Placental glucose uptake in a nonhuman primate model of western-style diet consumption and chronic hyperandrogenemia exposure.

Abstract

We reported that consumption of a western-style diet (WSD) with and without hyperandrogenemia perturbed placental perfusion and altered levels of glucose transporter proteins in rhesus macaques. Based on that result, we hypothesized that placental glucose uptake would be dysregulated in this model. In this study, female rhesus macaques were assigned at puberty to one of four groups: subcutaneous cholesterol implants + standard chow diet (controls, C); testosterone implants + chow (T); cholesterol implants + a high-fat, WSD; and T+WSD. After ~6 years of treatment, animals were mated, and pregnancies were delivered by C-section at gestational day (G) 130 (term is G168). Placental villous explants were immediately prepared for radiolabeled glucose assay. Linear glucose uptake was observed between 0 and 30 seconds. At 20 seconds, glucose uptake in placental villous explants did not differ across the four treatment groups with values as follows: C: 25.5 ± 6.33, T: 22.9 ± 0.404, WSD: 26.9.0 ± 3.71, T+WSD: 33.0 ± 3.12 (mean ± SD expressed in pmol/mg). Unlike our prior experiment, glucose transporter expression was reduced in WSD placentas, and our in vitro functional assay did not demonstrate a difference in glucose uptake across the transporting epithelium of the placenta. Notably, maternal blood glucose levels were significantly elevated in animals chronically fed a WSD. This disparity may indicate differences in glucose utilization and metabolism by the placenta itself, as glucose transporter expression and circulating fetal glucose concentrations were comparable across all four groups in this pregnancy cohort.

Introduction

Polycystic ovarian syndrome (PCOS) is a prevalent endocrine disorder in reproductive age women that is characterized by excessive androgen production [1] and metabolic disorders that lead to subfertility [2]. The underlying etiology is multifactorial [3], and many PCOS women display overweight/obesity that compounds the adverse metabolic phenotype [4]. Our collaborative research group has reported dysregulation of reproductive function using a translational animal model of elevated androgens, with and without consuming an obesogenic western-style diet (WSD). This model was established in young female rhesus macaques (2.5 years of age; prior to menarche), with four experimental cohorts that were fed either a control standard chow diet or a high fat, high sugar content diet (WSD) with or without subcutaneous testosterone (T) implants. A detailed description of the phenotypic characteristics of the model has been previously published [5], where it was demonstrated that greater metabolic impairments occur in the WSD with T group than either treatment alone, thus providing a good model of the human scenario. Significant metabolic and endocrine disruption were observed and studies from this model have reported impaired ovarian and uterine structure and function [6], reduced oocyte quality [7], and compromised fertility [8].

After approximately four years of WSD with or without T treatment, these animals underwent a pregnancy trial in which delayed and reduced conception rates were reported, along with a reduced placental blood volume in early gestation [8] and evidence of placental compromise in samples delivered in the mid-third trimester [9]. In that initial assessment of the effects of androgen excess in combination with a WSD, we found evidence to suggest altered placental nutrient regulation that required further investigation. Specifically, the perturbed expression of placental glucose transporters (GLUT) [9] has potential ramifications for fetal growth, as glucose is the primary energy substrate used by the fetus and the placenta throughout pregnancy [10, 11]. In humans, glucose transport is primarily regulated by the GLUT-1, GLUT-4, and GLUT-9 receptors [11], and we have made similar observations in the macaque [9]. Our prior data demonstrated altered localization of the GLUT receptors on the syncytiotrophoblast, which suggests differential glucose handling across the maternal-fetal exchange barrier in animals exposed to a WSD with and without T [9]. Additionally, it has been postulated that glucose dysregulation is associated with endothelial dysfunction [12], which may, in part, explain the reported changes in placental vascular function measured by contrast-enhanced ultrasound and stereological assessments of the maternal and fetal vascular compartments of the placenta [9].

We hypothesized that a maternal WSD would diminish placental glucose transport capacity and that hyperandrogenemia would exacerbate this detrimental effect. To test this, using the same cohort of animals after an additional two years of exposure to treatment conditions, we performed a second pregnancy trial. In this trial, the aim was to assess in vivo and in vitro placental function. We used contrast-enhanced ultrasound to examine placental perfusion characteristics in the first and third trimesters, and collected placental tissue post-delivery to perform an in vitro functional assay of glucose uptake in placental villous explants.

