Type I interferon signaling restrains IL-10R+ colonic macrophages and dendritic cells and leads to more severe Salmonella colitis

Type I interferons (IFNα, IFNβ) are key regulators of innate and adaptive immunity, modulating the severity of both viral and bacterial infections. While type I IFN signaling leads to improved outcomes in viral infections, its role in bacterial infections is more contextual and depends on the specific pathogen and route of infection. Given the limited evidence on whether type I IFN signaling affects enteric bacterial pathogens, we investigated the role of this signaling pathway in Salmonella enterica serovar Typhimurium (S. typhimurium)–induced colitis. Comparing mice deficient in IFNAR1- the common receptor for IFNα and IFNβ- with wild-type mice, we found that type I IFN signaling leads to more rapid death, more severe colonic inflammation, higher serum levels of pro-inflammatory cytokines, and greater bacterial dissemination. Specific ablation of plasmacytoid dendritic cells (pDCs), which are prominent producers of type I IFNs in antiviral responses, did not alter survival after infection. This result established that pDCs do not play a major role in the pathogenesis of S. typhimurium colitis. Flow cytometric analysis of macrophages and conventional dendritic cells (cDCs) during active colitis demonstrated an increase in CD11c- macrophages and CD103+ cDCs in the colon of Ifnar1-/- animals. Interestingly, cells expressing the anti-inflammatory cytokine receptor IL-10R are more abundant within these subsets in Ifnar1-/- than in wild-type mice. Moreover, blockade of IL-10R in Ifnar1-/- mice increased their susceptibility to S. typhimurium colitis, suggesting that altered numbers of these immunoregulatory cells may underlie the difference in disease severity. This cross-talk between type I IFN and IL-10R signaling pathways may represent a key host cellular mechanism to investigate further in order to unravel the balance between pathogenic inflammation and homeostasis of the colon. Taken together, our data clearly demonstrate that type I IFN signaling is pathogenic in S. typhimurium colitis.

Introduction Pathogen-induced inflammation is a natural response that is beneficial to the host as an initial mechanism of defense. However, excessive inflammation caused by an overwhelming immune response can lead to severe damage to the host, including organ dysfunction, septic shock, and death [1]. Thus, the balance of the inflammatory response must be tightly regulated to avoid excessive damage to the host. Type I interferon (IFN) encompasses a class of cytokines that play a key role in the balance between immunity and pathogenesis. Type I IFN's causal role in pathology has been demonstrated in the setting of various autoimmune diseases, such as lupus and multiple sclerosis [2,3]. Both of the major type I IFNs, α and β, signal through a common receptor, IFNAR1. Studies in mice deficient in IFNAR1 have elucidated both the immunityenhancing and pathogenic roles of type I IFN.
The role of type I IFNs in modulating infection has been most extensively studied in the context of viral infections, in which they promote immunity. While there is burgeoning evidence that type I IFNs also modulate the severity of bacterial infections, this effect appears to depend on context [4]. For example, type I IFNs are protective against infections with many extracellular bacteria (e.g., Streptococcus species, Pseudomonas aeruginosa) but are detrimental during infection with intracellular bacteria (e.g., Francisella tularensis, Brucella abortus) [5,6]. Moreover, the route of infection is important: in mice, type I IFN signaling during Listeria infections is pathogenic in parenteral infections but beneficial in enteric ones, the latter mimicking the natural route of infection for this pathogen in humans [7]. An important proof-ofconcept study recently demonstrated that type I IFN signaling is detrimental in Salmonella infection; however, this study used non-physiologic routes of infection-i.e., intravenous and intraperitoneal injection [8]. This study showed that during Salmonella enterica serovar Typhimurium (S. typhimurium) infection in vitro, IFN signaling induces necroptosis in infected macrophages; therefore, Ifnar1 -/macrophages are resistant to S. typhimurium-induced cell death. Murine infections with S. typhimurium can model either a typhoid-like illness with prominent systemic dissemination (with no antibiotic pre-treatment) or primarily gastroenteritis combined with systemic dissemination (after pre-treatment with streptomycin) [9]. This facultative intracellular bacterium causes an estimated 90 million cases of gastroenteritis per year [10], displaying an intricate interplay with the human immune system [11]. Using the typhoid model of Salmonella infection, investigators have recently shown that IFNβ promotes bacterial dissemination-an observation consistent with the idea that type I IFN signaling is detrimental under these conditions [12]. Given the discordant results for systemic and enteric infections with Listeria monocytogenes, we wanted to determine what role type I IFN signaling plays in Salmonella-induced colitis. Moreover, we wanted to explore the relevant immune-cell subsets involved in type I IFN signaling during Salmonella infection, about which only few details are known.
In this study, we found that type I IFN signaling is pathogenic in S. typhimurium-induced colitis, with wild-type (WT) mice dying faster than Ifnar1 -/mice. This greater susceptibility to mortality in WT mice was associated with increased dissemination of bacteria, increased serum levels of pro-inflammatory cytokines, and increased colonic inflammation. Numbers of CD11ccolonic macrophages, which may be similar to CD11c lo macrophages that reside in the muscularis mucosa region and regulate gut motility [13], as well as numbers of CD103 + dendritic cells, which play an important role in inflammatory diseases and antigen carriage from intestine to mesenteric lymph nodes (MLNs) [14], were higher in infected Ifnar1 -/mice than in infected WT mice. Notably, these cell types expressing the immunoregulatory interleukin 10 receptor (IL-10R) were similarly increased in Ifnar1 -/mice. These results suggest that type I IFN signaling leads to a loss of these cell types that is associated with worse overall outcomes. Indeed, blockade of IL-10R in Ifnar1 -/mice resulted in increased susceptibility to S. typhimurium colitis, eliminating the difference in mortality compared to WT mice. Thus, type I IFN signaling is pathogenic in Salmonella colitis, and its pathogenicity is likely to be a result of decreased numbers of immunoregulatory macrophages and dendritic cells.

