LAG3 Gene Therapy Attenuates Autoimmune Uveitis by Regulating the PI3K/Akt Signaling Pathway

Abstract

Purpose

Lymphocyte-activation gene 3 (LAG3) is a critical immune checkpoint molecule, yet its role and mechanisms in autoimmune uveitis remain inadequately understood. This study sought to elucidate the expression dynamics, therapeutic potential, and molecular mechanisms of LAG3 in the context of experimental autoimmune uveitis (EAU).

Methods

The EAU model was established in C57BL/6J mice through subcutaneous administration of interphotoreceptor retinoid binding protein (IRBP) peptide emulsified in complete Freund's adjuvant and intraperitoneal injection of pertussis toxin. Alterations in LAG3 expression throughout disease progression were monitored. Adeno-associated virus serotype 2 (AAV2)–LAG3 was administered via intravitreal injection. Disease severity and blood–retinal barrier (BRB) integrity were evaluated using clinical scoring, histopathological analysis, Evans blue leakage assay, and transmission electron microscopy. Quantitative reverse-transcription PCR, western blot analysis, and RNA sequencing (RNA-seq) were utilized to examine associated molecular expression and signaling pathways. Functional rescue experiments employing a phosphoinositide 3-kinase (PI3K) activator were conducted to further explore the underlying mechanism.

Results

The expression of retinal LAG3 was downregulated during the peak inflammatory phase of EAU. Administration of AAV2–LAG3 significantly mitigated both clinical and pathological manifestations of EAU, re-established the Th17/Treg cell balance, and preserved the integrity of the BRB. Analyses via RNA-seq and western blotting suggested that these protective effects were associated with modulation of PI3K/Akt signaling pathway activation. Notably, the protective effects conferred by LAG3 were largely attenuated by application of the PI3K activator 740Y-P.

Conclusions

This study elucidated the downregulation of LAG3 expression in EAU and suggests that AAV2–LAG3 attenuates disease, likely by regulating the PI3K/Akt pathway and restoring Th17/Treg cell balance and BRB integrity.

Autoimmune uveitis represents a significant cause of blindness worldwide, primarily impacting young and middle-aged adults.¹ The recurrent and chronic characteristics of this condition not only result in severe visual impairment but also place a substantial burden on patients’ families and socioeconomic systems.^2,3 The current first-line therapeutic interventions, which include corticosteroids, immunosuppressants, and biologics, are frequently linked to considerable adverse effects with prolonged use. These adverse effects encompass elevated intraocular pressure, cataract formation, increased susceptibility to infections, and potential hepatic or renal dysfunction. Moreover, a subset of patients exhibits inadequate response or poor tolerance to these treatment modalities.⁴

Experimental autoimmune uveitis (EAU) is widely recognized as a prototypical model for investigating human uveitis. Current research suggests that CD4⁺ T-cell subsets, particularly the disequilibrium between T helper 17 (Th17) cells and regulatory T (Treg) cells, are pivotal in the pathogenesis of EAU.⁵ Th17 cells compromise the integrity of the blood–retinal barrier (BRB) and promote inflammation through the secretion of pro-inflammatory cytokines such as IL-17, whereas Treg cells uphold immune tolerance by releasing inhibitory cytokines such as IL-10 and TGF-β.^6–9 Consequently, reestablishing the Th17/Treg cell immune equilibrium is considered a promising strategy for mitigating the development and progression of EAU.

Lymphocyte-activation gene 3 (LAG3), also referred to as CD223, has garnered significant attention in recent years as a member of the inhibitory receptor family. It is widely expressed across various immune cell types, including CD4⁺ T cells, CD8⁺ T cells, regulatory T cells, B cells, and natural killer cells.^10–12 As a pivotal inhibitory receptor, LAG3 plays a crucial role in maintaining T-cell homeostasis and preventing the onset of autoimmunity.^12,13 Within the domain of cancer immunology, LAG3 has emerged as a promising therapeutic target.^11,14–16 In the context of autoimmunity, research indicates that LAG3 deficiency exacerbates conditions such as experimental autoimmune encephalomyelitis and mercury-induced autoimmunity,^17,18 whereas the protective function of induced regulatory T cells is contingent upon LAG3 expression.¹⁹ These findings underscore the significance of LAG3 in the regulation of autoimmune processes. Nevertheless, the specific role, expression patterns, and molecular mechanisms of LAG3 in autoimmune uveitis remain unexplored.

This study revealed a downregulation of LAG3 expression in EAU and suggested that restoring LAG3 expression effectively mitigates inflammation associated with EAU while preserving the integrity of the BRB. Additionally, LAG3 restoration was found to suppress the expression of the pathogenic Th17 cell transcription factor retinoic acid–related orphan receptor gamma t (RORγt) and enhance the expression of the Treg transcription factor forkhead box P3 (FOXP3). Mechanistic investigations using in vivo models suggested that LAG3 may exert its protective effects through the modulation of the phosphoinositide 3-kinase (PI3K)/Akt signaling pathway. In conclusion, this research highlights the protective role of LAG3 in the progression of EAU and provides empirical evidence that suggests its potential as a novel therapeutic target.

Materials and Methods

Animals

Wild-type C57BL/6J mice, 6 to 8 weeks old, were obtained from Jiangsu Huachuang Xinnuo Pharmaceutical Technology Co., Ltd. (Jiangsu, China). The mice were maintained in a specific pathogen-free environment, adhering to a 12-hour light/dark cycle with ambient temperature controlled at 23°C ± 2°C and relative humidity maintained at 55% ± 10%. The experimental protocols involving animals received approval from the Animal Ethics Committee of Guilin Medical University (approval no. GLMU-IACUC-202510153). All surgical procedures were conducted under anesthesia, with rigorous efforts undertaken to minimize animal distress. The procedures conformed to the guidelines set forth in the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research.

