|Control of RelB during dendritic cell activation integrates canonical and noncanonical NF-κB pathways.|
|Jump to Full Text|
|PMID: 23086447 Owner: NLM Status: MEDLINE|
|The NF-κB protein RelB controls dendritic cell (DC) maturation and may be targeted therapeutically to manipulate T cell responses in disease. Here we report that RelB promoted DC activation not as the expected RelB-p52 effector of the noncanonical NF-κB pathway, but as a RelB-p50 dimer regulated by canonical IκBs, IκBα and IκBɛ. IκB control of RelB minimized spontaneous maturation but enabled rapid pathogen-responsive maturation. Computational modeling of the NF-κB signaling module identified control points of this unexpected cell type-specific regulation. Fibroblasts that we engineered accordingly showed DC-like RelB control. Canonical pathway control of RelB regulated pathogen-responsive gene expression programs. This work illustrates the potential utility of systems analyses in guiding the development of combination therapeutics for modulating DC-dependent T cell responses.|
|Vincent F-S Shih; Jeremy Davis-Turak; Monica Macal; Jenny Q Huang; Julia Ponomarenko; Jeffrey D Kearns; Tony Yu; Riku Fagerlund; Masataka Asagiri; Elina I Zuniga; Alexander Hoffmann|
Related Documents :
|24217607 - Roles of eosinophils in the modulation of angiogenesis.
23600877 - Immunomodulators in tick saliva and their benefits.
24551817 - Low concentrations of flavonoid - rich fraction of shallot extract induce delayed - typ...
23781227 - Pathogenicity of and plant immunity to soft rot pectobacteria.
20837477 - Entamoeba histolytica cysteine proteinase 5 binds integrin on colonic cells and stimula...
15833687 - Eosinophilic gastroenteritis.
|Type: Journal Article; Research Support, N.I.H., Extramural Date: 2012-10-21|
|Title: Nature immunology Volume: 13 ISSN: 1529-2916 ISO Abbreviation: Nat. Immunol. Publication Date: 2012 Dec|
|Created Date: 2012-11-19 Completed Date: 2013-01-17 Revised Date: 2013-07-11|
Medline Journal Info:
|Nlm Unique ID: 100941354 Medline TA: Nat Immunol Country: United States|
|Languages: eng Pagination: 1162-70 Citation Subset: IM|
|Signaling Systems Laboratory, Department of Chemistry and Biochemistry and San Diego Center for Systems Biology, University of California, San Diego, La Jolla, California, USA.|
|Data Bank Information|
Bank Name/Acc. No.:
|APA/MLA Format Download EndNote Download BibTex|
Cell Differentiation / genetics
Dendritic Cells / immunology*, metabolism*
Fibroblasts / metabolism
Gene Expression Regulation
I-kappa B Kinase / metabolism
Mice, Inbred C57BL
NF-kappa B / genetics, metabolism*
Toll-Like Receptor 2 / metabolism
Toll-Like Receptor 4 / metabolism
Toll-Like Receptor 9 / metabolism
Transcription Factor RelB / genetics, metabolism*
|AI081923/AI/NIAID NIH HHS; AI090935/AI/NIAID NIH HHS; GM071573/GM/NIGMS NIH HHS; GM085325/GM/NIGMS NIH HHS; GM085763/GM/NIGMS NIH HHS; P01 AI090935/AI/NIAID NIH HHS; P01 GM071862/GM/NIGMS NIH HHS; P50 GM085764/GM/NIGMS NIH HHS; R01 AI083453/AI/NIAID NIH HHS; R01 GM071573/GM/NIGMS NIH HHS|
|0/NF-kappa B; 0/Relb protein, mouse; 0/Tlr2 protein, mouse; 0/Tlr4 protein, mouse; 0/Tlr9 protein, mouse; 0/Toll-Like Receptor 2; 0/Toll-Like Receptor 4; 0/Toll-Like Receptor 9; 147337-75-5/Transcription Factor RelB; EC 22.214.171.124/I-kappa B Kinase|
|Nat Immunol. 2012 Dec;13(12):1139-41
Journal ID (nlm-journal-id): 100941354
Journal ID (pubmed-jr-id): 21750
Journal ID (nlm-ta): Nat Immunol
Journal ID (iso-abbrev): Nat. Immunol.
nihms-submitted publication date: Day: 30 Month: 8 Year: 2012
Electronic publication date: Day: 21 Month: 10 Year: 2012
Print publication date: Month: 12 Year: 2012
pmc-release publication date: Day: 01 Month: 6 Year: 2013
Volume: 13 Issue: 12
First Page: 1162 Last Page: 1170
PubMed Id: 23086447
|Control of RelB during dendritic cell activation integrates canonical and non-canonical NF-κB pathways|
|Vincent F.-S. Shih1|
|Jenny Q. Huang1|
|Jeffrey D. Kearns14|
|Elina I. Zuniga2|
1Signaling Systems Laboratory, Department of Chemistry and Biochemistry and San Diego Center for Systems Biology (SDCSB)
2Division of Biological Sciences
3San Diego Supercomputer Center University of California, San Diego 9500 Gilman Dr La Jolla, CA 92093-0375
5Center for Innovation in Immunoregulative Technology and Therapeutics, Graduate School of Medicine, Kyoto University, Yoshida-Konoe, Sakyoku, Kyoto 606-8501, Japan
|Correspondence: Correspondence should be addressed to A.H. (firstname.lastname@example.org). Tel. 858 822 4670 Fax 858 822 4671
4current address: Merrimack Pharmaceuticals, Cambridge, MA
Dendritic cells (DCs) are specialized sentinel immune cells essential in both innate and adaptive immunity. DC progenitors differentiate to become immature DCs that populate both non-lymphoid and lymphoid tissues and perform immune-surveillance functions. When encountering pathogens or pathogen-associated molecular patterns (PAMPs), immature DCs undergo a maturation program that determines their role in the adaptive immune response1. A hallmark of DC maturation is expression of major histocompatibility complex molecules (MHC), T cell costimulatory molecules (CD40, CD80 or CD86) and cytokines (for example, interleukin 23 (IL-23)) in addition to a gene expression program of intracellular factors that enable effective antigen uptake, processing and presentation, and T cell activation. In addition, inflammatory molecules such as nitric oxide and cytokines such as tumor necrosis factor (TNF) and interferon (IFN) underlies DC functions in innate immune responses2, 3. DCs have thus attracted attention for engineering or modulating immune-based therapies4.