S. typhimurium-induced infectious colitis
Streptomycin sulfate in PBS (20 mg/mouse, Sigma-Aldrich) was administered to mice via oral gavage. At 24 hours after treatment, 10 2 to 10 7 CFU of S. typhimurium strain SL1344 (grown overnight at 37˚C in LB Broth, Miller (Luria-Bertoni) with added streptomycin, washed and diluted in PBS) was administered to each mouse via oral gavage. A dose of 10 4 CFU was selected following dose response studies. For survival studies, body mass was measured daily and mice were observed for symptoms of pain or distress as per our institutional animal protocol until the end of the experiment. Mice were euthanized immediately and reported as dead once body condition score of less than 2 out of 5 was observed or they appeared moribund. Approximately less than 5% of animals were found dead beyond supervision and were considered to have died on the day of observation for the purpose of data analysis. For flow cytometric analysis, serum cytokine quantification, bacterial burden assessment, and histopathology studies, mice were euthanized on days 1, 3 or 5 after infection.

Cell isolation from primary tissues
For single-cell suspensions of mesenteric lymph nodes (MLNs), the first five MLNs (starting from the cecum) were treated with collagenase type IV (1 mg/ml; Sigma) for 30 minutes at 37˚C in an atmosphere of 5% CO 2 and passed through a 70-μm mesh. Single cells were isolated from colonic tissue as previously described [15]. In brief, the tissue was first cut transversely into small pieces and then cut longitudinally to expose the lumen. Pieces were washed in 1 mM dithiothreitol (Sigma) in PBS for 10 minutes and then in PBS. Three washes of 8 minutes each in 30 mM EDTA in PBS were then employed to strip off epithelial cells. After one more wash in PBS, tissue was treated with RPMI medium containing 5% fetal bovine serum (FBS) and collagenase type IV (1 mg/ml) for 1 hour at 37˚C in 5% CO 2 . Finally, tissue sections were passed through mesh (70-μm) to yield single-cell suspensions.

Serum cytokine quantification
Mice were treated with streptomycin and infected with 10 4 CFU of S. typhimurium strain SL1344 as described above. On day 5, mice were euthanized, and blood was collected into 1.1-ml Z-Gel Micro Tubes (Sarstedt). Blood was left at room temperature for 30 minutes before the coagulated cells were spun out (10,000 g, 10 minutes), and serum was stored at -80˚C. Levels of tumor necrosis factor α (TNFα) and interleukin-6 (IL-6) were quantified with ELISA Duoset kits (R&D Systems, Minneapolis, MN) according to the manufacturer's protocol.

Bacterial burden in major organs
For determination of bacterial burden, mice were euthanized on day 1, 3, or 5 after infection (10 4 CFU), and specific tissues (MLNs, liver, and spleen) were collected. Organs were weighed, hand-mashed, and homogenized with a Stomacher 80-paddle action blender (Seward, Port Saint Lucie, FL). Serial 10-fold dilutions were prepared with sterile PBS supplemented with 2% FBS. A 10-μl volume of each dilution was plated onto LB Strep plates and incubated at 37˚C in 5% CO 2 for 24 hours. Results are expressed as total CFU per tissue.

