CHROMATIN REMODELING AND TRANSCRIPTIONAL SILENCING DEFINE THE DYNAMIC INNATE IMMUNE RESPONSE OF TISSUE RESIDENT MACROPHAGES AFTER BURN INJURY

Abstract

Severe burn injury induces prolonged immune dysfunction, but the underlying molecular mechanisms remain poorly defined. We hypothesized that burn injury causes epigenetic and transcriptional training of innate immune cells. Splenic F4/80⁺ macrophages were isolated from mice at 2, 9, and 14 days after 20% total body surface area contact burn. Targeted transcriptomics and chromatin profiling revealed a biphasic response: early transcriptional silencing of inflammatory genes (e.g., Stat3, Traf6, and Nfkb1), followed by increased accessibility and expression of anti-inflammatory loci (Il-10 and Socs3). Metabolic genes showed persistent suppression of mitochondrial and oxidative phosphorylation programs. Canonical pathway analysis indicated early interleukin-10 signaling activation and long-term repression of classical macrophage activation. Chromatin remodeling included nucleosome repositioning events, supporting dynamic, and locus-specific regulation. These findings challenge the notion that burn-induced immune suppression is solely due to systemic inflammation and instead suggest durable, epigenetically programmed alterations in macrophage function.

INTRODUCTION

Severe burn injury triggers a complex immune response that can lead to life-threatening infections and sepsis. According to the American Burn Association, approximately 450,000 patients receive treatment for burns each year, with mortality rates of major burns being around 22.5% (1). The prevailing model proposes an initial exaggerated inflammatory cytokine storm driven by damage-associated molecular patterns that bind Toll-like receptor (TLR) ligands, such as hemagglutinin and circulating double-stranded deoxyribo-nucleic acid (DNA), after rapid glucocorticoid-mediated systemic death of immune cells. Patients who survive this phase often exhibit an immunological endotype of persistent inflammation, immunosuppression, and catabolism syndrome, which clinically presents as chronic immunosuppression, low-grade inflammation, and lean tissue wasting (2–11). Importantly, rather than representing sequential and discrete immune response phenotypes, pro- and anti-inflammatory mediators are simultaneously detected in both of these clinical phases (2,12–16), often described as a mixed antagonist response syndrome (MARS (11,12,17–21)). Many groups have attempted to delineate the underlying mechanisms of MARS to predict and prevent the onset of persistent inflammation, immunosuppression, and catabolism syndrome.

We have previously shown that compared with sham-injured mice, splenic tissue-resident F4/80⁺ macrophages isolated from mice two weeks after a 20% total body surface area (TBSA) burn produced significantly more interleukin (IL)-10 (approximately 1.8-fold increase) in response to TLR2 or TLR4 stimulation (22). This was accompanied by a reduction in tissue necrosis factor (TNF) production, indicating a shift toward an anti-inflammatory, tolerant profile in macrophages late after burn. Early after injury (e.g., day 3), cytokine release was not markedly different from sham, highlighting that innate immune tolerance (or at least a MARS-like innate cell response) develops over time postburn. A series of studies by Ogle et al. (23–28) also defined populations of tissue-resident and circulating inflammatory macrophages exhibiting very long-term immune changes, with the presence of “inflammatory monocytes” and myeloid-derived suppressor cells sustained in the periphery after burn injury.

The emerging concept of innate immune memory encompasses two opposing phenotypes: trained immunity versus immune tolerance (29). Trained immunity is characterized by a heightened inflammatory response upon rechallenge, due to sustained priming of innate cells. In contrast, tolerance is a state of refractoriness, with blunted pro-inflammatory cytokine production and a bias toward anti-inflammatory mediators (like IL-10). Trained immunity and tolerance are both driven by long-term reprogramming of innate immune cells through metabolic and epigenetic mechanisms. While the concepts of innate immune “training” and “tolerance” were initially defined using canonical ligands such as β-glucan polysaccharide, bacillus Calmette-Guérin, or repeated lipopolysaccharide (LPS) exposure (29), more recent studies demonstrate that sterile insults, including trauma, can similarly drive long-lived, damage-associated molecular pattern-mediated innate immune memory (30). In this broader framework, burn injury represents a potent inducer of both training-like and tolerance-like programs. Despite prior epigenomic analysis in sepsis models (31), relatively little is known about how epigenetic reprogramming contributes to inflammation and immune dysfunction after burn injury. This represents an important knowledge gap, as epigenetic remodeling is increasingly recognized as both a biomarker of immune dysfunction and a promising therapeutic target. To address this gap, we selected methyltransferase accessibility protocol for individual templates- flap-enabled next-generation capture (MAPit-FENGC) for this study because it provides single-molecule resolution of both chromatin accessibility and DNA methylation at predefined loci, enabling us to interrogate specific promoters and enhancers previously implicated by our group and others in burn- and trauma-induced immune dysfunction.

