Ligation of Prostate Cancer Cell Surface GRP78 Activates a Proproliferative and Antiapoptotic Feedback Loop

GRP78, a well characterized chaperone in the endoplasmic reticulum, is critical to the unfolded protein response. More recently, it has been identified on the cell surface, where it has many roles. On cancer cells, it functions as a signaling receptor coupled to proproliferative/antiapoptotic and promigratory mechanisms. In the current study, we demonstrate that ligation of prostate cancer cell surface GRP78 by its natural ligand, activated α2-macroglobulin (α2M*), results in a 2–3-fold up-regulation in the synthesis of prostate-specific antigen (PSA). The PSA is secreted into the medium as an active proteinase, where it binds to native α2M. The resultant α2M·PSA complexes bind to GRP78, causing a 1.5–2-fold increase in the activation of MEK1/2, ERK1/2, S6K, and Akt, which is coupled with a 2–3-fold increase in DNA and protein synthesis. PSA is a marker for the progression of prostate cancer, but its mechanistic role in the disease is unclear. The present studies suggest that PSA may be involved in a signal transduction-dependent feedback loop, whereby it promotes a more aggressive behavior by human prostate cancer cells.

Prostate-specific antigen (PSA) 2 is a serine proteinase that complexes with serum proteinase inhibitors, including ␣ 2macroglobulin (␣ 2 M) (1). Monitoring of PSA is generally recommended in men over 50 years old to screen for prostate cancer; however, there is significant controversy with respect to its use as a marker for the appearance of the disease (1-6). By contrast, substantial data support its use in monitoring progression and metastasis (1,2,5). Recently, a Phase II controlled trial suggested that immunization against PSA prolongs patient survival, suggesting a direct role for PSA in the pathobiology of the disease (7). Studies have shown that PSA produced by metastatic prostate cancer cells is involved in bone remodeling, a common feature in the bony metastasis of this disease (8). There is, however, little other evidence indicating a direct role for PSA in disease progression. ␣ 2 M is synthesized by many tissues, particularly the liver; however, it is also produced locally in prostate stromal tissue, where it is available to complex with PSA (9). When a proteinase attacks the so-called "bait region," thiol esters in each of the four ␣ 2 M subunits rupture, and the protein undergoes a very large conformational change, exposing receptor recognition sites in each subunit (10). Two receptors have been identified for activated forms of ␣ 2 M (␣ 2 M*), namely the LDL receptor-related protein (LRP) and cell surface-associated GRP78 (glucose-regulated protein of M r ϳ78,000) (10,11). In addition to proteinases, exposure of ␣ 2 M to small primary amines or ammonia, by direct attack on the thiol esters, produces ␣ 2 M* (10). Although GRP78 is primarily known as a resident endoplasmic reticulum chaperone, it appears on the cell surface of many types of malignant cells, including human prostate cancer (12)(13)(14)(15)(16). Here it has many functions, including serving as an ␣ 2 M* signaling receptor linked to proproliferative, promigratory, and antiapoptotic signaling cascades (12)(13)(14)(15)(16). Patients with many forms of cancer, including prostate, may develop autoantibodies that bind to the NH 2 -domain of GRP78 where ␣ 2 M* binds (16 -18). These antibodies are receptor agonists linked to a poor outcome for patients with prostate cancer (16). In the present study, we demonstrate that ligation of GRP78 on the surface of human prostate cancer cells causes increased production and secretion of PSA, which is functional and complexes with native ␣ 2 M. These complexes bind to cell surface GRP78, triggering ERK and Akt activation and increased DNA and protein synthesis. We hypothesize that ligation of cell surface GRP78 on prostate cancer cells may drive an autoregulatory feedback loop wherein reaction of ␣ 2 M with PSA generates more ␣ 2 M*, thus further activating proproliferative/antiapoptotic signaling mechanisms. ences, and [ 3 H]thymidine (specific activity 71.5 Ci/mmol) was from American Radiochemicals Inc. (St. Louis, MO). PSA was purchased from Innovative Research of America (Sarasota, FL). Other reagents of the highest available purity were procured locally.
