Abstract

   Extracellular vesicles (EVs) and their microRNA (miRNA) cargoes have
   garnered attention in the veterinary field for their regulatory role in
   various biological processes. This study aimed to (i) evaluate two
   techniques of EV isolation from sheep peripheral blood mononuclear cell
   (PBMC) supernatants using the ultracentrifugation (UC) and reagent
   (REA) methods and (ii) characterize the EV-miRNA profiles after an in
   vitro inflammatory environment mediated by lipopolysaccharides (LPS).
   Sheep peripheral blood was collected, and PBMCs were separated using a
   density gradient reagent. Subsequently, PBMCs were cultured at 37°C for
   24 h (5% CO[2]), and the supernatants were collected to perform the EV
   isolation. The presence of CD81^+ extracellular vesicle marker was
   determined, and the purity of isolated EVs was calculated as a ratio
   between the number of isolated EVs and the protein concentration.
   Moreover, the morphological characterization revealed mainly
   round-shaped structures with average sizes of 211 nm for EVs isolated
   by the UC method and 99 nm for EVs isolated by the REA method. Illumina
   NextSeq sequencing in a single-end mode was used to characterize the
   miRNA profile, and the differentially expressed (DE) miRNAs were
   analyzed using a combination of bioinformatics tools. The results
   revealed that the REA method is reliable for EV isolation from sheep
   supernatants. It was considered an improvement of the recovery rate and
   purity of EVs with the enhancement of the number and the expression
   levels of characterized miRNAs. The EVs isolated by the UC method after
   an LPS challenge showed 11 DE miRNAs, among which eight miRNAs were
   upregulated and three were downregulated. On the other hand, the REA
   method revealed an EV cargo in which eight DE miRNAs were upregulated
   and 21 DE miRNAs were downregulated. The master miRNA regulators of the
   biological process were identified by performing the MIRNA-mRNA network
   analysis, showing that, among the higher representative miRNAs based on
   the centrality and betweenness, the miR-26a-5p could have a crucial
   role in the resolution of inflammation. Moreover, the identification of
   the let-7 miRNA family in all the EVs showed potential targeted genes
   that regulate the inflammation and immune responses.

   Keywords: biomarkers, LPS, cell model, livestock, immune system, small
   ruminant

1. Introduction

   Recent studies in veterinary sciences have been showing a growing
   interest in extracellular vesicles (EVs) and their related molecular
   cargoes, particularly microRNAs (miRNA), which are considered potential
   disease biomarkers and therapeutic targets ([37]1–4), given their
   regulatory role in various biological processes ([38]5). EVs are
   defined as nanoparticles released by multiple cell types, and their
   biogenesis pathways include the endosome origin, with the release of
   exosome, and the plasma membrane-derived origin, with the release of
   microparticles/microvesicles ([39]6). EVs are classified according to
   their size as small EVs (<200 nm) or medium-large EVs (>200 nm)
   ([40]6). In a recent review by Moccia et al. ([41]7), the potential use
   of EVs as markers to diagnose diseases or as possible natural
   transporters of therapies or vaccines in veterinary studies is
   discussed. Indeed, EVs are largely involved in immune and cell–cell
   communication and facilitate the transfer of DNAs, mRNAs, microRNAs,
   and lipids to both nearby and distant recipient cargo cells in relation
   to the onset and development of many diseases ([42]8, [43]9).

   MiRNAs are described as small non-coding RNA (ncRNA) molecules of
   approximately 22 nucleotides in length that have a principal role in
   the post-transcriptional process as a gene repressor or a destroyer of
   targeted mRNA by binding with the complementary base pair on 3′UTR,
   5′UTR, or seeding and coding regions of targeted mRNA in plants,
   animals, and viruses ([44]10, [45]11). A large number of protein-coding
   genes are under the control of miRNA, which is tightly regulated
   ([46]10). Notably, miRNA species have been found to be significantly
   more abundant in EVs than in the cell of origin ([47]12). During the
   pathogen invasion or under injuries, miRNAs can play a pivotal role in
   the activation of immune and physiological processes that are important
   for the removal of pathogens and the maintenance of homeostasis.
   Excitingly, miRNAs have the ability to regulate the secretion of immune
   mediators by the immune cells together with the maturation and
   differentiation of dendritic cells, macrophages, granulocytes, and
   other immune cells. Additionally, during inflammation in the bone
   marrow, miRNAs also play a role in developmental processes ([48]13).
   Indeed, during inflammation and activation of the immune system, a
   class of miRNAs can work as important negative feedback loops in the
   immune system, whereas other miRNAs can be crucial for the
   amplification of the immune system’s response through repressing
   inhibitors of the response ([49]13).

   It is worth noting that the identification of the optimal technique to
   isolate EVs is essential in the purpose of searching for a novel
   biomarker. However, one of the main challenges in advancing our
   knowledge of EV functions is the lack of an efficient and standardized
   isolation strategy for isolating specific subpopulations and the
   absence of a gold standard method for isolating EVs ([50]14).
   Therefore, the International Society for Extracellular Vesicles (ISEV)
   proposed the Minimal Information for Studies of EV (MISEV) guidelines
   in 2014 ([51]14), upgraded in 2018, for improving EV research quality,
   recommending that there is no single optimal separation method;
   however, the choice ultimately depends on the specific downstream
   application and scientific question. Chen et al. ([52]15) reviewed the
   common EV separation techniques, demonstrating that each separation
   strategy has both advantages and disadvantages. As the most common
   method, ultracentrifugation is a mature technology that can be used for
   separating most samples at low operating expenses. However, it is
   considered time-consuming, with poor/unstable repeatability, and may
   lead to a possible co-purification with protein aggregates, which may
   affect the results of subsequent mass spectrometry or protein
   quantitation ([53]16–18). Furthermore, high-speed centrifugation may
   cause damage to EVs and reduce their biological activity ([54]16). On
   the contrary, the isolation method based on kits is an “easy and quick”
   procedure that can suffer from co-isolation among EVs; however, it can
   be considered an ideal choice for the identification of exosome-related
   disease biomarkers ([55]19).

