A phase I study of the heat shock protein 90 inhibitor alvespimycin (17-DMAG) given intravenously to patients with advanced solid tumors

Abstract

Purpose: A phase I study to define toxicity and recommend a phase II dose of the HSP90 inhibitor alvespimycin (17-DMAG; 17-dimethylaminoethylamino-17-demethoxygeldanamycin). Secondary endpoints included evaluation of pharmacokinetic profile, tumor response, and definition of a biologically effective dose (BED).

Patients and methods: Patients with advanced solid cancers were treated with weekly, intravenous (i.v.) 17-DMAG. An accelerated titration dose escalation design was used. The maximum tolerated dose (MTD) was the highest dose at which ≤ 1/6 patients experienced dose limiting toxicity (DLT). Dose de-escalation from the MTD was planned with mandatory, sequential tumor biopsies to determine a BED. Pharmacokinetic and pharmacodynamic assays were validated prior to patient accrual.

Results: Twenty-five patients received 17-DMAG (range 2.5-106 mg/m(2)). At 106 mg/m(2) of 17-DMAG 2/4 patients experienced DLT, including one treatment-related death. No DLT occurred at 80 mg/m(2). Common adverse events were gastrointestinal, liver function changes, and ocular. Area under the curve and mean peak concentration increased proportionally with 17-DMAG doses 80 mg/m(2) or less. In peripheral blood mononuclear cells significant (P < 0.05) HSP72 induction was detected (≥ 20 mg/m(2)) and sustained for 96 hours (≥ 40 mg/m(2)). Plasma HSP72 levels were greatest in the two patients who experienced DLT. At 80 mg/m(2) client protein (CDK4, LCK) depletion was detected and tumor samples from 3 of 5 patients confirmed HSP90 inhibition. Clinical activity included complete response (castration refractory prostate cancer, CRPC 124 weeks), partial response (melanoma, 159 weeks), and stable disease (chondrosarcoma, CRPC, and renal cancer for 28, 59, and 76 weeks, respectively).

Conclusions: The recommended phase II dose of 17-DMAG is 80 mg/m(2) weekly i.v.

Figure 1

The pharmacokinetics of 17-DMAG given over 1 hour IV, weekly, to patients with advanced solid tumors. A). Time-courses of concentration against time for individual patients treated with 80mg/m² (left panel) or 106mg/m² (right panel) of 17-DMAG. The mean IC₅₀ in the NCI human tumor cell line panel is marked with a dotted line on both graphs. B). Graphs with Cmax or AUC plotted against 17-DMAG dose, linear regression value (r²) shown above each plot. For AUC two graphs are: centre, all dose levels (r² = 0.66) with two patients who experienced DLT marked (arrows); and left, data from 106mg/m² have been excluded (r² = 0.88).

Figure 2

Pharmacodynamic changes in inducible HSP70 (HSP72) observed in PBMC and plasma. A). For each dose level of 17-DMAG, a western blot from a representative patient is shown. Samples were taken pre-dose (0), end of infusion (E), then 1, 8, 24, 48 and 96 hours after 17-DMAG; an additional sample was taken 24 hours after the fifth weekly infusion of 17-DMAG (#2). An HT29 human adenocarcinoma colon tumor sample is included as a positive control (+) and equal protein loading is confirmed by corresponding GAPDH expression. HSP72 levels were also detected in duplicate samples taken before and 24 hours after 17-DMAG dose. B). Histogram plotting HSP72 expression changes in PBMC as measured by ELISA. Values are the difference between pre- and post treatment samples expressed as mean ± SD. Values marked (*) were significantly, p < 0.05, different to the mean HSP72 expression change in the first dose level. C). Histogram showing HSP72 expression change in plasma samples. Values are change in HSP72 24 hours after 17-DMAG, expressed as mean ± SD. There was no statistically different (p > 0.05) change in mean HSP72 level compared to that of the first dose cohort.

Figure 3

Pharmacodynamic changes following 17-DMAG administration in tumor samples. A). Expression of ERBB2, inducible HSP70 (HSP72) and CDK4 detected by western blots from patients treated with 80 or 106mg/m² of 17-DMAG. GAPDH is included as a loading control. Samples are pre dose (0), 24 hours after 17-DMAG (24) or HT29 human colon adenocarcinoma positive control (+). B). Table summarizing pharmacodynamic changes in tumor. For each patient samples are marked positive (●) or negative (○) for induction of HSP70 and/ or depletion of a client protein (CDK4 or ERBB2). If both changes were detected then the sample was positive for detecting the molecular signature of HSP90 inhibition. * One sample set did not pass quality control.

Figure 4

Selected patient case histories. A). Graph of PSA changes in a patient with prostate adenocarcinoma treated with 17-DMAG. Time-points marked are: initial diagnosis and commencement of bicalutamide (1), radical radiotherapy (2), LHRH antagonist (3), bicalutamide withdrawal (4) and starting 17-DMAG (5), PSA and CT confirmed CR (6) and progression of disease (7). B). CT scans from a patient with metastatic melanoma at commencement of study and 30 months after starting 17-DMAG. Prior therapy had been adjuvant interferon and combination chemotherapy with dacarbazine and sorafenib. Patient received 159 weeks of 17-DMAG prior to PD.