Sex dimorphism in the associations of gestational diabetes with cord blood adiponectin and retinol-binding protein 4

Study design, subjects and specimens

This study was based on the recently described Shanghai Birth Cohort (SBC).23 The SBC is a large, carefully phenotyped pregnancy/birth cohort with linked biospecimen bank for studies on perinatal determinants of infant growth, development and health. Women at preconception or early pregnancy care clinics were recruited from six urban university-affiliated tertiary obstetric care hospitals in Shanghai between 2013 and 2016 and including a total of 4127 pregnancies. The women were followed up at the first, second and third trimesters of pregnancy and delivery. Data and specimens were collected at each study visit. All collected blood samples were kept on ice, stored temporarily in a 4°C refrigerator and centrifuged within 2 hours after the specimen collection. The separated serum and EDTA plasma samples were stored in multiple aliquots at −80°C until assays.

Maternal pre-pregnancy body mass index (BMI, in kg/m²) was calculated as pre-pregnancy weight (self-reported) divided by squared height (measured in the first prenatal care visit). GD was diagnosed according to the International Association of Diabetes and Pregnancy Study Groups criteria24—if any one of the blood glucose values was at or above the following thresholds in the 75 g oral glucose tolerance test at 24–28 weeks of gestation: fasting 5.1 mmol/L, 1-hour 10.0 mmol/L and 2-hour 8.5 mmol/L.

As part of the SBC project, we conducted a nested case–control study on cord blood biomarkers in GD.25 The present study reported the original data on the impacts of GD and fetal sex on cord blood leptin, adiponectin and RBP-4 levels (a priori planned study objective). Cases were the newborns of GD mothers, and controls were the newborns of euglycemic mothers. We did not have the data on whether GD mothers were hyperglycemic at the time of delivery (cord blood specimen collection). All cases and controls must meet the following eligibility criteria: (1) Han ethnicity (the majority ethnic group, >98%); (2) maternal age 20–45 years; (3) natural conception; (4) singleton pregnancy; (5) the mother was free of severe pre-pregnancy illnesses (eg, type 1 or type 2 diabetes, and chronic hypertension) and life-threatening pregnancy complications (eg, pre-eclampsia/eclampsia); (6) the infant was free of birth defects; (7) 5 min Apgar score ≥7; and (8) cord blood and placenta specimens available for future assays. All eligible newborns of GD mothers (n=153) were included. Controls were randomly sampled among all eligible subjects and matched to cases (1:1) by infant sex (the same) and gestational age (within 1 week). Therefore, a total of 306 subjects (153 GD; 153 controls) constituted the study sample.

Biochemical assays

Cord plasma RBP-4 was measured by an ELISA kit (R&D Systems, Minnesota, USA), and the absorbance was determined using a microplate spectrophotometer (Beckman CX7, USA). Plasma HMW and total adiponectin were measured by an ELISA kit from ALPCO (Salem, NH, USA), and plasma leptin by an ELISA kit from Invitrogen (Carlsbad, CA, USA), respectively. The detection limits were 0.22 ng/mL for RBP-4, 0.034 ng/mL for HMW and total adiponectin, and 3.5 pg/mL for leptin, respectively. Intra-assay and interassay coefficients of variation were in the ranges of 5.0%–8.6% for RBP-4, and 6.9%–10.4% for leptin, HMW and total adiponectin, respectively. In all biomarker assays, the laboratory technicians were blinded to the clinical status (GD or not) of study subjects.

