Brca1 and Brca2 expression patterns in mitotic and meiotic ... - Nature

Oncogene (1998) 16, 61 ± 68  1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00

Brca1 and Brca2 expression patterns in mitotic and meiotic cells of mice. Pamela E Blackshear1, Susan M Goldsworthy2, Julie F Foley1, Kimberly A McAllister1, L Michelle Bennett1, N Keith Collins1, Donna O Bunch1, Paula Brown1, Roger W Wiseman1 and Barbara J Davis1 1

National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina 27709; 2Pathology Associates International, Durham, North Carolina 27713, USA

The mouse homologues of the breast cancer susceptibility genes, Brca1 and Brca2, are expressed in a cell cycle-dependent fashion in vitro and appear to be regulated by similar or overlapping pathways. Therefore, we compared the non isotopic in situ hybridization expression patterns of Brca1 and Brca2 mRNA in vivo in mitotic and meiotic cells during mouse embryogenesis, mammary gland development, and in adult tissues including testes, ovaries, and hormonally altered ovaries. Brca1 and Brca2 are expressed concordantly in proliferating cells of embryos, and the mammary gland undergoing morphogenesis and in most adult tissues. The expression pattern of Brca1 and Brca2 correlates with the localization of proliferating cell nuclear antigen, an indicator of proliferative activity. In the ovary, Brca1 and Brca2 exhibited a comparable hormone-independent pattern of expression in oocytes, granulosa cells and thecal cells of developing follicles. In the testes, Brca1 and Brca2 were expressed in mitotic spermatogonia and early meiotic prophase spermatocytes. Northern analyses of prepubertal mouse testes revealed that the time course of Brca2 expression was delayed in spermatogonia relative to Brca1. Thus, while Brca1 and Brca2 share concordant cell-speci®c patterns of expression in most proliferating tissues, these observations suggest that they may have distinct roles during meiosis. Keywords: Brca1; Brca2; mitosis; meiosis

Introduction Most hereditary predisposition to breast and ovarian cancer can be attributed to germline mutations in the BRCA1 or BRCA2 breast cancer susceptibilty genes (Stratton, 1996). Germline mutations in the BRCA1 gene increase the risk for the development of earlyonset breast cancer and ovarian cancer (Miki et al., 1994). Mutations in the BRCA2 gene result in an increased risk for the development of breast cancer in both women and men (Wooster et al., 1995; Stratton, 1996), a moderate increased risk for the development of ovarian cancer in women (Stratton, 1996), and are associated with pancreatic, cervical, and laryngeal cancers (Wooster et al., 1995; Gudmundsson et al., 1995). Inherited mutations in either BRCA1 or BRCA2 are also associated with a signi®cant risk of

Correspondence: PE Blackshear Received 13 June 1997; revised 15 August 1997; accepted 15 August 1997

colon cancer and predispose men to the development of prostate cancer (Wooster et al., 1995; Stratton, 1996; Gudmundsson et al., 1995). Since inactivation of both alleles of either BRCA1 or BRCA2 are key features in neoplastic development in hereditary cancers, these genes are believed to act as tumor suppressor genes required for cell growth. Understanding how BRCA1 and BRCA2 function during normal growth and development should help to de®ne why their inactivation yields such profound predispositions to neoplasia. Several studies have demonstrated a relationship between BRCA1 and BRCA2 gene function and normal growth control. For example, BRCA1 and BRCA2 are induced at the G1/S boundary in normal human mammary epithelial cell and breast cancer cell lines stimulated to proliferate (Vaughn et al., 1996a, b; Gudas et al., 1996). BRCA1 message and protein expression are induced by the mitogenic activity of hormones, such as in estrogen responsive cell lines (Marks et al., 1997; Gudas et al., 1995). In human breast cancer cell lines and spermatocytes, BRCA1 protein binds to Rad51, which functions in DNA recombination and repair, supporting a role for BRCA1 speci®cally during mitosis and meiosis (Scully et al., 1997). Recent studies indicate a similar interaction between Rad51 and the Brca2 gene product (Sharan et al., 1997). The mouse homologues, Brca1 and Brca2, appear to function in normal growth control and their similar expression patterns imply that their regulatory pathways may be analogous. However, their phenotypic expressions in carcinogenesis associated with loss of gene function is not completely redundant, as demonstrated by the spectrum of tumors that develop in mutation carriers. Therefore, we investigated the relationship between Brca1 and Brca2 tissue expression patterns during embryonic development, puberty, and in adult mice. Additionally, we determined whether Brca2 was expressed independently of hormonal stimulation by comparing expression of Brca2 in ovaries of cycling mice, hypophysectomized mice, and in estrogen receptor knock-out (ERKO) mice as previously determined for Brca1 (Phillips et al., 1997). Our results demonstrate that Brca1 and Brca2 have concordant patterns of expression during embryogenesis, mammary gland morphogenesis, and adult tissues associated with proliferating cell populations, terminally differentiated neurons as well as hormonally altered ovaries, but there are dierences in their temporal expression in the testis. Thus, although there is considerable overlap, their regulatory pathways are not identical.

