Human Phosphatidylethanolamine-binding Protein 4 Promotes Transactivation of Estrogen Receptor α (ERα) in Human Cancer Cells by Inhibiting Proteasome-dependent ERα Degradation via Association with Src*

We identified human phosphatidylethanolamine-binding protein 4 (hPEBP4) as a human-derived novel member of the phosphatidylethanolamine-binding protein family, which is involved in apoptosis resistance of tumor cells. Because of its preferential expression in estrogen-related cancers, we wondered whether hPEBP4 plays a role in estrogen-induced cancer cell growth. Here, we demonstrated that hPEBP4 inhibited the 17β-estradiol (E2)-induced, proteasome-dependent estrogen receptor α (ERα) degradation to increase the protein level of ERα. Silencing of hPEBP4 inhibited the recruitment of ERα to the promoter of the ERα target gene pS2 in MCF-7 breast cancer cells after E2 treatment. E2-induced, ERα-mediated transcription via the estrogen-response element, as well as the cellular proliferation, was significantly suppressed in hPEBP4-silenced MCF-7 cells. We found that Src, whose association with ERα facilitates the ERα binding to components of proteolytic machinery, could associate with hPEBP4 and that overexpression of hPEBP4 prevented the E2-induced interaction between ERα and Src. ERα overexpression, proteasome inhibitor, or Src inhibitor could reverse the suppression of ERα-mediated transactivation by hPEBP4 silencing. The inhibition of the proteasome degradation and the promotion of transactivation of ERα by hPEBP4 via the Src pathway were further confirmed in HeLa cells. Finally, we found that the promoting effects of hPEBP4 on ERα-mediated transactivation and estrogen-induced proliferation of cancer cells did not depend on its regulation of Akt and ERK activity. Our data suggest that hPEBP4 inhibits proteasome-dependent ERα degradation through the Src pathway, thus enhancing ERα-mediated transactivation and promoting the proliferation of cancer cells in response to estrogen.

the pathogenesis of at least 40% of breast cancer in women (2). The effects of estrogen are largely exerted through the activation of ER␣, 3 the classic estrogen receptor that belongs to the steroid/thyroid nuclear receptor superfamily of ligand-regulated transcription factors (3). Binding of estrogens to ER␣ leads to transcription of a wide range of genes that stimulate proliferation of mammary cells. The increase in cell division and DNA synthesis then elevated the risk for replication errors, which may result in the acquisition of detrimental mutations that disrupt normal cellular processes such as apoptosis, cellular proliferation, or DNA repair (4). Because of the pivotal role of the ER␣ axis in breast cancer development, targeting ER or its ligands is a major strategy for breast cancer treatment (5). However, since the initial discovery and characterization of the ER␣ in the 1960s, mechanisms of ER signaling have become more complicated. Increasing evidence has revealed the participation of various coregulators in the binding process of ER␣ with the estrogen-response element (ERE) and demonstrated a membrane-located ER␣-mediated "nongenomic" signal pathway (6 -9). Elucidation of those molecular mechanisms and their interaction with the classical ER␣ signal pathway has become a major challenge in breast cancer research and would be of great significance in overcoming antiestrogen resistance in clinical therapy.
hPEBP4 (human phosphatidylethanolamine-binding protein 4) is a novel antiapoptotic protein identified by our laboratory in 2004 (10). Normally colocalizing with lysosome, hPEBP4 translocates to the cell membrane upon TNF-␣ stimulation, where it binds to Raf-1 and MEK1 (10). hPEBP4 appears to promote cellular resistance to TNF-␣/tumor necrosis factorrelated apoptosis-inducing ligand-induced apoptosis by inhibiting activation of JNK (c-Jun N-terminal kinase) and the Raf-1/MEK/ERK pathway (11)(12)(13). However, it is almost impossible to imagine that TNF-␣ or tumor necrosis factor-related apoptosis-inducing ligand can accumulate to a concentration high enough to kill tumor cells in vivo, let alone the determined association of TNF-␣ for patients with a poor prognosis of tumors. Therefore, the exact role of hPEBP4 in carcinogenesis still remains unknown. Interestingly, we find that all types of tumors, including breast and ovarian tumors, in which hPEBP4 is endogenously overexpressed, have been believed to associate with estrogen during their development. Therefore, we speculate that hPEBP4 may be involved in the responsiveness of cancer cells to estrogen in the tumor microenvironment.
