CRTAM Protects Against Intestinal Dysbiosis During Pathogenic Parasitic Infection by Enabling Th17 Maturation

The gastrointestinal tract hosts the largest collection of commensal microbes in the body. Infections at this site can cause significant perturbations in the microbiota, known as dysbiosis, that facilitate the expansion of pathobionts, and can elicit inappropriate immune responses that impair the intestinal barrier function. Dysbiosis typically occurs during intestinal infection with Toxoplasma gondii. Host resistance to T. gondii depends on a potent Th1 response. In addition, a Th17 response is also elicited. How Th17 cells contribute to the host response to T. gondii remains unclear. Here we show that class I-restricted T cell-associated molecule (CRTAM) expression on T cells is required for an optimal IL-17 production during T. gondii infection. Moreover, that the lack of IL-17, results in increased immunopathology caused by an impaired antimicrobial peptide production and bacterial translocation from the intestinal lumen to the mesenteric lymph nodes and spleen.


INTRODUCTION
The gastrointestinal tract hosts the largest collection of commensal microbes in the body, which impact host metabolism as well as development and regulation of the immune system (1). Pathogenic infections at this site can cause significant perturbations in the microbiota, known as dysbiosis, that facilitate the expansion of pathobionts, elicit inappropriate metabolic and immune responses, and damage the barrier function of the intestine, ultimately causing pathology (2)(3)(4)(5). This condition typically occurs during intestinal infection by Toxoplasma gondii (T. gondii), a widespread protozoan parasite of animals that also infects humans through the ingestion of oocysts contaminating water or food, or consumption of undercooked meat harboring tissue cysts (6). Host resistance to T. gondii depends on a potent IL-12-dependent IFN-γ response, which is largely mediated by Th1 cells (7)(8)(9)(10). However, following oral infection with tissue cysts, the Th1-induced IFN-γ response to T. gondii destroys Paneth cells and blunts their capacity to produce antimicrobial peptides (AMPs), thereby impeding control of invasive Enterobacteriaceae (11)(12)(13)(14)(15).
In addition to triggering the IL-12-Th1-IFN-γ axis, oral infection by T. gondii also elicits a Th17 response. How Th17 cells contribute to the host response to T. gondii remains unclear. In two studies, abrogation of IL-17 signaling in Il17ra −/− and Il17a −/− mice resulted in increased susceptibility to T. gondii infection (16,17). However, another study reported that Il17ra −/− mice as well as B6 mice treated with a blocking anti-IL-17A antibody are more resistant to T. gondii infection (18). Finally, infection with a high dose of T. gondii cysts induced gut immunopathology independent of IL-17, but dependent on IL-22 (19).
We previously found that Th17 response to T. gondii depends on the cell surface molecule class I-restricted T cell-associated molecule (CRTAM) (20). CRTAM was originally described on activated CD8 + T cells and NK cells (21). It binds cell adhesion molecule 1 (CADM1), which is expressed on many myeloid and epithelial cells (22)(23)(24)(25)(26)(27). We found that CRTAM is also expressed on intestinal intraepithelial CD4 + T cells upon activation. Moreover, we observed that Crtam −/− and Cadm1 −/− mice had a selective defect in Th17 compared to wild-type (WT) mice during infection with the type II T. gondii strain Prugnaud (Pru), which is relatively non-pathogenic. However, they developed an effective Th1 response and cleared gut infection as effectively as WT mice (20). Thus, Th17 deficiency had no obvious consequences in this model of infection.
Here we challenged Crtam −/− mice with an oral inoculation of tissue cysts of the type II ME49 strain of T. gondii, which is more pathogenic than the Pru strain we used previously, despite sharing the same genotype (28). Consistent with our previous study, Crtam −/− mice controlled intestinal infection; they developed an effective Th1 response, but their Th17 response was impaired. Despite being able to control infection, Crtam −/− mice suffered more pathology and many succumbed following infection. Remarkably, certain AMPs that are known to be induced by IL-17 including S100A8, S100A9, and beta defensins, were drastically reduced in the intestines of Crtam −/− mice. As a result, mice lacking CRTAM were unable to control T. gondii-induced dysbiosis and subsequent bacterial translocation to the mesenteric lymph nodes and spleen. Paneth cell-derived AMPs, such as alpha defensins, were not affected, suggesting that control of T. gondii-induced dysbiosis requires a broad spectrum of AMPs. Although IL-17-producing CD4 + T cells were less abundant in Crtam −/− mice than in WT mice, CD4 + T cells expressing RAR related orphan receptor t (Rorγt), the master transcription factor driving Th17, were equally represented; this suggests that CRTAM is required for terminal maturation of Th17 and acquisition of effector function rather than for Th17 lineage commitment. We conclude that CRTAM enables an optimal Th17 host response to pathogenic parasitic infections that is required for controlling dysbiosis and bacterial translocation associated with infection.