METHODS

Mouse cutaneous burn injury

Female C57BL/6 mice weighing 18–22 g (8–12 weeks old) were used for all experiments. Animals were anesthetized by an intraperitoneal injection of 0.024 mL/g of Avertin and had their dorsal and flank hair clipped. A subcutaneous injection of Ethiqa XR (buprenorphine extended-release injectable suspension, 1.3 mg/mL; Fidelis Animal Health, Inc., NJ) was administered before injury for pain control. To generate a full-thickness thermal burn of approximately 20% TBSA, a 65 g copper rod (1.9 cm in diameter) was heated to 100°C and applied to four separate areas, each for 10 seconds, to the animal’s dorsum and flank at a standard pressure that we have demonstrated produces a full-thickness burn (32). The mice were allowed to recover on a heating pad and were resuscitated with an intraperitoneal injection of lactated Ringer’s solution (2.5 mL). Once resuscitated, the mice were placed in individual cages, provided food ad libitum and maintained on Ethiqa XR, and continuously monitored until euthanasia. Sham controls underwent all procedures in parallel, except for the burn injury. Baseline (preinjury) samples were not collected, as sham procedures themselves alter immune status; sham-injured animals were therefore used as the appropriate comparator to specifically isolate burn-induced epigenetic changes. Mice were euthanized 2, 9, or 14 days after sham or burn injury. All animal work was performed under University of Florida Institutional Animal Care and Use Committee-approved protocols.

F4/80⁺ tissue-resident macrophage (trMø) purification

Spleens were removed from the euthanized mice, and following mechanical dissociation and standard ammonium-chloride-potassium red blood cell lysis, single-cell suspensionswere filtered through a 70μm meshand centrifugedat 300 × g for 10 minutes. Cells were resuspended in magnetic-activated cell sorting buffer (phosphate-buffer saline, pH 7.2), containing 0.5% bovine serum albumin and 2 mM ethylene-diaminetetraacetic acid (EDTA), and total cell counts were determined using a hemocytometer with trypan blue exclusion. Fifty million splenocytes were enriched for F4/80⁺ using mouse anti-F4/80 magnetic microbeads, ultrapure (130-110-443, Miltenyi Biotec, Teterow, Germany), selection through LS columns (130-042-401, Miltenyi Biotec) on a QuadroMACS setup (Miltenyi Biotec), according to the manufacturer’s instructions. One spleen routinely yielded approximately 2 x 10⁶ cells with >95% purity, as we have described previously (22,33). Enriched F4/80⁺ cells were used immediately for transcriptomic and epigenetic analysis.

Cytokine response to toll-like receptor stimulation

F4/80⁺ tissue-resident macrophages from each mouse (2 x 10⁶/mL) were stimulated with 10 μg/mL Bacillus subtilis peptidoglycan (PGN) or 1 μg/mL Ultra-Pure Esherichia coli 0111:B4 LPS (InvivoGen, San Diego, CA) in 1 mL of RPMI (Sigma Aldrich, St. Louis, MO), supplemented with 1 mM L-glutamine, 100 units/mL penicillin-streptomycin, and 1% noncommercial autologous mouse serum in 24-well flat-bottom plates. The cultures were incubated for 48 hours. Supernatants were then collected and assayed by multiplex mouse cytokine bead array.

Multiplex cytokine bead array

Samples were processed for analysis on a Bio-plex mouse chemokine 33-plex panel (Bio-Rad #12002231) or Bio-Plex Pro Mouse Cytokine Group 1 7-plex panel (Bio-Rad, #L6000004C6), according to the manufacturer’s protocol. Data was acquired on a Luminex 200 system running xPONENT software and analyzed using a 5-parameter logistic spline-curve fitting method using Milliplex Analyst software. Data are presented as pg/mL normalized to total protein or mg of tissue.

RNA extraction and transcriptomic analysis

Isolation of mRNA was performed using RNeasy Pure mRNA Bead kits according to manufacturer’s instructions. mRNA was quantified using a Nanodrop 2000 spectrophotometer (Waltham, MA). NanoString technology and the nCounter Mouse Immunology Panel (XT-CSO-MIM1-12, NanoString, Seattle, WA) and Mouse Metabolic Panel (XT-CSO-MMP1-12) were used to simultaneously evaluate 561 immune and 768 metabolic-related mRNAs in each sample, respectively (34). Each sample was analyzed in triplicate according to manufacturer’s instructions. nSolver v4.0 and ROSALIND integrated analysis platforms were used to generate appropriate data normalization as well as fold-changes, resulting ratios, and differential expression. Ingenuity Pathway Analysis (IPA: Qiagen) and R statistics were used to identify canonical pathway-specific responses (34).

Methyltransferase accessibility protocol for individual templates combined with flap-enabled next-generation capture epigenetic analysis of burn injury-induced tissue-resident macrophage

MAPit-FENGC is a validated method that allows for multiplexed, simultaneous single-molecular level analysis of methylation and accessibility at target promoter and enhancer regions (35). The full FENGC methodology can be found at Zhou et al. (35). Briefly, one million cells purified from each mouse were washed with ice-cold phosphate-buffer saline, pH 7.2. Cells were centrifuged at 1,000 × g for 5 minutes, then washed in ice-cold cell resuspension buffer (20 mM HEPES pH 7.5, 70 mM NaCl, 0.25 mM EDTA pH 8.0, 0.5 mM EGTA pH 8.0, 0.5% (w/v) glycerol, freshly supplemented with 10 mM DTT and 0.25 mM PMSF). Cells were centrifuged at 1,000 × g for 5 minutes before being resuspended with 92 μL of resuspension buffer with 0.5 % (w/v) digitonin. Cells were then stained with trypan blue to ascertain 100 % permeabilization. The cells were then treated with 100 U M.CviPI GpC methyltransferase (100 U/million cells; New England Biolabs, M0227B-HI) with fresh 160 μM S-adenosyl-L-methionine for 15 minutes at 37°C. The reaction was terminated by adding an equal volume of 10 mM EDTA, 100 mM NaCl, and 1% (w/v) sodium dodecyl sulfate followed by a quick vortex at medium speed. The nuclei were treated with 100 μg/mL RNase A for 30 minutes at 37°C followed with 100 μg/mL proteinase K treatment overnight at 50°C. Extraction of genomic DNA was conducted using phenol-chloroform-isoamyl alcohol (25:24:1, v/v) phase separation, which was followed by ethanol precipitation, then resuspension in molecular-grade H₂O.

