Up-regulation of Skp2 after Prostate Cancer Cell Adhesion to Basement Membranes Results in BRCA2 Degradation and Cell Proliferation*

Aberrant interaction of carcinoma cells with basement membranes (BM) is a fundamental pathophysiological process that initiates a series of events resulting in cancer cell invasion and metastasis. In this report, we describe the results of our investigations pertaining to the events triggered by the adhesion of normal (PNT1A) and highly metastatic (PC-3) prostate cells onto BM proteins. Unlike PNT1A, PC-3 cells adhered avidly to Matrigel BM matrix as well as to isolated collagen type IV, laminin, and heparan sulfate proteoglycan perlecan, main BM components. This aberrantly increased cancer cell adhesion resulted in sustained BRCA2 protein depletion and vigorous cell proliferation, a cascade triggered by β1 integrin-mediated phosphatidylinositol 3-kinase activation leading to BRCA2 degradation in the proteasome. This latter effect was orchestrated by phosphatidylinositol 3-kinase-dependent up-regulation of Skp2, a subunit of the Skp1-Cul1-F-box protein ubiquitin complex that directly associates with BRCA2 as demonstrated by coimmunoprecipitation assays, determines its ubiquitination, and ultimately targets it for proteasomal degradation. Inhibition of Skp2 expression by small interference RNA prevented BRCA2 depletion and inhibited the trophic effect upon cell proliferation. These results provide additional evidence on the role of BRCA2 as a modulator of cancer cell growth and elucidate the molecular mechanisms involved in its down-regulation in cancer cells when interacting with BM, a crucial step in the biology of metastasis. Furthering the understanding of this molecular pathway may prove valuable in designing new therapeutic strategies aimed at modifying the natural history of prostate carcinoma.

Basement membranes (BM) 2 are thin layers of specialized extracellular matrix (ECM) that surround and closely associate with epithelial and endothelial cells, muscle fibers, and nerves. They consist mostly of collagen type IV (COL4) admixed with laminins (LN), nidogens, and the heparan sulfate proteoglycan perlecan (PLN) and may contain small amounts of fibronectin (FN) (1,2). Although the BM structural role in defining tissue architecture and compartmentalization has long been recognized, its dynamic role in the modulation of cell behavior has only recently been documented (1).
Aberrant cancer cell interactions with BM proteins play a crucial role in the biology of metastasis (3). Cancer cells must be able to coordinately decrease cell-cell interactions and increase cell adhesion to an adjacent BM in order to become motile, which along with the capacity of degrading/remodeling a BM directly relates to their metastatic potential (4). This cell behavior is accompanied by changes in the expression and/or usage of various adhesion receptors, including integrins (5). Integrins are transmembrane adhesion receptors for ECM proteins that not only provide physical anchoring cell support but also play a pivotal role in triggering intracellular signaling in response to environmental changes through interactions with molecules such as growth factor receptor tyrosine kinases (6,7), MAPK/ERK 1/2 (8,9), and PI 3-kinase/AKT (10,11). These various interactions help in modulating the expression of genes exerting stringent control upon cell survival, motility, and cell proliferation (12).
Intracellular protein degradation via the ubiquitin-proteasome pathway is a prime pathway through which cells normally regulate processes involved in cell growth and proliferation (13,14). There is evidence that a number of growth inhibitory molecules and tumor suppressor proteins, such as p53, p21, p27, p130, the ␤ 1C integrin, and FOXO1, are preferentially degraded by the ubiquitin-proteasome system in carcinoma cells (14 -17). Furthermore, E3 ubiquitin ligase family members Skp2 and Mdm2 have been shown to play a role in prostate cancer development and progression (18 -20). In a previous report, we provided evidence for a novel pathological mechanism whereby prostate carcinoma cell adhesion to collagen type I (COL1), a major ECM protein at osseous metastatic sites, promotes cancer cell proliferation through depletion of BRCA2 protein, the product of a tumor suppressor gene whose inactivation accounts for an increased risk in cancer development (21)(22)(23). This newly described effect resulted from ␤ 1 integrin-dependent activation of the PI 3-kinase pathway, which promoted BRCA2 ubiquitination and degradation in the proteasome (24).
