Sleep deprivation exacerbates lens‑induced myopia in association with ERK/NF‑κB pathway activation

Introduction

Myopia is a global eye disease that can lead to visual impairment. The primary symptoms include decreased visual acuity, reduced accommodation and visual fatigue (1,2). An epidemiological study estimated that half of the global population will have myopia by 2025, with 9.8% developing high myopia (3). High myopia may trigger various severe complications, posing a public health concern requiring particular attention (4,5). Some studies have revealed negative impacts of myopia on adolescents' physical and emotional functioning, substantially reducing their quality of life and hindering physical and mental development (6–8). Therefore, investigating the causes of juvenile myopia is essential to address this issue. Jin et al (9) found that improved sleep quality, earlier bedtimes and later wake times exert protective effects against myopia. Chawla et al (10) also suggested that sleep duration, along with other environmental factors, influences myopia progression. Thus, increasing sleep duration may slow myopia progression. However, whether sleep deprivation (SD) is associated with myopia development and the underlying mechanisms remains unclear. In the present study, guinea pig lens-induced myopia (LIM) models were established to examine the effects of SD on diopter, axial length and retinal signaling factors related to myopia, such as dopamine (DA) and tyrosine hydroxylase (TH). The present study evaluated changes in related inflammatory factors induced by SD in the LIM model, aiming to clarify the specific molecular mechanisms involved. This research provides a theoretical molecular foundation for comprehending the association between SD and myopia development in adolescents.

Materials and methods

Establishment and grouping of guinea pig LIM and LIM + SD models

In April 2025, 60 healthy 21-day-old tricolor guinea pigs (30 male, 30 female), weighing 140–170 g, were obtained from the Hubei College of Chinese Medicine [Wuhan, China; license no. SCXK (E) 2026-0017]. Animals were randomly assigned to the four groups with balanced sex distribution (7–8 males and 7–8 females per group). Animals were housed at 25°C under 12 h light/dark cycles (7:00 AM-7:00 PM), with illumination levels between 480 and 520 lx. The SD device used in the present study adopts a motion rest cycle mode, which interrupts animal sleep by rotating a metal rod. In the first phase, 30 guinea pigs were randomized into the LIM and LIM + SD groups (15 per group) using the random number table method. The right eye (OD) of each guinea pig was treated as the defocus-induced eye and the contralateral untreated left eye (OS) served as the control. Guinea pigs in the LIM group wore −6.0 D concave lenses for 14 days to establish the LIM model (Fig. 1A). Guinea pigs in the LIM + SD group were placed in AUT-SD-1S animal SD chambers, in addition to wearing concave lenses, with five animals per chamber. The cylindrical chamber was made of organic glass (diameter, 50 cm; height, 40 cm). The bottom featured a rotating metal rod (46 cm in length) with adjustable speed (Fig. 1B). Chamber illumination met LIM modeling requirements, with the light intensity at guinea pig eye level measured using a digital luxmeter and controlled between 480 and 520 lx. Animals had free access to food and water inside the chamber. The rod rotated at a speed of one revolution every 15 sec, performing 3 consecutive revolutions every 10 min. Guinea pigs underwent continuous SD for 20 h each day (7:00 AM-3:00 AM) to establish the SD model. This duration was designed to achieve moderate-to-severe sleep deprivation, covering the full light phase and part of the dark phase of the animals' circadian cycle (11).

Figure 1. Establishment of guinea pig LIM and LIM + SD models. (A) Guinea pig LIM model diagram. (B) Animal SD device. LIM, lens-induced myopia; SD, sleep deprivation.

In the second phase, the remaining 30 guinea pigs were randomly divided into LIM + SD and LIM + SD + BVD-523 groups (15 per group) by adopting the random number table method. The LIM + SD group underwent modeling as aforementioned. Guinea pigs in the LIM + SD + BVD-523 group received a single oral dose of BVD-523 (cat no. HY-15816; MedChemExpress) at 30 mg/kg. Due to facility and resource constraints, the experiments were conducted in two sequential phases. Phase 1 compared LIM vs. LIM + SD to establish the baseline effect of SD. Phase 2 was subsequently designed to evaluate the pharmacological rescue with BVD-523 using a separate cohort. To minimize potential batch effects, all animals were age and weight-matched, housed under identical environmental conditions (12 h light/dark cycle, 25°C) and experiments were performed using the same equipment and standard operating procedures by the same personnel. Furthermore, baseline diopter and axial length did not differ significantly between the LIM + SD groups of Phase 1 and Phase 2 (P>0.05), confirming cohort comparability.