Brca1 and Brca2 expression PE Blackshear et al

62

Results Brca1 and Brca2 expression during mouse embryogenesis During embryonic development, in situ hybridization studies demonstrated that Brca2 was ubiquitously expressed throughout the three germ layers of day 7.5 embryos which evolved into a tissue speci®c pattern through day 19. The expression pattern of Brca2 was analogous to that of Brca1 at the same time points, and in most instances, correlated with PCNA positive staining (Figure 1). In gestational day 14 embryos, intense in situ hybridization staining, which represents relatively high expression for both Brca1 and Brca2 transcripts, was localized to neuroectodermal tissues, including the ventricular zone, and medial and lateral walls of the cerebral hemispheres, trigeminal ganglia, primodial spinal nerve roots and ganglia, Rathke's pouch, and neuroepithelium of the eye. These neuroectodermal cells also demonstrated PCNA positivity except for the medial and lateral walls of the cerebral hemispheres. Similarly high staining intensity, and thus relative expression, was observed in ectodermallyderived tissues including the stratum basalis of the epidermis, tongue, otic cochlea, ameloblasts of the tooth bud, and lens of the eye. High staining intensity was also observed in mesodermally-derived tissues including somites, brown fat, hematopoetic cells in the liver, and cartilage. Within the same embryo section, moderate staining intensity was localized to mesodermally-derived tissues of the genital ridge, including the overlying mesothelium and the Mullerian ducts, and mesenchymal tissues of the limb. Low staining intensity was detected in the heart, which is of mesodermal origin. In endodermally-derived tissues, moderate staining intensity was found in hepatocytes and intestinal epithelium. In the lung (endodermal origin), the alveolar buds stained moderately for both

1A

1B

transcripts while the bronchial epithelial cells stained with low intensity. By day 19, when most tissues had reached their de®nitive prenatal morphology, both Brca1 and Brca2 were expressed with the same relative staining intensity as described for the day 14 embryos except there was no evidence of Brca1 or Brca2 expression in the adrenal gland or urinary bladder. Brca1 and Brca2 expression in adult mouse tissues The expression patterns of Brca1 was similar to Brca2 in the adult mouse. In situ hybridization of Brca1 and Brca2 speci®cally localized expression to several tissues, which in most instances, correlated with PCNA positive staining (not shown). Brca1 and Brca2 expression was widely disseminated in lymph nodes containing primary follicles and was speci®cally con®ned to germinal centers of secondary follicles which contain proliferating B lymphocytes. Brca1 and Brca2 expression was localized to pancreatic acinar and islet cells and also correlated well with PCNA labeling (not shown). RNase protection assays con®rmed Brca1 and Brca2 expression in the pancreas. Expression of Brca1 and Brca2 was detected in uterine endometrial glands, the glandular mucosa of the stomach, duodenal crypt epithelial cells, lymphocytes of the periarterial sheaths and lymphatic nodules in the spleen, the outer rim of the thymic cortex which contains proliferating T lymphocytes, and in epithelial cells in the outer root sheath of the hair follicles. Staining for Brca1 and Brca2 transcripts was low in hepatocytes, myocardial cells, and proximal tubular epithelial cells of the kidney. The neurons were the only cells that expressed both Brca1 and Brca2 but the expression pattern was not correlated with cell proliferation. Low expression in the brain was con®rmed by RNase protection assay. Brca1 and Brca2 expression was not detected in lung or skeletal muscle.

1C

1D

Br

Brca2+

Brca2(negative control)

Brca1+

PCNA

Figure 1 In situ hybridization analysis of Brca1 and Brca2 in a gestational day 14 embryo from a CD-1 mouse. (a) Mid-sagittal section of formalin-®xed, paran embedded embryo with Brca2 cRNA digoxigenin-labeled antisense (+) probe. Transcript is visualized as a blue product from alkaline phosphatase-conjugated antidiogoxigenin antibody reacted with NBT-BCIP detection system. The counterstain is Nuclear Fast Red. (b) Serial section of embryo shown in (a) hybridized with Brca2 sense (7) probe to serve as a negative control. (c) Brca1 cRNA digoxigenin-labeled antisense probe hybridized to same embryo as in (a) but dierent plane of section. (d) Serial section of embryo shown in (c) stained for nuclear PCNA protein and visualized as brown by a DAB color reaction. Bar=1 mm for 1(a ± d). Symbols: (Br)-brain, (Tg)-trigeminal ganglia, (Ln)-lung, (Lv)-liver


