Absence of the Full-Length Breast Cancer–Associated Gene-1 Leads to Increased Expression of Insulin-Like Growth Factor Signaling Axis Members

Aging has been defined in humans as the age-related deterioration of physiologic functions necessary for the survival and fertility of an organism (1). Multiple environmental- or genetic-related factors have been identified. For instance, caloric or food restriction in mice has been clearly associated with delayed aging (2). On the other hand, altered expression of members of the helicase gene family has been linked to premature aging syndromes, such as Werner or Cockayne syndromes. Indeed, use of mice models has confirmed and reinforced the role of several proteins involved in DNA replication, transcription, and repair or cell growth (1, 3–5). For example, hyperactivity of p53 or chronic excess of growth hormone (GH)/insulin-like growth factor (IGF) I signaling has been associated with various “aging” symptoms (6).

Recently, we have shown that mice, which are homozygous for a targeted deletion of the full-length Brca1, survive to adulthood in a p53 heterozygous mutation background (Brca1^Δ11/Δ11p53^+/−) but exhibit premature aging characterized by decreased life span, reduced body fat deposition, osteoporosis, skin atrophy, and decreased wound healing (7). The Brca1^Δ11/Δ11p53^+/− mice also suffer increased tumorigenesis when the remaining wild-type (WT) p53 allele is lost (7, 8). More importantly, loss of p53 rescues embryonic and cellular senescence (7). These observations are consistent with recent findings that p53 activation plays an important role in aging process (9, 10).

Alterations of IGF-I signaling have been linked to increased adult life span (11, 12). It has been shown that knockout of the insulin receptor in the adipose tissue or deletion of one allele of the IGF-I receptor (IGF-IR) in the whole organism increases the adult life span significantly (11, 12). Because IGF-I is predominantly produced by the liver, we did microarrays using RNA isolated from the liver of Brca1-mutant and control mice to investigate the molecular mechanisms associated with aging of Brca1^Δ11/Δ11p53^+/− mice. We provide evidence that breast cancer–associated gene-1 (BRCA1) deficiency triggers increased expression of several IGF signaling axis members, suggesting that these alterations play a role in the aging process of mutant mice.

Mice. Brca1^Δ11/Δ11p53^+/− mice were generated as described previously (13). All experiments were approved by the Animal Care and Use Committee of National Institutes of Diabetes, Digestive and Kidney Diseases (NIDDK).

IGF-I serum levels. IGF-I serum levels were measured using radioimmunoassay kits (Amersham Life Science) as described (14). Briefly, blood samples were collected from the retro-orbital sinus of 3-month-old male and female mice. Ten mice (5 females and 5 males) have been used per group.

Total RNA preparation and microarrays. Total RNA from cells or tissues has been prepared using RNA STAT-60 (Tel-Test, Inc., Friendswood, TX). RNAs have been processed on Mouse Expression 430A chips (Affymetrix) by the NIDDK Microarray Facility.

Reverse transcription-PCR experiments. After total RNA purification, reverse transcriptions were done using the Cells-to-cDNA II kit (Ambion, Inc.). PCR was done using the Taq DNA polymerase (0.625 unit/reaction; GeneChoice, Inc.), 0.1 μmol/L deoxynucleotide triphosphate mix, 1.5 μmol/L MgCl₂ buffer, and the following PCR program [94°C for 5 minutes, 20-30 cycles (94°C for 10 seconds, 60°C for 20 seconds, 72°C for 30 seconds), 72°C for 10 minutes]. Quantification was done after gel agarose migration and scanning using Quantity One software (Bio-Rad, Inc.).

