Glucocorticoid receptor coordinates transcription factor-dominated regulatory network in macrophages
- Yurii Chinenov†1, 4Email author,
- Maddalena Coppo†1,
- Rebecca Gupte2,
- Maria A Sacta3 and
- Inez Rogatsky1, 2, 4Email author
© Chinenov et al.; licensee BioMed Central Ltd. 2014
Received: 24 April 2014
Accepted: 25 July 2014
Published: 6 August 2014
Inflammation triggered by infection or injury is tightly controlled by glucocorticoid hormones which signal via a dedicated transcription factor, the Glucocorticoid Receptor (GR), to regulate hundreds of genes. However, the hierarchy of transcriptional responses to GR activation and the molecular basis of their oftentimes non-linear dynamics are not understood.
We investigated early glucocorticoid-driven transcriptional events in macrophages, a cell type highly responsive to both pro- and anti-inflammatory stimuli. Using whole transcriptome analyses in resting and acutely lipopolysaccharide (LPS)-stimulated macrophages, we show that early GR target genes form dense networks with the majority of control nodes represented by transcription factors. The expression dynamics of several glucocorticoid-responsive genes are consistent with feed forward loops (FFL) and coincide with rapid GR recruitment. Notably, GR binding sites in genes encoding members of the KLF transcription factor family colocalize with KLF binding sites. Moreover, our gene expression, transcription factor binding and computational data are consistent with the existence of the GR-KLF9-KLF2 incoherent FFL. Analysis of LPS-downregulated genes revealed striking enrichment in multimerized Zn-fingers- and KRAB domain-containing proteins known to bind nucleic acids and repress transcription by propagating heterochromatin. This raises an intriguing possibility that an increase in chromatin accessibility in inflammatory macrophages results from broad downregulation of negative chromatin remodelers.
Pro- and anti-inflammatory stimuli alter the expression of a vast array of transcription factors and chromatin remodelers. By regulating multiple transcription factors, which propagate the initial hormonal signal, GR acts as a coordinating hub in anti-inflammatory responses. As several KLFs promote the anti-inflammatory program in macrophages, we propose that GR and KLFs functionally cooperate to curb inflammation.
KeywordsTranscriptional regulation Glucocorticoid receptor Inflammation Feed forward loops Gene regulatory network KLF transcription factors
Preserving homeostasis is the primary function of the innate immune system that detects “danger and stranger” signals and eliminates invading microorganisms, responds to irritation and injury and eventually initiates tissue repair. Innate immune cells, such as macrophages, neutrophils and dendritic cells constantly sample their environment for lipopolysaccharides (LPS), double and single-stranded nucleic acids, microbial proteins and other broad molecular patterns that are not normally present in eukaryotes and, in response, produce cytokines and chemokines that attract additional immune cells to the site of infection or injury. Normally a protective response, excessive or persistent inflammation is associated with tissue damage and needs to be regulated. Indeed, numerous mechanisms have evolved that control inflammation at multiple levels. Systemically, inflammatory stimuli activate neuro-endocrine circuitry that triggers the production of glucocorticoids by the adrenal glands, which ultimately attenuate the expression of inflammatory cytokines . These potent anti-inflammatory properties made glucocorticoids exceedingly common for managing a wide range of autoimmune and inflammatory conditions such as rheumatoid arthritis, systemic lupus erythematosus, inflammatory bowel disease, psoriasis and multiple sclerosis .
Glucocorticoids signal through the glucocorticoid receptor (GR) – a ubiquitously expressed ligand-dependent transcription factor (TF) of the nuclear receptor (NR) superfamily. GR regulates transcription by either binding directly to specific DNA sequences known as glucocorticoid response elements (GREs) or by tethering to other DNA-bound regulators, such as Activator Protein (AP)1, Nuclear Factor (NF) κB and Signal Transducers and Activators of Transcription (STAT) family members . Although GR is expressed in all immune cells, the physiological outcomes of GR activation are highly cell type-specific: for example, glucocorticoids are anti-apoptotic in neutrophils, but pro-apoptotic in eosinophils, dendritic and some T-cells . Moreover, prolonged glucocorticoid treatment induces cells polarization toward a new phenotype with either pro- or anti-inflammatory properties [5, 6].
The analyses of glucocorticoid-regulated transcriptomes paint a complicated picture encompassing hundreds of up- and down-regulated genes that vary in different cell types and populations and over time of glucocorticoid exposure [7, 8]. Although shared GR target genes certainly exist, system-specific regulators and pathways drastically affect transcriptional outcomes, response dynamics and relative activities of such shared genes and their products. The existence of intricate inter-protein and inter-pathway interactions contributing to the NR-mediated gene regulation has been proposed almost 30 years ago . The structural analysis of NR transcriptional networks, however, was lagging due to the lack of genome-wide data and limited availability of analytical tools. More recently, studies in bacteria and yeast have defined specific patterns of functional interactions (“network motifs”) between interdependent TFs and provided a computational framework for the analysis of gene expression data to identify such motifs [10–12]. The simplest motif - a positive or negative auto-regulatory loop - consists of a single TF that regulates its own expression. More complicated feed forward loops (FFL) involve three factors: a signal-responsive master regulator TF, an intermediate TF controlled by the master TF and a jointly regulated gene under the control of both the master and intermediate TFs . Depending on the specific activities of the master and intermediate TFs (activation or repression) and the response thresholds of participating genes, the dynamics of the FFL transcriptional outputs vary, yielding unique expression patterns for various TF and target gene combinations. Dynamic responses elicited by FFLs deviate from simple gene regulation providing for unusual control mechanisms that are responsible for noise filtering, fold-change sensing, pulse generation and transcription response acceleration [11, 13, 14].
Gene expression networks can be represented as graphs with TFs and other expression regulators acting as nodes and functional interactions between regulators and between regulators and their targets as directional edges [12, 15]. Of particular interest are the networks and network motifs in which TFs regulate each other, or themselves act as transcriptome organizers and ultimately determine the topology of the entire network . Several computational algorithms and experimental approaches have been successfully applied to map global transcriptional networks and identify novel functional motifs in various organisms [16, 17]. In metazoans, this analysis is often complicated by the lack of information about the edge identity (not all targets for a given TF are known, some known “targets” are not regulated directly) and direction (a TF can either activate or repress the same gene in a tissue-specific manner). To complicate matters further, the role of intermediate TFs can be fulfilled by miRNAs or regulators of RNA translation and stability [18, 19]. Thus, dissecting regulatory networks requires a combination of computational and experimental approaches.
We reasoned that a highly branched response to glucocorticoids is determined by the early transcriptional events. Here, we focused on the regulatory network elicited by an acute stimulation of mouse macrophages with glucocorticoids and/or LPS. Combined with high-resolution kinetic experiments and dynamics modeling, this analysis enabled us to dissect early post-stimulation events prior to extensive signal propagation, which usually masks the bona fide response to GR activation by a web of secondary effects.
