Intermittent maternofetal oxygenation during late gestation improved birthweight, neonatal growth, body symmetry, and muscle metabolism in intrauterine growth-restricted lambs

Abstract

In humans and animals, intrauterine growth restriction (IUGR) results from fetal programming responses to poor intrauterine conditions. Chronic fetal hypoxemia elevates circulating catecholamines, which reduces skeletal muscle β2 adrenoceptor content and contributes to growth and metabolic pathologies in IUGR-born offspring. Our objective was to determine whether intermittent maternofetal oxygenation during late gestation would improve neonatal growth and glucose metabolism in IUGR-born lambs. Pregnant ewes were housed at 40 °C from the 40th to 95th day of gestational age (dGA) to produce IUGR-born lambs (n = 9). A second group of IUGR-born lambs received prenatal O₂ supplementation via maternal O₂ insufflation (100% humidified O₂, 10 L/min) for 8 h/d from dGA 130 to parturition (IUGR+O2, n = 10). Control lambs (n = 15) were from pair-fed thermoneutral ewes. All lambs were weaned at birth, hand-reared, and fitted with hindlimb catheters at day 25. Glucose-stimulated insulin secretion (GSIS) and hindlimb hyperinsulinemic-euglycemic clamp (HEC) studies were performed at days 28 and 29, respectively. At day 30, lambs were euthanized and ex vivo HEC studies were performed on isolated muscle. Without maternofetal oxygenation, IUGR lambs were 40% lighter (P < 0.05) at birth and maintained slower (P < 0.05) growth rates throughout the neonatal period compared with controls. At 30 d of age, IUGR lambs had lighter (P < 0.05) hindlimbs and flexor digitorum superficialis (FDS) muscles. IUGR+O2 lambs exhibited improved (P < 0.05) birthweight, neonatal growth, hindlimb mass, and FDS mass compared with IUGR lambs. Hindlimb insulin-stimulated glucose utilization and oxidation rates were reduced (P < 0.05) in IUGR but not IUGR+O2 lambs. Ex vivo glucose oxidation rates were less (P < 0.05) in muscle from IUGR but not IUGR+O2 lambs. Surprisingly, β2 adrenoceptor content and insulin responsiveness were reduced (P < 0.05) in muscle from IUGR and IUGR+O2 lambs compared with controls. In addition, GSIS was reduced (P < 0.05) in IUGR lambs and only modestly improved (P < 0.05) in IUGR+O2. Insufflation of O₂ also increased (P < 0.05) acidosis and hypercapnia in dams, perhaps due to the use of 100% O₂ rather than a gas mixture with a lesser O₂ percentage. Nevertheless, these findings show that intermittent maternofetal oxygenation during late gestation improved postnatal growth and metabolic outcomes in IUGR lambs without improving muscle β2 adrenoceptor content.

Introduction

Placental insufficiency induces developmental programming that leads to intrauterine growth restriction (IUGR) and reduced postnatal muscle mass (reviewed by Yates et al., 2018). These programming mechanisms also impair metabolic function, which predisposes IUGR-born humans to obesity, diabetes, and other lifelong metabolic disorders (Hales and Barker, 2013). In livestock, low birthweight due to IUGR is associated with poor growth efficiency and carcass merit (Greenwood et al., 2005). Placental insufficiency causes progressive fetal hypoxemia during late gestation that in turn chronically elevates circulating catecholamines (Leos et al., 2010; Benjamin et al., 2017). Studies in IUGR fetal sheep have implicated both hypoxemia and hypercatecholaminemia as factors in poor skeletal muscle growth, glucose metabolism, and pancreatic β-cell function (Yates et al., 2014, 2016; Brown et al., 2015; Macko et al., 2016; Davis et al., 2020). When exposure to elevated catecholamines is sustained, fetal tissues develop changes in adrenergic sensitivity (Chen et al., 2010, 2014, 2017; Macko et al., 2016; Yates et al., 2018; Davis et al., 2020). In muscle, this includes reduced expression of the β2 adrenoceptor, which persists postnatal. Because β2 adrenergic activity enhances insulin signaling and stimulates muscle growth and glucose metabolism (Arp et al., 2014; Cadaret et al., 2017), we have postulated that this is a key mechanism underlying the programming of growth and metabolic deficits in IUGR skeletal muscle (Yates et al., 2018; Posont and Yates, 2019).

Skeletal muscle is a primary target for nutrient-sparing developmental programming in the IUGR fetus due to its high glucose consumption and metabolic plasticity (Brown, 2014). Studies in sheep show that phosphorylation of the insulin signaling pathway hub, Akt, is impaired in IUGR skeletal muscle before and after birth (Thorn et al., 2009; Cadaret et al., 2019; Posont et al., 2021), which helps to explain insulin resistance associated with IUGR. Stimulation of the β2 adrenoceptor in normal skeletal muscle enhances insulin-stimulated Akt phosphorylation (Cadaret et al., 2017), and thus the reduction in β2 adrenoceptor content observed in IUGR skeletal muscle may play a direct role in reduced insulin responsiveness. Muscle growth in late gestation and after birth requires properly functioning progenitor cells called myoblasts, but IUGR fetal myoblasts exhibited intrinsic functional deficits near term, which coincided with reduced fiber hypertrophy (Yates et al., 2014, 2016; Posont et al., 2018). Insulin promotes proliferation and differentiation of myoblasts (Brown et al., 2016) via Akt-mediated signaling pathways (Vandromme et al., 2001), and thus the potential impact of β2 adrenoceptor deficits on insulin action would help to explain muscle growth deficits. These same programming mechanisms may also underlie the metabolic dysfunction and poor glucose homeostasis that characterizes IUGR-born offspring (Jornayvaz et al., 2004; Vind et al., 2012). Reduced whole-body glucose oxidation and greater lactate production observed in IUGR fetal sheep (Limesand et al., 2007; Thorn et al., 2009; Brown et al., 2015) indicates a nutrient-sparing shift from oxidative to glycolytic metabolism. Our earlier studies in IUGR lambs found that impaired capacity for glucose oxidation was skeletal muscle-centric and persisted after birth (Cadaret et al., 2019; Yates et al., 2019; Posont et al., 2021). Moreover, β2 adrenoceptor stimulation increased basal and insulin-stimulated glucose oxidation in primary skeletal muscle from rats (Cadaret et al., 2017). Pancreatic β-cell dysfunction is another hallmark of IUGR offspring (Jensen et al., 2002; Camacho et al., 2017; Yates et al., 2019), and chronic elevation of fetal catecholamines produces adrenergic adaptations in pancreatic islets as well (Limesand et al., 2006; Chen et al., 2014). Although hypoxemia-induced hypercatecholaminemia reduced fetal islet function (Yates et al., 2012; Macko et al., 2016), ablation of adrenal-derived catecholamines via demedullation impaired glucose-stimulated insulin secretion (GSIS) and β-cell mass in normal fetuses (Yates et al., 2012; Davis et al., 2015; Macko et al., 2016). Thus, we postulated that basal adrenergic tone necessary for proper islet development and function may be lost due to adrenergic programming (Yates et al., 2012).

Studies in humans and animals have shown that maternal O₂ insufflation increases fetal O₂ status (Khaw et al., 2002; Tomimatsu et al., 2006, 2007) and can improve fetal survival in severe IUGR pregnancies (Vileisis, 1985; Nicolaides et al., 1987; Battaglia et al., 1992; Say et al., 2003). Fetal hypoxemia is primarily responsible for chronic hypercatecholaminemia. In the sheep fetus, plasma norepinephrine is highly correlated with blood O₂ (Limesand et al., 2006), and we have shown that acutely increasing the O₂ fraction of inspired air in the pregnant ewe results in recovery of normal O₂ partial pressures in IUGR fetal sheep (Macko et al., 2016). Thus, we hypothesized that better growth and metabolic outcomes in IUGR offspring could be achieved via improved fetal oxygenation, which we expected would reduce circulating catecholamines and therefore circumvent the reductions in muscle β2 adrenoceptor content. Because static maternal hyperoxygenation sustained for extended periods of time can affect cortical and cardiac development in the fetus (Edwards et al., 2019; Markert et al., 2020), our objective was to assess the effects of intermittent O₂ supplementation. We specifically sought to identify benefits for postnatal growth, body composition, skeletal muscle glucose metabolism, and β-cell function in the neonatal IUGR lamb.

