Exendin-4 averts all-trans-retinal-driven damage to photoreceptors and the retina via the GLP-1R/PKA/CREB1 signaling axis

Abstract

Background

Atrophic macular degeneration comprises dry age-related macular degeneration (AMD) and autosomal recessive Stargardt disease (STGD1). These disorders lead to irreversible blindness and still lack effective therapies. The rise of all-trans-retinal (atRAL) brought on by visual cycle disruption closely links to retinal atrophy in both conditions, yet the key downstream targets remain obscure. Exendin-4 (EX-4) is a natural glucagon-like peptide-1 receptor (GLP-1R) agonist. Recent clinical retrospective studies indicate that GLP-1R agonists such as exenatide (synthetic EX-4) can markedly lower the 5-year risk of developing dry AMD. Here, we sought to clarify the protective effect of natural EX-4 against retinal degeneration in atrophic macular degeneration linked to impaired clearance of atRAL.

Methods

Cell and animal paradigms of STGD1 and dry AMD were generated by atRAL-loaded 661W cells and light-exposed Abca4^−/−Rdh8^−/− mice, respectively. RNA-sequencing, cell viability assays, morphometric analysis, annexin V/propidium-iodide staining using flow cytometry, quantitative polymerase chain reaction (qPCR), western blotting, immunofluorescence, electroretinography (ERG), fundus photography, hematoxylin and eosin (H&E) histology, and TUNEL staining were integrated to delineate the anti-apoptotic actions of EX-4 and to uncover its underlying protective mechanism.

Results

GLP-1R/cyclic adenosine monophosphate (cAMP)-dependent protein kinase (PKA)/cAMP response element–binding protein 1 (CREB1) signaling was markedly downregulated in atRAL-challenged 661W cells and in neural retina of light-exposed Abca4^−/−Rdh8^−/− mice. EX-4 reinstated this pathway, suppressed caspase-3 activation and DNA damage, and curtailed apoptosis in both cell and tissue contexts. Silencing of Glp1r or the PKA catalytic subunits by small interfering RNA (siRNA) abrogated EX-4-induced activation of the PKA/CREB1 axis in atRAL-loaded 661W cells. Pharmacologic blockade of CREB1 phosphorylation with the PKA inhibitor H-89 or the CREB1 inhibitor 666-15 largely nullified the DNA-protective and anti-apoptotic benefits conferred by EX-4 in 661W cells following atRAL exposure, suggesting that the GLP-1R/PKA/CREB1 signaling axis contributes to its cytoprotection action. More importantly, intraperitoneal injection of EX-4 significantly preserved retinal structure and function in Abca4^−/−Rdh8^−/− mice after exposure to light, and mitigated punctate lesions in the fundus.

Conclusions

EX-4 exerted anti-apoptotic and DNA-protective effects against atRAL-induced photoreceptor loss and retinal degeneration at least partially through activating the GLP-1R/PKA/CREB1 pathway. These findings suggest that GLP-1R agonists could serve as potential preventive therapeutics for atrophic macular degeneration associated with atRAL toxicity, including dry AMD and STGD1.

Supplementary Information

The online version contains supplementary material available at 10.1186/s11658-026-00952-6.

Introduction

Atrophic macular degeneration includes dry age-related macular degeneration (AMD) and autosomal recessive Stargardt disease (STGD1), which are closely associated with dysfunction of the visual cycle [1–3]. ATP-binding cassette transporter 4 (ABCA4) and retinol dehydrogenase 8 (RDH8) are key components of the visual cycle responsible for clearing all-trans-retinal (atRAL), and their mutations lead to disrupted clearance of atRAL in the retina and contribute to age-dependent retinal degeneration [1, 4]. There is already evidence that Abca4 is the pathogenic gene of STGD1, and Abca4 and Rdh8 are the risk genes for dry AMD [5–7]. As previously reported, mice lacking both Abca4 and Rdh8 (Abca4^−/−Rdh8^−/− mice) show a significant rise of atRAL in the retina, and recapitulate key features of STGD1 and dry AMD, particularly progressive photoreceptor/retinal pigment epithelium (RPE) atrophy [1, 8]. Light exposure speeds up the release of atRAL from rhodopsin, leading to a rapid surge of atRAL in the retina, which in turn accelerates the degeneration of photoreceptors and the RPE [2, 3, 8, 9]. These findings implicate visual cycle dysfunction-induced atRAL overload as a key driver of atrophic macular degeneration.

To date, neither STGD1 nor dry AMD has a curative or vision-restoring therapy. However, recent retrospective cohort studies have reported that treatment with glucagon-like peptide-1 receptor (GLP-1R) agonists, such as exenatide, liraglutide, and semaglutide, is linked to a substantially reduced 5-year risk of developing dry AMD [10, 11]. It should be mentioned here that exendin-4 (EX-4) is a natural GLP-1R agonist, and synthetic EX-4 is called exenatide. GLP-1R is a class B1 G protein-coupled receptor of the incretin family that transduces signals from endogenous glucagon-like peptide-1 (GLP-1) and its analogs [12]. GLP-1R transcripts and protein are widely expressed in the pancreas, central nervous system, and retinal neurons and vasculature [13–16]. Upon binding of GLP-1 or GLP-1R agonists such as EX-4, GLP-1R primarily couples to the stimulatory heterotrimeric G protein Gs (composed of Gα, Gβ, and Gγ), thereby activating adenylyl cyclases and increasing intracellular cyclic adenosine monophosphate (cAMP) levels [17]. The accumulated cAMP binds the regulatory subunits of cAMP-dependent protein kinase (PKA), inducing a conformational change and the release of catalytic subunits [17]. These subunits phosphorylate cAMP response element–binding protein 1 (CREB1) and other downstream effectors, launching transcriptional programs that promote cellular survival [17]. Previous studies in animal models of Alzheimer-like neurodegeneration and ischemic neuronal injury have shown that GLP-1R agonists, including EX-4 and liraglutide, exert neuroprotective effects and ameliorate cognitive or neuronal deficits, at least in part through activation of the cAMP/PKA/CREB signaling pathway [18–20]. Furthermore, there is already evidence that GLP-1R activation attenuates retinal function deterioration, preserves retinal thickness, and supports the survival of retinal neurons in models of diabetic retinopathy and ocular hypertension [15, 21, 22].