Cancer Cell Lines-In preliminary experiments, we used two prostate cancer cell lines, 1-LN and DU-145, which express GRP78 on their cell surface, and the PC-3 prostate cancer line, which does not express GRP78 on the cell surface (11)(12)(13)(14)(15). The highly metastatic 1-LN cell line is derived from the less metastatic PC-3 line and was a kind gift of Dr. Phil Walther (Duke University Medical Center, Durham, NC). The PC-3 and DU-145 cell lines were obtained from the American Type Culture Collection. These cells were grown in 6-well plates (500 ϫ 10 3 cells/well) to confluence in RPMI 1640 medium containing 10% fetal bovine serum, 2 mM glutamine, 12.5 units/ml penicillin, 6.5 g/ml streptomycin, and 10 mM insulin in a humidified CO 2 (5%) incubator. At 90% confluence, the medium was aspirated, the monolayers were washed with ice-cold HHBSS, a fresh volume of medium was added, and the cells used for the experiments described below.
Measurement of ␣ 2 M⅐Methylamine-induced Expression of PSA Protein in 1-LN, DU-145, and PC-3 Prostate Cancer Cells-In initial experiments, we determined the effect of varying concentrations of ␣ 2 M⅐methylamine on PSA levels by Western blotting. 1-LN DU-145 and PC-3 cells grown to confluence were incubated in the RPM1640 medium used above in 6-well plates (3 ϫ 10 6 cells/well). The cells were stimulated with varying concentrations of ␣ 2 M⅐methylamine and then incubated at 37°C in a humidified CO 2 (5%) incubator for 20 min. The reaction was terminated by aspirating the medium. The cells were lysed in lysis buffer A containing 50 mM Tris⅐HCl (pH 7.5), 12 mM NaCl, 1% Nonidet P-40, 2.5 mM sodium fluoride, 1 mM sodium pyrophosphate, 0.1 mM sodium orthovanadate, 1 mM PMSF, 1 mM benzamidine, and leupeptin (20 g/ml) over ice for 20 min. Cell lysates were scraped into new Eppendorf tubes and centrifuged for 5 min at 800 ϫ g at 4°C, and protein contents of lysates were determined (19). To equal amounts of lysate protein, a volume of 4ϫ sample buffer was added, and samples were boiled for 5 min. Samples were electrophoresed on 10% gel, transferred to PDVF membrane, and membranes were immunoblotted with anti-PSA antibodies. Protein bands on the immunoblot were visualized and quantified by ECF and phosphorimaging. The membrane was reprobed for actin as a protein loading control.
Measurement of PSA mRNA Levels by Reverse Transcription-Total RNA from 1-LN, DU145, and PC-3 prostate cancer cells treated with buffer or ␣ 2 M* was extracted by a single method using an RNeasy minikit (Qiagen, Chatsworth, CA) according to the manufacturer's instructions. Total RNA was reversetranscribed with 1 g of RNA in a 20-l reaction mixture, using Moloney murine leukemia virus reverse transcriptase (200 units) and oligo(dT) as primer for 1 h at 4°C. The resulting cDNA (5 g) was used as a template, and a 225-bp segment of the PSA cDNA was amplified using a 20-mer upstream primer (5Ј-CCA ACA CCC GCT CTA CGA TA-3Ј) and a 22-mer downstream primer (5Ј-ACC TTC TGA CGG TGA ACT TGC G-3Ј). A 302-bp segment of mouse ␤-actin (constitutive internal control) cDNA was co-amplified using a set of PCR primers provided in an R&D Systems kit (Minneapolis, MN). Amplification was carried out in a Biometra T3 thermocycler for 28 cycles (one cycle: 94°C for 45 s, 60°C for  45 s, and 70°C for 45 s). PCR products were analyzed on a 1.2% (w/v) agarose/ethidium bromide) gel. The gels were photographed, and the intensity of individual PSA and ␤-actin mRNA bands was quantified as PSA/␤-actin ratios.
