Impact of chorionic somatomammotropin RNA interference on uterine blood flow and placental glucose uptake in the absence of intrauterine growth restriction

Abstract

Chorionic somatomammotropin (CSH) is one of the most abundantly produced placental hormones, yet its exact function remains elusive. Near-term [135 days of gestational age (dGA)], CSH RNA interference (RNAi) results in two distinct phenotypes: 1) pregnancies with intrauterine growth restriction (IUGR), and 2) pregnancies with normal fetal and placental weights. Here, we report the physiological changes in CSH RNAi pregnancies without IUGR. The trophectoderm of hatched blastocysts (9 dGA) were infected with lentiviral-constructs expressing either a scrambled control (Control RNAi) or CSH-specific shRNA (CSH RNAi), prior to transfer into synchronized recipient ewes. At 126 dGA, Control RNAi (n = 6) and CSH RNAi (n = 6) pregnancies were fitted with maternal and fetal catheters. Uterine and umbilical blood flows were measured at 132 dGA and nutrient uptakes were calculated by the Fick’s principle. Control RNAi and CSH RNAi pregnancies were compared by analysis of variance, and significance was set at P ≤ 0.05. Absolute (mL/min) and relative (mL/min/kg fetus) uterine blood flows were reduced (P ≤ 0.05) in CSH RNAi pregnancies, but umbilical flows were not impacted. The uterine artery-to-vein glucose gradient (mmol/L) was significantly (P ≤ 0.05) increased. The uteroplacental glucose uptake (μmoL/min/kg placenta) was increased (P ≤ 0.05), whereas umbilical glucose uptake (μmoL/min/kg fetus) was reduced. Our results demonstrate that CSH RNAi has significant physiological ramifications, even in the absence of IUGR, and comparing CSH RNAi pregnancies exhibiting both IUGR and non-IUGR phenotypes may help determine the direct effects of CSH and its potential impact on fetal development.

INTRODUCTION

The placenta-specific hormone chorionic somatomammotropin (CSH) is one of the most abundantly produced placental hormones and belongs to the growth hormone/prolactin gene family. Syncytiotrophoblast cells in the human placenta are responsible for the production and secretion of CSH, whereas chorionic binucleate cells in the sheep serve the same purpose (1). Recently, we demonstrated, through lentiviral-mediated RNA interference (RNAi) of CSH, that CSH deficiency resulted in the development of intrauterine growth restriction (IUGR) in sheep, both at near-term [135 days of gestational age (dGA)] (2) and at the end of the first one-third of pregnancy (50 dGA) (3). These recent studies confirm an established association between reduced CSH in maternal circulation and IUGR in both human (4, 5) and sheep pregnancies (6), thus emphasizing the importance of CSH in maintaining healthy pregnancies. In spite of being discovered over 60 years ago, the exact biological role of CSH remains to be elucidated.

As reported previously (2), CSH RNAi does not always result in IUGR. Two phenotypes emerged in response to CSH RNAi: 1) pregnancies with placental and fetal growth restriction, and 2) pregnancies with fetal and placental weights similar to control RNAi pregnancies (7). Although perplexing, this phenomenon is not unique to sheep. In the human, pregnancies have been observed to possess gene mutations or deletions that either resulted in IUGR (8–10) or non-IUGR phenotypes (11, 12, 13). Recently (7), we expanded the comparison of CSH RNAi pregnancies (2) to include both phenotypes. With both near-term phenotypes, CSH RNAi resulted in increased uterine artery to uterine vein glucose gradients, reduced umbilical artery IGF1 concentrations, and increased fetal liver insulin receptor concentrations, among other commonalities between the two CSH RNAi phenotypes. Although these near-term pregnancies (2) were harvested under fasted and anesthetized conditions, some of the commonalities (7) between the two phenotypes (e.g., uterine artery to uterine vein glucose gradients) led us to hypothesize that CSH RNAi may be impacting uterine blood flow and uteroplacental glucose uptake, regardless of fetal growth restriction. As little is known about the basic in vivo CSH physiology, the objective of our current research was to determine the in vivo impacts of CSH RNAi on uteroplacental blood flow and nutrient transport in the absence of IUGR.

MATERIALS AND METHODS

All animal procedures were approved by the Colorado State University Institutional Animal Care and Use Committee (Protocol # 18–7866 A), the Institutional Biosafety Committee (18-029B), and the University of Colorado Anschutz Medical Campus Institutional Animal Care and Use Committee (Protocol #00714).

Lentiviral Generation

Lentiviral generation of hLL3.7 tg6 (target 6; CSH RNAi) and hLL3.7 NTS (scramble control/nontargeting sequence; control RNAi) was described previously (2, 3). Briefly, both the NTS and tg6 sequences (2, 3) were cloned into the LL3.7 vector in accordance with the procedures described extensively (2). The shRNA sequences for the hLL3.7 NTS and hLL3.7 tg6 constructs are summarized in Supplemental Table S1 (all Supplemental material is available at https://doi.org/10.6084/m9.figshare.12711776.v1). All virus generation and titering were completed in accordance with our previously described procedures (2).

