CD4+ T Cell-derived IL-10 Promotes Brucella abortus Persistence via Modulation of Macrophage Function

Evasion of host immune responses is a prerequisite for chronic bacterial diseases; however, the underlying mechanisms are not fully understood. Here, we show that the persistent intracellular pathogen Brucella abortus prevents immune activation of macrophages by inducing CD4+CD25+ T cells to produce the anti-inflammatory cytokine interleukin-10 (IL-10) early during infection. IL-10 receptor (IL-10R) blockage in macrophages resulted in significantly higher NF-kB activation as well as decreased bacterial intracellular survival associated with an inability of B. abortus to escape the late endosome compartment in vitro. Moreover, either a lack of IL-10 production by T cells or a lack of macrophage responsiveness to this cytokine resulted in an increased ability of mice to control B. abortus infection, while inducing elevated production of pro-inflammatory cytokines, which led to severe pathology in liver and spleen of infected mice. Collectively, our results suggest that early IL-10 production by CD25+CD4+ T cells modulates macrophage function and contributes to an initial balance between pro-inflammatory and anti-inflammatory cytokines that is beneficial to the pathogen, thereby promoting enhanced bacterial survival and persistent infection.


Introduction
Persistent bacterial infections have a significant impact on public health [1]. While evasion of host immune responses is a prerequisite for these chronic infections, the underlying mechanisms are not fully understood. Human brucellosis, caused by the intracellular gram-negative coccobacilli Brucella spp., is considered one of the most important zoonotic diseases worldwide, with more than 500,000 new human cases reported annually [2]. The disease is characterized by a long incubation period that leads to a chronic, sometimes lifelong, debilitating infection with serious clinical manifestations such as fever, arthritis, hepatomegaly, and splenomegaly [3,4]. Human and animal brucellosis share many similarities, such as persistence in tissues of the mononuclear phagocyte system, including spleen, liver, lymph nodes, and bone marrow [4]. Therefore, the use of animal models such as mice has been an important tool to better characterize the immune response to Brucella infection that leads to long-term bacterial persistence and chronic disease.
There is general agreement that the initial interferon gamma (IFN-c) mediated Th1 immune response is crucial for the control of Brucella infection, since absence of IFN-c results in decreased control of bacterial growth [5,6] and IFN-c-deficient C57BL/6 mice succumb to overwhelming disease [7]. However, the inflammatory response induced by Brucella spp. in vivo is much milder than that observed with pyogenic infections such as salmonellosis, suggesting the stealth of Brucella as a possible reason for the absence of early proinflammatory responses [8,9]. Recent studies have shown that Brucella spp. use both passive and active mechanisms to evade initial innate immune recognition through toll-like receptors (TLRs) [10]. Although avoidance of TLR recognition is a key factor in the lack of initial inflammation during Brucella infection, how subsequent interactions of Brucella with the host immune system result in chronic disease is poorly understood.
Interleukin-10 (IL-10) is an immunoregulatory cytokine produced by most T cell subsets, B cells, neutrophils, macrophages, and some dendritic cell subsets [11]. It is suggested that by acting on antigen-presenting cells such as macrophages, IL-10 can inhibit the development of Th1 type responses [12]. In the context of infectious diseases, it is believed that the host uses IL-10 to control over-exuberant immune responses to pathogenic microorganisms in order to limit tissue damage [11]. Interestingly, studies using chronic pathogens such as Leishmania major [13], human cytomegalovirus [14], or Mycobacterium tuberculosis (reviewed in [15]) have demonstrated that the absence of IL-10 leads to a better clearance of these pathogens, with variable degrees of immunopathology. These studies suggest that pathogens have developed mechanisms to take advantage of the host immune-regulation in order to persist for longer periods and establish chronic infection.
Similar to other chronic pathogens, B. abortus infection induces IL-10 production [5,6,16]. Moreover, IL-10 gene polymorphisms have been associated with increased susceptibility to human brucellosis [17]. However, questions regarding the impact of IL-10 in B. abortus persistence and establishment of chronic infection, as well as the cell types responsible for this cytokine production remain to be answered. Therefore, we used IL-10 deficient mice to determine the role of IL-10 in modulating the initial immune response to Brucella infection. Furthermore, using cell-specific knock-out mice, we elucidated the immunological mechanisms underlying IL-10 induced immune-regulation during Brucellosis.

