Single-nuclei RNA Profiling Reveals Disruption of Adipokine and Inflammatory Signaling in Adipose Tissue of Burn Patients

Abstract

Objective:

We conducted a large-scale investigation of the systemic and adipose tissue-specific alterations in a clinical population of burn patients to identify factors that may influence hypermetabolism.

Background:

Previous research has identified chronic disturbances in adipose tissue inflammation, lipolysis, and browning, which may drive the perpetuation of hypermetabolism following the severe adrenergic stress of a burn injury. Give that adipose tissue is thought to be a central node in the regulation of systemic metabolism, we believe that by systematically delineating the pathological role of adipose tissue post-burn, this will lead to the identification of novel interventions to mitigate morbidity and mortality from severe burns.

Methods:

This was a single-institution cohort study, which obtained plasma and subcutaneous adipose tissue samples from severely burn adult patients over various timepoints during acute hospitalization. Whole-body clinical, metabolic, and inflammatory mediators were assessed in plasma, while genetic analyses via RT-qPCR and single-nuclei RNA sequencing (snRNA-seq) were conducted in adipose tissue.

Results:

Systemic inflammation and adrenergic stress increases IL-6 signaling, lipolysis, browning and adipokine dysfunction in the adipose tissue of adult burn patients, which may further propagate the long-term hypermetabolic response. Moreover, using snRNA-seq, we provide the first comprehensive characterization of alterations in the adipose tissue microenvironment occurring at acute and chronic stages post-burn.

Conclusion:

We provide novel insight towards the effect of burns on adipokine release, inflammatory signaling pathways, and adipose heterogeneity over the trajectory of acute and chronic stages.

Mini-Abstract:

We conducted a large-scale investigation of the systemic and adipose tissue-specific alterations in adult patients with severe burns to identify factors that may influence hypermetabolism. Through clinical and metabolic measurements and single-nuclei RNA sequencing, we provide novel insight towards the effect of burns on adipokine release, inflammatory signaling pathways, and adipose heterogeneity over the trajectory of acute and chronic stages.

Introduction

Critical illnesses invoke profound adrenergic stress responses resulting in hypermetabolic and hypercatabolic processes, including elevated energy expenditure, fat and lean muscle breakdown and systemic inflammation. While these symptoms usually plateau and decline within hours to days after the initial insult, the hypermetabolic response of a severe burn covering greater than 20% of the total body surface area may persist for up to 3 years, well after the wound has healed, indicating that alternative factors other than the healing process are at play(1). Thus, severe burns present a unique clinical challenge in determining the cause of why patients remain chronically hypermetabolic.

Several recent studies have suggested that adipose tissue may be a central regulator of burn-induced hypermetabolism(2, 3). Focused research efforts on adipose tissue over the past few decades have revealed that it has a strong influence on insulin and glucose tolerance, extensive cross-talk with distant organs mediated by vascular and nerve networks as well as secreted adipokines, and the capability to modulate local and systemic immune cell signaling(4, 5). In addition, white adipose tissues can undergo the process of “browning”, in which white adipocytes transition into energy expending beige adipocytes. Together, these characteristics implicate adipose tissue as a mediator of whole-body homeostasis. Based on these attributes, and given the urgency to discover novel therapeutic interventions to mitigate burn-induced morbidity and mortality, we have focused our efforts on targeting pathological adipose tissue. However, the vast majority of our understanding of adipose tissue metabolism post-burn has derived from murine models(2, 6–8). Therefore, it is crucial that we delineate the role of adipose tissue in a clinical burn patient population in order to effectively mitigate long-term hypermetabolism.

While previous studies have characterized the prolonged consequences of hypermetabolism, primarily in pediatric patients(1) and in adults with less severe burns(9), no one has yet determined the post-burn adipose-specific alterations relative to systemic dysfunction within a large clinical population and on a long-term scale. In this cohort study, we obtained plasma and adipose tissue samples from adult burn patients to confirm whether systemic dysfunction results in known consequences, such as adipose tissue lipolysis and browning, as well as provide novel insight towards the effect of burns on adipokine release, inflammatory signaling pathways, and adipose heterogeneity.

Methods

Study Population

This was a single institutional cohort study, which enrolled burn patients admitted to the Ross Tilley Burn Centre at Sunnybrook Health Sciences Centre (Toronto, Canada) between 2011—2020 who required surgery. Non-burn healthy patients undergoing elective surgery (n=16) were enrolled as controls. All patients consented to blood and tissue collection preoperatively. The study protocol was approved by the Sunnybrook Research Ethics Board (Approval #194–2010). Burn patients were selected based on the following inclusion criteria: adults (18— 64 years); severe burn size (≥20% total body surface area (TBSA)); non-futile i.e. survival ≥72-hours post-admission; ≥1 serum and subcutaneous white adipose tissue (sWAT) sample collected during acute hospitalization. Overall, we enrolled 65 severe burn patients, from which 86 sWAT samples and 214 plasma samples were collected over various timepoints post-injury. Samples were stored at −80°C until processed for analyses.

