Abstract

   Ischemia-reperfusion (I/R) injury is associated with a variety of
   retinal diseases, resulting in loss of the number of ganglion cells
   (RGCs), retinal structural disorders, and retinal dysfunction. The
   Reelin protein is an important regulator of neuronal migration and
   synaptogenesis, and the Reln signaling pathway plays an essential role
   in regulating the targeted projection of RGC dendrites and neuronal
   survival, which has not been reported in retinal I/R injury. The aim of
   this study was to investigate the expression, role and mechanism of
   Reln in retinal I/R injury. By establishing Reln-CreERT2 mTmG
   transgenic mice, it was observed that the expression of Reln initially
   decreased and then increased after retinal I/R injury. After
   supplementing exogenous Reelin protein and adeno-associated virus
   (AAV)-targeted regulation of Reln in vivo, morphological and functional
   experiments demonstrated its effectiveness in protecting RGCs survival,
   maintaining morphological integrity of the retina, and inhibiting
   post-injury retinal dysfunction. Furthermore, integrin β1 (Itgb1) was
   identified as the main receptor through which Reelin exerts
   neuroprotective effects while regulating retinal I/R injury repair
   through the Dab1-PI3K-Akt pathway. These findings provide evidence
   supporting Reln pathway’s role in maintaining retinal homeostasis and
   facilitating injury repair. Moreover, these findings have significant
   implications for identifying new targets for preventing and treating
   various retinal diseases.

   graphic file with name 41419_2025_7742_Figa_HTML.jpg

   Subject terms: Retina, Apoptosis, Stroke

Introduction

   Retinal ischemia-reperfusion (I/R) injury is the pathological basis of
   a diverse spectrum of retinal diseases [[42]1], including glaucoma,
   retinal vascular occlusion, diabetic retinopathy and retinopathy of
   prematurity [[43]2–[44]4]. These ocular diseases cause serious vision
   disorders and have a significant socio-economic impact worldwide
   [[45]5]. After each episode of I/R, there is a progressive
   deterioration of the retinal structure and retinal cell loss,
   culminating in retinal ganglion cell (RGC) death and potential
   blindness. However, due to the lack of a clinically approved
   therapeutic regimen for retinal I/R injury, there is an urgent need for
   novel and efficacious treatment strategies [[46]6, [47]7].

   Obstruction of retinal blood circulation can lead to a decrease in
   oxygen and nutrient supply to retinal cells. Blood reperfusion induces
   oxidative stress, inflammation, immune response, and calcium overload
   in the retinal tissue, ultimately culminating in neuronal cell death
   [[48]8–[49]10]. Neuronal death can be triggered by different
   mechanisms, including apoptosis, iron death, and autophagy [[50]11].
   Apoptosis plays a crucial role in retinal I/R injury. It primarily
   operates via the mitochondria-mediated endogenous apoptosis pathway
   [[51]12]. Thus, targeting the expression of key apoptotic proteins in
   this pathway may offer preventative and therapeutic benefits against
   retinal I/R damage.

   Reelin, an extracellular matrix protein, is recognized for its role in
   promoting neuronal survival through the inhibition of apoptosis
   [[52]13, [53]14]. Reelin-deficient mice show impaired neurogenesis and
   increased stroke size [[54]15], whereas those with Reln gene knock-in
   display activation of the downstream protein target Dab1 [[55]16].
   Reelin signals via receptors such as VLDLR, ApoER2 and Itgb1 [[56]17,
   [57]18], thereby initiating intracellular pathways through Dab1
   phosphorylation and PI3K-Akt activation [[58]19, [59]20]. Within the
   Reelin signaling pathway, Dab1 activation can trigger a cascade of
   events that recruit or activate various downstream factors, including
   Src family tyrosine kinases (SFK), which ultimately counteract neuronal
   apoptosis [[60]21]. Despite these insights, the potential of Reelin to
   protect central neurons in the retina following I/R injury remains
   unexplored. This study, therefore, endeavors to elucidate the
   cytoprotective role of Reelin in the retinal I/R process and to dissect
   the underlying mechanisms.

Materials and methods

   More details are provided in the online-only Data Supplement.

Animals

   Healthy male C57BL/6J wild-type mice, aged 6–8 weeks and weighing
   16–22 g, were obtained from Beijing Spefo Laboratory Animal Technology
   Co., Ltd, China. They were housed in a pathogen-free, temperature
   (22–24 °C) and humidity (40–70%) controlled environment with a 12-h
   light/dark cycle, and provided with water and food ad libitum.

Generation of genetically modified mice

   Reln-CreERT2 mTmG Mice: Generated using CRISPR-Cas9 and Cre-LoxP
   systems, these mice were sourced from Beijing Biocytogen Co., Ltd,
   China, and feature a Reln-CreERT2 line interbred with ROSAmT/mG for
   cell tracing. Initially expressing tdTomato fluorescence, Cre-mediated
   recombination switches expression to green fluorescent protein (GFP) in
   Reln-expressing cells. Male mice aged 6–8 weeks were used, with
   tamoxifen administered intraperitoneally 2 weeks prior to experiments
   to induce fluorescence shift.

   Retinal Localized Reln knockdown Mice: A localized Reln gene knockdown
   was induced in C57BL/6J wild-type mice using adeno-associated virus
   (AAV-2/9-shRNA-Reln; HANBIO, Wuhan, China). Healthy males aged 6–8
   weeks were injected with AAV solution (2 μL, 1.0 × 10^12 vg/mL) into
   the vitreous body 3 weeks before experiments. The AAV vector encoded
   EGFP for tracking, with the siRNA sequence targeting Reln being
   CCAGGATACATGATGCAATTT.

Retinal I/R model establishment

   We induced retinal I/R injury in mice using a 33-gauge needle to
   maintain intraocular pressure (IOP) at 110 mmHg for 60 min, following a
   balanced salt solution protocol [[61]22]. A sham group served as a
   control. After IOP normalization, mice were monitored with a heat pad
   and treated with tobramycin ointment (Alcon, Texas, USA) to prevent
   infection. Eyes were harvested at 24 h, 3, 7, 14, and 28 days post-I/R
   for analysis.

