RGS12 Is a Novel Tumor-Suppressor Gene in African American Prostate Cancer That Represses AKT and MNX1 Expression

African American (AA) men have a significantly higher incidence of prostate cancer compared with European American (EA) men (1) and are twice as likely to die from prostate cancer compared with EA men. The biological basis for this difference in prostate cancer mortality is unclear. Because AA men account for a significant fraction of all prostate cancer–related deaths in the United States, it is important to understand the basis for this higher mortality in order to optimize prevention and treatment strategies for this higher risk group of men.

There have been a number of studies comparing prostate cancer tissues from AA and EA men. Several studies have compared gene expression in AA and EA prostate cancer using large-scale expression microarrays (2–5) including a study from our group (6). A number of studies have focused on a smaller set of preselected genes (7–9). All of these studies indicate that there is differential gene expression between AA and EA prostate cancers. The TMPRSS2/ERG fusion gene is much less frequent in AA prostate cancer based on studies of DNA, RNA, and protein (8–16). Elevated SPINK1 expression appears to be more common in AA prostate cancer (8, 9, 17–19). Among other genes upregulated in AA prostate cancer, inflammatory genes are prominent (2, 4, 7). We have recently identified MNX1 as an oncogene that is expressed at significantly higher levels in AA prostate cancer compared with EA prostate cancer (6). We further demonstrated that MNX1 is regulated by AKT and androgen receptor activity and upregulates lipid synthesis, which has been linked to aggressive disease (20, 21), and thus, MNX1 may contribute to disease aggressiveness in AA prostate cancer.

We have published (22) a study of allelic loss and gain in 20 AA prostate cancers using Affymetrix 500k SNP arrays to define regions of recurrent copy-number gain and loss in localized prostate cancer and compared the pattern of copy-number alterations (CNA) with that of a similar cohort of EA men (23). We found multiple cytobands with a statistically higher frequency of CNAs in our AA cohort over the EA cohort. The only unique CNA identified in this initial analysis that had not been previously linked to prostate cancer was loss of chromosome 4p16.3.

We have now extended our original CNA studies to a new set of 40 highly tumor-enriched primary prostate cancers and matched benign prostate tissues from AA men using high-resolution Affymetrix 6.0 SNP arrays and expression array analysis using RNAs from the same tissues. We have confirmed the specific loss of 4p16.3 described previously (22) and identified a novel tumor-suppressor gene, RGS12, at this locus that shows significantly decreased expression in AA prostate cancer but not EA prostate cancer. Both in vitro and in vivo data show that RGS12 is a tumor-suppressor gene, as would be predicted from its known ability to negatively regulate pro-oncogenic signal transduction. Furthermore, we have found that loss of RGS12 increases expression of MNX1 at least in part by regulating AKT protein levels. Our findings establish a novel oncogenic axis in AA prostate cancer.

Tissue samples were obtained from the Human Tissue Acquisition and Pathology Core of the Dan L. Duncan Cancer Center and were collected from fresh radical prostatectomy specimens after obtaining informed consent under a Baylor College of Medicine Institutional Review Board–approved protocol and as such followed the principals of the Declaration of Helsinki and the Belmont Report. Cancer tissues include at least 70% tumor tissue, and benign tissues were free of cancer on pathologic examination. DNAs and RNAs were extracted using a Qiagen DNA/RNA mini kit following the manufacturer's protocol.