Induction of the EAU Mouse Model

The synthetic interphotoreceptor retinoid binding protein (IRBP) peptide 651-670 (sourced from NovoPro Bioscience, Shanghai, China) was emulsified with complete Freund's adjuvant (CFA; Chondrex, Woodinville, WA, USA) in a 1:1 volume ratio. Each C57BL/6J mouse was administered a subcutaneous injection of 200 µL of this emulsion, containing 500 µg of the IRBP peptide, at the base of the tail, back, and both thighs. Simultaneously, an intraperitoneal injection of 1 µg pertussis toxin (Sigma-Aldrich, St. Louis, MO, USA) was given to potentiate the immune response.^20,21

Viral Vector Construction and Preparation

Adeno-associated virus serotype 2 (AAV2) vectors, encoding either the LAG3 gene or the humanized Renilla reniformis green fluorescent protein (hrGFP) reporter gene, were constructed and packaged by Hunan Fenghui Biotechnology Co., Ltd. (Hunan, China). These vectors were driven by the cytomegalovirus (CMV) promoter, with a titer of 1 × 10¹² vector genomes/mL. The viral preparations were stored at −80°C and diluted to the appropriate working concentrations with sterile phosphate-buffered saline (PBS) prior to use.

Experimental Groups

Mice were randomly assigned to six independent studies as follows: (1) In a study on LAG3 expression dynamics, the mice were allocated into a normal control group and EAU groups, with evaluations conducted on days 7, 14, and 21 following immunization. Each group had eight mice. (2) In a validation study of AAV2–LAG3 overexpression, the mice were allocated into three distinct groups: normal control group, AAV2–hrGFP group, and AAV2–LAG3 group. Each group had eight mice. (3) For the efficacy study of AAV2–LAG3, the mice were allocated into four groups: normal control group, EAU group, AAV2–hrGFP+EAU group, and AAV2–LAG3+EAU group. The viral vectors were administered 3 weeks before the induction of EAU. All groups, each consisting of 18 mice, were assessed on the 14th day following immunization. (4) In the in vivo dose-finding study of 740Y-P, the mice were allocated into four distinct groups: normal control group, PBS group, 100-µM 740Y-P group, and 1-mM 740Y-P group. Intravitreal injections were administered at a volume of 1 µL per eye. Each group had three mice (Shanghai Yuanye Bio-Technology Co., Ltd., Shanghai, China). (5) For the toxicity study of 740Y-P, the mice were allocated into two groups: a normal control group and a group receiving an intravitreal injection of 1-mM 740Y-P (1 µL per eye). Each group had five mice. (6) For the in vivo rescue experiment, all mice were administered an AAV2–LAG3 injection, and EAU was induced 3 weeks subsequently. Seven days following the induction of EAU, the mice were subjected to an intravitreal injection of either PBS or 1-mM 740Y-P. Each group had 11 mice.

Intravitreal Injection

The AAV2 viral vector was diluted to a concentration of 5 × 10⁹ viral genomes/mL. Following the induction of anesthesia in the mice, pupil dilation was achieved using compound tropicamide eye drops (Shenyang Xingqi, Shenyang, China). Utilizing a dissecting microscope, an intravitreal injection of 2 µL was administered with a 34-gauge needle, approximately 1 mm posterior to the limbus. To prevent infection, topical levofloxacin hydrochloride eye drops (Jiangsu Giblee Pharmaceutical Co., Ltd., Jiangsu, China) were applied daily following the injection. Mice that developed severe corneal edema, cataract, or intraocular hemorrhage were excluded from the study.

EAU Clinical and Histological Scoring

Starting on the seventh day following immunization, ocular inflammation was assessed utilizing a slit-lamp microscope and a fundus camera. Clinical scoring was independently conducted by two observers who were blinded to the group assignments, following established criteria.^5,22 For histological evaluation, mice were euthanized at predetermined time points, and their eyeballs were enucleated. The samples were then fixed, embedded in paraffin, and sectioned into 3-µm-thick slices along the optic nerve axis, followed by staining with hematoxylin and eosin (H&E). Histopathological scoring of retinal inflammation was performed in accordance with established standards.⁵

Quantitative Real-Time PCR

Total RNA was isolated from retinal tissues utilizing the TRIzol reagent (Invitrogen, Carlsbad, CA, USA). Complementary DNA (cDNA) synthesis was conducted using the All-in-One First-Strand Synthesis MasterMix kit (YuGong Biotech, Jiangsu, China). Quantitative real-time PCR (qRT-PCR) assays were executed on an ABI Prism 7500 system (Applied Biosystems, Waltham, MA, USA), employing SYBR Green Premix (YuGong Biotech) as the fluorescent dye. The thermal cycling conditions were as follows: initial denaturation at 95°C for 30 seconds, followed by 40 amplification cycles of 95°C for 10 seconds and 60°C for 30 seconds. The relative expression levels of target mRNAs were determined using the 2^−ΔΔCt method, with β-actin serving as the internal control. Primer sequences are provided in the Table.

Table.

Primer Sequence Table

Primer	Sequence (5′→3′)
M-LAG3-F	AGTGACTCCCAAATCCTTCG
M-LAG3-R	CCTCCTGAATCTCCAGCACA
M-IL-10-F	TTCAAACAAAGGACCAGC
M-IL-10-R	GGATCATTTCCGATAAGG
M-IL-17-F	ACCGCAATGAAGACCCTGAT
M-IL-17-R	TCCCTCCGCATTGACACA
M-TNFα-F	AGGCGCCACATCTCCCTCCA
M-TNFα-R	CGGTGTGGGTGAGGAGCACG
M-β-actin-F	ATCATTGCTCCTCCTGAGCG
M-β-actin-R	CAGCTCAGTAACAGTCCGCC

Viral Vector Transduction Efficiency and Tissue Localization

To assess AAV2 vector transduction and expression within the retina, both retinal flatmounts and frozen sections were subjected to analysis. For the retinal flatmounts, the eyeballs were enucleated 3 weeks after intravitreal injection of either AAV2–hrGFP or AAV2–LAG3. The intact retinas were meticulously dissected under a dissecting microscope, radially incised with the optic disc serving as the center, mounted flat on a slide with PBS, and directly examined under a fluorescence microscope to observe macroscopic hrGFP expression. For the frozen sections, the eyeballs were fixed in an eyeball fixation solution for 24 hours, subsequently dehydrated overnight in 30% sucrose, embedded in optimal cutting temperature compound, and sectioned at a thickness of 10 µm. These sections were mounted with 4′,6-diamidino-2-phenylindole (DAPI) and examined using a confocal microscope to ascertain the cell-specific localization of the viral vector across different retinal layers.

