Fifteen specimens were obtained from patients with psoriasis vulgaris (ten males and five females, aged 25–59 years). They had received no systemic treatments, phototherapy or externally used drugs for at least 1 month before skin biopsies and blood specimens collected. Ten control skin biopsies were obtained from healthy volunteers (five males and five females, aged 22–57 years). We obtained biopsies (1 × 0.5 cm) by surgical operations and whole blood (5–10 ml). Psoriasis Area Severity Index (PASI) was determined at the condition when samples were collected by at least two dermatologists. The study was approved by the ethics committee of Shandong University, China, and all patients provided written informed consent.

Whole blood (5–10 ml) from patients and healthy controls were centrifuged at 3,000 rpm for 10 min to separate serum. The serum was collected, centrifuged at 300 g for 10 min, then at 16,500 g for 30 min, and later filtered through a 0.22 mm pore size filter to remove cell debris. The final supernatant was ultracentrifuged at 100,000 g for 70 min and recentrifuged at the same speed. The purified EVs were resuspended in PBS.

Morphology of serum EVs was observed by transmission electron microscopy (TEM). The numbers and diameters were detected using nanoparticle tracking analysis (NTA, Zeta-View PMX 110; Particle Metrix, Meerbusch, Germany). Phenotypic was assessed by western blot using antibodies (Abcam) specific for CD9 (1:2,000, ab92726), CD63 (1:1,000, ab134045), and CD81 (1:1,000, ab109201).

Serum EVs miRNAs and mRNAs in skin lesions were detected using Agilent Human miRNA microarray V 21.0 (Sinotech Genomics, Shenzhen, China) and SBC Human (4 × 180K) competing endogenous (ce)RNA microarray (Shanghai Biotechnology Corporation, Shanghai, China), respectively. The fold change >2 in these analyses was an absolute value of fold change.

A diluted double CY3-labelled miR-1305 probe (8 ng/μl, 5′-CY3-AAAAGTTGAGATTACCCTCTCT-CY3-3’; Servicebio, Wuhan, China) and miR-6785-5p probe (8 ng/μl, 5′-CY3- ACCCTCCCGCACCTACTACCAC-CY3-3’; Servicebio, Wuhan, China) were used for hybridization. 4′,6-Diamidino-2-phenylindole (Servicebio, Wuhan, China) was used for nuclear staining.

Serum EVs were labelled with PKH26 Red Fluorescent Cell Linker Mini kit (MINI26-1 KT; Sigma-Aldrich, St. Louis, MO) according to the instructions. Then, the suspension was centrifuged at 100,000 g for 70 min, and the supernatant was discarded. The labelled EVs were resuspended in PBS and centrifuged once again at the same speed. Next, the EVs were co-cultured with normal human epidermal keratinocytes (NHEKs) cells for 12 h. Then washed the cells 3 times with PBS and fixed with 4% formaldehyde. Subsequently, the cells were stained using phalloidin-iFluor 488 reagent, and DAPI was added to stain the nuclei. The cells were then observed under a confocal microscope (LSM800; Zeiss).

Serum EVs were digested with RNaseA (100 ng/ml, Qiagen, Australia). Then, total RNA from serum EVs was extracted using TRIzol reagent (Invitrogen). The expression levels of serum EVs miRNAs and miRNAs in skin lesions were detected according to the manual of the All-in-One miRNA qRT-PCR Detection System (GeneCopoeia, Guangzhou, China). U6 was used as the internal control for miRNA. Total RNAs from skin samples were extracted using an RNeasy mini kit (cat. no. 74106; Qiagen GmBH, Hilden, Germany) according to the manufacturer’s protocols. Then the expression levels of mRNAs were detected using microarray analysis.

Western blot was performed as described in our previous experiments (Jiang et al., 2017; Zhao et al., 2017). The following primary antibodies were used in this study: rabbit monoclonal antibodies (procured from Abcam) specific for CD9 (1:2,000, ab92726), CD63 (1:1,000, ab134045) and CD81 (1:1,000, ab109201) and GAPDH (1:1,000, ab8245).

