IL-23 Priming Enhances the Neuroprotective Effects of MSC-Derived Exosomes in Treating Retinal Degeneration

Abstract

Purpose: Neuroinflammation is a characteristic feature of neurodegenerative diseases. Mesenchymal stem cell-derived exosomes (MSC-exo) have shown neuroprotective effects through immunoregulation, but the therapeutic efficacy remains unsatisfactory. This study aims to enhance the neuroprotective capacity of MSC-exo through IL-23 priming for treating retinal degeneration in mice.

Methods: MSC were primed with IL-23 stimulation in vitro, and subsequently, exosomes (MSC-exo and IL-23-MSC-exo) were isolated and characterized. Two retinal degenerative disease models (NaIO₃-induced mice and rd10 mice) received intravitreal injections of these exosomes. The efficacy of exosomes was assessed by examining retinal structural and functional recovery. Furthermore, exosomal microRNA (miRNA) sequencing was conducted, and the effects of exosomes on the M1 and M2 microglial phenotype shift were evaluated.

Results: IL-23-primed MSC-derived exosomes (IL-23-MSC-exo) exhibited enhanced capability in protecting photoreceptor cells and retinal pigment epithelium (RPE) cells against degenerative damage and fostering the restoration of retinal neural function in both NaIO₃-induced retinal degeneration mice and rd10 mice when compared with MSC-exo. The exosomal miRNA suppression via Drosha knockdown in IL-23-primed MSC would abolish the neuroprotective role of IL-23-MSC-exo, highlighting the miRNA-dependent mechanism. Bioinformatic analysis, along with further in vivo biological studies, revealed that IL-23 priming induced a set of anti-inflammatory miRNAs in MSC-exo, prompting the transition of M1 to M2 microglial polarization.

Conclusions: IL-23 priming presents as a potential avenue for amplifying the immunomodulatory and neuroprotective effects of MSC-exo in treating retinal degeneration.

Retinal degenerative diseases encompass heterogeneous and multifactorial conditions, such as age-related macular degeneration (AMD), retinitis pigmentosa (RP), Stargardt's disease (STGD), and so on.¹ They have become the dominant causes of irreversible vision impairment and blindness worldwide.¹ The common pathogenesis of these diseases is RPE and/or photoreceptor dysfunction,² in which neuroinflammation plays a pivotal role.³ Therefore, inhibition of neuroinflammation has been a promising strategy for these diseases.⁴ Nevertheless, effective interventions to mitigate neuroinflammation in retinal degeneration remain elusive. 

Increasing evidence shows the MSC-based therapies provide great potential for protecting RPE and/or photoreceptors through nutritional support and immunomodulation via paracrine mechanisms.^1,5 The exosomes derived from MSC (MSC-exo) exhibit immunoregulatory functions similar to their parent cells.⁶ MSC-exo contain functional cargos, such as micro RNAs (miRNAs), mRNA, peptides, proteins, and lipids, capable of modulating the functionality of recipient cells.⁷ The distinct advantage of MSC-exo over MSC lies in their cell-free feature, their ability to easily cross biological barriers, and so on.⁸ With improved safety, enhanced targeting, and ease of production, MSC-exo presents a promising avenue for immunotherapy⁹ and holds tremendous potential for immunomodulation in various degenerative diseases. 

MSC-exo have been applied to various animal models of retinal neurodegeneration, including diabetic retinopathy (DR), glaucoma, RP, etc. In a streptozotocin-induced diabetic rat model, MSC-exo suppressed retinal inflammation via the miR-126/HMGB1 pathway.¹⁰ Mead et al. demonstrated that MSC-exo enhanced retinal ganglion cell survival in glaucoma models involving mice and rats.^11,12 Additionally, MSC-exo prevented MNU-induced photoreceptor apoptosis through the miR-21/Pdcd4 pathway.¹³ Besides, exosomes derived from neural progenitor cells and some other cells have also shown some neuroprotective effects in the retina.¹⁴ These findings underscore the potential of exosome-based therapies for retinal protection. 

It has been proposed that MSC would be regulated by an inflammatory niche in the injured environment, namely, MSC priming.¹⁵ This process could enhance their reparative and immunomodulatory functions and optimize the efficacy of MSC-based treatments.¹⁶ Diverse priming strategies have been explored, encompassing pro-inflammatory cytokines, growth factors, hypoxic conditions, pharmacological drugs, and bioactive molecules.¹⁶ In this study, we focused on IL-23, a pro-inflammatory cytokine associated with retinal inflammation,^17,18 as the stimulant for immune priming. The study aims to determine the neuroprotective effect of the exosomes derived from IL-23-primed MSC (IL-23-MSC-exo). 