During the experiment, animal health and behavior were monitored three times a day (every 8 h: 7:00 AM, 3:00 PM, 11:00 PM). Each monitoring focused on observing the following aspects: Overall body condition, motor function, food and water intake, eye condition, and whether there are any abnormalities in mental behavior. After 2 weeks, 60 guinea pigs were euthanized for tissue sampling. Throughout the entire experiment, no animals died, and no animals were euthanized prematurely due to meeting humane endpoint criteria (including weight loss exceeding 20% of baseline weight, complete refusal to eat or drink, severe drowsiness without response to stimuli, severe pain that cannot be relieved, severe impairment of motor function, severe respiratory distress, or obvious eye damage infection). The cause of death for all animals was anesthesia induced euthanasia performed at the endpoint of the experiment. All guinea pigs were euthanized by carbon dioxide (CO2) inhalation. The procedure was strictly performed in accordance with the following specifications: A compressed CO2 cylinder with a purity of ≥99% was used, with a slow filling rate of 30–70% of the chamber volume per min and ventilation was continued for at least 1 min after the animals stopped breathing. After euthanasia, the criteria for independent confirmation by two trained observers are complete cessation of heartbeat and breathing for at least 5 min, and permanent disappearance of corneal reflex and toe pinch reflex. The Animal Welfare and Ethics Review Committee of Hubei Beiente Experimental Center approved the study (approval no. IACUC-BNB-2025-024). All animal procedures adhered to the Regulations on the Management of Laboratory Animals issued by the State Scientific and Technological Commission of the PRC.

Measurement of diopter and axial length

All measurements were conducted by a technician who had no knowledge of group allocation (single blind design). Refractive power and axial length were measured at the endpoint (14 days after modeling). Diopter was measured using a computerized refractometer (Topcon KR-800; Topcon Medical Shanghai Co., Ltd.). Before retinoscopy, bilateral cycloplegia was induced by topical administration of compound tropicamide eye drops every 5 min for three applications. Diopter was calculated according to the spherical equivalent principle and the mean of three measurements was recorded. Axial length was measured using an ophthalmic A-scan biometer (10 MHz). After local anesthesia with one drop of 0.4% oxybuprocaine hydrochloride eye drops (12–15), the guinea pig's head was horizontally fixed, and the probe was aligned perpendicular to the corneal surface and directed toward the center of the pupil to obtain a stable and clear waveform. Axial length was recorded from 10 consecutive measurements, with the mean value obtained for analysis.

High-performance liquid chromatography (HPLC) detection of DA

Retinal tissues were isolated and stored in liquid nitrogen for measurement of retinal DA levels. Subsequently, 20 µl homogenization buffer per mg of tissue was added and samples were homogenized four times at −40°C. The homogenates were centrifuged at 38,000 × g at 4°C for 30 min and the supernatants were collected. Chromatographic conditions: Hypersil ODS2 column (5 µm; 250×4.6 mm; Thermo Fisher Scientific, Inc.); mobile phase encompassing 20 mmol/l trisodium citrate (5 mmol/l sodium 1-heptanesulfonate and 0.1 mmol/l EDTA; pH 3.7) and methanol (90:7, v/v); flow rate of 1.0 ml/min; column temperature of 35°C; electrochemical detector voltage of 750 mV; injection volume of 20 µl. Standard solutions were injected sequentially from low to high concentrations to obtain chromatograms. Sample solutions were placed in the autosampler (1260 Infinity Autosampler; Agilent Technologies, Inc.) for analysis. A standard curve was constructed using linear regression, with concentration and peak area as the x-axis and the y-axis, respectively. Results were normalized to tissue weight and expressed as ng/mg.

