Dietary environmental factors shape the immune defence against Cryptosporidium infection

Cryptosporidium is a leading cause of diarrheal-related deaths in children, especially in resource-poor settings. It also targets the immunocompromised, chronically infecting people living with HIV and primary immunodeficiencies. There is no vaccine or effective treatment. While it is known from human cases and animal models that CD4 + T-cells play a role in curbing Cryptosporidium , the role of CD8 + cells remains to be defined. Using a Cryptosporidium tyzzeri mouse model, we show that gut-resident CD8 + intraepithelial lymphocytes (IELs) confer resistance to parasite growth. CD8 + IELs express, and are dependent on, the ligand-dependent transcription factor aryl hydrocarbon receptor (AHR). AHR deficiency reduced CD8 + IELs, decreased their cytotoxicity, and worsened infection. Transfer of CD8 + IELs rescued severely immunodeficient mice from death following Cryptosporidium challenge. Finally, dietary supplementation of the AHR pro-ligand indole-3-carbinol to new-born mice promoted resistance to infection. Therefore, common dietary metabolites augment the host immune response to cryptosporidiosis, protecting against disease.


INTRODUCTION
Diarrhoea contributes to nearly 11% of early childhood mortality worldwide (Liu et al., 2012). Cryptosporidium is an apicomplexan parasite that invades epithelial cells of the small intestine (Guerin and Striepen, 2020) and its infections are the second-leading cause of severe diarrhoeal events in young children in resource-poor regions (Kotloff et al., 2013). Recurrent infections are associated with malnutrition, leading to lasting effects such as growth stunting and impaired cognitive development (Kabir et al., 2021;Khalil et al., 2018;Molbak et al., 1997). Cryptosporidium is also an important opportunistic pathogen in immunocompromised individuals such as people living with HIV, transplant and chemotherapy recipients, and patients undergoing treatment for haemodialysis and cancer (Manabe et al., 1998;Sepahvand et al., 2022). Human cryptosporidiosis is usually caused by anthroponotic Cryptosporidium hominis or zoonotic Cryptosporidium parvum (Hossain et al., 2019). There is no vaccine against cryptosporidiosis. Nitazoxanide, the only FDA approved drug to treat cryptosporidiosis, has limited efficacy in immunocompetent individuals and is ineffective in HIV-AIDS patients and malnourished individuals (Amadi et al., 2002;Amadi et al., 2009).
Although some progress has been made towards the development of new therapeutics (Manjunatha et al., 2017), this has been hindered by a lack of physiologically relevant model systems of cryptosporidiosis (Manjunatha et al., 2017;Marzook and Sateriale, 2020). We previously developed a mouse model of cryptosporidiosis with a natural mouse-infecting species, Cryptosporidium tyzzeri, which recapitulates the natural course of infection and intestinal pathology of human disease (Sateriale et al., 2019).
Increased susceptibility of patients with primary immunodeficiencies and experimental infections of human volunteers suggests an immune system-mediated protection from cryptosporidiosis (Cohn et al., 2022). Furthermore, most children living in endemic regions develop protective immunity to subsequent infections (Kabir et al., 2021;Kattula et al., 2017). Early studies on athymic mice revealed T-cells as important regulators of C. parvum infection (Heine et al., 1984). Increased infection burden in HIV-AIDS patients with low CD4+ T cell counts also highlights the importance of interferon-g (IFNg)-producing CD4 + T cells (Kaplan et al., 2009). Mice lacking T-cells (but not B cells) are unable to control a C. tyzzeri infection (Sateriale et al., 2019). Furthermore, mice without mature T and B cells, and those lacking IFNg, cannot elicit vaccination-mediated protection during a secondary parasite infection (Sateriale et al., 2019). Mice harbouring a commensal strain of C. tyzzeri elicit both innate and adaptive immune responses, with early evidence hinting at elevated CD8 + T cells in infected mice (Russler-Germain et al., 2021). CD8 + T cells are highly enriched in the intestine and notably heterogeneous in phenotype and function. 