Mesenchymal stromal cells-derived small extracellular vesicles protect against UV-induced photoaging via regulating pregnancy zone protein

Abstract

Ultraviolet (UV) radiation is the primary extrinsic factor in skin aging, contributing to skin photoaging, actinic keratosis (AK), and even squamous cell carcinoma (SCC). Currently, the beneficial role of mesenchymal stromal cell-derived small extracellular vesicles (MSC-sEVs) in cutaneous wound healing has been widely reported, but the field of photoaging remains to be explored. Our results suggested that human umbilical cord MSC-derived sEVs (hucMSC-sEVs) intervention could effectively alleviate skin photoaging phenotypes in vivo and in vitro, including ameliorating UV-induced histopathological changes in the skin and inhibiting oxidative stress and collagen degradation in dermal fibroblasts (DFs). Mechanistically, pretreatment with hucMSC-sEVs reversed UVA-induced down-regulation of pregnancy zone protein (PZP) in DFs, and achieved photoprotection by inhibiting matrix metalloproteinase-1 (MMP-1) expression and reducing DNA damage. Clinically, a significant decrease in PZP in AK and SCC in situ samples was observed, while a rebound appeared in the invasive SCC samples. Collectively, our findings reveal the effective role of hucMSC-sEVs in regulating PZP to combat photoaging and provide new pre-clinical evidence for the potential development of hucMSC-sEVs as an effective skin photoprotective agent.

Graphical Abstract

Graphical Abstract.

Significance statement.

Our study demonstrated that hucMSC-sEVs were effective in preventing skin photoaging. The mechanisms were that sEVs inhibited the expression of MMP-1 and reduced DNA damage by up-regulating PZP. Therefore, hucMSC-sEVs may be a potential strategy against skin photoaging.

Introduction

Photoaging is the result of long-term exposure to ultraviolet (UV) radiation and accounts for 80% of facial aging,^1,2 leading to symptoms such as skin dryness, wrinkles formation, and hyperpigmentation.³ In more severe cases, photoaging can progress to actinic keratosis (AK), which is the most commonly diagnosed and treated skin disorder in Western countries and has the potential to develop into squamous cell carcinoma (SCC).^4,5 Furthermore, one study reported that UV irradiation-induced skin aging reduced hippocampal neurogenesis and increased depression-like behavior in mice.⁶ Consequently, effective prevention of skin photoaging not only improves facial aesthetics but is also important for preventing various skin-related diseases.

Dermal fibroblasts (DFs), the primary type of dermal cells, are accountable for synthesizing and secreting various extracellular matrix (ECM) and maintaining normal skin structure.⁷ The accumulation of senescent fibroblasts in the dermis is the main characteristic of skin aging,⁸ and UV radiation-induced increase in reactive oxygen species (ROS) generation is considered to be a central mechanism accelerating this process.⁹ ROS can activate the mitogen-activated protein kinases-related pathway, resulting in increased expression of matrix metalloproteinase (MMP) family proteins and promoting the degradation of collagen and elastin in ECM.¹⁰ In addition, excess ROS causes DNA damage, which in turn activates the DNA damage response (DDR), causing cell cycle arrest and ultimately inducing cellular senescence.¹¹

Mesenchymal stromal cells (MSCs) have garnered significant interest in tissue regeneration due to their self-renewal and multi-directional differentiation potential.¹² Recent studies have revealed that the positive effects of MSCs therapy are largely attributed to extracellular vesicles (EVs) released through the paracrine pathway.^13,14 EVs are nanoscale membrane vesicles secreted by most cells, which can selectively wrap proteins and nucleic acids,¹⁵ playing an important role in mediating intercellular communication.¹⁶ According to international guidelines, EVs with a diameter of less than 200 nm are identified as small EVs (sEVs).¹⁷ Our earlier work found that human umbilical cord MSC-derived sEVs (hucMSC-sEVs) carrying 14-3-3ζ protein could mitigate oxidative stress and improve the acute skin damage caused by H₂O₂ by up-regulating sirtuin 1 (SIRT1) in keratinocytes.¹⁸ EVs from adipose-derived stem cells have been reported to improve skin photoaging by reducing ROS production and inflammation.¹⁹ However, the roles and mechanisms of hucMSC-sEVs in skin photoaging need to be further explored.

Pregnancy zone protein (PZP), a member of the α2M macroglobulin family,²⁰ was found to be significantly increased in the serum of pregnant women during the third trimester and was initially considered as an immunosuppressive factor to protect the fetus from the attack of the mother’s immune system.²¹ Further investigation has revealed that PZP may alleviate Alzheimer’s disease and preeclampsia by inhibiting the accumulation of misfolded proteins.²² Nevertheless, there is limited literature available regarding its impact on cutaneous diseases. A recent study demonstrated that α2M treatment could ameliorate irradiation-induced fibroblast damage and mitochondrial dysfunction.²³ Therefore, PZP might play an equally important role in combating oxidative damage due to its similar properties to α2M.

