The sst1 Resistance Locus Regulates Evasion of Type I Interferon Signaling by Chlamydia pneumoniae as a Disease Tolerance Mechanism

The sst1, “supersusceptibility to tuberculosis,” locus has previously been shown to be a genetic determinant of host resistance to infection with the intracellular pathogen, Mycobacterium tuberculosis. Chlamydia pneumoniae is an obligate intracellular bacterium associated with community acquired pneumonia, and chronic infection with C. pneumoniae has been linked to asthma and atherosclerosis. C. pneumoniae is a highly adapted pathogen that can productively infect macrophages and inhibit host cell apoptosis. Here we examined the role of sst1 in regulating the host response to infection with C. pneumoniae. Although mice carrying the sst1 susceptible (sst1S) locus were not impaired in their ability to clear the acute infection, they were dramatically less tolerant of the induced immune response, displaying higher clinical scores, more severe lung inflammation, exaggerated macrophage and neutrophil influx, and the development of fibrosis compared to wild type mice. This correlated with increased activated caspase-3 in the lungs of infected sst1S mice. Infection of sst1S macrophages with C. pneumoniae resulted in a shift in the secreted cytokine profile towards enhanced production of interferon-β and interleukin-10, and induced apoptotic cell death, which was dependent on secretion of interferon-β. Intriguingly macrophages from the sst1S mice failed to support normal chlamydial growth, resulting in arrested development and failure of the organism to complete its infectious cycle. We conclude that the sst1 locus regulates a shared macrophage-mediated innate defense mechanism against diverse intracellular bacterial pathogens. Its susceptibility allele leads to upregulation of type I interferon pathway, which, in the context of C. pneumoniae, results in decreased tolerance, but not resistance, to the infection. Further dissection of the relationship between type I interferons and host tolerance during infection with intracellular pathogens may provide identification of biomarkers and novel therapeutic targets.


Introduction
The susceptibility of a host to infection, as well as the immunologic response to a given infection, is heterogeneous within a population. While epidemiologic studies may suggest inherited factors can influence immunity to infectious diseases, it is clear that resistance to infection is a complex, multifactorial genetic trait in which many genetic polymorphisms contribute to the phenotype. Thus, the unique host-pathogen interaction of an individual that determines the outcome of infection is under multigenic control. Inbred strains of mice are well suited for studying the genetic control of infectious diseases as gene effects may have become fixed in inbred, recombinant inbred, and recombinant congenic strains of mice (reviewed in [1,2]). For example, using an experimental mouse model for Mycobacterium tuberculosis infection in C57BL/6 mice (inherently resistant to M. tuberculosis infection) and C3HeB/FeJ mice (inherently susceptible to M. tuberculosis), we previously identified a locus on chromosome 1 that has a significant effect on the development of disease following M. tuberculosis infection, particularly within the lung. We designated this as the sst1 locus, for ''super susceptibility to tuberculosis'', and demonstrated that the presence of the C57BL/ 6 sst1 locus imparts resistance to M. tuberculosis infection when expressed on the C3HeB/FeJ background [3], while the C3HeB/ FeJ sst1 locus imparts susceptibility to M. tuberculosis when expressed on the C57BL/6 background [4]. Thus, C57BL/6 mice are deemed ''resistant'' at the sst1 locus (or sst1 R ), while C3HeB/FeJ mice are ''susceptible'' (or sst1 S ).
The sst1 locus appears to participate in control of intracellular multiplication of virulent M. tuberculosis within the macrophage, and has an effect on the infected macrophages' mechanism of cell death. However, the effects of the sst1 locus are not specific to M.
tuberculosis as it has also been shown to be important in infections due to another bacterial species with a significant intracellular phase, Listeria monocytogenes [5]. Subsequent studies identified a candidate gene, Intracellular Pathogen Resistance 1 (Ipr1), within the sst1 locus as playing a role in controlling innate immunity and cell death in response to both M. tuberculosis and Listeria [6]. The closest human homologue to Ipr1 is Sp110, and mutations in Sp110 have been associated with hepatic veno-occlusive disease with immunodeficiency (VODI) [7,8]; however, the role of Sp110 in human tuberculosis is still being debated [9,10,11]. The identification of additional pathogens that are controlled by genes within the sst1 locus may shed light on the sst1/Sp110-mediated mechanism of innate host resistance to divergent intracellular pathogens, as well as their shared pathogenesis and immune evasion strategies.
Chlamydia pneumoniae is an obligate intracellular pathogen and the etiologic agent of atypical, community acquired pneumonia. While respiratory infection is often subclinical or mildly symptomatic, chronic infection with C. pneumoniae has been suggested as a trigger or promoter of a number of inflammatory conditions, including asthma (reviewed in [12,13]), chronic obstructive pulmonary disease (COPD) [14], and atherosclerosis (reviewed in [15]). The mechanism by which C. pneumoniae can disseminate from the lung remains unproven, but one leading model suggests that replication within lung macrophages provides a unique niche for transport to distant sites, such as chronic vascular lesions (reviewed in [15]), and case reports have identified C. pneumoniae within atherosclerotic lesions [16,17].
Like all Chlamydia species, C. pneumoniae has a unique dimorphic developmental cycle that sets it apart from other bacterial species. There are two major developmental forms that are recognized: the infectious, but metabolically inert form, known as the elementary body (EB); and the intracellular, replicative form known as the reticulate body (RB). The developmental cycle begins with cellular attachment and entry of the EB, which then converts into the RB form and begins to replicate by binary fission. These RBs grow and divide within a unique membrane-bound cellular inclusion, and at some point mid to late in the infectious process, RBs differentiate back to EBs, which are eventually released coincident with cell lysis into the extracellular space, around 40 to 72 hours post infection, depending on the species of Chlamydia. Then, a new round of infection and development begins. While this acute or lytic lifestyle has been extensively studied, Chlamydia species can also enter a less-understood persistent state, defined as the presence of large, irregular, noncultivatable developmental forms that fail to mature; these have been referred to as aberrant bodies or aberrant reticulate bodies (reviewed in [18]). In vitro, a number of seemingly unrelated triggers have been shown to induce persistence, including antibiotic exposure [19], nutritional deprivation (e.g., amino acid deficiency [20], iron limitation [21]), and exposure to interferon (IFN)-c, which is known to upregulate cellular indoleamine 2,3dioxygenase (IDO) and thus deplete intracellular tryptophan stores [22]. The role of chlamydial persistence and these aberrant developmental forms in the pathogenesis of any of the clinical syndromes associated with chlamydia infections remains unclear.
Given the limited innate immune evasion strategies of an obligate intracellular pathogen, and the similarities between the lifestyle of Mycobacteria and Chlamydia, we examined the hostpathogen interaction between C. pneumoniae and eukaryotic macrophages in the context of the sst1 locus with the goal of identifying key pathways involved in pathogenesis. We observed that the sst1 locus regulates the innate immune response of macrophages to C. pneumoniae, altering the cytokine balance towards IL-10 and IFN-b, and overriding the chlamydial antiapoptotic program, which resulted in arrested bacterial development. Paradoxically, this translated in vivo into a more severe clinical course in a pneumonia model, with enhanced apoptosis and fibrosis within lung tissue of infected mice, in spite of their ability to control bacterial replication. Our data demonstrate that sst1 is a key host susceptibility locus for intracellular pathogens, regulating inflammation and cell death in response to infection.

