Developmentally Regulated Intestinal Expression of IFN- and Its Target Genes and the Age-Specific Response to Enteric Salmonella Infection1
http://www.100md.com
免疫学杂志 2005年第14期
Abstract
Young infants are highly susceptible to systemic dissemination of enteric pathogens such as Salmonella typhimurium when compared with older individuals. The mechanisms underlying this differential susceptibility have not been defined clearly. To better understand this phenomenon, we examined the responses of adult mice and preweaned pups to oral infection by S. typhimurium. We found clear age-specific differences, namely, an attenuated intestinal inflammatory response and a higher systemic bacterial burden in the pups compared with the adults. To elucidate the molecular basis for these differences, we obtained a microarray-based profile of gene expression in the small intestines of uninfected adult and preweaned animals. The results indicated a striking age-dependent increase in the intestinal expression of a number of IFN--regulated genes involved in antimicrobial defense. This finding was confirmed by real-time quantitative PCR, which also demonstrated an age-dependent increase in intestinal expression of IFN-. The developmental up-regulation of the IFN--regulated genes was dependent on both IFN- and a normal commensal microflora, as indicated by experiments in IFN--knockout mice and germfree mice, respectively. However, the increase in expression of IFN- itself was independent of the commensal flora. The functional importance of IFN- in the immunological maturation of the intestine was confirmed by the observation that the response of adult IFN--knockout animals to S. typhimurium infection resembled that of the wild-type pups. Our findings thus reveal a novel role for IFN- in the developmental regulation of antimicrobial responses in the intestine.
Introduction
One of the most common infectious agents isolated from cases of acute gastroenteritis is Salmonella, with serovar Typhimurium being the most frequent isolate. The incidence of infection is highest in infants < 1 year of age with a rate of 134 cases/100,000 children (4).
S. typhimurium that is introduced into the gastrointestinal (GI)3 tract via contaminated food or water adheres to and invades epithelial cells of the intestine, particularly the M cells overlying Peyer’s patches. The process of invasion is an active one in which bacterial effector proteins introduced into host cells by a type III secretion system induce cytoskeletal rearrangements that result in engulfment of the Salmonella by large membrane protrusions (5). The effector proteins also activate a number of cellular signaling pathways, leading to the acute inflammatory response that is responsible for the manifestations of gastroenteritis (6). In most individuals, this response is sufficient to clear the infection so that systemic dissemination of the organism is unusual. However, young infants, particularly those <3 mo of age, have a high risk for complications resulting from extraintestinal spread of Salmonella, including sepsis, osteomyelitis, meningitis, and death (7, 8, 9).
It is not clear why young infants are so prone to spread of S. typhimurium from the gut. It may be part of a more generalized impairment of intestinal defense against enteric pathogens because other systemic infections in this age group often originate in the GI tract (10). Inadequate development of epithelial barrier functions and mucosal immunity may underlie this phenomenon (11), but the exact molecular basis for this immaturity, as well as the factors involved in the normal maturation of protective responses, are yet to be elucidated.
To better understand the age-specific differences in the intestinal response to bacteria, we have made use of the mouse model of Salmonella infection. Although infection of mice with S. typhimurium typically results in a systemic infection without overt intestinal inflammation, recent studies by Barthel et al. (12) have shown that following pretreatment of the animals with oral streptomycin, infection with a virulent, streptomycin-resistant strain of Salmonella results in a robust and reproducible inflammation of the cecum. The pathologic changes demonstrated in this murine model closely resembled the intestinal inflammation caused by S. typhimurium infection in humans. We have used this approach to compare the responses of young mouse pups (preweaned, 16–18 days old) and their adult (5–6 wk old) counterparts to oral infection with Salmonella.
Materials and Methods
Mice
C57BL/6 (B6) and B6.129S7-Ifngrtm1Agt/J mice ranging in age from 5 days to 10 wk old were obtained from The Jackson Laboratory and Charles River Laboratories. BALB/c mice ranging in age from 5 days to 6 wk old were obtained from Charles River Laboratories. Tissue obtained from 6-wk-old BALB/c germfree mice were kindly provided by Dr. N. Nanthakumar (Massachusetts General Hospital, Charlestown, MA). These mice were obtained from the Gnotobiotic Animal Core at the University of North Carolina. Animals were sacrificed within the germfree facility, intestinal tissue was removed under sterile conditions, snap frozen in liquid nitrogen, and stored at –80°C until use. All animal studies were approved by the Massachusetts General Hospital Subcommittee on Research Animal Care and conformed to the U.S. Department of Agriculture Animal Welfare Act, Public Health Service Policy on Humane Care and Use of Laboratory Animals, and other applicable laws and regulations.
Quantitation of commensal bacteria
Sixteen- to 18-day-old and 6-wk-old B6 mice (three of each age group) were treated with streptomycin (12 and 20 mg, respectively) via the orogastric route for 24 h and then sacrificed. The terminal ileum and colon from these animals, as well as from age-matched, untreated controls, were removed, rinsed free of contents, homogenized in 1 ml of sterile PBS, and serial dilutions of the homogenates plated on Luria-Bertani agar for culture under aerobic and anaerobic conditions. Culturable commensal flora were quantified as colony-forming units per milligram of tissue. The limit of detection equaled 250 CFU/mg tissue.
Salmonella typhimurium infection
Five- to 6-wk-old B6 mice were housed in groups of up to four animals under standard barrier conditions in a specific pathogen-free facility. Sixteen- to 18-day-old B6 litters of up to eight were housed with a lactating female. Water and food were supplied ad libitum. Adult mice were pretreated with 20 mg of streptomycin, and pups were pretreated with 12 mg of streptomycin via the orogastric route. At 24 h following streptomycin treatment, 5- to 6-wk-old and 16- to 18-day-old animals were infected with 1 x 108 and 3 x 107 CFU of the streptomycin-resistant S. typhimurium strain SL1344, respectively. At 48 h postinfection, mice were sacrificed by CO2 asphyxiation. Tissue samples from the small intestine, cecum, colon, and spleen were removed for additional analysis. Small intestine and cecum were rinsed of fecal contents with sterile PBS. Intestinal tissue and spleen were homogenized using 1 ml of sterile 1% Triton X-100. The numbers of colony-forming units were determined by plating serial dilutions on Luria-Bertani streptomycin and incubating at 37°C for 24 h.
Histologic analysis
Segments of ileum, cecum, and colon were rinsed free of contents and embedded in Tissue Tek OCT (Sakura Finetek), snap frozen in liquid nitrogen, and stored at –80°C. Tissue samples were then transferred and fixed in paraffin. Sections (5 or 10 μm) were mounted on glass slides, air dried, and stained with H&E. Immunoperoxidase staining was performed using a goat polyclonal IgG anti-mouse IDO Ab (1/50 dilution) (Santa Cruz Biotechnology).
Myeloperoxidase (MPO) assay
Whole cecum was resected, rinsed free of contents in sterile PBS, and homogenized in 1 ml of 0.05% hexadecyltrimethylammonium bromide (HTAB) buffer. Samples were snap frozen in liquid nitrogen for 1 min and then immediately thawed at 37°C for 10 min for a total of three such cycles. Samples were then sonicated on ice for 20 s and centrifuged at 16,000 x g for 10 min. MPO activity in the clarified sonicate was then estimated by a colorimetric assay as described in detail earlier (13)
Gentamicin protection assay
Overnight cultures of SL1344 were collected by centrifugation at 6000 x g for 10 min, washed with sterile PBS, and resuspended in antibiotic-free tissue culture medium. Macrophages were harvested from the peritoneum of adult B6 mice on the fourth day after injection of sterile thioglycollate broth as described previously (14). MODE-K cells (kindly provided by Dr. R. Blumberg (Brigham and Women’s Hospital, Boston, MA) were cultured in DMEM with 10% FCS. Cells were left untreated (control) or were pretreated overnight with 10 ng/ml recombinant murine IFN- (R&D Systems). The cells were infected with 107 bacteria (multiplicity of infection of 10:1). After 1 h of infection, cells were washed with sterile PBS and placed in DMEM with 100 μg/ml gentamicin for 2 h to kill extracellular bacteria. The cells were then lysed either immediately or after an additional 18 h in 10 μg/ml gentamicin, and the number of intracellular bacteria was determined at each time point by culturing serial dilutions of cell lysates, as described in detail earlier (15).