Immunofluorescence Staining

The mouse eyeballs were fixed using an oculus fixation solution (Servicebio Technology, Wuhan, China) and subsequently embedded in paraffin. Retinal sections, approximately 3 µm in thickness, were prepared along the optic nerve axis. These sections underwent deparaffinization in xylene, followed by rehydration through a graded ethanol series, and they were then subjected to heat-induced antigen retrieval using sodium citrate buffer. The sections were then blocked with 5% bovine serum albumin (BSA) at room temperature for 30 minutes. Following this, they were incubated with primary antibodies targeting CD31 (1:100; Servicebio Technology), glutamine synthetase (1:500; HUABIO, Woburn, MA, USA), and phosphorylated Akt (p-Akt; 1:100; Affinity Biosciences, Cincinnati, OH, USA) within a humidified chamber at 4°C overnight. After extensive washing with PBS, the sections were incubated with Cy3-conjugated goat anti-rabbit IgG (1:200; Beijing Zhongshan Golden Bridge Biotechnology Co. Ltd., Beijing, China) and Cy5-conjugated goat anti-mouse IgG secondary antibodies for 1 hour at room temperature in the dark. Finally, the nuclei were counterstained with DAPI, and the sections were mounted using an antifade mounting medium. Images were acquired using a confocal microscope.

Immunohistochemistry

The mouse eyes were fixed and embedded in paraffin, and retinal sections were prepared along the optic nerve axis with a thickness of approximately 3 µm. These sections underwent deparaffinization in xylene, followed by rehydration through a graded ethanol series, and were then subjected to heat-induced antigen retrieval using a sodium citrate buffer. Subsequently, the sections were blocked with 5% BSA at room temperature for 30 minutes and then incubated with an anti-LAG3 primary antibody (1:100; HUABIO) at 4°C overnight. After they were washed with PBS, the sections were incubated with a horseradish peroxidase (HRP)-conjugated secondary antibody at room temperature for 1 hour. Following additional PBS washes, immunoreactivity was visualized using a diaminobenzidine (DAB) substrate solution, with the reaction being monitored under a microscope. The sections were subsequently counterstained with hematoxylin, dehydrated through a graded ethanol series, cleared in xylene, and mounted with neutral balsam.

Evans Blue Leakage Assay

At the height of the inflammatory phase of EAU, occurring on day 14 post-immunization, a 2% Evans Blue solution (100 µL) was administered intravenously via the tail vein. After 2 hours, the mice were euthanized, and their eyeballs were enucleated and subsequently fixed in 4% paraformaldehyde for 2 hours. The retinas were then carefully isolated, flatmounted on slides using glycerol, and examined under a confocal microscope to capture images of Evans Blue leakage from the retinal vessels.

Transmission Electron Microscopy Sample Preparation and Observation

To assess the ultrastructure of the BRB, mouse eyeballs were swiftly enucleated on the 14th day following EAU induction. Full-thickness tissue blocks from the posterior pole were promptly placed in a pre-cooled fixative specific for transmission electron microscopy (TEM), consisting of 2.5% glutaraldehyde and 2% paraformaldehyde in phosphate buffer, and were immersed in fresh fixative for overnight fixation at 4°C. Following phosphate buffer washes, the tissue blocks underwent post-fixation in 1% osmium tetroxide at 4°C for 2 hours. Subsequently, the fixed samples were dehydrated through a graded series of ethanol and acetone, infiltrated with a mixture of epoxy resin and acetone, transitioned to pure resin, embedded, and polymerized. Ultrathin sections, measuring 80 nm, were prepared using an ultramicrotome, double-stained with uranyl acetate and lead citrate, and examined via TEM. At a magnification of 15,000×, the structural integrity and continuity of tight junctions between retinal capillary endothelial cells and retinal pigment epithelial cells were evaluated to determine BRB integrity.

Western Blot Analysis

Total protein was extracted from retinal tissue lysates, and the protein concentration was quantified utilizing a bicinchoninic acid (BCA) assay kit (Yeasen Biotechnology, Shanghai, China). Subsequently, equal amounts of protein (30 µg) were resolved via 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto nitrocellulose membranes (MilliporeSigma, Burlington, MA, USA). Following a blocking step with 5% non-fat milk at room temperature for 1 hour, the membranes were incubated with diluted primary antibodies at 4°C overnight. The primary antibodies employed included LAG3, FOXP3, phosphorylated PI3K (p-PI3K), zonula occludens-1 (ZO-1), occludin, and claudin-5 (Shandong Hua'an Biotechnology Co., Ltd., Shandong, China); PI3K, Akt, and RORγt (Servicebio Technology); p-Akt (Affinity Biosciences); and β-actin (Beijing Zhongshan Golden Bridge Biotechnology Co. Ltd.). On the subsequent day, after washing with Tris-buffered saline with Tween 20 (TBST), the membranes were incubated with the appropriate HRP-conjugated secondary antibodies at room temperature for 1 hour. Signal detection was performed using an enhanced chemiluminescence (ECL) kit (Yeasen Biotechnology). The grayscale values of the target bands were analyzed using ImageJ (National Institutes of Health, Bethesda, MD, USA) and normalized to β-actin. The original full-length blots for all key antibodies are provided as Supplementary File (western blot.pdf).

