Dual transcriptomics to determine interferon-gamma independent host response to intestinal Cryptosporidium parvum infection


 Background: Protective immune responses to Cryptosporidium parvum, a zoonotic, gastrointestinal parasite, are primarily dependent on the presence of interferon-gamma (IFNγ). We discovered that treatment with soluble T. gondii antigen (STAg) reduces Cryptosporidium parvum shedding in the absence of IFNγ. To identify the protective IFNγ independent responses elicited by STAg, we conducted a transcriptomic analysis of intestinal sections of IFNγ-deleted, C. parvum infected or uninfected mice treated with STAg or PBS.
Results: STAg treatment reduced oocyst shedding in C. parvum infected IFNγ deleted mice. Gene ontology analysis of the intestinal transcriptomes suggested that both C. parvum infection and STAg treatment changes the transcript abundance of genes involved in the host cell membrane, intracellular and extracellular transport, and immune responses. We found in high abundance 37 genes related to IFN type I response in infected mice treated with STAg. Among these genes, members of the oligoadenylate synthetase and Schlafen family were identified.
Conclusions: STAg treatment of C. parvum infected mice induced both host immune and metabolism changes associated with a reduction in shedding. Several components of the type I interferon immune response were more abundant in the ileum of C. parvum infected in IFNγ-deleted mice. Future studies will explore the role of type I IFN mediated immune responses in controlling C. parvum infections. STAg treatment appears to only affects the host transcriptome while the parasite transcriptome remains unaffected. Several C. parvum genes, including mucin genes, are more abundant during infection of animals, which opens new avenues in C. parvum research.


4
The life cycle culminates in the production of oocysts that are released into the environment in the feces (17,22).
During C. parvum infection the cytokine interferon-gamma (IFNγ) plays a central role in controlling infection in both innate and cell-mediated immune response (23,24) through a variety of mechanisms. IFNγ inhibits parasite invasion, changes the intracellular iron (Fe 2+ ) concentration in enterocytes that the parasite requires (25), and participates in the clearance of the parasite (24). Patients with IFNγ deficiency are more likely to develop chronic C. parvum infection (26). While wild-type mice are resistant to C. parvum infection, IFNγ-deficient mice are highly susceptible (23,27,28). Treatment of intact mice with IFNγ neutralizing antibodies renders these mice susceptible to C. parvum infection (29). IFNγ also induces up-regulation of the host enzyme indoleamine 2,3-deoxygenase (IDO), inducing tryptophan starvation (30) and killing the parasite.
To identify novel C. parvum immune responses, we tested a non-infectious extract of soluble Toxoplasma gondii antigens (STAg) for efficacy against C. parvum infection.
STAg was previously found to protect against viral, bacterial, and parasitic infections by 5 induction of various innate immune responses. In mice infected with the avian influenza virus H5N1, STAg treatment reduced viral titers in the lungs, and induced a strong gamma interferon (IFNγ) response resulting in increased survival (38). STAg treatment was also able to prevent experimental cerebral malaria in mice challenged with Plasmodium berghei via induction of IL-12 and IFNγ and subsequent reduction of parasitemia (39). Additionally, STAg treatment provided resistance against Listeria monocytogenes bacterial infection by reducing the bacterial burden in the spleen and liver by stimulation of TLR11 and promoting recruitment of Ly6C CCR2+ inflammatory monocytes (40).
Because STAg exhibited immunomodulatory activity protective against a variety of pathogens, for these studies, we tested STAg for anti-cryptosporidial activity. To mimic immunocompromised patients, we used IFNγ-deletion mice to define the IFNγ independent host response to C. parvum. Even in the absence of IFNγ, STAg treatment reduced C. parvum oocyst shedding, indicating that IFNγ independent immune responses were effective against the parasite. We examined the effect of STAg treatment on host gene expression to identify components of the immune response responsible for alleviating C. parvum infection in an IFNγ independent scenario.