Results
Lack of IL-10 production during early B. abortus infection results in lower bacterial survival and increased pathology in vivo IL-10 has an important role in controlling the immune response induced by different inflammatory processes [15]. Moreover, Brucella infection has been shown to induce IL-10 production by splenocytes in vitro [5] and during intravenous in vivo infection [6,16]. To determine the time-course of IL-10 production during B. abortus infection, C57BL/6 mice were infected intraperitoneally (IP) with 5610 5 CFU of the virulent B. abortus strain 2308 and IL-10 production was determined at 3, 9, 15, and 21 days postinfection (d.p.i.). Infected mice exhibited significantly higher levels of IL-10 in the serum (Fig. 1A), which was associated with increased IL-10 transcript levels in the spleen (Fig. 1B) and liver (Fig. 1C) as early as 3 d.p.i. Importantly, significantly increased levels of IL-10 in serum and infected organs was only detected until 15 d.p.i., suggesting a possible regulatory function for this cytokine during acute Brucella infection.
The typical tissue response to Brucella infection is granulomatous inflammation [4]. However, in the spleen and liver of IL-10 2/2 mice B. abortus infection resulted in development of an acute inflammatory response characterized by vasculitis and thrombosis, necrosis, and influx of neutrophils ( Fig. 2E and 2F). Collectively, these data demonstrate a critical role for IL-10 in modulating the initial inflammation and pathology in response to B. abortus infection, which in turn benefits the pathogen due to enhanced bacterial survival.
CD4 + CD25 + T cells are the main IL-10 producers during early B. abortus infection in vivo IL-10 can be produced by different T cell subsets, as well as by B cells, neutrophils, macrophages, and some DC subsets [18]. To determine the cell types responsible for IL-10 production during early B. abortus infection, Il10-GFP reporter mice [19] were infected IP with 5610 5 CFU of B. abortus and IL-10 producing cells were identified at 3 and 9 d.p.i. by flow cytometry. A significant increase in the number of IL-10 producing T cells was observed in infected mice at 3 and 9 days post infection, whereas macrophages presented increased production of IL-10 only at 3 d.p.i. Moreover, the number of IL-10 producing B cells, neutrophils, and dendritic cells did not change significantly when compared to uninfected mice ( Fig. 3A and 3B, and data not shown). Importantly, even though a significant increase in the number of IL-10 producing CD8 + T cells was observed (Fig. 3C), a tenfold higher number of IL-10 producing T cells was observed in the CD4 + T cell population (Fig. 3C). No IL-10 production by cd T cells was observed (data not shown).
Various IL-10 producing CD4 + T cells have been described, including the CD4 + CD25 + subset [20]. Interestingly, an expansion of the CD4 + CD25 + T cell population was observed in the spleen of B. abortus infected mice at 9 d.p.i. (Fig. S1). Moreover, a 5-fold greater number of IL-10 producing CD4 + CD25 + T cells was detected at 9 d.p.i., compared to IL-10 producing CD4 + CD25 2 T cells ( Fig. 3D and 3E), suggesting that the CD4 + CD25 + T cell subset is the major population responsible for IL-10 production during acute brucellosis in the mouse.

IL-10 production by T cells is required for B. abortus persistence and for pathology regulation in vivo
Our previous data suggested that T cells and possibly macrophages would be the main cell types producing IL-10 during early Brucella infection. Therefore, to further investigate the importance of macrophage derived IL-10 during Brucella infection, we generated Il-10 flox/flox LysMCre +/2 (IL-10flox/LysMCre) mice, which have macrophages and neutrophils that are unable to produce IL-10. IL-10flox/LysMCre mice and littermate Il-10 flox/flox LysMCre 2/2 controls were infected IP with 5610 5 CFU of B. abortus 2308, and at 3, 9 and 21 d.p.i. disease progression was evaluated. Interestingly, IL-10flox/LysMCre infected mice exhibited decreased levels of IL-10 in the serum (Fig. 4A) when compared to control animals only at 3 d.p.i., Moreover, IL-10flox/LysMCre mice exhibited increased ability to control B. abortus infection in both spleen (Fig. 4B) and liver (Fig. 4C), only at initial stages of infection. This increased host resistance was accompanied by significantly higher levels of the pro-inflammatory cytokines IL-6, TNF-a and IFN-c in serum (Fig. 4D), spleen (Fig. 4E) and liver (data not shown) as well as increased histopathological lesions in infected organs (Fig. S2) of IL-10flox/LysMCre mice when compared to littermate controls