'Conventional' mucosal T cells express TCRαβ together with CD4 or CD8αβ as TCR coreceptors and reside in the lamina propria. 'Nonconventional' mucosal T cells, expressing either TCRαβ or TCRγδ and typically also CD8αα homodimers, are prevalent in the mucosal epithelium (Hayday et al., 2001;Sujino et al., 2016). Thymic-derived natural intraepithelial lymphocytes (IELs) express CD8αα (TCRαβ + CD8αα, TCRαβ + CD4 + CD8αα and TCRγδ + CD8αα), whereas peripherally induced IELs express CD8αβ (TCRαβ + CD8αβ) (Cheroutre et al., 2011;Hoytema van Konijnenburg and Mucida, 2017).
CD8 + IELs cells constantly scan epithelial cells for injury or infection and are considered primary responders to epithelial damage (Hoytema van . CD8 + IELs are elevated in calves and mice infected with C. parvum (Chai et al., 1999;Pasquali et al., 1997). Nevertheless, factors that influence CD8 + IEL-mediated immunity and mechanisms by which CD8 + IELs confer resistance to Cryptosporidium are unknown. Most CD8 + T cells exhibit a tissue-resident memory phenotype (Sasson et al., 2020) and their maintenance depends on stimulation of the aryl hydrocarbon receptor (AHR) by endogenous ligands such as tryptophan-derived phytochemicals, microbial metabolites, and indole derivatives (Li et al., 2011) . AHR is a ligand-dependent transcription factor abundantly expressed in barrier tissues such as the gut, skin, and lungs. AHR is expressed by various cells in the gut including barrier epithelium, endothelium, and immune cells, and is hence a key factor in maintaining gut barrier integrity (Stockinger et al., 2021). AHR deficiency or depletion of AHR ligands increases susceptibility to bacterial infection in the colon (Stockinger et al., 2021) and also contributes to colon tumorigenesis (Metidji et al., 2018).
Cryptosporidium infects and replicates inside small intestinal epithelial cells, causing villus blunting and crypt hyperplasia and thereby significant gut damage (Sateriale et al., 2019). In human volunteers infected with C. hominis or C. parvum, increased fecal indole levels prior to infection correlated with decreased parasite burden (Chappell et al., 2016), indicating a protective function of indoles in cryptosporidiosis through an unknown mechanism. We wondered if these indoles were working via AHR-mediated gut protection.
To test this, we created a novel C. tyzzeri reporter strain expressing luminescent and fluorescent proteins utilizing an isolate of C. tyzzeri from the Czech Republic (Ct-CR2206, shortened here to Ct-CR) (Kvac et al., 2013). Here we show that in immunocompetent wildtype mice, Ct-CR altered epithelial differentiation and triggered an expansion of CD8 + IELs, and IELs conferred protection against infection when transferred to immunodeficient mice.
Immune cell-specific deletion of AHR or the deprivation of AHR ligands in mice greatly depleted CD8 + IELs. Furthermore, dietary supplementation of AHR ligands to nursing mothers and their weaned pups provided prophylactic defence against infection. This highlights the opportunity for diet-based therapeutic interventions to treat this debilitating disease in humans and ruminant animals.

