Mice Lacking Core Components of the Inflammasome Are Susceptible to Acute C. tyzzeri Infection.

To determine the role of the inflammasome in innate responses to C. tyzzeri, we infected knockout mice lacking inflammasome activity with two transgenic C. tyzzeri strains. The first strain expresses a red-shifted luciferase (strain Ct-FluC) to visualize parasite burden in the tissue, which we recently reported (6). The second was a newly derived strain (CT-NluC), which allowed rapid measurement of fecal parasite burden. For this new strain, we used Cas9-directed homologous repair (14) to insert a cassette next to the thymidine kinase locus (Methods) to constitutively express Nanoluciferase and the fluorescent protein mCherry (Fig. 1A). The Nanoluciferase activity measured in the feces of mice infected with these parasites is tightly correlated in a linear fashion with levels of Cryptosporidium-specific DNA measured by qPCR from the same sample. Thus, Nanoluciferase activity serves as a reliable indicator of parasite shedding (Fig. 1B).

Fig. 1.

Mice deficient in core components of the inflammasome have a higher parasite burden. (A) Schematic map of the strategy used to derive the Ct-NluC strain. (B) Graph of parasite burden measured by qPCR (x axis) and luminescence (y axis). Each measurement represents the average of three technical replicates from the same mouse fecal sample, and the linear relationship has an R² value of 0.94. (C–E) C57BL/6 (WT) and Casp1/11^−/− mice were infected with 50,000 Ct-FluC oocysts each, and parasite burden was measured by whole-animal imaging (C and D) and qPCR (E). Quantitative PCR was performed on pooled collected samples with two technical replicates; mean ± SD and n = 6 mice in total. (F–H) Parasite shedding in C57BL/6 (WT) versus Casp1/11^−/− (F), Casp11^−/− (G), and Asc^−/− (H) mice infected with 50,000 Ct-NluC oocysts per mouse. Fecal luminescence measurements are from pooled cage samples shown as mean ± SD from two technical replicates. (I) Comparative infection of Casp1^fl/fl mice and Casp1^fl/fl -vilCRE mice that are deficient in caspase 1 specifically in their intestinal epithelial cells. Mice were infected with 50,000 Ct-NluC parasites, and fecal luminescence is from pooled collections shown as mean ± SD from two technical replicates per time point. Data are from two separate experiments, each with n = 8 mice. For F and G, n = 8 mice. For H, n = 7 mice in total. Samples in F–I are reported for every 2 d, but the x axis is abbreviated to 4-d increments for clarity.

Both transgenic strains readily infect wild-type (WT) C57BL/6 mice with peak parasite burden occurring at day 6 of infection and resolution after 10 d. In contrast, Casp1/11^−/− mice demonstrated a much higher early burden (Fig. 1 C–F) regardless of parasite strain or measurement used. During this acute phase, whole-body luminescence measured from Casp1/11^−/− mice was nearly 10-fold higher than that of matched C57BL/6 controls (Fig. 1D). Quantitative PCR using Cryptosporidium-specific primers on fecal pellets collected from WT and Casp1/11^−/− mice revealed higher shedding (Fig. 1E), as did fecal luciferase measurements from Casp1/11^−/− mice infected with the Ct-NluC strain (Fig. 1F). These results are consistent with experiments that used C. parvum (13) and suggest that the inflammasome is broadly required for resistance to diverse Cryptosporidium species.

Casp1/11^−/− mice lack two caspases with distinct but overlapping functions (15). Therefore, to determine whether caspase-1 or -11 (or both) were responsible for the differences in parasite burden observed, we measured infection in mice solely lacking caspase 11 (15). In contrast to the Casp1/11^−/− mice, Casp11^−/− mice were no more susceptible to infection than control mice, indicating that caspase-1 is sufficient to provide innate resistance to cryptosporidiosis and suggesting that the canonical inflammasome is likely involved (Fig. 1G). ASC is a core component of canonical inflammasome complexes that contain an NLR and caspase-1. ASC contains an N-terminal pyrin domain and a C-terminal caspase activation and recruitment domain and functions as an adaptor that bridges caspases-1 and the pathogen/danger-sensing NOD-like receptors. We note that Asc^−/− mice had an increased parasite burden during the acute phase of infection but eventually were able to control Cryptosporidium, similarly to Casp1/11^−/− mice (Fig. 1H).

