Catalytic neural stem cell exosomes for multi-stage targeting and synergistical therapy of retinal ischemia-reperfusion injury

Abstract

Neuronal damage of the retina is a leading cause of visual impairment in patients with retinal ischemia-reperfusion injury (RIRI). Building on our clinical and experimental findings, the substantial decrease in catalase activity correlates with increased hydrogen peroxide (H₂O₂)-mediated oxidative stress that is primarily localized to the outer nuclear layer (ONL) situated in the posterior segment of the retina. Accordingly, we design a neural stem cell exosome with polylysine (K₁₀) decoration and catalase expression, named CataKNexo, which reaches the ONL and exerts synergistic antioxidant and neuroprotective therapy. Utilizing an in vitro retinal model recapitulating the layered architecture of the retina, we confirm that CataKNexo reaches the ONL through K₁₀-mediated transcytosis. In RIRI model mice, CataKNexo prevents the retina from H₂O₂-induced cell death, exerts neuroprotection, and restores vision function to near-normal levels. Moreover, CataKNexo shows promising antioxidative, neuroprotective, and safety profiles in RIRI model Bama miniature pigs, highlighting its potential for clinical translation.

Keywords: catalytic antioxidants; exosomes; retinal ischemia-reperfusion injury; synergistical therapy.

Graphical abstract

Figure 1

Aberrant CAT activity and H₂O₂ accumulation in the retinal ONL are prominent pathological features of RIRI (A) Schematic for extraction of the vitreous humor and subsequent analyses of antioxidase activities (SOD, CAT, and GPX) and ROS accumulation (O₂⁻, H₂O₂, and LPO) in the RVO group (50 patients with retinal vein occlusion, a typical RIRI disease) and control group (50 patients with idiopathic macular epiretinal membrane or idiopathic macular hole). (B) Measurement of antioxidant activities and ROS concentrations in the vitreous humor using Micro Antioxidase Activity Assay Kits and ROS Content Assay Kits. Control: n = 50 (1 eye from each of 50 patients); RVO: n = 50 (1 eye from each of 50 patients). (C) Ratios of mean antioxidase activities and ROS concentrations between the RVO and control group. (D) Spearman’s rank correlation analyses examining relationships among antioxidase activities, ROS concentrations, and BCVA in patients with RVO (n = 50). Each red dot represents a vitreous humor sample extracted from an individual patient’s eye. The XY plane (orange dots) represents the negative correlation values from an analysis between the antioxidase activity and ROS concentration. The XZ plane (cyan dots) represents the positive correlation values between ROS and BCVA. The YZ plane (purple dots) represents the negative correlation values between antioxidase activity and BCVA. (E) Correlation coefficients of BCVA with antioxidase activities and ROS concentrations in the RVO group. (F) Schematic for the extraction of retina and subsequent assays for antioxidase activities and ROS accumulation, as well as the accumulation of H₂O₂ in sham and RIRI model mice (sham: only the needle was inserted into the anterior chamber without saline infusion; RIRI model: retinal ischemia was induced by the instillation of saline into the anterior chamber of mice for 1 h, and then the high pressure was removed to restore perfusion of the retina). (G) Measurement of antioxidase activities (SOD, CAT, and GPX) and ROS concentrations (O₂⁻, H₂O₂, and LPO) in retinal samples of mice (left). Ratios of the mean antioxidase activities and mean ROS concentrations between the sham and RIRI model mice (right) (n = 6, 1 eye from each of 6 mice). (H) Representative images of CAT immunostaining and quantitative analysis of the expression of CAT in histological sections of retinas from RIRI model mice (n = 3, 1 eye from each of 3 mice). Scale bar: 50 μm. (I) Representative image from a two-photon fluorescence microscopy analysis of H₂O₂ and quantitative analysis of H₂O₂ accumulation in the GCL, INL, and ONL of retinas from RIRI model mice (n = 3, 1 eye from each of 3 mice). Scale bar: 50 μm. (J) Spearman’s rank correlation analyses examining relationships between CAT expression and H₂O₂ accumulation in the GCL, INL, and ONL of retinas from RIRI model mice (n = 3, 1 eye from each of 3 mice). (K) Quantitative analysis of the extent of cell injury in three retinal layers from RIRI model mice at the indicated time points (n = 3 biologically independent experiments). Data in (B) are presented as medians (25th–75th quartiles) and were compared using Mann-Whitney U test. Data in (G), (H), (I), and (K) are mean ± SD and were compared using one-way ANOVA. The experiments in (H) and (I) were independently repeated three times with similar results. See also Figure S1.

