Opposing Effects of PPARα Agonism and Antagonism on Refractive Development and Form Deprivation Myopia in Guinea Pigs

Abstract

Purpose: To determine if drug-induced peroxisome proliferator-activated receptor α (PPARα) signal pathway modulation affects refractive development and myopia in guinea pigs.

Methods: Pigmented guinea pigs were randomly divided into normal vision (unoccluded) and form deprivation myopia (FDM) groups. Each group received daily peribulbar injections of either a vehicle or (1) PPARα agonist, GW7647, clofibrate, or bezafibrate or (2) PPARα antagonist, GW6471, for 4 weeks. Baseline and posttreatment refraction and ocular biometric parameters were measured. Immunofluorescent staining of PPARα and two of its downstream readouts, cytosolic malic enzyme 1 (ME1) and apolipoproteinA II (apoA-II), was undertaken in selected scleral sections. Western blot analysis determined collagen type I expression levels.

Results: GW6471 induced a myopic shift in unoccluded eyes, but had no effect on form-deprived eyes. Conversely, GW7647 inhibited FDM progression without altering unoccluded eyes. Bezafibrate and clofibrate had effects on refraction similar to those of GW7647 in unoccluded and form-deprived eyes. GW6471 downregulated collagen type I expression in unoccluded eyes whereas bezafibrate inhibited collagen type I decreases in form-deprived eyes. GW6471 also reduced the density of ME1- and apoA-II–stained cells in unoccluded eyes whereas bezafibrate increased apoA-II–positive cell numbers in form-deprived eyes.

Conclusions: As GW7647 and GW6471 had opposing effects on myopia development, PPARα signaling modulation may be involved in this condition in guinea pigs. Fibrates are potential candidates for treating myopia since they reduced both FDM and the associated axial elongation. Bezafibrate also inhibited form deprivation–induced decreases in scleral collagen type I expression and the density of apoA-II expressing cells.

Myopia is the most common ocular problem worldwide whose prevalence is very markedly increasing with variable significance among different racial and ethnic groups. Between 15% and 40% of the population is afflicted in the Western hemisphere^1–3 whereas in East Asian countries such as mainland China, its prevalence increases to 78% in 15-year-olds,⁴ 62% in 12-year-olds in Hong Kong, China,⁵ 82% in a young male adult population in Singapore,⁶ as well as 66% and 97% among young adults in Japan and South Korea,^7,8 respectively. Individuals with high myopia (≥ −6.00 diopters [D]) are prone to more severe sight-compromising complications such as macular degeneration, glaucoma, and retinal detachment, which can result in irreversible loss of vision.^9,10 Even though myopia is a non–life-threatening disorder, its increasing prevalence imposes a burgeoning burden on the public health care system and the economy. Therefore, there is a pressing need to identify novel targets for improving therapeutic management of this condition. 

Myopia results from excessive ocular elongation mainly along the central visual axis, which causes visual blurring of distant images due to a mismatch with the ocular refractive power, with the focal plane of the eye located in front of the retinal photoreceptors.^11–13 Excessive elongation is accompanied by active scleral extracellular matrix (ECM) remodeling resulting in a thinner, weaker, and more distensible sclera particularly at the posterior pole in mammals^14–16 and in the fibrous scleral layer of chicks.¹⁷ This remodeling process includes declines in scleral proteoglycans and collagen content along with increased collagen degradation.^{15,16,18–20} Multiple modulators have been implicated to induce scleral ECM remodeling in animal studies including cAMP,²¹ cGMP,²² and TGF-β.^23,24 

