A kinome-wide screen identifies the insulin/IGF-I receptor pathway as a mechanism of escape from hormone dependence in breast cancer

Abstract

Estrogen receptor α (ER)-positive breast cancers adapt to hormone deprivation and become resistant to antiestrogens. In this study, we sought to identify kinases essential for growth of ER(+) breast cancer cells resistant to long-term estrogen deprivation (LTED). A kinome-wide siRNA screen showed that the insulin receptor (InsR) is required for growth of MCF-7/LTED cells. Knockdown of InsR and/or insulin-like growth factor-I receptor (IGF-IR) inhibited growth of 3 of 4 LTED cell lines. Inhibition of InsR and IGF-IR with the dual tyrosine kinase inhibitor OSI-906 prevented the emergence of hormone-independent cells and tumors in vivo, inhibited parental and LTED cell growth and PI3K/AKT signaling, and suppressed growth of established MCF-7 xenografts in ovariectomized mice, whereas treatment with the neutralizing IGF-IR monoclonal antibody MAB391 was ineffective. Combined treatment with OSI-906 and the ER downregulator fulvestrant more effectively suppressed hormone-independent tumor growth than either drug alone. Finally, an insulin/IGF-I gene expression signature predicted recurrence-free survival in patients with ER(+) breast cancer treated with the antiestrogen tamoxifen. We conclude that therapeutic targeting of both InsR and IGF-IR should be more effective than targeting IGF-IR alone in abrogating resistance to endocrine therapy in breast cancer.

Figure 1

RNAi screening identifies InsR is required for hormone-independent growth. A) MCF-7/LTED cells were screened with a siRNA library targeting 779 kinases. Cell viability was measured after 4 days using Alamar Blue. Data are presented as log₂ median % cell growth (from 4 independent experiments) for each siRNA relative to controls. B) Lysates from 10 ER+ tumors from patients treated with letrozole were analyzed by RPPA using 190 antibodies. Antibody signal intensities were log₂-normalized and correlated to the post-letrozole Ki67 score. Antibodies with a correlation (r) >0.4 (n=51) were filtered and used to generate a heatmap of signal intensities. These 51 total and phosphoproteins were analyzed using KEGG Pathway Analysis. Proteins involved in insulin signaling are indicated (red arrows).

Figure 2

Knockdown of InsR and/or IGF-1R inhibits LTED cell growth, but dual knockdown suppresses PI3K/AKT. Cells were transfected with siRNA targeting InsR, IGF-1R, both, or a non-silencing control (siCon) and re-seeded the next day for growth in monolayer (A) or immunoblot analyses (B). A) Cells were treated with 10% DCC-FBS for 6–8 days, then trypsinized and counted. Data are presented as % of siCon; each bar, mean ± SEM (n=3). *p<0.05 vs. siCon, one-way ANOVA. B) Transfected cells were maintained in 10% DCC-FBS for 3 days; protein lysates were analyzed by immunoblot using the indicated antibodies.

Figure 3

Pharmacological blockade of InsR/IGF-1R inhibits hormone-independent growth and PI3K/AKT. A) MCF-7 cells in serum-free medium were treated overnight ± 0–5 μM OSI-906 or 1 μM AEW-541. Cells were treated for 15 min ± 100 ng/ml IGF-1 or 10 μg/ml insulin. Protein lysates were analyzed by immunoblot using the indicated antibodies. B) LTED cells in 10% DCC-FBS were treated ± 4 μM OSI-906 for 3–24 h. Cell lysates were prepared and analyzed using phospho-RTK arrays. C) Left: Cells in 10% DCC-FBS were treated for 24 h ± 4 μM OSI-906. Lysates were analyzed by immunoblot. Right: Cells were treated overnight ± 4 μM OSI-906, then treated for 15 min ± 10% DCC-FBS, IGF-1, or insulin. Cell lysates were immunoprecipitated with a p85 antibody; lysates and immunoprecipitates were analyzed by immunoblot. D) Cells in 10% DCC-FBS were treated ± 4 μM OSI-906. Media and drugs were replenished every 3 days. Cells were counted after 5–10 days. Data are presented as % of parental control; each bar, mean ± SEM (n=3). *p<0.01, **p<0.0001, two-way ANOVA.

