IRF7 Regulates TLR2-Mediated Activation of Splenic CD11chi Dendritic Cells

Members of the Interferon Regulatory Factor (IRF) family of transcription factors play an essential role in the development and function of the immune system. Here we investigated the role of IRF7 in the functional activation of conventional CD11chi splenic dendritic cells (cDCs) in vitro and in vivo. Using mice deficient in IRF7, we found that this transcription factor was dispensable for the in vivo development of cDC subsets in the spleen. However, IRF7-deficient cDCs showed enhanced activation in response to microbial stimuli, characterised by exaggerated expression of CD80, CD86 and MHCII upon TLR2 ligation in vitro. The hyper-responsiveness of Irf7 −/− cDC to TLR ligation could not be reversed with exogenous IFNα, nor by co-culture with wild-type cDCs, suggesting an intrinsic defect due to IRF7-deficiency. Irf7 −/− cDCs also had impaired capacity to produce IL-12p70 when stimulated ex vivo, instead producing elevated levels of IL-10 that impaired their capacity to drive Th1 responses. Finally, analysis of bone marrow microchimeric mice revealed that cDCs deficient in IRF7 were also hyper-responsive to TLR2-mediated activation in vivo. Our data suggest a previously unknown function for IRF7 as a component of the regulatory network associated with cDC activation and adds to the wide variety of situations in which these transcription factors play a role.


Introduction
Interferon regulatory factor (IRF) 7, a member of the IRF family of transcription factors, is the 'master regulator' of type I interferon (IFN) -dependent immune responses and underpins their critical role in host defence [1]. However, IRFs are also essential for the full development of many components of the immune system [2], including dendritic cells (DCs). DCs are critical for the initiation and control of adaptive immunity [3], responding to infectious insult by changes in expression of MHCII, costimulatory molecules (notably of the B7 family) and cytokines involved in directing CD4 + and CD8 + T cell differentiation [4,5]. In the murine spleen, CD11c hi 'conventional' DCs (cDCs) can be segregated into three distinct subsets based upon surface expression of CD4 and CD8a [6,7]. Several members of the IRF family are known to be critical for the faithful development of these murine cDC subsets in vivo. Splenic CD8a + cDC development is exquisitely dependent on the expression of IRF8 [8,9] whereas differentiation of splenic CD4 + cDCs depends on IRF4 [10,11]. DN cDCs express and at least partially rely on both IRF4 and IRF8 for their full development [10]. Non-lymphoid DCs such as Langerhans cells and dermal DCs also require IRF8 expression for normal development in vivo [12]. In addition, Irf2 2/2 mice selectively lack splenic CD4 + cDCs and epidermal DC subsets [13], whereas IRF1-deficient animals have a reduced number of splenic CD8a + cDCs [14].
Although Type I IFNs are known to augment costimulatory molecule expression by DCs and monocytes in mice and humans [15,16,17,18], the specific IRFs involved are less clearly defined. PD-L1 and CD40 expression on endothelial cells is critically dependent on IRF1 [19,20], with this transcription factor also required for CD80 expression by monocytes in vitro -a situation where IRF7 appears to be redundant [21]. Nevertheless, there is some evidence supporting an interaction between IRF7 and costimulatory molecules, with an IRF7 binding site identified in the CD80 promoter that is involved in regulation of CD80 expression in LPS stimulated human monocytes [22]. Furthermore, large scale analysis of genes regulated by IRF7 in response to viral infection have identified CD80 as a potential target of this transcription factor in an in vitro system, suggesting a link between IRF7 and expression of certain costimulatory molecules [23]. However, whether IRF7 regulates the activation of splenic cDCs, and in particular their expression of costimulatory molecules, is currently unknown.
Here, we first sought to determine whether this transcription factor was required for the development of cDC subsets in the murine spleen. Furthermore, in light of the proposed link between IRF7 and the regulation of CD80 expression and possible modulation of Toll-like receptor (TLR) signalling by type I IFNs [24], we next sought to elucidate the role of this transcription factor in the functional activation of splenic cDCs in response to TLR ligation. We show that IRF7-deficient cDCs are hyper-responsive to TLR2, responding with heightened CD86 expression in vitro and in vivo, in addition to altered cytokine profiles that impact on their CD4 + T cell polarising capacity. These data implicate IRF7 as a component of a previously unknown regulatory pathway initiated in DCs during their response to microbial stimuli.

