Effects of low and high maternal protein intake on fetal skeletal muscle miRNAome in sheep

1. Introduction

In all mammals, maternal nutrition in pregnancy plays a critical role in intrauterine development, fetal survival, growth, and organogenesis [1,2]. Therefore, maternal under- or over-nutrition during gestation may cause low offspring growth and performance [3] with significant influence on the economic viability of livestock industry [4,5].

Dietary nutrients, especially macronutrients, can influence the rate of gene transcription directly through interactions with the genomic regulatory elements or indirectly affecting and/or modulating important signaling pathways [6,7]. The understanding of how nutrition affects an individual’s phenotype has greatly benefited from developments in molecular biology and genomics, and this area is called nutrigenomics. The term “epigenetics” describes heritable phenotypic alterations that occur without a change in the DNA sequence [8]. A few examples of epigenetic mechanisms are those that modify the chromatin structure, such as DNA methylation and histone modifications, and as well as those that control the activity of proteins, like microRNAs (miRNAs). The term “Nutrimiromics” has been used to refer to the study of interactions between nutrients and food components with the organism’s genome via modification of gene expression due to epigenetic processes related to miRNAs [9].

Protein intake has a variety of nutritional and biological effects. Protein plays a vital role in the regulation of food intake, glucose and lipid metabolism, blood pressure, bone metabolism, and immunological function in addition to nutritional importance as a supply of amino acids for protein synthesis [10]. Proteins’ physiological functions are influenced by their physico-chemical characteristics, amino acid make-up, and bioactive peptides encoded in their amino acid sequences [10]. Furthermore, research has been done on how proteins affect fetal and post-natal development [11]. To date, more experiments have focused on protein restriction compared high protein about maternal nutritional effects on the fetus and infant [12,13]. Blumfield and Collins [13], stated that altered maternal protein intake during pregnancy may have detrimental effects to the offspring, and optimal macronutrient ratio is crucial for the pre- and post-natal environment. Additionally, protein content is one of the most expensive and limiting feed characteristics for livestock nutrition. As a feed ingredient, protein content is usually provided from human-edible ingredients like soybean, other oilseeds, and grains [14]. Thus, careful management of these limiting feedstuffs is critical for sustainable agriculture in the context of growing world population and increasing livestock product demand [15]. Furthermore, balanced protein consumption may help to control nitrogen pollution to aid both environmental and economical sustainability [16].

Previous research has examined extreme undernutrition or overnutrition conditions to test the effects of maternal nutrition on fetal development in livestock and model animals, including fetal programming of muscle and fat tissues [12,17,18]. However, few studies have examined potential effects of maternal protein intake on the overall miRNA expression of fetal skeletal muscle. We have previously reported that maternal diets differing in protein content alter fetal skeletal transcriptome, and maternal high protein compared standard protein triggered pathways related to the immune system and disease in sheep skeletal muscle [18]. On the other hand, the mechanisms by which these diets affect fetal miRNAome in sheep skeletal muscle are not known. The objective of this study was to investigate the impact of different maternal dietary protein levels during mid-to-late gestation on the total miRNA expression of fetal skeletal muscle in sheep.

2. Materials and Methods

2.1. Experimental Animals and Diets

Details of ewes and diets were explained detailed previously [18]. Briefly, three distinct protein diets (standard, high, and low protein) were used in the study to assess the effect of maternal dietary protein on the miRNAome of fetal skeletal muscle. Twelve Akkaraman ewes aged two were divided into three treatment groups at random, with four ewes in each group. Ewe specifications are presented in Sohel et al. [18]. After a 15-day adaptation period, the ewe’s daily dry matter (DM) feed consumption was assessed and compared with the US National Research Council (NRC) [19] recommendations for ewes. Three alternative feed regimens with variable protein content were created based on content of wheat straw, alfalfa, soybean meal, barley, and maize grain, alfalfa and soybean meal were the main protein sources in the diet, whereas barley and corn were the main sources of energy. Wheat straw and alfalfa were additional sources of cellulose and energy. The protein content of the diets was changed by substituting different amounts of soybean meal and alfalfa. As recommended by NRC [19], the ewes were fed from 0 to 30 days of gestation. Ewes were fed the following diets from day 31 to day 105 of gestation: diet 1 was a conventional diet with an NRC level of the nutritional content of feeds (daily DM feed intake ranged from 1252.3 to 1862.4 g, crude protein ranged from 119.4 to 198.6 g, and kcal ME/kg ranged from 2506.0 to 3909.3 g), diet 2: low-protein diet (1247.9–1726.4 g DM feed intake daily; 73.1–112.2 g crude protein; 2520.1–3649.0 kcal ME/kg); 3: high-protein diet (1252.4-629.8 g DM feed intake daily; 160.0–271.6 g crude protein; 2513.0–3880.9 kcal ME/kg), and diet 3: high-protein diet (daily, DM feed intake 1252.4 to 1829.8 g, crude protein 160.0 to 271.6 g and 2513.0 to 3880.9 kcal ME/kg). The diets included alfalfa, wheat straw, barley, corn, and soybean meal. The chemical composition of the materials for feed was ascertained in the feed analysis laboratory in accordance with A.O.A.C. (1990). Diets were made in an isoenergetic fashion. Specific information about diets was reported previously [18].