The findings showed that IL-23 priming significantly augmented the neuroprotective effects of MSC-exo in both the NaIO₃-induced model and rd10 model. IL-23 pretreatment notably enhanced the role of MSC-exo in protecting the RPE and photoreceptors against oxidative stress. Mechanistically, IL-23-MSC-exo was specifically internalized by retinal microglia in vivo. IL-23 priming induced a distinct cluster of anti-inflammatory miRNAs within the exosomes, promoting the shift of M1 microglia to M2 by targeting TNF-a, CD86, Nos2, and IL-1β. These results presented an unrecognized therapeutic potential of IL-23-primed MSC-exo in retinal degenerative diseases. 

Methods

MSC Culture, Priming, Characterization, and Transfection

Murine bone marrow-derived MSC (MUBMX-01001; Cyagen Biosciences Inc., Santa Clara, CA, USA) were cultured in Dulbecco’s Modified Eagle Medium/Nutrient Mixture F-12 (DMEM/F-12) medium containing 10% exosome-free fetal bovine serum (FBS), penicillin (100 U/mL), and streptomycin (100 mg/mL). The MSC were passaged three to five times and the supernatants were collected for exosome extraction. At 70% to 80% confluence, the supernatant was changed to medium containing 20 ng/mL IL-23 protein (PeproTech). After 24 hours, the supernatants were collected for ultracentrifugation to obtain IL-23-MSC-exo. The MSCs were characterized by flow cytometry (positive for CD44 and CD29, negative for CD45 and CD11b) and adipogenic and chondrogenic differentiation assays (Supplementary Fig. S1). 

Small interfering RNA targeting Drosha (Drosha siRNA) and negative control (NC siRNA) were purchased from RiboBio Co. Ltd. (Guangzhou). After 48 hours of transfection, the supernatant of MSC was changed to medium containing 20 ng/mL IL-23. After 24 hours, the supernatants were collected for ultracentrifugation to obtain IL-23-MSC-siDrosha-exo/IL-23-MSC-siNC-exo. 

Exosome Isolation, Characterization, and Labeling

Exosomes were isolated as previously described.^19,20 Briefly, the exosomes were obtained by sequential centrifugation of supernatants at 200 g for 10 minutes, 2000 g for 20 minutes, 10,000 g for 30 minutes, and 100,000 g for 2 hours at 4°C. The pellets were washed twice with PBS and passed through the 0.22 µm filter. Exosome size and morphology were measured by transmission electron microscopy (TEM; FEI Tecnai 12, Philips, The Netherlands). Exosome diameter distribution and concentration were identified by the NanoSight Tracking Analysis (NTA; NanoSight NS300, Malvern, UK). Typical exosome markers, including CD63, CD9, CD81, and Grp94, were detected by Western blot analysis. To label the exosomes, PKH-67 (Sigma-Aldrich, St. Louis, MO, USA) dye staining was performed. 

Animals and Treatments

All animal procedures followed the guidelines of Animal Care and Use Committee of Zhongshan Ophthalmic Center (Ethics ID: 2018-096) and ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. For the NaIO₃ model, only male mice were used because female mice exhibit a higher susceptibility to NaIO3-induced damage. Tail vein injections of 20 mg/kg NaIO₃ were performed in 6- to 8-week-old male C57 BL/6J mice (GemPharmatech Co., Ltd., San Diego, CA, USA). Consistent with a previous report, we observed moderate RPE damage at day 7.²¹ Rd10 mice (female and male at 1:1) purchased from the Jackson Laboratory (Strain #:004297), were also utilized. 

NaIO₃-induced mice were intravitreally injected with MSC-exo or IL-23-MSC-exo (5 × 10¹⁰ particles/mL in 2 µL) just after NaIO₃ administration. After 7 days, the mice were euthanized with eyes enucleated for further investigations. Rd10 mice received intravitreal injections of MSC-exo or IL-23-MSC-exo (5 × 10¹⁰ particles/mL in 1 µL) at postnatal day (P)16, P19, and P22, and were euthanized at P25 for further investigations. PKH-67-labeled exosomes were injected intravitreally to evaluate in vivo uptake. 

