Modulation of Apolipoprotein D levels in human pregnancy and association with gestational weight gain

Human pregnancy is associated with profound changes in the maternal lipid, carbohydrate and protein metabolisms. These changes are aimed at favoring the maintenance of pregnancy, sustaining fetal growth and brain development, and facilitating parturition [1]. Maternal metabolism is intimately linked with prepregnancy body mass index (BMI) and gestational weight gain (GWG), which greatly influence the outcome of pregnancy [2]. In association with an unhealthy lifestyle, suboptimal prepregnancy BMI and GWG also increase the risk of birth defects and chronic health problems in the children [3].

The maternal lipid metabolism during pregnancy is characterized by progressive increases in plasma cholesterol and triglyceride levels accompanied by increases in low density lipoprotein (LDL) and very low density lipoprotein (VLDL), leading to maternal hyperlipidemia during late pregnancy [4]. Cholesterol is essential for optimal embryonic and fetal development. It is an important component of plasma membranes, required for cell proliferation, differentiation and morphogenesis modulation [5]. It is also used by the placenta for the synthesis of steroid hormones [6]. Triglycerides are also crucial as they are used as a source of essential fatty acids for the fetus, although they do not directly cross the placental barrier [7].

Plasma lipid metabolism is regulated in part by the specific apolipoprotein constituents of the various lipoprotein classes. The pregnancy induced hyperlipidemia is accompanied by a rise in the plasma levels of some apolipoproteins, namely ApoA-1, ApoB, ApoC-II and ApoC-III. Apolipoprotein D (ApoD) is a secreted lipocalin assigned with many putative functions including lipid transport. Its macromolecular distribution extends from VLDL to very high density lipoproteins (VHDL) with a maximum concentration in high density lipoproteins 3 (HDL₃) [8]. It is considered an atypical apolipoprotein as both its structure and major sites of synthesis differ from the other apolipoproteins. Indeed, in human, ApoD is poorly expressed in the liver and intestine while it is highly expressed in adrenal glands, spleen, kidneys, pancreas, placenta, nervous system, lungs, ovaries and testes [[9], Rassart et al., unpublished results]. Several hydrophobic molecules were identified as its potential ligands including cholesterol [9], bilirubin [10], pregnenolone, progesterone, estrogens [11, 12] and arachidonic acid [13].

ApoD expression is modulated in many situations (for review see [14, 15]), namely cellular growth and differentiation [16‐19], cancer [19, 20], nervous system pathologies [21‐24] inflammation [25, 26] and oxidative stress [27‐29]. ApoD is also involved in gestation and fetus development. In human, it is up-regulated in the endometrium during the window of uterine receptivity for embryonic implantation [30], is highly expressed in placenta [9] and is found in colustrum and milk [31]. Its levels are also increased in fallopian tubes and ovaries of gestating compared to non-gestating guinea pigs [32] but are decreased in lactating compared to virgin mouse mammary gland [33]. During mouse development, ApoD expression is selectively modulated from embryonic day 8.5 (E8.5) to birth [34] and in brain, its expression coincides with the period of active myelination and synaptogenesis [35]. The role of ApoD in embryogenesis is not limited to mammals as it is found in the yolk of the rapidly growing chicken oocyte, where it might transport lipids or regulatory molecules such as vitamin A and thyroid hormones [36, 37], and expressed during late chicken embryogenesis [38].

In spite of the important role of ApoD in lipid transport, its expression in gestation-related tissues and during fetus development, and also despite numerous studies on other apolipoproteins, there is no available information concerning ApoD modulation throughout human pregnancy.

To address this question, we measured ApoD levels in plasma at each trimester of pregnancy, in cord blood, two months after delivery and in placenta. We also examined how ApoD levels are modified by factors affecting lipid metabolism such as maternal prepregnancy BMI, gestational weight gain and hypercholesterolemia.

To our knowledge, this study is the first to demonstrate that plasma ApoD levels decrease during pregnancy. Plasma ApoD levels were already significantly lower during first trimester of pregnancy than values detected in non-pregnant women and continued to decrease until the end of pregnancy. The mechanisms governing this decrease are currently unknown, but many potential reasons could explain it.

As for albumin, which concentration is lowered by 30% during pregnancy [45], the decrease in plasma ApoD levels could be explained in part by hemodilution. Maternal blood volume increases until term and although dependant on many factors, the average increase in volume at term is 45-50% [46]. Still, although it could explain the decrease in plasma ApoD levels at the first and second trimesters of pregnancy, hemodilution alone cannot explain the observed 70% decrease of plasma ApoD at delivery.

Low levels of plasma ApoD could also result from decreased ApoD synthesis or secretion. Plasma ApoD can originate from many sources, one of which being the placenta. Since ApoD mRNA levels are almost as high in placenta as in adrenal glands and spleen, major sources of ApoD in non-pregnant humans [9], decreased ApoD transcription is unlikely the origin of low plasma ApoD concentration. Another argument against ApoD transcriptional downregulation is the presence of oxidative and inflammatory markers mostly in complicated but also in normal pregnancies [47]. Oxidative stress [26‐29] and inflammation [25, 26] were both reported as intensifying ApoD mRNA production. ApoD decrease could also be caused by the inhibition of its translation or the increase of its degradation rate. More likely, the decline in ApoD levels could be imputable to its reduced secretion by placenta and other tissues during pregnancy. In addition, as prolonged exposure to physiological concentrations of 17 beta-estradiol inhibited markedly (70-90%) ApoD secretion in human breast cells [19], high estrogen concentrations, as seen during gestation [48, 49], could be the root of plasma ApoD decrease.