Materials and Methods

Animals and experimental design

This study was approved by the Institutional Animal Care and Use Committee at the University of Nebraska-Lincoln, which is accredited by AAALAC International. Thirty-four commercial Polypay ewes were purchased from a single source and were timed-mated to a single sire at 24 to 48 months of age. Ewes were penned individually and nutrition was managed as previously described (Cadaret et al., 2019; Posont et al., 2021). Ewes were assigned via simple randomization to produce control lambs (n = 15 lambs from 9 different ewes, 3 singleton-born and 12 twin-born, 8 males and 7 females), intrauterine growth-restricted lambs (IUGR; n = 9 lambs from 6 different ewes, 3 singleton born and 6 twin born, 5 males and 4 females), or IUGR lambs + maternofetal oxygenation (IUGR+O2; n = 10 lambs from 8 different ewes, 3 singleton born and 7 twin born, 4 males and 6 females). IUGR lambs were produced using the maternal hyperthermia model for placental insufficiency as previously described (Beede et al., 2019; Yates et al., 2019). Briefly, ewes were exposed to heat-stress conditions (40 °C for 12 h/d, 35 °C for 12 h/d; 35% relative humidity) from the 40th to the 95th day of gestational age (dGA) and returned to thermoneutral conditions (25 °C, 15% relative humidity) thereafter. Ewes carrying control lambs were maintained under thermoneutral conditions for the entirety of gestation. Tracheal catheters were surgically placed in all ewes at 130 dGA. From dGA 131 to parturition (average of 14 d), ewes carrying IUGR+O2 lambs were supplemented with 100% USP Medical Grade O₂ (Airgas, Inc., Lincoln, NE) as previously described (Macko et al., 2016). The O₂ was infused through the tracheal catheter, which contained an in-line humidifier, at a constant rate of 10 L/min for 8 h/d. Blood gas analysis was performed for maternal venous blood samples collected from the first two ewes receiving O₂ supplementation in order to estimate changes in maternal blood gases. Ewes were allowed to deliver lambs naturally. Term was 145.5 ± 0.3, 145.8 ± 1.1, and 145.1 ± 0.6 dGA for controls, IUGR, and IUGR+O2 lambs, respectively. Lambs were weaned on the day of birth, administered colostrum (collected from a pool of previous ewes and stored at −20 °C) for the first 36 h, and then reared on ad libitum commercial milk replacer (Milk Specialities Co., Dundee, IL) as previously described (Yates et al., 2019; Posont et al., 2021). Lambs were weighed daily and biometrics were assessed weekly. Surgical hindlimb catheterizations were performed at 25 ± 1 d of age, square-wave hyperglycemic clamps were performed at 29 ± 1 d of age, and hyperinsulinemic-euglycemic clamps (HEC) were performed at 30 ± 1 d of age as previously described (Posont et al., 2021). Daily arterial blood samples were collected from d 26 to 31, and lambs were euthanized via barbiturate overdose at 31 ± 1 d of age. Organs were weighed and primary skeletal muscle was collected for ex vivo metabolic studies. Additionally, three pilot ewes (2 singleton pregnancies and 1 twin pregnancy) were housed under thermoneutral conditions and underwent tracheal and fetal arterial catheterization surgeries at dGA 128 as previously described (Cadaret et al., 2019). To estimate the average changes in fetal blood gases during maternal O₂ infusion (8 h/d at 10 L/min; dGA 130 to 133), fetal blood samples were collected 2, 4, and 6 h after beginning the O₂ infusion (i.e., during the insufflation period) on each of the 3 d and were averaged for each fetus. Additional fetal blood samples were collected 20 min before and 1 h after O₂ infusion (i.e., outside the insufflation period) on each of the 3 d and averaged for each fetus. These pilot ewes were euthanized prior to parturition.

Surgical preparations

Maternal tracheal catheterizations

Indwelling tracheal catheters were surgically placed in all ewes under general anesthesia as previously described (Yates et al., 2012). Briefly, a fistula was made on the centerline of the trachea eight rings below the larynx. The leading end of the Tygon catheter (US Plastics, Lima, OH) was inserted through the fistula and advanced approximately 5 cm down the trachea. The catheter was anchored via suture to the fascia surrounding the trachea, and the external end was tunneled subcutaneously to the back of the neck, exteriorized through a small incision in the skin, capped, and stored in a mesh pouch sutured to the skin.

Offspring hindlimb catheterizations

Surgical hindlimb preparations were performed in lambs as previously described (Yates et al., 2019; Posont et al., 2021). Under general anesthesia, indwelling Tygon catheters filled with heparinized saline were placed into the descending aorta via the femoral artery for arterial blood sampling and into the inferior vena cava via the femoral vein for infusions during the metabolic studies. In the contralateral hindlimb, a catheter was placed into the external iliac vein via the distal femoral vein for venous blood sampling, and a Precision S-series Flow Probe (Transonic Systems, Inc., Ithaca, NY) was placed around the external iliac artery. The deep circumflex iliac artery and vein of this hindlimb were ligated and severed. Catheters and flow probe cables were tunneled subcutaneously to the flank, exteriorized, and kept in a mesh pouch sutured to the skin.

In vivo metabolic studies

Glucose-stimulated insulin secretion

A square-wave hyperglycemic clamp was performed in order to estimate insulin secretion under resting (i.e., basal insulin secretion) and steady-state hyperglycemic conditions (i.e., GSIS) as previously described (Yates et al., 2019; Posont et al., 2021). Briefly, unfasted lambs were placed in Panepinto slings, and basal blood glucose and plasma insulin concentrations were determined from the average of three arterial blood samples collected in 5-min intervals. Hyperglycemic conditions equivalent to 2-fold basal blood glucose concentration (±10%) were induced via an intravenous bolus and variable-rate infusion of 33% dextrose solution. Once steady-state hyperglycemia had been achieved for a minimum of 20 min, three additional blood samples were collected at 5-min intervals to determine second-phase plasma insulin concentrations. Due to catheter failures in some lambs, GSIS was measured in 12 controls (3 singleton-born/9 twin-born, 5 male/7 female), 7 IUGR (2 singleton-born/5 twin-born, 4 male/4 female), and 9 IUGR+O2 (3 singleton-born/6 twin-born, 3 male/6 female) lambs.

Hindlimb glucose metabolism

Hindlimb-specific glucose metabolic rates were determined under basal and hyperinsulinemic conditions during a HEC as previously described (Yates et al., 2019; Posont et al., 2021). Briefly, lambs were infused with [¹⁴C(U)]-d-glucose (37.2 μCi/ml; PerkinElmer, Boston, MA) at 2 ml/h. After 40 min, four pairs of simultaneous arterial and venous blood samples were collected in 10-min intervals under these resting (i.e., basal) conditions. Hyperinsulinemia was then induced by infusing insulin (250 mU/kg; Humulin R; Lilly, Indianapolis, IN) at 4 mU/kg/min. Euglycemia (basal blood glucose ± 10%) was concurrently maintained by variable-rate infusion of 33% dextrose. After 60 min, four additional pairs of simultaneous arterial and venous blood samples were collected at 10-min intervals under these HEC conditions. Blood flow rates into the hindlimb through the exterior iliac artery were measured throughout the study. Hindlimb-specific metabolic fluxes were estimated via the Fick Principle (Camacho et al., 2017; Cadaret et al., 2019). Hindlimb glucose utilization and oxidation rates were estimated from arteriovenous differences in the concentrations of glucose and radiolabeled ¹⁴CO₂, respectively, which were determined as previously described (Yates et al., 2019; Posont et al., 2021). Differences were normalized to blood flow rate and hindlimb mass. Due to catheter and/or flow probe failures in some lambs, HEC studies were performed for 11 controls (3 singleton-born/8 twin-born, 4 male/7 female), 6 IUGR (2 singleton-born/4 twin-born, 3 male/3 female), and 9 IUGR+O2 (3 singleton-born/6 twin-born, 3 male/6 female) lambs.