Cryptosporidium tyzzeri infects the ileum and affects epithelial cell differentiation.
The Cryptosporidium tyzzeri strain Ct-CR2206 was originally isolated from a wild mouse in the Věrušičky municipality of the Czech Republic (Kvac et al., 2013). We sequenced the genome of Ct-CR2206 (shortened here to 'Ct-CR') using Illumina short read sequencing. Reads were mapped using the full C. tyzzeri-UGA55 genome as a reference ( Figure 1A). In total, 13139 SNPs and 2983 insertion-deletion events were detected between the two strains (see Supplementary Table S1 for a further breakdown). We genetically engineered Ct-CR to express the fluorescent mNeonGreen protein, along with a Nanoluciferase and Neomycin resistance cassette for easy parasite detection and transgenic selection, respectively ( Figures   1B and 1C). Parasite replication over time could be tracked by assaying Nanoluciferase activity in mouse fecal pellets ( Figure 1D).
Ct-CR infects the small intestine but not the colon ( Figure 1E). To determine which specific region in the small intestine harboured parasites, we infected C57BL/6 wild-type (WT) mice and measured luciferase levels from duodenal, jejunal, and ileal tissue every other day up to day 14 post-infection . This revealed the ileum to be the major site of Ct-CR expansion within the small intestine ( Figure 1F). This infection location is similar to previously described C.tyzzeri strains and to what is seen during human cryptosporidiosis (Russler-Germain et al., 2021;Sateriale et al., 2019). Luciferase levels in fecal samples from the same mice reflected the parasite burden in the ileum ( Figure 1G). Therefore, we used readings from fecal samples as representative measures of the Ct-CR burden in the mouse ileum.
IFNg is key to controlling Cryptosporidium growth, for both C. parvum (Griffiths et al., 1998;Gullicksrud et al., 2022;Robinson et al., 2001) and the previously isolated US C. tyzzeri strain, UGA55 (Sateriale et al., 2019). Consistently, infection with the Ct-CR strain led to a gradual increase in Ifng transcripts in the ileum ( Figure 1H), suggesting that the Ct-CR strain is pathogenic and induces inflammation in the infected area. Sca-1 (also known as Ly6a) is a marker of mouse intestinal epithelial injury, initially identified in the context of colitis (Yui et al., 2018) and staining of ileal sections showed a significant increase in Sca-1 expression in the villi of infected mice ( Figure 1I). To study how Cryptosporidium changes the epithelial cell composition in the small intestine, we purified epithelial cells at the peak of infection (DPI-6) and compared their gene expression with that of uninfected controls. Epithelial cells from infected mice showed reduced expression of the stem cell signature (Olfm4, Clu, Sox9) and reduced expression of markers for enteroendocrine cells (Chga) and Tuft cells (Dclk1) ( Figure   1J). However, no change was seen in expression of the goblet cell marker Muc2 ( Figure 1J).
In contrast, there was an increase in the enterocyte marker (Car4) ( Figure 1J). Therefore, the Ct-CR strain used in this study is pathogenic in nature and causes inflammation, epithelial cell injury and alters the cellular composition of the small intestinal epithelium.
Haematopoietic cell specific aryl hydrocarbon receptor-deficient mice are susceptible to C. tyzzeri infection.
AHR is a transcriptional regulator of genes involved in anti-microbial defence and intestinal epithelial differentiation. It has been shown to play a protective role during colonic bacterial infection and chemically-induced colon damage (Metidji et al., 2018;Shah et al., 2022;Stockinger et al., 2021). To determine whether AHR executes similar disease protective functions in the small intestine, we infected full body AHR knockout (AHRKO Body ) mice and cohoused WT littermate controls with Ct-CR. AHRKO Body mice displayed increased parasite burdens from the beginning and maintained high parasite loads throughout the course of infection (Figure 2A left panel), also assessed by area under the curve (AUC) (Figure 2A right panel). Having established the importance of AHR for Cryptosporidium infection control, we next wanted to know which cells were playing a role in this defence. Using AHR-reporter mice expressing AHR td-Tomato, we determined that the majority of epithelial cells express AHR ( Figure 2B), consistent with previous observations (Diny et al., 2022). Since Cryptosporidium only infects epithelial cells, we first challenged mice with an intestinal epithelial cell (IEC)specific AHR knockout (Vil-Cre AHR fl/fl termed AHRKO Epithelium ) with Ct-CR to probe whether IEC-intrinsic AHR expression influenced parasite burden. Surprisingly, we found that epithelial AHR signalling was dispensable for control of parasite growth ( Figure 2C). AHR is also expressed by immune cells in the gut ( Figure 2D). We next asked whether AHR expression in immune cells is required to limit Ct-CR. To achieve this, we crossed hematopoietic cell-specific Vav-Cre mice with AHR fl/fl mice to produce immune cell-specific AHR-deficient mice (AHRKO Immune ). Consistently, mice with AHR deficiency in their immune cells were highly susceptible to Ct-CR infection compared with littermate controls ( Figure 2E). Vav-Cre is also expressed by endothelial cells (Joseph et al., 2013), which express high levels of AHR (Diny et al., 2022). Therefore, AHRKO Immune mice could simultaneously delete AHR in both their immune and endothelial cells. To account for a potential role for AHR in anti-Cryptosporidial defence via the endothelium only, we made use of a tamoxifen-inducible Cre line controlled by the cadherin 5 (Cdh5) promoter to delete AHR selectively in endothelial cells.
Cdh5 CreERT2 AHR fl/fl (AHRKO Endo ) and littermate control mice were administered tamoxifen orally. Five days post-tamoxifen treatment, (AHRKO Endo ) mice were infected with Ct-CR, alongside a cohort of AHRKO Immune mice in the same experiment. In endothelial cell-specific AHR knockout mice, C. tyzzeri levels were comparable to those in littermate controls, whereas the infection burden was again increased in AHRKO Immune mice ( Figure 2F). C. tyzzeri levels in the ileum of AHRKO Immune mice were also similar to the levels in total body-AHR deficient mice (AHRKO Body ) ( Figure 2G). Hence, these results narrowed down the importance of AHR expression specifically in immune cells to control an intestinal C. tyzzeri infection. We also noted that the increased parasite burden in AHRKO Immune mice was associated with increased expression of enterocyte markers at the expense of stem cell and Tuft cell markers ( Figure   2H). Taken together, AHR signalling in immune cells is vital to regulate the ability of Cryptosporidium to grow in intestinal epithelial cells of the small intestine.

