Background: The presence of the adipokines adiponectin and leptin in cord blood and placental and fetal tissues suggests a possible role in fetal development.
Methods: We measured concentrations of adiponectin and leptin in maternal serum, cord blood, and breast milk and examined their correlations within a large, population-based study. Between November 2000 and November 2001, we recruited all mothers and their newborns after delivery at the University of Ulm (Ulm, Germany). The current analysis included 766 mothers with available breast milk samples collected 6 weeks postpartum. Adipokine concentrations were measured with commercially available ELISAs (R&D Systems).
Results: Median adiponectin concentrations in maternal serum (n = 713), cord blood (n = 709), and breast milk (n = 766) were 8.6 mg/L, 30.6 mg/L, and 10.9 μg/L, respectively. Median leptin concentrations were 12.8 μg/L in maternal serum, 7.8 μg/L in cord blood, and 174.5 ng/L in breast milk. Whereas increases in leptin concentrations with increasing birth weight, birth weight according to gestational age, and ponderal index were statistically significant in cord blood (all P values <0.0001), cord blood adiponectin was clearly related only to birth weight (P = 0.0004). Concentrations of both adipokines were moderately correlated in breast milk and maternal serum (both Spearman ρ values were 0.43; P <0.0001).
Conclusions: Concentrations of adiponectin and leptin vary strongly in maternal serum, cord blood, and breast milk, with only moderate correlations between both adipokines in maternal serum and breast milk. The health implications of these patterns warrant further investigation.
Adiponectin and leptin are members of the adipose-secreted proteins termed adipocytokines or adipokines. Leptin was discovered as a protein involved in the development of obesity, and although it is now recognized as a hormone that is produced by several tissues, adipose tissue is the principal site of leptin production and the major determinant of the concentration of circulating hormone (1). Adiponectin, discovered in 1995, is postulated to play a role in the modulation of glucose and lipid metabolism in insulin-sensitive tissues. Plasma adiponectin concentrations exceed those of any other hormone by a factor of 1000 and decrease in insulin-resistant states, including type 2 diabetes mellitus, suggesting a positive relationship to whole-body insulin sensitivity (2). Recent findings have indicated that adiponectin has antiatherogenic properties and antiinflammatory effects (3).
In addition to regulation of whole-body metabolism, leptin and adiponectin are known to be produced within the intrauterine environment (4)(5)(6)(7)(8)(9). The findings that leptin and adiponectin are present in cord blood and the positive correlation of their concentrations with neonatal birth weight (10)(11)(12)(13), as well as the high production of both leptin and adiponectin in the placenta and the fetus, suggest that these adipokines may play key roles in fetal development (14)(15). Furthermore, adipokines are potent regulators of the immune response and are possible risk determinants for many diseases that may involve dysregulation of the immune response, such as asthma, obesity, and other inflammatory conditions (16).
The presence of leptin in breast milk and its possible relationships with leptin in maternal serum or cord blood has been addressed in a small number of studies (17)(18)(19)(20), but to the best of our knowledge the corresponding roles of adiponectin have not been described. Therefore we measured adiponectin and leptin concentrations and investigated their relationships in 3 compartments, maternal serum, cord blood, and breast milk.
Materials and Methods
study design and study population
This analysis was carried out in the context of a large-scale birth cohort study that primarily investigated intrafamilial transmission of Helicobacter pylori infection. Details of the study design have been described elsewhere (21). In brief, all women and their infants who underwent delivery at the Department of Gynecology and Obstetrics at the University of Ulm between November 2000 and November 2001 were screened for eligibility for the study. According to a priori defined exclusion criteria, we excluded cases with fetal gestational age <32 weeks, infant birth weight <2500 g, or transfer of the infant to inpatient pediatric care immediately after delivery. Overall, 1066 mothers and their newborns were included in this study (67% of all 1593 eligible participants who fulfilled the inclusion criteria).
All woman participants gave informed consent for themselves and their infants. The study was approved by the Ethics Boards of the University of Ulm and of the Physicians’ Boards of the states of Baden-Wuerttemberg and Bavaria.
