MicroRNA (miRNA) in Plasma Small Extracellular Vesicles (sEV) as Potential Early Indicators of Dairy Cow Subfertility.

Conceptualisation: Pevindu Abeysinghe and Murray D. Mitchell. Investigation: Pevindu Abeysinghe. Writing—original draft preparation: Pevindu Abeysinghe. Writing—review and editing: Leila Cuttle, Natalie Turner and Murray D. Mitchell. Visualisation: Pevindu Abeysinghe. Supervision: Leila Cuttle and Murray D. Mitchell. All authors have read and agreed to the published version of the manuscript.

The presence of Western blot (WB) sEV characteristic markers Flot‐1, CD81, CD9, TSG101 and absence of non‐sEV marker BSA and GAPDH in the sEV samples confirmed they contained sEV (Figure S1 ). Furthermore, the TEM images confirmed the isolated samples contain spherical cup‐shaped particles (Figure S2 ), and NTA results confirmed the size of particles in the range from 30–200 nm (Figure S3 ). Further, purity estimate of the samples between the two groups demonstrates a similar distribution (Figure S4 ). Overall, the particles isolated with combined UC and size‐exclusion chromatography (SEC) adhere to the definition of exosomes/sEV. Supporting Information, File 1 contains sEV characterisation data. The relative fold‐changes (FC) of trimmed mean of the M‐values (TMM) normalised sEV miRNA read counts between sEV isolated from high fertile (HF) and low fertile (LF) primiparous dairy cows plasma samples revealed a total of 209 DE miRNAs above the fold change (FC) threshold value (−logFC > 2) (Figure  1 ). Interestingly, 97 of those DE miRNAs Volcano plot of miRNA expression. A total of 209 miRNAs were identified as differentially expressed (‐log fold‐change > 2). As the red dots denote, 112 miRNAs were highly expressed in the low fertile group (log fold change 2). Were relatively highly expressed in HF cow plasma‐derived sEV samples, and in contrast, 112 miRNAs were relatively highly expressed in LF‐derived sEV samples. The HF group later identified as pregnant, and the LF group were non‐pregnant after the first round of AI; thus, these DE miRNAs reflect miRNA profiles in blood plasma of animals that also had successful and failed conception. However, the majority of miRNAs identified did not show a DE, rather displaying a common distribution pattern regardless of fertility status. The complete DE miRNA data are available in Supporting Information, File 2 , and highly expressed miRNA lists are in Supporting Information, File 3 . Next, we explored high confidence differentially expressed miRNAs which passed a false discovery rate (FDR) threshold value of a 0.05. FDR correction was applied to control for multiple hypothesis testing, limiting false positives to no more than 5% and ensuring a balance between true discoveries and statistical accuracy. There were 6 sEV miRNAs (miR‐17‐5p, miR‐2335‐5p, miR‐29e‐3p, miR‐2284c‐5p, miR‐2285af‐1‐5p and miR‐2285dd‐3p) which were upregulated in most of the HF sEV samples, and 8 sEV miRNAs (miR‐2285j‐1‐3p, let‐7e‐5p, miR‐3613‐5p, miR‐12054‐5p, miR‐2285aw‐3p, miR‐199a‐2‐3p, miR‐2284e‐5p and miR‐181b‐2‐3p) upregulated in most of the LF sEV samples (Figure  2 ). Further, the average miRNA counts of the HF and LF sEV groups showed a distinct DE between the two groups (Figure  2 ). The hierarchical clustering demonstrates HF sEV and LF sEV cluster together, separate from the plasma miRNA profiles, and there was separate clustering of the 6 DE miRNAs upregulated in HF sEV from the 8 DE miRNAs upregulated in LF sEV. Differential expression of top 14 miRNAs (FDR< 0.05) (data generated using miRNA next‐generation sequencing). (a) all replicates of plasma‐derived sEV isolated from the high fertile (HF_SEV) and low fertile (LF_SEV) primiparous dairy cows. (b) average miRNA expression between crude plasma (Plasma) versus high fertile sEV (HF_SEV, n  = 10) versus low fertile sEV (LF_SEV, n  = 10). Plasma samples used for sEV isolation were selected from cows based on extreme fertility breeding values (FBV) and artificial insemination (AI) outcomes. HF_SEV represents cows with high FBV that conceived on first AI, and LF_SEV represents cows with low FBV that did not conceive on first AI ( n  = 10/group). Plasma represents the control: a plasma miRNA sample, isolated from a pooled plasma composed of samples from one randomly selected animal from each of the HF and LF cohorts. This comparison illustrates a clear distinction between the differentially expressed sEV miRNAs and circulating plasma miRNAs. (The colour keys of the heatmaps represent sEV miRNA abundance levels, with blue indicating the lowest abundance, red the highest, and white/light yellow representing moderate levels). Interestingly, the miRNA sequencing determined the absence of most of the DE miRNAs in the blood plasma control. Only miR‐2285dd‐3p was present at higher levels in the plasma compared to the HF or LF sEV groups. Collectively, the identification of the expression of 13 DE miRNAs exclusively in sEV indicates a unique packaging of miRNA into sEV. Relative quantification by qRT‐PCR on plasma sEV derived from young heifers revealed a significant upregulation of miR‐181b‐2‐3p in LF sEV group compared to HF sEV group, with a statistically significant difference ( p value = 0.0093) (Figure  3a ). The relative expression ratio of LF sEV to HF sEV was 2.665 (2 −ΔΔCT ; LF = 19.6095, HF = 7.35636). This result validates the results from the miRNA next‐generation sequencing (Figure  1 and Supporting Information, File 2 ), which also demonstrated that miR‐181b‐2‐3p was significantly upregulated in LF sEV derived from the plasma of primiparous dairy cows (FDR = 0.0129, log FC = −2.48704). Further, effect size estimation of miR‐181b‐2‐3p indicates a large effect between HF and LF sEV groups (Cohen's d  = 1.48) (Figure  3b ). This supports a biologically meaningful difference in expression levels between the two groups. ROUT test identified one outlier result in Plasma HF sEV group, thus there were n  = 7 in HF sEV group, and n  = 8 in LF sEV group, in the final data analysis (Figure  3 ). Additionally, CT calculations and statistical analysis results are available in the Supporting Information, File 4 . miRNA qRT‐PCR validation for miR‐181b‐2‐3p. (a) It confirmed the significant upregulation ( p  = 0.0093) of miR‐181b‐2‐3p in low‐fertile (LF) compared to high‐fertile (HF) plasma small extracellular vesicles (sEV) derived from young heifers consistent with the next‐generation sequencing