Human macrophages undergo NAIP/NLRC4- and NLRP3-dependent inflammasome activation during Salmonella infection (Bierschenk et al., 2019; Gram et al., 2021; Naseer, Egan, et al., 2022), which restricts intracellular Salmonella replication (Naseer, Egan, et al., 2022). Nonetheless, how inflammasome activation controls Salmonella replication in human macrophages remains unclear. Inflammasomes recruit and activate caspase-1 in both murine and human cells (Franchi et al., 2006; Man, Hopkins, et al., 2014; Mariathasan et al., 2004; Miao et al., 2006; Ross et al., 2022; Zamboni et al., 2006), and caspase-1 promotes the control of Salmonella both in vivo and in vitro (Broz et al., 2010, 2012; Crowley et al., 2020; Hausmann et al., 2020; Holly et al., 2020; Lara-Tejero et al., 2006; Miao, Leaf, et al., 2010; Rauch et al., 2017; Raupach et al., 2006; Sellin et al., 2014; Thurston et al., 2016). Thus, we sought to test whether caspase-1 restricts Salmonella replication in human macrophages.

To first interrogate the impact of caspase activity on Salmonella replication in human macrophages, we pretreated macrophages derived from the human monocytic cell line, THP-1, with Ac-YVAD-cmk (YVAD), a chemical inhibitor of caspase-1 activity, or Z-VAD-FMK (ZVAD), a pan-caspase inhibitor. Upon infection with wild-type (WT) Salmonella, WT THP-1 cells pretreated with either YVAD or ZVAD had significantly reduced levels of IL-1β secretion and cell death compared to infected WT cells pretreated with DMSO, the vehicle control (Figure 1 – figure supplement 1A-B). Since pretreatment with YVAD largely phenocopies ZVAD, these data suggest that caspase-1 is the primary caspase that responds to Salmonella infection in THP-1 cells. We next assessed intracellular Salmonella burdens by determining bacterial colony forming units (CFUs). The increase in bacterial CFUs was significantly higher in WT cells pretreated either YVAD or ZVAD compared to cells pretreated with DMSO (Figure 1 – figure supplement 1C). Microscopic analysis revealed that WT cells treated with either YVAD or ZVAD prior to infection harbored significantly higher intracellular Salmonella burdens compared to cells pretreated with DMSO (Figure 1 – figure supplement 1D). Overall, these results suggest that caspase activity, primarily caspase-1 activity, controls Salmonella burdens in human macrophages.

Next, to genetically test the requirement of caspase-1, we used two independent CASP1^-/- THP-1 single cell clones generated through CRISPR/Cas9-mediated deletion (Okondo et al., 2017). In agreement with previous reports (Bierschenk et al., 2019; Gram et al., 2021; Naseer, Egan, et al., 2022), WT THP-1 cells infected with WT Salmonella exhibited high levels of IL-1β, IL-18, and IL-1α release as well as cell death at 6 hpi (Figure 1A-B; Figure 1 – figure supplement 2A-B). In contrast, CASP1^-/- THP-1 cells released negligible levels of inflammasome-dependent cytokines and did not undergo substantial cell death upon infection (Figure 1A-B; Figure 1 – figure supplement 2A-B). Infected WT and CASP1^-/- THP-1 cells released similar levels of the inflammasome-independent cytokine TNF-α (Figure 1 – figure supplement 2C). These results indicate that caspase-1 is required for inflammasome responses to Salmonella infection in human macrophages.

We next tested whether caspase-1 is required to restrict Salmonella replication in human macrophages. At 1 hour post-infection (hpi), we did not observe any significant differences in bacterial uptake between WT and CASP1^-/- THP-1 cells (Figure 1 – figure supplement 2D). However, at 6 hpi, CASP1^-/- cells harbored significantly higher bacterial burdens and a significant fold-increase in bacterial CFUs compared to WT cells (Figure 1C; Figure 1 – figure supplement 2E). To assess intracellular bacterial burdens on a single-cell level, we quantified the amount of WT Salmonella per cell by microscopy. WT cells contained relatively low numbers of Salmonella on average (∼5 bacteria per cell), while CASP1^-/- cells harbored significantly higher numbers of Salmonella on average (∼30 bacteria per cell) at 6 hpi (Figure 1D-E). Overall, these data indicate that caspase-1 limits intracellular Salmonella replication in human macrophages.

