In This Issue

See article, p. 993.

Biochemical and functional studies have shown that oncogenic RAS mutations drive transformation in a dominant, gain-of-function manner. However, genetic data demonstrating loss of the normal RAS allele in some tumors have also suggested that wild-type RAS has tumor suppressor function in some tumors. NRAS is commonly mutated in human chronic myelomonocytic leukemia, juvenile myelomonocytic leukemia, and acute myeloid leukemia (AML). Xu and colleagues studied the contributions of mutant and wild-type NRAS during myeloid transformation through the use of conditional alleles that allowed simultaneous deletion of wild-type Nras and activation of oncogenic Nras^G12D in the hematopoietic lineage after birth. In contrast to homozygous Nras^G12D mice, which develop overt myeloproliferative neoplasms and T-lineage acute lymphoblastic leukemia, age-matched Nras^G12D hemizygous mice lacking the wild-type allele had no hematologic abnormalities. Moreover, enforced expression of wild-type Nras had no effect on Nras^G12D-induced myeloid progenitor colony growth. Instead, increased levels of expression of Nras^G12D promoted cytokine-independent colony growth, increased basal activation of downstream RAS effectors, and conferred sensitivity to MEK inhibition. Consistent with these data, loss of the normal Nras allele in primary AML cells was associated with duplication of oncogenic Nras^G12D. Together, these findings suggest that increased NRAS dosage, and not loss of wild-type NRAS, drives hematopoietic transformation. Given that NRAS does not act as a tumor suppressor in this context, attempts to restore or augment wild-type NRAS function will likely not be beneficial in NRAS-mutant human leukemias.

See article, p. 1002.

The silencing of tumor suppressor genes via aberrant DNA hypermethylation is a common feature of many tumors, including diffuse large B-cell lymphoma (DLBCL), and contributes to chemotherapeutic resistance. DNA methyltransferase (DNMT) inhibition has been proposed as a means to reverse the inactivation of these genes and modulate chemosensitivity, but the clinical efficacy of DNMT inhibitors in DLBCL is unknown. Clozel and colleagues found that low-dose treatment with the DNMT inhibitor 5-aza-2′-deoxycytidine (decitabine) triggered DNA demethylation and apoptosis in a subset of DLBCL cell lines and synergized with doxorubicin to inhibit the growth of chemosensitive DLBCL cells. In addition, decitabine induced DNA hypomethylation and a senescence-like phenotype in chemorefractory DLBCL cells and enhanced the chemosensitivity of these cells to doxorubicin without toxicity in vivo, suggesting that DNMT inhibitor-mediated epigenetic reprogramming overcomes chemotherapy resistance. Consistent with this idea, decitabine upregulated the expression of several genes that were silenced by hypermethylation in refractory DLBCL cell lines and primary tumors, in particular SMAD1, which encodes a mediator of TGFβ signaling. SMAD1 reactivation sensitized resistant cells to growth inhibition by doxorubicin, whereas SMAD1 depletion augmented chemoresistance. Furthermore, in a phase I clinical trial of newly diagnosed, high-risk patients with DLBCL, DNMT inhibitor pretreatment followed by standard chemoimmunotherapy was well tolerated and resulted in a high rate of complete remission; this antilymphoma effect was associated with decreased SMAD1 methylation and increased chemosensitivity ex vivo, supporting further investigation of this therapeutic combination in DLBCL.

See article, p. 1020.

A hallmark of castration-resistant prostate cancer (CRPC) is continued dependence on androgen receptor (AR) signaling. The second-generation anti-androgens enzalutamide (MDV3100) and ARN-509 are effective in patients with refractory CRPC, but resistance almost invariably develops through unknown mechanisms. Joseph and colleagues generated enzalutamide- and ARN-509–resistant prostate cancer cell lines and noted that in some cell lines enzalutamide and ARN-509 actually stimulated cell proliferation and AR transcriptional activity. Because resistance to first-generation anti-androgens can arise through AR mutations that confer ligand-specific agonist activity, the authors sequenced the AR ligand-binding domain and identified a recurring, previously uncharacterized F876L mutation. Mutation of the F876 residue significantly increased the affinity of AR for enzalutamide and ARN-509 and promoted agonist-like conformational changes. Moreover, expression of AR^F876L in prostate cancer cell lines induced AR binding to target gene promoters and conferred resistance to enzalutamide and ARN-509 both in vitro and in xenograft tumors growing in castrated immunodeficient mice. To determine whether AR^F876L is a clinically relevant resistance mutation, the authors screened circulating tumor DNA from participants in a phase I trial of ARN-509. The AR^F876L mutation was absent from pretreatment samples but was observed in 3 patients after ARN-509 treatment, all of whom had rising prostate-specific antigen levels indicative of progressive disease. Next-generation agents that maintain antagonist activity in the context of AR^F876L may therefore benefit a population of patients with CRPC who develop resistance to enzalutamide or ARN-509.

See article, p. 1030.