Scotopic Electroretinogram Recording and Optical Coherence Tomography Analysis

An electroretinogram (ERG) was performed following the previous procedure.²² The amplitudes of a- and b-waves were measured using the Celeris-Diagnosys system. Optical coherence tomography (OCT) images were acquired using SPECTRALIS-OCT (Heidelberg, Germany) with a mouse objective lens. The scanning process and the ONL thickness analysis were performed as previously described.²² Six mice in each group were analyzed. 

Hematoxylin and Eosin Staining and Thickness Analysis

Hematoxylin and eosin (H&E) staining was carried out as previously described.²³ The outer nuclear layer (ONL) thickness was measured using Image J software. Six mice were analyzed in each group, and at least three sections were analyzed and averaged for each area of each eyeball. 

Immunofluorescence Staining

The immunofluorescence staining of cryosections and whole mounts was performed as previously described.²⁴ The primary antibodies used included anti-CD206 antibody (R&D Systems), anti-Arg1 antibody (Santa Cruz Biotechnology), anti-CD86 antibody (Cell Signaling Technology), anti-iNOS antibody (Cell Signaling Technology), anti-Iba1 antibody (Wako), and anti-Rhodopsin antibody (Santa Cruz Biotechnology). When the source of primary antibodies was the same, the multiple fluorescent immunohistochemical staining kit (Absin) was utilized. TUNEL staining (In Situ Cell Death Detection Kit, Fluorescein; Roche, Indianapolis, IN, USA) was performed according to the manufacturer's instructions. For whole mounts, at least six to nine areas from each whole mount were analyzed. For cryosections, at least three slides per eyeball were subjected to staining. Subsequently, at least three areas were selected from each slide for imaging and analysis. The sample size for most cryosections and whole mounts was three eyeballs, whereas six eyeballs were used for TUNEL staining. 

Western Blot Analysis

The Western blot process was conducted as previously mentioned.²⁴ The primary antibodies used included anti-CD63 antibody (Abcam), anti-CD9 antibody (Abcam), anti-CD81 antibody (Cell Signaling Technology), anti-GRP94 antibody (Santa Cruz Biotechnology), anti-CD206 antibody (R&D Systems), anti-CD86 antibody (Cell Signaling Technology), anti-Arg1 antibody (Abcam), anti-Drosha antibody (Cell Signaling Technology), and anti-IL-23 antibody (Santa Cruz Biotechnology). 

MiRNA Library Construction, Sequencing, and Bioinformatic Analysis

Three batches of IL-23-MSC-exo and MSC-exo were utilized for miRNA sequencing. Library preparation and miRNA sequencing were performed by NovelBio Co., Ltd. Low-quality reads were filtered out, and the data was mapped to miRBase version 21.0 and mouse genome (GRCm38, NCBI) using the BWA software. Differential expression analysis was performed using the DESeq2 R package, with miRNAs with |log2(FC)| ≥ 1 and P value ≤ 0.05 as threshold. Target genes of these miRNAs were predicted using TargetScan8.0. Gene ontology (GO) enrichment analysis of the target genes was implemented using the clusterProfiler R package. Significantly differentially expressed miRNAs were investigated for enrichment analysis using the miEAA tool. To construct a protein-protein interaction (PPI) network, the predicted target genes of upregulated miRNAs were submitted to the stringApp plugin of Cytoscape, followed by functional enrichment. A new network was formed from selected nodes constituting the immune system process pathway using Cytoscape Network Analyzer. The same process was performed for the predicted target genes of downregulated miRNAs to establish the other network. Two molecular networks were constructed based on the overlapping of M1 microglia-targeting miRNAs and upregulated differentially expressed miRNAs or M2 microglia-targeting miRNAs and downregulated differentially expressed miRNAs. 

Statistical Analysis

Statistical analyses were performed using GraphPad Prism software 8.0 (GraphPad Software Inc., La Jolla, CA, USA). The data were presented as mean ± standard error of the mean (SEM). Statistical analysis was conducted using 1-way ANOVA with Turkey's post hoc test, or a two-tailed Student’s t-test. The P values < 0.05 were considered significant. 

Results

Exosomes From IL-23-Primed MSC Exhibited More Potent Neuroprotective Effects in the NaIO₃-Induced Retinal Degeneration Model

IL-23-MSC-exo and MSC-exo were identified using TEM, Western blot, and NTA analysis. As shown in Supplementary Figure S2, IL-23-MSC-exo presented similar exosomal characteristics to MSC-exo. Although IL-23 priming did not affect MSC characterization markers (Supplementary Fig. S3A), it could increase IL-23 expression in MSC (Supplementary Fig. S3B). To determine the optimal dosage, various concentrations of IL-23-MSC-exo were injected intravitreally in NaIO₃-induced retinal degeneration mice (Supplementary Fig. S4), and 5 × 10¹⁰ particles/mL was selected as the lowest effective concentration for further study. 

