Francisella tularensis Elicits IL-10 via a PGE2-Inducible Factor, to Drive Macrophage MARCH1 Expression and Class II Down-Regulation

Francisella tularensis is a bacterial pathogen that uses host-derived PGE2 to subvert the host's adaptive immune responses in multiple ways. Francisella-induced PGE2 acts directly on CD4 T cells to blunt production of IFN-γ. Francisella-induced PGE2 can also elicit production of a >10 kDa soluble host factor termed FTMØSN (F. tularensis macrophage supernatant), which acts on IFN-γ pre-activated MØ to down-regulate MHC class II expression via a ubiquitin-dependent mechanism, blocking antigen presentation to CD4 T cells. Here, we report that FTMØSN-induced down-regulation of MØ class II is the result of the induction of MARCH1, and that MØ expressing MARCH1 “resistant” class II molecules are resistant to FTMØSN-induced class II down-regulation. Since PGE2 can induce IL-10 production and IL-10 is the only reported cytokine able to induce MARCH1 expression in monocytes and dendritic cells, these findings suggested that IL-10 is the active factor in FTMØSN. However, use of IL-10 knockout MØ established that IL-10 is not the active factor in FTMØSN, but rather that Francisella-elicited PGE2 drives production of a >10 kDa host factor distinct from IL-10. This factor then drives MØ IL-10 production to induce MARCH1 expression and the resultant class II down-regulation. Since many human pathogens such as Salmonella typhi, Mycobacterium tuberculosis and Legionella pneumophila also induce production of host PGE2, these results suggest that a yet-to-be-identified PGE2-inducible host factor capable of inducing IL-10 is central to the immune evasion mechanisms of multiple important human pathogens.


Introduction
F. tularensis is a bacterial pathogen that infects macrophages and uses host-derived PGE 2 to enhance bacterial growth and subvert the adaptive immune response [1]. Other clinically relevant human pathogens such as Salmonella typhi [2,3], Mycobacterium tuberculosis [4] and Legionella pneumophila [5] also elicit host PGE 2 , and thus may employ a similar strategy to promote infection. Francisella-induced PGE 2 acts directly on CD4 T cells, reprogramming them to restrict production of anti-Francisella cytokines such as interferon-c (IFN-c) [1]. Moreover, blockade of Francisella-induced PGE 2 production in vivo allows for more robust IFN-c production and better control of pulmonary Francisella infection in mice [6].
Francisella-induced PGE 2 , which is elicited by either the live vaccine strain (LVS) or human virulent SchuS4 strain of Francisella, also acts in an autocrine/paracrine fashion to drive production of a soluble MØ factor that elicits the ubiquitindependent down-regulation of MHC class II and CD86 molecules expressed by IFN-c activated MØ [7] (Figure 1). The factor within this F. tularensis MØ supernatant (termed FTMØSN, pronounced foǒt-mō-sin) drives MØ class II down-regulation by eliciting ubiquitination of the class II cytoplasmic tail, which results in class II trafficking to degradative intracellular compartments [7,8]. The resulting class II negative MØ are thus greatly impaired in their ability to present antigens to CD4 T cells.
Until now, the identity of the active factor in FTMØSN and the mechanism of class II ubiquitination were unknown. Here, we establish that the mechanism of FTMØSN induced class II ubiquitination is via up-regulation of the ubiquitin ligase MARCH1. Studies with IL-10 knockout (IL-10D) MØ revealed that Francisella (via elicitation of PGE 2 ) is inducing production of a yet-to-be identified factor that is able to elicit MØ IL-10 production to then induce MØ class II down-regulation via a MARCH1-dependent mechanism.

Ethics Statement
Mice were house and used in strict accordance to the guidelines established by the Albany Medical College Institutional Animal Care and Use Committee. Animal protocols were reviewed and approved by the Albany Medical College Institutional Animal Care and Use Committee (Protocol # 603276).

Measurement of IL-10
IL-10 was measured by cytometric bead array (CBA, BD Biosciences, San Jose, CA catalog number 558300) IL-10 Flex Set according to the manufacturer's directions.

