Negative energy balance by feed deprivation affects the adipose miRNome in the lactating goat

Abstract

One of the main functions of ruminant adipose tissue (AT) is to store lipids for use in productive functions. Body fat mobilization is required during periods of negative energy balance (NEB) such as early lactation or undernutrition. Ruminant nutrition modifies the expression of adipose genes, the regulation of which is not fully understood. The expression of more than 60% of protein-coding genes is post-transcriptionally regulated by microRNAs (miRNAs, small non-coding RNAs, 18 to 25 nucleotides targeting messenger RNAs). Our aim was to characterize miRNA whose expression is regulated by nutrition in the lactating goat AT. Using high-throughput sequencing technology, miRNomes of the lactating AT were established from lactating goats fed a control diet ad libitum and after 48 h of food deprivation (FD) leading to an NEB. MiRNAs sequencing revealed 637 known miRNAs in omental AT of which 16 showed an expression modulated by FD; 14 were upregulated and 2 were downregulated. The network of miRNA-target enrichment identified 2 major miRNAs, miR-223-3p and miR-21-5p which were upregulated by FD and suggested an increase in inflammation of the AT with an NEB obtained after fasting during lactation. The target gene predictions of the differentially expressed miRNAs in AT indicated a significant enrichment in gene ontology functional categories of cell life including apoptosis, cell proliferation, and differentiation as well as in gene expression machinery including regulation of translation and transcription. These data suggest that miRNAs may play a key role in the regulation of AT remodeling.

An overview of goat adipose miRNA profiles (miRNome) and the identification of miRNAs differentially expressed under 2 different diets (ad libitum vs 48 h food deprivation) using RNA sequencing could improve the understanding of the mechanisms underlying the regulation of adipose gene expression in response to nutritional factors in dairy ruminants.

Introduction

One of the major functions of ruminant adipose tissue (AT) is to synthesize fatty acids and store triglycerides for use in productive functions such as lactation. Several rearing factors (including genetic, nutritional, physiological, and environmental) can modify AT lipid metabolism. For example, postpartum is a period of metabolic stress, with a rapid increase in milk secretion and a slower increase in feed intake, resulting in a negative energy balance (NEB) and body fat mobilization. Dairy ruminants respond to this postpartum NEB by decreasing lipogenesis and increasing lipolysis in ATs of bovine (McNamara, 1989; Chilliard et al., 1991; Sumner and McNamara, 2007), ovine (Vernon et al., 1987), and caprine (Chilliard et al., 1987, 2003; Dunshea et al., 2000) species. Transcriptomic analyses of bovine AT showed that the expression of most of the genes controlling lipogenesis declined in early lactation, whereas those coding for lipolytic control did not change during this period and then increased until peak lactation, with a concomitant increase in milk yield (Sumner and McNamara, 2007; Sumner-Thomson et al., 2011). Similarly, changes in AT metabolism occur in underfed ruminants, with a decrease in fatty acid synthesis, lipogenesis, and expression of some genes encoding key lipogenic enzymes in ruminant AT (Bonnet et al., 1998; Chilliard et al., 2000, 2003; Faulconnier et al., 2001). High-throughput RNA-Sequencing (RNA-Seq) provides a sensitive method that does not require a priori information. However, until now, RNA-Seq data on ruminant AT with NEB are rather scarce. In growing cattle, diets with different energy and protein content affected the expression of protein-coding genes in subcutaneous AT (Waerp et al., 2018). In dairy cows, the severe NEB resulted in altered expression of protein-coding genes related to lipid metabolism in subcutaneous AT (Mellouk et al., 2019; Salcedo-Tacuma et al., 2020). The effects of NEB obtained by food deprivation (FD) during lactation on AT expression of coding genes were also reported in goats. Indeed, transcriptomic analyses of lactating caprine AT demonstrated that a 48 h FD induced a change in the expression of 456 and 199 genes in omental and perirenal ATs, respectively, with more pronounced effects in omental than perirenal AT (Faulconnier et al., 2011). However, the mechanisms underlying the nutritional regulation of these adipose genes remain poorly documented. Lately discovered participants in the regulation of protein-coding gene expression are microRNAs (miRNAs). They are small single-stranded non-coding RNAs of ~22 nucleotides, in length that regulate coding gene expression by altering mRNA stability or translation. MiRNAs are thought to regulate at least ~60% of genes and are involved in many biological processes (Bartel, 2004; Friedman et al., 2009).