We next sought to further define the cell type in which inflammasome activation was required in vivo with the consideration that Cryptosporidium specifically infects the epithelial cell layer within the intestine. To specifically isolate these cells, we used Casp1^fl/fl-vilCRE mice that lack caspase 1 only in their enterocyte populations (16). Importantly, these mice demonstrated an 18.7-fold increase in parasite burden over the unfloxed Casp1^fl/fl strain. This increase is indistinguishable from that observed in whole-body Casp1/11^−/−-deficient mice, suggesting that caspase-1 activation in enterocytes is required to enable early control of Cryptosporidium in vivo.

In order to characterize the physiological impact of inflammasome deficiency on the local response to Cryptosporidium in the gut, we conducted histological analysis using a previously established grading rubric (6). At day 5 post infection, the small intestines were removed from uninfected and infected mice, and hematoxylin-and-eosin–stained sections were evaluated by microscopy (Fig. 2A). At baseline, the structure and organization of the villi of uninfected Casp1/11^−/− mice and WT C57BL/6 controls were not significantly different, and there were no overt signs that loss of the inflammasome resulted in changes in intestinal structure. In response to infection, WT mice showed a reduction in the villus-to-crypt ratio, a marker of intestinal inflammation, but no increase in the mitotic index in the crypts (Fig. 2 A and B). In contrast, parasites were more readily detected in Casp1/11^−/− mice (arrowheads), and these mice demonstrated a lower villus-to-crypt height ratio and an increase in crypt mitotic rate, indicating a greater level of tissue turnover (Fig. 2B).

Fig. 2.

Loss of the inflammasome leads to more severe intestinal pathology. C57BL/6 (WT) mice and Casp1/11^−/- mice were infected with 50,000 Ct oocysts each, followed by histological examination of the small intestine on day 5 post infection. (A) Representative images of infected and uninfected mice with parasites highlighted by arrowheads. (B) Histological scoring from a veterinary pathologist; n = 30, with the scoring from each mouse represented as a single measurement. Mean ± SD with significance determined using one-way ANOVA and Tukey’s multiple comparisons test (*P < 0.05, **P < 0.005, ***P < 0.0005, and ****P < 0.00005).

IL-18 Release Is Required for the Innate Control of C. tyzzeri.

Inflammasome activation can lead to the processing and release of the cytokines IL-1β and IL-18. To determine the contribution of IL-1 family members in resistance to C. tyzzeri, mice that lack IL-1α, IL-1β, or IL-18 were infected with C. tyzzeri, and parasite burden was measured. WT mice and those lacking IL-1β or IL-1α showed comparable burden; however, IL-18–deficient mice showed a marked increase in acute parasite burden. Similar to Casp1/11^−/− mice, Il18^−/− mice were able to eventually control the infection (Fig. 3A). The kinetics and magnitude of infection in Il18^−/− mice closely resembled that of mice lacking core inflammasome components, suggesting that inflammasome-dependent IL-18 release is an important effector mechanism required for innate resistance to Cryptosporidium. To test this hypothesis, we next asked whether treatment with recombinant IL-18 (rIL-18) would rescue innate immune resistance in Casp1/11^−/− mice. We used a version of recombinant IL-18 that is not susceptible to inhibition by IL-18–binding protein (17). A dose of 2 μg of rIL-18 was administered intraperitoneally to WT or Casp1/11^−/− mice every 2 d following infection. In WT mice, this treatment resulted in enhanced protection. Notably, in Casp1/11^−/− mice, rIL-18 treatment was sufficient to rescue innate control of Cryptosporidium (Fig. 3B).

Fig. 3.

