NRG1 Emerges as a Potentially Actionable Target in Lung Cancer and Solid Tumors

Stephen Liu, MD

Stephen Liu, MD

Alterations of theNRG1gene that are present at low rates in certain solid tumors have emerged as potentially actionable oncogenic drivers, with targeted therapies making their way to early clinical studies and into some basket trials.

When a patient with no smoking history presents with non—small cell lung cancer (NSCLC) and adeno­carcinoma histology, oncologists typically suspect a tumor that is positive for an oncogenic driver. Classic driver mutations—in genes such asEGFR (HER1),ALK,andROS1—are now regularly tested for in most patients with advanced lung disease based on guideline recommendations and the current standard-of-care treatment options. And, althoughNRG1alterations are rare, the potential actionability of that type of change in a patient population could change the course of their treatment and serve as reasonable motivation to con­sider this driver going forward.

NRG1fusions are present in 0.3% of NSCLC cases and in approximately 0.2% of solid tumors overall, according to work published inClinical Cancer Research. This gene has multiple potential fusion partners, and the specific genes found fused toNRG1have been shown to vary widely within and across tumor types. Among solid tumor types, NSCLC has one of the highest rates ofNRG1fusions; they are also found in renal cell, breast, ovarian, pancreatic, bladder, gallbladder, and colorectal cancers, as well as neuroendocrine tumors and sarcomas.¹

TheCD74-NRG1gene fusion was first described in 2014 by Lynnette Fernandez-Cuesta, PhD, and col­leagues in a study of lung adenocarcinoma; awareness of this genomic alteration has since increased.²Now, investigators likeStephen Liu, MD, of the Developmen­tal Therapeutics Program at Lombardi Comprehensive Cancer Center and MedStar Georgetown University Health in Washington, DC, are using experimental and currently approved agents in patients withNRG1fusions to determine whether these agents should be more widely applied in clinical practice.

Robert Doebele, MD, PhD

Robert Doebele, MD, PhD

Liu said that in his experience, patients withNRG1fusions do not respond well to standard immunother­apy or immunochemotherapy. This driver represents a viable therapeutic target, and identifying it prompts a corresponding change in treatment plan.

NRG1fusions belong to a class of oncogenes termed gene fusions, with large-scale chromosomal rearrange­ments bringing all or part of 2 different genes together to create 1 hybrid gene. “This is a different biological mechanism than other gene fusions that we commonly talk about, [such as those involving]ROS1,ALK,andNTRK,” saidRobert Doebele, MD, PhD, associate pro­fessor of medicine and director of the Thoracic Oncol­ogy Research Initiative at the University of Colorado Cancer Center in Aurora.

Gene fusions involving receptor tyrosine kinases (RTKs) eliminate the RTK regulatory domain and make the kinase domain constitutively activated.³ALK,ROS1,andNTRKare constitutively active as fusion proteins, for example, butNRG1fusions are different, said Doe­bele. The genetic rearrangement is structurally similar, but the biochemical mechanism leading to cancer cell growth is slightly different. TheNRG1gene encodes neuregulin 1, a ligand for the RTK human epidermal growth factor receptor 3 (HER3); binding results in indirect activation of downstream tumorigenic signal­ing pathways.⁴The gene fusion overexpresses part of the NRG1 protein at the cell surface, which then binds to HER3,¹said Doebele. HER3 then dimerizes with HER2, with subsequent effects similar to those of most other oncogenes: the activation of growth, prolifera­tion, and prosurvival signaling pathways in the cell.1 For many oncogenes, “the mutation causes direct acti­vation of the gene itself,” he said. Instead, theNRG1fusion overexpresses a ligand for other receptors, which is important to know for targeting.⁴

Current Status of Testing

Traditional DNA-based sequencing methods like Foun­dation Medicine’s FoundationOne assays will not find mostNRG1fusions, said Doebele. A small number of fusions, includingCD74-NRG1, can be identified because the panel covers theCD74gene. “As far as we know, these occur only or primarily in lung cancer,” he said.1 However,NRG1has multiple fusion partners beyondCD74, and not all can be detected with current DNA sequencing-based assays. Even more specialized molecular tests do not cover all mutations, Doebele said, so it is important to be aware of a test’s limitations.

Investigators like Liu and Doebele recommend anchored multiplex polymerase chain reaction assays that are agnostic to the partner to find rarer mutations.NRG1fusions are a key example of an instance in which RNA sequencing is a detection method superior to DNA sequencing, Liu said. When treating a patient for whom no mutations have been identified using a standard-of-care assay for NSCLC, Liu tries to acquire tissue for RNA sequencing to test for all possible drivers. Ideally, he does upfront dual extraction of RNA and DNA for patients with NSCLC.

