Adipocytokines levels at delivery, functional variation of TFAP2β, and maternal and neonatal anthropometric parameters

Authors


  • Disclosure: There is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported. The founding sources had no further role in study design; in the collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the paper for publication.

    Funding: This study was supported by grants from the Swedish Research Council (VR) (2009-3640), the Swedish Council for Working Life and Social Research (FAS) (2011-0627), the Swedish Medical Council (2011, SLS-172231), the Söderström-Köningska Foundation (2010-SLS-156831), the Åke-Wiberg Foundation (302298937), and the Marianne and Marcus Wallenberg Foundation (MMW 2011.0115).

    Contributors: AS, ISP, and MO designed the study and wrote the protocol; AL carried out the endocrinological measurements; EC and LO managed the genetic analyses; EC and AS undertook the statistical analyses; EC and SI wrote the first draft of the manuscript. All authors contributed to and have approved the final manuscript.

Abstract

Objective

Adipocytokines participate in the regulation of glucose metabolism and fetal development. The transcription factor activating protein 2B (TFAP2β) has been associated with adipocytokine regulation, and gene variations with type 2 diabetes and obesity. This study investigated associations between maternal TFAP2B variation, adipocytokine levels, and maternal and neonatal anthropometric characteristics.

Design and Methods

A population-based sample of women was followed from delivery to 6 months postpartum. Adiponectin, leptin, and interleukin-6 levels at delivery, and maternal as well as neonatal anthropometric variables were assessed. The TFAP2β intron 1 variable number tandem repeat (VNTR) was genotyped.

Results

Maternal interleukin-6 correlated positively with leptin at delivery, with peripartum weight changes and weight of newborn males, adjusted for potential confounders. Leptin at delivery was associated with TFAP2β intron 1 VNTR genotype, adjusted for confounders, maternal weight and negatively with birth weight among female neonates. A path model suggested a link between TFAP2β genotype, leptin levels, and newborn females' weight.

Conclusions

The present results stress a role for the TFAP2 β in adiposity-related conditions and intrauterine growth. The association between neonatal birth weight and maternal adipocytokine levels, together with the observed sex effect, call for further studies on the mechanisms behind neuroendocrine fetal programming.

Introduction

The adipose tissue acts as a dynamic organ, with differentiated adipocytes functioning in an endocrine manner to secrete adipocytokines. These adipocytokines, in turn, play a role in regulating energy expenditure, glucose and fat metabolism, inflammation, and immune response [1]. Adipocytokine dysregulation is a typical feature of obesity-related conditions and is influenced by several factors, including genetic predisposition [2]. Prenatal neuroendocrine malprogramming of mechanisms regulating body weight and food intake control, in which adipocytokine dysregulation plays a crucial role, might contribute to the risk of developing obesity later in the child lifetime [3]. In addition, adipocytokines have been shown to participate in the regulation of fetal development [4].

Leptin, coded by the OB gene, is an adipocyte-secreted protein that plays a crucial role in several functions such as appetite, energy expenditure, immune function, fertility, and inflammatory processes. Plasma leptin concentration fluctuates in order to prevent any weight change and is strongly correlated with BMI [2]. Elevated leptin levels have been associated with a negative energy balance state where food intake exceeds energy consumption. In case of weight loss, the subsequent fall in leptin levels initiates the reverse reaction. Additionally, leptin deficiency has been shown to be related to insulin resistance and diabetes [1], as supported by results showing that these effects can be reversed when leptin is administrated [2]. Leptin levels rise during pregnancy [5], presumably because of both increased leptin resistance and placental contribution, and then decrease after delivery [6]. Furthermore, leptin levels during pregnancy are higher in women carrying a female fetus or during multiple gestations [7].

