A novel ALK secondary mutation and EGFR signaling cause resistance to ALK kinase inhibitors

Abstract

Anaplastic lymphoma kinase (ALK) tyrosine kinase inhibitors (TKI), including crizotinib, are effective treatments in preclinical models and in cancer patients with ALK-translocated cancers. However, their efficacy will ultimately be limited by the development of acquired drug resistance. Here we report two mechanisms of ALK TKI resistance identified from a crizotinib-treated non-small cell lung cancer (NSCLC) patient and in a cell line generated from the resistant tumor (DFCI076) as well as from studying a resistant version of the ALK TKI (TAE684)-sensitive H3122 cell line. The crizotinib-resistant DFCI076 cell line harbored a unique L1152R ALK secondary mutation and was also resistant to the structurally unrelated ALK TKI TAE684. Although the DFCI076 cell line was still partially dependent on ALK for survival, it also contained concurrent coactivation of epidermal growth factor receptor (EGFR) signaling. In contrast, the TAE684-resistant (TR3) H3122 cell line did not contain an ALK secondary mutation but instead harbored coactivation of EGFR signaling. Dual inhibition of both ALK and EGFR was the most effective therapeutic strategy for the DFCI076 and H3122 TR3 cell lines. We further identified a subset (3/50; 6%) of treatment naive NSCLC patients with ALK rearrangements that also had concurrent EGFR activating mutations. Our studies identify resistance mechanisms to ALK TKIs mediated by both ALK and by a bypass signaling pathway mediated by EGFR. These mechanisms can occur independently, or in the same cancer, suggesting that the combination of both ALK and EGFR inhibitors may represent an effective therapy for these subsets of NSCLC patients.

Figure 1. The ALK mutation L1152R causes the ALK tyrosine kinase inhibitor resistance

A. Sequence tracing from post-treatment tumor specimens. There is a T to G mutation in codon 3455 in exon 23 resulting in the L1152R mutation. B. Ba/F3 cells were treated with crizotinib at the indicated concentrations, and viable cells were measured after 72 hours of treatment and plotted relative to untreated controls. There is a significant effect of the L1152R mutation at 300 nM (p < 0.001). C. Ba/F3 cells with indicated genotypes were treated with increasing concentrations of crizotinib for 6 hours. Cell extracts were immunoprecipitated with an anti-FLAG antibody followed immunoblotting to detect the indicated proteins. D. Ribbon diagram depicting the crystal structure of Alk kinase in complex with crizotinib. The sidechains of residues that are sites of resistance mutations, including the L1152R mutation described here, are shown in green. Note that L1152 and C1156 are not in contact with the ATP-binding cleft, but are adjacent to each other and to the C-helix (pink). Both the L1152R and C1156Y mutations could introduce hydrogen bond interactions with E1161 on the C-helix. All of the resistance mutations identified to date cluster around the conformationally sensitive C-helix and activation loop, suggesting that they may affect kinase activity and inhibitor binding through alterations in the structure or stability of these elements. The activation loop (A-loop, colored orange) is partially disordered in this structure as denoted by the dashed line. Figure is drawn from PDB ID 2XP2.

Figure 2. DFCI076 cells are co-dependent on ALK and EGFR

A. The H3122 or DFCI 076 cells were treated with crizotinib, alone or concurrently with 1 μM PF299804, at the indicated concentrations, and viable cells were measured after 72 hours of treatment and plotted relative to untreated controls. B. Downregulation of ALK or EGFR using shRNAs in H3122 and DFCI076 cells. Further growth inhibition is observed with the addition of 1 μM PF2990804. Cell viability is measured relative to the non-targeting (NT) control shRNA. C. A phospho-RTK array reveals that the DFCI076 cells contain strong phosphorylation of both EGFR and MET. Cells were grown in medium and the cell lysates were hybridized to a phospho-RTK array. Hybridization signals at the corners serve as controls. D. H3122 and DFCI076 cells are treated with indicated concentrations of crizotinib for 6 hours. Immunoblotting was used to detect the indicated proteins.

