Blood cell transcript levels in 5-year-old children as potential markers of breastfeeding effects in those small for gestational age at birth

Intrauterine and early life events are thought to be very important in the development of chronic disorders, contributing to the developmental origins of cardiometabolic disease mainly in adults [1]. Much evidence supports the fact that low birth weight children have an increased risk to develop cardiovascular illnesses compared with normal weight ones [2]. Poor growth in utero contributes to insulin resistance, significant increased risk for type 2 diabetes mellitus, obesity, hypertension, dyslipidemia, and coronary heart disease [2, 3]. The time immediately before and after birth may be a sensitive period related to programming cardiometabolic risk [4]. Therefore, a critical window of opportunity to modify programming may exist during pregnancy and throughout the first years of life. Conditions or dietary interventions during the lactation period may even potentially reverse metabolic malprogramming due to adverse in utero conditions, as demonstrated in animal models of maternal undernutrition [5, 6].

In this regard, nutrition of the newborn during the early postnatal period seems to be of capital importance [7]. Different lines of clinical evidence show the protective effect of breastfeeding against childhood obesity and its comorbidities [8, 9]. Several theories have been proposed to explain this protective effect, including behavioral and nutritional explanations [10], such as differences in milk energy density and macronutrient composition [11], the impact of breastfeeding on microbiota composition [12], or the implication of specific bioactive compounds [13], particularly those that occur naturally in human breast milk and are absent in infant formula, such as leptin [14‐17]. It is likely that beneficial effects of breastfeeding are particularly relevant in low birth weight infants [18].

To be able to predict the increased risk of metabolic-syndrome related alterations due to early programming events is a challenge. This may allow the application of strategies at a time when there are more opportunities to prevent the increased risk, and before phenotypic alterations become evident. Thus, it must be helpful to have reliable, early biomarkers of increased propensity to these alterations and sensitive to nutritional interventions at early life.

Peripheral blood cells, considering whole peripheral blood (PBC) or a subpopulation of white blood cells composed basically by lymphocytes and monocytes, the so-called peripheral blood mononuclear cells (PBMC), are easily accessible from blood samples, and they have been proposed as a useful source of markers to potentially predict some aspects of the pathological and physiological state of the organism [19, 20]. We have previously described the presence of low expression levels of energy metabolism-related genes, such as INSR, CPT1A and FASN, in PBC of obese children with high plasma triglyceride levels, whereas children with high expression levels displayed plasma levels of TG similar to normal-weight children [21]. In the case of CPT1A, low expression levels were also related to higher HOMA index in obese children [21]. Therefore, expression levels of these genes were proposed as potential biomarkers of the metabolic status of obese children as they are indicative of the risk for the insulin-resistant or dyslipidemic state associated with obesity. Moreover, in a previous study we also found a relationship between breastfeeding and an increased expression of some of these protective biomarkers; breastfed children displayed higher expression levels of INSR and FASN in PBC compared with formula-fed subjects [22]. Determination of expression levels of these and other potential candidate genes related with energy homeostasis in children born with low or adequate weight may be of interest, particularly as they may reflect the beneficial effects of breastfeeding.

Therefore, the objective of the present study was to analyze the impact of birth weight and the type of feeding during infancy (breastfeeding or formula feeding) on the expression levels in PBC of selected candidate genes involved in energy homeostasis (CPT1, FAS, FTO, INSR, LEPR, MC4R, and UCP2) in a subset of children followed from birth to the age of 5 years to find out potential biomarkers of metabolic programming effects during this period of metabolic plasticity.

The present prospective study performed in children born at term after a non-complicated pregnancy shows significant differences in transcript levels in PBC of genes related to different metabolic pathways at the age of 5 years, according to birthweight and the type of infant feeding. Notably, expression levels were particularly different as an effect of the type of infant feeding in SGA children. These results are in agreement with the potential interest of gene expression levels in blood cells at early ages as a marker of metabolic programming effects associated with birth weight and the type of infant feeding, which may thereby predict the susceptibility to metabolic alterations in later life. In contrast, no relevant differences in body weight, metabolic profile or BP were found at 5 years of age, with the exception of uric acid, HDL-cholesterol and insulin levels. Nevertheless, in all cases, values of these parameters were in the normal range for age.

Regarding gene expression, one of the most significant results of the present study is the different expression of UCP2 in PBC in terms of type of infant feeding. UCP2 was downregulated in FF children compared with BF ones, and the differences were more marked among the SGA subjects. Differences in UCP2 expression levels according to the type of infant feeding have been previously described in a group of 237 children aged 2–9 years from eight European countries [22]. We previously described that breastfed male subjects showed higher expression levels of this gene compared with formula-fed ones [22]. Notably, our present results, considering both male and female, confirm those findings. In addition, these results also show the importance of BF among children in the SGA group, since it normalizes UCP2 expression levels to those of the AGA + BF group, which could be considered as the reference group in terms of future metabolic health.

