MiRNA biogenesis is classified into the canonical and non-canonical pathways (41) ( Figure 1 ). In the canonical pathway, miRNA synthesis starts with a large primary transcript (pri-miRNA), an RNA stem-loop with a cap at 5’, and a polyadenine at 3’, usually transcribed by the RNA polymerase II (42). The 5’ cap and 3’ polyadenine of the pri-miRNA are cleaved in the cellular nucleus by a microprocessor complex formed by the type-III RNase Drosha and a double-stranded RNA binding protein, DiGeorge syndrome critical region 8 (DCGR8) in mammals, to produce the pre-miRNA, which only contains the stem-loop structure (43). The pre-miRNA is translocated to the cytoplasm by the Ran/GTP/Exportin 5 complex, where it is processed by Dicer, another type-III RNase, which breaks the stem-loop and forms a miRNA-miRNA duplex. One of the strands is the mature miRNA, while the other is the “passenger” strand (44). The non-canonical pathway usually starts with an ncRNA, such as an intron or a short hairpin RNA (shRNA). The shRNA is initially cleaved by the microprocessor complex and exported to the cytoplasm via Ran/GTP/exportin5. The pre-miRNAs are further cleaved through the Argonaute-2 (AGO2), but Dicer-independent, cleavage (45). Mirtrons, which are produced from the introns of mRNAs during splicing, and 7-methylguanosine (m⁷G)-capped pre-miRNAs are dependent on Dicer to complete their cytoplasmic maturation, but they differ in their nucleocytoplasmic shuttling. Mirtrons are exported via Exportin5/Ran/GTP, while m⁷G-pre-miRNA are exported via Exportin1. The mature miRNA generated from the canonical or non-canonical pathways is integrated into the RNA-induced silencing complex (RISC), which guides the miRNA to its target mRNA. The binding of miRNAs to their target mRNAs induces gene silencing and/or mRNA degradation. The binding of miRNAs to the 5’-untranslated region (5’UTR) and other DNA elements could enhance translation and transcription (46). miRNAs tend to associate with proteins and/or travel in serum and plasma encapsulated within vesicles, which makes them more resistant to degradation than other RNAs and allows them to be measured with high sensitivity (47).

The first study analyzing miRNA expression in MAP-infected animals did not show differential miRNA patterns between interferon-γ (IFN-γ) responders and unchallenged controls, probably due to limitations of the available miRNA annotation (48). However, the authors did identify differential expression of miR-205, miR-10b, miR-92b, miR-432, miR-27a, miR-127, miR-126, and miR-143 between controls at time point zero and controls at 6 months post-infection. These miRNAs are associated with proliferation and development roles in different mammal species, which suggests that miRNA expression also changes during development processes. Shaughnessy et al. did not detect differentially expressed (DE) miRNAs in serum at either the early (6 months) or late (43, 46, and 49 months) intervals between seropositive and seronegative animals. However, some miRNAs, such as bta-miR-92b, bta-miR-29a, bta-miR-205, and bta-miR-345-3p, were DE in seropositive animals at time point 0 and time points 6, 43, 46, and 49 months. These miRNAs are associated with immune functions, suggesting that they could be associated with immune maturation.

A summary of the miRNAs with differential expression between MAP-infected and uninfected cattle is shown in Table 2 . Wang et al. identified DE miRNAs in monocyte-derived macrophages (MDMs) infected with MAP 6 hours after infection (49). These DE miRNAs were associated with modulation of the MAPK pathway (miR-12023 and miR-1343-3p), the NF-kβ pathway (miR-214, bta-miR-133a, and bta-miR-1246), and apoptosis (bta-miR-1246, bta-miR-1306, bta-miR-454, and bta-miR-150. Using BoMac cells and MDMs, Wright et al. showed that MAP can modulate miR-19a, miR-129, miR-24, and miR-24-3p expression to induce macrophage polarization (50). MiR-19a and miR-129 are associated with an M1 macrophage phenotype and regulate lipid efflux and inflammation. MiR-24-3p, however, inhibits phagocytosis and promotes the M2 macrophage phenotype. These findings suggest that MAP can interfere with inflammation and macrophage function by altering miRNA expression. Additional studies in MAP-infected bovine macrophages found that bta-miR-133b, bta-miR-92a, and bta-miR-184 were downregulated upon MAP infection (51). Target gene function analysis suggested that miR-92a was involved in interleukin signaling, hence regulating the immune response against MAP.

