Maternal stress and undernutrition can occur together and expose the fetus to high glucocorticoid (GLC) levels during this vulnerable period. To determine the consequences of GLC exposure on fetal skeletal muscle independently of maternal food intake, groups of timed-pregnant Sprague-Dawley rats (n = 7/group) were studied: ad libitum food intake (control, CON); ad libitum food intake with 1 mg dexamethasone/l drinking water from embryonic day (ED)13 to ED21 (DEX); pair-fed (PF) to DEX from ED13 to ED21. On ED22, dams were injected with [3H]phenylalanine for measurements of fetal leg muscle and diaphragm fractional protein synthesis rates (FSR). Fetal muscles were analyzed for protein and RNA contents, [3H]phenylalanine incorporation, and MuRF1 and atrogin-1 (MAFbx) mRNA expression. Fetal liver tyrosine aminotransferase (TAT) expression was quantified to assess fetal exposure to GLCs. DEX treatment reduced maternal food intake by 13% (P < 0.001) and significantly reduced placental mass relative to CON and PF dams. Liver TAT expression was elevated only in DEX fetuses (P < 0.01). DEX muscle protein masses were 56% and 70% than those of CON (P < 0.01) and PF (P < 0.05) fetuses, respectively; PF muscles were 80% of CON (P < 0.01). Muscle FSR decreased by 35% in DEX fetuses (P < 0.001) but were not different between PF and CON. Only atrogin-1 expression was increased in DEX fetus muscles. We conclude that high maternal GLC levels and inadequate maternal food intake impair fetal skeletal muscle growth, most likely through different mechanisms. When combined, the effects of decreased maternal intake and maternal GLC intake on fetal muscle growth are additive.
- protein synthesis
- protein degradation
- fetal growth
- food intake
epidemiological and animal studies have identified that both the intrauterine and the early postnatal environments can program permanent changes in the structure and function of numerous physiological systems in offspring and that these time periods represent specific windows of development when the organism is especially vulnerable (2, 52). The fetal response to abnormal in utero environment defines the concept of “fetal programming”. Two major environmental factors that have been proposed to influence fetal programming are fetal malnutrition and fetal stress (31, 70). Exposure of the fetus to inappropriately high levels of endogenous glucocorticoids (GLCs) is thought to occur in both of these circumstances (29, 30). However, the extent to which the GLCs are responsible for the “programming” effects of fetal malnutrition and stress is still under debate. Stress and exogenous GLCs (in rodents) can alter maternal food intake, whereas maternal malnutrition alters fetal exposure to endogenous GLCs (54). Hence, it becomes difficult to delineate the specific effects of maternal malnutrition and GLCs, as they are closely linked to one another. One goal of the present experiment, therefore, was to define the relative contributions of maternal food intake vs. GLC exposure on fetal growth.
Low birth weight has been associated with changes in adult body composition, including altered fat distribution, low bone mineral content, and reduced muscle mass and strength (46). Numerous studies have shown a positive correlation between birth weight and muscle mass and strength throughout life (17, 20, 26, 42, 47–49, 57), and reduced skeletal muscle mass and strength in adult life have been linked to increased physical disability and mortality (23, 28). The contribution of GLCs to the etiology of various morbidities, such as hypertension, diabetes type 2, obesity, and ischemic heart disease has been investigated experimentally in animal models in which low birth weight has been a consistent outcome (53), and pregnant women treated with repetitive doses of antenatal GLCs give birth to offspring with reduced birth weight (4, 16). Moreover, increases in plasma cortisol levels in children and adults who were born with a low birth weight suggest that programming of the hypothalamic-pituitary-adrenal axis has occurred secondary to exposure to high GLCs at a critical developmental window. This, in turn, is associated with increased blood pressure, insulin resistance, glucose intolerance, and hyperlipidemia (32, 65).
It is thought that the programming effects of GLCs are brought about by their actions on cell division and differentiation and to permanent changes in gene expression, presumably as a result of epigenetic modifications to the genome (62, 69). However, the extent to which the observed effects of GLCs are entirely due to fetal exposure to GLCs themselves or are a result of other secondary effects of GLCs is unclear. Specifically, in rodents, GLC administration reduces maternal food intake (60, 72, 73), and in several species, maternal GLC administration compromises placental size and structure (14, 21). Both of these consequences of maternal GLC administration would limit nutrient availability to the fetus and secondarily contribute to the fetal growth retardation.
Although the “programming” effects of GLCs have been studied extensively in several tissues and organs, to our knowledge, no studies to date have assessed the impact of GLCs on fetal skeletal muscle growth. In mature muscle, GLC exposure decreases muscle protein synthesis and promotes protein degradation (27, 40, 68). In the immature muscle, we have demonstrated that protein synthesis is the primary regulator of protein accretion (10). Thus, we would predict that even minor alterations in protein synthesis rates would reduce muscle protein accretion and would be further exacerbated by increases in protein degradation.
