MC4R and ENPP1 gene polymorphisms and their implication in maternal and neonatal risk for obesity

The aims of this study were to establish the role of MC4Rrs17782313 and ENPP1rs1044498 gene polymorphisms on pre-pregnancy BMI and the newborn’s status. We performed a cross-sectional study on 185 mothers and their offspring. The groups were divided into: control group- underweight or normal mothers with BMIinitial < 25 kg/m2 (n1 = 134) and study group-overweight/obese mothers with BMIinitial ≥ 25 kg/m2 (n2 = 51). All subjects underwent demographic, anthropometric, paraclinical, bioimpedance and genetic parameters. We found association between initial BMI and gestational weight gain (GWG), and a higher frequency of excessive GWG in overweight/obese women (p = 0.037). Higher values of anthropometric and bioimpedance parameters were observed in overweight/obese versus underweight/normal women. The MC4R rs17782313 and ENPP1 rs1044498 variant genotypes had an increased risk of pre-pregnancy overweight (OR = 1.41; 95% CI:[0.72; 2.78]; OR = 1.34; 95% CI:[0.65; 2.75]). The newborns from mothers with excessive GWG had a higher birth weight (BW) (p = 0.001). Higher MUAC values were noticed in newborns with MC4R rs17782313 wild-type genotype. Also, BW was correlated with GWG status smoking in pregnancy, gestational age and neonatal ENPP1rs1044498 variant genotype (p = 0.026). Our study pointed out the role of MC4R rs17782313 and ENPP1 rs1044498 genotypes in obesity determinisms in mothers and their newborns in correlation with BMI, MUAC, TST and bioimpedance parameters.

Recently, other genes that are thought to have an impact in the development of obesity were discovered, such as: melanocortin receptor (MCR) MC4R, MC3R, MC2R, and also nucleotide pyrophosphatase/phosphodiesterase (ENPP1). The variant rs17782313 of MC4R is associated with obesity in both children and adults, regulating the control of energetic balance 15 . The ENPP1 gene inhibits the sensitivity of insulin receptor, being considered a gene with a potential role in determining insulin-resistance and type 2 diabetes mellitus 16 .
Therefore, the aims of this study were to establish (i) the association between both MC4R rs17782313 and ENPP1 rs1044498 gene polymorphisms and pre-pregnancy BMI and (ii) the impact of these 2 gene polymorphisms on excessive, respectively insufficient gestational weight gain, but also (iii) to establish if the MC4R rs17782313 and ENPP1 rs1044498 single nucleotide polymorphisms (SNPs) were multivariate predictors correlated with birth weight, neonatal mid-upper-arm circumference and tricipital skinfold thickness in newborns.

Results
The demographic description of the studied groups regarding anthropometric, clinical, paraclinical and bioimpedance characteristics. Of all 225 mother-newborn couples that were examined, after applying the exclusion criteria, only 185 remained in the present study.
The mean age of mothers included in the study was 28.10 ± 5.80 years, over 50% of them having superior level education, 22.2% with education level <8 years, 17.8% with educational level between 9 and 12 years, and only 5.9% had never frequented school.
Most of the mothers included in our study were primipara (56.2%), with a GWG mean value of 17.1 ± 6.2 kg. Of all 185 pregnant women 11.4% were smokers, 5.90% presented preeclampsia and only 2.7% were diagnosed with gestational diabetes ( Table 1).
The comparison between maternal and neonatal characteristics reported to the body mass index at the beginning of the pregnancy is presented in Table 1. Thus, we found a significant association between initial BMI and maternal age (MAge) (p < 0.05), the overweight/obese mothers having a median age higher than those included in the control group.
We encountered a significant association between initial BMI and GWG, the frequency of excessive GWG being higher in overweight or obese women (72.5% versus 59%; p = 0.037) as compared to control group. Moreover, the incidence of diabetes mellitus was higher in obese mothers (7.8% versus 0.7%; p = 0.021).
