Nutrient supply alters transcriptome regulation in adipose tissue of pre-weaning Holstein calves

Abstract

Performance of dairy cows can be influenced by early life nutrient supply. Adipose tissue is diet sensitive and an important component in that process as it is involved in the regulation of energetic, reproductive and immunological functions. However, it is not clear how early life nutrition alters the molecular regulation of adipose tissue in calves and potentially adult individuals. This study aimed at determining how differences in pre-weaning nutrient supply alter gene expression profiles and physiology in omental adipose tissue. A total of 12 female Holstein calves were fed two levels of milk replacer supply: a restricted amount of 11.72 MJ of metabolizable energy (ME) intake per day (n = 6) or an enhanced amount of 1.26 MJ ME intake per kg of metabolic body weight (BW0.75), resulting in supply from 17.58 to 35.17 MJ ME intake per day (n = 6). All calves had ad libitum access to a commercial calf starter and water. Analysis of the transcriptome profiles at 54 ± 2 days of age revealed that a total of 396 out of 19,968 genes were differentially expressed (DE) between groups (p < 0.001, FDR < 0.1). The directional expression of DE genes through Ingenuity Pathway Analysis showed that an enhanced nutrient supply alters adipose tissue physiology of pre-weaned calves. Several biological functions were increased (Z-score > +2), including Lipid Metabolism (Fatty Acid Metabolism), Cell Cycle (Entry into Interphase, Interphase, Mitosis and Cell Cycle Progression), Cellular Assembly and Organization (Cytoskeleton Formation and Cytoplasm Development) and Molecular Transport (Transport of Carboxylic Acid). These changes were potentially orchestrated by the activation/inhibition of 17 upstream regulators genes. Our findings indicate that adipose tissue of calves under an enhanced nutrient supply is physiologically distinct from restricted calves due to an increased development/expansion rate and also a higher metabolic activity through increased fatty acid metabolism.

About the Authors:

Leonel N. Leal

Contributed equally to this work with: Leonel N. Leal, Josue M. Romao

Roles Conceptualization, Data curation, Funding acquisition, Investigation, Methodology, Project administration, Writing – original draft

* E-mail: [email protected] (LNL); [email protected] (MAS)

Affiliation: Trouw Nutrition R&D, Amersfoort, Utrecht, the Netherlands

ORCID logo http://orcid.org/0000-0002-7798-3995

Josue M. Romao

Contributed equally to this work with: Leonel N. Leal, Josue M. Romao

Roles Conceptualization, Data curation, Formal analysis, Methodology, Validation, Visualization, Writing – original draft, Writing – review & editing

Affiliation: Department of Agricultural, Food and Nutritional Science, University of Alberta, Edmonton, Alberta, Canada

Guido J. Hooiveld

Roles Data curation, Formal analysis, Investigation, Methodology, Writing – review & editing

Affiliation: Nutrition, Metabolomics and Genomics group, Division of Human Nutrition, Wageningen University, Wageningen, Gelderland, the Netherlands

Fernando Soberon

Roles Conceptualization, Methodology, Writing – review & editing

Affiliation: Trouw Nutrition USA, Highland, Illinois, United States of America

Harma Berends

Roles Conceptualization, Project administration, Writing – review & editing

Affiliation: Trouw Nutrition R&D, Amersfoort, Utrecht, the Netherlands

Mark V. Boekshoten

Roles Data curation, Formal analysis, Investigation, Methodology, Writing – review & editing

Affiliation: Nutrition, Metabolomics and Genomics group, Division of Human Nutrition, Wageningen University, Wageningen, Gelderland, the Netherlands

Michael E. Van Amburgh

Roles Conceptualization, Funding acquisition, Investigation, Methodology, Supervision, Writing – review & editing

Affiliation: Department of Animal Science, Cornell University, Ithaca, New York, United States of America

Javier Martín-Tereso

Roles Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Writing – review & editing

Affiliation: Trouw Nutrition R&D, Amersfoort, Utrecht, the Netherlands

Michael A. Steele

Roles Conceptualization, Investigation, Project administration, Resources, Supervision, Visualization, Writing – review & editing

* E-mail: [email protected] (LNL); [email protected] (MAS)

Affiliation: Department of Agricultural, Food and Nutritional Science, University of Alberta, Edmonton, Alberta, Canada

ORCID logo http://orcid.org/0000-0001-6941-6205

Introduction

Early life is a critical period in which environmental factors such as nutrition can induce developmental plasticity towards altered biological functions in adulthood. Any event that occurs during this period can permanently influence the physiology of organisms: a process termed ‘programming’ or more recently ‘imprinting’[1]. The concept of early-life nutrition impacting long-term physiological responses in medical research has been well-established but only recently investigated in ruminants [2]. Studies on dairy calves have shown that pre-weaning intake, and thus gain, are positively correlated with lifetime milk production [3,4]. On the other hand, there is also evidence that excessive feeding can be detrimental to animal performance [5,6].

It is important to consider that dairy calves are traditionally restricted in milk or milk-replacer supply to approximately 10% of bodyweight as milk, which represents less than half of their voluntary consumption [7,8,9]. This contrasts with human and rodent early-life studies that typically investigate effects of excessive calorie consumption. More recently, calf studies have focussed on characterizing how milk supply influences metabolism and in particular insulin sensitivity [10,11]. So far, only one study investigated the epigenetic effects of pre-weaning nutrient supply in dairy calves, in which it was found that increased milk feeding induced the expression of genes and transcription factors associated with growth and development in rumen tissue [12].

Adipose tissue is critical in animal development as it participates in several biological processes, including energy balance, endocrine, thermogenic, reproductive, and immune functions [13,14,15]. It has been shown that its molecular regulation and physiology is responsive to diet in adult animals [16,17]. However, it is not clear how early life nutrition affects adipose tissue physiology and how early life adipose tissue modulation can impact the productive and reproductive performance of dairy cows.

