The protein factor β2-microglobulin (β2M), purified from the conditioned medium of human prostate cancer cell lines, stimulated growth and enhanced osteocalcin (OC) and bone sialoprotein (BSP) gene expression in human prostate cancer cells by activating a cyclic AMP (cAMP)–dependent protein kinase A signaling pathway. When β2M was overexpressed in prostate cancer cells, it induced explosive tumor growth in mouse bone through increased phosphorylated cAMP-responsive element binding protein (CREB) and activated CREB target gene expression, including OC, BSP, cyclin A, cyclin D1, and vascular endothelial growth factor. Interrupting the β2M downstream signaling pathway by injection of the β2M small interfering RNA liposome complex produced an effective regression of previously established prostate tumors in mouse bone through increased apoptosis as shown by immunohistochemistry and activation of caspase-9, caspase-3, and cleavage of poly(ADP-ribose) polymerase. These results suggest that β2M signaling is an attractive new therapeutic target for the treatment of lethal prostate cancer bone metastasis. (Cancer Res 2006; 66(18): 9108-16)
Prostate cancer bone metastasis is lethal, and currently, there is no effective therapy (1). To develop new approaches targeting prostate cancer bone metastasis, we sought to understand the reciprocal interaction between prostate cancer and bone cells at the molecular level to identify soluble factors that may be shuttled between these cells that are responsible for the maintenance of cell-specific phenotypes and behaviors. We focused specifically on the regulation of bone-restricted osteocalcin (OC) and bone sialoprotein (BSP) expression in prostate cancer cells because the expression of these proteins has been proposed to confer the ability of prostate cancer cells to grow and survive in the bone microenvironment (2, 3). OC and BSP could contribute in part to the adhesive properties of prostate cancer cells through the binding of their cell surface integrin receptors and a RGD motif of extracellular matrices (4, 5). The ability of prostate cancer cells to mimic bone by expressing OC, BSP, osteopontin, osteonectin/SPARC, and the receptor activator of nuclear factor-κB ligand (RANKL) could further aid prostate cancer bone colonization as shown by the ability of prostate cancer cells to form mineralized bone nodules in culture (2, 3, 6). This understanding of prostate cancer and bone cell interaction has been the basis for developing novel and promising therapies for the treatment of prostate cancer bone metastasis in the clinic. Bisphosphonates, agents known to decrease bone turnover, were shown to reduce overall bone loss and improve bone pain associated with bone metastasis when applied clinically in patients with prostate or breast cancer bone metastases (7, 8). Atrasentan, an endothelin-1 (ET-1) receptor A antagonist, was found to improve quality of life and bone pain in prostate cancer patients by interfering with the interaction between ET-1, secreted by prostate cancer cells, and its receptor ET-A, located on the cell surface of osteoblasts (9). In addition, data from experimental cell lines, animal models, and clinical studies reveal that insulin-like growth factor (IGF)-I and its receptor IGFR-I, platelet-derived growth factor (PDGF) and its receptor PDGFR, and RANKL and RANK interaction are promising targets for therapeutic intervention (10–12).
β2-microglobulin (β2M) is a nonglycosylated protein composed of 119 amino acid residues with a secreted form of 99 amino acids and a molecular mass of 11,800 Da (13, 14). β2M is synthesized by all nucleated cells and forms complexes with the heavy chain of MHC class I antigen through noncovalent linkage on cell surfaces (15). MHC I is essential for the presentation of protein antigens recognized by cytotoxic T cells (16). On recognition of foreign peptide antigens on cancer cell surfaces, T cells actively bind and lyse the antigen-presenting cells. Down-regulation of MHC I occurs frequently in cancer cells, and this contributes to the immune evasiveness of cancer cells that allows them to escape from elimination by the attacking cytotoxic T cells (17). The biological functions of β2M in cancer are not clear. Increased β2M levels in bone marrow and blood specimens are correlated with a poor prognosis and the failure of multiple myeloma patients to respond to therapy (18). Urine β2M levels are elevated in advanced prostate cancer patients and correlate negatively with patient survival (19). Concentrations of serum β2M are also increased in gastrointestinal (20) and breast cancer patients (21). Hence, β2M may be useful as a prognostic and therapeutic response indicator for cancer patients. β2M is a mitogen and is capable of increasing the growth of human osteoblasts (22), human prostate cancer PC3 cells, and rat stromal cells (23) and to regulate the expression of hormone/growth factor receptors (epidermal growth factor receptor, insulin receptor, and IGF-I and IGF-II receptors) and the interaction with their ligands (24–26).
