Abstract
Chromatin remodeling seems to regulate the patterns of proinflammatory genes. Our aim was to provide new insights into the epigenetic mechanisms that control transcriptional activation of early- and late-response genes in initiation and development of severe acute pancreatitis as a model of acute inflammation. Chromatin changes were studied by chromatin immunoprecipitation analysis, nucleosome positioning, and determination of histone modifications in promoters of proinflammatory genes in vivo in the course of taurocholate-induced necrotizing pancreatitis in rats and in vitro in rat pancreatic AR42J acinar cells stimulated with taurocholate or TNF-α. Here we show that the upregulation of early and late inflammatory genes rely on histone acetylation associated with recruitment of histone acetyltransferase CBP. Chromatin remodeling of early genes during the inflammatory response in vivo is characterized by a rapid and transient increase in H3K14ac, H3K27ac, and H4K5ac as well as by recruitment of chromatin-remodeling complex containing BRG-1. Chromatin remodeling in late genes is characterized by a late and marked increase in histone methylation, particularly in H3K4. JNK and p38 MAPK drive the recruitment of transcription factors and the subsequent upregulation of early and late inflammatory genes, which is associated with nuclear translocation of the early gene Egr-1. In conclusion, specific and strictly ordered epigenetic markers such as histone acetylation and methylation, as well as recruitment of BRG-1–containing remodeling complex are associated with the upregulation of both early and late proinflammatory genes in acute pancreatitis. Our findings highlight the importance of epigenetic regulatory mechanisms in the control of the inflammatory cascade.
Introduction
Chromatin dynamic changes toward open and more accessible euchromatin involve the disruption of interactions with histone-DNA, which in turn include DNA demethylation and histone modification, mainly acetylation, methylation, and phosphorylation (1–8). These modifications create docking sites for the recruitment of proteins or complexes leading to active gene transcription (1–3, 5–8). The epigenetic balance of histone acetylation/deacetylation is regulated by the writers, histone acetyltransferases (HATs), which incorporate the acetyl group to site-specific lysines using acetyl-CoA as donor, and the erasers, histone deacetylases (HDACs), which hydrolyze the acetamide bond. One of the primary functions of HDACs is to keep chromatin transcriptionally silent due to the removal of acetyl groups from the N terminal tails of histones H3 and H4 (8). Although other modifications of core histones have similar effects, the acetylation of histones H4 and H3, and the trimethylation of lysine 4 of H3 (H3K4me3), generally promote gene activation, whereas di- or trimethylation of lysine 27 of histone H3 (H3K27me2 and H3K27me3) lead to gene silencing (9, 10).
HATs and HDACs play a central role in gene regulation in inflammation (11). Indeed, histone acetylation via CBP/p300 HATs [also known as KAT3A (CBP) and KAT3B (p300)] coordinates the expression of proinflammatory cytokines, particularly through NF-κB and STAT pathways (12–14). It is noteworthy that NF-κB is able to access regions of condensed chromatin and trigger transcriptional activation because TNF-α rapidly and substantially reduced HDAC1 protein levels through protein degradation (15).
In addition, a shift in epigenetic marks seems to occur in H3 to change the repressed state of transcription to competent transactivation of proinflammatory genes. The basal, transcriptionally inactive state of chromatin in the promoters of these genes is characterized by the presence of methylated H3 lysine 9 coupled with unphosphorylated H3 serine 10, which shifts to the transcriptionally active state of demethylated H3 lysine 9 and phosphorylated H3 serine 10 (16, 17).
During the inflammatory cascade, many genes seem to be activated following two different patterns. The primary response (early) genes can be induced without de novo protein synthesis, although some exhibit delayed kinetics (18, 19). The secondary response (late) genes require de novo protein synthesis and are induced more slowly because that synthesis needs signaling molecules or cytokines (20). Chromatin remodeling regulates the expression of primary response genes with delayed kinetics and of secondary response genes at least through the recruitment of SWI/SNF remodeling complexes that contain either BRG1 or BRM as the catalytic subunit (19). Promoters of rapid primary response genes exhibit a high level of histone acetylation and H3K4me3 under basal conditions, which results in an open chromatin structure (19–21). Acetylation of histone H4 and H3K4me3 are found at the promoters of both primary and secondary response genes in macrophages stimulated with LPS (22). Nevertheless, the large number of acetylations and the complexity of the regulatory effects hinder the assignment of direct biological roles to a single acetylation event (23).
The present work focuses on severe acute pancreatitis as an experimental model of acute inflammation that is initially localized in the pancreatic gland but evolves into a systemic inflammatory response, and eventual multiple organ failure (24–26). Previous evidence from our group suggests that epigenetic mechanisms control TNF-α expression and may have a key role in this disease (14, 27). Indeed, activation of Tnf-α gene was accompanied by a strictly ordered recruitment or release of transcriptional factors (ELK-1, SP1, NF-κB and EGR-1), and chromatin modification complexes [HDAC1, HDAC2, GCN5 (KAT2A), PCAF (KAT2B) and CBP (KAT3A)], as well as by an ordered, increased level of histone H3K9ac, H3K14ac, H3K18ac, H4K5ac, and H3K4me. In the current study, our aim has been to gain new insights into the common epigenetic mechanisms that control transcriptional activation of early- and late-response genes in the initiation and development of severe acute pancreatitis.
