ABSTRACT
Recent studies have determined that inflammasome signaling plays an important role in driving intestinal epithelial cell (IEC) responses to bacterial infections, such as Salmonella enterica serovar Typhimurium. There are two primary inflammasome pathways, canonical (involving caspase-1) and noncanonical (involving caspase-4 and -5 in humans and caspase-11 in mice). Prior studies identified the canonical inflammasome as the major pathway leading to interleukin-18 (IL-18) release and restriction of S. Typhimurium replication in the mouse cecum. In contrast, the human C2Bbe1 colorectal carcinoma cell line expresses little caspase-1 but instead utilizes caspase-4 to respond to S. Typhimurium infection. Intestinal enteroid culture has enabled long-term propagation of untransformed IECs from multiple species, including mouse and human. Capitalizing on this technology, we used a genetic approach to directly compare the relative importance of different inflammatory caspases in untransformed mouse and human IECs and transformed human IECs upon S. Typhimurium infection in vitro. We show that caspase-1 is important for restricting intracellular S. Typhimurium replication and initiating IL-18 secretion in mouse IECs but is dispensable in human IECs. In contrast, restriction of intracellular S. Typhimurium and production of IL-18 are dependent on caspase-4 in both transformed and untransformed human IECs. Notably, cytosolic replication in untransformed cells from both species was less pronounced than in transformed human cells, suggesting that transformation may impact additional pathways that restrict S. Typhimurium replication. Taken together, these data highlight the differences between mouse and human IECs and the utility of studying transformed and untransformed cells in parallel.
INTRODUCTION
Inflammasomes are an important innate immune pathway for sensing and responding to infection (1). Inflammasome formation is initiated following detection of bacterial products, viral replication, or sterile stressors, which leads to the activation of proinflammatory caspases: caspase-1 (CASP1), caspase-4 (CASP4), and caspase-5 (CASP5) in humans. Inflammasome activation culminates in pyroptosis, an inflammatory form of cell death, as well as the caspase-dependent processing and secretion of proinflammatory cytokines, interleukin-1β (IL-1β) and interleukin-18 (IL-18). The mouse and human proinflammatory caspase repertoires differ (2, 3). CASP1 is expressed in both species, while mice express both caspase-11 (CASP11), a homolog of human CASP4 and CASP5, and full-length caspase-12 (CASP12). Notably, the majority of humans express a truncated noncatalytic form of CASP12 involved in the regulation of inflammasome activation rather than direct catalytic activity (4). In both species, “canonical” caspase-1 is activated by diverse signals sensed by Nod-like receptor (NLR) proteins as well as other pathways, while human CASP4 and CASP5 and mouse CASP11 “noncanonical” caspases directly bind to, and are activated by, cytosolic lipopolysaccharide (LPS) (5, 6).
The roles played by inflammasomes in host defense have been defined by extensive studies in innate immune cells and in mice (6, 7), through infection by pathogens such as Salmonella enterica serovar Typhimurium. S. Typhimurium is a significant cause of gastroenteritis in humans and other mammals (8). Upon infection, S. Typhimurium colonizes the gastrointestinal tract, primarily the distal ileum and proximal colon in humans, leading to acute inflammation at the site of infection. Notably, S. Typhimurium can disseminate systemically in mice, where it primarily infects macrophages and dendritic cells; however, streptomycin pretreatment of mice leads to prolonged infection of the cecum (9).
While only limited knowledge exists regarding inflammasome activation in epithelial cells, particularly for humans (10–12), we previously reported that CASP4 is critical, but CASP1 and CASP5 are dispensable, for production of IL-18 and for pyroptosis following S. Typhimurium infection of a transformed human epithelial cell line (C2Bbe1) derived from a colorectal adenocarcinoma (13). In the same study, we showed that CASP11 is required to restrict the intracellular replication of S. Typhimurium in the mouse cecum at day 7 postinfection (p.i.). In contrast, a related study in mice found that CASP1, but not CASP11, is important for restricting replication of S. Typhimurium in the cecal epithelium shortly after infection (18 h p.i.) (14).
At present, it remains unclear whether the differences in inflammasome activation in murine versus human intestinal epithelial cells (IECs) truly reflect species differences or, alternatively, result from complicating issues of in vivo studies and transformed cells. To address this issue, we focused on enteroid technology. This approach allows for the culture and propagation of untransformed IECs as “mini-guts” that recapitulate the epithelial cellularity of the gastrointestinal tract (15). Moreover, the expression profile of enteroids closely resembles that of the intestinal epithelium and is not significantly impacted by the microbiome (16). While enteroids are three-dimensional (3D) structures with an internal lumen, we utilized a modified approach where enteroids are disrupted and the individual IECs are plated to form a monolayer. We subsequently generated these enteroid-derived monolayers from humans and mice and infected them with S. Typhimurium to compare the relative importance of different inflammatory caspases in untransformed mouse and human IECs.
RESULTS
Expression of inflammasome components in intestinal epithelial cells.To compare to the responses elicited in the transformed human C2Bbe1 colorectal carcinoma cell line, ileal enteroids were generated from normal adult human tissue from a deidentified donor (17). We refer here to these cells as human intestinal epithelial (hIE) cells. We used lentiCRISPR to knock out (KO) the gene for each proinflammatory caspase in hIE and C2Bbe1 cells and generated lines of each CRISPR target in both cell systems. In hIE, there was restriction fragment length polymorphism and Sanger sequencing evidence of editing for CASP1, CASP4, and CASP5 KO clones. To identify the genetic lesions present in these cells and to assess their clonality, we sequenced amplicons of the targeted genes using next-generation sequencing (NGS). For C2Bbe1 cells, we similarly chose six to eight clones of each CRISPR target for NGS. Sequencing data were analyzed using the CRISPResso2 computational pipeline (18) at moderate stringency (using a value of 10 for the minimum average Phred value for the entire read and for any single base pair), yielding an average mapped read depth of approximately 11,000. All of the hIE and C2Bbe1 lines were composed of more than two alleles, indicating that they were mixed cultures derived from multiple edited cells. All alleles detected by NGS within each cell population at a frequency >1% are listed in Table 1 for each targeted caspase gene for the wild type (WT) and one gene-edited clone used for subsequent studies. Notably, there was an unanticipated single nucleotide polymorphism (SNP) in the target sequence for the CASP5 guide RNA (gRNA) in one allele of both C2Bbe1 and hIE that was not edited, and so all of the CASP5 lines are heterozygous KOs. However, we were able to identify both hIE and C2Bbe1 lines containing minimal or undetectable WT alleles for CASP1 and CASP4.
Alleles present at a frequency of >1% in gene-edited hIE and C2Bbe1 cellsa
We then determined which components of the inflammasome are expressed in WT C2Bbe1 and hIE cells and their relative levels. At the transcriptional level, hIE and C2Bbe1 cells expressed all three proinflammatory caspases (Fig. 1A and B), although their relative expression was lower in C2Bbe1 cells. The genes encoding inflammasome-associated cytokines IL-1β and IL-18 were also expressed in both human cell types at similar levels. At the protein level, C2Bbe1 cells produced CASP4 but not CASP1 (Fig. 1C and D), consistent with our previous findings (13). Notably, in contrast to the transformed cell line, hIE strongly expressed both CASP1 and CASP4 (Fig. 1C and D). Immunoblotting confirmed that the genetic lesions in the edited lines abrogated expression of CASP1 and CASP4. CASP5 was not produced at detectable levels in either hIE or C2Bbe1 cells (Fig. 1E), making it impossible to assess the effects of editing and leaving it unlikely that CASP5 contributes to the sensing of S. Typhimurium. Similarly to hIE, mouse intestinal epithelial (mIE) cells express Casp1, Casp11, and Il18, but they do not express Il1b (11, 16, 19). Thus, both hIE and mIE cells are suitable systems for studying inflammasome activation during pathogenic infection.
