Human umbilical cord tissue-derived MSCs (huc-MSCs) were isolated from discarded umbilical cords and prepared following an established protocol. All procedures were conducted under aseptic, standardized conditions. Fresh umbilical cord specimens were obtained from healthy term neonates and stored in PBS at 4°C. In the ultra-clean table, the cords were cut into 3 to 4 cm pieces and were rinsed with 5% penicillin–streptomycin PBS buffer to remove the cord blood and other impurities. The outer membrane and the three blood vessels in cords were separated in turn, and Wharton jelly was obtained from the viscous substance. Wharton jelly was cut into 1-mm³ pieces and was dispersedly transferred into a Petri dish with DMEM-F12 containing 10% fetal bovine serum, 1% penicillin–streptomycin, and 10 ng/ml bFGF at 37°C in a cell incubator with 5% CO₂. In the primary culture, half of the medium was changed 4 days later. Then, the culture medium was refreshed every 3 days. Generally, fibroblast-like cells can be observed to crawl out after 14 days. When they reached 70–80% confluence, the cells were subcultured to a new Petri dish for further expansion.

DMEM-F12 supplement with 10% fetal bovine serum, 5% penicillin–streptomycin, and 10 ng/ml bFGF was used to culture huc-MSC. To obtain more and purer sEVs, huc-MSC with passages two to three was transferred to the DMEM-F12 medium supplemented with 0.5% fetal bovine serum and 10 ng/ml IFN-γ. After 48 h, the culture medium was collected and then centrifuged at 1,000×g for 10 min, 4,000×g for 20 min, and 10,000×g for 40 min successively to remove cells, cells debris, and microvesicles. Subsequently, huc-MSC-sEVs were pelleted by ultracentrifugation at 100,000×g for 70 min and then resuspended in PBS. Moreover, all centrifugation processes were performed at 4°C.

The size distribution and zeta potential of sEVs were measured using a NanoBrook 90 Plus PALS (Brookhaven) instrument. The morphology of sEVs was detected using a transmission electron microscope (TEM).

Cell and sEV lysates were separated using the SDS-PAGE and were then transferred to polyvinylidene fluoride membranes (Millipore, Darmstadt, Germany). Then the membrane was sealed with 5% non-fat milk for 1.5 h at room temperature and incubated with the desired primary antibodies (CD63, CD9, Alix, TSG101, STAT6, LYN, and β-actin) overnight at 4°C. Post-incubation with HRP-conjugated secondary antibodies was performed for 1.5 h at room temperature, and then detection was carried out using an enhanced chemiluminescence reagent (ECL) (Protein Tech, China).

Total RNAs from cells, skin tissue, and spleen tissues were extracted using the TRIzol reagent according to the manufacturer’s protocol. The skin tissue and spleen were added to the TRIzol reagent and grounded. The RNA concentration was measured using a NanoDrop One (Thermo Fisher Scientific). Then, according to the manufacturer’s manual, the RNAs were reversely transcribed into cDNA using a QuantiTect RT Kit (Qiagen, Valencia, CA, United States), according to the manufacturer’s protocols. All the primers used in this study are listed in Table 1.

All the antisense miRNA-210 oligonucleotides were designed and purchased from GenePharma. The sequence of ASO-210 is 3′ GAC​ACG​CAC​ACU​GUC​GCC​GAC​U 5’. To encapsulate ASO-210 into sEVs, 1 mg (protein weight) purified IFNγ-sEVs and 100 nmol ASO-210 were gently mixed in 2 ml electroporation buffer (1.15 mM potassium phosphate pH 7.2, 25 mM potassium chloride) at 4°C. The mixtures were subjected to electroporation at 400V and 150 μF in to 0.4-cm electroporation cuvettes using a MicroPulser Electro-operator (Bio-Rad, United States). After electroporation, the mixtures were incubated at room temperature for 30 min for the membrane recovery. Unloaded ASO-210 was removed by ultracentrifugation at 100,000×g for 70 min. Then supernatants were removed, and the pellets were resuspended in 2 ml PBS. At last, the sample was diluted 50 times with PBS for test concentration. FAM-labeled ASO-210 was electroporated with sEVs as described before. The amount of FAM-ASO-210 was quantified using a fluorometer with excitation at 490 nm and emission at 520 nm. The loading efficiency was calculated as follows: Loading efficiency (%) = (Amount of AMO-210 after electroporation)/(Total amount of ASO-210 before electroporation) ×100.

