WO2000070944A1

WO2000070944A1 - Genetically defined animal models and genetic markers for estrogen-induced cancer

Info

Publication number: WO2000070944A1
Application number: PCT/US2000/014302
Authority: WO
Inventors: James D. Shull
Original assignee: Board Of Regents Of The University Of Nebraska
Priority date: 1999-05-24
Filing date: 2000-05-24
Publication date: 2000-11-30
Also published as: AU5589200A

Abstract

Animal models for susceptibility and resistance to estrogen-induced cancers are disclosed. One animal model is susceptible to estrogen-induced mammary cancer, but resistant to estrogen-induced pituitary tumor development. A number of other genetically defined animal models are disclosed, based on the identification and localization of several genetic loci that modify susceptibility or resistance to estrogen-induced mammary cancer or pituitary tumor formation. Three loci modifying susceptibility or resistance to estrogen-induced mammary cancer are described: Emca1, Emca2 and Emca3. Nine loci modifying susceptibility or resistance to estrogen-induced pituitary tumor development are also described: Ept 1-9. Also described are cell lines derived from the animal models, as well as uses of the animal models and cell lines derived therefrom.

Description

GENETICALLY DEFINED ANIMAL MODELS AND GENETIC MARKERS FOR ESTROGEN-INDUCED CANCER

Pursuant to 35 U.S.C. §202(c), it is acknowledged that the U.S. Government has certain rights in the invention described herein, which was made in part with funds from the National Cancer Institute, Grant Nos. CA77876 and CA68529. This application claims priority to U.S. Application No. 09/316,093, filed May 24, 1999, and U.S. Provisional Application No. [not yet assigned], filed May 12, 2000, the entireties of which are each incorporated by reference herein.

FIELD OF THE INVENTION This invention relates to the field of hormone-associated cancer development. In particular, the invention relates to the identification and characterization of the genetic basis for susceptibility to estrogen-induced mammary cancer and estrogen-induced pituitary tumors and pituitary' tumor-associated hypeφrolactinemia. The invention further relates to the development of a genetically defined animal model for estrogen-induced mammary cancer and estrogen-induced pituitary tumors.

BACKGROUND OF THE INVENTION

Several publications are referenced in this application in order to more fully describe the state of the art to which this invention pertains. The disclosure of each of these publications is incorporated by reference herein.

Estrogens play a central role in the regulation of cell proliferation in mammalian tissues and are inextricably implicated in the etiology of several major human cancers, including those of the breast and uterus. Numerous epidemiologic studies indicate the importance of estrogens in the etiology of human breast cancer. Both early onset of menarche and late onset of menopause are associated with increased risk of developing breast cancer. It is generally believed that this increased risk results from a prolonged exposure of the mammary tissues to estrogens produced by the ovaries throughout the extended period of reproductive viability. Supporting this assertion are data indicating that bilateral ovariectomy prior to menopause markedly reduces breast cancer risk. The antiestrogen, tamoxifen, is widely used in the treatment of breast cancer, and is presently being evaluated clinically as a preventive agent in women at high risk for developing the disease. Nevertheless, the molecular mechanisms through which estrogens contribute to development of breast cancer are not presently defined. It is often stated that only 10-15% of all breast cancers in humans arise due to an inherited predisposition. This statement is often misinteφreted by both the scientific and lay communities as indicating that "sporadic" or nonfamilial breast cancers lack a genetic component. It is highly probable that many genes influence susceptibility of breast cancer, but only a few do so with the high degree of phenotypic penetrance required to reveal themselves in epidemiologic studies. At the present time, two such genes, BRCA1 and BRCA2, have been identified and cloned. In addition, the tumor suppressor gene, p53, has been shown to be involved in the etiology of breast cancer.

The mammary tissue of the human and rat are similar in moφhology, and mammary gland development in these species appears, for the most part, to be regulated by the same hormonal signals (Russo et al., Lab Invest 62:244-278, 1990; Lippman et al., J Steroid Biochem 34:107-121, 1989). Consequently, the study of mammary cancer in rats has yielded information relevant to breast cancer in humans. Treatment of rats with genotoxic agents such as 7J2-dimethylbenz(α) anthracene (DMBA), N-methylnitrosourea (MΝU), or ionizing radiation results in development of mammary cancers. Ovariectomy dramatically inhibits development of spontaneous, carcinogen-induced and radiation-induced mammary cancers in the rat, and treatment of the ovariectomized animals with estrogen restores tumor development to approximately that observed in ovary-intact animals (Russo et al., 1990, supra). Like early breast cancers in humans, spontaneously arising, carcinogen- induced and radiation-induced mammary cancers in rats are estrogen dependent; ovariectomy leads to tumor regression, whereas subsequent administration of estrogen leads to tumor reappearance.

The susceptibility of various rat strains to spontaneous, carcinogen- induced, radiation-induced and/or estrogen-induced mammary cancers has been investigated. The ACI rat strain (also known as AxC, Irish) is an inbred line derived from a cross between the August (AUG) and Copenhagen (COP) strains. The ACI strain is unique from most other strains in that rats of this strain develop mammary adenocarcinoma when treated with estrogens in the absence of exposure to chemical carcinogens or ionizing radiation. The propensity of the ACI rat to develop mammary cancers in response to estrogen treatment has been confirmed in numerous studies. In earlier studies either the synthetic estrogen, diethylstilbestrol (DES), or the semi- synthetic estrogen, 17 -ethynylestradiol (EE2), was used as the inducing hormone. Shull et al. (Carcinogenesis 18:1595-1601, 1997) demonstrated that the naturally occurring estrogen, 17β-estradiol (E2), administered chronically at near physiologic levels, rapidly induces development of mammary cancers in female ACI rats.

Whereas ACI rats are unique in their propensity to develop mammary cancers when chronically treated with estrogens, they rarely develop spontaneous mammary tumors, and the tumors that do arise spontaneously are fibroadenomas, not adenocarcinomas. Similarly, ACI rats develop relatively few mammary cancers when treated with chemical carcinogens, such as DMBA or MNU, or ionizing radiation.

Male ACI rats also develop mammary adenocarcinoma when chronically treated with estrogens. However, relative to the female of this strain, estrogen treated male ACI rats develop these cancers at a lower incidence and display a longer latency. The COP rat strain is resistant to development of mammary adenocarcinomas when treated chronically the DES, a phenotype diametrically opposed to that of the ACI strain. The COP rat strain, like the ACI strain, is also uniquely resistant to development of spontaneous mammary cancers. The inbred ACI strain discussed above was generated from a cross between the COP and AUG strain; therefore, approximately one-half of the ACI genome is derived from the COP strain. Estrogens also contribute to development of prolactin-producing pituitary tumors in humans, and these benign tumors represent a significant health problem. In spite of many years of intense study by numerous research groups, the molecular mechanisms through which estrogens regulate cell proliferation in the pituitary gland are only poorly understood.

It is known that estrogens promote development of PRL -producing pituitary tumors in several different inbred rat strains. For instance, Fisher F344 rats, both male and female, rapidly develop PRL-producing pituitary tumors when chronically treated with estrogens such as E2 or DES. In this strain, pituitary tumors arise in virtually 100% of the treated animals following only 8-10 weeks of DES treatment. Moφhologically, these tumors are grossly enlarged pituitary masses that can exceed 20 times normal weight and often exhibit a hemorrhagic appearance. Histologically, these benign tumors display diffuse lactotroph hypeφlasia and hypertrophy while lacking distinct adenomatous foci. These tumors are initially estrogen dependent and regress if estrogen treatment is discontinued; however, they can become estrogen independent following prolonged estrogen treatment. Although the animals harboring these tumors die from tumor associated pathologies before the tumors progress to carcinoma, transplantation of the tumors into an estrogen treated host will cause malignant transformation to occur.

Estrogens, both naturally occurring and synthetic, induce PRL- producing pituitary tumors in ACI rats, both male and female. As is observed in F344 rats, these benign tumors in ACI rats display diffuse lactotroph hypeφlasia and hypertrophy while lacking distinct adenomatous foci. Dietary energy restriction does not inhibit estrogen induced pituitary tumor development in the ACI rat, in contrast to the observations in F344 rats.

As discussed above, the COP strain is noted for its unique and extreme resistance to development of mammary cancers, spontaneously arising, chemically induced and radiation induced. However, this strain heretofore has not been characterized for its susceptibility or resistance to estrogen-induced pituitary tumors.

The growth response of the anterior pituitary gland of the Brown Norway (BN) rat strain to estrogen has been characterized. The strain is highly resistant; it shows only a small (statistically significant, however) increase in pituitary mass in either male or female rats following 8 to 10 weeks of treatment with DES. Holtzman is an outbred line of rats derived from the Sprague-Dawley strain. Although resistant to the pituitary growth promoting effects of estrogens over the short and intermediate terms, pituitary tumors will develop in this line following many months of treatment with DES. As discussed above, rats of the F344 strain rapidly develop PRL- producing pituitary tumors in response to chronic estrogen treatment, whereas these tumors develop at a much slower rate in animals of the Holtzman strain. (Wiklund, J. et al., Endocrinology, 109:1700-1707, 1981). Genetic studies of FI, F2 and Be progeny from reciprocal crosses between the inbred F344 and outbred Holtzman strains led to the proposal of a model in which three genes, all of equal importance, confer susceptibility to estrogen induced pituitary tumor development and homozygosity of F344 alleles at any of the three loci would result in pituitary tumor development (Wicklund et al, 1981, supra).

