Plasma leptin levels are related to body composition, sex, insulin levels and the A55V polymorphism of the UCP2 gene

Objective: Circulating leptin levels show a high degree of individual variability even after the main effect of body fatness is accounted for. We therefore wanted to determine the roles of variation in body composition, age, sex and polymorphisms of the UCP2 gene and promoter region on levels of circulating leptin.

Subjects: One hundred and fifty Caucasian subjects, which represented a cross-section of the population from NE, Scotland, were recruited.

Measurements: Body composition was measured using dual X-ray absorptiometry. Fasted circulating leptin, insulin, T3 and T4 levels were measured, and all individuals were genotyped for the UCP2 polymorphisms A55V, -866G > A and exon-8 ins/del. ==Results: The results indicate that circulating leptin was significantly related to sex and principle component (PC) scores representing overall adipose tissue mass and a second representing the contrast of central to peripheral bone mineral content. Residual leptin was associated with the A55V polymorphism (P < 0.001) explaining 11.3% of the residual variance. There was a marginal effect associated with exon-8 ins/del (P = 0.045) explaining 4.4% of the residual variance in leptin. Log(e) transformed circulating fasting insulin was related to PC scores representing general adiposity and sex. Residual Loge insulin was associated with the A55V and exon-8 ins/del polymorphisms explaining 5.7% (P = 0.015) and 5% (P = 0.026) of the residual variation, respectively. The -866G > A polymorphism was not significantly associated with residual leptin or insulin. Leptin and insulin were significantly (P = 0.007) correlated. Statistically removing the effect of insulin on leptin still showed association between leptin and A55V (P = 0.002). Removing the effect of leptin on insulin, the A55V polymorphism was no longer significant (P = 0.120). After accounting for the correlation between insulin and leptin, the exon-8 ins/del was no longer significant for residual leptin (P = 0.119) or Loge insulin (P = 0.252).

Conclusion: These data suggest that the A55V polymorphism directly affected the levels of leptin but not via an effect on insulin.