mTORC1 Signaling and Negative Lens-Induced Axial Elongation

Abstract

Purpose: The mechanism underlying axial elongation during myopia progression remains unknown. Epidermal growth factor receptor (EGFR) signaling is associated with axial elongation. We explored whether mammalian target of rapamycin complex 1 (mTORC1) signaling acts as the downstream pathway of EGFR and participates in negative lens-induced axial elongation (NLIAE).

Methods: Three-week-old male pigmented guinea pigs underwent binocular NLIAE. (1) To investigate whether EGFR is the upstream regulator of mTORC1, an EGFR inhibitor (20 µg erlotinib) was intravitreally injected once a week for three weeks. (2) To assess the effect of mTORC1 inhibition on NLIAE, an mTORC1 inhibitor (2 µg, 10 µg, and 20 µg everolimus) was intravitreally injected once a week for three weeks. (3) To explore the long-term effect of mTORC1 overactivation on axial elongation, an mTORC1 agonist (4 µg MHY1485) was intravitreally injected once a week for three months. Biometric measurements included axial length and choroidal thickness were performed.

Results: Compared with the guinea pigs without NLIAE, NLIAE was associated with activation of mTORC1 signaling, which was suppressed by intravitreal erlotinib injection. Intravitreally injected everolimus suppressed NLIAE-induced axial elongation, mTORC1 activation, choroidal thinning, and hypoxia-inducible factor-1α expression in the sclera. Immunofluorescence revealed that the retinal pigment epithelium was the primary location of mTORC1 activation during NLIAE. Combining NLIAE and MHY1485 intravitreal injections significantly promoted axial elongation, choroidal thinning, and peripapillary choroidal atrophy.

Conclusions: The mTORC1 signaling is associated with increased axial elongation, as in NLIAE, raising the possibility of inhibiting mTORC1 as a novel treatment for slowing myopia progression.

Myopia is the most common refractive vision disorder globally and has become the most common cause of reversible vision impairment.¹ In myopia, the axial length of the eye becomes longer than necessary for emmetropia, and light rays from the distance are focused in front of the retinal plane. High myopia, the advanced form of myopia, can lead to myopic maculopathy and high myopia-associated optic neuropathy.² High myopia-related complications have become the most frequent causes of irreversible vision impairment and blindness in East Asia. It was estimated that 5.2% of the global population had high myopia in 2020, which will increase to 9.8% worldwide by 2050.³ The development and progression of myopia are characterized by axial elongation. Thus exploring the mechanism of axial elongation is a priority research focus for preventing high myopia and related complications.⁴ 

Emmetropization is presumably regulated by a feedback mechanism containing afferent sensory and efferent parts.⁵ Optic nerve section does not affect the development of myopia, indicating that regulation of axial elongation is mediated primarily by local mechanisms.⁶ Understanding the molecular basis underlying emmetropization and axial elongation may provide targets for myopia control. In previous studies, blockade of the epidermal growth factor receptor (EGFR) via intravitreal application of EGFR antibodies significantly reduced axial elongation in guinea pigs that underwent myopia induction with negative lenses.⁷ A similar effect was achieved with intravitreally injected antibodies to amphiregulin, a major EGFR ligand.^7–10 These results indicate that EGFR signaling might play a role in axial elongation in experimental myopia. 

The mTORC1 acts as a direct second messenger of EGFR. EGFR can activate mTORC1 signaling by inactivating the mTORC1 inhibitor tuberous sclerosis complex.¹¹ Thus mTORC1 may mediate EGFR-mediated enhanced axial elongation, as seen with NLIAE. The retinal pigment epithelium (RPE) anatomically connects the neural retina and the choroid and may play a pivotal role in regulating axial elongation. Previous immunofluorescence studies have revealed that staining for amphiregulin (an EGF family member) is localized mainly in RPE cells.^8,9 mTORC1 is associated with basic physiological and pathophysiological processes of RPE cells, including cell growth, metabolism, and survival.^12,13 Thus we hypothesized that mTORC1 activation in RPE cells might be involved in the downstream signaling of EGFR during negative lens-induced axial elongation. 

Methods

The study design, including the treatment and care of the animals, was approved and supervised by the Ethics Committee of Capital Medical University. All research protocols and procedures followed the Association for Research in Ophthalmology statement for the use of animals in ophthalmic and vision research. The study included male pigmented three-week-old guinea pigs that were reared in cycles of 12-hour light (450–500 lux) and 12-hour dark (∼0 lux) with room temperature maintained at 25°C. All animals had free access to food and water. 

