1st Page

Journal of Animal Science - Rapid Communication

Poor maternal nutrition during gestation alters the expression of genes involved in muscle development and metabolism in lambs1


This article in JAS

  1. Vol. 94 No. 7, p. 3093-3099
    unlockOPEN ACCESS
    Received: Apr 22, 2016
    Accepted: May 10, 2016
    Published: June 15, 2016

    2 Corresponding author(s):

  1. M. L. Hoffman 2*,
  2. K. N. Peck*,
  3. J. L. Wegrzyn,
  4. S. A. Reed*,
  5. S. A. Zinn* and
  6. K. E. Govoni*
  1. * Department of Animal Science
     Center for Genome Innovation, Institute for Systems Genomics, University of Connecticut, Storrs 06269


Poor maternal nutrition during gestation can result in reduced muscle mass and increased adiposity of the muscle tissue in the offspring. This can have long-lasting consequences on offspring health and productivity. However, the mechanisms by which poor maternal nutrition affects postnatal muscle development are poorly understood. We hypothesized that poor maternal nutrition during gestation would alter expression of key pathways and genes involved in growth, development, and maintenance of the muscle of lambs. For this study, beginning at d 31 ± 1.3 of gestation, ewes were fed 100 (control), 60 (restricted), or 140% (overfed) of the NRC requirements. Within 24 h of birth, lambs were necropsied and semitendinosus muscle tissue was collected for gene expression analysis. Using RNA sequencing (RNA-seq) across dietary treatment groups, 35 and 10 differentially expressed genes were identified using the Bos taurus and Ovis aries reference annotations, respectively. Maternal overfeeding caused changes in the expression of genes involved in regulating muscle protein synthesis and growth as well as metabolism. Alternately, maternal nutrient restriction affected genes that are involved in muscle cell proliferation and signal transduction. That is, despite a similar phenotype, the genes identified differed between offspring born to restricted- or overfed, ewes indicating that the mechanism for the phenotypic changes in muscle are due to different mechanisms.


Poor maternal nutrition, as a result of under- or overfeeding of the mother during pregnancy, can negatively impact offspring development including growth, tissue accretion, and organogenesis (Barker, 2007; Hoffman et al., 2014, 2016). Skeletal muscle tissue is particularly vulnerable as nutrients are partitioned away from muscle in favor of organ development (Wu et al., 2006), resulting in alterations in muscle mass, muscle fiber type, connective tissue content, and adiposity (Daniel et al., 2007; Yan et al., 2013; Reed et al., 2014). As a result, these changes impact meat and carcass quality and carcass yields. The effects of poor maternal nutrition on offspring development are well established; however, the mechanisms involved remain poorly understood, especially those involved with changes in global gene expression. It has been determined that overfeeding ewes during gestation causes a reduction in gene expression of several myogenic factors and Wnt signaling factors as well as increased expression of genes involved in connective tissue development and inflammation in the muscle (Tong et al., 2009; Huang et al., 2011). However, these analyses are limited to only a few genes. Furthermore, there is limited information on genes and pathways that may contribute to altered postnatal muscle development in the offspring. Therefore, we hypothesized that poor maternal nutrition would alter key pathways and global gene expression of factors involved in the growth and development of the muscle tissue of lambs.



All procedures were approved by the University of Connecticut Institutional Animal Care and Use Committee. A detailed description of experimental design, animals, and diets used was previously reported (Reed et al., 2014; Hoffman et al., 2016). Briefly, pregnant ewes were assigned to 1 of 3 treatment diets (100, 60, and 140% of the NRC [1992] requirements for total digestible nutrients) at d 31 ± 1.3 of gestation and remained on that diet until parturition. Eighteen lambs (n = 6 per treatment) were sampled within 24 h of birth. For RNA sequencing, samples from 3 Dorset lambs per treatment group were used (2 rams and 1 ewe born to control-fed ewes [CON], 3 rams born to restricted-fed ewes [RES], and 2 ewes and 1 ram born to overfed ewes [OVER]).