Generation of CSH RNAi Pregnancies

Sheep were group housed in pens at the Colorado State University Animal Reproduction and Biotechnology Laboratory and provided access to hay, trace mineral, and water to meet or slightly exceed their National Research Council (14) requirements. Animal management, estrus synchronization, and embryo transfers were done as previously described extensively (2). In summary, after synchronization and subsequent breeding (Columbia–Rambouillet rams), fully expanded and hatched blastocysts were collected by flushing the uteri at 9 days postconception (dpc). Each blastocyst was infected with 100,000 transducing units of either control RNAi or CSH RNAi lentivirus. After infection for ∼5 h, each blastocyst was washed thoroughly in media and a single blastocyst was transferred surgically (2, 3) into a synchronized Dorper ewe. Dorper ewes were used as recipients to meet size requirements for the metabolic studies. Each recipient ewe was monitored daily for return to standing estrus and confirmed pregnant at 50 days of gestational age (dGA) by ultrasound (ALKOA SSD-500V, Wallingford, CT).

Surgical Instrumentation of Fetus and Ewe

At ∼115 dGA, pregnant recipient ewes were transported to the University of Colorado Anschutz Medical Campus, Perinatal Research Center (Aurora, CO). Animals had access to ad libitum alfalfa pellets (Standlee Hay, Kimberly, ID) and water. Six control RNAi (4 males and 2 females) and six CSH RNAi (2 males and 4 females) ewes underwent surgical placement of fetal and maternal catheters at 126 dGA, to determine blood flow and nutrient flux as previously described (15, 16, 17, 18, 19). Briefly, the uterus was exteriorized via midventral laparotomy and the fetus was exposed through a 6-cm incision in the uterine wall (17, 19). Fetuses were instrumented with indwelling catheters in their descending aorta (representing umbilical artery blood), femoral vein, and umbilical vein (16, 17, 19, 20). Maternal femoral artery (representing uterine artery blood), femoral vein, and uterine vein were also catheterized (17, 19, 20). All catheters were tunneled subcutaneously to the paralumbar fossa, where they were exteriorized into a pouch and maintained with 5% heparinized saline flushes while the ewe was given a minimum of 5 days for recovery from surgery (17, 19).

Blood Flow Calculations and Tissue Collection

At 132 dGA, uterine and umbilical blood flows were determined by the steady -state ³H₂O transplacental diffusion technique (21, 22). Baseline samples (draw 0) were collected from maternal femoral artery (A), uterine vein (V), umbilical vein (γ), and fetal descending aorta (α) simultaneously. Then, a 3 mL bolus of ³H₂O was infused into the fetal femoral vein, and isotopic steady state was reached by continuous infusion at 3 mL/h (15 µCi/mL) for 90 min (18). Four samples (draws 1–4) were then collected from the four catheters simultaneously at 20-min intervals for analysis of blood gas, nutrient content, ³H₂O, and CSH concentrations (18, 19).

Uterine and umbilical blood flows were calculated by the steady-state diffusion technique described previously (21). Plasma flows for either the uterine or umbilical artery were calculated by using the corresponding vessel blood flows, respectively, multiplied by one minus the fractional maternal or fetal hematocrits (19). The diffusion rate of ³H₂O was calculated as a ratio by dividing the transplacental ³H₂O diffusion rate by the ³H₂O concentrations in umbilical artery to uterine artery (19). Oxygen content was determined in accordance with calculations previously described (19). Uterine and umbilical uptakes of oxygen, glucose, lactate, and amino acids were calculated by the Fick principle (23) and reported as an average of draws 1 through 4. Uteroplacental uptakes of the aforementioned nutrients were calculated as the difference between uterine and umbilical uptakes (20). The fraction of uterine glucose uptake taken up by the placenta is calculated as uteroplacental glucose uptake divided by uterine glucose uptake. The fraction of uterine glucose uptake being taken up by the fetus is calculated by the umbilical glucose uptakes divided by the uterine glucose uptakes. Maternal glucose to oxygen quotients are defined as the maternal artery to uterine vein (A-V) glucose gradient × 6 carbons/glucose, divided by the A-V oxygen gradient (24). Maternal lactate to oxygen quotients are defined as the A-V lactate gradient × 3 carbons/lactate, divided by the A-V oxygen gradient. Maternal amino acid to oxygen quotients are defined as the A-V gradient of an individual amino acid × number of carbons that can be oxidized for that amino acid, divided by the A-V oxygen gradient. Fetal umbilical nutrient to oxygen quotients are calculated the same way as maternal, except the umbilical vein to fetal artery (γ-α) difference replaced the (A-V) difference. To calculate the overall amino acid: oxygen quotient, each individual amino acid quotient is added together. Uteroplacental lactate production is defined as the addition of uterine lactate uptake and umbilical lactate uptake.

Ewes and fetuses were euthanized at 132 dGA by a lethal dose of pentobarbital sodium (390 mg/mL, Fatal Plus, Vortech Pharmaceuticals, Dearborn, MI), and the gravid uterus (uteroplacenta) was removed and weighed (19). Placentomes were trimmed from the endometrium and after recording a total placentome weight and number, 10 placentomes were selected from each placenta and separated into cotyledonary (fetal) and caruncular (maternal) components then snap frozen in liquid nitrogen (N) and stored at −80°C (19). Excess fluid was removed from fetal membranes and weighed, then the uterus was weighed after removal of all placentomes, fluids, and fetal membranes. Fetal crown rump length (CRL) was recorded (25). Fetal weight and dissected organ weights were recorded, and tissues were snap frozen in liquid nitrogen. Ponderal index was calculated as fetal weight (g) × 100/CRL (cm)³ and fetal brain (g)/liver weight (g) ratios were calculated as indicators of asymmetric fetal growth (19, 26).