Results
Faithful Splenic cDC Development and Equivalent Steady-state TLR2 Expression in the Absence of IRF7 Mice deficient in IRF7 had a normal complement of hematopoietic cells in the liver [25], normal splenic architecture [26] and had no major alterations in the frequencies of CD4 + and CD8 + T cells, CD19 + B cells, CD11b + macrophages, Gr-1 + neutrophils and DX5 + NKp46 + NK cells, compared to wild type control mice (Supplementary Fig. S1 and data not shown). Analysis of collagenase-digested spleens from C57BL/6 and B6.Irf7 2/2 mice revealed normal populations of CD11c hi MH-CII hi cDCs in both strains (Fig. 1A). cDCs were present in comparable frequencies (1.1560.06% and 0.9660.08% of total splenocytes in C57BL/6 and B6.Irf7 2/2 mice, respectively;  Fig. 1C). All CD11c hi cDCs subsets were present in the absence of IRF7 (Fig. 1D) and were at the expected frequencies [6,7] (Fig. 1E). CD4 + cDCs comprised 42.3261.01% of CD11c hi MHCII hi cells in C57BL/6 and 40.2062.43% in B6.Irf7 2/2 mice, respectively. CD4 -CD8a -(double negative, DN) cDCs were similarly unaffected by IRF7 deficiency, making up 18.5760.62% of splenic cDCs in wildtype mice and 18.1160.81% in those lacking IRF7. The CD8a + subset developed normally in the absence of IRF7, comprising 17.1960.38% of steady state splenic cDCs in C57BL/6 mice and 16.4860.79% in B6.Irf7 2/2 mice. To determine whether IRF7 deficiency altered the capacity for TLR2 expression by cDCs, CD11c hi MHCII hi cDCs from steady-state wildtype and B6.Irf7 2/2 mice were assessed for surface expression of TLR2. cDCs from both strains expressed equivalent amounts of TLR2 ( Fig. 1F and G). Therefore, IRF7 is not a critical transcription factor for the development of splenic cDCs subsets in vivo and expression of TLR2 is not affected by its absence.

IRF7 Deficiency Leads to Dendritic Cell Hyper-activation in vitro
We next sought to determine the impact of a lack of IRF7 on the activation state of splenic cDCs in response to TLR2 ligation in vitro. We sorted CD11c hi cDCs to .98% purity from the spleens of C57BL/6 and B6.Irf7 2/2 mice and cultured them for a total of 24 hrs in the presence of the TLR2/6 agonist PAM 3 CSK 4 ( Fig. 2A). Cells were monitored at intervals by flow cytometry for activation, as measured by fold changes in surface expression of MHCII, CD80 and CD86 (Fig. 2B). Stimulation of cDCs from both strains led to their activation, as indicated by progressively increasing expression of all three surface markers. However in all cases, cDCs which lacked expression of IRF7 were hyper-activated in response to TLR2 stimulation, with significantly greater fold increases in expression of CD80 (Fig. 2C), CD86 (Fig. 2D) and MHCII (Fig. 2E) when compared with IRF7-sufficient cDCs stimulated in the same way. Such hyper-activation was not consistently observed after stimulation with agonists for TLR3, TLR4 or TLR9 ( Supplementary Fig. S2). Hyper-activation in response to TLR2 triggering was evident from 2h post treatment in the case of CD80 and MHCII expression, when assessed on the bulk population. In all experiments there was some heterogeneity in cDC upregulation of co-stimulatory molecules in response to TLR triggering, however enhanced MHCII and CD86 expression by B6.Irf7 2/2 cDCs was also apparent when gating only on the more activated population, and the proportion of activated cDCs within the whole population was consistently higher in the IRF7deficient groups at all time points after stimulation (Supplementary Fig. S3). By 24 h post activation, the increase in expression of these markers of activation on cDCs from B6.Irf7 2/2 (as measured by changes in MFI on the whole population) was ,2fold (CD80) and ,4-fold (MHCII) greater than seen with cDC isolated from C57BL/6 mice. For example, whereas TLR2 activated cDCs from C57BL/6 mice increased expression of CD80 by 3.7960.12 fold compared to resting levels by 24 h, cDCs isolated from B6.Irf7 2/2 mice increased expression of CD80 by 8.3860.18 fold after the same stimulus (p,0.01). In the case of CD86 expression, an impact of IRF7-deficiency was only evident at later time points, and was most pronounced when considering the more activated of the cDC subpopulations ( Fig. 2B and D, and Supplementary Fig. S3D). Collectively, these data indicate that in the absence of IRF7, splenic cDCs become hyper-activated in response to TLR2 stimulation in vitro, showing significantly exaggerated expression of the costimulatory molecules CD80 and CD86, as well as of MHCII.
Exaggerated CD86 Expression by Irf7 2/2 cDCs Occurs in the Presence of Wildtype cDCs and Exogenous IFNa IRF7 regulates the production of Type I IFNs and potentially other soluble factors that could normally act in a negative feedback loop to self-limit activation. We tested this hypothesis in two ways. First, we first sorted CD11c hi cDCs from B6J.CD45.1 and congenic (CD45.2) B6.Irf7 2/2 mice (Fig. 3A). Sorted cDCs were then co-cultured at an approximately 50:50 ratio (Fig. 3B) and stimulated with PAM 3 CSK 4 , as before. Gating on CD45.1 or CD45.2 during flow cytometric analysis allowed an assessment of activation (measured as the fold increase in co-stimulatory molecule expression comparing stimulated to unstimulated cells) on wildtype and IRF7-deficient cDCs cultured in the same microenvironment. Even in the presence of wildtype cells, cDCs from IRF7-deficient mice remained hyper-responsive to TLR2 activation, as assessed for CD80, CD86 and MHCII expression ( Fig. 3C and data not shown). Next, we directly examined the potential role of Type I IFN as a negative regulator of CD86 expression. Addition of a biologically active amount of exogenous IFNa [26] potentiated the response to TLR2 ligation in wildtype cDC, inducing an increase in surface expression approximately twice that observed by TLR2 stimulation alone ( Fig. 3C and D). Importantly, IRF7-deficient cDCs retained their exaggerated response to TLR2 stimulation even in the presence of exogenous IFNa, as measured by CD86 (Fig. 2D), as well as CD80 and MHCII expression (data not shown). Therefore, addition of neither wildtype cDCs nor exogenous IFNa was able to prevent the hyper-activation of B6.Irf7 2/2 cDCs in response to TLR2 stimulation, suggesting that IRF7 may act as a cell intrinsic endogenous regulator of TLR-mediated activation in splenic cDCs.