ELISA

Relevant kits included: Guinea Pig IL-1β ELISA Kit (cat no. ml037059; Shanghai Enzyme-Linked Biotechnology Co., Ltd.), Guinea Pig IL-6 ELISA Kit (cat no. ml037056; Shanghai Enzyme-Linked Biotechnology Co., Ltd.), Guinea Pig MMP-2 ELISA Kit (cat no. AE90268Gu; Shanghai Lianshuo Biotechnology Co., Ltd.) and Guinea Pig TNF-α ELISA Kit (cat no. ml103499; Shanghai Enzyme-Linked Biotechnology Co., Ltd.). Microplate strips were equilibrated to room temperature (RT) for 10 min before use. Subsequently, 50 µl of scleral tissue homogenate supernatant or standard was added to each well and 50 µl diluent was added to blank wells. After the addition of 50 µl antibody working solution, the plate was sealed and incubated at 37°C for 1 h.

When the incubation was completed, the liquid was discarded. Each well was washed in 300 µl wash buffer for 1 min and the procedure was repeated three times. Next, 100 µl enzyme conjugate working solution was added to each well for a 30 min incubation at 37°C. The washing procedure was then repeated five times. Subsequently, with 90 µl TMB substrate added, the plate was incubated for 15 min at 37°C in the dark. Finally, each well was supplemented with 50 µl stop solution and a microplate reader (cat. no. VA000010C; Thermo Fisher Scientific, Inc.) was employed for the measurement of the optical density at 450 nm.

H&E staining

Eyes were enucleated and fixed in 40 g/l paraformaldehyde (4% PFA) at RT overnight. After fixation, specimens were dehydrated through graded ethanol, cleared in xylene, and embedded in paraffin. Paraffin-embedded sections were cut at a thickness of 5 µm and placed in a 60–65°C oven for 15–30 min. Slides were then immersed in xylene I and II for 5–10 min each to remove paraffin. Rehydration was performed sequentially in 100, 95, 85 and 70% ethanol, followed by distilled water. All subsequent staining procedures were performed at RT. Sections were stained with hematoxylin for 3–10 min and rinsed under running tap water. Differentiation was performed using 1% hydrochloric acid-ethanol, followed by rinsing until sections turned bright blue.

The slides were then immersed in eosin solution for 2 min. Subsequently, the slides were sequentially transferred to 70, 85, 95% ethanol I, 95% ethanol II, 100% ethanol I and 100% ethanol II for dehydration, for 1–3 min each. The slides were then immersed in xylene I for 3–5 min and in xylene II for an additional 3–5 min. After removal from xylene, excess reagent was carefully blotted off with absorbent paper. A drop of neutral balsam was applied to the tissue section, and a clean coverslip was gently lowered onto the mounting medium to avoid air bubbles. The stained sections were imaged with an Olympus bright-field microscope (BX63; Olympus Corporation).

Immunofluorescence (IF) staining

Sections taken from the −80°C freezer received 15–30 min of thawing treatment at RT. An immunohistochemistry oil-based pen was used to draw circles to form a hydrophobic barrier. Frozen sections with a thickness of 8 µm were immersed in PBS for 10 min to remove optimal cutting temperature medium. The sections were then immersed in preheated sodium citrate antigen retrieval solution and received 15 min of heating treatment at 92–95°C, followed by natural cooling to RT. Next, sections underwent 20 min of permeabilization with PBS containing 0.3% Triton X-100 at RT. After one PBS wash, sections were blocked for 1–2 h in blocking buffer (cat. no. P0261; Beyotime Biotechnology) at RT. After removing the blocking buffer, sections were incubated at 4°C in a humid chamber with TH Polyclonal Antibody (cat. no. 25859-1-AP; Wuhan Sanying Biotechnology) diluted 1:200 in blocking buffer. Sections were then equilibrated at RT and underwent 4 washes with PBS containing 0.1% (v/v) Tween-20 (PBST), each for 10 min. Subsequent to 1–2 h of incubation with CoraLite488-conjugated Goat Anti-Rabbit IgG (H+L) diluted 1:500 in blocking buffer (cat no. SA00013-2; Wuhan Sanying Biotechnology) in a humidified chamber covered with aluminum foil at RT in the dark, sections received four PBST washes in the dark, each for 10 min. Nuclei were counterstained with DAPI solution (cat no. C1006; Shanghai Biyuntian Biotechnology Co., Ltd.) for 5–10 min at RT in the dark and were washed three times with PBS for 5 min each to remove excess DAPI. Finally, excess liquid around the tissue was blotted away with filter paper. Sections were covered with anti-fade mounting medium and coverslips were placed gently to avoid air bubbles. Edges of the coverslip were sealed and sections were stored at 4°C in the dark. Images of stained sections were acquired using an Olympus fluorescence microscope (BX63; Olympus Corporation).

