Chronic maternal infusion of full-length adiponectin in pregnant mice down-regulates placental amino acid transporter activity and expression and decreases fetal growth

Abstract

Maternal adiponectin levels are inversely correlated to birth weight, suggesting that maternal adiponectin limits fetal growth. We hypothesized that full-length adiponectin (fADN) infusion in pregnant mice down-regulates placental amino acid transporters and decreases fetal growth. Starting at embryonic day (E) 14.5, fADN (0.62 ± 0.02 μg (g body weight)⁻¹ day⁻¹, n= 7) or vehicle (control, n= 9) were infused in pregnant C57/BL6 mice by mini-osmotic pump. At E18.5, dams were killed and placental homogenates and trophoblast plasma membrane (TPM) vesicles were prepared. Infusion of fADN elevated maternal serum fADN by 4-fold and decreased fetal weights by 18%. Adiponectin receptor 2, but not adiponectin receptor 1, was expressed in TPM. fADN infusion decreased TPM System A (–56%, P < 0.001) and System L amino acid transporter activity (–50%, P < 0.03). TPM protein expression of SNAT1, 2 and 4 (System A amino acid transporter isoforms) and LAT1 and LAT2, but not CD98, (System L amino acid transporter isoforms) was down-regulated by fADN infusion. To identify possible mechanisms underlying these changes we determined the phosphorylation of proteins in signalling pathways known to regulate placental amino acid transporters. fADN decreased phosphorylation of insulin receptor substrate-1 (Tyr-608), Akt (Thr-308 and Ser-473), S6 kinase 1 (Thr-389), eukaryotic initiation factor 4E binding protein 1 (Thr-37/46 and Thr-70) and ribosomal protein S6 (Ser-235/236) and increased the phosphorylation of peroxisome proliferator-activated receptor α (PPARα) (Ser-21) in the placenta. These data suggest that maternal adiponectin decreases fetal growth by down-regulation of placental amino acid transporters, which limits fetal nutrient availability. This effect may be mediated by inhibition of insulin/IGF-I and mTOR signalling pathways, which are positive regulators of placental amino acid transporters. We have identified a novel physiological mechanism by which the endocrine functions of maternal adipose tissue influence fetal growth.

Key points

• Fetal growth is positively correlated to maternal adiposity, but the underlying mechanisms remain largely unknown.

• Maternal circulating levels of adiponectin, a hormone secreted by adipose tissue, are negatively correlated to maternal adiposity and fetal growth, suggesting that maternal adiponectin may limit fetal growth.

• Here we report that chronic administration of adiponectin to pregnant mice inhibits placental insulin and mammalian target of rapamycin (mTOR) signalling, down-regulates the activity and expression of key placental nutrient transporters, and decreases fetal growth.

• We have identified a novel physiological mechanism by which the endocrine functions of maternal adipose tissue influence fetal growth by altering placental function.

• These findings may help us better understand the factors determining birth weight in normal pregnancies and in pregnancy complications associated with altered maternal adiponectin levels such as obesity and gestational diabetes.

Introduction

Normal fetal growth is critical for short and long-term health because both restricted growth in utero and fetal overgrowth increase the risk for perinatal complications and predisposes for the development of obesity, diabetes and cardiovascular disease later in life (Barker et al. 1993; Boney et al. 2005). Fetal growth is strongly influenced by maternal nutrition and metabolism and there is a well-established positive correlation between maternal body mass index and birth weight (Ehrenberg et al. 2004). However, the mechanisms linking the amount of maternal adipose tissue to fetal growth remain largely unknown. Adiponectin is the most abundant protein secreted by white adipose and has well-established insulin-sensitizing effects (Kadowaki & Yamauchi, 2005). Full length adiponectin (fADN), which constitutes the predominant circulating form of adiponectin in the human, is a 28 kDa protein that forms multimeric complexes in vivo (Kadowaki & Yamauchi, 2005). A truncated form containing only the C-terminal portion (globular adiponectin) is produced by proteolytic cleavage and is biologically active. Pregnant women who are obese (Jansson et al. 2008) or have gestational diabetes mellitus (GDM) (Ranheim et al. 2004; Ategbo et al. 2006) typically have low circulating levels of adiponectin (ADN), which is correlated to increased fetal growth (Ategbo et al. 2006; Segal et al. 2008). Lean women, on the other hand, have high circulating levels of adiponectin (Jansson et al. 2008). In a recent report from Wang and coworkers, maternal serum adiponectin was found to be inversely correlated to fetal growth across the full range of birth weights with mothers giving birth to growth restricted fetuses having 60% higher levels than in pregnancies with normal birth weights and 4-fold higher serum adiponectin levels as compared to mothers giving birth to macrosomic babies (Wang et al. 2010). Collectively, these data are consistent with the possibility that maternal ADN may limit fetal growth.

Nutrient supply, which is dependent on placental nutrient transport, is an important determinant of fetal growth. The activity of key placental amino acid transporters is decreased in intrauterine growth restriction (IUGR) (Mahendran et al. 1993; Glazier et al. 1997; Jansson et al. 1998; Norberg et al. 1998) and up-regulated in fetal overgrowth (Jansson et al. 2002), indicating that changes in the activity of placental nutrient transporters may directly contribute to abnormal fetal growth (Sibley et al. 2005; Jansson & Powell, 2006, 2007). Furthermore, animal experiments suggest that the effect of altered maternal nutrition on fetal growth may be mediated by changes in placental nutrient transporters (Jansson et al. 2006; Jones et al. 2009; Coan et al. 2010; Rosario et al. 2011a).