Figure 3. H3122 TR3 cells contain are ALK inhibitor resistant and contain co-activation of EGFR signaling

A. H3122 and TR3 cells were treated with indicated concentrations of TAE684. Viable cells were measured after 72 hours of treatment and plotted relative to untreated controls. B. ALK shRNA has minimal effect on H3122 TR3 cell viability. Cell viability is measured relative to non-targeting (NT) control. H3122 and A549 cells serve as positive and negative controls, respectively. Successful ALK knockdown in H3122 TR3 cells is confirmed by Western blotting. C. Summary of RTK array from H3122 and H3122 TR3 cells with and without TAE684 (100 nM; 6 hrs) treatment. The protein lysates were exposed to the RTK array (R&D). Each spot of membrane were calculated as signal intensity and shown in the bar graph. D. ALK and EGFR phosphorylation in H3122 and H3122 TR3 cells with and with TAE684 and gefitinib treatment. Phosphorylation was measured using a bead based assay (methods). E. Proliferation of H3122 TR3 but not H3122 cells is effected following EGFR knockdown using two different shRNAs. NT; non-targeting. F. Results of colony formation assay after 14 days treatment with indicated compounds with both H3122 and H3122TR3. The combination of crizotinib (1 μM) and PF299804 (1 μM) effectively inhibited colony formation in H3122 TR3 cells. G. H3122 and H3122TR3 cells were treated with indicated compounds for 6 hours, and immunoblotting was used to detect the indicated proteins. For the PARP blot, cells were treated for 24 hours.

Figure 3. H3122 TR3 cells contain are ALK inhibitor resistant and contain co-activation of EGFR signaling

A. H3122 and TR3 cells were treated with indicated concentrations of TAE684. Viable cells were measured after 72 hours of treatment and plotted relative to untreated controls. B. ALK shRNA has minimal effect on H3122 TR3 cell viability. Cell viability is measured relative to non-targeting (NT) control. H3122 and A549 cells serve as positive and negative controls, respectively. Successful ALK knockdown in H3122 TR3 cells is confirmed by Western blotting. C. Summary of RTK array from H3122 and H3122 TR3 cells with and without TAE684 (100 nM; 6 hrs) treatment. The protein lysates were exposed to the RTK array (R&D). Each spot of membrane were calculated as signal intensity and shown in the bar graph. D. ALK and EGFR phosphorylation in H3122 and H3122 TR3 cells with and with TAE684 and gefitinib treatment. Phosphorylation was measured using a bead based assay (methods). E. Proliferation of H3122 TR3 but not H3122 cells is effected following EGFR knockdown using two different shRNAs. NT; non-targeting. F. Results of colony formation assay after 14 days treatment with indicated compounds with both H3122 and H3122TR3. The combination of crizotinib (1 μM) and PF299804 (1 μM) effectively inhibited colony formation in H3122 TR3 cells. G. H3122 and H3122TR3 cells were treated with indicated compounds for 6 hours, and immunoblotting was used to detect the indicated proteins. For the PARP blot, cells were treated for 24 hours.

Figure 4. Concurrent ALK and EGFR signaling leads to drug resistance

A. H3122 cells were treated with crizotinib at the indicated concentrations in the presence or absence of EGF (10 ng/ml) and PF299804 (1 μM). Viable cells were measured after 72 hours of treatment and plotted relative to untreated controls. B. H3122 cells were treated with crizotinib for 6 hrs at indicated concentrations in the absence or presence of EGF (10 ng/ml). Immunoblotting was used to detect the indicated proteins. C. H3122 cells stably expressing GFP or EGFR E746_A750 were treated with crizotinib in the absence or presence of gefitinib (1 μM). Viable cells were measured after 72 hours of treatment and plotted relative to untreated controls. D. Results of colony formation assay after 14 days treatment with indicated compounds using H3122 and H3122 EGFR E746_A750. The combination of crizotinib and PF299804 effectively inhibited colony formation. E. EGFR sequence tracing (left) and ALK FISH analysis (right) from patient 1. Asterix, T to G mutation resulting in L858R; arrows, split red and green FISH signals. F. Immunohistochemistry for EGFR L858R (left) and ALK (right) from patient 1.

Figure 4. Concurrent ALK and EGFR signaling leads to drug resistance

A. H3122 cells were treated with crizotinib at the indicated concentrations in the presence or absence of EGF (10 ng/ml) and PF299804 (1 μM). Viable cells were measured after 72 hours of treatment and plotted relative to untreated controls. B. H3122 cells were treated with crizotinib for 6 hrs at indicated concentrations in the absence or presence of EGF (10 ng/ml). Immunoblotting was used to detect the indicated proteins. C. H3122 cells stably expressing GFP or EGFR E746_A750 were treated with crizotinib in the absence or presence of gefitinib (1 μM). Viable cells were measured after 72 hours of treatment and plotted relative to untreated controls. D. Results of colony formation assay after 14 days treatment with indicated compounds using H3122 and H3122 EGFR E746_A750. The combination of crizotinib and PF299804 effectively inhibited colony formation. E. EGFR sequence tracing (left) and ALK FISH analysis (right) from patient 1. Asterix, T to G mutation resulting in L858R; arrows, split red and green FISH signals. F. Immunohistochemistry for EGFR L858R (left) and ALK (right) from patient 1.