BRAFV600E-dependent Mcl-1 stabilization leads to everolimus resistance in colon cancer cells

Abstract

mTOR activation is commonly caused by oncogenic mutations in RAS/RAF/MAPK and PI3K/AKT pathways, and promotes cancer progression and therapeutic resistance. However, mTOR inhibitors show limited single agent efficacy in patients. mTOR inhibitors suppress tumor cell growth and angiogenesis, and have recently been shown to induce death receptor/FADD-dependent apoptosis in colon cancers. Using a panel of BRAF V600E and WT colorectal cancer cell lines and in vitro selected resistant culture, and xenograft models, we demonstrate here that BRAFV600E confers resistance to mTOR inhibitors. Everolimus treatment disrupts the S6K1-IRS-2/PI3K negative feedback loop, leading to BRAF V600E-dependent activation of ERK and Mcl-1 stabilization in colon cancer cells, which in turn blocks the crosstalk from the death receptor to mitochondria. Co-treatment with inhibitors to Mcl-1, PI3K, RAF or MEK restores mTOR inhibitor-induced apoptosis by antagonizing Mcl-1 or abrogating ERK activation in BRAFV600E cells. Our findings provide a rationale for genotype-guided patient stratification and potential drug combinations to prevent or mitigate undesired activation of survival pathways induced by mTOR inhibitors.

Keywords: BRAF V600E; ERK; Mcl-1; everolimus; mTOR.

Figure 1. BRAF V600E colon cancer cells are resistant to Everolimus

(A) isogenic pairs of BRAF WT and V600E (E) RKO and VACO432 cells were treated with 20 and 25 μM Everolimus, respectively. Attached cells after 48 h were stained by crystal violet. (B) cells treated as in A were analyzed for apoptosis by counting condensed and fragmented nuclei. **P < 0.01, 600E vs. WT. (C) cells treated as in A for 24 h were analyzed by western blotting. β-Actin was used as a loading control. (D) cells were treated as in A, stained with Annexin V/propidium iodide, and analyzed by flow cytometry (Right). Left, quantitation of Annexin V+ cells. (E) the growth of 10 colon cancer cell lines was determined by MTS assay following 72 h treatment with varying doses of Everolimus (10 nM to 20 μM). (F) apoptosis was analyzed after 48 h of 20 μM Everolimus. (G) cells treated as in F for 24 h were analyzed by western blotting.

Figure 2. Elevated Mcl-1 in BRAF 600E cells inhibits Everolimus-induced apoptosis

(A) isogenic RKO (top) and VOCO432 (bottom) cells were treated with 20 μM Everolimus for 24 h and analyzed by western blotting. (B) BRAF 600E RKO (Top) and VOCO432 (Bottom) cells were transfected with either scrambled or Mcl-1 siRNA for 24 h, then treated with Everolimus for 48 h, and analyzed for apoptosis. **P < 0.01, 600E vs. WT. (C) BRAF WT (+/−) RKO cells were transfected with either pcDNA or V5-Mcl-1 for 24h, then treated with Everolimus for 48 h and analyzed for apoptosis. (D) BRAF WT (+/−) RKO (top) or VACO431 cells (bottom) were transfected by either a scrambled, DR5, Bid siRNA for 24 h, treated with Everolimus for 48 h, and analyzed for apoptosis. **P < 0.01, *P < 0.05, V5-Mcl-1, DR5 and Bid siRNA vs. control. *DR5 specific band. (E) ten CRC cell lines with BRAF WT or 600E were treated with Everolimus for 24 h and analyzed by western blotting. B, C and D, apoptosis was analyzed by counting condensed and fragmented nuclei. Western blotting confirmed target depletion or overexpression. β-Actin was used as a loading control.

Figure 3. Everolimus treatment disrupts the S6K1-IRS-2/PI3K negative feedback and leads to MAPK activation and Mcl-1 stability in BRAF 600E cells

(A) BRAF 600E RKO cells were transfected with either a scrambled, or BRAF siRNA for 24 h, treated with 20 μM Everolimus for 48h, and analyzed for apoptosis. **P < 0.01, BRAF vs. scramble siRNA. (B) cells treated as in A for 24 h and analyzed by western blotting. (C) isogenic RKO cells were treated with Everolimus for 6 h plus 5 μM proteasome inhibitor MG132 and 20 μg/ml Cycloheximide (CHX). Mcl-1 was immunoprecipitated (IP) and probed for p-Mcl-1. (D) isogenic RKO cells were treated with Everolimus and 20 μg/ml cycloheximide (CHX), and analyzed by western blotting. (E) relative Mcl-1 band intensity in D normalized to β-Actin compared to that of the t = 0 as 100%. (F) BRAF 600E RKO cells were transfected with either S6K1 (HAS6K1) or a constitutively active S6K1 (HAS6K1 E389 ΔCT) for 24 h, then treated with Everolimus for 24 h, and analyzed by western blotting. (G) RKO cells were transfected with either a scrambled, or IRS-2 siRNA for 24 h, then treated with Everolimus for 24 h, and analyzed by western blotting.

Figure 4. Pathway inhibitors restore apoptosis in BRAF600E cells

(A) BRAF 600E RKO (Left) and VOCO432 (Right) cells were treated with 20 μM Everolimus, 5 μM TW-37, or their combination for 48 h, and analyzed for apoptosis. **P < 0.01, combination vs. single agent. (B) cells treated as in A for 24 h were analyzed by western blotting. (C) cells were treated with 20 μM Everolimus, 10 μM LY249002 (PI3Ki), 10 μM Sorafenib (RAFi) or 10 μM AS703026 (MEKi), or their combination for 48h, and analyzed for apoptosis. **P< 0.01, combination vs. single agent. The untreated and Everolimus bars are the same in each set. (D–F) cells were treated with indicated single agents or combinations as in C for 24 h and analyzed by western blotting.

Figure 5. BRAF 600E tumors are resistant to Everolimus in vivo with reduced apoptosis and elevated p-ERK and Mcl-1

(A) BRAF WT or 600E RKO tumors were established in nude mice. Mice were randomized into two groups when tumors reach 50 mm³ to receive Everolimus (7.5 mg/kg/day) or vehicle for 10 consecutive days. Tumor volume was monitored 3 times per week and plotted. N = 8 mice/group. *P-value was calculated on the last measurements. (B) representative images of tumors at the end of experiments. (C) representative images of cleaved caspase-3 staining in tumors and quantitation (right). *P < 0.05, 600E vs. WT. (D) representative images of p-ERK staining in tumors. (E) two randomly chosen WT or 600E tumors from each group were harvested the day after the last treatment, and analyzed by western blotting.

Figure 6. Mcl-1 induction and BRAF600E and acquired Everolimus resistance

Three parental BRAF WT cell lines, RKO (+/−), VACO432 (+/−) Lim1215 (+/+) were subjected to multiple cycles of Everolimus treatment to select for RAD001 resistant (RR) culture. (A) the parental and resulting RR cell lines were treated with 20 μM Everolimus for 48 h, and analyzed for apoptosis. **P < 0.01, RR vs. parental culture. (B) cells treated as in A for 24 h were analyzed by western blotting. (C) detection of BRAF V600E in RKO-RR culture that started from BRAF WT cells. Red arrows denotes the affected nucleotide and codon. (D) APC or KRAS genotype and Temsirolimus sensitivity in digestive cancer cell lines obtained from Genomics of Drug Sensitivity in Cancer Project. (E) a proposed model for Everolimus resistance in BRAF 600E cells via ERK-mediated Mcl-1 stabilization that blocks the crosstalk between the death receptor and mitochondrial pathways.