However, the therapeutic potential of GLP-1R activation for atrophic macular degeneration remains unexplored. In the present study, we investigated both the protective effects and mechanistic basis of the GLP-1R agonist EX-4 against retinal injury in atRAL-driven retinopathies, such as dry AMD and STGD1.

Methods and materials

Study design

Motivated by the clinical potential of GLP-1R agonists in dry AMD, we hypothesize that GLP-1R signaling attenuates atRAL-induced retinal degeneration. RNA sequencing (RNA-seq) analysis of atRAL-loaded 661W cells uncovered core signaling perturbations. Mechanistic experiments employed 661W cells with silencing of Glp1r or the PKA catalytic subunits by small interfering RNA (siRNA) or pretreated with EX-4 for 3 h before atRAL exposure, with or without the PKA inhibitor H89 and the CREB1 inhibitor 666-15, respectively. In vivo efficacy was assessed in light-exposed Abca4^−/−Rdh8^−/− mice. Sample sizes were at least six for primary endpoints and at least three for molecular assays based on power analysis. Electroretinography (ERG), histology, immunofluorescence, and other assays were performed with investigators blinded to group allocation throughout measurement and analysis.

Reagents

A list of reagents was provided in the Supplementary Materials Table S1.

Animals

Abca4^−/−Rdh8^−/− mice on a C57BL/6 J background were generated as previously described [2]. All procedures were approved by the Institutional Animal Care and Use Committee of Xiamen University School of Medicine. Beginning at 4 weeks of age, mice received daily intraperitoneal injections of EX-4 (20 µg/kg body weight). Days 1−5 were spent under a standard 12 h light/12 h dark cycle. After the fifth injection, animals were kept in complete darkness for 48 h. On day 7, mice were exposed to 10,000 lx white light-emitting diode (LED) light for 1 h, commencing 3 h after EX-4 administration. Following light exposure, animals were returned to darkness and continued receiving daily EX-4 treatment for an additional 5 days. Vehicle-injected littermates served as controls, and a separate cohort of light-free mice remained under the 12 h light/12 h dark cycle for 7 days followed by 5 days in darkness. Finally, the mice were subjected to scotopic full-field ERG and fundus photography, and their eyes were enucleated for subsequent analyses. It was worth noting that, to determine the optimal in vivo dosage of EX-4 (20 µg/kg body weight), a dose–response study was conducted to evaluate its protective effects against light-induced retinal damage in Abca4^−/−Rdh8^−/− mice. A pilot comparison of 10, 20, and 50 µg/kg body weight EX-4 using ERG and fundoscopy identified 20 µg/kg body weight as the optimal dose for subsequent ERG, fundoscopy, and histological evaluation (Supplementary Materials Fig. S1).

Cell culture and treatment

The 661W cell line is a cone photoreceptor cell line derived from a mouse retinal tumor, serving as an in vitro model for studying retinal degeneration [23]. The murine photoreceptor cell line 661W was purchased from Shanghai Zishi Biotechnology (catalog no. SNL-252) and has been previously characterized [24]. Cells were maintained in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin–streptomycin at 37 °C in a humidified 5% CO₂ incubator.

Cell viability and morphology

MTS-based cytotoxicity assessment was carried out as described [25]. First, 661W cells were seeded into 96-well plates at a density of 6.5 × 10³ cells per well and cultured overnight. Cells were pretreated with EX-4 for 3 h, then exposed to 5 µM atRAL or an equivalent volume of dimethyl sulfoxide (DMSO) vehicle for 6 h. Then, 20 µl of CellTiter 96^® AQueous One Solution Reagent (Promega, Madison, WI, USA) was added to each well and incubated for 1 h at 37 °C. Absorbance was read at 490 nm with a 1510 Multiskan GO microplate spectrophotometer (Thermo Fisher Scientific). Cell morphology was evaluated on a Leica DMi8 inverted microscope (Wetzlar, Germany).

RNA-seq analysis and visualization

RNA sequencing was performed on an Illumina NovaSeq 6000 platform by Shanghai OE Biotech Co., Ltd. (Shanghai, China). Gene Set Enrichment Analysis (GSEA) was conducted on OECLOUD using the Reactome.v2023.1. Heat maps displaying fragments per kilobase of transcript per million mapped reads (FPKM) values of differentially expressed genes were generated with GraphPad Prism 10.1.0.

RNA interference assay

661W cells were seeded into six-well plates at 1.5 × 10⁵ cells per well, cultured overnight, and then transfected with 20 μM siRNA (2 μl per well) using Lipofectamine RNAiMAX (9 μl per well) according to the manufacturer’s instructions. The siRNA sequences of Glp1r-1, Glp1r-2, Glp1r-3, Prkaca, Prkacb, and negative control (NC) are shown in Table 1.