Transcriptomic studies of key metabolic tissues offer a unique opportunity to understand how dietary manipulations can affect the physiology and development of farm animals by revealing a comprehensive view of the cellular and molecular regulation at a transcriptome level. We hypothesized that nutrient supply alters adipose tissue development and function in pre-weaning calves. Therefore, the objective of the present study was to determine how lifting the traditional nutrient supply in pre-weaning calves affects adipose tissue transcriptome expression and physiology.

Materials and methods

Experimental design, feeding and housing

The experimental protocols describing the management and animal care have previously been reported [18]. They were reviewed and approved by the Cornell University Animal Care and Use Committee, Ithaca, USA, (Protocol number 2009–0120). In brief, 12 Holstein heifer calves were randomly assigned at birth to one of two treatments. Calves assigned to the enhanced group were fed 1.26 Megajoules (MJ) of metabolizable energy (ME) per kg BW0.75 in three daily feedings. The supply of milk replacer (MR) was adjusted weekly according to changes in BW. Daily amounts of energy supply to each calf increased from 17.58 to 35.17 MJ per day. Calves in the restricted group were fed 11.72 MJ of ME per day in two daily feedings throughout the study, reflecting traditional industry standards. From day 18 onwards, calves were also offered a commercially available concentrate. All calves had access to water and were housed in individual hutches. Information on the composition of milk replacer and concentrate are provided in S1 Table.

Tissue sample collection and preservation

At 54 days of age, calves were fed at 0700h, weighed and transported to the Cornell University abattoir (Ithaca, NY). Calves were harvested approximately 4h after feeding by stunning with a captive bolt, followed directly by exsanguination. Immediately after exsanguination, a representative sample of omental fat was placed in foil bags and snapped frozen on liquid nitrogen. Samples were then transferred into a -80°C freezer until the gene expression profiling analyses were conducted.

RNA isolation

Total RNA was prepared by homogenizing the frozen tissues in TRIzol reagent (ThermoFisher Scientific) using an Ultra-Turrax homogenizer (IKA laboratory technology, Staufen, Germany), and purified using Qiagen RNeasy columns (Qiagen, Holden, Germany). RNA integrity was checked on chip analysis (Agilent 2100 Bioanalyzer, Agilent Technologies, Amsterdam, The Netherlands) according to the manufacturer's instructions. RNA was judged as suitable for array hybridization only if samples exhibited intact bands corresponding to the 18S and 28S ribosomal RNA subunits, and displayed no chromosomal peaks or RNA degradation products (RNA Integrity Number > 7.0).

Affymetrix GeneChip oligoarray hybridization, quality control, and analyses

Total RNA (100 ng) was used for whole transcript cDNA synthesis and terminal labelling using the Ambion WT expression Kit (Life Technologies, Bleiswijk, The Netherlands). Samples were analysed on Affymetrix Bovine Gene 1.1 ST arrays (Affymetrix, Santa Clara, CA). Hybridization, washing, staining and scanning of the arrays were performed automatically on an Affymetrix GeneTitan instrument, according to the manufacturer's instructions. Packages from the Bioconductor project [19], integrated in an online pipeline [20], were used to analyse the array data. Quality of the arrays was determined as described in the S 1 File. Because of insufficient quality of RNA, 2 arrays (1 from each group) were excluded from the final gene expression analysis.

The 553,263 probes on the Bovine Gene ST 1.1 array were redefined utilizing current genome information. In this study, probes were reorganized based on the gene definitions as available in the NCBI database (Bos taurus NCBI Build 6.1; custom CDF version 22), which represented 20,526 unique genes. Normalized expression estimates were obtained from the raw intensity values using the robust multiarray analysis (RMA) preprocessing algorithm available in the library ‘affyPLM’ using default settings [21].

Differentially expressed probe sets were identified using linear models, applying moderated t-statistics that implemented intensity-based empirical Bayes regularization of standard errors (library ‘limma’). The moderated t-test statistic has the same interpretation as an ordinary t-test statistic, except that the standard errors have been moderated across genes (i.e. shrunk to a common value, using a Bayesian model [22,23]). P-values were corrected for multiple testing using a false discovery rate (FDR) method proposed by Storey et al. [24]. Probe sets that satisfied the criterion of FDR < 10% (q-value < 0.10) and p-value < 0.001 were considered to be significantly regulated.

Bioinformatics analysis

A dataset containing 20,526 gene IDs with respective fold change and p-values was uploaded into the Ingenuity Pathway Analysis software (IPA). Core Analysis was performed using default settings with a p < 0.001 and FDR < 0.1 cut-off, which yielded a total 353 mapped genes ready for analysis.

The Downstream Effects Analysis was used to determine which Molecular and Cellular Functions were most relevant to the DE genes dataset (p < 0.05, B-H correction) and which functions were significantly predicted as decreased or increased (z-scores ≤−2 or ≥+2, respectively).

Upstream Regulator Analysis was used to predict which endogenous regulator genes (e.g. transcriptional regulators, cytokines, ligand dependant nuclear receptors, G-protein coupled receptors, growth factors, kinases, transmembrane receptors, and others) control the expression of downstream DE genes by measuring an overlap p-value (p < 0.05) with Fisher’s Exact test. Their predicted activation/inhibition state (z-scores ≤−2 or ≥2) was also verified based on the collective directional expression of their target genes. Pathway Designer tool was used to generate network illustrations derived from Downstream Effects Analysis (Fatty Acid Metabolism and Mitosis).

The Regulator Effects Analysis (customized settings: max. 1 regulator and 1 function per network generated) was used to determine how the predicted active/inhibited Upstream Regulators impacted downstream DE genes and eventually sub functions. The Upstream Regulators, sub functions and their predicted states were used to generate a customized network to directly link Upstream Regulators states with affected functions in the cellular context. Data visualization was created using the Pathway Designer tool.