Cellular Src (c-Src), the 60-kDa tyrosine kinase, regulates cellular proliferation and tumor metastasis (14). Recent studies have shown that Src promotes cell proliferation by driving ER␣ transcriptional activity and targeting ER␣ for ubiquitin-dependent proteolysis in human breast cancers (15). The cross-talk between Src and liganded ER␣ may affect the therapeutic efficacy and prognosis in patients with breast cancer.
In this study, we focused on the role of hPEBP4 in estrogeninduced ER␣ degradation in human cancer cells. To do this, RNA interference was used to reduce hPEBP4 expression in MCF-7 breast cancer cells, and the effects of hPEBP4 on ER␣ degradation and transcriptional activity were investigated. Our results indicated that hPEBP4 inhibits proteasome-dependent degradation of ER␣ via the Src pathway and thus promotes ER␣-mediated transactivation. This phenomenon is also further confirmed in HeLa cells that are forced to express hPEBP4. The inhibitory effects of hPEBP4 may be attributed to its association with Src, which may disrupt the interaction between Src and ER␣. We also found that the promoting effects of hPEBP4 on ER␣-mediated transactivation and estrogen-induced proliferation of cancer cells are independent of its roles in regulating ERK or Akt kinase activity. These results suggest that hPEBP4 may represent a novel regulatory mechanism of the ER␣ signal pathway in cancer cells, which offers a new therapeutic target for hormone-responsive cancers.
Cell Culture-Human breast cancer cell line MCF-7 and human cervix cancer cell line HeLa were obtained from American Type Culture Collection (ATCC) and grown in phenol red-free DMEM supplemented with 5% DCC-FCS (FCS treated with dextran-coated charcoal) (Hyclone), 4.5 g/liter D-glucose, nonessential amino acids (100 mol/liter each), 100 units/ml penicillin, 100 g/ml streptomycin, and 2 mmol/liter glutamine at 37°C in a 5% CO 2 atmosphere.
RNA Interference of hPEBP4-hPEBP4 transient RNA interference with chemically synthesized siRNA duplex (GeneChem, Shanghai, China) and mutation control was performed using Lipofectamine 2000 reagent (Invitrogen) as described previously (10). Efficacy of interference was determined by Western blot 48 h after transient transfection of siRNAs.
Assay for Luciferase Reporter Activity-Cells were seeded at 5 ϫ 10 4 cells/well in 24-well plates 16 h prior to transfection. Each well was transfected with 3ϫERE-TATA-Luc plasmid and pGL-TK control construct and treated as indicated. Luciferase activity was measured using the Dual-Luciferase reporter assay kit (Promega). Data were normalized for transfection efficiency by dividing firefly luciferase activity with that of Renilla luciferase (18,19).
Real Time PCR-Cells were grown for 48 h in phenol red-free DMEM supplemented with 5% DCC-FCS serum after RNA interference and treated with E 2 at a final concentration of 10 nM. RNA was collected with TRIzol (Invitrogen) as described previously (20,21). cDNA was synthesized with 1 g of total RNA using first strand cDNA synthesis kit-ReverTra Ace-␣ (Toyobo) at 42 and 99°C for 20 and 5 min, respectively. Quantification of inducible expression of pS2 was performed using SYBR master mix (Toyobo) on an iCycler fitted with an optical assembly unit (Bio-Rad). The relative expression of pS2 mRNA was calculated using the 2 Ϫ⌬⌬Ct method with ␤-actin mRNA as an internal control (22). The primers used were as follows: 5Ј-ATACCATCGACGTCCCTCCA-3Ј (forward) and 5Ј-AAGCGTGTCTGAGGTGTCCG-3Ј (reverse) for pS2; 5Ј-TGCGTGACATTAAGGAGAAG-3Ј (forward) and 5Ј-GCTCGTAGCTCTTCTCCA-3Ј (reverse) for ␤-actin.
Chromatin Immunoprecipitation-MCF-7 cells were subjected to RNA interference and grown for 48 h in phenol redfree DMEM supplemented with 5% DCC-FCS. Cells were then treated with E 2 at 10 nM for the indicated time and then fixed with formaldehyde. Cells were lysed and sonicated, and 10 l of purified chromatin samples was used as inputs, and the remaining amount was subjected to the chromatin immunoprecipitation procedure. Purified chromatin samples were immunoprecipitated with 1 g of anti-ER␣ antibody or antibodies against AIB-1, SRC-1, GRIP1. The immune complexes were recovered using protein A/G-Sepharose (Calbiochem) and processed as described (23). The immunoprecipitated DNA was amplified by PCR with the pS2 promoter-specific primer, and the pS2 upstream primer pair was used as negative control (24).