Assuming that ApoD synthesis and degradation rates remain unchanged, a reduced secretion will result in ApoD accumulation inside tissues. In the placenta, in collaboration with other molecules, ApoD could play a key role in the transport of essential nutrients, regulation factors and toxic metabolites between the mother and the fetus. Those could include cholesterol [9], arachidonic acid [13, 50], steroids [11, 12], vitamin A and thyroid hormones [37] as well as the removal of bilirubin [10].

It was already reported that ApoD levels in newborns represent about 37% of the levels found in adults [51, 52]. Accordingly, we observed that the concentration of circulating ApoD in the venous cord blood was very similar to that of the mother at delivery, which is about 30% of the levels in non-pregnant women, although the levels found in the cord blood and in the mother were not statistically correlated. This lack of correlation suggests that ApoD does not pass through the placental barrier. Since ApoD mRNA expression occurs as early as E8.5 during mouse embryogenesis [34], the fetus might not be dependent on a maternal ApoD supply. Furthermore, the similarity in mother and venous cord blood ApoD concentrations suggests that even if maternal plasma ApoD does not directly cross the placental barrier, both might be subjected to the same regulatory mechanisms.

We also observed that, while prepregnancy maternal BMI had no incidence on plasma ApoD concentrations on both mother and cord blood, high GWG (> 18 kg) significantly decreased ApoD values. This decrease is very likely multifactorial. In this particular case, it could, among others, be related to defects in lipid metabolism. Indeed, high GWG groups displayed lower levels of total cholesterol, LDL-cholesterol, and ApoB-100. This was present in normal and high BMI groups but the tendency, although not significant, was also observed in the low BMI group. This was also observed in the group composed of all women with high GWG regardless of their BMI. In addition, in this group, ApoD and lipid markers were strongly correlated. An involvement of ApoD in lipid metabolism was already suggested. Because of its capacity to bind cholesterol and its presence on lipoprotein fractions [8], ApoD, in collaboration with ApoA-I, lecithin-cholesterol acyltransferase (LCAT) and cholesteryl ester transfer protein (CETP) [53], may participate in the cholesterol transport pathway. Furthermore, genetic variation in ApoD affects the serum lipid levels [54].

Nevertheless, lipid levels are probably not the major factors responsible for low ApoD levels in these groups. Indeed, high GWG groups are not the only ones presenting levels of total cholesterol, LDL-cholesterol, and ApoB-100 different from the control group as well as correlations between ApoD and lipid levels. In addition, the low BMI group does not show a significant decrease of these lipid parameters. Finally, the ApoD concentration is similar in the control, HC and LC groups, the latter showing similarities with the high GWG groups.

Our results also suggest that a tight regulation of placental ApoD transcription and secretion is probably involved in the lower circulating ApoD concentrations in women with excessive GWG. In the view that ApoD may play a key role in the control of fetal homeostasis, the regulation of ApoD levels becomes vital for the pregnancy outcome and the health of the newborn. Thus, in women with high GWG, a decrease in circulating ApoD is desirable in order to reduce lipid transfer to fetus. This could be achieved both by reducing the ApoD transcription and by preventing ApoD from leaving the placenta. Consistently, high GWG have low ApoD mRNA levels and normal or high placental ApoD content. According to this hypothesis, the lower production of ApoD in low GWG women is disadvantageous. It would then be compensated by an increased secretion or lower degradation by the placenta and possibly other tissues resulting in adequate ApoD protein concentration in the circulation. In overweight women (high BMI), this control is probably altered, as evidenced by the high ApoD mRNA and low ApoD protein levels in the placenta independently of the GWG. This may explain why the plasma ApoD levels are similar in high GWG and in the control group. Still, plasma ApoD levels are not solely dependent on placental regulation as high BMI groups, in spite of high ApoD mRNA and low ApoD protein levels in the placenta, did not show higher plasma ApoD concentration than the control group. Similarly, despite low ApoD transcript levels and high protein content in the placenta, the low BMI, low GWG group had plasma ApoD levels similar to the control group. This suggests that other tissues have the ability to secrete or trap ApoD and that they may not be subjected to the same regulatory pathways than the placenta.

After delivery, the return to basal, non-pregnant, plasma ApoD levels, such as observed in breast-feeding women was expected. However, the fact that bottle-feeding women still have the same plasma ApoD levels than at delivery is intriguing and suggests again that the regulation of ApoD levels is complex. A possible explanation of this discrepancy is that breast-feeding mothers have high levels of prolactin. Prolactin was reported to exert an antagonistic action on estrogen production in order to maintain lactation [48, 55] and could thus favor ApoD production. In addition, breast-feeding induces the production of cytokines [56, 57], which may contribute to the higher levels of circulating ApoD.

Plasma ApoD levels decrease during normal uncomplicated human pregnancy. ApoD is further decreased in women with excessive GWG and their newborns. The cause of this decrease, controlled in part at the transcriptional level in the placenta, is probably multi-factorial and may be related to pregnancy-induced changes in lipids and hormone levels. Still, the connection between lipids, estrogen and apoD levels is speculative and requires further studies.