Diosgenin Attenuates Photoreceptor Degeneration in an N-Methyl-N-Nitrosourea–Induced Mouse Model of Retinal Degeneration

Abstract

Purpose

Retinitis pigmentosa (RP) is a hereditary retinal disorder distinguished by progressive photoreceptor cell (PRC) loss, in which glial activation can accelerate degeneration. Diosgenin, a natural steroidal sapogenin with potent anti-inflammatory properties, has shown therapeutic potential for ocular and neurodegenerative diseases but has not been explored for retinal degeneration. This study examined the protective role of diosgenin against PRC degeneration and explored potential anti-inflammatory mechanisms in an N-methyl-N-nitrosourea (MNU)–induced mouse model of retinal degeneration.

Methods

The model was established by intraperitoneal injection of 50 mg/kg MNU, followed by oral gavage of diosgenin/lutein (positive control)/vehicle. The effects of diosgenin on PRC structure, apoptosis, retinal function, and glial activation were evaluated. Network pharmacology and molecular docking were used to investigate potential mechanisms.

Results

Diosgenin preserved retinal integrity and function in MNU-induced mice, with relative preservation of the outer nuclear layer thickness and outer segment (OS) length of rods and cones. Disorganization of OS membrane discs and abnormal morphology of organelles were attenuated, while fundus photographs showed fewer lesions. Furthermore, diosgenin mitigated PRC apoptosis and maintained retinal light responses. Mechanistically, diosgenin ameliorated reactive gliosis of Müller glial cells (MGCs), attenuated the expression levels of inflammatory cytokines and chemokines, and mitigated activation of the IL6ST/JAK2/STAT3 pathway.

Conclusions

Diosgenin attenuated PRC degeneration in MNU-induced mice, and one of its mechanisms may involve the attenuation of reactive gliosis in MGCs and inflammatory responses, with possible involvement of the IL6ST/JAK2/STAT3 pathway. Diosgenin may serve as a potential intervention candidate for retinal degenerative diseases, such as RP.

Retinal degeneration is a chronic, progressive neurodegenerative disorder,¹ among which retinitis pigmentosa (RP) is the predominant hereditary retinal degeneration.² Until now, over 90 pathogenic genes associated with RP have been identified. Mutations in retinal genes encoding structural and functional proteins lead to the progressive degeneration and death of photoreceptor cells (PRCs), the predominant cause of vision loss in patients with RP.² Clinically, patients experience night blindness and reduced visual field, which may eventually lead to complete blindness. Although emerging therapeutic modalities, encompassing gene therapy and stem cell transplantation, have achieved partial success in specific RP types, these approaches remain costly and have limited applicability. Therefore, pharmacologic interventions that can broadly delay retinal degeneration are still of significant clinical interest.

Inflammation has a critical role in the pathogenesis of RP.³ Dead or degenerated PRCs release endogenous molecules, known as damage-associated molecular patterns, which trigger the innate immune system and promote the activation of inflammasomes in Müller glia cells (MGCs) and retinal microglia.⁴ MGCs are the primary glial cells in the retina, spanning its entire thickness and maintaining contact with virtually all neuronal types.⁵ Meanwhile, microglia are indigenous immune cells of the retina. They dynamically monitor the microenvironment to maintain homeostasis.⁶ Activated MGCs and microglia can secrete inflammatory and cytotoxic factors, which exacerbate the degeneration of PRCs, creating a vicious cycle of inflammation and neuronal loss. These two glial populations also interact dynamically during retinal inflammation. They play vital roles in maintaining retinal homeostasis, regulating cellular signals, and providing metabolic and trophic support to neurons.^7,8 Modulating retinal glial activity represents a promising anti-inflammatory therapeutic strategy independent of genetic background to attenuate retinal degeneration.

Diosgenin, a natural steroidal sapogenin predominantly found in plants of the genus Dioscorea,⁹ exhibits various pharmacologic characteristics, including anti-inflammatory, immunomodulatory, neuroprotective, hepatoprotective, and antiatherosclerotic effects.^10,11 Diosgenin is also preferred in industries as a precursor molecule for the synthesis of several steroid drugs, including progesterone, testosterone, and corticosteroids.¹² Diosgenin can delay the progression of galactosemia-induced cataract in rats¹³ and attenuate diabetic retinopathy in mice, possibly through antiapoptotic mechanisms.¹⁴ However, the potential impact of diosgenin on retinal degeneration has not been fully elucidated. In light of these findings, it can be speculated that diosgenin may exert similar regulatory effects in retinal degeneration. Lutein, a macular pigment component, has demonstrated beneficial effects in clinical studies of retinal degenerative diseases.^15,16 In addition, lutein has been shown to maintain photoreceptor survival in rd10 mice, attenuate reactive gliosis in MGCs, and exert anti-inflammatory effects.¹⁷ Therefore, based on previous studies,^18,19 lutein was used as a positive control in the present study to evaluate the effects of diosgenin.