The androgen receptor (AR) antagonist enzalutamide (also known as MDV3100) is initially effective in a significant proportion of patients with metastatic castration-resistant prostate cancer (CRPC), but many patients ultimately develop resistance. To identify potential resistance mechanisms, Korpal and colleagues created a model of spontaneous enzalutamide resistance using LNCaP prostate cancer cells and found a recurring AR mutation resulting in an F876L substitution in the AR ligand-binding domain that was associated with enzalutamide resistance both in vitro and in vivo. Computational modeling suggested that the F876L mutation would alter the interaction between enzalutamide and AR such that enzalutamide would no longer act as an AR antagonist and would potentially act as an AR agonist. Consistent with a mechanism of enzalutamide resistance involving an “agonist switch,” the F876L mutant selectively increased AR-dependent transactivation of target genes in the presence of enzalutamide and conferred resistance to enzalutamide in long-term culture. Moreover, F876L-expressing prostate cancer cells grown in vivo in the castrate setting were dependent on enzalutamide for growth. Of note, enzalutamide-resistant cells with the AR^F876L mutation retained sensitivity to other antiandrogens such as bicalutamide and were sensitive to inhibitors of cyclin-dependent kinases that act downstream of AR. Screening for the AR^F876L mutation may therefore identify those patients with CRPC who are likely to be resistant to enzalutamide, could benefit from enzalutamide withdrawal, or respond to alternative therapeutic strategies.

See article, p. 1044.

Chromosome 3q26 is amplified in various cancer types, including ovarian, breast, and non–small cell lung cancer, and is associated with poor prognosis. To identify the cancer driver genes targeted by this recurrent amplification, Hagerstrand and colleagues systematically evaluated the function of each of the 20 candidate genes contained within the minimal commonly amplified region. Gain- and loss-of-function studies showed that SEC62 homolog (also known as TLOC1) was selectively required for the proliferation of cancer cell lines with 3q26 amplification and enhanced anchorage-independent growth in immortalized mammary epithelial cells, whereas SKI-like oncogene (SKIL) induced cell invasion. Furthermore, both TLOC1 and SKIL stimulated xenograft tumor formation in vivo, and combined expression of these genes cooperatively augmented transformation independent of the adjacent oncogenes PIK3CA, SOX2, and TP63. Interaction of TLOC1 with DEAD box helicase 3, X-linked (DDX3X) was necessary for TLOC1-mediated transformation; TLOC1 and DDX3X formed a complex with translational regulatory proteins and decreased eukaryotic translation initiation factor 4E binding protein 1 phosphorylation, thereby modulating the ratio of cap-dependent translation. In contrast, SKIL-driven cell invasion was dependent on upregulation of proinvasive genes, in particular snail family zinc finger 2 (SNAI2, which encodes SLUG), a master regulator of epithelial–mesenchymal transition. These findings identify TLOC1 and SKIL as coamplified driver genes that promote malignant transformation via regulation of distinct tumor phenotypes.

See article, p. 1058.

Aberrant activation of fibroblast growth factor receptor (FGFR) signaling is a common event in multiple tumors including bladder cancer, supporting the clinical development of FGFR inhibitors. However, the factors that determine sensitivity to FGFR inhibition and whether mechanisms of resistance vary in cells expressing different FGFR family members remain unclear. To investigate these questions, Herrera-Abreu and colleagues performed parallel RNA interference screens using a panel of cancer cell lines with representative genetic alterations in FGFR1, FGFR2, or FGFR3. Intriguingly, depletion of EGF receptor (EGFR) enhanced the sensitivity of FGFR3-activated bladder cancer cells but not other FGFR-mutant cells to pan-FGFR blockade, suggesting that intrinsic EGFR activation confers oncogene-specific or tumor type–specific resistance. A subset of FGFR3-mutant cells was dependent on FGFR3 and initially responded to FGFR inhibition, but later restored MAPK pathway activity via upregulation of EGFR signaling. EGFR activation in these cells was mediated by relief of FGFR3-driven negative feedback and impaired signal termination in part due to defective receptor internalization. In contrast, a second group of FGFR3-mutant cells was primarily dependent on EGFR and intrinsically resistant to FGFR inhibition due to EGFR-mediated repression of FGFR3 transcription; delayed derepression of FGFR3 expression in response to gefitinib limited the sensitivity of these cells to EGFR blockade. However, combined inhibition of EGFR and FGFR3 synergistically reduced colony formation and diminished tumor growth in vivo, suggesting that this therapeutic strategy may overcome these resistance mechanisms in FGFR3-mutant bladder cancer.

See article, p. 1072.

Alpha-difluoromethylornithine (DFMO) is a former chemotherapeutic agent that has recently been repurposed as part of a colorectal cancer chemoprevention regimen in clinical trials. DFMO is a direct inhibitor of the polyamine biosynthetic enzyme ornithine decarboxylase (ODC), but mucosal polyamine levels do not directly correlate with colorectal cancer risk in DFMO-treated patients. To determine the chemopreventive mechanism of DFMO, Witherspoon and colleagues performed unbiased metabolite screens in DFMO-treated human colorectal cancer cells and intestinal tumors from Apc-mutant mice and found that, in addition to increasing levels of the ODC substrate ornithine and decreasing levels of ODC product polyamines, DFMO treatment induced a significant decrease in levels of the methionine cycle intermediate S-adenosylmethionine (SAM), which is also needed for polyamine biosynthesis, and led to an almost complete loss of cellular thymidine. Unexpectedly, thymidine supplementation counteracted the cytostatic effects of DFMO in colorectal cancer cells without restoring cellular polyamine levels, indicating that the effects of DFMO on cell proliferation are attributable to thymidine depletion, not inhibition of polyamine biosynthesis. Decreased SAM levels activate methylenetetrahydrofolate reductase, which converts the essential thymidine synthase cofactor 5,10-methylene tetrahydrofolate to 5-methyltetrahydrofolate, suggesting that thymidine depletion occurs following DFMO treatment due to diversion of a required tetrahydrofolate cofactor away from the key thymidine biosynthetic enzyme. These findings reveal a link between the polyamine and thymidine biosynthesis pathways and indicate that inhibition of thymidine synthase activity may be a shared mechanism of chemopreventative and chemotherapeutic agents used for colorectal cancer.

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