In This Issue

See article, p. 369

Hepatosplenic T-cell lymphoma (HSTL) is a rare lymphoma that occurs at an early age and has a poor prognosis; due to the rarity of HSTLs, the genetic drivers of HSTL are predominantly unknown. To define the genetic landscape of HSTL, McKinney, Moffitt, and colleagues performed whole-exome sequencing of a discovery set of 70 HSTLs including a discovery set of 20 HSTLs with matched germline DNA and two HSTL cell lines. HSTLs exhibited a genetic landscape distinct from those of other non-Hodgkin lymphoma subtypes: The most frequent gene alterations in HSTLs were loss-of-function mutations in the chromatin-modifying genes SETD2, INO80, and TET3, and activating mutations in the signaling pathway–related genes STAT5B, STAT3, and PIK3CD. Further, copy-number analysis identified frequent gains of chromosome 1q and loss of chromosome 10p in addition to the known chromosomal events occurring in chromosome 7 (isochromosome) and chromosome 8 (trisomy). Knockdown of SETD2, which was the most frequently silenced gene in HSTLs, resulted in the upregulated expression of proliferation-related genes and increased proliferation and colony formation in vitro. Similarly, expression of three PIK3CD mutants resulted in increased cell viability and proliferation in vitro. Together, these results describe the distinct mutational landscape of HSTLs, identify potential drivers of HSTL pathogenesis, and suggest that SETD2 is a potential tumor suppressor in HSTLs.

See article, p. 380

PTEN is a frequently mutated tumor suppressor gene with a well characterized role as a lipid phosphatase. PTEN inactivation reduces DNA repair and increases glucose metabolism, but its role in metabolism remains incompletely understood. Mathur and colleagues generated Pten^−/− mouse embryonic fibroblasts (MEF) to investigate the role of PTEN in cell growth and metabolism. Pten^−/− MEFs exhibited a growth advantage that depended on the presence of glutamine, a precursor for pyrimidine synthesis. Steady-state metabolic profiling revealed that Pten loss altered nucleotide synthesis and DNA metabolism, including upregulating components of the de novo pyrimidine synthesis pathway such as pyrimidine 2-deoxyribonucleotides. These findings suggest that Pten^−/− MEFs utilize glutamine to synthesize the nucleotides required for cell growth. The fourth step of de novo pyrimidine synthesis is catalyzed by dihydroorotate dehydrogenase (DHODH), and DHODH inhibitors reduced the proliferation of Pten^−/− MEFs. Further, in human and mouse breast cancer, glioblastoma, and prostate cancer cell lines, PTEN-mutant and Pten^−/− cells exhibited increased sensitivity to the DHODH inhibitor leflunomide, indicating that DHODH may be a therapeutic target in tumors with PTEN alterations. Consistent with these findings, leflunomide inhibited PTEN-mutant glioblastoma cell neurosphere formation, and reduced the growth of breast cancer xenografts. Mechanistically, PTEN-deficient cells were incapable of activating the ATR–CHK1 checkpoint at replication forks; thus, leflunomide-induced pyrimidine depletion promoted DNA damage and chromosomal breaks that resulted in cell death. These findings indicate that PTEN regulates glutamine-dependent pyrimidine synthesis, and suggest DHODH as a potential therapeutic target in patients with PTEN-mutant tumors.

See article, p. 391

Chemoresistance often develops in patients with triple-negative breast cancer (TNBC), resulting in a poor prognosis. Thus, strategies are needed to sensitize TNBC cells to chemotherapy. Brown and colleagues investigated the metabolic changes induced by chemotherapy to identify potential therapeutic targets. Changes in the steady-state metabolomics profile of a TNBC cell line treated with doxorubicin were measured by LC/MS-MS. Doxorubicin increased the levels of the pyrimidine nucleotides deoxycytidine triphosphate and deoxythymidine triphosphate. Pyrimidine nucleotides can by generated by recycling through a salvage pathway or synthesized by the glutamine-dependent de novo pyrimidine synthesis pathway. Metabolic flux analysis revealed an increase in the incorporation of stable isotope–labeled glutamine into metabolites in the de novo pyrimidine synthesis pathway, suggesting that an adaptive response to chemotherapy increases the synthesis of nucleotide precursors required for DNA synthesis and repair. Mechanistically, chemotherapy induced ERK-mediated phosphorylation of Thr456 of CAD, a multifunctional enzyme that controls metabolic flux through the de novo pyrimidine synthesis pathway, thereby enhancing CAD activity and increasing the generation of pyrimidine nucleotides. Inhibition of CAD or dihydroorotate dehydrogenase (DHODH), which catalyzes the fourth step of de novo pyrimidine synthesis, sensitized TNBC cells to doxorubicin and other genotoxic chemotherapeutic agents. In vivo, a DHODH inhibitor in combination with chemotherapy induced tumor regression in TNBC xenografts. Altogether, these findings demonstrate that the de novo pyrimidine synthesis pathway is aberrantly activated by chemotherapy, and its inhibition may potentially enhance the efficacy of chemotherapy in patients with TNBC.