Peroxisome proliferator-activated receptor α (PPARα) is a potential target to inhibit myopia progression since Bertrand et al.²⁵ first showed that intraocular injection of a PPARα agonist, GW7647, resulted in a significant reduction of lens-induced myopia in chicks. This PPAR isotype is a ligand-activated transcription factor belonging to the nuclear hormone receptor superfamily that binds to DNA as a heterodimer with its obligatory coreceptor retinoid X receptors (RXRs). This complex recognizes specific DNA sequences that regulate transcription of its target genes. PPARα regulates the expression of multiple genes including those involved in regulating lipid metabolism, glucose homeostasis, inflammation, energy homeostasis, and cellular differentiation.^26–28 Some of these pathways are associated with myopia development since it was shown that excessive consumption of saturated fats and cholesterol is associated with longer axial length,²⁹ while hyperinsulinemia,³⁰ hyperglycemia,³¹ and inflammation³² are associated with myopic refractive errors in humans. Meanwhile, numerous other studies demonstrated that PPARα activation attenuated experimentally induced lung,³³ liver,³⁴ kidney,³⁵ and cardiac³⁶ fibrosis by modifying ECM protein deposition especially through declines in collagen synthesis and increases in its degradation. Such effects account for increases in collagen turnover along with ECM remodeling. These considerations prompted us to hypothesize that PPARα modulation could affect ocular growth through effects on scleral collagen turnover and ECM remodeling in normal and form-deprived eyes. 

Increases in PPARα transcriptional activity enhance lipid metabolism by stimulating the expression of downstream genes such as apolipoproteinA II (apoA-II)^37,38 and cytosolic malic enzyme 1 (ME1),^39,40 which serve as readouts of variations in PPARα activity. Increases in their expression levels are also indicative of rises in mitochondrial fatty acid oxidation, fatty acid uptake, and triglyceride catabolism. However, it is unknown if myopia progression is affected by changes in ME1 and apoA-II expression levels. Fibrates are PPARα agonists that have been used for many years to treat hyperlipidemia in humans. This therapeutic effect along with their inhibition of collagen synthesis and ECM remodeling as well as fibrosis^41–43 suggested that PPARα ligands including fibrates warrant assessment as candidates for controlling myopia progression. 

We describe here the drug-induced effects of PPARα activation and inhibition on refraction and ocular growth in pigmented guinea pigs exposed to normal and form-deprived environments. The location of ME1- and apoA-II–immunostained cells as well as changes in scleral collagen type I expression levels provides insight into how PPARα modulation affects myopia progression. 

Materials and Methods

Animals and Experimental Design

The animal research was approved by the Animal Care and Ethics Committee at Wenzhou Medical University (Wenzhou, China). Animal treatment and care were conducted according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Three-week-old pigmented guinea pigs were reared in 12-hour light-dark cycles and had ad libitum access to food and water. Guinea pigs were randomly assigned to two different groups: unoccluded and FD groups. Each of these groups had noninjection control, vehicle-, and drug-injection subgroups (Table). Form deprivation myopia (FDM) was induced using a unilateral facemask as previously described.¹¹ The right eye served as the experimental treated eye (T) whereas the contralateral left fellow eye was left untreated (F). Ocular refraction and biometric parameters along with intraocular pressure (IOP) were recorded in both eyes of individual animals before and after 2 and 4 weeks of treatment. Electroretinograms (ERG) were recorded in treated eyes after 4 weeks of treatment. 

Table

 

Treatment Groups and Sample Sizes

 

Drugs

GW6471 (Tocris Bioscience, Glasgow, UK), GW7647 (Tocris Bioscience), clofibrate (Sigma-Aldrich Corp., St. Louis, MO, USA), and bezafibrate (Sigma-Aldrich Corp.) were dissolved in dimethyl sulfoxide (DMSO). Stock solutions were diluted with 0.9% saline immediately before use. The final drug dosage and sample sizes under each condition are shown in the Table. The final vehicle DMSO concentration of 0.1% vol/vol was the same in all the injection samples. Drugs were injected once daily (9:00 AM) in 100-μL aliquots into the right eye in the peribulbar space after applying one drop of 0.5% proparacaine hydrochloride (Alcon, Puurs, Belgium).⁴⁴ The left eye remained untreated. All injections were administered within 10 seconds under dim red light to minimize photoinduced decomposition of the drugs. 

Biometric Measurements

All measurements were carried out under minimal lighting after facemask removal. Refraction was measured in the vertical meridian with a custom-built eccentric infrared photoretinoscope.⁴⁵ Three readings were recorded for each measurement and the mean value was considered as the final refractive error. 

