The colonic pathogen Entamoeba histolytica activates caspase-4/1 that cleaves the pore-forming protein gasdermin D to regulate IL-1β secretion

A hallmark of Entamoeba histolytica (Eh) invasion in the gut is acute inflammation dominated by the secretion of pro-inflammatory cytokines TNF-α and IL-1β. This is initiated when Eh in contact with macrophages in the lamina propria activates caspase-1 by recruiting the NLRP3 inflammasome complex in a Gal-lectin and EhCP-A5-dependent manner resulting in the maturation and secretion of IL-1β and IL-18. Here, we interrogated the requirements and mechanisms for Eh-induced caspase-4/1 activation in the cleavage of gasdermin D (GSDMD) to regulate bioactive IL-1β release in the absence of cell death in human macrophages. Unlike caspase-1, caspase-4 activation occurred as early as 10 min that was dependent on Eh Gal-lectin and EhCP-A5 binding to macrophages. By utilizing CRISPR-Cas9 gene edited CASP4/1, NLRP3 KO and ASC-def cells, caspase-4 activation was found to be independent of the canonical NLRP3 inflammasomes. In CRISPR-Cas9 gene edited CASP1 macrophages, caspase-4 activation was significantly up regulated that enhanced the enzymatic cleavage of GSDMD at the same cleavage site as caspase-1 to induce GSDMD pore formation and sustained bioactive IL-1β secretion. Eh-induced IL-1β secretion was independent of pyroptosis as revealed by pharmacological blockade of GSDMD pore formation and in CRISPR-Cas9 gene edited GSDMD KO macrophages. This was in marked contrast to the potent positive control, lipopolysaccharide + Nigericin that induced high expression of predominantly caspase-1 that efficiently cleaved GSDMD with high IL-1β secretion/release associated with massive cell pyroptosis. These results reveal that Eh triggered “hyperactivated macrophages” allowed caspase-4 dependent cleavage of GSDMD and IL-1β secretion to occur in the absence of pyroptosis that may play an important role in disease pathogenesis.

Introduction Entamoeba histolytica (Eh) is a protozoan parasite that infects about 10% of the world's population resulting in 10 6 deaths/year [1]. In approximately 90% of infected individuals, Eh colonizes the colon and results in a non-invasive and asymptomatic infection [2]. The protective host factors as well as those that contribute to the onset of pathology remain poorly understood. However, under conditions that are not well characterized, Eh breaches innate mucosal barriers, disrupts the epithelium and invades the lamina propria and submucosa where it can disseminate through the portal circulation and cause extra intestinal infections. As Eh is large, between 20-60 μm in diameter, it is too big to be phagocytosed by neutrophils or macrophages and remains extracellular throughout infection [3].
Macrophages are considered to be essential in the innate immune response to invasive Eh by killing the parasite directly and driving a pro-inflammatory response by recruiting inflammatory/immune cells to combat the infection [3,4]. Direct contact of Eh with macrophages is a critical "cue" that host cells use to detect Eh and initiate host defense [5]. We have shown that the major Eh surface adhesin, the Gal/GalNAc lectin (Gal-lectin) mediates an "adhesive" signal at the intercellular junction with macrophages to activate the NLRP3 inflammasome [5,6]. In addition to surface Gal-lectin, the Eh genome encodes numerous genes for cysteine proteases (CPs) that play important roles in Eh virulence and invasiveness [7,8]. When Eh contacts macrophage, it activates caspase-1 by the recruitment of the NLRP3 inflammasome complex in a Gal-lectin and Eh cysteine proteases 5 (EhCP-A5)-dependent manner, resulting in the maturation and secretion of interleukin (IL)-1β and IL-18 [5,6]. Inflammasomes are a group of multiprotein cytosolic receptors that are formed to mediate host immune responses to microbial infection and cellular damage [9]. Upon activation by pathogen or damage associated molecular patterns (PAMPs and DAMPs), the nucleotide-binding oligomerization domain (NOD)like receptor-pyrin containing 3 (NLRP3) signaling triggers oligomerization of the adaptor

E. histolytica-macrophage interaction regulates the activation of caspase-4
Direct interaction between Eh and macrophage via the Gal-lectin and engagement of EhCP-A5 RGD motif to α 5 β 1 integrin are important for triggering outside-in signaling to activate the NLRP3 inflammasome characterized by assessing caspase-1 processing and secretion in the extracellular media [5,6]. Activated caspase-1 in turn, cleaves the precursors of IL-1β and IL-18 into bioactive fragments and mediates their release, inducing cell pyroptosis [22]. More recently [35], we discovered that Eh activated caspase-4 that interacted with caspase-1 in a protein complex to enhance the cleavage of caspase-1 CARD proteins for IL-1β secretion [5,35]. Unfortunately, we do not know mechanistically how Eh activates caspase-4 and to address this deficiency, we first determined the kinetics of caspase-4/1 activation following contact with Eh and in response to the positive control lipopolysaccharide (LPS) + Nigericin (NGC) in THP-1 cells. Eh activated caspase-4 in a time (Fig 1A) and dose-dependent fashion with 1:20 Eh to macrophage ratio being the optimal dosage (S1A Fig). This time and dosage were used in all subsequent studies. Eh activation of caspase-4 was quantified by the appearance of the 30-34 kDa intermediate forms because the cysteine catalytic site was discovered on the large subunit and processing of the pro-form of caspase-4 generates several different intermediate products [41]. Even though both caspase-4/1 were activated a time-dependent fashion in response to Eh (Fig 1A), unlike caspase-1, higher amount of caspase-4 was activated and released within 10 min of incubation and accumulated in the cell supernatant up until 60 min (Fig 1B). These findings suggested that intracellular caspase-4 was cleaved from its inactive pro-form and carried out its bioactive function and subsequently secreted in the cell supernatant prior to caspase-1. Full activation for caspase-1 occurred between 20 and 30 min in response to Eh stimulation (Fig 1A). In comparison, LPS + NGC vigorously activated and secreted mostly caspase-1 and IL-1β (Fig 1A and 1B). Bioactive IL-1β levels gradually increased in a time-and dose-dependent manner, as quantified by HEK-Blue IL-1β reporter cells via the measurement of secreted embryonic alkaline phosphatase (SEAP) (Figs 1C and S1B). Unlike LPS + NGC, the release of inflammatory caspases and IL-1β in response to Eh (Eh-macrophage ratio: 1:20) was not due to significant cellular damage as confirmed by the release of cytosolic lactate dehydrogenase (LDH) into the cell supernatant (Fig 1D). Cytosolic LDH release into the extracellular media was used as an indicator for loss of membrane integrity to drive lytic cell death, including pyroptosis, because it is too large to exit through GSDMD NT pores and relies on cell lysis for its secretion. Following prolonged treatment with Eh up to 60 min, only 20% of cells death were noted whereas, LPS + NGC killed 80% of the cells as compared to unstimulated controls (Fig 1D). As expected, with increasing Eh to macrophage ratio, LDH release was significantly increased (S1C Fig). To assess whether IL-1β release was caspase-4/1-dependent, macrophages were pretreated with the pan-caspase inhibitor Z-VAD-fmk and the caspase-1-specific inhibitor Z-YVAD-fmk. Inhibition of both caspase-4/1 prevented the maturation and release of IL-1β (Fig 1E and 1F). To confirm a role for caspase-4 enzymatic activity in caspase-1 activation, Z-LEVD-FMK was used to inhibit caspase-4 activity in wild type (WT) macrophages stimulated with Eh and it also inhibited caspase-1 activation [35]. To test specificity of this inhibitor, CRISPR-Cas9 CASP4 KO cells were treated with Z-LEVD-FMK prior to incubation with Eh and it also inhibited caspase-1 activation. Based on these finding we did not use this caspase to inhibit caspase-4 [42]. In summary, these data pinpoint the importance of caspase-4/1 in processing and release of IL-1β with low cell death, suggesting that IL-1β was actively regulated and released from macrophages in response to Eh.