More than fifteen amino acid transporters with distinct but often overlapping substrate specificity are expressed in the syncytiotrophoblast, the transporting epithelium of the human placenta (Jansson, 2001; Cleal & Lewis, 2008). System A is a sodium-dependent transporter mediating the uptake of non-essential neutral amino acids into the cell (Mackenzie & Erickson, 2004). All three known isoforms of System A, SNAT1 (SLC38A1), SNAT2 (SLC38A2), and SNAT4 (SLC38A4), are expressed in the placenta (Desforges et al. 2006; Novak et al. 2006). System L is a sodium-independent amino acid exchanger mediating cellular uptake of essential amino acids including leucine (Verrey et al. 2004). The System L amino acid transporter is a heterodimer, consisting of a light chain, typically LAT1 (SLC7A5) or LAT2 (SLC7A8), and a heavy chain, 4F2hc/CD98 (SLC3A2). In addition, LAT3 (SLC43A1) and/or LAT4 (SLC43A2) may contribute to the efflux of amino acids from the syncytiotrophoblast cell across the basal plasma membrane into the fetal compartment by non-exchange mechanisms (Cleal et al. 2011). The mechanisms by which placental nutrient transport is regulated remain to be fully established, but insulin, insulin-like growth factor I (IGF-I) and leptin have been shown to stimulate placental amino acid transporters (Karl et al. 1992; Jansson et al. 2003; Sferruzzi-Perri et al. 2006). Mammalian target of rapamycin (mTOR) is a ubiquitously expressed serine/threonine kinase that controls cell growth, which is primarily mediated by effects on protein translation (Hay & Soneneberg, 2004). We have recently shown that mTOR signalling is a positive regulator of trophoblast System A and L amino acid transporters (Roos et al. 2007, 2009a) and that glucose and growth factors are up-stream regulators of trophoblast mTOR (Roos et al. 2009b).

In previous studies, infusion of ADN in pregnant rats caused a decrease in mRNA expression of glucose transporter 3 and lipoprotein lipase in the placenta (Caminos et al. 2005). We recently demonstrated that fADN abolishes insulin stimulated amino acid uptake in cultured human primary trophoblast cells by modulating insulin receptor substrate (IRS) phosphorylation (Jones et al. 2010). These observations suggest that fADN causes insulin resistance in the human placenta, in contrast to the well-established insulin-sensitizing effect in skeletal muscle and liver (Kadowaki & Yamauchi, 2005). Furthermore, these data provide evidence for a link between maternal ADN and placental nutrient transport capacity and suggest that ADN may down-regulate placental nutrient transport functions. In this study, we tested the hypothesis that fADN infusion in pregnant mice inhibits placental insulin and mTOR signalling, down-regulates placental amino acid transporters and limits fetal growth.

Methods

Animals

All protocols were approved by the Institutional Animal Care and Use Committee at the University of Cincinnati. Eight-week-old female C57BL/6 mice (The Jackson Laboratory, Bar Harbor, ME, USA) were fed standard lab chow diet (Harlan Laboratories, Inc., Indianapolis, IN, USA) and water ad libitum. At 12–18 weeks of age, mice were mated and the presence of a vaginal plug was defined as embryonic day (E) 0.5.

Infusion of adiponectin

At E14.5, mice were anaesthetized using isoflurane and an osmotic pump (Model 1003D; Alza Corporation, Palo Alto, CA, USA) with a constant infusion rate of 1 μl h⁻¹ was implanted subcutaneously below the left scapular region. Pumps contained 0.75 μg μl⁻¹ (in phosphate-buffered saline (PBS)) of full-length murine recombinant adiponectin (Alexis Biochemicals, San Diego, CA, USA), giving an infusion rate of 0.62 ± 0.02 μg (g body weight)⁻¹ day⁻¹ (fADN-infused, n= 7) or vehicle only (Control, n= 9). The fADN dose and infusion rate were selected on the basis of earlier reports showing that 1–1.5 μg (g body weight)⁻¹ day⁻¹ administration of adiponectin (ADN) by mini osmotic pumps to non-pregnant mice results in increased circulating ADN (60–250%) (Xu et al. 2004; Shore et al. 2006).

Collection of blood and tissue samples

Dams were killed at E18.5 (by carbon dioxide inhalation using compressed gas in a cylinder as the source of CO₂), and maternal blood was collected by cardiac puncture, allowed to clot, and spun at 1000 g to collect serum. Serum was frozen in liquid nitrogen and stored at −80°C until analysis. After laparotomy, fetuses and placentas were removed and weighed, and placentas in each litter were pooled. Approximately 0.5 g of pooled placental tissue was washed in saline and transferred to 3 ml of buffer D (250 mm sucrose, 10 mm Hepes-Tris, and 1 mm EDTA (pH 7.4) at 4°C), protease and phosphatase inhibitor cocktails (Sigma-Aldrich Corp., St Louis, MO, USA) were added at a dilution of 1:1000, and the mixture was homogenized using a Polytron homogenizer (Kinematica, Bohemia, NY, USA). Homogenates were used for isolation of trophoblast plasma membrane (TPM) vesicles, as described below, or snap-frozen for subsequent analysis.

Serum analysis

Maternal serum levels of endogenous adiponectin at E18.5 were measured using an enzyme-linked immunosorbent assay (ELISA) by ALPCO (Salem, NH, USA). In addition, serum recombinant adiponectin protein levels were analysed by Western blotting as described previously (Yamauchi et al. 2001; Qiao et al. 2008), using an antibody raised against recombinant mouse full-length adiponectin (Alexis). The concentration of recombinant fADN in the serum samples was calculated using standards of known concentrations of recombinant fADN. For western blotting serum was diluted 1:5 with PBS. Maternal serum levels of insulin at E18.5 were also measured using an ELISA (Cystal Chem Inc., Downers Grove, IL, USA).