Biochemical Analysis of Blood Samples

Whole blood O₂ content, hemoglobin O₂ saturation (SO₂), partial pressure of oxygen (PO₂), partial pressure of carbon dioxide (PCO₂), pH, and hematocrit measurements were analyzed by an ABL 800 Blood Gas Analyzer (radiometer) (16, 18, 19). Plasma glucose and lactate were measured by Yellow Springs Instrument 2900 (YSI Incorporated, Yellow Springs, OH) as described previously (2, 16, 27). Plasma amino acids were measured by HPLC (28, 29). CSH concentrations in maternal and fetal plasma were analyzed by radioimmunoassay as previously described (2, 3, 6, 20, 31).

Western Blot Analysis

Protein isolation and analysis was done in accordance with methods described previously (2, 29–31). Cotyledonary or caruncular tissue (100 mg) was lysed in 500 µL of lysis buffer (0.48 M Tris, pH 7.4; 10 mM EGTA, pH 8.6; 10 mM EDTA, pH 8; 0.1 mM PMSF; 0.1 mM AEBSF; 0.0015 mM pepstatin A; 0.0014 mME-64; 0.004 mM bestatin; 0.002 mM leupeptin; and 0.00008 mM aprotinin) and sonicated on ice. For cotyledonary CSH analysis, 5 µg of protein from each sample was electrophoresed through a 4%–15% Tris-Glycine Stain-Free gel (BioRad, Hercules, CA) and transferred to a 0.45-µm pore nitrocellulose membrane. The protein in each lane was visualized after transfer to the nitrocellulose and protein loading was imaged using the ChemiDoc XRS+ chemiluminescence system (Bio-Rad) to use for normalization. To visualize CSH, the blot was incubated in a 1:25,000 dilution (in 5% nonfat dry milk/1X Tris-Bis Solution+ 1% Tween) of rabbit α- oPL-S4 (30) for 24 h at 4°C. After the membrane was washed, the membrane was transferred into a 1:100,000 dilution (in 5% non-fat dry milk/1X Tris-Bis Solution+ 1% Tween) of mouse α-rabbit IgG conjugated to horse radish peroxidase (sc-2357; Santa Cruz Biotechnology Inc., Dallas, TX). Nitrocellulose membranes were developed using an ECL Western Blotting Detection Reagent chemiluminescent kit (Amersham, Pittsburgh, PA) and imaged using the ChemiDoc XRS+ chemiluminescence system. Densitometry calculations were performed using Image Lab Software (Bio-Rad). To account for technical error between membranes, a common sample was included in each Western immunoblot and densitometry measurements were adjusted based on the average densitometry measurement of the common sample.

For analysis of caruncular and cotyledonary concentrations of NOS3, 20 µg of each sample was electrophoresed through 4%–15% Tris-glycine stain-free gels (Bio-Rad), transferred and analyzed as described for CSH. NOS3 was detected using a 1:2,000 dilution of mouse α-NOS3 (BD 610297; BD Biosciences, San Jose, CA) and a 1:5,000 dilution of goat α-mouse IgG conjugated to horse radish peroxidase (sc-2005; Santa Cruz Biotechnology Inc.). For analysis of caruncular (1 µg/sample) and cotyledonary (2.5 µg/sample) concentrations of SLC2A1 (GLUT1), samples were electrophoresed, transferred, and analyzed as described above. SLC2A1 was detected using a 1:2,000 dilution of rabbit α-SLC2A1 (07–1401; EMD Millipore) and a 1:5,000 dilution of goat α-rabbitt IgG conjugated to horseradish peroxidase (ab205718; Abcam). As described above, densitometry of SLC2A1 was normalized on total protein/lane, using Image Lab Software (BioRad).

To analyze the concentration of SLC2A3 (GLUT3), we generated a rabbit α-sheep SLC2A3 antiserum, following the protocol of Erhardt and Bell (32). The COOH-terminal sequence of sheep SLC2A3 (SIQPTKDTNA) was synthesized and conjugated to keyhole limpet hemocyanin, and used to immunize rabbits as described (30, 32). The resulting antiserum (CSU-α-SCL2A3-22) was titered and validated by detection of sheep SLC2A3 expressed in immortalized wild-type sheep OTR cells (33), and cells transduced to overexpress sheep SLC2A3. With both, a single immunoreactive band was detected at an apparent M_r of 50,000. In addition, CSU-α-SCL2A3-22 detected SLC2A3 in human placental extracts, equivalent to an α-human SLC2A3 antisera that we had previously used (3). Subsequently, 10-µg samples of caruncular or cotyledonary tissue were electrophoresed through NuPAGE 4%–12% Bis-Tris Gels (Life Technologies), transferred to nitrocellulose, and the resulting blots stained with Ponceau S to assess total/lane using the ChemiDoc XRS+ (Bio-Rad). Subsequent procedures as are described above, using a 1:1,000 dilution of CSU-α-SLC2A3-22, and a 1:5,000 dilution of a goat α-rabbit IgG conjugated to horseradish peroxidase (ab97051; Abcam). As described above, densitometry of SLC2A3 was normalized on total protein/lane, using Image Lab Software (Bio-Rad).

