Molecular Mechanisms of Resistance to First- and Second-Generation ALK Inhibitors in ALK-Rearranged Lung Cancer

Abstract

Advanced, anaplastic lymphoma kinase (ALK)-positive lung cancer is currently treated with the first-generation ALK inhibitor crizotinib followed by more potent, second-generation ALK inhibitors (e.g., ceritinib and alectinib) upon progression. Second-generation inhibitors are generally effective even in the absence of crizotinib-resistant ALK mutations, likely reflecting incomplete inhibition of ALK by crizotinib in many cases. Herein, we analyzed 103 repeat biopsies from ALK-positive patients progressing on various ALK inhibitors. We find that each ALK inhibitor is associated with a distinct spectrum of ALK resistance mutations and that the frequency of one mutation, ALK^G1202R, increases significantly after treatment with second-generation agents. To investigate strategies to overcome resistance to second-generation ALK inhibitors, we examine the activity of the third-generation ALK inhibitor lorlatinib in a series of ceritinib-resistant, patient-derived cell lines, and observe that the presence of ALK resistance mutations is highly predictive for sensitivity to lorlatinib, whereas those cell lines without ALK mutations are resistant.

Significance: Secondary ALK mutations are a common resistance mechanism to second-generation ALK inhibitors and predict for sensitivity to the third-generation ALK inhibitor lorlatinib. These findings highlight the importance of repeat biopsies and genotyping following disease progression on targeted therapies, particularly second-generation ALK inhibitors. Cancer Discov; 6(10); 1118-33. ©2016 AACRSee related commentary by Qiao and Lovly, p. 1084This article is highlighted in the In This Issue feature, p. 1069.

Figure 1

Overview of on-target mechanisms of resistance among ALK-positive specimens obtained from patients progressing on: A) crizotinib, B) ceritinib, and C) alectinib. Pie charts depict the frequency and distribution of ALK resistance mutations and ALK fusion gene amplification in each cohort. Four patients underwent two separate biopsies while on crizotinib; one patient underwent two separate biopsies while on ceritinib. Note: If a specimen is listed as having ≥2 ALK resistance mutations, the individual mutations are not listed separately. ^aOne post-crizotinib specimen harbored ALK G1269A and 1151Tins mutations. Four post-ceritinib samples contained ≥2 ALK resistance mutations. These included: I1171N+C1156Y, D1203N+F1174C, F1174L+G1202R, C1156Y+G1202del+V1180L mutations. ^bALK fluorescence in situ hybridization (FISH) to assess for fusion gene amplification was performed in only crizotinib-resistant specimens (N=36), of which 8% had amplification. Ceritinib- and alectinib-resistant specimens were not assessed for ALK amplification by FISH. ALK, anaplastic lymphoma kinase; WT, wild-type.

Figure 2

ALK resistance mutations are more common after treatment with second-generation ALK inhibitors compared to crizotinib. A) Comparison of the frequency and distribution of ALK resistance mutations in biopsy specimens obtained after disease progression on crizotinib (blue) or second-generation ALK inhibitors (red). Frequencies are expressed based upon the total numbers of biopsies in each cohort. B) Breakdown of specific ALK resistance mutations in ALK-positive patients progressing on crizotinib, ceritinib, alectinib or brigatinib. ^aFor patients with ≥2 ALK resistance mutations in a biopsy, each individual mutation is incorporated into the frequencies above. ^bEach specimen with ≥2 ALK resistance mutations is considered only once in determining the total number of specimens with ALK resistance mutations. WT, wild-type.

Figure 3

Clonal evolution of resistance to sequential ALK inhibitor therapy. A) Panel A depicts the treatment course of patient MGH086. Of note, the patient received several lines of therapy, including the EGFR inhibitor erlotinib, before identification of an ALK rearrangement. The points at which the patient underwent biopsies are indicated in red. B) Fused positron emission tomography (PET)-computed tomography (CT) images demonstrate a hypermetabolic, left axillary lymph node (white arrow) that developed at the time of disease progression on crizotinib. Whole-exome sequencing (WES) revealed an ALK E1210K mutation (cancer cell fraction [CCF] 0.82). This axillary lymph node initially responded to brigatinib but recurred after 12 months. WES of this brigatinib-resistant lesion (MGH086-0) demonstrated continued presence of the ALK E1210K mutation and a new ALK S1206C mutation. The patient remained on brigatinib. After an additional 9 months on brigatinib, he developed another recurrence in the left axilla (slightly more inferior than the prior lesion). Repeat WES (MGH086-1) revealed a new compound ALK E1210K+D1203N mutation. C) Panel C shows a model of clonal evolution of resistance to sequential ALK inhibitor therapy in patient MGH086. Using whole-exome sequencing, we determined that a founder ALK E1210K subclone was not present in a pre-crizotinib biopsy but later developed on crizotinib. When the patient was switched to brigatinib, the ALK E1210K subclone expanded and ultimately acquired a new ALK mutation, S1206C. Surgical excision of this site of progression may have depleted the compound mutant (ALK E1210K+S1206C), but microscopic parental E1210K clones may have persisted, ultimately acquiring ALK D1203N in combination with ALK E1210K.