The N-methyl-N-nitrosourea (MNU)–induced retinal degeneration model reproduces key pathologic and inflammatory features, including PRC apoptosis and glial activation, and is therefore widely used to evaluate neuroprotective compounds.²⁰ In the study, we investigated the capacity of diosgenin to attenuate PRC degeneration in MNU-induced mice. We further examined its effects on glial cells and inflammation. Additionally, network pharmacology and molecular docking analyses were used to predict potential molecular mechanisms. Collectively, this research provides insight into the anti-inflammatory and neuroprotective effects of diosgenin and suggests its potential as an intervention for retinal degeneration, such as RP.

Methods

Chemicals

Diosgenin (Sigma-Aldrich, St. Louis, MO, USA; D1634), MNU (Aladdin, Shanghai, China; N136701), and lutein (Yuanye, Shanghai, China; S24939) were used in this study. The primary antibodies used included rabbit anti-red/green opsin (1:500; Sigma-Aldrich, AB5405), anti-blue opsin (1:500; Sigma-Aldrich, AB540S), anti–BCL-2 (1:1000; ABclonal, Wuhan, China; A19693), anti-BAX (1:2000; ProteinTech, Wuhan, China; 50599-2-Ig), anti-GFAP (1:250 for immunofluorescence, 1:10,000 for Western blot; Abcam, Cambridge, UK; ab68428), anti-IBA1 (1:500; Abcam, ab289370), anti-IL6ST (1:200; Abcam, ab259927), anti-STAT3 (1:10,000; ProteinTech, 10253-2-AP), anti-phospho-STAT3 (1:2000; CST, Danvers, MA, USA; #9145), anti-phospho-JAK2 (1:2000; CST, #3776), anti-GAPDH (1:10000; ProteinTech, 10494-1-AP), mouse antibodies of anti-rhodopsin (1:200; Abcam, ab5417), and anti-JAK2 (1:1000; SANTA, Dallas, TX, USA; sc-390539). Secondary antibodies were Alexa Fluor-594–conjugated goat anti-rabbit (1:200; Invitrogen, Carlsbad, CA, USA; A32740) and anti-mouse IgG (1:200; Abcam, ab150116), horseradish peroxidase (HRP)–conjugated goat anti-rabbit and anti-mouse IgG (1:10,000; ProteinTech, SA00001-2; SAB, Greenbelt, MD, USA; #L3032), and 4′6-diamidino-2-phenylin-dole (DAPI)–containing antifade medium (Beyotime, Shanghai, China; P0131) for nuclear staining.

Animal Treatments

Male C57BL/6J mice (6–8 weeks old, 18–20g) were purchased from Charles River Laboratory Animal Technology (Beijing, China). Animals were acclimated for 1 week in a specific pathogen–free facility before the experiments and received food and water ad libitum. All experimental procedures were conducted following the ARVO guidelines for the use of animals. The study protocol was approved by the Experimental Animal Ethics Committee of Changchun Weishi Testing Technology Service Co., Ltd. (approval No. 20250507-01).

To induce retinal degeneration, mice received a single intraperitoneal injection of MNU (50 mg/kg). After MNU injection, diosgenin (50, 100, 200, or 400 mg/kg) was administered by oral gavage once daily for 7 days for dose screening. The vehicle group received an equal volume of vehicle (corn oil) under the same schedule. Retinal sections were then subjected to hematoxylin and eosin (H&E) staining. Mice were subsequently divided into five groups: normal control (NC), MNU-induced model, vehicle, diosgenin (200 mg/kg), and lutein (100 mg/kg)^18,21 groups. Oral gavage was performed once daily. Animals were evaluated on days 3 and 7 after the MNU injection. The difference between endpoint and baseline body weights was recorded for each group. A schematic diagram of experimental design is shown in Figure 1A.

Figure 1.

Diosgenin preserves ONL thickness in MNU-induced mice. (A) Molecular structures of MNU and diosgenin (left) and schematic timeline of the animal experiment (right). (B) H&E-stained retinal sections from MNU-induced mice treated with vehicle or increasing doses of diosgenin. Scale bar: 50 µm. (C) Quantification of the ONL/INL ratio, reflecting relative photoreceptor layer thickness, shows the dose-dependent effect of diosgenin in MNU-induced mice. Data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001 (n = 5). (D) H&E-stained images of retinal sections from each group. Scale bar: 50 µm. (E) Quantification of ONL thickness among different groups. Data are presented as mean ± SD. ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 vs. NC; *P < 0.05, **P < 0.01, ***P < 0.001 vs. vehicle (n = 5). GCL, ganglion cell layer; i.p., intraperitoneal injection; qRT-PCR, quantitative reverse transcription polymerase chain reaction; TEM, transmission electron microscopy.

Electroretinogram

Mice were dark-adapted for 12 hours before recording. Under low red light conditions, the pupils of the mice were dilated with atropine ophthalmic gel, and the animals were anesthetized after 30 minutes. Conductive cream was applied to the tail of the mice, which was then connected to a ground electrode. A reference electrode was inserted subcutaneously between the ears, and a corneal electrode was placed in contact with the corneal surface. Measurements were performed using a visual electrophysiology system (Chongqing Aierxi Medical Equipment Co., Ltd, China). The a-wave amplitude was defined as the maximum negative deflection from baseline following the stimulus, and the b-wave amplitude was defined as the maximum positive peak measured from the a-wave trough.

