Nfkb1 Inhibits LPS-Induced IFN-β and IL-12 p40 Production in Macrophages by Distinct Mechanisms

Background Nfkb1-deficient murine macrophages express higher levels of IFN-β and IL-12 p40 following LPS stimulation than control macrophages, but the molecular basis for this phenomenon has not been completely defined. Nfkb1 encodes several gene products including the NF-κB subunit p50 and its precursor p105. p50 is derived from the N-terminal of 105, and p50 homodimers can exhibit suppressive activity when overexpressed. The C-terminal region of p105 is necessary for LPS-induced ERK activation and it has been suggested that ERK activity inhibits both IFN-β and IL-12 p40 following LPS stimulation. However, the contributions of p50 and the C-terminal domain of p105 in regulating endogenous IFN-β(Ifnb) and IL-12 p40 (Il12b) gene expression in macrophages following LPS stimulation have not been directly compared. Methodology/Principal Findings We have used recombinant retroviruses to express p105, p50, and the C-terminal domain of p105 (p105ΔN) in Nfkb1-deficient murine bone marrow-derived macrophages at near endogenous levels. We found that both p50 and p105ΔN inhibited expression of Ifnb, and that inhibition of Ifnb by p105ΔN depended on ERK activation, because a mutant of p105ΔN (p105ΔNS930A) that lacks a key serine necessary to support ERK activation failed to inhibit. In contrast, only p105ΔN but not p50 inhibited Il12b expression. Surprisingly, p105ΔNS930A retained inhibitory activity for Il12b, indicating that ERK activation was not necessary for inhibition. The differential effects of p105ΔNS930A on Ifnb and Il12b expression inversely correlated with the function of one of its binding partners, c-Rel. This raised the possibility that p105ΔNS930A influences gene expression by interfering with the function of c-Rel. Conclusions These results demonstrate that Nfkb1 exhibits multiple gene-specific inhibitory functions following TLR stimulation of murine macrophages.


Introduction
The transcription factor NF-kB plays a central role in regulating innate immune gene expression [1]. While NF-kB is generally considered a pro-inflammatory factor, LPS induces significantly higher levels of the genes for IFN-b (Ifnb) and IL-12 p40 (Il12b) in macrophages derived from mice with targeted deletions in Nfkb1, the gene coding for the p50 subunit of NF-kB, than in WT macrophages [2][3][4][5][6][7][8]. Thus Nfkb1 is an inhibitor of key immunoregulatory genes, and this is manifest in the observation that Nfkb1 2/2 mice are highly susceptible to microflora-induced colitis [9]. However, the molecular basis for selective inhibition of Ifnb and Il12b by Nfkb1 has not been completely defined.
Nfkb1 has multiple gene products. Its full-length gene product p105 is a bipartite molecule consisting of an N-terminal Rel homology domain and a C-terminal ankyrin repeat domain [10][11][12][13][14][15]. Constitutive processing of p105 by the proteosome removes the C-terminal domain of p105 to generate p50 monomers, which then homodimerize or heterodimerize with other NF-kB family members [16]. In contrast to the constitutive processing of p105 to p50, LPS induces IkB kinase (IKK)-dependent complete degra-dation of p105 [17][18][19]. IKK mediates the phosphorylation of p105 at serine 930, which is located within a C-terminal amino acid sequence partially homologous to the N-terminal IKKdependent phospho-acceptor sites in IkB-a, and mutations of S930 inhibit IKK-mediated inducible degradation of p105 [18,20,21].
It has previously been suggested that p50 homodimers are inhibitory based on their ability to prevent binding of more active NF-kB dimers to kB sites necessary for inflammatory gene expression [22,23]. Nevertheless, this does not explain why the absence of Nfkb1 selectively affects expression of Ifnb and Il12b. Recently, it has been proposed that p50 binds to a G-rich sequence within the Ifnb interferon stimulated response element (ISRE) and interferes with recruitment of IRF-3. This potentially explains the selective increase in IFN-dependent gene expression observed in Nfkb1 2/2 macrophages [3], although it is not yet clear whether expression of p50 alone is sufficient to inhibit Ifnb expression in Nfkb1 2/2 macrophages. Further, while expression of Il12b does not appear to be dependent on IRF-3 [24], it does contain an ISRE in its proximal promoter region [25,26], but it has not yet been determined whether p50 is sufficient to inhibit LPS-induced Il12b.
In contrast to the suggestion that p50 is responsible for Nfkb1mediated inhibition of Ifnb, we have proposed that the ability of the C-terminal region of p105 to facilitate ERK activation is responsible for inhibition of LPS-induced Ifnb expression [8]. ERK inhibitors enhance LPS-induced expression of Ifnb and Il12b in macrophages [27][28][29]. The C-terminal portion of p105 binds to and stabilizes Tpl-2, a MAP3K that is essential for ERK activation [17,30], and as a result, LPS-induced ERK activation is blocked in Nfkb1 2/2 bone marrow-derived macrophages (BMDM) [19]. It has been suggested that the basis of ERK-mediated inhibition is its ability to enhance induction of c-Fos, which inhibits both Ifnb and Il12b expression [27][28][29]. We have demonstrated that expression of a C-terminal fragment of p105 that does not contain the Rel homology domain (p105DN) rescues LPS-induced ERK activation and inhibits LPS-induced Ifnb expression [8]. However, whether the ability of p105DN to inhibit Ifnb expression depends upon ERK activation and whether p105DN is able to inhibit LPSinduced Il12b expression has not been determined.
To further delineate the molecular basis for Nfkb1-mediated inhibition, we have directly compared the ability of p105, p50, and p105DN to inhibit LPS-induced expression of Ifnb and Il12b in Nfkb1 2/2 murine bone marrow-derived macrophages. In addition, we have used a mutant of p105DN (p105DNS930A) that is unable to support ERK activation, to determine whether the inhibitory function of p105DN depends on its ability to activate ERK.