An analysis of whole-blood samples collected from MAP-infected cows and uninfected cows from a MAP-free herd based on ELISA results and absence of clinical cases revealed differential expression of miRNAs with roles in the regulation of the immune response and inflammation upon bacterial or viral infection (52). In the infected cows, downregulation of bta-miR-100 and bta-miR-301a and upregulation of bta-miR-32 were observed. The targets of these miRNAs associated with phagocytosis, macrophage function and cytokine production were the Tweety Family Member 3 ( TTYH3), HIC ZBTB Transcriptional Repressor 1 (HIC1), Inosine Monophosphate Dehydrogenase 1 (IMPDH1), ARF Like GTPase 2 (ARL2), Adaptor Related Protein Complex 2 Subunit Alpha 1 (AP2A1), and Zinc Finger and BTB Domain Containing 4 (ZBTB4). Other differentially expressed miRNAs were bta-mir-19b, bta-mir-19b-2, bta-mir-1271, and a novel miRNA (Novel:14_7917). The authors hypothesize that the identified miRNAs may be involved in regulating early immune responses.

MAP infection can also result in miRNA upregulation in the ileal segments of newborn calves, such as miR-146b, which inhibits proteins in the TLR signaling pathway such as interleukin-1 receptor-associated kinase 1 (IRAK1), interleukin-1 receptor-associated kinase 2 (IRAK2), and tumor necrosis associated factor 6 (TRAF6) (25). Therefore, MAP could be interfering with TLR signaling and downstream inflammatory responses. This study also found an upregulation of bta-miR-202, bta-miR-184, bta-miR-196b, and bta-miR-1247, associated with lymphocyte activation, muscle and epithelium development, inflammatory response, the proliferation of endothelial cells, phagocytosis, and epigenetic regulation. Downregulation of bta-miR-133b, bta-miR-433, bta-miR-105a, and bta-miR-137, associated with lymphocyte activation, the proliferation of endothelial cells, muscle and epithelium development, and vesicle transport, was also detected.

Fecal samples have also been used to identify DE miRNAs between cows with clinical signs of PTB and with a bacteriological culture and ELISA-positive result vs. healthy cows (53). The results showed downregulation of miR-501-5p and miR-658, and upregulation of miR-92a-3p. When comparing miRNA-Seq data obtained from serum samples of cows in different stages of PTB based on ELISA and PCR results and clinical signs vs. control cows, Choi et al. detected differential expression of the bta-miR-363, bta-miR-196a, bta-miR-147, bta-miR-655, bta-miR-346, and bta-miR-2426, associated with the Ras and Wnt pathways (54). Two miRNAs (bta-miR-374b and bta-miR-2887) were identified in cows with PCR-positive and ELISA-negative results and with key molecules such as interleukin-10 (IL-10) and transforming growth factor beta (TGF-β1).

