Phosphatase protector alpha4 (α4) is involved in adipocyte maintenance and mitochondrial homeostasis through regulation of insulin signaling

Abstract

Insulin signaling is mediated via a network of protein phosphorylation. Dysregulation of this network is central to obesity, type 2 diabetes and metabolic syndrome. Here we investigate the role of phosphatase binding protein Alpha4 (α4) that is essential for the serine/threonine protein phosphatase 2A (PP2A) in insulin action/resistance in adipocytes. Unexpectedly, adipocyte-specific inactivation of α4 impairs insulin-induced Akt-mediated serine/threonine phosphorylation despite a decrease in the protein phosphatase 2A (PP2A) levels. Interestingly, loss of α4 also reduces insulin-induced insulin receptor tyrosine phosphorylation. This occurs through decreased association of α4 with Y-box protein 1, resulting in the enhancement of the tyrosine phosphatase protein tyrosine phosphatase 1B (PTP1B) expression. Moreover, adipocyte-specific knockout of α4 in male mice results in impaired adipogenesis and altered mitochondrial oxidation leading to increased inflammation, systemic insulin resistance, hepatosteatosis, islet hyperplasia, and impaired thermogenesis. Thus, the α4 /Y-box protein 1(YBX1)-mediated pathway of insulin receptor signaling is involved in maintaining insulin sensitivity, normal adipose tissue homeostasis and systemic metabolism.

Fig. 1. Loss of α4 in brown preadipocytes impaired insulin signaling and adipogenesis.

a Abundance of α4 mRNA in mouse brown preadipocyte cells during the differentiation process before or on days 2, 4, and 6 after the induction of differentiation. Box plots are defined in terms of minima and maxima by whiskers, and the center and bounds of box by quartiles (one-way ANOVA post hoc Bonferroni test, n = 6 technical replicates per group, *p = 0.01, ***p < 0.0001). b Immunoblotting analysis of α4 in lysates from brown preadipocyte cells during the differentiation process before or 2 and 6 days after induction of the differentiation. (n = 4). c Immunoblotting for α4 and Flag in lysates from Control mouse brown preadipocytes or cells overexpressed 3XFlag-α4 (α4-OE). Densitometric analysis of α4 and Flag in mouse brown preadipocytes. Data are mean ± SEM of n = 5 (Two-tailed Student’s t-test, ***p < 0.0001). d Adipocyte markers in differentiated brown adipocytes from Control (n = 8) and 3XFlag-α4 transfected (α4-OE) cells (n = 8). Data are mean ± SEM (Two-tailed Student’s t-test, Leptin: *p = 0.02, UCP1: ***p = 0.0004). e Mitochondrial oxidative phosphorylation activity in mouse differentiated brown adipocytes from Control (n = 9) and α4-OE (n = 10) cells. Data are mean ± SEM (Two-tailed Student’s t-test, *p < 0.05). f Quantitation of basal respiration (*p = 0.02), maximal respiration capacity, ATP production and Spare Respiratory Capacity (*p = 0.01). Data are presented as mean ± SEM (Two-tailed Student’s t-test, Control (n = 9), α4-OE (n = 10)). g Levels of adipocyte markers in differentiated brown adipocytes from Control cells and cells subjected to knockdown of α4 by shRNA (α4 KD) (n = 8 biologically independent cell clones per group). Data are mean ± SEM (Two-tailed Student’s t-test, Adiponectin: **p = 0.003, PPARγ: *p = 0.02, Glut4: *p = 0.003, Adrb3: **p = 0.07, AP2: ***p = 0.0009, UCP1: **p = 0.006). h Immunoblotting analysis results showing insulin-dependent signaling molecules for IR, Akt, ERK, and S6 phosphorylation in lysates from α4 KD-treated or Control adipocytes after 100 nM insulin stimulation for the indicated durations. i–l Relative changes in protein phosphorylation based on the densitometric immunoblotting analysis (Supplementary Fig. 1h) of cell lysates of adipocytes from α4KD and Control for 5 min using 100 nM insulin. Data are presented as mean ± SEM (One-way ANOVA post hoc Bonferroni test; three clones were used in three independent experiments, p-IR: ***p = 0.0007, p-Akt (S473): **p = 0.001, p-S6 (S235/S236): ***p = 0.0001). Source data are provided as a Source Data file.

