Apelin is a novel regulator of human trophoblast amino acid transport

Abstract

Apelin is an insulin-sensitizing hormone increased in abundance with obesity. Apelin and its receptor, APJ, are expressed in the human placenta, but whether apelin regulates placental function in normal body mass index (BMI) and obese pregnant women remains unknown. We hypothesized that apelin stimulates amino acid transport in cultured primary human trophoblast (PHT) cells and that maternal circulating apelin levels are elevated in obese pregnant women delivering large babies. Treating PHT cells with physiological concentrations of the pyroglutamated form [Pyr¹]apelin-13 (0.1–10.0 ng/ml) for 24 h dose-dependently increased System A amino acid transport (P < 0.05) but did not affect System L transport activity. Mechanistic target of rapamycin (mTOR), extracellular signal-regulated kinase-1/2 (ERK1/2), and AMP-activated protein kinase-α (AMPKα) signaling were unaffected by apelin (P > 0.05). Plasma apelin was not different in obese women (BMI 35.8 ± 0.7, n = 21) with large babies compared with normal-BMI women (23.1 ± 0.5, n = 16) delivering normal birth weight infants. Apelin was highly expressed in placental villous tissue (20-fold higher vs. adipose), and APJ was present in syncytiotrophoblast microvillous membrane, but neither differed in abundance between normal-BMI and obese women. Phosphorylation (Thr¹⁷²) of placental AMPKα strongly correlated with microvillous membrane APJ expression (P < 0.01, R = 0.63) but negatively correlated with placental apelin abundance (P < 0.01, R = −0.62). Neither placental APJ nor apelin abundance correlated with maternal BMI, plasma insulin, birth weight, or mTOR or ERK1/2 signaling (P > 0.05). Hence, apelin stimulates trophoblast amino acid uptake, establishing a novel mechanism regulating placental function. We found no evidence that apelin constitutes an endocrine link between maternal obesity and fetal overgrowth.

INTRODUCTION

Apelin is a peptide hormone secreted by adipose tissue that has potent insulin-sensitizing effects (5, 11). Apelin circulates in biologically active 36-, 17-, and 13-amino acid mature forms, along with the pyroglutamated form [Pyr¹]apelin-13 (30). Both apelin and its G protein-coupled receptor APJ are expressed in the nutrient-transporting syncytiotrophoblast of the human placenta (10, 17, 36, 40). Placental apelin and APJ abundance are reduced in pregnancy complications associated with impaired placental function and restricted fetal growth, like preeclampsia and intrauterine growth restriction (12, 17, 36, 40). Maternal apelin concentrations and placental APJ expression are also reduced in experimentally undernourished rats that deliver growth-restricted young (28). Conversely, exogenous apelin administration enhances maternal-fetal glucose transport in vivo (28), and APJ agonists have recently been shown to regulate placental angiogenesis in genetically modified mice and enhance invasiveness of choriocarcinoma cells (15). Taken together, these observations suggest that apelin may regulate placental function.

In nonplacental tissues, apelin stimulates intracellular Akt, mechanistic target of rapamycin (mTOR), extracellular signal-regulated kinase-1/2 (ERK1/2), and AMP-activated protein kinase-α (AMPKα) signaling (30). The activities of these signaling pathways in the human placenta have been shown to link maternal nutrient availability, placental nutrient transport capacity, and fetal growth (9, 20, 31). In particular, trophoblast mTOR signaling is activated by insulin (18, 21) and stimulates System A-mediated secondary active transport of neutral amino acids and System L-mediated exchange of essential amino acids (33), both key determinants of fetal growth (37). Thus, it is plausible that apelin stimulates trophoblast amino acid transport by activating insulin and mTOR signaling. However, the effects of apelin on nutrient transport in human trophoblasts have not been studied.

Plasma apelin concentrations and adipose tissue apelin expression are increased in nonpregnant obese humans and experimental animals (5, 7, 14, 35). In obese pregnant women, perturbed concentrations of insulin and adipose tissue-derived hormones like leptin and adiponectin may promote excessive fetal growth by overstimulating placental function and nutrient transport (3, 18, 19, 32, 37). Plasma apelin, placental apelin, and APJ expression are all increased in pregnant rodents with diet-induced obesity (13), but it remains unknown whether maternal obesity alters circulating apelin or placental apelin signaling in pregnant women.