The transcription factor NF-κB protein RelB is highly expressed in antigen-presenting cells5 and critical for DC maturation, their functions as antigen-presenting cells6 and DC-mediated immunity. Specifically, siRNA-mediated silencing of RelB expression radically altered the DC maturation process and resulted in blunted antigen-specific T cell responses in vitro and in vivo7. RelB-deficient mice revealed deficiencies in splenic DC subsets8, 9, but other critical roles of RelB in DCs may be masked by other cell types, notably T cells that are misregulated in these null animals. DC-specific deletion of the RelB-controlling kinase NIK resulted in deficient T cell responses10. Indeed, the extent of RelB activation determined the tolerance or rejection of allogenic organ transplants by determining the balance of associated activated or regulatory T cells7. These insights have prompted investigations of cell-based therapies for autoimmune diseases using RelB-silenced DCs11.
Despite the potential clinical importance of RelB, the molecular mechanisms that control its activity in DCs have remained unclear. Mouse embryonic fibroblasts (MEFs) have served as a useful model system for many signaling studies. Detailed biochemical studies in MEFs showed that unlike classical NF-κB (the RelA-p50 dimer), RelB is not activated from a latent cytoplasmic pool via the NEMO-dependent, so-called “canonical” signaling pathway, but via the so-called “non-canonical” NF-κB pathway that involves proteolysis and processing of newly synthesized NF-κB2 (p100)12-14. Consistent with the critical role of RelB in DCs, non-canonical signaling pathway components such as NIK and Nf-b2 were reported to be required for proper DC functions10, 15. However, RelB was also found to be rapidly activated in DCs by canonical pathway stimuli TNF and lipopolysaccharide (LPS)16-19 and the canonical signaling pathway component TRAF6 was shown to be essential9. These reports suggest that RelB control in DCs may be different than what has been described in MEFs. In DCs, the molecular control mechanisms must provide for constitutive RelB expression to enable rapid and decisive induction of maturation programs following exposure to pathogens or PAMPs, but must limit spontaneous maturation of DCs in their absence.
In this study, we elucidated the molecular mechanisms responsible for regulating RelB in DCs. We used a Systems Biology approach of iterative computational modeling and quantitative experimental analyses of the NF-κB signaling network in DCs to reveal that RelB activity was limited by classical IκBs, IκBα and IκBε, and regulated via the canonical pathway. Modeling studies identified two DC-specific control points that render RelB subject to regulation by the canonical pathway, and we demonstrated their sufficiency by engineering MEFs accordingly to produce DC-like RelB control. Finally, gene expression profiling revealed that RelB-dependent gene expression programs regulated by the canonical pathway activity control DC-orchestrated immune responses.
The established view of NF-κB signaling comprises two separate pathways (Fig. 1a)12. The canonical pathway, involving the NEMO-dependent kinase IKK, triggers degradation of NF-κB inhibitors, the classical IκBs, IκBα, -β, -ε. Resulting activation of latent RelA- and c-Rel-containing NF-κB dimers controls inflammatory and proliferative gene expression programs. The non-canonical pathway, involving the kinases NIK and IKK1, triggers processing of p100 to p52 and generation of the RelB-p52 transcription factor, which is implicated in cell survival and maturation. To examine NF-κB RelB signaling in DCs in a quantitative manner, we developed a mathematical model that describes the formation and regulation of RelA and RelB dimers in terms of mass action kinetics (Supplementary Notes). The first version of the model involves 41 molecular species, 132 reactions and 53 unique kinetic parameters based on published and newly made measurements that constrain the model to a single parameter set ensemble; it recapitulates well-documented NF-κB control in MEFs20-22, such as prompt LPS-induced RelA activation and delayed lymphotoxin β-mediated RelB activation (Fig. 1b).
To adapt the model to DCs, we first measured the expression of key NF-κB proteins in bone marrow-derived DCs (BMDCs) in comparison to mouse embryonic fibroblasts (MEFs) and bone marrow-derived macrophages (BMDMs). Relative to the housekeeping gene β-actin (Actb), expression of Rela mRNA was found to be similar in BMDCs, BMDMs and MEFs, and the relative amount of RelA protein in these cell types correlated (Fig. 1c, top,). In contrast, 3- to 6-fold more Relb mRNA and protein expression were observed in BMDCs than MEFs and BMDMs (Fig. 1c, middle, and Supplementary Fig. 1a). p100, encoded by the Nf-b2 gene, is known to inhibit RelB. We therefore tested if p100 expression correlated with enhanced RelB expression in BMDCs. We did observe 3.5-fold more Nf-b2 mRNA in BMDCs, but quantitative immunoblotting showed little difference in the p100 protein abundance among the cell types analyzed (Fig. 1c, bottom, and Supplementary Fig. 1b). Lack of correlation between the relative p100 protein and RNA abundance suggested that p100 degradation may be elevated in BMDCs. We noted a 2.5-fold increase of p52 protein in BMDCs, which suggests that both complete p100 degradation and p100 processing to p52 may be occurring in BMDCs (Fig. 1c, bottom, and Supplementary Fig. 1b). Consistent with this hypothesis, protein expression of IKK1, the kinase determining the activity of non-canonical NF-κB pathway, gradually increased during DC differentiation with concomitant p100 processing to p52 (Fig. 1d), potentially via the control of miRNAs23. Our data indicate that DC differentiation involves not only increased expression of RelB, but also elevated constitutive activity of the non-canonical NF-κB signaling pathway.
Based on the measurements, we made specific modifications to the computational model to recapitulate RelB control in DCs (Supplementary Notes). First, we increased Relb and Nf-b2 expression 3-fold, which increased the abundance of RelB but not its nuclear localization. Then we destabilized p100 by the IKK1-dependent pathway to achieve comparable p100 expression as in MEFs (Fig. 1c, bottom right). This change resulted in a substantial increase of nuclear RelB activity (Fig. 1e). To test experimentally if RelB in DCs primarily localizes into the nucleus, we separated BMDCs into cytoplasmic and nuclear extracts but found that more than 75% of the total RelB protein was cytoplasmic (Fig. 1f and Supplementary Fig. 1c,d). Indeed, whereas RelB was more abundant in the cytoplasm of BMDCs than MEFs or BMDMs, p100 was not (Fig. 1g). The fact that the mathematical model, which encodes the known mechanisms of RelB control, failed to reproduce our experimental observations suggested that there may be as yet undescribed regulatory mechanisms that sequester RelB in the cytoplasm.
