based on these data, one might hypothesize that TGF-β/BMP

signaling resulting in Smad4 activation is protective/growth-inhibitory. However,

the roles of BMPs are quite diverse in cancers.

BMP downstream targets include multiple cell cycle check point proteins and

proto-oncogenes, including c-myc (53) and p53/p21. Alterations in these

pathways have been observed when expression of BMP antagonists is increased

(70-72). In some tumors, BMP ligand expression is decreased, in others there is

overexpression (73). In some systems, overexpression of BMPAs is beneficial

(74), in others it seems to accompany a more transformed phenotype (75). At

this point the picture is quite complex, but there is a growing body of evidence

that the BMPs and their antagonists may be pivotal in our understanding of

cancer.

As mentioned above, array experiments in collaboration with Human Genome

revealed a novel gene product that is downregulated in breast and kidney cancer

tissues compared to adjacent normal tissue, SOSTDC1. This protein is a

putative dual-regulator of both BMP and Wnt signaling. A brief discussion of the

other known dual-pathway antagonists will precede more information about

SOSTDC1.

20

Dual BMP/Wnt Pathway Antagonists

Cerberus/coco

Cerberus acts in Xenopus embryos as a tri-pathway inhibitor with separate

binding sites for Nodal, Wnts, and BMPs (63). Identification of the rodent or

mammalian orthologs for cerberus has been difficult. One candidate is coco,

whose expression is induced by Smad-7 and shows BMP/Wnt pathway

modulation (76). Much remains to be determined regarding these proteins.

Sclerostin

Sclerostin was first identified as the genetic link to the rare bone overgrowth

disorder known as sclerosteosis, which is a continued deposition of new bone

due to increased activity of osteoblasts (77;78). It was determined that families

suffering from sclerosteosis showed mutations in the SOST gene resulting in loss

of function mutations. Additionally, deletion of SOST regulatory elements causes

a related bone dysplasia known as van Buchem disease (79). By either

mechanism, the resulting lack of active sclerostin being secreted from osteocytes

causes a loss of inhibitory paracrine signal to the osteoblasts (80;81).

Originally, sclerostin was thought to be a non-classical antagonist of the BMP

pathway as it has sequence similar to members of the DAN subfamily of BMP

antagonists (82). Some groups reported that Sclerostin binds to BMPs,

particularly BMP-6 and -7, and inhibits Smad phosphorylation (83-85). However,

further reports showed sclerostin was unable to inhibit ALP production (alkaline

21

phosphatase, a marker of osteocyte activity induced by BMP signaling) as most

BMP antagonists do (80). Thus, sclerostin was hypothesized to be a non-

classical BMP antagonist. The basis for the diverse findings has been under

scrutiny. Additionally, sclerostin was shown to antagonize the workings of noggin

when the two are co-expressed in a novel method of pathway regulation (86).

More recent work has shown that sclerostin interferes with both Wnt and BMP

signaling, in a tissue-context dependent manner (32;87;88). As mentioned

above, sclerostin binds LRP 5 and 6, known co-receptors in the Wnt/Fzd receptor

complex. Sclerostin is also unique from other members of the DAN family of

BMPAs in that it is missing an additional Cys residue that is thought to play a role

in multimerization of the BMPAs (Figure 2). Current data show that sclerostin is

secreted and active as a monomer.

SOSTDC1- A “novel” dual-pathway inhibitor

The human SOSTDC1 gene is located on the short arm of chromosome 7. The

known protein product of this locus is encoded over 2 exons and is 206 amino

acids in length. Analysis of the protein’s primary structure revealed a high usage

of cysteine residues (in the top 5% of all known proteins). SOSTDC1 protein is

positively charged (Arg and Lys residues outnumber Glu and Asp by a differential

of 18-Figure 1) and the computed pI of the entire chain is 9.9 (Proteomics and

Sequencing Tools, Primary Structure Analysis, ExPASy, www.expasy.ch). This

protein was predicted to be secreted from the cell with a high probability that the

first 23 residues comprised a secretory signal peptide (SignalP prediction server-

22

ExPASy, www.expasy.ch). Furthermore, the many cysteine residues within the

protein are arranged in a c-terminal cystine knot pattern (CTCK domain).

Figure 1. Amino acid profile of the 206 amino acid SOSTDC1 propeptide.

Secondary structure is likely influenced by the high incidence of cysteine

(10 residues) and proline (12 residues). The protein will have an overall

positive charge due to the incidence of basic residues (17 arginine and 14

lysine) compared to acidic (5 aspartate and 8 glutamate).

Acidic6%

Hydrophilic21%

Hydrophobic17%

Proline6%

Other30%

Cystine5%

Arginine8%

Lysine7%

Acidic6%

Hydrophilic21%

Hydrophobic17%

Proline6%

Other30%

Cystine5%

Arginine8%

Lysine7%

23

The CTCK domain motif is comprised of at least 6 cysteine residues that form 3

disulfide bonds in such a manner where a two dimensional “ring” formed by

cysteines #2,3,5, and 6 in the primary motif signature is crossed in a third plane

by the disulfide bond between cysteines 1 and 4. The result is a “knot” at the

center of the domain. This arrangement forces amino acids interspersed

between the cysteines into three loops or “fingers” from the knot. In crystallized

proteins containing the CTCK (noggin, PDGF, and BMP7), these loops often

assume short β-sheets. In many cases, the cystine knot is a rigid core which

initiates folding of the more flexible loops around it.

In addition to the role in protein folding, the CTCK domain is thought to play a

large role in protein-protein interactions for two reasons: first, hydrophobic

residues that might otherwise be buried in the protein core can be exposed by

the knot structure and secondly, the CTCK domain usually occurs in a very

cysteine-rich area, so other cysteine residues are available for binding between

monomers; many cysteine-knot proteins are active as homo- or heterodimers.

The superfamily of cystine-knot proteins contains a wide array of secreted factors

with essential roles in development, growth, and differentiation. The CTCK

domain is not seen in yeast, but is contained in proteins in C. elegans and

Drosophila, suggesting the roles of CTCK-containing proteins are geared

towards intercellular communication. There are three types of cystine-knot

motifs: the growth factor knot, the inhibitor knot seen in many types of naturally

24

occurring toxins, and the cyclic knot found in cyclotides (89). All motifs are based

around six core cysteine residues, however only the growth factor knot has the

bond between cysteine #1 and cysteine #4 as the transverse bond (90). The

family containing the growth-factor motif is the largest group of proteins: including

glycoprotein hormones (such as luteinizing hormone, human chorionic

gonadotropin, and follicle stimulating hormone), the TGF-β superfamily (including

TGF-β, activins, and BMPs), the PDGF family, von Willebrand factor (vWF),

nerve growth factor (NGF), norrie disease factor (NDF), the CCN family, mucins,

and slit-like homologues.

Among the cysteine knot superfamily, homology searches and alignments

(Figure 2) identified SOSTDC1 as bearing the most similarity to the CCN family

(28%) and the BMP antagonist (BMPA) family, particularly sclerostin (54%-Figure

3). The highly similar region between sclerostin and SOSTDC1 (SclerOSTin

Domain-Containing 1) is now being referred to as a “sclerostin domain”.

Recently, a solved NMR crystal structure of sclerostin defined probable LRP

binding sites, showed a previously un-indentified heparin-binding domain, and

identified greater flexibility than originally anticipated in the cystine knot structure

(91).

25

Figure 2. Alignment of 8-membered ring cystine knot proteins. * = protein

with NMR crystal structure known. Shaded boxes show the 8 residues in

the ring of the cystine knot. Di-sulfide bonds exist between Cys1 and Cys4,

Cys2 and Cys 5, and Cys3 and Cys6. Arrow points to extra cysteine

residue present in many family members, thought to potentiate

dimerization. This residue is missing in sclerostin and SOSTDC1.

26

Figure 3. Alignment of SOST (sclerostin) and SOSTDC1 protein sequences.

Orange box shows largest area of 100% identity. Red lines show end of

signal peptide. Yellow circles highlight the 6 conserved cysteines and

green box shows the entirety of the cystine knot (CTCK) domain. Overall

homology between the two is 54% with homology in the CTCK at 64%

27

From these structural similarities, it can be inferred that SOSTDC1 may have

common functional activities to sclerostin. Additionally, extended homology

searches to the proteomes of other organisms identifies the following orthologs:

ectodin (Mus musculus, 95% homology); USAG-1 (Uterine Sensitization

Associated Gene, Rattus norvegicus, 94%), wise (Xenopus laevis 78%), and

zgc:110293 (Danio rerio, 72%).

Known Functions of SOSTDC1 homologs in Animal Models

Wise (Xenopus)

Wise was first identified as both an inhibitor and an activator of Wnt signaling that

interacted with Wnt ligand and was able to bind to LRP6, causing some Wnt-

independent signaling (92). Later, a mechanism for both BMP and Wnt binding

was observed (93). It has lately been shown that another mechanism for the

dual agonist/antagonist pathway is the cellular localization of Wise as it is able to

bind LRP6 in the endoplasmic reticulum, sequestering it away from Wnt pathway

signaling (94). This represents a novel Wnt regulatory paradigm.

Ectodin/USAG-1/SOSTDC1 (Murinae)

In rodents, USAG-1 was first identified as a circulating factor involved in the pre-

implantation changes of the uterine epithelium. Shortly thereafter, USAG-1 was

shown to be highly expressed in the adult kidney (95) while ectodin was noticed

in the developing tooth buds of embryos (96) and was able to affect multiple

pathways, including BMP and Wnt signaling simultaneously. Knock-out models

28

have confirmed the importance of ectodin in these tissues as these mice exhibit

supernumerary teeth (97;98) and greater resistance to healing after kidney injury

(through antagonism of the pro-healing molecule BMP-7) (99). Little is known

about the effects of SOSTDC1 in normal human homeostasis and disease.

There exist no published data concerning SOSTDC1 in cancer. Collaborative

data with Human Genome Sciences for this project showed that there was less

SOSTDC1 in tumors compared to normal tissues (particularly in breast and

kidney), so it is logical to hypothesize that it might have tumor suppressor roles

supporting normal homeostasis in adult tissues. While diverse in their etiologies

and properties, breast and kidney cancers remain difficult to detect at an early

stage. This is crucial as both have the propensity for devastating metastasis as

the disease progresses. Thus, the identification of secreted protein products that

may be useful in earlier diagnosis or treatment, such as SOSTDC1, is a crucial

area for further exploration.

Thesis Goal and Hypothesis

This work examines facets of the BMP and Wnt pathways within the framework

of two diseases: renal and breast cancer. The hypothesis is that one molecule,

SOSTDC1, has the ability to modulate both of these pathways in breast and

kidney tissue. Furthermore, this ability of SOSTDC1 prevents improper

activation of either (or both) pathways. Thus, the end result of SOSTDC1’s

actions is to suppress the transformation of normal cells through the spectrum of

29

carcinogenic changes (i.e. hyperproliferation, evasion of death, invasiveness,

and metastatic potential).

The hypothesis is addressed with a three-pronged approach. 1) The expression

of SOSTDC1 in kidney and breast normal and tumor tissues is evaluated as an

early step to explore the potential of SOSTDC1’s use as an extracellular marker

or therapeutic target of cancer. 2) Characterization of the SOSTDC1 locus and

possible dysregulation is addressed through studies of the SOSTDC1 gene and

subsequent mRNA expression in kidney and breast cancers. 3) Cell culture

models of SOSTDC1 are developed and the phenotypes of SOSTDC1

manipulation in vitro will be assessed to start elucidating SOSTDC1’s functional

capabilities.

Through these studies, the goal of this work is to address the consequences of

SOSTDC1 downregulation in cancer progression, assess the possible affects of

the downregulation of Wnt/BMP signaling, and determine whether SOSTDC1

might be a useful marker or therapeutic target in breast and/or renal lesions.

30

Reference List

(1) Sherr CJ. Principles of Tumor Suppression. Cell 2004;116:235-46.

(2) Hanahan D, Weinberg RA. The hallmarks of cancer. Cell 2000;100(1):57-70.

(3) Reya T, Morrison SJ, Clarke MF, Weissman IL. Stem cells, cancer, and cancer stem cells. Nature 2001 Nov 1;414(6859):105-11.

(4) Katoh M. Networking of WNT, FGF, Notch, BMP, and Hedgehog signaling pathways during carcinogenesis. Stem Cell Rev 2007 Jan;3(1):30-8.

(5) Yin Y, Shen WH. PTEN: a new guardian of the genome. Oncogene 2008 Sep 18;27(41):5443-53.

(6) Aoki K, Taketo MM. Adenomatous polyposis coli (APC): a multi-functional tumor suppressor gene. J Cell Sci 2007 Oct 1;120(Pt 19):3327-35.

(7) Lindstrom E, Shimokawa T, Toftgard R, Zaphiropoulos PG. PTCH mutations: distribution and analyses. Hum Mutat 2006 Mar;27(3):215-9.

(8) Miyaki M, Kuroki T. Role of Smad4 (DPC4) inactivation in human cancer. Biochem Biophys Res Commun 2003 Jul 11;306(4):799-804.

(9) Miller JR. The Wnts. Genome Biol 2002;3(1):REVIEWS3001.

(10) Barker N. The canonical Wnt/beta-catenin signalling pathway. Methods Mol Biol 2008;468:5-15.

(11) Kestler HA, Kuhl M. From individual Wnt pathways towards a Wnt signalling network. Philos Trans R Soc Lond B Biol Sci 2008 Apr 12;363(1495):1333-47.

(12) Chien AJ, Conrad WH, Moon RT. A Wnt Survival Guide: From Flies to Human Disease. J Invest Dermatol 2009 Jan 29.

(13) Mason JO, Kitajewski J, Varmus HE. Mutational analysis of mouse Wnt-1 identifies two temperature-sensitive alleles and attributes of Wnt-1 protein essential for transformation of a mammary cell line. Mol Biol Cell 1992 May;3(5):521-33.

(14) Hoppler S, Brown JD, Moon RT. Expression of a dominant-negative Wnt blocks induction of MyoD in Xenopus embryos. Genes Dev 1996 Nov 1;10(21):2805-17.

31

(15) Clevers H. Wnt/beta-catenin signaling in development and disease. Cell 2006 Nov 3;127(3):469-80.

(16) Chen X, Yang J, Evans PM, Liu C. Wnt signaling: the good and the bad. Acta Biochim Biophys Sin (Shanghai) 2008 Jul;40(7):577-94.

(17) Ling L, Nurcombe V, Cool SM. Wnt signaling controls the fate of mesenchymal stem cells. Gene 2009 Mar 15;433(1-2):1-7.

(18) Kispert A, Vainio S, McMahon AP. Wnt-4 is a mesenchymal signal for epithelial transformation of metanephric mesenchyme in the developing kidney. Development 1998 Nov;125(21):4225-34.

(19) Malanchi I, Huelsken J. Cancer stem cells: never Wnt away from the niche. Curr Opin Oncol 2009 Jan;21(1):41-6.

(20) Guillen-Ahlers H. Wnt signaling in renal cancer. Curr Drug Targets 2008 Jul;9(7):591-600.

(21) Jin T, George F, I, Sun J. Wnt and beyond Wnt: multiple mechanisms control the transcriptional property of beta-catenin. Cell Signal 2008 Oct;20(10):1697-704.

(22) van NM, Meeldijk J, van der ZR, Destree O, Clevers H. Wnt signaling controls the phosphorylation status of beta-catenin. J Biol Chem 2002 May 17;277(20):17901-5.

(23) Yost C, Farr GH, III, Pierce SB, Ferkey DM, Chen MM, Kimelman D. GBP, an inhibitor of GSK-3, is implicated in Xenopus development and oncogenesis. Cell 1998 Jun 12;93(6):1031-41.

(24) Tamai K, Semenov M, Kato Y, Spokony R, Liu C, Katsuyama Y, et al. LDL-receptor-related proteins in Wnt signal transduction. Nature 2000 Sep 28;407(6803):530-5.

(25) Holmen SL, Robertson SA, Zylstra CR, Williams BO. Wnt-independent activation of beta-catenin mediated by a Dkk1-Fz5 fusion protein. Biochem Biophys Res Commun 2005 Mar 11;328(2):533-9.

(26) Angers S, Thorpe CJ, Biechele TL, Goldenberg SJ, Zheng N, Maccoss MJ, et al. The KLHL12-Cullin-3 ubiquitin ligase negatively regulates the Wnt-beta-catenin pathway by targeting Dishevelled for degradation. Nat Cell Biol 2006 Apr;8(4):348-57.

(27) Giles RH, Lolkema MP, Snijckers CM, Belderbos M, van der GP, Mans DA, et al. Interplay between VHL/HIF1alpha and Wnt/beta-catenin pathways during colorectal tumorigenesis. Oncogene 2006 May 18;25(21):3065-70.

32

(28) Angers S. Proteomic analyses of protein complexes in the Wnt pathway. Methods Mol Biol 2008;468:223-30.

(29) Bovolenta P, Esteve P, Ruiz JM, Cisneros E, Lopez-Rios J. Beyond Wnt inhibition: new functions of secreted Frizzled-related proteins in development and disease. J Cell Sci 2008 Mar 15;121(Pt 6):737-46.

(30) Pinzone JJ, Hall BM, Thudi NK, Vonau M, Qiang YW, Rosol TJ, et al. The role of Dickkopf-1 in bone development, homeostasis, and disease. Blood 2009 Jan 15;113(3):517-25.

(31) Mao B, Niehrs C. Kremen2 modulates Dickkopf2 activity during Wnt/LRP6 signaling. Gene 2003 Jan 2;302(1-2):179-83.

(32) Semenov M, Tamai K, He X. SOST is a ligand for LRP5/LRP6 and a Wnt signaling inhibitor. J Biol Chem 2005 Jul 22;280(29):26770-5.

(33) Major MB, Camp ND, Berndt JD, Yi X, Goldenberg SJ, Hubbert C, et al. Wilms tumor suppressor WTX negatively regulates WNT/beta-catenin signaling. Science 2007 May 18;316(5827):1043-6.

(34) Hendrickx M, Leyns L. Non-conventional Frizzled ligands and Wnt receptors. Dev Growth Differ 2008 May;50(4):229-43.

(35) Ilyas M. Wnt signalling and the mechanistic basis of tumour development. J Pathol 2005 Jan;205(2):130-44.

(36) Morin PJ, Sparks AB, Korinek V, Barker N, Clevers H, Vogelstein B, et al. Activation of beta-catenin-Tcf signaling in colon cancer by mutations in beta-catenin or APC. Science 1997 Mar 21;275(5307):1787-90.

(37) McCrea PD, Turck CW, Gumbiner B. A homolog of the armadillo protein in Drosophila (plakoglobin) associated with E-cadherin. Science 1991 Nov 29;254(5036):1359-61.

(38) Chan TA, Wang Z, Dang LH, Vogelstein B, Kinzler KW. Targeted inactivation of CTNNB1 reveals unexpected effects of beta-catenin mutation. Proc Natl Acad Sci U S A 2002 Jun 11;99(12):8265-70.

(39) Morin PJ, Weeraratna AT. The APC tumor suppressor pathway. Methods Mol Biol 2003;222:21-40.

(40) Aoki K, Taketo MM. Adenomatous polyposis coli (APC): a multi-functional tumor suppressor gene. J Cell Sci 2007 Oct 1;120(Pt 19):3327-35.

33

(41) Arnold HK, Zhang X, Daniel CJ, Tibbitts D, Escamilla-Powers J, Farrell A, et al. The Axin1 scaffold protein promotes formation of a degradation complex for c-Myc. EMBO J 2009 Mar 4;28(5):500-12.

(42) Rui Y, Xu Z, Lin S, Li Q, Rui H, Luo W, et al. Axin stimulates p53 functions by activation of HIPK2 kinase through multimeric complex formation. EMBO J 2004 Nov 24;23(23):4583-94.

(43) Polakis P. The many ways of Wnt in cancer. Curr Opin Genet Dev 2007 Feb;17(1):45-51.

(44) Morin PJ, Weeraratna AT. Wnt signaling in human cancer. Cancer Treat Res 2003;115:169-87.

(45) Suzuki H, Toyota M, Carraway H, Gabrielson E, Ohmura T, Fujikane T, et al. Frequent epigenetic inactivation of Wnt antagonist genes in breast cancer. Br J Cancer 2008 Mar 25;98(6):1147-56.

(46) Pukrop T, Binder C. The complex pathways of Wnt 5a in cancer progression. J Mol Med 2008 Mar;86(3):259-66.

(47) Ying J, Li H, Yu J, Ng KM, Poon FF, Wong SC, et al. WNT5A exhibits tumor-suppressive activity through antagonizing the Wnt/beta-catenin signaling, and is frequently methylated in colorectal cancer. Clin Cancer Res 2008 Jan 1;14(1):55-61.

(48) Wozney JM, Rosen V, Celeste AJ, Mitsock LM, Whitters MJ, Kriz RW, et al. Novel regulators of bone formation: molecular clones and activities. Science 1988;242(4885):1528-34.

(49) Nissinen L, Pirila L, Heino J. Bone morphogenetic protein-2 is a regulator of cell adhesion. Exp Cell Res 1997;230(2):377-85.

(50) Baade Ro T, Utne Holt R, Brenne AT, Hjorth-Hansen H, Waage A, Hjertner O, et al. Bone morphogenetic protein-5, -6 and -7 inhibit growth and induce apoptosis in human myeloma cells. Oncogene 2004;23(17):3024-32.

(51) Deckers MM, van Bezooijen RL, van der HG, Hoogendam J, van Der BC, Papapoulos SE, et al. Bone morphogenetic proteins stimulate angiogenesis through osteoblast-derived vascular endothelial growth factor A. Endocrinology 2002 Apr;143(4):1545-53.

(52) Miyazawa K, Shinozaki M, Hara T, Furuya T, Miyazono K. Two major Smad pathways in TGF-beta superfamily signaling. Genes to Cells 2002;7:1191-204.

34

(53) Shi Y, Massague J. Mechanisms of TGF-beta Signaling from Cell Membrane to Nucleus. Cell 2003;113:685-700.

(54) Jonk LJC, Itoh S, Heldin C-H, ten Dijke P, Kruijer W. Identification and Functional Characterization of a Smad Binding Element (SBE) in the JunB Promoter That Acts as a Transforming Growth Factor-beta , Activin, and Bone Morphogenetic Protein-inducible Enhancer. J Biol Chem 1998;273(33):21145-52.

(55) Guo X, Wang XF. Signaling cross-talk between TGF-beta/BMP and other pathways. Cell Res 2009 Jan;19(1):71-88.

(56) Yanagita M. BMP antagonists: their roles in development and involvement in pathophysiology. Cytokine Growth Factor Rev 2005 Jun;16(3):309-17.

(57) Itoh S, ten DP. Negative regulation of TGF-beta receptor/Smad signal transduction. Curr Opin Cell Biol 2007 Apr;19(2):176-84.

(58) Ebisawa T, Fukuchi M, Murakami G, Chiba T, Tanaka K, Imamura T, et al. Smurf1 interacts with transforming growth factor-beta type I receptor through Smad7 and induces receptor degradation. J Biol Chem 2001 Apr 20;276(16):12477-80.

(59) Shi W, Chen H, Sun J, Chen C, Zhao J, Wang YL, et al. Overexpression of Smurf1 negatively regulates mouse embryonic lung branching morphogenesis by specifically reducing Smad1 and Smad5 proteins. Am J Physiol Lung Cell Mol Physiol 2004 Feb;286(2):L293-L300.

(60) von BA, Cho KW. Intracellular BMP signaling regulation in vertebrates: pathway or network? Dev Biol 2001 Nov 1;239(1):1-14.

(61) Miyazono K, Suzuki H, Imamura T. Regulation of TGF-beta signaling and its roles in progression of tumors. Cancer Sci 2003 Mar;94(3):230-4.

(62) Groppe J, Greenwald J, Wiater E, Rodriguez-Leon J, Economides AN, Kwiatkowski W, et al. Structural basis of BMP signalling inhibition by the cystine knot protein Noggin. Nature 2002;420(6916):636-42.

(63) Piccolo S, Sasai Y, Lu B, De Robertis EM. Dorsoventral Patterning in Xenopus: Inhibition of Ventral Signals by Directing Binding of Chordin to BMP-4. Cell 1996;86:589-98.

(64) Wardle FC, Welch JV, Dale L. Bone morphogenetic protein 1 regulates dorsal-ventral patterning in early Xenopus embryos by degrading chordin, a BMP4 antagonist. Mech Dev 1999 Aug;86(1-2):75-85.

35

(65) Nakamura Y, Ozaki T, Nakagawara A, Sakiyama S. A Product of DAN, a Novel Candidate Tumour Suppressor Gene, is Secreted into Culture Medium and Suppresses DNA Synthesis. Eur J Cancer 1997;33(12):1986-90.

(66) Sudo S, Avsian-Kretchmer O, Wang LS, Hsueh AJW. Protein related to DAN and cerberus (PRDC) is a BMP antagonist that participates in ovarian paracrine regulation. J Biol Chem 2004;M402376200.

(67) Chen Bo, Athanasiou M, Gu Q, Blair DG. Drm/Gremlin transcriptionally activates p21/cip1 via a novel mechanism and inhibits neoplastic transformation. Biochemical and Biophysical Research Communications 2002;295:1135-41.

(68) Hahn SA, Schmiegel WH. Recent discoveries in cancer genetics of exocrine pancreatic neoplasia. Digestion 1998 Aug;59(5):493-501.

(69) Howe JR, Roth S, Ringold JC, Summers RW, Jarvinen HJ, Sistonen P, et al. Mutations in the SMAD4/DPC4 gene in juvenile polyposis. Science 1998 May 15;280(5366):1086-8.

(70) Hanaoka E, Ozaki T, Nakamura Y, Moriya H, Nakagawara A, Sakiyama S. Overexpression of DAN causes a growth suppression in p53-deficient SAOS-2 cells. Biochem Biophys Res Commun 2000;278(1):20-6.

(71) Chen B, Athanasiou M, Gu Q, Blair DG. Drm/Gremlin transcriptionally activates p21(Cip1) via a novel mechanism and inhibits neoplastic transformation. Biochem Biophys Res Commun 2002;295(5):1135-41.

(72) Chen B, Blair DG, Plisov S, Vasiliev G, Perantoni AO, Chen Q, et al. Cutting edge: bone morphogenetic protein antagonists Drm/Gremlin and Dan interact with Slits and act as negative regulators of monocyte chemotaxis. J Immunol 2004;173(10):5914-7.

(73) Alarmo EL, Rauta J, Kauraniemi P, Karhu R, Kuukasjarvi T, Kallioniemi A. Bone morphogenetic protein 7 is widely overexpressed in primary breast cancer. Genes Chromosomes Cancer 2006 Apr;45(4):411-9.