   In the present study, based on the assumption that isolation methods
   could affect EV cargo, two methods were selected for EV isolation from
   sheep peripheral blood mononuclear cell (PBMC) supernatants to compare
   the miRNomic profile after an in vitro inflammatory challenge mediated
   by LPS.

   Hence, the present study aimed to (i) isolate EVs from sheep PBMC
   supernatants by using ultracentrifugation (UC) and total exosome
   isolation reagent (REA) methods and (ii) characterize the miRNomic
   profile after an inflammatory in vitro challenge, as a sheep model of
   inflammation. To the best of our knowledge, this is the first study
   that isolated EVs from sheep PBMC supernatant to analyze the miRNomic
   profile. Our hypothesis was focused on a potential connection between
   miRNAs, inflammation, and immune responses in sheep to define new
   predictive biomarkers associated with health or production phenotypes.

2. Materials and methods

2.1. Animals and experimental treatments

   The Gentile di Puglia dairy sheep breed reared in the grazing system
   was used in this study (n = 3). All procedures were conducted according
   to the guidelines of the EU Directive 2010/63/EU (2010) on the
   protection of animals used for experimental and other scientific
   purposes. The animals were carefully examined by veterinarians
   throughout the trial to exclude the presence of any signs of disease
   (N. 12,917 del 20220303 2022-UNFGCLE-0012917).

2.2. Peripheral PBMC isolation

   Blood samples (15 mL) were collected in Na-heparinized vacuum tubes
   from the jugular vein of sheep through a 21G needle, and the PBMCs were
   isolated by Histopaque®-1077 density gradient (Sigma Aldrich, Milan,
   Italy) according to Wattegedera et al. ([56]20) and modified as
   previously reported by Ciliberti et al. ([57]21). Briefly, whole blood
   diluted at a 1:1 ratio with cold PBS was centrifuged, and the white
   cell rings recovered after centrifugation were diluted in Hanks
   Balanced Salt Solution (HBSS, Thermo Fisher Scientific, Waltham, United
   States) and slowly layered on the Histopaque®-1077 solution (10 mL,
   Sigma Aldrich, Milan, Italy). The tubes were centrifuged at 400 × g for
   30 min at 20°C, and the buffy coat containing the PBMCs layered on the
   upper layer of Ficoll-Paque was recovered. The PBMC suspension was
   washed three times with HBSS wash buffer containing 2% fetal bovine
   serum (FBS) exosome-depleted (Thermo Fisher, Waltham, United States)
   and penicillin/streptomycin mix antibiotics (Biowest, Riverside, United
   States). Finally, the PBMCs were resuspended in RPMI 1640 medium
   without calcium and magnesium (Sigma Aldrich, Milan, Italy), containing
   10% FBS, penicillin/streptomycin antibiotics, and L-Glutamine solution.
   The PBMCs were counted using the trypan blue exclusion method on a
   Countess™ II Automated Cell Counter (Thermo Fisher Scientific, Waltham,
   United States). The viability of cells obtained after the PBMC
   isolation was higher than 98%. A final concentration of 1 × 10^6
   cells/mL was seeded into six well plates and cultured for 24 h at 37°C.
   The PBMCs were either stimulated with lipopolysaccharides (LPS) from
   Escherichia coli (1 μg/mL) or left unstimulated (CON).

2.3. Extracellular vesicle isolation from PBMC supernatants

   After the supernatant free-cells collection from PBMCs, the EV
   isolation was performed using two different methods as depicted in the
   flowchart of the experiment ([58]Figure 1).

Figure 1.

   [59]Figure 1
   [60]Open in a new tab

   Flowchart representing the two methods (UC, Ultracentrifugation; REA,
   Reagent) used for EV isolation from the supernatant of sheep PBMCs.

   The UC method was conducted according to Baharlooi et al. ([61]22) with
   some modifications. Briefly, the collected supernatants were
   centrifuged at 300 g for 5 min to avoid cell or cellular contamination.
   Subsequently, the supernatant was filtered using a 0.22 μm filter and
   centrifuged at 17,000 g for 30 min to remove microvesicles. The
   collected supernatant was then transferred to a polypropylene
   ultracentrifuge tube and subjected to first ultracentrifugation at
   100,000 × g for 120 min at 4°C with an Optima LE-80 K ultracentrifuge
   (Beckman Coulter, United States). Afterward, the supernatant was
   removed, and the EV pellet was suspended in PBS. Furthermore, the
   sample was ultracentrifuged at 100,000 × g for 60 min. All
   centrifugation steps were performed at 4°C. Finally, the supernatant
   was discarded and the pellet containing EVs was suspended in PBS and
   stored at −20°C for further analysis. The REA method was performed
   according to the manufacturer’s instruction of Total Exosome Isolation
   Reagent (from cell culture media, Thermo Fisher, Waltham, United
   States), and the isolation ratio was based on binding the water
   molecule and forcing the less-soluble molecules (EVs) out of the
   solution, resulting in a concentration of intact EVs and avoiding the
   time-consuming ultra-centrifugation procedure.