Results
The sst1 locus controls the innate immune response to lung infection with C. pneumoniae To examine the effect of the sst1 locus on susceptibility of mice to C. pneumoniae (Cp) infection, we compared standard inbred C57BL/6 (B6) mice, which are resistant at the sst1 locus, and sst1 susceptible congenic mice, B6.C3H-sst1. The B6.C3H-sst1 strain is 99.5% genetically identical to the resistant B6 strain with the exception of a 12-cM interval on mouse chromosome 1 encompassing the sst1 locus, which has been introgressed from the susceptible C3HeB/FeJ strain [23]. Mice were infected with C. pneumoniae as described in the Methods section, and their clinical state was scored on a scale of 0 to 4, with a score of 0 for mice that did not appear clinically ill and a score of 4 for mice that were moribund and required euthanasia. As shown in Table 1, B6.C3H-sst1 mice overall displayed a higher clinical score compared to wild type B6 mice, with a statistically significant difference between the mean scores of 1.64 in B6 mice vs. 2.60 in the B6.C3H-sst1 mice (p = 0.0014). Moreover, 15 out of 25 mice in the B6.C3H-sst1 group reached a score of 3-4, corresponding to severe illness or moribund state, compared to only 5 out of 25 mice in the B6 group (Table S1). Although both groups experienced weight loss as a result of infection, there was no statistically significant difference between the B6 and B6.C3H-sst1 mice over the first week of infection ( Figure S1A). Quantitative culture of lung homogenates similarly revealed no statistically significant differences in clearance of the bacterial load in the lungs between the B6 and B6.C3H-sst1 mice ( Figure S1B), nor was there any