Ocular biometrics such as the anterior chamber depth (ACD), lens thickness (LT), vitreous chamber depth (VCD), and axial length (AL) were measured with an A-scan ultrasonograph (AVISO Echograph Class I-Type Bat; Quantel Medical, frequency: 11 MHz; Clermont-Ferrand, France) after applying one drop of 0.5% proparacaine hydrochloride (Alcon). Predicted sound velocities across each ocular component were the same as those previously described: 1557.5 m/s (aqueous humor), 1723.3 m/s (lens), and 1540 m/s (vitreous humor).⁴⁶ Six readings were recorded for each measurement and the mean value was considered as the final ACD, LT, VCD, and AL. 

IOP Measurements

IOP was measured with TonoVet TV01 Tonometer (TioLat, Helsinki, Finland) without anesthesia according to the manufacturer's recommended protocol.^47,48 The probe was approximately positioned at the center of the corneal surface at a vertical angle. The TonoVet acquires six measurements and displays an average IOP. In this study, three mean values were obtained from each eye, and the mean of these readings was taken as a single data point. 

ERG Recording

ERG was measured with Roland Electrophysiological Test Unit (Q450SCUV; Roland Consult, Wiesbaden, Germany) as previously described.⁴⁹ After overnight dark adaptation, guinea pigs were anesthetized with an intraperitoneal injection of a mixture of ketamine hydrochloride (60 mg/kg) and xylazine hydrochloride (9 mg/kg). Body temperature was maintained at 37°C by placing animals atop a warming pad. The pupils were dilated with tropicamide eye drops (Santen Pharmaceutical Co. Ltd., Osaka, Japan). A small amount of 2.5% methylcellulose gel was applied to the eye, and a custom gold wire loop electrode was placed over the central cornea as the active electrode. Needle reference and ground electrodes were inserted into the cheek and right foot, respectively. Four levels of stimulus intensity (−3.699, −2.201, 0.301, and 0.799 log cd·s/m²) were used for the dark-adapted ERG recording. For intensities −3.699 and −2.201 log cd·s/m², ERGs were averaged from five single flashes and the interstimuli interval was 15 seconds. For intensities 0.301 and 0.799 log cd·s/m², ERGs were averaged from three single flashes and the interstimuli interval was 40 seconds. Following 10 minutes of light adaptation (1.398 log cd/m²), three photopic stimuli (−0.699, 0.301, and 1.301 log cd·s/m²) were superimposed on this background white lighting used for light adaptation, with an interstimuli interval of 0.4 seconds. Each cone photopic ERG response represents the average of 50 responses. 

Immunofluorescent Staining

The animals were overdosed with ketamine and xylazine followed by eye enucleation and 1-hour fixation in 4% paraformaldehyde. The anterior segment tissues were removed and the posterior eyecup was dehydrated in 30% sucrose for 24 hours. Specimens were embedded in frozen section medium, Neg-50 (Thermo Fisher Scientific, Waltham, MA, USA), and rapidly frozen in liquid nitrogen. Vertical sections 12 μm thick were cut on a freezing cryostat and mounted onto glass slides. 

Frozen sections were allowed to thaw and dry until a pap-pen smear could be placed around pieces of the tissue and rehydrated for 5 minutes in phosphate-buffered saline (PBS) thrice. Slides were blocked for 30 minutes at room temperature with 10% normal donkey serum in a mixture of PBS and 0.3% Triton X-100, followed by incubation for 24 hours at 4°C with anti-PPARα polyclonal antibody (cat. no. sc-9000; Santa Cruz Biotechnology, Dallas, TX, USA), anti-ME1 polyclonal antibody (cat. no. sc-135303; Santa Cruz Biotechnology), or anti-apoA-II polyclonal antibody (cat. no. ab24241; Abcam, Cambridge, MA, USA) at dilutions of 1:50, 1:100, and 1:250, respectively. After rinsing the sections with PBS, either CY3 or FITC conjugated donkey anti-rabbit IgG (1:400; Invitrogen, Waltham, MA, USA; cat. no. A31570 and A21208, respectively) was applied and the tissues were incubated for 2 hours at room temperature. Sections without primary antibody exposure were used as negative controls. 