We hypothesized that 1) apelin activates Akt, mTOR, ERK1/2, and AMPK signaling and stimulates System A and System L amino acid transport in isolated primary human trophoblast cells; 2) APJ is present in the maternal facing microvillous membrane of human placenta; and 3) placental insulin and mTOR signaling are activated, and plasma apelin, placental apelin, and placental APJ are increased in obese pregnant women in association with fetal overgrowth.

METHODS

Blood and tissue collection.

All samples were collected with written informed consent and ethical approval from the Institutional Review Boards of the University of Colorado (maternal blood, placenta) or Oslo University Hospital (adipose tissue). Maternal venous blood and placentae were collected from normal body mass index (BMI; 19.7–28.8, n = 16) and obese (BMI 29.9–39.6, n = 21) women at term, following spontaneous vaginal delivery or planned cesarean section. Blood samples from the same cohort of women were collected into an EDTA-coated tube and centrifuged, and the plasma was separated and stored at −80°C. Clinical characteristics of the study cohort are given in Table 1. Maternal omental and subcutaneous adipose tissues were collected from a separate cohort of normal-BMI (n = 5) and obese (n = 4) women during planned cesarean section, as part of another study (29).

Table 1.

Clinical characteristics of the study cohort

	Normal BMI	Obese
N	16	21
Maternal age, yr	30.1 ± 1.3	30.9 ± 1.1
Ethnicity (no. of Hispanics)	1	8
Maternal BMI	23.1 ± 0.5	35.8 ± 0.7*
Gestational age at delivery, wk	39.0 ± 0.3	39.2 ± 0.2
Delivery type (no. vaginal)	3	3
Placenta weight, g	598 ± 35	742 ± 44*
Sex (no. of female infants)	10	9
Birth weight, kg	3.23 ± 0.09	3.78 ± 0.15*
Plasma insulin, μIU/ml	9.8 ± 1.1	25.1 ± 10.6
Plasma apelin, ng/ml	0.462 ± 0.027	0.493 ± 0.056

Values are means ± SE. BMI, body mass index.

^*P < 0.05 vs. lean (Student’s t-test).

Isolation, culture, and treatment of primary human trophoblast.

Primary human trophoblasts were isolated from fresh placental villous tissue collected from normal-weight women (n = 8) by trypsin digestion and discontinuous Percoll gradient separation, as described previously (25, 32). Trophoblasts were plated at 1.4 × 10⁶ per well in six-well plates and cultured in a 1:1 mixture of Dulbecco’s modified Eagle’s medium (DMEM; 25 mM glucose) and Ham’s F-12 medium (10 mM glucose), with 10% fetal bovine serum, glutamine, and antibiotics (32). At 66 h in culture, to allow for differentiation, cells were treated with [Pyr¹]apelin-13 (Abcam, Cambridge, UK) at a final concentration of 0.1, 1.0, or 10.0 ng/ml, or vehicle (DMEM). Twenty-four hours later, trophoblasts were washed with PBS and either lysed in radioimmunoprecipitation assay buffer with protease and phosphatase inhibitors (1:100, P8340, P2850, P0044, Sigma-Aldrich) and then snap-frozen or used for measurement of amino acid uptake, as described below. Trophoblast purity and absence of mesenchymal contamination were ascertained by Western blot analysis of cell lysate cytokeratin 7 (OV-TL 12/30; Dako, Santa Clara, CA) and vimentin (ab20346, Abcam) expression, respectively. Compared with crude placental homogenate, the relative abundance of cytokeratin 7 in syncytialized trophoblast cells was 3.24 ± 0.25. Vimentin was undetectable by Western blot in syncytialized trophoblast cell cultures.

Trophoblast System A- and System L-mediated amino acid transport was determined in syncytialized trophoblast at 90 h in culture as the rate of Na⁺-dependent uptake of [¹⁴C]methylaminoisobutyric acid (MeAIB) and 2-amino-2-norbornanecarboxylic acid (BCH)-inhibitable [³H]leucine (Leu) uptake over 8 min, as described previously (32). Preliminary experiments established that uptake of both tracers was linear with respect to time over 8 min.

Placental homogenization and isolation of syncytiotrophoblast microvillous and basal plasma membranes.

