Activated α2-Macroglobulin Binding to Human Prostate Cancer Cells Triggers Insulin-like Responses

Background: Ligation of cancer cell surface GRP78 by activated α2-macroglobulin (α2M*) promotes proliferation and blocks apoptosis. Results: α2M* treatment of prostate cancer cells enhances the Warburg effect and up-regulates lipid metabolism in an insulin-like manner. Conclusion: α2M* exerts insulin-like effects on prostate cancer cells. Significance: Targeting cell surface GRP78-dependent cancer cell regulation, such as by antibodies, offers a unique potential therapeutic approach. Ligation of cell surface GRP78 by activated α2-macroglobulin (α2M*) promotes cell proliferation and suppresses apoptosis. α2M*-treated human prostate cancer cells exhibit a 2–3-fold increase in glucose uptake and lactate secretion, an effect similar to insulin treatment. In both α2M* and insulin-treated cells, the mRNA levels of SREBP1-c, SREBP2, fatty-acid synthase, acetyl-CoA carboxylase, ATP citrate lyase, and Glut-1 were significantly increased together with their protein levels, except for SREBP2. Pretreatment of cells with α2M* antagonist antibody directed against the carboxyl-terminal domain of GRP78 blocks these α2M*-mediated effects, and silencing GRP78 expression by RNAi inhibits up-regulation of ATP citrate lyase and fatty-acid synthase. α2M* induces a 2–3-fold increase in lipogenesis as determined by 6-[14C]glucose or 1-[14C]acetate incorporation into free cholesterol, cholesterol esters, triglycerides, free fatty acids, and phosphatidylcholine, which is blocked by inhibitors of fatty-acid synthase, PI 3-kinase, mTORC, or an antibody against the carboxyl-terminal domain of GRP78. We also assessed the incorporation of [14CH3]choline into phosphatidylcholine and observed similar effects. Lipogenesis is significantly affected by pretreatment of prostate cancer cells with fatostatin A, which blocks sterol regulatory element-binding protein proteolytic cleavage and activation. This study demonstrates that α2M* functions as a growth factor, leading to proliferation of prostate cancer cells by promoting insulin-like responses. An antibody against the carboxyl-terminal domain of GRP78 may have important applications in prostate cancer therapy.

Prostate cancer, the most commonly diagnosed malignancy of men, may acquire androgen independence, a poor prognostic indicator associated with a more metastatic phenotype (1)(2)(3). Growth factors play a role in the progression of these cancers by ligating cell surface-binding sites to induce receptor autophosphorylation at specific tyrosine residues. This results in the assembly of multiprotein complexes, which activate sig-nal transduction mechanisms, promoting tumor cell proliferation (4 -9). Activation of PI 3-kinase results in its recruitment to the inner surface of the plasma membrane, where it generates 3Ј-phosphorylated phosphoinositides, which regulate downstream signal transduction by activating Akt, mTORC1, and mTORC2 (4 -12).
␣ 2 -Macroglobulin (␣ 2 M) 2 is a broad spectrum proteinase inhibitor synthesized by hepatocytes and macrophages (13)(14)(15)(16)(17); however, tumors and their associated stroma also produce this protein (18). Prostate cancer cells secrete prostate-specific antigen, a chymotrypsin-like serine proteinase, metalloproteinases, and other proteinases, which activate ␣ 2 M in the tumor microenvironment (18). When ␣ 2 M reacts with a proteinase, the "bait region" in each subunit is cleaved, and consequently, ␤-cysteinyl-␥-glutamyl thiol esters in all four of its subunits rupture, followed by a large conformational change in the molecule (13). This exposes a receptor recognition site in the carboxyl-terminal domain of each subunit (19). Small nucleophiles, such as methylamine or ammonia, can directly rupture the thiol esters, triggering a similar conformational change exposing these sites (13). Two receptors bind activated forms of ␣ 2 M (␣ 2 M*), the LDL receptor-related protein (LRP) and cell surface-associated GRP78 (glucose-regulated protein of M r ϳ78,000) (20 -23). Although GRP78 is primarily a resident endoplasmic reticulum chaperone (24), it appears on the cell surface of many types of malignant cells, including human prostate cancer (20 -23), where it functions as a very high affinity ␣ 2 M* signaling receptor linked to pro-proliferative, pro-migratory, and anti-apoptotic signaling cascades (6 -9, 22, 23, 25-39). Patients with many forms of cancer, including prostate, may develop auto-antibodies that bind to the amino-terminal domain of GRP78 close to or at the ␣ 2 M*-binding site (40 -42). These antibodies are receptor agonists in contrast to antibodies directed against the carboxyl-terminal domain of GRP78, which are receptor antagonists and promote cell death and apo-ptosis by impairing PI 3-kinase/PDK/Akt/mTORC1/mTORC2 signaling (18, 23, 25, 29, 36, 38 -45).
Cancers undergo metabolic reprogramming, most importantly effecting energy metabolism (46 -54). Cancer cells show an increased uptake of glucose that is decoupled from pyruvate oxidation via aerobic glycolysis, resulting in lactic acid accumulation, known as the Warburg effect (55). Cancer cells also demonstrate alterations of fatty acid, triglyceride, and cholesterol metabolism (56). Irrespective of the concentration of extracellular lipids, cancer cells synthesize and accumulate fatty acids de novo in part because of the elevated expression of fatty-acid synthase, a key metabolic enzyme catalyzing the synthesis of long chain fatty acids (46 -54, 57-64). Furthermore, fatty acid oxidation is a dominant pathway for energy generation in many tumors (65). PI 3-kinase/Akt/mTORC signaling stimulates fatty acid synthesis by activating ATP citrate lyase, and it stimulates lipogenic gene expression via activation and nuclear localization of the transcription factor SREBP1 (sterol regulatory element-binding protein) (64, 66 -71). Inhibition of ATP citrate lyase induces growth arrest and apoptosis in prostate cancer cells (72). Cholesterol accumulation also occurs in prostate cancer, and dysregulation of its biosynthetic pathway is associated with malignant transformation (59,(73)(74)(75). Cholesterol is an important component of biological membranes because it modulates the fluidity of lipid bilayers and forms lipid rafts that coordinate the activation of certain signal transduction pathways (59,(73)(74)(75). The intracellular pool of cholesterol esters is a storage form of cholesterol that prevents its toxic effects (76). The accumulation of cholesterol esters is induced by the loss of PTEN, up-regulation of the PI 3-kinase/Akt/ mTORC pathway, and activation of SREBP. Whereas SREBP1 mainly regulates fatty acid, phospholipid, and triacylglyceride biosynthesis, SREBP2 regulates cholesterol biosynthesis (77). SREBPs traffic to the Golgi apparatus where they are processed by two proteinases to liberate a soluble fraction that translocates to the nucleus. Here, SREBPs activate transcription by binding to sequences in the promoters of target genes. Insulinmediated stimulation of SREBP1-c processing and SREBP1-c mRNA induction requires PI 3-kinase/Akt/mTORC1 signaling, and either rapamycin or PI 3-kinase inhibitors block its activation (47,49,59,64,68,69,78,79).
