Distinct GSDMB protein isoforms and protease cleavage processes differentially control pyroptotic cell death and mitochondrial damage in cancer cells

Gasdermin (GSDM)-mediated pyroptosis is functionally involved in multiple diseases, but Gasdermin-B (GSDMB) exhibit cell death-dependent and independent activities in several pathologies including cancer. When the GSDMB pore-forming N-terminal domain is released by Granzyme-A cleavage, it provokes cancer cell death, but uncleaved GSDMB promotes multiple pro-tumoral effects (invasion, metastasis, and drug resistance). To uncover the mechanisms of GSDMB pyroptosis, here we determined the GSDMB regions essential for cell death and described for the first time a differential role of the four translated GSDMB isoforms (GSDMB1-4, that differ in the alternative usage of exons 6-7) in this process. Accordingly, we here prove that exon 6 translation is essential for GSDMB mediated pyroptosis, and therefore, GSDMB isoforms lacking this exon (GSDMB1-2) cannot provoke cancer cell death. Consistently, in breast carcinomas the expression of GSDMB2, and not exon 6-containing variants (GSDMB3-4), associates with unfavourable clinical-pathological parameters. Mechanistically, we show that GSDMB N-terminal constructs containing exon-6 provoke cell membrane lysis and a concomitant mitochondrial damage. Moreover, we have identified specific residues within exon 6 and other regions of the N-terminal domain that are important for GSDMB-triggered cell death as well as for mitochondrial impairment. Additionally, we demonstrated that GSDMB cleavage by specific proteases (Granzyme-A, Neutrophil Elastase and caspases) have different effects on pyroptosis regulation. Thus, immunocyte-derived Granzyme-A can cleave all GSDMB isoforms, but in only those containing exon 6, this processing results in pyroptosis induction. By contrast, the cleavage of GSDMB isoforms by Neutrophil Elastase or caspases produces short N-terminal fragments with no cytotoxic activity, thus suggesting that these proteases act as inhibitory mechanisms of pyroptosis. Summarizing, our results have important implications for understanding the complex roles of GSDMB isoforms in cancer or other pathologies and for the future design of GSDMB-targeted therapies.


Mitochondria purification from cell lines
Mitochondrial isolation method was done following the protocol detailed in (2). Briefly, 9 x 10 6 HEK293T cells were culture in 150 mm plates at 80-90% confluency and transfected with different GSDMB constructs during 48h. Cells were harvested by cell scrapper, pelleted at 600g for 5 min and washed twice in PBS. The cell pellets were placed on ice and broken by adding one volume of hypotonic homogenization buffer (IB 0.1x: 3.5 mM Tris-HCl, pH 7.8, 2.5 mM NaCl, 0.5 mM MgCl2), and homogenized by 10 strokes using a Thomas homogenizer with a motordriven Teflon pestle. Immediately after, 1/10 of the packed cell volume of hypertonic buffer was added to make the medium isotonic. Homogenate was centrifuged at 1200g for 3 min at 4 °C to pellet unbroken cells, debris, and nuclei. Supernatant was collected and centrifuged again at low speed in the same conditions (1200g for 3 min at 4 °C). Mitochondria contained on the supernatant were pelleted in Eppendorf tubes (adding approximately 1 ml of supernatant per tube) by centrifugation in microfuge at 15 000 g during 2 min at 4°C. Pellets were washed using homogenization buffer A (0.32 M sucrose, 1 mM EDTA, and 10 mM Tris-HCl, pH 7.4) and again, resuspended in the appropriate buffer and kept at 4 ºC until used.

Mitochondrial DNA release assay
Human mitochondrial DNA was isolated from the cytosolic fraction of transfected HEK293T with different GSDMB constructs using the DNeasy Blood &Tissue Kit (QIAGEN). The standard of mitochondrial DNA was obtained from mitochondria isolated from HEK293T cells using the protocol previously described in (2). Mitochondrial DNA standard curve was obtained in each assay for their absolute quantification. Quantitative PCR was employed to measure mitochondrial