Results
BMDM were infected with recombinant retroviruses that expressed either HA tagged full-length p105, p50 (amino acids 1-433), a C-terminal fragment of p105 which lacks the N-terminal Rel Homology Domain (p105DN, amino acids 434-971), or a mutant of p105DN in which S930, one of the serines in the Cterminal phosphorylation site necessary for IKK-dependent degradation and ERK activation, was mutated to alanine (p105DNS930A) [17,20] (Fig. 1). Western blotting demonstrated expression of HA tagged proteins of appropriate sizes ( Fig. 2A), and as expected we observed both p50 and p105 in macrophages infected with the vector expressing p105. Western blotting with an N-terminal p105 antibody indicated that levels of p50 and p105 in virally infected Nfkb1 2/2 cells were similar to endogenous levels observed in control cells ( Fig. 2A). We observed lower protein levels of p105DN and p105DNS930A than p105, and it is likely that this is secondary to more rapid turnover, as we observed similar mRNA levels in BMDM expressing p105, p105DN, or p105DNS930A using a C-terminal Nfkb1 Taqman RT-PCR probe (Fig. 2B, right). An N-terminal Nfkb1 probe revealed similar RNA levels in cells expressing p105 and p50 (Fig. 2B, left).
As expected, LPS induced significantly higher expression of Ifnb mRNA in Nfkb1 2/2 BMDM infected with vector alone than in control BMDM infected with the vector alone ( Fig 3A). The presence of p105, p50, and p105DN in Nfkb1 2/2 BMDM all significantly inhibited expression of Ifnb mRNA at the point of peak induction 2 hours after LPS treatment, and these differences were mimicked by differences in IFN-b accumulation in the cell supernatants (Fig. 3B). These results indicate that both p50 and the C-terminal fragment of p105 independently inhibit expression of Ifnb.
It has previously been demonstrated that phosphorylation of S930 in p105 is essential for LPS-induced ERK activation [20,21]. It has been proposed that ERK activation is associated with suppression of LPS-induced Ifnb and Il12b expression because it is necessary for induction of c-Fos, an inhibitor of both Ifnb and Il12b [29]. However, it has not been demonstrated whether phosphorylation at S930 is necessary for inhibition of Ifnb expression by p105DN. In contrast to p105DN, p105DNS930A did not inhibit LPS-induced Ifnb mRNA expression ( Fig. 3A) or accumulation of IFN-b in cell supernatants (Fig. 3B), and in fact Ifnb mRNA was expressed at significantly higher levels in Nfkb1 2/2 BMDM expressing p105DNS930A than in Nfkb1 2/2 BMDM infected with vector alone (Fig. 3A). Consistent with a role for phosphorylation of serine 930 in LPS-induced ERK activation and stabilization of c-Fos, expression of p105DN but not p105DNS930A rescued the defect in ERK activation and c-Fos induction observed in Nfkb1 2/2 BMDM (Fig. 4). As expected, LPS-induced degradation of p105DNS930A was impaired (Fig. 4). These results demonstrate that phosphorylation of S930 plays an essential role in mediating the ability of the C-terminal region of p105 to inhibit LPS-induced Ifnb expression, and that this is strongly associated with activation of ERK and induction of c-Fos. Further, these results suggest that in the absence of IKK-mediated phosphorylation at S930, p105DN demonstrates an activity that enhances LPS-induced expression of Ifnb mRNA.
Consistent with previous results [4,6,7], LPS also induced higher levels of Il12b in Nfkb1 2/2 cells than in control cells following infection with vector alone (Fig. 5). To determine  whether Nfkb1 inhibited Il12b expression using similar mechanisms employed for inhibition of Ifnb, we evaluated expression of Il12b in RNA samples from the virus-infected macrophages described above. As expected, p105 significantly inhibited expression of Il12b at both 2 (data not shown) and 4 hours after treatment with LPS (Fig. 5A). However, contrary to results with Ifnb, p50 did not significantly influence Il12b expression, suggesting that p50 is not an important inhibitor of Il12b following LPS treatment (Fig. 5A).
In contrast, p105DN significantly inhibited expression of Il12b. However, we were surprised to find that inhibition did not depend on ERK activation or the induction of c-Fos, as p105DNS930A also significantly inhibited Il12b expression (Fig. 5A), demonstrating that inhibition of Il12b expression by p105DN was proceeding in an ERK and c-Fos independent fashion. Indeed, inhibition by p105DNS930A consistently appeared somewhat more robust than inhibition by p105DN, although this did not reach statistical  significance. Differences observed in Il12b mRNA expression were closely correlated with differences observed in the accumulation of IL-12 p40 in cell supernatants (Fig. 5B).
The experiments described above employing p105DNS930A demonstrate that in the absence of the ability to activate ERK, the C-terminal region of p105 retains the ability to inhibit Il12b but enhances expression of Ifnb mRNA. In addition to facilitating LPSinduced ERK activation, the C-terminal region of p105 contains ankyrin repeats similar to those found in the classical IkBs [14,15]. It has been shown that this region mediates association between p105 and both p65 and c-Rel [12,15,31,32]. c-Rel is necessary for LPSinduced Il12b expression in macrophages [33,34] and therefore inhibition of c-Rel function by the C-terminal region of p105 could explain the ability of p105DNS930A to inhibit Il12b. Conversely, inhibition of c-Rel function could explain the ability of p105DNS930A to augment Ifnb induction if c-Rel is an inhibitor of Ifnb. In fact, we observed that Rel 2/2 BMDM expressed significantly higher levels of Ifnb mRNA 2 hours after LPS stimulation than control macrophages, while, as expected, they expressed markedly lower levels of Il12b mRNA (Figs. 6A and 6B). These results were closely mimicked by the accumulation of IFN-b and IL-12 p40 in cell supernatants (Figs. 6C and 6D). These results suggest that inhibition of c-Rel function by the C-terminal region of p105 could have differential effects on the expression of Il12b and Ifnb.

