Abstract

   Glaucoma is a leading cause of irreversible blindness, primarily driven
   by the progressive degeneration of retinal ganglion cells (RGCs). In
   this study, we report the novel application of bone marrow mesenchymal
   stem cells (BMSCs)-derived apoptotic extracellular vesicles (ApoEVs) in
   a glaucomatous ischemia/reperfusion (IR) model. ApoEVs exhibited
   remarkable anti-inflammatory properties, were efficiently internalized
   by retinal neurons, promoted RGC survival, and preserved visual
   function. Transcriptomic analysis revealed that ApoEV treatment
   significantly downregulated Irgm1, an inflammation-related gene.
   Notably, this study also established a new drug delivery strategy by
   successfully loading Berberine (Ber)—a natural compound with
   well-documented anti-inflammatory and neuroprotective effects—onto
   ApoEVs. The combination further enhanced their protective effects on
   RGCs, with synergistic suppression of inflammation and improved
   neuronal viability. Mechanistically, this enhancement was mediated
   through the coordinated inhibition of the MAPK signaling pathway via
   Irgm1, which is identified here for the first time as a potential
   therapeutic target in glaucoma. Collectively, our findings highlight
   the dual function of Berberine-loaded ApoEVs as a potent cell-free
   therapeutic strategy that integrates targeted anti-inflammatory and
   neuroprotective effects, offering a promising new avenue for the
   treatment of glaucomatous neurodegeneration.

Graphical abstract

   [40]graphic file with name 12951_2025_3714_Figa_HTML.jpg

Supplementary Information

   The online version contains supplementary material available at
   10.1186/s12951-025-03714-2.

   Keywords: Apoptotic extracellular vesicles (ApoEVs), Berberine (Ber),
   Retinal ganglion cells (RGCs), Ischemia/reperfusion (IR) injury,
   Glaucoma, Neuroinflammation

Introduction

   Glaucoma remains the leading cause of irreversible blindness worldwide.
   According to the World Health Organization (WHO), approximately 76
   million individuals were affected by glaucoma in 2020, with over 10%
   progressing to complete blindness. This number is projected to rise to
   111.8 million by 2040 [[41]1]. The resulting vision loss imposes
   considerable physical, psychological, and socioeconomic burdens on both
   patients and society. The hallmark pathological feature of glaucoma is
   the progressive and selective degeneration of retinal ganglion cells
   (RGCs) [[42]2]. Elevated intraocular pressure (IOP) is the primary risk
   factor for glaucoma, and the resultant retinal ischemia/reperfusion
   (IR) injury is a key contributor to RGC death [[43]3].

   Mesenchymal stem cells (MSCs) are multipotent stromal cells capable of
   differentiating into various mesodermal lineages, including
   osteoblasts, chondrocytes, myocytes, and adipocytes [[44]4]. Due to
   their potent immunomodulatory and regenerative capacities, MSCs have
   been extensively studied in the context of cell-based therapy [[45]5,
   [46]6]. In recent years, MSC-derived secretory products—particularly
   extracellular vesicles (EVs)—have garnered attention as a promising
   cell-free therapeutic approach with low immunogenicity for treating
   inflammatory and degenerative diseases [[47]7, [48]8]. EVs are
   heterogeneous membrane-bound nanoparticles released by nearly all cell
   types and are essential mediators of intercellular communication in
   both physiological and pathological settings [[49]9]. Increasing
   evidence suggests that EVs, either as therapeutic agents themselves or
   as drug delivery vehicles, may represent a novel and effective strategy
   for the treatment of glaucoma [[50]10–[51]12].

   A growing body of evidence indicates that programmed cell death
   pathways critically regulate RGC degeneration in glaucoma [[52]13]. The
   proper regulation and clearance of dying cells are essential for
   maintaining immune homeostasis and preventing pathological inflammation
   or malignancy. Among these pathways, apoptosis plays a fundamental role
   in tissue homeostasis by eliminating superfluous, damaged, or
   immunogenic cells during development and immune responses [[53]14].
   Apoptosis, traditionally associated with cell death and tissue damage,
   is now understood as a tightly regulated physiological process that
   contributes to tissue equilibrium and can potentially elicit reparative
   or anti-inflammatory responses through the release of bioactive
   components [[54]15, [55]16].During apoptosis, cells release a distinct
   subtype of extracellular vesicles known as apoptotic extracellular
   vesicles (ApoEVs), which carry nuclear, cytoplasmic, and
   membrane-derived components—including nucleic acids, proteins, and
   lipids. These vesicles can be internalized by nearby or distant
   recipient cells, thereby exerting functional regulatory effects
   [[56]17, [57]18]. Among them, MSC-derived ApoEVs have garnered
   attention as potent mediators of tissue regeneration and repair
   [[58]19, [59]20]. For example, MSC-ApoEVs have been shown to facilitate
   regeneration across various tissues, including skin, hair follicles,
   bone, muscle, and vasculature [[60]21, [61]22]. Mechanistically, they
   enhance hair follicle growth by activating the Wnt/β-catenin signaling
   pathway in dermal papilla cells [[62]20], and promote bone formation in
   osteoporotic models via the Ras/Raf1/Mek/Erk signaling cascade
   [[63]23]. Moreover, MSC-ApoEVs exhibit therapeutic potential in
   modulating inflammation and immune responses [[64]24, [65]25]. However,
   the role of MSC-ApoEVs in ocular disease—particularly in
   neuroprotection against glaucomatous injury—remains largely unexplored.

   In this study, we hypothesize that bone marrow-derived MSCs (BMSCs)
   ApoEVs exert neuroprotective effects on RGCs in a glaucomatous IR
   injury model. We observed that ApoEVs distributed efficiently into the
   retina and were internalized by endogenous retinal cells, including
   neurons. Treatment with ApoEVs significantly enhanced RGCs survival and
   functional integrity, thereby attenuating glaucomatous
   neurodegeneration. Transcriptomic analysis of retinal tissues treated
   with ApoEVs identified a set of differentially expressed genes, among
   which Irgm1 emerged as a key target. Irgm1 has been implicated in
   modulating immune responses and maintaining cellular homeostasis in the
   central nervous system, and its down-regulation may contribute to the
   observed neuroprotection [[66]26, [67]27].

   To further enhance the therapeutic efficacy of ApoEVs, we explored the
   co-administration of Berberine, a natural isoquinoline alkaloid derived
   from Coptis chinensis Franch, known for its potent anti-inflammatory
   properties [[68]28]. Previous studies have shown that Berberine can
   suppress IRGM, the human immunity-related GTPase family M gene
   [[69]29]. In addition, it was reported that in the mouse
   atherosclerotic plaque model, Irgm1 (the murine homolog of human IRGM)
   deficiency suppresses macrophage apoptosis by inhibiting ROS generation
   and MAPK signaling transduction [[70]26]. In our model, intravitreal
   Berberine administration alone promoted RGCs survival. More
   importantly, its combination with ApoEVs resulted in a synergistic
   protective effect, significantly enhancing RGCs survival following IR
   injury.

   Taken together, these findings provide compelling evidence for the
   therapeutic potential of Berberine-loaded BMSCs-derived ApoEVs in
   neuroprotection against glaucomatous injury. This work lays a
   theoretical foundation for the future development of ApoEV-based
   interventions in glaucoma and other retinal neurodegenerative diseases.

Methods and materials

Cell lines and cell culture

Isolation and primary culture of human bone marrow-derived cells

   Bone marrow aspirates were obtained from five patients undergoing
   spinal fracture fixation or hip arthroplasty at Xiangya Hospital,
   Central South University. All procedures were approved by the
   Institutional Ethics Committee of Xiangya Hospital (Approval number:
   2024091259), and written informed consent was obtained from all
   participants. Aspirates were centrifuged at 110 g for 5 min at room
   temperature, and the supernatant was discarded. The cell pellets were
   treated with red blood cell lysis buffer (Cat.AWC0358a, Abiowell) and
   then resuspended in high-glucose DMEM(Cat. C11995500BT, Gibco). After
   24 h, the floating cells were washed away, and the medium was changed
   with F12(Cat.C11330500BT, Gibco), which contained 15% FBS (Cat.FSP500,
   ExCell), and 1% penicillin-streptomycin (Cat. C100C5, Ncmbio) after
   48 h.

