In This Issue

See article, p. 1182

Mutations in RAS/RAF/MEK/ERK pathway members that constitutively activate ERK lead to increased signaling through this pathway, driving tumor growth. Allosteric MEK inhibitors have been developed and used in combination with RAF inhibitors to treat BRAF^V600 mutant—driven cancers. Resistance mechanisms to RAF and MEK combination therapy have been defined; they mostly function by decreasing the pharmacologic activity of RAF inhibitors. The mechanisms that underlie resistance to MEK inhibitors in the clinic remain largely unknown. Gao and colleagues report a patient with colon cancer with a MEK1 mutation (V211D) acquired during treatment with the MEK inhibitor binimetinib and the anti-EGFR antibody panitumumab. This mutation was not seen in any of more than 30,000 sequenced clinical specimens and appeared to be treatment-emergent. Unlike wild-type MEK, the phosphorylation of MEK1^V211D was only partially dependent on the upstream RAF kinase and exhibited elevated baseline catalytic activity. MEK1^V211D was not susceptible to allosteric MEK inhibitors in vitro or in a mouse fibroblast line; however, MEK1^V211D was sensitive to ATP-competitive inhibitors, which inhibited the activity of wild-type and V211D MEK1 equipotently in vitro and in vivo. Experiments using a patient-derived xenograft (PDX) model revealed that ATP-competitive MEK inhibitors caused tumor regression, inhibited ERK signaling, and induced cleaved caspase-3, a marker of apoptosis. Based on the biochemical characterization of this activating MEK mutation as a gatekeeper mutation for allosteric MEK inhibitors, investigation of selective ATP-competitive inhibitors of MEK or MEK inhibitors with different binding sites is warranted in patients with resistance to allosteric MEK inhibitors.

See article, p. 1192

The only potentially curative treatment for myeloproliferative neoplasms (MPN) is hematopoietic stem cell transplantation, and this approach is of limited use in this patient population due to factors including lack of donors, patients' advanced age, and comorbidities. Frameshift mutations affecting the C terminus of the MPN driver calreticulin (CALR) are common in MPN and generate a 36-amino-acid tail dissimilar to that of the wild-type CALR, making it an attractive neoantigen target. Cimen Bozkus and colleagues found that some patients with CALR-mutant MPN exhibited T-cell responses—predominantly the CD4⁺ T-cell phenotype—that were specific to mutant CALR. One possible explanation for the fact that the response was seen in only some patients is expression of programmed cell death protein 1 (PD-1) or cytotoxic T lymphocyte antigen 4 (CTLA4); indeed, in vitro experiments showed that blockade of checkpoint receptors restored the immune response to mutant CALR, and in vivo PD-1 blockade with pembrolizumab promoted a T-cell response in one patient with CALR-mutant MPN. Promisingly, it was possible to stimulate CD4⁺ and CD8⁺ T cells from healthy donors with no prior exposure to mutant CALR to specifically react to mutant CALR, ignoring wild-type CALR. Additionally, T cells recognize several endogenously processed and presented epitopes in the unique C terminus of mutant CALR. Collectively, these findings indicate that, in patients with CALR-mutant MPN, immunotherapies targeting the mutant CALR protein via neoantigen-specific vaccines or adoptive T-cell therapies may be worth investigating, as may immune checkpoint blockade therapies.

See article, p. 1208

Although more than half of patients with breast cancer are 60 years of age or older at the time of diagnosis, little is known about how age affects treatment responses. Using a mouse triple-negative breast cancer (TNBC) model, Sceneay, Goreczny, and colleagues discovered that agedanimals had diminished responses to immune checkpoint blockade (ICB) with anti—PD-L1 or anti-CTLA4. ICB therapy caused enrichment of interferon and inflammatory response pathways in responding tumors from young mice, but not those of aged mice. Increased age correlated with enhanced CD8⁺ T-cell effector memory and exhaustion phenotypes, and ex vivo stimulation of CD8⁺ T cells led to enhanced expression of checkpoint proteins (PD-1, CTLA4, LAG3, and TIM3) on the cells from aged mice. Even without treatment, there were marked differences in the immune contexture of the tumor microenvironment between young and aged mice; interferon and inflammatory responses and antigen presentation and processing genes were reduced in tumors from aged mice. Similarly, in patients with TNBC, those 40 years or younger in age had increased expression of antigen-processing and interferon and inflammatory response pathway genes relative to those 65 years or older. Intratumoral injection of a mouse-specific stimulator of interferon genes agonist, DMXAA, in combination with anti—PD-L1 or anti-CTLA4 led to reduced tumor growth and increased survival in aged mice, whereas addition of DMXAA did not increase efficacy of ICB in young mice. These results suggest that patients may benefit from assessment of interferon pathway status to determine need for additional immunotherapy in combination with ICB.