To search for inhibitors of RelB in DCs, we immunoprecipitated RelB from BMDC whole cell lysates and analyzed the associated proteins (Fig. 2a). As expected, p100, the known RelB inhibitor and non-canonical regulator, was found to be associated with RelB. Unexpectedly, IκBα and IκBε, the classical IκB inhibitors regulating the canonical NF-κB pathway, were also immunoprecipitated with RelB, but IκBβ and p105 were not. Interestingly, substantial amounts of p50, known as the binding partner of RelA in the canonical pathway, were found in RelB immunoprecipitates, and this complex was primarily cytoplasmic (Supplementary Fig. 2a,b). Reciprocal immunoprecipitation of various NF-κB inhibitors further confirmed that RelB not only directly interacts with p100 but also associates with IκBα and IκBε in BMDCs (Fig. 2b and Supplementary Fig. 2c), and RelA is associated with IκBα as expected (Supplementary Fig. 2d). Furthermore, the observations that IκB immunoprecipitates did not contain other IκB isoforms confirmed the specificity of the antibodies used and that only one IκB isoform associates with each RelB molecule. Analyses of the amounts of RelB captured (IP) and remaining in the flowthrough (FT) following immunoprecipitation with various IκB antibodies provides a quantitative understanding of RelB protein distribution in BMDCs (Fig. 2c). This analysis revealed that 37% to 45% of RelB was associated with p100 and 12% to 17% with IκBε. Further, we found a substantial proportion of RelB (19% to 34%) associated with IκBα, which prompted us to investigate the function of this interaction further.
To test whether IκBα may inhibit RelB activity in BMDCs, we took advantage of IκBα-deficient mice22 and developed two strategies to focus our experimental analysis on RelB activity. First, we bred the mice onto a c-Rel-deficient background (Rel−/−), then we modified the standard electrophoretic mobility shift assay with κB-site containing probes (κB-EMSA) to include shift-ablating antibodies for RelA, resulting in a specific RelB-EMSA. Using these tools, we found that RelB activity was more than two-fold elevated in IκBα-deficient BMDCs (Fig. 2d). Supershift analysis with antibodies, that were shown to be specific for p50 and p52 (Supplementary Fig. 2e), revealed that while control BMDCs contained primarily constitutive RelB:p52 activity, ablation of IκBα resulted in a substantial increase in active RelB:p50 dimer, rendering RelB:p50 the predominant NF-κB activity in IκBα-deficient BMDCs (Fig. 2e). We examined the functional consequences of RelB misregulation by monitoring the frequency of matured DCs as indicated by surface expression of the activation markers CD86 and MHC II. Strikingly, IκBα-deficiency resulted in an increased percentage (42% vs. 28%) of MHCIIhiCD86hi BMDCs in the absence of external stimuli (Fig. 2f). Although RelB-deficiency does not affect the frequency of MHCIIhiCD86hi BMDCs prior to exposure to maturation stimuli (Supplementary Fig. 2f), the inappropriate spontaneous DC maturation phenotype of IκBα-deficient BMDCs was dependent on RelB, as compound deletion of the Relb gene fully reversed the phenotype (Fig. 2f). We then examined the antigen-presenting functions of DCs by testing their ability to activate proliferation and cytokine production of antigen-specific T cells in DC–T cell co-cultures (Fig. 2g-i). We found that IκBα-deficiency increased the antigen-presenting functions in BMDC co-cultures with OVA-responsive T cells exposed to ovalbumin peptide, and this effect was largely but not entirely dependent on RelB (Fig. 2h), correlating with the partial RelB-dependence of surface MHC expression (Fig. 2f). However, when these co-cultures were exposed to ovalbumin protein, which must be taken up and processed before being presented, T cell activation showed a near absolute dependence on RelB (Fig. 2i), correlating with previous studies of RelB-deficient DCs6, and suggesting a specific function for RelB in regulating the antigen uptake and processing program of antigen-presenting cells. Together, these data demonstrate that the classical NF-κB inhibitor, IκBα, not only restrains the expression of RelB by controlling RelA or c-Rel24, in immature DCs it also has a critical functional role in restraining RelB activity to prevent inappropriate spontaneous maturation.
To explore the regulatory consequences of RelB-p50 interactions with IκBα and IκBε proteins during dendritic cell maturation, we incorporated them into the mathematical model as kinetic rate equations, and used the quantitative immunoprecipitation results as constraints in a multi-dimensional parameter optimization protocol (Supplementary Notes). We simulated NF-κB regulation during Toll-like receptor (TLR)-induced dendritic cell maturation using experimentally measured time-course data of the NEMO-dependent IKK kinase activity as an input. Such simulations indicated rapid and substantial activation not only of RelA but also RelB (Fig. 3a). To test this prediction experimentally, BMDCs and MEFs were stimulated with the TLR9 ligand CpG, the TLR2 ligand Pam3CSK4 and the TLR4 ligand LPS as well as an agonistic antibody to LTβR to induce the non-canonical NF-κB pathway. To specifically examine the activation profiles of RelA- and RelB-containing NF-κB dimers, we employed the newly developed RelA-EMSA22 and RelB-EMSA using shift-ablating antibodies for activation domain-containing Rel proteins (Supplementary Fig. 3b). RelA activation was similar in BMDCs and MEFs stimulated with TLR ligands (Fig. 3b, left). Interestingly, we found rapid RelB activation in response to TLR stimuli in BMDCs but not in MEFs, although MEFs do activate RelB at later time points when stimulated with anti-LTβR (Fig. 3b, right, and Supplementary Fig. 3c). Similarly, rapid activation of RelB was observed in splenic DCs stimulated with CpG or Pam3CSK4 (Supplementary Fig. 3d). Further, computational simulations suggested that this induced RelB activity consists of RelB-p50 rather than RelB-p52 dimer (Fig. 3c, top). Experimentally, supershift analyses of nuclear extracts revealed that both RelB-p50 and RelB-p52 activities were present in unstimulated conditions, but that CpG stimulation primarily increased RelB-p50 activity (Fig. 3c, bottom, and Supplementary Fig. 3e), unlike LTβR stimulation of MEFs which induces RelB-p52. These data suggest that during DC maturation RelB activation is regulated by the canonical pathway.