To better understand the mechanisms underlying the regulation of adipose protein-coding gene expression in response to dietary factors, the aim of this study was to obtain an overview of goat adipose miRNA profiles and to identify miRNAs whose expression was regulated during NEB. To mimic NEB, a 48 h-FD, previously studied at the mRNA level, was investigated at the miRNA level. Therefore, the adipose miRNA profiles of lactating goats fed either a control diet ad libitum or after 48-h FD were compared by RNA-Seq.

Materials and Methods

Ethics statement

This study was carried out at the experimental unit of the INRAE Research Center in Theix. To limit the animal experiment, we used samples from a previous study (Ollier et al., 2007). Thus, we did not have to submit each animal experiment to the ethics committee, the institute had its recommendations for Animal Care as well as the ethics committee for animal experiments of our region (Auvergne: CEMEAA number 02) which were strictly followed. To minimize suffering, all goats were euthanized using a captive bolt gun followed by exsanguination at the slaughterhouse of the INRAE Research Center in Theix under accreditation number 63 345 001. All omental AT samples were collected after the animals were slaughtered.

Animals and sampling

Eleven lactating Alpine goats from the Lusignan experimental station (France) were based on the homogeneity of the lactation period and parity. The goats were studied at the peak of lactation (48 ± 2 d postpartum, at the beginning of the experiment), a period during which AT lipids are strongly mobilized. For a 2-wk pre-experimental period, the goats were fed a standard diet of orchardgrass hay with a 35:65 ratio of forage-to-concentrate. During the 48 h prior to slaughter, 5 goats received the standard diet ad libitum (control, CTRL) while the remaining 6 goats were food deprived (FD_d). The goats were housed in individual stalls, had free access to water, and were fed twice daily (except during the 48-h fasting period). Omental AT samples were collected within 30 min after euthanasia, immediately snap frozen, and stored at −80 °C.

RNA preparation and RNA-Seq analysis

Total RNAs were extracted from an average of 1 g of AT powder (n = 5 CTRL and n = 6 FD_d) plus 750 µl of Trizol LS, using a miRVana kit (Thermo Fisher Sciences, USA) according to the manufacturer’s instructions. The concentration and quality of the RNAs were estimated by spectrophotometry NanodropTH (ND-1000, NanoDrop Technologies LLC, Wilmington, DE, USA) and using the Bioanalyzer 2100 (Agilent Technologies Inc., Santa Clara, CA, USA), respectively.

Libraries preparation and RNA-sequencing were performed by the IGBMC GenomEast platform, a member of the ‘France Génomique’ consortium (ANR-10-INBS-0009; Strasbourg, France). Small RNA-Seq libraries were generated from 2 μg of total RNA using TruSeq Small RNA Library Prep Kit (Illumina, San Diego, CA), according to the manufacturer’s instructions. Briefly, during the first step, RNA adapters were sequentially ligated to each end of the RNA; firstly the 3′ RNA adapter (5′ TGGAATTCTCGGGTGCCAAGG 3′) which is specifically designed to target miRNAs and other small RNAs, then the 5′ RNA adapter (5′ GTTCAGAGTTCTACAGTCCGACGATC 3′). Small RNA ligated with 3′ and 5′ RNA adapters were reverse transcribed, and PCR amplified (30 s at 98 °C; [10 s at 98 °C, 30 s at 60 °C, 15 s at 72 °C] × 13 cycles; 10 min at 72 °C) to obtain cDNA. Acrylamide gel purification of 140 to 160 nt amplified cDNA (corresponding to cDNA obtained from small RNA + 120 nt from the adapters) was performed. The final cDNA libraries were checked for quality and quantified by capillary electrophoresis. Libraries were loaded into the flow cell at 2.8 nM and clusters were generated using Cbot and sequenced on HiSeq 4000 (Illumina) as single-end 50 base reads, according to the manufacturer’s instructions. After trimming adapter sequences and removing reads containing ambiguous base calls (FASTX-Toolkit, https://github.com/agordon/fastx_toolkit), reads were filtered according to their size (15 to 40 nt). The sequence reads were aligned against the Bos taurus btau5.0.1 genome as miRBase_v22.1 using the miRDeep2 mapper.pl module (2.0.1.2; Friedlander et al., 2012). Precursors and mature miRNAs were identified using the miRDeep2 core module, miRDeep2.pl. Potential miRNAs were annotated against goat, sheep, and human orthologous miRNAs (miRBase release 22.1). We used a miRDeep2 score ≥0 as a cutoff threshold. The accession number of the RNA-Seq data is GSE236131.

