In This Issue

See article, p. 1230.

Germline mutations in the proofreading subunit of polymerase ϵ, DNA polymerase epsilon catalytic subunit (POLE), predispose to colon and endometrial cancers, and biallelic POLE mutations result in a hypermutator phenotype. While somatic POLE mutations have been identified in patients with hypermutated glioblastomas (GBM), only one patient with GBM was shown previously to harbor a germline POLE mutation. Johanns and colleagues screened an adult patient with GBM who had a personal and family history of colon polyps for the presence of mismatch repair deficiency and identified a germline POLE mutation. The patient had initially been treated with temozolomide and radiotherapy, and subsequently experienced progression of the primary lesion and developed metachronous spinal metastases. DNA and RNA sequencing revealed that the pre- and post-treatment GBM and metastases were hypermutated and presented a high neoantigen burden, and clonal analysis identified a linear evolutionary pattern. Because patients with hypermutated tumors have responded to immune checkpoint blockade, the patient received the immune checkpoint inhibitor pembrolizumab after the initial relapse. Post-pembrolizumab radiographic assessment and evaluation of lymphocytic infiltration of the tumor showed decreased primary tumor burden and increased immune cell infiltration of the metastases. Together, these results describe a second adult patient with germline POLE deficiency and GBM, and highlight the potential for the replacement of alkylating agents with immune checkpoint inhibition as a therapeutic approach for patients with hypermutated GBM.

See article, p. 1237.

Chronic myelogenous leukemia (CML) can be treated with tyrosine kinase inhibitors (TKI) targeting BCR–ABL1, but disease eradication is limited by leukemic stem cell (LSC) resistance to TKIs and by the emergence of BCR–ABL1 resistance mutations. In order to identify additional potential therapeutic targets in CML, Xie and colleagues compared the gene expression profiles of human CML LSCs and normal hematopoietic stem cells (HSC) and found that the histone methyltransferase EZH2 was upregulated in LSCs. Depletion of EZH2 in human CML cell lines reduced cell growth and enhanced apoptosis irrespective of the BCR–ABL1 mutation status, and in a mouse model of CML, inactivation of Ezh2 delayed leukemia development and extended survival in both BCR–ABL1 wild-type and mutant CML. EZH2 depletion reduced the number of LSCs, and EZH2 inhibition reduced colony formation in human CML LSCs, but not normal HSCs, indicating that CML LSCs are specifically dependent on EZH2 and sensitive to EZH2 depletion. Further, EZH2 loss mimicked the complete loss of polycomb repressive complex 2 (PRC2) and led to derepression of PRC2 target genes that suppress cell proliferation and promote apoptosis. Inactivation of Ezh2 via gene editing induced disease regression in mice with established CML, supporting further investigation of EZH2 inhibitors for the treatment of patients with CML. The finding that CML LSCs are specifically sensitive to EZH2 inhibition also suggests that combining TKIs with EZH2 inhibitors may be a strategy to eliminate LSCs in patients with CML.

See article, p. 1248.

Tyrosine kinase inhibitors (TKI) targeting the BCR–ABL1 fusion protein in chronic myeloid leukemia (CML) have had great clinical success. However, the cure rate of CML is low, and many patients with CML require lifelong treatment with TKIs or develop TKI resistance, likely because TKIs fail to target BCR–ABL1-independent leukemic stem cells (LSC). Therefore, therapies that target CML LSCs are urgently needed. Through mRNA profiling of normal and CML samples, Scott, Korfi, and colleagues discovered that CML LSCs exhibited deregulated levels of PRC2 components, which corresponded with increased promoter H3K27me3 and changes in EZH2 target gene expression, suggesting that LSCs undergo epigenetic reprogramming that might be therapeutically targetable. Consistent with this idea, treating primary CML cells with an EZH2 inhibitor (EZH2i) resulted in reduced colony-forming ability, whereas normal progenitor cells were not affected, indicating that EZH2i may selectively target LSCs. The effects of combined TKI and EZH2i treatment of CML LSCs were more than additive and led to reduced cell viability, decreased colony-forming ability, and increased apoptosis in vitro, and nearly eliminated patient-derived LSCs in the bone marrow of a mouse xenograft model. Moreover, the combination of TKI and EZH2i enhanced activation of apoptotic genes as well EZH2 target genes. This study indicates that combined treatment with TKI and EZH2i may be more effective in depleting the LSC population than TKI alone, and suggest that EZH2 inhibitors should be studied further in patients with CML.

See article, p. 1258.

