3D spheroids of umbilical cord-derived MSCs protect retinal pigment epithelium against oxidative and inflammatory injury by activating autophagy

Abstract

Background

Age-related macular degeneration (AMD) is characterized by progressive retinal pigment epithelium (RPE) dysfunction driven by oxidative stress and chronic inflammation, in which NLRP3 inflammasome activation plays a critical role. Mesenchymal stem cells (MSCs) exhibit therapeutic potential, but their efficacy is limited by poor survival and reduced paracrine activity in hostile microenvironments. Here, we investigated whether three-dimensional (3D) spheroid culture enhances the protective effects of umbilical cord-derived MSCs (UC-MSCs) on RPE cells by promoting autophagy and suppressing inflammasome activation.

Methods

Human UC-MSCs were cultured as 3D spheroids or conventional 2D monolayers and applied in sodium iodate (NaIO₃)-induced oxidative injury models both in vitro and in vivo. Retinal morphology and function were assessed via histology and electroretinography, while NLRP3/caspase-1 activation, LC3-II/I ratios, and autophagy flux were quantified using immunofluorescence and Western blot. GO/KEGG enrichment was performed to identify pathways associated with 3D MSCs efficacy. Mechanistic involvement of autophagy was validated using 3-methyladenine (3-MA) and rapamycin.

Results

3D MSCs formed compact spheroids exhibiting enhanced paracrine potential and significantly outperformed 2D MSCs in protecting RPE cells against NaIO₃-induced injury. In vivo, 3D MSC treatment preserved retinal structure, reduced RPE cell loss, and improved retinal function. In vitro, co-culture with 3D MSCs markedly improved ARPE-19 viability, reduced apoptosis, and modulated autophagy-related marker expression, as evidenced by increased LC3-II/I ratios. 3D MSCs significantly inhibited NLRP3 inflammasome activation and pro-inflammatory cytokine release, effects reversed by 3-MA and further enhanced by rapamycin.

Conclusions

3D spheroid culture substantially augments the therapeutic efficacy of UC-MSCs by boosting autophagy and suppressing NLRP3 inflammasome signaling, resulting in enhanced protection of RPE cells from oxidative and inflammatory injury. These findings provide preclinical evidence supporting 3D MSCs as a promising therapeutic strategy for AMD.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s13287-026-05043-z.

Introduction

Age-related macular degeneration (AMD), a leading cause of irreversible blindness in individuals over 60 years old, is characterized by the progressive dysfunction and loss of retinal pigment epithelium (RPE) cells [1, 2]. Chronic oxidative stress (OS) and inflammation in the RPE, particularly the aberrant activation of the NLRP3 (Nod-like receptor protein 3) inflammasome, are recognized as major drivers of AMD pathogenesis [3]. Prior investigations have illuminated that excessive reactive oxygen species (ROS) can overwhelm the RPE’s defenses, causing organelle damage and protein misfolding, which further amplify complement activation and inflammatory responses, thereby accelerating RPE injury [4–6]. Moreover, oxidative insults can directly activate the NLRP3 inflammasome in RPE cells, activating caspase-1 and releasing pro-inflammatory cytokines such as interleukin-1β (IL-1β) and IL-18, which amplifies retinal inflammation [7]. These findings suggest that mitigating RPE oxidative injury and chronic inflammation could be a promising strategy for preventing vision loss in AMD.

Beyond ROS and inflammation, growing evidence highlights that impaired autophagy in RPE cells contributes to AMD development. Autophagy is an essential homeostatic mechanism that removes damaged proteins and organelles. Disruptions in autophagic flux exacerbate oxidative damage and promote inflammatory signaling. Recent studies demonstrate that defective autophagy in RPE cells leads to increased NLRP3 inflammasome activation, amplifying retinal inflammation and accelerating cell death [8]. Given this evidence, enhancing autophagy has emerged as a promising approach to protect RPE cells from OS-induced injury and inflammasome activation in AMD [9].

Recently, mesenchymal stromal cells (MSCs) have gained significant attention as therapeutic agents for a wide range of degenerative disorders, including ocular conditions, due to their immunomodulatory and tissue-reparative effects [10, 11]. These effects are largely attributed to MSC-secreted bioactive factors that reduce inflammation, promote angiogenesis, and enhance tissue repair [12]. Among MSC sources, umbilical cord-derived MSCs (UC-MSCs) offer practical advantages such as non-invasive collection and robust expansion, and have been widely used in spheroid-based three-dimensional (3D) culture systems, which can enhance their paracrine and immunomodulatory profiles. These properties motivated us to select UC-MSCs for 3D spheroid preconditioning in the present study [13–15]. However, the therapeutic benefits of MSC-based therapy are often limited by poor cell survival and functional decline in the hostile retinal microenvironment, which has motivated exploration of strategies to enhance MSC potency [16].In various disease models, including inflammatory and degenerative conditions, 3D MSCs outperform their 2D counterparts [17, 18]. However, there is limited data on their application in AMD models or their influence on RPE autophagy and inflammasome signaling.

Based on these observations, we hypothesized that 3D MSCs exert enhanced therapeutic effects by promoting autophagy and suppressing NLRP3 inflammasome activation in RPE cells under oxidative stress. To test this hypothesis, we utilized sodium iodate (NaIO₃)-induced oxidative injury, a widely accepted model for dry AMD pathology, both in vitro and in vivo. We compared the protective effects of 3D- versus 2D- cultured umbilical cord MSCs (UC-MSCs) on RPE morphology, function, and molecular markers of autophagy and inflammation. Our findings demonstrate that 3D MSCs confer superior cytoprotection through autophagy activation, suggesting a novel therapeutic strategy for AMD management.