A baseline examination was performed on all mothers during postpartum hospitalization, with standardized interviews conducted by trained interviewers. A standardized form was used to collect laboratory and anthropometric data during pregnancy from the mother’s medical record. Infant weight and length at birth were obtained from the delivery records. The ponderal index at birth was calculated with the standard formula of weight (kg) divided by the cube of the length at birth (m3). We used a sex-specific, population-based German reference for birth weight for gestational age to classify infants as small, adequate, or large for gestational age (22). Cord blood samples were collected immediately after delivery, and maternal serum samples were collected along with routine hospital sample collection postpartum. Cord blood and serum samples were divided into aliquots and frozen at −80 °C until analysis. All mothers included at baseline were contacted by phone after 6 weeks and asked whether they were breastfeeding at that time. We were able to contact 1024 mothers (96%); 786 (76.7%) were still breastfeeding their infants at that time, and we collected 769 breast milk samples (from 97.8%). Six months after delivery we were able to contact 738 (96.0%) of these 769 mothers. At that time, 484 (66%) of the women were still exclusively or partly breastfeeding, and we collected 471 breast milk samples (from 97%). Milk samples (10 mL) were collected by a trained nurse who visited the homes of the participants. The breast milk samples were immediately cooled, and then frozen at −80 °C within 24 h. In most cases, breast milk samples were collected by the nurse from both breasts by manual expression before feeding. In rare cases, breast milk samples were collected by the mothers themselves and/or by means of a pump.
measurement of adiponectin and leptin
Adiponectin and leptin concentrations in maternal serum, cord blood, and breast milk were measured with a commercially available ELISA (R&D Systems). Measurements of adiponectin and leptin in maternal serum and in cord blood were performed according to the manufacturer’s instructions. The protocols were slightly modified for adiponectin and leptin in breast milk: to remove fat, we centrifuged breast milk for 5 min at 12 350g in a Heraeus Biofuge (Kendro Laboratory Products GmbH), and we used the fat-free phase for determination of adiponectin and leptin. The intra- and interassay CVs were both <7.0%.
The reported analysis data were from samples from 766 mothers from whom milk samples for both adipokine measurements were collected 6 weeks postpartum. Adiponectin and leptin concentrations were also measured in all available samples of maternal serum and cord blood from mothers. We also measured adiponectin and leptin concentrations in a randomly selected subsample of 42 samples of breast milk collected 6 months postpartum.
We first carried out descriptive analyses of sociodemographic characteristics of the mothers. We then analyzed adiponectin and leptin distributions in the 3 compartments according to the sex and anthropometric measurements of the neonates. Data are presented as mean (SD). Differences between groups within the same compartment were determined by the Kruskall–Wallis test. Correlations between concentrations of each adipokine in the 3 compartments and between the 2 adipokines within each compartment were determined by Spearman correlation coefficients. All analyses were carried out with the SAS statistical software package (SAS Institute, Inc.).
Characteristics of the mothers and infants in the study sample are shown in Table 1⇓ . Of the infants, 50.7% were male. Mean (SD) age of the mothers at delivery was 31.3 (4.8) years, and 19.6% had <10 years of school education; 86.6% of the mothers were of German nationality.
The concentrations of adiponectin in the 3 compartments according to neonate sex and anthropometric measurements at birth are shown in Table 2⇓ . Overall, adiponectin concentrations were significantly higher in cord blood than in maternal serum, but significantly lower in breast milk than in maternal serum (all P values <0.0001; data not shown). Adiponectin concentrations in maternal serum decreased with increasing ponderal index at birth (P = 0.03), whereas infant sex, birth weight, and birth weight according to gestational age showed no relationship with adiponectin concentrations in maternal serum. Adiponectin in cord blood was related to birth weight (P = 0.0004) but was not related to the other anthropometric measurements at birth or to infant sex. Concentrations of adiponectin in breast milk were higher among mothers whose infants were large for gestational age compared with mothers whose infants were small or adequate for gestational age (P = 0.004), but no relationship existed with infant sex or other anthropometric measurements at birth.
