Simian immunodeficiency virus–induced mucosal interleukin-17 deficiency promotes Salmonella dissemination from the gut

T_H17 responses dominate the Salmonella gene expression profile

Four young adult healthy rhesus macaques (MK5, MK7, MK8 and MK9) underwent loop surgery. Loops of each macaque were inoculated by injecting either a S. typhimurium culture or sterile culture medium (mock infection) into the intestinal lumen, and individual loops of each treatment group were removed surgically at 2, 5 and 8 h after inoculation. We monitored the host response to S. typhimurium infection in vivo at defined early time points after infection and compared it to responses in mock-infected tissue collected from the same macaque at the same time point. S. typhimurium infection resulted in marked enteritis and fluid accumulation at 5 and 8 h after inoculation. To characterize mucosal responses to S. typhimurium infection, we performed gene expression profiling by comparing mRNA levels from S. typhimurium–infected or mock-infected loops (Fig. 1). The gene expression profile induced by S. typhimurium was dominated by responses linked to immunity and inflammation. The strongest up-regulation of gene expression was observed for IL-22, followed by IFN-γ and IL-26 (Fig. 1c and data not shown). IL-22 and IFN-γ synergize to induce inducible nitric oxide synthase (iNOS) expression¹³, whose mRNA levels were markedly increased after S. typhimurium infection (Fig. 1c). Furthermore, S. typhimurium induced a marked upregulation of IL-17 expression and the expression of genes induced by IL-17, including those encoding lipocalin-2 (LCN2)¹⁴, CCL20 (MIP-3α)¹⁵, CXCL8 (IL-8)¹⁶ and CXCL10 (ref. ¹⁷). To verify changes in cytokine expression detected by gene expression profiling, we determined the mRNA levels of IL-17 and IL-17–regulated genes (those coding for MIP-3α, lipocalin-2 and IL-8) in samples from all four macaques by real-time PCR. Cytokines whose expression is not affected by IL-17, including interferon-γ (IFN-γ), tumor necrosis factor-α (TNF-α), IL-6, IL-23 p19, IL-12 and IL-23 p40, and transforming growth factor-β (TGF-β) were analyzed for comparison (Fig. 2). These results show a marked increase in the mRNA levels of all proinflammatory cytokines after S. typhimurium infection.

All macaques used in this study were raised in a Salmonella-free colony and tested negative for Salmonella serotypes. Although there was no evidence of previous exposure of the macaques to this pathogen, mRNAs whose expression was upregulated most strongly at a very early time point (that is, at 5 h) after S. typhimurium infection included IL-22, IL-26 and IL-17 (Fig. 1c). The main cellular sources of these cytokines are T cells^18–20. The importance of T cells in generating early inflammatory changes in response to S. typhimurium infection has not been apparent from previous studies^10,21. Thus, the increased expression of these cytokines attracted our attention because T cell depletion has been linked to an inability of HIV-affected individuals to contain S. typhimurium at the site of infection (that is, the intestinal mucosa), resulting in the development of a life-threatening bacteremia²². Therefore, we wanted to determine whether the IL-17 responses elicited by S. typhimurium were altered by an underlying SIV infection.

Salmonella-induced T_H17 responses are blunted by SIV infection

Four young adult healthy rhesus macaques (MK3, MK4, MK6 and MK10) were infected with SIV (SIV mac251), and blood samples were collected to monitor the number of CD4⁺ lymphocytes in the peripheral blood and the plasma viral load (Fig. 3). SIV-infected macaques underwent surgery between 6 and 10 weeks after inoculation. Compared to the SIV-negative macaques (see above), SIV-infected macaques had similar CD8⁺ counts in the peripheral blood at the time of surgery, whereas the numbers of CD4⁺ lymphocytes were lower. Loops of each macaque underwent S. typhimurium infection or mock infection as described above.

