Targeted inhibition of the molecular chaperone Hsp90 overcomes ALK inhibitor resistance in non-small cell lung cancer

Abstract

EML4-ALK gene rearrangements define a unique subset of patients with non-small cell lung carcinoma (NSCLC), and the clinical success of the anaplastic lymphoma kinase (ALK) inhibitor crizotinib in this population has become a paradigm for molecularly targeted therapy. Here, we show that the Hsp90 inhibitor ganetespib induced loss of EML4-ALK expression and depletion of multiple oncogenic signaling proteins in ALK-driven NSCLC cells, leading to greater in vitro potency, superior antitumor efficacy, and prolonged animal survival compared with results obtained with crizotinib. In addition, combinatorial benefit was seen when ganetespib was used with other targeted ALK agents both in vitro and in vivo. Importantly, ganetespib overcame multiple forms of crizotinib resistance, including secondary ALK mutations, consistent with activity seen in a patient with crizotinib-resistant NSCLC. Cancer cells driven by ALK amplification and oncogenic rearrangements of ROS1 and RET kinase genes were also sensitive to ganetespib exposure. Taken together, these results highlight the therapeutic potential of ganetespib for ALK-driven NSCLC.

Significance: In addition to direct kinase inhibition, pharmacologic blockade of the molecular chaperone Hsp90 is emerging as a promising approach for treating tumors driven by oncogenic rearrangements of ALK. The bioactivity profi le of ganetespib presented here underscores a new therapeutic opportunity to target ALK and overcome multiple mechanisms of resistance in patients with ALK-positive NSCLC.

Figure 1

Cellular viability, EML4-ALK expression and pathway modulation in ALK-rearranged NSCLC cell lines following ganetespib and crizotinib treatment. A, H2228 cells were treated with increasing concentrations of ganetespib or crizotinib and cell viability assessed at 72 h. B, H2228 cells were exposed to graded concentrations of ganetespib (3.3-100 nM) for 24 h and cell lysates immunoblotted with the indicated antibodies. v represents vehicle control. C, H3122 cells were treated with increasing concentrations of ganetespib or crizotinib and cell viability assessed at 72 h. D, H3122 cells were exposed to graded concentrations of ganetespib (3-500 nM) or crizotinib (10-500 nM) for 24 h and cell lysates immunoblotted with the indicated antibodies.

Figure 2

Ganetespib suppresses tumor growth and extends survival in ALK+ NSCLC xenografts. A, SCID mice bearing H3122 xenografts (n=7/group) were i.v. dosed with 50 mg/kg ganetespib once weekly, or on a weekly 2×25 mg/kg consecutive day dosing regimen, as indicated, for 3 weeks. % T/C values are indicated to the right of each growth curve and the error bars are the SEM. B, SCID mice bearing H3122 xenografts (n=7/group) were p.o. dosed with 50 mg/kg crizotinib 5X/week over a 3 week cycle. C, Representative images of tumors from vehicle, ganetespib (2X/week) or crizotinib (50 mg/kg) treated animals at day 50. D, Pharmacodynamic analysis of client protein modulation in H3122 xenografts. SCID mice bearing established H3122 tumors were treated with vehicle or a single bolus injection of ganetespib at 50 mg/kg at the indicated time points between 24 and 96 h. Mice were also treated a single bolus injection of crizotinib (50 mg/kg) for 24 h. Tumors were resected and the levels of the indicated proteins determined by immunoblotting. E, Kaplan-Meier analysis of overall survival in the H3122 xenograft model. Beginning 12 days after tumor cell implantation SCID mice bearing H3122 xenografts (n=7/group) were dosed with vehicle, 50 mg/kg ganetespib 1x/week i.v, or 50 mg/kg crizotinib 5X/week p.o and animal survival monitored until day 75.

Figure 3

Enhanced activity of ganetespib in combination with ALK inhibitors in vitro and in vivo. A, H3122 cells were treated with the indicated concentrations of ganetespib, crizotinib, ASP3026, CH5424802 either as single agents or in combination. Cell viability was determined at 72 h. B, Combination of ganetespib and crizotinib induces enhanced antitumor efficacy in vivo. SCID mice bearing H3122 xenografts (n=7/group) were i.v. dosed with 25 mg/kg ganetespib once weekly, 100 mg/kg crizotinib 5x/week p.o., or the combination, as indicated. % T/C values are indicated to the right of each growth curve and the error bars are the SEM.

Figure 4

Ganetespib retains potency against crizotinib-resistant NSCLC tumor phenotypes. A, Parental H3122 and crizotinib-resistant H3122 CR1 cells were treated with ganetespib at either 25 or 100 nM for 24 h and the levels of EML4-ALK protein determined by immunoblotting. B, H3122 and H3122 CR1 cells were treated with increasing concentrations of ganetespib or crizotinib and cell viability was assessed after 72 h. C, Table of IC₅₀ cytotoxicity values in H3122 and H3122 CR1 cells in response to ganetespib or ALK inhibitor exposure. D, Light micrographs of cellular morphology of H3122 and H3122 CR1 cells. Scale bar, 50 μM.

Figure 5

Crizotinib-resistant NPM-ALK mutants retain sensitivity to ganetespib. A, Ganetespib and crizotinib sensitivity was assessed in crizotinib-sensitive NPM-ALK BaF3 cells and crizotinib-resistant NMP-ALK mutants. Relative IC₅₀ values are plotted based on the sensitivity of the parental NPM-ALK BaF3 line to each compound. Clinically relevant mutations in NSCLC are indicated by asterisks. B, Dose-response analysis of ganetespib. Crizotinib-sensitive NPM ALK/BaF3 cells (upper left panel) together with NPM-ALK/BaF3 containing the indicated amino acid substitutions known to confer crizotinib resistance were incubated for 24 h with the range of ganetespib concentrations indicated, and the stability of the NPM-ALK protein expressed in each line assessed by immunoblotting. C, Kinetics of ganetespib-associated NPM-ALK protein degradation. The NPM-ALK protein degradation response of crizotinib-sensitive NPM-ALK/BaF3 and NPM-ALK/BaF3 containing the L1196M gatekeeper mutation to incubation in 250 nM ganetespib for the indicated times is shown.

Figure 6

Response of ALK-rearranged, crizotinib-resistant NSCLC after one cycle of ganetespib. CT scans taken at A, baseline and B, after one cycle (200 mg/m², 1X/week for 3 weeks) of ganetespib. Arrows depict locations of tumor masses.

Figure 7

Ganetespib displays potent cellular activity in models of ALK amplification, ROS1 kinase fusion and RET kinase fusion proteins. A, NB-39-nu neuroblastoma cells were treated with increasing concentrations of ganetespib or crizotinib and cell viability assessed at 72 h. B, NB-39-nu cells were exposed to graded concentrations of ganetespib (1-500 nM) for 24 h and cell lysates immunoblotted with the indicated antibodies. v represents vehicle control. C, HCC78 NSCLC and U118-MG glioma cells were treated with increasing concentrations of ganetespib and cell viability assessed at 72 h. D, HCC78 cells were treated with ganetespib at 5, 10, 50 or 100 nM for 24 h and the levels of phosphorylated and total SLC34A2-ROS1 fusion protein determined by immunoblotting. E, TPC1 thyroid carcinoma cells were treated with increasing concentrations of ganetespib and cell viability assessed at 72 h. F, TPC1 cells were treated with increasing concentrations of ganetespib as indicated for 24 h and the levels of phosphorylated and total CCDC6-RET fusion protein, phosphorylated ERK and cleaved PARP expression determined by immunoblotting.