A second goal of this experiment was to determine the extent to which fetal skeletal muscle growth is impaired when exposed to GLCs precociously, and if this occurs independently of any effects of GLCs on maternal food intake. To differentiate direct GLC effects from those attributable to altered maternal food intake, the responses of fetuses from dexamethasone (DEX)-treated dams were compared with those of pair-fed dams. Our objectives were to quantify the degree of fetal exposure to GLCs by measuring the expression in fetal liver of the GLC-inducible gene, tyrosine aminotransferase (TAT), to determine the effects on fetal muscle protein synthesis and degradation and their consequences for fetal skeletal muscle growth.
MATERIALS AND METHODS
Animals and Study Design
Timed pregnant Sprague-Dawley female rats were obtained from Harlan Laboratories on day 9 of gestation (ED9) and housed individually in wire-bottom cages in a climate-controlled room with temperature at 74°F and a 12:12-h light-dark cycle. Dams were fed a semipurified diet based on AIN-93G (Research Diets, New Brunswick, NJ). Body weight, food, and water intake were measured daily from ED9 to ED22. The study was approved by the Animal Care and Use Committee of Baylor College of Medicine and conducted in accordance with the National Research Council's Guide for the Care and Use of Laboratory Animals.
On ED12, dams were assigned to one of 3 groups: control (CON), dexamethasone-treated (DEX), and pair-fed (PF) (n = 7 dams/treatment). Each DEX dam was paired to a weight-matched PF and CON dam. To minimize the potential confounding effects of handling stress, dexamethasone was administered via the drinking water. Dexamethasone (as dexamethasone 21-acetate; Sigma Aldrich, St. Louis, MO) was dissolved in ethanol and diluted in the drinking water at a concentration of 1 mg/l (ethanol concentration 0.001%). The solution was prepared fresh daily and provided ad libitum from ED13. The DEX and CON dams were allowed to eat food ad libitum. The daily food ration for PF dams was provided in two aliquots, 12 h apart. Beginning on ED14, the food intake of each DEX dam during the previous day was measured, and the weight-matched PF dam was fed the same amount on a body weight basis. Food intake was measured from the daily change in the weight of the food jar, minus any spillage that was collected and weighed daily. CON and PF dams were provided water ad libitum.
On ED22, 20 min prior to the surgical delivery of the fetuses, each dam was injected with l-[4-3H]phenylalanine to measure protein synthesis. Two minutes prior to planned surgical delivery, the dam was anesthetized with isoflurane, the uterus was exposed through a vertical abdominal midline incision, and a blood sample was collected by cardiac puncture. The horns of the uterus were dissected open via a longitudinal incision, and each fetus with its placenta was dissected rapidly, weighed, and placed in ice. The second fetus from the tip of the right horn and the second fetus from the cervix on the left horn were designated for protein synthesis measurements and were immersed immediately in ice-cold PBS solution for 1 min and then dissected. Samples of liver, placenta, diaphragm, and the hindlimb muscles were collected and frozen in liquid nitrogen. An additional three fetuses per dam were designated for quantitative dissection; the fetuses selected were those with body weights closest to the average fetal weight of the entire litter. For the latter fetuses, the quadriceps and diaphragm were dissected quantitatively, frozen in liquid nitrogen, and weighed. The sex of the fetuses was not determined.
Protein Synthesis Measurements
Protein synthesis was measured by the “flooding dose” technique (18, 33). A bolus of l-[4-3H]phenylalanine (American Radiolabeled Chemicals, St. Louis, MO) was injected through the dam's lateral tail vein at a dose of 10 ml/kg body wt (750 μCi/kg body wt; 15 mM phenylalanine). Labeling time was taken as the time from the midpoint of injection to immersion of each fetus in the iced saline. The blood collected from each dam was acidified immediately to 0.2 M perchloric acid (PCA) and centrifuged, and the supernatant was collected for determination of [4-3H]phenylalanine specific radioactivity.
Frozen tissues for protein synthesis analysis were processed as described previously (9, 11, 18). Briefly, each sample was homogenized in water, and an aliquot was dissolved in 0.1 M NaOH for determination of the protein content using the bicinchoninic acid reagent (59). The remainder of the homogenate was acidified to 0.2 M PCA and centrifuged; the supernatant containing the tissue free amino acid pool was collected and neutralized. The PCA-insoluble precipitates were washed and assayed for total RNA, as described by Munro and Fleck (37). The remaining pellet was hydrolyzed in 6 M HCl, and the hydrolysate was resuspended in MilliQ water. The homogenates and blood supernatants were purified through an ion exchange resin (AG 50W-W8, 100–200 mesh; Bio-Rad, Hercules, CA), and resuspended in MilliQ water.