We report here that the protein factor β2M is required to maintain the bone phenotypes or osteomimicry exhibited by prostate cancer cells (2, 3). We showed that β2M activates phosphorylated cyclic AMP (cAMP)–responsive element binding protein (p-CREB) with increased expression of its target genes. This activation could enhance tumor growth and angiogenesis and facilitate the recruitment of osteoblasts and osteoclasts to the site of tumor colonization in bone. This signaling axis offers an opportunity for improved clinical targeting of prostate cancer bone metastasis. Recently, we (Huang et al., 2006, AACR Annual Meeting, Abstract 4822) and others (Yang et al., 2006, AACR Annual Meeting, Abstract 2218) have shown that β2M expression is linked to the increased growth, migration, and invasion of human prostate, breast, lung, renal cancer, and myeloma cells and promoted their epithelial to mesenchymal transition (EMT). Our results established for the first time the growth and signaling roles of β2M in human prostate cancer bone metastasis and confirmed β2M/protein kinase A (PKA)/CREB signaling axis as a potential new target for therapy.
Materials and Methods
Cell lines and cell culture. Human prostate cancer cell lines LNCaP, C4-2B4, DU145, PC3, and ARCaP and osteosarcoma cell line MG63 were cultured in T-medium (Life Technologies, Inc., Rockville, MD) supplemented with 5% fetal bovine serum (FBS) and 1% penicillin/streptomycin as described previously (27). The cells were maintained at 37°C in 5% CO2. For conditioned medium collection, cells were cultured in T-medium with serum until 80% confluence. The cells were washed subsequently twice in PBS (10 mmol/L phosphate buffer and 137 mmol/L NaCl) and incubated in T-medium without serum. After 2 days of additional incubation, conditioned media were collected, centrifuged, and stored at −20°C until use.
Purification and identification of β2M. All purification procedures were done at 4°C. Total proteins from 100 mL of ARCaP conditioned medium were precipitated by 0% to 100% saturation of ammonium sulfate. After centrifugation, the protein precipitates were dissolved in 1 mL of PB [10 mmol/L sodium phosphate (pH 7.6) containing 0.2 mmol/L phenylmethylsulfonyl fluoride, 0.5 mmol/L DTT, protease inhibitor cocktail (Roche Molecular Biochemicals, Indianapolis, IN)] and dialyzed against 500 mL of PB twice overnight. After centrifugation of the dialyzed solution, the supernatant was passed through Centricon (YM-30, Millipore Corp., Billerica, MA) to remove the proteins with high molecular mass (>30 kDa). Protein filtrate (100 μL, ∼100 μg of total proteins) was loaded into an anion exchange column (4 × 250 mm, ProPac Wax-10, Dionex, Sunnyvale, CA) preequilibrated with PB at a flow rate of 1 mL/min. The bound form proteins were eluted with a liner gradient of 0 to 1 mol/L NaCl. After analysis of human OC (hOC) promoter-luciferase activity (2) for each fraction, the purified protein fractions having promoter-luciferase activity were collected and analyzed by SDS-PAGE with silver staining (Invitrogen, Carlsbad, CA). After SDS-PAGE analysis, the NH2-terminal amino acid sequence of this homogenous protein was determined by the Edman degradation method (28). Human β2M protein was purchased from Sigma (St. Louis, MO).