Table 1.

siRNA sequences

Gene	Sense	Antisense
Glp1r-1	GAGACGGUGCAGAAAUGGATT	UCCAUUUCUGCACCGUCUCTT
Glp1r-2	GGUUCCUUCGUGAAUGUCATT	UGACAUUCACGAAGGAACCTT
Glp1r-3	GGAACUACAUCCACCUGAATT	UUCAGGUGGAUGUAGUUCCTT
Prkaca	CGAGUAACUUUGACGACUATT	UAGUCGUCAAAGUUACUCGTT
Prkacb	GCCCAGAUUGUGCUAACAUUU	AAAUGUUAGCACAAUCUGGGC
NC	UUCUCCGAACGUGUCACGUTT	ACGUGACACGUUCGGAGAATT

Quantitative polymerase chain reaction (qPCR)

The method of qPCR was executed as described previously [25]. Total RNA was extracted from cells using the total RNA extraction reagent and reverse-transcribed into cDNA using the qPCR RT Master Mix kit. qPCR was performed using the Taq Pro Universal SYBR qPCR Master Mix on a Roche Light Cycler 96 system. The primer sequences used for qPCR are provided in Table 2.

Table 2.

Primer sequences

Gene	Forward primer	Reverse primer
Glp1r	TCTGCAACCGGACCTTTGAT	TGTGCAGAACCGGTACACAT
Prkaca	GGAAAGTCCCGCTCAGAACA	CGAGTTTGACGAGGAACGGA
Prkacb	GGCAGCTTATAGAGAACCACCT	GAAGGCTGTCATTTTCCAGAGC
Actb	AGATCAAGATCATTGCTCCTCCT	GGACTGTTACTGAGCTGCGT

Annexin V/PI apoptosis detection

After rinsing with phosphate-buffered saline (PBS), cells were resuspended in binding buffer and stained with ABflo^® 488–Annexin V and propidium iodide for 15 min at 4 °C in the dark. Samples were immediately analyzed on a Beckman flow cytometer, and data were processed with FlowJo software (v10.8.1).

Immunofluorescence

Cells were fixed with 4% paraformaldehyde for 15 min at 4 °C, permeabilized with 0.1% Triton X-100 in PBS for 10 min at room temperature, and blocked with 5% bovine serum albumin (BSA). They were then incubated overnight at 4 °C with primary antibodies (1:100 dilution), followed by Alexa Fluor 594-conjugated secondary antibodies (1:200 dilution) for 1 h at room temperature. Alternatively, retinal sections from mouse eyeballs were immunostained using a Tyramide Signal Amplification (TSA) kit (AiFang Biological). Briefly, retinal sections were incubated at 4 °C overnight with primary antibodies (1:100 dilution) and then exposed for 1 h to an HRP-conjugated goat anti-rabbit (1:1000 dilution) secondary antibody at room temperature, followed by incubation with the TSA working solution for 10 min. Samples were mounted in Fluoroshield with DAPI and imaged on an Olympus FV4000 confocal microscope. For immunofluorescence analysis of retinal sections, two nonconsecutive sections containing the optic nerve head per eye were selected; in each slice, two images were acquired bilaterally within regions approximately 500–1250 μm from the optic nerve head. Fluorescence intensity was quantified using ImageJ software, and the mean value of images of each eye was used for statistical analysis.

Western blotting

Cells and tissues were lysed in RIPA buffer containing protease and phosphatase inhibitors (Thermo Fisher Scientific) and subjected to intermittent sonication on ice. Protein concentrations were determined with a BCA assay kit (Thermo Fisher Scientific). Equal amounts of protein were resolved on 8−12% SDS–PAGE gels and electro-transferred to PVDF membranes (Roche Applied Science). After blocking with 5% skim milk or 5% BSA in TBST for 1 h at room temperature, membranes were incubated overnight at 4 °C with primary antibodies, followed by 2 h incubation with HRP-conjugated secondary antibodies at room temperature. Immunoreactive bands were visualized with ECL reagents (NCM Biotech) and captured on a ChemiDoc XRS+ imaging system (Bio-Rad) or an eBlot Touch Imager. Band intensities were quantified using ImageJ.

Hematoxylin and eosin (H&E) staining

Mouse eyeballs were fixed, embedded in paraffin, and sectioned into 3-µm thick slices. Slices were routinely stained with H&E. Images of H&E staining were obtained from three nonconsecutive sections containing the optic nerve head from each eye using a Leica Aperio Versa 200 microscope, and the outer nuclear layer (ONL) thickness was measured at 250-µm intervals from the optic disc toward the nasal and temporal regions. The mean value from the three sections was used as one biological replicate for each eye.

ERG

Full-field ERG was used to evaluate retinal function. Mice were dark-adapted overnight and anesthetized with isoflurane under dim red light. Pupils were dilated with tropicamide, and responses were recorded with an ERG system (Diagnosys LLC, Lowell, MA, USA) across a series of flash intensities. Representative waveforms obtained at 1 cd s/m² were used to measure a- and b-wave amplitudes for statistical comparison.

Fundus photography

After intraperitoneal injection of 150 μl 1% pentobarbital sodium, mice were kept under stable anesthesia and their pupils were dilated with 1% tropicamide. To maintain optical clarity and prevent cataract formation, 0.2% carbomer gel was applied to the corneal surface. High-resolution fundus photographs were then acquired with a small-animal retinal imaging system (Optoprobe, OPIMG-L).

TUNEL staining assay

After deparaffinization and rehydration, sections were fixed in 4% paraformaldehyde (15 min, 4 °C). Apoptotic cells were detected with a TUNEL assay kit (Promega) according to the manufacturer’s instructions. Following counterstaining with DAPI, slides were imaged on an Olympus FV4000 confocal microscope.

Statistical analysis

All data are presented as mean ± standard deviation (S.D.) from at least three independent experiments. Group differences were assessed by Student’s t-test or one-/two-way ANOVA, followed by appropriate multiple-comparison correction (e.g., Tukey’s post hoc test) as specified in the figure legends. Exact p-values are shown to indicate statistical significance, except in Fig. 6C and F, in which significance is denoted by symbols. A p value less than 0.05 was considered statistically significant.

