Introduction

Peroxisome proliferator-activated receptors (PPARs) are involved in the body’s energy metabolism and are classified into three subtypes: PPAR-α, PPAR-β, and PPAR-γ. PPAR-α is distributed mainly in metabolic tissue, including the heart, skeletal muscle, liver, and retina, and it is activated by free fatty acids and leukotriene B4 as physiological ligands, which play a pivotal role in fatty acid metabolism [1,2,3]. PPAR-α has been shown to inhibit oxidative stress and degeneration in retinal neuronal cells in retinopathy model mice induced by high oxygen [4]. Fibrates are artificial ligands of PPAR-α that promote the beta-oxidation of fatty acids and lower triglycerides in the blood [5,6,7]. PPAR-α has recently attracted attention for its anti-inflammatory effects, but the precise mechanism by which PPAR-α confers such protection has not been elucidated. PPAR-α can inhibit the pro-inflammatory gene expression, leading to vascular inflammation reduction [5,6,7].

Fenofibrate is a PPAR-α agonist that has been used in the treatment of dyslipidemia. Large clinical trials, i.e., the Fenofibrate Intervention and Event Lowering in Diabetes trial [8] and the Action to Control Cardiovascular Risk in Diabetes trial [9,10,11], have shown that administration of fenofibrate inhibits the diabetic retinopathy progression [8,9,10,11]. A recent study has shown that fenofibrate reduces reactive oxygen species formation and ameliorates diabetic retinopathy [12]. Pemafibrate has been used as a novel treatment for fibrate dyslipidemia, and recently it has become clear that it has various functions in vivo activities [13, 14]. It has been reported to be safer than fenofibrate [15] and to be highly selective for PPAR-α [16].

c-Jun is a transcription factor belonging to the transcription factor activator protein 1 group that is phosphorylated to become p-c-Jun and is linked to the expression of apoptosis-related genes. Previous studies have reported that upregulated p-c-Jun and p-JNK are accompanied with neuronal death in the retina of diabetic rats [17], and that inhibition of p-c-Jun activity might be neuroprotective in Parkinson’s disease model rats [18].

Excess extracellular glutamate has been linked to diabetic retinopathy. For example, increase in vitreous glutamate has been shown in diabetes patients [19]. Other reports have shown that glutamate levels were increased in retinal Muller cells from diabetic rats and that such elevation might cause glutamate excitotoxicity [20, 21]. N-methyl-D-aspartate (NMDA) is a glutamate receptor agonist that is implicated in neuronal death and has been used in various studies to model neuronal cell death. Intravitreal administration of NMDA has been exhibited to activate nuclear factor-κB, inflammatory cytokines, and apoptosis-related factors and to produce retinal neurotoxicity [22]. Since diabetic retinopathy causes amacrine cell death and retinal ganglion cell (RGC) death [23,24,25,26], clarifying the mechanism of such cell death may be useful for understanding diabetic retinopathy.

In the current study, we investigated the effect of pemafibrate in rats with retinal neurotoxicity induced by NMDA and examined its alteration on p-c-Jun as a possible mechanism.

Materials and methods

Animals

A total of 41 8-week-old male Wistar rats were used for the study. The animals were housed in a room in which temperature (23 ± 1 °C), humidity (55 ± 5%), and lighting (light from 06:00 to 18:00) were controlled. Labeling of RGCs, administration of pemafibrate, intravitreal injection and measurement of RGC counts were performed on the schedule shown in Fig. 1. The study was conducted according to the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research.

Fig. 1
figure 1

Flow diagram of the experimental procedure

Full size image

Labeling of RGCs

Rats’ RGCs were labeled retrogradely with a neurotracer, as described previously [27]. The animals were anesthetized by intramuscular injection of a mixture of ketamine and xylazine. The cerebrum was aspirated, and pieces of gelfoam soaked in a 6% solution of Fluoro-Gold (Fluorochrome, Denver, CO, USA) were placed at the superior colliculus bilaterally.

Division of rats into groups

One day after the RGC labeling, rats were randomly assigned to administration of pemafibrate (pemafibrate group; n = 5) or to non-administration of pemafibrate (non-pemafibrate group; n = 5). For those in the pemafibrate group, pemafibrate 2.5 mg (10 mg/kg; provided by Kowa Co. Ltd., Nagoya, Japan) was dissolved in methylcellulose and administered orally once a day for 7 days. Non-pemafibrate group was also made.