Reverse transcription-PCR. Total RNA was isolated from the confluent monolayer of cells using RNAzol B (TelTest, Inc., Friendswood, TX). The total RNA was used as the template for reverse transcription according to the manufacturer's instructions (Invitrogen). The oligonucleotide primer sets used for PCR analysis of cDNA were β2M [5′-ACGCGTCCGAAGCTTACAGCATTC-3′ (forward) and 5′-CCAAATGCGGCATCTAGAAACCTCCATG-3′ (reverse)], OC, BSP, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as described previously (2). The thermal profile for β2M amplification is 30 cycles starting with denaturation for 1 minute at 94°C followed by 1 minute of annealing at 64°C and 30 seconds of extension at 72°C. Reverse transcription-PCR (RT-PCR) products were analyzed by agarose gel electrophoresis. Quantity One 4.1.1 Gel Doc gel documentation software (Bio-Rad, Hercules, CA) was used for quantification of β2M, OC, and BSP mRNA expression and normalized by GAPDH mRNA.
Western blot analysis and ELISA. Western blot was done using the Novex system (Invitrogen). Primary antibody anti-β2M and vascular endothelial growth factor (VEGF; Santa Cruz Biotechnology, Inc., Santa Cruz, CA) was used at a 1:500 dilution, and secondary antibody (Amersham Biosciences Corp., Piscataway, NJ) was used at a 1:5,000 dilution. Detection of protein bands was done with Enhanced Chemiluminescence Western Blotting Detection Reagents (Amersham Biosciences). For apoptosis analysis, rabbit polyclonal anti-caspase-9, anti-caspase-3, and anti–poly(ADP-ribose) polymerase (PARP) antibodies (1:500 dilution) were ordered from Cell Signaling Technology, Inc. (Beverly, MA). Anti-CREB, p-CREB (1:1,000), cyclin A, and cyclin D1 (1:2,000) antibodies were purchased from Cell Signaling Technology. The concentration of protein was determined by the Bradford method using Coomassie plus protein reagent (Pierce, Rockford, IL). β2M protein concentration was assayed by Quantikine IVD human β2M ELISA kit (R&D Systems, Inc., Minneapolis, MN) according to the manufacturer's instructions.
Cell proliferation assay. Prostate cancer cell lines were plated in 96-well plates in T-medium containing 5% FBS. After 24 hours of incubation, media were replaced by serum-free T-medium and incubated for an additional 24 hours. The cells were exposed with conditioned medium or reagents for 3 or 4 days of incubation. Cell numbers were measured every 24 hours using CellTiter 96 AQueous One Solution Cell Proliferation Assay (Promega, Madison, WI).
In vivo animal studies. All of the animal experiments were approved and done in accordance with institutional guidelines. Four-week-old male athymic nu/nu mice (National Cancer Institute, Frederick, MD) were inoculated into the bone marrow space of the femurs with 1 × 106 cells per mouse of Neo or β2M-overexpressing C4-2B cells, respectively. Blood specimens were harvested for prostate-specific antigen (PSA) assay biweekly. Serum PSA levels were determined by microparticle ELISA using the Abbott IMx machine (Abbott Laboratories, Abbott Park, IL).
β2M small interfering RNA and antiprostate tumor study. The specific β2M and control scramble small interfering RNA (siRNA) sequences were 5′-UUGCUAUGUGUCUGGGUUU(dT)(dT)-3′ and 5′-UUCAUGUGUCUGUGGUGUU(dT)(dT)-3′, respectively. For efficient RNA delivery, we used a cationic liposome formulation (29) to deliver β2M and scramble siRNA into cell lines or living mice. To test the prostate tumor growth inhibition by β2M siRNA, 4-week-old male athymic nu/nu mice were inoculated either with 2 × 106 C4-2B cells in mouse femur or in the s.c. space of the chest regions with 2 × 106 cells of PC3-Luc or C4-2B mixed with 20 mg of bone powder and 35 μL of Matrigel matrix (BD Biosciences, Bedford, MA), respectively. Three (PC3-Luc) or 4 (C4-2B) weeks later, the tumor-bearing mice were randomly divided into two groups. The mice in the treatment group received an intratumoral injection of the β2M siRNA liposome complex thrice weekly continuously for 4 weeks at a dose of 0.8 μg of siRNA mixed with 19.2 μL of liposome for each mouse. The control group was injected with the same dose of the scramble siRNA liposome complex. To assay the antitumor efficacy of β2M siRNA, real-time bioluminescence images acquired by CCD camera using a cryogenically cooled IVIS system with analysis software (Xenogen Corp., Alameda, CA; ref. 30) and serum PSA levels were used to monitor PC3-Luc and C4-2B tumor burden in nude mice, respectively.