A total of 588 primers for a total of 196 selected promoter and enhancer regions of immune and trauma-related genes (“BurnMAP”; Supplemental Materials 2, https://links.lww.com/SHK/C653) were used to concurrently profile chromatin accessibility and methylation. To maximize translational relevance, our BurnMAP promoter/enhancer panel was informed in part by recent human scATAC-seq studies in trauma and burn patients that identified injury-induced repression of inflammatory and interferon programs (36). Regions with less than three amplified samples in sham, day 2, or 9, or <100 reads (locus copies or molecules of cellular chromatin) per condition were not considered. Transcription start site (TSS) location was determined by RefSeq annotations for each gene. Regions were trimmed to 100 bp upstream of the TSS site. The average accessibility of this region in each sequence was then calculated for each read. Regions with low GCH resolution around the TSS site that also did not pass Levene’s test for equality of variances were omitted. Combined with QC filtering, this resulted in 98 final gene loci. A t test was conducted for P values as well as log₂ fold-changes of accessibility. Target regions with log₂ fold-changes that had significant P values (<0.05) were graphed using PRISM (GraphPad Software, Inc. v10.1). A two-way ANOVA with Geisser-Greenhouse correction and post hoc Tukey’s honestly significant difference test was conducted for the clustering analysis across the sham, D2, and D9 mice.

Target gene promoters and enhancers of interest were located using the TSS region and ENCODE candidate of cis-regulatory element annotations, respectively, from the mm10 genome assembly. Flap oligos 1 and 2 as well as nested oligos 3 (4 nmole each) were ordered as 0.3 mL each as salt-free in 96-well format (Eurofins Genomics, KY). Purified amplicons were submitted to the University of Florida Interdisciplinary Center for Biotechnology Research (UF-ICBR, RRID: SCR_019152) for barcoding and SMRT bell library construction. Libraries were sequenced on a PacBio SEQUEL IIe (Menlo Park, CA) instrument by UF-ICBR. The library pool was loaded at 120 pM, using diffusion loading and 20- to 30-hour movies with HiFi generation and demultiplexing. Sequencing Kit 2.0 (PacBio, 101-389-001) and Instrument Chemistry Bundle Version 11.0 were used. All other steps were performed using the recommended protocol by the PacBio sequencing calculator. For epigenetic analysis, high-fidelity circular consensus sequencing (HiFi CCS) was generated using default parameters. CCS reads were filtered for ≥5 single polymerase read passes and aligned to the reference sequences using the Python reAminator pipeline (37). Cutoffs of ≥95% conversion rate and alignment of ≥95% length of each reference sequence were applied. In the MAPit-FENGC assay, endogenous CpG methylation (HCG sites, where H=A, C, or T) is measured in parallel with GpC methylation (GCH sites) introduced by the M.CviPI methyltransferase. To unambiguously distinguish endogenous CG methylation and M.CviPI-probed GC methylation, GCG sites were removed, resulting in the calculation and plotting of HCG and GCH methylation. HCG methylation therefore reflects native 5-methylcytosine status at CpG dinucleotides, providing information about endogenous DNA methylation. In contrast, GCH methylation serves as a proxy for chromatin accessibility, as M.CviPI can only methylate protein-free and therefore accessible GpC sites. Together, these complementary readouts allow simultaneous resolution of endogenous methylation patterns (HCG) and nucleosome positioning/accessibility (GCH) at single-molecule resolution. Averaged heatmaps were constructed based on averaged percentages HCG and GCH per nucleotide post-methylscaper analysis (38).

RESULTS

Altered cytokine secretion by macrophages following toll-like receptor stimulation after burn injury

Wild-type C57BL/6 mice underwent a 20% TBSA full-thickness cutaneous contact burn or sham injury (n = 6 per group). All animals survived to sacrifice at the pre-planned intervals. To assess the functional responsiveness of macrophages in response to burn, trMø were harvested from sham- or burn-injured mice at acute (day 2, D2) and chronic (day 14, D14) postinjury. The purified cells were subsequently stimulated ex vivo with the TLR2 ligand PGN or the TLR4 ligand LPS. Cytokine secretion was quantified by multiplex bead array (Fig. 1). At D2, postburn injury macrophages displayed a hyperinflammatory phenotype characterized by significantly increased secretion of IL-6, monocyte chemoattractant protein (MCP)-1, and TNFα compared with sham controls, following stimulation with either PGN or LPS.

Fig. 1. Burn injury induces biphasic alterations in splenic tissue-resident macrophage cytokine secretion following TLR stimulation.

Splenic F4/80+ tissue-resident macrophages were isolated 2 (top) or 14 (bottom) days after 20% total body surface area burn injury (solid bars) or sham (open bars) injury, and stimulated ex vivo with the TLR2 ligand peptidoglycan (PGN) or the TLR4 ligand lipopolysaccharide (LPS). Cytokine secretion was measured by a multiplex bead array. Data are presented as mean ± SEM (n = 5/group); *P ≤ 0.05, **P ≤ 0.005 versus sham by Student’s t test. Results are representative of two independent experiments. SEM, standard error of the mean; TLR, toll-like receptor.