Results

Differentially expressed genes

When remapping the probes to the BOS TAURUS genome assembly, the expression of a total of 19,968 genes was detected (miRNA genes excluded). Expression varied widely among genes due to treatments from very stable genes with expression fold changes (FC) lower than 1.001 up to genes with FC > 13 such as lipocalin-1 and 3-hydroxybutyrate dehydrogenase 1. However, for most genes (94.48%) expression varied less than 50% (FC < +1.5 or FC > -1.5) (Table 1). Still, 1.98% of the genes (396) were considered differentially expressed, satisfying a cut-off level of p-value < 0.001 and FDR < 0.1.

[Figure omitted. See PDF.]

Table 1. Transcriptome level alterations in adipose tissue of pre-weaned calves under restricted or enhanced feeding regimes.

https://doi.org/10.1371/journal.pone.0201929.t001

Functional analysis of DE genes between restricted and enhanced diet treatments

A total of 353 genes out of 396 DE genes (p < 0.001, FDR < 0.1) identified in this study were successfully mapped to the Ingenuity Knowledge Base (IPA). Of these, 220 were up-regulated and 133 down-regulated. Based on this set of genes, IPA Core Analysis revealed that genes differentially expressed due to dietary treatment were involved (p<0.05) in nine Molecular and Cellular Functions (Fig 1). The analysis showed how DE genes are involved in a variety of biological processes. Their roles included several aspects of adipose cell metabolism, with the top functions ranging from Cell Cycle, which consists mainly of DNA replication and cell division, to Lipid Metabolism, considered the prime function of adipose tissue. The amount of genes involved in each function varied considerably, with the most in Cell Death and Survival (118 genes) and the least in Energy Production (14 genes).

[Figure omitted. See PDF.]

Fig 1. Molecular and cellular functions in DE genes in adipose tissue.

The likelihood of the association between the genes in the dataset and a biological function is represented as–log (p-value), with larger bars being more significant than shorter bars. The vertical line indicates the cut-off for significance (p-value of 0.05).

https://doi.org/10.1371/journal.pone.0201929.g001

Data was further processed through Downstream Effects Analysis in order to predict which biological processes were increased or decreased based on the directional expression of the involved DE genes. A total of eight predictions of sub functions activation states were statistically significant (Z-score > +2). They derived from the following main Molecular and Cellular Functions categories: Lipid Metabolism, Small Molecule Biochemistry, Cellular Transport, Cell Cycle, and Cellular Assembly and Organization (Table 2). Z-scores for sub functions were all positive (Z-score > +2), indicating functional increase. Fatty Acid Metabolism was the only sub function from Lipid Metabolism category that was significantly predicted to be increased and it was the process with the overall strongest activation prediction (Z-score = + 2.605) with a total of 26 DE genes involved. The directional expression of most of these 26 genes indicated that calves under the enhanced nutrient supply had a significant increase in Fatty Acid Metabolism function (Fig 2). This Downstream Effect Analysis result (Fatty Acid Metabolism increase) is based on the Ingenuity knowledge Base that indicated that out of these 26 genes, 15 genes showed a directional expression compatible with an increased function (ABCA7, CAMP, CRAT, CROT, CYP2J2, FASN, FCGR2B, LIPA, NPC2, UCP1, PLA2G5, SCD, SLC13A3, SLC1A1, and VLDLR), 3 genes had expressions compatible with a decreased function (GSN, LCAT, and TRIB3) and 8 gene are involved in this function (ACADSB, CERS6, CPT2, EHHADH, ME1, NQO1, SUCLG1, and THEM4) but literature results, in Ingenuity Knowledge Base, are not consistent whether they increase or decrease it. (Fig 2).

[Figure omitted. See PDF.]

Fig 2. Effects of DE genes in Fatty Acid Metabolism.

Values under gene symbols represent direction expression of DE genes (fold change). Relationship lines in orange indicate that genes showed a directional expression compatible with an increased function. Yellow lines indicate that the gene expression is compatible with a decreased function. Gray lines mean that the genes are involved in this function, but literature results in Ingenuity Knowledge Base do not indicate whether they increase or decrease Fatty Acid Metabolism.

https://doi.org/10.1371/journal.pone.0201929.g002

[Figure omitted. See PDF.]

Table 2. Biological functions predicted to be activated by Downstream Effects Analysis and DE genes involved.

https://doi.org/10.1371/journal.pone.0201929.t002

Cell Cycle was the functional category most represented, with four sub functions significantly predicted to be increased: Interphase (30 DE genes), Cell Cycle Progression (43 DE genes), Entry into Interphase (10 DE genes), and Mitosis (23 DE genes). Ingenuity knowledge Base indicated that out of these 23 DE genes involved in Mitosis, 11 genes showed a directional expression compatible with an increased function (BUB1B, ABL1, CCNE1, KIF11, STAT6, IGFBP5, FGF1, NUSAP1, EDN3, TIMP2, and CDK1), 2 genes (CDC20 and MET) had expressions compatible with a decreased function, and 10 genes (MKI67, CKAP2, DLGAP5, CENPA, CCNB1, AATF, TCERG1, ANLN, CENPF, and CENPE) were involved in this function but literature results, in Ingenuity Knowledge Base, do not indicate whether they increase or decrease Mitosis (Fig 3).

[Figure omitted. See PDF.]

Fig 3. Effects of DE genes in Mitosis.

Values under gene symbols represent direction expression of DE genes (fold change). Relationship lines in orange indicate that genes showed a directional expression compatible with an increased function. Yellow lines indicate that the gene expression is compatible with a decreased function. Gray lines mean that the genes are involved in this function, but literature results in Ingenuity Knowledge Base do not indicate whether they increase or decrease Fatty Acid Metabolism.

https://doi.org/10.1371/journal.pone.0201929.g003

Upstream regulators and effects

Upstream Regulator Analysis identified a total of 346 molecules as potential upstream regulators for the set of 353 DE genes, including transcriptional regulators, cytokines, ligand dependant nuclear receptors, G-protein coupled receptors, growth factors, kinases, transmembrane receptors, and others, but excluding drugs and chemicals. However, the analysis was able to reveal the activation status of only 23 Upstream Regulators based on the directional expression of their controlled DE genes (p-value of overlap < 0.05). Most of them (13 genes) were predicted to be activated (PPARGC1A, CSF2, PPARG, AHR, FOXM1, PTGER2, FGF21, KLF15, RABL6, E2F3, MED1, PPARA, and SMOC2) with a Z-score higher than +2, while 10 were predicted to be inhibited (RICTOR, MAP4K4, INSIG1, LGR4, LGR5, SMARCE1, SMARCB1, NUPR1, TCF3, and BNIP3L) with a Z-score lower than -2 (Table 3). This set of Upstream Regulators accounted for the molecular regulation of 108 out of 353 DE genes identified in this study. These transcriptional factors varied on how they influence DE genes regulation, from 4 DE genes by SMOC2 up to 22 DE genes by PPARG (Table 3).

