Antibodies for GSDME (ab215191), GSDMD (ab209845) were purchased from Abcam (Cambridge, Cambridgeshire, Britain). Antibodies for caspase-1 (3866T), caspase-9 (9502T), caspase-3 (9668T), caspase-8 (9746T) and γH2AX (9718T) were purchased from Cell Signaling Technology (Danver, MA, USA), and GAPDH (60004-1-lg) was from Proteintech Group (Wuhan, Hubei, China). The secondary antibodies including HRP Goat Anti-Mouse IgG (H+L) (AS003), HRP Goat Anti-Rabbit IgG (H+L) (AS014) and FITC Goat Anti-Rabbit IgG (H+L) (AS011) were bought from Abclonal (Wuhan, Hubei, China).

Other reagents were purchased as follows: caspase-3 inhibitor (Z-DEVD-FMK) (HY-12466) and ROS scavenger N-acetyl cysteine (NAC) (HY-B0215) were from MedChemExpress (Monmouth Junction, NJ, USA); protease inhibitor cocktail tablets (BL612A) were from Biosharp (Beijing, China).

Human colorectal cancer cell line HCT116 cells were obtained from Shanghai Institute of Cell Biology, Chinese Academy of Sciences. Human colonic epithelial cell line FHC cells were purchased from Kunming Cell Bank, Kunming Institute of Zoology. HCT116 and FHC cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM)-high glucose supplemented with 10% fetal bovine serum (FBS) (Gibco, Grand Island, NY, USA) and 1% penicillin-streptomycin (Gibco, Grand Island, NY, USA). The cells were incubated at 37°C incubator containing 5% CO_2.

For CDT treatments, cells were pre-seeded in 12-well plates overnight until the cells reached 50% confluency. To inhibit the activity of caspase-3 or ROS, cells were pre-treated with Z-DEVD-FMK for 2 h or NAC for 3 h, respectively. Unless otherwise specified, the final action concentrations were: 10 μg/ml for CDT, 40 μM for Z-DEVD-FMK and 5 μM for NAC.

Both cells and supernatants were collected for Western blot. After PBS washing twice, the cell sediments were lysed with radio-immuno-precipitation assay (RIPA) buffer (Beyotime, Shanghai, China) with protease inhibitor cocktail and the protein concentrations were measured by Bicinchoninic Acid (BCA) Protein Assay Kit (Biosharp, Beijing, China) finally. Total 40 μg of protein from each sample was separated by SDS-PAGE and transferred to a PVDF membrane followed by blocking with 5% non-fat dry milk for 1 h at room temperature. The membrane was then incubated with primary antibodies at 4°C overnight, followed by incubated with HRP-conjugated secondary antibodies for 1 h at room temperature. The protein bands were identified using Clarity Max Enhanced Chemiluminescence (ECL) Western Blotting Substrates (Bio-Rad, Hercules, CA, USA).

For detecting cell viability, cells were pre-seeded into 96-well plates (1.2 × 10⁴ cells/well) and treated with CDT for 24 h. Cell viability was measured using Cell Counting Kit-8 (CCK-8) assay (Biosharp, Beijing, China) according to the manufacturer’s protocol.

The release rate of lactate dehydrogenase (LDH) was examined using LDH Cytotoxicity Assay Kit (Beyotime, Shanghai, China) according to the manufacturer’s protocol. The absorbance at 490 nm was detected after incubation. All experiments were repeated three times.

The intracellular ROS levels were detected by ROS Assay Kit (Beyotime, Shanghai, China). After CDT exposure for 24 h, cells were incubated in the medium containing DCFH-DA (10 μM) at 37°C for 30 min in dark, followed by washing three times with non-FBS medium. The ROS levels were measured by a flow cytometer (BD Bioscience, Bedford, MA, USA).

