Diet composition influences the metabolic benefits of short cycles of very low caloric intake

Obesity driven by sedentary lifestyle and Western dietary habits is linked to metabolic disorders and increased mortality. Caloric restriction (CR), time-restricted feeding, and intermittent fasting improve metabolic health and longevity in animal models, likely via reductions in circulating glucose and growth factors, ketone body utilization, circadian modulation, and enhanced metabolic flexibility. Periodic 4:10 cycles (4 days of energy restriction followed by 10 days ad libitum) using a plant-based fasting-mimicking diet (FMD) have shown benefits, but it is unclear whether these arise from diet composition, degree of caloric reduction, fasting duration, or their combination. The study aims to dissect how diet composition (standard chow vs plant-based FMD) during short VLCI cycles affects metabolic health, physical performance, and metabolomic remodeling, and whether such cycles prevent diet-induced obesity or induce lasting metabolic memory, including under an obesogenic high-fat diet (HFD).

Prior rodent and human studies demonstrate benefits of CR and fasting-related regimens (IF, alternate-day energy restriction, time-restricted feeding), including improved insulin sensitivity, reduced inflammation, and disease risk, with lifespan extension in animals. FMD-based 4:10 cycles were reported to promote multi-system regeneration and improve risk markers for aging and chronic diseases. However, intermittent strategies may be hard to sustain and could cause compensatory hyperphagia and altered eating behaviors. Clinical comparisons of intermittent versus continuous energy restriction show similar weight and cardiometabolic risk reductions, but long-term benefits, safety, and the roles of fasting duration, caloric depth, and diet composition remain uncertain, particularly in overweight/obese subjects. This work builds on these findings to isolate the contribution of diet composition within short VLCI cycles.

Design: Middle-aged male C57BL/6J mice (12 months; n≈18–19/group initially) underwent 11 cycles (5 months) of 4 days of very low-calorie intake (VLCI) followed by 10 days ad libitum (AL) refeeding (RF). Diet groups under standard diet (SD): SD-AL controls; LCC (low-calorie cycles) received SD at 50%, 70%, 70%, 70% of daily calories on days 1–4; FMD received a plant-based FMD at 33%, 54%, 54%, 54% of daily calories on days 1–4. After day 4, both LCC and FMD returned to SD-AL for 10 days. Obesogenic cohorts: HFD-AL controls; HFD+LCC (4-day reductions of 50/70/70/70% on HFD, then 10 days HFD-AL); HFD+FMD (FMD at 33/54/54/54% for 4 days, then HFD-AL 10 days). VLCI food was provided in a 2-h afternoon window (3:30–5:30 pm). Measurements: - Body weight and food intake recorded on days 1–4, 5, 8, 11 of each cycle. - Body composition (lean, fat, fluid) by NMR (Minispec LF90) at baseline and cycles 4 and 10. - Indirect calorimetry (CLAMS) over 6 days spanning 1 day fed, 4 days VLCI, 1 day RF during cycles 3–4: respiratory exchange ratio (RER), heat production, and ambulatory activity. - Blood biomarkers and HOMA-IR2: 6-h fasted glucose (glucometer), insulin (ELISA), leptin (ELISA), 3-hydroxybutyrate; sampling on day 4 (VLCI) and day 10/11 (RF) during cycles 3 and 9. - Glucose tolerance (OGTT) per Supplementary methods. - Physical performance during RF phase (days 10–14): inverted cage top (grip/strength-endurance), rotarod (motor coordination/balance), treadmill endurance (separate young cohort at 26 weeks for treadmill protocol). - Untargeted metabolomics (serum and liver) at cycle 11: sampled in morning on day 3 (VLCI; prior-day 2-h feeding) or day 10 (6 days AL refeeding). Samples processed at UC Davis West Coast Metabolomics Center; volcano plots (FC ≥1.2 or ≤0.83, p≤0.05), Venn analyses, heatmaps; pathway analysis via MetaboAnalyst 4.0. - Statistics: mean ± SEM; two-tailed t-tests or one-way ANOVA with Fisher’s LSD; Spearman correlations; ANCOVA for energy expenditure vs body weight; significance p<0.05. Husbandry: Single-housed, controlled environment; SD (AIN-93G) and HFD (60% fat kcal). FMD plant-based formulation (carb 29%, fat 65.3%, protein 5.46%) pelletized with glycerol/hydrogel; 2.9 kcal/g.