[Figure omitted. See PDF.]

Table 3. Predicted activation status of upstream regulators that control differentially expressed genes and their targets.

https://doi.org/10.1371/journal.pone.0201929.t003

A total of 17 upstream regulators were further analyzed through IPA’s Regulator Effects feature in order to reveal not only what genes they are likely to control, but ultimately what functions the identified Upstream Regulators are affecting (Fig 4). By combining the activation status of the Upstream Regulators with the directional expression change of the genes they regulate and Ingenuity Knowledge Base information, this analysis provided evidence indicating that TCF3, SMARCE1, RABL6, BNIPL3, and PTGER2 were predicted to be involved in the regulating of Cell Cycle Progression. FOXM1, MED1, SMARCB1, NUPR1, and CSF2 were involved in Entry into Interphase. INSIG1, KLF15, and PPARGC1A were regulating Fatty Acid Metabolism. E2F3 was regulating Interphase related functions. SMOC2 was regulating Mitosis. PPARG was involved in Transport of Carboxylic Acid. AHR was predicted to be involved in Fibrosis of Liver aspects, however, this result was not included as samples were derived from adipose tissue. PPARA did not appear in any result of Regulator Effects feature (Fig 4).

[Figure omitted. See PDF.]

Fig 4. Predicted impact of upstream regulators controlling DE genes on adipose tissue biological functions.

Only upstream regulators with overlap p-value < 0.05 and Z-score > +2 or < -2 are displayed.

https://doi.org/10.1371/journal.pone.0201929.g004

Discussion

Adipose tissue is a dynamic organ and plays a crucial role in energy balance. Several factors have been reported to impact the molecular regulation of adipose tissue in bovine such as age [25], depot location [17] and diet [16]. However, it is not clear how early life dietary manipulation can affect adipose tissue transcriptome and function. In the present study, two feeding regimes were used, one simulating a restricted nutrient supply, a practice that is still used by part of the dairy industry, while an alternative feeding regimen provided an enhanced nutrient supply (higher amount of milk replacer). Calves in the first group were heavily restricted in their caloric intake over maintenance–based on NRC 2001 [26]–compared with calves under the enhanced diet (3.77 MJ/day vs. 15.91 MJ/day, respectively).

As previously reported [18], calves in the enhanced nutrient supply group had higher total milk replacer intakes than restricted calves (69.5 and 32.6 kg dry matter, respectively) and were larger at harvest at 54±2 days old (83.2 and 61.0 kg, respectively). The large difference in nutrient supply affected not only the overall body development of calves, but also increased (p<0.01) mammary gland and liver growth, as a percentage of body weight [18].

In this manuscript we further show that an enhanced nutrient supply also alters adipose tissue molecular regulation. The dietary treatment altered the adipose transcriptome molecular regulation as gene expression changed by more than 1.5 FC for more than a thousand genes (Table 1). Adipose tissue seems to be highly responsive to dietary treatments. Transcriptome data from the same animals indicate that adipose tissue genes were more affected by the treatment compared to muscle and bone marrow[27]. Also, in a study in which pre-weaning calves were fed a restricted vs. an ad libitum milk replacer supply, transcriptome changes in the jejunum were more discrete than the changes observed in the present study, as 244 genes were differentially expressed (p-value < 0.001) [28] while at the same cut-off adipose tissue in our study had 396 genes differentially expressed. However, samples in our experiment were collected at 54 days, while jejunum samples were collected only at 80 days of age, four weeks after our experimental diets.

White adipocytes are considered the main cellular type in adipose tissue; however this organ is also composed of other cells such as preadipocytes, brown adipocytes, beige adipocytes, macrophages, endothelial cells and stem cells [29]. Therefore, the results of the functional analysis on DE genes in this study reflect not only white adipocyte cells, but also the biology of the whole organ. Besides, each adipose tissue depot has a distinctive physiology and specific molecular regulation [17,30]. Omental fat, for example, is considered a visceral depot. However; unlike other major fat depots (e.g. subcutaneous and perirenal fat) it contains lymphoid aggregates, known as milky spots, which allow omental fat tissue to play a role in peritoneal immunity [31]. The functional analysis revealed DE genes in omental fat where involved in a variety of cellular processes (Fig 1). Cell Cycle appeared as the most significant function for the DE genes set. This function is related to the process of cell reproduction. Adipose tissue development happens through two distinct processes: hypertrophy (increase in cell size) and hyperplasia (increase in cell number) [32]. The later process corresponds to the enriched Cell Cycle function. Not surprisingly, Lipid Metabolism appeared as one of the top functions, as adipose tissue is the premier body location for lipid processing.

Downstream Effects Analysis (Table 2) further revealed the predicted activation state of sub functions based on the directional expression of DE genes involved. Fatty Acid Metabolism, which belongs to the Lipid Metabolism category, was the function with the highest predicted activation state (z-score). This finding is based on the directional expression of the 26 DE genes involved (Fig 2), which indicates that adipose tissue of calves under an enhanced nutrient supply are processing fatty acids at a higher rate compared to restricted animals. Another study with similar treatments reported that the gene network ‘energy production, lipid metabolism and small molecule biochemistry network’ was enriched in subcutaneous adipose tissue of calves (126 day old) fed an enhanced diet [33]. Older animals (197-day-old Angus steers) under an enhanced nutrient supply (high concentrate diet vs. pasture) also displayed altered transcriptome expression in subcutaneous adipose tissue (1113 DE genes), indicating increased lipid metabolism functions, including Fatty Acid Metabolism [34].

