Animal experiments.

Specific-pathogen-free (SPF) female mice were from Harlan Winkelmann (C57BL/6; 6 to 8 weeks old; Borchen, Germany) or Charles River (LTβR^−/− mice, C57BL/6 background; 6 to 8 weeks old; Sulzfeld, Germany [25]). Genotypes were verified by PCR typing with the primers 5′-CGG GTC TCC GAC CTA GAG ATC-3′ and 5′-GAG GTG GGT GGA TTG GAA AGA G-3′.

For the experiments, animals were housed individually or in groups of up to five animals under standard barrier conditions in individually ventilated cages (Tecniplast, Buguggiate, Italy) equipped with steel grid floors and autoclaved filter paper at the Max von Pettenkofer Institute (Munich, Germany) or BZL (Zurich, Switzerland).

Water and food were withdrawn 4 h before per os (p.o.) treatment with 20 mg of streptomycin (75 μl of sterile solution or 75 μl of sterile water [control]). Afterward, animals were supplied with water and food ad libitum. At 20 h after streptomycin treatment, water and food were withdrawn again for 4 h before the mice were infected with 10⁸ CFU of serovar Typhimurium (50-μl suspension in PBS p.o.) or treated with sterile PBS (control). Thereafter, drinking water ad libitum was offered immediately and food 2 h postinfection (p.i.). At the indicated times p.i., mice were sacrificed by CO₂ asphyxiation, and tissue samples from the intestinal tracts, mesenteric lymph nodes, spleens, and livers were removed for analysis.

Animal experiments were approved by the German and Swiss authorities and performed according to the legal requirements.

Analysis of serovar Typhimurium loads in intestines, mesenteric lymph nodes, spleens, and livers.

Two fresh fecal pellets were placed in 500 μl of 4°C cold PBS and suspended homogeneously on ice by vortexing and pipetting. Intestinal contents from the ileum, cecum, or colon were collected at the indicated times p.i. and weighed before resuspending them in 500 μl of 4°C cold PBS. The numbers of CFU were determined by plating appropriate dilutions on MacConkey agar plates (streptomycin at 50 μg/ml). The minimum detectable value was 5 CFU/fecal pellet and 10 CFU/sample (between 25 and 150 mg) of intestinal content.

Intestinal colonization by Lactobacillus spp. was determined by resuspending 50 mg of cecal content in 500 μl of ice-cold PBS, plating appropriate dilutions on MRS agar (Biolife, Milano, Italy) without delay, and incubation for 48 h at 37°C under an anaerobic CO₂-enriched atmosphere (Anaerocult A).

To analyze the colonization, mesenteric lymph nodes, spleens, and livers were removed aseptically and homogenized in 4°C cold PBS (0.5% Tergitol, 0.5% bovine serum albumin) by using a Potter homogenizer. The numbers of CFU were determined by plating appropriate dilutions on MacConkey agar plates (streptomycin at 50 μg/ml). The minimal detectable values were 20 CFU/organ in the spleen, 100 CFU/organ in the liver, and 10 CFU/organ in the mesenteric lymph nodes.

Histological procedures.

Segments of the ileum, cecum, and colon were fixed and embedded in paraffin according to standard procedures. Alternatively, tissue samples were embedded in O.C.T. (Sakura, Torrance, Calif.), snap-frozen in liquid nitrogen, and stored at −80°C. Cryosections (5 or 30 μm) were mounted on glass slides, air dried for 2 h at room temperature, and stained with hematoxylin and eosin (H&E). Pathological evaluation was performed by two pathologists in a blinded manner.

Based on an earlier study (49), we developed a scoring scheme for quantitative pathological analysis of cecal inflammation. H&E-stained sections (5 μm) were scored independently by two pathologists in a blinded manner as follows.

(i) Submucosal edema.

Submucosal edema was scored as follows: 0 = no pathological changes; 1 = mild edema (the submucosa is <0.20 mm wide and accounts for <50% of the diameter of the entire intestinal wall [tunica muscularis to epithelium]); 2 = moderate edema; the submucosa is 0.21 to 0.45 mm wide and accounts for 50 to 80% of the diameter of the entire intestinal wall; and 3 = profound edema (the submucosa is >0.46 mm wide and accounts for >80% of the diameter of the entire intestinal wall). The submucosa widths were determined by quantitative microscopy and represent the averages of 30 evenly spaced radial measurements of the distance between the tunica muscularis and the lamina mucosalis mucosae.