In this study, we extended our investigations to elucidate the mechanisms by which ␤ 1 integrin signaling in prostate cancer cells resulted in BRCA2 protein degradation in the proteasome. We also provide evidence demonstrating that the BRCA2-associated trophic effect is not restricted to the osseous environment but is quite active in mediating cancer cell proliferation after interaction with BM proteins.

EXPERIMENTAL PROCEDURES
Cell Culture-PNT1A cells (a human prostate normal cell line established by immortalization of normal adult prostate epithelial cells) and PC-3 cells (a human prostate carcinoma cell line derived from a bone metastasis) were kept in culture as described previously (16,24).
Adhesion to the BM matrix Matrigel (Sigma) was tested in 96-well plates coated with 50 l/well of a 1:3 dilution in RPMI medium (Invitrogen) before cell plating. Cells were starved in serum-free methionine/cysteine-deficient RPMI 1640 (Sigma) for 45 min at 37°C before labeling with 100 Ci/ml 35 S proteinlabeling mix (Amersham Biosciences) in 1 ml of methionine/ cysteine-free RPMI medium containing 5% fetal bovine serum. After 24 h, 100 l of a 0.2 ϫ 10 6 cell suspension were allowed to adhere for 1 h onto Matrigel or bovine serum albumin (10 g/ml) at 37°C and were washed three times. Adherent cells were lysed in 100 l of 150 mM NaCl, 50 mM Tris-HCl, pH 7.5, and 2 mM EDTA containing 1% (v/v) Triton X-100 in phosphate-buffered saline. Radioactivity was measured in a scintillation counter (Beckman Instruments).
Inhibition assays were performed by incubating cells for 1 h on ice in the presence of either P4C10, a monoclonal antibody to human ␤ 1 integrin (ascites 1:200; Chemicon, Temecula, CA), or the monoclonal antibody 1C10 against a vascular endothelial surface protein (ascites 1:200; Invitrogen) used as a negative control. Triplicate observations were recorded for each experiment.
Analysis of BRCA2 ubiquitination was performed as described previously (24). To analyze BRCA2 association with Skp2, cell extracts were precleared and incubated overnight with 2 g of polyclonal antibody to BRCA2. Immunocomplexes were recovered with protein A-Sepharose (Sigma), washed five times with phosphate-buffered saline containing 1% Nonidet P-40, 2 mM phenylmethylsulfonyl fluoride, 10 mM sodium fluoride, 1 mM sodium orthovanadate, 5 mM sodium pyrophosphate and were subjected to 10% SDS-PAGE under reducing conditions followed by transfer to polyvinylidene difluoride membranes. Filters were immunoblotted using 1 g/ml monoclonal antibody to Skp2 (Zymed Laboratories Inc., San Francisco, CA) or 2 g/ml monoclonal antibody to BRCA2 (clone 5.23; Chemicon) following the manufacturers' instructions. Alternatively, whole cell extracts were immunoprecipitated with 2 g of monoclonal antibody to Skp2 (Zymed Laboratories Inc.) and separated by 6% SDS-PAGE, and filters were immunoblotted using 1 g/ml anti-BRCA2 polyclonal antibody or 1 g/ml polyclonal antibody to Skp2.

Transient Transfections and [ 3 H]Thymidine
Incorporation-Transient transfections with wild-type BRCA2 cDNA (a kind gift from Dr. M. C. Hung, University of Texas M. D. Anderson Cancer Center, Houston), ⌬p85, a dominant negative form of PI 3-kinase (generously provided by Dr. R. Freeman, University of Rochester, Rochester, NY), or empty vector (pcDNA3; Invitrogen) were performed as described previously (24). A pool of siRNAs for human Skp2, Mdm2, and nonspecific siR-NAs was purchased from Santa Cruz Biotechnology and used for transient transfections according to the manufacturer's instructions. Thymidine incorporation assays were performed in 96-well plates as described previously (24).
