ENDOCRINOLOGY
`
`OE BREAST CANCER
`
`Edited by
`
`ANDREA MANNI, MD
`The Penny/[0472121 State University
`College ofMedicine, Hers/794 PA
`
`InnoPharma Exhibit 1069.0001
`
`HUMANA PRESS
`TOTOWA, NEW JERSEY
`
`l \
`
`\w
`/
`
`
`
`

`

`
`
`© 1999 Humana Press Inc.
`999 Riverview Drive, Suite 208
`Totowa, New Jersey 07512
`
`For additional copies, pricing for bulk purchases, and/or information about other Humana titles,
`contact Humana at the above address or at any ofthe following numbers: Tel: 973-256-1699;
`Fax: 973-256-8341; E—mail: humana@humanapr.c0m or visit our website at http://hutnanapress_com
`
`All rights reserved. No part ofthis book may be reproduced, stored in a retrieval system, or transmitted in any form or
`by any means, electronic, mechanical, photocopying, microfilming, recording, or otherwise without written permission
`from the Publisher.
`
`All articles, comments, opinions, conclusions, or recommendations are those of the author(s), and do not necessarily reflect
`the views ofthe publisher.
`
`This publication is printed on acid-free paper. ®
`ANSI 239.48-1984 (American National Standards lnstitute)
`Permanence ofl’aper for Printed Library Materials.
`
`Photocopy Authorization Policy:
`Authorization to photocopy items for internal or personal use, or the internal or personal use ofspecifie clients, is granted
`by Humana Press lnc., provided that the base fee of US $10.00 per copy, plus US $00.25 per page, is paid directly to
`the Copyright Clearance Center at 222 Rosewood Drive, Danvers, MA 01923. For those organizations that have been
`granted a photocopy license from the CCC, a separate system ofpayment has been arranged and is acceptable to Humana
`Press Inc. The fee code for users ofthe Transactional Reporting Service is: [0-89603-591-3/99 $10.00 + $00.25}
`
`Printed in the United StatesofAmerica.
`
`10
`
`9
`
`8
`
`7
`
`6
`
`5
`
`4
`
`3
`
`7
`
`1
`
`Library of Congress Cataloging-in-Publication Data
`
`Endocrinology of breast cancer/ edited by Andrea Manni
`p.
`cm.
`-(Contemporary endocrinology; 1 1)
`Includes index.
`ISBN 0-89603-591-3
`
`1. Manni, Andrea.
`l. Breast~Cancer—Endocrinc aspects.
`Contemporary endocrinology (Totwa, N. 1.):
`l l.
`
`11. Series:
`
`[DNLMz 1. Breast Neoplasms—physiopathology. 2. Breast Neoplasms—etiology.
`3. Breast Neoplasms—therapy. 4. Hormones. WP 87013565 1999]
`RC280.B8E54
`19999
`616.99‘449—dc21
`
`DNLM/DLC
`for Library of Congress
`
`98-33989
`ClP
`
`InnoPharma Exhibit 1069.0002
`
`