R28 cell culture

   R28 cells, a retinal precursor cell line with differentiation
   potential, are commonly employed in vitro to investigate
   neuroprotective mechanisms and the pathophysiology of RGCs. It was
   provided by the Department of Anatomy and Neurobiology of Central South
   University (Changsha, China). Cells were cultured in low-glucose DMEM
   (Cat.C11885500BT, Gibco) supplemented with 10% FBS (Cat. FSP500,
   ExCell) and 1% penicillin-streptomycin (Cat.C100C5, Ncmbio) at 37 °C in
   a humidified atmosphere containing 5% CO₂.

Induction of BMSCs apoptosis

   Undifferentiated human BMSCs were washed twice with phosphate-buffered
   saline (PBS, Cat.G4202, Servicebio), after which the culture medium was
   replaced with complete medium containing 250 nM staurosporine (STS,
   Cat.HY-15141, MCE) and FBS depleted of exosome
   (Cat.C38010050.VivaCell). Cells were incubated for 12 h to induce
   apoptosis. Apoptosis was confirmed by morphological assessment and
   Western blot analysis.

Isolation and characterization of apoptotic extracellular vesicles (ApoEVs)

   ApoEVs were isolated according to previously established protocols
   [[71]30]. Briefly, 12 h after STS-induced apoptosis, the culture
   supernatant was collected and subjected to centrifugation at 800 × g
   for 10 min to remove cellular debris. The supernatant was then
   centrifuged at 16,000 × g for 30 min to pellet ApoEVs, which were
   subsequently washed twice with filtered PBS. Protein concentration was
   determined using a BCA protein assay kit (Cat.WB6501.Ncmbio) for
   quantification purpose.

   Nanoparticle tracking analysis (NTA) was performed using a NanoSight
   NS300 instrument (Malvern Instruments) equipped with a 405 nm laser and
   an sCMOS camera to assess size distribution. ApoEV samples were
   resuspended and diluted in filtered PBS. Each sample was captured for
   15 s using a camera level of 14 and a detection threshold of 8, and
   measurements were performed in triplicate. Filtered PBS alone was
   imaged to confirm the absence of contaminating particles. A total of
   1,498 frames were captured and analyzed using NTA 3.2 Dev Build 3.2.16
   software (Malvern Instruments).

   To assess apoptotic markers, purified ApoEVs were analyzed by Western
   blotting using antibodies against cleaved caspase-3 and β-actin.
   Additionally, ApoEVs were stained with Annexin V-Fluos and examined
   under a confocal laser scanning microscope. Furthermore, the expression
   of phosphatidylserine (PS) on the vesicle surface was assessed by
   Annexin V labeling using a nanoflow cytometry system (Flow
   NanoAnalyzer, Xiamen Folo Biotechnology Co., Ltd., China). Briefly,
   purified ApoEVs were incubated with fluorophore-conjugated Annexin V in
   binding buffer according to the manufacturer’s instructions. After
   incubation, samples were analyzed directly without further washing. The
   nanoflow cytometer was calibrated using standardized 100 nm and 200 nm
   reference beads, and data acquisition was performed under optimized
   threshold and gain settings. Annexin V-positive events were quantified
   and expressed as a percentage of total vesicles. The enrichment of
   Annexin V on ApoEVs confirmed their apoptotic vesicle identity.

Scanning electron microscopy (SEM) and transmission electron microscope(TEM)

SEM

   ApoEVs in PBS were fixed with 2.5% glutaraldehyde at 4 °C for 2 h,
   followed by post-fixation with 1% osmium tetroxide(Ted Pella Inc,
   Product Number.18456). Samples were dehydrated through graded ethanol
   and acetone series, then subjected to tert-butyl alcohol and isoamyl
   acetate replacement. After vacuum drying, vesicles were mounted on
   conductive tape, gold-coated, and examined using a Hitachi S-3400 N
   scanning electron microscope.

TEM

   ApoEVs were fixed in 2.5% glutaraldehyde (0.1 M phosphate buffer, pH
   7.4) at 4 °C for at least 2 h, followed by post-fixation in 1% osmium
   tetroxide for 1–2 h. After phosphate-buffered rinses, samples were
   dehydrated through a graded acetone series (50%, 70%, 90%, and 100%,
   each concentration for 15 min), then infiltrated with a mixture of
   acetone and embedding resin(1: 1), and embedded in pure resin for 12 h.
   Polymerization was performed at 37 °C overnight and then at 60 °C for
   12–24 h. Ultrathin Sects. (50–100 nm) were cut using a Leica UC-7
   ultramicrotome, stained with uranyl acetate and lead citrate, and
   examined using a Hitachi HT7700 transmission electron microscope.

Labeling and tracking of ApoEVs

   The protein concentration of ApoEVs was first determined using a BCA
   protein assay kit. A fluorescent dye (Cat.40758ES25, Yeasen) was then
   added to the ApoEV suspension and incubated for 10 min at room
   temperature. After incubation, 10 mL of 1× PBS was added to the
   ApoEV–dye mixture to dilute unbound dye, followed by gentle mixing. The
   labeled ApoEVs were re-isolated using the same centrifugation protocol
   as previously described to remove excess dye. The final pellet was
   resuspended in 200 µL of 1× PBS to obtain the fluorescently labeled
   ApoEVs.

   For in vitro tracking, R28 cells were incubated with the labeled ApoEVs
   for 24 h. The cells were then fixed and processed for
   immunofluorescence staining according to standard protocols. Cellular
   uptake of labeled ApoEVs was visualized under a fluorescence
   microscope.

   For in vivo tracking, 1.5 µL of the labeled ApoEV suspension was
   intravitreally injected into the vitreous cavity. After 72 h, mice were
   sacrificed, and retinas were dissected and flat-mounted. The
   distribution of labeled ApoEVs was observed using a confocal laser
   scanning microscope.

Western-blot

   Retinal tissues were harvested at the indicated time points, and total
   protein was extracted using RIPA lysis buffer (Cat.WB3100, Ncmbio)
   supplemented with a protease and phosphatase inhibitor cocktail
   (Cat.P002, Ncmbio) (100:1, v/v). Protein concentration was determined
   using a BCA protein assay kit. Equal amounts of protein were separated
   on 8%, 10%, or 15% SDS-PAGE gels, depending on molecular weight, and
   subsequently transferred onto polyvinylidene difluoride (PVDF)
   membranes.

   Membranes were blocked with 5% skim milk or bovine serum albumin (BSA,
   Cat.4240, Biofroxx) in in PBS with 0.1%Triton X-100 (PBST, Cat.T8200,
   Solarbio) for 2 h at room temperature, followed by overnight incubation
   at 4 °C with primary antibodies. After five washes with PBST (5 min
   each), membranes were incubated with appropriate HRP-conjugated
   secondary antibodies (Cat.SA00001-1or SA00001-2,Proteintech)for 2 h at
   room temperature. Protein bands were visualized using enhanced
   chemiluminescence (ECL, Cat.[72]P10100, Ncmbio) detection reagents and
   quantified using ImageJ software (National Institutes of Health,
   Bethesda, MD, USA).

   The following antibodies were used in this study: p44/42 MAPK (Erk1/2)
   antibody (Cat. #4695, Cell Signaling Technology), phospho-p44/42 MAPK
   (Erk1/2) antibody (Cat. #9101, Cell Signaling Technology), phosphor-p38
   antibody (Cat. #4511, Cell Signaling Technology), p38 antibody (Cat.
   #8690, Cell Signaling Technology), IRGM1(Cat. #14979, Cell Signaling
   Technology), and ACTIN (Cat. 66009-1, Proteintech).

qRT-PCR

   Total RNA was isolated from retinal tissues using Trizol reagent, and
   RNA concentration/purity was assessed spectrophotometrically.
   High-quality RNA samples were reverse-transcribed into cDNA using a
   cDNA Synthesis SuperMix kit with 1 µg total RNA as template (Cat.11141,
   Yeasen). Target gene amplification was performed using SYBR Green
   Master Mix (Cat.11202, Yeasen) under manufacturer- recommended
   conditions, with β-ACTIN serving as the internal control. Primer
   sequences are provided in Supplementary Table 1.