Author Summary
Chlamydia pneumoniae is a highly adapted intracellular pathogen and a common cause of atypical, community acquired pneumonia. It has also been suggested as a trigger or promoter of asthma and atherosclerosis. In this study, we examined the role of a genetic locus on mouse chromosome 1 that has been associated with susceptibility to another intracellular pathogen, Mycobacterium tuberculosis, in the pathogenesis of respiratory infections secondary to Chlamydia pneumoniae. We have determined that a variant at this locus, known as sst1 and associated with destructive pulmonary tuberculosis, makes mice dramatically more sensitive in vivo to the inflammatory changes following respiratory infection with C. pneumoniae. This appears to arise from activation of type I interferons and apoptotic cell death, two signaling pathways that are normally silent during productive C. pneumoniae infection. Despite a noted inability of sst1 susceptible macrophages to support chlamydial development, exuberant lung tissue damage resulted in overall more severe disease in vivo. We conclude the sst1-mediated control of lung tissue damage is an important determinant of the genetic susceptibility of a given host to a number of diverse intracellular bacterial pathogens, which may provide predictors of outcomes to infectious diseases as well as possible target for novel therapeutics.
difference in dissemination of bacteria to the spleen (data not shown). This suggested clinically more severe disease in the infected B6.C3H-sst1 mice that could not be explained by impaired control of bacterial replication.
When histological sections of the lung were examined from infected mice, we observed qualitative differences in the inflammatory response. Infected B6 mice displayed patchy, focal areas of inflammation, while the B6.C3H-sst1 mice developed more diffuse, extensive inflammation. This was true on both day 3 ( . Furthermore, we observed increased immunostaining for vimentin in the lung parenchyma of the B6.C3H-sst1 mice as early as day 3, along with enhanced collagen deposition consistent with fibrosis, as detected by Masson's trichrome staining, and increased staining for thyroid transcription factor-1 (TTF-1), a nuclear protein expressed in type II pneumocytes ( Figure S2). These data are consistent with increased tissue damage and repair in response to infection in the B6.C3H-sst1 mice.
When we examined lung homogenates for the presence of inflammatory mediators (Figure 2), we observed a bias towards an anti-inflammatory response in the B6.C3H-sst1 mice, with exaggerated IL-10, a classic anti-inflammatory cytokine, and IL-6, which can have both pro-and anti-inflammatory activity [24]. Although there was a trend towards higher IFN-b in the infected B6.C3H-sst1 mice, it did not reach statistical significance. The B6.C3H-sst1 mice also displayed higher levels of MCP-1 (also known as CCL-2) in the lungs, consistent with the exaggerated influx of monocytes, while the IFN-c response was elevated but not significantly different between the two infected mouse strains. Together, these data demonstrate a dysregulated inflammatory and fibrogenic response in Cp-infected B6.C3H-sst1 mice, which carry the sst1 susceptible locus.
The sst1 locus regulates IFN-b and IL-10 production in C. pneumoniae-infected macrophages Type I IFNs are known to play a critical role in the immune response to viruses and intracellular bacteria, but they have also been shown to possess anti-inflammatory effects as well (reviewed in [25,26]). For example, IFN-b was recently reported to inhibit IL-1b production and inflammasome activation [27]. To better understand the sst1 regulated anti-inflammatory cytokine response, we prepared bone marrow derived macrophages (BMDM) from B6.C3H-sst1 mice and wild type B6 mice, and infected them in vitro with Cp in the presence or absence of IFN-c priming. As shown in Figure 3 A-D, we observed significantly higher levels of IL-6 and IL-10, with lower IL-1b production, in IFN-c primed B6.C3H-sst1 macrophages compared to wild type B6 macrophages, similar to what we observed in vivo. The differences were not as significant in the unprimed macrophages. The most striking effect, however, was observed in type I interferon response. As shown in Figure 3C, wild type B6 BMDM secrete barely detectable levels of IFN-b in response to Cp infection regardless of IFN-c priming. In contrast, Cp infection of IFN-c primed B6.C3H-sst1 BMDM resulted in a three to fourfold increase in IFN-b induction compared to B6 BMDM ( Figure 3C).
It has been shown that type I IFN induction is required for LPSinduced IL-10 production [28]. To investigate whether IFN-b and IL-10 induction was linked in the response of B6.C3H-sst1 BMDM to Cp infection, we performed IFN-b and IL-10 neutralization assays. As shown in Figure 3 E-F, neutralization of IL-10 significantly enhanced IFN-b production while neutralization of IFN-b resulted in a modest but not statistically significant reduction in IL-10 production, which might reflect residual signaling by other type I IFNs. Thus, while IFN-b and other type I IFNs may be upstream of IL-10, IL-10 itself appears to negatively regulate IFN-b in B6.C3H-sst1 BMDM.
To further examine the mechanism by which Cp infection of B6.C3H-sst1 BMDM resulted in the upregulation of IL-10 and IFN-b, we examined the effect of specific pathway inhibitors on this response. Two related IkB kinase homologues, IkB kinase-j (IKKj) [29,30] and TANK-binding kinase 1 (TBK1) [31,32], are essential components of the IRF3 signaling pathway upstream of type I interferon induction [33]. We observed that the pharmacological inhibitor, BX795, shown to be a potent inhibitor of IKKj and TBK1 [34], was capable of inhibiting Cp-induced secretion of both IL-10 and IFN-b to background levels in B6.C3H-sst1 BMDM in a dose-dependent manner (Figure 4 A-B). Similarly, 2-aminopurine (2-AP), a potent inhibitor of doublestranded RNA (dsRNA)-activated protein kinase (PKR) [35] that has also been shown to inhibit nuclear translocation of phosphorylated-IRF-3 [36], had a modest effect on IFN-b but only at the highest concentration used without significantly altering IL-10. In contrast, inhibition of the TLR adaptor MyD88, which is shared by all the TLRs except TLR3, had no effect on IL-10 or IFN-b induction in response to Cp (Figure 4 C-D). These data suggest that induction of type I IFNs in response to Cp infection of the B6.C3H-sst1 BMDM occurs in an IRF3 dependent manner, upstream of IKKj and TBK1, but independent of MyD88. Partial dependence on PKR suggests involvement of a cytosolic receptor although we cannot rule out involvement of endosomal TLR3.