The sections were mounted with phenylenediamine and examined with a laser (Z1; Carl Zeiss Vision, Göttingen, Germany) and a confocal microscope (LSM510 META; Carl Zeiss Meditec, Göttingen, Germany). The number of PPARα-, ME1-, and apoA-II–immunostained cells in the posterior segment of the sclera was estimated at four scleral locations, two each from either side of the optic nerve; the average number of positively stained cells in the four scleral locations was recorded as a single data point. 

Western Blot Analysis

Following 4 weeks of treatment, the eyes in the normal, GW6471, FDM, and FDM + bezafibrate groups were enucleated to isolate their sclera.²¹ The scleral samples were stored at −80°C until protein extraction. Disruption and homogenization of the tissue was carried out in RIPA Lysis buffer (Beyotime Biotechnology, Shanghai, China), PMSF (Beyotime Biotechnology), and EDTA-free Protease Inhibitor Cocktail (4693159001; Roche, Basel, Switzerland), using a ball mill (MM400; Retsch, Düsseldorf, Germany) at a frequency of 30 Hz for 10 minutes in a 4°C cold room. The homogenate was further lysed at 200 watts for 15 seconds in an ultrasonic crusher (JY92-2D; Ningbo Scientz Biotechnology, Ningbo, China) on ice. After centrifuging the samples at 13,000g for 10 minutes at 4°C, the supernatant was collected and its protein concentration was estimated using the BCA protein assay (Beyotime Biotechnology). 

Electrophoretic separation of protein samples was performed on BG-subMINI horizontal electrophoresis apparatus (BAYGENE, Beijing, China). Protein (50 μg) was separated on 8% SDS-PAGE and blotted onto a nitrocellulose membrane (Merck Millipore, Darmstadt, Germany). Nonspecific membrane binding was blocked with 5% nonfat milk for 2 hours at room temperature. The membranes were then incubated with primary antibodies against collagen type I (1:2000; Abcam, cat. no. ab88147) and β-actin (1:2000; Sigma-Aldrich Corp., cat. no. a5441) overnight at 4°C. After rinsing the membranes with PBST (1000 mL PBS and 1 mL Tween), they were incubated with the secondary antibody (goat anti-mouse, 1:10,000, LI-COR Biosciences, Lincoln, NE, USA, cat. no. 926-32210) for 2 hours at room temperature. Protein bands were visualized using an Odyssey Infrared Imaging System (LI-COR Biosciences), with the signal intensities normalized to β-actin and densitometric analysis carried out using ImageJ software (National Institutes of Health, Bethesda, MD, USA). 

Statistical Analysis

Significance was evaluated using the Statistical Package for the Social Sciences (SPSS version 16.0, Chicago, IL, USA) software. All results are expressed as mean ± standard deviation (SD). Intergroup differences of the biometric parameters, IOP, and ERG parameters were estimated by 2-way ANOVA and corrected for multiple testing using the post hoc Bonferroni correction unless otherwise stated. For immunohistochemistry and protein expression experiments, paired t-tests were performed to assess the changes between treated and fellow eyes; unpaired t-tests were used to compare the changes between the various treatment groups. 

Results

Ocular refraction, ACD, LT, VCD, and AL values were not significantly different between animals in the DMSO-injected and the control groups (normal versus DMSO and FDM versus FDM + DMSO) (P > 0.05, Figs. 1–4, Supplementary Tables S1–S8), implying a lack of a vehicle effect and no apparent corneal surface change due to administration of an anesthetic. Similarly, all drug injections did not affect the ACD or LT compared to the DMSO-treated group (P > 0.05, Supplementary Fig. S1). 

Figure 1

 

Effects of GW6471 on refraction, VCD, and AL in unoccluded eyes. GW6471 injections (4.5 μg) for 4 weeks induced myopic refractive errors (A), increased VCD (B) and AL (C) in a time-dependent manner (*P < 0.05, **P < 0.01, ***P < 0.001). Data expressed as mean difference between treated and fellow eyes at each time point.

Figure 2

 

Effects of GW6471 on refraction, VCD, and AL in form-deprived eyes. Injections of GW6471 (0.45 and 4.5 μg) for either 2 or 4 weeks did not affect the refractive error (A), VCD (B), or AL (C) in form-deprived eyes. Data expressed as mean difference between treated and fellow eyes at each time point.