See article, p. 1228

Although epigenetics and metabolism both contribute to cancer pathogenesis, little is known about how epigenetics may influence metabolism in cancer. Gu, Liu, and colleagues found that mutations in some subunits of polycomb repressive complex 2 (PRC2) and an activating mutation in neuroblastoma RAS (NRAS^G12D) worked together to cause progression of myeloproliferative neoplasms (MPN) to primary myelofibrosis and promote leukemic transformation in human hematopoietic cells and mice. However, complete loss of PRC2 function via inactivation of the PRC2 component enhancer of zeste homolog 1 (EZH1) in EZH2-KO NRAS^G12D+/− mice abolished MPN progression. Because EZH1 function is dispensable in normal hematopoietic stem cells, EZH1-targeted therapies may be worth investigating for EZH2-mutant leukemia. RNA-sequencing experiments revealed that the branched-chain amino acid (BCAA) metabolism pathway was upregulated in hematopoietic/stem progenitor cells of EZH2-KO NRAS^G12D+/− mice relative to NRAS^G12D+/− mice, with BCAT1—the first enzyme involved in catalyzing BCAA transamination—being specifically epigenetically reactivated in EZH2-deficient cells to produce excess BCAAs. NRAS^G12D cooperated with loss of EZH2 to provide substrate for BCAT1-mediated transamination, and EZH2-deficient, NRAS-mutant leukemia-initiating cells had increased BCAA pools and were dependent on BCAA-driven mTOR signaling for survival. Because BCAT1 loss does not affect normal HSPCs and BCAAs can be restricted, BCAT1 and BCAAs may be useful targets to investigate in hematologic malignancies, particularly those with EZH2 loss and/or NRAS mutation.

See article, p. 1248

Glioma stem cells (GSC) contribute to pathogenesis and treatment inefficacy in glioblastoma. To determine molecular processes and potential vulnerabilities unique to GSCs, Gimple and colleagues identified GSC-specific superenhancer-associated genes that were associated with poor prognosis in glioblastoma and had elevated expression in glioblastoma compared with other brain cells. Among such genes, the long-chain polyunsaturated fatty acid (LC-PUFA) synthesis gene ELOVL2 supported GSC proliferation by promoting LC-PUFA synthesis, which GSCs relied on to maintain membrane composition and integrity and to support efficient EGFR signaling at the cell membrane. Demonstrating the LC-PUFA synthesis pathway's clinical relevance, high-grade gliomas had more direct products of ELOVL2 than low-grade gliomas, and lipid-metabolism synthesis gene signatures correlated with poor clinical outcome. Further, FADS2, the enzyme immediately downstream of ELOVL2 in LC-PUFA synthesis, was upregulated in glioblastoma and GSCs, and FADS2 depletion diminished GSC proliferation. GSCs treated with an EGFR inhibitor (lapatinib) and a FADS2 inhibitor (SC-26196) reduced proliferation more than either drug alone. Additionally, in a mouse xenograft model, ELOVL2 knockdown slowed tumor formation, and treatment with lapatinib, SC-26196, or both prolonged survival. Collectively, these results imply there may be a positive feedback loop in which LC-PUFA synthesis supports cell membrane structure, which promotes EGFR signaling. Because therapies that target EGFR alone have failed in glioblastoma, testing a combined approach that also targets LC-PUFA synthesis may be warranted.

