Chromosomal rearrangements involving the gene that encodes the RET tyrosine kinase are known oncogenic drivers in 1% to 2% of patients with non–small cell lung cancer. These RET rearrangements occur with characteristic partners, most commonly KIF5B, but also CCDC6, NCOA, TRIM33, CUX1, KIAA1217, FRMD4A, and KIAA1468.
Christine M. Lovly, MD, PhD
Chromosomal rearrangements involving the gene that encodes the RET tyrosine kinase are known oncogenic drivers in 1% to 2% of patients with nonsmall cell lung cancer (NSCLC). These RET rearrangements occur with characteristic partners, most commonly KIF5B, but also CCDC6, NCOA, TRIM33, CUX1, KIAA1217, FRMD4A, and KIAA1468. They are typically identified in young patients with adenocarcinoma histology and minimal smoking history. Therapeutic targeting of RET-fusion–driven NSCLCs has taken the form of treatment with broad-spectrum tyrosine kinase inhibitors with anti-RET activity, such as cabozantinib (Cabometyx; Cometriq), vandetanib (Caprelsa), lenvatinib (Lenvima), RXDX-105, and sunitinib (Sutent). Cabozantinib and vandetanib have been the most heavily studied multi-kinase inhibitors (MKIs), with response rates of 20% to 50% in largely pretreated patients with RET-rearranged NSCLC. Sunitinib has been used in fewer patients to date with initial results demonstrating a 22% response rate. RXDX-105 has exhibited uniquely impressive response rates (75%) in patients with non–KIF5B-RET-fusion NSCLC, compared with 0% response in patients with KIF5B-RET-fusion–positive NSCLC. BLU-667 has demonstrated an objective response rate of 50% in patients with RET-fusion positive NSCLC, and LOXO-292 reported a 74% ORR in patients with RET-fusion positive NSCLC. Notably, RXDX-105, BLU- 667, and LOXO-292 have all demonstrated some central nervous system activity in these early phase trials. Future directions of RET inhibition in patients with RET-rearranged NSCLC include additional clinical validation of the next generation RET-selective inhibitors RXDX-105, BLU-667, and LOXO-292 and comparing multikinase inhibitors with RET-selective inhibitors to determine the optimal sequencing of RET-targeted therapies.
The RET gene is a receptor tyrosine kinase protooncogene that can acquire oncogenic activity through mutation or rearrangement.1-4RET is normally expressed on neurons, sympathetic and parasympathetic ganglia, testis germ cells, urogenital tract cells, adrenal medullary cells, and thyroid C cells.5-7RET ligands are members of the glial cell linederived neurotrophic factor family, and ligand binding results in RET autophosphorylation and activation of downstream cellular proliferation, cell migration, and differentiation pathways including RAS/MAPK/ERK, PI3K/AKT, and phospholipase C-gamma.8 While loss-of-function mutations in RET are associated with Hirschsprung disease, gain-offunction mutations are associated with a variety of human malignancies, including non–small cell lung cancer (NSCLC).9-11Gain-of-function point mutations in RET are associated with medullary thyroid carcinoma, 12 but in NSCLC, oncogenic changes in RET take the form of chromosomal rearrangements.11
The most common fusion partner for RET rearrangements in patients with NSCLC is KIF5B,although RET fusions with CCDC6, NCOA, TRIM33, CUX1, KIAA1217, FRMD4A, and KIAA1468 have been identified.11,13-17Importantly, oncogenic RET rearrangements in NSCLC result in constitutive activation of RET and consistently preserve the RET tyrosine kinase domain.18-21Since RET rearrangements typically do not co-occur with other well-established oncogenic mutations in NSCLC such as EGFR, KRAS, ALK, HER2, and BRAF, they are believed to harbor independent oncogenic driver potential.11,18,19,21-23
RET rearrangements are found in 1% to 2% of patients with NSCLC,11typically younger patients with adenocarcinoma histology and minimal smoking history.23Although there is no goldstandard method to identify RET rearrangements, rearranged RET has been detected in NSCLC tumor tissue using a variety of methods including immunohistochemistry (IHC),21,23fluorescence in-situ hybridization,18,20,23,24real-time polymerase chain reaction (RT-PCR),14,15,18-21,23-25and next-generation sequencing.19-21,25,26Importantly, despite the observation that RET is minimally expressed in normal pulmonary tissue,27 RET IHC has not yet proven to be an effective screening tool for detecting RET rearrangements in patients with NSCLC.21
Building on the observations that lung cancer cell lines harboring RET rearrangements are sensitive to multikinase tyrosine kinase inhibitors (MKIs) with anti-RET activity, such as sunitinib, sorafenib (Nexavar), and vendatinib, but not to MKIs without RET activity, such as gefitinib (Iressa) or crizotinib (Xalkori), a variety of TKIs with anti- RET activity have been applied in patients with RET-rearranged NSCLC.18,21
The 5 most heavily studied agents in patients with RET-rearranged NSCLC have been cabozantinib, vandetanib, lenvatinib, RXDX-105, and sunitinib. Less-specific MKIs with some anti-RET activity that have been used in patients with RET-rearranged NSCLC include sorafenib, alectinib (Alecensa), nintedanib (Ofev; Vargatef), ponatinib (Iclusig), and regorafenib (Stivarga).28More recently, early clinical results with RET-selective inhibitors RXDX-105,29 BLU-667,30 and LOXO-292,31have also been reported.
