Bovine hepatic miRNAome profiling and differential miRNA expression analyses between beef steers with divergent feed efficiency phenotypes

The study addresses how hepatic microRNAs may contribute to genetic and molecular regulation of feed efficiency in beef cattle, measured by residual feed intake (RFI). Feed costs comprise up to 75% of variable production expenses, and improving feed efficiency can also reduce methane emissions and environmental impact. Prior genome-wide association and transcriptomic studies have identified genes and pathways related to feed efficiency, and liver tissue has been implicated through differential gene expression associated with lipid, energy, carbohydrate, and amino acid metabolism. MicroRNAs (~22 nt) are key post-transcriptional regulators that can affect up to 60% of mammalian transcripts and are involved in cell proliferation, cell cycle, apoptosis, and nutrient metabolism. In the liver, miRNAs regulate hepatocyte proliferation, energy metabolism, detoxification, and nutrient metabolism. However, miRNA studies related to feed efficiency in beef cattle are limited. The purpose of this study was to profile the hepatic miRNAome across three beef breeds (Angus, Charolais, and Kinsella Composite) and identify miRNAs associated with RFI by differential expression analysis, then predict target genes and enriched biological functions to elucidate regulatory mechanisms.

The authors summarize prior work indicating: (1) feed efficiency is economically and environmentally important; (2) multiple GWAS and transcriptomic studies have linked hepatic pathways (lipid, energy, carbohydrate, amino acid metabolism) to feed efficiency; (3) miRNA biogenesis involves DROSHA and DICER, miRISC formation with Argonaute, and predominant binding to 3' UTR seed sites leading to downregulation; (4) miRNAs regulate many biological functions and a substantial portion of the transcriptome; (5) in liver, miRNAs modulate cell proliferation, detoxification, energy and nutrient metabolism. Previous miRNA studies on feed efficiency in cattle are limited, with some reports in Angus and Nelore, and results not fully concordant across populations. This sets the context for a comprehensive hepatic miRNAome profiling and DE analysis across breeds with divergent RFI.

Animals and phenotypes: 256 steers (Angus, Charolais, Kinsella Composite) raised under standardized management at the Roy Berg Kinsella Ranch (University of Alberta). Individual feed intake measured via GrowSafe over ~70–73 days on a finishing diet; DMI standardized to 12 MJ ME/kg DM. Body weights recorded serially for ADG; metabolic mid-weight computed; RFI calculated as actual standardized DMI minus expected DMI from regression on ADG and metabolic mid-weight. From these, 60 steers (20 per breed) were selected for liver miRNA sequencing; among them, 12 per breed (6 low-RFI, 6 high-RFI) were used for DE analysis (selected to maximize divergence). Tissue collection: liver sampled post-slaughter (standardized location), snap frozen and stored at −80°C. RNA extraction: total RNA including small RNAs extracted (Qiagen RNeasy Plus Universal), quantified (NanoDrop), quality assessed (TapeStation), all RIN > 8. Library prep and sequencing: Illumina TruSeq Small RNA Library Prep from 1 µg total RNA; adapters ligated, reverse transcribed, indexed PCR, gel-purified (145–160 bp), pooled, and sequenced on Illumina HiSeq 2500 (single-end 1x50 bp). Data processing: Quality control with FastQC; adapter trimming with cutadapt; reads <15 bp or >28 bp removed; reads aligning to rRNA, tRNA, snRNA, snoRNA removed using Rfam/RNAcentral references; post-QC quality reassessed. miRNA profiling: miRDeep2 v2.0.0.8 with Bowtie 1.1.1; mapped to UMD3.1 (Ensembl v93); known bovine/human mature miRNAs and precursors from miRBase v22 used to quantify known miRNAs; miRDeep2 used to predict novel miRNAs (hairpin structure via ViennaRNA RNAfold, mirDeep2 score, Randfold P-values). DE analysis: For each breed, miRNAs with <10 total counts across all 20 samples filtered out; for DE, considered all expressed known miRNAs and top 25 expressed novel miRNAs; retained miRNAs with ≥1 CPM in ≥6 of 12 samples; TMM normalization; edgeR GLM under negative binomial with high-RFI as reference; DE threshold P < 0.05 and fold-change > 1.5; additional stringent reporting with Bonferroni FDR < 0.05. Validation: TaqMan qPCR for selected DE miRNAs (Charolais: miR-2415-3p, miR-133a, miR-2419-5p; KC: miR-424-5p, miR-223, miR-155), with stable endogenous controls (KC: bta-miR-2284x, bta-let-7b; Charolais: bta-miR-192, bta-miR-93). Amplification efficiencies calculated; reactions run in triplicate; differential expression assessed by t-test; concordance with RNAseq evaluated by Pearson correlation. Target prediction and functional enrichment: TargetScan 7.0 scripts used to predict conserved/non-conserved targets and compute context++ scores (14 features, with RNAplfold for structural accessibility). Targets selected at ≥99th percentile context++ score (and also >50th percentile for overlap with prior DE genes). Ingenuity Pathway Analysis (IPA Spring 2019) used for functional enrichment (biological functions and canonical pathways). Correlations between miRNA and mRNA expression were computed where mRNA data from the same animals were available. Network analysis: Cytoscape 3.7.1 used to visualize DE miRNA–DE mRNA interaction networks per breed using targets with context++ percentile >50.