(74) Schwaninger R, Rentsch CA, Wetterwald A, van der HG, van Bezooijen RL, van der PG, et al. Lack of noggin expression by cancer cells is a determinant of the osteoblast response in bone metastases. Am J Pathol 2007 Jan;170(1):160-75.

(75) Sneddon JB, Zhen HH, Montgomery K, van de RM, Tward AD, West R, et al. Bone morphogenetic protein antagonist gremlin 1 is widely expressed by cancer-associated stromal cells and can promote tumor cell proliferation. Proc Natl Acad Sci U S A 2006 Oct 3;103(40):14842-7.

36

(76) Bell E, Munoz-Sanjuan I, Altmann CR, Vonica A, Brivanlou AH. Cell fate specification and competence by Coco, a maternal BMP, TGFbeta and Wnt inhibitor. Development 2003 Apr;130(7):1381-9.

(77) Balemans W, Van HW. Identification of the disease-causing gene in sclerosteosis--discovery of a novel bone anabolic target? J Musculoskelet Neuronal Interact 2004 Jun;4(2):139-42.

(78) Brunkow ME, Gardner JC, Van NJ, Paeper BW, Kovacevich BR, Proll S, et al. Bone dysplasia sclerosteosis results from loss of the SOST gene product, a novel cystine knot-containing protein. Am J Hum Genet 2001 Mar;68(3):577-89.

(79) Staehling-Hampton K, Proll S, Paeper BW, Zhao L, Charmley P, Brown A, et al. A 52-kb deletion in the SOST-MEOX1 intergenic region on 17q12-q21 is associated with van Buchem disease in the Dutch population. Am J Med Genet 2002 Jun 15;110(2):144-52.

(80) van Bezooijen RL, Roelen BA, Visser A, van dW-P, de WE, Karperien M, et al. Sclerostin is an osteocyte-expressed negative regulator of bone formation, but not a classical BMP antagonist. J Exp Med 2004 Mar 15;199(6):805-14.

(81) Winkler DG, Sutherland MK, Geoghegan JC, Yu C, Hayes T, Skonier JE, et al. Osteocyte control of bone formation via sclerostin, a novel BMP antagonist. EMBO J 2003 Dec 1;22(23):6267-76.

(82) Brunkow ME, Gardner JC, Van NJ, Paeper BW, Kovacevich BR, Proll S, et al. Bone dysplasia sclerosteosis results from loss of the SOST gene product, a novel cystine knot-containing protein. Am J Hum Genet 2001 Mar;68(3):577-89.

(83) Kusu N, Laurikkala J, Imanishi M, Usui H, Konishi M, Miyake A, et al. Sclerostin Is a Novel Secreted Osteoclast-derived Bone Morphogenetic Protein Antagonist with Unique Ligand Specificity. J Biol Chem 2003;278(26):24113-7.

(84) Winkler DG, Sutherland MS, Ojala E, Turcott E, Geoghegan JC, Shpektor D, et al. Sclerostin inhibition of Wnt-3a-induced C3H10T1/2 cell differentiation is indirect and mediated by bone morphogenetic proteins. J Biol Chem 2005;280(4):2498-502.

(85) Kusu N, Laurikkala J, Imanishi M, Usui H, Konishi M, Miyake A, et al. Sclerostin Is a Novel Secreted Osteoclast-derived Bone Morphogenetic Protein Antagonist with Unique Ligand Specificity. J Biol Chem 2003;278(26):24113-7.

37

(86) Winkler DG, Yu C, Geoghegan JC, Ojala EW, Skonier JE, Shpektor D, et al. Noggin and sclerostin bone morphogenetic protein antagonists form a mutually inhibitory complex. J Biol Chem 2004;M400521200.

(87) Ellies DL, Viviano B, McCarthy J, Rey JP, Itasaki N, Saunders S, et al. Bone density ligand, Sclerostin, directly interacts with LRP5 but not LRP5G171V to modulate Wnt activity. J Bone Miner Res 2006 Nov;21(11):1738-49.

(88) Kamiya N, Ye L, Kobayashi T, Mochida Y, Yamauchi M, Kronenberg HM, et al. BMP signaling negatively regulates bone mass through sclerostin by inhibiting the canonical Wnt pathway. Development 2008 Nov;135(22):3801-11.

(89) McDonald NQ, Hendrickson WA. A structural superfamily of growth factors containing a cystine knot motif. Cell 1993 May 7;73(3):421-4.

(90) Craik DJ, Daly NL, Waine C. The cystine knot motif in toxins and implications for drug design. Toxicon 2001 Jan;39(1):43-60.

(91) Weidauer SE, Schmieder P, Beerbaum M, Schmitz W, Oschkinat H, Mueller TD. NMR structure of the Wnt modulator protein Sclerostin. Biochem Biophys Res Commun 2009 Feb 27;380(1):160-5.

(92) Itasaki N, Jones CM, Mercurio S, Rowe A, Domingos PM, Smith JC, et al. Wise, a context-dependent activator and inhibitor of Wnt signalling. Development 2003 Sep;130(18):4295-305.

(93) O'Shaughnessy RF, Yeo W, Gautier J, Jahoda CA, Christiano AM. The WNT signalling modulator, Wise, is expressed in an interaction-dependent manner during hair-follicle cycling. J Invest Dermatol 2004 Oct;123(4):613-21.

(94) Guidato S, Itasaki N. Wise retained in the endoplasmic reticulum inhibits Wnt signaling by reducing cell surface LRP6. Dev Biol 2007 Oct 15;310(2):250-63.

(95) Yanagita M, Oka M, Watabe T, Iguchi H, Niida A, Takahashi S, et al. USAG-1: a bone morphogenetic protein antagonist abundantly expressed in the kidney. Biochemical and Biophysical Research Communications 2004;316:490-500.

(96) Laurikkala J, Kassai Y, Pakkasjarvi L, Thesleff I, Itoh N. Identification of a secreted BMP antagonist, ectodin, integrating BMP, FGF, and SHH signals from the tooth enamel knot. Developmental Biology 2003;264:91-105.

38

(97) Kassai Y, Munne P, Hotta Y, Penttila E, Kavanagh K, Ohbayashi N, et al. Regulation of mammalian tooth cusp patterning by ectodin. Science 2005 Sep 23;309(5743):2067-70.

(98) Murashima-Suginami A, Takahashi K, Sakata T, Tsukamoto H, Sugai M, Yanagita M, et al. Enhanced BMP signaling results in supernumerary tooth formation in USAG-1 deficient mouse. Biochem Biophys Res Commun 2008 May 16;369(4):1012-6.

(99) Yanagita M, Okuda T, Endo S, Tanaka M, Takahashi K, Sugiyama F, et al. Uterine sensitization-associated gene-1 (USAG-1), a novel BMP antagonist expressed in the kidney, accelerates tubular injury. J Clin Invest 2006 Jan;116(1):70-9.

39

CHAPTER II:

A HUMAN BONE MORPHOGENETIC PROTEIN ANTAGONIST IS DOWNREGULATED IN RENAL CANCER

Kimberly Rose Blish, Wei Wang, Mark C. Willingham, Wei Du, Charles E. Birse, Surekha R. Krishnan, Julie C. Brown, Gregory A. Hawkins, A. Julian Garvin,

Ralph B. D’Agostino, Jr. Frank M. Torti, Suzy V. Torti

The following manuscript was published in Molecular Biology of the Cell

(19(2):457-464, February 2008). Variations in style are due to the requirements

of the journal. Figure and manuscript preparation by K.R. Blish. Experiments

performed and data gathered by K.R. Blish and Wei Wang (Wnt luciferase assay)

with assistance and expertise from W. Du (immunostaining), M.C. Willingham

(immunostaining quantitation and analysis) and J.C. Brown (cell culture). Dot

blot and preparation of rhSOSTDC1 by C.E. Birse and S.R. Krishnan at Human

Genome Sciences, Inc. A.J. Garvin and G.A. Hawkins served advisory roles in

experimental design. Statistical analyses for immunostaining by R.B. D’Agostino,

Jr. F.M. Torti and S.V. Torti served advisory and editorial roles related to

experimental design, data analysis, and manuscript preparation.

40

ABSTRACT

We analyzed expression of candidate genes encoding cell surface or secreted

proteins in normal kidney and kidney cancer. This screen identified a BMP

antagonist, SOSTDC1 (SclerOSTin Domain-Containing-1) as downregulated in

kidney tumors. To confirm screening results, we probed cDNA dot blots with

SOSTDC1. SOSTDC1 message was decreased in 20/20 kidney tumors

compared to normal kidney tissue. Immunohistochemistry confirmed significant

decrease of SOSTDC1 protein in clear cell renal carcinomas relative to normal

proximal renal tubule cells (p<0.001). Expression of SOSTDC1 was not

decreased in papillary and chromophobe kidney tumors. SOSTDC1 was

abundantly expressed in podocytes, distal tubules, and transitional epithelia of

the normal kidney. Transfection experiments demonstrated that SOSTDC1 is

secreted and binds to neighboring cells and/or the extracellular matrix.

SOSTDC1 suppresses both BMP-7-induced phosphorylation of R-Smads-1, -5

and -8 and Wnt-3a signaling. Restoration of SOSTDC1 in renal clear carcinoma

cells profoundly suppresses proliferation. Collectively, these results

demonstrate that SOSTDC1 is expressed in the human kidney and decreased in

renal clear cell carcinoma. Since SOSTDC1 suppresses proliferation of renal

carcinoma cells, restoration of SOSTDC1 signaling may represent a novel target

in treatment of renal clear cell carcinoma.

41

INTRODUCTION

Bone morphogenetic proteins (BMPs) are members of the transforming

growth factor-β (TGF-β) superfamily that function as extracellular signaling

proteins (Urist, 1965; Wozney, 1992; Reddi, 2001). BMPs are pleiotropic whole-

body morphogens involved in proliferation, differentiation, and maintenance of

diverse tissue types (ten Dijke et al., 1994; Yamashita et al., 1995). BMP

signaling is primarily mediated though binding to a heterotetrameric complex of

cognate type I and type II receptors (BMPRs). Upon ligand binding, the

serine/threonine kinase on the Type II receptors phosphorylates the Type I

receptors, triggering intracellular Smad signal transduction cascades.

Phosphorylation of the BMP-specific intracellular Smads (Smad-1, -5, and -8)

allows them to bind the partnering co-Smad, Smad4 (Yamashita et al., 1995;

Kawabata et al., 1998; Miyazono et al., 2000; Miyazono et al., 2001; Moustakas

et al., 2001; Chen et al., 2004). This complex of Smads and other DNA binding

proteins undergoes nuclear translocation and triggers transcriptional activation of

target genes, including cell cycle regulators or apoptosis mediators such as

p21/Cip1/Waf1, bax, and p53 (Chen et al., 2002; Pouliot and Labrie, 2002;

Fukuda et al., 2006).

Regulation of BMP signaling is accomplished in a tissue-specific manner

via BMP antagonists (Yanagita, 2005; 2006). Classical BMP antagonists bind

BMPs in the extracellular space, preventing the association of BMPs with their

receptors. BMP antagonists may also prevent the activity of BMPs by fostering

42

their retention in the ER (Guidato and Itasaki, 2007). BMP antagonists have

been postulated to act as the key ‘on/off’ switches that spatially and temporally

regulate local BMP activity. Thus, BMP antagonists may represent local targets

to control specific tumors.

In addition to their effects on BMP pathways, BMP antagonists have also

been reported to impinge on Wnt signaling, For example, sclerostin exerts its

inhibitory effects on bone formation in mice primarily through inhibition of Wnt

signaling (van Bezooijen et al., 2007); in Xenopus, the BMP antagonist Wise

inhibits or stimulates Wnt signaling in a manner dependent on context (Itasaki et

al., 2003) and cellular localization (Guidato and Itasaki, 2007).

Several genes have been implicated in the development of renal clear cell

carcinoma. These include VHL, an ubiquitin ligase and tumor suppressor, as

well as c-Met and BHD (Linehan et al., 2005; Brugarolas, 2007). To identify

novel genes whose expression is altered in renal cancer, we probed arrays of

secreted and cell surface cDNAs with cDNA libraries prepared from tumor tissue

and matched normal tissue from 20 kidney cancer patients. This screen

identified SOSTDC1, a BMP antagonist, as a candidate gene with reduced

expression in renal cancer.

SOSTDC1 is orthologous to a recently characterized murine antagonist of

BMPs-2, -4, and -7 termed ectodin/USAG-1 (Mus musculus) (Yanagita et al.,

43

2004). In other species, this protein is known as USAG-1 (Rattus norvegius)

(Yanagita et al., 2004), ectodin (Mus musculus) (Laurikkala et al., 2003; Kassai

et al., 2005), and wise (Xenopus laevis) (Itasaki et al., 2003). Recent mouse

knock-outs of ectodin/USAG-1 have shown that this molecule is important in

tooth formation and modulation of the renoprotective effects of BMP-7 (Kassai et

al., 2005; Yanagita et al., 2006). The human protein is 98% identical to the

rodent protein, and like most members of the BMP and BMP antagonist families,

is predicted to contain a C-terminal cystine-knot motif (Avsian-Kretchmer and

Hsueh, 2004). However, the role of SOSTDC1 in human tissues or human

cancers has not been reported.

In this report, we demonstrate that SOSTDC1 is highly expressed in cells

of the normal human kidney, but is significantly decreased in clear cell renal

carcinoma.

MATERIALS & METHODS

Recombinant proteins and reagents

Recombinant human BMP-7 (rhBMP-7), noggin-Fc chimera, recombinant

mouse Wnt-3a (rmWnt-3a), and DKK1 were purchased from R & D Systems

(Minneapolis, MN). Mammalian expression vectors encoding the open reading

frame (ORF) for BMP-7 (pReceiver-M2-BMP7), dickkopf-1 (DKK1; pReceiver-

M2-DKK1), noggin (pReceiver-M02-noggin), and Wnt-3a (pReceiver-M02-Wnt3a)

were purchased from GeneCopoeia, Inc. (Germantown, MD). The TCF/LEF

44

luciferase reporter construct Super 8X TOPFlash and the corresponding mutant

control Super 8X FOPFlash vectors were obtained from the Addgene Repository

(Cambridge, MA). Plasmids deposited by Dr. Randall Moon (Veeman et al.,

2003).

The following commercial antibodies were obtained: anti-phospho-Smad-1,-5,-8

(Cell Signaling Technologies, Danvers, MA), anti-glyceraldehyde-3-phosphate

dehydrogenase (GAPDH, Research Diagnostics International, Flanders, NJ), and

mouse monoclonal anti-FLAG (Sigma Aldrich, St. Louis, MO).

Production of anti-hSOSTDC1 antisera

Production and affinity purification of an anti-SOSTDC1 peptide polyclonal

rabbit anti-sera directed against the 18 C-terminal amino acids of the SOSTDC1

sequence was performed by Quality Control Biochemicals Custom Services

(Hopkinton, MA).

Cells and culture conditions

Human embryonic kidney cells, HEK293, were maintained in Dulbecco’s

Modified Eagle Medium (DMEM; Gibco, Invitrogen,) supplemented with 10%

fetal bovine serum (FBS; Gibco, Invitrogen) and 1% penicillin/streptomycin

(Invitrogen). The following kidney cancer cell lines were obtained from the

American Type Culture Collection (ATCC, Rockville, MD): 786-O (CRL-1932,

renal clear cell adenocarcinoma), ACHN (CRL-1611, renal cell adenocarcinoma,

45

pleural effusion), 769-P (CRL-1933, derived from primary renal clear cell

adenocarcinoma), A-704 (HTB-45, adenocarcinoma), and Caki-1 (HTB-46, clear

cell carcinoma skin metastasis). All kidney cancer cell lines were grown in

McCoy’s 5A basal media (Gibco) supplemented with 10% FBS, 1%

penicillin/streptomycin, and 200mM L-glutamine. Renal epithelial proximal tubule

cells (RPTECs) were purchased from Cambrex (Hopkinton, MA) and were grown

in Renal Epithelial Basal Medium (REBM) with supplements (Cambrex).

Mammalian Vector Construction and preparation of recombinant, human

SOSTDC1 (rhSOSTDC1)

The open reading frame encoding the SOSTDC1 protein (NCBI:

NM_015464) without the final stop codon was amplified from cDNA by

polymerase chain reaction (PCR) and cloned into the BamH1/Asp718 restriction

sites of the pFLAG-CMV5a vector (Invitrogen). The resulting recombinant

human SOSTDC1 (rhSOSTDC1) has the FLAG (DYKDDDDK) epitope added to

its C-terminus.

HEK293 cells were transiently transfected with the pFLAG-CMV5A

SOSTDC1 vector using LipofectAMINE™ reagent (Invitrogen). Conditioned

media was harvested after 2 days and purified using an anti-FLAG M2 affinity

column (Sigma, St. Louis, MO).

46

Tissue Samples and Immunohistochemistry

All tissue samples were used with approval from the Institutional Review

Board at Wake Forest University. Paraffin blocks from the pathology archives

were sectioned at 5 micron intervals. After heat-induced antigen retrieval with

Antigen Unmasking Solution (AUS, Vector Labs, Burlingame, CA), slides were

stained with anti-SOSTDC1 polyclonal sera followed by affinity-purified goat anti-

rabbit-IgG-HRP (Jackson ImmunoResearch) and diaminobenzidine (DAB) HRP

substrate. The slides were counterstained with hematoxylin, viewed under

brightfield conditions, and photographed at the same magnification.

Intracellular SOSTDC1 Relative Quantification

SOSTDC1 staining was quantified in Image J (NIH freeware (Rasband,

1997-2007) as mean pixel intensity. Brightfield digital images of the peroxidase-

stained tissues were collected at equal exposure conditions. To avoid bias of cell

selection, an identical grid consisting of 9 equal squares was superimposed over

every image digitally. Only tumor cells crossing predetermined lines of the grid

were included in analysis. Analysis in Image J was completed via the “measure

line” function as follows: a 5 pixel straight line selection was drawn in the

perinuclear region of each candidate cell. Then, the “Measure” command was

chosen from the Analyze menu. The mean pixel intensity value was averaged

for 30 separate cells and reported as a percentage of maximal density. Mix pixel

intensity was also quantified using Adobe Photoshop (Lehr et al., 1997), which

gave virtually identical results.

47

Statistical Analysis

Statistical analysis was performed by the Biostatistics core of the

Comprehensive Cancer Center of Wake Forest University. A one-way ANOVA

was fit to compare the mean percent maximal density of the five groups. Pair

wise comparisons were then performed between groups using a conservative p-

value of 0.005 for significance (using a Bonferroni adjustment to allow for all 10

pair wise comparisons between groups to be made). Standard deviations for the

outcome were compared across the five groups to assure that they were

comparable.

cDNA microarray and RNA dot blot

Microarray analyses were performed on tissues from matched normal and

tumor tissues from 20 kidney cancer patients and cDNAs from other normal and

tumor tissue types. Labeled SOSTDC1 probes were hybridized to the array and

resulting signals were scanned and quantified. To independently verify these

results, the BD Clontech™ Cancer Profiling Array I, also having 20 matched

normal and tumor kidney tissues, was probed with radiolabeled SOSTDC1. The

majority of these tissues (~70%) are from patients with renal cell carcinoma, clear

type.

rhSOSTDC1 transient transfection and immunofluorescence

pFLAG-CMV5a-SOSTDC1 was transiently transfected into HEK293 cells

via LipofectAMINE™ reagent. Twenty-four hours after transfection, cells were

48

washed once in room temperature PBS, and then fixed in 100% methanol at 4°C

for 15 minutes. Cells were labeled with anti-FLAG mouse monoclonal antibody

(Sigma Aldrich) conjugated to rhodamine. Nuclei were stained with 4',6-

Diamidino-2-phenylindole (DAPI, CalBiochem, EMD Biosciences, Inc., San

Diego, CA) and cells were visualized through fluorescent microscopy.

Stimulation of phosphorylated Smad (p-Smad) production with BMP-7 and

p-Smad immunoblot

Renal cancer cells were plated in phenol red-free McCoy’s 5A with 5%

FBS and 1% penicillin/streptomycin and allowed to attach overnight. Media was

replaced with phenol-red free McCoy’s 5A (Promocell GmBH, Heidelberg,

Germany) containing 0.5% FBS and 1% penicillin/streptomycin with additions of

rhBMP-7 (40 ng/mL), rhnoggin-Fc chimera (150 ng/mL) and/or rhSOSTDC1 (150

ng/mL). Cells were placed at 37°C for either 1 or 3 hours. At harvest, cells were

washed with PBS, and 150 uL of sodium dodecyl sulfate polyacrylamide gel

electrophoresis (SDS-PAGE) loading buffer added (0.5M Tris pH 6.8, 20%

glycerol, 10% SDS, and 5% beta-mercaptoethanol). Lysates were

electrophoresed on 10% SDS-PAGE gels, transferred to nitrocellulose, and p-

Smad detected by Western blot analysis according to the instructions provided

by the antibody supplier. p-Smad-1,-5,-8 was visualized with PicoWest

chemiluminescent substrate (Pierce Biotechnology, Rockford, IL). GAPDH was

used as a loading control.

49

Wnt signaling luciferase assay

Transfections were performed using Fugene6 (Roche Applied Science,

Indianapolis, IN) according to the manufacturer’s instructions. For each condition,

300,000 HEK293 cells were transfected with 1.76 μg of total plasmid DNA. Each

group was some combination of: 0.25 μg TOPFlash reporter plasmid, 0.01 μg of

internal control pRL-Tk, 0.5 μg pReceiver-M02-Wnt3a expression vector, 1 μg

pFLAG-CMV5a-SOSTDC1, or 1 μg pReceiver-M02-DKK1. Empty pFLAG-

CMV5a vector was used to maintain a consistent level of plasmid DNA in each

transfection. Luciferase activity was measured twenty-four hours after

transfection using a Dual Luciferase Assay Kit according to the manufacturer’s

instructions (Promega, Madison, WI).

Transient transfection of SOSTDC1 in kidney cancer cell lines

pFLAG-CMV5a (empty) and pFLAG-CMV5a-SOSTDC1 were transiently

transfected into 769-P renal cancer cell lines using LipofectAMINE™ 2000

(Invitrogen) via manufacturer’s instructions. Twenty-four hours after transfection,

FBS was added to the transfected cell cultures to a final concentration of 5%. To

evaluate the success of transient SOSTDC1 overexpression, cells were

harvested for real-time RT PCR 36 hours after transfection.

Real-time RT PCR evaluation of SOSTDC1 transfected cells

Approximately 2.5 x 106 transfected cells were trypsinized, washed twice

with PBS, and then suspended in Trizol solution (Invitrogen) for preparation of

50

total RNA. RNA was treated with RQ1 RNase-free DNase (Promega) and then

purified with the Absolutely RNA® MiniPrep kit (Stratagene, La Jolla, CA).

Reverse transcriptase reactions were performed with the Taqman® Reverse

transcription kit (Applied Biosystems, Foster City, CA) per kit instructions.

Resulting cDNA was used in real-time PCR reactions on the ABI Prism 7000

Sequence Detection System machine with SYBR Green PCR Master Mix

(Applied Biosystems) according to manufacturer’s instructions. SOSTDC1

primers used in the real-time analysis were: Forward; 5’-

TGGATTGGAGGAGGCTATGGAACA-3’; Reverse, 5’-

ACTTGCAGGCAGTGACTACTGTGA-3’. Primers to the housekeeping gene

GAPDH were used to standardize for total RNA levels: Forward, 5’-

GAAGGTGAAGGTCGGAGTC- 3’; Reverse, 5’- GAAGATGGTGATGGGATTTC-

3’.

Effect of SOSTDC1, noggin, and DKK1 on proliferation

50,000 cells of the 769-P cell line were plated per well of a 24 well plate.

Cells were transfected with 0.4 μg either pFLAG-CMV5a empty vector control,

pFLAG-CMV5a-SOSTDC1, pReceiver-M02-Noggin, or pReceiver-M02-DKK1

with LipofectAMINE™ 2000 as described above. Twenty-four hours after

transfection, FBS was added to the transfected cell cultures to a final

concentration of 5%. Cells were allowed to recover from transfection for an

additional 2 hours before trypsinization and reseeding at 5,000 cells per well in

an E-plate for use on the ACEA RT-CES® system (ACEA Biosciences, Inc., San

51

Diego, CA). This system allows real-time monitoring of cultures grown in a 96-

well plate format. Each type of transfected cell was replated in its own

conditioned media.

RESULTS

Downregulation of SOSTDC1 in renal tumors.

To identify novel genes whose expression is altered in cancer, we probed

cDNA arrays of secreted and cell surface genes with cDNA libraries prepared

from tumor tissue and matched normal tissue from 20 renal cell carcinoma

patients. High SOSTDC1 expression was observed in normal kidney tissue, but

mean SOSTDC1 expression decreased by more than one standard deviation in

renal tumors (data not shown). To confirm this finding, a Cancer Profiling Array

(Clontech, Mountain View, CA) containing 500 cDNAs from a variety of tissues,

tumors, and cancer cell lines was probed with radiolabeled SOSTDC1

oligonucleotide. As seen in Figure 1, SOSTDC1 was downregulated in all 20 of

matched normal and tumor kidney sample pairs on this array. Of note, these

matched samples were from a different set of kidney cancer patients than the 20

patients included in the microarray analysis. These results suggest that

SOSTDC1 mRNA is downregulated in a majority of kidney cancers.

52

Figure 1. Figure 1. Downregulation of human SOSTDC1 mRNA in kidney tumors.

cDNA of matched normal and tumor kidney tissue from the same patient

was hybridized with radiolabeled DNA oligonucleotides to human

SOSTDC1. N, normal (lane 1); T, tumor (lane 2). SOSTDC1 cDNA

decreases in normal versus tumor tissue in 20 out of 20 patients; 17 out of

20 (85%) of tumors show virtually no detectable SOSTDC1 mRNA.

53

Expression of SOSTDC1 within the normal kidney and in kidney cancer.

To test whether expression of the SOSTDC1 protein was similarly

affected, we prepared an anti-SOSTDC1 antibody and performed

immunohistochemical analyses. We first evaluated the distribution of SOSTDC1

protein within the normal human kidney. IHC revealed the presence of

intracellular SOSTDC1 within most epithelial cells of the kidney, including

proximal and distal tubules and transitional urothelium (Figure 2, top panels).