Based on promising antitumor activity observed in murine models of RET-rearranged lung cancer, cabozantinib was used in a phase II clinical trial of patients with RET-rearranged NSCLC.32
From 2012 to 2016, 26 patients with metastatic or unresectable RET-rearranged lung cancer were treated with cabozantinib at 60 mg daily. A quarter of the patients were treatment-naïve, half had received 1 line of chemotherapy, and the remaining quarter had received 2 or more lines of therapy prior to study enrollment. Most patients (16 of 26; 62%) had the KIF5B-RET fusion, and among the 25 patients evaluated for response, the objective response rate (ORR) was 28% (7 of 25). The responses occurred within 4 weeks of therapy initiation in 5 of the 7 (71%) patients who responded, and the median duration of response was 7 months (95% CI, 3.7-38.9 months). The median progression-free survival (PFS) was 5.5 months (95% CI, 3.8-8.4 months), and the median overall survival (OS) was 9.9 months (95% CI, 8.1-not reached). Toxicity was manageable, with no grade 4 or 5 adverse events (AEs), and the only grade 3 toxicity to occur in greater than 10% of patients was lipase elevation (Table).32
In order to systematically capture outcomes data for patients with RET-rearranged NSCLC treated with RET inhibitors outside the context of a clinical trial, the Global RET Registry (GLORY) was launched in 2015. This collective experience has provided additional data regarding response rates to cabozantinib, along with other RET inhibitors that we will discuss later, in patients with RET-rearranged NSCLC.28 The GLORY database includes data from 53 patients with relapsed stage III or stage IV NSCLC treated with RET inhibitors between June 2015 and April 2016 outside a clinical trial at 29 centers: 15 in Europe (51%), 3 in Asia (11%), and 11 in the United States (38%). In 21 patients from this cohort treated with cabozantinib, the ORR was 33% (7/21); median PFS was 3.6 months (95% CI, 1.3-7 months); and median OS was 4.9 months (95% CI, 1.9-14.3 months).28
Based on its in vitro anti-RET activity, including against a lung adenocarcinoma cell line harboring the CCDC6-RET fusion, vandetanib has been used in patients with RET-rearranged NSCLC.33-36
In a phase II multicenter clinical trial in South Korea that enrolled between 2013 and 2015, 18 patients with metastatic or recurrent RET-rearranged NSCLC were treated with vandetanib 300 mg daily.37All patients had progressed through at least 1 line of systemic therapy, and 72% had received 2 or more lines of systemic therapy. Among 17 patients evaluated, 3 had a partial response (PR; 18%) and an additional 8 had stable disease (overall disease control rate [DCR], 65%). All 3 patients with a PR had disease control for more than 6 months, and 4 of the 8 patients with stable disease (SD; 50%) had disease control for more than 6 months. The most common grade 3 AE was hypertension (18%), and there were no grade 4 or 5 AEs.37
An analogous phase II multicenter clinical trial in Japan was reported in 2017 (LURET).33Between 2013 and 2015, 19 patients with RET-rearranged NSCLC were treated with vandetanib at 300 mg daily. Approximately one-third (37%) of patients had progressed through 1 prior line of systemic therapy, and the remaining 63% had received 2 or more lines of systemic therapy. Ten patients (53%) had the KIF5B-RET fusion, and 6 (31%) had the CCDC6-RET fusion. Nine (47%) patients had a PR, and 7 patients (37%) had a greater than 50% reduction in tumor size. The median PFS was 4.7 months (95% CI, 2.8-8.5 months) and the median duration of response was 5.6 months (95% CI, 2.1-9.1 months). The most common grade 3 AE was hypertension (58%), and 3 grade 4 AEs occurred (bacterial pneumonia, prolonged QT, and rash).33