Fundus Photography

We dilated the pupils of the mice following anesthesia, and fundus photographs were captured using a small-animal fundus imaging system (IIScience, Pleasanton, CA, USA). The extent of retinal damage was assessed based on the fundus photographs.

Histology

Eyeballs were fixed in a formalin/acetic acid/saline fixative overnight at 4°C, embedded in paraffin, and sectioned along the sagittal plane. Retinal sections (2 µm thick) containing the optic nerve head (ONH) were stained with H&E. For drug safety assessment, diosgenin (200 mg/kg/d) was administered to normal mice by oral gavage for 7 days. Major organs (heart, liver, spleen, lungs, and kidneys) were then collected and fixed in 4% paraformaldehyde solution overnight at 4°C to prepare 2-µm-thick paraffin sections for H&E staining. Tissue was observed under a light microscope (Olympus, Tokyo, Japan), and the thicknesses of the outer nuclear layer (ONL) and inner nuclear layer (INL) were measured using ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Immunofluorescence and TUNEL Assay

Eyeballs were freshly embedded in optimal cutting temperature compound, frozen, and sectioned at a 10-µm thickness, with sections including the ONH. Immunofluorescence staining was then performed. Apoptotic cells were detected using a TUNEL assay kit (Absin, Shanghai, China; abs50047) according to the manufacturer's instructions. Nuclei were counterstained with DAPI-containing mounting medium. A confocal microscope (S3000; HOOKE, Changchun, China) was used to capture the images. To minimize spatial variability across different retinal regions, mid-peripheral retinal areas located 600 to 1200 µm from the ONH at equivalent eccentricities were consistently selected for analysis. All measurements were obtained from corresponding regions on both sides of the ONH in each section and averaged. For length-related parameters, multiple equidistant points within the selected region were sampled and averaged. All image analyses were performed using ImageJ software.

Transmission Electron Microscopy

After overnight fixation of tissues in 2.5% glutaraldehyde at 4°C, they were rinsed thoroughly and then treated with 1% osmium tetroxide for 2 hours at ambient temperature. The samples were dehydrated through an ascending ethanol series, embedded in epoxy resin, and cut into 100-nm-thick ultrathin sections. Sections were double-stained with uranyl acetate and lead citrate, and ultrastructural images were captured using a transmission electron microscope (HT7800; Hitachi, Tokyo, Japan).

Western Blot Analysis

Retinas were dissected, homogenized, and lysed in RIPA buffer (Beyotime; P0013B) supplemented with phenylmethylsulfonyl fluoride (Beyotime, ST506) and phosphatase inhibitors (Seven, Beijing, China; SW107-02). Protein concentrations were determined using a BCA Protein Assay Kit (Beyotime, P0010). Lysates were boiled for 10 minutes and stored at −20°C. Subsequently, the lysates were subjected to SDS-PAGE, transferred to polyvinylidene difluoride membranes, and blocked before incubation with primary antibodies overnight at 4°C. Membranes were then incubated with HRP-conjugated secondary antibodies, and immunoreactive bands were visualized by chemiluminescence. Band intensities were quantified using ImageJ software.

Quantitative RT-PCR

Total RNA was extracted from mouse retinas using a Total RNA Extraction Kit (Seven, SM130-01), and cDNA was synthesized with the cDNA Synthesis Kit (Seven, SM134-01), according to the manufacturer's instructions. Quantitative RT-PCR was performed using SYBR Green qPCR MasterMix II (Seven, SM143-01). Relative mRNA expression levels were calculated using the 2^−ΔΔCT method, with normalization to an internal control gene. Primer sequences are provided in Supplementary Table S1.

Network Pharmacology and Molecular Docking

Detailed procedures for network pharmacology and molecular docking are described in Supplementary Text S1.

Statistical Analyses

Data distribution was assessed before analysis. Normally distributed data were analyzed using one-way ANOVA when the assumption of homogeneity of variance was met, followed by Bonferroni post hoc testing; otherwise, Welch's ANOVA, followed by Games–Howell post hoc testing, was used. Nonnormally distributed data were analyzed using the Kruskal–Wallis test. A P value <0.05 was considered statistically significant. Statistical analyses were performed using R (R Project for Statistical Computing, Vienna, Austria), and graphs were generated using GraphPad Prism 9 (GraphPad Software, La Jolla, CA, USA).

Results

Diosgenin Preserves the Structure and Morphology of MNU­Induced Mice Retinal PRCs

H&E staining was used to evaluate the effects of different doses of diosgenin (50, 100, 200, and 400 mg/kg) (Fig. 1B). Compared with the vehicle group, diosgenin-treated groups showed higher ONL/INL ratios across retinal regions. Among them, the 200-mg/kg dose exhibited the most pronounced effect (Fig. 1C). Therefore, 200 mg/kg was selected for subsequent experiments. ONL thickness was compared among the NC, MNU-induced model, vehicle, diosgenin, and lutein groups. Compared with the NC group, ONL thickness was reduced in the MNU-induced model and vehicle groups. In contrast, diosgenin and lutein attenuated the reduction in ONL thickness compared with the vehicle group, with diosgenin showing a more pronounced effect at most sites (Figs. 1D, 1E).