A hallmark of canonical signaling is the release of a pre-existing NF-κB dimer, whereas non-canonical signaling involves the stimulus-responsive de novo generation of the dimer12, 25. In CpG-responding DCs we did not detect increases in protein expression of RelB or p50, or Relb mRNA, whereas Nfkbia mRNA, encoding IκBα, was induced more than four-fold (Supplementary Fig. 3f). Furthermore, inhibition of protein synthesis by cycloheximide did not block CpG-induced RelB activation whereas resynthesis of IκBα protein was blocked (Supplementary Fig. 3g), suggesting that de novo RelB protein synthesis is not required for CpG-inducible RelB activation. In contrast, immunoblotting confirmed that in DCs nuclear accumulation of RelB was accompanied by disappearance of cytoplasmic RelB after CpG stimulation, indicative of stimulus-responsive nuclear translocation of a pre-existing pool of RelB (Supplementary Fig. 3h). Further, inhibition of IKK2 activity, a hallmark of the canonical pathway, by the inhibitor PS-1145 (ref. 26) resulted not only in reduced RelA activity and IκBα protein degradation but also in reduced RelB activation (Supplementary Fig. 3g,i), suggesting that IKK2 signaling is required for RelB activation. We monitored the abundance of known NF-κB inhibitor proteins during the CpG time course: the abundance of the potential RelB inhibitors p100 and p105 remained unaltered; however, IκBα and IκBε were rapidly degraded, correlating with the activation kinetics of RelB activation (Fig. 3d). Importantly, in coimmunoprecipitation assays the amount of IκBα associated with RelB decreased in response to CpG (Fig. 3e). Together, these data suggest that degradation of IκBα allows for the release of RelB from pre-existing IκBα–RelB complexes.
To investigate the role of IκBα in TLR-induced RelB activation, we utilized the mathematical model to computationally simulate the effect of IκB deletions on RelB activation. We found that in silico deletion of individual inhibitors had little effect, except in the case of IκBα (Supplementary Fig. 3j). Even compound deficiency of IκBβ, IκBε and IκBδ (which elevated basal RelB activity; Supplementary Fig. 3k), showed robust RelB activation in response to canonical pathway activation, as opposed to greatly diminished activation in an IκBα-deficient model (Fig. 3f). To test these computational modeling predictions, we utilized IκBα-deficient mice22 and generated Nfkbib−/−Nfkbie−/−Nfkb2−/− mice. The lack of protein products was confirmed by immunoblotting (Supplementary Fig. 3l). Indeed, RelB activation was found to be robust in Nfkbib−/−Nfkbie−/−Nfkb2−/− BMDCs, whereas IκBα-deficient BMDCs showed a diminished increase and delayed kinetics (Fig. 3g and Supplementary Fig. 3m). Together, these data provide genetic and mechanistic evidence that IκBα is required for CpG-induced RelB activation in DCs.
We previously showed that hallmarks of the NF-κB signaling system in mature but unstimulated DCs are abundant basal RelB expression and basal non-canonical pathway activity. To investigate whether these mechanisms are sufficient and what their relative contributions may be, we performed computational simulations of RelB activation for a range of parameter values governing basal RelB mRNA synthesis and NIK protein half-life. These in silico analyses showed that activation of RelB mildly increased when either parameter was increased, however, substantial enhancement occurred only when both parameters were concomitantly elevated (Fig. 4a). Our simulations suggest that the DC-specific, rapid RelB activation upon canonical pathway stimulation can be explained by DC-specific, constitutively elevated RelB mRNA synthesis and non-canonical IKK activity.
To test this model-derived hypothesis experimentally, we asked whether genetically engineering these two mechanisms into MEFs may be sufficient to allow for DC-like canonical regulation of RelB. We took advantage of MEFs deficient in TRAF3, an E3-ligase controlling NIK degradation27, to increase constitutive non-canonical signaling. As further suggested by the model simulations, we then transduced a retroviral Relb transgene to increase RelB expression about 3-fold relative to untransduced MEFs (Fig. 4b and Supplementary Fig. 4a). Remarkably, the engineered MEFs did indeed show substantial RelB activation in response to LPS (Fig. 4c) or TNF (Supplementary Fig. 4b), whereas the parental control MEFs did not, and RelA activation by these stimuli remained unchanged. Furthermore, neither single genetic alteration produced substantial RelB activation, indicating that enhanced RelB expression and non-canonical pathway activity function synergistically, as predicted by the model, to push RelB into the canonical pathway and render it responsive to TLR agonists. Further, antibody supershift and depletion analysis (Fig. 4d and Supplementary Fig. 4c) confirmed that canonical signaling primarily activated the RelB:p50 dimer (7-fold) rather than the RelB-p52 dimer (2-fold) as observed in DCs and predicted by the computational model (Fig. 3c). Overexpression of a RelB-GFP fusion protein retrovirally transduced into single cells also revealed nuclear translocation upon TNF stimulation in the Traf3−/− context but not in control cells (Fig. 4e).
These iterative computational-experimental studies support a model in which the NF-κB protein RelB may function in either non-canonical or canonical pathways (Fig. 4f). When dimerized to p100 or p52, RelB is subject to control by the non-canonical pathway; when dimerized to p50, RelB may be bound by IκBα and IκBε and is regulated by NEMO-dependent canonical signals. Our analysis indicates that low constitutive RelB expression and non-canonical pathway activity characterizes one steady state (found in MEFs) and allows for RelB-p52 activation by stimuli such as LTβ that engage the non-canonical pathway (Supplementary Fig. 4d). High constitutive RelB expression and non-canonical pathway activity characterizes another steady state (found in DCs) and allow for RelB:p50 activation by stimuli such as CpG that engage the canonical pathway (Supplementary Fig. 4e).
Given that RelB-p50 is induced by PAMPs during DC maturation, we wondered whether it controls the expression of inflammatory regulators or DC activation markers. Following stimulation with the TLR9 ligand CpG or the TLR2 ligand Pam3CSK4 for 24 h, we indeed observed reduced surface expression of DC activation markers MHCII, CD86, CD80 and CD40 in Relb−/− DCs (Fig. 5a and Supplementary Fig. 5a,b). Further, we found that expression of pro-inflammatory genes, Tnf and Il23a, correlated with the kinetics of CpG or Pam3CSK4-induced RelB activation and were reduced in Relb−/− DCs (Fig. 5b). In EMSAs, activated RelB-p50 was found to be able to bind to DNA probes containing the κB sites found in the Tnf and Il23a promoters (Fig. 5c), indicating that RelB-p50 can directly interact with these regulatory regions. In vivo, recruitment of RelB to the promoter regions of Tnf and Il23a genes following DC maturation with CpG was also observed using the chromatin immunoprecipitation assay (Fig. 5d).