(ii) PMN infiltration into the lamina propria.

Polymorphonuclear granulocytes (PMN) in the lamina propria were enumerated in 10 high-power fields (×400 magnification; field diameter of 420 μm), and the average number of PMN/high-power field was calculated. The scores were defined as follows: 0 = <5 PMN/high-power field; 1 = 5 to 20 PMN/high-power field; 2 = 21 to 60/high-power field; 3 = 61 to 100/high-power field; and 4 = >100/high-power field. Transmigration of PMN into the intestinal lumen was consistently observed when the number of PMN was >60 PMN/high-power field.

(iii) Goblet cells.

The average number of goblet cells per high-power field (magnification, ×400) was determined from 10 different regions of the cecal epithelium. Scoring was as follows: 0 = >28 goblet cells/high-power field (magnification, ×400; in the cecum of the normal SPF mice we observed an average of 6.4 crypts/high-power field and the average crypt consisted of 35 to 42 epithelial cells, 25 to 35% of which were differentiated into goblet cells); 1 = 11 to 28 goblet cells/high-power field; 2 = 1 to 10 goblet cells/high-power field; and 3 = <1 goblet cell/high-power field.

(iv) Epithelial integrity.

Epithelial integrity was scored as follows: 0 = no pathological changes detectable in 10 high-power fields (×400 magnification); 1 = epithelial desquamation; 2 = erosion of the epithelial surface (gaps of 1 to 10 epithelial cells/lesion); and 3 = epithelial ulceration (gaps of >10 epithelial cells/lesion; at this stage, there is generally granulation tissue below the epithelium).

We averaged the two independent scores for submucosal edema, PMN infiltration, goblet cells, and epithelial integrity for each tissue sample. The combined pathological score for each tissue sample was determined as the sum of these averaged scores. It ranges between 0 and 13 arbitrary units and covers the following levels of inflammation: 0 = intestine intact without any signs of inflammation; 1 to 2 = minimal signs of inflammation (this was frequently found in the ceca of SPF mice; this level of inflammation is generally not considered as a sign of disease); 3 to 4 = slight inflammation; 5 to 8 = moderate inflammation; and 9 to 13 = profound inflammation.

Statistical analysis.

Statistical analysis of the cecum weight, the individual pathological scores for submucosal edema, PMN infiltration, goblet cells, and epithelial integrity and for the combined pathological score was performed by using the exact Mann-Whitney U test and SPSS version 11.0 software. P values of <0.05 were considered statistically significant. This procedure was adopted from that of Madsen et al. (49).

To allow the statistical analysis of the bacterial loads, values for animals yielding “no CFU” were set to the minimal detectable value (fecal pellet = 5 CFU; spleen = 20 CFU; liver = 100 CFU; mesenteric lymph nodes = 10 CFU; intestinal content = between 67 and 400 CFU [see above]). Afterward, the median values were calculated by using Microsoft Excel XP, and statistical analysis was performed by using the exact Mann-Whitney U test and the SPSS version 11.0 software. P values of <0.05 were considered statistically significant.

Quantitative analysis of serovar Typhimurium colitis in streptomycin-pretreated mice.

To explore whether streptomycin-pretreated mice might provide a useful model for studying serovar Typhimurium colitis, we performed a more detailed analysis. A total of 24 mice were divided into four groups of 6 mice. The first group was pretreated with streptomycin 24 h prior to p.o. infection with 10⁸ CFU of serovar Typhimurium (Fig. ​(Fig.3).3). One control group was pretreated with water 24 h prior to infection p.o. with 10⁸ CFU of serovar Typhimurium. The two other control groups were pretreated either with streptomycin or with water 24 h prior to mock infection by oral treatment with sterile PBS. The mice were killed 2 days p.i., and we analyzed the bacterial colonization (Fig. ​(Fig.3)3) and pathological changes in the ceca (Fig. ​(Fig.4,4, 5, and 6).

Open in a separate window

FIG. 3.