Target genes of miRNAs were predicted using TargetScanHuman 7.1 and miRDB-MicroRNA Target Prediction Database. Then, the genes were intersected with mRNAs detected by mRNA microarray. Enrichment calculations were performed using Fisher’s exact test. Further, we conducted GO and pathway enrichment analysis of the target genes. The specific principle was to carry out annotation mapping of differentially expressed genes in GO and KEGG database entries, calculate the number of the target genes in each GO and pathway entry, and then use hypergeometric test for statistics. Select the GO and KEGG entries that were significantly enriched in the differentially expressed genes. After the calculated p-value was corrected by multiple hypothesis tests, the p-value 0.05 was taken as the threshold, and the GO and KEGG term meeting this condition was defined as the GO and KEGG term significantly enriched in the target genes. MiRNA-mRNA network analysis was visualized using Cytoscape 3.9.0 software.

The proliferation rates of NHEK Cells were measured at different time intervals using CCK-8 assays (Beyotime, Shanghai, China) after co-cultured with serum EVs obtained from psoriasis patients (n = 15) and healthy controls (n = 10) for 12 h. Optical density (OD) values at 450 nm were measured using a Synergy H1 Microplate Reader (BioTek, United States).

After 24 h co-cultured with serum EVs, the culture supernatants were collected. Then, the expression levels of CCL20 and IL-8 secreted from NHEK cells were measured using specific ELISA kits (Elabscience, China). Absorbance at 450 nm was measured using an ELISA plate reader (Bio-Rad).

All data are presented as mean ± standard deviation. All experiments were performed at least 3 times. The differences between 2 groups were analyzed using Student’s t-test with SPSS 17_0 analytical software (IBM, Armonk, NY, United States) and GraphPad Prism 7.00 (GraphPad Software, La Jolla, CA, United States). The target genes in GO and pathway were using hypergeometric test for statistics and p-value was corrected by multiple hypothesis tests. p-values < 0.05 were considered significant.

To investigate the characterization of serum EVs in psoriasis, we isolated serum EVs from psoriasis patients and healthy controls using differential ultracentrifugation. Using the methods of transmission electron microscopy (TEM) and nanoparticle tracking analysis (NTA), we confirmed that purified serum EVs obtained from psoriasis patients and healthy controls both exhibited a typical cup-shaped structure (Figure 1A), and their average diameters were around 100 nm (Figure 1B). In addition, as shown in Figure 1C, the expression of CD9, CD63, and CD81 was identified in serum EVs by western blot analysis.

To determine miRNA profiling of serum EVs in psoriasis, the expression of EVs miRNAs was detected using miRNA microarray analysis (n = 4). The results showed remarkable differences in the serum EVs miRNA profiles, with 16 differentially expressed miRNAs in serum EVs from psoriasis compared with healthy controls (GSE200637). Among 16 differentially expressed miRNAs, 2 were upregulated and 14 were down-regulated (fold change>2, p < 0.05, Figures 2A,B), (Supplementary Table S1).

To further investigate the potential role of serum EVs miRNAs in the activation of keratinocytes, we performed mRNA microarray analysis in psoriatic skin lesions and healthy controls (n = 3). According to the mRNA microarray results, a total of 1,725 dysregulated mRNAs were identified. Among 1,725 dysregulated mRNAs, 1,157 mRNAs were upregulated and 568 mRNAs were downregulated (fold change>2, p < 0.05, Figures 2C,D), (Supplementary Table S2). The mRNA microarray results suggested that there were a greater number of upregulated mRNAs in psoriatic skin lesions compared with healthy controls, which corresponded to the majority of miRNAs that were downregulated in serum EVs.

To validate the miRNA microarray results, we performed qRT-PCR (n = 10) and Fluorescence in situ hybridization (n = 10) to detect the expression of miR-1305 and miR-6785-5p which predicted closely related with inflammation responses. We digested EVs with RNaseA and reperformed qRT-PCR, the results indicated that miR-1305 and miR-6785-5p are present inside of EVs not on the surface of EVs (Figure 3A). The results of Figures 3B,C showed that the expression of miR-1305 was down-regulated in both serum EVs and skin lesions (fold change = 0.2 in EVs, p < 0.01; fold change = 0.17 in skin lesions, p < 0.001) and miR-6785-5p was upregulated (fold change = 2.24 in EVs, p < 0.001; fold change = 2.76 in skin lesions, p < 0.001). As shown in Figures 3D,E, the results of fluorescence in situ hybridization (FISH) were consistent with qRT-PCR and showed that miR-1305 and miR-6785-5p were differentially expressed in the epidermis.