Author Summary
Brucella spp. are pathogens causing chronic intracellular infections that evade detection by pattern recognition receptors of the innate immune system. In this work, we tested the hypothesis that, in addition to eliciting a weak proinflammatory response during infection, induction of an immunoregulatory environment early during infection could promote persistent infection. Our results show that IL-10 produced at early time points is important for blunting inflammatory responses to B. abortus in infected tissues. CD4 + T cells are an important source of this cytokine, since mice lacking T cell-derived IL-10 exhibited increased inflammatory pathology and also were better able to control B. abortus infection. A target of this CD4 T cell-derived IL-10 is macrophages, since treatment of these cells with IL-10 in vitro supported intracellular replication of B. abortus, while blocking IL-10 restricted the ability of B. abortus to exit the phagolysosomal compartment and replicate intracellularly. Further, mice conditionally deficient for IL-10 receptor on macrophages were better able to control infection with B. abortus. Taken together, our results support a model in which IL-10 production by T cells promotes chronic infection by rendering macrophages more permissive for intracellular replication of B. abortus.  Collectively, this data suggests that macrophage derived IL-10 plays a limited role in development of the chronic disease caused by B. abortus and raises the possibility that T-cells could, indeed, be the main cell type responsible for IL-10 production during early Brucellosis.
Previous results using IL-10 2/2 and IL-10flox/LysMCre mice ( Fig. 2 and 4) suggested that IL-10 was important for initial B. abortus persistence in the host. Therefore, we investigated whether the lack of T cell-derived IL-10 would affect bacterial persistence. Indeed, IL-10flox/CD4Cre mice exhibited significantly improved control of B. abortus infection in the spleen and liver at 9, 21 and 42 d.p.i. (Fig. 5A and 5B).
Controlling B. abortus replication could be beneficial to the host, since the ability to survive for longer periods is a key mechanism for chronic pathogens to thrive. However, loss of IL-10 driven immune modulation has been shown to cause severe and sometimes lethal inflammatory responses in different infectious disease models [15]. To determine the disease progression, weight changes in IL-10flox/CD4Cre and control mice were followed at 3, 9, 15, 21, 28, 35 and 42 d.p.i. (Fig. 5C). To ensure that any detectable change in weight was the result of B. abortus infection, uninfected IL-10flox/CD4Cre and littermate control mice were also used (data not shown). IL-10flox/CD4Ce mice exhibited significantly decreased weight gain during early infection and, by 21 days post-infection, they started to behave like control animals. IL-10flox/CD4Cre uninfected mice did not exhibit any slower weight gain and behaved like uninfected control mice (data not shown). Moreover, increased levels of the pro-inflammatory cytokines IFN-c, IL-6, and TNF-a were observed in serum ( To determine if the significantly reduced weight gain and higher induction of pro-inflammatory cytokines were associated with a detrimental inflammatory response, spleen and liver sections from IL-10flox/CD4Cre and control mice were blindly evaluated by veterinary pathologists (MNX and TMS). Interestingly, IL-10flox/ CD4Cre mice exhibited increased pathology, characterized by marked influx of neutrophils and histiocytes in the spleen ( Fig. 6A and 6C), as well as tissue necrosis and multifocal neutrophilic vasculitis and thrombosis in the liver ( Fig. 6B and 6C) at 3 and 9 d.p.i., suggesting an acute inflammatory response. Importantly, a hallmark of B. abortus infection is a mild initial pro-inflammatory response, which leads to a chronic infection characterized by formation of granulomas in infected organs [4,22]. However, by 21 d.p.i., IL-10flox/CD4Cre mice exhibited decreased granuloma formation in the spleen ( Fig. 6A and 6C) when compared to littermate control mice. Taken together, these data demonstrate that T cell-derived IL-10 production during early B. abortus infection is crucial for the development of the chronic disease and morbidity caused by B. abortus, and limits the production of the pro-inflammatory response necessary to control the infection. However, the cell types affected by the IL-10 production during Brucella infection remained unclear.
Lack of IL-10 results in lower B. abortus survival in infected macrophages due to bacterial inability to escape late endosome Since Brucella spp. are known to invade and survive inside phagocytic cells such as macrophages [23], we hypothesized that macrophages could be the cell type affected by IL-10 during Brucella infection. To determine the effect of IL-10 on B. abortus survival during macrophage infection in vitro, bone marrow derived macrophages (BMDM) from C57BL/6 and IL-10 2/2 mice were infected with B. abortus 2308 (MOI = 100) and the bacterial survival was measured at 1, 8, and 24 h post-infection (h.p.i.) (Fig. 7A). B. abortus infected wild-type BMDM produced significant amounts of IL-10 ( Fig. S4A). B. abortus exhibited a significantly decreased ability to survive inside BMDM from IL-10 2/2 mice at 8 and 24 h.p.i. when compared to BMDM from wild-type mice. Importantly, B. abortus infected IL-10 2/2 BMDM did not exhibit increased cell death, determined by LDH assay (data not shown). Moreover, to ensure that the observed effect resulted from an absence of IL-10 production, recombinant IL-10 (rIL-10) was added to the IL-10 2/2 BMDM media during infection (Fig. 7A). As expected, the addition of rIL-10 restored B. abortus survival inside macrophages. Results similar to those described above (Fig. 7A) were also observed in IFNc-activated BMDM (data not shown).