Previous studies have compared the miRNA profiles of cattle with positive and negative ELISA and fecal PCR results (52–55). In cattle, however, focal lesions in gut tissues can be detected before fecal shedding and ELISA. Therefore, examining the miRNA profiles of MAP-infected animals selected according to the presence of focal histopathological lesions facilitates the identification of miRNAs in animals in the subclinical stage of the infection when the amount of MAP and antibodies cannot be detected with current pre-mortem diagnostic methods. Our research team recently used miRNA-Seq to quantify and compare the expression of miRNAs in ICV and whole-blood samples collected from cows with focal or diffuse PTB-associated lesions in gut tissues vs. control cows without lesions (56). To our knowledge, this was the first study that used sRNA-Seq to compare miRNA expression in blood and gut tissues of cattle classified according to histopathological data. None of the identified miRNAs were differentially expressed in both the PB and ICV samples, and, therefore, our results showed PB and ICV-specific miRNAs expression. Specifically, in blood samples, we identified eight miRNAs in the comparison of cows with diffuse lesions vs. controls (bta-miR-2425-3p, bta-miR-144, bta-miR-19a, bta-miR-32, bta-miR-139, bta-miR-101, bta-miR-27a-5p, and bta-miR-181a) and three in the comparison of cows with diffuse vs. focal lesions (bta-miR-19a, bta-miR-144, and bta-miR-32). Among the eight miRNAs differentially expressed in blood samples from cows with diffuse lesions vs. controls, three (bta-miR-19a, bta-miR-144, and bta-miR-32) were also found downregulated in PB samples of cows with diffuse vs focal lesions. Interestingly, bta-miR-32 was also downregulated in a previous study where the miRNA profiles of Holstein cattle positive and negative for MAP antibodies were compared (52). In the ICV samples, we identified 4, 5, and 18 miRNAs that were differentially expressed in cows with focal lesions vs. controls, diffuse lesions vs. controls, and diffuse vs. focal lesions, respectively. This suggests that the miRNA expression in ICV changes more as PTB-associated lesions become more severe. In the comparison of cows with focal lesions vs. controls, two upregulated (bta-miR-2478 and bta-miR-150) and two downregulated (bta-miR-23a and bta-iR-23b-3p) miRNAs were identified. DE miRNAs in the comparison of cows with diffuse lesions vs. controls (bta-miR-433, bta-miR-146a, and bta-miR-99a-5p) were also dysregulated in the comparison of cows with diffuse vs. focal lesions. Interestingly, the human ortholog of bta-miR-146a is an inhibitor of the pro-inflammatory immune response (57). We, therefore, speculated that the upregulation of miR-146a would negatively regulate TRAF6 and control the pro-inflammatory immune response (58). Two miRNAs (bta-miR-23a and bta-miR-23b-3p) were downregulated in both comparisons, i.e., in cows with focal lesions vs. controls and diffuse vs. focal lesions. Some of the miRNAs identified in our study had been previously identified in studies describing miRNA expression in response to mycobacterial infections, including bta-miR-27a-5p, bta-miR-32, bta-miR-23a, bta-miR-135b, bta-miR146a, bta-miR146b, bta-miR-433, bta-let-7e, and bta-let-7c (25, 52, 59, 60).

This process is mediated by the spliceosome, a complex formed by highly dynamic small nuclear ribonucleoproteins (snRNPs) (U1, U2, U4, U5, and U6) and accessory proteins (61–63) ( Figure 2A ). The same mRNA can be spliced in different ways to produce different proteins. Splicing events can be categorized depending on whether they are always (constitutive) or only sometimes (alternative) recognized by the spliceosome and spliced in the mature mRNA. The most frequently used splice sites contain highly conserved sequences easily recognized by the spliceosome and lead to constitutive splicing (CS), whereas less frequently used splice sites lead to AS and are only used when specific splicing regulatory elements are activated. Depending on the splice site locations, AS events can be classified into different classes ( Figure 2B ). AS events can have multiple detrimental consequences in the mature transcript, such as a decrease in mRNA stability, a change in protein function or target, a change in final protein location, and the generation of premature termination codons, which are natural targets of the nonsense-mediated decay pathway and can cause changes in gene expression (64, 65). Changes in the frequency of AS events can be the cause of specific diseases, modify their severity, or be associated with their susceptibility (66).