Isolation of trophoblast plasma membranes

TPM were isolated from mouse placenta using differential centrifugation and Mg²⁺ precipitation as described previously (Jones et al. 2009; Kusinski et al. 2010). This protocol results in the isolation of the maternal-facing plasma membrane of trophoblast layer II of mouse placenta and accumulating evidence suggests that this membrane is functionally analogous to the syncytiotrophoblast microvillous plasma membrane of the human placenta (Kusinski et al. 2010). Briefly, centrifugation steps were carried out at 4°C and all other steps were performed on ice. The homogenates were centrifuged at 10,000 g for 15 min, the supernatant was collected, and the pellets were resuspended, homogenized in 1 ml of buffer D, and centrifuged again at 10,000 g for 10 min. The two resulting supernatants were combined and centrifuged at 125,000 g for 30 min. The pelleted crude membrane fraction was resuspended in 2 ml of buffer D, and 12 mm MgCl₂ was added. The mixture was stirred slowly for 20 min on ice and then centrifuged at 2500 g for 10 min. The supernatant containing TPM vesicles was centrifuged at 125,000 g for 30 min. The final pellet was resuspended in buffer D using a Dounce homogenizer to yield the vesicle suspension. Protein concentration was determined using the Bradford assay and TPM purity was assessed by TPM/homogenate enrichments of alkaline phosphatase activity. Alkaline phosphatase enrichments for TPM isolated from placentas of control (12.4 ± 1.0, n= 7) and from fADN infused animals (12.1 ± 0.6, n= 7) were not different (P= 0.8, Student's unpaired t test).

TPM amino acid transporter activity measurements

The activity of System A and L amino acid transporters in TPM was determined using radiolabelled amino acids and rapid filtration techniques slightly modified from procedures previously described for human syncytiotrophoblast plasma membranes (Jansson et al. 2002) and mouse TPM (Kusinski et al. 2010). TPM vesicles were preloaded by incubation in 300 mm mannitol and 10 mm Hepes-Tris, pH 7.4 overnight at 4°C. Subsequently, TPM vesicles were pelleted and resuspended in a small volume of the same buffer (final protein concentration ∼5–10 mg ml⁻¹). Membrane vesicles were kept on ice until immediately prior to transport measurements when samples were warmed to 37°C. At time zero 30 μl vesicles were rapidly mixed (1:2) with the appropriate incubation buffer containing µ¹⁴C½methyl-aminoisobutyric acid (MeAIB, 150 μm) with or without Na⁺ or l-µ³H½leucine (0.375 μm). Based on initial time course studies (0–30 s, Fig. 2A and B), uptake at 15 s was used in all subsequent experiments. Uptake of radiolabelled substrate was terminated by addition of 2 ml of ice cold PBS. Subsequently, vesicles were rapidly separated from the substrate medium by filtration on mixed ester filters (0.45 μm pore size, Millipore Corporation, Bedford, MA, USA) and washed with 3 × 2 ml of PBS. In all uptake experiments, each condition was studied in duplicate for each membrane vesicle preparation. Filters were dissolved in 2 ml liquid scintillation fluid (Filter Count, PerkinElmer, Waltham, MA, USA) and counted. Appropriate blanks were subtracted from counts and uptakes expressed as pmol (mg protein)⁻¹. Na⁺-dependent uptake of MeAIB (corresponding to system A activity) was calculated by subtracting Na⁺-independent from total uptakes. For leucine, mediated uptake was calculated by subtracting non-mediated transport, as determined in the presence of 20 mm unlabelled leucine, from total uptake.

Figure 2. System A and System L amino acid transport activity in TPM.

System A activity was measured as Na⁺-dependent uptake of µ¹⁴C½MeAIB and System L activity was determined using mediated uptake of µ³H½leucine. A and B, in time course studies using TPM from control animals only, uptakes were linear over 30 s (System A; r= 0.93, P < 0.0001, n= 6; System L; r= 0.89, P < 0.0001; n= 6; Spearman's rho). C and D, System A and L amino acid transporter activity, measured at 15 s, in TPM isolated from control (C) and fADN (A) infused mice. Values are given as means ± SEM (A and B) or means + SEM (C and D). *P < 0.05 in fADN group (n= 6) versus control (n= 6); Student's unpaired t test.

Western blot analysis

Protein expression of total and phosphorylated S6K1 (Thr-389), 4E-BP1 (Thr-37/46 or Thr-70), S6 ribosomal protein (Ser-235/236), Akt (Thr-308), Akt (Ser-473), AMP-activated protein kinase α subunit (AMPKα) (Thr-172), and LKB1 (Ser-428) was analysed in placental homogenates using commercial antibodies (Cell Signaling Technology, Boston, MA). In addition, the expression of total and phosphorylated insulin receptor substrate-1 (IRS-1(Tyr-608)) and total and phosphorylated peroxisome proliferator-activated receptor α (PPARα) (Ser-12 and 21) was determined in placental homogenates using antibodies purchased from Millipore (Billerica, MA, USA) and Abcam (Cambridge, MA, USA), respectively. Adiponectin receptor1 (AdipoR1; Alpha Diagnostics, San Antonio, TX, USA) and Adiponectin receptor 2 (AdipoR2; Phoenix Pharmaceuticals Inc., Burlingame, CA, USA) expression were analysed in placental TPM preparations. Furthermore, we determined TPM protein expression of the System A amino acid transporter isoforms (SNAT) 1, 2 and 4, the System L amino acid transporter isoforms LAT1 and LAT2 and CD98, the heavy chain associated with LAT1 and 2. The SNAT1 antibody was produced as described previously (Gu et al. 2001). A polyclonal SNAT2 antibody generated in rabbits (Ling et al. 2001), was received as a generous gift from Dr V. Ganapathy and Dr P. Prasad at the University of Georgia, Augusta. Affinity-purified polyclonal anti-SNAT4 antibodies were produced in rabbits using the epitope YGEVEDELLHAYSKV in human SNAT4 (Eurogentec, Seraing, Belgium). Antibodies targeting the LAT1 and LAT2 were produced in rabbits as described previously (Park et al. 2005b). The CD98 antibody was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA) and anti-β-actin was from Sigma-Aldrich.

