Asc-1 regulates white versus beige adipocyte fate in a subcutaneous stromal cell population

Abstract

Adipose tissue expansion, as seen in obesity, is often metabolically detrimental causing insulin resistance and the metabolic syndrome. However, white adipose tissue expansion at early ages is essential to establish a functional metabolism. To understand the differences between adolescent and adult adipose tissue expansion, we studied the cellular composition of the stromal vascular fraction of subcutaneous adipose tissue of two and eight weeks old mice using single cell RNA sequencing. We identified a subset of adolescent preadipocytes expressing the mature white adipocyte marker Asc-1 that showed a low ability to differentiate into beige adipocytes compared to Asc-1 negative cells in vitro. Loss of Asc-1 in subcutaneous preadipocytes resulted in spontaneous differentiation of beige adipocytes in vitro and in vivo. Mechanistically, this was mediated by a function of the amino acid transporter ASC-1 specifically in proliferating preadipocytes involving the intracellular accumulation of the ASC-1 cargo D-serine.

Fig. 1. Single-cell sequencing reveals major differences between adolescent and adult preadipocytes.

Single-cell sequencing of cells of the subcutaneous stromal vascular fraction from 2- to 8-week-old male C57Bl/6 mice. Cells were pooled from the SCF of seven (2 weeks old) and five (8 weeks old) mice. T-stochastic neighborhood embedding (t-SNE) of union of adolescent and adult cells of the SCF SVF (n = 9344) with mouse age (a) and Louvain group (b) superimposed. The circle indicates clusters containing preadipocytes. c Log1p expression [log1p = log(X + 1), where X is the expression of gene of interest in particular cell. Note that X, the mRNA counts from the alignment, are integer numbers] of different cell type markers in the 11 louvain clusters by cell (n₀ = 1733, n₁ = 1379, n₂ = 1222, n₃ = 948, n₄ = 892, n₅ = 816, n₆ = 799, n₇ = 518, n₈ = 514, n₉ = 421, n₁₀ = 102). Leukocytes (Ptprc), T cells (CD3d, CD3g), NK cells (Nkg7, Klrd1), myeloid cells (Fcgr3, S100a8), macrophages (Adgre1, Lyz2), dendritic cells (CD74, CD83), B cells (CD19), megakaryocytes (Ppbp), erythrocytes (Gypa), and preadipocytes (Pdgfra, Fbn1, Col4a1, CD34, CD24a, Dlk1, Asc-1). d t-SNE of union of adolescent and adult cells of the SCF SVF (n = 9344) with log1p Pdgfra and Asc-1 expression superimposed. e t-SNE of preadipocytes from adolescent and adult SCF (n = 3,062) colored by age. f Volcano plot of comparison between adolescent and adult preadipocytes with differential expression (false discovery rate-corrected P value <0.01, Welch’s t test) indicated by log1p expression. Each dot is a gene. Black: not differentially expressed, yellow: differentially expressed, red: Asc-1. g t-SNE of preadipocytes from adolescent and adult SCF (n = 3062) colored by Asc-1 expression.

Fig. 2. Loss of Asc-1 function induces beiging in white adipocytes.

a Immunofluorescence staining of ASC-1 (green), F-Actin (gray), and lipids (red) in SCF of 2 week old Asc-1 wt and ko mice (n = 1). Size bar 100 µm. b Correlation between Asc-1 and Ppary expression in 52 immortalized preadipocyte clones derived from adult subcutaneous SVF. c Relative Ucp1 expression before (day 0) or after 8 days of differentiation of MACS-sorted ASC-1⁺ and ASC-1⁻ preadipocytes with either insulin alone, the normal differentiation protocol (Adipo) or the differentiation protocol plus 1 µM rosiglitazone (n = 4). Asc-1 knockdown (shAsc-1) and control (shScr) immortalized subcutaneous preadipocytes were grown to 100% confluence, and RNA or protein was taken from the day of differentiation start (preadipocyte) and after 8 days of differentiation with (Rosi) or without rosiglitazone (adipocyte). d Relative gene expression of Asc-1 (n = 6, only shAsc-1 preadipocyte n = 5), Fabp4 (n = 6, only shScr adipocyte n = 5), Ucp1 (n = 10, only shScr preadipocyte and shAsc-1 Rosi n = 9), and Tfam (n = 6). e Western blot of UCP1 (33 kDa), PPARy (54 kDa), and β-actin (42 kDa). f Basal respiration (n = 20 shScr and 24 shAsc-1 technical replicates) and proton leakage (n = 19 shScr and 24 shAsc-1 technical replicates) calculated from the OCR measured by a Seahorse flux analyzer after differentiation with rosiglitazone, normalized to DNA content (n = 3 biologicalreplicates). g Balb/c nude mice were injected with shAsc-1 and shScr preadipocytes over the sternum. Histology (n = 1; size bar 20 µm) and h–i immunofluorescence staining of engrafted tissue 6 weeks post transplantation. Green: UCP1 (n = 2), red: lipids, blue: DAPI, and gray: PERILIPIN-1 (n = 1 shScr-2 shAsc-1). Size bar 100 µm. Statistics were calculated using ordinary two-way ANOVA with Tukey’s multiple comparison post hoc test or two-tailed unpaired Students t test (f). Data are shown as mean ± SEM.