Glucose-derived carbons are channeled into fatty acids, which are incorporated into glycerolipids (46 -59, 80, 81). Fatty-acid synthase inhibition decreases tumor growth by suppressing the synthesis of phosphatidylcholine and other phospholipids necessary for membrane biogenesis, lipid raft formation, and the production of proactive lipids (80,81). The hydrolysis of phosphatidylcholine mediates mitogenic signal transduction events in cells, and the products of its metabolism, such as diacylglycerol and arachidonic acid metabolites, are second messengers essential for mitogenic activity. Previous studies demonstrate that ␣ 2 M* up-regulates the synthesis and activity of cPLA 2 , phospholipase D, and COX-2 (82)(83)(84).
We previously reported that binding of ␣ 2 M* to GRP78 on the surface of various tumor cells, including prostate cancer, induces proliferation and survival by activating PI 3-kinase/ Akt/mTORC signaling. In this study, we determined whether ␣ 2 M* enhances the Warburg effect in prostate cancer cells causing proliferation. We report here that ␣ 2 M* up-regulates aerobic glycolysis in prostate cancer cells as determined by increased glucose uptake, increased lactate secretion, and upregulation of Glut-1 in the presence of oxygen. The synthesis of fatty acids, cholesterol, triglycerides, and phosphatidylcholine with corresponding increases in the expression of SREBP1-c, SREBP2, ATP citrate lyase, and acetyl-CoA carboxylase is observed. Treatment with an antibody directed against the carboxyl-terminal domain of GRP78 (anti-CTD) inhibits ␣ 2 M*-induced cell proliferation and lipogenesis as determined by studies with 1-[ 14 C]acetate, 6-[ 14 C]glucose, and [ 14 CH 3 ]choline. A similar effect was seen with inhibitors of PI 3-kinase, Akt-1, mTORC1, mTORC2, fatty-acid synthase, and SREBP activation with either ␣ 2 M* or insulin treatment, the latter was used as a positive control. The effects of ligating cell surface GRP78 with ␣ 2 M* are strikingly similar to the effects induced by insulin.

EXPERIMENTAL PROCEDURES
Materials-Culture media were purchased from Invitrogen. ␣ 2 M* was prepared as described previously (32). Antibodies against fatty-acid synthase, acetyl-CoA carboxylase, and SREBP1-c were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The antibody against the carboxyl-terminal domain of GRP78 (anti-CTD) was purchased from Aventa Biopharmaceutical Corp. (San Diego). Antibodies against ␤-actin were from Sigma. Cancer Cell Lines-We used two prostate cancer cell lines that express GRP78 on their cell surface, 1-LN and DU-145 (32). 1-LN cells originally were the kind gift of Dr. Phillip Walther, Department of Surgery, Duke University Medical Center. They now reside in our laboratory and are available on request. The DU-145 cell line was obtained from ATCC (Manassas, VA). These cells were grown as needed in multiwell plates (300 ϫ 10 3 cell/well) in RPMI 1640 medium containing 10% fetal bovine serum, 2 mM glutamine, 12.5 units of penicillin/ml, 6.5 g/ml streptomycin, and 10 nM insulin (RPMI-S) at 37°C in a humidified CO 2 (5%) incubator. At ϳ90% confluency, the medium was aspirated, and the monolayers were washed with ice-cold Hanks' balanced salt solution containing 10 mM HEPES (pH 7.4) and 3.5 mM NaHCO 3 (HHBSS) before use in the studies described below. Although the focus of this and previous studies has been androgen-independent tumors, in some experiments we employed hormone-dependent LnCap prostate cancer cells that, like 1-LN and DU-145 cells but unlike PC-3 cells, express GRP78 on their cell surface. LnCap cells were obtained from ATCC and grown as above. Fatty-acid Synthase Inhibitors-1-LN and DU-145 cells grown  in 48-well plates were preincubated with the Akt1 inhibitor  MKK-2206 (8 nM/60 min), fatty-acid synthase inhibitors C-75 (20 g/ml/2 h), orlistat (10 M/2 h), and mTORC1 and mTORC2 inhibitors KU0036794 (1 M/1 h) followed by the addition of ␣ 2 M* (100 pM) or insulin (200 nM) to the respective wells. [ 3 H]Leucine (2 Ci/ml) was added to the cells and incubated overnight. The reactions were terminated by aspirating the medium, and the cells were washed twice with ice-cold 5% TCA, followed by three washes with ice-cold PBS. The cells were lysed in a volume of NaOH (40°C/2 h). Protein was estimated in an aliquot (85), and the lysates were counted in a liquid scintillation counter (32).

Protein Synthesis in Prostate Cancer Cells Stimulated with ␣ 2 M* or Insulin and the Effect of PI 3-Kinase/Akt/mTORC and
[ 3 H]Thymidine Incorporation in Prostate Cancer Cells Stimulated with ␣ 2 M* or Insulin and the Effect of PI 3-Kinase/Akt/mTORC or Fatty-acid Synthase Inhibitors-1-LN and DU-145 cells in 48-well plates were grown and processed as described above except that [ 3 H]thymidine (2 Ci/ml) was added to the cells. The reactions were stopped by aspirating the medium, and the cells were washed twice with ice-cold 5% TCA, followed by three washes with ice-cold PBS. The cells were lysed in a volume of 1 M NaOH, the protein estimated in an aliquot (85), and samples were counted in a liquid scintillation counter. The effects of inhibitors on protein and DNA synthesis were evaluated by incubating prostate cancers with these inhibitors alone, and incorporation of [ 3 H]leucine or [ 3 H]thymidine into protein and DNA was determined as above. Inhibitor effects on prostate cancer cell proliferation was also evaluated by counting the cells in a hemocytometer at "0" time and after 16 h of incubation.
