Lysosomal damage due to cholesterol accumulation triggers immunogenic cell death

Abstract

Cholesterol serves as a vital lipid that regulates numerous physiological processes. Nonetheless, its role in regulating cell death processes remains incompletely understood. In this study, we investigated the role of cholesterol trafficking in immunogenic cell death. Through cell-based drug screening, we identified two antidepressants, sertraline and indatraline, as potent inducers of the nuclear translocation of TFEB (transcription factor EB). Activation of TFEB was mediated through the autophagy-independent lipidation of MAP1LC3/LC3 (microtubule associated protein 1 light chain 3). Both compounds promoted cholesterol accumulation within lysosomes, resulting in lysosomal membrane permeabilization, disruption of autophagy and cell death that could be reversed by cholesterol depletion. Molecular docking analysis indicated that sertraline and indatraline have the potential to inhibit cholesterol binding to the lysosomal cholesterol transporters, NPC1 (NPC intracellular cholesterol transporter 1) and NPC2. This inhibitory effect might be further enhanced by the upregulation of NPC1 and NPC2 expression by TFEB. Both antidepressants also upregulated PLA2G15 (phospholipase A2 group XV), an enzyme that elevates lysosomal cholesterol. In cancer cells, sertraline and indatraline elicited immunogenic cell death, converting dying cells into prophylactic vaccines that were able to confer protection against tumor growth in mice. In a therapeutic setting, a single dose of each compound was sufficient to significantly reduce the outgrowth of established tumors in a T-cell-dependent manner. These results identify sertraline and indatraline as immunostimulatory agents for cancer treatment. More generally, this research shed light on novel therapeutic avenues harnessing lysosomal cholesterol transport to regulate immunogenic cell death.Abbreviation: ATG5: autophagy related 5; ATG13: autophagy related 13; DKO: double knockout; ICD: immunogenic cell death; KO: knockout; LAMP1: lysosomal associated membrane protein 1; LAMP2: lysosomal associated membrane protein 2; LGALS3: galectin 3; LDL: low-density lipoprotein; LMP: lysosomal membrane permeabilization; MAP1LC3/LC3: microtubule associated protein 1 light chain 3; MTX: mitoxantrone; NPC1: NPC intracellular cholesterol transporter 1; NPC2: NPC intracellular cholesterol transporter 2; TFE3: transcription factor E3; TFEB: transcription factor EB; ULK1: unc-51 like autophagy activating kinase 1.

Keywords: Autophagy; NPC intracellular cholesterol transporter 1 and 2; TFEB; cancer; lipid transport; lysosomal membrane permeabilization.

Figure 1.

Identification of novel TFEB inducers by cell-based high-content screening. U2OS GFP-TFEB RFP-LC3 cells were treated with each of 1200 compounds from the Prestwick Chemical library at 10 μM for 6 h and then subjected to high-content microscopy and analysis of TFEB and LC3 subcellular distribution. Torin 1 (300 nM) was used as a positive control for TFEB translocation and LC3 puncta formation. (A) Representative fluorescence confocal images of cells treated with torin 1 illustrate the nuclear translocation of GFP-TFEB and the formation of RFP-LC3 dots as indicators of TFEB activation and LC3 lipidation, respectively. Scale bars: 10 μm. (B) The screening data illustrates the ratio of nuclear:cytoplasmic TFEB fluorescence intensity following the treatment with the Prestwick Chemical library in U2OS cells. The cutoff threshold for hit selection was set at 0.5 and the color code represents therapeutic classes. (C) The dot plot shows for the selected hits, the nuclear:cytoplasmic TFEB intensity ratio versus LC3 dots area quantification. (D) Representative micrographs depict endogenous TFEB in U2OS cells treated with the selected TFEB inducers (10 μM, 6 h) using a TFEB antibody. Scale bars: 10 μm. (E) Quantification of endogenous TFEB nuclear fluorescence intensity versus cytoplasmic fluorescence intensity. Individual data points correspond to the mean TFEB nuclear to cytoplasmic fluorescence intensity ratio per cell from each image (n = 15–16 images, corresponding to a minimum of 4,000 cells per condition). Statistical significance was determined by an unpaired Mann-Whitney U test (two-tailed). (F) To evaluate cytotoxicity, cells were incubated with the selected agents for 24 h and then subjected to cell death assessment by flow cytometry analyzing the loss of mitochondrial membrane potential (Δψm) and cytoplasmic membrane permeabilization. Values are expressed as percentages ± SD. Statistical significance analysis is detailed in the materials and methods section.

Figure 2.