Western blotting was carried out as described (Roos et al. 2009a). Briefly, protein concentration was determined using the Bradford assay. Twenty micrograms of total protein from placental homogenate/TPM was loaded and separated on Bis-Tris gels (7% for S6K1, Akt, AMPKα, IRS1, LKB1, PPARα, AdipoR1/R2 and LAT1, 2 and CD98; 12% for 4E-BP1 and S6 ribosomal protein) and transferred onto nitrocellulose membranes. For SNAT1, 2 and 4, 2 μg TPM protein per well was separated on precast 4–12% Bis-Tris gels (Invitrogen). Preincubation of the primary antibody with its blocking peptide was used to confirm the specificity of the SNAT2 and 4 bands. Subsequent to membrane blocking and incubation in primary antibody, membranes were incubated with the appropriate secondary peroxidase-labelled antibodies for 1 h. After washing, bands were visualized using enhanced chemiluminescence (ECL) detection reagents (GE Healthcare, Chalfont St Giles, UK). Blots were stripped and re-probed for β-actin as a loading control. Analysis of the blots was performed by densitometry using an Alpha Imager (Alpha Innotech Corp., San Leandro, CA, USA). The expression of the target proteins was normalized to β-actin expression. For each protein target the mean density of the control (C) sample bands was assigned an arbitrary value of 1. Subsequently, all individual C and fADN density values were expressed relative to this mean. The rationale for this procedure is that it facilitates comparison between C and fADN groups, making it easier to assess the magnitude of changes.

Data presentation and statistics

Data are presented as means + SEM or ± SEM. For fetal and placental data, means of each litter were calculated and used in statistical analysis. Thus, n represents the number of litters. In the TPM amino acid uptake experiments, each condition was studied in duplicate. Spearman's rho correlation coefficient was used to determine whether the time courses of amino acid uptake were linear. Statistical significance of differences between control and fADN groups was assessed using Student's unpaired t test. A P value <0.05 was considered significant.

Results

Maternal, fetal and placental weights

Administration of fADN from E14.5 to E18.5 did not significantly affect maternal weight gain (data not shown). Maternal, fetal and placental weights are presented in Table 1. Fetal weights in mice receiving continuous infusion of fADN from E14.5 to E18.5 was 18% lower than in animals given infusion of vehicle (P= 0.006). Placental weight and litter size was not statistically different between the two groups.

Table 1.

Maternal, fetal and placental weights and serum adiponectin and insulin concentrations

Parameter	Control (n= 9)	fADN (n= 7)	P value
Maternal
Age at mating (weeks)	13.2 ± 0.8	13.6 ± 0.9	0.711
Pre-pregnancy weight (g)	19.9 ± 0.5	20.0 ± 0.9	0.913
Weight at E14.5 (g)	28.7 ± 0.4	29.1 ± 0.7	0.644
Weight at E18.5 (g)	35.2 ± 0.4	36.0 ± 1.3	0.585
Serum endogenous ADN (μg ml−1)	12.0 ± 1.1	10.2 ± 0.6	0.166
Serum recombinant ADN (μg ml−1)	ND (n= 8)	44.6 ± 2.6 (n= 7)
Serum insulin (pg ml−1)	85.9 ± 0.7	85.7 ± 0.4	0.700
Fetal
Litter size	7.3 ± 0.4	8.3 ± 0.4	0.113
Fetal weight (g)	1.18 ± 0.04	0.96 ± 0.05*	0.006
Placenta weight (g)	0.058 ± 0.002	0.052 ± 0.002	0.095
Fetal/placental weight ratio	20.4 ± 1.1	18.0 ± 1.5	0.199

Values are means ± SEM. Mice received infusion of full-length adiponectin (fADN) from E14.5 to E18.5. All serum measurements were non-fasting at E18.5. ND: not detectable. Endogenous adiponectin serum concentrations were determined by ELISA and levels of recombinant adiponectin by Western blot. All fetal variables were measured at E18.5. Statistical significance was determined using Student's unpaired t test;

^*P < 0.05 was considered significant.

Serum adiponectin and insulin levels

Serum adiponectin levels at E18.5 were not different between groups when measured using ELISA (Table 1, endogenous adiponectin). This is likely to be due to an inability of the antibody to recognize the infused recombinant full length ADN. Therefore, we confirmed the presence of recombinant fADN in maternal serum by Western blots, using an antibody raised against recombinant murine fADN and with different amounts of recombinant murine fADN loaded as standards. This approach has commonly been used in the literature (Yamauchi et al. 2001). No significant signal could be detected in serum from control animals (Fig. 1), suggesting that the anti-adiponectin antibody did not recognize endogenous ADN. In contrast, recombinant fADN standards were recognized in a dose-dependent manner (Fig. 1). Using these standards the levels of recombinant fADN in infused animals were calculated to be 44.6 ± 2.6 μg ml⁻¹ (Fig. 1 and Table 1). Thus, total adiponectin levels (i.e. endogenous + recombinant) were increased in fADN infused animals by approximately 4-fold as compared to controls. Serum insulin levels were not significantly different between groups (Table 1).