Fig. 6.

EX-4 ameliorates retinal degeneration and photoreceptor apoptosis in light-exposed Abca4^−/−Rdh8^−/− mice. A Schematic diagram of the experimental design, illustrating the timeline for EX-4 administration, light exposure, and subsequent evaluations. B and C Assessment of retinal function using full-field ERG under scotopic conditions. Representative ERG waveforms recorded at a light intensity of 1 cd s/m². Quantification of scotopic a- and b-wave amplitudes elicited at increasing stimulus intensities (0.01, 0.1, 1, and 10 cd s/m²; n = 6). D Typical fundus photographs. E and F Histological analysis of mouse retina. Retinal sections stained with H&E. Quantification of ONL thickness measured along the temporal–nasal meridian (n = 6). G TUNEL staining (green) of mouse retina. H Immunoblots of cleaved caspase-3, PARP, cleaved PARP, and γH2AX, together with quantification of cleaved caspase-3, cleaved PARP, and γH2AX in neural retina (n = 3). Two-way ANOVA followed by Tukey’s post hoc test in C, F, and H. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 versus DMSO-treated, light-exposed Abca4^−/−Rdh8^−/− mice; ^#p < 0.05, ^##p < 0.01, ^###p < 0.001, ^####p < 0.0001 versus DMSO-treated, light-free Abca4^−/−Rdh8^−/− mice. RPE, retinal pigment epithelium; OS, outer segment; IS, inner segment; ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer; GCL, ganglion cell layer

Results

GLP-1R/PKA/CREB1 signaling is impaired in atRAL-loaded 661W cells

To examine whether atRAL disrupts GLP-1R-coupled signaling in 661W cells, we performed GSEA on RNA-seq data. Gene sets linked to GLP-1R signaling were significantly negatively enriched, including “GLP-1 regulates insulin secretion” (NES, Normalized Enrichment Score =  −1.55, p = 0.031) and “PKA-mediated phosphorylation of CREB1” (NES =  −1.46, p = 0.05) (Fig. 1A). Correspondingly, key genes of the pathway were coordinately downregulated: heterotrimeric Gs subunits (Gnas, Gnb1, Gnb4, Gng2, and Gng7), adenylyl cyclases (Adcy6, Adcy7, and Adcy9), PKA catalytic subunits (Prkaca and Prkacb), and downstream effectors (Rapgef3, Calm1, Calm2, and Calm3) (Fig. 1B). qPCR analysis showed that Glp1r mRNA expression was downregulated in atRAL-challenged 661W cells (Fig. 1C). Consistently, western blotting detected two GLP-1R-immunoreactive bands (approximately 50 and 175 kDa), both of which were markedly reduced upon atRAL challenge (Fig. 1D and E). To confirm the identity of the two bands, we transfected 661W cells with three independent siRNAs against murine Glp1r. All three siRNAs abolished both the 175 and 50 kDa signals, verifying that both bands represent GLP-1R (Fig. 1F). Consistent with the transcriptomic data, PKA catalytic subunits (PRKACA and PRKACB, collectively referred to as PKA), CREB1 phosphorylated at Ser133 (p-CREB1), and the p-CREB1/CREB1 ratio were all decreased (Fig. 1D and E). These data suggest that atRAL disrupts the PKA/CREB1 signaling axis in 661W cells.

Fig. 1.

GLP-1R/PKA/CREB1 signaling is downregulated in atRAL-loaded 661W cells. A GSEA of two GLP-1R-related gene sets: “GLP-1 regulates insulin secretion” and “PKA-mediated phosphorylation of CREB1”. B Expression of related genes in A. 661W cells were treated with 5 μM atRAL or DMSO for 6 h. C Glp1r mRNA levels in 661W cells exposed to atRAL or DMSO alone. D Verification at the protein level for A and B. Immunoblots of GLP-1R (175 and 50 kDa), PKA, p-CREB1, and CREB1. E Quantification of protein levels in D (n ≥ 3). Levels of each protein were normalized to those of β-actin or GAPDH. F Validation of GLP-1R antibody specificity via siRNA knockdown. Immunoblots of GLP-1R (175 and 50 kDa) in 661W cells transfected with a negative control siRNA and three different siRNAs targeting mouse GLP-1R, respectively (n = 3). Student’s t-test in E and one-way ANOVA with Tukey’s post hoc test in F

EX-4 protects 661W cells from atRAL exposure by mitigating apoptosis and DNA damage

MTS assays revealed that EX-4 rescued the viability of atRAL-loaded 661W cells in a dose-dependent manner, with a modest improvement already evident at 1 nM and significant protection achieved at 10 and 100 nM (Fig. 2A). Treatment with 100 nM EX-4 not only restored cellular morphology but also reduced the fraction of late-apoptotic/necrotic cells from 34.12% (DMSO + atRAL) to 5.32% (EX-4 + atRAL) in 661W cells following atRAL exposure (Fig. 2B and C). Mechanistically, EX-4 markedly attenuated the accumulation of cleaved caspase-3, cleaved poly(ADP-ribose) polymerase (PARP), and γH2AX in atRAL-exposed 661W cells (Fig. 2D). Caspase-3 is the central executioner protease that integrates extrinsic and intrinsic apoptotic signals [26]. Once activated, it cleaves PARP to commit cells to death [27]. Moreover, γH2AX flags double-strand DNA breaks [28]. Our previous work has demonstrated that DNA damage and caspase-3-dependent apoptosis involve atRAL-induced photoreceptor injury [2]. These results imply that pharmacological activation of GLP-1R blunts atRAL-elicited photoreceptor cytotoxicity by alleviating apoptosis and DNA damage.

Fig. 2.