Statistical Analysis

Data were analyzed by two-way analysis of variance using GraphPad Prism (8.3.1) to analyze the main effects of treatment and fetal sex as well as the treatment × sex interaction. There were no treatment by fetal sex interactions, potentially due to the limited n/group, therefore the data are presented as the main effect of treatment only. Statistical significance was set at P ≤ 0.05 and a statistical tendency at P ≤ 0.10. Data are reported as the means ± standard error (SE). The data figures are presented as scatter plots, with the horizontal line representing the mean, and the capped vertical lines representing the SE.

RESULTS

Fetal and Placental Assessments

A summary of all fetal measurements at necropsy (132 dGA) are provided in Table 1. CSH RNAi fetal weights, organ weights, and measurements of fetal growth did not differ from control RNAi. Although, measures of fetal growth were unchanged, placental and fetal membrane weights from CSH RNAi pregnancies tended (P ≤ 0.10) to be lighter compared to control RNAi. Furthermore, uteroplacental weight also tended (P ≤ 0.10) to be reduced in CSH RNAi pregnancies, whereas uterine weight was unchanged. Placental efficiency tended (P ≤ 0.10) to be increased in CSH RNAi pregnancies. Placentome number was not affected by treatment. Reductions in uterine vein CSH concentrations (ng/mL) did not reach statistical significance, but were reduced by 24%, mirroring the reduction (26%) in cotyledonary CSH concentration (P ≤ 0.10; Fig. 1).

Table 1.

Fetal and maternal measurements collected at necropsy (132 dGA)

	Control RNAi (n = 6)	CSH RNAi (n = 6)	% Change	P Value
Fetal body weight, kg	3.74 ± 0.14	3.82 ± 0.19	2.14	0.87
Maternal weight, kg	58.97 ± 1.94	62.75 ± 2.97	6.41	0.31
Uterine vein oCSH, ng/mL	538.87 ± 129.39	406.66 ± 51.75	24.53	0.31
Umbilical vein oCSH, ng/mL	41.90 ± 5.05	40.08 ± 3.58	4.34	0.46
Uteroplacental weight, kg	2.60 ± 0.29	1.99 ± 0.12	23.37	0.10
Total uterus weight, kg	0.64 ± 0.03	0.60 ± 0.04	5.11	0.46
Placental weight, kg	0.43 ± 0.02	0.38 ± 0.01	11.63	0.06
Placental efficiency	8.75 ± 0.50	9.92 ± 0.35	13.37	0.09
Total membrane weight, kg	0.69 ± 0.07	0.49 ± 0.04	29.07	0.06
Total placentomes	69 ± 5	66 ± 7	5.06	0.89
Crown-rump length, cm	50.8 ± 0.59	52.23 ± 0.68	2.81	0.22
Ponderal index*	2.86 ± 0.12	2.67 ± 0.05	6.64	0.17
Lower leg length (BCB), cm	38.7 ± 1.34	35.63 ± 0.63	7.93	0.11
Brain weight:liver weight	0.45 ± 0.04	0.51 ± 0.05	13.88	0.34

Data are shown as means ± SE for all ewes in each treatment group. *Ponderal index = fetal body wt (g) × 100/crown-rump length (cm³). CSH, chorionic somatomammotropin; dGA, days of gestational age; RNAi, RNA interference.

Figure 1.

Cotyledonary chorionic somatomammotropin (CSH) concentration, relative to total protein transferred, in control RNA interference [RNAi (n = 6)] vs. CSH RNAi (n = 6) pregnancies. Individual data points are presented with the horizontal line representing the mean and the capped vertical lines representing the SE. #P ≤ 0.10 for CSH RNAi pregnancies compared with control RNAi pregnancies.

Uterine and Umbilical Blood Flows

At 132 dGA, uterine and umbilical blood flows were determined by the transplacental diffusion technique. Uterine (mL/min) blood flow (Fig. 2) and uterine blood flow relative to fetal weight (mL/min/kg fetus) were reduced (P ≤ 0.05) in CSH RNAi pregnancies by 22% and 24%, respectively. Uterine blood flow relative to placental weight (435.70 ± 28.28 vs. 377.79 ± 24.72 mL/min/100 g placenta; control RNAi vs. CSH RNAi, respectively) did not statistically differ (P > 0.10) between treatments but was decreased by 13%. A summary of uterine and umbilical plasma flows is presented in Table 2. Uterine plasma flow (mL/min) and uterine plasma flow relative to fetal weight (mL/min/kg fetus) were also reduced (P ≤ 0.05) in CSH RNAi pregnancies, however uterine plasma flow relative to placental weight (mL/min/100 g placenta) did not differ. In contrast, umbilical (mL/min) blood flow (Fig. 3), umbilical blood flow relative to fetal weight (mL/min/kg fetus), and umbilical blood flow relative to placental weight (187.94 ± 20.77 vs. 204.88 ± 8.44 mL/min/100 g placenta; control RNAi vs. CSH RNAi, respectively) were not altered by CSH RNAi. Neither were umbilical plasma flow (Table 2) or relative umbilical plasma flow (mL/min/kg fetus) or plasma flow relative to 100 g placental weight (mL/min/100 g placenta). However, the uterine-to-umbilical flow ratio, tended to be reduced (P ≤ 0.10) in CSH RNAi pregnancies. Neither uterine nor umbilical hematocrits differed by treatment. To assess potential mediators of altered blood flow, NOS3 was assessed in the maternal caruncles and fetal cotyledons. As evidenced in Fig. 4, in response to CSH RNAi, caruncular expression of NOS3 was reduced (P ≤ 0.05) by 24%. In contrast, cotyledonary NOS3 (Fig. 4) was increased 35% in response to CSH RNAi, but this change did not reach statistical significance (P ≈ 0.14).