In this study, we identified a beneficial effect of hucMSC-sEVs in mitigating skin photoaging. Further RNA-seq analysis explained the specific mechanism by which hucMSC-sEVs positively regulated PZP to inhibit matrix metalloproteinase-1 (MMP-1) expression and reduce DNA damage for anti-photoaging efficacy, unveiling the previously unrecognized beneficial function of PZP in cellular senescence.

Materials and methods

Clinical samples

Fresh umbilical cord tissues from parturients at Zhenjiang Fourth People’s Hospital were used to separate hucMSCs. Skin biopsy samples from 30 patients initially diagnosed with AK and SCC were collected from the Affiliated Hospital of Jiangsu University to analyze PZP expression. These studies were approved by the Ethics Committee of Zhenjiang Fourth People’s Hospital (201701) and the Ethics Committee for Experimental Animals of Jiangsu University (2020161 and 2022264).

Animals

Thirty male Balb c/nude mice (18 ± 2 g, 5-week-old) were used for in vivo experiments and housed at 22 ± 2 °C with enough food and water (12-hour light/dark cycle). For in vitro experiments, Sprague Dawley (SD) rats aged 1-2 days were killed after systemic disinfection, and then the entire back skin was peeled off to isolate DFs. All animals were purchased from the Laboratory Animal Center of Jiangsu University.

Cell culture and identification

HucMSCs were isolated as previously described,²⁴ and cultured at 37 °C with 5% CO₂ in minimum essential medium α (α-MEM, Gibco) containing 15% fetal bovine serum (FBS, Gibco). Cells from the 2nd to 4th generation were used to identify the biological characteristics of MSCs, and the cell supernatant from the 3rd to 7th generation was collected to extract hucMSC-sEVs.

Dorsal skins of SD rats were cut into 1-2 mm³ small pieces and cultured in α-MEM medium supplemented with 10% FBS for one week, waiting for DFs to climb out. The first-generation DFs were used to detect surface markers, and P3-P5 were plated for in vitro experiments.

Separation, identification, and tracking of HucMSC-sEVs

HucMSC-sEVs were extracted from collected supernatant by ultracentrifugation, as previously described.²⁵ The precipitate was resuspended in phosphate buffer saline (PBS, Meilunbio) and filtered with 0.22 μm filter (Millipore). Nanoparticle tracking analyzer (NTA, Particle Metrix) was used to detect the size distribution and particle number of vesicles. The morphology of sEVs was observed by transmission electron microscopy (TEM, FEI Tecnai 12, Philips). The protein concentration of the purified sEVs was measured by a BCA protein quantification kit (Vazyme). According to the manufacturer’s protocol, hucMSC-sEVs were labeled with CM-Dil membrane dye (Invitrogen) and co-incubated for 1 hour in the dark. The mixed solution was then transferred to a 100 kDa MWCO ultrafiltration centrifuge tube (Millipore) to wash the unbound dye with PBS. Subsequently, the back skin of mice was treated with 5 × 10¹⁰ labeled Dil-sEVs combined with microneedle roller (MEVOS) and syringe, respectively, and a live imaging instrument (Cri Maestro) was used to detect the fluorescence. In vitro, DFs were co-incubated with Dil-sEVs for 24 hours, and internalization was observed under a laser confocal microscope (Nikon).

UV-irradiation photoaging model and treatment

Balb/c nude mice mentioned above were fed adaptively for a week before the formal experiment and then randomly divided into 5 groups, 6 animals per group: Control (Ctrl), UV radiation (UV), UV + 0.05% retinoic acid (UV + RA), UV + microneedle roller + PBS (UV + PBS), UV + microneedle roller + hucMSC-sEVs (UV + sEV). Nude mice in the Ctrl group were routinely raised without any treatment, and the other groups were exposed to UV for 7 or 10 rounds, respectively (5 days per round, including 1 day of pretreatment, 3 days of exposure, and 1 day of interval). On the first day of each round, the UV + RA group animals were smeared with 0.05% RA, and the last 2 groups were treated with microneedle roller and then applied PBS or 5 × 10¹⁰ sEVs to the dorsal skin (encapsulation with plastic wrap after skin treatment). During the exposure period, only the back skin of mice was exposed. The first round of light dose was 1 minimum erythema dose (MED), and then every 2 rounds will be increased by 1 MED. According to the pre-experiment, the final duration of 1 MED was determined to be 5 minutes, the UVA intensity was about 4 μW/cm², and the UVB intensity was about 0.6 μW/cm².

In vitro, DFs were seeded in six-well plates (ExCell) and pretreated with 1 mM NMN (>98.0% HPLC, TCI), hucMSC-sEVs of 1 × 10¹⁰/mL or a mixture of NMN and hucMSC-sEVs for 24 hours. Subsequently, the medium was discarded, a thin layer of PBS was spread, and DFs were placed 5 cm below the 365 nm UVA lamp (Philips, Shanghai, China) for irradiation. At the same time, the probe of the ultraviolet radiation meter (LINSHANG) was used to observe the irradiation dose in real time.

Cell viability

1 × 10³ DFs irradiated by UV were seeded in 96-well plates (ExCell) and incubated overnight at 37 °C. After cell adhesion, the cell counting kit-8 (CCK8, Vazyme) reagent was added to each well under dark conditions and co-incubated with cells for 2 hours, and the OD value at 450 nm was measured using a microplate reader (BioTek).