A higher nutrient intake allows adipocytes to increase fat stores by either an increased uptake of free fatty acids or de novo lipogenesis [32,35]. This condition provides the requirements for an increased adipose tissue expansion. Downstream Effects Analysis (Table 2) further showed that four sub functions belonging to Cell Cycle category were predicted to be increased: Entry into Interphase, Interphase, Cell Cycle Progression, and Mitosis. These results indicate that omental fat in the calves under the enhanced nutrient supply experienced an increased tissue expansion through hyperplasia. However, hypertrophy might have also contributed to adipose tissue development since the sub functions Development of Cytoplasm and Formation of Cytoskeleton (Cellular Assembly and Organization category) were predicted to be increased. These findings are evidence that early life nutrient supply influences the genetic regulation of hyperplastic and hypertrophic processes in calves’ adipose tissue. Our findings are supported by another study investigating low vs. high plane of nutrition in calves’ subcutaneous adipose tissue histology and transcriptome [33]. Histology results indicated that at 126 days of age, calves in a low plane of nutrition displayed less adiposity (lower quantity of adipocytes, smaller adipocyte diameter and more preadipocytes) when compared to calves in a high plane of nutrition [33].

In another study, pre-weaned Friesian calves fed an enhanced vs. restricted nutrient supply displayed increased adipose cell hypertrophy, but hyperplasia did not differ at 95 days; however after weaning treatment diets were paired until slaughter at 553 days where no differences in adipose cellularity were observed [36], which might represent a compensatory developmental mechanism.

The regulation of gene expression in bovine adipose tissue is complex. There is a multilayered system with different control levels to ensure that gene expression is fine tuned, including epigenetic modifications (e.g DNA methylation and histone modifications) [1], transcriptional regulation (e.g transcriptional factors) [37] and post-transcriptional regulation (microRNAs) [38]. Transcriptional factors are major players in regulating the expression of downstream genes. However, their expression (mRNA level) is not a safe indication of its activity since their activity is mainly controlled by a post-translational modification such as protein phosphorylation or ligand binding [39]. The Upstream Regulator Analysis identified a high amount of potential upstream regulators for the DE genes, including not only transcriptional factors, but also cytokines, G-protein coupled receptors and others (n = 436). However, only 23 Upstream Regulators obtained a significant prediction of activation or inhibition considering the expression profile of the DE genes they control (Table 3). Even though none of these 23 upstream regulators were differentially expressed, they are likely to orchestrate a significant part of the effects of the enhanced nutrient supply on adipose tissue regulation, since 108 DE genes were predicted to be under their control.

The combination of customized Regulator Effects analysis results with a Network construction with exclusion of regulated genes provided a simplified view of the impact of the dietary intervention on Upstream Regulators and consequentially on Molecular and Cellular functions of adipose cells (Fig 4). The dietary effects can be grouped mainly into Lipid Metabolism and Cell Cycle functions. PPARG, PPARGC1A, KLF15 and INSIG1 are known to affect not only Lipid Metabolism but also adipogenesis. PPARG is the master regulator of adipogenesis and regulates several aspects of Lipid Metabolism [37,40], such as the Transport of Carboxylic Acid, which comprise fatty acids. This transcription factor has binding sites in most lipid metabolism genes [41]. Its predicted activated state is an indicator that adipose tissue is developing and expanding through adipogenesis and lipogenesis. The transcriptional factor with the highest activation score was another lipid metabolism gene, PPARGC1A, which is a co-activator of PPARG and is involved in Fatty Acid Metabolism regulation as well as in the activation of adaptive thermogenesis [42]. KLF15 was also a transcription factor predicted to be active that is involved in adipogenesis and Lipid Metabolism aspects. The deletion of this gene in adipose tissue was shown to decrease adiposity and mechanistic studies indicate it regulates lipogenesis genes and inhibits lipolysis genes in adipocytes [43]. INSIG1 was the only Upstream Regulator involved in Lipid Metabolism aspects that was predicted to be inhibited. This gene was reported to inhibit lipogenesis and preadipocyte differentiation [44].

The other Upstream Regulators (BNIPL3, CSF2, E2F3, FOXM1, MED1, NUPR1, PTGER2, RABL6, SMARCB1, SMARCE1, SMOC2, and TCF3) were affecting functions related to Cell Cycle (Entry in Interphase, Interphase, Cell Cycle Progression, and Mitosis) (Fig 4). Three transcriptional factors predicted as active (MED1, FOXM1, and E2F3) are known to be involved in the regulation of Interphase steps. E2F3 activates genes that control the rate of proliferation of cells [45] and its immunodepletion inhibits the induction of the S phase (stage of chromosomes replication) [46]. Similarly, FOXM1 induces the expression of several genes that regulate G1/S transition and DNA replication [47], and moreover, influences G2/M transition and Mitosis [48]. Other transcriptional factors (NUPR1, SMARCE1, SMARCEB1, and TCF3) affecting Cell Cycle functions were predicted as inhibited. TCF3, for example, is a transcriptional factor that inhibits the Wnt target genes [49], and in somatic cells, the activation of Wnt pathway induce cell proliferation [50,51]. In addition to transcriptional factors, other categories of Upstream Regulators located outside the nucleus have also been reported to stimulate Cell Cycle Progression such as RABL6 in cytoplasm [52] and PTGER2 in the plasma membrane [53].