At D14, this pattern was reversed. Macrophages from burn-injured mice exhibited attenuated cytokine production relative to sham controls. In response to PGN, the increase in MCP-1 relative to sham observed at D2 was lost at D14, and secretion of TNFα was significantly decreased. IL-10 secretion was significantly increased relative to sham. Similarly, LPS-stimulated trMø from burn-injured mice at D14 relative to D2 showed reduced secretion of MCP-1 and TNFα relative to sham. In contrast, at D14 versus D2, IL-6 secretion remained comparable to sham, whereas IL-10 levels increased significantly. Taken together, burn stimulus elicits initial secretion of pro-inflammatory IL-6 and TNFα cytokines at D2 that subsides and is superseded by secretion of anti-inflammatory IL-10 at D14.

Transcriptomic reprogramming of splenic tissue-resident macrophage following burn injury

Wild-type C57BL/6 mice underwent a 20% TBSA full-thickness cutaneous contact burn or sham injury (n = 6 per group). F4/80⁺-enriched cells were purified from spleens collected at acute (D2), subchronic (D9), and chronic (D14) time points after burn injury. mRNA was subsequently purified and used to perform immune and metabolic gene transcriptomic analysis (Fig. 2), followed by corresponding IPA (Fig. 3) at each time point versus sham injury. In turn, Figure 2A illustrates the immune and metabolic gene expression changes in the burn groups compared to sham-injured mice at D2. Enriched F4/80⁺ cells at all time points after burn injury exhibit significant (P < 0.05) up- and downregulation of a broad range of immune and metabolic genes (DEGs) compared with sham mice (the top 50 significantly altered DEGs are shown in Supplemental Table 1, https://links.lww.com/SHK/C653).

Fig. 2. Splenic tissue-resident macrophages (trMø) harvested after burn injury exhibit specific immune and metabolic gene changes compared with trMø from sham-injured mice.

mRNA isolated from splenic trMø 2 (A), 9 (B), or 14 (C) days after 20% total body surface area burn injury were tested for expression of immune and metabolic genes by nanoString analysis compared with trMø harvested from sham-injured mice. Data are presented as the log₂-transformed differential fold change in immune or metabolic gene expression as shown in each label, with associated P value significance (using Welch’s t test). Tgfb1, Socs3, and IL-10 are highlighted as referred to in the Results section. IL, interleukin.

Fig. 3. Splenic tissue-resident macrophages (trMø) harvested after burn injury exhibit specific canonical pathway changes compared with trMø from sham-injured mice.

mRNA isolated from splenic trMø 2 (A), 9 (B), or 14 (C) days after 20% total body surface area burn injury were tested for expression of immune and metabolic genes by nanoString analysis compared to trMø harvested from sham-injured mice. From these data, canonical signaling pathways with associated Z scores and significance (using Welch’s t test) were derived by Ingenuity Pathway Analysis (IPA); all pathways shown are significantly altered (P < 0.05).

We then analyzed the complete set of all significant (P < 0.05) metabolic and immune DEGs by IPA to interrogate canonical signaling pathways at D2 after burn injury compared with sham-injured mice (Fig. 3A). Figures 2B and 3B present the same pathway analysis for D9; D14 data are represented in Figures 2C and 3C. As can be observed, multiple canonical signaling pathways were significantly impacted. For example, significant increases in IL-10 Signaling and MSP-RON Macrophage Signaling pathway Z scores were seen at either burn D2 and/or D14 versus sham-injured mice. Notable decreases in expression were seen in S100 Family Signaling, IL-6 Signaling, and Macrophage Classical Activation Signaling pathways at D2. D14 postburn expression exhibited significant decreases in Macrophage Classical Activation Signaling Interleukin-1 Family Signaling, and INOS Signaling, amongst others. At D9 postburn, the pathways significantly affected were less numerous and centered mainly on upregulation of Cell Cycle and Cell Cycle Regulation-related pathways. Direct comparisons between D9 and D14 with D2 are shown in Figure 4. Taken together, these transcriptomic data suggest significant differences in immune and metabolic reprogramming in F4/80⁺-enriched splenocytes (trMø) compared with sham-injured mice and between different timepoints after burn injury.

Fig. 4. Splenic tissue-resident macrophages (trMø) harvested at different time points after burn injury exhibit differential specific immune and metabolic gene, and canonical pathway changes.

mRNA isolated from splenic trMø harvested 2, 9, or 14 days after 20% total body surface area burn injury, was tested for expression of immune and metabolic genes by nanoString analysis. A–B, Data are presented as the log₂-transformed differential fold change in (A) immune or (B) metabolic gene expression as shown in each label, with associated P value significance (using Welch’s t test); C, From these data, canonical immune and metabolism pathways with associated Z scores and significance (using Welch’s t test) were derived by Ingenuity Pathway Analysis (IPA).