Methods

Animals

All protocols were approved by the Institutional Animal Care and Use Committee (IACUC) of the Oregon National Primate Research Center (ONPRC), and guidelines for humane animal care were followed. Female rhesus macaques (Macaca mulatta) were randomly assigned at puberty (~2.5 years of age) to one of four treatment groups: 1) C: control standard chow diet with a subcutaneous cholesterol (vehicle) implant (n=7), 2) T: control diet with a subcutaneous T + cholesterol implant (n=3), 3) WSD: cholesterol implants with a high-fat, calorie-dense diet (n=3), and 4) T+WSD: T + cholesterol implants in combination with a WSD (n=3). The standard diet contained 15% calories from fat, 27% protein and 58% carbohydrate (Fiber-balanced monkey diet #5052, LabDiet, St. Louis, MO). The WSD contained 36% calories from fat, 18% protein and 46% carbohydrate (TAD primate diet #5L0P, LabDiet, St. Louis, MO). The two diets were well matched for vitamin and micronutrient content. T-releasing implants were prepared by packing medical grade Silastic tubing (Dow Corning, Midland MI) with a mixture of cholesterol and T at a ratio of 9:1 to generate sustained elevated T levels as previously described in detail [5]. Animals were indoor, pair housed for the duration of the study. After approximately six years of treatment, pregnancies were generated by time-mated breeding. In brief, hormone levels were monitored by daily blood sampling to detect the mid-cycle peak in estradiol, which immediately precedes the surge of luteinizing hormone responsible for ovulation. Females were paired with a male several days before the estradiol surge then separated the day after peak levels were observed. Day 1 of gestation was considered to be 48 hours post-estradiol surge. This timed-mating scheme allows the gestational age to be determined within 24 to 48 hours of conception.

Contrast-Enhanced Ultrasound

At gestational days 30 (G30) and 130 (G130), animals underwent a contrast-enhanced ultrasound (CEUS) study. Following an overnight fast, animals were initially sedated with 10mg/kg ketamine IM before intubation and maintenance under anesthesia with inhaled 1-2% isoflurane gas. Physiological monitoring was carried out throughout each exam. CEUS was performed using a multiphase amplitude-modulation and phase-inversion algorithm on a Sequoia system (Siemens Medical Systems, Mountain View, CA) equipped with a 15L8 transducer at a transmit frequency of 7MHz with a 0.18 mechanical index (MI) and a 55dB dynamic range as previously described [13]. Lipid-shelled octofluoropropane microbubble contrast reagent (Definity®, Lantheus Medical Imaging, Billerica, MA) was prepared in 0.9% saline at a final concentration of 5%. Intravenous infusion via a cephalic catheter was performed at an infusion rate of 60ml/hr to visualize uteroplacental blood flow. The acoustic beam was centered over individually identified maternal spiral artery sources, and the microbubbles within the path of the beam were destroyed by a brief (2 seconds) increase in MI to 1.9. Microbubble re-entry in the spiral artery and the intervillous space was recorded until the area of interest reached video intensity (VI) signal saturation (VImax). Three replicates of all digital video recordings were obtained and mean data later calculated for each placental cotyledon.

Data analysis

Digital imaging data were analyzed using a custom-designed CEUS analysis program (Narnar©). For placental assessment, regions of interest (ROI) were drawn over the area of each cotyledon. The data was fit to the function y=A(1-e^−βt), where y is the VI at the pulsing interval t, A is the VI plateau and β is the flux rate constant. The β-value determines the rate constant of the post-destructive VI recovery and reflects the microvascular blood flux rate.

Placental Tissue Collection

All pregnancies were delivered by cesarean section in the mid third trimester at G130, where full term in the rhesus macaque is G168. Following delivery, fetal and placental weights were measured and recorded, and placental tissue was collected. Full-thickness sections (maternal decidua to chorionic plate) were collected in cassettes and fixed in 10% zinc formalin for paraffin embedding. Villous tissue was dissected and either flash-frozen in liquid nitrogen for later analysis or utilized in explant experiments as described below.