Indatraline and sertraline regulate autophagy-lysosomal gene expression levels and autophagic flux. (A) Changes in the expression levels of autophagy-lysosomal genes assessed by qPCR analysis of cells exposed to either indatraline (10 μM) or sertraline (10 μM) for 6 h. Torin 1 (300 nM) was used as a positive control for inducing lysosomal genes dependent on TFEB. Statistical significance was determined by an unpaired Mann-Whitney U test (two-tailed). The p-values are provided in figure S2A. (B) Left panel U2OS GFP-RFP-LC3 cells were treated with either indatraline (10 μM) or sertraline (10 μM) for the indicated times. Torin 1 (300 nM) as positive control of autophagic flux and bafilomycin A₁ (50 nM) as control of autophagy blockade were used. Representative confocal images are shown for 6 h post treatment with the indicated compounds. Scale bars: 10 μm. Right top panel, a schematic representation of the tandem GFP-RFP-LC3 reporter as a tool for autophagic flux assessment. (C) Line plots show the kinetic of autophagosome and autolysosome formation in response to the indicated compounds. The surface overlap coefficient (SOC) of GFP and RFP signals (autophagosomes) was calculated while the RFP signal is indicative of the presence of autolysosomes (red dots only). Data at each time point corresponds to the average value per cell from each image (n = 16 images, corresponding to a minimum of 2,000 cells per condition). Statistical significance was determined by an unpaired Mann-Whitney U test (two-tailed), p-values were calculated for each time point and the maximum p-values for each time point are depicted. (D) Representative confocal images of endogenous LC3 dots in U2OS cells (WT or ATG 13 and ULK1 depleted) treated with torin 1 (300 nM, 6 h), EBSS (2 h), indatraline (10 μM, 6 h) or sertraline (10 μM, 6 h). Scale bars: 10 μm. (E) Immunoblot analysis of ATG13 and ULK1 in WT cells and ATG13 or ULK1 depleted U2OS cells. (F) Quantification of the total area of dots per cell of endogenous LC3 puncta performed in D. (G) A schematic representation of the distinct phases of autophagosome formation. (H) Quantification of the SOC of LC3 and the lysosomal membrane protein LAMP2 is shown. For F and H, individual data points correspond to the average value per cell from each image (n = 16 images, corresponding to a minimum of 2,000 cells per condition). Statistical significance was determined by an unpaired Mann-Whitney U test (two-tailed).

Figure 3.

Lipidation of LC3 is indispensable for TFEB nuclear translocation triggered by indatraline and sertraline. (A) Schematic representation of proteins involved in the lipidation process of LC3. (B) Immunoblot analysis of LC3 lipidation in WT and in ATG5-depleted U2OS cells after 6 h of treatment with either indatraline (10 μM) or sertraline (10 μM). ACTB was detected to control equal loading. (C) Representative confocal images of endogenous TFEB staining in U2OS cells and cells lacking ATG5 and ULK1 after 6 h of treatment with either indatraline (10 μM), sertraline (10 μM) or positive controls of TFEB nuclear translocation (torin 1 at 300 nM, 6 h and EBSS for 2 h). Scale bars: 10 μm. (D) Box plots show the ratio of nuclear to cytoplasmic TFEB fluorescence intensity. For each cell line, the results were normalized to the values of untreated cells. Individual data points correspond to the ratio of nuclear to cytoplasmic TFEB fluorescence intensity per cell from each image (n = 21–27 images, corresponding to a minimum of 3,000 cells per condition). Statistical significance was determined by an unpaired Mann-Whitney U test (two tailed).

Figure 4.