Sequencing and alignment: On average >9 million raw reads per sample for Angus and Charolais, >11 million for KC. After trimming and filtering, an average of 5.5 million reads/sample retained for miRNA profiling; most reads 20–24 bp (mean 21 bp); high Phred quality scores. Mean genome mapping rate 74.78% (SD 2.31%): Charolais 72.47% (SD 1.2%), KC 77.09% (SD 1.34%). Known miRNAs: Identified 541 (Angus), 551 (Charolais), 575 (KC) expressed known miRNAs; 528 unique known miRNAs (~90%) common across breeds. Ten miRNAs dominated expression across breeds (accounting for ~78% of aligned reads): bta-miR-192 (most abundant; mean counts per sample: Angus 867,342; Charolais 1,060,828; KC 1,272,798), bta-miR-143, bta-miR-148a, bta-miR-26a, bta-miR-30a-5p, bta-miR-22-3p, bta-miR-27b, bta-let-7f, bta-miR-27a-3p, bta-miR-101. Novel miRNAs: At miRDeep2 score ≥4 with significant Randfold P-values, identified 126 (Angus; from 129 precursors), 101 (Charolais; from 103 precursors), and 119 (KC; from 125 precursors) novel miRNAs; only 31 novel miRNAs common to all three breeds (241 unique total), indicating breed specificity. Highly expressed novel miRNAs included bta-miR-AB-10 and bta-miR-AB-9 (mean aligned reads ~52,817 and ~46,932 across breeds). Differential expression by RFI (FC >1.5, P <0.05): Angus: 12 DE miRNAs (10 known, 2 novel); 7 up- and 5 downregulated in low-RFI. Charolais: 18 DE miRNAs (16 known, 2 novel); 12 up- and 6 downregulated in low-RFI. KC: 13 DE miRNAs (10 known, 3 novel); 8 up- and 5 downregulated in low-RFI. Overlap: Most DE miRNAs were breed-specific; only two common across all breeds: bta-miR-449a (upregulated in low-RFI in all breeds) and bta-miR-AB-2 (upregulated in low-RFI Charolais; downregulated in low-RFI Angus and KC). Stringent significance (Bonferroni FDR <0.05, FC >1.5): Angus: bta-miR-11985 (down in low-RFI), bta-miR-AB-2 (down). Charolais: bta-miR-2415-3p (up), bta-miR-2419-5p (up), bta-miR-AB-2 (up). KC: bta-miR-190a (down), bta-miR-449a (up), bta-miR-155 (up), bta-miR-424-5p (up), bta-miR-223 (up), bta-miR-AB-63 (down). Target prediction (most abundant miRNAs): For 18 highly expressed miRNAs (16 known + novel bta-miR-AB-10, bta-miR-AB-9), 1,022 targets identified at ≥99th percentile context++ score; 1,008 mapped in IPA and enriched for functions including cell morphology, cellular assembly and organization, growth/proliferation, free radical scavenging, and cell death/survival; enriched pathways included sirtuin signaling, IL-12 signaling/production in macrophages, tryptophan degradation III, receptor recognition of bacteria/viruses, and senescence. Target prediction (DE miRNAs): Targets at ≥99th percentile context++: Angus 767, Charolais 1,667, KC 787. Enriched functions/pathways: Angus targets related to cell cycle, gene expression modulation, cellular assembly, DNA repair, RNA post-transcriptional modification; pathways included HIPPO, GADD45, ATM, ER stress, granulocyte adhesion/diapedesis. Charolais targets enriched for protein synthesis/trafficking, RNA post-translational modification, lipid metabolism, molecular transport; pathways included estrogen receptor signaling, granzyme A, sirtuin signaling, protein ubiquitination, coagulation. KC targets enriched for cell signaling, RNA post-translation modification, protein synthesis, cell-to-cell signaling, cellular growth/proliferation; pathways included RhoGDI, RhoA, Rho family GTPases signaling, oxidative phosphorylation, diphthamide biosynthesis. Integration with previously reported DE genes: Using prior liver mRNA DE datasets from the same animals, at ≥99th percentile context++ only 1, 2, and 5 DE genes were targeted in Angus, Charolais, and KC, respectively. At >50th percentile, 44/72 (61.1%) Angus, 31/41 (75.6%) Charolais, and 129/175 (73.7%) KC DE genes were predicted targets of DE miRNAs. Many DE target genes were regulated by multiple DE miRNAs, and many DE miRNAs targeted multiple DE genes. For the common DE miRNA bta-miR-449a, predicted DE targets included 16 (Angus; 12 down, 4 up in low-RFI), 11 (Charolais; 4 down, 7 up), and 35 (KC; 23 down, 12 up), with SERPINA3, TP53INP1, and LPIN1 common across breeds. qPCR validation: Six miRNAs showed expression direction concordant with RNAseq; Pearson correlations between methods were high (0.73–0.99; P from 6.79E-03 to 6.00E-10), though qPCR P-values were generally higher than RNAseq.