Human enteroid-derived monolayers express components of inflammasomes. C2Bbe1 (A) and hIE (B) cells were harvested at 4 and 5 days postplating, respectively, for gene expression analysis. Data are the ΔCT of the target gene relative to that of HPRT. (C) CASP1 protein levels in THP-1 monocytes, WT hIE, CASP1 KO hIE clone 2-20, and WT C2Bbe1 (C2B) cells. Loading control is actin. (D) CASP4 protein levels in THP-1 monocytes, WT hIE, CASP4 KO hIE clones 1-6 and 1-7, and WT C2B cells. Loading control is GAPDH. CASP4 protein levels in CASP1 KO (clone 1-5) and CASP4 KO (clone 4-15) C2Bbe1 cells were assessed separately. Loading control is actin. (E) CASP5 protein levels in WT hIE, CASP5 Het hIE clones 3-01 and 3-08, and THP-1 cells. Loading control is actin.
S. Typhimurium replication in intestinal epithelial cells.To first establish the susceptibility to infection and replication kinetics of S. Typhimurium in IECs, we infected hIE, mIE, and C2Bbe1 cells with S. Typhimurium and enumerated intracellular CFU at 1.5, 3, 5, and 7 h p.i. using a gentamicin protection assay (Fig. 2A). As was reported previously, S. Typhimurium readily infected C2Bbe1 cells and replicated steadily between 1.5 and 7 h p.i. Intriguingly, both hIE and mIE cells were less susceptible to infection by S. Typhimurium, with fewer intracellular bacteria recovered at early time points than from C2Bbe1 cells. Note that the hIE and C2Bbe1 infections were initiated in parallel from the same inocula for each replicate. Interestingly, S. Typhimurium infection of hIE cells plateaued between 3 and 7 h p.i., while the number of intracellular S. Typhimurium cells in C2Bbe1 and mIE cells increased at a constant rate over the same time period. Although the initial infection (1.5 h p.i.) of the untransformed hIE and mIE cells was less than that of C2Bbe1 cells, the fold increases in intracellular CFU at 7 h were similar between the three cell culture systems (3.3- and 3.2- versus 3.1-fold, respectively). Thus, although bacterial entry is less efficient and results in lower initial numbers of intracellular bacteria, S. Typhimurium can replicate in untransformed mIE and hIE cells to a similar extent as in C2Bbe1 cells.
S. Typhimurium replicates in and induces IL-18 secretion from intestinal epithelial cells. (A) S. Typhimurium replication in C2Bbe1 (black squares; MOI = 100 to 200), hIE (blue circles; MOI = 100 to 200), and mIE (purple triangles; MOI = 50 to 200) cells. Data are means ± standard deviations (SDs) from the log transformation of intracellular CFU of n = 4 (mIE) or n = 3 (hIE and C2Bbe1) biological replicates. (B) Pro-IL-18 levels in uninfected WT C2Bbe1 (C2B), WT hIE, and CASP4 KO HIE5 clone 1-7 (C4 KO). Loading control is actin. (C) IL-18 secretion in response to S. Typhimurium infection was measured by ELISA. Samples are colored as described for panel A. Data are means ± SDs from n = 4 (mIE) or n = 3 (hIE and C2Bbe1) biological replicates. (D) Area under the curve (AUC) analysis of the individual replicates of IL-18 secretion in panel C. *, P = 0.01 to 0.05; **, P = 0.001 to 0.01; ***, P = 0.0001 to 0.001; ****, P < 0.0001. Color coding of asterisks corresponds to the groups that were compared.
S. Typhimurium infection of intestinal epithelial cells elicits robust and rapid IL-18 secretion.IL-18 secretion resulting from inflammasome activation is thought to protect against tissue damage as well as recruit NK cells to the site of gastrointestinal S. Typhimurium infection in vivo (20). We quantified IL-18 secretion by IECs in response to S. Typhimurium infection by enzyme-linked immunosorbent assay (ELISA) over a time course of infection in all three systems (Fig. 2C and D). C2Bbe1 cells began to secrete IL-18 at 3 h p.i., which accumulated in the medium for the duration of the experiment to a maximum concentration of 100 pg/ml. In contrast, hIE and mIE cells released much higher levels of IL-18 by 1.5 h and 3 h p.i., reaching concentrations of ∼225 pg/ml and ∼600 pg/ml, respectively. One potential explanation for the discrepancy in the kinetics and amount of IL-18 secretion between hIE and C2Bbe1 cells is a difference in basal pro-IL-18 levels. However, uninfected hIE did not contain more pro-IL-18 than C2Bbe1 cells (Fig. 2B). Therefore, the differences in kinetics of IL-18 release are potentially due to differential sensing of S. Typhimurium or differences in the physiology of the two cell types.
Caspase-dependent IL-18 secretion from intestinal epithelial cells in response to S. Typhimurium infection is species dependent.We previously showed that S. Typhimurium-induced IL-18 release from C2Bbe1 cells is CASP4 dependent (13). To determine whether IL-18 secretion in untransformed IECs is also dependent on proinflammatory caspase activity, we treated C2Bbe1 and hIE cells with two peptide inhibitors of CASP1/4/5, Z-WEHD-FMK and Z-YVAD-FMK. Although both inhibitors reduced IL-18 secretion in the two cell lines, they did not entirely abolish IL-18 secretion in either (see Fig. S1 in the supplemental material). Furthermore, there were no synergistic or additive effects of the two inhibitors in combination. These results suggest that proinflammatory caspases are involved in IL-18 secretion from primary human IECs but that a more efficient depletion of caspase activity was needed.
To definitively determine which caspase is responsible for IL-18 secretion during S. Typhimurium infection, we compared WT and CRISPR-edited hIE and C2Bbe1 cells. In the mouse system, we compared mIE cells derived from the ileal tissues of WT, Casp1−/−, Casp11−/−, and Casp1/11−/− mice. All cells were assayed 7 h p.i. for IL-18 secretion by ELISA. Deletion of CASP4 in both hIE and C2Bbe1 cells completely abrogated IL-18 secretion, whereas IL-18 secretion from CASP5 heterozygous (Het) and CASP1 KO hIE and C2Bbe1 cells was not reduced compared to that in WT cells (Fig. 3A and C). In mIE cells, Casp1 deficiency had the most profound effect on IL-18 secretion (Fig. 3B); however, near complete loss of IL-18 secretion was only observed when both Casp1 and Casp11 were absent. Thus, CASP4 is responsible for IL-18 secretion in human IECs, whereas CASP1 is dominant in mIE cells, with CASP11 contributing to a lesser extent.