Epigenetic reprogramming of immune and metabolic genes in tissue-resident macrophage following burn injury

We therefore hypothesized that epigenetic changes would occur in key immune and metabolic genes in trMø at D2 and D9 postburn compared with sham-injured mice. In parallel to the transcriptomic studies, we also purified F4/80⁺-enriched splenocyte nuclei from the same wild-type C57BL/6 mice that underwent the sham or burn injury at D2 and D9 after injury. Following cell permeabilization with digitonin, we applied MAPit single-molecule footprinting followed by FENGC target enrichment to concurrently profile chromatin accessibility and methylation over 180 selected promoter TSSs and enhancer regions of the BurnMAP library. Epigenetic profiles were compared between burn and sham conditions at D2 (acute) and D9 (subchronic) postinjury. At D2 postburn, significant increases in promoter/enhancer accessibility for genes such as Cxcl15, Hsfy2, Ccl7, and Cxcr2 (Fig. 5A) were observed, suggesting an early pro-inflammatory transcriptional poising, particularly for chemokines and stress-associated genes. In contrast, a subset of genes, including Hsp90ab1, Hspa1b, and Nfkbia showed reduced accessibility, potentially reflecting early repression of heat shock proteins and NF-κB regulatory feedback elements (as previously described in burn patient studies (39)). By D9 postburn (Fig. 5B), promoters of Mt2, Cd36, and Pik3r6 displayed significantly increased accessibility, consistent with a transition toward stress adaptation and metabolic or anti-inflammatory signaling. Conversely, regulatory regions associated with Nfkb1, Tgfb1, Traf6, Stat3, and Irf5 showed significant reductions in accessibility, indicating persistent epigenetic silencing of key inflammatory and interferon response pathways. Comparison between D9 and D2 trMø (Fig. 5C) revealed a significant time-dependent decline in accessibility at promoters of Nfkb1, Traf6, Tgfb1, Ccr6, and Fyb, pointing to progressive epigenetic repression of inflammatory signaling and chemotaxis-related genes. Only Mt2 and Pik3r6 exhibited increased accessibility over time, implicating their potential role in sustained metabolic or regulatory programs in chronically reprogrammed macrophages.

Fig. 5. Chromatin accessibility at selected promoter and enhancer regions of immune-related genes in splenic tissue-resident macrophages (trMø) following burn injury, which is associated with significant transcriptomic changes.

A–C, Flap-Enabled Next-Generation Capture (FENGC) of target sequences from F4/80⁺-enriched splenocytes at Day 2 (D2) and Day 9 (D9) was performed following 20% TBSA cutaneous burn injury in C57BL/6 mice. Gene regions were selected based on TSS proximity and mm10 genome annotations. Data shown are log₂ fold-changes in chromatin accessibility. A, D2 burn mice versus sham-injured control. B, D9 burn mice versus sham. C, D9 burn mice versus D2 burn mice. D–E, Matching transcriptomic data for selected genes, which significant log₂ fold-changes in chromatin accessibility. mRNA isolated from splenic trMø harvested from sham-injured mice, 2, 9, or 14 days after 20% TBSA burn injury, was tested for expression of immune and metabolic genes by nanoString analysis. Data are presented as the log₂-transformed differential fold change versus sham, with associated P value significance (using Welch’s t test; *P < 0.05, **P < 0.01, ***P < 0.005). TBSA, total body surface area; TSS, transcription start site.

Significant log₂ fold-changes in chromatin accessibility were examined in genes whose transcription was affected by burn injury (Fig. 5D). Genes with reduced chromatin accessibility at D2 and D9, such as Nfkbia, Nfkb1, Traf6, Stat3, and Tgfb1, exhibited sustained downregulation of expression by D14 (e.g., Tgfb1; Fig. 5D and highlighted in Fig. 2), suggesting that early epigenetic silencing led to persistent transcriptional repression. We also observed transcriptional upregulation of Socs3 and Il-10 following burn injury (Fig. 5E and highlighted in Fig. 2), which may reflect a shift toward an immunosuppressive macrophage phenotype, where Socs3 inhibits pro-inflammatory signaling pathways and Il-10 promotes sustained anti-inflammatory responses. We and others have shown that alteration in these genes to be biologically associated with the dysfunctional immune response after burn injury (22,26,40,41).

Epigenetic transcriptional silencing of Stat3 following burn injury

To determine whether burn injury alters chromatin structure at key immune regulatory loci, we performed high-resolution accessibility profiling of various promoters from splenic F4/80⁺-enriched splenocytes using GCH methylation-based MAPit analysis. GCH methylation levels were used as a proxy for chromatin openness, with increased methylation indicating accessibility in nucleosome-free regions. Figure 6A shows that at baseline, the Stat3 promoter (transcription shown in Fig. 5D) exhibited moderate to high accessibility across an ~500 bp region surrounding the TSS, with background levels of endogenous HCG methylation, consistent with an open, epigenetically permissive state. In contrast, F4/80⁺-enriched splenocytes from mice at D2 postburn showed a clear reduction in accessibility upstream and downstream of the TSS. By D9 postburn, accessibility was further diminished, with the lowest GCH methylation levels observed across the entire promoter region. These data indicate that Stat3 promoter accessibility is progressively reduced following burn injury, with a temporal decline in chromatin openness from day 2 to 9. Given Stat3’s central role in cytokine signaling, macrophage polarization, and immune homeostasis, these findings suggest that burn-induced epigenetic silencing at the Stat3 locus may contribute to persistent transcriptional suppression in trMø during the subacute and chronic postinjury phases.

Fig. 6. Average chromatin accessibility of the Stat3 and Il-10 promoter regions, as well as single molecular clustering analysis of Il-10.