Isolation of intestinal epithelial cells
Epithelial cells were isolated from B6 mice ranging in age from 14 days to 8 wk old as per the methods described by Evans et al. (16). In brief, the small intestine (from the terminal ileum to the ligament of Treitz) was resected, dissected open, cut into 2- to 3-mm pieces, and rinsed clean in 50 ml of Ca2+- and Mg2+-free HBSS at room temperature. Tissue was then diced into <1-mm3 pieces, incubated with 20 ml of enzyme solution (0.1 mg/ml dispase type I and 300 U/ml collagenase type XI (Sigma-Aldrich)) for 30 min with agitation at 80 cycles/min. The cell suspension was pipetted vigorously for 3 min and left to sediment under gravity for 1 min. The supernatant was collected and added to 10 ml of DMEM + 2.5% FCS with 2% sorbitol and centrifuged at 1000 x g for 2 min. This process was repeated until the supernatant was clear, and the pellet was well defined. The cell pellet was resuspended in DMEM + 2.5% FCS. Viability was assessed by staining cells with trypan blue solution (0.4%) and quantified with a hemacytometer. Purity of the preparation was analyzed by FACS analysis after staining with the TROMA-1 epithelial cell-specific marker (Developmental Studies Hybridoma Bank, University of Iowa) (17). We routinely obtained preparations containing 90% epithelial cells.
Microarray-based gene expression profiling
Experiment design. To investigate the molecular basis of the age-dependent differences in response to Salmonella described above, a microarray-based gene expression profiling analysis was performed on RNA prepared from intestinal epithelial cells of uninfected preweaned pups and postweaned adult mice to detect changes in the expression of genes involved in the immune response. A direct comparison of gene expression at different stages of development was made. Hybridizations were performed in triplicate.
Sample RNA preparation. Total RNA was extracted and purified from intestinal epithelial cells (isolated as described in Materials and Methods) from 16- and 33-day-old B6 mice (TRIzol reagent, Invitrogen Life Technologies; RNeasy column, Qiagen). Samples were pooled from five separate animals to yield a total quantity of 30 μg of RNA for each group. RNA quantity and quality were assessed by using a Beckman Coulter DU 640 Spectrophotometer (Beckman Coulter) and Agilent 2100 Bioanalyzer (Agilent), according to the manufacturer’s protocols.
Extract preparation and labeling. Briefly, two samples (a reference and a test) of isolated total RNA were reverse transcribed into amine modified cDNA. This cDNA was labeled via a coupling reaction to the Cy3 or Cy5 dyes, then purified and concentrated for hybridization. The protocol used was a modified version of the Atlas Powerscript Fluorescent Labeling kit (Clontech 634712) protocol, and full details are available online (https://dnacore.mgh.harvard.edu/microarray/protocol-labeling.shtml).
Hybridization procedures and parameters. For direct comparison, Cy3- and Cy5-labeled cDNA samples were resuspended in hybridization buffer and then hybridized to a PGA Mouse v1.0 array (MGH_Mm_01) slide made available through the Microarray Core Facility at the Massachusetts General Hospital. Both preweaned and postweaned samples were hybridized to the same slide, and each set of hybridizations was performed in triplicate. Full details regarding the hybridization protocol are available online (https://dnacore.mgh.harvard.edu/microarray/protocol-hybridization.shtml).
Measurement data and specifications. BioArray Software Environment (18) and GenMAPP (19, 20) software programs were used for data analysis. Statistical significance (p value) was calculated by Student’s t test based on the results of three separate hybridizations.
Array design. A PGA Mouse v1.0 array (MGH_Mm_01) was made available through the Microarray Core Facility at the Massachusetts General Hospital. A total of 14,609 oligonucleotides was arrayed onto 384-well Genetrix polystyrene V-bottom plates (X6004).
Full details regarding the microarray printing protocol are available online (https://dnacore.mgh.harvard.edu/microarray/protocol-printing.shtml). Specific details characterizing the individual spots on the array are available at the following web site: https://dnacore.mgh.harvard.edu/microarray/files/anno_dist-MGH_Mm_01.txt.
Results
Age-specific differences in the response to oral infection with Salmonella
Based on the protocol described by Barthel et al. (12), 16- to 18-day-old (preweaned) and 5- to 6-wk-old (adult) B6 mice were pretreated with a single dose of streptomycin per os and then inoculated orally 24 h later with the streptomycin-resistant strain of S. typhimurium, SL1344. The doses of both the streptomycin and the bacteria were adjusted to the weight of the animals. The basal number of gut commensal bacteria that could be cultured from the adult animals was 270-fold higher than in the pups. The streptomycin treatment significantly reduced the number of culturable commensals in both age groups (from 6350 CFU/mg intestinal tissue to <250 CFU/mg tissue—the limit of detection—in the case of the pups; from 1.73 x 106 to 1.13 x 104 CFU/mg in the case of the adults). The animals were sacrificed 48 h after the infection, and the inflammatory response in the cecum was assessed by histopathology, increase in MPO activity (as a measure of neutrophil recruitment), and up-regulation of the proinflammatory cytokine TNF-. At the same time, homogenates of the ileum and spleen were cultured to determine the number of bacteria in these tissues.
The pretreatment with streptomycin alone did not appear to elicit an inflammatory response in the intestine, either by gross appearance or by histopathology, nor were there any appreciable differences between the adult mice and the pups (Fig. 1A). However, following 48 h of infection with SL1344, there was a marked thickening of the cecal wall in the infected 5- to 6-wk-old mice compared with the streptomycin-pretreated, uninfected age-matched controls upon gross inspection. Microscopically, an inflammatory response was evidenced by marked submucosal edema and cellular infiltrate of the lamina propria (Fig. 1, B and C). In contrast, infection did not appreciably alter the gross appearance of the cecum in the 16- to 18-day-old mice. On histologic analysis in the younger animals, both the degree of submucosal edema and inflammatory infiltration were less than in the adults (Fig. 1, B and C).
To obtain a more quantitative assessment of the intestinal inflammation, we measured tissue MPO activity as an indicator of neutrophil infiltration, one of the characteristic features of Salmonella-induced enteritis (13, 21, 22). The results of this analysis demonstrated that there was a 37.6-fold (±11.3, p = 0.007) increase in activity in the infected adult mice compared with uninfected controls (Fig. 1D). In comparison, there was only a 1.06-fold (±0.44, p > 0.05) increase in MPO activity in the infected 18-day-old mice, indicating a lower level of neutrophil recruitment than in the adults, consistent with the histopathological findings.
Given the role of TNF- in the recruitment of neutrophils to sites of Salmonella infection (23), and in the defense against this organism (24), we examined cecal expression of this cytokine by quantitative real-time PCR. Following infection, there was a 14.6-fold (±0.28, p = 0.02) increase in TNF- transcript levels in the 5- to 6-wk-old mice. In contrast, infection-induced up-regulation of TNF- was significantly less (p = 0.01) in the 16- to 18-day-old pups: there was only a 7.06-fold (±1.57, p = 0.02) increase in the level of the transcript (Fig. 1E).
Taken together, the results of the histopathologic evaluation, MPO assay, and TNF- real-time PCR analysis all indicated that there was a decreased intestinal inflammatory response in the preweaned pups compared with the adult mice following infection with S. typhimurium. This result was not altered even if the adult doses of streptomycin, and Salmonella were administered to the pups (data not shown).
As mentioned above, S. typhimurium infection in the mouse is typically associated with invasion of the epithelium via the Peyer’s patches in the small intestine, particularly the ileum, and dissemination of the organism to extraintestinal sites such as the spleen. Therefore, to determine whether there were age-dependent differences in local and systemic infection, cultures of the ileum and spleen were performed 48 h after oral infection of preweaned pups or adult mice with SL1344. The results demonstrated that there was a significantly greater number of bacteria in both the ileum (3161 ± 2065 vs 434 ± 318, p = 0.04) and the spleen (9410 ± 5656 vs 872 ± 160, p = 0.0007) of the pups compared with the adult mice, respectively (Fig. 1, F and G). Thus, the less pronounced intestinal inflammatory response in the pups was associated with a greater local and systemic burden of bacteria.
Age-dependent changes in intestinal gene expression
To investigate the molecular basis of the age-dependent differences in response to Salmonella described above, a microarray-based gene expression profiling analysis was performed on RNA prepared from small intestinal epithelial cells of preweaned pups and adult mice to detect changes in the expression of genes involved in the immune response. The RNA samples were reverse transcribed, then hybridized to a PGA Mouse v1.0 array (MGH_Mm_01) slide (made available through the Microarray Core Facility at the Massachusetts General Hospital) containing 14,609 oligonucleotides. Following quality control analysis, 6791 usable signals were identified. Data from the microarray were further analyzed using a web-based software program BioArray Software Environment (https://base.mgh.harvard.edu) (18). A direct comparison of expression was made between the preweaned and the adult stages of development for each gene and was quantified as a fold change. A fold change > 1.0 indicated an increase in expression of that particular gene in the adult, whereas a value < 1.0 represented a decrease in expression.