Wendell et al. (Proc. Natl. Acad. Sci. USA, 93:8112-8116, 1996; Mamm. Genome, 1997) have investigated the genetic bases of estrogen induced pituitary tumorigenesis in an F344 x BN intercross. Ten weeks of treatment with DES increased pituitary weight 9.7- and 3.1 -fold in female F344 and FI progeny, respectively, while having no growth promoting effect in female BN rats. Genotypic analysis of 227 phenotypically defined F2 progeny for segregation of polymoφhic genetic markers distributed across the rat genome indicated the existence of a minimum of 5 quantitative trait loci (QTL) that modulate the pituitary growth response to DES. Two of these QTL were localized to rat chromosome 2; 1 to chromosome 3; 1 to chromosome 5; and 1 QTL was localized to chromosome 9. Each of these QTL appeared to account for a portion of the pituitary growth response to DES. From these data it is apparent that the pituitary growth response of the F344 rat to synthetic estrogen is a complex heritable trait conferred by multiple genes, most of which contribute to pituitary growth in an additive manner. However, it is as yet unclear whether such a result would be obtained upon exposing F344 rats to natural estrogens, such as 17β-estradiol. The foregoing is merely a summary of the numerous studies conducted to date in various rat strains to determine the genetic basis for estrogen-induced mammary cancer and pituitary tumors. It is clear from this summary that the genetic basis for each condition is complex, involving several genes that may act additively to produce the observed phenotype in rats. As a result, the elucidation of a defined genetic basis for these two forms of cancer has not been achieved, due at least in part to shortcomings in currently available animal models for the disease. One shortcoming is that the characterization of susceptibility or resistance in these models traditionally has been done using synthetic estrogens, wherein the resulting phenotypes can differ markedly from those observed when natural estrogens are used. Another shortcoming involves the confounding effect of estrogen-induced prolactin secretion on the development of mammary carcinomas, wherein no animal model currently exists in which the animal is susceptible to mammary cancer upon exposure to natural estrogen, but is resistant to pituitary tumor development.

Accordingly, a need exists for an animal model of estrogen-induced mammary cancer and/or pituitary tumor development in which the underlying genetic basis for each condition is clearly defined, and may be manipulated in a rational manner. Moreover, the genetically defined animal model for mammary cancer should comprise animals that are highly resistant to estrogen-induced pituitary tumor formation and concomitant prolactin secretion. Such animal models would enable the isolation and characterization of the genetic differences associated with resistant and susceptible phenotypes, allowing for the independent study of the etiology of these estrogen-induced neoplasms. Such models would further enable the identification, isolation and characterization of the genes responsible for resistance or susceptibility to estrogen-induced tumor formation.

SUMMARY OF THE INVENTION

The present invention fills a need in the art by providing a variety of well-defined animal model systems for estrogen-induced cancers. These animal models, many of which are defined to precise genetic loci that confer a defined phenotype, will be used to great advantage to dissect the genetic undeφinnings of estrogen-induced cancer, and to identify useful therapeutic agents for the treatment of estrogen-induced cancers. According to one aspect of the invention, an animal model displaying susceptibility to estrogen-induced mammary cancer and resistance to estrogen- induced pituitary tumor development is provided. The model comprises a rat strain which, when administered a physiological amount of estrogen, develops a phenotype of one or more estrogen-induced mammary cancers, but does not develop a phenotype of estrogen-induced pituitary tumor development.

For puφoses of the present invention, a "physiological amount" is intended to be correlated with an amount of estrogen observed during the course of a normal estrous cycle or during pregnancy. In rats, this amount is approximately 250 pg/ml or less. The amount can vary depending on the species, but such amounts have been measured and are well known. The type of estrogen preferred for administration to the rat strain isl 7β-estradiol. However, other estrogens, such as estrone, estriol or metabolic conjugates of estradiol, estrone or estriol, may be substituted.

Preferably, the phenotype of one or more estrogen-induced mammary cancers is selected from the group consisting of intraductal carcinomas, papillary carcinomas, adenocarcinomas, cribiform carcinomas and invasive carcinomas, and the phenotype of estrogen-induced pituitary tumor development is selected from the group consisting of pituitary weight gain and increased circulating prolactin.

The rat strain used for the model typically is obtained by conventional selective breeding and phenotypic analysis. In a preferred embodiment, the rat strain is a progeny of an ACI rat strain; particularly preferred are crosses between the ACI strain and a COP or BN strain. It is preferred the rat strain's susceptibility to estrogen-induced mammary cancer, and/or its resistance to pituitary tumor development, is genetically defined, as described in greater detail below. According to another aspect of the invention, an animal model for estrogen-induced mammary cancers, which is based on a rat strain having the features of the ACI rat strain, which are: (1) a propensity to develop mammary cancers when chronically treated with estrogens, (2) but rarely developing spontaneous mammary tumors, and the tumors that do arise spontaneously being fibroadenomas, not adenocarcinomas, and (3) rarely developing mammary cancers when treated with chemical carcinogens such as DMBA or MNU, or ionizing radiation. This model comprising an ACI or similar rat strain which has been genetically modified at one or more loci selected from the group consisting oϊ Emcal, Emcal and EmcaS, such that the strain is either heterozygous or homozygous for a resistance-conferring allele at the one or more loci. Preferably, the one or more resistance-conferring alleles is obtained from a BN or COP rat strain.

According to another aspect of the invention, an animal model for estrogen-induced mammary cancers is provided, based on a strain such as the COP or BN rat that is resistant to mammary cancer when chronically exposed to estrogen. This model comprises a BN, COP, or similar rat strain which has been genetically modified at one or more loci selected from the group consisting oϊ Emcal, Emcal and EmcaS, such that the strain is either heterozygous or homozygous for a susceptibility- conferring allele at the one or more loci. Preferably, the one or more susceptibility- conferring alleles is obtained from an ACI rat strain.

According to other aspects of the invention, various animal models for estrogen-induced pituitary tumor development are provided. One model comprises an ACI or similar rat strain which has been genetically modified at one or more loci selected from the group consisting oϊEptl, Eptl, Ept5, Ept 7, Ept8 and Ept9, such that the strain is either heterozygous or homozygous for a resistance-conferring allele, preferably obtained from a BN or COP strain, at the one or more loci. Another animal model for estrogen-induced pituitary tumor development comprises a COP, BN or similar rat strain which has been genetically modified at one or more loci selected from the group consisting oϊEptl, Eptl, Ept 5, Ept 7, Ept8 and Ept9, such that the strain is either heterozygous or homozygous for a susceptibility-conferring allele (preferably obtained from an ACI strain) at the one or more loci. Yet another animal model for estrogen-induced pituitary tumor development, comprising a BN, COP or similar rat strain which has been genetically modified at one or more loci selected from the group consisting oϊEpt3, Ept4 and Ept6, such that the strain is either heterozygous or homozygous for a resistance-conferring allele (preferably from ACI) at the one or more loci. Still another animal model for estrogen-induced pituitary tumor development comprises an ACI or similar rat strain which has been genetically modified at one or more loci selected from the group consisting of Ept 3, Ept4 and Eptό, such that the strain is either heterozygous or homozygous for a susceptibility- conferring allele (preferably from a BN or COP strain) at the one or more loci.

According to another aspect of the present invention, sets of congenic strains comprising one or more of the aforementioned defined complements of genetic loci are provided. The sets of congenic strains may contain single or multiple genetic modifications at the various Emca and Ept loci.

According to another aspect of the invention, a genetic locus that modifies susceptibility or resistance to estrogen-induced mammary cancer in rats, which is linked to a physically identifiable chromosomal marker, is provided. Preferably, the locus is Emcal, Emcal or Emca3, as described herein. Also provided is a genetic locus that modifies susceptibility or resistance to estrogen-induced pituitary tumor development, also linked to a physically identified chromosomal marker, is provided. Preferably, this locus is Eptl, Eptl, Ept3, Ept4, Ept 5, Ept6, Ept7, Ept8 oτ Ept9. A "physically identifiable chromosomal marker" refers to a physical location on a chromosome that has been, or can be, distinguished in some way from other locations on the chromosome. Such markers may comprise defined genes, but preferably they comprise sequence polymoφhisms, such as simple sequence repeats (SSRs), that are polymoφhic between two alleles of a genetic locus. The polymoφhisms can be physically identified by isolating and sequencing them, according to standard methods. Preferably, PCR amplification is used, employing defined primers that flank the polymoφhic region.

According to still another aspect of the invention, cell lines derived from the animal models of the invention are provided. Preferably, the cell lines are derived from mammary tissue or pituitary tissue.

Other features and advantages of the present invention will be understood from the description and examples that follow. DETAILED DESCRIPTION OF THE INVENTION

In accordance with the present invention, several inbred rat strains have been characterized according to their genetic predisposition for estrogen induced tumors, resulting in the identification of novel genetic loci associated with estrogen induced cancers. Genetic markers have been characterized with alleles that confer susceptibility to mammary cancers or pituitary tumor development. This includes the identification of three novel loci, Emcal, Emcal and Emca3, each of which has an allele that confers susceptibility to estrogen-induced mammary cancer in rats. This breast cancer susceptibility is novel and distinct from Brcaλ , Brcal and p53, and represents a previously unknown genetic determinant for the study and etiology of breast cancer. One non limiting identifiable source of the Emca susceptible alleles is the inbred rat strain ACI. Also identified from various rat strains are nine novel loci, Eptl - Ept9, which have distinct alleles that confer resistance or susceptibility of rats to estrogen-induced pituitary tumors. The susceptible alleles of these loci exist in one ore more of the ACI, BN and COP strains.