Pyroptosis and apoptosis assays in THP1 cells
The human non-adherent monocyte-like cell line THP1 were infected with lentiviral particles containing the following myc-tagged GSDMB plasmid constructs: 1-416, 1-416D6,7 or empty pLVX plasmid (as control). Stable expressing cells were maintained in the presence of the selection antibiotic, puromycin (12.5 μg/μL). For inducing canonical pyroptosis, 1x10 7 cells were cultured in P100 plates with 6 mL RMPI medium in the absence of puromycin. To differentiate THP1 cells into macrophages, 0.08 μL/mL of PMA (forbol-12-miristate-13-acetate; SIGMA) were added and incubated for 24 h. Then, medium was replaced and 1 μL/mL of LPS (Lipopolysaccharide, SIGMA). After 24 hours, medium was replaced with RPMI without phenol red and containing 3 μL/mL of Nigericin (Cayman) to induce pyroptotic cell death. Control cells did not have LPS nor Nigericin. After 4 hours, medium containing dead cells was collected and centrifuged 5 minutes at 1200 rpm. The cell pellet was combined with the cells adhered to the plate, for subsequent protein extraction, as described in "protein extraction method". The supernatant was stored at -80ºC. To quantify pyroptosis, the enzymatic activity of the released LDH (Lactate Dehydrogenase) was measured from 1 mL of the supernatant with the Cytotoxicity Detection KitPLUS (LDH) (Roche). For apoptosis induction, cells were treated with either 1 μL/mL of 10 μM etoposide (SIGMA) or DMSO, as control. After 24 hours, the supernatant containing dead cells was processed as described in the pyroptosis assays for Western blotting.

Neutrophil Elastase protease assay
Transient transfected cells were lysed in 0,1% Triton X-100 lysis buffer without protease inhibitors and subsequently sonicated twice during 30s (Soniprep 150). Lysates were centrifugated at 10 000 g for 10 min at 4°C and the supernatant was collected to determine the protein concentration by BCA assay (Thermo Fisher Scientific™). Cell lysates (12,5 ug for hNE) were mixed at indicated concentrations of recombinant human neutrophile elastase (Sigma), followed by incubation at 37°C for 1h. Elastase reactions were carried out in the lysis buffer containing 0,1% Triton X-100. Where indicated, BAY-678 inhibitor was incubated with the protease treatment at the same conditions. Reactions were stopped by adding 1x Leammli Buffer with DTT (25mM) and incubated for 5 min at 95°C prior to loading on SDS-PAGE.

Cell death Annexin/ PI detected by flow cytometry
Cell death analysis was performed using FITC Annexin V Apoptosis Detection Kit

Immunofluorescence, confocal microscopy and Cell Observer
Cells were seeded in 24-well cell culture plate (1,5 x 10 5 cell/well) and transfected for 72 hours. Immunofluorescence preparations were visualized in a confocal microscopy LSM710 (Zeiss) and images were processed by Fiji software (Image J 1.52).
To assess GSDMB intracellular localization in real time, HEK293T cells were transiently transfected with Doxycycline-inducible vectors expressing GFP-tagged GSDMB constructs.
Cells were pre-induced with Doxycycline at 200ng for 2h. Additionally, to assess the dynamics of cell death induction, cells were cultured in the presence of 0.2µg/ml Propidium Iodide. Live videos (1 frame every 10 minutes for at least 20h) were recorded with either Cell Observer Microscopy or LSM710 confocal microscopy (Zeiss). Videos were processed by Microscope Software ZEN lite (Zeiss) and Fiji software (Image J 1.52), respectively.

Correlative light and electron microscopy (CLEM)
For

Structural modeling
Structural model of the monomeric form of N-terminal of both 1-275 and 1-275D6 human GSDMB protein after proteolytic cleavage were obtained using procedures of homology modeling (4), using as templates the cryo-EM structures of the membrane pore formed by the Nterminal domains of murine Gasdermin A3, Protein Data Bank (PDB) accession number 6CB8 (5) and human Gasdermin D, PDB accession number 6VFE (6). The model of the dimer formed by two N-terminal domains was constructed using consecutive monomers belonging to these same membrane pore structures as a template.
Models were subjected to 100ns of unrestrained molecular dynamics (MD) simulations using the AMBER18 molecular dynamics package (http://ambermd.org/; University of California), essentially as previously described (7). 3D structures were solvated with a periodic octahedral pre-equilibrated solvent box using the LeaP module of AMBER, with 12 Å as the shortest distance between any atom in the protein subdomain and the periodic box boundaries. MD simulation was performed using the PMEMD program of AMBER18 and the ff14SB force field (http://ambermd.org/), applying the SHAKE algorithm, a time step of 2 femtoseconds (fs) and a non-bonded cut-off of 12Å. Systems were initially relaxed over 10,000 steps of energy minimization, using 1,000 steps of steepest descent minimization followed by 9,000 steps of conjugate-gradient minimization. Simulations were then started with 20 picoseconds (ps) heating phase, raising the temperature from 0 to 300 K in 10 temperature change steps, after each of which velocities were reassigned. During minimization and heating, the Cα trace dihedrals were