Experimental groups

   The experimental animals were randomly divided into three groups:

   Reelin Protein Group: Received 2 μL recombinant Reelin protein
   (300 ng/μL; R&D Systems, Minn, USA) or PBS in the vitreous cavity 24 h
   pre-I/R, forming Reelin and PBS groups, with post-modeling designations
   as -I/R counterparts. Dosage selection of Reelin protein is shown in
   Fig. [62]S1A, B.

   Reln knockdown Group: Received 2 μL AAV-shReln or AAV-EGFP in the
   vitreous cavity 3 weeks pre-I/R, forming AAV-shReln and AAV-EGFP
   groups, with post-modeling designations as -I/R counterparts.

   Alpelisib group: Mice were orally gavaged with placebo (0.5% Sodium
   carboxymethyl cellulose; MedChem Express, NJ, USA) or Alpelisib
   (50 mg/kg) 14 days prior to I/R. The treatment was administered once
   daily until tissue collection. This regimen was selected based on
   previous in vivo studies that demonstrated its efficacy and safety
   (Fig. [63]S1C) [[64]23].

Immunohistochemistry

   Mouse eyes were enucleated and fixed in a solution containing 4%
   paraformaldehyde at room temperature (RT) for 2 h followed by the
   incubation in 30% sucrose solution at 4 °C overnight and then embedded
   in an OCT compound. Cryosections of 5 µm-thickness were incubated with
   goat serum at RT for 1 h. After incubation with primary antibodies at
   4 °C overnight, the sections were incubated with secondary antibodies
   at RT for 2 h. Antibody information is described in Table [65]S1.
   Anti-fluorescence quenching sealer containing DAPI was used to seal the
   films (ab104139; Abcam, Cambridge, UK). Mouse eyeball sections across
   the optic nerve plane were selected. Images were taken with the Olympus
   VS200 Virtual Slide Scanning microscope (Olympus, Tokyo, Japan).
   Quantitative analysis of the inner retinal layer (from the internal
   limiting membrane to the inner edge of the outer plexiform layer) was
   performed using ImageJ software (National Institutes of Health, USA).

Quantitative real-time polymerase chain reaction (RT-qPCR)

   Mouse retinal tissues were processed for mRNA extraction using Trizol
   reagent (Invitrogen, Grand Island, NY, USA). The mRNA was subsequently
   reverse-transcribed into cDNA with the iScript cDNA Synthesis Kit
   (Bio-Rad, Hercules, CA, USA). RT-qPCR was conducted utilizing the iTaq
   Universal SYBR Green Supermix (Bio-Rad, Hercules, CA, USA) on a CFX96
   RT-qPCR detection system. Relative gene expression was calculated using
   the comparative 2-ΔΔCt method, with primer sequences detailed in
   Supplementary Table [66]S2.

Hematoxylin-eosin staining (H&E)

   The excised eyes were promptly fixed in 4% paraformaldehyde and
   subsequently embedded in paraffin. Retinal morphology was assessed
   through hematoxylin and eosin staining of 5-μm-thick cross-sections
   along the optic nerve axis. Measured the thickness of the inner retinal
   layer. The images were taken with a light microscope (Olympus).

Stretched preparation of the retina

   Retinas were fixed in 4% paraformaldehyde for 3 h, incised at the
   corneoscleral rim, and dissected to remove the cornea, iris, lens, and
   vitreous body. The optic nerve was separated from the sclera, and the
   retina was detached from the uvea. After removing pigmented tissue and
   rinsing in 0.1% PBST, retinas were dehydrated in methanol, blocked with
   goat serum, and incubated with the RGC marker RBPMS polyclonal antibody
   (Proteintech, Wuhan, China) overnight at 4 °C. Following secondary
   antibody incubation, retinas were mounted and sealed with
   DAPI-containing sealer. Photographic documentation and subsequent cell
   count analysis were conducted using a high-content imaging system
   (Operetta CLS, PerkinElmer, UK). The calculation of the entire retinal
   area was performed using ImageJ software (National Institutes of
   Health, USA). Antibody information is described in Table [67]S1.

Intravitreal injection

   Anesthetized mice received intraocular injections of 2 μL recombinant
   Reelin protein, PBS, AAV-shReln, or AAV-EGFP using a 34-gauge
   microinjector (Hamilton, Bonaduz, Switzerland). The injection site was
   the superior temporal quadrant, 1 mm from the corneal limbus, with care
   to avoid the lens and target the optic disc. Agents were delivered
   gradually into the vitreous cavity.

Optical coherence tomography (OCT)

   Pupils were dilated with compound tropicamide eye drops (Santen
   Pharmaceutical Co., Ltd, Osaka, Japan), followed by ophthalmic surface
   anesthesia using oxybuprocaine hydrochloride. A carbomer-based eye drop
   gel (Bausch & Lomb, New York, USA) was applied to the cornea. The
   ophthalmic ultramicrographic imaging system (Optoprobe, Leeds, UK) was
   calibrated for optimal retinal imaging via OCT. Retinal thickness,
   measured from the internal limiting membrane to the retinal pigment
   epithelium, was evaluated using proprietary software (version 2.0) from
   OptoProbe Research Ltd.

Flash electroretinogram (FERG)

   After a 12-h dark adaptation, mice were anesthetized, immobilized and
   mydriasis dilated. A gold-plated wire loop, in contact with the corneal
   surface, served as the active electrode, while stainless-steel needle
   electrodes were inserted between the ears and into the caudal skin,
   serving as reference and ground leads, respectively. The OPTO-III
   visual electrophysiology instrument (Optoprobe) was operated according
   to its standard protocol. FERG waveform analysis involved measuring the
   a-wave amplitude from baseline to the first negative peak and the
   b-wave amplitude from the a-wave trough to the subsequent positive
   peak.