Western blot analysis

The following antibodies were used: ERK (1:1,000; cat no. AF0147; Shanghai Biyuntian Biotechnology Co., Ltd.), phosphorylated (p)-ERK (1:200; cat no. sc-81492; Santa Cruz Biotechnology, Inc.), IκBα (1:1,000; cat no. AI096; Shanghai Biyuntian Biotechnology Co., Ltd.), p-IκBα (1:200; cat no. sc-8404; Santa Cruz Biotechnology, Inc.), NF-κB (1:10,000; cat no. 80979-1-RR; Wuhan Sanying Biotechnology), p-NF-κB (1:5,000; cat no. 82335-1-RR; Wuhan Sanying Biotechnology) and GAPDH (1:10,000; cat no. 60004-1-Ig; Wuhan Sanying Biotechnology). Retinal tissue samples were lysed on ice for 30 min using RIPA lysis buffer (cat no. P0013B; Beyotime Biotechnology) containing 1 mM PMSF (cat no. BL507A; Biosharp Life Sciences) and phosphatase inhibitor cocktail (cat no. P1081; Beyotime Biotechnology), the resulting lysates underwent 15 min centrifugation (13,000 × g; 4°C), with the supernatant harvested. The BCA method was applied for the protein concentration measurement. The mixture of samples with 5X loading buffer was denatured by boiling for 5 min. After cooling, 30 µg total protein and protein markers (cat no. 26617; Thermo Fisher Scientific, Inc.) were loaded onto 10% precast SDS-PAGE gels (cat no. ET12420Gel; ACE Biotechnology Co., Ltd.). Electrophoresis was performed at a constant voltage of 80 V until the target proteins were fully separated. Protein transfer onto 0.45 µm PVDF membranes (cat no. IPVH00010; MilliporeSigma) was carried out in an ice bath at a constant current of 200 mA for ~120 min. Membranes were blocked for 1 h in 5% skim milk in TBS-Tween [TBST; containing 0.1% (v/v) Tween-20] at RT with gentle agitation before TBST rinsed, followed by overnight incubation with primary antibodies diluted in blocking buffer at 4°C with gentle agitation. Later, membranes were subjected to three times of TBST washes at RT, each for 10 min. Then, membranes were incubated for 1 h with HRP-labeled secondary antibody (1:1,000; cat no. A0208 or A0216; Shanghai Biyuntian Biotechnology Co., Ltd.) diluted in blocking buffer at RT with gentle shaking before three TBST washes, each for 10 min. Finally, ECL chemiluminescent substrates (cat no. G2074; Wuhan Servicebio Technology Co., Ltd.; solutions A and B mixed at 1:1) were applied to membranes after removing excess TBST. Membranes were incubated for 1–2 min and images were captured with a Bio-Rad ChemiDoc MP chemiluminescence imaging system. The experimental results were obtained using ImageJ (v1.53; National Institutes of Health) software to open the strip image, selecting the same area to measure the grayscale value of the strip and subtracting the adjacent background. The ratio of the grayscale value of the target protein to the grayscale value of the internal reference (GAPDH) was used as the relative expression level of the protein.

Statistical analysis

Statistical analyses were carried out using GraphPad Prism 9.0 (Dotmatics) and SPSS 25.0 (IBM Corp.). Quantitative data are presented as mean ± standard deviation. Normally distributed data with equal variance were compared with independent samples t-tests. Non-normally distributed data or data with unequal variance were analyzed using non-parametric tests such as the Mann-Whitney U test. All quantitative data were obtained from at least three independent replicate experiments, with statistical significance defined as P<0.05.

