We investigated whether (1) subcutaneous adipose tissue (SAT) mitochondrial capacity predicts weight loss success and (2) weight loss ameliorates obesity-related SAT mitochondrial abnormalities.
SAT biopsies were obtained from 19 clinically healthy obese subjects (body mass index (BMI) 34.6±2.7 kg/m^2) during a weight loss intervention (0, 5 and 12 months) and from 19 lean reference subjects (BMI 22.7±1.1 kg/m^2) at baseline. Based on 1-year weight loss outcome, the subjects were divided into two groups: continuous weight losers (WL, n=6) and weight regainers (WR, n=13). Main outcome measures included SAT mitochondrial pathways from transcriptomics, mitochondrial amount (mitochondrial DNA (mtDNA), Porin protein levels), mtDNA-encoded transcripts, oxidative phosphorylation (OXPHOS) proteins, and plasma metabolites of the mitochondrial branched-chain amino-acid catabolism (BCAA) pathway. SAT and visceral adipose tissue (VAT) glucose uptake was measured with positron emission tomography.
Despite similar baseline clinical characteristics, SAT in the WL group exhibited higher gene expression level of nuclear-encoded mitochondrial pathways (P=0.0224 OXPHOS, P=0.0086 tricarboxylic acid cycle, P=0.0074 fatty acid beta-oxidation and P=0.0122 BCAA), mtDNA transcript COX1 (P=0.0229) and protein level of Porin (P=0.0462) than the WR group. Many baseline mitochondrial parameters correlated with WL success, and with SAT and VAT glucose uptake. During WL, the nuclear-encoded mitochondrial pathways were downregulated, together with increased plasma metabolite levels of BCAAs in both groups. MtDNA copy number increased in the WR group at 5 months (P=0.012), but decreased to baseline level between 5 and 12 months (P=0.015). The only significant change in the WL group for mtDNA was a reduction between 5 and 12 months (P=0.004). The levels of Porin did not change in either group upon WL.
Higher mitochondrial capacity in SAT predicts good long-term WL success. WL does not ameliorate SAT mitochondrial downregulation and based on pathway expression, may paradoxically further reduce it.
The transcriptomics data generated in this study have been deposited to the Gene Expression Omnibus public repository, accession number GSE103769.