To determine APJ membrane localization in the placenta, villous tissue collected from subsets of both lean (n = 10) and obese (n = 12) women was dissected and homogenized in ice-cold buffer (10 mM HEPES-Tris, 250 mM sucrose, pH 6.95) with protease and phosphatase inhibitors. Syncytiotrophoblast microvillous (MVM) and basal (BM) plasma membranes were isolated from the homogenized villous tissue by differential centrifugation, Mg²⁺ precipitation, and sucrose gradient separation, as described previously (16) The relative enrichment of MVM and BM preparations was measured by alkaline phosphatase activity kinetic assay and ferroportin immunoblot, respectively, and was determined to be more than 10-fold in all preparations used.

Western blotting.

The protein abundance and phosphorylation of readouts of mTOR, ERK1/2, and AMPK signaling were determined by Western blot in lysates from [Pyr¹]apelin-13-treated trophoblasts (n = 7 placentas) and placental homogenates (n = 9 normal-BMI, n = 12 obese). Total expression and phosphorylation of AMPKα (Thr¹⁷²), ERK1/2 (Thr^202/204), and S6 ribosomal protein (Ser^235/236), as a functional readout of mTOR complex 1 signaling, were determined using primary antibodies generated in rabbits (Cell Signaling Technology, Danvers, MA). Placental homogenates and adipose tissue were also immunoblotted for full-length apelin, detected as a 16-kDa dimer, (mouse-derived, sc-293441; Santa Cruz Biotechnology, Dallas TX), while APJ expression was determined in MVM and BM (rabbit-derived, ABD43, EMD Millipore). Immunoblots were visualized using enhanced chemiluminescence.

Maternal plasma apelin and insulin measurement.

Maternal plasma samples from normal-BMI (n = 14) and obese (n = 16) women were analyzed using commercial ELISAs for insulin (80-INSHU-E01.1; Alpco, Salem, NH) and apelin (EKE-057-15; Phoenix Pharmaceuticals, Burlingame, CA). For the insulin assay, the limit of detection was 0.4 μIU/ml, and the intra- and interassay coefficients of variability were 6 and 11%, respectively. For the apelin assay, the limit of detection was 0.11 ng/ml, and the mean intra- and interassay coefficients of variability were 10 and 15%.

Statistics.

Results are presented as means ± SE. The effect of exogenous [Pyr¹]apelin-13 on amino acid transport and signaling in trophoblast cells was determined by one-way ANOVA with repeated measures and post hoc pairwise comparisons of treatments by Holm-Sidak test. APJ protein abundance in MVM and BM samples matched within individual placentae was compared by paired Student’s t-test. Clinical characteristics, plasma hormone concentrations, and placental protein abundances of normal-weight and obese women were compared by unpaired Student’s t-test. Linear interdependence of variables in the study subjects was determined by Pearson’s product-moment correlation. In all cases, significance was accepted at the level P < 0.05.

RESULTS

Effect of exogenous [Pyr¹]apelin-13 on cultured primary human trophoblast.

Exogenous [Pyr¹]apelin-13 stimulated System A amino acid transport activity in primary human trophoblast cells in a dose-dependent manner, such that sodium-dependent [¹⁴C]MeAIB uptake was significantly increased from control values by 37 and 42% in the presence of 1.0 and 10.0 ng/ml [Pyr¹]apelin-13 concentrations, respectively (Fig. 1A). By contrast, [Pyr¹]apelin-13 did not affect System L-mediated [³H]leucine uptake (Fig. 1B). Moreover, neither the total abundance nor the phosphorylation of S6, ERK1/2, or AMPKα were altered in primary human trophoblast after 24 h of [Pyr¹]apelin-13 treatment (Fig. 2).

Fig. 1.

Effect of treatment with physiological concentrations of the pyroglutamated form [Pyr¹]apelin-13 on System A and System L amino acid transport in cultured primary human trophoblast. MeAIB, methylaminoisobutyric acid; BCH, 2-amino-2-norbornanecarboxylic acid. Effect of treatment determined by one-way ANOVA with repeated measures and significance determined at the level P < 0.05. *Significantly different from control (Con) by Holm-Sidak post hoc test; n = 8 placentas. Values are means ± SE.