Cell viability assay

   The Cell Counting Kit-8 (Cat. EK-5103; ECOTOP SCIENTIFIC) was applied
   to measure the viability of cells according to the manufacturer’s
   instructions. R28 cells were seeded in 96-well plates at a density of
   5000 cells/well and incubated for 24 h. Then, they were divided into
   four groups: Ctrl/PBS group, Ctrl/ApoEV group, H[2]O[2] treatment
   group, and H[2]O[2] + ApoEV treatment group. The H[2]O[2] treatment
   group was cultured in the medium containing H[2]O[2] (200 µmol, Cat.
   10011218,Sinopharm Chemical Reagent Co.,Ltd) for 2–4 h. Subsequently,
   cell absorbance was measured at 450 nm using a microplate reader
   (Synergy LX; BioTek, USA) after 2–4 h of CCK8 addition.

Calcein AM and Propidium iodide (PI) Live/Dead cell staining

   R28 cells were seeded into 24-well plates and cultured under standard
   conditions. After the designated treatments, the culture medium was
   removed, and cells were gently washed once with PBS. A fresh staining
   solution containing Calcein acetoxymethyl ester (Calcein AM) and
   propidium iodide (PI) (Cat.KGAF001, KeyGEN Biotech), each diluted
   1:2000 in low-glucose basal medium, was prepared. Cells were incubated
   with the staining solution at 37 °C in the dark for 30 min. After
   incubation, the staining solution was removed and replaced with PBS.
   Fluorescence images were then captured using a fluorescence microscope
   (Leica DMi8, Wetzlar, Germany) to assess cell death in each group.

Intracellular reactive oxygen species (ROS) detection

   The supernatant of each group of cells was discarded, and the cells
   were washed twice with serum-free medium. DCFH-DA (Cat.E-BC-K138-F,
   Elabsciences)was diluted with serum-free medium at a ratio of 1:1000 to
   a final concentration of 10µM and added to the cells. The cells were
   incubated in the dark at 37 °C in a cell culture incubator. After the
   treatment was completed, the cells were washed with serum-free medium 3
   times. Fluorescence was detected using a fluorescence microscope (Leica
   DMi8, Wetzlar, Germany).

Animal model and drug administration

   In this study, wild-type C57BL/6J mice (8 weeks old, weighing 20–25 g)
   were used as experimental subjects. Based on our previous research, an
   acute angle-closure glaucoma model was established by elevating IOP to
   120 mmHg for 60 min [[73]31, [74]32].

   Mice were anesthetized via intraperitoneal injection of 1%
   pentobarbital sodium (10 mg/kg). Following anesthesia, 0.5% tropicamide
   phenylephrine (Santen Pharmaceutical Co, Ltd, Shiga Plant) followed by
   0.4% obtained hydrochloride (Benoxil; Santen Pharmaceutical Co, Ltd)
   were then applied to the eyes for mydriasis and cornel topical
   anesthesia. The animals were placed on a heating pad to maintain body
   temperature at 37 °C. A 30-gauge needle connected to a saline infusion
   system was carefully inserted into the anterior chamber of the eye. The
   IOP was gradually increased to 120 mmHg and maintained for 60 min. In
   the control group, a sham procedure was performed without IOP elevation
   in the left eye.

   Intravitreal drug delivery was conducted as previously described
   [[75]31]. Using a Hamilton syringe equipped with a 30-gauge glass
   micropipette, intravitreal injections were performed under a surgical
   microscope. A total volume of 1.5 µL of PBS, ApoEVs, Berberine (Ber),
   or ApoEVs co-incubated with berberine (Ber-ApoEVs) was slowly injected
   into the vitreous cavity. The IR model was induced 24 h after
   injection.

   All animal procedures were conducted in accordance with the ARVO
   Statement for the Use of Animals in Ophthalmic and Vision Research. The
   experimental protocols were reviewed and approved by the Institutional
   Animal Care and Use Committee of the Laboratory Animal Research Center,
   Xiangya hospital, Central South University (Approval No:2024091259).

Histological staining and retinal morphometry

   Histological evaluations were performed on retinal cross-sections.
   H&E-stained retinal morphology was acquired post IR injury, to
   visualize and obtain a count of cells in the ganglion cell layer (GCL).
   Eyes were immediately enucleated post-execution and fixed in FAS eye
   fixation fluid (Servicebio, China) for 24 h at 4 °C. The eyes were
   embedded in paraffin(Servicebio, Cat No. G1101) and were cut into
   sections of 4 μm thickness through the optic nerve, prepared in the
   standard way, then stained with hematoxylin and eosin [[76]33]. Retinal
   morphology was observed, and photomicrographs were taken using a light
   microscope. Sections containing the optic nerve stump were chosen for
   consistency, and at least three discontinuous sections (per animal)
   were analyzed using CaseViewer software.

Whole-mount retina immunostaining

   Mouse eyes were enucleated and fixed in 4% paraformaldehyde (PFA, w/v)
   for 2 h at room temperature. After removing the cornea and lens, the
   retinas were further fixed in 4% PFA for an additional 1 h at RT. The
   retinas were then dissected, flat-mounted onto glass slides, and
   divided into four equal quadrants. Following permeabilization with 0.3%
   PBST for 10 min, the tissues were blocked with 5% BSA for 1 h and
   incubated overnight at 4 °C with an anti-Brn3a antibody (Cat.ab245230,
   Abcam,1:500 dilution). After three PBS washes, the retinas were
   incubated with Alexa Fluor 488-conjugated goat anti-rabbit secondary
   antibody(Cat.#4412, Cell Signaling Technology) for 2 h at RT, followed
   by mounting.

   Fluorescence microscopy was used to capture images of each retinal
   quadrant (with a shared central region). Brn3a-positive cells were
   counted using ImageJ to determine RGCs density (RGCs/mm²). The mean
   RGCs density across all groups was calculated, and relative RGCs
   density was expressed as a percentage of the control group. Statistical
   analysis of Brn3a + cell density was performed using GraphPad Prism
   9.0.

VEP&ERG

Visual electrophysiological assessments

Flash visual evoked potential (FVEP)

   Mice were anesthetized with 1% sodium pentobarbital, and then
   0.5%tropicamide phenylephrine (Santen Pharmaceutical Co, Ltd, Shiga
   Plant) followed by 0.4%obtained hydrochloride (Benoxil; Santen
   Pharmaceutical Co, Ltd) were then applied to the eyes for mydriasis and
   cornel topical anesthesia.The animals were placed on a heating pad to
   maintain body temperature at 37 °C. GenTeal lubricant eye gel(GenTeal^®
   Tears Lubricant Eye Gel, Alcon, Fort Worth, TX, USA) were applied to
   prevent corneal dehydration. Three needle electrodes were
   subcutaneously inserted at the following locations:

Recording electrode

   Midpoint between the ears, in contact with the occipital bone.

Reference electrode

   Subcutaneous insertion at the nasal region.

Ground electrode

   Subcutaneous insertion at the tail.

   One eye was occluded, and FVEP recordings were performed following
   international standards ([77]http://www.iscev.org/standards). The
   procedure was repeated for the contralateral eye. The N1-P1 amplitude
   was analyzed for each group.

Flash electroretinography (FERG)

   Prior to testing, mice were dark-adapted for > 12 h. Preparation was
   similar to FVEP, except recordings were conducted under dim red light.
   Two corneal ring electrodes and three needle electrodes were used:

Recording electrodes

   Placed on both corneas, with carboxymethylcellulose eye drops enhancing
   conductivity and moisture.