In-Gel Digestion and Reverse phase-liquid chromatography RP-LC-MS/MS analysis
After drying, gel bands or spots were distained in acetonitrile:water (ACN:H2O, 1:1), were reduced and alkylated (disulphide bonds from cysteinyl residues were reduced with 10 mM DTT for 30 min at 56 ºC, and then thiol groups were alkylated with 10 mM iodoacetamide for 30 min at room temperature in darkness) and digested in situ with sequencing grade trypsin (Promega) as described by (8) with minor modifications (9). The gel pieces were shrunk by removing all liquid using sufficient ACN. Acetonitrile was pipetted out and the gel pieces were dried in a speedvac. The dried gel pieces were re-swollen in 100 mM Tris-HCl pH 8, 10mM CaCl2 with 12.5 ng/μl trypsin for 1h in an ice-bath. The digestion buffer was removed, and gels were covered again with 100 mM Tris-HCl pH 8, 10mM CaCl2 and incubated for 12 h at 37°C. Digestion was stopped by the addition of 1% TFA. Whole supernatants were dried down and then desalted onto ZipTip C18 Pipette tips (Millipore) until the mass spectrometric analysis. The desalted protein digest was dried, resuspended in 10 µl of 0.1% formic acid and analyzed by RP-LC-MS/MS in an Easy-nLC II system coupled to an ion trap LTQ-Orbitrap-Velos-Pro hybrid mass spectrometer (Thermo Fisher Scientific™). The peptides were concentrated (on-line) by reverse phase chromatography using a 0.1mm × 20 mm C18 RP precolumn (Thermo Fisher Scientific™), and then separated using a 0.075mm x 250 mm C18 RP column (Thermo Fisher Scientific™) operating at 0.3 μl/min. Peptides were eluted using a 60-min dual gradient. The gradient profile was set as follows: 5−25% solvent B for 68 min, 25−40% solvent B for 22 min, 40−100% solvent B for 2min and 100% solvent B for 18 min (Solvent A: 0,1% formic acid in water, solvent B: 0,1% formic acid, 80% acetonitrile in water). ESI ionization was done using a Nano-bore emitters Stainless Steel ID 30 μm (Proxeon) interface at 2.1 kV spray voltage with S-Lens of 60%. The Orbitrap resolution was set at 30 000 (10). Using an anti-GSDMB-CT antibody Ab-GB (11) a CT fragment (asterisk) of 35 KDa (1-275D6 ,7) and 37 KDa (1-416) was detected in both pyroptosis and apoptosis assays. For comparison, GSDMD showed a differential cleavage pattern between apoptosis (p43 CT fragment) and pyroptosis (p30 NT fragment and secondary p43 CT fragment), as reported before (12). Procaspase 1 and activated caspase-1 were detected in the supernatant from pyroptotic cells (A). Bars represents mean values ± SEM from three independent experiments. Differences between control (empty vector) and each condition was tested by two-tailed unpaired t-test: **p<0,01 and ***p<0,001. GSDMD-NT was used as a positive control. B) Immunofluorescence and confocal microscopy analysis in 23132/87 cells transiently transfected with indicated GSDMB-NT constructs. GSDMB-NT (green; NT antibody SIGMA, HPA023925), co-localizes with mitochondrial marker TOM20 in red. pLVX (empty vector) was used as a negative control. C)

SUPPLEMENTARY FIGURES
Immunoblotting analysis of GSDMB cleavage by recombinant human neutrophil elastase (rhNE).

A) GSDMB sequence and cleavage sites representation indicating the exon localization. B)
Schematic representation of GSDMB cleavage by GZMA, NE and caspases.

SUPPLEMENTARY TABLES:
Supplementary