Mouse primer information were as follows: Igf-I, 5′-TGCTGTGTAAACGACCCG-3′ (A) and 5′-CAACACTCATCCACAATGCC-3′ (B); Igf-Ir, 5′-TCTTCCCCAACCTCACAGTC-3′ (A) and 5′-GGCGTTCTTCTCAATCCTGA-3′ (B); Irs-1, 5′-AGCGAGCTCGAGCATGGCGAGCCCTC-3′ (A) and 5′-ATCGTCGACTCGAGATCTCCGAGTCA-3′ (B); Igfbp1, 5′-CCAGGGATCCAGCTGCCGTGCG-3′ (A) and 5′-GGCGTTCCACAGGATGGGCTG-3′ (B); and Igfbp2, 5′-CAACTGTGACAAGCATGGCCG-3′ (A) and 5′-CACCAGTCTCCTGCTGCTCGT-3′ (B).

Human primer information were as follows: IGF-IR, 5′-CGCCTGGAAAACTGCACG-3′ (A) and 5′-AGCTGCCCAGGCACTCCG-3′ (B) and IRS-1, 5′-AAGGCCAGCACCTTACCTCG-3′ (A) and 5′-AGCCATGGTGGCCCTGGGCA-3′ (B).

Real-time PCR experiments. Real-time PCRs were prepared with SYBR Green PCR Master Mix and done on a 7500 real-time PCR system according to the manufacturer's protocol (Applied Biosystems, Inc., Warrington, United Kingdom).

Mouse primer information were as follows: Igf-I, 5′-CTTCAACAAGCCCACAGGCTAT-3′ (A) and 5′-GCTCCGGAAGCAACACTCAT-3′ (B); Igf-Ir, 5′-GCCGACGAGTGGAGAAATC-3′ (A) and 5′-CTTGGAGATGAGCAGGATGTG-3′ (B); Irs-1, 5′-GCTCTAGTGCTTCCGTGTCC-3′ (A) and 5′-GATATAATTACTCAGCTCCTC-3′ (B); and Brca1, 5′-CGTGGGCTACCGGAACC-3′ (A) and 5′-TCTTCACTGATCTCACGATTCCA-3′ (B).

Human primer information were as follows: IRS-1, 5′-TGCACTGTGACACCAGAATAAT-3′ (A) and 5′-GTACACGTTTCAGCAGCAG-3′ (B) and BRCA1, 5′-CAGAGGACAATGGCTTCCATG-3′ (A) and 5′-CTACACTGTTCCAACACCCACTCTC-3′ (B).

Transfections, RNA interference, and Western blots. Mouse embryonic fibroblast (MEF) cells were obtained as described (7). UBR60 cells (15) were cultured in DMEM in the presence of tetracycline (5 μg/mL) to repress exogenous BRCA1 expression. For Brca1 induction, cells were washed with 1× PBS and cultured with tetracycline-free medium for 48 hours. For RNA interference (RNAi) experiments, tumor cells were seeded to 25,000 per well in six-well plate and then transfected with Irs-1 short hairpin RNA (shRNA; Open Biosystems, Inc.) according to the manufacturer's protocol. For each transfection, a mock control experiment was also done (n = 3). Cells were harvested and counted each day under microscope. Proteins were harvested on day 3, and Western blots were done according to Cao et al. (7). The following antibodies were used for Western blot analysis: IGF-IR, mitogen-activated protein kinase (MAPK), phosphorylated MAPK, AKT and phosphorylated AKT (Cell Signaling Technology, Inc.), and IRS-1 (Upstate, Inc.).

Luciferase assay. Natural killer and human UBR60 cells were cultured on fibronectin-coated six-well plates. On the following day, cells were transfected with 2 μg of a pGL2/IRS-1 construct containing a 3.4-kb fragment of the 5′-flanking region of the IRS-1 gene cloned into a pGL2 expression vector containing a luciferase gene. Renilla luciferase reporter vector (0.5 μg) was used as a transfection standard to normalize transfection efficiency. After 48 hours, the activities of the two types of luciferase were measured using the Dual-Luciferase Reporter Assay System according to the supplier's instructions (Promega).