Brca1 and Brca2 expression during mammary gland ductular morphogenesis in mice Because the loss of either BRCA1 or BRCA2 gene function leads to the development of human breast cancer, we determined the cell type speci®c in vivo expression pattern of Brca1 and Brca2 during normal mammmary gland development (Figure 2). During postnatal ductular proliferation and glandular development during pregnancy, Brca1 and Brca2 mRNA expression was localized to both epithelial and stromal cells (®broblasts). Mammary glands from prepubertal, 2-week-old mice consisted of primordial ducts containing few epithelial cells that expressed Brca1 and Brca2, or that were PCNA immunopositive (not shown). Mammary glands from pubertal, 5-week-old virgins consisted of terminal end buds and branching epithelial ducts and ductules in which Brca1 and Brca2 were expressed in the epithelial cells of terminal end buds, ductal epithelial cells, and the immediately adjacent stromal ®broblasts (Figure 2a). Mammary glands from sexually mature, virgin mice consisted of minimally branching ducts in which staining for Brca1 and Brca2 mRNA was detected in few epithelial and stromal cells (Figure 2b). During early pregnancy to day 13.5 when

Brca2+

Brca2-

lateral alveolar buds were forming, Brca1 and Brca2 were prominently expressed in the epithelial cells and adjacent mammary stromal ®broblasts (Figure 2c). However, Brca1 and Brca2 expression was no longer apparent in adjacent stromal cells at day 14 of pregnancy (not shown). By gestational day 19 when mammary alveoli were well developed, the expression of Brca1 and Brca2 was primarily con®ned to epithelial cells of the ducts and were present in a few alveolar epithelial cells (not shown). In lactating mice, the alveolar epithelial cells were well dierentiated with minimal to undetectable expression of Brca1 or Brca2 (Figure 2d). At one week post-weaning while the alveoli were regressing, there was a notable reexpression of epithelial cell and stromal cell Brca1 and Brca2 transcripts (Figure 2e). These expression patterns correlated well with PCNA immunopositivity (Figure 2a ± e). RNase protection assays using Brca1 and Brca2 probes on total RNA isolated from mammary glands collected from similar time points showed an increase in transcript levels of both genes associated with mammary gland morphogenesis (Figure 3). Mammary glands from 4 week old virgins were used to determine the level of expression associated with terminal end

Brca1+

PCNA

2A 5 wk

2B 12 wk

2C gd 13.5

2D Ld 9

2E 1 wk reg

Figure 2 In situ hybridization analysis of Brca1 and Brca2 compared with PCNA immunohistochemistry during mammary gland growth, dierentiation and regression in the CD-1 mouse. Tissues were hybridized with Brca2 antisense probe (left panel); Brca2 sense probe (middle left panel); Brca1 antisense probe (middle right); or stained for PCNA by immunohistochemistry (right panel). (a) Terminal endbud in 5-week-old virgin (5 wks). Bar=37mm. (b) Duct from adult nulliparous mouse (12 wks). Bar=37mm. (c) Beginning alveolar buds, gestational day 13.5 (gd 13.5). Bar=37mm. (d) Lactation day 9 (Ld 9). Bar=37mm. (e) Mammary gland regression one week post-weaning (1 wk reg). Bar=37mm

63


64

bud formation which occurs in 3 ± 5 week old virgins. Brca1 and Brca2 expression increased in early pregnancy, peaked at gestational day 10.5, and decreased to nearly undetectable levels during lactation associated with alveolar epithelial dierentiation. At 1 week post weaning, during mammary gland regression, expression returned to levels seen in glands of virgin mice and during early pregnancy. These expression levels correlated with the intensity of staining as well as the change in volume of stained cells associated with cell proliferation as demonstrated by in situ hybridization (Figures 2 and 3).

4d). The expression pattern from a normal cycling adult was compared to prepubertal ovaries from 2 week old virgins, the ovaries from hypophysectomized mice, and ovaries from mice lacking the estrogen

Brca1 and Brca2 expression in mouse ovary Since Brca1 is expressed independently of hormonal stimulation in the mouse ovary (Phillips et al., 1997), we determined whether Brca2 expression was dependent or independent of hormonal stimulation, including estrogen, by comparing expression of Brca2 in ovaries of cycling mice, hypophysectomized mice, and in ERKO mice that do not have a normal response to estrogen. In the normal adult ovary, Brca1 and Brca2 transcripts were localized speci®cally to granulosa cells, thecal cells and oocytes of developing follicles (Figure 4a ± c) as well as luteal cells of recently formed corpora lutea (not shown) and surface epithelium. The expression pattern of Brca1 and Brca2 correlated closely with PCNA immunohistochemistry (Figure