Salmonella-induced IL-18 secretion is dependent on CASP4 in humans and CASP1 in mice. IL-18 secretion from S. Typhimurium-infected hIE (MOI = 100 to 200) (A), mIE (MOI = 50 to 200) (B), or C2Bbe1 (MOI = 50 to 100) (C) cells was measured by ELISA at 7 h p.i. Human cells (hIE and C2Bbe1) are wild type (WT), complete knockouts of CASP1 (C1; hIE clone 2-20 and C2Bbe1 clone 1-5) or CASP4 (C4; hIE clone 1-7 and C2Bbe1 clone 4-15), or heterozygous knockouts of CASP5 (C5; hIE clone 3-08 and C2Bbe1 clone 5-14). mIE cells are WT or complete knockouts of Casp1 (C1), Casp11 (C11), or both Casp1 and Casp11 (C1/11). Each data point is an independent biological replicate. Bars are the means ± SDs of secreted IL-18. Dotted lines are the average values for mock-infected samples. *, P = 0.01 to 0.05; **, P = 0.001 to 0.01; ***, P = 0.0001 to 0.001. Comparisons lacking annotation are not significant.
Caspase activation restricts intracellular replication of S. Typhimurium in intestinal epithelial cells.Previous studies have shown increased intraepithelial S. Typhimurium burdens in the ceca of infected Casp1/11−/− mice (14). In contrast, we showed that CASP4 is critical, but CASP1 and CASP5 are dispensable, for restricting intracellular S. Typhimurium replication in C2Bbe1 cells (13). These data indicate that proinflammatory caspases are key proteins involved in controlling intracellular replication of S. Typhimurium. We therefore examined the effects of caspase deficiency on intracellular S. Typhimurium levels in mouse and human IECs at 7 h p.i. CASP5 Het and CASP1 KO hIE and C2Bbe1 cells carried similar intracellular bacterial burdens to those of WT cells (Fig. 4A and C; see also Fig. S2A). In contrast, deletion of CASP4 in either cell type resulted in a significantly higher (∼5-fold) S. Typhimurium burden. Similarly, S. Typhimurium replicated to 3-fold higher levels in Casp1−/− and Casp1/11−/− mIE cells than in WT or Casp11−/− cells (Fig. 4B).
Caspase-dependent restriction of S. Typhimurium replication differs between mice and humans. Intracellular S. Typhimurium replication was assessed in hIE (A, D, and G), mIE (B and E), or C2Bbe1 (C, F, G, and H) cells at the indicated time points postinfection by enumerating total intracellular CFU from the population (A to C), scoring the number of S. Typhimurium organisms per cell by fluorescence microscopy (D to F), or quantifying bacterial burden on a per-cell basis by measuring the total fluorescence intensity of S. Typhimurium on an INCell analyzer 2000 (G). In panel H, the percentages of infected C2Bbe1 cells with ≥50 bacteria per cell from the data shown in panel F are graphed. MOIs are as follows: panel A, 50 to 300; B and E, 50 to 200; C, F, and H, 50 to 100; and D, 750. In panel G, the C2Bbe1 MOI is 150 to 200, while that of the hIE cells is 750. Genotype nomenclature and clones are the same as for Fig. 3. Bar graphs (A to C) show the means ± SDs from the log transformation of CFU counts, and the data for WT cells in panel B are reproduced from Fig. 2A. InCell data (G) are also log transformed. For fluorescence microscopy analysis (D to F), individual cell counts from n = 2 (mIE, 1 h), n = 6 to 7 (mIE, 7 h), n = 3 (hIE), or n = 4 (C2Bbe1) biological replicates were pooled, but replicates were compared individually for statistical analysis. A cutoff value of 30 was used for mIE cells containing ≥30 bacteria (E), and a cutoff value of 100 was used for C2Bbe1 cells containing ≥100 bacteria (F). Bar graph in panel H shows the means ± SDs. ns, P > 0.05; *, P = 0.01 to 0.05; **, P = 0.001 to 0.01. Unannotated comparisons were not analyzed.
As an alternative measure, we quantified the number of bacteria per infected epithelial cell by fluorescence microscopy. At 1 h p.i., the numbers of internalized bacteria per cell (± standard deviation) were similar for WT hIE (3.3 ± 2.4), CASP4 KO hIE (3.6 ± 3.1), WT mIE (1.6 ± 1.0), Casp1−/− mIE (2.1 ± 1.3), Casp11−/− mIE (1.7 ± 1.0), Casp1/11−/− mIE (1.8 ± 1.4), WT C2Bbe1 (4.3 ± 3.3), CASP1 KO C2Bbe1 (5.4 ± 3.5), CASP4 KO C2Bbe1 (5.2 ± 3.6), and CASP5 Het (5.5 ± 3.9) (Fig. 4D to F; Fig. S2B). Thus, despite slight differences in multiplicities of infection (MOIs) across cell types, the numbers of internalized bacteria are comparable. At 7 h p.i., WT C2Bbe1 cells and all of the gene-edited lines harbored a wide range (1 to ≥100) of S. Typhimurium organisms per cell (Fig. 4F; Fig. S2B), as was previously reported (13, 21). On average, bacterial numbers per cell were higher in C2Bbe1 cells than in hIE or mIE cells at 7 h p.i. (Fig. 4D to F). As we previously reported for CASP4 knockdown in C2Bbe1 cells (13), the percentage of infected cells containing a bacterial load of ≥50 bacteria was increased for CASP4 KO cells compared to that for WT C2Bbe1 cells (Fig. 4H; Fig. S2C). Even though untransformed hIE or mIE rarely harbored more than 20 bacteria per cell at 7 h p.i. (Fig. 4D and E), deletion of CASP4 in hIE and Casp1 in mIE resulted in a greater average number of S. Typhimurium per cell, a similar phenotype to that seen in C2Bbe1 cells. To expand our findings to a larger sample size, we also analyzed WT C2Bbe1 cells and WT and CASP4 KO hIE cells through high-content imaging and an automated image analysis pipeline. This method does not enumerate individual bacteria but quantifies the total bacterial burden of each cell as the integrated fluorescence intensity (IFI) of the chromosomally encoded mCherry expressed by the bacteria; however, akin to flow cytometry, it allows for a more comprehensive picture of the entire infected cell population. Overall, the results were consistent with the data from the fluorescence microscopy studies. In all cell types, the IFI increased between 1 and 7 h p.i., indicative of intracellular bacterial replication (Fig. 4G). At 7 h p.i., the IFI of WT C2Bbe1 cells was much higher than that of WT hIE cells, consistent with the data obtained by fluorescence microscopy (Fig. 4D and F) and gentamicin protection assays (Fig. 4A and C). However, deletion of CASP4 in hIE led to an increase in IFI, indicating that CASP4 plays a major role in restricting S. Typhimurium expansion in IECs. Thus, by multiple measures, CASP4 and CASP1 play the principal roles in human and murine IECs, respectively, in directing the ability of IEC-intrinsic inflammasomes to limit intracellular S. Typhimurium replication.
Caspase activation restricts cytosolic replication of S. Typhimurium in intestinal epithelial cells.While our data showed that inflammasome signaling limits replication of S. Typhimurium within untransformed human and murine IECs, the intracellular localization of the bacteria remained unclear. Previously, we and others showed that upon invasion of transformed IECs, S. Typhimurium is initially internalized within a membrane-bound vacuole, the Salmonella-containing vacuole (SCV) (8, 22). S. Typhimurium can either remain contained within the SCV or lyse its membrane and colonize the cytosol. Prior studies of C2Bbe1 cells correlated intracellular bacterial growth with vacuolar and cytosolic proliferation, of which only the cytosolic population increased upon CASP4 knockdown (13, 21, 23).