Measurement of the Effects of Silencing GRP78 Gene Expression by RNAi in 1-LN, DU-145, and PC-3 Cells on ␣ 2 M⅐methylamine-induced PSA Synthesis-To determine the requirement of cell surface GRP78 for ␣ 2 M⅐methylamineinduced up-regulation of PSA, we silenced the expression of GRP78 by RNAi. In our earlier publications we have used two GRP78 targeting mRNA sequences for silencing its gene expression and found identical effects on GRP78 expression as well as downstream signaling (15). We also found that transfection of cells with GRP78 dsRNA down-regulates both the total cellular GRP78 pool and cell surface-localized GRP78 (20). Therefore, in this investigation, we have used only one GRP78 targeting mRNA sequence for silencing the expression of GRP78. The chemical synthesis of dsRNA homologous in sequence to the target GRP78 370 K1QQLVK 376 mRNA sequence. 5Ј-AAA ATA CAG CAA TTA GTA AAG-3Ј (Swiss Prot GRP primary sequence accession number P11021) were performed by Ambion (Austin, TX). For making GRP78 dsRNA, sense (5Ј-AAU ACA GCA AUU AGU AAA GTT-3Ј) and antisense (5Ј-CUU UAC UAA UUG CUG UAU UTT-3Ј) oligonucleotides were annealed by Ambion. Transfection of 1-LN, DU-145, and PC-3 cells with GRP78 dsRNA was done as described previously (15). Briefly, confluent 1-LN DU-145 and PC-3 cell monolayers (1.5 ϫ 10 6 cells/well in 6-well plates) were incubated as described above and transfected with 75 nM annealed GRP78 dsRNA and control cells were transfected with Lipofectamine as described previously (15). Forty-eight h after transfection, the control cells were stimulated with either buffer or ␣ 2 M⅐methylamine (50 pM for 20 min). Cells transfected with scrambled dsRNA and treated with ␣ 2 M* were used as the control. The reactions were terminated by aspirating the medium, and cells were lysed in lysis buffer A over ice for 15 min. Lysates were then transferred to tubes and centrifuged, and proteins in the supernatant were determined. Equal amounts of lysate protein were electrophoresed, transferred to membrane, and immunoblotted with PSA antibodies. Protein bands on the membrane were detected and quantified by ECF and phosphorimaging. The membrane was reprobed for protein loading control actin.
Measurement of ␣ 2 M*-induced Synthesis of PSA in Prostate Cancer Cells-Confluent 1-LN cells incubated in the RPMI 1640 medium described above in 6-well plates (4 ϫ 10 6 cells/ well) were labeled with [ 35 S]-Express S35S35 protein labeling mixture (PerkinElmer Life Sciences) (250 Ci/ml) for 2 h at 37°C in a humidified CO 2 (5%) incubator. The labeling was stopped by aspirating the medium; monolayers were washed four times with cold Hanks' balanced salt solution containing 10 mM HEPES, pH 7.4, and 3.5 mM NaHCO 3 buffer; a volume of the RPMI medium was added; and cells were incubated at 37°C for temperature equilibration. Cells in the wells were treated with anti-CTD antibody (2 g/ml) or buffer for 1 h before adding ␣ 2 M* (50 pM for 20 min), and the cells were incubated as above. The reactions were stopped by aspirating the medium; a volume of lysis buffer containing 50 mM Tris⅐HCl (pH 7.5), 120 mM NaCl, 1% Nonidet P-40, 25 mM sodium fluoride, 1 mM sodium pyrophosphate, 0.1 mM sodium orthovanadate, 1 mM PMSF, 1 mM benzamidine, and leupeptin (20 g/ml) was added to each well; and cells were lysed for 15 min over ice. Cell lysates were scraped into respective tubes and centrifuged for 5 min at 800 ϫ g at 4°C, and protein contents of lysates were determined (19). Equal amounts of lysate proteins were immunoprecipitated with anti-PSA antibodies (1:50) in the presence of Protein A-agarose overnight at 4°C with rotation. PSA immunoprecipitates were washed three times with the above lysis buffer by centrifuging at 2500 rpm at 4°C for 5 min. Thirty l of 4ϫ sample buffer was added to each immunoprecipitate; samples were boiled for 5 min and centrifuged for 4 min; an equal volume of supernatant was electrophoresed (10% gel) and transferred onto PVDF membranes; and the membranes were autoradiographed in a PhosphorImager and quantified.