Blood analyses

Total white blood cells, lymphocytes, monocytes, granulocytes, hematocrit, hemoglobin, red blood cells, platelets, mean red blood cell volumes, mean corpuscular hemoglobin concentrations, red blood cell distribution widths, and mean platelet volumes were determined in EDTA-treated whole blood with a HemaTrue Veterinary Hematology Analyzer (Heska, Loveland, CO) as previously described (Barnes et al., 2019; Swanson et al., 2020). Glucose, lactate, HCO₃, pH, partial pressure of O₂ (pO₂), partial pressure of CO₂ (pCO₂), oxyhemoglobin, and carboxyhemoglobin were determined in heparinized whole blood with an ABL90 FLEX blood gas analyzer (Radiometer, Brea, CA) as previously described (Cadaret et al., 2019; Posont et al., 2021). Blood plasma insulin concentrations were determined with a commercial ELISA Kit (Ovine Insulin, Alpco Diagnostics, Windham, NH) as previously described (Cadaret et al., 2019; Swanson et al., 2020). Intra- and inter-assay coefficients of variance were less than 10%. Total blood volume collected from lambs did not exceed 20 ml for any 24-h period.

Skeletal muscle analyses

Ex vivo glucose metabolism

Glucose uptake and oxidation rates were determined in intact longitudinal strips (769 ± 19 mg) of flexor digitorum superficialis (FDS) muscles as previously described (Cadaret et al., 2017; Posont et al., 2021). Briefly, muscle strips were incubated for 1 h at 37 °C in O₂-saturated Krebs–Henseleit bicarbonate buffer with 0.1% bovine serum albumin (KHB) spiked with 5 mM d-glucose (Millipore Sigma) and 0 (i.e., basal) or 5 mU/ml insulin (Humulin R). Strips were then washed in glucose-free KHB with 0 or 5 mU/ml insulin for 20 min at 37 °C. To estimate glucose uptake, muscle strips were incubated for 20 min at 37 °C in KHB spiked with 1 mM [³H]2-deoxyglucose (300 µCi/mmol; PerkinElmer), 39 mM [1-¹⁴C] mannitol (1.25 µCi/mmol; PerkinElmer), and 0 or 5 mU/ml insulin. Specific activities for ³H and ¹⁴C were determined in muscle strip lysates by liquid scintillation to estimate glucose uptake and extracellular fluid, respectively. To estimate glucose oxidation, muscle strips were incubated in a sealed dual-well chamber for 2 h at 37 °C in KHB spiked with 5 mM [¹⁴C-U]-d-glucose (0.25 µCi/mmol) and 0 or 5 mU/ml insulin. CO₂ produced by the muscle strip was captured by NaOH in the adjacent well, and specific activity for ¹⁴CO₂ was determined in this NaOH via liquid scintillation.

Protein immunoblots

Total protein was isolated from FDS muscle strips incubated in KHB spiked with 5 mM d-glucose and 0 or 5 mU/ml insulin and from semitendinosus muscle snap-frozen at necropsy, and protein immunoblots were performed with antibodies previously validated for sheep (Yates et al., 2019; Posont et al., 2021). Briefly, dried muscle samples were homogenized in RIPA buffer containing manufacturer-recommended concentrations of protease and phosphatase inhibitors (Thermo Fisher), sonicated for 15 s, and centrifuged (14,000 × g, 5 min, 4 °C). Total protein concentrations were determined for each supernatant with a Pierce BCA Protein Assay Kit (Thermo Fisher), and a 30-μg protein aliquot for each sample was mixed with Bio-Rad 4× Laemmli sample buffer. This 1× solution was then heated for 5 min at 95 °C and separated by SDS–PAGE, and transferred to poly-vinylidene fluoride low-fluorescent membranes (Bio-Rad Laboratories, Hercules, CA). Membranes were incubated in Odyssey blocking buffer (LI-COR Biosciences, Lincoln, NE) for 1 h at room temperature and washed with 1× TBS-T. Membranes containing protein from FDS samples were incubated with rabbit antiserum raised against Akt (1:1,000; Cell Signaling Technologies, Danvers, MA; CAT# 9272) and phosphorylated Akt (Ser⁴⁷³) (1:2,000; Cell Signaling; CAT# 4060) at room temperature for 1 h (Akt) or at 4 °C overnight (phosphorylated Akt) as previously described for lambs (Posont et al., 2021). Membranes containing proteins from semitendinosus samples were incubated with rabbit antiserum raised against β2 adrenoceptor (1:500; Cohesion Biosciences, London, UK; CAT# CPA1027) and glucose transporter 4 (Glut4; 1:1000; Millipore Sigma) at 4 °C overnight as previously described for lambs (Yates et al., 2019). All membranes were subsequently incubated with goat anti-rabbit IR800 IgG secondary antiserum (LI-COR) at room temperature for 1 h, scanned with the Odyssey Infrared Imaging System, and analyzed with Image Studio Lite Software Ver 5.2 (LI-COR).

Statistical analysis

Data collected at necropsy including organ masses and biometrics were analyzed by ANOVA using the mixed procedure of SAS 9.4 (SAS Institute, Cary, NC) for the fixed effects of experimental group and sex. Where no significant effects of sex were observed, it was removed from the final model. Fisher’s LSD test was used for mean separation. Daily growth and blood component data were analyzed using the mixed procedure with repeated measures to analyze the effects of experimental group, day of age, and the interaction. For in vivo and ex vivo metabolic studies, study period and incubation condition were treated as the respective repeated measure, and thus the model included experimental group, period/incubation condition, and the interaction (technical replicates within each period or incubation were averaged) with lamb as the random effect. As with necropsy data, sex was treated as a fixed effect but was removed if no effect was observed. Best-fit statistics were used to select appropriate covariance structures. The main effects for birth number were not tested due to insufficient power, but birth number categories (i.e., singleton born, twin born) were represented in all groups and are reported in the methods. Because of the very low occurrence of placental anastomosis in sheep (Dain, 1971), the experimental condition of placental insufficiency was assumed to be applied to each fetus individually, and thus fetus/lambs was considered the experimental unit. For data from the pilot fetuses, all technical replications within O₂ infusion periods across the 3 d were averaged for each fetus. Likewise, those collected outside of the infusion periods were averaged for each fetus, and differences between periods were assessed by ANOVA. Significance for all analyses was indicated by a P-value of ≤ 0.05, and tendencies were indicated by P-values of ≤ 0.10. All data are presented as LS means ± standard error of the mean. Potential limitations to be considered for this study include the grouping of males and females together, the grouping of singleton-born lambs and twin-born lambs together, the assumed independent effects of placental insufficiency on twins, and the use of a single sire for all lambs.