Demographics, Clinical Outcomes, and Functional Biomarkers

Demographics were obtained with patient consent. Clinical outcomes were prospectively recorded by the burn care team during acute hospitalization (Supplementary Table 1). Clinical outcomes included full TBSA, TBSA with 3rd-degree burns (TBSA-3), the presence or absence of inhalation injury, etiology, complications (acute kidney injury (AKI), urinary tract infections (UTI), pneumonia, wound infection, and sepsis), length of stay (LOS), and mortality. Functional biomarkers of systemic and organ dysfunction were measured up to 8-weeks post-admission, including: Creatine kinase (CK), creatinine, blood urea nitrogen (BUN), alkaline phosphatase (ALP), alanine aminotransferase (ALT), aspartate aminotransferase (AST), bilirubin, troponin T, lactate, glucose, and pH.

Isolation of nuclei and single-nuclei RNA sequencing (snRNA-seq) analysis on frozen sWAT from burn patients

sWAT was collected from the burn site of patients at early or late time points post-burn (Supplementary Table 2). To isolate nuclei, frozen adipose tissues were cut into 1–2mm³ pieces over dry ice and transferred into lysis buffer (1M sucrose, 1M CaCl₂, 1M Mg(Ac)₂. 1M Tris-HCl, 0.1% Triton X-100, 0.5M EDTA, 40U/mL RNase Inhibitor and H₂O) for 5 minutes while homogenizing tissues over ice using glass douncers. The nuclei quality were assessed under a microscope using SYBR green (ThermoFisher Scientific #S7564). Samples were then centrifuged at 800 g for 10 minutes and washed with 2mL of wash buffer (1X PBS, 10% BSA, 0.2U/uL RNase Inhibitor) three times before suspending in 1mL of wash buffer and filtering nuclei through a 40uM Flowmi cell strainer (Sigma-Aldrich #BAH136800040). Nuclei were stained with 7AAD and sorted by FACS with a BD cell sorter for 7-AAD positive nuclei to exclude any debris or nuclei aggregates. Finally, nuclei were centrifuged at 800 g for 10 minutes, resuspended in cold wash buffer, counted, and immediately processed with the 10X Genomics platform following the Chromium Single Cell 3′ v3.1 kit protocol. A library was generated and sequenced on an Illumina NovaSeq 6000 with a sequencing depth of 50,000 reads per nuclei. Reads were aligned to the GRCh38–2020-A human transcriptome reference and nuclei were de-multiplexed, filtered for background noise, and counted using 10X CellRanger version 6.1.2.

The filtered count matrices were then uploaded to Cellenics^® (https://scp.biomage.net) for further processing, analysis, visualization and clustering. Data were then processed through a custom pipeline, which filtered for mitochondrial content, number of genes vs. UMIs using a spline fit curve, and doublets. Samples were then integrated using the Harmony method and embedded by the UMAP method. Louvain clustering based on 23 principal components was conducted with the resolution set to 0.5. Pathway analysis of differentially expressed genes (DEG; logFC>0.5; FDR>0.05) were performed by using Reactome terms in Panther. snRNA-seq data was deposited in the NCBI under submission PRJNA915460.

Inflammatory Responses

Circulating cytokines and chemokines were detected in the serum of burn patients collected at several timepoints during acute hospitalization. Cytokine profiling of select mediators were determined with the Milliplex MAP human cytokine/chemokine/growth factor panel A (cat# HCYTA-60K) using the Luminex 200 instrument (EMD Millipore). The cytokines analyzed in this study are as follows: eotaxin, epidermal growth factor (EGF), fibroblast growth factor-2 (FGF-2), granulocyte colony stimulating factor (GCSF), interferon-γ (IFN-γ), interleukin-1 beta (IL-1β), IL-6, IL-8, IL-10, IL-13, macrophage chemoattractant protein-1 (MCP-1), tumor necrosis factor-α (TNF-α), and vascular endothelial growth factor (VEGF).

