Determining the metabolic effects of dietary fat, sugars and fat-sugar interaction using nutritional geometry in a dietary challenge study with male mice

The study addresses how dietary fats and sugars, individually and in combination, influence obesity-related metabolic outcomes, a central issue in the carbohydrate versus fat debate. Two predominant models frame the debate: the energy balance model (EBM) posits that obesity results from excess caloric intake regardless of macronutrient source, while the carbohydrate–insulin model (CIM) emphasizes the role of refined carbohydrates and insulin responses in driving fat storage and subsequent increased intake. Ultra-processed foods often contain both sugars and fats, complicating attribution of metabolic effects to either nutrient per se versus energy density. Using a multi-nutrient approach and nutritional geometry (NG), the authors aim to disentangle the effects of fat, glucose, fructose, and their mixtures on energy intake, adiposity, glucose homeostasis, and liver metabolism, and to test whether outcomes align with EBM, CIM, or elements of both. Protein was fixed at 20% energy and fat varied (10–30%) to evaluate fat–sugar interactions across dietary contexts, including fat source (soy oil vs lard).

Background literature highlights that carbohydrates typically provide 45–70% of dietary energy and fats are the most energy-dense macronutrient. The single-nutrient focus in nutrition has fueled controversy over the roles of fats and sugars in obesity. The EBM implicates energy-dense, palatable, low-protein/fiber foods in excess caloric intake and weight gain, while the CIM implicates high-glycemic carbohydrates and insulin-driven nutrient partitioning to adipose tissue causing hyperinsulinemia, internal energy deprivation, and compensatory intake. Prior mouse NG work by the authors showed carbohydrate type matters: low protein–high carbohydrate diets are harmful when carbohydrate is mainly HFCS, but beneficial with resistant starch. Human studies suggest mixed sugar (50:50 fructose:glucose) can blunt satiety more than other ratios, and co-ingestion of fructose and glucose enhances hepatic de novo lipogenesis relative to either sugar alone. Conversely, some rodent studies found that altering fat–carbohydrate ratios in isocaloric diets does not change ad libitum energy intake, and fat source (lard vs plant oils) may have limited effects on weight/composition, suggesting total fat content can be more influential than saturation level.

Study design and animals: Male C57BL/6J mice (n=245), 8 weeks old at diet start, were housed 4 per cage under 12-h light/dark, 24–26°C, 44–46% humidity, and fed ad libitum for 18–19 weeks. Metabolic assessments occurred at weeks 5–6, 12–14, 15–16, and terminal sampling at weeks 18–19. Ethics approval: University of Sydney (protocol 2018/1362). Diets: Eighteen isocaloric diets (~14.3 kJ/g net metabolizable energy) based on AIN93G with fixed protein at 20% energy. Fat energy was set at 10%, 20%, or 30% (with corresponding carbohydrate at 70%, 60%, 50%). Isocaloricity was maintained by adjusting cellulose. Carbohydrates: 30% of carbohydrate energy from native wheat starch; remaining 70% from glucose, fructose, or mixtures (fructose:glucose 100:0, 75:25, 50:50, 25:75, 0:100). Fifteen diets used soy oil as fat source; three used lard (all at 20% fat, 60% carbohydrate) across glucose, fructose, or 50:50 mixes. The design enabled systematic evaluation of sugar composition, fat level, and fat source via NG. In vivo assessments: Weekly body weight and food intake. Body composition (fat and lean mass) by EchoMRI at weeks 5–6 and 12–14. Metabolic cage studies (Promethion, Sable Systems) on 5–7 mice/diet for 48 h after 24 h acclimation measured VO2, VCO2, respiratory quotient (RQ), energy expenditure, and physical activity; energy expenditure normalized to lean mass; data analyzed with CalR. Glucose tolerance test (GTT): Oral gavage 2 g glucose/kg lean mass after 6 h fast; blood glucose measured at 0, 15, 30, 45, 60, 90 min; AUC computed; fasting and 15-min post-gavage insulin measured (ELISA). Insulin tolerance test (ITT): Weeks 15–16; intraperitoneal 0.75 U insulin/kg lean mass; blood glucose at 0, 15, 30, 45 min; higher AUC indicates lower insulin sensitivity. Fasting glucose and insulin product used as a surrogate of insulin resistance (HOMA-IR-like). Circulating factors: Plasma insulin (Ultrasensitive Insulin ELISA, Crystal Chem) during GTT; plasma FGF21 (BioVendor) at study end; plasma triglycerides via clinical chemistry analyzer. Liver analyses: Histology (formalin-fixed, paraffin-embedded, H&E; blinded scoring 0–3 by three observers) for steatosis; hepatic triglycerides quantified via chloroform:methanol extraction and colorimetric assay (Triglyceride-GPO-PAP, Roche). Liver gene expression by RT-qPCR (Roche LightCycler; SYBR Green); housekeeping gene Rpl13a; targets included Khk (isoform C), Acly, Fasn, Scd1, Gpat3, Apob, Mcp1; ΔΔCt method. Statistics: Nutritional geometry analyses using generalized additive models (GAMs) with thin-plate splines in R (v4.1.1) to model responses over nutrient intake space; model validation via residuals; log-transformation as needed. Liver histology scores modeled by ordinal regression (proportional odds). Soy vs lard comparisons by ANOVA (GraphPad Prism). Data reported as mean ± s.e.m.; P<0.05 considered significant. Custom R scripts available on GitHub (https://github.com/Nidane/Sugar-Fat-Study).