Methods

Isolation and culture of UC-MSCs

Human UC-MSCs were isolated and cultured using a standard protocol. Briefly, the umbilical cord from a parturient was thoroughly cleaned, and Wharton’s jelly was peeled off and cut into small pieces of approximately 1 mm². These pieces were then seeded into MSC culture medium containing 10% fetal bovine serum (FBS). After 48 h, non-adherent cells were removed, and adherent cells were expanded through multiple passages. The obtained cells were tested for MSC surface markers by flow cytometry, confirming high expression of CD44, CD90, and CD105, and no expression of the hematopoietic markers CD34 and CD45, consistent with the characteristics of mesenchymal cells (data not shown). To generate three-dimensional MSC spheroids, approximately 80–90% confluent UC-MSCs were harvested with trypsin and resuspended in complete medium. The cells were then seeded onto the lid of a cell culture dish at 30 µl per droplet. The lid was inverted onto the culture dish for hanging drop culture and incubated to promote self-assembly. Within 24 h, MSCs self-assembled into tight spherical clusters (3D MSCs). MSCs maintained in conventional adherent monolayer culture served as a two-dimensional control (2D MSCs). All experiments were performed using MSCs at passages 3 to 5.

In vivo experiments

Animals

The work has been reported in line with the ARRIVE guidelines 2.0. Eight-week-old female Kunming mice (22–24 g) were obtained from the Shandong Experimental Animal Center (Certification No.: SYXK 2020-0022). All animal procedures were approved by the Laboratory Animal Ethics and Welfare Committee of Qilu Hospital, Shandong University (Approval No.: DWLL-2022-173 Date: Sep 20, 2022) and adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Mice were housed in a specific pathogen-free environment under a 12-hour light/12-hour dark cycle, with ad libitum access to standard chow and water.

Sodium iodate-induced retinal degeneration

To induce RPE oxidative damage resembling dry AMD, mice were randomly assigned to five experimental groups (n = 10 per group): Control, NaIO₃-induced Model, 3D MSCs-treated (NaIO₃+3D MSCs -treated), 2D MSCs-treated (NaIO₃+2D MSCs -treated), and Vehicle control (NaIO₃+ vehicle treatment) groups. Except for the control group, all mice received a single tail-vein injection of 35 mg/kg NaIO₃ to induce retinal injury; control animals received an equivalent volume of PBS via tail vein [19]. 7 days post- NaIO₃ administration, 3D MSCs-treated, 2D MSCs-treated, and vehicle control groups underwent intravitreal injection in the right eye with their respective treatments. For the MSC-treated groups, approximately 2 × 10⁵ UC-MSCs suspended in 2 µL of PBS were injected into the vitreous. The Vehicle group received an equal volume of PBS intravitreally. The NaIO₃ model group received sham injection without fluid after the injury.

Electroretinography

14 days after the intravitreal treatments, retinal function was evaluated by flash electroretinography (FERG). To ensure unbiased analysis, ERG measurements were conducted by blinded observers unaware of treatment group assignments. Mice were dark-adapted overnight and then anesthetized by intraperitoneal injection of a ketamine/xylazine mixture. Pupils were dilated with topical 0.5% tropicamide and 0.5% phenylephrine. Each mouse was placed on a pre-warmed stage (37 °C) of a visual electrophysiology system (RetiMINER-C ERG, IRC, Chongqing, China). A ground electrode was attached to the tail base, a reference electrode to the cheek, and gold wire loop electrodes were placed on the corneas as active electrodes. During recordings, eyes were kept moist with sterile saline. Light stimuli of calibrated intensity were presented, and ERG tracings were recorded. The amplitudes of the a-wave and b-wave were measured, as well as the implicit times to these wave peaks, for each experimental group. The amplitudes of the a-wave and b-wave were measured, as well as the implicit times to these wave peaks, for each experimental group (n = 4 mice per group from the total n = 10 per group used in the study).

Histological analysis

Immediately after ERG recording, mice were humanely sacrificed using carbon dioxide. The mice were placed in a sealed chamber, and carbon dioxide was gradually introduced at a flow rate that displaced 30–70% of the chamber volume per minute. Death was confirmed by the absence of respiratory movement for at least 2 min, followed by cervical dislocation as a secondary method to ensure death. Then the eyes were enucleated for morphological and molecular analysis. Eyeballs were fixed in a dedicated eye fixation solution (#G1109, Wuhan, China), embedded in paraffin, and sectioned at 4 μm thickness through the central retina. Three non-consecutive sections per eye were analyzed to ensure representative sampling. Hematoxylin and eosin (H&E) staining was performed on retinal sections for general morphological evaluation. Images of the stained retinal sections were captured under a light microscope (OPTIKA IM-3, Italy). For quantitative histological analysis, the thickness of the outer nuclear layer (ONL) was measured at 0.4 mm from the optic disc (nasal and temporal sides) using ImageJ in a masked fashion. Measurements were taken at three positions per section (nasal, central, and temporal) to assess regional variability. RPE cell nuclei in each section were manually counted under 100× magnification to assess RPE cell loss.

