Beiging of perivascular adipose tissue regulates its inflammation and vascular remodeling

Abstract

Although inflammation plays critical roles in the development of atherosclerosis, its regulatory mechanisms remain incompletely understood. Perivascular adipose tissue (PVAT) has been reported to undergo inflammatory changes in response to vascular injury. Here, we show that vascular injury induces the beiging (brown adipose tissue-like phenotype change) of PVAT, which fine-tunes inflammatory response and thus vascular remodeling as a protective mechanism. In a mouse model of endovascular injury, macrophages accumulate in PVAT, causing beiging phenotype change. Inhibition of PVAT beiging by genetically silencing PRDM16, a key regulator to beiging, exacerbates inflammation and vascular remodeling following injury. Conversely, activation of PVAT beiging attenuates inflammation and pathological vascular remodeling. Single-cell RNA sequencing reveals that beige adipocytes abundantly express neuregulin 4 (Nrg4) which critically regulate alternative macrophage activation. Importantly, significant beiging is observed in the diseased aortic PVAT in patients with acute aortic dissection. Taken together, vascular injury induces the beiging of adjacent PVAT with macrophage accumulation, where NRG4 secreted from the beige PVAT facilitates alternative activation of macrophages, leading to the resolution of vascular inflammation. Our study demonstrates the pivotal roles of PVAT in vascular inflammation and remodeling and will open a new avenue for treating atherosclerosis.

Fig. 1. Endovascular injury-induced macrophage accumulation and inflammation in PVAT followed by upregulation of BAT markers.

a Gene expressions of immune cell markers in PVAT 24 h after vascular injury (sham, n = 4 for Cd8, 6 for others; injury, n = 9, two-tailed t tests with Holm-Sidak’s correction for multiple comparisons). b Haematoxylin and eosin (H&E), Elastica van Gieson (EVG) and immunohistochemical staining for F4/80 in the early (day 3) and late (day 14) phases after injury. Blue dashed lines indicate the external elastic lamina. Scale bars represent 100 μm (thick bars) and 50 μm (thin bars). Images are representative of three independent experiments. c Time course of F4/80⁺ cell accumulation in arteries (vessel) and outer tissues mainly composed of PVAT (n = 3, 4, 7, 7, 6 (from left to right) at each group, respectively, one-way analysis of variance (ANOVA) and Tukey–Kramer post-hoc test). d Gene expressions of inflammatory cytokine markers in PVAT 48 h after vascular injury (sham, n = 12, 14, 12, 12, 12, 8, 8, respectively; injury, n = 10, 12, 7, 7, 8, 8, 8, respectively, two-tailed t tests with Holm-Sidak’s correction for multiple comparisons. e Gene expressions of BAT (Ucp1 and Elovl3) and WAT [Rstn (Resistin) and Cfd (Adipsin)] markers in PVAT at 24 and 48 h after injury (sham, n = 6, 8, 8, 8, respectively; injury 24 h, n = 6, 6, 8, 8, respectively; injury 48 h, n = 11, 7, 6, 6, respectively, one-way ANOVA followed by Tukey–Kramer post-hoc test). f In situ hybridization showing Ucp1 mRNA expression 14 days after vascular injury in wild type mice. Scale bars represent 50 μm (thick bars) and 20 μm (thin bars). Images are representative of three independent experiments. g H&E and immunohistochemical staining for UCP1 in injured (3 and 14 days after injury) or sham-operated FAs and outer tissue. Scale bars represent 100 μm (thick bars) and 50 μm (thin bars). Images are representative of three independent experiments. h Time course of UCP1⁺ cells accumulation in PVAT (n = 3, 4, 7, 7, 6 at each group, respectively, one-way ANOVA followed by Tukey–Kramer post-hoc test). i Histograms of adipocyte area of sham-operated or injured PVAT (14 days after injury). Three images of each biological replicate were analyzed and combined to create the histogram. The size distribution between each group was compared using Kolmogorov-Smirnov test. Each bin was normalized to a percent of the total count for that individual tissue. Adipocytes of the bin size in the range of 20-500 μm² were included for the analysis. j Representative images of western blots for UCP1 in arteries (vessel) and outer tissues mainly composed of PVAT harvested 48 h after injury. Intra-scapular BAT was used as a positive control. GAPDH was used for internal control (n = 4 for each group, representative images are shown). k Immunohistochemical staining for F4/80 and UCP1 in outer tissue surrounding FAs 14 days after injury was performed in mice treated with either clodronate or vehicle. Scale bars represent 50 μm. Images are representative of three independent experiments. l F4/80⁺ and UCP1⁺ area in PVAT (vehicle, n = 9; clodronate, n = 4, unpaired two-tailed Student’s t test). Data represent mean ± SEM. Source data are provided as a Source Data file.