   The number of EVs isolated was determined using ExoELISA-ULTRA Complete
   Kit CD81 detection (SBI System Bioscience, Embarcadero Way Palo Alto,
   CA, United States), and a standard curve was generated to determine the
   number of EVs positive to tetraspanin CD81 (CD81^+) following the
   manufacturers’ recommendations.

   Both the REA method and the Elisa kit used to count the final number of
   isolated EVs were calibrated on the NanoSight LM10 instrument as
   reported in the respective manufacturer’s instructions. The isolated
   EVs were analyzed for the protein content after lysis in RIPA buffer
   following the procedure described by Subedi et al. ([62]23). The
   protein concentration was determined using the BCA Protein Assay kit
   (Thermo Scientific, Waltham, United States) according to the
   instruction manual, calibrated against the bovine serum albumin as a
   standard.

2.4. Zeta potential analysis and transmission electron microscopy

   The zeta potential of EVs was measured by laser Doppler electrophoresis
   (LDE) with a Nanosizer ZS (Malvern Instruments, Malvern, United
   Kingdom), as previously reported by De Leo et al. ([63]24).

   Transmission electron microscopy analyses were performed by a Jeol
   Jem-1011 microscope (JEOL USA, Inc., Pleasanton, CA, United States)
   operating at 100 kV, equipped by a high-contrast objective lens, and a
   W filament as an electron source, with an ultimate point resolution of
   0.34 nm. Images were acquired by a Quemesa Olympus CCD 11 MP camera.
   The samples were prepared by casting 3 μL of the aqueous dispersions of
   the isolated EVs onto 300 mesh amorphous carbon-coated Cu grids and
   then leaving the solvent to evaporate at room temperature. After
   deposition onto the grid, the samples were stained with a 2% aqueous
   solution of phosphotungstic acid hydrate. The size statistical analysis
   (EV average size and size distribution) of each sample (CON EVs
   isolated by UC and REA methods, and LPS EVs isolated using UC and REA
   methods) was performed on 100 nanostructures using a freeware Image J
   analysis program (National Institutes of Health, United States). The
   results are reported as mean ± standard deviation.

2.5. Library preparation and next-generation sequencing

   The small RNA-Seq kit [Bioo Scientific NextflexTM (v2, v3)] was used
   for the library preparation following the manufacturer’s instructions.
   Then, the libraries were sequenced in single-end mode on the Illumina
   NextSeq. Before further analysis, a quality check was performed on the
   raw sequencing data using the FastQC tool available on
   [64]http://www.bioinformatics.babraham.ac.uk/projects/fastqc for high
   throughput sequencing data. The quality trimming and adapter removal
   were done using the selected bioinformatics tool sRNAbench, following
   the procedure recommended by Aparicio-Puerta et al. ([65]25).

2.6. miRNA identification and differential expression analysis

   The sequences detected were analyzed using sRNAtoolbox ([66]25), to
   identify miRNA expression profiles using an ovis aries species data set
   contained in miRBase (version 22) ([67]26). The expression files
   generated with sRNAbench listed all copies of miRNAs. The NOISeq-Sim
   method was used to determine the differently expressed miRNAs of CON
   vs. LPS of the two EV isolating methods, and the probability of being
   differently expressed was calculated and set at >0.85.

2.7. Statistical analysis and bioinformatics

   Data on EV CD81^+ marker and protein quantification were checked by
   normality tests and analyzed with a one-way ANOVA of SAS ([68]27). The
   significance of the differences was assessed using the Tukey post hoc
   test for multiple comparisons and a p value lower than 0.05 was
   considered statistically significant. The data were presented as
   mean ± SEM.

   R was used to create a matrix of all transcripts expressed in all
   samples with the corresponding read counts, and the Bioconductor
   package NOISeq ([69]28, [70]29) was used to normalize the data using
   the RSEM method and then to perform the differential expression
   analysis using NOISeq-Sim ([71]28, [72]29).

   The data on downregulated and upregulated differentially expressed
   miRNAs were analyzed using a functional enrichment analysis tool
   (FunRich, [73]http://funrich.org/index.html) fed with human miRNA
   homologs retrieved through miRbase (version 22.2,
   [74]https://mirbase.org). The FunRich tool performs statistical
   analysis for different lists of proteins, peptide analysis, clustering,
   or more complex proteomic analysis ([75]30). The cutoff of the
   enrichment analyses was set to <0.05. The miRNAs-gene target
   interaction network analysis was performed and visualized using the
   miRNet web-based platform (version 2.0,
   [76]https://www.mirnet.ca/miRNet/home.xhtml), and the KEGG functional
   enrichment analysis was performed based on miRNet network analysis of
   the downregulated and upregulated EVs-miRNAs.