Transcriptome analysis of mouse macrophages exposed to acute glucocorticoid and LPS stimulation
To analyze early regulatory events initiated by glucocorticoids and inflammatory stimuli we treated BMMФ with either ethanol vehicle (U), LPS (L), Dex (D), or a combination of the two (L + D) for 1 h, isolated and sequenced PolyA-enriched RNA as described in (Additional file 1). The sequencing results are summarized in Additional file 2: Table S1.
I) Genes activated by either LPS (Cluster 2) or glucocorticoids (Cluster 7)
The majority of genes in these two clusters are upregulated by either LPS or Dex independently, with little evidence for inter-treatment interactions at 1 h. The LPS-induced Cluster 2 (Figure 1, Mann–Whitney, PU-L = 4.11*10-13) contains genes encoding pro- and anti-inflammatory cytokines and chemokines (Il10, Cxcl1, 3, 5 and 7, Ccl7 and Tnfsf9), TFs involved in stress response (Maff, Ets2, Fosl2 and Kdm6b) and proteins involved in TLR signaling (Tlr2, Cd14 and Cd40) and signal transduction (Itpkc, Rabgef1, Gbp5a).
Cluster 7 contains Dex-induced genes (PU-D = 0.0013), including several well-characterized GR targets such as TFs Per1 and Klf9, immunophilin Fkbp5, potassium channel Kcnk6. In addition, this cluster includes several genes whose regulation by Dex has not been previously reported: Interleukin 15 receptor alpha (Il15ra), the Wnt pathway receptor Fzd4, the TF Klf2 and chemokine Ccl17.
II) Genes co-activated by LPS and Dex
These genes display either predominantly additive (cluster 1) or synergistic (cluster 6) activation by LPS and Dex. Several genes in these clusters are previously characterized GR targets including Dusp1, Nfil3 and Cited2 (cluster 1) and Mt2 and Pfkfb3 (cluster 6), whereas the glucocorticoid and LPS responsiveness of several others, such as histone chaperon Jdp2 (cluster 1), has not been reported previously.
III) Genes induced by LPS and repressed by Dex
This large group of genes is represented by clusters 3, 4 and 5. Cluster 3 contains LPS-induced genes (PU-L = 1.95*10-11) expressed at relatively high level in resting BMMФ. The basal expression of these genes is significantly more sensitive to hormonal treatment (PU-D = 0.0107) than their LPS-induced expression. This cluster encompasses a number of inflammatory cytokines (Ccl2, 3 and 4, Tnf, Tnfaip2), TFs (Ier5, Junb, Bcl6, Prdm1 and Irf1) and proteins involved in signal transduction (Gadd45b, Dusp5, Rasgef1b). Interestingly, several genes in this cluster (Ccl2, 3 and 4, Tnf) are characterized by the presence of the stalled RNA Pol II near the transcription start site in uninduced conditions and are activated primarily at the level of the Pol II pause release during early elongation [23–25]. Cluster 4 combines a heterogeneous group of genes with low basal expression (PUcluster 4<Ucluster3 = 0.0003) that are strongly induced by LPS and includes inflammatory cytokines (Il1a, Il1b, Cxcl10, Ifnb1, Tnfsf4 and Il1f9) and other direct mediators of inflammation (Ptgs2, mIR-155 host gene, Hsp1a), TFs (Egr3) and signaling molecules (Gpr84, Areg). Cluster 5 contains many genes whose LPS induction is strongly inhibited by Dex treatment (PL-(L+D) = 0.02) including several cytokines and chemokines (Il12b, Lif, Il1rn and Il17ra ) and TFs (Mxd1, Etv3, Klf7 and Ets1).
Cluster 8 contains statistically heterogeneous previously reported (Atf3, Egr2 and Ier3) as well as novel (Enc1 and Bhlhe40) targets for GR-mediated repression that are largely unaffected by LPS treatment. Thus, this cluster is formally outside of group III.
IV) Genes downregulated by LPS
LPS-repressed genes were separated into 3 clusters based on the combined effect of Dex and LPS on gene expression. The LPS-mediated downregulation is either weakly potentiated (cluster 12) or antagonized (cluster 11) by Dex. The functions of the majority of these genes are poorly understood or unknown, however, 32% of genes in cluster 11 and 20% in cluster 12 encode uncharacterized C2H2 Zn-finger proteins implicated in transcription and chromosome maintenance.
Expression of genes in cluster 10 is downregulated by Dex and LPS in an additive manner. At least one gene in this cluster (Angptl4) is a known GR target. Several genes encode regulators of immune cells activities including Rit1 and Cd300lb. Cluster 9 contains Dex-induced genes that are weakly repressed by LPS including previously reported Ddit4, Arl4d and Sik1.
Cluster validation by RT-qPCR
Two representative genes in each cluster were chosen for validation. Total RNA was isolated from treated (D, L, or L + D for 1 h) and control BMMФ and transcript levels for indicated genes (Figure 1) were determined by RT-qPCR using Act1 or Hprt as housekeeping control genes. As several LPS-induced, Dex-repressed genes coding for various cytokines that we found in clusters 3 (Tnf, Ccl2, 3 and 4), 4 (Il1a and b) and 5 (Lif, Niacr1, Mmp13) have been previously characterized by us and others [20–22], we focused on uncharacterized LPS/Dex targets. The expression patterns determined by RT-qPCR closely resembled those determined by RNA-seq, with the exception of several weakly expressed genes that were not detectably repressed by Dex following a 1-h treatment (e.g., Bcl2, Med21; data not shown).
Glucocorticoid-regulated genes form a highly interconnected association network with distinct response-specific modules
To determine the prevalent functional patterns in groups of co-regulated genes, we performed gene enrichment analysis and visualization using GeneMANIA plugin  for Cytoscape 2.8 and Exploratory Gene Association Networks (EGAN) software .
Glucocorticoid response-specific modules are functionally distinct
Conversely, the majority of enriched GO categories in Module 1, which contains predominantly Dex-repressed genes, are related to immune and inflammatory responses, signaling and regulation of signal transduction and metabolic regulation including immune response (GO:6955), immune system process (GO:2376), inflammatory response (GO:6954) and regulation of cytokine production (GO:1817) (Figure 3B, 3C). The fraction of gene expression-related GO categories in this module is significantly smaller (22/425, χ2 = 8.05, df = 1, p = 0.00455) than in Module 2.