Jurkat cells and HaCaT cells were, respectively, seeded in a confocal plate at a density of 10⁶ cells and cultured for 24 h. Jurkat T cells and HaCaT cell membranes were pre-stained with WGA 350 for 10 min at 37°C. The sEVs and ASO-210@sEVs were stained with DiD for 10 min at 37°C. Then, the DiD-labeled sEVs were added to both Jurkat cells and HaCaT cells for 30–60 min. The cells were observed using a confocal microscope (Zeiss, LSM880).

PBMCs were harvested from healthy donor’s blood and isolated by Ficoll gradients. Fresh anticoagulated whole blood was added onto the half volume of Ficoll-Paque PLUS (Sigma-Aldrich) in a centrifuge tube, and centrifuged at 1500 rpm for 40 min. The PBMC interface layer was gently collected and washed twice with filter-sterilized PBS by centrifugation at 1000 rpm for 10 min. Next, CD3⁺ T cells were separated from freshly PBMCs by using a MojoSort Human CD3⁺ T-cell Isolation Kit (Biolegend, United States), cultured in CD3-coated plates (clone OKT3; Biolegend), and then cultured in medium (RPMI1640 with 10% FBS, penicillin/streptomycin, and 2 ng/ml IL-2).

In total, 6- to 8-week-old male BALB/c mice were purchased from the Laboratory Animal Center of Sun Yat-Sen University and kept in a pathogen-free facility. To induce psoriasis, the mice were treated with a daily topical dose of 62.5 mg of imiquimod (IMQ) cream on the shaved back for 7 consecutive days.

DiD-labeled native sEVs (150 μg) and IFNγ-sEVs (150 μg) were intradermally injected into the BALB/c mice. After 12 h, accumulation of sEVs in various organs was observed using a Night OWL imaging system (LB983).

The severity of back skin inflammation in the IMQ-induced psoriasis-like mouse model was scored by an objective scoring system, which was based on clinical PASI for psoriasis patients. Thickening, scaling, and erythema were scored independently on a scale from 0 to 4: 0, none; 1, slight; 2, moderate; 3, marked; and 4, very marked. The total score was obtained by accumulating the three index scores (scores 0–12).

Skin tissue samples from the psoriasis site were collected and fixed with 3 ml 4% paraformaldehyde for 24 h before transferring into 70% ethanol. Then, the samples were embedded with paraffin and cut into 4-μm-thick sections. The sections were stained with hematoxylin and eosin (H&E) following the standard procedure. The inflammatory cell infiltration condition was observed using a light microscope with ×100 magnification. Immunohistochemistry was performed to confirm the counts of Ki67⁺, CD3⁺, IL-6⁺, and TNF-α⁺ cells in psoriasis skin lesions. In brief, all formalin-fixed paraffin-embedded tissue samples were subjected to deparaffinization, rehydration, and heat-induced epitope retrieval and were subsequently incubated with Ki67⁺, CD3⁺, IL-6⁺, and TNF-α⁺ antibodies (1:50 in 1.5% BSA) overnight. Afterward, secondary antibodies were incubated at room temperature in the dark for 1 h. After washing with PBS, Ki67⁺, CD3⁺, IL-6⁺, and TNF-α⁺ cells were observed by microscopy.