The ability of Brucella spp. to persist and replicate within macrophages involves a temporary fusion of Brucella-containing vacuole with the late endosome/lysosome during the initial hours post-infection, and subsequent exclusion of the endosomal/ lysosomal proteins from the Brucella-containing vacuole [24]. To determine if B. abortus inability to persist inside macrophages was due to changes in the pathogen's intracellular trafficking, we infected wild-type and IL-10 2/2 BMDM with B. abortus expressing mCherry (MOI = 100) and bacterial co-localization with the late endosome marker LAMP1 was determined at 24 h.p.i. by confocal microscopy. Interestingly, while the majority of B. abortus was found to lack LAMP1 in wild type BMDM, the opposite was found in BMDM derived from IL-10 2/2 mice, in which over 80% of bacteria were co-localized with LAMP1 ( Fig. 7B and 7C). This phenotype was IL-10-dependent, since addition of rIL-10 to the IL-10 2/2 BMDM restored the ability of the pathogen to escape the late endosome ( Fig. 7B and 7C). Results similar to those described ( Fig. 7B and Fig. 7C) were also observed in IFNc activated BMDM (data not shown).
Lack of IL-10 results in higher levels of NF-kB activation and pro-inflammatory cytokine production in B. abortus infected macrophages It has been demonstrated that IL-10 can inhibit a protective immune response, possibly by blocking NF-kB activation [25] and downstream production of pro-inflammatory cytokines by antigenpresenting cells such as macrophages [15,26]. To investigate if IL-10 production would have an effect on NF-kB activation in Brucella infected macrophages, we used an NF-kB reporter RAW murine (*) represents P,0.05 when compared to control. (**) represents P,0.05 when compared to days 15 and day 21 for (C) using unpaired t-test statistical analysis. doi:10.1371/journal.ppat.1003454.g001 T Cell Derived IL-10 Promotes Brucella Persistence PLOS Pathogens | www.plospathogens.org  S4B). Blockage of IL-10R resulted in decreased B. abortus intracellular survival inside RAW-Blue macrophages when compared to untreated controls at 24 h.p.i., as previously observed in IL-10 2/2 BMDM (data not shown). B. abortus infection did not result in significant NF-kB activation by any of the treatment groups at 8 h.p.i. (Fig. 8A). The use of IL-10R Ab to block the response of B. abortus infected RAW-Blue macrophages to IL-10 resulted in a significant increase in NF-kB activation at 24 h.p.i. when compared to untreated infected macrophages. Conversely, addition of exogenous rIL-10 resulted in a significant inhibition of NF-kB activation in B. abortus infected cells (Fig. 8A). The results above described were also observed in IFNc activated RAW blue cells (data not shown).
To confirm that IL-10 affects pro-inflammatory cytokine production by infected macrophages, bone marrow-derived macrophages (BMDM) from C57BL/6 and IL-10 2/2 mice were infected with B. abortus 2308 (MOI = 100) and cytokine expression was measured at 24 h.p.i. by ELISA and quantitative real-time PCR. Significantly, an absence of IL-10 resulted in higher expression levels of the pro-inflammatory cytokines IL-6 and TNFa by infected macrophages (Fig. 8B and 8C). Moreover, the phenotype observed was shown to be IL-10 dependent, since the addition of rIL-10 to IL-10 2/2 infected BMDM restored IL-6 and TNF-a expression to wild-type levels. Results similar to those described in Fig. 8B and Fig. 8C were also observed in IFN-c activated BMDM (data not shown). Our results demonstrate that IL-10 production during B. abortus infection in vitro affects macrophage function by modulating NF-kB activation and the production of pro-inflammatory cytokines by infected cells.
Inability of macrophages to respond to IL-10 results in severe acute pathology and decreased B. abortus survival in vivo To further investigate the possibility that macrophages are the cell type responding to the IL-10 produced during B. abortus infection, we used Il-10R flox/flox LysMCre +/2 (IL-10Rflox/ LysMCre) mice, which are unable to express the IL-10 receptor 1 chain (IL-10R1) specifically in monocytes/macrophages and/or neutrophils [27]. IL-10Rflox/LysMCre and Il-10R flox/flox -LysMCre 2/2 control mice were infected intraperitoneally with 5610 5 CFU of B. abortus 2308 for 3, 9, 21 and 42 days and bacterial survival in infected organs was evaluated. Our results from in vitro infection suggested that IL-10 affects the ability of B. abortus to survive inside macrophages. Remarkably, IL-10Rflox/ LysMCre mice showed lower CFU counts in both spleen (Fig. 9A) and liver (Fig. 9B) at 9, 21 and 42 d.p.i. when compared to littermate control mice. This data provided strong support for the idea that macrophage responsiveness to IL-10 is necessary for optimal initial B. abortus colonization of the host as well as longterm persistence.
The results shown above (Fig. 5) demonstrated that a lack of IL-10 production by T cells during B. abortus in vivo infection resulted in increased pro-inflammatory responses and evident clinical signs of disease in mice. Moreover, our in vitro results suggested that blockage of IL-10R played a role in the control of NF-kB activation and pro-inflammatory cytokine production by B. abortusinfected macrophages. Therefore, we sought to determine the effect of macrophage responsiveness to IL-10 in the early host response to B. abortus infection. B. abortus-infected IL-10R/ LysMCre mice exhibited decreased weight gain at 9, 15, and 21 d.p.i. when compared to wild-type infected mice (Fig. 9C). Furthermore, levels of IFN-c, IL-6 and TNF-a were significantly increased in serum (Fig. 9D) spleens (Fig. 9E) and livers (Fig. S5) of IL-10R/LysMCre at 3, 9 d.p.i. and, to a lesser extent, at 21 d.p.i.
To determine if macrophage non-responsiveness to IL-10 would result in detrimental pathologic changes, spleen and liver sections from infected IL-10Rflox/LysMCre and control mice were blindly evaluated by veterinary pathologists (MNX and TMS). As expected, IL-10Rflox/LysMCre showed severe acute lesions characterized by influx of neutrophils and histiocytes, as well as tissue necrosis and multifocal neutrophilic vasculitis and thrombosis in the spleen at 9 d.p.i. (Fig. 10A and 10C) and in the liver ( Fig. 10B and 10C) at 9 and 21 d.p.i. However, at 21 and 42 days post-infection, IL-10Rflox/LysMCre mice exhibited decreased granuloma formation in spleen ( Fig. 10A and 10C) when compared to littermate control mice, suggesting that macrophage responsiveness to IL-10 is important for development of chronic pathological lesions in spleens of infected animals. These data provide the direct support for the idea that induction of IL-10 during B. abortus in vivo infection plays a key role in modulation of macrophage function, which, in turn, provides the ideal initial immunological environment for bacterial colonization and development of chronic infection.