The study profiled the hepatic miRNAome in three beef cattle populations and identified miRNAs differentially expressed between steers divergent for residual feed intake, addressing the hypothesis that miRNA regulation contributes to feed efficiency. The high overlap (~90%) of expressed known miRNAs across breeds indicates a conserved hepatic miRNA landscape, while the breed-specific nature of most DE miRNAs mirrors prior observations for mRNA DE genes, suggesting population-specific regulatory mechanisms. Highly abundant miRNAs (e.g., miR-192, miR-143, miR-148a, miR-26a) likely maintain core hepatic functions such as metabolic homeostasis, cell proliferation, and detoxification, consistent with literature in other species. The consistent upregulation of bta-miR-449a in low-RFI animals across all breeds highlights a potential conserved regulator of efficiency-related pathways; its predicted targets include genes involved in stress response and metabolism (e.g., TP53INP1, LPIN1). Functional enrichment of targets for both abundant and DE miRNAs revealed processes previously implicated in feed efficiency (lipid metabolism, protein synthesis/trafficking, cell signaling, growth/proliferation, cell death/survival), supporting a role for miRNAs in modulating these pathways. Integration with prior liver mRNA DE data showed that a substantial proportion (~61–76%) of DE genes are predicted targets of DE miRNAs (at context++ >50th percentile), and observed negative miRNA–mRNA correlations support canonical repression; cases of positive correlation or unchanged mRNA despite miRNA upregulation could reflect translational repression without mRNA decay, complex feedback, or prediction limitations. qPCR validation supported the RNAseq-based expression patterns. Overall, the findings support that miRNAs are part of the regulatory architecture underlying hepatic contributions to feed efficiency, with both shared (e.g., miR-449a) and breed-specific components.

This work provides a comprehensive profile of bovine hepatic miRNAs across Angus, Charolais, and Kinsella Composite steers and identifies 39 miRNAs (including five novel) associated with residual feed intake. Most DE miRNAs were breed-specific, while bta-miR-449a and bta-miR-AB-2 were common across breeds (with bta-miR-449a consistently upregulated in efficient animals). Predicted targets of abundant and DE miRNAs are enriched in biological processes central to hepatic function and previously linked to feed efficiency, and DE miRNAs potentially regulate a large fraction of previously identified DE mRNAs. The study advances understanding of miRNA-mediated regulation in liver related to feed efficiency and highlights candidate miRNAs and pathways for further validation. Future research should include larger cohorts for increased statistical power, cross-population validations, experimental validation of miRNA–mRNA interactions (e.g., reporter assays, perturbation experiments), and integration with proteomics to capture translational effects.

• Differential expression thresholds primarily used raw P < 0.05 with FC > 1.5 for discovery; only a subset met stringent Bonferroni FDR < 0.05, indicating potential false positives and the need for larger sample sizes to improve power.
• Most target gene interactions were predicted in silico (TargetScan) and not experimentally validated; prediction accuracy and context-dependent regulation (mRNA decay vs translational repression) limit interpretability.
• Breed specificity of DE miRNAs may reflect genetic background, environmental, or physiological differences; generalizability across populations requires replication.
• Correlation analyses between miRNA and mRNA capture only transcript-level effects and may miss translational regulation; positive correlations observed suggest complex regulation beyond simple repression.
• The study focused on liver; other tissues relevant to feed efficiency were not assessed for miRNA contributions.