Study Protocol

We performed five experimental protocols (the details for the doses applied are described in the Supplementary Materials and Methods). 

• 1) To investigate the upstream regulation of mTORC1, guinea pigs were subjected to binocular NLIAE and intravitreally injected with an EGFR inhibitor (erlotinib, 20 µg) into the right eye once a week for three weeks (n = 10).
• 2) To assess the effect of mTORC1 inhibition on axial elongation, guinea pigs were subjected to binocular NLIAE and intravitreally injected with an mTORC1 inhibitor (everolimus, 2 µg [n = 10], 10 µg [n = 10], or 20 µg [n = 10]) into the right eye once a week for three weeks. The applied doses of everolimus were based on the results of previous investigations and took into account the vitreous volume of guinea pig eyes.¹⁴ We applied everolimus instead of rapamycin because everolimus is less lipophilic and potentially has a longer half-life in the vitreous body.¹⁵ In addition, clinical studies have suggested that everolimus has similar efficacy and might be better tolerated than rapamycin.¹⁶
• 3) To explore the effect of long-term mTORC1 overactivation, guinea pigs were subjected to binocular NLIAE and intravitreally injected with an mTORC1 agonist (MHY1485, 4 µg) into the right eye once a week for three months. Previous studies have revealed that MHY1485 can increase P70S6K phosphorylation and suppress mTORC2 activity through a negative feedback pathway.^17,18
• 4) To determine whether everolimus can induce retinal apoptosis, guinea pigs were intravitreally injected with everolimus (10 µg) into the right eye once a week for three weeks (n = 10).
• 5) Guinea pigs without NLAE were included as negative controls (n = 10). Guinea pigs that underwent only binocular NLIAE were included as positive controls (n = 10).

Drug Preparation and Delivery

Everolimus and erlotinib, and MHY1485 (GlpBio Technology, Montclear, CA, USA) are insoluble in water. To keep a constant drug concentration during injection, suspensions were prepared by adding 5% polysorbate 80 (Sigma-Aldrich Co., St. Louis, MO, USA) to phosphate-buffered saline solution (PBS; Sigma-Aldrich Co.). All suspensions were immediately aliquoted and stored at −20°C. Intravitreal injections were performed under topical anesthesia using 0.5% proxymetacaine hydrochloride eye drops. The suspensions were thawed and shaken well before use. A 26-gauge needle (0.26 mm inner diameter) was used as a trocar needle to first make a vitreous entry port 2 mm posterior to the limbus. A 32-gauge Hamilton microsyringe (outer diameter: 0.235 mm, Hamilton Microliter syringe; Sigma-Aldrich Co.) delivered 5 µL of the solution into the eyes through the 26-gauge trocar needle. 

All guinea pigs that received intravitreal injections of drugs into their right eyes received intravitreal injections of 5 µL of PBS with 5% polysorbate 80 into their left eyes. The guinea pigs in the negative control group and the NLIAE-only group received 5 µL of PBS with 5% polysorbate 80 once a week in both eyes. 

Negative Lens-Induced Axial Elongation

To induce axial elongation, goggles were fitted with −10.0 diopter lenses (polymethyl methacrylate; diameter: 12.7 mm; Supplementary Fig. S1a) and taped onto the orbital rims of both eyes of the guinea pigs. Care was taken to ensure that the guinea pigs could open their eyes and blink freely while wearing the goggles. The refractive power of the lens and its centration were measured and verified before application. The goggles were examined daily to ensure that the lenses were clean and in place; otherwise, the goggles were detached and replaced with new goggles. The goggles were removed weekly for biometric examinations of the eyes. 

Choroid Thickness Assessment, Axial Length Measurement, and Fundus Photography

All guinea pigs underwent optical coherence tomography (OCT) imaging of the ocular fundus (SS-OCT, VG200D, SVision Imaging, Ltd., Guangdong, China) without anesthesia. The OCT scans were performed using a star scan pattern centered on the optic disc center. The horizontal and vertical scan images were exported. For each guinea pig, the choroidal thickness was measured in the horizontal (3 o'clock and 9 o'clock positions) and vertical meridians (12 o'clock and 6 o'clock positions) at distances of one and three horizontal disc diameters from the optic disc center. The mean choroidal thickness was calculated as the average of the four measurement points at each distance. Under topical anesthesia, we measured the axial length by ocular ultrasonography (A-scan mode scan; oscillator frequency: 11 MHz; Quantel Co., Les Ulis, France). The ultrasound velocities used were 1557.5 m/s for the cornea and aqueous humor, 1723.3 m/s for the lens, and 1540 m/s for the vitreous cavity.^19,20 For each guinea pig, five measurements were performed, and the mean values were recorded. Fundus photography was performed using a wide-field fundus camera (ZEISS CLARUS 500; Carl Zeiss Meditec AG, Jena, Germany). 