RNA Sequencing and Bioinformatics Analysis

To elucidate the mechanism of action of AAV2–LAG3, we conducted RNA sequencing (RNA-seq) analysis on retinal samples from four distinct groups (n = 3 per group). Following the extraction of total RNA and ensuring quality control with a RNA integrity number (RIN) greater than 7.0, library construction and sequencing were executed by Shanghai Majorbio Bio-Pharm Technology Co., Ltd. (Shanghai, China). The data analysis process encompassed sequence alignment using hierarchical indexing for spliced alignment of transcripts 2 (HISAT2), quantification of fragments per kilobase of transcript per million mapped reads, identification of differentially expressed genes (defined by |log₂ fold change| > 1.5 and an adjusted P < 0.05), and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis. The enrichment results for key pathways, specifically PI3K/Akt, were subsequently validated at the protein level through western blot analysis.

Statistical Analysis

All experiments were replicated at least three times for reliability and reproducibility. Data are expressed as mean ± standard deviation and were analyzed utilizing Prism 8.0 (GraphPad Software, Boston, MA, USA). For comparisons between two groups, the unpaired Student's t-test was employed. Analyses involving multiple groups were conducted using one-way analysis of variance (ANOVA). The Mann–Whitney U test was applied to clinical and pathological scores. Statistical significance was set at P < 0.05.

Results

LAG3 Expression Was Downregulated in the EAU Model

To elucidate the natural progression of EAU, we established the model and conducted a comprehensive assessment of its clinical and histopathological alterations. Clinical scoring based on anterior segment and fundus photography indicated that ocular inflammation reached its zenith on day 14 post-immunization, as indicated by pronounced anterior segment changes, such as ciliary congestion and aqueous flare, alongside severe fundus pathologies, including optic disc edema, vascular tortuosity and dilatation, and exudates (Figs. 1A–D). Histological analysis corroborated these findings, revealing the most intense intraocular inflammation on day 14 post-immunization, characterized by extensive inflammatory cell infiltration and retinal folds (Figs. 1E, 1F). In alignment with the clinical and histopathological scoring trends, inflammation-related factors within retinal tissue exhibited synchronous dysregulation. The qRT-PCR results showed a significant upregulation of pro-inflammatory cytokines IL-17 and TNF-α in retinal tissue during disease progression, concomitant with a reduction in the expression of the anti-inflammatory cytokine IL-10 (Fig. 1G). Further analyses through qRT-PCR, western blot, and immunohistochemistry consistently revealed a downregulation of LAG3 mRNA and protein expression levels.

Figure 1.

LAG3 expression is downregulated in EAU. (A–D) Clinical scores of mice at 0, 7, 14, and 21 days post-immunization (mean ± SD; n = 5/group). Red arrow indicates ciliary congestion; blue arrow, optic disc edema; black arrow, exudate; green arrow, vascular tortuosity. **P < 0.01 (Mann–Whitney U test). (E, F) Histopathological scores of the same groups as in A to D (mean ± SD; n = 5/group). Red arrow indicates inflammatory cells; black arrow, retinal folds. **P < 0.01 (Mann–Whitney U test). (G, H) Relative mRNA expression levels of IL-10, IL-17, TNF-α, and LAG3 in the same groups mentioned above (mean ± SD; n = 3/group). *P < 0.05, **P < 0.01, ***P < 0.001 (one-way ANOVA). (I) LAG3 protein levels and quantification in retinal tissue (mean ± SD; n = 5/group). *P < 0.05, **P < 0.01, ***P < 0.001 (one-way ANOVA). (J) Retinal immunohistochemical staining (mean ± SD; n = 4/group). Brown indicates LAG3. **P < 0.01, ***P < 0.001 (one-way ANOVA).

AAV2–LAG3 Alleviated Clinical and Pathological Damage in EAU

To explore the therapeutic potential of LAG3 in the context of autoimmune uveitis, we engineered AAV2 vectors to carry either the LAG3 gene or the reporter gene hrGFP (Fig. 2A). The intravitreal administration of AAV2–hrGFP demonstrated effective transduction of retinal tissue, with expression primarily localized to the ganglion cell layer (GCL) and the INL (Fig. 2B, 2D). Autofluorescence was detected in the photoreceptor layer, inner plexiform layer (IPL), and outer plexiform layer (OPL) (Fig. 2C). Similarly, 3 weeks after intravitreal administration of AAV2–LAG3, there was marked upregulation of LAG3 expression at both the mRNA and protein levels within the retina (Figs. 2E, 2F).

Figure 2.

AAV2–LAG3 gene therapy alleviates EAU injury and ameliorates inflammation. (A) Schematic diagram of AAV2–LAG3 vector construction. (B) Retinal whole-mount after intravitreal injection of AAV2–hrGFP. Green represents hrGFP; blue, DAPI. (C) Frozen section of a normal eye. Green represents tissue autofluorescence; blue, DAPI. (D) Frozen section after intravitreal injection of AAV2–hrGFP. Green represents hrGFP; blue, DAPI. PRL, photoreceptor layer. (E, F) Retinal LAG3 mRNA and protein levels and quantification 3 weeks after AAV2–LAG3 injection (mean ± SD; n = 3–5/group). ***P < 0.001 (one-way ANOVA). (G–J) Clinical scores of mice in the normal (N), EAU, AAV2–hrGFP+EAU, and AAV2–LAG3+EAU groups (mean ± SD; n = 5/group). Red arrows indicate ciliary hyperemia; blue arrows, optic disc edema; black arrows, exudation; green arrows, vascular tortuosity. **P < 0.01 (Kruskal-Wallis test). (K, L) Histopathological scores of the same groups (mean ± SD; n = 5/group). Red arrows indicate inflammatory cells; black arrows, retinal folds. **P < 0.01 (Kruskal-Wallis test). (M) Relative mRNA expression levels of IL10, IL17, and TNF-α in the same groups (mean ± SD; n = 3/group). *P < 0.05, **P < 0.01, ***P < 0.001 (one-way ANOVA).

The therapeutic efficacy of AAV2–LAG3 was assessed using the EAU model. On day 14 post-injection, both the EAU group and the AAV2–hrGFP+EAU group demonstrated severe ocular inflammation, characterized by ciliary congestion, optic disc edema, vascular tortuosity, retinal exudates, and partial retinal detachment, with no significant differences observed between these two groups (Figs. 2G, 2I). Conversely, the AAV2–LAG3+EAU treatment group exhibited a marked reduction in ocular clinical symptoms (Figs. 2G, 2I). Clinical scoring analysis further corroborated that the scores in the AAV2–LAG3+EAU treatment group were significantly lower than those in both the AAV2–hrGFP+EAU and untreated EAU groups (Figs. 2H, 2J).