2.2. Estrus Synchronization, Semen Preparation, and Artificial Insemination

During the annual mating season, estrus synchronization was achieved using progesterone-infused pessaries (EaziBreed CIDR; 1.38 mg progesterone; Zoetis, Turkey) for 12 days, 2 mL of PGF2 (5 mg/mL Dinoprost tromethamine; Dinolytic; Zoetis, Turkey) treatment on day 12, and 400 IU of eCG (Folligon, 1000 IU/vial PMSG, MSD Animal Health, Turkiye) at the time of pessary removal. Four 1- to 3-year-old rams were utilized in the investigation, and the semen was extracted using the artificial vaginal procedure while an estrus ewe was present. A phase-contrast microscope with a heated plate set at 37°C was used to measure the motility of the artificial vaginal method-obtained sperm. For the insemination of ewes, the ejaculates of four rams with a motility of 70% or higher were combined. To achieve homogeneity, the appropriate ejaculates were blended based on their spermatological motility. The density of mixed ejaculates was determined using the hemocytometry method. To create 0.5 ml of semen containing 150×10⁶ motile spermatozoa, the semen was diluted with Trisma base, D-Fructose, Citric acid, and skimmed milk [20]. Then, a water bath set at 35 °C was used to keep the semen. Using diluted fresh semen, artificial insemination was carried out intravaginally 48 hours following sponge removal.

2.3. Necropsy, Sample Collection and Isolation of miRNA

On the 105th day of pregnancy, the dams had necropsies. A total of 14 fetuses were gathered from 12 dams. To avoid growth bias resulting from multiple births, twin birth fetuses were not included in the experiment. In the conventional protein group and the low protein group, respectively, there were two twin pregnancies. To make sure that each experimental group contained 3 fetuses, a fetus from the HP group was also removed. A sample of muscle tissue was taken from the fetal left longissimus dorsi muscles. Prior to miRNA extraction, all tissues were snap-frozen in liquid nitrogen and stored at a temperature of −80 °C. Thirty milligrams of each skeletal muscle tissue were homogenized in QIAzol Lysis reagent (Qiagen, Hilden, Germany) with liquid nitrogen using a homogenizer (Precellys Bertin, France). Using a miRNeasy mini kit (Qiagen, Hilden, Germany) in accordance with the manufacturer’s instructions, total RNA containing miRNA was isolated from tissue homogenates. The manufacturer’s recommendations were followed for the on-column DNA digestion stage using the RNase-Free DNase Set (Qiagen, Hilden, Germany). With the help of a BioSpec-nano micro-volume UV-Vis spectrophotometer (Shimadzu Co., Kyoto, Japan), the purity and concentration of the total RNA were assessed.

