Abstract
The lumen-dwelling protozoan Giardia is an important parasitic cause of diarrheal disease worldwide. Infection can persist over extended periods with minimal intestinal inflammation, suggesting that Giardia may attenuate host responses to ensure its survival, although clearance eventually occurs in most cases. IL-10 is an anti-inflammatory regulator critical for intestinal homeostasis and controlling host responses to bacterial exposure, yet its potential role in coordinating antiprotozoal host defense in the intestine is not known. In this study, we found that murine infection with the natural enteric pathogen Giardia muris induced a transient IL-10 response after 2–4 wk at the primary site of infection in the upper small intestine, but parasite colonization and eradication were not affected by the absence of the cytokine in gene-targeted mice. However, IL-10 was critical for controlling infection-associated immunological sequelae in the colon because severe and persistent diarrhea and colitis were observed in IL-10–deficient mice within 1–2 wk postinfection but not in uninfected littermate controls. Inflammation was characterized by epithelial hyperplasia, neutrophil and macrophage expansion, and Th1 induction and could be prevented by blockade of IL-12/IL-23 p40 but not depletion of CD11c+ dendritic cells. Furthermore, the intestinal microbiota underwent characteristic shifts in composition and was required for disease because antibiotics and loss of TLR signaling in MyD88-deficient mice protected against colitis. Together, our data suggest that transient infection by a luminal and seemingly noninflammatory pathogen can trigger sustained colitis in genetically susceptible hosts, which has broader implications for understanding postinfectious syndromes and other chronic intestinal inflammatory conditions.
Introduction
Interleukin-10 is a critical immune regulator that is secreted by hematopoietic cells, particularly macrophages and T cells as well as nonhematopoietic cells. The cytokine acts on both myeloid and T cells to stimulate anti-inflammatory factors and suppress proinflammatory mediators (1). The importance of IL-10 in regulating homeostasis was first demonstrated by the observation that IL-10–deficient (Il10−/−) mice, which are phenotypically normal after birth and weaning, develop spontaneous, progressive enterocolitis upon aging (2–5). Colitis in this model is driven by commensal bacteria, as inflammation is not observed under germ-free conditions (4), and antibiotic treatment can prevent and attenuate disease (6, 7). Similarly, mice with impaired IL-10R signaling develop severe spontaneous colitis, driven by expansion of Th17 cells normally restricted by regulatory T cells and dendritic cells (1, 8, 9). Consistent with the findings in murine models, genome-wide association studies of inflammatory bowel disease (IBD) patients have identified IL-10 and IL-10R polymorphisms that are associated with severe, early-onset forms of the disease (1, 10). Despite its importance in immune regulation and association with IBD, trials of IL-10 treatment did not reveal clinical benefits, perhaps because disease had progressed beyond IL-10–dependent stages (1).
Although loss of IL-10 signaling can cause excessive immune and inflammatory responses to commensals, its impact on host interactions with pathogens is more complex. Thus, on the one hand, IL-10 deficiency can promote clearance of bacterial pathogens, such as Salmonella, Klebsiella pneumoniae, Listeria monocytogenes, and Chlamydia trachomatis, as well as certain viruses, parasites, and fungi, by induction of heightened Th1 responses, thereby improving host survival (11–14). On the other hand, enhanced Th1 responses can lead to exacerbation of inflammation postinfection. For instance, intestinal colonization of Il10−/− mice with the enteric pathogens Helicobacter hepaticus or Campylobacter jejuni causes chronic inflammatory lesions in the cecum and colon (15, 16). Inflammation under these conditions is not only caused by excessive Th1 responses but is also related to blooms of distinct families of intestinal commensals that correlate with increased susceptibility to colitis (17, 18).
Intestinal infection with the common protozoan parasite Giardia leads to diarrhea and abdominal pain, and engages adaptive immunity, but causes only minimal mucosal inflammation (19). The parasite colonizes the small intestine but does not invade the mucosa, indicating that it can impact host responses from an “off-shore” location in the lumen (20). Infection elicits both Th1 and Th17 responses at the site of infection, yet only IL-17–producing cells are required for parasite elimination (21, 22). Importantly, this effector response does not appear to be associated with significant inflammation, raising the possibility that any potential proinflammatory responses that are typically associated with microbial infections may be attenuated by effective suppression during Giardia infection.
Given the importance of IL-10 in regulating intestinal inflammation and immunity, we set out to test the hypothesis that IL-10 is essential for dampening effector T cell responses induced by Giardia infection. Unexpectedly, we found that IL-10 is not critical for limiting T cell responses at the site of infection in the small intestine but is required later for controlling excessive immune responses to the microbiota in the colon. The results suggest that a seemingly innocuous pathogen has the potential to trigger severe colitis in genetically susceptible hosts, which has broader conceptual implications for understanding the onset of IBD in humans.