Fig. 3. Overexpression of Asc-1 in brown (pre-)adipocytes suppresses thermogenic gene expression.

a Asc-1-HA-2a-4F2HC or EGFP overexpressing brown preadipocytes were grown to 100% confluence, and RNA or protein was collected from the day of differentiation start (preadipocyte), and after 8 days of differentiation with (Rosi) or without rosiglitazone (adipocyte) and relative gene expression of Asc-1 measured (n = 3). b Western blot of ASC-1-HA (45 kDa) and β-actin (42 kDa; n = 2). c Relative gene expression of Ppary (n = 3), Ucp1 (n = 3), Pgc1α (n = 3), and Prdm16 (n = 3). d Quantification of PPARy1 and 2 and UCP1 (33 kDa) protein levels relative to β-actin (42 kDa; n = 6, only preadipocyte in Ppary blot of Rosi n = 4). The same preadipocyte samples were used as controls for the different differentiation conditions. Statistics were calculated using two-way ANOVA with Tukey’s or Sidak’s (for a, Pgc1α in c, and Ucp1 and Ppary1–2 for adipocyte in d) or two-way ANOVA mixed-effects analysis (Ppary1–2 for Rosi in d) multiple comparison post hoc test. Data are shown as mean ± SEM.

Fig. 4. Asc-1 regulates white versus beige adipocyte lineage commitment in a subset of proliferating precursors.

a Ppary, Pgc1α, and Prdm16 gene expression of subconfluent proliferating shAsc-1 and shScr preadipocytes (n = 4). b Preadipocytes were grown to 100% confluence and differentiation was induced. RNA was taken from the day of differentiation start (preadipocyte d0) and after 8 days (preadipocyte d8), as well as after 8 days of insulin treatment (insulin), insulin and rosiglitazone treatment (insulin + rosi), or differentiation (adipocyte). Relative gene expression of Ppary (n = 3), Ucp1 (n = 6; for shScr insulin n = 3), Pgc1α (n = 6; for shScr insulin n = 4), and Prdm16 (n = 6; for shScr insulin n = 4). c Western blot of UCP1 (33 kDa), PPARy (54 kDa), and β-actin (42 kDa; n = 1). d shScr preadipoctyes were plated in presence of 10 µM Asc-1 inhibitor (BMS-466442) or DMSO as control. After 2 days, when cells reached 100% confluency, differentiation was induced with Asc-1 inhibitor or DMSO. Relative gene expression of Ucp1 (n = 3; one-way ANOVA performed with log transformed data), Ppary (n = 3), Pat2 (n = 3), Prdm16 (n = 3), and Pgc1α (n = 3). e Intracellular D-serine levels of subconfluent and confluent preadipocytes (n = 2). f Srr was knocked down using DsiSrr in proliferating shAsc-1 preadipocytes and differentiated upon confluency. Ucp1 and Ppary mRNA levels of shAsc-1 cells at day of induction (preadip) or cells differentiated following the regular differentiation protocol, including rosiglitazone (adip + Rosi). A scrambled DsiRNA was used as negative Control (Control) for knockdown (n = 4). Statistics were calculated using two-tailed t test (a) or ordinary one-way or two-way (f) ANOVA with Tukey’s or Sidak’s multiple comparison post hoc test. Data are shown as mean ± SEM.