Caspase-11 is another member of the inflammatory caspase family, and plays a role in host defense against Salmonella in mice (Aachoui et al., 2013; Crowley et al., 2020; Knodler, Crowley, et al., 2014; Sellin et al., 2014). Humans harbor two orthologs of murine caspase-11: caspase-4 and caspase-5 (Shi et al., 2014). Caspases-11/4/5 recognize cytosolic LPS and form the noncanonical inflammasome to mediate downstream inflammatory responses (Casson et al., 2015; Hagar et al., 2013; Kayagaki et al., 2011, 2013; Lagrange et al., 2018; Schmid-Burgk et al., 2015; Shi et al., 2014). Caspase-4 contributes to inflammasome responses during Salmonella infection of THP-1 macrophages and primary human macrophages (Casson et al., 2015; Naseer, Egan, et al., 2022), and caspases-4/5 are required for inflammasome responses to Salmonella in THP-1 monocytes (Baker et al., 2015). In human intestinal epithelial cells (IECs), caspase-4 drives inflammasome responses during Salmonella infection and limits intracellular bacterial replication (Holly et al., 2020; Knodler, Crowley, et al., 2014; Naseer, Zhang, et al., 2022). However, whether caspase-4 contributes to restriction of intracellular Salmonella replication within human macrophages is unclear.

To genetically test the contribution of caspase-4 during Salmonella infection in THP-1 macrophages, we used CRISPR/Cas9-mediated deletion to disrupt the CASP4 gene. We selected and sequence-validated two independent CASP4^-/- THP-1 single cell clones (Figure 2 – figure supplement 1). We observed a slight decrease in secreted IL-1β levels in CASP4^-/- THP-1 macrophages infected with WT Salmonella compared to infected WT THP-1 macrophages at 6 hpi (Figure 2 – figure supplement 2A). We also observed a slight decrease in cell death at 6 hpi in infected CASP4^-/- clone #2 compared to infected WT cells, whereas cytotoxicity levels were unaffected in infected CASP4^-/- clone #6 (Figure 2 – figure supplement 2B). However, at 24 hpi, we observed a significant decrease in IL-1 release and cell death in both CASP4^-/- clones infected with WT Salmonella compared to infected WT cells (Figure 2A-B; Figure 2 – figure supplement 3A-B). WT and CASP4^-/- cells also released similar levels of the inflammasome-independent cytokine TNF-α upon infection with WT Salmonella at 24 hpi (Figure 2 – figure supplement 3C). Overall, these data indicate that caspase-4 contributes to inflammasome responses in THP-1 macrophages later during Salmonella infection.

Since caspase-4 restricts Salmonella replication in human epithelial cells (Holly et al., 2020; Knodler, Crowley, et al., 2014; Naseer, Zhang, et al., 2022), we then asked whether caspase-4 contributes to the control of Salmonella replication in human macrophages. Upon examination of the fold-change in bacterial CFUs at 6 hpi, we did not observe any significant differences between WT and CASP4^-/- THP-1 cells (Figure 2 – figure supplement 2C). However, at 24 hpi, CASP4^-/- cells harbored higher bacterial burdens and a significant fold-increase in bacterial CFUs compared to WT cells (Figure 2C; Figure 2 – figure supplement 3D). To confirm these findings, we enumerated the amount of WT Salmonella per cell by microscopy. At 6 hpi, WT and CASP4^-/- THP-1 cells contained comparable numbers of Salmonella per cell (Figure 2 – figure supplement 2D-E). In contrast, at 24 hpi, CASP4^-/- cells harbored significantly higher burdens of Salmonella per cell on average (∼20 bacteria per cell) compared to WT cells (∼10 bacteria per cell) (Figure 2D-E). Altogether, these results suggest that caspase-4 plays a larger role in controlling Salmonella burdens later during infection.

Inflammasome activation triggers a lytic form of cell death known as pyroptosis (Lamkanfi & Dixit, 2014). Death of the infected host cell eliminates Salmonella’s intracellular replicative niche and thus, may contribute to restricting intracellular bacterial replication. Upon its cleavage by inflammatory caspases, GSDMD forms pores in the host plasma membrane, resulting in pyroptosis (Agard et al., 2010; Ding et al., 2016; Kayagaki et al., 2015; Shi et al., 2015). Whether GSDMD contributes to the control of Salmonella replication in human macrophages is unknown.

First, we asked whether GSDMD pore formation plays a role in controlling Salmonella replication in human macrophages. To do this, we pretreated THP-1 macrophages and primary human monocyte-derived macrophages (hMDMs) with the chemical inhibitor disulfiram. Disulfiram prevents cleaved GSDMD from inserting into the host plasma membrane, thereby limiting GSDMD-mediated pore formation (J. J. Hu et al., 2020). Disulfiram treatment led to the loss of IL-1β release and cytotoxicity in macrophages infected with WT Salmonella, compared to treatment with the vehicle control, DMSO (Figure 3 – figure supplement 1A-B, D-E). Next, we examined the effect of GSDMD-mediated pore formation on intracellular Salmonella burdens. The fold-increase in bacterial CFUs at 6 hpi was significantly higher in cells treated with disulfiram compared to cells treated with DMSO (Figure 3 – figure supplement 1C, F). Collectively, these results suggest that GSDMD-mediated pore formation promotes the restriction of intracellular Salmonella replication.