The sst1 locus regulates chlamydial development in macrophages
Chlamydiae have a unique dimorphic developmental cycle, with vegetative EBs converting to replicative RBs, and back again. In addition, a third developmental form is recognized, the ''aberrant body'', which appears to be an arrested RB which takes on an irregular shape and stops dividing. A number of triggers have been shown to induce the production of aberrant bodies, including IFN-c, iron or tryptophan starvation, and exposure to penicillins (reviewed in [18]). In order to determine if the sst1 locus had any impact on Cp replication or development within macrophages, we infected BMDM from B6 and B6.C3H-sst1 mice with Cp in the absence or presence of IFN-c, and quantified the burst of recovered EBs over time. As shown in Figure 5A, we observed significantly reduced recovery of EBs from Cp grown in B6.C3H-sst1 BMDM compared to B6 BMDM. Furthermore, while IFN-c priming reduced the recovery of viable organisms from B6 BMDM by over 1 log, in the case of B6.C3H-sst1 BMDM we were unable to recover any viable EBs at the end of the culture period ( Figure 5B). Similarly, we observed significantly fewer inclusions in Cp-infected B6.C3H-sst1 BMDM compared to wild type B6 BMDM ( Figure S3A). This defect in Cp development in macrophages was confirmed by electron microscopy. As shown in Figure 5C, Cp development within B6.C3H-sst1 BMDM resulted in smaller inclusions with fewer intracellular bacteria and more aberrant appearing forms, compared to B6 BMDM. Under the stress of IFN-c priming, this difference was even more exaggerated, and the abnormal inclusions and chlamydial forms observed was quite striking. In fact, we were able to identify just 3-4 intact inclusions in the monolayer of infected, primed B6.C3H-sst1 macrophages, and the ones that were found were small, surrounded by an irregular membrane with enhanced membrane blebbing, and contained no morphologically normal-appearing forms within them ( Figure 5C, bottom right). The mechanism behind this impaired development was not clear, and we found no evidence that steady-state mRNA or protein levels of IDO differed between the B6 and B6.C3H-sst1 macrophages, nor did we observe any recovery of organisms when primed B6.C3H-sst1 macrophages were supplemented with tryptophan (data not shown). Moreover, in contrast to what we observed with BMDM, lung fibroblasts derived from the B6.C3H-sst1 mice had no defect in their ability to support chlamydial development ( Figure S3), nor did they have any alteration in cytokine induction ( Figure S4). Thus, the phenotype appears to be specific to macrophages or macrophage-like cell types.
Enhanced IFN-b secretion leads to increased cell death and decreased Cp replication in macrophages from B6.C3H-sst1 mice carrying the sst1 S locus Development of Chlamydia pneumoniae is closely linked to the host cell. Although the intracellular lifestyle is ultimately a lytic process, the ability of this pathogen to actively inhibit cell death through manipulation of apoptotic signaling pathways during development is clearly advantageous to its survival [37,38]. For example, Chlamydia pneumoniae has been shown to protect cells against mitochondria-induced cell death by inhibiting cytochrome c release in response to known apoptosis inducers, and this requires bacterial protein synthesis and the apparent generation of an antiapoptotic factor in the cytosol of infected cells [39,40]. We wanted to know if the decreased capacity of the B6.C3H-sst1 BMDM to support chlamydial development might reflect loss of this antiapoptotic effect of productive chlamydia infection. We first examined cytotoxicity using propidium iodide staining as a measure of overall cell death. As shown in Figure 6A, B6.C3H-sst1 BMDM had significantly more cytotoxicity following Cp infection at both 48 and 72 hpi compared to wild type B6 BMDM, particularly when subjected to IFN-c priming. In order to determine if the Cp induced cytotoxicity in the B6.C3H-sst1 strain was linked to the exaggerated production of IFN-b by infected macrophages, we performed IFN-b neutralization assays. As shown in Figure 6B, neutralization of IFN-b, but not IL-10, rescued B6.C3H-sst1 BMDM from Cp-induced death, regardless of IFN-c priming, and enhanced the recovery of viable bacteria from infected BMDM ( Figure 6C). To confirm the mechanism of cell death in the B6.C3H-sst1 BMDM was due to induction of apoptosis, we repeated these studies using TUNEL staining, which detects DNA fragmentation [41]. We found that Cp infection of IFN-c primed B6.C3H-sst1 BMDM consistently led to 10-15% TUNEL positive cells, and that this was significantly inhibited in the presence of neutralizing IFN-b Ab (Figure 7). These data demonstrate that the Cp-induced anti-apoptotic program is not effective in blocking cell death in the B6.C3H-sst1 BMDM, and that this is apparently linked to the exaggerated IFN-b response.
Cp infected B6.C3H-sst1 mice carrying the sst1 susceptible locus display increased apoptosis in the lungs The final question was whether we could link Cp induced cell death and loss of the anti-apoptotic program in vitro to the striking in vivo phenotype observed in the Cp-infected B6.C3H-sst1 mice. We examined lungs from infected B6 and B6.C3H-sst1 mice for evidence of apoptosis, probing for cleaved caspase-3, a known apoptosis effector protein, by immunohistology. As shown in  , we saw a significant increase in the number of cells staining for cleaved caspase-3 by in the B6.C3H-sst1 mice compared to the B6 control, particularly within the infiltrating neutrophils and macrophages, as well as the endothelial cells lining the blood vessels. The areas of cleaved caspase-3 correlated with the areas of alveolar wound repair, as detected by TTF-1 staining shown in Figure S2 and suggested an association between enhanced apoptosis and the tissue damage and repair process observed in the lungs of the B6.C3H-sst1 mice.

