Oncogenic NRG1 fusions: A new hope for targeted therapy in pancreatic cancer

SUMMARY:

Approximately 8-10% of pancreatic ductal adenocarcinoma cases are KRAS wild-type. In a subset of these tumors, NRG1 gene fusions have been identified as targetable oncogenic drivers, a discovery that highlights the importance of deep molecular characterization for KRAS wild-type pancreatic cancers and provides a novel treatment strategy in this disease.

In this issue of Clinical Cancer Research, Jones and colleagues report the identification of oncogenic Neuregulin 1 (NRG1) gene fusions in KRAS wild-type pancreatic cancer patients (1). As part of a prospective clinical trial, the authors performed whole genome sequencing (WGS) and whole transcriptome analysis on 47 patients with metastatic pancreatic ductal adenocarcinoma and identified KRAS mutations in 94% (44/47) of cases. In all three patients with KRAS wild-type tumors, the authors discovered translocations affecting the NRG1 gene that were predicted to be in-frame and preserved the EGF-like domain of the NRG1 protein (Figure 1A). Two patients harbored a recurrent ATP1B1 fusion partner, and the third patient demonstrated a complex structural rearrangement in which exons 6 and 7 of NRG1 (harboring the EGF-like domain) were inserted between exons 15 and 16 of the APP gene. All three NRG1 fusion partners were predicted to donate trans-membrane domains to the NRG1 fusion protein, and gene expression analysis demonstrated that relevant NRG1 transcripts were upregulated in all three cases. Given that NRG1 is known to bind the ERBB3 receptor which heterodimerizes with ERBB2 to activate downstream signaling pathways, the authors treated two patients with the pan-ERBB receptor inhibitor afatinib and observed partial responses to therapy. This report describes NRG1 fusion proteins as an important oncogenic driver in a subset of KRAS wild-type pancreatic cancers and suggests a new therapeutic strategy for patients harboring these lesions.

Figure 1:

NRG1 fusions are targetable alterations in KRAS wild-type pancreatic ductal adenocarcinoma. (A) Model for how NRG1 fusion proteins harboring the EGF-like domain bind ERBB3, which heterodimerizes with ERBB2 and activates oncogenic levels of RAS effector signaling pathways (see also ref. 2). RTK, receptor tyrosine kinase; EGF, EGF-like domain of NRG1 (B) Actionable oncogenic driver events in KRAS wild-type pancreatic ductal adenocarcinoma.

NRG1 belongs to a well described set of ligands for the ERBB family of transmembrane receptor tyrosine kinases (RTKs). NRG1 binds to ERBB3, which is kinase-deficient but heterodimerizes with ERBB2. The ERBB2:ERBB3 heterodimer potently activates wild-type RAS and the Map-kinase (MAPK) and PI3-kinase (PI3K) mitogenic signaling pathways (2). NRG1 rearrangements were first described in invasive mucinous adenocarcinoma of the lung (3), where they occur in approximately 30% of cases, but these fusions have now been identified in multiple other tumor types, including cancers of the pancreas, bladder, liver, head and neck, kidney, ovary, uterus and prostate (1,4,5). NRG1 rearrangements have been identified with multiple fusion partners, including CD74, ATP1B1, APP, CDH6, SARAF, ROCK1 and others, with the occurrence of complex structural rearrangements with multiple fusion partners in some tumors (1,3,4,6). The EGF-like domain of the protein is maintained in essentially all fusions that have been identified (1,4,6), and is thought to mediate ERBB3 binding and subsequent oncogenic signaling (Figure 1A). Additionally, NRG1 transcript levels have been shown to be elevated in fusion-positive cancers (1,4). Functional studies of several NRG1 fusions have supported that these events are oncogenic drivers across a range of in vitro and in vivo models (3,4,6). In addition to NRG1 translocation events, NRG1-mediated autocrine and paracrine signaling loops have also been implicated in cancer (2).

Heining and colleagues have previously described a striking incidence of targetable oncogenic gene fusions in young adults with pancreatic cancer (6). In a small cohort of 17 patients, four tumors were KRAS wild-type. Three of the KRAS wild-type pancreatic cancers harbored a NRG1 fusion and the tumor from the fourth patient had an NCOA4-RET fusion. Two patients with NRG1 fusions were treated with therapy directed at ERBB signaling (afatinib or erlotinib/pertuzumab) and demonstrated clinical benefit. Thus, combined analysis of available published cases demonstrates that NRG1 fusions occur in a subset of KRAS wild-type pancreatic cancer patients, with ATP1B1 and APP being recurrent fusion partners, and that ERBB receptor-directed therapy may have clinical efficacy.

The work of Jones et al. and Heining et al. add important data to the growing appreciation for the relevance of alternative oncogenic driver events that occur in KRAS wild-type pancreatic cancer (Figure 1B). Analyses done by The Cancer Genome Atlas (TCGA) project (7), the International Cancer Genome Consortium (ICGC)(8) and others have suggested that multiple targetable oncogenic events occur in these KRAS wild-type pancreatic cancers, including mutations or in-frame deletions in BRAF (8,9), and mutations or amplifications in ERBB2, MET, FGFR1 and other RTKs (7,8). Additionally, other oncogenic rearrangements have been identified in known oncogenes such as ROS1 (9), ALK1 (10) and RET (6). Notably, in the report by Jones and colleagues (as well as in the prior report by Heining et al.), no other alternative drivers were identified in these KRAS wild-type tumors, further suggesting that the NRG1 fusion proteins are oncogenic drivers of these cancers. Collectively, these studies indicate that KRAS wild-type tumors can be driven by multiple different potentially targetable oncogenes, many of which provide an alternative pathway to activate oncogenic levels of RAS signaling. Thus, deep genomic and transcriptomic analyses – preferably utilizing WGS and whole transcriptome approaches to identify fusion events – should be pursued for KRAS wild-type tumors in a prospective manner to discover targetable events early in a patient’s disease course. Moreover, given the relatively small numbers of KRAS wild-type pancreatic cancer patients with any particular actionable alteration, development of multi-center umbrella trials with treatment arms targeting distinct oncogenic events should be prioritized to test targeted therapy strategies in sufficient numbers of patients.

Given the solid evidence that NRG1 fusions are oncogenic and targetable, the field must further study these NRG1 translocated tumors to understand the best approach to therapy. Multiple examples have now been published of pancreatic cancer patients with NRG1 fusions demonstrating partial responses to ERBB family receptor-directed therapy with afatinib (1,6), and these data should fuel further work in relevant patient-derived models to understand how to maximize the efficacy of single-agent and combination therapy approaches with afatinib or other ERBB-directed kinase inhibitors. Drilon and colleagues recently described a small series of lung adenocarcinoma patients with NRG1 fusions that did not readily respond to afatinib monotherapy; however, they also reported an exceptional response for one of these patients on a phase I trial of a novel monoclonal antibody that directly targets ERBB3 (4). Development of additional ERBB3-targeting therapies or combination strategies that target the ERBB-RAS-MAPK signaling pathway also remains an important priority. Given that treatment responses may vary based on the specific NRG1 fusion protein or the type of tumor in which it arises, appropriately powered basket trials that include a variety of lineages and genetic contexts may be useful to investigate the efficacy of novel therapies directed at these NRG1 fusions.

The discovery of NRG1 gene fusions in KRAS wild-type pancreatic cancers identifies a tractable therapeutic target in a disease that has long been thought to lack such actionable events, thus providing much needed hope for therapeutic benefit in a subset of patients suffering from this difficult disease.