The results of this study revealed that an early life enhanced diet changes adipose tissue physiology by increasing the metabolism of fatty acids and cell proliferation. Fat mass of calves from enhanced diet was likely to be increased as Leptin expression was higher (S2 Table). Another study also found that calves supplied with an enhanced diet display increased Letpin levels compared to calves in a restricted diet [33] and it is known that Leptin expression correlates positively with fat mass [54,55] and. However, the expression of Tumor Necrosis Factor α (TNF) and Interleukin 6 (IL6), which are pro-inflammatory cytokines [56,57], did not differ (S2 Table), indicating that although the amount of adipose tissue of calves under an enhanced diet was higher, this organ was healthy, not inducing inflammation and its consequences. Underfeeding or overfeeding animals is a concern. A sustained energy imbalance with deficient or excessive energy intake compared to energy expenditure, may lead to abnormal amounts of adipose tissue and health problems. Excessive fat mass can lead to local and systemic health problems such as hypoxia in adipose tissue cells [58], inflammation and insulin resistance [59,60]. In adult cows it may also lead to fertility [5] and metabolic problems during lactation [6]. Conversely, a restricted calorie intake in young animals may lead not only to a reduced fat mass, but also to a delayed immune system maturation [61] and even a lower milk yield at adult life [3].

An interesting finding that might represent another functional difference between calves treated with an enhanced nutrient supply was the increased expression of Uncoupling Protein 1 (UCP1) (S2 Table). This mitochondrial membrane protein is expressed only in brown and beige adipocytes [62,63], which are likely to be part of the omental adipose tissue of calves in this experiment. Beige adipocytes are inducible and widely distributed in various white adipose tissues such as subcutaneous, epididymal and mesenteric depots [64]. Like brown adipocytes, beige adipocytes have been shown to dissipate chemical energy in the form of heat and play an important role in non-shivering thermoregulation and glucose/insulin sensitivity [63,65]. Uncoupling Protein 1 expression is controlled by the transcriptional factor PPARGC1A, which was predicted to be activated. PPARGC1A induces the adaptive thermogenesis program responsible for turning chemical energy into heat (non-shivering thermogenesis) [40,66] and UCP1 is the key player in this processes [67]. However, it is not clear whether this finding indicates that calves fed an enhanced diet developed an increased thermoregulation control through non-shivering thermogenesis or if this gene expression response was a diet induced thermogenesis phenomenon in which the purpose was not thermoregulation, but to regulate the amount of lipid stores through heat generation to control body weight [68].

In conclusion, adipose tissue of pre-weaning female calves is developmentally and metabolically responsive to early life nutrient supply (enhanced vs. restricted). However, it is still not clear if it imprints adipose physiology and under what conditions it positively or negatively impact future animal performance. Additional studies considering both growing and productive stages will lead to a better understanding on how neonatal nutrient supply alters long term adipose function and how that relates to animal performance.

Supporting information

[Figure omitted. See PDF.]

S1 Table. Milk replacer and starter grain chemical composition as reported by manufacturer.

https://doi.org/10.1371/journal.pone.0201929.s001

(DOCX)

S2 Table. Overall genes expression.

https://doi.org/10.1371/journal.pone.0201929.s002

(XLSX)

S1 File. Array quality.

https://doi.org/10.1371/journal.pone.0201929.s003

(DOCX)

Citation: Leal LN, Romao JM, Hooiveld GJ, Soberon F, Berends H, Boekshoten MV, et al. (2018) Nutrient supply alters transcriptome regulation in adipose tissue of pre-weaning Holstein calves. PLoS ONE 13(8): e0201929. https://doi.org/10.1371/journal.pone.0201929

References

1. Mazzio EA, Soliman KFA (2014) Epigenetics and Nutritional Environmental Signals. Integrative and Comparative Biology 54: 21–30. pmid:24861811

2. Fisher AG, Brockdorff N (2012) Epigenetic memory and parliamentary privilege combine to evoke discussions on inheritance.

3. Soberon F, Raffrenato E, Everett RW, Van Amburgh ME (2012) Preweaning milk replacer intake and effects on long-term productivity of dairy calves. Journal of Dairy Science 95: 783–793. pmid:22281343

4. Soberon F, Van Amburgh ME (2013) Lactation Biology Symposium: The effect of nutrient intake from milk or milk replacer of preweaned dairy calves on lactation milk yield as adults: a meta-analysis of current data. J Anim Sci 91: 706–712. pmid:23296823

5. Adamiak SJ, Mackie K, Watt RG, Webb R, Sinclair KD (2005) Impact of Nutrition on Oocyte Quality: Cumulative Effects of Body Composition and Diet Leading to Hyperinsulinemia in Cattle. Biology of Reproduction 73: 918–926. pmid:15972884

6. Rukkwamsuk T, Kruip TAM, Wensing T (1999) Relationship between overfeeding and overconditioning in the dry period and the problems of high producing dairy cows during the postparturient period. Veterinary Quarterly 21: 71–77. pmid:10427630

7. Jasper J, Weary DM (2002) Effects of ad libitum milk intake on dairy calves. J Dairy Sci 85: 3054–3058. pmid:12487471

8. Kertz AF, Prewitt LR, Everett JP Jr. (1979) An Early Weaning Calf Program: Summarization and Review. Journal of Dairy Science 62: 1835–1843.

9. Berends H (2014) Nutrient utilization, dietary preferences, and gastrointestinal development in veal calves: interactions between solid feed and milk replacer. Wageningen: Wageningen University.

10. Yunta C, Terré M, Bach A (2015) Short- and medium-term changes in performance and metabolism of dairy calves offered different amounts of milk replacers. Livestock Science 181: 249–255.