In 11 patients from the GLORY cohort treated with vandetanib, the ORR was 18% (2 of 11), median PFS was 2.9 months (95% CI, 1.0-6.4 months), and median OS was 10.2 months (95% CI, 2.4-not reached) (Table).28
Based on its in vitro anti-RET activity, lenvatinib has been applied in a multicenter, combined US and Japanese phase II clinical trial. In this trial, 25 patients with RET-rearranged NSCLC were treated with lenvatinib at 24 mg daily.38Of these 25 patients, 40% had progressed through 1 line of systemic therapy, while the remaining 60% had received 2 or more lines of systemic therapy. Importantly, 7 patients (28%) had received previous RET-targeted therapy. Four patients had a partial response (16%), and an additional 15 patients had stable disease, for an overall DCR of 76%. Lenvatinib proved highly toxic in these patients, with 92% experiencing a grade 3 or higher AE; the most common grade 3 or higher AE was hypertension (68%), and there were 3 fatal AEs, 1 of which was pneumonia attributed to lenvatinib (Table).38
Based on its in vitro anti-RET activity, 9 patients from the GLORY cohort were treated with sunitinib, and the ORR was 22% (2/9), median PFS was 2.2 months (95% CI, 0.7-5.0 months), and median OS was 6.8 months (95% CI, 1.1-not reached).28
Next Generation RET-selective Inhibitors
The previously discussed MKIs have biochemical activity against RET but were not designed to selectively target RET. RET-driven NSCLC are not optimally RET-selective. More recently, the results of early clinical trials with RET-selective inhibitors, RXDX-105,29BLU-667,30and LOXO-292,31have been reported.
As recently reported at the European Society of Medical Oncology 2017 Congress, RXDX-105 is a VEGFR-sparing RET inhibitor with activity against patient-derived xenograft tumor models harboring RET rearrangements39; it has been applied to patients with varying tumor histologies harboring RET rearrangements and RET mutations.29In a multicenter phase I/Ib clinical trial, an established phase II dose of 275 mg daily was determined. No patients had received previous RET inhibitor therapy. A total of 21 patients with RET-rearranged NSCLC were treated in the phase Ib expansion cohort, and among 8 patients with nonKIF5B-RET fusions, the ORR was 75% (6/8), including a central nervous system (CNS) response in 1 patient. However, among 13 patients harboring KIF5B-RET fusions, none had an objective response. The most common grade 3 AE was rash (10%), and no grade 4 or 5 AEs occurred.29 Experimental studies have not yet identified a unique basis for KIF5B-RET insensitivity to RET inhibition compared with non–KIF5B-RET sensitivity (Figure).
The results of a phase I, multi-histology basket trial treating patients with RET-driven malignancies (papillary thyroid cancer (PTC), NSCLC, and medullary thyroid cancer (MTC)) were orally presented at the American Association of Cancer Research Annual Meeting in 2018 and simultaneously published. BLU-667 was designed to maximize on-target and minimize off-target effects, and it was validated in both cell-line and patient-derived xenograft models.30
In the dose escalation cohort the maximum tolerated dose was determined to be 400 mg daily. The ORR for all 40 response evaluable patients was 45%, and of these 40 patients 53% (n=21) had received prior RET-directed MKI treatment. In patients with RET-mutated PTC the ORR was 40% (10/25), and in patients with RET-fusion NSCLC the ORR was 50% (7/14). While PFS and OS data are not mature, adverse events were reported, and the only grade >2 adverse event that occurred in more than 5% of patients was hypertension (>grade 2 in 8% of patients). Other grade 3 adverse events included neutropenia in 2 patients, leukopenia in 1 patient, ALT increase in 1 patient, fatigue in 1 patient, and diarrhea in 1 patient. There were no grade 4/5 adverse events. Importantly, CNS activity was reported in one patient with KIF5B-RET fusion NSCLC.