AHR-expressing CD8a intraepithelial lymphocytes respond to C. tyzzeri parasite infection.
Lymphocytes in the small intestine are spatially organised into intraepithelial and lamina propria layers. The intraepithelial layer is enriched with cytotoxic CD8 + intraepithelial lymphocytes (IELs), which are the primary responders to epithelial damage due to their close proximity to epithelial cells (Cheroutre et al., 2011). Immunofluorescence images of mouse ileal sections show CD3 + IELs in close contact with C. tyzzeri-infected epithelial cells ( Figure   3A). Using AHR-tdTomato reporter mice, we found that all the major IEL subsets, including TCRab + CD8aa + , TCRab + CD8ab + , and TCRgd + CD8aa + express AHR-tdTomato (Figures 3B and Figure S1). TCRab and TCRgd IELs from Ct-CR infected mice had increased expression of Ki67, indicative of a hyperproliferative state ( Figure 3C), which corresponded with increased numbers of IELs in Ct-CR infected mice ( Figure 3D). Moreover, IFNg expressing CD8 + IELs were significantly increased following Ct-CR infection ( Figure 3E). CD8 + IELs are quiescent at steady state with minimal proliferation (Figures 3C), however they rapidly mount a cytotoxic response to kill target cells through effector proteins such as granzymes (Konjar et al., 2018). Indeed, all 3 major types of CD8 + IELs express granzyme-B at steady state ( Figure 3F).
Therefore, IELs sense epithelial invasion by C. tyzzeri in a similar manner to other infection settings (Hoytema van Konjar et al., 2018) and respond by producing IFNg and granzyme-B, which are effector mediators of IEL cytotoxicity.