The concentrations of leptin in the 3 compartments according to infant sex and anthropometric measurements at birth are shown in Table 3⇓ . Overall, leptin concentrations were significantly lower in cord blood than in maternal serum, and concentrations in breast milk were ∼70-fold lower than in maternal serum (all P values <0.0001; data not shown). Leptin concentrations in maternal serum showed no significant association with infant sex or anthropometric measurements at birth. Leptin concentrations in cord blood were significantly higher in females than in males, and concentrations of leptin increased significantly with increasing birth weight, birth weight according to gestational age, and ponderal index (all P values <0.0001). Leptin concentrations in breast milk were lower in women who gave birth to a boy than in women who gave birth to a girl (P = 0.04), but no relationship existed with the anthropometric measurements at birth.
The correlations of adiponectin and leptin in different compartments are shown in Table 4⇓ . Adiponectin concentrations were moderately correlated in breast milk and maternal serum (Spearman ρ, 0.43; P <0.0001), but no correlations were found between concentrations in maternal serum and cord blood or between breast milk and cord blood. Leptin concentrations, likewise, showed the highest correlation between breast milk and maternal serum (Spearman ρ, 0.43; P <0.0001) and weak positive correlations between the other compartments.
Within the same compartment, adiponectin and leptin concentrations were positively associated in cord blood (Spearman ρ, 0.25; P <0.0001) and in breast milk collected 6 weeks postpartum (Spearman ρ, 0.20; P <0.0001), but no correlation was found for adiponectin and leptin in maternal serum (see Table 5⇓ ).
In the 42 breast milk samples collected 6 months postpartum, adiponectin concentration was 1.6–50.2 g/L, and leptin concentration was 5.8–723.0 ng/L (data not shown). The correlations between adiponectin and leptin in breast milk collected 6 weeks and 6 months postpartum are shown in Table 6⇓ . As in breast milk collected 6 weeks postpartum, adiponectin and leptin were positively correlated in breast milk collected 6 months postpartum. Also, adiponectin in breast milk collected 6 weeks postpartum and in breast milk collected 6 months postpartum were positively correlated (Spearman ρ, 0.47; P = 0.002), and an even stronger correlation was seen for leptin in breast milk collected 6 weeks postpartum and in breast milk collected 6 months postpartum (Spearman ρ, 0.64; P <0.0001).
In this large, population-based study of mothers and their newborns, for both adipokines we found a moderate correlation between concentrations in maternal serum and breast milk. In contrast, only cord blood leptin concentrations were weakly correlated with maternal serum leptin concentrations, and adiponectin concentrations in cord blood were considerably higher than those in maternal serum, suggesting that the fetal tissue may produce high amounts of adiponectin. In addition, breast milk was found to be a further source of adipokines for breastfed infants.
leptin and adiponectin in breast milk
Recent data about leptin in breast milk showed variations with respect to concentrations, mainly according to the breast milk fractions and the sample treatments used (20)(23)(24)(25). In general, leptin concentrations were much higher in whole breast milk than in skim breast milk (18)(26)(27). These results may indicate that leptin is linked to the fat globules in breast milk (26), or it could be an artifact of interference between milk fat and the RIA used (24)(27).
In breast milk collected 6 weeks postpartum in our study population, the mean concentration of leptin was comparable to that previously reported for leptin in skim breast milk also prepared by centrifugation of whole breast milk in studies conducted in relatively small populations consisting of 4 to 34 mothers (17)(18)(26)(27). However, the wide variation in breast milk leptin concentrations (0.0–4119.0 ng/L in our study population) makes it clear that enormous random error is possible if the distribution of leptin concentrations in breast milk is based on a small sample of mothers.