¹⁴C-Glucose Uptake

Uptake of ¹⁴C-Glucose into fresh villous explants was measured in glucose-free Tyrodes buffer over a 120-second time course. In brief, ~2mm³ pieces of villous tissue were tied to specially designed hooks that facilitate consistency with replicates and quick transfer between incubation solutions. After an initial pre-incubation in glucose-containing media at 37°C, fragments were washed in glucose-free Tyrodes before incubation with radiolabeled ¹⁴C-Glucose for a time course of 10, 20, 30, 60, and 120 seconds, washed and lysed overnight in H₂O to release ¹⁴C-Glucose taken up by the tissue, followed by overnight incubation in 0.3M NaOH to lyse the tissue for protein determination. Glucose uptake in each explant was determined as the fraction of ¹⁴C-Glucose uptake expressed as pmol/mg protein.

Placental protein Expression

Protein was extracted from pooled villous tissue sampled from four different cotyledons of each placenta following homogenization in lysis buffer containing 20mM Tris (pH 7.5), 1mM EDTA, 1mM EGTA, 20mM sodium flouride, 0.15M sodium chloride, 0.5% Nonidet P-40, 0.5% Triton X-100, 200μM sodium orthovanadate, 2μM leupeptin, 5.8μM pepstatin, 200μM 4-(2-aminoethyl) benezenesulfonyl fluoride hydrochloride (AEBSF) and 5μM N-tosyl-L-lysine chloromethyl ketone (TLCK). Homogenate was centrifuged for 5 minutes at 20,000xg to remove cell debris from the placental samples. The pellet was discarded and supernatant stored at −80°C until use. Protein quantification was determined by BCA assay (BioRad) to standardized protein loading concentration. Protein samples were heat-denatured at 95°C for 5 minutes in Laemmli buffer (125M Tris, 4.1% SDS, 40mM urea, 20% glycerol, 0.002% bromophenol blue) and separated on 4-20% Tris-glycine pre-cast gels (Invitrogen) using a mini-gel electrophoresis system (Bio-Rad Laboratories) at 35mA per gel for 80 minutes. Protein was transferred to nitrocellulose membranes using the iBlot2 dry blotting system (Thermo Fisher Scientific). Equal protein loading was assessed using Ponceau S staining (Sigma-Aldrich). Following the transfer, membranes were blocked for 1 hour with Tris buffered saline/0.1% Tween 20 (polyoxyethylene sorbitan monolaurate, TBST) containing 5% bovine serum albumin or milk protein before overnight incubation with primary antibodies at 4°C. The following primary antibodies were used: anti-glucose transporter 1 (GLUT-1, cat# ab128033, Abcam), anti-glucose transporter 4 (GLUT-4, cat# ab33780, Abcam), anti-glucose transporter 9 (GLUT-9, cat# PA5-22971, Invitrogen), and anti-actin (cat# ab8224, Abcam).

Membranes were washed in TBST before incubation with the appropriate secondary antibody (anti-rabbit antibody, cat# ab97051, Abcam; anti-mouse antibody, cat# sc-2314, Santa Cruz) for 1 hour at room temperature. Antibody binding was detected using chemiluminescence (Pierce ECL Western Blotting Substrate, cat# 32209, Thermo Fisher Scientific), and images were acquired on a Fluorchem M digital darkroom system (ProteinSimple, San Jose, CA). Band intensity of each protein of interest was normalized to actin and quantified using scanning densitometry in ImageJ software.

Placental Immunohistochemistry

Formalin fixed paraffin-embedded tissue sections were deparaffinized with xylene, washed with ethanol and rehydrated. Slides were boiled for 10 minutes in a citrate buffer (pH 6.0) for antigen retrieval and then incubated with 2% normal donkey serum to block nonspecific antibody binding. Slides were incubated with primary antibody (as detailed for western blotting) overnight at 4°C. Incubation with biotinylated secondary antibody (Cat# 711-065-152, Jackson Immunoresearch) was performed for 1 hour at room temperature before washing and application of peroxidase substrate (SK-4100, Vector Laboratories). Nuclei were counterstained with Modified Mayer’s hematoxylin (Cat# 72804, Thermo Scientific).

