Targeting drug-resistant mutations in ALK

Abstract

Therapy resistance limits the clinical success of tyrosine kinase inhibitors (TKIs) in ALK-positive non-small cell lung cancer. A study now proposes a framework to identify compound resistance mutations to the lorlatinib TKI and provides structure-based drug design approaches to overcome resistance mediated by ALK(G1202R) or ALK(I1171N/S/T).

ATP-competitive kinase inhibitors have emerged as a highly successful class of targeted therapy, with more than 70 examples now used clinically¹. The remarkable efficacy of several of these drugs comes from the inhibition of ATP binding and the catalytic activity of driver kinases on which tumors are strongly dependent on or ‘addicted’ to². However, clinical responses can be of limited duration owing to mutations within the active-site pocket that can either sterically occlude drug binding or enhance catalytic efficiency of the driver kinase^3,4. Other types of mutation that further restrict enzyme flexibility, alter drug–receptor signaling, or disfavor specific conformational states of a kinase that are required for drug binding are also well documented¹.

The identification of several resistance mechanisms has led to various next-generation approaches, including the development of drugs intended to inhibit or anticipate a drug-resistant mutant kinase^5–7. However, possibly more challenging is the emergence of compound mutations — that is, two or more mutations within a single kinase oncoprotein⁸. The effects of compound mutations and how they can be addressed remains an open area of investigation. In this regard, the study by Shiba-Ishii et al.⁹ in this issue of Nature Cancer is notable in providing a framework in which to classify heterogeneous compound mutations and further identify strategies for how such mutations may be overcome using inhibitor analogs and structure-based drug design (Fig. 1a).

Fig. 1 |.

Mechanism of ALK compound mutations and inhibition by LA9 and LA7. a, Classification and functional characterization of mutations in EML4–ALK that drive resistance to active site TKIs, including lorlatinib, has led to the development of analogs LA9 and LA7 that inhibit G1202R- and I1171N-based compound mutations, respectively. b, 2D depiction of the macrocyclic inhibitor loratinib, and the more flexible acyclic analogs LA9 and LA7, highlighting crucial binding contacts in ALK including at the G1202R solvent-front position and the I1171N alteration at the base of helix αC and directly adjacent to the gatekeeper residue L1196.

Lung cancer is the leading cause of cancer-related death worldwide¹⁰. Non-small cell lung cancer (NSCLC), one of two major types of lung cancer, contributes to 80–85% cases. Up to 8% of patients with NSCLC possess rearrangements in the anaplastic lymphoma kinase (ALK) enzyme, mostly in younger patients or in non- or light-smokers¹¹. The most common ALK rearrangement in NSCLC is a fusion between ALK and the echinoderm microtubule-associated protein-like 4 (EML4). The resulting EML4–ALK fusion drives cell growth, proliferation and ultimately tumor formation owing to hyperactive enzymatic activity from the ALK kinase domain¹². Several ALK tyrosine kinase inhibitors (TKIs), including crizotinib (first generation), ceritinib, alectinib and brigatinib (second generation), have been approved to treat patients with ALK-positive NSCLC. However, resistance to first- and second-generation ALK-TKIs invariably occurs after treatment. Recently, lorlatinib, a third-generation ALK inhibitor and an ATP-competitive drug, has been approved by the US Food and Drug Administration (FDA) and demonstrated clinical efficacy against most of the acquired resistance caused by both the first- and second-generation ALK inhibitors¹³. Loralatinib is in a unique class of macrocyclic kinase inhibitors designed as a small and non-protruding drug within the ALK active site pocket (Fig. 1b).

However, despite initial responses, drug resistance to lorlatinib can eventually occur owing to acquired compound ALK resistance mutations⁸. Shiba-Ishii et al.⁹ start by identifying the most common EML4–ALK compound mutations resulting from treatment with lorlatinib. They find that most cases involve one of two key mutations, including either G1202R or I1171N in ALK. Notably, these two mutations were not found together, which suggests that they represent two distinct categories of alteration. In particular, the authors found that among 47 patients with lorlatinib-resistant tissue biopsies, almost half contained one or more ALK mutation⁹. Among 14 cases (29%) with double and triple ALK mutations, ALK(G1202R) (57%) and ALK(I1171N) (21%) were the most and second-most common ALK resistance mutations after treatment with second-generation ALK-TKIs, respectively. The investigators further confirmed this finding in an independent Guardant Health database including 194 plasma specimens from 167 patients with ALK-positive NSCLC.

To determine the activity of inhibitors, the authors generated Ba/F3 models containing distinct ALK compound mutations to screen FDA-approved first- and second-generation ALK-TKIs. All first- and second-generation ALK inhibitors demonstrated a marked decrease in potency against either ALK(G1202R)- or ALK(I1171N)-based mutations. By contrast, those containing ALK(L1198F) remained sensitive to crizotinib. Notably, crizotinib also demonstrated strong potency against the ALK(I1171N/L1198F) double mutation, which suggests that patients with this set of compound mutations may regain sensitivity and therefore benefit from treatment with a first-generation ALK inhibitor.