Results from the microarray demonstrated an age-dependent increase in a number of antimicrobial, IFN--regulated genes. In contrast, there were no statistically significant differences in expression of housekeeping genes such as GAPDH, pyruvate carboxylase, 2-microglobulin, and cyclophilin A. There were a total of 460 genes in which there was at least a 2-fold, statistically significant increase in expression with aging. Of these, there were 33 genes that were classified by the BioArray Software Environment and GenMAPP programs (18, 19, 20) under the biologic process of "immune response." More strikingly, of these 33 immune response genes, 10 were IFN- regulated. In addition, there were at least another five IFN--regulated genes identified by inspection whose expression increased significantly in an age-dependent manner (Table I). The IFN--regulated genes included several chemokines involved in innate and adaptive immunity (25), members of a GTPase family implicated in resistance to intracellular infection (26), and other proteins such as the polyimmunoglobulin receptor, which is required for the transport of secretory Ig across mucosal surfaces (27), and IDO, which has been shown to be involved in both innate defense against several types of pathogens and in the regulation of lymphocyte activation (28, 29). Intriguingly, one of the genes in this list was IFN-induced protein with tetratricopeptide repeats 1, the orthologue of which was recently identified as being up-regulated in the intestine of gnotobiotic zebrafish following colonization with bacteria (30), suggesting that the increased expression of IFN--induced genes may be an evolutionarily conserved feature of gut maturation. Of note, the results of the microarray analysis did not reveal any differences in the expression of IFN- or the IFN-R
One important change that occurs during the weaning period is the colonization of the intestine with commensal flora, and it is well-known that commensal bacteria can influence gene expression in the intestinal epithelium, both in mammals and in lower vertebrates (30, 31). To determine whether the commensal flora was involved in the age-dependent increase in expression of the IFN--regulated genes in the intestine, quantitative real-time PCR for CCL5, CXCL9, and IDO was performed on RNA prepared from the ileum of adult germfree and conventionally raised mice. Because the only germfree mice available to us were of the BALB/c strain, we first demonstrated that the expression of these genes did in fact increase in an age-dependent manner in BALB/c mice as in the B6 animals (Fig. 3B). When expression of CCL5, CXCL9, and IDO was analyzed in adult germfree BALB/c mice, it was found to be significantly less than in the age-matched conventionally reared counterparts and was similar to that of the conventionally raised preweaned BALB/c pups (Fig. 3B). Thus, there was a significant attenuation of the normal developmental increase in IFN--regulated gene expression in the germfree mice, suggesting that the commensal flora is involved in regulating expression of these genes.
We also examined the effect of the commensal flora on IFN- expression in the intestine. In contrast to the results for CCL5, CXCL9, and IDO, the level of IFN- expression in the intestine of adult germfree mice was not significantly different from the conventionally reared controls (Fig. 3C), suggesting that the commensal flora did not play a major role in the age-dependent increase in IFN- in the gut.
IFN- enhances antimicrobial mechanisms of intestinal epithelial cells and macrophages in vitro
The results presented so far indicate an important role for IFN- in the developmental changes in gene expression in the intestine. Some of these IFN--regulated genes have been implicated in antimicrobial defense at the cellular level (26, 28), suggesting that one of the effects of IFN- may be to increase epithelial resistance to bacterial infection. To evaluate this possibility, we examined the effect of IFN- on the murine small intestinal epithelial cell line MODE-K, which has been shown previously to be responsive to this cytokine (32). MODE-K cells, pretreated overnight with 10 ng/ml rIFN- or left untreated, were infected with S. typhimurium strain SL1344, and the number of bacteria present inside control or IFN--treated cells was determined at 2 and 20 h after infection by gentamicin protection assay (15). Thioglycollate-elicited peritoneal macrophages from adult B6 mice were evaluated similarly in parallel. As shown in Fig. 4A, the IFN- treatment decreased the number of bacteria surviving in the MODE-K cells by 50% at the 20-h time point. This effect was seen as early as 6 h postinfection (data not shown). There was no difference in the number of intracellular bacteria at 2 h postinfection, suggesting that the IFN- is unlikely to be affecting the number of bacteria that initially entered the cells. IFN- treatment had a similar, even more dramatic, effect on the survival of Salmonella inside the thioglycollate-elicited peritoneal macrophages (Fig. 4B). Thus, IFN- acts on two cell types present in the intestinal mucosa, enterocytes, and macrophages to enhance microbicidal mechanisms that limit the intracellular growth of Salmonella.
IFN- influences the intestinal response to Salmonella and systemic growth of bacteria in vivo
We addressed the in vivo role of IFN- in the age-specific differences in response to enteric Salmonella infection by infecting adult IFN--knockout mice with S. typhimurium and comparing their response to those of the wild-type adult animal. As before, the inflammatory response in the cecum was assessed by histology, MPO assay, and TNF- quantitative real-time PCR. Histologic analysis revealed less submucosal edema and decreased lamina propria inflammatory infiltrate in the adult IFN--knockout animals compared with the age-matched, wild-type controls (Fig. 5, A and B). Similarly, both the MPO assay and the TNF- real-time PCR demonstrated a significantly attenuated response in the adult IFN--knockout mice compared with wild-type animals (Fig. 5, C and D). Compared with uninfected controls, there was only a 13.3-fold (±3.2, p = 0.02) increase in MPO activity in the IFN--knockout mice as opposed to a 37.6-fold (±11.27, p = 0.007) increase in the wild-type animals. With regard to TNF- gene expression, there was only a 5.85-fold (±0.51, p = 0.01) increase in transcript levels following infection in the IFN--knockout mice compared with a 13.8-fold (±0.51, p < 0.0001) increase in the wild-type animals. To determine whether there were IFN--dependent differences in local and systemic bacterial numbers, cultures of the ileum and spleen were performed. The results demonstrated that there was a significantly greater number of bacteria in the spleen of IFN--knockout mice compared with the wild-type controls (Fig. 5E). There was also a trend toward greater numbers of bacteria in the ileum of the IFN--knockout mice, although this did not reach statistical significance (data not shown). Overall, the responses of the adult IFN--deficient mice resembled those of the wild-type pups, an observation that is consistent with the idea that the cytokine is required for the normal maturation of the response to enteric Salmonella infection.
Discussion
The experiments reported in this article reveal a novel role for IFN- in the immunological maturation of the gut. The mechanistic explanation for this role of IFN- is suggested by the identity of the IFN--regulated genes that are up-regulated in the intestine in an age-dependent manner. IFN--induced GTPases such as LRG-47 have been shown to inhibit intracellular multiplication of Salmonella and other pathogens (26). Similarly, IDO has been shown to inhibit growth of several types of bacteria and certain parasites (28). The net effect of increased expression of these molecules would be to make cells of the intestinal mucosa more resistant to bacterial multiplication, an outcome that is consistent with our in vitro studies with IFN--treated macrophages and intestinal epithelial cells (Fig. 4), as well as in the in vivo experiments with wild-type and IFN--deficient mice (Figs. 1 and 5). The age-dependent increase in expression of other IFN--regulated genes, such as the polyimmunoglobulin receptor (Table I) and the chemokines CCL5, CXCL9, and CXCL10 (Table I and Fig. 2), is also likely to contribute to making the adult intestine more resistant to microbial entry and multiplication by promoting, respectively, the transepithelial transport of protective secretory Abs (27) and the recruitment of NK cells and macrophages to augment innate protection against infection (33, 34, 35). Finally, the age- and IFN--dependent enhancement of intestinal TNF- expression in response to infection (Figs. 1 and 5) results in greater recruitment of inflammatory cells such as neutrophils that are able to deal with bacterial invaders. Specifically relevant to our observations is the fact that IFN- has been shown to have a protective role in Salmonella infection and to enhance the expression of TNF- in response to LPS (36, 37).
Our quantitative real-time PCR analyses indicate that both intestinal epithelial cells and lamina propria cells such as macrophages are likely sites of expression of the IFN--regulated genes because similar results were obtained regardless of whether RNA from purified epithelial cells or whole intestinal tissue was used. Indeed, immunohistochemical localization of IDO in the intestine indicates expression in both the epithelium and in subepithelial cells (Fig. 2). Chemokines such as CCL5, CXCL9, and CXCL10 have been shown to be expressed by intestinal epithelial cells, especially in response to infection (34, 38, 39, 40). The cellular source of IFN- in the intestine is at present unclear. The results of quantitative real-time PCR indicate that it is likely to be in the nonepithelial compartment because an age-dependent increase in expression of the cytokine was observed only when RNA from whole intestinal tissue, rather than purified epithelial cells, was used. Candidate IFN--producing cells include macrophages, NK cells, NKT cells, and T lymphocytes, and experiments to determine which of these cell types is the site of the developmentally regulated expression of IFN- in the intestine are ongoing.