2.4. Microarray Hybridization and Data Analysis

The miRNA profile was determined from the Affymetrix Microarray system using the GeneChip miRNA 4.0 Array (Affymetrix, USA). The miRNA microarray contains 30,434 mature miRNAs from all organisms, miRBase set Release 20.0. Following the manufacturer’s instructions, 500 ng of total RNA was labeled using an Affymetrix FlashTag Biotin HSR RNA Labeling Kit. Following hybridization, the chip arrays were washed and stained using AGCC Fluidics Control Software on a Fluidics Station 450. After that, a high-resolution Affymetrix GeneChip Scanner 3000 was used to scan the array’s fluorescence. The Affymetrix GeneChip Command Console program produced probe cell intensity files (*CEL files), which were imported into probe-level summary files (*CHP files) for data extraction. Software called Expression Console and Transcriptome Analysis Console were used to analyze these *CHP files. After image analysis, the data were normalized. Data was compared using tests including the t test, analysis of variance (ANOVA), and significance analysis of microarray after log₂ translation of the normalized signal intensities of arrays. The identification of miRNA-targeted genes was performed using TargetScan [21] (http://www.targetscan.org; accessed on 08 January 2024) with selection criteria of cumulative weighted value ≤ −0.75 for Targetscan. Genes targeted by multiple miRNAs were assembled from individual miRNA target lists. ConsensusPathDB [22] (http://cpdb.molgen.mpg.de/; accessed on 10 January 2024) was used to perform gene ontology (GO) annotation and enrichment analysis from three ontologies (molecular function, biological process, and cellular component) on the frequently identified target genes. The threshold for significance was set at p < 0.01 for both of these analyses. The predicted target gene lists were examined using the default collections of the KEGG, Reactome, and BioCarta (http://cgap.nci.nih.gov/Pathways/BioCarta_Pathways) databases.

3. Results

Microarray analysis was used to compare the expression profiles of miRNAs in the fetal longissimus dorsi muscle samples collected from ewes fed with different protein (standard-SP, high-HP, and low protein-LP) diets. As presented in Table 1, 12 different miRNAs were differentially expressed (>2-fold, p < 0.01, FDR <0.05) in the different diet groups as per the Affymetrix GeneChip miRNA 4.0 Array. Chromosomal locations and mature miRNA sequences of these 12 prominent miRNAs identified using Transcriptome Analysis Console commercial software are shown in Table 1.

We identified 2 significantly up-regulated miRNAs in the HP group compared to the SP group. (Table 1). The 2 miRNAs were oar-miR-3957-5p and oar-miR-329a-3p. The fold change of those miRNAs was 4.58, and 2.96, respectively. miRNA oar-miR-3957-5p was also upregulated in SP group compared to HP group with the fold change 4.25. One of the remarkable results of our miRNAome analysis was the identification of nine miRNAs on OAR18 in which they clustered together in a tight locus (Figure S1). The miRNA oar-miR-200c showed up-regulation in SP group compared to LP group. The differentially expressed miRNAs were grouped using a hierarchical clustering technique, and the outcomes demonstrated that the various diet groups could be readily distinguished by the hierarchical grouping of all covered ovine miRNAs (Figure S2).

Next, we analyzed the target genes that may be regulated by these 12 miRNAs in two layers. First, we identified genes targeted by multiple miRNAs (Table S1). In silico analyses identified that 32 genes found to be targeted by two miRNAs and no mRNA found to be targeted by three or more miRNAs (Table S1). In the second layer, a total of 65 genes were targeted by at least one of 12 miRNAs (Table S2). Hypergeometric distribution testing was used to examine the gene ontology (GO) of their anticipated target genes to investigate the roles of differentially regulated miRNAs among 3 distinct diet groups. GO terms for genes that targeted by 2 miRNAs were identified as being related to 19 biological processes and 6 molecular functions (Table 2). GO terms for genes that are targeted by single miRNA were identified as being related to 20 biological processes, involved in the creation of 5 molecular functions, and expected to take part in 3 cellular components (Table 3).

GO functional enrichment analysis and pathway enrichment analysis were performed on 32 genes targeted by multiple miRNAs and separately on 65 genes targeted by at least one miRNA (Table 2 and Table 3). GO terms showed that genes targeted by multiple miRNAs were mainly involved in biological processes (BP) lymphocyte chemotaxis, monocyte chemotaxis, regulation of nitrogen compound metabolic process, granulocyte chemotaxis and other categories that control immune system (Table 2). In terms of molecular function (MF), targeted mRNAs were involved in cytokine activity, cytokine receptor binding, protein kinase regulator activity (Table 2). We conducted a pathway enrichment study using public pathway databases to better understand the roles of these projected target genes, and 31 pathways were found (Table 3). Pathway analysis showed that targeted genes by multiple miRNAs were significantly enriched in the Chemokine receptors bind chemokines, Rheumatoid arthritis - Homo sapiens (human), Viral protein interaction with cytokine and cytokine receptor - Homo sapiens (human) and Toll-like receptor signaling pathway - Homo sapiens (human) (Table 3).