Strong SOSTDC1 signal was observed within the distal tubules, collecting ducts

and podocytes of the glomerulus, with more moderate staining in the cytoplasm

of proximal tubule epithelial cells. Quantification of cytoplasmic SOSTDC1

staining intensity in each of these cell types demonstrated that an increase in

expression of SOSTDC1 was generally correlated with progression through the

nephron and into the collecting ducts and urothelium (Table 1). These results

are consistent with the strong distal tubule expression observed for USAG-1, the

murine homolog of SOSTDC1 (Yanagita et al., 2004).

54

Figure 2. Immunohistochemical analysis of SOSTDC1 expression in

normal kidney and renal cancer subtypes. Representative images of IHC

staining with anti-SOSTDC1 antiserum are shown. All images are obtained at

equal magnification and exposure. Left panel represents procedural background

control (without anti-SOSTDC1 primary antibodies) and the right panel depicts

the same tissue stained with anti-SOSTDC1 antiserum. Top row: normal kidney.

G = glomerulus, arrowhead = distal tubule, arrow = proximal tubule. Bottom

rows: renal epithelial or urothelial cancers. RCC, renal cell carcinoma; TCC,

transitional cell carcinoma.

55

Figure 2.

56

Table 1. Quantification of intracellular anti-SOSTDC1 by

immunohistochemical staining.

Normal Cell Type Staining Intensity

(% of Maximum)

Podocyte 45 ± 6

Proximal Tubule 25 ± 4

Distal Tubule 49 ± 4

Collecting Duct 53 ± 4

Renal Urothelium 48 ± 3

Bladder Urothelium 61 ± 5

Table 1. Quantification of intracellular anti-SOSTDC1 by

immunohistochemical staining. Staining intensity was analyzed as

described in Materials and Methods. Means and standard deviations are

reported.

We then assessed the expression of SOSTDC1 in a variety of kidney

tumor types. Over 75% of kidney tumors are of the clear cell type. Papillary and

chromophobe types are also seen. These represent approximately 15 and 5% of

all kidney cancers, respectively, and have a more favorable prognosis than clear

cell (Iliopoulos, 2006). We investigated SOSTDC1 protein expression in these

kidney tumor types and compared them to normal controls. In addition, we

57

examined expression of SOSTDC1 in transitional cell carcinoma of the bladder.

Representative images are shown in Figure 2, and results are quantified in

Figure 3.

As seen in Figure 2, SOSTDC1 expression is markedly diminished in clear

cell renal carcinoma (RCC-clear cell) when compared to normal kidney. In

papillary (RCC-papillary), transitional cell carcinoma (TCC) and chromophobe

types (data not shown), SOSTDC1 staining was intracellular and of higher signal

intensity than the staining in clear cell renal carcinoma. Results of SOSTDC1

protein quantification in various renal cell carcinomas are shown in Figure 3.

Because multiple comparisons were performed, we required a conservative p

value <0.005 for significance (see Materials and Methods). ANOVA analysis

demonstrated a large group difference, p<0.001. RCC-clear cell had significantly

lower SOSTDC1 immunostaining than any other group, including normal renal

urothelium (p < 0.001 for all comparisons). Additionally, no RCC-clear cell

sample had higher staining than its likely cell of origin, the proximal tubule cell (p

value also <0.001 see also Table 1). Normal renal urothelium and bladder TCC

groups were not significantly different from one another (p=.137), but were

significantly higher than the proximal tubule and RCC-clear cell groups. The

RCC-Papillary group was significantly higher than RCC-clear cell (p<0.001) and

lower than normal renal urothelium (p=0.0008), but not significantly different from

proximal tubules (p=0.03) or bladder TCC (p=.008) based on the Bonferroni

correction. Comparison of standard deviations across the five groups indicated

58

that variability was comparable between groups. Thus, immunohistochemical

analysis is concordant with SOSTDC1 mRNA analysis, and indicates that

SOSTDC1 is profoundly decreased in RCC-clear cell tumors.

Figure 3.

Figure 3. Quantification of SOSTDC1 staining. Relative SOSTDC1 staining

of normal and tumor tissues was quantified as described in Materials and

Methods in 3-9 individuals per group; means and standard deviations are

shown. RCC-clear cell has significantly less SOSTDC1 than the other

tissues. * = p < 0.001.

59

SOSTDC1 is secreted and antagonizes signaling of BMP-7 in kidney cancer

cells

We next assessed functional consequences of SOSTDC1 expression. As

classic BMP antagonists function extracellularly, we first tested whether the

localization of SOSTDC1 was compatible with an extracellular mode of action.

cDNA for C-terminally FLAG-tagged SOSTDC1 was cloned into an expression

vector and transiently transfected into HEK293 cells. rhSOSTDC1-FLAG was

detected via immunofluorescence after binding of anti-FLAG-rhodamine

antibodies to fixed cells (Figure 4). This experiment revealed the presence of

SOSTDC1 predominantly in the secretory apparatus of transfected, highly

expressing cells. We transfected cells at low efficiency for this experiment so

that each highly expressing transfectant was relatively isolated, enabling

visualization of extracellular SOSTDC1-specific signal on the outside of low-

expressing cells (Figure 4, right panels). Greatest staining intensity was seen

closest to highly expressing cells, with staining intensity decreasing with distance

(Figure 4, center panel). Western blotting also demonstrated the presence of

SOSTDC1 in the media of cultured cells (data not shown). These results

demonstrate that SOSTDC1 is secreted and binds to the extracellular matrix

and/or cell surface, consistent with an ability to function in an autocrine or

paracrine manner.

60

Figure 4.

Figure 4. Normal human SOSTDC1 is secreted and binds to matrix or cell

surface proteins around neighboring cells. HEK293 cells were transfected

with SOSTDC1-FLAG and expression analyzed using rhodamine-

conjugated anti-FLAG antibody (mouse anti-FLAG-rh). Nuclei were stained

with DAPI. Top row: color fluorescent image. Bottom row: Corresponding

black and white channel of same images. All images obtained at the same

magnification and exposure. Left panels: a highly expressing cell (arrow)

shows strong signal in the cytoplasm, especially the endoplasmic

reticulum and Golgi apparatus. Center panels: secreted SOSTDC1 (small

arrowheads) accumulates around both expressing and non-expressing

cells. Right panels: Empty vector control transfected cells show no

background rhodamine fluorescence.

61

We next tested whether SOSTDC1 acts as a functional antagonist of BMP-7

signaling. BMP-7 binds to BMPRs leading to receptor activation and

phosphorylation of intracellular R-Smads (Smad-1,-5,-8). We predicted that if

SOSTDC1 functions as an extracellular BMP antagonist, it should bind to BMP-7,

prevent the BMP-receptor interaction, and stop BMP-7 induced Smad

phosphorylation. To test this, cells were treated with BMP-7 in presence and

absence of SOSTDC1, and phosphorylation of the BMP-specific R-Smads

measured.

As shown in Figure 5, treatment with rhBMP-7 alone caused a rapid and

robust increase in phosphorylation of R-Smads-1,-5, and -8 (p-Smad) in cultured

kidney cancer (786-O) cells. A similar rapid increase in p-Smad levels was also

observed in other kidney cell lines, including ACHN, 769-P, A-704, Caki-1, as

well as normal RPTEC cells (data not shown). To test the ability of rhSOSTDC1

to antagonize this effect, rhSOSTDC1 was added to the cells simultaneously with

rhBMP-7. Noggin, a classical antagonist of BMP-7, was used as a positive

control (Avsian-Kretchmer and Hsueh, 2004; Kawabata et al, 1998). At both 1

hour and 3 hours, rhSOSTDC1 antagonized the BMP-7 induced production of p-

Smad at least as efficiently as noggin (Figure 5). Thus, SOSTDC1 effectively

antagonizes BMP-7 signaling.

62

Figure 5.

Figure 5. rhSOSTDC1 antagonizes the production of phosphorylated Smad

in response to BMP-7 signaling. 786-O cells were treated BMP-7,

SOSTDC1, or noggin as indicated and Smad phosphorylation analyzed by

western blotting. SOSTDC1 inhibited the production of p-Smad-1, -5, and -

8 by 3.5 fold. GAPDH was used as a loading control. Note: for clarity,

lanes from the same blot have been rearranged. Similar results were

observed in the 769-P and A-704 renal adenocarcinoma cell lines (data not

shown).

63

SOSTDC1 suppresses Wnt signaling in kidney cells

Since Wise, an ortholog of SOSTDC1, can both stimulate and inhibit the

Wnt pathway (Itasaki et al., 2003), we asked whether SOSTDC1 would affect

Wnt signaling in human kidney cells. Cells were transfected with TOPFlash, a

construct containing 8 TCF/LEF transcription factor binding sites upstream of the

luciferase gene. Cells were co-transfected with Wnt-3a (to activate the classical

Wnt/β-catenin/TCF signaling pathway) with and without co-transfection with

SOSTDC1 expression vector. Co-transfection with DKK1, a known Wnt inhibitor,

was used as a control. As seen in Figure 6, Wnt-3a increased expression of the

TOPFlash luciferase, and this response was blocked by DKK1. SOSTDC1

inhibited Wnt signaling to an extent comparable to DKK1. Thus, SOSTDC1

suppresses Wnt-3a signaling as well as BMP signaling in human kidney cells.

64

Figure 6.

Figure 6. SOSTDC1 antagonizes Wnt-3a signaling in renal carcinoma cells.

HEK293 cells were transiently transfected with Super 8X TOPFlash

luciferase reporter construct plus Wnt-3a and antagonist expression

constructs as outlined in Materials & Methods. After 24 hours, resulting

luciferase activity was measured for each transfection group and reported

as relative luciferase units. Means and standard errors of triplicate

experiments are shown.

65

SOSTDC1 suppresses proliferation of RCC-clear cell cultures

Clear cell tumors are the most prevalent histopathologic type of kidney

cancer. Because these tumors showed the greatest change in SOSTDC1

protein expression (Figure 2,3), we asked what functional effects SOSTDC1

might have on these cells. To address this question, the renal cancer clear cell

line 769-P was transiently transfected with the pFLAG-CMV5a-SOSTDC1

expression plasmid, and the effect of SOSTDC1 on proliferation measured. As

shown in Figure 7, SOSTDC1 strikingly inhibited proliferation of 769-P cells:

SOSTDC1 was cytostatic to these cells, while cells transfected with empty vector

proliferated normally. Since SOSTDC1 inhibits BMP and Wnt signaling (Figure

5, 6), we asked whether other inhibitors of these pathways would exert similar

effects. Cells were transfected with an expression vector for noggin, an inhibitor

of the BMP pathway, or a vector encoding DKK1, an inhibitor of the Wnt

pathway, and effects on proliferation measured. As seen in Figure 7, inhibition of

either of these pathways also inhibited proliferation, although the effect was more

modest than that seen with SOSTDC1.

66

Figure 7.

Figure 7. Overexpression of SOSTDC1 suppresses proliferation of 769-P

RCC-clear cell cultures. 769-P cells were transiently transfected with

pFLAG-CMV5a-SOSTDC1 (open circles), pFLAG-CMV5a empty vector

control (darkened circles), pReceiver-M02-noggin (open triangles), or

pReceiver-M02-DKK1 (darkened triangles). After transfection recovery,

cells were reseeded onto the 96 well ACEA E-plate and loaded into the

ACEA RT-CES system for continuous cell growth monitoring. All treatment

groups were seeded in triplicate, means and standard deviations for each

group at selected times for a representative experiment are shown here.

67

DISCUSSION

Differential expression of genes in normal and cancer tissue has been

used to study processes involved in malignant change, to identify new tumor

markers, and to identify new targets in tumor therapy. Motivated by these goals,

we sought to identify secretory or cell surface proteins altered in kidney cancer.

Several genes have been implicated in the development of renal clear cell

carcinoma, including VHL, a ubiquitin ligase and tumor suppressor, as well as c-

Met, BHD and sFRP1 (Linehan et al., 2005; Brugarolas, 2007); (Gumz et al.,

2007); however current molecular understanding of the disease is incomplete. To

identify novel genes whose expression is altered in renal cancer, we probed

arrays of secreted and cell surface cDNAs with cDNA libraries prepared from

tumor tissue and matched normal tissue from 20 kidney cancer patients. This

screen identified SOSTDC1, a BMP antagonist, as a candidate gene with

reduced expression in renal cancer.

Studies have previously linked BMPs, the target of BMP antagonists, as

well as BMP receptors, to cancer. For example, aberrant BMP or BMPR

expression has been noted in cancers, including osteosarcomas, breast, kidney,

colon, prostate (Miyazaki et al., 2004; Hsu et al., 2005; Alarmo et al., 2006).

Smad4, the main intracellular target of both BMP and TGF-β signaling, is a

known tumor suppressor in pancreatic and intestinal cancer (Alberici et al.,

2006).

68

SOSTDC1 is a BMP antagonist, as evidenced both by its sequence (Vitt et

al., 2001; Avsian-Kretchmer and Hsueh, 2004)and function (Figure 5). Although

BMP antagonists have not been previously implicated in kidney cancer, they

have been implicated in malignant processes in other tissue types. For example,

noggin (a BMP antagonist) suppresses growth of implanted prostate cancer cells

(Feeley et al., 2006). DAN, another BMP antagonist discovered in v-mos

transformed cells, inhibits neoplastic transformation (Chen et al., 2002).

However, the activities of BMP antagonists are complex and may vary. For

example the BMP antagonist gremlin 1 is expressed by stromal cells associated

with esophageal, pancreatic and other cancers (not including the kidney), and

may promote tumor cell proliferation (Sneddon et al., 2006). Our data

demonstrate that in renal clear cell cancer, decreases in SOSTDC1 mRNA and

protein are associated with malignant change. These data are the first link

between the BMP antagonist SOSTDC1 and the process of carcinogenesis.

A number of intersections between BMP signaling and regulatory

pathways important in carcinogenesis have been recently identified. These

include such key cancer pathways as Wnt/β-catenin, PI3-K/PTEN, and

MAPK/ERK (Aubin et al., 2004; He et al., 2004; Moustakas and Heldin, 2005;

Pardali et al., 2005). Further, BMP signaling directly impinges on the cell cycle

regulatory proteins p21 and Rb: BMP signaling regulates p21/Cip1/Waf1

expression in prostate cancer (Haudenschild et al., 2004), thyroid carcinomas

69

(Franzen and Heldin, 2001), and inhibits phosphorylation of Rb protein in breast

cancer (Ghosh-Choudhury et al., 2000).

We found that in human kidney cells, SOSTDC1 inhibits both BMP-7

(Figure 5,6) and Wnt-3a (Figure 6) signaling. Similarly, previous reports have

indicated that orthologs of SOSTDC1 can affect both BMP and Wnt pathways in

complex ways that are dependent on context and cellular localization. USAG-1,

the murine ortholog of SOSTDC1, binds to BMPs-2, -4 and -7 and antagonizes

BMP activity in Xenopus embryos (Yanagita et al., 2004). Wise, another ortholog

of SOSTDC1 identified in chick and Xenopus, can both activate and antagonize

Wnt signaling during Xenopus development (Itasaki et al., 2003; Guidato and

Itasaki, 2007).

Recent work has directly implicated the Wnt pathway in kidney cancer.

WTX, a negative regulator of Wnt signaling, was identified as a new tumor

suppressor in Wilms tumor, a pediatric form of kidney cancer (Major et al., 2007).

In a separate study using genomic profiling of adult renal cell carcinoma tumors,

loss of sFRP1, a secreted form of the frizzled receptor that acts as an inhibitor of

the Wnt pathway, was identified in 15/15 renal cell carcinoma patients (Gumz et

al., 2007). Our results indicate that SOSTDC1, a BMP antagonist that can also

negatively regulate the Wnt pathway (Figure 6), is similarly downregulated in

kidney cancer. Collectively, these results suggest that upregulation of Wnt

signaling via reduction of sFRP1, SOSTDC1 or perhaps other as yet unidentified

70

mechanisms, may make a critical contribution to the development of kidney

cancer.

We observed that SOSTDC1 exerts an antiproliferative effect on clear cell

renal carcinoma cells (Figure 7). Similarly, sFRP1 was also reported to

negatively affect cell proliferation (Gumz et al., 2007). Considering that

SOSTDC1 is downregulated in virtually all of the 20 cancer specimens we

analyzed (Figure 1), we speculate that the pervasive decrease in SOSTDC1 in

kidney cancers may arise due to the requirement of the malignant cell to

overcome this antiproliferative activity of SOSTDC1. Our evidence further

suggests that inhibition of both BMP and Wnt pathways may underlie the

antiproliferative effect of SOSTDC1. Thus, the BMP antagonist noggin as well as

the Wnt antagonist DKK1 also inhibited proliferation in RCC cells (Figure 7).

The ability of SOSTDC1 to simultaneously antagonize both Wnt and BMP

signaling may explain its enhanced antiproliferative activity when compared to

noggin and DKK alone (Figure 7).

In this manuscript, we identify SOSTDC1 as a BMP antagonist that is

down regulated in renal cancer, particularly clear cell carcinoma, the most

common kidney cancer. In contrast, expression of SOSTDC1 was not decreased

in papillary and chromophobe kidney tumors, which have a more favorable

prognosis than clear cell carcinoma. We show a pattern of extracellular secretion

in human tissue sections. We observe that reestablishment of SOSTDC1

71

expression in clear cell renal cancer cells profoundly inhibits proliferation of these

cells, suggesting that SOSTDC1 is involved in regulating the proliferative

capacity of these cancer cells. Finally, we show that SOSTDC1 expression leads

to inhibition of both BMP and Wnt signaling. The observations that SOSTDC1 is

significantly downregulated in clear cell renal carcinoma and that overexpression

of SOSTDC1 inhibits proliferation of clear cell carcinoma cells suggest that

modulation of SOSTDC1 may represent a novel therapeutic approach for renal

cancers.

72

ACKNOWLEDGEMENTS

1. KRB is supported by the Department of Defense Breast Cancer Research

Program under award number W81XWH-05-1-0287. Views, opinions, and

endorsements by the author(s) do not reflect those of the US Army or the

Department of Defense.

2. Supported by NIH R21 CA119181 (SVT). The authors also gratefully

acknowledge support from the Ben Mynatt family and the Brown

Foundation.

73

References

Alarmo, E.L., Rauta, J., Kauraniemi, P., Karhu, R., Kuukasjarvi, T., and Kallioniemi, A. (2006). Bone morphogenetic protein 7 is widely overexpressed in primary breast cancer. Genes Chromosomes Cancer 45, 411-419.

Alberici, P., Jagmohan-Changur, S., De Pater, E., Van Der Valk, M., Smits, R., Hohenstein, P., and Fodde, R. (2006). Smad4 haploinsufficiency in mouse models for intestinal cancer. Oncogene 25, 1841-1851.

Aubin, J., Davy, A., and Soriano, P. (2004). In vivo convergence of BMP and MAPK signaling pathways: impact of differential Smad1 phosphorylation on development and homeostasis. Genes Dev 18, 1482-1494.

Avsian-Kretchmer, O., and Hsueh, A.J. (2004). Comparative genomic analysis of the eight-membered ring cystine knot-containing bone morphogenetic protein antagonists. Mol Endocrinol 18, 1-12.

Brugarolas, J. (2007). Renal-cell carcinoma--molecular pathways and therapies. N Engl J Med 356, 185-187.

Chen, B., Athanasiou, M., Gu, Q., and Blair, D.G. (2002). Drm/Gremlin transcriptionally activates p21(Cip1) via a novel mechanism and inhibits neoplastic transformation. Biochem Biophys Res Commun 295, 1135-1141.

Chen, D., Zhao, M., and Mundy, G.R. (2004). Bone morphogenetic proteins. Growth Factors 22, 233-241.

Feeley, B.T., Krenek, L., Liu, N., Hsu, W.K., Gamradt, S.C., Schwarz, E.M., Huard, J., and Lieberman, J.R. (2006). Overexpression of noggin inhibits BMP-mediated growth of osteolytic prostate cancer lesions. Bone 38, 154-166.

Franzen, A., and Heldin, N.E. (2001). BMP-7-induced cell cycle arrest of anaplastic thyroid carcinoma cells via p21(CIP1) and p27(KIP1). Biochem Biophys Res Commun 285, 773-781.

Fukuda, N., Saitoh, M., Kobayashi, N., and Miyazono, K. (2006). Execution of BMP-4-induced apoptosis by p53-dependent ER dysfunction in myeloma and B-cell hybridoma cells. Oncogene 25, 3509-3517.

Ghosh-Choudhury, N., Woodruff, K., Qi, W., Celeste, A., Abboud, S.L., and Ghosh Choudhury, G. (2000). Bone morphogenetic protein-2 blocks MDA MB 231 human breast cancer cell proliferation by inhibiting cyclin-dependent kinase-mediated retinoblastoma protein phosphorylation. Biochem Biophys Res Commun 272, 705-711.

74

Guidato, S., and Itasaki, N. (2007). Wise retained in the endoplasmic reticulum inhibits Wnt signaling by reducing cell surface LRP6. Dev Biol 310, 250-263.

Gumz, M.L., Zou, H., Kreinest, P.A., Childs, A.C., Belmonte, L.S., LeGrand, S.N., Wu, K.J., Luxon, B.A., Sinha, M., Parker, A.S., Sun, L.Z., Ahlquist, D.A., Wood, C.G., and Copland, J.A. (2007). Secreted frizzled-related protein 1 loss contributes to tumor phenotype of clear cell renal cell carcinoma. Clin Cancer Res 13, 4740-4749.

Haudenschild, D.R., Palmer, S.M., Moseley, T.A., You, Z., and Reddi, A.H. (2004). Bone morphogenetic protein (BMP)-6 signaling and BMP antagonist noggin in prostate cancer. Cancer Res 64, 8276-8284.

He, X.C., Zhang, J., Tong, W.G., Tawfik, O., Ross, J., Scoville, D.H., Tian, Q., Zeng, X., He, X., Wiedemann, L.M., Mishina, Y., and Li, L. (2004). BMP signaling inhibits intestinal stem cell self-renewal through suppression of Wnt-beta-catenin signaling. Nat Genet 36, 1117-1121.

Hsu, M.Y., Rovinsky, S., Penmatcha, S., Herlyn, M., and Muirhead, D. (2005). Bone morphogenetic proteins in melanoma: angel or devil? Cancer Metastasis Rev 24, 251-263.

Iliopoulos, O. (2006). Molecular biology of renal cell cancer and the identification of therapeutic targets. J Clin Oncol 24, 5593-5600.

Itasaki, N., Jones, C.M., Mercurio, S., Rowe, A., Domingos, P.M., Smith, J.C., and Krumlauf, R. (2003). Wise, a context-dependent activator and inhibitor of Wnt signalling. Development 130, 4295-4305.

Kassai, Y., Munne, P., Hotta, Y., Penttila, E., Kavanagh, K., Ohbayashi, N., Takada, S., Thesleff, I., Jernvall, J., and Itoh, N. (2005). Regulation of mammalian tooth cusp patterning by ectodin. Science 309, 2067-2070.

Kawabata, M., Imamura, T., and Miyazono, K. (1998). Signal transduction by bone morphogenetic proteins. Cytokine Growth Factor Rev 9, 49-61.

Laurikkala, J., Kassai, Y., Pakkasjarvi, L., Thesleff, I., and Itoh, N. (2003). Identification of a secreted BMP antagonist, ectodin, integrating BMP, FGF, and SHH signals from the tooth enamel knot. Dev Biol 264, 91-105.

Lehr, H.A., Mankoff, D.A., Corwin, D., Santeusanio, G., and Gown, A.M. (1997). Application of photoshop-based image analysis to quantification of hormone receptor expression in breast cancer. J Histochem Cytochem 45, 1559-1565.

Linehan, W.M., Grubb, R.L., Coleman, J.A., Zbar, B., and Walther, M.M. (2005). The genetic basis of cancer of kidney cancer: implications for gene-specific clinical management. BJU Int 95 Suppl 2, 2-7.

75

Major, M.B., Camp, N.D., Berndt, J.D., Yi, X., Goldenberg, S.J., Hubbert, C., Biechele, T.L., Gingras, A.C., Zheng, N., Maccoss, M.J., Angers, S., and Moon, R.T. (2007). Wilms tumor suppressor WTX negatively regulates WNT/beta-catenin signaling. Science 316, 1043-1046.

Miyazaki, H., Watabe, T., Kitamura, T., and Miyazono, K. (2004). BMP signals inhibit proliferation and in vivo tumor growth of androgen-insensitive prostate carcinoma cells. Oncogene 23, 9326-9335.

Miyazono, K., Kusanagi, K., and Inoue, H. (2001). Divergence and convergence of TGF-beta/BMP signaling. J Cell Physiol 187, 265-276.

Miyazono, K., ten Dijke, P., and Heldin, C.H. (2000). TGF-beta signaling by Smad proteins. Adv Immunol 75, 115-157.

Moustakas, A., and Heldin, C.H. (2005). Non-Smad TGF-beta signals. J Cell Sci 118, 3573-3584.

Moustakas, A., Souchelnytskyi, S., and Heldin, C.H. (2001). Smad regulation in TGF-beta signal transduction. J Cell Sci 114, 4359-4369.

Pardali, K., Kowanetz, M., Heldin, C.H., and Moustakas, A. (2005). Smad pathway-specific transcriptional regulation of the cell cycle inhibitor p21(WAF1/Cip1). J Cell Physiol 204, 260-272.

Pouliot, F., and Labrie, C. (2002). Role of Smad1 and Smad4 proteins in the induction of p21WAF1,Cip1 during bone morphogenetic protein-induced growth arrest in human breast cancer cells. J Endocrinol 172, 187-198.

Rasband, W.S. (1997-2007). ImageJ, Bethesda, Maryland, USA: National Institutes of Health.

Reddi, A.H. (2001). Bone morphogenetic proteins: from basic science to clinical applications. J Bone Joint Surg Am 83-A Suppl 1, S1-6.

Sneddon, J.B., Zhen, H.H., Montgomery, K., van de Rijn, M., Tward, A.D., West, R., Gladstone, H., Chang, H.Y., Morganroth, G.S., Oro, A.E., and Brown, P.O. (2006). Bone morphogenetic protein antagonist gremlin 1 is widely expressed by cancer-associated stromal cells and can promote tumor cell proliferation. Proc Natl Acad Sci U S A 103, 14842-14847.

ten Dijke, P., Yamashita, H., Sampath, T.K., Reddi, A.H., Estevez, M., Riddle, D.L., Ichijo, H., Heldin, C.H., and Miyazono, K. (1994). Identification of type I receptors for osteogenic protein-1 and bone morphogenetic protein-4. J Biol Chem 269, 16985-16988.

Urist, M.R. (1965). Bone: formation by autoinduction. Science 150, 893-899.