The results of our studies indicate an important role for the commensal flora in the induction of IFN--regulated genes in the intestine (Fig. 3), which is consistent with the age-dependent increase in number and complexity of these organisms and with their ability to induce alterations in gene expression and function in cells of the gut (26, 27, 38, 39, 40, 41). The TLR family of cell surface proteins, which recognize microbial components such as LPS, flagellin, and bacterial lipopeptide, constitute one set of candidate molecules that might be involved in sensing and responding to the commensal flora (41). Interestingly, the absence of a normal commensal flora did not affect the age-dependent increase in intestinal IFN- expression (Fig. 3). It is possible that other factors associated with the weaning period, such as the introduction of solid food or the surge in circulating corticosteroids, might be involved in regulating expression of IFN- in the intestine (11). Regardless, our results suggest that signals from the commensal flora and IFN- act independently to influence the expression of CCL5, CXCL9, and IDO. It is possible that the two sets of signals cooperate at the level of transcription to up-regulate these genes, as has been reported in other systems (42).
It is important to note that factors other than the change in IFN--regulated gene expression could contribute to the age-specific differences in immune response seen in our experiments. The GALT, including Peyer’s patches, the mesenteric lymph nodes, and the lymphocyte population of the lamina propria, are all underdeveloped in preweaned mice (43). This immaturity of organized lymphoid structures could place constraints on the kind of local immune response that develops. In addition, although we were able to detect Salmonella in the mesenteric lymph nodes even in the preweaned mice (data not shown), the lack of fully mature Peyer’s patches in these animals may favor the translocation of the organism by alternate routes, similar to what has been described for the uptake and extraintestinal dissemination of noninvasive bacteria by a population of CD18-positive phagocytes (44). The greater number and complexity of the commensal flora of the adult animals could also modify the response to Salmonella by amplifying local inflammation or by limiting colonization, and therefore spread, of the pathogen. The fact that the preweaned animals were still on breast milk is another factor that has to be considered. The presence of antimicrobial and immunomodulating factors in breast milk, such as secretory IgA, lactoferrin, IL-10, TGF-, and IL-1R antagonist may act to alter the inflammatory response (45). Finally, changes in intestinal expression of genes other than those that we have analyzed might play a role in the age-specific differences in immune response. For instance, our microarray data indicated that the expression of TLR3 was higher in the adult mice than in the pups (10.5-fold difference, p = 0.02). We have not pursued this observation further in the present series of experiments, but it is possible that alterations in TLR expression could contribute to differences in the inflammatory and antimicrobial responses between the adult and preweaned animals.
Although our studies have focused on the response to Salmonella, the results are likely to be relevant to other infections. Neonatal and weanling mice are more susceptible than adult animals to several intestinal pathogens, including Vibrio cholerae, enterotoxigenic Escherichia coli, Campylobacter jejuni, and Cryptosporidium parvum (46, 47, 48, 49). In the case of Cryptosporidium infection, lack of IFN- has been shown to confer susceptibility to adult mice (50). It would be worth examining whether developmental changes in the expression of IFN--regulated genes in the intestine contribute to the age-specific differences in susceptibility to these infections.
The findings reported here may also be relevant to more basic aspects of immune development in the GI tract. The chemokines CCL5, CXCL9, and CXCL10 are generally considered to be of the type 1 variety in that they preferentially attract Th1-type activated and memory T cells (25). IFN- is well-known to promote the differentiation of Th1 cells (51), and there is some evidence to suggest that certain chemokines, including CCL5, can also promote Th1 differentiation, probably via indirect effects on APCs (52). Thus, the age-dependent increase in expression of IFN-, CCL5, CXCL9, and CXCL10 in the intestine may contribute to the development of a local environment that favors Th1-type responses. This trend toward a Th1 bias is consistent with observations indicating a similar shift systemically, with neonatal mice tending to make a more Th2-dominated response to immunization than adults (53, 54). Interestingly, this age-dependent shift to Th1-dominated responses appears to be influenced by the commensal microflora (55), much like our observations on the expression of IFN--regulated genes in the intestine (Fig. 3). The presence of a Th1-promoting intestinal environment may be important in preventing the development of food allergies, which are characterized by inappropriately Th2-biased responses to food Ags. The age-dependent increase in IDO expression in the intestine is also likely to be relevant in this regard because it has been shown recently that induction of this enzyme in the lung suppresses the inflammation and airway hyperreactivity of asthma (56), consistent with its role in suppressing lymphocyte responses (29). Again, the importance of the intestinal commensal flora in preventing allergic responses has been indicated by both epidemiologic and laboratory data (57, 58, 59). Indeed, it has been shown clearly that antibiotic-induced alteration of the gut commensal flora can promote the development of a pulmonary allergic response to fungal spores (60, 61). It may be interesting to determine whether the reported beneficial effects of probiotic bacteria in allergy (62) might be related to their ability to induce the expression of IFN- and its target genes in the intestine.
It remains to be seen whether the expression of IFN- and its target genes increases in an age-dependent fashion in the human intestine. The overall sequence of development of the GALT is similar in mice and humans, and the period of weaning is associated with marked changes in this tissue in both species (11, 43). This was one reason for our choice of the preweaned mouse pup for our experiments, but it remains to be seen whether the responses of a 16- to 18-day-old mouse correspond to those of a young human infant. If it can be shown that our observations in the mouse model can be extended to humans, inadequate levels of IFN- and IFN--regulated genes could provide an explanation for the unusual susceptibility of young infants to extraintestinal spread of Salmonella and other bacteria (5, 6, 7). IFN--based interventions could then be considered as a way of boosting intestinal resistance to infection in high-risk neonates and young babies and may also be one approach to preventing the development of food allergies.
Acknowledgments
We thank Danny Park and Glenn Short for microarray analysis, Tricia Della Pelle and Atul Bhan for histopathologic analysis, Hai Ning Shi for assistance with FACS analysis, and Mohamed Bashir for assistance with animal studies.
Disclosures
The authors have no financial conflict of interest.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by the National Institutes of Health RO1AI48815 (to B.J.C.), RO1DK070260 (to W.A.W.), and T32DK07477 (to S.J.R.). S.J.R. is the recipient of a Harvard Medical School Fellowship in Pediatric Gastroenterology and Nutrition.
2 Address correspondence and reprint requests to Dr. Bobby J. Cherayil, Massachusetts General Hospital and Harvard Medical School, Building 114, Room 3400, 16th Street, Charlestown, MA 02129. E-mail address: cherayil@helix.mgh.harvard.edu
3 Abbreviations used in this paper: GI, gastrointestinal; MPO, myeloperoxidase; Ct, threshold cycle.
Received for publication February 17, 2005. Accepted for publication May 12, 2005.
References
Kosek, M., C. Bern, R. L. Guerrant. 2003. The global burden of diarrhoeal disease, as estimated rom studies published between 1992 and 2000. Bull. World Health Organ 81: 197-204.
Black, R. E., S. S. Morris, J. Bryce. 2003. Where and why are 10 million children dying every year?. Lancet 361: 22-26.
King, C. K., R. Glass, J. S. Bresee, C. Duggan. 2003. Managing acute gastroenteritis among children: oral rehydration, maintenance, and nutritional therapy. MMWR Recomm. Rep. 52: 1-16.
Preliminary FoodNet data on the incidence of food-borne illnesses: selected sites, United States, 2001. MMWR Morb. Mortal. Wkly. Rep. 51:2002 325-329.
Brumell, J. H., O. Steele-Mortimer, B. B. Finlay. 1999. Bacterial invasion: force feeding by Salmonella. Curr. Biol. 9: R277-R280.
Wallis, T. S., E. E. Galyov. 2000. Molecular basis of Salmonella-induced enteritis. Mol. Microbiol. 36: 997-1005.
Kuppermann, N.. 1999. Occult bacteremia in young febrile children. Pediatr. Clin. North Am. 46: 1073-1109.