GO functional enrichment analysis for genes targeted by single mRNAs in terms of BP, the primary categories for the functions were cellular macromolecule biosynthetic process, macromolecule biosynthetic process, regulation of biosynthetic process and cellular nitrogen compound biosynthetic process cell (Table 4). In terms of MF, targeted mRNAs were involved in DNA binding, regulatory region nucleic acid binding and metal ion binding (Table 4). In terms of CC, targeted mRNAs were involved in nuclear lumen, nucleus, nucleoplasm (Table 4). Pathway analysis showed that targeted genes by single miRNAs were significantly enriched in the Rheumatoid arthritis - Homo sapiens (human), Transferrin endocytosis and recycling, Oncogene Induced Senescence and G1 to S cell cycle control (Table 5).

4. Discussion

Proteins are vital for the structure, function, and regulation of the body’s tissues and organs in all living organisms; in addition, they are the most expensive and often most limiting ingredients in livestock feed formulation [23]. Excess fed protein can contribute to excreted nitrogen with undesirable nitrate leaching to groundwater and atmospheric ammonia emissions [24]. Despite the importance of protein for all living organisms, ruminants use dietary protein for production rather inefficiently (as defined by g N in product/g N intake) [25]. This creates opportunity to utilize technologies such as dynamically alter nutrient intakes via mid-infrared spectroscopy of milk composition [24,26]. Additionally, feedomics may assist in modifying nutrition for monetary and long-term gains. The key components of the feedomics study are nutrigenomics, nutrigenetics, and nutritional epigenetics, which aid in establishing a link between genetic variation and nutrient-driven epigenetic changes that are proposed as the main barrier to meeting nutritional demands [27]. Maternal nutrition is a major intrauterine environmental factor that alters expression of the fetal genome in ways that may have lifelong consequences. Such fetal programming has been shown to be regulated by epigenetic modifications, including DNA methylation, histone modifications, and as well as miRNAs [12].

In addition, we have identified that miR-3957-5p was upregulated in both LP and SP groups vs HP group (Table 1). Pokharel et al. [28] detected miR-3957-5p in the tissue biopsies of corpus luteum and endometrium of pregnant ewes. Prior work demonstrated an abundance of miRNAs in packed form as extracellular vesicles to maintain feto-maternal communication in multiple mammals [29,30,31]. In the current study, together with miR-3957-5p, other miRNAs revealed a large cluster of miRNAs on chromosome 18 (Table 1) which may co-regulate different biological processes [32]. This is consistent with Pokharel et al. [28] who also detected similar miRNAs in the same clustered locus. This miRNA cluster is highly conserved among placental mammals and known to be regulated by a maternally imprinted DLK1-DIO3 region, which has been associated with severe muscle disease caused by impaired expression of dystrophin called Duchenne muscular dystrophy [33,34,35]. Additionally, two growth QTL associated with sheep body weight [36] and average daily gain [37] were found coinciding with the miRNA cluster on OAR18. In the same experimental design subjected to current study, Sohel et al. [18] showed that a high-protein diet consumed by the mother predominantly changed the expression of mRNAs involved in myogenesis and immunity in the developing fetus’s skeletal muscle. Both Shandilya et al. [38] and Sharma et al. [39,40] showed that sheep exposed to bacterial lipopolysaccharide challenge exhibited differentially regulated miRNAs such as miR-3957-5p, miR-329a-3p, miR-379-5p (Table 1) which regulate biological processes such as stress and immune responses.

