In This Issue

See article, p. 147.

The efficacy of targeted therapies is limited by the development of acquired resistance, which is generally assessed via molecular analysis of single tumor lesions. However, heterogeneity in resistance mechanisms between tumor lesions may affect the response to subsequent targeted therapies. To evaluate this hypothesis, Russo, Siravegna, Blaszkowsky, and colleagues analyzed tumor biopsies and serial circulating tumor DNA (ctDNA) profiles isolated from a patient with metastatic colorectal cancer who initially responded to treatment with the anti-EGFR antibody cetuximab. Sequencing analysis of the postprogression liver metastasis identified a mutation in MAP2K1, resulting in a K57T substitution in the MEK1 protein, which was not present prior to cetuximab therapy. Expression of mutant MEK1 was sufficient to confer resistance to EGFR inhibition, whereas resistance was overcome by dual inhibition of EGFR and MEK, prompting treatment of the patient with the combination of the anti-EGFR antibody panitumumab and the MEK inhibitor trametinib. Serial analysis of ctDNA profiles revealed a persistent decline in MAP2K1^K57T levels, which was accompanied by a radiologic reduction in the corresponding metastatic lesion. However, levels of a previously unrecognized KRAS^Q61H mutation, which was detected in a distinct liver metastasis that progressed on panitumumab and trametinib treatment, increased during therapy. These findings demonstrate that the emergence of distinct resistance mechanisms in individual metastatic lesions can result in lesion-specific responses to targeted therapy and underscore the value of integrating multiple tumor biopsies and ctDNA analysis to guide selection of subsequent therapies.

See article, p. 154.

Systemic histiocytic neoplasms are hematopoietic disorders that can be divided into Langerhans cell histiocytoses (LCH) and non-Langerhans histiocytoses (non-LCH), which differ in histology and type of infiltrating cells. About half of patients harbor activating BRAF^V600E mutations; however, other driving mutations have not been identified in non-LCH. To identify other genomic alterations, Diamond, Durham, Haroche, and colleagues performed whole-exome and transcriptome sequencing of tumor biopsies from patients with LCH or non-LCH. Recurrent activating mutations in MAP2K1 and ARAF were identified in non-LCH tumors, and activating MAPK alterations were found in all patients. In-frame kinase fusions were also identified in BRAF^V600E–wild-type non-LCH, including fusions involving BRAF, ALK, and NTRK1. In particular, the RNF11–BRAF fusion transformed cells to cytokine-independent growth and showed sensitivity to MEK inhibition. LCH samples were enriched for gene sets in late-state myeloid progenitors, granulocyte-monocyte progenitors, and classic dendritic cells, as well as cell-cycle and IL1 signaling genes, whereas non-LCH cells were enriched for hematopoietic stem cell, common myeloid progenitor, and monocyte gene sets, as well as lipid metabolism and adipogenesis genes. Based on these findings, three patients with refractory non-LCH were treated with targeted therapies; two patients with MAP2K1 mutations experienced rapid clinical responses to MEK inhibition with trametinib or cobimetinib, and a patient with an ARAF mutation experienced tumor regression following treatment with the RAF inhibitor sorafenib. This study identifies previously uncharacterized genetic alterations in histiocytic neoplasms and provides preliminary evidence demonstrating the efficacy of targeting these alterations.

See article, p. 166.

Approximately 10% of pancreatic ductal adenocarcinomas (PDAC) are familial, occurring in families with two or more affected first-degree relatives. However, the genetic basis of susceptibility in the majority of familial pancreatic cancers (FPC) is unknown. Roberts and colleagues used whole-genome sequencing of germline DNA from a large cohort of patients with FPC, along with whole-exome sequencing of select tumors, to identify FPC susceptibility genes. A focused analysis of rare premature truncating variants (PTV) found 6,114 private heterozygous PTVs in 4,553 genes. The majority of genes had only 1 heterozygous PTV; however, 1,077 genes had 2 or more heterozygous PTVs. The 16 most promising candidates contained 3 or more heterozygous PTVs, and have known roles in DNA repair and cancer (DNMT3A, BUB1B, and FANCC), or were previously identified as FPC susceptibility genes (ATM, BRCA2, CDKN2A, and PALB2). A more detailed analysis of 87 genes previously associated with hereditary cancer included deleterious variants of all types. Together, these data validate the role of previously reported FPC susceptibility genes, identify new candidate FPC susceptibility genes, and reveal the significant genetic heterogeneity underlying susceptibility in patients with FPC. Furthermore, the identification of susceptibility genes may allow for improved early detection of FPC and guide the development of personalized therapies.

See article, p. 176.