Adiponectin is an adipocyte-derived protein encoded by a gene located on chromosome 3q27, near the locus linked to adiposity and type II diabetes, suggesting an association with these conditions. Indeed, low plasma adiponectin is related to higher BMI and obesity [8]. Secretion of maternal adiponectin has been reported to be unaltered or decreased during normal pregnancy [5]. Neonatal adiponectin levels have been shown to be lower in the presence of fetal macrosomia, whereas maternal adiponectin levels at delivery do not seem to be correlated with birth weight and neonatal BMI [9]. However, maternal adiponectin levels appear to decrease in women with gestational diabetes mellitus (GDM), which may raise the risk of fetal macrosomia [10].

Interleukin-6 (IL-6) is a multifunctional cytokine, produced by several cell types, including adipocytes. It has a regulatory function in hematopoiesis, immune response, and acute phase reactions [11]. IL-6 interferes with insulin signaling and may contribute to the development of insulin resistance. Increased adipose and plasma protein IL-6 levels are associated with obesity [12]. Some studies report upregulation of IL-6 in subjects with GDM, indicating a direct association between insulin resistance and circulating IL-6 levels [13]. IL-6 levels increase during pregnancy [14], with a peak during the first day after delivery and a subsequent decrease [15].

Adipocytokine genes, such as those of leptin, adiponectin, and IL-6, have been found to contain binding sites for the transcription factor activating enhancer binding protein 2B (TFAP2β) in their promoter region [16, 17]. Additionally, relevant expression levels of the TFAP2B protein have been observed in human differentiated adipocytes and placenta [18]. The TFAP2B gene is located on chromosome 6p.12, and presents several SNPs, VNTRs, and ins/del polymorphisms [18]. Maeda and colleagues [18], in a genome-wide association study (GWAS), reported an association between polymorphic variations in the TFAP2B gene and type 2 diabetes. Among others, a single polymorphic locus in the intron 1, identified as a variable number tandem repeat (VNTR), was strongly associated with type 2 diabetes. The minor allele consisting of 9 repeats (R) has been indicated as the risk allele in comparison with the alleles consisting of 8 or 10R [18]. In addition, Tsukada and colleagues [19] have reported that the minor allele of this VNTR and the adjacent single nucleotide polymorphism (SNP) +774 have a positive enhancing transcription activity in differentiated adipocytes compared with the major alleles, 10R and G, respectively. Samples of human adipose tissue heterozygous for the intron I +774 TG SNP (which is in almost complete linkage disequilibrium with the VNTR [18]) displayed increased TFAP2β expression compared with homozygous GG samples [19], thus implying that the VNTR risk allele containing 9R is also associated with higher TFAP2β expression. Moreover, variation in the TFAP2β has been associated with insulin resistance and adiposity in adolescent males [20], as well as with binge eating disorder in women [21], with the 5R allele of a VNTR in intron 2 (in moderate linkage disequilibrium with the 9R allele of the VNTR in intron I [19]) being the risk allele.

Although several studies have investigated maternal and neonatal adipocytokine levels in relation to anthropometric measures, no previous study has investigated the genetic variation in the TFAP2β gene in relation to maternal adipocytokine levels and birth weight. This study aimed to investigate possible associations between maternal polymorphic variation of the TFAP2β gene, levels of adipocytokines at delivery, and maternal and neonatal anthropomorphic characteristics. We hypothesized that carriers of the TFAP2β risk allele would display altered adipocytokine levels at delivery and, as a consequence, this would have an effect on newborn birth weight.

Methods and Procedures

The study protocol was approved by the Independent Research Ethics Board of Uppsala University. This study was undertaken as part of the UPPSAT project, a population-based cohort study in the County of Uppsala, investigating correlates of postpartum depression [22-24].

Between May 2006 and June 2007, all women giving birth at Uppsala University Hospital were asked to take part in a longitudinal study of maternal, paternal, and infant well-being. Exclusion criteria for the study were [1] not being able to adequately communicate in Swedish, [2] women whose personal data were kept confidential on their own request, and [3] women with intrauterine demise or with infants immediately admitted in the neonatal intensive care unit.