DNAs from AA prostate cancer tissues and matched benign tissue were analyzed using Affymetrix 6.0 SNP arrays by the Albert Einstein College of Medicine Genomics Core. SNP array data were processed using the crlmm package in Bioconductor, with the preprocessing steps for copy-number estimation as follows: (1) quantile normalization of the raw intensities (quantile normalizing the SNPs and nonpolymorphic markers separately), (2) genotyping, and (3) for total copy number, translating the normalized intensities to an estimate of raw copy number by adding the allele-specific summaries. For each of the 1M SNP probes, each tumor profile was centered on the paired normal, in order to generate tumor:normal ratios. Tumor:normal logged values were averaged by gene, and each profile was centered on the median of log ratios across all genes. For heat map presentation, gene-level tumor:normal values were further collapsed into cytobands. When combining datasets from multiple studies, values for each dataset were binned as gain or loss or no change, using a similar approach to that of our previous study (22). For the Lapointe dataset, the SD of the tumor profile with the smallest SD across cytobands was used as the reference for defining gain or loss events within each cytoband; cytobands with average values greater than +3SD were called as gain, and cytoband values less than −3 SD were called as loss. Gene-level copy alterations for the Taylor dataset were previously binned in that study, with average cytoband log₂ ratio >0.6 or ←0.6 being called here as gain or loss, respectively. For our own SNP array datasets [present study and Castro and colleagues (22)], a log₂ of 0.2 was used as the cutoff [similar to that of Castro and colleagues (22)]. Expression array analysis has been described previously (6). GEO accession number is pending.

Gene expression levels were tested using quantitative real-time PCR (qRT-PCR) on an Applied Biosystem (StepOne, Lifetechnologies). Total RNAs were extracted using the RNasy Kit (Qiagen). cDNAs were synthesized as described previously. TagMan probes used are listed in Supplementary Table S1. Differences in mRNA levels were analyzed using the ΔΔCt method normalized to β-actin expression. Each measurement point was repeated at least in triplicate.

Human immortalized normal prostate epithelial cell line PNT1A and prostate cancer cell lines LNCaP, DU145, and PC3 were all maintained in RPMI-1640 medium (Invitrogen) supplemented with 10% FBS (Invitrogen). LAPC4 cells were cultured in RPMI-1640 medium with 10% FBS supplemented with 10 nmol/L R1881 (Sigma). VCaP and 293T cells were maintained in DMEM (Invitrogen) with 10% FBS. All cell culture medium contained 1x Antibiotic-Antimycotic (Gibco). PNT1A cells were obtained from the European Type Culture Collection. PNT1A with myristoylated AKT and controls have been described previously (24). All other cell lines were obtained from the American Type Culture Collection. Cell were obtained between 2001 and 2012, expanded, frozen, and stored as stocks in liquid nitrogen. All cell lines are authenticated by STR analysis at MD Anderson Cancer Center Characterized Cell Line Core Facility. Cells are tested monthly for mycoplasma contamination.

LNCaP cells with stable knockdown of RGS12 were produced by utilizing RGS12-shRNAs in lentiviral vector pGFP-C-shLenti (Origene). Four unique human shRNAs (A–D) for RGS12 constructs in lentiviral GFP vector were purchased from Origene (Cat # TL302015). Another 3 unique shRNA-RGS12 constructs (V3LHS_310594; 310595; and 310599) were purchased from the Baylor College of Medicine C-Bass core. Lentiviruses carrying these stable shRNAs were produced in 293T cells using a Lenti-vpak Packaging kit (Origene) following the manufacturer's instruction. LNCaP and PNT1A cells were infected by these viruses and were selected with 0.5 μg/mL puromycin (Sigma).

Primers used to amplify three human RGS12 isoforms (GenBank accession NM_198229, NM_002926, and NM_198227) are listed in Supplementary Table 2. Three RGS12 isoforms were cloned into pcDNA3.1/V5-His-TOPO vector (Invitrogen) containing CMV promoter. Constructs were sequenced and confirmed their accuracies before transfection into cells. The transfection was performed using Fugene 6 reagent (Promega), and transfected cells were selected and maintained in media containing 200 μg/mL G418.

Cells (5 × 10³) were plated in each well of 96-well plates. Proliferation was determined using the Cell Counting Kit-8 Cell Proliferation Assay Kit (Dojindo Molecular Technologies) as described by the manufacturer. The absorbance was read at 450 nm with VERSAmax Tunable microplate reader.

