Puberty Status Modifies the Effects of Genetic Variants, Lifestyle Factors and Their Interactions on Adiponectin: The BCAMS Study

Figures

• 

  Figure 1. Changes in ln-adiponectin levels during puberty according to sex. Data were shown as mean and SE. Puberty status was reclassified into prepuberty (Tanner stage I), midpuberty (Tanner stage II-III) and postpuberty (Tanner stage ≥ IV). Difference between boys and girls in the same pubertal group was indicated *P < 0.001 and N.S. as non-significant value after adjusted for age. Difference between diverse pubertal groups of the same sex was indicated as different letters with P < 0.05 after adjusted for age, that was, difference between pubertal groups of the same sex with the same letter were not statistically significant.

• 

  Figure 2. Genetic and lifestyle associations with adiponectin levels according to the different puberty stages. Figure (A) shows the effect (histograms) and SEs (error bar) of SNPs on adiponectin levels (% change in adiponectin levels per effect allele) at different puberty stages. Figure (B) shows the effect (histograms) and SEs (error bar) of lifestyle factors on adiponectin levels (% change in adiponectin levels when walking to school for the transportation variable and % change in adiponectin levels per assigned score increase for other variables) at different puberty stages. The results were adjusted for age, sex, and BMI. *P < 0.05; ** P < 0.05/6 = 0.008 (after Bonferroni correction); *** P < 0.001.

• 

  Figure 3. Effects of the interactions of the weighted genetic score with lifestyle factors on the % change in adiponectin levels according to the different puberty stages. The figure shows the effect and 95% CI of the interactions of wGPSall and wGPSno CDH13 with lifestyle factors on adiponectin levels (% change in adiponectin levels per wGPSall or wGPSno CDH13 per diet score or per lifestyle factors assigned score increase) in the entire population of children at different puberty stages. The results for the diet items were adjusted for age and sex. The results for the diet score were adjusted for age, sex, and activities (including exercise and transportation type). The results for the activities were adjusted for age, sex, and diet score; * P < 0.05.

• 

  Figure 4. The associations of diet and exercise with adiponectin levels according to the categories of genetic risk. The figure shows the main effects (histograms) and SEs (error bars) of diet score and exercise on adiponectin levels (% change in adiponectin levels per assigned score increase for diet and exercise) according to the genetic risk for decreased adiponectin levels. The data for diet scores were adjusted for age, sex, transportation type and exercise. The data for exercise were adjusted for age, sex and diet score. As we reported an interaction between diet score and wGPSall and an interaction for exercise with wGPSno CDH13, we used wGPSall to identify the genetic modification of diet effect and used wGPSno CDH13 to identify the genetic modification of the exercise effect. Genetic risk was divided into low genetic risk (wGPSno CDH13 or wGPSall > mean +1SD), intermediate genetic risk (wGPSno CDH13 or wGPSall ≥ mean -1SD but ≤ mean+1SD) and high genetic risk (wGPSno CDH13 or wGPSall < mean-1SD). N.S. means the effect of exercise at low genetic risk is not significant.

• 

  Figure 5. Adiponectin levels according to genetic risk and categories of diet and exercise. The figure shows multivariable-adjusted means (histograms) and SEs (error bar) of the natural logarithm transformed adiponectin levels according to the categories of lifestyle and genetic risk for decreased adiponectin levels. The P-values are the results of an ANCOVA comparing the adiponectin levels among the genetic risk groups. Data for exercise were adjusted for age, sex, and diet score, while data for diet were adjusted for age, sex, and exercise frequency. As we reported an interaction between diet score and wGPSall and an interaction for exercise with wGPSno CDH13, we used (A) wGPSall to identify genetic risk categories in the diet subgroups and (B) used wGPSno CDH13 to identify genetic risk categories in the exercise subgroups. Genetic risk was divided into low genetic risk (wGPSno CDH13 or wGPSall > mean + 1SD), intermediate genetic risk (wGPSno CDH13 or wGPSall ≥ mean - 1SD but ≤ mean + 1SD) and high genetic risk (wGPSno CDH13 or wGPSall < mean - 1SD). Similarly, diet and exercise were divided into high (diet score> mean + 1SD; exercise frequency ≤ 2 times per week), intermediate (diet…

• 

  Figure 6. Hypothesis for the changes in adipocyte metabolism during puberty. The figure shows our hypothesis, which is that the modification effect of puberty on the SNP-adiponectin association is based on the development of adipose tissue during puberty. The SNPs shown in this figure are SNP1: ADIPOQ-rs10937273; SNP2: ADIPOQ-rs6773957; SNP3: CDH13-rs4783244; SNP4: WDR11-FGFR2-rs3943077; and SNP5: PEPD-rs889140. The puberty categories that we used in the current study were as follows: prepuberty (Tanner stage I), midpuberty (Tanner stage II-III), and postpuberty (Tanner stage ≥ IV). Prepuberty is a period during which the processes of puberty have not yet been activated completely. Midpuberty is the phase during which puberty has been activated but is not finished. Postpuberty is the stage at which the processes of puberty are nearly completed, and adolescents at postpuberty are similar to adults. The development of adipose tissue includes hyperplasia, related to WDR11-FGFR2-rs3943077, and the hypertrophy of adipocytes. The number of adipocytes increases quickly at midpuberty but remains relatively consistent after postpuberty. The size of adipocytes increases during puberty, which makes…