Spleen and lymph nodes collected from mice were immersed in PBS in 1.5-ml centrifuge tubes and kept in ice. For cell isolation, the spleen and lymph nodes were repeatedly grounded, washed with PBS, and filtered through a 100-μm filter. Cells were collected by centrifuging at 1,500 rpm for 5 min, before resuspending in cell staining buffer (Biolegend, United States), and were incubated with the FITC anti-mouse CD3 antibody (Biolegend, catalog 317306) and APC anti-mouse CD4 antibody on ice for 15 min in the dark. For intracellular staining, the cells were fixed in fixation buffer (Biolegend, United States) in the dark for 20 min at room temperature, permeabilized, and stained in the dark for 15 min at room temperature with Brilliant Violet 510™ anti-mouse IL-17A (Biolegend, catalog 504127), PerCP-Cy5.5 anti-mouse IFN-γ (Biolegend, catalog 505822), and Brilliant Violet 421™ anti-mouse IL-4 (Biolegend, catalog 504127). The cells were washed with PBS, centrifuged, and filtered through a 100-μm filter before analysis.

Data analysis was performed by GraphPad Prism version 7.0 software. Data are presented as mean ± SEM. We assessed data for normal distribution and similar variance between groups. Statistical significance (*p < 0.05, **p < 0.01, and ***p < 0.001) was assessed using 2-tailed unpaired Student’s t-test for comparisons between two groups and one-way analysis of variance (ANOVA) with relevant post hoc tests for multiple comparisons. When the data were not normally distributed or displayed unequal variances between two groups, we used the 2-tailed Mann–Whitney U test for statistical analysis.

In order to test the potential of MSC-sEVs as the potential immunosuppressant drug to prevent the pathology of psoriasis, MSCs were isolated from fresh human umbilical cords (huc-MSCs) and expanded for three or four generations in vitro for sEV purification (Figure 1A). Flow cytometry analysis was used to confirm that the isolated huc-MSCs were positive for mesenchymal stem cell markers CD90, CD105, and negative for hematopoietic stem cell markers CD34, CD9, or CD45 (Figure 1B). IFN-γ, a potent multifunctional cytokine which is secreted primarily by activated NK cells and T cells, has recently been reported to promote immunosuppressive effects of MSCs (Riazifar et al., 2019). We found that the mRNA expression of a panel of immunosuppressive markers including IDO1, PGE2, and IL10 was significantly increased in huc-MSCs with IFN-γ stimulation (Figure 1C ).

Next, we prepared and purified sEVs derived from native and IFN-γ-stimulated huc-MSCs by differential centrifugation. Dynamic light scattering (DLS) analysis showed that the particle size distribution is Gaussian, with a peak value of 91.86 and 97.06 nm, and the zeta potentials were -22.33 and -21.57 mV, respectively, indicating that the extracellular vesicles are well dispersed and stable (Figures 1D,E). When visualized under a TEM, the particles were spherically shaped with a typical bilayer membrane structure within the size range of sEVs in diameter (Figure 1F). Western blotting confirmed the presence of sEV markers, namely, CD63, CD9, TSG101, and Alix in isolated vesicles (Figure 1G).

ASO-210 has been reported to alleviate psoriasis by blocking the immune imbalance and inhibiting the development of psoriasis-like inflammation (Wu et al., 2018). However, instability and delivery inefficiency to target cells of ASO-210 hinder its clinical application. Therefore, we speculate that sEVs from primary huc-MSCs may be used as a targeted drug delivery vehicle to improve the therapeutic effects of ASO-210. In order to test the potential of IFNγ-sEVs as a synergistic, targeted drug delivery system of ASO-210 to effector cells, FAM-labeled ASO-210 was loaded into the MSC-derived IFNγ-sEVs by electroporation (ASO-210@IFNγ-sEVs), with an average loading efficiency of 13.35% and a release efficiency of 94.02% (Figure 1I). Similar to native and IFN-γ-stimulated sEVs, the average size of ASO-210@sEVs was 115 nm, and the mean zeta potential was -22.48 mV (Figures 1D,E), with membrane-bounded round shape within the size range of sEVs (Figure 1F). In addition, we also used confocal analysis to ensure that FAM-labeled ASO-210 is maintained in DiD-labeled sEVs, indicating that the ASO-210 was entrapped into sEVs (Figure 1H). To sum up, all the aforementioned experiments demonstrated the success of obtaining sEVs from primary huc-MSCs, which could carry ASO-210 for the following experiments.