Reference electrodes

   Subcutaneously inserted near the nasal region.

Ground electrode

   Subcutaneously inserted at the tail.

   Stimulation and recording followed the International Society for
   Clinical Electrophysiology of Vision (ISCEV) standards
   ([78]http://www.iscev.org/standards). The amplitudes and implicit times
   of the a-wave and b-wave were analyzed.

mRNA-seq and bioinformatic analysis

    1. Bulk RNA sequencing

   The total RNA of mouse retina tissue was extracted using Trizol reagent
   (Invitrogen, Carlsbad, CA, USA). The concentration, quality, and
   integrity were determined using a NanoDrop spectrophotometer (Thermo
   Scientific). 3 micrograms of RNA were used as input material for the
   sequencing library preparation. Sequencing libraries were generated
   according to the following steps. The sequencing library was then
   sequenced on a NovaSeq 6000 platform (Illumina) by Shanghai Personal
   Biotechnology Cp. Ltd.
     * b)
       Transcriptome analysis

   The original data in FASTQ format (Raw Data) was filtered using fastp
   (0.22.0) software to get a high-quality sequence (Clean Data) for
   further analysis. The reference genome (Mus_musculus. GRCm39) and gene
   annotation files were downloaded from the genome website. The filtered
   reads were mapped to the reference genome using HISAT2 (v2.1.0). HTSeq
   (v0.9.1) statistics was used to compare the Read Count values on each
   gene as the original expression of the gene. RSEM (version 1.2.31) was
   used for the transcript abundance estimation. The following
   bioinformatic analysis is conducted in R (4.2.3). The RSEM-expected
   counts were then subjected to DESeq2 (1.38.3) [[79]34] for the
   identification of differentially expressed genes (DEGs), with a
   criterion of |log2 fold change| >1 and adjusted p-value of < 0.05. Venn
   diagrams showing the number of common DEGs across the experimental
   comparisons were drawn by using Jvenn [[80]35]. Meanwhile,
   ComplexHeatmap (v2.16.0) was used to perform a bidirectional clustering
   analysis of all different genes of samples [[81]36]. A heatmap was
   generalized according to the expression level of the same gene in
   different samples and the expression patterns of different genes in the
   same sample with the Euclidean method to calculate the distance and the
   Complete Linkage method to cluster. ClusterProfiler (v4.6.0) software
   was used to carry out the enrichment analysis of the KEGG (Kyoto
   Encyclopedia of Genes and Genomes) pathway of differential genes,
   focusing on the significant enrichment pathway with a P-value < 0.05
   [[82]37]. Other R package used to declaratively creating graphics
   includes ggplot2 (3.4.4), cowplot (1.1.3), RColorBrewer (1.1.3),
   ggrepel (0.9.5) [[83]38–[84]40].

Preparation of ApoEV-Berberine complex (Ber-ApoEV)

   ApoEVs were incubated with berberine (Ber) at a mass ratio of 1:1.5
   (ApoEV protein: berberine) under gentle agitation at room temperature
   overnight. After incubation, unbound berberine was removed using the
   same isolation protocol as described above. The concentration of
   berberine encapsulated within ApoEVs was determined using
   high-performance liquid chromatography (HPLC) analysis conducted on an
   isocratic liquid chromatography system (JASCO), which included a dual
   pump (PU-1580), a Rheodyne injector with a 20 µL loop, and a photodiode
   array detector (MD-1510). Chromatographic data were recorded using
   Jasco-Borwin software version 1.50. Separation was performed on a
   Cosmosil C8 column (150 × 4.6 mm, 5 μm particle size) using a mobile
   phase consisting of water and acetonitrile (70:30, v/v) containing 0.2%
   trifluoroacetic acid, at a flow rate of 1.0 mL/min. The column was
   equilibrated prior to sample injection, and detection was carried out
   at a wavelength of 348 nm. The HPLC method was developed and validated
   according to ICH guidelines to ensure reliable quantification of
   Berberine.

Image processing using Imaris

   Colocalization analysis and quantification of IBA-1 and OX-6 signals
   were performed using Imaris10.2.0 (Bitplane). Confocal z-stack images
   were first converted into Imaris file format, and Regions of Interest
   (ROIs) were defined based on IBA-1 + signals to identify microglial
   cells. A 3D surface rendering was generated for both IBA-1 and OX-6
   using consistent parameters [[85]41]. The “Coloc” module in Imaris was
   used to assess signal overlap within the ROIs, and the colocalized
   volume (IBA-1 + OX-6+) was quantified for each dataset. For microglial
   morphology analysis, sphericity of IBA-1 + cells was quantified by
   generating individual 3D surfaces, followed by automatic measurement of
   sphericity values using the “Statistics” function in Imaris 10.2.0. All
   output data were exported and subjected to statistical analysis using
   GraphPad Prism 9.0.

Statistical analysis

   All data are presented as mean ± standard deviation (SD). Statistical
   analyses were conducted using GraphPad Prism 9.0 (GraphPad Software,
   San Diego, CA, USA). For comparisons between two groups, unpaired
   two-tailed Student’s t-test was used. For multiple group comparisons,
   one-way or two-way analysis of variance (ANOVA) followed by Tukey’s
   post hoc test was applied when data conformed to a normal distribution.
   Normality was assessed using the Shapiro–Wilk test. When data did not
   meet normality assumptions, nonparametric tests were applied.
   Specifically, the Kruskal–Wallis test followed by Dunn’s multiple
   comparison test was used for multiple non-normally distributed group
   comparisons. A P value of < 0.05 was considered statistically
   significant. All experiments were performed in triplicate or as
   otherwise indicated. Details of sample size and statistical tests used
   are provided in the figure legends.

Results

Isolation and validation of BMSCs-derived apoevs

   To isolate and characterize ApoEVs derived from BMSCs(Fig. [86]1A), a
   combination of TEM, SEM, NTA, Western-blotting, confocal microscopy and
   flow cytometry was employed, following standard protocols previously
   described for EV analysis [[87]23]. TEM & SEM imaging revealed that
   ApoEVs displayed the classical cup-shaped or spherical morphology (Fig.
   [88]1B and C). NTA demonstrated a peak-shaped curve, with most of the
   area under the curve in the characteristic ApoEV size range of 100–300
   nm and a significant peak at around 200 nm (Fig. [89]1D).

Fig. 1.

   [90]Fig. 1
   [91]Open in a new tab

   Isolation, characterization, and uptake of BMSCs-derived ApoEVs (A)
   Schematic workflow illustrating the isolation and validation process of
   BMSCs-ApoEVs. (B) SEM image showing the spherical morphology of
   ApoEVs.(C) TEM image showing the characteristic cup-shaped structure of
   ApoEVs. (D) NTA revealing a bell-shaped size distribution with a peak
   around 200 nm, consistent with the typical size range of ApoEVs.(E)
   Western blot analysis showing cleaved caspase-3 expression in ApoEVs.
   (F) Flow cytometry analysis demonstrating Annexin V expression on the
   surface of ApoEV. (G) Immunofluorescence staining further confirming
   Annexin V positivity in ApoEVs. (H) In vitro uptake assay showing
   internalization of DiD-labeled ApoEVs by recipient cells. (I) In vivo
   retinal distribution of DiD-labeled ApoEVs after intravitreal
   injection, confirming localization in the neural retina. Scale bars as
   indicated

   Furthermore, Western blotting revealed the presence of cleaved
   caspase-3,a hallmark of apoptosis in the ApoEV fraction but not in
   parental BMSCs, confirming the apoptotic origin of these vesicles
   (Fig. [92]1E). Flow cytometry analysis and immunofluorescence staining
   confirmed that BMSCs-ApoEVs express classical apoptotic marker Annexin
   V, a hallmark of membrane phosphatidylserine exposure during apoptosis
   (Fig. [93]1F and G). These molecular features collectively validate the
   vesicular identity and apoptotic nature of the isolated ApoEVs.

