Dual Inhibition of Autophagy and PI3K/AKT/MTOR Pathway as a Therapeutic Strategy in Head and Neck Squamous Cell Carcinoma

Abstract

Genomic analyses of head and neck squamous cell carcinoma (HNSCC) have highlighted alterations in the phosphatidylinositol 3-kinase (PI3K) signaling pathway, presenting a therapeutic target for multiple ongoing clinical trials with PI3K or PI3K/MTOR inhibitors. However, these inhibitors can potentially increase autophagy in HNSCC and indirectly support cancer cell survival. Here, we sought to understand the relationship between the PI3K signaling pathway and autophagy during their dual inhibition in a panel of HNSCC cell lines. We used acridine orange staining, immunoblotting, and tandem sensor Red Fluorescent Protein- Green Fluorescent Protein-, microtubule-associated protein 1 light chain 3 beta (RFP-GFP-LC3B) expression analysis to show that PI3K inhibitors increase autophagosomes in HNSCC cells, but that chloroquine treatment effectively inhibits the autophagy that is induced by PI3K inhibitors. Using the Bliss independence model, we determined that the combination of chloroquine with PI3K inhibitors works in synergy to decrease cancer cell proliferation, independent of the PIK3CA status of the cell line. Our results indicate that a strategy focusing on autophagy inhibition enhances the efficacy of therapeutics already in clinical trials. Our results suggest a broader application for this combination therapy that can be promptly translated to in vivo studies.

Keywords: HNSCC; PI3K inhibitor; PI3K signaling pathway; autophagy; buparlisib; cancer; chloroquine; combination therapy; omipalisib; oral tongue.

Figure 1

PI3Ki induce autophagy flux in head and neck squamous cell carcinoma (HNSCC) cell lines. (A) PI3Ki increase the number of acidic vacuoles. HNSCC cell lines SCC-9 and FaDu were incubated with vehicle dimethyl sulfoxide (DMSO (CTL)) or 1 nM Omi for one day and stained for 15 min. with acridine orange. Nuclei stain green and acidic organelles stain orange. Images are representative of three independent experiments. Scale bar = 50 µm. (B) Western blot analysis of autophagy markers and AKTser473 phosphorylation. Top: Autophagy flux is indicated by SQSTM1 degradation and MAP1LC3B conversion, and inhibition of phosphatidylinositol 3-kinase (PI3K) signaling pathway is indicated by decreased AKTser473 phosphorylation. Equal loading of proteins is shown with a stain-free blot. SCC-9 and FaDu were incubated for three days with vehicle, 0.5 µM Bup or 1 nM Omi. Bottom: Bar graphs show densitometric analysis. Data are presented as mean ± SD, n = 3 independent experiments and analyzed by ANOVA test. Left: SQSTM1 protein levels relative to stain-free blot or TUBA. For FaDu, * p < 0.05 for Bup vs CTL and * p < 0.05 for Omi vs CTL. Middle: MAP1LC3B-II protein levels relative to MAP1LC3B-I for SCC-9 (* p < 0.05 for Bup vs CTL and Omi vs CTL) and FaDu (*** p < 0.001 for Bup vs CTL, * p < 0.05 for Omi vs CTL). Right: AKTser473 protein levels relative to stain-free blot or TUBA, for SCC-9: * p < 0.05 for Bup vs CTL, ** p < 0.01 for Omi vs CTL. For FaDu: **** p < 0.0001 for Bup vs CTL, *** p < 0.001 for Omi vs CTL. (C) Autophagy flux analysis with tandem sensor RFP-GFP-LC3B. PI3Ki increase autophagolysosome (AL) formation. Top: GFP is sensitive to the acidic pH of AL. Red = AL; yellow = autophagosomes (AV). Left: merged stack of confocal images. SCC-9 and FaDu were transduced with the tandem sensor 24 h before the 24 h incubation with inhibitors (1 µM Bup, 5 nM Omi). Cells were stained with 4′,6-Diamidino-2-Phenylindole (DAPI). Scale bar = 50 µm. Right: AL and AV counts per cell of SCC-9 and FaDu using the specialized Image J macro analysis (https://imagejdocu.tudor.lu/plugin/analysis/colocalization_analysis_macro_for_red_and_green_puncta/start). Data are mean ± SD of six images from two independent experiments (three images by experiment per condition with 15–30 cells per picture). For SCC-9 * p < 0.05 for Bup vs CTL, and for FaDu ** p < 0.01 for Omi vs CTL by unpaired Student’s t-test. Uncropped blots are shown in Figure S2.

