Modulation of Acute Diarrheal Illness by Persistent Bacterial Infection

Male and female 5- to 12-week-old C57BL/6J (The Jackson Laboratory, Bar Harbor, ME) mice were used for all studies. All experiments were approved by the Massachusetts Institute of Technology Committee on Animal Care. Mice were fed a rodent diet and water ad libitum and housed in microisolator cages that were maintained in facilities approved by the Association for Assessment and Accreditation of Laboratory Animal Care, International. Until experimentally inoculated, the mice were maintained specific pathogen free of known murine bacterial, viral, and parasitic infections including all known Helicobacter spp. For each experiment, the mice were divided into four treatment groups: uninoculated (n = 10), inoculated with H. hepaticus (n = 10), inoculated with C. rodentium (n = 10), and inoculated with H. hepaticus followed by C. rodentium (n = 10). Five independent experiments were conducted: two with necropsy at 1 week after C. rodentium inoculation, one with necropsy at 2 wpi, one with necropsy at 3 wpi, and one with necropsy at 4 wpi.

H. hepaticus 3B1 (ATCC 51449) was grown on tryptic soy agar supplemented with 5% sheep red blood cells or in tryptic soy broth (TSB) supplemented with 5% fetal calf serum at 37°C in a microaerobic environment (80% N₂, 10% H₂, and 10% CO₂). H. hepaticus inocula were prepared from 3-day liquid cultures from which ∼2 × 10⁸ bacteria (estimated from the optical density at 600 nm) were administered in 200 μl of TSB via intragastric gavage to individual mice. Uninoculated mice were gavaged with 200 μl of sterile TSB. For C. rodentium infections, mice were gavaged with ∼2 × 10⁹ bacteria from an overnight culture of Kan^rC. rodentium (DBS120) in 100 μl of Luria-Bertani broth, 7 to 8 weeks after H. hepaticus infection. Helicobacter status was confirmed with an all-Helicobacter PCR of fecal DNA as previously described (48). C. rodentium fecal shedding was determined by plating serial dilutions of fecal slurries (10% [wt/vol] in phosphate-buffered saline) on Luria-Bertani agar with selection for kanamycin.

Body weights were monitored every 3 to 4 days. Mice were euthanized and excluded from the study if they lost >20% of their body weight.

At necropsy fecal and tissue samples were collected. Distal colon (∼0.5 cm) was snap-frozen in liquid nitrogen and stored at −80°C until it was used for RNA isolation. The remaining colon was fixed in 10% formalin, paraffin embedded, sectioned at 5 μm, and stained with hematoxylin and eosin for histologic evaluation. Colonic tissue sections were scored on a scale of 0 to 4 (0, no lesion; 1, minimal; 2, mild; 3, moderate; and 4, severe) for inflammation, edema, hyperplasia, dysplasia, and epithelial defects by a board-certified blinded pathologist. Lesion scores are presented as histologic colitis indices that are a sum of all five categorical scores (maximum of 20). Foxp3 immunohistochemistry was performed as previously described (32), using Foxp3 antibody (FJK-16S; eBiosciences, San Diego, CA). Cells expressing Foxp3⁺ were counted in the distal to mid-colon at a magnification of ×20 (1 field is 1.00 mm²), excluding gut-associated lymphoid tissue. Ten fields were counted per mouse, and results are presented as the average number of Foxp3⁺ cells/mm² of colon. F4/80 immunohistochemistry was performed as described for Foxp3 but using F4/80 antibody (CI:A3-1; Abcam, Cambridge, MA). Cells expressing F4/80 were counted in 20 fields of distal colon at ×40 magnification (1 field is 0.26 mm²), also excluding gut-associated lymphoid tissue. Results are presented as the average number of F4/80⁺ cells/mm² of colon.