   To evaluate the biodistribution and uptake potential of BMSCs-ApoEVs,
   vesicles were labeled with the lipophilic fluorescent dye DiD and
   subsequently applied to both in vitro cultures and in vivo via
   intravitreal injection. Fluorescence microscopy revealed robust DiD
   signal in recipient cells in vitro, as well as in retinal tissue
   following in vivo administration, confirming the efficient
   internalization and retinal penetration of BMSCs-ApoEVs (Fig. [94]1H
   and I). These findings support the notion that ApoEVs can effectively
   reach neural retinal cells and are bioavailable for mediating local
   functional effects.

Neuroprotective effects of apoevs in acute glaucoma

   The neuroprotective potential of ApoEVs was first assessed in vitro
   using the R28 retinal precursor cell line subjected to the Hydrogen
   Peroxide (H₂O₂)-Induced Oxidative Stress injury model [[95]42]. As
   outlined in the experimental workflow (Fig. [96]2A), R28 cells were
   treated with ApoEVs at various concentrations. Cell viability,
   evaluated via CCK-8 assay, demonstrated a significant, dose-dependent
   improvement in survival. The application of 1.5 µL of ApoEVs at these
   concentrations resulted in pronounced enhancement in cell viability
   (Fig. [97]2B). ROS levels measured after H₂O₂ stimulation were markedly
   decreased in ApoEV-treated cells relative to PBS controls, indicating
   that ApoEVs mitigate oxidative stress-induced injury (Fig. [98]2C and
   D). ApoEV treatment significantly reduced PI-positive cell proportions
   from 59.32 ± 17.10% to 8.56 ± 7.23%, representing an 85.6% decrease in
   cell death following H₂O₂ exposure. Consistent findings were observed
   in PI staining assays, with the ApoEV-treated groups exhibiting reduced
   cell death at both 24 and 48 h post H₂O₂ stimulation compared to the
   untreated controls: Compared to the H₂O₂-injured group (H₂O₂/PBS, mean
   ± SD: 59.32 ± 17.10%), ApoEV treatment significantly reduced
   PI-positive cell percentages (H₂O₂/ApoEV, mean ± SD: 8.56 ± 7.23%),
   reflecting an 85.57% decrease in cell death rate. (Figure [99]2E and
   F).

Fig. 2.

   [100]Fig. 2
   [101]Open in a new tab

   BMSCs-ApoEVs protect R28 retinal precursor cells from oxidative
   stress-induced injury in vitro. (A) Schematic illustration of the
   experimental workflow. (B) CCK-8 assay in cell viability following
   ApoEV treatment. (C, D) Quantification and representative images of
   intracellular reactive oxygen species (ROS) levels. (E, F) Propidium
   iodide (PI) staining assay results at 2 and 4 h post H₂O₂ exposure.
   Data are presented as mean ± SD from at least three independent
   experiments. Statistical analysis was performed following a normality
   test; for non-normally distributed data, the Kruskal–Wallis test
   followed by pairwise comparisons was used to determine significance.
   *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Scale bars as
   indicated

   In vivo, the neuroprotective efficacy of ApoEVs was evaluated in a
   murine model of acute glaucoma induced by retinal IR injury.
   Intravitreal injections of ApoEVs were administered as depicted in the
   schematic diagram (Fig. [102]3A). To evaluate the neuroprotective
   effect of ApoEVs on retinal structure, HE staining was performed on
   retinal sections at 3 and 7days post-injury (dpi) (Fig. [103]3B-E). In
   the IR/PBS group, a marked reduction in total retinal cell number was
   observed compared to the Ctrl/PBS group, indicating significant retinal
   degeneration induced by IR injury. Notably, treatment with ApoEVs
   partially rescued this cell loss. Quantitative analysis revealed a
   significant increase in the number of retinal cells in the IR/ApoEV
   group compared to the IR/PBS group at both 3 dpi and 7 dpi (P < 0.05),
   suggesting that ApoEV administration mitigated IR-induced retinal
   damage and preserved cellular integrity. To further evaluate RGCs
   survival, retinal wholemounts were immune-stained for Brn3a at 3dpi and
   7dpi. Quantitative analysis demonstrated a significantly higher RGCs
   survival rate in the ApoEV-treated group compared to IR group, as
   compared to the IR injury group at 3dpi (PBS/IR, mean ± SD:
   60.65 ± 5.27%), ApoEV treatment significantly increased RGCs survival
   (ApoEV/IR, mean ± SD: 77.84 ± 7.37%), corresponding to a 28.3%
   improvement in relative RGC viability. The result is consistent at 7dpi
   as RGCs survival in the untreated IR group (PBS/IR, mean ± SD:
   28.82 ± 6.11%) was significantly increased in the ApoEV-treated group
   (ApoEV/IR, mean ± SD: 63.46 ± 10.05%) (Fig. [104]3F-I). The above data
   indicated that ApoEVs confer robust neuroprotection against ischemic
   damage.

Fig. 3.

   [105]Fig. 3
   [106]Open in a new tab

   BMSCs-ApoEVs confer structural and functional neuroprotection in a
   murine retinal IR model (A) Schematic of intravitreal ApoEV
   administration in the IR model. (B, C) Representative images of H&E
   staining and quantification at 3 days post-reperfusion(3dpi). (D, E)
   Representative images of H&E staining and quantification at 7 days
   post-reperfusion(7dpi). (F, G) Brn3a immunostaining of retinal
   wholemounts at 3 dpi. (H, I) Brn3a immunostaining of retinal
   wholemounts at 7 dpi. (J–N) Functional assessments using ERG and VEP.
   Statistical analysis was performed using the Kruskal–Wallis test with
   pairwise comparisons after confirming data distribution. All data are
   presented as mean ± SD; *P < 0.05, **P < 0.01, ***P < 0.001,
   ****P < 0.0001. Scale bars as indicated

   In addition to assessing the morphological changes, retinal visual
   function was evaluated using F-ERG and F-VEP. The combined application
   of these two techniques provides a more comprehensive assessment of
   retinal functional status. F-ERG, which reflects retinal function,
   showed that both a-wave and b-wave amplitudes were significantly
   decreased following IR injury, whereas treatment with intravitreal
   ApoEV administration led to a recovery of these amplitudes
   (Fig. [107]3J–L). Furthermore, F-VEP was employed to objectively
   evaluate the functionality of the retinal–cortical neural pathway. The
   IR group exhibited a significant reduction in N1–P1 amplitude,
   indicating impaired functional output of RGCs. Notably, ApoEV treatment
   partially restored the N1–P1 amplitude compared to the IR group,
   suggesting a protective effect of ApoEVs on RGCs function and their
   contribution to maintaining normal visual signal transmission
   (Fig. [108]3M and N).

   Collectively, these data highlight the therapeutic potential of
   MSC-ApoEVs in preserving retinal structure, reducing apoptosis, and
   restoring neuronal function in acute glaucoma models.

Transcriptomic profiling reveals ApoEV-mediated regulation of
neuroinflammatory responses in IR-injured retina

   To investigate the molecular mechanisms underlying the neuroprotective
   effects of ApoEVs in retinal IR injury, we performed bulk RNA
   sequencing on retinal tissues collected from Ctrl/PBS, IR/PBS, and
   IR/ApoEV groups. Violin plots of transcript abundance (TPM)
   demonstrated consistent data quality across biological replicates
   (Fig. [109]4A). DEG analysis revealed substantial transcriptional
   changes induced by IR, with 1,630 upregulated and 631 downregulated
   genes in IR/PBS versus Ctrl/PBS (Fig. [110]4B). Notably, ApoEV
   treatment partially reversed this expression profile, with 458 genes
   upregulated and 67 downregulated in IR/ApoEV versus IR/PBS.

Fig. 4.