Single-cell data analysis

   Sequencing data for this study were obtained from the Genome Sequence
   Archive at the Beijing Institute of Genomics (BIG) Data Center under
   project PRJCA008174 (GSA registration CRA006042). Data processing was
   performed using CellRanger (version 7.1.0), followed by quality control
   and analysis with Seurat (version 4.3.0.1). Cells with over 200
   detected genes and less than 20% mitochondrial gene content were
   included. After data merging, SCTransform normalization and PCA were
   conducted, with batch effects corrected using the harmony function.
   Clustering and visualization were based on the first 20 principal
   components using t-SNE. Marker genes were identified with Seurat’s
   FindAllMarkers function. DEG enrichment and pathway analysis were
   performed using the ClusterProfiler package (version 4.6.2) with the
   KEGG database. Single-cell data from [68]GSE137400 were used to
   validate the co-expression of Reln, Itgb1, and Dab1 in RGCs, visualized
   using Seurat’s FeaturePlot. Protein interactions were predicted with
   the STRING database and visualized using Cytoscape (version 3.10.1).
   Intercellular communication was analyzed using the CellChat package
   (version 2.1.0) in R, which quantifies communication networks and
   predicts dominant signaling pathways. Detected ligand-receptor pairs
   were integrated into a protein-protein interaction network for further
   analysis. CellChat estimated the probabilities of cell interactions,
   enabling the identification of significant biological communication
   patterns. The intercellular signaling data were synthesized into a
   circular graph for visual representation.

TUNEL labeling

   Retinal cryosections were fixed, permeabilized, and subjected to TUNEL
   staining using the Beyotime kit (C1090; Jiangsu, China). After
   incubation with TDT enzyme and fluorescent labeling solution, sections
   were mounted with DAPI-containing medium and visualized with an Olympus
   VS200 microscope. Cell quantification was performed with ImageJ
   software (National Institutes of Health, USA).

Western blot (WB) analysis

   At 3 days post-I/R injury, retinas were harvested and lysed in RIPA
   buffer containing protease inhibitors. Protein extracts were quantified
   using a BCA assay kit (Promega, Madison, WI, USA). Proteins were
   separated by 10% SDS-PAGE and transferred to nitrocellulose membranes.
   Following a 30-min block with 5% skimmed milk, membranes were incubated
   with primary antibodies overnight at 4 °C, then with HRP-conjugated
   secondary antibodies for 2 h at RT. Immunoreactive bands were
   visualized using the ChemiDoc™ MP Imaging System (Bio-Rad), and band
   intensities were quantified with ImageJ software (National Institutes
   of Health, USA). Antibody information is described in Table [69]S1.

Statistical analysis

   Sample sizes are indicated in the figure legends and were consistent
   with previous studies. Before data analysis, normal distribution test
   and variance homogeneity test were performed. All data were expressed
   as mean ± SD. The comparison between the two groups is involved. If the
   data conform to normal distribution, the t-test is used. If the data
   does not conform to the normal distribution, the nonparametric
   Mann–Whitney test is used. Comparisons between multiple groups (n ≥ 3)
   were involved. If the data were normally distributed, one-way ANOVA
   (with Dunnett’s multiple-comparisons test) was used. If the data does
   not conform to the normal distribution, the nonparametric
   Kruskal-Wallis test is used. Statistical analyses and graphs were
   generated using GraphPad Prism 9 statistical software (GraphPad
   Software, Inc., San Diego, CA, USA). Statistical significance was
   considered at P < 0.05, and all experiments were repeated at least
   three times.

Results

Characteristics of expression changes of Reln gene after injury

   The successful establishment of a retinal I/R injury model in mice was
   validated through multiple assessment methods. Immediately post-I/R,
   pronounced corneal edema was observed (Fig. [70]S2A), and H&E staining
   indicated maximal retinal thinning at 7 days, followed by stabilization
   (Fig. [71]1A, B). Fundus photography captured intermittent retinal
   blood flow, peripheral atrophy, and visible choroid 7 days post-I/R
   (Fig. [72]1C). Immunofluorescence revealed a gradual decline in RGC
   numbers, stabilizing by 7 days post-injury (Fig. [73]S2B, C).

Fig. 1. Expression of Reln in the retina after I/R injury.

   [74]Fig. 1
   [75]Open in a new tab

   A Histopathological examination of retinal tissue stained with H&E
   showed significant edema in the inner retina layer 24 h after I/R
   injury, with loose and disordered cell arrangement. After 3 days, the
   inner retinal layer gradually became thinner, and the number of cells
   decreased. Scale bar: 80 μm. B Quantification of inner retinal
   thickness. C Fundus photography shows intermittent blood supply to
   retinal blood vessels after I/R injury, peripheral retinal atrophy, and
   visible choroid membrane. D Reln-CreERT2 mTmG mice expressed
   Reln-positive cells and their progeny at various time points after
   retinal I/R injury. Scale bar: 40 μm or 10 μm. E Representative images
   of Reelin immunofluorescence staining at various time points after
   retinal I/R injury in wild-type mice, which can confirm the expression
   of reelin protein in RGCs. Arrows indicate Reelin-positive cells. Scale
   bar: 40 μm. F Percentage of Reln-positive cells in the inner retina. G
   The expression level of Reelin protein in the inner retinal layer was
   assessed by the intensity of the fluorescence signals. H Following
   retinal I/R injury, the mRNA expression of the Reln gene initially
   decreased and subsequently increased, but it remained lower than the
   normal baseline level. Data are represented as means ± SD (n = 6).
   *P < 0.05, **P < 0.01, ***P < 0.001.

