Targeting ER stress-induced autophagy overcomes BRAF inhibitor resistance in melanoma

Abstract

Melanomas that result from mutations in the gene encoding BRAF often become resistant to BRAF inhibition (BRAFi), with multiple mechanisms contributing to resistance. While therapy-induced autophagy promotes resistance to a number of therapies, especially those that target PI3K/mTOR signaling, its role as an adaptive resistance mechanism to BRAFi is not well characterized. Using tumor biopsies from BRAF(V600E) melanoma patients treated either with BRAFi or with combined BRAF and MEK inhibition, we found that BRAFi-resistant tumors had increased levels of autophagy compared with baseline. Patients with higher levels of therapy-induced autophagy had drastically lower response rates to BRAFi and a shorter duration of progression-free survival. In BRAF(V600E) melanoma cell lines, BRAFi or BRAF/MEK inhibition induced cytoprotective autophagy, and autophagy inhibition enhanced BRAFi-induced cell death. Shortly after BRAF inhibitor treatment in melanoma cell lines, mutant BRAF bound the ER stress gatekeeper GRP78, which rapidly expanded the ER. Disassociation of GRP78 from the PKR-like ER-kinase (PERK) promoted a PERK-dependent ER stress response that subsequently activated cytoprotective autophagy. Combined BRAF and autophagy inhibition promoted tumor regression in BRAFi-resistant xenografts. These data identify a molecular pathway for drug resistance connecting BRAFi, the ER stress response, and autophagy and provide a rationale for combination approaches targeting this resistance pathway.

Figure 1. BRAFi induces autophagy in the tumors of BRAF mutant melanoma patients.

(A) Representative images of pretreatment and resistance samples from patients treated with the BRAF inhibitor vemurafenib. IHC staining (brown) was conducted for the autophagy marker LC3. Original magnification, ×10. (B) IHC scores for pretreatment (Pre) and resistance (Res) samples for 15 patients. *P < 0.05, Res>Pre vs. Res=Pre and Res<Pre. (C) RECIST responses based on a 2+ increase in LC3B staining in resistance versus pretreatment samples. PR, partial response; SD, stable disease. *P < 0.05. (D) PDF in BRAF mutant melanoma patients based on LC3B IHC staining score. P = 0.0487, log-rank test.

Figure 2. BRAFi induces cytoprotective autophagy.

(A) Dose-dependent growth impairment of 6 BRAF mutant cell lines treated in 6 replicates with PLX4720 (72-hour MTT assay). Results are mean ± SEM. (B) Immunoblots and gel density quantifications (mean ± SEM for 3 separate experiments) directed against autophagy markers in 48-hour cell lysates from the indicated cell lines treated with vehicle or PLX4720. (C) A375P mCherry-GFP-LC3 cells were treated with DMSO, 1 μM PLX4720, 500 nM rapamycin, or 10 μM HCQ. Representative merged images of red and green channels after 8 hours of treatment are shown. Original magnification, ×40. Autophagic flux (measured by red puncta) and distal blockade of autophagy (measured by yellow puncta) was quantified (mean ± SD of triplicate experiments). (D) Immunoblotting against autophagy markers (mean ± SEM) in untreated A375PshNT and 3 distinct A375PshBRAF clones. (E) Percent growth inhibition after 48 hours of 10 μM HCQ treatment compared with control in A375PshNT and A375PshBRAF cells (MTT assay). Results are mean ± SEM. (F) A375P cells were treated with vehicle, the BRAF inhibitor GSK2118436, the MEK inhibitor GSK1120212, 10 μM HCQ, or the indicated combinations. Shown are 48-hour immunoblots directed against the indicated proteins. *P < 0.05.

Figure 3. Autophagy inhibition augments growth impairment associated with BRAFi.

(A) Percent growth inhibition (mean ± SEM) compared with control for 1 μM PLX4720 treatment with or without 10 μM HCQ (left), or for 1 μM PLX4720 treatment of MEL624shNT, MEL624shATG5, A375PshNT, and A375PshATG5 clones (right). (B) 14 day clonogenic growth assays for BRAFi-sensitive A375P and BRAFi-resistant MEL1617R and MEL624. Cells were plated in triplicate and immediately treated with PBS vehicle, 500 nM PLX4720, 5 μM HCQ, or the combination. Colony counts (mean ± SEM) are presented. (C) 72-hour MTT assay was conducted after treating the indicated cell lines with PBS, 1 μM PLX4720, 3 μM Lys05, or the combination. Results are mean ± SEM normalized to control. (D) MEL624 and A375P cells were treated with 30 nM GSK2118436, 20 nM GSK1120212, 10 μM HCQ, and the indicated combinations. 72-hour MTT (mean ± SEM) was performed. *P < 0.05.

Figure 4. BRAFi induces an early ER stress response.

(A) Immunoblotting against the indicated signaling markers after treatment of MEL624 cells with vehicle or 1 μM PLX4720 for the indicated times. P-, phospho-. (B) MEL624 and A375P cells were treated with 1 μM PLX4720 for the indicated times. Whole-cell and nuclear lysates were subjected to immunoblotting. Tg, thapsigargin. (C) PLA performed on MEL624 cells treated for 30 minutes with DMSO or 10 μM PLX4720 and immunofluorescence for the ER resident protein disulfide isomerase (green), which reflects total ER area. Red fluorescence reflects a protein-protein interaction. Original magnification, ×40. Quantification of red, green, and yellow (colocalization) signals (mean ± SD) reflects at least 3 separate experiments.

Figure 5. Targeting the BRAFi-induced ER stress response blocks autophagy and enhances cell death.

(A and B) The indicated cells were treated with vehicle, 1 μM PLX4720, 1 μM of the PERK inhibitor GSK2606414, or both for 24 or 48 hours. Immunoblots against the indicated ER stress and apoptosis markers are shown. c-, cleaved; CC3, cleaved caspase-3. (C) Cell viability (Trypan blue staining). Results are mean ± SEM. *P < 0.05. (D) Immunoblot against LC3 in MEL624 cells treated with vehicle, 1 μM PLX4720, 1 μM GSK2606414, or both for 24 hours. (E and F) MEL624 and A375P cells were treated with PLX4720 (24 hours) and/or siRNA (48 hours) against nontarget control and PERK. Representative immunoblots and LC3II/LC3I ratios (mean ± SEM; based on gel density quantification from 2 separate experiments) are shown. du, duplex.

Figure 6. Autophagy inhibition augments BRAFi-induced cell death and antitumor activity in vivo.

(A and B) Mice bearing MEL624 melanoma xenografts were treated with PBS or Lys05 (40 mg/kg) i.p. daily and fed either control or PLX4720-infused chow (4–6 per cohort). Tumors were harvested after 22 days of treatment. (A) Change in tumor volume from baseline. Results are mean ± SEM. *P < 0.05. (B) Immunoblots and gel quantification (mean ± SEM) of individual tumors. (C) Representative electron micrographs. Red arrows denote AVs. Scale bars: 2 μm.