Effect of the streptomycin pretreatment on the colonization of the intestine and internal organs by serovar Typhimurium. Four groups of six mice (C57BL/6) were pretreated with streptomycin (solid symbols) or sterile water (open symbols) and infected with 10⁸ CFU of serovar Typhimurium SL1344 (circles) or mock infected with sterile PBS (triangles; see Materials and Methods). The mice were sacrificed 2 days p.i., and we analyzed the bacterial colonization. (A) Bacterial loads in the intestinal contents of the medial ileum (left panel), the terminal ileum (middle panel), and the cecum (right panel). (B) Bacterial loads in the mesenteric lymph nodes (left panel), the spleen (middle panel), and the liver (right panel; see Materials and Methods). The dashed line indicates the limit of detection; the bars indicate the median bacterial load; and “P” indicates the P value (Mann-Whitney U test; differences between the water and streptomycin-pretreated mice infected with serovar Typhimurium). NS, not statistically significant. Symbols: , water-pretreated mice mock infected with PBS; ○, water-pretreated mice infected with serovar Typhimurium; , streptomycin-pretreated mice mock infected with PBS; •, streptomycin-pretreated mice infected with serovar Typhimurium.

Open in a separate window

FIG. 4.

Serovar Typhimurium induces macroscopic changes in the cecum of streptomycin-pretreated mice at 2 days p.i. We removed the ceca of mice from the experiment shown in Fig. ​Fig.33 for macroscopic analysis. (A) Cecum morphology. The cecum of one mouse from each group was taken for photographic documentation. The photograph is representative of all six animals from each group. The centimeter scale indicates the magnification. (B) Total weight of the cecum. We determined the weight of the ceca of the remaining five mice from each group. P, P value (Mann-Whitney U test; differences between animals which were infected with serovar Typhimurium or mock infected with sterile PBS); NS, not statistically significant. “S. Tm + or −” and “sm + or −” indicate whether the mice were pretreated with streptomycin (sm) or water and whether the animals were infected with serovar Typhimurium (S. Tm). Bars indicate the median weight of the cecum. Symbols: , water-pretreated mice mock infected with PBS; ○, water-pretreated mice infected with serovar Typhimurium; , streptomycin-pretreated mice mock infected with PBS; •, streptomycin-pretreated mice infected with serovar Typhimurium.

As expected, serovar Typhimurium was only detectable in the intestine and organs of the mice that had been infected (Fig. ​(Fig.3).3). We detected no serovar Typhimurium in the contents of the medial and terminal ilea of most of the water-pretreated mice infected with serovar Typhimurium. Low bacterial loads (≤10⁵ CFU/g) were detected in the medial ileum of one, in the terminal ilea of two, and in the ceca of five of the six water-pretreated mice (Fig. ​(Fig.3A3A).

In streptomycin-pretreated mice we observed a slight but not significant (P = 0.24; P = 0.065) increase in the bacterial loads in the medial (P = 0.24) and terminal (P = 0.065) ilea and significantly increased (ca. 10⁵-fold) bacterial loads in the cecum contents (P = 0.002; Fig. ​Fig.33).

We observed significantly increased loads of serovar Typhimurium in the mesenteric lymph nodes of streptomycin-pretreated mice (P = 0.002; Fig. ​Fig.3B).3B). Slightly increased colonization was also detected in the spleens (P = 0.002) and livers (P = 0.009) of streptomycin-pretreated mice. However, even in the streptomycin-pretreated mice bacterial loads in the spleens and livers were only ca. 10-fold above the detection limit. Overall, these data are in line with earlier observations (57) and suggest that even in streptomycin-pretreated mice the systemic infection is still at a very early stage at 48 h p.i. Therefore, the systemic infection in these mice is unlikely to affect the pathological changes observed in the lower intestine.

During the dissection we did not observe macroscopic signs of inflammation of the ilea, ceca, and colons of the water-pretreated mice that were infected with serovar Typhimurium or mock infected with sterile PBS. In accordance with earlier results (69), streptomycin-pretreated mice that were mock infected with sterile PBS had enlarged ceca. In contrast to the three control groups, fecal pellet formation was impaired in the proximal colons of all six streptomycin-pretreated mice infected with serovar Typhimurium, and the tissues of the proximal colons appeared pale and swollen in five of the six animals (data not shown). Furthermore, the ceca of all six streptomycin-pretreated serovar Typhimurium-infected mice were shriveled to a small size, pale, and filled with purulent exudate (Fig. ​(Fig.4A,4A, right side). To document these macroscopic differences, we removed the cecum from one mouse of each group for photographic documentation (Fig. ​(Fig.4A)4A) and determined the weight of the cecum, including the contents of the remaining five animals from each group (Fig. ​(Fig.4B).4B). In water-pretreated mice, infection with serovar Typhimurium did not significantly affect the total weight of the cecum (Fig. ​(Fig.4B;4B; P = 0.458). However, in streptomycin-pretreated mice we observed a significant reduction of the cecal weights in mice infected with serovar Typhimurium (Fig. ​(Fig.4B,4B, right panel; P = 0.002).

