Selective RET kinase inhibition for patients with RET-altered cancers

Abstract

Background: Alterations involving the RET kinase are implicated in the pathogenesis of lung, thyroid and other cancers. However, the clinical activity of multikinase inhibitors (MKIs) with anti-RET activity in RET-altered patients appears limited, calling into question the therapeutic potential of targeting RET. LOXO-292 is a selective RET inhibitor designed to inhibit diverse RET fusions, activating mutations and acquired resistance mutations.

Patients and methods: Potent anti-RET activity, high selectivity, and central nervous system coverage were confirmed preclinically using a variety of in vitro and in vivo RET-dependent tumor models. Due to clinical urgency, two patients with RET-altered, MKI-resistant cancers were treated with LOXO-292, utilizing rapid dose-titration guided by real-time pharmacokinetic assessments to achieve meaningful clinical exposures safely and rapidly.

Results: LOXO-292 demonstrated potent and selective anti-RET activity preclinically against human cancer cell lines harboring endogenous RET gene alterations; cells engineered to express a KIF5B-RET fusion protein -/+ the RET V804M gatekeeper resistance mutation or the common RET activating mutation M918T; and RET-altered human cancer cell line and patient-derived xenografts, including a patient-derived RET fusion-positive xenograft injected orthotopically into the brain. A patient with RET M918T-mutant medullary thyroid cancer metastatic to the liver and an acquired RET V804M gatekeeper resistance mutation, previously treated with six MKI regimens, experienced rapid reductions in tumor calcitonin, CEA and cell-free DNA, resolution of painful hepatomegaly and tumor-related diarrhea and a confirmed tumor response. A second patient with KIF5B-RET fusion-positive lung cancer, acquired resistance to alectinib and symptomatic brain metastases experienced a dramatic response in the brain, and her symptoms resolved.

Conclusions: These results provide proof-of-concept of the clinical actionability of RET alterations, and identify selective RET inhibition by LOXO-292 as a promising treatment in heavily pretreated, multikinase inhibitor-experienced patients with diverse RET-altered tumors.

Figure 1.

Pre-clinical characterization of RET inhibitor selectivity and antitumor activity. As shown in the left portion of (A), four RET-altered (filled black symbols) and 83 RET-wild-type (open red symbols) human cancer cell lines were treated with LOXO-292, cabozantinib or vandetanib for 72 h. Cell survival was determined by DAPI staining and cell counting. Half-maximal effective concentration (EC₅₀) values were normalized to the EC₅₀ value for LOXO-292-treated LC-2/ad (CCDC6-RET) non-small-cell lung cancer cells. For each cell line, ratios equal to 1 correspond to the same cytotoxicity as for LC-2/ad, ratios <1 correspond to greater cytotoxicity than for LC-2/ad and ratios >1 correspond to lower cytotoxicity than for LC-2/ad. As shown in the right portion of Panel A, HEK-293 cells harboring different mutant versions of RET (filled blue symbols) or other targets (open red symbols) were treated with each inhibitor, followed by determination of RET kinase or target activity in cell lysates as described in supplementary Appendix, available at Annals of Oncology online. Half-maximal inhibitory concentration (IC₅₀) values were normalized to the IC₅₀ value for LOXO-292-treated KIF5B-RET cells. For each target, ratios equal to 1 correspond to the same cellular potency as for KIF5B-RET cells, ratios less than 1 correspond to greater potency than for KIF5B-RET cells, and ratios greater than 1 correspond to lower potency than for KIF5B-RET cells. As shown in (B), immunodeficient mice xenografted with the indicated RET-altered tumor models were treated orally with vehicle, cabozantinib (40-60 mg/kg daily) or LOXO-292 (30 mg/kg twice daily) for 14 (NIH-3T3 allografts), 22 (cell line xenografts) or 28 (patient-derived xenografts/PDX) days. For each model, tumor volume from 8 to 10 animals at the end of the treatment period was normalized to tumor volume before the first treatment; 0% corresponds to no net effect of treatment on tumor size, <0% corresponds to net tumor regression with treatment, and >0% corresponds to net tumor growth despite treatment. As shown in (C), immunodeficient mice (10 per treatment) were injected intracranially with CCDC6-RET fusion-positive PDX tumor suspensions and treated orally with vehicle, ponatinib (40 mg/kg daily) or LOXO-292 (30 mg/kg twice daily). Animals were sacrificed if they displayed significant morbidity, and survival was compared with Kaplan–Meier analysis before and after dose reduction by 10-fold on day 52 (indicated with a black arrow). EC₅₀, half-maximal effective concentration; IC₅₀, half maximal inhibitory concentration; Cabo, cabozantinib; Vande, vandetanib; PDX, patient-derived xenograft; MTC, medullary thyroid cancer; NSCLC, non-small cell lung cancer; CRCA, colorectal cancer.