IRF7 Modulates TLR-mediated Cytokine Production by Splenic cDCs ex vivo
To extend this functional assessment of the role of IRF7 in cDC activation, CD11c hi MHCII hi splenic cDCs were sorted from naïve C57BL/6 and B6.Irf7 2/2 mice and cultured for 24 IL-12p70 production by cDCs due to spontaneous maturation in culture was impaired when cDCs lacked IRF7 (17.163.86 pg/ml vs. 3.9460.82 pg/ml in C57BL/6 and B6.Irf7 2/2 cDCs respectively; p,0.01). In response to stimulation with LPS, IRF7-deficiency severely impaired IL-12p70 production by cDCs (127.06619.08 pg/ml vs. 16.5165.85 pg/ ml after 24 h culture; p,0.01). The addition of exogenous IFNa affected neither the response of wild type cDCs to LPS stimulation nor the defective response of B6.Irf7 2/2 cDCs (Fig. 4A). Similarly impaired IL-12p70 production was also observed in IRF7-deficient cells cultured with PAM 3 CSK 4 and again exogenous IFNa did not compensate for this deficiency. In contrast to the defective IL-12p70 production observed in IRF7-deficient cDCs, culture supernatants of cDCs isolated from B6.Irf7 2/2 mice contained significantly higher levels of IL-10 when stimulated with either LPS or PAM 3 CSK 4 (Fig. 4B).
As with TLR induced IL-12p70, we found no evidence to suggest that exogenous IFNa affected production of IL-10 in either strain of mice. Therefore, a lack of IRF7 in splenic cDCs leads to reciprocal effects on two key cytokines involved in T cell activation and differentiation and produced by cDCs in response to TLR2 and TLR4 ligands in vitro.