For histopathologic analysis, intestinal tissue samples were cryoembedded, and 5-μm thin sections were stained with H&E (see Materials and Methods). The intestines of all three control groups appeared histologically normal (Fig. 5A to C and E to G and data not shown). However, in line with our initial observations (Fig. ​(Fig.2),2), pronounced inflammation was detected in the ceca and, to a lesser extent, also in the colons of all streptomycin-pretreated mice colonized with serovar Typhimurium (Fig. 5D and H; data not shown). This included pronounced edema in the submucosa and the lamina propria, crypt elongation, disruption of the crypt architecture, reduced numbers of goblet cells, epithelial erosion, and pronounced PMN infiltration of the submucosa, the lamina propria, and the epithelial layer, as well as transmigration of PMN into the intestinal lumen. No signs of severe inflammation were observed in the ilea of these mice (data not shown).

Open in a separate window

FIG. 5.

Cecal inflammation at 2 days p.i. The ceca of five of the six animals from each group of the experiment shown in Fig. ​Fig.33 were cryoembedded, and 5-μm thin sections were stained with H&E for histopathological analysis. (A and E) Cecum of a water-pretreated mouse mock infected with PBS; (B and F) cecum of a water-pretreated mouse infected with serovar Typhimurium; (C and G) cecum of a streptomycin-pretreated mouse mock infected with PBS; (D and H) streptomycin-pretreated mouse infected with serovar Typhimurium. The images are representative for all animals from each group. Boxes in panels A, B, C, and D indicate the area shown at the higher magnification. L, intestinal lumen; e, edema; p, PMN; er, erosion of the epithelial layer; g, goblet cell; sa, submucosa. Magnifications are indicated by the black bars.

To improve comparison of the inflammatory responses between the different groups of mice, we devised a histopathologic scoring scheme for H&E-stained cecal tissue sections (see Materials and Methods). This scheme considers the extent of the submucosal edema (0 to 3 [arbitrary units]), PMN infiltration (0 to 4), loss of goblet cells (0 to 3), and the epithelial integrity (0 to 3) and yields a total pathological score of 0 to 13 U (Materials and Methods). Using this scoring scheme, we have found no significant differences between the water-pretreated mice that had been infected with serovar Typhimurium or mock infected with sterile PBS (Fig. ​(Fig.6,6, left panel). The minimal signs of cecal inflammation observed in these mice and in the streptomycin-pretreated mice that had been mock infected with sterile PBS are frequently found in untreated SPF mice and are generally not considered as a sign of disease. In contrast, the streptomycin-pretreated mice infected with serovar Typhimurium displayed profound inflammation that was significantly stronger than in the animals of the control group (Fig. ​(Fig.6,6, right panel; P = 0.008). These data are in line with the macroscopic observations (Fig. ​(Fig.4)4) and demonstrate that streptomycin-pretreated mice offer a robust model for studying serovar Typhimurium colitis.

Open in a separate window

FIG. 6.

Scoring of inflammatory changes at 2 days p.i. In order to quantify pathological changes, we scored H&E-stained sections of the ceca from five of the six mice from each group of the experiment shown in Fig. ​Fig.33 as described in Materials and Methods. Edema in the submucosa (black), PMN infiltration (medium gray), reduction of the number of goblet cells (dark gray), and erosion or ulceration of the epithelial layer (light gray) were scored separately and plotted as stacked vertical bars. The combined score equals the sum of the separate scores. “S. Tm + or −” and “sm + or −” indicate whether the mice were pretreated with streptomycin (sm) or water and whether the animals were infected with serovar Typhimurium (S. Tm). Statistical analyses are shown for the separate scores and for the combined score. P, P value (Mann-Whitney U test; see Materials and Methods; difference between mice infected with serovar Typhimurium and mice mock infected with sterile PBS); NS, not statistically significant.

Lactobacillus spp. do not induce colitis in streptomycin-pretreated mice.