Chromatin immunoprecipitation assay. UBR60 cells were grown on 150-mm dishes. After 24 hours, proteins were cross-linked with DNA by using 1% formalin for 10 minutes at 37°C. Cells were washed twice with ice-cold PBS, harvested, and lysed with radioimmunoprecipitation assay buffer in the presence of a mixture of protease inhibitors (Sigma). The lysates were sonicated to shear DNA to lengths between 500 and 600 bp. Proteins were incubated with 40 μL protein A+G beads (Santa Cruz Biotechnology) and 100 μg salmon sperm DNA (Stratagene) for 1 hour at 4°C with rotation to reduce nonspecific background. Immunoprecipitations were carried out overnight at 4°C with mixing by using 6 μg of Brca1, anti-acetyl histone3-K9, anti-trimethyl histone3-K9, anti-trimethyl histone3-K4, anti-trimethyl histone4-K20, anti-acetyl histone4-K16, and control IgG monoclonal antibodies. DNA-protein complexes were eluted from the beads with Tris-EDTA buffer containing 1% SDS. The cross-links were reversed by incubating the eluates for overnight at 65°C. Proteinase K (Roche) was added for 1 hour at 45°C, and the DNA was recovered by phenol/chloroform extraction and ethanol precipitation. Immunoprecipitated DNA was analyzed for the presence of the IRS-1 gene promoter sequence by PCR of 35 cycles at 95°C for 10 seconds, 60°C for 20 seconds, and 72°C for 30 seconds with a pair of primer: 5′-CTCCTCCTTCGCCTCCTC-3′ and 5′-CGTCACGTGTTTTTCTCCTC-3′. These primers amplify a fragment of 524 bp (−529 to −1,053) of IRS-1 promoter, which shares high homology with IGF-IR. Additionally, the cluster of four Sp1-binding sites (−593 to −618) in this region is required for efficient expression of the gene (16).

Statistical analysis. Statistical analysis was done with Excel software (Microsoft, Inc.).

IGF-I plasma levels are increased in Brca1^Δ11/Δ11p53^+/− mice. To study the potential relationship between IGF-I and Brca1, we measured the plasma IGF-I levels in Brca1^Δ11/Δ11p53^+/− mice and found that IGF-I is significantly increased by 18% in these mice than in control (Brca1^Δ11/+p53^+/−) mice (Fig. 1A). Although modest, this value is higher than 11%, a previously described increase, which has been correlated to a bigger risk of tumor development (17). In addition, partial inactivation or deletion of one of the IGF-IR alleles in mice leads to growth defects and life span modifications (12), indicating that subtle changes of expression or activity of one of the IGF system member can have dramatic consequences. We assume that this 18% increase observed in young mice could have significant biological effects.

Disruption of Brca1 leads to change in mRNA expression of several members of the IGF system. Despite recent debates on the role of IGF-I in normal postnatal growth and development (14, 18, 19), the liver is the major contributor to circulating IGF-I levels (20). To understand the potential effect of Brca1 on the regulation of IGF-I and its signaling axis members, we did a microarrays analysis using RNAs isolated from livers of Brca1^Δ11/Δ11p53^+/− and Brca1^+/Δ11p53^+/− mice. We chose to use 9-month-old mice for this experiment because Brca1^Δ11/Δ11p53^+/− mice at this age have started to display early onset of premature aging (7). Our data detected alterations in several Igf or insulin-related genes (Supplementary Fig. S1A). Notably, Igf-I mRNA expression increased to the same level compared with that previously observed for plasma IGF-I (21). This is concordant with the concomitant increase of growth hormone receptor mRNA. Indeed, stimulation of this receptor has been reported to favor IGF-I secretion (14). Of note, we also detected increased Irs-1 mRNA levels and decreased Igfbp1 expression in the BRCA1-mutant liver. IRS-1 plays important roles in relaying the intracellular signal from the IGF-IR, and IGFBP1 controls IGF-I bioavailability and inhibits IGF-I-dependent cellular growth and differentiation (22–25). Another significant change is the expression of Igfbp2, which was increased by 6-fold (Supplementary Fig. S1A). Next, we validated these data using reverse transcription-PCR (RT-PCR; Fig. 1B) and Western blot analysis (Fig. 1C), and we were able to confirm the changes revealed by microarrays. We also have measured quantitative expression levels of Igf-I and Irs-1 using real-time PCR on mRNA samples isolated from 3-, 9-, and 13-month-old animals (Fig. 1D). Our data indicated that the increased expression of both genes reached significant levels at all these time points.