Accessibility line graphs based on average GCH methylation are provided (n = 5). Groups based on days after burn are colored per the figure key. Available HCG methylation data of all 15 mice are plotted by a dashed black line. Genomic coordinates and positions of GCH sites provided by vertical lines at the top of each graph. TSS position and transcription orientation are provided by the vertical dashed red line and arrow. Averaged data is presented for the A, Stat3 promoter and B, Il-10 promoter. C–D, Single-molecule clustering of the Il-10 promoter based on GCH methylation (accessibility) patterns. Black shading indicates nucleosome-protected (inaccessible) DNA, whereas yellow shading indicates nucleosome-free (accessible) DNA. Clusters 7–4 represent transcriptionally repressed states with nucleosome coverage over the TSS, while clusters 3–1 show progressive downstream nucleosome sliding that exposes the TSS and transcription factor binding motifs, consistent with a transcription-ready state. Burn mice at Day 9 display an increased proportion of molecules in the open, transcription-ready clusters (1–2) and a reduced proportion in the fully repressed cluster (7). Error bars represent standard deviation. TSS, transcription start site.

Burn-induced chromatin reorganization at the Il-10 locus reflects a shift toward macrophage tolerance

Accessibility levels across the Il-10 promoter region remain relatively consistent among all different conditions (Fig. 6B). Using a single-molecule clustering method, we pooled all Il-10 reads across all samples to visualize dynamic nucleosome positions across the Il-10 promoter (Fig. 6, C and D). Molecules in clusters 7–4 show nucleosome occupancy of the Il-10 TSS, consistent with promoter repression (Fig. 6C). Clusters 4–1, meanwhile, show incremental, downstream nucleosome sliding, first localizing the TSS to the edge of a nucleosome (cluster 3), and sequentially exposing two actively bound transcription factor binding sites (clusters 2 and 1). A significantly lower proportion of reads (P < 0.05) in the chronic condition populated the transcriptionally inactive cluster 7 (Supplemental Fig. 1, https://links.lww.com/SHK/C653). Furthermore, a trend towards increased average percentages of clusters 1 and 2 in the burned mice at D9 suggests that a higher proportion of cells could be in a transcription-ready epigenetic conformation. We posit that these shifts in nucleosome arrangements may indicate precursor epigenetic reprogramming that permits the robust Il-10 transcriptional activity observed in D14 mice (Fig. 6D). Finally, the D2 condition presents an even distribution of nucleosome states, indicating heterogeneity of nucleosome remodeling events in earlier time points. The induction of downstream nucleosome sliding and transcription-ready chromatin at the Il-10 promoter in F4/80⁺-enriched splenocytes by D9 postburn suggests an epigenetically driven shift toward an anti-inflammatory state, contributing to macrophage-mediated immune suppression and chronic dysfunction after severe injury.

Nucleosome dynamics define epigenetic silencing at Il2ra, Irak4, Tgfb1, and the Socs3 enhancer

In addition to Stat3 and Il-10, averaged accessibility plots for Il2ra, Irak4, Tgfb1, and a Socs3 cis-regulatory enhancer are presented (Fig. 7A–D). Il2ra presents decreased average accessibility for the D9 condition around the TSS site. However, accessibility remains consistent in an upstream region of the promoter containing elevated levels of HCG methylation. Irak4, which plays a role in TLR4 signaling, presents decreased promoter accessibility in both D2 and D9 conditions. Tgfb1 has decreased average accessibility for the D9 condition around the TSS. Finally, the Socs3 enhancer region maintains relatively consistent accessibility upstream of the enhancer site but decreased D9 accessibility within the enhancer region.

Fig. 7. Average chromatin accessibility of Il2ra, Irak4, and Tgfb1 promoter regions, Socs3 enhancer regions, as well as single-molecule clustering of Tgfb1 and the Socs3 enhancer.

Accessibility line graphs based on average GCH methylation are provided (n = 5). Groups based on days after burn are colored per the figure key. Available HCG methylation data of all 15 mice are plotted by a dashed black line. Genomic coordinates and positions of GCH sites provided by vertical lines at the top of each graph. TSS position and transcription orientation are provided by the vertical dashed red line and arrow. Averaged data is presented for the A, Il2ra promoter, B, Irak4 promoter, C, Tgfb1 promoter, and D, Socs3 enhancer. Single-molecule clustering for E, Tgfb1 and F, the Socs3 enhancer is provided with clustered group labels on the right. A 147 bp nucleosome reference is provided at the bottom of the graph, along with the relative position from the TSS. Black shading indicates nucleosome-protected (inaccessible) DNA, whereas yellow shading indicates nucleosome-free (accessible) DNA. TSS, transcription start site.

Clustering of Tgfb1 was of note as clusters 1–4 depict a nucleosome sliding event that regulates two transcription binding sites (Fig. 7E). Stepwise loss of Tgfb1 promoter openness and transcription factor binding by D9 postburn is consistent with the observed decrease in Tgfb1 transcript at D14 (Fig. 7D), given that changes in promoter chromatin architecture precede changes in accumulated transcript. The proportion of reads in a given cluster by condition is provided for Tgfb1 in Supplemental Figure 2, https://links.lww.com/SHK/C653. Of note, the most accessible cluster 1 is significantly downregulated (P < 0.05) in the D9 condition compared with sham, while a less accessible cluster 6 is significantly upregulated (P < 0.05) in the D9 condition compared with D2.

Our single-molecular analysis predicts transcriptional upregulation of Il-10 and transcriptional downregulation of Tgfb1 in D9 mice. Although the transcript changes of Il-10 and Tgfb1 are negligible at D9 (Fig. 5, D and E), we observe significant changes by D14. Therefore, our epigenetic analysis suggests early epigenetic changes that can predict transcriptomic outcomes at later time points.