Sample Collection and RNA Preparation

Euthanasia, semitendinosus muscle sample collection, and RNA isolation are described by Reed et al. (2014). Ribosomal RNA was removed using a human and mouse Ribo-Zero Gold kit (Epicentre, Madison, WI). Complementary DNA libraries were prepared and sequenced using the Ion Torrent Proton sequencer (Li et al., 2015).

Data Analysis

Quality control (QC) was performed using Sickle (Joshi and Fass, 2011) to eliminate sequences that were ≤35 bp in length and had a Phred score ≤20. Sequences were mapped to the Ovis aries (Oar_V.3.1/OviAri3;; Accessed 5/13/14) and the Bos taurus (Btau_4.61/BosTau7;; Accessed 5/13/14) reference genomes using the TopHat aligner (Trapnell et al., 2009). Differential gene expression was then determined using Cufflinks, Cuffmerge, and Cuffdiff packages (Trapnell et al., 2010). Genes were considered to be differentially expressed when P ≤ 0.00005 and the corresponding false discovery rate–corrected q-value was ≤0.05. False discovery rates were determined using the Benjamini–Hochberg multiple testing correction (Benjamini and Hochberg, 1995). Functional classifications, gene ontology, and gene enrichment analysis were performed on differentially expressed genes for CON vs. OVER and CON vs. RES comparisons that were identified with the Bos taurus reference via the PANTHER classification system (Mi et al., 2013) using default parameters.


Differential Gene Expression

Gene names corresponding to abbreviations are listed in Table 1. Using the Ovis aries reference annotation, 10 unique differentially expressed genes were identified (Table 1). Alternately, using the Bos taurus reference, 35 differentially expressed genes were identified (Table 1). The differentially expressed genes FOS, MSTN, MTE1, PSPH, RASD1, and SLC25A33 were identified using both reference annotations with similar changes in gene expression (Table 1). The majority of differentially expressed genes were identified comparing expression between CON and OVER lambs and between RES and OVER lambs. Only 4 genes were upregulated whereas 15 were downregulated in CON vs. OVER lambs (Table 1). Specifically, the ARRDC2, ARID5B, MSTN, and MYF6 genes were 3.46-, 2.87-, 1.96-, and 1.67-fold greater, respectively, in OVER lambs relative to CON lambs (Table 1; q ≤ 0.05). Expression of small RNA genes SNORD70, SNORD133, and U1 were reduced 2.05-, 2.55-, and 2.77-fold, respectively, in OVER lambs compared with CON lambs (q ≤ 0.05). Lambs from overfed ewes also exhibited reduced expression (≥2-fold) of ANKRD1, JUNB, MTE1, PPARGC1A, PSAT, PSPH, RGS16, and THBD genes relative to CON lambs (q ≤ 0.05; Table 1). Four genes, TRIM63, FOS, SNORD113, and suppressor of cytokine signal 3 (SOCS3), exhibited a 1.41-, 1.59-, 2.68-, and 2.83-fold reduction, respectively, in RES lambs compared with CON lambs (q ≤ 0.03). No genes were found to be upregulated in response to underfeeding during gestation compared with CON lambs. Fifteen genes were upregulated and 5 genes downregulated in RES vs. OVER lambs (Table 1). Expression of PSPH, ANKRD1, SNORD113, and U1 were reduced 2.04-, 2.11-, 2.40-, and 3.47-fold, respectively, compared with OVER lambs (Table 1; q ≤ 0.05). In RES lambs, Y_RNA expression was 8.47-fold less than in OVER lambs (Table 1; q ≤ 0.05). In contrast to other small nuclear RNA, expression of SNORA72 was 7.46-fold greater in RES lambs than in OVER lambs. Expression of MYF6, RASD1, SLC25A33, and ZFAND5 were increased 1.72-, 1.88-, 2.76-, 2.43-, and 2.03-fold, respectively, in RES lambs compared with OVER lambs (q ≤ 0.05; Table 1).