Indatraline and sertraline induce lysosomal membrane permeabilization. (A) Representative fluorescence images of U2OS LAMP1-GFP cells treated with amlodipine (10 μM), for indicated times. The intrinsic fluorescence of amlodipine (assessed with a DAPI filter set) is shown together with LAMP1-GFP. Scale bars: 10 μm. (B) Quantification of the area of amlodipine⁻positive dots after treatment with amlodipine (10 μM) for either 30 min or 18 h. (C) The dot plot represents the surface overlap coefficient (SOC) of LAMP1-GFP positive dots and amlodipine positive dots after treatment with amlodipine. For B and C, individual data points correspond to the average value per cell from each image (n = 3 images, corresponding to a minimum of 500 cells per condition). Statistical significance was determined by an unpaired Mann-Whitney U test (two tailed). (D) The dot plot shows the distribution of selected agents based on their basic pKa value and LogP values. LLOMe and chloroquine were included as positive controls for assessing lysosomotropism while torin 1 serves as a negative control. (E) Representative confocal images of U2OS cells that co-express mCherry-LGALS3 (galectin 3) and LAMP1-GFP after treatment with the indicated agents for 18 h. LLOMe (500 μM, 2 h) was used as a positive control of lysosomal membrane permeabilization (LMP). Scale bars: 10 μm. (F) Quantification of lysosomal membrane permeabilization based on the relocalization of LGALS3 from the cytosol to lysosomes. Results are expressed as LGALS3 dots area per cell. Individual data points correspond to the average value per cell from each image (n = 16–18 images, corresponding to a minimum of 10,000 cells per condition). Statistical significance was determined by one-way ANOVA with Tukey’s multiple comparison test. (G) Representative confocal images of U2OS cells that co-express mCherry-LGALS3 (galectin 3) and LAMP1-GFP pretreated with either β-cyclodextrin (1 mM), N-acetylcysteine (4 mM) or the pan-caspase inhibitor QVD-OPH (25 μM) for 2 h and then exposed to indatraline (10 μM) or sertraline (10 μM) for 18 h. Scale bars: 10 μm. (H) Top panel, a schematic representation of the LGALS3 relocalization from the cytosol to lysosomes. Bottom panel, quantification of LMP was performed as described in F. Individual data points correspond to the average value per cell from each image (n = 12–16 images, corresponding to a minimum of 4,000 cells per condition). Statistical significance was determined by an unpaired Mann-Whitney U test (two-tailed).

Figure 5.

Indatraline and sertraline promote the accumulation of cholesterol within lysosomes. (A) Representative confocal images of U2OS LAMP1-GFP cells treated with either indatraline (10 μM) or sertraline (10 μM) for 18 h. U18666A (1 μM) was used as a positive control of cholesterol accumulation in lysosomes. Cells were then stained with filipin complex, which permits the visualization of free cholesterol. Scale bars: 10 μm. (B) The average area of filipin positive dots indicative of the presence of cholesterol was scored. (C) The surface overlap coefficient (SOC) of filipin positive dots and the lysosomal membrane protein LAMP1 is presented. In the box plot, mean (+), median as well as the minimum and maximum values are shown. For B and C, individual data points correspond to the average value per cell from each image (n = 16–64 images, corresponding to a minimum of 3,000 cells per condition). Statistical significance was determined by an unpaired Mann-Whitney U test (two-tailed). (D) Representative confocal images of BODIPY-cholesterol staining after treatment of U2OS cells with indatraline (10 μM) or sertraline (10 μM) for 18 h. Scale bars: 10 μm. The average BODIPY-cholesterol dots area, indicative of the presence of cholesterol, was scored (n = 9 images, corresponding to a minimum of 200 cells per condition). Statistical significance was calculated by means of an unpaired Mann-Whitney U test (two-tailed). (E) Representative electron micrographs of U2OS cells after treatment with either indatraline (10 μM) or sertraline (10 μM) for 6 h. Arrows indicate the presence of multilamellar vesicles (MLVs). In the box plot, each dot represents the number of MLVs per cross-section. The mean (+), median as well as the minimum and maximum values are shown. (F) Prediction of the interaction between indatraline and the NTD (N terminal domain domain) of the NPC1 protein and with the NPC2 protein. For each case, a zoom on the active site with the interactive residue is indicated in the box. (G) Prediction of the interaction of sertraline and the NTD (N terminal domain domain) of the NPC1 protein and with the NPC2 protein. For each case, a zoom on the active site with the interactive residue is indicated in the box. (H) Quantification of filipin dots area in non-targeting control (NTC) cells and TFEB TFE3 DKO U2OS cells following treatment with indatraline (10 μM) or sertraline (10 μM) for the indicated times points. For each time point, the data are expressed as fold change relative to untreated cells. Statistical significance was obtained by an unpaired Mann-Whitney U test (two-tailed). (I) Immunoblot analysis of NPC1, NPC2, PLA2G15 and LAMP1 levels in NTC cells and TFEB TFE3 DKO cells after treatment with indatraline (10 μM) or sertraline (10 μM) for 16 h. Protein levels were quantified after normalization to ACTB and are presented as fold change relative to control cells.

Figure 6.