Immunohistochemical staining. Detection of β2M expression and cell death of normal and tumor specimens was conducted using Dako Autostainer Plus system (Dako Corp., Carpinteria, CA). Mouse monoclonal antibody against β2M (used at 1:1,000 dilution; Santa Cruz Biotechnology) and against M30 CytoDeath (1:600; DiaPharma Group, Inc., West Chester, OH) were used. Tissues were deparaffinized, rehydrated, and subjected to pressure cooking antigen retrieval at 125°C and 20 p.s.i. for 30 seconds, 10 minutes of double endogenous enzyme block, 30 minutes of primary antibody reaction, and 30 minutes of EnVision+ Dual Link or Biotinylated Link and Streptavidin-Peroxidase System incubation. Signals were detected by adding substrate hydrogen peroxide using diaminobenzidine as the chromogen and counterstained with hematoxylin. All reagents were obtained from DAKO.
Statistical analysis. Statistical analyses were done as described previously (2). Student's t test and two-tailed distribution were applied in the analysis of statistical significance.
Purification and identification of a soluble factor conferring enhanced hOC promoter activity in human prostate cancer cells. We previously reported that conditioned medium of ARCaP cells can induce hOC and human BSP (hBSP) promoter activities and their respective mRNA expression through a cAMP-dependent PKA signaling pathway by human prostate cancer cells (2). To identify the responsible factor, ARCaP conditioned medium was subjected to purification by ammonia sulfate [(NH4)2SO4] precipitation, membrane filtration (YM-30), anion exchange column chromatography, and NH2-terminal amino acid sequencing of the final purified product. The respective fractions collected from the various purification steps were subjected to hOC promoter-luciferase activity analysis in C4-2B cells, an androgen-independent human prostate cancer cell line of the LNCaP lineage (31). Figure 1A shows that anion exchange chromatography fractions eluted between 0.20 and 0.22 mol/L NaCl seem to contain the active factor capable of stimulating hOC promoter-luciferase activity. This fraction was then analyzed by SDS-PAGE and silver stain. Figure 1B (left) shows that a homogenous protein band migrating 11 kDa was observed. In addition, we used mass spectrometry to determine the exact molecular mass of this factor, which is 11,802.6 Da. NH2-terminal amino acid sequencing of this fraction revealed that this biologically active protein to be β2M (IQRTPKIQVYSRHPA). Western blot analysis subsequently using anti-β2M antibody confirmed that this protein is β2M (Fig. 1B,, right). β2M, a known housekeeping gene, was further shown to be expressed uniformly among a series of human prostate cancer and bone cell lines as assessed by semiquantitative RT-PCR (Fig. 1C). As expected, β2M protein, conducted by Western blot (Fig. 1D), correlated directly with the ability of the conditioned medium to stimulate hOC promoter-luciferase activity, which was higher in the conditioned medium collected from the more aggressive ARCaP than from the less aggressive LNCaP cells with the soluble factor presented as the secreted form (2).