Discussion
C. pneumoniae is a common respiratory pathogen, and acute complications are rare except in the extremes of age and severely immunocompromised hosts. Thus, it seems likely that host factors play an important role in determining the outcome of infection. The B6.C3H-sst1 mouse strain is essentially an immunocompetent host, and differs from the congenic C57BL/6 strain only at a 12-cM interval on mouse chromosome 1 where the sst1 susceptibility locus has been introgressed from the C3HeB/FeJ inbred mouse strain. This sst1 locus has primarily been studied as a resistance locus for M. tuberculosis [3,4,6,42], where it appears to impart a macrophage-specific phenotype that controls development of lung pathology and bacterial replication.
Like M. tuberculosis, respiratory infection with C. pneumoniae leads to the development of lung inflammation and granuloma formation to contain the infectious process. The C57BL/6 inbred mouse strain has been widely studied as a model for C. pneumoniae infection, and while infected mice display systemic signs of infection, they are largely tolerant of the inoculum size used in our report. We and others have found historically that C57BL/6 mice will clear the infection in the lungs over a 2-3 week period, with complete resolution of lung inflammation without intervention. However, when we examined the response of B6.C3H-sst1 mice to intranasal infection with C. pneumoniae, the impact of the sst1 susceptibility locus on the progression of disease, in terms of lung pathology and tissue repair, was quite striking, and mice were clinically so impaired that they generally required euthanasia within a week. Histological analysis of the lung tissue from infected sst1 susceptible B6.C3H-sst1 mice revealed a significant increase in the influx of neutrophils and macrophages, and marked tissue damage and repair, exemplified by the increased immunostaining for vimentin and TTF-1. Despite this severe clinical course and evidence of significant tissue injury, mice were not apparently impaired in their ability to control the bacterial load, suggesting the outcome was a consequence of the host immune response and not unchecked bacterial replication. This phenotype was quite unlike that reported in known immunocompromised hosts, including mice deficient in MyD88 [43], NOD/RIP2 [44], caspase-1 [45], IL-1b [45,46], and XIAP [47], where enhanced pathology and worse clinical outcome following C. pneumoniae infection could be ascribed to impaired bacterial clearance. Thus, the effect of the sst1 locus appears to represent an alteration in the host inflammatory response rather than a failure in the host antibacterial response.
Because in the tuberculosis and liseriosis models the sst1 phenotype is primarily a macrophage-specific phenomenon, we conducted a number of studies using bone marrow derived macrophages as an in vitro model. While there is some conflicting data with human peripheral blood mononuclear cells [48], C. pneumoniae is capable of productively infecting murine macrophages and Shimada et al. demonstrated that, in the lungs of C. pneumoniae infected mice, the majority of the cells harboring chlamydia were of the macrophage lineage [44]. Thus, these cells are relevant to the pathogenesis of C. pneumoniae induced pneumonia, at least in the mouse model. Given that Chlamydia species are known to upregulate IFN-c production during infection, and that macrophages in the lung are likely to be primed during pneumonia due to local cytokine production, we felt a comparison of resting BMDM vs. IFN-c primed the BMDM would enable us to more closely replicate the lung environment during infection in vitro. Similar to what was observed in vivo, the in vitro data again demonstrated an altered immune response with exaggerated IL-10 and IFN-b secretion observed during infection of primed macrophages. This was unexpected as these two cytokines are not typically upregulated to a significant degree in response to C. pneumoniae. Our investigation of the sst1 locus and the altered cytokine response on chlamydial development provided us with another unique experimental tool with which to study host-pathogen interactions. As chlamydial development is closely tied to the host cell which it parasitizes, the carefully orchestrated cycle of EBs and RBs is largely driven by the intracellular milieu. We observed that inclusion development was clearly impaired, as was the recovery of infectious EBs, in the infected sst1 susceptible B6.C3H-sst1 macrophages compared to the wild type C57BL/6 macrophages. This was even more pronounced with IFN-c primed macrophages. The mechanism by which the IFN-c primed sst1 susceptible macrophages fail to support chlamydial development remains unclear, although we did not find any evidence that it was a result of differential IDO expression. It remains to be established whether aberrant maturation of the bacteria within the sst1 susceptible macrophages reflects their impaired metabolic status and/or possibly altered regulation of autophagy. This link is suggested by a recent finding in the tuberculosis model, where autophagy in infected macrophages was activated by mycobacterial DNA and regulated via the type I interferon pathway [49].
The connection between the sst1 susceptible locus and activation of the normally silent type I IFN pathway in the context of C. pneumoniae infection is most intriguing. Indeed, Byrne and Rothermel had data in 1983 suggestive of the ability of type I IFNs to inhibit growth of some Chlamydia species based on studies that used a crude ''fibroblast interferon'' preparation [50]. We observed that the sst1 locus regulated the type I IFN response to C. pneumoniae infection as IFN-c primed B6.C3H-sst1 macrophages secreted large amounts of IFN-b upon C. pneumoniae infection while wild type C57BL/6 macrophages did not. Activation of the type I IFN pathway correlated with impaired chlamydial development since IFN-b neutralization improved bacterial recovery. We also found that the sst1 susceptible locus regulated cell death, both in vitro and in vivo, with the induction of apoptosis in vitro also dependant on the upregulation of IFN-b. Our inhibitor data suggests the mechanism by which the type I IFN pathway is activated by C. pneumoniae in the B6.C3H-sst1 macrophages is dependent on IRF3 and IKKj/TBK1, and may involve PKR. This, combined with the impaired chlamydial development and fragile-appearing inclusion membranes, suggests to us a model whereby leakage of microbial nucleic acid antigens out of the inclusion might either directly activate cytosolic DNA sensing pathways or target the inclusion to phagosomes that are normally avoided. However, the specific receptors engaged upstream of Figure 5. In vitro C. pneumoniae growth. BMDMs prepared from C57BL/6 (B6) or B6.C3H-sst1 mice were infected with C. pneumoniae at an MOI of 3:1, in the absence or presence of IFN-c (5 U/ml). A-B: Cells were lysed at the indicated time points and recovered EBs were quantitatively cultured, as described in the Methods. Each time point was run in triplicate and data is reported as the mean recovered bacteria per well 6 SEM. C: Cells were harvested at 69 hpi and processed for electron microscopy, as described in the Methods. Shown above are representative images from triplicate infected wells. IRF3 and IKKj/TBK1 remain undefined, as does the explanation as to why this pathway is not engaged in the IFN-c primed C57BL/6 macrophages, which carry the sst1 resistant locus.
The genetic dissection of the sst1 locus to identify causal genetic variation has been previously performed using a positional cloning approach in the tuberculosis model. These studies revealed complex structure of the candidate region, as it encompasses the largest repeat region in the mouse genome HSR (homogeneously stained region). This repeat was originally described in populations of wild mice, where its maintenance has been attributed to positive selective pressure [51]. We determined that the tuberculosis susceptibility allele was located within the HSR repeat region, Figure 6. Neutralization of IFN-b protects cells from death and enhances C. pneumoniae replication. BMDMs prepared from C57BL/ 6 (B6) or B6.C3H-sst1 mice were infected with C. pneumoniae (MOI 3:1) in the presence or absence of IFN-c (5 U/ml). The ratio of dead cells to total cell number was calculated using propidium iodide and Hoechst staining, as described in the Methods section, and is reported as the percent cytotoxicity. Where indicated, antibody neutralization was carried out using neutralizing antibody against IL-10 or IFN-b, or isotype control at a concentration of 250 ng/ml prior to infection or IFN-c treatment. A: Cytotoxicity was determined over time at 24, 48 and 72 hpi in Cp infected BMDM, and is reported as percent cell death. B: Effect of IL-10 and IFN-b neutralization on Cp-induced cytotoxicity at 72 hpi. C: Effect of IL-10 and IFN-b neutralization on Cp growth. Infected BMDM were lysed at 69 hpi, and recovered EBs were quantitatively cultured, as described in the Methods. All data shown above represents the mean 6 SEM from triplicate wells, and is representative of at least two independent experiments. Significance: *, p,0.05; **, p#0.01; ***,  which encodes multiple copies of Ifi75 (interferon-inducible 75) gene (human homologue is known as Sp110). One particular variant of Ifi75 encoding a full length protein corresponding to the Sp110b isoform, which we named Ipr1 (for intracellular pathogen resistance), was found to be absent in the lungs of M. tuberculosis infected sst1 susceptible mice compared to sst1 resistant mice, and expression of Ipr1 in macrophages both in vitro and in vivo restored many key functions of tuberculosis resistance on the sst1 susceptible background [6]. However, generation of a single Ipr1/Sp110 knockout mouse was not feasible due to the repeat structure. Therefore, the specific role of Ipr1/Sp110 in control of C. pneumoniae remains to be explored. Recently the mouse Sp110 gene has been identified as a regulator of the type I interferon response to cytosolic DNA [52]. Thus while we have no direct data to suggest Ipr1 is involved in our C. pneumoniae model, the similarities of the sst1 susceptible phenotype in these various infection models with regard to cell death and type I interferon production in macrophages suggests that it is an excellent candidate for future studies. Nevertheless, other polymorphic genes within the sst1 region may also contribute to the observed differences between the sst1 susceptible and sst1 resistant animals. Indeed, clustering of functionally related genes in mammalian genomes, as well as multiple functional polymorphisms within candidate chromosomal loci, has been documented in other forward genetic analyses of inflammatory diseases [53].
Finally, while our data suggest the sst1 locus regulates a shared defense mechanism against intracellular pathogens, the effects can be quite different amongst these pathogens, likely reflecting differences in their intracellular lifestyles. For example, the sst1 susceptible mouse strain is clearly compromised in its ability to control the bacterial burden during infection with M. tuberculosis and L. monocytogenes. In the case of M. tuberculosis, we observed unrestricted mycobacterial growth in the sst1 susceptible mice which correlated with a shift from an apoptotic to a necrotic cell death pathway [3]. Likewise, in the case of L. monocytogenes, sst1 susceptible macrophages were impaired in their ability to kill the bacteria in vitro and in vivo [5]. Thus, in the context of tuberculosis and listeriosis, the sst1 susceptible locus appears to also control anti-bacterial resistance.
However, in the case of C. pneumoniae we observed a different phenotype. We found no significant difference in the bacterial burden in vivo between the sst1 susceptible and sst1 resistant mouse strains. In fact, in vitro, we found the sst1 susceptible macrophages were actually impaired in their ability to support bacterial growth and thus bacterial replication was higher in the sst1 resistant macrophages compared to the sst1 susceptible macrophages. Yet, the sst1 susceptible macrophages demonstrated significantly higher IFN-b induction, and in vivo the sst1 susceptible mice displayed more severe inflammation and fibrosis, which correlated with activation of an apoptotic cell death pathway that is normally actively suppressed during infection. This suggests that in the context of C. pneumoniae infection the sst1 susceptible mice are not impaired in their resistance to the infection, but they are impaired in their tolerance to the infection-triggered lung inflammation. That is to say, they develop an inappropriately balanced inflammatory response and, as a result, display more severe disease. This concept of the difference between resistance and tolerance was recently reviewed by Medzhitov et al. [54].
We conclude that sst1 is an important host susceptibility locus which regulates a shared macrophage defense mechanism against diverse intracellular pathogens. Alteration at this locus from a ''resistant'' allele to a ''susceptible'' allele results in a severe phenotype in response to respiratory infection with C. pneumoniae, which is characterized by arrested bacterial development, activation of the normally silent type I IFN pathway, and induction of apoptosis. The enhanced pathological inflammatory response to the pathogen in this setting demonstrates the important role of host tolerance as a defense mechanism [54], and demonstrates that an effective antibacterial response does not necessarily correlate with an effective immune response. These data could be clinically relevant to acute C. pneumoniae infections in the context of a high type I IFN response in the lungs, such as in the setting of a postviral bacterial pneumonia. Adjunctive therapies to modulate the type I IFN response in the context of intracellular bacteria could improve survival and decrease the morbidity associated with these infections. Finally, further dissection of the sst1-mediated pathways using this unique mouse model may identify biomarkers that can be used to predict increased susceptibility to an array of intracellular bacterial infections, and suggest universal therapeutic targets for ameliorating their pathogenesis.