NMDA and PBS intravitreal administration

One day after administration/non-administration of pemafibrate, NMDA was administered intravitreally as described previously [28]. Briefly, rats were anesthetized by intraperitoneal injection of sodium pentobarbital, and 2 μl of 1 × 10−2 M NMDA (Sigma-Aldrich) was then injected intravitreally into the left eye of each animal under microscopy, and phosphate-buffered saline (PBS) was injected intravitreally into the rats’ right eye.

Counting RGCs

Seven days after the intravitreal injection, the rats were killed by pentobarbital overdose, and their eyes were removed. The eyes were fixed and flat mounted. The tracer-labeled RGCs were counted at 500 μm (center area) and 2 mm (mid-periphery area) from the edge of the optic disc under a microscope (Carl Zeiss, Jena, Germany) with the use of UV filters at 8 different areas of 151,875 μm2 each (two areas per retinal quadrant). Distinct glial cells (bright, small, spindle-shaped cells) were excluded from the counts.

Immunoblotting

Expression of p-c-Jun in the sensory retinas that had been removed from the enucleated eyes of 11 rats was examined by Western blotting. The retinas were homogenized in protein extraction buffer. Centrifugation was performed at 15,000 g for 15 min at 4 °C, and supernatants were obtained. The protein concentration was determined with use of a commercial protein assay kit (Bio-Rad Laboratories). Samples (8 μg each) were loaded onto SDS polyacrylamide gels (Bio-Rad Laboratories) and transferred to PVDF membranes (Immobilon-P, Merck Millipore, Tullagreen, Ireland). The membranes were incubated in Tris-buffered saline containing 0.1% Tween 20, and one of the following primary antibodies: anti-phospho-c-Jun antibody (1:200 dilution) (Cell Signaling Technology, Inc., Beverly, MA, USA) or beta-actin antibody (1:5000 dilution) (Sigma). The membranes were then reacted with peroxidase-conjugated anti-rabbit IgG or anti-mouse IgG antibody (1:5000 dilution) (Cappel, Aurora, OH, USA) and visualized with a chemiluminescence detection system (ECL Plus Western Blotting Detection Reagents). The bands were then scanned and analyzed quantitatively with the use of NIH Image software.

Immunohistochemical analysis

Twelve hours after vitreous injection of NMDA or PBS, removed eyes of 10 rats were fixed by immersion in 4% paraformaldehyde for 24 h. After paraffin processing and sectioned, deparaffinized sections were blocked with 1% bovine serum (Roche Diagnostics GmbH, Mannheim, Germany) and incubated with anti-p-c-Jun antibody (Cell Signaling Technology, Inc.) or anti-Thy-1 antibody (Santa Cruz Biotechnology) diluted 1:100 in 1% BSA in PBS overnight. The secondary antibodies were FITC-labeled anti-rabbit antibody, or rhodamine-labeled anti-mouse antibody (Cappel, Aurora, OH, USA). Photomicrographs of the sections were obtained with a confocal microscope (LSM510, Carl Zeiss).

TUNEL assay

TUNEL assay was used to detect retinal cells undergoing apoptosis. The assay was performed 12 h after the NMDA and PBS injections and applied in situ to 10 rats with a fluorescein apoptosis detection system (Promega, Madison, WI) [29]. TUNEL-positive cells were counted at 1.0–1.5 mm from the edge of optic disc on photos obtained with a confocal microscope (Carl Zeiss). A positive control group was also made using DNase (Promega, Madison, WI) in the PBS group.

TUNEL combined with p-c-Jun Immunohistochemistry

After TUNEL assay, slides were double-stained with anti-p-c-Jun antibody, as described above.

Statistical analysis

Data are presented as mean ± SEM. Differences between groups were analyzed by one-way analysis of variance (ANOVA) followed by Dunnett’s test or Mann-Whitney U test. A P value of less than 0.05 was considered statistically significant.