See article, p. 1268

Activating mutations in KRAS are nearly ubiquitous in pancreatic ductal adenocarcinoma (PDAC), but oncogenic KRAS has not been effectively targeted. Humpton, Alagesan, and colleagues discovered that in mouse embryonic fibroblasts and pancreatic ductal organoids, oncogenic KRAS expression reduced cytoplasmic and mitochondrial reactive oxygen species (ROS) levels and resulted in a reduced mitochondrial network, as measured by mitochondrial mass and membrane potential per cell as well as quantitation of transmission electron micrographs. Oncogenic KRAS expression also led to increased expression of Bnip3l/Nix and an increase in NIX protein in mitochondria, causing increased selective removal of damaged mitochondria (mitophagy), which supports cancer progression in the context of cytotoxic stress by reducing mitochondrial ROS. Further, in a mouse PDAC model, NIX deletion delayed cancer progression and restored normal levels of functional mitochondria. In human pancreatic cancer cell lines, NIX depletion decreased mitophagic flux and diminished proliferation in glucose-limited conditions—similar to those in PDAC—likely by increasing the requirement for NADPH. Notably, however, increased mitochondrial clearance content in NIX-ablated KPC mice was eventually overcome by activating alternative pathways of mitochondrial clearance, so mitochondrial content was ultimately reduced to normal levels, allowing cancer progression despite a substantial extension of median survival. Demonstrating the potential clinical relevance of these findings, analysis of data from The Cancer Genome Atlas revealed that elevated NIX expression was associated with reduced survival in patients with PDAC. These results imply that therapy targeting mitophagy, possibly in combination with ROS-generating drugs, may be worth investigating for PDAC.

See article, p. 1288

Disease outcome in pancreatic ductal adenocarcinoma (PDA) is affected by the composition of inflammatory cells in the tumor microenvironment (TME), and unconventional T-lymphocyte populations are now being recognized as regulators of tumor immunity and disease. Hundeyin, Kurz, and colleagues found that one such T-lymphocyte population, TCRαβ⁺CD4⁻CD8⁻NK1.1⁻ innate αβ T cells (iαβT), comprised 10% of the infiltrating T-lymphocytes in a mouse model of PDA and in human PDA tumors. iαβTs in the TME were highly activated and exhibited a unique phenotype characterized by expression of the adhesion ligand JAML and the cytotoxic marker CD107a along with upregulation of the C-type lectin receptor Dectin-1 and several checkpoint receptors, ectonucleotidases, and costimulatory receptors. iαβT transfer also exerted protective effects against PDA in mice, reducing tumor growth, activating CD4⁺ and CD8⁺ T cells in the TME, and increasing expression of CD44, ICOS, and TNFα. Additionally, the TME of iαβT-treated tumors was immunogenically reprogrammed, and CD4⁺ and CD8⁺ T-cell populations increased substantially. Consistent with these mouse findings, autologous iαβT treatment in patient-derived organotypic tumor spheroids caused conventional T-cell activation. iαβTs did not reduce proliferation or promote lysis or apoptosis of PDA tumor cells, indicating that they are not directly cytotoxic to the PDA cells. However, iαβTs did indirectly aid the adaptive immune response by activating CCR5, thus promoting immunogenic macrophage polarization. Together, these results imply that cell therapies based on iαβTs may be of interest in PDA.

See article, p. 1306

PTEN, a tumor suppressor often disabled in cancers, is a lipid and protein phosphatase that has a variety of functions, including in repair of DNA double-strand breaks (DSB). In multiple cell lines, Zhang, Lee, Dang, and colleagues found that the presence of DNA DSBs increased PTEN phosphorylation at T/S398 by ATM kinase, promoting PTEN's interaction with mediator of DNA damage checkpoint protein 1 (MDC1), which mediated PTEN's interaction with the histone-lysine N-methyltransferase NSD2 (also known as MMSET/WHSC1). K349 of PTEN was then dimethylated by NSD2; this dimethylated lysine was subsequently recognized by the tudor domain of p53-binding protein 1, promoting PTEN accumulation on the DNA DSB sites. Moreover, NSD2-mediated K349 dimethylation of PTEN was required for efficient DNA DSB repair. The potential clinical relevance of this discovery is illustrated by the fact that inhibition of NSD2-mediated PTEN dimethylation made human colorectal carcinoma cells and cancer cells in xenografted mice more susceptible to PI3K-inhibitor treatment and DNA damage caused by etoposide or irradiation. This effect appeared to be dependent on PTEN protein phosphatase activity—mediated dephosphorylation of the histone protein γH2AX, a biomarker for DNA DSBs. In summary, this work provides insight into the molecular mechanism of PTEN-mediated DNA DSB repair and its possible role in cancer.

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