`

`This material may be protected by Copyright law (Title 17 US. Code)
`
`
`
`12
`
`Role of Angiogenesis in the
`
`Transition to Hormone
`
`Independence and Acquisition
`of the Metastatic Phenotype
`
`
`Francis G. Kern, PHD
`
`INTRODUCTION
`
`Breast cancer becomes a life-threatening illness when the malignant tumor cells have
`acquired the capability of invading through the basement membrane and into the circu—
`lation, where they are then disseminated to distant sites in the body. Therefore it may
`seem intuitive that tumors that have acquired a capability of stimulating the growth of
`new blood vessels would possess a greater threat to the patient, and indeed a good amount
`of correlative clinical data now link tumors that have a higher number of microvessels
`with poor prognosis (1—5). However, while neoangiogenesis is likely to be important in
`the dissemination of tumor cells, it is becoming more apparent that additional factors are
`required to allow proliferation of disseminated tumor cells at the distant sites.
`When breast cancer presents as an invasive disease, it has been useful to use estrogen
`receptor (ER) content as a basis for choice oftherapy. This relates to the fact that estrogen
`itself can stimulate the growth of breast cancer cells that contain ER through still
`unknown mechanism that involves binding of estrogen to nuclear ER and subsequent
`stimulation of its transcriptional activation functions. Consequently, a highly useful
`treatment for most patients with ER-positive tumors is hormonal therapy with
`antiestrogens, the most common being tamoxifen (6-8). Although other ER-independent
`mechanisms may also contribute to the therapeutic effect of this agent (9—11), it is
`commonly assumed that the benefit of this drug derives from its ability to bind to ER
`without stimulating the same type oftranscriptional activity as an estrogen-bound receptor.
`An invariable and unfortunate outcome of treatment of patients with tumors that
`initially respond to tamoxifen is the outgrowth of populations of cells that no longer
`respond to this treatment. In addition, some ER-positive tumors fail to Show any initial
`response to this relatively benign form of treatment. What may not seem so intuitive is
`how expression of angiogenic growth factors might also be involved in either acquired
`or de novo antiestrogen resistance, yet recent studies with transfected breast cancer cell
`lines (discussed in greater detail below) raise this as a possibility. If this turns out to be
`
`From: Contemporary Endocrinology: Endocrinology ofBreast Cancer
`Edited by: A. Manni © Humana Press Inc., Totowa, NJ
`
`169
`
`This material was {copied
`at the NLM and man: be
`
`InnoPharma Exhibit 1069.0003
`
`

`

`170
`
`Part III / Tumor Progression
`
`substantiated by clinical studies, then in the appropriate context, treatment with agents
`that inhibit angiogenesis may afford a means of restoring antiestrogen sensitivity to
`patients with acquired resistance or a means ofconferring sensitivity to patients with de
`novo resistance.
`
`CLINICAL DILEMMAS
`
`ASSOCIATED \WITH TAMOXIFEN TREATMENT
`
`A major dilemma for the clinician currently confronted with a patient with an ER—
`positive breast tumor is trying to make the best guess as to whether a patient will respond
`to this type of antiestrogen therapy in either an adjuvant or advanced breast cancer
`treatment setting. Since approximately 60% of patients with invasive ductal carcinoma
`present with ER-positive tumors, this particular question is relevant for a significant
`portion of women with breast cancer. It is also known that around 65—70% of patients
`with ER-positive tumors with advanced stage disease will show at least some response
`to tamoxifen (12,13). Among those that eventually fail on this therapy, 50—70% of these
`patients still have tumors that contain ER when these sites are biopsied (12—15). There—
`fore loss ofER is not the primary mechanism involved in the acquisition of antiestrogen
`resistance.
`
`Two different types of second-line antihormonal therapies are currently in the final stages
`of clinical trials for use in patients who have failed on tamoxifen. One approach involves the
`use of inhibitors of the enzyme aromatase, which is involved in the biosynthesis of estradiol
`produced at extraovarian sites (16—23). A second form of second-line therapy is the use of
`what have been called pure antiestrogens (24). These are compounds that bind to the ER,
`result in no transcriptional activation, and may increase the degradation of the receptor (25).
`They also do not possess any of the partial agonist properties of tamoxifen.
`Various recent studies have shown that 20~40% of patients who have failed on tam-
`oxifen therapy show some response to these second-line therapies and that an additional
`30% show no immediate disease progression when switched to these forms of therapy
`(19—23,26——28). Therefore these second-line approaches are likely to be of some value,
`but they also raise another dilemma for the oncologist. In addition to needing help from
`molecular oncologists and epidemiologists in determining who is most likely to respond
`initially to tamoxifen, the clinician will soon also need assistance in determining which
`patients who do not respond are likely to benefit from these alternative hormonal therapies.
`
`MECHANISMS OF TAMOXIFEN RESISTANCE
`
`A number of different mechanisms have been proposed for tamoxifen resistance.
`Based on what is known about some of these mechanisms, some educated guesses can
`be made as to whether the patient with a tamoxifen-resistant tumor that is due to one of
`these mechanisms is likely to respond to a second-line therapy. However, for many of
`these mechanisms for which a molecular basis for tamoxifen resistance has been demon-
`strated with cell lines or with in vitro assays, subsequent examination of patient samples
`has revealed that these types of alterations are unlikely to be responsible for most
`tamoxifen—resistant tumors. This includes point mutations in the hormone binding domain
`of the estrogen receptor, alterations in splicing of the ER mRNA that can result in con-
`stitutiver active forms of the receptor, or metabolism of tamoxifen to metabolites with
`estrogenic activity (8,29—35).
`
`This material was {a pied
`at the NLM and mav bie-
`
`InnoPharma Exhibit 1069.0004
`
`

