Prostate Specific Membrane Antigen associates with Filamin A, beta1 integrin, pp130CAS and pSrc regulating beta1 integrin activation and survival of prostate cancer cells.

Prostate cancer (PCa) is the second cause of cancer-related death in males of developed countries. Pre-clinical evidence has revealed that virtually all specimens of advanced PCa express great amounts of Prostate Specific Membrane Antigen (PSMA), a transmembrane folate-hydrolase/ carboxypeptidase, associated in 60% of the samples with activation of survival pathways, with an increase of p130CAS, the major scaffolding protein of beta1 integrin signaling complex. Up-regulated PSMA expression in aggressive prostate tumors provides a selective advantage for tumor cells and a role for PSMA in cancer cell growth and progression is therefore envisaged. We have previously reported that PSMA clustering induces RAS-RAC-MAPK pathway activation, NF-kB translocation, IL-6 gene expression and proliferation in LNCaP cells, a widely used cellular model of advanced PCa cells. We here investigated some of the mechanisms required for signaling activity of PSMA. PSMA has a short intracellular cytodomain, devoid of phopshorylation or docking sites for kinase and adaptor and therefore unable to signal directly. However, it exposes a functional recognition site for Filamin A (FLNa). Recent evidence indicates that FLNa associates also with beta1 integrin (beta1) regulating its intracellular trafficking. This led us to make the hypothesis that FLNa could mediate a cooperation between PSMA and beta1 in the assembly of a signaling platform which may mediate, among others, the activation of anti apoptotic pathways. Our cellular system was constituted by the LNCaP cell line, derived from a lymphnodal metastasis and expressing PSMA endogenously, the PC3-PSMA cell line, originated from bone metastasis of PCa and the MCF7-PSMA cell line, derived from breast carcinoma. PC3-PSMA and MCF7-PSMA were stably transfected with the human PSMA gene. In the first part of the work we have investigated the signaling pathways engaged by PSMA activation (PSMA cross-linking), demonstrating that PSMA induced the parallel activation of both AKT/mTOR/BAD and p38 and ERK1/2 MAPKs pathways in the cell lines, irrespectively of their different origin. The relative role of the two pathways in the basal cell survival was evaluated next in LNCaP by using different doses of specific inhibitors of Wortmannin (inhibitor of PI3K, the main activator of AKT) or a mix of SB202190 (inhibitor of p38) and PD 98059 (inhibitor of MEK1, the ERK-activating kinase). Reversible and irreversible apoptosis was quantified respectively by FACS analysis of Annexin V exposure or ELISA measuring the cytosolic release of DNA–hyston complexes. The most significant and synergic induction of apoptosis was observed when both pathways were inhibited simultaneously with a mix of all drugs used at maximal doses. As PSMA cross-linking regulates the activation of the two anti apoptotic pathways we next assayed if PSMA cross-linking could rescue LNCaP from apoptosis induced by serum withdrawal. Results clearly indicated that this was the case. Moreover, PSMA-induced rescue could be abrogated, by inhibiting MAPK p38 or ERK1/2 or PI3K pathways, thus demonstrating the PSMA activity depended on its ability to activate AKT, p38 and ERK1/2. Among MAPKs, ERK1/2 seemed to prevail on p38. As previously mentioned, PSMA cytodomain is by structurally unable to start the assembly of a signaling complex. We therefore considered FLNa, beta1, Src and p130CAS as potential molecular partners. Confocal analysis indicated that PSMA and beta1 were located in close proximity on the membrane of the majority of LNCaP cells. Immunoblotting of immunoprecipitates prepared from cell lysate of LNCaP cells after PSMA activation, showed that beta1 and FLNa could be pulled down from PSMA IPs, together with phopsphorylated p130CAS and Src. Conversely, PSMA, FLNa, pp130CAS and pSrc could be pulled down from beta1 IPs. Noteworthy, PSMA activation was rapidly followed by Src and p130CAS phosphorylation. In investigating possible location of the various molecules in the lipid rafts of untreated cells we found PSMA in Lubrol WX-detergent resistant membranes (DRMs) with FLNa, beta1, phosphorylated Src and p130CAS, but not with not phosphorylated Src and p130CAS, thus suggesting that DRMs could be a privileged membrane location for the complex formation. The first and perhaps most important effect of this complex assembly was the beta1 activation. As assessed by cytofluorometry, the very low expression of HUTS-21 activation epitope on beta1, in all the three PSMA-positive cell lines analysed, raised dramatically 5 min after PSMA cross-linking, peaked at 10 min and declined at 20 min whereas the total expression of beta1 remained stable throughout the experimental time, thus indicating that PSMA and beta1 entertained a very remarkable relationship in PCa cells. The role of the different partners of the PSMA-signaling complex, in the anti-apoptotic activity, was determined by siRNA silencing or chemical blockage. The phosphorylation of AKT and ERK1/2, induced by PSMA cross-linking, was greatly and significantly reduced at all points of the time courses in FLNa or beta1 or p130CAS silenced LNCaP if compared to the wt cells, thus demonstrating that both molecules were required for PSMA signaling activity. Src activity was as well needed, as specific Src inhibition, induced by SU6656 or PP1 drug, inhibited ERK phosphorylation and AKT phosphorylation, induced by PSMA activation. The effects of FLNa or beta1 or p130CAS silencing on PSMA-mediated rescue from apoptosis, were assayed by using 3D cultures. LNCaP cells do formed colonies in Matrigel at 4% FBS. The number of colonies was strongly increased upon PSMA activation, halved by serum deprivation but was fully restored if PSMA was activated. Colonies in starved samples showed the “shrinking” phenotype of apoptotic cells. FLNa silencing did not affect the ability of LNCaP cells to form colonies at 4% FBS, nor aggravated the effects of starvation if compared to wt cells. However, it abrogated the promoting/rescuing effect of PSMA cross-linking in the two conditions. Beta1 or p130CAS silencing changed dramatically the LNCaP growth in 3D culture. The number of colonies at 4% FBS and at 0% FBS dropped down up to one tenth of that of wt cells and PSMA cross-linking lacked activity in both conditions. Very unexpected results were instead observed regarding the beta1 activation in LNCaP, PC3-PSMA and MCF7-PSMA cells silenced for FLNa or p130CAS. PSMA activation was followed by a strong activation of beta1. However, this activation was not anymore time-dependent, nor followed by activation of ERK or AKT. More investigation is needed on this matter. In all, our results demonstrate for the first time that PSMA expression gives a remarkable advantage to PCa cells by recruiting a signaling complex, including FLNa, beta1, p130CAS and Src, and thereby supporting the activation of antiapototic and pro-proliferative signaling pathways.