EX-4 increased the viability of atRAL-loaded 661W cells by mitigating apoptosis and DNA damage. A Cell viability was assessed by MTS assay. 661W cells were pretreated with 1, 10, and 100 nM EX-4 for 3 h and then exposed to 5 μM atRAL for 6 h. Control cells were incubated with DMSO alone (n = 6). B−D 661W cells were pretreated with 100 nM EX-4 for 3 h, followed by 6 h exposure to 5 μM atRAL. B Cellular morphology (n = 3). Scale bars, 100 μm. C Apoptosis was assessed by annexin V/PI staining and flow cytometry (n = 3). D Immunoblots of cleaved caspase-3, PARP, cleaved PARP, and γH2AX, as well as the quantification of cleaved caspase-3, cleaved PARP, and γH2AX (n = 3). Two-way ANOVA with Tukey’s post hoc test in A and D

EX-4 promotes GLP-1R/PKA/CREB1 signaling activation in 661W cells

GLP-1R knockdown impaired the viability of 661W cells (Fig. 3A and B), suggesting its pro-survival function. It also attenuated the PKA/p-CREB1 signaling axis (Fig. 3C and D). Notably, siGLP-1R treatment abolished the EX-4-mediated increases in the levels of GLP-1R, PKA, and p-CREB1 (Fig. 3C and D). In atRAL-loaded 661W cells, EX-4 enhanced the expression of both GLP-1R bands (50 and 175 kDa), as evidenced by immunoblotting (Fig. 3E and F). Concomitantly, EX-4 visibly restored PKA and p-CREB1 levels in 661W cells following exposure to atRAL (Fig. 3E and F). Immunofluorescence staining confirmed these findings, and revealed a pronounced nuclear accumulation of both PKA and p-CREB1 signals in atRAL-treated 661W cells (Fig. 3G–L). Together, these results suggest that EX-4 can restore the GLP-1R/PKA/CREB1 signaling in 661W cells following atRAL exposure.

Fig. 3.

EX-4 promotes GLP-1R/PKA/CREB1 signaling activation in atRAL-treated 661W cells. A and B 661W cells transfected with a negative control siRNA (siNC) and a siRNA targeting mouse GLP-1R (siGlp1r) for 24 h. A Glp1r mRNA levels (n = 3). B Cell viability was assessed by MTS assay (n = 4). C and D 661W cells were transfected with siNC and siGlp1r for 24 h, followed by incubation with 100 nM EX-4 or DMSO for 9 h. C Immunoblots of GLP-1R (175 and 50 kDa), PKA, p-CREB1, and CREB1. D Quantification of GLP-1R (175 and 50 kDa), PKA, p-CREB1, and the ratio of p-CREB1/CREB1 (n = 3). E−L 661W cells were pretreated with 100 nM EX-4 for 3 h and subsequently challenged with 5 μM atRAL for 6 h. E Immunoblots of GLP-1R (175 and 50 kDa), PKA, p-CREB1, and CREB1. F Quantification of GLP-1R (175 and 50 kDa), PKA, p-CREB1, and the ratio of p-CREB1/CREB1 (n = 5). G and H Immunofluorescence analysis and quantification of GLP-1R (n = 4). I−L Immunofluorescence analysis and quantification of PKA (I and J) and p-CREB1 (K and L) within cells and nuclei (n = 4). Student’s t-test in A and B, two-way ANOVA with Tukey’s post hoc test in D, F, H, J, and L

PKA/CREB1 signaling underpins the anti-apoptotic protection conferred by EX-4 in 661W cells

To investigate whether PKA participates in transmitting downstream p-CREB1 signaling following EX-4-induced GLP-1R activation, we silenced the catalytic PKA subunits PKACα and PKACβ, which are respectively encoded by Prkaca and Prkacb [29, 30]. Individual knockdown of Prkaca or Prkacb affected the mRNA expression of the reciprocal subunit, whereas co-silencing both genes achieved more efficient suppression (Fig. 4A). PKA knockdown impaired 661W cell viability, indicating that PKA contributes to cell survival (Fig. 4B), and was accompanied by a reduction in p-CREB1 signaling (Fig. 4C and D). Notably, the EX-4-induced increase in p-CREB1 levels was markedly attenuated by siPKA treatment, suggesting that PKA is involved in transmitting the downstream signal from EX-4 to p-CREB1 (Fig. 4C and D). Given the essential role of PKA in maintaining basal cell survival and the baseline level of PKA/p-CREB1 signaling under resting conditions, we next employed pharmacological inhibition approaches to better control this pathway and to determine whether the cytoprotective effect of EX-4 in atRAL-loaded 661W cells is mediated via the PKA/CREB1 axis [31].

Fig. 4.

Inhibition of PKA attenuates EX-4-induced CREB1 activation and anti-apoptotic protection in atRAL-exposed 661W cells. A and B 661W cells were transfected with siNC, siPrkaca, siPrkacb, or combined siPrkaca + siPrkacb for 24 h. A The mRNA validation of Prkaca and Prkacb knockdown efficiency (n = 4). B Cell viability of 661W cells transfected with siNC or combined siPrkaca + siPrkacb (n = 5). C and D 661W cells were transfected with siNC or combined siPrkaca + siPrkacb for 24 h, followed by incubation with 100 nM EX-4 or DMSO for 9 h. C Immunoblots of PKA, p-CREB1, and CREB1 in siNC or combined siPrkaca + siPrkacb transfected cells with or without EX-4 treatment. D Quantification of protein levels of PKA and p-CREB1, and the ratio of p-CREB1/CREB1 (n = 3). E The nontoxic concentration window of the PKA inhibitor H89 was delineated by cytotoxicity profiling: 661W cells were treated with H89 (0.5, 1, 2, or 4 µM) for 9 h (n = 3). F H89 dose-dependently attenuated EX-4-induced CREB1 phosphorylation within the safe concentration range. Cells were co-incubated with 100 nM EX-4 and escalating concentrations of H89 (0.5, 1, 2, or 4 µM) for 9 h (n = 3). G and H 661W cells were pretreated with 100 nM EX-4 together with 1 µM PKA inhibitor H89 for 3 h, followed by exposure to 5 µM atRAL for 6 h. G Immunoblots of p-CREB1, CREB1, cleaved caspase-3, PARP, cleaved PARP, and γH2AX. H Quantification of p-CREB1, the p-CREB1/CREB1 ratio, cleaved caspase-3, cleaved PARP, and γH2AX (n = 3). One-way ANOVA with Tukey’s post hoc test in A, E, F, and H, Student’s t-test in B, and two-way ANOVA with Tukey’s post hoc test in D