Histopathology
Cecum, and colon tissues were fixed and stored in Bouin's solution (VWR Scientific, West Chester, PA). Fixed tissues were embedded in paraffin, sectioned, mounted onto slides, and stained with hematoxylin and eosin. Sections were evaluated in blinded fashion by a single pathologist (Dr. R. T. Bronson, Harvard Medical School); samples were scored from 0 to 4 for degree of observable inflammation.

Statistical analysis
Statistical analysis was performed using the standard appropriate tests. For survival studies the log-rank test was used to compare differences between the survival of two groups. All other experiments were statistically analyzed with the Mann-Whitney non-parametric test.

IFNAR1 signaling contributes to mortality in S. typhimurium-induced colitis
To assess the effect of type I IFN signaling on susceptibility to S. typhimurium infection after streptomycin treatment, we initially investigated the survival of WT and Ifnar1 -/animals after infection with various doses of S. typhimurium. When we used inocula of S. typhimurium (10 2 and 10 4 CFU) that approximate the infectious dose in humans [16], WT mice died significantly faster than Ifnar1 -/mice (Fig 1A and 1B, respectively). Interestingly, when we used a much larger inoculum (10 7 CFU), there was no difference in survival between the groups ( Fig  1C). This finding suggests that, with a higher infectious burden, another pathway overwhelms the effects of IFNAR1 signaling. These results demonstrate that at lower, clinically relevant doses of Salmonella, IFNAR1 signaling is pathogenic in Salmonella-induced colitis.

Colonic and systemic inflammation are augmented by IFNAR1 signaling
To assess the effects of IFNAR1 signaling on S. typhimurium-induced colonic inflammation [9,17], we examined and scored the histopathological condition of the colon and cecum of WT and Ifnar1 -/mice on days 3 and 5 after infection. The cecum was highly inflamed on both day 3 and day 5, with no difference between WT and Ifnar1 -/mice (Fig 2A and 2B, respectively).
In contrast, although no difference in colonic inflammation was observed on day 3 (Fig 2C), on day 5 WT mice had significantly greater inflammation (as manifested by edema, immune cell infiltrations, and loss of normal colonic architecture) than did Ifnar1 -/mice (Fig 2D and  2E).
To obtain a better understanding of the dynamics of bacterial dissemination during Salmonella gastroenteritis, we examined bacterial burden in several major organs throughout the course of infection. The bacterial burden in the MLNs, liver, and spleen was significantly Type I interferon signaling in Salmonella infection higher in WT than in Ifnar1 -/animals on days 3 and 5 after infection (Fig 3B and 3C, respectively). In contrast, bacterial dissemination to these organs was not different on day 1 after infection; thus, these groups of mice had a similar disease burden at an early time point ( Fig  3A). To gain insight into the systemic response to infection, we measured serum levels of the pro-inflammatory cytokines TNFα and IL-6 on day 5 after infection. Consistent with the greater degree of bacterial dissemination in WT mice, we found that WT mice had higher serum levels of TNFα and IL-6 than did Ifnar1 -/animals ( Fig 3D). Overall, these studies demonstrate that WT mice have more severe disease than do Ifnar1 -/mice, as evidenced by greater colonic inflammation, systemic dissemination of bacteria, and faster death.
To assess whether there were differences in bacterial colonization that were driving the differential outcome of the disease, we measured bacterial CFU in the feces of WT and Ifnar1 -/mice on day 1 after infection. As in distal organs, where we observed no significant difference in bacterial numbers on day 1 (Fig 3C), we found that bacterial numbers were overall similar in Ifnar1 -/and WT feces at this early time point (S1 Fig). These results suggest that better overall disease outcome could not be attributed to lower-level initial colonization in Ifnar1 -/mice.