Materials and Methods
Mice and infection protocol
Wild-type, Il10−/− (5), Ccr7−/− (23), and Tcra−/− (24) mice on a C57BL/6 background were obtained from The Jackson Laboratory and bred and maintained under specific pathogen-free conditions. To generate double mutant mice, Il10−/− mice were crossed with either Myd88−/− (25) or Cd11c-diphtheria toxin receptor (DTR) transgenic mice (26). Littermates served as controls. At 7–9 wk of age, mice were infected by oral gavage with G. muris cysts (104 cysts per mouse) prepared from infected wild-type mice as previously described (27). Parasite burden in infected mice was determined by enumerating trophozoites in the small intestine (27). All mice were monitored daily for weight loss, stool consistency, and rectal bleeding. No significant differences were observed in the course of infection or disease severity between males and females, so results were combined and are reported together. All animal studies were approved by the Institutional Animal Care and Use Committees of the University of California, San Diego and the University of Texas Medical Branch.
Histological analysis
Jejunum, ileum, and colon were removed, opened longitudinally, and processed as Swiss rolls before overnight fixation in either Bouin’s solution or 10% formalin. Fixed tissues were embedded in paraffin, and 5-μm sections were prepared and stained with H&E. Histological scores (range 0–20) were obtained in a blinded fashion by assessing the following parameters: 1) mucosal architecture (0, normal; 1, focally abnormal; 2, diffusely abnormal; 3, severely abnormal), 2) inflammatory cell infiltration of mucosa (0, normal; 1, mild infiltration; 2, moderate infiltration; 3, severe infiltration), 3) epithelial hyperplasia (0, normal; 1, mild; 2, moderate; 3, extensive), 4) epithelial erosions and ulcerations (0, absent; 1, present), 5) crypt abscesses (0, absent; 1, present), 6) goblet cells (0, present; 1, absent), 7) muscle thickening (0, normal; 1, mild; 2, moderate; 3, severe), and 8) extent of damage and inflammation (0, normal; 1, damage present in <25% of the tissue; 2, damage in 25–75%; 3, damage in >75%). Cell numbers in the lamina propria were quantified per 40× visual field in an area equivalent to 0.1 mm2. Crypt depth was measured using NIS-Elements software (Nikon). For each colon section, quantification was performed on the two most affected areas at least 10 crypts apart, and the scores were averaged.
Analysis of mRNA expression
Full-thickness pieces of jejunum, ileum, and colon were treated with RNAlater (Ambion) at 4°C for 24 h, snap-frozen, and stored at −80°C until use. Total RNA was isolated with TRIzol Reagent (Thermo Fisher) and analyzed individually for each tissue by RT-PCR (21). Briefly, RNA was reverse-transcribed into cDNA using the High Capacity cDNA Reverse Transcription kit (Applied Biosystems), and quantitative RT-PCR (qRT-PCR) analysis was done with Mesa Green 2 x SYBR mix (Eurogentec) in a real-time PCR machine. Primers and expected PCR product size were as follows: IL-10, 5′-GCT CTT ACT GAC TGG CAT GAG-3′ (sense), 5′-CGC AGC TCT AGG AGC ATG TG-3′ (antisense), 108 bp; IL-17A, 5′-ACC AGC TGA TCA GGA CGC GC-3′ (sense), 5′-CCA GGC TCA GCA GCA ACA-3′ (antisense), 84 bp; IFN-γ, 5′-TCA AGT GGC ATA GAT GTG GAA GAA-3′ (sense), 5′-TGG CTC TGC AGG ATT TTC ATG-3′ (antisense), 84 bp; IL-4, 5′-AGA AGG GAC GCC ATG CAC GG-3′ (sense), 5′-ATG CGA AGC ACC TTG GAA GCC C-3′ (antisense), 105 bp; IL-9, 5′-CAG CTG ACC AAT GCC ACA CAG AAA-3′ (sense), 5′-GAA ATG ACA GTG TGT TGC CTG CCA-3′ (antisense), 151 bp; IL-12 p40, 5′-CGG ACG GTT CAC GTG CTC ATG G-3′ (sense), 5′-TCC ACA TGT CAC TGC CCG AGA GT-3′ (antisense), 103 bp; CCL2, 5′-GCT CAG CCA GAT GCA GTT AAC GC-3′ (sense), 5′-TGG GGT CAG CAC AGA CCT CTC T-3′ (antisense), 169 bp; CCL3, 5′-GGT CTC CAC CAC TGC CCT TGC-3′ (sense), 5′-AGG CAG TCG GGG TGT CAG CT-3′ (antisense), 95 bp; CXCL5, 5′-GCA TTT CTG TTG CTG TTC ACG CTG-3′ (sense), 5′-TGC AGG GAT CAC CTC CAA ATT AGC-3′ (antisense), 150 bp; MPO, 5′-CAG CGA GGA CCC CCT AGC CA-3′ (sense), 5′-GGC ATC TCG CTG GAG CGC AT-3′ (antisense), 198 bp; and GAPDH, 5′-TGT GAT GGG TGT GAA CCA CGA CAA-3′ (sense), 5′-AGT GAT GGC ATG GAC TGT GGT CAT-3′ (antisense), 209 bp. Relative changes in target mRNA levels were calculated by the 2∆∆Ct method, with GAPDH mRNA as the reference standard.