Standard diet context (SD): - Body weight: Over 5 months, SD-AL gained ~6.9%, LCC lost ~4.3%, FMD lost ~0.9%. Within each 4-day VLCI, mice lost ~8–12% (LCC ~12%; FMD ~8%), regaining most upon refeeding. - Intake/behavior: LCC and FMD showed compensatory hyperphagia during RF, yet total 5-month calorie intake per mouse fell by ~12.4% (LCC) and ~7.7% (FMD) vs SD-AL; BW-normalized cumulative intake similar across groups. - Body composition: At cycle 4 day 4, LCC had significantly lower fat and lean mass vs SD; FMD reduced both but with less effect on lean-to-fat ratio. After 7 days RF, partial regain occurred. By cycle 10, LCC maintained significant reductions in fat and lean mass; FMD had less consistent fat loss and lower lean mass. - Metabolic health: At VLCI day 4 (cycle 9), glucose, insulin, leptin, and HOMA-IR2 were significantly reduced in LCC and FMD vs SD-AL; after 7 days RF, LCC showed sustained lower leptin (and trends for lower glucose/insulin), whereas FMD returned toward SD-AL. OGTTs were not significantly altered. - Physical performance (RF phase): LCC outperformed FMD and SD-AL in inverted cage top (significant at 2 and 4 months), with latency to fall negatively correlating with percent body fat; LCC improved rotarod performance vs SD-AL after 4 months; both LCC and FMD improved treadmill endurance vs SD-AL. - Energy metabolism: LCC exhibited larger RER amplitude (~0.70 day/fasting to ~0.95 night/feeding), indicating enhanced metabolic flexibility; FMD showed smaller postprandial RER amplitudes during VLCI, consistent with composition-driven substrate use. VLCI reduced heat production; ambulatory activity trended higher. - Metabolomics (serum): VLCI altered 52 (LCC) and 54 (FMD) metabolites vs SD, with 31 shared (lipids, ketone bodies, amino acids). LCC-RF retained 22 altered metabolites vs SD-AL, 16 overlapping with LCC (persistent ‘fasting’ signature), whereas FMD-RF showed only 7 altered, 5 overlapping (weaker persistence). Beta-hydroxybutyrate and lipid species were elevated; AAs were reduced during VLCI. - Metabolomics (liver): VLCI altered 46 (LCC) and 54 (FMD) hepatic metabolites vs SD; 26 shared. Top pathways up-modulated: purines, lipids/ketone bodies, ascorbate, redox. LCC-RF vs SD-AL showed 22 altered metabolites (14 shared with VLCI), preserving lipids/ketone bodies and purine up-modulation with decreased lysine degradation and redox; FMD-RF showed 8 altered with minimal overlap (2 shared). Obesogenic diet context (HFD): - Body weight and composition: The efficacy of 4:10 cycles to reduce weight declined over time; HFD+FMD maintained weight loss better than HFD+LCC during study, though neither prevented overall diet-induced obesity. Within-cycle losses ~11% (HFD+LCC) and ~9% (HFD+FMD), with regain on RF. Cumulative intake comparable across HFD groups. Cycle 4 day 4: both HFD+LCC and HFD+FMD reduced fat and lean mass vs HFD-AL; lean-to-fat ratio increased in HFD+LCC at day 4 and in HFD+FMD after RF. - Metabolic health: At VLCI day 4, fasting glucose, insulin, and HOMA-IR2 were reduced vs HFD-AL; leptin unchanged. After RF, parameters returned to HFD-AL, with glucose and leptin surging in HFD+LCC. OGTTs unchanged. - Physical performance and metabolism: Minor improvements in cage top for HFD+LCC with normalization; no rotarod differences. All HFD groups showed low RER indicative of FA oxidation; VLCI further reduced heat and increased locomotor activity; RER dynamics similar across HFD groups. - Metabolomics (serum): VLCI altered 30 (HFD+LCC) and 42 (HFD+FMD) metabolites vs HFD-AL; 17 shared (lipids/ketone bodies, AAs, redox). After RF, serum VLCI signature was largely lost: HFD+LCC-RF had only 2 altered metabolites (none overlapping), HFD+FMD-RF had 13 (3 overlapping). - Metabolomics (liver): VLCI altered 29 (HFD+LCC) and 49 (HFD+FMD) hepatic metabolites vs HFD-AL; 15 shared, highlighting up-regulated ketone bodies, ascorbate, short-chain fatty acids, with AA biosynthesis down-regulated. After RF, HFD+LCC-RF showed 5 altered (4 shared with VLCI); HFD+FMD-RF had 7 altered (2 shared). Overall: Short VLCI cycles improved metabolic flexibility, glucose-insulin markers, and physical performance under SD regardless of diet composition, but only SD-LCC induced a persistent ‘metabolic memory’ after refeeding in serum and liver. Under HFD, VLCI did not prevent obesity nor induce lasting metabolic memory, though modest flexibility benefits appeared.