Statistical analysis

The primary outcomes were cord plasma leptin, HMW adiponectin and RBP-4 concentrations. Paired t-test was used to compare continuous variables in GD and control groups, and McNemar’s χ² test was used for comparisons of dichotomous variables. Pearson partial correlation analysis was used to evaluate the correlations between biomarkers adjusting for gestational age at delivery/blood sampling. Log-transformed biomarker data were used in t-tests, correlation and regression analyses. Generalized linear models were applied to assess the differences in cord blood RBP-4, leptin and adiponectin concentrations (by GD status or fetal sex) adjusted for maternal and neonatal characteristics, and in tests for interactions (as multiplicative terms). Maternal and neonatal characteristic variables included maternal age, ethnicity (all Chinese Han), parity, education (university: yes/no), pre-pregnancy BMI (kg/m²), gestational weight gain (z score), smoking or alcohol use in pregnancy (yes/no), GD, gestational hypertension, family history of diabetes, family history of hypertension, mode of delivery (cesarean section/vaginal), infant sex, gestational age (weeks) and birth weight (z score). Parsimonious final regression models were fitted in evaluating the adjusted effects excluding covariables that were not associated with the outcomes (p>0.2) and did not affect the comparisons. Data management and analyses were performed in SAS V.9.4 (SAS Institute). Two-tailed p value <0.017 was considered statistically significant in testing the hypothesis concerning the differences in the three primary outcomes comparing GD versus control groups (Bonferroni correction for three comparisons). A relatively lenient p value <0.10 was considered significant in regression tests of interactions to allow better chance of identifying potentially important interactions.

Maternal and neonatal characteristics

Characteristics of the study subjects were described recently.25 Briefly, women with GD had higher pre-pregnancy BMI than euglycemic women (mean±SD: 23.6±4.3 vs 21.6±2.9 kg/m², p<0.001), and were more likely to have gestational hypertension (5.2% vs 0.6%, p=0.036), and tended to be more likely to have a family history of diabetes (16.5% vs 9.5%, p=0.072). There were no significant differences in other maternal characteristics including maternal age (30.4±3.7 vs 29.9±3.4 years), education (university 62.8% vs 60.5%), parity (primiparous 81.1% vs 81.7%), alcohol use (7.1% vs 11.3%) and smoking during pregnancy (only three smokers) (all p>0.1). Newborns of GD mothers were more likely to be delivered by cesarean section than newborns of euglycemic mothers (54.3% vs 33.1%, p<0.001). Despite the matching for infant sex and gestational age at delivery (within 1 week), mean gestational age at delivery was slightly shorter comparing GD versus euglycemic pregnancies (39.1±1.2 vs 39.3±1.2 weeks, p<0.001), while average birth weight tended to be higher (3435.9±488.7 vs 3354.6±435.6 g, p=0.093).

Cord plasma RBP-4, leptin and adiponectin in GD and euglycemic pregnancies

Adjusting for maternal and neonatal characteristics including pre-pregnancy BMI, family history of hypertension, gestational hypertension, family history of diabetes, cesarean section, and gestational age at delivery (other covariates did not affect the comparisons), cord plasma HMW adiponectin concentrations and HMW/total adiponectin ratios were lower in GD versus euglycemic pregnancies in the total study sample (table 1). However, there were no significant differences in cord plasma RBP-4, total adiponectin and leptin concentrations between the two groups.

Sex-specific associations of GD with RBP-4, leptin and adiponectin

Adjusted for maternal and neonatal characteristics, there were no significant differences in all observed cord blood biomarkers between GD and non-GD groups in male newborns (table 2). In contrast, for female newborns, GD was associated with higher cord plasma RBP-4 concentrations (adjusted p=0.007), but lower HMW adiponectin (adjusted p=0.004) and total adiponectin (adjusted p=0.048, not significant if accounting for multiple tests) concentrations, and lower HMW/total adiponectin ratios. The p values in tests for interactions of GD with infant sex were 0.094 for RBP-4, 0.035 for HMW adiponectin, 0.114 for total adiponectin, and 0.198 for HMW/total adiponectin ratio, respectively. The adjusted differences (in log scale) are presented in table 3, and the findings were similar as in table 2.

Sex differences in cord blood RBP-4, leptin and adiponectin

In GD pregnancies, female newborns had significantly higher cord plasma RBP-4 concentrations than male newborns (21.7±5.4 vs 18.8±4.5 ng/mL, p<0.001), while there was no sex difference in euglycemic pregnancies (p=0.43). In euglycemic pregnancies, there were higher HMW adiponectin concentrations (p=0.014) and HMW/total adiponectin ratios (p=0.006) in female versus male newborns, while these differences disappeared in GD pregnancies (p=0.59 for HMW adiponectin; p=0.47 for HMW/total adiponectin ratio). Cord plasma leptin levels were higher in female versus male newborns regardless of GD status (p=0.001 in GD; p=0.045 in controls).