Cells were pre-seeded into 12-well plates (2.5 × 10⁵ cells/well) and treated with CDT for 24 h. CDT-induced cell death was detected with annexin V-fluorescein isothiocyanate (FITC) and propidium iodide (PI) Apoptosis Detection Kit (Yeasen, Shanghai, China). After treatment, cells were harvested, washed with PBS and incubated in annexin V and PI for 10 min at room temperature in dark. 1.0 × 10⁴ cells were collected and detected with flow cytometer (FACSVerse, Bioscience, CA, USA), and data was analyzed by FlowJo software. Before the analysis, the gating was performed to exclude doublets and debris. Most of cells should be gated, and greater than 90% of surviving cells in the control group should be ensured, while dividing four rectangular gates. All assays were repeated three times.

Recombinant CdtA, CdtB and CdtC from C. jejuni were overexpressed respectively as a fusion protein with histidine-tag (His-tag) in E. coli BL21 (DE3). The E. coli BL21 (DE3) containing expression plasmid was cultured and induced by 0.8 mM isopropyl-β-D-thiogalactopyranoside (IPTG) at 18°C for 20 h until an OD₆₀₀ of 0.5–0.6. The cells were lysed, and recombinant proteins were purified via Ni-NTA affinity chromatography (GE healthcare, Marlborough, MA, USA) followed by SDS-PAGE and Western blotting. Three subunits were reconstituted as a holotoxin at 25°C for 1 h for subsequent research.

The GSDME-KO cell lines were generated using CRISPR-Cas9 technology. In brief, the guide RNA (gRNA), 5ʹ-TAAGTTACAGCTTCTAAGTC-3ʹ to target GSDME was subcloned into lentiCRISPR v2 vector. The lentiviruses were collected after co-transfection of HEK293T cells with lentiviral vector and packaging plasmids for 48 h and then used to infect HCT116 cells. The infected HCT116 cells were selected with puromycin (1 μg/ml) for 3 days and the single clones were cultured in 96-well plates for another 10–14 days or longer. Anti-GSDME immunoblotting and sequencing were used to identify GSDME-KO cell lines.

For siRNA knockdown, 80 pmol siRNA was transfected into HCT116 cells using jetPRIME^® transfection reagent (Polyplus, Paris, France) following the manufacturer’s instruction. Cells were used for assay 48 h after transfection.

DNA damage in single cells was detected using comet assay according to the article (Lu et al., 2017). Briefly, after the formation of agarose coat on the slide, cell suspension at a concentration of 2.0 × 10⁵/ml was mixed with 1% low-melting agarose at a ratio1:10 (v/v), and the mixture was immediately pipetted onto the slide. Following a solidification at 4°C for 30 min and lysis in lysis solution at 4°C for 1 h, the slides were transferred to alkaline electrophoresis solution for 1 h to allow DNA unwinding. Then, alkaline electrophoresis was performed for 30 min at 26 V and 300 mA. Slides were washed twice in dH₂O and one time in 70% EtOH for 5 min each at room temperature, dried at 37°C for 15 min, and stained with 100 μl SYBR Green for 15 min. Comet images were taken using a fluorescence microscope (Olympus, Tokyo, Japan) and the images of comet assay from the experiment with at least 50 cells each group was analyzed with OpenComet plug-in in ImageJ. The tail moments are calculated as follow: TM=Tail length × Tail% DNA/100.

To examine the morphological characteristics of pyroptosis, cells were pre-seeded into 12-well plates overnight until reaching approximately 50% confluency. After CDT exposure, static bright field images of cells were captured using Ti-U microscope (Nikon, Tokyo, Japan).

To examine the γH2AX foci, cells were pre-seeded into 12-well plates overnight and subjected to different treatments. Afterward, the cells were incubated with 4% paraformaldehyde for 30 minutes, 0.5% Triton X-100 in PBS for 15 min, and blocked with 5% non-fat dry milk for 1 h at room temperature. The cells were then incubated with primary antibodies at 4°C overnight, followed by incubated with FITC-conjugated secondary antibodies for 1 h and DAPI for 15 min at room temperature. The fluorescence intensity was detected with the laser confocal microscope (Nikon, Tokyo, Japan).