Glucose Uptake by Prostate Cancer Cells Stimulated with ␣ 2 M* or Insulin-1-LN and DU-145 cells were grown in 6-well plates as described above in serum-free medium. The serumfree medium was aspirated, and DMEM containing 25 mM glucose was added followed by the addition of buffer, ␣ 2 M*, insulin, or anti-CTD prior to the addition of ␣ 2 M*. A final concentration of 1 Ci/ml [ 14 C]deoxyglucose was then added. The cells were incubated as described above for 0 and 60 min. The reactions were terminated by aspirating the medium. The cells were washed with ice-cold PBS, and the cells were lysed in lysis buffer containing 50 mM Tris⅐HCl (pH 7.5), 120 mM NaCl, 1% Nonidet P-40, 2.5 mM sodium fluoride, 1 mM sodium phosphate, 0.1 mM sodium orthovanadate, 1 mM PMSF, 1 mM benzamidine, and leupeptin (20 g/ml) (32). The samples were counted in a scintillation counter.
Lactate Secretion in 1-LN and DU-145 Cells Stimulated with ␣ 2 M* or Insulin-Lactate secretion (3 ϫ 10 6 cells/well) was determined in 6-well plates according to the manufacturer's instructions using a kit (Enzo Life Sciences, Farmingdale, NY). The cells incubated overnight in DMEM were stimulated with buffer, ␣ 2 M*, insulin, or anti-CTD prior to ␣ 2 M*. Lactate secretion was measured in aliquots of the medium removed at 0 and 60 min of incubation.
Real Time PCR of SREBP1, SREBP2, Fatty-acid Synthase, ATP Citrate Lyase, Acetyl-CoA Carboxylase, and Glut-1 in Prostate Cancer Cells Stimulated with ␣ 2 M* or Insulin-1-LN and DU-145 prostate cancer cells were grown in 6-well plates and washed twice with HHBSS. The cells were exposed to either buffer, ␣ 2 M*, or insulin and incubated as described above. The reactions were terminated by aspirating the medium. Total cellular RNA was extracted using the RNeasy minikit (Qiagen, Valencia, CA) according to the manufacturer's instructions and quantified by absorbance at 260 nm. Total RNA (500 ng) was pre-digested with DNase I (Invitrogen) for 15 min and used for DNA synthesis with RNase H reverse transcriptase (Invitrogen) and random primers according to the manufacturer's instructions. Real time formation of product was monitored in a Stratagene Mx3005P system (Agilent Technologies Inc., Santa Clara, CA) using SYBR Green I detection under the following conditions MxPro QPCR software (Stratagene Agilent Technologies, Santa Cruz, CA) was used to select a threshold level of fluorescence in the linear log phase of the PCR product accumulation. The expression of target genes was normalized to the expression of ␤-actin to determine the fold change over buffer-treated cells.
Effect of Silencing GRP78 Gene Expression in 1-LN and DU-145 Cells on ␣ 2 M*-induced Fatty-acid Synthase and ATP Citrate Lyase Expression-To determine the requirement of cell surface GRP78 for ␣ 2 M*-induced up-regulation of lipogenesis, we silenced the expression of GRP78 by RNAi. In our earlier publications, we used two GRP78-targeting RNA sequences for gene expression silencing and found identical effects on GRP78 expression and downstream signaling (29). We also found that transfection of cells with GRP78 dsRNA down-regulates both the total cellular GRP78 pool and cell surface-localized GRP78 by ϳ60 -65% (29). Therefore, in this study, we used only one GRP78-targeting RNA sequence to silence the expression of GRP78. For the GRP78 dsRNA, sense (5Ј-AAA ACA GCA AUU AGU AAA GTT-3Ј) and antisense (5Ј-CUU UAC UAA UUG CUG UAU UTT-3Ј) oligonucleotides were prepared and annealed by Ambion (Austin, TX). Transfection of 1-LN and DU-145 cells with GRP78 dsRNA was performed as described previously (29). Down-regulation of GRP78 upon transfection with GRP78 dsRNA was evaluated by Western blotting. Briefly, 1-LN and DU-145 cell monolayers (1.5 ϫ 10 6 cells/well in 6-well plates) were incubated as described above and transfected with 100 nM annealed GRP78 dsRNA, and control cells were transfected with Lipofectamine as described pre-viously (29). Forty eight hours after transfection, the control cells were stimulated with either buffer or ␣ 2 M*. Cells transfected with scrambled dsRNA (100 nM) and treated with ␣ 2 M* were used as the control. The reactions were terminated by aspirating the medium, and the cells were lysed in lysis buffer on ice for 15 min. The lysates were then transferred to tubes and centrifuged, and the protein concentration in the supernatant was determined. Equal amounts of lysate protein were electrophoresed, transferred to a membrane, and immunoblotted with anti-fatty-acid synthase or ATP citrate lyase antibodies. The protein bands were detected and quantified by enhanced chemifluorescence and phosphorimaging. The membrane was reprobed for the protein loading control, actin. Consistent with previous studies, the GRP78 protein was reduced 60 -70%.