Schutze, G. E., S. E. Schutze, R. S. Kirby. 1997. Extraintestinal salmonellosis in a children’s hospital. Pediatr. Infect. Dis. J. 16: 482-485.
nella bacteremia in infants with Salmonella gastroenteritis. J. Pediatr. 108:(5 Pt. 1): 718-721.(Sue J. Rhee, W. Allan Wal)
Young infants are highly susceptible to systemic dissemination of enteric pathogens such as Salmonella typhimurium when compared with older individuals. The mechanisms underlying this differential susceptibility have not been defined clearly. To better understand this phenomenon, we examined the responses of adult mice and preweaned pups to oral infection by S. typhimurium. We found clear age-specific differences, namely, an attenuated intestinal inflammatory response and a higher systemic bacterial burden in the pups compared with the adults. To elucidate the molecular basis for these differences, we obtained a microarray-based profile of gene expression in the small intestines of uninfected adult and preweaned animals. The results indicated a striking age-dependent increase in the intestinal expression of a number of IFN--regulated genes involved in antimicrobial defense. This finding was confirmed by real-time quantitative PCR, which also demonstrated an age-dependent increase in intestinal expression of IFN-. The developmental up-regulation of the IFN--regulated genes was dependent on both IFN- and a normal commensal microflora, as indicated by experiments in IFN--knockout mice and germfree mice, respectively. However, the increase in expression of IFN- itself was independent of the commensal flora. The functional importance of IFN- in the immunological maturation of the intestine was confirmed by the observation that the response of adult IFN--knockout animals to S. typhimurium infection resembled that of the wild-type pups. Our findings thus reveal a novel role for IFN- in the developmental regulation of antimicrobial responses in the intestine.
Introduction
One of the most common infectious agents isolated from cases of acute gastroenteritis is Salmonella, with serovar Typhimurium being the most frequent isolate. The incidence of infection is highest in infants < 1 year of age with a rate of 134 cases/100,000 children (4).
S. typhimurium that is introduced into the gastrointestinal (GI)3 tract via contaminated food or water adheres to and invades epithelial cells of the intestine, particularly the M cells overlying Peyer’s patches. The process of invasion is an active one in which bacterial effector proteins introduced into host cells by a type III secretion system induce cytoskeletal rearrangements that result in engulfment of the Salmonella by large membrane protrusions (5). The effector proteins also activate a number of cellular signaling pathways, leading to the acute inflammatory response that is responsible for the manifestations of gastroenteritis (6). In most individuals, this response is sufficient to clear the infection so that systemic dissemination of the organism is unusual. However, young infants, particularly those <3 mo of age, have a high risk for complications resulting from extraintestinal spread of Salmonella, including sepsis, osteomyelitis, meningitis, and death (7, 8, 9).
It is not clear why young infants are so prone to spread of S. typhimurium from the gut. It may be part of a more generalized impairment of intestinal defense against enteric pathogens because other systemic infections in this age group often originate in the GI tract (10). Inadequate development of epithelial barrier functions and mucosal immunity may underlie this phenomenon (11), but the exact molecular basis for this immaturity, as well as the factors involved in the normal maturation of protective responses, are yet to be elucidated.
To better understand the age-specific differences in the intestinal response to bacteria, we have made use of the mouse model of Salmonella infection. Although infection of mice with S. typhimurium typically results in a systemic infection without overt intestinal inflammation, recent studies by Barthel et al. (12) have shown that following pretreatment of the animals with oral streptomycin, infection with a virulent, streptomycin-resistant strain of Salmonella results in a robust and reproducible inflammation of the cecum. The pathologic changes demonstrated in this murine model closely resembled the intestinal inflammation caused by S. typhimurium infection in humans. We have used this approach to compare the responses of young mouse pups (preweaned, 16–18 days old) and their adult (5–6 wk old) counterparts to oral infection with Salmonella.
Materials and Methods
Mice
C57BL/6 (B6) and B6.129S7-Ifngrtm1Agt/J mice ranging in age from 5 days to 10 wk old were obtained from The Jackson Laboratory and Charles River Laboratories. BALB/c mice ranging in age from 5 days to 6 wk old were obtained from Charles River Laboratories. Tissue obtained from 6-wk-old BALB/c germfree mice were kindly provided by Dr. N. Nanthakumar (Massachusetts General Hospital, Charlestown, MA). These mice were obtained from the Gnotobiotic Animal Core at the University of North Carolina. Animals were sacrificed within the germfree facility, intestinal tissue was removed under sterile conditions, snap frozen in liquid nitrogen, and stored at –80°C until use. All animal studies were approved by the Massachusetts General Hospital Subcommittee on Research Animal Care and conformed to the U.S. Department of Agriculture Animal Welfare Act, Public Health Service Policy on Humane Care and Use of Laboratory Animals, and other applicable laws and regulations.
Quantitation of commensal bacteria
Sixteen- to 18-day-old and 6-wk-old B6 mice (three of each age group) were treated with streptomycin (12 and 20 mg, respectively) via the orogastric route for 24 h and then sacrificed. The terminal ileum and colon from these animals, as well as from age-matched, untreated controls, were removed, rinsed free of contents, homogenized in 1 ml of sterile PBS, and serial dilutions of the homogenates plated on Luria-Bertani agar for culture under aerobic and anaerobic conditions. Culturable commensal flora were quantified as colony-forming units per milligram of tissue. The limit of detection equaled 250 CFU/mg tissue.
Salmonella typhimurium infection
Five- to 6-wk-old B6 mice were housed in groups of up to four animals under standard barrier conditions in a specific pathogen-free facility. Sixteen- to 18-day-old B6 litters of up to eight were housed with a lactating female. Water and food were supplied ad libitum. Adult mice were pretreated with 20 mg of streptomycin, and pups were pretreated with 12 mg of streptomycin via the orogastric route. At 24 h following streptomycin treatment, 5- to 6-wk-old and 16- to 18-day-old animals were infected with 1 x 108 and 3 x 107 CFU of the streptomycin-resistant S. typhimurium strain SL1344, respectively. At 48 h postinfection, mice were sacrificed by CO2 asphyxiation. Tissue samples from the small intestine, cecum, colon, and spleen were removed for additional analysis. Small intestine and cecum were rinsed of fecal contents with sterile PBS. Intestinal tissue and spleen were homogenized using 1 ml of sterile 1% Triton X-100. The numbers of colony-forming units were determined by plating serial dilutions on Luria-Bertani streptomycin and incubating at 37°C for 24 h.
Histologic analysis
Segments of ileum, cecum, and colon were rinsed free of contents and embedded in Tissue Tek OCT (Sakura Finetek), snap frozen in liquid nitrogen, and stored at –80°C. Tissue samples were then transferred and fixed in paraffin. Sections (5 or 10 μm) were mounted on glass slides, air dried, and stained with H&E. Immunoperoxidase staining was performed using a goat polyclonal IgG anti-mouse IDO Ab (1/50 dilution) (Santa Cruz Biotechnology).
Myeloperoxidase (MPO) assay
Whole cecum was resected, rinsed free of contents in sterile PBS, and homogenized in 1 ml of 0.05% hexadecyltrimethylammonium bromide (HTAB) buffer. Samples were snap frozen in liquid nitrogen for 1 min and then immediately thawed at 37°C for 10 min for a total of three such cycles. Samples were then sonicated on ice for 20 s and centrifuged at 16,000 x g for 10 min. MPO activity in the clarified sonicate was then estimated by a colorimetric assay as described in detail earlier (13)
Gentamicin protection assay
Overnight cultures of SL1344 were collected by centrifugation at 6000 x g for 10 min, washed with sterile PBS, and resuspended in antibiotic-free tissue culture medium. Macrophages were harvested from the peritoneum of adult B6 mice on the fourth day after injection of sterile thioglycollate broth as described previously (14). MODE-K cells (kindly provided by Dr. R. Blumberg (Brigham and Women’s Hospital, Boston, MA) were cultured in DMEM with 10% FCS. Cells were left untreated (control) or were pretreated overnight with 10 ng/ml recombinant murine IFN- (R&D Systems). The cells were infected with 107 bacteria (multiplicity of infection of 10:1). After 1 h of infection, cells were washed with sterile PBS and placed in DMEM with 100 μg/ml gentamicin for 2 h to kill extracellular bacteria. The cells were then lysed either immediately or after an additional 18 h in 10 μg/ml gentamicin, and the number of intracellular bacteria was determined at each time point by culturing serial dilutions of cell lysates, as described in detail earlier (15).