results. Values are presented as mean ± SEM. Mann–Whitney U test was used to identify statistically significant differences between the HF sEV and LF sEV groups. p values of * p  ≤ 0.01 were considered statistically significant. Two data points in Bovine Endometrial Cells (BEC, n  = 2) represents miR‐181b‐2‐3p expression level in each bovine endometrial epithelial (bEEL) and stromal (bCSC) cells. (b) Effect size estimation of miR‐181b‐2‐3p expression between HF and LF sEV groups. Violin plot shows the bootstrapped distribution of the mean difference (LF−HF). The black dot and whiskers represent the observed mean difference (12.25) and its 95% CI (3.63 to 20.88). The effect size (Cohen's d  = 1.48) indicates a large effect between HF and LF sEV groups. Bovine mature miR‐181b‐2 of hairpin miR‐181b‐2‐3p consists of 24 nucleotides (nt) (Figure  4a ). It targets a total of 58 genes, identified as shared in at least two of the miRNA prediction tools: TargetScan, miRwalk and miRanda (Figure  4b and Table  4 ). GSK3B‐interacting protein (GSKIP) and Synaptogyrin‐2 (SYNGR2) were predicted as a gene targets of miR‐181b‐2‐3p by all three target prediction tools, confirming considerable miRNA‐mRNA binding capacity (Target Scan repression ratio >20%, miRanda Score > 170, miRwalk binding P value of 1 and accessibility < 0.05). Functional analysis of bovine miR‐181‐2‐3b (a) structure of hairpin bta‐miR‐181b‐2 precursor miRNA ( MIMAT0003793, Accession MI0005013 ). (b) A total of 58 gene targets were identified as shared between at least 2 of the miRNA target prediction tools. (c) Gene Ontology (GO) analysis of the 58 identified target genes of bta‐miR‐181b‐2‐3p predominantly associated with reproduction‐related pathways. Reproduction‐related pathways and genes regulated by bta‐miR‐181b‐2‐3p. Interestingly, most of the cellular pathways regulated by the target genes of bta‐miR‐181b‐2‐3p are associated with reproduction, such as Gonadotropin‐releasing hormone (GnRH) receptor pathway, integrin signalling, EGF and FGF signalling. Further, bta‐miR‐181b‐2‐3p regulates immune‐related pathways such as inflammation mediated by chemokine and cytokine, and beta‐cell activation (Figure  4c ). Thus, alterations to pathways associated with both reproduction and immunity, achieved through the inhibition of specific genes by the highly expressed bta‐miR‐181b‐2‐3pw in low fertile dairy cows, may have played a role in the subfertility observed within the LF dairy cow group. The qRT‐PCR results for the other DE miRNAs (miR‐12054‐5p, miR2285aw‐3p, miR‐17‐5p, miR‐2285dd‐3p, miR‐2335‐5p, miR‐3613a‐5p, let‐7e‐5p, 2285j‐1‐3p, miR‐3613b‐5p, miR‐2285af‐1‐5p, miR‐199a‐2‐3p, miR‐2284e‐5p, miR‐2284c‐5p and miR‐29e‐3p) did not comply with the next‐generation sequencing results, as the miRNA expressions between the HF sEV and LF sEV groups were not statistically significant ( p  > 0.01) (Figure  5 ) (CT calculations and statistical analysis results are available in the Supporting Information, File 4 and data repository [Abeysinghe 2024 ]). Validation qRT‐PCR data for the rest of the DE miRNA demonstrates absence of statistically significant ( p  > 0.01) difference in the miRNA expression levels between HF sEV and LF sEV groups [(a) miR‐12054‐5p (b) miR2285aw‐3p (c) miR‐17‐5p (d) miR‐2285dd‐3p (e) miR‐2335‐5p (f) miR‐3613a‐5p, (g) let‐7e‐5p, (h) 2285j‐1‐3p, (i) miR‐3613b‐5p, (j) miR‐2285af‐1‐3p, (k) miR‐199a‐2‐3p, (l) miR‐2284e‐5p, (m) miR‐2284c‐5p and (n) miR‐29e‐3p] which did not comply with the miRNA next‐generation sequencing results. Values are presented as mean ± SEM. Mann–Whitney U test was used to identify statistically significant differences between the HF sEV and LF sEV groups. p values of * p  ≤ 0.01 were considered statistically significant. Two data points in bovine endometrial cells (BEC, n =2) represent miRNA expression level in each bovine endometrial epithelial (bEEL) and stromal (bCSC) cells.

The Ruakura Animal Ethics Committee (RAEC) in New Zealand and the University Animal Ethics Committee (UAEC) at the Queensland University of Technology (QUT), Australia, approved all the research conducted in this study. It includes approval for the cows used in the discovery study (RAEC approval #14200 and UAEC QV ref #83807) and for the young heifers used in the validation study (RAEC approval #14654 and UAEC QV ref #83810). The study was conducted in compliance with Australian code for the care and use of animals for scientific purposes (Australia aNHaMRCA, author 2013 ), ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines, and all methods were performed in accordance with the relevant guidelines and regulations (Ahluwalia et al. 2020 ). Animals used in this study were Holstein–Friesian dairy cows ( Bos taurus ), and they were part of a larger experiment being conducted at DairyNZ, New Zealand, to establish Fertility Breeding Value (FBV). The complete description of the establishment of the research herd is available elsewhere by Meier, McNaughton et al. 2021 , Meier, Kuhn‐Sherlock et al. 2021 and Roche et al. 2007 . Briefly, a divergently fertile dairy cow herd representing a 10‐point difference in their FBV, ranging from extremely high (+5%; HF) to extremely low (−5%; LF), were determined utilising a breeding strategy using the MateSel algorithm (Kinghorn 2011 ). It restricted the averages for milk volume, fat yield, protein yield, birth weight, and ancestry (North American Holstein Friesian) to within 1 standard deviation between the two groups. Inbreeding of the offspring was limited (mean ± SD: 2.8 ± 1.44%) to balance out other critical traits. Additionally, New Zealand's FBV is the proportion of a sire's daughters that calve in the first 42 days of the seasonal calving period (Meier, Kuhn‐Sherlock et al. 2021 ). For the initial discovery study (next‐generation miRNA sequencing), a population of dairy cows from the above‐mentioned research herd was randomly selected from Tokanui Farm (Waikato), New Zealand (38°08′ S, 175°33′ E, 40 m above sea level). Cow blood samples were collected around an artificial insemination (AI) mating during seasonal breeding, in the same season of the year (October 2017). All cows were at their first calving/primiparous in their first lactation and 830 ± 10.6 days (∼27.5 months) of age at the time of sampling. This larger cohort ( n  = 80) contained dairy cows with four different fertility groups ranging from extremely high to extremely low fertility (Table  1 ). Groups were