Phenylalanine in the protein hydrolysate, homogenate supernatant, and blood supernatant was isolated by anion exchange HPLC (AminoPac1; Dionex, Sunnyvale, CA), postcolumn derivatized with o-phthalaldehyde reagent and detected with an online fluorimeter. The concentration was determined by comparison with an amino acid standard (amino acid standard H; ThermoScientific, Rockford, IL). The fraction of the eluent that included the phenylalanine peak was collected, and the associated radioactivity was measured in a liquid scintillation counter (Packard Tricarb, Perkin Elmer, Waltham, MA).
The fractional rate of protein synthesis (FSR), i.e., the percentage of protein mass synthesized in a day, was calculated as FSR (%/day) = [(SP/SF) × (1,440/t)] × 100, where SP is the specific radioactivity of the protein-bound phenylalanine; SF is the muscle-free phenylalanine specific radioactivity at the time of tissue collection; t is the labeling time in minutes. Equilibration of the tracer between dam and fetus has been demonstrated previously (25, 33) and was verified by comparing the [3H]phenylalanine specific radioactivity in the dam's blood (SB) with that in the fetuses. The average dam SB /fetal SB ratio was 93.6 ± 0.6% and that of fetal SB/fetal SF was 99.9 ± 0.4%, demonstrating excellent equilibration. These values were similar among treatments and consistent with the values reported by Lewis et al. (33). Absolute synthesis rates (ASR) were calculated as the product of muscle protein mass and FSR.
Because most of the RNA in tissues is ribosomal RNA, the RNA-to-protein ratio (mg RNA/g protein) provides an estimate of ribosomal content and represents the tissue's protein synthetic capacity (Cs). The synthetic efficiency of the ribosomes (KRNA) was estimated as the total protein synthesized per total RNA [g protein/(g RNA/day)].
The mean total protein mass was calculated by multiplying the mean weight of the quantitatively dissected muscles (quadriceps and diaphragm) by the protein concentration.
Messenger RNA Expression
The expression of TAT mRNA in the fetal liver and the expression of the muscle-specific ubiquitin E3 ligases, MuRF-1, and atrogin-1 (MAFbx), in the hindlimb muscles were measured by quantitative RT-PCR with β-actin or 18S rRNA as the internal control. Total RNA from the liver and hindlimb muscles of two fetuses from each litter was extracted using a commercially available kit (RNeasy isolation kit; Qiagen, Valencia, CA). RNA concentration was determined using a Nanodrop spectrophotometer (Thermo Scientific, Wilmington, DE). RNA integrity was determined using an Experion gel electrophoresis system (Bio-Rad Laboratories, Hercules, CA) to determine the 28S:18S ratio. After ascertaining RNA quality, a 2-μg aliquot was reverse transcribed using a high-capacity RNA-to-cDNA kit, which uses random primers (Applied Biosystems, Carlsbad, CA). Quantitative PCR was performed using Taqman gene expression master mix (Applied Biosystems).
Four previously designed, assay-on-demand Taqman primer and probe sets were used (Applied Biosystems): MuRF1 (#Rn00590197_m1), atrogin-1(MAFbx) (#Rn00591730_m1), tyrosine aminotransferase (#Rn00562011_m1), and the housekeeping gene, β-actin (Rn00667869-m1). For another housekeeping gene, 18S rRNA, probes were custom made (Eurogentec, San Diego, CA). The primers were made by Sigma (forward: ACGAGACTCTGGCATGCTAACTAGT; reverse: CGCCACTTGTCCCTCTAAGAA). Each Taqman primer and probe set was validated by performing real-time PCR with a four-fold series of cDNA template dilutions to obtain standard curves of cycle threshold number (CT) against log relative concentration (data not shown). All genes were found to be amplified with equal efficiency allowing for the subsequent use of the comparative CT method (ΔΔCT) for the relative quantification of gene expression. For each gene tested, a control identical to the test assay, but omitting the RT reaction (no RT control), was included. Reactions were performed using an ABI Prism 7900 Sequence Detection System (Applied Biosystems), and data were analyzed using the 2−ΔΔCT method (34).
Data Analysis and Statistics
Minitab Statistical Software (release 13.31, State College, PA) and SPSS (version 19.0, IBM, Armonk, NY) were used to analyze the data. The data were subjected to ANOVA using the general linear model, taking into account the pairings between DEX, PF, and CON groups. When a treatment effect was identified (P < 0.05 allowing for multiple comparisons), post hoc pair-wise comparisons among treatment groups were performed using either Tukey's test for comparison among all groups or Dunnett's test for comparisons to the CON. Values are presented as least-square means with pooled standard error of the mean. For the analyses that involved fetal and placental weights, individual weights were used, taking into consideration that values from fetuses and placentas within a dam are not independent of each other. The relationship between fetal and placental weights was examined using linear regression analysis, also taking into account the interdependence of values within each dam.