Rhodopsin, a light-sensitive protein expressed in the rod outer segments (OSs),²² showed preserved structure in the NC group but was significantly degenerated and shortened in the MNU-induced model and vehicle groups. Compared with the vehicle group, both diosgenin and lutein attenuated the reduction in rod OS length, with diosgenin showing a more pronounced effect (Figs. 2A, 2B; Table 1). Opsins are light-sensitive proteins expressed in cone OSs.²³ Similarly, the green and blue cone OS lengths were significantly shortened in the MNU-induced model and vehicle groups compared with the NC group. The green and blue cone OS lengths were better preserved in the diosgenin group compared with the vehicle group, showing greater efficacy than the lutein group (Figs. 2C–F; Table 1).

Figure 2.

Diosgenin preserves photoreceptor OS length in MNU-induced mice. (A) Immunofluorescence staining of retinal sections showing rod OS (anti-rhodopsin). The boxed region is enlarged (right). White reference lines indicate the measured rod OS length. (B) Quantification of rod OS lengths among different groups. (C) Immunofluorescence staining of retinal sections showing green cone OS (anti-red/green opsin). The boxed region is enlarged (right). White reference lines drawn along the longitudinal axis of the cone OS indicate the measured green cone OS length. (D) Quantification of green cone OS lengths among different groups. (E) Immunofluorescence staining of retinal sections showing blue cone OS (anti-blue opsin). The boxed region is enlarged (right). White reference lines drawn along the longitudinal axis of the cone OS indicate the measured blue cone OS length. (F) Quantification of blue cone OS lengths among different groups. Nuclei were stained with DAPI. Scale bar: 100 µm. Data are presented as mean ± SD. ns, not significant (P ≥ 0.05); ^###P < 0.001 vs. NC; *P < 0.05, ***P < 0.001 vs. vehicle (n = 5).

Table 1.

Photoreceptor OS Length and GFAP Length Across Groups

Characteristic	NC	MNU	MNU + Vehicle	MNU + Diosgenin	MNU + Lutein
Rhodopsin	27.33 ± 4.26	10.68 ± 1.83	9.65 ± 1.95	20.14 ± 1.85	15.78 ± 2.86
Red/green opsin	14.22 ± 1.52	7.35 ± 1.24	6.21 ± 1.57	12.63 ± 1.75	10.04 ± 2.25
Blue opsin	14.58 ± 1.90	5.69 ± 1.09	5.71 ± 1.28	11.02 ± 1.25	8.67 ± 1.91
GFAP	13.48 ± 1.70	111.86 ± 13.88	111.43 ± 8.67	60.92 ± 6.21	90.48 ± 6.49

Data are presented as mean ± SD (µm).

Transmission electron microscopy images revealed the ultrastructure of PRCs. Rounded and swollen mitochondria, with shortened cristae, swollen and deformed endoplasmic reticulum (ER), and Golgi apparatus, were observed in the inner segments (ISs) of PRCs in the vehicle group. Diosgenin intervention ameliorated the morphology of these organelles (Fig. 3A–D). Cross-sectional views of the connecting cilia with 9+0 microtubule arrays were observed in all groups (Figs. 3E, 3F). The OS membrane disc structure was disordered in the vehicle group, but diosgenin intervention delayed OS membrane disc degeneration (Fig. 3G). Fundus photographs showed numerous dense yellow-whitish punctate lesions in the MNU-induced model and vehicle groups. In contrast, the extent of the lesions was alleviated in the lutein and diosgenin groups (Fig. 3H). Collectively, these findings suggest that diosgenin effectively preserves the PRCs and retinal structure and morphology in MNU-induced mice.

Figure 3.

Diosgenin preserves PRC ultrastructure and fundus morphology in MNU-induced mice. (A–G) Transmission electron microscopy images of retinas from each group. (A) IS overview. Scale bar: 5 µm. (B) Mitochondria (white arrows). Scale bar: 500 nm. (C) Endoplasmic reticulum (white arrows). Scale bar: 500 nm. (D) Golgi apparatus (white arrows). Scale bar: 500 nm. (E) The boundary between IS and OS. Scale bar: 5 µm. (F) Connecting cilia (white arrows). Scale bar: 500 nm. (G) Membrane disc. Scale bar: 500 nm. (H) Fundus photographs from each group.