Brca2+

Brca2-

4A

4B

Figure 3 Relative expression for Brca1 and Brca2 during mammary growth, dierentiation and regression as determined by RNase protection assay. Each data point represents percentage of Brca1 (^) and Brca2 (*) expression +/- standard error in the mouse mammary gland from 4 week virgin females (4 wk virgin), at gestational day 9 (gd 9), gestational day 10.5 (gd 10.5), gestational day 13 (gd 13), gestational day 19 (gd 19), lactation day 9 (Ld 9), and 1 week regression (1 wk reg)

Brca1+

4C

PCNA

ERKO/Brca2+

4D

4E

Figure 4 In situ hybridization analysis of Brca1 and Brca2 compared to PCNA positive staining in the ovary from adult CD-1 mouse and ERKO mouse: (a) Tissue was hybridized with Brca2 cRNA digoxigenin-labeled antisense probe. (b) Serial section of ovary shown in (a) hybridized with Brca2 sense probe to serve as a negative control. (c) Adjacent section of ovary shown in (a) hybridized with Brca1 cRNA digoxigenin-labeled antisense probe. (d) Adjacent section of ovary shown in (c) stained for PCNA by immunohistochemistry. (e) Ovary section from an ERKO mouse hybridized with Brca2 cRNA digoxigenin-labeled antisense probe. Bar=75mm (a ± e). Symbols: (G)-granulosa cells and (T)-theca cells of developing follicles, (P)-primary follicles


receptor (ERKO). Ovaries from the 2-week-old virgins contained primordial, small and medium sized follicles with no corpora lutea as expected in the prepubertal state. Brca1 and Brca2 were expressed in the oocytes, granulosa cells and thecal cells (not shown) as in the adult cycling mice. Brca2 was also expressed in oocytes, thecal cells, and granulosa cells of small growing follicles in hypophysectomized mice (not shown). Ovaries from homozygous mutant ERKO (7/7) mice contained variably sized follicles and hemorrhagic follicular cysts lined by granulosa cells but corpora lutea were absent. Brca2 expression again was localized to the oocytes, granulosa cells of small and medium sized follicles and thecal cells (Figure 4e). Expression of Brca1 and Brca2 in ovaries from wildtype ERKO (+/+) animals was indistinguishable from cycling adult CD-1 mice. Brca1 and Brca2 expression in mouse testes The human BRCA1 and mouse Brca2 proteins interact with the Rad51 protein during mitosis and meiotic recombination (Scully et al., 1997; Sharan et al., 1997) and BRCA1 has been shown to be expressed in meiotic

male germ cells (Zabludo et al., 1996). We examined the mouse testes at various stages of development to speci®cally identify and correlate Brca1 and Brca2 expression in mitotic and meiotic germ cells. Both transcripts were localized in the spermatogonia and spermatocytes of meiotic prophase (Figure 5a ± c). Sertoli cells and Leydig interstitial cells were consistently negative for Brca1 and Brca2 transcripts. Brca1 expression was more readily apparent during the early stages of spermatogenesis (spermatogonia and preleptotene spermatocytes) than Brca2 in similarly staged seminiferous tubules. Brca1 and Brca2 expression was most intense in the late pachytene and diplotene spermatocytes. In contrast, PCNA positive staining was prominent in spermatogonia and pre-leptotene spermatocytes, which comprise the mitotic phase of germ cell development. In addition, PCNA positive nuclear staining was prominent in leptotene and zygotene spermatocytes with minimal to no staining in early pachytene spermatocytes dependent on the stage of spermatogenesis for a given seminiferous tubule (Figure 5d). In eort to evaluate the dierence in temporal expression between Brca1 and Brca2 in the testes, we

Brca2+

Brca2-

5B

5A

Brca1+

PCNA

4D

5C

5D

Figure 5 In situ hybridization analysis of Brca1 and Brca2 compared with PCNA immunohistochemistry in testes from adult CD-1 mouse: (a) Tissue was hybridized with Brca2 cRNA digoxigenin-labeled antisense probe. Leptotene spermatocytes (arrow head) and pachytene spermatocytes (arrow) (b) Serial section of testes shown in (a) hybridized with Brca2 sense probe. (c) Adjacent section of testes shown in (a) hybridized with Brca1 cRNA digoxigenin-labeled antisense probe. (d) Adjacent section of testes shown in (a) stained for PCNA. Bar=37mm (a ± d)