The histopathological evaluation corroborated the clinical findings. H&E staining (Fig. 2K) indicated extensive infiltration of inflammatory cells, pronounced retinal folds, and detachment in both the EAU and AAV2–hrGFP+EAU groups. In contrast, the AAV2–LAG3+EAU group exhibited only minimal scattered inflammatory cells and mild structural changes with focal detachment (Fig. 2K). Quantitative histological scoring, conducted in accordance with established criteria, revealed that the damage score in the AAV2–LAG3+EAU treatment group was significantly lower compared to the AAV2–hrGFP+EAU and EAU groups (Fig. 2L). Thus, intravitreal administration of the AAV2 vector successfully facilitated targeted LAG3 expression in the retina, resulting in a significant reduction in both the clinical severity and pathological damage associated with EAU.

The study investigated the molecular mechanisms underlying AAV2–LAG3 treatment in EAU. Retinal tissues were collected on day 14 post-immunization for analysis. The qRT-PCR results demonstrated that the AAV2–LAG3+EAU treatment group exhibited significantly lower mRNA expression levels of the pro-inflammatory cytokines IL-17 and TNF-α in comparison to the AAV2–hrGFP+EAU and EAU groups. Conversely, there was a significant upregulation of the anti-inflammatory cytokine IL-10 (Fig. 2M). Additionally, western blot analysis revealed that AAV2–LAG3+EAU treatment significantly decreased the protein expression of RORγt, a critical transcription factor for Th17 cells, while increasing the protein expression of FOXP3, a key transcription factor for Treg cells (Fig. 3A). These findings suggest that intravitreal administration of AAV2–LAG3 not only mitigates the clinical and histopathological manifestations of EAU but also exerts its therapeutic effects by potentially inhibiting pro-inflammatory cytokine expression (IL-17, TNF-α), enhancing anti-inflammatory cytokine expression (IL-10) and restoring the Th17/Treg cell balance.

Figure 3.

AAV2–LAG3 regulates the Th17/Treg cell balance and protects the BRB. (A) Protein levels and quantification of RORγt and FOXP3 in the normal, EAU, AAV2–hrGFP+EAU, and AAV2–LAG3+EAU groups (mean ± SD; n = 3/group). *P < 0.05, **P < 0.01 (one-way ANOVA). (B, D) Evans Blue extravasation assay and quantification in the previous four groups (mean ± SD; n = 5/group). White arrows indicate leakage. ***P < 0.01 (one-way ANOVA). (C) TEM images of retinal capillary tight junction ultrastructure in the same four groups. Black arrows indicate tight junctions. (E) Protein levels and quantification of ZO-1, occludin, and claudin-5 in the same four groups (mean ± SD; n = 3/group). *P < 0.05, **P < 0.01 (one-way ANOVA).

AAV2–LAG3 Protected the Integrity of the BRB in EAU

The BRB serves as an essential protective mechanism for preserving the homeostasis of the retinal microenvironment. This barrier is comprised of tight junction proteins that function as a selective physical barrier, preventing the infiltration of harmful substances and immune cells from the bloodstream while simultaneously enabling the transport of nutrients and the removal of waste products through active transport mechanisms.^23–26 In pathological conditions such as uveitis, the integrity of the BRB is compromised, resulting in increased permeability. This disruption facilitates the infiltration of inflammatory cells and consequently leads to retinal damage.^27–30

To assess the protective effects of AAV2–LAG3 on the BRB, we conducted an Evans Blue leakage assay. Retinal vascular leakage was significantly increased in the EAU and AAV2–hrGFP+EAU groups compared to the normal control group, with no significant difference observed between these two groups (Figs. 3B, 3D). In contrast, treatment with AAV2–LAG3+EAU effectively reduced this leakage (Figs. 3B, 3D). TEM analysis further corroborated at the ultrastructural level that the tight junction structure of the BRB was compromised in the EAU and AAV2–hrGFP+EAU groups, whereas AAV2–LAG3 treatment enhanced the integrity of this structure (Fig. 3C). At the molecular level, western blot analysis revealed a significant reduction in the expression of the key tight junction proteins ZO-1, occludin, and claudin-5 in the EAU and AAV2–hrGFP+EAU groups, with no significant difference between them. Conversely, AAV2–LAG3+EAU treatment increased the expression levels of these barrier-associated proteins (Fig. 3E).

AAV2–LAG3 Attenuated EAU by Inhibiting the PI3K/Akt Signaling Pathway

To further elucidate the molecular mechanisms underlying the therapeutic effects of AAV2–LAG3 on EAU, we conducted RNA-seq analysis on retinal tissues from four groups: normal control, EAU, AAV2–hrGFP+EAU, and AAV2–LAG3+EAU. Through the analysis of differentially expressed genes (DEGs) between the normal control and EAU groups, we preliminarily identified key alterations in signaling pathways associated with the disease state (Figs. 3E, 3G). Additionally, by examining DEGs between the AAV2–hrGFP+EAU and AAV2–LAG3+EAU groups, we identified pathways potentially linked to the therapeutic effects of LAG3 (Figs. 3F, 3H). In both comparative analyses, the PI3K/Akt signaling pathway was notably enriched. In the comparison between the EAU group and the normal control group, this enrichment was primarily characterized by the downregulation of X-inactive specific transcript (Xist) mRNA and the upregulation of fibroblast growth factor 2 (FGF2) and placental growth factor (PGF) mRNA (Figs. 4A, 4C). Conversely, in the AAV2–LAG3+EAU group compared to the AAV2–hrGFP+EAU group, the enrichment was indicated by the upregulation of Xist mRNA and the downregulation of angiopoietin-1 (ANGPT1), G protein subunit gamma transducin 2 (GNGT2), GNGT1, serum- and glucocorticoid-regulated kinase 3 (SGK3), and integrin subunit beta 8 (ITGB8) mRNA, collectively contributing to the enrichment of the PI3K/Akt signaling pathway (Figs. 4B, 4D). These findings suggest that the PI3K/Akt signaling pathway not only plays a significant role in the pathogenesis of EAU but may also serve as a crucial target for mediating the therapeutic effects of AAV2–LAG3.