Dex-responsive transcription regulators
The dynamics of expression for several Dex-regulated TFs suggests that they are under combinatorial controls that involve GR and additional GR targets which either cooperate with or antagonize GR actions. As such a model implies transcription/protein production of these putative GR targets, we first examined the expression of Dex-regulated genes in the presence of a protein synthesis inhibitor cycloheximide (Chx). Treatment with Chx up for to 3 h had no dramatic effects on Dex-mediated regulation of Klf9, Tcs22d3, Tgfb3 and Bcl3 (Figure 5B), suggesting a direct regulation by GR that does not rely on synthesis of additional proteins. Conversely, the expression of Atf3 became refractory to Dex in the presence of Chx (Figure 5B) suggesting that additional proteins induced by Dex rather than GR itself are likely to directly regulate this gene. For several Dex-responsive genes (Klf2, Klf4, Nfil3, Tiparp, Tgif1), however, treatment with Chx dramatically upregulated their basal expression, complicating the assessment of the relative contribution of direct vs. indirect effects of GR to their regulation (Additional file 3: Figure S2) and necessitating an alternative approach.
Temporal dynamics of hormone-regulated gene expression is consistent with feed forward logic
The dynamics of the transcriptional response of several genes to Dex imply the existence of some feedback mechanism that limits activation by GR yet, at the same time, is GR-dependent. Because Chx elicits many off-target effects and does not enable us to discriminate between the secondary targets of GR and those jointly controlled by GR and a GR-regulated TF, we performed dynamic modeling of expression data in an attempt to discern specific regulatory patterns. Several mechanisms including positive and negative autoregulation, positive and negative feedback and feed forward loops (FFL) could account for deviations from a simple model with a single TF regulating gene expression via a single DNA binding site . The kinetics of GR expression following Dex stimulation is not consistent with auto-regulatory models (Figure 5A). Depending on the overall regulatory outputs and activities of individual FFL components, two types of FFL have been recognized – coherent (C-FFL) and incoherent (I-FFL). In type 1 I-FFL (I1-FFL), the activating master TF and a repressing intermediate regulator have opposite effect on a jointly regulated gene. Dynamic modeling and experimental studies of I1-FFL dynamics have demonstrated several properties of this network motif, including its ability to produce sharp pulse-like activation of a jointly regulated gene with a fast relaxation time [11, 13]. Several GR-regulated genes, including Klf2, Tiparp, Tgfb3, Mt2 exhibit pulse-like kinetics at constant Dex exposure (Figure 5A). Near-baseline relaxation of Klf2 expression suggests that this gene is under control of both GR and a strong Dex-induced repressor. Because GR is largely inactive in the absence of ligand, glucocorticoids act as a low-latency on-off switch eliminating the need to correct for a baseline activity of GR. The dynamics of Klf2 and repressor (R) expression is described by the ordinary differential equations (2) and (3) (see Additional file 1) [11, 13].
Assuming equal degradation rates of Klf2 and R, these equations can be solved analytically (equation (1) and ) and used to fit the expression data for Klf2. When limited to the early data points (up to 4 h), the expression data fit very well to the predicted expression pattern (Figure 5C, R 2 = 0.9225), however, at the later time, when the contribution of degradation rates becomes significant, the quality of fit decreases (R 2 = 0.5491). Using parameter estimates derived from equation (1) fitting, we solved equations (2) and (3) numerically. Figure 5C shows a good concordance between the experimental data for Klf2 expression as determined by RT-qPCR and the numerical solution (R 2 = 0.8317) that describes the dynamics of the I1-FFL. This result strongly suggests that Klf2 and, potentially, several other GR targets that exhibit similar expression dynamics (Tiparp, Tgfb3 and Mt2) are jointly regulated by GR and a GR-induced repressor via the I1-FFL network motifs.
Numerical solutions of equations 2 and 3 also provide a theoretical prediction of an intermediate repressor dynamics in the GR-R-Klf2 I-FFL. Interestingly, among several known transcription repressors activated by GR, the expression kinetics of Klf9 fits the best to the predicted model (Additional file 3: Figure S3). The dynamics of the GR-R-KLF2 FFL can be tested by perturbing the concentration of a hypothetical intermediate repressor which should uncouple the FFL thus shifting peak-like FFL-mediated kinetics to simple monotonous kinetics eventually converging to a steady-state level. To test the role of Klf9 as a potential GR-activated repressor of Klf2 transcription, we derived M from Klf9-KO mice  kindly provided by Dr. Simmen and treated them with 100 nM Dex as above. Interestingly, in Klf9 null BMMФ the peak-like Klf2 induction profile “degenerated” to monotonous activation kinetics that plateaued at a steady state level by 3 h (Figure 5D), replicating the profiles of previously reported uncoupled experimental I-FFLs [11, 31], consistent with the proposed role of KLF9 as an intermediate repressor in the GR-KLF9-KLF2 I-FFL. At the same time, deletion of Klf9 did not affect the expression dynamics of either GR itself or the GR target gene Tsc22d3 with a simple monotonous activation profile (Figure 5D).
GR is recruited to binding sites associated with Dex-regulated genes
We then used a set of Dex/LPS-regulated genes to analyze the distribution and enrichment of binding sites for TFs that were present in Chip-Enrichment Analysis (ChEA) database . The ChEA database contains curated published genome-wide datasets of TF binding sites in human, mouse and rat. After filtering out TFs that were not expressed in MФ (RPKM < 1), we noted that binding sites for several Dex-responsive TFs, such as KLF2, KLF4, ATF3, EGR1, CEBPβ and IRF1 are enriched among Dex/LPS-regulated genes. Interestingly, binding sites for PPARγ, whose expression was inhibited upon prolonged Dex treatment, were found near the majority (22/37) of Dex-responsive gene expression regulators (Figure 4) and highly enriched among Dex/LPS-regulated genes in general.
We then determined the frequency of genes associated with binding sites for several TFs identified by ChEA and induced by Dex in individual clusters of Dex/LPS-regulated genes (Figure 1). We defined a binding site as ‘gene-associated’ if its genomic intervals overlapped with genomic intervals encompassing all mouse genes annotated in mm9 +/- 15 Kb by at least one nucleotide. In good correlation with the RNA-seq data, acute GR recruitment peaks previously identified by ChIP-seq in Dex-treated adipocytes  were enriched in Dex-induced clusters 1, 6, 7 and 9 (Figure 7A, arrow up). Among Dex-regulated TFs, mouse genome-wide binding datasets are currently available for KLF4 (ChIP-seq), KLF2 (chip-on-chip), PPARγ (ChIP-seq) and NFIL3 (chip-on-chip) [36–40]. Although KLF4 sites are enriched in the entire Dex/LPS dataset compared to non-expressing genes, only in cluster 8 (Dex-repressed genes) the enrichment level attains significance (Figure 7A). Conversely, KLF4 binding sites are underrepresented in the LPS-induced cluster 2. Although the KLF2 chip-on-chip dataset available was relatively small, several KLF2 targets were regulated by Dex and contained GR binding sites, including Klf2 itself, Klf4, Klf9 and Tgfb3 .