Western blot analysis

Cells were fully lysed on ice using RIPA lysis buffer (Pierce) containing protease inhibitors (Pierce). After centrifugation at 12 000 × g for 15 min, the supernatant was collected, mixed with LDS sample buffer (Invitrogen), and boiled for 10 min. The protein samples were separated by SDS-PAGE and transferred to polyvinylidene fluoride membranes (Millipore). The membranes were blocked with 5% skimmed milk at room temperature (RT) for 2 hours, incubated with primary antibody at 4 °C overnight, and then incubated with horseradish peroxidase (HRP)-conjugated secondary antibody (1:2000, Invitrogen) at RT for 2 hours. The labeled protein was quantified with ECL (Vazyme) and detected by a chemiluminescence gel imager. The primary antibodies are listed below: β-actin (1:5000, Abclonal), COL1A1 (1:500, Abclonal), MMP-1 (1:500, Abclonal), SIRT3 (1:500, Wanleibio), γ-H2AX (1:500, CST), PZP (1:500, ABclonal, Catalog No.: A3324), ataxia telangiectasia-mutated protein (ATM, 1:500, Abclonal), p-ATM (1:500, Abclonal), 53BP1 (1:500, Proteintech), MMP-9 (1:500, Wanleibio), FMO2 (1:500, Proteintech), Alix (1:500, CST), CD9 (1:500, CST), CD81 (1:500, Proteintech), CD63 (1:500, Proteintech), TSG101 (1:500, Proteintech), and Calnexin (1:500, CST).

RNA extraction and reverse transcription-quantitative polymerase chain reaction (RT-qPCR)

Total RNA was extracted from cell samples with Trizol reagent (Invitrogen), and the concentration and purity were measured by Nanodrop 1000 spectrophotometer (Thermo Fisher Scientific). cDNA was synthesized using the HiScript Ⅲ First Strand cDNA Synthesis Kit (Vazyme) according to the manufacturer’s instructions. RT-qPCR was performed in 96-well plates using the AceQ qPCR SYBR green master mix (Vazyme), and the mRNA expression levels of col1a1, mmp-1, sirt1, sirt3, p16, p21, p53, and all differentially expressed genes were analyzed by the Quantitative 3 Real-Time PCR Detection System (Thermo Fisher Scientific), with β-actin used as endogenous control. All primers were produced by Sangon Biotech, and the primer sequences used in this study are listed in Supplementary Table S1.

Protein docking

Protein structures were downloaded from the UniProt database, and HDOCK (http://hdock.phys.hust.edu.cn/) was used to perform protein–protein docking. The docking results were then visualized using Pymol.

siRNA and plasmid transfection

pzp-siRNAs and the pzp-negative control (nc) were purchased from GenePharma. pzp overexpression (oe) plasmid was constructed by Fenghbio. 1 × 10⁵ DFs in 6-well-plate were cultured in serum-free medium for 24 hours, then transfected with 200 nM pzp-siRNA and pzp-nc using Lipofectamine 2000 (Invitrogen) for 48 hours. The plasmid transfection method was the same as above, but the applied concentration was 2 μg. RT-qPCR and Western blot were used to detect transfection efficiency. The related sequences are listed in Supplementary Table S2.

Statistical analysis

The statistical analysis was performed with GraphPad Prism 8.0 software (San Diego), and all data were presented as mean ± standard deviation (SD). Differences between 2 groups were analyzed using unpaired t-test, while multiple groups were compared using one-way ANOVA. P value < .05 indicated statistical significance.

Other methods can be found in the supplementary information.

Results

Characterization of hucMSCs and hucMSC-sEVs

The hucMSCs obtained were short fusiform and arranged like a school of fish (Figure 1A). After induction by osteogenic and adipogenic medium, staining revealed that calcium nodules and lipid droplets appeared (Figure 1B and C). Flow cytometry analysis showed that hucMSCs highly expressed surface markers CD29, CD44, and CD166, while CD14, CD45, and HLA-DR expression ratios were lower than 2% (Figure 1D). The above results indicated the successful isolation of hucMSCs with multidifferentiation capability. HucMSC-sEVs extracted from the conditioned medium of hucMSCs had a typical membrane-type vesicle-like structure under TEM (Figure 1E). NTA and Western blot revealed that hucMSC-sEVs mainly enriched at 104.8 nm (Figure 1F), expressed characteristic markers such as Alix, CD9, CD81, CD63, and TSG101, and negatively expressed Calnexin (Figure 1G). To determine the optimal delivery method of sEVs in the in vivo experiments, we conducted a comparative analysis between intradermal injection and microneedle roller. The bioluminescence image demonstrated that the fluorescence signal of hucMSC-sEVs followed by microneedle roller treatment was stronger and more widely distributed than intradermal injection treatment (Figure 1H). In addition, immunofluorescence showed that sEVs delivered by the syringe gathered in a certain area of the dermis, while sEVs treated with microneedle roller were uniformly distributed in the dermis tissue (Figure 1I). Therefore, we chose microneedle roller as the delivery method for sEVs.