Cells (2.5 × 10⁴) were plated in the top chamber of Matrigel-coated membrane (24-well insert; pore size, 8 mm; BD Biosciences). The cells on the apical side of each insert were then scraped off after 24 hours. The wells were washed with PBS, fixed with 100% methanol, and stained with DAPI. After staining, membranes were removed from the insert and mounted on slides, and the invading cells were counted under the Nikon Eclipse TE2000-U microscope. Matrigel assays were performed in triplicate.

Six-well plates with 0.5% base agar layer mixed with 1X culture media plus 10% FBS were prepared before the seeding of cells. PNT1A cells (5 × 10⁴) with stable knockdown of RGS12 or vector controls were plated in 0.35% top agar layer each agar dishes. Cell colonies were counted after incubation at 37°C in an incubator for 3 weeks and staining with 1 mg/mL of idonitrotetrazolium chloride for 8 hours. This experiment was repeated twice.

Total cellular protein lysate was prepared as described previously. Anti–β-actin was obtained from Sigma-Aldrich (dilution 1:5,000). Anti-MNX1 was purchased from Origene (Cat# TA337035) and used at a dilution of 1:3,000. Anti-RGS12 was purchased from Santa Cruz (sc-514173). Antibodies to total AKT and phospho-AKT-T308 and S473 were from Cell Signaling. Western blotting procedures were described previously (6).

All procedures were approved by the Baylor College of Medicine Institutional Animal Use and Care Committee. Experiments were carried out on 8- to 10-week-old male SCID mice. Tumor xenografts were established by subcutaneous injection over each flank in 50 μL volume mixed with 50 μL Matrigel (BD Bioscience). Tumors were harvested 8 weeks after inoculation, and the tumor weights were recorded. Tumor tissues were snap frozen for further mRNA and protein expression studies.

Numerical values from two groups were compared by the Mann–Whitney test, with P < 0.05 considered significant. For more than two groups, ANOVA was used followed by pairwise comparison to controls, which were considered significant if P < 0.05.

We have extended our original CNA studies of AA prostate cancer (22) to a new set of 40 highly tumor-enriched primary prostate cancers and matched benign prostate tissues from AA men using high-resolution Affymetrix 6.0 SNP arrays (906K SNPs). We combined these new data, our published data [Castro and colleagues (22)] and the predominantly EA data sets of Lapointe and colleagues (23) and Taylor and colleagues (total AA: n = 89; EA: n = 169; ref. 25) for analysis. Data organized by race are shown in Supplementary Fig. S1. Cluster analysis of these data revealed that the vast majority of AA prostate cancers clustered in two major groups (Fig. 1A, right), indicating that AA and EA prostate cancers have different patterns of CNAs than EA prostate cancer. The AA cluster to the far right shows a distinct pattern of CNAs. A smaller cluster of AA prostate cancers clusters adjacent to the metastatic samples and has significant similarities to CNAs in these samples.

We then compared frequencies of CNAs at all cytobands between AA and EA patients after excluding metastatic samples and samples without data on race. A total of 32 cytobands showed significantly higher loss or gain (P < 0.01, one-sided Fisher exact test) in AA prostate cancer when compared with EA prostate cancer (Fig. 1B and Supplementary Table S3). Of note, we confirmed the specific loss of 4p16.3 described previously (P < 0.001). It is well known that hereditary cancer loci often show somatic alterations as well, so it is noteworthy that 6 of the 32 cytobands we identified have been implicated in hereditary AA prostate cancer by linkage analysis, including 8q24 (26–28),11q13 (28), 12q24 (29), 14q32 (30), 17p11 (31), and 17q21 (32). We then carried out a cluster analysis of the cytobands that were significantly different between AA and EA prostate cancer. Most of the AA prostate cancers cluster into four groups in this analysis as indicated in Fig. 1B. Group A has no losses in these cytobands. Group B shows multiple gains and some losses, whereas Group D shows more focal gains. Group C shows loss in multiple cytobands preferentially lost in AA prostate cancer. Most of the AA prostate cancer cases with loss of 4p16.3 cluster in this group. Of note, Group C clusters adjacent to the metastatic prostate cancers, which are predominantly derived from EA men, and the metastatic cases also show loss at 4p16.3, indicating that loss of 4p16.3 is likely to be associated with aggressive disease in primary prostate cancer.