Measurement of Formation of ␣ 2 M*⅐PSA Complex from Secreted PSA and Exogenous Native ␣ 2 M* in 1-LN Cells Stimulated with ␣ 2 M*-1-LN cells incubated overnight in serumfree RPMI 1640 medium in 6-well plates (4 ϫ 10 6 cells/well) were stimulated with either buffer or ␣ 2 M* (50 pM) and incubated as described above. At 12 h of incubation, another dose of ␣ 2 M* (50 pM) was added, and the cells were incubated for an additional 11 h. At this time 50 nM native ␣ 2 M was added to the medium, and cells were incubated for 1 h. At the end of the incubation, medium from both groups was carefully collected and centrifuged at 800 ϫ g for 5 min at 4°C. The respective supernatants were pooled and concentrated to small volume by centrifugation in Amicon ultracentrifuge units (Millipore Corp., Billerica, MA) at 2500 rpm at 4°C for the desired period of time. The concentrates were carefully removed, a volume of the above lysis buffer was added to each concentrate, and their protein contents were determined (19). Equal amounts of concentrate proteins were immunoprecipitated with anti-PSA antibody (1:50) in the presence of Protein A-agarose overnight with 4°C with gentle buffer. The PSA immunoprecipitates were washed three times with lysis buffer at 2500 rpm for 5 min at 4°C. A volume of Laemmli 2ϫ sample buffer without DTT was added to each immunoprecipitate; each sample was boiled for 5 min and centrifuged for 4 min; and an equal volume of supernatant was electrophoresed (10% gel), transferred to PVDF membrane, and immunoblotted with anti-PSA antibodies overnight at 4°C with gentle shaking. The detection and quantitation of immunoblots were performed by ECF and phosphorimaging. The membranes were reprobed for ␣ 2 M.
Synthesis of ␣ 2 M⅐PSA Complex-The synthesis of human ␣ 2 M⅐PSA complexes was performed according to Otto et al. (21). ␣ 2 M was isolated and purified from human plasma as described previously (11). Native ␣ 2 M was incubated with a 5-fold molar excess of PSA at 37°C for 1 h in PBS. The ␣ 2 M⅐PSA complex was isolated by chromatography on a HiPrep Sephacryl 16/60 S300 HR column (GE Healthcare) equilibrated in PBS. The ␣ 2 M⅐PSA complex, which eluted first, was collected, and fractions were pooled and concentrated in a Pierce (Rockford, IL) iCon concentrator to a final concentration of 2-4 mg/ml. Eluates of unincorporated PSA, which eluted later, were also collected, pooled, and concentrated. The formation of ␣ 2 M⅐PSA complex was analyzed by native gel electrophoresis. Two-g samples of methylamineactivated ␣ 2 M or ␣ 2 M⅐PSA complex were loaded onto a 7% Tris-acetate gel and electrophoresed with TBE extended range (Bio-Rad) running buffer for 1.5 h. The protein bands were detected with Coomassie staining.
Determination of the Effects of ␣ 2 M⅐PSA Complex on DNA Synthesis in 1-LN, DU-145, or PC-3 Cells-In preliminary experiments, we determined the effect of incubating 1-LN cells with varying concentrations of ␣ 2 M⅐methylamine and premade ␣ 2 M⅐PSA on incorporation of [ 3 H]thymidine into DNA, as described previously (22,23). After determining the optimal concentrations of ␣ 2 M⅐methylamine and ␣ 2 M⅐PSA required for maximal incorporation of [ 3 H]thymidine into cells, we used this concentration to examine the modulation of ␣ 2 M*-induced DNA synthesis. 