Results

Maternal and fetal responses to O₂ infusion

Blood gas analyses performed on the first two ewes receiving O₂ supplementation showed that maternal venous blood pO₂ and oxyhemoglobin (Supplementary Figure S1A and B, respectively) exhibited polynomial increases (R² = 0.31 and 0.37, respectively) during the infusion period. Maternal venous blood pCO₂ (R² = 0.06), pH (R² = 0.00), glucose (R² = 0.14), and lactate (R² = 0.01) were minimally affected during the O₂ infusion period (Supplementary Figure S1C–F, respectively). Blood gas analyses performed for the three pilot ewes with catheterized fetuses and averaged over the 3-d period showed that fetal arterial blood pO₂ and oxyhemoglobin increased (P < 0.05) by 34% and 43%, respectively, during O₂ infusion (Supplementary Figure S2A and B, respectively). Fetal arterial blood pCO₂ did not differ due to O₂ infusion, and pH tended to be modestly reduced (P = 0.06; Supplementary Figure S2C and D, respectively). Maternal arterial blood pO₂ was increased (P < 0.05), oxyhemoglobin did not differ, pCO₂ was increased (P < 0.05), and pH was decreased (P < 0.05) during O₂ infusion in these pilot ewes (Supplementary Figure S2E–H, respectively).

Growth and biometry in lambs

An experimental group × day interaction was observed (P < 0.05) for lamb bodyweight (BW), which was less (P < 0.05) in IUGR lambs but not in IUGR+O2 lambs compared with controls on each of the 30 d between birth and necropsy (Figure 1A). Bodyweight in male lambs tended to be greater (P = 0.06) than in female lambs after 18 d of age (Supplementary Table S1). Average daily gain over this 30-d period was also less (P < 0.05) for IUGR lambs (0.253 ± 0.013 kg/d) but not for IUGR+O2 lambs (0.307 ± 0.018 kg/d) compared with controls (0.288 ± 0.017 kg/d). Average daily gain was also greater (P < 0.05) for male lambs than for female lambs (0.30 ± 0.02 vs. 0.26 ± 0.02 kg/d, respectively). Experimental group × day interactions were observed (P < 0.05) for crown circumference/BW and cannon bone length/BW but not for crown circumference, body length, body length/BW, abdominal circumference, abdominal circumference/BW, or cannon bone length. Crown circumference, body length, and abdominal circumference were less (P < 0.05) for IUGR lambs but not for IUGR+O2 lambs compared with controls, regardless of day (Supplementary Figure S3A–C, respectively). Cannon bone length was less (P < 0.05) for IUGR lambs and greater (P < 0.05) for IUGR+O2 lambs than for controls, and regardless of day (Supplementary Figure S3D). Crown circumference/BW was greater (P < 0.05) for IUGR lambs than controls on days 0, 14, and 21 and less (P < 0.05) for IUGR+O2 lambs than controls on days 0, 7, 14, and 28 (Figure 1B). Body length/BW and abdominal circumference/BW were greater (P < 0.05) for IUGR lambs but not for IUGR+O2 lambs compared with controls, regardless of day (Figure 1C and D, respectively). Cannon bone length/BW was greater (P < 0.05) for IUGR lambs than controls on days 0, 7, 21, and 28 but did not differ between IUGR+O2 lambs and controls on any day (Figure 1E). Cannon bone length was greater (P < 0.05) for male lambs than for females, but no other biometrics differed between sexes (Supplementary Table S2).

Figure 1.

Growth and body symmetry in IUGR-born lambs following maternofetal oxygenation during late gestation. Biometrics were assessed in control lambs (n = 15), IUGR lambs (n = 9), and IUGR lambs following maternal insufflation of O₂ for 8 h/d for the final 2 wk of gestation (IUGR+O2; n = 10). Data are presented for daily bodyweight (A) and for weekly crown circumference/BW (B), body length/BW (C), abdominal circumference/BW (D), and cannon bone length/BW (E). Effects of experimental group (GRP), day (DAY), and the interaction (G × D) were evaluated and are noted where significant (P < 0.05) or tending toward significance (P < 0.10). ^a,b,cMeans with different superscripts within the individual day differ (P < 0.05).

At necropsy, hindlimb mass, hindlimb mass/BW, and FDS mass were less (P < 0.05) for IUGR lambs and greater (P < 0.05) for IUGR+O2 lambs than for controls (Table 1). Hindlimb mass/body length was less (P < 0.05) and FDS/body length tended to be less (P < 0.10) for IUGR but not IUGR+O2 lambs than for controls. Heart mass was less (P < 0.05) for IUGR but not IUGR+O2 lambs compared with controls. Heart mass/BW was less (P < 0.05) for IUGR and IUGR+O2 lambs than controls. Heart mass/body length was less (P < 0.05) for IUGR lambs and greater (P < 0.05) for IUGR+O2 lambs compared with controls. Lung mass/BW did not differ between IUGR lambs and controls but was less (P < 0.05) for IUGR+O2 lambs than controls. Liver mass/BW and brain mass/BW were greater (P < 0.05) for IUGR lambs but not IUGR+O2 lambs compared with controls. Lung mass/body length and liver mass/body length did not differ among groups. Brain mass did not differ between IUGR lambs and controls but was greater (P < 0.05) for IUGR+O2 lambs than controls. Brain mass/body length was greater (P < 0.05) for IUGR but not IUGR+O2 lambs than for controls. FDS mass/BW, lung mass, lung mass/body length, liver mass, liver mass/body length, kidney mass, kidney mass/BW, and kidney mass/body length did not differ among groups. Hindlimb mass, liver mass, FDS mass/body length, and liver mass/body length were greater (P < 0.05) and lung mass/body length tended to be greater (P = 0.06) in male lambs than female lambs, but no other organ metrics differed between sexes.

Table 1.

Total and relative organ mass in IUGR-born neonatal lambs at 30 d of age following intermittent daily maternofetal oxygenation during late gestation

Organ	Experimental group			P-value
	Control	IUGR	IUGR+O2
Mass, g
Brain	77.1 ± 1.4a	77.7 ± 1.0a	84.4 ± 2.0b	0.02
Heart	114 ± 8a	72 ± 5b	103 ± 6a	<0.01
Liver	396 ± 20	395 ± 25	401 ± 18	NS
Lung	250 ± 16	228 ± 15	248 ± 12	NS
Kidney	49.9 ± 4.6	43.3 ± 5.5	47.9 ± 3.7	NS
Hindlimb	1214 ± 87a	829 ± 43b	1358 ± 37c	<0.01
FDS muscle	14.7 ± 0.8a	11.9 ± 0.6b	17.9 ± 0.8c	<0.01
Mass/BW, g/kg
Brain	6.16 ± 0.37a	7.84 ± 0.25b	6.04 ± 0.30a	<0.01
Heart	9.55 ± 0.62a	7.13 ± 0.52b	7.66 ± 0.19b	0.01
Liver	30.3 ± 0.9a	38.6 ± 1.3b	27.7 ± 1.0a	<0.01
Lung	21.2 ± 1.2a	22.8 ± 1.4a	17.2 ± 0.9b	<0.01
Kidney	4.00 ± 0.32	4.34 ± 0.60	3.42 ± 0.20	NS
Hindlimb	92.5 ± 3.5a	84.0 ± 2.7b	100.3 ± 2.8c	<0.01
FDS muscle	1.12 ± 0.04	1.17 ± 0.04	1.20 ± 0.04	NS
Mass/body length, g/cm
Brain	0.93 ± 0.02a	1.02 ± 0.02b	0.95 ± 0.04a	0.04
Heart	1.41 ± 0.08a	0.97 ± 0.06b	1.23 ± 0.08c	<0.01
Liver	4.78 ± 0.22	5.37 ± 0.33	4.63 ± 0.18	NS
Lung	3.11 ± 0.17	3.08 ± 0.20	3.10 ± 0.14	NS
Kidney	0.59 ± 0.05	0.54 ± 0.07	0.55 ± 0.04	NS
Hindlimb	14.4 ± 0.9a	9.7 ± 0.2b	15.3 ± 0.4a	<0.01
FDS muscle	0.18 ± 0.0x	0.15 ± 0.01y	0.19 ± 0.01z	0.10

^{a, b, c}Means with different superscripts differ (P < 0.05). BW, bodyweight; FDS, flexor digitorum superficialis muscle; IUGR, intrauterine growth-restricted lambs; IUGR+O2, maternal O₂-supplemented intrauterine growth-restricted lambs; NS, not significant.