To identify mucosal responses elicited by S. typhimurium that were modified by an underlying SIV infection, gene expression profiling was performed with mRNA from S. typhimurium–infected and mock-infected loops collected from an SIV-infected macaque (MK6; Fig. 1d). The mRNA levels for genes involved in epithelial repair and maintenance were markedly upregulated in S. typhimurium–infected tissue from the SIV-negative macaque (MK5), and these responses were blunted by SIV infection (MK6) (Supplementary Fig. 1 online). Decreased expression levels of genes regulating epithelial barrier maintenance are also observed in people with HIV²³. However, mRNA expression of genes linked to cell death and stress response were induced more strongly during S. typhimurium challenge of the SIV-infected macaque than the control macaque. Notably, mRNA expression of IL-22, IL-17 and genes regulated by IL-17 was blunted in the SIV-infected macaque compared to the SIV-negative macaque (Fig. 1c). To verify changes in mRNA levels detected by gene expression profiling, samples from all four SIV-infected macaques taken throughout the time course of S. typhimurium infection were analyzed by real-time PCR (Fig. 2). No significant differences between SIV-infected and SIV-negative macaques were observed for the S. typhimurium–mediated induction of IFN-γ, TNF-α, TGF-β, IL-6, IL-23 p19 or IL-12 and IL-23 p40 expression. In contrast, S. typhimurium challenge resulted in significantly (P < 0.05) higher mRNA levels of IL-17, IL-22, MIP-3α, LCN2 and IL-8 in SIV-negative compared to SIV-infected macaques. These data suggest that SIV-positive macaques develop a cytokine deficiency during S. typhimurium infection because IL-17 responses, which dominate the gene expression profile elicited by this bacterial pathogen, are blunted in these macaques.

SIV-induced CD4⁺ depletion reduces IL-17 production

IL-17 production detected by fluorescence microscopy of loop tissue suggested a cellular localization for IL-17 in mock-infected loops (Supplementary Fig. 2 online). Staining with antibody to Il-17 was markedly increased in S. typhimurium–infected loops of control macaques but produced a diffuse background signal, presumably because much of this cytokine is released during infection (Supplementary Fig. 2). To study the cellular sources of IL-17, we performed in vitro experiments with lymphocyte populations in which the release of IL-17 was prevented by treatment with brefeldin A.

Because T cells are a major source of IL-17, we investigated whether SIV-mediated T cell depletion affects IL-17 production in mucosal T cells. Memory CD4⁺ T cell depletion in the intestinal mucosa occurs within the first few weeks after HIV infection in humans or SIV infection in rhesus macaques^24–28. Lymphocyte populations isolated from the ileal mucosa of four SIV-infected macaques (MK3, MK4, MK6 and MK10) and six control macaques (MK1, MK2, MK5, MK7, MK8 and MK9) were analyzed for expression of surface markers (CD3, CD4, CD8 and CD95) and for their capacity to produce IL-17 and IFN-γ. To this end, one 10-cm segment of the ileum that had not been exposed to S. typhimurium was collected from SIV-negative and SIV-infected macaques, and lamina propria lymphocytes were isolated and analyzed by flow cytometry. The CD4⁺ T cell population was markedly depleted in the intestinal mucosa of SIV-infected macaques (P = 0.0005), and this was accompanied by a relative increase in the fraction of CD8⁺ T cells (Fig. 4a). For the eight macaques that underwent loop surgery (MK3–10), the severity of CD4⁺ T cell depletion in the ileal mucosa (measured in tissue that had not been exposed to S. typhimurium) correlated directly with the magnitude of IL-17 mRNA expression induced in the same macaque 5 h after in vivo challenge of the loops with S. typhimurium (Fig. 4b).

We next investigated whether ileal T cell populations in SIV-negative macaques (n = 6) differed from those in SIV-infected macaques (n = 4) in their capacity to produce IL-17. To this end, lamina propria lymphocytes from each macaque (isolated from tissue that had not been exposed to S. typhimurium) were stimulated ex vivo with phorbol 12-myristate 13-acetate (PMA) and ionomycin in the presence of brefeldin A (to prevent cytokine secretion) and then analyzed by multicolor flow cytometry to assess IL-17 and IFN-γ production by T cells (Fig. 4c,d). The majority of lamina propria CD3⁺ T cells showed high expression of CD95, indicative of a memory phenotype. The fraction of CD3⁺ T cells that showed a T_H1 phenotype (that is, IFN-γ–producing CD4⁺ T cells) was significantly (P = 0.019) smaller in SIV-infected macaques (1.9% on average) compared to control macaques (9.2% on average; Fig. 4d). This depletion of T_H1 cells did not, however, result in reduced IFN-γ expression during S. typhimurium infection (Fig. 2), presumably because these cells are not an important source of this cytokine at early times after infection. No significant differences between SIV-infected and SIV-negative macaques were observed in IL-17 production by CD8⁺ T cells in response to PMA-ionomycin stimulation. In contrast, the average percentage of lamina propria CD3⁺ T cells that was IL-17–producing and CD4⁺ (T_H17 cells) was significantly (P = 0.005) lower in SIV-infected macaques (0.12% on average) compared to control macaques (3.7% on average; Fig. 4c). These data show that SIV infection significantly lowers the number of T_H17 cells in the intestinal mucosa, which is predominantly the result of the overall CD4⁺ T cell depletion caused by the infection. The finding that SIV-mediated memory CD4⁺ T cell depletion significantly (P < 0.01) lowers the number of T_H17 cells provides a plausible explanation for the blunted IL-17 response elicited by S. typhimurium in SIV-infected macaques (Fig. 2).