We noted that RelB bound to consensus κB site sequences28 associated with the known canonical NF-κB pathway effectors, RelA and c-Rel, rather than the unconventional sequences previously ascribed to RelB in splenic stromal cells29 or MEFs20. Because single knockouts did not show overt defects in CD11c+ cell generation in BM cultures (Supplementary Fig. 5d), we tested whether c-Rel and RelB have overlapping functions in regulating the DC maturation program by examining gene expression in c-Rel and RelB doubly-deficient DCs. Genome-wide expression profiling activated by TLR ligands CpG and Pam3CSK4 revealed a group of 157 genes that were statistically significantly down-regulated in Rel−/−Relb−/− BMDCs (Fig. 5e and Supplementary Table 1). To delineate the contribution of RelB in activating these genes, we examined the expression phenotype of the 50 most severely c-Rel–RelB-dependent genes in Relb−/− BMDCs stimulated with TLR ligands. Expression phenotypes in fold induction were calculated between wild-type and null DCs, and the order of genes was sorted in increasing degree of RelB-dependency (Fig. 5f and Supplementary Table 2). This analysis revealed a continuous spectrum of RelB-dependency rather than two distinct classes (of RelB-dependent and RelB–independent genes), suggesting an overlap in DNA interaction specificities between c-Rel and RelB dimers. Interestingly, Tnf and Il23a were identified in this analysis as regulated by both RelB and c-Rel. Quantitative RT-PCR further validated the requirements of RelB and c-Rel in activating Cxcl2, Cd40 and Il1b gene expression (Supplementary Fig. 5c).
Given overlapping functions of c-Rel and RelB in regulating DC gene expression programs, we investigated their relationship within the signaling system. Whereas RelA and c-Rel proteins were found to be expressed similarly in wild-type BMDCs and those lacking RelB, Rel−/− BMDCs showed decreased RelB protein expression (Fig. 6a). Further, Relb transcripts were reduced ~40% in Rel−/− BMDC (Fig. 6b). This reduction resulted in severely diminished activation of RelB DNA binding activity in Rel−/− BMDCs in response to LPS (Fig. 6c). These data indicate that one of the key determinants of RelB control by the canonical pathway, namely RelB expression, is in fact controlled by c-Rel (Fig. 6d). The feed-forward circuit architecture suggests that expression of RelB in differentiated but immature DCs may reflect the exposure of differentiating cells to c-Rel-inducing stimuli. We therefore tested whether c-Rel-deficient DCs may also be defective in RelB-responsive gene expression by comparing the expression of RelB-target genes in Rel−/− and Rel−/−Relb−/− DCs. Indeed, Rel−/− BMDCs showed reductions of surface marker and inflammatory cytokine expression (Supplementary Fig. 6a,b). These data support a model in which RelB acts as a downstream mediator of c-Rel in DC activation programs.
The present study revealed how canonical and non-canonical NF-κB signaling pathways integrate to control the RelB transcription factor during DC development and maturation. NF-κB signaling is generally classified as either “canonical” (i.e. NEMO-dependent and involving classical IκBα, -β, and -ε) when activated by acute inflammatory agents or “non-canonical” (i.e. NEMO-independent and involving Nf-b2/p100) when activated by developmental signals12. NFκB family members were thought to follow this distinction: RelA and c-Rel as effectors of the canonical pathway, and RelB as the effector of the non-canonical pathway, based on its role as a RelB:p52 transcription factor in secondary lymphoid organogenesis. However, we show here that RelB is also an effector of the canonical pathway in DCs.
Specifically, we show that during DC differentiation RelB expression is increased, and that elevated steady-state non-canonical pathway activity results not only in the expected RelB-p52 dimer, but in formation of the RelB-p50 dimer. Unlike RelB-p52, which is mostly nuclear in immature DCs, RelB-p50 is inhibited by the IκB proteins, IκBα and IκBε, which allows for rapid activation of RelB-p50 activity via the canonical pathway upon exposure to maturation stimuli. Conversely, with the recent discovery of IκBδ, 21, 22 chronic inflammatory conditions were found to render RelA an effector of the non-canonical signaling pathway. Thus, both RelA and RelB are potential effectors of the canonical and non-canonical signaling pathways; whether they are functionally relevant effectors is determined by the physiological steady state of the NF-κB signaling system.
Our observations imply that RelB-p50 and RelB-p52 present different molecular surfaces to IκB proteins, providing physiological relevance to previous studies of protein interaction specificities30, 31. Similarly, the DNA interaction characteristics of RelB-p50 and RelB-p52 may be distinct32, 33. RelB-Arg125 within the RelB-p52 dimer makes an additional base contact with DNA that allows RelB-p52 to recognize a broader range of κB sites. This may account for the RelB-p52 specific function in regulating chemokines involved in secondary lymphoid organeogenesis, such as secondary lymphoid tissue chemokine (SLC), EBI1 ligand chemokine (ELC), B lymphoblastoid cell chemokine (BLC), and stromal cell-derived factor 1α (SDF-1α)20, 29. In contrast, RelB-p50 interacts with DNA sequences similarly to RelA-p50 and a role for RelB in TNF production, GM-CSF and Bcl-xl expression has been reported34, 35. Together, these studies suggest that the dimerization partner of RelB determines not only the signaling pathway that RelB is responsive to, but also the RelB target gene program.
Why then, would DCs employ RelB as an effector of the canonical NF-κB signaling pathway along with RelA and c-Rel? One possibility is that RelB-p50 target genes are in fact distinct from those controlled by c-Rel or RelA. Our transcriptomic profiling suggests overlap between c-Rel- and RelB-dependent gene programs, but c-Rel turned out to control RelB expression, thus other tools are required to address the question of RelB-p50 vs. cRel-p50 specificity. A second possibility is that the stimulus-responsive dynamic control of RelB is distinct from RelA or c-Rel. Although RelB-p50 is inhibited by IκBα in resting cells, it may make for a poorer substrate for IκB feedback control than RelA, which is efficiently stripped off the DNA by IκBα 36. We speculate that the involvement of RelB-p50 in DC biology, ensures irreversible execution of a terminal maturation/activation program in response to transient PAMP exposure.