The sections immediately below summarize the research results that support the present invention (these results are presented in greater detail in the Examples). Following that, aspects and preferred embodiments of the invention are described in detail.

The discoveries made in accordance with the present invention build upon the inventor's previous demonstration that chronic treatment with the naturally occurring estrogen, 17β-estradiol (E2), induces development of mammary cancers in female ACI rats at an incidence approaching 100% and a median latency of approximately 20 weeks (Shull et al., Carcinogenesis, 18:1595-1601, 1997) . By contrast, it has been learned in connection with this invention that female Copenhagen (COP) and Brown Norway (BN) rats are relatively resistant to E2-induced mammary cancers. When examined following 12 weeks of E2 treatment, both ACI and COP rats exhibited lobuloalveolar hypeφlasia; however, the response in the ACI strain was more robust than in COP, and focal regions of atypical epithelial hypeφlasia were observed in treated ACI, but not COP, rats. Susceptibility to E2-induced mammary cancers in ACI x COP and ACI x BN intercrosses was examined in detail, using the following protocol.

1. ACI rats were mated to either the Copenhagen (COP) or Brown Norway (BN) rats to generate FI progeny. 2. FI siblings were mated to generate F2 progeny. FI males were also mated to ACI females to generate backcross (BC) progeny.

3. Female F2 and BC rats were treated with 17-β estradiol (E2) to induce mammary cancers. Several phenotypes were assessed in each F2 and BC rat including: 1 ) latency to appearance of first mammary cancer; 2) number of tumors present following 28 weeks of E2 treatment; and 3) number of tumors present at sacrifice.

4. Using polymerase chain reaction, the genotype of each F2 and BC rat was defined at approximately 100 polymoφhic markers (nucleotide sequence differences between ACI and COP or BN strains) distributed across the rat genome. At each marker, genotype was defined as ACI/ ACI, ACI/COP (or ACI/BN in cross to BN strain) and COP/COP (BN/BN in cross to BN strain).

5. Mapmaker QTL (see Example 4) was used to determine the correlation between phenotype and genotype. Highly significant correlations were revealed on chromosmes 5, 18 and 2. The phenotypes of the ACI/COP and ACI/BN F 1 progeny indicated incompletely dominant inheritance of susceptibility. Mapping analysis of F2 progeny revealed on chromosome 5 a locus, Emcal, that conferred susceptibility in both the ACI x COP and ACI x BN crosses. Two additional loci, Emcal and Emca3, modifying susceptibility to E2-induced mammary cancers were identified on chromosomes 18 and 2, respectively. Emcal and Emca3 were revealed in the ACI x BN intercross, but not in the ACI x COP intercross. The majority of mammary cancers induced in ACI rats by treatment with E2 exhibited aneuploidy. Analysis of polymoφhic microsatellites indicated that loss of heterozygosity (LOH) was a common event in mammary cancers induced in female ACI x COP FI rats. Table 1 below summarizes the Emca loci and their actions as modifiers of susceptibility to E2-induced mammary cancers. Table 1.

Locus Marker LOD Phenotype for Each Genotype' Latency Tumor Pos. at 28 wks.

Emcal 5Rat37 4.4² /A: 1 84 days A/A: 0.84 A/C: 225 days A/C: 0.67 C/C: 265 davs C/C: 048

5 at95 3J- A' A: 190 days A A: 0.62 A/B: 267 days A/B: 0.52 B/B: 362 days B/B: 0.27

Emcal 18Rat30 4.0³ AA: 218 days A A: 0.71 A/B: 221 days A/B: 0.65 B/B: 367 davs B/B: 0.17

Emca3 2Ratl 6 3.0³ AA: 1 87 days A/A: 0.71 A/B: 242 days A/B: 0.60 B/B: 342 days B/B: 0.32

Footnotes: 1. Latency represents the length of E2 treatment preceding the appearance of the first palpable mammary carcinoma. Tumor positive at 28 weeks represents the fraction of the E2 treated F2 population exhibiting palpable mammary tumors by this time point. A, ACI allele; C, COP allele; B, BN allele. 2. Data presented are from analysis of E2-induced mammary cancer development in a total of 224 female F2 progeny, 124 from an ACI female x COP male cross and 120 from a COP female x ACI male cross. LOD score was calculated using MapMaker QTL software and tumor status at 28 weeks as the analyzed phenotype. 3. Data presented are from analysis of E2-induced mammary cancer development in a total of 84 female F2 progeny from an ACI female x BN male cross. Additional ACI/BN F2 progeny are currently being evaluated.

It is important to note that in above-described genetic mapping analyses, particular attention was paid to regions of the rat genome that contain homologs to known human breast cancer susceptibility genes as well as to regions demonstrated by others to contain modifiers of susceptibility to mammary cancers induced by the chemical 5 carcinogen dimethylbenz[α] anthracene (DMBA). The data indicate that rat chromosome 10, on which Brcal and p53 reside, and rat chromosome 12, on which Brcal resides, do not harbor detectable modifiers of susceptibility to E2-induced mammary cancers. Similarly, the region of rat chromosome 2 demonstrated by Gould and colleagues (Hus et al.. Cancer Research 54:2765-2770, 1994) to harbor Mcsl, a 10 modifier of DMBA-induced mammary cancers, lies far proximal to the Emca3 locus identified in accordance with this invention. Finally, the Mcsl, Mcs3 and Mcs4 modifiers of DMBA-induced mammary carcinogenesis, mapped by Gould and colleagues (Shepel et al.. Genetics 149:289-299, 1998) to rat chromosomes 7, 1 and 8, respectively, do not appear to modify susceptibility to E2-induced mammary cancers. These data indicate that E2-induced mammary cancers develop through molecular mechanisms that are distinct from those responsible for development of DMBA- induced mammary cancers. In parallel experiments performed in connection with the present invention, several loci that confer pituitary tumor development in response to estrogen treatment have been mapped, using a similar protocol to that described above.

3. Male F2 and BC rats were treated with the synthetic estrogen, diethylstilbestrol (DES) to induce pituitary growth. Following 12 weeks of DES treatment, pituitary growth was quantified by measuring pituitary weight, which correlates with pituitary cell number.

4. Using polymerase chain reaction, the genotype of each F2 and BC rat was defined at approximately 100 polymoφhic markers distributed across the rat genome. At each marker, genotype was defined as ACI/ACI, ACI/COP (or ACI/BN in cross to BN strain) and COP/COP (BN/BN in cross to BN strain). 5. Mapmaker QTL was used to determine the correlation between phenotype and genotype.

Nine loci have been identified that modify the pituitary growth response to estrogen. Eptl, Eptl and Eptό were identified in the ACI x COP cross.

The remaining Ept loci were identified in the ACI x BN cross. The loci that confer or modify development of estrogen-induced pituitary tumors appear distinct from those conferring susceptibility to E2-induced mammary cancers, indicating that these two estrogen-induced tumorigenic processes are genetically separable. Table 2 summarizes the Ept loci and their actions as modifiers of susceptibility to DES- induced pituitary tumor development. Table 2.

Locus Cross Chromosome Linked Growth Marker Promoting

Allele

Eptl ACI x COP 6 6Rat80 ACI

Eptl ACI x COP 3 3Rat26 ACI

Ept3 ACI x BN 6 6Rat22 BN

Ept4 ACI x BN 12 12Rat53 BN

Ept5 ACI x BN 18 18Rat57 ACI

Eptό ACI x COP 3 3Mgh9 COP

Ept 7 ACI x BN 2 2Ratl9 ACI

Ept8 ACI x BN 2 2Ratl47 ACI

Ept9 ACI x BN 10 10Mit7 ACI

Together, the experimental results set forth above have provided the foundation for creating the genetically defined, physiologically relevant, animal models for elucidating the etiology of estrogen-induced breast cancer and pituitary tumor formation, as described below.

ANIMAL MODELS

A variety of animal models are provided in accordance with the present invention. These are described and exemplified below.

I. Animal models for estrogen-induced mammary cancer

According to a basic aspect of the invention, an animal model with susceptibility to estrogen-induced mammary cancer and resistance to estrogen- induced pituitary tumors is provided. The animal model is a rat strain which, when administered a physiological amount of 17 β-estradiol or other estrogen, develops a phenotype of one or more estrogen-induced mammary cancers including intraductal carcinomas, papillary carcinomas, adenocarcinomas and invasive carcinomas, but does not develop a phenotype of estrogen-induced pituitary weight gain. Furthermore, the genetic basis for the susceptibility to estrogen-induced mammary cancer in this strain is distinct from that induced by dimethylbenz- -anthracene, or that associated with Brcal, Brcal, p53 or the Mcs loci (1 - 4). Selected progeny of crosses between the ACI strain and the COP or

BN strain provide exemplary embodiments of this basic animal model. Progeny are selected that possess one or more of the Emcal, 1 or 3 alleles from the ACI parent that confer susceptibility to estrogen-induced mammary cancers, but are lacking one or more of the growth promoting alleles oϊ Ept 1-9 from the ACI parent or other sources (for Eptl, Eptl, Ept 5 and Ept 7-9, the growth non-promoting alleles are found in the COP or BN strains; for Ept3, Ept4 and Ept6, the growth non-promoting alleles are present in the ACI strain).