Results
STAg treatment reduces the oocyst shedding in C. parvum infected IFNγ-deleted mice. 6 To evaluate if STAg treatment would affect C. parvum infection, IFNγ-deleted female mice were infected with 1x10 7 C. parvum oocysts Iowa II strain by oral gavage then treated intraperitoneally with 1 mg STAg or PBS control on 1, 3, and 5 days postinfection (dpi). Fecal samples were collected every other day from 0 to 14 dpi. Oocysts shedding over time was quantified by qPCR (18) and the area under the curve was calculated. We observed that STAg treatment significantly reduced oocyst shedding compared with the PBS treated group (p=0.030), indicating that STAg has a therapeutic or immunomodulatory effect against C. parvum infection (Fig 1A). The experiment was repeated with IFNγ-deleted mice infected with 1x10 5 Nano luciferase (Nluc) C. parvum oocysts Iowa II strain and oocyst shedding over time was quantified by NLuc expression (42). STAg treatment again significantly reduced the oocyst shedding (Fig 1B; p=0.034) confirming our first results (Fig 1A). From these experiments, we concluded that STAg treatment reduces C. parvum oocyst shedding in the absence of IFNγ.

Murine intestinal sections were sequenced to determine the interferon-gamma independent transcriptomic response to C. parvum infection
To determine the underlying mechanisms by which STAg treatment successfully reduced C. parvum shedding in an IFNγ independent manner, we obtained a global view of the transcriptomic changes of the host and parasite elicited by STAg treatment via RNAseq of intestinal tissue. To optimize the collection of intestinal tissue with the highest concentration of parasites, we infected IFNγ-deleted female mice with NLuc C.
parvum and visualized the parasites throughout the mouse intestinal tract using an in 7 vivo imaging system (IVIS). Using this technique, Nluc C. parvum was localized mainly in the cecum and ileum at 9 dpi ( Fig 1C).
To investigate the changes in the transcriptome attributable to STAg in the setting of C. parvum infection, mice were infected, treated and samples collected as shown in Fig   2A. Infected mice treated with PBS shed more oocysts than STAg treated infected mice ( Fig 2B), indicating that results in this experiment were analogous to experiments shown in Fig 1. As C. parvum infects and replicates in enterocytes (15,43,44), we collected the epithelial cell layer in the ileum and caecum for transcriptomic analysis. This methodology allowed us to remove the mucus layer and to concentrate the parasite transcriptome relative to the host transcriptome. RNA was isolated by the Trizol method and libraries prepared for polyA enrichment. Raw reads were processed and analyzed for differential expression and gene ontology enrichment (Fig 2C and Table S1).
Intestinal transcriptomic response to C. parvum oral challenge is the most prominent in the host's ileum.
A total of 12 differential gene expression comparisons between tissues, treatments, and infection status were conducted for Mus musculus (Table S2) and C. parvum (Table S3) transcripts. Principal Component Analysis (PCA) was used to visually compare the ileal and cecal transcriptomes of mice treated with STAg or PBS with and without C. parvum infection (Fig 3). Clustering of normalized values calculated by DESeq2 indicates a clear separation between ileal and cecal samples. Ileal samples from STAg treated animals (n=6, n=3 per infection group), clustered by infection status (Fig 3, right side). One of the ileal samples from the PBS treated infected group does not cluster with the other infected samples, likely due to the percent of reads mapped to the C. parvum genome (5.56% vs 1.21 and 1.66%). The cecal samples did not cluster significantly based on infection or treatment.
In terms of the host transcriptomic response to infection and STAg treatment, we observed a low number of differentially expressed genes (DEGs) when comparing the PBS and STAg treatment while maintaining tissue and infection status constant, (Table   S2 comparisons 1-4 and Fig 4 A-B). In the ileum, STAg treatment had 13 significant DEGs compared to PBS treatment (Fig 4A). Only a single gene, Myosin Binding Protein C2 (Mybpc2), is shared between comparisons 1 and 2 (Fig S1), highlighting that it is less abundant with STAg treatment with or without infection.
We also compared infection status while maintaining treatment type constant (Table S2 comparisons [5][6][7][8] and observed the largest number of DEG in these comparisons, which is not surprising as in these comparisons we are capturing the host cell response to infection. Whether STAg or PBS treated, more than 100 genes were significantly differentially expressed within the ileum (Fig 4C) between infected and uninfected animals. Within the STAg treated mice, a total of 129 genes for ileum ( Fig   4C) and 78 for the cecum (Fig 4D) were classified as differentially expressed. Out of those 129 genes within the ileum of STAg treated mice, 100 were exclusive to the ileum, and 7 were shared with the cecum of STAg treated mice. The remaining 22 genes were shared with PBS treated, 20 within the ileum and 2 within the cecum ( Fig   S2). Out of 78 cecal DEGs between infected and uninfected STAg treated mice ( Fig   4D), 61 were exclusive to the cecum, 7 were shared with the ileum of STAg treated mice. The remaining 10 genes were not exclusive to STAg treatment and shared with PBS treated groups: 5 ileum PBS and 5 cecum PBS treated groups (Fig S2). Only 2 genes were shared between every single comparison when treatment was maintained Overall, the number of DEGs was higher for the ileum (Fig 4A and 4C) than for the cecum (Fig 4B and 4D). These results are not due to unmapped reads, as we observed, on average, ~79% of the reads of both tissues to be uniquely mapped to the host genome. Because the percentage of reads mapped to the C. parvum genome is considerably higher in the ileum than the caecum (Fig 3), it is not surprising that this tissue had the highest number of significant DEGs regardless of treatment.