   Utilizing Reln-CreERT2 mTmG mice, which exhibit permanent GFP
   fluorescence in Reln-expressing cells and their progeny post-tamoxifen
   induction, we assessed Reln gene expression under normal and
   I/R-injured conditions. Confocal microscopy of frozen sections
   demonstrated GFP signal from Reln-positive cells in the inner nuclear
   layer, inner plexiform layer, and ganglion cell layer (GCL). Analysis
   of fluorescence in these mice at various post-I/R time points showed a
   transient decrease in Reln-positive cells in the inner retina, nadir at
   7 days, with a subsequent increase yet remaining below control levels
   up to 28 days post-I/R (Fig. [76]1D, F). This trend aligned with the
   results of Reelin immunofluorescence staining in wild-type mice. And
   the localization of Reelin protein in RGCs could be determined by
   observing the distribution of fluorescence signals (Figs. [77]1E, G,
   [78]S2D). In addition, RT-qPCR confirmed a significant downregulation
   of Reln gene expression at 24 h and 3 days post-I/R, with a subsequent,
   albeit insufficient, recovery by 7, 14, and 28 days (Fig. [79]1H).
   These findings suggest an acute reduction in Reln gene expression
   following retinal I/R injury, implicating its role in injury response
   and warranting further investigation.

Reelin protein promotes retinal tissue damage repair

   Considering the initial downregulation of Reelin expression during the
   early stages of retinal I/R injury, we introduced exogenous recombinant
   Reelin protein into the vitreous cavity 24 h before injury induction to
   assess its role in retinal repair. Control mice received PBS injections
   (Fig. [80]2A). Retinal assessment at the peak damage period (7 days
   post-I/R) revealed significant inner retinal thinning in I/R-injured
   mice, which was mitigated in Reelin-supplemented mice as demonstrated
   by H&E staining (Fig. [81]2B, C). OCT further confirmed Reelin’s
   protective effect on overall retinal thickness post-injury (Fig.
   [82]2D, E). Immunofluorescence staining indicated a preservation of
   RGCs in Reelin-treated mice, evidenced by increased RGC counts compared
   to the PBS-I/R group (Fig. [83]2F, G). FERG assessments reflected a
   decline in a-wave and b-wave amplitudes post-I/R injury, which were
   significantly improved in the Reelin-I/R group (Fig. [84]2H, I).
   Moreover, the b-wave amplitude was slightly increased in the uninjured
   Reelin group compared to the PBS group, but this difference did not
   exhibit statistical significance. In conclusion, the supplementation of
   Reelin protein demonstrated therapeutic potential in facilitating
   retinal repair and may serve as a promising intervention for retinal
   ischemic diseases.

Fig. 2. Reelin protein rescues retinal damage following I/R injury both
structurally and functionally.

   [85]Fig. 2
   [86]Open in a new tab

   A Schematic representation of the experimental design (Created with
   BioRender.com). B Histopathological examination of retinal tissue
   stained with H&E showed that the inner retinal thickness of mice was
   significantly reduced after I/R injury. This I/R-induced inner retinal
   thinning was significantly improved in mice treated with exogenous
   Reelin protein. Scale bar: 80 μm. C Quantification of inner retinal
   thickness (n = 6). D Representative OCT images of retinal tissue
   demonstrate changes after I/R injury. E Quantification of total retinal
   thickness (n = 6). F Representative fluorescent staining of retinal
   patches after I/R injury showed the RBPMS-positive RGCs (orange)
   distribution. Supplementation with Reelin protein exerted a significant
   protective effect on I/R-induced RGC injury. Scale bars: 100 μm or
   1 mm. G Number of RBPMS-positive RGCs (n = 4). H Representative
   oscillatory potential waveforms of visual function were assessed via
   FERG 7 days after I/R injury. I Quantification of the magnitude of a
   and b waves (n = 6). Data are represented as means ± SD. *P < 0.05,
   **P < 0.01, ***P < 0.001, ****P < 0.0001.

Reln signaling pathway regulates retinal tissue damage repair

   To investigate the Reln signaling pathway’s role in I/R injury, we
   developed a mouse model with localized retinal Reln knockdown. We
   employed AAV-mediated intravitreal injection to locally knock down the
   retinal Reln gene. Three weeks post-injection, immunofluorescence
   staining and RT-qPCR confirmed effective knockdown of Reln expression
   in retinal tissues (Fig. [87]3A, B). The knockdown model was validated
   (Fig. [88]3C, D), and the impact of I/R injury was assessed at 7 days
   post-injury.

Fig. 3. AAV-shReln-mediated local knockdown of Reln aggravated retinal
thickness thinning, the reduction in RGC count, and retinal dysfunction in
mice caused by I/R injury.

   [89]Fig. 3
   [90]Open in a new tab

   A Schematic representation of the experimental design (Created with
   BioRender.com). B Frozen slides showing retinal infection 3 weeks after
   intravitreal injection of AAV. EGFP carried by AAV was expressed in
   nearly all the layers of the retina. Scale bar: 40 μm. C The expression
   level of Reelin protein in the inner retinal layer was assessed by the
   intensity of the fluorescence signals (n = 6). D The mRNA expression of
   the Reln gene (n = 6). E Histopathological examination of retinal
   tissue via H&E staining showed that the inner retinal thickness in
   AAV-shReln mice was further reduced after I/R injury. The damage to
   I/R-induced RGCs was further aggravated in AAV-shReln mice. Scale bar:
   80 μm. F Quantification of inner retinal thickness (n = 6). G
   Representative OCT images of retinal tissue changes after I/R injury. H
   Quantification of total retinal thickness (n = 6). I Representative
   fluorescent staining of retinal patches after I/R injury showed the
   distribution of RBPMS-positive RGCs (orange). Scale bars: 100 μm or
   1 mm. J Number of RBPMS-positive RGCs (n = 4). K Representative
   oscillatory potential waveforms of visual function were assessed by
   FERG after I/R injury. L Quantification of the magnitude of a and b
   waves (n = 6). Data are represented as means ± SD. *P < 0.05,
   **P < 0.01, ***P < 0.001.

   H&E staining at this time point revealed a significant decrease in
   inner retinal thickness in I/R-injured mice, which was exacerbated in
   the Reln knockdown group (Fig. [91]3E, F). OCT confirmed a reduction in
   overall retinal thickness post-injury, particularly in Reln knockdown
   mice (Fig. [92]3G, H). Immunofluorescence staining demonstrated a
   decreased RGC density and count in the AAV-shReln-I/R group compared to
   the AAV-EGFP-I/R group (Fig. [93]3I, J). Although a minor decrease in
   retinal thickness and RGC count was observed in the uninjured
   AAV-shReln group, it was not statistically significant. Functional
   assessment using FERG showed a significant reduction in a and b wave
   amplitudes in the AAV-shReln-I/R group compared to the AAV-EGFP-I/R
   group. Interestingly, a-wave and b-wave amplitudes were also reduced in
   the uninjured AAV-shReln group, indicating a role for Reln in
   maintaining retinal function (Fig. [94]3K, L). Collectively, these
   results underscore the Reln pathway’s regulatory function in retinal
   homeostasis and its protective role in I/R injury.