Immunofluorescence analysis

Subsequently, for inflammatory marker detection, deparaffinized sections underwent heat-mediated antigen retrieval in citrate buffer (pH 6.0) for 20 min, then were blocked with 2% bovine serum albumin (BSA) for 1 h at room temperature. Sections were incubated overnight at 4 °C with primary antibodies against NLRP3 (rabbit anti-NLRP3, Abmart #T55651, 1:100) or caspase-1 (rabbit anti-caspase-1, Abmart #P79884-2R, 1:100). Negative control sections (omitting primary antibodies) were included to validate antibody specificity. After washing in PBS, sections were incubated with Alexa Fluor 488-conjugated goat anti-rabbit IgG secondary antibody (1:500) for 1 h at room temperature. Nuclei were counterstained with DAPI. Fluorescent images were acquired with an Olympus BX53F fluorescence microscope. The fluorescence intensity of NLRP3 and caspase-1 in the RPE layer (and other retinal layers) was quantified using ImageJ, and mean intensities were compared across groups.

In vitro experiments

ARPE-19 cell culture

The ARPE-19 cell line (a human RPE cell model) was obtained from a cell bank (Zhong Qiao Xin Zhou Biotechnology, Shanghai, China; Cat# ZQ0460). Cell line identity was verified by short tandem repeat profiling and matched to the DSMZ database. ARPE-19 cells were maintained in DMEM/F12 medium (Thermo Fisher #10565018) supplemented with 10% FBS (Thermo Fisher, Gibco) and 1% penicillin-streptomycin (Beyotime, China). Cultures were kept at 37 °C in a humidified 5% CO₂ incubator. Cells between passages 10 and 20 were used for experiments. Prior to experimental treatments, ARPE-19 cells were grown to 90–95% confluence and then serum-starved (0% FBS) for 24 h to synchronize the cells.

Co-culture and treatments

For co-culture experiments, ARPE-19 cells were exposed to oxidative injury with or without MSC support. ARPE-19 cells were seeded in the lower chamber of Transwell plates (Corning, 0.4 μm pore size inserts) and allowed to reach confluence. 3D MSCs or 2D MSCs were then added to the Transwell insert (upper chamber) in serum-free medium, so that ARPE-19 and MSCs shared media without direct cell–cell contact. Injury was induced by treating ARPE-19 monolayers with sodium iodate (NaIO₃, 10 mM) for 24 h. Control ARPE-19 cultures received no NaIO₃. In some co-culture experiments, additional modulators were used to probe the mechanism: 3-methyladenine (3-MA, 5 mM) was added to inhibit autophagy, and rapamycin (Rap, 100 nM) was added to stimulate autophagy. These agents were applied 1 h before NaIO₃ treatment and maintained throughout the 24-hour exposure.

Cell viability assessment

ARPE-19 cell viability was assessed using the Cell Counting Kit-8 (CCK-8, Abcam #ab228554) colorimetric assay. ARPE-19 cells were plated in 96-well plates at 1 × 10^4 cells per well and incubated overnight. For co-culture experiments, Transwell inserts containing MSCs were carefully removed prior to CCK-8 assay performance to ensure measurements exclusively reflected ARPE-19 viability in the lower chamber. After treatments (such as 24-hour exposure to NaIO₃ with or without MSC co-culture), 10 µL of CCK-8 reagent was added to each well and incubated at 37 °C for 2 h. Absorbance at 450 nm was measured using a microplate reader (Tecan Sunrise). Cell viability of each group was normalized to the untreated control group and expressed as a percentage. Each experiment included three technical replicate wells per condition, and all assays were independently repeated three times with distinct biological samples (n = 3 biological repeats).

Western blot

After experimental treatments, ARPE-19 cells were lysed in ice-cold RIPA buffer (Beyotime, with protease and phosphatase inhibitors) for protein extraction. Cell lysates were sonicated briefly on ice and clarified by centrifugation at 12,000×g for 20 min at 4 °C. Protein concentrations were determined using a BCA assay. Equal amounts of protein from each sample (20–30 µg) were separated by SDS-PAGE and transferred onto PVDF membranes. Membranes were blocked with 5% nonfat milk in TBS-T buffer and incubated overnight at 4 °C with primary antibodies against: NLRP3 (1:1000, Abcam, #T55651), caspase-1 (total and cleaved) (1:1000, Abcam, #P79884-2R), LC3B (microtubule-associated protein 1 light chain 3B, 1:1000, Abmart, #T55992F), p62 (1:1000, Abcam, #T55546F)pro-IL-1β (1:1000, Abcam, #TA5103), and GAPDH (1:5000, Cell Signaling Technology, #8884). After thorough washes, membranes were incubated with appropriate HRP-conjugated secondary antibodies at room temperature for 1 h. Protein bands were detected using an enhanced chemiluminescence (ECL) system and imaged. Densitometric analysis of band intensity was performed using ImageJ, and target protein levels were normalized to GADPH. Each experiment was repeated at least three times.

Annexin V-FITC/PI assay

Apoptosis was assessed using the Cell Apoptosis Detection Kit (#556547, BD Bioscience, San Jose, CA, USA) following the manufacturer’s instructions. ARPE-19 cells were detached using 0.25% trypsin, harvested by centrifugation at 300 × g for 5 min, and resuspended in binding buffer at a concentration of 1 × 10⁶ cells/mL. Cells were then incubated with Annexin V-FITC and propidium iodide (PI) (both at 1:20 dilution) in the dark for 15 min at room temperature. Flow cytometry analysis was performed using the Accuri C6 Plus cytometer (BD Biosciences, USA), and data were analyzed using FlowJo software (BD, USA).

ROS quantification

For ROS quantification, cells were washed and subsequently maintained in a phenol red-free medium containing 2.5 µM 2’,7’-dichlorodihydrofluorescein diacetate (Beyotime, #S0033S). The incubation was performed at 37 °C for 30 min in the dark. Fluorescence images were acquired using an Olympus BX53F microscope (Tokyo, Japan), and the mean fluorescence intensity was analyzed with ImageJ software (NIH).