Isolation and analysis of immune cells
Single-cell suspensions were prepared from mesenteric lymph nodes (MLN) and intestinal tissues by collagenase digestion as previously described (21). After removing the epithelial layer with EDTA/DTT, tissues were digested in RMPI 1640 medium containing 1 mg/ml Collagenase D (Roche Applied Science) and 100 μg/ml DNase I (Worthington Biochemical Corporation). Liberated cells were collected and pooled. To assess changes in dendritic cells and macrophages, cells were stained with Alexa Fluor 700–conjugated rat anti-mouse MHC II (clone M5/114.15.2; eBioscience), Horizon V450–conjugated Armenian hamster anti-mouse CD11c (clone HL3; BD Biosciences), Texas Red–conjugated rat anti-mouse CD11b (clone M1/70; Thermo Fisher), PE-Cy7–conjugated rat anti-mouse B220 (clone RA3-6B2; eBioscience), and FITC-conjugated Armenian hamster anti-mouse CD103 (clone 2E7; eBioscience). Viability was assessed by staining with Fixable Aqua Dead Cell Stain (Thermo Fisher). For functional evaluation of T cell subsets, cells were stimulated for 6 h with 50 ng/ml PMA and 750 ng/ml ionomycin in complete RPMI 1640 medium at 37°C. After an additional 2 h of incubation with GolgiStop (BD Biosciences), stimulated cells were washed and stained with PE-Cy5–conjugated rat anti-mouse CD4 (clone RM4-5; eBioscience). Following fixation, cells were permeabilized and stained with FITC-conjugated rat anti-mouse IFN-γ (clone XMG1.2; eBioscience) and PE-conjugated rat anti-mouse IL-17A (clone eBio17B7; eBioscience). Cells were analyzed on a FACSCalibur flow cytometer (BD Biosciences) followed by data analysis with FlowJo software (Tree Star).
Luminex immunoassays
For quantitation of tissue CCL2 and CCL3 levels, colon tissue was harvested, snap-frozen, and stored at −80°C until use. The tissue was weighed and homogenized in PBS containing 1% Triton X-100 and Complete Protease Inhibitor (Roche). Following centrifugation, supernatants were collected and assayed by Luminex xMAP technology (Thermo Fisher) according to the manufacturer’s instructions. Results were normalized to total protein as determined by bicinchoninic acid assay (Bio-Rad).
Generation of bone marrow chimeras and dendritic cell depletion
To generate bone marrow chimeric mice, Il10−/− recipients were exposed to a single lethal dose of 9 Gy total body irradiation from a caesium-137 source. Within 6–8 h after irradiation, mice were given 4–6 × 106 bone marrow cells from sex-matched Il10−/− × Cd11c-DTR mice via tail vein injection and were rested for 8–10 wk to allow reconstitution. To deplete dendritic cells, Cd11c-DTR mice were treated with 4 ng/g diphtheria toxin i.p. (Sigma-Aldrich) 24 h prior to infection and every 48 h thereafter. Cell depletion was confirmed by FACS.
Neutralization of IL-12 p40
Anti–IL-12 p40 (clone C17.8, rat IgG2a) and its isotype control were purchased from Bio X Cell (West Lebanon, NH). Mice were given 1 mg of Ab s.c. 2 d prior to infection and every 7 d for up to 3 wk.