11. MacPherson JAR, Berends H, Leal LN, Cant JP, Martin-Tereso J, et al. (2016) Effect of plane of milk replacer intake and age on glucose and insulin kinetics and abomasal emptying in female Holstein Friesian dairy calves fed twice daily. J Dairy Sci 99: 8007–8017. pmid:27522426

12. Naeem A, Drackley JK, Lanier JS, Everts RE, Rodriguez-Zas SL, et al. (2014) Ruminal epithelium transcriptome dynamics in response to plane of nutrition and age in young Holstein calves. Funct Integr Genomics 14: 261–273. pmid:24318765

13. Choe SS, Huh JY, Hwang IJ, Kim JI, Kim JB (2016) Adipose Tissue Remodeling: Its Role in Energy Metabolism and Metabolic Disorders. Frontiers in Endocrinology 7:30. pmid:27148161

14. Poulos SP, Hausman DB, Hausman GJ (2010) The development and endocrine functions of adipose tissue. Molecular and Cellular Endocrinology 323: 20–34. pmid:20025936

15. Perez-Perez A, Sanchez-Jimenez F, Maymo J, Duenas JL, Varone C, et al. (2015) Role of leptin in female reproduction. Clin Chem Lab Med 53: 15–28. pmid:25014521

16. Romao JM, Jin W, He M, McAllister T, Guan LL (2012) Altered microRNA expression in bovine subcutaneous and visceral adipose tissues from cattle under different diet. PLoS ONE 7: e40605. pmid:22815773

17. Romao JM, Jin W, He M, McAllister T, Guan LL (2013) Elucidation of Molecular Mechanisms of Physiological Variations between Bovine Subcutaneous and Visceral Fat Depots under Different Nutritional Regimes. PLoS One 8: e83211. pmid:24349463

18. Soberon F, Van Amburgh ME (2017) Effects of preweaning nutrient intake in the developing mammary parenchymal tissue. Journal of Dairy Science 100: 4996–5004. pmid:28390726

19. Gentleman RC, Carey VJ, Bates DM, Bolstad B, Dettling M, et al. (2004) Bioconductor: open software development for computational biology and bioinformatics. Genome Biol 5: 2004–2005.

20. Lin K, Kools H, de Groot PJ, Gavai AK, Basnet RK, et al. (2011) MADMAX—Management and analysis database for multiple ~omics experiments. J Integr Bioinform 8: 2011–2160.

21. Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, et al. (2003) Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 4: 249–264. pmid:12925520

22. Smyth GK (2004) Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Stat Appl Genet Mol Biol 3: 1544–6115.

23. Sartor MA, Tomlinson CR, Wesselkamper SC, Sivaganesan S, Leikauf GD, et al. (2006) Intensity-based hierarchical Bayes method improves testing for differentially expressed genes in microarray experiments. BMC Bioinformatics 7: 1471–2105.

24. Storey JD, Tibshirani R (2003) Statistical significance for genomewide studies. Proceedings of the National Academy of Sciences 100: 9440–9445.

25. Romao JM, He ML, McAllister TA, Guan LL (2014) Effect of age on bovine subcutaneous fat proteome: Molecular mechanisms of physiological variations during beef cattle growth1. Journal of Animal Science 92: 3316–3327. pmid:24894005

26. NRC (2001) Nutrient requirements of dairy cattle. 7th Revised Edition, Subcommittee on Dairy Cattle Nutrition, Committee on Animal Nutrition, Board on Agriculture and Natural Resources. National Academy Press 2001, 7th Revised Edition.

27. Leal LN, Hooiveld GJ, Soberon F, Berends H, Boekschoten MV, et al. (2017) Pre-weaning nutrient supply affects gene expression profiles in bone marrow and muscle in calves. In: committee Es, editor. 68th Annual Meeting of the European Federation of Animal Science. Tallinn, Estonia: Wageningen Academic Publishers. pp. 447.

28. Hammon HM, Frieten D, Gerbert C, Koch C, Dusel G, et al. (2018) Different milk diets have substantial effects on the jejunal mucosal immune system of pre-weaning calves, as demonstrated by whole transcriptome sequencing. Scientific Reports 8: 1693. pmid:29374218

29. Cinti S (2005) The adipose organ. Prostaglandins, Leukotrienes and Essential Fatty Acids 73: 9–15.

30. Gerhard GS, Styer AM, Strodel WE, Roesch SL, Yavorek A, et al. (2013) Gene expression profiling in subcutaneous, visceral and epigastric adipose tissues of patients with extreme obesity. International Journal of Obesity 38: 371. pmid:23949615

31. Meza-Perez S, Randall TD (2017) Immunological Functions of the Omentum. Trends in Immunology 38: 526–536. pmid:28579319

32. Rutkowski JM, Stern JH, Scherer PE (2015) The cell biology of fat expansion. The Journal of Cell Biology 208: 501–512. pmid:25733711

33. English A-M, Waters SM, Cormican P, Byrne CJ, Fair S, et al. (2018) Effect of early calf-hood nutrition on the transcriptomic profile of subcutaneous adipose tissue in Holstein-Friesian bulls. BMC Genomics 19: 281. pmid:29690861

34. Joseph SJ, Robbins KR, Pavan E, Pratt Sl, Duckett SK, et al. (2010) Effect of Diet Supplementation on the Expression of Bovine Genes Associated with Fatty Acid Synthesis and Metabolism. Bioinformatics and Biology Insights 2010: 19.

35. Laliotis GP, Bizelis I, Rogdakis E (2010) Comparative Approach of the de novo Fatty Acid Synthesis (Lipogenesis) between Ruminant and Non Ruminant Mammalian Species: From Biochemical Level to the Main Regulatory Lipogenic Genes. Current Genomics 11: 168–183. pmid:21037855

36. Robelin J, Chilliard Y (1989) Short-Term and Long-Term Effects of Early Nutritional Deprivation on Adipose Tissue Growth and Metabolism in Calves. Journal of Dairy Science 72: 505–513. pmid:2703573

37. White UA, Stephens JM (2010) Transcriptional factors that promote formation of white adipose tissue. Molecular and Cellular Endocrinology 318: 10–14. pmid:19733624

38. Romao JM, Jin W, Dodson MV, Hausman GJ, Moore SS, et al. (2011) MicroRNA regulation in mammalian adipogenesis. Experimental Biology and Medicine 236: 997–1004. pmid:21844119

39. Calkhoven CF, Ab G (1996) Multiple steps in the regulation of transcription-factor level and activity. Biochemical Journal 317: 329–342. pmid:8713055

40. Medina-Gomez G, Gray S, Vidal-Puig A (2007) Adipogenesis and lipotoxicity: role of peroxisome proliferator-activated receptor gamma (PPARgamma) and PPARgammacoactivator-1 (PGC1). Public Health Nutr 10: 1132–1137. pmid:17903321

41. Siersbæk R, Nielsen R, Mandrup S (2010) PPARγ in adipocyte differentiation and metabolism–Novel insights from genome-wide studies. FEBS Letters 584: 3242–3249. pmid:20542036

42. Liang H, Ward WF (2006) PGC-1α: a key regulator of energy metabolism. Advances in Physiology Education 30: 145–151. pmid:17108241

43. Matoba K, Lu Y, Zhang R, Chen ER, Sangwung P, et al. (2017) Adipose KLF15 Controls Lipid Handling to Adapt to Nutrient Availability. Cell Reports 21: 3129–3140. pmid:29241541

44. Li J, Takaishi K, Cook W, McCorkle SK, Unger RH (2003) Insig-1 “brakes” lipogenesis in adipocytes and inhibits differentiation of preadipocytes. Proceedings of the National Academy of Sciences 100: 9476–9481.