Fig. 2. α4 regulates IR Tyr phosphorylation via YBX1 and PTP1B.

a Schematic diagram showing the regulation of insulin signaling in brown preadipocytes via α4. Specifically, α4 binds to the catalytic subunit of protein phosphatase 2A (PP2Ac) and is known to regulate Ser/Thr phosphorylation. For the mass spectroscopic proteomic analysis. we compared the Control (n = 1) and α4 overexpressing (n = 2, biological replicates) samples. b α4-complex containing YBX1. Flag-tagged α4 immunoprecipitation showed the interaction between YBX1 with α4 as well as that between PP2Ac and α4 in brown preadipocytes (n = 3 biologically independent cell clones/group). c Western blotting of phosphorylation of insulin signaling molecules in Control and α4KO brown preadipocytes following indicated concentrations of insulin stimulation for 10 min. d Effect of α4 on insulin signaling-based densitometric immunoblotting analysis with antibodies to α4 and phosphorylated IR, Akt, YBX1, as well as PTP1B in lysates of α4 Knockout brown preadipocytes (α4 KO) before and 10 min after insulin stimulation (see c). Data are presented as mean ± SEM (one-way ANOVA post hoc Bonferroni test, n = 3, each cell was generated from biologically independent animals from three independent experiments, Statistical significance is shown as p < 0.05 (*), p < 0.01 (**), and p < 0.001 (***)). e Luciferase reporter assay for PTP1B transcription was performed as described by Fukada and Tong. Cells were transfected with the expression vectors of α4 and/or YBX1 in the presence of Ptp1b-Luc reporter or mock plasmid DNA in HEK293cells. Data are presented as mean ± SEM (One-way ANOVA post hoc Bonferroni test, n = 8 technical replicates per group, ***p < 0.0001). f Model of the regulation of α4 in insulin-activated IR phosphorylation. Insulin-bound IR phosphorylates itself and IRS1/2, and activates the PI3K-AKT and S6K pathways. α4 stabilizes PP2A and dephosphorylates mTORC2 with Rictor. α4 also binds to Y-box protein 1 (YBX1), which functions as a transcription factor for Tyr phosphatase PTP1B. α4 blocks YBX1 nuclear translocation and inhibits PTP1B expression, thus in feedback regulation, promoting the enhanced signaling through the insulin receptor. Source data are provided as a Source Data file.

Fig. 3. Role of α4 in the regulation of lipid metabolism in mature adipocytes.

a Tissue weights (g) of iWAT, eWAT, and BAT from Control (n = 6) and Ai-α4KO (n = 5) male mice on day 3 after treatment with tamoxifen (Two-tailed Student’s t-test). b HE-stained sections of iWAT, eWAT, and BAT from Control and Ai-α4KO mice on day 3. Scale bars = 100 μm. c Diameter distribution of isolated iWAT and eWAT adipocytes in Control and and Ai-α4KO at day 3. Data are mean ± SEM (two-tailed Student t-test, iWAT; Control (n = 8), Ai-α4KO (n = 7), eWAT; (n = 9/group). d Immunoblotting results showing the insulin signaling of phosphorylated IR, Akt, and PTP1B in interscapular BAT of 12-wk-old fasted Control and Ai-α4KO mice 10 min after i.v. insulin stimulation (5 IU per mouse) or control saline (n = 4). e Densitometric comparison of the phosphorylation signals shown as fold increases in relation to the control without insulin stimulation (1.0). Data are presented as mean ± SEM (one-way ANOVA post hoc Bonferroni test: n = 4, *p < 0.05, **p < 0.01 and ***p < 0.001, biologically independent animals per group from one experiment). f Staining of iWAT and BAT sections from Control and Ai-α4KO mice on day 3 for F4/80 or cleaved caspase-3. The experiments were repeated independently four times. Scale bars = 50 μm. g HE-stained sections of iWAT, eWAT, and BAT from Control and Ai-α4KO mice on days 6. Scale bars = 100 μm. h Diameter distribution of isolated iWAT from Control and Ai-α4KO mice on day 6. Data are presented as mean ± SEM (Two-tailed Student’s t-test: n = 8/group). i Lipolysis assessed by FFA release from iWAT of Control and Ai-α4KO mice on day 6. Samples were incubated ex vivo in the presence or absence of 10 mM isoproterenol, and FFA release into the medium was quantified. (Two-tailed Student’s t-test, **p = 0.002: n = 6/group). j Heatmap showing Z-scores of lipid species in iWAT and BAT from Control and Ai-α4KO mice (n = 5/group) at 1 wk. Green or red represents a decrease and increase, respectively. Comparison of lipid classes (k), fatty acids with different chain lengths (l) and Ceramide species with the indicated chain lengths m in BAT from Ai-α4KO mice with respect to the 1.0 level corresponding to the Control (n = 5/group). Data are presented as mean ± SEM (Two-tailed Student’s t-test). Statistical significance is shown as p < 0.05 (*), p < 0.01 (**), and p < 0.001 (***). Source data are provided as a Source Data file.