`

`
`Chapter 12 / Angiogenic Growth Factors and Breast Cancer Progression
`171
`
`There is now recent evidence that the balance in the amounts of recently discovered
`steroid receptor coactivators and corepressors can influence whether tamoxifen will act
`as an antagonist or agonist of the estrogen receptor. At this point, work with in vitro
`systems suggests that the pure antiestrogens will still act as antagonists when tamoxifen
`is changed to an agonist as a result ofalterations in the ratio ofcoactivators to corepressors
`(36). Therefore other mechanisms are also probably responsible for those ER-positive
`tamoxifen-resistant breast tumors that fail to respond to second—line therapies.
`Two potential mechanisms involve the downstream signaling events associated with
`activation of growth factor receptors. Work from a number of laboratories has demon-
`strated what has been termed cross talk between steroid receptors and either growth factor
`tyrosine kinase receptors or G-protein-coupled receptors (3 7—44). In examples relevant
`to the question oftamoxifen resistance, phosphorylation ofthe estrogen receptor by MAP
`kinases, which are downstream effectors of a number of different signaling pathways
`including those mediated by growth factor receptors, or activation of G-protein-coupled
`receptors can result in tamoxifen acquiring agonistic properties (3 7,38, 43). However, the
`same studies show that the ER is not activated by these mechanisms when pure
`antiestrogens are present (3 7—43).
`Therefore at least one additional mechanism probably remains underlying the de novo
`or acquired tamoxifen resistance of ER-positivc tumors that do not subsequently respond
`to a second—line treatment with the pure antiestrogens. Recent data suggest that one such
`mechanism may also involve growth factor signaling, which provides the breast tumor
`cell with an alternative pathway to the growth stimulation mediated by ER (45). This
`would bypass the requirement for activation ofthe ER by an agonist for growth promotion
`. to occur and would account for the failure of second-line antihormonal therapies.
`
`ROLE OF GROWTH FACTORS IN THE
`
`ACQUISITION OF TAMOXIFEN RESISTANCE
`
`It can now be seen that growth factor signaling can affect the resistance of a breast
`tumor cell to tamoxifen in a number of different ways, with two of these mechanisms
`potentially leading to tamoxifen acting as an agonist. In one scenario, growth factor
`signaling can affect the level of coactivators or corepressors. Although no experimental
`data currently exist indicating that this type of regulation is indeed occurring in breast
`cancer cells, the field is very new, the number of coactivators and corepressors being
`identified is continually expanding, and the experiments remain to be performed (46). In
`a second scenario, growth factor signaling could result in activation of ER via phospho—
`rylation, leading to increased agonistic activity for tamoxifen. In the third scenario,
`growth factor signaling could bypass the requirement for ER activation for growth. These
`three scenarios are not mutually exclusive, and all three mechanisms may be operating
`simultaneously within a cell to provide additive or perhaps synergistic growth-stimula—
`tory effects. As such, these mechanisms may account for the clinical response observed
`in some patients when tamoxifen is withdrawn (4 7,48). However the available evidence
`to date suggests that a better understanding is required of what particular growth factor
`signaling pathways can bypass the need for ER since these pathways may also be respon-
`sible for resistance to second—line therapies.
`The initial indication of the potential importance growth factors might have in the
`process of the development of hormone independence came from work by Dickson and
`
`This material was tooled
`at the NLM and man: be
`
`InnoPharma Exhibit 1069.0005
`
`

