Proteomic profiling of milk small extracellular vesicles from bovine leukemia virus-infected cattle

Milk small extracellular vesicles (sEV) contain proteins that provide potential information of host physiology and immunology. Bovine leukemia virus (BLV) is an oncogenic virus that causes progressive B-cell lymphosarcoma in cattle. In this study, we aimed to explore the proteomic profile of milk sEV from BLV-infected cattle compared with those from uninfected cattle. Milk sEV were isolated from three BLV-infected and three uninfected cattle. Proteomic analysis was performed by using a comprehensive nanoLC-MS/MS method. Furthermore, gene ontology (GO) annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were used to evaluate the candidates for uniquely or differentially expressed proteins in milk sEV from BLV-infected cattle. Proteomic analysis revealed a total of 1330 common proteins in milk sEV among BLV-infected cattle, whereas 118 proteins were uniquely expressed compared with those from uninfected cattle. Twenty-six proteins in milk sEV were differentially expressed proteins more than two-fold significant difference (p < 0.05) in BLV-infected cattle. GO and KEGG analyses indicated that the candidates for uniquely or differentially expressed proteins in milk sEV had been involved in diverse biological activities including metabolic processes, cellular processes, respond to stimulus, binding, catalytic activities, cancer pathways, focal adhesion, and so on. Taken together, the present findings provided a novel insight into the proteomes of milk sEV from BLV-infected cattle.

. Clinical status of BLV-infected and uninfected cattle. + , positive; -, negative; ± , suspect; NT, not tested; 1 no., number; 2 Age at the time of blood sampling; 3 ELISA, anti-BLV antibody enzyme-linked immunosorbent assay; 4 copies/10 5 WBCs DNA; 5 WBCs, white blood cells; 6 Key of EC, leucosis-key of the European Community; 7  www.nature.com/scientificreports/ Proteomic profiling of milk sEV. Isolated sEV were analysed using nano LC-MS/MS method to compare the milk sEV proteins of BLV-infected cattle with that of uninfected cattle. The proteomic analysis revealed a large number of proteins in milk sEV from both BLV-infected and uninfected cattle ( Table 2). A total of 1330 proteins in milk sEV were found as common proteins among three BLV-infected cattle. Whereas, a total of 1512 proteins in milk sEV were found as common proteins among three uninfected cattle. From the result, a total of 1212 proteins in milk sEV were overlapped after comparing BLV-infected cattle with those of uninfected cattle (Fig. 2). Most significantly, our study identified a total of 118 proteins that were the candidates for uniquely expressed proteins in milk sEV from BLV-infected cattle (Supplement Table 1). The volcano plot indicated the differential encapsulation of milk sEV proteins between BLV-infected and uninfected cattle (Fig. 3A). The heat map showed that a total of 26 proteins in milk sEV showed more than two-fold up/down-regulation with a significant difference (p < 0.05) in BLV-infected cattle ( Fig. 3B and Table 3).

Figure 1.
Milk sEV characterization. The peak (mode) of particle size distribution of the milk sEV from BLVinfected and uninfected cattle were showed by NTA. Representative data from B1 and U1 cattle were shown (A). Similar spherical morphology of milk sEV between BLV-infected and uninfected cattle were observed by TEM analysis. Scale bar shows 200 nm (B). Milk sEV were successfully detected by WB analysis using antibodies against sEV-surface-marker protein CD63 and internal protein HSP70 (C).   4A-4D) to obtain a comprehensive image of the changes in proteins in milk sEV from BLV-infected cattle. The candidates for uniquely or differentially expressed proteins in milk sEV were engaged in a broad range of biological processes such as cellular process, metabolic processes, response to stimulus, signalling, biogenesis, and so on. As for the molecular function, the majority of the proteins appeared to participate in binding and catalytic activity. In the cellular component category, proteins were mainly localised in the cell, membrane, and organelles. Furthermore, the majority of the proteins were classified as cytoskeletal proteins, protein modifying enzymes, translational proteins, and metabolite interconversion enzyme. The candidates for uniquely or differentially expressed proteins in milk sEV were illustrated in the protein-protein interaction (PPI) network by STRING analysis. Importantly, the unique proteins in milk sEV have been illustrated a connectivity PPI network ( Supplementary Fig. 5A) relating to several KEGG pathways including ribosome, pathways in cancer, regulation of actin cytoskeleton, chemokine signalling pathway, focal adhesion, and so on (top 10 KEGG pathways were shown in Supplementary Fig. 5B). The candidates for differentially expressed proteins in milk sEV made a strong cluster of networking in PPI (Fig. 5A) relating to several KEGG pathways including pathways in cancer, cytokine-cytokine receptor interaction, JAK-STAT signalling pathways, proteoglycans in cancer, focal adhesion, and so on (top 10 KEGG pathways were shown in Fig. 5B).