Tissue Collection

Guinea pigs were anesthetized by intraperitoneal injection of urethane (1000 mg/kg). After the animals were killed, the eyes were enucleated. For Western blot examination, the cornea, lens, and vitreous were first removed, and then the retina-choroid tissue and the sclera were harvested separately under a microscope. All samples were immediately stored in liquid nitrogen and transferred to a −80°C freezer. The frozen tissues were analyzed within one week. For histopathological examination, the enucleated eyeballs were immediately fixed in 10 mL of FAS Eyeball Fixative Solution (Wuhan Servicebio Technology Co. Ltd. Wuhan, China) for 24 hours and then embedded in paraffin. 

Terminal Deoxynucleotidyl Transferase dUTP Nick-End Labeling (TUNEL) Staining

The details of the TUNEL staining have been described previously.⁷ After fixation and embedding in paraffin, histological slides (thickness: 8 µm) of three eyes were prepared following a routine protocol. We performed TUNEL staining (Cell Death Detection kit; Kaiji Biotechnology Co. Ltd., Jiangsu, China) to detect apoptotic cells in the retina. The histological sections were deparaffinized, and 200 µL of Proteinase K (10 µg/mL) were added to completely cover each section before incubation for 10 minutes at room temperature followed by rinsing in PBS three times. The sections were incubated with terminal deoxyribonucleotide transferase enzyme mixture solution (45 µL of equilibration buffer, 1.0 µL of biotin-11-dUTP, 4.0 µL of TdT enzyme) at 37°C for one hour and washed with PBS solution three times. Afterward, the sections were labeled with streptavidin-fluorescein (50 µL) for 30 minutes and counterstained with 4,6-diamidino-2-phenylindole (300 nM, 50–100 µL) for five minutes. Three sections from each eye were photographed, and the TUNEL-positive cells in the retina were counted. The mean of the counts obtained from three images representing the eyes of three different animals was recorded. 

Immunofluorescence

After deparaffinization and antigen retrieval of histological slides, the sections were treated with PBS containing 0.25% Triton X-100 for 10 minutes and washed with PBS solution three times. Nonspecific bindings were blocked by incubation in 50 to 100 µL goat serum for 20 minutes at room temperature before incubation overnight with anti-phosphorated P70S6 kinase (Phospho T389, ab2571; Abcam, Cambridge, MA, USA; at 1:200 dilution) as the primary antibody. The sections were then incubated with a FITC-conjugated goat anti-rabbit secondary antibody (1:1000 dilution, 50-100 µL; AiFang Biology, China) for one hour in a dark chamber at 37°C. Finally, the sections were repeatedly immersed in PBS (three minutes) three times. The cell nuclei were counterstained with 4,6-diamidino-2-phenylindole for five minutes in dim light at room temperature. The sections were examined using a biological fluorescence inverted microscope (Olympus-CKX53; Olympus Co., Tokyo, Japan). 

Western Blot Analysis

Frozen retina-choroid and scleral tissues were homogenized and lysed in cold lysis buffer (RIPA; Amresco, Solon City, OH, USA) supplemented with protease inhibitors (Roche 11697498001; Roche, Basel, Switzerland) and phosphatase inhibitors (Roche 04906837001; Roche). The tissue extracts were separated on 8% SDS‒PAGE gels and transferred to nitrocellulose membranes according to a standard protocol. The membranes were blocked with 5% skimmed milk in TBST (Tris-HCl, NaCl, and Tween 20) for two hours and sequentially incubated with primary antibodies overnight and with secondary antibodies for two hours on the following day. Signals were assessed with an enhanced chemiluminescence kit (Millipore, Burlington, MA, USA), and images were taken with the Total Lab Quant V11.5 (TotalLab Ltd., Gosforth, UK). The target bands were quantified and analyzed using ImageJ (NIH, Bethesda, MD, USA) with β-tubulin as an internal control. The antibody details are listed in the Table. 

Table.