Figure 3

 

Effects of GW7647 on refraction, VCD, and AL in unoccluded eyes. Injections of GW7647 (0.5 and 1.5 μg) for either 2 or 4 weeks did not affect the refractive error (A), VCD (B), or AL (C) in unoccluded eyes. Data expressed as mean difference between treated and fellow eyes at each time point.

Figure 4

 

Effects of GW7647 on refraction, VCD, and AL in form-deprived eyes. GW7647 injections (1.5 μg) reduced myopia progression (A), VCD (B), and AL (C) in a time-dependent manner (*P < 0.05, **P < 0.01, ***P < 0.001). Data expressed as mean difference between treated and fellow eyes at each time point.

PPARα Antagonist GW6471 Promotes Myopia Progression in Unoccluded But Not Form-Deprived Eyes

Two weeks of GW6471 injections (4.5 μg) during exposure to a normal visual environment resulted in a myopic shift relative to the DMSO-treated animals (interocular differences: −1.45 ± 0.52 vs. −0.21 ± 0.64 D, P < 0.001, Fig. 1A, Supplementary Table S1). After extending the treatment to 4 weeks, the myopic shift increased more relative to the DMSO-treated eyes (interocular differences: −1.77 ± 1.02 vs. −0.22 ± 1.02 D, P = 0.001, Fig. 1A, Supplementary Table S1). In parallel with refractive changes, GW6471 injections increased the VCD (interocular differences: 0.05 ± 0.03 vs. 0.01 ± 0.03 mm, P = 0.001, Fig. 1B, Supplementary Table S1) and AL (interocular differences: 0.05 ± 0.03 vs. 0.01 ± 0.03 mm, P = 0.002, Fig. 1C, Supplementary Table S1). 

On the other hand, after 4 weeks of antagonist injections, the refractive error, VCD, and AL of form-deprived eyes were not significantly different from those injected with DMSO (Fig. 2, Supplementary Table S2). 

PPARα Agonist GW7647 Inhibited FDM Progression But Had No Effect on Unoccluded Eyes

In unoccluded eyes, after 4 weeks of daily GW7647 injections (0.5 or 1.5 μg), neither the refraction, VCD, nor AL (Fig. 3, Supplementary Table S3) changed relative to that in DMSO-treated eyes. 

On the other hand, after 2 weeks of injections, this PPARα agonist inhibited myopia progression as compared to the DMSO-treated form-deprived eyes (interocular differences: −3.68 ± 1.92 vs. −6.00 ± 2.57 D, P = 0.012, Fig. 4A, Supplementary Table S4). Significant relative inhibition was also observed after 4 weeks of treatment (interocular differences: −5.94 ± 1.12 vs. −8.81 ± 2.48 D, P = 0.001, Fig. 4A, Supplementary Table S4). This difference in refractive error changes at 4 weeks accompanied reduced changes in VCD (interocular differences: 0.14 ± 0.05 vs. 0.20 ± 0.07 mm, P = 0.008, Fig. 4B, Supplementary Table S4) and AL (interocular differences: 0.15 ± 0.05 vs. 0.23 ± 0.07 mm, P = 0.004, Fig. 4C, Supplementary Table S4). Taken together, GW7647 had no effect on unoccluded eyes but reduced the development of FDM and inhibited ocular elongation in these eyes. 

Fibrates Suppress Myopia in Form-Deprived Eyes

Similar to GW7647, after 4 weeks of clofibrate (2 μg) injections the myopic refractive errors of form-deprived eyes decreased compared to DMSO-injected form-deprived eyes (interocular differences: −7.96 ± 2.41 vs. −10.62 ± 2.55 D, P = 0.023, Fig. 5A, Supplementary Table S6). This inhibition was accompanied by reduced changes in VCD (interocular differences: 0.15 ± 0.04 vs. 0.22 ± 0.06 mm, P = 0.006, Fig. 5B, Supplementary Table S6) and AL (interocular differences: 0.15 ± 0.04 vs. 0.22 ± 0.05 mm, P = 0.004, Fig. 5C, Supplementary Table S6). 