Consistent with our findings with disulfiram, we observed a significant decrease in IL-1 cytokine release in GSDMD^-/- THP-1 macrophages (Okondo et al., 2017; Taabazuing et al., 2017) infected with WT Salmonella compared to infected WT THP-1 macrophages (Figure 3A; Figure 3 – figure supplement 2A-B). These results suggest that human macrophages undergo GSDMD-dependent IL-1 release during infection. Interestingly, loss of GSDMD did not completely abrogate the release of IL-1, suggesting that there is also GSDMD-independent IL-1 release (Figure 3A; Figure 3 – figure supplement 2A-B). Infected WT and GSDMD^-/- THP-1 cells secreted similar levels of the inflammasome-independent cytokine TNF-α (Figure 3 – figure supplement 2C). Importantly, infected GSDMD^-/- cells failed to undergo substantial cell death, in contrast to infected WT cells, suggesting that GSDMD facilitates cell death during Salmonella infection (Figure 3B). Overall, these results indicate that GSDMD plays an important role in mediating inflammasome responses during Salmonella infection of human macrophages.

We next examined whether GSDMD controls intracellular Salmonella replication in human macrophages. While we did not observe any differences in bacterial uptake between WT and GSDMD^-/- cells at 1 hpi (Figure 3 – figure supplement 2D), we did observe significantly higher bacterial CFUs in GSDMD^-/- cells compared to WT cells at 6 hpi (Figure 3 – figure supplement 2E). Moreover, the increase in bacterial CFUs at 6 hpi was significantly higher in GSDMD^-/- cells than in WT cells (Figure 3C). Next, we used microscopy to further interrogate intracellular Salmonella burdens. As expected, WT cells contained low numbers of Salmonella per cell (∼5 bacteria per cell) (Figure 3D-E). However, GSDMD^-/- cells harbored significantly higher numbers of Salmonella per cell on average (∼20 bacteria per cell) at 6 hpi (Figure 3D-E), suggesting that GSDMD restricts intracellular Salmonella burdens. Altogether, these data indicate that GSDMD promotes the control of Salmonella replication in human macrophages.

Terminal cell lysis is regulated downstream of GSDMD cleavage and pore formation (Bjanes et al., 2021; Kayagaki et al., 2021). While GSDMD pores release a subset of molecules, including IL-1 family cytokines, the extensive release of cellular contents is mediated by plasma membrane rupture (Ding et al., 2016; Ruan et al., 2018). Recently, NINJ1 was identified as an executioner of terminal cell lysis (Bjanes et al., 2021; Borges et al., 2022; Degen et al., 2023; Kayagaki et al., 2021, 2023). NINJ1 is a transmembrane protein that oligomerizes in the host plasma membrane to induce lytic rupture of the cell after the initiation of pyroptosis and other forms of regulated cell death (Degen et al., 2023; Kayagaki et al., 2021). Cells can be protected from plasma membrane rupture through treatment with the amino acid glycine (Fink & Cookson, 2006; Frank et al., 2000; Heilig et al., 2018; Verhoef et al., 2005), which interferes with the clustering of NINJ1 (Borges et al., 2022). Thus, upon incubation with glycine, cells can still form GSDMD pores but cannot undergo terminal cell lysis (Fink & Cookson, 2006; Heilig et al., 2018; Tsuchiya et al., 2021; Verhoef et al., 2005). Whether NINJ1-dependent cell lysis might contribute to control of Salmonella replication in human macrophages remains unknown.

First, to determine whether glycine exerts a cytoprotective effect on THP-1 macrophages, we pretreated WT and NAIP^-/- THP-1 cells with glycine prior to infection with WT Salmonella and assayed for downstream inflammasome responses. Notably, WT and NAIP^-/- cells pretreated with glycine exhibited significantly decreased cell death and a small defect in IL-1β release following infection compared to infected cells treated with the vehicle control (Figure 4 – figure supplement 1A-B). Release of the inflammasome-independent cytokine TNF-α was unaffected by glycine treatment (Figure 4 – figure supplement 1C). Altogether, these results indicate that glycine prevents cell lysis and limits IL-1 release in THP-1 macrophages upon Salmonella infection.

Given glycine’s cytoprotective effect on infected THP-1 macrophages, we next sought to determine whether glycine impacts intracellular Salmonella burdens in human macrophages. THP-1 macrophages pretreated with glycine retained significantly higher intracellular bacterial burdens at 6 hpi compared to cells given the vehicle control (Figure 4 – figure supplement 1D). We did not detect any significant differences in bacterial uptake (Figure 4 – figure supplement 1E). Moreover, WT and NAIP^-/- THP-1 cells pretreated with glycine exhibited a greater increase in Salmonella CFUs compared to cells treated with the vehicle control (Figure 4 – figure supplement 1F), suggesting that glycine treatment impedes intracellular control of Salmonella. Therefore, cytoprotection by glycine appears to interfere with the restriction of Salmonella burdens in human macrophages.