Concentrations of 0.5–2 µM H89, a competitive inhibitor of PKA that targets the ATP-binding site of the catalytic subunit, did not significantly compromise 661W cell viability (Fig. 4E). Within this window, H89 suppressed EX-4-induced CREB1 phosphorylation in a dose-dependent manner in 661W cells; 1 µM H89 virtually eliminated the EX-4-mediated increase in p-CREB1 levels (Fig. 4F). Cells were therefore pre-incubated with 100 nM EX-4 together with 1 µM H89 for 3 h and then challenged with 5 µM atRAL for 6 h. H89 co-treatment visibly prevented the EX-4-mediated suppression of cleaved caspase-3, cleaved PARP, and γH2AX in atRAL-loaded 661W cells, and drove p-CREB1 levels and the p-CREB1/CREB1 ratio back to the atRAL-alone level (Fig. 4G and H). To further establish the requirement for p-CREB1, we treated atRAL-exposed 661W cells with 666-15, a selective CREB1 phosphorylation inhibitor that is virtually nontoxic at 100 nM in retinal cells [32]. At this concentration, 666-15 virtually blocked EX-4-driven CREB1 phosphorylation and restored cleaved caspase-3, cleaved PARP, and γH2AX, as well as returned p-CREB1 levels and the p-CREB1/CREB1 ratio to baseline in atRAL-loaded 661W cells (Fig. 5). Collectively, these findings imply that EX-4 protects 661W cells against atRAL-evoked apoptosis and DNA damage at least partly through the PKA-mediated phosphorylation of CREB1.

Fig. 5.

CREB1 signaling contributes to the anti-apoptotic protection conferred by EX-4 in atRAL-loaded 661W cells. Immunoblot analysis of p-CREB1, CREB1, cleaved caspase-3, PARP, cleaved PARP, and γH2AX, as well as the quantification of p-CREB1, the ratio of p-CREB1/CREB1, cleaved caspase-3, cleaved PARP, and γH2AX (n = 3). Cells were pretreated with 100 nM EX-4 together with 100 nM CREB1 inhibitor 666–15 for 3 h, followed by exposure to 5 µM atRAL for 6 h. Two-way ANOVA with Tukey’s post hoc test

EX-4 ameliorates retinal degeneration and photoreceptor apoptosis in light-exposed Abca4^−/−Rdh8^−/−mice

As reported previously, 48 h dark-adapted Abca4^−/−Rdh8^−/− mice exposed to 10,000 lx white LED light for 1 h exhibit rapidly increased atRAL levels in the retina, and develop acute retinal degeneration [33]. To test the in vivo therapeutic potential of EX-4, an experimental protocol was designed (illustrated in Fig. 6A). EX-4 alone left ERG responses of light-free Abca4^−/−Rdh8^−/− mice unchanged, yet significantly preserved scotopic function in Abca4^−/−Rdh8^−/− mice after light insult, yielding visibly larger a- and b-wave amplitudes at every flash intensity (0.01, 0.1, 1, and 10 cd s/m²) (Fig. 6B and C). At a flash intensity of 1 cd s/m², EX-4-treated, light-exposed Abca4^−/−Rdh8^−/− mice exhibited a 28.6% increase in a-wave amplitudes and a 30.6% elevation in b-wave amplitudes relative to vehicle-treated, light-exposed Abca4^−/−Rdh8^−/− mice (Fig. 6B and C). Funduscopy showed that EX-4 markedly reduced retinal edema and RPE mottling (Fig. 6D) in light-exposed Abca4^−/−Rdh8^−/− mice, while H&E staining revealed significant preservation of ONL thickness (Fig. 6E and F). TUNEL staining confirmed a pronounced drop in apoptotic photoreceptors within the ONL of EX-4-treated, light-exposed Abca4^−/−Rdh8^−/− mice (Fig. 6G). Moreover, immunoblotting of neuroretinal lysates showed strong suppression of cleaved caspase-3, cleaved PARP, and γH2AX in EX-4-treated versus vehicle-treated light-challenged Abca4^−/−Rdh8^−/− mice (Fig. 6H). Together, these results demonstrate that intraperitoneal injection of EX-4 robustly sustains retinal function and architecture, and suppresses photoreceptor apoptosis in light-exposed Abca4^−/−Rdh8^−/− mice.