Isolation of intestinal epithelial cells
Epithelial cells were isolated from B6 mice ranging in age from 14 days to 8 wk old as per the methods described by Evans et al. (16). In brief, the small intestine (from the terminal ileum to the ligament of Treitz) was resected, dissected open, cut into 2- to 3-mm pieces, and rinsed clean in 50 ml of Ca2+- and Mg2+-free HBSS at room temperature. Tissue was then diced into <1-mm3 pieces, incubated with 20 ml of enzyme solution (0.1 mg/ml dispase type I and 300 U/ml collagenase type XI (Sigma-Aldrich)) for 30 min with agitation at 80 cycles/min. The cell suspension was pipetted vigorously for 3 min and left to sediment under gravity for 1 min. The supernatant was collected and added to 10 ml of DMEM + 2.5% FCS with 2% sorbitol and centrifuged at 1000 x g for 2 min. This process was repeated until the supernatant was clear, and the pellet was well defined. The cell pellet was resuspended in DMEM + 2.5% FCS. Viability was assessed by staining cells with trypan blue solution (0.4%) and quantified with a hemacytometer. Purity of the preparation was analyzed by FACS analysis after staining with the TROMA-1 epithelial cell-specific marker (Developmental Studies Hybridoma Bank, University of Iowa) (17). We routinely obtained preparations containing 90% epithelial cells.
Microarray-based gene expression profiling
Experiment design. To investigate the molecular basis of the age-dependent differences in response to Salmonella described above, a microarray-based gene expression profiling analysis was performed on RNA prepared from intestinal epithelial cells of uninfected preweaned pups and postweaned adult mice to detect changes in the expression of genes involved in the immune response. A direct comparison of gene expression at different stages of development was made. Hybridizations were performed in triplicate.
Sample RNA preparation. Total RNA was extracted and purified from intestinal epithelial cells (isolated as described in Materials and Methods) from 16- and 33-day-old B6 mice (TRIzol reagent, Invitrogen Life Technologies; RNeasy column, Qiagen). Samples were pooled from five separate animals to yield a total quantity of 30 μg of RNA for each group. RNA quantity and quality were assessed by using a Beckman Coulter DU 640 Spectrophotometer (Beckman Coulter) and Agilent 2100 Bioanalyzer (Agilent), according to the manufacturer’s protocols.
Extract preparation and labeling. Briefly, two samples (a reference and a test) of isolated total RNA were reverse transcribed into amine modified cDNA. This cDNA was labeled via a coupling reaction to the Cy3 or Cy5 dyes, then purified and concentrated for hybridization. The protocol used was a modified version of the Atlas Powerscript Fluorescent Labeling kit (Clontech 634712) protocol, and full details are available online (https://dnacore.mgh.harvard.edu/microarray/protocol-labeling.shtml).
Hybridization procedures and parameters. For direct comparison, Cy3- and Cy5-labeled cDNA samples were resuspended in hybridization buffer and then hybridized to a PGA Mouse v1.0 array (MGH_Mm_01) slide made available through the Microarray Core Facility at the Massachusetts General Hospital. Both preweaned and postweaned samples were hybridized to the same slide, and each set of hybridizations was performed in triplicate. Full details regarding the hybridization protocol are available online (https://dnacore.mgh.harvard.edu/microarray/protocol-hybridization.shtml).
Measurement data and specifications. BioArray Software Environment (18) and GenMAPP (19, 20) software programs were used for data analysis. Statistical significance (p value) was calculated by Student’s t test based on the results of three separate hybridizations.
Array design. A PGA Mouse v1.0 array (MGH_Mm_01) was made available through the Microarray Core Facility at the Massachusetts General Hospital. A total of 14,609 oligonucleotides was arrayed onto 384-well Genetrix polystyrene V-bottom plates (X6004).
Full details regarding the microarray printing protocol are available online (https://dnacore.mgh.harvard.edu/microarray/protocol-printing.shtml). Specific details characterizing the individual spots on the array are available at the following web site: https://dnacore.mgh.harvard.edu/microarray/files/anno_dist-MGH_Mm_01.txt.
Results
Age-specific differences in the response to oral infection with Salmonella
Based on the protocol described by Barthel et al. (12), 16- to 18-day-old (preweaned) and 5- to 6-wk-old (adult) B6 mice were pretreated with a single dose of streptomycin per os and then inoculated orally 24 h later with the streptomycin-resistant strain of S. typhimurium, SL1344. The doses of both the streptomycin and the bacteria were adjusted to the weight of the animals. The basal number of gut commensal bacteria that could be cultured from the adult animals was 270-fold higher than in the pups. The streptomycin treatment significantly reduced the number of culturable commensals in both age groups (from 6350 CFU/mg intestinal tissue to <250 CFU/mg tissue—the limit of detection—in the case of the pups; from 1.73 x 106 to 1.13 x 104 CFU/mg in the case of the adults). The animals were sacrificed 48 h after the infection, and the inflammatory response in the cecum was assessed by histopathology, increase in MPO activity (as a measure of neutrophil recruitment), and up-regulation of the proinflammatory cytokine TNF-. At the same time, homogenates of the ileum and spleen were cultured to determine the number of bacteria in these tissues.
The pretreatment with streptomycin alone did not appear to elicit an inflammatory response in the intestine, either by gross appearance or by histopathology, nor were there any appreciable differences between the adult mice and the pups (Fig. 1A). However, following 48 h of infection with SL1344, there was a marked thickening of the cecal wall in the infected 5- to 6-wk-old mice compared with the streptomycin-pretreated, uninfected age-matched controls upon gross inspection. Microscopically, an inflammatory response was evidenced by marked submucosal edema and cellular infiltrate of the lamina propria (Fig. 1, B and C). In contrast, infection did not appreciably alter the gross appearance of the cecum in the 16- to 18-day-old mice. On histologic analysis in the younger animals, both the degree of submucosal edema and inflammatory infiltration were less than in the adults (Fig. 1, B and C).
To obtain a more quantitative assessment of the intestinal inflammation, we measured tissue MPO activity as an indicator of neutrophil infiltration, one of the characteristic features of Salmonella-induced enteritis (13, 21, 22). The results of this analysis demonstrated that there was a 37.6-fold (±11.3, p = 0.007) increase in activity in the infected adult mice compared with uninfected controls (Fig. 1D). In comparison, there was only a 1.06-fold (±0.44, p > 0.05) increase in MPO activity in the infected 18-day-old mice, indicating a lower level of neutrophil recruitment than in the adults, consistent with the histopathological findings.
Given the role of TNF- in the recruitment of neutrophils to sites of Salmonella infection (23), and in the defense against this organism (24), we examined cecal expression of this cytokine by quantitative real-time PCR. Following infection, there was a 14.6-fold (±0.28, p = 0.02) increase in TNF- transcript levels in the 5- to 6-wk-old mice. In contrast, infection-induced up-regulation of TNF- was significantly less (p = 0.01) in the 16- to 18-day-old pups: there was only a 7.06-fold (±1.57, p = 0.02) increase in the level of the transcript (Fig. 1E).
Taken together, the results of the histopathologic evaluation, MPO assay, and TNF- real-time PCR analysis all indicated that there was a decreased intestinal inflammatory response in the preweaned pups compared with the adult mice following infection with S. typhimurium. This result was not altered even if the adult doses of streptomycin, and Salmonella were administered to the pups (data not shown).
As mentioned above, S. typhimurium infection in the mouse is typically associated with invasion of the epithelium via the Peyer’s patches in the small intestine, particularly the ileum, and dissemination of the organism to extraintestinal sites such as the spleen. Therefore, to determine whether there were age-dependent differences in local and systemic infection, cultures of the ileum and spleen were performed 48 h after oral infection of preweaned pups or adult mice with SL1344. The results demonstrated that there was a significantly greater number of bacteria in both the ileum (3161 ± 2065 vs 434 ± 318, p = 0.04) and the spleen (9410 ± 5656 vs 872 ± 160, p = 0.0007) of the pups compared with the adult mice, respectively (Fig. 1, F and G). Thus, the less pronounced intestinal inflammatory response in the pups was associated with a greater local and systemic burden of bacteria.
Age-dependent changes in intestinal gene expression
To investigate the molecular basis of the age-dependent differences in response to Salmonella described above, a microarray-based gene expression profiling analysis was performed on RNA prepared from small intestinal epithelial cells of preweaned pups and adult mice to detect changes in the expression of genes involved in the immune response. The RNA samples were reverse transcribed, then hybridized to a PGA Mouse v1.0 array (MGH_Mm_01) slide (made available through the Microarray Core Facility at the Massachusetts General Hospital) containing 14,609 oligonucleotides. Following quality control analysis, 6791 usable signals were identified. Data from the microarray were further analyzed using a web-based software program BioArray Software Environment (https://base.mgh.harvard.edu) (18). A direct comparison of expression was made between the preweaned and the adult stages of development for each gene and was quantified as a fold change. A fold change > 1.0 indicated an increase in expression of that particular gene in the adult, whereas a value < 1.0 represented a decrease in expression.