Data were presented as mean ± standard error of the mean (SEM) from at least three experiments and were compared by one-way ANOVA followed by Tukey’s multiple comparison. Statistical analyses were performed using GraphPad Prism version 7.0 (CA, USA). *P < 0.05 was considered statistically significant, **P < 0.01 and ***P < 0.001 indicated extremely significant differences.

To investigate the cytotoxicity of C. jejuni CDT on human colonic epithelial cells, a CDT-deficient mutant of C. jejuni strain NCTC11168 was engineered via electroporation of a gene targeting vector into the NCTC11168 wild type (WT) strain, in which the consecutive genes cdtA, cdtB and cdtC, were replaced by kanR. Later, we prepared bacterial lysates from C. jejuni WT (C. jejuni) and mutant strain (C. jejuni mut), and stimulated human colonic epithelial cell line FHC and human colon cancer cell line HCT116, with these extracts (10 μg/ml) for 12 hours or recombinant protein CdtA, CdtB, CdtC and holotoxin CDT (CjCDT) (10 μg/ml) for 24 hours. As shown in Figure 1A, CjCDT and extracts from C. jejuni WT strain treatments increased the number of γH2AX foci, a marker for DNA damage signaling, in both HCT116 and FHC cells, when compared with untreated cells. Conversely, extracts generated from C. jejuni mutant strain and the single subunits lacked a notable effect on augmenting γH2AX level (Figure 1A). Evaluation of DNA damage via comet assay revealed an increase in DNA damage in HCT116 and FHC cells exposed to CjCDT and extracts from C. jejuni WT strain when compared with cells exposed to C. jejuni mutant extracts and the single subunits (Figure 1B). Moreover, CCK-8 assay was performed to further evaluate the effect of CDT on cytotoxicity in human colonic epithelial cells. We found that exposure of cells to CjCDT attenuated the cell viability in a concentration- and time- dependent manner (Figures 1C, D). Interestingly, the cells primed with CdtA, CdtB or CdtC protein alone did not exhibit significant decline in cell viability unless with CDT holotoxin, indicating that the catalytic subunit CdtB requires the assistance of CdtA and CdtC to perform a cytotoxic function (Figure 1E). Collectively, these data strongly support the notion that CDT induces cytotoxicity.

Firstly, followed by the exposure of CDT, the morphological changes were observed. Morphologically, HCT116 and FHC cells exhibited microscopic features of cell swelling and balloon-like bubbles emerging from plasma membrane following CDT infection, which were different from the shrinkage of apoptotic cells (Figure 2A).The previous report found that pyroptosis is characterized by large bubbles blown from the plasma membrane, increase of LDH release and annexin V and PI positive cells (Rogers et al., 2017). It is plausible to postulate that pyroptosis is involved in CDT-mediated cytotoxicity. Afterward, HCT116 and FHC cells were stimulated with different concentrations (0, 1, 5 and 10 μg/ml) of CDT for 24 h or with a concentration of 10 μg/ml CDT for different periods of time (0, 6, 12 and 24 h). LDH release assay indicated that CDT increased LDH release in a dose- and time- dependent manner, suggesting the cell membrane rupture (Figures 2B, C). Moreover, flow cytometry analysis revealed that CDT treatment remarkably augmented the proportion of annexin V and PI double positive cells in a dose- and time- dependent manner, indicating the cell death through triggering plasma membrane permeabilization (Figures 2D, E). Hence, these results demonstrate that CDT changes cellular morphology and induces cell death in human colonic epithelial cells through triggering plasma membrane permeabilization.