Figure 1. Serum levels of ADN in control and fADN infused mice measured by Western blot.

Proteins in serum samples (2 μl of 1:5 diluted sample) from dams infused with fADN (A) or vehicle (Control, C) and standards of recombinant fADN were separated by electrophoreses and probed using an antibody recognizing murine recombinant fADN. Standards (S) were 0.0185 (S1), 0.075 (S2), 0.15 (S3) and 0.3 μg (S4).

Decreased activity of System A and l-amino acid transporters in TPM isolated from fADN placentas

The Na⁺-dependent uptake of µ¹⁴C½MeAIB (corresponding to System A activity) and the mediated uptake of µ³H½leucine (corresponding to System L activity) were linear up to 30 s in TPM isolated from mouse placenta (Fig. 2A and B). Based on these time course experiments, an incubation time of 15 s was chosen in subsequent studies. As shown in Fig. 2C and D, full length adiponectin infusion in mice from E14.5 to E18.5 inhibited TPM System A uptake by 56% (P= 0.001), and System L uptake by 50% (P= 0.03) as compared to TPM isolated from placenta of control mice.

Maternal fADN infusion down-regulates System A and System L transporter isoform protein expression in TPM

We determined the effect of fADN infusion on TPM System A amino acid transporter isoform protein expression using Western blot. The SNAT1 antibody detected a band at approximately 54 kDa in TPM isolated from mice placenta. SNAT1 protein expression was significantly decreased (–42%, P= 0.04) in placental TPM of fADN mice (Fig. 3A and B). TPM protein expression of SNAT2 and 4 was identified at approximately 52 kDa. TPM SNAT2 and SNAT4 expression were down regulated by 40% (P= 0.006) and 47% (P= 0.0008), respectively, in the fADN infused group as compared to control mice (Fig. 3A and B).

Figure 3. Relative protein expression of System A amino acid transporter isoforms in TPM isolated from control and fADN infused mice.

A, representative Western blots for SNAT1, SNAT2 and SNAT4 in TPM isolated from mice placenta at E18.5. B, bar graph summarizing the Western blot data from control (C, n= 6) and fADN (A, n= 6) infused groups. After normalization to β-actin, the mean density of control samples was assigned an arbitrary value of 1. Individual C and A density values were expressed relative to this mean. Values are given as means + SEM. *P < 0.05 versus control; Student's unpaired t test.

TPM LAT1 protein expression was identified at 45 kDa (Fig. 4A). fADN infusion significantly reduced the TPM LAT1 protein expression by 24% (P= 0.03; Fig. 4A and D) as compared to control. Two distinct LAT2 bands at approximately 30 and 50 kDa were observed in TPM. These two bands correspond to different transcripts produced by alternative splicing (Pineda et al. 1999), and the two bands were analysed separately. TPM LAT2 (30 and 50 kDa) protein expression was significantly down-regulated (51% and 54%, respectively; P= 0.04) in TPMs from animals infused with fADN (Fig. 4B and D). CD98 expression was detected at 80 kDa in placental TPM and was comparable between control and fADN mice (Fig. 4C and D).

Figure 4. Relative protein expression of System L amino acid transporter isoforms in TPM isolated from control and fADN infused mice.

A–C, representative Western blots are shown for LAT1 (A), LAT2 (B) (30 kDa and 50 kDa) and CD98 (C) in TPM isolated from mice placenta at E18.5. D, bar graph summarizing the Western blot data from control (C, n= 6) and fADN (A, n= 6) infused groups. After normalization to β-actin, the mean density of control samples was assigned an arbitrary value of 1. Individual C and A density values were expressed relative to this mean. Values are given as means + SEM. *P < 0.05 versus control; Student's unpaired t test.

Inhibition of placental insulin signalling in response to maternal infusion of fADN

Insulin/IGFs act, at least in part, through the IRS-1/PI3-kinase/Akt signalling pathway (Kulik & Weber, 1998). Despite similar maternal insulin levels in the two groups (Table 1), we observed decreased phosphorylation of IRS-1 at Tyr-608 in homogenates of fADN placentas (–34%, P= 0.005) as compared to control (Fig. 5A and F). Phosphorylation of Akt at Thr-308 expression was reduced by 20% (P < 0.03) in fADN placentas compared to control (Fig. 5C and F). Similarly, phosphorylation of Akt at Ser-473 was decreased by 24% (P= 0.03) in the fADN group as compared to control (Fig. 5D and F). However, there were no significant differences between the fADN and control groups when analysing total placental IRS-1 and Akt expression (Fig. 5B, E and F).

Figure 5. Relative placental IRS-1 and Akt phosphorylation in control and fADN infused mice.

A–E, representative Western blots for phosphorylated IRS-1 (Tyr-608) (A) and total IRS-1 (B), phosphorylated Akt Thr-308 (C), phosphorylated Akt Ser-473 (D) and total Akt (E) in homogenates of pooled placentas from control and fADN infused mice. F, bar graph summarizing the Western blot data from control (C, n= 9) and fADN (A, n= 7) mice. After normalization to β-actin, the mean density of control samples was assigned an arbitrary value of 1. Individual C and A density values were expressed relative to this mean. Values are given as means + SEM. *P < 0.05 versus control; Student's unpaired t test.