Reverse Transcriptase Quantitative PCR (RT-qPCR) and Gene Expression Analyses

Total RNA was isolated from human sWAT samples using RNAzol (Sigma-Aldrich). Reverse transcription of normalized RNA was then performed to obtain cDNA using a high-capacity cDNA reverse transcriptase kit (Applied Biosciences). RT-qPCR was performed using the Step One Plus Real-Time PCR System (Applied Biosciences) and gene expression was calculated using 2^-ΔΔCT method expressed as a fold change, using 18S or Gapdh as housekeeping controls. Primer sequences are listed in Supplementary Table 3.

Metabolic Responses

Resting energy expenditure (REE) was measured under a steady-state using a Sensor Medics 2900 metabolic measurement cart, as previously described(10, 11). Percent predicted REE (pREE), calculated using the Harris-Benedict equation, and measured REE (mREE) were both used for these analyses. Normal mREE and pREE were derived from Willis et al.(12).

Statistical Analyses

All data were reported as mean ± SEM. For correlations, simple linear regressions were conducted. For REE, gene expression, and plasma cytokine analyses, a Shapiro-Wilk test was first conducted to test for normal or log-normal distribution. Then, a mixed effect analysis with either a Tukey’s (REE, gene expression) or Dunnett’s (plasma cytokines) multiple comparisons test or Mann-Whitney U non-parametric tests were conducted, accordingly. Statistical comparisons were conducted using GraphPad Prism software 9.3 (GraphPad Software, La Jolla, CA, USA). Significance was set at p<0.05, where: *p<0.05, **p<0.01, ***p<0.001, and ****p<0.0001.

Results

Resting Energy Expenditure Peaks Approximately 2-Weeks Following a Severe Burn

Hypermetabolism is a well-established hallmark following a severe burn, and a primary consequence of this response is increases in energy expenditure. We found burn patients exhibited increased pREE and mREE over the study period (Figure 1A-B). Burn patients exhibited peak increases in REE at approximately 2-weeks (10–15 days) post-burn (Figure 1A-B). It was interesting to note that pREE and mREE did not significantly increase over time during acute hospitalization (p=0.87 and p=0.93, respectively; Supplementary Figure 1A-B), This demonstrates that our cohort of burn patients remain in a hypermetabolic state long-term, maintaining REE above the average of normal control levels.

FIGURE 1. Alterations in energy expenditure, β-adrenergic receptors, and UCP1 gene expression in adipose tissue of burn patients.

(A) pREE and (B) mREE, as presented over a range of acute and long-term time points postburn, collected during acute hospitalization while patients were intubated. Adult severe burn patient white adipose tissue gene expression of the β-adrenergic receptors, (C) ADRβ-1, (D) ADRβ-2, (E) ADRβ-3, and (F) UCP1 expressed as a fold change over controls. (G) UCP1 gene correlated to days postburn. mREE indicates measured REE; pREE, predicted REE.

In addition to REE, we assessed functional biomarkers of systemic and organ dysfunction for up to 8-weeks post-injury (Supplemental Figure 2). Acutely (1–2 weeks post-injury), pH levels were reduced while lactate, glucose, CK, and AST were elevated above normal levels, which may indicate alterations in cardiovascular and hepatic function leading to an increased need for tissue respiration. These acute changes were followed by long-term elevations above normal ranges in markers of cardiac (troponin T), hepatic (ALP, ALT), and renal dysfunction (BUN). Bilirubin and creatinine showed an increased level in the first 2–3 weeks postburn with subsequent steady decline, but they did not fall outside normal ranges.

β1-Adrenergic Stress Correlates With Increased UCP1 Expression in Burn Patient Adipose Tissue

Severe burns trigger the activation of the sympathetic nervous system leading to an adrenergic stress response(13). This likely drives not only whole-body increases in energy expenditure, but also tissue-specific response. In sWAT of burn patients, ADRB1 gene expression significantly increased at 1–3 (p<0.05), 4–9 (p=0.0694), and ≥10 (p<0.05) days post-burn, compared with controls (Figure 1C). No significant changes in ADRB2 (Figure 1D) or ADRB3 (Figure 1E), with the exception of a signal towards an increase in ADRB3 between control and burn at 1–3 days post-burn (p=0.0995). Moreover, ADRB1 displayed the highest relative gene expression amongst the β-adrenergic receptors (Supplemental Figure 3). β-adrenergic activation led to the induction of the chief regulator of browning, UCP1, in the sWAT of burn patients at 4–9 and ≥10 (p<0.01) days post-burn compared with controls (Figure 1F). Additionally, UCP1 gene expression was significantly correlated with longer hospitalization (p=0.0003; Figure 1G). Together, these data indicate that WAT browning is initiated as a delayed response, which may be triggered by β1-adrenergic activation, and its expression is sustained long-term.