Western Blot and CHX Chase-A BCA protein assay reagent kit (Pierce) was used to measure protein concentration. Sam-ples containing equal amounts of protein were prepared, separated by 12% SDS-PAGE, and transferred to Protran nitrocellulose membranes. The blots were probed with specific antibodies as indicated with appropriate horseradish peroxidase-conjugated antibodies as secondary antibodies. Supersignal West Femto maximum sensitivity substrate (Pierce) was used for the chemiluminescent visualization of membranebound proteins. The ER␣ t1 ⁄ 2 was determined by CHX chase, with addition of 100 g of CHX considered t ϭ 0. Cells were lysed at the times as indicated, and equal samples were analyzed by immunoblotting.
Detection of ER␣ Ubiquitylation-MCF-7 cells were starved in 0.1% DCC-FCS for 48 h and then treated 5% DCC-FCS plus 10 nM E 2 for 6 h with or without prior addition of Src inhibitor PP1. Total cell lysates were collected, and ER␣ was determined by Western blot. To analyze ubiquitylated ER␣, ER␣ was immunoprecipitated, and precipitates were blotted with antibody against ubiquitin (Abcam).
Coimmunoprecipitation-Cellular lysates of stably transfected HeLa cells were precleared with protein A-Sepharose beads (Calbiochem), and immunoprecipitation was performed as described previously (11,20) using ER␣ (Santa Cruz Biotechnology), Src, and Myc (Cell Signaling Technology) antibodies cross-linked to protein-A-Sepharose beads. Resulting elution was subjected to Western blot to detect coimmunoprecipitated proteins.
BrdUrd Incorporation Assay-The protocol was performed as described previously (18,19). Briefly, MCF-7 cells were seeded in 6-well plate, stimulated with 10 nM E 2 and incubated with 10 M BrdUrd (Calbiochem) for 2 h before collection. The cells were then fixed, permeabilized, incubated with fluorescein isothiocyanateconjugated antibody, and subjected to fluorescence-activated cell sorting analysis.
Statistical Analysis-The Student's t test was used to determine the statistical significance of the data obtained and to compare the means between groups. A p value of less than 0.05 represented a statistically significant difference.

RESULTS
Silencing of hPEBP4 Attenuates the Expression of ER␣ by Increasing ER␣ Degradation-We first determined the effect of hPEBP4 downregulation on ER␣ levels in MCF-7 breast cancer cells upon E 2 stimulation. The result confirmed that siRNA could efficiently reduce hPEBP4 protein levels in MCF-7 cells (Fig. 1A). The cells were then grown in phenol red-free DMEM with 5% DCC-FCS for 2 days before treating with E 2 or ethyl alcohol as a vehicle control. As shown in Fig. 1B, E 2 stimulation elicited an up-regulation of ER␣ protein level, which is in agreement that reported previously (9). Targeted reduction of hPEBP4 with siRNA dramatically inhibited the increase of ER␣ expression as compared with siRNA mutation control (Fig. 1B). Additionally, hPEBP4 down-regulation-mediated depletion of ER␣ was associated with attenuation of the progrowth cyclin D1 protein (Fig. 1B), a downstream target of ER␣. The changes of ER␣ protein level were independent of mRNA, which did not show any concomitant change within the duration of E 2 stimulation (supplemental Fig. S1), suggesting that hPEBP4 may affect the ER␣ protein expression at the posttranscriptional level. As it has been determined that proteasome-mediated degradation acts as an important mechanism in fine-tuning the ER␣-initiated signal pathway (25), we used MG-132, a specific proteasome inhibitor, to test whether the effect of hPEBP4 on ER␣ protein level was caused by affecting proteasome-dependent proteolysis activity. The results showed that inhibiting proteasome activity with MG-132 led to an accumulation of the ER␣ protein and that MG-132 treatment almost completely abolished the reduction of ER␣ caused by hPEBP4 silence (Fig. 1C). In support of this, we examined the effect of hPEBP4 on ER␣ stability by the CHX chase assay. As shown in Fig. 1D, a decrease in ER␣ stability was found in hPEBP4-silenced cells when compared with that of mutation control cells (Fig. 1D, top). A more significant effect of hPEBP4 on ER␣ stability was observed in the presence of E 2 (Fig. 1D,  bottom).
