This research was approved by the Ethics Committee of the Nanchang University (Approval No. NCULAE-20221031185). BALB/C mice were procured from Cavens Laboratory Animal Co., Ltd. (Changzhou, China) and maintained in a germ-free facility at the Animal Laboratory Center for Nanchang University. The mice were housed in a controlled environment with a temperature of 23 $\pm$ 2 ℃, humidity of 55 $\pm$ 10%, a 12-hour light/dark cycle, and provided ad libitum access to standard laboratory chow and water. The mice were randomly allocated into two groups: control and experimental colitis (3% DSS [MW: 36–50 kDa, 60316ES25, YEASEN Biotechnology, Shanghai, China] dissolved in drinking water for 7 days [27]). Mice were then randomly assigned into four groups: control, experimental colitis, experimental colitis with adeno-associated virus (AAV)- short hairpin RNA negative control (shNC), and experimental colitis with AAV-shHMGB1. Two weeks before the onset of DSS treatment, a 0.1 mL dose of HMGB1 interference virus (AAV-shHMGB1, 4 $\times$ 10¹⁰) or control virus (AAV-shNC, 4 $\times$ 10¹⁰) was administered into the colon of mice. All animals were euthanized (anaesthetized by isoflurane and sacrificed by cervical dislocation) on day 8, and colon tissue was harvested for measurement of colon length. The Disease Activity Index (DAI) is computed based on two criteria: weight loss, with scores assigned as follows—1% to 5% loss scores 1 point, 5% to 10% loss scores 2 points, 10% to 20% loss scores 3 points, and losses greater than 20% score 4 points; stool consistency and rectal bleeding, where normal stool scores 0 points, mild diarrhea scores 1 point, watery stool scores 2 points, the presence of shallow blood and loose stool scores 3 points, and distinct rectal bleeding scores 4 points.

The colon specimens were fixed in 4% paraformaldehyde for 12 h, followed by dehydration, paraffin embedding, sectioning (3 µm), and staining with an H&E staining kit (Beijing Solaibao Technology Co., Ltd., Beijing, China). Subsequently, images were captured using an optical microscope.

Dewaxed and rehydrated slides of colon tissue were blocked for endogenous peroxidase activity using an H₂O₂ solution post-antigen retrieval, followed by blocking for 30 min with 10% normal goat serum. Tissue sections were then exposed to anti-F4/80 (ab300421, Abcam, Cambridge, MA, USA) or anti-HMGB1 (ab79823, Abcam, Cambridge, MA, USA) overnight at 4 °C, and subsequently incubated with secondary antibody (S0001/2, Affinity, Cincinnati, OH, USA) at 37 °C for 60 min. After 3,3’-diaminobenzidine (DAB) staining for 5 min, hematoxylin counterstaining was performed before dehydration. Imaging was conducted using an optical microscope (BX63, Olympus, Tokyo, Japan).

Mice were euthanized post-experimentation under anesthesia. Serum was collected from blood after centrifugation at 3000 rpm for 15 min and the levels of tumor necrosis factor $\alpha$ (TNF-$\alpha$, E-EL-M3063) and interleukin-6 (IL-6, E-EL-M0044c, Elabscience, Houston, TX, USA) were quantified utilizing ELISA kits as per the manufacturer’s instructions.

Colon tissues were collected, diced, enzymatically digested, filtered, purified, and centrifuged to isolate individual cells. These were stained with specific antibodies including PerCP-Cy5.5 rat anti-CD11b (550993, BD, New York, NY, USA), fluorescein Isothiocyanate (FITC)-leukocyte common antigen (CD45) (553080, BD, New York, NY, USA), APC-iNOS (696807, BioLegend, San Diego, CA, USA), PE-CD206 (568273, BD, New York, NY, USA), and BV421-F4/80 (554722, BD, New York, NY, USA) at the recommended dilutions for 30 min at 4 °C in the dark. The single cell suspensions were then analyzed using flow cytometry (CytoFLEX S, Beckman Coulter, Brea, CA, USA) and FlowJo LLC software (Ver.10, Ashland, OR, USA), and the percentages of macrophages (CD45⁺F4/80⁺), M1-macrophages (CD45⁺F4/80⁺iNOS⁺) and M2-macrophages (CD45⁺F4/80⁺CD206⁺) were quantified.