Caspase-4/1 activation parallels each other in response to E. histolytica
To determine if live Eh activated caspase-4 similar to caspase-1, macrophages were stimulated with live Eh, dead Eh, glutaraldehyde fixed Eh, and equivalent amount of freeze thawed whole lysates of Eh. As predicted, only live Eh in direct contact with macrophages activated caspase-4/1 and induced IL-1β secretion in the cell supernatant (Fig 2A and 2B). To define if Eh Gallectin-mediated adhesion was required for the activation of caspase-4, cells were stimulated Lipopolysaccharide (LPS) (50 ng/mL) and nigericin (NGC) (10 μM) stimulation for 60 min was used as a positive control. Immunoblot analysis was performed for caspase-4 and caspase-1 in supernatants (SN). (B) Quantifications of activated caspase-4 and caspase-1 were performed by densitometric analysis from three independent experiments and the negative control (cells only) acted as an internal control. Statistical significance was calculated between caspase-4 and caspase-1 at each time point. (C) Cell free supernatant was added to HEK-Blue IL-1β reporter cells to detect bioactive IL-1β via measurement of SEAP levels. (D) Cell death was quantified by lactate dehydrogenase (LDH) release into the culture supernatant and is shown as a percentage of LDH release compared to non-stimulated cells (control). (E, F) Macrophages were pre-incubated with the pan-caspase inhibitor Z-VAD-fmk (100 μM) and caspase-1 specific inhibitor Z-YVAD-fmk (100 μM) for 45 min followed by stimulation with Eh for 30 min. Caspase-4/1 activation as well as IL-1β secretion in the cell supernatant were assessed via immunoblotting. Cell free supernatant was added to HEK-Blue IL-1β reporter cells to detect bioactive IL-1β using the SEAP assay. Data and immunoblots are representative of at least three experiments (n = 3) and statistical significance was calculated with ANOVA and Bonferroni's post-hoc test ( � p < 0.05, �� p < 0.01, ���� p < 0.0001). Bars represent mean ± SEM. with Eh in the presence or absence of exogenous galactose that competitively blocked Eh from binding to macrophages via the Gal-lectin [43]. Inhibition of contact between Eh and macrophages abrogated caspase-4/1 activation as compared to glucose, the osmotic control and Eh treatment alone (Fig 2C). As caspase-1 activation required both Gal-lectin and EhCP-A5 [5,6], we determined if EhCP-A5deficient parasites could activation caspase-4 and indeed, it activated less caspase-4 ( Fig 2D) and IL-1β release as compared to WT Eh (Fig 2E). These results suggest that EhCP-A5 is required for triggering caspase-4 activation and IL-1β processing.
The activation of caspase-1 requires two-signals to safeguard against unintentional caspase activation. The priming signal upregulates the transcription of NLRP3 and pro-IL-1β, and the second signal recruits NLRP3, ASC, and pro-caspase-1 into a complex for the cleavage of pro- Live Eh were fixed with 1.5% glutaraldehyde for 1 h at 4˚C and washed 3 times with sterile cold PBS before use. LPS (50 ng/mL) and NGC (10 μM) stimulation for 60 min acted a positive control. Post incubation, the cell supernatant (SN) was TCA precipitated and equal amount was loaded onto the SDS-PAGE gel to detect caspase-4/1 activation with indicated antibodies. (B) Cell supernatant was added to HEK-Blue IL-1β reporter cells to detect bioactive IL-1β using the SEAP assay. Statistical significance was calculated between live Eh and fixed (dead) Eh, and between live Eh and Eh lysate (C) Macrophages were pretreated for 5 min with 55 mM D-galactose (Gal), or glucose (Glu) as an osmotic control and then incubated with Eh for 30 min at a 20:1 ratio. (D) Macrophages were incubated with Eh, and Eh deficient in CP5 (EhCP-A5 -) for 60 min and 90 min, respectively. LPS (50 ng/mL) and NGC (10 μM) stimulation for 60 min were used as a positive control. (E) Cell supernatant was added to HEK-Blue IL-1β reporter cells to detect bioactive IL-1β using the SEAP assay. Statistical significance was calculated between WT Eh and EhCP-A5 -Eh. Data and immunoblots are representative of at least three independent experiments (n = 3) and statistical significance was calculated with one-way ANOVA, followed by Bonferroni's post-hoc test ( ��� p < 0.001, ���� p < 0.0001). Bars represent mean ± SEM. https://doi.org/10.1371/journal.ppat.1010415.g002

PLOS PATHOGENS
E. histolytica-induced caspase4/1 regulates gasdermin D cleavage caspase-1 into active caspase-1 [44][45][46][47]. We next interrogated how caspase-4 activation is regulated in response to Eh and whether it also requires two-signals for its activation. To determine if a priming signal is necessary for caspase-4 activation, macrophages were exposed to Eh for increasing time points and pro-caspase-4 transcription and translation were determined through quantitative PCR (qPCR) and western blot analysis, respectively. Pro-caspase-4 transcription (Fig 3A) was upregulated within 10 mins and protein expression significantly increased temporally (Fig 3B and 3C). Since Eh Gal-lectin provides a critical adhesive signal to mediate Eh-macrophage attachment, we determined if soluble native Gal-lectin could initiate the priming step to upregulate caspase-4 expression in macrophages. Native Gal-lectin significantly enhanced pro-caspase-4/1 expression after 2 h exposure as compared to Eh stimulation Equal amount of lysate (LYS) was resolved on SDS-PAGE and immunoblotted for pro-caspase-4 detection. Blots were reprobed for GAPDH. Densitometry was performed to assess pro-caspase-4 proteins and the negative (cells only) acted as an internal control. (D) Macrophages were treated with native Eh Gal-lectin (500 ng/mL) for 2 h and pro-caspase-4, pro-caspase-1 and GAPDH levels were determined by western blot. (E-H) Caspase-4 activation requires ATP signaling via the P2X 7 receptor and pannexin-1 channels. (E) Macrophages were pretreated with oxidized ATP (oATP) for 2 h and then stimulated with Eh for 30 min. (F) Immunoblot analysis of active caspase-4 and caspase-1 in macrophages stimulated for 30 min with Eh with the addition of apyrase (20 U/mL). (G) Inhibition of caspase-4 with carbenoxolone (CBX), connexin/pannexin channel dual inhibitor, or pannexin channel inhibitor probenecid (PB). LPS and NGC were used as a positive control. Left: LPS (100 ng/ml) priming for 30 min and NGC (5 μM) stimulation for 30 min; right: LPS (50 ng/ml) priming for 30 min and NGC (10 μM) stimulation for 30 min. (H) Macrophages were incubated with CBX and PB for 30 min, prior to Eh stimulation for 30 min. Cell supernatant (SN) was TCA precipitated and cells were washed and lysed. Equal amount of supernatants and lysed cell lysates was loaded onto SDS-PAGE and immunoblot analysis was performed for caspase-4, caspase-1 and IL-1β. Data and immunoblots are representative of at least three independent experiments (n = 3) and statistical significance was calculated with one-way ANOVA, followed by Bonferroni's post-hoc test ( � p < 0.05, �� p < 0.01, ��� p < 0.001 ���� p < 0.0001). Bars represent mean ± SEM. https://doi.org/10.1371/journal.ppat.1010415.g003

