Mechanisms of resistance to crizotinib in patients with ALK gene rearranged non-small cell lung cancer

Abstract

Purpose: Patients with anaplastic lymphoma kinase (ALK) gene rearrangements often manifest dramatic responses to crizotinib, a small-molecule ALK inhibitor. Unfortunately, not every patient responds and acquired drug resistance inevitably develops in those who do respond. This study aimed to define molecular mechanisms of resistance to crizotinib in patients with ALK(+) non-small cell lung cancer (NSCLC).

Experimental design: We analyzed tissue obtained from 14 patients with ALK(+) NSCLC showing evidence of radiologic progression while on crizotinib to define mechanisms of intrinsic and acquired resistance to crizotinib.

Results: Eleven patients had material evaluable for molecular analysis. Four patients (36%) developed secondary mutations in the tyrosine kinase domain of ALK. A novel mutation in the ALK domain, encoding a G1269A amino acid substitution that confers resistance to crizotinib in vitro, was identified in two of these cases. Two patients, one with a resistance mutation, exhibited new onset ALK copy number gain (CNG). One patient showed outgrowth of epidermal growth factor receptor (EGFR) mutant NSCLC without evidence of a persistent ALK gene rearrangement. Two patients exhibited a KRAS mutation, one of which occurred without evidence of a persisting ALK gene rearrangement. One patient showed the emergence of an ALK gene fusion-negative tumor compared with the baseline sample but with no identifiable alternate driver. Two patients retained ALK positivity with no identifiable resistance mechanism.

Conclusions: Crizotinib resistance in ALK(+) NSCLC occurs through somatic kinase domain mutations, ALK gene fusion CNG, and emergence of separate oncogenic drivers.

Conflict of interest statement

Conflict of Interest Statement: RCD has research grants from Pfizer, Eli Lilly, and ImClone. RCD and MVG have received speaker’s fees from Abbott Molecular. RCD, MVG, DRC, KLK, AJW, and DLA have served as consultants for Pfizer. MVG has served as a consultant for Abbott Molecular. DRC has served as a consultant/advisory board member for Chugai, Ariad, and Eli Lilly. DLA has served as a consultant for GSK.

Figures

Figure 1. Identification and characterization of a novel ALK Kinase domain mutation in patients

Surface (A) and ribbon (B) models of the ALK kinase domain in complex with crizotinib (PDB 2XP2). Insets to the right represent a magnified view of the ATP- binding pocket where crizotinib is located. (A) and (B) were generated using the PyMol molecular graphics system. (C) Ba/F3 cells expressing wild-type EML4-ALK (E6;A20), or the same EML4-ALK construct with the specified ALK kinase domain mutations or empty vector were treated with the indicated concentration of crizotinib and viable cells were measured after 72 hours and then plotted relative to untreated controls. Ba/F3 cells expressing empty vector were grown in the presence of IL-3. (D) SDS-PAGE and immunoblotting analysis to detect the indicated proteins in cell lysates from NIH3T3 cells expressing wild-type EML4-ALK (E6;A20) and EML4-ALK (E6;A20) G1269A treated with the indicated doses of crizotinib for 5 hours. NIH3T3 with empty vector are also shown as a control. (E) Quantitation of western blot in (D) is graphically represented using LiCor image analysis software. (F) Ba/F3 cells expressing wild-type EML4-ALK (E6;A20), or the same EML4-ALK construct with the specified ALK kinase domain mutations or empty vector (with and without IL-3) were plated and viable cells were measured at the indicated time points and plotted.

Figure 2. ALK FISH pattern changes from pre- to post-crizotinib tumor samples

FISH analysis of patients #6 (A) and #7 (B) before crizotinib treatment (left) and following progression on crizotinib (right) demonstrating a gain of split green (5’) and red (3’) ALK signals per each tumor cell. FISH analysis of patients #8a (C) and 11 (D) before crizotinib treatment (left) and following progression on crizotinib (right) demonstrating loss of split green (5’) and red (3’) ALK signals.

Figure 3. Alternate activating oncogenes in patients with ALK+ NSCLC

(A) H3122, H2228, and CUTO-1 cells (from patient #10) were treated with the indicated concentration of crizotinib and viable cells were measured after 72 hours and then plotted relative to untreated controls. (B) H3122 cells with expression of KRAS G12V or empty vector were treated with the indicated concentration of crizotinib and viable cells were measured after 72 hours and then plotted relative to untreated controls.

Figure 4. Relative frequencies of crizotinib resistance mechanisms in ALK+ NSCLC patients and models for potential mechanisms of alternate oncogene acquisition

(A) The wedges represent different molecular mechanisms of resistance identified in ALK+ NSCLC patients in this study. The blue arc represents presumed or confirmed presence of an alternate oncogene. The yellow arc represents copy number gain (CNG). The red arc represents the presence of an ALK kinase domain mutation. The grey wedge represents those patients where an ALK gene rearrangement was observed, but no mechanism of resistance was identified. *Denotes inclusion of one patient with intrinsic resistance within this category. (B) Model #1 depicts the low level presence of a second oncogenic driver in the same cell as an ALK gene rearrangement, which following treatment with crizotinib becomes the dominant clone. Model #2 depicts the presence of separate clonal populations, some with an ALK gene rearrangement as the driver and others with an alternate oncogene driver (e.g., KRAS or EGFR). Following treatment with crizotinib, the non-ALK clones become the dominant clone.