Figure 4

Summary of genetic alterations in resistant biopsies among patients progressing on ceritinib or alectinib. Specimens underwent targeted, next generation sequencing (NGS) using the Massachusetts General Hospital (MGH) NGS assay or the FoundationOne platform. Only genes with at least one genetic alteration detected in resistant specimens are depicted. If a particular gene was not evaluated in a given specimen due to the type of sequencing platform used, it is represented in dark gray. Within this cohort, the most commonly mutated genes were ALK and TP53. Recurrent alterations in other genes were uncommon. In addition, six ceritinib-resistant, patient-derived cell lines underwent NGS using a 1000 gene panel (See Genotype Assessments in Methods). Genes from this panel are depicted in Figure 4 if the following criteria are met: (1) a genetic alteration was present in at least one of the patient-derived cell lines and (2) the gene was also included in either the MGH NGS or FoundationOne panels. Please see Table S11 for a comprehensive assessment of all genetic alterations identified within these cell lines.

Figure 5

Epithelial-mesenchymal transition (EMT) is associated with ceritinib resistance. A) Panel A depicts the clinical course of patient MGH067. B) Pre-crizotinib and post-ceritinib biopsies (lung and subcutaneous lesions, respectively) from MGH067 underwent hematoxylin and eosin staining, and immunostaining for E-cadherin, and vimentin. The post-ceritinib biopsy shows a loss of E-cadherin staining and gain of vimentin expression, consistent with EMT. Black arrows indicate a lack of vimentin staining of tumor cells in the pre-crizotinib biopsy. Red arrows depict vimentin staining of alveolar macrophages in the same specimen. C) Twelve ceritinib-resistant biopsy specimens underwent E-cadherin and vimentin staining to assess for EMT. In total, five specimens demonstrated immunohistochemical features consistent with EMT.

Figure 6

Lorlatinib potently inhibits ALK resistance mutations, including ALK G1202R. Absolute IC₅₀ values of crizotinib, ceritinib, alectinib, brigatinib, and lorlatinib on cellular ALK phosphorylation in Ba/F3 cells harboring wild-type EML4-ALK variant 1 or various EML4-ALK resistance mutants are depicted. ^aIn Ba/F3 cells, ALK F1174C and ALK I1171T appear sensitive to ceritinib and alectinib, respectively; however, these mutations may not be susceptible to these agents in vivo based upon prior clinical reports.

Figure 7

ALK resistance mutations predict for sensitivity to lorlatinib in patient-derived cell line models of acquired resistance to ceritinib. A–B) Cell viability assays of two representative ceritinib-resistant, patient-derived cell lines harboring ALK resistance mutations (MGH051-2C [EML4-ALK^G1202R] and MGH084-1D [EML4-ALK^{I1171N,C1156Y}]) treated with ceritinib, alectinib and lorlatinib. The number of cells seeded and the duration of treatment were adjusted for each cell line in order to have a consistent proliferation index (3.5 to 5) at the end of treatment. Values are presented as means (N=3). C–D) Cell viability assays of two representative, ceritinib-resistant, patient-derived cell lines without ALK resistance mutations (MGH049-1A [EML4-ALK^WT] and MGH075-2E [EML4-ALK^WT]) treated with ceritinib, alectinib and lorlatinib. The number of cells seeded and the duration of treatment were adjusted for each cell line in order to have a consistent proliferation index (3.5 to 5) at the end of treatment. Values are presented as means (N=3). E) Comparison of cell viabilities of ceritinib-resistant, patient-derived cell lines treated with lorlatinib based upon ALK resistance mutation status. F) Proposed schema for the clinical approach to ALK-positive patients with acquired resistance. This paradigm incorporates repeat biopsies and decision-making based upon ALK resistance mutation status following disease progression on second-generation ALK inhibitors. ^aChoice of second-generation ALK inhibitors may be impacted by identification of specific ALK resistance mutations, such as G1202R and I1171N/S/T, which can be rarely seen after crizotinib.