Cryptosporidium inflammasome activation is independent of the microbiome and drives IL-18 secretion. (A) Parasite burden from C57BL/6 (WT), Il1α ^−/−, Il1β ^−/−, and Il18^−/− mice and (C) C57BL/6 and Gsdmd^−/− mice infected with 50,000 Ct-NluC oocysts per mouse. Four mice per group; n = 16 or n = 8 mice total, respectively. Fecal luminescence was measured from pooled cage samples, and mean ± SD is shown from two technical measurements per time point. (B) Parasite burden from rIL-18–treated and untreated C57BL/6 (WT) and Casp1/11^−/− mice. Mice were infected with 50,000 Ct-NluC parasites each, and 2 μg of rIL-18 was administered by intraperitoneal injection every second day starting on day 1 post infection. Fecal luminescence was measured from pooled cage samples, and mean ± SD is shown from two technical measurements per time point. Four mice per group with n = 20 mice total. One group received a single rIL-18 injection which alone decreased parasite burden by 64.4%. (D) C57BL/6 and Casp1/11^−/− mice were cohoused for 2 wk prior to infection with 50,000 Ct-NluC oocytes per mouse. Fecal luminescence shown as mean ± SD from two technical replicates per time point; n = 12 from two separate experiments. (E) C57BL/6 and Casp1/11^−/− mice were treated for 1 wk with an antibiotic mixture (Methods) ad libitum in their cage water and then infected with 50,000 Ct-NluC oocysts per mouse. Fecal luminescence shown as mean ± SD from two technical replicates per time point; n = 16 from one experiment. (F) C57BL/6 germfree mice and matched controls were infected with 50,000 Ct-Nluc parasites by gavage, and fecal luminescence was measured from pooled cage fecal collections. Fecal luminescence shown as mean ± SD from two technical replicates per time point; n = 8 from one experiment. (G and H) ELISA measurements of IL-1β and IL-18 and from intestinal segments of infected and uninfected mice. Mice from G and H were a part of a cohort; therefore, significance was determined using one-way ANOVA with Tukey’s multiple comparisons test. Four mice per condition with n = 16 mice total and six intestinal segments harvested per mouse on day 5 post infection. Each mouse was infected with 50,000 Ct-NluC oocysts. Samples in A and C–F are reported for every 2 d, but the x axis is abbreviated to 4-d increments for clarity.

Like other members of the IL-1 cytokine family, IL-18 lacks an N-terminal signal peptide typically associated with export via the canonical secretory pathway. Instead, it is secreted via an unconventional pathway that requires processing by caspase-1 and the pore-forming activity of the gasdermin family of proteins (18). To test whether gasdermin D (GSDMD) might contribute to control of Cryptosporidium, Gsdmd^−/− and WT mice were infected and parasite burden was measured. In these experiments, Gsdmd^−/− mice had a 10-fold increase in parasite burden compared to WT controls (Fig. 3C), an increase that is comparable to transgenic mice deficient in inflammasome core components and those lacking IL-18. Together, these experiments support the model that inflammasome-deficient mice lack IL-18 activation and release through GSDMD, leading to increased susceptibility to Cryptosporidium infection.

Cryptosporidium Inflammasome Activation Does Not Require the Intestinal Microbiome.

While the inflammasome can be activated in response to pathogens, exposure to the natural flora of the gut can also engage this pathway, and the lack of inflammasome components can lead to alterations in the microbiome that subsequently affect susceptibility to intestinal inflammatory insults (19). Therefore, we conducted experiments to evaluate the impact of the microbiome on inflammasome activation in the context of cryptosporidiosis. To minimize microbiome differences, Casp1/11^−/− mice and C57BL/6 controls were cohoused for 2 wk prior to infection. This did not affect parasite burden measurements in either strain (Fig. 3D), suggesting that the observed susceptibility phenotype is not driven by transferrable differences in the intestinal microbiome of these mice. Next, we treated Casp1/11^−/− and WT C57BL/6 mice with a mixture of broad-spectrum antibiotics prior to infection to reduce the intestinal microbiome. As previously shown, mice treated with antibiotics show an increase in the total parasite burden (6, 14) (Fig. 3E). However, this increase in parasite burden was inflammasome independent, as Casp1/11^−/− mice displayed the same increase in parasite burden from antibiotic pretreatment as C57BL/6 controls (Fig. 3E). Finally, we infected normal and germfree C57BL/6 mice, thereby ensuring that the only microbe present in the latter was Cryptosporidium. Germfree mice show higher parasite shedding than normal mice, and the increase resembles the increase associated with antibiotic treatment (4.6- and 3.6-fold, respectively) (Fig. 3 F and E). Sections of the small intestine were removed from uninfected and infected mice, and these intestinal explants were assayed for secretion of cytokines. In samples from uninfected mice, there were basal levels of IL-1β and IL-18, and, in response to infection, IL-1β levels were not altered (Fig. 3G) whereas levels of IL-18 increased (Figs. 3H and 4E). Infected germfree mice also demonstrated increased IL-18 secretion when compared to uninfected germfree controls (Fig. 3H). Taken together, these data show that, in the absence of the microbiome, Cryptosporidium infection alone is capable of driving processing and secretion of IL-18 in the small intestine. Furthermore, while the presence of the intestinal microbiota antagonizes Cryptosporidium growth, this activity appears to be inflammasome independent.