Gene Ontology analysis shows differentially expressed immune response
To further explore the transcriptomic response to C. parvum infection and STAg treatment, we used Gene Ontology (GO) enrichment analysis through the Database for Annotation, Visualization, and Integrated Discovery (DAVID), v6.8 (45,46). Genes that were differentially expressed when treatment was constant and infection was compared (comparisons 5-8 in Table S2) were analyzed for enrichment of three GO term categories: cellular component (Fig S3), molecular function (Fig S4), and biological process ( Fig 5). In comparison to the cecum of uninfected STAg treated mice, the cecum of infected STAg treated mice was enriched for cellular component GO terms was not observed in biological process or molecular function GO terms ( Fig S4 and Fig   5). Due to the localization of the parasite on the periphery of the host cell, it is encouraging that we observe GO terms that are related to the extracellular space and cellular trafficking when comparing infection status. We observed major enrichment of molecular function GO terms related to purine metabolism and RNA binding in the ileum of mice infected and STAg treated, which was not observed in the ileum-of mice that were PBS treated ( Fig S4).
As expected, the biological process GO terms showed enrichment for host immunity in the ileum of mice treated with STAg ( Fig 5). As we tested our hypothesis in IFNγ-deleted mice, we were expecting signaling that was IFNγ independent. Overall, the innate immune response was significantly associated with type I interferon response, including IFN-alpha signaling (GO:0035457), response to IFN-beta signaling (GO:0035458), downstream targets like defense response to virus (GO:0051607), and response to virus (GO: 0009615) (Fig 5).

Transcript abundance comparisons show immune response and host metabolism changes
As the DEGs in the caecum were highly variable among biological replicates, we focused on ileal samples for further analysis. To compare the abundance of these significant genes across the 6 biological replicates per treatment group, we generated heatmaps charting the normalized reads per animal (Fig 6A and S5). We found 37 genes related to IFN type I in mice infected and treated with STAg (Table 1 and Fig 6A) but not present in mice infected and treated with PBS ( Fig S5). Among them were eight oligoadenylate synthetase (OAS) genes, the DDx60 (47), the Dhx58 also known as Lpg2 (48), Mtx2, Tap1, Trim 30a-b-c, the ubiquitin-specific peptidase 18, seven interferon-induced proteins, the Immunity-related GTPase, the immunity-related GTPase family M member 1 and 2; and the Ring finger proteins 213 and 225 (49,50).
Similarly, 6 members of the Schlafen gene family (SLFN) were present in higher abundance in STAg treated infected mice versus STAg treated uninfected mice (Comparison 5 and Table 1).
Curiously, 2 transcripts related to gluconeogenesis were identified in lower abundance in the ileum of infected and STAg treated mice when compared to their uninfected counterparts (Fig 6B), Glucose-6-Phosphatase Catalytic Subunit (G6pc), and Fructose 1,6-bisphosphatase (Fbp1). There have been sporadic reports that children (51) or senior individuals (52) with C. parvum infection seem to have a lower tolerance for lactose, and we are the first to observe lactase (Lct) transcript to be in lower abundance in infected mice ( Fig 6B). In the PBS-treated mice, transcripts for G6pc, Fbp1, and Lct were also less abundant, but they did not achieve the set threshold limits. For Fpp1, the log2 fold change was -1.8 and for Lct, the log2 fold change was -1.9. While G6pc did make the -2 log change cutoff (-2.7), due to variability, G6pc did not meet the padj threshold (p=0.07). These results suggest that infection with C. parvum may modulate host carbohydrate metabolism away from gluconeogenesis and disaccharides like lactose. It will be interesting to further investigate if C. parvum infection reduces the expression of these enzymes in other hosts or in vitro models.