We next performed rescue experiments to further confirm the effect of hPEBP4 silencing on ER␣ degradation. The siRNA-resistant hPEBP4 expression vector (hPEBP4-R) was constructed and transfected into hPEBP4 siRNA-transfected FIGURE 1. Silencing of hPEBP4 increases ER␣ degradation to repress ER␣ expression. A and B, MCF-7 cells were transfected with hPEBP4-specific siRNA (hPEBP4 siRNA) or mutated hPEBP4 siRNA (mutation control) and cultured in the absence of E 2 for 48 h. Then the cells were treated without or with E 2 (10 nM) for the indicated time and subjected to Western blot analysis using indicated antibodies. C, hPEBP4-silenced MCF-7 cells were cultured in the absence of E 2 for 48 h and treated with MG-132 (10 M) or DMSO 1 h prior to 10 nM E 2 stimulation for 4 h. Western blot was then performed with anti-ER␣ antibody. D, hPEBP4-silenced MCF-7 cells were cultured in the absence of E 2 for 48 h. ER␣ t1 ⁄2 was assayed by CHX chase at 2, 4, and 6 h after treatment with or without E 2 . E, hPEBP4-silenced MCF-7 cells were transfected with siRNA-resistant hPEBP4 expression vector (hPEBP4-R) or mock vector, cultured for 24 h, and then treated with E 2 (10 nM) for 6 h. Western blot was performed with total cell lysates using indicated antibodies.

hPEBP4 Inhibits ER␣ Degradation via Src
MCF-7 cells followed by E 2 treatment, and the protein levels of ER␣ and cyclin D1 were then determined by Western blot. As shown in Fig. 1E, the decrease of ER␣ and cyclin D1 levels induced by hPEBP4 silencing could be completely reversed by re-expression of siRNA-resistant hPEBP4. These data suggested that silencing of hPEBP4 attenuates the expression of ER␣ by increasing ER␣ degradation.
Silencing of hPEBP4 Suppresses Transcriptional Activity of ER␣ and Cellular Proliferation-ER␣-mediated transcriptional activation is the most important mechanism by which estrogen exerts its carcinogenic effect (7). To investigate the hPEBP4 effect on ER␣ signaling, we examined ER␣-mediated transactivation through ERE in the hPEBP4-silenced MCF-7 cells, which were transfected with the vectors containing a luciferase reporter construct of ERE and the control pGL-TK reporter. Treatment with 10 nM E 2 showed a marked increase in luciferase transcription and activity, which was significantly suppressed by hPEBP4 silencing (Fig. 2A). The E 2 -induced expression of pS2, the endogenous target gene transcriptionally activated by ER␣, was also determined, and the results showed that the pS2 mRNA level was down-regulated in the hPEBP4-silenced MCF-7 cells (Fig. 2B). The amount of ER␣ as well as its cofactors recruited to estrogen-responsive promoter of pS2 was also examined by chromatin immunoprecipitation. As shown in Fig. 2C, we found that E 2 stimulation resulted in a dramatic increase of ER␣ binding with the estrogenresponsive promoter, and silencing of hPEBP4 significantly reduced the amount of ER␣ as well as its cofactors (SRC-1, AIB-1, and GRIP1) bound to the estrogen-responsive promoter of pS2 after E 2 treatment. We went further to evaluate the effect of hPEBP4 on E 2 -induced cancer cell proliferation using BrdUrd incorporation assay. We found that the percentage of proliferative MCF-7 cells was markedly reduced after hPEBP4 was silenced (Fig. 2D). In addition, to determine whether it is ER␣ degradation that causes the difference in cancer cell proliferation, we transfected MCF-7 cells with ER␣ plasmid. We found that forced expression of ER␣ completely reversed the decreased proliferation of hPEBP4-silenced MCF-7 cells, suggesting that hPEBP4 promotes E 2 -induced proliferation of cancer cells as a consequence of inhibiting ER␣ degradation (Fig. 2D). Therefore, depletion of hPEBP4 suppressed both ER␣-mediated transcriptional activity and cellular proliferation induced by E 2 .