Dam Weight Gain
The cumulative and daily weight gain of the dams from ED10 to ED21 is illustrated in Fig. 1A and 1B, respectively. The cumulative weight gain of DEX dams from ED13 (at the start of the dexamethasone treatment) to ED21 was only 20% of weight gained by their PF counterparts, and 14% of the weight gained by the CON dams; the PF dams gained 70% of the weight gained by CON dams (Table 1). To determine how the patterns of weight gain varied among the groups, the average daily weight gains were compared (Fig. 1B). Weight gain among treatment groups differed after ED12 (time × treatment, P < 0.001). On average, DEX dams lost weight for the 24 h after the start of the dexamethasone treatment (P < 0.001), and their daily weight gain remained lower than that of CON dams until ED22. There was a trend toward diminished weight gain in PF dams compared with CON (P < 0.08) within the first 24 h of restricted feeding (ED14–ED15). For the subsequent 3 days, weight changes were similar for PF and DEX dams and significantly less than that of CON dams (P < 0.05). However, from ED18 to ED21, the daily weight gain of DEX dams was significantly lower (P < 0.001) than PF and CON dams.
Dam Food Intake
The average daily food intakes of all the groups from ED10 to ED21 expressed in grams per day and grams per kilogram per day are summarized in Fig. 2A and 2B, respectively. Values averaged for the entire experimental period (ED13 to ED21) are summarized in Table 1.
Absolute food intake (g/day) was similar among all dams from ED10 to ED13 of pregnancy. Absolute daily food intake of the CON dams that was considered across the entire experiment remained constant throughout gestation (time effect, NS). However, there was a trend for intake to decrease toward the end of gestation; when the average intake for ED20 and ED21 was compared with the average intake up to ED19, the difference was statistically significant (P < 0.001). DEX dams reduced their food intake by 22% within 24 h of starting the treatment (ED13 to ED14), with no further change to ED19; over the last 2 days, there was a further 25% reduction. Intakes of PF dams lagged behind the DEX dams by 24 h; thus, their intakes were significantly higher than for the DEX dams on ED14. The total amount of food consumed over the 9 days of treatment was not statistically different among DEX and PF dams, and both were significantly lower than for CON dams (Table 1). The reduction in food intake between ED19 and ED21 was similar for CON, DEX, and PF dams (treatment effect, NS).
As differences in food intake are dictated, in part, by metabolic mass (which changed significantly during gestation), the data were also adjusted for body weight and are illustrated in Fig. 2B and averaged from ED13 to ED21 in Table 1. Between ED13 and ED21, DEX and PF dams consumed 87% of the amount eaten by the CON dams (P < 0.05). Although the amount of food consumed per body weight decreased significantly over time among all treatments (P < 0.001), this change was significantly different among groups (time × treatment, P < 0.001). After the initial 24-h lag, the weight-adjusted daily intakes of the PF and DEX dams were not significantly different and differed significantly from CON dams only on ED19.
Dam Water and Dexamethasone Intake
Figure 3 illustrates the dams' average daily water intake for each treatment group (expressed as ml/day) from ED10 to ED21. The average water intake changed significantly over time in all dams (P < 0.001), but with no significant difference among treatment groups. The mean oral dexamethasone intake of the DEX dams from ED13 to ED21 was 84.6 ± 4.1 μg·kg−1·day−1, and as the dexamethasone was administered via the drinking water, the variation in dexamethasone dose over time is reflected by the variation in water intake.
Placental Weights on ED 22
The mean placental weights are given in Table 1. The average values for the DEX group were significantly smaller (P < 0.001) than those of PF and CON, with no difference between the CON and PF. Fetal-to-placental weight ratios were significantly lower in DEX animals than CON (P < 0.04), while the average value for the PF animals was intermediate and not different from either CON or DEX groups.
Fetal Body Weights on ED 22
There was no difference in the average number of fetuses in CON and DEX dams, but PF had significantly fewer fetuses than CON; the difference between PF and DEX was not statistically significant (Table 1). There was a significant treatment effect on the average individual fetal weight (P < 0.001; Table 2): the CON fetuses were the largest and the DEX fetuses were the smallest (56% of CON). There was a strong tendency for body weights of the PF pups to be smaller than the CON pups (P = 0.07). The variation in DEX fetus weights can be explained, in part, by the variation in placental weight (P = 0.004 for regression of fetus vs. placenta weights), but this was not the case among CON or PF groups. However, even after adjusting for differences in placental weight, DEX fetuses were significantly lighter than CON and PF fetuses (P < 0.03). Total fetal mass (sum of all fetal weights per dam, Table 1) was significantly lower in DEX and PF dams compared with CON dams (P < 0.001), with no difference between DEX and PF. However, when the number of fetuses in each dam was included as a covariate, the DEX total fetal mass was significantly smaller than for CON and PF masses (P ≤ 0.001) with no difference between PF and CON, indicating that the lower total fetal mass of PF dams likely was due to the reduced number.