Mathematical modeling, which is used here to describe biochemical reactions in terms of kinetic rate equations, lends itself as a tool for studying the regulation of signaling networks. Iteratively refined mathematical models of the NF-κB–IκB system have addressed the dynamic and homeostatic control of the NF-κB RelA-p50 dimer by IκB proteins in fibroblasts21, 22, 37-41. Within this study, we have developed the first kinetic model that accounts for the generation and regulation of multiple NF-κB dimers, namely RelA- and RelB- containing dimers. Further, we contrasted the steady-state and dynamic control mechanisms in two cell types, MEFs and DCs, and found that the key biochemical differences are two kinetic rate constants (Relb mRNA synthesis and NIK halflife); a 3-fold increase was sufficient to shift the in silico model from MEF to DC-like regulation of the NF-κB signaling system. This prediction was confirmed experimentally by genetically engineering MEFs to produce DC-like RelB control. Importantly, there was no need to invoke cell-type-specific protein interaction specificities or any other cell-type-specific molecular mechanism. The results indicate that cell-type-specific quantitative control of the steady state of a signaling system may determine seemingly qualitative cell-type-specific properties, such as DC-specific RelB activation by TLRs. As such, kinetic modeling and a quantitative analysis of signaling systems may serve to generate hypotheses not only for mechanistic studies but also for the development of DC-mediated therapeutics.
GM-CSF and IL-4 were from Peprotech. 0.1 μM CpG (Invivogen), 500 ng/ml Pam3CSK4 (Invivogen), 100 ng/ml LPS (Sigma, B5:055) and 0.5 μg/ml LTβR agonist (Biogen, Inc.) were used to stimulate cells. Cycloheximide and IKK2 inhibitor (PS-1145) were from Sigma. Antibodies to RelA (sc-372), RelB (sc-226), c-Rel (sc-70), IκBα (sc-371), IκBβ (sc-945), IκBε (sc-7155), IKK1 (sc-7606), TRAF3 (sc-6933), USF-2 (sc-861), α-tubulin (sc-5286), β-actin (sc-1615) and CD16/CD32 (sc-18867) were from Santa Cruz Biotechnology. p105/p50, p100/p52 and antibody to p100 C-terminus were from National Cancer Institute, Biological Resources Branch, Frederick, MD. NIK antibody (4994) was from Cell Signaling. Immunoprecipitation beads and HRP-conjugated anti-rabbit secondary antibody were from eBioscience.
Wild-type and gene-deficient C57BL/6 mice were maintained in accordance with the Animal Care Program at UCSD. Nfkbia−/−Tnf−/−Rel−/− and Nfkbia−/−Tnf−/−Rel−/−Relb−/− mice were generated by cross-breeding Nfkbia−/−Tnf−/− with Rel−/− and Relb−/− mice. Nfkbib−/−Nfkbie−/−Nfkb2−/− mice were generated by cross-breeding Nfkbib−/−, Nfkbie−/−, and Nfkb2−/− mice. Rel−/−Relb−/− mice were generated by cross-breeding Rel−/− and Relb−/− mice. Primary MEFs were generated from E12.5-14.5 embryos. Bone marrow-derived macrophages (BMDMs) and bone marrow-derived dendritic cells (BMDCs) were made from bone marrow suspensions prepared from mouse femurs. 2 × 106 BM cells seeded on 10 cm plate were cultured for one week with L929-conditioned DMEM medium to derive BMDMs or cultured for 6–11 days with 20 ng/ml GM-CSF and 10 ng/ml IL-4 to derive BMDCs. BMDC medium was replaced on day 3, 6, 8 and floating cells were collected and subjected to experimental analyses as previously described42. Typically, day 6-7 BMDCs were used to investigate TLR-induced DC maturation and day 9-11 BMDCs for spontaneous DC maturation studies.
Spleens were cut into small fragments and digested with collagenase D (2 mg/ml, Roche) for 30 min at 37°C followed by incubation with 10 mM EDTA pH 8.0 for 5 min. Single-cell suspension of splenocytes were enriched for CD11c+ cells by immunomagnetic cell sorting using MACS CD11c microbeads (Miltenyi Biotec) according to manufacturer’s protocol.
Single-cell suspension were collected and blocked with anti-mouse CD16/CD32 in PBS containing 5% FCS for 10 min. Cells were stained with 7-AAD to exclude dead cells and indicated antibodies for DC maturation analyses. All antibodies were purchased from BD Pharmingen: anti-CD11 (HL3), anti-CD40 (3/23), anti-CD80 (16-10A1), anti-CD86 (GL-1), and anti-IAb (AF6-120.1). Stained cells were acquired in either a FACSCalibur (BD Biosciences) or an Accuri C6 and data analysis was performed with FlowJo software.
GM-CSF derived bone marrow DCs were pulsed with 200 μg whole ovalbumin (Sigma, 5 μM OVA 323-339 (OT-II) peptide (Anaspec, Inc.), or media alone for 2 h at 37 °C. Naive CD4+T cells (5 × 104 cells/well) were obtained by negative enrichment (>90% purity; Stem Cell Technologies) from spleens of B6.Cg-Tg (TcraTcrb)425 Cbn/J mice transgenic for ovalbumin 323-339 specific αβTCR (Jackson Laboratory) and labeled with CFSE (Sigma). DCs were washed and cultured with CFSE-labeled CD4+T cells (5 × 104 T cells/well) at the indicated DC:T cell ratios as described43. 72 h later, T cells were re-stimulated with 5 μM OT-II peptide for 5 h in the presence of brefeldin A and examined for CFSE dilution and production of TNF, and IL-2 by flow cytometry (BD LSRII). Data were analyzed using FlowJo software (Treestar, Inc.).
Whole cell extracts were prepared in RIPA buffer and normalized for total protein or cell numbers before immunoblot analysis. Cytoplasmic and nuclear extracts from BMDMs and BMDCs were prepared by high salt extraction buffer (Buffer A: 10 mM HEPES pH 7.9, 10 mM KCl, 0.1 mM EGDA, 0.1 mM EDTA; Buffer C: 20 mM HEPES pH 7.9, 420 mM NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 25% glycerol). Immunoprecipitation-immunoblotting analysis, electrophoretic mobility shift assay (EMSA), chromatin immunoprecipitation (ChIP) were performed as previously described21, 44. In EMSA focusing on RelB-DNA binding activity, nuclear extracts were ablated of RelA and c-Rel-containing DNA binding activities by pre-incubating them with RelA and c-Rel antibodies (Fig. 3b and Supplementary Fig. 3b). Similarly, nuclear extracts were pre-incubated with RelB and c-Rel antibodies when RelA DNA binding activity was the focus (Fig. 3b). Antibody-shift ablation analysis (for RelB, p50 and p52) was performed as previously described22 and the specificities of antibodies were confirmed (Supplementary Fig. 2e). RNA extraction was performed with RNAeasy Mini Kit (Qiagen) and purified RNA was used for first strand cDNA synthesis with oligo dT and SuperScript RT II (Invitrogen). Quantitative RT-PCR was performed with SYBR Green PCR Master Mix reagent (Stratagene) and Eppendorf Mastercycler realplex system using the Δ(ΔCt) method with β-actin as normalization control and relative to signals in MEFs or respective basal levels to derive fold induction. qRT-PCR and ChIP data shown are representative of three independent experiments (mean ± SD). Quantitation of mRNA and protein abundance are representative of four independent experiments.