EX-4 upregulates GLP-1R/PKA/CREB1 signaling in the retina of light-challenged Abca4^−/−Rdh8^−/− mice

Light exposure reduced 175 and 50 kDa GLP-1R, PKA, CREB1, and p-CREB1 in neuroretinal lysates of Abca4^−/−Rdh8^−/− mice, whereas intraperitoneally injected EX-4 fully or partially restored the levels of each protein (Fig. 7A and B). EX-4 raised GLP-1R levels in both light-free and light-stressed neuroretinas of Abca4^−/−Rdh8^−/− mice, with the 175 kDa band showing the most pronounced gain and the 50 kDa band increasing modestly (Fig. 7A and B). In light-free, vehicle-treated Abca4^−/−Rdh8^−/− mice, GLP-1R immunoreactivity peaked in the ganglion cell layer, consistent with previously reported retinal distribution profiles of GLP-1R [14], whereas it was present at lower but readily detectable levels in the inner and outer nuclear layers (Fig. 7C). After illumination, the remarkable rise in the 175 kDa band in neural retina lysates from EX-4-treated, light-exposed Abca4^−/−Rdh8^−/− mice coincided with a striking amplification of GLP-1R fluorescence specifically in the ONL (Fig. 7C). In light-free Abca4^−/−Rdh8^−/− mice, EX-4 boosted PKA expression in photoreceptors, concentrating the signal within the inner and outer segments (Fig. 7D and E). Following light exposure, PKA immunostaining intensified and extended into the ONL of EX-4-treated Abca4^−/−Rdh8^−/− mice (Fig. 7D and E). CREB1, a nuclear protein, declined in the ONL of Abca4^−/−Rdh8^−/− mice after light exposure, but was preserved in the ONL of EX-4-treated, light-challenged Abca4^−/−Rdh8^−/− mice (Fig. 7F). The levels of p-CREB1 were elevated by EX-4 in the ONL of Abca4^−/−Rdh8^−/− mice without exposure to light (Fig. 7F and G). Light exposure also lowered protein levels of p-CREB1 in the ONL of Abca4^−/−Rdh8^−/− mice, yet EX-4 clearly reversed the decrease of p-CREB1 protein levels in the ONL of light-exposed Abca4^−/−Rdh8^−/− mice (Fig. 7F and G). Moreover, protein levels of both PKA and p-CREB1 were clearly elevated by EX-4 in the ONL of light-free Abca4^−/−Rdh8^−/− mice (Fig. 7D−G). The data show that intraperitoneal administration of EX-4 enhances the GLP-1R/PKA/CREB1 signaling axis in Abca4^−/−Rdh8^−/− mice following exposure to light.

Fig. 7.

EX-4 activates the GLP-1R/PKA/CREB1 signaling pathway in Abca4^−/−Rdh8^−/− mice retinas. A and B Western blot analysis of GLP-1R (175 and 50 kDa), PKA, p-CREB1, and CREB1, together with quantification of these proteins and the p-CREB1/CREB1 ratio in neural retina (n ≥ 6). C Immunofluorescence staining and quantification of GLP-1R (magenta) in the ONL (n = 4). D and E Immunofluorescence staining and quantification of PKA (red) in the ONL (n = 4). F and G Immunofluorescence co-staining of p-CREB1 (red) with CREB1 (green), and quantification of p-CREB1 in the ONL (n = 4). Two-way ANOVA followed by Tukey’s post hoc test in B, E, and G

Discussion

Recent retrospective cohort studies have shown that treatment with GLP-1R agonists, e.g., exenatide, liraglutide, and semaglutide, is associated with a markedly lower 5-year incidence of dry AMD [10, 11], suggesting GLP-1R agonists may confer benefit in atrophic macular degeneration. In this study, we demonstrated that the GLP-1R agonist EX-4 significantly blunted retinal degeneration via reducing apoptosis and DNA damage in cell and animal models of dry AMD and STGD1. RNA-seq analysis, corroborated by western blotting, revealed a coordinated downregulation of the GLP-1R/PKA/CREB1 signaling axis in atRAL-loaded 661W cells (Fig. 1A−E). EX-4 restored the protein levels of GLP-1R, PKA, and p-CREB1; rescued cell viability and morphology; and inhibited apoptosis in 661W cells after exposure to atRAL (Figs. 2 and 3). Knockdown of Glp1r or combined silencing of Prkaca and Prkacb impaired 661W cell survival and concurrently blocked the EX-4-induced increase in PKA/p-CREB1 signaling (Figs. 3A−D and 4A−D). Treatment with the PKA inhibitor H89 or the CREB1 inhibitor 666‑15 largely abolished the anti‑apoptotic effects of EX‑4 in atRAL‑loaded 661W cells, and reduced p‑CREB1 levels to baseline (Figs. 4G and H and 5). These lines of evidence indicate that EX-4-mediated protection of 661W cells following atRAL exposure is at least partly GLP-1R/PKA/CREB1-dependent. Abca4^−/−Rdh8^−/− mice that undergo light exposure serve as an acute model of dry AMD and STGD1 that shows severe defects of atRAL clearance in the retina [3, 7]. Importantly, intraperitoneal injection of EX-4 activated the GLP-1R/PKA/CREB1 axis in the retina of light-exposed Abca4^−/−Rdh8^−/− mice, and shielded retinal function and structure against apoptosis (Figs. 6 and 7). GLP-1R immunoblotting of retinal and 661W cell lysates revealed two predominant bands at 50 and 175 kDa (Figs. 1D and E, and 7A and B), both of which correspond to GLP-1R, as evidenced by siRNA interference assay (Fig. 1F). atRAL treatment in 661W cells, or light challenge in Abca4^−/−Rdh8^−/− mice, reduced both bands, whereas EX-4 preserved their levels (Fig. 3E and F and Fig. 7A and B).

Previous studies have shown that the function of GLP-1R is further shaped by post-translational modification, most notably N-linked glycosylation, as well as higher-order receptor organization [34, 35]. GLP-1R contains multiple N-linked glycosylation motifs within its N-terminal ectodomain, and glycosylation enhances receptor folding, stability, and surface expression [36]. Enzymatic deglycosylation with PNGase F collapses GLP-1R to lower apparent molecular weights, confirming that N-glycans substantially decrease its electrophoretic mobility [37]. Our data indicated that the 175 kDa band likely represent a glycosylated form of GLP-1R (Supplementary Materials Fig. S2), implying that the glycosylation status of GLP-1R in photoreceptors may contribute to the protective efficacy of GLP-1R agonists.