Discussion
The balance between pro-inflammatory and anti-inflammatory cytokine production appears to be crucial for the ability of the host to eradicate an infection, as well as for the clinical presentation and/or pathology resulting from the infection. This balance appears to shift in the case of persistent pathogens such as Brucella spp., which are able to evade TLR signaling during the early stages of infection, thereby preventing development of an immune response that is appropriate to clear the infection [28]. It is known that during the acute phase of B. abortus infection in mice, neutralization of IL-10 reduces bacterial colonization [5]. Here, we provide support to the idea that during this phase of B. abortus infection, early production of IL-10 by T cells is key to promoting persistent intracellular infection.
In a number of infectious disease models, several cell types, including T cell subsets, B cells, neutrophils, macrophages and some DC subsets have been shown to be able to produce IL-10 [18]. Even though B. abortus infected macrophages are capable of producing IL-10 in vitro, only B cells have been implicated as a potential source of IL-10 during in vivo infection [29]. In this study, however, we have demonstrated that macrophages play a limited role in IL10 production during early acute B. abortus infection. Additionally, we identified T cells, more specifically CD4 + CD25 + T cells, as the major source of this cytokine during acute brucellosis. Indeed, Svetić and collaborators [1] have suggested a measured by ELISA in serum of C57BL/6 and IL-10 2/2 mice infected with B. abortus 2308 for 9 days. n = 5. Values represent mean 6 SEM. *P,0.05 using unpaired t-test statistical analysis.( E) Histopathology score of spleen and liver from C57BL/6 and IL-10 2/2 mice infected with B. abortus 2308 for 9 days. (F) Representative pictures from (E). Black arrow shows microgranulomas, white arrowheads shows neutrophilic infiltrate and upper case N shows areas of coagulative necrosis (620). n = 5. Values represent individual mice (black circles) and geometric mean (black dash). *P,0.05 using Mann-Whitney statistical analysis. doi:10.1371/journal.ppat.1003454.g002 possible role for CD4+ T cells in IL-10 production during acute murine Brucellosis. Interestingly, previous studies have demonstrated elevated numbers of CD4+CD25+ T cells in PBMCs from human patients with acute brucellosis [30] as well as in draining lymph nodes from B. melitensis infected sheep [28]. Moreover, Pasquali and collaborators demonstrated that depletion of CD4 + CD25 + T cells resulted in increased control of B. abortus infection due to elevated activation of effector T cells and higher production of pro-inflammatory cytokines such as IFN-c by infected mice [31]. Most likely, the effects seen in this latter study were an indirect effect of the decreased IL-10 production by T cells, resulting in elevated macrophage activation in CD25depleted mice upon B. abortus infection. Taken together, these results point to CD4 + CD25 + T cells as important players in modulating the early immune response to B. abortus in vivo.
There is a general agreement that macrophages represent a critical niche for Brucella persistence in the host [4]. Importantly, they have been described as one of the cell types responding to IL-10 production in other infection models [11,27]. Here we have demonstrated that macrophages are a main cell type responding to the immunomodulatory functions of IL-10 during both in vitro and in vivo Brucella infection. Moreover, IL-10 signaling had a Figure 3. CD4 + CD25 + T cells are the main producers of IL-10 during early Brucella abortus in vivo infection. (A) Flow cytometry measurement of IL-10 expression in splenic T cells, B cells, and macrophages from C57BL/6 IL-10 GFP-reporter mice infected with B. abortus 2308 for 3 and 9 days. (B) Representative data plot of IL-10 expression in splenic T cells (CD3), B cells (B220), and macrophages (CD11b + F4/80 + cells) from C57BL/ 6 IL-10 GFP-reporter mice. (C) Flow cytometry measurement of IL-10 expression in splenic CD8 + T cells and CD4 + T cells from C57BL/6 IL-10 GFPreporter mice infected with B. abortus 2308 for 9 days. (D) Flow cytometry measurement of IL-10 expression in splenic CD4 + CD25 + T cells and CD4 + CD25 2 T cells from C57BL/6 IL-10 GFP-reporter mice infected with B. abortus 2308 for 9 days. (E) Representative data plot of IL-10 expression in splenic CD4 + CD25 + (upper panel) and CD4 + CD25 2 (lower panel) T cells from C57BL/6 IL-10 GFP-reporter mice. Values represent mean 6 SEM. n = 4. (*) represents P,0.05 relative to uninfected control (day 0), (**) represent P,0.05 relative to day 3 infection for (A), relative to CD8 + day 9 infected for (C) and relative to CD4 + CD25 2 for (D) using unpaired t-test statistical analysis. FMO = fluorescence minus one. doi:10.1371/journal.ppat.1003454.g003 significant impact on the ability of B. abortus-infected macrophages to produce pro-inflammatory cytokines and to permit intracellular growth of B. abortus. Interestingly, O'Leary and collaborators have demonstrated that IL-10 production by immune cells can affect the ability of M. tuberculosis to escape the LAMP1 + late endosomal compartment and to establish infection in human macrophages in vitro [32]. In agreement with this previous study, we have demonstrated that the capacity of macrophages to respond to IL-10 impacts intracellular survival of Brucella by decreasing the pathogen's ability to escape the LAMP1 + late endosome, a prerequisite for replication in an endoplasmic reticulum-associated compartment.
It is important to note that the LysM promoter used to drive Cre expression in IL10Rflox mice is expressed in both neutrophils and macrophages [27]. Although our understanding of the role of neutrophils during brucellosis is still evolving [8,33], we and others have described recruitment of these cells to both spleens and livers of B. abortus-infected mice during the acute infection phase ( [33], Fig. 6 and Fig. 9). Moreover, the neutrophil recruitment was more evident in the absence of IL-10 (Fig. 2, Fig. 6 and Fig. 9). Therefore, although Brucella resistance to neutrophil killing has been well described [34], it is possible that the increased cytokine expression and pathology observed in IL10Rflox/LysMCre mice could also be due in part to a failure of neutrophils recruited to the site of infection to respond to IL-10.
It should be pointed out that, although IL-10 contributes to persistence of B. abortus in vivo, abrogation of IL-10 production (Fig. 2) or neutralization of IL-10 in vivo [5] did not result in eradication of B. abortus from tissues, contrary to what has been shown for Leishmania [13]. Therefore, factors in addition to IL-10 production must also contribute to chronic persistence of B. abortus. Although TGF-b has been shown to be produced by B cells and macrophages in BALB/c mice [29], consistent with this report, we did not observe any increase in circulating TGF-b1, TGF-b2, or TGF-b3 at 21 or 42 days post infection of C57BL/6 mice (data not shown). However, these results do not rule out a role for local activation of TGF-b in promoting chronic infection. An important factor in persistence of Brucella is the transient nature of IFNc production in infected mice, which subsides by 21 d post infection in mice [35], therefore our observed lack of a role of IL-10 later in infection could suggest that its role is to antagonize the activity of IFNc at earlier stages of infection. Finally, the possibility should be considered that during chronic infection, B. abortus may reside in a cell type that has inherently low microbicidal activity, as has been found for M. tuberculosis [36] and B. melitensis [37].
At first glance, our data suggest that inhibition of IL-10 signaling would be beneficial to the host, since IL-10 2/2 , IL-10flox/CD4Cre and IL-10Rflox/LysMcre showed increased ability to control B. abortus infection at both acute and chronic stages of infection. Moreover, both IL-10flox/CD4Cre and IL-10Rflox/LysMcre exhibited reduced formation of granulomas, a potential niche for B. abortus persistence [4], during the chronic stage of infection. However, in spite of the increased bacterial clearance, we demonstrated that lack of IL-10 during Brucella infection could potentially be detrimental to the host, since B. abortus infected IL-10 deficient mice presented evident signs of acute disease, characterized by changes in weight gain and marked histopathological lesions in both spleen and liver. Indeed, studies on other chronic pathogens such as Leishmania major [38], human cytomegalovirus [14], and M. tuberculosis (reviewed in [15]) have demonstrated that even though absence of IL-10 leads to better clearance of these pathogens, it can also result in severe and sometimes lethal pathologic changes. Therefore, although modulation of the IL-10 signaling pathway could be a potential target to avoid the establishment of chronic infection, more studies are needed to elucidate the optimal activation of the immune system necessary to improve clearance of chronic pathogens without a great cost to the host.