Temporal Alterations in Plasma Cytokine Expression

Another key consequence of hypermetabolism is the systemic immune and inflammatory response, which can occur immediately post-burn and result in a cytokine storm. In our cohort, we show that several cytokines and chemokines involved in immune cell recruitment and activation increased immediately (0–3 days post-burn), including MCP-1 (p<0.0001; Figure 2A), GCSF (p<0.001; Figure 2B), IL-10 (p<0.0001; Figure 2C), EGF (p<0.01; Figure 2D), and IL-8 (p<0.05; Figure 2E). MCP-1, GCSF, and IL-10 expression levels dropped dramatically at 4–9 days post-burn and remained non-significantly different from controls for the duration of the study period. However, EGF expression significantly increased at 10–15 days post-burn (p<0.01). Typically classified as pro-inflammatory cytokines, IL-8 (4–9, p<0.001; 10–15, p<0.0001; 16–30, p<0.01) and TNFα (4–9, p<0.01; 10–15, p<0.001; 16–30, p=0.06; Figure 2F) were elevated until 30 days post-burn, and appeared to reach peak significant at 10–15 days post-burn. IL-13, an anti-inflammatory cytokine, was significantly downregulated immediately post-burn and did not recover to normal levels by the end of the study period (p<0.0001; Figure 2G). Additional cytokines measured showed time-dependent alterations following a severe burn (Supplemental Figure 4).

FIGURE 2. Acute and chronic upregulation of systemic inflammatory mediators.

Circulating concentrations of inflammatory mediators (A) MCP-1, (B) GCSF, (C) IL-10, (D) EGF, (E) IL-8, (F) TNFα, and (G) IL-13 in burn patient plasma samples compared with normal controls. EGF indicates epidermal growth factor; GCSF, granulocyte colony-stimulating factor; MCP-1, macrophage chemoattractant protein-1.

Plasma IL-6 Initiates Adipose Tissue Signaling Cascade

The pleotropic cytokine IL-6 has been previously correlated with injury severity and REE in severe burn patients(14, 15). Moreover, IL-6 has been reported to be a key mediator of adipose tissue browning in mice and human burn patients(16). In the current study, we found plasma IL-6 increased immediately post-burn (p<0.0001) and remained significantly upregulated at 4–9 days post-burn, compared with controls (p<0.001; Figure 3A). Alterations in plasma IL-6 were followed by increased IL-6 gene expression in sWAT, which was upregulated at 1–3 days post-burn (p<0.05), peaked at 4–9 days post-burn (p<0.01), and then dropped below the significance threshold at ≥10 days post-burn (Figure 3B). The downstream mediator of IL-6, suppressor of cytokine signaling 3 (SOCS3), was similarly upregulated at 4–9 days post-burn (p<0.01; Figure 3C). At ≥10 days post-burn, sWAT expression of Signal transducer and activator of transcription 3 (STAT3), another downstream target of IL-6, signaled towards a decrease (p=0.0597; Figure 3D), while glucose transporter-1 (GLUT1) signaled towards an increase (p=0.0614; Figure 3E). Overall, this suggests that circulating IL-6 induces adipose tissue IL-6 expression, leading to alterations in adipocyte metabolism.

FIGURE 3. Circulating and adipose IL-6 signaling following a severe burn.

(A) Circulating concentrations of the inflammatory mediator IL-6 in burn patient plasma compared with controls. Gene expression of (B) IL-6, (C) SOCS3, (D) STAT3, and (E) GLUT1 in adult burn patient white adipose tissues expressed as a fold change over controls. GLUT1 indicates glucose transporter-1; SOCS3, suppressor of cytokine signaling 3; STAT3, Signal transducer and activator of transcription 3.

Adipokine Dysregulation in Burn Patient Adipose Tissue

To explore adipose-specific alterations further, we then focused on select adipokines that can be released from adipose tissue and may influence systemic metabolism. Here, we found that leptin (LEP; Figure 4A), adiponectin (ADIPOQ; Figure 4B) and transforming growth factor-β1 (TGF-β1; Figure 4C) were all significantly reduced at 1–3 (p<0.0001), 4–9 (p<0.0001) and ≥10 days post-burn (Lep, Adipoq, p<0.0001; TGF-β1, p<0.001). These data indicate that the consequences of hypermetabolism extend to adipokine dysregulation, which may play a role in mediating systemic organ dysfunction.

FIGURE 4. Disruption of expression in adipose-derived hormones.

Gene expression of (A) LEP, (B) ADIPOQ, and (C) TGF - β-1 in adult burn patient WAT expressed as a fold change over controls. ADIPOQ indicates adiponectin; LEP, leptin.