pDCs do not play a significant role in pathogenesis during S. typhimurium infection
Having demonstrated that type I IFN signaling is pathogenic in Salmonella infections, we wanted to define which immune cells contribute to this phenotype. pDCs are an important source of type I IFNs and a major initial source of IFNα: their early removal during viral infection delays antiviral immunity while in a murine model of lupus similar removal reduces the disease [2,18]. In addition, pDCs are involved in the immune response to bacterial pathogens, including both extracellular and intracellular organisms [19][20][21]. Given their contribution to other viral and bacterial infections and their involvement in other autoimmune and inflammatory diseases, we examined whether pDCs play a role in S. typhimurium infection. By administering diphtheria toxin (DT) to the well-defined BDCA2-DTR mice, in which pDCs are the only cells that express the diphtheria toxin receptor (DTR), we were able to specifically deplete pDCs. Depletion of pDCs (!95%) in uninfected mice using this system has previously been demonstrated [15]. We confirmed depletion of pDCs in the spleen and MLN of Salmonellainfected mice using flow cytometric analysis (S2 Fig). As a control, we first confirmed that diphtheria toxin administration did not influence mortality in Salmonella-infected mice that do not express DTR (Fig 4A). Similarly, depletion of pDCs had no effect on mortality during S. typhimurium infection (Fig 4B). Thus, the absence of pDCs does not affect the outcome of Salmonella colitis.
Salmonella-infected Ifnar1 -/mice have more CD11cmacrophages and CD103 + DCs than WT mice Since pDCs are dispensable for disease severity, we explored the role of other immune-cell subsets by enumerating innate immune cells in the colon and MLNs on day 3 after infection. We focused our analysis on macrophages (CD45 + MHCII + CD11b + F4/80 + CD103 -) and conventional DCs (cDCs; CD45 + MHCII + CD11c + F4/80 -) (Fig 5A), which are major producers of type I IFN [22][23][24]. Indeed, as indicated previously, macrophages at other sites have been demonstrated to play a major role in Salmonella infection [25]. Compared with naïve animals, infected mice had significantly higher total colonic macrophage numbers, with no difference between WT and Ifnar1 -/mice (Fig 5B, left panel). CD11c expression by colonic macrophages helps distinguish cell subsets that are functionally and anatomically distinct; CD11c + macrophages are present in the lamina propria and CD11cmay be like CD11c lo macrophages in the muscularis mucosa region or draining the tissue during inflammation [13]. While the CD11c + macrophage population did not differ between infected WT and infected Ifnar1 -/mice, CD11cmacrophages (which represent the majority of colonic macrophages) were more numerous in the colon of Ifnar1 -/mice (Fig 5B, center and right panels). Similar to our findings with colonic macrophages, total numbers of colonic cDCs did not differ between infected WT mice and infected Ifnar1 -/mice (Fig 5C). Intestinal cDCs can be distinguished by CD103 expression; CD103 + cDCs have important roles in carrying antigens to MLNs and regulating the inflammatory response [14]. Interestingly, there was a marked increase in the colonic CD103 + cDC population in Ifnar1 -/mice over that in WT mice, while the two groups did not differ in numbers of CD103 -cDCs (Fig 5C).
Given that the MLN is an important draining lymph node in enteric Salmonella infection and is likely to represent the gateway to systemic dissemination, we enumerated macrophages and cDCs in this compartment. The results were quite different from those in the colon. Increase in tolerogenic IL-10R + colonic antigen-presenting cells (APCs) in the absence of IFNAR1 signaling Colonic macrophages expressing the receptor of anti-inflammatory cytokine IL-10 (IL-10R) have anti-inflammatory properties and protect against colitis [26,27]. Given that numbers of Type I interferon signaling in Salmonella infection specific macrophage and dendritic cell subsets were increased in Ifnar1 -/mice, which have less severe disease than WT animals, we investigated whether these cell types correspond to tolerogenic subsets marked by IL-10R. Indeed, while the number of IL-10R + CD11c + macrophages increased similarly in infected WT and Ifnar1 -/mice, Salmonella-infected Ifnar1 -/mice had more IL-10R + CD11cmacrophages than did infected WT mice (Fig 6A). Similarly, Salmonella-infected Ifnar1 -/mice had more IL-10R + CD103 + cDCs than infected WT mice because of a decrease in this cell type in the setting of type I IFN signaling (Fig 6B). The number of total IL-10R + cDCs, CD103 + IL-10R + cDCs, and CD103 -IL-10R + cDCs were enumerated in WT and Ifnar1 -/mice. Results represent the mean with SD. Statistical analysis was performed using the Mann-Whitney test. *p<0.05, **p<0.005. (C) Survival curve of WT and Ifnar1 -/mice treated with anti-IL-10R mAb (anti-IL-10R) or isotype control antibody (isotype) on days -3, 0, and +3 and infected with 10 4 CFU Salmonella. Statistical analysis was performed using a log-rank test. ns, not significant; *p<0.05. https://doi.org/10.1371/journal.pone.0188600.g006 Type I interferon signaling in Salmonella infection

Blockade of IL-10R increases susceptibility of Ifnar1-/-mice to S. typhimurium
Based on the observed increase in IL-10R expressing CD11cmacrophages and CD103 + cDCs in S. typhimurium infected Ifnar1 -/mice compared to WT mice, we sought to investigate the role of IL-10R in the reduced susceptibility of these mice to S. typhimurium colitis. Blockade of IL-10R in Ifnar1 -/mice significantly reduced the time to death of these mice compared to isotype control treated Ifnar1 -/mice ( Fig 6C). Notably, with blockade of IL-10R, the difference in susceptibility of WT and Ifnar1 -/mice is no longer observed (Fig 6C). Taken together, these data suggest that type I IFN signaling may lead to more severe disease in Salmonella infections by downregulating these immunoregulatory, IL-10R expressing APC subsets.