PLOS PATHOGENS
E. histolytica-induced caspase4/1 regulates gasdermin D cleavage for 60 min (Fig 3D). These results support the hypothesis that Eh-induced caspase-4 activation required binding of Gal-lectin to macrophages as a priming signal that resulted in the upregulation of pro-caspase-4 transcription and expression. As ATP-P2X 7 receptor signaling is required for caspase-1 activation in response to Eh [5], we next investigated whether ATP gated P2X 7 receptor signaling was required for the activation of caspase-4. To determine if Ehinduced ATP release activated caspase-4 in an autocrine fashion similar to caspase-1 [5], macrophages were pre-treated with the specific antagonists of the P2X 7 receptor, oxidized ATP (oATP), prior to incubation with Eh. Surprisingly, increasing the concentrations of oATP upregulated caspase-4, whereas caspase-1 activation was inhibited (Fig 3E). To better define the role of ATP in mediating the activation of caspase-4, macrophages were cultured with apyrase that hydrolyses ATP to AMP and PPi. Upon addition of apyrase, the activation of caspase-4 was increased with a corresponding decrease in caspase-1 activation in response to Eh. These data validate the importance of ATP in activating caspase-4, and in cells treated with apyrase, caspase-4 activation was increased indicating that ATP acted as a second signal to induce its activation (Fig 3F). Given that the P2X 7 receptor is involved in mediating the activation of caspase-4 in response to Eh, and ATP is conducted into the extracellular space by non-junctional (hemi) pannexin-1 channels, we next delineated if pannexin-1 channels was required for the activation of caspase-4. To test whether Eh-induced ATP release through either the pannexin-1 or connexin channels [5,48] activated caspase-4, two inhibitors were used to block the channels, carbenoxolone (CBX), a dual antagonist of connexin/pannexin channels, and the specific pannexin antagonist, probenecid (PB). PB abolished caspase-4/1 activation and inhibited IL-1β release following stimulation with either LPS + NGC (Fig 3G) or Eh (Fig 3H), indicating that the release of ATP through the pannexin-1 channels is a critical signal for activating caspase-4. These results confirm that ATP acted as the second signal through the pannexin-1 channels to signal back onto the P2X 7 receptor to activate caspase-4 in response to Eh.
Ligation of EhCP-A5 RGD sequence to α 5 β 1 integrin on macrophage is an essential trigger to induce ATP release to activate the NLRP3 inflammasome [5]. To quantify the function of ATP as a potential second signal for Eh-induced caspase-4 activation, we added exogenous ATP to both WT and CRISPR-Cas9 CASP1 KO cells for the indicated time points. Eh stimulation alone was used as a positive control. Exogenous ATP modestly rescued caspase-4 activation in CRISPR-Cas9 CASP1 KO but not in WT macrophages; inflammasome activation was not restored (S2A and S2B Fig). To address if EhCP-A5 RGD sequence induces ATP release to activate the NLRP3 inflammasome is the singular function of EhCP-A5, we applied exogenous ATP to macrophage culture following stimulation with EhCP-A5 -Eh. Immunoblot analysis of secreted active caspase-4/1 cleavage product was performed and IL-1β release was quantified by SEAP assay from macrophages stimulated with WT Eh or EhCP-A5 -Eh for 60 or 90 min, with or without the addition of exogenous ATP, respectively (S2C and S2D Fig). Strikingly, exogenous ATP modestly restored caspase-4 activation and rescued IL-1β maturation and secretion in CRISPR-Cas9 CASP1 cells in response to EhCP-A5 -Eh, whereas it did not restore inflammasome activation (caspase-1 activation and IL-1β release) with EhCP-A5 -Eh in both cell types (S2E and S2F Fig). These data indicate that EhCP-A5initiates additional signaling that is critical for caspase-4 and NLRP3 inflammasome activation.

Caspase-4 activation does not require the NLRP3 inflammasome assembly in response to E. histolytica
We next quantified if Eh-induced caspase-4 activation is regulated via the canonical NLRP3 inflammasome, since the non-canonical caspase-4 inflammasome is considered inextricably linked to the canonical NLRP3 inflammasome, but crosstalk between these two pathways PLOS PATHOGENS E. histolytica-induced caspase4/1 regulates gasdermin D cleavage remains unclear. To dissect if Eh-induced caspase-4 activation is regulated by different NLRP3 inflammasome components, we first investigated if caspase-1 was required for Eh-induced caspase-4 activation. WT and CRISPR/Cas9 CASP1 KO macrophages were incubated with Eh for 10 and 30 min using LPS + NGC as a positive control. Unexpectedly, caspase-4 activation but not bioactive IL-1β secretion was significantly enhanced in CASP1 KO macrophages, indicating that caspase-4 activation was independent of caspase-1 in response to Eh (Fig 4A-4C). Caspase-4 activation was also enhanced in response to the positive control, LPS + NGC ( Fig  4B). To interrogate whether caspase-1 was dependent on caspase-4 for its activation, CRISPR/ Cas9 CASP4 KO macrophages were stimulated with Eh for 10 and 30 min. Intriguingly, caspase-1 activation (Fig 4D and 4E) and IL-1β secretion (Fig 4F) were significantly decreased in CASP4 KO macrophages as compared to WT counterparts, suggesting a critical role of caspase-4 in regulating caspase-1 activation in response to Eh. Consistent with Eh stimulation, LPS + NGC exhibited vigorously less caspase-1 activation and IL-1β secretion in CASP4 KO cells. Collectively, these data revealed that caspase-4/1 interacted to synergize pro-inflammatory responses elicited by Eh. WT and CRISPR/Cas9 CASP1 KO macrophages were stimulated with Eh for 10 and 30 min and unstimulated cells were used as an internal control. Macrophages stimulated with LPS (50 ng/mL) and NGC (10 μM) was used as a positive control. Quantifications of active caspase-4 protein were performed by densitometric analysis and negative (cells only) acted as an internal control. Statistical significance was calculated between WT and CASP1 KO macrophages at each time point (C) Bioactive IL-1β secretion in the histogram was quantified by the SEAP assay and statistical significance was calculated between WT and CASP1 KO macrophages at each time point. (D, E) WT and CRISPR/Cas9 CASP4 KO macrophages were incubated with Eh (20:1) at increasing time points. Caspase-1 CARD densitometry was measured and statistical significance was calculated between WT and CASP4 KO macrophages at each time point. Cell supernatant (SN) was TCA precipitated and equal amount of cell supernatants was loaded onto SDS-PAGE and immunoblot analysis was performed for caspase-4 and caspase-1. Active caspase-1 was quantified by densitometric analysis, and negative (cells only) acted as an internal control. (F) Cell supernatant was added to HEK-Blue IL-1β reporter cells to detect bioactive IL-1β secretion via measuring the SEAP and statistical significance was calculated between WT and CASP4 KO macrophages at each time point. Data and immunoblots are representative of at least three separate experiments (n = 3) and statistical significance was calculated with with Student's t-test between KO and WT, ( � p < 0.05, �� p < 0.01, ���� p < 0.0001). Bars represent mean ± SEM. The mechanisms governing crosstalk between the NLRP3 sensor and caspase-4/5/11 in response to Eh remains poorly understood. As the NLRP3 inflammasome requires the recruitment of ASC, NLRP3, and pro-caspase-1 into a high multimeric complex for activating caspase-1, we determined if the NLRP3 inflammasome components are also involved in activating caspase-4. To do this, macrophages deficient in ASC (ASC def) were stimulated with Eh and caspase-4 activation was unaffected, whereas, less caspase-1 CARD proteins and IL-1β secretion were evident in ASC def macrophages (Fig 5A and 5B). Similarly, in CRISPR/ Cas9 NLRP3 KO macrophages, caspase-4 activation was not significantly decreased (Fig 5C).

PLOS PATHOGENS
E. histolytica-induced caspase4/1 regulates gasdermin D cleavage However, there was less caspase-1 CARD proteins and IL-1β secretion as compared to WT macrophages, supporting the notion that NLRP3 is the central inflammasome activated by Eh to regulate IL-1β maturation and release (Fig 5C and 5D). These data demonstrate that caspase-4 activation is not dependent on the components of canonical NLRP3 inflammasome in response to Eh. Eh stimulation of macrophages for 30 mins did not cause significant bioactive IL-1β secretion in ASC def and NLRP3 KO macrophages as compared to WT (Fig 5E and 5F). This was in marked contrast to LPS + NGC stimulated WT cells were caspase-1 activation and IL-1β release were prominent, supporting the notion that LPS + NGC is highly dependent on the NLRP3 inflammasome pathway to elicit pro-inflammatory responses.