segregated by considering the success of conception in the first round of AI and FBV value of the parent progeny. All the animals in the sub‐cohort were of similar physical attributes and genetic merit (e.g., body weight, milk production, and percentage of North American genetics) except for the FBV. From these four groups, cow plasma samples were selected based on extreme phenotypes and AI outcome, specifically Group 1 (High FBV, conceived first AI) and Group 4 (Low FBV, did not conceive first AI) ( n  = 10/group) for next‐generation sEV miRNA sequencing in the discovery study (Meier, McNaughton et al. 2021 , Meier, Kuhn‐Sherlock et al. 2021 ). Classification of primiparous dairy cow groups used for the discovery study. The inclusion criteria for this study cohort were if calving occurred between week 28 and 31 (inclusive) (week = year week) n  = 96 (inclusive), which was between 9th July and 5th August 2017 (inclusive) and the exclusion criteria were if a collection did not have both milk and EDTA plasma samples, if there was no sample date recorded, if the cows received a reproductive treatment, or if the cows had censored postpartum anovulatory intervals. Classification of primiparous dairy cows into 4 different groups based on fertility breeding values (FBV) and conception after the first artificial insemination (AI). Blood samples taken for sEV isolations were from Group 1 (HF‐sEV) and Group 4 (LF‐sEV). Cows were allowed to graze in a grazing property on the farm during the blood collection, where a consistent grazing system was maintained across the herd. After blood sampling, cows had access to a fresh allocation of pasture twice daily, and cows were only returned to the same area when a minimum of 2 leaves appeared on the majority (>66%) of the grazing plants. It includes perennial ryegrass ( Lolium perenne L.) tillers, including kikuyu and chicory. Supplementary feed (palm kernel extract and pasture bailage/silage) was used by the grazier as required to achieve targeted weight gains (190 to 200 kg). Further information on grazing and breeding management of this animal sample is available in two publications by Meier, Kuhn‐Sherlock et al. 2021 , Meier et al. 2017 ). A cattle cohort of young heifers was used for the validation studies (miRNA qRT‐PCR analysis) were also from the same progeny of a divergent fertile FBV dairy cow model as described in Animal management. However, these heifers were reared in a different farm facility—DairNZ Lye Farm, Hamilton, New Zealand (37.78°S, 175.28°E). The young heifers were born between July and September 2018, where they were peripubertal, 311 ± 0.8 days (∼10 months) of age during the plasma collection (which was conducted between April and September 2019). These animals were part of a large experiment (Flay et al. 2022 ) that investigated reproductive endocrine axis in heifers with divergent FBV values. The control animal group ( n  = 80) of that experiment comprised blood plasma samples collected before and after saline administration (instead of the treatment, which involved a challenge of Kisspeptin and Gonadotropin‐releasing hormone). However, for this validation qRT‐PCR study, blood plasma samples collected 30 min before the saline administration from two distinct FBV groups, HF ( n  = 8) and LF ( n  = 8), were selected. The HF heifers were born to HF cows sired by bulls with HF FBV. Conversely, LF heifers were born to LF cows sired by bulls with LF FBV. Complete information on heifer rearing and the feed is included elsewhere, by Meier, McNaughton et al. 2021 . Briefly, heifers grazed on ryegrass ( L. perenne L.) pasture, with the sward including kikuyu ( Pennisetum clandestinum ) and chicory ( Cichorium intybus ). Blood collection from cows used in the discovery study and the heifers used in the validation study followed a similar protocol. Blood was collected by coccygeal venepuncture into evacuated blood tubes containing lithium heparin (BD Vacutainers, BD New Zealand, Auckland, New Zealand) anticoagulant and was placed on ice. The blood collection team was blinded to all other information about the cows at that time. At the end of the day, blood samples were transported to the laboratory, briefly centrifuged for 1900 ×  g for 12 min at 4°C, aliquoted, frozen, and stored at −80°C. They were then transferred to Queensland University of Technology, Centre for Children's Health Research Facility in Brisbane, Australia, where they remained at ‐80°C until sEV isolation procedures were conducted (Meier, McNaughton et al. 2021 , Meier et al. 2017 , Flay et al. 2022 , Reed et al. 2017 , DairyNZ 2020 ). EV were isolated from 20 cow plasma samples (Discovery Study) and 16 young heifer plasma samples (Validation Study) using an optimised sequential centrifugation protocol as previously described (Turner et al. 2022 ). Briefly, 40 mL starting volume of plasma per sample was pre‐treated at 3000 ×  g for 10 min at 4°C to remove any cellular debris. Then the supernatant was centrifuged at 12,000 ×  g for 30 min 4°C to eliminate apoptotic cellular bodies. The resultant supernatant was next filtered through a 0.22 µm filter (Corning Inc., Corning, NY, USA) and ultracentrifuged at 100,000 ×  g for 2 h at 4°C in fixed angle ultracentrifugation (UC) rotor (Type 50.2, Beckman Coulter, Brea, CA, USA). Finally, the pellets containing the extracellular vesicles were resuspended in 500 µL of filtered Dulbecco's Phosphate Buffered Saline (DPBS, pH 7.0–7.2; Gibco, Life Technologies Australia Pty Ltd) and stored at −80°C for further analysis. Fractionation of EV isolated from above steps was conducted by qEV original size exclusion columns (qEV original/70 nm, Izon Science, Christchurch, New Zealand) as per manufacturer's instructions. A 500 µL EV sample was introduced to the column, and 16 individuals' 500 µL fractions were eluted from the column using gravity into separate 1.5 mL microcentrifuge tubes. Next, fractions 7–10 were pooled according to the previous reports published by our group (Turner et al. 2022 , Koh et al. 2018 ) to obtain the best representation of sEV (particles < 200 nm). A micro BCA Protein Assay Kit (cat number 23235, Thermo Fisher Scientific, Brisbane, Australia) was used to quantify total protein in pooled sEV fractions 7–10. The microplate assay procedure was conducted with a linear range of 2–200 µg/mL and followed the manufacturer's protocol as described previously (Turner et al. 2022 ). Three technical replicates of