Figure 5

 

Effects of 4 weeks of clofibrate and bezafibrate injections on refraction, VCD, and AL in form-deprived eyes. Injections of 2 μg clofibrate (high dose) and 3.5 μg bezafibrate (high dose) reduced myopia progression (A), VCD (B), and AL (C) after 4 weeks of treatment. (*P < 0.05, **P < 0.01, ***P < 0.001), whereas lower doses of clofibrate (0.2 μg) and bezafibrate (0.35 μg) had no effect on these parameters. Data expressed as mean difference between treated and fellow eyes at each time point.

Similar to clofibrate, after 4 weeks of bezafibrate (3.5 μg) injections, myopia progression decreased in form-deprived eyes compared to the DMSO-injected form-deprived eyes (interocular differences: −6.14 ± 1.86 vs. −9.14 ± 2.37 D, P = 0.001, Fig. 5A, Supplementary Table S8). Bezafibrate injections inhibited the effect of FD on VCD (interocular differences: 0.11 ± 0.03 vs. 0.17 ± 0.03 mm, P < 0.001, Fig. 5B, Supplementary Table S8) and AL (interocular differences: 0.12 ± 0.04 vs. 0.18 ± 0.04 mm, P < 0.001, Fig. 5C, Supplementary Table S8) in form-deprived eyes. 

On the other hand, after 4 weeks of clofibrate and bezafibrate injections neither the refraction, VCD, nor AL changed in the unoccluded eyes relative to the corresponding values in the DMSO-treated unoccluded eyes (Supplementary Tables S5, S7). Taken together, similar to GW7647, both of these fibrates had no effect on unoccluded eyes, but inhibited the effects of FD on refraction, VCD, and AL in form-deprived eyes. 

ERG and IOP Unaltered by GW7647 and GW6471

After 4 weeks of injections with either DMSO, GW7647, or GW6471, IOP was unchanged in both eyes by either drug or vehicle injections for 2 and 4 weeks (P > 0.05, Supplementary Figs. S2, S3). Similarly, scotopic and photopic ERGs were unaltered and similar to one another in the drug-injected and DMSO-injected groups (P > 0.05, Supplementary Figs. S4, S5) indicating that these drugs had no obvious effect on retinal function. 

Drug-Induced PPARα Modulation Inhibits Myopia Progression by Inhibiting Form Deprivation–Induced Decreases in Collagen Type I Expression

Scleral collagen type I expression levels were reduced in unoccluded eyes after 4 weeks of GW6471 injections (4.5 μg) relative to that in the fellow eyes (P = 0.001, paired-samples t-test, Figs. 6A, 6B). In contrast, bezafibrate injections (3.5 μg) inhibited FD-induced decrease in collagen type I expression in the sclera (form-deprived eye versus fellow eyes: P = 0.004, paired-samples t-test, Figs. 6C, 6D). 

Figure 6

 

Opposing effects of GW6471 and bezafibrate on collagen type I expression. (A) Scleral collagen type I expression levels decreased with GW6471 injections. (B) Summary of Western blot collagen type I expression densitometric analysis in normal and GW6471-treated eyes (n = 4/group). (C) Bezafibrate inhibited the decreases in collagen type I expression in the sclera of form-deprived eyes. (D) Summary of Western blot collagen type I expression densitometric analysis in form-deprived and bezafibrate-treated form-deprived groups (n = 4/group, **P < 0.01; R, right eye; L, left eye; T, treated eye; F, fellow eye).

Opposing Effects of GW6471 and Bezafibrate on ApoA-II Expression in the Posterior Scleral Segment

Even though PPARα immunofluorescent staining was similar in the form-deprived and their fellow control eyes, the expression of its downstream mediators, ME1 (P = 0.014, paired-samples t-test) and apoA-II (P = 0.029, paired-samples t-test), decreased in form-deprived eyes (Fig. 7B). Similarly, their expression was reduced in the GW6471-treated unoccluded eyes compared to that in the fellow eyes (ME1: P = 0.002, and apoA-II: P = 0.035, paired-samples t-test, Fig. 7C). 