Single-nuclei transcriptomics reveal alterations in adipocyte expression and immune cell composition at acute and chronic stages post-burn

To gain further insight into the post-burn alterations in adipose tissue and to confirm our findings thus far, we conducted single-nuclei RNA sequencing (snRNA-seq) on sWAT collected from the burn site of 4 separate patients at early and late timepoints (Supplementary Table 2). After filtering, the combined dataset included 11,502 nuclei isolated from sWAT. We first confirmed that the datasets from different samples showed overlapping cell types (Supplementary Figure 5A). Initial clusters were then manually annotated based on the nuclei expression pattern of signature genes consistent with adipose progenitor cells (PDGFRA), pericytes (NOTCH3), adipocytes (PLIN1), T cells (CD247), macrophages (MERTK), endothelial cells (VWF), lymphatic endothelial cells (MMRN1) and mesenchymal stem cells (MSCs; cKIT) (Figure 5A-B and Supplementary Figure 5B), supported by the identification of the top 5 marker genes per integrated cluster (Supplementary Figure 5C).

Figure 5: Single-nuclei transcriptomics reveals alterations in adipose tissue at acute and chronic stages post-burn.

(A) UMAP plot of nuclei sub-populations combined from burn patient subcutaneous white adipose tissues at early and late timepoints post-burn. (B) UMAP plots of signature genes used for classifying cell populations as adipose progenitor cells (PDGFRA), pericytes (NOTCH3), adipocytes (PLIN1), T cells (CD247), macrophages (MERTK), endothelial cells (VWF), lymphatic endothelial cells (MMRN1) and mesenchymal stem cells (MSCs; cKIT). (C) Proportion of cell populations between early and late samples. (D) Dot plot of the expression of genes related to a pro-inflammatory/M1-like phenotype (PPARG, NFKB1, EREG, COL1A1, IL1B, IL6) or an anti-inflammatory/M2-like phenotype (RBJ, CD163, MRC1, CD86, IL10) in the macrophage sub-population between early and late samples. (E) Dot plot of the expression of genes related to adipogenesis (FASN), IL-6 signaling (SOCS3, IL6R, IL6), adipokines (LEP, ADIPOQ, TGFB1), browning (UCP1, CIDEA, DIO2), lipolysis (MGLL, LIPE, PNPLA1) and fatty acid trafficking (LPL, CD36, PPARG) within the adipocyte sub-population between early and late samples.

We discovered that the proportion of APCs, adipocytes and endothelial cells decreased while T cells, macrophages and MSCs increased from early to late timepoints post-burn (Figure 5C and Supplementary Figure 5D). Further, we found that percent expression of genes associated with an M1-like pro-inflammatory macrophage phenotype (PPARG, NFKB1, EREG, COL4A1, IL1B, and IL6) were increased within integrated macrophage sub-population of early samples, whereas genes associated with an M2-like anti-inflammatory macrophage phenotype (RPBJ, CD163, MRC1, CD86 and IL10) were increased within late samples (Figure 5D). These data are in line with previous work from our lab indicating the burn injuries provoke an increase in macrophage migration and polarization to sWATs(17).

We next profiled the expression of select markers in the adipocyte sub-population (Figure 5E). We found that markers associated with fatty acid synthesis (FASN) and IL-6 signaling (SOCS3, IL6R, IL6) were upregulated in early samples, consistent with our previous results. Of the previously measured adipokines, only a small proportion of adipocytes expressed LEP in early samples and TGFB1 in late samples. However, while ADIPOQ expression was negligent in early samples, it appeared to increase at later timepoints. We also detected a subtle increase in genes associated with adipose tissue browning (UCP1, DIO2, CIDEA) and a substantial increase in genes associated with lipolysis (MGLL, LIPE, PNPLA2) and fatty acid trafficking (LPL, CD36, PPARG) at late timepoints. Collectively, this supports the notion of a lack of adipokine signaling and an increase in late-stage adipose tissue remodelling, which likely fuels catabolism and may contribute to the hypermetabolic state.

Discussion

In this single-institution cohort study, we assessed the pathophysiology of clinical and adipose tissue metabolic outcomes in a large population of severe burn patients. Our primary findings are as follows; first, burn-induced increases in adrenergic stimulation within subcutaneous WAT may be driven by the β1-AR to induce and maintain adipose tissue browning. Additionally, systemic IL-6 expression induces alterations in the IL-6 signaling axis in adipose tissue, which leads to alterations in glucose homeostasis and further stimulates the browning response. Moreover, burn patients lack the expression of key adipokines involved in regulating fat mass and insulin sensitivity. These findings were supported by snRNA-seq data, which revealed previously unidentified insight into the alterations in adipocyte and immune cell compositions during the acute and chronic stages post-burn. The cumulative derangements in adipose tissue metabolism demonstrated here contribute to systemic inflammation, insulin resistance and catabolism that impedes the long-term recovery process (Figure 6).