Though selected ACI x COP or BN progeny are preferred for use in the invention, they are not the only strains that can be used. Now that the genetic basis for the ACI, BN and COP phenotypes has been defined, other rat strains having those same genetic features may be identified or generated through conventional breeding techniques combined with molecular genetic analysis, and these may serve equally well as an animal model having the basic features described above.

It has been demonstrated in accordance with the present invention that, even though ACI rats exhibit estrogen induced pituitary weight gain, the hypeφrolactinemia associated therewith is not a significant influence on mammary tumor development. Indeed, it has been clearly demonstrated, using ovary-intact versus ovariectomized female ACI rats, that the ovary modulates estrogen-mediated mammary carcinogenesis in this rat strain (see Example 1 and Shull et al., 1997, supra). For this reason, the ACI strain, or other strains with equivalent genotypes, may be used as defined models for estrogen-induced mammary cancer, even though they may possess one or more pituitary tumor growth promoting alleles of Ept 1-9.

Accordingly, an important aspect of the present invention is a genetically defined animal model system for estrogen-induced mammary cancer, in which the three loci, Emcal, Emcal and Emca3 may be studied separately or in various combinations of susceptibility and resistance alleles. This animal model system comprises a set of congenic rat strains, wherein the background genotype remains constant, and modifications are made specifically to the Emcal, Emcal and/or Em ca 3 loci.

An exemplary set of congenic rat strains comprises an ACI genetic background, into which is bred to homozygosity the resistance alleles oϊ Emcal, Emcal and/or Emca3 from COP or BN as follows:

1. Emcal from COP or BN

2. Emcal from BN

3. Emca3 from BN 4. Emca 1 from COP or BN and Emcal from BN

5. Emcal from COP or BN and Emca3 from BN

6. Emcal and Emca3 from BN

7. Emcal from COP or BN and Emcal and Emca3 from BN

Another exemplary set of congenic rat strains comprises a BN or COP genetic background, into which is bred to homozygosity the susceptibility alleles of Emcal, Emcal and/or Emca3 from ACI as follows:

1. Emcal

2. Emcal

3. Emca3 4. Emca 1 and Emcal

5. Emcal and Emca3

6. Emcal and Emca3

7. Emcal and Emcal and Emca3

Persons of skill in the art will be able to devise other useful congenic strains for estrogen-induced mammary cancer, based on the principles used to devise the two exemplary congenic sets of strains described above. It will also be appreciated by one of skill in the art that each member of a set of congenic strains may be used alone or in combination with one or more other members of that set, or with one or more members of a different set of congenic strains. II. Animal models for estrogen-induced pituitary tumors

A variety of useful animal models can be generated, based on the discovery of the nine Ept loci that contribute to the susceptibility or resistance to pituitary tumor development in rats. The ACI, COP and BN strains are of particular utility in this regard. However, inasmuch as the loci have been identified and mapped with particularity, any rat strain may prove appropriate in the generation of a genetically defined model for pituitary tumor development. Males or ovariectomized females are preferred for use in this aspect of the invention, so as to avoid complications arising from mammary cancers that develop in response to estrogen treatment.

Because the Ept loci appear to contribute additively to the pituitary tumor phenotype, sets of congenic strains in which specific alleles oϊ Ept 1-9 are isolated or combined are expected to be of particular utility. Some examples of useful sets of congenic strains include the following: A. An ACI genetic background, into which is bred to homozygosity the growth non-promoting alleles oϊEptl, Eptl and/or Ept 5 from COP (or BN) as follows:

1. Eptl

2. Eptl 3. Ept5

4. Eptl and Ept2

5. Eptl and Ept 5

6. Eptl and Ept 5

1. Ept land Eptl and Ept 5

B. An ACI genetic background into which is bred to homozygosity the growth non-promoting alleles of Ept 7, Ept8 and Ept9 from BN in all combinations as described for the previous example.

C. A large congenic set comprising an ACI genetic background into which is bred to homozygosity the non-growth promoting alleles oϊEptl, Eptl, Ept5,

Ept 7, Ept 8 and or Ept9 from BN or COP, in all combinations as described for the previous examples.

D. A BN genetic background into which is bred the growth non- promoting alleles oϊEpt3 and 'or Ept4 in all combinations.

E. Various sets of congenic strains that start with a resistant genetic background (e.g., BN for Epts 1, 2, 5, 7, 8, 9 or ACI for Epts 3 and 4), into which is bred the respective growth promoting alleles from, e.g., ACI (or BN in the case oϊ Epts 3 and 4), in all combinations.

Again, persons of skill in the art will be able to devise other useful congenic strains for estrogen-induced pituitary tumor development, based on the principles used to devise the several exemplary congenic sets of strains described above. It will also be appreciated by one of skill in the art that each member of a set of congenic strains may be used alone or in combination with one or more other members of that set, or with one or more members of a different set of congenic strains, including the Emca congenic strains described above. The identification of specific genetic markers associated with each of the Emca or Ept alleles renders the production of the aforementioned congenic strains straightforward. The following basic steps are used, wherein "A" is used to define the background genotype and "B" is used to define the specific allele to be bred into the background genotype. 1. Mate a B female with an A male to generate Nl progeny (also referred to as FI); this fixes the Y chromosome in the Nl progeny and all subsequent generations as arising from the A strain.

2. Backcross males of the Nl strain with females of the A strain to create N2 progeny. 3. Genotype the N2 progeny across the entire genome to select progeny heterozygous for the B allele at a desired locus, and homozygous for A alleles a disproportionately large fraction of markers at other loci distributed throughout the rest of the genome. This latter step effectively decreases the number of backcross generations required to generate the desired congenic rat strains. 4. Selected N2 males are mated to A females to generate N3 progeny, and the step 3 genotyping step is repeated. 5. Selected N3 males are mated to A females, and the step 3 genotyping is repeated. By the N5 generation, B alleles should be eliminated at all loci except the desired loci. If this is not the case, then further backcrosses are made, to generate N6, N7, to NX progeny. 6. Male and female NX siblings are mated. The resulting male and female progeny are genotyped to identify those that are homozygous for B at the locus of interest.

7. Animals that are homozygous for the locus of interest are mated brother to sister to maintain the desired congenic rat strain. 8. Each congenic strain may be mated to other congenic strains to generate strains carrying the desired alleles for multiple Emca and Ept loci. For example, a congenic strain homozygous for BN alleles oϊ Emcal on an ACI background are mated to a strain homozygous for BN alleles oϊ Emcal on an ACI background to generate progeny that are heterozygous for Emcal and Emcal. Males are mated to siblings and the progeny are genotyped to identify animals that are homozygous for BN alleles at both Emcal and Emcal (expected frequency about 6.25%o) on an ACI background.

As an example of a preferred embodiment, a congenic strain comprising a BN allele oϊ Emcal on an ACI genetic background is made by the following steps:

1. A BN female is mated to ACI male to generate Nl progeny.

2. A male Nl (from step 1) is mated to ACI females to generate N2. The resulting male progeny are genotyped across genome to select animals heterozygous for BN alleles at Emcal and homozygous for ACI alleles at disproportionately high fraction of markers at other loci distributed across rat genome.

3. Selected N2 males are mated to ACI females to generate N3 progeny. Genotypic analyses are performed on male progeny to select animals heterozygous for BN alleles at Emcal and homozygous for ACI alleles at disproportionately high fraction of markers at other loci distributed across rat genome. 4. Selected N3 males are mated to ACI females to generate N4 progeny. Genotypic analyses are performed on male progeny as described in step 2. 5. Selected N4 males are mated to ACI females to generate N5 progeny. Genotypic analyses are performed on male and female progeny as described in step 2. By the N5 generation, BN alleles should be eliminated at all genetic loci other than Emcal. If this is not the case, then selected N5 males are mated to ACI females to generate N6 progeny for genotypic analysis; this step is repeated until male and female NX progeny of desired genotype are obtained.

6. Male and female N5 siblings are mated. The resulting male and female progeny are genotyped to identify those that are homozygous for BN alleles at Emcal . 7. Animals that are homozygous for the Emcal locus are mated brother to sister to maintain the desired congenic rat strain.

GENETIC MARKERS

Tables 1 and 2 above set forth specific markers, which are polymoφhisms or other markers associated with each of the Emca or Ept loci. As determined by computer analysis, each marker is linked with its respective Emca or Ept locus with an LOD (logarithm of the likelihood ratio for linkage) greater than 3.0 which is generally considered to be a statistically significant indicator of linkage.

It will be understood by one of skill in the art that a physical interval containing the gene of interest may be estimated from the LOD score. The physical interval is generally defined by a 1 or 2 LOD score decrease from the peak marker. Thus, to define the precise locations oϊ Emca and Ept loci, additional polymoφhic markers in the regions of interest are identified and the genotypes of phenotypically defined F₂ and BC animals at each additional locus are determined. The nearest polymoφhic marker at which the LOD score is 2 or more units less than that observed at a peak marker (highest LOD score) is considered to define the boundaries of the interval containing the gene.