Partial correlations

Adjusting for gestational age at blood sampling, cord plasma RBP-4 was weakly and positively correlated to leptin and birth weight z score, but not correlated to adiponectin (table 4). As expected, HMW and total adiponectin were strongly correlated. Leptin was positively correlated to HMW adiponectin. Sex-stratified analyses showed that RBP-4 was positively correlated with leptin in female newborns but not in male newborns (table 5).

Main findings

The study is the first to reveal sex-dimorphic associations of GD with cord blood RBP-4 and adiponectin concentrations. GD was associated with elevated cord blood RBP-4 and decreased HMW adiponectin concentrations in female newborns but not in male newborns. The study is the first to demonstrate that GD may affect the normal presence (HMW adiponectin) or absence (RBP-4) of sex dimorphism in some insulin sensitivity regulation-relevant adipokines in early life.

Data interpretation and comparisons to findings in previous studies

Adipokines play a critical role in the regulation of glucose homeostasis and insulin sensitivity,12 26 and may provide an important link between obesity-related disorders and insulin resistance.27 28 RBP-4 is involved in regulating glucose metabolism and insulin sensitivity.12 29 Some studies have reported higher maternal serum RBP-4 levels in GD women.9 13 14 Two studies9 13 reported no significant differences in cord serum RBP-4 concentrations between GD and control groups, but they did not report sex-specific comparisons. In contrast, our data showed that GD was associated with elevated cord plasma RBP-4 concentrations in female newborns only, suggesting that GD may have a sex-dimorphic impact on metabolic health during fetal life.

Two studies reported that total or HMW adiponectin concentrations in cord serum did not differ between GD and controls,30 31 whereas two other studies reported lower cord serum or plasma adiponectin concentrations in GD.8 9 In contrast, we observed that GD was associated with decreased cord plasma adiponectin (more substantially for HMW adiponectin) concentrations in female newborns, but not in male newborns. Previous studies did not provide sex-specific data. Our data from a nested case–control study in a much larger cohort suggest a sex-dimorphic impact of GD on cord blood adiponectin concentration.

Some studies in adolescents or adults reported sex differences in circulating RBP-4 levels,29 32 while others did not.33 Serum RBP-4 concentrations were observed to be lower in female than male subjects in children and adults.29 34 35 There is a lack of data on sex difference in cord blood RBP-4. We observed no sex difference in cord plasma RBP-4 concentrations in euglycemic pregnancies, but higher concentrations in female newborns in GD pregnancies. The mechanisms are unclear. We speculate that this may be due to the altered metabolic milieu in GD with pronounced insulin resistance.36

Higher total adiponectin circulating concentrations in females versus males have been reported in children, adolescents and adults.37–39 This sex dimorphism appears to be primarily related to the HMW adiponectin complexes.39 40 Our data in euglycemic pregnancies suggest that this sex dimorphism in HMW adiponectin may be established in utero in normal (euglycemic) pregnancies (higher cord plasma HMW adiponectin concentrations in female vs male neonates). HMW adiponectin levels may vary by fetal sex even at the early stage of human development. We speculated that this sex dimorphism could be related to the effects of androgens. Testosterone may selectively reduce HMW adiponectin secretion from adipocytes,41 42 and higher testosterone concentrations in male versus female fetuses43 may contribute to the lower HMW concentrations in male newborns. Androgens may indirectly regulate adiponectin distribution through a decrease in adipocyte number and size in male fetuses.44 45 Importantly, our study is the first to reveal that such physiological sex difference in cord blood HMW adiponectin levels may be lost in GD pregnancies.

Strengths and limitations

The main strengths include the large birth cohort, timely collection and processing of blood specimens and high-quality biochemical assays (low intra-assay and interassay coefficients of variation), and sex-specific data. The main limitation is the observational nature of the study. We could not rule out the possibility of reverse causality. The study was limited to Chinese Han ethnicity subjects. While such study design confers the advantage of enhanced power to uncover biological relationships in the absence of the potential confounding effects of race/ethnicity, it also calls for more studies in other race/ethnic groups to understand the generalizability of the study findings.