3. Results

3.1. EV quantification and characterization

   The presence of tetraspanin CD81^+, located in the EV membrane, was
   used to calculate the number of isolated particles, showing a
   significant difference among treatments (p < 0.001, [77]Figure 2A);
   indeed, the EV isolation performed using the REA method resulted in a
   higher quantity of EV isolation, apart from PBMC stimulation with LPS.
   Additionally, to calculate the purity of isolated EVs, the ratio of the
   number of EVs to protein concentration was calculated, as suggested by
   Tang et al. ([78]31), and significant differences were found between
   the two isolation techniques, resulting in a higher ratio in the REA
   method than in the UC method (p = 0.006, [79]Figure 2B). By means of
   TEM analysis, numerous round-shaped structures were clearly identified
   in the two types of samples, although slight polydispersion in particle
   size distribution emerged. In addition, a slight difference in the mean
   size of the EV samples was observed ([80]Figure 3). The statistical
   analysis returned an average diameter value of 211 ± 26 nm for the EVs
   isolated with UC from the CON sample and 99 ± 40 nm for the EVs
   isolated with REA from the CON sample ([81]Figures 3A,[82]B).
   Stimulation with LPS did not result in significant changes in the
   observed dimensions, and the mean diameter was 260 ± 29 nm for EVs
   isolated by the UC method and 67 ± 28 nm for EVs isolated by the REA
   method ([83]Figures 3C,[84]D). Zeta potential values of all EV
   preparations were negative, as previously observed for EVs isolated
   from biological fluids ([85]32). In particular, the measured values
   were − 12.3 ± 3.26 and − 27.2 ± 1.9 mV for the sample EVs from both UC
   and REA CON samples, respectively. Zeta potential values remained
   negative after stimulation with LPS, namely, −19.3 ± 6.8 mV for the EVs
   from the UC samples and 24.7 ± 0.78 mV for the EVs from the REA
   samples, suggesting good colloidal stability.

Figure 2.

   [86]Figure 2
   [87]Open in a new tab

   (A) Extracellular vesicle abundance (# of isolated extracellular
   vesicles ± SEM) using the two methods (UC, Ultracentrifugation; and
   REA, Reagent) in the presence and absence of LPS stimulus of
   inflammation; (B) Ratio of the number of extracellular vesicles
   isolated/protein concentration (mg/mL) of the ultracentrifugation (UC)
   and reagent (REA) methods in the presence and absence of LPS stimulus
   of inflammation. Data were presented as means ± SE.

Figure 3.

   [88]Figure 3
   [89]Open in a new tab

   Representative TEM images of samples of extracellular vesicles (EVs)
   and size distribution measured on 100 EVs for each sample. (A) CON EVs
   isolated by the ultracentrifugation (UC) method; (B) CON EVs isolated
   by the reagent (REA) method; (C) LPS EVs isolated by the UC method; and
   (D) LPS EVs isolated by the REA method.

3.2. Differentially expressed miRNAs between isolation methods

   To compare the overlapping miRNAs between EV-miRNA cargoes from UC and
   REA isolation techniques, regardless of the presence of a stimulant, a
   Venn diagram was built ([90]Figure 4). A total of five miRNAs for CON
   UC and 12 miRNAs for LPS UC were identified; the unique miRNAs found in
   REA CON and REA LPS were 42 and 24, respectively. The expression level
   for each mature miRNA characterized in all samples was reported in
   [91]Supplementary Figure S1.

Figure 4.

   Figure 4
   [92]Open in a new tab

   Venn diagram of extracellular vesicles-miRNAs characterized in PBMC
   supernatants by UC and REA methods and in the presence and absence of
   LPS stimulus of inflammation using the Funrich tool.

   The significant differentially expressed (DE) miRNAs of EVs from PBMCs
   isolated by the UC method (CON vs. LPS) showed 11 DE miRNAs in total,
   among which eight miRNAs were upregulated (oar-miR-26b, oar-miR-23a,
   oar-let-7i, oar-let-7f, oar-miR-16b, oar-miR-30c, oar-miR-21, and
   oar-miR-27a), with fold change (FC) > 1.5 and probability > 0.85, and
   three were downregulated (oar-miR-125b, oar-miR-143, and oar-let-7a),
   with FC < 1.5 and probability > 0.85 ([93]Figure 5A). On the contrary,
   using the REA method (CON vs. LPS) showed that among the DE miRNAs,
   eight were upregulated (oar-miR-369-3p, oar-miR-199a-3p, oar-miR-194,
   oar-miR-10b, oar-let-7b, oar-let-7d, oar-let-7i, and oar-let-7a), with
   FC > 1.5 and probability > 0.85 ([94]Figure 5B), and 21 miRNAs were
   downregulated (oar-miR-27a, oar-miR-10a, oar-miR-152, oar-miR-181a,
   oar-miR-17-5p, oar-miR-30c, oar-miR-19b, oar-miR-99a, oar-miR-374a,
   oar-miR-106b, oar-miR-200c, oar-miR-221, oar-miR-25, oar-miR-30d,
   oar-miR-150, oar-miR-29b, oar-miR-191, oar-miR-26b, oar-miR-29a,
   oar-miR-23a, and oar-miR-148a), with FC < 1.5 and probability > 0.85
   ([95]Figure 5B). Among the DE miRNAs, the REA and UC methods had in
   common six miRNAs among which oar-miR-26b, oar-miR-30c, oar-miR-27a,
   and oar-miR-23a were upregulated in UC and downregulated in REA, the
   miRNA oar-let-7i was upregulated in both the UC and REA methods, and
   oar-let-7a was upregulated in the REA method and downregulated in the
   UC method.

Figure 5.