We have also carried out expression array analysis using RNAs from the same cancers used for CNA analysis (6). We identified a total of 4,341 probes altered in prostate cancer versus benign (P < 0.01) in AA prostate cancer, and the overall quality of the data has been confirmed as described previously (6). To identify the potential tumor suppressor on 4p16.3 that was preferentially lost in AA prostate cancer, we systematically examined expression of genes on 4p16.3 in AA prostate cancer in the expression array data. We found two genes that are adjacent on 4p16.3 (HTT and RGS12), which both show downregulation of mRNA in AA prostate cancers. Detailed examination of deletions showed that losses are more concentrated in RGS12, and CNAs correlate with expression levels for RGS12 (r = 0.59; P = 0.016) but not HTT (Supplementary Table S4).

RGS12 has three alternatively spliced protein coding isoforms as shown in Supplementary Fig. S2. The proteins encoded by isoforms 1 and 2 are almost identical and encode full-length proteins. Isoform 3 lacks the amino terminal PTB and PDZ domains. Because of the extensive overlap of the three isoforms, we were not able to measure the mRNA levels of all three isoforms individually. We were able to analyze total RGS12 using a probe from the common region of all three isoforms, total isoforms 1+2, total isoforms 1+3, and isoform 3 alone. qRT-PCR analysis using a TaqMan probe that detects all three RGS12 isoforms showed significantly decreased RGS12 expression in AA prostate cancer (Fig. 2A; P < 0.001, Mann–Whitney). Analysis using primers detecting both isoforms 1 + 2 or isoform 3 only also showed significantly decreased expression levels in AA prostate cancers compared with benign prostate tissues (Fig. 2B and C; both P < 0.001; Mann–Whitney). However, no loss was seen in EA prostate cancer using primers detecting all three isoforms (Fig. 2D). Scatter plots of this data are shown in Supplementary Fig. S3. Overall, the CNA and expression analysis data show that there is loss of RGS12 alleles and/or gene expression in AA prostate cancer that is not seen in EA prostate cancer.

We then compared expression of RGS12 in PNT1A cells, an immortalized normal prostate epithelial cell line to LNCaP, VCaP, LAPC4, DU145, and PC3 prostate cancer cell lines. As shown in Fig. 2E, all five prostate cancer cell lines showed decreased RGS12 relative to PNT1A. The differences were highly statistically significant (P < 0.001; ANOVA) in all prostate cancer cell line except DU145 where the decrease was relatively small (P < 0.05, ANOVA). Analysis of isoform expression in the same cell lines revealed that isoform 1+2 and isoform 1+3 expression was decreased in all prostate cancer cell lines except DU145 (Fig. 2F). Isoform 3 had relatively low expression in PNT1A compared with isoforms 1 and 2, and although there was a trend for lower isoform 3 expression in the prostate cancer cell lines, this was not statistically significant.

In order to study RGS12's potential tumor-suppressor function, we knocked down RGS12 expression in PNT1A cells with three different RGS12-shRNAs using a lentiviral vector (pGFP-C-shLenti). After stable selection, knockdown of RGS12 in each group was confirmed using a TaqMan probe that detects all three isoforms (Fig. 3A). Knockdown of RGS12 significantly increased cell proliferation (P < 0.001 at 4 days, ANOVA) in all three groups compared with scrambled control (Fig. 3B). We also tested colony formation in soft agar, a hallmark of transformation. PNT1A cells are not fully transformed and form only rare small colonies in this assay. As shown in Fig. 3C, knockdown of RGS12 markedly increased colony formation in soft agar (P < 0.001, ANOVA). The overall colony sizes in RGS12 knockdown lines were also much larger compared with controls (Fig. 3D). As a positive control, we used PNT1A cells expressing Huntington-interacting protein-1, which formed colonies in soft agar (data not shown) as we have shown previously (33). We also knocked down RGS12 in LNCaP using four different shRNAs (Fig. 3E). As in PNT1A, proliferation was significantly increased with all four shRNAs (Fig. 3F).

