Predictive markers of obesity and glucose metabolism dysfunction in adult common marmosets (Callithrix jacchus)

Abstract

Objective: Characterize the effects of obesity on common marmoset glucose metabolism and develop predictive markers of glucose metabolism dysfunction.

Methods: Body size, weight, lean mass, fat mass, %fat, resting energy expenditure (REE), and glycosylated hemoglobin (HbA1c) were measured on 51 adult marmosets. Physical activity was assessed using actimeter collars (n = 50). A body mass-per-length parameter (BML) was constructed. Animals were classified as without obesity or with obesity (%fat >10%) and by the age they obtained maximum weight (Maxwt). Correlation, MANOVA, and binary logistic regression were used to examine relationships between parameters; path analysis to explore directional relationships.

Results: Body fat and BML were correlated (r = 0.565, p < 0.001). Both were correlated with HbA1c (r = 0.658; r = 0.764, p < 0.001). Activity was negatively correlated with %fat and REE (r = -0.437, p = 0.002; r = -0.473, p < 0.001). REE was correlated with %fat, BML, and HbA1c (r > 0.5, p < 0.001). Marmosets with obesity were more likely to have elevated HbA1c (>5.7%; odds ratio = 8.25, p = 0.003). BML above 3.4 g/mm predicted obesity (OR = 6.25 [95% CI 1.62-24.02], p = 0.008) and high HbA1c (OR = 29.47 [95% CI 6.21-139.72], p < 0.001). Early Maxwt predicted increased fat mass (F = -0.476, p = 0.015) and high %fat (F = -0.084, p = 0.014).

Conclusion: Both %fat and BML were markers for high HbA1c. Early maximum adult weight predicts increased adiposity and risk of glucose dysfunction.

Fig. 1. The relationship between body fat% and body mass divided by suprasternal-pubic length (BML).

The two marmosets with diabetes indicated by open circles. BML explained 43.1% of the adjusted variance in body fat%. For every one-unit increase in BML (g/mm), the value of body fat% was predicted to increase by 6.38 (95% CI 4.32–8.44) p < 0.001.

Fig. 2. The relationship between body fat% and glycosylated hemoglobin (HbA1c).

The two marmosets with diabetes indicated by open circles. Body fat% explained 26.3% of the adjusted variance in HbA1c. For every one-unit increase in body fat%, the value of HbA1c was predicted to increase by 0.169 (95% CI 0.091–0.248) p < 0.001.

Fig. 3. The relationship between HbA1c and body mass divided by suprasternal-pubic length (BML).

The two marmosets with diabetes indicated by open circles. BML explained 28.3% of the adjusted variance in HbA1c. For every one-unit increase in BML (g/mm), the value of HbA1c was predicted to increase by 1.678 (95% CI 0.937–2.42) p < 0.001.

Fig. 4. Node diagram for the path analysis model.

Marmosets that achieved maximum body mass earlier, ended up with a larger body size (B = −0.25, p < 0.001). Larger body size predicted higher body mass in proportion to body size (B = 0.02, p = 0.003), and lower activity (B = −1.97, p < 0.001). In turn, higher body mass in proportion to body size (B = −3.81, p = 0.002) and lower activity (B = 0.04, p = 0.01), predicted lower REE per gram of fat mass. Additional statistics available in Table 2.