Figure 4

OSI-906 prevents the emergence and inhibits growth of established hormone-independent tumors. A–B) MCF-7 cells were injected s.c. into athymic mice supplemented with 14-day release E2 pellets. Mice with no tumors (A) or bearing tumors ≥ 150 mm³ (B) were randomized to vehicle or OSI-906 (50 mg/kg/day, p.o.) for six weeks. Data are presented as number of tumors formed; p=0.02, Fisher’s Exact (A), or mean tumor volume ± SEM; *p<0.05 vs. vehicle, two-way ANOVA (B). C) Tumor-bearing mice from (A) were treated with vehicle or OSI-906 for 3 days. Xenografts were harvested 4 h after the last dose. Tumor lysates were precipitated with a p-Tyr antibody; p-Tyr pulldowns and tumor lysates were analyzed by immunoblot with the indicated antibodies. D) Tumor-bearing mice were imaged before and 4 h after the initial dose of OSI-906 by [¹⁸F]FDG-PET. Images from a representative mouse show [¹⁸F]FDG uptake pre- and post-OSI-906 (T-tumor). Quantification is shown below. *p<0.0001, two-way ANOVA.

Figure 5

Dual blockade of InsR/IGF-1R prevents emergence of hormone-independent breast cancer cells. A) MCF-7/LTED cells in serum-free medium were treated overnight ± 4 μM OSI-906, 3 μg/ml MAB391, or both. The next day, cells were treated for 15 min ± 100 ng/ml IGF-1 or 10 μg/ml insulin; protein lysates were analyzed by immunoblot using the indicated antibodies. B) Cells in 10% DCC-FBS were treated for 24 h ± the indicated inhibitors. Lysates were analyzed as above. C) Parental cells in 10% DCC-FBS were treated ± the indicated inhibitors. Media and drugs were replenished every 2–3 days. When control wells reached 60–70% confluence, cells were fixed and stained with crystal violet. Representative images and quantification of integrated intensity (% control) are shown. *p<0.05 vs. control, t-test. D) MCF-7 cells in 2% DCC-FBS were treated ± inhibitors and ligands as indicated. Media and drugs were replenished every 3 days. Cells were counted after 5 days. Data are presented as % control, each bar, mean ± SEM (n=3). *p<0.05, two-way ANOVA.

Figure 6

Blockade of InsR/IGF-1R suppresses hormone-independent tumor growth. A) Mice bearing MCF-7 xenografts ≥ 150 mm³ were randomized to the indicated treatments. Data are presented as mean tumor volume ± SEM. *p<0.05 vs. vehicle at each time point, two-way ANOVA. B) IHC for total IGF-1R and P-InsRβ_Y1146/IGF-1Rβ_Y1131. Left: Representative images of tumors from (A). Right: Quantitative comparison of membrane histoscores. **p<0.0001, *p<0.05 vs. vehicle, one-way ANOVA. C) Blood was collected via submandibular bleeding; serum insulin was measured in duplicate. Data are presented as insulin levels ± SEM. *p<0.05 vs. vehicle; ^#p<0.05 vs. OSI-906, t-test. D) Tumor-bearing mice were randomized to the indicated treatments as in (A). *p<0.05 vs. vehicle; ^#p<0.05 vs. OSI-906, two-way ANOVA.

Figure 7

Insulin/IGF-1-induced gene expression correlates with patient outcome after endocrine therapy. MCF-7 cells were serum-starved for 24 h, then treated ± 10 μg/ml insulin for 4 or 24 h. RNA was isolated and analyzed using gene expression microarrays. A) A tumor signature of insulin-induced gene expression correlates inversely with RFS in patients with ER+ breast cancer treated with tamoxifen. B) Genes altered by insulin or IGF-1 stimulation were evaluated by GSA. Genesets were grouped by hierarchical clustering and shown as a heatmap. C) An insulin/IGF-1 gene expression signature predicts RFS in two cohorts of patients with ER+ breast cancer treated with tamoxifen.