1-[ 14 C]Acetate Incorporation into Lipids of Prostate Cancer Cells Stimulated with ␣ 2 M* or Insulin and the Effect of Pretreatment with PI 3-Kinase/Akt/mTORC and Fatty-acid Synthase
Inhibitors-1-LN and DU-145 cells (3 ϫ 10 6 cells/well) incubated overnight in 6-well plates were washed with RPMI-S medium, and RPMI-S medium was added prior to the addition of 1-[ 14 C]acetate (1 Ci/well). The cells were incubated for 4 h as described above. The reactions were terminated by aspirating the medium, and the cells were washed three times with ice-cold PBS, and RPMI-S medium was added to each well. The cells were pretreated with either LY294002, rapamycin, torin, KU0036795, C-75, or anti-CTD and then stimulated with buffer, ␣ 2 M*, or insulin. The reactions were terminated by aspirating the medium, and the cells were washed three times with ice-cold buffer and lysed in lysis buffer for 15 min over ice. The cell lysates were transferred into Eppendorf tubes and centrifuged at 4°C for 5 min at 1000 rpm; the supernatant was transferred to new Eppendorf tubes, and their protein content was determined. Equal amounts of lysate proteins (250 g) were used for lipid extraction. The lipids were extracted in screw cap glass tubes by adding chloroform/methanol/water (2:1:0.5 v/v/v) to the lysates. The mixture was vortexed and incubated at 4°C overnight (86). The tubes were centrifuged at 1000 rpm for 5 min at 4°C, and the aqueous phase was removed. The organic phase was washed with 1 ml of water by centrifugation as described above. The aqueous phase was removed, and the organic phase was dried under N 2 . The lipid residues were dissolved in chloroform, and an aliquot was counted in a liquid scintillation counter. The effect of inhibitors alone on lipogenesis from 1-[ 14 C]acetate in 1-LN and DU-145 cells was evaluated by incubating these cells with the inhibitors under identical conditions.

1-[ 14 C]Acetate Incorporation into Nonsaponifiable (Cholesterol) and Saponifiable (Fatty Acid) Fractions of Prostate Cancer Cells Treated with ␣ 2 M* or Insulin and the Effect of PI 3-Kinase/Akt/mTORC and Fatty-acid Synthase
Inhibitors-1-[ 14 C]Acetate (1 Ci/ml/4 h)-labeled 1-LN and DU-145 cells were pretreated with inhibitors of PI 3-kinase/Akt/mTORC and fatty-acid synthase and stimulated with ␣ 2 M* or insulin as described above. Lysate protein, 250 g, was used for lipid extraction as above. The lipid extracts in glass tubes were evaporated to dryness under N 2 . The dried lipid extracts were saponified by refluxing with 30% alcoholic KOH for 30 min at 70°C in a water bath. The saponification was terminated by adding water, and the nonsaponifiable lipid was extracted three times with petroleum ether (bp. 40 -60°C). The combined lipid extracts were washed with water. The lipid extracts were evaporated to dryness under N 2 , and the dried lipid residue was dissolved in chloroform, and an aliquot of the lipid extract was counted in a liquid scintillation counter. The alkali digest remaining after extraction of the nonsaponifiable fraction was acidified to pH 2 with 6 N H 2 SO 4 , and the fatty acids were extracted three times with petroleum ether (bp 40 -60°C). The combined extracts were washed with water. The lipid extracts were evaporated to dryness under N 2 , and the dried lipid residue was dissolved in chloroform, and an aliquot was counted in a liquid scintillation counter.
In a separate set of studies, the lipid extracts were evaporated to dryness under N 2 , and the lipid residues were dissolved chloroform and subjected to TLC fractionation on Silica Gel H-coated plates, preheated at 110°C for 1 h. An aliquot of the respective lipid extracts and the lipid standards, cholesterol ester, tripalmitin, palmitic acid, cholesterol, and phosphatidylcholine were applied to the TLC plates. The plates were developed to half the distance in CHCl 3 /CH 3 OH/H 2 O (60:30:5 v/v/ v). The plates were removed, air-dried, and again developed to half-distance in the same solvent system. The plates were airdried and developed to the full distance in hexane/ether/acetic acid (80:20:1.5 v/v/v/). The plates were dried and exposed to I 2 in a closed glass chamber in a chemical hood until the lipid spots appeared. The lipid spots on the plates corresponding to the lipid standards, cholesterol ester (R f value 0. . The reactions were terminated by aspirating the medium and washed three times with ice-cold PBS. The monolayers were lysed in lysis buffer containing 50 mM Tris⅐HCl (pH 7.5), 120 mM NaCl, 1% nonidet-P-40, 2.5 mM sodium fluoride, 1 mM sodium phosphate, 0.1 mM sodium orthovanadate, 1 mM PMSF, 1 mM benzamidine, leupeptin 20 g/ml) and proteaseinhibitor tablets (1 tablet/10 ml) (Roche Applied Science) over ice for 15 min. The lysates were transferred to Eppendorf tubes, centrifuged at 1000 rpm/5 min/4°C; supernatants were removed into new tubes and their protein contents determined. Equal amounts of lysate proteins were electrophoresed on 10% gels, and protein bands were transferred to PVDF membranes that were immunoblotted with anti-SREBP1 antibody. SREBP1 cleavage was determined by the generation of cleaved SREBP1, and the molecular mass of ϳ74 kDa from was SREBP1 (molecular mass ϳ151 kDa).

Effect of Incubation Time on Expression of SREBP1-c, ATP Citrate Lysate, and Fatty-acid Synthase in Prostate Cancer Cells
Stimulated with ␣ 2 M*-1-LN and DU-145 cells were stimulated with buffer or ␣ 2 M* and incubated for 0, 10, 20, 40, and 60 min as described above. The reactions were terminated by aspirating the medium, and the cells were lysed in lysis buffer for 15 min on ice. The cell lysates were transferred to Eppendorf tubes and briefly centrifuged. Then the supernatants were transferred to new tubes and their protein contents determined. A volume of 4ϫ sample buffer was added to equal amounts of protein lysates, contents were mixed and boiled for 5 min, electrophoresed, and then transferred to PVDF membranes. The respective membranes were immunoblotted with antibodies against SREBP1-c, ATP-citrate lyase, or fatty-acid synthase. The bands were visualized and quantified with a PhosphorImager. The membranes were reprobed for actin as the protein loading control. The specificity of the antibodies used in this study was determined by treating the cells with nonimmune antibodies of the appropriate species and isotype, and the cell lysates were processed as described above. Under the experimental conditions, no reactivity of these antibodies was observed.

Effect of Inhibiting the Proteolytic Cleavage of SREBPs by Fatostatin A on Cell Proliferation in Prostate Cancer
Cells Stimulated with ␣ 2 M* or Insulin-1-LN, DU-145, and LnCap cells in 48-well plates were grown as described above. Fatostatin A (20 M/2 h) or anti-CTD was added to the wells, and the cells were incubated at 37°C in a humidified incubator (5% CO 2 ). At certain time points, the cells were stimulated with buffer, ␣ 2 M*, or insulin, followed immediately by the addition of [ 3 H]leucine (2 Ci/ml) or [ 3 H]thymidine (2 Ci/ml) and incubated overnight. The incorporation of [ 3 H]leucine into the cell proteins and of [ 3 H]thymidine into DNA was measured as described above.