The study addressed whether diet composition during short VLCI cycles modulates physiological and metabolic benefits and whether these cycles induce a lasting metabolic memory. Both LCC (standard chow under severe energy restriction) and FMD (plant-based, milder restriction) improved metabolic markers and physical performance under SD; however, LCC more robustly enhanced coordination/strength and metabolic flexibility. Untargeted metabolomics revealed a shared hepatic ‘core’ (purines, lipids/ketone bodies, ascorbate, redox) under VLCI irrespective of diet, consistent with fuel switching to fat oxidation and ketogenesis and enhanced purine turnover. Critically, after 6 days AL refeeding, only LCC preserved a substantial fraction of the VLCI-associated metabolomic signature in serum and liver, indicating a selective persistent metabolic ‘memory’, potentially linked to sustained hypoleptinemia and adipose-driven nutritional memory. In contrast, FMD’s signature largely reverted to SD-AL. In the obesogenic HFD context, repeated VLCI cycles did not prevent weight gain or maintain a fasting signature after refeeding, although HFD+FMD retained weight loss better than HFD+LCC within cycles and serum 3-HB was elevated. Energy metabolism analyses (RER, heat, activity) corroborated substrate shifts and reduced energy expenditure during VLCI. These findings underscore that diet composition strongly shapes the durability and magnitude of metabolic remodeling induced by short VLCI cycles and that an obesogenic environment abrogates long-term benefits despite transient improvements.

Short 4:10 cycles of very low caloric intake enhance metabolic health and physical performance in middle-aged male mice. VLCI induces a shared hepatic and serum ‘core’ remodeling (purines, lipids/ketone bodies, ascorbate, redox), but the persistence of this remodeling after refeeding is diet-dependent: standard chow-based LCC elicits a lasting metabolic footprint, whereas plant-based FMD shows weaker persistence. Under HFD, VLCI fails to prevent obesity or sustain a metabolic memory. These results highlight the importance of diet composition in VLCI protocols and suggest that optimizing macronutrient content and caloric depth could maximize durable benefits. Future work should dissect mechanisms underlying the persistent LCC signature (e.g., adipose-leptin axis, hepatic pathways), test iso-caloric comparisons of LCC vs FMD, and evaluate sex, strain, age, duration, and resilience outcomes, as well as long-term healthspan and lifespan impacts.

Direct comparison between LCC and FMD is constrained by differing caloric reductions (non-isocaloric). Independent validation of the biochemical pathways implicated in the sustained LCC metabolomic signature is needed. Generalizability is limited by use of only male C57BL/6J mice and a specific age; effects of sex, strain, age at onset, degree and length of VLCI require testing. Longer-term, longitudinal studies in both sexes are needed to assess durability of effects, aging phenotypes, resilience, and lifespan outcomes.