In agreement with previous studies, we found a significant positive correlation between leptin in maternal serum and breast milk (17)(18)(19)(20), with much lower concentrations in breast milk than in serum, in accordance with findings by Ucar et al. (19) but not described by others (17)(20). The presence of leptin in breast milk raises questions concerning the ability of the mammary epithelial cells to transfer leptin from the blood or to synthesize it before its secretion (24). Although 1 study has suggested the existence of leptin transfer from the blood to the breast milk (17), probably involving leptin receptors produced by the mammary gland (28)(29), other reports have shown mammary synthesis of leptin (26)(30)(31).
Martin et al. (32) recently demonstrated the presence of adiponectin in breast milk in 2 distinct small populations from Cincinnati and Mexico and reported adiponectin concentrations measured by RIA of 4.2 to 87.9 μg/L in skim breast milk, results comparable to those of our study. Because our study is the first description of the relationships of adiponectin concentrations in human breast milk with adiponectin in maternal serum and cord blood, it is impossible to put all of our findings in context. To the best of our knowledge, no information is available as to whether adiponectin in breast milk is associated with fat globules, as has been suggested for leptin (26)(29). Although the mean concentration of adiponectin in the fat-free phase of breast milk was much lower than that of adiponectin in cord blood, for breastfed neonates, breast milk might be an additional source of adiponectin, whose role in infant development warrants further investigation. Because adiponectin concentrations were positively associated in breast milk and maternal serum, and because a positive correlation exists between adiponectin and leptin in breast milk, it can be speculated that mammary epithelial cells are able to transfer adiponectin from the blood or to synthesize it before its secretion, as has been suggested for leptin (31).
Adiponectin and leptin concentrations in breast milk samples collected 6 months postpartum were in comparable ranges, and the agreement was quite high compared with breast milk samples collected 6 weeks postpartum. Therefore, prolonged breastfeeding might be an additional source of these adipokines in addition to adipokine production by infant adipose tissue.
The effects of breast milk leptin and adiponectin on the human neonate are not known. However, some evidence exists in nursing rats that leptin can be transferred via milk to the stomach and afterward into the neonatal rat circulation (17), and that neonatal rats given orally administered leptin showed several metabolic adaptations such as lower food intake, lower leptin production in the stomach and subcutaneous adipose tissue, and lower thermogenic capacity (33). Studies on mice suggest that leptin has a different role in neonates than in adults. High leptin concentrations lead to appetite suppression in adult mice, but in mouse neonates, despite their obvious need for maximum nutrition, there is a surge in the concentration of circulating leptin. Exogenous administration of leptin to mouse neonates altered the production of neuropeptides known to affect appetite in adult mice but did not alter appetite in neonatal mice. Studies on leptin deficient (ob/ob) mice further suggest that leptin concentrations influence hypothalamic neurons involved in regulation of food intake and thus play a role in the programming of appetite control during early postnatal life (34).
leptin in maternal serum and cord blood
Leptin is released into the circulatory system by the adipose tissue in proportion to the amount of lipid stores and acts at the hypothalamic receptors, decreasing food intake and increasing energy expenditure (35). The maternal blood concentration of leptin increases in the 3rd trimester and dramatically decreases to prepregnancy concentrations around parturition (4)(36)(37). In pregnancy, as in some forms of obesity, leptin resistance may result from inhibited transport across the blood-brain barrier or sequestration of bioactive leptin in the circulation by a soluble receptor (35). This increase in maternal leptin concentrations during pregnancy could result from the contribution of adipose depots (31) or from the synthesis of leptin by the placenta (4)(6)(38). Leptin produced in the human placenta is secreted into both maternal and fetal circulation (4)(6). Placental leptin production, however, makes a substantial contribution only to maternal circulating leptin concentrations. Umbilical leptin concentrations were found to be independent of placental leptin production and closely reflected ponderal index at birth (6). The positive associations in our study of cord blood leptin concentrations with birth weight, birth weight according to gestational age, and ponderal index might confirm the finding in previous investigations that leptin in fetal circulation is mainly produced by the adipose tissue of the fetus (6).