All cell lines were validated by STR profiling and tested negative for mycoplasma. Cells were all cultured in a humidified incubator at 37 °C and 5% CO₂. RAW264.7 cells (YS1879C, Shanghai Yaji Biotechnology Co., Ltd., Shanghai, China) were cultured in DMEM medium (PM150210B, Wuhan Pricella Biotechnology Co., Ltd., Wuhan, China) supplemented with 10% FBS, 100 U/mL penicillin and streptomycin in disposable plastic cell culture flasks at 37 °C with 5% CO₂. To investigate the effect of HMGB1 on macrophage polarization, cells were divided into four groups: overexpression negative control (NC), HMGB1, small interfering RNA targeing negative control (si-NC), and si-HMGB1. The first two groups were polarized into M2 macrophages using 50 ng/mL IL-4 (90105ES08, YEASEN Biotechnology, Shanghai, China), while the latter two groups were polarized into M1 macrophages using 100 ng/mL interferon-$\gamma$ (IFN-$\gamma$, 91212ES08, YEASEN Biotechnology, Shanghai, China) and 250 ng/mL lipopolysaccharide (LPS, 00-4976-93, eBioscience, San Diego, CA, USA). To explore the role of coenzyme A (CoA) in si-HMGB1-mediated macrophage polarization, the cells were subsequently divided into si-NC+vehicle, si-HMGB1+vehicle, and si-HMGB1+CoA groups. Further investigation of the effect of Cpt1a on si-HMGB1-mediated macrophage polarization involved the transfection of cells with Cpt1a and HMGB1 interference sequences, followed by polarization into M1 macrophages using IFN-$\gamma$ and LPS.

Small interfering RNA (siRNA) that targeted HMGB1 and Cpt1a, as well as their corresponding negative controls, were obtained from GenePharma (Shanghai, China). Transfection was conducted using Lipofectamine 3000 (Invitrogen, Carlsbad, CA, USA).

Total RNA was extracted from cells using TRIzol reagent (15596026CN, Invitrogen, Carlsbad, CA, USA). The mRNA expression levels for inducible nitric oxide synthase (iNOS), Arginase 1 (Arg-1) and Carnitine palmitoyltransferase 1A (Cpt1a) were quantified with a SweScript One-Step RT-PCR Kit (G3335-100, Servicebio, Wuhan, China), as per the manufacturer’s instructions. $\beta$-actin served as the internal control. Primer sequences for iNOS, Arg-1, Cpt1a and $\beta$-actin are shown in Table 1. Relative mRNA expression was calculated with the 2^{-Δ⁢Δ⁢Ct} method.

Cellular proteins were extracted and quantified using a BCA kit (E112, Vazyme, Nanjing, China). Subsequently, the proteins were separated by Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred onto Polyvinylidene Fluoride (PVDF) membranes (IPVH00010, Immobilon, Millipore Co., Bedford, MA, USA). Following blocking, the membranes were incubated overnight at 4 °C with primary antibodies against iNOS (1:1000, 18985-1-AP, ProteinTech Group, Chicago, IL, USA), Arg-1 (1:5000, 16001-1-AP, ProteinTech Group, Chicago, IL, USA), Cpt1a (1:10,000, 15184-1-AP, Proteintech), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). After washing, horseradish peroxidase (HRP)-conjugated Immunoglobulin G (IgG) secondary antibodies were then applied for 2 h at 37 °C. Protein detection was performed using an enhanced chemiluminescence kit (BL523A, Biosharp, Shanghai, China), and Band Scan software (ImageJ, National Institutes of Health, Bethesda, MD, USA) was used to analyze the optical density of bands.

Cellular glycolysis activity was assessed using a glycolysis assay kit (ab197244, Abcam, Cambridge, MA, USA). Briefly, harvested cells were pelleted and the supernatant discarded. Cells were disrupted by ultrasonication at a cell: distilled water ratio of 500:1 to 1000:1, followed by incubation at 95 °C for 10 min. After cooling, the cell lysate was centrifuged and the resulting supernatant collected. A spectrophotometer/plate reader was preheated, and the detection wavelength set at 505 nm for analysis.

The assessment of FAO in cells was performed using a FAO assay kit (BR00001, AssayGenie, Wuhan, China) according to the manufacturer’s instructions. About 10⁶ cells were resuspended in 50–100 µL of cold 1$\times$cell lysate solution. This was centrifuged at 14,000 rpm for 5 min in the cold. The Bicinchoninic Acid Assay (BCA) assay was used to determine protein concentration. Each sample was treated with 50 µL of control solution and 50 µL of reaction solution, gently stirred for 10 sec, and then incubated at 37 °C for 60–120 min. The gradual appearance of cherry red indicates a reaction, and absorbance at 492 nm was subsequently determined using a tablet reader.