OCT images in Figure 1A demonstrated that aberrant hyper-reflective retinal lesions were significantly reduced and ONL thicknesses were recovered after exosome injection compared with PBS controls. Particularly, the retinal structure in IL-23-MSC-exo group was mostly restored (see Fig. 1A). H&E staining revealed that exosome treatments, particularly IL-23-MSC-exo administration, preserved RPE cells, and ameliorated photoreceptor cells degeneration, with reduced abnormal pigment agglomerations in NaIO₃-induced mice (Fig. 1B). Notably, IL-23-MSC-exo treatment better rescued the neuroretinal function with increased amplitudes of both a-wave and b-wave in comparison to MSC-exo treatment (Fig. 1C). These results indicated that with IL-23 licensing, MSC-exo exhibited significant improvement of neuroprotective activity for treating NaIO₃-induced retinal degeneration mice. 

Figure 1.

 

IL-23 priming promoted MSC-exo to preserve the neuroretinal structure and function in NaIO₃-induced retinal degenerative mice. (A) Representative images of OCT scanning. Compared with C57 normal control (NC), the PBS-treated NaIO₃ model displayed obvious RPE layer discontinuity with high-reflective deposits (red arrows) and decreased ONL thickness (as indicated by the distance between the yellow lines). Upon the administration of exosomes, the RPE layer became continuous and flat, and the thickness of ONL was preserved, particularly evident in the IL-23-MSC-exo group (Scale bar = 200 µm, n = 6 mice, 1-way ANOVA and Tukey’s post hoc test). (B) H&E staining images revealed a decrease in the thickness of ONL and the accumulation of melanin in the RPE layer (yellow arrows) in PBS-treated NaIO₃-retina. Exosomes could significantly rescue the retina structure, especially the IL-23-MSC-exo. The ONL thickness of the central, mid-peripheral, and peripheral retinas was calculated and compared (Scale bar = 100 µm, n = 6 mice, 1-way ANOVA, and Tukey’s post hoc test). (C) ERG showed that intravitreal injection of IL-23-MSC-exo could significantly increase the amplitudes of a- and b- waves in comparison with the MSC-exo group (n = 6 mice, 1-way ANOVA, and Tukey’s post hoc test). GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer. All data are shown as mean ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001, ns: no significance.

MiRNAs in Exosomes Mediated the Neuroprotective Role of IL-23-MSC-exo

Exosomes usually carry various molecules to facilitate intercellular communications.⁷ Among these cargos, miRNAs can be easily transferred and participate in multiple biological processes, such as immune regulation, cellular survival, proliferation, differentiation, and death.^19,25,26 To find out whether miRNAs mediated the neuroprotective role of IL-23-MSC-exo, MSC were transfected with siRNA-Drosha. Previous studies have reported that Drosha is crucial for miRNA maturation and Drosha-knockdown MSC could generate the exosomes with global suppression of miRNA biogenesis.²⁷ The knockdown efficiency of Drosha in MSC was identified by Western blot (Supplementary Fig. S5). Then, we utilized IL-23-MSC-siDrosha-exo to treat NaIO₃-induced mice by intravitreal injection. Remarkably, IL-23-MSC-siDrosha-exo group (labeled as Drosha siRNA) showed more dome-shaped hyperreflective foci along the irregular RPE layer and reduced ONL thickness in OCT images in comparison to IL-23-MSC-siNC-exo group (labeled as NC siRNA in Fig. 2A). Consistently, H&E images exhibited the recovery of RPE and ONL layers was abrogated by IL-23-MSC-siDrosha-exo (Fig. 2B). ERG analysis demonstrated that IL-23-MSC-siDrosha-exo impaired the improvement of neuroretinal function by IL-23-MSC-exo (Fig. 2C). Thus, miRNA suppression by Drosha knockdown would abolish the preservative effects of IL-23-MSC-exo in NaIO₃-induced retinal degeneration mice. 

Figure 2.