Streptomycin treatment is known to diminish the intestinal flora and to render mice susceptible to intestinal colonization by various microorganisms. It was conceivable that any dramatic change in the composition of the intestinal microflora might lead to inflammation and that the colitis induced by serovar Typhimurium might not be attributable to specific virulence factors. To test this hypothesis, we pretreated two groups of mice with 20 mg of streptomycin given p.o. 24 h prior to oral infection with 10⁸ CFU of nonpathogenic Lactobacillus spp. or serovar Typhimurium SL1344. The mice were sacrificed 2 days p.i., and we verified efficient colonization by both bacterial species by plating samples of the cecum contents on suitable agar plates (see Materials and Methods; 10⁴ CFU/g of Lactobacillus spp. or Salmonella spp.). Intestinal tissue samples were cryoembedded, and we analyzed 5-μm thin H&E-stained sections for signs of inflammation. In the Lactobacillus-infected mice we did not observe significant inflammation in the small or large intestine (Fig. 7A, C, and E; data not shown), whereas serovar Typhimurium-infected mice showed severe inflammation of the cecum and the proximal colon (Fig. 7B, D, and E). This indicates that colitis in streptomycin-pretreated mice is a specific response to colonization by serovar Typhimurium.

Open in a separate window

FIG. 7.

Infection of streptomycin-pretreated mice with Lactobacillus spp. does not lead to colitis. Groups of three streptomycin-pretreated mice (C57BL/6) were infected with 10⁸ CFU of Lactobacillus spp. or 10⁸ CFU of serovar Typhimurium. At 48 h p.i. the animals were sacrificed and 5-μm cryosections of the cecum were stained with H&E. (A and C) Cecum of streptomycin-pretreated mouse infected with Lactobacillus spp.; (B and D) cecum of streptomycin-pretreated mouse infected with serovar Typhimurium. The images are representative for all three animals from each group. The boxes in panels A and B indicate the areas shown at the higher magnification in panels C and D. L, intestinal lumen; e, edema; p, PMN; er, erosion of the epithelial layer; g, goblet cell; sa, submucosa. Magnifications are indicated by black bars. (E) Scoring of inflammatory changes as described in Materials and Methods. Edema in the submucosa (black), PMN infiltration (medium gray), reduction of the number of goblet cells (dark gray), and erosion or ulceration of the epithelial layer (light gray) were scored separately and are plotted as stacked vertical bars. The combined score equals the sum of the separate scores. S. Tm, mice infected with serovar Typhimurium; sm, mice were pretreated with streptomycin.

Time course of serovar Typhimurium colitis in streptomycin-pretreated mice.

To analyze the early stages of serovar Typhimurium colitis, we investigated the time course of the infection. Groups of five mice were pretreated with streptomycin, infected with 10⁸ CFU of serovar Typhimurium SL1344 (see Materials and Methods), and sacrificed at 2, 8, 20, and 48 h p.i. We analyzed pathological changes, as well as bacterial loads in the liver, mesenteric lymph nodes, and the intestinal content.

At 2 h p.i. the highest bacterial loads were present in the small intestine, and only low loads (<10⁴ CFU/g) were present in the cecal contents (Fig. 8A and B). In accordance with earlier results (66), bacterial loads in the cecum increased significantly between 2 h and 8 h p.i. (P = 0.008). Even higher bacterial loads were detected in the cecum at 20 h p.i. (P = 0.008; median = 4.4 × 10⁷ CFU/g) (Fig. ​(Fig.8B;8B; Table ​Table1).1). Between 20 h and 48 h p.i. we did not observe any further increase in the serovar Typhimurium loads in the cecum (P = 0.841; Table ​Table1),1), which indicated that colonization of the ceca of streptomycin-pretreated mice by serovar Typhimurium had reached a steady-state level within 20 h after oral infection. We did not observe a specific enrichment of serovar Typhimurium at the intestinal epithelium at any time between 8 h and 48 h p.i. and observed that only a small minority of the bacteria (5%) were located in crypts (Fig. 9J, K, and L).

Open in a separate window

FIG. 8.