Statistical Analysis-Data are reported as the mean Ϯ S.E. Statistical analysis was performed by the Student's t test. All experiments were repeated at least twice.

RESULTS
Prostate Normal and Carcinoma Cells Adhere Differently to BM-We investigated the adhesive properties of PNT1A and PC-3 cells to LN, PLN, and COL4, major components of BM, and to FN, a widely expressed extracellular matrix protein (25) enriched at the prostatic stroma but a minor component in BM.
As shown in Fig. 1A, PNT1A cells adhered efficiently to FN, to a lesser extent to LN and PLN (ϳ60% of FN levels), and showed no adhesion to COL4. Highly metastatic PC-3 cells exhibited a reversed affinity pattern, with partial loss of adhesion to FN and newly gained adhesion capabilities to BM proteins. Because cell adhesion to FN, LN, COL4, and PLN (26) is mediated by integrin receptors that share a common ␤ 1 subunit (␣ 5 ␤ 1 to FN, ␣ 6 ␤ 1 to LN, ␣ 1 ␤ 1 , ␣ 2 ␤ 1 , and ␣ 3 ␤ 1 to COL4, ␣␤ 1 to PLN), we tested the effect of the ␤ 1 integrin-blocking antibody P4C10 on cell adhesion. As shown in Fig. 1B, adhesion to FN was reduced by 42 Ϯ 6% ( p Ͻ 0.002) and 39 Ϯ 7% ( p Ͻ 0.001) in PNT1A and PC-3 cells, respectively. Adhesion to LN was inhibited by ϳ25% in PNT1A and ϳ75% in PC-3 cells, adhesion to PLN was almost completely inhibited in both cell types, and onto COL4 PC-3 cell adhesion was almost completely inhibited (90 Ϯ 7%, p Ͻ 0.0001).
To investigate whether these findings could be reproduced onto naturally occurring BM layers, we tested the adhesive properties of normal and cancer prostate cells to Matrigel, a reconstituted BM preparation (27). As shown in Fig. 1C, whereas PC-3 cells adhered efficiently to Matrigel in a manner that was significantly inhibited by the anti-␤ 1 integrin-blocking antibody (ϳ70% inhibition, p Ͻ 0.004), normal prostate cell adhesion was nil in the time period considered.
Adhesion of PC-3 Cells to BM Proteins Decreases BRCA2 Protein Levels in a ␤ 1 Integrin-dependent Manner-We asked whether PC-3 cell adhesion to BM proteins exercised a modu-latory effect upon BRCA2 expression. To this effect, PNT1A and PC-3 cells were grown onto plates coated with FN, LN, PLN, COL4, or Matrigel and BRCA2 levels were assessed. As depicted in Fig. 2A, whereas PNT1A cells transiently increased BRCA2 protein after adhesion to BM proteins (2.2 Ϯ 0.3-fold onto FN after 2.5 h, p Ͻ 0.002), PC-3 cells exhibited almost complete (PLN) or complete (LN, COL4, Matrigel) loss of detectable BRCA2 after adhesion for 6 h. The strongest downregulatory effect was observed onto COL4, which caused BRCA2 protein levels to decrease after only 2.5 h by 68 Ϯ 4% ( p Ͻ 0.01). BRCA2 protein levels did not recover after 12 h of cell adhesion (data not shown). The ␤ 1 integrin-blocking antibody P4C10 partly rescued BRCA2 protein to 45 Ϯ 6% ( p Ͻ 0.001) after 6 h onto COL4 (Fig. 2B) as well as onto other BM proteins (data not shown). As COL4 is the most abundant component of the BM, most of the subsequent experiments were performed onto this BM protein.