 

Western Blot Antibodies

 

Statistical Analysis

Using a website-based tool (http://powerandsamplesize.com/), we calculated the sample size before the start of the study. The standard deviation of axial length measurements was <0.05 mm. We assumed an axial length difference of 0.2 mm between eyes with NLIAE and eyes of the negative control group at the end of the study period of three weeks.⁷ Fewer than ten guinea pigs were needed to detect a 0.08 mm intergroup difference with 90% power. Thus each group contained 10 guinea pigs. The other statistical analyses were conducted using the Stata 17.0 software program (StataCorp, College Station, TX, USA) and GraphPad Prism 9.3.1 (GraphPad Software, San Diego, CA, USA). Unless stated otherwise, continuous variables are presented as the mean ± standard error. Comparisons of two samples were performed using the two-tailed Student's t-test. Paired t-tests were used to compare data from the two eyes of individual animals, whereas unpaired t-tests were used for data from different groups. Comparisons of multiple measurements from the same set of animals at different time points were performed by applying repeated-measures ANOVA and then using a Tukey honestly significant difference post hoc analysis to identify which differences between pairs of means were significant. For Western blot analysis, protein abundance is expressed as the ratio of each protein to β-tubulin, and the first control values (relative intensities of each protein/β-tubulin signal ratios) are set to 1. P values <0.05 were considered to indicate statistical significance. 

Results

NLIAE and Activation of mTORC1 in the Retina-Choroid Tissue of Guinea Pigs

Three weeks of negative lens induction resulted in an increase in axial length by 0.23 ± 0.04 mm (P < 0.001; Supplementary Fig. S1a). The NLIAE did not affect the body weight of the guinea pigs (P > 0.20 for all time points; Supplementary Fig. S1b). Compared to the negative control, NLIAE was associated with activation of the downstream signaling pathway of EGFR, including ERK1/2 (0.775 ± 0.128 in negative control vs. 1.256 ± 0.197 in NLIAE, relative abundance to β-tubulin, P = 0.045) and PI3K (2.707 ± 1.046 in negative control vs. 4.833 ± 1.010 in NLIAE, relative abundance to β-tubulin, P = 0.006) in the retina-choroid tissue of the guinea pigs (Figs. 1a, 1b). The activated ERK1/2 and PI3K pathways were associated with increased phosphorylation of ribosomal protein S6 kinase beta-1 kinase (p70S6K), a key substrate of mTORC1 (1.737 ± 0.234 in negative control vs. 2.848 ± 0.172 in NLIAE, P < 0.001, relative abundance to β-tubulin, Fig. 1c).¹¹ Compared to the contralateral eye, the unilateral intravitreal injection of the EGFR inhibitor erlotinib (dose: 20 µg) significantly reduced axial elongation by 0.09 ± 0.02 mm (P < 0.001; Fig. 1d). The injection of erlotinib was associated with a reduction in the phosphorylation of p70S6K in the eyes with NLIAE compared with the contralateral eyes (0.882 ± 0.073 in NLIAE vs. 0.569 ± 0.072 in NLIAE + erlotinib, relative abundance to β-tubulin, P = 0.004; Fig. 1e). NLIAE was not significantly associated with AKT (protein kinase B) phosphorylation (0.606 ± 0.205 in negative control vs. 0.661 ± 0.153 in NLIAE, relative abundance to β-tubulin, P = 0.43; Supplementary Fig. S1c). Because mTORC2 can activate AKT, the unchanged AKT phosphorylation levels suggested that only mTORC1, not mTORC2, was involved in NLIAE.¹¹ 

Figure 1.

 

NLIAE and activation of the mTORC1 pathway in the retina-choroid complex in guinea pigs. Guinea pigs underwent NLIAE for three weeks. The retina-choroid tissue was harvested at the end of the study. Binocular NLIAE activated ERK1/2 (a), PI3K (b), and mTORC1 (here shown as an increased phosphorylation of p70S6 kinase) compared to their levels in normal control eyes (c). Intravitreal injection of the EGFR inhibitor erlotinib into right eyes with NLIAE significantly attenuated NLIAE (d) and inhibited the activation of mTORC1 (e). Relative protein levels were defined by the ratio of each protein to β-tubulin.