View Full Table | Close Full ViewTable 1.

Differential gene expression treatment comparisons1

Fold change2
Gene name Description Bovine Ovine
CON vs. OVER lambs
    Arrestin domain containing 2 (ARRDC2) Protein coding gene 3.46
    AT rich interactive domain 5B (ARID5B) Adipogenesis and liver development 2.87
    Myostatin (MSTN) Regulator of skeletal muscle growth 1.96 1.95
    Myogenic factor 6 (MYF6) Muscle differentiation 1.67
    BTG family, member 2 gene (BTG2) Antiproliferative protein −1.48
    Glutamine-Fructose-6-Phosphate Transaminase 2 (GFPT2) Regulates hexosamine pathway −1.84
    Jun B proto-oncogene (JUNB) Regulates gene expression −2.02
    Small Nucleolar RNA 70 (SNORA70) RNA gene −2.05
    Thrombomodulin (THBD) Atrophy of skeletal and cardiac muscle −2.07
    Regulator of G-protein signaling 16 (RGS16) Inhibits signal transduction −2.11
    Metallothionein 1E (MT1E) Bind heavy metals −2.20 −3.16
    PPARγ coactivator 1 alpha (PPARGC1A) Coactivator for nuclear receptors −2.31
    Phosphoserine aminotransferase 1 (PSAT1) Phosphoserine aminotransferase −2.50
    Small Nucleolar RNA 113 (SNORD113) Cytospin-A family −2.55
    Phosphoserine Phosphatase (PSPH) Biosynthesis of serine −2.70 −2.54
    U1 spliceosomal RNA (U1) Novel miscellaneous RNA −2.77
    Ankryin repeat domain 1 (ANKRD1) Myofibrillar stretch-sensor system −3.10
RES vs. OVER lambs
    Small Nucleolar RNA 72 (SNORA72) RNA gene 7.46
    Arrest7in domain containing 2 (ARRDC2) Protein coding gene 3.85
    F-Box Protein 32 (FBXO32) Muscle atrophy 3.72
    Tripartit motif containing 63 (TRIM63) Small nuclear RNA 3.47
    6-Phosphofructo-2-Kinase/Fructose-2,6-Biphosphatase 3 (PFKFB3) Cell cycle progression 3.24
    RAS, Dexamethasone-Induced 1 (RASD1) GTPase 3.13 2.76
    Sestrin 1 (SESN1) Dendritic cell endocytosis 3.07
    Yippee-Like 3 (YPEL3) Cellular proliferation and apoptosis 2.88
    Kelch-Like Family Member 38 (KLHL38) Protein coding gene 2.73
    Solute carrier family 25, member 33 (SLC25A33) RNA gene 2.57 2.43
    Growth Arrest and DNA-Damage-Inducible, Gamma (GADD45G) Growth and apoptosis 2.56
    Zinc finger, AN1-type domain 5 (ZFAND5) Protein degradation 2.03
    BTG family, member 1 (BTG1) Cell growth and differentiation 1.94
    LOC443255 Hormone signaling 1.88
    Myogenic factor 6 (herculin; MYF6) Muscle differentiation 1.72
    MGC165715 Protein coding gene 1.65
    KIAA0408 Protein coding gene 1.62
    DNA damage inducible transcript 4-ike (DDIT4L) Inhibits cell growth 1.48
    Phosphoserine Phosphatase (PSPH) Biosynthesis of serine −2.04
    Ankryin repeat domain 1 (ANKRD1) Myofibrillar stretch-sensor system −2.11
    Small Nucleolar RNA 113 (SNORD113) Cytospin-A family −2.40
    U1 spliceosomal RNA (U1) Novel miscellaneous RNA −3.47
    Y_RNA Proliferation and apoptosis −8.47
1Lambs born to control-fed ewes (100% of the NRC (1992) requirements; 2 rams and 1 ewe born to control-fed ewes [CON]), born to restricted-fed ewes (60% of the NRC (1992) requirements; 3 rams born to restricted-fed ewes [RES]), and lambs born to overfed ewes (140% of the NRC (1992) requirements; 2 ewes and 1 ram born to overfed ewes [OVER]).
2Data are presented as relative fold change from CON.