bovine serum albumin (BSA) protein from each standard provided by the micro‐BCA assay kit and two technical replicates of samples from each sEV isolation method were added to a 96‐well flat‐bottom plate (Greiner CELLSTAR, Sigma‐Aldrich (Merck), Melbourne, Australia). Volume/weight 1% of sodium deoxycholate (SDC; Sigma‐Aldrich (Merck), Melbourne, Australia)/deionised H 2 O was used to solubilise samples and standards (which were made as a 10 mL master solution for the plate). For characterisation of sEV and non‐sEV protein markers, three representative pooled fraction 7‐10 sEV samples were selected from each HF and LF group. Volumes required for WB were calculated according to the micro BCA total protein assay results to include a protein content of 5 µg in each sample. Samples were vacuum evaporated in a vacuum concentrator (Eppendorf Concentrator plus, Sydney, Australia) and resuspended with 19.5 µL milliQ water. Initial concentration 10× NuPAGE sample reducing agent (NP0004, Thermo Fisher Scientific, Brisbane, Australia) and 4× NuPAGE LDS sample buffer (NP0007, Thermo Fisher Scientific, Brisbane, Australia) were added to the samples which were placed on ice to make up to a final concentration of 1x and reduced for 10 min at 70°C according to the manufacturer's protocol. To resolve the proteins in the samples, NuPAGE 4 to 12%, Bis‐Tris, 1.0 mm, Mini Protein Gels, 10‐well (NP0321BOX, Thermo Fisher Scientific, Brisbane, Australia) with Chameleon Duo Pre‐stained Protein Ladder (928‐60000, Li‐COR, Mulgrave, Australia) were used at 110 V for 30 minutes. For the positive control, a crude plasma sample was used, and milliQ water was used as the negative control. The Trans‐Blot Turbo Transfer System (Bio‐Rad Laboratories Pty Ltd., Sydney, Australia) was used for the transfer of proteins in the gel to a polyvinylidene fluoride membrane (Bio‐Rad Laboratories Pty Ltd., Sydney, Australia). Next, the PVDF membrane was blocked using a 1:1 solution of Odyssey Intercept blocking buffer (927‐70001, Li‐COR, Mulgrave, Australia) and phosphate‐buffered saline (PBS) (Sigma‐Aldrich (Merck), Melbourne, Australia) for 1 h in a cold room (4°C) with agitation. Primary antibodies, recombinant anti‐flotillin‐1 (1:500 dilution, Rabbit monoclonal ab133497, Abcam, Melbourne, Australia), recombinant anti‐CD9 (1:250 dilution, Mouse monoclonal NB500‐494, Novus Biological, USA), recombinant anti‐CD‐81 (1:500 dilution, NBP1‐77039, Rabbit polyclonal Novus Biological, USA), and recombinant‐anti‐BSA (1:5000 dilution, Rabbit polyclonal ab192603, Abcam, Melbourne, Australia) were used for primary blotting. The primary antibodies were diluted in a 1:1 diluted solution of Odyssey Blocking buffer and PBS and incubated overnight in a cold room (4°C) with agitation. The PVDF membrane was then rinsed four times in PBST (PBS + 0.1% Tween‐20) for 5 min in each wash, and the membrane was incubated in the respective secondary antibody for 1 h in a cold room (4°C). Goat anti‐Rabbit IgG (1:10,000 dilution, Li‐COR, Mulgrave, Australia) was used for Flot‐1, CD‐81 and BSA; Goat anti‐Mouse IgG (1:10,000 dilution, Li‐COR, Mulgrave, Australia) was used for CD‐9. PVDF membranes were then again washed with PBST four times for 5 min in each wash. Finally, the membranes were imaged in a Li‐COR Odyssey fluorescent scanner at 700 and 800 nm and were processed using Image Studio Lite v5.2 (Li‐COR Biosciences, Lincoln, NE, USA). A pooled sample comprising sEV fraction 7–10 was passed through Nanosight NS500 instrument (NanoSight NTA 3.1 Build 3.1.46, Malvern Panalytical, Sydney, Australia) to measure particle size and concentration. Samples were diluted 1:10 in filtered DPBS before being analysed in the nanoparticle tracking analysis (NTA). Three technical replicates were run for 30 s interval each for all biological replicates and captured using the Scientific CMOS camera setting according to the manufacturer's guidelines (Camera level–14, Detection threshold–5). Synthetic (latex) beads of size 100 nm were used to perform instrument calibration at a 1:250 dilution in deionised water. Transmission electron microscopy (TEM) was used for visualisation of sEV at QUT Central Analytical Research Facility (CARF). A 5 µL aliquot of sEV sample was added onto glow‐discharged copper grids (200 mesh) for 3 min and was negatively stained with 1% uranyl acetate for 2 min, then briefly blotted with blotting paper to remove excess liquid. The voltage setting of the JEOL 1400 transmission electron microscope (JEOL, Sydney, Australia) was 100 kV, and images were captured using a 2K TVIPS CCD camera (TVIPS, Gauting, Germany). SEV samples, which represented around 25 µg of total protein, were used to enrich miRNA using an optimised sEV miRNA isolation methodology as previously described in Abeysinghe et al. ( 2021 ). Briefly, total RNA was isolated from sEV samples using TRIzol LS reagent (CAT#15596026; Invitrogen, Carlsbad, CA, USA) (2.5 volume of sEV sample: 7.5 volume of TRIzol; incubation for 5 min at room temperature (RT)). Next, chloroform (cat number 388306, Sigma‐Aldrich (Merck), Melbourne, Australia) was added for 2 min at RT (2.5 volume of sample: 1.5 volume of Chloroform) for the phase separation of RNA. The sample was then centrifuged at 12,000 ×  g for 15 min, and the transparent upper aqueous layer with RNA was then purified further using miRNeasy mini spin columns (miRNeasy mini kit, 217004, QIAGEN) following the manufacturer's protocol for seV miRNA isolation. RNase‐free water was passed through the column twice consecutively (each 40 µL) to collect 80 µL of sEV miRNA sample. The sEV isolation method was conducted within a day, and the isolated sEV miRNA samples were sent on the same day to the Australian Genome Research Facility (AGRF; Melbourne, Australia) for sequencing. Samples enriched with sEV miRNA were first analysed for primary quality control (QC) by quantifying total miRNA concentrations using an Agilent 2100 Bioanalyzer Small RNA Series II (Agilent, Santa Clara, California, USA). NEXTFLEX Small RNA‐Seq Kit v3 was then used for miRNA library preparation, and later the next‐generation sequencing was conducted using a Novaseq S1 platform, utilising single‐end 100 bp sequencing, covering up to 10 million sequencing depth. A plasma miRNA sample was used as a control, isolated from a pooled plasma composed of samples from one randomly selected animal from each of the HF and LF cohorts. The resulting fastq.gz files were downloaded from