Treatment of BMMØ with FTMØSN
FTMØSN was generated by exposure of producer BMMØ to F. tularensis LVS at an multiplicity of infection of 100:1 (Ft:MØ) as previously described [7]. FTMØSN was cleared of any bacteria and cellular debris by centrifugation and 0.2 mm filtration. Reporter MØ were exposed to 50% FTMØSN or control supernatants for 20-24 hours before washing and harvesting to generate whole cell lysates. The anti-IL-10 receptor blocking mAb 1B1.3A (BioXCell # BE005) was used at the indicated concentration.

Western Blot Analysis of MØ MHC Class II Levels
Whole cell lysates of treated or control BMMØ were analyzed for total MHC class II levels by SDS-PAGE and anti-class II b chain western blot as previously reported [7]. Blots were subsequently probed for GAPDH as a loading control.

RT-PCR Analysis of MARCH1 mRNA Levels
Total RNA was prepared using TRIzol (Invitrogen) and treated with DNase I to remove genomic DNA. RNA was converted to cDNA using standard procedures and reagents from Invitrogen. Real-time PCR primers used to amplify MHC class II and MARCH1 have been described previously [12] and primers for GAPDH were from Qiagen (QuantiTect primer #QT01658692). Real-time PCR was performed using an ABI Prism 7900HT Sequence detection system and QuantiTect SYBR green PCR kit (QIAGEN) according to the manufacturers instructions. In all experiments the normalized Ct values for the FTMØSN-treated macrophages were expressed relative to the normalized Ct values for the mock supernatant-treated macrophages.

Results
The Ubiquitin Ligase MARCH1 is Responsible for FTMØSN-induced Down-regulation of Macrophage Class II Expression Previous work established that upon Francisella infection, MØ make PGE 2 , which works in an autocrine/paracrine fashion to elicit the production of a .10 kDa soluble factor termed FTMØSN. FTMØSN drives the ubiquitin-dependent downregulation of MHC class II expressed by IFN-c pre-activated reporter MØ (Figure 1 and [7]). The goal of this report is to define the mechanism of class II down-regulation and characterize the active factor in FTMØSN.
One mechanism for the post-translational control of MHC class II expression in MØ and dendritic cells (DC) is the regulated expression of the ubiquitin ligase MARCH1, which ubiquitinates the cytoplasmic domain of class II molecules resulting in their intracellular sorting to degradative intracellular compartments [13,14]. To determine if MARCH1 might be involved in FTMØSN-elicited down-regulation of class II, the effect of FTMØSN treatment on reporter MØ MARCH1 mRNA levels was determined ( Figure 2). Overnight treatment of IFN-c preactivated reporter MØ with FTMØSN results in the up-regulation of MARCH1 mRNA, but no significant change in the level of Ab class II mRNA expression, consistent with a post-transcriptional mechanism for regulating MHC class II expression. The current lack of reliable anti-MARCH1 antibodies precludes direct analysis of the impact of FTMØSN treatment on MARCH1 protein levels. Nevertheless, the results presented in Figure 2 suggest that FTMØSN-induced up-regulation of MARCH1 may be responsible for the ubiquitin-dependent down-regulation of MHC class II molecules observed in FTMØSN treated MØ [7].
To further investigate this possibility, the level of class II expression by FTMØSN-treated wild type and MARCH1D reporter MØ was determined ( Figure 3). The results show that MARCH1D MØ are completely refractory to the class II downregulatory effects of FTMØSN, confirming the notion that FTMØSN-induced MARCH1 drives class II down-regulation. To further confirm this scenario and determine if class II is a direct target for MARCH1-mediated ubiquitination, the impact of FTMØSN treatment on class II expression by reporter MØ expressing class II molecules lacking the single cytoplasmic lysine ubiquitination site (i.e., Ab K225R [11]) was examined ( Figure 3). Consistent with the idea that class II is the direct target for MARCH1-mediated ubiquitination, MARCH1-positive MØ expressing these non-ubiquitinatable class II molecules are refractory to FTMØSN-induced MHC class II down-regulation. Taken together, the results presented in Figures 2 and 3 establish that MARCH1-mediated ubiquitination of the cytoplasmic tail of class II b chains is directly responsible for the FTMØSN-induced down-regulation of reporter MØ MHC class II expression [7].