LOXO-292 is another RET-selective TKI that was strategically designed to maximize on-target activity while minimizing activity against other kinases. Results from a multi-histology phase I basket trial of this novel agent in patients with RET-driven malignancies was recently presented at the American Society of Clinical Oncology Annual Meeting in 2018.In the dose escalation cohort, patients with NSCLC, PTC, MTC, and RET-fusion positive pancreatic cancer (2 patients) were included. In this cohort, 66% (n=55) of patients had previously received a RET-directed MKI. The maximum tolerated dose was not reached, and the maximum dose administered in trial was 240 mg twice daily. The confirmed ORR in 34 evaluable patients was 74% (25/34), including a 74% (20/27) confirmed ORR in patients with RET-fusion NSCLC and 33% (6/18) confirmed ORR in patients with RET-mutant MTC. Responses were seen in both patients with KIF5B-RET fusions and non-KIF5BRET fusion NSCLC. There were 10 patients with CNS metastases, and 3 of 3 patients with measurable CNS disease had CNS responses, including one patient with CLIP1-RET fusion NSCLC. The only grade >2 adverse events reported were grade 3 dyspnea, grade 3 tumor lysis syndrome, and grade 3 increase ALT.
Central Nervous System Penetration
The central nervous system (CNS) penetration of oncogene- targeted therapies in patients with NSCLC is an active area of research, including in patients with RET-rearranged NSCLC. Preclinical models suggest that the CNS penetration of vandetanib is decreased by P-glycoprotein and breast cancer resistance protein 1-mediated efflux,40and mTOR inhibition with agents such as everolimus overrides this efflux. Therefore, combination therapy with vandetinib and everolimus has been assessed. A recent case report has shown CNS activity of combination therapy with vandetanib and everolimus in a patient with KIF5B-RETfusion NSCLC.41Although alectinib has not been as heavily studied in patients with RET-rearranged NSCLC as have other agents described above, it is important to note that it, too, has demonstrated CNS activity in patients with KIF5B-RETfusion NSCLC.39
Importantly, all 3 next generation RET inhibitors RXDX-105, BLU-667, and LOXO-292 have demonstrated CNS activity. These data are described above under the relevant sections for each drug. While expected rates of CNS activity have not been defined in early clinical data presented for these agents, this will be an important clinical activity metric to follow in subsequent, larger reports. Acquired Resistance Anticipating the inevitable acquired resistance that accompanies oncogene-targeted therapies, in vitro studies have begun to analyze RET inhibitor resistance mechanisms. Several potential mechanisms of resistance to RET inhibition involve EGFR signaling, as it has been shown that EGF ligand binding to EGFR can blunt the pharmacologic activity of RET inhibitors on their target fusion kinase; adaptor protein binding can be shifted from the target kinase protein to EGFR; and EGFR can act as a bypass pathway to downstream oncogenic signaling through mitogen-activated protein kinase.42Additionally, in lung adenocarcinoma cell lines that develop resistance to RET inhibitors, oncogenic mutations and overexpression of NRAS have been identified, as have upregulations of EGFR and AXL.43
The 2 most heavily studied RET inhibitors have been cabozantinib and vandetanib, and these agents have demonstrated response rates between 20% and 50% in a majority of patients who have received 2 or more lines of systemic therapy. Tolerability has been manageable with these agents, with lipase elevation the most common grade 3 toxicity observed with cabozantinib and hypertension the most common grade 3 AE seen with vandetanib. While lenvatinib has proven excessively toxic and sunitinib has not been broadly applied in patients with RET-rearranged NSCLC, RXDX-105 has shown notably higher response rates in patients with non-KIF5B-RETfusion NSCLC, and phase I clinical trials have demonstrated promise for the RET-selective inhibitors, BLU-667 and LOXO-292. Future directions for this exciting avenue of clinical research include further testing of the RET-selective inhibitors in larger patient cohorts, identifying the optimal sequence of RET inhibitors, and defining mechanisms of acquired resistance to both MKIs and RET-selective inhibitors.