Materials and Methods
Animals
Male Wistar-Furth rats (250–300 g body weight) were used. They were fed a standard laboratory diet and tap water ad libitum, and were subjected to a 12 h light–dark cycle. All animals were cared for and handled according to the criteria outlined in the Guide for the Care and Use of Laboratory Animals prepared by the National Academy of Sciences and published by the National Institutes of Health (NIH Publication 86-23 revised 1985). The Research Committee of the School of Medicine (University of Valencia, Spain) approved the study protocol.
AR42J acinar cells
The acinar cell line AR42J, derived from an exocrine rat pancreatic tumor (CRL 1492; American Type Culture Collection, Manassas, VA) was grown in DMEM (Invitrogen, Carlsbad, CA) containing 25 mM glucose, 100 μg/ml penicillin, 100 μg/ml streptomycin, and 25 μg/ml fungizone, supplemented with 10% FBS. To activate the Egr-1 and Tnf-α genes, AR42J cells were treated with 0.3% of sodium taurocholate (Sigma) and 50 ng/ml of TNF-α (Sigma) for the indicated times. To inhibit the cell signaling pathways, AR42J cells were pretreated with cycloheximide (25 μg/ml) for 1 h, and incubated with the following compounds: PD98059 (100 μM), SB203580 (100 μM), Wortmanin (1 μM), H89 (10 μM), JNK (100 μM), ERKs (20 μM), NF-κB (25 μM), STAT3 (100 μM), and JAK-STAT1 (S14-95 5 μg/ml) inhibitors (Calbiochem). Subsequently, the cells were incubated with sodium taurocholate or TNF-α.
Acute necrotizing pancreatitis
Animals were anesthetized with an i.p. administration of ketamine (80 mg/kg body weight) and acepromazine (2.5 mg/kg body weight)). Acute necrotizing pancreatitis was induced by retrograde injection into the biliopancreatic duct of sodium taurocholate (3.5%) (Sigma) in a volume of 0.1 ml/100 g body weight using an infusion pump (Harvard Instruments) (27). In control rats, 0.9% NaCl (0.1 ml/100 g body weight) was infused into the biliopancreatic duct. For each group 6–12 animals were used for the in vivo experiments. Rats were anesthetized as previously described then sacrificed at the indicated times after the infusion of taurocholate or saline. Serum lipase activity was measured, and histological studies were performed to confirm the appropriate induction of necrotizing pancreatitis.
RNA isolation and analysis of mRNA level by real-time PCR and quantitative PCR
Total RNA from pancreatic tissue or AR42J cells was isolated by the guanidinium thiocyanate method (28). In the case of the pancreas, the tissue was previously immersed in 1 ml of RNA-later solution (Ambion) to stabilize the RNA. The isolated RNA (2 μg/lane) was size-fractionated by electrophoresis in a 1% agarose/formalin gel, and stained with ethidium bromide to assess the quality of the RNA. The cDNA used as template for amplification in the PCR assay was obtained by the reverse transcription reaction using SuperScript II (Invitrogen), with random hexamers as primers, starting with equal amounts of RNA. As a PCR internal control, 18S rRNA was simultaneously amplified. To obtain similar PCR band intensities, competitor oligonucleotides were added to the assay in a 3:7 proportion of normal 18S rRNA oligonucleotides/competitor 18S rRNA oligonucleotides. The competitors correspond to the same 18S rRNA oligonucleotide sequence, except that their 3′ termini were blocked with an amino group.
Real-time quantitative PCR was performed using dsDNA binding dye Syber Green PCR Master mix in an ABI GeneAmp 7000 Sequence Detection System (Applied Biosystems). Each reaction was run in triplicate, and the melting curves were constructed using Dissociation Curves Software (Applied Biosystems) to ensure that only a single product was amplified. As real-time quantitative PCR control 18S rRNA or ACTB was also analyzed. The specific primers used for real-time PCR (RT-PCR) analysis are shown in Table I. The threshold cycle (Ct) was determined, and the relative gene expression was expressed as follows: fold change = 2−Δ(ΔCt), where ΔCt = Cttarget−Cthousekeeping and Δ(ΔCt) = ΔCttreated−ΔCtcontrol.