16S rRNA sequencing and analysis
For metagenomic studies, total DNA was extracted from unwashed colon tissue using MO BIO PowerSoil DNA Isolation Kit (MO BIO Laboratories). The V4 region of the bacterial 16S rRNA gene was amplified and sequenced in the MiSeq platform (Illumina) using a 2 × 250 bp paired-end protocol. The primers used for amplification contained adapters for MiSeq sequencing and single-index barcodes, allowing the PCR products to be pooled and sequenced directly (28). Using the UPARSE algorithm (29), 16Sv4 rDNA sequences were clustered into operational taxonomic units (OTUs) at a similarity cutoff value of 97%, and the OTUs were subsequently mapped using the SILVA Database (30). Abundances were recovered by mapping the demultiplexed reads to the UPARSE OTUs. Analysis and visualization of the microbial composition data were conducted in R. Significance was determined by nonparametric tests that included the Mann–Whitney U test and Kruskal–Wallis test. Correlations were determined with R’s base “lm” function for linear regression models, in which p values indicate the probability that the slope of the regression line is zero. The Monte Carlo permutation test was used to estimate p values for principal coordinate analysis (pCOA) plots. All p values were adjusted for multiple comparisons with the false discovery rate algorithm (31).
Antibiotic treatment
For modulation of the intestinal microbiota, an antibiotic mixture of 500 mg/l vancomycin hydrochloride, 500 mg/l imipenem, and 1000 mg/l neomycin was added to the drinking water every 3 d for 2–3 wk.
Statistical analysis
Statistical analysis was performed using SigmaPlot12 (Systat Software). To compare parasite loads, trophozoite numbers were log10 transformed, and mean and SEM were calculated. Samples without detectable trophozoites were assigned a log value of 3, which represents half of the detection limit of the assay. Differences in trophozoite numbers were evaluated by Wilcoxon rank-sum test, survival data were examined by Kaplan-Meier analysis, and all other comparisons between groups were evaluated by Student t test. Differences with a value of p <0.05 were considered significant.
Results
Giardia infection triggers colonic inflammation in IL-10–deficient mice
Giardia infection induces IL-17A and IFN-γ expressing CD4+ T cells (21, 32), although only the former are important for parasite control (21, 22). Despite the proinflammatory potential of these effectors, wild-type mice and humans generally exhibit little mucosal inflammation in the small intestine after acute infection (19). To determine whether IL-10 may be involved in this outcome, we first examined expression of the cytokine during the course of G. muris infection in adult C57BL/6 mice. At different times postinfection, total RNA was extracted from the jejunum (the primary site of infection) and assayed for IL-10 mRNA by qRT-PCR. IL-10 expression was significantly increased between 2 and 4 wk, with a maximum >60-fold induction after 3 wk (Fig. 1A), a time that coincides with a steep decrease in trophozoite burden in wild-type mice (Fig. 1B) and a reduction of elevated IL-17A expression to preinfection levels (21).
To determine the importance of IL-10 in the host response to Giardia infection, we compared the time course of G. muris infection in IL-10–deficient (Il10−/−) mice with wild-type controls. In both groups, peak trophozoite numbers were observed in the small intestine by 1 wk, followed by gradual decline over the ensuing 6 wk and clearance by 7 wk (Fig. 1B). No significant difference was observed between the groups at any time postinfection, indicating that IL-10 was not required for antigiardial host defense. However, 80% of the infected Il10−/− mice, but very few (<10%) of the infected wild-type controls, developed significant clinical symptoms, including diarrhea, lethargy, and weight loss within 2 wk (Fig. 1C). Induction of symptoms was largely dependent on infection because only a minority (28%) of uninfected Il10−/− littermate controls spontaneously developed any symptoms during the same follow-up period. Furthermore, marked diarrhea, as measured by a significant increase in fecal water content, was evident as early as 1 wk postinfection of Il10−/− animals and persisted until the end of the 7 wk experiment but was not observed at any time in parallel cohorts of uninfected Il10−/− or infected wild-type mice (Fig. 1D).
For clues about the underlying pathophysiology of the clinical symptoms, we examined the histology of the intestinal tract postinfection. Il10−/− mice, but not the wild-type controls, displayed modest mucosal inflammation in the small intestine 1 wk after G. muris infection (Supplemental Fig. 1), but the changes were transient and thus not consistent with the observed symptom persistence. Surprisingly, the colon, which is not a major site of Giardia infection, showed marked and progressive inflammation in Il10−/− mice postinfection, whereas infected wild-type mice had no signs of inflammation (Fig. 1E, 1F). Inflammatory changes occurred in the knockout mice within 1 wk postinfection and showed steady progression for 7 wk. Inflammation was characterized by epithelial hyperplasia, lamina propria infiltration with leukocytes, crypt abscesses, and submucosal edema (Fig. 1E–G). Quantitative histological analysis confirmed a significant increase in lamina propria leukocytes by 4–7 wk postinfection (Fig. 1H), whereas increased neutrophil recruitment, as determined by tissue myeloperoxidase levels, was already apparent by 12 d (Fig. 1I, left). Preceding this recruitment was a strong increase in the major neutrophil chemokine, CXCL5, in the colon (Fig. 1I, right). In contrast, wild-type controls showed no colonic neutrophil recruitment and only modest CXCL5 induction postinfection. These pathological features were similar to the spontaneous colitis known to develop in Il10−/− mice with increasing age (5) and were confirmed in our animal facility in aged (>6 mo) Il10−/− mice with spontaneous colitis (Supplemental Fig. 2).