Inhibition of mTORC1 With Everolimus and Suppression of NLIAE

To explore the involvement of mTORC1 signaling during NLIAE, we assessed the effect of an intravitreally applied mTORC1 inhibitor (everolimus) on axial elongation. Compared to the contralateral eyes undergoing NLIAE alone, weekly monocular injections of everolimus in the eyes with NLIAE was associated with a significant and dose-dependent reduction in mTORC1 activation and a reduction in axial elongation (1.737 ± 0.234 in negative control, 2.848 ± 0.172 in NLIAE, 1.946 ± 0.157 in NLIAE + 2 µg everolimus, 2.061 ± 0.215 in NLIAE + 10 µg everolimus, 1.630 ± 0.230 in NLIAE + 20 µg everolimus, relative abundance to β-tubulin; Figs. 2a, 2b, Supplementary Fig. S2a). After three weekly unilateral injections of everolimus in doses of 2 µg, 10 µg, and 20 µg, axial elongation was decreased by 0.05 ± 0.01 mm (P = 0.04), 0.09 ± 0.01 mm (P < 0.001), and 0.14 ± 0.01 mm (P < 0.001), respectively, when compared to the contralateral eyes undergoing NLIAE alone. 

Figure 2.

 

Intravitreally applied everolimus and NLIAE. (a) Retina-choroid tissue was harvested at the end of the study. Compared to normal control eyes (first column), eyes with binocular NLIAE (second column) showed activated mTORC1 (here shown as increased phosphorylation of p70S6 kinase). Intravitreally applied everolimus was associated with a dose-dependent reduction in mTORC1 (third to fifth column). (b) Guinea pigs with binocular NLIAE and unilateral intravitreal everolimus (Eve) applications showed decreased unilateral axial elongation (i.e., a negative interocular axial length difference) in an everolimus dose-dependent manner. Intravitreally applied everolimus in eyes without NLIAE was not associated with a change in the interocular axial length difference. (c) TUNEL staining of the retina and choroid did not show a difference between control eyes and eyes with intravitreally applied everolimus in terms of the presence of apoptotic cells. Relative protein levels were defined by the ratio of each protein to β-tubulin. INL, inner nuclear layer; ONL, outer nuclear layer; Ch, choroid.

The changes in axial length were due mainly to changes in the depth of the vitreous compartment, with only minor changes in the anterior chamber depth and lens thickness (Fig. 3). After three weekly unilateral injections of everolimus in doses of 2 µg, 10 µg, and 20 µg, the increase in vitreous depth was reduced by 0.05 ± 0.01 mm (P = 0.18), 0.08 ± 0.02 mm (P < 0.001), and 0.10 ± 0.02 mm (P < 0.001), respectively, when compared to the contralateral eyes undergoing NLIAE alone. In contrast, the injections were not significantly associated with a significant difference in either anterior chamber depth (everolimus 2 µg: 0.02 ± 0.01 mm; P = 0.23; everolimus 10 µg −0.00 ± 0.01 mm P = 0.99; and everolimus 20 µg: 0.00 ± 0.01 mm; P = 0.86) or in lens thickness (everolimus 2 µg: −0.01 ± 0.02 mm, P = 0.93; everolimus 10 µg: −0.01 ± 0.01 mm, P = 0.92; and everolimus 20 µg: 0.03 ± 0.01 mm, P = 0.44), when compared to contralateral eyes undergoing NLIAE alone. Well-defined choroidal vessels at the posterior pole were observed in guinea pigs with NLIAE. In contrast, such well-defined choroidal vessels were almost undetectable in guinea pigs with both NLIAE and intravitreal mTORC1 inhibition (Supplementary Figs. S2b–d). 

Figure 3.

 

Ocular component changes in guinea pigs that underwent NLIAE and mTORC1 inhibition. Guinea pigs underwent binocular NLIAE and weekly monocular injections of everolimus (2 µg, 10 µg, and 20 µg). The changes in anterior chamber depth (a, b), lens thickness (c, d), and vitreous chamber thickness (e, f) were measured.

The contralateral eyes intravitreally injected with vehicle solution and the eyes intravitreally injected with 10 µg everolimus did not differ significantly in axial length or in the fundus appearance (Fig. 2b, Supplementary Figs. S2a, S3). TUNEL staining also did not show significant differences in the count of TUNEL-positive cells in the inner nuclear and outer nuclear layer of the retina (0.33 ± 0.33 cells vs. 0.33 ± 0.33 cells; P > 0.99) between the eyes receiving intravitreal everolimus injections and the contralateral eyes receiving vehicle injections (Fig. 2c). 