Discussion
We have shown that Nfkb1 inhibits LPS-induced gene expression by multiple gene-specific mechanisms. p50 inhibits Ifnb but not Il12b. In contrast, the C-terminal region of p105 plays a role in inhibition of both Ifnb and Il12b, but only inhibition of Ifnb requires the ability of this C-terminal region to activate ERK.
It was previously proposed that homodimers of p50 inhibit access of more active Rel dimers to kB sites present with the promoter regions of NF-kB target genes [22]. However, a recent study demonstrated that promoters of genes expressed at higher levels in Nfkb1-deficient cells are enriched for ISRE-like sequences that contain short G-rich stretches [3]. It was proposed that these G-rich stretches serve as cryptic binding sites for p50 homodimers, which prevent recruitment of basal IRF-3. The promoter for Ifnb contains such an ISRE [3], but it has not been previously determined whether p50 alone is capable of inhibiting LPSinduced expression of endogenous Ifnb in Nfkb1 2/2 macrophages. Here we show that expression of p50 in Nfkb1 2/2 BMDM can inhibit expression of Ifnb, which is consistent with a model in which binding of p50 to the Ifnb ISRE inhibits Ifnb expression.
In contrast to its ability to inhibit Ifnb expression, p50 was unable to inhibit expression of Il12b. The fact that p50 was expressed above endogenous levels but was still unable to inhibit Il12b indicates that this failure was not an artifact of low expression levels. The Il12b promoter has previously been demonstrated to contain an atypical NF-kB binding site that consists of a stretch of Gs, termed an NF-kB half site [35], as well as an ISRE [25,26]. However the presence of these sites is clearly not sufficient to support inhibition of Il12b by p50, indicating that p50 has highly selective inhibitory function. Whether this selectivity is solely based on the presence of G-rich ISRE sequences and inhibition of IRF-3 recruitment remains to be determined. In contrast to p50, expression of p105DN inhibited LPSinduced expression of both Ifnb and Il12b. Previous studies have shown that ERK inhibitors increase LPS-induced expression of both of these genes [27][28][29], leading us, and others, to hypothesize that p105 inhibits by facilitating activation of ERK [29,36]. It has been proposed that ERK-dependent inhibition is mediated by increased mRNA expression [29] and/or protein stabilization of c-Fos [27], an inhibitor of both Ifnb and Il12b [27][28][29]. The results reported here strongly suggest that the ability of p105DN to rescue ERK activation and induce c-Fos is necessary to inhibit Ifnb, but in contrast, is dispensable for inhibition of Il12b.
The C-terminal fragment of p105 contains multiple ankyrin repeats similar to those found in other IkBs [11,12,14,15], and it has been suggested that p105 may preferentially interact with c-Rel [12,15,31,32]. c-Rel is essential for LPS-induced expression of Il12b [33] and thus inhibition of c-Rel function by p105DN could explain inhibition of Il12b. This is supported by the observation that p105DNS930A, which is resistant to IKK-mediated degradation, may be a better inhibitor of Il12b expression than p105DN itself. Interestingly, we have found that in contrast to its effects on Il12b expression, c-Rel inhibits LPS-induced expression of Ifnb. Thus, more robust inhibition of c-Rel function by p105DNS930A than by p105DN could explain why p105DNS930A actually augments expression of Ifnb. While one potential mechanism for inhibition of c-Rel function by p105DNS930A is inhibition of nuclear localization following LPS stimulation, we have not observed increased c-Rel nuclear translocation in Nfkb1-deficient BMDM nor have we observed inhibition of c-Rel nuclear translocation by p105DNS930A (data not shown). Therefore the functional consequences of c-Rel interaction with p105 remain to be fully elucidated.
The studies presented here provide new insight into the mechanisms of gene inhibition by Nfkb1. They demonstrate that Nfkb1 has multiple inhibitory mechanisms that function in genespecific manners. Further they illustrate that rather than an on/off switch for inflammation, NF-kB is a complex modulator of the innate response that differentially influences key immunoregulatory cytokines. Understanding the molecular details of gene- specific regulation by NF-kB could lead to highly selective therapeutic modulation of immune and inflammatory based diseases.