Regarding anthropometric and bioelectrical impedance analysis (BIA) parameters, we found a significant difference between overweight/obese and underweight/normal weight women, identifying higher values of these parameters in overweight/obese women (Table 1).
Gestational weight gain status and maternal gene polymorphisms. We found no statistically significant associations between MC4R rs17782313 variant genotype and increased weight/obesity before pregnancy (Table 2). After adjusting for maternal covariates as MAge, education level and parity, the presence of MC4R rs17782313 variant genotype (CT + CC) was associated with an increased risk of pre-pregnancy increased weight (adjusted OR = The studied genotype frequencies of the two gene polymorphisms were consistent with Hardy-Weinberg equilibrium in the studied sample and GWG status (p > 0.05).
Neonatal gene polymorphisms correlated with birth weight, mid-upper-arm circumference and tricipital skinfold thickness. In the studied sample (n = 185 newborns), the mean BW was 3368 g ± 448, with 9.7% (18 cases) above 4000 grams. We found no significant difference between the means of BW in newborns whose mothers had an initial BMI ≥25 kg/m 2 (overweight/obese mothers) in comparison to those from mothers with an initial BMI <25 kg/m 2 . BW, middle upper arm circumference (MUAC) and tricipital skinfold thickness (TST) were significantly associated with maternal GWG status (one-way ANOVA, F(2; 182) = 8.19, p < 0.001). In the post-test analysis, the mean BW was significantly different in newborns whose mothers had insufficient or normal GWG in comparison to those from mothers with excessive GWG (Tukey test, p = 0.001). Therefore, the mothers with excessive GWG had newborns with higher BW (Table 3).
Our study showed a significant difference regarding the distribution of MUAC values in newborns with MC4R rs17782313 variant genotype in comparison to those with normal genotype, higher values being noticed in the newborns with normal genotype ( Table 3).
As a result of a linear multiple regression analysis, BW was significantly positively correlated with GWG status (p = 0.007), gestational age (GA) (p < 0.001) and neonatal ENPP1 rs1044498 variant genotype (p = 0.026), while negatively correlated with smoking during pregnancy (p < 0.001). BW was positively correlated with excessive GWG status. Thus, for each unit increment in GWG status, BW increased by 113 g. The presence of neonatal variant ENPP1 SNP was correlated with an increase in BW by 150 g. Moreover, an increase in BW by 132.1 g was www.nature.com/scientificreports www.nature.com/scientificreports/    www.nature.com/scientificreports www.nature.com/scientificreports/ observed per gestational week. The effect of GWG status, neonatal ENPP1 rs1044498 and MC4R rs17782313 variant genotypes, adjusted according to important covariates (as mother's age, educational level, smoking during pregnancy and gestational age at delivery) was illustrated in Fig. 1. Neonatal MUAC was significantly correlated with GWG status (p = 0.015), smoking in pregnancy (p = 0.024) and gestational age (GA) (p = 0.001). There was a tendency towards statistical association between neonatal ENPP1 rs1044498 variant genotype and MUAC (p = 0.092). Regarding neonatal TST there was a significant correlation with GWG status (p = 0.001), and neonatal ENPP1 rs1044498 variant genotype and MUAC (p = 0.045), the variant genotype being correlated with an increment in TST.

Discussions
Predictors for maternal weight gain. The studies reported in literature proved that maternal weight at the beginning of the pregnancy and GWG are important parameters that might result in diabetes mellitus, obesity and poor birth outcomes, with different obstetrical complications 2,3 . In addition, Farah et al. proved the fact that an excessive GWG leads to an increased neonatal BW with a subsequent risk of faster weight gain and even childhood obesity, but also other complications 4 . Similarly, in our study, BW was higher in newborns from overweight/ obese women (3.5 ± 0.5) compared to underweight/normal weight group (3.3 ± 0.5).
Excessive GWG increases the risk for hypertension 17 , hypertriglyceridemia, obesity, type 2 diabetes mellitus, insulin resistance 18 , and is associated with increased BMI, TST, and waist circumference. Similarly, in our study, we noticed that excessive GWG and gestational diabetes incidence were higher in overweight or obese women (p = 0.037/p = 0.021).
