In This Issue

See article, p. 657.

Oncogenic V600 mutations in BRAF are found in a variety of cancers, and BRAF inhibitors such as vemurafenib have achieved clinical success in the treatment of BRAF^V600-mutant melanoma. However, the clinical results of RAF inhibitors in colorectal cancer have been disappointing, prompting questions about the general efficacy of RAF inhibition across RAF-mutant cancers. Subbiah and colleagues evaluated the efficacy of vemurafenib in a phase II basket trial including 172 patients with 26 types of non-melanoma cancer. Vemurafenib was generally well tolerated, with the most common treatment-emergent adverse events being arthralgia, fatigue, and hyperkeratosis, and no deaths were deemed to have been treatment related. The overall response rate including all 26 cancers was 33%, with 3% of patients attaining complete responses and 30% attaining partial responses, and the median duration of response was 13.1 months. Responses occurred in patients with 13 of the 26 tumor types; notably, this included some tumor types typically considered to be treatment refractory, such as cholangiocarcinoma, sarcoma, glioma, neuroendocrine carcinoma, and salivary gland carcinomas. One limitation of this study is that it was initiated prior to the finding that combined BRAF and MEK inhibition may be more beneficial than BRAF inhibition alone, and before next-generation sequencing and blood-based circulating tumor DNA sequencing were widely available. In summary, this trial demonstrates that the clinical utility of BRAF inhibitors may not be limited to BRAF-mutant melanoma and suggests that larger trials of BRAF inhibitors in other BRAF-mutant cancers should be initiated.

See article, p. 664.

Cell-free DNA (cfDNA) circulating in plasma is fragmented, and the fragmentation patterns, such as fragment size and the genomic locations at which fragmentation most often occurs, may reveal information about the cells of origin. Jiang, Sun, Peng, and colleagues investigated whether plasma DNA-end motifs—that is, the short sequences of nucleotides at plasma DNA ends—could be used as markers for pathophysiologic states, such as cancer, pregnancy, and transplantation. Analysis of hepatocellular carcinomas (HCC) and adjacent healthy tissue revealed differences in nuclease expression between the two tissue types. Comparing plasma DNA-end motifs in patients with HCC, patients with chronic hepatitis B virus infection, and healthy controls also showed that there were significant differences among the groups, and the motif diversity score (MDS) was higher in patients with HCC, indicating that these patients had greater diversity in their plasma DNA-end motifs. The higher MDS was also observed in other cancer types, including colorectal cancer, lung cancer, nasopharyngeal cancer, and head and neck squamous cell carcinoma, indicating the broader relevance of the finding. Further experiments revealed that the frequencies of various plasma DNA-end motifs and the MDS could be used to determine the tissue of origin of the cfDNA, such as from a cancer or an organ (for example, placenta or liver). Together, the results of this proof-of-concept work demonstrate that plasma DNA-end motif profiling can be used to detect the presence of multiple types of cancer in the context of liquid biopsy, justifying larger studies to ascertain the clinical utility of this technique.

See article, p. 674.

Mutations or amplifications affecting the gene encoding HER2 (ERBB2) occur in a subset of lung cancers, but no HER2-targeted treatments for lung cancers are approved. Li, Michelini, Misale, and colleagues investigated the mechanism and utility of the antibody–drug conjugate trastuzumab emtansine (T-DM1), which links the HER2 antibody trastuzumab to the cytotoxic microtubule inhibitor emtansine, in lung cancers. In vitro experiments showed that ERBB2 mutations increased HER2 internalization and T-DM1 activity. Further, experiments using functional imaging confirmed that T-DM1 accumulates in ERBB2-mutant or -amplified patient lung tumors. Analysis of data from 49 patients with ERBB2-mutant or -amplified lung cancers from a phase II trial of T-DM1 revealed an overall response rate of 51%. In vitro experiments demonstrated that cotreatment of ERBB2-mutant or -amplified lung cancer cells with irreversible HER2 inhibitors increased ubiquitination and internalization of HER2, translating into increased efficacy of the combination with respect to single agents in vivo. Clinically, T-DM1 with the irreversible pan-HER inhibitor neratinib showed efficacy in a patient with ERBB2-amplified breast cancer that had recurred after multiple lines of HER2-targeted therapy, including T-DM1. After obtaining evidence of efficacy in a patient-derived xenograft model of ERBB2-amplified and -mutant lung cancer, the corresponding patient was treated with trastuzumab deruxtecan (T-DXd), which links trastuzumab to the topoisomerase I inhibitor deruxtecan, following progression on T-DM1. This produced a durable partial response. These results show that patients whose disease progresses on T-DM1 may benefit from the addition of an irreversible HER inhibitor or from switching to T-DXd, potentially informing future clinical trials.