Daily blood parameters in lambs

Total white blood cell and lymphocyte concentrations did not differ among experimental groups for any day (Supplementary Figure S4A and B, respectively). Monocyte concentrations tended to be greater (P = 0.08) for IUGR lambs and tended to be less (P = 0.08) for IUGR+O2 lambs compared with controls, regardless of day (Supplementary Figure S4C). An experimental group × day interaction was observed (P < 0.05) for granulocyte concentrations, which did not differ between IUGR lambs and controls but were less (P < 0.05) for IUGR+O2 lambs than controls (Supplementary Figure S4D). Red blood cell concentrations were less (P < 0.05) for IUGR lambs and greater (P < 0.05) for IUGR+O2 lambs compared with controls (Supplementary Figure S4E). Platelet concentrations were greater (P < 0.05) for IUGR lambs but not for IUGR+O2 lambs compared with controls (Supplementary Figure S4F). Red blood cell distribution width did not differ between IUGR lambs and controls but was greater (P < 0.05) for IUGR+O2 lambs than for controls (Supplementary Figure S4G). Mean corpuscular volume (32.5 ± 0.7, 33.2 ± 1.0, 32.7 ± 0.7 fl, respectively), mean platelet volume (5.06 ± 0.03, 5.04 ± 0.06, 5.07 ± 0.07 fl, respectively), and mean corpuscular hemoglobin concentration (37.0 ± 0.3, 36.9 ± 0.2, 37.3 ± 0.3 g/dl, respectively) did not differ among controls, IUGR, and IUGR+O2 lambs for any day. No differences in daily blood cell concentrations or hematological parameters were observed between sexes (Supplementary Table S3).

Experimental group × day interactions were observed (P < 0.05) for glucose, lactate, pO₂, pCO₂, oxyhemoglobin, and HCO₃. Glucose concentrations were less (P < 0.05) for IUGR and IUGR+O2 lambs than controls on day 26 and greater (P < 0.05) for IUGR+O2 lambs than controls on day 31 (Supplementary Figure S5A). Lactate concentrations were less (P < 0.05) for IUGR and IUGR+O2 lambs than controls on day 26, less (P < 0.05) for IUGR lambs and greater (P < 0.05) for IUGR+O2 lambs than controls on day 27, and greater (P < 0.05) for IUGR lambs than for controls on day 30 (Supplementary Figure S5B). Blood pO₂ did not differ between IUGR lambs and controls but was greater (P < 0.05) for IUGR+O2 lambs than controls on days 27, 29, and 31 (Supplementary Figure S5C). Blood pCO₂ was greater (P < 0.05) for IUGR lambs than controls on days 27, 28, and 29 and was greater (P < 0.05) for IUGR+O2 lambs than controls on days 27, 29, 30, and 31 (Supplementary Figure S5D). Oxyhemoglobin was less (P < 0.05) for IUGR lambs than controls on day 28 and was greater (P < 0.05) for IUGR+O2 lambs than controls on days 28 and 31 (Supplementary Figure S5E). Carboxyhemoglobin did not differ between IUGR lambs (3.6 ± 0.5%) and controls (4.2 ± 0.3%) but was less (P < 0.01) for IUGR+O2 lambs (2.4 ± 0.1%) than controls, regardless of day. Blood pH was less (P < 0.05) for IUGR and IUGR+O2 lambs than for controls, regardless of day (Supplementary Figure S5F). HCO₃ was less (P < 0.05) for IUGR and IUGR+O2 lambs than for controls on day 26 and was less (P < 0.05) for IUGR+O2 lambs than for controls on day 28 (Supplementary Figure S5G). Hemoglobin and hematocrit were less (P < 0.05) for IUGR lambs but not for IUGR+O2 lambs compared with controls, regardless of day (Supplementary Figure S5H and I, respectively). Daily blood glucose concentrations were greater (P < 0.05) for male lambs than female lambs (7.3 ± 0.2 vs. 6.7 ± 0.2 mM, respectively), but no other daily blood parameters differed between sexes.

Responses to hyperglycemia in lambs

Experimental group × study period interactions were observed (P < 0.05) for glucose, insulin, and glucose-to-insulin ratios but not for any other blood components during the GSIS study. Blood glucose concentrations were greater (P < 0.05) for IUGR and IUGR+O2 lambs than for controls during the basal period and were greater (P < 0.05) for IUGR lambs than for controls during the hyperglycemic period (Figure 2A). Plasma insulin concentrations did not differ among groups during the basal period and were less (P < 0.05) for IUGR lambs but not IUGR+O2 lambs than for controls during the hyperglycemic period (Figure 2B). Glucose-to insulin ratios were greater (P < 0.05) for IUGR and IUGR+O2 lambs than for controls during the basal period and did not differ among groups during the hyperglycemic period (Figure 2C). Blood lactate concentrations were less (P < 0.05) for IUGR and IUGR+O2 lambs than for controls, regardless of period (Figure 2D). Blood pO₂ did not differ among groups for either period (Figure 3A). Blood pCO₂ was greater (P < 0.05) for IUGR lambs but not IUGR+O2 lambs than for controls, regardless of period (Figure 3B). Oxyhemoglobin did not differ between IUGR lambs (92.8 ± 0.5%) and controls (90.3 ± 1.5%) and was greater (P < 0.05) for IUGR+O2 lambs (94.2 ± 0.2%) than for controls, regardless of period. Carboxyhemoglobin did not differ between IUGR lambs (3.74 ± 0.29%) and controls (3.53 ± 0.21%) and was less (P < 0.05) for IUGR+O2 lambs (2.32 ± 0.02%) than for controls, regardless of period. Blood pH was less (P < 0.05) for IUGR and IUGR+O2 lambs than for controls, regardless of period (Figure 3C). HCO₃ did not differ between IUGR lambs and controls and was less (P < 0.05) for IUGR+O2 lambs than for controls, regardless of period (Figure 3D). Hemoglobin and hematocrit were less (P < 0.05) for IUGR lambs (10.13 ± 0.08 g/dl and 31.0 ± 0.3%, respectively) but not for IUGR+O2 lambs (11.08 ± 0.09 g/dl and 34.0 ± 0.3%, respectively) than for controls (10.90 ± 0.15 g/dl and 34.1 ± 0.4%, respectively), regardless of period. No differences between sexes were observed for any blood components during the GSIS study (Supplementary Table S4).

Figure 2.

Glucose-stimulated insulin secretion in IUGR-born lambs following maternofetal oxygenation during late gestation. Arterial blood samples were assessed under basal and steady-state hyperglycemia at 28 d of age in control lambs (n = 12), IUGR lambs (n = 7), and IUGR lambs following maternal insufflation of O₂ for 8 h/d for the final 2 wk of gestation (IUGR+O2; n = 9). Data are presented for blood glucose concentrations (A), plasma insulin concentrations (B), glucose-to-insulin ratios (C), and blood lactate concentrations (D). Effects of experimental group (GRP), period (PER), and the interaction (G*P) were evaluated and are noted where significant (P < 0.05). ^a,b,cMeans with different superscripts within the individual period differ (P < 0.05).

Figure 3.