Chromatin immunoprecipitation assay and RNApol-chromatin immunoprecipitation
Cross-linking of chromatin, chromatin immunoprecipitation (ChIP), and RNApol-ChIP procedures were performed using the method of Sandoval et al. (27). Briefly, isolated nuclei from formaldehyde–cross-linked pancreas or AR42J cells were lysed and cross-linked chromatin was sonicated to yield fragments of ∼500 bp. Diluted soluble chromatin fragments were precleared with blocked Protein A/G–Sepharose, to discard non-specifically bound chromatin fragments. Immunofractionation of complexes was carried out by adding 2 μg of the corresponding Abs (see Table II) to aliquots containing 50 μg of DNA each. The immunocomplexes were recovered by centrifugation at 13,500 × g for 1 min after adding blocked Protein A/G–Sepharose and washing extensively. Immunoselected chromatin was eluted, and the formaldehyde cross-linking was reverted at 65°C overnight. The DNA from all samples was purified with a PCR purification kit (Qiagen), and used for PCR analysis. The analysis was carried out using the primers shown in Table III. PCR fragments were size-fractionated by 2% agarose gel electrophoresis, stained with ethidium bromide, and analyzed with an FLA3000 electronic autoradiography system (Fujifilm), using ImageJ software (http://rsbweb.nih.gov/ij/). The ChIP assay of histone H3 was used as an internal control for in vivo and in vitro experiments to confirm the lack of significant fluctuation (Supplemental Fig. 1).
Western blotting for total and nuclear protein
AR42J cell samples were resuspended in a PBS buffer supplemented with protease inhibitor mixture (Sigma), added immediately before its use, at a concentration of 5 μl/ml. For total protein isolation, samples were incubated in total extraction buffer [10 mM Tris–HCl (pH 7.5), 0.25 M sucrose, 5 mM EDTA, 50 mM NaCl, 30 mM sodium pyrophosphate, 50 mM sodium fluoride, 100 μM sodium orthovanadate, and 1% Igepal]. Debris was removed by centrifugation at 900 × g at 4°C for 10 min, and the supernatant obtained was used for Western blot analysis. For nuclear protein purification, AR42J cell samples were incubated with cell lysis buffer (5 mM HEPES pH 8, 85 mM KCl, 0.5% NP40 supplemented with protease inhibitor mixture) for 10 min, centrifuged at 1500 × g at 4°C for 10 min, and the nuclear pellet was subsequently incubated in nuclear lysis buffer (50 mM TrisHCl pH 8.1, 10 mM EDTA, 1% SDS) for 10 min.
SDS-PAGE was used to separate 30 μg of protein, which was transferred to nitrocellulose membranes (Schleicher & Schuell). EGR-1 was determined by Western blotting and chemiluminescence detection using the Phototope-HRP detection kit (Cell Signaling Technology). As loading control, β-tubuline (Abcam) was used.
Nucleosomal occupancy
To analyze nucleosomal occupancy, briefly, nuclei from AR42J cells were suspended in RBS buffer [15 mM TrisHCl pH 7.5, 60 mM KCl, 15 mM NaCl, 3 mM MgCl2, 20% (v/v ) glycerol, 5 mM β-mercaptoethanol and 1 mM PMSF]. Eighty micrograms of chromatin DNA were digested with micrococcal nuclease for 5 min and blocked with EDTA 50 mM and SDS 1% (v/v). DNA from digested nuclei were enriched in fragments of mononucleosomal size (150–200 bp). These fragments are used as a template for quantitative PCR using tiled amplicons encompassing the promoter region of Egr-1 gene (see Table IV). To normalize for the different efficiency of primers, sonicated genomic DNA was also amplified by PCR. To evaluate the affinity score for nucleosomes located in the Egr-1 promoter region, the nucleosome positioning prediction software (NuPoP) was used (http://nucleosome.stats.northwestern.edu/).
Immunolocalization
For immunolocalization analysis AR42J cells were cultured in chamber slides and fixed with formalin 10% (v/v), permeabilized with 10% Triton X-100 (v/v) in PBS and blocked with goat normal serum at 10% (v/v) in PBS. EGR-1 was determined by immunofluorescence. AR42J cells were incubated with DAPI 3 μM (Invitrogen). Visualization images were acquired with a Laser Scanning Spectral Confocal Microscope (Leica TCS-SP2) connected with a Leica DM1RB phase contrast microscope.
Statistical analysis
Results are expressed as mean ± SD with the number of experiments given in parentheses. Statistical analysis was performed in two steps. First, a one-way ANOVA was carried out to find significance in the overall comparison of groups. Then, differences between individual groups were investigated by the T (Scheffé) test. Differences were considered to be significant at p < 0.05.
Results
Gene expression profiles for early and late genes in the pancreas during acute pancreatitis
The different profiles of gene expression for early (Egr-1, Atf-3, Il-1β, MyD118) or late genes (Nos-2, Il-6, Pap, Tnf-α) in the head of pancreas during acute pancreatitis are depicted in Fig. 1. Similar findings were also obtained in the pancreas tail (data not shown). Early genes exhibited a rapid upregulation, reaching the maximum mRNA stationary level 1 h after taurocholate infusion (Fig. 1A), and progressively decreasing thereafter. However, late genes increased their mRNA stationary levels progressively, reaching a maximum at 3 h or after (Fig. 1B). Gene expression was studied using 18S rRNA as a reference. No significant differences in the expression of 18S rRNA occurred in the course of pancreatitis using ACTB as reference in this case (Fig. 1C, Tables I–IV).