Reln is primarily expressed in RGCs and cone bipolar cells (CBCs) and
participates in the PI3K-Akt pathway after I/R injury

   Single-cell transcriptomic analysis was performed on retinal tissues
   from mice in the blank group and mice subjected to injury 3 days after
   retinal I/R. Based on the expression profiles of maturation markers, we
   identified 12 distinct cell clusters: rods, CBCs, cones, macroglia, rod
   bipolar cells (RBCs), horizontal cells, RGCs, amacrine cells,
   neutrophils, vascular endothelial cells (VECs), microglia (myeloid), T
   cells, and dendritic cells (Fig. [95]4A). The Reln gene was
   predominantly expressed in retinal RGCs and CBCs, and its overall
   expression level decreased after injury (Figs. [96]4B, [97]S3A).
   Furthermore, the expression of the Reln gene decreased in RGCs and CBCs
   following I/R injury, while it increased in VECs (Fig. [98]4C).
   Subsequently, we investigated the impact of I/R on cellular
   interactions and found that RGCs, CBCs, and VECs were the primary
   sources of ligands in the Reln signaling network, operating in both
   autocrine and paracrine modes. I/R led to a reduction in the Reln
   pathway gene expression between RGCs and CBCs and increased expression
   between RGCs and VECs (Fig. [99]4D). In addition, the extracellular
   matrix receptor, Integrin α3β1, is posited as a potential Reelin
   protein target in RGCs (Fig. [100]4E). KEGG pathway analysis indicates
   that the Reln gene is primarily involved in the PI3K-Akt pathway, which
   is associated with the heightened activities in pathways of responding
   to extracellular signals, promoting metabolism, proliferation, and cell
   survival following I/R injury (Fig. [101]4F). Afterward, we analyzed
   the interactions between Dab1, a key factor in the Reln pathway, and
   the related proteins mentioned above. The protein-protein interaction
   (PPI) network showed that there can be interactions between Reln,
   Itgb1, and Dab1, and both Itgb1 and Dab1 can activate the PI3K-Akt
   pathway (Fig. [102]S3B). Due to the small number of RGCs in this
   dataset, expanded analysis utilizing RGC single-cell transcriptome data
   identified 45 molecular clusters [[103]24] with widespread Reln
   expression and co-expression of Itgb1, Dab1, and Reln (Figs. [104]S3C,
   [105]4G). And we verified the expression of Itgb1 and Dab1 in RGCs by
   immunofluorescence (Fig. [106]4H).

Fig. 4. Reln was enriched in the PI3K-Akt pathway in RGC cells of the I/R
group.

   [107]Fig. 4
   [108]Open in a new tab

   A A dot plot based on the marker genes of each cluster (right). The
   t-SNE map shows 12 distinct clusters of retinal cells, each colored dot
   represents a cell. The size of each circle in the dot plot represents
   the proportion of gene expression, and the color represents the level
   of expression. B Violin Plot shows Reln gene expression in 12 different
   cell clusters. C Violin Plot shows Reln gene expression in RGC, CBC,
   and VEC clusters in the blank and I/R groups. D Circle chart shows the
   differences in the inferred Reln signaling networks between the blank
   and I/R groups. E Heatmap shows the role and importance of each cluster
   in the Reln signaling network (top). Bar charts show the relative
   contribution of each ligand-receptor pair in the Reln signaling pathway
   (bottom). F KEGG pathway enrichment analysis of RGCs in the I/R group
   (P < 0.05). G t-SNE charts show the expression of the Reln, Itgb1 and
   Dab1 genes in different RGC cell clusters, with the color representing
   the expression level. H Representative images of Itgb1 and Dab1
   immunofluorescence staining in normal mice retinas.

   In addition, we confirmed, through immunofluorescence staining, that
   the Reln gene was primarily expressed in CBCs rather than RBCs (Fig.
   [109]S3D). CBCs could be further classified into three subclusters
   (SCs), with differential subcluster expression patterns observed
   between control and I/R groups (Fig. [110]S3E–G). A novel subcluster,
   CBC-SC2, emerged post-I/R, potentially contributing to explaining the
   increase in Reln-positive cells in later stages of injury (Fig.
   [111]S3H). These results collectively indicate that the Reln gene,
   predominantly expressed in RGCs and CBCs, is involved in the PI3K-Akt
   signaling pathway, and its dysregulation post-I/R injury may disrupt
   retinal cell signaling, leading to associated retinal tissue damage.

Reelin protein inhibits apoptosis after I/R injury through the Dab1-PI3K-Akt
pathway

   As previously demonstrated, Reelin protein reduced RGC death in the
   mouse retina post-I/R (Fig. [112]2F, [113]G), and single-cell
   sequencing indicated activation of the PI3K-Akt pathway following I/R
   injury (Fig. [114]4F). To explore the mechanism of Reelin-induced
   neuroprotection, we evaluated its impact on apoptosis via TUNEL
   staining (Fig. [115]5A). TUNEL staining revealed peak retinal cell
   apoptosis at 24 h post-I/R (Fig. [116]S4A, B). In contrast to the
   negligible TUNEL-positive cells in the uninjured group, a significant
   increase was observed 24 h post-I/R, with a substantial decrease in
   TUNEL-positive cells in both the GCL and inner retina in the Reelin-I/R
   group (Fig. [117]5B). This suggests that Reelin effectively attenuates
   retinal cell apoptosis post-injury.

Fig. 5. After retinal I/R injury, Reelin protein inhibits apoptosis through
the Dab1-PI3K-Akt pathway.