To confirm that hPEBP4 does promote ER␣-mediated transactivation by inhibiting proteasome, we determined the effect of MG-132 on ER␣-mediated transactivation using reporter gene assay. The result showed that inhibiting proteasome activity with MG-132 reversed the decrease in transcriptional activity of hPEBP4-silenced cells (Fig. 2E). When MCF-7 cells were transiently transferred with pSG5-ER␣ plasmid after silencing of hPEBP4 with siRNA, overexpression of ER␣ completely reversed the inhibition of ER␣-mediated transactivation caused by hPEBP4 silence (Fig. 2E). Moreover, hPEBP4 silencing-induced down-regulation of ER␣-mediated transactivation could also be restored by siRNA-resistant hPEBP4 re-expression (Fig.   FIGURE 2. Silencing of hPEBP4 decreases ER␣-mediated transactivation and cellular proliferation. A, hPEBP4-silenced MCF-7 cells were transfected with luciferase reporter plasmids of ERE and pGL-TK vector, cultured for 8 h, and then stimulated with E 2 (10 nM) for 16 h. The luciferase activity was measured using Dual-Luciferase reporter assay system. B, hPEBP4-silenced MCF-7 cells were cultured for 48 h in the absence of E 2 . The cells were treated with E 2 (10 nM) for the indicated time, and real time PCR was used to quantify pS2 mRNA. C, hPEBP4-silenced MCF-7 cells were cultured for 48 h in the absence of E 2 . The cells were treated with E 2 (10 nM) for 1 h, and then chromatin immunoprecipitation was performed to detect the binding to the pS2 promoter sequence of ER␣ and its cofactors. PCR with pS2 promoter-specific primers was performed to quantify the amount of pS2 promoter sequence pulled down by the indicated antibodies. pS2 upstream primer pair was used as negative control. D, hPEBP4-silenced MCF-7 cells were transfected with or without pSG5-ER␣ plasmids, cultured for 24 h, and then treated with E 2 (10 nM) for 24 h. Cellular proliferation was then analyzed by BrdUrd (BrdU) incorporation assay. E and F, hPEBP4-silenced MCF-7 cells were transfected with luciferase reporter plasmids of ERE and pGL-TK vector or together with pSG5-ER␣ and siRNA-resistant vector (hPEBP4-R), cultured for 8 h, and then stimulated with E 2 (10 nM) for 16 h. Cells were treated with or without MG-132 for 1 h prior to E 2 stimulation. The luciferase activity was measured using Dual-Luciferase reporter assay system. Data show means Ϯ S.D. of triplicates from one experiment representative of three experiments. *, p Ͻ 0.05 versus mutation control. 2F). Together, these data suggest that hPEBP4-mediated inhibition of proteasome-dependent ER␣ degradation potentiates ER␣-mediated transactivation. hPEBP4

Overexpression Up-regulates ER␣-mediated Transactivation and Cellular Proliferation Induced by E 2 in HeLa
Cells-We further confirmed the above effects of hPEBP4 by using HeLa cells as a cell model. Because both hPEBP4 and ER␣ are not endogenously expressed in HeLa cells, we first established HeLa cells stably expressing hPEBP4 and then transfected with pSG5-ER␣ plasmid for 48 h to transiently express ER␣. Western blot was used to verify the forced expression of hPEBP4 and transient expression of ER␣ in HeLa cells (Fig. 3A). As shown in Fig. 3, B and C, overexpression of hPEBP4 led to increased levels of ER␣ protein, as well as the ER␣-mediated transcriptional activity. hPEBP4 overexpression also resulted in a higher proportion of proliferative HeLa cells (Fig. 3D). Once proteasome activity was blocked with MG-132, no promoting effect of hPEBP4 was observed on the ER␣ protein level as well as E 2 -induced transactivation (Fig. 3, C and E). Thus, consistent with the results of hPEBP4 silencing in MCF-7 cells, our findings indicated that overexpression of hPEBP4 in HeLa cells promoted ER␣-mediated transactivation and cellular proliferation induced by E 2 via inhibiting proteasome-dependent ER␣ degradation.
