Contribution of TMEM16F to pyroptotic cell death

Pyroptosis is a highly inflammatory form of programmed cell death that is caused by infection with intracellular pathogens and activation of canonical or noncanonical inflammasomes. The purinergic receptor P2X7 is activated by the noncanonical inflammasome and contributes essentially to pyroptotic cell death. The Ca2+ activated phospholipid scramblase and ion channel TMEM16F has been shown earlier to control cellular effects downstream of purinergic P2X7 receptors that ultimately lead to cell death. As pyroptotic cell death is accompanied by an increases in intracellular Ca2+, we asked whether TMEM16F is activated during pyroptosis. The N-terminal cleavage product of gasdermin D (GD-N) is an executioner of pyroptosis by forming large plasma membrane pores. Expression of GD-N enhanced basal Ca2+ levels and induced cell death. We observed that GD-N induced cell death in HEK293 and HAP1 cells, which was depending on expression of endogenous TMEM16F. GD-N activated large whole cell currents that were suppressed by knockdown or inhibition of TMEM16F. The results suggest that whole cell currents induced by the pore forming domain of gasdermin-D, are at least in part due to activation of TMEM16F. Knockdown of other TMEM16 paralogues expressed in HAP1 cells suggest TMEM16F as a crucial element during pyroptosis and excluded a role of other TMEM16 proteins. Thus TMEM16F supports pyroptosis and other forms of inflammatory cell death such as ferroptosis. Its potent inhibition by tannic acid may be part of the anti-inflammatory effects of flavonoids.


Introduction
Intracellular Ca 2+ is enhanced during many biological processes including inflammation. Ca 2+ mobilization is suggested to have a role in the regulation of NLRP3 (NOD, LRR, and pyrin domain-containing 3) inflammasome, a large supramolecular complex that activates caspase-1 during pyroptosis. Pyroptosis, a highly inflammatory form of programmed cell death, occurs upon infection with intracellular pathogens and is part of the antimicrobial response. In contrast to apoptosis, pyroptotic cell death results in plasma membrane (PM) rupture and release of so called damage-associated molecular pattern (DAMP) molecules 1 . Inflammasomes activate caspase-1 or caspase 11/4/5, which cleave the pore-forming N-terminal part of gasdermin D that drives the cell into lytic cell death [2][3][4] . Large gasdermin D pores are regarded as effectors of pyroptosis. These pores may lead to an increase in intracellular Ca 2+ by permeabilizing the plasma membrane and probably also intracellular membranes. Moreover, noncanonical inflammasomes lead to caspase-11-dependent pyroptosis due to activation of pannexin-1, release of ATP binding to purinergic P2X 7 receptors and consecutively increases intracellular Ca 2+ 5 . Notably the Ca 2+ activated phospholipid scramblase and ion channel TMEM16F has been shown to participate in the cellular effects downstream of P2X 7 receptors that finally lead to cell death 6 .
TMEM16F belongs to a family of 10 proteins (TMEM16A-K; anoctamin 1-10) 7 . These proteins are localized in the plasma membrane or in intracellular membrane compartments. Apart from TMEM16A and B, which are Ca 2+ activated chloride channels without scrambling activity, other TMEM16 proteins expose phosphatidylserine to the outer plasma membrane leaflet and conduct ions when activated by an increase in intracellular Ca 2+ 8-14 . Evidence has been provided that TMEM16F (i) participates in cell shrinkage and presumably apoptotic cell death [15][16][17] , (ii) forms an outwardly rectifying Cl − channel (ORCC) that is activated during death of immune cells 6,18,19 , and (iii) is activated during other forms of programmed cell death such as necroptosis and ferroptosis 20,21 . In the present study we asked whether TMEM16F is also activated during pyroptosis and, if so, whether it contributes to pyroptotic cell death.