The microarray analysis failed to detect expression of Igf-Ir, the Igf-I receptor. However, low levels of this gene were detected in previously reported studies in other mouse models (26, 27). Furthermore, recent reports have indicated that human BRCA1 could down-regulate rat IGF-IR promoter activity through interaction with the transcription factor Sp1 in cultured cells (28–30). Therefore, we decided to include Igf-Ir expression in our study; our data showed the increased expression of this gene by real-time PCR at all the developmental stages studied (Fig. 1D).

Altogether, our analyses indicated that the absence of the full-length Brca1 resulted in alterations of several IGF signaling members. However, as there is no report on Brca1 expression in liver, it raises the question whether such changes are secondary to aging instead of being a direct consequence of BRCA1 deficiency. To investigate this, we did RT-PCR analysis and found that Brca1 was indeed expressed in the liver of control (Brca1^Δ11/+p53^+/−) whereas absent in mutant (Brca1^Δ11/Δ11p53^+/−) mice (Fig. 1B). Furthermore, our analysis of Brca1-mutant and control MEF cells also confirmed the increased expression of IGF axis members in mutant cells (Supplementary Fig. S1B and C). Altogether, those results indicate that Brca1 deficiency in the liver and in MEF cells is correlated with the increased expression of IGF axis members.

Expression of BRCA1 represses expression of IGF axis members in human cells. Next, we determined whether the alteration of IGF axis members also occur in human cells. We decided to use a human BRCA1 cellular model, UBR60, whose BRCA1 transcription is controlled by a tetracycline-off system (15). On the induction of BRCA1 (48 hours after removal of tetracycline; Fig. 2A), we, respectively, detected a 2- and 3-fold decrease of IRS-1 mRNA and protein levels (Fig. 2A-C). Interestingly, the phosphorylation level of AKT, a downstream target of IGF-IR and IRS-1, was also decreased, whereas the total AKT levels were unchanged (Fig. 2C).

Next, we did an acute knockdown of BRCA1 expression using small interfering RNA (siRNA) throughout RNAi. Efficient down-regulation of BRCA1 was achieved 24 hours after siRNA transfection, and levels of IGF-IR and IRS-1 were, as a consequence, significantly up-regulated (Fig. 2D). Altogether, our data confirmed in both mouse and human cells (Figs. 1 and 2) that BRCA1 plays an important regulatory role in the expression of several IGF signaling members. We have also studied p53 expression levels and found that acute induction or suppression of BRCA1 did not affect p53 transcription (Fig. 2A and D). This observation suggests that BRCA1 affects expression of IGF signaling members independent of p53.

Increased expression of IGF axis members in normal mammary gland and mammary tumors in Brca1^Δ11/Δ11p53^+/− mice. Because Brca1^Δ11/Δ11p53^+/− mice undergo aging and tumorigenesis primarily in mammary tissue (7), we compared the expression of Igf-related members in normal mammary glands between Brca1^Δ11/Δ11p53^+/− and control mice. In 3-month-old Brca1^Δ11/Δ11p53^+/− mice, we detected a 1.8-fold increase of Igf-I mRNA levels (Fig. 3A and B), whereas expression of Irs-1 mRNA was even more increased (2.4-fold) primarily due to its very low levels in the control mice (Fig. 3A and B). Continuous increased levels of both Igf-I (52% increase) and Irs-1 (92% increase) mRNA levels were also observed when the comparison was done between 3- and 5-month-old Brca1^Δ11/Δ11p53^+/− mice (data not shown).