Figure 7

 

PPARα, ME1, and apoA-II cell number modulations in the posterior segment of the sclera induced by FD, GW6471, and bezafibrate. Cells expressing PPARα, ME1, and apoA-II (indicated by arrows) in four posterior scleral sections (as described in the pictorial representation) (A). The number of ME1- and apoA-II–labelled cells in the posterior sclera of FD eyes (5 eyes) was reduced compared to that in fellow and normal control eyes (5 eyes) (B). GW6471 suppressed scleral cell populations of ME1 and apoA-II in unoccluded eyes (6 eyes) (C). The cell counts of apoA-II increased, while decreases in ME1-stained cells were not inhibited by bezafibrate injections in form-deprived eyes compared to fellow eyes (6 eyes) (D). *P < 0.05, **P < 0.01; T, treated eye; F, fellow eye. Scale bar: 20 μm.

We also evaluated if bezafibrate could inhibit the decreases in ME1- and apoA-II–stained cells induced by FD. Interestingly, bezafibrate injections (3.5 μg) in form-deprived eyes increased the density of apoA-II–stained scleral cells as compared to those in form-deprived eyes without any injections. These increases in apoA-II–positive cells reached a level that was not significantly different from that of the fellow eyes (P = 0.739, paired-samples t-test, Fig. 7D). On the other hand, the decrease in ME1 counts was not inhibited by bezafibrate treatment, which continued to be still significantly lower than that in the fellow eyes (P = 0.038, Fig. 7D). 

Discussion

Opposing Effects of PPARα-Related Drugs on Myopia Development

We show here that PPARα agonist and antagonist treatments had opposing effects on myopia progression; specifically, GW7647 inhibited myopia progression in form-deprived eyes whereas GW6471 induced a myopic shift in unoccluded eyes. However, it is unclear if GW7647 (1.5 μg in 100 μL, 29 μM) solely modulated the PPARα signal pathway given that this concentration is high enough to also activate PPARβ and PPARγ, since their EC₅₀ values are 6.2 and 1.1 μM, respectively. Nevertheless, its inhibitory effect on FDM agrees with a previous study in which a similarly high dose of GW7647 was employed in chicks.²⁵ Furthermore, the fact that GW6471 promoted myopia development in a normal visual environment is supportive of the notion that PPARα contributes to ocular growth control since it is a selective inhibitor of this transcription factor.⁵⁰ Even though GW7647 and GW6471 had opposing effects on refractive error, it is unclear how much each of these changes is attributable to either interactions solely with PPARα or downstream in the PPARα signaling pathway. 

Off-target PPARα drug effects can also be involved. For example, fenofibrate suppressed the growth of human hepatocellular carcinoma cells, whereas neither the PPARα antagonist (GW6471) nor a PPARα-specific siRNA had opposing effects, suggesting a PPARα-independent mechanism.⁵¹ Furthermore, fenofibrate affects the survival rate of retinal endothelial cells instead through the MAPK signaling pathway.⁵² Given that different drug dosages were employed in the above studies, a detailed dose-dependent study may be warranted in the future to clarify the pathways involved in mediating observed changes. 

Fibrates suppress the pathogenesis of various metabolic disorders by significantly reducing stress-related inflammatory responses and fibrosis as well as inhibiting autophagy.^43,53 Our results show that clofibrate and bezafibrate to some extent inhibited myopia progression and reduced axial elongation induced by FD without altering normal ocular development. This apparent selective effect on FD eyes suggests that these fibrates may be developed for use in the treatment of myopia without interfering with emmetropization. However, future studies are still needed to assess their safety and effectiveness. 

Opposing Effects of PPARα-Related Drugs on Collagen Type I Expression

The mammalian sclera is mainly composed of different collagen subtypes that contribute to approximately 90% of its scleral tissue weight, with collagen type I being the major subtype (Zorn N, et al. IOVS 1992;33:ARVO Abstract S1053).⁵⁴ Hence, limiting the increased collagen turnover and degradation is a viable option toward myopia control as loss of collagen content contributes to reduced scleral biomechanical strength during myopia progression.⁵⁵ The finding that GW6471 injections reduced collagen type I expression and induced myopia in guinea pigs exposed to a normal visual environment, with bezafibrate injections exhibiting the opposite effect, suggests that PPARα could be a potential target for therapeutic management of myopia. 