Discussion
In this study, milk sEV were characterized by NTA, TEM, and WB analyses. The results by NTA, TEM and WB analyses were indicating the typical definition and enrichment of sEV according to the MISEV2018 guidelines 3 . Proteomic analysis revealed a total of 1330 proteins in milk sEV were common among three BLV-infected cattle, of which 118 proteins were the candidates for uniquely expressed proteins compared to those from uninfected cattle. The identified protein numbers in milk sEV from BLV-infected cattle were distinctive and relatively larger to those reported in previous studies 9, [17][18][19] . These results indicated that our current study was efficiently isolated milk sEV from BLV-infected cattle which allowed the protein count in milk sEV to increase. Moreover, Figure 3. Volcano plot demonstrated the expression of milk sEV proteins from BLV-infected and uninfected cattle. Up/down-regulated sEV proteins were indicated in orange and blue colours, respectively. Proteins not classified as up/down-regulated were plotted in grey colour. X-axis and Y-axis indicated a more than two-fold change (in log2 scale) and − log10 with significance (p < 0.05) (A). Heat map of hierarchical clustering indicated the candidates for differentially expressed proteins milk sEV from BLV-infected cattle. Up/down-regulated proteins were indicated by red/green colour, respectively (B). GraphPad Prism software, version 8 (https ://www. graph pad.com) was used to generate the volcano plot and heat map. www.nature.com/scientificreports/ nanoLC-MS/MS method had detected of low-abundance compounds in small amount of proteins resulting a high-throughput protein count in milk sEV from BLV-infected cattle 20 .
Although the mortality rate is low, the most noticeable negative outcome of BLV infection is that cattle develop to lymphosarcoma. However, it is difficult to suspect and detect which BLV-infected cattle could be developed with lymphosarcoma soon or later. Collection and analysis of milk sEV proteins hold to promise as a novel monitoring platform for BLV infection of cattle. Though sEV play a vital role in cell to cell communication 7 and influence the physiological processes 8 , their function in infectious diseases is still under progress. Huang et al. 21 reported that Staphylococcus aureus modified the protein expression in mammary tissue of cattle indicating that infectious diseases are capable of changing the normal proteomic profile of infected tissues. Recently, we demonstrated that mRNAs profile in milk sEV from BLV-infected cattle were altered compared with those of uninfected cattle 15 . However, to date, no research has been performed on identifying the changes of proteins in milk sEV from BLV-infected cattle. Therefore, the current study was conducted to evaluate the proteins encapsulation in milk sEV from BLV-infected cattle.
The present study detected a total of 26 proteins which were the candidates for differentially expressed proteins in milk sEV from BLV-infected cattle compared with those of uninfected cattle. Previous studies have been identified different types of proteins from blood or tumour tissues that were feasible in understanding the tumour progression in human. Some of the up/down-regulated proteins in our present study such as CNTFR, DDAH1, TUBB1, MARCH8, KRT1, KRT18, APOE, PDCD6, MMP2, ITGA6, MYOF, and TGFB2 were previously reported to be associated with certain tumorigenesis and progression in malignancies [22][23][24][25][26][27][28][29][30][31][32][33] . Taken together, these results indicated that as an oncogenic viral disease in cattle, BLV infection could alter the encapsulated sEV proteins; therefore, the new proteins are appearing within milk sEV. The results also suggested that these proteins could be crucial in obtaining information regarding BLV infection and its pathogenesis in cattle.