C. parvum transcript analysis shows no significant differences in gene expression between PBS or STAg treatments
An analysis was conducted to determine if STAg treatment had any effect on the transcriptome of C. parvum (Table S3). When a PCA exploration for reads that were aligned to the C. parvum genome was conducted, we observed clustering of the ileum, but not the cecum samples ( Fig S6). The percentage of uniquely mapped reads to the C. parvum genome was significantly lower when compared to host mapping, a result that we were expecting according to other similar studies (53). On average per sample, the ileum had more transcripts that mapped to the C. parvum genome than the cecum (Table S1). Differential expression analysis of the parasite transcriptome was performed in the same manner as for the host transcriptome but there were no significant differentially expressed genes between PBS and STAg treated samples. These results indicate that the effect of STAg to lower C. parvum oocyst shedding was more likely due to changing the host immune response than a direct effect on the parasite.
The most abundant C. parvum transcripts within the ileum, are listed in Tables 2 and 3.

Discussion
While nitazoxanide has some efficacy in immune-competent adults, it is not effective in severely immunocompromised patients (56)(57)(58) highlighting the need to develop interventions effective in the absence of normal immune responses. To mimic the human immune-compromised state, we used IFNγ-deletion mice to define the IFNγ independent host response to C. parvum. Our group and others have determined previously that STAg, through various immune mechanisms, has a protective effect against intracellular infections from multiple pathogens (38)(39)(40). In this study, we discovered that STAg treatment also reduced C. parvum infection in IFNγ-deleted mice (Fig 1).
Through our transcriptomic analysis, we have found that IFNγ-deleted mice have a higher number of significant DEGs when comparing infection status and maintaining tissue and treatment type constant (Fig 4C and D), suggesting that treatment alone is not sufficient to elicit a host gene expression change. C. parvum infection likely acts synergistically with STAg treatment to elicit the type I interferon immune response observed in this transcriptomic study. We suggest this immune response aids in the lower level of parasite shedding detected in the fecal samples of STAg treated mice (Fig   1 and Fig 2B).
Mice that were STAg treated had a total of 207 host DEGs (129 and 78 genes for the ileum and cecum, respectively). Most of the DEGs were of higher transcript abundance within both tissues (87% for ileum and 60% for cecum). Most of the DEGs in the ileum were classified as belonging to the type I interferon response and its downstream effectors, typically known for their anti-viral activity (Table 1 and Fig 6A). IFNs are a multigenic family of cytokines that regulate different aspects of the immune response.
There are three classes of IFN: type I, II, and III, which recognize distinct receptors on cell membranes (59). IFN signaling induces transcription of hundreds of IFN-stimulated genes that further the immune response against viral infection. The IFN type I response is essential in the immune response against viruses (60)  Among the IFN-stimulated genes, we found six Schlafen gene members ( Table 1). The murine Schlafen family has 9 proteins with 5 human homologs. Schlafen proteins have been described to have many potential roles in cell processes, including but not limited to cell differentiation, inhibition of anchorage-independent growth, and regulation of the immune response (67,68). SLFN2 has been shown to regulate type I interferon response via the NF-κβ pathway (69). SLFN3 and its human analog SLFN12 are critical in regulating intestinal epithelial differentiation (43,44). Of this family, the most abundant transcript in our transcriptomic dataset was Schlafen 4 (Slfn4) (70). A previous study has shown in SLFN4 over-expressing mice a reduction in the recruitment of inflammatory macrophages, suggesting SLFN4 as a negative regulator of monocytopoiesis (71). It will be interesting to further explore the relationship between SLFN4 and C. parvum infection.
Using GO enrichment analysis, we found an enrichment of immunological terms downstream of IFNα/β signaling (Fig 6). Curiously, no change in IFNα/β transcript abundance was observed. This change in the abundance of the effectors versus the signaling molecules themselves is probably due to the late time point selected. We chose the 9 dpi timepoint to coincide with the peak C. parvum oocysts shedding to maximize the number of C. parvum transcripts in the datasets. Future experiments will analyze the IFNα/β signaling pathways throughout the time course of infection to determine the level of protection type I signaling confers against C. parvum infection.
STAg treated mice had a lower abundance of IL-10, Ccl4, Ptger3, and granzymes B and K (Fig 7). IL-10 is an anti-inflammatory cytokine, its role is to modulate a balance between pathology and protection against a rampant immune response (72,73). IL-10 is elevated in the stool of Cryptosporidium infected Haitian children (72) as well as in patients co-infected with C. parvum and HIV (74). In animal models, IL-10 knockout mice are more resistant to C. parvum infection (75), but anti-IL-10 antibody treatment in calves does not reduce C. parvum shedding (76). Ccl4 is a chemokine independent of IFNγ, whose expression level increases during C. parvum infection when compared to control mice (33). A high level of prostaglandins has been observed in C. parvum infection, implicated in increased mucosal activity and downregulation of inflammatory cytokine production (77,78). Granzymes induced by NK cells or CD8+ cytotoxic T cells (79) have been found in high abundance in mice and humans infected with Cryptosporidium, inducing cell death and cytolysis of infected epithelial cells. They are also associated with apoptosis and colon lesions in AIDS patients co-infected with Cryptosporidium (80-83).
Even though we could not assess C. parvum differential expression, we report the most abundant C. parvum transcripts within the ileum. Among those transcripts, we found the GDP-fucose transporter (CGD3_500) to be abundant in our infected animals, regardless of treatment. In other apicomplexan parasites like Toxoplasma gondii, O-fucosylated glycoproteins can form assemblies that are associated with nuclear pore complexes (84). For both T. gondii and Plasmodium falciparum, O-fucosylation of thrombospondinlike repeats is required for processing proteins relevant to host cell invasion (85,86).
We also found transcripts in high abundance of parasite mucins. Mucins are glycoproteins that have an amino acid composition consisting of 20 to 55% serine, threonine, and proline residues, with extensive chains of O-linked carbohydrates (54) forming homo-oligomers. Mucins have been reported both in the host and in some gastrointestinal parasites. Cryptosporidium's mucins are important for attachment and invasion to the host cell (54) or evasion of the host immune system (87). Many essential Cryptosporidium proteins are mucin-like proteins such as gp40/15 (88), cpClec, gp30, and gp900 (89). There is a genetic locus of seven mucins denominated cpMuc1-7, which suggests that these seven mucins are regulated and expressed in a coordinated fashion with a role in the same biological processes (54). Cryptosporidium mucins have high polymorphism between species and genotypes, which suggests gene products might be important virulence determinants subject to immune pressure (54). Recently, it has been published that mucins play a role in tissue tropism between gastric and intestinal Cryptosporidium species and also determine differences in host range among Cryptosporidium species (90,91). Further investigation is required for the validation of the differential expression of C. parvum transcriptomic response to STAg treatment.