AHR-dependent CD8a IELs confer resistance to C. tyzzeri infection in immunodeficient mice.
AHR expression and signalling is essential for both TCRab + and TCRgd + IEL survival, maintenance, and effector function (Dean et al., 2023;Li et al., 2011;Panda et al., 2023). In line with these findings, the percentage, and total numbers of natural CD8aa IELs recovered from naïve AHRKO immune mice were significantly lower compared to WT littermates whereas the numbers of TCRab + CD8ab + IELs were similar ( Figures 4A-4D). During infection with Ct-CR, TCRab + CD8aa + and TCRgd + CD8aa + IELs were similarly reduced but TCRab + CD8ab + IELs increased in AHRKO immune mice. (Figures 4E and 4F). AHR-deficient CD8a IELs were hyperproliferative ( Figures 4G and 4H). However, they exhibited diminished cytolytic activity indicated by decreased Granzyme-B levels ( Figures 4I and 4J). Therefore, it is likely that a decrease in cytotoxic CD8 + IELs in AHR deficient mice contributed to the increased Ct-CR burden.
To directly assess the protective function of CD8 + IELs, we purified CD8 + IELs from the small intestine of naïve WT mice and transferred them intravenously into immunodeficient Rag2-IL2Rg-CD47 triple knockout mice which are usually highly susceptible to infections (Gullicksrud et al., 2022). Four weeks after IEL transfer, the mice were infected with Ct-CR ( Figure 5A). While untreated immunodeficient mice experienced increased severity with significant reduction in survival ( Figure 5B), triple knockout mice that received WT CD8 + IELs survived and had reduced parasite burdens in the feces ( Figure 5C) and ileum ( Figure 5C right panel). In conclusion, AHR expression in CD8 + IELs is required for their maintenance and cytotoxicity, and these CD8 + IELs alone are sufficient to control Ct-CR parasite growth in vivo.

Dietary AHR ligands confer resistance to C. tyzzeri infection.
AHR is a ligand-activated transcription factor. In the gut, a major source of these ligands are dietary tryptophan-derived phytochemicals and tryptophan metabolites produced by the microbiota(Gomez de Aguero et al., 2016). Since AHR expressing CD8 + IELs are key to anti-Cryptosporidial activity, we wanted to determine if this function is influenced by AHRligand availability. The phytochemical indole-3-carbinol (I3C), enriched in cruciferous vegetables, is an AHR pro-ligand that is converted to high-affinity AHR ligands upon exposure to stomach acid (Bjeldanes et al., 1991). We first asked whether dietary supplementation with I3C influences CD8 + IELs. 3-week-old WT mice were fed either control phytochemical-free synthetic AIN93M diet ('purified control diet') or I3C (1000 mg/kg) enriched diet ('I3C diet') for 2 weeks and infected with Ct-CR while they continued on the same diets ( Figure 5D). I3C diet fed mice had reduced parasite burdens compared with the mice fed the control diet ( Figure 5E). Enumeration of CD8 + IELs in these mice indicated that I3C diet supplementation robustly increased all subsets of CD8 + IELs in the small intestine ( Figure 5F). To demonstrate that I3C diet mediated protection from Ct-CR infection is T-cell dependent, we fed Rag2 knockout mice with purified control or I3C diet for 2 weeks. Infection with Ct-CR resulted in similar burden of parasite in both groups of mice independent of exposure to I3C diet ( Figures   5G and 5H). We then investigated the significance of an I3C-enhanced diet in ameliorating C. tyzzeri infection in a mouse model that recapitulates early childhood infection in humans.
Since AHR signalling is known to impact fertility (Hernandez-Ochoa et al., 2009), pregnant WT females were maintained on normal chow diet until they gave birth. We then changed the diet of WT dams from normal chow to either a purified control diet or an I3C-enriched diet. 3 weeks later, pups were weaned on the same diet that their mothers had received and were infected with Ct-CR ( Figure 5I). WT pups that grew up on the purified control diet were far more susceptible to infection than WT pups that grew up on the I3C diet ( Figure 5J). Thus, prior exposure to I3C protects mice from Cryptosporidium infection.
Taken together, these data underscore the potential of dietary AHR ligands to modulate cytotoxic immune defence against C. tyzzeri, leading to improved elimination of infected epithelial cells, and thus offering a therapeutic avenue to prevent or treat cryptosporidiosis by enhancing immunity to Cryptosporidium through dietary interventions.