In accordance with these principles, then, each marker associated with a selected Emca or Ept allele, along with additional markers defined as above, may be used to track that allele in breeding programs and ultimately to identify the gene's physical location so that it may be cloned and characterized. These genetic markers therefore comprise an important aspect of the present invention.

Moreover, the synteny that exists between the rat genome and genomes of other species, humans being a significant example, facilitates the identification of corresponding regions on the other species' genome, thereby providing a tool for identifying corresponding susceptibility and resistance alleles in the other species, and ultimately locating the corresponding genes. In this regard, the table below sets forth synteny between the rat and human genomes corresponding to the markers that have been linked to the Emca loci.

Table 3.

EEmmccaa llooccuuss RRaatt iinntteerrvvaall Syntenous human interval

Emcal 5rat37 9p21 and lp31-33

Emcal 18rat30 18ql 1-12 and 5q21-32

Emca3 2ratl6 5ql3.3

In view of the evidence set forth herein from the rat model, these Emca loci are expected to act as modifiers of breast cancer risk in humans, perhaps in a manner similar to the BRCA genes, mutations in which are highly correlated with increased risk of breast cancer in women. Accordingly, identification the human homologs of the Emca loci will provide a significant advance in the art of diagnosis, treatment and prevention of breast cancer in women. This identification will be greatly facilitated by the genetically defined animal model systems of the present invention.

ASSAYS AND CELL LINES

The genetically defined animal models of the invention will be of great utility in confirming the efficacy of novel anti-estrogenic compounds or other agents, for the treatment of estrogen-induced cancers. In this regard, it should be noted that anti-estrogenic agents, such as taximofen and raloxifene, are presently being used to prevent breast cancer in humans. The genetically defined animals of the invention provide a very useful set of standards for measuring the effect of a test compound on estrogen-induced cancer development. Thus, the efficacy of a test compound could be clearly elucidated from a comparison of susceptible and resistant animals. Moreover, the impact of the animal's defined genetic composition on the efficacy of a test compound could yield important information relating not only to the test compound and its structural analogs, but also to the importance of each defined genetic allele that contributes to the development of an estrogen-induced cancer.

In this regard, another aspect of the present invention comprises cell lines obtained from any of the animal models of the invention. Any cell line from these animals may be useful, but mammary or pituitary cell lines are expected to be particularly useful tools in screening estrogenic and ant-estrogenic compounds. The inventor has established immortalized mammary epithelial cells from the ACI and COP strains. Using similar techniques, which are well known in the art, mammary and pituitary cell lines from any of the rat strains can be established. One or more of the cell lines may be used in standard cell-based screening assays for the agents of interest, according to methods well known in the art.

The following examples are provided to illustrate the invention. They are not intended to limit the invention in any way.

EXAMPLE 1

Phenotypic Characterization of E2 Response in Crosses of ACI and COP Rat Strains, and Progeny

ACI and Brown Norway (BN) rats were obtained from Harlan Sprague-Dawley (Indianapolis IN). Copenhagen (COP) rats were obtained from the National Cancer Institute Animal Breeding Program (Bethesda MD). COP males were mated to ACI females to generate F, progeny. F, siblings were mated to generate F₂ progeny. F, males were mated to ACI females to generate BC progeny. ACI, COP, F,, F₂, and BC females were treated with E2, administered from subcutaneous Silastic tubing implants, beginning at 9 weeks of age. Silastic tubing implants containing 27.5 mg of E2 were prepared and surgically inserted over the scapulae. Thereafter, each animal was examined regularly for the presence of palpable mammary tumors. A tumor-bearing animal was sacrificed when its largest tumor reached a size approximating 1.0-1.5 cm in diameter. The E2-containing implant from each animal was inspected; all implants were observed to contain crystalline E2 at the time of sacrifice. Each mammary tumor was classified histologically. Latency to appearance of the first palpable tumor was used as the primary phenotypic indicator of susceptibility to E2-induced mammary cancers. Data on mammary tumor burden (number, size, location) and pituitary weight were also collected. Trunk blood sera were collected and stored for assay of circulating E2, prolactin (PRL), and possibly other hormones, including metabolites of E2. Untreated control ACI, COP, FI, F2 and BC animals were examined regularly and sacrificed at various intervals throughout the course of the experiment. No mammary tumors were observed in these untreated females.

ACI rats display susceptibility to E2-induced mammary cancer. A palpable mammary tumor was first detected 99 days after initiation of E2 treatment. The median latency (n=21) to appearance of the first palpable tumor was 143 days, and 100%) of the treated population displayed palpable tumors within 197 days of initiation of E2 treatment. The mean latency was 145 days (standard deviation = 26 days), and latency within the E2 treated population displayed a normal distribution. Upon necropsy, all but one of the E2 treated animals were determined to harbor multiple macroscopic mammary tumors (mean = 5.6 tumors per animal, range 1-18). These tumors appeared to be randomly distributed throughout the six pairs of mammary glands. No mammary tumors were observed in untreated, ovary intact, ACI rats monitored for up to 427 days relative to the time at which treatment of the E2 treated animals was initiated. All excised tumors were histologically classified. The majority of the tumors showed features of intraductal carcinoma, comedo type. Some of the carcinomas had areas of invasion with desmoplastic reaction. Some papillary carcinomas were also observed. Mitotic figures were often frequent. The surrounding mammary tissue displayed marked ductal and lobuloalveolar hypeφlasia, often with secretory changes. Microscopic foci of intraductal carcinoma were often observed within these grossly normal mammary tissues. Mammary tissues from untreated, ovary intact, ACI rats were poorly developed, relative to those from E2 treated animals, with virtually no lobuloalveolar tissue identified. Mitotic figures were virtually absent in the mammary tissues from untreated rats.

The amount of E2 in trunk blood sera from untreated, ovary intact, ACI rats was less than 50 pg/ml, the level of detection of the RIA as performed in this study. The mean level of E2 in trunk blood sera from E2 treated, ovary intact, animals averaged 185 ± 23 pg/ml. During the rat estrous cycle, circulating E2 levels oscillate between approximately 20 and 100 pg/ml, with peak levels occurring at midday on proestrus. These data demonstrate that continuous treatment with near physiologic levels of the naturally occurring estrogen, E2, induces rapid development of multiple and often invasive mammary cancers in the female ACI rat.

Copenhagen rats display resistance to E2-induced mammary cancers. A population of E2 treated female COP rats was examined for susceptibility to mammary cancer development. This population consisted of only 7 animals because published reports indicated the resistance of the COP strain to DES-induced, spontaneous and carcinogen-induced mammary cancers. One animal from this group was killed following 142 days of E2 treatment, a time at which 50% of the E2 treated ACI rats displayed palpable mammary tumors, and the mammary tissues were processed for histologic evaluation. A second animal was killed following 223 days of treatment. Both of these animals were free of mammary cancers at the time of sacrifice. The remaining 5 animals were monitored for up to 378 days following initiation of E2 treatment. A single palpable tumor was detected in one of the 5 remaining animals following 226 days of E2 treatment. Histologic evaluation indicated this tumor to be an intraductal carcinoma. Despite the appearance of a single carcinoma in this small population of E2 treated COP rats, the relative resistance of the COP strain to E2-induced mammary cancers is apparent.

FI Progeny Display a Susceptible Phenotype. A population of 34 E2 treated female FI animals was examined. In the first group of 17 FI females, E2 induced palpable mammary tumors at an incidence of 100% with a median latency of 197 days (range = 152-249 days). In the second group of 17 animals treated with E2 for 175 days, the first palpable tumor was detected following 163 days of E2 treatment. Thirteen of the 34 FI females were killed and their tumors classified as either adenocarcinomas or invasive carcinomas. These FI animals harbored, on average, 3J macroscopic cancers per animal. The data from these FI animals indicate that susceptibility to estrogen-induced mammary cancers is inherited as a dominant genetic trait. However, relative to the ACI parental strain, latency is prolonged and tumor yield appears to be reduced. FI animals not treated with E2 did not develop palpable tumors.

F2 Progeny Display Three Phenotypes. A population of 133 female F2 rats were treated with E2. The F2 progeny displayed three phenotypes: 1) highly susceptible (ACI-like); 2) susceptible (Fl-like); and 3) resistant (COP-like).

Highly susceptible. Fifteen of the 133 E2 treated F2 progeny (11.3%) developed a palpable mammary tumor within 147 days of E2 treatment, a time point preceding the first observation of a palpable tumor in the FI progeny. By comparing the ACI and F2 populations, it is possible to estimate the proportion of the F2 population that is as highly susceptible to E2-induced mammary cancers as the ACI population. Normalizing the proportion of F2 progeny (11.3%) displaying palpable tumors at the 147 day time point by the proportion of the ACI population displaying palpable tumors at this time point (13 to 21 animals, 61.9%) indicates that approximately 18%o of the F2 population is as highly susceptible to E2-induced mammary cancers as the parental ACI strain.