To rule out the possibility that any bystander microbes in the G. muris cyst preparations may be responsible for disease, we treated the preparations with broad-spectrum antibiotics in vitro prior to infection but did not observe significant differences in the course of infection or disease outcome. We also isolated and prepared cysts from in vivo antibiotic-treated animals and again saw no impact on G. muris infection or colitis induction. Together, these results suggest that in young adult Il10−/− mice, Giardia infection triggers and exacerbates colitis with 100% penetrance, which would have developed spontaneously in only a small subset of these mice in a delayed fashion with age.
TCRα deficiency renders mice susceptible to Giardia-induced colitis
Intrigued that Giardia, an enteric pathogen which typically causes only minor and transient enteritis, triggered severe persistent colitis in Il10−/− mice, we next sought to determine whether this observation was unique to that particular colitis-prone model. Mice lacking the TCR α (Tcrα−/−) develop aging-related spontaneous inflammation that resembles human ulcerative colitis (24), so we investigated the impact of Giardia infection in these mice at a young, predisease age. Parasite levels were similar in these mice and wild-type controls 1 wk postinfection, but Tcrα−/− mice failed to clear infection thereafter, with ∼10,000-fold higher trophozoite numbers at week 7 (Fig. 2A), which confirms that intact T cell immunity is required for Giardia eradication (32). Similar to what was observed in Il10−/− animals, fecal water content was significantly increased after 1 wk (62.1 ± 3.5% on week 1 versus 50.0 ± 2.3% preinfection; p = 0.01) and steadily rose thereafter (78.4 ± 3.5% by week 7). No increase in fecal water was found in uninfected littermate controls. Histological examination of infected Tcrα−/− mice revealed significant colon inflammation within 1 wk and steady progression to severe colitis, characterized by extensive thickening of the mucosa, striking crypt hyperplasia, marked leukocyte infiltration in the lamia propria, and numerous crypt abscesses over the 7-wk course of infection (Fig. 2B–D). As for Il10−/− mice, histological disease penetrance was 100% in Tcrα−/− mice at week 7 (Fig. 2C). These data show that Giardia-triggered colitis can occur with different underlying susceptibility mechanisms of inflammation.
Role of adaptive immune cells and IL-12/IL-23 p40 in Giardia-triggered colitis
Both Il10−/− and Tcrα−/− mice developed colitis after Giardia infection, yet the latter lack normal T cells, raising the question of whether T cells play any role in the colitis observed in the absence of IL-10. To address this, we generated double knockout mice for IL-10 and Rag2 (Il10−/− × Rag2−/−) and challenged them with G. muris. As expected, the double knockout mice and Rag2−/− single knockout controls were unable to clear infection (Fig. 3A), confirming that adaptive immune cells are required for effective host defense against Giardia. The double knockout mice had a significant reduction in the severity of Giardia-triggered colitis relative to Il10−/− mice, although they still exhibited modest inflammation (Fig. 3B). In contrast, Rag2−/− mice, like wild-type mice, had no signs of colon inflammation postinfection (Fig. 3B). These findings show that adaptive immune cells, most likely T cells, are major mediators of colitis postinfection, but also suggest that innate immune cells contribute to inflammation.
To characterize the involvement of T cells in Giardia-triggered colitis in more detail, we examined T effector subsets by flow cytometry. The frequency of both Th1 and Th17 cells increased in the colon of Il10−/− mice after 1–2 wk (Fig. 3C). Similarly, infected wild-type mice showed an increase of both Th subsets in the colon, although the increase in Th1 cells was significantly lower compared with Il10−/− mice. A differential Th1 response was also observed in the small intestine, with the increase in Th1 cells peaking at 1 wk postinfection in Il10−/− mice (12.3 ± 0.9 in Il10−/− mice versus 1.2 ± 0.4% in wild-type mice; p < 0.001).