Next, we determined whether IFNγ-sEVs can be used as a targeted drug delivery system to carry ASO-210 to effector T cells to enhance immune suppression. First, we examined the cellular uptake of IFNγ-sEVs and ASO210@IFNγ-sEVs in HaCaT keratinocytes and Jurkat T cells, two key target cell types of ASO-210 in psoriasis. The confocal images suggested that both IFNγ-sEVs and FAM-ASO210@IFNγ-sEVs could be used by both cell lines upon co-culture for 0.5 h (Figure 2A). A previous study has reported that ASO-210 can target miR-210 and thus activate the expression of STAT6 and LYN, two direct target genes of miR-210 (Wu et al., 2018). To investigate whether ASO210@IFNγ-sEVs can deliver ASO-210 into the cells to target the miR-210 pathway, Jurkat T cells were incubated with ASO-210@IFNγ-sEVs (1 nM ASO-210) for 0h, 24 h, 48 h, and 72 h. As expected, miR-210 expression in Jurkat T cells was dramatically reduced (Figure 2B), whereas both mRNA and protein levels of STAT6 and LYN increased significantly as the incubation time extended (Figures 2B,C). To investigate the biological role of IFNγ-sEVs and ASO-210 on the inhibition of T-cell activation, CFSE-labeled activated healthy human PBMCs or CD3⁺ T cells were treated with IFNγ-sEVs, ASO-210, ASO-210&IFNγ-sEVs, and ASO-210@IFNγ-sEVs for 5 days, respectively. The following flow cytometry analysis showed that IFNγ-sEVs, ASO-210, ASO-210&IFNγ-sEVs, and ASO-210@IFNγ-sEVs significantly inhibited the PBMCs proliferation by 28.28%, 18.58%, 29.98%, and 34.71%, respectively, suggesting that both IFNγ-sEV and ASO-210 presented immunosuppressive effects on activated PBMCs (Figure 2D). In addition, compared to ASO-210 mixed with IFNγ-sEVs (ASO-210&IFNγ-sEVs), ASO-210 loaded into IFNγ-sEVs (ASO-210@IFNγ-sEVs) greatly enhanced its immune inhibition function (Figures 2D,F). Furthermore, IFNγ-sEVs, ASO-210, ASO-210&IFNγ-sEVs, and ASO-210@IFNγ-sEVs have similar inhibition on proliferation of CD3⁺ T cells with PBMCs by 20.58%, 18.68%, 28.45%, and 33.58%, respectively (Figures 2E,G). This was supported by the corresponding quantitative analysis of CFSE results (Figures 2F,G).