E. histolytica-induced caspase-4/1 cleaves gasdermin D to mediate IL-1β release
Proteomics have identified GSDMD as the most efficient and selective substrate for inflammatory caspases [49]. The function of GSDMD and its cleavage remained unclear until the investigation of GSDMD as the key executor of inflammasome-induced pyroptosis was uncovered using a chemical mutagenesis screen in mice and a cell-based CRISPR screen [20,50]. In these studies, GSDMD was identified as the critical mediator for pyroptotic cell death initiated by caspase-1/11 activation [20,22,50]. However, caspase-4 as the human ortholog of caspase-11 was less characterized, especially in the context of parasitic infections. To explore this, we first investigated the kinetics of Eh-induced caspase-4/1 activation in cleaving GSDMD and IL-1β release. Macrophages were incubated with Eh for increasing amount of time and immunoblot analysis demonstrated that as early as 5 min, the 55 kDa GSDMD pro-form was cleaved into the NT p30 pore-forming fragment that accumulated steadily in a time-dependent manner that peaked at 30 min (Fig 6A and 6B). This time frame perfectly paralleled the activation of caspase-4/1 (Fig 1A and 1B). As EhCP-A5 is required for caspase-4/1 activation and IL-1β processing, we next explored whether GSDMD cleavage would be affected by stimulating macrophages with EhCP-A5 -Eh as compared WT Eh. Surprisingly, EhCP-A5 -Eh significantly enhanced the cleavage of GSDMD but not IL-1β secretion as compared to WT Eh (S3A- S3C  Fig). This unexpected finding hinted that other caspases or cysteine proteases might be involved in cleaving GSDMD. In support of this, we have found that both EhCP-A1 and EhCP-A4 were polarized to the site of contact with macrophages to initiate rapid caspase-6-dependent degradation of the cytoskeletal-associated proteins, paxillin, talin, and Pyk2 to induce downstream inflammatory signaling pathways [51].
To determine if activated caspase-4 cleaved GSDMD, CRISPR/Cas9 CASP1 KO macrophages were stimulated with Eh and GSDMD was cleaved into the p30 NT fragment in a timedependent manner, whereas, in LPS + NGC treated cells there was no cleavage of GSDMD suggesting complete inhibition of the NLRP3 inflammasome (Fig 6C and 6D). In Eh stimulated CRISPR/Cas9 CASP4 KO cells, GSDMD cleavage was markedly inhibited but was maintained in the positive control, LPS + NGC (Fig 6E and 6F). Similarly, cleavage of GSDMD was inhibited with the pannexin channel inhibitor, probenecid (PB) that blocked caspase-4 activation (Figs S4A, S4B, 3G and 3H). The pan-caspase inhibitor, Z-VAD-fmk and caspase-1 inhibitor, Z-YVAD-fmk also inhibited the cleavage of GSDMD cleavage by inhibiting both caspase-4/1 (S4C and S4D Fig). These results support a crucial role for activated caspase-4 in cleaving GSDMD in response to Eh. These findings are noteworthy as previously it was widely considered that only caspase-1/11 triggered the cleavage of GSDMD and this study shows clearly that caspase-4 can also promote the cleavage of GSDMD. Even though both canonical and non-canonical inflammasomes were discovered to mediate the cleavage of GSDMD [50],

PLOS PATHOGENS
E. histolytica-induced caspase4/1 regulates gasdermin D cleavage it is unclear how inflammatory caspases regulate GSDMD cleavage upon Eh stimulation, and if the cleaved products can even impact the activity of upstream effectors.

Caspase-4 is essential for gasdermin D-regulated pro-inflammatory cytokine release independent of the NLRP3 inflammasome in response to E. histolytica
To determine which caspase was more important in cleaving GSDMD, CRISPR/Cas9 CASP1 KO and CASP4 KO macrophages were stimulated with Eh for 30 min using LPS + NGC as a positive control. In the absence of caspase-4/1, cleavage of GSDMD was significantly reduced and the effect was significantly more pronounced in CASP4 KO in response to Eh (Fig 7A and  7B). These results indicate that cleavage of GSDMD was not solely dependent on only one caspase but rather, both caspase-4/1 interacted together for maximal cleavage. Thus, to better

PLOS PATHOGENS
E. histolytica-induced caspase4/1 regulates gasdermin D cleavage define the role of caspase-4/1 in mediating IL-1β release, WT and CRISPR/Cas9 CASP4/1 KO were stimulate with Eh using LPS + NGC as a potent agonist for IL-1β (Fig 7C). Consistent with decreased cleavage of GSDMD in CASP4 KO cells, IL-1β release was significantly decreased as compared to WT stimulated with Eh. Conversely, there was only marginal reduction in IL-1β secretion in CASP1 KO in response to Eh and complete inhibition with LPS + NGC. As 30 min was too short to accurately quantify and discern differences in bioactive IL-1β secretion in WT and CASP4/1 KO cells, the kinetics of IL-1β release was measured up to 60 mins (Fig 7C). As predicted, Eh-induced IL-1β secretion occurred in a time-dependent manner with peak secretion after 60 min with significant reduction in IL-1β release in CASP4/1 KO cells in the absence of cell death (Fig 7D). These data reveal that activation of caspase-4 was essential for cleaving GSDMD to mediate bioactive IL-1β secretion. Intriguingly, Eh-induced cleavage of GSDMD was not dependent on caspase-1 activation but rather, may synergize with

PLOS PATHOGENS
E. histolytica-induced caspase4/1 regulates gasdermin D cleavage caspase-4 for enhanced cleavage. The dependency for caspase-4 in regulating IL-1β secretion in response to Eh was measured and sustained in the absence of cell death in comparison to the positive control, LPS + NGC that caused excessive IL-1β release and massive pyroptosis (Fig 7D).
We next substantiated if NLRP3 inflammasome components were indispensably involved in the cleavage of GSDMD. Mechanistically, genetic deficiencies in inflammasome components (e.g., NLRP3 or ASC) prevent caspase-1 activation, resulting in insufficient cleavage of GSDMD and defects in pyroptosis [20]. To determine if Eh-induced cleavage of GSDMD requires the recruitment of NLRP3 inflammasome components, ASC def and CRISPR/Cas9 NLRP3 KO macrophages were exposed to Eh from 5 min to 60 min and the cleaved GSDMD p30 fragment was confirmed via immunoblot analysis. As predicted, comparable GSDMD cleavage was noted in both ASC def and NLRP3 KO macrophages (Fig 8A-8D) with

PLOS PATHOGENS
E. histolytica-induced caspase4/1 regulates gasdermin D cleavage significant decrease in IL-1β release at 60 min (Fig 8E) in the absence of cell death (Fig 8F) as compared to WT cells. These results suggest that both the non-canonical caspase-4 and the canonical NLRP3 inflammasome are required to initiate sufficient cleavage of GSDMD to mediate IL-1β secretion, whereas the NLRP3 inflammasome is dispensable in this process.
Cleavage of gasdermin D regulates IL-1β release in the absence of cell death in response to E. histolytica GSDMD pores breaks the normal permeability barrier of cell membranes to disrupt cellular electrochemical potential to cause cell death. To determine whether GSDMD regulates pore formation and IL-1β release in response to Eh, studies were performed using CRISPR/Cas9 GSDMD KO macrophages. Intriguingly, even though GSDMD KO macrophages preserved the ability to activate the NLRP3 inflammasome in response to Eh and LPS + NGC, caspase-4/1 activation (Fig 9A and 9B), IL-1β secretion (Fig 9C) and pyroptosis (Fig 9D) was significantly

PLOS PATHOGENS
E. histolytica-induced caspase4/1 regulates gasdermin D cleavage reduced. As IL-1β release was not completely blocked in GSDMD KO cells in response to Eh (Fig 9C), these results suggest that other secretory pathways could mediate low level of IL-1β release. Cells that release IL-1β in a GSDMD-dependent manner while maintaining viability, reach a state that is different from their activated or pyroptotic counterparts termed "hyperactivated" [25,52]. Our data indicate that Eh-stimulated macrophages achieved a "hyperactivated state" with IL-1β release in the absence of excessive cell pyroptosis while maintaining efficient cleavage of GSDMD by active caspase-4/1 (Fig 9A-9D). This was in marked contrast to the positive control, LPS + NGC, where caspase-1 was highly activated that enhanced the cleavage of GSDMD to induce high levels of IL-1β and caused massive cell death (Fig 9C and 9D). At present, it is unclear if Eh-induced activated caspase-4/1 compete with each another to cleave GSDMD or, if the early activation of caspase-4 acted upstream of caspase-1 to regulate its activation and cleave GSDMD or whether, caspase-4 is the major caspase that promote GSDMD pore formation.

