Effects of FGF21, soluble TGFBR2, and environmental temperature on metabolic dysfunction in lipodystrophic mice

Abstract

Metabolic health is influenced by adipose tissue, and obesity and lipodystrophy are characterized by inflammation and metabolic dysfunction. Whereas obesity and lipodystrophy treatments involve pharmacological approaches and lifestyle changes, these therapies require long-term, repeated dosing and are not successful for all patients. Gene therapy with targets such as FGF21 and soluble TGF-β receptor 2 (sTGFBR2) provides an alternative approach, specifically in lipodystrophy. Preclinical experiments in mice housed at 22°C are confounded by a mild cold stress not generally experienced by humans, which can negatively affect translation of metabolic therapeutics. In this study, we investigated effects of FGF21/sTGFBR2 combination gene therapy on obese and lipodystrophic mice and how housing temperature influences therapeutic efficacy. In obese mice, FGF21/sTGFBR2 improved insulin resistance and hyperlipidemia more dramatically at warmer temperatures. In lipodystrophic mice on a high-fat diet, combination therapy required adipose tissue to improve insulin resistance at 30°C, whereas FGF21 alone improved insulin resistance at 22°C. Transcriptomic analyses revealed that lipodystrophic mice had upregulated hepatic cell proliferation and fibrosis pathways and that FGF21 promoted hepatic metabolism. Thus, metabolic dysfunction caused by lipodystrophy is improved by targeting FGF21 and TGFB signaling, but effectiveness in preclinical models may be dependent upon environmental temperature and presence of adipose tissue.

Keywords: Adipose tissue; Bone biology; Bone marrow; Metabolism; Obesity.

Figure 1. FGF21/sTGFBR2 reduces body weight, improves insulin sensitivity, and reduces plasma TAG in HFD-fed mice, with greatest effects in mice housed at 26°C and 30°C.

(A) Schematic illustration of housing conditions and experimental timeline. (B) Concentrations of FGF21 in plasma pretransduction (Pre) and at euthanasia (Euth). (C) Concentrations of sTGFBR2 in plasma 3, 5, 7, and 9 weeks after transduction. Body weight over time in 60% HFD-fed female mice housed at (D) 22°C, (E) 26°C, or (F) 30°C (n = 6–7). Insulin tolerance tests (0.75 U/kg) in mice housed at (G) 22°C, (H) 26°C, or (I) 30°C (n = 6–7), data presented as mean ± SEM. #Mice in G were given intraperitoneal glucose because their blood glucose concentrations dropped below 30 mg/dL. (J) Plasma TAG concentrations 8 weeks posttreatment in mice housed at 22°C, 26°C, or 30°C (n = 6–7). *P < 0.05. Statistical analyses were performed using 2-way ANOVA, followed by Bonferroni’s post hoc test. TAG, triacylglycerol.

Figure 2. FGF21/sTGFBR2 decreases adiposity and plasma insulin concentrations, with highest efficacy at housing temperatures of 26°C and 30°C.

Female mice were euthanized 10 weeks after transduction with FGF21/sTGFBR2 (n = 5–7). (A) Body weight of mice transduced with FGF21/sTGFBR2 at euthanasia. (B) Posterior subcutaneous white adipose tissue (psWAT) and (C) brown adipose tissue (BAT) weights. (D) Representative histological images of mouse BAT stained with H&E. Scale bar = 50 μm. Fed (E) blood glucose concentrations, (F) plasma TAG, and (G) insulin concentrations prior to euthanasia. Tibial (H) cortical thickness (Ct.Th), (I) cortical area (Ct.Ar), and (J) cortical bone volume fraction (Ct.BV/TV) analyzed via nanoCT. *P < 0.05. Statistical analyses were performed using 2-way ANOVA, followed by Bonferroni’s post hoc test.

Figure 3. FGF21/sTGFBR2 decreases liver weights and hepatic steatosis, with highest efficacy at housing temperatures of 26°C and 30°C.

(A) Liver weights in vehicle- and FGF21/sTGFBR2-treated mice at all housing temperatures 10 weeks after transduction (n = 6–7). (B) Liver TAG concentration normalized to protein amount (n = 6–7). (C) Representative histological images of mouse livers stained with H&E. Scale bar = 50 μm. (D) Representative images of mouse livers stained with Picrosirius red to identify collagen. Scale bar = 50 μm. *P < 0.05. Statistical analyses were performed using 2-way ANOVA, followed by Bonferroni’s post hoc test.

Figure 4. Adipose tissue is required for FGF21/sTGFBR2 to mediate metabolic improvements at thermoneutrality.