Fig. 4.

NLRP6 is responsible for resistance to Cryptosporidium infection. (A) Parasite shedding from a comparative infection of C57BL/6, Aim2^−/-, Nlrp3^−/−, Nlrp1b^−/−, Nlrc4^−/−, and Nlrp6^−/− mice infected with 50,000 Ct-Nluc oocysts. Four mice per group with n = 24 mice total. Fecal luminescence was measured from pooled cage collections with mean ± SD shown from two technical replicates. Samples in A are reported for every 2 d, but x axis is abbreviated to 4-d increments for clarity. (B–D) C57BL/6 (WT) and Nlrp6^−/− mice were infected with 50,000 Ct-FluC oocysts, and parasite burden was measured by whole-animal imaging (B and C) and qPCR (D). qPCR was performed on pooled collected samples with two technical replicates; mean ± SD and n = 6 mice in total. (E) ELISA measurements of IL-18 from intestinal segments of infected and uninfected C57BL/6 and Nlrp6^−/− mice. Note that lack of NLRP6 ablates the increase in IL-18 production from Cryptosporidium infection. Significance was determined using one-way ANOVA with Tukey’s multiple comparisons test. Four mice per condition with n = 16 mice total and four intestinal segments harvested per mouse on day 5 post infection (*P < 0.05).

The Enterocyte NOD-like Receptor NLRP6 Is Required to Control Cryptosporidium.

We next sought to discover the specific pathogen receptor(s) that detect Cryptosporidium infection. In the related apicomplexans Toxoplasma and Plasmodium, the NOD-like receptors NLRP1b and NLRP3, as well as the pyrin domain containing absent in melanoma protein 2 (AIM2) have been associated with parasite control (202122–23). We tested a comprehensive set of mouse strains deficient in the expression of different NOD-like receptors and AIM2 for susceptibility to C. tyzzeri. Mice that lack AIM2, NLRP1b, NLRP3, and NLRC4 showed parasite burdens that were indistinguishable from WT mice (Fig. 4A). In contrast, NLRP6-deficient mice demonstrated increased parasite shedding and marked susceptibility to Cryptosporidium infection when compared to controls (Fig. 4A). To confirm that higher levels of parasite shedding in the feces coincided with an increase of parasite infection in the tissue, we next infected Nlrp6^−/− mice with the Ct-FLuC strain to allow for quantification of parasite burden in the gut (Fig. 4B). This revealed that Nlrp6^−/− mice had a nearly 20-fold higher tissue burden compared to WT C57BL/6 controls, yet were eventually able to control infection (Fig. 4 C and D). Finally, we tested the impact of lack of NLRP6 on intestinal IL-18 production. Comparison of uninfected and infected intestinal segments from WT or Nlrp6^−/− mice showed that, while infection in WT mice resulted in increased processing and secretion of IL-18, Nlrp6^−/− mice failed to release IL-18 at a higher level than uninfected controls (Fig. 4E). In conclusion, these data show that NLRP6 is the NOD-like receptor required to trigger an inflammasome response that enables early control of Cryptosporidium replication.

Mouse Models of Infection.