Fig. 4. Role of α4 in the regulation of gene expression and mitochondrial biogenesis in mature adipocytes.

a Principal component analysis plots of the transcriptome profiles of Control iWAT (Blue), Ai-α4KO iWAT (purple), Control BAT (Gold), and Ai-α4KO BAT (Green). b Venn diagram showing the numbers of significantly regulated genes in Ai-α4KO iWAT and BAT compared with the Control iWAT and BAT (FDR < 0.05, |FC| > 2.0). Heatmap of top 50 (up and down: top 25 each) differentially regulated genes between Control and Ai-α4KO in BAT (c) and iWAT (d). e Top directionally downregulated KEGG pathways between Control and Ai-α4KO in BAT (upper) and iWAT (bottom). f Genes involved in oxidative phosphorylation (OXPHOS) in BAT (left) and iWAT (right) listed in the heatmap. The color intensities indicate the Z-score of each gene. g UCP1 expression in iWAT and BAT from Control and Ai-α4KO mice at 1 wk (n = 3 biologically independent animals per group). Scale bars = 100 μm. h Representative electron microscopic images of mitochondria in white and brown adipocytes from Control and Ai-α4KO mice 1 wk after tamoxifen administration. Scale bars = 500 nm. i Quantification of the average mitochondrial size from Control and Ai-α4KO brown adipocytes 1 wk after tamoxifen administration. Data are mean ± SEM (Two-tailed Student’s t-test: *p = 0.01; n = 3). j Oxygen consumption (VO2) corresponding to Control (n = 5) and Ai-α4KO (n = 6) mice housed in metabolic cages 1 wk after tamoxifen administration. The dark phase represents the 12-h period of a day during which the lights were turned off. Data are mean ± SEM (Two-tailed Student’s t-test, Dark: **p = 0.003, Light: **p = 0.002). k Interscapular temperature of Control (n = 6) and Ai-α4KO (n = 5) mice on day 10 during a 3-h exposure to an environment at 4 °C. Data are mean ± SEM (Two-tailed Student’s t-test: *p < 0.05; **p < 0.01; ***p < 0.001). l Thermal images showing the comparison of the surface temperature over interscapular BAT from Control and Ai-α4KO mice after 2-h exposure to a temperature of 4 °C on days 10 and 90. Source data are provided as a Source Data file.

Fig. 5. Adipose tissue inflammation in Ai-α4KO mice.

a Top upregulated KEGG pathways between Control and Ai-α4KO in BAT (left) and iWAT (right). b Heatmap of genes involved in the cytokine-cytokine receptor interaction, with the color intensities indicating the Z-score of each sample. Expression levels of inflammatory cytokines, chemokines (c), and macrophage (d) markers in BAT from Control and Ai-α4KO mice on day 9 (n = 12/group). (Two-tailed Student’s t-test, **p < 0.01; ***p < 0.001). e Staining of iWAT and BAT sections from Control and Ai-α4KO mice on day 9 for F4/80. Scale bars =50 μm. f Flow cytometric analysis results showing infiltrated immunocytes with CD11c and CD206 on day 9. Representative data corresponding to iWAT are indicated as the FACS profiles (left). The ratio of M1 (CD11c⁺) to M2 (CD11c⁺) macrophages infiltrated in iWAT and BAT are shown in the right graph (n = 8/group). (Two-tailed Student’s t-test, *p = 0.03, ***p < 0.0001). g TUNEL staining results corresponding to iWAT and BAT sections from Control and Ai-α4KO mice on day 9. Scale bars = 50 μm. The arrow shows TUNEL-positive cells. h Immune-stained iWAT and BAT sections from Control and Ai-α4KO mice for the identification of cleaved caspase-3 on day 9. Scale bars = 50 μm. The arrow shows cleaved caspase-3 cells. i Expression levels of genes related to NF-κB signaling pathways and inflammasome components measured via real-time qPCR using BAT from Control and Ai-α4KO mice on day 9 (n = 12/group). Data are presented as mean ± SEM (Two-tailed Student’s t-test, ***p < 0.0001). Source data are provided as a Source Data file.