DISCUSSION
While ubiquitously prevalent, Cryptosporidium is particularly problematic in resource-poor settings. Malnutrition is endemic to those same settings, which is in turn a risk factor for Cryptosporidium infections and chronic diarrhoea (Iannotti et al., 2015). Malnutrition and immunodeficiency increase the risk of recurrent Cryptosporidium infections, which contributes to increased morbidity and mortality (Checkley et al., 2015). The findings presented in this study using a mouse model of infection offers a proof of concept for the potential use of dietary AHR ligands such as I3C to curb this vicious cycle of chronic infections, diarrhoea, and malnutrition in humans and farm animals. Many resource-poor regions across the world already rely on ready-made food formulations such as RUTF (Ready-to-Use Therapeutic Food) to treat children suffering from severe wastage and malnutrition, often as the result of infections, with varying degrees of success (Schoonees et al., 2019). Several human clinical trials show that I3C is safe for human consumption in adults (Naik et al., 2006;Reed et al., 2005;Wong et al., 1997). I3C supplements are commercially available on the market and thus it is conceivable for them to be included in RUTF formulations following controlled human challenge trials confirming their therapeutic potential. Intriguingly, it has also been shown that nursing mice can pass on AHR ligands to their newborn (Lu et al., 2021).
Therefore, there is also scope for dietary prophylactic interventions to be made at the level of nursing mothers, especially since Cryptosporidium infections are most severe in children less than one year old.
In laboratory settings, younger mice are also more susceptible to Cryptosporidium infections.
Here, we have shown the importance of cytotoxic CD8 + IELs in controlling this infection in an IFNg is a key cytokine responsible for early protection from C. tyzzeri infection (Gullicksrud et al., 2022;Sateriale et al., 2019). Interestingly, natural CD8aa IELs do not produce as much IFNg as their induced CD8αβ counterparts although both robustly respond to C. tyzzeri infection. However, both TCRab + and TCRgd + IELs expressed cytotoxic granular protein granzyme-B in their quiescent and activated states. A heightened state of activation with increased cytotoxic potential and rapid proliferative capacity could facilitate swift killing of C.
tyzzeri-infected epithelial cells by IELs. Indeed, AHR deficient mice (AHRKO Body and AHRKO Immune ) that lack these IELs have elevated parasite burdens from DPI-2 onwards, suggesting that in the absence of cytotoxic CD8 + IELs C. tyzzeri robustly establish an infection in the small intestine. In contrast to CD8aa IELs, induced CD8αβ IELs can form antigen-specific long-lived memory cells (Soerens et al., 2023). Future studies should explore the antigenspecificity of conventional TCRab + CD8αβ IELs during primary and secondary infections and narrow down their role in long-term protection against subsequent Cryptosporidium infections.
We have previously established C. tyzzeri as a model of Cryptosporidium infection to study the relationships between a pathogen and its natural host (Sateriale et al., 2019). Our work showed parasite infections caused characteristic villus blunting, a decrease in villus-crypt height ratios, as well as an increase in mitotic events in the epithelium of the small intestine (Sateriale et al., 2019). Here, we further characterised the pathological effects of the C. tyzzeri-CR strain on the mouse small intestine and found that it produces many hallmarks of epithelial damage. There is increased expression of Ifng at the tissue level, a loss of markers of differentiated cells such as Chga and Dclk1, as well as an increase in the enterocyte marker Car4. Most notably, Sca-1 protein levels significantly increased in infected epithelial tissue, a notable marker of epithelial damage previously seen in mouse models of colitis (Shah et al., 2022;Yui et al., 2018). Increased Sca-1 expression is also caused during infection by the intestinal parasitic helminth H. polygyrus in an IFNg-dependent manner and is indicative of a reversion to a fetal-like state in tissue (Nusse et al., 2018) . We hope that this study paves the way for future work understanding how this parasite specifically damages the small intestine, potentially triggering a re-wiring of the epithelial regenerative response.
Taken together, we have further uncovered host responses to C. tyzzeri and revealed the role of environmental sensor AHR and its natural ligands in conferring protection from Cryptosporidium infection by modulating gut-resident cytotoxic lymphocytes using a welldefined genetically tractable mouse model. This study extends our understanding of AHR signalling and its importance in maintaining gut barrier integrity in the small intestine and suggests a way forward towards future therapeutic strategies to control the severity of Cryptosporidium infections.