We sought to determine whether S. Typhimurium also reached the cytosol of untransformed IECs using a S. Typhimurium reporter strain (PuhpT-gfp S. Typhimurium) in which green fluorescent protein (GFP) is expressed under the control of the uhpT promoter, which is specifically activated by the presence of a cytosolic metabolite, glucose-6-phosphate (22). We quantified the proportion of infected cells harboring GFP-positive S. Typhimurium and the number of GFP-positive S. Typhimurium per cell at 7 h p.i. by fluorescence microscopy. Consistent with our previous data (13), cytosolic bacteria were detected in all C2Bbe1 genotypes (Fig. 5C, F, and H; see also Fig. S2D to F), while CASP4 deficiency led to significantly higher numbers of cytosolic bacteria per cell (Fig. 5F and H; Fig. S2E and F). Strikingly, GFP-positive bacteria were absent in WT hIE cells yet were found in 10.9% ± 2.3% of infected CASP4 KO hIE cells (Fig. 5A). When present, there was a vast range of GFP-positive bacteria per cell (Fig. 5D and G), indicating that cytosolic access alone does not necessarily equate with rapid bacterial proliferation. The phenotype in mIE cells was similar to that in hIE cells in that GFP-positive S. Typhimurium was absent in WT mIE cells (Fig. 5B), whereas the frequencies of cells with GFP-positive bacteria were highest in Casp1−/− and Casp1/11−/− mIE cells. Similar results were obtained when scoring for the number of GFP-positive bacteria per cell (Fig. 5E and G). Taken together, these results indicate that the inflammatory caspase required for bacterial restriction in the epithelial cytosol differs between mice (CASP1) and humans (CASP4). This restriction is absolute in untransformed hIE and mIE cells over the time periods examined. Moreover, these data suggest that S. Typhimurium can escape the SCV in untransformed IECs but that caspase activation more effectively restricts S. Typhimurium cytosolic survival and proliferation than in C2Bbe1 cells.
Caspase-dependent restriction of cytosolic S. Typhimurium replication differs between mice and humans. Cytosolic S. Typhimurium was assessed in hIE (MOI = 750) (A and D), mIE (MOI = 50 to 200) (B and E), or C2Bbe1 (MOI = 50 to 100) (C, F, and H) cells at 7 h p.i. In these S. Typhimurium strains, plasmid-borne gfpmut3.1 (hIE) or gfp_ova (mIE and C2Bbe1) are under the control of a glucose-6-phosphate-responsive gene promoter (PuhpT). Data are the percentages of infected cells containing GFP-positive intracellular S. Typhimurium (A to C) or counts of GFP-positive intracellular S. Typhimurium per cell by fluorescence microscopy (D to F). In panel H, the percentages of infected C2Bbe1 cells with ≥50 bacteria per cell from the data shown in panel F are graphed. Genotype nomenclature and clones are the same as for Fig. 3. Bar graphs are the means ± SDs. In panels D to F, individual cell counts from n = 3 (mIE), n = 3 (hIE), or n = 4 (C2Bbe1) biological replicates were pooled. *, P = 0.01 to 0.05; **, P = 0.001 to 0.01; ****, P < 0.0001. Comparisons lacking annotation are not significant. (G) Representative images of infected hIE and mIE cells. Red, mCherry; green, GFP; blue, Hoechst 33342 (hIE) or DAPI (mIE); gray, MemBrite Fix dye 640/660 (hIE) or Alexa Fluor 680 phalloidin (mIE). Bars, 5 μm.
DISCUSSION
Inflammasome activation in response to pathogen-associated molecular patterns from S. Typhimurium and other intracellular pathogens elicits a robust host response commensurate with recognition that host cells have been invaded (6). In myeloid cells, this response includes processing and secretion of potent inflammatory cytokines from the IL-1 family (IL-1β and IL-18) as well as incompletely understood cell fate decisions resulting in either pyroptotic cell death or more recently described hyperactivation (24, 25). Notably, we have now shown that similar responses are seen in infected untransformed IECs; however, the inflammatory caspase that mediates these responses is species dependent. The human-derived hIE and C2Bbe1 cell responses require CASP4 rather than the CASP1-dependent responses seen in mIE cells. Our expression results reflect prior studies indicating that IECs express abundant levels of IL-18 (11–13, 26). Thus, our finding that CASP1 is critical for IL-18 secretion in mIE cells is consistent with the known roles for this caspase in IL-1 family cytokine processing and with prior studies demonstrating the importance of CASP1 in how mice respond to S. Typhimurium infection (7, 14, 27, 28). In contrast, Casp11 deletion had a modest effect in our assays in that Casp11−/− mIE cells had a similar phenotype to that of WT mIE cells, and Casp1/11−/− mIE cells had a similar phenotype to that of Casp1−/− mIE cells. CASP11 impacts immune responses and inflammation in mice by driving macrophage cell death, which contributes to the control of Aspergillus fumigatus at late time points during infection (7, 27). CASP11 signaling also promotes the lethality of LPS endotoxemia (7, 29). Although there is good agreement on separable roles of CASP1 and CASP11 in mice, conclusions regarding the importance of CASP1 versus the human homologs of CASP11, CASP4 or CASP5, have varied significantly, largely due to different experimental setups and the cell types employed (13, 30, 31). As in our prior studies of C2Bbe1 cells (13), we found that CASP4 is critical for restricting intracellular replication of S. Typhimurium in hIE cells, specifically, within the cytosol, and for the secretion of IL-18. Unlike C2Bbe1 cells, hIE cells constitutively express significant levels of CASP1, allowing us to definitively rule out a role for CASP1 in IL-18 secretion upon infection of these cells with S. Typhimurium. CASP1 has been convincingly established as the enzyme that processes both pro-IL-18 and pro-IL-1β in phagocytic cells (2, 32). CASP4 is not known to directly process these cytokines in vivo, although studies have demonstrated cleavage of pro-IL-1β (33) and pro-IL-18 (34) by CASP4 in vitro. Additionally, we and others previously demonstrated that IL-18 is processed in CASP1-deficient epithelial cells (13, 35). Thus, another processor of pro-IL-18 must function in human epithelial cells. In addition, prior studies have shown that mouse CASP11 and human CASP4 and CASP5 are able to efficiently cleave gasdermin D (GSDMD) (36–38). Cleaved GSDMD serves as a portal for IL-1β and IL-18 secretion across lipid bilayers (36, 37, 39, 40). In macrophages, CASP11 activation in response to LPS requires recruitment of the NLRP3/PYCARD/CASP1 inflammasome to mediate IL-1β secretion (24, 29, 41, 42). In summary, our results support the existence of an alternative pathway of IL-18 activation independent of CASP1 in hIE and C2Bbe1 cells (32), reflecting both species- and cell type-dependent differences in inflammasome pathways.