Statistical and bioinformatics analyses

Normalization and differential expression analysis of the miRNAs between CTRL and FD_d goats were conducted with the DESeq2 R package v1.18.1, including a Benjamini–Hochberg (Benjamini and Hochberg, 1995) multi-testing correction. Significance was considered at P_adj ≤ 0.1. In silico functional annotation using OmicsNet tools with Panther database and P_adj ≤ 0.1 was performed to identify the biological processes potentially regulated by the differentially expressed miRNAs (DEMs) in the caprine AT. Due to the lack of predictive tools for ruminants and the high conservation of miRNAs through evolution, bioinformatic analyses were performed using tools developed for humans with miRTarBase (V8.0) which provides experimentally validated miRNA-gene interactions with a threshold P-value ≤ 0.10. Target genes of the studied miRNAs and potential networks were identified using Mienturnet software (http://userver.bio.uniroma1.it/apps/mienturnet/; [Licursi et al., 2019]) and miRTarBase with 2 as the threshold for the minimum number of miRNA-target interactions. Venn diagrams were generated using Venny 2.1 software (https://bioinfogp.cnb.csic.es/tools/venny/index2.0.2.html).

Results and Discussion

Global description of AT miRNomes in lactating goats

Eleven libraries were constructed from the omental AT of 5 CTRL and 6 FD_d goats. RNA-Seq analyses showed an average of about 23.8 and 27.5 million reads from the CTRL and FD_d AT libraries, respectively (Table 1). After removing sequencing adapters and filtering reads by their size, 22.6 and 25.8 million clean reads (corresponding to an output rate of 94.8% for both) of 17 to 28 nt were obtained for the CTRL and FD_d libraries, respectively. The reads were mapped and clustered into an average of 85,005 and 101,058 unique sequences, in the CTRL and FD_d libraries, respectively (Table 1). The knowledge of genome and functional annotations was greater in bovine than in caprine, and previous studies on the caprine transcriptome (Mobuchon et al., 2015; Faulconnier et al., 2022) showed that the annotations from Capra hircus were included in those obtained with Bos taurus, thus bovine was used for the annotations and analyses of the sequences described below. A total of 637 known miRNAs were detected at least once in the omental AT miRNome (Supplementary Figure 1A). Liu et al. (2022a) showed 233 miRNAs specifically expressed in goat perirenal AT, which is lower than those we detected. The study by Gu et al. (2007) was among the first to describe bovine adipose miRNome, identified 135 miRNAs from the AT of non-pregnant and non-lactating Holstein cows? Similarly, 222 known miRNAs were detected in the subcutaneous fat of Italian Large White pigs (Gaffo et al., 2014). The different number of known miRNAs detected could be explained by the experimental design (number of samples, statistical methods for prediction, and number of reads per sample), the location of the AT (omental vs. perirenal), the physiological status studied, or the version of miRBase used for miRNA annotation (present study: v22. vs. v21) in the study of (Liu et al., 2022a) and (Gu et al., 2007). However, the number of 637 known miRNAs detected was closer to that of other recent publications. For example, (Liu et al., 2022b) identified 776 and 752 miRNAs in the subcutaneous AT of Duolang and Small Tail Han sheep, respectively, including 660 common to both breeds. Similarly, a study of subcutaneous tail AT from periparturient cows allowed the detection of a total of 744 miRNAs (Sadri et al., 2022). Likewise, 780 miRNAs were identified in perirenal AT from rabbits (Wang et al., 2020).

Table 1.

Summary of sequencing data

	CTRL1	FDd2
Raw reads	23.882.220	27.489.863
Cleaned reads3	22.713.103	25.962.138
Sized reads4	22.574.355	25.758.623
Reads mapped5	20.176.549	22.362.078
Unique sequences corresponding to mapped reads	85.005	101.058

¹means of data for the 5 control (CTRL) goat libraries.

²means of data for the 6 food deprived (FD_d) goat libraries.

³sequencing adapters removed.