`

`172
`
`Part III / Tumor Progression
`
`Lippman (49). These two investigators formulated the ideas that the genes for growth
`factors were part of the set of genes that were transcriptionally activated in ER-positive
`cells in response to estrogen treatment and that the subsequent growth stimulation of
`ER-positive breast cancer cell line xenografts in nude mice was the result of either the
`autocrine or paracrine effects ofthese factors (49). This line ofreasoning was supported
`by findings that continuous administration of epidermal growth factor (EGF) could
`facilitate the hormone-independent growth of an estrogen-dependent cell line as xe-
`nografts in ovariectomized athymic nude mice (50) and that transforming growth fac—
`tor—0t (TGF-oc), another ligand for the EGF receptor, was induced by estrogen in the
`same cell line (51—54).
`Subsequently a number of laboratories have used a transfection approach to express
`constitutively growth factors or growth factor receptors whose overexpression was impli-
`cated by various clinical studies to be associated with estrogen-independent tumors or
`induced by estrogen in vitro (39, 55—66). Molecules such as protein kinase c (PKC)-OL,
`ras“, and raf-l, which are downstream effectors of many growth factor signaling path-
`ways, have also been overexpressed in estrogen-dependent cell lines (6 7—72). The trans-
`fected cell lines were then tested to determine whether the overexpression had any effect
`on the estrogen responsiveness or antiestrogen sensitivity in vitro or in vivo.
`Many of these studies utilized the MCF-7 breast cancer cell line as a recipient since
`it is estrogen responsive in vitro and sensitive to treatment with antiestrogens (73). In
`vivo, there is a strict estrogen dependence for tumor formation, and tumors will not
`develop in ovariectomized nude mice unless they are supplemented with estrogen (74, 75).
`Even though this cell line was derived from the pleural effusion of a patient with breast
`cancer (76), the tumors that do develop with this cell line and most other breast cancer
`cell lines capable of in vivo growth are noninvasive and nonmetastatic (69, 77). This in
`vivo phenotype seen with MCF-7 cells is in line with the poor invasive ability seen in in
`vitro Matrigel basement membrane invasion assays (78). Consequently, this cell also
`represents a useful model to determine the effects ofoverexpression on the metastatic and
`invasive phenotypes. To facilitate this process, we developed a clonal line of MCF-7
`cells, designated as the ML-20 cell line, which stably expresses a transfected bacterial
`lacZ gene (77, 79). In many of our studies we have used this cell line as a recipient for
`retransfection since this allows a chromogenic substrate for B-galactosidase to be used
`to stain metastatic deposits of cells blue.
`
`EFFECTS OF ANGIOGENIC GROWTH FACTOR
`
`OVEREXPRESSION ON THE METASTATIC PHENOTYPE
`
`Varying degrees of increased estrogen-independent in vitro or in vivo growth can be
`conferred to ER-positive breast cancer cell lines by transfection with vectors for EGFR
`(56,5 7, 64) HER-2/neu (62, 63), heregulin (39,66), activated rasH (67—70), constitutively
`activated raf—l (71), PKC-(x (72), TGF-B (61), insulin-like growth factor (IGF)-II (59, 60),
`fibroblast growth factor (EGF)-l (65) and FGF-4 (55). In those transfection studies in
`which estrogen-independent in vivo growth was a consequence ofoverexpression, it was
`only with the PKC-a and FGF-1 and FGF-4 transfectants that an increased metastatic
`phenotype was also consistently observed (55, 65, 72). F GF-overexpressing cells acquire
`the ability to form tumors in ovariectomized athymic nude mice or beige mice without
`estrogen supplementation. They can also form tumors in either nude or beige mice treated
`
`ThiS material was rapied
`at the NLM arm max: ma-
`
`InnoPharma Exhibit 1069.0006
`
`