 

IL-23-MSC-exo exerted the neuroprotective effect in a miRNAs-dependent manner. (A) The OCT images of NaIO₃ mice treated with IL-23-MSC-siDrosha-exo (Drosha siRNA group) showed more hyper-reflective foci (red arrows) along the RPE layer and thinner thickness of ONL (distance between the yellow lines) than those mice treated with IL-23-MSC-siNC-exo (negative control NC siRNA group) (Scale bar = 200 µm, n = 6 mice, unpaired Student's t-test). (B) H&E staining of retinal sections showed that Drosha siRNA group presented more melanin deposited (yellow arrows) along RPE layer and also less ONL thickness (Scale bar = 100 µm, n = 6 mice, unpaired Student's t-test). (C) ERG showed decreased amplitudes of a- and b- waves after IL-23-MSC-siDrosha-exo treatment compared with IL-23-MSC-siNC-exo treatment (n = 6 mice, unpaired Student's t-test). All data are shown as mean ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001, ns: no significance.

Exosomal miRNA Sequencing Revealed a Set of Anti-Inflammatory MiRNAs in IL-23-MSC-exo Involved in the Microglial M1/M2 Polarization Process

Because miRNAs mediated the neuroprotective role of IL-23-MSC-exo, we further performed miRNA sequencing from IL-23-MSC-exo and MSC-exo. A total of 117 differentially expressed miRNAs were identified between IL-23-MSC-exo and MSC-exo (Fig. 3A). There were 6116 putative target genes of these miRNAs that were predicted using TargetScan8.0. GO enrichment analysis showed that pathways related to neuron death, regulation of immune system process, response to oxidative stress, etc., were involved in the therapeutic process of IL-23-MSC-exo (Fig. 3B). Given the importance of immune responses in retinal degenerative diseases,³ we performed further GO analysis in the term of regulation of immune system process. Various inflammation-related terms, such as macrophage activation and microglial cell activation, contributed to the IL-23-MSC-exo function (Fig. 3C). GO pathway enrichment was also performed using miEAA to analyze the 117 differentially expressed miRNAs. Terms relevant to the immune system process included microglial cell activation, inflammatory response, positive regulation of IL-6 production, and other pathways (Fig. 3D). Furthermore, the Cytoscape's plug-in stringApp was used to perform the functional enrichment analysis of the predicted target genes of upregulated miRNAs (Fig. 3E) or downregulated miRNAs (Fig. 3F) to explore the potential relationships between the target genes. Based on the genes enriched in the “immune system process,” two new PPI networks were constructed with node degree as the filter standard. The PPI network highlighted the upregulated M1 microglial markers (see Fig. 3E), whereas the other emphasized downregulated M2 markers (see Fig. 3F), suggesting modulation of microglial M1/M2 polarization by miRNAs from IL-23-MSC-exo. Then, the putative miRNAs targeting M1 or M2 markers were analyzed by TargetScan8.0 to find out the potential miRNAs (Figs. 3G, 3H). Together, these bioinformatic analyses indicated that a set of anti-inflammatory miRNAs in IL-23-MSC-exo may contribute to the regulation of M1/M2 microglia. 

Figure 3.

 

Exosomal miRNA profile and analysis. (A) miRNA sequencing was performed on IL-23-MSC-exo and MSC-exo (n = 3 for each group). Volcano plot [x-axis = log2 (fold change); y-axis = −log10 (P value)] showing 117 significantly differentially expressed miRNAs between IL-23-MSC-exo and MSC-exo. Red dots indicate the upregulated miRNAs in IL-23-MSC-exo; and the blue dots indicate the downregulated miRNAs in IL-23-MSC-exo with fold change >1. (B) Putative target genes of differentially expressed miRNAs were analyzed for GO biological process enrichment (P < 0.05). The enriched pathways included neuron death, regulation of immune system process and other biological processes. (C) Selected pathways from the GO biological process enrichment of the predicted target genes belong to immune process system (P < 0.05), which included macrophage/microglia cell activation. (D) The miRNA enrichment analysis was performed using the experimentally validated miRNA Enrichment Analysis and Annotation Tool (miEAA). The miRNAs enriched in IL-23-MSC-exo were related to microglia cell activation and inflammatory response. (E) Cytoscape and plug-in StringApp were used to construct the protein-protein interaction (PPI) network ground on the predicted genes of up-regulated miRNAs. The hub genes of this network included M1 markers CD86, IL-1β, and IFN-y. (F) The similar process was performed for the target genes of downregulated miRNAs to establish the other PPI network. The core genes of this network included M2 markers IL-4 and IL-10. (G) The molecular network was established to show the putative miRNAs targeting M1 microglial markers (IL-1β, TNF-α, CD86, and Nos2) among the upregulated miRNAs. (H) The molecular network showed the putative miRNAs targeting M2 microglial markers (Arg1, IL-4, and IL-10) among the downregulated miRNAs.