Ethics statement
All animal use protocols were approved by the Institutional Animal Care and Use Committee (IACUC) of Boston University, approved protocol number AN-14832, in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. Every effort was made to minimize discomfort, pain and distress in the animals. Boston University is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC).

Propagation of Chlamydiae
Chlamydia pneumoniae strain AR39 was obtained from Dr. Li Shen (Louisiana State University, New Orleans, LA). Chlamydia pneumoniae was propagated in L929 fibroblasts growing in RPMI-1640 medium supplemented with 10% FBS at 35uC, in a 5% CO2 environment. Following infection, cells were harvested, disrupted by glass beads or sonication (Sonicator 4000, Misonix Sonicators, Newtown, CT, USA), and chlamydiae were separated from cell debris by ultracentrifugation through 32% Renografin-60 (Bracco Diagnostics Inc., Princeton, NJ, USA). Chlamydial EBs were further purified by ultracentrifugation through 40% Renografin-60. After washing twice, the EB pellets were suspended in SPG (sucrose-phosphate-glutamate buffer, pH 7.2) and stored at 280uC. Bacterial titers were calculated as inclusion forming units (IFU) per ml. All the chlamydia stocks used in this study tested negative for Mycoplasma contamination by PCR [55]. All experimental procedures involving pathogenic bacteria were carried out with approval from the Institutional Biosafety Committee (IBC) at Boston University Medical Center.

Mice
C57BL/6 (abbreviated as B6) mice were purchased from Jackson Laboratory. The BL6.C3H-sst1 transgenic mice were generated from our lab. All animals were housed in groups of 3-5 mice per cage in a controlled environment (temperature 20-22uC, 12:12 hours light:dark cycle), given free access to food and water, and maintained under the supervision of veterinary staff from the Laboratory Animal Science Center (LASC) at Boston University Medical Center. All experimental procedures were carried out with approval from the Institutional Animal Care and Use Committee (IACUC) and the Institutional Biosafety Committee (IBC) at Boston University Medical Center.