1-LN cells (300 ϫ 10 3 cells/ well) in 48-well plates were grown in RPMI medium in a humidified CO 2 (5%) incubator at 37°C. At ϳ90% confluence, the medium was aspirated, and a volume of RPMI was added, followed by the addition of either buffer, native ␣ 2 M (50 pM), PSA (50 pM), cell-free medium, ␣ 2 M⅐methylamine (50 pM), premade ␣ 2 M⅐PSA (50 pM) or complex obtained by adding ␣ 2 M (5 g/ml) to 1-LN tumor cells in culture, anti-CTD antibody (2 g/ml), anti-CTD antibody and then ␣ 2 M⅐methylamine, anti-CTD antibody and then premade ␣ 2 M⅐PSA, anti-CTD antibody and then ␣ 2 M⅐PSA made in culture, actinomycin D and then premade ␣ 2 M⅐PSA, actinomycin D (5 g/1 ml) and then ␣ 2 M⅐methylamine (50 pM), actinomycin D and then premade ␣ 2 M⅐PSA, or actinomycin D and then ␣ 2 M⅐PSA made in the medium as above. [ 3 H]thymidine (2 Ci/ml) was added to each well, and the cells were incubated overnight. The reactions were terminated by aspirating the medium, and monolayers were washed twice with ice-cold 5% TCA, followed by three washings with ice-cold PBS. Cells were lysed in a volume of 1 N NaOH (40°C/2 h), lysate protein was estimated, and the samples were counted in a liquid scintillation counter (14). To further examine the requirement of cell surface GRP78 expression for cell-proliferative effects of ␣ 2 M-NH 2 , we measured [ 3 H]thymidine uptake in DU-145 cells, which express cell surface GRP78, and on PC-3 cells, which do not. DU-145 and PC-3 cells were grown in 48-well plates in RPMI medium as described above. At about ϳ90% confluence, the medium was aspirated, and a volume of RPMI medium was added, followed by the addition of either buffer, native ␣ 2 M (50 pM), ␣ 2 M⅐methylamine (50 pM), premade ␣ 2 M⅐PSA (50 pM) complex obtained by adding ␣ 2 M (5 g/ml) to 1-LN tumor cells in culture or anti-CTD antibody (2 g/ml) added 1 h before adding ␣ 2 M⅐methylamine, or ␣ 2 M⅐PSA complexes as above. The effects of anti-CTD antibody alone were also studied. [ 3 H]thy-midine (2 Ci/ml) was added to each well, and the cells were incubated overnight. The incubations were then processed as described for 1-LN cells.
Determination of the Effects of ␣ 2 M⅐PSA Complex on Protein Synthesis in 1-LN Cells-1-LN cells (300 ϫ 10 3 cells/well in 48-well plates were grown in RPMI medium in a humidified CO 2 (5%) incubator at 37°C. At ϳ90% confluence, the medium was aspirated, and a volume of RPMI was added, followed by the addition of either buffer, native ␣ 2 M (50 pM), PSA (50 pM), cell-free medium, ␣ 2 M⅐methylamine (50 pM), ␣ 2 M⅐PSA complexes as above, anti-CTD antibody (2 g/ml), anti-CTD antibody and then ␣ 2 M⅐methylamine (50 pM), anti-CTD antibody and then ␣ 2 M⅐PSA as above, actinomycin D (5 g/ml) and then ␣ 2 M⅐methylamine (50 pM), or actinomycin D and then ␣ 2 M⅐PSA as above. To the wells was then added [ 3 H]leucine (2 Ci/ml), and cells were incubated overnight as above. The reactions were terminated by aspirating the medium, and monolayers were washed twice with ice-cold 5% TCA, followed by three washings with ice-cold PBS. Cells were lysed in a volume of 1 N NaOH (40°C/2 h), protein was estimated, and the lysates were counted in a liquid scintillation counter (14). -well plates were grown in the RPMI medium used above in a humidified CO 2 (5%) incubator at 37°C. At ϳ90% confluence, the medium was aspirated, and a volume of RPMI was added, followed by the addition of either buffer, native ␣ 2 M (50 pM for 20 min), ␣ 2 M⅐methylamine (50 pM for 20 min), premade ␣ 2 M⅐PSA (50 pM for 20 min), or tumor tissue culture-derived ␣ 2 M⅐PSA (␣ 2 M, 5 g/ml for 20 min) in the respective wells. The reaction was terminated by aspirating the medium, and a volume of lysis buffer A was added. Cells were lysed over ice for 15 min and scraped into tubes, and tubes were centrifuged for 5 min at 800 ϫ g at 4°C. The supernatants were transferred to new tubes, their protein contents were estimated, and equal amounts of protein were electrophoresed (10 or 12.5% gels). Protein bands were transferred to PVDF membrane and membrane-immunoblotted for PSA, GRP78, phospho-MEK1/2, phospho-ERK1/2, phospho-p38 MAPK, phospho-Akt Thr-308 , phospho-Akt Ser-473 , and phospho-S6K with their specific antibodies. The respective immunoblots were probed for their protein loading control actin and respective unphosphorylated protein. Protein bands were detected by ECF and quantified using a PhosphorImager.