To explore the potential functions of differentially expressed miRNAs in serum EVs, we predicted the target genes of the validated miR-1305 and miR-6785-5p using bioinformatics analysis (TargetScanHuman 7.1 and miRDB-MicroRNA Target Prediction Database). As miR-1305 was downregulated and miR-6785-5p was upregulated in serum EVs and psoriatic skin lesions, the predicted target genes of miR-1305 and miR-6785-5p were intersected with upregulated and downregulated mRNAs which were detected by mRNA microarray (GSE181318), respectively. The results showed that 77 upregulated mRNAs might be regulated by miR-1305 and 83 downregulated mRNAs might be regulated by miR-6785-5p (Figures 4A,B); Supplementary Tables S3, S4). In addition, we mapped the miRNA-mRNA networks of miR-1305 and miR-6785-5p to study their regulatory mechanisms with the 77 upregulated mRNAs and 83 downregulated mRNAs, respectively (Figures 4C,D); Supplementary Tables S5,S6).

To clarify the potential biological functions of the 77 upregulated mRNAs and 83 downregulated mRNAs and the predicted involved signalling pathways, we annotated each gene of the intersection based on the Gene Ontology and KEGG database. Then, GO analysis and KEGG pathway analysis were performed to predict the function of miR-1305 and miR-6785-5p, respectively. The results showed that both miR-1305 and miR-6785-5p were involved in regulating the inflammation response and the signalling pathway related to psoriasis, including positive regulating of type 1 interferon production, autophagy, cytokine-cytokine receptor interaction, Th17 cell differentiation, IL-17 receptor activity, Wnt signalling pathway, MAPK signalling pathway and so on (Supplementary Figures S1A,D). These results indicated that miR-1305 and miR-6785-5p might be involved in the pathogenesis of psoriasis.

To investigate whether serum EVs were involved in the activation of keratinocytes, we performed a serum EVs-keratinocytes interaction experiment. Purified serum EVs were isolated and labelled with PKH26 Red Fluorescent. Next, the labelled EVs were co-cultured with NHEK cells for 12 h. Then, the cells were fixed with 4% formaldehyde and stained with phalloidin-iFluor 488 reagent and DAPI. Using a confocal microscope, we verified that serum EVs could be taken up by NHEK cells (Figure 5A).

Next, we detected the expression of miR-1305 and miR-6785-5p in NHEK cells after co-cultured with serum EVs obtained from psoriasis patients and healthy controls. The results showed that miR-1305 was downregulated and miR-6785-5p was upregulated after co-cultured with serum EVs obtained from psoriasis compared with controls (Figures 5B,C). As the intrinsic cell targeting properties of EVs, our findings assumed that serum EVs might not only directly communicate with neighbouring cells, such as T cells or B cells, but also communicate with the activation of keratinocytes in psoriasis via the loaded miRNAs, including miR-1305 and miR-6785-5p.

Then, we explored the role of serum EVs in keratinocyte proliferation and inflammatory response. CCK-8 assay showed that serum EVs obtained from psoriasis significantly promoted the proliferation of keratinocytes compared with that of control groups (Figure 5D). Similarly, the results of ELISA analysis showed the secretion of CCL20 and IL-8 was significantly increased compared with that of control groups (Figures 5E,F). These findings suggested that serum EVs might be involved in the activation of keratinocytes in psoriasis.

Serum EVs miR-1305 and miR-6785-5p are related to the disease severity and might be biomarkers of psoriasis.

To further confirm the relationship of serum EVs miR-1305 and miR-6785-5p with the severity of psoriasis, we measured the correlation of miR-1305 and miR-6785-5p in serum EVs with PASI scores. As Figures 5G,H showed, the increase of miR-6785-5p was positively correlated with the PASI score. In contrast, the expression of miR-1305 in serum EVs was negatively correlated with disease severity. These results indicated that serum EVs miR-1305 and miR-6785-5p might serve as biomarkers of psoriasis.