⁴17 to 28 nt size filter, used by the miRDeep2 software.

⁵reads with at least one and at most 5 reported alignments, used by the miRDeep2 software.

The distribution of miRNAs according to the mean of their reads in all the libraries showed that 4 miRNAs (miR-10a, miR-10b, miR-26a, and miR-143) had a mean of more than 1,000,000 reads. Seventeen miRNAs had an average of reads between 100,00 and 1,000,000 and 155 miRNAs had less than 1,000 reads (Figure 1). Therefore, the 21 most abundant miRNAs with more than 100,000 reads in mean in the AT represented 90% of the total reads with the first 4 (miR-143, miR-10a, miR-10b, miR-26a) representing nearly 65% of the total reads (Figure 2). These data are consistent with the TOP 21 most abundant miRNAs observed in perirenal AT of goats (Liu et al., 2022a) and in subcutaneous AT from pigs (Gaffo et al., 2014), with 7 miRNAs in common with our study, including miR-10a and miR-10b which were highly detected (Figure 3).

Figure 1.

Distribution of miRNAs according to the read means in adipose tissue of lactating goats.

Figure 2.

Relative abundance of miRNAs with an average of more than 100,000 reads led to the identification of the TOP 21 miRNAs in the omental adipose tissue of lactating goats. Representation of the average percentage of normalized reads of each miRNA relative to the total reads of control (CTRL) and food-deprived (FD_d) goats (n = 11).

Figure 3.

Venn diagram of the TOP 21 most highly expressed miRNAs in different adipose tissue (AT) studies. The list of the comparison between omental (OM, present study) and perirenal (PR, Liu et al, 2022a) ATs of goats and subcutaneous (SC) AT from pigs (Gaffo et al., 2014). The name of the common miRNAs between the 3 studies is indicated.

Identification of miRNA differentially expressed in goat AT after FD

To identify miRNAs significantly affected by FD, a differential analysis between FD_d compared to CTRL miRNomes was performed using the DESeq2 package (Table 2). Sixteen DEMs were identified including (Supplementary Figure 1B), 11 of which showed a fold change ≥2. One miRNA (miR-21-5p) was among the TOP 21 most highly expressed miRNAs in goat omental AT with an average of more than 184,000 reads. Six (miR-16a, miR-25, miR-30e-5p, miR-100, miR-146b, and miR-186) were detected with more than 5,000 reads on average.

Table 2.

miRNA whose expression was affected by food deprivation in lactating goat omental adipose tissue

miRNA	Séquence	CTRL*1	FDd*2	log2FC3 (FDd/CTRL)	P-adj4
miR-21-5p	UAGCUUAUCAGACUGAUGUUGACU	134,545.07	234,664.97	0.80	0.02
miR-27a-5p	AGGGCUUAGCUGCUUGUGAGCA	153.71	362.61	1.24	0.14
miR-132	UAACAGUCUACAGCCAUGGUCG	442.80	1078.74	1.29	0.03
miR-142-5p	CAUAAAGUAGAAAGCACUAC	16,568.18	28,703.74	0.79	0.00
miR-196b	UAGGUAGUUUCCUGUUGUUGGGA	29.77	62.40	1.08	0.00
miR-212	ACCUUGGCUCUAGACUGCUUACU	69.81	172.76	1.31	0.00
miR-223	UGUCAGUUUGUCAAAUACCCCA	294.15	1221.73	2.06	0.01
miR-410	AAUAUAACACAGAUGGCCUGU	398.87	691.98	0.80	0.00
miR-431	UGUCUUGCAGGCCGUCAUGCAGG	1.34	18.48	3.82	0.15
miR-541	UGGUGGGCACAGAAUCCGGCCU	8.21	50.49	2.61	0.06
miR-543	AAACAUUCGCGGUGCACUUCUU	16.54	44.07	1.42	0.14
miR-665	ACCAGUAGGCCGAGGCCCCU	124.46	349.81	1.49	0.01
miR-1247-5p	ACCCGUCCCGUGCGUCCCCGGA	56.57	153.1	1.43	0.10
miR-1260b	AUCCCACCACUGCCACCA	158.59	459.29	1.53	0.13
miR-2285bb	GAAAGUUUGUUGGGGUUUUUCU	267.83	44.00	−2.61	0.00
miR-6715	ACAGGCACGGCCAGUUUGAGC	88.45	43.54	−1.01	0.07

^*Mean of normalized read counts.