Gene Ontology and Enrichment Analysis

For gene classifications, 42.9 and 45.5% of differentially expressed genes in CON vs. OVER lambs were classified into the subcategories of binding (gene ontology [GO]: 0005488) and cell part (GO: 0044464) for the main categories of molecular function and cellular processes (Table 2). Additionally, for the main classification of biological processes, 26.70% of genes identified were categorized into the subcategories of cellular processes (GO: 0009987) and metabolic processes (GO: 0008152). In CON vs. RES lambs, 33.30% of differentially expressed genes identified in the category of molecular function belonged to the subcategory of binding (GO: 0005488) and catalytic activity (GO: 0003824). For the main category of biological processes, 33% of differentially expressed genes were also classified into the metabolic processes subcategory (GO: 0008152). For protein classifications, CON vs. OVER lamb differentially expressed genes were predominantly subcategorized into the transcription factor group (PC00218; 40%). Differentially expressed genes for the CON vs. RES lamb comparison were evenly distributed across all protein classification subcategories identified (Table 2). Gene enrichment analysis did not identify any genes or subsequent classification categories that were affected by maternal diet (q > 0.05).

View Full Table | Close Full ViewTable 2.

Gene classification and ontology1

Classification CON vs. OVER CON vs. RES
Molecular function
    Binding (GO3:0005488) 42.9% 33.3%
    Catalytic activity (GO:0003824) 21.4% 33.3%
    Enzyme regulator activity (GO:0030234) 7.1% 16.7%
    Nucleic acid binding transcription factor activity (GO:0001071) 21.4% 16.7%
    Protein binding transcription factor activity (GO:0000988) 7.1%
Biological processes
    Apoptotic process (GO:0006915) 3.3%
    Biological regulation (GO:0065007) 20.0% 20.0%
    Cellular component organization or biogenesis (GO:0071840) 6.7%
    Cellular process (GO:0009987) 26.7% 20.0%
    Developmental process (GO:0032502) 10.0% 20.0%
    Localization (GO:0051179) 10.0%
    Metabolic process (GO:0008152) 26.7% 30.0%
    Multicellular organismal process (GO:0032501) 3.3%
    Response to stimulus (GO:0050896) 3.3%
Cellular component
    Cell part (GO:0044464) 45.5%
    Extracellular region (GO:0005576) 9.1%
    Membrane (GO:0016020) 9.1%
    Organelle (GO:0043226) 36.4%
Protein class
    Enzyme modulator (PC00095) 10.0% 20.0%
    Ligase (PC00142) 20.0%
    Nucleic acid binding (PC00171) 20.0% 20.0%
    Signaling molecule (PC00207) 10.0% 20.0%
    Transcription factor (PC00218) 40.0% 20.0%
    Transferase (PC00220) 20.0%
1Differentially expressed genes were classified into 1 of 3 major categories and correlating gene ontology classifications. Genes were also classified by protein class. Classification was performed using the PANTHER classification system (Mi et al., 2013). Percentages were determined by the number of input genes that corresponded with a given classification and corresponding ontologies. Ontology subclassifications are listed in alphabetical order.
2CON = 2 rams and 1 ewe born to control-fed ewes; OVER = 2 ewes and 1 ram born to overfed ewes; RES = 3 rams born to restricted-fed ewes.
3GO = gene ontology.