E. histolytica-induced gasdermin D pores function as a conduit to regulate IL-1β release
GSDMD pore formation in hyperactivated macrophages functions as gatekeepers to regulate IL-1β release [20,50]. As GSDMD pores regulate both IL-1β secretion and pyroptosis, we inhibited p30-GSDMD oligomerization [53] with necrosulfonamide (NSA), a chemical antagonist that binds GSDMD to block pores formation and measured IL-1β release and cell death in macrophages stimulated with Eh (Fig 10A). Increasing concentration of NSA significantly inhibited Eh-induced IL-1β secretion and pyroptotic cell death as compared to untreated Ehstimulated macrophages (Fig 10B-10D). NSA also significantly inhibited pyroptotic pore formation quantified by IL-1β release in CRISPR/Cas9 CASP4/1 KO cells in a dose-dependent manner and in response to LPS + NGC (Fig 10D). Under these conditions, cleavage of GSDMD was inhibited with increasing concentrations of NSA in Eh-stimulated macrophages as compared to LPS + NGC treated cells implying that both canonical and noncanonical inflammasome activation were affected (Fig 10E). In the presence of NSA in Eh-stimulated macrophages at 15 and 20 mins, CASP4 KO displayed robust downregulation in the cleavage of GSDMD as compared to CASP1 KO cells (Fig 10F and 10G). These data indicate that higher concentrations of NSA correlated with greater decrease in caspase-4 activation, supporting the importance of caspase-4 in cleaving GSDMD in response to Eh. These studies also provide a link and potential crosstalk between caspase-4/1 and cleavage of GSDMD that controls pyroptosis.

Gasdermin D is a proteolytic substrate for caspase-4/1
To explore mechanistically how Eh-induced caspase-4/1 regulates the cleavage of GSDMD, we determined the efficiency of the enzymes to cleave recombinant GSDMD. To do this, GST tagged GSDMD-NT and His-tagged CT human recombinant GSDMD (rGSDMD) were incubated with active recombinant caspase-1 (rC-1) and caspase-4 (rC-4) and the degraded fragments visualized by immunoblots (Fig 11A). rC-1 cleaved rGSDMD as early as 30 min whereas, the cleavage with rC-4 was gradually as depicted by the presence of GST-tagged NT-rGSDMD as compared to rC-1 (Fig 11A). Incubation with rC-1 and rC-4 at later time points (2-16h) revealed that rC-1 cleaved rGSDMD more rapidly than rC-4 with two prominent degraded cleavage fragments at 26 and 17kDa (Figs 11A and S5A). After 16 h enzymatic digest with rC-4, the GST-GSDMD-NT and His-GSDMD-CT were completely degraded whereas, both the GST and His tags were intact in the presence of rC-1. Silver-stained gels of higher sensitivity (S5B Fig) revealed that incubation with rC-1 resulted in rapid cleavage of rGSDMD PLOS PATHOGENS E. histolytica-induced caspase4/1 regulates gasdermin D cleavage within 5 min with degraded fragments at 58 and 25 kDa that correlated with the GST-GSDMD NT and His-GSDMD CT shown on the immunoblots. rC-4 cleavage of rGSDMD revealed comparable fragments at 58 kDa (NT), 25 kDa (CT) and 17 kDa (arrows). However, when immunoblots were probed with either GST or His antibody, only one cleavage product at 25 kDa was shown with either rC-1 or rC-4, suggesting that both caspase-4/1 cleave GSDMD only at one specific site (S5C Fig). Specificity for caspase-4/1 cleavage was shown with the pan-caspase inhibitor Z-VAD-fmk, and the caspase-1-specific inhibitor Z-YVAD-fmk, that completely inhibited the degradation of rGSDMD (Fig 11B). After Eh stimulation, cell free supernatant was added to HEK-Blue IL-1β reporter cells to detect bioactive IL-1β using the SEAP assay. (C) Pyroptotic pore formation and cell death were assessed through LDH release, cell free supernatants from the same experiments were used to quantify LDH released into the culture supernatant and is shown as a percentage of LDH release compared to non-stimulated cells. (B, C) Eh treatment only as a positive control and statistical significance was calculated between Eh 30 min and various concentration of NSA treatments. (D) Cell free supernatant was added to HEK-Blue IL-1β reporter cells to detect bioactive IL-1β using the SEAP assay to detect NSA inhibition in GSDMD pore formation in WT, CASP1 KO, CASP4 KO macrophage. LPS (50 ng/mL) and NGC (10 μM) stimulation for 60 min acted as the positive control. Statistical significance was calculated between WT and KO macrophages and between CASP1KO and CASP4 KO macrophages at each time point. (E-G) Immunoblot analysis was performed for GSDMD p30 cleavage in cell lysates (LYS), and blots were reprobed for GAPDH. WT, CRISPR/ Cas9 CASP1 KO and CRISPR/Cas9 CASP4 KO macrophages were pre-incubated with NSA for 60 min before stimulation with LPS + NGC. Data and immunoblots are representative of six experiments (n = 6) and statistical significance was calculated with Student's t-test and one-way ANOVA followed by post hoc Bonferroni test, ( � p < 0.05, �� p < 0.01, ��� p < 0.001, ���� p < 0.0001). Bars represent mean ± SEM. https://doi.org/10.1371/journal.ppat.1010415.g010

PLOS PATHOGENS
E. histolytica-induced caspase4/1 regulates gasdermin D cleavage rC-1 and rC-4 also completed degraded native GSDMD that was immunoprecipitated (IP) from macrophage cell lysates incubated for 16 h at 37˚C in a similar fashion as the recombinant protein (Fig 11C). We also expressed full length GSDMD in human embryonic kidney (HEK) 293T cells containing a Myc-DDK tag (Fig 11D) that was used to pull down overexpressed GSDMD. HEK 293T cells do not constitutively express GSDMD (Fig 11E) and the IP GSDMD containing the Myc-DDK tag (Fig 11F) incubated with active human rC-4 or rC-1 for 16 h was also completely degraded (Fig 11G). Shorter incubation times (30 min and 2 h) did not significantly degrade GSDMD (S5D Fig). Of interest, rC-1 cleaved DDK tagged GSDMD CT (C-terminus) was detectable in the cell lysate, whereas, it was absent in the presence of rC-4, suggesting that caspase-1 enzymatic activity might be more efficient at cleaving

Identification of caspase-4 cleavage site on gasdermin D
The caspase-1 cleavage site within the variable linker region of GSDMD is well-studied, however, it is not known where the cleavage site is for caspase-4. To identify the cleavage site of caspase-4 in the linker region of GSDMD, the two major cleaved 26 and 17 kDa fragments (Fig 12A) were excised and sequenced by Edman degradation that identified several amino acid calls. Schematic representation of full length rGSDMD (Fig 12B) shows the amino acid sequence for the red GST tag on the NT and the grey His tag on the CT. To uncover the cleavage site of the 26 kDa CT fragment, a cleavage specificity preference "logo" for P 4 -P 4' position

PLOS PATHOGENS
E. histolytica-induced caspase4/1 regulates gasdermin D cleavage was generated from "MEROPS" (www.ebi.ac.uk), which is a peptidase database based on a cleavage site specificity matrix table (Fig 12C). The cleavage site "logo" is a graphic representation of characteristic preference of individual amino acid presented at the P 4 -P 4' positions during caspase-4 cleavage. The Single-Letter Amino Acid Code was used to represent the preferable amino acid residues. By multiple sequence alignment analysis, the amino acids were arranged with the GSDMD sequence at asparagine-271 aa position and within these sequences we presumed that the one cleavage site for caspase-4 occurred at Aspartic Acid (D) position 275 (DGVPAE) (Fig 12D). By analyzing the cleavage site "logo" and using Expasy (expasy.org) to compute the molecular weight, the predicted cleaved peptide fragments was 58 and 25 kDa, respectively (Fig 12D). These data indicate that caspase-4 cleaves GSDMD at a single position that is identical to the caspase-1 cleavage site on human GSDMD (Fig 12E).