   [111]Fig. 4
   [112]Open in a new tab

   Transcriptomic analysis reveals that ApoEV treatment modulates
   IR-induced immune responses (A) Violin plot showing the distribution of
   gene expression (TPM) across replicates from each group, demonstrating
   data quality and consistency. (B) Bar graph summarizing the number of
   differentially expressed genes (DEGs) across comparisons. DEGs were
   defined as those with adjusted p-value < 0.05 and absolute log₂ fold
   change > 1. (C, D) Volcano plots illustrating significantly altered
   genes between groups. Notably, Irgm1 was highly upregulated in IR/PBS
   vs. Ctrl/PBS and downregulated following ApoEV treatment (IR/PBS vs.
   IR/ApoEV). (E, F) Heatmaps displaying top DEGs associated with
   inflammatory and apoptotic responses. Rows represent genes, columns
   represent biological replicates. Color scale indicates Z-score
   normalized expression (red: high expression, blue: low expression) (G,
   H) GO enrichment analysis of DEGs in IR/PBS vs. Ctrl/PBS (G) and IR/PBS
   vs. IR/ApoEV (H). (I, J) KEGG pathway analysis revealing upregulation
   of MAPK signaling pathway and apoptosis in IR/PBS group, which were
   partially reversed by ApoEV administration

   Volcano plots further highlighted these differences, showing robust
   upregulation of immune and apoptosis-related genes such as Irgm1,
   Tnfaip3, and Cxcl10 in the IR/PBS group [[113]43], which were
   significantly downregulated following ApoEV intervention (Fig. [114]4C
   and D). Heatmap clustering confirmed the normalization of these
   inflammatory gene signatures in the IR/ApoEV group, including marked
   suppression of Irgm1, a key regulator of inflammation and autophagy
   under stress conditions (Fig. [115]4E and F).

   Gene Ontology (GO) enrichment analysis revealed that IR insult
   activated numerous immune-related biological processes, including
   “immune system process,” and “response to cytokine stimulus”
   (Fig. [116]4G, Sup Fig. [117]1A-1D). These responses were largely
   attenuated in the IR/ApoEVs group, which exhibited reduced enrichment
   of innate immune terms and increased expression of genes related to
   cellular homeostasis (Fig. [118]4H).

   Similarly, KEGG pathway analysis showed that IR triggered pathways
   related to inflammation and apoptosis, especially MAPK signaling
   (Fig. [119]4I, Sup Fig. [120]1E&1 F). Following ApoEVs treatment,
   activation of these pathways was significantly dampened (Fig. [121]4J).
   In particular, the MAPK signaling cascade, known to mediate stress and
   apoptotic responses in retinal neurons, was significantly
   downregulated, supporting the anti-inflammatory and neuroprotective
   effects of ApoEVs.

   Collectively, these transcriptomic findings suggest that ApoEVs
   mitigate IR-induced retinal injury by suppressing immune and apoptotic
   signaling, particularly via the inhibition of Irgm1 expression and MAPK
   pathway activation.

ApoEVs impresses neuroinflammation in acute glaucoma

   To validate the transcriptomic alterations observed in our RNA-seq
   analysis, we first examined the expression of Irgm1, a key regulator of
   immune responses and autophagy. Both Western Blot (Fig. [122]5A and B)
   and qRT-PCR(Fig. [123]5C) analyses demonstrated a marked upregulation
   of IRGM1/Irgm1 in the retina following IR injury.

Fig. 5.

   [124]Fig. 5
   [125]Open in a new tab

   BMSCs-derived ApoEVs mitigate neuroinflammation and microglial
   activation in the IR-injured retina. (A,B) Western blot analysis and
   quantification of IRGM1 protein expression in retinas after IR
   injury.(C) qRT-PCR analysis of Irgm1 mRNA levels after IR injury at
   several timepoints. (D) qRT-PCR analysis of Il-6, Caspase1, Igf1, Bcl2,
   and Bst2 mRNA levels in retinal tissues at 3 dpi. (E,F) Representative
   immunofluorescence images showing IBA-1 and OX-6 staining in the
   retina. Scale bar = 20 μm.(G) Quantification of percentage of
   OX-6stained ROI co-localized with IBA-1⁺ microglial cell in each group.
   (H,I) 3D morphological reconstruction and analysis of IBA-1⁺ microglial
   sphericity using Imaris software (v10.2.0). All quantitative data are
   shown as mean ± SD. Group comparisons were analyzed using the
   Kruskal–Wallis test followed by Dunn’s multiple comparisons. *P < 0.05,
   **P < 0.01, ***P < 0.001, ****P < 0.0001. Scale bars as indicated

   Building upon this observation, we sought to assess whether ApoEVs
   impact the broader neuroinflammatory environment in the retina. We
   conducted qRT-PCR analysis of retinal tissue harvested 3dpi to quantify
   expression of selected inflammation-related cytokines. Among these,
   Il-6, Igf1, Caspase1, and Bcl2 [[126]44–[127]47] were chosen for their
   established roles in regulating immune responses, apoptosis, and
   cellular stress in neural tissue(Fig. [128]5D). ApoEVs treatment
   resulted in a significant downregulation of Il-6 and Caspase1, both of
   which are typically associated with pro-inflammatory and pro-apoptotic
   signaling. In contrast, the levels of Igf1 and Bcl2 were significantly
   upregulated, consistent with enhanced neuroprotective and
   anti-apoptotic signaling. The expression of Bst2, a multifunctional
   cytokine-like molecule implicated in innate immune regulation, also
   exhibited a trend toward normalization following ApoEV treatment (Fig.
   [129]5D).

   To further examine the effect of ApoEVs on microglial activation, we
   performed immunofluorescence staining using IBA-1 and OX-6 to label
   microglial cells in vivo (Fig. [130]5E and F). Quantitative analysis
   revealed a significant reduction in the number of IBA-1⁺ microglia in
   the retina of ApoEV-treated animals compared to vehicle-treated
   controls (Fig. [131]5G) indicating reduced microglial recruitment or
   proliferation in response to injury. In addition, using Imaris 10.2.0
   software, we performed detailed 3D morphometric analysis to quantify
   the morphology change of IBA-1⁺ microglia(Fig. [132]5H and I). These
   analyses showed that IR injury markedly altered microglial morphology
   toward a more amoeboid, activated state, characterized by increased
   sphericity and reduced branching. ApoEVs treatment significantly
   reversed these changes, as reflected by a reduction in sphericity and
   restoration of a more ramified morphology. These findings suggest that
   microglial activation and morphological changes are key components of
   the IR-induced neuroinflammatory response, which can be attenuated by
   ApoEV treatment.

   Collectively, these findings demonstrate that ApoEVs mitigate
   neuroinflammation after IR injury by downregulating pro-inflammatory
   cytokines, enhancing neuroprotective gene expression, and reducing
   microglial activation.

Protective effects of Berberine on RGCs survival following retinal IR

   Berberine has been implicated in modulating inflammatory responses via
   IRGM1, raising the possibility of its relevance to neurodegenerative
   diseases such as glaucoma. To further investigate the neuroprotective
   effects of Berberine in the context of acute glaucoma, we administered
   Berberine via intravitreal injection in a mouse model and assessed its
   therapeutic impact (Fig. [133]6A).

Fig. 6.

   [134]Fig. 6
   [135]Open in a new tab

   Protective effects of Berberine on RGCs survival following retinal IR
   injury and characterization of Berberine-loaded ApoEVs (Ber-ApoEVs).
   (A) Experimental schematic depicting intravitreal injection of
   Berberine. (B) Representative H&E stained retinal sections after
   Berberine treatment at different doses. (C) Quantification of
   percentage of GCL cell number compared to Ctrl/PBS group from H&E
   stained sections. (D) Representative immunofluorescence images of
   Brn3a-positive RGCs in whole mounts after Berberine treatment, scale
   bar = 100 μm. (E) Quantitative analysis of Brn3a + RGCs density
   compared to Ctrl/PBS group. (F) Molecular docking model predicting
   stable binding of Berberine within the active site of IRGM1. (G)
   Schematic workflow for loading Berberine into ApoEVs. (H) HPLC
   quantification of Berberine concentration. (I) NTA showing particle
   size distribution and concentration of Ber-ApoEVs. (J) TEM images of
   Ber-ApoEVs. All quantitative data are shown as mean ± SD. Group
   comparisons were analyzed using the Kruskal–Wallis test followed by
   Dunn’s multiple comparisons. *P < 0.05, **P < 0.01, ***P < 0.001,
   ****P < 0.0001

   A dose-response evaluation was first performed to determine the optimal
   Berberine concentration. H&E staining revealed that Berberine treatment
   preserved cell density of the inner retinal layers, particularly the
   GCL, which is typically reduced in following IR injury (Fig. [136]6B
   and C). As shown in Fig. [137]6D and E, Brn3a immunostaining revealed a
   significant preservation of RGCs following treatment with Berberine at
   a dose of 2.5 mM, compared to controls.