Our single-molecule clustering approach was also used to visualize nucleosome conformation changes in a cis-regulatory enhancer region. Average HCG levels indicate high CpG methylation at the 5’ end of the selected region (Fig. 7D). Indeed, single-molecular analysis reveals a randomly positioned array of nucleosomes with accessible linkers upstream of the enhancer (Fig. 7F, clusters 3–7). Meanwhile, the cis-regulatory enhancer for Socs3 (EM10E0527575) presents a nucleosome sliding event in the 3’ direction between clusters 7–2 followed by nucleosome eviction in cluster 1. This sliding event regulates a high-occupancy transcription factor binding site, which could play a role in Socs3 transcription (Fig. 5D). However, there were no significant changes in the proportion of reads from sham, D2, and D9 as presented in Supplemental Figure 3, https://links.lww.com/SHK/C653.

DISCUSSION

This study provides the first high-resolution, time-resolved integration of transcriptomic and epigenetic profiling of tissue-resident F4/80⁺ macrophages after burn injury, revealing a novel paradigm of innate immune reprogramming. Using MAPit-FENGC single-molecule accessibility mapping along-side nanoString-based transcriptional analysis, we uncover temporally coordinated chromatin remodeling events that likely drive gene expression patterns associated with immune memory. We identify epigenetic silencing of Stat3, Tgfb1, and Nfkb1, as well as transcriptional and structural poising of Il-10 and Socs3, supporting a shift toward an anti-inflammatory, tolerogenic macrophage phenotype. These features emerge progressively postburn, implicating early chromatin remodeling as a key determinant of immune dysfunction after burn injury. Our functional data support the concept that burn injury induces a biphasic innate immune phenotype in macrophages. At day 2, burn macrophages exhibited a hyperresponsive state, secreting elevated levels of inflammatory mediators (IL-6, MCP-1, and TNFα) as well as the immunoregulatory cytokine IL-10 following TLR2 and TLR4 stimulation. By day 14, secretion of MCP-1 and TNFα was markedly attenuated, consistent with a tolerant phenotype; however, IL-10 production was significantly elevated compared to sham, highlighting a sustained immunoregulatory axis. These cytokine dynamics mirror our MAPit-FENGC results, where early chromatin remodeling at Il-10 preceded transcriptional divergence at later timepoints.

In more detail, our transcriptomic analysis revealed distinct temporal “waves” of gene expression following burn injury, including early adhesive signatures (e.g., VCAM1), pro-inflammatory mediators such as S100A8/9, and later sustained inflammatory cytokines including IFNA1, IL-25, and C-C motif chemokine ligand 2. However, we focused our mechanistic epigenetic analysis on IL-10 because of its established role in burn-induced immune suppression in both mouse and human studies. Notably, IL-10 was significantly upregulated in our transcriptomic dataset and emerged as one of only three immune canonical pathways enriched at day 14 postburn. MAPit-FENGC analysis revealed early changes in accessibility and DNA methylation at the Il-10 locus on day 9, preceding the transcriptional increase at day 14, illustrating how locus-specific epigenetic remodeling can predict later gene expression. This underscores a central value of epigenetic interrogation: remodeling reflects shifts in regulatory potential rather than immediate expression amplitude. While our study centered on IL-10 as a sentinel immunoregulatory locus, the same framework can be extended to other genes highlighted in our dataset (e.g., C-C motif chemokine ligand 2 and IL-6) that are strongly linked to burn outcomes, providing a path for future investigation.

Therefore, we have revealed a programmed and locus-specific epigenetic architecture that may shape macrophage immune and metabolic function long after the acute phase. This work expands current models of innate immune memory by demonstrating that trauma can also drive functional epigenetic imprints in macrophages. Our findings challenge the classical view of burn-induced immune suppression as solely a byproduct of systemic inflammation or myeloid turnover. Instead, they reveal a composite form of injury-induced innate immune memory, in which multiple epigenetic trajectories coexist. In our dataset, the elevation of IL-25 and related mediators reflects a training-like hyperinflammatory axis, IL-10 upregulation represents a regulatory/training-like feature, and locus-specific repression of Stat3, Tgfb1, and Nfkb1 exemplifies tolerance-associated remodeling (Fig. 8). Taken together, these data suggest that burn injury does not induce exclusively pathogenic inflammation or exclusively tolerance, but rather establishes a hybrid state in which hyperinflammatory (IL-25), immunoregulatory (IL-10), and tolerogenic (Stat3, Tgfb1, and Nfkb1) mechanisms are all epigenetically engaged. This provides a mechanistic framework for the paradoxical clinical phenotype of burn patients, who often experience exaggerated systemic inflammation in the acute phase followed by profound immunosuppression and infection susceptibility during recovery.

Fig. 8. Burn injury induces a hybrid innate immune memory phenotype in macrophages.

Schematic representation of tissue-resident macrophage (trMø) responses after burn injury. The balance of epigenetic and functional programs is depicted as a balance between hyperinflammatory training-like mechanisms (red; IL-25, S100A8/9, CCL2) and tolerance-associated repression (blue; Stat3, Tgfb1, Nfkb1). In addition, a regulatory axis (green) marked by IL-10, which demonstrates both early chromatin remodeling and elevated secretion at day 14 compared with sham, highlights the engagement of an immunoregulatory program. CCL2, C-C motif chemokine ligand 2; IL, interleukin.