`

`Chapter 12 / Angiogenic Growth Factors and Breast Cancer Progression
`
`173
`
`with tamoxifen pellets. In either type ofmice, and in either untreated or tamoxifen-treated
`mice, X-gal-stained cells can be readily detected in either the lungs or lymph nodes of
`tumor-bearing mice injected with FGF-l- or FGF-4 overexpressing cells (45, 65, 77). By
`contrast, no cells are detected in lungs or lymph nodes of mice bearing control vector-
`transfected ML-20 cell tumors in estrogen-supplemented mice, even when these tumors
`are the same size as the tumors that form in mice injected with FGF-overexpressing cells
`(65). However, if the same ML-20 cells are coinjected with FGF-4-overexpressing cells
`that do not themselves express B—galactosidase, X—gal-stained cells can be detected in the
`tumors of either untreated or tamoxifen~treated ovariectomized nude mice and in the
`
`lungs and lymph nodes of mice bearing tumors composed of the mixture (55).
`The underlying mechanism responsible for the increased metastatic phenotype asso-
`ciated with PKC-oc overexpression remains to be elucidated. Both FGF-1 and FGF-4 are
`angiogenic growth factors and both can therefore stimulate the growth of new blood
`vessels in a tumor producing either factor (80,81). Thus the mechanism likely to be
`responsible for the increased diSSemination of tumor cells is likely to be related to the
`increased neoangiogenesis that is occurring in these tumors.
`Although deposits of X-gal-stained cells can be seen in essentially 100% ofthe mice
`bearing tumors of either FGF-4- or FGF- 1 ~0verexpressing cells, we never observed the
`development of tumor nodules at either of these sites (65, 77). Since the disseminated
`tumor cells express the cDNAs for both hygromycin and G418, it is a relatively simple
`task to establish cell lines from the tumor cells present in the lungs or lymph nodes. These
`cells continue to express FGF and remain capable of forming tumors in either untreated
`or ovariectomized mice. However, these cells also do not appear to be able to develop into
`macrometastases. When inoculated via a tail vein route, these tumor cells disappear from
`the lungs with approximately the same kinetics as FGF—overexpressing cells that had not
`been established from lungs of tumor-bearing mice. With both cell types, X-gal-stained
`cells can no longer be detected by 96 h after injection (82).
`Long-lived angiogenic inhibitors produced by the primary tumor cells have been
`shown to be capable of inhibiting the development of macrometastases from cells that
`have been disseminated from the primary site. In these model systems, macrometastases
`are found when the primary tumor is resected (83—86). However this was not the case
`when the primary tumors of FGF-l-overexpressing cells were resected. Tumors were
`allowed to develop for 3 weeks, at which point the tumors were of sufficient size that
`100% of the animals sacrificed at this point had large numbers of disseminated cells
`present in both the lungs and lymph nodes. Tumors were resected from the remaining
`animals and after 3 weeks, this group of animals was sacrificed. Examination of the lungs
`and lymph nodes did not reveal any X-gal-stained cells in these organs (82).
`Taken together, these results indicate that the increased neoangiogenesis brought
`about by the production of an angiogenic growth factor facilitates the constant shedding
`of tumor cells into the circulation. However, the cells that accumulate within the lungs
`and lymph nodes of tumor-bearing mice do not possess any selective advantage at these
`distant sites compared with the cells remaining in the primary tumor. These cells are
`lacking in the ability to either attach, extravasate, or proliferate at these distant sites; as
`a consequence, they are cleared from these organs within a relatively short period. The
`mechanism of clearance remains to be elucidated but may involve an increased rate of
`apoptosis relative to tumor cells at the primary tumor site or attack by cells of the remain-
`ing immune system present in the athymic nude mice.
`
`This material was {copied
`at the NLM and man: be
`
`InnoPharma Exhibit 1069.000? i
`
`