IL-23-MSC-exo Effectively Induced Microglial Polarization From M1 to M2 via MiRNA Transport

Next, we evaluated that whether IL-23-MSC-exo was internalized by microglia and augmented M1 to M2 microglial polarization in NaIO₃-induced retinal degeneration mice. Exosomes labeled with PKH67 were injected intravitreally. PBS with PKH67 dye served as control. Immunofluorescence images of retinal whole-mounts revealed that both MSC-exo and IL-23-MSC-exo may be captured and internalized by microglia (Fig. 4A). On cryosections, there were more Arg1⁺Iba-1⁺ cells and fewer iNOS⁺/CD86⁺Iba-1⁺ cells in IL-23-MSC-exo treatment group in comparison to MSC-exo group (Figs. 4B, 4C, 4D). Western blot results showed that NaIO₃ induction would increase CD86 expression and decrease CD206 expression, which was reversed by exosome treatment, particularly in IL-23-MSC-exo group (Fig. 4E). Together, these results suggested that IL-23 pretreatment facilitated a greater transition of the M1 pro-inflammatory to M2 anti-inflammatory microglia in retinal degeneration mice. 

Figure 4.

 

IL-23-MSC-exo promoted microglia polarization from M1 to M2 phenotype via miRNAs transport in the NaIO₃-induced mice. (A) Representative confocal images showed that the PKH67-labeled IL-23-MSC-exo or MSC-exo (green) were co-located with Iba-1⁺ microglia (red), indicating the exosomes were internalized by microglia (Scale bar = 50 µm, n = 3 retinas). (B–D) Intravitreal injection of exosomes increased the Arg1⁺Iba-1⁺ microglia and decreased the iNOS⁺/CD86⁺Iba-1⁺microglia, particularly evident in the IL-23-MSC-exo treatment (Scale bar = 50 µm, n = 3 eyeballs). (E) Application of exosome especially the IL-23-MSC-exo could significantly upregulate protein expression of CD206 and downregulate CD86 (n = 3 retinas, 1-way ANOVA and Tukey's post hoc test). (F) IL-23-MSC-siDrosha-exo suppressed the expressions of CD206 and Arg1 in the NaIO₃-retinas (n = 3 retinas, unpaired Student's t test). All data are shown as mean ± SEM, **P < 0.01, ***P < 0.001. (G–I) Intravitreal injection of IL-23-MSC-siDrosha-exo decreased Arg1⁺Iba-1⁺ as well as CD206⁺Iba-1⁺ microglia and increased CD86⁺ Iba-1⁺ microglia (n = 3 eyeballs, Scale bar = 50 µm).

Furthermore, we examined the microglial M1/M2 phenotype after miRNA suppression in IL-23-MSC-exo by Drosha knockdown. The miRNA suppression in IL-23-MSC-exo (IL-23-MSC-siDrosha-exo group) significantly reduced retinal levels of CD206 and Arg1 (Fig. 4F). Immunofluorescence images revealed a decrease in Arg1⁺/CD206⁺Iba-1⁺ M2 microglia and an increase in CD86⁺Iba-1⁺ M1 microglia in the IL-23-MSC-siDrosha-exo treatment group (Figs. 4G, 4H, 4I). Together, IL-23-MSC-exo may improve the M1 to M2 microglia in a miRNA-dependent manner. 

IL-23-MSC-exo Could Also Protect Photoreceptors Through Promoting Microglial M2 Shift in Spontaneous Retinal Degeneration Model of rd10 Mice

We also evaluated the therapeutic effects of IL-23-MSC-exo in rd10 mice. H&E staining images showed that IL-23-MSC-exo remarkably counteracted the substantial ONL loss in PBS-rd10 mice compared with MSC-exo (Fig. 5A). They also significantly reduced the number of TUNEL+ cells in ONL layer and increased the rhodopsin expression (Fig. 5B), indicating that IL-23-MSC-exo could protect the photoreceptors against death in rd10 mice. 

Figure 5.