Newly delivered women completed a self-administered structured questionnaire 5 days after delivery. A second and third questionnaire were sent by mail at 6 weeks and 6 months after delivery. The questionnaires contained various questions on anthropometric, social, and lifestyle characteristics.

Coded blood samples, collected in conjunction with routine intravenous catheterization before delivery, were stored at 4°C for a maximum of 24 hours and then centrifuged. Buffy coat was prepared and stored at -70°C. The final sample comprised 275 women.

Adipocytokines measurement

Samples were analyzed using commercially available ELISA kits (Leptin kit DY398, adiponectin/Acrp30 kit DY1065, and high-sensitivity IL-6 kit [HS600B, R&D Systems, Minneapolis, MN, USA]), by use of quantitative sandwich enzyme immunoassay technique. The immunoassays were calibrated against highly purified recombinant human leptin, adiponectin, or IL-6. The assays had a total coefficient of variation of approximately 6-7%. No significant variations in leptin or IL-6 levels were detected in relation to time of blood sampling, number of hours fasting, or cervical status at blood sampling (data not shown [22]).

TFAP2β intron 1 VNTR genotyping

The TFAP2β intron 1 VNTR polymorphism was amplified using the following primers designed in-house using the software Primer3 (http://frodo.wi.mit.edu/) based on the NCBI public database: forward 5′-XCAGACGGCTCCGCTACTC-3′ and reverse 5′-CCAGACCATTCCGCTTAAAA-3′ (chr6:50895458+50895639). The forward primer was labeled with the fluorescent dye 5'-fam. PCR was performed in a 25 μL reaction mixture containing 1 mM PCR Buffer10× with MgCl2; GC Buffer 2 μL, 7% DMSO; 0.5 μM dNTPs; 2.0 μM of the two primers, and 1.1 U Fast Start Taq DNA polymerase (Roche Diagnostics, Mannheim, Germany). The PCRs were performed on a GeneAmp 9700 (Applied Biosystems, Carlsbad, CA, USA) at the following profile: 95°C for 7 min, followed by 37 cycles at 95°C for 45 s, 59°C for 30 s, and 72°C for 30 s, and ended at 72°C for 7 min. The PCR products were analyzed using capillary electrophoresis ABI PRISM@3700 DNA Analyzer (Applied Biosystems) and allele size was determined manually on chromatograms using Gene Marker 1.5® AFLP/Genotyping software (SoftGenetics LLC®2004. State College, PA, USA). As control material, one-third of the samples were analyzed and genotyped twice, and the comparison indicated no inconsistencies. Genotyping was completed on 269 of 275 subjects (98%) and genotype frequencies fell within expected ranges, based on multiple published reports for Caucasians. Genotype frequencies were AA 1.9%; AB 4.9%; AC 14.1%; BB 3.4%; BC 25.5%; and CC 50.2%. The genotype frequencies were in Hardy Weinberg Equilibrium. The TFAP2β intron 1 VNTR polymorphism was categorized as carriers of the 9R risk-allele versus the rest (9R9R; 9R10R; 8R9R vs. 8R8R; 8R10R; 10R10R), according to Maeda et al. [18].