`

`174
`
`Part III / Tumor Progression
`
`In addition to facilitating the spread of tumor cells, the production of an angiogenic
`growth factor would presumably confer some growth advantage over parental MCF-7
`cells at these distant sites if the other missing components of the metastatic process were
`otherwise present. Consequently, the FGF-transfected cell lines are likely to be a useful
`model in determining the ability of the various proteases, attachment factors, apoptosis
`modulators, and other putative regulators ofthe metastatic process to facilitate the devel-
`opment of macrometastases in this system.
`
`MECHANISMS MEDIATING ESTROGEN—INDEPENDENT
`
`GROWTH IN CELLS OVEREXPRESSING GROWTH FACTORS
`
`A critical question is whether the estrogen-independent growth conferred by FGF
`overexpression is the result of a mechanism that activates ER or bypasses the need for
`an ER-regulated growth pathway. Heregulin overexpression can also confer on MCF-7
`cells the ability to form tumors in ovariectomized nude mice (39,66). Heregulin is a
`ligand for homodimers of erbB-3/HER-3 and erbB-4/HER-4 and heterodimers of
`erbB-Z/HER-Z/neu with either of these two transmembrane tyrosine kinase receptors
`(87,88). Recent clinical evidence shows that patients with tumors overexpressing
`HER-2/neu respond poorly to tamoxifen therapy (89), implying that activation of the
`erbB family of receptors is involved in resistance to this antiestrogen. In heregulin-
`overexpressing MCF-7 cells, the ER becomes tyrosine phosphorylated, and the num-
`bers ofER are downregulated (39, 66). There is also evidence from transient expression
`assays with estrogen response element (ERE)-containing reporter constructs that the
`ER is constitutively activated in heregulin-overexpressing cells and that this effect can
`be reversed by treatment with the pure antiestrogen ICI 182,780 (39). This raises the
`possibility that cells that have acquired resistance to tamoxifen through the activation of
`the erbB family of transmembrane tyrosine kinase receptors may be able to respond to
`this second-line antihorrnonal therapy.
`By contrast, the evidence available to date suggests that signaling mediated by FGF
`receptors may instead confer on ER-positive breast cancer cells a mechanism that can
`bypass the need for ER activation. FGF-overexpressing cells have obtained the ability to
`form soft agar colonies with high efficiency in media depleted ofestrogens. In media that
`contain estrogen, the effect of added 4-hydr0xytamoxifen is diminished compared with
`control transfected cells. The same holds true in anchorage-dependent growth assays in
`which FGF-overexpressing cells have faster doubling times in estrogen-depleted media
`or in estrogen-containing media with added 4-hydroxytamoxifen (65). The enhanced
`anchorage-dependent or anchorage-independent growth in estrogen-depleted media is
`not abrogated by treatment with the pure antiestrogen ICI 182,780, as might be expected
`if the effect were due to ligand-independent activation of ER mediated by FGF signaling
`pathways. Similarly, the FGF-transfected cells continue to show enhanced growth in
`either anchorage-dependent or anchorage-independent growth assays when ICI 182,780
`is added to estrogen-containing media. There is no downregulation of ER in FGF-trans-
`fected cells, and there is no evidence from analysis ofthe basal levels ofestrogen-induced
`genes or from transient assays with ERE-containing reporter constructs that FGF-
`overexpressing cells contain a constitutively activated ER. Furthermore, FGF-over-
`expressing cells continue to be able to form tumors in mice that have been treated with
`ICI 182,780 (45).
`
`This material was {sawed
`at the NLM and max Ea
`
`InnoPharma Exhibit 1069.0008
`
`