“While these [fusions] are uncommon events, if we detect [just] one, it changes everything,” he said, including the understanding of disease biology, how to approach the patient’s cancer, and which therapies will be ineffective. Identification of a fusion is critical because it guides therapy from the beginning. He tells patients that finding a driver mutation is unlikely, but if one is found, it makes a big difference in management.

TreatingNRG1Fusions

As a heterogeneous group of drivers,NRG1fusions are challenging not only to detect but also to target. No therapies inhibitNRG1specifically, so investigators have tested agents with pan-HER inhibitory activity, such as afatinib (Gilotrif).^5,6

“The real heart of the initiation point of the sig­naling node is the HER2/HER3 heterodimer,” said Doebele. Targeting HER2, HER3, or both is the strategy behind antibody—drug conjugates such as ado-trastuzumab emtansine (T-DM1; Kadcyla) and trastuzumab deruxtecan (DS-8201) for HER2, for example, and U3-1402 for HER3, he said.

In his pulmonary oncology practice, Liu has anecdotally observed that patients with NSCLC andNRG1fusions, who are primarily nonsmok­ers with adenocarcinoma, do not respond well to chemoimmunotherapy or immunotherapy, regardless of PD-L1 status. This is consistent with the literature on similarROS1andALKdriver events, he said. The HER2 pathway is the most promising target, he said, and some patients have had success with afatinib. “We’ve seen dramatic responses and durable responses with afatinib,” he said, some lasting more than 2 years. At this point, Liu said it is important for investigators to determine why afatinib works so well for some patients, but not others.

To date, there are only a few published reports of patients receiving afatinib as targeted therapy forNRG1fusions; afatinib has been used almost exclusively to treatEGFR-mutated adenocarci­noma of the lung. “There have been [isolated] case reports in the literature where patients withNRG1fusions responded to afatinib, but there’s never been a clinical trial of afatinib to know the response rate,” Doebele said, adding that he and Liu collaborated on 7 new case stud­ies of patients withNRG1fusion-positive lung adenocarcinoma given afatinib after multiple previous lines of therapy, the results of which were presented at the European Society for Medical Oncology 2019 Congress.³

The TAPUR (Targeted Agent and Profiling Utilization Registry) study is one of several umbrella trials matching patients with various driver mutations to appropriate targeted thera­pies or immunotherapies. One treatment group in this prospective multiarm basket study comprises patients withNRG1fusions, who will receive afatinib (NCT02693535).

New Agents Under Examination

Other targeted therapies being evaluated for the treatment ofNRG1fusion—positive cancers include the antibody-dependent cell-mediated cytotoxicity–enhanced Biclonics agent zenocutu­zumab (MCLA-128),7 which targets the HER3 pathway. MCLA-128 data presented at the Interna­tional Conference on Molecular Targets and Can­cer Therapeutics showed tumor shrinkage in all 3 patients studied, 2 with pancreatic cancer and 1 with NSCLC and brain metastases.⁸

Tarloxotinib is a prodrug of an EGFR inhibitor that is activated under the hypoxic conditions common in tumor tissue.⁹It was developed by Rain Therapeutics, a company Doebele cofound­ed in 2017. “We designed [this agent because the] drugs targeting EGFR are toxic. That’s one of the limitations of afatinib,” said Doebele. Although afatinib is a good inhibitor of EGFR and HER2, its adverse effects (AEs) can include rash and diarrhea, he said. The incidence of AEs is significantly reduced with tarloxotinib because of its limited activity in healthy tissue.

Future Directions forNRG1

Obtaining the proper molecular testing is key for all patients, particularly those in whom the presence of a molecular driver is strongly sus­pected. “We’re on our way to identifying the best way to treat patients withNRG1fusion—positive cancer...the real challenge is identifying the patients [with the mutations]. There’s a fairly low frequency,” said Liu.

Doebele recommends using a targeted DNA panel for common drivers such asEGFR,KRAS,andHER2alterations and an anchored multiplex panel for the majority of gene fusions. With lung tumors, it is relatively easy to justify panel test­ing test­ing for many oncogenes. “We think most lung adenocarcinomas have an oncogenic driver,” he said. Pancreatic ductal adenocarcinoma (PDAC), however, is known to have almost exclusively activatingKRASmutations.¹⁰“One strategy is to say if they haveKRAS, then you stop testing. If they don’t, test forNRG1,” he said. One study showed that patients withKRASwild-type PDAC and anNRG1fusion benefited from afatinib.¹¹This was a small trial, however, and further investigation is needed, Doebele said.