Genes and pathways associated with AS event changes in MAP-infected cattle have been identified and are summarized in Table 3 . Using surgically isolated intestinal segments, significant changes in the frequency of two AS events between MAP-infected and non-infected tissues were detected 1 month post-infection (25). There were 16 genes that displayed significantly different AS events between the control and infected compartments of control and infected animals. RT-qPCR was performed to verify the differences between MAP-infected and control tissues in alternative spliced forms of the adenosine deaminase (ADA) and monocyte-to-macrophage differentiation-associated (MMD) genes. The identified AS events affected ADA and MMD mRNAs and appeared more frequently in MAP-infected intestinal segments. The event affecting ADA is a retained intron (RI) event, in which the intron 4 was retained significantly more frequently in MAP-infected segments. The event affecting MMD consisted of a mutually exclusive exon (MXE) event, in which the alternative first exon was detected more frequently in infected segments. ADA is involved in purine metabolism, and its deficiency results in the accumulation of adenosine in the lysosome, as well as lysosome enlargement, alkalinization, and dysfunction (67). MMD is expressed in mature macrophages and participates in macrophage activation through TNF-α and nitric oxide production (68). MAP infection may cause these AS events to disrupt macrophage maturation and lysosome function. The retention of intron 4 in ADA could decrease transcript stability, gene expression, and impair macrophage lysosome function. An abnormal first exon of MMD, however, could disrupt protein synthesis or function and interfere with macrophage maturation and MAP clearance during the early stage of the infection.

RNA-Seq data from Holstein cows classified as PTB-positive or negative according to ELISA and fecal PCR results allowed the identification of 119 (90 genes), 150 (89 genes), and 68 (45 genes) differential AS events between MAP-infected and healthy Holstein cows in whole-blood, jejunum, and salivary gland samples, respectively (69). Furthermore, 14 genes (CD3G, SCARB1, MYO5A, RABEPK, KIF2C, EXOC7, COPG2, REPS1, NFYC, SYK, EEA1, CARD9, DNASE1, TAC1, and RHBDD3) were significantly enriched for immune-related pathways, including vesicle-mediated transport, regulation of acute inflammatory response, and tuberculosis. Two of the identified AS events may alter protein structure, with an event in exocyst complex component 7 (EXOC7), which skipped exon 8 and appeared less frequently in MAP-infected animals, and an event in kinesin family member 2C (KIF2C), which skipped exon 10 and appeared more frequently in MAP-infected animals. EXOC7 plays an important role in vesicular trafficking and the secretory pathway, and it has been reported that a lack of EXOC7 causes a decrease in the number of phagosomes, reducing the efficiency of cytokine transport. KIF2C encodes for a protein associated with mitotic chromosome segregation and with immune cell migration and infiltration in humans. Thus, the skipping of exon 10 may affect the immune response against MAP by decreasing immune cell migration.

Our research team has recently used RNA-Seq to identify differential AS profiles between whole blood and ICV samples collected from Holstein cattle with focal and diffuse PTB-associated lesions in gut tissues vs cows without lesions in gut tissues (70). The role of AS in the regulation of the immune response was studied while keeping in mind its potential application in the diagnosis of subclinical MAP infection. To the best of our knowledge, our study is the first genome-wide profiling of AS events in whole-blood and ICV samples from Holstein cows with PTB-associated lesions of distinct severity vs. uninfected cattle without lesions in gut tissues. In the PB of cows with focal lesions vs. controls, differential AS was identified in genes involved in endocytosis (bta041144) including BOLA, LOC509006, JSP.1, Dynamin-2 (DNM2), CLTA, CLTA Adaptor complex AP2, Mu subunit (AP2M1), Cytohesin 1 (CYTH1) RAB11 Family Interacting Protein 5 (RAB11FIP5), the apoptosis regulator (PRKC1), ArfGAP With RhoGAP Domain, and Ankyrin Repeat And PH Domain 1 (ARAP1). CLTA is the main structural component of the lattice-type cytoplasmic face of coated pits and vesicles generated during receptor-mediated endocytosis. DNM2 participates in the release of CLTA-coated vesicles from the plasma membrane, and AP2M1 is required for the activity of a vacuolar ATPase, which pumps protons into endosomes and lysosomes to acidify them. CYTH1 regulates the adhesiveness of integrins at the plasma membrane and is involved in membrane trafficking, and PRKC1 negatively regulates apoptosis. After entry into macrophages by endocytosis, MAP can survive within phagosomes by inhibiting apoptosis, phagosome acidification, and antigen presentation (71). In this context, our findings suggested that the AS of genes involved in endocytosis may contribute to phagosome maturation arrest and MAP persistence within the macrophages of cows with focal lesions in gut tissues.