All protocols for animal care were approved by the Institutional Animal Care and Use Committee of the University of Pennsylvania. Unless otherwise specified, mice were 4 to 12 wk of age at the time of experimentation. While age and sex were controlled for each experiment, male and female mice were used indiscriminately for experiments as we have not seen a difference in infection. C57BL/6, Casp1/11^−/−, Asc^−/−, Casp11^−/−, Aim2^−/−, Nlrp1b^−/−, Nlrp3^−/−, and Il18^−/− mice were purchased from Jackson Laboratories. Nlrc4^−/−, Il1α^−/−, and Il1β ^−/− mice were provided by Sunny Shin, University of Pennsylvania, Philadelphia, PA. Nlrp6^−/− mice were obtained from Maayan Levy, University of Pennsylvania, Philadelphia, PA, but originated from the laboratory of Richard Flavell, Yale University, New Haven, CT. Casp1^fl/fl and Casp1^fl/fl-vilCRE mice were bred in-house.

Parasite Strains.

C. tyzzeri used in this experiment were originally isolated from wild house mice in Athens, GA (6), and propagated in Ifng^−/− mice initially obtained from Jackson Laboratories and bred in-house. Parasites were purified using sucrose flotation and cesium chloride gradient centrifugation.

Generation of Transgenic Parasites.

Transgenic parasites were generated using CRISPR-directed homologous recombination and described in detail (14, 52). Briefly, oocysts were induced to excyst through treatment with sodium taurodeoxycholate and incubation at 37 °C. Excysted sporozoites were transfected with vectors to drive production of Cas9 and a parasite-specific guide RNA (gRNA) and another containing the desired genetic manipulation flanked by DNA homologous to the regions surrounding the gRNA cut site. Parasites were then used to infect Ifng^−/− mice by oral gavage, which was preceded by oral gavage with sodium bicarbonate to neutralize stomach acid and maintain viability of transfected sporozoites. After 24 h, parasites were selected for by treating infected mice with paromomycin (16 mg/mL ad libitum in cage water).

Measuring Parasite Burden.

For parasites expressing Nanoluciferase, 20 mg of collected fecal sample was disrupted in 1 mL of Nanoluciferase lysis buffer by 1 min of vortexing with glass beads. Samples were centrifuged briefly at 10,000 x g to pellet debris, and 100 μL of supernatant was moved to a 96-well round bottom plate (Corning). A total of 100 μL from a 1:50 mixture of Nanoluciferase substrate:buffer (Promega) was then added to each sample. Luminescence was recorded using a ProMax plate reader with standard settings (Promega).

To measure parasite DNA in feces, DNA was purified from 100 mg of collected fecal samples using a Quick-DNA Fecal/Soil Microbe Kit (Zymo Research). We then used a Cryptosporidium-specific probe and primer set (SI Appendix, Table S1) and SsoAdvanced Universal Probes Supermix (BioRad) was used with a ViiA 7 Real-Time System (Thermo Fisher Scientific) to perform qPCR.

To quantify infection with CtFLuc parasites, mice were injected with D-luciferin subcutaneously (Gold Biotechnology) at 125 mg/kg and then anesthetized in an induction chamber using 2.5% isoflurane. After 5 min, mice were moved into an IVIS Lumina II instrument (Caliper Life Sciences), and luminescence was measured using 5 min exposure, medium binning, and 1/16 F-stop.

Histology of Infected Tissue.

Mice were killed at day 5 post infection, and the small intestine was dissected and rinsed in cold phosphate-buffered saline (PBS). This intestine was flushed with 10% neutral buffered formalin (Sigma) and then “Swiss-rolled” and fixed overnight. Fixed samples were paraffin-embedded, sectioned, and stained with hematoxylin and eosin. Slides were evaluated by a board-certified veterinary pathologist in a blinded fashion and scored using a previously establish matrix (6). Five consecutive and complete villus-crypt structures, meaning villus and crypt lumens within the same plane of section, were selected for scoring.

Measurement of IL-18 and IL-1b from Mouse Intestinal Segments.

On day 5 post infection, mice were killed, and their small intestine was dissected and opened length-wise. Intestinal segments 1 cm in diameter were then harvested, rinsed in PBS, and moved to a 24-well plate (Corning) filled with 200 μL RPMI with antibiotics and 10% fetal bovine serum. Segments were incubated at 37 °C for 12 h, and then 50 μL of the supernatant was used for IL-18 and IL-1β detection by enzyme-linked immunosorbent assay (ELISA) according to the manufacturer’s protocols (Thermo Fisher Scientific).