RNA-seq data acquisition and differential expression analysis

RNA-seq data for human MSCs cultured under 2D and 3D conditions were obtained from the NCBI Gene Expression Omnibus (GEO) database (GSE185874). The dataset included MSC samples from nine independent donors cultured as 2D monolayers and nine donors cultured as 3D spheroids. Raw sequencing data were processed using standard quality control pipelines, and differential gene expression analysis was performed using DESeq2. Genes with an absolute log₂ fold-change (|log₂FC|) > 1 and a P-value < 0.05 were considered significantly differentially expressed.

Functional enrichment analysis

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed to assess the biological significance of differentially expressed genes (DEGs). Enrichment analyses were conducted using the clusterProfiler R package. Significantly enriched terms and pathways were defined based on an adjusted P-value < 0.05 (Benjamini–Hochberg correction).

Protein–Protein Interaction (PPI) network construction

To further investigate the functional relationships among DEGs enriched in autophagy-related pathways, a protein–protein interaction (PPI) network was constructed using the STRING database (https://string-db.org). DEGs mapped to the KEGG category “Autophagy – animal” were uploaded to STRING, and interactions with a confidence score > 0.7 were included. The resulting network was visualized and analyzed to identify key autophagy regulators and hub genes.

Statistical analysis

All experiments, both in vitro and in vivo, were independently repeated at least three times. Data are presented as mean ± SD and analyzed using GraphPad Prism version 8.0 software. Normality was assessed using the Shapiro-Wilk test. For normally distributed data, statistically significant differences were determined using one-way ANOVA and unpaired t-tests. For non-normally distributed data, statistically significant differences were assessed using Mann-Whitney U tests and the Kruskal-Wallis test, with a significance level set at p < 0.05.

Results

Characterization of UCB-MSCs and formation of 3D spheroids

UC-MSCs were isolated from umbilical cord tissue samples and rapidly grew after adhering to the surface of culture flasks (Fig. 1A), exhibiting a typical spindle-shaped, fibroblast-like morphology and adherent growth. Flow cytometry confirmed robust expression of mesenchymal stem cell markers: CD90 (100% positive), CD44 (99.74% positive), and CD105 (99.81% positive), while showing negligible expression of hematopoietic markers CD45 (0.10% negative) and CD34 (0.00% negative) (Fig. 1B). When cultured in non-adherent 6-well plates, the cells spontaneously formed spherical aggregates that progressively increased in size (Fig. 1C). These 3D MSCs spheroids were compact in structure and had smooth edges, indicating strong cell–cell adhesion. The average sphere diameters at 12 h, 24 h, and 48 h were 38.85 ± 28.35 μm, 53.42 ± 21.76 μm, and 95.43 ± 36.28 μm, respectively. At 48 h, the average number of cells per sphere was 0.35 ± 0.08 × 10³. Collectively, these findings confirm successful generation of UC-MSC-derived 3D spheroids with preserved mesenchymal stem cell characteristics, providing a solid foundation for subsequent functional studies.

Fig. 1.

Characterization of UC-MSCs and formation of 3D spheroids. A Light microscope images of UC-MSCs. B Flow cytometry analysis expression of CD44, CD90, CD105, CD45, and CD34. C Images of 3D-MSCs

3D MSCs treatment attenuates NaIO₃-induced retinal degeneration in vivo

To evaluate the protective effects of MSCs transplantation in vivo, we used a sodium iodate-induced retinal injury model. Figure 2 illustrates retinal morphology changes in each group as revealed by H&E staining. In the Control group (no NaIO₃), the retinal layers were well-organized and intact, with a regular monolayer of RPE cells visible. In contrast, NaIO₃-treated eyes (model group) showed notable retinal damage. There was disorganization and degeneration of cells in the ganglion cell layer (GCL) and a pronounced thinning of the outer nuclear layer (ONL) (Fig. 2A). Most strikingly, the RPE layer in NaIO₃-injured eyes exhibited signs of severe damage, with RPE cells appearing hypertrophic or completely lost in regions (Fig. 2B). These histological changes confirm that NaIO₃ injection induces significant retinal degeneration, particularly targeting the RPE and photoreceptor layers, consistent with an AMD-like pathology. Treatment with MSCs markedly protected the retinal structure from NaIO₃ toxicity. Eyes receiving an intravitreal injection of 3D MSCs showed much better preservation of the RPE layer and overall retinal architecture compared to eyes treated with 2D MSCs. While not achieving statistical significance, the data exhibited a favorable trend toward improved outcomes. In the 3D MSCs-treated group, the RPE remained largely intact (only minimal cell loss or atrophy), whereas the 2D MSCs-treated group showed some improvement over the Model but less RPE protection than 3D MSCs (RPE gaps and thinning were still observed). Quantification of ONL thickness and RPE cell counts (Fig. 2C) confirmed these observations: NaIO₃ caused a significant reduction in ONL thickness and RPE cell number relative to Control, and both MSC treatments mitigated these effects. However, the 3D MSCs-treated mice had a thicker ONL and a higher RPE cell count on average than the 2D MSCs-treated mice, indicating a superior protective effect. Together, these data suggest that 3D MSCs more effectively preserve retinal histology in the face of oxidative injury than conventional 2D MSCs.

Fig. 2.