Statistical Analysis-Statistical significance of the data were determined by Student's t test.

␣ 2 M⅐Methylamine Up-regulates Expression of PSA Protein in 1-LN and DU-145 but Not in PC-3 Prostate Cancer Cells-
Treatment of 1-LN or DU-145 prostate cancer cells with varying concentrations of ␣ 2 M⅐methylamine for 20 min caused a concentration-dependent increase in PSA protein, as determined by Western blotting (Fig. 1A). The maximum increase occurred at about 50 pM ␣ 2 M⅐methylamine; therefore, we have used this concentration of ␣ 2 M⅐methylamine in the experiments described below. This same treatment had no effect on PC-3 cells.    the buffer-treated cells (Fig. 1A). Pretreatment of 1-LN cells with anti-CTD antibody also inhibited ␣ 2 M*-induced transcriptional up-regulation of PSA as judged by mRNA levels (Fig. 1B). Likewise, ␣ 2 M⅐methylamine treatment of 1-LN and DU145 cells also increased PSA levels as determined by Western blotting (Fig. 2). Pretreatment of these cells with anti-CTD antibody significantly attenuated the expression of PSA protein (Fig. 2).  proliferation and prevents apoptosis as a result of activating several antiapoptotic signaling pathways (10 -16). We have previously shown that under our experimental conditions, silencing GRP78 gene expression reduces GRP78 mRNA and cellular protein levels by about 60 -65% (15). We hypothesized that if activated ␣ 2 M* promotes PSA synthesis consequent to its binding to cell surface GRP78, then down-regulation of cell surface GRP78 by RNAi in cancer cells expressing GRP78 on their cell surface should decrease PSA synthesis in these cells. Indeed, silencing GRP78 expression by RNAi in 1-LN and DU-145 cells, but not PC-3 cells, significantly decreased ␣ 2 M⅐methylamine-induced synthesis of PSA (Fig. 1D).

Down-regulating GRP78 Expression by RNAi Inhibits
␣ 2 M⅐Methylamine Induces Increased Synthesis of PSA in 1-LN Prostate Cancer Cells-To further demonstrate ␣ 2 M⅐methylamine-induced up-regulation of PSA transcription in 1-LN cells, we next determined PSA synthesis by measuring the incorporation of 35 S-labeled precursor amino acids (Fig. 3). ␣ 2 M⅐methylamine-stimulated cells caused a 2-3-fold increased synthesis of PSA compared with buffertreated cells (Fig. 3). Pretreatment of cells with anti-CTD antibody nearly abolished ␣ 2 M*-induced PSA synthesis (Fig. 3). Anti-CTD antibody exerted a similar effect on PSA mRNA levels (Fig. 1C).
Secretion of Newly Synthesized PSA into Media and Its Interaction with Exogenous Added Native ␣ 2 M* in ␣ 2 M⅐methylamine-treated 1-LN Cells-In the next series of experiments, we measured the secretion of PSA by 1-LN cells treated with ␣ 2 M*⅐methylamine and interaction of secreted PSA with exogenously added native ␣ 2 M (Fig. 4, A and B). Secreted PSA reacted with added ␣ 2 M, as is evident from the immunoblot showing the reactivity of PSA (ϳ33 kDa) and native ␣ 2 M subunits (ϳ200 kDa) protein bands with antibody against PSA (Fig. 4A). The PSA band on the immunoblot upon reprobing did not react with anti-␣ 2 M antibody, but the ␣ 2 M band did (Fig. 4B), confirming ␣ 2 M⅐PSA complex formation between secreted PSA and ␣ 2 M. Pretreatment of cells with anti-CTD antibody inhibited ␣ 2 M⅐PSA complex formation. This effect appears to be dependent on anti-CTD antibody-induced inhibition of PSA synthesis, resulting in its diminished secretion and thus ␣ 2 M⅐PSA complex formation (Fig. 4A).