TMEM16F supports gasdermin D-induced cell death
In order to examine cell death induced by gasdermin D we expressed the amino-terminal pore-forming domain of gasdermin D (GD-N) in HEK293 cells. Cells were examined by flow cytometry after 24 h of expression, which indicated a high percentage of death, i.e., 7-AADpositive cells, when compared to mock transfected cells (Fig. 1a, b). Interestingly, when GD-N-transfected cells were grown in the presence of the TMEM16F-inhibitor tannic acid (TA), the cell death-inducing effect of GD-N was completely abolished, suggesting that TMEM16F contributes to GD-N induced cell death. LDH-release was assessed after 24 h expression of full − length gasdermin (GD) and GD-N. While GD expressing cells showed only a small increase in LDH release, LDH release by GD-N expressing cells was remarkable, and was significantly inhibited by three different inhibitors of TMEM16F, CaCCinhAO1 (AO1), TA or niflumic acid (NFA) (Fig. 1c). Moreover, knockdown of TMEM16F, expressed endogenously in HEK293 cells, suppressed cell death induced by GD and GD-N (Fig. 1d, f). Expression of full − length gasdermin D (GD) and N-terminal fragment of gasdermin D (GD-N) was demonstrated by immunocytochemistry using gasdermin D antibody. While GD was found to be distributed homogenously throughout the cytosol, GD-N was localized as spots in the plasma membrane (Fig. 1e). Finally, GD-N induced LDH release was reduced in Scott B-lymphocytes (Scott-BL), which lack of expression of TMEM16F 19 , when compared to wt B-lymphocytes expressing TMEM16F (Scott-BL) (Fig. 1g, h). Taken together the data strongly suggest support of gasdermin D-induced cell death by TMEM16F.
Increase in intracellular Ca2 + and activation of whole cell Cl − currents by expression of pore forming gasdermin D As the N-terminal pore-forming domain of gasdermin D (GD-N) is likely to lead to an increase in intracellular Ca 2+ , we examined intracellular Ca 2+ levels in cells expressing GD-N or in mock transfected cells, using the Ca 2+ sensitive dye Fura2. The data indicate that baseline Ca 2+ levels were enhanced in cells expressing GD-N.
Interestingly, increase in intracellular Ca 2+ in GD-N expressing cells was completely inhibited by tannic acid (Fig. 2a, b). This suggests that TMEM16 proteins, which are inhibited by tannic acid, have a role in GD-N induced Ca 2+ increase. In fact, an earlier report demonstrated Ca 2 + permeability of TMEM16F 14 . Because GD-N increased intracellular Ca 2+ , we examined ion currents activated during pyroptotic cell death (Fig. 2c, d). In whole cell patch clamp experiments we detected a whole cell current in HEK293 cells expressing GD-N, but not in cells expressing full length GD or in mock transfected cells (Fig. 2d, f). GD-N induced whole cell currents were outwardly rectifying and showed no consistent time dependence (Figs. 2d, e, 3). The enhanced whole cell currents found in GD-N expressing cells were potently inhibited by removal of extracellular Cl − , indicating a permeability of the current for Cl − (Fig. 2e). The GD-N induced currents were also inhibited by the TMEM16F inhibitors CaC-CinhAO1 (AO1) and tannic acid (TA; Fig. 3a, b). Moreover, knockdown of endogenous TMEM16F significantly inhibited GD-N induced whole cell currents (Fig. 3c, d).
The results therefore suggest that whole cell currents induced by the pore forming domain of gasdermin-D are at least partially due to activation of TMEM16F.
Knockdown of TMEM16F in HAP1 cells inhibits Ca 2 + -activated exposure of phosphatidylserine and GD-N induced cell death We examined TMEM16F and its role for gasdermininduced cell death in the haploid leukemia cells line HAP1 22 . This cell line was chosen because it was also available as TMEM16F-knockout cell line (horizon, Cambridge, UK). Moreover, it was straightforward to knockout additional TMEM16 paralogues expressed in HAP1 cells by CRISPR/Cas9, and to examine their potential contribution to gasdermin-induced cell death. RT-PCR analysis indicated expression of TMEM16D,F,H, K in HAP1 parental cells (Fig. 4a). Expression of these TMEM16 paralogues in HAP1 parental cells was confirmed by Western blotting (Fig. 4b). No expression of TMEM16F was detected in TMEM16F-knockout HAP1 cells (KO-T16F). Moreover, CRISPR/Cas9 knockout of the remaining TMEM16 paralogues D,H,K (KO-T16all) was also confirmed by Western blotting (Fig. 4b).
TMEM16A was not detected in HAP1 cells. FACS analysis demonstrated Ca 2+ -induced exposure of phosphatidylserine (phospholipid scrambling) in parental cells, which was absent in KO-T16F and KO-T16all cells (Fig. 4c, d). Knockout of TMEM16F or TMEM116D,F,H, K progressively attenuated cell proliferation (Fig. 4e). An anti-proliferative effect of TMEM16F knockout has also been observed in another study 23 . Moreover, similar to the study by Schenk et al, we observed slightly enhanced phospholipid scrambling after knockdown of TMEM16F and other TMEM16 proteins (Fig. 4d). The scramblase Xkr8 24 was also found to be expressed in HAP1 cells (in semiquantitative (sq) RT-PCR and Western blots, not shown). Interestingly, by sqRT-PCR we found an upregulation of Xkr8 expression relative to GAPDH from 0.49 ± 0.031 (parental) to 0.86 ± 0.058 (KO_16F) and 1.1 ± 0,37 (KO_T16all; all n = 3). This may explain the lower basal scrambling activity in TMEM16F expressing cells. This was further supported by experiments in which we knocked down expression of Xkr8 by siRNA (by 95.1 %, parental; 96.2%, KO_T16F; 66.9%, KO_T16all), which significantly reduced basal scrambling in the three different cell lines. Finally, we reported earlier that TMEM16 proteins have a significant impact on cytosolic Ca 2+ levels, which may affect cell proliferation and basal scrambling activity 25 . In fact, knockdown of TMEM16K had a pronounced inhibitory effect on TNFα-induced cell death 26 .
Knockout of TMEM16F or all TMEM16 proteins did not affect basal Ca 2+ levels (Fig. 5a). We examined enhanced intracellular Ca 2+ concentrations by stimulation with ATP (100 µM), or cyclopiazonic acid (CPA, 10 µM) in the presence of a Ca 2+ free extracellular bath solution. ATP-induced rise in intracellular Ca 2+ was significantly reduced in KO_T16F and KO_T16all cells, corresponding to earlier observations in other cell types 25 (Fig. 5b). Using HAP1 cells, we examined the role of TMEM16F for gasdermin-induced cell death. Cell death induced by expression of GD-N was twice as frequent in parental cells when compared to cells lacking expression of TMEM16F (KO-T16F). Additional knockdown of TMEM16D,H,K did (KO_T16all) did not further reduce GD-N induced cell death (Fig. 5c). Single additional knockdown of TMEM16D,H,K did not show additional effects. Similar to GD-N transfected HEK293 cells, we also observed enhanced baseline Ca 2+ levels in GD-N expressing HAP1 cells, which was potently suppressed by tannic acid (Fig. 5d). We also examined other forms of programmed cell death, such as ferroptosis (using the cystine import inhibitor erastin and the GPX4 inhibitor RSL3) or apoptosis (by incubation with TNFα). Both ferroptosis and apoptosis were significantly attenuated in cells lacking expression of TMEM16F (Fig. 5e, f). Analysis of cell morphology by quantitative holographic phase microscopy demonstrated the impressive attenuation of apoptosis in cells lacking expression of TMEM16F (Fig. 5g). These results are in line with earlier observations that suggested an impact of TMEM16F on apoptotic and ferroptotic cell death, respectively, and an activation of TMEM16F during necroptosis 6,15,16,20,21 . The present report now adds pyroptosis to this list, as another regulated cell death pathway that includes simultaneous activation of TMEM16F. Future studies should examine whether TMEM16F could serve as a drug target to control both inflammatory and non-inflammatory cell death.