¹means of data for the 5 control (CTRL) goat libraries.

²means of data for the 6 food deprived (FD_d) goat libraries.

³Fold Change: A positive log2 fold change means the corresponding miRNA is more highly expressed in FD goats than in CTRL and vice versa. P-values were adjusted using the Benjamini–Hochberg (Benjamini and Hochberg, 1995) correction at 10%.

⁴ P adjusted.

To the best of our knowledge, this is no study on the effect of FD on the ruminant miRNome AT. However, among these nutriregulated miRNAs, 5 (miR-196b, miR-223, miR-431, miR-541, and miR-1247), upregulated in the present study, have been previously described to be affected by caloric restriction or when animals are in NEB in other species and/or matrices (tissues or plasma). Indeed, in the lactating goat mammary gland, the expressions of miR-196a-5p and miR-223-3p were increased by FD, whereas miR-541-5p decreased (Mobuchon et al., 2015). The expression of miR-223-3p was also increased in the skeletal muscle of monkeys receiving a restricted diet (Mercken et al., 2013). Recently, Veshkini et al. (2022) observed that miR-223 and miR-431 were upregulated while miR-1247 was downregulated in the plasma of dairy cows during the transition period from pregnancy to lactation when the animals are in NEB.

The nutriregulation of miRNAs in AT was reported using high-fat diets. The expression of miR-142 and miR-196b both increased by FD in the present study, were altered in ovine AT by the fatty acid composition of the diet (Meale et al., 2014). The expression of miR-142-5p was downregulated in subcutaneous and visceral ATs of lambs fed a diet rich in n-3 docosahexaenoic acid (DHA) compared to the CTRL diet whereas miR-196 was upregulated but only in subcutaneous AT (Meale et al., 2014). The expression of miR-142-5p was higher in steers fed the high-fat diet than those fed the CTRL diet in subcutaneous and visceral ATs (Romao et al., 2012).

Due to the role of AT in lipid storage, we also compared the results of the present study with those related to body fat deposits. Indeed, the body fat deposits have also been reported to modulate the expression of miRNAs (miR-21-5p, miR-27a-5p, miR-132, miR-142-5p, miR-196b). A correlation between backfat thickness and the expression of miR-142-5p and miR-196b was reported in bovine subcutaneous AT with miR-142-5p downregulated in cattle with thicker subcutaneous fat whereas miR-196b was upregulated (Jin et al., 2010). In porcine subcutaneous AT, the expression of miR-132 and miR-21-5p were downregulated in fat compared to lean pigs (Davoli et al., 2018). In humans, the expression of miR-132 was decreased between obese and non-obese omental (Heneghan et al., 2011) and subcutaneous (Giardina et al., 2018) ATs and correlated with changes in fat mass. However, an increase in the expression of miR-21 and miR-142-5p were observed in subcutaneous AT of obese compared to lean subjects (Keller et al., 2011; Chartoumpekis et al., 2012; Kurylowicz, 2021). The literature revealed discrepant results regarding miR-27, which was upregulated by FD, in AT of obese vs controls. Indeed, the expression of miR-27 decreased (Keller et al., 2011; Chartoumpekis et al., 2012; Kurylowicz, 2021) or did not differ (Torres et al., 2022) in subcutaneous or visceral AT of obese subjects. In the epididymal AT of ob/ob mice, the expression of miR-27 was upregulated compared to lean mice and identified as a strong regulator inhibiting adipogenesis (Lin et al., 2009). Taken together these studies indicated the influence of miR-27 in adipogenesis but its mechanism of action needs to be investigated.