Statistical Analysis

Numeric data were analyzed using statistical analysis software (GraphPad Prism, La Jolla, CA). Two-way Analysis of Variance (ANOVA) was used to compare continuous data among treatment groups, and a p-value of <0.05 used to indicate statistical significance for both treatment effects and treatment interactions between groups. For statistically significant complex interactions, Tukey’s multiple comparisons test was used to compare individual groups.

Results

Animal cohort demographics

Data on maternal and fetal body weight, placental weight, fetal sex, and maternal and fetal glucose levels at G130 are presented in Table 1. Despite variability within each of the four cohorts, there was a strong trend towards increased weight in animals chronically fed a WSD. Fetal body weight and placental weight were unchanged across the four treatment groups resulting in no significant difference in the calculated measure of placental efficiency, the fetal:placental weight ratio. Maternal fasting glucose levels at G130 were significantly elevated in animals consuming a WSD (p<0.01 by Two-way ANOVA). Fetal glucose measured in umbilical vein blood samples immediately prior to cord clamping and delivery of the fetus, demonstrates variability within the small sample cohort. No overall difference across the four groups was seen despite a slight increase in WSD alone (Table 1).

Table 1:

Animal cohort demographics.

Parameter╲Animal group (n)	C (7)	T (3)	WSD (3)	T+WSD (3)	P value
Maternal Weight (kg)	8.7 ± 1.3	9.1 ± 2.6	10.9 ± 1.1	10.6 ± 2.7	0.0857#
Fetal weight (g)	349.3 ± 36.0	330.1 ± 60.6	372.2 ± 72.7	385.8 ± 9.5	0.1393
Placental weight (g)	92.5 ± 14.7	103.5 ± 12.8	114.8 ± 24.6	97.1 ± 15.5	0.3858
Fetal:Placental ratio	3.8 ± 0.6	3.2 ± 0.5	3.3 ± 0.2	4.1 ± 0.8	0.6425
Maternal glucose (mg/dL)	55.7 ± 5.7	47.3 ± 9.1	78.0 ± 22.7	64.7 ± 11.0	0.0080a
Fetal glucose (mg/dL)	39.1 ± 7.8	35.0 ± 11.5	51.7 ± 14.0	39.0 ± 5.3	0.1288
Fetal sex (M:F)	1:6	2:1	2:1	2:1	-

Data are presented as mean ± SD.

^#Diet effect

^aStatistically significant diet effect

Contrast-enhanced ultrasound analysis of placental blood flow

CEUS with Definity® microbubbles permitted visualization of maternal spiral arteries supplying individual placental cotyledons of the primate placenta. Figure 1A shows a representative image from the early first trimester (G30). Figure 1C shows a representative image from the mid-third trimester (G130). Microvascular flux rate was calculated from individual sources (2-12 per animal), and mean values for each animal per cohort are shown in Figure 1B and D for G30 and G130 time points respectively. We demonstrate variability within the data sets but overall, no change at G30 (Figure 1B) and a trend towards increased flux rate at G130 in WSD-fed animals (Figure 1D).

Figure 1: Contrast-enhanced ultrasound.

Representative image of one maternal spiral artery (arrow) source perfusing a cotyledon (white outline) at G30 (A) and G130 (C). Flux rate in control and WSD-fed animals with (black bars) and without (open bars) T treatment at G30 (B) and G130 (D). Data are mean±SD from C (n=7), T (n=3), WSD (n=3) and T+WSD (n=3) animals.

Placental glucose uptake

A radiolabeled glucose assay was used to assess glucose uptake over a 120-second time course. We observed linear glucose uptake between 0 and 30 seconds (Figure 2A). At 20 seconds, there was no difference in glucose uptake in placental villous explants across the four treatment groups (Figure 2B).

Figure 2: Placental glucose uptake.

(A) Time course of glucose uptake in placental villous explants in C (open circles), T (closed circles), WSD (open triangles), and T+WSD (closed triangles) animals over a 120-second time course. (B) Glucose uptake at 20 seconds, corresponds to the linear portion of the uptake curve in control and WSD-fed animals with (black bars) and without (open bars) T treatment. Data are mean+SD from C (n=7), T (n=3), WSD (n=3) and T+WSD (n=3) animals.