Bacterial strains, media and culture conditions
Bacterial strains used in this study were the virulent strain Brucella abortus 2308 and its isogenic mutant strain MX2 which has an insertion of pKSoriT-bla-kan-PsojA-mCherry plasmid [37]. For strain MX2, positive clones were kanamycin resistant and fluorescent, as previously described [37]. Strains were cultured on tryptic soy agar (Difco/Becton-Dickinson, Sparks, MD) or tryptic soy broth at 37uC on a rotary shaker. Bacterial inocula for mouse infection were cultured on tryptic soy agar plus 5% blood   [39]. For cultures of strain MX2, kanamycin (Km) was added to the culture medium at 100 mg/mL. All work with B. abortus cells was performed at biosafety level 3.

Bone marrow derived macrophage infection
Bone marrow-derived macrophages were differentiated from bone marrow precursors from femora and tibiae of female, 6 to 8 weeks old, C57BL/6J and IL-10 2/2 mice obtained from The Jackson Laboratory (Bar Harbor) following a previously published procedure [40]. For BMDM experiments, 24-well microtiter plates were seeded with macrophages at concentration of 5610 5 cells/ well in 0.5 mL of RPMI media (Invitrogen, Grand Island, NY) supplemented with 10% FBS and 10 mM L-glutamine (RPMI supl) incubated for 48 h at 37uC in 5% CO 2 . Preparation of the inoculum and BMDM infection was performed as previously described [40]. Briefly, for inoculum preparation, B. abortus 2308 was grown for 24 h and then diluted in RPMI supl, and about 5610 7 bacteria in 0.5 mL of RPMI supl were added to each well of BMDM, reaching multiplicity of infection (MOI) of 100. Microtiter plates were centrifuged at 2106 g for 5 min at room temperature in order to synchronize infection. Cells were incubated for 20 min at 37uC in 5% CO 2 , free bacteria were removed by three washes with phosphate-buffered saline (PBS), and the zero-time-point sample was taken as described below. After the PBS wash, RMPI supl plus 50 mg gentamicin per mL was added to the wells, and the cells were incubated at 37uC in 5% CO 2 . For cytokine production assays, supernatants from each well were sampled at 0, 8, 24, or 48 h after infection, depending on the experiment performed. In order to determine bacterial survival, the medium was aspirated at the time points described above, and the BMDM were lysed with 0.5 mL of 0.5% Tween 20, followed by rinsing of each well with 0.5 mL of PBS. Viable bacteria were quantified by serial dilution in sterile PBS and plating on TSA. For gene expression assays, BMDM were resuspended in 0.5 mL of TRI-reagent (Molecular Research Center, Cincinnati) at the timepoints described above and kept at 280uC until further use. When necessary, 1 ng/mL of mouse rIL-10 (eBioscience, San Diego, CA) or 1 ng/mL of mouse rIFN-c (BD Bioscience, San Jose, CA) was added to the wells and kept throughout the experiments. All experiments were performed independently in triplicate at least three times and the standard error for each time point calculated.