Results from the microarray demonstrated an age-dependent increase in a number of antimicrobial, IFN--regulated genes. In contrast, there were no statistically significant differences in expression of housekeeping genes such as GAPDH, pyruvate carboxylase, 2-microglobulin, and cyclophilin A. There were a total of 460 genes in which there was at least a 2-fold, statistically significant increase in expression with aging. Of these, there were 33 genes that were classified by the BioArray Software Environment and GenMAPP programs (18, 19, 20) under the biologic process of "immune response." More strikingly, of these 33 immune response genes, 10 were IFN- regulated. In addition, there were at least another five IFN--regulated genes identified by inspection whose expression increased significantly in an age-dependent manner (Table I). The IFN--regulated genes included several chemokines involved in innate and adaptive immunity (25), members of a GTPase family implicated in resistance to intracellular infection (26), and other proteins such as the polyimmunoglobulin receptor, which is required for the transport of secretory Ig across mucosal surfaces (27), and IDO, which has been shown to be involved in both innate defense against several types of pathogens and in the regulation of lymphocyte activation (28, 29). Intriguingly, one of the genes in this list was IFN-induced protein with tetratricopeptide repeats 1, the orthologue of which was recently identified as being up-regulated in the intestine of gnotobiotic zebrafish following colonization with bacteria (30), suggesting that the increased expression of IFN--induced genes may be an evolutionarily conserved feature of gut maturation. Of note, the results of the microarray analysis did not reveal any differences in the expression of IFN- or the IFN-R
One important change that occurs during the weaning period is the colonization of the intestine with commensal flora, and it is well-known that commensal bacteria can influence gene expression in the intestinal epithelium, both in mammals and in lower vertebrates (30, 31). To determine whether the commensal flora was involved in the age-dependent increase in expression of the IFN--regulated genes in the intestine, quantitative real-time PCR for CCL5, CXCL9, and IDO was performed on RNA prepared from the ileum of adult germfree and conventionally raised mice. Because the only germfree mice available to us were of the BALB/c strain, we first demonstrated that the expression of these genes did in fact increase in an age-dependent manner in BALB/c mice as in the B6 animals (Fig. 3B). When expression of CCL5, CXCL9, and IDO was analyzed in adult germfree BALB/c mice, it was found to be significantly less than in the age-matched conventionally reared counterparts and was similar to that of the conventionally raised preweaned BALB/c pups (Fig. 3B). Thus, there was a significant attenuation of the normal developmental increase in IFN--regulated gene expression in the germfree mice, suggesting that the commensal flora is involved in regulating expression of these genes.
We also examined the effect of the commensal flora on IFN- expression in the intestine. In contrast to the results for CCL5, CXCL9, and IDO, the level of IFN- expression in the intestine of adult germfree mice was not significantly different from the conventionally reared controls (Fig. 3C), suggesting that the commensal flora did not play a major role in the age-dependent increase in IFN- in the gut.
IFN- enhances antimicrobial mechanisms of intestinal epithelial cells and macrophages in vitro
The results presented so far indicate an important role for IFN- in the developmental changes in gene expression in the intestine. Some of these IFN--regulated genes have been implicated in antimicrobial defense at the cellular level (26, 28), suggesting that one of the effects of IFN- may be to increase epithelial resistance to bacterial infection. To evaluate this possibility, we examined the effect of IFN- on the murine small intestinal epithelial cell line MODE-K, which has been shown previously to be responsive to this cytokine (32). MODE-K cells, pretreated overnight with 10 ng/ml rIFN- or left untreated, were infected with S. typhimurium strain SL1344, and the number of bacteria present inside control or IFN--treated cells was determined at 2 and 20 h after infection by gentamicin protection assay (15). Thioglycollate-elicited peritoneal macrophages from adult B6 mice were evaluated similarly in parallel. As shown in Fig. 4A, the IFN- treatment decreased the number of bacteria surviving in the MODE-K cells by 50% at the 20-h time point. This effect was seen as early as 6 h postinfection (data not shown). There was no difference in the number of intracellular bacteria at 2 h postinfection, suggesting that the IFN- is unlikely to be affecting the number of bacteria that initially entered the cells. IFN- treatment had a similar, even more dramatic, effect on the survival of Salmonella inside the thioglycollate-elicited peritoneal macrophages (Fig. 4B). Thus, IFN- acts on two cell types present in the intestinal mucosa, enterocytes, and macrophages to enhance microbicidal mechanisms that limit the intracellular growth of Salmonella.
IFN- influences the intestinal response to Salmonella and systemic growth of bacteria in vivo
We addressed the in vivo role of IFN- in the age-specific differences in response to enteric Salmonella infection by infecting adult IFN--knockout mice with S. typhimurium and comparing their response to those of the wild-type adult animal. As before, the inflammatory response in the cecum was assessed by histology, MPO assay, and TNF- quantitative real-time PCR. Histologic analysis revealed less submucosal edema and decreased lamina propria inflammatory infiltrate in the adult IFN--knockout animals compared with the age-matched, wild-type controls (Fig. 5, A and B). Similarly, both the MPO assay and the TNF- real-time PCR demonstrated a significantly attenuated response in the adult IFN--knockout mice compared with wild-type animals (Fig. 5, C and D). Compared with uninfected controls, there was only a 13.3-fold (±3.2, p = 0.02) increase in MPO activity in the IFN--knockout mice as opposed to a 37.6-fold (±11.27, p = 0.007) increase in the wild-type animals. With regard to TNF- gene expression, there was only a 5.85-fold (±0.51, p = 0.01) increase in transcript levels following infection in the IFN--knockout mice compared with a 13.8-fold (±0.51, p < 0.0001) increase in the wild-type animals. To determine whether there were IFN--dependent differences in local and systemic bacterial numbers, cultures of the ileum and spleen were performed. The results demonstrated that there was a significantly greater number of bacteria in the spleen of IFN--knockout mice compared with the wild-type controls (Fig. 5E). There was also a trend toward greater numbers of bacteria in the ileum of the IFN--knockout mice, although this did not reach statistical significance (data not shown). Overall, the responses of the adult IFN--deficient mice resembled those of the wild-type pups, an observation that is consistent with the idea that the cytokine is required for the normal maturation of the response to enteric Salmonella infection.
Discussion
The experiments reported in this article reveal a novel role for IFN- in the immunological maturation of the gut. The mechanistic explanation for this role of IFN- is suggested by the identity of the IFN--regulated genes that are up-regulated in the intestine in an age-dependent manner. IFN--induced GTPases such as LRG-47 have been shown to inhibit intracellular multiplication of Salmonella and other pathogens (26). Similarly, IDO has been shown to inhibit growth of several types of bacteria and certain parasites (28). The net effect of increased expression of these molecules would be to make cells of the intestinal mucosa more resistant to bacterial multiplication, an outcome that is consistent with our in vitro studies with IFN--treated macrophages and intestinal epithelial cells (Fig. 4), as well as in the in vivo experiments with wild-type and IFN--deficient mice (Figs. 1 and 5). The age-dependent increase in expression of other IFN--regulated genes, such as the polyimmunoglobulin receptor (Table I) and the chemokines CCL5, CXCL9, and CXCL10 (Table I and Fig. 2), is also likely to contribute to making the adult intestine more resistant to microbial entry and multiplication by promoting, respectively, the transepithelial transport of protective secretory Abs (27) and the recruitment of NK cells and macrophages to augment innate protection against infection (33, 34, 35). Finally, the age- and IFN--dependent enhancement of intestinal TNF- expression in response to infection (Figs. 1 and 5) results in greater recruitment of inflammatory cells such as neutrophils that are able to deal with bacterial invaders. Specifically relevant to our observations is the fact that IFN- has been shown to have a protective role in Salmonella infection and to enhance the expression of TNF- in response to LPS (36, 37).
Our quantitative real-time PCR analyses indicate that both intestinal epithelial cells and lamina propria cells such as macrophages are likely sites of expression of the IFN--regulated genes because similar results were obtained regardless of whether RNA from purified epithelial cells or whole intestinal tissue was used. Indeed, immunohistochemical localization of IDO in the intestine indicates expression in both the epithelium and in subepithelial cells (Fig. 2). Chemokines such as CCL5, CXCL9, and CXCL10 have been shown to be expressed by intestinal epithelial cells, especially in response to infection (34, 38, 39, 40). The cellular source of IFN- in the intestine is at present unclear. The results of quantitative real-time PCR indicate that it is likely to be in the nonepithelial compartment because an age-dependent increase in expression of the cytokine was observed only when RNA from whole intestinal tissue, rather than purified epithelial cells, was used. Candidate IFN--producing cells include macrophages, NK cells, NKT cells, and T lymphocytes, and experiments to determine which of these cell types is the site of the developmentally regulated expression of IFN- in the intestine are ongoing.