Figure 1.

Characterization of hucMSCs and hucMSC-sEVs. (A) Morphological identification of hucMSCs, P0 (left) and P1 (right). Scale bar = 100 μm. (B) Alizarin Red S staining detected the osteogenic differentiation capacity of hucMSCs. Scale bar = 20 μm. (C) Oil-red-O staining detected the adipogenic differentiation capacity of hucMSCs. Scale bar = 20 μm. (D) Flow cytometry analysis of surface markers of hucMSCs. (E) Morphology of hucMSC-sEVs under TEM. Scale bar = 50 nm. (F) NTA showed the size distribution of hucMSC-sEVs. (G) Western blot analysis of hucMSC-sEVs marker proteins. (H) In vivo imaging detection of fluorescence intensity of Dil-sEVs after intradermal injection or microneedle delivery. (I) Immunofluorescence showed the distribution of Dil-labeled sEVs in skin tissue after intradermal injection or microneedle treatment for 24 h. Scale bar = 200 μm.

HucMSC-sEVs effectively prevent skin photoaging in nude mice

To investigate whether hucMSC-sEVs have a positive effect on skin photoaging, we established a UV-induced photoaging model of nude mice (Figure 2A). After 35 days of UV exposure, the dorsal skins of the UV group exhibited severe desquamation, thinning, redness, and wrinkling (Figure 2B). And compared with other groups, the mice in the UV group showed significant weight loss (Figure 2C). HE staining suggested that UV exposure led to stratum corneum thickening but thinning of the overall skin (Figure 2D and E). Meanwhile, Masson staining displayed a decrease in collagen volume fraction, accompanied by irregular and denatured arrangement of some collagen bundles, as well as broken ones (Figure 2F and G). Whereas exogenous supplementation of hucMSC-sEVs improved skin appearance and histopathological changes induced by UV, including wrinkle formation, desquamation, dermis thinning, and collagen structure disorder, and the effect was better than the positive control 0.05% RA, a treatment for photoaging approved by FDA (Figure 2B-G). Moreover, hucMSC-sEVs significantly increased the expression of COL1A1 and SIRT3 (Figure 2H). Notably, under longer UV exposure, hucMSC-sEVs treatment also exhibited an excellent anti-aging effect (Supplementary Figure S1A-1C). These findings collectively indicated that hucMSC-sEVs could effectively suppress the photoaging process caused by UV.

Figure 2.

HucMSC-sEVs effectively prevent skin photoaging in nude mice. (A) Schematic diagram of UV-induced photoaging model. (B) Gross view of dorsal skin of nude mice on day 35 of modeling. (C) Weight changes in nude mice on day 35 of modeling. (D, E) HE staining (D) and quantification analysis (E) of skin tissue sections. Scale bar = 20 μm. (F, G) Masson staining (F) and quantification analysis (G) of collagen volume fraction. Scale bar = 20 μm. (H) Immunofluorescence detection of COL1A1 and SIRT3 expression in skin tissues. Scale bar = 20 μm. Ctrl: untreated mice, UV: mice subjected to UV, UV + RA: mice treated with retinoic acid before UV irradiation, UV + PBS: mice treated with PBS by microneedle roller before UV irradiation, UV + sEV: mice treated with hucMSC-sEVs by microneedle roller before UV irradiation. Data are presented as mean ± SD. n = 3, **P < .01, ***P < .001, ns: no significant difference.

Establishment of photoaging model of DFs by UVA irradiation

In vitro, we established a cellular photoaging model applying UVA radiation. DFs isolated from the rat dorsal skin displayed a short spindle shape (Figure 3A), expressed positive markers vimentin and α-SMA, but negatively expressed keratinocyte-related CK10 (Figure 3B). After exposure to UVA three times, the number of dead DFs, the level of ROS, and the expression of p53, p21, and mmp-1 were increased in a dose-dependent manner. The most significant changes were observed at the dose of 5 J/cm² (Figure 3C-E).

Figure 3.

Construction of the UV-induced photoaging model of DFs. (A) Morphological identification of DFs. Scale bar = 5 μm. (B) Fluorescence images showed the DFs biomarkers (Vimentin, α-SMA) and the negative biomarker CK10. Scale bar = 25 μm. (C) Number of living cells exposed to 1 J/cm², 2 J/cm², or 5 J/cm² UVA for 3 times. (D) Detection of ROS fluorescence intensity in DFs. Scale bar = 300 μm. (E) RT-qPCR detection of p53, p21, p16, and mmp-1. (F) CCK8 assay for the proliferation ability of DFs exposed to 5 J/cm² UVA for 1, 3, and 5 times. (G) SA β-gal staining images and their quantification diagram. Scale bar = 10 μm. (H) RT-qPCR analysis of the expression of col1α1, mmp-1, sirt1, and sirt3 at different time points after 5 J/cm² UVA irradiation. Data are presented as mean ± SD. n = 3, *P < .05, **P < .01, ***P < .001, ns: no significant difference.