Two mouse PPARγ datasets, one from differentiated adipocytes and one from resting MФ are currently available [38, 39]. Consistent with the previously reported role of PPARγ in repression of inflammatory genes, PPARγ binding sites are overrepresented in several LPS-induced clusters (clusters 2, 3 and 5; Figure 7A) in both datasets. In addition, in the MФ dataset, PPARγ binding sites were enriched in Dex-induced and -repressed clusters, 7 and 8, respectively (Figure 7A).
The only available genome-wide dataset of Nfil3 binding was acquired in a neuronal cell line . Among Nfil3-occupied genes identified in that study, only four genes overlap with Dex/LPS-regulated dataset reported here; however, one of them, Tsc22d3 (GILZ), is a well-characterized GR target.To identify genes that might be under combinatorial control by GR and another Dex-responsive TF, we searched for loci that contained GR, KLF and/or PPARγ binding sites located close to each other within a gene +/- 15 Kb. Intriguingly, several GR targets including Klf2, Klf9, Cited2 and Mt2 contained tight clusters of binding sites for GR, KLF(4) and PPARγ (Figure 7B) suggesting that these TFs may interact functionally or physically at a subset of GR-regulated genes.
LPS downregulates a unique class of genes encoding the C2H2-KRAB gene expression regulators
Glucocorticoids- and LPS-regulated gene expression programs
GR is a ubiquitous ligand-dependent TF capable of eliciting highly divergent transcription programs with up to a third of protein-coding genes differentially expressed following a 24-h glucocorticoid treatment . Establishing the hierarchy of regulatory events upon prolonged hormonal exposure in individual cell types is challenging, which complicates both accurate mechanistic predictions and clinical decisions. Multiple GR ligands have been designed in an attempt to create a highly specific compound that selectively regulates desired subsets of genes. Mechanistic analyses of these ligands usually focus on a specific group of disease-relevant genes and often involve long-term treatments, which obscure primary and transient responses to GR activation by a plethora of secondary pathways. In the context of inflammation, both immediate and delayed regulatory events are clinically relevant as they reflect typical glucocorticoid treatment modalities. We reasoned that by analyzing early transcriptomes elicited by the inflammatory and glucocorticoid exposure in MФ, a clinically relevant cell type, we will be able to isolate a key set of immediate GR targets responsible for the delayed gene expression patterns. Our results indicate that early glucocorticoid- and LPS-dependent changes establish a highly organized program of gene expression with distinct groups of genes following cooperative and antagonistic regulation. As expected from previous work [22, 34, 45] a large group of LPS-induced genes that included among others inflammatory cytokines, was rapidly downregulated by glucocorticoids. Another group encompassing glucocorticoid-induced genes, some of which encode TFs (Cited2, Nfil3, Jdp2) or signaling proteins (Dusp1, Tsc22d3), are involved in curbing inflammatory signaling . We identified several previously unreported glucocorticoid-induced genes whose products are involved in signaling (Ccl17, Il15ra and Fzd4), regulation of transcription (Klf2, Jdp2, Ncoa5 and Tiparp) and mRNA stability (Zfand5). Several of these genes add to the arsenal of anti-inflammatory mediators regulated by GR. For example, KLF2 interferes with AP1- and NFκB-mediated transcription of Tnf and several chemokines including Ccl2, 3 and 4 . Furthermore, Klf2 haploinsufficiency in mice results in an exaggerated inflammatory response and more severe disease in arthritis models . CCL17, another previously unreported glucocorticoid-induced chemokine, is a marker and promoter of the polarization of ‘alternatively activated’ M2 MФ, which are considered anti-inflammatory and mediate tissue repair and wound healing [49, 50]. In addition to repressing cytokine gene transcription, glucocorticoids downregulate expression of several TFs including Atf3, Junb, Irf1, Bcl3, Tgif1, some (e.g., Rela, Nfkb2, Myc, Ets1) in the context of LPS induction, and, unexpectedly, Pparg An enrichment in positive regulators of inflammation and cell proliferation among Dex-downregulated TFs is consistent with the anti-inflammatory and anti-proliferative effects of glucocorticoids. The role of GR in repression of the Pparg gene in MФ has not been previously reported, the effect might be indirect and mediated by a well-established GR target GILZ , which may also account for the delay (Figure 5A). Finally, we described a previously overlooked group of LPS-downregulated genes encoding proteins with the C2H2 Zinc-fingers adjacent to the KRAB domain. Despite being one of the largest TF family, KRAB proteins remain poorly characterized. Among those whose functions were described, several are involved in transcriptional regulation, RNA and DNA binding and splicing [42–44, 52]. KRAB domains interact with a scaffolding co-repressor TRIM28 (KAP1, TIF1β) which in turn binds the heterochromatin protein 1, chromatin remodeler MI2A and H3K9-specific methyltransferase . Indeed, some KRAB proteins reportedly repress transcription by heterochromatin spreading . Interestingly, several KRAB proteins have been linked to NR actions [54, 55]. The role of KRAB proteins in inflammation is essentially unknown; however, genomic studies indicate that inflammatory signaling increases accessibility of large sections of the genome . It is tempting to speculate that a broad downregulation of proteins involved in heterochromatin maintenance and spreading serves to increase DNA accessibility and inflammatory gene transcription.
The dynamic response to GR activation is consistent with feed forward logic
Functional relationships between GR and its targets are often classified as “direct”, that involve GR recruitment to genomic binding sites associated with regulated genes, and “indirect”, whereby primary GR-regulated factors, rather than GR itself, are responsible for activation of the downstream targets. Thus, the activation of these secondary targets is often described as sequential or delayed. Such a model, however, cannot explain many instances of non-monotonous expression dynamics (see Figure 5) and non-linear response to varying hormone concentration of many GRE-driven genes . The large number of shared neighbors, overrepresentation of TFs and their high interconnectivity in GR regulatory networks (Figure 2) are consistent with more intricate regulatory modalities such as FFL. Variations in kinetic parameters for participating TFs, target gene structure and activation/repression thresholds often lead to paradoxical responses to stimulation of the master TF with profound functional implications. C-FFLs serve as delayed response organizers that detect the duration/strength of a signal that activates the initiating TF [11, 12]. Interestingly, the dynamics of Fkbp5 induction by Dex, characterized by a substantial post-exposure delay followed by a robust expression (Figure 5A), is reminiscent of the C-FFL in which the jointly regulated gene is activated by both the master and intermediate TFs . Although additional experiments are required to establish the precise mode of Fkbp5 regulation, this gene is a known direct GR target that recruits GR to several GREs (Additional file 3: Figure S5).
