1. Role of TET2 in prostate and other solid tumors. Somatic, non-synonymous TET2 alterations have been frequently observed in myeloid cancers, but the mutation frequency in solid tissue cancers, including PCa, is unclear. We reviewed TET2 status in published exomes, The Cancer Genome Atlas, and the Catalogue of Somatic Mutations in Cancer and identified 97 confirmed somatic non-synonymous alterations in tumors from 13 tissues, including endometrium, kidney, lung , and PCa. These introduced 13 nonsense mutations, 7 truncating insertions-deletions, and altered 2 splice junctions; 52 (54%) were likely deleterious; and 18 (19%) altered the catalytic oxoglutarate- and iron-binding (OGFE) domain, including an iron-binding catalytic residue, D1384 (Uniprot Database). We sought to confirm TET2 was the causal PCa gene identified by rs7679673, associated with increased PCa risk by GWAS although it resides 5 kilobases upstream of the TET2 promoter. We performed fine mapping of PCa risk across TET2 by selecting this region from genotyping 2.5 million SNPs in 4,838 cases and 3,053 controls in the PEGASUS study, which is an extension of the Cancer Genetic Markers of Susceptibility Study (DCEG, NCI, Bethesda, MD). We identified seven variants significantly associated with PCa risk (p=10-4), including six new SNPs in introns 1 and 2. SNP rs7679673 retained the highest association with risk (p=1.6x10-6) followed by rs1015521 (p=8.6x10-5). Significantly, two new PCa risk variants, rs6825684 and rs17508261, have a more protective homozygous OR (0.56-0.83, 95% CI) than rs7679673 (homozygous OR=0.63-0.82, 95% CI). Germline risk SNPs in introns 1 and 2 definitively identify TET2 as a causal PCa gene. With Dr. Steve Anderson and the CIP Genetics Core we performed electrophoretic mobility shift assays for 6 PCa risk SNPs in LNCaP and PC3 cell lines to examine TF binding. Nuclear proteins bound strongly to allelic probes containing rs17508261-C and rs7655890 G and T, with no detectable binding to rs17508261-T and stronger binding to the rs7655890-T allele in LNCaP cells. In silico analysis of TF binding sites showed rs17508261 was located in an octamer-binding protein 1 (Oct1) binding motif. We confirmed Oct1 binding to rs17508261-C and rs7655890-T by supershift analysis with a series of antibodies specific for the indicated TFs. This suggests Oct1-risk SNP interactions may modulate TET2 expression to increase PCa risk. We analyzed the genotypes of Oct1-binding risk SNPs, rs17508261 and rs7655890 and found Oct1-binding alleles, rs17508261-C and rs7655890-T, were risk and protective (prot) alleles, respectively. The (rs17508261-T prot)-(rs7655890-T prot) haplotype is in linkage disequilibrium with the protective risk allele, rs7679673- C, and binds Oct1 at rs7655890-T, which appears to be dominant for protection. We determined the genotypes of rs17508261 and rs7655890 in 11 PCa cell lines and observed reduced TET2a expression associated with haplotypes containing the rs17508261-C risk allele; however only a limited number of samples were analyzed (not shown). Decreased TET2 expression associated with a germline PCa risk SNP agrees with genetic observations of somatic TET2 loss in PCa tumors. To examine TET2 protein function, we used affinity chromatography and mass spectrometry to identify binding partners in HEK293T kidney cells, which we confirmed by immunoprecipitation in LNCaP prostate cells. IP for endogenous OGT, PSPC1, NONO and SFPQ revealed interactions with endogenous TET2 in LNCaP cell lysates. Reciprocal IP of endogenous TET2 with anti-TET2 confirmed endogenous OGT, PSPC1, NONO, and SFPQ were detected after WB with the appropriate antibodies. Since PSPC1 binds the AR, we examined whether TET2 and AR interacted in LNCaP cells. Anti-AR co-precipitated TET2 and NONO indicating that TET2 forms a multi-protein complex containing AR and NONO, and by inference, PSPC1 and SFPQ. We subjected LNCaP cell lysates to IP with anti-AR and demonstrated OGT and AR interact. 2. Functional analysis of BRCA1 and BRCA2 variants. Especially in minority populations, there is a large fraction of variants of unknown significance that confound interpretation of sequencing studies. To address this issue we have made use of available databases such as ClinVar and LOVD to classify most variants as either benign or likely benign. To further explore the function of the remaining variants we are collaborating with Dr. Sharan/CCR to use his extensive experience in BRCA1 and BRCA2 gene function. 3. The role of helical mutations in PIK3CA in cervical cancer. To begin to determine the potential functional consequences of the mutations we described in the PIK3CA gene in cervical tumors, we used site-directed mutagenesis to create specific mutations in a wild-type vector. We generated several know functional alleles as well as newly described alleles in our study and alleles that have not been functionally tested. Transfection of these constructs into cells allowed determination of phosphorylation at Ser473 and Thr308 of AKT. Most of the mutations do indeed increase AKT phosphphorylation consistent with a role in activating the AKT pathway. However, helical domain mutations such as E542K and E545K do not dramatically increase PIK3CA catalytic activity and may contribute to tumorigenesis by alternate mechanisms as kinase domain mutations.
Showing the most recent 10 out of 30 publications