While bezafibrate reduced myopia progression by inhibiting decreases in scleral collagen expression in form-deprived eyes, in various other tissues this PPARα agonist had an antifibrotic effect.^56,57 Another PPARα agonist, WY14643, inhibited the severity of steatohepatitis with reductions in mRNA levels of type I collagen.⁵⁸ Similarly, fenofibrate markedly suppressed liver fibrosis and reduced collagen deposition by 50%. This was associated with a significant downregulation of collagen type I and a marked decrease in α-SMA expression.³⁴ Besides these pharmacologic effects, in PPARα transgenic mice TGF-β levels declined along with reduced ECM protein content, including collagen type I, fibronectin, α-SMA in an unilateral ureteral obstruction (UUO)-mediated renal fibrosis model.³⁵ While the reasons for this lack of agreement with the current study are unclear, the mode of drug administration, as well as species and pathologic differences could be contributing factors. 

Bezafibrate Subverts Decreases of ApoA-II–Stained Cells in Form-Deprived Eyes

To elucidate a possible role of PPARα signaling pathway mediators in facilitating the effects of PPARα agonist and antagonists, we evaluated the changes in density of PPARα-, ME1-, and apoA-II–immunostained cells following pharmacologic and/or visual manipulations. Even though neither FD nor GW6471 altered the density of PPARα-immunostained cells, ME1 and apoA-II immunofluorescent intensities were reduced in the posterior segment of the sclera in both form-deprived and GW6471-injected eyes as compared to those in fellow or normal control eyes. Furthermore, bezafibrate inhibited the decrease in apoA-II intensity induced by FD. These results suggest that modulation of apoA-II levels may occur through changes in PPARα activity. However, the cell count estimates alone may not be adequate to conclusively prove the involvement of PPARα signaling pathway as our target tissue included only the posterior sclera rather than the whole sclera. 

Bertrand et al.²⁵ demonstrated that apoA-I acts as a STOP signal in regulating the expression of plasminogen and Wnt signaling, which are both involved in mediating axial elongation during myopia development. Since GW7647 induced apoA-I upregulation, this effect supports the notion that PPARα activation contributes to inhibiting myopia progression.²⁵ This conjecture is consistent with our finding that the PPARα agonist, bezafibrate, inhibited form deprivation–induced apoA-II expression reduction. Such reversal accompanied declines in FDM, which suggests a possible role for apoA-II in mediating PPARα control of myopia. 

The involvement of PPARα activity in lipid metabolism combined with its role in various myopia-associated metabolic processes points to the possibility that PPARα-mediated regulation of gene products in lipid metabolism may also contribute to myopia.^29,31,59,60 Besides its possible role in lipid and glucose homeostasis, PPARα activity mediates various events involving nitric oxide,^61,62 retinoic acid,^63,64 and vitamin D,⁶⁵ all of which have been linked to eye growth regulation in some way.^66–68 

Conclusions

GW7647 and GW6471 had vision environment-dependent opposing effects on refractive development in pigmented guinea pigs. These responses are accompanied by changes in apoA-II, a PPARα downstream signaling pathway mediator, and in the expression of a major scleral ECM component, collagen type I. While further research into the contributory role of PPARα is necessary, we provide possible insights into the importance of this transcription factor in mediating changes in scleral collagen expression and refractive development in the guinea pig model. That PPARα agonists reduced myopia development by 20% to 30% in form-deprived guinea pigs raises the possibility that fibrates may provide an option for improved management of myopia in humans. 

Acknowledgments

Supported by Grant 81670886, 81371047, 81470659 from the National Natural Science Foundation of China; Grant 81422007 from the National Science Foundation for Excellent Young Scholars of China; Grants LZ14H120001 from Natural Science Foundation of Zhejiang Province; the Zhejiang Provincial Program for the Cultivation of High-Level Innovative Health Talents; and the National Young Excellent Talents Support Program. 

Disclosure: M. Pan, None; S. Jiao, None; P.S. Reinach, None; J. Yan, None; Y. Yang, None; Q. Li, None; N. Srinivasalu, None; J. Qu, None; X. Zhou, None