The RPE and the Primary Site of mTORC1 Activation

To explore the location of mTORC1 activation during NLIAE, we examined the pattern of immunofluorescence staining of eyes undergoing NLIAE. The eyes undergoing NLIAE exhibited phosphorylated P70S6K primarily in the RPE layer (Fig. 4). Eyes receiving NLIAE and everolimus injections (10 µg) showed less intense immunofluorescence for phosphorylated P70S6K than eyes undergoing NLIAE alone. 

Figure 4.

 

The mTORC1 was activated in the RPE cell layer during NLIAE. Immunofluorescence staining of phosphorylated P70S6K in (a) normal controls, (b) NLIAE, and (c) NLIAE plus intravitreal everolimus (10 µg) injection. Red, p-P70S6K; Blue, 4,6-diamidino-2-phenylindole; ONL, outer nuclear layer; Ch, choroid.

mTORC1 Activation During NLIAE, Scleral HIF-1α, and Nuclear Factor-κB (NF-κB)

Sclera hypoxia and local inflammation have been found to be associated with the development and progression of myopia.^21,22 To better understand the mechanisms of mTORC1 signaling, we investigated the association of the mTORC1 signaling pathway with scleral hypoxia and local inflammation. As evaluated by in vivo OCT imaging (Figs. 5a, 5b), the intravitreal application of 10 µg everolimus partially attenuated NLIAE-induced choroidal thinning by 2.95 ± 0.94 µm (P = 0.002) and 9.20 ± 2.08 µm (P = 0.003), measured at a distance of one and three horizontal disc diameters from the optic disc center (Figs. 5c, 5d, Supplementary Figs. S4a, S4b). Compared to NLIAE alone, intravitreally injected MHY1485 combined with NLIAE promoted NLIAE-induced choroidal thinning by 3.28 ± 0.94 µm (P = 0.040) and 9.42 ± 2.10 µm (P = 0.008), measured at a distance of one and three horizontal disc diameters from the optic disc center (Figs. 5c 5d, Supplementary Fig. S4a, S4b). As the choroid contributes to the oxygen supply of the sclera, choroidal thinning may cause scleral hypoxia. Therefore, we examined the expression of hypoxia-inducible factor-1α (HIF-1α) in the scleral tissue. Compared to the negative control group, the NLIAE group showed an upregulation of HIF-1α expression in the sclera (1.603 ± 0.395 in negative control group vs. 3.288 ± 0.295 in the NLIAE group, relative abundance to β-tubulin, P = 0.022) (Figs. 5e, 5f). Intravitreal mTORC1 inhibition significantly downregulated scleral HIF-1α in a dose-dependent manner (3.288 ± 0.295 in the NLIAE group vs. 2.086 ± 0.198 in the NLIAE + 10 µg everolimus group, P = 0.021; 3.288 ± 0.295 in the NLIAE group vs. 0.546 ± 0.346 in the NLIAE + 20 µg everolimus group, relative abundance to β-tubulin, P = 0.038) (Figs. 5e, 5f). In addition, after three weeks of NLIAE treatment, NF-κB was significantly upregulated, but this effect was partially reversed by mTORC1 inhibition (1.246 ± 0.277 in the negative control group vs. 2.166 ± 0.192 in the NLIAE group, P = 0.045; 2.166 ± 0.192 in the NLIAE group vs. 1.552 ± 0.283 in the NLIAE + 10 µg everolimus group, P = 0.024; 2.166 ± 0.192 in the NLIAE group vs. 1.308 ± 0.110 in the NLIAE + 20 µg everolimus group, relative abundance to β-tubulin, P = 0.014) (Supplementary Fig. S5a). Monocyte chemoattractant protein–1 expression in the retina-choroid tissue was independent of mTORC1 activity and intravitreal application of everolimus (1.303 ± 0.234 in the negative control group, 1.316 ± 0.207 in the NLIAE group, 1.025 ± 0.228 in the NLIAE + 2 µg everolimus group, 1.194 ± 0.149 in the NLIAE + 10 µg everolimus group, 1.047 ± 0.224 in the NLIAE + 20 µg everolimus group, relative abundance to β-tubulin, P > 0.20 for all comparisons) (Supplementary Fig. S5b). 

Figure 5.