β2M is responsible for the activation of hOC and hBSP promoter-luciferase activities in selective prostate cancer cell lines. To validate that β2M is the active soluble factor responsible for the stimulation of hOC and hBSP promoter-luciferase activities in prostate cancer cells, we exposed C4-2B cells to ARCaP conditioned medium either in the presence or absence of anti-β2M neutralizing antibody. Both hOC and hBSP promoter-luciferase activities were induced by ARCaP conditioned medium, and the fold of the promoter activation seemed to be dependent on the concentrations of β2M protein in the conditioned medium (Fig. 2A). We observed that the activated hOC and hBSP promoter-luciferase activities were blocked by a neutralizing anti-β2M antibody but not by the control IgG (Fig. 2A). This study was also confirmed by a similar experiment where β2M conditioned medium was collected from C4-2B cells stably transfected with a β2M expression vector (Fig. 2B). As shown, β2M conditioned medium stimulated hOC and hBSP promoter-luciferase activities, which were blocked by the presence of anti-β2M antibody but not IgG. Finally, we ratified our data using β2M protein and confirmed a dose-dependent up-regulation of the promoter-luciferase activity in C4-2B cells, which were also suppressed by anti-β2M but not control antibody (Fig. 2C).
β2M activates OC and BSP mRNA expression and stimulates the growth of human prostate cancer cells through activation of CREB. To further evaluate whether β2M up-regulates the endogenous OC and BSP mRNA, we assessed the expression of these noncollagenous bone matrix proteins using a semiquantitative RT-PCR in C4-2B cells exposed to ARCaP or β2M conditioned medium (from C4-2B cells stably transfected with a β2M expression vector). The endogenous OC and BSP mRNA expression increased by 6- to 8-fold on the exposure of C4-2B cells to ARCaP or β2M conditioned medium (Fig. 3A,, left). Consistent with the hOC and hBSP promoter-luciferase activity data, anti-β2M antibody also inhibited the mRNA expression induction by ARCaP or β2M conditioned medium. Furthermore, the ability of β2M to induce OC and BSP mRNA expression was also confirmed in an additional study, in which C4-2B cells were stably transfected with a β2M cDNA expression vector. As shown in Fig. 3A (right), both the endogenous OC and BSP mRNA levels were increased in cells stably transfected with β2M cDNA (B2 and C2, the two highest β2M expression clones) but not with a control neomycin-resistant empty vector (Neo).
β2M has been shown previously to enhance the growth of PC3 cells (23). To address whether β2M also stimulates the proliferation of other human prostate cancer cells, we tested β2M on the growth of a broad range of human prostate cancer cell lines, including androgen-dependent (LNCaP), androgen-independent (C4-2B, DU145, and PC3), and androgen-repressed (ARCaP) cells. All the prostate cancer cell lines responded to β2M-induced cell proliferation. In a 3-day proliferation assay, the relative cell growth increased were as follows: ARCaP, 129 ± 8%; C4-2B, 121 ± 10%; DU145, 120 ± 7%; LNCaP, 117 ± 7%; and PC3, 111 ± 3%. We also compared the cell growth of B2-β2M-transfected, C2-β2M-transfected, and Neo-transfected C4-2B cell clones on plastic dishes and in soft agar. B2 and C2 clones that expressed the highest levels of β2M had the highest growth rate compared with Neo and the parental cells (−), which expressed only the endogenous levels of β2M (Fig. 3B,, top). Anti-β2M antibody inhibited parental and Neo-transfected C4-2B cell growth in a dose-dependent manner (Fig. 3B,, bottom). Cells expressing high levels of β2M, such as the B2 clone, required a higher amount of anti-β2M antibody (10 μg/mL) to inhibit cell growth induced by endogenous β2M than did the Neo-transfected clone (1 μg/mL). Control IgG (20 μg/mL) failed to exert a growth-inhibitory effect. We also conducted the three-dimensional anchorage-independent growth of Neo and β2M-overexpressing C4-2B cells in soft agar. The results showed a direct correlation between the size and the number of colonies formed in the soft agar and the levels of β2M expression by the C4-2B cells (Fig. 3C). These data collectively suggest that β2M is a potent stimulator for OC and BSP expression and the proliferation of prostate cancer cells in vitro.