In the comparison of PB samples from cows with focal lesions vs. controls, AS events in genes associated with the pathogenesis of herpesviruses that infect humans at an early age but remain asymptomatic during for long time, such as the Kaposi sarcoma-associated herpesvirus infection (bta05167) and Epstein–Barr virus infection (bta05169), were identified. These genes affected by AS events include the BOLA, LOC509006, Mitogen-Activated Protein 2 Kinase 7 (MAP2K7), TNF Receptor Associated Factor 3 (TRAF3), Interferon regulatory factor 7 (IRF7), Phosphatidylinositol-4,5-Bisphosphate 3-Kinase Catalytic Subunit Delta (PIK3CD), and JSP.1. IRF7 plays an important role in the regulation of Epstein–Barr virus latency. TRAF3 can interact with Epstein–Barr virus-encoded latent infection membrane protein-1 (LMP1) along with other members of the TRAF family, acting as a negative NF-kβ regulator and inhibiting the inflammatory response (72). Our findings suggest that the AS of the BOLA, LOC509006, MAP2K7, TRAF3, IRF7, PIK3CD, and JSP.1 genes may contribute to the development of focalized PTB-associated lesions and establishment of the latent stage of MAP infection, as it has been observed in Epstein–Barr virus and Kaposi sarcoma-associated herpesvirus latency.

In the comparison of the PB samples from cows with diffuse lesions vs. controls, genes involved in the CLTA-mediated endocytosis, neutrophil degranulation, and platelet activation pathways were affected by differential AS. More specifically, AS of genes belonging to the neutrophil degranulation and platelet activation pathways, including the Cluster of differentiation 53 (CD53), CLEC12A C-type lectin domain family 12 member A (CLEC12A), Solute Carrier Family 11 Member 1 (SLC11A1), and T Cell-Interacting, Activating Receptor on Myeloid Cells 1 (TARM1) genes was observed. CD53 contributes to transduction signals in T cells and natural killer cells, and TARM1 increases the production of pro-inflammatory cytokines in macrophages and neutrophils via the toll-like receptor. Neutrophils phagocytose Mycobacteria from dying infected macrophages within the granuloma (73, 74). CLEC12A and SLC11A1 are both associated with innate host resistance. SLC11A1 is a membrane transporter involved in iron metabolism expressed in the late endosomal/lysosomal compartments of macrophages. We hypothesize that the AS of genes involved in neutrophil degranulation and platelet activation pathways may, at least in part, contribute to granuloma formation and pro-inflammatory immune response in MAP-infected cattle with diffuse lesions.

In PB samples from cows with diffuse lesions vs. focal lesions, several genes (ACTN1, TGFβ2, ITGβ2, and C8G) with AS events associated with the lysosome (bta:04142) and amoebiasis (bta:05146) were identified. Amoebiasis is caused by Entamoeba histolytica, a human extracellular protozoan parasite that, like MAP, invades the intestinal epithelium and over time disrupts the intestinal mucus layer, followed by apoptosis of host epithelial cells (75). MAP infection also impacts the intestinal mucosa and, over time, causes serious damage to the ileum and jejunum, diarrhea, progressive wasting, and the eventual death of the infected animal in the more advanced stages of clinical PTB (76). The AS of several genes associated with the activation of the lysosomes and amoebiasis may also cause disruption of the intestinal mucus layer, apoptosis, and MAP dissemination to other organs and tissues, as seen in humans infected with Entamoeba histolytica.

In the ICV samples from the infected cows, genes that encode proteins with RNA-binding domains and coiled-coil domains and those involved in AS showed changes in AS. As RNA splicing is regulated by cis-regulatory elements in pre-mRNA and trans-regulatory elements, mainly RNA-binding proteins, AS of RNA-binding proteins may in turn affect the splicing of other pre-mRNAs and contribute to the various onsets of PTB.