Consecutive Day HSP90 Inhibitor Administration Improves Efficacy in Murine Models of KIT-Driven Malignancies and Canine Mast Cell Tumors

Abstract

Purpose: STA-1474, prodrug of the heat shock protein 90 inhibitor (HSP90i) ganetespib, previously demonstrated activity in canine preclinical models of cancer; interestingly, prolonged infusions were associated with improved biologic activity. The purpose of this study was to identify the ideal treatment schedule for HSP90i in preclinical models of KIT-driven malignancies and in dogs with spontaneous mast cell tumors (MCT), where KIT is a known driver.

Experimental design: In vitro and murine xenograft experiments and clinical studies in dogs with MCTs were used to define the effects of HSP90i-dosing regimen on client protein downregulation and antitumor activity.

Results: Continuous HSP90 inhibition led to durable destabilization of client proteins in vitro; however, transient exposure required >10× drug for comparable effects. In vivo, KIT was rapidly degraded following a single dose of HSP90i but returned to baseline levels within a day. HSP90 levels increased and stabilized 16 hours after HSP90i and were not elevated following a subsequent near-term exposure, providing a functional pool of chaperone to stabilize proteins and a means for greater therapeutic activity upon HSP90i reexposure. HSP90i administered on days 1 and 2 (D1/D2) demonstrated increased biologic activity compared with D1 treatment in KIT or EGFR-driven murine tumor models. In a trial of dogs with MCT, D1/D2 dosing of HSP90i was associated with sustained KIT downregulation, 50% objective response rate and 100% clinical benefit rate compared with D1 and D1/D4 schedules.

Conclusions: These data provide further evidence that prolonged HSP90i exposure improves biologic activity through sustained downregulation of client proteins.

Figure 1.. Effects of ganetespib on HSP90 client protein turnover in vitro.

H1975 cells (A) and Kasumi cells (B) were exposed to increasing concentrations of ganetespib, either continuously or for 1 hour, followed by drug washout. Lysates were prepared 24 hours later and immunoblotted with the indicated antibodies. H1975 (C) and Kasumi (D) cells were treated with vehicle or ganetespib (100 nM) for the indicated times. Lysates were immunoblotted with the indicated antibodies.

Figure 2.. Two consecutive-day dosing as a strategy to saturate baseline and ganetespib-induced HSP90 binding sites.

(A-B) Pharmacodynamic assessment of HSP90 client turnover. (A) Mice bearing H1975 xenografts were dosed once with vehicle or ganetespib (150 mg/kg), tumors were harvested at the indicated times post-dose, and lysates were immunoblotted with the indicated antibodies. (B) Mice bearing Kasumi xenografts were dosed once with vehicle or ganetespib (150 mg/kg), tumors were harvested at the indicated times post-dose, and lysates were immunoblotted with the indicated antibodies. (C) Comparison of HSP90 induction for Day 1 (D1) vs. D1/D2 dosing regimens in vitro. For D1 dosing, H1975 cells were treated with ganetespib (100 nM) for 1 hour followed by drug washout. For D1/D2 dosing, a second dose of ganetespib was added to cells 24 hours after D1 drug washout. Lysates were prepared at the indicated times post-dose and immunoblotted with the indicated antibodies. For quantification, HSP70 and HSP90 levels were normalized to GAPDH loading control. (D-E) HSP90 binding sites are rapidly occupied by ganetespib in a dose dependent manner in H1975 cells. (D) Cells were treated with increasing concentrations of ganetespib for 24 hours. (E) Binding kinetics were assessed by treating cells with ganetespib (100 nM) for the indicated times. The amount of ganetespib bound to HSP90 was quantified using the competitive HSP90 binding assay. Results plotted as percent of total HSP90 binding sites occupied by ganetespib. (F) Kinetics of ganetespib-induced HSP90 binding sites and dependence on protein synthesis. Cells were treated with ganetespib (100 nM) −/+ cycloheximide (1 μg/mL) for the indicated times, followed by HSP90 binding assay. The fold change in HSP90 binding sites was calculated from the total amount of drug (ganetespib + D3-ganetespib) detected after saturation of available HSP90 binding sites. (G) Cellular levels of ganetespib in D1 vs. D1/D2 dosing. Cells were treated with ganetespib (100 nM) using D1 and D1/D2 dosing schedules, and the total amount of ganetespib was measured 4 hours post-dose following size exclusion filtration. (H) Assessment of degree and durability of HSP90 binding site occupancy with single and two consecutive day dosing. Cells were treated with ganetespib (100 nM) using D1 and D1/D2 dosing schedules. HSP90 binding assays were performed at the indicated time post-dose. Results plotted as percent of total HSP90 binding sites occupied by ganetespib.

Figure 3.. Dosing BID or two consecutive day’s results in greater therapeutic activity compared to once weekly treatment.

(A) SCID mice bearing established human GIST882 tumors (n=8 mice/group) were administered two weekly cycles of vehicle or ganetespib, i.v., as indicated starting on Day 22. Data are expressed as mean tumor volumes ± SEM. Treatment/control values (T/C), as a measure of percent tumor growth inhibition or regression, were as follows: 48% for ganetespib dosed 1x/week; −2.2% for ganetespib dosed D1/D2; 0.4% for ganetespib dosed BID. All doses were well tolerated with the exception of a single animal in the ganetespib BID group whose body weight dropped to −21% on Day 31 and rapidly recovered. (B) SCID mice bearing established human H1975 tumors (n=4 mice/group) were administered two weekly cycles of vehicle or ganetespib, i.v., as indicated, starting on Day 14. Data are expressed as mean tumor volumes ± SEM. T/C values were as follows: −13% for ganetespib dosed 1x/week; −53% for ganetespib dosed D1/D2; −61% for ganetespib dosed BID. All doses were well tolerated during the course of the study. **, p=0.001; ns, not significant for statistics between once weekly and D1/D2 ganetespib groups.

Figure 4.. Administration schedule affects clinical activity and target modulation in canine MCT.

(A) Waterfall plot showing percent change in tumor size for each subject at the end of the study period. Partial responses (≥30% reduction) are illustrated by the black dotted line. Diagonal marks denote patients with PD at week 4 due to progression of non-target lesions. (B) Tumor biopsy samples at 0, 24 and 72 hours post STA-1474 treatment were analyzed by immunoprecipitation and western blotting using anti-KIT antibody as indicated. The blots corresponding to individual patient responses in the waterfall plot above are indicated. (C) Response to therapy following consecutive-day treatment (Cohort C).