Synthesis of ␣ 2 M⅐PSA Complex-We also premade ␣ 2 M⅐PSA and purified and characterized the complex. Nearly all of the PSA reacted with ␣ 2 M. The ␣ 2 M⅐PSA complex eluted first from a HiPrep Sephacryl 16/60 S300 HR column, followed by unreacted PSA (Fig. 4C). Upon electrophoresis, the synthesized ␣ 2 M⅐PSA complex showed faster mobility than the native ␣ 2 M, similar to that observed with ␣ 2 M⅐methylamine (Fig. 4D). These are the expected properties of an ␣ 2 M-proteinase complex (12).

Secreted and Synthesized ␣ 2 M⅐PSA Complex Enhances [ 3 H]Thymidine and [ 3 H]Leucine Uptake by 1-LN Cells-In
preliminary experiments, we first studied the effect of varying concentrations of ␣ 2 M⅐methylamine and synthesized ␣ 2 M⅐PSA complexes on [ 3 H]thymidine incorporation into 1-LN prostate cancer cells (Fig. 5A). The maximum incorporation of [ 3 H]thymidine (fmol/mg cellular protein) into cells was observed at about a 50 pM concentration of both ␣ 2 M⅐methylamine and premade ␣ 2 M⅐PSA (Fig. 5A). Treatment of cells with higher concentrations of both forms of ␣ 2 M* resulted in loss of [ 3 H]thymidine incorporation (Fig.  5A). The observed kinetics of [ 3 H]thymidine incorporation into cells is similar to that reported earlier (22). Therefore, in experiments described below, we employed 50 pM ␣ 2 M*. In the next series of experiments, we measured DNA synthesis in 1-LN, DU-145, and PC-3 prostate cancer cells stimulated with premade ␣ 2 M⅐PSA or ␣ 2 M⅐PSA derived from the 1-LN cultured cells to which native ␣ 2 M was added (Fig. 5, B-D). Both ␣ 2 M⅐PSA preparations up-regulated DNA (Fig. 5, B and C) by about 2-fold in 1-LN and DU-145 cells compared with cells treated with buffer, native ␣ 2 M, PSA, or medium (Fig. 5, B-D). In contrast, DNA synthesis in PC-3 cells was unaltered upon treatment with ␣ 2 M* or anti-CTD antibody (Fig. 5E). Furthermore, the effects of ␣ 2 M⅐PSA complexes on DNA and protein synthesis were nearly identical to those of ␣ 2 M⅐methylamine (Fig. 5, B-D) observed here and reported elsewhere (11)(12)(13)(14). Pretreatment of cancer cells with anti-CTD antibody nearly abolished DNA and protein synthesis induced by all activated ␣ 2 M* preparations in prostate cancer cells (Fig. 5, B-D). Likewise, pretreatment of cells with actinomycin D abrogated all activated ␣ 2 M-induced DNA synthesized (Fig. 5B). A similar effect of ␣ 2 M⅐PSA preparations on protein synthesis comparable with that of ␣ 2 M⅐methylamine was observed in these cells (Fig. 5C). Like DNA synthesis, increased protein synthesis in 1-LN cells induced by all activated ␣ 2 M preparations was abolished by pretreatment of cells with either anti-CTD antibody or actinomycin D (Fig. 5C).