Relative mRNA Expression of TAT Activity in Fetal Liver
The mean CT values for the housekeeping gene, β-actin, were 24.0 ± 0.2, 24.1 ± 0.2, and 24.2 ± 0.2 in the CON, DEX, and PF groups, with no difference among the groups. The fetal liver TAT mRNA abundance was increased in the DEX group only (P < 0.001) with no difference between the CON and PF groups (Table 2).
Total Protein Mass in Quadriceps and Diaphragm of Fetuses
There was a significant treatment effect on quadriceps and diaphragm total protein masses (Table 2) (P < 0.01). The total protein mass of quadriceps of the DEX group and PF group was 56% (P < 0.01) and 77% (P = 0.05) of the value for the CON group, respectively; the difference between PF and DEX groups was significant (P < 0.01). For the diaphragm, the total protein mass of the DEX and PF fetuses was 56% (P < 0.001) and 82% (P < 0.03) of the diaphragm mass of the CON fetuses, respectively; values for the DEX fetuses were 69% (P < 0.03) of those in PF fetuses.
Total RNA Mass in Quadriceps and Diaphragm of Fetuses
The treatment effects on total RNA were significant for both muscles (P < 0.001; Table 2). However, the responses differed between muscles. The total RNA in the hindlimb muscles was only reduced in the DEX fetuses, whereas, for the diaphragm, the total amount was lower in both PF (P < 0.04) and DEX (P < 0.01) fetuses compared with CON. The reduction was greater for DEX than PF diaphragms (P < 0.01).
Fractional Protein Synthesis Rate of Hindlimb Muscles and Diaphragm of Fetuses
FSR for the hindlimb muscles were significantly higher than for the diaphragm (P < 0.001), and this difference was preserved across treatments (muscle × treatment, NS). In DEX fetuses, FSR of both muscles were significantly lower than for CON and PF (P < 0.001, Table 2). There was no difference in the FSR of muscles from CON and PF fetuses. For both muscles, ASR in the DEX fetuses were significantly lower than in the CON and PF fetuses (P < 0.001). There was no difference in diaphragm protein ASR between CON and PF fetuses, whereas for the hindlimb muscle, values were lower in the PF fetuses (P < 0.05).
Total RNA concentration is determined largely by ribosomal abundance and provides a measure of the maximal protein synthetic capacity of the muscle (Cs), whereas the ribosomal synthetic efficiency is reflected by protein synthetic efficiency (KRNA); these values are summarized in Table 2. For the hindlimb muscles, Cs was the same for the DEX and CON groups, but values were on average 28% higher for the PF group (P < 0.05). For the diaphragm, the Cs was 15% lower in the DEX group (P < 0.05) with no significant difference between the PF and CON groups. For the hindlimb muscles, the KRNA was significantly lower (P < 0.001) in the DEX and PF groups (P = 0.044) compared with CON, with no difference between the DEX and PF groups (P = 0.08). In the diaphragm, the KRNA was significantly lower only in the DEX group (P < 0.05) compared with the PF group with no difference between the PF and CON. There was also no difference between DEX and CON groups.
Relative mRNA Expression of Atrogin-1(MAFbx) and MuRF-1 in Hindlimb Muscles of Fetuses
The mean CT values for the housekeeping gene, 18S rRNA were 19.6, 19.6, and 19.7 in the CON, DEX, and PF groups with no difference among the groups. Atrogin-1(MAFbx) mRNA expression was significantly elevated only in the DEX group (P = 0.01) with no difference between CON and PF groups. There was no difference in MuRF-1 mRNA expression among groups (Fig. 4).
GLCs play a primary role in the maturation of the fetus (15). Maternal nutritional status is also a major factor regulating placental and fetal growth and can account for up to 60% of variation in fetal growth (60). In addition to the immediate effects on fetal growth and development, GLC levels and maternal nutrition during gestation have consequences for the postnatal health of the individual. The rodent has been used extensively to study fetal programming, and while in nonpregnant animals, GLC administration (usually dexamethasone) increases food intake, in pregnant animals, it decreases food intake (35, 71).
There is uncertainty as to whether the programming effects of GLCs observed in such animal models is a result of its direct effects or whether it is a result of the effects of decreased maternal food intake secondary to GLC administration. This issue was addressed in studies by Woods (72) and Woods and Weeks (73), in which they concluded that the effects of maternal GLC administration on the offspring may be attributable, in large part, to the associated reduction in maternal food intake. They further postulated that the programming effects of GLCs are possibly a consequence of increased fetal exposure to endogenous maternal GLC when maternal food or protein intakes are restricted during gestation. Nonetheless, some studies report that maternal and fetal GLC levels during modest protein restriction are not different from controls (13). Hence, the precise contribution of maternal GLCs vs. nutrient insufficiency on fetal development remains unclear, and this is the focus of the present study.