   [118]Fig. 5
   [119]Open in a new tab

   A Representative TUNEL staining of frozen retinal sections 24 h after
   I/R injury. Scale bar: 40 μm. B Quantification of TUNEL-positive cells
   in the GCL (left) and inner retinal layer (right), (n = 6). C, D
   Representative blot images of the effects of Reelin protein therapy on
   retinal Itgb1, Src, p-Src, Fyn, p-Fyn, Dab1, p-Dab1, PI3K, Akt, p-Akt,
   GSK3b, p-GSK3b, Bcl-2, Bax and Cle-Caspase3 after I/R injury. E
   Quantitative analysis of the levels of the aforementioned proteins. F
   Representative blot images showing the effects of PI3K inhibitors on
   Akt, p-Akt, GSK3b, p-GSK3b, Bcl-2, Bax, and Cle-Caspase3 in the retina
   following I/R injury treated with Reelin protein. G Quantitative
   analysis of the levels of the aforementioned proteins (n = 3). Data are
   represented as the means ± SD. *P < 0.05, **P < 0.01, ***P < 0.001,
   ****P < 0.0001.

   To further explore the anti-apoptotic mechanism of Reelin protein,
   combined with single-cell transcriptome analysis results (Fig. [120]4E,
   F), we performed WB analysis of the samples, revealing increased Itgb1
   expression post-injury in the presence of exogenous Reelin (Fig.
   [121]5C, E). Activation of Fyn, Src, Dab1 and PI3K was observed in the
   Reelin-I/R group, with no change in Akt and GSK3b levels, but a notable
   increase in phosphorylated proteins p-Akt and p-GSK3b(Ser9) compared to
   the PBS-I/R group. Additionally, the Reelin-I/R group exhibited
   significantly lower expression of apoptosis-related protein Caspase-3
   and a reduced Bax/Bcl-2 ratio (Fig. [122]5D, E). Collectively, these
   findings indicate that Reelin protein supplementation post-retinal I/R
   injury is neuroprotective, promoting Dab1 phosphorylation, activating
   the PI3K-Akt-GSK3b pathway, and inhibiting retinal cell apoptosis. In
   addition, we performed a detailed analysis of the protein expression
   levels of p35 and p25, two key regulatory proteins of cyclin-dependent
   kinase 5 (Cdk5). Our results indicated that I/R injury significantly
   increased the P25/P35 ratio, suggesting the aberrant activation of the
   Cdk5 pathway. However, Reelin protein failed to counteract this
   increase. This suggests that under I/R-induced injury conditions,
   Reelin did not exert repair or regulatory functions via the Cdk5
   pathway (Fig. [123]S5A, B).

The PI3K inhibitor Alpelisib reverses the anti-apoptotic effect of Reelin
protein

   To further elucidate whether Reelin protects retinal cells from
   I/R-induced damage via the PI3K-Akt pathway, the PI3K inhibitor
   Alpelisib was administered orally to mice, and its effects on
   Reelin-mediated protection of retinal cells against I/R-induced damage
   were investigated. Our study revealed that Alpelisib administration
   significantly downregulated the expression of p-Akt protein (Fig.
   [124]5F, G), thereby confirming its inhibitory effect on the PI3K-Akt
   pathway. Furthermore, Alpelisib reversed the inhibitory effects of
   Reelin protein on I/R-induced pro-apoptotic factors, including GSK3b,
   Bax, and Cleaved-Caspase3. Conversely, the expression of the
   anti-apoptotic protein Bcl-2 was significantly reduced (Fig. [125]5F,
   G). These findings demonstrate that Alpelisib administration markedly
   diminished the protective effects of Reelin protein on I/R-induced
   retinal cells and promoted cell apoptosis. Collectively, these results
   further support the notion that Reelin protein exerts its protective
   role against I/R-induced retinal cell damage by activating the PI3K-Akt
   pathway.

Reln deficiency promotes apoptosis through the Dab1-PI3K-Akt pathway

   As previously shown, Reln gene knockdown intensifies RGC loss in the
   mouse retina post-I/R injury (Fig. [126]3I, J). To explore the
   mechanism by which the Reln signaling pathway influences I/R injury, we
   examined the effect of Reln knockdown on apoptosis using TUNEL staining
   (Fig. [127]6A). Consistent with prior findings, TUNEL-positive cells
   were negligible in the uninjured group but markedly increased 24 h
   post-I/R injury, with an exacerbated increase in the context of Reln
   gene knockdown (Fig. [128]6B). These observations indicate that Reln
   knockdown exacerbates apoptosis following I/R injury.

Fig. 6. Local knockdown of Reln promotes apoptosis after I/R through the
Dab1-PI3K-Akt pathway.

   [129]Fig. 6
   [130]Open in a new tab

   A Representative TUNEL staining of frozen retinal sections of
   AAV-shReln mice 24 h after I/R injury. Scale bar: 40 μm. B
   Quantification of TUNEL-positive cells in the GCL (left) and inner
   retinal layer (right), (n = 6). C, D Representative blot images of the
   effects of Reln local knockdown on retinal Dab1, p-Dab1, PI3K, Akt,
   p-Akt, GSK3b, p-GSK3b, Bcl-2, Bax, cle-caspase3 and Itgb1 after I/R
   injury. E Quantitative analysis of the levels of the aforementioned
   proteins (n = 3). Data are represented as the means ± SD. *P < 0.05,
   **P < 0.01, ***P < 0.001.

   Further mechanistic investigation via WB analysis disclosed no
   significant changes in Dab1, Akt and GSK3b expression levels in the
   AAV-shReln-I/R group. However, expression of PI3K and phosphorylated
   proteins p-Dab1, p-Akt and p-GSK3b(Ser9) was significantly reduced
   compared to the AAV-EGFP-I/R group. Additionally, the expression of
   apoptotic markers, Caspase-3 and the Bax/Bcl-2 ratio, was significantly
   elevated in the AAV-shReln-I/R group. Notably, Itgb1 protein expression
   was lower in the AAV-shReln-I/R group than in the AAV-EGFP-I/R group
   (Fig. [131]6C–E). In summary, Reln gene knockdown results in reduced
   Dab1 phosphorylation and impedes the activation of the PI3K-Akt-GSK3b
   pathway following retinal I/R injury, thereby promoting retinal cell
   apoptosis post-injury.