UCP2 is involved in a broad range of pathological processes. Remarkably, we have found an inverse association between UCP2 expression levels in PBC and uric acid levels. This relationship can be explained by the function of UCP2. This protein is present in the inner mitochondrial membrane and it mainly acts in the protection against oxidative stress, decreasing the ATP levels and reactive oxygen species (ROS) produced by electron transport [39]. Uric acid overproduction may occur as a result of the acceleration of ATP degradation to AMP, a precursor of uric acid [40] and UCP2 could decrease the ATP level and thereby reduce AMP levels for uric acid formation [41]. Therefore, although a cause-effect relationship cannot be established from our results, the presence of increased UCP2 expression levels as mirrored in PBC might reflect an increased potential to regulate ROS and the ATP/AMP ratio, and thereby a decreased risk of hyperuricemia and other related metabolic complications. The inverse association between UCP2 expression levels in PBC and uric acid levels is also of interest regarding the relevance of uric acid levels in relation to cardiometabolic risk factors, particularly high blood pressure and elevated insulin and triglycerides [42]. Lurbe et al. have previously described that uric acid levels in children at 5 and 10 years were significantly and independently associated to birth weight [23, 43], and they were also positively correlated with blood pressure, insulin and triglycerides in overweight and obese youths [42].

Other signs on the potential connection between UCP2 expression and uric acid levels have been previously suggested from studies in adult Chinese population showing a relation between the risk of hyperuricemia and genetic variants, such as the − 866G/A (rs659366) polymorphism located in the promoter region of UCP2 [41]. Females with the A allele, which is associated with higher UCP2 mRNA expression, had lower serum urate and a decreased risk of hyperuricemia, but no relation was found in males [41]. Therefore, the relationship between transcript levels of UCP2 in PBC and uric acid levels in blood merits further consideration in studies regarding early prevention of cardiometabolic risk factors.

MC4R has a well-established role in feeding behavior and energy homeostasis [44]. Transcript levels of MC4R were higher in FF children compared with BF children, particularly in SGA infants. It is noticeable that, as mentioned above for UCP2, breastfeeding normalized expression levels of MC4R in SGA children to the “reference” levels of the AGA + BF group. The meaning of differences in the expression levels of MC4R in PBC is not known, and, to our knowledge, this is the first study addressing MC4R expression levels in PBC in relation to birth weight and type of infant feeding. In a previous study, high expression levels of this gene in PBC were associated with the intake of high percentage of energy from fat, but also with lower body fat percentage and BMI z-score in children [45]. Here, although dietary intake was not assessed, no differences were found in terms of anthropometric parameters between BF and FF children of the SGA group, and no significant associations were found between expression levels of this gene and anthropometric and biochemical measurements performed at 5 years of age. However, the small sample size could be the reason for the lack of correlations. In any case, present findings suggest that greater MC4R expression levels in PBC in children may be indicative of adverse programming events during intrauterine life, while normalization of transcript levels of this gene in children who were SGA may be indicative of adequate phenotypic reprogramming by adequate nutritional supply during lactation. Notably, considering the inverse trends between the expression levels of MC4R and UCP2 in SGA children according to the type of infant feeding, besides the potential interest of individual gene expression values by itself, the MC4R/UCP2 mRNA ratio could be of greater relevance as a potential, more sensitive marker of the beneficial effects of breast feeding in SGA infants.

Transcript levels of FASN in PBC have also been raised as potential biomarkers of metabolic alterations, particularly related with lipid metabolism. Specifically, in children, upregulation of FASN expression in PBC has been associated with lower triglyceride levels [21]. Moreover, FASN transcript levels in PBC have also been reported to be higher in children who were BF compared with FF ones [22]. Here, no relationship has been found between expression levels of this gene and triglyceride levels at 5 years of age (data not shown), which could be tentatively attributed to the limited number of subjects. Although triglycerides were within the normal range and no differences were found between groups. Nevertheless, the presence of higher FASN expression levels in BF compared with FF children, but only in SGA subjects, is also described here, in agreement with previous results [22]. Notably, association studies have revealed that FASN expression levels in PBC are positively related with insulin levels at 5 years of age, which were found to be higher in BF compared with FF children and the difference was apparently greater in children that were SGA at birth. The biological meaning of this association could be interpreted according to the known function of insulin activating FASN expression [46], and the fact that PBC may reflect changes in gene expression occurring in different tissues, such as liver, which has been described in animal models [20]. On the other hand, the negative association found between FASN expression levels and HDL-cholesterol is more difficult to interpret, but it could be tentatively related with the essential role of insulin regulating HDL-cholesterol metabolism, since it favors HDL2 to HDL3 conversion, through its action on hepatic lipase [47]. However, in both cases, the possible clinical interest of these associations, if any, remains to be established.