Total RNA was isolated from distal colon using TRIzol reagent (Invitrogen, Foster City, CA), cleaned up with an RNeasy Kit (Qiagen Sciences, MD), and reverse transcribed (Invitrogen) following the manufacturers' protocols. Quantitative real-time PCR was performed on cDNA using TaqMan Gene Expression Assays (Applied Biosystems, Foster City, CA) specific for glyceraldehyde-3-phosphate dehydrogenase (GAPDH; assay no. Mm99999915_g1), IL-6 (Mm00446190_m1), MCP-1 (Mm00441242_m1), TNF-α (Mm99999068_m1), IFN-γ (Mm99999071_m1), IL-10 (Mm00439616_m1), transforming growth facter β (TGF-β; Mm00498234_m1), IL-1β (Mm00434228_m1), IL-12/IL-23p40 (Mm99999067_m1), IL-12p35 (Mm01208555_m1), IL-23p19 (Mm00518984_m1), and IL-17 (Mm00439619_m1). Each sample was calibrated to internal GAPDH levels and normalized to the average value of control (uninoculated mice) samples at the same time point.

Statistical significance in bacterial counts, weight change, disease indices, mRNA expression, and F4/80⁺ and Foxp3⁺ cell numbers was determined by two-way analysis of variance (ANOVA) followed by Bonferroni posttests. A Spearman correlation was used to evaluate the correlation between disease indices and cytokine mRNA expression levels. All analyses were done with GraphPad Prism software, version 4.0. P values of <0.05 were considered significant.

In agreement with previous reports where the Helicobacter spp. status of mice was not specified (21, 43, 44), fecal shedding of C. rodentium in Helicobacter-free C57BL/6J mice reached a maximum of 10⁹ CFU/g of feces between 1 and 2 wpi (Fig. 1a). Over the next week C. rodentium was cleared, and fecal shedding was undetectable (<10³ CFU/g of feces) by 3 wpi. Because the severity of C. rodentium-induced disease can be decreased by as little as a 1 log reduction in peak bacterial load (19), we carefully compared fecal shedding with and without H. hepaticus infection. Persistent infection with H. hepaticus, confirmed by PCR on DNA isolated from fecal pellets, did not affect fecal shedding of C. rodentium (Fig. 1a). Body weight change in mice infected with H. hepaticus alone was not different from uninoculated mice, confirming the subclinical nature of the infection. In contrast, body weight loss was apparent in Helicobacter-free C. rodentium-inoculated mice by 1 wpi. These mice lost an average of 4% of their initial body weight by 12 days postinoculation (Fig. 1b) (P < 0.001). By 3 wpi these mice recovered, and their body weights were not significantly different from those of uninoculated mice. The extent of weight loss in C. rodentium-inoculated mice with a persistent subclinical H. hepaticus infection was comparable to that of Helicobacter-free C. rodentium-inoculated mice (Fig. 1b). However, concurrently infected mice did not regain body weight at 3 wpi (P < 0.001 compared to Helicobacter-free C. rodentium-inoculated mice) and remained underweight compared to control mice through 4 wpi (P < 0.001). Given that persistent subclinical H. hepaticus infection caused delayed recovery from weight loss, colonic lesions were evaluated to ascertain whether this morbidity was associated with an enhanced chronicity of mucosal disease.

H. hepaticus infection alone did not cause intestinal lesions, and histologic colitis in these animals was not significantly different from that of uninoculated mice (Fig. 2a, d, and g). By 1 wpi Helicobacter-free C. rodentium-inoculated mice developed colonic lesions consisting of inflammation, edema, epithelial destruction, and hyperplasia. Categorical lesion scores for each of these parameters were summed to form a histologic colitis index. At 1 wpi the median histologic colitis index in Helicobacter-free C. rodentium-inoculated mice was 4.0 (range, 0.0 to 6.0) (Fig. 2g) (P < 0.001 compared to uninoculated mice). Over the following week, histologic colitis reached maximal severity, characterized by gland elongation, goblet cell depletion, mucosal and submucosal inflammation, and epithelial defects, including erosions, crypt atrophy, and mild dysplasia (Fig. 2b). At peak severity, Helicobacter-free C. rodentium-inoculated mice had an index of 9.0 (range, 7.0 to 12.0) (Fig. 2b and g). Disease resolved in these mice during 3 and 4 wpi (at 4 wpi, index of 2.0; range, 0.5 to 4.5) (Fig. 2e and g). Concurrent infection did not significantly alter the maximal severity of histologic colitis at 2 wpi (index of 8.5; range, 5.5 to 10.0) (Fig. 2c and g), but concurrent H. hepaticus infection did delay the resolution of C. rodentium-induced lesions. At 4 wpi the index of concurrently infected mice was 4.5 (range, 2.5 to 7.5) compared with 2.0 (range, 0.5 to 4.0) in Helicobacter-free C. rodentium-inoculated mice (Fig. 2e to g) (P < 0.001). Although lesions in both Helicobacter-free and concurrently infected mice were resolving at 4 wpi, inflammation and hyperplasia were more severe in concurrently infected mice (Fig. 2 h) (P < 0.001 and P < 0.05, respectively). Thus, delayed recovery from weight loss was associated with colitis of a longer duration. Morphometrics were used to determine if the chronicity of disease in concurrently infected mice was associated with an increase in the number of macrophages, which are a hallmark of chronic inflammation (26).

