Investigating Genetic Alterations in Bladder Cancer
BAP1 mutations contribute to BRCA pathway alterations in bladder cancer. A, Somatic BAP1 mutations in 8/54 (15%) bladder tumors. B, A ribbon diagram showing the recurrently-altered active site proton donor residue, p.H169 (green). C, Altered residue, p.R718 (*), in the nuclear localization signal and BRCA1 binding domain is highly conserved. D, Papillary features in H&E stained tumor sections from tumors with BAP1 mutations.
Bladder cancer (BC) is the fifth most common cancer worldwide and the sixth most common cancer in the U.S. Mutations in a number of oncogenes and tumor suppressor genes were previously associated with invasive or noninvasive forms of the disease. More recently, next generation sequencing (NGS) of bladder tumors from over 100 Chinese patients revealed alterations in additional genes, including those encoding chromatin remodeling enzymes, like lysine specific histone demethylase 6A (KDM6A) and ARID1A, and spindle checkpoint proteins. Because the NGS studies analyzed tumors from patients of a single ethnicity, the results may not be representative of alterations in other patients. To expand on these findings, Michael Nickerson, Ph.D., and Michael Dean, Ph.D., of CCR’s Cancer and Inflammation Program and their colleagues, including Dan Theodorescu, M.D., Ph.D., Director of the University of Colorado NCI-Designated Comprehensive Cancer Center, performed exome sequencing of 14 tumors and targeted sequencing of another 40 tumors all from non-Asian patients diagnosed with BC in the U.S.
Exome sequencing of the 14 tumors produced an average of more than 7.6 gigabases per sample, and each exon base was read an average of 92 times. The variants predicted by NGS were enriched in C-to-T and G-to-A changes, which have been observed in other cancers, including BC. Removing common variants and known duplications, the researchers identified 9,312 missense, 348 nonsense, 216 splice junction, and 653 insertion/deletion candidate somatic variants. They compared matched tumor and normal tissue DNA and confirmed 228 of 316 variants; 112 were somatic, and 116 were germline, 43 of which were novel.
The validated somatic variants occurred in 67 genes, many known to be mutated in BC. Frequently mutated genes, which were altered in three or more tumors, included KDM6A, ARID1A, and four other chromatin remodeling genes previously unrecognized as mutated in BC, GCN1L1, CHD1, CHD1L, and, BAP1 suggesting an important role for this class of genes in BC. The somatic changes to GCN1L1 observed by the investigators likely caused a loss of protein function. In one tumor, for example, deletion of a base resulted in a frameshift and protein truncation. Mutations to CHD1 and the related CHD1L also resulted in protein truncation or altered conserved residues. Fifteen percent of the bladder tumors from Chinese patients also showed changes in CHD-family genes.
The scientists then looked more closely at BAP1, and sequenced the gene in an additional 40 tumors and matched normal tissue. Fifteen percent of 54 tumors showed BAP1 variants predicted to alter protein function. One introduced a premature stop codon, while two others altered a key residue in the ubiquitin hydrolase active site. Two additional variants affected the protein’s nuclear localization signal, which blocked its deubiquitination activity in previous studies. Since BAP1 binds BRCA1, the researchers wondered whether BAP1 mutation may contribute to alterations in the BRCA DNA repair pathway. In fact, they found that a majority of tumors had BRCA pathway mutants, suggesting these BCs may be vulnerable to targeted therapies, such as PARP inhibitors. When they compared their BAP1 results to the previous NGS studies, the investigators found that BAP1 was altered much more frequently in tumors from U.S. than Chinese patients (15 percent vs. 1 percent). However, three other frequently mutated genes, ARID1A, KDM6A, and STAG2, were similarly altered in the two populations, indicating preferential BAP1 mutation in U.S. patients.
Sequencing revealed copy number variations in tumor versus normal tissue DNA. However, of the BC genes identified, the scientists found that most somatic alterations, around 80 percent, were sequence changes rather than copy number variants. They also showed that two or more alterations potentially affecting both alleles of a BC gene were fairly common and impacted an average of two genes per tumor. Focusing on the germline variants, the researchers identified 11 nonsense, frameshift, and splice junction alterations in known and novel cancer-related genes, suggesting a significant contribution of these likely rare germline variants to apparently sporadic cases of BC. Indeed, they observed germline TERT promoter alterations in 56 percent of tumors, some of which were identical to somatic variants in distinct tumors. In addition, 69 percent of tumors contained somatic TERT promoter changes. A subset of the somatic and germline variants created new transcription factor binding sites, while others altered known transcription factor recognition sites or likely sites of CpG methylation, potentially causing dysregulated gene expression.
The investigators next assessed somatic alterations in five BC genes, ARID1A, BAP1, KDM6A, STAG2, and TERT, and their association with tumor histology and clinical factors. Only BAP1 and KDM6A mutations were correlated, and TERT promoter alterations were independent of changes in the other genes. The scientists found that BAP1 mutations were associated with papillary features in BC tumors. However, they observed no significant associations between clinical factors and mutations in these five genes.
To further examine the interactions between BC genes, the researchers performed network analysis and identified several high-scoring networks. Intriguingly, their analysis suggested that KDM6A is a driver of BC at the center of one network. Using a previously developed gene mutation signature for KDM6A, the investigators found higher scores in bladder tumors compared to normal tissue in three patient cohorts. They then tested the functional role of KDM6A in two bladder cancer cell lines, MGHU3 and T24T, which express wild-type and truncated KDM6A, respectively. Over-expressing wild-type KDM6A in T24T cells reduced anchorage-independent growth and cell migration. Depleting KDM6A with small interfering RNA (siRNA), on the other hand, enhanced these activities in MGHU3 cells. Similarly, MGHU3 cells treated with the KDM6A siRNA formed significantly larger tumors than control cells when implanted in mice.
Together these studies have uncovered novel BC-related genes, revealed previously unappreciated interactions between these genes and with other pathways that might be targeted by therapeutics, and demonstrated the importance of TERT, KDM6A, and BAP1 in BC.Summary Posted: 09/2014
Nickerson ML, Dancik GM, Im KM, Edwards MG, Turan S, Brown J, Ruiz-Rodriguez C, Owens C, Costello JC, Guo G, Tsang SX, Li Y, Zhou Q, Cai Z, Moore LE, Lucia MS, Dean M, Theodorescu D. Concurrent alterations in TERT, KDM6A, and the BRCA Pathway in Bladder Cancer. Clin. Cancer Res. September 15, 2014 PubMed Link