Statistical analyses

SPSS version 20.0 was used for the statistical analyses (®IBM). T-tests were used to test for significant differences in maternal anthropometric and background characteristics by the TFAP2β intron 1 VNTR polymorphism. Correlations between maternal and newborn characteristics and adipocytokine levels were assessed with the Spearman correlation coefficient. Stepwise backward linear regression analyses were performed with the natural logarithm (ln) of leptin as the outcome variable, in order to account for non-normality, and TFAP2β intron 1 VNTR polymorphism, ln adiponectin, ln IL-6, and maternal and newborn characteristics as the predictor variables. One-way ANOVA and Bonferroni post hoc tests were used to test differences in maternal serum adipocytokine levels at delivery in relation to the newborn weight by gestational age. Newborns were defined as small for gestational age (SGA) and large for gestational age (LGA) when deviating more than two standard deviations from the mean birth weight for newborns of the same sex, born during the same gestational week, according to the national standardized birth weight curves used in clinical practise in Sweden [25]. Linear backward stepwise regression was used to assess the effects of maternal adipocytokines (as continuous ln transformed measures), TFAP2B intron 1 VNTR polymorphism, and potential confounding factors on birth weight. Path analyses were used to investigate the relationship among the investigated variables, and results of the best-fit models have been illustrated by a path diagram using the statistical system R. Unstandardized regression coefficients have been reported to indicate how many units a dependent variable will change, per unit increase, in another predictor variable. However, the numbers relative to the coefficient cannot be interpreted in terms of units because the variables have been ln transformed. Error terms, which equal the residual variance of the variable, have been included for exogenous variables (not dependent on other variables in the model) and for endogenous variables (dependent on other variables in the model). As indicators of good model fit, chi-square (χ2), goodness-of-fit, and adjusted goodness-of-fit, RMSEA, Bentler–Bonnett NFI, Tucker–Lewis NNFI, Bentler CFI, and SRMR have been used as indices.

Results

Anthropometric and background characteristics of the participants in relation to TFAP2β intron 1 VNTR genotype are described in Tables 1 and 2. Obesity-related conditions were present among a few of the women: 1 woman had diabetes mellitus; 3 women developed gestational diabetes; 10 women had high blood pressure before pregnancy; and 20 during pregnancy. As potential confounding factors, first trimester data on alcohol use (N = 12) and cigarette smoking (N = 8) were also taken into consideration.

Table 1. Background characteristics of the mothers and their offspring by TFAP2β intron 1 VNTR
 TFAP2β intron 1 VNTR
9R9R; 9R10R; 8R9R8R8R; 8R10R; 10R10R
NMeanSDMedianNMeanSDMedian
  1. IL-6, interleukin-6; R, repeat; Spearman's rho Correlation Coefficient; TFAP2β, transcription factor activating protein 2B; VNTR, variable number tandem repeat.
MOTHERS        
Leptin at delivery (ng/mL)85275482325321515164346013256825026
Adipo at delivery (pg/mL)85502917334886164489520854461
IL-6 at delivery (pg/mL)858.110.44.91648.010.13.7
Weight before pregnancy (kg)7467.312.665.014568.413.565.0
Weight at delivery (kg)8182.113.881.015282.514.880.0
Weight gained during pregnancy (kg)7214.25.814.013614.65.114.0
Weight 6 weeks postpartum (kg)6671.212.670.513371.113.469.0
Weight 6 months postpartum (kg)6167.912.867.011768.613.567.0
Weight lost at 6 months postpartum (kg)5912.95.113.010613.56.413.0
BMI before pregnancy (kg/m2)7423.94.223.214324.64.623.3
BMI at delivery (kg/m2)7929.34.728.414929.64.828.7
BMI 6 weeks postpartum (kg/m2)6225.14.124.512725.54.424.5
BMI 6 months postpartum (kg/m2)6124.04.023.511624.65.223.6
Age (years)8830.94.931.017030.84.231.0
Parity (n)890.70.90.01730.60.80.0
NEWBORNS        
Sex (males)890.40.50.01740.60.51.0
Gestational age (weeks)8939.91.540.017439.81.540.0
Weight (g)863558576353716836035213590
Apgar score at 1 min898.80.79.01738.80.79.0

Correlations between maternal adipocytokines levels at delivery and anthropometric characteristics are presented in Table 3. As expected, maternal weight and BMI values over the peripartum period were positively correlated with maternal leptin levels, and negatively correlated with adiponectin levels at delivery. Leptin and adiponectin levels were inversely correlated. Maternal IL-6 levels at delivery were positively correlated with leptin levels, weight gain during pregnancy, and weight loss during the first 6 months postpartum.

