(a) Hepatic triglyceride content, (n = 5–10/group). (b) Thiobarbituric acid reactive substances (TBARS)/Malonaldehyde (MDA) levels in liver tissue, (n = 5–8/group). (c) Protein expressions of oxidative stress-related proteins: glyoxalase 1 (GLO1) and superoxide dismutase 1 (SOD1), (n = 3–8/group). Actin is used as loading control. All graphs represent mean ± S.E.M expressed as arbitrary units normalized to the mean of the WT control mice. * = p ≤ 0.05, ** ^## = p ≤ 0.01, *** ^### = p ≤ 0.001 (* WT vs. STAT6-deficient mice; ^#Control vs. RSG-treated mice) determined by one way analysis of variance (ANOVA) test. All runs were done in parallel with the same experimental conditions.

(a) Western blot and (b) real-time PCR analysis of PKM2 expressions in livers of control and RSG-treated mice, (n = 5/group). Graphs represent mean ± S.E.M expressed as arbitrary units normalized to the mean of control WT mice. ^# = p ≤ 0.05, ** = p ≤ 0.01, ^### = p ≤ 0.001 (* WT vs. STAT6-deficient mice; ^#Control vs. RSG-treated mice) determined by one way analysis of variance (ANOVA) test. (PKM2 protein Student's t-test p = 0.003, control WT vs. STAT6-deficient mice). All runs were done in parallel with the same experimental conditions.

(a) Hematoxylin-eosin (HE) and PKM2 immunohistochemical stainings in livers of WT (upper panels) and STAT6-deficient (lower panels) mice. Images are representative of pictures obtained from five mice/genotype. Arrows point to PKM2 positive non-parenchymal cells in: 1) association with inflammatory foci, 2) individual localization and 3) surrounding hepatocytes. (b) Serial immunohistochemical or histological stainings from liver sections of WT RSG-treated mice. The middle section was used for PKM2 labelling and the two adjacent sections were used for markers for different non-parenchymal liver cell types. Upper row: the macrophage markers CD206 (M2 type) and F4/80 (M1 type); lower row: GFAP: Glial fibrillary acidic protein and Masson trichrome staining (collagen).

Western blot analysis of CD206 expression in livers of control and RSG-treated mice, (n = 5/group). Graphs represent mean ± S.E.M expressed as arbitrary units normalized to the mean of control WT mice. * ^# = p ≤ 0.05, (* WT vs. STAT6-deficient mice; ^#Control vs. RSG-treated mice) determined by one way analysis of variance (ANOVA) test. Actin is used as loading control. All runs were done in parallel with the same experimental conditions.

(a) Epididymal (WATe) and intra-abdominal (WATi) white adipose tissue weights of control and RSG-treated mice expressed as percentage of total body weight, (n = 5–8/group). (b) Representative hematoxylin-eosin stained images of WATe derived from WT and STAT6-deficient mice (upper panels) and their computer-assisted transformation (lower panels). Adipocytes are colored according to their size. (c) Histogram of adipocyte size distribution and the bar graph of adipocyte diameter of control and RSG-treated mice. Data were obtained from 6 images/mouse derived from two sections of three different levels, n = 5 mice/group. (d) Serum leptin levels measured by ELISA analysis (n = 4–5/group). Graph represents mean ± S.E.M; ** = p ≤ 0.01, *** ### = p ≤ 0.001 (* WT vs. STAT6-deficient mice; #Control vs. RSG-treated mice) determined by one-way analysis of variance (ANOVA) test.

(a) Western blot of PKM2 in WAT of control and RSG-treated mice, (n = 5/group). (b) Serial immunohistochemical and histological stainings from adipose tissue sections of RSG-treated STAT6-deficient mice. The middle section was used for PKM2 labelling and the two adjacent sections were used for markers for the macrophage markers CD206 (M2 type) and F4/80 (M1 type) (c) Real-time PCR analysis of MCP1and IL-1β expression (d) Intraperitoneal insulin tolerance test (ipITT). Mice (n = 4–5/group) were starved overnight and injected intraperitonelly by insulin (1 U/Kg) and glucose levels were monitored for a period of one hour. Graph represents glucose levels expressed as the percentage of initial level. (e) AUC = area under the curve. All graphs represent mean ± S.E.M expressed as arbitrary units normalized to the mean of control WT mice. Results were analyzed by one-way ANOVA test, * # = p ≤ 0.05, (* WT vs. STAT6-deficient mice; ^#Control vs. RSG-treated mice). All runs were done in parallel with the same experimental conditions.

(a) Under normal conditions STAT6 cooperates with PPARγ to induce adipose tissue accumulation. In the absence of STAT6 adipose tissue mass is decreased and the excess fat is stored in the liver resulting in hepatic oxidative stress and inflammation. (b) Induction of PPARγ in the absence of STAT6 results in enhanced adipogenesis and adipose tissue lipid storage, resulting in decreased secretion of inflammatory cytokines (MCP1, IL-1β). In parallel, in liver, lipid accumulation decreases, along with diminished oxidative stress and inflammation. These changes are accompanied by attenuated immune cell PKM2 expression, indicating differences in metabolism and activation. The positive effect on hepatic homeostasis can be related to 1.) decreased adipose tissue inflammatory cytokine production, or 2.) a crosstalk between liver inflammatory cells and hepatocytes. In addition, 3.) immunmodulatory effects of adipose tissue cytokines on liver immune cells might results in STAT6-dependent upregulation of liver PKM2 expression indicating differences in activation profile. These three potential crosstalk events are indicated by dotted arrows.