An important strength of our approach is that human and mouse enteroids recapitulate the epithelial cellularity of the gut and offer an opportunity to explore pathogen-IEC interactions/responses under equivalent conditions across species and in direct comparison to prior experiments using transformed epithelial cells such as C2Bbe1 and HeLa. We previously demonstrated the expression of genes associated with specific subsets of differentiated IECs in the line of hIE cells used in this study when cultured both as enteroids and as monolayers (17). Thus, human enteroid-derived monolayers are a suitable system for investigating host-pathogen interactions. Although S. Typhimurium can infect many different cell lines (21) as well as mouse enteroids (43), human intestinal organoids (44), and human enteroids (45), we show for the first time the specific contribution of inflammatory caspases to multiple aspects of the interplay between S. Typhimurium replication and host responses in IECs derived from adult ileal tissue. Although prior studies have reported editing of human enteroids (46–48), they pose greater technical challenges than standard cell lines. Human ileal enteroids grow slowly (∼7-day doubling time). Thus, outgrowth of clonal lines takes weeks to months. Difficulties with stably transducing human enteroid stem cells have also been reported (47), and human enteroids are capable of fusing with neighboring enteroids (49). Unlike standard cell lines in which cloning discs or rings can be employed to isolate single colonies, individual human enteroids were manually picked and transferred to individual wells. Moreover, human enteroids do not thrive in isolation, and clonal selection results in a high attrition rate. However, improved medium conditions published subsequent to our efforts may ameliorate this challenge (47).
Despite the fact that the critical inflammatory caspase is species dependent, the conclusions made using hIE and mIE cells mirror some but not all of the findings obtained previously with C2Bbe1 cells, suggesting specific effects of cellular transformation. For example, both hIE and C2Bbe1 cells rely on CASP4 to control IL-18 secretion and limit the intracellular replication of S. Typhimurium. However, both hIE and mIE cells proved less susceptible to S. Typhimurium infection than C2Bbe1 cells, and at a population level, the replication kinetics of intracellular S. Typhimurium in hIE and mIE cells followed similar patterns that were distinct from that of infected C2Bbe1 cells. Previous studies in transformed cell lines have identified two replication phenotypes within epithelial cells, slower replicating bacteria contained with the SCVs and hyperreplicating bacteria present in the cytosol (13, 14, 21, 23, 50, 51). Cytosolic replication has also been noted in epithelial cells of the mouse intestine and gallbladder in vivo (23, 52); however, cytosolic replication is more extreme in transformed cell lines (≥100 bacteria/cell) (13, 21, 23) than in vivo (>20 bacteria/cell) (23, 52). We found that S. Typhimurium is capable of escaping the SCV and reaching the cytosol in all IEC types tested, although the fluorescent reporter for cytosolic access was only activated in the absence of CASP4 in hIE cells and in the absence of CASP1 in mIE cells. Notably, cytosolic replication was absent in WT hIE and mIE cells but prominent in C2Bbe1 cells. Thus, our single-cell analyses were consistent with the population average derived from gentamicin-resistant CFU. Although the MOIs varied among experiments, we observed a very narrow and consistent range of uptake across the different cell culture systems, as assessed by microscopy. Since these are the hIE experiments in which we used an MOI of 750, we can anticipate no more than 3 to 4 bacteria/hIE cell in experiments where we used lower MOIs. Thus, differences in the amount of cytosolic LPS as a function of internalized bacteria is a very unlikely explanation for the significant differences in cell-autonomous inflammasome activation that we observed across cell types and species. Additionally, since these experiments were performed with the same strain of S. Typhimurium and in some cases with the same inocula, these discrepancies cannot be explained by strain or culture variation. These results likely reflect some combination of decreased binding of S. Typhimurium to the cells, reduced invasion, reduced replication within the SCV, or greater host restriction of vacuolar and/or cytosolic replication in untransformed cells than in C2Bbe1 cells.
Inflammasome activation cannot account for all of the restriction observed in untransformed IECs, because even when the critical caspase was deleted, intracellular bacterial numbers did not approach those in WT C2Bbe1 cells. This is similar to the modest increase in cytosolic replication observed in vivo when the inflammasome activator, Nlrc4, is deleted in mice (52). An absence of cytosolic replication of S. Typhimurium has also been noted in mouse macrophages (53) and fibroblasts (54, 55). This is at least partially explained by the high levels of CASP11 in these cell types (56). A possible noninflammasome mechanism that could differ between transformed and untransformed epithelial cells that is important in fibroblasts is the activation of autophagy. Mouse embryonic fibroblasts (MEFs) lacking a key component of the autophagy pathway, ATG5, harbor more bacteria than WT MEFs at late time points (55). Autophagy also promotes repair of SCVs damaged by the type III secretion system I (57, 58), and two forms of autophagy are activated in rat and human fibroblast cell lines that target both damaged SCVs and a subset of intact SCVs, resulting in the persistence of a vacuolar population of S. Typhimurium (50, 54). It is thus possible that autophagy responses in untransformed cells are more capable of restricting cytosolic S. Typhimurium proliferation than those in transformed IECs.
Another difference between untransformed IECs and C2Bbe1 cells was the kinetics of IL-18 secretion. We showed that S. Typhimurium infection of hIE and mIE cells led to robust and rapid secretion of IL-18, whereas C2Bbe1 cells were delayed in secreting IL-18 until 3 h p.i. and failed to achieve levels of IL-18 secretion comparable to those of hIE or mIE cells at any point. The difference in IL-18 secretion was not due to increased levels of intracellular pro-IL-18 in resting cells but may reflect levels of pore formation by GSDMD or differences in membrane repair and pore closure. Although canonically thought of as concurrent with pyroptosis, cleavage of GSDMD by CASP4 or CASP1 was recently shown to promote pore formation that allows IL-18 release prior to complete cell death (40, 59) or sustained IL-1β release from cells with activated inflammasomes that do not undergo pyroptosis and remain viable and hyperactive (24, 25, 60–62). Our data are consistent with the concept of hyperactivation or delayed pyroptosis in C2Bbe1 cells and rapid pyroptosis in hIE and mIE cells. Thus, parallel studies of C2Bbe1 and hIE cells may provide information on inflammasome-dependent cell fate decisions.
An outstanding question is the mechanism(s) by which inflammasome activation leads to the restriction of S. Typhimurium replication within the epithelium (13, 14, 21) or, more precisely, within epithelial cells. Activated caspases themselves are not directly antibacterial, but they mediate processes that function to limit bacterial replication. One inflammasome-dependent mechanism through which IECs are thought to restrict intraepithelial bacterial loads is by initiating extrusion of infected cells from the epithelium (13, 14, 23, 63). Knockdown of CASP4 in C2Bbe1 cells led to an increased percentage of cells with cytosolic bacteria concomitant with reduced IEC extrusion (13). However, cytosolic replication may be sufficient but not necessary for activation of the noncanonical inflammasomes, as LPS can be released into the cytosol via damage to the SCV membrane by guanylate-binding proteins (50). Alternatively, S. Typhimurium could be transiently exposed to cytoplasmic sensors but restricted prior to replication. Thus, while IEC extrusion is likely a mechanism to limit the intraepithelial burden of S. Typhimurium, it is not necessarily dependent on cytosolic replication. A second possible mechanism is that activated CASP1 in mIE cells or CASP4 in hIE/C2Bbe1 cells cleaves GSDMD, exposing the N-terminal domain (36, 59). Expression of the GSDMD N-terminal domain alone is sufficient to induce cytotoxicity by oligomerizing and forming pores in the inner leaflet of cellular membranes (36, 40, 59). Cleaved GSDMD can also kill bacteria in suspension (59, 64, 65), although it is unknown whether this direct bactericidal activity is physiologically relevant. Alternatively, GSDMD-dependent pore formation could promote gentamicin uptake, which is supported by our previous demonstration that extruding C2Bbe1 cells containing cytosolic S. Typhimurium have a compromised plasma membrane (23). Finally, caspase-dependent control of cytosolic S. Typhimurium replication in macrophages was found to be independent of cell death and GSDMD but instead dependent on caspase catalytic activity, suggesting that other substrates of activated caspases in the cytosol may be directly antibacterial (56). Although we have not investigated the contribution of these potential mechanisms, the more robust restriction of intracellular S. Typhimurium replication in untransformed cells could be explained by reduced GSDMD expression or activation in C2Bbe1 cells. In conclusion, our parallel studies have illuminated aspects of host responses that differ between species and between untransformed and transformed cells. Future efforts are likely to uncover novel mechanisms to explain these differences, and our studies provide a strong rationale for parallel studies of both untransformed and transformed cells to provide complementary information.