A critical layer of analysis involves miRNA targeting the same genes (Table S2). This is an established part of miRNA biology and a useful way to highlight important genes and pathways [41,42]. Multiple miRNAs regulating the same target gene are naturally cooperative since miRNAs act to inhibit targets; this heterotypic regulation is believed to offer extra regulatory specificity [43]. And specially, there is ongoing debate over whether clustered miRNAs with different seed sequences co-evolved due to shared functions [43]. In the current study, two members of miRNA cluster that are oar-miR-3957-5p and oar-miR-432 (Table S2; Figure S1) are cooperatively targeting Calcium/calmodulin dependent protein kinase IG (CAMK1G) and Ras protein specific guanine nucleotide releasing factor 1 (RASGRF1) genes. Calcium/calmodulin-dependent protein kinases are known to be key regulators of calcium signaling in health and disease [44] and it has been observed that T cells and B cells produce less IFN-γ when CaMKIV is inhibited or silent [45]. Futhermore experiments showed that in myoblasts and myotubes treated with IFN-γ, calcium/calmodulin-dependent protein kinase IV specifically suppresses the production of co-stimulatory molecules and pro-inflammatory cytokines/chemokines [46]. In terms of RASGRF1, it is an intrinsic key mediator for brain-derived neurotrophic factor (BDNF)-induced R-Ras activation and R-Ras-mediated axonal morphological regulation [47]. One of the nerve growth factor family’s members, BDNF, is primarily produced by the brain and its main role involves synaptic modulation, neurogenesis, neuron survival, immune regulation, myocardial contraction, and angiogenesis in the brain [48]. On the other hand, previous experiments had shown that BDNF has a significant regulatory role in myogenic differentiation in skeletal muscle which acts as a retrograde survival factor for motor neurons in the neuromuscular system [49,50]. Additionally, on the same miRNA cluster, oar-miR-379-5p and oar-miR-134-5p were found to be cooperatively targeting Glutamyl-TRNA Amidotransferase Subunit C (GATC) (Table S2). GATC was associated with mitochondrial respiratory dysfunction which cause skeletal muscle atrophy, is a state of uncontrolled inflammation and oxidative stress [51]. Multiple target analysis showed a complex network of mutual interactions between let-7c, miRNAs on various chromosomes and mRNAs (Table S2). In literature role of Let-7c and other Let-7 family members’s in myogenesis [52] and ovine fetal skeletal muscle development related to maternal obesity had been shown [53].

A second outcome involves immune genes. Chemokine ligands (CCLs), CCL3, CCL3L1, and CCL3L3 are all targeted by both oar-miR-432 and let-7c (Table S2). CCL3 is critical for trophoblast invasion and maintenance of pregnancy [54,55]. While others have studied dietary impacts on sheep, to our knowledge, this is the first demonstration that maternal diet protein intakes in the context of isoenergetic dietary formulations could impact CCL3 and maintenance of pregnancy. In addition, IL7 is targeted by both oar-miR-432 and oar-miR-200c (Table S2). This is particularly striking because IL7 influences development of the offspring’s T cell immunity [56,57]. These examples highlight the importance of maternal protein intake on multiple critical outcomes for fetal health and lifetime wellbeing.

Rheumatoid arthritis - Homo sapiens (human) pathways were identified in different protein intake groups among both layers of target prediction (Table 3 and Table 5). Rheumatoid arthritis (RA) is a chronic, inflammatory, systemic autoimmune disease that causes skeletal muscle loss and weakness in RA individuals [58,59]. Similarly, transferrin endocytosis and recycling pathway is associated with iron homeostasis, is a key regulator of innate and adaptive immunity and plays a decisive role in inflammatory processes [60]. These findings were consistent with the findings of transcriptome analysis [18], which revealed that many detected genes are involved in pathways related to the immune system and diseases.

5. Conclusions

In summary, muscle biology is crucial for economic and nutritional value of sheep, and long-term optimization of sheep production in requires enhanced understanding of the molecular mechanisms that control muscle growth and metabolism. To the best of our knowledge, this is the first study that investigates the nutrimiromics of sheep fetal skeletal muscle based on differing maternal protein intake. In addition, experiments that investigate effect of nutrients and bioactive compounds that regulate miRNAs in the fetal period are very limited. The findings of our research show that maternal diet in terms of different protein levels may affect fetal miRNAome that epigenetically affect the skeletal muscle transcriptome in vivo [18]. The discovery of multiple miRNAs that target mRNAs is a significant development that may enable us to further comprehend the network that connects miRNAs and their targets. Since the maternal high-protein diet specifically changed the expression of miRNAs involved in number of pathways and biological processes that are unique to disease and immunity, these results may have provided insight into the reversible nature of epigenetic modifications, which hold promise for the use of miRNAs in the diagnosis, prevention, and treatment of health conditions in the offspring.