Diosgenin Mitigates PRC Apoptosis in MNU-Induced Mice

TUNEL staining revealed no apoptotic cells in the retina of the NC group. In contrast, numerous apoptotic cells were observed in the MNU-induced model and vehicle groups, concentrated in the ONL. Compared with the vehicle group, both lutein and diosgenin attenuated the proportion of apoptotic cells, with diosgenin showing a lower level (Figs. 4A, 4B). The BAX/BCL-2 ratio reflects the relative balance between proapoptotic and antiapoptotic signals and serves as an indicator of apoptotic tendency. Western blot analysis showed that the BAX/BCL-2 ratio was increased in the MNU-induced model and vehicle groups, consistent with the above findings. In contrast, diosgenin attenuated this increase (Figs. 4C, 4D). BAX and BCL-2 levels, normalized to GAPDH, are presented in Supplementary Figures S1A and S1B. These findings indicate that diosgenin effectively mitigates PRC apoptosis in MNU-induced mice.

Figure 4.

Diosgenin mitigates PRC apoptosis in MNU-induced mice. (A) TUNEL staining images of retinal sections showing apoptotic cells. Nuclei were stained with DAPI. Scale bar: 100 µm. (B) Quantification of apoptotic PRCs among different groups (n = 5). (C) Western blot images of BAX and BCL-2 expression in the retinas from each group. (D) Quantification of the ratio of BAX to BCL-2 expression in Western blot among different groups (n = 5). Data are presented as mean ± SD. ns, not significant (P ≥ 0.05); ^##P < 0.01, ^###P < 0.001 vs. NC; *P < 0.05, ***P < 0.001 vs. vehicle. BAX, BCL-2–associated X protein; BCL-2, B-cell lymphoma 2; GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Diosgenin Preserves Retinal Light Responses in MNU­Induced Mice

ERG analysis revealed strong light responses in the NC group, whereas those in the MNU-induced model and vehicle groups were significantly reduced at all tested flash intensities. Under scotopic conditions, at a flash intensity of 0.1 cd·s/m², compared with the vehicle group, lutein attenuated the reduction in b-wave amplitude, whereas diosgenin attenuated the reduction in both a- and b-wave amplitudes (Figs. 5A, 5B). However, at a flash intensity of 1.0 cd·s/m², a- and b-wave amplitudes in the lutein and vehicle groups did not differ significantly, whereas those in the diosgenin group were partially preserved (Figs. 5C, 5D). At a flash intensity of 10.0 cd·s/m², the results were similar to those obtained at 1.0 cd·s/m² (Figs. 5E, 5F). Under photopic conditions, at a flash of 1.0 cd·s/m², compared with the vehicle group, lutein partially preserved the a-wave amplitude, whereas diosgenin partially preserved both a- and b-wave amplitudes (Figs. 5G, 5H). The a- and b-wave amplitudes from different groups at the tested flash intensities are shown in Table 2. These results indicate that diosgenin exerts a stronger protective effect than lutein. Given that bipolar cell responses depend on PRC input, diosgenin appears to better preserve retinal function in both rod- and cone-mediated light pathways in MNU-induced mice.

Figure 5.

Diosgenin preserves retinal light responses in MNU-induced mice. (A) Representative retinal electrical waveforms in response to a flash under dark-adapted 0.1 ERG. (B) Quantification of amplitudes of the a- and b-waves under dark-adapted 0.1 ERG. (C) Representative retinal electrical waveforms in response to a flash under dark-adapted 1.0 ERG. (D) Quantification of amplitudes of the a- and b-waves under dark-adapted 1.0 ERG. (E) Representative retinal electrical waveforms in response to a flash under dark-adapted 10.0 ERG. (F) Quantification of amplitudes of the a- and b-waves under dark-adapted 10.0 ERG. (G) Representative retinal electrical waveforms in response to a flash under light-adapted 1.0 ERG. (H) Quantification of amplitudes of the a- and b-waves under light-adapted 1.0 ERG. Data are presented as mean ± SD. ns, not significant (P ≥ 0.05); ^##P < 0.01, ^###P < 0.001 vs. NC; *P < 0.05, **P < 0.01, ***P < 0.001 vs. vehicle (n = 5).

Table 2.

Amplitudes of the a- and b-Waves From Different Groups of the ERG

Characteristic	NC	MNU	MNU + Vehicle	MNU + Diosgenin	MNU + Lutein
Dark-adaptation 0.1 ERG
a-wave	56.44 ± 9.13	13.48 ± 1.01	14.18 ± 2.04	35.14 ± 4.31	19.88 ± 6.40
b-wave	133.28 ± 31.76	27.22 ± 2.98	26.42 ± 3.71	80.58 ± 8.06	62.02 ± 11.09
Dark-adaptation 1.0 ERG
a-wave	75.74 ± 14.81	19.22 ± 4.23	18.20 ± 2.99	53.64 ± 8.16	27.74 ± 5.05
b-wave	191.50 ± 21.58	48.20 ± 8.96	52.34 ± 6.95	97.58 ± 9.29	74.46 ± 18.00
Dark-adaptation 10.0 ERG
a-wave	86.16 ± 12.25	22.50 ± 4.73	24.60 ± 5.88	66.76 ± 6.83	33.26 ± 6.81
b-wave	255.20 ± 47.62	49.44 ± 5.63	56.32 ± 13.80	134.36 ± 20.49	83.46 ± 18.43
Light-adaptation 1.0 ERG
a-wave	35.10 ± 6.55	6.3 ± 4.16	5.72 ± 0.80	25.38 ± 4.04	15.16 ± 3.35
b-wave	72.08 ± 7.12	16.02 ± 5.73	12.92 ± 3.00	46.14 ± 14.97	24.48 ± 1.97

Data are presented as mean ± SD (µV).