DISCUSSION
Controversy exists with respect to the value of measuring PSA levels to screen men for the occurrence of previously undetected prostate cancer (1-6). However, PSA levels do correlate with disease progression (1,6), and targeting PSA may have value in treating the disease (7). The latter observation is intriguing in that it suggest that PSA synthesis by prostate cancer cells represents more than a disease marker. In the present study, we provide a hypothesis as to why production of PSA may correlate with disease progression. In a number of studies, we have demonstrated that cell surface GRP78 is a receptor whose activation triggers proproliferative, promigratory, and antiapoptotic signaling cascades in prostate cancer cells (11)(12)(13)(14)(15)(16)22). Moreover, autoantibodies to GRP78 in the serum of men with prostate cancer indicate a poor prognosis (16 -18). Mechanistically, these autoantibodies are agonists that bind to a region in the NH 2 -terminal domain of GRP78 where the natural ligand ␣ 2 M* also binds (16). Binding of either ␣ 2 M* or these autoantibodies promotes autophosphorylation of GRP78. This then triggers Ras/MAPK, phosphatidylinositol 3-kinase/Akt, cAMP-dependent, and other signaling pathways that are proproliferative, promigratory, and antiapoptotic (11)(12)(13)(14)(15)(16)(22)(23)(24)(25)(26). Interestingly, ligation of cell surface GRP78 promotes its own up-regulation (27). At least one transcription factor, TFII-I has been identified as a promoter of GRP78-dependent up-regulation of various cellular mechanisms (27). By contrast, antibodies directed against the COOH-terminal domain of GRP78 activate pro-apoptotic and anti-proliferative signaling events (23)(24)(25)(26). Here we show that ligation of GRP78 by ␣ 2 M* promotes the synthesis and secretion of PSA in an active form that binds to ␣ 2 M, producing ␣ 2 M⅐PSA complexes. These complexes, as previously shown with ␣ 2 M⅐methylamine, promote DNA and protein synthesis by activation of Akt-dependent and other signaling cascades (11)(12)(13)(14)(15)(16)22). Thus, ligation of the NH 2 -terminal domain of GRP78 by ␣ 2 M* leads to the synthesis of more ␣ 2 M*, which in turn drives GRP78 receptor-dependent proproliferative, promigratory, and antiapoptotic signaling mechanism. We propose that PSA is involved in an autoregulatory feedback loop to the advantage of the prostate cancer but to the disadvantage of the host. These observations are summarized in Fig. 9. We have previously demonstrated that there is considerable "cross-talk" between Akt, NFB, and unfolded protein response signaling, all of which are activated in 1-LN prostate cancer cells treated with ␣ 2 M* (15). Up-regulation of NFB signaling is a well characterized mechanism that up-regulates PSA synthesis (28). By contrast, up-regulation of p53 downregulates PSA synthesis (29). Our results with antibody directed against the COOH-terminal domain in the present study are consistent with these results. Treatment of 1-LN prostate cancer cells with this antibody up-regulates p53 (23,25), which should down-regulate PSA synthesis as noted above (29).
These studies may shed light on the role of PSA in the progression of prostate cancer. As noted above, the proteinase activity of PSA may be involved in bone remodeling, that characteristic of prostate cancer metastasis (8). PSA is a typical chymotrypsin-like serine proteinase (30). However, other studies suggest that PSA may regulate pro-survival signaling by an unknown mechanism (31). The current study suggests a mechanism for these observations.  These observations raise several other questions. The concentration of ␣ 2 M in the serum ranges from 1 to 5 M (10). It is produced by many tissues, including prostate stroma (9). Patients with prostate cancers often demonstrate ␣ 2 M levels at the lower end of this range, presumably reflecting proteinase-dependent turnover, not a change in synthetic rate (32)(33)(34). Consistent with this view, ␣ 2 M⅐PSA complexes are detected in the serum of patients with prostate cancer (1, 3). Because the t1 ⁄ 2 for the clearance of ␣ 2 M-proteinase complexes is on the order of 2 min (10), the detection of ␣ 2 M⅐PSA complexes in serum suggests a very high rate of synthesis.
Taken together, these observations and the results of the present study suggest a high likelihood that the proposed proproliferative/antiapoptotic feedback loop exists in prostate cancers. Finally, we have demonstrated that the binding of ␣ 2 M* to GRP78 demonstrates a K d of 50 -100 pM, whereas binding to LRP occurs with a K d of ϳ1-5 nM (10,11). Our studies demonstrate that LRP is not required for GRP78-dependent signal transduction (11). However, recently, Gonias and colleagues (35) have demonstrated that at very high concentrations of ␣ 2 M*, on the order of 100 nM, proproliferative signaling may be seen. Although one could conceive of such high concentrations in the vicinity of growing prostate cancer cells, they are very high. We therefore suggest that this mechanism is not likely to be a major impact in the biology of prostate cancer. It should be noted that this pathway is irrelevant in the present studies because the 1-LN prostate cell line employed in this study expresses GRP78 in its surface but not LRP (11,36).