IL-10-dependent Impairment in Th1 Polarisation by IRF7deficient cDCs in vitro
To determine the functional significance of the altered activation state and cytokine profile of IRF7-deficient cDCs after TLR2 triggering, we employed a TCR transgenic co-culture approach (Fig. 5). CD11c + cells were enriched from spleens of C57BL/6 or B6.Irf7 2/2 mice, stimulated (or not) with PAM 3 CSK 4 and cultured with CFSE-labelled CD4 + OTII.Rag2 2/2 splenic T cells. After 7 days, cultures were restimulated and OTII T cells assessed for proliferation and cytokine production by flow cytometry. IRF7-deficient cDCs stimulated significantly more proliferation in OTII cells than cDC from wild-type mice, likely reflecting a degree of spontaneous activation during culture that was regulate by IRF7, but this difference was lost after in vitro TLR2 stimulation, (data not shown). Similarly, resting B6.Irf7 2/2 cDCs had an enhanced capacity to drive OTII T cells toward a Th1 phenotype compared to wild-type cells (Fig. 5A&C). However, after cDCs were stimulated with PAM 3 CSK 4 , cells lacking IRF7 exhibited a significantly impaired capacity to drive Th1 polarisation and IFNc production by OTII T cells, when compared to those from wildtype mice (p,0.001, Fig. 5B&C).

In vivo Administration of PAM 3 CSK 4 to Microchimeric Mice Reveals Enhanced CD86 Expression on IRF7deficient cDCs
We next sought to address the role of IRF7 in regulating cDC activation by TLR2 in vivo. To investigate this, we generated microchimeric mice in which a minor population of donor derived hematopoietic cells develop under physiological conditions. B6J.CD45.1 mice were injected i.p. with the stem cell depleting drug busulfan [27,28], and the next day bone marrow cells from congenic CD45.2 + donor C57BL/6 or B6.Irf7 2/2 mice were adoptively transferred (Fig 6A). After 7-14 days to allow engraftment and the development of donor derived cDCs, microchimeric mice were injected with PBS or PAM 3 CSK 4 i.v. and 24h later, cDCs of recipient (i.e. untouched endogenous cDCs) and donor origin were assessed for expression of CD80 and CD86 by flow cytometry. Spleens of microchimeric mice contained CD11c hi MHCII hi cDCs derived from both host (CD45.1) and donor (CD45.2) genotypes (Fig. 6B). There was no difference in the ability of C57BL/6 or B6.Irf7 2/2 bone marrow to engraft, with equivalent frequencies of endogenous (,70-80%) vs. chimeric (,20-30%) cDCs regardless of which strain of donor mice was used (data not shown). For each population of cDCs, we calculated the fold change in expression for CD80 or CD86 in mice that had received PAM 3 CSK 4 compared to those that had received saline alone injection. Although in vivo administration of PAM 3 CSK 4 increased the expression of CD80 on the endogenous cDCs by approximately two fold compared to untreated mice, there was no significant difference in extent of upregulation of CD80 between wildtype and IRF7-deficient chimeric cDCs (Fig. 6C). As expected, in vivo administration of PAM 3 CSK 4 -enhanced CD86 expression on the endogenous population of cDCs (by ,3.5 fold), and expression was als increased to the same extent on wild type chimeric cDCs. In contrast, chimeric IRF-7-deficient cDCs displayed a hyperactivated phenotype, with enhanced CD86 expression relative to the endogenous population of cDCs (Fig. 6D), of similar magnitude to that seen in vitro (Fig. 2D). Hence, cDCs lacking IRF7 can be hyper-activated either in vitro or in vivo (where they have developed and been activated in a fully IRF7-sufficient environment), but environment (i.e. in vitro or in vivo) determines the extent to which CD80 and CD86 are affected.