Fig. 6. Adipocyte metabolic dynamics in Ai-α4KO mice.

a Tissue weights (g) of iWAT, eWAT, and BAT from Control and Ai-α4KO on days 9 and 17 after treatment with tamoxifen (Two-tailed Student’s t-test: n = 8/group, *p = 0.02, **p = 0.001, ***p < 0.0001). b Representative pictures of the adipose tissues of Control and Ai-α4KO on day 17. Scale bars = 1 cm. c mRNA expression levels in iWAT from Control (n = 11) and Ai-α4KO (n = 12) mice on day 17. Box plots are defined in terms of minima and maxima by whiskers, and the center and bounds of box by quartiles (Two-tailed Student’s t-test, *p < 0.05; **p < 0.01; ***p < 0.001). d mRNA abundance in BAT (n = 12/group) on day 17. Box plots are defined in terms of minima and maxima by whiskers, and the center and bounds of box by quartiles (Two-tailed Student’s t-test, *p < 0.05; **p < 0.01; ***p < 0.001). e Results of HE staining of iWAT sections on days 9, 17, 30, and 74. The arrow shows crown-like structures. Scale bars = 100 μm. f Results of HE staining of BAT sections from Control and Ai-α4KO on days 9, 17, 30, and 74. The arrow shows crown-like structures. Experiments in e–f were repeated in at least three independent experiments. Scale bars = 50 μm. g UCP1 expression in BAT from Control and Ai-α4KO on days 74. The experiments were repeated independently three times. Scale bars = 50 μm. h mTmG lineage tracing of adipocytes in iWAT from Control and Ai-α4KO mice carrying a Rosa-mTmG transgene as shown in Supplementary Fig. 6a before and at wks 1, 2, 4, 8, and 12 after tamoxifen treatment. Scale bars = 100 μm. (i) mTmG lineage tracing of adipocytes in eWAT. j mTmG lineage tracing of adipocytes in BAT. Experiments in h–j were repeated in at least three independent experiments. k Interscapular temperature in male Control and Ai-α4KO at 30-min intervals during the 3-h exposure of mice 12 wks after tamoxifen administration to an environment at 4 °C (Two-tailed Student’s t-test, **p = 0.005: n = 5/group). l Comparison of rectal and interscapular temperature drop in Control and Ai-α4KO after 3 h of the 4 °C challenge on day 10 (Control (n = 6) and Ai-α4KO (n = 5)) with 12 wks after tamoxifen administration (n = 5/group) mice (One-way ANOVA post hoc Bonferroni test, ****p < 0.001). Data are represented as mean ± SEM. Source data are provided as a Source Data file.

Fig. 7. Vulnerability of continued α4-deficient adipose tissues to the development of diabetes and progressive NAFLD owing to a HFD.