Generation of transgenic parasites
Transgenic C. tyzzeri-CR parasites were created using methods described previously (Pawlowic et al., 2017;Sateriale et al., 2020). Briefly, 1.5 x 10 6 oocysts were bleach-treated and incubated at 37 o C in 0.75 % sodium taurocholate for one hour to promote excystation.
Sporozoites were transfected with a plasmid expressing Cas9 and guide targeting the C. tyzzeri TK gene, along with a repair template carrying Nanoluciferase, mNeonGreen, and Neomycin resistance genes. IFNg -/mice were orally gavaged with these transfected sporozoites, following a sodium bicarbonate oral gavage a few minutes prior. Mice were given Paromomycin in their drinking water to select for transgenic parasites and Nanoluciferase levels in fecal samples were tracked over time.

Measuring parasite burden in tissues and feces
Nanoluciferase readings from fecal samples were performed as a proxy for parasite burdens as previously described using the NanoGlo Luciferase reporter assay (NanoGlo Luciferase kit, Promega, N1150) (Vinayak et al., 2015). Small intestinal tissue sections were cut, weighed, and processed in a similar manner to fecal samples.

Isolation of parasites
Parasites were isolated from daily mouse fecal collections by sucrose flotation followed by cesium chloride-mediated density gradient centrifugation as described previously (Pawlowic et al., 2017). Briefly, 5-7 days' worth of fecal samples were made into a slurry in tap water and filtered through a sieve of 250 um pore size. Filtrates were mixed 1:1 with a sucrose solution (1.33 specific gravity) and centrifuged at 1000 g at 4 o C. The supernatant was washed in cold water (1:100), centrifuged again at 1500 g at 4 o C and resuspended in cold saline. This suspension was carefully laid over a cesium chloride solution (1.25 M) in a 1:1 ratio and centrifuged at 16,000 g at 4 o C to produce an interphase containing suspended oocysts. These were collected and stored in cold saline for future use.

Isolation and transfer of intraepithelial lymphocytes
IELs were isolated from the small intestine.

Bioimage analysis of Sca-1 immunofluorescence staining:
OCT embedded cryosections were stained with Sca-1 as mentioned above. Quantification of Sca-1+ areas in the small intestinal epithelium was performed blindly using QuPath software (Bankhead et al., 2017). The epithelial layer of the villi in the small intestine was delineated manually using the polygon or brush tool in QuPath. 6 villi per small intestine were analyzed.
Using the automated quantification tool, mean fluorescence intensity per area unit was estimated. This was repeated in sections stained without primary antibody as controls. The average fluorescence intensity values from the control sections were subtracted from the fluorescence intensity values obtained from each villi in the original samples. The fluorescence intensity values from 6 villi per small intestine were averaged, to report the mean GFP intensity per mouse.

Quantitative Real-time Polymerase Chain Reaction:
RNA from the gut tissues and epithelial cells was extracted using TRIzol RNA purification kit (Invitrogen; Cat# 9738G). cDNA was synthesized with high-Capacity cDNA Reverse Transcription Kit (Thermofisher; Cat# 4368814) and qRT-PCR was performed using Taqman 2x universal PCR master mix (Thermofisher; Cat# 4318157) with appropriate primer sets. Data was normalized to housekeeping genes HPRT or B2M.

Acknowledgements:
We would like to acknowledge the Science Technology Platforms at the Francis Crick Institute.
We thank the Biological Research Facility for breeding and maintenance of our mouse strains,  Luciferase Units. Error bars, mean + SEM. ns-not significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 as calculated by t-test or one-way ANOVA with Tukey post-test.