Next, to genetically test the role of NINJ1, we transfected WT THP-1 macrophages with small interfering RNA (siRNA) targeting NINJ1 or control scrambled siRNA. WT cells treated with control siRNA exhibited robust IL-1β release and cytotoxicity upon infection with WT Salmonella (Figure 4A-B). However, knockdown of NINJ1 resulted in a significant decrease, yet not complete abrogation, of cytotoxicity in infected cells (Figure 4B). Knockdown of NINJ1 also led to defect in IL-1β secretion (Figure 4A). In these experiments, we observed efficient siRNA-mediated knockdown of NINJ1 ranging from 77% to 85%. Collectively, these data suggest a critical role for NINJ1 in contributing to IL-1 release and cell death during Salmonella infection in human macrophages.

We then enumerated bacterial CFUs to assess whether NINJ1 impacts intracellular Salmonella replication. At 1 hpi, we did not observe any differences in bacterial uptake between WT THP-1 cells treated with control siRNA or NINJ1 siRNA (Figure 4C). However, at 6 hpi, we found that cells treated with NINJ1 siRNA contained higher intracellular bacterial burdens than cells treated with control siRNA (Figure 4D). Furthermore, there was a greater increase in bacterial CFUs at 6 hpi in NINJ1 siRNA-treated cells compared to control siRNA-treated cells (Figure 4E). Together, these findings suggest that NINJ1 contributes to intracellular bacterial control.

In epithelial cells, WT Salmonella can replicate in both vacuolar and cytosolic compartments, specifically hyperreplicating in the cytosol (Knodler, Crowley, et al., 2014; Knodler et al., 2010; Knodler, Nair, et al., 2014; Malik-Kale et al., 2012). However, in murine macrophages, WT Salmonella appears to replicate exclusively in vacuoles (Beuzón et al., 2002; Thurston et al., 2016). Salmonella lacking the SPI-2 effector SifA (ΔsifA), which is required for SCV membrane stability, frequently enter the cytosol and cannot replicate efficiently in murine macrophages (Beuzón et al., 2000, 2002; Thurston et al., 2016). Whether Salmonella replicates within vacuoles or the cytosol of human macrophages remains largely unknown. One study suggested that a small but significant proportion of Salmonella are exposed to the cytosol in THP-1 macrophages (Fisch et al., 2020). Our recently published data and current findings indicate that when human macrophages fail to undergo robust inflammasome responses, Salmonella hyperreplicates to large numbers within these cells (Naseer, Egan, et al., 2022). The hyperreplication that we observed in human macrophages was reminiscent of previous reports describing Salmonella hyperreplication within the cytosol of IECs (Knodler, Crowley, et al., 2014; Knodler et al., 2010; Knodler, Nair, et al., 2014; Malik-Kale et al., 2012). Thus, we hypothesized that inflammasome activation limits Salmonella replication in both SCVs and the host cell cytosol, and that in the absence of inflammasome responses, Salmonella hyperreplicates to large numbers within the host cell cytosol and also exhibits increased replication within SCVs.

First, to test whether inflammasome activation restricts the ability of Salmonella to replicate in vacuolar and cytosolic compartments of human macrophages, we used a chloroquine (CHQ) resistance assay. As a weak base, CHQ accumulates in endosomal compartments, including the SCV, without entering the cytosol of host cells (Klein, Powers, et al., 2017; Knodler, Nair, et al., 2014; Steinberg, 1994). Thus, vacuolar bacteria are CHQ-sensitive, while cytosolic bacteria are CHQ-resistant (Knodler, Nair, et al., 2014). We infected THP-1 macrophages with WT Salmonella and treated the infected cells with CHQ. Then, we determined the bacterial CFUs at 6 hpi, quantifying the numbers of vacuolar bacteria (CHQ-sensitive) and cytosolic bacteria (CHQ-resistant). We found that WT THP-1 cells contained mostly vacuolar Salmonella while also retaining some cytosolic Salmonella (Figure 5A), suggesting that a subset of Salmonella dwells in the cytosol of THP-1 macrophages, in agreement with a previous study (Fisch et al., 2020). In NAIP^-/- and CASP1^-/- THP-1 cells, we observed slightly higher burdens of vacuolar Salmonella compared to WT cells (Figure 5A). Strikingly, we also observed significantly larger numbers of cytosolic Salmonella in NAIP^-/- and CASP1^-/- cells compared to WT cells (Figure 5A). These data indicate that inflammasome signaling restricts Salmonella primarily within the host cell cytosol but also within SCVs in human macrophages.