We then cloned all three major isoforms of RGS12 from LNCaP cells into pCDH-CMV-MCS-EF1-Neo lentiviral vector. As seen in Fig. 4A, V5 antibody detected all three bands with correct estimated sizes using lysate from 293T cells following transient transfection of the three isoforms. Using an anti-RGS12 antibody, we were able to confirm the overexpression of different RGS12 isoforms, although based on the relative band intensity in anti-V5 and anti-RGS12 Westerns, the affinity of the anti-RGS12 antibody for isoform 3 was higher than for isoforms 1 and 2. After stable transfection of all three isoforms into LNCaP cells, we confirmed significantly increased RGS12 expression at RNA level in LNCaP cells (Fig. 4B). Overexpression of each isoform significantly decreased cell proliferation (Fig. 4C). Isoform 3 had a lower effect on growth compared with the other two isoforms (P < 0.05 vs. Iso1 or Iso2, ANOVA). Overexpression of each isoform dramatically inhibited cell invasion (Fig. 4D), and again isoform 3 showed a weaker effect on invasion (P < 0.01 vs. Iso1 or Iso2, ANOVA).

To evaluate tumor-suppressive activities in vivo, we carried out xenograft experiments in SCID mice. In the first experiment, two groups of mice (10 mice/group) were injected subcutaneously with LNCaP with vector or LNCaP-shD, respectively. Tumor growth was monitored twice weekly. At end of 5 weeks, mice were euthanized and primary tumors were excised, weighed, and a portion of the tumor was frozen in liquid nitrogen for molecular analysis and another portion fixed and paraffin-embedded. The difference of tumor weight between shD group and controls was statistically significant (Fig. 5A, P < 0.05, Mann–Whitney), with higher tumor weights in tumors with RGS12 knockdown. In the second experiment, we used LNCaP cells overexpressing RGS12 isoform 2 or 3 or control cells with scrambled vector. Tumor weights were significantly decreased in both isoform expressing groups compared with control cells (Fig. 5B, P < 0.001, Mann–Whitney). Surprisingly, isoform 3 tumors were smaller than isoform 2–expressing tumors given that isoform 3 appeared to be less tumor suppressive than isoform 2 in vitro. Analysis of RGS12 mRNA expression in the final tumors revealed that isoform 3 tumors had approximately 4-fold higher levels of RGS12 compared with isoform 2 tumors (data not shown). This implies there may have been preferential growth of tumor cells with lower RGS12 knockdown during the in vivo growth, and this effect was more profound in the isoform 2–expressing cells than the isoform 3–expressing cells. Finally, because PNT1A cells with RGS12 knockdown formed colonies in soft agar, we injected PNT1A-shD or shE cells or control PNT1A cells into mice. After 8 weeks, there were no tumors found in the PNT1A control group, whereas obvious tumor masses were seen in both groups with RGS12 knockdown. The average weight of tumors collected was 68 and 84 mg in the shD and shE groups, respectively (Fig. 5C). To confirm that the tumors were from PNT1A cells, we used SV40 T-antigen immunohistochemistry, because PNT1A cells were originally immortalized with SV40 large T-antigen. As shown in Fig. 5D, tumor cells were positive for SV40 T-antigen. Overall, our data show that RGS12 is tumor-suppressor gene in vivo for prostate cancer and prostate epithelial cells.

We have recently shown that MNX1 is an oncogenic transcription factor whose expression is preferentially increased in AA prostate cancer. Examination of the gene expression data in our AA prostate cancers (6) revealed a significant negative correlation between MNX1 and RGS12 mRNA expression in cancer tissues (−0.278; P = 0.028, Spearman). We saw a similar negative correlation (−0.230; P = 0.01, Spearman) in the dataset of Grasso and colleagues (34).These data are from a predominantly EA cohort but have a mixture of localized (59 cases) and metastatic prostate cancer (35 cases), unlike our data, which are all from localized disease. These correlations are shown in Supplementary Fig. S4. We hypothesized that RGS12 may repress MNX1 expression.