Discussion

   In this study, we demonstrate for the first time the neuroprotective
   role of Reelin protein in promoting the survival of retinal neurons, a
   significant advancement in addressing retinal ischemia. Retinal
   ischemia-reperfusion, implicated in the pathogenesis of glaucoma, is
   characterized by increased IOP and RGC death, leading to irreversible
   vision impairment globally [[132]25]. Current treatments, including eye
   drops, intraocular injections (e.g., anti-VEGF), and surgery, aim to
   halt disease progression but have inherent limitations, prompting the
   search for safer and more effective alternatives. We found basal Reelin
   expression in normal mouse retinas, with a decrease during the acute
   phase of I/R injury. Prophylactic administration of exogenous
   recombinant Reelin protein was observed to significantly reduce RGC
   loss, retinal thinning, and partially ameliorate visual function
   impairments. Conversely, Reln gene knockdown exacerbated RGC death and
   visual dysfunction post-I/R injury. Mechanistic insights reveal that
   Reelin’s neuroprotection is mediated through the activation of the
   Dab1-PI3K-Akt pathway, underscoring its potential as a therapeutic
   target for retinal ischemic diseases.

   Previous research indicates that Reelin expression peaks during retinal
   development, essential for the migration, positioning, and
   differentiation of retinal cells—a physiological upregulation
   [[133]26]. In adulthood, Reelin expression subsides to baseline, only
   to re-emerge in response to injury or degeneration [[134]27]. To
   monitor cells expressing Reln and their progeny in the retina, we
   developed the Reln-CreERT2 mTmG transgenic mouse model. Following
   tamoxifen induction, the tomato sequence is excised by Cre recombinase,
   and a membrane-tagged green fluorescent protein is expressed, as
   previously described [[135]28]. Our study tracked Reln expression
   post-retinal I/R injury, noting an initial decline in both Reln gene
   expression and protein levels, followed by an increase that still fell
   short of baseline levels. The early decrease in Reelin post-I/R may
   reflect a reduction in Reelin-expressing RGCs or CBCs due to injury.
   Conversely, the late-stage increase may result from the activation of
   Reelin expression in VECs, CBC-SC2, and blood or stromal cells,
   indicative of an injury-repair response aimed at maintaining retinal
   integrity.

   Existing literature on Reelin and I/R injury has predominantly focused
   on the central nervous system. In vitro studies have reported that
   recombinant Reelin protein confers significant neuroprotection against
   oxidative stress in neurons. Corresponding in vivo experiments have
   illustrated that pre-administration of exogenous Reelin protein into
   the ventricles prior to middle cerebral artery occlusion reduces
   cerebral infarct volume and promotes neuronal survival [[136]13,
   [137]14]. Our study pioneers the investigation of the neuroprotective
   effects of exogenous Reelin protein on the retina. Our study selected a
   well-established model of in vivo induced retinal I/R injury in which
   the entire retina is induced to ischemia, resulting in damage to
   multiple retinal cell types, primarily RGCs [[138]11]. The retina
   demonstrates distinct sensitivity to ischemia in different regions,
   with the outer layer exhibiting less sensitivity than the inner layer
   [[139]29]. Therefore, our research mainly focuses on the inner layer of
   the retina. Our findings underscore the role of the Reln gene and
   Reelin protein in preserving the inner retinal layer’s morphology and
   the retina’s overall structural integrity during I/R. The principal
   cellular damage post-retinal I/R injury is characterized by a decline
   in RGC count, which is crucial for visual information transmission to
   the brain. Notably, Reelin protein provides substantial protection for
   RGC survival.

   Additionally, we utilized FERG to evaluate visual function, revealing a
   significant positive correlation between the reduction in FERG
   amplitude and the deterioration of visual function post-I/R injury.
   This correlation is indicative of compromised electrophysiological
   function among retinal neurons. FERG parameters, notably the Max wave,
   offer an overview of general retinal functionality, while the a and b
   waves specifically reflect the status of the outer and inner retina,
   respectively. Our study demonstrated that Reelin treatment partially
   restored the amplitudes of both a-wave and b-wave in I/R-injured mice,
   signifying an amelioration of retinal dysfunction. Consequently,
   exogenous Reelin protein not only fosters neuronal survival and
   preserves retinal morphology but also significantly enhances retinal
   function in mice following I/R injury.

   To further validate the pivotal role of the Reln signaling pathway in
   retinal I/R injury, we examined the consequences of I/R-induced retinal
   injury in AAV-shReln mice. Given the survival challenges and
   neurological abnormalities associated with Reln deficiency in Reeler
   mice [[140]30], we opted for a localized retinal Reln knockdown model
   using AAV, which offers high target specificity alongside low
   carcinogenicity and immunological risk [[141]31]. We found that Reln
   knockdown resulted in a more pronounced thinning of both the inner and
   overall retinal layers following injury, suggesting a more severe
   alteration in retinal tissue morphology in retinal tissue. Strikingly,
   the inner retinal layer in uninjured Reln knockdown mice already
   displayed significant thinning, whereas the overall retinal thinning,
   while present, was not statistically significant, underscoring the Reln
   gene’s pronounced influence on the inner retinal layer in the uninjured
   state. Furthermore, a more pronounced reduction in RGC count and
   increased visual dysfunction were observed in Reln knockdown mice
   following I/R injury, with even uninjured knockdown mice demonstrating
   significant visual impairment. These outcomes not only emphasize the
   regulatory role of the Reln signaling pathway in maintaining normal
   retinal homeostasis but also further validate its neuroprotective role
   in promoting injury repair.