To determine the possible signaling link between β2M expression and prostate cancer cell growth, we chose to evaluate the activation status of CREB, a target of β2M/PKA downstream signaling (2). We assessed the expression of CREB target genes, cyclin A, cyclin D1, and VEGF (32), which are known to control tumor cell growth and angiogenesis in nontransfected parental, Neo, and β2M-overexpressing C4-2B cells. Figure 3D shows that increased β2M expression (B2 and C2 clones) greatly promoted the protein expression of p-CREB (2.3- to 2.5-fold), cyclin D1 (3.9- to 4.1-fold), cyclin A (2.6- to 3.0-fold), and VEGF (3.3- to 3.8-fold) over that of the Neo control C4-2B cells.
β2M supports the growth of prostate cancer cells in mouse bone in vivo. Because OC and BSP accumulation by human prostate cancer cells confers increased growth and survival of cancer cells and their expression of growth regulatory CREB target genes in hosts, we further compared the growth of Neo and β2M-overexpressing C4-2B cells in mouse bone. β2M-overexpressing C4-2B cells exhibited a 16-fold increased rate of growth over that of the Neo-transfected cells, as assessed by serum PSA (33), when injected into mouse skeleton (serum PSA concentration: Neo, 103 ± 17 ng/mL and β2M, 1,697 ± 500 ng/mL) at 17 weeks after tumor cell inoculation (Fig. 4A, and B, top). β2M does not significantly affect the PSA expression in C4-2B cells (PSA levels in Neo and β2M clones are 21.9 ± 2.1 ng/μg and 25.3 ± 2.5 ng/μg cellular protein, respectively; P = 0.15). Histomorphologic analysis of β2M tumors showed a more intense osteoblastic response in mouse femur compared with the specimens obtained from Neo tumors and normal control samples (Fig. 4B,, bottom). Immunohistochemical analysis shows that β2M tumors highly expressed β2M proteins compared with Neo and normal tissues harvested from mouse femurs using anti-β2M antibody (Fig. 4C). The observed increase in β2M expression in C4-2B tumors in mouse femurs corresponded with the increased β2M expression in primary and metastatic human prostate cancer tissues. Figure 4D shows representative immunohistochemical analysis of β2M staining in primary and bone metastatic human prostate cancer specimens. Abundant β2M staining was seen in primary prostate cancers (139 of 153 or 91% positive) and bone metastases (4 of 4 or 100% positive). Note that heterogeneity exists in the primary human prostate cancer specimens, where some of the specimen cancer areas stained stronger than the normal, whereas in others the reverse was observed.
β2M is an attractive new therapeutic target for the treatment of human prostate cancer bone metastasis. Because β2M conferred increased prostate cancer cell growth in mice bone, we tested the possibility that β2M-induced intracellular signaling may be a therapeutic target. We devised a sequence-specific β2M siRNA and compared the activity of this siRNA with its control scramble β2M siRNA, delivered as a cationic liposome complex (29), to cultured cancer cells or preestablished prostate tumors in mice.
First, we assessed β2M and scramble siRNA for antagonizing β2M mRNA expression in C4-2B cells. Figure 5A shows that, in a transient transfection assay, β2M, but not scramble siRNA, effectively blocked the endogenous level of β2M mRNA. β2M protein expression, detected by Western blot and ELISA, was also inhibited by β2M siRNA but not scramble siRNA (data not shown). Corresponding with decreased β2M mRNA expression, β2M siRNA, as expected, eliminated the mRNA expression of OC and BSP (Fig. 5A). β2M siRNA significantly inhibited the growth of C4-2B cells. In a 4-day proliferation assay, relative cell growth increased to 139 ± 6% and 138 ± 8% in parental nontransfected and scramble siRNA-transfected cells, respectively, whereas the growth of β2M siRNA-C4-2B cells remained at only 109 ± 3%.