Gene expression pattern during taurocholate-induced acute necrotizing pancreatitis in rats. (A) Early gene expression (Egr-1, Atf-3, IL-1β, and Myd118), (B) late gene expression (Nos-2, IL-6, Pap, and Tnf-α), and (C) control gene expression (β-actin). Steady-state mRNA levels were measured by quantitative RT-PCR in the pancreas of mice treated with saline (control) or taurocholate acid. Quantitative RT-PCR values were normalized against those obtained for rRNA 18S analysis, giving an arbitrary value of 1 to the control sample. The error bars correspond with the SD of three to four independent RT-PCR measurements. The statistical significance is indicated as ***p < 0.001, **p < 0.01, *p < 0.05.
We further analyzed the real-time transcriptional activity of early- or late-response genes by measuring the actual transcription rate using RNApol ChIP, a technique that detects elongating RNA polymerase II at the desired genes, avoiding interference with polymerase paused at the promoter. Fig. 2 shows the recruitment of RNA polymerase to the coding regions (elongating), and to the promoter of early (Fig. 2A) and late genes (Fig. 2B). The maximum binding of RNA polymerase II to the coding region was found at 0.5–1 h for early genes, whereas the maximum recruitment for late genes was at 3–6 h. Hence, the upregulation of early and late genes was due to active transcription following the profile of gene expression shown in Fig. 1. ACTA1 and ACTB were used, respectively, as negative and positive controls of gene transcription (Fig. 2C).
Actual transcription rate for selected early, late, and control genes in taurocholate-induced acute necrotizing pancreatitis in rats. Upper panels show steady-state mRNA levels were measured by semiquantitative RT-PCR for (A) early (Egr-1 and Atf-3), (B) late (Nos-2, IL-6), and (C) control genes (α-actin as negative and β-actin as positive). The rRNA 18S was used as an internal control, and α- and β-actin genes were used as negative and positive control, respectively, of the RT-PCR analysis. The middle and lower panels depict RNApol ChIP analysis for all selected genes in promoter and transcriptional regions, respectively.
Early- and late-response genes were also upregulated in other tissues, such as liver and lung, although with a delayed profile (Supplemental Fig. 2). These findings show the different time-course between the local and the systemic inflammatory response characteristics of severe acute pancreatitis.
Histone modifications in early and late genes in the pancreas during pancreatitis
The levels of relevant histone modifications in the promoter region of early and late genes were estimated by ChIP assay. These histone modifications were selected according to their prominent role in chromatin remodeling (9, 10, 16, 17, 19–22), as mentioned in the Introduction. Fig. 3A shows acetylation or methylation of histones H3 and H4 in two representative genes for each group, i.e., Egr-1 and Atf3 for early genes, and Nos-2 and Il-6 for late genes. The modification characteristics of early genes were the transient acetylations H3K14ac, H3K27ac, and H4K5ac with a maximum at 1 h. In addition, H3K4me3, and H3K9ac were markedly reduced at 6 h in early genes (Fig. 3A, upper panel). The major features of late genes were the increased levels of H3K14ac, and H3K4me3, the latter particularly at 6 h. Nevertheless, Nos-2 and Il-6 exhibited different profiles concerning H3K4me2, H3K4ac, and H4K5ac. Il-6 exhibited a high H3K4me2 level under basal conditions decreasing at 6 h, whereas this modification was completely absent in the promoter of Nos-2. The H3K9ac mark increased at 6 h in Nos-2; however, it decreased markedly at 6 h in Il-6. H4K5ac increased transiently at 1 h in Il-6 but it did not change in Nos-2 (Fig. 3A, middle panel). Neither of these changes was found in the gene of α-actin, which was used as a negative control (Fig. 3A, lower panel).
Epigenetic promoter characterization by ChIP of selected early, late, and control genes during taurocholate-induced acute necrotizing pancreatitis in rats. (A) Histone modification occupancy analyzed with indicated Abs were identified using specific primers in promoters of early (upper panel), late (middle panel), and control genes (lower panel). (B) Recruitment of histone modification complexes were analyzed by ChIP using Abs recognizing histone acetyltransferase complex CBP, and histone deacetylase complex Sin3A in the aforementioned promoter genes.
The recruitment of HATs or HDACs was also studied in early and late genes. The HAT CBP (KAT3A) was the histone modifier complex mainly associated with upregulation of early and late genes, except in the case of Egr-1. It should be highlighted that binding of HAT CBP occurred at 1 h in the early gene Atf3 but at 3–6 h in the late genes Nos-2 or Il-6. None of the HAT complexes was found at the α-actin gene (Fig. 3B).