Putative functions of the 16 DEMs

In silico functional annotation by OmicsNet showed that the 16 DEMs are mainly involved in the regulation of transcription and RNA metabolism (42%) and in cell life (29%) including apoptosis and cell cycle. The second class of affected biological processes were those involved in immune response (9%) and lipid metabolism (3%; Figure 4). The putative effects of FD on the regulation of transcription and RNA metabolism and cell life may suggest a link to FD-induced remodeling of AT. Future investigations including on chromatin will validate these effects of FD and may refine our understanding of the mechanisms underlying this regulation. These data are in line with those reported in human AT and in 3T3-L1 cells describing a central role of miRNAs in the regulation of adipogenesis (cell proliferation and differentiation) and in lipid metabolism (lipolysis, lipogenesis, and lipid droplet formation; Lin et al., 2009; Kurylowicz, 2021; Heo et al., 2022). Our results showed that FD-induced significant changes in miRNA expressions via their action on protein-coding genes related to AT remodeling might alter AT function, as observed in the AT of obese subjects (Lee et al., 2010; Kurylowicz, 2021) and in patients with undergo chronic energy restriction (Verma et al., 2020). To complete the functional annotations, target gene predictions were performed using Mienturnet and miRTarBase. A total of 88 validated miRNA-gene interactions were found to have at least 2 miRNA-target interactions (Supplemental Table 1). These target protein-coding gene predictions of the DEMs in AT indicated a significant enrichment in gene ontology functional categories of cell life including apoptosis, cell proliferation, and differentiation as well as gene expression machinery including regulation of translation and transcription, confirming OmicsNet analyses described above.

Figure 4.

Biological processes identified by OmicsNet software with Panther database to be influenced by the 16 candidate miRNAs whose expression was affected by food deprivation in lactating goat omental adipose tissue with FDR ≤ 0.1.

The network of miRNA-target enrichment, conducted using Mienturnet to complete the overview of the effects of FD, identified 2 upregulated miRNAs, miR-223, and miR-21, which are predicted to target 50 and 23 genes, respectively (Figure 5). The miR-223 is a multifunctional miRNA regulated by several transcription factors, and its expression is significantly increased during cellular or tissue inflammation, especially during AT dysfunction or under pathological conditions (Kim et al., 2019; Estrella Ibarra et al., 2021; Jiao et al., 2021). Indeed, upregulation of miR-223-3p was observed in human subcutaneous, omental, and visceral ATs of obese (Macartney-Coxson et al., 2020; Estrella Ibarra et al., 2021) and type 2 diabetes (Sanchez-Ceinos et al., 2021) patients. miR-21 was an important miRNA frequently regulated in many types of chronic diseases and involved in the cell cycle (cell death and cell migration) and inflammatory processes (Jenike and Halushka, 2021). It was overexpressed in the subcutaneous AT of obese subjects (Keller et al., 2011) and in omental and subcutaneous ATs of type 2 diabetes patients (Guglielmi et al., 2017). Elsewhere, miR-21 was shown to promote human adipocyte differentiation (Kim et al., 2009). Thus, the upregulation of miR-223 and miR-21, in the present study, could indicate an increased inflammation of the AT after fasting during lactation, as usually observed in human diseases. An inflammatory status of the goat AT after fasting was also suggested previously by the upregulation in perirenal and omental goat ATs of several genes encoding mediators of local inflammatory processes (Faulconnier et al., 2011).

Figure 5.

Network identified by Mienturnet software as affected by miRNAs regulated by food deprivation. The larger blue and smaller yellow circles represent the targeted miRNAs and mRNAs, respectively.