Effect of Fatostatin A on ␣ 2 M* or Insulin-stimulated Lipogenesis from 1-[ 14 C]Acetate in Prostate Cancer Cells-1-LN, DU-145, and LnCap cells in 6-well plates were prelabeled with 1-[ 14 C]acetate (1 Ci/ml/4 h) and washed twice with ice-cold PBS, followed by the addition of RPMI-S medium to each well. Fatostatin A or anti-CTD was added to the wells and incubated. Then buffer, ␣ 2 M*, or insulin was added, and the cells were incubated. The reactions were terminated by aspirating the medium, and the cells were washed with PBS and lysed in lysis buffer on ice for 15 min. The incorporation of 1-[ 14 C]acetate into cellular lipids was measured as described above.
Effect of Fatostatin A Treatment on Phosphatidylcholine Synthesis in Prostate Cancer Cells Stimulated with ␣ 2 M* or Insulin-1-LN and DU-145 cells in 6-well plates were labeled with [ 14 CH 3 ]choline (4 Ci/ml/4 h) (87). The reactions were terminated by aspirating the medium, and the cells were washed twice with cold PBS, and RPMI-S medium was added to each well. In separate experiments, fatostatin A or anti-CTD was added to the wells. After incubation, the cells were stimulated with either buffer, ␣ 2 M*, or insulin and incubated as described above. The reactions were terminated by aspirating the medium, and the cells were lysed in lysis buffer on ice for 15 min. Lipids from the cell lysates were extracted as described above. The incorporation of [ 14 CH 3 ]choline into phosphatidylcholine was measured by TLC as described above.
Statistical Analysis-The statistical significance of the data was determined by Student's t test. The values significantly different at the 5% level are denoted by an asterisk.

RESULTS
Fatty-acid Synthase or PI 3-Kinase/Akt/mTORC Signaling Pathway Inhibitors Suppress ␣ 2 M* or Insulin-induced Protein and DNA Synthesis-We first assessed whether lipogenesis is involved in ␣ 2 M*-induced cell proliferation by pretreating 1-LN and DU-145 cells with specific inhibitors of Akt1, mTORC1, mTORC2, or fatty-acid synthase, because these regulate cell proliferation by modulating SREBP processing and fatty-acid synthase activity. We used MKK-2206, a specific inhibitor of Akt1, KU0063794, an inhibitor of mTORC1 and mTORC2, C-75, which inactivates the ␤-ketoacyl reductase and thiol esterase activities of fatty-acid synthase, and orlistat, which inactivates fatty-acid synthase by forming a covalent adduct with the thiol esterase domain of the enzyme. Stimulation of 1-LN and DU-145 cells with ␣ 2 M* caused an ϳ1.5-2fold increase in protein and DNA synthesis, which was similar to insulin up-regulated protein (Fig. 1, A and B) and DNA (Fig.  1, C and D) synthesis. The data are corroborated by cell growth studies that show similar results to these findings (Fig. 1, E and  F). These inhibitors at the concentration used did not induce cell death as determined by incorporation of [ 3 H]leucine into protein and [ 3 H]thymidine into DNA as compared with buffertreated cells (Fig. 1, G and H). Down-regulation of ␣ 2 M* and insulin-dependent protein and DNA synthesis in these cells by inhibitors demonstrates that Akt1 and mTORC signaling and fatty-acid synthase up-regulation are necessary for macromolecular synthesis.
␣ 2 M* Up-regulates the Expression of SREBP1-c, Fatty-acid Synthase, and ATP Citrate Lyase in 1-LN and DU-145 Prostate Cancer Cells-Because inhibition of fatty-acid synthase suppresses protein and DNA synthesis in 1-LN and DU-145 cells stimulated with ␣ 2 M* or insulin, we next evaluated SREBP1-c, ATP citrate lyase, and fatty-acid synthase expression in these cells by Western blotting (Fig. 2, A and B). Stimulation of 1-LN and DU-145 cells with ␣ 2 M* caused a 1.5-2.5-fold up-regulation of SREBP1, SREBP2, fatty-acid synthase, and ATP citrate lyase in 1-LN ( Fig. 2A) and DU-145 cells (Fig. 2B). We do not understand the decline in SREBP1 and ATP citrate lyase expression in ␣ 2 M*-stimulated 1-LN and DU-145 cells after 15 min of incubation. Most likely it is due to their rapid turnover and/or precursor-product relationship in a downstream cascade causing activation of lipogenic responses. We also measured the transcriptional up-regulation of SREBP1-c, SREBP2, fatty-acid synthase, acetyl-CoA carboxylase, and ATP citrate lyase by assaying their mRNA levels by real time PCR. Treatment of 1-LN and DU-145 cells with ␣ 2 M* and insulin increased the mRNA expression relative to ␤-actin by ϳ1.5-2.5-fold compared with buffer-treated cells (Fig. 2, C and D). Although ␣ 2 M* and insulin caused a 1.5-2.5-fold increase in the mRNA levels, the mRNA expression varied among the genes studied. The expression level of SREBP1-c protein was very low compared with the other gene products (Fig. 2, C and  D). These results show that similar to insulin, ␣ 2 M* up-regulates lipogenic pathways in prostate cancer cells.
Elevated Glucose Uptake in Prostate Cancer Cells Treated with ␣ 2 M*-Prostate carcinoma cells show increased expression of Glut-1 and increased glucose uptake and utilization to support androgen-independent proliferation (53,88 fore, we measured glucose uptake in 1-LN and DU-145 prostate cells stimulated with ␣ 2 M* or insulin by incubating cells with 1-[ 14 C]deoxyglucose (Fig. 3, A and B). At 60 min of incubation, ␣ 2 M* and insulin-stimulated cells showed a 2-3-fold increase in glucose uptake compared with buffer-treated cells (Fig. 3, A  and B). Pretreatment of 1-LN and DU-145 cells with anti-CTD significantly reduced ␣ 2 M*-mediated glucose uptake. Glut-1 mRNA levels in both 1-LN and DU-145 cells treated with ␣ 2 M* or insulin increased 1.5-4-fold compared with buffer-treated cells (Fig. 3C). Expression of Glut-1 was more pronounced in 1-LN cells compared with DU-145 cells stimulated with ␣ 2 M* or insulin (Fig. 3C).