76

van Bezooijen, R.L., Svensson, J.P., Eefting, D., Visser, A., van der Horst, G., Karperien, M., Quax, P.H., Vrieling, H., Papapoulos, S.E., ten Dijke, P., and Lowik, C.W. (2007). Wnt but not BMP signaling is involved in the inhibitory action of sclerostin on BMP-stimulated bone formation. J Bone Miner Res 22, 19-28.

Veeman, M.T., Slusarski, D.C., Kaykas, A., Louie, S.H., and Moon, R.T. (2003). Zebrafish prickle, a modulator of noncanonical Wnt/Fz signaling, regulates gastrulation movements. Curr Biol 13, 680-685.

Vitt, U.A., Hsu, S.Y., and Hsueh, A.J. (2001). Evolution and classification of cystine knot-containing hormones and related extracellular signaling molecules. Mol Endocrinol 15, 681-694.

Wozney, J.M. (1992). The bone morphogenetic protein family and osteogenesis. Mol Reprod Dev 32, 160-167.

Yamashita, H., ten Dijke, P., Huylebroeck, D., Sampath, T.K., Andries, M., Smith, J.C., Heldin, C.H., and Miyazono, K. (1995). Osteogenic protein-1 binds to activin type II receptors and induces certain activin-like effects. J Cell Biol 130, 217-226.

Yanagita, M. (2005). BMP antagonists: their roles in development and involvement in pathophysiology. Cytokine Growth Factor Rev 16, 309-317.

Yanagita, M. (2006). Modulator of bone morphogenetic protein activity in the progression of kidney diseases. Kidney Int 70, 989-993.

Yanagita, M., Oka, M., Watabe, T., Iguchi, H., Niida, A., Takahashi, S., Akiyama, T., Miyazono, K., Yanagisawa, M., and Sakurai, T. (2004). USAG-1: a bone morphogenetic protein antagonist abundantly expressed in the kidney. Biochem Biophys Res Commun 316, 490-500.

Yanagita, M., Okuda, T., Endo, S., Tanaka, M., Takahashi, K., Sugiyama, F., Kunita, S., Takahashi, S., Fukatsu, A., Yanagisawa, M., Kita, T., and Sakurai, T. (2006). Uterine sensitization-associated gene-1 (USAG-1), a novel BMP antagonist expressed in the kidney, accelerates tubular injury. J Clin Invest 116, 70-79.

77

CHAPTER III

LOSS OF HETEROZYGOSITY AFFECTING SOSTDC1 IN RENAL TUMORS

Kimberly R. Blish, Abdoulaye Diallo, Shelley Smith, Seon-Hi Jang, Julie C.

Brown, A. J. Garvin, M.C. Willingham, Gregory A. Hawkins, Frank M. Torti, Suzy V. Torti

The following manuscript is for submission to Cancer Epidemiology, Biomarker,

and Prevention. Stylistic variations are due to the requirements of the journal.

Figure and manuscript preparation by K.R. Blish. Sample preparation,

experiments performed and data gathered by K.R. Blish with DNA sequencing

assistance from S. Jang, J.C. Brown, A. Diallo and LOH analysis from S. Smith.

A.J. Garvin and M.C. Willingham served as expert consultants on Wilms Tumor

and experiment design. G.A. Hawkins, F.M. Torti, and S.V. Torti served advisory

and editorial roles related to experimental design, data analysis, and manuscript

preparation.

78

Abstract Loss of heterozygosity (LOH) at 7p21-22 affects 10% of pediatric Wilms renal

tumors. SOSTDC1, a bone morphogenetic protein antagonist dysregulated in

adult renal tumors, lies within this area of observed LOH. We hypothesized that

LOH or other genetic abnormalities in this area may affect SOSTDC1 in Wilms

and adult renal tumors. Normal and tumor tissue from 25 Wilms patients, 34

adult renal clear cell carcinoma (RCC) patients, and 2 adult oncocytoma patients

were used to obtain genomic DNA. Fifty-one single nucleotide polymorphisms

(SNPs) were genotyped across approximately 2 million bases (MB) surrounding

SOSTDC1 to assess LOH. Direct sequencing of SOSTDC1 exons was also

performed.

We observed LOH within all or part of the 2 MB region in 4/25 (16%) of Wilms

tumors. Additionally, LOH was observed in 5/36 (13.8 %) of adult RCC samples.

Of these samples, LOH within SOSTDC1 was observed in 3 Wilms tumors and 2

adult samples- 1 RCC and 1 oncocytoma. Interestingly, LOH was discontinuous

throughout this region. Direct sequencing confirmed LOH observations. No

additional genetic variations were found from direct sequencing. We conclude

from these studies that LOH in the short arm of chromosome 7 occurs in adult as

well as pediatric renal tumors, and includes SOSTDC1.

79

Introduction

Renal tumors affecting both adults and children are often idiopathic in origin. The

clinical presentation, disease history, and treatments of renal tumors differ

between children and adults. In children, the majority of renal masses are

pediatric Wilms tumors (referred to less often as nephroblastomas). Wilms tumor

is the 6th most common malignancy of childhood; affecting approximately 500

children in the U.S. per year (1). Children most often present with palpable

abdominal mass. While lesions respond quite well to treatment, the causes

contributing to the tumor are not fully known (2). 1-2% of tumors are associated

with familial predisposition and are associated with loci at 17q, FWT1 (3), and

19q FWT2 (4). Over 90% of Wilms tumor cases are sporadic in origin, with the

remaining 5-10 % of tumors arising from developmental syndromes which

include Beckwith-Wiedemann syndrome, hemihypertrophy, Denys-Drash

syndrome, and WAGR (Wilms tumor, Aniridia, Genitourinary malformation, and

mental Retardation) syndrome. These syndromes have been instrumental in

identifying Wilms tumor suppressor loci (5): WT1 gene, a zinc-finger transcription

factor at 11p13 (6, 7) and the putative WT2 locus containing multiple genes at

11p15 (8). Additional studies have highlighted an increased incidence of loss of

heterozygosity (LOH) at 16q (9), 1p (10), and 7p (11-13) as other potential loci

for Wilms tumor suppressor genes.

Unlike pediatric tumors, adult renal cancers can be difficult to detect and treat.

The most prevalent type of renal tumor is renal clear cell carcinoma (RCC, 80-85

80

% of cases); followed by papillary (5-10%), chromophobe, medullary, and

oncocytic (<5%) types. Incidence and mortality from renal carcinoma has

increased in the past two decades, in part due to the asymptomatic nature of the

tumor and lack of reliable early detection strategies (14). Similar to Wilms

tumor, familial tumor syndromes have aided in the discovery of certain genes

involved in the pathogenesis of RCC. These include von Hippel-Lindau on 3p14

(VHL, (15); hereditary papillary renal carcinoma, c-met, on 7q (16); Birt-Hogg-

Dube on 17p (BHD; (17); and fumarate hydratase (FH, (18)). Other genetic

abnormalities noted in kidney cancer include trisomy of 5q, 12, and 20; and

partial or total loss of chromosomes 7, 8, 9, 13q, 14q (19).

Regions of LOH associated with both pediatric and adult renal cancers represent

candidate regions for the location of tumor suppressors whose inactivation may

be critical for the initiation or progression of renal cancer. In both adult and

pediatric tumors, cytogenetic changes have been noted on the short arm of

chromosome 7. These include a 10% incidence of LOH on 7p in Wilms Tumor

(12), loss of 7p and duplication of 7q (20) and consistent gains of chromosome 7

in adult late stage RCC (21) and RCC-papillary subtypes (22, 23). In Wilms

Tumor, the smallest consensus region of LOH has been narrowed to a 2 million

base pair (MB) segment within 7p21 (12) containing 10 known genes, including

the aryl hydrocarbon receptor (AHR), the mesenchyme homeobox 2 gene

(MEOX2), and SOSTDC1.

81

In this study, we focused on SOSTDC1. SOSTDC1 encodes an extracellular

human bone morphogenetic protein (BMP) antagonist (24, 25) that is

predominantly expressed in the renal epithelia of the distal tubules, collecting

ducts, and urothelium (26). We have previously shown the SOSTDC1 product is

downregulated in approximately 90% of clear cell RCC tumors (14). SOSTDC1

affects two key signaling pathways in the kidney: BMP and Wnt. As changes in

BMP signaling have been noted in a variety of tumors (27-29), including renal

tumors (30), an extracellular modulator of BMP signaling could have potential

tumor suppressor roles within normal kidney epithelia. Similarly, dysregulation of

the Wnt pathway often plays a role in tumorigenesis (31). In particular, in Wilms

tumor, mutations have been observed in β-catenin, the main intracellular effector

of classical Wnt signaling (5); alterations in Wnt signaling have also been

implicated in adult renal carcinoma (32).

As SOSTDC1 is a secreted molecule with the potential to modify two cell

signaling pathways that are critical to renal development and function, we asked

whether SOSTDC1 is contained within a region of LOH at 7p noted in kidney

tumors.

82

Materials and Methods

Collection of Tissues

Approval was obtained from the Institutional Review Board at Wake Forest

University for the retrieval of matched normal and tumor tissues from the Tumor

Bank of the Wake Forest University Comprehensive Cancer Center. Matched

tissues were collected for 36 adult kidney cancer patients and 7 pediatric Wilms

tumor patients. Information concerning the patients’ primary diagnosis was

collected; however no patient identifiers were obtained. An additional 18

matched normal and Wilms tumor tissues were obtained from the Cooperative

Human Tissue Network (CHTN), which is funded by the National Cancer

Institute. Other investigators may have received samples from these same

tissues. Diagnosis and treatment information were made available for each

tissue. All tissues were collected as snap frozen specimens stored at -80°C.

Sample preparation and Genomic DNA isolation

Each snap frozen tissue was sectioned on a bed of dry ice to ensure minimal

thawing during sample preparation. An approximately 30-50 mg piece of tissue

was cut and an adjacent piece of tissue was cut and embedded in OCT

cryomounting material (Tissue-Tek, Elkhart, IN) to make frozen blocks for future

cryosectioning and histology if desired. Genomic DNA was isolated from tissue

samples via homogenization in ice cold lysis buffer (10mM Tris pH 8.0, 0.1M

EDTA, 0.5% SDS, 100 µg/ml Proteinase K, 25 µg/ml RNAase); followed by

subsequent phenol chloroform extraction as has been previously described (33).

83

Integrity and concentration of each resulting DNA sample was assessed by

agarose gel electrophoresis.

Sequencing Primer Design

The known coding region of SOSTDC1 is contained in two exons. However,

other potentially transcribed areas have been identified in the University of

California Santa Clara Genome database (UCSC; http://genome.ucsc.edu, (34,

35)). Two of these areas occur upstream of the coding region and one small

area occurs between the known coding exons for a total of 5 putative exons or

regulatory regions at this locus (Figure 1). Primers were designed for direct

sequencing of each putative exon or regulatory region for a total of 14 pairs direct

sequencing primers (Table S1). All primers were synthesized by IDT (Coralville,

Iowa).

PCR Amplification and Direct Sequencing

Each direct sequencing primer pair was used to amplify all 5 putative regions of

interest in each normal and tumor sample using PCR. PCR and sequencing

reactions were performed with 60 ng of genomic DNA template at 60°C for 40

cycles. Resulting data from the ABI 3730XL was analyzed using Sequencher

software (GeneCodes, Ann Arbor, MI v. 4.7).

84

Loss of Heterozygosity Analysis

To examine the area surrounding SOSTDC1 for loss of heterozygosity, we

genotyped single nucleotide polymorphisms (SNPs). All SNPs within the 2.4 MB

region as found at HapMap (www.hapmap.org) were examined and 51 SNPs

with high percentage of heterozygosity (>0.45) were chosen. These SNPs are

fairly well distributed over the region (Figure 2). Primers for each SNP were

designed for analysis on the MassARRAY system (Sequenom, San Diego, CA

(Table S2). All primers were synthesized by IDT. 5 ng genomic DNA from each

sample was used in the genotyping reactions.

85

Figure 1.

Figure 1. SOSTDC1 locus. Summarized from the Genome Browser hosted

at UCSC. Diagram shows relative positions of designed primer pairs to

potential regions of interest within the gene (arrows). Gene translation

start codon is in exon 3 and stop codon is in exon 5. All known coding

sequence in contained in exons 3 and 5.

(Next page)

Figure 2. Map of known gene loci over 2 million base pair region of interest

in LOH study. Approximate locations of genes and the 51 SNP markers

used in this study shown on chromosome 7p21 map from physical location

15400000 to 18000000. Terminal location is at the top, while centrosomal

end of map is at the bottom.

86

Figure 2.

87

Results and Discussion

Loss of Heterozygosity in 7p21 Wilms tumors

In Wilms tumors, SNP genotyping over the 2 MB region at 7p21.1 to 7p21.2

revealed LOH in 3 of the 25 tumors we genotyped (Figure 3). These samples

included a patient with hemihypertrophy being evaluated for Beckwidth-

Wiedemann syndrome with a Stage II tumor that showed complete LOH at every

informative SNP in the region (W733); a patient with multifocal Wilms also

showing complete LOH at every informative SNP (W8188); and a patient with

anaplastic Wilms (W8194), showing one instance of LOH at SNP rs15711849,

near MEOX2.

Direct sequencing of the SOSTDC1 allele revealed one additional patient,

W8197, with one instance of LOH affecting the 3’ UTR in Exon 5 of SOSTDC1;

all other sequences in this patient showed no informative SNPs. Direct

sequencing also confirmed that LOH directly affects SOSTDC1 in patients W733

and W8188, as every heterozygous SNP in the normal was lost in the tumor

(Table 1). Patient W8194 had no informative SNPs seen in the direct sequence

of SOSTDC1, so it was not possible to ascertain whether this patient exhibited

LOH at SOSTDCI. Sequence analysis revealed no mutations within known

exons (3 and 5) or candidate exons (1, 2, and 4) of the remaining SOSTDC1

allele.

88

Figure 3.

Figure 3. LOH analysis in 2MB region. Results from positive samples

aligned with the 7p21.1 to 7p21.2 SNP map from Figure 2. For each

patient’s row, black boxes indicate regions where all genotyped SNPS

show LOH in the tumor samples. Gray blocks indicate regions of

uninformative SNPs in between observed LOH; therefore, these areas may

be regions of LOH as well. Unmarked areas of each sample indicate

informative SNPs where no LOH was observed. The dotted lines highlight

the region covered by SOSTDC1. We note 3 samples (2 Wilms and 1 RCC

show a large region of LOH that includes either the entire genotyped

region, or a ~1 MB region including SOSTDC1 (RCC-614). LOH does not

appear to center around a particular gene. RCC = adult renal carcinoma

samples; W = pediatric Wilms tumor samples.

89

Table 1. Results of direct sequencing of SOSTDC1 locus. All SNPs found

in the direct sequences are summarized here. All other samples

sequences showed no LOH or other mutations. SNP location relative to

sequenced exons and chromosome 7 base pair location provided where

known. The existence of heterozygous SNPs (informative, but no LOH

present) in the sample is shown via yes/no designation. RCC = adult renal

carcinoma samples, W = pediatric Wilms tumors.

Table 1. Direct Sequencing of SOSTDC1 Results Sample Location Informative SNPs

without LOH Normal Tumor

RCC-129 End of Exon 1- rs35324397 Yes A/G G RCC-614 Beginning of Exon 1:

16,536,670; 16,536,667 between rs10240242 and

rs35324397

Yes G/T, A/G T, A

RCC-614 Beginning of Exon 1: 16,536,641 between

rs10240242 and rs35324397

Yes C/G C

RCC-614 End of Exon 1- rs35324397 Yes C/G C RCC-614 End of Exon 1- 5 bp

downstream of rs35324397 Yes A/G G

RCC-635 Beginning of Exon 1: 16,536,641 between

rs10240242 and rs35324397

Yes C/G C

RCC-737 Exon 5: 16468252 closest to rs 6959246

Yes G/T T

W-733 Before Exon 1: rs7781903 No C/T C W-733 Beginning of Exon 1 between

rs10240242 and rs35324397 No C/G G

W-733 Beginning of Exon 2: rs7801569

No C/T C

W-8188 Beginning of Exon 2: rs7801569

No C/T C

W-8197 Exon 5: 16468252 closest to rs 6959246

No G/T T

90

To assess whether expression levels of SOSTC1 are affected in Wilms tumor

patients, we queried the Oncomine database (36). As seen in Figure 4, there is

a highly significant (p< 0.00001) loss of SOSTDC1 in Wilms Tumors compared to

normal fetal kidney and adult clear cell sarcomas (median of normal tissue =

1.116 compared to Wilms media value of -0.948, Figure 4).

Loss of Heterozygosity in adult renal tumors

Three of the 36 adult patients we analyzed showed LOH in the 2MB region of

interest (Figure 3). Two of these patients had clear cell renal carcinoma (RCC-1,

RCC-614); while one had a far less common oncocytoma (RCC-635). Patient

RCC-614 showed LOH over much of the area, while RCC-1 and RCC-635

showed LOH at approximately 15-20% of informative SNPs.

Direct sequencing of SOSTDC1 exons in adult tumors also showed LOH in

patients RCC-614 and RCC-635 in several locations of exon 1. Additionally,

patients RCC-129 and RCC-737 also showed LOH in one SNP each. Similar to

what was observed on the large scale LOH study, all of the adult tumors showed

LOH at some heterozygous loci, but not others- even within the SOSTDC1 gene

itself- while all of the Wilms tumors showed complete LOH at every heterozygous

allele. Among all samples (adult and pediatric), LOH within SOSTDC1 was

observed mostly in exon 1; with no observed heterozygosity in the known

transcribed parts of the gene. Whether adult or Wilms, for each SNP that

91

Figure 4. Oncomine database shows significant SOSTDC1 downregulation

in adult renal clear cell tumors and pediatric Wilms tumors. The

ONCOMINE database was queried for all studies involving markers in

SOSTDC1. Results of studies by Cutcliffe et al (Wilms tumors vs. normal or

adult tumors, (43)) and Bittner et al (Adult clear cell carcinoma vs. other

RCC subtypes,

http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE2109 ) were

compared using the software available on the site. Dots above and below

the boxes show sample maximum and minimum values respectively. The

horizontal lines show the spread of the values from starting at the 10%

value through the 90%, with the box highlighting the range of 25% to 75%.

Dark boxes show the normal or control tissues for each study, white boxes

show adult clear cell RCC and Wilms tumor respectively. The horizontal

black bar through each box shows the median value for the sample. ** = p<

0.00001, normal or control tissue compared to adult RCC or Wilms tumors.

92

Figure 4.

93

showed LOH in more than one sample, the same allele was lost. For example,

at the beginning of exon (position 16,536,641) the G is lost out of the C/G.

Comparison between LOH in Wilms and adult renal tumors

Overall, we observed LOH at the SOSTDC1 gene in 4/25 (16%) of Wilms tumor

patients. This frequency is comparable to WT1, the prototype of Wilms tumor

suppressor genes, which occurs in sporadically in Wilms tumor, often associated

with WAGR Syndrome. We also observed LOH including the SOSTDC1 in adult

RCC at a similar frequency (5/36 or 14%). It should be borne in mind that current

treatment of Wilms tumors sometimes involves pre-surgical chemotherapy or

radiation. Differences in treatment status within the patient population may have

effects on the resulting tissues used to obtain genomic DNA and thus the results

of the LOH studies.

LOH in Wilms tumors appears to occur in large sections on the short arm of

chromosome 7, as seen in W733 and W8188. This is concordant with previous

studies. Interestingly, two patients (W8194 and W8197) showed examples of

just one instance of LOH each. Due to distances between LOH markers for

patient W8194 (approximately 100 KB), and a lack of informative SNPs in

SOSTDC1, it is hard to estimate whether this region of LOH extends beyond the

SOSTDC1 locus. Patient W8197 showed 1 incidence of LOH in the direct

sequence; since no other informative SNPs were found within the direct

sequence, this may represent either LOH affecting SOSTDC1 or a point

94

mutation. It is noteworthy that tumor size, stage, histology, and treatment status

varied among these patients.

In contrast to what we observed in Wilms tumors, in adult tumors, regions of LOH

were noncontiguous or “patchy” as SNPs showing LOH were broken up by

heterozygous alleles. Due to the high incidence of aneuploidy in these tumors,

this phenomenon may be partially explained by chromosomal copy number

variation. Indeed, multiple studies referenced in the Database of Genomic

Variants show variations in copy number that affect parts of the 2 MB region;

including the area around SOSTDC1 (http://projects.tcag.ca/cgi-

bin/variation/tbrowse?source=hg18&table=Locus&rnum=50&rstart=101; 37).

We have previously reported downregulation of both the message (90% of

patients) and protein of SOSTDC1 in RCC-clear cell tumors. Results presented

here demonstrate for the first time that LOH in SOSTDC1 affects pediatric as well

as adult (Fig. 3) renal tumors, and that SOSTDC1 expression is downregulated in

pediatric tumors (Fig. 4). However, while LOH may play a role in the regulation

of this locus in some patients, other mechanisms, including epigenetic regulation,

must also be considered. For example, promoter methylation has been shown to

have an important role in regulation of the IGF2 gene (38-40) and loci at 11p13

and 11p15 (41) in Wilms tumor. Improper splicing, a mechanism that contributes

to dysregulation of the Wilms tumor suppressor gene WT1, must also be

considered (42). Future studies on the role and regulation of SOSTDC1 in

95

pediatric and adult kidney cancer may be particularly important given the

potential utility of SOSTDC1 (or SOSTDC1 mimetics) as a therapeutic agent that

targets both BMP and Wnt pathway signaling.

96

Supplemental Data

Table S1. Direct Sequencing Primers for SOSTDC1. All primers designed

to potential exons or regulatory regions of SOSTDC1 and were optimized

for 60°C reaction temperatures. Target exon, forward and reverse primer

sequence and amplicon size shown.

Table S1. Primers for Direct Sequencing of SOSTDC1

Title (Exon, #) Sequence Amplicon

1-1 F CCAGCCATTCTACCTCCAGG 1-1 R TGAAGTGTGTGCATTTTGTATTCA 785 1-2 F TGCATAGTGTTTGGGGTGG 1-2 R TGAAGTGTGTGCATTTTGTATTCA 877 2-1 F GGCAGTTCCCCTGCACAT 2-1 R GGCCTGAAGGGAGGTGAAG 604 2-2 F TGGTTTGACCAGTCCCCACT 2-2 R ATAGTTCTCCACACAATCTCCTCA 656 3-1 F TAGATTCAGGAAAGGAAATGGC 3-1 R ACTTACTGTTCCGATCCAGTCC 572 3-2 F TCCCACCCCTTCTCTGTGTT 3-2 R ATGGTCATTTTGCATGATTTTG 680 3-3 F ACACCTGAATGAACGCCAAACCTC 3-3 R TAGGGAAGAATGCCAACCTGCACA 495 4 F CTTACACAAATCTTTTGCCTCTCC 4 R ATCAGGAGTTTCACTTCATCTCTG 520

5-1 F CATGAAAGTGTCCCTATACTATCCA 5-1 R CTAACTCATGCTGTGCTTGCT 597 5-2 F GTACTGGAGCAGGAGGAGCT 5-2 R AGGAAGATCACTCATGGCTGC 750 5-3 F TCAGGACCTTCTTTGGGAATAG 5-3 R GGTCAAGACACCTTCTGATTGC 810 5-4 F CGCTTGGAATGGAATGCC 5-4 R AATGAGCAGCAGACTTGGCA 668 5-5 F CCTGCCAGTGCTCCCTAACT 5-5 R CATTCCAAGCGAGGGTCAG 902

97

Table S2. Primers for LOH SNP Genotyping. All primers designed for use

on the Sequenom MassARRAY platform. Percentage of heterozygosity

(informative SNPs) shown in two populations from the International

HapMap Project: CEU = Utah residents with Northern and Western

European Ancestry; YRI = Samples from Yoruban descent Ibadan, Nigeria.

UEP = Unextended Primer.