(A) Schematic illustration of housing conditions and experimental timeline; Lmna^fl/fl = control mice and Lmna^ADKO = lipodystrophic mice. Plasma (B) FGF21 and (C) sTGFBR2 concentrations in mice 12 weeks after transduction (n = 4–7). (D) Body weight in vehicle and FGF21/sTGFBR2 mice 4 weeks after transduction (n = 5–7). (E) Fat and (F) lean mass measured by EchoMRI 4 weeks after transduction (n = 4–7). (G) Plasma TAG and (H) blood glucose concentrations 4 weeks posttreatment (n = 5–7). (I) Insulin tolerance tests (0.75 U/kg; n = 5–7), data presented as mean ± SEM. *P < 0.05. Statistical analyses were performed using 2-way ANOVA, followed by Bonferroni’s post hoc test.

Figure 5. FGF21/sTGFBR2 does not improve metabolic dysfunction of lipodystrophic mice housed at thermoneutrality.

Female mice were euthanized 12 weeks after FGF21/sTGFBR2 transduction (n = 4–7). (A) Body, (B) psWAT, (C) parametrial white adipose tissue (pmWAT), (D) BAT, (E) spleen, and (F) heart weights in mice at euthanasia. Fed (G) blood glucose, (H) plasma TAG, and (I) plasma insulin concentrations prior to euthanasia. (J) Representative tibial osmium-nanoCT images. (K) rBMAT normalized to marrow volume. (L) cBMAT normalized to marrow volume. *P < 0.05. Statistical analyses were performed using 2-way ANOVA, followed by Bonferroni’s post hoc test.

Figure 6. FGF21/sTGFBR2 reduces liver weight, total hepatic TAG, and liver collagen deposition in lipodystrophic mice housed at 30°C.

Female mice were euthanized 12 weeks after FGF21/sTGFBR2 transduction (n = 4–7). (A) Liver weights at euthanasia. (B) Liver TAG normalized to protein content. (C) TAG amount normalized to total liver weight. (D) Representative histological images of mouse livers stained with H&E. Scale bar = 50 μm. (E) Representative images of mouse livers stained with Picrosirius red for collagen. Scale bar = 50 μm. (F) Machine learning quantification of area stained by Picrosirius red in the liver. Statistical analyses were performed using 2-way ANOVA, followed by Bonferroni’s post hoc test.

Figure 7. FGF21 improves insulin sensitivity of lipodystrophic mice housed at 22°C.

(A) Schematic illustration of housing conditions and experimental timeline. (B) Plasma FGF21 concentrations 8 weeks post-transduction (n = 3–16). (C) Body weight, (D) fat mass, and (E) lean mass measured by EchoMRI 7 weeks posttransduction (n = 3–13). (F) Glucose tolerance test (1 mg/kg; n = 3–13), data presented as mean ± SEM. (G) Insulin tolerance test (0.75 U/kg; n = 3–13), data presented as mean ± SEM. #Mice in G received intraperitoneal glucose because their blood glucose dropped below 30 mg/dL. Indirect calorimetry was measured for 3 days using the Promethion system. (H) Energy expenditure (EE) averaged during dark or light cycles, (I) fat oxidation averaged during the first 4 hours of dark or light cycles, and (J) distance traveled averaged during dark or light cycles. *P < 0.05. Statistical analyses were performed using 2-way ANOVA, followed by Bonferroni’s post hoc test.

Figure 8. FGF21 does not improve metabolic dysfunction of lipodystrophic mice housed at 22°C.

Female mice were euthanized 8 weeks after FGF21 administration (n = 3–15). (A) Body, (B) psWAT, (C) pmWAT (D) BAT, and (E) heart weights at euthanasia. Fed (F) blood glucose, (G) plasma TAG, (H) plasma insulin, and (I) plasma leptin concentrations prior to euthanasia. *P < 0.05. Statistical analyses were performed using 2-way ANOVA, followed by Bonferroni’s post hoc test.

Figure 9. FGF21 reduces liver weight, total hepatic TAG content, and liver collagen deposition in lipodystrophic mice housed at 22°C.

Female mice were euthanized 8 weeks after FGF21 administration (n = 3–15). (A) Liver weights at euthanasia. (B) Liver TAG normalized to protein content. (C) Total liver TAG normalized to liver weights. (D) Representative histological images of H&E-stained mouse livers. Scale bar = 50 μm. (E) Representative images of mouse livers stained with Picrosirius red for collagen. Scale bar = 50 μm. (F) Machine learning quantification of area stained by Picrosirius red in the liver. *P < 0.05. Statistical analyses were performed using 2-way ANOVA, followed by Bonferroni’s post hoc test.

Figure 10. RNA-Seq analyses of livers reveal that livers from lipodystrophic mice have increased cell proliferation and increased fibrosis and that FGF21 treatment increases mitochondrial and translation pathways.

(A) Summary of key GSEA pathways in hepatic mRNA of CTRL versus ADKO mice 8 weeks after treatment with FGF21 or vehicle. (B) GSEA pathways changed in CTRL mice dependent on treatment, padj < 0.05. (C) GSEA pathways changed in ADKO mice dependent on treatment, padj < 0.05. (D) GSEA pathways changed in vehicle-treated mice dependent on genotype, padj < 0.05. (E) GSEA pathways changed in FGF21-treated mice dependent on genotype, padj < 0.05.