Next, we moved on to explore whether pyroptosis is involved in CDT-mediated cytotoxicity and the mechanism by which CDT induces pyroptosis. First, to investigate whether GSDMD is involved in CDT-induced cytotoxicity, full-length GSDMD and cleaved N-terminus of GSDMD were detected using immunoblotting. We found that although GSDMD was expressed in HCT116 cells, no GSDMD cleavage was observed upon CDT treatment (Figure 3A). To further determine whether GSDMD participates in CDT-induced cytotoxicity, GSDMD was stably knocked down in HCT116 cells by transfecting siRNA targeting GSDMD (Figure 3B). Cell viability and LDH release assay revealed no obvious difference among the groups transfected with control siRNA or GSDMD targeting siRNA followed by CDT treatment, demonstrating that GSDMD is not involved in CDT-induced pyroptosis in HCT116 cells (Figures 3C, D). Given that GSDME is a newly recognized executor of cell pyroptosis, we postulated that GSDME can trigger CDT-induced pyroptosis. As expected, we found that CDT treatment (CjCDT and bacterial lysates from C. jejuni WT strain) led to elevated levels of N-terminus of GSDME in a dose- and time- dependent manner in HCT116 and FHC cells (Figures 3E, F). To further confirm this hypothesis, we knocked out the GSDME gene in HCT116 cells (Figure 3G). Knockout of GSDME resulted in significantly lessened LDH release and microscopic feature of large bubbles (Figures 3H, I), compared with negative control (NC) cells in response to CDT. Flow cytometry analysis revealed that GSDME knockout noticeably decreased the percentage of annexin V and PI double positive of HCT116 cells following CDT treatment (Figure 3J). These results suggest that lack of GSDME diminished CDT-induced pyroptosis. In conclusion, these results demonstrate that GSDME is a key executive protein in CDT-mediated pyroptosis.

We next sought to explore the potential molecular mechanism of GSDME-mediated pyroptosis induced by CDT. To this end, firstly, HCT116 cells were exposed to CDT in the absence or presence of the caspase-3 specific inhibitor, Z-DEVD-FMK, and incubated for 24 h. A notable attenuation was observed in terms of GSDME cleavage (Figure 4A), LDH release (Figure 4B), microscopic feature of large bubbles (Figure 4C) and cell death (Figure 4D) in combined treatment group, compared with the cells stimulated with CDT separately, demonstrating the role of caspase-3 in the process. It is well known that caspase-3 is an executor caspase that can be activated either by caspase-9 via mitochondrial pathway or caspase-8 via death receptor pathway (Chen et al., 2006). In the present study, we revealed that caspase-8 was not activated after HCT116 and FHC cells were exposed to CjCDT (Figure 4E). Besides, no evident activation of pro-inflammatory caspase, caspase-1, was found as well, indicating that CDT might not lead to pro-inflammatory effect in HCT116 and FHC cells directly (Figure 4E). Next, the siRNA technology was employed to knock down the expression of caspase-3 and caspase-9 for further investigation. Knocking down either caspase-3 or caspase-9 efficiently suppressed the cleavage of GSDME (Table 1). Additionally, caspase-9 knock-down severely compromised the activation of caspase-3, whereas caspase-3 knock-down lacked a notable inhibition on caspase-9 activation (Figures 4F, G). Together, these results demonstrate that caspase-9/caspase-3/GSDME axis is associated with CDT-induced pyroptosis.

Excessive accumulation of ROS could increase cellular oxidative stress, leading to several types of cell death (Zhou et al., 2018). However, whether ROS is related to CDT-induced pyroptosis remains currently unclear. First, we performed flow cytometry to detect the level of intracellular ROS following CDT treatment. Indeed, we found that CDT exposure significantly increased cellular ROS levels reflected by the fluorescence of DCFH-DA. Of note, the effect was impeded by NAC, a scavenger of ROS (Figures 5A, B). Then, we asked whether cellular ROS may influence CDT-induced cleavage of caspase-9/caspase-3/GSDME in HCT116 cells. As expected, the cells pretreated with NAC followed by CDT stimulation exhibited the attenuation of LDH release (Figure 5C), GSDME cleavage, caspase-3 and caspase-9 activation (Figure 5D). Moreover, CDT-induced pyroptotic morphology and cell death were rescued by NAC incubation (Figures 5E, F). These results indicate that the removal of ROS efficiently suppressed CDT-triggered pyroptosis. Hence, our results suggest that ROS, serving as an upstream of caspase-9/caspase-3/GSDME pathway, plays an essential role in CDT-mediated pyroptosis.