GO and KEGG analyses identified several functional terms that were enriched by the candidates for uniquely or differentially expressed proteins in milk sEV. Attention was paid to the aforementioned milk sEV proteins to unveil diverse GO terms and numerous KEGG pathways for BLV infection. In the biological process, where proteins were predominantly associated with the cellular process, metabolic process, response to stimulus, and developmental process suggesting cellular proliferation, alteration, and attachment, extensively participated during BLV infection. Binding and catalytic activity were the most prevalent molecular functions, which indicated that direct regulation of protein-protein interactions might be the key process during BLV infection and its pathogenicity in the host. The cellular component indicated that the identified proteins were enriched in the cell and membrane, thus suggesting that the contents of sEV shared close ties with their host cells. The KEGG www.nature.com/scientificreports/ analysis revealed that the candidates for uniquely or differentially expressed proteins involved in many pathways including pathways in cancers, cytokine-cytokine interaction, PI3K-Akt signalling pathway, Jak-STAT signalling pathway, and focal adhesion. Several of these KEGG pathways played an important role in tumour cell proliferation, metastasis, leukemogenesis as well as formation of solid tumours reported previously in human 34 . The results suggested that the biologically important pathways are probably involved in development and/or progression of BLV-induced leukemogenesis and tumour formation in cattle. Previous studies reported that mRNAs in milk and blood from BLV-infected cattle involved in diverse biological functions and many other pathways that are consistent with our current study 15,35 .
In conclusion, proteomic analysis identified a total of 1330 common proteins in milk sEV from BLV-infected cattle, from which 118 proteins were the candidates for uniquely expressed proteins compared with those of www.nature.com/scientificreports/ uninfected cattle. Besides, GO and KEGG analyses showed a new biological phenomenon and distinct pathways that may have a great contribution into the BLV pathogenesis. The results of this study could be solid ground to facilitate future development of milk sEV-based disease monitoring. This is the first study to present the proteomic analysis of milk sEV from BLV-infected cattle that could have wide applicability in molecular biology. In this study, since verification of the data by ELISA or other, equivalent means has not been performed, further studies are required.

Materials and methods
All experiments were performed in accordance with relevant guidelines and regulations of the Gifu University Animal Care and Use Committee (approval number 17046 and approved on 4 September, 2017). Additionally, experiments using cows were carried out in compliance with the standards of animal rights, welfare, and with minimum distress by following relevant guidelines and regulations of the Gifu University Animal Care and Use Committee.
Blood samples. Blood samples of 10 ml of each of the 16 Holstein cows were collected from two different farms and directly allocated to vacuum blood collection tubes with or without an anti-coagulant (VP-AS076K, Figure 5. PPI network and KEGG pathway analysis. The PPI network was analysed by STRING software to evaluate the candidates for differentially expressed proteins in milk sEV from BLV-infected cattle (A). In the network, nodes and edges indicated milk sEV proteins and physical and/or functional interactions of the milk sEV proteins, respectively. Empty nodes represent the proteins of unknown three-dimensional structure, and filled nodes represent the proteins with some known or predicted three-dimensional structure. Different coloured lines between the proteins represent the various types of interactions in STRING (databases, experiments, neighbourhood, gene fusion, co-occurrence, text mining, co-expression, and homology). The candidates for differentially expressed proteins in milk sEV from BLV-infected cattle were analysed using KEGG software, and the top 10 KEGG pathways were demonstrated in the bar diagram (B). www.nature.com/scientificreports/ VP-NA050K, and VP-H070K, Terumo, Tokyo, Japan). Total WBCs and lymphocyte counts were measured by an automatic cell counter Celltac α (Nihon Kohden, Tokyo, Japan). The increased lymphocyte count was checked based on the European Community's leukosis key 16 . After WBCs and lymphocyte counts, 1.3 ml of each of the anticoagulated blood samples were centrifuged at 2500 × g for 15 min at 25 °C for plasma separation by a centrifuge, MAX-307 (Tomy Seiko, Tokyo, Japan). Plasma samples were collected from the top portion of the tube and used for lactate dehydrogenase (LDH) isozymes measurement later.