The SPI-1 T3SS can damage the SCV (Roy et al., 2004), and the SPI-1 T3SS and its effectors contribute to Salmonella replication in the cytosol of epithelial cells (Chong et al., 2019; Klein, Grenz, et al., 2017; Knodler, Nair, et al., 2014). So, we next assessed whether the SPI-1 T3SS was required for the cytosolic exposure of Salmonella in human macrophages. We infected WT and CASP1^-/- THP-1 macrophages with Salmonella lacking the SPI-1 T3SS translocon protein SipB (ΔsipB), thus preventing SPI-1 T3SS effector translocation into host cells. While we observed a significant increase in the amount of vacuolar and cytosolic ΔsipB Salmonella in CASP1^-/- cells compared to WT cells at 6 hpi, the numbers of cytosolic ΔsipB were significantly lower than vacuolar ΔsipB in both WT and CASP1^-/- cells (Figure 5B). We also observed lower cytosolic ΔsipB Salmonella burdens compared to WT Salmonella in both WT and CASP1^-/- THP-1 cells (Figure 5A-B). Therefore, these results suggest that there is a partial dependence on the SPI-1 T3SS for Salmonella’s cytosolic access in human macrophages.

Since CFU assays are population-based, we next used single-cell methods to interrogate the subcellular localization of WT Salmonella in human macrophages. We relied on a strain of WT Salmonella that constitutively expresses mCherry and maintains a reporter plasmid, pNF101, that expresses gfp-ova under the control of a promoter responsive to the host cytosolic metabolite glucose-6-phosphate (Lau et al., 2019). We scored the number of GFP-positive/mCherry-positive bacteria (cytosolic) and GFP-negative/mCherry-positive bacteria (vacuolar) in WT, NAIP^-/- and CASP1^-/- THP-1 macrophages at 8 hpi by microscopy. We observed low numbers of both vacuolar and cytosolic populations of Salmonella in WT THP-1 cells (Figure 5C-D). In contrast, NAIP^-/- and CASP1^-/- THP-1 cells maintained significantly larger numbers of vacuolar Salmonella, with a substantial increase in the numbers of cytosolic Salmonella (Figure 5C-D). Taken together, these results suggest that inflammasome responses primarily restrict Salmonella replication within the host cell cytosol and also control bacterial replication within SCVs in human macrophages.

We next asked whether inflammasome responses also restrict the replication of Salmonella within the cytosol and SCV in primary human macrophages. We pretreated hMDMs with either ZVAD, to inhibit inflammasome responses mediated by caspase activity, or the vehicle control DMSO, and then infected the cells with WT Salmonella constitutively expressing mCherry and harboring pNF101. We observed that hMDMs treated with ZVAD contained significantly more cytosolic Salmonella than hMDMs treated with DMSO (Figure 5 – figure supplement 1A-B). We observed no significant differences between the vacuolar burdens of Salmonella in hMDMs pretreated with ZVAD or DMSO (Figure 5 – figure supplement 1A-B). Thus, inflammasome responses appear to primarily control Salmonella replication within the cytosol of primary human macrophages.

Finally, to characterize the effect of inflammasomes on the subcellular niches of Salmonella in human macrophages at higher resolution, we used transmission electron microscopy (TEM). In WT THP-1 cells, TEM analysis revealed that the majority of bacteria resided in a membrane-bound compartment (vacuolar, white arrows) (Figure 6A; Figure 6 – figure supplement 1B). In CASP1^-/- THP-1 cells, we observed a more mixed population of vacuolar bacteria and bacteria that were exposed to the host cell cytosol to varying extents (Figure 6B; Figure 6 – figure supplement 1B). We observed Salmonella free-living in the cytosol without a vacuolar membrane (fully cytosolic, black asterisk) as well as Salmonella exposed to the cytosol within discontinuous vacuolar membranes (partially cytosolic, cyan arrows) (Figure 6B; Figure 6 – figure supplement 1B). We observed a greater proportion of cytosol-exposed Salmonella in CASP1^-/- cells compared to WT cells (Figure 6 – figure supplement 1A). Strikingly, these TEM results revealed distinct intracellular populations of Salmonella in human macrophages, providing further evidence that inflammasome signaling controls the replicative niches that Salmonella occupies in human macrophages.

We next sought to further characterize the cytosolic exposure of Salmonella in CASP1^-/- THP-1 cells using electron tomography (ET), a technique that can reveal three-dimensional (3D) detail about SCV membranes. Indeed, we confirmed the presence of Salmonella in SCVs with varying degrees of vacuolar membrane discontinuities, yielding various extents of cytosolic exposure (Figure 6C; Figure 6 – figure supplement 1C; Video 1). Altogether, our TEM and ET data reveal the full spectrum of cytosolic exposure for Salmonella in CASP1^-/- cells as a result of vacuolar membrane discontinuities. Overall, our findings indicate that inflammasome responses play a role in modulating the subcellular populations of Salmonella, thereby controlling the number of Salmonella able to replicate within the cytosol and SCV in human macrophages.