Results

SD-associated changes in refractive state in guinea pigs with LIM

Guinea pig LIM and LIM + SD models were successfully established. Measurements revealed significant myopic shifts and axial elongation in defocused eyes of both LIM and LIM + SD groups vs. control eyes (LIM group: all P<0.0001; LIM + SD group: all P<0.0001). Defocused eyes (OD) in the LIM + SD group showed greater myopic shift and axial elongation when compared with the LIM group (diopter, P<0.0001; axial length, P=0.001). However, control eyes (OS) in the LIM + SD and LIM groups did not differ significantly (P>0.05; Fig. 2A and B). H&E staining showed that control eyes (OS) had intact, compact and well-organized retinal structures, while defocused eyes (OD) exhibited retinal thinning, reduced cell density and disorganized nuclear layers, especially evident in the LIM + SD group (Fig. 2C).

Figure 2. SD-associated changes in refractive state in guinea pigs with LIM. (A) Diopter of retina in the LIM and LIM + SD group. (B) Axial length in the LIM and LIM + SD groups. (C) H&E staining of retina in the LIM and LIM + SD groups. **P<0.01, ****P<0.0001. LIM, lens-induced myopia; SD, sleep deprivation; OD, oculus dexter (defocused eye); OS, oculus sinister (control eye); GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; ONL, outer nuclear layer.

SD-associated changes in retinal DA and its rate-limiting enzymes in guinea pigs with LIM

HPLC results indicated significantly reduced levels of DA in defocused eyes (OD) compared with control eyes (OS) in both LIM and LIM + SD groups (LIM group, P=0.0009; LIM + SD group, P<0.0001). The LIM + SD group demonstrated significantly lower DA levels in defocused eyes (OD) vs. the LIM group (P=0.0005). The DA levels in control eyes (OS) did not differ greatly between the two groups (P>0.05; Fig. 3A). IF staining revealed red fluorescence indicating TH-positive cells, primarily localized in the inner plexiform and inner nuclear layers, with slight staining in the outer plexiform layer, identifying dopaminergic amacrine cells. Average fluorescence intensity of TH-positive cells significantly decreased in defocused eyes (OD) of LIM and LIM + SD groups vs. control eyes (OS) (LIM group, P<0.0001; LIM + SD group, P<0.0001). The LIM + SD group exhibited significantly lower intensity vs. the LIM group (P=0.003). Control eyes (OS) did not differ greatly between LIM + SD and LIM groups (P>0.05; Fig. 3B and C).

Figure 3. SD-associated changes in retinal DA and its rate-limiting enzymes in guinea pigs with LIM. (A) DA in the LIM and LIM + SD group. (B) Representative fluorescence images and (C) the average fluorescence intensity of TH positive cells in the LIM and LIM + SD group. **P<0.01, ***P<0.001, ****P<0.0001. SD, sleep deprivation; LIM, lens-induced myopia; DA, dopamine; OD, oculus dexter (defocused eye); OS, oculus sinister (control eye); GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; ONL, outer nuclear layer.

SD-associated changes in retinal inflammatory factor expression in guinea pigs with LIM

Analysis of ELISA assays revealed significantly elevated inflammatory factors (IL-1β, IL-6, TNF-α and MMP-2) in defocused eyes (OD) compared with control eyes (OS) in both LIM and LIM + SD groups (LIM group: IL-1β P=0.0001, IL-6 P<0.0001, TNF-α P=0.006, MMP-2 P<0.0001; LIM + SD group, all P<0.0001). The LIM + SD group presented increased inflammatory factor levels in defocused (OD) and control eyes (OS) vs. the LIM group (defocused eye, all P<0.0001; control eye: IL-1β P=0.054, IL-6, TNF-α and MMP-2 all P<0.0001) (Fig. 4A-D).

Figure 4. SD-associated changes in retinal inflammatory factor expression in guinea pigs with LIM. (A) Expression of IL-1β in the LIM and LIM + SD group. (B) Expression of IL-6 in the LIM and LIM + SD group. (C) Expression of TNF-α in the LIM and LIM + SD group. (D) Expression of MMP-2 in the LIM and LIM + SD group. **P<0.01, ***P<0.001, ****P<0.0001. SD, sleep deprivation; LIM, lens-induced myopia; OD, oculus dexter (defocused eye); OS, oculus sinister (control eye).