MATERIALS AND METHODS
C2Bbe1 cells.C2Bbe1 cells (ATCC) were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 0.01 mg/ml human transferrin (Sigma) and 10% fetal bovine serum (FBS). Cells, 2.5 × 104 cells/well, were plated on 96-well plates previously coated with human placental collagen, type IV (Fig. 1 to 3 and 4G). The cultures were incubated for 4 days prior to infection, with a medium change every day. For Fig. 3C and 4C, 6.0 × 104 cells/well were seeded in 24-well plates on rat tail collagen-coated wells, and for Fig. 4F and 5C, F, and H, 8.0 × 104 cells/well were seeded on acid-washed glass coverslips in 24-well plates 24 h prior to infection.
hIE cells.Human ileal enteroids (HIE5) were originally cultured from normal healthy ileal tissue from one patient acquired from surgical resection, as previously described (17). For continuous culture, human ileal enteroids were maintained in Matrigel domes in human complete crypt culture medium (hCCCM): 50% Wnt3a conditioned medium (CM), 10% R-spondin-1 CM, 10% noggin CM, 1× B-27 supplement, 10 mM HEPES, 1× Glutamax, 1× antibiotic-antimycotic, 1 mM N-acetyl-l-cysteine (Sigma), 10 μM Y-27632, 50 ng/ml epidermal growth factor (EGF; Peprotech), 10 nM gastrin (Sigma), 50 ng/ml fibroblast growth factor 2 (basic fibroblast growth factor; Peprotech), 100 ng/ml insulin-like growth factor 1 (BioLegend), and 10 mM nicotinamide in DMEM. CM production and quality control were performed as previously described (17). All medium components were from Thermo Fisher Scientific, except where noted.
Monolayers of hIE cells derived from human ileal enteroids were generated as previously described (17). Briefly, human ileal enteroids were removed from Matrigel using cell recovery solution (Thermo Fisher Scientific), dissociated in 0.05% trypsin at 37°C for 5.5 min, quenched with DMEM containing 10% FBS, 10 mM HEPES, and 1× Glutamax, and mechanically dissociated with a pipette. Cells were passed through a 40-μm cell strainer, centrifuged at 400 × g for 5 min, and resuspended in hCCCM. Cells, 3.0 × 105 cells/well, were plated on 96-well plates previously coated with human placental collagen, type IV (Sigma), for all experiments except for those shown in Fig. 4D and 5A, D, and G, where 1.8 × 105 cells/well were plated on μ-slide angiogenesis slides (Ibidi). The medium was changed to hCCCM lacking nicotinamide, Y-27632, and antibiotics 1 day after seeding. The cultures were incubated for 5 days prior to infection, with medium changes every 2 days.
mIE cells.Mouse enteroids were derived from ileal tissues. WT C57BL/6N mice were purchased from Charles River (stock number 027). Casp1/11−/− and Casp11−/− mice were obtained from Genentech (13), while Casp1−/− (63) mice were a kind gift from Isabella Rauch and Russell Vance. All mice were repeatedly backcrossed on the C57BL/6N background and were housed under specific-pathogen-free conditions. Mouse enteroids were cultured in Matrigel domes in mouse complete crypt culture medium (mCCCM): 50% CM containing Wnt3a, R-spondin-3, and noggin derived from L-WRN cells (ATCC) in advanced DMEM/F12 supplemented with 1× Glutamax, 10 mM HEPES, 1× penicillin streptomycin, 1× B-27 supplement, 1× N-2 supplement, 1.25 mM N-acetyl-l-cysteine (Sigma-Aldrich), 10 mM nicotinamide (Sigma), 50 ng/ml murine epidermal growth factor, 0.5 μM A83-01 (Tocris), 10 μM SB 202190 (Sigma-Aldrich), and 10 μM Y-27632 (AbMole). All medium components were from Thermo Fisher Scientific, except where noted.
mIE monolayers cells were generated as previously described (66, 67), with modifications. Briefly, mouse enteroids were resuspended in cell recovery solution, dissociated in 0.05% trypsin at 37°C for 10 min, and mechanically dissociated with a pipette. Cells were centrifuged at 400 × g for 5 min and resuspended in mCCCM. Cells, 5.0 × 104 cells/well, were plated on 96-well plates previously coated with collagen I (Corning) for all experiments except for those shown in Fig. 4E and 5B, E, and G, where 2.0 × 105 cells/well in 24-well plates were plated onto glass coverslips previously coated with human embryonic stem cell (hESC)-qualified ready-to-use Geltrex (Thermo Fisher Scientific) for 1 h at 37°C. The medium was changed to mCCCM lacking nicotinamide, A83-01, and SB 202190 1 day after seeding. The cultures were incubated for 5 days prior to infection, with medium changes every 2 days.
Ethics statement.All mouse experiments were performed in strict accordance with the guide for the care and use of laboratory animals of the National Institutes of Health and in accordance with the international guiding principles for biomedical research involving animals. Protocols were approved by the Institutional Animal Care and Use Committee of the University of British Columbia.
Bacterial strains.S. enterica serovar Typhimurium SL1344 was the wild-type strain used in this study (68). SL1344 glmS::Ptrc-mCherryST::FRT, which constitutively expresses mCherry codon optimized for S. Typhimurium under the control of the trc promoter (13), was used for experiments shown in Fig. 2 to 4 with the exception that SL1344 pFPV-mCherry (69), which expresses plasmid-borne mCherry, was used with hIE and C2Bbe1 cells shown in Fig. 2. SL1344 glmS::Ptrc-mCherryST::FRT ΔPlac-PuhpT-gfpmut3.1, which expresses GFPmut3.1 (half-life > 24 h) under the control of the S. Typhimurium uhpT promoter, was used in the experiments shown in Fig. 5A, D and G (hIE cells). To construct this strain, oligonucleotides uhpT-XbaF (5′-GC TCT AGA ACG GCA ACC GCG GAC CGA TGA-3′) and uhpT-SmaR (5′-GCC CCC GGG TCG GCT TGC GCA CCT GGT-3′) were used to amplify a 573-bp fragment encompassing the uphT promoter from SL1344 genomic DNA. The amplicon was digested with XbaI and SmaI and ligated into XbaI- and SmaI-digested pGFPmut3.1 (Clontech). An XbaI-ApaI fragment encompassing PuhpT-gfpmut3.1 was then ligated into the corresponding sites of pMPMA3 ΔPlac (70). SL1344 glmS::Ptrc-mCherryST::FRT pNF101 (PuhpT-gfpova), which expresses unstable GFP_OVA (half-life of ∼88 min) under the control of the Shigella flexneri uhpT promoter (22), was used in the experiments shown in Fig. 5B, C, and E to G (mIE and C2Bbe1 cells) and Fig. 5H.