T. gondii Infection Causes Marked
Intestinal Pathology in Crtam −/− Mice Given our previous observation that CRTAM has a limited impact on the response to a non-pathogenic strain of T. gondii, we wanted to re-examine CRTAM function in the context of intestinal infection by the more pathogenic ME49 strain. Crtam −/− and WT mice were infected orally with 10 cysts of T. gondii ME49, which is capable of causing death at higher inoculum. This line of ME49 also expresses luciferase and therefore can be visualized by bioluminescence imaging of the whole mouse (29). Crtam −/− mice lost more weight and more of them died following infection (Figures 1A,B), although not statistically significant, Crtam −/− mice show a trend to control parasite replication better than WT ( Figure 1C). Histological analysis of the intestine revealed that Crtam −/− mice had more ileal pathology than did WT mice after infection with T. gondii; Crtam −/− mice displayed loss of epithelial integrity, mucosal and submucosal edema, altered villus architecture, and moderate inflammatory infiltrates (Figures 1D,E), the colon was equally affected in WT and Crtam −/− mice ( Figure 1E; Figure S1). We conclude that the cause of increased lethality in Crtam −/− mice is exacerbated pathology rather than uncontrolled infection.

CRTAM Deficiency Impairs the Th17 Response to T. gondii
We next sought to investigate the mechanisms by which CRTAM counteracts intestinal pathology. No significant differences were observed in the frequencies of CD4 + T cells producing IFN-γ and IL-22 that were present within intraepithelial (IEL) and lamina propria (LPL) lymphocytes isolated from the small intestine of WT and Crtam −/− mice after infection with T. gondii; however, considerably fewer CD4 + T cells producing IL-17 were detected in Crtam −/− mice (Figures 2A-C). Moreover, cytokine profiles of the ilea of WT and Crtam −/− mice after infection with T. gondii confirmed a strong reduction in the abundance of IL-17A and IL-17F mRNAs in Crtam −/− mice, while the amounts of IFN-γ, IL-22, and IL-10 mRNAs were not significantly different between WT and Crtam −/− strains. Together, these results suggest that CD4 + T cells require CRTAM in order to produce IL-17 during T. gondii infection.