the AGRF data repository account and further processed for the bioinformatics analysis. The optimised bioinformatics pipeline including the codes are available in the GitHub repository https://github.com/PevinduA/sEV‐miRNA‐annotation/tree/main . Briefly, the fastq.gz files were extracted and unpacked using WinZip v27 (Mansfield, Connecticut, USA), and unitas v1.6.0 ( https://sourceforge.net/ projects/unitas/) was used to map B. taurus miRNA using miRbase release 22.1 ( https://www.mirbase.org/ ). The reads were also screened for the presence of any Illumina adapter/ overrepresented sequences (AGATCGGAAGAGCACACGTCTGAACTCCAGTCAC) and the species was selected as B. taurus . The resultant bovine miRNA count tables from different sEV miRNA methodologies and the control bovine plasma sample were used in edgeR (version 3.30.3) of R package 4.2.2 ( https://bioconductor.org/packages/release/bioc/html/edgeR . html) to perform DE of miRNAs between the HF and LF groups. False discovery rate (FDR) analysis was performed to correct for multiple hypothesis testing and set to 0.05 (FDR < 0.05). Plasma samples from young heifers (HF n  = 8 and LF n  = 8) were from the Fertility FBV experimental study from DairyNZ Lye Farm, New Zealand, and the young heifers were peripubertal, around 10 months old during the plasma collection, as previously described in the Animal, Management, and Blood collection section. Plasma sEV total RNA was isolated using previously validated exoEasy Maxi and Midi spin columns (exoRNeasy Serum/Plasma Starter Kit, 77023, QIAGEN) according to the manufacturer's protocol and verified protocol by Dunlop et al. 2021 . 4 mL of plasma were passed through the spin columns, and final eluted total RNA samples were checked for RNA concentration and purity using a NanoDrop One Microvolume UV‐Vis Spectrophotometer (Thermo Fisher Scientific, Wilmington, Delaware). For First‐Strand cDNA synthesis, 4 µL of total RNA were used according to Dunlop et al. 2021 in a miRCURY LNA RT Kit (339340, QIAGEN) according to the manufacturer's protocol. Next, real‐time PCR was conducted for three technical replicates per sample using Individual miRCURY LNA miRNA PCR Assays (339306, QIAGEN) and miRCURY LNA SYBR Green PCR Kit (339346, QIAGEN) according to the manufacturer's protocol for the DE miRNAs between LF and HF identified using next‐generation sequencing (RT‐qPCR cycling conditions are in Table  2 ). Custom‐made miRCURY LNA miRNA PCR Assays were designed with the assistance of Qiagen technical team for each of 15 DE miRNAs, preserving the miRNA sequence and the bovine species specificity. Individual miRNA primer sets were designed using miRCURY LNA miRNA PCR Assay, and the list of miRNA sequences is in Table  3 . RT‐qPCR cycling conditions. miRNA sequence information in miRNA LNA PCR Assays used for DE miRNA identified in next‐generation sequencing. U6 snRNA (v2) miRCURY LNA miRNA PCR Assay was used as the house‐keeping endogenous miRNA control to normalise the RT‐qPCR results when analysing miRNA expression using the comparative 2 −ΔΔCT method (Schmittgen and Livak 2008 , Erdem et al. 2023 ). Further, undetermined CT values (i.e. samples which did not show respective miRNA expression signal) were considered as zero (0). RNA lysates from bovine endometrial epithelial (bEEL) and stromal (bCSC) cells were used as the controls during this study. EV RNA was isolated separately from each bEEL and bCSC cell line sample, and miRCURY LNA miRNA PCR Assay was performed for each miRNA in 3 qRT‐PCR replicates ( n  = 3 technical replicates). The relative abundance of each miRNA in bEEL and bCSC cells was assessed using 2 −ΔΔCT method. The results from both cell lines were then combined during the statistical analysis as a single control reference, represented as bovine endometrial cells (BEC) in the manuscript. Therefore, BEC in the manuscript represents two samples (BEC; n  = 2). Bovine endometrial epithelial (bEEL) and stromal (bCSC) cell lines were a kind gift from Professor Michel A. Fortier (Université Laval, Québec). The cells were grown in RPMI media (Gibco, Thermo Fisher Scientific Australia Pty Ltd, Scoresby, Vic) containing sEV/exosome‐depleted 10% foetal bovine serum (Bovorgen, Interpath Services Pty Ltd, Australia) and incubated at 37°C and 5% CO 2 . These cells were not derived from the animals which were used in any of the discovery or validation studies. Data analysis for qRT‐PCR gene expression results was performed first by identifying outliers using the ROUT method with a Q value of 5%. For the cleaned data, a Mann–Whitney U test was performed to identify statistically significant differences between the HF_sEV and LF_sEV groups. Data are presented as sample means ± SEM. p values of * p  ≤ 0.01. were considered statistically significant. Further, undetermined CT values (i.e., samples which did not show respective miRNA expression signal) were considered as zero (0). Three different miRNA target prediction tools (1) miRanda from miRNet version 2.0 ( https://www.mirnet.ca/ ) (Fan et al. 2016 ), (2) miRmap ( http://mirmap.ezlab.org .) (Vejnar et al. 2013 ), and (3) miRWalk ( http://mirwalk.umm.uni‐heidelberg.de/search_mirnas/ ) were used to identify multiple target binding sites of the validated 6 DE miRNAs identified from the next‐generation sequencing results. In miRNet target prediction, the B. taurus was used as organism, miRBase ID as ID type, and target type was miRanda genes. For Degree filter, “All network nodes” were selected with a cutoff value of greater than 3, a greater than 100 cutoff was used for betweenness, and all networks were to connect with the shortest path (Herrnreiter et al. 2021 ). B. taurus was used as the organism for miRmap, and 10 DE miRNAs were checked for targets individually. Targets which were above 95% of power exact were chosen as confident miRNA targets. All the individual miRNA targets were pooled together to make a total list of miRmap targets for the 10 DE miRNAs (Vejnar et al. 2013 ). MiRWalk, putative target genes for the 10 DE miRNAs were predicted with a cutoff binding probability > 0.98 (Li et al. 2021 ). Target genes which intersected at least with two of the target prediction tools out of the three miRNA target prediction tools were used as the confident target sites of the RT‐qPCR validated 6 DE miRNAs. Finally, PANTHER v17.0 (Thomas et al. 2022 ) was used to perform Gene Ontology (GO) enrichment analysis for the target genes of DE sEV miRNAs.