IL-17 deficiency accelerates Salmonella dissemination

At the end of the experimental procedure (8 h after inoculation of loops), samples of the draining mesenteric lymph nodes of rhesus macaques were processed for quantification of S. typhimurium. SIV-infected macaques had significantly higher numbers (330-fold increase, P = 0.019) of S. typhimurium in the mesenteric lymph node than SIV-negative macaques (Fig. 5a). These data provide a noteworthy illustration of the magnitude by which SIV-induced defects in mucosal barrier function increase the ability of S. typhimurium to disseminate within its host.

To determine whether there is a causal link between the defect in IL-17 mRNA expression (Fig. 2) and increased bacterial dissemination (Fig. 5a), we used a well-established mouse model of S. typhimurium infection. Whereas mice normally develop few inflammatory changes in the intestinal mucosa during S. typhimurium infection, pretreatment with streptomycin drastically exacerbates neutrophil influx and inflammation in the cecal mucosa²⁹. By using this model, we have previously shown that Il-17 mRNA expression is increased 150-fold in the cecal mucosa of streptomycin–pretreated mice by 48 h after S. typhimurium infection³⁰. To assess the importance of IL-17 in controlling S. typhimurium infection, we pretreated two groups of six IL-17 receptor–deficient mice (Il17ra^−/− mice) bred on a C57BL/6 background with streptomycin and inoculated them with S. typhimurium or sterile LB broth. As a control, two groups of C57BL/6 mice were treated identically. Organs were collected 48 h after infection (Fig. 5b,c). A significantly increased ability of S. typhimurium to disseminate to the mesenteric lymph nodes (40-fold, P = 0.001) and the spleens (42-fold, P = 0.009) was detected in Il17ra^−/− mice compared to C57BL/6 mice (Fig. 5c). These data provide direct evidence that the IL-17–IL-17RA axis limits bacterial translocation from the intestinal mucosa to systemic sites of infection.

We further investigated the host response elicited by S. typhimurium in Il17ra^−/− mice by measuring cytokine responses in the cecal mucosa by real-time PCR. S. typhimurium infection resulted in similar IFN-γ and IL-23 p19 mRNA levels in C57BL/6 mice as in Il17ra^−/− mice (Fig. 6). A compensatory increase in Il-17 mRNA levels was observed during S. typhimurium infection of Il17ra^−/− mice. Compared to infection of Il17ra^−/− mice, S. typhimurium infection of C57BL/6 mice elicited significantly increased mRNA levels of lipocalin-2 (eightfold, P = 0.003). These data suggest a cause-and-effect relationship between blunted Il-17 responses in SIV-positive macaques and the inability of these macaques to increase lipocalin-2 expression in response to S. typhimurium infection (Fig. 2). S. typhimurium elicited significantly higher mRNA levels of the keratinocyte-derived chemokine (Kc; threefold, P = 0.016), a murine neutrophil chemoattractant related to CXCL1, in C57BL/6 mice than in Il17ra^−/− mice (Fig. 6). Increased expression of Kc mRNA in the cecal mucosa was accompanied by an increased neutrophil influx detected in histological sections from the cecal mucosa of C57BL/6 mice compared to Il17ra^−/− mice. In summary, our data show that IL-17 is required for the full induction of responses that lead to neutrophil influx and the production of antimicrobials in the cecal mucosa during S. typhimurium infection.

Animal experiments

All experiments were approved by the Institutional Animal Care and Use Committee at the University of California, Davis. We used eight healthy, Salmonella-free, male rhesus macaques ranging from 2 to 4 years of age (MK3–10) for ligated ileal loop surgery. We isolated lamina propria T cells from all macaques undergoing loop surgery and two additional macaques (MK1 and MK2). We inoculated four macaques (MK3, MK4, MK6 and MK10) intravenously with previously titrated frozen stocks of SIV mac251. We determined plasma viral loads as previously described⁴⁹. Macaques were pre-anesthetized with ketamine (10 mg/kg; Parke-Davis), followed by placement of an endotracheal tube and maintenance of the anesthesia with isofluorane. When needed, we kept the macaques under a positive-pressure respirator. We performed a laparotomy, exposing the ileum and ligating 13 loops with an average of 4 cm in length, leaving 1-cm spacer loops in between. We inoculated the loops by intralumenal injections of 1 ml of either sterile LB broth or a logarithmically grown culture containing 1 × 10⁹ colony-forming units (CFU) of wild-type S. typhimurium (IR715)³⁰. We collected loops at 2, 5, or 8 h after inoculation.