FIGURE 6. Summary of study findings.

A severe burn induces profound adrenergic stress, leading to an increase in sympathetic activation that triggers the release of catecholamines such as norepinephrine (NE). This then binds to primarily β1-adrenergic receptors on the adipocyte membrane to upregulate UCP1 gene expression and stimulates adipose tissue browning. Concomitantly, adrenergic stress also induces systemic inflammation, leading to the immediate yet prolonged release of several inflammatory mediators, most notably IL-6. This then binds to its receptors on adipocyte membranes and may also indirectly promote adipose tissue browning while also activating its downstream negative regulator SOCS3, which inhibits IL-6/STAT3 signaling and IRS1 to prevent insulin from binding to its receptors and promote insulin resistance. This may also have inhibitory actions on adipokine production. Basal glucose transport mediated by GLUT1 slightly increases to compensate, potentially due to disruptions in insulin-stimulated glucose uptake. Together, the dysregulation of normal adipocyte metabolism may contribute to the catabolic, insulin-resistant state, chief hallmarks of a severe burn. ADIPOQ indicates adiponectin; GLUTI, glucose transporter-1; LEP, leptin.

Several studies have previously defined the long-term trajectories of the hypermetabolic response to severe burns. Jeschke et al. comprehensively detailed the persistent increase in fat and lean catabolism, inflammatory and hormonal stress markers, and resting metabolic rate in severely burned pediatric patients, which remained elevated for up to 3-years post-injury(1). Subsequently, Stanojcic et al. described a similar phenomenon in adult burn patients(9). However, the significance of adipose tissue metabolism to burn-induced catabolism was only recently recognized within the last decade(18). We(14) and others(19) uncovered molecular and functional evidence of delayed (>10 days) adipose tissue browning and remodeling occurring in the sWAT of mice, pediatric and adult burn patients. Further, sWAT browning was shown to positively correlates with REE(14, 19). While these data, which are supported by several murine studies, suggest that WAT remodeling may, in part, induce and perpetuate hypermetabolism, the evidence in adult burn patients is held up only by small sample sizes, which limits the interpretative value of the data. In our cohort of 65 severely burned adult patients, from which 86 sWAT and 216 plasma samples were collected, we present robust support for previous findings.

Consistent with Stanojcic et al(9), we found that peak REE occurs 1–2 weeks post-burn. Increases in REE are coupled with alterations in circulating biomarkers, such as a reduction in pH and increases in lactate, glucose, CK, and AST, which likely contributes to the increased demand for oxygen consumption and tissue respiration and leads to long-term organ damage(13).

The severe adrenergic stress response immediately following a burn injury provokes a surge in circulating catecholamines, which then bind to β-AR on adipocyte membranes to initiate catabolic responses. We demonstrate novel evidence that the β1-AR is most responsive β-AR subtype to a severe burn, while β2- and β3-ARs did not significantly increase. Given that β1-ARs and β2-ARs are the most prominent subtypes in human white adipocytes(20), and unlike mice, humans largely lack β3-ARs expression(21, 22), it seems plausible that β1-ARs may primarily drive the burn-induced remodeling process in human sWAT. Our findings are in support of previous work in cancer cachexic patients. Cao et al. found that, in comparison to weight-stable and non-malignant cancer patients, gene and protein expressions of ADRB1 and the lipolytic enzyme hormone sensitive lipase (HSL) were significantly upregulated and correlated with one another in sWAT of cancer cachectic patients(23). Further, ADBR1 was significantly correlated with the rate of glycerol and free fatty acid release, indicating that the activation of ADBR1 may account for fat catabolism in cancer cachectic patients.

In support of the papers led by Patsouris(14) and Sidossis(19), we demonstrate that UCP1 gene expression in sWAT increases as a delayed response with significant increases detected as early as 4-days post-burn, which continuously increased throughout hospitalization. To our knowledge, our study demonstrates the longest duration of sWAT browning in humans, with one patient exhibiting >280-fold increase in UCP1 at 42 days post-burn. This underscores the importance and unique opportunity to study adipose tissue metabolism in a clinical population of prolonged adrenergic stress, and highlights the potential translational value to apply towards mitigating obesity and other metabolic disorders.