 

IL-23-MSC-exo also protected rd10 mice against retinal degeneration. (A) The H&E staining images showed that intravitreal injection of IL-23-MSC-exo preserved the photoreceptors in the ONL layer and increased the thickness of ONL layer in the rd10 mice (Scale bar = 100 µm, n = 4 mice, 1-way ANOVA, and Tukey’s post hoc test). (B) The rhodopsin was highly preserved in the IL-23-MSC-exo-treated rd10 retinae, whereas the least Tunel+ cells in the ONL layer in comparison to PBS-treated or MSC-exo-treated rd10 mice (Scale bar = 50 µm, n = 6 eyeballs, 1-way ANOVA, and Tukey’s post hoc test). Data are shown as mean ± SEM, *P < 0.05, **P < 0.01, ns: no significance.

Next, both MSC-exo and IL-23-MSC-exo could be captured and internalized by microglia in rd10 mice (Fig. 6A). Subsequent analysis of CD206 and Arg1 expressions after exosome treatment showed that the protein levels were the highest in IL-23-MSC-exo treatment group (Fig. 6B). Immunofluorescence images revealed that IL-23-MSC-exo treatment could induce CD206⁺Iba-1⁺ M2 microglia (Fig. 6C). Furthermore, miRNA suppression in IL-23-MSC-exo (Drosha siRNA group) remarkably reduced CD206 and Arg1 (Fig. 6D), indicating the requirement of miRNA in microglial M2 induction by IL-23-MSC-exo. 

Figure 6.

 

IL-23-MSC-exo induced microglia to adopt M2 phenotype via miRNAs transport in rd10 mice. (A) Representative confocal images depicted the co-localization of PKH67-labeled IL-23-MSC-exo or MSC-exo (green) with Iba-1+ microglia (red), suggesting the internalization of exosomes by microglia (n = 3 retinas, Scale bar = 50 µm). (B) Western blot shows that exosome treatment could significantly increase the protein levels of CD206 and Arg1 in rd10 mice, especially in the IL-23-MSC-exo group (n = 3 retinas, 1-way ANOVA and Tukey's post hoc test). (C) Intravitreal injection of exosomes promoted the CD206 expression on the Iba-1⁺ microglia, especially the IL-23-MSC-exo injection (n = 3 eyeballs, Scale bar = 50 µm). (D) Intravitreal injection of IL-23-MSC-siDrosha-exo decreased the protein levels of CD206 and Arg1 compared with IL-23-MSC-siNC-exo in rd10 mice (n = 3 retinas, unpaired Student's t test). Data are shown as mean ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001.

Discussion

Neuroinflammation is a hallmark of various neurodegenerative diseases.²⁸ MSC-exo could exert neuroprotective effects via immunomodulation, presenting as a potential therapeutic approach for neurodegenerative diseases.²⁹ However, they still need to overcome the hurdles to improve their efficacy, which is crucial for realizing the full therapeutic potential of MSC-exo in clinical settings. Our study showed that after IL-23 priming, the exosomes from MSC possessed more potent neuroprotective function for treating retinal degenerative diseases. IL-23-MSC-exo induced increasing retinal microglia to polarize from M1 to M2 phenotype and orchestrated a more anti-inflammatory condition for photoreceptors and RPE cells recovery via a miRNA-dependent manner. This MSC priming approach by IL-23 stimulation was first reported and the IL-23-MSC-exo presented a promising therapy with enhanced efficacy for retinal degenerative diseases. The comprehensive understanding of their mode of action would optimize their therapeutic applications. 

Many studies have been conducted to enhance the therapeutic potential of MSC-exo, encompassing approaches such as genetic engineering or surface modification to improve targeting,^30,31 electroporation for loading with therapeutic cargo,³² scale-up production to optimize dosage,³³ and so on. The bioactive molecules carried by MSC-exo, such as miRNAs, proteins, and lipids, play a vital role in mediating therapeutic effects.⁷ The composition and abundance of cargo can vary depending on donor MSC and culture conditions, influencing the therapeutic outcome.³⁴ Here, we found exosomes derived from IL-23-primed MSC carried an altered cargo enriched with a set of anti-inflammatory miRNAs, leading to enhanced immunomodulatory properties for the treatment of retinal degenerative diseases. This is distinct from methods that enhance the secretion of exosomes by some mechanical stimulation of MSC.³⁵ As we know, other inflammatory factors, like TNF-α or IL-1β, have been attempted for immune priming for MSC.^36,37 Here, we focus on the IL-23, a pro-inflammatory cytokine implicated in tissue repair and regeneration processes.³⁸ The cargo compositions of the MSC-exo after different inflammatory cytokines priming are varied. IL-23-MSC-exo were characterized by the enrichment of a set of anti-inflammatory microRNAs including miR-493-3p, miR-3154, miR130b-3p, etc., which were involved in the regulation of M1/M2 microglia. The effects of IL-23-MSC-exo may be attributed to a network of miRNAs rather than a single one. Actually, miR-493-3p has been reported to suppress macrophage recruitment by targeting macrophage inhibitory factor.³⁹ The miR-130b-3p could induce M2 macrophage through the PTEN/PI3K/Akt pathway.⁴⁰ Furthermore, clarifying this microRNA interaction network contributes to a better understanding of the benefits of IL-23 pretreatment and facilitates the optimization of IL-23 priming strategy. 