Discussion
This study reveals a novel role for IRF7 in regulating TLR2induced costimulatory molecule expression and cytokine production by splenic cDCs in vitro and in vivo. Unlike many other members of this transcription factor family, IRF7 is redundant with respect to the in vivo development of cDC subsets. However, cDCs lacking IRF7 respond to TLR2 activation with exaggerated costimulatory molecule expression and altered cytokine production in vitro, impacting on their capacity to affect CD4 + T cell polarisation. Furthermore, administration of PAM 3 CSK 4 to microchimeric mice indicates a key role for IRF7 in the regulation of CD86 expression by splenic cDCs in vivo.
IRF7 appears to play a limited role in the development of cDCs subsets and other splenic immune cell types, supporting previous data showing normal cell distribution in the spleens of B6.Irf7 2/2 mice by immunohistochemistry [26] and in the livers of IRF7deficient mice by flow cytometry [25]. This is in stark contrast to mice lacking expression of other IRF family members, with normal cDC, CD8a + T cell [29,30] and NK cell [31,32,33] development relying on the expression of several IRFs. These data suggest either that compensatory mechanisms for immune cell development exist in the absence of IRF7, or that the functions of this transcription factor are more restricted than other members of the same family.
As IRF7 is critical for the optimal induction of type I IFN expression, it was surprising that a deficiency in this transcription factor led to alterations in the response of cDCs to ligation of TLR2; signalling through which was, until recently, not thought to lead to type I IFN expression [34]. However it is now becoming clear that TLR2 signalling can lead to type I IFN production under some conditions. This has been observed during the response of mice to infection with vaccinia virus, where a subset of Ly6C hi inflammatory monocytes produces type I IFN in an IRF3 and IRF7-dependent manner after TLR2-mediated recognition of viral ligands [35]. This was proposed to be restricted to both inflammatory monocytes and virus-derived stimuli. However, more recent studies have revealed a capacity for type I IFN production in other cell types and in response to diverse TLR2 ligands, including PAM 3 CSK 4 [36,37]. This is thought to be dependent upon the localisation of TLR2 to endolysosomal membranes, with subsequent signalling from this compartment leading to IFNa/b production -more akin to the mechanisms employed by TLR3, TLR7 and TLR9 [38,39]. The location of TLR2 to endosomal compartments and subsequent IFNa/b production may, as with TLR4, be dependent upon TRAF3, as forced localisation of TRAF3 to the plasma membrane enables IFNb production after PAM 3 CSK 4 stimulation, rather than the classical plasma membrane-restricted signalling that initiates proinflammatory cytokine production in response to TLR2 [40]. However it is still unclear as to the precise molecular mechanisms leading to TLR2-induced IFNa/b production.
Although IFNa production is critically dependent upon IRF7, IFNb production in response to LPS stimulation occurs at normal levels in IRF7-deficient cells [1]. However, the recent studies showing type I IFN production in response to endosomal TLR2 signalling all revealed a key requirement for IRF7 in the induction of IFNb [36,37,41]. Therefore, in addition to a lack of IFNa, cDCs stimulated with TLR2 ligands would also likely lack the capacity to produce IFNb. Contradictory findings have been reported regarding the role of type I IFNs in regulating CD80/86 expression. IFNb has been shown to be required for maximal costimulatory molecule expression on peritoneal macrophages in response to LPS stimulation [16], whereas splenic cDCs which are deficient in the IFNab receptor (IFNabR) and thus cannot respond to either type I IFN, show greatly enhanced expression of CD80 and CD86 in response to infection with Listeria monocytogenes in vivo [42]. The data presented here show that IRF7 deficiency led to exaggerated co-stimulatory molecule expression on cDCs after TLR2 signalling. In our hands, we have not been able to manipulate this response with exogenous IFNa in vitro or by placing IRF7-deficient cDCs within a type I IFN (and IRF7) sufficient host. Hence, our data support a model whereby IRF7 acts as a cell intrinsic regulator of cDCs activation independent of the availability of exogenous type I IFNs.