Fig. 2. Modification of PVAT beiging alters vascular inflammatory response and remodeling after injury.

a EVG staining in FAs 14 days after injury of Adipo^Cre+;Prdm16 mice and Adipo^Cre−;Prdm16 (control) mice. The ratio of intima to media area (intima/media) and the % intima–medial area in the area surrounded by the external elastic lamina (area of intima–media) at 14 days after vascular injury were analyzed (Adipo^Cre−;Prdm16, n = 10; Adipo^Cre+;Prdm16, n = 13, unpaired two-tailed Student’s t test). Scale bars represent 100 μm. b Gene expression of inflammatory cytokine markers in PVAT 14 days after vascular injury in Adipo^Cre+;Prdm16 mice and Adipo^Cre−;Prdm16 (control) mice (Adipo^Cre-;Prdm16 – sham, n = 6; Adipo^Cre+;Prdm16 – sham, n = 7; Adipo^Cre-;Prdm16 – injury, n = 6; Adipo^Cre+;Prdm16 – injury, n = 6, one-way ANOVA followed by Tukey–Kramer post-hoc test).　c, d EVG and immunohistochemical staining for F4/80, iNOS (classically activated) and CD206 (alternatively activated) in FAs 14 days after injury in wild type mice treated with pluronic gel containing siRNA [Prdm16 or scrambled, (c)] or β3AR agonist [CL316243 or vehicle, (d)] applied to the PVAT surrounding the FA. The ratio of intima to media area (intima/media) and the % intima–medial area in the area surrounded by the external elastic lamina (area of intima–media) at 14 days after vascular injury were analyzed ((c) scrambled siRNA, n = 12; Prdm16 siRNA, n = 11, unpaired two-tailed Student’s t test, (d) vehicle, n = 11; CL316243, n = 9, unpaired two-tailed Student’s t test). The positive area of immunostaining and ratio of iNOS to CD206-positive area in arteries (vessel) or surrounding tissues (PVAT) were analyzed (c scrambled siRNA, n = 16; Prdm16 siRNA, n = 15, unpaired two-tailed Student’s t test, d vehicle, n = 11; CL316243, n = 9, unpaired two-tailed Student’s t test). Scale bars represent 100 μm. Data represent mean ± SEM. Source data are provided as a Source Data file.

Fig. 3. Beige-differentiated PVAT shifts macrophage phenotypes to alternatively activated state.

a–d Cultured preadipocytes isolated from the PVAT stromal vascular fraction were differentiated into beige adipocytes, and media conditioned by the cells were added to RAW 264.7 macrophage cells (a, b) or bone marrow-derived monocytes (BMDMs) (c, d). After 48 h, mRNA was extracted from the RAW 264.7 cells or BMDMs and qRT-PCR was performed. The gene expression levels of macrophage phenotype markers [Cd86 (classically activated), Mrc1 (CD206, alternatively activated) and ratio of Cd86 to Mrc1, (a, c)] and inflammatory cytokines [Tnf, Serpine1, Ccl2, and Il1b, (b, d)] were analyzed. Control cells were treated with media cultured in the absence of adipocytes (a n = 6 biological replicates, representative data of three different culture lines are shown, two-tailed t tests with Holm-Sidak’s correction for multiple comparisons, b–d n = 4 biological replicates, representative data of three different culture lines are shown, two-tailed t tests with Holm-Sidak’s correction for multiple comparisons). e–f PVAT-preadipocytes were introduced by siRNA of Prdm16 or scrambled and stimulated with beige differentiation factors. Culture media conditioned by these cells were added to RAW 264.7 cells and mRNA was extracted 48 h after treatment. The results of the qRT-PCR analyses of the gene expression levels of macrophage phenotype markers (e) and inflammatory cytokines (f) are shown (n = 4, biological replicates, representative data of three different culture lines are shown, two-tailed t tests with Holm-Sidak’s correction for multiple comparisons). g Cell growth analysis of RAW 264.7 cells 72 h after treatment with culture media conditioned by PVAT-derived beige adipocytes. RAW 264.7 cells were pre-treated with IFNγ (20 ng/ml) and LPS (20 ng/ml) or IL4 (10 ng/ml) for 24 h and treated with media cultured in the absence of adipocytes (gray dots), culture media conditioned by adipocytes (red dots) and culture media conditioned by adipocytes introduced by Prdm16 siRNA (blue dots) (n = 4 biological replicates, representative data of three different culture lines are shown, one-way ANOVA followed by Tukey–Kramer post-hoc test). h EVG staining in the FA 14 days after injury in Ucp1^−/− and wild type (Ucp1^+/+) mice. Scale bars represent 100 μm. The ratio of intima to media area (intima/media) and % intima–medial area in the area surrounded by the external elastic lamina (area of intima–media) were analyzed (Ucp1^+/+, n = 8; Ucp1^−/−, n = 10, unpaired two-tailed Student’s t test). Data represent mean ± SEM. Source data are provided as a Source Data file.

