Endothelial cell-derived stem cell factor promotes lipid accumulation through c-Kit-mediated increase of lipogenic enzymes in brown adipocytes

Abstract

Active thermogenesis in the brown adipose tissue (BAT) facilitating the utilization of lipids and glucose is critical for maintaining body temperature and reducing metabolic diseases, whereas inactive BAT accumulates lipids in brown adipocytes (BAs), leading to BAT whitening. Although cellular crosstalk between endothelial cells (ECs) and adipocytes is essential for the transport and utilization of fatty acid in BAs, the angiocrine roles of ECs mediating this crosstalk remain poorly understood. Using single-nucleus RNA sequencing and knock-out male mice, we demonstrate that stem cell factor (SCF) derived from ECs upregulates gene expressions and protein levels of the enzymes for de novo lipogenesis, and promotes lipid accumulation by activating c-Kit in BAs. In the early phase of lipid accumulation induced by denervation or thermoneutrality, transiently expressed c-Kit on BAs increases the protein levels of the lipogenic enzymes via PI3K and AKT signaling. EC-specific SCF deletion and BA-specific c-Kit deletion attenuate the induction of the lipogenic enzymes and suppress the enlargement of lipid droplets in BAs after denervation or thermoneutrality in male mice. These data provide insight into SCF/c-Kit signaling as a regulator that promotes lipid accumulation through the increase of lipogenic enzymes in BAT when thermogenesis is inhibited.

Fig. 1. snRNA-seq analysis reveals distinct ligand-receptor pairings between ECs and BAs in BAT of adult mice.

a Diagram depicting the procedures for snRNA-seq on the isolated nuclei of BAs, ECs, and SCs of the BAT in adult C57BL/6 J mice (N = 20). snRNA-seq was performed in nuclei isolated from a floating layer consisting of brown adipocytes (BAs) and fragmented vasculature (FV) after BAT digestion. b UMAP plot showing 5 main clusters, Ucp1^high and Ucp1^low BAs, white adipocytes (WAs), endothelial cells (ECs), and pericytes (PCs) of a total of 10,841 cells in BATs. c UMAP plots showing the expression of the distinct genes in each cluster of BAs, ECs, and PCs. d Violin plots showing the expressions of indicated genes in Ucp1^high and Ucp1^low BAs and WAs. e Dot plots showing indicated signaling pathways from source cells (outgoing communication) to target cells (incoming communication) according to ligand-receptor pairings among BAs, ECs, and PCs. The signaling pathways from ECs to BAs are highlighted by dark red dotted-lined boxes. Relative contribution degrees are present as different sizes of circles. f UMAP plot showing ligand-receptor pairing for Kit, Efnb, and Bmp signaling pathways in 5 main clusters.

Fig. 2. Denervation increases the number of c-Kit⁺ BAs and protein level of c-Kit in BAT.