Incoherent loops are responsible for negative and positive pulse generation, accelerated response and fold change sensing [11, 13, 14]. Here, we observed that several GR target genes exhibit both positive (Klf2, Tiparp, Tgfb3 and Mt2) and negative (Tgif1, Junb and Bcl3) pulse-like dynamics consistent with the I-FFL. In keeping with the role of a potential master regulator, GR binds to the GREs in regulatory regions of many of these genes (Figure 6). Furthermore, using a system of ordinary differential equations which describe FFLs in the “fold sensors” model , we showed that Klf2 expression is consistent with that of a gene under joint control of GR and a strong GR-activated repressor (Figure 5C). Several GR-activated genes are either known transcription repressors (e.g., Klf4, 9, Nfil3, Per1 and Jdp2) or may downregulate gene expression by destabilizing RNA transcripts (Zfand5, ). Curiously, the expression dynamics of Klf9 fits closely with the computational prediction of an intermediate repressor in the GR-R-Klf2 I-FFL (Additional file 3: Figure S3). GR is recruited to the Klf9 and Klf2 GREs as early as 40 min of Dex treatment. Both Klf9 and Klf2 regulatory regions also contain functional GAGGCGTGG KLF sites (, Figure 7B) which can be occupied by various TFs of the KLF family  including KLF9. Finally, in KLF9-KO macrophages, the induction profile of Klf2 loses the early peak followed by a decrease and acquires monotonous kinetics (Figure 5D) strongly suggesting a collapse of the I1-FFL to simple GR-dependent activation. Interestingly, KLF binding sites are overrepresented in glucocorticoid-regulated genes and are located near GREs in several bona fide GR target genes suggesting that these factors may co-regulate a number of GR targets.
KLF proteins in inflammation
Both KLF2 and KLF4 have been implicated in myeloid cell biology. KLF2 inhibits monocyte activation by inhibiting NFκB activity, which correlates with decreased expression of multiple cytokines and HIF1α, a TF that regulates myeloid cell response to bacterial infection and reactive oxygen species . Consistent with the anti-inflammatory role of KLF2, mice hemizygous for Klf2 have elevated levels of inflammatory mediators, such as CCL2 and PTGS2 (COX-2). By extension, in KLF2-deficient mice, the manifestations of both Me-BSA- and IL1β-experimentally induced inflammatory arthritis are more severe .
KLF4 is involved in inflammatory monocyte differentiation [61, 62] and in MФ polarization toward the M2 anti-inflammatory phenotype . KLF9 can act as either a transcriptional activator or a repressor [64, 65], however its role, if any, in inflammation has not been described. We showed here that GR regulates Klf genes with distinct temporal dynamics and proposed that KLF9 may act as a GR-induced Klf2 repressor. Thus, it is tempting to speculate that GR anti-inflammatory activities rely in part on the activation of Klf genes whose products regulate transcription of additional targets in concert with GR. Indeed, glucocorticoids and KLF4 regulate partially overlapping set of genes during epidermal barrier establishment in embryogenesis . The proximity of the GREs and KLF binding sites in the genome suggests an intriguing possibility that GR and KLFs interact functionally or physically. Curiously, a functional interaction with the I-FFL logic has been reported for GR and another member of KLF family, KLF15 . Although KLF15 is not expressed in MФ, our studies strongly suggest extensive crosstalk between GR and other KLF family members in the innate immune cells.
Anti-inflammatory activities of glucocorticoids involve downregulation of inflammatory mediators and activation of various anti-inflammatory genes. The early glucocorticoid-driven transcriptome in MФ contains an unusually large number of genes coding for transcriptional regulators. Temporal dynamics of hormone-regulated gene expression is consistent with feed forward logic suggesting that GR and GR-induced TFs jointly regulate GR target genes. In particular, our data suggest that GR is rapidly recruited to and activates genes encoding several members of the KLF family of TFs with profound anti-inflammatory activities, such as Klf2 and Klf4. Furthermore, GR appears to regulate Klf2 expression via the GR-Klf9-Flf2 I1-FFL. We propose that by acting as a hub for highly branched regulatory networks and activating genes encoding TFs to propagate the initial signal, GR coordinates anti-inflammatory responses.
Mice and macrophage cultures
C57BL/6 mice (NCI, Charles River Laboratories) were maintained in the Hospital for Special Surgery Animal Facility in full compliance with institutional guidelines approved by the HSS Animal Care and Use Committee. Klf9-KO mice  were generously provided by Dr. R. Simmen (U. of Arkansas). BMMФ were prepared from 8–12 weeks old mice as in . For RNA-seq, BMMФ from two independent mice were treated with vehicle, Dex (100 nM), LPS (10 ng/ml) or LPS + Dex for 1 h. For qPCR and ChIP analyses, BMMФ were treated as above for time indicated in Figure Legends.
RNA isolation, RT-qPCR and RNA-seq
Total RNA was isolated using the RNeasy kit (Qiagen). 0.25 μg of total RNA was used for random primed cDNA synthesis which was performed with M-MuLV reverse transcriptase (NEB) according to the manufacturer’s recommendations. Quantitative PCR (qPCR) was performed using Maxima Sybr Green/ROX/ 2x master mix (Fermentas) on StepOne Plus real time PCR system (ABI) and analyzed using δδCt method as described previously  with Hprt or Act1 as a normalization control. Primer pairs are listed in Additional file 2: Table S3. RNA-seq is described in (Additional file 1).
ChIP-qPCR and ChIP-seq
BMMФ were incubated -/+100 nM Dex for 40 min and ChIPs were performed as in  using N499  and sc1004 (Santa Cruz Biotechnology) anti-GR antibodies, or normal rabbit IgG as a background control. The data for each recruitment site was normalized to non-specific signals at the unrelated 28S ribosomal gene. Primer pairs are listed in (Additional file 2: Table S3). ChIP-seq is detailed in (Additional file 1).
Gene association network construction, analysis and expression data modeling
GeneMANIA algorithm was used to build gene association networks. Starting from a gene list of interest (e.g., combined list of Dex up- and downregulated genes), geneMANIA algorithm creates a consensus network and predicts gene functions based on integration of multiple prebuilt gene association networks. The detailed description of Gene association network construction, analysis and expression data modeling is in the Supplemental Information section.
Glucocorticoid response element
Bone-marrow derived macrophages
Feed forward loop
Reverse transcription coupled with quantitate polymerase chain reaction
We are deeply grateful to R. Simmen (U. of Arkansas) for Klf9-KO mice. We thank D. Rollins, L. Ivashkiv and M. Shtutman for insightful discussion and helpful comments on the manuscript. This work was supported by the National Institutes of Health grant (NIH R01DK099087) and The Hospital for Special Surgery David Rosensweig Genomics Center (to I.R.), by the American Heart Association Grant 11SDG5160006 (to Y.C.). M.C. is a Research Fellow of The Hospital for Special Surgery David Rosensweig Genomics Center. M.A.S. was supported by the NIH MSTP T32GM007739.