Based on the RNA sequencing (RNaseq) analysis, the most important and novel aspect of these data is that the mechanisms mediating changes in the muscle tissue of RES and OVER offspring are fundamentally different despite similar phenotypic changes observed in the muscle tissue (as presented in Reed et al. [2014]). Identifying this difference is key, as the potential for the involvement of these differential mechanisms needs to be taken into consideration when developing future studies as well as potential intervention strategies.

Differential expression of several genes involved in the regulation of muscle metabolism, hypertrophy, nutrient uptake, and protein turnover were identified including MSTN, ANKRD1, JUNB, ARID5B, RGS16, and GFPT2. Myostatin is responsible for suppressing myogensis by inhibiting the activation of MYOD and increasing protein degradation in muscle (Argilés et al., 2012). However, no change in MSTN protein expression was identified (Reed et al., 2014). This disparity between mRNA and protein expression could be due to the prevalence of specific microRNA that regulate MSTN gene expression (Hitachi and Tsuchida, 2014). The prevalence of microRNA in muscle can be affected by nutritional status (Yan et al., 2013); however, the prevalence of MSTN microRNA within muscle tissue of lambs born to poorly nourished mothers needs to be determined.

The ANKRD1 gene is a target for several myogenic factors (i.e., MyoD and myogenin) and is involved in the differentiation of muscle tissue, muscle fiber–type switch, and muscle hypertrophy (Blais et al., 2005; Arimura et al., 2009). This gene was downregulated in OVER lambs, and similar to our findings, Peñagaricano et al. (2014) reported that ANKRD1 was downregulated in the muscle of lambs born to ewes fed a corn-based diet compared with lambs from ewes fed a distillers’- or haylage-based diet. Although these diets fed were isoenergetic, the corn-based diet contained 30% less protein than the other diets (Peñagaricano et al., 2014), suggesting that maternal protein restriction and overfeeding have similar effects on ANKRD1 expression. Similar to ANKRD1, the transcription factor JUNB, which is responsible for maintaining muscle mass and inducing muscle hypertrophy (Raffaello et al., 2010), was reduced in OVER lambs, potentially leading to reduced muscle fiber cross-sectional area (Reed et al., 2014). Therefore, a reduction in the expression of these genes could affect postnatal muscle hypertrophy and response to physical loading in these animals.

Expression of several genes involved in regulating muscle nutrient uptake (e.g., RGS16, GFPT2, ARID5B, PPARCG1A) was also affected by poor maternal nutrition. The RGS16 gene encodes a G protein regulator that controls fatty acid oxidation in response to glucose production (Rudkowska et al., 2013), and GFPT2 is responsible for controlling flux of glucose into the hexoseamine biosynthetic pathway (Zhang et al., 2004). Reduced expression may lead to alterations in nutrient partitioning in the muscle tissue. The ARID5B gene is required for the accumulation of triglycerides in adipose tissue and is involved in the differentiation of smooth muscle and adipose cells (Whitson et al., 2003; Wang et al., 2012). Similar to our findings, Peñagaricano et al. (2014) observed that ARID5B gene expression was greater in the subcutaneous adipose tissue of fetuses from ewes fed distillers’ grain compared with those fed corn or haylage, demonstrating that maternal diet can alter the expression of this gene. In general, changes to these factors suggest that the regulation of nutrient accretion and utilization is altered within the muscle tissue of OVER offspring, and therefore, this may be a mechanism contributing to altered postnatal muscle development.

Expression of PPARGC1A can vary depending on muscle fiber composition and mitochondrial function (Handschin, 2010). Specifically, Type I and Type IIa muscle fibers are classified as oxidative fibers and exhibit greater expression of PPARGC1A than glycolytic Type IIb fibers (Handschin, 2010). Muscle tissue from both RES and OVER lambs exhibited a reduction in Type I fibers and an increase in Type IIb fibers at birth (Reed et al., 2014). This reduction in Type I fibers could explain the reduction in PPARGC1A gene expression observed in OVER animals. However, it would be expected that changes in the expression of PPARGC1A should have been observed in both RES and OVER lambs, given the similar phenotype. Therefore, these data provide additional evidence that the mechanisms facilitating the changes with maternal under- or overfeeding are different. Additionally, PPARG1A regulates mitochondrial biogenesis (Handschin, 2010), with high fat feeding during gestation causing reduced expression of PPARGC1A in postnatal liver and muscle tissue of rat pups (Borengasser et al., 2014). Therefore, exposure to overfeeding during gestation may influence the mitochondrial biogenesis and metabolism of postnatal muscle tissue.