SD-associated changes in ERK/NF-κB signaling pathways in guinea pigs with LIM

Western blot analysis showed significantly elevated levels of p-ERK/ERK, p-IκBα/IκBα and p-NF-κB/NF-κB in defocused eyes (OD) compared with control eyes (OS) in both LIM and LIM + SD groups (LIM group: p-ERK/ERK P=0.12, p-IκBα/IκBα P=0.10, p-NF-κB/NF-κB P=0.001; LIM + SD group: p-ERK/ERK P=0.0009, p-IκBα/IκBα P=0.0017, p-NF-κB/NF-κB P=0.013). Both defocused (OD) and control eyes (OS) presented increased phosphorylation levels in the LIM + SD group vs. the LIM group (defocused eye: p-ERK/ERK P=0.0003, p-IκBα/IκBα P=0.014, p-NF-κB/NF-κB P=0.012; control eye: p-ERK/ERK P=0.101, p-IκBα/IκBα P=0.003, p-NF-κB/NF-κB P=0.14) (Fig. 5A-D).

Figure 5. SD-associated changes in ERK/NF-κB signaling pathways in guinea pigs with LIM. (A) Western blot and relative expression levels of (B) p-ERK/ERK, (C) p-IκBα/IκBα and (D) p-NF-κB/NF-κB protein in the LIM and LIM + SD groups. Samples were run on parallel gels and equal amounts of protein from the same lysates were loaded for all runs. *P<0.05, **P<0.01, ***P<0.001. SD; sleep deprivation, LIM; lens-induced myopia; p-, phosphorylated; OD, oculus dexter (defocused eye); OS, oculus sinister (control eye).

BVD-523-associated changes in refractive state of myopic eyes in LIM + SD guinea pig model

Measurements showed a reduced myopic shift and shorter axial length in defocused eyes (OD) of the LIM + SD + BVD-523 group vs. the LIM + SD group (LIM + SD group, all P<0.0001; LIM + SD + BVD-523 group, all P<0.0001). Control eyes (OS) demonstrated no significant differences (P>0.05; Fig. 6A and B). H&E staining revealed relatively intact, compact and well-organized retinal structures in control eyes (OS). After BVD-523 treatment, defocused eyes (OD) exhibited increased retinal thickness, increased cell number and improved nuclear layer organization (Fig. 6C).

Figure 6. BVD-523-associated changes in refractive state of myopic eyes in LIM + SD guinea pig model. (A) Diopter of the retina in the LIM + SD and LIM + SD + BVD-523 groups. (B) Axial length in the LIM + SD and LIM + SD + BVD-523 groups. (C) H&E staining of the retina in the LIM + SD and LIM + SD + BVD-523 groups. ****P<0.0001. LIM; lens-induced myopia; SD; sleep deprivation; OD, oculus dexter (defocused eye); OS, oculus sinister (control eye); GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; ONL, outer nuclear layer.

BVD-523-associated changes in retinal dopamine and its rate-limiting enzymes in LIM + SD guinea pig model

HPLC analysis showed significantly increased DA levels (Fig. 7A) and immunofluorescence revealed a significant increase in average fluorescence intensity of TH-positive cells in defocused eyes (OD) of the LIM + SD + BVD-523 group vs. the LIM + SD group (DA, P=0.0005; average fluorescence intensity, P=0.016). Control eyes (OS) demonstrated no significant differences (P>0.05; Fig. 7B and C).

Figure 7. BVD-523-associated changes in retinal dopamine and its rate-limiting enzymes in LIM + SD guinea pig model. (A) DA in the LIM + SD and LIM + SD + BVD-523 groups. (B) Representative fluorescence images and (C) the average fluorescence intensity of TH positive cells in the LIM + SD and LIM + SD + BVD-523 groups. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. DA; dopamine; LIM; lens-induced myopia; SD, sleep deprivation; TH, tyrosine hydroxylase; OD, oculus dexter (defocused eye); OS, oculus sinister (control eye); GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; ONL, outer nuclear layer.