RelB- or RelB-GFP expressing pBabe-puro constructs were generated by standard methods, and co-transfected with pCL.Eco into 293T cells with Lipofectamine™ 2000 transfection reagent (Invitrogen) for 48 h. Supernatant was filtered and used to infect MEFs. Transduced cells were selected with puromycin hydrochloride (Sigma). Images were acquired with a Zeiss Axio Z1 microscope.
RNA was collected from one set of timecourse experiments (1, 8, 24 h) using WT, Relb−/−, and Rel−/−Relb−/− BMDC stimulated with 0.1 μM CpG (Invivogen) or 500 ng/ml Pam3CSK4. Labeling and hybridization to the Illumina v.2 gene expression chip was performed by UCSD Biogem core facility. The raw data was preprocessed and normalized by mloess method45. Genes differentially regulated between WT and Rel−/−Relb−/− BMDCs during TLR stimulation time courses were analyzed by two-class paired SAM (Significant Analysis of Microarray)46 implemented in the MeV program (Multiple expression Viewer)47. Class pairing was defined by corresponding time points between WT and Rel−/−Relb−/− BMDCs. Differentially expressed genes identified at the false discovery rate below 5% were deemed significant. Genes with at least two-fold induction during TLR-elicited DCs maturation are listed in Supplementary Table 1. In heatmaps, expression values of each gene were normalized to its maximum fold induction and clustered by hierarchical clustering with Euclidian distance (Fig. 5e). For phenotyping analyses (Fig. 5f and Supplementary Table 2), the average-fold induction26 in log2 scale across either timecourse (CpG and Pam3CSK4) was calculated for different genotypes, e.g. FIWT, FIRel−/−, and FIRel−/−Relb−/−. The RelB phenotype was defined as FIWT-FIRel−/−, the c-Rel-RelB phenotype was defined as FIWT- FIRel−/−Relb−/−.
The RelA–RelB mathematical model (Version 5.0) involving mass action kinetic equations was developed based on a previously published model (Version 3.1)22 and experimental data20 that allowed for constraints-based parameterization. Refinement of the model (Version 5.1) and MEF- and DC-specific parameterization were based on experimental data presented in this paper. Computational simulations were performed in Matlab using the ode15s solver. Detailed descriptions are included in Supplemental Notes and biochemical reactions are listed in Supplementary Tables 3, 4.
Statistical significance was calculated by two-tailed Student’s t test with Prism software (GraphPad). Error bars were shown as either s.d. or s.e.m. as indicated.
FN2Accession codes The microarray data presented is accessible at GEO, accession number GSE 34990.
FN3Author Contributions V.F-S.S. and A.H. designed the study. V.F-S.S. and M.M. carried out all experimental work with assistance from J.Q.H., T.Y., R.F. and M.A. and guidance from E.I.Z. and A.H. J.D.-T. and J.D.K. carried out the computational modeling work and J.P. the bioinformatic analysis. V.F-S.S. and A.H. wrote the manuscript with contributions from all authors.
We thank Z. Tao and G. Ghosh for plasmids and recombinant proteins, S. Basak, A. Wu, P. Loriaux, R. Tsui for computational modeling advice, and C. Brown and M. Karin for Traf3−/− embryos. This study was supported by GM085763 (AH), GM071573 (AH), AI090935 (AH), GM085325 (JP) and AI081923 (EIZ).
|1.||Banchereau J,Steinman RM. Dendritic cells and the control of immunityNatureYear: 19983922452529521319|
|2.||Liu YJ. IPC: professional type 1 interferon-producing cells and plasmacytoid dendritic cell precursorsAnnu Rev ImmunolYear: 20052327530615771572|
|3.||Serbina NV,Salazar-Mather TP,Biron CA,Kuziel WA,Pamer EG. TNF/iNOS-producing dendritic cells mediate innate immune defense against bacterial infectionImmunityYear: 200319597012871639|
|4.||Steinman RM,Banchereau J. Taking dendritic cells into medicineNatureYear: 200744941942617898760|
|5.||Carrasco D,Ryseck RP,Bravo R. Expression of relB transcripts during lymphoid organ development: specific expression in dendritic antigen-presenting cellsDevelopmentYear: 1993118122112318269849|
|6.||Zanetti M,Castiglioni P,Schoenberger S,Gerloni M. The role of relB in regulating the adaptive immune responseAnnals of the New York Academy of SciencesYear: 200398724925712727647|
|7.||Li M,et al. Immune modulation and tolerance induction by RelB-silenced dendritic cells through RNA interferenceJ ImmunolYear: 20071785480548717442929|
|8.||Wu L,et al. RelB is essential for the development of myeloid-related CD8alpha-dendritic cells but not of lymphoid-related CD8alpha+ dendritic cellsImmunityYear: 199898398479881974|
|9.||Kobayashi T,et al. TRAF6 is a critical factor for dendritic cell maturation and developmentImmunityYear: 20031935336314499111|
|10.||Hofmann J,Mair F,Greter M,Schmidt-Supprian M,Becher B. NIK signaling in dendritic cells but not in T cells is required for the development of effector T cells and cell-mediated immune responsesThe Journal of experimental medicineYear: 20112081917192921807870|
|11.||Yang H,et al. Suppression of ongoing experimental autoimmune myasthenia gravis by transfer of RelB-silenced bone marrow dentritic cells is associated with a change from a T helper Th17/Th1 to a Th2 and FoxP3+ regulatory T-cell profileInflamm ResYear: 20105919720519768385|
|12.||Pomerantz JL,Baltimore D. Two pathways to NF-kappaBMol CellYear: 20021069369512419209|
|13.||Oeckinghaus A,Hayden MS,Ghosh S. Crosstalk in NF-kappaB signaling pathwaysNature immunologyYear: 20111269570821772278|
|14.||Derudder E,et al. RelB/p50 dimers are differentially regulated by tumor necrosis factor-alpha and lymphotoxin-beta receptor activation: critical roles for p100J Biol ChemYear: 2003278232782328412709443|
|15.||Lind EF,et al. Dendritic cells require the NF-kappaB2 pathway for cross-presentation of soluble antigensJ ImmunolYear: 200818135436318566401|