   We further performed molecular docking analysis to predict the
   potential interaction between Berberine and IRGM1, which revealed that
   Berberine stably fits into the active pocket of IRGM1, forming multiple
   hydrogen bonds and hydrophobic interactions, suggesting a potential
   direct regulatory effect on IRGM1 function (Fig. [138]6F).

   Taken together, these results demonstrate that intravitreal
   administration of Berberine provides significant structural and
   functional protection to RGCs in the IR-induced acute glaucoma model,
   likely through modulation of IRGM1-related pathways and inhibition of
   apoptosis.

Neuroprotection of apoevs combined with Berberine (Ber-ApoEVs) contributes to
treatment of IR injury

   Given the individual neuroprotective properties of ApoEVs and
   Berberine, we hypothesized that their combination might exert additive
   effects in attenuating retinal IR injury. Based on the shared
   modulation of IRGM1—a central mediator of inflammation—we generated
   Ber-ApoEVs for intravitreal delivery.

   Berberine was loaded into ApoEVs using a mild sonication protocol
   followed by ultrafiltration (Fig. [139]6G). Drug loading efficiency was
   determined via HPLC, revealing a final Berberine concentration in the
   Ber-ApoEV suspension (Fig. [140]6H). Successful encapsulation was
   confirmed by NTA, which showed no significant changes in size or
   particle concentration compared to native ApoEVs (Fig. [141]6I). TEM
   analysis confirmed the typical cup-shaped morphology of Ber-ApoEVs,
   comparable to native ApoEVs, with no signs of aggregation or structural
   damage (Fig. [142]6J).

   Mice received intravitreal co-administration of Berberine and ApoEVs
   (as separate or combined Ber-ApoEVs) (Fig. [143]7A). H&E staining
   demonstrated superior preservation of retinal architecture,
   particularly in the GCL, in the combination group relative to
   single-agent groups. Cell density in GCL was significantly greater,
   suggesting effective structural protection against IR-induced damage
   (Fig. [144]7B and D). Furthermore, Brn3a immunostaining revealed that
   the combination treatment led to a significantly higher density of
   surviving RGCs compared to either ApoEVs or Berberine alone
   (Fig. [145]7C and E).

Fig. 7.

   [146]Fig. 7
   [147]Open in a new tab

   Combined neuroprotective effects of BMSCs-derived ApoEVs and Berberine
   in retinal IR injury. (A) Experimental timeline of intravitreal
   injection of ApoEVs, Berberine, or Ber-ApoEVs. (B) Representative H&E
   stained retinal sections. (C) Representative immunofluorescence images
   of Brn3a-positive retinal ganglion cells in treated groups. Scale bar
   = 100 μm. (D) Quantification of percentage of GCL cell number compared
   to Ctrl/PBS group from H&E stained sections at 3dpi. (E) Quantification
   of Brn3a-positive RGCs density of treatment groups compared to Ctrl/PBS
   group at 3dpi. (F) Representative VEP waveforms recorded at 3dpi. (G)
   Quantification of VEP signal amplitudes. (H,I) Western blot analysis
   and quantification of ERK1/2 and p-38 protein expression in retinas
   after treatments. (J) qRT-PCR analysis of inflammatory and
   neuroprotective gene expression levels in retina tissue across
   treatment groups. All quantitative data are shown as mean ± SD. Group
   comparisons were analyzed using the Kruskal–Wallis test followed by
   Dunn’s multiple comparisons. *P < 0.05, **P < 0.01, ***P < 0.001,
   ****P < 0.0001

   Electrophysiological assessments using VEP further confirmed functional
   improvements. The combined treatment group exhibited the highest VEP
   signal amplitudes at 3dpi, suggesting improved retinal signal
   transmission and RGCs excitability (Fig. [148]7F and G).

   At the molecular level, Western blot analysis showed inhibition of p38
   after both ApoEV/Ber alone as well as combined treatment (Fig. [149]7H
   and I), while qRT-PCR indicated a more pronounced downregulation of
   Caspase1, Bst2 and Il-6, and a concomitant upregulation of Bcl-2 and
   Igf-1 in the combination group, compared to either treatment alone
   (Fig. [150]7J). These data suggest that Berberine and ApoEVs may act on
   overlapping but complementary anti-inflammatory and anti-apoptotic
   pathways to mediate retinal protection, particularly through the
   p38/MAPK pathway.

   Collectively, these results demonstrate that Ber-ApoEVs offers enhanced
   neuroprotection in the IR-injured retina through structural
   preservation, functional rescue, and transcriptional modulation of key
   inflammatory and apoptotic regulators.

Discussion

Characteristics of apoevs compared to other extracellular vesicles

   Apoptosis is a major form of programmed cell death, orchestrated by
   highly conserved molecular mechanisms that play a pivotal role in
   development, immune regulation, and tissue homeostasis, as well as in
   various pathological processes [[151]14]. Notably, apoptosis has been
   consistently observed as a predominant mode of RGCs death in glaucoma,
   highlighting its relevance in disease progression [[152]13]. During
   apoptosis, cells undergo a sequence of regulated events including
   membrane blebbing, caspase activation, chromatin condensation, and DNA
   fragmentation, ultimately leading to the formation of ApoEVs [[153]48].
   These vesicles carry fragmented DNA, RNAs, lipids, and specific
   proteins, and are actively released into the extracellular space, where
   they influence the behavior and fate of neighboring cells [[154]49].

   Human BMSCs represent a reliable and clinically relevant source of EVs,
   owing to their low immunogenicity, strong immunomodulatory capacity,
   and established regenerative potential, making them highly suitable for
   allogeneic therapeutic applications [[155]50]. EVs can be broadly
   classified into exosomes, microvesicles, and ApoEVs, based on their
   size and biogenesis [[156]51]. Among these, ApoEVs are the largest
   (ApoEVs: 50–5000 nm, exosomes: 50–150 nm, micro-vesicles: 50–1000 nm)
   [[157]49]. Beyond size, they differ in origin, morphology, membrane
   composition, and cargo content [[158]22]. While exosomes and
   micro-vesicles are typically released by viable cells through endosomal
   pathways or membrane budding, ApoEVs are uniquely released during
   apoptosis, often containing apoptotic markers such as
   phosphatidylserine [[159]52]. Importantly, due to their derivation from
   dying cells, ApoEVs possess inherently lower immunogenicity compared to
   EVs from viable cells, reducing the risk of triggering detrimental
   immune responses when used therapeutically.

   Interestingly, despite their origin from apoptotic processes, ApoEVs do
   not simply function as passive by-products of cell death; rather, they
   are active mediators of intercellular communication and tissue
   remodeling [[160]22, [161]53]. This phenomenon may be attributed to the
   fact that apoptosis is a tightly regulated and non-inflammatory
   process, during which ApoEVs selectively package and transmit bioactive
   signals that facilitate immune resolution, cellular adaptation, and
   tissue repair. Emerging evidence has revealed their pleiotropic
   effects, including anti-tumor, anti-inflammatory, and tissue
   repair-promoting functions [[162]20, [163]48, [164]54]. These seemingly
   paradoxical properties likely stem from their unique molecular cargo,
   which allows ApoEVs to convey regulatory signals that promote immune
   resolution and cellular regeneration, rather than propagating
   apoptosis. In this context, we investigated BMSCs-derived ApoEVs in a
   glaucoma-related IR model, focusing on their capacity to preserve RGCs
   viability.