RAW-Blue macrophages experiments
RAW-Blue cells (Invivogen, San Diego, CA) were derived from RAW-264.7 macrophages with chromosomal integration of a SEAP reporter construct inducible by NF-kB and AP-1. RAW-Blue cells were maintained in Zeocin (Invivogen, San Diego, CA) selective medium. For RAW-Blue experiments, 24-well microtiter plates were seeded with macrophages at concentration of 2610 5 cells/well in 0.5 mL of DMEM media (Invitrogen, Grand Island, NY) supplemented with 10% FBS and 10 mM L-glutamine (DMEM supl). Preparation of the inoculum and RAW-Blue infection was performed as previously described [40], using MOI = 100. For NF-kB activation assays, supernatant from each well was sampled at 8 h and 24 h after infection and secretion of the substrate SEAP was detected and measured in a spectrophotometer at 650 nm with QUANTI-Blue (Invivogen, San Diego, CA) according to manufacturer's instructions. When necessary, 1 ng/mL of mouse rIL-10 (eBioscience, San Diego, CA),1 ng/mL of mouse rIFN-c (BD Bioscience, San Jose, CA),1 mg/mL of antimouse IL-10R antibody (R&D Systems, Minneapolis, MN) or anti-mouse IgG isotype antibody control (R&D Systems, Minneapolis, MN) were added to the wells and kept throughout the experiments. All experiments were performed independently in triplicate at least three times and the standard error for each time point calculated.

Ethics statement
Experiments with mice were carried out in strict accordance with the recommendations in the Guide for Care and Use of Laboratory Animals of the National Institute of Health and were approved by the Institutional Animal Care and Use Committees at the University of California at Davis (protocol number: 16468).

ELISA
The presence of IL-10, IL-6 and TNF-a in BMDM supernatant and in serum samples from C57BL/6, IL-10flox/CD4Cre, IL-10flox/LysMCre and littermate control mice infected with B. abortus 2308 was determined by indirect enzyme-linked immunosorbent assay (ELISA) (eBioscience, San Diego, CA) according to the manufacturer's instructions. The ELISA test was read at 450 nm with an ELISA microplate reader (MR5000; Dynatech). The sensitivity of the ELISA used was 7.8 pg/mL. Data points are the averages of duplicate dilutions, with each measurement being performed twice.

RT-PCR and real time PCR analysis
Eukaryotic gene expression was determined by real-time PCR as previously described [40]. Briefly, eukaryotic RNA was isolated  Table  S1 in Text S1. Data were analyzed using the comparative Ct method (Applied Biosystems, Carlsbad). Transcript levels of Il10, Il6, Ifng and Tnfa were normalized to mRNA levels of the housekeeping gene bactin.

Histopathology
Formalin fixed spleen and liver tissue sections were stained with hematoxylin and eosin, and two veterinary pathologists (MX and TS) performed a blinded evaluation using criteria described in Table S2 in Text S1. Representative images were obtained using an Olympus BX41 microscope and the brightness adjusted (Adobe Photoshop CS2).

