Decreased adipose tissue oxygenation associates with insulin resistance in individuals with obesity

Abstract

BACKGROUNDData from studies conducted in rodent models have shown that decreased adipose tissue (AT) oxygenation is involved in the pathogenesis of obesity-induced insulin resistance. Here, we evaluated the potential influence of AT oxygenation on AT biology and insulin sensitivity in people.METHODSWe evaluated subcutaneous AT oxygen partial pressure (pO2); liver and whole-body insulin sensitivity; AT expression of genes and pathways involved in inflammation, fibrosis, and branched-chain amino acid (BCAA) catabolism; systemic markers of inflammation; and plasma BCAA concentrations, in 3 groups of participants that were rigorously stratified by adiposity and insulin sensitivity: metabolically healthy lean (MHL; n = 11), metabolically healthy obese (MHO; n = 15), and metabolically unhealthy obese (MUO; n = 20).RESULTSAT pO2 progressively declined from the MHL to the MHO to the MUO group, and was positively associated with hepatic and whole-body insulin sensitivity. AT pO2 was positively associated with the expression of genes involved in BCAA catabolism, in conjunction with an inverse relationship between AT pO2 and plasma BCAA concentrations. AT pO2 was negatively associated with AT gene expression of markers of inflammation and fibrosis. Plasma PAI-1 increased from the MHL to the MHO to the MUO group and was negatively correlated with AT pO2, whereas the plasma concentrations of other cytokines and chemokines were not different among the MHL and MUO groups.CONCLUSIONThese results support the notion that reduced AT oxygenation in individuals with obesity contributes to insulin resistance by increasing plasma PAI-1 concentrations and decreasing AT BCAA catabolism and thereby increasing plasma BCAA concentrations.TRIAL REGISTRATIONClinicalTrials.gov NCT02706262.FUNDINGThis study was supported by NIH grants K01DK109119, T32HL130357, K01DK116917, R01ES027595, P42ES010337, DK56341 (Nutrition Obesity Research Center), DK20579 (Diabetes Research Center), DK052574 (Digestive Disease Research Center), and UL1TR002345 (Clinical and Translational Science Award); NIH Shared Instrumentation Grants S10RR0227552, S10OD020025, and S10OD026929; and the Foundation for Barnes-Jewish Hospital.

Keywords: Adipose tissue; Glucose metabolism; Metabolism; Obesity.

Figure 1. Flow diagram of study participants.

Figure 2. AT oxygen tension is reduced in MUO and is positively associated with insulin sensitivity.

(A and B) AT oxygen partial pressure (AT pO₂) (A) and HIF1A gene expression (B) in metabolically healthy lean (MHL; n = 11), metabolically healthy obese (MHO; n = 15), and metabolically unhealthy obese (MUO; n = 20) groups. Data are means ± SEM. One-way ANOVA and Fisher’s least significant difference procedure were used to identify significant mean differences between groups. *Value significantly different from the MHL value, P < 0.001. ^†Value significantly different from the MHO value, P < 0.05. ^#Linear trend, P < 0.001. (C and D) Relationship between AT pO₂ and hepatic insulin sensitivity index (HISI), assessed as 1000 divided by the product of endogenous glucose rate of appearance (in μmol/kg fat-free mass/min) and plasma insulin concentration (in μU/mL) in the overnight fasting state (C), and between pO₂ and whole-body insulin sensitivity, assessed as glucose rate of disposal (Rd; in nmol/kg fat-free mass/min) divided by plasma insulin (I) concentration (in μU/mL) during a hyperinsulinemic-euglycemic clamp procedure (D), in MHL (white circles; n = 11), MHO (gray circles; n = 15), and MUO (black circles; n = 20) participants. Associations between AT pO₂ and HISI and between AT pO₂ and glucose Rd/I were determined using Pearson’s correlation coefficient.

Figure 3. AT biological pathways associated with AT oxygen tension.