Recruitment of transcription factors to promoters of early and late genes in the pancreas during pancreatitis
It is well known that histone modifier complexes are recruited to specific regions of genes through transcriptional factors, or even through other histone modifier complexes, allowing gene activation in response to cellular signals. After in silico analysis of the putative binding sites for transcription factors (TRANSFAC analysis), the candidate transcription factors were analyzed by ChIP assay. Fig. 4A shows that the rapid upregulation of Egr-1 was associated with recruitment of ELK1 and SP1, in particular. It is noteworthy that EGR-1 was recruited to its promoter under basal conditions but was rapidly released when Egr-1 was upregulated, returning to the promoter when Egr-1 expression progressively decreased. Hence, EGR-1 seems to exert a negative feedback mechanism in the regulation of its own expression. The recruitment of transcription factors was different for the early gene Atf3 (Fig. 4), which was characterized by binding of EGR-1 and ATF3 at early time points. Therefore, in opposition to EGR-1, ATF3 seems to exert a positive feedback mechanism in the control of its expression.
Transcriptional factor recruitment by ChIP in selected early, late, and control genes during taurocholate-induced acute necrotizing pancreatitis in rats. ChIP analysis of transcriptional factors occupancy in (A) early, (B) late, and (C) control genes.
The upregulation of late genes was associated with recruitment of C/EBPβ and NF-κB, in particular, concomitant with the highest expression of these genes (Fig. 4B). None of the transcription factors studied were recruited to the promoter of α-actin as control gene (Fig. 4C).
Chromatin remodeling of Egr-1 promoter upon stimulation of pancreatic acinar cells in vitro
The bile salt taurocholate also triggered a rapid upregulation of Egr-1 in vitro in AR42J pancreatic acinar cells (Fig. 5), similar to the induction found in vivo in the pancreas upon taurocholate-induced acute pancreatitis. The upregulation found in vitro was also due to active transcription, as RNA polymerase II was intensely recruited to the promoter and to the coding region of the gene. Fig. 6A shows that recruitment of transcription factors EGR-1, SP1, and CREB as well as remodeling complex containing BRG1 was associated with induction of Egr-1 expression in vitro. Importantly, the negative feedback regulation mediated by EGR-1 in vivo was not observed in AR42J cells, which additionally exhibited a different profile in the recruitment of transcription factors and modifier complexes. Fig. 6B shows nucleosomal remodeling measured by micrococcal nuclease protection assay, which characterized Egr-1 upregulation in AR42J cells in vitro. It is worth noting that nucleosomal movement or eviction at 30 min and 1 h favored the accessibility of transcription factors and therefore transcriptional activation of Egr-1.
The early Egr-1 gene expression in taurocholate-activated AR42J acinar cells. (A) and (B) RNApol ChIP analysis in transcriptional and promoter regions, respectively. (C) and (D) Semiquantitative RT-PCR and quantitative RT-PCR analysis of the Egr-1 mRNA level. rRNA 18S was used as an internal control and β-actin as negative control of the RT-PCR analysis, and α-actin as negative control of ChIP assay. The statistical significance is ***p < 0.001.
Transcriptional factor occupancy and nucleosomal remodeling in EGR-1 during taurocholate-activated AR42J acinar cells. (A) ChIP assay of transcription factors and chromatin remodeling complex binding in Egr-1 promoter in taurocholate-activated AR42J cells. The α-actin gene has been included as negative control of the ChIP experiment. (B) Changes in nucleosome occupancy in the promoter of the Egr-1 gene at different times after taurocholate treatment. (B) (upper panel) Sequence-based prediction of nucleosome affinity carried out using the NuPoP software tool. (B) (middle panel) The micrococcal nuclease protection was determined and plotted against the distance to TSS. The plotted experimental points correspond to the means ± SE of three determinations. (B) (lower panel) Map of the Egr-1 region under study, showing the location of cis elements, and the experimental positions of the nucleosomes, in non-induced (solid line) and induced (dashed line) Egr-1 gene. At the bottom, the position of their centers in relation to TSS are shown.
ERK1/2 and JNK mediate Egr-1 upregulation leading to EGR-1 nuclear translocation
To identify the signaling network that drives the inflammatory stimulus to the Egr-1 promoter, we analyzed the major signaling pathways related to inflammation by blocking them with specific inhibitors. Fig. 7A shows that taurocholate-induced Egr-1 upregulation in AR42J cells was blocked by inhibition of MEK or JNK, but not by inhibition of ERK, p38, PI3K, protein kinase A, NF-κB or STAT3 pathways. In addition, the induction of Egr-1 expression was associated with nuclear translocation of EGR-1 in AR42J cells as shown by Western blot (Fig. 7B–C) and immunohistochemistry (Fig. 7D).