Table S2. Primers for Loss of Heterozygosity Analysis

Marker % Informative Sequence CEU YRI

rs4598143

53.4 64.5

1-ACGTTGGATGCAGTGTTTTTCTTTACAGAG 2- ACGTTGGATGGAGAACTCAATTTGCTTCCTC

UEP- TTGCTTCCTCTTGTTTATCCTTC rs6944475

55 69.5

1- ACGTTGGATGGAGACTATATTTTCTAGCCCC 2- ACGTTGGATGGTTATTATCACTCACAGTGC

UEP- gggAACAGTTCCGTAAACCTATCAAG rs11982985

45.6 90.8

1- ACGTTGGATGGCATTACTTATCTGTAGTTTC 2- ACGTTGGATGACCAGAGCTGGCATTCTTTC

UEP- ctGTGTTTGCCCTCTCTTT rs7783003

47.4 81.4

1- ACGTTGGATGGTCCGGATTCCATACTTTGC 2- ACGTTGGATGTCCCTGGTAACCTTACCTAC

UEP- gcACCTTACCTACAATTGTTTACATA rs10950562

47.4 76.3

1- ACGTTGGATGATGGTACCAGCTCGTCTTTG 2- ACGTTGGATGCCTACCAACCAAAAAACGCC

UEP- AAACGCCTAGGATCAAAC rs6946099

45.8 87.5

1- ACGTTGGATGGACAAACTCAGGTATTATG 2- ACGTTGGATGCGACCCCAATTACTTCCTAC

UEP- TGTCAGACTATATTTTCCTCTTCA rs6942413

48.3 27.6

1- ACGTTGGATGCTCCCGAGAAAACTTGTTGG 2- ACGTTGGATGCCTCCCTCCCTCAATTTATG

UEP- cccCCCTCAATTTATGAATTACAGAA rs37410

50.8 17.5

1- ACGTTGGATGGAGTTCATAGTGAATGATTCC 2- ACGTTGGATGCCCCAAACCCAATAAACCAC

UEP- AACCACTAATCTGCTTTCAGTC rs38176

53.3 95.8

1- ACGTTGGATGCTCAACGTTAAGCATGGTTC 2- ACGTTGGATGGAGCTCTGCAGAATGGAAAC

UEP- GAAACCCACTTCAAGTTGT rs13231687

47.5 60

1- ACGTTGGATGAGTTGAACTGTAGCATCCGC 2- ACGTTGGATGCTGCGAATTCACAGCAGTAG

UEP- CAGCAGTAGGTTGGGAT

98

rs9648211

55 88.3

1- ACGTTGGATGTGTTGCCCTATCTCATAAGC 2- ACGTTGGATGCCTTTGCTAACCTTACGGAC

UEP- GGACCATTTCCCAATACTG rs10277151

50 75.4

1- ACGTTGGATGCATTGCAGAATGTTCTCTCC 2- ACGTTGGATGCCATAATAAAGCTGGAAGAG

UEP- ttgGCTGGAAGAGTATTGACTCAG rs7777150

53.4 70.8

1- ACGTTGGATGATAAAAGGGTGGTACAAAGG 2-ACGTTGGATGTCTTTACATCCCAGAGTAGG

UEP- TCCCAGAGTAGGAATCACCTAT rs6963052

54.6 76.8

1- ACGTTGGATGCATGTCTAAACCACTGGAAA 2- ACGTTGGATGTAGCAGGTACACTTGCTCAC

UEP- cTGCTCACTGATATCCCTCTT rs12699778

47.5 96.7

1- ACGTTGGATGGAAGAAAGGAATGGAAGCAG 2- ACGTTGGATGCTTCCTGGATCTTTGGTTGG

UEP- TTGGTTGGTTTCTAACATTGATT rs2178598

46.6 47.1

1- ACGTTGGATGGGGTATCTCTGAGTTCTGTC 2- ACGTTGGATGGCCACAATAATCTCAACAGG

UEP- CTCAACAGGAAAGATTTCATATAA rs2704674

52.5 46.7

1- ACGTTGGATGTGGAAGTTGTCCATGTGCTC 2- ACGTTGGATGTGGTAGACACAGAGATGCAC

UEP- cCAAGGAAGGGCTTGTTG rs10270965

56.6 48

1- ACGTTGGATGCTATTTATAATGCAGAAACC 2- ACGTTGGATGTACAGGCATGAGCCATCGT

UEP- atAGTTTTCCTGGTTTTTTTATATTG rs6959566

45 23.3

1- ACGTTGGATGGGGACAGTGTAAAGCACAAT 2- ACGTTGGATGCTGGAGATAGTCTGACTAGC

UEP- TGTTTGACCTCAAGAAAAATT rs1524362

54.4 55.9

1- ACGTTGGATGTCACTTTTGTCATGTCTTG 2- ACGTTGGATGTGTTAATAGGCCACATGACC

UEP- TCTGTCTGTAATTTTCCATGA rs6968649

55 40

1- ACGTTGGATGGTATGGAGCAGAAGTAGAAG 2- ACGTTGGATGGCTTGCTAGGCTTTCCAAAC

UEP- cACCTGGCCTAGCCCTTTTGC rs1524358

45.8 40.8

1- ACGTTGGATGTTCGTCTTCTCTTTCCCCTC 2- ACGTTGGATGGTTACTTAACCAAGGATTAGC

UEP- AACCAAGGATTAGCAGAAAAA rs578621

49.2 62.5

1- ACGTTGGATGGGCGGAGTGATTCAAAATAG 2- ACGTTGGATGCAAAACAGTTTTGACAGGATG

UEP- cACAGTTTTGACAGGATGTTCTCAT rs479202

53.9 /

1- ACGTTGGATGTGTCTTTGGCTTGGAAGTGG 2- ACGTTGGATGATGAATGTGAGAGTGGCAGG

UEP- CCAAATGCCTAGAGAGAA rs818488

45.2 8.5

1- ACGTTGGATGTATGTCCCCTTGGGAATGTG 2- ACGTTGGATGCCTGGTGATTTGGATGCAAG

UEP- ATGCAAGCTTGGGCTTA rs706059 46.6 5.3 1- ACGTTGGATGCCTGGCCTGTAGGCATTTTT

99

2- ACGTTGGATGGTACAGTGAAAAATACGAGACUEP- acAAATACGAGACAAATGTGAGAGAC

rs12112188

47.3 86.7

1- ACGTTGGATGCCTTACACAGTAACAGGGAC 2- ACGTTGGATGAGCCCACCTGGGTAATAATC

UEP- GAGACCCTTAACCACAC rs2389710

50.8 /

1- ACGTTGGATGGGCTGCAATAGATACACCAC 2- ACGTTGGATGGACATTTCCTCATTCCTCCC

UEP- cTCCTCATTCCTCCCATTATACT rs860196

45.5 0

1- ACGTTGGATGGGCACTAATCCAGGAGTTTG 2- ACGTTGGATGTGCTATGACTTACTCTTTGG

UEP- TGACTTACTCTTTGGAACTAT rs17137106

52.8 90

1- ACGTTGGATGGTTCCTATGTTATCTCTCTG 2- ACGTTGGATGTACCTATGTTTTTCTTCAC

UEP- ggTAGGTGGTACCAGGCTCA rs7799920

50.8 /

1- ACGTTGGATGTTCCTGATGTTGGCAATTAG 2- ACGTTGGATGACGAAAAGCTGTTTCTTTG

UEP- gaTGTTTCTTTGAATGATCAATAAGA rs4719491

47.5 77.5

1- ACGTTGGATGCCTACATTCTTTTTACCCAG 2- ACGTTGGATGTTGGGATGGAGTGTTGAGTG

UEP- tgTCTGAAAGTTTTCTCTTTACTCTC rs6957270

52.5 74.2

1- ACGTTGGATGAGACTTGGGAAAGCTAAGTG 2- ACGTTGGATGAGGAGAGGACTCTATGAGAC

UEP- CCAAACCTGAACTCTACCA rs11971406

45.5 72.9

1- ACGTTGGATGCAGATCCTCAGATCATCTCC 2- ACGTTGGATGAAGGAGAAGCCAAGTAGAGG

UEP- aGGAACCATGGAGCCAAGT rs12531256

52.5 70.3

1- ACGTTGGATGTTTGTGGGATTCAGCTGAAC 2- ACGTTGGATGGTTAGAAGTAAAACAGCGCC

UEP- CAGCGCCTGATTTGACC rs687265

53.3 45.8

1- ACGTTGGATGCAACAAATCTAGGAGTTG 2- ACGTTGGATGCATCATTTCTAATTGTCTTA

UEP- tgATTAGTCTATTTAGTGGCCTAT rs524715

45.7 70.3

1- ACGTTGGATGACAGTCCTTCTAACCTTCCC 2- ACGTTGGATGTAGGACTGGACCCTTAACTG

UEP- GTCAGGGTAGGCAAGAA rs652187

55 20

1- ACGTTGGATGGGAGAACAAATAAACACTG 2- ACGTTGGATGTGGTCACTTCTCTCAATTGC

UEP- ctACTCATCATTACTGTTTTTAGT rs9638747

48.3 59.1

1- ACGTTGGATGCTTAGGAATCCTCTTCTCGC 2- ACGTTGGATGACTTCCTACTTGAGAGACCC

UEP- TTGAGAGACCCCTAACATAG rs6964052

55 84.2

1- ACGTTGGATGTTGACAAGCTAATCCATAG 2- ACGTTGGATGTATCTGTAGCCTTCTCTGGG

UEP- ccgAAATGTAGACTTCCTCTCAGTTAT rs9785042

45.8 55.9 1- ACGTTGGATGGATTCTTTATCTGACATGCC

2- ACGTTGGATGGGAACAGAACAGAAAGCCCA

100

UEP- cAAGCCCAAAAATAAATCCACGCT rs12699872

54.2 90

1- ACGTTGGATGCCCTCTCCCTCTCAAATATA 2- ACGTTGGATGGTGGATGGAATTATTCAGA

UEP- GATGGAATTATTCAGAAAGTTGAG rs1562632

54.2 94.8

1- ACGTTGGATGTCTACTTGTGCCTAGATGCC 2- ACGTTGGATGACATCTTATTGCCAAGAGTG

UEP- gATTGCCAAGAGTGGCACCAGACC rs7788582

54.3 45

1- ACGTTGGATGCAGAAGGGATTATCAAGGTG 2- ACGTTGGATGAAGACTTTAGTCCTTTCACC

UEP- AAATGACCACACTAGCCAAG rs2192073

54.2 17.2

1- ACGTTGGATGTTGTCTTCTACAAGGCCTAT 2- ACGTTGGATGGGGAGGTTGAAGTGGATTAC

UEP- ccTACCCTGCTAGTGGGATTGTA rs10235024

46.7 13.3

1- ACGTTGGATGTCTAAAAAGTAGCACTTTC 2- ACGTTGGATGGTTTTGGGAGGTTATTTGAG

UEP- =TAAGGGAAGAATTAGCAATAGTTA rs12672575

47.4 33.3

1- ACGTTGGATGTTGCTCACTTGCTTGCTCTC 2- ACGTTGGATGCCTCACTTTCTGGCCAAATG

UEP- GGCAACTTACAACATGG rs10950672

45.8 61.7

1- ACGTTGGATGAGGAGATCGTCCTTGATAAC 2- ACGTTGGATGCTACTACCTCACAGAACAAG

UEP- ggTCTGGAGAAAGTTGGTAGCA rs2723548

54.2 59.3

1- ACGTTGGATGCGTCATCCTCCATTACCTTG 2- ACGTTGGATGATGGGCGAGATGAGCAGAAA

UEP- GCGAGATGAGCAGAAAACAATA rs2110013

48.3 17.9

1- ACGTTGGATGACAGCACTCTAGCCTGTGT 2- ACGTTGGATGGCTCTATTGTTTAGGATGAG

UEP- CTTCAGACAGGGTCTTG rs1051603

45.8 1.7

1- ACGTTGGATGGACCAGATACAGATGAGAAG 2- ACGTTGGATGCTGCACTCTACAATTAACGTC

UEP- ATTAACGTCTCCAAAAGTTAATA

101

References

1. Aune, G. J. Wilms tumor. Pediatr Rev, 29: 142-3; discussion 143, 2008. 2. Metzger, M. L., and Dome, J. S. Current therapy for Wilms' tumor. Oncologist, 10: 815-26, 2005. 3. Rahman, N., Abidi, F., Ford, D., Arbour, L., Rapley, E., Tonin, P., Barton, D., Batcup, G., Berry, J., Cotter, F., Davison, V., Gerrard, M., Gray, E., Grundy, R., Hanafy, M., King, D., Lewis, I., Ridolfi Luethy, A., Madlensky, L., Mann, J., O'Meara, A., Oakhill, T., Skolnick, M., Strong, L., Stratton, M. R., and et al. Confirmation of FWT1 as a Wilms' tumour susceptibility gene and phenotypic characteristics of Wilms' tumour attributable to FWT1. Hum Genet, 103: 547-56, 1998. 4. McDonald, J. M., Douglass, E. C., Fisher, R., Geiser, C. F., Krill, C. E., Strong, L. C., Virshup, D., and Huff, V. Linkage of familial Wilms' tumor predisposition to chromosome 19 and a two-locus model for the etiology of familial tumors. Cancer Res, 58: 1387-90, 1998. 5. Dome, J. S., and Coppes, M. J. Recent advances in Wilms tumor genetics. Curr. Opin. Ped., 14: 5-11, 2002. 6. Call, K. M., Glaser, T., Ito, C. Y., Buckler, A. J., Pelletier, J., Haber, D. A., Rose, E. A., Kral, A., Yeger, H., Lewis, W. H., and et al. Isolation and characterization of a zinc finger polypeptide gene at the human chromosome 11 Wilms' tumor locus. Cell, 60: 509-20, 1990. 7. Gessler, M., Poustka, A., Cavenee, W., Neve, R. L., Orkin, S. H., and Bruns, G. A. Homozygous deletion in Wilms tumours of a zinc-finger gene identified by chromosome jumping. Nature, 343: 774-8, 1990. 8. Steenman, M., Westerveld, A., and Mannens, M. Genetics of Beckwith-Wiedemann syndrome-associated tumors: common genetic pathways. Genes Chromosomes Cancer, 28: 1-13, 2000. 9. Grundy, R. G., Pritchard, J., Scambler, P., and Cowell, J. K. Loss of heterozygosity on chromosome 16 in sporadic Wilms' tumour. Br J Cancer, 78: 1181-7, 1998. 10. Grundy, P. E., Telzerow, P. E., Breslow, N., Moksness, J., Huff, V., and Paterson, M. C. Loss of heterozygosity for chromosomes 16q and 1p in Wilms' tumors predicts an adverse outcome. Cancer Res, 54: 2331-3, 1994.

102

11. Powlesland, R. M., Charles, A. K., Malik, K. T., Reynolds, P. A., Pires, S., Boavida, M., and Brown, K. W. Loss of heterozygosity at 7p in Wilms' tumour development. Br J Cancer, 82: 323-9, 2000. 12. Sossey-Alaoui, K., Vieira, L., David, D., Boavida, M. G., and Cowell, J. K. Molecular Characterization of a 7p15-21 Homozygous Deletion in a Wilms Tumor. Genes, Chromosomes & Cancer, 36: 1-6, 2003. 13. Grundy, R. G., Pritchard, J., Scambler, P., and Cowell, J. K. Loss of heterozygosity for the short arm of chromosome 7 in sporadic Wilms tumor. Oncogene, 17: 395-400, 1998. 14. Linehan, W. M., Rubin, J. S., and Bottaro, D. P. VHL loss of function and its impact on oncogenic signaling networks in clear cell renal cell carcinoma. Int J Biochem Cell Biol, 2008. 15. Latif, F., Tory, K., Gnarra, J., Yao, M., Duh, F. M., Orcutt, M. L., Stackhouse, T., Kuzmin, I., Modi, W., Geil, L., and et al. Identification of the von Hippel-Lindau disease tumor suppressor gene. Science, 260: 1317-20, 1993. 16. Schmidt, L., Duh, F. M., Chen, F., Kishida, T., Glenn, G., Choyke, P., Scherer, S. W., Zhuang, Z., Lubensky, I., Dean, M., Allikmets, R., Chidambaram, A., Bergerheim, U. R., Feltis, J. T., Casadevall, C., Zamarron, A., Bernues, M., Richard, S., Lips, C. J., Walther, M. M., Tsui, L. C., Geil, L., Orcutt, M. L., Stackhouse, T., Lipan, J., Slife, L., Brauch, H., Decker, J., Niehans, G., Hughson, M. D., Moch, H., Storkel, S., Lerman, M. I., Linehan, W. M., and Zbar, B. Germline and somatic mutations in the tyrosine kinase domain of the MET proto-oncogene in papillary renal carcinomas. Nat Genet, 16: 68-73, 1997. 17. Schmidt, L. S., Warren, M. B., Nickerson, M. L., Weirich, G., Matrosova, V., Toro, J. R., Turner, M. L., Duray, P., Merino, M., Hewitt, S., Pavlovich, C. P., Glenn, G., Greenberg, C. R., Linehan, W. M., and Zbar, B. Birt-Hogg-Dube syndrome, a genodermatosis associated with spontaneous pneumothorax and kidney neoplasia, maps to chromosome 17p11.2. Am J Hum Genet, 69: 876-82, 2001. 18. Kiuru, M., Launonen, V., Hietala, M., Aittomaki, K., Vierimaa, O., Salovaara, R., Arola, J., Pukkala, E., Sistonen, P., Herva, R., and Aaltonen, L. A. Familial cutaneous leiomyomatosis is a two-hit condition associated with renal cell cancer of characteristic histopathology. Am J Pathol, 159: 825-9, 2001. 19. Meloni-Ehrig, A. M. Renal cancer: cytogenetic and molecular genetic aspects. Am J Med Genet, 115: 164-72, 2002.

103

20. Rubin, B. P., Pins, M. R., Nielsen, G. P., Rosen, S., Hsi, B. L., Fletcher, J. A., and Renshaw, A. A. Isochromosome 7q in adult Wilms' tumors: diagnostic and pathogenetic implications. Am J Surg Pathol, 24: 1663-9, 2000. 21. Pavlovich, C. P., Padilla-Nash, H., Wangsa, D., Nickerson, M. L., Matrosova, V., Linehan, W. M., Ried, T., and Phillips, J. L. Patterns of aneuploidy in stage IV clear cell renal cell carcinoma revealed by comparative genomic hybridization and spectral karyotyping. Genes Chromosomes Cancer, 37: 252-60, 2003. 22. Jiang, F., Richter, J., Schraml, P., Bubendorf, L., Gasser, T., Sauter, G., Mihatsch, M. J., and Moch, H. Chromosomal imbalances in papillary renal cell carcinoma: genetic differences between histological subtypes. Am J Pathol, 153: 1467-73, 1998. 23. Waldert, M., Haitel, A., Marberger, M., Katzenbeisser, D., Ozsoy, M., Stadler, E., and Remzi, M. Comparison of type I and II papillary renal cell carcinoma (RCC) and clear cell RCC. BJU Int, 2008. 24. Laurikkala, J., Kassai, Y., Pakkasjarvi, L., Thesleff, I., and Itoh, N. Identification of a secreted BMP antagonist, ectodin, integrating BMP, FGF, and SHH signals from the tooth enamel knot. Developmental Biology, 264: 91-105, 2003. 25. Yanagita, M., Oka, M., Watabe, T., Iguchi, H., Niida, A., Takahashi, S., Akiyama, T., Miyazono, K., Yanagisawa, M., and Sakurai, T. USAG-1: a bone morphogenetic protein antagonist abundantly expressed in the kidney. Biochem Biophys Res Commun, 316: 490-500, 2004. 26. Yanagita, M. BMP antagonists: Their roles in development and involvement in pathophysiology. Cytokine Growth Factor Rev, 2005. 27. Hardwick, J. C., Kodach, L. L., Offerhaus, G. J., and van den Brink, G. R. Bone morphogenetic protein signalling in colorectal cancer. Nat Rev Cancer, 8: 806-12, 2008. 28. Katsuno, Y., Hanyu, A., Kanda, H., Ishikawa, Y., Akiyama, F., Iwase, T., Ogata, E., Ehata, S., Miyazono, K., and Imamura, T. Bone morphogenetic protein signaling enhances invasion and bone metastasis of breast cancer cells through Smad pathway. Oncogene, 27: 6322-33, 2008. 29. Lai, T. H., Fong, Y. C., Fu, W. M., Yang, R. S., and Tang, C. H. Osteoblasts-derived BMP-2 enhances the motility of prostate cancer cells via activation of integrins. Prostate, 68: 1341-53, 2008.

104

30. Kim, I. Y., Lee, D.-H., Lee, D. K., Kim, B. C., Kim, H. T., Leach, F. S., Linehan, W. M., Morton, R. A., and Kim, S. J. Decreased Expression of Bone Morphogenetic Protein (BMP) Receptor Type II Correlates with Insensitivity to BMP-6 in Human Renal Cell Carcinoma Cells. Clin Cancer Res, 9: 6046-6051, 2003. 31. Reya, T., and Clevers, H. Wnt signalling in stem cells and cancer. Nature, 434: 843-50, 2005. 32. Guillen-Ahlers, H. Wnt signaling in renal cancer. Curr Drug Targets, 9: 591-600, 2008. 33. Chomczynski, P., and Sacchi, N. Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem, 162: 156-9, 1987. 34. Karolchik, D., Kuhn, R. M., Baertsch, R., Barber, G. P., Clawson, H., Diekhans, M., Giardine, B., Harte, R. A., Hinrichs, A. S., Hsu, F., Kober, K. M., Miller, W., Pedersen, J. S., Pohl, A., Raney, B. J., Rhead, B., Rosenbloom, K. R., Smith, K. E., Stanke, M., Thakkapallayil, A., Trumbower, H., Wang, T., Zweig, A. S., Haussler, D., and Kent, W. J. The UCSC Genome Browser Database: 2008 update. Nucleic Acids Res, 36: D773-9, 2008. 35. Kent, W. J., Sugnet, C. W., Furey, T. S., Roskin, K. M., Pringle, T. H., Zahler, A. M., and Haussler, D. The human genome browser at UCSC. Genome Res, 12: 996-1006, 2002. 36. Rhodes, D. R., Yu, J., Shanker, K., Deshpande, N., Varambally, R., Ghosh, D., Barrette, T., Pandey, A., and Chinnaiyan, A. M. ONCOMINE: a cancer microarray database and integrated data-mining platform. Neoplasia, 6: 1-6, 2004. 37. Iafrate, A.J., Feuk, L. Rivera, M.N., Listewnik, M.L., Donahoe, P.K., Qi, Y., Scherer, S.W., and Lee, C. Detection of large-scale variation in the human genome. Nat. Genet, 36:949-951. 38. Bjornsson, H. T., Brown, L. J., Fallin, M. D., Rongione, M. A., Bibikova, M., Wickham, E., Fan, J. B., and Feinberg, A. P. Epigenetic specificity of loss of imprinting of the IGF2 gene in Wilms tumors. J Natl Cancer Inst, 99: 1270-3, 2007. 39. Cerrato, F., Sparago, A., Verde, G., De Crescenzo, A., Citro, V., Cubellis, M. V., Rinaldi, M. M., Boccuto, L., Neri, G., Magnani, C., D'Angelo, P., Collini, P., Perotti, D., Sebastio, G., Maher, E. R., and Riccio, A. Different mechanisms cause imprinting defects at the IGF2/H19 locus in Beckwith-Wiedemann syndrome and Wilms' tumour. Hum Mol Genet, 17: 1427-35, 2008.

105

40. Haruta, M., Arai, Y., Sugawara, W., Watanabe, N., Honda, S., Ohshima, J., Soejima, H., Nakadate, H., Okita, H., Hata, J., Fukuzawa, M., and Kaneko, Y. Duplication of paternal IGF2 or loss of maternal IGF2 imprinting occurs in half of Wilms tumors with various structural WT1 abnormalities. Genes Chromosomes Cancer, 47: 712-27, 2008. 41. Brown, K. W., Power, F., Moore, B., Charles, A. K., and Malik, K. T. Frequency and timing of loss of imprinting at 11p13 and 11p15 in Wilms' tumor development. Mol Cancer Res, 6: 1114-23, 2008. 42. Baudry, D., Hamelin, M., Cabanis, M. O., Fournet, J. C., Tournade, M. F., Sarnacki, S., Junien, C., and Jeanpierre, C. WT1 splicing alterations in Wilms' tumors. Clin Cancer Res, 6: 3957-65, 2000. 43. Huang, C. C., Cutcliffe, C., Coffin, C., Sorensen, P. H., Beckwith, J. B., and Perlman, E. J. Classification of malignant pediatric renal tumors by gene expression. Pediatr Blood Cancer, 46: 728-38, 2006.

106

Chapter IV

SOSTDC1, AN ANTI-PROLIFERATIVE BMP ANTAGONIST, IS POORLY

EXPRESSED IN PATIENTS WITH ADVANCED BREAST CANCER

Kimberly R. Blish, Matthew A. Triplette, Wei Du, Mark C. Willingham, Timothy E.

Kute, Charles E. Birse, Surekha R. Krishnan, Julie C. Brown, Ralph B.

D’Agostino, Frank M. Torti, and Suzy V. Torti

The following manuscript was prepared for future submission to Breast Cancer

Research. Stylistic variations are due to the requirements of the journal. Figure

and manuscript preparation by K.R. Blish. Experiments performed and data

gathered by K.R. Blish and M.A. Triplette (Western blots from cell lines and

tissue microarray quantitation) with assistance and expertise from W. Du

(immunostaining) and J.C. Brown (cell culture). Dot blot and preparation of

rhSOSTDC1 by C.E. Birse and S.R. Krishnan at Human Genome Sciences, Inc.

Collaboration and analysis of tissue microarray data accomplished with T.E. Kute

and M.C. Willingham; statistical analyses by R.B. D’Agostino, Jr. F.M. Torti and

S.V. Torti served advisory and editorial roles related to experimental design, data

analysis, and figure preparation.

107

Abstract Introduction: Bone morphogenetic proteins (BMPs) are extracellular signaling

molecules that accomplish important tasks in development and maintenance of

normal adult tissues. Although the data are varied, multiple BMPs and receptors

are associated with breast cancer, demonstrating the need for regulation of this

pathway. This project focuses on the BMP antagonist sclerostin domain-

containing-1 (SOSTDC1). Current knowledge concerning SOSTDC1 has focused

on its expression in the murine kidney, bones, and teeth and potential roles in

progression of nephropathies and there are no published studies on SOSTDC1’s

expression or actions in breast cancer.

Methods & Results We analyzed the expression of SOSTDC1 protein using in

vitro cell culture models of breast cancer progression, and found that less

SOSTDC1 is secreted from transformed cells. A significant decrease of

SOSTDC1 message is observed in ~90% of breast cancer patients, with the

greatest loss in metastatic disease. We studied SOSTDC1 protein expression in

tissue microarrays containing samples equally from recurrent and non-recurrent

disease in African American and non-African American populations. Data from

these tumor microarrays has shown that SOSTDC1 negatively correlates with

tumor size and stage; demonstrating that less expression is associated with more

advanced disease. Additionally, there are strong correlations between

SOSTDC1 expression and the expression of known breast cancer prognostic

markers EGFR, IGFR1 and PTEN.

108

We made recombinant SOSTDC1 protein and report that SOSTDC1 is a

functional extracellular antagonist of BMP-7 signaling in MCF7 adenocarcinoma

cells. Finally, we observed that exogenous SOSTDC1 suppresses proliferation

in MCF7 cells.

Conclusions: Taken together these data suggest that proper levels of secreted

SOSTDC1 are important to the homeostasis of normal cell signaling and

proliferation. Restoration of SOSTDC1 to breast tumors may have future

therapeutic potential.

109

Introduction

Breast cancer is a group of related neoplastic diseases, each with their own

variations in etiology, pathology, and clinical course. The majority of cases

(>90%) are epithelial tumors including: adenomas, intraductal papilloma, Paget’s

disease, ductal carcinoma in situ (DCIS) and lobular carcinoma in situ (LCIS),

infiltrating ductal carcinomas (including inflammatory disease), and infiltrating

lobular carcinomas. As of 2008, there is an annual age-adjusted incidence rate

of 99.8 per 100,000 and mortality rate of 20.3 per 100,000 for all types of breast

cancer (1). While advances in academic knowledge and technologic savvy

caused a decrease in mortality rates by 2.2% from 1990-2004 (2;3); much

remains to be discovered regarding the pathology of breast tumors to develop

more efficacious treatment and prevention strategies.

One approach for identifying new therapeutic targets has been to study common

regulatory and developmental signaling pathways in cancer cells (4-8). Many of

these pathways include known oncogenes or tumor suppressor genes and affect

cell cycle, apoptosis, or proliferation regulation. Either inappropriate activation or

disintegration of these pathways can provide causal impetus for carcinogenesis.

One pathway of interest involves members of the transforming growth factor

(TGF-β) superfamily, the bone morphogenetic proteins (BMPs, 9).

After first identification as ectopic bone formation inducers in vivo (10), BMPs

have been shown to be organism-wide morphogens and are expressed in many

110

tissues (11-13). The mammalian BMP family members largely function as

extracellular cell-cell communication molecules. Each is expressed in unique

patterns throughout development and into adulthood and has varying roles in the

adult organism. Changes in cellular morphology or function are induced through

the activation of intracellular R-Smad proteins (Smads-1,-5, and -8) after BMP

ligands bind hetero-tetrameric BMP receptor complexes at the cell surface.