Results

Effects of pemafibrate on NMDA-induced neurotoxicity

Using retrograde labeling with Fluoro-Gold, we found a significant decrease in the number of RGCs in the NMDA vitreous injection group compared to the control group (Fig. 2). The number of RGCs at the center in the control group was 2963 ± 349/mm2, while the number was 962 ± 65/mm2 in the NMDA group. This decrease was significantly ameliorated in the systemic pemafibrate treatment group (RGC number was 1350 ± 61/mm2; p < 0.05; Fig. 2b). Similarly, the number of mid-periphery RGCs was 2072 ± 139/mm2 in the control group, while it was 752 ± 77/mm2 in the NMDA group (Fig. 2c). This decrease was significantly prevented by the pemafibrate treatment (RGC number was 1106 ± 48/mm2; p < 0.05; Fig. 2c).

Fig. 2
figure 2

Effects of pemafibrate on RGC loss induced by NMDA. (a) Neurotracer-labeled RGCs in flat preparations. Scale bar = 50 μm. (b) Number of RGCs at center area. n = 4–5. *P < 0.05. (c) Number of RGCs at mid-periphery area. n = 4–5. *P < 0.05

Full size image

Effects of pemafibrate on p-c-Jun protein levels in the retina

Immunoblotting showed a significant increase in the p-c-Jun protein levels in the retina 12 h after NMDA injection compared to the control group (Fig. 3). However, this increase was significantly prevented by the systemic pemafibrate treatment (Fig. 3).

Fig. 3
figure 3

Effects of pemafibrate on p-c-Jun expression in retina. (a) Representative immunoblotting of p-c-Jun. (b) Band density. Data are normalized with β-actin. n = 5–6. *P < 0.05

Full size image

Localization of p-c-Jun in the rat retina

Immunohistochemical analysis showed that faint p-c-Jun immunoreactivity was observed in the control group, but 12 h after NMDA injury, p-c-Jun immunoreactivity was apparent (Fig. 4e). They were observed in mainly RGC layer (RGCL) and inner nuclear layer (INL) and were colocalized with Thy-1 immuno-positive cells in the RGCL (Fig. 4d-f). In the pemafibrate group, p-c-Jun immunoreactivity tended to be decreased compared to NMDA group, and they were also observed in Thy-1 positive cells in the RGCL (Fig. 4g-i).

Fig. 4
figure 4

Immunohistochemistry of p-c-Jun (b, e, h) and Thy-1 (a, d, g) in retina. (a-c) PBS- treated retina, (d-f) NMDA-treated retina and (g-i) NMDA + pemafibrate treated retina. Arrows indicate colocalizations of Thy-1 and p-c-Jun immunopositive cells. Scale bar = 50 μm. n = 6

Full size image

Effects of pemafibrate on p-c-Jun in NMDA-induced apoptosis

TUNEL staining showed abundant apoptotic cells in the RGCL and INL in the NMDA vitreous injection group and they were decreased in the pemafibrate treatment group (Fig. 5a). The number of TUNEL-positive cells in the RGCL was significantly lower in the pemafibrate oral administration group compared to the NMDA group (Fig. 5b). Although the number of TUNEL-positive cells in the INL tended to be lower in the pemafibrate group compared to the NMDA group, the difference was not statistically significant (p = 0.075; Fig. 5b). Moreover, the double staining revealed that p-c-Jun immuno-positive cells were co-localized with TUNEL-positive cells after NMDA injection, indicating that they are consistent with apoptotic cells (Fig. 5c). Furthermore, such colocalized cells were not as abundant in the pemafibrate group as they were in the NMDA group (Fig. 5c).

Fig. 5
figure 5

TUNEL assay in retina. (a) TUNEL-positive cells 12 h after NMDA injection with or without systemic pemafibrate treatment. DNase was used for positive control. TUNEL-positive cells were not seen 12 h after PBS injection. Scale bar = 50 μm. (b) TUNEL-positive cell number in the RGCL. n = 5. *P < 0.05. (upper panel). TUNEL-positive cell number in the INL. n = 5. (lower panel). (c)Double-labeling for TUNEL and p-c-Jun immunoreactivity 12 h after NMDA injection with or without systemic pemafibrate treatment. PBS is also shown as a control. Arrows indicate colocalizations of TUNEL-positive cells and p-c-Jun immunopositive cells. Scale bar = 50 μm