S. Typhimurium infections.For all experiments, S. Typhimurium cells were grown shaking overnight at 37°C at 220 rpm. The overnight culture was diluted 16- to 33-fold into LB-Miller broth and incubated for 3 h to 3.5 h at 37°C at 220 rpm to late log phase, centrifuged at 2,000 × g for 3 min, washed twice with Dulbecco’s phosphate-buffered saline (DPBS), and resuspended in 10% of their original volume in DPBS. S. Typhimurium cells were then diluted into medium specific to each cell type and added to cells. S. Typhimurium was allowed to bind and internalize for 10 min, cells were washed with PBS three times to remove noninternalized bacteria, and fresh medium was added to the cells. At 30 min p.i., cells were incubated with medium containing 100 μg/ml gentamicin for 1 h at 37°C to kill extracellular bacteria. Thereafter, cells were incubated with medium containing 10 μg/ml gentamicin for the duration of the experiment.
Secreted cytokine measurements.Cytokines in culture supernatants were quantified using commercial ELISA kits for human IL-18 (R&D Systems and MBL) and murine IL-18 (Sigma) according to the manufacturers’ instructions. The human IL-18 ELISA from MBL only detects processed IL-18 (71), and the R&D Systems human IL-18 ELISA is specific for processed IL-18, with only 0.5% cross-reactivity to pro-IL-18.
Enumeration of gentamicin-resistant CFU.Infected cells were washed once with DPBS, lysed in 200 μl of 1% (vol/vol) Triton X-100 in water (hIE and mIE) or 0.2% sodium deoxycholate (C2Bbe1), and serially diluted in DPBS. Each serial dilution was plated in triplicates on LB agar plates and incubated at 37°C for ∼24 h.
Fluorescence microscopy.hIE cells (and C2Bbe1 cells for experiments shown in Fig. 4G) were washed once with Hanks’ balanced salt solution, stained with MemBrite Fix dye 640/660 (Biotium) according to the manufacturer’s instructions, fixed in 4% paraformaldehyde (PFA) for 20 min, and stained with 10 μg/ml Hoechst 33342 (Thermo Fisher Scientific) for 15 min at room temperature (RT). Images were acquired using a Nikon A1R point scanning confocal system (Fig. 4A and D and 5A, D, and G) or a GE Healthcare INCell Analyzer 2000 (Fig. 4G). Z projections of images from at least 4 (confocal) or 10 (INCell) fields of view per sample were segmented using Cell Profiler 3.1.9 (72). Bacteria per cell were enumerated manually for confocal images. For INCell images, a global threshold for mCherry was determined on uninfected control wells, and the integrated mCherry fluorescence above threshold was calculated for each cell using Cell Profiler.
mIE cells seeded on glass coverslips were fixed with 2.5% PFA at 37°C for 10 min and then permeabilized with 0.1% Triton X-100 and 0.05% Tween 20 in PBS for 15 min. Fixed and permeabilized cells were blocked with 2% donkey serum, 0.01% Triton X-100, and 0.05% Tween 20 in PBS overnight. Coverslips were then stained with Alexa Fluor 680-phalloidin (1:2,000; Thermo Fisher Scientific) for 30 min at RT and mounted onto glass slides using ProLong Gold Antifade reagent containing 4′,6-diamidino-2-phenylindole (DAPI) to stain the nuclei (Thermo Fisher Scientific). Monolayers were viewed using a Zeiss AxioImager microscope and AxioCam HRm Camera (Fig. 4B and E and 5B, E, and G). The numbers of bacteria per cell were enumerated manually.
Infected C2Bbe1 monolayers were fixed with 2.5% PFA at 37°C for 10 min, blocked, and permeabilized in 10% normal goat serum with 0.2% saponin in PBS containing Alexa Fluor 488 or Alexa Fluor 647 phalloidin (1:400 dilution) for 20 min at RT. Coverslips were mounted in Mowiol, and monolayers were viewed using a Leica DM400 upright fluorescence microscope (Fig. 4C, F, and H and 5C, F, and H).
LentiCRISPR.CRISPR guide RNAs (gRNAs) were designed using the CRISPR design tool in Benchling Biology Software 2017 (Benchling, San Francisco, CA). gRNAs were cloned into LentiCRISPR v2 (73) or pRRL LentiCRISPR (74). LentiCRISPR v2 was a gift from Feng Zhang (Addgene plasmid 52961, RRID Addgene_52961; Addgene, Watertown, MA). 293β5 cells (75) were transfected with LentiCRISPR vectors, pCMV-VSV-G (envelope), and psPAX2 (packaging) using Lipofectamine 2000 (Thermo Fisher Scientific). pCMV-VSV-G was a gift from Bob Weinberg (Addgene plasmid 8454, RRID Addgene_8454). psPAX2 was a gift from Didier Trono (Addgene plasmid 12260, RRID Addgene_12260). Lentivirus-containing supernatant was harvested 72 h posttransfection and filtered through a 0.45-μm polyethersulfone (PES) syringe filter.
Lentiviral transduction and clonal selection.Human ileal enteroids were cultured in 2× infection media (modified hCCCM with 20 μM Y-27632, 20 mM nicotinamide, and 80% Wnt3a CM) for 3 day prior to transduction, dissociated as described above, and pelleted at 300 × g for 5 min at 4°C. The pellet was resuspended in 2× infection medium supplemented with 16 μg/ml Polybrene and 20 μM Y-27632, mixed with an equal volume of lentivirus-containing supernatant, plated in a 48-well plate, and centrifuged at 600 × g for 60 min at RT. The plate was transferred to a 37°C tissue culture incubator for 6 h. Transduced enteroids were transferred to a 1.5-ml tube, pelleted at 600 × g for 10 min at 4°C, and resuspended in Matrigel. Matrigel domes were overlaid with 500 μl of 2× infection medium and incubated at 37°C for 2 days. After 2 days, the medium was changed to hCCCM supplemented with 2 μg/ml puromycin. Selection medium was changed every 3 to 4 days until all untransduced control cells that were plated in parallel died. Puromycin-resistant enteroids were removed from Matrigel using cell recovery solution, centrifuged at 600 × g for 10 min at 4°C, and resuspended in DMEM containing 10% FBS, 10 mM HEPES, 1× Glutamax, 10 μM Y-27632, and 0.67 μM Jag-1. Individual enteroids were collected using a micropipette under a dissecting scope and plated in Matrigel domes in individual wells of a 96-well plate. Matrigel domes were overlaid in hCCCM without puromycin.
C2Bbe1 cells were seeded at 2.5 × 105 cells per well in a 6-well plate 1 day prior to lentivirus transduction. Lentivirus-containing supernatant was added, and the cells were centrifuged at 300 × g for 30 min, cultured at 37°C for 24 h, transferred to a 10-cm dish, and cultured for an additional 24 h. Transduced C2Bbe1 cells were selected with 20 μg/ml puromycin, and selection medium was changed every 3 to 4 days until all cells in an untransduced control died. C2Bbe1 colonies were picked using cloning discs (Bel-Art) and transferred to individual wells in a 24-well plate containing medium without puromycin. Wells were coated with collagen, as described above, to facilitate adhesion of the isolated cells. Collagen coating was discontinued after outgrowth in the 24-well plate.