Conclusions
Our results show that STAg treatment reduces the oocyst shedding in C. parvum infected IFNγ-deleted mice. We evaluated the effect of STAg treatment on both the mouse and in the parasite transcriptomes. The detailed analysis of the transcriptome suggests that STAg treatment induces both host inflammatory responses and metabolism changes relevant to gluconeogenesis and disaccharide breakdown.
Particularly notable is the induced type I interferon immune response in the ileum of C.
parvum infected in IFNγ-deleted mice. We hypothesize that this immune response can lower the level of parasites shed in the fecal samples of STAg treated mice, but further experiments are needed to confirm this observation.
In terms of the parasite transcriptomic regulation, we did not find significant differences in transcript abundances when comparing STAg vs PBS treatment. It is possible that STAg treatment only affects the host transcriptome while the parasite transcriptome remains unaffected. This lack of differential C. parvum expression could also be due to the limitations of coverage and sequence depth of our study. While several of the C. parvum transcripts were more abundant in the ileum than the cecum, there was also significantly higher mapping of C. parvum reads to the ileum. Thus, we focused on the most abundant transcripts within the ileum with both treatments in hopes that the C. parvum community can use this in vivo data for comparisons with in vitro systems. By discovering C. parvum genes that are more abundant during infection of animals that lack IFNγ signaling, we are discovering targets that could serve as an alternative route to control infection. A significant addition to the field that opens new avenues in C.
parvum disease halting research.     (Tables S2 and S3). Visualization of all possible comparisons can be found in Figure S1 and Figure S2.           supplemented with 10% fetal bovine serum, 1% penicillin-streptomycin, and 2 mM Lglutamine. When the parasites were beginning to lyse the host cells, monolayers were scraped, passed twice through a 27-gauge needle, and pelleted at 500 g. Parasites were washed in phosphate-buffered saline (PBS) without divalent cations and resuspended to 4 x 10 8 parasites per mL. After sonication with five 30-s pulses, parasites were centrifuged at 100,000 g for 45 min and supernatants were collected and stored at 80°C until use.