Susceptible. Seventy-nine of the 133 F2 females were treated with E2 for at least 231 days. Fifty one of these 79 (64.6%) F2 progeny developed palpable mammary tumors with a latency ranging from 148-224 days, and resemble the FI progeny in their susceptibility to E2-induced mammary cancers. Resistant. Twenty of the 79 F2 animals (25.3 %>) that were treated with E2 for a minimum of 231 days remained free of palpable cancers. Although several of these 20 resistant F2 animals remained tumor free following 252-308 days, others became moribund, apparently due to pituitary tumors, and were sacrificed. These data suggest that approximately 25%> of the F2 progeny are as resistant to E2- induced mammary cancers as the COP parental strain. Data Summary and Interpretation. Data summarized above indicate that approximately 75% of E2 treated F2 population develop mammary cancers, whereas approximately 25% remain tumor free. Of the 75%₀ that develop mammary tumors, approximately 11% can clearly be defined as highly susceptible (ACI phenotype), with the remaining 64%) being defined as susceptible (FI phenotype). Following normalization, it is estimated that approximately 18% of the F2 population are as highly susceptible as the ACI strain, approximately 57% are as susceptible as the FI progeny, and 25% are as resistant as the COP strain.

Backcross Progeny Display Highly Susceptible and Susceptible Phenotypes. A population of E2 treated female BC animals, derived from a cross between F, males and ACI females, was examined. Each of these BC animals developed palpable tumors with latencies ranging from 1 17-216 days. These data are consistent with a model in which Emcal acting in either a dominant or codominant manner confers susceptibility to E2-induced mammary cancers. This BC population was small by design; in gene mapping studies, BC animals provide one-half the number of informative meioses as F₂ animals. Table 4 illustrates a model in which Emcal acts as a dominant or co-dominant susceptibility gene for estrogen-induced mammary cancer. Other models will also explain the same phenotypic data, as additional Emca loci are defined (see Example 4).

Table 4. Emcal acts in a dose-dependent manner to confer susceptibility. The

ACI and COP alleles oϊ Emcal are represented as EMCA1 and emcal, respectively.

Group Genetic Cross Genotype(s) Predicted Predicted

(Male x female) (pat/mat) Phenotype(s) Frequencies (%)

ACI parental strain EMCA 1 /EMCA 1 highly susceptible 100

COP parental strain emcal /emcal resistant 100

Fl (b) COP x ACI emcal /EMCA I susceptible 100

F2(b) Fl (b) x Fl (b) emcal /emcal resistant 25 emcal ^'/EMCA 1 susceptible 25

EMCA 1 /emcal susceptible 25

EMCA 1 /EMCA 1 highly susceptible 25 EXAMPLE 2

Phenotypic Characterization of Estrogen-Induced

PituitarvTumor Development in the ACI x Copenhagen Intercross

While characterizing the ACI and COP strains for susceptibility to E2 induced mammary cancers, we recognized that these strains display quantitatively different pituitary growth responses to administered estrogens. Therefore, the male progeny from the ACI x COP intercross described in Example 1 were utilized to characterize the genetic bases underlying the differing pituitary growth responses. Male FI , F2 and backcross (BC) progeny, as well as male ACI and COP rats, were treated with DES beginning at 9 weeks of age. DES was administered from subcutaneous Silastic implants containing 5 mg of this synthetic estrogen. This same treatment protocol was used by Wiklund et al. (1981, supra) and Wendell et al. (1996, supra; 1997, supra) in their studies of the genetic bases underlying development of estrogen induced pituitary tumors in the F344 rat. After 12 weeks of treatment, the animals were killed, the pituitary glands were removed and weighed, and trunk blood sera were harvested for assay of circulating prolactin (PRL). Each animal was injected with BrdU 4 hours prior to sacrifice so that S phase fraction could be assessed in the lactotroph and nonlactotroph populations.

Anterior pituitary weight correlates with cell number and DNA content per gland, and is, therefore, widely used as a surrogate indicator of a cellular growth response to estrogen. DES administered for 12 weeks from Silastic implants induced pituitary tumor development in both the ACI and COP rat strains, as well as in their derived F,, F₂ and BC progeny. However, the ACI strain displayed an approximately 2-fold greater pituitary growth response to DES than the COP strain, and the ACI phenotype was conferred upon the derived progeny as a dominant genetic trait. Average pituitary weight in male ACI rats was increased 6.9-fold in response to 12 weeks of DES treatment; from 9.2 ± 0.2 mg (mean ± SD) in untreated rats to 63.7 ± 12.6 in DES-treated rats. In male COP rats, DES increased pituitary weight 3-fold, from 12.7 ± 0.9 to 38J ± 8.2 mg. The difference in mean pituitary weight displayed by the ACI and COP populations after DES treatment was highly significant (P = 3.3 x 10^"6). In F, progeny resulting from a cross between ACI females and COP males, DES increased pituitary weight 5.8-fold, from 10.2 ± 1.2 to 58.8 ± 1 A mg. The mean pituitary weight observed in the DES-treated F, population was indistinguishable from that in the treated ACI population (P = 0.20), but was significantly greater than that in the treated COP population (P = 8J x 10 ⁸). In the F₂ population, DES increased the mean pituitary weight 6.0-fold, from 10J ± 1.0 to 60.9 ± 23.9 mg. The large SD in the treated F₂ population reflected the presence of both the ACI and COP phenotypes within this genetically heterogeneous population of animals. Utilizing statistical methods developed by Wright (In: Anonymous evolution and the genetics of populations: genetics and biometrical foundations, pp. 373-420, University of Chicago Press, 1968), we estimated that 84% of the total phenotypic variance in the DES-treated F₂ population was genetically conferred, whereas the remaining 16%o was due to environmental factors. Pituitary weights in a population of DES-treated male BC progeny averaged 68.2 ± 12.5 mg, which is equivalent to that observed in the treated ACI population (P = 0.34) but greater than that observed in the treated COP population (P = 4.4 x 10^"9).

To equalize variances among the populations and facilitate the phenotypic characterization of individuals within the F F₂ and BC populations, the data for pituitary weight were log transformed. After transformation, the quantitative difference between the pituitary growth response of the ACI strain and that of the COP strain remained apparent (P = 3.7 x 10^~7). Moreover, the distribution of pituitary weights within the F, and BC populations closely resembled that in the ACI strain (F[ vs. ACI, P = 0.2268; BC vs. ACI, P = 0.3011), but not that in the COP strain (F, vs. COP, P = 1.0 x 10-6; BC vs. COP, P = 1.0 x 10^"8). Each individual in the F„ F₂ and BC populations was defined as being of the ACI phenotype when its pituitary weight was greater than or equal to the mean pituitary weight displayed by the DES-treated ACI population minus 1 SD (≥ 1.72 log units; 51 J mg). An individual was classified as COP in phenotype when its pituitary weight was less than 1 SD greater than that displayed by the DES-treated COP population (< 1.66 log units; 46.3 mg). Using these criteria, 80% (24 of 30), 61% (63 of 103) and 100% (19 of 19) of the DES- treated F F₂ and BC populations, respectively, were classified as ACI in phenotype; 3% (lof 30), 25% (26 of 103) and 0% of the F„ F₂ and BC populations were classified as COP in phenotype, and 17% (5 of 30), 14% (14 of 103) and 0% of the F„ F₂ and BC populations had a pituitary weight intermediate between those observed in the ACI and COP populations and consequently were not classified as being either ACI or COP in phenotype. DES increased circulating PRL to a greater extent in ACI males than in

COP males (P - 4.0 x 10^~4), probably reflecting the differing pituitary weights and absolute lactotroph numbers displayed by these strains. Circulating PRL levels in the DES-treated F, and BC populations did not differ significantly from those in the treated ACI population (P = 0.064 and 0.798, respectively), but did differ significantly from those in treated COP populations (P = 1.5 x 10^~3 and 1.0 x ION respectively).

Like pituitary weights, PRL levels in the F₂ population reflected the genetic variation within this population. The positive correlation (correlation coefficient = 0.882) between pituitary weight and circulating PRL across the different DES-treated populations indicates that circulating PRL provides an accurate surrogate indicator of estrogen-induced pituitary mass.

EXAMPLE 3

Phenotypic Characterization of Estrogen-Induced

PituitarvTumor Development in the ACI x Brown Norway Intercross

To confirm that the genetic bases underlying the growth response of the ACI pituitary gland to DES differ from those of the F344 rat, the ACI strain was crossed with the BN strain and the ability of DES to stimulate pituitary growth in the F, progeny was examined. Wendell et al. utilized a F344 x BN cross to identify QTL that modulate the pituitary growth response of the F344 rat to DES. In that study, F344/BN F, progeny displayed approximately 30%o of the pituitary growth response of the parental F344 strain to administered DES. Following 12 weeks of DES treatment, the average pituitary weight in male ACI/BN F, progeny was increased 6.1- fold; pituitary weights averaged 9.7 mg (sd = 1.5,n=7) and 59J mg (sd = 6.0,n=16) in untreated and DES treated F, progeny, respectively. The observed 6J-fold induction of pituitary weight is virtually identical to that observed in males of the ACI strain as well as in male ACI/COP F, progeny. Together, these data indicate that the ACI alleles of the genes which confer the tumorigenic response of the ACI rat pituitary gland to estrogens act in a dominant manner over the BN alleles and confirm the hypothesis that the genetic bases underlying the growth response of the ACI pituitary to DES differ from those of the F344 rat.