Breast and ovarian cancer, in whichNRG1fusions are sometimes found, do not have many targeted therapies, making it harder to justify an anchored multiplex panel, Doebele said. “I believe in broad testing. If you look, you will find them. [Finding anNRG1fusion is] a relatively rare event—0.3% is 1 in 300 patients. [But] as we find more and more rare targets and you have a comprehensive panel strategy, your cumulative odds of finding something relevant go up.” For instance, 8 targets with an incidence of 0.3% each means more than a 2% chance of finding 1 of them, which clearly increases the justifica­tion for ordering an anchored multiplex panel.

Doebele acknowledges that some push­back exists against testing for rare drivers, with questions about whether it is worth the investment. “A lot of drugs we prescribe don’t have good response rates. When one works, it’s great. But immunotherapy doesn’t work in a lot of patients. A single dose of that [type of] drug costs more than one of these tests,” he said, noting that there are good therapies forNRG1fusions and trials investigating new treatments.

AsNRG1fusions are a relatively new target, knowledge of them “probably hasn’t filtered down to the community [setting] at this point,” Doebele said, adding that oncologists need to learn what the molecular panels cover.

“Oncologists today have been forced to become molecular pathologists, and it’s chal­lenging to interpret some of these results,” said Liu. He encourages oncologists to call the test­ing companies to walk them through a patient’s results and, at the very least, to carefully review the entire report. He said that critical findings are sometimes presented in bold print on page 1 and easy to interpret; however, the results may be “buried” in a long report in other instances. “What a tragedy [it would be] to have anNRG1fusion detected and miss it because you’re only looking at the top-line results,” he declared.

With no FDA-approved drugs forNRG1fusions, clinicians’ options include referring patients to an appropriate clinical trial or prescribing off label, said Doebele.NRG1fusions are less straightforward than some other drivers, but there may be multiple ways to target them.

References

1. Jonna S, Feldman RA, Swensen J, et al. Detection ofNRG1gene fusions in solid tumors. Clin Cancer Res. 2019;25(16):4966-4972. doi: 10.1158/1078-0432.CCR-19-0160.
2. Fernandez-Cuesta L, Plenker D, Osada H, et al.CD74-NRG1fusions in lung adenocarcinoma. Cancer Discov. 2014;4(4):415-422. doi: 10.1158/2159-8290.CD-13-0633.
3. Medves S, Demoulin J-B. Tyrsoine kinase gene fusions in cancer: translating mechanisms into targeted therapies. J Cell Mol Med. 2012;16(2):237-248. doi: 10.111/j.1582-4934.2011.01415.x.
4. Yun S, Koh J, Nam SK, et al. Clinical significance of overexpression of NRG1 and its receptor HER3 and HER4, in gastric cancer patients.Gastric Cancer. 2018;21(2):225-236. doi: 10.1007/s.10120-017-0732-7.
5. Liu SV, Duruisseaux M, Tolba K, et al. Targeting NRG1-fusions in multiple tumour types: afatinib as a novel potential treatment option.Ann Oncol. 2019;30(suppl 5; abstr 1969P). doi: 10.1093/annonc/mdz268.096.
6. Duruisseaux M, Laskin JJ, Tolba K, et al. Targeting NRG1-fusions in lung adenocarcinoma: afatinib as a novel potential treatment strategy. Presented at: 2019 International Association for the Study of Lung Cancer World Conference on Lung Cancer; September 7-10, 2019; Barcelona, Spain. Abstract P1.14-25. doi: 10.1016/j.jto.2019.08.1176.
7. Clinical Pipeline. Merus website. bit.ly/305JTpv. Accessed January 6, 2020.
8. Bispecific antibody MCLA-128 shows clinical activity in patients with solid tumors harboring NRG1 gene fusions [news release]. Boston, MA: American Association for Cancer Research; October 27, 2019. bit.ly/2Zey93h. Accessed December 20, 2019.
9. Tarloxotinib. Rain Therapeutics, Inc, website. bit.ly/2PLoZZf. Accessed December 20, 2019.
10. Waters AM, Der CJ. KRAS: the critical driver and therapeutic target for pancreatic cancer.Cold Spring Harb Perspect Med. 2018;8(9):a031435. doi: 10.1101/cshperspect.a031435.
11. Jones MR, Williamson LM, Topham JT, et al. NRG1 gene fusions are recurrent, clinically actionable gene rearrangements in KRAS wild-type pancreatic ductal adenocarcinoma.Clin Cancer Res. 2019;25(15):4674-4681. doi: 10.1158/1078-0432.CCR-19-0191.