BRCA2 Protein Depletion after Cancer Cell Adhesion to BM Increases DNA Synthesis-To investigate whether the signaling cascade initiated by BM proteins had any effect on cell proliferation, we measured DNA incorporation of [ 3 H]thymidine in PNT1A and PC-3 cells onto various substrates (Fig. 3A). Whereas PNT1A cell proliferation showed no variations irrespective of the adhesive substrate, adhesion to COL4 enhanced PC-3 cell proliferation compared with plastic by 291 Ϯ 24% ( p Ͻ 0.001). A somewhat less strong proliferative response was elicited upon adhesion onto LN (ϳ213%) and PLN (ϳ175%). No significant change was detected onto FN (data not shown). Preincubation of PC-3 cells with P4C10 inhibited the response to COL4 by ϳ31% and to LN and PLN by ϳ35% (data not shown). These results were further replicated and confirmed onto native BM tissue. As shown in Fig. 3B, adhesion to Matrigel enhanced cell proliferation by 509 Ϯ 27% and P4C10 inhibited the response by 32 Ϯ 5% ( p Ͻ 0.0002). The increase in DNA synthesis after adhesion to BM proteins could be reversed by transfecting PC-3 cells with wild-type BRCA2 cDNA (Fig.  3C). In these experiments, PC-3 cells were transiently transfected with BRCA2 cDNA or empty vector for 36 h, after which we measured BRCA2 protein levels (upper panel) and [ 3 H]thymidine incorporation upon 12 h of cell adhesion to COL4 in the presence or absence of the proteasome inhibitor MG132 (lower panel). After transfection, BRCA2 protein levels increased 2.3fold compared with mock-transfected cells. Upon COL4 adhesion, BRCA2 decreased by 44 Ϯ 8% ( p Ͻ 0.03) in transfected cells, but this reduction had no effect in [ 3 H]thymidine incorporation, which remained at basal levels throughout the experiment. On the contrary, mock-transfected cells exhibited complete disappearance of BRCA2 protein upon adhesion to COL4, which resulted in a burst in [ 3 H]thymidine incorporation. This trophic effect could be reversed with the proteasome inhibitor MG132, which decreased [ 3 H]thymidine incorporation by 94% ( p Ͻ 0.003). Treatment of BRCA2-transfected cells with MG132 resulted in BRCA2 accumulation to 245 Ϯ 29% ( p Ͻ 0.001).
Skp2 Promotes BRCA2 Protein Depletion in PC-3 Cells Adherent to BM-Because it is known that F-box proteins Skp2 and Mdm2 mediate the ubiquitin-dependent degradation of several negative regulators of cell proliferation in cancer (17, 28, A, 1.5 ϫ 10 5 PNT1A or PC-3 cells were allowed to adhere to bovine serum albumin (BSA) (10 g/ml), FN (3 g/ml), LN (10 g/ml), PLN (10 g/ml), or COL4 (10 g/ml) at 37°C for 1 h and were washed and stained with crystal violet. Cell adhesion was quantified by measuring the absorbance at 630 nm in an enzyme-linked immunosorbent assay reader. B, 1.5 ϫ 10 5 PNT1A or PC-3 cells were incubated for 1 h on ice in the presence of the ␤ 1 integrin-blocking antibody P4C10 or 1C10 as control and allowed to adhere to bovine serum albumin (BSA), FN, LN, PLN, or COL4 as described in panel A. C, 2 ϫ 10 5 35 Slabeled PNT1A or PC-3 cells were incubated for 1 h on ice in the presence or absence of P4C10 or 1C10 before being allowed to adhere to Matrigel (ϳ0.2 mg of protein/well) for 1 h at 37°C. Following three phosphate-buffered saline washes, cells were lysed and the radioactivity was measured. Data are expressed as mean Ϯ S.E. of triplicate wells. A representative experiment of three is shown. 29), we hypothesized that they were likely candidates to be involved in BRCA2 ubiquitination and proteasomal degradation in prostate cancer cells. We first investigated whether PC-3 cell adhesion to COL4 had any effect upon Skp2 or Mdm2 protein levels. As depicted in Fig. 4A, PC-3 cell adhesion to COL4 triggered Skp2 and Mdm2 protein expression levels by 281 and 184% at 2.5 h and 393 and 203% at 6 h, respectively. This upregulatory effect was significantly inhibited by a ␤ 1 integrinblocking antibody for Skp2 but not Mdm2. We then proceeded to investigate whether manipulation of Skp2 or Mdm2 levels had an effect upon BRCA2 protein levels and DNA synthesis. As shown in Fig. 4B, Skp2 knock down by siRNAs resulted in Skp2 protein cell depletion (upper panel) and concomitant rescue of BRCA2 protein levels after adhesion to COL4 (lower panel). Furthermore, these newly induced changes blunted much of the new DNA synthesis upon PC-3 cell adhesion to COL4 (Fig. 4C). BRCA2 protein depletion was independent from any increase in Mdm2 protein levels upon adhesion to COL4, and Mdm2 knock down by siRNAs did not rescue BRCA2 protein levels (data not shown).