Figure 2.

Uterine blood flow (mL/min) and relative uterine blood flow (mL/min/kg fetus) in control RNA interference [RNAi (n = 6)] vs. chorionic somatomammotropin (CSH) RNAi (n = 6) pregnancies. Individual data points are presented with the horizontal line representing the mean and the capped vertical lines representing the SE. *P ≤ 0.05 and **P ≤ 0.01, when CSH RNAi pregnancies are compared with control RNAi pregnancies.

Table 2.

Uterine and umbilical plasma flow based on transplacental diffusion technique

	Control RNAi (n = 6)	CSH RNAi (n = 6)	% Change	P Value
Uterine
Uterine plasma flow, mL/min	1231.20 ± 79.40	940.59 ± 67.00	23.60	0.02
Uterine plasma flow, mL/min/kg fetus	328.61 ± 15.18	245.63 ± 10.12	24.02	0.006
Uterine plasma flow, mL/min/100 g placenta	286.95 ± 18.81	244.74 ± 15.90	14.71	0.12
Average uterine arterial hematocrit	0.34 ± 0.00	0.35 ± 0.01	3.07	0.39
Umbilical
Umbilical plasma flow, mL/min	533.6 ± 58.98	529.1 ± 37.32	0.80	0.88
Umbilical plasma flow, mL/min/kg fetus	141.08 ± 10.63	138.97 ± 7.82	1.49	0.80
Umbilical plasma flow, mL/min/100g placenta	125.48 ± 15.97	137.01 ± 5.97	9.19	0.51
Average umbilical arterial hematocrit	0.34 ± 0.02	0.33 ± 0.01	1.39	0.51
Uterine:umbilical flow	2.40 ± 0.19	1.87 ± 0.17	22.21	0.09

Data are shown as means ± SE for all ewes in each treatment group. CSH, chorionic somatomammotropin; RNAi, RNA interference.

Figure 3.

Umbilical blood flow (mL/min) and relative umbilical blood flow (mL/min/kg fetus) in control RNA interference [RNAi (n = 6)] vs. chorionic somatomammotropin (CSH) RNAi (n = 6) pregnancies. Individual data points are presented with the horizontal line representing the mean and the capped vertical lines representing the SE.

Figure 4.

Densiometric analysis of NOS3, relative to total protein transferred, in maternal caruncles and fetal cotyledons from control RNA interference (RNAi) and chorionic somatomammotropin (CSH) RNAi pregnancies (n = 6/treatment). Individual data points are presented with the horizontal line representing the mean and the capped vertical lines representing the SE. **P ≤ 0.01, when CSH RNAi pregnancies are compared with control RNAi pregnancies.

Blood Gas and Oxygen Uptakes

A summary of uterine, umbilical, and uteroplacental oxygen uptakes and concentrations can be found in Supplemental Table S2. The uterine artery to vein O₂ concentration gradient (mmol/min) was elevated (P ≤ 0.05) in response to CSH RNAi. However, because of the reduction in uterine blood flow in CSH RNAi, uterine O₂ uptake was not different. Limited differences were observed in uterine and umbilical blood gas measurements, which are provided in Supplemental Tables S3 and S4. Uterine artery methemoglobin levels were elevated (P ≤ 0.05) in CSH RNAi pregnancies and uterine venous PO₂ was reduced (P ≤ 0.01). No other uterine or umbilical blood gas measurements were different. The lack of response in oxygen uptake continued on the fetal side, as umbilical uptakes of oxygen did not differ from control RNAi. Uteroplacental O₂ uptakes (mmol/min), oxygen utilization and fraction of oxygen being transferred to the fetus did not change as a result of CSH RNAi.

Glucose Uptakes

Uterine, umbilical, and uteroplacental glucose gradients and concentrations are summarized in Table 3. Uterine arterial plasma glucose tended to be elevated (P ≤ 0.10) in CSH RNAi pregnancies, however uterine venous glucose was unchanged. The uterine artery to vein glucose gradient (mmol/L) was elevated (P ≤ 0.01) in CSH RNAi pregnancies, yet the uterine uptake of glucose was not altered (Table 3; Fig. 5). CSH RNAi had no effect on umbilical plasma arterial or venous glucose concentrations as well as the umbilical vein to artery glucose gradient. Because of the lack of response to CSH RNAi on umbilical blood flow and the umbilical glucose gradient, the umbilical uptake of glucose (Fig. 5) was not impacted by CSH RNAi. In contrast to the uterine and umbilical measures, CSH RNAi resulted in a 27% increase in (P ≤ 0.05) uteroplacental glucose uptake relative to placental weight (µmol/min/kg; Fig. 5) and a tendency (P ≤ 0.10) for uteroplacental glucose uptake (µmol/min) to be increased (Table 3). This observation is supported by the increased fraction of uterine glucose uptake being utilized by the placenta (P ≤ 0.05; Fig. 6) in CSH RNAi pregnancies. Because more glucose was actually being utilized by the placenta, the fraction of uterine glucose uptake being transferred to the fetus was decreased (P ≤ 0.05; Fig. 6). To further assess the changes in glucose uptake and transport, SLC2A1 and SLC2A3 concentrations in maternal caruncles and fetal cotyledons were determined. As evidenced in Fig. 7, caruncular SLC2A1 concentration was not impacted by CSH RNAi, but cotyledonary SLC2A1 tended (P ≤ 0.10) to be reduced (25%). By contrast, there appeared to be no significant impact of CSH RNAi on SLC2A3 concentration (Fig. 8) in either the maternal caruncles or fetal cotyledons.