Figure 2

Construction and characterization of CataKNexo (A) Schematic of the construction of CataKNexo and the combination of scavenging H₂O₂ with repairing cell damage. CataNexo are derived from mNSC (mouse neural stem cells) transfected with CAT-expressing lentiviral vectors (pLVX-CD63-CAT) and subsequently conjugated with K₁₀ on the membrane. Among them, CD63 fusion protein can ensure the expression of CAT on the membrane. Wherein, CataKNexo collaboratively reduce the H₂O₂ accumulation and oxidative stress injury based on the effect of CAT and neurotrophic factors and functional miRNAs. (B) TEM image of CataKNexo stained with uranyl acetate. Scale bar: 200 nm. (C) Nanoparticle tracking analysis (NTA) of CataKNexo. (D) CLSM images of CataKNexo showing co-localization of CAT (Cy3 labeled, red) and K₁₀ (FITC labeled, green). The insets are STED images of the outlined region in the main images. The images (scale bar: 10 μm) and insets on the right (scale bar: 100 nm) are presented with same magnification, respectively. (E) Immunoblotting of Nexo (1), CataNexo (2), and CataKNexo (3) with β-actin as control. Alix, CD63, and TSG101 are exosome markers. SOX2 and Nestin are NSC exosome markers. CAT is an expressed protein of CataNexo and CataKNexo. (F) CAT activities of Nexo, CataNexo, and CataKNexo using a Micro CAT Assay Kit (n = 3 biologically independent experiments). (G) H₂O₂ scavenging efficiency of CataNexo and CataKNexo compared to commercial CAT (EC 1.11.1.6) (n = 3 biologically independent experiments). (H) Zeta potentials of Nexo, CataNexo, and CataKNexo (n = 3 biologically independent experiments). (I) Agarose gel electrophoresis image of Nexo, CataNexo, and CataKNexo. (J) Proteomic analysis of fresh and Lyo/Reh CataKNexo. The abundance of characteristic proteins associated with exosomes, NSC, chemokine receptor, neuroprotection, and CAT was analyzed. (K) miRNA sequencing of fresh and Lyo/Reh CataKNexo. Representative miRNAs related to antioxidants, neuroprotection, and anti-apoptosis were analyzed. Data in (F), (G), and (H) are mean ± SD and were compared using one-way ANOVA (F, H). The experiments in (B), (D), (E), and (I) were independently repeated three times with similar results. See also Figure S2.

Figure 3

In vivo evaluation for the retention and multi-stage targeting of CataKNexo in RIRI model mice (A) Schematic for evaluation of the intraocular retention and multi-stage targeting (Nexo, CataNexo, and CataKNexo) in RIRI model mice. (B) Representative fluorescence images and quantitative analysis of the intraocular retention at the indicated time points after single intravitreal injection of Nexo, CataNexo, or CataKNexo (n = 3, 1 eye from each of 3 mice). (C) Representative fluorescence images and quantitative analysis of excised eyeballs at 24 h after treated with various exosomes (n = 3, 1 eye from each of 3 mice, posterior: the right area of the dotted line represents the ocular structure behind the lens, anterior: the left area of the dotted line represents the ocular structure in front of the lens). (D) Representative fluorescence images by CLSM and statistical analysis of the targeting in different groups (n = 3, 1 eye from each of 3 mice). Scale bar: 50 μm. The images are presented at the same magnification. Data in (B) and (C) are mean ± SD and were compared using one-way ANOVA (C). The experiments in (B)–(D) were independently repeated three times with similar results. See also Figure S3.

Figure 4

In vitro investigation of the penetration of CataKNexo into distinct cell layers of a simulated ocular structure (A) Experimental design schematic for evaluating the penetration of CataKNexo using an in vitro model. In this iterative Transwell coculture system, retinal ganglion cells (RGCs, representing the GCL) were cultured in the top layer, Müller cells (representing the INL) were cultured in the middle layer, and 661W cells (representing the ONL) were cultured in the bottom layer. (B) Representative CLSM images and analysis of the distribution after treatment with the indicated exosomes. The images of XZ (scale bar: 50 μm) and XY (scale bar: 100 μm) planes are presented with same magnification, respectively. Cyan: phalloidin-labeled cytoskeleton, red: PKH26-labeled exosomes. (C) Experimental design schematic for the flow cytometry analysis. (D) Flow cytometry and corresponding quantitative analysis after endocytosis of Nexo, CataNexo, and CataKNexo in RGCs (n = 3 biologically independent experiments). (E) Flow cytometry and corresponding quantitative analysis after endocytosis of CataKNexo in RGCs with the application of the indicated inhibitors for distinct endocytosis pathways (n = 3 biologically independent experiments). Amiloride is an inhibitor of macropinocytosis; chlorpromazine is an inhibitor of clathrin-mediated endocytosis; genistein is an inhibitor of caveolae-mediated endocytosis. (F) Schematic for the anticipated transport of CataKNexo in RGCs. (G) Representative CLSM fluorescence images showing co-localization of PKH26 with LysoTracker fluorescence in RGCs. Blue: DAPI, red: PKH26-labeled CataKNexo, green: LysoTracker-labeled lysosome. The images on the left (scale bar: 10 μm) are presented with same magnification. (H) Representative CLSM fluorescence images showing colocalization of PKH26 with GolgiTracker fluorescence in RGCs. Blue: DAPI, red: PKH26-labeled CataKNexo, green: GolgiTracker-labeled Golgi apparatus. The images on the left (scale bar: 10 μm) are presented with same magnification. (I) Corresponding Pearson correlation coefficients of PKH26 with LysoTracker Green or GolgiTracker Green fluorescence calculated by pixel intensity using ImageJ software (n = 3 biologically independent experiments). (J) TEM image of Au nanoparticle-labeled CataKNexo stained with uranyl acetate. Scale bar: 50 nm. (K) TEM images of endocytosis, intracellular transport, and exocytosis of Au nanoparticle-labeled CataKNexo in RGCs. The images (scale bar: 200 nm) and insets on the right (scale bar: 100 nm) are presented with same magnification, respectively. Data in (D), (E), and (I) are mean ± SD and were compared using one-way ANOVA (D, E) or two-tailed unpaired Student’s t test (I). The experiments in (B), (G), (H), (J), and (K) were independently repeated three times with similar results. See also Figure S4.