Discussion
Increase in intracellular Ca 2+ by initial Ca 2+ store release and subsequent store operated Ca 2+ entry (SOCE) is essential to drive a number of processes that operate in immune cells, such as degranulation and cytotoxicity by CD8 + T cells and natural killer (NK) cells, production of reactive oxygen species in neutrophils via NADPH oxidase, and phagocytosis by macrophages 27 . We demonstrated earlier that Ca 2+ activated TMEM16F mediates a number of functions in macrophages, downstream of P2X 7 receptors 6 . Initial cell shrinkage and subsequent cell swelling was related to progressive activation of TMEM16F. Gasdermin D binds to phosphatidylinositol species and phosphatidylserine and forms large nonselective pores 4 . Expression of TMEM16F is particularly high in macrophages, where it supports phagocytic activity and cell death. The present data now imply that TMEM16F may have also a significant impact on pyroptotic cell lysis and subsequent phagocytosis by neutrophils. Consequently, TMEM16F is likely to affected inflammation. Notably, TMEM16F was found recently to be activated during ferroptotic cell death, induced by lipid peroxidation 21,28 . Taken together, TMEM16F contributes to different forms of regulated cell death such as apoptosis, ferroptosis and pyroptosis. It may therefore contribute to inflammation, hypersensitivity and generation of pain, and might represent a pharmacological target during inflammatory diseases".
Although it is emerging that TMEM16F and probably other TMEM16 paralogues have a role in programmed cell death, it is not clear whether they do so by operating as a channel/scramblase or by affecting intracellular Ca 2+ levels 6,15,16,29 . In fact, recent work suggests that TMEM16F and other TMEM16 proteins control intracellular Ca 2+ signals, thereby affecting numerous cellular functions 25,30 . TMEM16F may operate as a plasma membrane localized nonselective Ca 2+ influx channel 14 or as a passive Cl − /nonselective channel that counterbalances charge movements occurring during ER Ca 2+ store release or Ca 2+ influx 25 .
Inflammatory stimuli that enhance intracellular Ca 2+ levels or modify plasma membrane phospholipids also activate TMEM16F 15,21,28 . Thus PLA 2 not only acts as a central switch that turns on synthesis of numerous proinflammatory mediators 31 , but also serves as a powerful activator of TMEM16F 15 . Similarly, proinflammatory and lipid peroxidizing reactive oxygen species (ROS) are also powerful activators, and both PLA 2 and ROS do not require increase in cytosolic Ca 2+ to activate TMEM16F 15,21 . In contrast, TMEM16F did not contribute to necroptotic cell death, although it might be activated during necroptosis 20 . In contrast, the present data suggest a significant contribution of TMEM16 to pyroptosis. In this context is might be of interest that membrane stabilizing tannic acid and other flavonoids 32 are potent inhibitors of TMEM16F and strong antiinflammatory/ antioxidant molecules 15,[33][34][35] . Remarkably, tannic acid completely suppressed ATP-induced membrane blebbing in mouse macrophages 6 . Finally, TMEM16F is essential for activation of ADAM17 (Disintegrin and Metalloproteinase 17), which in turn is required for release of interleukin 6 receptors causing proinflammatory IL-6 trans-signaling 36,37 . The present data expand the proinflammatory role of TMEM16F to another highly inflammatory form of programmed cell death.