Lysosomal cholesterol accumulation is critical for the induction of immunogenic cell death (ICD) hallmarks. To characterize the cell death modality, murine fibrosarcoma MCA205 cells were exposed to various inhibitors of cell death pathways as described in (A-C) and then treated with either indatraline (20 μM) or sertraline (20 μM) for 24 h. Cell death was evaluated using DAPI staining followed by flow cytometry analysis as described in the methods section. (A) MCA205 cells were pre-incubated with various cathepsin inhibitors during 2 h, namely odanacatib (5 μM), E64d (20 μM), and pepstatin a methyl ester (20 μM) before the addition of indatraline (20 μM) and sertraline (20 μM). (B) MCA205 cells pre-treated during 2 h with specific inhibitors for cell death pathways, namely apoptosis (Q-VD-Oph, 25 μM), ferroptosis (ferrostatin-1, 10 μM), or necroptosis (necrostatin-1, 10 μM) before the addition of indatraline (20 μM) and sertraline (20 μM). (C) MCA205 cells were incubated with the cholesterol-depleting agent β-cyclodextrin (1 mm) or LDL-depleted medium (serum-free medium supplemented with 1% BSA) for 2 h before treatment with indatraline (20 μM) or sertraline (20 μM). In A, B and C, results were expressed as percentage of cell death ± SD. Statistical significance analysis is detailed in the materials and methods section. (D) U2OS cells were treated with the cholesterol-depleting agent β-cyclodextrin (1 mM) or LDL-depleted medium for 2 h before treatment with either indatraline (10 μM) or sertraline (10 μM) for 18 h. U18666A (1 μM) was used as a positive control of cholesterol accumulation in lysosomes. Cells were then stained with filipin complex, which permits the visualization of free cholesterol within the cells. Representative images of filipin staining are shown. Scale bars: 10 μm. (E) The average area of filipin positive dots indicative of the presence of cholesterol was scored. (F) Evaluation of plasma membrane exposure of calreticulin in MCA205 cells after treatment with either indatraline (20 μM) or sertraline (20 μM) for 24 h. When indicated, cells were pretreated for 2 h with β-cyclodextrin (1 mM) or LDL-depleted medium prior to the addition of indatraline or sertraline. The percentage of calreticulin-positive cells among the live population of cells was quantified. (G) MCA205 cells were treated as mentioned in (F) and the concentration of ATP release in the cell supernatant was quantified. Mitoxantrone (2 μM) was used as positive control for the induction of ATP release. (H) MCA205 cells were treated as mentioned in (F), and the release of HMGB1 into the culture media was assessed. For F, G and H, statistical significance was determined by an unpaired Mann-Whitney U test (two-tailed).

Figure 7.

Anti-tumor prophylactic activity of sertraline and indatraline-treated malignant cells. (A) Experimental schedule of prophylactic vaccination assay. MCA205 cells were treated in vitro with either indatraline (20 μM), sertraline (20 μM) or mitoxantrone (2 μM) as a positive control, to reach about 70% cell death. Dying MCA205 cells were injected by subcutaneous injection (s.c.) in the left flank of the mice. Ten days later, mice vaccinated or not with tumor cell lysates were challenged with living MCA205 cells injected s.c. into the right flank of the mice. Tumors were measured three times per week and their volume was calculated with the following formula: 4/3 * π * L/2 * I/2 * h/2. The curves show tumor growth evolution. (B-D) Individual (B) and mean (C) tumor growth, as well as tumor-free survival (D) of mice treated with indatraline. (E-G) Individual (E) and mean (F) tumor growth, as well as tumor-free survival (G) of mice treated with sertraline. For (C) and (F) the error bars represent the standard error of the mean (SEM). For tumor-free mice graphs, the statistical significance was determined by a log-rank (Mantel-Cox) test. (H) Graphical representation of tumor rechallenge in surviving mouse. Mice that did not develop tumor after the initial vaccination-tumor challenge procedure were rechallenged by s.c. injection of living and untreated MCA205 cells in the right flank and B16-F10 melanoma cells in the opposite flank. (I-J) Volumes of MCA205 and B16-F10 tumors 24 days post-tumor rechallenge in mice initially vaccinated with extracts of MCA205 cells treated with indatraline (I) or sertraline (J). Data were analyzed using TumGrowth (https://github.com/kroemerlab).

Figure 8.

Therapeutic efficacy of indatraline and sertraline. (A) Experimental schedule of implantation and treatment of fibrosarcoma. (B-D) MCA205 cells (300,000) were injected subcutaneously in the right flank of the mice. Once the tumor was established, mice were randomized and treated intraperitoneally with indatraline (10 mg/kg) or sertraline (10 mg/kg). T cells were depleted or not by i.p. injections of anti-CD4 and anti-CD8 (100 μg per mouse of each antibody) at the indicated days. The groups of mice that did not received injections of anti-CD4 and anti-CD8 antibodies were instead administrated an isotype control (IgG). Tumors were measured three times per week and their volume was calculated with the following formula: 4/3 * π * L/2 * I/2 * h/2. Data were analyzed using TumGrowth (https://github.com/kroemerlab). (E) Overall survival curves from the experiment performed in B-D are reported. Statistical significance was determined by a log-rank (Mantel-Cox) test.