Figure 5

Therapeutic efficacy and safety of CataKNexo in RIRI model mice (A) Schematic for evaluating the therapeutic benefit of CataKNexo for the treatment of RIRI model mice. (B) Transcriptome analysis (left, n = 3, 1 eye from each of 3 mice) of retinas from RIRI mice after intravitreal injection with PBS or CataKNexo, and gene set enrichment analysis (GSEA) (right) of the PBS group versus the CataKNexo group. (C) Representative neuroprotein fluorescent images obtained by CLSM and quantitative analysis of retinal sections from the various groups (n = 6, 1 eye from each of 6 mice). Scale bar: 50 μm. The images are presented at the same magnification. (D) Representative CLSM images for cell damage and quantitative analysis of retinal sections in the indicated groups (n = 6, 1 eye from each of 6 mice). Scale bar: 50 μm. The images are presented at the same magnification. (E) Representative ERG images presenting the retinal photoelectric conversion function and quantitative analysis of a-wave and b-wave amplitudes in the various groups (n = 6, 1 eye from each of 6 mice). Data in (C)–(E) are mean ± SD and were compared using one-way ANOVA. The experiments in (C)–(E) were independently repeated three times with similar results. See also Figure S5.

Figure 6

Distribution, therapeutic efficacy, and safety of CataKNexo in RIRI model Bama miniature pigs (A) Schematic for evaluating the therapeutic benefit and safety of CataKNexo in a Bama miniature pig RIRI model. The exosomes are derived from human neural stem cells. (B) TEM image of CataKNexo stained with uranyl acetate. Scale bar: 200 nm. (C) Fluorescence image of excised eyeball at 72 h after single intravitreal injection of CataKNexo. (D) Retinal section imaged using CLSM to evaluate the retinal penetration of CataKNexo. Scale bar: 50 μm. (E) Measurement of H₂O₂ concentration using a H₂O₂ Content Assay Kit after treatment with PBS or CataKNexo (n = 3, 1 eye from each of 3 pigs). (F) PCR analysis of retinal βIII-Tubulin in the PBS and CataKNexo groups (n = 3, 1 eye from each of 3 pigs). (G) Representative TUNEL staining of retinal sections from the PBS and CataKNexo groups (n = 3, 1 eye from each of 3 pigs). Scale bar: 50 μm. The images are presented at the same magnification. (H) Representative H&E staining of retinal sections in PBS and CataKNexo groups (n = 3, 1 eye from each of 3 pigs). Scale bar: 100 μm. The images are presented at the same magnification. (I) Representative OCT images and quantitative analysis of ONL thickness in the PBS and CataKNexo groups (n = 3, 1 eye from each of 3 pigs). Scale bar: 100 μm. The images are presented at the same magnification. (J) Representative images of full-field ERG and quantitative analysis of amplitudes of a-wave and b-wave in the PBS and CataKNexo groups (n = 3, 1 eye from each of 3 pigs). (K) Color fundus photograph (left) and intraocular pressure (right) using an Icare TonoVet rebound tonometer for the assessment of local responses to the intravitreal injection of CataKNexo (n = 3, 1 eye from each of 3 pigs). (L) Complete blood count (CBC, left) and blood biochemistry (right) for evaluation of side effects following intravitreal injection of CataKNexo (n = 3, 1 eye from each of 3 pigs). Data in (E), (F), (K), and (L) are mean ± SD and were compared using one-way ANOVA (E, F) or two-tailed unpaired Student’s t test (K, L). The experiments in (B)–(D) and (G)–(K) were independently repeated three times with similar results. See also Figure S6.