Detection of antibodies. Serum was separated from blood by centrifugation at 3000 × g for 15 min at 25 °C by using a centrifuge, MAX-307. Anti-BLV antibodies in serum were measured by using an anti-BLV antibody ELISA kit (JNC, Tokyo, Japan) according to the manufacturer's instructions.
Detection of BLV. WBC was isolated from blood by hemolysis of red blood cells with 0.83% ammonium chloride followed by washing twice with phosphate buffer saline (PBS). Total amount of DNA was extracted from WBCs by using QIAamp DNA Mini Kit (51304, Qiagen, Hilden, Germany) according to the manufacturer's instructions. After measurement of DNA concentration of WBC DNA by a spectrophotometer Nan-oDropLite (Thermo Fisher Scientific, Waltham, MA, USA). Primers to amplify the envelope or pX region of BLV were used for nested PCR according to the protocols by Fechner et al. 36 and Murakami et al. 37 . PCR was carried out in a total reaction volume of 20 µl containing 0. Measurement of proviral load. BLV-infected cattle with HPL in blood were selected for this study. It was reported that BLV-infected cattle with HPL in blood were considered as cattle at high risk to be BLV spreaders and might be one of the factors of disease progression 38 . BLV proviral load was measured by using 100 ng of WBC DNA by a quantitative real-time PCR (qRT-PCR). The amplification was carried out in a reaction mixture containing 10 µl of THUNDERBIRD Probe qPCR Mix (A4250K, Toyobo, Osaka, Japan), 0.3 µl of CoCoMo-BLV Primer/Probe (A803, Riken Genesis, Tokyo, Japan), 5 µl of a template DNA sample, and PCR grade water to increase the volume up to 20 µl. For the proviral quantification, BLV BoLA-DRA gene Plasmid DNA was used from the kit (A804, Riken Genesis) and BLV proviral DNA was measured by a Thermal Cycler Dice Real Time System III (TP970, Takara Bio) according to the manufacturer's instructions. After the measurement, BLV proviral copies of > 30,000 in 10 5 WBCs DNA was considered as HPL in BLV-infected cattle (Table 1). Hematological test, detection of serum antibodies against BLV, detection of BLV provirus, and measurement of BLV proviral load were conducted by the Gifu Central Livestock Hygiene Service Center (Gifu, Japan).

Measurement of lactate dehydrogenase (LDH) isozymes. A previous study reported that LDH
activities in the serum, mainly, increased LDH 2 and 3 isozyme percentages, reflected progression of EBL, thereby making it a key parameter for the diagnosis of lymphosarcoma 39 . Therefore, we also focused on serum LDH isozyme activity. BLV-infected cattle with LDH 2 + 3 > 30% was selected. LDH isozymes were measured by a Hydrasys 2 Scan (Sebia, Paris, France) using HYDRAGEL 7 ISO-LDH (Sebia), which was conducted by a clinical laboratory testing company, Fujifilm Vet Systems (Tokyo, Japan).
Collection of milk samples. Milk samples were collected from three BLV-infected and three uninfected healthy cattle. After collection, both milk samples were transported quickly to the laboratory in a cool box to maintain the temperature and were stored at 4 °C for further use.

Isolation of sEV.
For the isolation of milk sEV from BLV-infected and uninfected cattle, we followed the procedure previously described by Yamauchi et al. 40 and Rahman et al. 41 with slight modifications. Importantly, after defatting of milk, milk sEV were purified using acetic acid followed by sequential filtration through 1.0, 0.45, and 0.2 μm filters (GA-100, C045A047A, and C020A047A, Advantec, Tokyo, Japan). Subsequently, milk sEV were concentrated by ultracentrifugation (UC) at 100,000 × g at 4 °C for 1 h using a P42A angle rotor (Hitachi Koki, Tokyo, Japan) in a Himac CP80WX ultracentrifuge (Hitachi Koki). After the first UC, the supernatant was discarded and the pellet was resuspended with PBS up to 10 ml into a 13PET tube (Hitachi Koki) followed by another UC at 100,000 × g at 4 °C for 1 h using a P40ST swing rotor (Hitachi Koki). sEV pellet was collected and resuspended with 100 µl of PBS for further use.