Gallagher et al. proved that mothers with increased weight or obesity presented higher GWG and their newborns had a higher BW, but also a greater fat-free mass (FFM) evaluated by BIA 19 . Farah et al. 4 and Butte et al. 20 noticed a direct relationship between FFM and fat mass (FM) for GA under 37 weeks, with accurate predictions on BW. In our study, all anthropometric (MUAC, TST) and BIA (FM, MM, bone mass -BM, total body water -TBW and also basal metabolism rate -BMR) measurements were significantly higher in overweight/obese women than in underweight/normal weight women (p < 0.0001), similar to another study of the same authors 21 . Even though other studies 22 proved that BW was correlated with total body water and FFM, but not with FM, in our study we noticed that BW was correlated with both FFM and FM.
There are studies proving that BW is correlated with GA at delivery 4,21 , but our study failed in proving this correlation. Contradictory data regarding the correlation between BW and smoking were reported by different studies. Thus, Farah et al. found a direct relationship between BW and smoking, and parity 4 , respectively, while Mărginean et al. in a previous study 21 pointed out a reverse relationship between these parameters. The present www.nature.com/scientificreports www.nature.com/scientificreports/ study also underlined the negative impact of smoking during pregnancy on the newborn's BW. Even though certain studies 11,21 established a correlation between arterial hypertension (AHT) and higher GWG, in our study we did not notice this fact, probably due to the small number of cases with AHT.
Among the multiple MCR genes that proved to be involved in the etiology of obesity, MC4R is associated with monogenic obesity, regardless of age 15 . Therefore, the MC4R rs17782313 gene polymorphisms was associated in certain studies with obesity in both adults and children 15,23 . Thus, Bordoni et al. 24 observed that the C/C genotype of MC4R rs17782313 gene polymorphism was associated with higher BMI and obesity risk in young Italian population. Similarly, Loos et al. emphasized that the C allele of the same gene was associated with an increased risk for developing obesity 15 , while Garcia-Solis et al. underlined a positive correlation between FTO rs9939609 homozygotes and MC4R rs17782313 heterozygotes and increased risk for both obesity and high blood pressure values in Mexican school-aged children 25 . Lazopoulou et al. 26 proved that MC4R rs17782313 C allele was associated with a higher BMI and BW, but also that three or more high-risk alleles of the combined FTO and MC4R genotypes result in a 4 folds increased risk for obesity in Greek children and teenagers. Wu et al. noticed on Chinese children an association between MC4R rs17782313 SNP and adiponectin 27 .
Also, Mejía-Benítez et al. 28 found associations between obesity risk, BMI and several SNPs, among which we recall MC4R rs17782313 and ENPP1 rs7754561. Contrariwise, Albuquerque et al. 29 , Martins et al. 30 , similar to our study, identified no significant associations between MC4R rs17782313 SNP and GWG, maternal body weight, or other characteristic parameters for obesity.
ENPP1 gene polymorphisms are associated with obesity in children, inhibiting insulin receptors and resulting in a higher incidence of type 2 diabetes mellitus and a lower tolerance for glucose in carriers of ENPP1 K121Q (rs1044498) and A/G_1044TGA haplotypes 31,32 . Contrariwise, Morandi et al. stated that the Q121 variant allele of ENPP1 K121Q (rs1044498) SNP presents a protective role for obesity in a study performed on 453 Italian children 33 . Other authors also underlined the relationship between the Q allele of ENPP1 K121Q (rs1044498) SNP and type 2 DM or obesity 31,34 . In contradiction to the afore-mentioned studies, we found no correlations between ENPP1 rs1044498 variant genotype and GWG status. Our findings are similar to those of Lyon et al. 35 who did not notice any correlation between this SNP and obesity or diabetes mellitus.