Glucose transporter expression and localization

Protein expression of the primary GLUT transporters, GLUT-1, GLUT-4, and GLUT-9 was assessed by western blot. There was a slight, non-significant effect of T on GLUT-1 expression (Figure 3A) and a modest decrease in GLUT-4 in WSD fed animals (Figure 3B) but overall, no significant differences in GLUT protein expression across the four treatment groups (Figure 3). Supplemental figure 1 shows the full western blot images.

Figure 3: Placental glucose transporter expression.

Representative western blot images (upper panels) and semi-quantitative densitometry of (A) GLUT-1, (B) GLUT-4, and (C) GLUT-9 protein expression in placental homogenate from control and WSD-fed animals with (black bars) and without (open bars) T treatment (lower graphs). Densitometry data are mean+SD from C (n=7), T (n=3), WSD (n=3) and T+WSD (n=3) animals.

GLUT-1 was localized on the syncytiotrophoblast basal membrane with some positive staining in the fetal capillaries (Figure 4 top row). The GLUT-4 isoform was expressed throughout the syncytiotrophoblast and not detected in the fetal vasculature (Figure 4 middle row). GLUT-9 was predominantly localized to the basal membrane of the syncytiotrophoblast and to endothelial cells of the fetal capillaries, which was especially observed in the T treated groups (Figure 4 lower row).

Figure 4: Placental glucose transporter localization.

Representative immunolocalization images of GLUT-1 (upper panels), GLUT-4 (middle panels) and GLUT-9 (lower panels) in placental sections from C, T, WSD and T+WSD-fed animals. Negative control images were acquired with no primary antibody. Scale bar is 20μm. SYN; syncytiotrophoblast, BM; basal membrane, MVM; microvillus membrane, FC; fetal capillary.

Discussion

Here we report the results of a follow-up study to our initial investigation of WSD and hyperandrogenemia effects on pregnancy outcomes and, more specifically, placental function in a nonhuman primate model. In the first pregnancy trial, consumption of a WSD resulted in altered vascular impedance and placental glucose transporter expression and localization. The effect of testosterone treatment was predominantly on fetal vascular development in the placenta with a 40% reduction in capillary volume [9], which could suggest an associated reduction in nutrient transfer availability. Based on these earlier findings of perturbed glucose transporter expression and transport capacity, we anticipated that continued treatment would exacerbate these detrimental outcomes. However, our in vitro functional assay did not demonstrate a difference in glucose uptake in placental villous explants from this model. Similarly, despite our prior observations of reduced GLUT protein levels in placentas from animals fed a WSD, in this study, similar transporter expression between groups was found. There may be several plausible explanations for the discrepancy in our outcomes, some of which prove to be limitations of the current study.

With the continuation of this model, in keeping with the spectrum of hyperandrogenemia effects that are reported in women, we have observed a range of maternal phenotypes within our animal groups that manifest in response to prolonged and chronic exposure to treatment. Metabolic profiling after five years of treatment, prior to this second pregnancy trial, demonstrated heterogeneity within individual animals from the same treatment cohort, in addition to treatment-specific effects [14]. Of note, control animals are fed ad libitum and therefore over time these animals displayed a steady increase in body weight, yet fat mass is still significantly greater in WSD-fed animals. There is no evidence of hyperglycemia with fasting glucose levels being comparable across cohorts, but increased fasting insulin levels are found in T+WSD animals suggesting elevated insulin resistance. Comprehensive assessment of reproductive hormones and monthly cycling characteristics in these animals demonstrates that none of the treatments result in amenorrhea [14]. However, there is evidence of subfertility, and in the first trial, pregnancy rates were lower in both the WSD and T+WSD cohorts [8]. With this in mind, it is essential to note that in our current study, pregnancy was not achieved in animals that exhibited the most severe impacts of their treatment group. This is reflected in the number of animals available in each treatment cohort, which was seven in our control group, and only three pregnant animals in the WSD alone and the two T treatment groups. In addition, some animals have been removed from the overall study cohort due to secondary complications of chronic treatment, predominantly severe endometriosis which prevented any possible further mating. Smaller sample sizes present a challenge when variability in outcomes occurs, and they prohibit in-depth analysis that accounts for confounding factors such as fetal sex, which may affect functional outcomes as demonstrated in other aspects of placental and fetal development [15, 16]. The imbalance between fetal sex ratios must be acknowledged as a limitation of the study and a potential complication for data interpretation. However, in combination between our two pregnancy trials, the ratio of male to female offspring was similar across treatment groups and no correlations were found between fetal sex and our measured parameters (data not shown).