65


66

examined the expression of these genes during the ®rst wave of spermatogenesis in male postnatal mice when spermatogenesis is synchronous. Northern analysis of RNA isolated from the testes of mice starting at postnatal day 8 demonstrated that both genes were expressed as early as postnatal day 8 when the testes contains 27% spermatogonia and 73% Sertoli cells (Bellve et al., 1977). However, the intensity of expression appeared signi®cantly greater for Brca1 than for Brca2 on postnatal day 8, 10 and 12 (Figure 6). By day 14 when the testes contains 12% spermatogonia and 36% preleptotene, leptotene, and zygotene spermatocytes, and 15% of newly emergent pachytene spermatocytes, both Brca1 and Brca2 expression were signi®cantly up-regulated to a comparable degree. Northern analysis of mixed germ cells, pachytene spermatocytes, and round spermatids con®rmed the expression of Brca2 in pachytene spermatocytes that contributed to the increased intensity observed in the testes on postnatal day 14 (not shown). Discussion Brca1 and Brca2 mRNAs are expressed in a concordant cell-type speci®c manner associated with cell proliferation. Both genes are coordinantly expressed from gestational day 7.5 to day 19 of embryogenesis, a process which is characterized by rapid cell growth and morphogenesis. In adulthood, Brca1 and Brca2 transcripts are localized to tissues that maintain high intrinsic rates of proliferation such as the ovary, testes, intestinal crypts, and the mammary

a

Days Post-natal 8 10 12 14 16 18 20 22 24 26 28 30

Brca 1 28S b

Brca 2 28S c Actin

18S

Figure 6 Expression of Brca1 and Brca2 during postnatal development of mouse testes. Northern analyses using the same ®lter hybridized with probes for Brca1 and Brca2, and b-actin are shown in panels a, b, and c, respectively. Days of age are shown above each lane. Fifteen ug of total RNA was loaded per lane. Exposure times were 12 days for Brca1 and Brca2 and 5 days for b-actin. Brca1 (a) is expressed at day 8 and continues throughout the prepubertal period with maximal intensity at day 14 and 16 coincident with appearance of the ®rst wave of pachytene spermatocytes. Brca2 (b) is only expressed at a low level on day 8 and expression increases to reach a maximum on days 14, 16 and 18 coincident with the appearance of pachytene spermatocytes. Actin expression (c) is shown as an RNA loading control

gland during pregnancy. Our results are similar to that reported by Conner and co-workers which established tissue distribution of Brca1 and Brca2 in mouse tissues by RT ± PCR (Conner et al., 1997), and other in situ hybridization studies (Marquis et al., 1995; Lane et al., 1995; Rajan et al., 1996, 1997). Additionally, our studies demonstrate that the expression patterns of Brca1 and Brca2 are closely correlated in most tissues with PCNA nuclear staining, used as a marker for proliferating cells in S-phase. Exceptions to this correlation between Brca1 and Brca2 expression and PCNA positivity are found in the testes and terminally dierentiated neurons. Moreover, during the mitotic phase of spermatogenesis, Brca1 expression appears to precede Brca2 expression. Thus, our data comparing Brca1 and Brca2 in the testes provides the ®rst evidence of temporal dierences in the expression pattern of these genes during spermatogenesis. Spermatogenesis is characterized by waves of mitotic divisions of the spermatogonia stem cells of which a subset commits to meiotic divisions to form the haploid spermatozoa. The stages of the ®rst meiosis occur as a distinct progression from leptotene through diplotene. The second meiotic division occurs rapidly in secondary spermatocytes which in turn give rise to haploid spermatids which then dierentiate into spermatozoa. The initial wave of spermatogenesis is synchronized and the stages readily distinguishable in the postnatal mouse. In the adult, the mitotic and meiotic cells can be distinguished by morphology and cellular association. Thus, the testes proved useful for delineating the dierences between Brca1 and Brca2 in proliferating cell populations. In a previous study Brca1 expression in the testes was found to be limited to meiotic cells (Zabludo et al., 1996). In contrast, our data demonstrates that Brca1 and Brca2 are expressed in mitotic spermatogonia in addition to meiotic spermatocytes through late pachytene and diplotene stages. Perhaps the dierence in results are due to the relative sensitivity of the techniques employed in each study. Most recently in vivo studies indicate that Brca2 expression is upregulated during spermatogenesis from day 12 onward in postnatal mice while Brca1 expression remains relatively constant (Rajan et al., 1997). We ®nd by in situ hybridization and Northern analysis that Brca1 expression was greater than Brca2 during the mitotic phase of spermatogenesis, but that both genes had a comparable increase in intensity during late meiotic prophase. Therefore, we suggest that Brca1 expression is up-regulated prior to Brca2 during the mitotic phase of spermatogenesis. The lag in Brca2 expression in mitotic cells during spermatogenesis implies that the regulation of these gene are not identical and that they may even be interdependent. For example, each gene may be independently regulated or the induction of Brca2 may require the initial expression of Brca1 during spermatogenesis. The expression patterns of Brca1 and Brca2 during spermatogenesis also suggests these genes function in the cell beyond DNA replication, since Brca1 and Brca2 transcripts are expressed for longer periods than the PCNA protein. Additionally, Brca1 and Brca2 expression patterns parallel that of the mouse rad51 protein which is expressed in spermatogonia and early and mid-meiotic prophase spermatocytes (Yamamoto et al., 1996). Human BRCA1 and mouse Brca2