FTMØSN Contains a Soluble Factor the Induces MØ IL-10 Production
IL-10 is the only cytokine that has been demonstrated to induce MARCH1 expression in both DC and human monocytes [13,14]. FTMØSN contains a soluble factor of .10 kDa molecular mass that elicits the ubiquitin-dependent down-regulation of ''reporter'' MØ MHC class II molecules [7]. In this report, we establish that the active factor in FTMØSN is distinct from IL-10, but that ''reporter'' MØ class II downregulation is driven by induction of reporter MØ IL-10 production, which drives MARCH1 expression and class II ubiquitination. This means that FTMØSN contains a factor distinct from IL-10, which is induced by F. tularensis-elicited PGE 2 and which is able to drive reporter MØ IL-10 production. doi:10.1371/journal.pone.0037330.g001 Moreover, PGE 2 has been reported to induce MØ production of IL-10 [15] and FTMØSN contains some level of IL-10 (see below). Together, this suggests that IL-10 is likely the PGE 2induced factor in FTMØSN responsible for inducing MØ MARCH1 expression. To test this possibility, the ability of FTMØSN generated from both wild type and IL-10D ''producer'' MØ ( Figure 1) to drive MHC class II down-regulation by wild type reporter MØ was determined ( Figure 4). Surprisingly, FTMØSN from both wild type and IL-10D producer MØ drive the same degree of reporter MØ MHC class II down-regulation, demonstrating that the active factor in FTMØSN is not IL-10 (The absence of IL-10 in FTMØSN generated by IL10D BMMØ was confirmed by CBA, see below).
The ability of FTMØSN from IL-10D producer MØ to drive MHC class II down-regulation either means that FTMØSN contains a factor distinct from IL-10 that is able to elicit MARCH1 expression and drive class II down-regulation or that the active factor in FTMØSN acts by inducing IL-10 production by reporter MØ which then induces MARCH1 expression and class II down-regulation. To determine which of these scenarios is correct, BMMØ from IL-10D mice were used as reporter MØ ( Figure 5). IL-10D reporter MØ fail to down-regulate MHC class II expression in response to either wild type or IL-10D FTMØSN, whereas wild type reporter MØ down-regulate class II in response to both forms of FTMØSN. These results indicate that IL-10 produced by reporter MØ is central to FTMØSN's ability to induce class II down-regulation.
Consistent with the possibility that FTMØSN is acting through induction of IL-10, analysis of both FTMØSN and secondary supernatants from FTMØSN-treated reporter MØ revealed that these secondary supernatants contain approximately 10 fold more IL-10 than FTMØSN itself ( Figure 6). Moreover, treatment of reporter MØ with high dose recombinant IL-10 (100 ng/ml) results in down-regulation of MØ class II (Figure 7), whereas lower doses of recombinant IL-10 had a limited effect on MØ class II   levels (not shown). Finally, treatment of wild type reporter MØ with a blocking anti-IL-10 receptor mAb blocks FTMØSNinduced class II down-regulation ( Figure 8). Taken together, these results indicate that FTMØSN contains a PGE 2 -inducible factor distinct from IL-10 that acts to induce IL-10 release by reporter MØ. This reporter MØ-produced IL-10 then acts in an autocrine/ paracrine fashion to induce MARCH1 expression to drive ubiquitin-dependent MHC class II down-regulation.