Skp2 Directly Interacts with BRCA2, Promoting Its Ubiquitination-Next we examined whether changes in Skp2 expression had any effect upon BRCA2 ubiquitination in PC-3 cells after COL4 adhesion. To this effect, cell extracts from PC-3 cells transfected with Skp2 siRNAs or mock transfectants were prepared at various time points and immunoprecipitated with anti-BRCA2 antibody, followed by Western blotting with an antibody against ubiquitin (Fig. 5A). Before adhesion (0 h), PC-3 cells showed minimal BRCA2 ubiquitination, as previously described (24). After 2.5 h of adhesion to COL4, cells transfected with nonspecific siRNAs exhibited steady increase in BRCA2 ubiquitination at a pace that mirrored the reduction in BRCA2 protein levels shown in Fig. 2A. In contrast, knock down of Skp2 expression resulted in a dramatic reduction in BRCA2 ubiquitination at all time points. This effect results from a direct interaction between the two proteins as demonstrated by coimmunoprecipitation assays. In the experiment shown in Fig. 5B, after COL4 adhesion amounts of Skp2-BRCA2 complex immunoprecipitated with anti-BRCA2 sharply increased from detectable levels at 0 h to maximal amounts at 2.5 h to subsequently decrease after 6 h, reflecting virtual BRCA2 protein depletion. This behavior mirrored the changes in Skp2 depicted in Fig.  4A. BRCA2 was also detected when the anti-Skp2 antibody was used for immunoprecipitation (data not shown).

PC-3 Cell Adhesion to BM Triggers PI 3-Kinase Activation in
a ␤ 1 Integrin-dependent Manner-We also investigated the involvement of the MAPK/ERK and the PI 3-kinase signaling pathways in modulating BRCA2 protein depletion upon cancer cell adhesion to BM. In Western blotting with anti-phospho-ERK, MAPK/ERK activity in PC-3 cells was nil at rest and did not increase after adhesion to COL4 (Fig. 6A). On the contrary, PI 3-kinase activity increased by 2-fold, remaining highly phosphorylated for as long as 6 h as measured by Western blotting with anti-phospho-Ser-473 AKT. This response was dependent on ␤ 1 integrin as pretreatment with the blocking antibody P4C10 resulted in ϳ80% inhibition in AKT phosphorylation onto COL4 (Fig. 6B). AKT phosphorylation affected BRCA2 levels as demonstrated by inhibition of PI 3-kinase activity with wortmannin ( Fig. 6C) or LY294002 (data not shown), which increased BRCA2 protein by 4.3-fold after 2.5 h of adhesion to COL4.