   Previous studies have identified apoptosis as a prominent mechanism of
   injury in I/R-associated retinopathy [[142]32]. We aimed to further
   investigate the regulatory role of the Reln gene and the Reelin protein
   in RGC apoptosis through TUNEL staining. Additionally, we quantified
   key proteins involved in the mitochondrial apoptotic pathway via WB
   analysis to elucidate the potential mechanisms underlying Reelin’s
   contribution to I/R injury. The Bcl-2 protein family, which represents
   the first documented group of apoptosis-related proteins, is known to
   play a key role in mediating mitochondrial apoptosis [[143]33].
   Numerous studies have demonstrated that Bcl-2 and Bax are major
   regulators in the Caspase-3-dependent apoptotic pathway that determines
   neuronal fate. In the context of retinal I/R injury, Caspase-3 is
   regarded as the primary executor of apoptosis in the inner layer of the
   rodent retina [[144]34]. Our findings indicate an increase in Caspase-3
   protein expression and the Bax/Bcl-2 ratio following I/R injury.
   However, pretreatment with Reelin protein led to a decrease in both
   metrics, which in turn increased following Reln knockdown. These
   results corroborate the neuroprotective mechanism of Reelin in
   attenuating RGC apoptosis.

   Classical Reelin pathway triggers include the VLDLR, ApoER2 and Itgb1
   [[145]17, [146]18]. In our study, single-cell transcriptomic analysis
   revealed that Reln-Itgα3β1 ligand-receptor pair is a signaling driver
   of the Reln signaling pathway. Integrin, a heterodimeric cell surface
   receptor for extracellular matrix proteins, comprises α and β subunits,
   with Itgb1 being the largest integrin subfamily involved in various
   cellular processes, including receptor-mediated activity, cell
   signaling, and defense mechanisms [[147]35, [148]36]. Previous study
   has found that Itgb1 can be activated by Reelin and induce the
   recruitment and activation of SFK [[149]37]. Our findings indicate that
   exogenous Reelin protein significantly upregulates Itgb1 expression in
   retinal tissues post-I/R injury, and its expression is suppressed
   following Reln knockdown, suggesting that Itgb1 is an important
   signaling molecule in the Reln signaling pathway in the retina.
   Additionally, previous studies have demonstrated that SFK, particularly
   Fyn and Src, function as the primary physiological kinases for Dab1 and
   are activated by receptor clustering during Reelin signaling. Their
   activity is subsequently upregulated in a Dab1-dependent manner through
   a positive feedback mechanism [[150]21, [151]38, [152]39]. Our findings
   reveal that Reelin protein robustly enhances the phosphorylation of Fyn
   and Src, thereby indicating the activation of the SFK signaling
   pathway. In vitro studies demonstrate that Reelin activates the
   PI3K-Akt pathway in a manner that is dependent on SFK activity and Dab1
   phosphorylation. Furthermore, the activated Akt phosphorylates Ser9 of
   GSK3b, leading to the inhibition of GSK3b activity and the promotion of
   cell survival [[153]40, [154]41]. In this study, we found that the
   Reelin protein significantly enhances Dab1 phosphorylation after
   retinal I/R injury, suggesting the activation of Reln signaling
   pathway. Phosphorylated Dab1 further activates the PI3K-Akt pathway,
   Akt phosphorylation at serine 473 is recognized for its neuroprotective
   effects, with p-Akt typically maintained at high levels in stable cells
   [[155]42]. However, Akt phosphorylation is significantly reduced
   post-I/R injury. Our study demonstrates that Reelin protein promotes
   Akt phosphorylation, a critical process for sustaining cell survival
   after ischemic retinal injury. In addition, we found that PI3K
   expression increases in Reelin protein-pretreated mice following
   retinal I/R. The activation of the PI3K-Akt pathway induces
   phosphorylation of GSK3b at the Ser9 residue, thereby inhibiting its
   activity and subsequently reducing neuronal apoptosis. Conversely, the
   PI3K inhibitor Alpelisib effectively reverses the anti-apoptotic
   effects of Reelin mediated through the PI3K-Akt pathway. In addition to
   the aforementioned findings, our results also revealed that although
   I/R injury led to an increase in the P25/P35 ratio, the administration
   of exogenous Reelin protein did not reverse this elevation. Both Cdk5
   and the Reelin pathway target tau protein as downstream substrate. The
   overphosphorylation of tau protein is a critical hallmark of various
   neurodegenerative diseases [[156]43]. We speculate that this outcome
   may stem from Reelin and Cdk5 functioning in parallel rather than
   through a simple linear interaction [[157]44]. In contrast, Reln
   knockdown mice exhibit contrasting trends. Reln knockdown resulted in
   marked attenuation of Dab1 phosphorylation and a marked inhibition of
   the PI3K-Akt pathway in the mouse retina, thereby activating GSK3b and
   inducing cell apoptosis. These findings substantiate that Reelin
   signaling activates the intracellular junction protein Dab1 via
   membrane receptors Itgb1, and subsequent activation of the PI3K-Akt
   signaling pathway. Activated p-Akt inhibits GSK3b activity and
   modulates the expression of Bcl-2 and Bax, as well as caspase-3,
   thereby inhibiting apoptosis and offering neuroprotection.

   Our findings reveal that the Reln-Dab1 signaling pathway is a critical
   regulatory and therapeutic target in the context of retinal I/R injury,
   exerting its effects through the modulation of neuronal apoptosis. With
   the novel introduction of this pathway into the study of retinal I/R
   injury, we enhance the understanding of disease mechanisms.
   Furthermore, our findings contribute to the development of targeted
   therapies with potential applications in glaucoma and other I/R-related
   retinal conditions.

Supplementary information

   [158]Supplementary Materials^ (65.9KB, doc)
   [159]Figure S1^ (68.4MB, tif)
   [160]figureS2^ (187.7MB, tif)
   [161]figureS3^ (11.1MB, tif)
   [162]figureS4^ (68.1MB, tif)
   [163]Figure S5^ (31.6MB, tif)
   [164]Full and uncropped western blots^ (2.1MB, pdf)
   [165]41419_2025_7742_MOESM8_ESM.xlsx^ (908.2KB, xlsx)

   Table of DEGs between Blank and IR groups in single-cell sequencing

Acknowledgements