The optimal approach to differentiate between the effects of GLCs and reduced maternal food intake is by pair feeding, which we performed in this study. The fundamental objective in using pair-feeding is to produce an equivalent negative nutrient balance in both groups of dams. Because nutrient requirements depend on metabolic mass, it is important that the feeding take this into account. This concept is especially important in the present model because GLCs increase catabolism and promote the loss of metabolic mass independent of food intake, and the nutrient requirement to sustain a smaller metabolic mass will be lower. Thus, the daily food intake of the DEX dams, if fed to a larger PF dam, would be relatively more restrictive for the latter. To reproduce the same nutrient deficit experienced by the DEX dams in the PF dams, food intake of PF dams was established daily and adjusted for the differences in body weight. Although food intakes of DEX and PF dams from ED13 to ED21 were not different from one another and ∼25% less than that of CON dams, when adjusted for the differences in body weight, the cumulative deficit was only 13% (Table 1). The reduced intake of DEX and PF dams was limited to the days at the start of the treatment as differences in weight-adjusted intakes among all three treatments were minimal from ED16 (Fig. 2B). Despite the similar degree of food restriction, the DEX dams gained only 20% of the weight gained by PF dams, which, in turn, gained 70% of the weight gained by CON dams. Thus, the decrease in food intake of DEX dams accounted for only 35% of their decreased weight gain and occurred soon after the start of DEX administration, with the remainder attributable to independent effects of DEX, especially toward the end of gestation. At this stage, the PF dams were able to gain weight on the identical food intake of DEX dams, suggesting that the dexamethasone alone was substantially compromising the use of dietary nutrients for anabolic processes (either maternal or fetal).
The effect of global maternal nutrient restriction on fetal growth is dependent on the degree of food restriction. We saw a marginal effect (∼11%) on ED22 fetal body weights with a 13% reduction in food intake; this effect is consistent with findings from other studies where the restriction of maternal intake to 50% or more (7, 12, 56) of ad libitum food intake reduced birth weight by 15–25% of normal. Our observation that litter size was smaller for the PF group compared with the CON (P = 0.02) and DEX groups contrasts with other studies that have demonstrated no reduction in litter size even with a 50% reduction in food intake (7). It is unlikely that the reduction in litter size in the PF group was secondary to the food restriction because the intervention was performed during the last half of gestation and we found no evidence of fetal resorption. This observation, therefore, most likely was a random effect due to the selection of dams for the PF group and not a treatment effect.
The DEX fetuses weighed 40% less than the CON and PF fetuses with no statistical difference between CON and PF groups, although a trend was present. Although maternal GLC administration has been shown to decrease birth weight across all species (14), for reasons that are unclear, our study appears to have demonstrated the greatest impact on fetal growth among the numerous other studies reviewed, where similar dosing regimens were used (3, 22, 38, 39, 58, 63, 64, 66, 73, 75). The biological effects of GLCs depend not only on the dose, the length of time for which they are administered, and their bioavailability, but also on the route of administration. The effect of dexamethasone on fetal growth appears to be more pronounced when it is given orally (25–30% reduction in birth weight) (22, 58, 66, 74) vs. parenterally (∼10% reduction) (38, 39, 63, 64). The oral dose of dexamethasone we aimed to achieve in this study was 100 μg·kg−1·day−1 by providing it in the drinking water at 1 μg/ml, and the actual measured dose was fairly close (84.6 ± 4.1 μg·kg−1·day−1). The results of our study would suggest that the bioavailability of dexamethasone is higher when given orally.
We measured the expression in the fetal liver of TAT, a GLC-inducible gene as an index of fetal GLC exposure. TAT is a well-studied enzyme that reflects a prototype response in gene-mediated steroid induction (43) and has been used as a marker for the biological activity of GLC in other studies (6). Thus, TAT mRNA expression provides a way to compare the GLC exposure and its biological activity in all of the three treatment groups, even though the GLC species to which they were exposed differed in DEX vs. the PF and CON fetuses (endogenous corticosterone). Moreover, the level of TAT expression in the PF fetuses would ascertain whether the maternal food restriction increased exposure of the fetus to endogenous GLCs (30). We found increased TAT mRNA expression in the livers of DEX fetuses, but there was no difference between PF and CON groups. This suggests that the degree of food restriction in PF dams was not sufficient to increase endogenous GLCs in utero to levels that induce fetal TAT mRNA expression.