We examined expression of MNX1 in the LNCaP cell lines with overexpression of RGS12 isoforms 1 to 3. As seen in Fig. 6A, MNX1 protein expression was completely lost in isoform 1–expressing cells and markedly diminished in isoform 2– and 3–expressing cells. Conversely, knockdown of RGS12 markedly increased MNX1 expression in LNCaP cells in vitro (Fig. 6B). Similar increased levels of MNX1 protein were seen in xenograft tumors from mice with RGS12 knockdown (Fig. 6C). Examination of MNX1 in LNCaP xenografts expressing RGS12 isoform 3 or vector controls significant downregulation of MNX1 protein (Fig. 6D). Thus, RGS12 significantly decreases MNX1 protein expression.

We have shown previously using inhibitors of the PI3K (LY294002) and AKT (AZD5363) that MNX1 is strongly regulated by AKT activity (6). We have now confirmed this observation using PNT1A cells expressing myristoylated AKT. As shown in Fig. 7A, these cells show increased levels of total and phosphorylated AKT compared with vector controls and also show increased MNX1 protein levels. We then examined the impact of RGS12 expression on AKT activity. As shown in Fig. 7B, isoform 1 almost completely abolishes AKT protein expression as well as expression of phosphorylated AKT. Expression of isoforms 2 and 3 resulted in lesser decreases in total and phosphorylated AKT. Knockdown of RGS12 in PNT1A cells showed increase in S473 phosphorylated AKT (Supplementary Fig. S5). Of note, PNT1A are PTEN wild type.

Examination of AKT mRNA in cells overexpressing RGS12 showed increased mRNA that was statistically significant for isoforms 1 and 2 (Fig. 7C). This strongly suggests that RGS12 regulates AKT posttranscriptionally, with a feedback upregulation of AKT mRNA. Consistent with this, we observed a positive correlation between RGS12 and AKT mRNA levels in our expression microarray data (0.338, P = 0.009; Supplementary Fig. S6).

In this report, we have shown that RGS12 is preferentially deleted in AA compared with EA prostate cancer, and there is significantly lower RGS12 mRNA expression in prostate cancer compared with benign tissues in AA but not EA prostate cancer. We observed a significant correlation of genomic deletion and decreased expression (r = 0.59; P = 0.016) in AA prostate cancer, but at this level of correlation, it is likely that other factors may also affect RGS12 mRNA expression in AA prostate cancer in addition to genomic deletion. Such factors will require further studies to elucidate.

Our in vitro and in vivo data show that RGS12 is a tumor-suppressor gene. Of note, decreased RGS12 is by itself capable of fully transforming immortalized normal prostate epithelial cells such that they form large colonies in soft agar and tumors in SCID mice, indicating that it has a strong tumor-suppressive effect in this context. It can also affect tumorigenesis and transformation-related cellular phenotypes in vitro and in vivo in fully transformed prostate cancer cells.

RGS12 has not been previously implicated as a tumor-suppressor gene. A SNP in RGS12 has been shown to be associated with overall survival in lung cancer (35), but the mechanism for this association is unknown. RGS12 is a negative regulator of G-protein signaling that acts via enhancing GTP hydrolysis (36). As such, it can inhibit signal transduction from G-protein–coupled receptors, a number of which have been implicated in the pathogenesis of prostate cancer (37–42), although the exact targets of RGS12 are not clear. RGS12 has been shown to interact with the IL8 receptor (43). It also contains a phospho-tyrosine binding domain and has been shown to interact with MAPK/ERK and PI3K signaling in various contexts (44–46) and thus may act as link between G-protein–coupled signaling and other signaling pathways (36).