Constitutively active serine/threonine kinases on BMP Type II receptors

phosphorylate the Type I receptors, leading to subsequent phosphorylation and

activation of the R-Smads (14). This pathway is analogous to TGF-β signaling

and also shares the use of the “common” Smad protein (15), Smad-4 in the

formation of the Smad signaling complex that then translocates to the nucleus

and causes changes in expression of genes bearing Smad binding regulatory

elements (SBEs, (16). Alterations in BMP pathway signaling have been

documented to affect cell survival, proliferation rates, differentiation,

invasiveness, and migration; all processes that can contribute to the formation

and aggressiveness of cancer (17-20).

In normal breast tissues, few BMPs are expressed at detectable levels and the

expression of BMP ligands, receptors, and signaling molecules in breast cancers

varies (21-24). There are many reports linking BMP activity or lack of BMP

regulation to invasive breast cancer. However, the overall contributions of BMP

signaling to breast carcinogenesis remain complex and at times contradictory, as

the exact in vivo contributions of the BMP ligands are difficult to understand. For

111

example, BMP-2 is anti-proliferative on breast epithelial cells at low

concentrations, affecting cell cycle proteins p21, Rb, and the tumor suppressor

PTEN (21;25;26). Both BMP-2 and BMP-6 counteract proliferation induced by

estrogen (27).

However, other evidence supporting the pro-cancer actions of BMP signaling

include: Overexpression of BMP-4 (28) and BMP-7 and BMP type I receptor has

been documented in tumors (22;29). Additionally, BMP-2 has been shown to be

pro-angiogenic with the ability to aid in cell survival during hypoxia (30;31).

Nuclear staining of activated phospho-Smad-1,-5,-8 is increased in primary

breast tumors, and bony metastasis (32). Overexpression of dominant-negative

TGF-β type II receptors or BMP type II receptors limits BMP signaling and inhibits

the invasive potential of breast cancer cell models (32-34).

While the collective contribution of BMPs signaling in the etiology and

progression of tumors remains complex, the fact remains that careful regulation

of this pathway is necessary for normal tissue homeostasis. Thus, the

extracellular regulators or antagonists of BMP signaling become important foci of

research and may have therapeutic potential (35).

This study focuses on the regulation of BMPs in breast cancer via a new

antagonist named SOSTDC1 (SclerOSTin Domain-Containing-1). Current

knowledge concerning SOSTDC1 (and its orthologs USAG-1 (36), ectodin ((37),

112

or wise (38)) focuses on its expression in the murine kidney, bones, and teeth.

SOSTDC1 knock-out mouse models show supernumerary teeth formation

indicating a role for SOSTDC1 as an integrator of multiple developmental

pathways (39-41). In humans, the BMP-7/SOSTDC1 relationship is being

evaluated for potential roles in modulation of nephropathies (42;43). We have

previously reported on the downregulation of and anti-proliferative capability of

SOSTDC1 in renal carcinomas (44). However, no published reports currently

exist for the expression or activity of SOSTDC1 in breast tissue.

As SOSTDC1 is an antagonist of BMP-2, 4, and 7 (45), these findings suggest

that SOSTDC1 could have a role in suppression of the cancer phenotype, we

hypothesized that SOSTDC1 is a repressor of breast carcinogenesis through

extracellular antagonism of BMP signaling. We tested this hypothesis through

the following studies: 1) Expression of SOSTDC1 in breast tissues and whether

that expression is changed during carcinogenesis; 2) Ability of SOSTDC1 to

module BMP signals in breast cells; and 3) whether SOSTDC1 affects the

transformed phenotype of in vitro models of breast cancer.

113

Materials & Methods Recombinant proteins and reagents

Recombinant human BMP-7 (rhBMP-7) and noggin-Fc chimera were purchased

from R & D Systems (Minneapolis, MN). The following commercial antibodies

were obtained: anti-phospho-Smad-1,5,8 (Cell Signaling Technologies, Danvers,

MA), anti-glutaraldehyde-3-phosphate dehydrogenase (GAPDH, Research

Diagnostics International, Flanders, NJ), horseradish peroxidase (HRP)-

conjugated anti-rabbit IgG (Biorad, Hercules, CA), and goat anti-rabbit IgG-HRP

(Jackson Immunoresearch, West Grove, PA). The generation of anti-

hSOSTDC1 antiserum, the construction of rhSOSTDC1 mammalian expression

vectors and production of rhSOSTDC1 protein have been previously described

(44).

Cells and culture conditions

The mammary epithelial adenocarcinoma cell line MCF7 (ATCC HTB-22) was

maintained in DMEM/F12 media supplemented with 10% fetal bovine serum

(FBS, Gibco/Invitrogen, Carlsbad, CA) and 1% penicillin/streptomycin (P/S,

Invitrogen). The ductal carcinoma cell line T47D (HTB-133) was maintained in

RPMI 1640 media with supplements as described by the ATCC. Human

mammary epithelial cells (HMECs) were purchased from Cambrex and

maintained in MEGM media (Cambrex, MA). The HMEC stepwise-transformed

cell lines were a gift from the lab of R. Weinberg and were maintained according

to previously published protocol (46).

114

Detection of cell-associated and secreted SOSTDC1 in breast cells

HMECs, four lines of HMEC-transformed cells from the stepwise model of breast

cancer progression (hTERT, SV40 T antigen, and H-ras overexpressing cells),

MCF7, and T47D cell lines were all weaned to a defined media system over a

period of two-weeks. All cell lines were seeded at equal density, and were

maintained in FBS-free media for 72 hours until conditioned media removed and

cells harvested by scraping and lysed. Conditioned medias were filter-sterilized

and snap-frozen until concentration. 30 mLs of each conditioned media were

concentrated ~300 fold with Amicon and Microcon 10,000 kDa cutoff spin filters

(Millipore, Billerica, MA). Total protein concentrations of each cell lysate and

concentrated media samples were determined by Bradford or BCA assay

colorimetric assays. Equal amounts of cellular lysates or concentrated media

were electrophoresed via SDS-PAGE. Nitrocellulose blots were stained with

Ponceau S as a loading control and then probed with anti-human-SOSTDC1.

RNA dot blot

BD Clontech™ Cancer Profiling Array I was probed with radiolabeled SOSTDC1

per manufacturer’s instructions. Tissues in the array included matched normal

and tumor tissues from 50 breast cancer patients (including three patients with

samples from both primary tumor and metastatic site. Resulting signals were

scanned and quantified with UnScanIt (Silk Scientific, Orem, UT).

115

Tissue microarrays and immunohistochemistry

The creation of tissue microarrays of tumor cores from 198 breast cancer

patients with recurring tumors has been previously described (3). The blocks of

arrayed tissues were sectioned and fixed on slides. Each section was

deparaffinized with xylene, rehydrated, treated with 0.3% hydrogen peroxide, and

antigen unmasking solution (AUS, Jackson Labs). After blocking, slides were

incubated with anti-SOSTDC1 primary antibody in 1% BSA-PBS, followed by

(HRP) - conjugated secondary antibody to rabbit IgG. DAB substrate was used

to indicate location of antibody binding, followed by counterstaining with

hematoxylin. Bright field digital images of each core in the analysis (n = 81) were

obtained using 20x magnification at equal exposure conditions.

Intracellular SOSTDC1 quantification and statistical analysis

The intracellular staining intensity of these images was quantified in Adobe

Photoshop as mean pixel density based on previously reported protocol (47).

Briefly, bright field digital images were adjusted identically to full RBG “Red-Blue-

Green” ranges. Random areas of cytoplasm from 10 separate cells were

selected with the “magic wand tool” and mean pixel intensity was quantitated

using the “histogram” function of total luminosity. Mean pixel intensity was

averaged for the 10 separate cells and reported as a percentage of maximal

intensity. The intensity for the control image (no primary antibody) was

subtracted as background from the percentage of intensity for its corresponding

116

SOSTDC1 image for tissue, thus obtaining an absolute intensity reading for each

sample.

The resulting 81 SOSTDC1 absolute intensity values were imported into SAS

statistical analysis software (Cary, NC) along with samples values for patient

age, race (African-American or other), recurrence and survival info, tumor size,

grade, stage, lymph node involvement, metastasis occurrence and metastasis

location. SOSTDC1 staining intensity was individually correlated to each of

these variables by calculation of Pearson’s coefficient. Additionally, SOSTDC1

staining intensity was compared to survival times (length in years) via

construction of Kaplan-Meier estimates. Finally, staining intensity of SOSTDC1

was correlated to staining of known biomarkers: PTEN, EGFR, and IGFR-1

previously obtained in these microarray samples (48).

Stimulation of phosphorylated Smad (pSmad) production with BMP-7

MCF7 breast cancer cells were plated in DMEM/F12 with 5% FBS and 1% P/S

and allowed to attach overnight. Treatment media was plain DMEM/F12 plus

some combination of rhBMP-7 (75 ng/mL), rhnoggin-Fc chimera (150 ng/mL), or

rhSOSTDC1 (150 ng/mL). After 4 hours, cells were harvested directly into SDS-

PAGE loading buffer. Lysates were run on SDS-PAGE gels followed by pSmad

and GAPDH detection by Western blot.

117

Proliferation experiments

MCF7 cells were transiently transfected in OptiMEM media (Gibco, Invitrogen)

with pFLAG-CMV5a (empty, Sigma) and pFLAG-CMV5a-SOSTDC1 using

LipofectAMINE™ 2000 (Invitrogen) via manufacturer’s instructions. 6 hours after

transfection, FBS was added to the transfected cell cultures to a final

concentration of 5%. 24 hours after transfection, cells were trypsanized and

reseeded at 5,000 cells per well in an E-plate for use on the ACEA RT-CES®

system (ACEA Biosciences, Inc., San Diego, CA). This system allows real-time

monitoring of cultures grown in a 96-well plate format. Each type of transfected

cell was re-plated in its own conditioned media. The remaining cells from each

transfection were plated at 200,000 per well in a 6-well plate for timelapse bright

field photography study of cell viability and morphology.

118

Results & Discussion

SOSTDC1 expression in cell culture models and normal breast samples

As SOSTDC1 levels have not previously been reported in normal breast tissues

or available cell models of breast cancer, we first examined SOSTDC1 protein

via immunohistochemistry and immunoblotting. Normal human mammary

tissues express SOSTDC1 in both ductal and lobular epithelial cells (Figure 1).

Additionally, human mammary epithelial cells (HMECs) also express SOSTDC1

endogenously and secrete mature protein into the culture media (Figure 2).

Figure 1.

Figure 1. Immunohistochemistry of SOSTDC1 in normal and breast tumor

tissues. Representative pictures shown. Rabbit-anti-human SOSTDC1

serum was used to stain paraffin sections of normal (left) and breast tumor

tissue (right) from the same patient. Moderate intracellular SOSTDC1

staining is seen in the epithelia and ducts of normal tissue while the tumor

tissue shows approximately 60% less SOSTDC1 in the tumor cells.

119

Loss of mature, extracellular SOSTDC1 expression in conditioned media of

breast cancer cell lines

We next examined endogenous SOSTDC1 in a panel of breast cancer cells

including a model of breast cancer progression made by step-wise stable

expression of telomerase, SV40 small and large T antigen, and h-ras in HMECs

(46). SOSTDC1 was detected at comparable levels in the whole-cell lysates in

every cell line studied (Figure 2, top panel). As cultures were harvested by

scraping, it is assumed that all cell membrane-associated proteins would be in

the cellular lysate.

All known activities of SOSTDC1 occur in the extracellular space, thus the

amount of mature SOSTDC1 in the conditioned media is biologically relevant.

We found that secreted SOSTDC1 was undetectable in the supernatants from all

transformed cell lines; including HMECs expressing SV40 T antigens and ras

oncogene (Figure 2, middle panel). Interestingly, HMECs with over-expression

of telomerase still expressed some extracellular SOSTDC1 (although less than

the normal HMECs) but once another oncogene was introduced (SV40 T

antigens) no secreted SOSTDC1 was detected.

We wanted to ensure the lack of secreted SOSTDC1 was not solely attributable

to changes in the secretory pathway characteristic of cancer cells (49), so the

conditioned medias were analyzed by SDS-PAGE with two different loading

strategies. First, equal amounts of total media proteins were analyzed and

120

immunoblot only revealed the SOSTDC1 band doublet in the conditioned media

of HMECs and HMEC-hTERT cells (data not shown). In the event that the

secretion machinery of the transformed cells was affecting outcome, media

proteins were loaded on a second gel as a function of total intracellular protein.

Ponceau S stain for this blot showed some variation in the amount of total protein

in the conditioned media, presumably showing different secretion rates of the

cultures (Figure 2, bottom panel). The observable amounts of SOSTDC1 in the

media did not differ (Figure 2, middle panel), so we conclude the observed

decrease of secreted SOSTDC1 is not due solely to changes in the overall

secretion mechanism.

The explanation for the lack of free, extracellular SOSTDC1 is a subject of

ongoing study in our laboratory. One possibility is that a change in the

expression or secretion of proteolytic factors is affecting the rate of SOSTDC1

degradation in more transformed cells. Current experiments with protease

inhibitors are addressing this. Genetic experiments or RNA studies would be

necessary to determine whether mutated SOSTDC1 or alternative splice variants

exist that are not efficiently secreted. Another possibility is that other regulatory

molecules, such as binding partners or chaperones, are affecting the expression,

sorting, or secretion of SOSTDC1. Feedback loops commonly exist where BMP

signaling causes the upregulation of inhibitors (50) and we have observed

changes in SOSTDC1 expression in response to BMP-7 treatment (unpublished

data). Expression of the known BMP ligands that SOSTDC1 binds (BMPs -2, -4,

121

-7, -6) and BMP receptors as well as co-localization studies would start to

address these issues.

Figure 2.

Figure 2. Secreted SOSTDC1 decreases in mammary cells with increasing

transformation. A rabbit anti-human SOSTDC1 antibody was used to probe

Western blots of various normal and cancer cell lysates and culture

supernatants. Top row, cell-associated protein: 26 μg of total cellular

protein for each cell lysate were loaded and immunoblotted with anti-

SOSTDC1. Recombinant human SOSTDC1 (rSOSTDC1, 20 ng) was used as

a positive control for bands at 32 KDa. All normal and cancer cell lines

show comparable amounts of endogenous, cell-associated SOSTDC1.

Middle panel: SOSTDC1 has sites for N-glycosylation and is usually

detectable extracellularly as a doublet of bands at 32 kDa. Conditioned

culture media representing equal amounts of total intracellular culture

122

protein were concentrated for SDS-PAGE- the Ponceau S stain of this blot

(bottom panel) shows some differences in total protein loaded, possibly

reflecting different secretion rates of the cultures. No secreted SOSTDC1

was detected in transformed cells or cancer cell lines. HMECs = Human

Mammary Epithelial Cells; hTERT = HMECs stably overexpressing human

telomerase; LT = hTERT cells also stably expressing the large and small

SV40 T antigens causing inactivation of p53, Rb pathways; H-ras V21- lo

and –hi = LT cells also stably expressing mutant H-ras at low and high

levels (HMERs). MCF7, T47D breast cancer cells also used in this study.

SOSTDC1 message is downregulated in breast tumors

We next asked whether downregulation of SOSTDC1 expression also occurs in

clinical samples. A commercially-available cDNA dot blot containing matched

normal and tumor samples from breast cancer patients (Cancer Profiling Array I,

Clontech) was probed using radiolabeled SOSTDC1 oligonucleotide (Figure 3A).

SOSTDC1 mRNA levels noticeably decrease in 47/53 (89%) of tumors compared

to matched normal tissue. We also observe that baseline levels of SOSTDC1

expression can vary greatly from one individual to another. SOSTDC1

expression is lowest in those patients with metastatic disease (Figure 3B). This

observation is consistent with recent theories concerning the roles of BMPs in

promoting invasion and metastasis in breast cancer. It has been suggested that

like TGF-β (51;52), BMPs when properly expressed inhibit proliferation of normal

123

breast epithelia or early stage lesions; however, when BMPs are overexpressed

in malignant cells a pro-invasive state is achieved (32).

Figure 3.

Figure 3. Quantification of SOSTDC1 message in breast cancer patients.

A. Clontech’s Cancer Profiling Array I of 53 matched normal and tumor

tissues (both primary and distant metastases) was probed with anti-

SOSTDC1 radiolabled primers. SOSTDC1 message visibly decreases

between normal and tumor tissue in at least 47 (89%) sample pairs and

disappears almost entirely in 16 (30%) sample pairs. Lanes 1 & 3, N =

Normal tissue; Lanes 2 & 4, T = tumor tissue. B. The resulting blot was

scanned and dot intensity quantitated with UnScanIt Image Quantification

124

software. Results are summarized in the graph on the right. Average

staining intensity of tumors is ~60% of all normal tissues (* = p < 0.05),

while metastasis show even less SOSTDC1 (** = p <0.01).

SOSTDC1 protein is lowest in patients with advanced disease

As SOSTDC1 message is downregulated in many tumors, we asked whether

protein levels are also decreased. To study this as thoroughly as possible we

performed immunohistochemistry (IHC) on breast cancer tissue microarrays

(TMAs). The samples on these TMAs were evenly divided between recurrence

rates and African American race status, to ascertain whether differentially

expressed proteins predict the imbalance in recurrence rates frequently seen in

African American populations (3). A total of 5 normal breast samples and 81

breast cancer tissues from the microarrays stained with anti-SOSTDC1 serum

were included in the analysis. Average intensity of SOSTDC1 staining was

compared to a variety of clinical parameters including: recurrence rate, race

(African American vs. non), tumor size, tumor stage, tumor grade, hormone

receptor status, cell cycle percentage information, number of positive lymph

nodes, occurrence and site of metastasis, time to recurrence, survival time, and

disease-metastasis free survival.

We observed a decrease in SOSTDC1 protein from normal breast tissue to

breast tumors in 3 matched pairs (Figure 1 for an example). More matched

samples would need to be examined to draw further conclusions about the

125

frequency of loss of SOSTDC1 protein in the population. We found a large range

of expression of SOSTDC1 in the tumors, which fell in a fairly normal distribution

across the sample set (Figure 4). This agrees with the SOSTDC1 dot blot, which

also showed a wide range of baseline mRNA expression levels.

Comparing SOSTDC1 to other parameters, SOSTDC1 did not correlate

significantly with race (p = 0.99 , recurrence (p = 0.12) , or hormone receptor

status (p = 0.62 for progesterone receptor, p = .28 for estrogen receptor). We

found significant associations between: SOSTDC1 expression and tumor size,

tumor stage. An association between SOSTDC1 and tumor grade approached

significance (Table 1). Larger, more advanced tumors have less SOSTDC1. As

there was no association with SOSTDC1 expression and race or recurrence

rates, we conclude that loss of SOSTDC1 may be a negative prognostic factor

across all racial groups. These data leads us to hypothesize that loss of

SOSTDC1 may allow the development of more aggressive tumors.

126

Figure 4.

Figure 4. SOSTDC1 protein levels by immunohistochemistry for 81 breast

tumors. Frequency distribution histogram of SOSTDC1 staining intensities

over all samples quantified. Staining intensity was calculated as a function

SOSTDC-1 stained pixel density minus control (no primary antibody)

sample pixel density.

Relationship of SOSTDC1 to disease metastasis-free survival (DMFS)

We next compared SOSTDC1 expression to incidence of metastasis and survival

times (ranging from <1 to 15 years) for the patient samples in the TMAs. We

found that those patients with the highest levels of SOSTDC1 had a median

survival time of 8.5 years (n = 19) compared to those with the lowest SOSTDC1

expression, median survival time of 5.4 years. However, Kaplan-Meier analysis

0

2

4

6

8

10

12

14

-5-0 0-5 5-10

10-15

15-20

20-25

25-30

30-35

35-40

40-45

45-50

50-55

55-60

Range Density Values

# of

tiss

ues

127

of these data did not result in a significant difference (p= 0.15). It is possible that

for this analysis, the sample size was too small.

In this population, SOSTDC1 expression did not significantly correlate with

overall occurrence of metastasis, site-specific metastasis or DMFS. Again,

further studies in a larger group will be necessary to ascertain whether sample

size is an issue here, or whether there is discordance between the

downregulation of SOSTDC1 message seen in metastasis (Figure 3) and the

amount of SOSTDC1 protein levels.

Table 1: Correlations Between SOSTDC1 Staining Intensity and Tumor Parameters

Stage Mean SOSTDC1

N Tumor size

(mm3)

Mean SOSTDC1

N Grade Mean SOSTDC1

N

1 37.3 12 0-20 36.9 18 1 43.0 4

2 35.3 41 20-50 33.2 42 2 33.0 15

3 26.5 12 >50 20.6 12 3 31.03 37

4 25.7 3

Pearson: -0.289 P = 0.012

Pearson: -0.351 p = 0.0025

Pearson: -0.203 p = 0.105

Table 1. SOSTDC1 staining of TMAs- significant correlations. Pearson’s

correlation coefficient determined for average SOSTDC1 intensity versus

tumor stage, size, and grade. The Pearson score and p values are shown

for each comparison.

128

Relationship of SOSTDC1 to known prognostic markers of breast cancer

Through the TMA study, we evaluated SOSTDC1’s potential as a prognostic

marker through comparisons to other known prognostic markers and/or proposed

tumor suppressors such as IGFR-1 (insulin-like growth factor receptor-1), EGFR,

and PTEN (phosphatase and tensin homolog). SOSTDC1 expression

significantly correlates with expression of all of these markers, but in different

manners (Table 2). SOSTDC1 expression positively correlates with PTEN and

IGFR-1 expression (higher levels of SOSTDC1 are found in samples with higher

levels of these gene products); grouping SOSTDC1 with protective markers and

known tumor suppressors. In contrast, SOSTDC1 has a strong negative

correlation with EGFR expression. As high levels of EGFR are a known negative

prognostic marker, the negative association still supports the hypothesis that high

SOSTDC1 levels are protective.

As discussed above, SOSTDC1 expression did not significantly correlate with

survival. We have speculated this could be due to small sample size in this TMA

staining. Lending some support to this idea, both PTEN and EGFR expression

are known to be significantly associated with survival rates in breast cancer (53-

55); however, in our population, neither PTEN nor EGFR were associated

significantly with survival (p=0.5597 for EGFR and .7250 for PTEN).

129

Table 2: Correlations Between SOSTDC1 Staining Intensity and Staining Intensity of Various Prognostic Markers EGFR IGFR-1 PTEN

SOSTDC1 PC: -0.323 p:0.0048 N: 75

PC: 0.331 p: 0.0067 N: 66

PC: 0.395 p: 0.0006 N: 72

Table 2. Comparisons between SOSTDC1 staining intensity and other

tumor markers in the TMA tissues. PC = Pearson’s Coefficient, p values,

and number of observations used in each comparison are shown.

SOSTDC1 antagonizes the signaling of BMPs in breast cancer

As the effects of BMPs on breast cancer cells are diverse, we next asked

whether SOSTDC1 antagonizes BMP signaling in breast cancer cells and what

the functional consequences of such antagonism might involve. We established

a model of BMP/SOSTDC1 activity in a breast cancer cell system. Several

breast cancer cell lines were cultured and treated with varying concentrations of

recombinant, human BMP-7 (rhBMP-7) ligand for 0, ½, 1, 2, 4, 6, and 10 hours.

rhBMP-7 ligand was chosen both because breast cancers have been shown to

overexpress BMP-7 (22;56) and SOSTDC1 binds to BMP-7 with highest affinity

of BMP ligands -2, -4, -6, and -7 (45). After treatment, the presence of BMP-

response Smads (R-Smads-1,-5,-8) was assessed via Western blot with anti-

phospho-Smad-1-5-8. MCF7, T47D, and HMEC-Ras/HMER cells (see

description in Figure 2) all responded to BMP-7 with a noticeable increase in

130

phosphorylated R-Smad between 2 and 8 hours (data not shown). MCF7 cells

showed the most robust response at 4 hours of treatment time and were used in

further experiments to test the ability of SOSTDC1 to antagonize the BMP-7-

induced rise in phospho-R Smads-1,-5,-8.

Co-treatment with rhBMP-7 and recombinant, human SOSTDC1 (rhSOSTDC1)

showed a great reduction in the amount of phosphorylated R-Smad produced

compared to BMP-7 alone (Figure 5). rhSOSTDC1 was able to suppress

phospho-Smad formation more effectively than the “classical” BMP antagonist

noggin. The efficiency of SOSTDC1’s antagonism of BMP-7 in a breast cancer

model led us to hypothesize:if breast epithelial cells lost their expression of

SOSTDC1, the proper balance of BMP signals maintaining homeostasis could be

jeopardized. As BMPs have various effects on the cell viability and proliferation

rates, we explored the functional consequences of overexpression of SOSTDC1.

131

Figure 5.

Figure 5. SOSTDC1 inhibits BMP-7-initiated phosphorylation of intracellular

R-Smads-1,-5, and -8. Western blots of phosphorylated R-Smad-1,-5,-8 and

GAPDH from MCF7 breast adenocarcinoma cell lysates after treatment with

recombinant human BMP-7 (75 ng/mL) for 4 hours with or without

inhibitors. Cells treated with BMP-7 showed transmission of BMP signal

with an increase in pSmad (lane 2 compared to lane 1 control). Cells

concurrently treated with a 2-fold excess of SOSTDC1 or the known BMP

antagonist noggin showed lesser amounts of pSmad (lanes 3 and 4);

SOSTDC1 co-treatment inhibited pSmad formation by BMP-7 signaling to a

greater extent than noggin.

132

SOSTDC1 suppresses proliferation of MCF7 breast cancer cells

MCF-7 cells transiently transfected with either pFLAG-CMV5a-SOSTDC1 vector

or empty vector control were plated on an E-plate and culture growth monitored

on an ACEA biosystems RT-CES machine for 6 days. Empty vector control cells

adhered to the plate within the first 10-12 hours after plating and then started to

proliferate over the rest of the time period. In contrast, SOSTDC1-expressing

cells showed some adherence to the plate during the first 12-hours, but then cells

showed no signs of proliferation (Figure 6A). Additionally, this phenomenon of

proliferation suppression was not rescued with addition of BMP-2 or BMP-7 to

the culture media at the time of plating (data not shown). Transfected cells not

plated in the E-plate were observed for morphogenetic changes via time lapse

microscopy. Representative pictures show rounding up of SOSTDC1-MCF7

cells with occasional blebs and spiking (Figure 6B). After 3 days of culture, the

cells in the time lapse experiment were harvested and viable cells counted with

trypan blue. Control cultures proliferated 5.6 fold, SOSTDC1 overexpressing

cultures only 1.4 fold.