Then we found that 5 J/cm² UVA irradiation for 1, 3, or 5 times could significantly inhibit the proliferation of DFs (Figure 3F). A remarkable increase in the number of β-gal positive cells was evident subsequent to 5 times UVA irradiation (Figure 3G). Furthermore, the expression of col1α1 in DFs was dramatically down-regulated within 48 h, and the expression of mmp-1 was up-regulated within 12 h, while the expression of deacetylase sirt1 and sirt3 involved in the regulation of cell stress was also considerably reduced within 24 h (Figure 3H). Based on the above results, 5 times irradiation with 5 J/cm² UVA was used for establishing the photoaging model of DFs.

HucMSC-sEVs protect DFs against UVA-induced oxidative stress

HucMSC-sEVs, nicotinamide mononucleotide (NMN), or the mixture were added respectively 24 hours prior to UVA exposure (Figure 4A). Figure 4B showed that Dil-labeled hucMSC-sEVs could be effectively internalized by DFs. CCK8 assay revealed that 1 mM NMN did not produce cytotoxicity after co-incubation with DFs for 24 hours or 48 hours (Figure 4C). The number of β-gal positive cells increased after UVA irradiation, whereas the number of β-gal positive cells decreased in the other 3 intervention groups (Figure 4D). Moreover, pretreatment with hucMSC-sEVs or the mixture could significantly restore COL1A1 expression and inhibit ROS generation, and the mixture appeared to be superior (Figure 4E and F). Immunofluorescence presented a considerable increase in the level of γ-H2AX after UV exposure, which was reduced in the intervention groups (Figure 4G). In addition, we observed an obvious up-regulation of COL1A1, a corresponding down-regulation of MMP-1, and an increase in SIRT3 after pretreating with hucMSC-sEVs (Figure 4H). All these results suggested that hucMSC-sEVs might have a more effective role in protecting DFs against UVA-induced oxidative stress and cellular senescence compared to NMN.

Figure 4.

Pretreatment with hucMSC-sEVs mitigates the UV-induced oxidative stress and DNA damage in DFs. (A) Schematic diagram of DFs photoaging intervention model. (B) Fluorescence images showed the uptake of Dil-sEV after co-incubation with DFs for 24 hours. Scale bar = 25 μm. (C) CCK8 showed the proliferative ability of DFs after 24 hours or 48 hours treatment with NMN. (D) SA β-gal staining images of the DFs and their quantification diagram. Scale bar = 10 μm. (E) Immunofluorescence detection of COL1A1. Scale bar = 10 μm. (F) Histogram showed the levels of ROS detected by flow cytometry and their quantification. (G) Immunofluorescence exhibited the level of γ-H2AX. Scale bar = 20 μm. (H) Western blot analysis showed the expression of MMP-1, COL1A1, and SIRT3. Ctrl: untreated DFs, UV: UVA-irradiated DFs, UV + NMN: UVA-irradiated DFs after treatment with NMN, UV + sEV: UVA-irradiated DFs after treatment with hucMSC-sEVs, UV + NMN + sEV: UVA-irradiated DFs after treatment with NMN and hucMSC-sEVs. Data are presented as mean ± SD. n = 3, *P < .05, **P < .01, ***P < .001, ns: no significant difference.

Anti-aging effect of hucMSC-sEVs is closely related to the induced expression of PZP

We further performed an unbiased RNA-seq analysis to explore the mechanism of hucMSC-sEVs in preventing skin photoaging. Compared with the Ctrl group, UVA irradiation up-regulated 160 genes and down-regulated 169 genes in DFs. In the UVA-irradiated cells following pretreatment with hucMSC-sEVs, the expression of 53 genes changed, with 24 genes up-regulated and 29 genes down-regulated (Figure 5A and B). Based on the Venn and R language analysis, we preliminarily screened 5 top genes, namely pzp, trh, sall1, tex26, and fmo2 (Figure 5C and D). RT-qPCR results displayed that the changes of pzp and fmo2 were consistent with the transcriptome data: pzp was significantly down-regulated after UV irradiation and up-regulated following hucMSC-sEVs pretreatment, while the changing mode of fmo2 was opposite to pzp (Figure 5E and F). Further Western blot verified that pzp was the top-ranked DEG dynamically expressed among the three groups (Figure 5G). Gene Ontology analysis showed that pzp not only participated in pregnancy-related physiological processes but was mainly involved in the negative regulation process of peptidase activity as an endopeptidase inhibitor (Figure 5H). Immunohistochemical analysis demonstrated that hucMSC-sEVs pretreatment could effectively restore PZP levels in the skin tissues of nude mice (Figure 5I).

Figure 5.

Screening and verification of the differentially expressed gene pzp in DFs and skins based on RNA-seq. (A) DEGs thermograms of 3 independent duplicate samples of the Ctrl, UV, and UV + sEV groups. (B, C) Volcano maps (B) and Venn maps (C) of DEGs. (D) R language analysis of TOP DEGs. (E) RNA-seq analysis of DEGs, including pzp, sall1, fmo2, trh, and tex26. (F) RT-qPCR validation of the DEGs. (G) Western blot detection of the expression of PZP, FMO2, and γ-H2AX. (H) GO functional analysis of pzp. (I) Immunohistochemical detection of PZP expression in skin tissues of nude mice on day 35 of modeling. Scale bar = 20 μm. Data are presented as mean ± SD. n = 3, **P < .01, ***P < .001, ns, no significant difference.