Degradation of BRCA2 by Skp2 Requires PI 3-Kinase-To investigate whether the increase in Skp2 protein levels leading to BRCA2 ubiquitination after PC-3 cell adhesion to COL4 results from aberrant PI 3-kinase activation, we transiently transfected PC-3 cells with a dominant negative (⌬p85) form of the PI 3-kinase before adhesion to COL4 and analyzed protein levels of BRCA2, Skp2, and Skp2-BRCA2 complex and [ 3 H]thymidine incorporation. As depicted in Fig. 7, transfection with ⌬p85 completely prevented BRCA2 protein degradation and Skp2 protein up-regulation (panel A) as well as the increase in

DISCUSSION
The ability of carcinoma in situ to overcome the surrounding epithelial BM and spread into the underlying stroma has long defined the progression to invasive carcinoma (30). In turn, for local and distant metastasis to occur, locally invasive carcinoma cells must be able to colocalize and eventually invade muscle, nerves, and blood vessels, all of which are separated from the adjacent stroma by continuous sheets of BM. Hence, a fundamentally abnormal interaction of carcinoma cells with BM proteins must be at the core of their metastatic capabilities. Our experimental paradigm confirmed that unlike normal PNT1A, highly invasive PC-3 carcinoma cells exhibited inappropriate adhesion onto the main components of the BM as well as on a reconstituted BM matrix system. On the other hand, the ECM glycoprotein FN enriched at the prostatic stroma and a minor component of BM was a poor adhesive substrate for neoplastic  PC-3 cells when compared with normal PNT1A cells. This latter observation could be partly correlated with the lower expression of the FN receptor ␤ 3 integrin in PC-3 3 when compared with PNT1A cells.
Another crucial finding is that after adhesion to isolated BM proteins, particularly to COL4, highly invasive carcinoma cells exhibit a burst in [ 3 H]thymidine incorporation, a sensitive measurement of new DNA synthesis and cell proliferation. This trophic response seems to be elicited by ␤ 1 integrin adhesion, as a similar phenomenon was previously demonstrated onto COL1, a ligand for ␤ 1 integrin receptors (24). The proliferative response after adhesion to LN and PLN was slightly weaker than with COL4, and no effect was noticeable onto FN. When allowed to adhere onto the BM matrix Matrigel, prostate cancer cells exhibited a proliferation surge that was stronger than with isolated components, suggesting cooperation rather than competition among various proteins in activating proliferative signals.
The abnormal proliferative response following PC-3 cell adhesion onto BM was linked to a sustained BRCA2 protein depletion and could be completely averted by wild-type BRCA2 cDNA transfection. This pathway seems to switch cell proliferation on and off depending on a critical amount of BRCA2 protein, as suggested by the unresponsiveness to mild reductions in BRCA2 protein after PC-3 cell adhesion to FN. Conversely, normal cell adhesion to FN resulted in an increase in BRCA2 expression with no discernible effect upon DNA syn-3 L. Moro, unpublished data. FIGURE 5. Skp2 directly interacts with BRCA2, promoting its ubiquitination. A, PC-3 cells were transfected with a pool of siRNAs for Skp2 or nonspecific siRNAs (Ϫ); 36 h after transfection cells were allowed to adhere to COL4 for the indicated times and lysed, and protein extracts were immunoprecipitated (IP) with a BRCA2 antibody. Immunoprecipitates were analyzed by immunoblotting (IB) with anti-ubiquitin and anti-BRCA2 antibody. Blots are representative of two independent experiments. B, PC-3 cells were allowed to adhere to COL4 for the indicated times and lysed, and protein extracts were immunoprecipitated with anti-BRCA2 antibody and immunoblotted with antibodies for Skp2 and BRCA2. Blots are representative of three independent experiments. FIGURE 6. PC-3 cell adhesion to COL4 triggers PI 3-kinase activation in a ␤ 1 integrin-dependent manner. A, PC-3 cells were allowed to attach to COL4 at 37°C for 0 -6 h, washed and lysed, and 100 g of cell protein extracts were electrophoresed onto 10% SDS-polyacrylamide gel under reducing conditions. Membranes were probed with anti-phospho-Ser-473 AKT (P-AKT) or anti-phospho-ERK (P-ERK). Subsequently, membranes were stripped and blotted with anti-AKT 1/2 (AKT 1/2) or anti-ERK2 (ERK2) as control. The blots are representative of three independent experiments. B, PC-3 cells were preincubated with the ␤ 1 integrin-blocking antibody P4C10 or 1C10 as control before plating onto COL4 and being