The level of acetyl CoA in the cell was assessed using a commercial kit (ab87546, Abcam, Cambridge, MA, USA) according to the manufacturer’s instructions. Fifty µL of standard dilution solution was added to the standard well, and sample adjusted to 50 µL with assay buffer was added to the sample well. The sample adjusted to 50 µL with assay buffer was also added to the background sample well. CoA quencher was added to all wells and incubated for 5 min. Quencher remover (2 µL) was then added to all wells, followed by another 5 min incubation. Reaction mixtures were prepared within the range of the standard curve. Each well received 50 µL of reaction mixture, followed by incubation at 37 °C for 10 min. Measurements were then made using a microporous reader. The acetyl CoA concentration was calculated as Ay/Sv, where Ay is the amount of acetyl CoA in the sample well (pmol), and Sv is the volume of sample added to the well (µL). The molecular weight of acetyl CoA is 809.6 g/mol.

ChIP-PCR was conducted on cell samples in order to assess the binding of HMGB1 to the promoter region of Cpt1a. Cells were first incubated with 1% formaldehyde for 10 min to induce protein-DNA cross-linking. Pre-cleared chromatin was aliquoted equally and subjected to sonication. Immunoprecipitation was then performed using a monoclonal antibody against HMGB1 (MA5-31967, Invitrogen) or with a corresponding IgG isotype control. Subsequently, DNA was purified and prepared for quantitative PCR (qPCR) analysis.

The full-length human Cpt1a promoter (–2000 to +100) and three truncated fragments (–2000 to –1300, –1299 to –600, –599 to +100) were cloned into the pGL3 luciferase reporter vector. These reporter constructs, along with the HMGB1 expression vector, were co-transfected into cells using Lipofectamine 2000 (11668-027, Invitrogen). Luciferase activity was then quantified 48 h post-transfection using the luciferase reporter assay system.

Statistical evaluation was performed utilizing Prism 8 (GraphPad Software, GraphPad Software, Inc., San Diego, CA, USA). Experimental outcomes were presented as the mean $\pm$ standard deviation (SD) and were obtained from at least three independent replicates. Normally distributed data was analyzed with the Student’s t-test or one-way ANOVA, while data that failed the normality test was subjected to non-parametric tests. Specifically, the Mann-Whitney U test was applied for two independent samples, and the Kruskal-Wallis test for three or more groups. A p-value 0.05 was deemed to indicate statistical significance.

H&E staining (Fig. 1A) revealed an intact colon tissue structure in the control group, with no inflammation. In the UC group, the colon tissue showed scattered and missing epithelium, with significant edema and bleeding under the mucosa. The formation of ulcers and the shorter colon length observed in the UC group demonstrated the successful modeling of UC (Fig. 1B). Subsequent immunohistochemical analysis revealed upregulated HMGB1 expression and increased macrophage infiltration in the colon tissue of UC mice (Fig. 1C,D). Moreover, flow cytometry showed an imbalance in the M1/M2 macrophage ratio in the colon tissue of UC mice (Fig. 1E). These results suggest that upregulation of HMGB1 expression and infiltration of macrophages are involved in the pathogenesis of UC.

Fig. 1.

High mobility group box-1 protein (HMGB1) expression and macrophage infiltration in the colon mucosal tissue of ulcerative colitis (UC) mice. Mice were administered dextran sulfate sodium (DSS) to induce colitis, with the control group receiving no treatment. (A) Representative histological images of colon tissue sections stained with H&E, scale bars = 100μm. (B) Representative colon images, scale bars = 100μm. (C) F4/80 immunohistochemical images of colitis-affected tissues, scale bars = 100μm. (D) HMGB1 immunohistochemistry of colitis tissue, scale bars = 100μm. (E) Analysis of the M1/M2 macrophage ratio in colon tissue by flow cytometry. Statistical analysis was conducted using Student’s t-test. Data shown are the mean $\pm$ SD (n = 3) of three independent experiments, *p 0.05, ***p 0.001 vs control group. F4/80, Adhesion G protein-coupled receptor E1; CD206, mannose receptor C type 1; CD45, leukocyte common antigen; iNOS, inducible nitric oxide synthase.