Next, we validated the chronic effect of β2M siRNA on human prostate tumor growth both in bone powder xenografts (34) and in mouse skeleton (35). The bone powder was shown previously by Reddi and Huggins (34) to recapitulate bone morphogenesis and cytodifferentiation in a highly time-dependent and host microenvironment-dependent manner. Because of the known heterogeneity in human prostate cancer, we chose to test the proliferation of both C4-2B and PC3 cells and their requirement of β2M as a growth and survival signal. C4-2B and PC3 cells are documented to have different profiles of androgen receptor and response to β2M-induced hOC and hBSP promoter-luciferase activity (2). PC3 cells stably transfected with a luciferase gene (PC3-Luc) formed tumors in bone powder, with growth increase as a function of time as analyzed by a CCD Xenogen camera in mice during a 28-day siRNA treatment period (Fig. 5B). Although the growth of PC3-Luc tumors was not affected by injections of the scramble siRNA liposome complex, the β2M siRNA liposome complex delivered in the same manner shrank and eliminated preexisting PC3-Luc tumors grown as bone powder xenografts (Fig. 5B, and C, top). Likewise, the growth of C4-2B tumors in bone powder in vivo, as detected by monitoring serum PSA levels, was also markedly inhibited by β2M siRNA treatment compared with scramble siRNA (Fig. 5C , bottom). The effect of β2M siRNA in eliminating preexisting prostate tumor growth in bone powder xenografts was also observed in mice inoculated directly with C4-2B cells in bone (data not shown). These data support the conclusion that antagonizing β2M signaling by β2M siRNA resulted in efficient targeting of the growth of human prostate tumors in mouse bone.
To investigate the molecular mechanism by which β2M siRNA induced human prostate cancer cell death, we conducted Western blot analysis to examine the activation of caspases, a hallmark of apoptosis (36). As shown in Fig. 5D (left), in C4-2B cells, β2M siRNA induced the activation of the initiator caspase, cleaved caspase-9 (37 and 35 kDa), and caspase-3 (19 and 17 kDa), which is one of the downstream effector caspases. The cleaved form of PARP (89 kDa), a factor downstream of caspase-3, was also detected in β2M siRNA-treated C4-2B cells. However, in parental (−) and scramble siRNA-treated C4-2B cells, no activated caspase-9, caspase-3, or cleaved PARP proteins were observed. Consistent with these results, marked elevation of apoptosis by β2M siRNA, as assessed by a CytoDeath M30 stain, was detected in both PC3-Luc and C4-2B bone powder xenografts (Fig. 5D , right). These results suggest that β2M siRNA reduced human prostate cancer cell proliferation in vivo and induced apoptotic death in prostate cancer cells.
The mortality and morbidity associated with human prostate cancer bone metastasis remains a challenge in the clinical management of human prostate cancer. Recently, new approaches have been explored to target the interphase between prostate cancer and bone, and some success has been achieved in reducing pain and improving survival and morbidity in patients with hormone refractory metastatic disease. Specifically, the use of bisphosphonate has been shown to reduce bone loss and fracture in patients treated with hormone withdrawal therapy (8). A bone-directed cotargeting strategy, killing prostate cancer cells with chemotherapy and modulating host bone cells with strontium-89, was shown to improve the overall survival of patients with hormone refractory prostate cancer (37). Atrasentan was shown to reduce bone pain and improve the quality of life of prostate cancer patients with hormone refractory bone metastasis (9). These advances illustrate the importance of understanding the cellular interaction between prostate cancer and bone at the molecular level. In the present communication, we showed for the first time that β2M regulates the signaling pathway that confers the expression of bone matrix proteins OC and BSP and induces growth of human prostate cancer cells through the activation of the angiogenesis factor and cell cycle regulatory cyclins. By interrupting β2M-mediated downstream signaling, we observed considerable cell death and the shrinkage of preexisting prostate tumors in bone powder and mouse femur models. Cell death was confirmed by immunohistochemistry of the apoptotic M30 marker and the activation of caspases and PARP, markers known to be associated with programmed cell death.
β2M, a known housekeeping gene, expresses at a constant level with respect to its mRNA in many mammalian tissues and cells (38). β2M is a key protein involved in the presentation and stabilization of MHC I antigen on the cell surface. However, the role of β2M in cancer and bone metastasis is unexplored and unclear. Our present investigation revealed that β2M protein levels are variably in prostate cancer cell lines, with the levels of protein expression corresponding positively with the malignant status of the prostate cancer cells. These results suggest that the translation or stability of β2M protein must be tightly controlled to maintain cancer cell growth and survival.
