Page 82 - The prevalence of the Val66Met polymorphism in musicians: Possible evidence for compensatory neuroplasticity from a pilot study

Validation of genes in childhood obesity | 5

rs17782313, MAF was lower than in both populations. Discussion
Genotype and allele frequencies were not different be-
tween ethnic groups (whites and non-whites) in our The advent of GWAS has brought several advances to
sample (Supporting Information Table S1). No differ- the identification of new susceptibility loci for several
ences were also observed regarding either anthropo- complex characteristics, such as obesity. However,
metric or dietary measurements among ethnicities. these studies have been focused in populations from
Therefore, the individuals from both ethnicities were North America and Europe, while the South American
grouped in further analyses. populations remained unexplored in this context. ORIGINALRESEARCH
Mean weight and length were compared among In the present study, we were able to validate the as-
different genotypes at birth, and other anthropomet- sociations of five of the 10 analyzed gene variants in a
ric and dietary variables were compared at 1year cohort of children from the south of Brazil followed from
and 3.5 years old (Table 3 and Supporting Informa- birth until 3.5years of age. The TMEM18, BDNF and
tion Tables S2, S3 and S4). Significant associations NEGR1 gene variants investigated were associated with
of individual SNPs located at three of the nine genes BMI-Z and subcutaneous fat in our setting, and the
investigated with anthropometric measurements SEC16B SNP, although not associated with these phe-
were found, and they are summarized in Table 3. notypes, was shown to be associated with higher daily
For TMEM18 rs6548238, we could observe that intake of energy provided from carbohydrates and lipids
C/C homozygotes presented higher BMI-Z com- at the age of 3.5years. Our results also show that the
pared with T-allele carriers (P =0.007; Pcorr =0.049) BDNF variants effect begins earlier, at the age of 1year,
at the age of 3.5 years. Carriers of the minor allele than the effect of the NEGR1 and TMEM18 gene vari-
(A) of BDNF rs6265 presented lower BMI-Z at 1 year ants, which seem to begin a little later, being detected
(P =0.026; Pcorr=0.091) and sum of skin-folds at after the age of 3.5years.
3.5 years old (P =0.011; Pcorr=0.077) than G/G ho- A more recent GWAS, published in 2013 by Wheeler
mozygotes. At this same age, the minor T-allele of et al. (19), performed analyses with a high number of
BDNF rs10767664 was associated with a reduction severely obese children and controls, identified nine
in the sum of skin-folds (P =0.009; Pcorr =0.063) genome-wide significant signals in eight loci, named:
when compared with the A/A genotype. For the FTO, MC4R, TMEM18, NEGR1, PRKCH, LEPR,
NEGR1 rs2815752, the A/A homozygotes presented PACS1 and RMST, the last four being detected in
higher BMI-Z (P =0.042; Pcorr =0.098) and sum of GWAS for the first time. Although their sample is not to-
skin-folds at 3.5 years old (P =0.023; Pcorr =0.098). tally comparable with ours, because their study focused
The post hoc Tukey test revealed that these differ- on children in the extreme tail of the BMI distribution
ences were due to a higher BMI and amount of sub- (BMI-Z more than 3 SD from the mean), we were able
cutaneous fat in the A/A homozygotes than in the to replicatehalfofthese findings in the present
A/G heterozygotes. Longitudinal analyses demon- (TMEM18 and NEGR1) and in previous studies per-
strated that individuals bearing the A/A genotype formed with the same sample of children (FTO (20)
had also a lower variation in BMI from 1 year to and LEPR (21)). Of course, we must take in consider-
3.5 years (P =0.030; Pcorr =0.098). All results ation that the specificgenevariantswerenot thesame,
showed earlier remained significant when corrected although located in or near the same genes, and that
for multiple comparisons using an FDR= 0.10. three of them (PRKCH, PACS1 and RMST)werenot
Only one of the investigated genes presented as- evaluated by us.
sociations with parameters related to dietary pat- From the 10 polymorphisms included in the present
terns: SEC16B rs10913469 (Supporting Information study, we identified significant associations for NEGR1
Table S2). The T/T genotype was associated with rs2815752, TMEM18 rs6548238, BDNF rs6265 and
higher intake of total energy, carbohydrates and lipids rs10767664 with anthropometric phenotypes (Table 3)
compared with C-carriers at 3.5 years old (P =0.006, and for SEC16B rs10913469 with dietary parameters
Pcorr =0.048; P =0.016, Pcorr= 0.064; P =0.036, (Supporting Information Table S2). A common feature
Pcorr =0.096; respectively). between these genes is that they are expressed or
We did not find any significant associations between known to act in the central nervous system, highlighting
SH2B1, HOXB5, KCTD15, OLFM4, SEC16B, MC4R the importance of the neuronal component to the obe-
and anthropometric parameters (Supporting Informa- sity susceptibility, although they have also peripheral
tion Table S3), neither for NEGR1, SH2B1, HOXB5, functions related to adipose tissue (6,22).
KCTD15, BDNF, TMEM18, OLFM4 and MC4R gene The associations reported herein were directionally
variants with dietary phenotypes (Supporting Informa- consistent with results from GWAS: TMEM18
tion Table S4). rs6548238 C-allele (23,24), BDNF rs6265 G-allele (25),

© 2016 World Obesity. Pediatric Obesity ••, ••–••