Blood gas responses to hyperglycemia in IUGR-born lambs following maternofetal oxygenation during late gestation. Arterial blood samples were assessed under basal and steady-state hyperglycemia at 28 d of age in control lambs (n = 12), IUGR lambs (n = 7), and IUGR lambs following maternal insufflation of O₂ for 8 h/d for the final 2 wk of gestation (IUGR+O2; n = 9). Data are presented for blood O₂ partial pressures (A), blood CO₂ partial pressures (B), blood pH (C), and blood HCO₃ concentrations (D). Effects of experimental group (GRP), period (PER), and the interaction were evaluated and are noted where significant (P < 0.05) or tending toward significance (P < 0.10).

Responses to hyperinsulinemia in lambs

Experimental group × study period interactions were observed (P < 0.05) for hindlimb glucose uptake, hindlimb glucose oxidation, hindlimb lactate secretion, and arterial glucose concentrations but not for arterial plasma insulin concentrations, glucose-to-insulin ratios, or arterial lactate concentrations. Hindlimb glucose uptake rates did not differ among groups during the basal period but were less (P < 0.05) for IUGR lambs and greater (P < 0.05) for IUGR+O2 lambs than for controls during the hyperinsulinemic period (Figure 4A). The difference in hindlimb glucose uptake rates between basal and hyperinsulinemic periods (i.e., ∆ glucose uptake) was less (P < 0.05) for IUGR but not IUGR+O2 lambs compared with controls. Hindlimb glucose oxidation rates were less (P < 0.05) for IUGR lambs and greater (P < 0.05) for IUGR+O2 lambs than for controls during both basal and hyperinsulinemic periods (Figure 4B). The difference in hindlimb glucose oxidation rates between basal and hyperinsulinemic periods (i.e., ∆ glucose oxidation) was less (P < 0.05) for IUGR but not IUGR+O2 lambs than for controls. Hindlimb lactate secretion did not differ between IUGR lambs and controls but was less (P < 0.05) for IUGR+O2 lambs than for controls during the basal period (Figure 4C). Hindlimb lactate secretion was also less (P < 0.05) for IUGR but not IUGR+O2 lambs compared with controls during the hyperinsulinemic period. The difference in hindlimb lactate secretion rates between basal and hyperinsulinemic periods (i.e., ∆ lactate secretion) was less (P < 0.05) for IUGR but not IUGR+O2 lambs than for controls. Arterial blood glucose concentrations were greater (P < 0.05) for IUGR lambs (7.4 ± 0.2 mM) and IUGR+O2 lambs (7.5 ± 0.2 mM) compared with controls (6.6 ± 0.2 mM) during the basal period and were also greater (P < 0.05) for IUGR lambs (9.6 ± 0.2 mM) and IUGR+O2 lambs (7.8 ± 0.2 mM) compared with controls (6.9 ± 0.2 mM) during the hyperinsulinemic period. Plasma insulin concentrations did not differ among IUGR lambs (2.7 ± 0.8 ng/ml at basal, 13.0 ± 4.7 ng/ml at hyperinsulinemia), IUGR+O2 lambs (4.2 ± 0.8 ng/ml at basal, 16.9 ± 3.1 ng/ml at hyperinsulinemia), and controls (4.4 ± 0.8 ng/ml at basal, 17.0 ± 2.8 ng/ml at hyperinsulinemia). Glucose-to-insulin ratios were greater (P < 0.05) for IUGR (2.37 ± 0.32) but not IUGR+O2 lambs (1.48 ± 0.34) than for controls (1.27 ± 0.26), regardless of period. Arterial blood lactate concentrations were less (P < 0.05) for IUGR lambs (0.32 ± 0.05 mM) but not for IUGR+O2 lambs (0.58 ± 0.05 mM) compared with controls (0.61 ± 0.05 mM). Arterial and venous blood glucose concentrations were greater (P < 0.05) for male lambs than for female lambs, but no other parameters differed between sexes during the HEC study (Supplementary Table S5).

Figure 4.

Hindlimb glucose metabolism in IUGR-born lambs following maternofetal oxygenation during late gestation. Arterial and venous blood samples were simultaneously assessed under basal and steady-state hyperinsulinemic-euglycemic clamp (HEC) conditions at 29 d of age in control lambs (n = 11), IUGR lambs (n = 6), and IUGR lambs following maternal insufflation of O₂ for 8 h/d for the final 2 wk of gestation (IUGR+O2; n = 9). Data are presented for hindlimb-specific glucose utilization rates (A), hindlimb-specific glucose oxidation rates (B), and hindlimb-specific lactate secretion rates (C). Effects of experimental group, period, and the interaction (G × P) were evaluated and are noted where significant (P < 0.05). ^a,b,cMeans with different superscripts within the individual period differ (P < 0.05).

Arterial blood pO₂ did not differ between IUGR lambs and controls but was greater (P < 0.05) for IUGR+O2 lambs than for controls, regardless of period (Supplementary Figure S6A). Venous blood pO₂ did not differ among groups for either period (Supplementary Figure S6B). Arteriovenous ∆pO₂ did not differ between IUGR lambs and controls but was greater (P < 0.05) for IUGR+O2 lambs than for controls, regardless of period (Supplementary Figure S6C). Arterial oxyhemoglobin was less (P < 0.05) and venous oxyhemoglobin tended to be less (P = 0.09) for IUGR but not IUGR+O2 lambs compared with controls, regardless of period (Supplementary Figure S6D and E, respectively). Arteriovenous ∆ oxyhemoglobin did not differ among groups, regardless of period (Supplementary Figure S6F). Arterial and venous blood pCO₂ were greater (P < 0.05) for IUGR and IUGR+O2 lambs than for controls, regardless of period (Supplementary Figure S6G and H, respectively). An experimental group × study period interaction was observed (P < 0.05) for arteriovenous ∆pCO₂, which was greater (P < 0.05) for IUGR lambs than controls during both basal and hyperinsulinemic periods but was greater (P < 0.05) for IUGR+O2 lambs than for controls during the basal period only (Supplementary Figure S6I). Experimental group × study period interactions were observed (P < 0.05) for arterial and venous blood pH, which were less (P < 0.05) for IUGR and IUGR+O2 lambs than for controls during both basal and hyperinsulinemic periods (Supplementary Figure S7A and B, respectively). Arterial HCO₃ did not differ among groups (Supplementary Figure S7C), but venous HCO₃ tended to be greater (P = 0.09) for IUGR and IUGR+O2 lambs than for controls, regardless of period (Supplementary Figure S7D). Arterial and venous hemoglobin and hematocrit were less (P < 0.05) for IUGR but not IUGR+O2 lambs compared with controls, regardless of period (Supplementary Figure S7E–H, respectively). None of these blood parameters differed between sexes during the HEC study.

Ex vivo skeletal muscle glucose metabolism and protein analyses

An experimental group × incubation condition interaction was observed (P < 0.05) for ex vivo glucose uptake and Akt phosphorylation but not for glucose oxidation. Ex vivo glucose uptake rates were greater (P < 0.05) for IUGR but not IUGR+O2 muscle compared with control muscle when incubated in basal media but were less (P < 0.05) for IUGR and IUGR+O2 muscle than for control muscle when incubated in insulin-spiked media (Figure 5A). Ex vivo glucose oxidation rates were less (P < 0.05) for IUGR and IUGR+O2 muscle than for control muscle but were greater (P < 0.05) for IUGR+O2 than for IUGR muscle, regardless of incubation condition (Figure 5B). Ex vivo glucose uptake and oxidation rates did not differ between sexes (Supplementary Table S6). Ex vivo Akt phosphorylation did not differ among groups when muscle was incubated in basal media but was less (P < 0.05) for IUGR and IUGR+O2 muscle compared with control muscle when incubated in insulin-spiked media (Figure 5C). Total Akt did not differ among groups or between incubation conditions. Representative micrographs for phosphorylated and total Akt, β2 adrenoceptor, and Glut4 gels are presented in Supplementary Figures S8A–C, respectively. Semitendinosus muscle protein concentrations of β2 adrenoceptor were less (P < 0.05) for IUGR and IUGR+O2 lambs than for controls (Figure 6A). Semitendinosus muscle protein concentrations of Glut4 did not differ among groups (Figures 6B). Phosphorylated Akt was less (P < 0.05) in muscle from male lambs than in muscle from female lambs (0.147 ±.022 vs. 0.226 ± 0.019, respectively), but no other muscle protein concentrations differend between sexes.