Induction of a transient Th1 response in the small intestine and a more persistent response in the colon suggested that Th1 cells may contribute to Giardia-induced colitis. IL-12/IL-23 p40 is a key cytokine involved in inducing and sustaining effector Th1 cells (33). Consistent with Th1 involvement, we found that IL-23 p40 mRNA expression was significantly increased in Il10−/− mice after 2–4 wk (Fig. 3D). To determine if IL-12/IL-23 p40 was important in mediating the host response to infection, we treated G. muris–infected Il10−/− mice with anti-p40 neutralizing or control Abs. Anti-p40 significantly attenuated the development of diarrhea (Fig. 3E) and colitis (Fig. 3F) 3 wk postinfection compared with control Abs, whereas it had no significant impact on host defense against the parasite (log10 trophozoites = 5.8 ± 0.3 in Il10−/− mice treated with anti-p40 versus 6.1 ± 0.3 in untreated controls; p = 0.7). These results indicate that colonic Th1 responses are unchecked upon Giardia infection in the absence of IL-10, and that IL-12/IL-23 p40 is major mediator of the ensuing colitis.
Dendritic cells are not required for Giardia-triggered colitis.
Development of T effector cells is controlled to a large extent by dendritic cells, so we questioned whether these cells played a role in colitis after Giardia infection of susceptible mice. Characterization of dendritic cell numbers by flow cytometry revealed only minimal changes in the numbers of CD11c+ MHCII+ B220− dendritic cells in the colon, small intestine, and MLN (Fig. 4A), suggesting that dendritic cells were not particularly impacted by Giardia infection or the infection-induced inflammation. To directly determine dendritic cell functions in the model, we crossed Il10−/− mice to mice expressing the DTR under the control of the dendritic cell–specific Cd11c promoter. Administration of diphtheria toxin resulted in effective CD11c+ cell deletion (>99% in the spleen and MLN). Depletion of CD11c+ dendritic cells impaired parasite clearance, as shown by 100-fold higher trophozoite numbers at week 3 (Fig. 4B), indicating that the cells played a role in mounting an effective host defense against the parasite. However, depletion had no significant impact on the development or severity of colitis induced by G. muris infection (Fig. 4B), indicating that colitis was not controlled by CD11c+ dendritic cells.
Mucosal macrophage expansion and role of CCR7 signaling in Giardia-induced colitis.
The unexpected lack of dendritic cell involvement in Giardia-triggered colitis prompted us to explore other innate immune cells in the model. Macrophages in particular not only serve as a link to adaptive immunity but also promote inflammation independent of T cells. Flow cytometry of mucosal leukocyte populations revealed that G. muris infection induced a marked 4-fold increase in CD11b+ CD11c− macrophages in the colon of Il10−/− mice after 2 wk (Fig. 5A), which was paralleled by induction of macrophage chemokines CCL2 and CCL3 (Fig. 5B). By comparison, no change in macrophage numbers was observed in the small intestine or MLN of G. muris–infected Il10−/− mice.
The marked macrophage expansion in the colon raised the possibility that their local activation might induce or sustain colon inflammation. Although multiple pathways have been proposed for the inflammation-promoting functions of certain macrophage subsets, two mechanisms are particularly relevant for intestinal homeostasis. One involves downregulation of inflammation-inducing genes by IL-10–dependent macrophage deactivation (34). In this context, loss of IL-10 leads to intestinal accumulation of activated macrophages and subsequent inflammation, which are findings consistent with our observations in Il10−/− mice. Another mechanism relates to CCR7-dependent migration of activated macrophages out of the mucosa (35). In the absence of CCR7 signaling, activated cells accumulate in the intestine and may cause inflammation. To explore whether this process might be relevant to Giardia-triggered colitis, we infected CCR7 knockout mice (Ccr7−/−) with G. muris. Surprisingly, Ccr7−/− mice were relatively protected against acute infection, with a 100-fold lower peak trophozoite burden after 7 d compared with wild-type controls (Fig. 6A). Despite the reduced infectious load, mice developed marked colitis that commenced within 1 wk and reached maximal severity by 3–7 wk (Fig. 6B). These results are consistent with the notion that CCR7-dependent cell migration of macrophages and perhaps other potentially colitogenic immune cells out of the mucosa normally attenuate inflammation, but if this homeostatic process is disturbed, inflammation can ensue upon mucosal microbial challenge.