All experiments on primary human monocyte-derived macrophages (hMDMs) were performed in compliance with the requirements of the US Department of Health and Human Services and the principles expressed in the Declaration of Helsinki. hMDMs were derived from samples obtained from the University of Pennsylvania Human Immunology Core, and they are considered to be a secondary use of deidentified human specimens and are exempt via Title 55 Part 46, Subpart A of 46.101 (b) of the Code of Federal Regulations.

THP-1 cells (TIB-202; American Type Culture Collection) were maintained in RPMI supplemented with 10% (vol/vol) heat-inactivated FBS, 0.05 nM β-mercaptoethanol, 100 IU/mL penicillin, and 100 μg/mL streptomycin at 37°C in a humidified incubator. Two days prior to experimentation, the cells were replated in media without antibiotics in a 48-well plate at a concentration of 2 × 10⁵ cells/well or in a 24-well plate at a concentration of 3.5 × 10⁵ cells per well and incubated with phorbol 12-myristate 13-acetate (PMA) for 24 hours to allow differentiation into macrophages. Then, macrophages were primed with 100 ng/mL Pam3CSK4 (Invivogen) for 16 to 20 hours prior to bacterial infections. For fluorescence microscopy experiments, cells were plated on glass coverslips in a 24-well plate.

Purified human monocytes from deidentified healthy human donors were obtained from the University of Pennsylvania Human Immunology Core. The monocytes were cultured in RPMI supplemented with 10% (vol/vol) heat-inactivated FBS, 2 mM L-glutamine, 100 IU/mL penicillin, 100 μg/ml streptomycin, and 50 ng/ml recombinant human M-CSF (Gemini Bio-Products). Cells were cultured for 4 days in 10 mL of media in at a concentration of 4-5 × 10⁵ cells/mL in 10 cm-dishes. Then, 10 mL of fresh growth media was added for an additional 2 days to promote complete differentiation into macrophages. One day prior to infection, the cells were rinsed with cold PBS, gently detached with trypsin-EDTA (0.05%) and replated in media without antibiotics and with 25 ng/mL human M-CSF in a 48-well plate at a concentration of 1 × 10⁵ cells per well or in a 24-well plate at a concentration of 2 × 10⁵ cells per well. For fluorescence microscopy, cells were replated on glass coverslips in a 24-well plate. For experiments involving LPS, cells were primed with 500 ng/mL LPS (Sigma-Aldrich) for 3 hours prior to bacterial infections.

Human macrophages were treated 1 hour prior to infection at the indicated concentrations with the following inhibitors: 20 μM of the pan-caspase inhibitor Z-VAD(OMe)-FMK (SM Biochemicals; SMFMK001), 25 μM of the caspase-1 inhibitor Ac-YVAD-cmk (Sigma-Aldrich; SML0429), and 40 μM of the GSDMD inhibitor disulfiram (Sigma). DMSO treatment was used as a vehicle control with these inhibitors. To prevent cell lysis, cells were treated with 20 mM glycine (Fisher Scientific) for 30 minutes prior to infection. Distilled water was used as a vehicle control.

Salmonella enterica serovar Typhimurium SL1344 WT was routinely grown shaking overnight at 37°C in Luria-Bertani (LB) broth with streptomycin (100 μg/mL). For infection of cultured cells, the overnight culture was diluted in LB with streptomycin (100 μg/mL) containing 300 mM NaCl and grown standing for 3 hours at 37°C to induce SPI-1 expression (Lee & Falkow, 1990). SL1344 WT glmS::Ptrc-mCherryST::FRT pNF101 (Lau et al., 2019) was kindly provided by Dr. Leigh Knodler. This strain constitutively expresses mCherry that is chromosomally encoded. It also harbors the PuhpT-gfpova plasmid, pNF101, where expression of GFP is under the control of the glucose-6-phosphate responsive uhpT promoter derived from Shigella flexneri. SL1344 WT glmS::Ptrc-mCherryST::FRT pNF101 was grown shaking overnight at 37°C in Luria-Bertani (LB) broth with streptomycin (100 μg/mL) and ampicillin (100 μg/mL). For SPI-1 induction prior to infection, the overnight culture was diluted in LB with streptomycin (100 μg/mL) and ampicillin (100 μg/mL) that also contained 300 mM NaCl and then grown standing for 3 hours at 37°C (Lee & Falkow, 1990).