Figure 5.

Ex vivo skeletal muscle glucose metabolism in IUGR-born lambs following maternofetal oxygenation during late gestation. Intact strips of flexor digitorum superficialis muscle from control lambs (n = 15), IUGR lambs (n = 9), and IUGR lambs following maternal insufflation of O₂ for 8 h/d for the final 2 wk of gestation (IUGR+O2; n = 10) were collected at 30 d of age and incubated in basal and insulin-spiked media. Data are presented for skeletal muscle glucose uptake rates (A), glucose oxidation rates (B), and ratios of phosphorylated-to-total Akt protein (C). Effects of experimental group (GRP), media, and the interaction (G × M) were evaluated and are noted where significant (P < 0.05). ^a,bMeans with different superscripts within the individual period differ (P < 0.05).

Figure 6.

Skeletal muscle protein expression in IUGR-born lambs following maternofetal oxygenation during late gestation. Semitendinosus muscles were collected at 30 d of age from control lambs (n = 15), IUGR lambs (n = 9), and IUGR lambs following maternal insufflation of O₂ for 8 h/d for the final 2 wk of gestation (IUGR+O2; n = 10), and protein content was assessed by immunoblot. Data are presented for skeletal muscle β2 adrenoceptor content (A) and Glut4 content (B). Effects of experimental group (GRP) were evaluated and are noted where significant (P < 0.05). ^a,bMeans with different superscripts differ (P < 0.05).

Discussion

In this study, we found that intermittent maternofetal O₂ supplementation during late gestation improved subsequent growth capacity and metabolic outcomes in IUGR-born lambs. Placental insufficiency creates progressive fetal hypoxemia in late gestation that contributes to the programming of the IUGR phenotype (Limesand et al., 2013; Macko et al., 2013, 2016). When we induced hyperoxemia in ewes carrying IUGR fetuses via intratracheal O₂ infusion for 8 h/d over the final 2 wk of gestation, we recovered birthweight, neonatal growth, and body symmetry in their offspring. Deficits in skeletal muscle glucose metabolism and in GSIS observed in IUGR-born lambs at 1 mo of age were also improved by maternofetal oxygenation. The fetal adrenal medulla is highly responsive to hypoxemia (Yates et al., 2012), and the resulting chronic hypercatecholaminemia in the IUGR fetus alters programming of adrenergic tone in several tissues by reducing expression of the β2 adrenoceptor (Chen et al., 2010, 2014, 2017; Macko et al., 2016; Yates et al., 2018; Davis et al., 2020). Activation of β2 adrenergic pathways increases skeletal muscle hypertrophy (Miller et al., 2012; Gibbs et al., 2020) and insulin-stimulated glucose utilization and oxidation (Cadaret et al., 2017; Davis et al., 2020). Thus, we expected improvements in growth and metabolic outcomes following maternofetal oxygenation to coincide with restored β2 adrenergic activity, which in turn was expected to improve insulin sensitivity. However, skeletal muscle growth, glucose utilization, and glucose oxidation rates were improved without recovery of β2 adrenoceptor content. Moreover, deficient insulin-stimulated Akt phosphorylation observed in IUGR skeletal muscle, indicative of insulin resistance, was also not improved by oxygenation. Therefore, although growth capacity and metabolic function in IUGR-born neonates were improved by intermittent maternofetal oxygenation, it appears that this was independent of our presumed mechanistic target of β2 adrenergic programming. Previous studies in which growth, metabolism, and insulin responsiveness in IUGR-born lambs were improved with β2 adrenergic agonists (Yates et al., 2019; Gibbs et al., 2020) demonstrate a clear role for β2 adrenergic programming and thus make the present findings somewhat unexpected. However, these findings underscore the likelihood of multiple independent programming mechanisms for IUGR pathologies.

Maternal hyper-oxygenation effectively increased the O₂ status of the fetus, but not without consequences to the dam. When our pilot ewes with uncompromised pregnancies and catheterized fetuses were infused at 10 L/min for 8 h/d, we observed a 33% average increase in fetal blood O₂ partial pressure and a 44% average increase in O₂-bound hemoglobin. This was comparable to increases previously reported for IUGR fetuses during a single 1-h period of maternal O₂ insufflation (Macko et al., 2016). Because O₂ diffuses passively through the placenta and does not require facilitation, the increase in fetal O₂ following infusion initiation and the return to resting levels following cessation were essentially immediate. Oxygenation had little or no effect on fetal blood CO₂, glucose, lactate, or pH levels. Conversely, maternal blood pH dropped markedly during the infusion periods, which coincided with increases in CO₂ partial pressure. Brief periods of maternal hyper-oxygenation did not affect maternal blood pH or CO₂ in previous studies of uncompromised ewes or humans (Khaw et al., 2002; Tomimatsu et al., 2007) but did reduce pH and increase CO₂ in ewes that had been maintained at very high altitudes (Tomimatsu et al., 2006). Thus, maternal responses to hyper-oxygenation on may depend upon environmental conditioning, as chronic hyperthermia in the present study and altitude-induced chronic hypoxemia in the previous study almost certainly altered maternal hematology and respiration patterns. Additionally, we used 100% humidified O₂ as our infusate compared with the 40–55% O₂ mixtures used by others (Khaw et al., 2002; Tomimatsu et al., 2006, 2007). Hematology indicators in offspring were also affected by maternofetal oxygenation. Concentrations of red blood cells, hematocrit, and hemoglobin, which were reduced in unsupplemented IUGR-born lambs, were recovered by maternofetal oxygenation to or beyond control levels. Moreover, O₂ partial pressures were increased in IUGR+O2 lambs despite not being deficient in IUGR lambs and despite a paradoxical reduction in their proportional lung mass. A more robust assessment of neonatal pulmonary function would be warranted to fully understand the off-target effects of this intermittent oxygenation strategy, but the influence of O₂ on fetal lung development and growth has been demonstrated in other studies (Leslie et al., 2021; Mundo et al., 2021).

Deficits in body symmetry and growth capacity observed in our IUGR-born lambs were resolved by maternofetal oxygenation. At birth, non-oxygenated IUGR lambs were 38% lighter than controls, which was consistent with the progressive fetal growth restriction known to occur over the last trimester of pregnancy (Limesand et al., 2013; Macko et al., 2013; Yates et al., 2019). The less severe reductions in body length, crown circumference, and other skeletal growth indicators in these lambs at birth demonstrated the asymmetry of their fetal growth deficits. When placental insufficiency diminishes fetal nutrient supply, the fetus develops survival mechanisms that spare nutrients and O₂ for vital tissues such as brain and liver by limiting growth capacity of skeletal muscle and other tissues (Yates et al., 2018; Posont and Yates, 2019). These programming events become liabilities after birth, as demonstrated in the present study by persistent deficits in neonatal bodyweight and average daily gain. The tissue-specific nature of growth-limiting developmental programming also manifested in smaller hindlimbs and flexor digitorum superficialis muscles and proportionally larger brain and liver masses for our IUGR-born lambs at necropsy. Placental insufficiency creates numerous adverse conditions for the fetus, but previous evidence indicates that chronic hypoxemia is perhaps the greatest driver of the asymmetric body composition that is hallmark to the IUGR phenotype (Turan et al., 2017; Cahill et al., 2019). Intermittent targeting of intrauterine hypoxemia in our IUGR+O2 lambs during late gestation resulted in proper size and body symmetry at birth and restoration of neonatal growth capacity. Although robust assessment of β2 adrenergic signaling was not an objective of this study, we must assume that the observed reduction in β2 adrenoceptor content in IUGR skeletal muscle limited signaling activity. Because intermittent resolution of fetal hypoxemia did not improve β2 adrenoceptor content, we must also assume that improved growth outcomes were independent of this mechanism. A potential alternative mechanism for improved growth is moderation of inflammatory programming. Neonatal monocyte, granulocyte, and platelet profiles in the present study provide modest indications of heightened inflammatory tone in our IUGR lambs as well as modest improvements in IUGR+O2 lambs. However, more direct evidence of enhanced cytokine signaling pathways was observed in a previous study of IUGR fetal myoblasts and muscle tissues (Posont et al., 2018), and experimental enhancement of maternofetal inflammation during the 3^rd trimester of pregnancy reduced fetal and postnatal growth capacity in sheep (Cadaret et al., 2019; Posont et al., 2021).