Maternal fADN infusion inhibits placental mTOR signalling

We studied the phosphorylation of S6K1, 4E-BP1 and S6 ribosomal protein as functional read-outs for mTOR activity in placental homogenates of control and fADN infused mice (Figs 6–8). Maternal fADN infusion decreased the phosphorylation of S6K1 at Thr-389 by 38% in the placenta (P= 0.0001) as compared to control (Fig. 6A and B). Total S6K1 expression in placental homogenate was not significantly different between the control and fADN mice (Fig. 6A and B). Phosphorylation of 4E-BP1 occurs at multiple sites in an ordered manner. Phosphorylation by mTOR at Thr-37 and Thr-46 of 4E-BP1 may prime it for subsequent phosphorylation at sites including Ser-65 and Thr-70. Figure 7A and B shows representative Western blots using antibodies directed against 4E-BP1 phosphorylated at Thr-37/46 or at Thr-70. Both phosphorylation at Thr-37/46 (–55%, Fig. 7A and C; P= 0.001) and at Thr-70 (–58%, Fig. 7B and C; P= 0.0001), were significantly decreased in placentas of fADN infused mice. In contrast, total placental 4E-BP1 expression was comparable between control and fADN mice (Fig. 7B and C). Phosphorylation of ribosomal protein S6 (Ser-235/236), a component of the 40S ribosome and a physiologically relevant S6K1 substrate, was inhibited by 55% (P= 0.0001) in fADN placentas as compared to control (Fig. 8A and C). Figure 8B and C shows that there was no significant difference in the total S6 ribosomal protein expression level between control and fADN placentas.

Figure 6. Relative placental S6K1 phosphorylation in control and fADN infused mice.

A, representative Western blots of phosphorylated S6K1 (Thr-389) and total S6K1 in homogenates of mice placenta from control and fADN group at E18.5. C, summary of Western blot data from control (C, n= 9) and fADN (A, n= 7) infused mice. After normalization to β-actin, the mean density of control samples was assigned an arbitrary value of 1. Individual C and A density values were expressed relative to this mean. Values are given as means + SEM. *P < 0.05 versus control; Student's unpaired t test.

Figure 8. Relative placental S6 ribosomal protein phosphorylation in control (C) and fADN (A) infused mice.

A and B, representative Western blots of phosphorylated S6 ribosomal protein (Ser-235/236) (A) and total S6 ribosomal protein (B) in homogenates of mice placenta from control and fADN group at E18.5. C, summary of Western blot data from control (C, n= 9) and fADN (A, n= 7) infused mice. After normalization to β-actin, the mean density of control samples was assigned an arbitrary value of 1. Individual C and A density values were expressed relative to this mean. Values are given as means + SEM. *P < 0.05 versus control; Student's unpaired t test.

Figure 7. Relative placental 4E-BP1 phosphorylation in control and fADN infused mice.

A and B, representative Western blots of 4E-BP1 (Thr-37/46) (A), phosphorylated (Thr-70) and total 4E-BP1 (B) in homogenates of mice placenta from control and fADN group at E18.5. Equal loading was performed. C, summary of Western blot data from control (C, n= 9) and fADN (A, n= 7) infused mice. After normalization to β-actin, the mean density of control samples was assigned an arbitrary value of 1. Individual C and A density values were expressed relative to this mean. Values are given as means + SEM. *P < 0.05 versus control; Student's unpaired t test.

Maternal fADN infusion does not alter placental LKB or AMPK phosphorylation

AMP-activated protein kinase (AMPK) has been shown to be activated by adiponectin in other tissues (Kadowaki & Yamauchi, 2005) and LKB plays a key role as the upstream regulator of AMPK activity (Hardie, 2007). To examine whether maternal fADN infusion activates the placental LKB1/AMPK signalling pathway the levels of phosphorylation of LKB at Ser-428 and of AMPK at Thr-172 and total AMPK were assessed by Western blot. LKB phosphorylation was not significantly different in fADN and control groups (Supplementary Fig. 1). Similarly, AMPK phosphorylation and the expression of total AMPK were unaltered by infusion of fADN (Fig. 9A and B), which is consistent with a lack of effect of fADN on AMPK phosphorylation in cultured primary human trophoblast cells (Jones et al. 2010).

Figure 9. Relative placental AMPK phosphorylation in control and fADN infused mice.

A, representative Western blots of phosphorylated AMPK (Thr-172) and total AMPK in homogenates of mice placenta from control and fADN group at E18.5. B, the expression levels of phosphorylated AMPK (Thr-172) and total AMPK from control (C, n= 9) and fADN (A, n= 7) infused group are shown in the bar graph. After normalization to β-actin, the mean density of control samples was assigned an arbitrary value of 1. Individual C and A density values were expressed relative to this mean. Values are given as means + SEM. *P < 0.05 versus control; Student's unpaired t test.

Maternal fADN infusion stimulates placental PPARα signalling

We have previously shown that fADN activates PPARα in cultured primary syncytiotrophoblast cells (Jones et al. 2010). We therefore tested the hypothesis that fADN phosphorylates placental PPARαin vivo. Phosphorylation of PPARα at Ser-21 was increased by 3-fold (P= 0.01) in fADN placentas as compared to control. In contrast, phosphorylation of PPARα at Ser-12 and total PPARα expression were comparable between control and fADN mice (Fig. 10A and B).

Figure 10. Relative placental PPARα phosphorylation in control and fADN infused mice.