Even though the DEX fetuses were exposed to high levels of GLCs, the extent to which their growth impairment can be ascribed to a primary effect of GLCs on the fetus remains uncertain. This uncertainty arises from the observation that placental size was seriously compromised in the DEX dams, and this would limit the exchange of nutrients and metabolites between mother and fetus (21). Nonetheless, even after adjusting for the smaller placentas, there remained an independent effect of GLCs on fetal weight. Thus, from our results, we propose that maternal dexamethasone treatment during the second half of gestation affects fetal growth indirectly by reducing maternal food intake and decreasing placental size, and directly by antagonizing anabolic processes responsible for fetal growth.
There is paucity of data on the effects of prenatal GLCs on fetal skeletal muscle growth. Thus, the second main objective of this study was to determine the consequences of fetal GLC exposure on skeletal muscle mass and the extent to which these are secondary to the GLC effects on maternal nutrient intake. As the degree of food restriction was similar between DEX and PF dams, we assume that the effects of the decrease in maternal food intake on muscle growth in DEX fetuses is represented by the deficit in muscle mass of the PF fetuses. Thus, in DEX fetuses, 50% of the reduction in total protein mass in both diaphragm and quadriceps muscles could be attributed to the effects of decreased maternal food intake. The fact that the muscle protein masses in the PF group were smaller in spite of only a minor effect on birth weight indicates that fetal tissues and organs are not equally affected by maternal diet and that birth weight is not an accurate indicator of fetal skeletal muscle protein accretion. Moreover, given the temporal pattern of food restriction, the smaller muscle protein masses of PF fetuses were likely the result of the restriction in maternal nutrient intake at the start of the treatments, i.e., at midgestation.
Measurements of muscle protein metabolism were performed to identify the mechanisms responsible for the blunted muscle growth. Total protein mass reflects a balance between protein synthesis and degradation rates, and a decrease in total protein mass could be the result of decreased protein synthesis, increased protein degradation, or a combination of both. The protein masses of both quadriceps and diaphragm muscles of PF fetuses were lower than for the CON group, even though their FSRs were similar. In the PF hindlimb muscles, the attainment of normal FSR values was enabled by a higher CS (with a tendency for KRNA to fall), which resulted from the absence of any change in total RNA content despite the lower protein mass. This did not occur in the diaphragm where the reduction in total protein and RNA content were proportional and CS remained unaltered. There are a number of possible explanations for the smaller protein masses of the muscles of PF fetuses. One possibility is that protein degradation rates were higher than in the CON fetuses. Johnson et al. (25) observed in fetuses from dams that were food deprived from ED18 to ED22 that the reduction in fetal skeletal muscle protein mass could be attributed largely to increased protein degradation. In the mature muscle, the ubiquitin proteosome system (UPS) and autophagy are the major pathways that facilitate muscle protein breakdown (8). The canonical functions of atrogin-1(MAFbx) and MuRF-1 are as muscle-specific E3 ubiquitin ligases that target proteins to the UPS for degradation in conditions that promote muscle protein catabolism. In adult animal models atrogin-1(MAFbx) is upregulated by nutrient restriction (8, 19). Hence, we measured the mRNA expression of atrogin-1(MAFbx) and MuRF-1 as evidence of increased UPS activity in the PF fetuses; our results did not support this possibility. Another alternative is that the UPS system may not be a major pathway for protein degradation in the immature muscle. Autophagy is rapidly activated upon fasting, especially in neonatal tissues, including the skeletal muscle (51). However, whether autophagy is responsible for increased muscle protein breakdown when nutrients are limiting in utero as observed by Johnson and colleagues (24, 25) is unknown. An additional possibility is that FSR and/or UPS activity in the PF group were altered only during the time when the dams were effectively restricted compared with CON dams, i.e., between approximately ED14 and ED17 (Figs. 1B and 2B). Subsequently, when the dams' metabolic mass was reduced to a degree that could be supported by the smaller nutrient supply (so that they effectively were not food-restricted), the muscle protein FSR in the fetus normalized. Measurement of FSR at earlier time points would identify such a possibility.
As observed for adult muscle (27, 40, 50), muscle protein FSR for the DEX fetuses was significantly lower than for CON fetuses. Given the significant role of the protein synthesis in the regulation of protein deposition in the immature muscle (10), it is likely that this difference was a dominant factor in mediating the impaired growth of the muscle in DEX fetuses. The comparison with the PF suggests that the response is specifically attributable to the dexamethasone rather than a decrease in food intake. In both the hindlimb and diaphragm of DEX fetuses, the reduction in synthesis rate appears to be due primarily to a reduction in translational efficiency, KRNA, and is consistent with studies in adult muscle that demonstrate that dexamethasone inhibits translation, primarily through repression of mTOR signaling (55, 67). In addition, in the diaphragm of DEX pups, total RNA was reduced to a greater extent than protein content, and this decrease in CS added to the reduction in KRNA to inhibit protein synthesis. In the hindlimb, protein and RNA were affected equally, so that Cs did not contribute to the reduced FSR.