Lack of CRTAM Curtails the Production of AMPs That Restrain the Microbiota
Given that the dysbiosis incurred during infection with T. gondii has previously been attributed to IFN-γ-induced death of Paneth cells and consequent loss of AMPs that results in an increase of Enterobacteriaceae (11), we assessed expression of various AMPs in the ilea of Crtam −/− and WT mice infected with ME49. Significantly less S100a8, S100a9, Regenerating isletderived 3 gamma (Reg3g) and Defensin beta 3 (Defb3) mRNA was detected in Crtam −/− mice than in WT mice at day 5 post-T. gondii infection. In contrast, Defensin alpha 1 and alpha 2 were equally expressed ( Figure 3A). 16S rRNA sequencing analysis further showed that although naïve WT and Crtam −/− mice had a similar microbiome (Figures 3B-D; Table S1), an increase in Enterobacteriales ( Figure 3C; Table S1) in particular γ-proteobacteria ( Figure S2) was observed in infected Crtam −/− mice compared to WT ( Figure 3B). Moreover, more bacteria were present in mesenteric lymph nodes and spleens from Crtam −/− mice than in those from WT mice at day 8 postinfection ( Figures 3E,F), suggesting that this partial defect in AMPs is sufficient to allow breach of the barrier and bacterial translocation, which may contribute to the death of Crtam −/− mice. Remarkably, S100A8, S100A9, and Defb3 are primarily induced by IL-17, whereas defensin alpha 1 and alpha 2 are chiefly produced by Paneth cells. Our results suggest that the Th17 response to pathogenic challenge with T. gondii prevents dysbiosis by inducing AMPs for the containment of commensal bacteria.

IL-17 Prevents Bacterial Translocation During T. gondii Infection
To validate the importance of IL-17 for the production of AMPs and control of bacterial translocation during T. gondii infection, WT mice were treated with either anti-IL-17A and anti-IL-17F blocking antibodies or an isotype control and infected with 10 cysts of T. gondii ME49 (Figure 4). Mice that received the anti-IL-17 antibodies expressed less AMP mRNA and had more bacterial translocation to the mesenteric lymph nodes, confirming that IL-17 is essential for preventing dysbiosis. Of note, bacteria were not found in the spleens of mice treated with the anti-IL-17 blocking antibodies, suggesting that other cytokines or chemokines may contribute to barrier protection.

CRTAM Engagement Is Required for Terminal Differentiation of Th17
Th17 differentiation is a multi-step process that requires induction of the master Th17 transcription factor Rorγt in naïve CD4 + T cells (30,31), followed by maturation of Rorγt + CD4 + T cells into IL-17 secreting cells, which depends on exposure to cytokines, such as IL-23 (32). To understand when CRTAM is required during Th17 development, we assessed expression of Rorγt in lamina propria CD4 + T cells during T. gondii infection in WT and Crtam −/− mice. Rorγt + T cells were present at similar frequencies in WT and Crtam −/− mice (Figures 5A,B), indicating that CRTAM is dispensable for Th17 lineage specification.
Since the effector function of Rorγt + CD4 + T cells is controlled by a subset of regulatory T cells that express both Forkhead box p3 (Foxp3) and Rorγt and suppress intestinal inflammation (33,34), we asked whether lack of CRTAM facilitates the development of Foxp3 + Rorγt + T cells. However, we didn't see an increase in Foxp3+Rorγt+T cells in Crtam −/− mice (Figures 5A,C). These results suggest that CRTAM does not impact modulation of Th17 through Treg induction.