Table 2.  
 TFAP2β intron 1 VNTR
 9R9R; 9R10R; 8R9R8R8R; 8R10R; 10R10R
 NYes%NYes%
MOTHERS      
Diabetes mellitus before pregnancy8811.116900
High blood pressure before pregnancy8844.516963.6
Gestational diabetes880016931.8
Gestational high blood pressure8866.8169148.3
Alcohol use during pregnancy7756.515874.4
Cigarette smoking during pregnancy7045.712943.1
Table 3. Correlation between adipocytokines and anthropometric characteristics
  LeptinAdiponectinInterleukin-6
 NrPrPrP
Leptin at delivery (ng/mL)255  −0.20.0010.40.000
Adipo at delivery (pg/mL)255−0.20.001  −0.10.205
IL-6 at delivery (pg/mL)2550.40.000−0.10.205  
Weight before pregnancy (kg)2120.20.000−0.20.005−0.10.375
Weight at delivery (kg)2260.20.000−0.20.0120.00.842
Weight gained during pregnancy (kg)2030.10.062−0.10.3160.20.005
Weight 6 weeks postpartum (kg)1930.20.001−0.10.056−0.10.435
Weight 6 months postpartum (kg)1730.30.001−0.10.108−0.00.792
Weight lost at 6 months postpartum (kg)1620.10.460−0.10.1460.20.014
BMI before pregnancy (kg/m2)2100.30.000−0.20.017−0.10.407
BMI at delivery (kg/m2)2210.30.000−0.10.0730.00.512
BMI gain at delivery (kg/m2)2030.10.096−0.00.7240.20.013
BMI 6 weeks postpartum (kg/m2)1830.30.000−0.10.073−0.10.413
BMI 6 months postpartum (kg/m2)1720.30.000−0.10.290−0.00.945
BMI loss 6 months postpartum (kg/m2)1630.00.571−0.10.3210.20.033
Newborn weight (gr)2470.00.468−0.10.1770.00.469

Because the TFAP2β gene intron I VNTR has been associated with type 2 diabetes [18], presence of diabetes mellitus was an exclusion criterion. Group differences by TFAP2β intron 1 VNTR genotype were observed regarding leptin (P = 0.058; t = 1.9) and newborn sex (P = 0.001; t = 3.4) but not regarding birth weight (Table 1). A stepwise backward linear regression model with leptin levels at delivery as the outcome variable suggested a borderline significant effect of TFAP2β intron 1 VNTR genotype (P = 0.097). The covariates adiponectin, IL-6, weight before pregnancy, age at delivery, cigarette smoking during pregnancy, as well as newborn sex contributed significantly to the model (model F = 10.6; P < 0.001) (Table 4). As sex of the newborn had a significant main effect on the model, the analyses were repeated after stratification by sex of the child. The effect of TFAP2β intron 1 VNTR genotype (P = 0.022) was significant only among women carrying female fetuses (model F = 10.12; P < 0.001) (Table 5). Among pregnant women with female neonates, carriers of the 9R allele showed the lowest leptin levels at delivery (24,878 ± 19,286 ng/mL) compared with carriers of the rest of the genotypes (43,334 ± 34,686 ng/mL).

Table 4. Backward stepwise linear regression model on leptin levels at delivery considering TFAP2β intron 1 VNTR, adjusted for covariates and confounding factors
Leptin at deliveryβP
  1. β: Standardized coefficient.
  2. Dependent variable: ln leptin at delivery.
  3. Entered variables: TFAP2β intron 1 VNTR; ln adipo at delivery; ln IL-6 at delivery; ln weight before pregnancy; ln weight at delivery; gestational week; age; gestational diabetes; high blood pressure before pregnancy; gestational high blood pressure; alcohol use during pregnancy; cigarette smoking during pregnancy; newborn sex; newborn weight.
TFAP2β intron 1 VNTR−0.120.097
Adiponectin at delivery−0.180.009
Interleukin-6 at delivery0.420.000
Weight before pregnancy0.170.015
Cigarette smoking during pregnancy−0.210.002
Age at delivery−0.130.062
Newborn sex−0.140.044
Final model (F = 10.6; P < 0.001) 
Table 5. Backward stepwise linear regression model on leptin levels at delivery considering TFAP2β intron I VNTR, adjusted for covariates and confounding factors, by newborn sex
Leptin at deliveryβp
  1. Dependent variable: ln leptin at delivery.
  2. Entered variables for box sexes: TFAP2β intron 1 VNTR; ln adipo at delivery; ln IL-6 at delivery; ln weight before pregnancy; ln weight at delivery; gestational week; age; gestational diabetes; high blood pressure before pregnancy; gestational high blood pressure; alcohol use during pregnancy; cigarette smoking during pregnancy; newborn weight.
Newborn   
FemalesTFAP2β intron 1 VNTR−0.230.022
 Interleukin-6 at delivery0.430.000
 Weight before pregnancy0.210.037
    