We report no difference in fetal body weight across the four animal cohorts. This outcome is not atypical of our published reports from several different macaque models [17-19] where delivery in the mid-third trimester may not capture the final significant fetal growth during the last month of gestation. Similarly, measurement of glucose uptake at this point in gestation may not reveal nutritional requirement differences associated with the late accelerated growth in the final month of gestation. However, with a lack of fetal body weight differences in combination with the finding that glucose uptake is unaltered by either WSD or T treatment, it seems plausible to suggest that the primate placenta has sufficient capacity to compensate for an adverse in utero environment to maintain regular nutrient transport. Indeed, this concept is supported by previous animal model studies. Of note, graded reductions of uteroplacental blood flow in a pregnant sheep model demonstrate the ability to maintain fetal hemodynamics despite a 50% reduction in placental perfusion [20]. In addition, our own work in the pregnant NHP has demonstrated placental plasticity in response to surgical ligation of the bridging vessels that supply the fetal side vasculature to the structurally independent secondary placental lobe (effectively reducing the placenta by 40-50%). When this acute insult occurs in early gestation, there is compensatory growth of the primary placental lobe and maintenance of fetal growth at the time of delivery in the mid third trimester [21].

The complex regulation of glucose utilization during pregnancy is intended to keep metabolism relatively constant [22] and to maintain the maternal delivery to fetal supply balance regulated by the placenta. Maternal-fetal glucose exchange is dependent on several dynamic factors including the maternal blood supply to the placenta, the concentration gradient between the two circulatory systems, transporter expression and glucose utilization as an energy substrate. Our assessment of GLUT expression and localization did not indicate significant effects of either diet or T treatment, which suggests relatively stable transport characteristics across the syncytiotrophoblast, and uptake in the fetal circulation predominantly orchestrated via GLUT-9, which was discretely localized to the fetal endothelial cells. There appears to be good modulation of glucose transfer despite the higher maternal circulating glucose concentration in WSD-fed animals, as fetal glucose concentrations remained consistent across the four cohorts. It is also conceivable that there may be differences in glucose utilization by the placenta itself in response to an adverse uterine milieu. Further investigation of placental metabolism in real-time using advanced in vivo imaging modalities will be required to address these differences.

Using contrast-enhanced ultrasound, we can generate a semi-quantitative measure of vascular impedance in the placental intervillous space. This methodology has increased sensitivity over standard Doppler ultrasound and allows assessment of maternal perfusion of the placenta [23]. Perturbations in our output function, the flux rate, suggest alterations in microvascular blood flow that may affect materno-fetal nutrient transport capabilities. In this study, contrast-enhanced ultrasound data did not fully recapitulate the findings from the first pregnancy trial conducted in this model. Variability in small animal cohorts is likely to explain this discrepancy, but it seems plausible to suggest that there may be developmental differences associated with parity. During pregnancy, the uterus undergoes extensive vascular remodeling and development. This highly orchestrated process functions to support the growth and development of the placenta and fetus in the ongoing pregnancy, and permanently alters the maternal uterine vasculature. We have made this observation in a histological assessment of hysterectomy samples from nulliparous and multiparous post-menopausal women [24] and could apply this evidence to our understanding of vascular function in our animals. More specifically, endothelial dysfunction is more prevalent in nulliparous women [12], and this may be relevant to the interpretation of our findings. An advantage of our animal model is the well-controlled experimental conditions and, in particular, the known fact that this was the second pregnancy for all of the animals included in the present data set, as well as access to comprehensive reproductive hormone profiles. Endothelial dysfunction is associated with aberrant glucose regulation, which was more apparent in our first round of pregnancies [9]. In addition, testosterone is known to induce mechanisms of vascular dysfunction in pregnancy [25]. In conclusion, we can speculate that the lack of functional evidence of alterations in glucose uptake, and the consistency in placental perfusion characteristics exhibited across our treatment groups, is due in part to the physiological adaptations of previous pregnancy, and perhaps secondarily to the less severe maternal metabolic phenotype exhibited in the animals that achieved pregnancy in this second trial as the most severely affected animals failed to conceive.