a, b Diagram depicting the experimental procedure for the unilateral denervation (D-BAT) and contralateral sham-operation (intact BAT, I-BAT) in the interscapular BAT of 8-week-old c-Kit^iTR mice, tamoxifen administrations, and sampling 3 or 5 days later for the analyses. c, d Representative immunoblotting and comparisons of tyrosine hydroxylase (TH) in I-BAT and D-BAT at day 3 after unilateral denervation (I-BAT3 or D-BAT3). The same amount of protein loading in each lane is verified by immunoblotting of GAPDH. Each dot indicates a value from one mouse and n = 6 mice/group from two independent experiments. Vertical bars indicate mean ± SD. ***P < 0.001 versus I-BAT3 by two-tailed t-test. Protein sizes are indicated as kilodalton (kDa). e–g, Representative images and comparisons of populations of different sizes of lipid droplets in I-BAT and D-BAT and the number of c-Kit⁺ BAs in BAT from c-Kit^iTR mice at day 3 or 5 after unilateral denervation (I-BAT3, D-BAT3, I-BAT5, or D-BAT5). Scale bars, 20 µm. Each dot indicates a value from one mouse and n = 4 mice/group from two independent experiments. Vertical bars indicate mean ± SD. ***P < 0.001 versus I-BAT and *P < 0.05 versus D-BAT3 by one-way ANOVA test followed by Tukey’s post-hoc test (d), *P < 0.05, **P < 0.01 and ***P < 0.001 versus I-BAT by two-tailed t-test (e). h–j Unilateral denervation in the interscapular BAT of 8-week-old C57BL/6 J mice was performed and BATs were isolated 3 or 5 days later for the analyses (I-BAT3, D-BAT3, I-BAT5, or D-BAT5). Representative immunoblotting and comparisons of relative density (RD) of c-Kit and UCP1 in BAT of sham-operated (SO), I-BAT, and D-BAT. The same amount of protein loading in each lane is verified by immunoblotting of tubulin. Each dot indicates a value from one mouse and n = 6 mice/group from two independent experiments. Vertical bars indicate mean ± SD. **P < 0.01 and ***P < 0.001 versus I-BAT3 or I-BAT5 by one-way ANOVA test followed by Tukey’s post-hoc test. NS, not significant. Protein sizes are indicated as kilodalton (kDa). k Comparisons of SCF concentration in the BAT of sham-operated (SO), I-BAT, and D-BAT at day 3 or 4 after unilateral denervation (I-BAT3, D-BAT3, I-BAT4, or D-BAT4). Each dot indicates a value from one mouse and n = 6 mice/group from two independent experiments. Vertical bars indicate mean ± SD. **P < 0.01 versus I-BAT3 or I-BAT4 by one-way ANOVA test followed by Tukey’s post-hoc test. NS not significant.

Fig. 3. SCF increases DNL enzymes by activating c-Kit signaling pathway in cultured BAs.

a, b Representative immunoblotting and comparisons of c-Kit, pAKT1 (Ser473), AKT1, pAKT2 (Ser474), AKT2, pERK1/2 (Thr202/Tyr204), ERK1/2, ACL, ACC, FASN, and SCD1 before and 10 min (10 m), 6 h, and 12 h after either bovine serum albumin (BSA, 100 ng/ml) or SCF (100 ng/ml) treatment in Control (transfected with empty lentiviral plasmid) or c-Kit over-expressed (transfected with c-Kit cDNA inserted lentiviral plasmid) cultured BAs. The cultured BAs were pre-incubated in high glucose DMEM containing 1% FBS for 12 h before and after the treatment. The same amount of protein loading in each lane is verified by immunoblotting of tubulin. Dots and bars indicate mean ± SD from n = 3/group from two independent experiments. *P < 0.05, **P < 0.01, and ***P < 0.001 versus BSA by two-tailed t-test. Protein sizes are indicated as kilodalton (kDa). c, d Representative immunoblotting and comparison of effect of indicated signaling inhibitor on the SCF-induced increased protein level of SCD1 in the c-Kit over-expressed and differentiated cultured BAs. The cultured BAs were pre-incubated in high glucose DMEM containing 1% FBS for 12 h before and after the treatment. For the inhibitor experiments, either Dynasore (20 µM), SU5402 (1 µM), wortmanin (30 nM), AKT1/2 inhibitor (25 nM), PD 0325901 (1 µM), ruxolitinib (10 µm) or Y27632 (10 mM) was treated to cultured cells for 1 hr, and then SCF (100 ng/ml) or VEGF (100 ng/ml) was added and incubated for 12 h. Each dot indicates a value from one experiment and n = 3/group from two independent experiments. Vertical bars indicate mean ± SD. *P < 0.05 versus SCF by one-way ANOVA test followed by Tukey’s post-hoc test. Protein sizes are indicated as kilodalton (kDa).