MalesAdiponectin at delivery−0.260.004
 Interleukin-6 at delivery0.380.000
 Cigarette smoking during pregnancy−0.320.001
 High blood pressure before pregnancy−0.160.079
Final model: females (F = 10.12; P < 0.001); males (F = 10.67; P < 0.001)

There were 17 (6.5%) newborn babies LGA and 5 (1.9%) SGA. When maternal serum adipocytokine levels at delivery and differences in newborn weight were compared, maternal adiponectin levels were borderline significantly lower among LGA than normal weight newborns (4201.7 ± 2567.7 vs. 4991.6 ± 1940.5 pg/mL; Bonf. corr. P = 0.086). The effect of adipocytokines and TFAP2β intron 1 VNTR genotype on birth weight was assessed by linear backward stepwise regression analysis. Potential confounders were also included in the multivariate model, that is, maternal age and weight at delivery, nicotine and alcohol use at first prenatal visit, gestational diabetes, and gestational age. In the final models, the birth weight of newborn females was negatively related to maternal leptin (P = 0.019; β = −0.21), whereas birth weight of newborn males was positively related to maternal IL-6 (P = 0.035; β = 0.18). As expected, maternal weight at delivery and gestational age were significant covariates in both models.

Best-fit models obtained by path analysis among newborn females were in line with the hypothesis that carriers of the TFAP2β risk allele would display altered leptin levels at delivery, and as a consequence, this would have an effect on newborn birth weight (Figure 1a). Goodness of fit indices indicated a good fit of the data to the proposed model (Adj. goodness-of-fit: 0.96 among females and 0.94 among males; see also Supporting information). Among newborn males, consistent with the results from regression analyses, there was no effect of the TFAP2β genotype (Figure 1b).

Figure 1.

Path diagram indicating the relationships among the selected variables by the best-fit model, and their respective unstandardized regression coefficients (→ and ←) or correlation coefficients (←→). Error terms are indicated for each variable (←→). Squared multiple correlation coefficients (R2) are reported to indicate the variance in the variables accounted for by the factors. Further details are reported in the Supporting information. The numbers relative to the coefficients are based on ln transformed variables. A: female newborns, B: male newborns.

Discussion

In this study, we examined the association between TFAP2β polymorphisms, adipocytokines at delivery, and maternal as well as neonatal anthropometric characteristics. Our findings indicate an effect of the polymorphism TFAP2β intron 1 VNTR on maternal leptin levels at delivery. After stratification for newborn sex, a significant effect of TFAP2β intron 1 VNTR genotype was found only among women carrying female newborns. In line with previous findings, we found that maternal weight and BMI over the peripartum period were positively correlated with maternal leptin levels and negatively correlated with adiponectin levels. Interestingly, in this study, we show a positive association between maternal IL-6 levels at delivery and weight gain during pregnancy as well as weight loss during the first six months postpartum, but not directly with weight and BMI.