Macrophages expressing F4/80 (a macrophage-specific marker) were present in low numbers in the colon of uninoculated mice (Fig. 3a and b) (average of all time points, 20 ± 18 cells/mm²). No significant change in macrophage numbers was seen in colons of H. hepaticus-infected animals, consistent with a lack of disease (Fig. 3a and c) (average of all time points, 16 ± 20 cells/mm²). Macrophages did accumulate in the colon during C. rodentium infection (Fig. 3a) (P < 0.001). Macrophage numbers were maximal in Helicobacter-free C. rodentium-inoculated mice at 1 wpi (175 ± 36 cells/mm²), decreased by 2 wpi (72 ± 23 cells/mm²), and remained constant at 3 and 4 wpi (98 ± 34 and 125 ± 28 cells/mm², respectively) (Fig. 3a and d). Macrophage numbers in concurrently infected mice were not significantly different from those in mice infected with C. rodentium alone at 1, 2, or 3 wpi (Fig. 3a). However, at 4 wpi, instead of staying constant at 2 and 3 wpi, as in Helicobacter-free C. rodentium-infected mice, the number of colonic macrophages significantly increased to 256 ± 40 cells/mm² in concurrently infected mice (Fig. 3a, d, and e) (P < 0.01), indicating chronic mucosal inflammation. Natural Foxp3⁺ T_reg cells play a key role in limiting immunopathology during infection and chronic inflammatory diseases as well as in contributing to peripheral tolerance (1). Morphometrics were used to enumerate Foxp3⁺ cells to determine if greater chronicity of histologic colitis was associated with accumulation of natural T_reg cells.

As expected in the absence of histologic colitis, H. hepaticus infection alone caused no colonic accumulation of Foxp3⁺ T_reg cells (Fig. 3f) compared to uninoculated mice (1 ± 1 Foxp3⁺ cells/mm²). Early after C. rodentium inoculation (1 wpi) there was no significant increase in the number of colonic Foxp3⁺ cells in Helicobacter-free mice, suggesting that natural T_reg cells accumulate in response to colitis, not infection per se. With the development of histologic colitis at 2 and 3 wpi, there were increased numbers of Foxp3⁺ T_reg cells in the colon of Helicobacter-free C. rodentium-inoculated mice (10 ± 6 and 21 ± 5 Foxp3⁺ cells/mm², respectively) (Fig. 3f and g) (P < 0.001). By 4 wpi, colonic numbers of Foxp3⁺ T_reg cells in Helicobacter-free C. rodentium-inoculated mice had declined and were not significantly different from basal numbers in uninoculated or H. hepaticus-infected mice (Fig. 3f and i). Colonic numbers of Foxp3⁺ T_reg cells in concurrently infected mice began to accumulate earlier (1 wpi) than in mice infected with C. rodentium alone. At the peak of histologic colitis at 2 and 3 wpi, accumulation of Foxp3⁺ T_reg cells in concurrently infected mice was comparable to the level in Helicobacter-free C. rodentium-inoculated mice (Fig. 3f to h). However, at 4 wpi Foxp3⁺ T_reg-cell numbers remained elevated in concurrently infected mice, in contrast to the return to basal levels in mice infected with C. rodentium alone (Fig. 3f and j) (P < 0.001). The continued presence of natural T_reg cells is consistent with a delayed recovery from histologic colitis, resulting in greater chronicity of disease.