Time course of serovar Typhimurium colitis in streptomycin-pretreated mice. Groups of five streptomycin-pretreated mice were infected for 2, 8, 20, or 48 h with 10⁸ CFU of serovar Typhimurium SL1344 p.o. We determined bacterial colonization and pathological changes as described in Materials and Methods. (A) Bacterial loads in different regions of the intestine at 2 h p.i. (left panel) and 8 h p.i. (right panel). (B) Time course of the bacterial loads in the cecum content. (C) Time course of bacterial loads in the mesenteric lymph nodes. (D) Time course of bacterial loads in the liver. (E) Time course of the total weight of the cecum (determined before the removal of samples for bacteriological or histopathologic analysis). (F) Time course of the histopathologic changes. Samples of the cecal tissue were cryoembedded and 5-μm thin sections were stained with H&E. Edema in the submucosa (black), PMN infiltration (medium gray), reduction of the number of goblet cells (dark gray), and erosion or ulceration of the epithelial layer (light gray) were scored separately (see Materials and Methods) and are plotted as stacked vertical bars. The combined score equals the sum of the separate scores. For the statistical analysis, see Table ​Table1.1. S. Tm, serovar Typhimurium; dashed line, limit of detection; bars, median bacterial load; gray arrows, values determined for the two mice without cecal inflammation at 8 h p.i. The black arrows indicate the pathological scores of the tissue sections shown in Fig. ​Fig.99.

At 2 h p.i. macroscopic inspection did not reveal any signs of inflammation in the small and the large intestine (median weight of the cecum = 0.59 g; Fig. ​Fig.8D;8D; data not shown). This was confirmed by histopathological analysis (Fig. ​(Fig.8F8F and 9A and E). At 8 h p.i. two of the mice appeared to be normal, and histopathological analysis did not reveal signs of inflammatory disease (Fig. ​(Fig.8F).8F). Interestingly, these two mice also carried the lowest loads of serovar Typhimurium in their cecal contents (Fig. ​(Fig.8C),8C), suggesting that a slower bacterial colonization might result in a delayed onset of the inflammatory response. The ceca of the three other animals sacrificed 8 h p.i. were pale, shriveled to a small size, and filled with purulent exudate (Fig. ​(Fig.8E).8E). Histopathological analysis confirmed that the ceca of these mice were inflamed, as judged by submucosal edema, PMN infiltration into the lamina propria, epithelial erosion or ulceration, and a reduced number of goblet cells (Fig. ​(Fig.8F8F and 9B and F). Substantial infiltration of PMN (CD18⁺, segmented nuclei) into the lamina propria was confirmed by immunofluorescence microscopy (Fig. ​(Fig.9N).9N). In addition, we observed increased expression of ICAM-1 (compare Fig. 9M and N), a key transmembrane receptor that binds CD18-β₂-integrin dimers and is involved in the extravasation of leukocytes from capillary blood vessels into tissues.

At 20 h p.i. the ceca of all five mice showed characteristic signs of inflammation, including a significantly reduced cecum size (Fig. ​(Fig.8E8E and Table ​Table1)1) and histopathologic changes such as submucosal edema, a reduced number of goblet cells, and epithelial erosion or ulceration (Fig. ​(Fig.8F8F and 9C and G). In contrast to the three mice that showed intestinal inflammation already at 8 h p.i., PMN were present in high numbers not only in the lamina propria but also in the submucosa and in the intestinal lumen (Fig. 9G and O), and we observed a concomitant increase in ICAM-1 expression (compare Fig. 9O and N).

In accordance with our first observations, the ceca of all five mice sacrificed at 48 h p.i. showed signs of profound inflammation (Fig. 8E and F and 9D, H, and P). Compared to the findings at 20 h p.i., we observed a further progression of the disruption of the villus architecture (compare Fig. 9C and G with 9D and H), increased ICAM-1 expression, and a significant exacerbation of the PMN infiltration (P = 0.032) and the loss of goblet cells (P = 0.008) (Fig. ​(Fig.8F;8F; Table ​Table1;1; see also Fig. 9L and P).

To assess a possible correlation between intestinal inflammation and the onset of systemic infection, we monitored the time course of colonization of mesenteric lymph nodes and liver. Serovar Typhimurium could not be detected in the mesenteric lymph nodes and liver of any mice up to 8 h p.i. (Fig. 8C and D). However, at 8 h p.i. three of the five mice already had an inflamed cecum (see above). This suggests that colonization of the mesenteric lymph nodes or spread to internal organs is not a prerequisite for intestinal inflammation. At 20 h p.i. we detected a significant increase in the median serovar Typhimurium load in the mesenteric lymph nodes (P = 0.032) and a further increase at 48 h p.i. (P = 0.008; Fig. ​Fig.8C;8C; Table ​Table1).1). In the livers we detected serovar Typhimurium in only one of five animals at 20 h p.i. (P = 0.690; Fig. ​Fig.8D;8D; Table ​Table1).1). However, bacterial loads in the liver were significantly increased at 48 h p.i. (Fig. ​(Fig.8D8D and Table ​Table1).1). In conclusion, colonization of the mesenteric lymph nodes and liver does not seem to be required for the initiation of cecal inflammation. However, we cannot rule out that colonization of the mesenteric lymph nodes or the early systemic infection might play a role in the exacerbation of the inflammatory response observed between 8 h and 48 h p.i.