MMP-1 serves as a protein of interest for interaction with PZP

To determine the specific mechanism by which PZP alleviated skin aging, we searched for its target proteins. Through the STRING database analysis of the PZP interaction protein network, it was found that there is an interaction relationship between PZP and matrix metalloproteinase-9 (MMP-9) (Figure 6A). Therefore, we speculated that MMP-1, belonging to the same family as MMP-9, may also interact with PZP. Through protein docking, the specific binding patterns of PZP and MMP-9 were modeled (Figure 6B), and the binding of PZP and MMP-1 was predicted (Figure 6C). Co-immunoprecipitation assay validated this hypothesis (Figure 6D). Furthermore, immunofluorescence analysis revealed the co-localization of PZP and MMP-1 in both DFs and skin tissues of nude mice (Figure 6E and F).

Figure 6.

Validation of PZP and MMP-1 interaction. (A) Known proteins interacting with PZP identified by the STRING database. (B) Predicted protein docking pattern of PZP and MMP-9. (C) Predicted protein docking pattern of PZP and MMP-1. (D) Validation of the interaction between PZP and MMP-1 via co-immunoprecipitation. (E) The co-localization of PZP and MMP-1 in DFs was assessed using immunofluorescence assay. Scale bar = 20 μm. (F) Immunofluorescence co-localization analysis of PZP and MMP-1 in skin tissues of nude mice. Scale bar = 200 μm.

PZP ameliorates skin photoaging by mainly inhibiting MMP-1 expression

Subsequently, a sequence of overexpression and knockdown experiments in SD rat DFs was conducted to investigate the impact of PZP on MMP-1. As shown in Supplementary Figure S2A, transfection of pzp-overexpression plasmid successfully up-regulated the level of PZP in DFs. SA β-gal staining showed that increased PZP reduced the number of senescent cells (Figure 7A). Moreover, the UV + sEV + oe-pzp group exhibited lower MMP-1 expression and higher COL1A1 expression in comparison to the UV + sEV + vector group, indicating that PZP may decelerate collagen degradation by inhibiting MMP-1 (Figure 7B). Fluorescence and Western blot analysis also demonstrated an increase in SIRT3 expression with the elevation of PZP (Figure 7B and C). Furthermore, our results indicated that PZP overexpression reduced DNA damage and decreased the levels of DDR-related proteins, including p-ATM, γ-H2AX, and 53BP1 (Figure 7B and D). After that, the effects of PZP knockdown on UV-irradiated DFs were also investigated. The transfection efficiency of pzp-siRNA fragments was assessed by immunofluorescence, RT-qPCR, and Western blot (Supplementary Figure S2B-2D). Knockdown of the pzp gene diminished the impact of hucMSC-sEVs on cellular senescence (Figure 7E) and reversed the suppression of MMP-1, leading to a subsequent decrease in COL1A1 levels (Figure 7F). Similarly, si-pzp transfection attenuated the protective effect on SIRT3 and the inhibition of DNA damage (Figure 7F-H).

Figure 7.

PZP ameliorates photoaging by mainly inhibiting MMP-1 expression. After overexpression of the pzp gene (A) SA β-gal staining and its quantization diagram. Scale bar = 100 μm. (B) Detection of the expression of PZP, COL1A1, MMP-1, SIRT3, and ATM signaling pathway-related proteins by Western blot. (C) Fluorescence images of mitotracker staining and SIRT3. Scale bar = 200 μm. (D) Immunofluorescence detection of γ-H2AX expression. Scale bar = 200 μm. After knockdown of the pzp gene (E) SA β-gal staining and its quantization diagram. Scale bar = 100 μm. (F) The levels of COL1A1, MMP-1, and DDR-related proteins were quantified by Western blot. (G) Fluorescence images of mitotracker staining and SIRT3. Scale bar = 200 μm. (H) Analysis of γ-H2AX expression by immunofluorescence. Scale bar = 200 μm. Ctrl + vector: DFs transfected with empty vector, UV + vector: UVA-irradiated DFs after transfection with empty vector, UV + sEV + vector: UVA-irradiated DFs after transfection with empty vector and treatment with hucMSC-sEVs, UV + sEV + oe-pzp: UVA-irradiated DFs after transfection with pzp-overexpression plasmid and treatment with hucMSC-sEVs. Ctrl + si-nc: DFs transfected with unordered fragments, UV + si-nc: UVA-irradiated DFs after transfection with unordered fragments, UV + sEV + si-nc: UVA-irradiated DFs after transfection with unordered fragments and treatment with hucMSC-sEVs, UV + sEV + si-pzp: UVA-irradiated DFs after transfection with siRNA of pzp and treatment with hucMSC-sEVs. Data are presented as mean ± SD. n = 3, *P < .05, ***P < .001, ns, no significant difference.