Overnight cultures of Salmonella were diluted into LB broth with streptomycin (100 μg/mL) containing 300 mM NaCl and grown for 3 hours standing at 37°C to induce SPI-1 gene expression (Lee & Falkow, 1990). Bacterial cultures were then pelleted at 6,010 × g for 3 minutes, washed once with PBS, and resuspended in PBS. Human macrophages were infected with Salmonella at a multiplicity of infection (MOI) of 20. Infected cells were centrifuged at 290 × g for 10 minutes and incubated at 37°C. At 30 minutes post-infection, the cells were treated with 100 ng/mL of gentamicin to kill any extracellular Salmonella. Then, the infection proceeded at 37°C for 6 to 8 hours, as indicated. For all experiments, control cells were mock-infected with PBS.

Cells were infected with WT Salmonella as described above at an MOI of 20. Then, 30 minutes post-infection, cells were treated with 100 μg/ml of gentamicin to kill any extracellular bacteria. 1 hour post-infection, the media was replaced with fresh media containing 10 μg/ml of gentamicin. At the indicated time points, the infected cells were lysed with PBS containing 0.5% Triton to collect all intracellular Salmonella. Harvested bacteria were serially diluted in PBS and plated on LB agar plates containing streptomycin (100 μg/ml) to enumerate colony forming units (CFUs). Plates were incubated overnight at 37°C and CFUs were subsequently counted.

THP-1 macrophages were infected in 48-well plates as described above. For each timepoint, triplicate wells were incubated in the presence of CHQ (500 µM) and gentamicin (100 ng/mL) for 1 hour to quantify the CHQ-resistant bacteria (Bârzu et al., 1997; Fernandez et al., 2001; Klein, Powers, et al., 2017; Knodler, Nair, et al., 2014; Zychlinsky et al., 1994). Another triplicate wells were incubated with gentamicin (100 ng/mL) only to quantify the total intracellular bacteria. At the indicated time points, the infected cells were lysed with PBS containing 0.5% Triton to collect intracellular Salmonella. Harvested bacteria were serially diluted in PBS and plated on LB agar plates containing streptomycin (100 μg/ml) to enumerate colony forming units (CFUs). Plates were incubated overnight at 37°C and CFUs were subsequently counted.

Harvested supernatants from infected human macrophages were assayed for cytokine levels using ELISA kits for human IL-1α (R&D Systems), IL-18 (R&D Systems), IL-1β (BD Biosciences), and TNF-α (R&D Systems).

Harvested supernatants from infected human macrophages were assayed for cytotoxicity by quantifying the loss of cellular membrane integrity via lactate dehydrogenase (LDH) activity. LDH release was measured using an LDH Cytotoxicity Detection Kit (Clontech) according to the manufacturer’s instructions and normalized to mock-infected cells.

All Silencer Select siRNA oligos were purchased from Ambion (Life Technologies). Individual siRNA targeting NINJ1 (ID# s9556) was used. The two Silencer Select negative control siRNAs (Silencer Select Negative Control No. 1 siRNA and Silencer Select Negative Control No. 2 siRNA) were used as a control. Three days prior to the infection, 30 nM of siRNA was transfected into the human macrophages using Lipofectamine RNAiMAX transfection reagent (Thermo Fisher Scientific) following the manufacturer’s protocol. Then, 24 hours after transfection, the media was replaced with fresh media containing antibiotics. Finally, 16 hours before infection, the media was replaced with fresh antibiotic-free media containing 100 ng/ml Pam3CSK4.

RNA was isolated using the RNeasy Plus Mini Kit (Qiagen) following the manufacturer’s protocol. Human macrophages were lysed in 350 μL RLT buffer with β-mercaptoethanol and centrifuged through a QIAshredder spin column (Qiagen). cDNA was synthesized from isolated RNA using SuperScript II Reverse Transcriptase (Invitrogen) following the manufacturer’s protocol. Quantitative PCR was conducted with the CFX96 real-time system from Bio-Rad using the SsoFast EvaGreen Supermix with Low ROX (Bio-Rad). To calculate knockdown efficiency, mRNA levels of siRNA-treated cells were normalized to the housekeeping gene HPRT and control siRNA-treated cells using the 2^−ΔΔCT (cycle threshold) method (Livak & Schmittgen, 2001). The following primers from PrimerBank were used. The PrimerBank identifications are NINJ1 (148922910c1), CASP4 (73622124c2) and HPRT (164518913c1); all primers listed as 5′–3′:

NINJ1 forward: TCAAGTACGACCTTAACAACCCG

NINJ1 reverse: TGAAGATGTTGACTACCACGATG

CASP4 forward: TCTGCGGAACTGTGCATGATG

CASP4 reverse: TGTGTGATGAAGATAGAGCCCAT

HPRT forward: CCTGGCGTCGTGATTAGTGAT

HPRT reverse: AGACGTTCAGTCCTGTCCATAA

Cell lysates were harvested for immunoblot analysis by adding 1X SDS/PAGE sample buffer to cells. All protein samples (lysates) were boiled for 5 minutes. Samples were separated by SDS/PAGE on a 12% (vol/vol) acrylamide gel and transferred to PVDF Immobilon-P membranes (Millipore). Primary antibodies specific for caspase-4 (4450S; Cell Signaling) and β-actin (4967L; Cell Signaling) and HRP-conjugated secondary antibodies anti-mouse IgG (F00011; Cell Signaling) and anti-rabbit IgG (7074S; Cell Signaling) were used. ECL Western Blotting Substrate (Pierce Thermo Scientific) was used as the HRP substrate for detection.