This profound block of culture proliferation warrants further investigation to

identify the mechanisms. As discussed previously, BMPs have published roles

controlling proliferation via cell-cycle regulation, prevention of apoptosis, and

changing the morphology/invasive potential of breast cancer cells. Any of these

mechanisms may be involved in producing the cell phenotypes seen here. This

is particularly true as SOSTDC1 binds to more than one BMP, with the potential

133

to concurrently antagonizing multiple ligands by sequestering them away from

the receptors. Additionally, studies in our lab and others with other tissues have

shown that SOSTDC1 is a potential dual-pathway inhibitor, also antagonizing

Wnt pathway signaling. Exploration of these mechanisms will provide fertile

research fields for future studies.

Figure 6.

134

Figure 6 (Previous Page). Cells overexpressing hSOSTDC1-FLAG show

proliferation suppression. MCF7 cells were transiently transfected with

either the pFLAG-CMV-5a-SOSTDC1 vector (SOSTDC1) or empty pFLAG

vector (EMPTY CONTROL). Culture proliferation was measured on the

ACEA RT-CES system as a function of electric resistance over time

(reported as the Cell Index, CI). Results shown are representative of three

separate experiments. A. The ACEA E-plate was prepared at time = 0 and

background reading of media only obtained. Cells were plated at 7 hours

(gold arrow) and had adhered to the plate by 10-12 hours (red arrow).

Increases in CI over the next 5 days showed steady proliferation of control

cells compared to complete suppression of proliferation in cells

overexpressing SOSTDC1. B. Phase contrast microscopy (20X) of cells at

Day 1 and Day 3 of experiment confirm lack of live cell expansion in

SOSTDC1 cells. Red arrows show a cell undergoing dying with blebbing

and apoptotic echinoid spike formation (middle and right bottom panel,

40X).

135

Conclusions

This is the first report of the distribution of the BMP antagonist SOSTDC1 in

human breast tissues and its subsequent loss in many breast tumors. Loss of

expression of mature, secreted SOSTDC1 was identified in breast cancer cells.

The downregulation was explored at the message and protein level in clinical

samples with the use of tissue microarrays. We identified SOSTDC1 as a

potential biomarker of breast cancer (negatively associated with tumor stage and

size) with correlations to other known prognostic markers. Additionally, our

findings that overexpression of SOSTDC1 causes proliferation suppression in an

invasive breast cancer model suggests that SOSTDC1 may have tumor

suppressive activities. The potential roles of SOSTDC1 uncovered by this work

increase our enthusiasm to understand more about this molecule.

136

Reference List

(1) American Cancer Society. Breast Cancer Facts & Figures 2007-2008. American Cancer Society 2008 [cited 2009 Mar 20];Available from: URL: http://www.cancer.org/docroot/STT/content/STT_1x_Breast_Cancer_Facts__Figures_2007-2008.asp

(2) Berry DA, Cronin KA, Plevritis SK, Fryback DG, Clarke L, Zelen M, et al. Effect of screening and adjuvant therapy on mortality from breast cancer. N Engl J Med 2005 Oct 27;353(17):1784-92.

(3) Stackhouse BL, Williams H, Berry P, Russell G, Thompson P, Winter JL, et al. Measurement of glut-1 expression using tissue microarrays to determine a race specific prognostic marker for breast cancer. Breast Cancer Res Treat 2005 Oct;93(3):247-53.

(4) Hanahan D, Weinberg RA. The hallmarks of cancer. Cell 2000;100(1):57-70.

(5) Logan CY, Nusse R. The Wnt signaling pathway in development and disease. Annu Rev Cell Dev Biol 2004;20:781-810.

(6) Bolos V, Blanco M, Medina V, Aparicio G, az-Prado S, Grande E. Notch signalling in cancer stem cells. Clin Transl Oncol 2009 Jan;11(1):11-9.

(7) Zardawi SJ, O'Toole SA, Sutherland RL, Musgrove EA. Dysregulation of Hedgehog, Wnt and Notch signalling pathways in breast cancer. Histol Histopathol 2009 Mar;24(3):385-98.

(8) Rubin LL, de Sauvage FJ. Targeting the Hedgehog pathway in cancer. Nat Rev Drug Discov 2006 Dec;5(12):1026-33.

(9) Li B. Bone morphogenetic protein-Smad pathway as drug targets for osteoporosis and cancer therapy. Endocr Metab Immune Disord Drug Targets 2008 Sep;8(3):208-19.

(10) Wozney JM, Rosen V, Celeste AJ, Mitsock LM, Whitters MJ, Kriz RW, et al. Novel regulators of bone formation: molecular clones and activities. Science 1988;242(4885):1528-34.

137

(11) Chen D, Zhao M, Mundy GR. Bone morphogenetic proteins. Growth Factors 2004;22(4):233-41.

(12) Reddi AH. Bone morphogenetic proteins: from basic science to clinical applications. J Bone Joint Surg Am 2001;83-A Suppl 1(Pt 1):1-6.

(13) Wozney JMRV. Bone Morphogenetic Protein and Bone Morphogenetic Protein Gene Family in Bone Formation and Repair. Clin Orthop 1998;346:26-37.

(14) Shi Y, Massague J. Mechanisms of TGF-beta Signaling from Cell Membrane to Nucleus. Cell 2003;113:685-700.

(15) Miyazono K, Kusanagi K, Inoue H. Divergence and convergence of TGF-beta/BMP signaling. J Cell Physiol 2001 Jun;187(3):265-76.

(16) Jonk LJC, Itoh S, Heldin C-H, ten Dijke P, Kruijer W. Identification and Functional Characterization of a Smad Binding Element (SBE) in the JunB Promoter That Acts as a Transforming Growth Factor-beta , Activin, and Bone Morphogenetic Protein-inducible Enhancer. J Biol Chem 1998;273(33):21145-52.

(17) Waite KA, Eng C. From developmental disorder to heritable cancer: it's all in the BMP/TGF-beta family. Nat Rev Genet 2003 Oct;4(10):763-73.

(18) Ten Dijke P, Goumans M-J, Itoh F, Itoh S. Regulation of Cell Proliteration by Smad Proteins. J Cell Phys 2002;191:1-16.

(19) Blanco CM, Bolos F, V, Medina V, V, Aparicio GG, Diaz PS, Grande PE. Biology of BMP signalling and cancer. Clin Transl Oncol 2009 Mar;11(3):126-37.

(20) Li B. Bone morphogenetic protein-Smad pathway as drug targets for osteoporosis and cancer therapy. Endocr Metab Immune Disord Drug Targets 2008 Sep;8(3):208-19.

(21) Arnold SF, Tims E, Mcgrath BE. Identification of bone morphogenetic proteins and their receptors in human breast cancer cell lines: importance of BMP2. Cytokine 1999 Dec;11(12):1031-7.

138

(22) Alarmo EL, Kuukasjarvi T, Karhu R, Kallioniemi A. A comprehensive expression survey of bone morphogenetic proteins in breast cancer highlights the importance of BMP4 and BMP7. Breast Cancer Res Treat 2007 Jun;103(2):239-46.

(23) Bobinac D, Maric I, Zoricic S, Spanjol J, Dordevic G, Mustac E, et al. Expression of bone morphogenetic proteins in human metastatic prostate and breast cancer. Croat Med J 2005 Jun;46(3):389-96.

(24) Davies SR, Watkins G, Douglas-Jones A, Mansel RE, Jiang WG. Bone morphogenetic proteins 1 to 7 in human breast cancer, expression pattern and clinical/prognostic relevance. J Exp Ther Oncol 2008;7(4):327-38.

(25) Ghosh-Choudhury N, Woodruff K, Qi W, Celeste A, Abboud SL, Ghosh CG. Bone morphogenetic protein-2 blocks MDA MB 231 human breast cancer cell proliferation by inhibiting cyclin-dependent kinase-mediated retinoblastoma protein phosphorylation. Biochem Biophys Res Commun 2000 Jun 16;272(3):705-11.

(26) Ghosh-Choudhury N, Ghosh-Choudhury G, Celeste A, Ghosh PM, Moyer M, Abboud SL, et al. Bone morphogenetic protein-2 induces cyclin kinase inhibitor p21 and hypophosphorylation of retinoblastoma protein in estradiol-treated MCF-7 human breast cancer cells. Biochimica et Biophysica Acta (BBA) - Molecular Cell Research 2000;1497(2):186-96.

(27) Takahashi M, Otsuka F, Miyoshi T, Otani H, Goto J, Yamashita M, et al. Bone morphogenetic protein 6 (BMP6) and BMP7 inhibit estrogen-induced proliferation of breast cancer cells by suppressing p38 mitogen-activated protein kinase activation. J Endocrinol 2008 Dec;199(3):445-55.

(28) Montesano R, Sarkozi R, Schramek H. Bone morphogenetic protein-4 strongly potentiates growth factor-induced proliferation of mammary epithelial cells. Biochem Biophys Res Commun 2008 Sep 12;374(1):164-8.

(29) Helms MW, Packeisen J, August C, Schittek B, Boecker W, Brandt BH, et al. First evidence supporting a potential role for the BMP/SMAD pathway in the progression of oestrogen receptor-positive breast cancer. J Pathol 2005 Jul;206(3):366-76.

139

(30) Raida M, Clement JH, Ameri K, Han C, Leek RD, Harris AL. Expression of bone morphogenetic protein 2 in breast cancer cells inhibits hypoxic cell death. Int J Oncol 2005 Jun;26(6):1465-70.

(31) Raida M, Clement JH, Leek RD, Ameri K, Bicknell R, Niederwieser D, et al. Bone morphogenetic protein 2 (BMP-2) and induction of tumor angiogenesis. J Cancer Res Clin Oncol 2005 Nov;131(11):741-50.

(32) Katsuno Y, Hanyu A, Kanda H, Ishikawa Y, Akiyama F, Iwase T, et al. Bone morphogenetic protein signaling enhances invasion and bone metastasis of breast cancer cells through Smad pathway. Oncogene 2008 Oct 23;27(49):6322-33.

(33) Dumont N, Arteaga CL. A kinase-inactive type II TGFbeta receptor impairs BMP signaling in human breast cancer cells. Biochem Biophys Res Commun 2003 Jan 31;301(1):108-12.

(34) Pouliot F, Blais A, Labrie C. Overexpression of a dominant negative type II bone morphogenetic protein receptor inhibits the growth of human breast cancer cells. Cancer Res 2003 Jan 15;63(2):277-81.

(35) Yanagita M. BMP antagonists: their roles in development and involvement in pathophysiology. Cytokine Growth Factor Rev 2005 Jun;16(3):309-17.

(36) Simmons DG, Kennedy TG. Uterine sensitization-associated gene-1: a novel gene induced within the rat endometrium at the time of uterine receptivity/sensitization for the decidual cell reaction. Biol Reprod 2002 Nov;67(5):1638-45.

(37) Laurikkala J, Kassai Y, Pakkasjarvi L, Thesleff I, Itoh N. Identification of a secreted BMP antagonist, ectodin, integrating BMP, FGF, and SHH signals from the tooth enamel knot. Dev Biol 2003 Dec 1;264(1):91-105.

(38) Itasaki N, Jones CM, Mercurio S, Rowe A, Domingos PM, Smith JC, et al. Wise, a context-dependent activator and inhibitor of Wnt signalling. Development 2003 Sep;130(18):4295-305.

(39) Kassai Y, Munne P, Hotta Y, Penttila E, Kavanagh K, Ohbayashi N, et al. Regulation of mammalian tooth cusp patterning by ectodin. Science 2005 Sep 23;309(5743):2067-70.

140

(40) Murashima-Suginami A, Takahashi K, Kawabata T, Sakata T, Tsukamoto H, Sugai M, et al. Rudiment incisors survive and erupt as supernumerary teeth as a result of USAG-1 abrogation. Biochem Biophys Res Commun 2007 Aug 3;359(3):549-55.

(41) Murashima-Suginami A, Takahashi K, Sakata T, Tsukamoto H, Sugai M, Yanagita M, et al. Enhanced BMP signaling results in supernumerary tooth formation in USAG-1 deficient mouse. Biochem Biophys Res Commun 2008 May 16;369(4):1012-6.

(42) Yanagita M. Modulator of bone morphogenetic protein activity in the progression of kidney diseases. Kidney Int 2006 Sep;70(6):989-93.

(43) Yanagita M. Balance between bone morphogenetic proteins and their antagonists in kidney injury. Ther Apher Dial 2007 Oct;11 Suppl 1:S38-S43.

(44) Blish KR, Wang W, Willingham MC, Du W, Birse CE, Krishnan SR, et al. A human bone morphogenetic protein antagonist is down-regulated in renal cancer. Mol Biol Cell 2008 Feb;19(2):457-64.

(45) Yanagita M, Oka M, Watabe T, Iguchi H, Niida A, Takahashi S, et al. USAG-1: a bone morphogenetic protein antagonist abundantly expressed in the kidney. Biochem Biophys Res Commun 2004 Apr 2;316(2):490-500.

(46) Elenbaas B, Spirio L, Koerner F, Fleming MD, Zimonjic DB, Donaher JL, et al. Human breast cancer cells generated by oncogenic transformation of primary mammary epithelial cells. Genes Dev 2001 Jan 1;15(1):50-65.

(47) Lehr HA, Mankoff DA, Corwin D, Santeusanio G, Gown AM. Application of photoshop-based image analysis to quantification of hormone receptor expression in breast cancer. J Histochem Cytochem 1997 Nov;45(11):1559-65.

(48) Winter JL, Stackhouse BL, Russell GB, Kute TE. Measurement of PTEN expression using tissue microarrays to determine a race-specific prognostic marker in breast cancer. Arch Pathol Lab Med 2007 May;131(5):767-72.

141

(49) Coppe JP, Patil CK, Rodier F, Sun Y, Munoz DP, Goldstein J, et al. Senescence-associated secretory phenotypes reveal cell-nonautonomous functions of oncogenic RAS and the p53 tumor suppressor. PLoS Biol 2008 Dec 2;6(12):2853-68.

(50) Benazet JD, Bischofberger M, Tiecke E, Goncalves A, Martin JF, Zuniga A, et al. A self-regulatory system of interlinked signaling feedback loops controls mouse limb patterning. Science 2009 Feb 20;323(5917):1050-3.

(51) Tang B, Vu M, Booker T, Santner SJ, Miller FR, Anver MR, et al. TGF-beta switches from tumor suppressor to prometastatic factor in a model of breast cancer progression. J Clin Invest 2003 Oct;112(7):1116-24.

(52) Chang CF, Westbrook R, Ma J, Cao D. Transforming growth factor-beta signaling in breast cancer. Front Biosci 2007;12:4393-401.

(53) Depowski PL, Rosenthal SI, Ross JS. Loss of expression of the PTEN gene protein product is associated with poor outcome in breast cancer. Mod Pathol 2001 Jul;14(7):672-6.

(54) Tsutsui S, Inoue H, Yasuda K, Suzuki K, Higashi H, Era S, et al. Reduced expression of PTEN protein and its prognostic implications in invasive ductal carcinoma of the breast. Oncology 2005;68(4-6):398-404.

(55) Magkou C, Nakopoulou L, Zoubouli C, Karali K, Theohari I, Bakarakos P, et al. Expression of the epidermal growth factor receptor (EGFR) and the phosphorylated EGFR in invasive breast carcinomas. Breast Cancer Res 2008;10(3):R49.

(56) Alarmo EL, Rauta J, Kauraniemi P, Karhu R, Kuukasjarvi T, Kallioniemi A. Bone morphogenetic protein 7 is widely overexpressed in primary breast cancer. Genes Chromosomes Cancer 2006 Apr;45(4):411-9.

142

CHAPTER V

GENERAL DISCUSSION

SOSTDC1 in Cancer

K. BLISH

143

The introductory chapter suggested that models of cancer involving key

pathways of cell viability and behavior are necessary for the development of

novel anti-neoplastic treatments. The work presented here developed models of

BMP and Wnt signaling in breast and kidney cancers. The goal was to better

understand regulation of these pathways through studies of the dual-pathway

inhibitor, SOSTDC1. The strategy for accomplishing this goal involved: looking

for loss-of-heterozygosity (LOH) and mutations at the SOSTDC1 locus, studying

SOSTDC1 expression and actions in vitro, and evaluation of SOSTDC1

expression in patient samples. In this discussion, the success of meeting these

goals will be evaluated and an examination of the challenges encountered during

this work will be presented.

Success in Accomplishing the Goals of this Project

Review: SOSTDC1 in Cancer, What Has Been Accomplished

This body of work has yielded new and relevant information to the understanding

of SOSTDC1 as well as BMP and Wnt signaling in breast and renal tumors.

From the initial array observations showing SOSTDC1 as downregulated in a

wide range of tissues and tumors, the potential tumor suppressor actions of

SOSTDC1 have been explored.

144

Current evidence shows that:

1) LOH affects SOSTDC1 in a ~16% of Wilms Tumors. This rate is comparable

to LOH seen at other Wilms tumor suppressor genes. LOH along the same

section of chromosome 7p was described for the first time in adult renal cancers.

Direct sequencing confirmed LOH and unveiled some previously unpublished

single nucleotide polymorphisms (SNPs).

2) SOSTDC1 message and protein are downregulated in renal and breast

tumors. Mature SOSTDC1 protein is secreted from overexpressing cells and

distributes to nearby cells in a concentration gradient. SOSTDC1 expression

was examined in a step-wise model of breast cancer progression (where normal

HMECs receive 4 genetically defined hits causing, in order, unlimited replicative

potential, immortality, increasing growth factor signaling, insensitivity to apoptotic

signals, and the ability to invade surrounding tissue). Mature SOSTDC1 was

secreted from normal HMECs and HMECs overexpressing telomerase, but once

immortality and full transformation is achieved, extracellular SOSTDC1 was lost

from the cultures. Additional work with cell culture kidney models (data not

shown) revealed that many renal carcinoma cell lines do not make routinely

detectable amounts of SOSTDC1 message or protein.

Functional consequences of this loss may affect BMP pathway regulation as

exogenous SOSTDC1 is able to effectively inhibit BMP-7 signaling in breast and

renal cancer cells and BMP-2 signaling in breast cancer cells (pilot study data not

145

shown). Additionally, the ability of SOSTDC1 to antagonize Wnt signaling was

shown for the first time in human tissues. SOSTDC1 inhibited Wnt-3a induced

TCF/LEF promoter activation almost as well as the Wnt specific inhibitor DKK.

This is the first evidence of the dual-pathway effects of SOSTDC1 in human

tissues, normal or cancerous. Finally, restoration of SOSTDC1 to a variety of

breast and renal clear cell cancer lines blocked the proliferation of these cultures

better than DKK-1 or noggin alone (renal) and was not rescued by BMP-2 or

BMP-7 treatment (breast).

3) Within clinical kidney tissue samples, clear cell renal carcinomas show the

lowest amounts of SOSTDC1 protein. In breast samples, loss of SOSTDC1

significantly correlates with markers of more advanced disease: larger tumor

volume, higher stage, and prognostic markers such as EGFR. Early studies

have shown that SOSTDC1 can be detected in urine and serum (data not

shown).

While these data are quite general observations about SOSTDC1, putting them

together builds a model that is thus far consistent with the early hypothesis of this

work. SOSTDC1 does have extracellular actions to inhibit BMP signaling in

kidney and breast tissues and furthermore, the lost of SOSTDC1 is seen with

high penetrance in renal and breast tumors and appears to correlate with more

advanced disease. This hypothesis is broad and much remains to be determined

146

regarding the specifics of SOSTDC1’s protective mechanisms and the

phenotypic consequences of SOSTDC1 loss in cancer.

Challenges Encountered in this Project

As this project started with an observation concerning (at that time) a hypothetic

protein in humans with little to no data about its nature or mode of actions, the

goals for this project were by necessity broad. Therefore, the roadblocks

encountered during these studies were of two varieties: technical and

conceptual.

Technical Challenges

Production of rhSOSTDC1: Technically, reagent development was a limiting step

as the production of rhSOSTDC1 was at best a tedious process and at worst,

unsuccessful. Overexpression within bacterial systems never yielded mature

soluble protein and the process for protein collection from HEK293 cells yields

very little amount of protein. This is not uncommon as most cystine-knot

containing proteins are notoriously insoluble and hard to produce/purify. Work

continues on this front. In the meantime, investigation of the production and

experimental usefulness of SOSTDC1 conditioned media might be a short-term

solution.

In vitro models and transient transfection issues: In addition to having proper

reagents, it was important to develop appropriate cell models systems for the

147

experimental questions. For example, cells in which to test the effects of

SOSTDC1 overexpression should also express appropriate BMP and/or Wnt

pathway receptors and ligands in order to respond to SOSTDC1’s proposed

activities. Determining this could be time-consuming and yet not yield useful

information about whether SOSTDC1 will cause a phenotypic change in those

cells. For this project, kidney cell lines and breast cell lines were screened for

endogenous SOSTDC1 production (Kidney data not shown, all cell lines had low

to non-detectable amounts of SOSTDC1, Breast cell data in Chapter 4). Then,

the choice was made to test the effects of SOSTDC1 overexpression on

phenotype of these cells. If cancer cells responded to SOSTDC1, further studies

of ligand, receptor, and signaling protein availability in the cell lines of interest

would be undertaken. Cancer cell lines from kidney and breast were identified

that did respond to overexpression of SOSTDC1, but characterization of

interacting proteins from the BMP and Wnt pathways is still ongoing.

By nature, in vitro cell models are prone to additional challenges associated with

gene delivery method and dosage issues. As recombinant protein was scarce,

overexpression of SOSTDC1 from plasmids was a necessary step. This

introduces another layer of artificial set-up to the model as overexpression via

transient transfections are notoriously hard to reproduce at similar levels, have

varying toxicities to the cells, and likely express the target gene at levels beyond

physiologic significance. To deal with some of these disadvantages, cell-lines

stably overexpressing SOSTDC1 were attempted. Overexpression of a growth

148

inhibitor, such as SOSTDC1, poses further technical difficulty, as successfully

transfected cells by definition will not replicate as quickly as non-transfected

counterparts. The use of an inducible expression system is one solution to this

problem and is being developed for use in future models. Even with successful

creation of stables cells lines with inducible SOSTDC1 expression, the accurate

assessment of the roles of SOSTDC1 at physiologic levels is necessary via

alternative approaches including identification of likely physiologic levels and

treatment with rhSOSTDC1.

The need for in vivo experiments: Having a good in vitro model in which to

determine a novel protein’s actions is a beneficial tool; however, even a good

model has its conceptual drawbacks as it is still a very artificial system. This is

particularly true with studies of secreted proteins that interact in paracrine

manner. There are many examples of Wnt/BMP interactions between cell types,

including tumor and stroma (1-5). Thus, an in vivo model provides the only way

to truly study the full range of effects of SOSTDC1.

The most immediate line of inquiry would be experiments comparing the in vivo

characteristics of cancer cell lines with or without overexpression of SOSTDC1,

such as a tumor flank or mammary fat pad tumorigenesis model. Such

experiments would allow for a closer approximation of the effects of SOSTDC1 in

an environment where there are multiple cell types. Effects could be observed

149

on many processes including: tumor incidence after cell injection, tumor

growth/size, angiogenesis, tumor invasiveness and metastasis.

Mouse knockouts of SOSTDC1 have been created by several groups (6-8).

Reported phenotypes of living mice include: supernumerary and malformed

teeth, craniofacial abnormalities affecting the trigeminal nerve, and increased

resistance to kidney injury (9). While Kassai et al. report knockout mice are fully

fertile and viable out to 15 months, Shigetani et al. more recently report that

some homozygous knockouts die shortly after birth, and suggest that full

penetrance of the knock-out gene may be hard to obtain. This phenotype

suggests that as a solitary event, loss of SOSTDC1 is not causative for

tumorigenesis, which is not surprising considering the number of BMP and Wnt

antagonists and their overlapping functions. However, because the loss of

SOSTDC1 is a frequent event in human tumors, further investigation is

worthwhile. It remains to be determined whether loss of SOSTDC1 makes mice

more susceptible in response to other carcinogenic exposures such

environmental carcinogens or cross-breeding with other mouse models

expressing oncogenes.

Conceptual Challenges

Even when all technical issues are addressed and all desired model systems

obtained, there still remain a variety of conceptual challenges. These are the

issues at the heart of experimental design and finding efficient ways to address

150

these challenges raises fundamental questions about the nature of SOSTDC1.

The conceptual challenges include: 1) addressing the proof-of-principle studies,

2) considering the multiple potential levels of SOSTDC1’s regulation, 3) attending

to SOSTDC1’s unknown mechanism(s) of action, and 4) grappling with the

complexities inherent to a dual-pathway antagonist orchestrating more than one

signaling network in a tissue-specific manner.

1) Proof-of-Principle Studies

RNA/cDNA and protein studies have shown that SOSTDC1 expression is greatly

downregulated at a high penetrance in many tumor types, particularly breast and

renal. This suggests that SOSTDC1 has a protective effect in normal tissue that

must be overcome for cancer to form or progress. As discussed above, current

mouse knockout strains show that SOSTDC1 loss is not sufficient to cause

transformation. The question remains whether SOSTDC1 loss contributes to the

tumorigenic process. The main proof-of-principle experiment to address this

question lies in knock-down of SOSTDC1 expression from normal cells, then

testing whether these cells have greater sensitivities to oncogenic agents or

develop a more transformed phenotype. Cell knock-down SOSTDC1 models

have been attempted but successful knock-down has yet to be accomplished

(data not shown).

151

These experiments might also address a secondary question: whether

SOSTDC1 loss occurs as a cause that then predisposes cells to cancer, or

whether loss is a result of other changes. Supporting the latter argument, the cell

line step-wise model of breast cancer discussed in Chapter 4 showed a loss of

mature, secreted SOSTDC1 after the addition of SV40 T antigens to the

immortalized HMECs. The expression of SV40 large T antigen causes disruption

to p53 as well as pRb, p300, and CBP, leading to large-scale changes in genome

integrity. It is reasonable to speculate such changes could affect SOSTDC,

either by inducing mutations or affecting its regulation. However, little is known

about what regulates the expression or activity of SOSTDC. More effort on this

front would inform future experiments concerning any causal relationship

between SOSTDC1 and carcinogenesis.

2) Regulation of Expression

There is very little published data about what affects SOSTDC1 expression.

Examples of regulatory mechanisms affecting other BMP/Wnt antagonists as well

as pilot data inform of possible mechanisms worth investigation. These

examples are diverse and fall into genetic regulatory mechanisms, epigenetic,

transcriptional (splicing), gene dosage, as well as protein regulation.