The influence of IGF-I on tumorigenesis prompted us to examine its levels in tumors compared with normal tissues. Our data indicated that mammary tumors exhibited increased levels of Igf-I and Irs-1 mRNA levels by 1.7- and 3.8-fold, respectively (Fig. 3C and D). Interestingly, we also observed a dramatic increase of Igf-Ir mRNA expression when we compared three independent mammary gland tumors with normal epithelium from Brca1^Δ11/Δ11p53^+/− mice (Fig. 3C and D). Because all these primary mammary tumors were p53 deficient as previously determined (31), these data are consistent with our earlier observation that activation of IGF signaling by Brca1 deficiency could occur in a p53-independent fashion.

BRCA1 interacts with IRS-1 promoter and inhibits its activity. Previous investigations showed that BRCA1 represses IGF-IR expression through the promoter of IGF-IR (31–33). IGF-IR and IRS-1 promoters share high homology in the 5′-end 374-nucleotide regulatory sequence (−1,350 to −986) with 56% identity (32). Those observations suggest that BRCA1 could be involved in IRS1 transcriptional regulation as previously shown for IGF-IR.

To assess this hypothesis, we transfected a mouse Brca1 WT mammary tumor cell line derived from a mouse mammary tumor virus (MMTV)-cNeu transgenic mice (31) with Brca1-specific shRNA. We showed that the Brca1-specific shRNA was able to knockdown Brca1 mRNA levels 48 hours after transfection as revealed by real-time PCR, and in parallel, Irs-1 mRNA levels increased (Fig. 4A). We then assessed the activity of Irs-1 promoter throughout a classic transient transfection experiment using a luciferase reporter vector (33). We showed that Irs-1 promoter activity was increased (>3-fold) on the acute suppression of Brca1 by the Brca1-specific shRNA construct but not by a green fluorescent protein (GFP)-specific shRNA construct (Fig. 4B). We further extended this study to human UBR60 cells, where the expression of BRCA1 can be regulated with tetracycline (15). Transfection of UBR60 cells in the presence or absence of tetracycline with BRCA1 siRNA was efficient to down-regulate both the endogenous and induced BRCA1 mRNA (Fig. 4C). Concomitant with the down-regulation of BRCA1, IRS-1 mRNA levels were significantly increased in these cells (Fig. 4C). IRS-1 promoter (luciferase reporter) activity showed similar up-regulation in the absence of BRCA1 (Fig. 4D).

Next, we tested whether BRCA1 was recruited to the IRS-1 promoter using chromatin immunoprecipitation (ChIP) assay. We detected that an antibody to BRCA1 could pull down the IRS-1 promoter sequence from UBR60 cells in both the presence and the absence of tetracycline (Fig. 4E). However, the intensity of the band from cells under BRCA1 induction condition is significantly stronger, whereas using an antibody to IgG failed to pull down any IRS-1 promoter sequence (Fig. 4E).

Histone modification plays an important role in controlling gene expression (34). For example, acetylation of histone H3 at lysine 9 and acetylation of histone H4 at lysine 16 are associated with transcriptional activation, whereas methylation at histone H3 at lysine 9 is associated with transcriptional repression. We next did ChIP to evaluate whether BRCA1 has an effect on histone H3 and histone H4 modification. Our data indicated that under the low level, uninduced condition of BRCA1, the IRS-1 promoter is associated with transcriptionally favorable histone modifications (high levels of histone H3-acetylated K9, histone H3-trimethylated K4, histone H4-trimethylated K20, and histone H4-acetylated K16 and low levels of histone H3-trimethylated K9; Fig. 4E), whereas in the higher level, induced condition of BRCA1, the IRS-1 promoter is associated with a transcriptional repression chromatin histone modification (Fig. 4E). These results provide evidence that increased expression of IRS-1 in BRCA1-deficient systems is directly due to BRCA1 recruitment at the promoter level and associated transcriptional down-regulation.