(A and B) RNA sequencing was conducted in AT samples obtained from MHL (n = 11), MHO (n = 14), and MUO (n = 20) participants. Functional enrichment analyses were performed to identify biological pathways significantly (false discovery rate < 0.01) enriched with genes positively (red) (A) and negatively (blue) (B) associated with AT oxygen tension.

Figure 4. AT oxygen tension is negatively associated with plasma BCAA concentrations and positively associated with AT expression of genes involved in BCAA catabolism.

(A) Fasting plasma leucine, isoleucine, and valine concentrations in MHL (n = 11), MHO (n = 15), and MUO (n = 20) groups. Data are means ± SEM. One-way ANOVA and Fisher’s least significant difference procedure were used to identify significant differences between groups. *Value significantly different from the corresponding MHL value, P < 0.05. ^†Value significantly different from the corresponding MHO value, P < 0.05. ^#Significant linear trend for the 3 groups, P < 0.05. (B and C) Relationship between AT pO₂ and plasma leucine, isoleucine, and valine concentrations (B) and AT gene expression of key enzymes involved in BCAA catabolism (C) in MHL (white circles; n = 11), MHO (gray circles; n = 15 in B and n = 14 in C), and MUO (black circles; n = 20) participants. Associations between AT pO₂ and plasma BCAA concentrations and between AT pO₂ and AT gene expression were determined using Pearson’s correlation coefficient.

Figure 5. AT oxygen tension is positively associated with AT expression of genes involved in regulating protein synthesis.

Relationship between AT pO₂ and expression of AT genes involved in regulating protein synthesis in MHL (white circles; n = 11), MHO (gray circles; n = 14), and MUO (black circles; n = 20) participants. Associations between AT pO₂ and AT gene expression were determined using Pearson’s correlation coefficient.

Figure 6. AT oxygen tension is inversely associated with AT expression of genes related to inflammation.

Relationship between AT pO₂ and AT expression of genes related to inflammation in MHL (white circles; n = 11), MHO (gray circles; n = 14), and MUO (black circles; n = 20) participants. Associations between AT pO₂ and AT gene expression were determined using Pearson’s correlation coefficient.

Figure 7. AT oxygen tension is inversely associated with AT expression of genes involved in ECM remodeling.

Relationship between AT pO₂ and AT expression of genes involved in ECM remodeling in MHL (white circles; n = 11), MHO (gray circles; n = 14), and MUO (black circles; n = 20) participants. Associations between AT pO₂ and AT gene expression were determined using Pearson’s correlation coefficient.

Figure 8. AT oxygen tension is positively associated with AT VEGFA expression.

(A) AT VEGFA gene expression in MHL (n = 11), MHO (n = 14), and MUO (n = 20) groups. Data are means ± SEM. One-way ANOVA and Fisher’s least significant difference procedure were used to identify significant mean differences between groups. *Value significantly different from the MHL value, P < 0.001. ^†Value significantly different from the MHO value, P < 0.05. ^#Linear trend, P < 0.001. (B) Relationship between AT pO₂ and AT VEGFA gene expression in MHL (white circles; n = 11), MHO (gray circles; n = 14), and MUO (black circles; n = 20) participants. Associations between AT pO₂ and AT gene expression were determined using Pearson’s correlation coefficient.

Figure 9. Proposed effects of decreased AT oxygen tension on AT biology and insulin sensitivity in people with obesity.

An increase in AT mass associated with obesity decreases AT pO₂, which increases AT HIF-1α expression and triggers a cascade of alterations in AT biology, including a decrease in BCAA catabolism and an increase in inflammation and fibrosis. These changes lead to an increase in circulating BCAAs and PAI-1 that impair systemic insulin action. Differences in AT pO₂ among people with obesity provide a potential mechanism to help explain the differences in insulin sensitivity between people with metabolically healthy and metabolically unhealthy obesity.