 

The mTORC1 activation, choroidal thickness and HIF-1α. OCT image of normal control (a) and a guinea pig after three weeks of NLIAE (b). (c, d) Interocular choroidal thickness difference (right − left) measured at 1 disc diameter (c) and 3 disc diameters (d) distant to the optic disc. Everolimus-attenuated choroidal thinning, thus generating a positive interocular difference. (e) Increased expression of HIF-1α in scleral tissue in eyes with NLIAE compared to eyes with NLIAE and additional intravitreal application of everolimus (Eve). (f) Differences in scleral HIF-1α expression levels. Relative protein levels were defined by the ratio of each protein to β-tubulin. Black arrow head, inner and outer choroidal boundaries.

Long-Term mTORC1 Overactivation and Fundus Changes in Guinea Pigs With NLIAE

Compared to NLIAE alone, weekly intravitreal injections of MHY1485 and NLIAE significantly increased axial elongation. After three months of NLIAE and MHY1485 injections, the guinea pigs showed an additional axial elongation of 0.43 ± 0.03 mm (P < 0.001; Supplementary Fig. S6). The fundus images and OCT scan of negative controls and eye received NLIAE only were exhibited in Figures 6a through 6d. In parallel to the increased axial elongation, fundus photography showed large well-defined choroidal vessels at the posterior pole in all guinea pigs. After three months of NLIAE and MHY1485 injections, diffuse choroidal atrophy was found in the peripapillary region of 2 out of 8 (25%) guinea pigs (Fig. 6e). The diffuse choroidal atrophy areas corresponded to marked choroidal thinning on the OCT scans (Fig. 6f). 

Figure 6.

 

The mTORC1 overactivation and fundus changes. Fundus photographs and optical coherence tomographic images of the eyes of guinea pigs that underwent 3 months of binocular NLIAE. The right eyes received weekly intravitreal injections of 5 µg of MHY1485. At week 12, images were taken of eyes of the negative control group (a, b), eyes undergoing NLIAE (c, d), and eyes undergoing NLIAE combined with MHY1486 injections (e, f). Red arrow, peripapillary diffuse choroidal atrophy; black arrow head, inner and outer choroidal boundaries.

Discussion

In this study, NLIAE was found to be associated with an activation of mTORC1, which could be suppressed by an intravitreally applied mTORC1 inhibitor (everolimus). The mTORC1 inhibition attenuated choroidal thinning during NLIAE. Immunofluorescence revealed that the RPE cell layer was the primary location of mTORC1 activation after NLIAE treatment. OCT examination showed that mTORC1 inhibition can partially attenuated NLIAE-induced choroidal thinning and upregulation of sclera HIF-1α expression. Compared with NLIAE alone, combining NLIAE and intravitreal injections of an mTORC1 agonist (MHY1485) further enhanced axial elongation, choroidal thinning, and parapapillary choroidal atrophy. 

The results of our study agree with observations made in previous investigations suggesting that the EGFR signaling pathway may play a role in the process of axial elongation. Previous studies have linked EGFR activation with axial elongation of eyes with NLIAE in a guinea pig model of myopia.^7–9 Correspondingly, a single-gene polymorphism of amphiregulin (rs12511037), an EGFR ligand, is significantly associated with myopic refractive error, and this association interacts with high educational attainment in Asians.²³ In agreement with previous studies showing that the RPE has receptors for EGF and EGF family members such as amphiregulin,²⁴ we found that RPE cells were the primary site of mTORC1 activation. These findings indicate that the RPE and mTORC1 signaling may play a role in axial elongation. 

The upstream regulation of mTORC1 in axial elongation may involve several signaling pathways. In this study, we applied a dose of 20 µg (0.051 µmol) of erlotinib in a vitreous volume of guinea pig eyes of approximately 90 µL, resulting in an erlotinib concentration of 5.65*10² µM.¹⁴ Considering that the IC50 (half maximal inhibitory concentration) of a cell-based assay for EGFR inhibition is 0.42 µM, 20 µg of erlotinib results in an intraocular concentration 1000-fold higher than the IC50 in cell-based assays for EGFR inhibition.²⁵ However, the magnitude of attenuation of axial elongation was considerably lower than that achieved with the mTORC1 inhibitor (NLIAE + 20 µg everolimus vs. NLIAE: −0.14 ± 0.01 mm, NLIAE + 20 µg erlotinib vs. NLIAE −0.09 ± 0.02 mm for erlotinib). One of the reasons for the difference could be the lower molecular weight of erlotinib (393.4 g/mol) compared to everolimus (958.2 g/mol), which might result in faster clearance of erlotinib from the vitreous.¹⁵ Other mechanisms may also have contributed to mTORC1 activation. Local inflammation has been reported to be associated with myopia.^26,27 In guinea pigs eyes undergoing form deprivation-induced myopia, the retinal TNFα- and IL-17–related proinflammatory pathways are activated.²⁶ After treating hamster eyes that underwent form deprivation with the anti-inflammatory agent diacerein, retinal proinflammatory cytokine levels and axial elongation decreased.²⁸ In contrast, administrating proinflammatory cytokines, such as TNF-α and IL-6, to the conjunctival sacs of hamsters induced myopia ranging from −2.1D to −3.1D after 21 days.²² In our study, we also found that in the retina-choroid tissue of guinea pigs, the expression of NF-κB was significantly upregulated after NLIAE treatment. Proinflammatory cytokines can activate mTORC1 through inhibition of the tuberous sclerosis complex.¹¹ Because of its immunosuppressive effect, everolimus may suppress local proinflammatory cytokines in the retina and choroid. This effect may contribute to the inhibitory effect of everolimus on axial elongation.²⁹ 