   [96]Figure 5
   [97]Open in a new tab

   (A) log[2] fold change of upregulated (green) and downregulated (red)
   miRNAs from extracellular vesicles (EVs) of CON vs. LPS isolated by the
   ultracentrifugation (UC) method. (B) log[2] fold change of upregulated
   (green) and downregulated (red) miRNAs from EVs CON vs. LPS isolated by
   the Reagent (REA) method.

3.3. Functional and pathway enrichment analysis for identified miRNA target
genes

3.3.1. Ultracentrifugation method

   To gain a better understanding of the function and mechanism of miRNAs,
   we analyzed the data on the downregulated and upregulated DE miRNAs. We
   used the online enrichment tool Funrich to carry out this analysis, and
   the biological processes (BPs), biological pathways (BPath), and
   molecular functions (MFs) were analyzed using this tool. After the LPS
   challenge, the downregulated miRNAs did not result in any significantly
   enriched genes in the BP term ([98]Figure 6A). On the contrary, the
   upregulated miRNAs resulted in enriched genes in the BP term related to
   the regulation of nucleobase, nucleoside, nucleotide, and nucleic acid
   metabolism ([99]Figure 6B). Regarding the BPath term, the significant
   genes enriched for the downregulated miRNAs ([100]Figure 6C) were
   implicated in the p53 pathway (7.3%, p = 0.004), direct p53 effectors
   (5.9%, p = 0.007), and plasma membrane estrogen receptor signaling
   (28.8%, p = 0.038). The upregulated miRNAs ([101]Figure 6D) resulted in
   the six main representative significantly enriched genes related to
   plasma membrane estrogen receptor signaling (28.8%, p = 0.005), ErbB
   receptor signaling network (29%, p = 0.005), signaling events mediated
   by VEGFR1 and VEGFR2 (28.8%, p = 0.004), TRAIL signaling pathway
   (29.4%, p = 0.004), sphingosine 1-phosphate (S1P) pathway (29.2%,
   p = 0.003), and the VEGF and VEGFR signaling network (29.2%,
   p = 0.002).

Figure 6.

   [102]Figure 6
   [103]Open in a new tab

   Biological process (A,B), biological pathways (C,D), and molecular
   function (E,F) of down and up regulated extracellular vesicles miRNAs
   isolated from PBMC supernatants by using ultracentrifugation (UC)
   method.

   Regarding the MF term, the downregulated miRNAs did not result in any
   significantly enriched genes ([104]Figure 6E), whereas the upregulated
   miRNAs resulted in the enriched genes involved in the significant
   transcription factor activity (7.5%, p < 0.001, [105]Figure 6F).

3.3.2. Reagent method

   The main enriched genes involved in the downregulated ([106]Figure 7A)
   and upregulated ([107]Figure 7B) miRNAs of the BP terms characterized
   in EVs isolated using the REA method (CON vs. LPS) were implicated in
   cell communication and signal transduction. In addition, the enriched
   genes of the downregulated miRNAs were also implicated in the
   regulation of nucleobase, nucleoside, nucleotide, and nucleic acid
   metabolism of the BP term (20%, p < 0.001).

Figure 7.

   [108]Figure 7
   [109]Open in a new tab

   Biological process (A,B), biological pathways (C,D), and molecular
   function (E,F) of down and up regulated extracellular vesicles miRNAs
   isolated from PBMC supernatants by using Reagent (REA) method.

   The enriched genes of the BPath terms involved in both downregulated
   ([110]Figure 7C) and upregulated ([111]Figure 7D), miRNAs were all
   significant (p < 0.001). In particular, the enriched genes with the
   highest percentages in both downregulated and upregulated miRNAs were
   implicated in the beta1 integrin cell surface interactions of the BPath
   term (33.7 and 33%, respectively).

   The significantly enriched genes of the main MF term were related to
   the transcription factor activity for both downregulated (8.7%,
   p < 0.001, [112]Figure 7F) and upregulated miRNAs (7.1%, p = 0.011,
   [113]Figure 7E). Moreover, the enriched genes of the downregulated
   miRNAs were also implicated in the significant transcription regulator
   activity (6.4%, p = 0.006), followed by ubiquitin-specific protease
   activity (3.7%, p < 0.001).

3.4. Identification of extracellular vesicle miRNAs from PBMC

3.4.1. miRNA-gene target interaction networks and KEGG pathways

   The interaction network between miRNAs and their potential target genes
   is crucial for identifying the biological relevance of miRNAs.
   Accordingly, in the UC method, [114]Figure 8A shows the ForceAtlas of
   the downregulated EV human homolog miRNAs in which hsa-let-7a-5p had a
   higher centrality and betweenness than hsa-miR-125b-5p and
   hsa-miR-143-3p. Among the upregulated EVs-miRNAs from UC, the human
   homolog hsa-miR-26a-5p showed higher centrality and betweenness than
   hsa-let-7i-5p ([115]Figure 8B).

Figure 8.

   [116]Figure 8
   [117]Open in a new tab

   MiRNA-gene target network analysis. The centric target network of
   upregulated (A) and downregulated (B) target genes in
   ultracentrifugation (UC) samples of EVs by miRNet software. The boxes
   represent miRNAs (blue) with the size resulting from the betweenness of
   the miRNAs in constructing the network. The nodes represent the genes
   (red).