In a previous study, we observed that 48 h-FD altered the expression of 391 known protein-coding genes (mRNAs) in the omental AT of lactating goats (Faulconnier et al., 2011). A comparison of the 88 miRNA predicted target genes in the present study and differentially expressed protein-coding genes pointed out in the previous study (Faulconnier et al., 2011) using the same samples, identified 7 (APAF1, CDK2, CFTR, HIF1A, IGF1R, PTPN14, and SPRED1) genes in common (Figure 6). Five (APAF1, CDK2, CFTR, IGF1R, and SPRED1) were downregulated by FD and 2 (HIF1A and PTPN14) were upregulated. HIF1A, hypoxia-inducible factor 1 subunit alpha, is a transcription factor involved in the NF-κB signaling pathway, which has long been recognized as a “master switch” in regulating the expression of several cytokines (such as osteopontin) and a wide range of genes involved in inflammation and immune responses (Hayden et al., 2006). He et al. (2011) observed an increase of both HIF1A mRNA and protein in the obese mice AT. The upregulation of HIF1A is also known to increase the expression of inflammation-related adipokine genes and to decrease adiponectin in mouse AT (Higami et al., 2004). It should be noted that a feedback regulatory loop between HIF-1A and miR-21 (identified above as a key regulatory miRNA) has been previously reported (Liu et al., 2014). The second common upregulated gene, the Protein Tyrosine Phosphatase Non-Receptor Type 14 (PTPN14), is known to be a signaling molecule that regulates a variety of cellular processes including cell growth and differentiation. Among the commonly downregulated genes, APAF1, an apoptotic protease-activating factor 1, encoding a cytoplasmic protein, and CDK2, encoding a member of a family of serine/threonine protein kinases, both participate in cell cycle regulation including apoptosis (Fan et al., 2007; Colon-Mesa et al., 2021). The expressions of APAF1 and CDK2 were downregulated and upregulated, respectively, in the AT of diet-induced obese rats (Fan et al., 2007) and mice (Colon-Mesa et al., 2021). In the present study, these 2 genes were also altered but both were decreased by FD in goat omental AT. IGF1R, Insulin-Like Growth Factor 1 Receptor, which mediates the effects of insulin-like growth factor 1 (IGF1), is involved in cell growth. Its expression in perirenal, subcutaneous, and omental ATs of cattle was greater after exposure to a high compared to a low level of maternal nutrition during the second trimester of pregnancy (Micke et al., 2011), consistent with the downregulation of IGF1R in the present study in restricted goats. The CFTR gene encodes a member of the ATP-binding cassette (ABC) transporter superfamily. It was known to control ion secretion in epithelial cells and therefore its loss impairs host defense (Shah et al., 2016). Elsewhere, CFTR knockout mice exhibited insulin resistance and glucose tolerance (Gu et al., 2021). In addition, the treatment of cystic fibrosis caused by mutations in the CFTR gene can improve weight and body mass index (Stallings et al., 2018), thus affecting AT functions. The SPRED1 gene encodes a member of the Sprouty family, which has been shown to be localized in lipid raft membranes, and involved in cholesterol transport (Nonami et al., 2005).

Figure 6.

Comparison of the miRNA-target genes in the present study and differentially expressed genes (DEG) identified in the same caprine omental adipose tissue (Faulconnier et al., 2011). The names of the genes shared with the present study are indicated.

Taken together, the expression profiles of the aforementioned genes suggest an inflammatory state and a remodeling of the AT after a 48-h fasting in lactating goats.

Conclusion

To the best of our knowledge, this is the first report on miRNA profiles of omental AT of lactating goats using high-throughput miRNA sequencing. MiRNA sequencing revealed 637 known miRNAs in omental AT, of which 16 showed a modulated expression after 48 h-FD. Their target prediction analyses revealed a significant enrichment in gene ontology of functional categories of cell life including apoptosis, cell proliferation, and differentiation as well as in gene expression machinery including regulation of translation and transcription suggesting a remodeling of AT after FD. The miRNA-target enrichment network identified 2 major miRNAs, miR-223 and miR-21 upregulated by FD, which may indicate an increase in inflammation of the AT with a NEB obtained after fasting during lactation. Further studies on the role of these 2 major miRNAs are needed to support our results and to decipher their precise influence as major regulators of AT function.

Glossary

Abbreviations

APAF1

apoptotic peptidase activating factor 1

AT

adipose tissue

CDK2

cyclin-dependent kinase 2

CFTR

cystic fibrosis transmembrane conductance regulator

CTRL

control

DEM

differentially expressed miRNA

DEG

differentially expressed gene

DHA

n-3 docosahexaenoic acid

FD

food deprivation

FD_d

food deprived

GO

Gene Ontology

HIF1A

hypoxia inducible factor 1 subunit alpha

IGF1

insulin-like growth factor 1

IGF1R

Insulin Like Growth Factor 1 Receptor

Log2FC

Log2 Fold Change

NF-κB

nuclear factor-kappa B

miRNAs

microRNAs

miRNA profile

MiRNome

NEB

negative energy balance

OM

omental

PR

perirenal

PTPN14

Protein Tyrosine Phosphatase Non-Receptor Type 14

RNA-Seq

RNA-sequencing

SC

subcutaneous

SPRED1

sprouty related EVH1 domain containing 1

Conflict of interest statement

The authors declare to have no conflict of interest.

Author Contributions

Yannick FAULCONNIER (Conceptualization, Formal analysis, Investigation, Methodology, Project administration, Writing—original draft), Tao Ye (Formal analysis, Methodology), and Christine Leroux (Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Writing—original draft)

Data availability

RNA-sequencing data were deposited in the Gene Expression Omnibus: GSE236131.