proteins bind to Rad51 which is thought to be involved in repair of DNA double strand breaks, and may act to monitor the process of DNA replication and recombination (Scully et al., 1997; Sharan et al., 1997). Therefore, our results substantiate a role for Brca1 and Brca2 in meiotic as well as mitotic recombination and replication. The expression of Brca1 and Brca2 in the nervous system suggests a further complexity of function. An earlier study investigating the tissue distribution of Brca1 using Northern analysis of adult mouse tissues reported low but detectable expression in brain (Lane et al., 1995). However, Rajan and co-workers recently found only Brca2 is expressed in the adult brain, embryonic brain and spinal cord, whereas Brca1 is found in particular areas of embryonic brain but not the adult brain (Rajan et al., 1997). Our in situ hybridization studies in embryonic and adult brains, and RNase protection assays of adult brain demonstrate that both Brca1 and Brca2 transcripts are expressed in neural tissues. Thus, while Brca1 and Brca2 appear to function in association with cell proliferation in other tissues, in the brain these gene products may be involved in neuronal growth and maintenance perhaps by interaction with proteins that are common to both neurons and mitotic cells, such as some structural proteins (Cleveland, 1990; Kuriyama and Nislow, 1992). Since loss of function of the BRCA1 and BRCA2 genes results in a dramatic predisposition to breast and ovarian cancer in humans, a central question focuses on the biology of these tissues and why they are more susceptible to tumor formation. In the mammary gland, we have demonstrated that Brca1 and Brca2 transcripts are coordinately expressed and that they are readily detected in both epithelial and stromal cells during periods of ductal and glandular proliferation. Mammary ductal elongation and alveolarization are dependent on both epithelial and stromal cell proliferation and epithelial and stromal cell interaction (Daniel and Silberstein, 1987; Howlett and Bissell, 1993). The localization of Brca1 and Brca2 transcripts in both epithelial and stromal cell populations during mammary gland development is consistent with the hypothesis that both of these cell populations are involved in the development of mammary cancers (Smith et al., 1991). Additionally, the temporal expression of Brca1 and Brca2 during mammary gland development clearly correlates with ductular proliferation and morphogenesis. These results are consistent with in vitro studies that show expression of BRCA1 and BRCA2 mRNA is regulated by the cell cycle and that these two genes are induced with similar kinetics at the G1/S boundary in normal human mammary epithelial cells and breast cancer cell lines (Vaughn et al., 1996a,b; Gudas et al., 1996). Furthermore, recent in vitro and in vivo studies demonstrate that Brca1 and Brca2 are coordinately regulated in mouse mammary epithelial cells and the mammary gland during puberty and pregnancy (Rajan et al., 1996; 1997) which is consistent with our results. The presence of signi®cant Brca1 expression in hormonally responsive tissues such as the testes, mammary gland, ovary and uterus led to the hypothesis that Brca1 may be hormonally regulated (Marquis et al., 1995). Moreover, Brca1 and Brca2 may be regulated dierently by sex hormones (Rajan et al., 1997).

However, the expression of Brca2, like Brca1 (Phillips et al., 1997), is expressed in the ovary independently of various hormonal stimulations. The similar pattern of expression of both genes suggests that their expression is closely related to the cell cycle. In vitro studies have also demonstrated that BRCA1 expression is up-regulated due to mitogenic eects of several hormones such as estrogen, rather than a direct interaction with the gene (Marks et al., 1997; Gudas et al., 1995). In summary, we show that Brca2, like Brca1, is widely expressed in the embryo and a variety of adult tissues associated with the cell cycle. We also demonstrate that Brca1 and Brca2 have similar in vivo, hormone independent-patterns of cell speci®c expression during normal growth and dierentiation in the ovary. Moreover, Brca1 and Brca2 are expressed beyond the mitotic phase of spermatogenesis during meiotic recombination which is consistent with proposed functions for these breast cancer susceptibility genes related to maintaining genome integrity. Although Brca1 and Brca2 appear to be coordinately regulated in many tissues, the temporal divergence in expression between Brca1 and Brca2 in the testes suggests that the regulatory pathway for these two genes may not be identical. Materials and methods Animals Young virgin and timed pregnant CD-1 mice were obtained from Charles River (Raleigh, NC). The morning of a vaginal plug was counted as day 0.5 post-coitus. The following tissues were frozen or ®xed for 24 h in 10% neutral buered formalin and then embedded in paran: fetuses (from day 7.5, 9, 10.5, 13, 13.5, 14, 14.5, 19 postcoitus): the fourth mammary gland (from day 7.5, 9, 10.5, 13, 13.5, 14, 14.5, 19 post-coitus; lactation day 9; and 2-day, 1-week and 4-weeks post weaning; and from 2-week old, 5week old, 12-week old, and 16-week old virgins); ovary, uterus, liver, spleen, kidney, stomach, small and large intestine, pancreas, heart, lung, brain and skeletal muscle. RNA probe preparation Brca1 and Brca2 cRNA probes were generated from a 143 base pair cDNA fragment containing exon 7 of the murine Brca1 gene (nucleotides 418 ± 560, Accession # U32446) and a 163 base pair cDNA fragment from exon 11 of the murine Brca2 gene (nucleotides 3283 ± 3445, Accession # U89652). These PCR products were subcloned into the EcoRV site of pBluescript1 (Stratagene, La Jolla, CA). The resulting Brca1 plasmid was linearized by digestion with either BamHI or NotI, and the Brca2 plasmid was linearized with either HindIII or BamHI (New England Biolabs). RNase protection assays were performed on mouse testis and mammary gland tissues as previously described (McAllister et al., 1997). RNase protection analysis of mammary gland was performed with pooled mammary gland tissue from one to six mice to obtain 40 mg of total RNA for each time point. The percent relative expression for Brca1 and Brca2 was obtained by taking the raw counts from two samples of pooled total RNA for each data point and normalizing their value relative to the highest expression at gestation day (gd) 10.5 and then averaged. The relative percent for gd 10.5 is set at 100%. The digoxigenin-labeled probes were prepared as previously described (Phillips et al., 1997). For in situ hybridization, the tissue sections were de-paranized,