Epigenetic immune memory (tolerance or training) means that the immune system does not simply reset to baseline after the acute inflammatory phase of burn. Burn and trauma patients often enter a state of chronic inflammation and immune suppression in which they are less capable of preventing infections or healing wounds. Epigenetic silencing of key inflammatory genes (tolerance) provides a mechanistic basis for this devastating clinical syndrome. On the other hand, there is evidence, including our work (22,23,26,33,42), that elements of trained immunity can also emerge after injury. Thus, both the “tolerant” and “trained” outcomes of innate memory are likely relevant to burns and trauma, with an initial injury simultaneously dampening certain immune functions while amplifying others via epigenetic changes. This may underlie both the paradox of burn and trauma survivors who suffer from both recurrent infections and persistent inflammatory complications.

Several limitations of this study should be acknowledged. First, our enrichment strategy for F4/80+ cells using magnetic bead selection yields a mixed macrophage population that may include both bona fide tissue-resident red pulp macrophages and monocyte-derived macrophages that have infiltrated and differentiated within the spleen postinjury. Without high-resolution cell sorting or single-cell analysis, it remains challenging to fully disentangle the relative contribution of these subsets to the observed transcriptional and chromatin signatures. Our analyses were performed on a specific macrophage population, which limits the ability to capture heterogeneity in chromatin state or gene expression that may exist within subpopulations. Because our transcriptomic analysis revealed that the most dynamic shifts in immune and metabolic gene expression occurred between D2 and D9, we prioritized these two timepoints for MAPit-FENGC, reasoning that chromatin remodeling would be most evident during this transition. We acknowledge, however, that D14 represents a biologically important later phase, particularly in the case of Tgfb1, which shows transcript-level reduction only at this time point. Extending locus-specific accessibility and methylation profiling to D14 and beyond will therefore be an important direction for future studies to capture the full temporal arc of epigenetic remodeling after burn injury.

While MAPit-FENGC provides high-resolution single-molecule insight into chromatin dynamics, the targeted nature of our approach restricts epigenetic profiling to a curated subset of promoters and enhancers. Thus, global chromatin remodeling events, distal regulatory elements, and changes in 3D genome architecture remain unexamined. Future studies incorporating histone-based approaches such as ChIP-seq or CUT&RUN for key marks (e.g., H3K27ac and H3K27me3) will be critical to complement our MAPit-FENGC analysis, enabling integration of DNA methylation, chromatin accessibility, and histone modification landscapes to more fully define the epigenetic circuitry driving burn-induced immune dysfunction. Although our BurnMAP panel was designed to target promoters and enhancers, we acknowledge that intronic methylation can also play an important role in transcriptional regulation. Future iterations of BurnMAP will therefore be expanded to include intronic elements to more fully capture the spectrum of epigenetic control after burn injury.

The sample size per condition was modest, which, while sufficient for identifying robust changes, may limit the detection of more subtle effects and increases the risk of cohort-specific variation. Although we controlled for technical batch effects and performed all transcriptomic and epigenetic profiling in parallel, we cannot exclude biological variation related to animal handling, injury timing, or circadian influences. We focused on female C57BL/6 mice to reduce biological variability and because female mice are widely used in burn models, including the majority of our prior work, due to their heightened cytokine responses following injury (43). We acknowledge, however, that sex and sex hormones exert profound influences on immune function and outcomes after trauma and sepsis (44).

Finally, we infer functional outcomes based on gene expression and epigenetic configurations. However, definitive linkage to macrophage phenotype or systemic host defense was not directly assessed. Additional protein-level validation (e.g., by flow cytometry or immunoblotting) and functional assays of macrophage responsiveness or pathogen clearance will be needed in future studies to confirm these mechanistic interpretations. In addition, orthogonal epigenetic approaches such as ATAC-seq or ChIP-qPCR will be important for validating our MAPit-FENGC findings, particularly given that recent human scATAC-seq data in trauma and burn cohorts independently support injury-induced epigenetic repression of immune programs consistent with our observations (36). Our interpretations are also supported by prior data from our group and others showing that splenic macrophages from burn-injured mice exhibit significantly altered cytokine (e.g., elevated TNFα and IL-10) production in response to TLR stimulation (22,23,45), underscoring the functional relevance of the epigenetic signatures we describe. Nonetheless, integrating MAPit-FENGC with direct functional assays, such as cytokine secretion, phagocytosis, or pathogen clearance, represents a critical next step to confirm mechanistic causality.

Despite these limitations, our study provides novel evidence that burn injury drives locus-specific epigenetic remodeling in macrophages, laying important groundwork for future larger-scale and functional studies. These new insights have broad implications not only for burn/trauma care but also for understanding immune paralysis in other trauma contexts, and may offer new targets for immunomodulatory therapies such as interventions that recalibrate epigenetic programming.

Supplementary Material

Supplemental digital content is available for this article. Direct URL citation appears in the printed text and is provided in the HTML and PDF versions of this article on the journal’s Web site (www.shockjournal.com).

ABBREVIATIONS

DNA

deoxyribonucleic acid

EDTA

ethylenediaminetetraacetic acid

FENGC

flap-enabled next-generation capture

IL-6

interleukin-6

IL-10

interleukin-10

IL-25

interleukin-25

IPA

Ingenuity Pathway Analysis

LPS

lipopolysaccharide

MAPit-FENGC

methyltransferase accessibility protocol for individual templates–flap-enabled next-generation capture

MARS

mixed antagonist response syndrome

MCP-1

monocyte chemoattractant protein-1

PGN

peptidoglycan

TBSA

total body surface area

TLR

Toll-like receptor

TLR2

Toll-like receptor 2

TLR4

Toll-like receptor 4

TSS

transcription start site