Quantitative proteomics analysis of E. histolytica-induced hyperactivated macrophage
We recently reported a quantitative shotgun proteomic analysis from bone marrow-derived macrophages stimulated with Eh for 10 mins [54] with a focus on proteins that regulated the autophagy pathway. Based on this proteomics analysis (Fig 13A) database [54], we performed a pathway enrichment and protein networking analysis comparing uninfected macrophages with Eh-induced "hyperactivated macrophages" by meta-analysis (metascape.org) [55] that revealed several top downregulated and upregulated pathways [51] (Fig 13B). Consistent to our previous findings [51], apoptotic signaling pathway and membrane trafficking were downregulated as indicated in blue, whereas, regulation of proteolysis and secretion by cell were upregulated and marked in red, corresponding to individual downregulated and upregulated proteins that were identified in the pathway enrichment analysis, respectively ( Table 1). By protein-protein network (STRING-db) [56] analysis, the interactions between upregulated and downregulated pathways from protein-protein enrichment analysis revealed several interesting top hit proteins (S6A and S6B Fig). Among the downregulated proteins in regulating membrane trafficking in response to Eh were SNAP23, RAB8A and RAB1A. We previously reported that the SNARE vesicle-associated membrane protein (VAMP8) present on mucin granules regulate exocytosis in human goblet cells in the presence of Eh [57]. These SNARE complexes are made up of the synaptosome-associated proteins (SNAP) and some syntaxins, as well as many SNARE chaperones to mediate formation of this complex [58]. VAMP8 as the critical vesicle SNARE plays important role in mucin secretion from intestinal goblet cells in response to Eh [57], and various SNARE proteins regulate autophagosome formation [59], both ATG7 and SNAP23 were downregulated upon Eh stimulation [54]. At present, it is not known what determines whether GSDMD cleavage triggers pyroptosis or hyperactivation. A membrane repair mechanism possibly exists in confronting membrane damage by GSDMD pores, which is capable of rapidly restoring membrane integrity, indicating an important aspect of Eh-induced downregulation in SNARE proteins in hyperactivated macrophages. Protein enrichment analysis conducted on approximately 900 proteins that were altered upon Eh stimulation shows that comparable amounts of proteins that are upregulated or downregulated in the presence of Eh (S6C Fig). Among the downregulated proteins (Fig 13C), nerve injuryinduced protein 1 ninjurin-1 (NINJ1) is of specific interest as it mediates pyroptosis-associated plasma membrane rupture during lytic cell death [60]. The downregulation of the apoptotic pathway in response to Eh strongly suggests that "hyperactivated macrophages" switch from apoptosis to pyroptosis [61]. Indeed, macrophages exposed to Eh for 10 and 30 min showed marked downregulation of NINJ1 protein expression (Fig 13D). These findings suggest that NINJ1 is a mediator of plasma membrane rupture that can release pro-inflammatory cytokines PLOS PATHOGENS E. histolytica-induced caspase4/1 regulates gasdermin D cleavage and DAMPs from hyperactivated macrophages in response to Eh. Downregulation on NINJ1 in response to Eh treatment supports the notion that Eh do not cause pyroptosis but triggers GSDMD-regulated hyperactivation in macrophages to elicit a proper pro-inflammatory response. STRING analysis of NINJ1 protein-protein interaction with other pyroptosis-relevant proteins suggested that NINJ1 was not involved in direct interaction with inflammasomeregulated events (Fig 13E). Based on the cumulative results of this study, we proposed a model for E. histolytica-macrophage interaction with activation of primarily caspase-4 (solid red arrow) that interacted with caspase-1 to regulate the cleavage of GSDMD pores for sustained IL-1β secretion (Fig 14).

Discussion
The focus of this study was to elucidate the molecular mechanisms at the Eh-macrophage intercellular junction that activated caspase-4/1 to cleave GSDMD that regulated the secretion

PLOS PATHOGENS
E. histolytica-induced caspase4/1 regulates gasdermin D cleavage of IL-1β in the absence of cell death. We have previously shown that Eh-induced caspase-1 activation via the NLRP3 inflammasome was extremely important in the secretion of IL-1β and IL-18 [10]; however, the molecular events that governed the activation of caspase-4 and its cross talk with caspase-1 to regulate IL-1β secretion via GSDMD pores was not known. Here we reveal that Eh-induced activation of caspase-4 followed a similar pattern as caspase-1 required a priming signal to upregulate pro-caspase-4 mRNA and protein. Intriguingly, we

PLOS PATHOGENS
E. histolytica-induced caspase4/1 regulates gasdermin D cleavage found that following Gal-lectin binding, ligation of EhCP-A5 RGD sequence to macrophage α 5 β 1 integrin was an essential trigger to activate caspase-4 in the absence of the NLRP3 inflammasome or any inflammasome components. With the use of CRISPR/Cas9 CASP4/1 KO cells, we uncovered that GSDMD was a key downstream substrate for both caspases to regulate proinflammatory cytokine secretion and/or pyroptotic cell death. By Edman degradation we identified that the GSDMD enzymatic cleavage site for caspase-4 was identical to caspase-1, however caspase-4 was indispensable in cleaving GSDMD to mediate pore formation that regulated sustained IL-1β release in response to Eh. Remarkably, Eh-induced high output IL-1β secretion from caspase-4-mediated GSDMD pores was not due to cell pyroptosis, but rather, the parasite seems to program macrophages to reach a "hyperactivated" state for sustained secretion of IL-1β. This was in distinct contrast to LPS + NGC stimulated cells that induced robust expression of caspase-1, high output IL-1β secretion and extensive pyroptotic cell death (~80%). Thus, the findings of this study have uncovered a new role for Eh-induced activation of caspase-4 that regulated measured GSDMD pore formation and pro-inflammatory secretion in the absence of cell death that may play a role in disease pathogenesis and innate host defence. A surprising finding was that GSDMD-dependent IL-1β secretion following contact with Eh occurred in the absence of significant cell death even after 60 min interaction. HEK IL-1β

Fig 14. Proposed schematic representation of E. histolytica-macrophage interaction and induction of caspase-4/1 activation and IL-1β
secretion. The activation of caspase is initially triggered by Eh in contact with macrophage via the Gal-lectin to Gal/GalNAc residues on the surface of macrophage. EhCP-A5 is highly expressed on the surface of Eh and following Gal-lectin binding brings, EhCP-A5 RGD sequences ligate α 5 β 1 integrin on the macrophage surface to induce the generation of ATP and release through the opening of pannexin-1 channel that subsequently signals back onto the P2X 7 receptor to activate the NLRP3 inflammasome. Simultaneously, K + efflux and the production of ROS collaborate to activate the NLRP3 inflammasome. The NLRP3 inflammasome in turn activates caspase-1, whereas, the activation of caspase-4 is independent of the inflammasome complex. Whereas both caspase-4/1 acted together to induce the cleavage of GSDMD, caspase-4 played a dominant role in this process. The cleaved GSDMD initiates pore formation allowing bioactive IL-1β release without causing significant cell pyroptosis. https://doi.org/10.1371/journal.ppat.1010415.g014