   In the REA method, the downregulated EVs-miRNAs of the human homologs,
   hsa-miR-26a-5p and hsa-miR-17-5p, had the highest centrality and
   betweenness ([118]Figure 9A). Furthermore, among the upregulated
   miRNAs, hsa-let-7b-5p, hsa-let-7a-5p, and hsa-let-7i-5p resulted in
   higher centrality and betweenness than the other upregulated miRNAs
   ([119]Figure 9B). To further explore the biological function of the
   predicted target genes, the KEGG pathway analysis of the significantly
   downregulated and upregulated miRNAs in both UC and REA samples was
   reported ([120]Supplementary Figures S2, [121]S3).

Figure 9.

   [122]Figure 9
   [123]Open in a new tab

   MiRNA-gene target network analysis. The centric target network of
   upregulated (A) and downregulated (B) target genes in reagent (REA)
   samples of EVs by miRNet software. The boxes represent the miRNAs
   (blue) with size resulting from the betweenness of the miRNAs in
   constructing the network. The nodes represent the genes (red).

4. Discussion

   In recent years, the putative role of miRNAs as biomarkers in farm
   animal diseases and as a therapeutic approach based on their regulatory
   key role in disease has been explored ([124]33). To the best of our
   knowledge, this is the first study in which the characterization of the
   miRNA profile of EVs isolated from supernatants of sheep PBMCs,
   challenged in vitro with LPS, has been investigated. One of the most
   significant scientific challenges to date is the isolation and
   characterization of EVs. Understanding the importance of EVs-miRNAs is
   a crucial aspect of this research ([125]14, [126]34). Therefore, the
   identification of the optimal technique to isolate EVs is crucial when
   the main purpose is the characterization of miRNAs for further
   biomarker discoveries. Particularly, in the present study, we focused
   on two methods of EV isolation starting from cell culture supernatants.
   EV isolation using the UC method is a traditional technique, but it
   requires prior training and can be quite tedious and time-consuming.
   Additionally, the results may be sensitive to the technique used
   ([127]35). Our results demonstrated that the UC method decreases the
   recovery rate of CD81^+ EVs, as also confirmed by the low purity
   calculated as the ratio of particle number to protein concentration.
   According to the studies conducted by Helwa et al. ([128]35) and Lane
   et al. ([129]36), the UC isolation method resulted in a lower recovery
   of EVs compared to commercially available kits. This result can be due
   to the different centrifugation forces over multiple cycles, useful for
   removing cell debris and other contaminants, which can also cause a
   loss of EVs from the sample, leading to a lower and more variable EV
   yield ([130]37). According to the study conducted by Tang et al.
   ([131]31), the number of EVs isolated using the UC method was lower
   compared to that obtained using two commercial kits specifically
   designed for EV isolation (ExoQuick and Total Exosomes Isolation
   Reagent_TEI). However, even if the UC method resulted in a lower number
   of EVs, more enriched EV markers by using Western blotting were
   obtained ([132]31). In contrast with these findings, the use of a
   centrifuge at high speeds can affect the EVs’ physical properties and
   the sensitivity of the proteomic analysis ([133]38, [134]39).
   Accordingly, in the present study, the EV-miRNA characterization
   revealed that the UC method reduces the EV-miRNA cargoes, in terms of
   the number of miRNAs identified, as depicted in the Venn diagram. This
   result is consistent with the lower EV recovery rate and purity due to
   the UC isolation method. To date, this study is the first on EV
   characterization from sheep PBMC supernatants; therefore, it will be a
   crucial step to confirm our data with additional EV protein
   characterization based on the methods reported in the MISEV2018
   guidelines.

   MiRNAs are recognized as playing a key role in regulating gene and
   biological processes such as cell proliferation, apoptosis,
   differentiation, and tumorigenesis ([135]40, [136]41). Mature miRNAs
   regulate the expression of the target gene by recognizing and
   complementing completely or incompletely the target mRNA sequences
   through 5–8 nucleotides at its 5′ end and then inducing the degradation
   or translational inhibition of target mRNA; thus, mature miRNAs
   regulate the expression of the target gene ([137]42).

   In veterinary studies, miRNAs have been proven to mediate cellular
   immunity, apoptosis, signal transduction, and cell differentiation
   ([138]43, [139]44). In particular, the initiation of inflammatory
   stimuli triggers both miRNAs and protein-coding genes. On the one hand,
   the expression of many miRNAs is regulated by silencers, enhancers, and
   epigenetic modification such as hypomethylation and hypermethylation in
   miRNA promoter and transcription factors ([140]45, [141]46). Of these
   transcription factors, the nuclear factor-κB (NF-κB) is the one that
   regulates the expression of various inflammation-associated miRNAs
   (miR-146a and miR-155) ([142]47–49) and functions as a strong
   pro-inflammatory signal ([143]47), inducing the expression of
   proinflammatory genes, including cytokines and chemokines ([144]50). In
   dairy cows, miRNAs can exert a post-transcriptional regulation of
   endocrine, metabolic, and immune reactions activated by the energy
   imbalance occurring during the transition to lactation ([145]51,
   [146]52). In particular, in circulating miRNA profiles of transition
   dairy cows, a selection of miRNA subsets have been identified as
   putative biomarkers for metabolic disorders, with these biomarkers
   being the primary affected pathways linked to the metabolic and immune
   adaptation ([147]52). In the present study, the EVs-miRNAs,
   characterized using both methods and analyzed considering the in vitro
   LPS challenge, showed a different subset to upregulated and
   downregulated miRNAs. Notably, the REA and UC methods had in common,
   among the DEM, the miR-26b, which was upregulated in the UC method and
   downregulated in the REA method. The miR-26 family has been recognized
   to target regulators that control the development ([148]53), cell type,
   and differentiation of tumors ([149]54). Zhu et al. ([150]55) reported
   that, in both physiological and pathological conditions, miR-26b is
   expressed concomitantly with its host genes, cooperating to block the
   G1/S-phase transition of proliferation through the activation of the
   pRb protein ([151]55). The overexpression of miR-26b was directly
   related to the composition of milk fatty acids in an in vitro study on
   goat mammary epithelial cells ([152]56). In addition, the role of
   miR-26b was associated with numerous metabolic diseases, playing a
   crucial role in adipocyte development ([153]56) and lipogenesis
   ([154]57). The synthetic bta-miR-26b was found to downregulate three
   immune system-related genes in primary bovine endometrial epithelial
   cells (EECs); thus, bta-miR-26b has been a candidate miRNA among the
   potential miRNAs that regulate the immune system during
   peri-implantation periods ([155]58).