Tripartite motif-containing (TRIM) proteins comprise a family of type I IFN-inducible factors with diverse roles in the immune system [43]. Of particular relevance is TRIM30a, where bioinformatic analysis suggests the presence of a putative IRF7 binding sequence in its promoter region (Phillips and Kaye, unpublished). This protein acts to restrict NF-kB -mediated signalling downstream of TLR ligation by interacting with TAK1 and degrading TAB2 and TAB3 [44]. Interestingly TAB2 itself also contains a putative IRF7-binding element (Phillips and Kaye, unpublished), suggesting that IRF7 may be capable of regulating the activation of NF-kB directly and through the actions of TRIM30a. As enhanced costimulatory molecule expression is dependent on NF-kB [45,46], it is possible that when this pathway is not correctly regulated, as may be the case in IRF7-deficient Figure 5. IL-10 dependent impairment in Th1 polarisation by IRF7-deficient cDCs. Splenic CD11c hi cDCs were purified from C57BL/6 and B6.Irf7 2/2 mice and cultured for 3 hours in the presence/absence of 10 mg/ml PAM 3 CSK 4 . DCs were cultured at a 1:2 ratio with CFSE-labeled CD4 + OTII.Rag2 2/2 cells for 7 days. TCRb + CD4 + T cells were assessed for production of IFNc after in vitro restimulation from cultures containing unstimulated DCs (A&C) or DCs pre-activated for 3 hours with PAM 3 CSK 4 (B&C). Dot plots in A and B are representative, C shows mean percentage of IFNc + OTII.Rag2 2/2 T cells in indicated culture conditions 6 SEM. Each condition contained 3 replicate cultures, using DCs purified and pooled from 8-9 individual mice of each genotype. ** = p,0.01 *** = p,0.001. doi:10.1371/journal.pone.0041050.g005 cDCs, that this results in a failure to regulate CD86 expression and thus leads to enhanced accumulation of this costimulatory molecule on cDCs. Further work will be required to confirm whether IRF7 does directly interact with TRIM30a, and to provide a molecular basis for our current observations.
The altered cytokine profile of cDCs lacking IRF7 and stimulated in vitro suggests an uncoupling of costimulatory molecule expression and cytokine production in response to TLR stimulation. In particular, the greatly reduced production of IL-12p70 by IRF7-deficient cDCs in response to TLR2 signalling was surprising, given the exaggerated expression of costimulatory molecules occurring simultaneously on these cells. This was unexpected, as previous reports have described exaggerated inflammatory cytokine production in the absence of IRF7, with MCMV infected B6.Irf7 2/2 mice having significantly elevated levels of serum IL-12p70 compared to wildtype animals [47]. Exacerbated pro-inflammatory cytokine production has also been reported in IRF7-deficient mice in a model of liver pathology, although this appears to be due to a failure in the IFNa-mediated induction of soluble IL-1Ra rather than any direct regulation of cytokine gene expression by IRF7 [41]. However, these studies did not address cytokine production from defined cell populations, and so the enhanced levels of systemic cytokines could be due to differential responses of distinct immune cell populations when deficient in IRF7. Of note, IRF1 and IRF8 cooperate to generate maximal IL-12 production [48,49,50,51,52,53,54], and IRF7 has been shown to interact with IRF8 [55] and is predicted to bind both IRF1 and IRF8 [23,56]. Hence, IRF7-deficiency may affect IL-12 production via the direct and indirect activation of other IRFs.
Alongside their defective production of IL-12p70, cDCs deficient in IRF7 produced elevated levels of IL-10 in response to TLR ligation, sufficient to inhbit TH1 polarisation in vitro. This is similar to IRF1-deficient splenic DCs, and IRF5-deficient peritoneal macrophages [57] which show impaired IL-12 and enhanced IL-10 production in vitro [14]. However, splenic DCs from Irf1 2/2 mice have limited upregulation of costimulatory molecule expression in response to TLR ligation: a phenotype distinct from PAM 3 CSK 4 -stimulated splenic cDCs lacking IRF7 and indicating that differential mechanisms underlie these broadly similar observations.
In summary, this data reveals a novel role for IRF7 in the regulation of TLR2-induced costimulatory molecule expression by splenic cDCs in vitro and in vivo. This appears to be distinct from the regulation of pro-inflammatory cytokine production, as cDCs displaying exaggerated costimulatory molecule expression in vitro had a significantly impaired capacity for IL-12p70 production. In contrast IL-10 production was enhanced, indicating divergent effects of IRF7 on the regulation of cytokine production by cDCs. Although the molecular mechanisms remain to be experimentally determined, preliminary bioinformatic data suggest that interactions between IRF7 and regulators of TLR-induced NF-kB expression may underpin the altered expression of costimulatory molecules in IRF7-deficient cDCs.

Dendritic Cell Isolation
Splenic tissue was dissociated mechanically using a scalpel and digested in RPMI-1640 supplemented with 0.2mg/ml collagenase type IV/DNAse1 mix (Worthington Biochemical, NJ, USA) for 30 minutes at room temperature. A single cell suspension was generated and CD11c + cells were enriched using a modified MACS magnetic separation protocol as previously described [58]. CD11c hi or CD11c hi MHCII hi cDCs were sorted to ,98-99% purity on a BeckmanCoulter MoFlo cell sorter.

cDC Cytokine Production
CD11c hi MHCII hi cDCs were sorted from C57BL/6 and B6.Irf7 2/2 mice as described, plated in triplicate in complete RPMI at 1610 6 cells/ml and stimulated as indicated with 1 mg/ml LPS or 10 mg/ml PAM3CSK4 (Invivogen) for 24 hours, in the presence/absence of 1000 U/ml IFNa. Supernatants were assessed using a Quantikine ELISA (R&D Systems, Minneapolis, USA) for levels of IL-12p70 and IL-10.

Statistical Analysis
Statistical analysis was performed using a student's t test where p,0.05 was considered significant.