The contribution of protein degradation to the reduced size of DEX fetal muscles is uncertain. Although atrogin-1 (MAFbx) expression was increased, studies in mature mouse muscle (1) suggest that even though this E3 ligase is upregulated by GLCs, it may not contribute to GLC-mediated muscle atrophy in vivo. The absence of a response in MuRF-1 expression was unexpected and contrasts with studies done in cell cultures, as well as in mature muscle, where both atrogin-1(MAFbx) and MuRF-1 have been found to be upregulated by GLCs (1, 5, 36, 44, 61). The reason for our divergent observation is unclear: the upregulation of MuRF-1 expression may be dose-dependent, and it may respond only at doses higher than the one employed by us but that are normally employed in atrophy models. Additionally, the expression and responsiveness of MuRF-1 expression to GLCs could be developmentally regulated as in the heart and be less responsive to GLCs in fetal muscle (41). GLCs also stimulate autophagy (45, 50) that may have contributed additionally to increased protein breakdown. A further possibility is that, like the PF fetuses, stimulation of protein degradation, regardless of mechanism, was only transient and occurred earlier in gestation soon after the start of treatment; with prolonged GLC treatment, the effect on protein degradation gradually waned. This transient response, together with the observation that suppression of protein synthesis by chronic GLC administration is sustained, was carefully documented by Odedra et al. (40) and Kayali et al. (27). Thus, by performing measurements only at a single time point, we may not have captured all the potential mechanisms that ultimately contributed to the diminished muscle size. Taken together, our findings support the hypothesis that exposure of the fetus to GLCs impairs skeletal muscle growth, and this effect is independent of the effects of maternal nutrient food intake. Nonetheless, because of the separate effects on muscle mass observed for the PF group, we must infer that when the two insults are combined, the consequence for muscle growth is exacerbated.
Perspectives and Significance
In summary, we conclude that fetal exposure to GLCs reduces fetal growth independent of its effects on maternal food intake. Our data also indicate that a modest reduction in maternal nutrient intake (13%) can impair muscle protein accretion, a change that is not necessarily reflected by the fetal body weights but consistent with our past observations that skeletal muscle protein anabolism in the immature muscle is highly sensitive to nutrient supply. The inhibitory effects of reduced maternal food intake on skeletal muscle growth were additive when combined with the inhibitory effects of GLCs. We do not know the extent to which the reduction in DEX fetal growth was the consequence of reduced placental function and, thus, also the product of an inadequate supply of substrates for anabolic processes. Nonetheless, increased expression of TAT mRNA in the DEX fetus livers indicates that the fetuses were exposed of to high GLC levels. An unforeseen outcome of this study was the severity of the growth retardation in the DEX fetuses, even though the dose that we administered was one that is commonly used to study the effects of GLCs on fetal programming, and significantly lower than the dose used for studying GLC-induced muscle atrophy. Thus, one must be cautious in extrapolating these results to instances where the fetal exposure to GLC may not be as extreme, such as with severe maternal undernutrition or stress, or when GLCs are used therapeutically in pregnancy. These data illustrate that an elevation of GLCs during pregnancy can influence fetal growth both indirectly by altering maternal food consumption and placental function, and also by their direct effect on fetal muscle protein metabolism. Moreover, this finding highlights the necessity to evaluate more than body weight in the assessments of fetal outcomes. Although the reduction in muscle mass undoubtedly has adverse short-term functional consequences for the survival of the newborn, whether the muscle can recover over the long term from this prenatal insult will likely depend on the effects of high GLC exposure on muscle differentiation and maturation that occurs during the second half of gestation.
These studies were supported by grants from the National Institutes of Health NIAMS, AR46308 and USDA/ARS/CSREES, 6250-51000-043 (M.L.F.).
No conflicts of interest, financial or otherwise, are declared by the authors.
Author contributions: G.G. and M.L.F. conception and design of research; G.G., I.J.E., H.A.S., and M.L.F. performed experiments; G.G., I.J.E., H.A.S., and M.L.F. analyzed data; G.G. and M.L.F. interpreted results of experiments; G.G. prepared figures; G.G. and M.L.F. drafted manuscript; G.G., I.J.E., H.A.S., and M.L.F. edited and revised manuscript; G.G., I.J.E., H.A.S., and M.L.F. approved final version of manuscript.
The authors thank E. O'Brian Smith for assistance with statistical analyses, Adam Gillum for graphics, and Jerome Stubblefield for assistance with animal husbandry. The authors would also like to thank Dr. Steve Welty (Section of Neonatology, Department of Pediatrics, Baylor College of Medicine) and Dr. Teresa Davis (USDA/ARS Children's Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine) for their input and critical reading of the manuscript.
This work is a publication of the USDA, Agricultural Research Service Children's Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine, Houston, TX, USA. The contents of this publication do not necessarily reflect the views or politics of the USDA, nor does the mention of trade names, commercial products, or organizations imply endorsement by the U.S. government.