Increased Lactate Secretion in Prostate Cancer Cells Stimulated with ␣ 2 M* or Insulin-We next determined the secretion of lactate in prostate cells stimulated with ␣ 2 M* or insulin (Fig.  3D) (89 -91). ␣ 2 M* or insulin-stimulated cells showed an ϳ2-fold increase in lactate secretion compared with buffertreated cells. Importantly, anti-CTD blocked both the increased glucose uptake (Fig. 3, A and B) and lactate secretion (Fig. 3D) in 1-LN and DU-145 prostate cancer cells stimulated

GRP78, ␣ 2 -Macroglobulin, and Metabolic Regulation
with ␣ 2 M*. Previous studies of head and neck tumors have shown a correlation of high lactate levels with an increased incidence of metastasis (91). We hypothesize a similar behavior in prostate cancer.
Lipogenesis from 1-[ 14 C]Acetate Is Up-regulated in Prostate Cancer Cells Stimulated with ␣ 2 M* or Insulin-Next, we studied the activities of lipogenic enzymes in prostate cancer cells via the incorporation of the labeled precursors 1-[ 14 C]acetate and 6-[ 14 C]glucose into the total cellular lipids, cholesterol, fatty acids, and phosphatidylcholine. In both 1-LN and DU-145 cells, ␣ 2 M* and insulin up-regulated 1-[ 14 C]acetate incorporation into total lipids by 2-3-fold compared with buffer-treated cells (Figs. 3B and 4A). Pretreatment of cells with PI 3-kinase, mTORC1, mTORC2, or fatty-acid synthase inhibitors nearly abolished ␣ 2 M* and insulin-induced lipogenesis. Importantly, anti-CTD also abrogated ␣ 2 M*-induced lipogenesis in these cells. We also evaluated whether inhibitors used in this study caused reduced lipogenesis due to initiation of cell death by incubating cancer cells with inhibitors alone and processed cells for lipid extraction and counting for 1-[ 14 C]acetate incorporation. The incorporation was comparable with that observed in buffer-treated cells showing thereby that inhibitors did not suppress 1-[ 14 C]acetate incorporation into lipid by inducing cell death (data not shown).
Lipogenesis Is Up-regulated in Prostate Cancer Cells Stimulated with ␣ 2 M* or Insulin-We next studied the incorporation of 6-[ 14 C]glucose into the cellular lipids of 1-LN and DU-145 cells stimulated with ␣ 2 M* or insulin (Fig. 4, C and D). We also examined the modulation of 6-[ 14 C]glucose incorporation into cellular lipids using inhibitors of PI 3-kinase, Akt, mTORC1, mTORC2, fatty-acid synthase, or anti-CTD. Similar to lipogenesis from 1-[ 14 C]acetate, lipogenesis from 6-[ 14 C]glucose was up-regulated 2-3-fold in these cells. Pretreatment of cells prior to stimulation with ␣ 2 M* or insulin with PI 3-kinase/Akt/ mTORC1, mTORC2, or fatty-acid synthase inhibitors abrogated ␣ 2 M* or insulin-induced increased lipogenesis from 6-[ 14 C]glucose, consistent with the role of PI 3-kinase/Akt/ mTORC1 and mTORC2 signaling and fatty-acid synthase activity in lipogenic responses in these cells. Anti-CTD also showed a similar effect on ␣ 2 M* effects.

␣ 2 M* and Insulin Increase the Incorporation of 1-[ 14 C]Acetate into Nonsaponifiable and Saponifiable Lipids of 1-LN and DU-145
Prostate Cancer Cells-To access 1-[ 14 C]acetate conversion into fatty acids and cholesterol, we saponified cellular lipids and separated the nonsaponifiable fraction, which includes cholesterol, and the saponifiable fraction, which includes fatty acids from glycerides and phospholipids. To understand the role of PI 3-kinase/Akt/mTORC signaling in

GRP78, ␣ 2 -Macroglobulin, and Metabolic Regulation
the synthesis of these lipids, we pretreated prostate cancer cells with specific inhibitors of these signaling pathways prior to stimulation with ␣ 2 M* or insulin. Each caused a severalfold increase in 1-[ 14 C]acetate incorporation into both nonsaponifiable (Fig. 5, A and B) and saponifiable (Fig. 5, C and D) Fig. 6A. Cholesterol was the major lipid in both 1-LN and DU-145 prostate cancer cells representing ϳ75-80% of the total lipids, followed by phosphatidylcholine, which formed ϳ15% of total cellular lipids (Fig. 6A). The synthesis of esterified cholesterol from 1-[ 14 C]acetate-labeled cholesterol and fatty acyl-CoA was increased 2-2.5-fold in both 1-LN (Fig. 6, B and C) and DU-145 cells (Fig. 6, D and E) stimulated with ␣ 2 M* or insulin. Pretreatment of cells before stimulation with ␣ 2 M* or insulin combined with inhibitors of PI 3-kinase/Akt/mTORC and fatty-acid synthase significantly decreased the esterification of cholesterol (Fig. 6). Similarly, the ligation of GRP78 with anti-CTD reduced cholesterol esterification (Fig. 6).
Stimulation of 1-LN (Fig. 6, B and C) and DU-145 prostate cancer cells (Fig. 6, D and E) also caused an increased synthesis of triglycerides incorporating fatty acids synthesized from 1-[ 14 C]acetate (Fig. 6). However, the magnitude of synthesis varied in these cells stimulated with ␣ 2 M* or insulin. Synthesis of triglycerides and free fatty acids from 1-[ 14 C]acetate in 1-LN and DU-145 cells was significantly suppressed by inhibition of PI 3-kinase/Akt/mTORC signaling or fatty-acid synthase, as well as by anti-CTD treatment prior to ␣ 2 M* addition (Fig. 6, B  and C, and D and E).