Table 3.

In vivo measurements of glucose transfer and uptake based on the Fick principle

	Control RNAi (n = 6)	CSH RNAi (n = 6)	% Change	P Value
Uterine
Uterine artery plasma glucose concentration, mmol/L	3.40 ± 0.08	3.68 ± 0.10	8.24	0.10
Uterine vein plasma glucose concentration, mmol/L	3.15 ± 0.09	3.31 ± 0.08	5.17	0.30
Uterine artery to vein plasma glucose concentration, mmol/L	0.25 ± 0.02	0.37 ± 0.03	48.00	0.006
Uterine glucose uptake, µmol/min	431.77 ± 44.27	479.08 ± 34.16	10.96	0.42
Umbilical
Umbilical artery plasma glucose concentration, mmol/L	0.85 ± 0.05	0.92 ± 0.07	8.24	0.35
Umbilical vein plasma glucose concentration, mmol/L	1.06 ± 0.06	1.11 ± 0.07	4.72	0.55
Umbilical vein to artery plasma concentration, mmol/L	0.22 ± 0.02	0.19 ± 0.01	13.64	0.22
Umbilical glucose uptake, µmol/min	155.70 ± 12.37	139.29 ± 10.47	10.54	0.34
Uterine glucose uptake transferred to the fetus	0.37 ± 0.04	0.29 ± 0.02	21.54	0.05
Uteroplacental
Uterine artery-umbilical vein glucose gradient, mmol/L	2.53 ± 0.11	2.78 ± 0.07	9.88	0.14
Uteroplacental glucose uptake, µmol/min	276.07 ± 40.40	339.79 ± 28.57	23.08	0.06
Uterine glucose uptake utilized by the placenta	0.63 ± 0.04	0.71 ± 0.02	12.92	0.05

Data are shown as means ± SE for all ewes in each treatment group. CSH, chorionic somatomammotropin; RNAi, RNA interference.

Figure 5.

Relative uterine (µmol/min/kg uterus), umbilical (µmol/min/kg fetus), and uteroplacental (µmol/min/kg placenta) uptakes of glucose in control RNAi (n = 6) vs. chorionic somatomammotropin (CSH) RNA interference [RNAi (n = 6)] pregnancies. Individual data points are presented with the horizontal line representing the mean and the capped vertical lines representing the SE. *P ≤ 0.05, when CSH RNAi pregnancies are compared with control RNAi pregnancies.

Figure 6.

Fraction of uterine uptake of glucose being utilized by the placenta vs. transferred to the fetus in control RNA interference [RNAi (n = 6)] vs. chorionic somatomammotropin (CSH) RNAi (n = 6) pregnancies. Individual data points are presented with the horizontal line representing the mean and the capped vertical lines representing the SE. *P ≤ 0.05, when CSH RNAi pregnancies are compared with control RNAi pregnancies.

Figure 7.

Densiometric analysis of SLC2A1, relative to total protein transferred, in maternal caruncles and fetal cotyledons from control RNA interference (RNAi) and chorionic somatomammotropin (CSH) RNAi pregnancies (n = 6/treatment). Individual data points are presented with the horizontal line representing the mean and the capped vertical lines representing the SE. #P ≤ 0.10, when CSH RNAi pregnancies are compared with control RNAi pregnancies.

Figure 8.

Densiometric analysis of SLC2A3, relative to total protein transferred, in maternal caruncles and fetal cotyledons from control RNA interference (RNAi) and chorionic somatomammotropin (CSH) RNAi pregnancies (n = 6/treatment). Individual data points are presented with the horizontal line representing the mean and the capped vertical lines representing the SE.

Uptake of Other Nutrients

Uterine, umbilical, and uteroplacental lactate uptakes and concentrations are provided in Supplemental Table S5. CSH RNAi pregnancies tended (P ≤ 0.10) to have decreased total placental lactate production as well as umbilical lactate uptake relative to placental mass (µmol/kg/min). However, no other differences were observed in lactate uptakes or concentrations. Individual amino acid uptakes by the uterine and umbilical vasculature are provided in Supplemental Tables S6, S7 and S8. Relative uterine (µmol/min/kg uterine weight) uptake of taurine and glycine was reduced (P ≤ 0.05) in CSH RNAi pregnancies, without significant changes in the uterine uptake of other amino acids. Only relative uteroplacental (µmol/min/kg placenta) glycine uptake was reduced (P ≤ 0.05), with no changes in the uteroplacental uptake of other amino acids. CSH RNAi did not significantly impact relative umbilical uptake (µmol/min/kg fetus) of any amino acids.

Uterine and umbilical carbon and nitrogen uptakes as well as nutrient to oxygen quotients are provided in Supplemental Table S9. Overall, uterine carbon and nitrogen uptakes did not differ between CSH RNAi and control RNAi pregnancies. In CSH RNAi pregnancies, umbilical carbon and nitrogen uptake from amino acids tended (P ≤ 0.10) to be increased, but those effects were canceled out as carbon uptake from lactate tended (P ≤ 0.10) to be decreased. Umbilical carbon uptake from glucose was not altered by treatment. Although minor changes in the proportion of carbon sources being taken up by the umbilicus between different oxidative sources might exist, the total carbon uptake and the nutrient quotients available for oxidative metabolism did not differ between treatments.