Fat catabolism has a significant effect on the release of adipose-derived factors that play major roles in regulating systemic metabolism. Wade et al. unexpectedly found that circulating concentrations of leptin and adiponectin were suppressed in patients with severe burns while insulin and resistin were increased, suggesting abnormal adipokine signaling(24). Our study, which to our knowledge is the first to look at adipokine expression directly in adipose tissue, supports these findings. We report an immediate yet prolonged decrease of leptin, adiponectin and TGF-β1 gene expression in adipose tissue, indicating that the production of these adipokines into circulation is attenuated. Leptin, a satiety hormone, normally increases with fat mass while adiponectin conversely decreases. It appears that, following a severe burn, this reciprocal relationship is immediately dysregulated. These changes may be a result of a combination of several factors, including decreased expression due to wound healing(25, 26), insulin resistance(27), inadequate nutrition, or potentially a lack of sufficient adipose tissue to produce adipokines. Interestingly, leptin administration mitigated multi-organ damage (28), and stimulated wound angiogenesis (29), in rats at acute timepoints. Given our findings, future work should further investigate the effectiveness of leptin administration over long-term timepoints in murine models, and consider the potential therapeutic benefit to hypermetabolic burn patients. Moreover, TGF-β1 is an established regulator of extracellular matrix (ECM) remodeling and has been linked to lipogenesis and fat expansion(30). Our observation of decreased TGF-β1 expression indicates a lack of fibrosis and lipogenesis in adipose tissue, which is supportive of the increase in lipolysis and WAT browning demonstrated here and by others(8, 31).

Many studies have highlighted the role of inflammatory factors in the regulation of adipose tissue metabolism. Following a severe burn, immune cells including monocytes, macrophages and neutrophils are immediately activated. This leads to the systemic inflammatory response syndrome, in which several pro- and anti-inflammatory chemokines and cytokines are uncontrollably released into circulation and further recruit immune cells(13). In support of this, we found circulating inflammatory markers, such as MCP-1, GC-SF, IL-10 and IL-8, were upregulated within the first three days post-injury. Most notably, IL-6, which has been implicated as a key mediator of post-burn hypermetabolism(15, 16), increased in circulation immediately post-burn and remained upregulated for up to 30-days post-burn in our cohort. The immediate increase in circulating IL-6 was followed by increased gene expression of IL-6 and its downstream regulator SOCS3 in adipose tissue. SOCS3 mediates the inhibitory effect of IL-6/STAT3(32), as demonstrated here, while also leading to the disruption of insulin signaling by degrading and preventing insulin receptor substrates from binding to insulin receptors(32). Given that hyperglycemia and insulin resistance are key hallmarks of post-burn hypermetabolism(33), our data demonstrate that chronic systemic inflammation may drive IL-6-mediated insulin resistance in adipose tissue and may further compromise whole-body insulin sensitivity. While the role of IL-6 and adipose has been heavily studied under the context of burn-induced browning, to our knowledge, this is the first study investigating downstream alterations of IL-6 signaling in human burn patients.

With the exception of macrophages, a complete lack of knowledge currently exists regarding how the adipose tissue microenvironment behaves following a severe burn. To provide insight into the temporal alterations in adipose heterogeneity and to further elucidate our findings thus far, we performed snRNA-seq on whole adipose tissue of four burn patients at acute (1 and 2 days post-burn) and long-term (12 or 17 days post-burn) time points, which are reflective of the ebb and hypermetabolic flow phases occurring after injury(13). Consistent with our gene expression data, we showed that adipocytes exhibit and upregulation of IL-6 signaling at early time points, which was not observed during late stages. Concurrent with studies demonstrating adipose catabolism, the proportion of adipocytes was found to decrease from early to late time points. This was supported by the increased expression of lipolytic (MGLL, LIPE, PNPLA2) and WAT browning (UCP-1, DIO2, CIDEA) enzymes detected in late samples, and by pathway analysis demonstrating significant enrichment in mitochondrial bioenergetics, fatty acid and triglyceride metabolism, and heat production by uncoupling proteins. Of note, we observed a decrease in the proportion of two distinct endothelial cell populations from early to late samples. As vascular endothelial cells have been shown to differentiate into beige adipocytes(34), this may indicate the fate of adipose endothelial cells following burns.

Of the CD45⁺ leukocytes, we observed a proportional decrease in CD247⁺ (CD3⁺) T cells and a substantial increase in total macrophages, of which three sub-types were distinguished. Total macrophages in early samples displayed a more pro-inflammatory, lipid-laden phenotype, as determined by the upregulation of PPARG, NFKB1 and IL1B. However, we detected a clear phenotypic switch towards an M2-like, anti-inflammatory phenotype in late samples, which highly expressed CD163, MRC1, and CD86. Taken together, our snRNA-seq data represents the first characterization of adipose heterogeneity at a single nuclei resolution, which underscores the remarkably dynamic trajectory of sWAT cells in response to severe burns.