BVD-523-associated changes in inflammatory factor expression in LIM + SD guinea pig model

Analysis of ELISA assay results showed significantly reduced inflammatory factor levels in defocused eyes (OD) of the LIM + SD + BVD-523 group vs. the LIM + SD group (IL-1β, P=0.0016; IL-6, TNF-α and MMP-2, all P<0.0001). Similarly, in the control eyes (OS), inflammatory factor levels were also significantly lower in the BVD-523 vs. the LIM + SD group (IL-1β, P=0.0001; IL-6, P<0.0001; TNF-α, P<0.0001; MMP-2, P=0.0001) (Fig. 8A-D).

Figure 8. BVD-523-associated changes in inflammatory factor expression in LIM + SD guinea pig model. (A) Expression of IL-1β in the LIM + SD and LIM + SD + BVD-523 groups. (B) Expression of IL-6 in the LIM + SD and LIM + SD + BVD-523 groups. (C) Expression of TNF-α in the LIM + SD and LIM + SD + BVD-523 groups. (D) Expression of MMP-2 in the LIM+SD and LIM + SD + BVD-523 groups. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. LIM; lens-induced myopia, SD; sleep deprivation; OD, oculus dexter (defocused eye); OS, oculus sinister (control eye).

Discussion

Shorter sleep duration induces increased susceptibility to myopia and accelerated progression in adolescents who are already nearsighted (16,17). In addition, irregular sleep-wake patterns are positively associated with myopia (18,19). These findings suggest that SD may exacerbate myopia in adolescents. LIM guinea pigs are widely used as experimental models because of their large ocular size, short induction period and rapid response (20,21). Tobler et al (11) found that the sleep of guinea pigs exhibits obvious multiphase and fragmented characteristics, with a total sleep time of ~32.2% of the 24 h recording time, or ~7.7 h, which is at a relatively low level among rodents; A 20 h deprivation program can cause sleep loss to reach a level equivalent to ~2.6 times normal sleep time, which is sufficient intervention intensity.

In the present study, the LIM guinea pig model was used to evaluate the impact of SD on myopia progression. The LIM + SD group exhibited greater increases in diopter and axial length than the LIM group, indicating more severe myopia progression. These findings from an animal model complement and extend previous clinical and mechanistic reports (22–25). The LIM + SD group showed markedly lower retinal DA and TH levels vs. the LIM group. Hence, SD may be associated with aggravation of myopia progression, along with suppression of DA synthesis and release. Landis et al (26) reported that myopia in mice can be prevented by enhancing endogenous DA production. Lan et al (27) also demonstrated a lower amount of retinal DA released in deprivation-induced myopia progression. Furthermore, SD increased the expression of inflammatory factors in the LIM model, suggesting that SD may be associated with a low-grade inflammatory response and subsequent metabolic changes.

A previous study showed that quercetin attenuates myopia progression by restricting the protein kinase RNA-like ER kinase-eukaryotic translation initiation factor 2A signaling pathway and intensifying scleral remodeling in guinea pigs (19). Zhang et al (28) reported that microRNA (miR)-15b-5p/miR-379-3p modulates the scleral cell cycle and apoptosis via the insulin-like growth factor 1 receptor/PTEN/FOXO pathway, facilitating scleral remodeling during myopia progression. Based on these findings, we hypothesized that SD may also be associated with both myopia progression and alterations in signaling pathways that associate with DA and TH levels.

Small interfering RNA-mediated inhibition of ERK has been shown to suppress acidosis-induced NF-κB activation by reducing phosphorylation of IκBα (29). In addition, Mn2+-induced expression of proinflammatory mediators can be attenuated by ERK and NF-κB p65 inhibitors (30). Therefore, the present study proposed that the ERK/NF-κB signaling pathway may contribute to the effects of SD on myopia in LIM guinea pigs. In the present study, SD increased the levels of p-ERK/ERK, p-IκBα/IκBα and p-NF-κB/NF-κB, indicating activation of the ERK/NF-κB pathway. Administration of the ERK1/2 inhibitor BVD-523 attenuated the increase in diopter and axial length, restored DA and TH levels and reduced inflammatory factor expression. The present study demonstrated that SD may activate retinal ERK/NF-κB signaling through two intermediate pathways, systemic oxidative stress (increased reactive oxygen species) and sympathetic overactivation (increased cyclic catecholamines stimulating retinal β-adrenergic receptors). This hypothesis is supported by previous literature. Vaccaro et al (31) showed that sleep loss causes reactive oxygen species accumulation, consistent with the oxidative stress component of the present model. Ma et al (32) demonstrated that chronic sleep deprivation modulates ERK signaling and causes pathological alterations, supporting the ERK activation arm of our proposed pathway. In addition, the present study demonstrated that NF-κB may inhibit the potential mechanism of TH expression, potentially by direct binding to the putative κB site within the TH promoter, as well as indirect inhibition by inducing transcription inhibitory factors (33–35). The precise molecular mechanism proposed in the present study is still speculative and requires future experimental verification.