- Energy intake: The dietary fructose:glucose ratio was the primary driver of intake. Diets with a 50:50 fructose:glucose mixture produced the highest energy intake; intake was lower with glucose or fructose alone. Lower dietary fat modestly increased early energy intake, but fat effects were not significant long-term. - Adiposity and body weight: Co-ingestion of fructose and glucose at 50:50 maximized body weight and fat mass; gonadal and inguinal white fat pads were heaviest with 50:50. Increasing fat intake blunted sugar-ratio effects, with high-fat diets inducing generalized adiposity regardless of sugar composition. - Energy expenditure and FGF21: Energy expenditure (normalized to lean mass) decreased with increasing fat intake; interscapular brown fat mass was highest with 50:50 fructose:glucose. Circulating FGF21 increased with higher total carbohydrate and lower fat intake, with glucose more potent than fructose at raising FGF21. RQ increased with higher glucose and lower fat intake. Absolute energy expenditure and physical activity did not differ significantly across diets. - Glucose homeostasis: A 50:50 fructose:glucose mixture reduced insulin sensitivity (higher ITT AUC and higher fasting glucose×insulin). Fat mass correlated with reduced insulin sensitivity (e.g., linear regression R² ≈ 0.46). High fat intake caused generalized impairment of insulin sensitivity irrespective of sugar ratio, and over time (12–14 weeks) high fat was more detrimental to fasting glycemia/insulinemia than high carbohydrate. The combination of high fat with 50:50 fructose:glucose produced the worst glucose tolerance (higher GTT AUC) and highest peak blood insulin after glucose load. - Liver metabolism: Co-ingestion of fructose and glucose (especially 50:50) increased hepatic fat deposition by histology and biochemical triglyceride content. Increasing dietary fat had minimal effect on hepatic fat accumulation. Hepatic Khk (isoform C) expression rose with higher fructose intake, most in low-fat/high-carbohydrate diets. DNL genes Acly and Fasn were higher with fructose; Scd1 and Gpat3 peaked with 50:50 mixture, aligning with maximal liver triglycerides in that group. Expression of lipogenic genes (Khk, Acly, Fasn) decreased with increasing dietary fat, consistent with suppression of DNL by dietary fat; Apob expression also decreased with higher fat, while inflammatory Mcp1 was highest at high fat. Plasma triglycerides were not significantly altered. - Fat source: Substituting soy oil with lard (at 20% fat, 60% carbohydrate) did not significantly affect body weight, fat mass, energy intake, insulin sensitivity, glucose tolerance, peak insulin, or liver triglycerides. The 50:50 sugar mixture remained the most obesogenic and metabolically adverse irrespective of fat source. - Overall: Both dietary sugars and fat can drive adverse metabolic outcomes depending on context: 50:50 fructose:glucose increased intake, adiposity, and liver fat at low–medium fat, while higher fat intake broadly worsened adiposity, insulin sensitivity, and glucose tolerance regardless of sugar mix.

The findings clarify how fat–sugar interactions shape metabolic health and help reconcile EBM and CIM. Supporting EBM, the highest energy intakes (on 50:50 fructose:glucose) led to the greatest adiposity, and higher fat intake produced generalized impairments in insulin sensitivity and glucose tolerance independent of sugar type. Contrary to strict CIM predictions, diets of 100% glucose did not yield the worst adiposity despite higher glycemic index. Supporting aspects of CIM, co-ingestion of fructose and glucose increased hepatic de novo lipogenesis and liver fat compared with either sugar alone, and carbohydrate-rich diets raised FGF21. Importantly, the dietary context modulated effects: at low–medium fat, sugar composition (50:50) had large impacts on intake and adiposity; at higher fat, sugar-specific differences diminished and fat content dominated outcomes. The absence of differences between soy oil and lard suggests total fat content is more influential than fat saturation in this setting. Together, these results suggest a unified view in which both energy intake and carbohydrate quality/insulin dynamics contribute to obesity-related phenotypes, contingent on macronutrient balance.

Using nutritional geometry in male mice on 18 isocaloric diets with fixed protein, the study shows that a 50:50 fructose:glucose mixture (as in sucrose/HFCS) maximizes energy intake, adiposity, and hepatic fat at low–medium fat intakes, whereas increasing dietary fat content drives generalized adiposity and worsens insulin sensitivity and glucose tolerance irrespective of sugar mix. Energy expenditure decreased with higher fat intake; FGF21 rose with carbohydrate intake (especially glucose). Fat source (soy oil vs lard) did not alter outcomes. These results indicate that both EBM and CIM capture valid mechanisms under different dietary contexts and support moving toward a unified model of obesity. Future research should apply NG to human cohorts, test liquid sugar formulations, include both sexes and different mouse strains, examine very high-fat/high–energy-density diets, evaluate hypothalamic/hedonic signaling, and perform controlled pair-feeding and thermoneutral studies to refine translational relevance.

- Simple sugars only; not compared with low–glycemic index complex carbohydrates. - Postprandial glycemic and insulinemic responses and downstream effects on clearance, lipid metabolism, and appetite were not directly measured. - Hypothalamic appetite and hedonic signaling were not examined. - Very high-fat or highly energy-dense diets (>45% fat) were not used. - Diets were kept isocaloric by adjusting cellulose, which could be a confounder in other contexts. - Fructose and glucose were provided only in solid diets; liquid sugar effects (notably more obesogenic) were not tested. - No pair-feeding experiments to fully dissociate nutrient effects from caloric intake. - Single sex (male) and single strain (C57BL/6J); not replicated in females or other strains. - Experiments were not conducted under thermoneutral conditions.