EXAMPLE 4

Mapping Genetic Loci that Modify Susceptibility to Estrogen-

Induced Mammary Cancers and Pituitary Tumor Development

Using polymerase chain reaction, the genotype of the F2 and BC rats of

Examples 1 -3 were defined at approximately 100 polymoφhic markers (many or all being "simple sequence repeats", referred to as "SSR", which display nucleotide sequence differences between ACI and COP or BN strains) distributed across the rat genome (referencing the Rat Genome project, at http://www.genome.wi.mit.edu rat. and utilizing a set of PCR primers that frame the polymoφhic SSRs or markers, available from Research Genetics, Inc., Huntsville AL, http://www.resgen.com/ index.php3). At each marker, the genotype was defined as ACI/ ACI, ACI/COP (or ACI/BN in crosses to the BN strain) and COP/COP (BN BN in crosses to BN strain). At weaning, tails were snipped and stored at -80°C. At sacrifice, the liver was retained and stored at -80°C. DNA was isolated from these tissues using QIAmp Spin Columns, as directed by the manufacturer (QIAGEN).

PCR was performed as described by Wilder and colleagues (Remmers et al, Immunogenetics 41 :316-319, 1995; Zha et al., Mamm Genome 5:538-541, 1994; Du et al., Cytogenet Cell Genet 68:107-111, 1995; Mathern et al., Cytogenet Cell Genet 66:283-286, 1994, for example). Briefly, 50 ng of DNA was amplified in a 10 μl reaction containing 10 mM Tris (pH 8.3 at 25°C), 1.5 mM MgCl₂, 50 mM KC1, 01.1% gelatin, 120 nM primers, 200 μM dNTP, [α-32P]dCTP (1 μCi), and Taq polymerase (0J U). The reaction mixtures were denatured at 94°C for 5 minutes followed by 35 cycles of: 1) 94°C for 1 minute; 2) 55°C for 1 minute; 3) 72°C for 1 minute. Eleven hundred seventy-nine primer pairs for genotyping mapped SSR loci were available from Research Genetics, Inc. (http://www.resgen.com) at the time the experiments were conducted. PCR conditions were optimized for the different primer pairs as necessary. The radiolabeled DNA products were separated on an 8% denaturing polyacrylamide sequencing gel and visualized by autoradiography, on a Molecular Dynamics Phosphorlmager. The genotype of each animal at each polymoφhic SSR locus was defined independently by two researchers without knowledge of the animal's phenotype. In cases where the two individuals did not concur on genotype, the PCR analysis were repeated.

Mapmaker QTL (available at no charge to the public from Dr. Eric Lander, Whitehead Institute, Massachusetts Institute of Technology) was used to determine the correlation between phenotype and genotype. Identification of loci that modify estrogen-induced mammary cancers. We have analyzed E2-induced mammary cancers in FI, F2 and backcross (BC) progeny derived from three different genetic crosses (strain of female noted first by convention): 1) ACI x COP (previous Examples); 2) COP x ACI; and 3) ACI x BN. Data from these genetic studies indicate that susceptibility to E2-induced mammary cancers is inherited as an incompletely dominant genetic trait.

Summarizing from Example 1, we observed that F, progeny from the different crosses developed mammary cancers at an incidence approaching 100%, but latency was prolonged and tumor number was reduced relative to the ACI strain. The genetically diverse F₂ populations exhibited great phenotypic diversity, with some F2 progeny being as highly susceptible as the ACI strain and some as resistant as the COP or BN strains. The BC progeny approximated the ACI strain in susceptibility to E2-induced mammary cancers.

Genotypic analysis of several hundred phenotypically defined F2 and BC progeny from the different crosses revealed three loci that modify susceptibility. Emcal was localized to rat chromosome 5 in each of the ACI x COP and ACI x BN intercrosses. The ACI allele oϊ Emcal acts to lessen the duration of E2 treatment required to induce palpable mammary cancers; average latency in female F2 progeny of genotype A/A at 5Rat37 was 184 days compared to 265 days for F2 progeny of genotype C/C at this marker (Table 1 , Detailed Description). Emcal and Emca3 were localized to rat chromosomes 18 and 2, respectively, in the ACI x BN intercross. The ACI alleles oϊ Emcal and Emca3 increase susceptibility to E2-induced mammary cancers (Table 1, Detailed Description). Interestingly, neither Emcal nor Emca 3 impacted susceptibility to E2-induced mammary cancers in the ACI x COP and COP x ACI intercrosses. This observation can be inteφreted in two ways: 1) the ACI and COP strains share alleles for Emcal and Emca3; and 2) the ability of the ACI alleles oϊ Emcal and Emca3 to enhance susceptibility to E2-induced mammary cancers is strongly impacted by genetic background.

Two-thirds of mammary cancers induced in ACI rats by chronic E2 treatment exhibited aneuploidy. A genome- wide scan for loss of heterozygosity (LOH) has revealed several, apparently non-random, patterns of LOH. One explanation is that the extended latency observed in ACI/COP FI and ACI/BN FI progeny, relative to the parental ACI strain, results as a consequence of a required somatic LOH event in the mammary epithelium involving one or more Emca loci.

Identification of loci that modify estrogen-induced pituitary tumor development. The pituitary growth responses to DES observed in the intercrosses between the ACI and COP or BN strains are modified by several loci, Eptl - Ept9. Three loci, Eptl, Eptl and Eptό were found to modify DES-induced pituitary growth in ACI x COP and COP x ACI crosses. Unexpectedly, the COP allele of the Ept6 locus, rather than the ACI allele, contributes to the DES-induced pituitary growth.

The ACI x BN cross revealed six additional loci, Ept 3, 4, 5, 7, 8 and 9 that modify DES-induced pituitary tumor development. Again, it was unexpected to observe that the BN alleles of Ept3 and Ept4, rather than the ACI alleles, contributed to the DES-induced pituitary weight gain.

The foregoing data illustrate that the genetic control of estrogen action in the regulation of pituitary growth is much more complex than previously thought, with several loci contributing to the modification of the growth response to estrogen. Additional Ept loci may exist, which will be revealed from further genetic analysis of these strains.

The present invention is not limited to the embodiments described and exemplified above, but is capable of variation and modification without departure from the scope of the appended claims.

Claims

What is claimed:

1. An animal model displaying susceptibility to estrogen-induced mammary cancer and resistance to estrogen-induced pituitary tumor development, comprising a rat strain which, when administered a physiological amount of estrogen, develops a phenotype of one or more estrogen-induced mammary cancers, but does not develop a phenotype of estrogen-induced pituitary tumor development.

2. The animal model of claim 1, wherein the estrogen is 17β-estradiol.

3. The animal model of claim 1, wherein the phenotype of one or more estrogen-induced mammary cancers is selected from the group consisting of intraductal carcinomas, papillary carcinomas, adenocarcinomas, cribiform carcinomas and invasive carcinomas.

4. The animal model of claim 1, wherein the phenotype of estrogen- induced pituitary tumor development is selected from the group consisting of pituitary weight gain and increased circulating prolactin.

5. The animal model of claim 1 , wherein the rat strain is obtained by conventional selective breeding and phenotypic analysis.

6. The animal model of claim 5, wherein the rat strain is a progeny of an ACI rat strain.

7. The animal model of claim 6, wherein the rat strain is a progeny of a cross between an ACI rat strain and a COP rat strain.

8. The animal model of claim 6, wherein the rat strain is a progeny of a cross between and ACI rat strain and a BN rat strain.

9. The animal model of claim 1 , wherein the rat strain's susceptibility to estrogen-induced mammary cancer is genetically defined.

10. The animal model of claim 9, wherein the rat strain's susceptibility to estrogen-induced mammary cancer is defined by the presence of at least one susceptibility allele of at least one locus selected from the group consisting oϊ Emcal, Emcal, and Emca3.

1 1. The animal model of claim 1, wherein the rat strain's resistance to estrogen-induced pituitary tumor development is genetically defined.

12. The animal model of claim 1 1, wherein the rat strain's resistance to estrogen-induced pituitary tumor development is defined by absence of a sufficient number of tumor growth promoting alleles of loci selected from the group consisting oϊEptl, Eptl, Ept 3, Ept4, Ept 5, Ept6, Ept7, Ept8 and Ept9 such that the strain exhibits an amount DES-induced pituitary weight gain equivalent to that displayed by a COP or BN rat strain.

13. The animal model of claim 12, wherein the rat strain lacks tumor growth promoting alleles oϊEptl and Eptl.

14. An animal model for estrogen-induced mammary cancers, comprising an ACI rat strain which has been genetically modified at one or more loci selected from the group consisting oϊ Emcal, Emcal and Emca3, such that the strain is either heterozygous or homozygous for a resistance-conferring allele at the one or more loci.

15. The animal model of claim 14, wherein the one or more resistance- conferring alleles is obtained from a BN or COP rat strain.

16. A congenic set of animal models according to claim 14, comprising one or more genetically modified ACI rat strains selected from the group consisting of: a) a strain heterozygous or homozygous for resistance- conferring alleles oϊ Emcal; b) a strain heterozygous or homozygous for resistance- conferring alleles oϊ Emcal; c) a strain heterozygous or homozygous for resistance- conferring alleles oϊEmca3; d) a strain heterozygous or homozygous for resistance- conferring alleles oϊ Emcal and Emcal; e) a strain heterozygous or homozygous for resistance- conferring alleles oϊ Emcal and Emca3; ϊ) a strain heterozygous or homozygous for resistance- conferring alleles oϊ Emcal and Emca3; and g) a strain heterozygous or homozygous for resistance- conferring alleles oϊ Emcal, Emcal and Emca3.