Overall, the observed pathological changes correlated well with the kinetics of intestinal colonization of streptomycin-pretreated mice by serovar Typhimurium. In some animals we observed cecal inflammation as early as 8 h p.i. These animals carried bacterial loads of ca. 10⁷ CFU/g in the cecal contents. At 20 h p.i. the cecum was fully colonized (median = 4 × 10⁸ CFU/g; Fig. ​Fig.8B),8B), and the maximum levels of cecal inflammation were reached by between 20 and 48 h p.i. These observed symptoms are indicative of inflammation and a rapid regeneration of the cecal epithelium in response to colonization by serovar Typhimurium and resemble many aspects of human (nontyphoid) salmonellosis and bovine and rabbit models of Salmonella enterocolitis. Therefore, streptomycin-pretreated mice might offer a versatile alternative model for studying the pathogenesis of gastrointestinal serovar Typhimurium infections.

Role of SPI1 type III secretion in murine serovar Typhimurium colitis.

The SPI1 type III secretion system plays a key role in the early, gut-associated stages of the infection in susceptible animals (75, 81). In calves, the SPI1 type III secretion system plays a key role in the induction of inflammatory diarrhea. However, it has remained unclear whether the SPI1 type III secretion system can play a similar role in mice. To explore this question, we used SB161 (SL1344, ΔinvG), an isogenic mutant that lacks an essential subunit of the SPI1 type III apparatus and is incapable of secreting and translocating any proteins via this route (14, 43). Eight streptomycin-pretreated C57BL/6 mice were infected with wild-type serovar Typhimurium SL1344, and eight mice were infected with SB161 (10⁸ CFU p.o.). At 48 h p.i., the mice were sacrificed and analyzed. Bacterial loads in the cecum contents, the mesenteric lymph nodes, and in the livers did not differ significantly between mice infected with wild-type serovar Typhimurium and those infected with SB161 (Fig. 10A and B). However, significantly lower loads of SB161 than of wild-type serovar Typhimurium were detected in the spleens (P = 0.015; Fig. 10B, middle panel). These observations are in line with earlier findings (27, 57) and indicate that the SPI1 type III secretion system plays a role in establishing systemic infection in streptomycin-pretreated mice. However, in spite of similar interaction patterns with the intestinal wall during the early stages of colonization (Fig. 11G and H) and similar Salmonella loads in the cecal contents (Fig. 10A), only mild symptoms of colitis were detected in the ceca (and proximal colons) of all eight mice colonized with SB161. This included significantly higher cecal weights (P < 0.001; Fig. 10C) and reduced histopathological signs of inflammation (combined score = P < 0.001; Fig. 10D, compare Fig. 11A and C with 11B and D; Table ​Table2).2). Infiltration and transmigration of CD18⁺ cells/PMN and expression of ICAM-1 were also alleviated in the ceca of mice infected with the serovar Typhimurium mutant SB161 (compare Fig. 11E and F). This demonstrates that serovar Typhimurium requires a functional SPI1 type III secretion apparatus in order to elicit profound intestinal inflammation.

Open in a separate window

FIG. 10.

Role of the SPI1 type III secretion system in serovar Typhimurium colitis. Eight streptomycin-pretreated C57BL/6 mice were infected p.o. with 10⁸ CFU of wild-type serovar Typhimurium SL1344 (black circles in panels A, B, and C; left side of panel D) or with SB161 (SL1344, ΔinvG; open circles in panels A, B, and C; right side of panel D) and analyzed at 2 days p.i. (A) Loads of serovar Typhimurium present in the cecum content. (B) Colonization of mesenteric lymph nodes (left panel), spleen (middle panel), and liver (right panel). (C) Total weight of the cecum, including the cecal contents (determined before the removal of samples for bacteriological or histopathologic analysis). (D) Histopathological analysis. Samples of the cecal tissue were cryoembedded, and 5-μm thin sections were stained with H&E. Edema in the submucosa (black), PMN infiltration (medium gray), reduction of the number of goblet cells (dark gray), and erosion or ulceration of the epithelial layer (light gray) were scored separately (see Materials and Methods) and plotted as stacked vertical bars. The combined score equals the sum of the separate scores. Black and white arrows indicate the pathological scores of the tissue sections from the wild-type- and SB161-infected mice, respectively, shown in Fig. ​Fig.11.11. For the statistical analysis, see Table ​Table2.2. P, P value (Mann-Whitney U test; see Materials and Methods). NS, not statistically significant; S. Tm, serovar Typhimurium; dashed line, limit of detection; bars, median bacterial load or weight.