- Silverman MN, Sternberg EM: Glucocorticoid regulation of inflammation and its functional correlates: from HPA axis to glucocorticoid receptor dysfunction. Ann N Y Acad Sci. 2012, 1261: 55-63.PubMed CentralPubMedView ArticleGoogle Scholar
- Rhen T, Cidlowski JA: Antiinflammatory action of glucocorticoids–new mechanisms for old drugs. N Engl J Med. 2005, 353: 1711-1723.PubMedView ArticleGoogle Scholar
- Chinenov Y, Gupte R, Rogatsky I: Nuclear receptors in inflammation control: repression by GR and beyond. Mol Cell Endocrinol. 2013, 380: 55-64.PubMed CentralPubMedView ArticleGoogle Scholar
- Meagher LC, Cousin JM, Seckl JR, Haslett C: Opposing effects of glucocorticoids on the rate of apoptosis in neutrophilic and eosinophilic granulocytes. J Immunol. 1996, 156: 4422-4428.PubMedGoogle Scholar
- Sica A, Mantovani A: Macrophage plasticity and polarization: in vivo veritas. J Clin Invest. 2012, 122: 787-795.PubMed CentralPubMedView ArticleGoogle Scholar
- Varga G, Ehrchen J, Tsianakas A, Tenbrock K, Rattenholl A, Seeliger S, Mack M, Roth J, Sunderkoetter C: Glucocorticoids induce an activated, anti-inflammatory monocyte subset in mice that resembles myeloid-derived suppressor cells. J Leukoc Biol. 2008, 84: 644-650.PubMedView ArticleGoogle Scholar
- Chrousos GP: Stress and sex versus immunity and inflammation. Sci Signal. 2010, 3: pe36-PubMedView ArticleGoogle Scholar
- Maranville JC, Luca F, Richards AL, Wen X, Witonsky DB, Baxter S, Stephens M, Di Rienzo A: Interactions between glucocorticoid treatment and cis-regulatory polymorphisms contribute to cellular response phenotypes. PLoS Genet. 2011, 7: e1002162-PubMed CentralPubMedView ArticleGoogle Scholar
- Yamamoto KR: Steroid receptor regulated transcription of specific genes and gene networks. Annu Rev Genet. 1985, 19: 209-252.PubMedView ArticleGoogle Scholar
- Shen-Orr SS, Milo R, Mangan S, Alon U: Network motifs in the transcriptional regulation network of Escherichia coli. Nat Genet. 2002, 31: 64-68.PubMedView ArticleGoogle Scholar
- Mangan S, Alon U: Structure and function of the feed-forward loop network motif. Proc Natl Acad Sci U S A. 2003, 100: 11980-11985.PubMed CentralPubMedView ArticleGoogle Scholar
- Alon U: Network motifs: theory and experimental approaches. Nat Rev Genet. 2007, 8: 450-461.PubMedView ArticleGoogle Scholar
- Goentoro L, Shoval O, Kirschner MW, Alon U: The incoherent feedforward loop can provide fold-change detection in gene regulation. Mol Cell. 2009, 36: 894-899.PubMed CentralPubMedView ArticleGoogle Scholar
- Wall ME, Dunlop MJ, Hlavacek WS: Multiple functions of a feed-forward-loop gene circuit. J Mol Biol. 2005, 349: 501-514.PubMedView ArticleGoogle Scholar
- Zhu X, Gerstein M, Snyder M: Getting connected: analysis and principles of biological networks. Genes Dev. 2007, 21: 1010-1024.PubMedView ArticleGoogle Scholar
- Neph S, Stergachis AB, Reynolds A, Sandstrom R, Borenstein E, Stamatoyannopoulos JA: Circuitry and dynamics of human transcription factor regulatory networks. Cell. 2012, 150: 1274-1286.PubMed CentralPubMedView ArticleGoogle Scholar
- Haynes BC, Maier EJ, Kramer MH, Wang PI, Brown H, Brent MR: Mapping functional transcription factor networks from gene expression data. Genome Res. 2013, 23: 1319-1328.PubMed CentralPubMedView ArticleGoogle Scholar
- Re A, Cora D, Taverna D, Caselle M: Genome-wide survey of microRNA-transcription factor feed-forward regulatory circuits in human. Mol Biosyst. 2009, 5: 854-867.PubMed CentralPubMedView ArticleGoogle Scholar
- Siciliano V, Garzilli I, Fracassi C, Criscuolo S, Ventre S, di Bernardo D: miRNAs confer phenotypic robustness to gene networks by suppressing biological noise. Nat Commun. 2013, 4: 2364-PubMedView ArticleGoogle Scholar
- Ogawa S, Lozach J, Benner C, Pascual G, Tangirala RK, Westin S, Hoffmann A, Subramaniam S, David M, Rosenfeld MG, Glass CK: Molecular determinants of crosstalk between nuclear receptors and Toll-like receptors. Cell. 2005, 122: 707-721.PubMed CentralPubMedView ArticleGoogle Scholar
- Rao NA, McCalman MT, Moulos P, Francoijs KJ, Chatziioannou A, Kolisis FN, Alexis MN, Mitsiou DJ, Stunnenberg HG: Coactivation of GR and NFKB alters the repertoire of their binding sites and target genes. Genome Res. 2011, 21: 1404-1416.PubMed CentralPubMedView ArticleGoogle Scholar
- Chinenov Y, Gupte R, Dobrovolna J, Flammer JR, Liu B, Michelassi FE, Rogatsky I: Role of transcriptional coregulator GRIP1 in the anti-inflammatory actions of glucocorticoids. Proc Natl Acad Sci USA. 2012, 109: 11776-11781.PubMed CentralPubMedView ArticleGoogle Scholar
- Adelman K, Kennedy MA, Nechaev S, Gilchrist DA, Muse GW, Chinenov Y, Rogatsky I: Immediate mediators of the inflammatory response are poised for gene activation through RNA polymerase II stalling. Proc Natl Acad Sci U S A. 2009, 106: 18207-18212.PubMed CentralPubMedView ArticleGoogle Scholar
- Gupte R, Muse GW, Chinenov Y, Adelman K, Rogatsky I: Glucocorticoid receptor represses proinflammatory genes at distinct steps of the transcription cycle. Proc Natl Acad Sci U S A. 2013, 110: 14616-14621.PubMed CentralPubMedView ArticleGoogle Scholar
- Escoubet-Lozach L, Benner C, Kaikkonen MU, Lozach J, Heinz S, Spann NJ, Crotti A, Stender J, Ghisletti S, Reichart D, Cheng CS, Luna R, Ludka C, Sasik R, Garcia-Bassets I, Hoffmann A, Subramaniam S, Hardiman G, Rosenfeld MG, Glass CK: Mechanisms establishing TLR4-responsive activation states of inflammatory response genes. PLoS Genet. 2012, 7: e1002401-View ArticleGoogle Scholar