Fig. 4. Identification of NRG4 as beige PVAT-derived secretory factor that controls macrophage phenotypes and growth.

a, b Clustering analysis of UMAP dimensionality reduction using scRNA-seq data in publicly available datasets of mouse iWAT treated with CL316243 or control (GSE 133486), dividing adipose tissue cells into 15 clusters which are color-coded by cell type (a) and by stimulation (CL316243 treatment [pink] or control [turquoise]) (b). c Violin plots of representative beige/BAT marker genes and Nrg4 in each cluster indicated in a. d Gene ontology analysis for the top 100 regulated genes in C7 showing that several biological processes annotated by the genes were changed significantly. Nrg4 was listed as a gene involved in receptor-binding. P-values are determined by Fisher’s Exact test. e qRT-PCR analysis showing Nrg4 gene expression in PVAT 24 and 48 h after injury (sham, n = 6; injury 24 h, n = 6; injury 48 h, n = 11, one-way ANOVA followed by Tukey–Kramer post-hoc test). f Time course of gene expressions in PVAT-derived preadipocytes stimulated by beige differentiation factors. Relative expression levels of beige/BAT markers and Nrg4 are shown. (n = 4 biological replicates, representative data of three different culture lines are shown, one-way ANOVA followed by Tukey–Kramer post-hoc test). g Gene expression of Nrg4 in PVAT 14 days after vascular injury in Adipo^Cre+;Prdm16 mice and Adipo^Cre−;Prdm16 (control) mice (Adipo^Cre−;Prdm16, n = 7; Adipo^Cre+;Prdm16, n = 6, unpaired two-tailed Student’s t test). h qRT-PCR analysis showing Nrg4 gene expression in PVAT-derived preadipocytes introduced by scrambled, Prdm16 and Nfia siRNAs followed by stimulated with beige differentiation factors for 6 days (n = 4, biological replicates, representative data of three different culture lines are shown, one-way ANOVA followed by Tukey–Kramer post-hoc test). i–l PVAT-derived preadipocytes were introduced by Nrg4 or scrambled siRNA and stimulated with beige differentiation factors. Culture media conditioned by these cells were then added to RAW 264.7 cells (i, j) or BMDMs (k, l) and mRNA was extracted 48 h after treatment. The results of qRT-PCR for the gene expression levels of macrophage phenotype markers (i, k) and inflammatory cytokines (j, l) are shown ((i, j) n = 4 biological replicates, representative data of three different culture lines are shown, two-tailed t tests with Holm-Sidak’s correction for multiple comparisons, (k, l) n = 3 biological replicates, representative data of three different culture lines are shown, two-tailed t tests with Holm-Sidak’s correction for multiple comparisons). m Cell growth analysis of RAW 264.7 cells 72 h after treatment with conditioned media from PVAT-derived beige adipocytes. RAW 264.7 cells were pre-treated with IFNγ and LPS or IL4 for 24 h and treated with media cultured in the absence of adipocytes (gray dots), media conditioned by adipocytes (red dots) and media conditioned by adipocytes introduced by Nrg4 siRNA (blue dots) (n = 4 biological replicates, representative data of three different culture lines are shown, one-way ANOVA followed by Tukey–Kramer post-hoc test). n EVG staining in FAs 14 days after injury in wild type mice treated with pluronic gel, containing siRNA against Nrg4 or scrambled, that was applied to the PVAT surrounding the FA. The ratio of intima to media area (intima/media) and the % intima–medial area in the region surrounded by the external elastic lamina (area of intima–media) at 14 days after vascular injury were analyzed (scrambled siRNA, n = 10; Nrg4 siRNA, n = 11, unpaired two-tailed Student’s t test). Scale bars represent 100 μm. Data represent mean ± SEM. Source data are provided as a Source Data file.

Fig. 5. PVAT beiging in human and murine aorta with acute dissection.

a Aorta tissue samples from patients with or without AAD were collected and H&E and immunohistochemical staining for UCP1 were performed. Scale bars represent 5 mm (thick bars) and 100 μm (thin bars). Images are representative of three independent experiments. b The % of UCP1⁺ areas in PVAT from patients with or without AAD were analyzed (control, n = 5; AAD, n = 6, two-sided Mann–Whitney U-test. Box plots denote the median, 25–75th percentiles, and minimum and maximum values). c Survival curves comparing vehicle-treated and CL316243-treated male AAD mice (n = 32 for each group, Kaplan–Meyer analysis, followed by log-rank test). d Representative flow plots derived from the PVAT of AAD mice treated with CL316243 or vehicle. e Percentage of classically activated (M1; defined as CD11c⁺, CD206⁻) and alternatively activated (M2; defined as CD11c⁻, CD206⁺) macrophages, and the ratio of M1 to M2 macrophages measured by flow cytometric analysis are shown (n = 3 for each group, unpaired two-tailed Student’s t test). Data represent the mean ± SEM. Source data are provided as a Source Data file.