HSP90 empowers evolution of resistance to hormonal therapy in human breast cancer models

Abstract

The efficacy of hormonal therapies for advanced estrogen receptor-positive breast cancers is limited by the nearly inevitable development of acquired resistance. Efforts to block the emergence of resistance have met with limited success, largely because the mechanisms underlying it are so varied and complex. Here, we investigate a new strategy aimed at the very processes by which cancers evolve resistance. From yeast to vertebrates, heat shock protein 90 (HSP90) plays a unique role among molecular chaperones by promoting the evolution of heritable new traits. It does so by regulating the folding of a diverse portfolio of metastable client proteins, many of which mediate adaptive responses that allow organisms to adapt and thrive in the face of diverse challenges, including those posed by drugs. Guided by our previous work in pathogenic fungi, in which very modest HSP90 inhibition impairs resistance to mechanistically diverse antifungals, we examined the effect of similarly modest HSP90 inhibition on the emergence of resistance to antiestrogens in breast cancer models. Even though this degree of inhibition fell below the threshold for proteotoxic activation of the heat-shock response and had no overt anticancer activity on its own, it dramatically impaired the emergence of resistance to hormone antagonists both in cell culture and in mice. Our findings strongly support the clinical testing of combined hormone antagonist-low-level HSP90 inhibitor regimens in the treatment of metastatic estrogen receptor-positive breast cancer. At a broader level, they also provide promising proof of principle for a generalizable strategy to combat the pervasive problem of rapidly emerging resistance to molecularly targeted therapeutics.

Keywords: antiestrogen; drug resistance; estrogen receptor; tamoxifen; tumor progression.

Fig. 1.

Very modest pharmacologic inhibition of HSP90 function limits the emergence of Tam resistance in cell culture. (A) Dose–response testing. MCF-7 cells were cultured for 5 d with ganetespib (Gan), Tam, or a combination of Tam (1 µM) plus serial dilutions of Gan. Concentration-dependent inhibition of relative MCF-7 cell mass was measured by standard sulforhodamine B protein assay. (B) Clonogenic assays. MCF-7 cells were cultured continuously in the presence of the indicated concentrations of the HSP90 inhibitor, ganetespib, with or without concurrent Tam (1 µM). After 1 mo, cells were fixed with cold methanol, and the macroscopic outgrowth of resistant colonies was visualized by staining with Diff-Quik. Photomicrographs of representative wells are provided in Fig. S1A.

Fig. 2.

More uniform and profound cell cycle arrest is caused by exposure to a combination of Tam and low-dose HSP90 inhibitor than to either agent alone. (A) Photomicrographs of Diff-Quick-stained MCF-7 cell monolayers after 10 d in culture with Vehicle (DMSO 0.1% vol/vol and ethanol 0.01% vol/vol), Tam 1 µM, Gan 10 nM, or Gan +Tam (G+T, 1 µM + 10 nM). The white arrow indicates outgrowth of a Tam-resistant focus. All images were obtained at the same magnification. (Scale bar, 50 µm.) (B) Relative mRNA expression levels and gene set enrichment analysis for cells treated as in A. (C) Immunoblotting to assess levels of the indicated proliferation-associated proteins in lysates prepared from cultures treated as in A. Tubulin is used as a loading control.

Fig. 3.

Very modest pharmacologic inhibition of HSP90 limits the emergence of Fulv resistance in ER+ breast cancer cells. (A, Upper) Macroscopic images of Diff-Quick-stained wells after culture for 1 mo in the indicated concentrations of ganetespib, with or without concomitant Fulv (1 µM). (Inset) Relative cell number per well, expressed as percentage vehicle control, was monitored using Sytox green fluorescence as a measure of DNA content. MCF-7 and T47D are estrogen-responsive human breast cancer cell lines; HCC-38 is a triple-negative, estrogen-independent cell line. (Lower) Photomicrographs obtained from the same wells. All images were acquired at the same magnification. (Scale bar, 50 µm.) (B) Combined treatment enhances the inhibition of E2F transcriptional activating activity. The indicated cell lines were stably transduced with an E2F-regulated luciferase reporter construct and treated as indicated for 10 d: vehicle (0.1% DMSO + 0.01% ethanol), ganetespib (10 nM), Fulv (1 µM), and G+F (ganetespib 10 nM + Fulv 1 µM). Reporter activity is normalized to the protein concentration in each lysate. Mean of quadruplicate determinations is presented. Error bars, SEM; *P < 10⁻², **P < 10⁻³ (Student t test, two-sided, unpaired).

Fig. 4.

Low-level HSP90 inhibition, having no effect on its own, limits the emergence of Tam resistance in mice and prolongs survival. (A) Antitumor activity of single-agent and combination treatments in estrogen-supplemented mice bearing MCF-7 xenografts. Mean tumor volume (eight mice per treatment group) is plotted. Bars, SEM. P value for the difference between Tamoxifen and Tam + STA groups was determined by two-way repeated-measures ANOVA. (B) Heterogeneity in tumor size within treatment groups at day 39 after cell implantation is reduced by combination treatment. Horizontal bars indicate median with interquartile range. P values were determined by Student t test (two-sided, unpaired). ns, not significant. (C) Combination treatment markedly prolongs event-free survival. Kaplan-Meier analysis is presented, using death or euthanasia for body condition score <2 or excess tumor volume (>1,500 mm³) as event endpoints. The P values for differences from vehicle control were determined by Log-rank test. (D) H&E staining reveals a marked effect of combination treatment on tumor histology. Images were acquired at the same magnification. (Scale bar, 50 µm.) (E) Combination treatment increases expression of membranous MUC1 staining, as monitored by immunohistochemistry (IHC). (Scale bar, 50 µm.) (F) Combination treatment enhances depletion of Cyclin D1 levels detected by IHC. (Scale bar, 50 µm.)