Primary hMDMs or THP-1 cells were plated on glass coverslips in a 24-well plate as described above. Cells were either infected with WT Salmonella constitutively expressing GFP [Sl1344 harboring pFPV25.1 (Valdivia & Falkow, 1996)], or WT Salmonella constitutively expressing mCherry with a cytosolic GFP reporter [SL1344 glmS::Ptrc-mCherryST::FRT pNF101 (Lau et al., 2019)] at an MOI of 20 as described above. At the indicated timepoints following infection, cells were washed 2 times with PBS and then fixed with 4% paraformaldehyde for 10 minutes. Following fixation, cells were mounted on glass slides with DAPI mounting medium (Sigma Fluoroshield). Coverslips were imaged on an inverted fluorescence microscope (IX81; Olympus), and the images were collected using a high-resolution charge-coupled-device camera (FAST1394; QImaging) at a magnification of 100×. All images were analyzed and presented using SlideBook (version 5.0) software (Intelligent Imaging Innovations, Inc.) and ImageJ software. For experiments with WT Salmonella constitutively expressing GFP, the proportion of infected cells containing GFP-expressing Salmonella (green) were scored by counting 50 infected cells per coverslip. 150 total infected cells were scored for each condition. For experiments with WT Salmonella constitutively expressing mCherry with a cytosolic GFP reporter, the proportion of infected cells containing GFP-positive Salmonella (cytosolic) and GFP-negative, mCherry-positive Salmonella (vacuolar) were scored by counting 50 infected cells per coverslip. 150 total infected cells were scored for each condition.

Two days prior to experimentation, 2 × 10⁶ cells THP-1 cells were replated in 10-cm dishes in media without antibiotics and incubated with phorbol 12-myristate 13-acetate (PMA) for 24 hours to allow differentiation into macrophages. Then, macrophages were primed with 100 ng/mL Pam3CSK4 (Invivogen) for 16 hours prior to bacterial infection. Cells were infected with WT Salmonella at an MOI of 20 as described above. At 8 hpi, the media was aspirated, and the cells were fixed with 2.5% glutaraldehyde, 2.0% paraformaldehyde in 0.1M sodium cacodylate buffer, pH 7.4. Then, at the Electron Microscopy Resource Laboratory in the Perelman School of Medicine, after subsequent buffer washes, the samples were post-fixed in 2.0% osmium tetroxide with 1.5% K3Fe(CN)6 for 1 hour at room temperature, and rinsed in dH2O. After dehydration through a graded ethanol series, the tissue was infiltrated and embedded in EMbed-812 (Electron Microscopy Sciences, Fort Washington, PA). Thin sections were stained with uranyl acetate and SATO lead and examined with a JEOL 1010 electron microscope fitted with a Hamamatsu digital camera and AMT Advantage NanoSprint500 software.

Electron tomography was performed at room temperature on a ThermoFisher Krios G3i TEM equipped with a 300 keV field emission gun. Imaging was performed using the SerialEM software (Mastronarde, 2005) on a K3 direct electron detector (Xuong et al., 2007) (Gatan Inc., Pleasanton, CA, USA) operated in the electron-counted mode. We additionally used the Gatan Imaging Filter (Gatan Inc., Pleasanton, CA, USA) with a slit width of 20 eV to increase contrast by removing inelastically scattered electrons (Krivanek et al., 1995). After initially assessing the infected macrophages at lower magnifications, tilt series were collected at a magnification of 26,000X (with a corresponding pixel size of 3.38 Å) and a defocus of −6 µm. Precooking the target areas with a total dosage of 1000-1500 e/Ų was required to minimize sample shrinking and drifting during tilt series collection. A bi-directional tilt scheme was employed with 2° increments and 120° span (−60° to +60°). The cumulative dose of each tilt-series was in the order of ∼500 e−/Å2. Once acquired, tilt series were aligned (using the patch tracking function) and reconstructed into tomograms, both using the IMOD software package (Kremer et al., 1996). After careful assessment of the 3-dimensional (3-D) tomograms, optimal 2-D slices were chosen for figure presentations. Color overlays provide additional assistance with interpretation.

Prism 9.5.0 (GraphPad Software) was utilized for the graphing of data and all statistical analyses. Statistical significance for experiments with THP-1 macrophages and hMDMs was determined using the appropriate test and are indicated in each figure legend. Differences were considered statistically significant if the p value was <0.05.