Upstream Egr-1–activated pathways and nuclear translocation in taurocholate-activated AR42J acinar cells. (A) Egr-1 fold induction after taurocholate treatment compared with DMSO treated AR42J cells, and the abrogated activation or not using the indicated pathways inhibitors. (B) Coomassie gel with nuclear and total protein purification staining and different BSA protein at different concentrations. (C) Western blot analyzing EGR-1 protein levels at nuclear and total fractions. TUBULIN protein levels were used as reference. Lower panel depicts ImageJ analysis of Western blot levels. (D) Confocal laser scanning microscope images using Ab–recognizing EGR-1 (green) during taurocholate-activated AR42J acinar cells. Nuclei were immunostained using DAPI (red).
Epigenetic changes of Nos-2 promoter induced in vitro by TNF-α in pancreatic acinar cells
Taurocholate only upregulated Egr-1 in vitro but not the other inflammatory mediators typically induced in vivo. Indeed, Atf-3, Nos-2, Pap, and Tnf-α were not upregulated by taurocholate in vitro (data not shown). Hence, we decided to investigate the classical inflammatory cascade triggered by TNF-α in vitro. Fig. 8A–D shows that the TNF-α–induced upregulation of Nos-2 in AR42J cells was associated with recruitment of NF-κB, and C/EBPβ to its promoter, as well as with K14H3ac, and K18H3me3 epigenetic marks. These histone modifications were also associated with recruitment of HAT CBP together with release of HDAC1 and HDAC3. This in-depth and detailed study of epigenetic changes in vitro under very precise conditions provided similar findings to those observed in vivo with the exception of Egr-1.
Promoter characterization by ChIP of Nos-2 TNFa-activated AR42J acinar cells. The immunoprecipitated samples were analyzed by PCR using primers of the Nos-2 promoter region with indicated Abs that recognize (A) transcription factors, (B) histone modification and (C) histone modification complexes. A-actin was used as negative control of ChIP assay.
TNF-α triggers Nos-2 up-regulation through MAPK and NF-κB in pancreatic acinar cells in vitro
Fig. 9 shows that the slow and progressive Nos-2 upregulation induced by TNF-α in AR42J pancreatic acinar cells in vitro was also mediated by active transcription as RNA polymerase II was recruited to the coding region (Fig. 9A–C). TNF-α–induced upregulation of Nos-2 was completely blocked by the inhibition of p38 MAPK, and partially restrained by inhibition of MEK/ERK, JNK, NF-κB, or JAK-STAT1. However, inhibition of PI3K, STAT3, or PKA did not significantly affect Nos-2 upregulation (Fig. 9D).
The late Nos-2 gene expression and upstream pathways activated in TNF-α activated AR42J acinar cells. (A) RNApol ChIP analysis for Nos-2 and α-actin was used as negative control of ChIP assay. (B) and (C) Semiquantitative RT-PCR and quantitative RT-PCR analysis of the Nos-2 mRNA level. rRNA 18S was used as an internal control and β-actin as negative control of the RT-PCR analysis. (D) Nos-2 fold induction after TNF-α treatment compared with DMSO treated AR42J cells and the abrogated activation or not using the indicated pathways inhibitors. The statistical significance is indicated as follows: ***p < 0.001, **p < 0.01, *p < 0.05.
Chromatin remodeling profile for early and late genes in vitro
Fig. 10 summarizes the pattern of chromatin remodeling for early genes in vitro taking Egr-1 as reference. The rapid upregulation of early genes would be mediated by recruitment of transcription factors EGR-1 and SP1, as well as histone modifiers CBP and BRG1, together with phosphorylation of transcription factors mediated by MEK1/2 and JNK and changes in nucleosomal positioning. On the other hand, Fig. 11 shows that the slow upregulation of late genes would be first mediated by release of HDAC1 and HDAC3, and subsequently by recruitment of NF-κB, C/EBPβ, CBP, as well as phosphorylation events mediated by all MAPK, especially p38.
Schematic model for early Egr-1 gene activation during acute necrotizing pancreatitis. Pathways activation, transcription factor occupancy, RNApol II (red oval) and nucleosomes (gray ovals) are depicted at different times. Traffic lights on the right indicated gene expression status (red: silencing, and green: activated).
Schematic model for late Nos-2 gene activation during acute necrotizing pancreatitis. Pathway activation, transcription factor occupancy, RNApol II (red oval) and histone modifications are depicted at different times after TNF-α treatment. Traffic lights on the right indicate gene expression status (red: silencing, yellow: potentially activated, and green: activated).
Discussion
Common inflammatory cascades seem to be involved in the etiology of acute pancreatitis, and cytokines play a major role in it (14). IL-1β and TNF-α are considered initiators of the inflammatory cascade because they activate NF-κB, trigger the secretion of other proinflammatory mediators, and induce their own secretion by a positive feedback mechanism, leading to amplification of the inflammatory cascade. However, they follow different patterns of secretion through independent pathways. IL-1β is mainly involved in the early response associated with the inflammasome, whereas TNF-α is a major driver of the late response (14, 29).