Fig. 2.

Effect of treatment with physiological concentrations of the pyroglutamated form [Pyr¹]apelin-13 on mechanistic target of rapamycin (mTOR), extracellular signal-regulated kinase-1 and -2 (ERK1/2), and AMP-activated protein kinase-α (AMPKα) signaling in cultured primary human trophoblast. p/phospho, phosphorylatedA: representative Western blots. B–G: effect of treatment determined by one-way ANOVA with repeated measures and significance determined at the level P < 0.05; n = 7 placentas. Values are means ± SE.

Placental APJ and apelin in pregnancies complicated by maternal obesity.

Maternal prepregnancy BMI, birth weight, and placenta weight were significantly greater, whereas maternal age and gestational age at delivery were similar, in the obese, compared with normal-BMI, groups of women (Table 1). By proportion, the distributions of ethnicities, delivery types, and infant sexes did not differ significantly between the two groups (Fisher’s exact test, P > 0.05; Table 1). Maternal plasma insulin concentration did not differ significantly with BMI but tended to be greater in obese than in normal-weight women (Table 1). Maternal plasma apelin concentrations also did not differ between normal-BMI and obese women (Table 1) and did not correlate with maternal plasma insulin (P > 0.05, Pearson correlation).

APJ was detected in syncytiotrophoblast plasma membrane vesicles as a 42-kDa band by immunoblot and was present in greater abundance in the maternal-facing MVM than in the fetal-facing BM of the term placenta (Fig. 3). However, there was no significant difference in MVM APJ abundance between placentae of normal-weight and obese women (Fig. 4A). Apelin peptide abundance was 20-fold greater in placental villous tissue than in maternal adipose tissue (relative abundance: placenta 1.00 ± 0.42, n = 10; adipose 0.05 ± 0.01, n = 18, P < 0.05; t-test) but was also similar in the placentae of normal-BMI and obese women (Fig. 4B).

Fig. 3.

Apelin receptor APJ abundance in human syncytiotrophoblast microvillous (MVM) and basal plasma membrane (BM). Protein abundance compared between MVM and BM by paired t-test. *P < 0.05; n = 3 placentas. Values are means ± SE.

Fig. 4.

Apelin receptor APJ and apelin abundance in placentas of normal-BMI and obese women. A: APJ protein abundance in microvillous mebrane (MVM) vesicles. B: apelin protein abundance in placental homogenate. Cer., cerebellum (positive control); Normal, n = 9–10 (N, ○); obese, n = 12 (O, ▲). Normal-BMI and obese groups were compared with Student’s t-test. Values are means ± SE.

As previously reported (20), maternal obesity was associated with activation of placental mTOR signaling, as indicated by a significant increase in the ratio of phosphorylated to total S6 ribosomal protein, although the individual abundances of phosphorylated and total S6 did not differ significantly between normal-BMI and obese women (Fig. 5A). The ratio of phosphorylated to total ERK1/2 was also higher in placentas of obese compared with lean women (Fig. 5A). This was driven by a significant reduction in total ERK1/2 abundance in obese women, without a change in phosphorylated ERK1/2 between the groups. Maternal obesity did not affect AMPKα abundance or phosphorylation or the ratio of the two (Fig. 5A).

Fig. 5.

Ribosomal protein S6, extracellular signal-regulated kinase-1 and -2 (ERK1/2), and AMP-activated protein kinase (AMPK) signaling in placentas of normal-BMI and obese women. A: representative blots and total and phosphorylated (phospho)S6, ERK1/2, and AMPK abundance. Cer., cerebellum (positive control). Values are means ± SE. *P < 0.05 vs. normal by Student’s t-test. B and C: correlation of placental phosphorylated AMPK with microvillous mebrane (MVM) APJ (D) and placental apelin (E). P values and Pearson’s correlation coefficients given in figure. Normal, n = 9–10 (N, ○); obese, n = 12 (O, ▲).