Methods
Generation of TMEM16-knockout cell lines by Crispr/Cas9, RT-PCR, and cell culture HAP1 cells were transfected with the CRISPR_CD4 plasmids obtained through ThermoFisher Scientific® (Schwerte, Germany) using Lipofectamine3000 (Life technologies, Darmstadt, Germany) according to the manufacturer's instructions (Table 1). Cells were collected and selected 3 days after transfection. Selected cells were expanded and genomic DNA was isolated to determine the efficiency of locus-specific double-strand break formation using GeneArt® Genomic Cleavage Detection (Invitrogen). HAP1 cells (Thermo Fisher scientific) were grown in Iscove's modified Dulbecco's medium (IMDM) supplemented with 10% fetal bovine serum at 37°C in 5% CO 2 . Total RNA was isolated from CRISPR cells using RNeasy Mini-Kit (Qiagen; Hilden, Germany). Two μg of total RNA was reverse transcribed in 50 μl buffer for 1 h at 40°C, using a random primer and M-MLV reverse transcriptase (Promega, Mannheim, Germany). Thirty cycles of RT-PCR was performed using standard procedures (GoTaq DNA Polymerase, Promega), 1 μl RT using primers for TMEM16 proteins described earlier 38 (Table 2). PCR products were analyzed on 2% agarose gels.