a Body weights of 5–16 wk-old male Control and Aα4KO mice fed a CD (Two-tailed Student’s t-test, *p < 0.05; **p < 0.01; ***p < 0.001: n = 9/group). b Representative Micro-CT scan images displaying body fat distribution in Control and Aα4KO male mice at 3 months old. The white, yellow, and green arrows indicate the location of interscapular BAT, subcutaneous WAT, and visceral WAT, respectively (left). Total, Visceral, and subcutaneous adipose tissue weights calculated based on the micro-CT images of Control (n = 4) and Aα4KO (n = 4) mice at 3 months old (right). Box plots are defined in terms of minima and maxima by whiskers, and the center and bounds of box by quartiles (Two-tailed Student’s t-test, Total: ***p = 0.0007; Visceral: **p = 0.001; Subcutaneous: **p = 0.002). c Representative pictures of the livers of Control and Aα4KO mice at 3 months old. Scale bars = 1 cm (upper). Liver sections from Control and Aα4KO mice fed a CD stained with Oil Red O. Scale bars = 100 μm (bottom). d mRNA expression levels of genes involved in de novo lipogenesis, inflammation, and fibrosis, as well as those of gluconeogenic enzymes in CD-fed Control and Aα4KO mice at 4 months old. Box plots are defined in terms of minima and maxima by whiskers, and the center and bounds of box by quartiles (Two-tailed Student’s t-test, Srebp1c: ***p = 0.0007, G6pase: *p = 0.01, Scd1: **p = 0.001, Fasn: *p = 0.04, Tgfβ1: *p = 0.01, Ccl2: ***p = 0.0002: n = 10/group). e Blood glucose levels corresponding to Control (n = 12) and Aα4KO (n = 13) mice under fed or fasted at 4 months old. Data are presented as mean ± SEM (Two-tailed Student’s t-test, **p = 0.001, ***p < 0.0001). f Serum insulin levels corresponding to Control (n = 8) and Aα4KO (n = 8) mice under fasting (left). Serum leptin levels corresponding to Control (n = 5) and Aα4KO (n = 5) mice under fasting at 4 months old (right) (Two-tailed Student’s t-test, *p = 0.02, **p = 0.002). g Blood glucose levels corresponding to random-fed 12-wk-old Control and Aα4KO mice during 2 wks of leptin (10 μg/mouse/day) or saline treatment using Alzet osmotic minipumps (One-way ANOVA post hoc Bonferroni test, ***p < 0.0001: n = 6/ group). h Body weights of CD-fed and HFD-fed Control and Aα4KO mice for another 16 wks (Two-tailed Student’s t-test, *p < 0.05; **p < 0.01; ***p < 0.001; ^#p < 0.05; ^##p < 0.01: Control CD, n = 9; Aα4KO CD, n = 11; Control HFD, n = 9; and Aα4KO HFD, n = 9). (i) Blood glucose levels of Control and Aα4KO mice under fed (Control CD, n = 12; Aα4KO CD, n = 13; Control HFD, n = 11; and Aα4KO HFD, n = 9) and fasted (n = 9/ group, *p < 0.05; **p < 0.01; ***p < 0.001) (one-way ANOVA post hoc Bonferroni test,). j GTT (left) and GTT AUC corresponding to CD- or HFD-fed Control (n = 10) and Aα4KO (n = 10) after 14 wks of HFD (one-way ANOVA post hoc Bonferroni test, *p < 0.05; **p < 0.01; ***p < 0.001; ^# p < 0.05; ^##p < 0.01). k ITT (left) and the decrease in AUC (right) after 15 wks of HFD (One-way ANOVA post hoc Bonferroni test, CD: *p = 0.02, HFD: *p = 0.01, n = 10/group). (l) Pancreatic sections from CD- or HFD-fed Control and Aα4KO immunostained for insulin and Ki67 after 16 wks of HFD. Scale bars = 100 μm. Quantification of Ki67 + insulin + cells in pancreas sections from Control and Aα4KO (left). Quantitation of β-cell mass (right) from Control and Aα4KO (One-way ANOVA followed by Turkey’s multiple comparisons, *p < 0.05; **p < 0.01; ***p < 0.001, n = 6/group). Box plots are defined in terms of minima and maxima by whiskers, and the center and bounds of box by quartiles. m TG content in livers from CD (n = 7/group) or HFD (n = 7/group) fed mice after 16 wks of HFD (One-way ANOVA post hoc Bonferroni test, **p = 0.001, ***p < 0.0001). n mRNA expression levels of genes involved in inflammation and fibrosis after 16 wks of HFD (n = 9/group) or CD (n = 10/group). Box plots are defined in terms of minima and maxima by whiskers, and the center and bounds of box by quartiles (one-way ANOVA post hoc Bonferroni test, IL-1β: **p = 0.004, ***p = 0.0008, Tgfβ1: **p = 0.001). o Liver sections CD- or HFD-fed Control and Aα4KO immunostained for Iba-1 (upper). Representative Azan staining of liver samples (bottom). Scale bars = 100 μm. Quantification of Azan-positive/total area. Box plots are defined in terms of minima and maxima by whiskers, and the center and bounds of box by quartiles (one-way ANOVA post hoc Bonferroni test, *p < 0.05; **p < 0.01; ***p < 0.001, n = 6 per group) (right). Source data are provided as a Source Data file.