The effect of MSCs on retinal morphology induced by NaIO₃. A Representative H&E-stained cross-sections of mouse retina from each group, highlighting pathological changes induced by NaIO₃ and the protective effects of MSC treatment. In the NaIO₃ Model, disorganization of the retinal layers and RPE disruption are evident. 2D MSCs treatment partly preserved retinal structure, while 3D MSCs treatment largely maintained normal retinal anatomy, including a continuous RPE layer. B, C Quantification of ONL thickness and RPE cell count per field for the different groups. Data are mean ± SD. One-way ANOVA with post-hoc test. n = 4 per group. GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer; RPE, retinal pigment epithelium

At the functional level, NaIO₃-induced damage led to significant visual impairment as measured by electroretinography. Figure 3 shows example ERG waveforms and summarized a-wave and b-wave amplitudes for each group 14 days after intravitreal injection. The untreated NaIO₃ Model group exhibited greatly diminished ERG responses compared to the Control group, indicating that NaIO₃ injection severely disrupted retinal function. In Model mice, the a-wave amplitude (originating from photoreceptor activity) and b-wave amplitude (reflecting inner retinal function) were both markedly reduced relative to Control, while the a-wave latency (time to reach a-wave peak) and b-wave implicit time were significantly prolonged (indicating slowed retinal response) compared to controls. Both MSC-treated groups showed improved ERG outcomes relative to the Model group. Notably, eyes treated with 3D MSCs demonstrated larger a- and b-wave amplitudes than those treated with 2D MSCs, approaching values seen in uninjured controls (Fig. 3, bar graphs). The implicit times in MSC-treated eyes were also closer to normal. These results indicate that MSCs transplantation preserved retinal electrophysiological function after NaIO₃-induced injury, with the 3D MSCs providing a greater degree of functional rescue than 2D MSCs.

Fig. 3.

The effect of MSCs on retinal function induced by NaIO₃. A Representative scotopic ERG traces from Control, NaIO₃ Model, 3D MSCs-treated, 2D MSC-treated, and Vehicle control mice. B, C, D and E Quantification of ERG a-wave and b-wave amplitudes (mean ± SD, n = 4 per group). Statistical significance indicated: *Model vs. Control, #MSC-treated vs. Model; one-way ANOVA with post-hoc test

To determine whether MSCs therapy reduced inflammasome activation in vivo, we performed immunofluorescence staining for NLRP3 and (pro)-caspase-1 in retinal sections (Fig. 4). In NaIO₃-injured eyes without treatment, there was a strong upregulation of NLRP3 and pro-caspase-1 immunoreactivity throughout the retina compared to controls, reflecting an activated inflammatory state. The increase was most prominent in the RPE layer, but elevated signals were also observed in the neural retina layers. Both 2D MSCs and 3D MSCs treatments markedly attenuated NLRP3 and caspase-1 signals in the retina. Quantification of fluorescence intensity indicated that MSC-treated eyes had significantly lower NLRP3 and caspase-1 levels than the untreated NaIO₃ Model. While both 2D MSCs and 3D MSCs reduced inflammasome markers, the 3D MSCs-treated group tended toward even lower levels, especially in the RPE, although the difference between 2D MSCs and 3D MSCs in vivo was not statistically dramatic for these markers. Nonetheless, the overall anti-inflammatory effect of MSC transplantation in vivo is evident.

Fig. 4.

Anti-inflammatory effects of MSCs treatments in vivo. A Representative images of immunofluorescence staining for (pro)-caspase-1 and NLRP3 respectively (red color) in the eye sections from different groups. Nuclei are counterstained with DAPI (blue). Green brackets indicated the location of ONL; white brackets indicated the location of INL; green asterisks indicated the location of RPE layer. B and C Quantitative mean fluorescence intensity data of the immunofluorescence staining results. Data were expressed as mean ± SD. Statistical significance indicated: * Model vs. Control, # MSC-treated vs. Model; Kruskal–Wallis test. Each dot represented an individual animal

2D MSCs and 3D MSCs co-culture protects ARPE-19 cells from NaIO₃-induced injury in vitro

We next explored the protective effect of MSCs on ARPE-19 survival under oxidative condition. To assess the cytotoxic effect of NaIO₃ in RPE cells, ARPE-19 were incubated with increasing concentrations of NaIO₃ (3, 5, 10, 15 or 20 mM) for 24 h. The viability of the RPE cells was evaluated using CCK8 cell viability assay. As illustrated in Fig. 5C, treatment with NaIO₃ resulted in a concentration-dependent reduction in RPE cell viability. Exposure to 10 mM NaIO₃ for 24 h induced approximately 50% of cell viability loss, whereas NaIO₃ induced about 83.22% reduction in cell viability at the concentration of 15 mM. Based on this dose-response curve, our study selected 10 mM as the working concentration for the in vitro oxidative damage model to maintain approximately 50% cell viability while inducing significant damage.

Fig. 5.

Effect of MSCs coculture on the apoptosis of RPE cells treated with NaIO₃. A Phase contrast light microscopic images of ARPE-19 cells. B Example images of flow cytometry results and corresponding quantitative data showing the effects of NaIO₃ and MSCs on ARPE-19. C Concentration-dependent effects of NaIO₃ on the viability of ARPE-19 cells assessed by CCK-8 assay. D Effects of NaIO₃ and MSCs on the viability of ARPE-19 cells assessed by CCK-8 assay. E Apoptosis rates evaluated through flow cytometry. Data were expressed as mean ± SD. * Model vs. Control, # MSC-treated vs. Model; one-way ANOVA with post-hoc test. Each dot represented an independent biological replication