Placental phosphorylated S6 abundance and the ratio of phosphorylated to total S6 were both correlated with maternal BMI (R > 0.4, P < 0.05, n = 21 both cases). The phosphorylated-to-total S6 ratio also correlated with birth weight (R = 0.56, P < 0.05, n = 21). Total, but not phosphorylated, ERK1/2 abundance inversely correlated with birth weight (R = −0.46, P < 0.05, n = 21) but not maternal BMI (P > 0.05). Neither phosphorylated nor total AMPKα abundance correlated with maternal BMI or birth weight (P > 0.05). Placental APJ and apelin did not correlate with maternal BMI, plasma insulin, birth weight, or abundance of mTOR and ERK1/2 signaling proteins (P > 0.05 all cases). However, MVM APJ was strongly correlated with AMPKα phosphorylation (Fig. 5C) and the ratio of phosphorylated to total AMPKα (R = 0.69, P < 0.05, n = 21), whereas apelin abundance was negatively correlated with placental AMPKα phosphorylation (Fig. 5D) and the phosphorylated-to-total AMPKα ratio (R = −0.66, P < 0.05, n = 21).

DISCUSSION

This study is the first to determine the effect of the adipokine apelin on human trophoblast nutrient transport function and to assess placental apelin signaling in pregnant women. The results show that apelin stimulates trophoblast System A-mediated amino acid transport, a determinant of fetal growth. We also establish that placental APJ and apelin abundance are strongly related to placental AMPKα signaling in vivo but are not altered in pregnancies complicated by maternal obesity and fetal overgrowth.

Insulin stimulates trophoblast System A amino acid transport by activating mTOR signaling, whereas cytokines, such as tumor necrosis factor-α, stimulate System A through mitogen-activated protein kinase signaling (2, 32, 34). Apelin binding to APJ in vitro has been reported to activate intracellular mTOR and ERK1/2 signaling (30), consistent with a role for apelin as an insulin sensitizer. It was therefore unexpected that apelin increased System A-mediated uptake of MeAIB in primary human trophoblast cells without activating either mTOR or ERK1/2 signaling. System A and System L amino acid uptake in trophoblast is dependent on the abundance and activity of sodium-dependent neutral amino acid transporters (SNATs) and L-type amino acid transporters (LATs) in the plasma membrane. Activation of mTOR stimulates trophoblast amino acid uptake by promoting the translocation of specific SNAT and LAT isoforms to the cell membrane (33) while mitogen-activated protein kinase signaling enhances SNAT1 and SNAT2 protein translation (2). It has been established that apelin stimulates proliferation in an mTOR-dependent manner and increases ERK1/2 phosphorylation in nonplacental cell lines (4, 27). The absence of any change in S6 or ERK1/2 phosphorylation in apelin-treated cells in this study may reflect the fact that protein phosphorylation was assessed 24 h after trophoblast cell treatment, whereas the effects of apelin on kinase signaling pathways in vitro are typically short-lived (<3 h) (4). Alternatively, the effects of apelin on System A amino acid transport may be mediated by other mechanisms not investigated in this study. Apelin has in vivo cardioprotective effects that are independent of mTOR (23) and does not influence ERK1/2 signaling in primary human osteoblasts (39), suggesting that the kinase signaling pathways downstream of APJ are cell type specific and may be less important in primary cells. Differences in the specific pathways activated by APJ agonists may reflect cell type-specific heterodimerization of APJ with other membrane-bound receptors and differential G protein coupling (30).

Mean maternal plasma apelin concentrations measured here in both normal-BMI and obese study subjects were within the range of values previously reported in pregnant women (26, 36). However, the finding that maternal plasma apelin is similar in normal-BMI and obese women is in contrast with previous animal experimental data showing that apelin concentrations are elevated in obese pregnant rats (13). In nonpregnant humans, plasma apelin is substantially increased only in morbidly obese patients with type 2 diabetes (14, 35). Moreover, adipocyte apelin secretion is insulin dependent in vitro, and increased plasma apelin concentrations are observed only in obese mice with concomitant hyperinsulinemia (5). Therefore, the absence of an effect of maternal obesity on circulating apelin concentrations in the present study may be explained by the lack of significant difference in maternal plasma insulin concentrations between normal-BMI and obese women. Given that maternal plasma samples were collected both from women delivering vaginally and from those by cesarean section, intragroup variability in the time since the last prior meal might in turn explain interindividual variations in insulin and apelin concentrations, which are both suppressed by fasting (5). Indirect evidence also suggests that the placenta itself is an endocrine source of circulating apelin during pregnancy. Both rat and human placental tissues secrete apelin into the culture medium in vitro (28, 40), and maternal plasma apelin concentrations decrease in line with placental apelin protein abundance in pregnancies complicated by placental insufficiency (36). Our data also show that the relative abundance of apelin protein is greater in human placenta than in adipose tissue, suggesting that the placenta may be the main source of circulating apelin during pregnancy. Thus, differences in circulating apelin concentrations in normal-weight and obese women may also be obscured by similar placental secretion into the maternal circulation in both groups.