While the current study highlights several important findings, it is not without limitations. First, we solely conducted gene expression analysis on our adipose tissue samples. However, it is important to confirm translation with protein quantification and functional assays, such as respirometry. This is especially significant for quantifying the function of UCP-1, which is best determined by measuring the mitochondrial respiratory response of adipose tissue to GDP(35). Given the robust number of samples obtained in this study, and that they were kept frozen until analysis, we were unable to conduct functional measurements. However, increased mitochondrial respiratory capacity has been previously observed in sWAT of burn patients(36). In addition, our snRNA-seq data was limited to burn patient samples, thus we were unable to compare to healthy controls. Given the recent release of many publicly available single-cell and single-nuclei datasets of healthy human adipose tissues (37, 38), additional studies should use these existing data to better delineate burn pathophysiology. Finally, this was conducted as a single-institution cohort study, thus our findings should be validated by others.

Severe burns induce chronic, debilitating alterations in systemic and tissue-specific function that results in long-term pathologies, including fat and muscle catabolism, insulin resistance and systemic inflammation. In this study, we enrolled a large patient cohort from which we obtained robust tissue sample sizes which led to the confirmation of previous findings and provided novel insights into burn-induced pathophysiological responses, both systemically and within adipose tissue. Moreover, through snRNA-seq, we provide the opportunity for further research exploring the adipose niche at a single-nuclei resolution in a clinical population of extreme adrenergic stress. Further work should focus on integrating snRNA seq with functional and clinical outcomes to uncover key determinants that could predict the trajectory of burn patient sub-populations. For example, the ability to predict whether patients will survive or not, whether they will develop sepsis, and how patients respond to specific therapeutics on a single-cell level based would greatly enhance the quality of burn care. Moreover, while outside the scope of this paper, interaction analyses, such as pseudotime ordering (39), could be conducted to reconstruct trajectories of burn pathophysiology. This may infer novel cell-cell signaling interactions, which could pave the way to novel therapeutic targets.

Not only are these findings pertinent to severe burn patients, but may also be translated to other hypermetabolic conditions such as cancer cachexia and hypermetabolic amyotrophic lateral sclerosis (ALS)(40). Our study also underscores the differences in cold-induced compared with burn-induced adrenergic responses, and may also provide unique insight into the use of prolonged adrenergic stimulation to mitigate the consequences of obesity and its comorbidities.

Supplementary Material

Supplemental Data File 1. Supplementary Figure 1: Clinical biomarkers of organ function.

Several biomarkers of systemic and tissue-specific function were measured in the plasma of burn patients up to 8 weeks post-admission. These included (A) pH, (B) arterial lactate, (C) glucose, (D) CK, (E) Troponin T, (F) AST, (G) ALT, (H) ALP, (I) Bilirubin, (J) Creatinine, and (K) BUN. Greyed regions represent the normal range of values based on institutional guidelines.

Supplemental Data File 2. Supplementary Figure 2: Predicted and measured energy expenditure in burn patients.

(A) pREE and (B) mREE as shown in correlation with days post-burn as shown in correlation with days post-burn.

Supplemental Data File 3. Supplementary Figure 3: Comparison of relative expression of β-adrenergic receptors.

Relative gene expression of the β-adrenergic receptors, ADRβ1, ADRβ2, and ADRβ3 in adipose tissue of burn patients compared with controls.

Supplemental Data File 4. Supplementary Figure 4: Additional systemic inflammatory mediators.

Additional circulating inflammatory mediators (A) TGFα, (B) FGF2, (C) VEGF, (D) IFNγ, (E) IL-1β, and (F) Eotaxin, measured in the plasma of adult burn patients over several timepoints post-burn compared with controls.

Supplemental Data File 5. Supplementary Figure 5: Single-nuclei distribution between samples and integrated sub-population analysis.

(A) UMAP plot of all samples used demonstrating similar distributions among nuclei sub-populations. (B) UMAP plot of integrated nuclei sub-populations combined from burn patient subcutaneous white adipose tissues at early and late timepoints post-burn. (C) Heatmap of top 5 markers expressed in the integrated nuclei sub-populations, including macrophages, APCs, endothelial cells, adipocytes, T cells, lymphatic-endothelial cells and MSCs. (D) Proportion of nuclei sub-populations in early or late samples.