Figure 4.

Differential enrichment of the PI3K/Akt signaling pathway in EAU pathogenesis and AAV2–LAG3 therapy. (A, C) KEGG pathway enrichment analysis of DEGs between the normal and EAU groups. (B, D) KEGG pathway enrichment analysis of DEGs between the AAV2–hrGFP+EAU and AAV2–LAG3+EAU groups. (E) Protein levels and quantification of p-PI3K, PI3K, p-Akt, and Akt in the same four groups (mean ± SD; n = 3/group). *P < 0.05, **P < 0.01, ***P < 0.001 (one-way ANOVA).

To substantiate the bioinformatics findings, we investigated the activation status of the PI3K/Akt signaling pathway at the protein level in vivo. Western blot analyses revealed that, in comparison to the control group, the levels of p-PI3K and p-Akt were markedly elevated in the retinas of the EAU and AAV2–hrGFP+EAU groups, suggesting aberrant activation of this pathway in the pathological state. Subsequent to AAV2–LAG3 treatment, there was a significant reduction in the protein expression levels of both p-PI3K and p-Akt (Fig. 4E). These findings collectively suggest that the therapeutic efficacy of AAV2–LAG3 is closely linked to the suppression of the hyperactivated PI3K/Akt signaling pathway in the retina.

To elucidate the subcellular distribution of the activated form of p-Akt within the retinal architecture, we conducted a co-localization analysis employing the endothelial cell marker CD31 and the Müller cell marker glutamine synthetase (GS). The findings revealed that p-Akt was predominantly localized within the IPL, INL, and OPL. Notably, in comparison to the normal control group, the expression of p-Akt in the IPL was significantly elevated in both the EAU group and the adeno-associated virus serotype 2 expressing humanized green fluorescent protein (AAV2–hrGFP) plus EAU group. Conversely, treatment with AAV2–LAG3 substantially attenuated p-Akt levels within this layer. In contrast, no significant alterations in p-Akt expression were observed in the OPL. Additionally, the absence of discernible co-localization signals between p-Akt and either CD31 or GS indicates that p-Akt is not predominantly expressed in endothelial cells or Müller cells (Supplementary Figs. S1, S2).

PI3K Activator 740Y-P Reversed the Therapeutic Effect of AAV2–LAG3 on EAU

Building upon the previously discussed findings, we hypothesized that the therapeutic efficacy of AAV2–LAG3 is associated with its inhibition of the PI3K/Akt signaling pathway. To empirically evaluate this relationship, we performed a rescue experiment by co-administering the PI3K activator 740Y-P via intravitreal injection in conjunction with AAV2–LAG3 treatment.

On the seventh day following immunization, an intravitreal injection of 740Y-P was administered, with subsequent evaluations conducted on the 14th day post-immunization. Dose optimization studies, involving injection concentrations of 100 µM and 1 mM (both at a volume of 1 µL/eye), indicated that an intravitreal injection of 1-mM 740Y-P (1 µL/eye) elicited the most pronounced activation of the targeted pathway (Fig. 5A). At this concentration, ocular toxicity was evaluated 7 days post-injection. Compared to the control group, the administration of 1-mM 740Y-P (1 µL/eye) resulted in mild increases in IL-17 and TNF-α levels, whereas IL-10 levels remained unchanged. Furthermore, no significant differences were observed in anterior segment scores, fundus scores, or pathological scores (Supplementary Fig. S3). Consequently, 1-mM 740Y-P (1 µL/eye) was selected as the optimal concentration for intravitreal injection in subsequent experiments. An in vivo functional rescue study demonstrated that intravitreal injection of 1 µL/eye of 1-mM 740Y-P significantly attenuated the therapeutic effects of AAV2–LAG3.

Figure 5.

The PI3K agonist 740Y-P reverses the protective effects of AAV2–LAG3 on inflammatory infiltration and tissue structure. (A) Protein levels and quantification of p-PI3K, PI3K, p-Akt, and Akt in the normal, PBS, 100-µm, and 1-mM groups (mean ± SD; n = 3/group). *P < 0.05 (one-way ANOVA). (B–E) Clinical scores for the AAV2–LAG3+EAU, AAV2–LAG3+PBS+EAU, and AAV2–LAG3+740Y-P+EAU groups (mean ± SD; n = 5/group). Red arrows indicate ciliary hyperemia; black arrows, exudation; green arrows, vascular sheathing. **P < 0.01 (Kruskal–Wallis test). (F, G) Histopathological scores of the same groups (mean ± SD; n = 5/group). Red arrows indicate inflammatory cells; black arrows, retinal folds. *P < 0.05 (Kruskal-Wallis test). (H, I) Relative mRNA expression levels of IL10, IL17, and TNF-α in the same groups (mean ± SD; n = 3/group). *P < 0.05, **P < 0.01, ***P < 0.001 (one-way ANOVA). (J, K) Protein levels and quantification of RORγt and FOXP3 in the same groups (mean ± SD; n = 3/group). *P < 0.05, **P < 0.01 (one-way ANOVA).

Clinical scoring outcomes indicate that, in comparison to the AAV2–LAG3+PBS+EAU group, mice in the AAV2–LAG3+740Y-P+EAU group demonstrated exacerbated anterior segment inflammation, characterized by ciliary congestion (Figs. 5B, 5C), as well as more severe fundus lesions, including vascular inflammation and exudates (Figs. 5D, 5E). Histopathological scoring results revealed that H&E-stained retinal sections from the AAV2–LAG3+PBS+EAU group maintained a relatively intact retinal structure with minimal inflammatory cell infiltration. Conversely, the AAV2–LAG3+740Y-P+EAU group exhibited hallmark EAU pathological alterations, such as inflammatory cell infiltration, disrupted retinal architecture, pronounced retinal folds, and focal detachment (Figs. 5F, 5G).