152

Genetic

The LOH studies described in Chapter 3 show that loss of chromosomal data

around the SOSTDC1 locus affects 10-15% of kidney cancers. The remaining

allele appears to be normal in all sequenced sections. The observed LOH

includes other genes within the 2 Mbp region, so causality between changes in

SOSTDC1 genetic material and Wilms or adult RCC cannot be concluded.

However, the fact remains that as many as 90% of renal clear cell carcinomas

show a reduction in SOSTDC1 message. Therefore, it is presumed that

alternative mechanisms exist affecting the downregulation of SOSTDC1

message in kidney.

The first consideration when interpreting these results is the current incomplete

understanding of the SOSTDC1 allele. Direct sequencing to look for significant

mutations was undertaken at 5 potential exons. However, the promoters,

enhancing elements, and other regulatory regions are not defined. Thus,

mutations may exist in important regulatory regions for this allele that have not

been defined and may not have been included in the sequencing. If such

mutations exist, then loss of heterozygosity concurrent with such a mutation on

the remaining allele could cause loss of SOSTDC1 expression in these 10-15%

of cases where LOH occurs. This still leaves ~80% of cases showing decreased

SOSTDC1 expression to explain by other regulatory mechanisms.

153

Epigenetic

There are many examples of BMP and Wnt pathway proteins whose expression

is modulated by epigenetic mechanisms. One pertinent example is the secreted

Wnt antagonist, sFRP1. The sFRP1 locus shows hypermethylation in breast,

oral, hepatocellular, and colon cancers at an incidence of between 10-50% (10-

12). Loss of sFRP1 causes constitutive Wnt signaling upregulation and loss of

anti-proliferative and apoptotic responses in these cells (13). Analysis of the

SOSTDC1 locus reveals a potential CpG island near the 3’ end of what is termed

exon 1 (see Chapter 3 for locus map) and other potential methylation sites exist

around several of the putative transcription start sites. Methylation-specific PCR

in the Wilms and adult kidney cancer populations would address whether a

certain percentage of patients experiencing silencing at the SOSTDC1 allele.

Splice variants

Mutant splice variants, changes in ratios of variants, and changes in gene

dosage are all documented methods of tumor suppressor gene inactivation. For

example, in Wilms Tumor, the prototypical tumor suppressor allele, WT1 does

not often undergo LOH or other deletions of material. Instead, dysregulation of

WT1 function is accomplished by shifts in the transcript variants resulting from

alternative splicing over the two coding exons and use of two translation state

sites (14). mRNAs have been identified that show the existence of as many as 6

different transcripts from the complex SOSTDC1 locus (Aceview Project, NCBI,

(15). It remains to be determined whether these are detectable species in

154

tissues and whether these alternative transcripts produce variant proteins or play

regulatory roles.

Protein regulation

BMP/Wnt Negative Feedback Loop: It is a well-established paradigm in BMP and

Wnt regulation that ligand signaling induces upregulation of transcription of

pathway antagonists as a negative feedback mechanism. Pilot data for this

project has shown that treatment of SaOS-2 cells with BMP-7 increases

SOSTDC1 mRNA by 2-4 fold. Additionally, sequence analysis of the 5-exon

potential SOSTDC1 locus has identified possible Smad-responsive SBE

sequences; however, a more careful investigation of this is necessary, potentially

including mutational studies. SOSTDC1 upregulation in response to Wnt

signaling has not yet been studied. Other pathways could potentially regulate

SOSTDC1 as well; one possibility might be steroid hormone regulation as is

observed for sclerostin (16), but BMP and Wnt regulation remain the most likely

candidates.

Subcellular Localization and Modulation of Secretion from the Cell:

SOSTDC1 has a signal peptide and has been seen in the media of cell cultures

(Chapter 2, Chapter 4). It is hypothesized that all actions of SOSTDC1 occur

outside the cell; however, this has not been formally studied. First, regulation

can occur at the level of protein processing and secretion. Mutations which

affect the appropriate secretion of SOSTDC1 could exist and prevent the proper

155

localization of active molecules to the extracellular space. However, Itasaki et al

recently report different roles for Xenopus wise depending on subcellular

localization. When wise/SOSTDC1 is secreted, it binds to Wnt proteins and

either activates or inhibits Wnt signaling through either β-catenin or the planar

cell polarity pathways (17). However, if wise is maintained in the endoplasmic

reticulum, it binds to LRP6 and reduces its expression on the cell surface,

inhibiting Wnt signals. The mechanism for retention of wise in the ER is unclear

(18). Investigation of this phenomenon in cancer models is necessary and can

be initiated with studies of exogenous protein that lacks the N-terminal signal

peptide domain. Mammalian expression constructs have been prepared for

mutant SOSTDC1 that expresses only residues 24-206 and would likely not get

secreted from the cell.

Once SOSTDC1 is outside the cell, binding partners within the extracellular

matrix that modify its dispersal provide another potential level of regulation.

Studies to examine these issues have been complicated, as little is known about

the folding and partitioning of Wnt or BMP proteins (including SOSTDC1) due to

their insolubility (19). However, sclerostin is known to bind heparins (20) and the

use of glycoprotein scaffolding to provide concentration gradients of BMP and

Wnt proteins is a well-established phenomenon.

Ratio of SOSTDC1 to binding partners: The last issue to consider concerning

SOSTDC1 regulation is the ratio of SOSTDC1 to its binding partners. It is

156

currently unknown what happens to SOSTDC1 if there are no BMP/Wnts to bind,

or if such a scenario is possible, but the actions of the antagonists and ligands

are always interrelated. Several groups have shown that the stoichiometry of

various BMP and/or Wnt pathway members and antagonists changes the

formation of receptor complexes and subsequent downstream signaling (21-24),

at times changing the activated intracellular pathway entirely (such as from Smad

to p38/MAPK signaling) . There are also many examples of antagonists that

interfere with each other’s actions, such as noggin and sclerostin when co-

expressed (25), thus its possible that the actions of SOSTDC1 are dependent

upon the levels of other interacting proteins. This analysis awaits full

identification of SOSTDC1s binding partners.

3) SOSTDC1’s Unknown Mechanisms of Action

The observed downregulation of SOSTDC1 in a variety of cancer tissues and cell

culture models of cancer suggests that many of its in vivo actions may be tumor

suppressive in some capacity: whether anti-transformative, anti-proliferative or

anti-metastatic. Anti-proliferative activity has been established in our renal and

breast cancer cell culture models. It remains to be determined precisely how the

profound proliferation block of cancer cell culture is achieved. It is possible that

SOSTDC1 overexpression increases the rate of cancer cell apoptosis, or causes

a cell cycle block, or affects both processes. Furthermore, the therapeutic

potential of SOSTDC1 would need to be assessed to see if treatment with

157

recombinant protein at physiologic concentrations could obtain these anti-

proliferative results.

Some of the actions may be dependent on the binding partner(s) of SOSTDC1.

Yanagita and colleagues showed that USAG-1 (rat homologue of SOSTDC1)

binds to BMP-7 in vitro most tightly followed by , -2, -4, and -6 (26). Only

SOSTDC1’s interactions with BMPs-7 and -2 have been studied thus far. This is

significant particularly in regards to BMP-6, which often has contradictory roles to

BMP-7,-2, and -4. Additionally, all studies of BMP/SOSTDC1 interaction have

been in artificial systems, as it is currently unknown what physiologic

concentrations SOSTDC1 and interacting BMPs would achieve in vivo, or even

which BMP or Wnt ligands are co-expressed with SOSTDC1.

More thorough approaches are necessary to understand the interacting proteins

with SOSTDC1 - such as a yeast 2-hybrid system. Rationale for undertaking a

more thorough investigation of this matter comes from the example of sclerostin,

as it’s actions have been in part explained by its multiple binding partners (BMPs,

LRP-5, LRP-6). Sclerostin is able to inhibit both pathways and this is an

important regulatory mechanism of bone formation. Loss of BMP signaling

decreases sclerostin expression. The resulting loss of sclerostin leads to an

increase in canonical Wnt signaling (27).

158

Interestingly, Itasaki et al. and Ohazama et al. provide evidence that SOSTDC1’s

Xenopus ortholog binds to the Wnt co-receptors LRP-6 and LRP-4. LRP-4

(Lipoprotein related peptide-4) is homologous to the more thoroughly studied

LRP-5 and -6 Wnt co-receptors and is hypothesized to regulate sonic hedgehog

(Shh) signaling between epithelial and mesenchymal cells (28). The

hypothesized main role of these co-receptors is to stabilize the Wnt-Frizzled

complex. An interesting possibility raised by Itasaki and colleagues is that

SOSTDC1 is able to simultaneously bind BMPs and LRP-4. This could

effectively antagonizing both pathways concurrently- depending upon the

expression levels of BMP binding partners and LRP-4- as it would sequester

BMPs away from the cognate receptor complexes while disrupting the Wnt-

Frizzled interaction. The biologic consequences of SOSTDC1’s ability to interact

in this manner remain to be determined in various human tissues and disease

processes.

Early evidence points to SOSTDC1 having a greater effect on proliferation of

renal clear cell cultures than either DKK1 (Wnt only antagonist) or noggin (BMP

only antagonist). Whether this effect is due to a synergistic effect of concurrent

antagonism, sequential antagonism, or another phenomenon altogether will be a

subject of future investigation. Experiments will be necessary in the renal and

breast cells to determine with which proteins SOSTDC1 is interacting, and

whether both pathways are active in a particular tumor or cancer model. One

general starting point could be to check for SOSTDC1 association with

159

expressed BMPs, receptors, LRPs, and Wnts via fluorescent labeling and

confocal microscopy. Pathway activation could be established via staining for

phospho-Smad or active-β-catenin in clinical samples or use of luciferase-based

reporter assays in cell culture based models. Assuming the active intracellular

messangers exist, mutants or knockdowns of LRP proteins, disheveled, BMP

receptors, Smads, or other pathway players shown to interact with SOSTDC1

would be necessary to being characterizing each pathway’s contribution to the

overall phenotype.

When considering the possible mechanisms of action, it is necessary to take into

account the great potential for pathway cross-talk. The phenotypic effects of

SOSTDC1 may not be due to any primary action of the molecule itself, but a

result of expression level changes of molecules regulated by Wnt and/or BMP

signaling. This is not only inherent in the nature of SOSTDC1 as a dual-pathway

antagonist, but it is also compounded given the breadth of cellular changes

affected by BMP and Wnt signaling (29). The strong associations between

SOSTDC1 and EGFR, IGFR-1, and PTEN expression in breast cancer tumors

hints at the intertwined nature of these pathways. For example, inhibition of Wnt

signaling can transactivate EGFR (30). Microarray expression experiments in

cells treated plus or minus SOSTDC1 might begin to identify which pathways are

most critical.

160

4) Tissue Specificity: The Complexity of Multiple Pathway Involvement

In the previous section, the point was discussed that the mechanisms of

SOSTDC1 actions is likely dependent upon its binding partners and the result of

modulations in both Wnt and BMP pathways. As SOSTDC1 may have the

potential to concurrently inhibit both pathways, this introduces complexity into the

model of SOSTDC1’s actions in cancer. This complexity is illustrated when you

compare the differences in the models of SOSTDC1 in kidney cancers versus

breast cancers.

For example, the kidney is the predominant location for expression of BMP-7 in

the adult and BMP-7 is the main BMP ligand expressed in normal and disease

states. Additionally, Yanagita, et al. report that SOSTDC1 is the BMP antagonist

expressed at the highest levels in the kidney (26). In agreement, data in this

project (Chapter 2) showed fairly consistent levels of SOSTDC1 message and

protein staining in normal kidney samples, particularly in the distal tubules.

There was also a very penetrant significant decrease in SOSTDC1 in renal cell

carcinoma samples. Taken together, these data suggests that in the kidney, the

BMP7/SOSTDC1 axis is the primary BMP pathway control system and it is the

ratio between these two proteins that is important in the modulation of disease.

While data does not exist for any pro-carcinogenic actions of BMP-7 on the

kidney itself, BMP-7 does play a role in mediating TGF-β signaling, which is

161

connected with kidney tumors as it balances the epithelial-mesenchymal

transition step of dedifferentiation during transformation (4).

The connections between Wnt signaling and kidney cancer are newly expanded

with the discovery of WTX (31) and identification of the anti-proliferative

properties of sFRP (32). SOSTDC1 appears to have preferential antagonism of

Wnt ligands as Wnt3a antagonism has been observed in this work (Chapter 2),

while previous studies show no antagonism of Wnt1 (26). The significance of

these observations remains to be determined as little is known about the effects

of individual Wnt proteins in kidney tumors.

In contrast to the kidney, breast tissue expresses a many BMPs and their

expression and activities change in diverse ways throughout the spectrum of

transformation (discussed in Chapter 4). The model of the BMP-7/SOSTDC1

axis of control would not fit here; instead, more investigation is needed into the

relationship between the network of BMPs and Wnts expressed. Results of

these studies found widely varying basal expression levels of both SOSTDC1

message and protein in normal tissues. Additionally, SOSTDC1 levels changed

in varying amounts from normal to tumor tissues and did not correlate with a

specific pathologic type of breast cancer. This suggests that a more complex

network of SOSTDC1 regulation might exist in breast tissue. SOSTDC1

regulation of Wnt signaling in breast tissue is yet to be explored. This is likely a

162

promising area of future investigation as there are many reports of changes in

Wnt signaling that correlate with the progression of breast cancer.

Overall Conclusions

Simplification of carcinogenesis models by focusing on extracellular regulation of

BMP and Wnt signaling pathways led to the identification of SOSTDC1 as a

novel dual-pathway antagonist. Loss of heterozygosity has been shown to affect

10-15% of kidney cancers, the striking downregulation of SOSTDC1 in breast

and kidney tumors has been more carefully characterized, the extracellular

antagonistic capabilities of SOSTDC1 in cell culture models established, and

SOSTDC1’s potential to block proliferation identified.

These findings suggest that SOSTDC1 is an important regulator of both Wnt and

BMP pathways and that loss of SOSTDC1’s activities may contribute to more

advanced disease. SOSTDC1’s actions on multiple pathways increases its

attractiveness as a potential anti-cancer therapeutic. Future investigations of this

potential should occur by: greater characterization of the SOSTDC1 locus, study

of SOSTDC1’s actions on in vivo models, further exploration of SOSTDC1 as a

prognostic biomarker, and biochemical identification of the mechanism behind

proliferation suppression.

163

Reference List

(1) Macheda ML, Stacker SA. Importance of Wnt signaling in the tumor stroma microenvironment. Curr Cancer Drug Targets 2008 Sep;8(6):454-65.

(2) Sneddon JB, Zhen HH, Montgomery K, van de RM, Tward AD, West R, et al. Bone morphogenetic protein antagonist gremlin 1 is widely expressed by cancer-associated stromal cells and can promote tumor cell proliferation. Proc Natl Acad Sci U S A 2006 Oct 3;103(40):14842-7.

(3) Sudo S, Avsian-Kretchmer O, Wang LS, Hsueh AJW. Protein related to DAN and cerberus (PRDC) is a BMP antagonist that participates in ovarian paracrine regulation. J Biol Chem 2004;M402376200.

(4) Zeisberg M, Kalluri R. Reversal of experimental renal fibrosis by BMP7 provides insights into novel therapeutic strategies for chronic kidney disease. Pediatr Nephrol 2008 Sep;23(9):1395-8.

(5) Beppu H, Mwizerwa ON, Beppu Y, Dattwyler MP, Lauwers GY, Bloch KD, et al. Stromal inactivation of BMPRII leads to colorectal epithelial overgrowth and polyp formation. Oncogene 2008 Feb 14;27(8):1063-70.

(6) Yanagita M, Okuda T, Endo S, Tanaka M, Takahashi K, Sugiyama F, et al. Uterine sensitization-associated gene-1 (USAG-1), a novel BMP antagonist expressed in the kidney, accelerates tubular injury. J Clin Invest 2006 Jan;116(1):70-9.

(7) Kassai Y, Munne P, Hotta Y, Penttila E, Kavanagh K, Ohbayashi N, et al. Regulation of mammalian tooth cusp patterning by ectodin. Science 2005 Sep 23;309(5743):2067-70.

(8) Shigetani Y, Howard S, Guidato S, Furushima K, Abe T, Itasaki N. Wise promotes coalescence of cells of neural crest and placode origins in the trigeminal region during head development. Dev Biol 2008 Jul 15;319(2):346-58.

(9) Yanagita M. Modulator of bone morphogenetic protein activity in the progression of kidney diseases. Kidney Int 2006 Sep;70(6):989-93.

164

(10) Nojima M, Suzuki H, Toyota M, Watanabe Y, Maruyama R, Sasaki S, et al. Frequent epigenetic inactivation of SFRP genes and constitutive activation of Wnt signaling in gastric cancer. Oncogene 2007 Jul 12;26(32):4699-713.

(11) Takagi H, Sasaki S, Suzuki H, Toyota M, Maruyama R, Nojima M, et al. Frequent epigenetic inactivation of SFRP genes in hepatocellular carcinoma. J Gastroenterol 2008;43(5):378-89.

(12) Sogabe Y, Suzuki H, Toyota M, Ogi K, Imai T, Nojima M, et al. Epigenetic inactivation of SFRP genes in oral squamous cell carcinoma. Int J Oncol 2008 Jun;32(6):1253-61.

(13) Suzuki H, Toyota M, Carraway H, Gabrielson E, Ohmura T, Fujikane T, et al. Frequent epigenetic inactivation of Wnt antagonist genes in breast cancer. Br J Cancer 2008 Mar 25;98(6):1147-56.

(14) Wagner KD, Wagner N, Schedl A. The complex life of WT1. J Cell Sci 2003 May 1;116(Pt 9):1653-8.

(15) Thierry-Mieg D, Thierry-Mieg J. AceView: a comprehensive cDNA-supported gene and transcripts annotation. Genome Biol 2006;7 Suppl 1:S12-S14.

(16) Sutherland MK, Geoghegan JC, Yu C, Winkler DG, Latham JA. Unique regulation of SOST, the sclerosteosis gene, by BMPs and steroid hormones in human osteoblasts. Bone 2004 Aug;35(2):448-54.

(17) Itasaki N, Jones CM, Mercurio S, Rowe A, Domingos PM, Smith JC, et al. Wise, a context-dependent activator and inhibitor of Wnt signalling. Development 2003 Sep;130(18):4295-305.

(18) Guidato S, Itasaki N. Wise retained in the endoplasmic reticulum inhibits Wnt signaling by reducing cell surface LRP6. Dev Biol 2007 Oct 15;310(2):250-63.

(19) Clevers H. Wnt/beta-catenin signaling in development and disease. Cell 2006 Nov 3;127(3):469-80.

165

(20) Weidauer SE, Schmieder P, Beerbaum M, Schmitz W, Oschkinat H, Mueller TD. NMR structure of the Wnt modulator protein Sclerostin. Biochem Biophys Res Commun 2009 Feb 27;380(1):160-5.

(21) Lin SY, Chen CL, Wu YL, Yang YC, Hwu YM. Ratio of Wnt3a to BMP4 doses is critical to their synergistic effects on proliferation of differentiating mouse embryonic stem cells. Cell Prolif 2008 Jun;41(3):492-505.

(22) Tanaka M, Endo S, Okuda T, Economides AN, Valenzuela DM, Murphy AJ, et al. Expression of BMP-7 and USAG-1 (a BMP antagonist) in kidney development and injury. Kidney Int 2008 Jan;73(2):181-91.

(23) Nohe A, Keating E, Underhill TM, Knaus P, Petersen NO. Effect of the distribution and clustering of the type I A BMP receptor (ALK3) with the type II BMP receptor on the activation of signalling pathways. J Cell Sci 2003 Aug 15;116(Pt 16):3277-84.

(24) Nohe A, Hassel S, Ehrlich M, Neubauer F, Sebald W, Henis YI, et al. The mode of bone morphogenetic protein (BMP) receptor oligomerization determines different BMP-2 signaling pathways. J Biol Chem 2002 Feb 15;277(7):5330-8.

(25) Winkler DG, Yu C, Geoghegan JC, Ojala EW, Skonier JE, Shpektor D, et al. Noggin and sclerostin bone morphogenetic protein antagonists form a mutually inhibitory complex. J Biol Chem 2004;M400521200.

(26) Yanagita M, Oka M, Watabe T, Iguchi H, Niida A, Takahashi S, et al. USAG-1: a bone morphogenetic protein antagonist abundantly expressed in the kidney. Biochem Biophys Res Commun 2004 Apr 2;316(2):490-500.

(27) Kamiya N, Ye L, Kobayashi T, Mochida Y, Yamauchi M, Kronenberg HM, et al. BMP signaling negatively regulates bone mass through sclerostin by inhibiting the canonical Wnt pathway. Development 2008 Nov;135(22):3801-11.

(28) Ohazama A, Johnson EB, Ota MS, Choi HY, Porntaveetus T, Oommen S, et al. Lrp4 modulates extracellular integration of cell signaling pathways in development. PLoS ONE 2008;3(12):e4092.

166

(29) Guo X, Wang XF. Signaling cross-talk between TGF-beta/BMP and other pathways. Cell Res 2009 Jan;19(1):71-88.

(30) Schlange T, Matsuda Y, Lienhard S, Huber A, Hynes NE. Autocrine WNT signaling contributes to breast cancer cell proliferation via the canonical WNT pathway and EGFR transactivation. Breast Cancer Res 2007;9(5):R63.

(31) Major MB, Camp ND, Berndt JD, Yi X, Goldenberg SJ, Hubbert C, et al. Wilms tumor suppressor WTX negatively regulates WNT/beta-catenin signaling. Science 2007 May 18;316(5827):1043-6.

(32) Gumz ML, Zou H, Kreinest PA, Childs AC, Belmonte LS, LeGrand SN, et al. Secreted frizzled-related protein 1 loss contributes to tumor phenotype of clear cell renal cell carcinoma. Clin Cancer Res 2007 Aug 15;13(16):4740-9.

167

SCHOLASTIC VITA

KIMBERLY ROSE BLISH

BORN: June 18, 1981, Winston-Salem, North Carolina UNDERGRADUATE Eastern Nazarene College STUDY: Quincy, Massachusetts B.S., Chemistry, summa cum laude 2002 GRADUATE Wake Forest University Winston-Salem, North Carolina Ph.D. 2009 SCHOLASTIC AND PROFESSIONAL EXPERIENCE: MD/PhD Graduate Fellowship, 2002 -2009 Teaching and Laboratory Assistant, Eastern Nazarene College, 1999-2002.

Summer undergraduate research project: QSAR Modeling to Predict Blood-Brain Barrier Partitioning, 2001.

HONORS AND AWARDS: Predoctoral Traineeship Award: BC043274, “A Novel Bone Morphogenetic Protein

Antagonist as a Breast Cancer Growth Suppressor.” Breast Cancer Research Program, CDMRP, Department of Defense, 2005. 1st Place Student Presentation, WFU Cancer Biology Department, 2005 & 2008.

Senior Award in Biochemistry- American Institute of Chemists, 2002.

James Norris-Richard Flack Summer Undergraduate Research, Scholarship. Northeastern Section of the American Chemical Society, 2001. National Merit Scholarship Award, 1998.

PROFESSIONAL SOCIETIES:

2007-Present American Association for Cancer Research (AACR)

2004-06 American Physician Scientist Association (APSA) Wake Forest University Institutional Representative.

2002-04 American Medical Student Association (AMSA)

168

2002-03 American Institute of Chemists (AIC)

PUBLICATIONS: Articles: Blish, KR, Willingham MC, Du W, Birse CE, Krishnan SR, Brown JC, Wang W, Hawkins GA, Garvin AJ, D'Agostino RB, Torti SV, Torti FM. A Novel, Human Bone Morphogenetic Protein Antagonist is Down regulated in Renal Cancer. Mol. Biol. Cell, 2008 Feb; 19(2):457-464 E-published: 2007 Nov 21. Blish, K.R.; Ibdah, J.A. Maternal heterozygosity for a mitochondrial trifunctional protein mutation as a cause for liver disease in pregnancy. Med. Hypotheses. 2005. 64(1): 96-100. Rose, K.; Hall, LH. E-State Modeling of Fish Toxicity Independent of 3D Structure Information. SAR QSAR Environ. Res. 2003. 14 (2): 113-129 Rose, K.; Hall, LH; Kier, LB. Modeling blood-brain barrier partitioning using the electrotopological state. J. Chem. Inf. Comput. Sci. 2002. May – June; 42(3): 651-66. Abstracts: June 2008: Blish KR, Triplette, MA, Willingham MC, Kute TE, Du W, Birse CE, Krishnan, SR, Russell GB, Brown JC, Torti FM, and Torti SV. A Bone Morphogenetic Protein Antagonist as a Breast Cancer Growth Suppressor. Poster for DOD Era of Hope Meeting. Baltimore Convention Center, Baltimore, MD, June 26, 2008. April 2007: Annual Meeting of the American Association for Cancer Research (AACR), Los Angeles, CA. Kimberly R. Blish, Mark C. Willingham, Wei Du, Charles E. Birse, Surekha R. Krishnan, Julie C. Brown, Wei Wang, Gregory A. Hawkins, A. Julian Garvin, Ralph B. D'Agostino Jr., Frank M. Torti, Suzy V. Torti. The human bone morphogenetic protein antagonist BARC is downregulated in renal cancers In: American Association for Cancer Research Annual Meeting: Proceedings; 2007 Apr 14-18; Los Angeles, CA. Philadelphia (PA): AACR; 2007. Abstract #3785. April 2006: BARC: A Novel, Human BMP Antagonist Downregulated in Breast and Renal Cancers. K.R. Blish, M.C. Willingham, W. Du, C. Birse, J.C. Brown, M.A. Triplette, A.J. Garvin, F.M. Torti, S.V. Torti. Poster for ASCI/AAP Joint Meeting. The Fairmont Hotel, Chicago, Illinois, April 29, 2006. January 2004:SCGF in Breast Cancer Growth and Detection. K.R. Blish, J. Brown, M. Willingham, C. Birse, F.M. Torti, and S.V. Torti. Poster for the Breast Cancer Center of Excellence Retreat. Comprehensive Cancer Center of Wake Forest University. January 15, 2004. Winston-Salem, NC. May 2003 Rose, K.; Zhao, Y.; Ibdah, JA. Liver Disease in Pregnancy Associated with Maternal Heterozygosity for A Mitochondrial Trifunctional Protein Mutation and A Normal Fetal Genotype. For presentation at Digestive Disease Week, the joint conference of the AASLD, AGA, ASGE, and SSAT. May 17-22, 2003. Orange County Convention Center, Orlando, Florida.