We next examined the effects of MSCs on RPE cells under oxidative stress in vitro using an indirect Transwell co-culture system with ARPE-19 cells. ARPE-19 monolayers were exposed to NaIO₃ (10 mM, 24 h) to induce oxidative injury, and either no MSCs, 2D MSCs and 3D MSCs were co-cultured in the inserts. After 24 h of NaIO₃ exposure, ARPE-19 cell viability was severely compromised in the absence of MSCs, as evidenced by cell morphology and viability assays. Figure 5A shows phase-contrast images of ARPE-19 cells under different conditions. The Control ARPE-19 culture (no NaIO₃) reached full confluence with uniform cobblestone morphology. In stark contrast, ARPE-19 cultures treated with NaIO₃ alone had dramatically reduced cell density; many cells detached or died, and the remaining cells appeared enlarged and irregular in shape with evident cytoplasmic vacuoles and blebbing. Co-culture with MSCs noticeably improved ARPE-19 survival. In both 2D and 3D MSCs co-culture conditions, ARPE-19 cells were more numerous and had healthier morphology compared to NaIO₃ alone. Particularly in the 3D MSCs co-culture group, ARPE-19 cells maintained a morphology closer to normal, with clear cell borders and more regular arrangement, and the cell count was higher than in the 2D MSCs treared group.

To elucidate whether the protective effect of MSCs is associated with the alleviation of cell apoptosis, apoptosis was detected using an Annexin V-FITC/PI apoptosis detection kit and flow cytometry (Fig. 5B). This assay measured both early apoptotic cells (lower right square) and late apoptotic cells (upper right square). Annexin-V-positive cells which were shown to undergo apoptosis raised from 6.05% to 51.53% during treatment of NaIO₃. Coculturing with MSCs, the apoptosis rate of RPE cells decreased to 20.4% (for 3D MSCs) and 31.4% (for 2D MSCs). These results showed that MSCs exhibit the antiapoptotic effect on NaIO₃-triggered RPE.

To elucidate whether the protective effect of MSCs is associated with the suppression of oxidative stress, we evaluated intracellular ROS levels in ARPE-19 cells using fluorescence imaging and quantitative analysis (Figure S1). Exposure to NaIO₃ induced a marked increase in ROS production, with levels rising from 100% (control) to approximately 390%. Co-culture with MSCs significantly mitigated this elevation. These results demonstrate that MSCs, particularly when cultured in 3D, exert a potent antioxidant effect on NaIO₃-triggered RPE cells.

Gene expression profiling reveals enhanced autophagy pathway activity in 3D MSCs

To investigate potential mechanisms for the improved therapeutic effect of 3D MSCs, we analyzed global gene expression differences between 3D and 2D MSCs. We performed in silico analysis of an RNA-seq dataset (NCBI GEO: GSE185874) comprising human MSCs cultured in 2D (n = 9 donors) versus 3D spheroids (n = 9 donors). Differentially expressed genes (DEGs) were identified using a threshold of |log₂ fold-change| > 1 and P < 0.05. This yielded a total of 3,475 DEGs between 3D and 2D MSCs, of which 3,132 genes were upregulated and 343 were downregulated in 3D culture compared to 2D (Fig. 6A, B). This transcriptomic analysis highlights the broad extent of MSC reprogramming under 3D culture conditions.

Fig. 6.

Transcriptomic reprogramming and autophagy-related enrichment in 3D MSCs compared to 2D MSCs. A Volcano plot of DEGs between 3D and 2D MSCs. Red dots indicate significantly upregulated genes, and green dots indicate significantly downregulated genes. B Heatmap of hierarchical clustering of DEGs. C GO enrichment analysis of DEGs. D KEGG pathway enrichment analysis. E PPI network of genes in the “Autophagy–animal” KEGG category

To further elucidate the functional implications of these changes, we conducted GO and KEGG enrichment analyses. GO enrichment analysis of the DEGs indicated that the genes upregulated in 3D MSCs were significantly associated with cellular components and processes such as the plasma membrane region, intracellular vesicles, and cell–cell signaling. These outcomes suggest enhanced secretory function and intercellular communication in 3D MSCs (Fig. 6C). KEGG analysis demonstrated that multiple autophagy-related pathways were significantly enriched in 3D MSCs. In particular, the mTOR signaling pathway and the canonical autophagy pathway (KEGG category: Autophagy – animal) were prominently represented, implicating enhanced regulation of autophagic processes under 3D culture conditions (Fig. 6D).

To further investigate the interactions among genes enriched in the autophagy pathway, we constructed a protein–protein interaction (PPI) network based on the KEGG category “Autophagy – animal” using the STRING database (Fig. 6E). The resulting network revealed a highly interconnected module of autophagy regulators, including ULK1, ATG4B, ATG9A, LC3 (MAP1LC3B), and SQSTM1/p62, as well as signaling mediators such as AKT1, RRAGD, and RPS6KB2. This network highlights the central role of autophagy-related molecules in 3D MSCs, providing a framework for subsequent experimental validation.

3D MSCs activates autophagy and suppresses NLRP3 inflammasome activation in RPE cells

Given the indications from the gene analysis, we next examined autophagy and inflammasome activity in the RPE cells co-cultured with MSCs. We focused on key protein markers of the NLRP3 inflammasome pathway and the autophagy pathway in ARPE-19 cells after NaIO₃ injury, with or without 2D MSCs/3D MSCs support. Western blot analyses were performed for NLRP3, pro-caspase-1 and its cleaved p20 fragment (the activated form of caspase-1), and pro-IL-1β to assess inflammasome activation, as well as for LC3-I/II to assess autophagy-associated markers (Figs. 7 and 8).

Fig. 7.