Stimulation of 1-LN (Fig. 6, B and C) and DU-145 (Fig. 6, D and E) prostate cancer cells with ␣ 2 M* or insulin up-regulated cholesterol synthesis from 1-[ 14 C]acetate by about 2-3-fold compared with buffer-treated cells. This increased cholesterol synthesis causes the accumulation of cholesterol in prostate cancer cells under our experimental conditions (Fig. 6A). Cholesterol synthesis is regulated by PI 3-kinase/Akt/mTORC signaling and was demonstrated by inhibiting these pathways. Inhibition of fatty-acid synthase by C-75 nearly abrogated cholesterol synthesis. Down-regulation of GRP78 by anti-CTD also inhibited cholesterol synthesis. We next measured the synthesis of phosphatidylcholine in prostate cancer cells stimulated with ␣ 2 M* or insulin. We quantified the levels of 1-[ 14 C]acetate-labeled fatty acids incorporated into phosphatidylcholine by thin layer chromatography (Fig. 6B). Similar to insulin, ␣ 2 M*-treated prostate cancer cells up-regulated phosphatidylcholine synthesis 2-3-fold (Fig. 6). Pretreatment of prostate cancer cells with inhibitors of PI 3-kinase/Akt/mTORC or fatty-acid synthase significantly reduced ␣ 2 M*/insulin-induced increased phosphatidylcholine synthesis from 1-[ 14 C]acetate-labeled fatty acids.

Inhibition of SREBP1 Proteolytic Cleavage by Fatostatin A Suppresses Cell Proliferation, Lipogenesis, and Phosphatidylcholine Biosynthesis in Prostate Cancer Cells Stimulated with ␣ 2 M* or Insulin-Fatostatin
A inhibits proteolytic cleavage of SREBP1 and thus inhibits downstream signaling cascades (91,92). Stimulation of prostate cancer cells with ␣ 2 M*, similar to insulin, up-regulated protein synthesis 2-3-fold in 1-LN, DU-145, and LnCap cells compared with buffer-treated controls, although pretreatment of cells with fatostatin A nearly abolished these effects (Fig. 7, A and C). ␣ 2 M* and insulin also up-regulated [ 3 H]thymidine uptake, indicating increased DNA synthesis 2-3-fold, although pretreatment with fatostatin A nearly abrogated these effects (Fig. 7, B and  D). The effects of fatostatin treatment on SREBP1 cleavage is shown in Fig. 7, E and F. Fatostatin A inhibits proteolytic cleavage of SREBP1 and thus inhibits signaling cascades. We determined the effect of fatostatin A on SREBP1 cleavage by Western blotting in 1-LN and DU-145 cells. Cells stimulated with ␣ 2 M* or insulin demonstrated significant cleavage of SREBP1 (molecular mass ϳ151 kDa) to mature SREBP1 (molecular mass ϳ74 kDa).
Because of the critical regulation of lipogenesis by the transcription factor, SREBP, we next determined the total cellular lipogenesis from 1-[ 14 C]acetate in 1-LN and DU-145 prostate cancer cells pretreated with fatostatin A (Fig. 8, A and B). Fatostatin A significantly reduced 1-[ 14 C]acetate incorporation into total lipids of prostate cancer cells stimulated with ␣ 2 M* or insulin (Fig. 8, A and B). Pretreatment of 1-LN and DU-145 and LnCap prostate cancer cells with PI 3-kinase/Akt/mTORC and fatty-acid synthase inhibitors and fatostatin A inhibitor of SREBP cleavage significantly suppressed the ␣ 2 M* or insulininduced increase in total lipogenesis from 1-[ 14 C]acetate (Fig. 8,  A-C). PI 3-kinase/Akt/mTORC is upstream of SREBPs, and fatty-acid synthase, a target gene of SREBP1-c, is downstream; therefore, it is expected that impairment of SREBP1 activation and hence activation of target genes is significantly inhibited. These results are consistent with genetic manipulations of SREBPs and RNAi silencing of SREBPs, which impair lipogenesis and cell proliferation (46 -54).
Next, using thin layer chromatography, we showed that both ␣ 2 M* and insulin increased phosphatidylcholine synthesis 2-3fold from 1-[ 14 C]acetate-labeled acyl-CoAs in 1-LN and DU-145 prostate cancer cells, which was greatly suppressed by inhibitors of PI 3-kinase/Akt/mTOR and fatty-acid synthase (Fig. 6). SREBP is involved in de novo phosphatidylcholine synthesis (88,93,94). To further understand the regulation of de novo phosphatidylcholine by SREBP, we pretreated

DISCUSSION
Rapidly proliferating tumor cells exhibit increased glucose uptake and increased secretion of lactate in the presence of abundant oxygen, known as the Warburg effect, which is a hallmark of aggressive tumors (55). The yield of ATP by anaerobic glucose consumption is low, two molecules of ATP/molecule of glucose compared with the 38 ATP molecules/glucose obtained by the complete oxidation of glucose (55). The high glycolytic flux, however, allows the amount of cellular ATP produced from glycolysis to exceed that produced from oxidative phosphorylation (46 -64, 78). Activation of the PI 3-kinase/Akt/ mTORC pathways in tumor cells enhances the expression of genes and up-regulates enzyme activities causing increased glycolysis and lactate production (46 -64, 78).