Giardia-triggered colitis is dependent on the microbiota
Alterations in the intestinal microbiota are associated with and contribute to intestinal inflammation (36). To determine if Giardia infection impacts the microbiome in Il10−/− and Tcra−/− mice, DNA from unwashed colon tissues were analyzed using the Illumina MiSeq platform. pCOA to assess overall compositional differences between individual mice revealed that prior to Giardia infection, Il10−/− and Tcra−/− mice had distinct microbiome profiles compared with wild-type mice (Fig. 7A, left). Two weeks postinfection, clear shifts in the composition were observed, with the microbiota of Il10−/− and Tcra−/− mice converging in the coordinate space and becoming more similar to that in wild-type mice (Fig. 7A, right). Giardia infection of Il10−/− mice also led to a significant reduction in microbial diversity, which is a trend also observed in Tcra−/− but not wild-type mice (Fig. 7B). Analysis of the dominant phyla revealed that Giardia infection caused a reduction in Bacteroidetes and Firmicutes but caused increases in Proteobacteria in both Il10−/− and Tcra−/− mice (Fig. 7C). Expansion of Proteobacteria had previously been associated with intestinal inflammation (37). At the family level, major changes in abundance were observed for Provotellaceae, Rikenellaceae, Deferribacteraceae, Lachnospiraceae, and Helicobacteraceae, the most dominantly detected family in wild-type controls. Further analysis of the OTUs classified in the Helicobacteraceae family demonstrated >99% identity to two particular species, Helicobacter bilis and Helicobacter mastomyrinus, which were increased in abundance after Giardia infection, although they were also identified at lower levels in uninfected controls.
To determine whether the microbiota contributes to the development of Giardia-triggered colitis, we treated Il10−/− mice with a combination of broad-spectrum antibiotics before and during G. muris infection. The antibiotics were chosen to have no direct antigiardial activity, and sustained administration to wild-type mice had no significant impact on the course of G. muris infection (Supplemental Fig. 3). Il10−/− mice given antibiotics 2 d prior to Giardia infection showed significantly less colon inflammation, as evidenced by reductions in mucosal thickening, crypt hyperplasia, and leukocytic infiltration compared with infected but untreated Il10−/− mice (Fig. 8A). Administration of antibiotics 3 wk after G. muris infection also reduced the severity of colitis, although not as effectively as the preGiardia infection regimen (Fig. 8B). These findings strongly suggest that Giardia-triggered colitis can be prevented or at least attenuated by reducing or modifying the intestinal microbiota.
MyD88 is required for driving Giardia-triggered colitis but not antigiardial defense
To further define the role of the bacterial microbiota in mediating Giardia-triggered colitis, we explored the functions of the common TLR signal transducers, MyD88. This transducer is central to the recognition of bacterial products and in antibacterial host defense (25). We constructed double-mutant mice lacking both IL-10 and MyD88 (Il10−/− × Myd88−/−) and infected them with G. muris. The double-mutant mice were equally infected after 1 wk and cleared the parasites as effectively thereafter as Il10−/− controls (Fig. 8C), indicating that MyD88 signaling is not required for normal host defense against Giardia. However, infected double-mutants had strongly attenuated colitis after 5 wk compared with the infected Il10−/− single mutants (Fig. 8D). These results provide independent support for the conclusion of the antibiotics studies that the intestinal bacterial microbiota, as recognized by signaling through MyD88, mediates Giardia-triggered colitis in the absence of IL-10.
Discussion
Our studies demonstrate that colonization of the small intestine by a lumen-dwelling protozoan parasite, which by itself causes minimal mucosal inflammation in a normal host (19), can trigger chronic colitis in genetically susceptible mice. Giardia-triggered colitis in at least one of the models, Il10−/− mice, was associated with mucosal expansion of neutrophils and macrophages and induction of a persistent Th1 response that contributed to elevated IL-12/IL-23 p40 expression critical for the disease manifestation. Our findings parallel prior observations that subclinical norovirus infection can initiate chronic intestinal inflammation in Il10−/− mice (38). In both models, the acute pathogen infection is transient and minimally symptomatic, whereas the intestinal microbiota was required for disease induction and persistence beyond the initial infection. Important differences exist between these microbial triggers, as norovirus infects multiple cell types throughout the intestinal tract, whereas Giardia is noninvasive and largely restricted to the proximal small intestine. Nonetheless, the data suggest that the development of chronic inflammation can result in a susceptible host from complex temporal host-pathogen–microbiota interactions that are not unique to particular enteric pathogens.
Synergistic effects of genetic susceptibility and pathogen exposure have also been implicated in the development of IBD in humans. Several viral pathogens have been associated with IBD (38–41). Similarly, a number of bacterial pathogens, such as Salmonella and Campylobacter, have been linked to IBD (42–44). For example, adherent invasive Escherichia coli is commonly found in Crohn’s disease patients with ileal involvement (45), whereas Yersinia enterocolitica has been associated with ulcerative colitis in adults (46). Yet, the available evidence does not suggest that a single viral or bacterial pathogen is the definitive cause of IBD (4, 42, 47). Instead, and consistent with our findings, a range of microbial pathogens can initiate inflammatory processes in the intestinal tract that ultimately become self-perpetuating, independent of the original trigger.