PLOS PATHOGENS
E. histolytica-induced caspase4/1 regulates gasdermin D cleavage reporter cells were used to quantify IL-1β release, as it was specific in measuring bioactive IL-1β through the generation of SEAP and the assay was consistency between experiments [6]. More importantly, even though both caspase-4/1 cleaved GSDMD in response to Eh, we assumed there would be a temporal order for these caspases to execute their functions and therefore explored which caspase cleaved GSDMD more efficiently/rapidly to regulate IL-1β release. Although both caspase-4/1 were required to orchestrate a competent pro-inflammatory response towards Eh, our data suggest that caspase-1 played a dominant role in processing pro-IL-1β into its bioactive form, whereas caspase-4 played a principal indispensable role in cleaving GSDMD for pore formation that regulated IL-1β release (Fig 7). These findings are noteworthy, as prior to this study, it was not known if caspase-4 played a dominant role in pore-forming activity during protozoan infections and there were no reports on the caspase-4 cleavage site on GSDMD. By applying GSDMD antagonist to macrophages, an intense reduction in IL-1β secretion and LDH release was detected, especially comparing Eh treatment to the positive control, LPS + NGC. These findings suggest that pharmacologically inhibiting GSDMD may be clinically efficacious for treating inflammatory diseases such as familial cryopyrin-associated periodic syndromes, autoimmune conditions and gastrointestinal diseases [62]. Despite the importance of studying the mechanisms that regulate GSDMD pore-forming activity, it is also crucial to reveal the mechanisms that control auto-inhibition, prior to activation and processing of GSDMD. Structural studies have revealed that the crystal structures of active caspases-1/4/11 form a complex with GSDMD-CT [63]. Hence, GSDMD CT functions would not only auto-inhibit GSDMD-NT but might serve as the platform to trigger inflammatory caspases recruitment and GSDMD cleavage.
Although cytosolic LPS and some gram-negative bacteria can trigger the activation of caspase-4 [34,64], we discovered that Eh, as an extracellular parasite can induce outside-in signaling in macrophages to activate caspase-4. Intriguingly we found that Eh coupling with macrophages via the Gal-lectin and EhCP-A5 to α 5 β 1 integrin [5,6] resulted in pannexin-1 channel-regulated ATP release that subsequently signals back through ATP-gated P2X 7 receptors to deliver a co-stimulatory/second signal to trigger the activation of both the NLRP3 inflammasome and noncanonical caspase-4. In vitro cleavage assays and Edman sequencing confirmed that the caspase-4 cleavage site on GSDMD was the same as caspase-1 that cleaved aspartic acid at position D275 (26 kDa fragments). GSDMD functions as the gatekeeper to release IL-1β and mediate cell pyroptosis. If few GSDMD pores are generated, the cell might react by initiating compensatory mechanisms to recover its volume. Alternatively, if the number of GSDMD pores patches to cell membrane exceeds the recovery capability of the cell, cell volume in turn increases. Consequently, the opening of GSDMD pore breaks the normal permeability barrier of the plasma membrane, resulting in membrane disruption, leading to pyroptotic cell death. Based on the progressive (temporal) cleavage of GSDMD induced by Eh in contact with macrophages that was remarkably similar to what was observed in CASP1 KO cells (Fig 6A and 6B) with higher IL-1β secretion up to 60 mins in the absence of cells death (Fig 7C and 7D), our data suggests that caspase-4 played a dominant role in generating sufficient GSDMD pores that allowed the cells to initiate compensatory mechanisms for its survival.
Pyroptosis is not the only means by which IL-1β is secreted from cells. As shown in this study, GSDMD-dependent IL-1β release triggered by Eh was not due to massive membrane rupture and significant pyroptosis, but instead, stimulated macrophages to reach a "hyperactivated stage". This was somewhat surprising as the use of lytic factors to destroy host cells for nutrient acquisition and immune evasion is a common feature of many invasive pathogens. This could be the typical scenario with Eh, or from a broader perspective, extracellular parasites that depends on outside-in signaling not to cause phagocytes death. Hence, it is of critical importance to determine what triggers cell hyperactivation and what are the consequences and significance of this "hyperactivation status". Unravelling the underlying mechanisms on how Eh manipulate macrophages to be "hyperactivated" might shed light on new strategies to alleviate amebiasis. Additionally, as plasma membrane rupture (PMR) was widely believed to occur spontaneously and passively after cell pyroptosis, it was interesting to demonstrate a requirement for NINJ1 to trigger pyroptosis-related and -unrelated PMR in mouse bone marrow derived macrophages [60]. In future studies, it will be interesting to dissect the relationship between GSDMD-triggered IL-1β secretion and NINJ1-dependent PMR in response to Eh. Our data provide promising evidence that the release of IL-1β form macrophages is independent of PMR and probably occurs via the approximately 10-15 nm GSDMD pores (IL-1 family cytokines have a diameter of 4.5 nm) [16,18]. However, what defines whether GSDMD cleavage causes pyroptosis or hyperactivation in macrophages remains unclear. There is emerging evidence that indicate a membrane repair response is triggered to combat membrane damage by GSDMD pores, and cells can sense membrane disruption by an instant boost of intracellular Ca 2+ to trigger membrane repair by replenishing the endosomal sorting complexes required for transport (ESCRT) [65]. Proteomic analysis on Eh-induced hyperactivated macrophages revealed downregulation in membrane trafficking, suggesting Eh hijacks this process and elicit a vigorous inflammatory response. Cell survival and membrane disruption potentially reflects the competition between how severe and how rapidly membrane is disintegrated, versus the efficiency of the repair process, which is possibly relied upon by how much GSDMD pore are formed. Various intensity of inflammatory stimuli and degree of caspase expression and activation can be essential determinants to establish cell pyroptosis or hyperactivation. It is worth noting that CRISPR/Cas9 GSDMD KO cells still release detectable levels of IL-1β in response to Eh (Fig 9) thus, a more in depth understanding of unconventional cytokine and the regulation of cell pyroptosis should be explored.
There is still much to be learned in innate defenses that underpin pro-inflammatory responses initiated by Eh via the canonical inflammasome-dependent and -independent pathways. We still do not understand how the host recognize and elicits an adequate level of response by assessing the level of threat caused by Eh. In this regard, our study advances plausible mechanisms of how inflammatory caspase-4/1 are activated through outside-in contactdependent signaling at the synapse between Eh and host immune cells. A major finding was that Eh triggered a "hyperactivated macrophage" state for high output IL-1β production in the absence of extensive cell pyroptosis despite efficient cleavage of GSDMD by activated caspase-4/1. These findings unravel new concepts on how Eh-induced inflammatory caspases can shape the magnitude of host pro-inflammatory responses that can play a role in disease pathogenesis and innate host defense.

Cultivation, harvesting of E. histolytica
E. histolytica virulent strain, HM-1:IMSS were grown axenically in TYI-S-3 medium supplemented with 100 U/mL penicillin and 100 μg/mL streptomycin sulfate at 37˚C in sealed 15 mL borosilicate glass tubes [66]. In order to maintain virulence, trophozoites are regularly passed through gerbil livers [67]. E. histolytica were harvested after 72 h of log-phase growth by placing on ice for 5 min and then centrifuged at 200 ×g for 5 min at 4˚C. After centrifugation, Eh were resuspended in serum-free RPMI to count and prepared a final cell suspension of 1×10 6 Eh/mL. EhCP-A5 deficient Eh were a generous gift from Dr. David Mirelman (Weizmann Institute of Science, Rehovot, Israel) and cultured similarly.

CRISPR/Cas9 KO gene editing of THP-1 cells
Caspase-1 and caspase-4 CRISPR/Cas9 KO THP-1 cells were a gift from Dr. V. Hornung (Institute of Molecular Medicine, University Hospital, University of Bonn, Germany). To generate this cell, CMV-mCherry-CAS9 expression cassette encoded plasmid and a gRNA under the U6 promoter was used. The CRISPR target regions were: ATTGACTCCGTTATTCCGAA AGG (Caspase-1) and GCTCATCCGAATATGGAGGCTGG (Caspase-4), PAM regions in bold. All CRISPR KO THP-1 cells were cultured in complete RPMI media as described above.
ASC def THP-1 cells were purchased from InvivoGen (thp-dasc). These cells were originally obtained from THP-1 human monocytic cells with no expression of ASC, but express native levels of NLRP3 and pro-caspase-1. CRISPR/Cas9 NLRP3 KO THP-1 cells were a generous gift from Dr. D. Muruve (Department of Immunology, University of Calgary, Canada). They used the gRNA CTGCAAGCTGGCCAGGTACCTGG and TGTCATAGCCCCGTAAT CAACGG. CRISPR Cas9 GSDMD KO THP-1 cells were a generous gift from Dr. Bachovchin (Memorial Sloan Kettering Cancer Center, New York, USA). The sgRNA used was TGAGTGTGGACCC-TAACACC.

Human gasdermin D plasmid transfection to HEK 293T and in vitro caspase-cleavage assay
HEK 293T cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FBS and antibiotics. HEK 293T cells (1.0 × 10 6 /well) were transiently transfected using jetPRIME Polyplus transfection reagent with human pCMV6-Gsdmd (Origene) plasmid, encoding the full-length GSDMD containing both a DYKDDDDK (FLAG) tag and a Myc tag on the C-terminal of the pLenti-C-Myc-DDK Lentiviral Gene Expression Vector. After 24 h, cells were lysed, collected and immunoprecipitated with anti-DYKDDDDK (FLAG) tag antibody followed by protein A/G PLUS-Agarose (Santa Cruz) conjugation at 4˚C. Coupled protein bound beads were resuspended into caspase cleavage buffer (50 mM HEPES, pH 7.2, 50 mM NaCl, 0.1% CHAPS, 10 mM EDTA, 5% glycerol and 10 mM DTT). 4U active recombinant human caspase-4 (ENZO) and caspase-1 (ENZO) were incubated with the isolated proteins for 16 h at 37˚C. After 16 h of incubation, samples were examined by immunoblot analysis with both anti-GSDMD (cell signaling) and anti-DYKDDDDK (FLAG) tag antibody.