67


68

treated with 0.2 M HCL, 0.3% Triton-X 100, proteinase K and acetylated. The fetuses were treated with 20mg/ml proteinase K for 5 min at 378C; all other tissues were treated with 10 mg/ml for 12 min at 378C. The RNA probes were hybridized overnight at 558C for mammary glands and fetuses and 488C for all other tissues. The hybridized probe was detected using an alkaline phosphatase-conjugated antidigoxigenin as previously described (Phillips et al., 1997). To determine relative expression of transcripts during embryonic development, staining intensity for speci®c transcripts was compared for each organ or anatomical region within the same slide from at least three separate experiments and scored by two pathologists (PEB and BJD) as high, moderate, low, or no signal detected. Localization of Brca1 and Brca2 expression during mammary gland development and in adult tissues was determined by identi®cation of signal from at least three separate experiments and compared to appropriate, concurrently run negative controls. The proliferating cell nuclear antigen (PCNA) protein was detected by immunohistochemistry with 3,3'-diaminobenzidine (DAB) as previously described (Foley et al., 1993) in adjacent sections of tissues stained for Brca1 or Brca2. The dilution of primary antibody was 1 : 200 and secondary antibody was 1 : 800. PCNA immunopositivity was determined by visualizing the number of cells with dark brown to black nuclear Sphase staining per tissue area in adjacent sections.

Northern analysis The testes of prepubertal CD-1 mice from Charles River (Raleigh, NC) were collected on post-natal days 8 ± 30 and frozen in liquid nitrogen. Mixed spermatogenic cells, pachytene spermatocytes (91.5%), and round spermatids (92%) were isolated from adult CD-1 mice as described previously (Romrell et al., 1976; Bellve et al., 1977; O'Brien et al., 1989). Total RNA was isolated using Trizol reagent (Life Technologies, Grand Island, NY). RNA samples were separated on a 1.2% agarose/ 6.6% formaldehyde gel and transferred by capillary action to Genescreen membranes (New England Nuclear Research Products, Boston, MA) according to the manufacturer's instructions. Northern blotting was performed as described previously (Welch et al., 1995).

Acknowledgements The authors wish to thank E Mitch Eddy of the National Institute of Environmental Health Sciences (NIEHS) for his expert advice and the donation of mouse testes RNA; Debbrah A O'Brien of the University of North Carolina at Chapel Hill for her gift of isolated germ cells; Kenneth S Korach and John F Couse of NIEHS for the ERKO ovary sections; Gina Goulding, Cindy R Moomaw and Norris Flagler for their excellent technical assistance.