17. An animal model for estrogen-induced mammary cancers, comprising a BN or COP rat strain which has been genetically modified at one or more loci selected from the group consisting oϊ Emcal, Emcal and Emca3, such that the strain is either heterozygous or homozygous for a susceptibility-conferring allele at the one or more loci.

18. The animal model of claim 17, wherein the one or more susceptibility-conferring alleles is obtained from an ACI rat strain.

19. A congenic set of animal models according to claim 17, comprising one or more genetically modified BN or COP rat strains selected from the group consisting of: a) a strain heterozygous or homozygous for susceptibility- conferring alleles oϊ Emcal; b) a strain heterozygous or homozygous for susceptibility- conferring alleles oϊ Emcal; c) a strain heterozygous or homozygous for susceptibility- conferring alleles oϊEmca3; d) a strain heterozygous or homozygous for susceptibility- conferring alleles oϊ Emcal and Emcal; e) a strain heterozygous or homozygous for susceptibility- conferring alleles oϊ Emcal and Emca3; f) a strain heterozygous or homozygous for susceptibility- conferring alleles oϊ Emcal and Emca3; and g) a strain heterozygous or homozygous for susceptibility- conferring alleles oϊ Emcal, Emcal and Emca3.

20. An animal model for estrogen-induced pituitary tumor development, comprising an ACI rat strain which has been genetically modified at one or more loci selected from the group consisting oϊEptl, Eptl, Ept 5, Ept 7, Ept8 and Ept9, such that the strain is either heterozygous or homozygous for a resistance- conferring allele at the one or more loci.

21. The animal model of claim 20, wherein the one or more resistance- conferring alleles is obtained from a BN or COP rat strain.

22. A congenic set of animal models according to claim 20, comprising one or more genetically modified ACI rat strains selected from the group consisting of: a) a strain heterozygous or homozygous for resistance- conferring alleles oϊEptl; b) a strain heterozygous or homozygous for resistance- conferring alleles oϊEptl; c) a strain heterozygous or homozygous for resistance- conferring alleles of Ept5; d) a strain heterozygous or homozygous for resistance- conferring alleles of Ept 7; e) a strain heterozygous or homozygous for resistance- conferring alleles oϊEpt8; ϊ) a strain heterozygous or homozygous for resistance- conferring alleles of EptP; g) a strain heterozygous or homozygous for resistance- conferring alleles of any combination of two of Eptl, Eptl, Ept5, Ept7, Ept8 and Ept9; h) a strain heterozygous or homozygous for resistance- conferring alleles of any combination of three of Ept/, Eptl, Ept5, Ept7, Ept8 and Ept9; i) a strain heterozygous or homozygous for resistance- conferring alleles of any combination of four of Ept/, Eptl, Ept5, Ept 7, Ept8 and Ept9; j) a strain heterozygous or homozygous for resistance- conferring alleles of any combination of five of Eptl, Eptl, Ept5, Ept7, Ept8 and Ept9; and k) a strain heterozygous or homozygous for resistance- conferring alleles of Eptl, Eptl, Ept 5, Ept 7, Ept8 and Ept9.

23. An animal model for estrogen-induced pituitary tumor development, comprising a COP or BN rat strain which has been genetically modified at one or more loci selected from the group consisting of Eptl, Eptl, Ept5, Ept7, Ept8 and Ept9, such that the strain is either heterozygous or homozygous for a susceptibility-conferring allele at the one or more loci.

24. The animal model of claim 23, wherein the one or more susceptibility-conferring alleles is obtained from an ACI rat strain.

25. A congenic set of animal models according to claim 23, comprising one or more genetically modified BN or COP rat strains selected from the group consisting of: a) a strain heterozygous or homozygous for susceptibility- conferring alleles oϊEptl; b) a strain heterozygous or homozygous for susceptibility- conferring alleles oϊEptl; c) a strain heterozygous or homozygous for susceptibility- conferring alleles of Ept5; d) a strain heterozygous or homozygous for susceptibility- conferring alleles of Ept7; e) a strain heterozygous or homozygous for susceptibility- conferring alleles of EptS; f) a strain heterozygous or homozygous for susceptibility- conferring alleles of Ept9; g) a strain heterozygous or homozygous for susceptibility- conferring alleles of any combination of two of Eptl, Eptl, Ept5, Ept7, Ept8 and Ept9; h) a strain heterozygous or homozygous for susceptibility- conferring alleles of any combination of three of Eptl, Eptl, Ept5, Ept7, Ept8 and Ept9; i) a strain heterozygous or homozygous for susceptibility- conferring alleles of any combination of four of Eptl, Eptl, Ept5, Ept7, Ept8 and Ept9; j) a strain heterozygous or homozygous for susceptibility- conferring alleles of any combination of five of Eptl, Eptl, Ept5, Ept7, Ept8 and Ept 9; and k) a strain heterozygous or homozygous for susceptibility- conferring alleles of Eptl, Eptl, Ept5, Ept7, Ept8 and Ept9.

26. An animal model for estrogen-induced pituitary tumor development, comprising a BN or COP rat strain which has been genetically modified at one or more loci selected from the group consisting oϊEpt3, Ept4 and Eptό, such that the strain is either heterozygous or homozygous for a resistance-conferring allele at the one or more loci.

27. The animal model of claim 26, wherein the one or more resistance- conferring alleles is obtained from an ACI rat strain.

28. A congenic set of animal models according to claim 26, comprising one or more genetically modified BN or COP rat strains selected from the group consisting of: a) a strain heterozygous or homozygous for resistance- conferring alleles oϊEpt3; b) a strain heterozygous or homozygous for resistance- conferring alleles oϊEpt4; c) a strain heterozygous or homozygous for resistance- conferring alleles oϊEptό; d) a strain heterozygous or homozygous for resistance- conferring alleles oϊEpt3 and Ept4; e) a strain heterozygous or homozygous for resistance- conferring alleles oϊEpt3 and Eptό; f) a strain heterozygous or homozygous for resistance- conferring alleles oϊEpt4 and Ept7; and g) a strain heterozygous or homozygous for resistance- conferring alleles oϊEpt3, Ept4 and Eptό.

29. An animal model for estrogen-induced pituitary tumor development, comprising an ACI rat strain which has been genetically modified at one or more loci selected from the group consisting oϊEpt3, Ept4 and Eptό, such that the strain is either heterozygous or homozygous for a susceptibility-conferring allele at the one or more loci.

30. The animal model of claim 29, wherein the one or more susceptibility-conferring alleles is obtained from a BN or COP rat strain.

31. A congenic set of animal models according to claim 29, comprising one or more genetically modified ACI rat strains selected from the group consisting of: a) a strain heterozygous or homozygous for susceptibility- conferring alleles oϊEpt3; b) a strain heterozygous or homozygous for susceptibility- conferring alleles oϊEpt4; c) a strain heterozygous or homozygous for susceptibility- conferring alleles oϊEptό; d) a strain heterozygous or homozygous for susceptibility- conferring alleles oϊEpt3 and Ept4; e) a strain heterozygous or homozygous for susceptibility- conferring alleles oϊEpt3 and Eptό; f) a strain heterozygous or homozygous for susceptibility- conferring alleles oϊEpt4 and Eptό; and g) a strain heterozygous or homozygous for susceptibility- conferring alleles oϊEpt3, Ept4 and Eptό.

32. A genetic locus that modifies susceptibility or resistance to estrogen-induced mammary cancer in rats, the locus being linked to a physically identifiable chromosomal marker.

33. The genetic locus of claim 32, wherein the physically identifiable chromosomal marker comprises a polymoφhism that distinguishes a susceptibility- conferring allele of the linked locus from a resistance-conferring allele of the linked locus.

34. The genetic locus of claim 32, selected from the group consisting oϊ Emcal, Emca2 and Emca3.

35. The genetic locus of claim 34, wherein Emcal is linked to 5Rat37 or 5rat95, Emcal is linked tol8Rat30, and Emca3 is linked to 2Ratl6.

36. A genetic locus that modifies susceptibility or resistance to estrogen-induced pituitary tumor development in rats, the locus being linked to a physically identifiable chromosomal marker.

37. The genetic locus of claim 36, wherein the physically identifiable chromosomal marker comprises a polymoφhism that distinguishes a susceptibility- conferring allele of the linked locus from a resistance-conferring allele of the linked locus.

38. The genetic locus of claim 36, selected from the group consisting of Eptl, Eptl, Ept3, Ept4, Ept5, Eptό, Ept7, Ept8 and Ept9.

39. The genetic locus of claim 38, wherein Eptl is linked to 6Rat80, Eptl is linked to 3Rat26, Ept3 is linked to 6Rat22, Ept4 is linked to 12Rat53, EptS is linked to 18Rat57, Eptό is linked to 3Mgh9, Ept7 is linked to 2Ratl9, Ept8 is linked to 2Ratl47, and Ept9 is linked to 10Mit7.

40. A cell line derived from the animal model of claim 1, 14, 17, 20, 23, 26 or 29.

41. The cell line of claim 40, derived from mammary tissue.

42. The cell line of claim 40, derived from pituitary tissue.