- Mostafavi S, Ray D, Warde-Farley D, Grouios C, Morris Q: GeneMANIA: a real-time multiple association network integration algorithm for predicting gene function. Genome Biol. 2008, 9 (1): S4-PubMed CentralPubMedView ArticleGoogle Scholar
- Paquette J, Tokuyasu T: EGAN: exploratory gene association networks. Bioinformatics. 2010, 26: 285-286.PubMed CentralPubMedView ArticleGoogle Scholar
- Newman ME, Girvan M: Finding and evaluating community structure in networks. Phys Rev E Stat Nonlin Soft Matter Phys. 2004, 69: 026113-PubMedView ArticleGoogle Scholar
- Isserlin R, Merico D, Alikhani-Koupaei R, Gramolini A, Bader GD, Emili A: Pathway analysis of dilated cardiomyopathy using global proteomic profiling and enrichment maps. Proteomics. 2010, 10: 1316-1327.PubMed CentralPubMedView ArticleGoogle Scholar
- Simmen RC, Eason RR, McQuown JR, Linz AL, Kang TJ, Chatman L, Till SR, Fujii-Kuriyama Y, Simmen FA, Oh SP: Subfertility, uterine hypoplasia, and partial progesterone resistance in mice lacking the Kruppel-like factor 9/basic transcription element-binding protein-1 (Bteb1) gene. J Biol Chem. 2004, 279: 29286-29294.PubMedView ArticleGoogle Scholar
- Mangan S, Itzkovitz S, Zaslaver A, Alon U: The incoherent feed-forward loop accelerates the response-time of the gal system of Escherichia coli. J Mol Biol. 2006, 356: 1073-1081.PubMedView ArticleGoogle Scholar
- Yu CY, Mayba O, Lee JV, Tran J, Harris C, Speed TP, Wang JC: Genome-wide analysis of glucocorticoid receptor binding regions in adipocytes reveal gene network involved in triglyceride homeostasis. PLoS One. 2010, 5: e15188-PubMed CentralPubMedView ArticleGoogle Scholar
- Kuo T, Lew MJ, Mayba O, Harris CA, Speed TP, Wang JC: Genome-wide analysis of glucocorticoid receptor-binding sites in myotubes identifies gene networks modulating insulin signaling. Proc Natl Acad Sci U S A. 2012, 109: 11160-11165.PubMed CentralPubMedView ArticleGoogle Scholar
- Uhlenhaut NH, Barish GD, Yu RT, Downes M, Karunasiri M, Liddle C, Schwalie P, Hubner N, Evans RM: Insights into negative regulation by the glucocorticoid receptor from genome-wide profiling of inflammatory cistromes. Mol Cell. 2013, 49: 158-171.PubMed CentralPubMedView ArticleGoogle Scholar
- Lachmann A, Xu H, Krishnan J, Berger SI, Mazloom AR, Ma’ayan A: ChEA: transcription factor regulation inferred from integrating genome-wide ChIP-X experiments. Bioinformatics. 2010, 26: 2438-2444.PubMed CentralPubMedView ArticleGoogle Scholar
- Chen X, Xu H, Yuan P, Fang F, Huss M, Vega VB, Wong E, Orlov YL, Zhang W, Jiang J, Loh YH, Yeo HC, Yeo ZX, Narang V, Govindarajan KR, Leong B, Shahab A, Ruan Y, Bourque G, Sung WK, Clarke ND, Wei CL, Ng HH: Integration of external signaling pathways with the core transcriptional network in embryonic stem cells. Cell. 2008, 133: 1106-1117.PubMedView ArticleGoogle Scholar
- Jiang J, Chan YS, Loh YH, Cai J, Tong GQ, Lim CA, Robson P, Zhong S, Ng HH: A core Klf circuitry regulates self-renewal of embryonic stem cells. Nat Cell Biol. 2008, 10: 353-360.PubMedView ArticleGoogle Scholar
- Mikkelsen TS, Xu Z, Zhang X, Wang L, Gimble JM, Lander ES, Rosen ED: Comparative epigenomic analysis of murine and human adipogenesis. Cell. 2010, 143: 156-169.PubMed CentralPubMedView ArticleGoogle Scholar
- Lefterova MI, Steger DJ, Zhuo D, Qatanani M, Mullican SE, Tuteja G, Manduchi E, Grant GR, Lazar MA: Cell-specific determinants of peroxisome proliferator-activated receptor gamma function in adipocytes and macrophages. Mol Cell Biol. 2010, 30: 2078-2089.PubMed CentralPubMedView ArticleGoogle Scholar
- MacGillavry HD, Cornelis J, van der Kallen LR, Sassen MM, Verhaagen J, Smit AB, van Kesteren RE: Genome-wide gene expression and promoter binding analysis identifies NFIL3 as a repressor of C/EBP target genes in neuronal outgrowth. Mol Cell Neurosci. 2011, 46: 460-468.PubMedView ArticleGoogle Scholar
- Urrutia R: KRAB-containing zinc-finger repressor proteins. Genome Biol. 2003, 4: 231-PubMed CentralPubMedView ArticleGoogle Scholar
- Takahashi K, Sugi Y, Hosono A, Kaminogawa S: Epigenetic regulation of TLR4 gene expression in intestinal epithelial cells for the maintenance of intestinal homeostasis. J Immunol. 2009, 183: 6522-6529.PubMedView ArticleGoogle Scholar
- Huang C, Martin S, Pfleger C, Du J, Buckner JH, Bluestone JA, Riley JL, Ziegler SF: Cutting edge: a novel, human-specific interacting protein couples FOXP3 to a chromatin-remodeling complex that contains KAP1/TRIM28. J Immunol. 2013, 190: 4470-4473.PubMed CentralPubMedView ArticleGoogle Scholar
- Barde I, Rauwel B, Marin-Florez RM, Corsinotti A, Laurenti E, Verp S, Offner S, Marquis J, Kapopoulou A, Vanicek J, Trono D: A KRAB/KAP1-miRNA cascade regulates erythropoiesis through stage-specific control of mitophagy. Science. 2013, 340: 350-353.PubMed CentralPubMedView ArticleGoogle Scholar
- Glass CK, Saijo K: Nuclear receptor transrepression pathways that regulate inflammation in macrophages and T cells. Nat Rev Immunol. 2010, 10: 365-376.PubMedView ArticleGoogle Scholar
- Clark AR, Belvisi MG: Maps and legends: the quest for dissociated ligands of the glucocorticoid receptor. Pharmacol Ther. 2011, 134: 54-67.PubMedView ArticleGoogle Scholar