hPEBP4 Suppresses Src Activation and Disrupts the Interaction between Src and ER␣ via Association with Src-It has been reported that Src could promote ER␣ transcriptional activity and target ER␣ for ubiquitin-mediated proteolysis (15). We next tested the effect of hPEBP4 on Src activation. As shown in Fig. 4A, E 2 induced Src phosphorylation in a time-dependent manner. hPEBP4-silenced cells showed a more evident basal Src phosphorylation compared with that of mutation of control cells. The extent of Src phosphorylation was greater in hPEBP4-silenced cells compared with mutation of control cells during the entire time course, suggesting that silencing of hPEBP4 increased the activation of Src. Besides, the phosphorylation of ER␣ on Tyr-537, which is dependent on Src kinase and may then enhance E 2 binding to ER (26,27), was also increased after hPEBP4 silencing (Fig. 4A). We then evaluated the role of increased Src activity in the hPEBP4-mediated regulation of ER␣ protein expression. We found that treatment of MCF-7 cells with the Src inhibitor PP1 caused a dose-dependent increase of ER␣ protein levels over 36 h (supplemental Fig. S2). PP1 could restore the reduction of ER␣ and cyclin D1 induced by hPEBP4 depletion after E 2 stimulation in MCF-7 cells (Fig. 4B), indicating the involvement of Src in the regulation of ER␣ protein expression and its transcriptional activity by hPEBP4. As indicated in Fig. 4C, hPEBP4-silenced cells showed a more profound ubiquitination level of ER␣ after E 2 stimulation compared with mutation of control cells, which could be inhibited by PP1. Accordingly, the ER␣ protein level was inversely correlated with its ubiquitination level. More importantly, PP1 pretreatment could fully restore the decrease of ER␣ transcriptional activity induced by hPEBP4 silencing (Fig. 4D). These results strongly suggested that the decrease in ER␣ level induced by hPEBP4 silence may be attributed to the increased ubiquitination level because of the up-regulated Src activity. Unexpectedly, Src inhibition repressed ER␣ transcriptional activity in mutation control cells, suggesting that moderate activity of Src was required for efficient ER␣ transactivation. In HeLa cells, we found that overexpression of hPEBP4 significantly inhibited the activation of Src after E 2 stimulation (Fig. 4E). Forced expression of activated Src blocked the up-regulation of ER␣ induced by hPEBP4 overexpression (Fig. 4F). Therefore, the promoting effect of hPEBP4

hPEBP4 Inhibits ER␣ Degradation via Src
on ER␣-mediated transactivation via Src pathway is not restricted to MCF-7 breast cancer cells.
To further determine the correlation between hPEBP4 and Src, we performed coimmunoprecipitation to examine the association of the two molecules. Stably transfected HeLa cells were stimulated with E 2 and immunoprecipitated with either ER␣ or Src antibody. As shown in Fig. 5A, the association of ER␣ and Src was detected in both control and hPEBP4-overexpressing cells after E 2 stimulation. However, overexpression of hPEBP4 apparently reduced the interaction between ER␣ and Src. hPEBP4 associated with Src, but it failed to associate with ER␣ (Fig. 5A). We also found that p75, the construct lacking the phosphatidylethanolamine-binding domain of hPEBP5, could not immunoprecipitated with Src (Fig. 5B). These results suggest that hPEBP4 may disrupt the interaction between ER␣ and Src via association with Src.
hPEBP4 Promotes ER␣-mediated Transactivation and E 2 -induced Proliferation in a ERK-and Akt-independent Pathway-It is known that ER␣ activity can be regulated in a non-genomic pathway via protein-protein interactions. Several studies have shown that ER␣ can activate Src and its down stream kinases ERK and Akt (28,29). These pathways are critical in the proliferation and survival of breast cancer cells. Because hPEBP4 was found to be associated with Src, we then tested the effect of hPEBP4 on the activation of two major components, -ERK and Akt in estrogenmediated nongenomic signal pathway. Results of Western blot showed that E 2 could activate ERK but only slightly activate Akt in MCF-7 cells. Silencing of hPEBP4 significantly enhanced ERK activation but reduced Akt activation after treatment with E 2 (Fig. 6A). We then examined the ER␣-mediated transactivation after blocking the activity of ERK and Akt with U0126 and LY294002, respectively. As shown in Fig. 6B, we found that inhibiting ERK caused a reduction in ERE-luciferase activity (p Ͻ 0.05) but inhibiting Akt yielded no obvious effect. However, neither U0126 nor LY294002 could block the suppression effect of hPEBP4 silencing on ER␣-mediated transactivation. We also examined the accompanying ER␣ protein level after treatment with inhibitors for ERK and Akt. We found U0126 and LY294002 could not rescue the drop of ER␣ level caused by hPEBP4 silencing (Fig. 6C). These results imply that hPEBP4 also participates in the regulation of Akt and ERK in ER␣-mediated nongenomic signal pathway, but the promoting effect of hPEBP4 on the ER␣-mediated transactivation does not depend on its regulation of ERK and Akt activity.
