Histological improvements following energy restriction and exercise: The role of insulin resistance in resolution of MASH

Abstract

Background & aims: Metabolic dysfunction-associated steatohepatitis (MASH) is one of the most common liver diseases worldwide and is characterized by multi-tissue insulin resistance. The effects of a 10-month energy restriction and exercise intervention on liver histology, anthropometrics, plasma biochemistries, and insulin sensitivity were compared to standard of care (control) to understand mechanisms that support liver health improvements.

Methods: Following medical diagnosis of MASH, individuals were randomized to treatment (n = 16) or control (n = 8). Liver fat (magnetic resonance spectroscopy), 18-hour plasma biochemical measurements, and isotopically labeled hyperinsulinemic-euglycemic clamps were completed pre- and post-intervention. Body composition and cardiorespiratory fitness (VO₂peak) were also measured mid-intervention. Those in the treatment group were counseled to reduce energy intake and completed supervised, high-intensity interval training (3x/week) for 10 months. Controls continued physician-directed care.

Results: Treatment induced significant (p <0.05) reductions in body weight, fat mass, and liver injury, while VO₂peak (p <0.05) and non-esterified fatty acid suppression (p = 0.06) were improved. Both groups exhibited reductions in total energy intake, hemoglobin A1c, hepatic insulin resistance, and liver fat (p <0.05). Compared to control, treatment induced a two-fold increase in peripheral insulin sensitivity which was significantly related to higher VO₂peak and resolution of liver disease.

Conclusions: Exercise and energy restriction elicited significant and clinically meaningful treatment effects on liver health, potentially driven by a redistribution of excess nutrients to skeletal muscle, thereby reducing hepatic nutrient toxicity. Clinical guidelines should emphasize the addition of aerobic exercise in lifestyle treatments for the greatest histologic benefit in individuals with advanced MASH.

Impact and implications: The mechanisms that underpin histologic improvement in individuals with metabolic dysfunction-associated steatohepatitis (MASH) are not well understood. This study evaluated the relationship between liver and metabolic health, testing how changes in one may affect the other. We investigated the effects of energy restriction and exercise on the association between multi-tissue insulin sensitivity and histologic improvements in participants with biopsy-proven MASH. For the first time, these results show that an improvement in peripheral (but not hepatic) insulin sensitivity and systemic markers of muscle function (i.e. cardiorespiratory fitness) were strongly related to resolution of liver disease. Extrahepatic disposal of substrates and improved fitness levels supported histologic improvement, confirming the addition of exercise as crucial to lifestyle interventions in MASH.

Clinical trial number: NCT03151798.

Keywords: MASH; MASLD; NAFLD; energy restriction; exercise; histology; insulin resistance; lifestyle treatment.

Figure 1.

Experimental design and flow of inpatient metabolic studies (A) Recruitment of all participants was achieved after a liver biopsy demonstrated a NAFLD activity score (NAS)≥4/8. Subjects completed a maximal exercise tolerance test (ETT) and a DEXA in visit 3 (V3) and liver imaging and intermediary metabolism at V4. Some tests were repeated mid-intervention (V5). All baseline procedures were repeated (V6&7) and then participants had a follow-up liver biopsy (V8). Control subjects underwent standard of care. Treatment subjects underwent nutritional counseling and supervised high intensity interval training (HIIT) three times/week. (B) The inpatient protocol from V4 and V7 including MRS, 18h blood metabolite concentrations, and isotope infusion during fasting and a two-step hyperinsulinemic-euglycemic clamp to measure endogenous glucose production and glucose disposal.

Figure 2.

Body composition and cardiorespiratory fitness (A) Body weight, (B) lean mass, (C) fat mass, and (D) absolute VO₂peak are presented as mean±SEM with individual data as lines across the visits (control: n=8; treatment: n=16; Exception for treatment panel D: mid-testing n=14, post-testing n=15). Data were analyzed by mixed model ANOVAs, adjusted for pre-intervention values. Interaction terms with P<0.10 were followed by Tukey-corrected post-hoc analysis. Within a group, an asterisk (*) indicates significant differences from pre-intervention values (P<0.05).

Figure 3.

Inpatient blood biochemistries over 18h Plasma concentrations (mean±SEM, pre-testing [open circles] and post-testing [closed circles]) for control (n=8) and treatment (n=15–16). (A&B) Glucose, (D&E) insulin, and (G&H) NEFA; (C, F, I) AUC for each metabolite. Circadian data were split into three, six-hour phases: postprandial (PP), nighttime (NT), and clamp (CL). The calculated AUC for each phase was analyzed by mixed model ANOVAs (table J), adjusted for fasting values (P-values are presented).

Figure 4.

Multi-tissue insulin sensitivity (A) Basal hepatic insulin resistance (HepIR), (B) insulin-stimulated (low-dose insulin-7 mU/m²•min) HepIR, (C) peripheral insulin sensitivity (IS, high-dose insulin-50 mU/m²•min), (D) NEFA suppression (low-dose insulin) are presented as mean±SEM with individual data as lines graphs (control: n=8; treatment: n=16). Data were analyzed by mixed model ANOVAs, adjusted for pre- intervention values. Interaction terms with P<0.10 were followed by Tukey-corrected post-hoc analysis.

Figure 5.

Liver health (A) Intrahepatic triacylglycerol (IHTG; Treatment: pre-testing n=15, post-testing n=14), (B) steatosis, (C) lobular inflammation, (D) hepatocellular ballooning, (E) steatohepatitis activity (sum of inflammation and ballooning), and (F) fibrosis are presented as mean±SEM with individual data as lines graphs (control: n=8; treatment: n=16). Representative histology from (G) subject #1 and (H) subject #9. Data analyzed by (A) mixed model ANOVAs, adjusted for pre-intervention values (* P<0.05 versus pre-values) and by (B-F) cumulative link models to determine the effects of group and time on ranked outcome variables. Model significance and test statistics are presented in table I (χ² and T values) with T-values presented for the Group, Time, and GroupxTime effects. Following the model estimation, the estimated marginal means (EMMs) for each combination of ‘Group’ and ‘Time’ were computed using the emmeans() function. Values above bar graphs represent pairwise comparisons between the EMMs were conducted, adjusting for multiple comparisons using the Tukey method. Only liver fat (IHTG: P=0.032 and steatosis: P=0.073) was different between groups pre-intervention.

Figure 6.

Correlates of improvements in liver disease Correlations between (A, Spearman) change in NAS and HepIR, (B, Spearman) change in peripheral insulin sensitivity (IS) for control (n=8; light-filled) and treatment (n=16; dark-filled) subjects and (C, Pearson) change in VO₂peak and HepIR, (D, Pearson) change in VO₂peak and peripheral IS, and (E, Spearman) change in VO₂peak and NAS are shown for n=7 control and n=15 treatment.