A, representative Western blots of phosphorylated (Ser-12 and Ser-21) and total PPARα in homogenates of mice placenta from control and fADN groups at E18.5. B, summary of data from control (C, n= 9) and fADN (A, n= 7) infused animals. After normalization to β-actin, the mean density of control samples was assigned an arbitrary value of 1. Individual C and A density values were expressed relative to this mean. Values are given as means + SEM. *P < 0.05 versus control; Student's unpaired t test.

AdipoR1 and AdipoR2 expression in TPM

To demonstrate that adiponectin receptors are present in TPM we determined the expression of AdipoR1 and AdipoR2 in this membrane fraction using Western blot. As shown in Supplementary Fig. 2A and B, AdipoR2 protein, but not AdipoR1, was expressed in TPM.

Discussion

The factors regulating fetal growth remain to be fully established and, in particular, the mechanisms linking maternal BMI to the growth of the fetus are not well understood. The novel findings in this report are that maternal fADN modulates placental insulin/IGF-1 signalling pathways, down-regulates placental amino acid transport and reduces fetal growth in vivo. Thus, in contrast to its action in skeletal muscle and liver, fADN decreased insulin sensitivity in the placenta. We have identified a novel mechanism by which the endocrine functions of maternal adipose tissue influence fetal growth and we speculate that the strong influence of maternal body mass index on fetal growth is mediated, in part, by changes in maternal adiponectin levels.

Using Mg²⁺ precipitation and alkaline phosphatase as a specific marker, Kusinski et al. (2010) developed a protocol for isolation of trophoblast plasma membranes derived from syncytiotrophoblast layer II of mouse placenta. These authors provided evidence to support the idea that the maternal-facing plasma membrane of syncytiotrophoblast layer II in the rodent placenta is functionally analogous to the microvillous plasma membrane (MVM) in the human placenta (Kusinski et al. 2010). In support of this concept, we found that that all three SNAT isoforms and CD98, LAT1 and LAT2 proteins, which are associated with System L activity, are expressed in the trophoblast plasma membranes derived from syncytiotrophoblast layer II of mouse placenta, in agreement with the presence of these isoforms in human syncytiotrophoblast MVM (Cleal & Lewis, 2008; Desforges et al. 2009).

ADN is not produced by the placenta (Mazaki-Tovi et al. 2007; Pinar et al. 2008), and trophoblast function in vivo is therefore likely to be regulated by circulating ADN produced in adipose tissue. ADN exerts its cellular effects by binding to two plasma membrane receptors, AdipoR1 and AdipoR2 (Kadowaki & Yamauchi, 2005). AdipoR2 mRNA is expressed in the human placenta (Caminos et al. 2005), and AdipoR2 protein has been localized to the syncytiotrophoblast (Caminos et al. 2005). Furthermore, it was recently reported that both AdipoR1 and R2 proteins are expressed in cytotrophoblast cells freshly isolated from human placenta as well as in human primary trophoblast cells in culture (McDonald & Wolfe, 2009). We show that AdipoR2 is expressed in trophoblast plasma membranes isolated from mouse placenta. Because these plasma membranes are believed to be derived from the maternal facing plasma membrane of syncytiotrophoblast layer II, which represents the first major membrane barrier to maternofetal transport in the mouse placenta, (Kusinski et al. 2010), adiponectin receptors expressed in this plasma membrane are likely to be accessible to ligand circulating in the maternal circulation. However, the role of ADN in regulating placental function is only beginning to be appreciated. Recently it was reported that adiponectin attenuates mRNA expression and/or production of placental lactogen, chorion gonadotropin and progesterone in trophoblast cells (McDonald & Wolfe, 2009). These data, together with our previous in vitro studies (Jones et al. 2010) and the findings in the current report, are consistent with an inhibitory effect of adiponectin on placental endocrine and transport functions.

ADN stimulates glucose transport in skeletal muscle, mediated by increased GLUT4 translocation (Ceddia et al. 2005). In addition, adiponectin has been shown to increase glucose uptake into 3T3-L1 preadipocytes and C(2)C(12) myoblasts by activation of GLUT1 (Abbud et al. 2000). Recently it was reported that infusion of adiponectin in pregnant rats caused a decrease in mRNA expression of GLUT3 in the placenta (Caminos et al. 2005), suggesting that the effect of ADN on glucose transport may be fundamentally different in the placenta. It is conceivable that ADN may alter placental glucose transport by affecting insulin signalling since insulin or IGF-I has been shown to stimulate placental glucose uptake in some (Kniss et al. 1994; Gordon et al. 1995; Ericsson et al. 2005), albeit not in all (Challier et al. 1986), studies. We demonstrate for the first time that maternal infusion of fADN down-regulates the activity and expression of System A and System L amino acid transporter isoforms in the placental barrier. Since the System L transporter constitutes the primary route by which neutral essential amino acids, such as leucine, valine and threonine, are transported across the placental barrier, maternal full-length adiponectin may lead to a decreased fetal growth by limiting fetal supply of essential amino acids. These in vivo findings in the rodent are in contrast to the lack of effect of fADN on System L amino acid transport activity in cultured human primary trophoblast cells (Jones et al. 2010). These discrepancies could be due to species differences, or, alternatively, it could be that secondary changes to the altered levels of maternal fADN cause the in vivo inhibition of placental System L activity.