`

`Chapter 12 / Angiogenic Growth Factors and Breast Cancer Progression
`
`175
`
`ROLE OF PARACRINE AND AUTOCRINE EFFECTS
`
`OF FGF IN MEDIATING TAMOXIFEN RESISTANCE
`
`Although FGF—4 is not expressed in human breast tumor tissues, there is considerable
`evidence that both FGF-1 and all four of the FGF transmembrane tyrosine kinase receptors
`are expressed in many breast tumor tissues (65, 90—1 00). The transfection studies discussed
`above certainly suggest a potential role for FGF expression in the acquisition of hormone
`independence or antiestrogen resistance in ER-positive breast tumors. Since FGF-1 can
`have both paracrine and autocrine or intracrine effects, it is important to dissect the relative
`contributions of these effects to the phenotypes observed in FGF-l-overexpressing cells.
`This process can be performed by retransfecting the FGF-l-overexpressing cells with a
`dominant—negative FGF receptor. Receptor dimerization is required for transmission ofthe
`signal resulting from FGF binding (101,102). Therefore, a construct directing the expres-
`sion ofa transmembrane-anchored truncated FGF receptor lacking the cytoplasmic tyrosine
`kinase domain will effectively block the transmission of the growth signal through the
`formation of nonfunctional homo- and heterodimers. In such a pairing, the mutant lacking
`the tyrosine kinase domain will not be capable of phosphorylating the wild-type receptor
`on tyrosine residues in the cytoplasmic regulatory region of the receptor and thus will
`prevent the binding of downstream signaling molecules (103—105).
`FGF-l—overexpressing cells retransfected with the dominant negative FGF receptor no
`longer contained elevated levels ofphosphorylated MAPKs in cell extracts, indicating that the
`dominant negative receptor was blocking FGF signaling. As expected, these cells were also
`no longer capable of forming soft agar colonies in either estrogen-depleted or antiestrogen-
`containing media, confirming that autocrine or intracrine effects of FGF were important in
`conferring this phenotype. In line with these in vitro results, when they are injected into
`ovariectomized nude mice, the retransfected cells are no longer capable of forming tumors
`without estrogen supplementation. However, the tumors that form in estrogen-supplemented
`animals are typically much larger than those formed by MCF—7 cells, suggesting that the
`paracrine effects of FGF-1 production can act in an additive or synergistic manner with the
`mitogenic effect ofestrogen to stimulate tumor growth. An interesting finding from this study
`was the observation that in tamoxifen-treated animals, the FGF—overexpressing cell lines
`formed tumors at the same rate regardless of whether or not they were also expressing the
`dominant negative receptor. Thus the paracrine effects alone ofFGF were capable ofallowing
`antiestrogen-resistant in vivo tumor growth (106). These results also indicate that tamoxifen
`resistance can be dissociated from estrogen independence.
`It has been suggested that the partial agonistic properties of tamoxifen may be greater in the
`mouse than they are in the human (107). When tamoxifen-treated mice are injected with MCF—7
`cells that do not overexpress FGF, small tumors sometimes develop, but they usually remain
`static or regress (75, I 08). Thus it is possible that the paracrine effects ofFGF may cooperate with
`the weak mitogenic activity of tamoxifen to allow continued and progressive tumor develop-
`ment to occur. While this an intriguing possibility, it remains to be determined if these results
`seen in nude mice have any relevance to the tamoxifen resistance that occurs in human patients.
`
`EFFECTS OF VEGF OVEREXPRESSION ON ANTIESTROGEN
`
`SENSITIVITY AND THE METASTATIC PHENOTYPE
`
`As previously stated, FGF-1 is a strong angiogenic growth factor (80). Therefore, one
`of the paracrine effects that is likely to be a strong contributor to the observed tumor
`
`This material was {retried
`at the NLM and man: be
`
`InnoPharma Exhibit 1069.0009
`
`