Previous studies indicate that TFAP2β variations might affect regulation of genes involved in insulin responsiveness of differentiated adipocytes [18, 26], and that TFAP2β overexpression in adipocytes is associated with insulin resistance [27]. Thus, insulin response might be potentially under transcriptional regulation of the TFAP2β. Fuke and colleagues [28] have previously shown that overexpression of the TFAP2β in adipocytes directly down-regulates leptin protein expression and ultimately leptin secretion, through its role as a transcription factor regulating the leptin gene. Yeung and colleagues [29] found that female carriers of the risk variant (G) of SNP rs987237 (A/G) in the TFAP2β, previously associated with central obesity in a GWAS study, have lower leptin levels. These results, which are in line with our results and previous findings [18, 26, 27, 30], support that dysregulation of adipocytes, resulting in altered leptin levels, and consequently in insulin resistance and metabolic syndrome, may depend on TFAP2B genotype. The role of newborn gender as an effect modifier in the association between leptin levels and the TFAP2β polymorphism, which was demonstrated in this study, could potentially be due to the fact that leptin levels are higher among women carrying female fetuses [7], thus magnifying the genetic effect. Surprisingly, no direct association was found between TFAP2β intron 1 VNTR genotype and maternal adiponectin, or IL-6 levels, or on maternal peripartum weight, or BMI. This suggests that other factors might interact with TFAP2β in regulating adiponectin, IL6, and ultimately BMI. This also suggests that distinct signaling pathways of the various adipocytokines, their central and peripheral feedback regulatory mechanisms, and intermediate hormonal, paracrine, or autocrine mechanisms, might interfere with the role of the TFAP2β [1, 31]. In fact, in two previous studies, an association was observed between variation in the TFAP2β and insulin sensitivity, central adiposity [20], and leptin [29], but not with adiponectin levels [20, 29]. All together, these and the present findings suggest an effect of genetic variation of the TFAP2β on leptin and insulin sensitivity but not on adiponectin levels. However, further investigations are warranted since Ikeda et al. [32] reported that overexpression of TFAP2β in adipocytes decreases mRNA expression and secretion of adiponectin, and increases those of IL-6. Dysregulated leptin secretion has been repeatedly associated with obesity [1, 31]. In addition, the role of leptin in the inhibition of feeding, energy storage and expenditure, glucose and lipid metabolism, insulin sensitization and pancreatic B-cell function, as well as stimulation of fatty acid oxidation, is well established [1, 31]. Consequently, the genetic regulation of leptin by TFAP2β is of clinical relevance and further investigations of possible influences of maternal adipocytokine levels at delivery on subsequent child development are planned within this longitudinal study.

In this study, maternal levels of adipocytokines were also investigated in relation to maternal and neonatal anthropometric characteristics. Leptin and adiponectin levels were inversely correlated in our material, whereas IL-6 levels were positively correlated with leptin, in accordance with previous studies [33]. Maternal weight and BMI perinatally were positively correlated with maternal leptin levels, in accordance with previous studies [6, 34]. This might, at least in part, be due to accretion or loss of fat mass associated with pregnancy, rather than being the cause of changes in BMI.

Adiponectin levels showed a statistically significant negative correlation with perinatal weight and BMI at the beginning of pregnancy, whereas its correlation with BMI at delivery and postpartum was only borderline significant. Many studies have reported decreased maternal and umbilical cord levels of adiponectin in macrosomic pregnancies [35]. It is also suggested that levels of maternal adiponectin are decreased in women with GDM, which may increase the risk of fetal macrosomia [10]. A possible mechanism explaining the association between maternal adiponectin and LGA children could be the increased insulin resistance and glucose intolerance, which characterize obese mothers. Our study, while being in line with these previous reports, could demonstrate only a borderline statistically significant difference between LGA and normal weight infants in adiponectin levels. This may be because of various factors, involving timing of sampling during pregnancy or lack of power because of the small sample size.

A positive association between maternal IL-6 levels at delivery and weight gain during pregnancy, as well as weight loss during the first 6 months postpartum, was observed in the current material. A small study indeed found a positive correlation between maternal IL-6, weight, and BMI [34]. In pregnant women, IL-6 is detectable in early pregnancy and rises significantly after 25 gestational weeks [14], possibly as a consequence of the physiological insulin resistance of pregnancy. In trying to explain the biological mechanisms behind the present finding, the complex systems involved in both inflammatory processes and weight parameters need to be considered. Indeed, several studies reported on the associations between cortisol, sex steroids, pro-inflammatory cytokines, and weight changes in the literature.