The three most prevalent types of renal cell carcinomas (RCC) have been genomically characterized, but rarer RCC types, such as mucinous tubular and spindle cell carcinoma (MTSCC), have yet to be similarly analyzed. Although the majority of patients with MTSCC have a favorable prognosis, a small subset of patients with aggressive MTSCC faces a poor prognosis. Mehra, Vats, and colleagues performed whole-exome sequencing (WES) and capture transcriptome sequencing of 22 MTSCCs and matched adjacent normal tissues. While MTSCCs exhibited an overall low mutational burden, they exhibited highly recurrent mutations in the Hippo pathway–related tumor suppressor genes (TSG) protein tyrosine phosphatase non-receptor type 14 (PTPN14) and neurofibromin 2 (NF2). Consistent with these findings, subclonal somatic mutations in other known and putative Hippo pathway–related genes were identified in independent MTSCCs. Further, recurrent losses in chromosomes that harbor PTPN14, NF2, and other Hippo pathway TSGs, such as salvador homolog 1 (SAV1), result in biallelic loss of these TSGs. Consistent with these findings, 85% of MTSCCs with biallelic loss of Hippo pathway TSGs, and 90% of MTSCCs overall, exhibited nuclear immunopositivity of Yes associated protein 1 (YAP1), a downstream effector of the Hippo pathway. Taken together, these findings describe the mutational landscape of a rare type of renal cell carcinoma and identify MTSCC as another cancer characterized by dysregulated Hippo pathway signaling.

See article, p. 1267.

Familial risk of breast cancer is associated with loss-of-function mutations in homologous recombination repair genes, such as BRCA1/2, as well as common single-nucleotide polymorphisms. However, the contribution of genomic variants of lower penetrance to hereditary breast cancer risk remains unclear. Vijai, Topka, and colleagues identified a recurrent protein-truncating mutation in ERCC3 in individuals of Ashkenazi Jewish ancestry with BRCA1/2 wild-type breast cancer. The germline R109X mutation in ERCC3, which encodes an ATP-dependent DNA helicase that functions in nucleotide excision repair and as a component of the TFIIH transcription factor complex, was observed in three members of the same kindred, as well as in two other unrelated individuals. Complementation assays in ERCC3-deficient cells and genome-edited ERCC3-mutant human mammary epithelial cells showed that the ERCC3 R109X mutation resulted in decreased ERCC3 expression, less efficient DNA repair, and reduced survival in response to DNA damage–inducing agents compared with wild-type ERCC3, indicating that this mutation is functionally hypomorphic. In addition, heterozygosity for the ERCC3 R109X variant was significantly associated with increased risk of breast cancer in people of Ashkenazi Jewish ancestry, in particular patients with the ER-positive breast cancer subtype. These data demonstrate that the ERCC3 R109X mutation confers a moderate risk for breast cancer and suggest that ERCC3 may represent a potential therapeutic target in a subset of patients with hereditary breast cancer.

See article, p. 1276.

Telomerase reverse transcriptase (TERT), which encodes the catalytic protein component of telomerase, is epigenetically silenced in most adult cells but reactivated in the majority of human cancers. Mutations in the TERT promoter are common in several types of cancer and result in TERT reactivation. These mutations create de novo binding sites for ETS family transcription factors including GABPA, but the mechanism by which this promotes TERT reactivation remains unclear. Akıncılar and colleagues found that TERT promoter mutations were associated with enrichment of the active histone marks H3K4me3 and H3K9ac, and GABPA bound specifically to mutant promoters to form long-range chromatin interactions. Using CRISPR/Cas9 to convert the mutant TERT promoter to wild-type reduced activating histone marks and disrupted intrachromosomal interactions, and GABPA depletion had similar effects, suggesting that GABPA mediates the long-range interactions. Moreover, GABPA-mediated long-range interactions were required for transcription from mutant TERT promoters. Active histone marks can recruit chromatin remodelers, and accordingly, occupancy of the acetyl-lysine reader BRD4 was enhanced at active TERT promoters and associated with active histone marks. In melanoma and glioblastoma cell lines, reverting the mutant TERT promoter to wild-type reduced active histone marks, suppressed POL2, GABPA, and BRD4 binding to the TERT promoter, and disrupted long-range chromatin interactions. Further, BRD4 inhibition suppressed transcription from the mutant TERT promoter. Collectively, these findings reveal a mechanism by which GABPA selectively activates TERT in cells with cancer-specific promoter mutations, suggesting the possibility of developing drugs that target TERT expression specifically in cancer cells.

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