We have shown previously that MNX1 is an oncogenic transcription factor whose expression is markedly increased in AA prostate cancer and to a much lesser extent in EA prostate cancer (6). Remarkably, examination of gene expression data in our AA prostate cancers revealed a significant negative correlation between MNX1 and RGS12 mRNA expressions in AA prostate cancer tissues. Both knockdown and overexpression studies confirm that RGS12 can strongly inhibit expression of MNX1. AKT activity strongly regulates MNX1 expression based on our published data (6), and analysis of PNT1A cells expressing myristoylated AKT confirms this observation. Our data indicate that RGS12 significantly negatively regulates AKT protein levels and activity. It should be noted that levels of phosphorylated AKT were roughly proportional to AKT protein levels, indicating that activation of AKT was not inhibited, but with lower AKT protein, total phosphorylated AKT was also decreased. The exact mechanism by which RGS12 can regulate AKT protein levels is yet to be determined but appears to be posttranscriptional since AKT mRNA is actually increased by RGS12 overexpression.

The regulation of AKT protein levels has not been as intensively studied as its activation by phosphorylation, although increased AKT protein may enhance AKT signaling, particularly in the context of dysregulated AKT activation. As reviewed by Liao and Hung (47), there are multiple posttranscriptional mechanisms potentially affecting AKT protein levels. AKT can be phosphorylated at threonine-450, and this phosphorylation affects protein stability. This site is phosphorylated during AKT translation and may be important in modulating interactions with Pin1, which can regulate AKT stability. It should be noted that Pin1 is increased in prostate cancer and is associated with aggressive disease (48). Interactions with heat shock proteins can also increase AKT stability. On the other hand, ubiquitin-mediated proteolysis or degradation by caspases of AKT has also been described (47). Additional studies are needed to understand the mechanism by which AKT protein levels are controlled by RGS12.

Our data indicate that decreased RGS12 enhances transformed phenotypes at least in part via increasing expression of MNX1, establishing a novel oncogenic axis in AA prostate cancer. MNX1 enhances lipid synthesis, which has been shown to be associated with disease aggressiveness (20, 21). The increased MNX1 resulting from RGS12 loss is mediated at least in part by its effect on AKT protein levels, but other mechanisms may also be involved in RGS12 regulation of MNX1 expression. Of course, increased AKT protein will almost certainly affect other AKT targets as well. Whether decreased RGS12 also affects other cellular targets that can enhance transformed phenotypes is unclear. Further studies are needed to fully clarify the mechanism of action of RGS12 as a tumor-suppressor gene in prostate cancer.

Our current and previous CNA studies in AA prostate cancer have shown that there are significant quantitative and qualitative differences in CNAs between AA and EA prostate cancer. Of the 32 cytobands with CNAs identified as being more common in AA prostate cancer, 6 coincide with a region linked to familial AA prostate cancer. However, the majority of CNAs we have identified have not been previously linked to AA prostate cancer. Although we have focused on 4p16.3 in these studies, other areas of CNAs that are more frequently present in AA prostate cancer identified in our studies may also harbor novel tumor suppressors or oncogenes relevant to AA prostate cancer. Future studies will hopefully identify additional genes that affect initiation and/or progression in AA prostate cancer and provide new insights into the optimal strategies for prevention and treatment of prostate cancer in AA men.

No potential conflicts of interest were disclosed.

Conception and design: J. Wang, L. Zhang, M. Ittmann

Development of methodology: J. Wang, L. Zhang

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y. Wang, L. Zhang, O.F. Karatas, L. Shao, P. Castro, M. Ittmann

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Wang, J. Wang, L. Zhang, Y. Zhang, C.J. Creighton, M. Ittmann

Writing, review, and/or revision of the manuscript: Y. Wang, J. Wang, C.J. Creighton, M. Ittmann

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y. Wang, L. Zhang, M. Ittmann

Study supervision: J. Wang, C.J. Creighton, M. Ittmann

This work was supported by grants from the Department of Defense Prostate Cancer Research Program (W81XWH-12-1-0046 to M. Ittmann); the National Cancer Institute supporting the Dan L. Duncan Cancer Center (P30 CA125123) Human Tissue Acquisition and Pathology and Genomic and RNA Profiling Shared Resources; the Prostate Cancer Foundation (M. Ittmann), and by the use of the facilities of the Michael E. DeBakey VAMC.

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.