Data from human or animal studies on a possible association between maternal adipocytokines and birth weight are scarce. In this study, the birth weight of newborn females was negatively related to maternal leptin, and birth weight of newborn males was positively related to maternal IL-6, after adjusting for potential confounders, and there was no main effect of the TFAP2β intron 1 VNTR polymorphism. When pairing the finding that higher weight among female newborns is associated with lower maternal leptin levels at delivery with the finding that the TFAP2β intron 1 VNTR risk variant is associated with lower maternal leptin levels (as shown in Figure 1), and in light of human and animal findings [36, 37], it seems likely that maternal TFAP2β variation and leptin play a role for fetal programming of neuronal systems involved in control of feeding behavior, and on risk for obesity and related condition later in life. Thus, the present results can be considered showing the first evidence of a potential effect of maternal TFAP2β genotype, via genetic regulation of leptin levels, on offspring weight. The investigation on possible influences of such finding on subsequent child development in this current ongoing longitudinal study will contribute to the clinical relevance of the present finding.

Regarding the positive association of maternal IL-6 with birth weight among males found in this study, one other study showed a positive correlation between maternal IL-6 levels at delivery and neonatal weight and fat mass. However, this study included only 18 women, and effect of neonatal sex was not assessed [38]. Thus, the present findings need to be replicated in large cohorts as several potential explanations could be addressed regarding differences among the published studies, including diverse assessment measures, time-points, adjustment for different covariates, and sample selection. Moreover, other factors such as placenta adipocytokine secretion, the neonate TFAP2β genotype, as well as other factors, which can act in a paracrine way, such as the TNF-α, contribute to fetal weight and adiposity and should also be considered. Fetal programming in a sex-dependent manner has been shown in animal studies on prenatal exposure to cytokines [39], and is a probable explanation for the present findings.

Among the limitations of this study was the timing of the blood sampling, which, because of administrative reasons, was performed during labor. However, no significant variations in adipocytokine levels were detected in relation to time of blood sampling, number of hours fasting, or cervical status at blood sampling. Serial measurements of adipocytokines, at a specific time of day, during and after pregnancy would have been able to better demonstrate potential associations. On the contrary, the study has a population-based design, and an acceptable size for its kind. The detailed individual information on various perinatal characteristics, and especially weight-related parameters was valuable in the context of this study. To note is that maternal anthropometric measures were self-reported, which might have led to a random misclassification that would however most probably not affect the results of this study. The follow-up period, comprising the first 6 postpartum months, was also adequately long. Moreover, the fact that this study was carried out on the basis of a priori hypothesis, and that both a genetic and an endocrine approach have been used represents the main strength of the study.

Conclusion

Maternal obesity is a risk factor associated with maternal, neonatal, and childhood morbidity. High maternal BMI and altered adipocytokine levels contribute to a suboptimal intrauterine environment, which is responsible for obesity, metabolic syndrome, and diabetes in children [3]. Complex pathophysiological pathways link together genetic polymorphic variations, tissue-specific epigenetic modification processes, protein levels, and the final phenotype in both mother and child. Although this study focuses on only some of the above-mentioned entities, our results contribute with valuable information on different steps of the individual risk-shaping process, and reflect the impact of dysregulation of adipocytokines on maternal and neonatal health. Future aims of this ongoing longitudinal study will be to investigate possible influences of maternal altered adipocytokine levels at delivery on subsequent child development.

Acknowledgments

The authors sincerely thank all the mothers who participated in this study, as well as all the employees at the Department of Obstetrics, Uppsala University Hospital, who helped with the distribution of questionnaires and blood sampling. The authors thank Dr. Fotis Papadopoulos and Åsa Johansson for their statistical advice.

Ancillary