Cellular growth is dramatically reduced in tumors cells after Irs-1 knockdown by RNAi. The above observations raise the possibility that the IGF axis members play an important role in BRCA1-associated tumorigenesis. Indeed, Irs-1 has been associated with cell survival and proliferation (10, 11). To investigate it, we suppressed Irs-1 with shRNA in three BRCA1-deficient tumor cells (69, 525, and 780; ref. 31) to see if it could affect cellular proliferation. At the molecular level, after 3 days of transfection, IRS-1 levels were down-regulated without affecting other proteins (Fig. 5A). Significant inhibitory effects were detected in all three lines (Fig. 5B and C). Low power images of 69 cells revealed significantly decreased cell masses 3 days after transfection (Fig. 5B), whereas continuous counting of cell numbers up to 5 days revealed a persistent inhibition (Fig. 5C). More significant inhibitory effects were observed with 525 and 780 cells (Fig. 5C). In one case (780 cells), suppression of Irs-1 expression led to a complete inhibition of the cell growth. Interestingly, we observed that the two main pathways downstream of IRS-1 were also down-regulated with a dramatic reduction of AKT and MAPK phosphorylation levels (Fig. 5A), which may account for the decreased proliferation of the tumor cells.

As a control, we also treated a cell line (Ras; ref. 31) derived from MMTV-Ras transgenic mice. This cell line expresses Brca1 (data not shown). Our data indicated that suppression of Irs-1 in Ras cells resulted in a minor effect on cell growth (Fig. 5C). To understand this differential response, we measured the Irs-1 levels by real-time PCR. Our data indicated that Brca1-deficient cell lines, in general, expressed higher levels of Irs-1, especially 525 and 780 cells (Fig. 5D). Notably, the inhibitory knockdown effects are well correlated with Irs-1 expression levels. These data provide compelling evidence that overexpression of Irs-1 serves as a mitogenic force for the survival and proliferation of these cells and suggest that IRS-1 might be a good target to decrease the cellular proliferation of BRCA1-associated tumors that have high levels of IRS-1.

In this study, we have investigated the potential relationship between BRCA1 and IGF signaling in BRCA1-deficient mice, primary mammary tumors, and cultured human cells. Our data indicate that the absence of BRCA1 results in increased expression of IRS-1, IGF-IR, Igfals, Igfbp2, and Ghr and increased levels of serum IGF-I. Our analysis revealed that BRCA1 interacts with IRS-1 promoter and inhibits its activity. We further showed that BRCA1-deficient mammary cancer cells have high levels of IRS-1 expression, and acute suppression of IRS-1 by RNAi retarded the growth of these cancer cells. These observations uncover a genetic interplay between IGF-I signaling and BRCA1 during mammary tumorigenesis.

Indeed, over the last few years, investigations have shown the importance of IGF signaling in breast tumorigenesis (35, 36). The autocrine IGF-I is produced primarily in the stromal cells, which thus stimulates tumor growth in a paracrine fashion. Current experimental therapies used to treat breast cancer (tamoxifen, retinoic acid, and GH-releasing hormone analogues) have led to a decrease of circulating IGF-I levels, suggesting that the antitumor effect of these agents, at least in part, is due to reduced circulating IGF-I (35, 37, 38). It is shown that the activation of IGF-I/IGF-IR/IRS-1 leads to activation of two main proliferative pathways, the Ras-MAPK-extracellular signal-regulated kinase pathway and the phosphatidylinositol 3-kinase-AKT pathway, which stimulate cell cycle progression through cyclin D accumulation in the nucleus and inhibition of p27kip1 expression (39, 40). Stimulation of IGF-IR/IRS-1 can also destabilize adhesion and stimulate cytoskeleton reorganization through modulation of the phosphorylation signals from adhesion molecules (integrins, RACK1, focal adhesion kinase, Cas, paxillin, E-cadherin; refs. 41–45). IGF-I and IGF-II have been shown to promote endothelial cell migration and angiogenic outgrowth in vivo. In addition to the direct stimulating effects of IGFs on proliferation and migration of the endothelial cells, IGF signaling enhances transcription of vascular endothelial growth factor by hypoxia-inducible factor-1α/aryl hydrocarbon nuclear receptor translocator transcriptional complex (46).