The downstream mechanism of mTORC1 activation in axial elongation involved upregulation of scleral HIF-1α. In a previous study, myopization-associated hypoxia promoted myofibroblast transdifferentiation.²¹ A cell-based assay demonstrated that hypoxia downregulated collagen-1 expression in human scleral fibroblasts, suggesting that scleral hypoxia may reduce scleral collagen synthesis, thereby promoting axial elongation.²¹ Furthermore, the anti-hypoxia drugs salidroside and formononetin ameliorated form deprivation-induced axial elongation in mice, and scleral HIF-1α knock-down led to hyperopia in mice.^21,30 The choroid supplies nutrients and oxygen to the scleral stroma. The axial elongation–associated thinning of the choroid may lead to an insufficient oxygen supply to the sclera. In agreement with previous studies, we found that scleral HIF-1α expression was upregulated during NLIAE. Furthermore, we found that the upregulated scleral HIF-1α expression could be suppressed by intravitreal mTORC1 inhibition. These results suggests that the downstream mechanism of mTORC1 activation in axial elongation may involve sclera hypoxia. When mTORC1 agonists were administered in combination with NLIAE, guinea pigs showed enhanced axial elongation, choroidal thinning, and peripapillary choroidal atrophy compared to those seen with NLIAE alone. 

The limitations of our study should be considered. First, although mTORC1 signaling was associated with axial elongation in our experiments, the downstream pathway of mTORC1 activation in RPE cells has remained unclear. Considering that the complex regulatory network underlying myopia development involves the retina, choroid, and sclera, further studies are needed to explore the details of the downstream mechanism of mTORC1. Second, we did not assess either refractive error or corneal curvature in the current study. Although enhanced axial elongation is the most important risk factor for myopia-related complications, such measurements need to be included in future studies to comprehensively evaluate the effects of mTORC1 inhibition and overactivation on NLIAE.³¹ Third, we used −10D lenses for exploring the effect of long-term mTORC1 overactivation on axial elongation. However, axial elongation during three months of NLIAE would likely compensate for the imposed −10D defocus, thereby potentially confounding the interpretation of our results. For future studies, one may consider using defocusing lenses with incremental refractive power. 

In conclusion, NLIAE was associated with activation of mTORC1. Suppression of mTORC1 activation by intravitreal injection of the mTORC1 inhibitor, everolimus, was dose-dependently associated with a reduction in axial elongation and choroidal thinning during NLIAE. In contrast, compared to NLIAE alone, intravitreal injection of the mTORC1 activator MHY1485 was associated with greater axial elongation, advanced choroidal thinning, and parapapillary choroidal atrophy, compared to NLIAE alone. The findings suggest that mTORC1 may play a role in axial elongation, and that its blockade by everolimus could be helpful in reducing axial elongation and slowing myopia progression. 

Acknowledgments

Supported by the National Natural Science Foundation of China (82220108017, 82141128), the Capital Health Research and Development of Special (2020-1-2052) and the Science & Technology Project of Beijing Municipal Science & Technology Commission (Z201100005520045, Z181100001818003). 

Disclosure: R. Zhang, None; L. Dong, None; H. Wu, None; X. Shi, None; W. Zhou, None; H. Li, None; Y. Li, None; C. Yu, None; Y. Li, None; Y. Nie, None; L. Shao, None; C. Zhang, None; Y. Liu, None; J.B. Jonas, EP 3 271 392 (P), JP 2021-119187 (P), US 2021 0340237 A1 (P); W. Wei, None; Q. Yang, None