To explore the impact of HMGB1 on macrophage metabolism and polarization, mouse macrophages (RAW264.7) were subjected to HMGB1 overexpression or knockdown by HMGB1 overexpression plasmid and small interference RNA, and RT-PCR and western blot were used to validate the overexpression or knockdown efficiency (Supplementary Fig. 1). This was shown to induce the M1 or M2 macrophage phenotypes, respectively. The upregulation of iNOS and downregulation of Arg-1 in HMGB1-overexpressing cells showed that HMGB1 overexpression promotes the polarization of M1 macrophages (Fig. 2A). Furthermore, the downregulation of iNOS and upregulation of Arg-1 in HMGB1-knockdown cells demonstrated that knockdown of HMGB1 promotes M2 macrophage polarization (Fig. 2B). Next, macrophage glycolipid metabolism was studied by evaluating the ECAR, FAO activity and acetyl CoA levels. As shown in Fig. 2C–H, overexpression of HMGB1 increased ECAR, reduced FAO, and increased the acetyl CoA level. In contrast, knockdown of HMGB1 had the opposite effects, suggesting that overexpression of HMGB1 promotes glycolysis and inhibits FAO. The addition of acetyl CoA to HMGB1-knockdown cells, followed by induction of M1 macrophages using IFN-$\gamma$ and LPS, reversed the changes in expression of iNOS and Arg-1 (Fig. 2I,J). In summary, these results suggest that overexpression of HMGB1 may promote M1 macrophage polarization by increasing glycolysis and inhibiting FAO.

Fig. 2.

Effect of HMGB1 on macrophage metabolism and polarization. (A,B) The mRNA and protein expression levels of iNOS and Arg-1 were quantified using RT-PCR and Western blot, respectively. (C,F) Measurement of extracellular acidification rate (ECAR) in cells. (D,G) Evaluation of fatty acid oxidation (FAO) in cells. (E,H) Quantification of acetyl-CoA content in cells. (I,J) Quantification of iNOS and Arg-1 mRNA and protein expression. Statistical analysis was conducted using Student’s t-test or one-way ANOVA. Data shown are the mean $\pm$ SD (n = 3) of three independent experiments. **p 0.01, ***p 0.001 vs NC or si-NC/si-NC+vehicle group. ^#⁢#p 0.01, ^#⁢#⁢#p 0.001 vs si-HMGB1+vehicle. iNOS, inducible nitric oxide synthase; Arg-1, arginase-1; RT-PCR, reverse transcription-polymerase chain reaction; ANOVA, analysis of variance; HMGB1, high mobility group box 1; siHMGB1, small interfering RNA targeting HMGB1; NC, negative control; si-NC, small interfering RNA negative control; RFU, relative fluorescence units; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; CoA, coenzyme A.

In view of its important role in cellular glucose and lipid metabolism, we next evaluated the expression of Cpt1a in cells with HMGB1 overexpression or knockdown. RT-PCR and Western blot results showed that HMGB1 negatively regulates Cpt1a gene and protein expression (Fig. 3A,B). Moreover, ChIP-PCR experiments demonstrated the binding of HMGB1 protein to the Cpt1a promoter (Fig. 3C). The luciferase reporter gene method also showed that HMGB1 could inhibit transcriptional activity from the Cpt1a promoter (Fig. 3D). We conclude from these findings that HMGB1 negatively regulates Cpt1a expression by inhibiting transcription from its promoter.

Fig. 3.

HMGB1 inhibits Cpt1a transcription by binding to its promoter. (A,B) HMGB1-knockdown cells were induced into M1 macrophages by IFN-$\gamma$ and LPS, while HMGB1-overexpressing cells were induced into M2 macrophages by IL-4. RT-PCR and Western blot analysis of Cpt1a expression. (C) Chromatin Immunoprecipitation-quantitative Polymerase Chain Reaction (ChIP-qPCR) assays of HMGB1 occupancy at the Cpt1a promoter region. (D) Luciferase reporter assay of Cpt1a transcriptional activity. Statistical analysis was conducted using the Mann-Whitney U test. Data shown are the mean $\pm$ SD (n = 3) of three independent experiments. *p 0.05, ***p 0.001 vs NC/si-NC. Cpt1a, Carnitine palmitoyltransferase 1A; HMGB1, High Mobility Group Box 1; siHMGB1, small interfering RNA targeting HMGB1; NC, negative control; si-NC, small interfering RNA negative control; IFN-$\gamma$, interferon-$\gamma$; LPS, lipopolysaccharide; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; RT-PCR, reverse transcription-polymerase chain reaction.