Mouse Infections
Experiment 1: Eight 4-week-old females were infected by oral gavage with 1x10 7 C.
parvum Iowa II oocysts and given STAg (1 mg/dose via ip injection,) or PBS (4 mice per treatment group) at 1, 3, and 5 dpi. Fecal samples were collected every other day from day 0-day 14 and oocyst shedding was quantified by qPCR following the protocol previously published (41). Shedding overtime was graphed and the area under the curve was calculated for each mouse. The area under the curve was averaged for each group and compared by t-test two-tailed using GraphPad PRISM.  STAg or PBS 4 hours before infection. Then, 2.5 x10 5 nanoLuciferase Iowa II strain oocyst/mice were given by oral gavage to the infected groups. At 24, 48, and 72 hours post-infection, the mice were treated with either STAg or PBS. From 6 to 9 dpi, the mice were weighed, and individual fecal samples were collected to evaluate infection levels.
On 9 dpi, mice were euthanized. The intestine was isolated and rinsed using a gavage needle and scrapped using a single-use sterile cell scrapper with sterile 1X PBS. A section of 2 cm of ileum from each mouse was processed. Similarly, the cecum was isolated and rinsed to remove its content. The cecum bag was opened and scraped.
Scrapping was performed to isolate the epithelial layer. After microscopic verification, Stranded mRNA). Samples were run in duplicate in two lanes of a 2x100bp S1 Flowcell (Novaseq6000). On average a total of 41.5 million reads (min: 36 million, max: 49.6 million) paired-end 150-bp reads per sample were generated. The quality of the reads was determined (FastQC v0.11.9 (93), a threshold of 34 was selected, and only reads that met the threshold were used for further analysis.
Luciferase assay: Mouse infections were confirmed by luciferase measurement of mouse fecal samples following a protocol previously reported (42). Briefly, 20mg of fecal sample was weighed into a 1.5-ml microcentrifuge tube and homogenized in 1mL of lysis buffer (50mM Tris-HCl, 10% glycerol, 1% Triton-X, 2mM dithiothreitol (DTT), 2mM EDTA) using 10-15 glass beads (3mm) and a vortex mixer for 1 min, followed by clarification of lysate by brief centrifugation. One hundred microliters of lysate were mixed with an equal volume of NanoGlo Luciferase Buffer, prepared with 1:50 dilution of the substrate (Promega, corporation). Three technical replicates per sample were conducted. Luminescence was measured in a synergy HT luminator (BioTek).
Oocyst purification from mouse feces: Oocysts were purified from mouse fecal samples using sucrose suspension followed by a cesium chloride (CsCl) centrifugation as previously published (42). Mouse feces were suspended in water, homogenized, and passed through two filters. This filtered suspension was mixed 1:1 with aqueous sucrose solution (specific gravity 1.33) and centrifuged at 1,000g for 5 min. Oocysts were collected from the supernatant and suspended in 0.85% NaCl saline solution.
Then, 0.5 mL of this preparation was overlaid onto 0.        (Tables S2 and S3). Visualization of all possible comparisons can be found in Figure S1 and Figure S2. Only GO terms populated by 3 or more DEGs were included in these visualizations (p < 0.05). Bar lling color represents treatment: control treatment PBS (blue); Soluble T. gondii antigen, STAg (red).

Figure 6
Heatmap of differentially expressed genes in the ileum of mice treated with STAg (comparison 5). The top 25 more (A) and less (B) abundant transcripts (ranked by log2 fold change). Each column represents a biological replicate.

Figure 7
Heatmap of differentially expressed genes in the ileum of infected mice treated with STAg or PBS. 13 genes (ranked by log2 fold change) were found to be differentially expressed in the ileum between infected mice treated with PBS or STAg. Each column represents a biological replicate.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.