Colonic cytokine expression levels were measured to further characterize the mechanism by which persistent subclinical infection modulated the host response to acute diarrheal illness. Adaptive immunity is required for the clearance of C. rodentium but also contributes to histologic colitis (5, 13). IL-17-producing T cells have also been recently implicated in disease pathogenesis (23). Elevated expression of the Th1 cytokines IFN-γ, TNF-α, and IL-12 coincided with increasing histologic colitis, reaching maximal expression at 2 wpi (Fig. 4a and Fig. 5) (P < 0.01 compared with uninoculated or H. hepaticus-infected mice). Expression levels of these proinflammatory cytokines declined at 3 and 4 wpi as lesions resolved. The type 2 cytokines IL-4 and IL-13 were not consistently detected in distal colons of control or infected mice. Additionally, the IL-10 expression profile mimicked that of inflammatory cytokines, possibly as a feedback mechanism to limit collateral inflammatory damage (Fig. 5). No significant changes in TGF-β expression occurred although expression trended higher at 2 wpi (Fig. 5). IL-23p19 expression was unaltered by C. rodentium infection, indicating either constitutive expression or no transcriptional induction of IL-23, a Th17 maintenance (46, 49) and Foxp3⁺ T-cell inhibitory factor (14). IL-17 expression increased during C. rodentium infection, reaching a plateau at 3 wpi (Fig. 4a). The association between C. rodentium disease pathogenesis and Th17 described previously was based on observations made at 8 days postinoculation (23). We also found greater colonic expression of IL-17 than IFN-γ at 1 wpi; however, later in the progression of disease IFN-γ and MCP-1 were found to predominate. Over the course of 4 weeks of infection, the dominant, IFN-γ nature of C. rodentium disease was apparent.

Cytokine expression and histologic colitis were correlated for individual mice in all infection groups at each time point. C. rodentium colitis correlated with colonic expression of IFN-γ but not with IL-17 (Fig. 4b). Analysis revealed a strong positive correlation between histologic colitis severity and levels of the proinflammatory cytokines IL-6, MCP-1, TNF-α, IL-1β, IL-12p35, IL-12/23p40, and in particular IFN-γ (Spearman r for IFN-γ of 0.763; P = 0.0032). Anti-inflammatory IL-10, TGF-β, and transcription factor Foxp3 expression levels had weak (Spearman r values of 0.557, 0.333, and 0.362, respectively) but significant (P < 0.01) correlation with histologic colitis severity. These weak correlations are likely due to an anti-inflammatory response to counteract the inflammation in the tissue. In contrast to other proinflammatory cytokines, such as IFN-γ, neither IL-23p19 nor IL-17 was correlated with histologic colitis (Spearman r for IL-17 of 0.097; P > 0.05).

Persistent concurrent H. hepaticus infection did not affect colonic cytokine expression at 1 wpi (Fig. 4a). However, at 2 wpi, concurrently infected mice had significantly lower levels of IFN-γ (773-fold over uninoculated controls) and MCP-1 (45-fold) message than Helicobacter-free C. rodentium-inoculated mice (1,756-fold [P < 0.05] and 160-fold [P < 0.001], respectively) (Fig. 4a and 5). No significant differences in colonic cytokine expression were noted at 3 wpi between concurrently infected mice and mice inoculated with C. rodentium alone. By 4 wpi, when other inflammatory cytokines were returning to basal expression, IL-17 message continued to increase in concurrently infected mice, reaching a 202-fold increase over uninoculated mice. This elevated expression of IL-17 was significantly increased compared with Helicobacter-free C. rodentium-inoculated mice (92-fold over uninoculated mice) at 4 wpi (Fig. 4a) (P < 0.05). Increased IL-17 expression in concurrently infected mice was not accompanied by a change in IL-23p19 or IL-12/23p40. Overall, persistent H. hepaticus infection modulated C. rodentium-induced cytokine expression by slightly suppressing peak levels of both IFN-γ and MCP-1, as well as by modestly enhancing IL-17 expression later in the course of the disease. Decreased IFN-γ and MCP-1 expression at 2 wpi was not associated with reduced severity of colitis or body weight loss. However, the increase in IL-17 was associated with delayed recovery from weight loss, chronic colitis, and macrophage and natural T_reg-cell accumulation in the colon. Given that IL-17 contributes to chronic inflammation (47) and inflammatory bowel disease (11), the association of increased IL-17 expression with increased morbidity suggests that it may prevent resolution of acute disease during self-limiting infection and may promote chronic disease progression.