   The effect of miR-17-5p on inflammation mediated by LPS was tested in
   an in vitro study on RPMI2650 cells conducted by Huang et al.
   ([156]59), who found that the overexpression of miR-17-5p intensifies
   the inflammation through the negative regulation of Smad7 protein
   expression, which exerts a protective role by the inactivation of NF-κB
   and Wnt/β-catenin pathways. Moreover, miR-17-5p acts as a key regulator
   of the G1/S cell cycle de-coupling negative regulators of the MAPK
   signaling cascade ([157]60). In the present study, KEGG analysis
   revealed the most enriched pathways of the target genes of
   downregulated and upregulated miRNAs, the implication of the MAPK
   signaling pathway. The MAPK family has a central role in many
   physiological processes such as apoptosis and inflammation, with direct
   or indirect regulation of the NF-κB activity, which is the key
   transcription factor to promote the expression of inflammatory
   cytokines ([158]61). The recruitment of the MAPK pathways is activated
   by pathogen-associated molecular patterns (PAMPs), molecular portions
   present in invading pathogens, and by damage-associated molecular
   patterns (DAMPs), molecules produced endogenously in times of
   physiological stress, which support a robust inflammatory response
   including cytokine production and leukocyte activation ([159]62). In
   the present study, the implication of the MAPK pathway of the target
   genes of upregulated miRNAs could establish the relationship between
   the post-transcriptional information driven by EV-miRNAs and the
   response to the inflammation mediated by LPS. Previous studies stated
   that the cell injury induced by LPS can be mediated by the modulatory
   action of miRNAs ([160]63). Moreover, results from the present study,
   from the interaction network between miRNAs and their potential target
   genes, suggest that both miR-26b and miR-17-5p could be considered
   among the master regulators of the biological processes activated by
   LPS. No previous studies have been conducted on sheep PBMC supernatant
   miRNA characterization; however, it could be hypothesized that miR-26b
   and miR-17-5p can have a crucial role in the inflammation process
   mediated by LPS in blood cells and in its resolution, even if further
   studies to better elucidate this hypothesis are needed.

   From the interaction network between miRNAs and their potential target
   genes data analysis, it was found that there is a high number of let-7
   family miRNAs present, among which oar-let-7a, oar-let-7b, oar-let-7d,
   and oar-let-7i emerged as the most prominent. In the study of Izumi et
   al. ([161]64), the let-7a miRNA was found to be highly expressed in
   bovine raw milk exosomes (MEVs) and was also present in human and pig
   milk exosomes ([162]65, [163]66). The highly expressed miRNAs of milk
   exosomes from several species exhibited common functions, such as
   regulating immune function and intestinal maturation ([164]64). Indeed,
   the role of microRNAs in mammary gland development, health, and
   function of cattle, goats, and sheep was demonstrated ([165]67).
   Moreover, metabolomic and transcriptomic approaches applied to MEVs
   from cows, donkeys, and goats revealed that numerous metabolic pathways
   implicated in immunomodulation ([166]68) and miRNA targets with
   enriched terms related to the immunity modulation, protein synthesis,
   and cellular cycle regulation were common in the cow, donkey, and goat
   species ([167]69).

   The let-7f and let-7b miRNAs are highly expressed in both cow and sheep
   milk with established immune-related functions ([168]70). The let-7b
   targets toll-like receptors (TLR), regulating the activation of NF-κB,
   and its downstream genes were related to the inflammation and immune
   responses during an in vitro Helicobacter pylori infection of human
   gastric epithelial cells ([169]71). The miRNA profiling of sheep
   responding to the infection of small ruminant lentiviruses showed that
   DE miR-21, miR-148a, and let-7f regulated the seronegative and infected
   sheep, therefore demonstrating their potential implications for the
   host-virus interaction ([170]72). Moreover, the let-7a miRNA targets
   the IL-6 gene, resulting in a downregulation of tumors related to an in
   vitro cell model of inflammation ([171]73, [172]74). In the present
   study, the let-7 miRNA family was implicated under the LPS inflammatory
   challenge, showing both upregulated and downregulated miRNAs, thus
   demonstrating its regulatory biological function during inflammation in
   sheep PBMC. Moreover, a promising implication of the IL-6 gene may
   reveal a possible cross-link between inflammatory pathways and miRNAs,
   as recently reviewed by Chatterjee et al. ([173]75), which needs to be
   further explored. Indeed, IL-6 is well known as a potential biomarker
   of inflammation both in human and ruminant studies ([174]21, [175]76,
   [176]77).