Preparation of bone marrow-derived macrophages
Bone marrow derived macrophages (BMDM) were prepared as described previously [46]. Briefly, bone marrows were flushed from femurs and tibiae of the gender and age matched of C57BL/ 6 or BL6.C3H-sst1 congenic mice (6-8 wks age) and the cells were cultured in RPMI 1640 supplemented with 10% FBS, 20 mg/ml of gentamicin (Invitrogen, Life Technologies, Grand Island, NY), and 20-30% (v/v) of L929 condition medium (containing M-CSF). The cells were incubated at 37uC, 5% CO 2 incubator for 7-9 days to allow macrophage differentiation, and removed from gentamicin one day prior to infection with chlamydia.

Preparation of mouse lung fibroblasts
Mouse lung fibroblasts (MLF) were obtained by culture of lung single-cell suspension prepared from C57BL/6 or BL6.C3H-sst1 congenic mice. Single-cell suspension was prepared using a gentleMACS Dissociator (MACS Miltenyi Biotec. Auburn, CA). Briefly, lungs were removed from gender and age matched C57BL/6 or BL6.C3H-sst1 mice and briefly processed using a gentleMACS Dissociator, according to manufacturer's protocol. Lung connective tissue was then digested by incubation with Collagenase D and DNase I for 30 min at 37uC. After second homogenization step, single-cell suspensions were obtained by applying the homogenate to a 70 mm cell strainer filter and the cells were cultured in EMEM supplemented with FBS for 3 weeks. MLF purity was checked by staining for a fibroblast marker using ER-TR7 antibody, and was .90%.

Murine intranasal infection model
Groups of 6 (mock) or 19-20 (infected) gender and age matched of C57BL/6 and B6.C3H-sst1 congenic mice (7 to 8 weeks of age) were inoculated with C. pneumoniae AR39 or mock via the intranasal route under light anesthesia using ketamine/xylazine mix (60-100/5-10 mg/kg i.p.). Infected mice received 5610 6 IFU gradient purified AR39 in 20 mL of SPG followed by 20 ml of PBS; mock infected mice received 20 ml of SPG followed 20 ml PBS. Mice were weighed daily, observed and recorded the signs of distress. At day 3 and day 6, designated mice were euthanized by CO 2 inhalation. Following collection of the blood by cardiac puncture, the lungs and spleens were removed for cytokines, flow cytometry and bacterial quantification. Lungs and spleens were homogenized in PBS using a Medimachine System (BD Biosciences, San Jose, CA). Three mice from each group were designated for histopathology and immunohistochemistry and processed as follows: lungs were inflated with 10% neutral formalin via the trachea, removed en bloc for further formalin fixation, and embedded in paraffin. All in vivo experiments were repeated at least twice.

Clinical score
The clinical score of the infected mice was determined by observing the signs of mouse activity, ruffled fur, hunched posture and degree of the weight loss. Briefly, the mice with healthy, no signs of illness were scored as 0; with subtle ruffled fur, weight loss less than 10% were scored as 1; mice with mild or moderately decreased activity, evident ruffled fur, weight loss within 10-20% were scored as 2; the mice with severe decreased activity and ruffled fur, hunched posture, weight loss over 20%, scored as 3; the mice with no activity, severe hunched posture, scored as 4 (moribund). Both a mean and median clinical score were calculated.
Quantitative culture of C. pneumonia Quantitative culture of C. pneumoniae in lung homogenates, spleen samples, or cell lysates was carried out as described previously [46]. Briefly, for infected tissues, serially diluted samples were inoculated in duplicate onto L929 seeded in a 96-well plate. Infection was initiated by centrifugation at 20006g for 1 hour at 35uC. Time ''zero'' for infection was calculated mid-way through the centrifugation step. Infected cells were incubated at 35u/5% CO 2 . After 48-56 hours incubation, the cells were fixed in ice-cold methanol, and inclusions were stained using a Chlamydia-specific LPS monoclonal antibody (gift of Dr. You-Xun Zhang, Boston Medical Center), followed by FITC-conjugated secondary antibody; cells were counter stained with Evans blue and DAPI (Sigma, St. Louis, MO). The inclusions were counted under fluorescence microscopy, and calculated as the number of IFU per well/per mouse. For in vitro infections of BMDM, cells were disrupted, supernatants collected, and inoculated in duplicate onto L929 fibroblasts seeded in a 96-well plate. At 56-70 hours post inoculation, the cells were fixed and chlamydial inclusion staining was performed as described above. Where indicated, infectivity was quantified from at least 10 images using Image-J software.

Histopathology and immunohistochemistry
Embedded lung blocks were cut completely in 7-8 mm paraffin sections, and every 10th section stained using routine hematoxylin and eosin (H&E), chloroacetate esterase (CAE) or Masson's trichrome protocols. Immunohistochemistry was carried out on deparaffinized tissue using the following: monoclonal Ab directed against mouse vimentin (clone V9), CD68 (clone KP-1), and Thyroid Transcription Factor-1 (TTF-1) (Clone 8G7G3/1; all from Ventana Medical Systems, Tucson, AZ, USA); polyclonal Ab directed against cleaved caspase-3 (Asp175; Cell Signaling Technology, Danvers, MA, USA); or appropriate isotype controls. Detection was carried out using horseradish peroxidase. Staining for neutrophils and macrophages was quantified by averaging 25 random fields using Image-J software. All slides were reviewed independently by a veterinary pathologist who was blinded as to the experimental design.
Electron microscopy of Chlamydia-infected BMDM BMDMs from B6 WT or B6.C3H-sst1 congenic mice were plated onto 6-well plates at 2610 6 cells/well overnight, and inoculated with Cp as described in the text. At 69 hpi, the cells were washed twice with cold PBS and detached using a cell lifter. The cells were prefixed with fixative reagent [2.5% glutaraldehyde/2.0% paraformaldehyde in 0.1M sodium cacodylate buffer] at 4uC for overnight. The sample pellets were enrobed and solidified in 2% aqueous agarose (A2576, Sigma) and then dissected into 1 mm cubes. The sample cubes were fixed for 1 h in 1% osmium tetroxide in 0.15 M cacodylate buffer, dehydrated, and embedded in Eponate-Araldite (Ted Pella, Redding, CA, USA). After overnight curing, ultrathin sections were cut and stained with uranyl acetate and lead citrate. The sections were examined at 80 kV in a JEOL transmission electron microscope equipped with a Gatan digital camera.