Marked alterations in PZP levels of skin samples from patients with AK and SCC

To investigate whether PZP has clinical significance, we examined its expression in AK and SCC samples. Among the 30 skin samples with primary clinical diagnoses of AK and SCC, 24 were double-checked as AK and SCC by experienced doctors and used in this experiment. The 24 patients visited hospitals between 2013 and 2022, and a detailed description of their disease and age distribution can be found in Supplementary Table S3. Immunohistochemical staining from the 12 AK samples showed that the basal layer cells proliferated in a bud-like manner toward the superficial dermis, with a substantial number of lymphocytes infiltrating. The positive rate of PZP in the dermis and epidermis of 10 AK tissue pathological areas was significantly lower than that in the adjacent normal areas (Supplementary Figure S3A), while there was no significant change in the other 2 AK tissues. In 6 samples that progressed to SCC in situ, the pathological areas exhibited a full-thickness epidermal cells atypia with local dermal papilla compressed into thin ribbons, and PZP showed extremely low expression in the lesion areas (Supplementary Figure S3B). In other 6 samples that developed into invasive SCC (iSCC), cancerous cell mass penetrated the basal layer and invaded the dermis. Among them, 5 showed a significant increase in PZP positivity in the cancerous areas (Supplementary Figure S3C), while 1 did not. In the latest 2 skin lesions of AK and iSCC collected from the same patient, we analyzed the PZP expressions in the 2 different kinds of pathological areas respectively. Notably, the result showed that compared with the adjacent areas of AK samples, the positive rate of PZP in both dermis and epidermis of the pathological areas decreased. However, in the iSCC tissues, the positive rate in the cancerous area significantly recovered (Supplementary Figure S3D). In a tissue array comprising 12 AK samples, positive (score 1-3) staining of PZP was detected in 100% (12/12) of adjacent areas, and 83.3% (10/12) of pathological areas showed weak positive (score 1) or negative staining (score 0). Conversely, in 6 iSCC samples, weak (score 1) and negative (score 0) staining of PZP was detected in 66.7% (4/6) of adjacent areas, whereas 100% (6/6) of pathological areas showed positive staining (score 1-3) (Supplementary Figure S3E). The H-scores revealed a significantly lower immunoreactivity of PZP in pathological areas of AK, while a high expression level in the cancerous areas of SCC tissues (Supplementary Figure S3F). In consideration of our prior research indicating a protective function of 14-3-3ζ in photodamage, we also examined its expression in AK and SCC samples and found no changes in either of these conditions (Supplementary Figure S3G).

Discussion

Here, our research identified hucMSC-sEVs as a potential means to mitigate photoaging and further elucidated their underlying mechanisms in restoring PZP levels in DFs to delay various aging phenotypes, including inhibiting MMP-1 expression and reducing DNA damage.

Among the 3 types of UV rays, UVB with a short wavelength mainly causes damage to the epidermis, whereas UVA with a strong penetration can directly reach the dermis, which is the major culprit for photoaging.²⁶ Current strategies for preventing and treating skin photoaging include vitamin-based antioxidants, sunscreens, lasers, and dermal fillers, but all have certain limitations.²⁷ Consequently, there is a need to develop novel and efficient strategies to combat photoaging. MSC-sEVs, a cell-free therapy that avoids the risks of immune rejection and tumorigenicity of MSC treatment,²⁸ have gained increasing attention in tissue repair in recent years.^29,30 Hence, we sought to investigate their functions and molecular mechanisms in skin photoaging. Regarding the application of sEVs in the skin, their absorption rate is constrained due to the skin barrier.³¹ Microneedle is a new transdermal drug delivery device that can promote drug permeability through reversible skin cuticle destruction, garnering significant interest in tissue regeneration.³² Compared with intradermal administration, microneedle has a lower probability of skin irritation and inflammation.³³ Similarly, our findings indicated that microneedle roller had advantages over intradermal injection as a mode of administration for sEVs, and that hucMSC-sEVs combined with microneedle roller effectively reduced UV-induced collagen loss and the accumulation of abnormal collagen fibers, preserving dermal thickness.

ROS-induced increase in MMPs generation and DNA damage are important mechanisms involved in skin photoaging.³⁴ In the MMPs family, MMP-1 is the main enzyme that degrades type I collagen.³⁵ After UV irradiation, the expression of MMP-1 significantly increases, and inhibiting its expression may be one of the ways to delay the aging process.³⁶ SIRT3, as a member of the longevity protein family Sirtuins, can enhance the activity of endogenous antioxidant enzymes and prevent the production of ROS, thus protecting cells from oxidative damage.^37,38 Hence, increasing the expression of SIRT3 can confer anti-aging benefits. We found that hucMSC-sEVs intervention could achieve similar effects, including down-regulation of MMP-1 expression and increased COL1A1 and SIRT3 expression. UV radiation can directly or indirectly cause DNA damage.³⁹ DNA double-strand breaks (DSBs) are the most severe type of DNA damage, and when they occur, ATM is activated, causing rapid phosphorylation of the serine 139 site of H2AX to form γ-H2AX.⁴⁰ γ-H2AX, a marker of DSBs, recruits relevant repair proteins to the damaged site, initiating DDR to promote DNA damage repair.⁴¹ Our results suggested that DNA damage was exacerbated, and ATM phosphorylation levels were significantly increased after UVA irradiation. However, hucMSC-sEVs pretreatment reduced ROS generation and DNA damage, thereby down-regulating the expression of p-ATM, γ-H2AX, and 53BP1 associated with DDR. Taken together, hucMSC-sEVs demonstrated a significant ability to suppress UV-induced cellular senescence and functional decline. Additionally, we established a group in which hucMSC-sEVs were incubated with NMN to explore the combined intervention effects. NMN, a direct precursor of coenzyme I, has been shown to improve various aging phenotypes by raising NAD + levels.^42,43 Notably, the combined intervention exhibited superior effects in inhibiting cell aging, oxidative stress, and collagen degradation, however, the precise mechanism remains to be investigated.