To confirm that HMGB1 modulates the metabolism and polarization of macrophages through negative regulation of Cpt1a, we conducted simultaneous intervention of HMGB1 and Cpt1a. Knockdown of Cpt1a resulted in the upregulation of iNOS and downregulation of Arg-1. In contrast, knockdown of HMGB1 caused downregulation of iNOS and upregulation of Arg-1. Furthermore, knockdown of Cpt1a reversed the effects of si-HMGB1 on macrophage polarization (Fig. 4A,B). Interestingly, knockdown of Cpt1a increased ECAR, reduced FAO, and upregulated acetyl CoA, whereas knockdown of HMGB1 had the opposite effects. As expected, knockdown of Cpt1a reversed the effects of si-HMGB1 on macrophage metabolism (Fig. 4C–E). Therefore, we conclude that HMGB1 affects macrophage metabolism and polarization by regulating Cpt1a.

Fig. 4.

Inhibition of Cpt1a reverses the effects of si-HMGB1 on macrophage metabolism and polarization. RAW264.7 cells were transfected with interference sequences for Cpt1a and HMGB1, then induced into M1 macrophages using IFN-$\gamma$ and LPS. (A,B) RT-PCR and Western blot analysis of iNOS and Arg-1 expression. (C) Measurement of ECAR in cells. (D) Evaluation of FAO in cells. (E) Detection of acetyl-CoA content in cells. Statistical analysis was conducted using one-way ANOVA. Data shown are the mean $\pm$ SD (n = 3) of three independent experiments, ***p 0.001 vs si-NC1+si-NC2. ^#⁢#⁢#p 0.001 vs si-HMGB1+si-NC2. Cpt1a, carnitine palmitoyltransferase 1A; HMGB1, high mobility group box 1; siHMGB1, small interfering RNA targeting HMGB1; NC, negative control; si-NC, small interfering RNA negative control; IFN-$\gamma$, interferon-$\gamma$; LPS, lipopolysaccharide; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; RT-PCR, reverse transcription-polymerase chain reaction; iNOS, inducible nitric oxide synthase; Arg-1, arginase 1; ECAR, extracellular acidification rate; FAO, fatty acid oxidation.

This project uses AAV virus to infect the colon of mice. We first evaluated the in vivo knockout efficiency of AAV-shHMGB1 and its impact on mouse colon tissue. Normal mice were injected with AAV-shNC and AAV- shHMGB1, respectively. Supplementary Fig. 2 showed the expression of HMGB1 in mouse colon tissue. Compared with the AAV-shNC injection group, the expression of HMGB1 in the colon tissue of AAV- shHMGB1 injection group mice was significantly reduced. In addition, histological examination and colon morphology evaluation showed that AAV-shHMGB1 did not have any harmful effects on mouse colon tissue. Supplementary Fig. 3A showed the variations in animal body weight and Disease Activity Index (DAI) scores.

To determine the role of HMGB1 knockdown on UC progression, mice were administered DSS to induce colitis and were also treated with HMGB1 interference virus (AAV-shHMGB1). The variations in animal body weight and DAI scores are depicted in Supplementary Fig. 3B. Compared to the AAV-shNC group, mice in the AAV-shHMGB1 group exhibited less submucosal edema and bleeding in colon tissue (Fig. 5A). Additionally, the colon length in the AAV-shHMGB1 group was significantly longer than in the AAV-shNC group (Fig. 5B). Furthermore, reduced levels of the pro-inflammatory cytokines TNF-$\alpha$ and IL-6 (Fig. 5C,D) in the AAV-shHMGB1 group suggested that HMGB1 knockdown substantially improved symptoms in UC mice. Finally, a notable decrease was observed in the M1/M2 macrophage ratio of colon tissue in the AAV-shHMGB1 group (Fig. 5E,F). Taken together, these results indicate that inhibition of HMGB1 regulates macrophage polarization to ameliorate UC progression.

Fig. 5.

Inhibition of HMGB1 regulates macrophage polarization to ameliorate UC progression. (A) Representative histological analysis of the colon using H&E staining, scale bars = 100μm. (B) Representative images of the colon, and quantification of colon length. (C,D) ELISA assay results of TNF-$\alpha$ and IL-6 levels in mouse serum. (E,F) Flow cytometry analysis of the macrophage M1/M2 ratio in colon tissue. Statistical analysis was conducted using one-way ANOVA. Data shown are the mean $\pm$ SD (n = 5) of three independent experiments, **p 0.01, ***p 0.001 vs control; ^#p 0.05, ^#⁢#p 0.01 vs UC+AAV-shNC. UC, ulcerative colitis; AAV-shNC, adeno-associated virus-short hairpin RNA negative control; AAV-shHMGB1, adeno-associated virus-short hairpin RNA targeting HMGB1; TNF-$\alpha$, tumor necrosis factor $\alpha$; IL-6, interleukin-6; iNOS, inducible nitric oxide synthase; CD206, Mannose Receptor C Type 1.