Colitis induction after Giardia infection required the resident microbiota and MyD88 signaling, indicating that host interactions with the endogenous microbiota are essential for disease induction and persistence. Similarly, development of spontaneous colitis in aged Il10−/− mice relies on the intestinal microbiota and microbial sensing through MyD88-dependent pathways (48, 49). Besides the overall importance of the microbiota, we also found that the composition of the intestinal microbiota changed in parallel with disease development in both Il10−/− and Tcra−/− mice. Although additional studies are required to determine why certain commensal species expand following Giardia infection, the parasite promoted bloom of two colitogenic pathobionts, H. bilis and H. mastomyrinus. These two Helicobacter species have been associated with colitis in other models (50, 51), although it is not clear that either is sufficient to cause colitis. Inflammation induced in Il10−/− mice with another colitogenic Helicobacter species, H. hepaticus, is accompanied and modulated by the composition of the intestinal microbiota (4), yet the altered microbiota alone cannot induce inflammation after colonization of germ-free Il10−/− mice (52). These findings underline that the causal relationship between microbiota changes and inflammation observed in our studies and those prior remains to be firmly established. Monoassociation studies in gnotobiotic mice have shown that several different bacterial members of the normal microbiota can induce colitis in genetically susceptible hosts (52, 53), so it is likely that a constellation of commensal bacteria with relatively greater potential to act as pathobionts rather than a single dominant bacterial pathobiont are responsible for colitis induction upon microbial triggering in a susceptible host.
Our data on persistent mucosal inflammation upon transient Giardia infection in mice are consistent with clinical reports that G. lamblia infection can cause protracted intestinal symptoms and even mucosal inflammation (54–56). Following a large waterborne outbreak of giardiasis in Norway, over a third of patients presented with irritable bowel syndrome and chronic fatigue 6 y after successful treatment of the infection (56). Postinfectious syndromes can occur postinfection with other enteric pathogens, particularly Salmonella spp., E. coli O157, and C. jejuni (57). The underlying mechanisms are not well understood in most cases but may involve localized residual inflammation impacting the composition of the microbiota or functions of the enteric nervous systems and serotonin transporters (58, 59). In giardiasis, it has been postulated that disruptions in the microbiota predispose individuals to postinfectious gastrointestinal disorders (60). Our data suggest that postinfectious complications after giardiasis can result from interactions between the microbiota and dysregulated mucosal immunity, suggesting that functional and genetic testing of patients with protracted postgiardiasis intestinal disorders may reveal subtle immunological abnormalities.
Giardia infection caused a marked induction of IL-10 in the jejunum, the primary site of infection, yet the cytokine was not required for the development of protective immunity against giardiasis. Similar observations have been made in other infections with enteric pathogens. For example, infection with Salmonella enterica serovar Typhimurium led to increased IL-10 expression in the spleen, but neutralization of IL-10 did not modify the course of infection (61). For another enteric pathogen, Citrobacter rodentium, IL-10–deficient mice also had no clearance defect and could, in fact, eradicate intestinal infection moderately faster than wild-type controls, suggesting that the cytokine was not only not essential but modestly inhibitory for adaptive immunity against this pathogen (62). For Giardia infection, we can conclude that IL-10 had no positive or negative role in acute intestinal pathogen defense but was critical for controlling the long-term, indirect immunological sequelae of the infection. This intriguing concept of a key role of an immune regulator in maintaining long-term homeostasis upon transient microbial exposure without contributing to acute host defense may have relevance to understanding other protracted postinfectious syndromes.
Acknowledgements
We thank Kim Nguyen, Lauren Gima, MaChesha Banks, Raymond Dann, and the staff of the U.S. Department of Veterans Affairs San Diego Flow Cytometery Core Laboratory; the Alkek Center for Metagenomics and Microbiome Research at Baylor College of Medicine in Houston for technical assistance; and Lucia Hall and the staff of the University of California, San Diego and the University of Texas Medical Branch animal care programs for animal husbandry.
Footnotes
This work was supported by National Institutes of Health Grants AI112594 and AI125860. S.M.D. was supported by a Career Development award from the Crohn’s and Colitis Foundation of America and the Institute for Translational Sciences at the University of Texas Medical Branch, supported in part by a Clinical and Translational Science Award (UL1 TR001439) from the National Center for Advancing Translational Sciences, National Institutes of Health. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
The online version of this article contains supplemental material.
References
Disclosures
The authors have no financial conflicts of interest.