Ethics Statement
This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The animal protocol was approved by the Harvard Medical Area Standing Committee on Animals (OLAW Assurance number: A34131-01). All efforts were made to minimize suffering.

Retrovirally-mediated gene expression
BMDM were grown as previously described [36]. On the second day of culture cells were infected with the indicated retroviruses. 72 hours later macrophages were re-plated at 5610 5 cells/well in 0.5 ml of medium in a 24-well plate, or at 0.3-1610 5 cells/well in 100 ml of medium in a 96 well plate. The following day cells were stimulated with LPS from E. coli 0127:B8 (Sigma, St Louis, MO) at 1 ng/ml or left unstimulated.

RT-PCR
RNA was isolated in Trizol (Invitrogen, Carlsbad, CA) following the manufacturer's instructions. RT-PCR analysis was performed using probes from Applied Biosystems (Foster City, CA), as per the manufacturer's instructions. Expression for each sample was analyzed in duplicate and normalized to GAPDH using the DDCt method. Fold-change is reported as relative to the average of uninduced levels in unstimulated BMDM for each experiment.
ELISA IFN-b was analyzed in culture supernatants using an IFN-bspecific ELISA kit (PBL Biomedical Laboratories) according to the manufacturer's instructions. IL-12 p40 was analyzed via ELISA using C15.6 (Biolegend, San Diego, CA) as the capture antibody and C17.8 (Thermo Scientific, Rockford, IL) as the secondary antibody.

Statistical Analysis
All data analysis was performed using GraphPad Prism software (GraphPad Software, inc., San Diego, CA). Data generated by RT-PCR or ELISA was compared using One-way ANOVA with the Tukey's multiple comparison test. Error bars on graphs represent SEM. * indicates p,0.05, ** indicates p,0.01, and *** indicates p,0.001.