Figure 2.

Clinical activity of LOXO-292 in medullary thyroid cancer after acquired resistance to multikinase inhibitors. (A) The various treatments the patient received for metastatic RET M918T-mutated medullary thyroid cancer, together with the duration of and best response to each treatment. (B) Estimated RET target inhibition based on measured LOXO-292 plasma levels determined at the indicated times after 8 days of treatment with each dose of LOXO-292, compared with cabozantinib and vandetanib. The lower edge of each rectangle corresponds to the steady-state minimum concentration (C_min) of each inhibitor/target, the upper edge corresponds to the maximum concentration (C_max). The degree of target inhibition was modelled using actual patient (LOXO-292) or published (cabozantinib, vandetanib) human pharmacokinetic parameters (see supplementary Appendix, available at Annals of Oncology online). (C) Serial monitoring of (upper) carcinoembryonic antigen (CEA, open circles), calcitonin (filled squares), and (lower) plasma cell-free DNA allelic fraction of the founder M918T mutation and V804M, V804L and Y806C gatekeeper mutations during LOXO-292 treatment. (D) Computed tomographic images of the patient's metastatic liver disease before and at the indicated times after he initiated treatment with LOXO-292. A radiologic PR by RECIST was achieved after 4.2 months that deepened with treatment (best objective tumor regression 54% after 6.9 months of treatment), with concurrent resolution of hepatomegaly. PD, progressive disease; SD, stable disease; PR, partial response; Cabo, cabozantinib; Vande, vandetanib; mg, milligrams; QD, once daily; BID, twice daily; CEA, carcinoembryonic antigen; ng, nanograms; ml, milliliters; pg, picograms; Mo, month; PR, partial response; cPR, confirmed partial response. *The patient had an interruption in the dosing of LOXO-292 during ‘Hurricane Harvey’, the unprecedented 1000-year flood that affected Houston, Texas. There was a transient elevation in calcitonin and CEA that decreased when LOXO-292 was resumed. **Note infiltrative pattern of liver involvement by tumor.

Figure 3.

Clinical activity of LOXO-292 in KIF5B-RET fusion-positive lung cancer metastatic to the brain. (A) The various treatments the patient received for metastatic KIF5B-RET fusion-positive non-small-cell lung cancer together with the best response to each treatment. (B) Estimated target inhibition based on steady-state LOXO-292 C_min and C_max plasma levels determined at the indicated times after 8 days of treatment with the indicated doses of LOXO-292, compared with alectinib. The degree of target inhibition was modelled using actual patient (LOXO-292) or published (alectinib) human pharmacokinetic parameters (see supplementary Appendix, available at Annals of Oncology online). (C) T1, contrast-enhanced magnetic resonance imaging images of the patient’s metastatic brain disease before and at the indicated times after she initiated treatment with LOXO-292. A radiologic partial response was achieved extracranially and intracranially after 2 months of treatment that was sustained and deepened at 5 months of treatment (with tiny foci of nonspecific enhancement but no measurable target lesions remaining in the brain indicating a confirmed PR). See supplementary Figure S6, available at Annals of Oncology online for additional brain and extracranial imaging. PR, partial response; SD, stable disease; PD, progressive disease; WBRT, whole brain radiation; SRS, stereotactic radiosurgery; ag, agonist; Alect, alectinib; mg, milligrams; BID, twice daily; Mo, month; cPR, confirmed partial response.