Predictors for newborn's birth weight. One of the most important predictors for neonatal BW was proved to be excessive GWG. Therefore, the study of Ferrari et al. proved the fact that newborns coming from mothers with excessive GWG presented a higher BW in comparison to those whose mothers presented normal GWG 36 . The same study also underlined that the chances for these newborns to be macrosomic were 50% higher. Similarly, in our study, BW, MUAC and TST were correlated with GWG (p < 0.001), meaning that the newborns that came from mothers with excessive GWG had a higher weight than those from mothers with insufficient or normal GWG (p = 0.001). Some studies emphasized that newborns with higher BW are more predisposed to develop obesity and increased weight, but also more prone to cardiovascular and metabolic complications 37,38 . Previous studies performed by Farah et al. 4 and Marginean et al. 21 pointed out no correlation between initial BMI of the pregnant women and BW, similarly to the present study in which despite the fact that BW was with 9.7% higher in newborns from mothers with initial BMI ≥25 kg/m 2 versus normal BMI, we found no significant difference between the two groups.
The studies of Bordoni et al. 24 and Loos et al. 15 pointed out the association between the variant allele C of MC4R rs17782313 and both increased BMI and obesity risk in teenagers. Also, the study of Lazopoulou et al. 26 proved that BW and BMI were associated with the C allele of MC4R rs17782313, while Garcia-Solis et al. 25 established a significant correlation between heterozygous MC4R rs17782313 carriers and both blood pressure and obesity risk in Mexican school-aged children 25 . Moreover, similar findings were underlined by other studies performed on Chinese people revealing a correlation between the MC4R rs17782313 SNP and adiponectin in Chinese obese children 27 . On the contrary, in our study, we noticed an association of the wild-type MC4R rs17782313 gene polymorphism and higher MUAC values in newborns.
It is well-documented that ENPP1 gene is associated with obesity in pediatric patients resulting in an increased risk for glucose intolerance and type 2 diabetes mellitus 31 . Similarly, recent studies 32,39 proved that the variant allele of ENPP1 rs104449 increases the risk for obesity and type 2 diabetes mellitus, but no study established the role of this SNP in the newborn's BW and pregnant woman's nutritional status. On the other hand, Morandi et al. noticed that variant ENPP1 rs104449 owns a protective role for developing obesity 33 . In exchange, our study proved a positive association between the same variant gene and GWG status, GA and an increase of 150 g in the mean value of the newborn's BW. Moreover, a negative association was noticed regarding smoking during pregnancy and BW. Also, neonatal MUAC and TST were correlated with GWG and neonatal variant ENPP1 rs1044498.
Strengths and limitations of our study. The strengths of this study are the uniform population taken into consideration, the assessment of increased weight risk in mother-newborn couples, evaluating obesity risk in newborns and their correlations with other BIA and anthropometric parameters, and also evaluating the role of these genes in the determinism of BW. Some of the limitations of our study consist in the small number of mothers and newborns, the lack of assessment of other cytokines involved in the determinism of obesity, the type of diet, or the lack of BIA measurements in newborns.
There are a few data in literature that are meant to establish the role of these polymorphisms in the determinism of BW, therefore, we consider this study as a pilot one concerning the determinism of MC4R rs17782313 and ENPP1 rs1044498 variant genotypes on GWG status and also on the mother-newborn couple.

Conclusions
In our study we found a significant association between initial BMI and GWG, the frequency of excessive GWG being higher in overweight or obese women. The MC4R rs17782313 and ENPP1 rs1044498 variant genotypes presented an increased risk in pre-pregnancy increased weight. Also, BW, MUAC and TST were significantly associated with GWG status. Higher MUAC values were noticed in newborns with wild-type MC4R rs17782313. The results of multivariate analysis showed that BW was positively correlated with GWG status, GA and neonatal variant ENPP1 rs1044498, while negatively correlated with smoking during pregnancy.
Our study pointed out the role of MC4R rs17782313 and ENPP1 rs1044498 in maternal and neonatal obesity risk in correlation with BMI, MUAC and TST and BIA, which could be useful for diagnosing obesity in mother-newborn couples.