Flow cytometry
Flow cytometric analysis of IL-10 producing cells was performed in splenocytes from IL-10 GFP reporter mice infected for 3 and 9 days with B. abortus 2308. Briefly, after passing the TNF-a expression by BMDM from wild type mice (C57BL/6) and IL-10 deficient mice (IL-10 2/2 ) in the presence or absence of recombinant IL-10. (C) ELISA essay for IL-6 and TNF-a production in supernatant from wild type mice (C57BL/6) and IL-10 deficient mice (IL-10 2/2 ) in the presence or absence of recombinant IL-10. Results shown are representative of two independent experiments. n = 5. Values represent mean 6 SEM. *P,0.05 using unpaired t-test statistical analysis. doi:10.1371/journal.ppat.1003454.g008 The cells were washed with FACS buffer and fixed with 4% formaldehyde for 30 min at 4uC, and resuspended in FACS buffer prior to analysis. Flow cytometry analysis was performed using an LSRII apparatus (Becton Dickinson, San Diego, CA), and data were collected for 5610 5 cells/mouse. Resulting data were analyzed using Flowjo software (Treestar, inc. Ashland, OR). Gates were based on Fluorescence-Minus-One (FMO) controls.

Immunofluorescence microscopy
Immunofluorescence of Brucella infected BMDM was performed as previously described [24]. Briefly, B. abortus MX2 infected BMDM were grown on 12-mm glass coverslips in 24-well plates were washed three times with PBS, fixed with 3% paraformaldehyde, pH 7.4, at 37uC for 20 min, washed three times with PBS and then incubated for 10 min in 50 mm NH 4 Cl in PBS in order to quench free aldehyde groups. Samples were blocked and permeabilized in 10% goat serum and 0.1% saponin in PBS for 30 min at room temperature. Cells were labeled by inverting coverslips onto drops of primary antibodies diluted in 10% horse serum and 0.1% saponin in PBS and incubating for 45 min at room temperature. The primary antibody used was rat anti-mouse LAMP-1 (BD Pharmigen, San Jose, CA). Bound antibodies were detected by incubation with 1:500 dilution of Alexa Fluor 488 donkey anti-rat (Invitrogen, Grand Island, NY) for 45 min at room temperature. Cells were washed twice with 0.1% saponin in PBS, once in PBS, once in H 2 O and then mounted in Mowiol 4-88 mounting medium (Calbiochem). Samples were observed on a Carl Zeiss LSM 510 confocal laser scanning microscope for image acquisition (Carl Zeiss Micro Imaging). Confocal images of 102461024 pixels were acquired as projections of three consecutive slices with a 0.38-mm step and assembled using Adobe Photoshop CS2 (Adobe Systems). For quantification of Brucella MX2 and Lamp1+ compartment colocalization, at least 100 bacteria/sample were counted. All experiments were performed independently in quadruplicate at least two times.

Statistical analysis
Fold changes of ratios (bacterial numbers or mRNA levels) and percentages (flow cytometry and fluorescent microscopy) were transformed logarithmically prior to statistical analysis. An unpaired Student's t-test was performed on the transformed data to determine whether differences in fold changes between groups were statistically significant (P,0.05). Significance of differences in histopathology scores was determined by a one-tailed nonparametric test (Mann-Whitney).  Figure S3 (A) ELISA essay for IL-10 production in serum from littermate control mice (grey bars) compared with IL10flox/ CD4Cre mice (white bars) at 0, 3, 9, 21 and 42 d.p.i. (B) qRT-PCR analysis of IL-10 expression in spleen from littermate control (grey bars) compared with IL10flox/CD4Cre mice (white bars) at 3, 9, 21 and 42 d.p.i. (C) qRT-PCR analysis of pro-inflammatory cytokines genes (Ifnc, Il6 and Tnfa) in liver from littermate control (grey bars) compared with IL10flox/CD4Cre mice (white bars) at 3, 9, 21 and 42 d.p.i. n = 5. Values represent mean 6 SEM. (*) represents P,0.05 relative to uninfected control using unpaired ttest statistical analysis. (TIF) Figure S4 (A) ELISA essay for IL-10 production in supernatant from C57BL/6 wild type BMDM infected with B. abortus 2308 for 24 h. (B) ELISA essay for IL-10 production in supernatant from RAW-Blue macrophages infected with B. abortus 2308 for 8 h and 24 h in the presence of IL-10 receptor blocking antibody (IL-10R Ab), isotype control (IgG Ab) or exogenous IL-10 (rIL-10). (TIF) Figure S5 qRT-PCR analysis of pro-inflammatory cytokines genes (Ifnc, Il6 and Tnfa) in liver from littermate control (grey bars) compared with IL10Rflox/LysMCre mice (white bars) at 3, 9, 21 and 42 d.p.i. n = 5. Values represent mean 6 SEM. (*) represents P,0.05 relative to uninfected control using unpaired t-test statistical analysis.

(TIF)
Text S1 Supplementary Methods for real-time RT-PCR and histopathology scoring. Table S1: Real-time RT-PCR primers used in this study.