Heterogeneity of d-glucuronyl C5-epimerase expression and epigenetic regulation in prostate cancer.

Heparansulfate proteoglycans (HSPG) play an important role in cell-cell and cell-matrix interactions and signaling, and one of the key enzymes in heparansulfate biosynthesis is d-glucuronyl C5-epimerase (GLCE). A tumor suppressor function has been demonstrated for GLCE in breast and lung carcinogenesis; however, no data are available as to the expression and regulation of the gene in prostate cancer. In this study, decreased GLCE expression was observed in 10% of benign prostate hyperplasia (BPH) tissues and 53% of prostate tumors, and increased GLCE mRNA levels were detected in 49% of BPH tissues and 21% of tumors. Statistical analysis showed a positive correlation between increased GLCE expression and Gleason score, TNM staging, and prostate-specific antigen (PSA) level in the prostate tumors (Pearson correlation coefficients GLCE/Gleason = 0.56, P < 0.05; GLCE/TNM = 0.62, P < 0.05; and GLCE/PSA = 0.88, P < 0.01), suggesting GLCE as a candidate molecular marker for advanced prostate cancer. Immunohistochemical analysis revealed an intratumoral heterogeneity of GLCE protein levels both in BPH and prostate cancer cells, resulting in a mixed population of GLCE-expressing and nonexpressing epithelial cells in vivo. A model experiment on normal (PNT2) and prostate cancer (LNCaP, PC3, DU145) cell lines in vitro showed a 1.5- to 2.5-fold difference in GLCE expression levels between the cancer cell lines and an overall decrease in GLCE expression in cancer cells. Methyl-specific polymerase chain reaction (PCR), bisulfite sequencing, and deoxy-azacytidin (aza-dC) treatment identified differential GLCE promoter methylation (LNCaP 70-72%, PC3 32-35%, DU145, and PNT2 no methylation), which seems to contribute to heterogeneous GLCE expression in prostate tumors. The obtained results reveal the complex deregulation of GLCE expression in prostatic diseases compared with normal prostate tissue and suggest that GLCE may be used as a potential model to study the functional role of intratumor cell heterogeneity in prostate cancer progression.

The TCF4/ß-catenin pathway and chromatin structure cooperate to regulate D-glucuronyl C5-epimerase expression in breast cancer.

D-glucuronyl C5-epimerase (GLCE) is a potential tumor-suppressor gene involved in heparan sulfate biosynthesis. GLCE expression is significantly decreased in breast tumors; however, the underlying molecular mechanisms remain unclear. This study examined the possible epigenetic mechanisms for GLCE inactivation in breast cancer. Very little methylation of the GLCE promoter region was detected in breast tumors in vivo and in breast cancer cells (MCF7 and T47D) in vitro and GLCE expression in breast cancer cells was not altered by 5-deoxyazacytidine (5-aza-dC) treatment, suggesting that promoter methylation is not involved in regulating GLCE expression. Chromatin activation by Trichostatin A (TSA) or 5-aza-dC/TSA treatment increased GLCE expression by two to 3-fold due to an increased interaction between the GLCE promoter and the TCF4/ß-catenin transactivation complex, or H3K9ac and H3K4Me3 histone modifications. However, ectopic expression of TCF4/ß-catenin was not sufficient to activate GLCE expression in MCF7 cells, suggesting that chromatin structure plays a key role in GLCE regulation. Although TSA treatment significantly repressed canonical WNT signaling in MCF7 cells, it did not influence endogenous TCF4/ß-catenin mRNA levels and activated TCF4/ß-catenin-driven transcription from the GLCE promoter, indicating GLCE as a novel target for TCF4/ß-catenin complex in breast cancer cells. A correlation was observed between GLCE, TCF4 and ß-catenin expression in breast cancer cells and primary tumors, suggesting an important role for TCF4/ß-catenin in regulating GLCE expression both in vitro and in vivo. Taken together, the results indicate that GLCE expression in breast cancer is regulated by a combination of chromatin structure and TCF4/ß-catenin complex activity.

Genetic and epigenetic analysis of non-small cell lung cancer with NotI-microarrays.

This study aimed to clarify genetic and epigenetic alterations that occur during lung carcinogenesis and to design perspective sets of newly identified biomarkers. The original method includes chromosome 3 specific NotI-microarrays containing 180 NotI clones associated with genes for hybridization with 40 paired normal/tumor DNA samples of primary lung tumors: 28 squamous cell carcinomas (SCC) and 12 adenocarcinomas (ADC). The NotI-microarray data were confirmed by qPCR and bisulfite sequencing analyses. Forty-four genes showed methylation and/or deletions in more than 15% of non-small cell lung cancer (NSCLC) samples. In general, SCC samples were more frequently methylated/deleted than ADC. Moreover, the SCC alterations were observed already at stage I of tumor development, whereas in ADC many genes showed tumor progression specific methylation/deletions. Among genes frequently methylated/deleted in NSCLC, only a few were already known tumor suppressor genes: RBSP3 (CTDSPL), VHL and THRB. The RPL32, LOC285205, FGD5 and other genes were previously not shown to be involved in lung carcinogenesis. Ten methylated genes, i.e., IQSEC1, RBSP3, ITGA 9, FOXP1, LRRN1, GNAI2, VHL, FGD5, ALDH1L1 and BCL6 were tested for expression by qPCR and were found downregulated in the majority of cases. Three genes (RBSP3, FBLN2 and ITGA9) demonstrated strong cell growth inhibition activity. A comprehensive statistical analysis suggested the set of 19 gene markers, ANKRD28, BHLHE40, CGGBP1, RBSP3, EPHB1, FGD5, FOXP1, GORASP1/TTC21, IQSEC1, ITGA9, LOC285375, LRRC3B, LRRN1, MITF, NKIRAS1/RPL15, TRH, UBE2E2, VHL, WNT7A, to allow early detection, tumor progression, metastases and to discriminate between SCC and ADC with sensitivity and specificity of 80-100%.