Next, the investigators sought to find new analogs of lorlatinib that may retain binding on several of the identified compound mutations. Shiba-Ishii et al.⁹ used a set of 20 lorlatinib analogs including both strained macrocycles and more flexible acyclic analogs. The lorlatinib analogs were further chosen based on cellular potency and chemical modifications that were predicted to reduce potential steric clashes with the common solvent-front G1202R alteration. Although modest in size, this focused library of analogs provided key structure–activity relationship data in the context of specific compound mutations and several possible leads to consider for future clinical development.

By applying a three-step functional screening of the 20 lorlatinib analogs, including the screening for compounds with high potency against single ALK mutations and for compounds with non-ALK specific cell toxicity, distinct patterns of drug efficacy were identified, and the two candidates LA7 and LA9 were selected for further validation. Compared with lorlatinib, the more flexible LA7 and LA9 showed more than 10-fold increases in potency against ALK(I1171N)- or ALK(G1202R)-based compound mutations. Furthermore, LA7 and LA9 exhibited a distinct pattern of differential activity against compound ALK mutations. LA7 was more potent against ALK(I1171N)-containing compound mutations, whereas LA9 was the most potent inhibitor against ALK(G1202R)-containing compound mutations. Moreover, the authors compared the inhibition efficacy of LA7 and LA9 in patient-derived cell lines containing ALK compound mutations. Consistent with the results in Ba/F3 models, LA7 exhibited better inhibition against ALK compound mutations containing both I1171N and D1203N, and LA9 was most potent against the ALK(G1202R/L1196M) compound mutation. Western blot analysis demonstrated that LA7 and LA9 dose-dependently inhibited the phosphorylation of ALK and downstream signaling pathways including pAKT, pERK and pS6, which lorlatinib failed to inhibit in both Ba/F3 and patient-derived models. Furthermore, LA7 (20 mg kg⁻¹) and LA9 (40 mg kg⁻¹) induced the regression of patient-derived xenograft tumors expressing ALK compound mutations, had a favorable pharmacokinetic profile and were well tolerated with no significant changes in body weight. Collectively, LA7 and LA9 exhibited distinct selectivity profiles against ALK(I1171N)- and ALK(G1202R)-containing compound mutations, respectively.

To investigate the diverse mechanism of action of LA7 and LA9 in compound ALK mutations, Shiba-Ishii et al.⁹ solved several co-crystal structures of their inhibitors bound to wild-type ALK and also generated energy-minimized models for the mutant variants. Under the substitution of G1202 with arginine, modeling results revealed that LA9 can maintain identical binding without significant steric clash, but the two hydroxyl groups on the thiazole ring of LA7 introduce a significant steric clash upon G1202R substitution. Unlike the solvent-front mutation G1202R, I1171N is located within the ALK hydrophobic core at the base of helix αC and leads to resistance by increasing the catalytic activity of ALK¹⁴. In particular, LA7 exhibited better potency than LA9 and lorlatinib against the ALK(I1171N) mutant as a consequence of strong binding between the hydroxyl groups of the ligand thiazole ring and side-chain carboxylate of D1203, along with a hydrogen-bond network including the D1203 amino nitrogen, the G1201 backbone carbonyl and a structural water.

Despite a wide range of distinct ALK inhibitors that may be used serially, clinical relapse remains a challenge owing to the occurrence of new ALK-resistant mutations or the activation of bypass signaling pathways such as EGFR, SRC and c-KIT¹⁵. Shiba-Ishii et al.⁹ show that secondary compound mutations reduce the affinity between lorlatinib and ALK sufficiently to levels that eliminate inhibition, even at the highest drug concentrations that can be achieved in vivo, thus necessitating the design of next-generation inhibitors. Compounds LA7 and LA9 identified by the authors⁹ have potency against post-lorlatinib ALK compound mutations. Further studies will be needed to investigate whether long-term treatment, earlier therapy induction, or possibly combination therapies of these compounds will succumb to similar resistance mechanisms. In addition, a detailed investigation of treatment schedules and the timing of recurrence during treatment will be crucial. Overall, Shiba-Ishii et al.⁹ describe two distinct categories of heterogenous compound mutations that lead to resistance against lorlatinib. They further identified unique chemical solutions for each resistance category, which was achieved primarily by mechanisms of improved targeting of either G1202R- or I1171N/S/T-mutant kinases. Rational and systematic approaches such as those described by the authors for developing next-generation ALK inhibitors based on distinct resistance features will enhance the tailoring of treatment for individual patients.