Adenoid cystic carcinomas (ACC) are rare malignant neoplasms occurring primarily in the salivary glands, and the majority are characterized by a recurrent translocation that generates the oncogenic MYB–NFIB chimeric transcription factor. Identification of MYB–NFIB target genes and additional genetic alterations may be helpful in developing therapeutic approaches. Brayer and colleagues analyzed low-quality RNA from 20 archival formaldehyde-fixed, paraffin-embedded ACC tumor samples by RNA sequencing to molecularly characterize ACC and identify fusion genes. In 8 tumors, the 3′end of MYB was lost due to a translocation with NFIB. Several alternatively spliced MYB transcripts were also detected. In addition, 3 tumors had breaks in the related MYBL1 gene, suggesting a chromosomal translocation. Further analysis identified a previously uncharacterized translocation involving MYBL1, which encodes the MYB family transcription factor A-MYB, and RAD51B that led to A-MYB truncation, and a second translocation fusing MYLB1 to NFIB. The MYB and MYBL1 fusions encode truncated proteins that display increased transcription activity. Tumors with high levels of MYB/MYBL1 had distinct gene expression patterns from tumors with low levels of these genes, and based on gene expression patterns, MYB and MYBL1 fusions were interchangeable, suggesting that ACC tumors can be driven by activation of either MYB or MYBL1. Furthermore, high MYB/MYBL1 expression was associated with higher tumor stage and poor outcome. These results suggest that MYB or MYBL1 fusions are dominant oncogenic drivers in ACC and demonstrate that analysis of archival samples provides insight into ACC biology.

See article, p. 188.

miRNAs modulate gene expression programs to control cellular processes. Analyses of bulk tissue from epithelial cancers have indicated that the levels of miR-143 and miR-145, two miRNAs shown to repress the expression of several oncogenes, are decreased compared to adjacent normal tissue, suggesting a tumor-suppressive role. In contrast, in normal tissues, the miR-143/145 cluster is expressed in the stroma and promotes epithelial growth during injury responses. To clarify the role of miR-143/145 in tumorigenesis and to elucidate the cell-type specificity of miR-143/145 expression in cancer, Dimitrova, Gocheva, and colleagues evaluated an autochthonous mouse model of Kras-mutant/Trp53-null (KP)–driven lung adenocarcinomas. Tumor-specific deletion or induced overexpression of miR-143/145 in KP mice did not affect tumor development or overall survival; however, organism-wide loss of miR-143/145 in vivo resulted in decreased lung tumor burden, suggesting that stromal miR-143/145 supports lung tumor growth. Expression of miR-143/145 was specifically detected in normal lung endothelial cells, and lung endothelial cells from miR-143/145^−/−;Kras–mutant mice exhibited decreased proliferation and impaired neoangiogenesis compared to miR-143/145^+/+ littermates. Expression data analysis revealed that calcium/calmodulin–dependent protein kinase 1D (CAMK1D) was a prominent target of miR-143/145. Mechanistically, increased expression of CAMK1D, which contains two conserved miR-145 sites in its 3′-UTR, recapitulated the mitotic defects observed in miR-143/145^−/− endothelial cells. Together, these results in autochthonous mouse models of lung cancer support a non–cell-autonomous tumor-promoting role of stromal miR-143/145 and shed light on the implications of targeting miR-143/145 in human cancer.

See article, p. 202.

T cell–mediated immunotherapy elicits a durable antitumorigenic response in a subset of melanoma patients; however, little is known about the underlying molecular factors that drive resistance in nonresponsive patients. Recently, activating mutations in the BRAF oncogene have been shown to suppress tumor-infiltrating T-cell function, suggesting that protumorigenic signaling pathways may also contribute to immune evasion. In line with this idea, concurrent loss of the PTEN tumor suppressor gene and BRAF mutation have been linked with poor outcome and resistance to BRAF inhibitors in melanoma, prompting Peng and colleagues to evaluate whether PTEN loss modulates the immunotherapeutic response. Suppression of PTEN in BRAF-mutant human melanoma cells inhibited T cell–mediated antitumorigenic effects both in vitro and in vivo. Clinically, PTEN-positive tumors were more sensitive to anti–programmed cell death 1 (PD-1) antibodies compared with PTEN-negative melanomas, which were characterized by reduced CD8⁺ T-cell infiltration and decreased capacity for expansion of tumor-infiltrating T cells. Mechanistically, loss of PTEN inhibited T-cell function via upregulation of immunosuppressive cytokines, including chemokine (C-C motif) ligand 2 and VEGF, and downregulation of autophagy regulators, such as autophagy-related 16-like 1. Overexpression of autophagy-initiating factors or selective inhibition of the PI3Kβ isoform in PTEN-silenced melanoma cells restored sensitivity to T cell–mediated immunotherapy. Together, these results suggest that PTEN loss confers resistance to T cell–mediated immunotherapy in melanoma by promoting immunosuppressive cytokine signaling and inhibiting autophagy and provide a rationale for combining PI3K–AKT inhibitors with immunotherapies.

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