See article, p. 688.

Although HER2 mutation or overexpression is observed in some non–small cell lung cancers (NSCLC) and colorectal cancers, no HER2-targeted therapies are available for affected patients. Tsurutani and colleagues conducted a phase I trial of the antibody–drug conjugate trastuzumab deruxtecan (T-DXd), which combines the HER2 antibody trastuzumab with the topoisomerase I inhibitor deruxtecan, in patients with advanced solid tumors. Fifty-nine heavily pretreated patients with HER2-mutant or -overexpressing cancers (including 20 with colorectal cancer, 18 with NSCLC, and 21 with other solid tumors) received at least one dose of T-DXd. The best responses were observed in patients with HER2-mutant NSCLC, among whom the objective response rate (ORR) was 72.7% (8 of 11 patients), with a median progression-free survival of 11.3 months and a median overall survival of 17.3 months. One limitation of this analysis is the small number of patients with each tumor type; however, the especially promising ORR in patients with HER2-mutant NSCLC warrants further study. At least one treatment-emergent adverse event, most commonly gastrointestinal or hematologic, occurred in all patients, and two patients died of drug-related side effects. Interstitial lung disease or pneumonitis occurred in seven patients, indicating a need to monitor for these reactions. Given the manageable safety profile and early evidence of efficacy of T-DXd in tumor types other than breast and gastric cancers—the only diseases for which HER2-targeted treatment is currently approved—phase II studies of T-DXd in HER2-mutant or -overexpressing advanced NSCLC and HER2-expressing recurrent or metastatic colorectal cancer are now in progress.

See article, p. 702.

The potency of CAR-T cells relies on expression of antigens on target cells: Cancer cells expressing low levels of the target antigen may evade CAR-T cells, leading to treatment resistance. In B-cell malignancies, Majzner and colleagues found extensive inter- and intrapatient heterogeneity in expression of CAR-T cell target antigens, including CD19, on cell surfaces and showed that CAR-T cell efficacy is proportional to antigen density. In vitro and in vivo, CAR-T cells with CD28 costimulatory domains (as in axicabtagene ciloleucel) were more effective against low-antigen-density target cells than were CAR-T cells with 4-1BB costimulatory domains (as in tisagenlecleucel). The greater efficacy of CAR-T cells expressing CD28í compared with those expressing 4-1BBí was due to increased signal strength, and enhancing signal strength in the latter CAR-T cells increased their ability to recognize low-antigen- density target cells. Conversely, reduction of signal strength caused decreased killing of low-antigen-density tumor cells. Additionally, incorporation of the CD28 hinge–transmembrane (H/T) domain (as in axicabtagene ciloleucel) rather than the CD8 H/T domain (as in tisagenlecleucel) resulted in greater efficacy against low-antigen-expressing target cells in vitro and in vivo. Further analysis revealed that immunologic synapses formed by CD28 H/T domain–expressing CAR-T cells were more stable and efficient, and these CAR-T cells were better at forming clusters, a phenomenon that has been linked to T-cell activation. This work demonstrates that the sensitivity of CAR-T cells to antigen density on target cells is tunable, providing new avenues for development of CAR-based therapies.

See article, p. 724.

Much research has been devoted to understanding the roles of variants in protein-coding regions of the genome in cancer; however, less study has been devoted to characterizing the contributions of variants in noncoding parts of the human DNA sequences. Using targeted resequencing, Li, Zhang, Liu, Liu, and colleagues discovered 1,836 recurrently mutated cis-regulatory elements (CRE) that contained noncoding variants associated with leukemias. They engineered high-throughput CRISPR–dCas9-based functional screens and pinpointed 131 of these CREs as potentially oncogenic and 87 CREs as possibly tumor suppressive. Notably, one of the top-ranked oncogenic enhancers was located approximately 135 kb upstream of KRAS, and further experiments showed that this CRE variant was a gene-distal enhancer of KRAS in acute myeloid leukemia (AML) cells that was required for AML cell growth in vitro and in vivo. Also of note, another of the recurrently mutated enhancers in AML cells controlled PER2, which encodes a negative regulator of the core circadian-rhythm genes CLOCK and BMAL1, and deeper investigation confirmed that PER2 and this associated CRE variant were tumor suppressive in AML. Interestingly, the relevant noncoding variants in the KRAS and PER2 enhancers were found in proximity to nuclear receptor binding sites, and these variants were shown to influence nuclear receptor binding and consequent transcriptional activation. Experiments using mice xenografted with the knock-in cells confirmed the oncogenic role of the KRAS enhancer variant and the loss of the normal tumor-suppressive role of the PER2 variant. Together, these findings establish a noncoding variant–mediated connection between enhancer dysregulation and nuclear receptor signaling in leukemia.

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