The qRT-PCR analysis demonstrated that, in comparison to the AAV2–LAG3+PBS+EAU group, the AAV2–LAG3+740Y-P+EAU group exhibited a marked increase in the mRNA expression levels of the pro-inflammatory cytokines IL-17 and TNF-α within the retina (Fig. 5H). Conversely, there was a significant downregulation in the expression of the anti-inflammatory cytokine IL-10 (Fig. 5I). Concurrently, western blot analysis revealed a substantial elevation in the protein level of the Th17-associated transcription factor RORγt, but the protein level of the Treg cell–associated transcription factor FOXP3 was notably reduced (Figs. 5J, 5K). These findings suggest that activation of the PI3K/Akt pathway attenuates the protective effects conferred by AAV2–LAG3 against inflammatory infiltration and the preservation of tissue structure.

The Evans Blue leakage assay (Figs. 6A, 6D) and TEM observations (Figs. 6B, 6C) provided further evidence that 740Y-P treatment exacerbated retinal vascular leakage and disrupted the tight junction architecture of the BRB. Western blot analysis revealed a significant reduction in the expression of key BRB proteins, including ZO-1, occludin, and claudin-5, in the AAV2–LAG3+740Y-P+EAU group (Fig. 6E). These findings indicate that activation of the PI3K/Akt pathway attenuated the barrier-protective effects conferred by AAV2–LAG3.

Figure 6.

The PI3K agonist 740Y-P reverses the protective effects of AAV2–LAG3 on the BRB. (A, D) Evans Blue extravasation assay and quantification in the AAV2–LAG3+EAU, AAV2–LAG3+PBS+EAU, and AAV2–LAG3+740Y-P+EAU groups (mean ± SD; n = 5/group). White arrows indicate leakage. ***P < 0.01 (one-way ANOVA). (B, C) TEM images of retinal capillary and RPE tight junction ultrastructure in the same three groups. Black arrows indicate tight junctions. (E) Protein levels and quantification of ZO-1, occludin, and claudin-5 in the same three groups (mean ± SD; n = 3/group). *P < 0.05, **P < 0.01 (one-way ANOVA).

Discussion

Uveitis is a prevalent ocular condition that can lead to blindness, and its management continues to encounter substantial challenges. Presently, clinical practice lacks universally effective and specific therapeutic options, largely due to the intricate nature of its etiology and pathogenesis.¹ Although conventional treatments, such as corticosteroids and various immunosuppressants, can partially alleviate symptoms, they are frequently accompanied by significant adverse effects.^2,3 Consequently, there is an imperative need to investigate novel therapeutic agents that can prevent recurrence, minimize side effects, and sustain efficacy.

LAG3 plays a pivotal role in maintaining immune homeostasis. It is essential for Treg cells to achieve optimal inhibitory function,^31,32 and the absence or inhibition of LAG3 can disrupt self-tolerance. For example, in the non-obese diabetic (NOD) mouse model, the genetic knockout of LAG3 exacerbates type 1 diabetes progression and accelerates the infiltration of autoreactive T cells into the pancreatic islets.^33,34 Furthermore, LAG3 frequently operates synergistically with other inhibitory receptors, such as programmed cell death protein 1 (PD-1). The concurrent deficiency of LAG3 and PD-1 in mice leads to fatal autoimmune myocarditis,³⁵ underscoring the critical role of LAG3 in preventing autoimmune diseases. This evidence suggests that LAG3 dysfunction may be a significant factor in the pathogenesis of autoimmune diseases. Our study systematically examined the dynamic expression changes, therapeutic potential, and molecular mechanisms of LAG3 in EAU and found that LAG3 may represent a therapeutic target for autoimmune uveitis.

Our study showed that day 14 post-immunization marked the peak inflammatory phase in C57BL/6J mice. Initially, we examined the alterations in LAG3 expression within the retinal tissue of mice throughout the disease progression. The findings revealed that, during the peak inflammatory phase of EAU, both mRNA and protein expression levels of LAG3 in the retina were markedly downregulated. Furthermore, the extent of this downregulation was inversely correlated with the clinical and histopathological severity of the disease. This observation aligns with findings from other autoimmune models, such as multiple sclerosis, rheumatoid arthritis, and autoimmune diabetes,^17,18,33,34 suggesting that dysregulated LAG3 expression may be a common characteristic in the pathogenesis of various autoimmune diseases.

AAV vectors are extensively utilized in ocular gene therapy due to their minimal cytotoxicity and pathogenicity.³⁶ In order to assess the therapeutic efficacy of the LAG3 gene, we utilized an AAV2 serotype vector administered via intravitreal injection, which facilitated targeted overexpression within the local retina, specifically in the GCL and the INL. The findings indicate that intravitreal administration of AAV2–LAG3 successfully elevated LAG3 expression levels in retinal tissue and significantly mitigated both clinical and histological damage associated with EAU, without inducing discernible toxicity. These results support the feasibility and safety of this targeted therapeutic approach.

At the mechanistic level, the overexpression of LAG3 markedly inhibited the production of the pivotal pro-inflammatory cytokines IL-17 and TNF-α, while concurrently augmenting the expression of the anti-inflammatory cytokine IL-10. At the level of transcription factors, LAG3 downregulated RORγt, the principal regulator of Th17 cells, and upregulated FOXP3, the central regulator of Treg cells. These findings suggest that LAG3 has the capacity to effectively restore equilibrium to the ocular Th17/Treg cell axis, thereby alleviating autoimmune uveitis.