processed as described in panel A. The blots are representative of two independent experiments. C, PC-3 cells were pretreated for 1 h at 37°C with 0.1 M wortmannin (WM), a PI 3-kinase inhibitor, or the solvent alone (Me 2 SO). Cells were then allowed to adhere to COL4 at 37°C for 0 or 2.5 h and were washed, lysed, and processed for immunoblotting. Membranes were probed with anti-BRCA2 antibody and ␤-tubulin as loading control. The blots are representative of two independent experiments. Bottom, BRCA2 protein levels were quantitated and reported as percentage of the amount of protein at 0 h, set to represent 100%. Data are expressed as mean Ϯ S.E. of two independent experiments. thesis. Our findings add to an increasing body of evidence suggesting that BRCA2 also functions as modulator of cell proliferation and tumor growth. A previous report showed that recombinant BRCA2 overexpression in the BRCA2-deficient pancreatic adenocarcinoma cell line Capan-1 resulted in substantial decrease in tumorigenicity when inoculated into nude mice (31). More recently, Miyamoto et al. (32) observed that Capan-1 cells exhibited an increased proliferation after adhesion to COL4. The mechanisms involved are not yet well understood. Normally, part of the cellular response to DNA damage involves the activation of an ATM/p53/Mdm2 feedback loop that regulates cell cycle progression and/or apoptosis in response to relative amounts of a DNA repair complex containing BRCA2 (33)(34)(35). However, involvement of this mechanism seems unlikely because our experiments demonstrate no relationship between ␤ 1 integrin signaling and Mdm2 levels and suggest an alternative pathway involving Skp2 (see below). Furthermore, PC-3 (and Capan-1) cells exhibit inactivating mutations in the p53 gene (36,37).
Depletion of BRCA2 protein upon cancer cell adhesion to BM proteins and osseous COL1 (24) occurs through protein degradation in the proteasome. However, the precise mechanisms involved had not been elucidated before. In the present study, we have provided evidence that this event is mediated by Skp2, an F-box protein that associates with Skp1, Cul1, and Roc1/Rbx1 to form the SCF(Skp2) ubiquitin ligase complex (28). Indeed, down-regulation of Skp2 by siRNA was sufficient to rescue BRCA2 levels inhibiting the burst in cell proliferation upon cancer cell adhesion to BM. Evidence that this mechanism may be relevant in prostate cancer has also been provided in immunohistochemical studies performed on 622 radical prostatectomy specimens that demonstrate that Skp2 levels and cell-labeling frequency increase dramatically in both premalignant prostatic intraepithelial lesions and prostate carcinoma (19). Changes in Skp2 levels are dependent upon ␤ 1 integrin-mediated signaling for Skp2 up-regulation, and subsequent BRCA2 degradation could be prevented by a ␤ 1 integrin-blocking antibody. Evidence demonstrating that cell adhesion to the ECM results in Skp2 mRNA and protein up-regulation had been provided previously (38). We confirm those results and demonstrate for the first time that ␤ 1 integrin-mediated signaling is necessary for ECM-dependent changes in Skp2 expression.
The up-regulation of Skp2 that follows ␤ 1 integrin adhesion is signaled through PI 3-kinase/AKT and could be abrogated by transfecting a dominant negative form of PI 3-kinase. These observations are in agreement with recent reports implicating PI 3-kinase in Skp2-dependent degradation of p27 kip1 and FOXO1 (17,18,39,40). The involvement of PI 3-kinase/AKT in BRCA2 protein depletion was not surprising, as activation of this pathway has been consistently invoked in studies on cancer cell proliferation and survival (41)(42)(43)(44). Recently, compelling confirmatory evidence for increased PI 3-kinase/AKT phosphorylation in prostate intraepithelial neoplasia and at the invasive edge of prostate carcinoma has been obtained by reverse-phase protein microarrays combined with laser microdissection (45).
Overall, our studies have demonstrated that Skp2 directly interacts with and promotes the degradation of BRCA2. As depicted in Fig. 8, this process requires PI 3-kinase/AKT activation initiated by ␤ 1 integrin-mediated signaling following cancer cell interaction with BM proteins, a crucial physiopathological phenomenon at the beginning of the metastatic cascade to local and distant organs. Furthering the understanding of this molecular pathway may prove valuable in designing new therapeutic strategies aimed at modifying the natural history of prostate carcinoma.