MSCs suppressed NLRP3 inflammasome activation in ARPE-19 cells. A, B, C and D Representative western blot images and quantitative densitometry data showing that NaIO₃ (10 mM for 24 h) significantly increased the expression level of NLRP3 protein, while treatment with 3D MSCs inhibited this response. E, F, G and H Representative western blot images and quantitative densitometry data showing that 3-MA partially reversed the inhibitory effect of 3D MSCs on inflammasome activation. I, J, K, L and M Representative western blot images and quantitative densitometry data showing that rapamycin further suppressed NLRP3 and caspase-1 levels. Data were expressed as mean ± SD. * Model vs. Control, # MSC-treated vs. Model; a Drugs-treated vs. MSC. one-way ANOVA with post-hoc test. Each dot represented an independent biological replication

Fig. 8.

MSCs increased the level of autophagy-associated markers of RPE cells stimulated with NaIO₃. A Representative western blot images showing that NaIO₃ (10 mM for 24 h) significantly decreased the expression level of p62 and LC3-II/LC3-I, while treatment with MSCs inhibited this response. B and C Quantification of expressions of western blot results. Data were expressed as mean ± SD. * Model vs. Control, # MSC-treated vs. Model; a Drugs-treated vs. MSC. one-way ANOVA with post-hoc test. Each dot represented an independent biological replication

As shown in Fig. 7A, NaIO₃ exposure (10 mM, 24 h) markedly upregulated the protein levels of NLRP3 in ARPE-19 cells compared to untreated controls. This was accompanied by an increase in pro-caspase-1 and the appearance of the cleaved caspase-1 p20 subunit, indicating that the inflammasome was activated and processing caspase-1. Additionally, pro-IL-1β levels were elevated in NaIO₃-treated cells, consistent with an enhanced pro-inflammatory state. These results confirm that NaIO₃-induced oxidative stress can trigger NLRP3 inflammasome activation in RPE cells. Co-culture with MSCs attenuated these changes. In ARPE-19 cells co-cultured with 3D MSCs during NaIO₃ treatment, the levels of NLRP3, caspase-1 p20, and pro-IL-1β were all significantly lower than in cells treated with NaIO₃ alone (P < 0.05; see densitometry in Fig. 7B, C and D). 2D MSCs co-culture also reduced the expression of these inflammasome-related proteins, but the reduction was less pronounced than with 3D MSCs.

Notably, we probed the role of autophagy in this protective effect by using pharmacological modulators. Treatment with the autophagy inhibitor 3-MA partially reversed the inhibitory effect of 3D MSCs on inflammasome activation (in 3-MA treated samples, NLRP3 and cleaved caspase-1 levels were higher than with 3D MSCs alone) (Fig. 7E, F, G and H), whereas adding the autophagy activator rapamycin further suppressed NLRP3, pro caspase-1 and p20 levels (beyond what 3D MSCs alone achieved) (Fig. 7I, J, K, L and M). Although not all conditions are shown in the figure, these observations indicate that 3D MSCs’ suppression of inflammasome signaling in RPE cells is at least partly mediated through autophagy. In summary, 3D MSCs co-culture significantly blunted the NaIO₃-induced activation of the NLRP3 inflammasome pathway in RPE cells, correlating with its stronger cytoprotective effect.

NaIO₃ treatment altered autophagy-associated markers in ARPE-19 cells, including a reduced LC3-II/LC3-I ratio and decreased p62 abundance (Fig. 8). Co-culture with MSCs, particularly 3D MSCs, was associated with increased LC3-II/LC3-I and a relative increase in p62 compared with NaIO₃ alone, indicating that MSC treatment modulates autophagy-related marker expression under oxidative stress.

The inclusion of 3-MA in 3D MSCs co-culture negated the increase in LC3-II (as expected, since 3-MA blocks autophagosome formation), whereas rapamycin further boosted the LC3-II/LC3-I ratio in 3D MSCs co-culture. Thus, there is a strong correlation between autophagy activation and the cytoprotective, anti-inflammatory effects observed: conditions that enhanced autophagy (3D MSCs, rapamycin) yielded better protection and lower inflammasome activity, while blocking autophagy (3-MA) led to a loss of protection and higher inflammasome markers. These results support the notion that 3D MSCs protect RPE cells by activating autophagy, which in turn helps to limit oxidative damage and inflammasome activation. The autophagy substrate p62 exhibited a cumulative decrease upon NaIO₃ treatment, suggesting a potential of autophagy activity under oxidative stress. Co-culture with MSCs, particularly in 3D culture conditions, resulted in a relative restoration of p62 levels, further supporting the improvement of autophagic function.

Discussion

In this study, we investigated the efficacy of UC-MSCs in protecting RPE cells from oxidative stress and inflammation, using both an in vitro and an in vivo AMD-like models. Our findings demonstrated that MSCs treatment, particularly with MSCs cultured as 3D spheroids, significantly improved RPE cell viability and retinal function following NaIO₃-induced injury. Importantly, 3D MSCs showed stronger anti-oxidative, anti-inflammatory, and pro-autophagic effects than 2D MSCs. To our knowledge, this is the first study to directly demonstrate that 3D spheroid preconditioning substantially enhances the therapeutic efficacy of MSCs in AMD models, laying the groundwork for future clinical applications.

Chronic oxidative stress and low-grade inflammation, particularly through NLRP3 inflammasome activation, play central roles in AMD pathogenesis [20]. Consistent with this, our NaIO₃-induced model produced significant RPE cell loss in vitro and retinal degeneration in vivo, mirroring key features of dry AMD, including RPE cell loss and ONL thinning [19]. NaIO₃-exposed ARPE-19 cells also showed marked cell death and increased levels of ROS and a surge in NLRP3 inflammasome components.