|16.||O’Sullivan BJ,Thomas R. CD40 ligation conditions dendritic cell antigen-presenting function through sustained activation of NF-kappaBJ ImmunolYear: 20021685491549812023343|
|17.||Saccani S,Pantano S,Natoli G. Modulation of NF-kappaB activity by exchange of dimersMol CellYear: 2003111563157412820969|
|18.||Ammon C,Mondal K,Andreesen R,Krause SW. Differential expression of the transcription factor NF-kappaB during human mononuclear phagocyte differentiation to macrophages and dendritic cellsBiochem Biophys Res CommunYear: 20002689910510652220|
|19.||Gasparini C,Foxwell BM,Feldmann M. RelB/p50 regulates CCL19 production, but fails to promote human DC maturationEur J ImmunolYear: 2009392215222319655301|
|20.||Basak S,Shih VF,Hoffmann A. Generation and activation of multiple dimeric transcription factors within the NF-kappaB signaling systemMol Cell BiolYear: 2008283139315018299388|
|21.||Basak S,et al. A fourth IkappaB protein within the NF-kappaB signaling moduleCellYear: 200712836938117254973|
|22.||Shih VF,et al. Kinetic control of negative feedback regulators of NF-kappaB/RelA determines their pathogen- and cytokine-receptor signaling specificityProc Natl Acad Sci U S AYear: 20091069619962419487661|
|23.||Li T,et al. MicroRNAs modulate the noncanonical transcription factor NF-kappaB pathway by regulating expression of the kinase IKKalpha during macrophage differentiationNature immunologyYear: 20101179980520711193|
|24.||Wuerzberger-Davis SM,et al. Nuclear export of the NF-kappaB inhibitor IkappaBalpha is required for proper B cell and secondary lymphoid tissue formationImmunityYear: 20113418820021333553|
|25.||Shih VF,Tsui R,Caldwell A,Hoffmann A. A single NFkappaB system for both canonical and non-canonical signalingCell ResYear: 2011218610221102550|
|26.||Waterfield M,Jin W,Reiley W,Zhang M,Sun SC. IkappaB kinase is an essential component of the Tpl2 signaling pathwayMol Cell BiolYear: 2004246040604815199157|
|27.||He JQ,et al. Rescue of TRAF3-null mice by p100 NF-kappa B deficiencyThe Journal of experimental medicineYear: 20062032413241817015635|
|28.||Hoffmann A,Natoli G,Ghosh G. Transcriptional regulation via the NF-kappaB signaling moduleOncogeneYear: 2006256706671617072323|
|29.||Bonizzi G,et al. Activation of IKKalpha target genes depends on recognition of specific kappaB binding sites by RelB:p52 dimersEmbo JYear: 2004234202421015470505|
|30.||Lernbecher T,Kistler B,Wirth T. Two distinct mechanisms contribute to the constitutive activation of RelB in lymphoid cellsEmbo JYear: 199413406040698076601|
|31.||Dobrzanski P,Ryseck RP,Bravo R. Differential interactions of Rel-NF-kappa B complexes with I kappa B alpha determine pools of constitutive and inducible NF-kappa B activityEmbo JYear: 199413460846167925301|
|32.||Fusco AJ,et al. NF-kappaB p52:RelB heterodimer recognizes two classes of kappaB sites with two distinct modesEMBO RepYear: 20091015215919098713|
|33.||Moorthy AK,Huang DB,Wang VY,Vu D,Ghosh G. X-ray structure of a NF-kappaB p50/RelB/DNA complex reveals assembly of multiple dimers on tandem kappaB sitesJ Mol BiolYear: 200737372373417869269|
|34.||Weih F,Warr G,Yang H,Bravo R. Multifocal defects in immune responses in RelB-deficient miceJ ImmunolYear: 1997158521152189164938|
|35.||Sasaki CY,Ghosh P,Longo DL. Recruitment of RelB to the Csf2 promoter enhances RelA-mediated transcription of granulocyte-macrophage colony-stimulating factorJ Biol ChemYear: 20112861093110221071440|
|36.||Bergqvist S,et al. Kinetic enhancement of NF-kappaB×DNA dissociation by IkappaBalphaProc Natl Acad Sci U S AYear: 2009106193281933319887633|
|37.||Werner SL,Barken D,Hoffmann A. Stimulus specificity of gene expression programs determined by temporal control of IKK activityScienceYear: 20053091857186116166517|
|38.||O’Dea EL,Kearns JD,Hoffmann A. UV as an amplifier rather than inducer of NF-kappaB activityMol CellYear: 20083063264118538661|
|39.||O’Dea EL,et al. A homeostatic model of IkappaB metabolism to control constitutive NF-kappaB activityMol Syst BiolYear: 2007311117486138|
|40.||Kearns JD,Basak S,Werner SL,Huang CS,Hoffmann A. IkappaBepsilon provides negative feedback to control NF-kappaB oscillations, signaling dynamics, and inflammatory gene expressionJ Cell BiolYear: 200617365966416735576|
|41.||Hoffmann A,Levchenko A,Scott ML,Baltimore D. The IkappaB-NF-kappaB signaling module: temporal control and selective gene activationScienceYear: 20022981241124512424381|
|42.||Lutz MB,et al. An advanced culture method for generating large quantities of highly pure dendritic cells from mouse bone marrowJ Immunol MethodsYear: 1999223779210037236|
|43.||Boonstra A,et al. Flexibility of mouse classical and plasmacytoid-derived dendritic cells in directing T helper type 1 and 2 cell development: dependency on antigen dose and differential toll-like receptor ligationThe Journal of experimental medicineYear: 200319710110912515817|
|44.||Cheng CS,et al. The specificity of innate immune responses is enforced by repression of interferon response elements by NF-kappaB p50Sci SignalYear: 20114ra1121343618|
|45.||Sasik R,Woelk CH,Corbeil J. Microarray truths and consequencesJ Mol EndocrinolYear: 2004331915291738|
|46.||Tusher VG,Tibshirani R,Chu G. Significance analysis of microarrays applied to the ionizing radiation responseProc Natl Acad Sci U S AYear: 2001985116512111309499|
|47.||Saeed AI,et al. TM4 microarray software suiteMethods EnzymolYear: 200641113419316939790|
|48.||Giorgetti L,et al. Noncooperative interactions between transcription factors and clustered DNA binding sites enable graded transcriptional responses to environmental inputsMol CellYear: 20103741842820159560|
Keywords: NF-κB, dendritic cells, computational modeling, systems biology.
Previous Document: In vitro protoscolicidal effects of high-intensity focused ultrasound enhanced by a superabsorbent p...
Next Document: The CD46-Jagged1 interaction is critical for human TH1 immunity.