Fig. 4. BA-specific deletion of c-Kit inhibits the denervation-induced lipid accumulation and reduces SCD1 in the BAs of BAT.

a Diagram of the experimental procedure for the unilateral denervation of BAT in 8-week-old control (Con) and c-Kit^∆BA mice, and harvest BAT (I-BAT and D-BAT) 3 or 5 days later for the analyses. b, c Representative immunoblotting and comparisons of c-Kit, ACL, ACC, FASN, and SCD1 in the I-BAT and D-BAT 3 days after the unilateral denervation (I-BAT3 or D-BAT3). The same amount of protein loading in each lane is verified by immunoblotting of tubulin. Each dot indicates a value from one mouse and n = 6 mice/group from two independent experiments. Horizontal and vertical bars indicate mean ± SD. *,#P < 0.05 versus Con I-BAT3 or c-Kit^∆BA I-BAT3 by one-way ANOVA test followed by Tukey’s post-hoc test. Protein sizes are indicated as kilodalton (kDa). d, e Representative images and comparisons of the populations of different sizes of lipid droplets in the CD147⁺ BAs of BAT among the indicated groups. White arrowheads indicate the lipid droplets whose size is larger than 51 µm² in CD147⁺ BA. Scale bars, 20 µm. Each dot indicates the population (%) of the total of 1200–2400 droplets (100%) from three portions of BAT in one mouse and n = 5 mice/group from two independent experiments. Vertical bars indicate mean ± SD. *P < 0.05 by two-way ANOVA test followed by Sidak test. NS not significant.

Fig. 5. Thermoneutrality increases the number of c-Kit⁺ BAs and expression of c-Kit in BAT.

a Diagram depicting the experimental procedure for the exposure of 8-week-old c-Kit^iTR mice to RT (25°C) or thermoneutrality (TN, 30°C) for 3 or 5 days with tamoxifen administrations. b–d Representative of the number of c-Kit⁺ BAs, and the populations of different sizes of lipid droplets in BATs from the c-Kit^iTR mice at RT (25°C) or thermoneutrality (TN, 30°C) for 3 or 5 days (RT3, TN3, or TN5). Scale bars, 20 µm. Each dot indicates a value from one mouse and n = 4 mice/group from two independent experiments. Vertical bars indicate mean ± SD. ***P < 0.001 versus RT by one-way ANOVA test followed by Tukey’s post-hoc test (c). Each dot indicates the population (%) of the total of 1200–2400 droplets (100%) from three portions of BAT in one mouse and n = 4 mice/group from two independent experiments. Vertical bars indicate mean ± SD. *P < 0.05 and ***P < 0.001 versus R3 by one-way ANOVA test followed by Tukey’s post-hoc test (d). NS, not significant. e Diagram depicting the experimental procedure for the exposure of 8-week-old C57BL/6 J mice to RT (25 °C) or thermoneutrality (TN, 30°C) for 3 or 5 days (R5, T3, or T5). f–k Representative immunoblotting and comparisons of c-Kit and UCP1 in BATs from C57BL/6 J mice at R5, T3, and T5. Each dot indicates a value from one mouse and n = 6 mice/group from two independent experiments. Vertical bars indicate mean ± SD. *P < 0.05 versus R5 by one-way ANOVA test followed by Tukey’s post-hoc test. NS, not significant. Protein sizes are indicated as kilodalton (kDa).

Fig. 6. BA-specific c-Kit deletion reduces SCD1 and attenuates the growth of lipid droplets in BAs at thermoneutrality.