Both 2D and 3D MSCs significantly reduced oxidative injury and inflammation in vivo and in vitro, with 3D MSCs showing superior efficacy. This enhanced protection is likely attributed to strengthened cell–cell interactions and augmented paracrine signaling under 3D spheroid culture, which is critical for tissue protection and regeneration. Our findings are in agreement with other reports that have demonstrated superior outcomes using MSC spheroids in various inflammatory and degenerative disease models, such as traumatic brain injury and intervertebral disc degeneration [21–23]. Furthermore, recent studies confirm that 3D MSCs exert stronger anti-inflammatory effects by better modulating inflammasome activity [24].

To elucidate the mechanisms underlying the superior protective action of 3D MSCs, we focused on the role of autophagy. Autophagy is a fundamental cellular homeostatic process [25], especially important in RPE cells which face a high burden of phagocytosing photoreceptor outer segments and managing oxidative stress [8]. Dysregulated autophagy leads to toxic protein accumulation and RPE degeneration [1]. Interestingly, autophagy has also been identified as a key factor in MSC biology. Recent studies have shown that 3D MSCs activates autophagy within the MSCs themselves, improving their survival and therapeutic capacity [26]. In our study, we observed that co-culture with 3D MSCs improved level of autophagy-associated markers in NaIO₃-injured RPE cells, suggesting that MSC-secreted factors stimulate autophagy in recipient cells. This restoration of autophagy likely contributes to reduced cellular stress and improved RPE survival. Because p62 levels can reflect both transcriptional regulation and autophagic degradation, p62 changes should be interpreted cautiously. Future work using lysosomal inhibitors or tandem fluorescent LC3 reporters would further strengthen conclusions regarding autophagic flux.

We further demonstrated that 3D MSCs suppress NLRP3 inflammasome activation through autophagy induction, revealing a key mechanism underlying their therapeutic efficacy. This aligns with reports showing that MSC-derived exosomes regulate neuroinflammation via the NRF2/NF-κB/NLRP3 pathway [27], highlighting the broader relevance of MSC-mediated inflammasome control. Our results expand on these findings, emphasizing the protective role of 3D MSCs against inflammasome-driven inflammatory damage in retinal cells. Our results indicate that the protective effects of 3D MSCs are causally linked to autophagy-mediated inhibition of NLRP3 inflammasome activation in RPE cells. By restoring autophagic flux, 3D MSCs prevent the accumulation of damaged mitochondria and other danger signals that would otherwise trigger inflammasome assembly. Importantly, autophagy inhibition with 3-MA attenuated the protective and anti-inflammatory benefits of 3D MSCs, while activation of autophagy with rapamycin further enhanced their effects. These findings provide strong mechanistic evidence supporting autophagy as a critical mediator of 3D MSCs-induced retinal protection.

Looking forward, several important questions remain. First, identifying the specific paracrine factors or exosomal components secreted by 3D MSCs that drive autophagy activation in RPE cells will be key. Second, while our study demonstrates significant benefit at the 14-day endpoint after intravitreal administration, AMD-like degeneration is chronic and may require sustained or repeated treatment. Our previous study has performed a time-course comparison of spheroid morphology, viability, and phenotypic characteristics between 2D and 3D MSCs, confirming the advantages and stability of the 3D culture system in optimizing the in vitro preparation of MSCs [28]. Because 3D spheroids may undergo structural changes after injection, future studies should incorporate longitudinal cell-tracking and repeated-dosing designs to determine persistence, biodistribution, and durability of therapeutic effects.

In conclusion, 3D MSCs offer significant therapeutic advantages over conventional 2D MSCs by alleviating oxidative and inflammatory injury in RPE cells through activation of autophagy and suppression of NLRP3 inflammasome signaling. This study provides preclinical evidence supporting 3D MSCs as a promising therapeutic strategy for AMD. Future studies should aim to validate these findings and translate 3D MSCs therapy into practical treatments for AMD and other degenerative retinal diseases.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Author contributions

Qian Xu and Mengyao Li contributed equally to this work and should be considered co-first authors.

Funding

This study was supported by the research grant from Shandong Provincial Natural Science Foundation (#ZR2025QC1712 for Mengyao Li) and Jinan Health High-Caliber Talent Project (for Mengyao Li).

Data availability

RNA-seq data for human MSCs cultured under 2D and 3D conditions were obtained from the NCBI Gene Expression Omnibus database (GSE185874). All other relevant data supporting the findings of this study are included in the manuscript.

Declarations

Ethics approval and consent to participate

Cell Line Ethics: The original source (Zhong Qiao Xin Zhou Biotechnology) has confirmed that there was initial ethical approval for collection of human cells, and that the donors had signed informed consent. Human Ethics: The human cells (UC-MSCs) used in this study were approved by the Ethics Committee of Qilu Hospital, Shandong University. Title: Exploration and Comparison of UC-MSCs in Treating Dry Age-Related Macular Degeneration. Approval number: KYLL-202008-111. Date of approval: Aug 31, 2020. All participants provided informed consent, and this study adheres to the principles of the Declaration of Helsinki. Animal Ethics: The animal experimental procedures were approved by the Laboratory Animal Ethical and Welfare Committee of Qilu Hospital, Shandong University, and this study was reported in compliance with the ARRIVE guidelines 2.0. Title: Exploration and Comparison of UC-MSCs in Treating Dry Age-Related Macular Degeneration by Improving Retinal Pigment Epithelial Cell Dysfunction. Approval number: DWLL-2022-17. Date of approval: Sep 20, 2022.

Competing interests

The authors declare no competing interests.