Intrinsically impaired skeletal muscle glucose metabolism observed in IUGR lambs was improved in IUGR+O2 lambs. Much like fetal growth, fetal glucose utilization for oxidative metabolism is impaired by developmental programming responses to placental insufficiency that promote metabolic thrift (Limesand et al., 2007; Brown et al., 2015; Cadaret et al., 2019). Our earlier studies found that these programming responses are muscle-centric and persist after birth (Yates et al., 2019; Posont et al., 2021), and our present observations corroborate these previous findings. Intermittent maternofetal oxygenation over the last 2 wk of gestation improved subsequent hindlimb glucose utilization and oxidation rates in IUGR+O2 lambs, not only beyond the rates observed in IUGR lambs but beyond those observed in controls as well. It also mitigated about 50% of the deficit in ex vivo skeletal muscle glucose oxidation IUGR. β2 adrenergic stimulation is a potent catalyst for muscle glucose metabolism (Cadaret et al., 2017; Barnes et al., 2019; Yates et al., 2019), and thus improved muscle-specific metabolic outcomes in IUGR+O2 lambs despite no recovery of β2 adrenoceptor content were somewhat unexpected. However, a recent study indicated that preventing elevated catecholamines (i.e., the apparent impetus for adrenergic programming) in IUGR fetal sheep via adrenal demedullation did not resolve the changes in fluxes or metabolic fate of glucose near term (Davis et al., 2021). Impaired postnatal insulin signaling is a hallmark of IUGR (Dunlop et al., 2015; Xing et al., 2019), and yet the more favorable metabolic outcomes associated with maternofetal oxygenation in the present study did not appear to be due to improved skeletal muscle insulin sensitivity. Ratios of phosphorylated-to-total Akt in primary muscle incubated with insulin are commonly used to estimate insulin responsiveness (Sugita et al., 2005; Cadaret et al., 2017). In this study, insulin-stimulated Akt phosphorylation was deficient in skeletal muscle from both IUGR and IUGR+O2 lambs. In addition, although maternofetal oxygenation improved hindlimb glucose utilization only during hyperinsulinemia, glucose oxidation was improved under euinsulinemic and hyperinsulinemic conditions. Thus, in addition to being independent of β2 adrenoceptor content, the beneficial effects of maternofetal oxygenation were at least partially independent of insulin activity. Interestingly, the improvement in insulin-stimulated glucose utilization observed in vivo did not occur ex vivo. This might indicate altered responsiveness to potentiating factors such as adiponectin (Ruan and Dong, 2016) or testosterone (Pal et al., 2019), which may be present in the live lamb but absent in culture.

GSIS in the IUGR-born neonate was only modestly improved by prenatal maternofetal oxygenation. Pancreatic β-cell development and function is markedly impaired by IUGR fetal programming during late gestation (Limesand et al., 2006; Limesand et al., 2013). Moreover, experimental induction of hypoxemia reduces fetal GSIS even in the absence of other IUGR-associated pathologies (Yates et al., 2012; Benjamin et al., 2017), and greater adrenergic tone mediates much of this effect (Yates et al., 2012; Macko et al., 2016). Pancreatic islets from IUGR fetal sheep exhibit adrenergic adaptations as a product of chronic exposure to high circulating catecholamines (Limesand and Hay, 2003). In contrast to peripheral insulin sensitivity and glucose metabolism, insulin stimulus-secretion coupling in IUGR fetal sheep appears to be quite responsive to these changes in adrenergic regulation (Leos et al., 2010; Chen et al., 2014, 2017; Davis et al., 2021). We did not measure islet adrenoceptor profiles in this study but speculate that they might be comparable to profiles observed in the IUGR fetus, which would help to explain reduced GSIS in our IUGR-born lambs. In addition, if maternofetal oxygenation failed to resolve adrenergic programming in islets as it did in skeletal muscle, it could explain the modest nature of the improvement in insulin secretion relative to other metabolic outcomes. Although we speculate that inflammatory adaptations contributed to muscle-specific metabolic dysfunction, it is less likely that they play a role in the programming of islet dysfunction, as maternofetal inflammation impaired prenatal but not postnatal GSIS in IUGR lambs (Cadaret et al., 2019; Posont et al., 2021).

From these findings, we can conclude that intermittent maternofetal oxygenation during late gestation in ewes carrying IUGR fetuses improved neonatal growth and metabolic outcomes. Moderating intrinsic skeletal muscle deficits helped to restore body symmetry and rates of gain throughout the neonatal period and also helped to improve glucose metabolic homeostasis at 1 mo of age. To our surprise, maternofetal oxygenation elicited these benefits without recovering skeletal muscle β2 adrenoceptor content and without comprehensive improvements in indicators of skeletal muscle insulin responsiveness. This leaves us to presume that the benefits were mediated by the resolution of some other programming mechanism, which we speculate might be enhanced inflammatory tone based on the observed changes in circulating leukocyte profiles. Despite broad advantages for the fetus/offspring, the daily intermittent O₂ insufflation regimen did not appear to be well-tolerated by the dams. Maternal blood CO₂ was moderately elevated and pH was substantially reduced during O₂ infusion, which we speculate could be mitigated by the use of lower-percentage O₂ infusate.

Glossary

Abbreviations

BW

bodyweight

CO2

carbon dioxide

dGA

day of gestational age

EDTA

ethylenediaminetetraacetic acid

FDS

flexor digitorum superficialis

Glut4

glucose transporter 4

GRP

group effect

GSIS

glucose-stimulated insulin secretion

HCO3

bicarbonate

HEC

hyperinsulinemic/euglycemic clamp

IUGR

intrauterine growth restriction

IUGR+O2

intrauterine growth restriction + maternofetal oxygenation

KHB

Krebs–Henseleit buffer

O₂

oxygen

NS

not significantly different

pCO2

blood CO₂ partial pressure

pO2

blood O₂ partial pressure

PER

period effect

ST

semitendinosus

Funding

A subset of these data were included in the proceedings for the 2019 Western Section of the American Society of Animal Science. This manuscript is based on research that was supported in part by the USDA National Institute of Food and Agriculture Foundational Grants 2019-67015-29448 and 2020-67015-30825, the National Institute of General Medical Sciences Grant 1P20GM104320 (J. Zempleni, Director), the Nebraska Agricultural Experiment Station with funding from the Hatch Act (accession number 1009410) and Hatch Multistate Research capacity funding program (accession numbers 1011055, 1009410) through the USDA National Institute of Food and Agriculture. The Biomedical and Obesity Research Core (BORC) in the Nebraska Center for Prevention of Obesity Diseases (NPOD) receives partial support from NIH (NIGMS) COBRE IDeA award NIH 1P20GM104320. The contents of this publication are the sole responsibility of the authors and do not necessarily represent the official views of the NIH or NIGMS.

Conflict of Interest Statement

The authors declare no real or perceived conflicts of interest.