DISCUSSION

Baker et al. (2) previously reported two distinct phenotypes associated with lentiviral-mediated CSH RNAi in near-term pregnancies (135 dGA). The first phenotype had placental weight value 2 SDs below the mean of control pregnancies. In those pregnancies, significant reductions in fetal body weight (32%) and placental weight (52%) occurred. The second phenotype of “nonresponder pregnancies” did not exhibit the same degree of reductions in fetal and placental weight. These dual phenotypes agree with observations in human pregnancies. Daikoku et al. (4) reported suppressed maternal CSH concentrations in IUGR pregnancies at 25–28 wk of gestational age and again from 33–40 wk. Pregnancies with reduced CSH yielded babies ≤3% of birth weight, which tracks with the CSH RNAi IUGR phenotype described in sheep (2). Interestingly, Daikoku et al. (4) also observed that some pregnancies had suppressed maternal CSH from 33 to 40 wk of gestational age, but gave birth to non-low birth weight pregnancies. This second phenotype mimics what was observed in the non-IUGR CSH-deficient phenotype in the Baker et al. (2) study, as well as in the present study. In fact, numerous case studies report CSH loci deletions or mutations that either result in IUGR pregnancies (8, 9, 10) or non-IUGR phenotypes (11, 12, 13). One study reported that up to 40% of pregnancies that had low maternal levels of CSH had no obvious obstetric abnormality (related to birth weight) (34). More studies are needed to investigate the true prevalence of human IUGR and non-IUGR outcomes associated with CSH deficiencies, with emphasis placed on whether or not pregnancies and/or offspring are truly physiologically normal. This need is highlighted by our recent report (7) that the CSH RNAi pregnancies from the Baker et al. (2) study exhibited significant changes in umbilical concentrations of IGF1 and elevated fetal liver concentrations of the insulin receptor. These recent data (7) imply that the progression of fetal development was altered by CSH RNAi, even in the absences of placental and fetal growth restriction.

Although it is clear that two phenotypes can emerge in spite of a deficiency in CSH, no studies have yet to examine the physiological ramifications of CSH deficiency on uteroplacental function. This is the rationale for creating sheep pregnancies with CSH deficiency based on RNAi. Larger uterine artery-vein glucose gradients in both the IUGR and non-IUGR CSH RNAi pregnancies when compared with control pregnancies (2, 7) suggest the potential for metabolic perturbations in response to CSH deficiency regardless of fetal weight outcome. Therefore, the main objective of this study was to examine the impacts of CSH RNAi in the absence of IUGR on placental and fetal physiology. Our results support the idea that CSH RNAi impairs uteroplacental function even in the absence of IUGR.

In the present study, fetal weights did not differ between treatments, yet uteroplacental, placental, and fetal membrane weights tended to be reduced as a result of CSH RNAi. Experimentally, one factor that might explain the milder fetal growth phenotype in this study as compared with the previous report (2) is a lower efficiency in CSH RNAi in this cohort. Although uterine vein CSH and cotyledonary CSH concentrations were reduced by ∼25%, in the Baker et al. (2) study, the CSH RNAi pregnancies with IUGR had a 38% reduction in cotyledonary CSH concentrations. Several factors can change RNAi efficacy, including integration rate of RNAi vector, length of time between initiation of RNAi and sample harvest, and effective competition for RISC complexes (35, 36, 37,). In addition, if CSH RNAi was more robustly active during placental development, it is possible that a more severe phenotype would be produced like the ones we observed previously (2, 3). Unfortunately, no feasible methods exist to measure RNAi expression in vivo across pregnancy without terminating those pregnancies before their experimental end points. Another difficulty in determining CSH concentrations is the well-documented variability in CSH secretion (38, 39, 40). In the current cohort, uterine and umbilical vein samples were only collected at the time of metabolic study (132 dGA). Because CSH in circulation can fluctuate rapidly and lacks circadian rhythms for secretion (38, 39, 40), increased sampling frequency and duration maybe warranted. It is also worth noting that in two different models of IUGR in sheep (6, 31), maternal CSH was significantly depressed in the absence of any change in cotyledonary CSH mRNA or protein concentrations. This infers that there is not always a direct correlation between CSH mRNA and protein concentrations and what is secreted, suggesting that there is still a deficit in our knowledge of CSH regulation and secretion.

Even in the absence of IUGR, as a result of CSH RNAi, physiological changes were observed in this study. Previous studies have attempted to associate CSH concentrations and blood flow. In pregnant women, Gitsch et al. (41) and Qiao et al. (42) observed concomitant decreases in CSH in maternal circulation and decreases in placental perfusion or uteroplacental blood flow in cases of pregnancy-related hypertension or placental insufficiency. However, in goats (43), uterine blood flow was measured simultaneously with blood sampling to examine CSH concentrations, and no direct relationship in short-term uterine blood flow was observed. In the present study, we found that even in the absence of growth restriction, profound reductions in uterine blood and plasma flow occur as a result of CSH RNAi, suggesting a potential role for CSH in modulating uterine blood flow. To our knowledge, this is the first study to examine the direct impact of CSH deficiency on uterine and umbilical blood flow, and supports the concept that even in the absence of IUGR, CSH-deficient pregnancies have altered uteroplacental function. We also determined that CSH RNAi resulted in significant reductions in maternal caruncle NOS3 concentration, which coincides with the reduction in uterine blood flow, as the relationship between NOS3 and uterine blood flow has been demonstrated previously (44, 45). NOS3 null mice exhibit major reductions in uterine blood flow during mid- to late-gestation (44). This relationship was also observed in sheep using a NOS-specific antagonist (L-NAME) to investigate estrogen-mediated effects on NOS and subsequent changes in uterine blood flow (45). It is possible that the impact of CSH RNAi on uterine blood flow and NOS3 is indirect, and may be mediated by placental steroids. In contrast to uterine blood flow, there were no significant changes in umbilical blood flow. Perhaps, in more severe cases of CSH deficiency, where IUGR is apparent (2, 3), umbilical blood flows might be decreased as well.