IGF signaling also plays an important role in aging (reviewed in refs. 47, 48). It has been shown that knockout of the insulin receptor in the adipose tissue or deletion of one allele of the IGF-IR in the whole organism increases the adult life span significantly (11, 12). Our earlier observation that Brca1^Δ11/Δ11p53^+/− mice undergo both premature aging and increased tumor formation prompted us to search for a possible link between Brca1 and the IGF-I axis. Our observation that the absence of Brca1 causes increased mRNA levels of several major IGF signaling members is encouraging. However, the increase is moderate in most cases. We have selected IGF-I, Irs-1, and IGF-IR for further validation using multiple methods, including quantitative RT-PCR, acute suppression of BRCA1 by RNAi, and/or tetracycline-induced overexpression of BRCA1. These experiments are able to confirm increased expression of these genes in the absence of Brca1. Several possibilities may account for this low level of induction. For example, the expression of these genes could be subject to regulation of multiple factors, in addition to BRCA1. In this case, the effect of BRCA1 deficiency is largely compensated by the presence of other factors. It is known that regulation machinery for gene expression always exists as a complex containing many proteins (49, 50). It is also possible that the regulation of IRS-1 is achieved by multiple independent complexes, and BRCA1 is only a component of one of the complexes. Our further analysis by ChIP analysis revealed that BRCA1 interacts with the promoter of IRS-1 promoter, and this interaction is correlated with a chromatin modification toward a less active form. These observations suggest that BRCA1 may affect IRS-1 expression through changing chromatin configurations.

Interestingly, although it has been shown that p53 activation increases expression of IGF axis members (6, 10) and Brca1^Δ11/Δ11p53^+/− mice exhibit increased p53 levels primarily due to genetic instability (7), our data indicate that the regulation of BRCA1 to IGF members seems p53 independent, as our primary tumors, which are p53 deficient, also exhibited increased levels of IGF members. Moreover, our data indicate that acute induction or suppression of endogenous BRCA1 in human UBR60 cells results in alteration of IGF signaling without p53 alteration. Indeed, our data indicate that the IRS-1 promoter is a direct target of BRCA1 for chromatin modifications that negatively affect its expression. In tumor cells, we showed that IRS-1 plays a critical role for cell growth and survival. Based on these observations and also on published data (9–11), we propose the following speculative model explaining the phenotypes observed in Brca1^Δ/Δp53^+/− mice (Fig. 6). Absence of the full-length BRCA1 leads to genomic instability and subsequent p53 activation. This may result in p21 up-regulation and cell cycle arrest, leading Brca1-mutant embryos and cells to undergo dramatic senescence. Meanwhile, BRCA1 deficiency and p53 activation could both result in activation of IGF signaling pathway, AKT and MAPK activation, which should have effects on cell survival, senescence, and animal aging. Importantly, our data, showing that acute suppression of Irs-1 dramatically inhibits growth of BRCA1-deficient tumor cells, suggest a therapeutic approach for treating BRCA1-associated tumorigenesis.

Grant support: Intramural Research Program of National Institute of Diabetes, Digestive and Kidney Diseases, NIH.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank D. Haber for UBR60 cells and members of Deng laboratory for critical discussion.