References Bellve AR, Cavicchia JC, Millette DF and O'Brien DA. (1977). J. Cell Biol., 74, 68 ± 85. Cleveland D. (1990). Cell, 60, 701 ± 702. Conner F, Smith A, Wooster R, Stratton M, Dixon A, Campbell E, Tait TM, Freeman T and Ashworth A. (1997). Human Molecular Genetics, 6, 291 ± 300. Daniel CW and Silberstein GB. (eds) (1987). Postnatal development of the rodent mammary gland. Plenum Press: New York Foley J, Ton T, Maronpot R, Butterworth B and Goldsworthy TL. (1993). Environmental Health Perspectives, 101, 199 ± 206. Gudas JM, Li T, Nguyen H, Jensen D, Rauscher FJ III and Cowan KH. (1996). Cell Growth & Dier., 7, 717 ± 723. Gudas JM, Nguyen H, Li T and Cowen KH. (1995). Cancer Res., 55, 4561 ± 4565. Gudmundsson J, Johannesdottir G, Bergthorsson JT, Arason A, Ingvarsson S, Egilsson V and Barkardottir RB. (1995). Cancer Res., 55, 4830 ± 4832. Howlett AR and Bissell MJ. (1993). Epithelial Cell Biol., 2, 79 ± 89. Kuriyama R and Nislow C. (1992). BioEssays, 14, 81 ± 88. Lane TF, Deng C, Elson A, Lyu MS, Kozak CA and Leder P. (1995). Genes and Develop., 9, 2712 ± 2722. Marks JR, Huper G, Vaughn JP, Davis PL, Norris J, McDonnell DP, Wiseman RW, Futreal PA and Inglehart JD. (1997). Oncogene, 14, 115 ± 121. Marquis ST, Rajan JV, Wynshaw-Boris A, Xu J, Yin G-Y, Abel KJ, Weber BL and Chodosh LA. (1995). Nature Genetics, 11, 17 ± 26. McAllister KA, Haugen-Strano A, Hagevik S, Brownlee HA, Collins NK, Futreal PA, Bennett LM and Wiseman RW. (1997). Cancer Res., 57, 3121 ± 3125. Miki Y, Swensen J, Shattuck-Eidens D, Futreal P, Harshman K, Tavtigian S, Liu Q, Cochran C, Bennett LM, Ding W, Bell R, Rosenthal J, Hussey C, Tran T, McClure M, Frye C, Hattier T, Phelps R, Haugen-Strano A, Katcher H, Yakumo K, Gholami Z, Shaer D, Stone S, Bayer S, Wray C, Borgden R, Dayananth P, Ward J, Tonin P, Narod S, Bristow P, Norris F, Helvering L, Morrison P, Roseteck P, Lai M, Barrett JC, Lewis C, Neuhaussen S, Cannon-Albright L, Goldgar D, Wiseman R, Kamb A and Skolnick MH. (1994). Science, 266, 66 ± 71.

O'Brien DA, Gabel CA, Rockett DL and Eddy EM. (1989). Endocrinology, 125, 2973 ± 2984. Phillips KW, Goldsworthy SM, Bennett LM, Brownlee HA, Wiseman RW and Davis BJ. (1997). Lab. Invest., 76, 419 ± 425. Rajan JV, Marquis ST, Gardner HP and Chodosh LA. (1997). Dev. Biol., 184, 385 ± 401. Rajan JV, Wang M, Marquis ST and Chodosh LA. (1996). Pro. Nat. Acad. Sci. USA, 93, 13078 ± 13083. Romrell LJ, Bellve AR and Fawcett DW. (1976). Dev. Biol., 49, 119 ± 131. Scully R, Chen J, Plug A, Xiao Y, Weaver D, Feunteun J, Ashley T and Livingston DM. (1997). Cell, 88, 265 ± 275. Sharan SK, Morimatsu M, Albrecht U, Lim D-S, Regel E, Dinh C, Sands A, Eichele G, Hasty P and Bradley A. (1997). Nature, 386, 804 ± 810. Smith H, Stern R, Liu E and Benz C. (1991). Basic Life Sce., 57, 329 ± 340. Stratton MR. (1996). Hum. Mol. Genetics, 5, 1515 ± 1519. Vaughn J, Cirisano FD, Huper G, Berchuck A, Futreal P, Marks J and Inglehart J. (1996a). Cancer Res., 56, 4590 ± 4596. Vaughn J, Davis P, Jarboe M, Huper G, Evans A, Wiseman R, Berchuck A, Iglehart J, Futreal P and Marks J. (1996b). Cell Growth and Dierentiation, 7, 711 ± 715. Welch JE, Brown PR, O'Brien DA, Eddy EM. (1995). Developmental Genetics, 16, 179 ± 189. Wooster R, Bignell G, Lancaster J, Swift S, Seal S, Mangion J, Collins N, Gregory S, Gumbs C, Mickiem G, Barfoot R, Hamoudi R, Patel S, Rice C, Biggs P, Hashim Y, Smith A, Connor F, Arason A, Gudmundsson J, Ficenec D, Kelsell D, Ford D, Tonin P, Bishop DT, Spurr NK, Ponder BAJ, Eeles R, Peto J, Devilee P, Cornelisse C, Lynch H, Narod S, Lenoir G, Eglisson V, Barkadottir RB, Easton DF, Bentley DR, Futreal PA, Ashworth A and Stratton MR. (1995). Nature, 378, 789 ± 792. Yamamoto A, Taki T, Yagi H, Habu T, Yoshida K, Yoshimura Y, Yamamoto K, Matsushiro A, Nishimune Y and Morita T. (1996). Mol. Gen. Genet., 251, 1 ± 12. Zabludo SD, Wright WW, Harshman K and Wold BJ. (1996). Oncogene, 13, 649 ± 653.