Histone acetylation coordinates the induction of early and late inflammatory genes
HATs and HDACs coordinate the dynamic transcriptional regulation of genes involved in the inflammatory response mainly through NF-κB (11). Indeed, histone acetyltransferases CBP (KAT3A), and p300 (KAT3B) are coactivators of p65 (RelA) (30, 31). Thus, recruitment of CBP and/or p300 complexes and their corresponding histone acetyltransferase activity are required for the transcriptional activation of numerous NF-κB driven promoters, such as those of Il-6, Il-8, and Nos-2 (30, 32, 33). At the transcriptional level, selectivity is conferred by interactions between NF-κB and other transcription factors, and coactivators that form specific enhanceosome complexes in association with specific promoters (34). It was previously reported that NF-κB cooperates with STAT1 for Nos-2 upregulation induced by proinflammatory cytokines (35, 36). Here we show that NF-κB contributes to Nos-2 upregulation in pancreatic acinar cells together with JAK-STAT1, ERK, and JNK, but p38 seems to be the major driver of this induction.
In epithelial cells NF-κB induces acetylation, preferentially of lysines 8 and 12 of histone H4, targeting primarily at NF-κB responsive regulatory elements on proinflammatory genes (37). Sepsis induced by cecal ligation and puncture in mice increased both histone H3 and H4 acetylation levels in lungs, in parallel with the inflammatory response (38).
In our in vivo model of acute inflammation the induction of early genes, such as Egr-1 and Atf-3, was associated with increased histone acetylation, particularly H3K14, H3K27 and H4K5. In addition, c-Fos, MyD118, and Il-1β were also identified as early genes. The upregulation of late genes, such as Nos-2 and Il-6, was also associated with increased histone acetylation, particularly in H3K14. Pap and Tnf-α were also identified as late genes. In addition, the upregulation of both early and late genes, with the exception of Egr-1, was associated with recruitment of HAT CBP, which was rapid for early genes and slow for late genes.
The upregulation of early and late genes was due to active transcription following the profile of gene expression. mRNAs of most inflammatory mediators exhibit short half-lives often caused by the presence of AU-rich elements in the 3′untranslated region (3′UTR), which promotes mRNA decay (39, 40). In agreement, our findings show the profile of active transcription using RNApol ChIP was parallel with that of steady state levels obtained by RT-PCR.
Among the early response genes, Egr-1 and Atf-3 were more deeply studied due to their relevant role in the control of the initial phase of the inflammatory cascade. EGR-1 plays a major role in TNF-α expression in monocytes (41), and it is also capable of enhancing p300 expression by itself (42). We have shown that EGR-1 upregulation is associated with its nuclear translocation, which may induce not only early-expressed genes, such as Atf-3 as it binds to its promoter (Fig. 4), but also late genes, as EGR-1 binds to the Tnf-α promoter during gene activation (27). As we have previously shown, TNF-α in turn activates inflammatory late genes, such as Nos-2, Icam or Il-6, in the pancreas during acute pancreatitis, because their upregulation was abrogated in TNF-α knockout mice (27).
In contrast, ATF-3 is a member of the ATF-3/CREB subfamily of the basic-region leucine zipper family that may interact directly with HDAC1 and recruit HDAC1 to the ATF/NF-κB sites in the Il-6 gene promoter, leading to a blockade of NF-κB binding, and inhibition of inflammatory gene transcription (43).
Tnf-α is considered by some authors an early-response gene, because it is rapidly transcribed following exposure to pathogens or signals of inflammation and stress (44, 45). It has been reported that Tnf-α exhibits rapid transcription after stimulation, independent of new protein synthesis (44). However, our results clearly show that Tnf-α expression follows the profile of late-response genes in vivo. Consequently, the role of TNF-α as a strict early-response mediator during the inflammatory response in vivo should be revised.
Our in vitro experiments show that histone modifications change in the promoter of early-response genes, particularly for Egr-1, in comparison with the in vivo conditions, whereas they were maintained for late-response genes, with a general increase in H3K14ac. Importantly, the in vitro upregulation of Egr-1 was associated with recruitment of the HAT CBP– and the BRG-1–containing modifier complex, which drives a marked nucleosomal shift from a higher nucleosomal affinity position to a lower one, revealing the CREB binding site during gene activation.
The in vitro histone modifications of late genes were associated not only with the recruitment of HAT CBP but also with the release of HDAC1 and HDAC3. Interestingly, secondary genes appear to be more sensitive to changes in histone acetylation as the synthetic compound I-BET, which mimics acetylated histones, interferes with the recognition of acetylated histones by the BET family of proteins disrupting chromatin complexes, reducing the levels of H3 and H4 acetylation on LPS-induced gene promoters, and suppressing many secondary response genes (46).
Histone methylation during the inflammatory cascade
Specific histone methylation marks also arise during the inflammatory cascade. Thus in a monocytic cell line stimulated with TNF-α, methylation of H3R17 together with the acetylation of H3K9 and H3K14 was induced at the promoters of NF-κB target genes in a CBP/p300 dependent manner (47). Furthermore, stimulation of monocytes led to increased histone acetylation as well as to methylation of H3K4 (44).