Based on our published work and others, β2M-mediated signaling, but not necessarily β2M protein level in tissues or sera, contributes to enhanced prostate cancer growth and colonization in bone through four possible mechanisms (Fig. 6): First, β2M can activate cAMP-dependent PKA activity through binding to and activation of the seven-transmembrane G protein-coupled receptor or a yet-to-be-identified β2M receptor (2). This activation could induce p-CREB, which increases cell proliferation, survival, and angiogenesis (39–42) through elevated levels of cyclin A, cyclin D1, and VEGF (Fig. 3D). Second, β2M also enhances the synthesis and deposition of noncollagenous bone matrix proteins, such as OC and BSP, which serve as survival factors by binding to the cell surface integrin receptors (i.e., αvβ3 and αvβ5), and thus could sustain the growth and survival of prostate cancer cells in bone (43, 44). Third, OC and BSP have been shown to recruit osteoclasts and osteoblasts in the skeleton, resulting in increased bone turnover, which creates a growth factor-enriched niche that allows cancer growth and colonization in the newly “pitted” bone areas (3). Fourth, β2M (Huang et al., 2006, AACR Annual Meeting, Abstract 4822) or PKA/CREB activation (45) has been shown to promote EMT in a wide spectrum of human tumor cells, including prostate, breast, renal, and lung. Through the induction of EMT, cancer cells acquire enhanced ability to migrate, invade, and eventually gain access to the metastatic sites (46). Interrupting β2M-regulated signaling could block all of the above-mentioned pathways, which could contribute to the observed in vivo antitumor effects of β2M siRNA. The siRNA technique has been broadly used to investigate gene function, gene regulation, and gene-specific therapeutics (47). However, low RNA transfection efficiency and interference by serum has limited the application of this technology for gene delivery. An improved nonviral method using the cationic liposome formulation has gained popularity because of its high binding affinity with the negatively charged RNA or DNA (29) that form a complex and can efficiently carry siRNA into cancer cells or host mouse tissues with minimal immunogenicity and toxicity in immunocompromised mice (30). In the present communication, we used this cationic liposome formulation and evaluated the ability of the β2M siRNA liposome complex to inhibit the growth of preexisting human prostate tumors in bone powder and mouse femur in nude mice. We showed direct cytotoxic effects by interrupting β2M signaling in human prostate cancer cells because these tumors were grown in immunocompromised mice. Thus, the β2M-mediated signaling pathway may contribute directly to prostate tumor growth, survival, and transdifferentiation. By blocking β2M-regulated signaling, prostate tumor death through the apoptotic cascade pathway can result. Because normal cells are insensitive to growth inhibition by β2M siRNA or anti-β2M antibody (Yang et al., 2006, AACR Annual Meeting, Abstract 2218), this suggests that cancer cells develop dependence on β2M-mediated signaling and thus are particularly vulnerable to the β2M signal blockade-induced cytotoxicity.
In summary, we have shown that β2M signaling, via a cAMP-dependent PKA pathway, CREB activation, and the expression of bone-like properties by prostate cancer cells, contributes to prostate cancer cell growth and survival in bone. By interrupting β2M signaling with sequence-specific β2M siRNA, we observed a marked decrease of expression of OC and BSP, cancer cell death, and shrinkage of preexisting prostate tumors in mice. Targeting β2M either alone or in combination with other therapeutic modalities may be a promising new approach for the treatment of lethal human prostate cancer bone metastasis.
Grant support: 1 P01 CA98912, DAMD 17-03-2-0033, and PC040260 (L.W.K. Chung).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank Dr. Mien-Chie Hung (University of Texas M.D. Anderson Cancer Center, Houston, TX) for providing the liposome formulation, Dr. Hari Reddi (University of California at Davis, Davis, CA) for supplying the bone powder, Gary Mawyer for editing the article, and our colleagues at the Molecular Urology and Therapeutics Program for helpful suggestions and discussion.