The molecular mechanisms by which fADN inhibits placental amino acid transport in vitro (Jones et al. 2010) and in vivo (current study) remain to be identified, but it is likely that interaction with the insulin signalling pathway, well established to stimulate placental System A amino acid transport (Karl et al. 1992; Jansson et al. 2003), is involved. In cultured primary human trophoblast cells, fADN reversed the insulin stimulated phosphorylation of Akt at Thr-308 and of IRS-1 at Tyr-612 (Jones et al. 2010). In agreement with these in vitro findings, fADN decreased phosphorylation of IRS-1 at Tyr-608 (homologous to Tyr-612 in the human) as well as phosphorylation of Akt, supporting the hypothesis that fADN inhibits trophoblast insulin signalling. This effect appears not be mediated by decreased maternal insulin levels in response to infusion of fADN because maternal insulin levels were not altered in the fADN infused group. Indeed, fADN inhibited placental insulin signalling both in vitro (Jones et al. 2010) and in vivo, suggesting that this is a direct effect of fADN on trophoblast cells. We observed a marked inhibition of placental mTOR signalling in response to maternal administration of fADN, which could be a consequence of reduced insulin signalling because Akt/TSC2/Rheb signalling is upstream of mTOR. We have previously shown that mTOR signalling is a positive regulator of trophoblast System A and L amino acid transporters (Roos et al. 2007, 2009a) and it is possible that the marked inhibition of mTOR signalling in response to maternal infusion of fADN in the current study is responsible for the down-regulation of placental amino acid transport in vivo.

The protein expression of all three SNAT isoforms and LAT1 and 2 was decreased in isolated trophoblast plasma membranes, whereas CD98 expression was unaffected by fADN, suggesting that it is unlikely that fADN modulates trophoblast plasma membrane transport mediated by a general effect on membrane proteins. We speculate that changes in System A and System L activity are a result of both transcriptional and post-translational regulation. This hypothesis is based on the observation in the current study that fADN inhibits insulin and mTOR signalling and our previous reports demonstrating that fADN inhibits insulin stimulated increase in SNAT2 expression (Jones et al. 2010) and that mTOR signalling stimulates System A and L transport activity in trophoblast cells by modulating the cell surface abundance of SNAT2 and LAT1 (Rosario et al. 2011b).

The intracellular signalling pathways mediating the cellular effects of ADN are not fully established in any tissue. It is thought that globular adiponectin preferentially binds to AdipoR1, whereas fADN is believed to primarily interact with AdipoR2 (Kadowaki & Yamauchi, 2005). In skeletal muscle both receptors activate PPARα and AMPK (Kadowaki & Yamauchi, 2005). In contrast, Adipo R1 and R2 have distinct signalling pathways in the liver in that AdipoR1 signalling is believed to be mediated via AMPK and AdipoR2 activates PPARα. Emerging evidence suggests that PPARα may inhibit insulin signalling in muscle (Finck et al. 2005; Park et al. 2005a). Importantly, we show in the present study that fADN administration to pregnant mice activates PPARα, which is consistent with our previous report that fADN stimulates PPARα in cultured primary human trophoblast cells (Jones et al. 2010). These data, together with the observation that fADN does not activate AMPK in trophoblast cells and the demonstration of expression of AdipoR2 receptors in TPM, suggest that fADN may preferentially bind to AdipoR2 on trophoblast cells, resulting in an activation of PPARα, which inhibits insulin signalling (Supplementary Fig. 3).

Our data suggest that maternal adipose tissue regulates placental function and fetal growth by the secretion of adiponectin, which may explain the strong correlation between maternal BMI and birth weight. Furthermore, these findings may have important clinical implications for pregnancies complicated by obesity and/or gestational diabetes mellitus (GDM). Women who are obese or have GDM are more likely to give birth to babies with increased birth weight (Ehrenberg et al. 2004) and typically have low circulating levels of ADN (Ranheim et al. 2004; Ategbo et al. 2006; Jansson et al. 2008). In a recent study of pregnant women with pre-pregnancy BMI ranging from 17 to 44, serum ADN in the third trimester was approximately 7-fold higher in the most lean women as compared to the women with the highest BMI (Jansson et al. 2008). Thus, the 4-fold increase in serum fADN concentrations in our mouse model is physiologically and clinically relevant. The data in the current study and in the previous report (Jones et al. 2010) suggest that insulin and fADN interact in regulating placental transport function. In lean pregnant women the combination of relatively high ADN and low insulin levels would tend to decrease the activity of placental amino acid transporters. In contrast, the combination of low ADN and high insulin in obese women and in pregnancies complicated by GDM would tend to stimulate placental nutrient transport. An up-regulation of placental amino acid transport would contribute to fetal overgrowth in pregnancies of women with obesity or GDM (Jansson et al. 2002). Thus, our data are compatible with the possibility that the inverse correlation between maternal ADN and birth weight in women with obesity or GDM (Ategbo et al. 2006; Jansson et al. 2008; Segal et al. 2008) may represent a cause and effect relationship, because low adiponectin levels would release the inhibition of this adipokine on placental insulin signalling and amino acid transport, thereby promoting increased fetal growth.

Glossary

Abbreviations

AMPK

AMP-activated protein kinase;

ADN

adiponectin

fADN

full-length adiponectin

GDM

gestational diabetes mellitus

IRS

insulin receptor substrate

PPARα

peroxisome proliferator-activated receptor α

TPM

trophoblast plasma membrane

Author contributions

Conception and design of the experiments: F.J.R., M.A.S., T.L.P. and T.J. Collection, analysis and interpretation of data: F.J.R., M.A.S., T.L.P. and T.J. Draftingthe article or revising it critically for important intellectual content: F.J.R., M.A.S., J.J., Y.K., T.L.P. and T.J. All authors approved the final version of the manuscript.

Supplementary material

Supplementary Figure 1

Supplementary Figure 2

Supplementary Figure 3