Methods
Ethics approval and informed consent. The approval for this research was granted by the Ethics Committee of the University of Medicine, Pharmacy, Sciences and Technology from Târgu Mureș (No 26/7 th of April 2017), and it was performed according to the principles of the Helsinki Declaration. Informed consent was obtained from all individual participants included in the study (the mothers signed informed consent for them and their newborns).
Study sample selection. We developed a cross-sectional study on 185 mothers and their offspring, admitted in a Tertiary Hospital from Romania, in an Obstetrics Gynecology Clinic between May 2017 and October 2017. The groups were divided into: control group -the underweight or normal mothers with Initial BMI <25 kg/m 2 (n 1 = 134) and study group -the overweight/obese mothers with initial BMI ≥25 kg/m 2 (n 2 = 51).
The inclusion criteria were: MAge above 18 years and single pregnancies. The exclusion criteria were: intrauterine growth retardation as a result of congenital malformations, chronic disorders, intrauterine infections; the lack of complete clinical, paraclinical, anthropometric and genetic data, but also mothers who refused to sign the informed consent prior to the inclusion in the study.
Variables of interest. Body mass index (BMI), Middle upper arm circumference (MUAC), Tricipital skinfold thickness (TST), Fat mass (FM): We performed all measurements in both mothers and newborns. These measurements included: weight (kg), height (cm), MUAC and TST. Height was measured with a daily calibrated pedometer, being evaluated by SD (0.1-cm error). For MUAC determination, we measured the arm circumference with a tape measure calibrated in centimeters, at the midpoint between shoulder and elbow tips, while TST was assessed in the posterior area of the upper arm with a thickness caliper. BMI was calculated as the ratio between weight (kg) and squared standing height (m 2 ). According to Control Disease Center (CDC), a BMI between 25.0 and 29.9 classified mothers as overweight, while those with a BMI of 30.0 or higher were considered obese.
For the bioelectrical impedance analysis (BIA), we used a Tanita BC-420 MA body composition analyzer (Tanita Corp, Tokyo, Japan). Weight was provided automatically with a 0.5 kg adjustment for the clothes weight, while height, sex, and age were typed manually.
Genetic variables. The DNA was isolated from fresh peripheral blood using the PureLink Genomic DNA kit (ThermoScientific) according to the manufacturer instructions. The concentration and the purity (A 260 /A 280 ) of the DNA was quantified by spectrophotometry (BioSpectometer basic, Eppendorf).
For genotyping of MC4R rs17782313 and ENPP1 rs1044498 we used TaqMan technology. In this respect we used the following single tub TaqMan SNP Genotyping assay formats: C__32667060_10 for rs17782313, and C__1207994_20 for rs1044498 (from ThermoFisher Scientific). Genotyping was performed according to the standard protocol for genotyping for 7500 Fast Dx Real-Time PCR System from Applied Biosystems. For genotypes interpretation the 7500 Fast Software v2.3 (Applied Biosystems) was used. All gDNA samples isolated from whole blood collected from children included in patient and control groups were successfully genotyped.
Statistical analysis. Qualitative variables were described by relative frequency while description of quantitative variables were released using centrality measures as mean standard deviation or median (interquartile range).
Bivariate associations between the studied gene polymorphisms, maternal GWG and neonatal characteristics (BW, MUAC, TST) were tested by Chi-square test, Student-t test for independent samples, ANOVA test or nonparametric tests (as Mann-Whitney or Kruskal-Wallis).
The effect of the studied gene polymorphisms on GWG and neonates' characteristics (BW, MUAC, TST) was adjusted for other covariates using logistic regression analysis or linear multiple regression.
Statistical significance was achieved with the estimated significance level of p < 0.05. All statistical analysis was performed using environment for statistical computing R version 3.4.4.

What is known.
• Excessive gestational weight gain (GWG) is associated with diabetes, obesity, and dystocia, but also modifications of neonatal birth weight (BW) with afterwards consequences on their evolution. One of four Romanian children is obese or overweight, with percentages varying between 19.7% and 35.8%, depending on the geographic area • MC4R and ENPP1 gene polymorphism are associated with increased risk for developing obesity.