During retinal degenerating in rd10 mice, microglia inclined to be activated and migrate toward damaged photoreceptors, adopting a dual M1/M2 phenotype with a predominance of the M1 phenotype.^41,42 Approaches for modulating the M1 and M2 switch of microglia have been recognized as effective strategies for treating neuroinflammation-mediated neurodegenerative changes.⁴³ Here, IL-23-MSC-exo showed potential to modulate the M1 and M2 switch of retinal microglia, presenting as a promising approach for inhibiting neuroinflammation, thereby alleviating retinal degeneration. Although the M1/M2 categorization might oversimplify the complexity and dynamics of microglial activation states,⁴⁴ IL-23-MSC-exo treatment overall promotes microglia to adopt an anti-inflammatory and tissue-repairing state. For rd10 mice, the potential impact of gender should be considered, because the onset of the disease occurred earlier in female rd10 mice compared to their male counterparts, indicated by a lower b-wave amplitude in female mice at P18. However, the b-wave amplitudes were nearly identical by P25.⁴⁵ Here, no significant differences were observed between female and male mice within the same exosome treatment group at P25, indicating sex-specific differences in the effects of exosomes in rd10 mice are not noticeable. 

In this study, we applied different intervention strategies to the NaIO3-induced and rd10 mouse models respectively to elucidate the therapeutic effects of IL-23-MSC-exo on retinal degenerative diseases. The NaIO₃-induced mouse model represents an acute form of retinal degeneration, for which a single intravitreal injection is typically administered and usually obtains a definitive effect.^22,46 For the rd10 mouse model, these mice carry a spontaneous mutation of the rod-phosphodiesterase (PDE) gene, which initiates progressive rod degeneration starting around P18, followed by subsequent cone loss. The peak of rod cell death occurs around P25.⁴⁷ Taking into account the exosome degradation and disease progression, rd10 mice received multiple injections at three critical time points to evaluate the therapeutic potential of exosomes: P16, before the onset of rod degeneration, and subsequently at P19 and P22. These different strategies were based on the time course of pathological progression, which were consistent with previous studies.^22,48,49 Further exploration of the different injection strategies is important for improving the therapeutic potential. 

Although we found that IL-23 priming enhances the neuroprotective effects of MSC-exo by inducing a set of anti-inflammatory miRNAs and facilitating the transition from M1 to M2 microglial polarization, it is of interest to explore extensively the exact mechanism through which IL-23 primes MSC and alters the composition of their exosomes. Even though both the IL-23-MSC-exo and MSC-exo were isolated from culture media containing the same batch of exosome-free FBS, and all other conditions were identical except for IL-23 priming step between the 2 groups, ensuring the comparability, we still cannot rule out the influence of other FBS components on the results. In addition, because the sample size for some assays was three in this study, further increasing the sample size could strengthen our findings. 

In summary, our study proposed a novel approach for immune priming of MSC-derived exosomes. By pretreating MSC with IL-23, the resulting exosomes demonstrated enhanced efficacy in promoting neuroretinal recovery and treating retinal degenerative diseases. IL-23-MSC-exo exerted potent immunomodulatory and neuroprotective effects by delivering a set of anti-inflammatory miRNAs that modulate the M1/M2 phenotype switch of microglia. This MSC immune priming strategy for improving efficacy of MSC-exo by using IL-23 stimulation holds great promise for advancing their therapeutic application, especially in neuroinflammation-mediated neurodegenerative diseases. 

Acknowledgments

Supported by National Natural Science Foundation of China (No. 82322016, 82171068, and 82271095), and Guangdong Provincial Key Area R&D Program (No. 2023B1111050004). 

Disclosure: H. Zhou, None; Y. Liu, None; T. Zhou, None; Z. Yang, None; B. Ni, None; Y. Zhou, None; H. Xu, None; X. Lin, None; S. Lin, None; C. He, None; X. Liu, None