45. Humbert PO, Verona R, Trimarchi JM, Rogers C, Dandapani S, et al. (2000) E2f3 is critical for normal cellular proliferation. Genes & Development 14: 690–703.

46. Leone G, DeGregori J, Yan Z, Jakoi L, Ishida S, et al. (1998) E2F3 activity is regulated during the cell cycle and is required for the induction of S phase. Genes & Development 12: 2120–2130.

47. Kalin TV, Ustiyan V, Kalinichenko VV (2011) Multiple faces of FoxM1 transcription factor: lessons from transgenic mouse models. Cell Cycle 10: 396–405. pmid:21270518

48. Costa RH (2005) FoxM1 dances with mitosis. Nature Cell Biology 7: 108. pmid:15689977

49. Cole MF, Johnstone SE, Newman JJ, Kagey MH, Young RA (2008) Tcf3 is an integral component of the core regulatory circuitry of embryonic stem cells. Genes & Development 22: 746–755.

50. Holland JD, Klaus A, Garratt AN, Birchmeier W (2013) Wnt signaling in stem and cancer stem cells. Current Opinion in Cell Biology 25: 254–264. pmid:23347562

51. Niehrs C, Acebron SP (2012) Mitotic and mitogenic Wnt signalling. Embo J 31: 2705–2713. pmid:22617425

52. Hagen J, Muniz VP, Falls KC, Reed SM, Taghiyev AF, et al. (2014) RABL6A Promotes G1–S Phase Progression and Pancreatic Neuroendocrine Tumor Cell Proliferation in an Rb1-Dependent Manner. Cancer Research 74: 6661–6670. pmid:25273089

53. Lee J, Banu SK, Rodriguez R, Starzinski-Powitz A, Arosh JA (2010) Selective blockade of prostaglandin E2 receptors EP2 and EP4 signaling inhibits proliferation of human endometriotic epithelial cells and stromal cells through distinct cell cycle arrest. Fertil Steril 93: 2498–2506. pmid:20207353

54. Foote AP, Tait RG Jr., Keisler DH, Hales KE, Freetly HC (2016) Leptin concentrations in finishing beef steers and heifers and their association with dry matter intake, average daily gain, feed efficiency, and body composition. Domest Anim Endocrinol 55: 136–141. pmid:26851619

55. Geary TW, McFadin EL, MacNeil MD, Grings EE, Short RE, et al. (2003) Leptin as a predictor of carcass composition in beef cattle. J Anim Sci 81: 1–8.

56. Sindhu S, Thomas R, Shihab P, Sriraman D, Behbehani K, et al. (2015) Obesity Is a Positive Modulator of IL-6R and IL-6 Expression in the Subcutaneous Adipose Tissue: Significance for Metabolic Inflammation. PLoS One 10: e0133494. pmid:26200663

57. Cawthorn WP, Sethi JK (2008) TNF-α and adipocyte biology. FEBS Letters 582: 117–131. pmid:18037376

58. Hosogai N, Fukuhara A, Oshima K, Miyata Y, Tanaka S, et al. (2007) Adipose Tissue Hypoxia in Obesity and Its Impact on Adipocytokine Dysregulation. Diabetes 56: 901–911. pmid:17395738

59. Shoelson SE, Herrero L, Naaz A (2007) Obesity, Inflammation, and Insulin Resistance. Gastroenterology 132: 2169–2180. pmid:17498510

60. Mraz M, Haluzik M (2014) The role of adipose tissue immune cells in obesity and low-grade inflammation. J Endocrinol 222: 14–0283.

61. Walford RL, Liu RK, Gerbase-Delima M, Mathies M, Smith GS (1973) Longterm dietary restriction and immune function in mice: Response to sheep red blood cells and to mitogenic agents. Mechanisms of Ageing and Development 2: 447–454. pmid:4599309

62. Palou A, Picó C, Bonet ML, Oliver P (1998) The uncoupling protein, thermogenin. The International Journal of Biochemistry & Cell Biology 30: 7–11.

63. Shabalina IG, Petrovic N, de Jong JM, Kalinovich AV, Cannon B, et al. (2013) UCP1 in brite/beige adipose tissue mitochondria is functionally thermogenic. Cell Rep 5: 1196–1203. pmid:24290753

64. Frontini A, Cinti S (2010) Distribution and development of brown adipocytes in the murine and human adipose organ. Cell Metab 11: 253–256. pmid:20374956

65. Cannon B, Nedergaard J (2004) Brown adipose tissue: function and physiological significance. Physiol Rev 84: 277–359. pmid:14715917

66. Elabd C, Chiellini C, Carmona M, Galitzky J, Cochet O, et al. (2009) Human multipotent adipose-derived stem cells differentiate into functional brown adipocytes. Stem Cells 27: 2753–2760. pmid:19697348

67. Crichton PG, Lee Y, Kunji ERS (2017) The molecular features of uncoupling protein 1 support a conventional mitochondrial carrier-like mechanism. Biochimie 134: 35–50. pmid:28057583

68. Rothwell NJ, Stock MJ (1997) A role for brown adipose tissue in diet-induced thermogenesis. Obes Res 5: 650–656. pmid:9449154