See article, p. 400

Gene fusions involving the tyrosine kinase genes NTRK1/2/3, ROS1, or ALK occur in multiple tumor types and constitutively activate oncogenic pathways. In two phase I trials, Drilon, Siena, and colleagues evaluated the safety and antitumor activity of entrectinib, an oral inhibitor of TRKA/B/C (encoded by NTRK1/2/3), ROS1, and ALK. A total of 119 patients with a wide range of advanced or metastatic solid tumors were treated with entrectinib, including 60 patients with rearrangements involving NTRK1/2/3, ROS1, or ALK, 53 patients with other alterations affecting these genes, and 6 patients with no known alterations. Entrectinib was well tolerated; the majority of treatment-related adverse events were grade 1 or grade 2, and only 2 dose-limiting toxicities occurred. One patient with ALK^F1245V mutant neuroblastoma achieved a partial response, but no other patients lacking gene rearrangements responded to entrectinib. Further, none of the patients who had received prior therapy with tyrosine kinase inhibitors (TKI) targeting ROS1 or ALK achieved responses. Of the “phase II–eligible” patients who were TKI-naïve, overall responses were observed in all three patients with NTRK1 rearrangements, 12 of 14 patients with ROS1 rearrangements, including 2 complete responses, and 4 of 7 patients with ALK rearrangements. Further, 5 of 8 patients with primary or metastatic central nervous system (CNS) disease achieved responses. Collectively, these findings suggest that entrectinib is safe and effective in patients with TKI-naïve tumors harboring NTRK1/2/3, ROS1, or ALK rearrangements, including those with CNS disease, and thus warrants further clinical investigation in these patients.

See article, p. 410

Risk variants for different cancer types have been identified, but the role of these variants in determining the somatic events leading to cancer has not been fully elucidated. To identify interactions between germline variants and somatic events in cancer, Carter and colleagues performed association testing of genomic data from over 5,900 tumors, representing 22 cancer types, in The Cancer Genome Atlas (TCGA). Analyses that compared tumor-specific SNPs identified loci that were significantly associated with specific tumor types; for example, the chromosome locus 8q24.13, which contains genes that have been associated with shorter time to breast cancer recurrence, harbored markers that were associated with breast cancer and age of breast cancer onset. Association testing of SNPs in TCGA samples and the alteration status of 138 cancer driver genes identified 35 associations between 28 germline loci and 20 cancer driver genes that were replicated in the validation cohort. Consistent with these findings, the minor allele of an SNP shown to be significantly associated with a locus involved in regulating PIK3CA/mTOR signaling enhanced the effect of PTEN loss on inducing PIK3CA/mTOR signaling in vitro. Analysis of mutational heterogeneity based on germline loci identified 20 candidate cancer driver genes, 15 of which had not previously been reported as frequently mutated. These results show that inherited risk variants can influence the somatic evolution of cancer and provide a resource of germline–somatic interactions in cancer that may help elucidate the molecular mechanisms that underlie cancer susceptibility.

See article, p. 424

Fizzy-related protein 1 (FZR1) is a putative tumor suppressor that functions as an adaptor protein for the anaphase-promoting complex/cyclosome (APC/C; also known as APC) E3 ubiquitin ligase complex and is required for the degradation of various key cell-cycle regulators, including mitotic cyclins and kinases. Recent reports suggest that FZR1 can also serve as a scaffold protein that disrupts protein–protein interactions independent of APC. Wan and colleagues found that depletion of FZR1 in primary melanocytes triggered accumulation of BRAF protein and subsequent activation of ERK signaling, resulting in premature senescence. Consistent with this finding, FZR1 interacted with BRAF and specifically promoted APC-dependent ubiquitination and degradation of BRAF, but not ARAF or CRAF, in primary melanocytes. However, FZR1 did not suppress BRAF protein stability in cancer cells, but instead attenuated BRAF kinase activation in an APC-independent manner, in part via disruption of BRAF dimerization. In contrast to nontransformed cells, the E3 ligase activity of APC^FZR1 was inhibited by ERK- and/or cyclin D1/CDK4–mediated phosphorylation of the FZR1 N-terminus in cancer cells, in particular in melanoma cells expressing oncogenic BRAF^V600E. Pharmacologic inhibition of BRAF/MEK and CDK4 restored APC^FZR1 E3 ligase activity, destabilizing oncogenic APC^FZR1 substrates including BRAF. Combined depletion of FZR1 and PTEN synergistically enhanced melanocyte transformation in vitro and promoted oncogenic BRAF/ERK and AKT signaling in vivo. These findings identify FZR1 as a critical negative regulator of BRAF oncogenic signaling, and suggest that this function may contribute to the tumor-suppressive role of FZR1.

Note: In This Issue is written by Cancer Discovery editorial staff. Readers are encouraged to consult the original articles for full details.