Cytokine assay
Supernatants from lung homogenates from infected mice were assayed for cytokines and chemokine using MilliplexHMAP Kit (22 plex, Millipore, Billerica, MA), except for IL-1b for which an individual ELISA was performed. For in vitro cytokine assays of infected cells, commercially available ELISA kits were used according to the manufacturer's instructions as follows: IL-1b, IL-6, and IL-10 from R&D Systems (Minneapolis, MN, USA); TNF-a from eBioscience. For mouse IFN-b ELISA, the rat antimouse IFN-b mAb (I7662-10A) used as capture antibody was from US Biological (Swampscott, MA, USA); rabbit anti-mouse IFN-b polyclonal Ab (32400-1) used as detection antibody and mouse rIFN-b standard (12400-1), were both from PBL Biomedical Laboratories (Piscataway, NJ, USA); and HRP-conjugated donkey anti-rabbit IgG (711-036-152) was from Jackson ImmunoResearch Laboratories (West Grove, PA, USA). Plates were read in ELx800 Universal Microplate Reader (Bio-Tek, Instruments, Inc., Winooski, VT, USA); and data analyzed using SoftMax Pro 4.6 software. The cytokine induction experiments were performed in triplicate wells, and the data presented as the mean plus or minus SEM. mRNA analysis by Reverse Transcriptase (RT)-PCR Total RNA from 2610 6 cells was purified using RNeasy Mini Kit (QIAGEN) following the Manufacturer's instruction. Total RNA (1 mg) was reverse transcribed using the SuperScriptH First-Strand Synthesis System from Invitrogen. GoTaqH qPCR Master Mix for Real-Time PCR was purchased from Promega (Madison, WI, USA). The primers used for amplification of IDO were: 59-ATA TTG CTG TTC CCT ACT GCG A-39 and 59-CAT ACA GCA GAC CTT CTG GCA -39. Cycling condition was 94uC for 2 min (1 cycle); 94uC 30 s, 62uC 30 s, 72uC 1 min for 22 cycles, annealing temperature went down 0.5uC for every next cycle; 94uC 30 s, 52uC 30 s, 72uC 1 min for 30 cycles; 72uC for 7 min (1 cycle). PCR products were visualized using 1.5% agarose gel containing SYBR Safe DNA stain. Relative expression was determined by normalizing data to the housekeeping gene, GAPDH.

Cell death assay
For cell survival assay, BMDMs were plated onto 96-well plates at 2.5610 4 cells/well with 2% of L929 culture supernatant in triplicate. The cells were pre-treated as described in the text, and infected with C. pneumoniae at an MOI of 3:1. At the designated time points, the cells were stained with Hoechst and propidium iodide for 20 min. Total cell number (Hoechst positive) and dead cell number (propidium iodide positive) were counted by using Cyntellect Celigo cytometer (San Diego, CA). TUNEL staining was performed on BMDMs from BL6.C3H-sst1 mice. BMDMs were pre-incubated with IFN-c, isotype control or anti-IFN-b for 6 hr, and then infected with Cp for 10 hr. Apoptotic cells were determined by using Click-iT TUNEL Alexa Flour Imaging Assay (Invitrogen) according to the manufacturer's protocol.

Statistical analysis
Each experiment was performed 2-3 times to demonstrate the reproducibility of the observed trend. For the in vivo studies, mouse number is indicated in the figure or the figure legend. For the in vitro studies, each experimental condition was run in triplicate, allowing us to calculate a mean and standard error of the mean. Unless otherwise noted, data was analyzed by calculating the mean and SEM values, and compared using the Student's t test (two-tailed). GraphPad Prizm 4.0 (GraphPad, CA) software was used for the analysis. Values of p,0.05 were considered significant. Figure S1 Weight loss and bacterial clearance over time in C. pneumoniae infected B6 vs B6.C3H-sst1 mice. C57BL/6 (B6) or B6.C3H-sst1 congenic mice were intranasally inoculated with 5610 6 IFU C. pneumoniae (Cp) or SPG buffer (mock) as described in the Methods. (A) Recorded daily weights of mice over 6 days post infection. Shown above is the weight change relative to the starting weight of each individual mouse. (B) Recovered Cp by quantitative culture from lung homogenates prepared on day 3, 6 and 24 post infection. Data is reported as total IFU Cp per mouse lung. Significance: NS, No significant difference; ***, p#0.001 Cp infected vs. mock infected. The result is a representative of two independent experiments. (TIF) Figure S2 C. pneumoniae infected B6.C3H-sst1 mice display more evidence of tissue damage and repair compared to B6 mice. C57BL/6 (B6) or B6.C3H-sst1 congenic mice were infected with C. pneumoniae as described in the Methods. Lungs were removed at the indicated time and processed as follows: immunohistochemistry for detection of vimentin at day 3 (A and B); Masson's trichrome staining at day 6 (C and D); and immunohistochemistry for detection of thyroid transcription factor-1 (TTF-1) at day 6 (E and F). Original magnification: 1006. Shown above are images from one of four infected mice from each genetic background. The result is a representative of two independent experiments. (TIF) Table S1 Clinical scores were determined on day 6 post-infection, and mice were subsequently assigned to two outcome groups based on severity of illness: severe to moribund animals (score 3-4) vs. subtle to moderately ill animals (score 0-2). Data shown above represents the number of mice that fell into each outcome group. Significance was calculated from the 262 contingency table using Fisher's exact t-test (two-tailed). N = 25 mice per genotype, pooled from two independent experiments. (DOC)