A similar pattern to that of FASN has been observed regarding FTO expression. Breastfeeding was associated with FTO upregulation in PBC, but only in the SGA group. This fact also supports the hypothesis that the benefits of breastfeeding could be even more important in SGA subjects, and that such transcript-based biomarkers could reflect these early events more accurately than other clinical markers.

Despite the recognized relevance of FTO gene variants on obesity predisposition [33, 34] the relation between FTO expression in blood cells and biochemical or clinical parameters related to metabolic syndrome is not well-defined. In adult humans, higher mRNA expression levels of FTO have been described to be greater in adipose tissue of obese than normal weight individuals, but no differences have been found regarding FTO expression in peripheral blood mononuclear cells (PBMC) [48]. Here, we have found a positive correlation between FTO expression levels in PBC and HDL-cholesterol at 5 years of age. Evidence on the relationship between FTO and cardiovascular disease (CVD) exists from studies showing the association of FTO gene variants with CVD risk, independently of BMI [49]. Some studies have described an association between the FTO gene rs9939609 polymorphism and HDL-cholesterol. Individuals carrying the risk allele, displayed decreased HDL-cholesterol concentration [50]. Nevertheless, whether or not the association found in the present study involves causality remains to be ascertained.

Unlike the aforementioned genes, expression levels of other candidate genes studied in the present study, CPT1L, INSR and LEPR were not significantly different between groups. Results regarding CPT1L and LEPR agree with previous published results showing no significant differences in their transcript levels in PBC between BF and FF children [22]. However, increased expression levels of INSR have been previously described in BF children compared with FF children [22]. Here, no differences have been found in AGA children, but BF children from the SGA group showed a non-significant trend to higher INSR expression levels compared to FF children (P = 0.098). The limited sample size could be the reason for the lack of statistical significance. This is in accordance with the general pattern described for other genes studied. Moreover, the finding of a negative correlation between transcript levels of INSR and SBP at 5 years of age may be considered in the light of the known link between insulin resistance and hypertension [51, 52]. In this regard, it is plausible to consider changes in the expression levels of INSR in PBC as an early molecular marker of future alterations frequently associated with hypertension.

The strengths and limitations of the present study must be taken into account. The study has a prospective design, with subjects recruited from birth, and includes data on pregnancy, birth, and early childhood growth, together with BP values and metabolic data, as well as data on gene expression analysis in blood cells. The study is focusing on an area with a lack of research: The relationship between BW and type of infant feeding, with gene expression in PBC as potential biomarkers of future metabolic status. However, one of the limitations of the study is the small sample size, which reduces statistical power. Another limitation may be the use of insulin levels and HOMA index as surrogate markers of insulin resistance, which differs from the gold standard method, the hyperinsulinemic euglycemic clamp [53]. Nevertheless, the use of this method is mainly restricted to investigational facilities and is normally unavailable in the clinical setting.

In summary, present results show differences in gene expression in PBC in 5-year-old children due to the type of infant feeding, which, interestingly, are more marked in infants born SGA compared with those in the AGA group. This is particularly evident for MC4R and UCP2, whose expression levels were increased and decreased, respectively, in FF children and especially in SGA, which may be indicative of adverse programming during intrauterine life. Therefore, although it needs to be independently confirmed in future studies, increased MC4R/UCP2 mRNA ratio in PBC may be a hallmark of FF infants born SGA; whereas normalization of this ratio could be indicative of adequate reprogramming by appropriate nutrition during lactation. Results also show differences in FASN and FTO expression levels in PBC due to the type of infant feeding that were observed only in SGA children. The finding of a negative correlation between transcript levels of UCP2 and uric acid levels is also relevant, considering the relation of uric acid with cardiometabolic risk factors, such as blood pressure and circulating insulin and triglycerides [42]. Therefore, interventions to avoid the elevation of uric acid levels should be considered and breast feeding might play an important role. Overall, these results suggest that SGA infants are more sensitive to the type of feeding during lactation and gene expression of particular genes in PBC may precisely reflect these effects in the absence of clear differences in phenotypic traits.

In terms of future work, it would be interesting to increase the sample size and power of analysis to be able to validate and make further correlations between gene expression and phenotypes and risk factors.