Figure 2

PI3Ki and chloroquine (CQ) work in synergy to decrease proliferation of HNSCC cell lines. (A) Real-time cell proliferation of SCC-9 captured by IncuCyte S3 (Essen Bioscience, Ann Arbor, MI, USA) for 3 days with vehicle (CTL), 0.5 µM Bup or 1 nM Omi and ± 10 µM CQ. Data are the mean of triplicates ± SEM. and are representative of four independent experiments. (B) Evaluation of population doubling after 3 days in culture for five HNSCC cell lines with/without PI3Ki and CQ at 0.5 µM Bup, 1 nM Omi and 10 µM CQ (except for Detroit 562: 1 µM Bup, 10 nM Omi, and 30 µM CQ). Data are the mean ± SD; n = 4 independent experiments in triplicates for SCC-9, UM22A, and Detroit 562, and n = 5 in triplicates for FaDu and SCC-25. Data were analyzed with ANOVA test, * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001. * p without underscore bar represents significant difference compared with CTL (vehicle). (C) Three-dimensional (3D) diagrams representing Bliss score (z) versus PI3Ki concentration (y) and autophagy inhibitor concentration (x). SCC-9 cell line was incubated with different concentrations of the inhibitors alone or in combination (1, 10, 30 µM CQ; 0.1 or 0.5 µM Bup; 0.1 or 1 nM Omi). Real-time proliferation was recorded for four days with IncuCyte S3 (Essen Bioscience, Ann Arbor, MI, USA) before AUC determination and Bliss score evaluation. Top: Bup. Bottom: Omi. (D) Bliss synergy scores for treatment combinations and PIK3CA status of six HNSCC cell lines. Bliss synergy scores over 1 indicate synergy. Cell lines were incubated in IncuCyte S3 (Essen Bioscience, Ann Arbor, MI, USA), as described with different concentrations of the inhibitors alone or in combination (1, 10, 30 µM CQ; 0.1 or 0.5 µM Bup; 0.1 or 1 nM Omi with the exception of FaDu with 1 or 5 nM Omi, and of Detroit 562 with 0.2 or 1 µM Bup and 2 or 10 nM Omi). Real-time proliferation was recorded for 3-4 days with IncuCyte S3 (Essen Bioscience, Ann Arbor, MI, USA).

Figure 3

CQ inhibits PI3Ki-induced autophagy in HNSCC cell lines. (A) SCC-9 cells were incubated with vehicle DMSO (CTL), 1 nM Omi, 10 µM CQ or both inhibitors for one day and stained for 15 min. with acridine orange. Nuclei stain green and acidic organelles stain orange. Images are representative of three independent experiments. Scale bar = 100 µm. (B) Top: Western blot of autophagy markers SQSTM1 and MAP1LC3B, using alpha-tubuline as a loading control. FaDu cells were incubated for 3 days with vehicle, 0.5 µM Bup or 1 nM Omi and ± 10 µM CQ. Bottom: Densitometric analysis of SQSTM1 protein levels relative to TUBA and of MAP1LC3B-II protein levels relative to MAP1LC3B-I. Data are presented as mean ± SD, n = 3 independent experiments and analyzed by ANOVA test. * p < 0.05 for normalized SQSTM1 Bup vs CTL and for relative MAP1LC3B Bup + CQ vs CTL and Omi + CQ vs CTL. (C) Autophagy flux analysis with tandem sensor RFP-GFP-LC3B. Top: GFP is sensitive to acidic pH of AL. Red = AL, yellow = AV. Bottom: Confocal images and AL and AV counts per cell with specialized Image J macro analysis. SCC-9 cells were transduced with the tandem sensor 24 h before 24 h incubation with inhibitors: 1 nM Rapa, 10 µM CQ. Baf (0.3 uM) was added for the last 4 h. The cells were stained with DAPI. PI3Ki concentrations were 0.5 µM Bup and 1 nM Omi. Scale bar = 50 µm. Data are presented as mean ± SD, six images from two independent experiments (three images by experiment per condition with 15–30 cells per picture) and analyzed by unpaired Student’s t-test: * p < 0.05 for AL Rapa and AL Omi vs AL CTL and ** p < 0.01 for AL Bup vs AL CTL, * p < 0.05 for AV Rapa + Baf and AV CQ vs AV CTL and between AV Omi and AV Omi + CQ, ** p < 0.01 for AV Bup vs AV Bup + C. Uncropped blots are shown in Figure S2.