However, the present study has certain limitations. Firstly, BVD-523 is a selective ERK1/2 inhibitor and its inhibitory effect on NF-κB is indirect. Therefore, the present study cannot fully distinguish the independent roles of ERK and NF-κB. Future studies using specific NF-κB inhibitors or genetic approaches are needed to validate the direct role of NF-κB. Although a negative association was observed between retinal DA and TH levels and myopia severity, functional intervention experiments such as dopamine supplementation or dopamine receptor agonist treatment were not performed. Therefore, data produced from the present study cannot distinguish whether reduced DA/TH is a driver of myopia progression or merely a byproduct of myopia or stress. Future studies using L-DOPA or apomorphine are needed to establish the causal role of dopamine signaling in SD-exacerbated myopia. Thirdly, the prolonged systemic SD protocol (20 h/day) employed in the present study is likely to induce widespread physiological stress responses, including endocrine and metabolic disturbances (21). Therefore, the possibility that the observed retinal molecular changes and myopia exacerbation are partially attributable to systemic stress rather than solely to local retinal ERK/NF-κB pathway activation cannot be excluded. Future studies incorporating measurements of circulating stress hormones or the inclusion of non-sleep-deprived stress control groups would help to dissect the specific contribution of sleep loss vs. general stress to myopia progression. In addition, the present study lacks a completely untreated normal control group (without lens induction), which limits the ability to fully distinguish between normal eye development effects and experimental intervention effects. The inclusion of blank control cohorts in future research will help further elucidate the findings of the present study. Besides, the ability of oral BVD-523 to cross the blood retinal barrier (BRB) was not directly measured in the present study. Although the observed decrease in retinal p-ERK and p-NFκB levels provides functional evidence of target involvement, future pharmacokinetic studies, such as LC-MS/MS measurements of BVD-523 concentration in retinal homogenates, require direct confirmation of BRB penetration. The present study only evaluated ocular parameters at the endpoint (14 days after modeling). Continuous measurements at intermediate time points will provide a more dynamic view of the time-dependent effects of myopia progression. Future studies should include multiple time points. Furthermore, the baseline diopter and axial length were not recorded prior to group assignment. Although randomization was performed, future studies should include pre-intervention measurements of refractive and biometric parameters to more rigorously exclude the influence of baseline differences. The present study did not include direct phospho-protein validation through western blotting to biochemically confirm ERK/NF-κB pathway inhibition and the precise molecular mechanisms of BVD-523 in this specific context remain to be fully elucidated in future investigations. Finally, the experiments were performed in two sequential phases rather than a single parallel multi-group design, which may introduce batch effects despite efforts to standardize conditions. Future studies employing a fully randomized parallel design would provide stronger control for potential inter-cohort variability.

Collectively, these findings suggest that SD may exacerbate LIM in association with the ERK/NF-κB signaling pathway activation, providing a theoretical basis for potential therapeutic strategies in myopia treatment.

Acknowledgements

Not applicable.

Funding

Funding was received from Scientific Research Program of Wuxi Health Commission (grant no. Q202464).

Availability of data and materials

The data generated in the present study may be requested from the corresponding author.

Authors' contributions

XY and LG took charge of experiment conception and design. ZZ, CZ, JL and WG carried out the experiments. XY, ZZ and RY participated in data analysis. XY, ZZ, CZ and LG were responsible for writing the draft paper. All authors read and approved the final version of the manuscript. XY and LG confirm the authenticity of all the raw data.

Ethics approval and consent to participate

The Animal Welfare and Ethics Review Committee of Hubei Beiente Experimental Center approved the present study (approval no. IACUC-BNB-2025-024).

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References