Tissue culture experiments have suggested that mere secretion (and not direct injection into the host cell) of some effector proteins (i.e., SipA) might be sufficient to elicit proinflammatory responses in polarized epithelial monolayers (47). In order to assess whether protein secretion or direct injection (translocation) of bacterial effector proteins into host cells is required for the induction of colitis, we analyzed two serovar Typhimurium mutants that are incapable of injecting effector proteins into host cells but still retain the capacity to secrete some or all proteins across the bacterial cell envelope. In contrast to SB161, strain SB302 (SL1344, invJ::aphT) can secrete a few proteins (InvJ and SpaO [14]) and strain SB241 (SL1344, sipD::aphT) is even more efficient than wild-type serovar Typhimurium at secreting all known effector proteins into bacterial culture supernatants (44).

Groups of five C57BL/6 mice were infected with the isogenic mutants SB161, SB302 (SL1344, invJ::aphT), or SB241 (SL1344, sipD::aphT) or with wild-type serovar Typhimurium (10⁸ CFU p.o.). At 48 h p.i. the mice were sacrificed for analysis of bacterial loads in internal organs, intestinal colonization, and pathological symptoms. The capacity of SB302 and SB241 to colonize the cecum, the lymph nodes, and the liver and to cause inflammatory changes in the cecal tissue did not differ significantly from that of the secretion- and translocation-deficient serovar Typhimurium mutant SB161 (Fig. 12A to E and Table ​Table3).3). Altogether, these data are in accordance with observations from bovine models (75, 81, 86) and demonstrate that the SPI1 type III secretion system is dispensable for establishing colonization of the murine intestine and spread to the mesenteric lymph nodes. However, efficient dissemination to systemic sites and the elicitation of colitis are dependent on a functional SPI1 type III secretion apparatus. Furthermore, the results obtained with SB241 indicate that in vivo mere secretion is insufficient but that the SPI1 effector proteins inducing the inflammatory response must be translocated into host cells in order to exert their biological function.

Open in a separate window

FIG. 12.

Effect of different mutations compromising effector protein translocation on murine serovar Typhimurium colitis. Groups of five streptomycin-pretreated C57BL/6 mice were infected p.o. with 10⁸ CFU of wild-type serovar Typhimurium SL1344 (black circles), SB161 (SL1344, ΔinvG; white circles), SB302 (SL1344, invJ::aphT; gray circles), or SB241 (SL1344, sipD::aphT; striped circles), sacrificed 48 h p.i. and analyzed as described in Materials and Methods. (A) Bacterial load in the cecal content; (B) bacterial load in the mesenteric lymph nodes (mLN); (C) bacterial load in the liver; (D) total weight of the cecum, including the cecal contents (determined before the removal of samples for bacteriological or histopathologic analysis); (E) histopathological analysis. Samples of the cecal tissue were cryoembedded, and 5-μm thin sections were stained with H&E. Edema in the submucosa (black), PMN infiltration (medium gray), reduction of the number of goblet cells (dark gray), and erosion or ulceration of the epithelial layer (light gray) were scored separately (see Materials and Methods) and plotted as stacked vertical bars. The combined score equals the sum of the separate scores. S. Tm, serovar Typhimurium. For the statistical analysis, see Table ​Table3.3. Dashed line, limit of detection; bars, median bacterial load.

M.B. and S.H. contributed equally to this study.

We are grateful to Andrew Macpherson for scientific discussion, help with the histopathological evaluation, and comments on the manuscript; Burkhardt Seifert for advice on the statistical analysis; C. Künzel and S. Brinkmann for expert assistance with the animal experiments; and J. E. Galan and J. Heesemann for continued support. We thank R. Zinkernagel and H. Hengartner for generous support of our animal work at the BZL Zurich.

The work described here was funded in part by grants from the DFG and the Volkswagenstiftung (to W.-D.H.).