␣ 2 M* binds to LRP and cell surface GRP78 with very distinct binding characteristics and cellular responses (18, 20 -23, 25-39). The K d value for ␣ 2 M* binding to GRP78 is 50 -100 pM compared with ϳ5 nM for LRP, whereas the receptor number of GRP78 is ϳ5000/cell compared with ϳ200,000/cell for LRP (87)(88)(89). At ligand concentrations near the K d value of the ligand, binding of ␣ 2 M* to GRP78 but not LRP activates phospholipase ␤/␥, generating an increase in [Ca 2ϩ ] i , inositol 1,4,5trisphosphate, and diacylglycerol (18, 20 -23, 25-39). GRP78 is expressed on rheumatoid, but not normal, synovial fibroblasts (32), 1-LN and DU-145, but not PC-3 prostate cancer cells, and highly activated but not resident peritoneal macrophages (6 -8,  18, 22, 32-39, 45-47, 95, 96). In all these cases, the effects correlate with the presence of GRP78 on the cell surface. ␣ 2 M* binding to cell surface GRP78 up-regulates protein and DNA synthesis, which is blocked by inhibitors of the PI 3-kinase/Akt/ mTORC signaling pathway. In contrast to amino-terminal domain-dependent pro-proliferative signaling, down-regulation of GRP78 by RNAi or ligation by anti-CTD induces apoptosis by up-regulating p53 levels and caspase activation (38,39). Because of the similarities observed between signaling events elicited in various cancer cell types by growth factors resulting in the Warburg effect, we ascertained whether ␣ 2 M*stimulated prostate cancer cells also exhibit aerobic glycolysis and associated metabolic alterations. The primary focus of these studies is on androgen-independent human cancer cells as in most of our previous studies. These are the most aggressive cancers with the highest metastatic potential. Indeed, it is these patients who are most likely to develop autoantibodies to the amino-terminal domains, which are receptor agonists (18,40,41). Nevertheless, in this study we have performed studies of the androgen-dependent LnCAP human prostate cancer line with similar results. We focused on glucose uptake, lactate secretion, lipogenesis, and fatty-acid synthase-dependent activities. The salient findings of this study are as follows. 1) Inhibition of PI 3-kinase/Akt/mTORC or fatty-acid synthase activation significantly suppresses both protein synthesis as measured by  ]acetate. Pretreatment of cells with inhibitors of PI 3-kinase/Akt/mTORC or fatty-acid synthase significantly suppresses ␣ 2 M* or insulin-induced lipogenesis in these cells, and anti-CTD has a similar effect on ␣ 2 M*-treated cells. 7) ␣ 2 M* or insulin-induced lipogenesis from 6-[ 14 C]glucose in 1-LN and DU-145 is also inhibited by pretreatment of cells with inhibitors of PI 3-kinase/Akt/mTORC, and fatty-acid synthase, and anti-CTD has a similar effect on ␣ 2 M*-dependent effects. 8) ␣ 2 M* and insulin induce a severalfold increase in cholesterol and fatty acid synthesis as measured by the incorporation of 1-[ 14 C]acetate into the nonsaponifiable fraction and saponifiable fraction, respectively, which is suppressed by pretreatment of cells with inhibitors of PI 3-kinase/Akt/mTORC or fatty-acid synthase, as well as anti-CTD-treated ␣ 2 M* exposed cells. 9) Thin layer chromatographic fractionation of 1-[ 14 C]acetate-labeled cell lipids of 1-LN and DU-145 cells stimulated with ␣ 2 M* or insulin shows a significant increase in the synthesis of cholesterol esters, triglycerides, free fatty acids, cholesterol, and phosphatidylcholine. Pretreatment of labeled cells with inhibitors of PI 3-kinase/Akt/mTORC or fatty-acid synthase significantly suppressed ␣ 2 M* or insulin-induced synthesis of cholesterol esters, triglycerides, free fatty acids, cholesterol, and phosphatidylcholine, and anti-CTD suppresses ␣ 2 M*-dependent effects. 10) Inhibition of proteolytic cleavage of SREBP by fatostatin A significantly suppresses ␣ 2 M* or insulin-induced protein and DNA synthesis, lipogenesis from 1-[ 14 C]acetate, and de novo phosphatidylcholine synthesis from [ 14 CH 3 ]choline. 11) Down-regulation of cell surface GRP78 by RNAi significantly suppresses the expression of fatty-acid synthase and ATP citrate lyase in 1-LN and DU-145 cells compared with buffer-treated controls. These data demonstrate that ␣ 2 M*, similar to insulin, functions as a growth factor, inducing the Warburg effect and altering lipid metabolism in 1-LN and DU-145 prostate cancer cells. These effects promote prostate cancer cell proliferation, which is blocked by both inhibitors of PI 3-kinase/Akt/mTORC or fatty-acid synthase and silencing or ligation of cell surface GRP78 by RNAi or by anti-CTD, respectively.
Based on our data, we postulate the following mechanism of ␣ 2 M*-induced metabolic changes in cancer cells. Binding of ␣ 2 M* to cell surface GRP78 activates PI 3-kinase/Akt/mTORC signaling, which enhances aerobic glycolysis as evidenced by increased glucose uptake and lactate secretion in these cells. PI 3-kinase/Akt/mTORC signaling activates SREBPs and initiates cell proliferation as evidenced by increased lipogenesis, protein, and DNA synthesis. Increased protein expression of fatty-acid synthase, ATP citrate lyase, and SREBP1-c and increased mRNA levels of SREBP1-c, SREBP2, ATP citrate lyase, acetyl-CoA carboxylase, and fatty-acid synthase correlates with these data. This increased metabolic activity of prostate cancer cells is suppressed by pretreatment of cells with inhibitors of PI 3-kinase/Akt/mTORC and fatty-acid synthase. These effects are mediated by SREBP, which is supported by the observation that inhibition of SREBP proteolytic cleavage by fatostatin A significantly suppresses ␣ 2 M*-induced protein synthesis, DNA synthesis, and lipogenesis. The ligation of cell surface GRP78 by ␣ 2 M* is a key player in inducing metabolic alterations in prostate cancer cells because down-regulating GRP78 by RNAi significantly suppresses the activation of two lipogenic enzymes, acetyl-CoA carboxylase and fatty-acid synthase. The key role of GRP78 is also evident in experiments in which we ligated GRP78 with antibody directed against its (CTD. This results in significantly decreased protein and DNA synthesis accompanied by reduced glucose uptake, lactate secretion, lipogenesis, and phosphatidylcholine synthesis. This down-regulation of GRP78 functions to inhibit the Warburg effect and may be a potential target for cancer therapeutic development. A schematic representation of ␣ 2 M* and cell surface GRP78 in inducing aerobic glycolysis in prostate cancer cells is shown in Fig. 10. In conclusion, the metabolic effects on prostate cancer cells treated with ␣ 2 M* were highly consistent with insulin. The ␣ 2 M/complement gene family evolved at least 600 million years ago (16). In horseshoe crabs, ␣ 2 M functions as both a proteinase inhibitor and complement protein, but through gene duplication and divergent evolution, these functions became discrete to either ␣ 2 M or complement (16). The insulin gene evolved 500 million years ago as a regulator of cellular metabolism (97). Because of the similarities on cellular responses elicited upon binding of ␣ 2 M* and insulin to their receptors, we propose that although ␣ 2 M evolved as a proteinase inhibitor, during later evolutionary phases, it acquired functions as a growth factor. In this context, GRP78 on the cell surface may be viewed as a proto-oncogene (98).