Immunoprecipitation
For immunoprecipitation, anti-GSDMD (A305-736A-M, Bethyl Laboratories, Inc.) was used to pull down proteins from cell lysate. Immunoprecipitation of HEK 293T cells was described as above. Cells were lysed in lysis buffer (50 mM Tris-Cl pH 8.0, 150 mM NaCl, 1% NP40, 5 mM EDTA, 0.1% CHAPS) supplemented with protease inhibitor cocktail. 200 μg lysates were incubated with the relevant antibody at 4˚C before adding protein A/G PLUS-Agarose (Santa Cruz) for 2 h. Beads were washed three times with the same buffer and bound proteins were eluted with Laemmli buffer by boiling for 5 min. Post immunoprecipitation, complexes of beads-protein were incubated with recombinant caspase-1 (8U, ENZO) and recombinant caspase-4 (8U, ENZO) 16 h at 37˚C. Protein complexes were washed three times with lysis buffer and incubated at 95˚C for 5 min and resolved by immunoblotting. 24612). For inhibition studies, 100 μM pan caspase inhibitor Z-VAD-FMK and caspase-1 Z-YVAD-FMK was added for 10 min at room temperature prior to recombinant caspase-1 and caspase-4 incubation. After the indicated incubation times, samples were examined by western blot with anti GSDMD, anti-GST and anti-His antibodies.

IL-1β assay and HEK-Blue reporter cells
Following treatment of THP-1 macrophages, supernatants were kept on ice for immediate processing or frozen at -80˚C. In the following morning, HEK-Blue IL-1β cells were seeded onto a 96-well plate at 8 x 10 5 cells/mL with a total of 100 μL in each well. Next, 100 μL of supernatants from THP-1 macrophages were added undiluted into each well containing HEK-Blue IL-1β cells overnight at 37˚C in an incubator with 5% CO 2 . A standard curve using human recombinant IL-1β (200-01B, Peprotech) was made with serial dilutions (100 to 0.01 ng/μL). A total of three replicates were performed for each treatment condition. The following day, 100 μL of supernatant was transferred into a black 96-well plate. Next, 100 μL of the QUANTI-Blue solution containing reagent (rep-qbla, InvivoGen) and QUANTI-Blue buffer (rep-qblb, InvivoGen) were added into each well. The QUANTI-Blue solution is initially pink and eventually turns into blue with incubation, as indicative of SEAP levels. Bioactive IL-1β levels gradually increased and the intensity of the color reaction is proportional to the amount of IL-1β in the supernatant from stimulated THP-1 macrophages. The plate was incubated at 37˚C for 90 min and the SEAP levels were assessed using a spectrophotometer at 655 nm.

Cytotoxicity assay (LDH release)
Lactate dehydrogenase (LDH) released into extracellular media from macrophage culture was measured with the Promega CytoTox-ONE homogeneous membrane integrity assay (G7890, Fisher Scientific) following the instruction from the manufacturer. Relative LDH release was calculated using the equation: LDH (% release) = % of (LDH released from stimulation-background) / (maximum LDH released-background). This was performed to calculate percent cell death relative to a complete cell lysis control and the LDH (% released from control) was calculated using the LDH (% release) subtracted from non-stimulated cells.

Edman protein sequencing
Active recombinant caspase-4 were incubated with GST-tagged human recombinant GSDMD for 2 h at 37˚C. After 2 h of incubation, samples were resolved on SDS-PAGE gel and transferred to polyvinylidene difluoride (PVDF) membrane. Coomassie blue staining was used to visualize the degraded fragments. After detaining the membrane, each fragment band was cut and sent for Edman sequencing at Tufts University core facility.

Proteomic data and bioinformatics analysis
Uninfected macrophages and E. histolytica-induced hyperactivated macrophages were used for shotgun proteomics analysis. Detailed preparation and data collection process was as described previously [54]. MaxQuant [68] software was used for a peptide-spectrum match at a 1% false discovery rate (FDR), log2 of value was implemented to interpret changes in protein abundance.

mRNA expression analysis by real-time qPCR
Total RNA was extracted from snap-frozen tissue using E.Z.N.A. Total RNA Kit (Omega Bio-TEK) by following Trizol reagent protocol (Invitrogen; Life Technologies, Burlington, ON)

PLOS PATHOGENS
E. histolytica-induced caspase4/1 regulates gasdermin D cleavage per manufacturer's instructions. The purity and yield of the RNA was detected by the ratio of absorbance at 260/280 nm (NanoDrop, Thermo Scientific). qScript cDNA synthesis kit was used to prepare complementary DNA (cDNA). Rotor Gene 3000 real-time PCR system (Corbett Research) was used for mRNA expression analysis. Each reaction mixture contained 1:10 dilution of prepared cDNA, SYBR Green PCR Master Mix (Qiagen) and 1 μM of primers (F + R). Results were analyzed using the 2 −ΔΔCT methods and expressed as fold changes relative to housekeeping genes. The primer sequences and conditions used are listed below: Human GAPDH, F: GGATT TGGTCGTATTGGG, R: GGAAGATGGTGATGGGATT; Human caspase-4, F: AAGAGAA-GCAACGTATGGCAGGAC, R: GGACAAAGCTTGAGGGCAT CTGTA

Statistics
All experiments shown are representative of at least three independent experiments. Densitometry analysis was performed by the Image Lab software. GraphPad Prism 8 (Graph-Pad Software, San Diego, CA) was used for statistical analysis. Statistical significance between two groups was done by Student's t test and comparison between two or more groups were done by one-way analysis of variance (ANOVA), followed by post hoc Bonferroni test. Statistical significance was considered at p < 0.05. Results are displayed as mean ± standard error of the mean (SEM). ImageLab Software Version 6.0 was used for western blot analysis and to determine densitometric values from three independent experiments. Results were reported as mean ± SEM. (A) Macrophages were incubated with increasing Eh-macrophage ratios using LPS + NGC as a positive control. Unstimulated macrophages were used as a negative control. Cell supernatant (SN) was TCA precipitated and equal amount of proteins was resolved on SDS-PAGE following the investigation of activated caspase-4 and caspase-1 that was secreted into cell supernatant. (B) Cell supernatant from macrophages was added to HEK-Blue reporter cells to detect bioactive IL-1β via the SEAP assay macrophages that were incubated with Eh for increasing Eh-macrophage ratios. (C) Cell death was also determined by LDH released into cell culture and is shown as a percentage of LDH release compared to non-stimulated cells. Data and immunoblots are representative of at least three independent experiments (n = 3) and statistical significance was calculated with an ANOVA and Bonferroni's post-hoc test between each Eh-macrophage ratio and positive control treatment, ( �� p < 0.01, ���� p < 0.0001). Bars represent mean ± SEM. (TIFF)

S2 Fig. Exogenous ATP does not restore inflammasome activation in response to
EhC-P-A5 -Eh, but slightly increased caspase-4 activation in the absence of caspase-1. (A, B) Both WT and CRISPR/Cas9 CASP-1 KO macrophages were incubated with 5 mM exogenous ATP from 30 min to 4 h. Restored activation of caspase-4 was detected in CASP1 KO cells stimulated with ATP for 4 h (red box). (C, D) Immunoblot analysis was performed for active caspase-4 and caspase-1 products and IL-1β assay in HEK-Blue reporter cells from macrophages stimulated for 60 or 90 min with WT Eh and EhCP-A5 − Eh, respectively. Statistical significance was calculated between each treatment at the same time points (E, F) Exogenous ATP slightly restored caspase-4 activation to rescue IL-1β secretion in the absence of caspase-1 in response to EhCP-A5 -Eh. Cell supernatant was TCA precipitated and equal amount of PLOS PATHOGENS E. histolytica-induced caspase4/1 regulates gasdermin D cleavage supernatants (SN) was loaded onto SDS-PAGE and immunoblot analysis was performed for caspase-4, caspase-1 and IL-1β. IL-1β assay in HEK-Blue reporter cells from both WT and CASP1KO macrophages was quantified and statistical significance was calculated between each treatment under the same cell types. Data and immunoblots are representative of at least three independent experiments (n = 3) and statistical significance was calculated with one-way ANOVA, followed by Bonferroni's post-hoc test, ( � p < 0.05, ���� p < 0.0001, ns: not significant). Bars represent mean ± SEM.