a Diagram depicting experimental procedure for the exposure of 8-week-old control (Con) and c-Kit^∆BA mice to thermoneutrality (TN, 30 °C) for 3 or 5 days, and sampling and analyses. b–f Representative immunoblotting and comparisons of ACL, ACC, FASN, and SCD1 in BAT from Con and c-Kit^∆BA mice at the thermoneutrality for 3 or 5 days (TN3 or TN5). The same amount of protein loading in each lane is verified by immunoblotting of tubulin1 for ACL and ACC, tubulin2 for FASN, and tubulin3 for SCD1. Each dot indicates a value from one mouse and n = 6 mice/group from two independent experiments. Vertical bars indicate mean ± SD. *P < 0.05, **P < 0.01 versus Con by two-tailed t-test. Protein sizes are indicated as kilodalton (kDa). g Comparisons of the proportions of palmitate (16:0), palmitoleate (16:1), stearate (C18:0), and oleate (C18:1) in total FFAs of BATs between Con versus c-Kit^∆BA mice at the thermoneutrality for 3 days. SCD1 desaturation indices calculated by the ratio of 16:1/16:0 and 18:1/18:0 are shown. Each dot indicates a value from one mouse and n = 9–10 mice/group. Vertical bars indicate mean ± SD. *P < 0.05 versus Con by two-tailed t-test. NS, not significant. h, i Representative images and comparisons of the populations of different sizes of lipid droplets in BATs from Con and c-Kit^∆BA mice at the thermoneutrality for 5 days (TN5). Scale bars, 20 µm. Each dot indicates the population (%) of total 1200–2400 droplets (100%) from three portions of BAT in one mouse and n = 5 mice/group from two independent experiments. Vertical bars indicate mean ± SD. **P < 0.01 versus Con by two-way ANOVA test followed by Sidak test. NS not significant.

Fig. 7. EC-specific SCF deletion reduces SCD1 and attenuates the growth of lipid droplets in BAs at thermoneutrality.

a Diagram depicting an experimental procedure for the exposure of 8-week-old control (Con) and SCF^i∆EC mice to thermoneutrality (TN, 30 °C) for 3 or 5 days, and sampling and analyses. b,c, Representative immunoblotting and comparison of c-Kit in BAT in Con and SCF^i∆EC mice at RT or thermoneutrality for 3 or 5 days (RT5, TN3, or TN5). The same amount of protein loading in each lane is verified by immunoblotting of tubulin. Each dot indicates a value from one mouse and n = 5 mice for SCF^iΔEC group in TN3 and n = 6 mice for other groups from two independent experiments. Vertical bars indicate mean ± SD. *P < 0.05, ***P < 0.001 versus RT by one-way ANOVA test followed by Tukey’s post-hoc test. Protein sizes are indicated as kilodalton (kDa). d–h Representative immunoblotting and comparisons of ACL, ACC, FASN, and SCD1 in BAT of Con and SCF^i∆EC mice at thermoneutrality for 3 or 5 days (TN3 or TN5). The same amount of protein loading in each lane is verified by immunoblotting of tubulin. Each dot indicates a value from one mouse and n = 6 mice/group from two independent experiments. Vertical bars indicate mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001 versus Con by two-tailed t-test. NS, not significant. Protein sizes are indicated as kilodalton (kDa). i Comparisons of the proportions of palmitate (16:0), palmitoleate (16:1), stearate (C18:0), and oleate (C18:1) in total FFAs of BATs between Con versus SCF^i∆EC mice at the thermoneutrality for 3 days. SCD1 desaturation indices calculated by the ratio of 16:1/16:0 and 18:1/18:0 are shown. Each dot indicates a value from one mouse and n = 8 mice/group. Vertical bars indicate mean ± SD. *P < 0.05 versus Con by two-tailed t-test. NS, not significant. j, k Representative images and comparison of the populations of different sizes of lipid droplets in BAT from Con and SCF^i∆EC mice at the thermoneutrality for 5 days (TN5). Scale bars, 20 µm. Each dot indicates the population (%) of a total of 1200–2400 droplets (100%) from three portions of BAT in one mouse and n = 5 mice/group from two independent experiments. Vertical bars indicate mean ± SD. *P < 0.05 versus Con by two-way ANOVA test followed by Sidak test. NS not significant.

Fig. 8. Schematic diagram depicting the promoting role of SCF/c-Kit signaling in lipid accumulation in BAs.

Activation of c-Kit in BAs by SCF derived from ECs increases protein levels of DNL enzymes, which enhances the growth of lipid droplets. TN thermoneutrality, TG triglycerides, LD lipid droplet.