Will ERK and PI3K/Akt, frequently suggested as important players in E 2 -induced proliferation, possibly participate in the process of hPEBP4-mediated cancer cell proliferation? To answer this question, we blocked ERK and Akt signaling with U0126 and LY294002, respectively, before measuring cancer cell proliferation. As shown in Fig. 6D, blocking either ERK or Akt dramatically suppressed the proliferation of E 2 -treated MCF-7 cells, suggesting that ERK and Akt are essential in E 2 -induced proliferation of cancer cells. However, inhibition of ERK and Akt did not rescue the repression of MCF-7 cell proliferation caused by hPEBP4 silencing. These results suggest that the effect of E 2 -induced proliferation of cancer cells by hPEBP4 is independent of its regulation of ERK and PI3K/Akt signal pathway.

. Src inhibitor PP1 reverses the inhibition of ER␣ protein level and ER␣-mediated transactivation by hPEBP4 knockdown.
A and B, hPEBP4-silenced MCF-7 cells were cultured in the absence of E 2 for 48 h. The cells were then pretreated with or without Src inhibitor PP1 (10 M) before the stimulation of E 2 for the indicated times. Western blot was performed with total cell lysates using the indicated antibodies. C, MCF-7 cells were cultured in 0.1% DCC-FCS and estrogen-deprived conditions for 48 h to deplete both growth factors and estrogen, and then cells were treated with PP1 for 10 min prior to a 6-h stimulation with E 2 and 5% DCC-FCS. ER␣ and phospho-Src levels were determined by Western blot. For measurement of ER␣ ubiquitination, ER␣ was precipitated and blotted with anti-ubiquitin antibody. IP, immunoprecipitation; IB, immunoblot. D, hPEBP4-silenced MCF-7 cells were transfected with luciferase reporter plasmids of ERE and pGL-TK vector, cultured for 7 h, and then treated with or without PP1 for 1 h prior to stimulation with E 2 (10 nM) for 16 h. The luciferase activity was measured using the Dual-Luciferase reporter assay system. E and F, stably transfected HeLa cells were transiently transfected with pSG5-ER␣ together with or without the activated Src plasmid (Src Y530F). The cells were cultured in the absence of E 2 for 48 h and then stimulated with E 2 (10 nM) for the indicated time (E) or for 4 h (F). Western blots were performed with total cell lysates by using the indicated antibodies. Data show means Ϯ S.D. of triplicates from one experiment representative of three experiments. *, p Ͻ 0.05.

DISCUSSION
hPEBP4 was first identified by us as an anti-apoptotic molecule and a potential target for treatment of breast cancer (10 -13). By silencing hPEBP4 in MCF-7 cells and overexpressing hPEBP4 in HeLa cells, we have demonstrated that hPEBP4 suppresses the Src activation and disrupts the interaction between Src and ER␣ via association with Src, resulting in inhibition of proteasome-dependent degradation of ER␣ and promotion of ER␣-mediated transactivation of a targeted gene such as pS2 and cyclin D1. These molecular perturbations may be responsible for the promoting effect of hPEBP4 on E 2 -induced cellular proliferation mediated by ER␣. Degradation of ER␣ may be part of the physiological negative feedback during the interaction of estrogen and estrogen-responsive tissues. However, cancer cells with genetic instability will probably evolve various means to minimize the "side effect" of residual feedback in malignant development. Up-regulation of hPEBP4 in several estrogen-related tumors, as demonstrated previously by us (10,12), may be one of the means for ER-related tumors to avoid ER␣ degradation. By inhibiting the degradation of ER␣, hPEBP4 sustains the ER␣ at a high level and facilitates the carcinogenic effect of estrogen.