Discussion
Infection with S. typhimurium elicits a strong inflammatory response that benefits the bacteria by creating an environment that provides a growth advantage for Salmonella over members of the microbiota [11]. Therefore, the ability of the host to actively limit an excessive inflammatory response can be advantageous to the host even in the context of bacterial infection. Here, we demonstrate that type I IFN signaling is pathogenic in the murine model of Salmonella gastroenteritis, as evidenced by shorter time to death, greater bacterial dissemination, higher serum levels of pro-inflammatory cytokines, and greater colonic inflammation.
In light of the differences in the degree of colonic inflammation between WT and Ifnar1 -/animals, we characterized prominent colonic APCs, focusing on cell types known to produce type I IFN. We found that numbers of CD11cmacrophages but not CD11c + macrophages differed in the two groups, with higher counts in Ifnar1 -/mice. Since these macrophage subsets have different anatomic localizations and distinct functional characteristics [13], CD11cmacrophages whose numbers increase in the absence of IFNAR1 signaling may possibly contribute to the tolerogenic phenotype observed in Ifnar1 -/mice. Interestingly, Ifnar1 -/mice also had an increased number of CD11cmacrophages expressing IL-10R, a recently described immunoregulatory subset of macrophage. IL-10R + colonic macrophages are crucial in preventing colitis [26,27], and IL-10R mutations in humans are associated with the early onset of inflammatory bowel disease [28]. In addition, lack of IL10R mediated signaling has recently been demonstrated to change macrophage programming to an activated inflammasome pathway driven by mTORC1 [29]. Further investigation will be required to determine whether blocking IL10R in Ifnar1 -/mice causes earlier death mediated by mTORC1/inflammasome activation. Taken together, these findings suggest that IL-10R + macrophages are important for maintenance of colonic homeostasis and prevention of excessive intestinal inflammation.
Our results demonstrate that IFNAR1 signaling results in a decreased number of IL-10R + macrophages, although the mechanism by which these cell numbers are modulated is unclear. One possibility is that signaling by type I IFNs leads directly to a reduction of these immunoregulatory macrophages. Given recent data demonstrating that IL-10 levels after intraperitoneal infection with Salmonella are diminished in the absence of IFNβ [12], an alternative, mutually non-exclusive possibility is that IL-10 levels are decreased in Salmonella-infected Ifnar1 -/mice with a compensatory increase in the expression of IL-10R. Zigmond et al. found that IL-10R on macrophages-not IL-10 secreted by macrophages-was critical for the antiinflammatory properties of this macrophage subset [27], suggesting that it is the altered number of IL-10R macrophages that matters, irrespective of the underlying mechanism.
Although pDCs are involved in regulating immunity to viral and some bacterial infections, we found that they do not contribute significantly to the pathogenesis of Salmonella infection. Rather, we found that numbers of colonic CD103 + cDCs, which are known to be important for uptake and presentation of bacterial antigens during S. typhimurium infection [30,31], were lower in WT mice than in Ifnar1 -/mice. Moreover, CD103 + IL-10R + cDC numbers were lower in infected WT mice than in uninfected mice, with no change in Ifnar1 -/mice. These results suggest that Salmonella-induced IFNAR1 signaling leads to the decrease in CD103 + IL-10R + cDCs. While IL-10R expression in macrophages is crucial for immune tolerance, the connection between IL-10R expression and DC function has not yet been fully addressed. On the basis of our demonstration of an association between a numerical decrease in these cells and worse outcomes, it is intriguing to speculate that IL-10R + cDCs are similarly involved in immunoregulation.
In summary, we have demonstrated that type I interferon signaling is detrimental to the host in a model of Salmonella-induced colitis. Furthermore, CD11c -IL-10R + colonic macrophages are more abundant in Ifnar1 -/mice than in WT mice, and blockade of IL-10R eliminated the mortality differences between WT and Ifnar1 -/mice. These results suggest that these macrophages may not only play a role in maintaining intestinal homeostasis but may also have anti-inflammatory functions during severe bacterial infection. Future investigation of how these different signaling pathways interact with each other during invasion by a mucosal pathogen would lead to a better understanding of immune defense and might contribute to the design of novel approaches to mucosal vaccination for infectious diseases of the GI tract.