Verification of CRISPR knockouts.Genomic DNA from puromycin-resistant human ileal enteroids and C2Bbe1 cells was isolated using the GeneJET Genomic DNA purification kit (Thermo Fisher Scientific). Target regions were amplified by PCR using primers listed in Table 2. PCR products were subsequently gel purified using QIAquick gel extraction kit (Qiagen), evaluated for editing by restriction fragment polymorphism (RFP) analysis (Table 2), and subjected to Sanger sequencing (Genewiz).
PCR primers and enzymes used for knockout verification
CRISPR knockout variant analysis.Target regions of 350 of 450 bp were amplified using PCR primers containing Illumina Nextera XT adapter sequences (Table 3). Amplicons were purified and barcoded according to the Illumina 16S metagenomic sequencing library preparation protocol. Briefly, amplicons were purified using AMPure XP beads (Beckman Coulter Genomics). Illumina sequencing adapters were attached to purified amplicons using the Nextera XT Index kit (Illumina), and indexed amplicons were purified using AMPure XP beads. The resulting library was normalized to 4 nM. Data were acquired on an Illumina MiSeq and analyzed using CRISPResso2 (18).
PCR primers for Illumina Nextera XT kit
Gene expression.RNA was extracted from cells using RNA-Bee (Tel-Test), and cDNA was synthesized (Promega GoScript). Target gene expression was quantified by qPCR using the primers listed in Table 4. The threshold cycle method (ΔCT) was used to determine the gene expression of target genes relative to HPRT expression after 4 (C2Bbe1) or 5 (hIE) days in culture.
qPCR primers for gene expression
Immunoblotting.hIE and C2Bbe1 cells were lysed using RIPA buffer containing 150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate, 50 mM Tris, pH 8.0, and supplemented with 1× Halt protease inhibitor cocktail (Thermo Fisher) for 30 min on ice. Lysates were centrifuged at 25,000 × g for 10 min at 4°C to pellet insoluble debris. Supernatants were denatured by boiling in 1× SDS buffer and 0.1 M dithiothreitol (DTT). Proteins were separated using SDS-PAGE, transferred to a nitrocellulose membrane, and blocked in 3% bovine serum albumin (BSA) in 0.1% Tween 20-PBS for 1 h. Blots were probed with the following antibodies: goat polyclonal anti-human caspase-1 (1:1,000; R&D), mouse monoclonal anti-caspase-4 (4B9) (1:2,000; MBL International), rabbit monoclonal anti-caspase-5 (EP876Y) (1:2,000; GeneTex), rabbit polyclonal anti-human IL-18 (1:1,000; MBL International), mouse monoclonal anti-β-actin 8H10 (1:1,000; OriGene), mouse monoclonal anti-β-actin (8H10D10) (1:20,000; Cell Signaling), and goat anti-GAPDH (glyceraldehyde-3-phosphate dehydrogenase) (1:1000, R&D). Caspase-1, caspase-4, caspase-5, and the actin loading control for caspase-5 immunoblots were probed with a horseradish peroxidase (HRP)-conjugated secondary antibody, developed with Clarity Western (Bio-Rad) or Femto (Pierce) enhanced chemiluminescence (ECL), and imaged on the Li-Cor Odyssey Fc or Bio-Rad Gel Doc. All other immunoblots were probed with fluorescently conjugated secondary antibodies and imaged on a GE Healthcare Typhoon 9400.
Statistical analysis.Statistical analysis was performed using Prism 8.3.0 or by modeling uncorrelated replicate specific random effects using the Generalized Linear Mixed Models using Template Model Builder (glmmTMB) in R version 3.6.2. Data in Fig. 2A were analyzed by a linear mixed model of log-transformed CFU on time and cell type (no interaction) with Bonferroni’s multiple-comparison test. Data in Fig. 2D were analyzed by ordinary one-way analysis of variance (ANOVA) with Tukey’s multiple-comparison test. Data in Fig. 3 were analyzed by a linear mixed model of IL-18 on genotype with Dunnett’s test for multiple comparisons to compare each genotype to the WT. Data in Fig. 3C were log transformed prior to analysis, because analysis of the untransformed data showed heteroscedasticity. Transformed CFU data for knockout cells and matched WT controls in Fig. 4A were analyzed by a ratio paired t test. Data in Fig. 4B and C were analyzed by repeated measures one-way ANOVA with the Geisser-Greenhouse correction and Dunnett’s multiple comparisons to WT control with individual variances computed for each comparison. Data in Fig. 4D were analyzed by a truncated negative binomial mixed model and adjusted using Dunnett’s test comparing WT and CASP4 KO at 7 h. Data in Fig. 4E and H were analyzed by a logistic mixed model and adjusted using Dunnett’s test comparing each KO to WT at 7 h. We categorized cells containing at least 12 (Fig. 4E) or 50 (Fig. 4H) bacteria as highly infected. Data in Fig. 4G were log transformed and analyzed by a linear mixed model comparing WT and CASP4 KO at 7 h. To test for an association between genotype and presence of infection as shown in Fig. 5A and B, Fisher’s exact test was performed for each replicate, and P values were aggregated using Fisher’s method. Data in Fig. 5C were analyzed by mixed-effect analysis with the Geisser-Greenhouse correction and Dunnett’s multiple comparisons to WT control with individual variances computed for each comparison. Replicate specific P values for the data shown in Fig. 5D and E were calculated using a one-sample t test with 1-sided alternative and aggregated using Fisher’s method. Data in Fig. 5H were analyzed by a logistic mixed model and adjusted using Dunnett’s test comparing each KO to WT. We categorized cells containing at least 50 bacteria as highly infected. Data in Fig. S1 in the supplemental material were analyzed by repeated measures one-way ANOVA with the Geisser-Greenhouse correction and Dunnett’s multiple comparisons to the dimethyl sulfoxide (DMSO) control, with individual variances computed for each comparison.
ACKNOWLEDGMENTS
We thank Danielle Williams, Youngmee Sul, and Mackenzi Oswald at the University of Washington for assistance with creation of C2Bbe1 clones and hIE enteroid culture.
Research reported in this publication was supported by the National Institute of Allergy and Infectious Diseases of the National Institutes of Health under awards R01AI104920 (to J.G.S.) and R01AI134766 (to L.A.K. and B.A.V.), by the National Institute of General Medical Sciences of the National Institutes of Health under award T32GM007270 (to M.K.H.), by the Office of the Director of the National Institutes of Health under award S10OD026741 (to J.G.S.), by the Helen Riaboff Whiteley Endowment of the Department of Microbiology of the University of Washington (to M.K.H.), and by grants from the Canadian Institutes of Health Research (to B.A.V.) and Crohn’s and Colitis Canada (to B.A.V. and L.A.K.).
B.A.V. is the Children with Intestinal and Liver Disorders (CH.I.L.D.) Foundation Chair in Pediatric Gastroenterology. Imaging studies were supported in part by the Mike and Lynn Garvey Cell Imaging Core at the University of Washington Institute for Stem Cell and Regenerative Medicine. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
FOOTNOTES
- Received 6 January 2020.
- Returned for modification 15 February 2020.
- Accepted 6 April 2020.
- Accepted manuscript posted online 13 April 2020.
Supplemental material is available online only.
- Copyright © 2020 American Society for Microbiology.