Diosgenin Mitigates MGC Reactive Gliosis in MNU­Induced Mice

Immunofluorescence staining showed that glial fibrillary acidic protein (GFAP) expression in the NC group was confined to astrocytes in the ganglion cell layer, with inactive MGCs. In contrast, the MNU-induced model and vehicle groups showed increased MGC activation, with GFAP-positive processes extending into the ONL and significantly increasing in length. Compared with the vehicle group, both lutein and diosgenin exhibited shorter GFAP extension lengths, indicating attenuated MGC activation, with diosgenin showing a greater effect (Figs. 6A, 6B; Table 1). These findings were further supported by Western blot results (Figs. 6C, 6D). Furthermore, microglia in the NC group were primarily distributed in the inner retinal layers. In contrast, the ionized calcium-binding adaptor molecule 1 (IBA1)–positive area increased in the MNU-induced model and vehicle groups, and activated microglia increased in number and infiltrated the outer retinal layers. Compared with the vehicle group, the lutein or diosgenin groups showed fewer activated microglia in the retina; however, the difference was not significant (Figs. 6E, 6F). These results suggest that diosgenin significantly attenuates reactive gliosis in MGCs; however, its effects on microglia remain to be further validated.

Figure 6.

Diosgenin mitigates reactive gliosis in MGCs in MNU-induced mice. (A) Immunofluorescence staining of retinal sections showing the activation of MGCs stained with anti-GFAP. Nuclei were stained with DAPI. Scale bar: 100 µm. (B) Quantification of the GFAP length in the retina among different groups (n = 5). (C) Western blot images of GFAP expression in the retinas from each group. (D) Quantification of the ratio of GFAP to GAPDH expression in Western blot among different groups (n = 7). (E) Immunofluorescence staining of retinal sections showing the microglia stained with anti-IBA1. Nuclei were stained with DAPI. Scale bar: 100 µm. (F) Quantification of the area of IBA1⁺ of the retina among different groups (n = 5). (G) Quantitative RT-PCR analysis showing the mRNA levels of Tnfa, Il6, Il1b, and Ccl2 among different groups (n = 3). Data are presented as mean ± SD. ns, not significant (P ≥ 0.05); ^##P < 0.01, ^###P < 0.001 vs. NC; *P < 0.05, **P < 0.01, ***P < 0.001 vs. vehicle. Ccl2, c-c motif chemokine ligand 2; IBA1, ionized calcium-binding adapter molecule 1; Il6, interleukin-6; Il1b, interleukin-1 beta; Tnfa, tumor necrosis factor–α; β-actin, beta-actin.

The release of inflammatory cytokines and chemokines accompanies activation of MGCs.^24,25 Therefore, we assessed the mRNA expression levels of Tnfa, Il6, Il1b, and Ccl2. These inflammatory factors were significantly upregulated in the retinas of both the MNU-induced model and vehicle groups compared with the NC group. Compared with the vehicle group, the lutein group showed slight attenuation of inflammatory factor expression, whereas the diosgenin group exhibited a more pronounced effect (Fig. 6G). Together, these findings suggest that diosgenin significantly mitigates retinal inflammation in MNU-induced mice, and the attenuation of reactive gliosis in MGCs may represent one possible underlying mechanism.

Prediction of Potential Targets and Signaling Pathways Through Network Pharmacology and Molecular Docking

Figures 7A–E shows the results of network pharmacology. Supplementary Table S2 shows the degree values of targets. Among these signaling pathways, Janus kinase 2 (JAK2)/signal transducer and activator of transcription 3 (STAT3) has been proven to be closely linked to inflammation,²⁶ and IL-6 signal transducer (IL6ST) is also one of the cross-targets. Molecular docking showed the best binding conformation and potential binding sites of diosgenin with IL6ST, JAK2, and STAT3, and diosgenin could bind closely with them (Figs. 7F–H). We then focused on IL6ST/JAK2/STAT3 for subsequent experiments.

Figure 7.