Full size image

Discussion

Our study revealed a significant neuroprotective effect of systemic administration of pemafibrate on NMDA-induced retinal neuronal excitotoxicity. A previous study showed a neuroprotective effect of oral fenofibrate administration in a Parkinson’s disease model rat [30]. Other study demonstrated that fenofibrate treatment significantly ameliorated neuronal death in hippocampus following global cerebral ischemia/reperfusion [31]. In addition, it was shown that daily administration of fenofibrate led to neuroprotection in a transgenic mice model of amyotrophic lateral sclerosis [32]. On the other hand, some studies showed the effect of PPAR-α agonist on retinal vessels. For example, oral administration of pemafibrate has been shown to prevent pathological angiogenesis in the retina by increasing FGF21 levels in a mouse model of oxygen-induced retinopathy [33]. In a very recent study, pemafibrate was shown to ameliorate diabetes-induced retinal vascular leukostasis and leakage via thrombomodulin upregulation [34]. Taken together, it is likely that pemafibrate has a beneficial effect not only in terms of vascular permeability but also in terms of neuronal cell survival in retina.

The present study showed that vitreous administration of NMDA led to a significant elevation in the p-c-Jun protein level, and this increase was significantly prevented by systemic pemafibrate treatment. The increase in p-c-Jun 12 h after NMDA injection was consistent with our previous study showing significant increases 6 h and 12 h after NMDA administration [35]. It was shown that ET-1 increased the p-c-Jun protein level and this increase was markedly inhibited by fenofibrate in cultured cardiomyocytes [36]. It is interesting to note that fenofibrate reduced cardiac fibrosis via the inhibition of c-Jun activity [37]. There is a possibility that other signaling pathway participates in the protective effect of pemafibrate because of the discrepancy between large reduction in p-c-Jun and modest protection in RGCs. However, since we previously found that inhibition of p-c-Jun resulted in neuronal cell protection in the RGCL against NMDA injection [35], it is reasonable to speculate that the protective effect of pemafibrate is accompanied with p-c-Jun inhibition. As reported previously [35], p-c-Jun was colocalized with Thy-1 immunoreactivity, and the increased p-c-Jun was ameliorated by pemafibrate treatment, thus implicating decreased p-c-Jun in RGCs may be involved in RGC protection by pemafibrate.

It is well known that vitreous administration of NMDA causes abundant TUNEL-positive cells in the RGCL and INL [38]. The current study also confirmed the presence of abundant TUENL-positive cells in the RGCL and INL. The number of TUNEL-positive cells in the RGCL was significantly lessened by systemic administration of pemafibrate, while the reduction of TUNEL-positive cell number was not significant in the INL. One hypothesis posits that the effects of pemafibrate may vary dependent on cell types as the INL consists of the nucleus of amacrine cells, bipolar cells, horizontal cells, and Müller cells. Colocalization of TUNEL-positive cells and p-c-Jun immunopositive cells implicates the involvement of p-c-Jun in apoptosis, and these colocalized cells were decreased by pemafibrate treatment. In line with this finding, previous studies have suggested that PPAR-α agonists suppress apoptosis. For example, fenofibrate was shown to inhibit palmitate-induced myocardial apoptosis in mice [39]. Moreover, a recent study showed that daily intraperitoneal injection of fenofibric acid significantly decreased retinal apoptosis in STZ-induced diabetes [40]. Hence, long-term studies will be needed to support the idea that pemafibrate has potential for treatment of diabetic retinopathy since apoptosis is involved in the pathophysiology of this disease [41]. Furthermore, more beneficial effects of pemafibrate include usefulness for patients who have renal dysfunction because it is metabolized in the liver [42]. Other effects the pemafibrate could have are protective effects against atherosclerosis and inflammation in hyperlipidemic condition [42]. It has been reported that oral fenofibrate reduced caspase-3 immunoreactivity in a rat stroke model and that prior administration of fenofibrate resulted in a significant reduction in infarct size [43]. Therefore, it is possible that the neuroprotective effect of pemafibrate is accompanied by an anti-apoptotic effect. Further studies are needed to clarify the mechanism underlying the neuroprotection provided by pemafibrate.