Degradation of ER␣ is an important mechanism in the regulation of ER␣ activity. Previous studies have suggested that ER␣ phosphorylation by several kinases can promote its degradation. It has been reported that E 2 -induced rapid activation of ERK and Akt can promote ER␣-mediated transactivation by phosphorylating coactivators of ER␣ or ER␣ itself (30,31). However, our study showed that neither ERK nor Akt accounts for the promoting effect of hPEBP4 on ER␣-mediated transactivation because inhibiting ERK or Akt could not rescue the increased degradation of ER␣ caused by hPEBP4 silence, which is consistent with our previous study showing that blockage of PI3K or ERK does not alter the ER␣ level when bound to E 2 (32). Instead, our data suggest that Src may be responsible for hPEBP4-induced elevation of ER␣ levels. ER␣ phosphorylation by Src kinase can increase its affinity for estrogen (26) and therefore enhance the ER␣-coactivator binding and transcriptional activity of ER␣. Also, Src appears to promote ER␣ ubiquitylation because Src inhibition impaired cellular ER␣ ubiquitylation and proteolysis in vivo (33). Therefore, liganded ER␣ may recruit Src or Src-dependent kinases, leading to phosphorylation of ER␣ that facilitates ER␣ binding to coactivators and/or components of the proteolytic machinery (15). Our results suggested that hPEBP4 could inhibit the interaction between Src and ER␣, thereby inhibiting the Src-dependent degradation of ER␣. However, we are not sure whether the reduced interaction between Src and ER␣ can be attributed to down-regulated activity of

hPEBP4 Inhibits ER␣ Degradation via Src
Src itself induced by hPEBP4 overexpression. Glutathione S-transferase pulldown assay showed that GST-hPEBP4 fusion protein could not pull down either Src or ER␣ (data not shown), indicating the interaction between Src and hPEBP4 may be indirect or dynamic. Besides, Src or its downstream effectors may also regulate ligand-activated ER␣ coactivator phosphorylation to affect ER␣ degradation, because several ER␣ coactivators are known to be ubiquitin ligases (34,35). We have found that the E 2 -induced phosphorylation of MDM2, an E3 ligase and also a direct activator of ER␣ (35), was inhibited by hPEBP4 overexpression and increased by hPEBP4 silencing (supplemental Fig S3, A and B). Wild-type MDM2 overexpression could completely abolish the effect of hPEBP4 on ER␣mediated transactivation in HeLa cells, whereas mutant MDM2 S166A/S186A could not (supplemental Fig S3C), indicating the critical role of the E3 ligase MDM2 in the regulation of ER␣-mediated transactivation by hPEBP4. Our data support the notion that hPEBP4 regulates ER␣ activity via association with Src in estrogen-mediated nongenomic pathway. Nevertheless, how hPEBP4 affects Src-dependent ER␣ degradation remains to be further investigated.
The results showed here are in contrast to other reports that show an inverse relationship between receptor proteolysis and transcriptional efficiency (36,37) or a dissociation of ER␣ activation from degradation (38 -40). Our results are consistent with several recent reports demonstrating that proteasome inhibition increases hormone-induced activation of ER␣ (22). For example, inhibition of nuclear export with leptomycin B impairs E 2 -induced ER␣ proteasome degradation and then increases ER␣-mediated transcription (41). Increased proteasome-dependent ER␣ degradation and thus the depleted ER␣ level, induced by transforming growth factor-␤1 (42) or hydroxamic acid analogue pan-HDAC inhibitors (43), abrogated its transcriptional activity, E 2 -driven DNA synthesis and cell growth in E 2 -responsive cancer cells. One possible explanation that could account for the phenomenon is that preventing ER␣ degradation effectively increases the number of functional receptors available to participate in transcriptional processes. It has been demonstrated that inhibiting receptor proteolysis with proteasome inhibitors increases the estrogenbinding capacity (25). Thus, receptors that evade proteolysis retain certain ER␣ functions, including ligand binding. Similar results have been reported with glucocorticoid receptor in embryonic hippocampal neurons, which show that activated receptors that are not targeted to proteasome retain transactivation capacity (44). Studies of ER␣ in stable Chinese hamster ovary (45) and MCF-7 breast cancer cells (46) demonstrate that the magnitude of ER␣ transcriptional activity is directly related to the concentration of receptor. Thus, any regulatory mechanism that has the potential to elevate the concentration of functional ER␣ could similarly augment the transcriptional output of ER␣.
Key protein stabilities in malignancy and their destructive destination by the ubiquitin-proteasome or related pathways will no doubt be an important focus of future endeavors. The necessity of proteasomal targeting in the delicate control of cellular growth is only beginning to unfold. Here, for the first time, we have demonstrated the inhibition of proteasome-de-pendent degradation of ER␣ by hPEBP4 via the Src pathway, resulting in the promotion of E 2 -induced transactivation of ER␣ and cellular proliferation. This study provides a new therapeutic target as being helpful for further exploring new regulatory mechanisms in carcinogenesis, especially the hormonally responsive breast cancers.