In retinal photoreceptors, PKA fine-tunes phototransduction: PKA-catalyzed phosphorylation of phosducin gates transducin signaling, thereby controlling dark current and light sensitivity [38]. In line with early reports [39, 40], activation of the GLP-1R/adenylyl cyclase/cAMP axis by EX-4 liberates the PKA catalytic subunits PRKACA and PRKACB, which translocate to the nucleus and phosphorylate CREB1 (Fig. 3). Within the visual system, CREB1 couples photic input to circadian gene expression and its activation has been implicated in photoreceptor neuroprotection [41, 42]. It has been reported that PKA-dependent phosphorylation of CREB1 at Ser133 recruits the co-activator CREB binding protein (CBP)/p300, driving the transcription of cAMP response elements (CRE)-bearing genes such as brain-derived neurotrophic factor (Bdnf), PPARG coactivator 1 alpha (Pgc-1α), and B cell leukemia/lymphoma 2 (Bcl2) that collectively promote trophic support, mitochondrial biogenesis, and anti-apoptotic defense [43–45].

Limitations of this study include the undefined molecular architecture of the higher-order GLP-1 complex, which precludes a precise delineation of GLP-1R status on photoreceptor membranes. The agonist panel was also limited because we prioritized EX-4 for its extensive preclinical validation and well-documented ocular safety profile [46]. Whether the observed rescue extends to other GLP-1R agonists remains to be determined.

On the basis of these results, we proposed a plausible mechanism by which EX-4 repressed atRAL-induced photoreceptor degeneration (Fig. 8). Excess atRAL in photoreceptor cells inhibits the GLP-1R/PKA/CREB1 signaling axis, and disrupts mitochondria, triggering cytochrome c (Cyt c) release and the caspase-9/3 apoptotic cascade, as well as PARP cleavage and γH2AX accumulation. GLP-1R activation by EX-4 couples Gα to adenylyl cyclases, elevating intracellular cAMP. cAMP binding to PKA regulatory subunits unleashes catalytic subunits that translocate to the nucleus and phosphorylate CREB1. p-CREB1 recruits CBP/p300 to CRE, thereby antagonizing apoptosis and DNA damage. In summary, EX-4 exerts anti-apoptotic and DNA-protective effects against atRAL-induced photoreceptor cell death and retinal degeneration, at least in part, through activation of the GLP-1R/PKA/CREB1 signaling pathway. Our data may position the GLP-1R/PKA/CREB1 signaling as a druggable checkpoint against atrophic macular degeneration caused by atRAL.

Fig. 8.

Proposed mechanism underlying the protective effect of EX-4 against atRAL-induced photoreceptor apoptosis

Abbreviations

AMD

Age-related macular degeneration

STGD1

Autosomal recessive Stargardt’s disease

atRAL

All-trans-retinal

Abca4

ATP-binding cassette (ABC) transporter 4

Rdh8

Retinol dehydrogenase 8

EX-4

Exendin-4

GLP-1R

Glucagon-like peptide-1 receptor

Gs

Stimulatory G protein

Gα

Guanine nucleotide-binding protein alpha subunit

Gβ

Guanine nucleotide-binding protein beta subunit

Gγ

Guanine nucleotide-binding protein gamma subunit

cAMP

Cyclic adenosine monophosphate

PKA

CAMP-dependent protein kinase

PRKACA

CAMP-dependent protein kinase catalytic subunit alpha

PRKACB

CAMP-dependent protein kinase catalytic subunit beta

CREB1

CAMP response element–binding protein 1

Caspase-3

Cysteine-dependent aspartate-specific protease-3

PARP

Poly(ADP-ribose) polymerase

γH2AX

Phosphorylated histone H2A.X (Ser139)

RPE

Retinal pigment epithelium

ONL

Outer nuclear layer

GSEA

Gene set enrichment analysis

FPKM

Fragments per kilobase of transcript per million mapped reads

PI

Propidium iodide

TUNEL

Terminal deoxynucleotidyl transferase dUTP nick end labeling

H&E staining

Hematoxylin and eosin staining

ERG

Electroretinography

GIP

Glucose-dependent insulinotropic polypeptide

CBP

CREB binding protein

CRE

CAMP response elements

Bdnf

Brain-derived neurotrophic factor

Pgc-1α

PPARG coactivator 1 alpha

Bcl2

B cell leukemia/lymphoma 2

Cyt c

Cytochrome c

Author contributions

B.H., S.L., and Y.W. conceived the study and designed the experiments. B.H., Y.C., P.C., R.X., X.S., S.G., and Y.W. performed the experiments. B.H., S.L., and Y.W. analyzed the data. J.C., S.L., and Y.W. contributed reagents or analytic tools. B.H. and Y.W. wrote the paper. All authors reviewed the manuscript and approved of the final version for submission.

Funding

This work was supported by the grants from the National Key R&D Program of China (grant no. 2024YFA1108700), the National Natural Science Foundation of China (grant no. 82471093), the Shenzhen Science and Technology Program (grant no. JCYJ20250604122759011), and the Xiamen City Science and Technology Plan Project (grant no. 3502Z20256018).

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

All animal experiments were reviewed and approved by the Institutional Animal Care and Use Committee of Xiamen University School of Medicine (approval no. XMULAC20190021; 17 December 2019). The Local Ethics Committee confirms that all experiments were conducted in accordance with the ethical principles and guidelines of the International Council for Laboratory Animal Science (ICLAS).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.