DISCUSSION
Host control of T. gondii infection has been shown to rely on a strong Th1 response (6,7). Our study demonstrates that the Th17 response is important for host survival to intestinal infection by a more pathogenic type II strain of T. gondii. By secreting IL-17, Th17 support the production of AMPs, that prevents the dysbiosis and bacterial translocation associated with pathogenic infection. How Th17 response impacts oral infection by T. gondii has been controversial. While some studies have reported a protective role for IL-17 signaling during T. gondii infection (16,17), other studies have found that IL-17 can either promote or is irrelevant for immunopathology (18,19). We postulate that these discrepancies may be due to differences in the T. gondii strains used and doses administered in each study, or differences in the microbiota of the mice used. The ability of the various strains used to disrupt the microbiota and the susceptibility of the microbiota to perturbations induced by T. gondii clearly have a great impact; because the microbiota in genetically identical mice varies enormously in different mouse facilities, replicating any of these studies is challenging.
It has been reported that T. gondii-induced dysbiosis depends on the negative impact of IFN-γ produced by CD4 + T cells on Paneth cells and their ability to produce certain AMPs (11). Our study demonstrates that IL-17-induced AMPs contribute to provide a broad umbrella of protection from dysbiosis that is essential during pathogenic T. gondii infection; therefore a reduced Th17 response during infection results in exacerbated dysbiosis and pathology. Indeed, Crtam −/− mice had a very selective reduction in AMPs induced by IL-17, but expressed AMPs released by Paneth cells at normal levels. Given the role of IL-17 in inducing the differentiation and recruitment of neutrophils (35), we envision that Th17 could potentially act by mobilizing a neutrophilic source of AMPs that, together with Paneth cells, provides the broad spectrum of AMPs required for preventing perturbation of microbiota, particularly the expansion of Enterobacteriaceae. The relative impacts of Paneth cell-and neutrophil-derived AMPs in preventing dysbiosis may vary depending on the type and location of the intestinal infection.
Our data support a role for CRTAM in intestinal Th17 responses, whereas Th1 and Th22 responses seem to be relatively CRTAM-independent. This conclusion is consistent with previous studies focusing on the impact of CRTAM on T cell responses in various contexts (24,(36)(37)(38), though it does not concur with the originally proposed role for CRTAM in CD4 + T cell production of IL-22 and IFN-γ (39). Our study further advances our understanding of the role of CRTAM in Th17; clearly it is not required for the initial induction of Rorγt in naïve CD4 + T cells, as Crtam −/− and WT mice have similar percentages of intestinal Rorγt + CD4 + T cells. Rather, CRTAM seems to enable IL-17 production by Th17 cells. Since myeloid cells secrete IL-23 in response to changes in the intestinal microbiota (40,41) and express the CRTAM ligand CADM1 (26), we speculate that CRTAM-CADM1 interactions may sustain the exposure of Rorγt + CD4 + T cells to myeloid cells producing IL-23, thereby promoting their final maturation into IL-17producing CD4 + T cells.
In conclusion, our study demonstrates that a protective immune response against T. gondii must both effectively control invading pathogens through the Th1-IFN-γ axis and prevent dysbiosis, barrier leakage and microbial translocation through the Th17-AMP axis. Given that oral infection of mice with T. gondii causes intestinal inflammation and ileitis that resemble some aspects of Crohn's disease (42), our findings suggest that unbalanced representation of CD4 + T cell subsets may have a decisive impact in the onset and progression of human inflammatory bowel disease.

Mice
Crtam −/− and C57BL/6 mice were bred in a pathogen-free facility at Washington University. Age and sex matched animals were used throughout the experiments and were co-housed from birth. All animal experiments were conducted according to U.S.A. Public Health Service Policy of Humane Care and Use of Laboratory Animals. All protocols were

T. gondii Infection and Blocking Antibodies
The luciferase-expressing Me49-Luc type II T. gondii strain was provided by Laura Knoll (29). Tissue cysts were obtained from the brains of infected mice. Experimental mice were infected with 10 cysts of T. gondii Me49 orally. Weight was monitored every 48 h for the first 20 days. Parasite burden was analyzed by bioluminescence measurements. Mice were imaged every 48 h for 20 days by i.p. injection of 150 mg of luciferin D (Biosynth AG) per kg of body weight and using a Xenogen IVIS 100 (Caliper Life Sciences). Data was analyzed with the Living Image Software (Caliper Life Sciences) and is expressed in relative light units. Blocking antibodies anti IL-17A (IL-17F) and IgG1 isotype control were purchased from BioxCell. Mice were injected intraperitoneally with 350 µg at day −1 and +5 of T. gondii infection.

Bacterial Counts in Tissue
Mesenteric lymph nodes and spleens from infected mice were aseptically removed and homogenized in PBS in a Magnalyzer (Roche; 5,000 rpm × 30 s). The homogenized organs were plated in Luria Bertani (LB) agar and incubated at 37 • C for 48 h. Colonies in the plates were counted and the colony forming units per organ calculated.