The present finding that APJ is localized to the syncytiotrophoblast microvillous membrane indicates that apelin in the maternal circulation can influence human placental function. A strong positive correlation between microvillous membrane APJ and phosphorylated AMPKα abundance furthermore suggests that placental apelin signaling may be mediated by AMPKα activation. Indeed, [Pyr¹]apelin-13 treatment also tended to increase AMPKα phosphorylation in primary trophoblasts in this study. AMPK stimulates glucose uptake in nonplacental tissues (1), whereas AMPKa silencing inhibits glucose uptake, glycolysis, mitochondrial respiration, and proliferation in mouse trophoblasts (6, 38). Apelin may therefore be a regulator of trophoblast glucose transport in addition to amino acid transport. Certainly, apelin stimulates skeletal muscle glucose uptake in an AMPK-dependent manner (11), while acute apelin infusion in pregnant rats enhances fetal tracer glucose accumulation (28). However, the relationship between apelin and glucose transport in the human placenta remains unknown. Given our data showing that villous tissue also abundantly expresses apelin, it is also unclear whether apelin acting at the microvillous membrane is of maternal endocrine, paracrine, or autocrine origin. Nonetheless, the inverse relationship between phosphorylated AMPKα and apelin abundance in the placenta suggests that AMPKα inhibits local apelin secretion in a negative feedback manner, potentially via its antagonistic effect on intracellular insulin signaling.

The absence of a correlation of placental apelin or APJ abundance with maternal BMI or placenta or birth weight suggests that placental apelin signaling does not represent a critical link between maternal nutrient availability and fetal growth per se. As infant body composition was not measured in this study, we cannot exclude a subtler association of placental apelin signaling with neonatal adiposity, which predicts later-life metabolic disease risk and can increase independently of birth weight in obese and diabetic mothers (8, 22). Nevertheless, the apelin-APJ system contrasts with placental mTOR complex 1 signaling, measured as S6 phosphorylation, which correlates with both maternal BMI and birth weight, consistent with its role as a nutrient sensor involved in enhancing fetal nutrient supply in obese women (9, 20, 31). Placental apelin abundance may instead be regulated through mechanisms independent of maternal nutrient availability or mTOR signaling. For example, low placental apelin in intrauterine growth-restricted placentae may be secondary to impaired vascular growth, since endothelial cells express apelin (12, 17, 24, 36, 40). Apelin signaling may therefore link trophoblast function and placental blood supply.

Overall, the study demonstrates that physiological concentrations of apelin regulate amino acid transport in primary human trophoblast cells and that the local abundance of apelin and its receptor are strongly linked to AMPKα signaling in the placenta. In contrast with our original hypothesis, circulating apelin does not constitute an endocrine link between maternal obesity and fetal growth. High expression of apelin in the placenta itself may mean that its more important effects are mediated in a paracrine/autocrine manner, although the specific signals that regulate its availability remain unclear. Our finding that apelin is a positive regulator of human trophoblast System A amino acid transport suggests that activating placental apelin signaling could be a useful therapeutic intervention to enhance fetal nutrient delivery in pregnancies complicated by impaired placental transport and restricted fetal growth.

GRANTS

This study was supported by National Institutes of Health Grant R01 HD-68370 and by NIH/NCATS Colorado CTSA Grant UL1 TR-002535.

DISCLAIMERS

Contents are the authors’ sole responsibility and do not necessarily represent official NIH views.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

O.R.V. conceived and designed research; O.R.V. performed experiments; O.R.V. analyzed data; O.R.V., T.L.P., and T.J. interpreted results of experiments; O.R.V. prepared figures; O.R.V. drafted manuscript; O.R.V., T.L.P., and T.J. edited and revised manuscript; O.R.V., T.L.P., and T.J. approved final version of manuscript.