Isolation of RNA and detection of BLV genomic RNA and BLV proteins in milk sEV. Total RNA was extracted from milk sEV from BLV-infected cattle by using QIAamp viral RNA Mini Kit (52906, Qiagen, Hilden, Germany) according to the manufacturer's instructions. The RNA concentration was measured by a spectrophotometer NanoDropLite (Thermo Fisher Scientific). cDNA was synthesis by reverse transcription reaction using 5 × PrimeScript RT Master Mix (RR036A-1, Takara Bio) followed by nested PCR was performed to amplify the envelope or pX region of BLV 36,37 . WB analysis was performed as described in a previous study 41 . www.nature.com/scientificreports/ After gel electrophoresis and trans-blotting, membranes were blocked with 5% non-fat skim milk in Tris-buffered saline [0.1 M Tris-HCl (pH 8.0) and 0.03 M NaCl] containing 0.1% Tween-20 (TBST) at room temperature (RT) for 30 min. For the detection of BLV proteins gp51 in milk sEV from BLV-infected cattle, monoclonal antibody specific to a D-D' epitope on BLV gp51 (Env) (1:400, VMRD, Pullman, WA, USA) was used. The membrane was incubated for 1 h at RT, diluted in 1% non-fat skim milk in TBST, followed by washing thrice with TBST. The secondary antibody, anti-mouse IgG sheep antibody (1:1000, NA9310, GE Healthcare, Little Chalfont, UK) conjugated with horseradish peroxidase, were diluted with TBST incubated for 1 h at RT followed by washing thrice with TBST. Peroxidase activity was detected using a Pierce ECL Plus substrate (Thermo Fisher Scientific) and visualised by using a chemiluminesence apparatus (ChemiDoc XRS + , Bio-Rad Laboratories).
Characterization of milk sEV. NTA analysis of milk sEV from BLV-infected and uninfected cattle was performed using a NanoSight LM10V-HS, NTA 3.4 instrument by an assigning company (Quantum Design Japan, Tokyo, Japan). Morphological examination of isolated milk sEV from BLV-infected and uninfected cattle was carried out by TEM as described in a previous study 41 with slight modifications. The sEV pellet was diluted 10 times from its original concentration with distilled water, applied to glow-discharged carbon support films on copper grids followed by stained with 2% uranyl acetate. The samples were then visualized under an electron microscope, JEM-2100F (JEOL, Tokyo, Japan) at 200 kV. WB analysis was carried out as described in a previous study 41 44 were performed to assess the statistical significance of the data. Corrected p-value cut off of 0.05 was applied. From our data sets, common proteins present in all three samples from BLV-infected and those from three uninfected cattle were considered for listing up and the proteins without a gene name or those with a dual gene name were excluded.
Functional and STRING interaction network analysis. GO and KEGG pathway 45 analyses were performed using the Database for Annotation, Visualization and Integrated Discovery (DAVID version 6.8) (https ://david .ncifc rf.gov/) 46 . The biological process, molecular function, cellular component, and protein class were determined using Panther software, online version (http://www.panth erdb.org/). An adjusted p-value of less than 0.05 was considered statistically significant. In this study, we used the Search Tool for the Retrieval of Interacting Genes (STRING) database, version 10.5 (https ://strin g-db.org) to construct the interaction network of the candidates for uniquely or differentially expressed proteins in milk EVs from BLV-infected cattle.
Statistical analysis. Statistical analysis was performed using the Student's t-test to compare the milk sEV proteins from BLV-infected and uninfected cattle. The p value 0.05 was considered as statistically significant.

Data availability
Raw data of proteomic analysis of milk sEV from BLV-infected and uninfected cattle was deposited in the Mendeley data repository (