In our study population, leptin concentrations in cord blood were positively, but only weakly, associated with leptin concentrations in maternal serum. Lepercq et al. (6) did not find a statistically significant association between fetal and maternal leptin concentrations, but their study, which included only 74 infants, had very limited power. Because maternal leptin concentrations are related to maternal body mass index and cord blood concentrations are related to ponderal index at birth, the positive association between leptin concentrations in maternal serum and cord blood in our study population might reflect a positive association between maternal and fetal body weight.
Leptin concentrations found in cord blood in our study were higher than those previously reported (7)(39)(40) from studies in which the concentrations were measured by RIAs and the study populations were smaller than ours. Also, differences in characteristics of the study populations, such as ethnic background or anthropometric measures of newborns, might contribute to the observed differences. However, our results of leptin concentrations in cord blood and maternal serum are in agreement with those reported by Sivan et al. (11), who also found higher mean concentrations in maternal serum than in cord blood (30.9 μg/L vs 10.1 μg/L).
adiponectin in maternal serum and cord blood
In our study, adiponectin concentrations in cord blood were much higher than in maternal serum. This observation suggests that the fetal adipose tissue also produces adiponectin. Mean adiponectin concentrations in cord blood and maternal serum in our study population are in the same range as reported in previous much smaller studies (7)(9)(11)(41). As in previous studies, we found no correlation of adiponectin in cord blood and maternal serum (11)(41).
In accordance with other studies, we found a positive association between cord blood adiponectin and birth weight (11)(40)(41) and no clear relationship with degree of adiposity (7). In contrast, other studies found no relationship between cord blood adiponectin and birth weight (9)(39), but they did find a relationship between cord blood adiponectin and the sum of skinfold thickness as a measurement of adiposity (40).
Positive correlations of adiponectin and leptin concentrations in cord blood, as in our study population, were also reported by Sivan et al. (11) and Tsai et al.(40). Lindsay et al. did not report this correlation (7), possibly because of limited statistical power (n = 46 mothers). The positive correlation between adiponectin and leptin in cord blood, in contrast to the absence of any correlation between these adipokines in maternal serum, may indicate different biological mechanisms in production or regulation of these adipokines in fetal and adult individuals.
Limitations of our study include the lack of measurements of leptin and adiponectin in fore- and hindmilk, but Ucar et al. (19) reported for leptin that concentrations did not differ. Furthermore, weaning influences the concentrations of certain constituents of breast milk, as well as total breast milk production. However, in our study population the majority of mothers were still breastfeeding their children exclusively 6 weeks postpartum and had not started weaning. Therefore, it is unlikely that weaning influenced the concentrations of leptin and adiponectin in breast milk collected at that time. We were unable to take the time of sample collection into account, and therefore we were not able to give information about possible circadian cycles of these adipokines in the different compartments. Furthermore, results of recent studies have indicated that different adiponectin isoforms may have distinct biological functions (42)(43). We were not able to determine these adiponectin isoforms because our ELISA could not distinguish between low–molecular-mass trimer forms of adiponectin and high–molecular-mass complexes. In addition, it is not known whether the ELISA used has equal affinity for the different isoforms of adiponectin.
Despite these limitations, our data provide a simultaneous description of adiponectin and leptin in maternal serum, cord blood, and breast milk among a large, population-based group of mothers and their newborns. The different patterns suggest possible roles of adipokines in fetal and postnatal development. Such roles require clarification in further, particularly longitudinal, studies. Longitudinal studies are also necessary to clarify whether adipokines produced within the intrauterine and neonatal environment play an important role in fetal or neonatal programming of diseases occurring in adulthood, such as diabetes mellitus, hypertension, and coronary heart disease.
This study was supported by grants of the German Research Council (BR 1704/3-1, 3-2, 3-3). We thank Gisela Breitinger (Department of Clinical Epidemiology and Aging Research, German Cancer Research Center) for the collection of breast milk samples and Prof. Wolfgang Koenig and Gerlinde Trischler (Department of Internal Medicine II-Cardiology, Medical Clinic Ulm) for the measurement of adiponectin and leptin.
- © 2006 The American Association for Clinical Chemistry