LRRC3B gene is frequently epigenetically inactivated in several epithelial malignancies and inhibits cell growth and replication.

Chromosome 3 specific NotI microarrays containing 180 NotI linking clones associated with 188 genes were hybridized to NotI representation probes prepared using matched tumor/normal samples from major epithelial cancers: breast (47 pairs), lung (40 pairs) cervical (43 pairs), kidney (34 pairs of clear cell renal cell carcinoma), colon (24 pairs), ovarian (25 pairs) and prostate (18 pairs). In all tested primary tumors (compared to normal controls) methylation and/or deletions was found. For the first time we showed that the gene LRRC3B was frequently methylated and/or deleted in breast carcinoma - 32% of samples, cervical - 35%, lung - 40%, renal - 35%, ovarian - 28%, colon - 33% and prostate cancer - 44%. To check these results bisulfite sequencing using cloned PCR products with representative two breast, one cervical, two renal, two ovarian and two colon cancer samples was performed. In all cases methylation was confirmed. Expression analysis using RT-qPCR showed that LRRC3B is strongly down-regulated at the latest stages of RCC and ovarian cancers. In addition we showed that LRRC3B exhibit strong cell growth inhibiting activity (more than 95%) in colony formation experiments in vitro in KRC/Y renal cell carcinoma line. All these data suggest that LRRC3B gene could be involved in the process of carcinogenesis as a tumor suppressor gene.

High mutability of the tumor suppressor genes RASSF1 and RBSP3 (CTDSPL) in cancer.

BACKGROUND: Many different genetic alterations are observed in cancer cells. Individual cancer genes display point mutations such as base changes, insertions and deletions that initiate and promote cancer growth and spread. Somatic hypermutation is a powerful mechanism for generation of different mutations. It was shown previously that somatic hypermutability of proto-oncogenes can induce development of lymphomas. METHODOLOGY/PRINCIPAL FINDINGS: We found an exceptionally high incidence of single-base mutations in the tumor suppressor genes RASSF1 and RBSP3 (CTDSPL) both located in 3p21.3 regions, LUCA and AP20 respectively. These regions contain clusters of tumor suppressor genes involved in multiple cancer types such as lung, kidney, breast, cervical, head and neck, nasopharyngeal, prostate and other carcinomas. Altogether in 144 sequenced RASSF1A clones (exons 1-2), 129 mutations were detected (mutation frequency, MF = 0.23 per 100 bp) and in 98 clones of exons 3-5 we found 146 mutations (MF = 0.29). In 85 sequenced RBSP3 clones, 89 mutations were found (MF = 0.10). The mutations were not cytidine-specific, as would be expected from alterations generated by AID/APOBEC family enzymes, and appeared de novo during cell proliferation. They diminished the ability of corresponding transgenes to suppress cell and tumor growth implying a loss of function. These high levels of somatic mutations were found both in cancer biopsies and cancer cell lines. CONCLUSIONS/SIGNIFICANCE: This is the first report of high frequencies of somatic mutations in RASSF1 and RBSP3 in different cancers suggesting it may underlay the mutator phenotype of cancer. Somatic hypermutations in tumor suppressor genes involved in major human malignancies offer a novel insight in cancer development, progression and spread.

Functional characterization of the candidate tumor suppressor gene NPRL2/G21 located in 3p21.3C.

Initial analysis identified the NPRL2/G21 gene located in 3p21.3C, the lung cancer region, as a strong candidate tumor suppressor gene. Here we provide additional evidence of the tumor suppressor function of NPRL2/G21. The gene has highly conserved homologs/orthologs ranging from yeast to humans. The yeast ortholog, NPR2, shows three highly conserved regions with 32 to 36% identity over the whole length. By sequence analysis, the main product of NPRL2/G21 encodes a soluble protein that has a bipartite nuclear localization signal, a protein-binding domain, similarity to the MutS core domain, and a newly identified nitrogen permease regulator 2 domain with unknown function. The gene is highly expressed in many tissues. We report inactivating mutations in a variety of tumors and cancer cell lines, growth suppression of tumor cells with tet-controlled NPRL2/G21 transgenes on plastic Petri dishes, and suppression of tumor formation in SCID mice. Screening of 7 renal, 5 lung, and 7 cervical carcinoma cell lines showed homozygous deletions in the 3' end of NPRL2 in 2 renal, 3 lung, and 1 cervical (HeLa) cell line. Deletions in the 3' part of NPRL2 could result in improper splicing, leading to the loss of the 1.8 kb functional NPRL2 mRNA. We speculate that the NPRL2/G21 nuclear protein may be involved in mismatch repair, cell cycle checkpoint signaling, and activation of apoptotic pathway(s). The yeast NPR2 was shown to be a target of cisplatin, suggesting that the human NPRL2/G21 may play a similar role. At least two homozygous deletions of NPRL2/G21 were detected in 6 tumor biopsies from various locations and with microsatellite instability. This study, together with previously obtained results, indicates that NPRL2 is a multiple tumor suppressor gene.