- Das H, Kumar A, Lin Z, Patino WD, Hwang PM, Feinberg MW, Majumder PK, Jain MK: Kruppel-like factor 2 (KLF2) regulates proinflammatory activation of monocytes. Proc Natl Acad Sci U S A. 2006, 103: 6653-6658.PubMed CentralPubMedView ArticleGoogle Scholar
- Das M, Lu J, Joseph M, Aggarwal R, Kanji S, McMichael BK, Lee BS, Agarwal S, Ray-Chaudhury A, Iwenofu OH, Kuppusamy P, Pompili VJ, Jain MK, Das H: Kruppel-like factor 2 (KLF2) regulates monocyte differentiation and functions in mBSA and IL-1beta-induced arthritis. Curr Mol Med. 2012, 12: 113-125.PubMed CentralPubMedView ArticleGoogle Scholar
- Katakura T, Miyazaki M, Kobayashi M, Herndon DN, Suzuki F: CCL17 and IL-10 as effectors that enable alternatively activated macrophages to inhibit the generation of classically activated macrophages. J Immunol. 2004, 172: 1407-1413.PubMedView ArticleGoogle Scholar
- Kato T, Saeki H, Tsunemi Y, Shibata S, Tamaki K, Sato S: Thymus and activation-regulated chemokine (TARC)/CC chemokine ligand (CCL) 17 accelerates wound healing by enhancing fibroblast migration. Exp Dermatol. 2011, 20: 669-674.PubMedView ArticleGoogle Scholar
- Shi X, Shi W, Li Q, Song B, Wan M, Bai S, Cao X: A glucocorticoid-induced leucine-zipper protein, GILZ, inhibits adipogenesis of mesenchymal cells. EMBO Rep. 2003, 4: 374-380.PubMed CentralPubMedView ArticleGoogle Scholar
- Groner AC, Meylan S, Ciuffi A, Zangger N, Ambrosini G, Denervaud N, Bucher P, Trono D: KRAB-zinc finger proteins and KAP1 can mediate long-range transcriptional repression through heterochromatin spreading. PLoS Genet. 2010, 6: e1000869-PubMed CentralPubMedView ArticleGoogle Scholar
- Iyengar S, Farnham PJ: KAP1 protein: an enigmatic master regulator of the genome. J Biol Chem. 2011, 286: 26267-26276.PubMed CentralPubMedView ArticleGoogle Scholar
- Shin JH, Ko HS, Kang H, Lee Y, Lee YI, Pletinkova O, Troconso JC, Dawson VL, Dawson TM: PARIS (ZNF746) repression of PGC-1alpha contributes to neurodegeneration in Parkinson’s disease. Cell. 2011, 144: 689-702.PubMed CentralPubMedView ArticleGoogle Scholar
- Kino T, Pavlatou MG, Moraitis AG, Nemery RL, Raygada M, Stratakis CA: ZNF764 haploinsufficiency may explain partial glucocorticoid, androgen, and thyroid hormone resistance associated with 16p11.2 microdeletion. J Clin Endocrinol Metab. 2012, 97: E1557-E1566.PubMed CentralPubMedView ArticleGoogle Scholar
- Natoli G: Specialized chromatin patterns in the control of inflammatory gene expression. Curr Top Microbiol Immunol. 2011, 349: 61-72.PubMedGoogle Scholar
- Chen SH, Masuno K, Cooper SB, Yamamoto KR: Incoherent feed-forward regulatory logic underpinning glucocorticoid receptor action. Proc Natl Acad Sci U S A. 2013, 110: 1964-1969.PubMed CentralPubMedView ArticleGoogle Scholar
- He G, Sun D, Ou Z, Ding A: The protein Zfand5 binds and stabilizes mRNAs with AU-rich elements in their 3′-untranslated regions. J Biol Chem. 2012, 287: 24967-24977.PubMed CentralPubMedView ArticleGoogle Scholar
- McConnell BB, Yang VW: Mammalian Kruppel-like factors in health and diseases. Physiol Rev. 2010, 90: 1337-1381.PubMed CentralPubMedView ArticleGoogle Scholar
- Mahabeleshwar GH, Kawanami D, Sharma N, Takami Y, Zhou G, Shi H, Nayak L, Jeyaraj D, Grealy R, White M, McManus R, Ryan T, Leahy P, Lin Z, Haldar SM, Atkins GB, Wong HR, Lingrel JB, Jain MK: The myeloid transcription factor KLF2 regulates the host response to polymicrobial infection and endotoxic shock. Immunity. 2011, 34: 715-728.PubMed CentralPubMedView ArticleGoogle Scholar
- Alder JK, Georgantas RW, Hildreth RL, Kaplan IM, Morisot S, Yu X, McDevitt M, Civin CI: Kruppel-like factor 4 is essential for inflammatory monocyte differentiation in vivo. J Immunol. 2008, 180: 5645-5652.PubMed CentralPubMedView ArticleGoogle Scholar
- Kurotaki D, Osato N, Nishiyama A, Yamamoto M, Ban T, Sato H, Nakabayashi J, Umehara M, Miyake N, Matsumoto N, Nakazawa M, Ozato K, Tamura T: Essential role of the IRF8-KLF4 transcription factor cascade in murine monocyte differentiation. Blood. 2013, 121: 1839-1849.PubMed CentralPubMedView ArticleGoogle Scholar
- Liao X, Sharma N, Kapadia F, Zhou G, Lu Y, Hong H, Paruchuri K, Mahabeleshwar GH, Dalmas E, Venteclef N, Flask CA, Kim J, Doreian BW, Lu KQ, Kaestner KH, Hamik A, Clément K, Jain MK: Kruppel-like factor 4 regulates macrophage polarization. J Clin Invest. 2011, 121: 2736-2749.PubMed CentralPubMedView ArticleGoogle Scholar
- Velarde MC, Zeng Z, McQuown JR, Simmen FA, Simmen RC: Kruppel-like factor 9 is a negative regulator of ligand-dependent estrogen receptor alpha signaling in Ishikawa endometrial adenocarcinoma cells. Mol Endocrinol. 2007, 21: 2988-3001.PubMedView ArticleGoogle Scholar
- Mitchell DL, DiMario JX: Bimodal, reciprocal regulation of fibroblast growth factor receptor 1 promoter activity by BTEB1/KLF9 during myogenesis. Mol Biol Cell. 2010, 21: 2780-2787.PubMed CentralPubMedView ArticleGoogle Scholar
- Patel S, Xi ZF, Seo EY, McGaughey D, Segre JA: Klf4 and corticosteroids activate an overlapping set of transcriptional targets to accelerate in utero epidermal barrier acquisition. Proc Natl Acad Sci U S A. 2006, 103: 18668-18673.PubMed CentralPubMedView ArticleGoogle Scholar
- Sasse SK, Mailloux CM, Barczak AJ, Wang Q, Altonsy MO, Jain MK, Haldar SM, Gerber AN: The glucocorticoid receptor and KLF15 regulate gene expression dynamics and integrate signals through feed-forward circuitry. Mol Cell Biol. 2013, 33: 2104-2115.PubMed CentralPubMedView ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.