We streptomycin-pretreated and orally infected groups of eight female C57BL/6 mice or six female Il17ra^−/− mice (B6.IL-17R mice, Taconic), aged 8–10 weeks, with S. typhimurium as described previously³⁰. At 48 h after infection, we killed the mice and collected the organs for mRNA isolation and bacteriology. A veterinary pathologist performed blinded histopathological evaluation of H&E-stained sections from the cecum.

Bacteriology

We collected two 6-mm biopsy punches from each loop. At the end of the surgery, we collected two biopsy punches (6-mm) from the mesenteric lymph node. We incubated biopsy punches for 1 h in PBS containing 50 mg/l of gentamicin, homogenized them, serially diluted the homogenate and plated on LB agar plates containing the appropriate antibiotics.

RNA analysis

We collected two 6-mm biopsy punches from each loop for RNA isolation, which provided comparable tissue samples across different areas of the intestinal mucosa (gut-associated lymphoid tissue in rhesus macaques was composed of randomly distributed individual lymphoid follicles). We performed RNA extraction from intestinal biopsy punches from macaques or organs collected from mice as previously described³⁰.

We monitored gene expression profiles in ileal tissues with newly developed rhesus macaque genome–specific high-density oligonucleotide microarrays (Affymetrix). The arrays contained probe sets representing over 20,000 rhesus macaque genes. To minimize the occurrence of false positives in the array data, we used a minimum twofold difference in mRNA levels (P value <0.05; 95% confidence) between control and experimental samples as a criterion for identifying a change in gene expression. We determined statistical confidence through analysis of at least 11 independent 25-nucleotide oligonucleotide probes for each gene in each sample. We performed real-time PCR as described previously³⁰ with primers listed in Supplementary Table 1 online.

Fluorescence microscopy

We stained rhesus ileal sections with rabbit polyclonal antibody to human IL-17 (Santa Cruz Biotechnology) and FITC-conjugated antibody to rabbit IgG. We mounted sections with Slow Fade with DAPI (Invitrogen) and captured images by confocal laser microscopy using LSM 5 and PASCAL software (Zeiss)²³.

Cell isolation, stimulation and flow cytometry

For each macaque, we collected one 10-cm segment of the terminal ileum at surgery. We incubated the tissue in 2.5 mM EDTA, washed it twice with PBS and digested with collagenase type IV (Sigma). After filtering the cell suspension over glass wool, we washed the cells and incubated them with or without 50 ng/ml PMA and 1,000 ng/ml ionomycin in the presence of 0.01 mg/ml brefeldin A for 6 h. We blocked the cells with 2% human γ-globulin (Sigma) and then stained with Amcyan live amine dye (Invitrogen). We used antibodies to CD3 (allophycocyanin (APC)-Cy7–labeled; SP34.2), CD4 (Pacific Blue–labeled; OKT4), CD8 (APC-Cy5.5-labeled; 3B5), γδ T cell receptor (phycoerythrin-labeled; 5A6.E9) and CD95 (phycoerythrin-Cy5–labeled; DX2), followed by intracellular staining with FITC-labeled antibody to IL-17 (eBio64CAP17) and APC-labeled antibody to IFN-γ (4SB3) using 2 × Perm/Fix solution (BD). We performed multicolor immunophenotyping on a modified LSRII with a minimum of 500,000 events collected.

Statistical analyses

We analyzed microarray data using model-based algorithms (dChip, http://biosun1.harvard.edu/complab/dchip) and t-tests. We performed biological analysis of microarray data with the Affymetrix NetAFFX web interface and the DAVID (http://david.abcc.ncifcrf.gov/) annotation tool. Statistically over-represented (P < 0.05) biological processes within sub-clusters were identified with EASE (http://david.abcc.ncifcrf.gov/). Fold changes in mRNA levels measured by real-time PCR and CFU numbers underwent logarithmic transformation, and percentage values underwent angular transformation before ANOVA followed by either Student’s t-test or the Student-Newman-Keuls test.