Differential Mitochondrial Gene Expression in Adipose Tissue Following Weight Loss Induced by Diet or Bariatric Surgery

Abstract

Context: Mitochondria are essential for cellular energy homeostasis, yet their role in subcutaneous adipose tissue (SAT) during different types of weight-loss interventions remains unknown.

Objective: To investigate how SAT mitochondria change following diet-induced and bariatric surgery-induced weight-loss interventions in 4 independent weight-loss studies.

Methods: The DiOGenes study is a European multicenter dietary intervention with an 8-week low caloric diet (LCD; 800 kcal/d; n = 261) and 6-month weight-maintenance (n = 121) period. The Kuopio Obesity Surgery study (KOBS) is a Roux-en-Y gastric bypass (RYGB) surgery study (n = 172) with a 1-year follow-up. We associated weight-loss percentage with global and 2210 mitochondria-related RNA transcripts in linear regression analysis adjusted for age and sex. We repeated these analyses in 2 studies. The Finnish CRYO study has a 6-week LCD (800-1000 kcal/d; n = 19) and a 10.5-month follow-up. The Swedish DEOSH study is a RYGB surgery study with a 2-year (n = 49) and 5-year (n = 37) follow-up.

Results: Diet-induced weight loss led to a significant transcriptional downregulation of oxidative phosphorylation (DiOGenes; ingenuity pathway analysis [IPA] z-scores: -8.7 following LCD, -4.4 following weight maintenance; CRYO: IPA z-score: -5.6, all P < 0.001), while upregulation followed surgery-induced weight loss (KOBS: IPA z-score: 1.8, P < 0.001; in DEOSH: IPA z-scores: 4.0 following 2 years, 0.0 following 5 years). We confirmed an upregulated oxidative phosphorylation at the proteomics level following surgery (IPA z-score: 3.2, P < 0.001).

Conclusions: Differentially regulated SAT mitochondria-related gene expressions suggest qualitative alterations between weight-loss interventions, providing insights into the potential molecular mechanistic targets for weight-loss success.

Trial registration: ClinicalTrials.gov NCT00390637 NCT01312090 NCT01785134.

Keywords: adipose tissue; bariatric surgery; diet-induced; mitochondria; transcriptomics; weight loss.

© The Author(s) 2021. Published by Oxford University Press on behalf of the Endocrine Society.

Figures

Figure 1.

Gene enrichment for biological pathways using significantly differentially expressed global genes associated with the weight-loss percentage following the diet-induced weight-loss DiOGenes study (n = 121) (A) and the surgery-induced weight-loss KOBS study (n = 171) (B). The top 10 significant pathways from the ingenuity pathway analysis (IPA) tool (P < 0.001) for both cohorts are presented. For several pathways, IPA provided z-scores for the pathway directionality by calculating the observed number of “activated” genes (z-score > 0), “inhibited” genes (z-score < 0), or no directionality (z-score = 0). Z-scores > 2 or < −2 were considered statistically significant. Results are ranked according to statistical significance.

Figure 2.

Gene enrichment for biological pathways using significantly differentially expressed mitochondria-related genes associated with the weight-loss percentage following diet- and surgery-induced weight loss across all studies (in total, 6 postintervention vs baseline comparisons) (A) and in groups matched for 15% to 20% weight loss from the diet-induced weight-loss DiOGenes study (n = 24) and surgery-induced weight-loss KOBS study (n = 34) (B). In (A), we selected the top 10 significant pathways from the ingenuity pathway analysis (IPA) tool (P < 0.001) from each comparison; the pathway results are shown when present in at least 2 of 6 of the top 10 selections. In (B), the results are presented for the top 10 significant pathways from the IPA tool (P < 0.001) from both studies. For several pathways, IPA provided z-scores for the pathway directionality by calculating the observed number of “activated” genes (z-score > 0), “inhibited” genes (z-score < 0), or no directionality (z-score = 0). Z-scores > 2 or < −2 were considered significant. Results are ranked according to biological function. In (A), gradients indicated by blue bars (from dark to light blue) indicate diet-induced weight-loss comparisons: DiOGenes 2 months vs baseline (n = 261); DiOGenes 8 months vs baseline (n = 121); and CRYO 12 months vs baseline (n = 19). Gradients indicated by red bars (from dark to light red) indicate the surgery-induced weight-loss studies: KOBS, 12 months vs baseline (n = 171); DEOSH, 24 months vs baseline (n = 49); and DEOSH, 60 months vs baseline (n = 37). Abbreviation: m, months.

Figure 3.

Heat map showing the differential expression of the individual genes in the oxidative phosphorylation (OXPHOS) pathway associated with weight-loss percentage following diet- and surgery-induced weight loss. Log2-fold changes indicated for the 5 primary complexes of the electron transport chain genes. The color in the heat maps reflect the differential expressions associated with the weight-loss percentage before and after weight loss, where red indicates up- and blue indicates downregulation. Asterisks indicate statistically significant differential expressions: ***P < 0.001; **P < 0.01; *P < 0.05. Gray squares indicate unmeasured transcripts. The 3 left-most columns show the diet-induced weight-loss comparisons: DiOGenes, 2 months vs baseline (n = 261); DiOGenes, 8 months vs baseline (n = 121); and CRYO, 12 months vs baseline (n = 19). The 3 right-most columns show the surgery-induced weight-loss studies: KOBS, 12 months vs baseline (n = 171); DEOSH, 24 months vs baseline (n = 49); and DEOSH, 60 months vs baseline (n = 37). Abbreviation: m, months.

Figure 4.

SAT mitochondria-related proteins were upregulated following surgical intervention. Protein enrichment for biological pathways using differentially expressed proteins (nominal P < 0.1) associated with the weight-loss percentage following surgery-induced weight loss in KOBS (n = 29) (A), the top 10 significant findings from the ingenuity pathway analysis (IPA) tool (P < 0.001) are shown. For several pathways, IPA provided z-scores for pathway directionality by calculating the observed number of “activated” genes (z-score > 0), “inhibited” genes (z-score < 0), or no directionality prediction (z-score = 0). Z-scores > 2 or < −2 were considered statistically significant. Results ranked according to statistical significance. In (B), a heat map showing the differential expression of the individual proteins in the oxidative phosphorylation (OXPHOS) pathway associated with weight-loss percentage following surgery-induced weight loss. Log2-fold changes indicated for the 5 primary complexes of the electron transport chain proteins. Only measured proteins are shown. The color in the heat maps reflect the differential expressions associated with the weight-loss percentage before and after weight loss, where red indicates up- and blue indicates downregulation. Asterisks indicate statistically significant differential expressions: **P < 0.05; *P < 0.1. Abbreviations: c, cytochrome c; q, ubiquinone.