We next sought to investigate the therapeutic effect of IFNγ-sEVs and ASO-210@IFNγ-sEVs in treating psoriasis in a well-established IMQ-induced model, which closely resembles the human disease phenotype according to previously published studies. To establish IMQ-induced psoriasis-like mice, 6-week-old male BALB/c mice were treated with a daily topical dose of 62.5 mg of IMQ cream on the shaved back for six consecutive days. DiD-labeled native sEVs and IFNγ-sEVs were intradermally injected into the mice, and bioluminescent images of IFN-γ stimulated MSC-sEVs taken in vivo confirmed a marked accumulation of IFNγ-sEVs on the injured skin of IMQ-induced psoriasis-like mice, as compared to unstimulated native sEVs (Figure 3A). IMQ-induced psoriasis-like mice were injected with phosphate-buffered saline (PBS), IFNγ-sEVs (25 mg/kg), ASO-210 (8 nmol), ASO-210&IFNγ-sEVs (8 nmol +25 mg/kg), or ASO-210@IFNγ-sEVs (8 nmol +25 mg/kg), intradermally 4 times from day 2 to day 5 to test the therapeutic effect (Figure 3B). Halometasone (Hal) was topically applied daily as the positive control (Figure 3B). Simultaneously, toxicological experiments were also performed as biological safety is essential for the clinical application of sEV-loaded drugs. The intradermal administration of all groups of drugs did not cause obvious histopathological damage to the main organs, including the heart, liver, spleen, lung, and kidney in the mice (Figure 3C), demonstrating that IFNγ-sEVs do not cause systemic adverse effects when used for psoriasis therapy. As expected, the PBS control mice developed typical psoriasis-like lesions with evident clinical and pathological changes, together with obvious weight loss during the experiment period (Figures 3D–G). Consistent with the previous report, intradermal administration of ASO-210 ameliorated the pathological phenotype of IMQ-induced psoriasis-like mice (Figures 3D–G). Remarkably, IFNγ-sEVs alleviated the weight loss and disease severity including thickness, erythema, and scales of skin lesions following treatment (Figures 3D,E). In addition, the representative skin photographs and H&E staining also exhibited significant improvement in both clinical manifestations and pathological characteristics on day 7, following the treatment of IFNγ-sEVs (Figures 3F,G). Interestingly, IFNγ-sEVs also reduced the proliferation of keratinocytes as reflected by the decreased levels of Ki67 (Figures 3F,H). Consistent with in vitro results, IFNγ-sEVs could cooperate with ASO-210 (ASO-210&IFNγ-sEVs and ASO-210@IFNγ-sEVs) to achieve the inhibitory effect on disease development (Figures 3E–G). Interestingly, ASO-210@IFNγ-sEVs showed the better therapeutic effect than ASO-210&IFNγ-sEVs, indicating that loading into sEVs might improve the therapeutic effect of ASO-210 in vivo (Figures 3E–G). These results conclude that the IFNγ-sEVs are effective natural nanovesicles carrying antisense oligonucleotides to protect against psoriasis.

In order to further explore the mechanism for the therapeutic effect of IFNγ-sEVs, we carried out the immunohistochemistry analysis of CD3, IL-6, and TNF-α. The results showed that compared with the PBS group, the release of IL-6, TNF-α, and the infiltration of CD3 positive T cells was significantly reduced in all treatment groups, especially in the ASO-210@IFNγ-sEVs group (Figure 4A). Then splenic CD4⁺ T cells were obtained to assess the in vivo effect of IFNγ-sEVs and ASO-210 on the immune imbalance of Th cell subsets in a psoriasis-like IMQ mouse model. Notably, both IFNγ-sEVs and ASO-210 increased the percentage of CD4⁺IL-4⁺ T cells (Th2) and decreased the percentage of CD4⁺IL-17A⁺ T cells (Th17) in the spleen of IMQ-treated mice (Figures 4B,C). In addition, IMQ-treated mice injected with IFNγ-sEVs or ASO-210 displayed a prominent reduction of Il17a, Ifng, Il6, and Tnfa mRNA in both skin lesions and splenic CD4⁺ T cells (Figures 4D,E). Compared to IFNγ-sEVs, ASO-210, or ASO-210&IFNγ-sEVs, ASO-210@IFNγ-sEVs had more significant effects in preventing the immunopathological changes, which was consistent with milder disease severity in the ASO-210@IFNγ-sEVs group (Figures 4B–E). Together, these results indicated that IFNγ-sEVs exhibited a synergistic immunosuppressive function with ASO-210 in IMQ-induced psoriasis-like mice, while ASO-210@IFNγ-sEVs presented stronger inhibition (Figure 5).