Discussion
Previous studies have established that Francisella tularensisinfected MØ produce PGE2, which can both alter CD4 T cell cytokine production and drive the production of a soluble factor (i.e., FTMØSN) that elicits the ubiquitin-dependent downregulation of MØ MHC class II expression [7]. Since other clinically relevant human pathogens such as Salmonella typhi [2,3], Mycobacterium tuberculosis [4] and Legionella pneumophila [5] also elicit host PGE2, it was of interest to determine the mechanism of FTMØSN-induced MØ class II down-regulation.
Post-translation control of MHC class II expression in DC and MØ can be driven by MARCH1-mediated ubiquitination of the cytoplasmic tail of class II molecules [13,14], which results in altered intracellular class II trafficking away from recycling endosomes and toward degradative compartments [8]. To date, IL-10 is the only cytokine reported to induce antigen presenting cell MARCH1 expression and class II down-regulation [13,14]. Since PGE 2 can induce MØ IL-10 [13], we explored the possibility that a PGE 2 to IL-10 to MARCH1 axis is responsible for the FTMØSN-induced down-regulation of reporter MØ class II expression (Figure 1).
Analysis of reporter MØ exposed to FTMØSN revealed that they up-regulate MARCH1 mRNA expression. In addition, MARCH1D MØ fail to down-regulate class II expression in response to FTMØSN. Finally, MØ expressing class II molecules refractory to the action of MARCH1 (i.e., I-A b chain K225R [11]) fail to down-regulate class II in response to FTMØSN. Taken together, these results establish that the active factor in FTMØSN drives the ubiquitin-dependent down-regulation of MØ class II by increasing MØ MARCH1 expression, which results in the ubiquitination of the cytoplasmic domain of class II and its post-endocytic sorting into a degradative compartment.
To address the question of whether the MARCH1-inducing factor in FTMØSN is IL-10, IL-10D producer MØ were exposed  to F. tularensis and the resulting FTMØSN tested for activity. Surprisingly, FTMØSN from IL-10 knockout MØ is able to drive reporter MØ class II down-regulation, ruling out IL-10 as the active factor in FTMØSN. In contrast, IL-10D reporter MØ fail to down-regulate class II in response to FTMØSN (both wild type and IL-10D FTMØSN). These results indicate that PGE 2 is inducing IL-10 via an indirect mechanism. More precisely, PGE 2 is inducing the production of a .10 kDa soluble factor distinct from IL-10 (i.e. the active factor in FTMØSN), which then acts to induce reporter MØ IL-10 production, driving MARCH1 expression and class II down-regulation. While previous reports have demonstrated that PGE 2 can induce IL-10 production by MØ and DC [15,16], the molecular mechanism of IL-10 induction remains undetermined. These current findings suggest a complex mechanism, minimally involving the production of at least one yet to be defined soluble factor. These results also suggest that Francisella might act to control MØ responsiveness to this undefined soluble factor, since Francisella-infected producer MØ (which are bathed in FTMØSN) do not make as much IL-10 at non-infected reporter MØ exposed to FTMØSN. This possibility will be addressed in future studies.
Presently, the identity of the IL-10-inducing active factor in FTMØSN remains unclear. Preliminary experiments using either knockout MØ and/or neutralizing mAbs have ruled out TGF-b, IL-6, VEGF, MIP-1a and leukemia inhibitory factor. Moreover, generation of supernatants from Francisella-infected gene knockout producer MØ has established that production of FTMØSN activity is independent of TLR-2 (which recognizes F. tularensisderived lipoproteins [17]) and MØ caspase-1 activity (which drives IL-1b production/release). Therefore, it is likely that a combined genomics/proteomics approach will be necessary to conclusively identify and characterize the PGE 2 -induced active factor in FTMØSN that elicits MØ IL-10 expression to then drive MØ class II down-regulation.
MHC class II restricted antigen presentation by MØ is central to immune defense against many human pathogens and tumors. Moreover, many human pathogens and tumors elicit the production of significant levels of PGE 2 . The results presented in this report establish that PGE 2 drives the production of a soluble MØ factor distinct from IL-10 that induces MØ IL-10 production to drive MARCH1 expression, culminating in down-regulation of MØ class II molecules and limiting the ability of these cells to participate in an adaptive immune response. Future identification of the active factor in FTMØSN will provide greater insight into this mechanism of immune suppression common to both infectious agents and tumors, and will provide an additional target for new interventional therapies.