Cell proliferation assay
Cells were plated in 96-well plates at a density of 8 × 10 3 cells per well for the time duration as indicated (0, 24, 48, 72, 96 and 120 h). At the end of incubation, cells were washed with 150 μl of phosphate buffered solution (D-PBS Ca2+/Mg2+ ). Afterwards cells were incubated for 2 h in 100 μl of fresh media containing 0.5 mg/ml of the tetrazolium salt MTT. The dark blue formazan product was dissolved with DMSO and measured the absorbance at 540 nm.

Volume measurements using HoloMonitor TM
Cells were seeded in 35 mm dishes at a density of 300,000 cells/ dish. Cell morphology was observed after application of 10 µM ABT737 in IMDM medium containing 10% FBS or 100 ng/ml TNFα in OptiMEM for 24 h in a cell culture incubator (37°C, humidified air, 5% CO 2 ). The cell volume was calculated by quantitative phase microscopy in the HoloMonitor TM time-lapse cytometer (Phase Holographic imaging PHI, Lund, Sweden).

Flow cytometry
Cells were grown in 24-well plates for 24 h and were incubated afterwards with 10 μM ABT737 in IMDM medium containing 10% FBS or 100 ng/ml TNFα in OptimMEM medium for 24 h. Floating cells were collected initially, whereas adherent cells were collected by treatment with Accutase and centrifuged 500 × g 4°C for 10 min. Cell pellets were washed with cold DPBS and stained with FITC-labeled annexin V and 7-AAD for 10 min at room temperature in the dark. At least 10,000 cells were analyzed immediately with BD Accuri™ C6 flow cytometer.

Caspase assay
Cells were plated in 24-well plates for 24 h and then incubated with 10 μM ABT737 in IMDM medium containing 10% FBS or 100 ng/ml TNFα in OptiMEM for 24 h. Floating cells were collected initially, whereas adherent cells were collected by treatment with Accutase and centrifuged 500 × g 4°C for 10 min. Cell pellets were washed with cold DPBS and stained with 5 μM DEVD-NucView488 substrate for 30 min at room temperature in the dark. At least 10,000 cells were analyzed immediately with a BD Accuri™ C6 flow cytometer.

Patch clamping
Cells were grown on coated glass cover slips. Patch pipettes were filled with a cytosolic-like (standard) solution containing KCl 30, K -gluconate 95, NaH 2 PO 4 1.2, Na 2 HPO 4 4.8, EGTA 1, Ca -gluconate 0.758, MgCl 2 1.03, D-glucose 5, ATP 3, pH 7.2. In all experiments the bath was perfused for at least 5 min with control Ringer solution after establishing the whole cell configuration, in order to assure stable recording conditions. The intracellular (pipette) Ca 2+ activity was 0.1 µM. In some experiments, all except of 5 mM Cl − was replaced by impermeable gluconate (5Cl − ). Fast whole cell current recordings were performed as described recently 18 . In brief, the bath was perfused continuously with Ringer solution (containing (in mM) 145 NaCl, 0.4 KH2PO4, 1.6 K2HPO, 4.6 D-glucose, 1 MgCl2 1.3 Ca 2+ gluconate; pH 7.4) at a rate of 8 ml/min. Patch pipettes had an input resistance of 2-4 MΩ and measured whole cell currents were corrected for serial resistance. Currents were recorded using a patch clamp amplifier (EPC 7, List Medical Electronics, Darmstadt, Germany), the LIH1600 interface and PULSE software (HEKA, Lambrecht, Germany) as well as Chart software (AD Instruments, Spechbach, Germany). In regular intervals, membrane voltage (Vc) was clamped in steps of 20 mV from −100 to +100 mV from a holding voltage of −100 mV. Current densities at +100 mV clamp voltage were calculated by dividing the measured whole cell currents by cell capacitance.

Materials and statistical analysis
All compounds used were of highest available grade of purity and were from Sigma-Aldrich (Germany), Tocris Bioscience (Bristol, United Kingdom), or Merck (Darmstadt, Germany). Cell culture reagents were from Invitrogen and Capricorn Scientific (Ebsdorfergrund, Germany). FACS reagents were from BD biosciences and BioLegend GmbH (Koblenz, Germany). Data are presented as Mean±SEM. Student's t-test for paired or unpaired samples or Fisher's exact test and analysis of variance were used, for statistical analysis. P < 0.05 was accepted as significant difference.