DE patterns of sEV‐derived miRNA in reproductive disorders and healthy controls have previously been identified from various tissues, body fluids and cells (Ibañez‐Perez et al. 2022 , Zhou et al. 2020 ). Highly expressed miR‐2285aa and miR‐199a in LF plasma sEV were also detected by Wang et al. 2021 in plasma sEV from cows with endometritis. Exosome or sEV miRNA cargo in seminal plasma has been identified as regulators or biomarkers of fertility (Barceló et al. 2018 ); however, studies on association of circulating blood plasma‐derived EV miRNA cargo on subfertility are limited. Circulating plasma EV from endometriosis patients contains unique miRNA signatures in contrast to the healthy controls (Khalaj et al. 2019 ). Khalaj et al. 2019 identified the presence of miR‐139‐3p in plasma EV of healthy controls compared to the endometriosis patients, and presence of miR‐339 in plasma EV of endometriosis patients compared to healthy controls (Khalaj et al. 2019 ). Similarly, HF cow plasma sEV had a higher expression of miR‐139‐3p, and LF cow plasma sEV had a higher expression of miR‐339b‐3p and miR‐339b‐5p in our discovery study (data repository [Abeysinghe 2024 ]). Human and cow interspecies structural homologous miRNAs (Fromm et al. 2018 ) and functional relationships between bovine miRNAs and human gene expressions (Myrzabekova et al. 2021 ) have been identified. However, it requires extensive research and/or bioinformatic analysis to elucidate any functional similarity of bovine miR‐339b‐3p, miR‐339b‐5p and miR‐139‐3p to their identical human miRNAs. MiR‐17‐5p, which was highly expressed in HF cattle sEV, has been identified in oviductal fluid and cord blood, and is positively associated with embryo development (Aoki et al. 2022 ) and ovarian function (Ding et al. 2020 ), respectively. The anti‐inflammatory pathway of miR‐17‐5p is mediated through suppression of TLR4, and miR‐17‐5p has been shown to downregulate inflammatory mediators IL‐1β and TNF‐α (Ji et al. 2019 ). MiR‐2335‐5p, also highly expressed in sEV from HF cattle, acts as a tumour suppressor (Yu et al. 2021 ); however, its function in the reproductive context is unclear. Similarly, the miR‐29 family has been identified as a tumour suppressor for malignancies and the downregulation of miR‐29 correlates with aggressive forms of cancer (Kriegel et al. 2012 ). Although miR‐29e‐3p was highly expressed in HF cow sEV in our data, there is limited information on miR‐29e‐3p in the literature. Further, miR‐2284c‐5p, miR‐2285af‐1‐5p and miR‐2285dd‐3p were highly expressed in most of the HF cow sEV samples; however, there is limited or no research information on their functional aspects. The higher expression of miR‐17‐5p, miR‐2335‐5p, miR‐29e‐3p, miR‐2284c‐5p, miR‐2285af‐1‐5p and miR‐2285dd‐3p in HF cow sEV samples may suggest a possible activation of an anti‐inflammatory pathway that favours fertility status. Moreover, DE analysis identified 8 miRNAs which were highly expressed in most of the LF cow sEV samples. The let‐7 miRNA family and their role in inflammatory responses have been studied extensively (Lin et al. 2017 ). The expression of inflammatory molecules through activation of the NF‐κB pathway is among the pro‐inflammatory functional properties of let‐7e, and its inflammatory effects can be suppressed by binding to lncRNA lnc‐MKI67IP‐3 (Lin et al. 2017 ). MiR‐3613‐5p was upregulated in LF cow sEV, and it has been reported to possess oncogenic properties acting via activation of the NF‐κB inflammatory pathway (He et al. 2020 ). MiR‐199a‐2‐3p was highly expressed in LF cow sEV and is one of the predicted sequences in the primary precursor miR‐199a stem loop, which possesses anti‐inflammatory regulatory properties during labour and pregnancy (Williams et al. 2012 ). Human miR‐181b stimulates preeclampsia‐induced inflammation (Guo et al. 2022 ), demonstrating pro‐inflammatory function. The previous research on miR‐12054‐5p, 2285aw‐3p, miR‐2285j‐1‐3p and 2284e‐5p has been limited, and it is difficult to postulate the functional effects of higher expression in LF sEV. The miRNA candidates identified in this study are known to be associated with EVs and membranes. MiR‐17‐5p is relatively highly expressed in HF sEV and regulates endocytic trafficking through a complex regulatory pathway, which indicates miR‐17‐5p displays functional dynamics specifically related to sEV (Serva et al. 2012 ). Let‐7e‐5p has been identified in a circulating serum exosome biomarker panel for oesophageal adenocarcinoma (Chiam et al. 2015 ). There is evidence that exosomes can be used to deliver miR‐181b (Liu et al. 2021 ) and miR‐29 (Wang et al. 2019 ) as therapeutics. Further, miR‐3613‐5p, let‐7e‐5p, miR‐199a‐2 and miR‐181b‐2 have been reported in the Vesiclepedia database as miRNAs engulfed in extracellular vesicles (Pathan et al. 2019 ). Interestingly, our results report for the first time the presence of miR‐12054‐5p, 2285aw‐3p, miR‐2285j‐1‐3p, 2284e‐5p, miR‐2284c‐5p, miR‐2285af‐1‐5p and 2335‐5p in sEV samples compared to blood plasma. Additionally, we also report that miR‐2285dd‐3p is highly expressed in plasma compared to the HF and LF sEV groups. RNA content engulfed by EV differs profoundly from the RNA from non‐EV origin, indicating that RNA, including miRNAs, is selectively packaged into sEV during exosome biogenesis (O'Brien et al. 2020 ). The vesicular membrane serves as protection for fragile nucleotide sequences such as miRNA (Zhang et al. 2022 ); otherwise, they will be degraded by target complementary RNA sequences via 3′UTR direct binding (Kang et al. 2023 ). Therefore, the DE of the 14 miRNAs in sEV compared to the plasma may indicate a selective sEV miRNA packaging mechanism. Further, these 14 DE miRNAs may serve as putative sEV miRNA markers; however, this requires more research for validation. A higher expression of miR‐17‐5p has previously been identified in serum sEV from pregnant pigs compared to non‐pregnant controls (Zhou et al. 2020 ). MiR‐17‐5p is a broadly discussed biomarker in fertility, involved in luteal tissue vascularisation, implantation and during pregnancy (Reza et al. 2019 ). Higher abundance of miR‐181b‐2‐3p in plasma sEV in LF cows was confirmed by the validation study and the remaining 14 DE miRNAs that did not comply with the next‐generation sequencing data may indicate temporal aspects of their DE. Since the initial discovery phase was conducted in primiparous cattle, these 14 discordant sEV miRNAs may be age‐dependent and only quantifiably different in post‐pubertal cattle. While we acknowledge this limitation to our study, longitudinal data on pregnancy outcomes was critical to establishing extreme fertility phenotypes during the discovery phase. Our study therefore suggests an alternative route of inquiry that could focus on age‐dependent changes in cattle of divergent fertility status that may complement current methods. Alternatively, validation experiments from different animal cohorts and different dairy cow breeds, geographies, animal species and varied time‐points during the dairy cow life cycle may also be beneficial. Focusing only on the extreme fertility groups in this study may have overlooked key miRNA expression patterns present in intermediate groups, which could have provided valuable insights into early detection and stratification of dairy cows into different subfertility risk levels. These intermediate profiles might help in identifying subtle fertility issues and more tailored interventions across a broader range of reproductive states. However, it also highlights the limitations of the FBV scoring system in young cattle, which is simply the average of the