   Given previous studies highlighting the neuroprotective properties of
   EVs [[165]55, [166]56], including ApoEVs, and considering that glaucoma
   is characterized by progressive neurodegeneration marked by the loss of
   RGCs, we investigated the potential of ApoEVs to preserve RGCs
   viability in a glaucoma-related IR model.

Regulation of inflammation by apoevs and the mechanism of action

   Pathogenic mechanisms in glaucoma involve ischemia, oxidative stress,
   hypoxia, excitotoxicity, mitochondrial dysfunction and
   neuroinflammation [[167]57, [168]58]. Based on what was recently been
   documented, modulating local immune responses and targeting
   inflammatory pathways could offer new therapeutic strategies for
   glaucoma management.

   ApoEVs have been shown to enhance the expression of anti-apoptotic and
   pro-survival factors while simultaneously suppressing the activation of
   pro-inflammatory cytokines, which may play a critical role in
   mitigating the detrimental effects of IR injury in the retina. Our
   study found that ApoEVs promoted RGCs survival by modulating
   inflammatory responses, partly via targeting IRGM1, a protein
   intricately linked with both autophagy and immune regulation [[169]59].
   In disease models, IRGM1 exhibits context-dependent effects: while it
   negatively regulates NLRP3 inflammasome activation and maintains
   intestinal or immune homeostasis in certain inflammatory settings
   [[170]60], its overexpression has also been associated with tissue
   damage, including enhanced oxidative stress and promoting MAPK
   activation [[171]26, [172]27]. In our study, we found that IRGM1
   expression was markedly upregulated following IR injury and exerted
   detrimental effects on the RGCs, suggesting it may serve as a
   pathological target mediating inflammation and cell death in
   glaucomatous neurodegeneration. While the role of IRGM1 has been
   extensively studied in various contexts, its function in glaucoma has
   not been previously reported. To our knowledge, this study provides
   preliminary evidence suggesting a potential role of IRGM1 in the
   pathophysiology of glaucoma.

   In our study, ApoEVs could significantly modulate inflammatory pathways
   to promote RGCs survival. Specifically, pathway enrichment analysis
   based on our retinal RNA sequencing data suggests that ApoEVs may exert
   their effects, at least in part, by modulating the MAPK signaling
   pathway, which is critically involved in regulating inflammation, cell
   survival, and apoptosis [[173]61, [174]62]. The MAPK pathway has been
   implicated in the progression of glaucomatous optic neuropathy, where
   it regulates the inflammatory response and cellular damage in RGCs
   [[175]63, [176]64]. Moreover, A recent study demonstrated that MSCs-
   ApoEVs promote cartilage regeneration by modulating the immune
   microenvironment through MAPK signaling, supporting the notion that
   ApoEVs may exert anti-inflammatory effects via MAPK-related pathways
   [[177]65]. The precise mechanisms through which ApoEVs regulate these
   pathways remain to be fully elucidated, but our results indicate that
   they exert protective effects by suppressing inflammatory damage and
   promoting cellular homeostasis.

   Previous studies on ApoEVs have predominantly focused on their role in
   tissue regeneration and restoring homeostasis [[178]17, [179]27].
   However, our research highlights the anti-inflammatory effects of
   ApoEVs, suggesting their potential in mitigating neuroinflammation in
   other neurodegenerative diseases, such as Alzheimer’s disease and
   Parkinson’s disease. Future investigations should expand on these
   findings to explore the broader therapeutic applications of ApoEVs in
   modulating neuroinflammatory responses across various neurodegenerative
   disorders.

Synergistic effect of apoevs and Berberine in enhancing neuroprotection

   In this study, the combination of ApoEVs with Berberine, a natural
   alkaloid with well-established neuroprotective properties, was found to
   enhance the protective effects against RGCs damage in the IR model.
   Berberine has been shown to exert neuroprotection by modulating
   inflammation, oxidative stress, and apoptosis [[180]66, [181]67]. In
   addition to its neurological benefits, Berberine has emerged as a
   promising therapeutic agent for cardiovascular and metabolic diseases
   (CVMD), owing to its multi-target pharmacological activities validated
   in both basic and clinical studies [[182]68]. Its anti-inflammatory
   mechanisms primarily involve the suppression of inflammatory cytokine
   production (e.g., IL-1β, IL-6, TNF-α) and inhibition of key signaling
   pathways such as NF-κB and MAPK signaling [[183]28]. These findings
   provide important molecular insights into Berberine’s therapeutic
   potential in inflammatory diseases. In ophthalmologic research,
   Berberine has been reported to suppress inflammation and apoptosis via
   the MAPK pathway in dry eye disease [[184]69], however, its role in
   glaucoma has not yet been reported.

   When Berberine was co-incubated with ApoEVs, we observed a synergistic
   effect that was more pronounced than either treatment alone. This
   suggests that the two agents may act through complementary mechanisms.
   ApoEVs, by modulating inflammatory pathways and promoting cell
   survival, appear to work in concert with Berberine’s antioxidant and
   anti-inflammatory effects to enhance the overall neuroprotective
   response. The combination treatment may offer a promising therapeutic
   approach for conditions where neuroinflammation and cell death play a
   critical role, such as glaucoma.

   This finding opens up the potential for further studies into the use of
   ApoEVs in combination with other pharmacological agents to maximize
   their therapeutic benefits. Exploring the mechanisms underlying this
   synergistic effect will be crucial for optimizing the clinical
   application of such combination therapies.

Translational potential of apoevs in clinical practice

   The translational potential of ApoEVs in clinical settings is an
   exciting prospect for treating retinal diseases, particularly glaucoma.
   Given their ability to cross biological barriers and their
   immunomodulatory properties, ApoEVs could be used as a therapeutic
   strategy to promote cell survival and reduce inflammation in the
   retina. Current treatments for glaucoma, such as IOP-lowering drugs, do
   not directly target the underlying neurodegenerative processes or
   inflammation [[185]70]. Thus, the development of therapies based on
   ApoEVs could provide a novel approach to preserving RGCs function and
   halting disease progression.

   In addition to their use in glaucoma, ApoEVs may also have applications
   in other retinal diseases, such as diabetic retinopathy and age-related
   macular degeneration, where inflammation and retinal cell death are
   prominent features. The ability to deliver therapeutic molecules in a
   highly targeted manner using ApoEVs could also enhance the efficacy of
   treatment and minimize off-target effects.

   However, several challenges remain in the clinical translation of
   ApoEVs. These include the efficient and scalable isolation of ApoEVs,
   as well as ensuring their stability, safety, and optimal delivery to
   target tissues [[186]71, [187]72]. Furthermore, the potential for
   immune responses to ApoEVs must be carefully evaluated to ensure their
   safety in human patients.

Conclusion

   In conclusion, our study demonstrates that BMSCs-derived ApoEVs exert
   significant neuroprotective effects in a glaucoma-related IR model,
   likely by modulating inflammatory responses and promoting RGCs
   survival. Mechanistically, these effects may involve the
   IRGM1/p-38/MAPK signaling axis, as suggested by our transcriptomic and
   validation analyses. Moreover, the synergistic benefits observed with
   the addition of Berberine indicate that combination therapies could
   further enhance the therapeutic potential of ApoEVs.

   Future research should focus on further elucidating the molecular
   mechanisms underlying the protective effects of ApoEVs, particularly
   their interaction with inflammatory pathways. Additionally, exploring
   the clinical application of ApoEVs, both alone and in combination with
   other therapeutic agents, holds great promise for treating glaucoma and
   other neurodegenerative diseases. As the field of EVs continues to
   evolve, it is important to address the challenges related to their
   production, stability, and delivery to realize their full clinical
   potential.

Supplementary Information

   [188]Supplementary Material 1^ (346.2KB, docx)

Acknowledgements