Glucose is a key oxidative substrate for the fetus and many IUGR pregnancies are characterized by a reduction in placental glucose transport (46). Even in the absence of fetal growth restriction, CSH RNAi pregnancies exhibited altered uteroplacental glucose uptake. The maternal artery-uterine vein gradient was greater in CSH RNAi pregnancies, which agrees with previous findings (2, 7). However, the uterine uptake of glucose was not statistically different, likely due to the decrease in uterine blood flow in CSH RNAi pregnancies. Maternal caruncle SLC2A1 concentrations were not significantly impacted by CSH RNAi, but fetal cotyledon concentrations of SLC2A1 was reduced 25%, similar to what we observed with CSH RNAi pregnancies exhibiting fetal growth restriction (3). Neither caruncular or cotyledonary SLC2A3 concentrations were significantly altered in these CSH RNAi pregnancies, which is in contrast to the significant reduction of cotyledonary SLC2A3 observed in growth-restricted CSH RNAi pregnancies (3). This might infer a compensatory maintenance of SLC2A3 in response to CSH RNAi in our nongrowth-restricted pregnancies, not dissimilar to what was observed in late-onset IUGR human pregnancies (47). The placenta is a highly metabolic organ and can consume upward of 65% of the glucose it transports for its own oxidative needs (20, 24, 28). In this study, relative uteroplacental uptake of glucose was increased by 27% in CSH RNAi pregnancies, but the umbilical uptake of glucose was not altered. This may suggest that CSH deficiency, at least in the absence of IUGR, is actually increasing the glucose demands of the placenta. This is supported by the increased fraction of uterine glucose uptake being utilized by the placenta and reductions in the fraction of uterine glucose being transferred to the fetus. We have previously reported (3) altered placental gene expression in response to CSH RNAi. Perhaps in more severe cases of IUGR as a result of CSH deficiency, placental glucose consumption increases to the point where it can no longer support normal fetal growth. However, in order to address this question, IUGR phenotypes in response to CSH RNAi need to be studied in the same fashion.

In this cohort of CSH RNAi pregnancies, in the absence of IUGR, uteroplacental glucose uptake was significantly increased, suggesting metabolic perturbations. Adult onset disease risk is increased in offspring due to in utero exposure to adverse events, leading to long-term irreversible changes in physiology and metabolism (50, 51, 52, 53). In sheep, a variety of models exist to examine the impacts of placental insufficiency and/or altered fetal growth, ranging from elevated core body temperatures, maternal inflammation, uteroplacental embolization, uterine artery ligation or occlusion, carunclectomies, and nutritional manipulation (54). Sheep models of nutritional manipulation have been followed postnatally to study fetal programming consequences in adulthood (55). When ewes were either overfed (150% NRC from -60 dpc to term) or underfed (50% NRC from 28–78 dGA then realimented to 100% NRC from 78 dGA to term), their phenotypes diverged at midgestation (78 dGA), with overfed having fetuses larger than control and underfed smaller than controls (55). However, both phenotypes then converged by late gestation and had similar lamb weights at birth (55). In adulthood, offspring from both over and underfed dams displayed compensatory growth, increased insulin resistance, and increased adiposity (55). These effects proved to be multigenerational as F2 offspring from overfed ewes also had increased adiposity, glucose, and insulin at birth (56). In the current study, changes in uteroplacental glucose uptake and uterine blood flow, as well as changes in uterine A-V glucose gradients show evidence of metabolic perturbations. Taken together with our previous studies (2, 3, 7), it appears that CSH deficiency could predispose lambs to fetal programming outcomes regardless of birth weight. Furthermore, it potentially raises a concern about the adult outcomes of human pregnancies in the normal birthweight range, following a CSH-deficient pregnancy.

PERSPECTIVES AND SIGNIFICANCE

While CSH RNAi did not result in significant reductions in near-term fetal weight in this cohort, this was the first study to examine the direct impacts of in vivo CSH RNA interference on in vivo uteroplacental nutrient transport and hemodynamics. We established the relationship between CSH RNAi and reduced uterine blood flow and increased placental glucose utilization. This study highlights that significant physiological changes may occur during pregnancy without noticeable impacts on birthweight, likely predisposing the offspring to altered health and well-being. By examining in vivo placental physiology and fetal development in both IUGR and non-IUGR pregnancies in the future, we may potentially define the cause and effect relationship between CSH and normal pregnancy physiology.

GRANTS

This work was supported by National Institutes of Health Grants HD093701, HD094952, DK08813, and S10OD023553, and Agriculture and Food Research Initiative Grants 2016–38420-25289 and 2019–67011-29614 from the United State Department of Agriculture.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

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