parent FBVs and thus is not specific to individual cattle until they have undergone one mating cycle, and individual data become available. Therefore, it is crucial to recognise variability in FBV assessment across different countries. In New Zealand, FBV prediction metrics are based on a sire's daughters’ calving proportion in the first 42 days of the seasonal calving period, reinforcing the challenge of applying FBV to young cattle in this study (Meier, Kuhn‐Sherlock et al. 2021 ). In contrast, the United States employs fertility parameters such as daughter pregnancy rate, heifer and/or cow conception rate, and early first calving (Peñagaricano et al. 2023 ). The Netherlands utilises non‐return rate (NRR) to assess the fertility in Dutch Holstein dairy cows, defining it as a heifer that has not received a second AI after 52 days of the first insemination (van der Heide et al. 2020 ). In Brazil, fertility traits such as age at first calving and the first calving interval have been studied in Girolando dairy cattle in Brazil, identifying moderate and low heritability, respectively, underscoring the relevance of these phenotypic parameters in assessing FBV (Canaza‐Cayo et al. 2016 ). These examples emphasise the need for a better understanding of genetic factors contributing to divergent fertility and early predictors of fertility status in dairy herds. While we acknowledge that external factors such as age, environment, and circadian rhythms can affect miRNA expression, the use of distinct populations in the discovery and validation phases was fundamental to our hypothesis. By investigating whether miRNA patterns associated with subfertility in primiparous dairy cows are conserved across different age groups, particularly in younger heifers, we aim to establish the broad applicability of these miRNA signatures. If validated, these findings could facilitate early interventions, significantly enhancing reproductive efficiency in dairy farm management. Similarly, the rationale for utilising different EV isolation methods across the discovery and validation phases was to evaluate the robustness and transferability of the identified miRNA signatures and to consider the real‐world applicability of fertility management strategies in dairy operations. UC may be considered the gold standard for EV isolation in a laboratory setting; however, commercial precipitation‐based kits are more practical and accessible for potential on‐farm applications. The gene ontology enrichment analysis conducted for gene targets of bta‐miR‐181b‐2‐3p revealed critical reproduction‐related molecular signalling pathways regulated by the inhibition of its target genes. Gonadotropin‐releasing hormone (GnRH) is a crucial reproductive hormone that regulates luteinising hormone (LH) and follicle‐stimulating hormone (FSH), which are responsible for the maturation and release of oocytes within ovarian follicles, thus contributing to ovulation and cow oestrous cycle (Crowe 2016 ). GnRH releasing hormone receptor pathway represents the highest GO enrichment percentage (Table  4 ), thus indicates miR‐181b‐2‐3p contribution towards cow reproduction through GnRH regulating axis. Further, Wnt signalling pathway which also been regulated by the miR‐181b‐2‐3p, is vital for maintaining cow follicle development, ovulation, embryo development and optimal function of the endometrium (Gupta et al. 2014 , Xiao et al. 2021 , Tribulo et al. 2017 ). MiR‐181b‐2‐3p inhibits cadherin‐8 (CDH8) to disrupt Wnt signalling pathway (Table  4 ), in which CDH8 plays a central role in ovary development and cow fertility (Piprek et al. 2020 ). Moreover, miR‐181b‐2‐3p is implicated in the potential regulation of various other molecular signalling pathways related to bovine reproduction. Further, GO ontology analysis indicated miR‐181b‐2‐3p alters essential inflammatory pathways including chemokine and cytokine regulation, beta‐cell adhesion, TGF‐beta signalling and p38MAPK signalling, which is a critical indication of bovine immune suppression properties of mature miR‐181b‐2‐3p. Collectively, GO ontology analysis unveils disruptions of molecular signalling pathways essential for bovine overall reproduction by bta‐miR‐181b‐2‐3p. This indicates that upregulation of miR‐181b‐2‐3p in sEV from LF cattle may be reflective of, or actively contribute to their subfertility status. According to miRNA target prediction results, GSKIP protein was consistently identified as a potential target gene of bta‐miR‐181b‐2‐3p by all three miRNA target prediction tools. This designation was supported by its higher GSKIP mRNA binding complementarity, along with higher accessibility (0.025) (miRWalk), miRanda score of 182 and a significant repression ratio (38.44%) (TargetScan). GSKIP is a critical gene during parturition, and GSKIP deficiency causes lethality at birth in a GSKIP‐deficient mouse model (Deák et al. 2016 ). Significant upregulation of bta‐miR‐181b‐2‐3p in LF cattle may inhibit GSKIP and could be one mechanism through which subfertility is potentiated. MiR‐181b‐2‐3p acts as a precursor/hairpin miRNA to produce mature miR‐181b, which has already been identified as a critical gene regulator. Fibroblast growth factor (FGF) (Jiang and Price 2012 ) and epidermal growth factor (EGF) (Wang et al. 2023 ) signalling pathways dysregulated by mature miR‐181b are critical in bovine ovarian granulosa cell development. Negative interaction with the integrin signalling pathway by mature miR‐181b disrupts cell adhesion, vascularisation and intracellular signalling during bovine placentation (Bridger et al. 2008 ). Further, regulation of immune function by miR‐181b through various mechanisms is of importance in understanding its role in dairy cow reproduction and fertility. MiR‐181b inhibits the activation‐induced cytidine deaminase (AID) enzyme, impairing class switch recombination (CSR), which synthesises high‐affinity antibodies with specific antigen‐binding capabilities for effective antigen removal. Consequently, elevated miR‐181b expression in activated B cells impairs immunity (de Yébenes et al. 2008 ). The bovine miR‐181 family is elevated in degenerated bovine embryos in contrast to healthy blastocysts and was also detected in the culture media of pre‐implantation bovine embryos in the same study (Kropp et al. 2014 ). Interestingly, miR‐181a contains a complementary binding site for nucleoplasmin 2 (NPM2), an oocyte‐specific nuclear protein essential during early embryonic development, and the elevated levels of miR‐181a inhibit NPM2 protein expression (Lingenfelter et al. 2011 ). Treatment with an anti‐inflammatory decreases miR‐181c expression in bovine neutrophils stimulated with LPS, confirming pro‐inflammatory properties of the bovine miR‐181 family (Chuammitri et al. 2017 ). Further, miR‐181 family, including miR‐181b, has been differentially expressed in bovine granulosa cells of subordinate and dominant follicles during the development stage in the oestrous cycle (Salilew‐Wondim et al. 2014 ). Together, we can speculate that miR‐181b‐2‐3p, mature miR‐181b and other miRNAs in the miR‐181b family orchestrate the regulation of bovine reproduction, potentially contributing to the subfertility status observed in LF cows. We suggest that miR‐181b‐2‐3p could be a critical miRNA associated with dairy cow subfertility via the regulation of multiple molecular signalling pathways. This includes oocyte maturation and development, ovulation, embryo development, and regulatory control over essential immune and inflammatory processes. Further studies are warranted to validate miR‐181b‐2‐3p as a potential prognostic biomarker of dairy cow subfertility.