This study also revealed that LAG3 plays a protective role in maintaining the integrity of the BRB. The BRB is a critical structure, and its disruption is a key factor contributing to vision loss in uveitis.^24–27 The inner BRB is comprised of tight junctions among retinal capillary endothelial cells, whereas the outer BRB is composed of tight junctions between retinal pigment epithelium (RPE) cells^29,30 Our findings indicate that overexpression of LAG3 results in reduced vascular leakage, enhanced ultrastructural integrity of tight junctions, and increased expression of essential junctional proteins such as ZO-1, occludin, and claudin-5. These results suggest that LAG3 may exert an indirect effect on retinal endothelial cells and/or RPE cells to stabilize the barrier architecture. This indirect effect is likely mediated by paracrine factors secreted from LAG3-overexpressing retinal cells (such as retinal ganglion cells and Müller cells). Recent studies offer compelling mechanistic insights, demonstrating that the interaction between LAG3 and its ligand, major histocompatibility complex (MHC) class II molecules, can establish an exceptionally tight intercellular interface between immune cells and target cells, such as antigen-presenting cells. The “tight interface” exhibits a molecular exclusion effect, effectively preventing key costimulatory molecules, such as CD4, and other receptor–ligand complexes from entering the core interaction zone, thereby functioning as a “microscopic barrier.”³⁷ We hypothesize that AAV2–LAG3 treatment may partially restore or mimic this effect, synergistically enhancing the BRB at the cellular level. Nonetheless, this hypothesis requires further validation in subsequent studies.

To systematically elucidate the downstream signaling pathways, we conducted transcriptome sequencing and subsequent bioinformatics analysis. The KEGG pathway enrichment analysis identified significant enrichment of the PI3K/Akt signaling pathway in both the EAU pathogenesis and AAV2–LAG3 treatment comparison groups, albeit exhibiting divergent regulatory patterns. In the EAU group compared to the normal control group, pathway enrichment was characterized by the downregulation of Xist and the upregulation of FGF2 and PGF, suggesting activation of the PI3K/Akt signaling pathway during EAU pathogenesis, potentially contributing to the promotion of inflammatory responses. Conversely, in the AAV2–LAG3+EAU group compared to the AAV2–hrGFP+EAU group, pathway enrichment was associated with the upregulation of Xist and the downregulation of ANGPT1, GNGT2, GNGT1, SGK3, and ITGB8 (Figs. 4B, 4D). It is noteworthy that Xist is recognized as a negative regulator of the PI3K/Akt pathway,³⁸ and the observed downregulation of SGK3, a direct downstream target of this pathway, further supports the possibility of pathway activity inhibition.^39,40 The collective findings suggest that the PI3K/Akt signaling pathway may serve as a pivotal driver in the pathogenesis of EAU and may represent a significant therapeutic target modulated by AAV2–LAG3. AAV2–LAG3 potentially mitigates the progression of EAU through the upregulation of Xist and the inhibition of PI3K/Akt pathway activity. The PI3K pathway is integral to cellular processes such as growth, differentiation, and the cell cycle, with Akt serving as the direct downstream effector within the PI3K signaling cascade.^41,42 Additionally, the PI3K/Akt pathway has been associated with various autoimmune diseases, including experimental autoimmune prostatitis and autoimmune uveitis.^21,43,44 The findings of the western blot analyses were consistent with activation of the PI3K/Akt pathway in EAU, whereas LAG3 overexpression was found to effectively inhibit its phosphorylation and activation. Immunofluorescence assays further revealed that p-Akt is predominantly localized within the IPL, inner nuclear layer (INL), and OPL. The IPL has been recognized as a critical region where LAG3 modulates the activity of the PI3K/Akt signaling pathway. Furthermore, there was no significant co-localization of p-Akt with CD31, an endothelial cell marker, or with GS, a marker for Müller cells. Given the distribution pattern observed in the IPL, it is hypothesized that p-Akt–positive signals are predominantly localized to amacrine cells or the dendrites of ganglion cells. These findings suggest that LAG3 may exert its therapeutic effects on EAU by modulating the PI3K/Akt pathway within these specific cell types.

Importantly, the administration of the PI3K activator 740Y-P in vivo attenuated the protective effects imparted by LAG3, which included diminished inflammation, restored BRB integrity, and modulation of the Th17/Treg cell balance. This finding provides additional support that the inhibition of the PI3K/Akt signaling pathway constitutes the key mechanism by which LAG3 mediates its therapeutic effects.

This study is subject to certain limitations. Our research was confined to the PI3K/Akt signaling pathway, leaving open the question of whether LAG3 cross-regulates other associated pathways, such as signal transducer and activator of transcription 3 (STAT3), nuclear factor kappa B (NF-κB), AMP-activated protein kinase (AMPK), and hypoxia-inducible factor 1-alpha (HIF-1α), which merits further investigation. Additionally, the specific molecular mechanism connecting the cell surface LAG3 receptor to intracellular PI3K/Akt regulation remains to be elucidated. Furthermore, it should be acknowledged that AAV2–LAG3 predominantly transduces retinal neurons rather than directly targeting infiltrating CD4⁺ T cells; therefore, its protective effects are likely mediated indirectly through modification of the retinal microenvironment. LAG3 may act in a paracrine manner on adjacent endothelial cells to enhance BRB integrity, or it may regulate PI3K/Akt signaling in neurons to modulate the release of inflammatory mediators. In addition, in vitro supernatant transfer experiments using conditioned media from LAG3-overexpressing retinal cells co-cultured with endothelial cells or naïve T cells would be valuable to definitively test whether secreted factors mediate the observed protective effects. Such experiments are warranted in future investigations to further validate this hypothesis. Moreover, it should be noted that our RNA-seq analysis was performed on whole retinal tissue, which inherently contains a mixture of resident retinal cells and infiltrating immune cells. Consequently, we cannot definitively identify the specific cell types responsible for the observed changes in PI3K/Akt signaling and related molecular alterations. Future studies employing single-cell RNA sequencing or cell-sorting experiments are warranted to validate the exact cellular dynamics underlying these molecular alterations.

Conclusions

Our findings indicate that LAG3 expression is downregulated in EAU and that administration of AAV2–LAG3 can mitigate disease progression, likely by inhibiting the PI3K/Akt signaling pathway. This inhibition subsequently rebalances the Th17/Treg cell axis and preserves the integrity of the BRB. These results lay the preclinical groundwork for development of therapeutic strategies targeting LAG3.