The results of our studies indicate an important role for the commensal flora in the induction of IFN--regulated genes in the intestine (Fig. 3), which is consistent with the age-dependent increase in number and complexity of these organisms and with their ability to induce alterations in gene expression and function in cells of the gut (26, 27, 38, 39, 40, 41). The TLR family of cell surface proteins, which recognize microbial components such as LPS, flagellin, and bacterial lipopeptide, constitute one set of candidate molecules that might be involved in sensing and responding to the commensal flora (41). Interestingly, the absence of a normal commensal flora did not affect the age-dependent increase in intestinal IFN- expression (Fig. 3). It is possible that other factors associated with the weaning period, such as the introduction of solid food or the surge in circulating corticosteroids, might be involved in regulating expression of IFN- in the intestine (11). Regardless, our results suggest that signals from the commensal flora and IFN- act independently to influence the expression of CCL5, CXCL9, and IDO. It is possible that the two sets of signals cooperate at the level of transcription to up-regulate these genes, as has been reported in other systems (42).
It is important to note that factors other than the change in IFN--regulated gene expression could contribute to the age-specific differences in immune response seen in our experiments. The GALT, including Peyer’s patches, the mesenteric lymph nodes, and the lymphocyte population of the lamina propria, are all underdeveloped in preweaned mice (43). This immaturity of organized lymphoid structures could place constraints on the kind of local immune response that develops. In addition, although we were able to detect Salmonella in the mesenteric lymph nodes even in the preweaned mice (data not shown), the lack of fully mature Peyer’s patches in these animals may favor the translocation of the organism by alternate routes, similar to what has been described for the uptake and extraintestinal dissemination of noninvasive bacteria by a population of CD18-positive phagocytes (44). The greater number and complexity of the commensal flora of the adult animals could also modify the response to Salmonella by amplifying local inflammation or by limiting colonization, and therefore spread, of the pathogen. The fact that the preweaned animals were still on breast milk is another factor that has to be considered. The presence of antimicrobial and immunomodulating factors in breast milk, such as secretory IgA, lactoferrin, IL-10, TGF-, and IL-1R antagonist may act to alter the inflammatory response (45). Finally, changes in intestinal expression of genes other than those that we have analyzed might play a role in the age-specific differences in immune response. For instance, our microarray data indicated that the expression of TLR3 was higher in the adult mice than in the pups (10.5-fold difference, p = 0.02). We have not pursued this observation further in the present series of experiments, but it is possible that alterations in TLR expression could contribute to differences in the inflammatory and antimicrobial responses between the adult and preweaned animals.
Although our studies have focused on the response to Salmonella, the results are likely to be relevant to other infections. Neonatal and weanling mice are more susceptible than adult animals to several intestinal pathogens, including Vibrio cholerae, enterotoxigenic Escherichia coli, Campylobacter jejuni, and Cryptosporidium parvum (46, 47, 48, 49). In the case of Cryptosporidium infection, lack of IFN- has been shown to confer susceptibility to adult mice (50). It would be worth examining whether developmental changes in the expression of IFN--regulated genes in the intestine contribute to the age-specific differences in susceptibility to these infections.
The findings reported here may also be relevant to more basic aspects of immune development in the GI tract. The chemokines CCL5, CXCL9, and CXCL10 are generally considered to be of the type 1 variety in that they preferentially attract Th1-type activated and memory T cells (25). IFN- is well-known to promote the differentiation of Th1 cells (51), and there is some evidence to suggest that certain chemokines, including CCL5, can also promote Th1 differentiation, probably via indirect effects on APCs (52). Thus, the age-dependent increase in expression of IFN-, CCL5, CXCL9, and CXCL10 in the intestine may contribute to the development of a local environment that favors Th1-type responses. This trend toward a Th1 bias is consistent with observations indicating a similar shift systemically, with neonatal mice tending to make a more Th2-dominated response to immunization than adults (53, 54). Interestingly, this age-dependent shift to Th1-dominated responses appears to be influenced by the commensal microflora (55), much like our observations on the expression of IFN--regulated genes in the intestine (Fig. 3). The presence of a Th1-promoting intestinal environment may be important in preventing the development of food allergies, which are characterized by inappropriately Th2-biased responses to food Ags. The age-dependent increase in IDO expression in the intestine is also likely to be relevant in this regard because it has been shown recently that induction of this enzyme in the lung suppresses the inflammation and airway hyperreactivity of asthma (56), consistent with its role in suppressing lymphocyte responses (29). Again, the importance of the intestinal commensal flora in preventing allergic responses has been indicated by both epidemiologic and laboratory data (57, 58, 59). Indeed, it has been shown clearly that antibiotic-induced alteration of the gut commensal flora can promote the development of a pulmonary allergic response to fungal spores (60, 61). It may be interesting to determine whether the reported beneficial effects of probiotic bacteria in allergy (62) might be related to their ability to induce the expression of IFN- and its target genes in the intestine.
It remains to be seen whether the expression of IFN- and its target genes increases in an age-dependent fashion in the human intestine. The overall sequence of development of the GALT is similar in mice and humans, and the period of weaning is associated with marked changes in this tissue in both species (11, 43). This was one reason for our choice of the preweaned mouse pup for our experiments, but it remains to be seen whether the responses of a 16- to 18-day-old mouse correspond to those of a young human infant. If it can be shown that our observations in the mouse model can be extended to humans, inadequate levels of IFN- and IFN--regulated genes could provide an explanation for the unusual susceptibility of young infants to extraintestinal spread of Salmonella and other bacteria (5, 6, 7). IFN--based interventions could then be considered as a way of boosting intestinal resistance to infection in high-risk neonates and young babies and may also be one approach to preventing the development of food allergies.
Acknowledgments
We thank Danny Park and Glenn Short for microarray analysis, Tricia Della Pelle and Atul Bhan for histopathologic analysis, Hai Ning Shi for assistance with FACS analysis, and Mohamed Bashir for assistance with animal studies.
Disclosures
The authors have no financial conflict of interest.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by the National Institutes of Health RO1AI48815 (to B.J.C.), RO1DK070260 (to W.A.W.), and T32DK07477 (to S.J.R.). S.J.R. is the recipient of a Harvard Medical School Fellowship in Pediatric Gastroenterology and Nutrition.
2 Address correspondence and reprint requests to Dr. Bobby J. Cherayil, Massachusetts General Hospital and Harvard Medical School, Building 114, Room 3400, 16th Street, Charlestown, MA 02129. E-mail address: cherayil@helix.mgh.harvard.edu
3 Abbreviations used in this paper: GI, gastrointestinal; MPO, myeloperoxidase; Ct, threshold cycle.
Received for publication February 17, 2005. Accepted for publication May 12, 2005.
References
Kosek, M., C. Bern, R. L. Guerrant. 2003. The global burden of diarrhoeal disease, as estimated rom studies published between 1992 and 2000. Bull. World Health Organ 81: 197-204.
Black, R. E., S. S. Morris, J. Bryce. 2003. Where and why are 10 million children dying every year?. Lancet 361: 22-26.
King, C. K., R. Glass, J. S. Bresee, C. Duggan. 2003. Managing acute gastroenteritis among children: oral rehydration, maintenance, and nutritional therapy. MMWR Recomm. Rep. 52: 1-16.
Preliminary FoodNet data on the incidence of food-borne illnesses: selected sites, United States, 2001. MMWR Morb. Mortal. Wkly. Rep. 51:2002 325-329.
Brumell, J. H., O. Steele-Mortimer, B. B. Finlay. 1999. Bacterial invasion: force feeding by Salmonella. Curr. Biol. 9: R277-R280.
Wallis, T. S., E. E. Galyov. 2000. Molecular basis of Salmonella-induced enteritis. Mol. Microbiol. 36: 997-1005.
Kuppermann, N.. 1999. Occult bacteremia in young febrile children. Pediatr. Clin. North Am. 46: 1073-1109.
Schutze, G. E., S. E. Schutze, R. S. Kirby. 1997. Extraintestinal salmonellosis in a children’s hospital. Pediatr. Infect. Dis. J. 16: 482-485.
nella bacteremia in infants with Salmonella gastroenteritis. J. Pediatr. 108:(5 Pt. 1): 718-721.(Sue J. Rhee, W. Allan Wal)