Based on the beneficial role of hucMSC-sEVs in preventing photoaging, further RNA-seq analysis confirmed that PZP was closely related to UV resistance. The latest research indicated a reduction in PZP levels in aging BJ fibroblasts induced by ionizing radiation,⁴⁴ consistent with our results that UV radiation down-regulated the expression of PZP in DFs. PZP exhibits notable endopeptidase inhibitor activity and has been shown to be involved in the inhibition of MMPs, a group of zinc-dependent endopeptidases.⁴⁵ Hence, we deduced that PZP might mediate ECM degradation by regulating MMP-1. The interaction between PZP and MMP-1 was confirmed by co-immunoprecipitation and fluorescence co-localization experiments, and overexpression or knockdown of PZP in DFs could enhance or attenuate the inhibitory effect of hucMSC-sEVs on MMP-1. Therefore, we concluded that hucMSC-sEVs improved photoaging by regulating PZP to inhibit MMP-1, thereby impeding the degradation of COL1A1. In addition, we found that PZP inhibition increased the activation of ATM, but the exact mechanism needs further exploration.

The long-term effects of UV radiation on the skin include not only photoaging but also certain precancerous and cancerous skin lesions, such as AK and SCC.⁴⁶ Currently, the specific mechanism of AK is not entirely clear, and its accurate diagnosis usually necessitates experienced staff, causing difficulties in the diagnosis and treatment of this disease. Consistent with the alterations in photoaging tissues, immunohistochemical analysis of clinical skin samples revealed a significant reduction in the level of PZP in the vast majority of AK pathological areas, with extremely low expression in the SCC in situ. Conversely, in the cancerous areas of iSCC samples, PZP expression exhibited an increase. These results suggested that PZP might have a significant role in the occurrence and development of AK and SCC. Nevertheless, further clinical data is required to substantiate this hypothesis.

To investigate the downstream pathway of PZP more deeply, we performed mass spectrometry analysis on the interacting proteins of PZP. Among the identified proteins, the heat shock protein 70 (HSP70) family (HSPA9, HSPA5, HSPA8) had higher abundance values, which is essential in assisting the correct folding of proteins.^47,48 Subsequently, we will focus on elucidating the mechanisms through which PZP modulates the HSP70 family to mitigate skin photoaging. In addition, we aimed to further explore the precise mechanism by which hucMSC-sEVs up-regulated PZP in UVA-irradiated DFs. Based on RNA-seq results, we speculated that sEVs may carry transcription factors capable of binding to the promoter of PZP in DFs, thereby initiating transcription and consequently increasing PZP expression. Following this, we will use bioinformatics analysis and dual luciferase reporter gene assay to predict and validate the transcription factors of PZP.

In conclusion, our findings provide a kind of theoretical support for the beneficial application of hucMSC-sEVs in skin photoaging and reveal the potential of PZP as a novel target in anti-cell senescence and UV-induced cutaneous diseases.

Conclusion

Our research confirmed that hucMSC-sEVs effectively combated photoaging by up-regulating the expression of PZP, demonstrating the potential of PZP as a novel anti-aging target. The beneficial effects may be achieved by regulating PZP-mediated collagen degradation and the ATM pathway. These findings provide a kind of theoretical support for the application of hucMSC-sEVs in anti-aging and highlight the important role of PZP in the occurrence and development of AK and SCC.

Supplementary material

Supplementary material is available at Stem Cells Translational Medicine online.

Author contributions

Zixuan Sun, Tangrong Wang, Xiaomei Hou: conception and design, conducted the experiments, collected the data, and manuscript writing. Wenhuan Bai, Jiali Li: data analysis and interpretation, and the reversion of the manuscript. Yu Li: performed in vivo experiments. Jiaxin Zhang: created a graphical abstract, and data analysis and interpretation. Yuzhou Zheng, Zhijing Wu: performed protein docking, and checked the manuscript. Peipei Wu: conception and design, and provided experimental protocols. Hui Qian, Lirong Yan, Zixuan Sun: designed the project, financial support, administrative support, and final approval of manuscript.

Funding

This work was funded by the National Natural Science Foundation of China, grant number 82003379, Zhenjiang Key Laboratory of High Technology Research on Exosomes Foundation and Transformation Application, grant number SS2018003, and Geriatric Health Research Project of Jiangsu Health Commission, grant numbers LKZ2011012 and LR2021015.

Conflict of interest

The authors confirm that there are no conflicts of interest regarding the publication of this article.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.