Importantly, H3K4me2 and H3K4me3 occur at the transcription initiation site, with trimethylation correlating with active transcription and dimethylation signaling a state of readiness or competence (48, 49). The upstream trimethylation mark promotes phosphorylation of the C terminal domain of RNA polymerase II, which is required for transcription initiation (50). Nevertheless, exceptions to this pattern have been described (51, 52). Thus, Ash1l (KMT2H), an H3K4 methyltransferase, suppressed IL-6 and TNF-α production in TLR-triggered macrophages, and, in addition, protected mice from sepsis (53). Furthermore, trimethylation in other positions may be repressive, and thus Jumonji domain-containing protein 3 (JMJD3), also called lysine-specific demethylase 6B (KDM6B), enhances gene expression by demethylating repressive H3K27me3 epigenetic marks (54). In vitro experiments in macrophages showed that the pattern of chromatin remodeling varies between primary response genes and secondary response genes (19). Particularly, a high level of histone acetylation and H3K4me3 were found under basal conditions in primary response genes (19–21), whereas acetylation of histone H4 and H3K4me3 increased at promoters of both primary and secondary genes in response to LPS (22).
Here we show that chromatin remodeling of early genes during the inflammatory response in vivo is characterized by a rapid and transient increase in H3K14ac, H3K27ac, and H4K5ac, together with a late reduction in H3K4me3 and H3K9ac. In contrast, chromatin remodeling in late genes is characterized by a late but marked increase in H3K4me3. Nevertheless, marked differences in H3K4me2 were found among the different late genes, particularly specific profiles corresponding to Il-6 and Nos-2.
Our in vitro experiments show that histone modifications were maintained for late response genes with a general increase in H3K18me3. The in vitro upregulation of Egr-1 was associated with the recruitment of a modifier complex containing BRG-1. Importantly, methylation of H3K4 facilitates interactions with BRG-1 chromatin remodeling factors (55). BRG-1 is a catalytic component of the mammalian chromatin remodeling complex, which is recruited to the promoters of cell adhesion molecules (CAM) by NF-κB/p65, and promotes their transactivation as well as leukocyte adhesion in endothelial cells (56).
Recruitment of transcription factors is driven by JNK and p38 MAPK
The signaling pathways and transcription factors that drive the inflammatory cascade were elucidated in vitro. MEK/ERK and JNK would be responsible for the upregulation of early response genes, whereas the induction of late-response genes should be mainly ascribed to p38. Concerning the transcription factors, EGR-1 and SP1 were rapidly recruited in early genes, whereas NF-κB and C/EBPβ bound later to promoters of late genes. It is worth noting that EGR-1 seems to promote Egr-1 upregulation in vitro in pancreatic acinar cells, and yet by negative feedback it seems to reduce Egr-1 upregulation in vivo in the pancreas during acute inflammation. It was previously reported that early and secondary response genes rely on similar transcription factors (NF-κB, AP-1, and C/EBP), with the binding of NF-κB being delayed in the secondary-response genes (57). However, our findings support a different signaling network between early- and late-response genes.
In conclusion, here we show that the upregulation of early and late inflammatory genes relies on histone acetylation associated in most cases with the recruitment of HAT CBP. Chromatin remodeling of early genes during the inflammatory response in vivo is characterized by a rapid and transient increase in H3K14ac, H3K27ac, and H4K5ac, and by the recruitment of the histone remodeling complex containing BRG-1. Chromatin remodeling in late genes is characterized by a late and marked increase in histone methylation, particularly H3K4me3. The JNK and p38 MAPK drive the recruitment of transcription factors and the subsequent upregulation of early and late inflammatory genes, which is associated with nuclear translocation of the transcription factor EGR-1. Our findings highlight the importance of epigenetic regulation in the control of the inflammatory cascade and may contribute to identifying the mechanism of action of numerous anti-inflammatory drugs.
Disclosures
The authors have no financial conflicts of interest.
Acknowledgments
We thank Maria del Carmen Cebrian for her technical support.
Footnotes
This work was supported by Grants SAF2012-39694 and SAF2015-71208-R with Fondo Europeo de Desarrollo Regional funds from the Spanish Ministry of Economy and Competitiveness to J. Sastre and Consolider-Ingenio CSD2006-49 (Network Group Valencia) from the Spanish Ministry of Science and Technology to G.L.-R. I.F. received a fellowship from Programa de Pós-Doutorado no Exterior, which belongs to the Conselho Nacional de Desenvolvimento Científico e Tecnológico.
The online version of this article contains supplemental material.
Abbreviations used in this article:
- ChIP
- chromatin immunoprecipitation
- Ct
- threshold cycle
- HAT
- histone acetyltransferase
- HDAC
- histone deacetylase
- RT-PCR
- real-time PCR
- TSS
- transcriptional start site.
- Received November 13, 2015.
- Accepted September 16, 2016.
- Copyright © 2016 by The American Association of Immunologists, Inc.
References
In this issue