Prediction of potential targets and signaling pathways through network pharmacology and molecular docking. (A–E) Network pharmacology. (A) A total of 79 diosgenin targets, 5024 RP-related genes, and 3590 inflammation-related genes were identified, with 35 overlapping targets obtained using Venny2.1.0. (B) The 35 overlapping targets were used to construct a PPI network in STRING 12.0; after removing unconnected nodes, 21 targets remained. Edges represent protein–protein interactions. (C) The PPI network was reconstructed using Cytoscape 3.9.1. Seven targets with above-average degree values are shown in the inner circles and the others in the outer circles. Nodes represent proteins and edges represent interactions. Node color (red to yellow) and size (large to small) indicate decreasing degree values. (D–E) GO and KEGG enrichment analyses of 21 targets were conducted using DAVID. (D) A total of 174 BP, 42 MF, and 29 CC terms were significantly enriched, and the top 10 terms in each category are shown. (E) KEGG analysis identified 96 significantly enriched pathways, of which the top 25 with the lowest P values are presented. Node size represents the number of genes, and node color indicates the P value. (F–H) Molecular docking. (F) The binding pose of diosgenin is embedded in the binding pocket of IL6ST, with GLN-210 as the key binding residue. The binding energy is –9.7 kcal/mol. (G) The binding pose of diosgenin is embedded in the binding pocket of JAK2, with THR-998 as the key binding residue. The binding energy is –9.9 kcal/mol. (H) The binding pose of diosgenin is embedded in the binding pocket of STAT3, with GLN-232 as the key binding residue. The binding energy is −7.9 kcal/mol. ABL1, ABL proto-oncogene 1, nonreceptor tyrosine kinase; AKT1, AKT serine/threonine kinase 1; AR, androgen receptor; BP, covering biological process; CC, cellular component; CCND1, cyclin d1; CDK4, cyclin-dependent kinase 4; CDK6, cyclin-dependent kinase 6; CDKN1A, cyclin-dependent kinase inhibitor 1a; ESR1, estrogen receptor 1; HIF1A, hypoxia inducible factor 1 subunit alpha; IGF1R, insulin like growth factor 1 receptor; IL6ST, interleukin 6 signal transducer; INSR, insulin receptor; JAK2, Janus kinase 2; MAPK14, mitogen-activated protein kinase 14; MET, MET proto-oncogene, receptor tyrosine kinase; MF, molecular function; MTOR, mechanistic target of rapamycin; NTRK1, neurotrophic receptor tyrosine kinase 1; NTRK2, neurotrophic receptor tyrosine kinase 2; PDGFRB, platelet-derived growth factor receptor beta; SRC, SRC proto-oncogene, nonreceptor tyrosine kinase; STAT3, signal transducer and activator of transcription 3; TP53, tumor protein p53.

Diosgenin Attenuates IL6ST/JAK2/STAT3 Pathway Activation in the Retinas of MNU-Induced Mice

Immunofluorescence analysis showed that the average fluorescence intensity of IL6ST was significantly increased in the MNU-induced model and vehicle groups compared with the NC group, predominantly from the inner plexiform layer to the outer plexiform layer. Compared with the vehicle group, the diosgenin group exhibited a lower mean fluorescence intensity of IL6ST (Figs. 8A, 8B). Similarly, Western blot analysis showed lower p-JAK2/JAK2 and p-STAT3/STAT3 ratios in the diosgenin group compared with the vehicle group (Figs. 8C–F). These results suggest that diosgenin mitigates reactive gliosis in MGCs and attenuates the inflammatory response, which may be associated with the mitigation of IL6ST/JAK2/STAT3 pathway activation as a potential underlying mechanism.

Figure 8.

Diosgenin attenuates IL6ST/JAK2/STAT3 pathway activation in the retinas of MNU-induced mice. (A) Immunofluorescence staining of retinal sections stained with anti-IL6ST. Nuclei were stained with DAPI. Scale bar: 100 µm. (B) Quantification of the average fluorescence intensity of IL6ST among different groups (n = 5). (C) Western blot images of p-JAK2 and JAK2 expression in the retinas from each group. (D) Quantification of the ratio of p-JAK2 to JAK2 expression in Western blot among different groups (n = 7). (E) Western blot images of p-STAT3 and STAT3 expression in the retinas from each group. (F) Quantification of the ratio of p-STAT3 to STAT3 expression in Western blot among different groups (n = 7). Data are presented as mean ± SD. ns, not significant (P ≥ 0.05); ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 vs. NC; *P < 0.05, ***P < 0.001 vs. vehicle. ELM, external limiting membrane.

Additionally, body weight changes did not differ significantly among the groups (Supplementary Fig. S1C). No histopathologic abnormalities were observed in major organs after diosgenin administration (Supplementary Fig. S1D), which was harmless in traditional use.¹² These findings suggest no significant systemic toxicity associated with diosgenin administration.

Discussion

Diosgenin is known for its potent anti-inflammatory capabilities,²⁷ yet its protective potential in retinal degenerative diseases has remained largely unexplored. In this study, we investigated its effects using an MNU-induced mouse model of retinal degeneration, and the results, including ONL thinning, PRC OS loss, PRC apoptosis, MGC activation, microglial migration to the outer retina, and increased IBA1 expression, were consistent with previous reports.^20,28 However, in the present study, diosgenin tended to exhibit greater efficacy than lutein in most cases, suggesting its potential as a promising therapeutic option.

