Obesity and obesogenic growth are both highly heritable and modified by diet in a nonhuman primate model, the African green monkey (Chlorocebus aethiops sabaeus)

Abstract

Objective: In humans, the ontogeny of obesity throughout the life course and the genetics underlying it has been historically difficult to study. We compared, in a non-human primate model, the lifelong growth trajectories of obese and non-obese adults to assess the heritability of and map potential genomic regions implicated in growth and obesity.

Study population: A total of 905 African green monkeys, or vervets (Chlorocebus aethiops sabaeus) (472 females, 433 males) from a pedigreed captive colony.

Methods: We measured fasted body weight (BW), crown-to-rump length (CRL), body-mass index (BMI) and waist circumference (WC) from 2000 to 2015. We used a longitudinal clustering algorithm to detect obesogenic growth, and logistic growth curves implemented in nonlinear mixed effects models to estimate three growth parameters. We used maximum likelihood variance decomposition methods to estimate the genetic contributions to obesity-related traits and growth parameters, including a test for the effects of a calorie-restricted dietary intervention. We used multipoint linkage analysis to map implicated genomic regions.

Results: All measurements were significantly influenced by sex, and with the exception of WC, also influenced by maternal and post-natal diet. Chronic obesity outcomes were significantly associated with a pattern of extended growth duration with slow growth rates for BW. After accounting for environmental influences, all measurements were found to have a significant genetic component to variability. Linkage analysis revealed several regions suggested to be linked to obesity-related traits that are also implicated in human obesity and metabolic disorders.

Conclusions: As in humans, growth patterns in vervets have a significant impact on adult obesity and are largely under genetic control with some evidence for maternal and dietary programming. These results largely mirror findings from human research, but reflect shorter developmental periods, suggesting that the vervet offers a strong genetic model for elucidating the ontogeny of human obesity.

Figure 1

Chronic obesity outcome in adults assigned to heavy vs light growth clusters. Chronically obese individuals are defined as having a waist circumference above 40.5 cm for more than three consecutive measurements.

Figure 2

Measurements and NLME logistic growth model output for (I) body weight (BW) and (II) crown-to-rump length (CRL) in the VRC. Color-coding indicates sex and growth cluster assignment: dark blue=heavy males, light blue=light males, dark yellow=heavy females, light yellow=light females. In plot (a), each thin line connects individual measurement points for a single vervet, while thicker trend lines represent the average growth model for each sex/cluster. The boxplots show the mean and interquartile ranges of random effects deviations from the population average growth model, divided by cluster, for each growth parameter—(b) θ₁, the asymptote of growth, measured in kg; (c) −θ₂/θ₃, the midpoint of growth, measured in years; and (d) θ₃, the growth rate constant, measured in years⁻¹—that describe individual NLME logistic growth models grouped by sex/cluster and color-coded using the same system as (a). Parameter values are derived by adding random effects of subject identity to the mean parameter values for each sex/cluster.

Figure 3

Mean and standard error of residual differences by sex in age-adjusted (a) BW and (b) BMI for individuals fed a Standard diet (in black) and those who experienced a shift to ID (in gray) during (i) gestation (maternal shift to HF while gestating that individual), (ii) during the first 2 years after birth (PN1), (iii) during the subsequent 3 years after birth (PN2), and (iv) during adulthood. Residual values were attained after regressing out significant covariates, here including age and growth cluster assignment.

Figure 4

Results of the multipoint linkage analysis for all traits across the vervet genome (autosomal chromosomes with suggestive results only). Each gray horizontal line represents genome-wide suggestive (LOD ⩾ 1.9) and genome-wide significant (LOD ⩾ 3.3) levels.