We identified differentially expressed plasma sEV‐derived miRNA (FDR < 0.05) between HF and LF dairy cows and found that most of these DE miRNA signatures were uniquely packaged in sEV. Our results suggest sEV‐derived miRNAs are potential early indicators of subfertility status in divergent cattle groups and propose that miR‐181b‐2‐3p is associated with dairy cow subfertility.

Fertility can be defined as the rate of pregnancies in mammals, including humans (Vander Borght and Wyns 2018 ) and bovine species (VanRaden et al. 2004 ), with a more precise definition being the ability to carry a healthy offspring to full term. A successful pregnancy in mammals relies on the perfect synchronicity of embryo and endometrial development (Teh et al. 2016 ). Maintaining an optimal balance of hormones, cytokines and growth factors is essential to enhance endometrial receptivity and implantation (Binelli et al. 2022 ). Nonetheless, inflammation resulting from illness or disease reduces reproductive potential by promoting the release of pro‐inflammatory cytokines that can negatively impact the intrauterine microenvironment (Chen et al. 2018 ). Specifically, excessive and persistent inflammatory diseases in the bovine uterus, such as mastitis and metritis, depreciate the receptivity of the endometrium, leading to failure of artificial insemination (AI) during cow breeding (Chastant and Saint‐Dizier 2019 ). However, expression of anti‐inflammatory cytokines in the bovine endometrium favours successful embryo implantation (Vlasova and Saif 2021 ), and in humans, anti‐inflammatory interleukin‐4 (IL‐4), interleukin‐10 (IL‐10) and C‐X3‐C motif chemokine ligand 1 (CX3CL1) enable the symbiotic growth of the foetus (Chatterjee et al. 2014 ). Moreover, systemic anti‐inflammatory treatments have been used in dairy cow animal models to improve immune function (Pascottini et al. 2020 ), and non‐steroidal anti‐inflammatory drugs (NSAIDs) are utilised for gynaecological practices under supervision for humans to alleviate pain during pregnancy (Deussen et al. 2020 ). Extracellular vesicles (EVs) are nanosized particles that are known to be important mediators of cell‐cell communication. Defined as subcellular particles released from cells that are delimited by a lipid bilayer and cannot replicate on their own (Welsh et al. 2024 ), ‘small’ extracellular vesicles (sEV) are the most widely studied among the various EV classes. sEV can be categorised according to their size or pelleting gravitational ( g) force, such as EV particles < 200 nm, or those particles obtained after sequential differential centrifugation ending with 100,000 ×  g (Welsh et al. 2024 ). Exosomes (∼50–150 nm) are a subgroup of sEV that originate from the endosomal compartment, are incorporated into multivesicular bodies as intraluminal vesicles, and are subsequently released as exosomes into the extracellular environment (Welsh et al. 2024 ). Ectosomes and microvesicles (∼100–1000 nm) bud directly from the plasma membrane and thus have distinct cargo and functions compared to exosomes. Due to the technical limitations of purely separating exosomes from ectosomes/microvesicles, we refer to all EVs < 200 nm in size as sEV in this study. Current multi‐omics analyses reveal the role of EVs in embryo implantation and the therapeutic potential of EVs for reproductive biomedicine (Poh et al. 2023 ). Circulating sEV miRNA are uniquely packaged, and the epigenetic information preserved by the vesicular membrane reflects the metabolic status of the cells of origin (Cao et al. 2022 ). Thus, sEV miRNAs are an excellent candidate for diagnostic and prognostic biomarkers for various diseases (Min et al. 2019 ). In particular, sEV miRNAs have been found to regulate cytokine release in cows with endometritis (Wang et al. 2020 ) and have been utilised as markers of successful embryo implantation in humans (Tan et al. 2020 ), demonstrating their utility as a prognostic biomarker of reproduction‐related disorders. However, studies which have identified plasma‐derived sEV miRNA as biomarkers of pregnancy or fertility status are limited. Higher abundance of miR‐25, miR‐16b and miR‐3596 has been observed in sEV isolated from plasma of cows which did not conceive after artificial insemination (AI) (Pohler et al. 2017 ). These studies demonstrate that sEV miRNA presents an innovative approach that may enhance our understanding of reproductive status. However, classical methods such as ultrasound (Balhara et al. 2013 ) and measuring pregnancy‐associated glycoproteins (PAGs) (Szenci 2021 ) have proven effective for early pregnancy diagnosis in dairy cows. Therefore, integrating sEV miRNA could provide further insights into these established techniques by exploring underlying biological processes associated with subfertility and pregnancy outcomes (Pohler et al. 2016 ). Blood is a rich source of biological information, including epigenetic, proteomic and gene expression data which reflects a cross‐section of the cellular micro‐environment, and is a minimally invasive diagnostic biosample (Nisenblat et al. 2016 ). We recently demonstrated that proteomic profiling of sEV enriched from the blood plasma of divergent cattle groups identified putative biomarkers of fertility phenotype in primiparous and peripubertal cattle (Turner et al. 2023 ). Here, our initial discovery study focuses on the differential expression (DE) of sEV miRNA derived from blood plasma collected before AI from an established divergently fertile animal model developed by researchers in New Zealand (Meier, McNaughton et al. 2021 , Meier, Kuhn‐Sherlock et al. 2021 ). This model consists of two extreme fertility groups: high fertile (HF) and low fertile (LF) primiparous cattle, which became pregnant or non‐pregnant after a single AI. Next‐generation miRNA sequencing was used to investigate the DE of sEV miRNA profiles between HF and LF groups to understand the post‐transcriptional regulatory landscape of the fertility phenotype/status. Subsequently, we utilised a second set of blood plasma samples from young heifers (a different animal group from the same fertility breeding value study) to validate sEV miRNA, which were identified in the discovery study, using qRT‐PCR miRNA panels. The main objective of the study was to identify any association between blood plasma sEV‐derived miRNA on dairy cow subfertility, presenting these findings as putative candidates for prognostic and diagnostic biomarkers of fertility for future studies.

The authors declare no conflict of interest.

Supplementary material : jex270084‐sup‐0001‐SuppMatt.docx Supplementary material : jex270084‐sup‐0002‐SuppMatt.xlsx Supplementary material : jex270084‐sup‐0003‐SuppMatt.xlsx Supplementary material : jex270084‐sup‐0004‐SuppMatt.xlsx