   From the KEGG analysis emerged the involvement of the p53 signaling
   pathway from the gene target of all characterized DE miRNAs. It has
   been suggested that the proper balance between the p53 and MAPK
   signaling pathways is necessary after a stressful stimulus is applied
   to the cells as this will determine whether the cells survive or die.
   The involvement of the p53 signaling pathway in this balance has been
   proposed as a potential link between the two pathways ([177]78).
   Indeed, the activation of p53 causes growth arrest and apoptosis
   through the suppression of a number of genes that may favor cell
   survival ([178]78). MiR-143 is classified as a tumor suppressor by
   arresting the G0/G1 phase cells and promoting caspase-3-apoptosis
   ([179]79). Moreover, a specific apoptosis mechanism activated by
   miR-143 is involved in the targeting of MDM2 and, indirectly, in the
   activation of the p53 pathway, which in turn activates the
   transcription of miR-143, generating a short miR-143-MDM2-p53 feedback
   loop ([180]18, [181]80, [182]81). Previous evidence could explain the
   characterized downregulation of oar-miR-143 in the UC sample and the
   concomitant augmentation of the biological p-53 pathway.

   Additionally, the KEGG analysis showed that the most enriched pathways
   of the target genes of both upregulated and downregulated DE miRNAs
   from the UC and REA methods after LPS stimulation were associated with
   the JAK/STAT signaling pathway. In the study of Yang et al. ([183]82),
   it was found that miR-203a-3p induces the polarization of macrophage-M2
   by downregulating the suppressor of cytokine signaling 3 (SOCS3)
   expression and activating JAK/STAT3 signaling ([184]82). It is well
   known that SOCS is a protein family of eight members (SOCS1–7 and CIS)
   that inhibit STAT activation by acting on JAK/STAT activating
   receptors. A previous study highlighted that miRNAs are involved in the
   regulation of innate immune and inflammatory responses ([185]83). In
   ruminant studies, the nexus between circulating miRNAs and the cytokine
   profiles could be considered a key factor in studying their disease
   susceptibility. In a study conducted by Naylor et al. ([186]84) on
   lambs challenged with LPS at a systematic level, it was observed that
   there was a simultaneous increase in both immune and stress biomarkers.
   Some of the biomarkers that showed an increase were IL-6, cortisol, and
   certain miRNAs, including miR-145, miR-1246, and miR-223. Moreover, it
   was demonstrated that, following LPS-mediated inflammation in a murine
   dendritic cell line derived from bone marrow, miR-369-3p can suppress
   the inflammation by reducing the secretion of proinflammatory molecules
   (TNFα, IL-6, IL-12, IL-1α, and IL-1β) while increasing the levels of
   IL-10 and IL-1RA, with anti-inflammatory roles ([187]85). In the
   present study, miR-369-3p was upregulated in REA probably as a result
   of its suppressive action in inflammation mediated by LPS. In a study
   conducted by Gholami et al. ([188]86), some miRNAs were involved in the
   initiation of inflammatory response by modulating many immune responses
   including the secretion of cytokines and chemokines such as miR-26a.
   Accordingly, to select potential miRNAs as novel biomarkers in sheep,
   it will be fundamental to understand the specific functions and the
   mechanisms of interaction between miRNAs and other biomolecules
   involved in disease pathogenesis, which include proinflammatory
   cytokines and plasma proteins ([189]87).

5. Conclusion

   In the present study, two different isolation methods were used to
   compare the EV-miRNome from sheep supernatant PBMCs. The data
   demonstrated that the REA method is reliable for EV isolation,
   enhancing the recovery rate and the purity of CD81^+ EVs, and improving
   the number of miRNAs characterized. However, this study being the first
   on sheep supernatant EV characterization, additional methods for more
   in-depth protein characterization as reported in the MISEV2018
   guidelines are going to be performed in the future.

   Among the DE miRNAs, the EVs from PBMC isolated using the UC method
   displayed eight upregulated miRNAs and three downregulated miRNAs,
   while the REA EV isolation method was characterized by eight
   upregulated miRNAs and 21 downregulated miRNAs. Moreover, the tested
   isolation methods of the DE miRNAs had in common the presence of
   miR-26b and miR-17-5p, which could play a crucial role in the
   inflammation resolution mediated by LPS, as well as the let-7 miRNA
   family characterized in all the EV targeted genes that regulate the
   inflammation and immune responses.

   The present study’s data can provide insights into the putative role of
   EVs-miRNAs in a sheep model of inflammation mediated by an LPS
   challenge.

Data availability statement

   The datasets generated for this study are available on request to the
   corresponding author.

Ethics statement

   The animal study was approved by Scientific Ethics Commitee of the
   University of Foggia. The study was conducted in accordance with the
   local legislation and institutional requirements.

Author contributions

   MCi, MCa, and AgS contributed to the conception and design of the
   study. MCi supervised the overall project and performed the statistical
   analysis. MCi, VL, CI, and LC performed the analysis. MCi and AnS
   organized the database. MCi, MA, AnS, and MCa wrote the first draft of
   the manuscript. VL, CI, and LC contributed to the revision of the
   manuscript. All authors contributed to the article and approved the
   submitted version.

Acknowledgments