`

`176
`
`Part III / Tumor Progression
`
`growth of the dominant-negative receptor-transfected cell lines in tamoxifen-treated
`animals is the neoangiogenesis resulting from the continued production of this growth
`factor. To test the effect of neoangiogenesis directly, B-galactosidase-expressing MCF-7
`cells were transfected with a vector directing the overexpression of the l65-amino acid
`form of vascular endothelial growth factor (VEGF). There is a considerable amount of
`evidence that expression of this particular isoform of this growth factor is an important
`determinant in the establishment ofa neoangiogenic phenotype in many types ofcancers.
`VEGF has been shown to be expressed in breast tumors (93,109—113), and the extent of
`expression correlates with the number of microvessels present in the tumor (109). This
`is a growth factor that is relatively specific for endothelial cells. There are two transmem-
`brane tyrosine kinase receptors for VEGF. One is known as Flkl in the mouse and as KDR
`in the human; the second is called Flt-1 (114—11 7). Expression of both receptors is
`restricted to endothelial cells, hematopoietic stem cells, monocytes, megakaryocytes,
`and platelets (1 16—120). mRNA for these receptors is strongly expressed in the endothe—
`lial cells of small vessels adjacent to infiltrating ductal carcinoma or metastatic ductal
`carcinoma cells as well as in the endothelial cells adjacent to comedo-type ductal carci-
`noma in situ. However, there is no evidence ofexpression ofeither receptor in endothelial
`cells in the normal areas of the breast (113).
`There are a few reports of ectopic expression of VEGF receptors in some types of
`tumor cells (120—123), but an examination of the recipient ML-20 cell line using a
`sensitive reverse transcription polymerase chain reaction assay has failed to show any
`evidence of expression of either receptor in this MCF-7-derived breast cancer cell line.
`In line with this finding, cells overexpressing either the 121- or 165-amino acid forms of
`VEGF do not show any altered in vitro growth properties with regard to growth in either
`estrogen-depleted media or in antiestrogen-containing media (124). Both types ofVEGF
`transfectants are unable to form tumors in ovariectomized mice without estrogen supple—
`mentation. Overexpression ofthe 121-amino acid form ofVEGF was reported to increase
`the in vivo growth rate in estrogen-supplemented animals (125), but no statistical differ-
`ence was observed between control transfectants and transfectants overexpressing the
`l65-amino acid form of VEGF in this treatment group (124). This may relate to the fact
`that different sublines of MCF-7 cells have different in vivo growth rates in estrogen-
`supplemented mice and that the controls for the 121-amino acid transfection study grew
`poorly whereas the controls for the 165—amino acid study grew well in estrogen-supple—
`mented mice, making any further stimulation due to VEGF overexpression difficult to
`observe.
`
`No effect of overexpression of the 121-amino acid form of VEGF on either tamoxifen
`sensitivity or metastatic potential was seen in the one study. However, overexpression of
`the 165-amino acid form was found to increase the tumorigenicity of MCF-7 cells in
`tamoxifen-treated animals and to facilitate tumor cell dissemination. For the effect of
`
`tamoxifen to be observed, coinjection with Matri gel was required (124). This is a mixture
`of basement membrane components that has been demonstrated to increase the tumori-
`genicity of a number of human cancer cell lines, including MCF~7 cells (126,127). The
`12 1 -amino acid form ofVEGF does not bind to heparin, whereas the 165-amino acid form
`is a heparin-binding growth factor (128). It is difficult to compare studies with two
`different growth factor isoforms performed in two different laboratories. The difference
`in effects observed may simply relate to differences in expression levels. However, it is
`also possible that this difference in heparin binding capability may also explain the
`Th is material was {opiad
`at the NLM arm mav Ea-
`
`InnoPharma Exhibit 1069.0010
`
`

`

`
`
`Chapter 12 / Angiogenic Growth Factors and Breast Cancer Progression 177
`
`apparent difference between the 121- and l65-amino acid forms with regard to tamoxifen,
`since heparan-sulfate—containing proteoglycans present in the Matrigel mixture may
`protect the VEGF from inactivation due to binding to (x2—macroglobulin (129). Small
`amounts of other growth factors are present in the Matrigel mixture, and a cooperat