Two other differentially expressed genes that were identified in OVER lambs, JUNB and THBD, are regulated by insulin (Coletta et al., 2008; Raffaello et al., 2010). At this time point, changes to circulating insulin concentrations were not observed in OVER animals (Hoffman et al., 2016). Changes in insulin sensitivity in the muscle may affect the expression of these genes, but the effects of maternal nutrition on these genes have yet to be investigated. Our analyses indicated that expression of SOCS3 and FOS, 2 factors responsible for inhibiting myogenesis (Broholm et al., 2012), were affected by maternal nutrient restriction. The postnatal involvement of these factors could be a mechanism by which the muscle cross-sectional area of RES lambs did not grow as well as CON lambs and why lipid acumination in the muscle tissue of these animals was greater when evaluated at 3 mo of age (Reed et al., 2014). To better understand these changes, other factors involved in the Janus kinase/signal transducer of transcription pathway and targets of SOCS3 and FOS need to be evaluated.

Maternal under- and overnutrition during gestation affects several epigenetic mechanisms in the offspring including microRNA species (Ford and Long, 2012; Yan et al., 2013). Differentially expressed genes in the current study may indicate that poor maternal nutrition may affect other small RNA species in the offspring. The SNORD113 and SNORA70 are small, nuclear, noncoding RNA that are responsible for guiding RNA for post-transcriptional modifications (Xu et al., 2014), and Y RNA is a small cytoplasmic noncoding RNA that have a critical role in DNA replication (Kowalski and Krude, 2015). Additionally, U1 is a small nuclear RNA that is responsible for pre-mRNA splicing (O’Reilly et al., 2013). Therefore, poor maternal nutrition during gestation may decrease the expression of both small nuclear and cytoplasmic RNA in the muscle tissue of the offspring suggesting, this may be an epigenetic mechanism by which maternal nutrition alters muscle development. Although library preparations and bioinformatics analysis were not performed specifically for the identification of small RNA, given the number that have been identified as differentially expressed in the analysis, this is an area that needs to be further evaluated.

Results of the gene ontology analysis suggest that poor maternal nutrition alters factors involved in metabolic processes, transcription regulation, cytological function, and structure. However, in the current study, the number of differentially expressed genes identified within the analysis is low compared with other studies using RNA-seq analysis in sheep (Peñagaricano et al., 2014). This could be due to the number of replicates that were available for sequencing and the number of reads and read length lost during QC. The Ion Torrent Proton sequencer that was used has an increased error rate in base calling (Golan and Medvedev, 2013), and therefore, this could have contributed to the loss of read length and number after QC. Despite this, we were still able to identify key mechanisms that are altered as a result of maternal under- and overfeeding during gestation.


In conclusion, based on our analysis, we have found that despite exhibiting phenotypic changes similar to muscle development (Reed et al., 2014), the genetic mechanisms involved in the changes to offspring muscle development are different. Furthermore, these changes have been observed at an early postnatal time point, suggesting that these mechanisms may be involved in the long-lasting effects of poor maternal nutrition on muscle development. Identifying these mechanisms is the first step in developing intervention strategies for these offspring to improve productivity and efficiency. Additional work needs to be performed at multiple points during fetal development to better understand how the differences in these mechanisms develop as a result of maternal diet.




Be the first to comment.

Please log in to post a comment.
*Society members, certified professionals, and authors are permitted to comment.