Administration with 200 mg/kg diosgenin significantly attenuated MNU-induced PRC degeneration, preserved ONL thickness and OS length in rods and cones, mitigated OS membrane disc degeneration, and ameliorated morphologic abnormalities in the IS. Yellow-whitish punctate lesions were observed in the fundus photographs of the MNU-induced mice, although their cellular origin remains difficult to determine. Previous studies have reported widespread yellow-whitish lesions in MNU models, which, together with the concurrent loss of the ONL on optical coherence tomography (OCT) or histology, are consistent with outer retinal injury.²⁹ It should be noted that inflammatory cell accumulation may also contribute to the formation of such lesions, and findings from OCT combined with immunostaining support the presence of an immune cell component.^30,31 A “fundus lesion–OCT–histology” colocalization strategy may help further clarify their origin. PRC structure and morphology in the retina were preserved relatively intact. Diosgenin also significantly mitigated PRC apoptosis and maintained retinal light responses, which demonstrates functional neuroprotection. In the early stages of retinal injury, reactive gliosis of MGCs may be beneficial, as they release protective factors such as neurotrophic factors,⁵ scavenge and recycle neurotransmitters, phagocytose PRC debris, mitigate local inflammatory responses, and limit subsequent injury to neighboring PRCs.^32,33 However, chronic glial proliferation is detrimental, with key factors including edema formation, neuronal hyperexcitability, and glutamate toxicity, leading to secondary PRC loss.^8,34 Here, it was observed that diosgenin markedly mitigated the expression of GFAP and the levels of inflammatory cytokines (Tnfa, Il6, Il1b, and Ccl2), suggesting its potential to mitigate reactive gliosis of MGCs and inflammation effectively.

Studies have shown that after IL-6 family cytokines bind to their specific receptors, they further form a complex with IL6ST and then activate the downstream JAK2/STAT3 signaling cascade, thereby regulating the transcription of multiple genes.³⁵ Our analysis demonstrated that diosgenin mitigated the increases in IL6ST, p-JAK2, and p-STAT3 protein levels. This suggests that attenuation of the aberrant activation of the IL6ST/JAK2/STAT3 pathway, which is a critical upstream regulator of reactive gliosis of MGCs, may be one of the mechanisms underlying the protective effects of diosgenin. An earlier report revealed that early activation of MGCs is abolished in IL6ST/JAK/STAT-deficient mice.³⁶ Inhibition of JAK/STAT and MAPK pathways attenuates TNF-α–induced MGC proliferation in vitro.³⁷ Clinically, reduced activation of the JAK pathway may contribute to lowering the risk of age-related macular degeneration.³⁸ The JAK/STAT3 signaling pathway also mediates crosstalk between MGCs and PRCs.³⁹ For example, intervention with wolfberry extract has been proven to delay retinal degeneration in rd10 mice, partly through the mitigation of STAT3 and GFAP expression in MGCs.⁴⁰ Similarly, systemic administration of STAT3 inhibitors exerts strong therapeutic effects in rats with diabetic retinopathy, largely due to their anti-inflammatory effects.⁴¹ Conversely, activation of the JAK/STAT and MAPK pathways by mesenchymal stem cell–derived retinal progenitor cells helps functional recovery in rd12 mice,⁴² and ciliary neurotrophic factor protects against retinal degeneration in rds mice by maintaining the protective properties of MGCs through IL6ST/STAT3 signaling.⁴³ However, this activation can also result in loss of cone immunoreactivity and disorganization of MGCs.⁴⁴ The reason for these contrasting outcomes may be that MGC gliosis can exert beneficial or detrimental effects depending on the microenvironment. Furthermore, as this pathway is also expressed in other retinal cells, including PRCs, its function exhibits cell-type–specific differences. Therefore, modulation of the IL6ST/JAK2/STAT3 pathway may lead to divergent outcomes depending on contextual factors and cellular.

This study has several limitations. Owing to the short interval between MNU and diosgenin administration, diosgenin is more likely to have exerted preventive effects on MNU-induced injury and downstream responses rather than reversing established damage. Future studies using delayed administration protocols are needed to further clarify its therapeutic potential in preexisting retinal injury. Our results showed that diosgenin attenuated PRC degeneration and preserved retinal integrity, and we speculate that modulation of glial activity and inflammation may be one of its underlying mechanisms. However, the current experimental design does not permit definitive causal inference, and further investigation in specific cell types is required. In addition, the effects of diosgenin on microglia remain insufficiently characterized and require further rigorous validation, for example, by incorporating additional markers. Moreover, diosgenin may also exert protective effects through multiple other mechanisms, including direct actions on PRCs and regulation of oxidative stress,⁴⁵ ER stress,⁴⁶ and other pathways. Other signaling molecules may serve as potential targets, and their specific roles remain to be further explored. Despite these limitations, our findings provide evidence that diosgenin confers structural and functional protection in MNU-induced retinal degeneration and offer mechanistic insights into its anti-inflammatory effects.

In conclusion, this study demonstrates that diosgenin attenuates PRC degeneration and preserves retinal structure, morphology, and function in MNU-induced mice. One of its protective mechanisms may involve the attenuation of reactive gliosis in MGCs and the inflammatory response, possibly through modulation of the IL6ST/JAK2/STAT3 pathway. These findings provide new insight into the neuroprotective effects of diosgenin and suggest that it may serve as a potential candidate for the intervention of retinal degenerative diseases such as RP.