Histological, morphometric, protein and gene expression analyses of rat retinas with ischaemia–reperfusion injury model treated with sildenafil citrate

Summary

The aim of this study was to better understand the role of apoptosis in a retinal ischaemia–reperfusion injury model and to determine whether sildenafil citrate treatment can prevent retinal cell apoptosis. Thirty‐six rats were divided into a control group (n = 6) and two experimentally induced ischaemia–reperfusion groups (7 and 21 days; n = 15 per group). The induced ischaemia–reperfusion groups were treated with sildenafil for 7 and 21 days (n = 10 per group), and 10 animals were treated with a placebo for the same period (n = 5 per group). Paracentesis of the anterior chamber was performed with a 30‐G needle attached to a saline solution (0.9%) bag positioned at a height of 150 cm above the eye for 60 min. Intraocular pressure was measured by rebound tonometer (TonoVet^®). The eyes were analysed by histology and morphometry, and by immunohistochemistry and qRT‐PCR for expression of Caspase‐7, Caspase‐6, Caspase‐9, Tnf‐r2, Fas‐l, Bcl‐2 and Bax. Sildenafil‐treated animals showed lower levels of histopathological changes (inflammatory, cellular and tissue) than their placebo‐treated counterparts at both 7 and 21 days. The retinal ganglion cell layer (RGC) was preserved in the sildenafil groups (SG), with a cell count closer to control than in the placebo groups (PG). Caspase‐7 expression was significantly higher in both treated groups at 7 days compared to controls. Gene expression levels in both treatment groups differed from the controls, but in SG Bax and Caspase‐6 expression levels were similar to control animals. These results suggest that the main mechanism of retinal cell death in this model is apoptosis, as there is an increase in pro‐apoptotic factors and decrease in the anti‐apoptotic ones. Also, sildenafil seems to protect the retinal ganglion cell layer from apoptosis. Cell survival was evident in the histological and morphometric analyses, and sildenafil treatment had a protective effect in the apoptosis pathways, with gene expression levels in SG similar to the controls.

Glaucoma is a class of optic neuropathies characterized by gradual degeneration of neuronal tissue due to the loss of retinal ganglion cells (RGCs) and their axons, the optic nerve fibres, with accompanying loss of visual field over time (Hayreh et al. 1999; Quigley 1999; Avisar et al. 2006; Nickells 2007a,b). It is one of the world's major causes of blindness in humans, with elevated intraocular pressure (IOP) being a well‐known major risk factor. In addition, there is growing evidence that local ischaemia, vascular factors and reduced perfusion of the optic nerve may play a role in glaucomatous neuropathy pathogenesis (Yanagi et al. 2001; Quigley & Broman 2006; Flammer & Mozaffarieh 2007; Quigley 2007; Wostyn et al. 2010; Lin et al. 2014).

Several risk factors have been proposed to contribute to glaucoma progression, including elevated IOP, age, genetic background, decreased corneal thickness and vascular deregulation (Nickells 2007a,b; Leske 2009). The presence of any of these factors determines an individual's risk of developing glaucoma, but they are not necessarily the cause of this condition (Weinreb 2007; Leske 2009).

Research into the pathogenesis of glaucoma has been aided by the development of glaucoma models (Johnson & Tomarev 2010). In this context, the retinal ischaemia–reperfusion (I/R) injury occurs in clinical situations such as retinal vascular occlusion diseases and acute glaucoma and has been extensively studied in order to understand the retinal neuronal cell damage which occurs after ischaemic insult, and there is much evidence to suggest that retinal neuronal cells were damaged by apoptosis and necrosis after retinal I/R injury (Lam et al. 1999; Osborne et al. 2004; Kaur et al. 2008; Johnson & Tomarev 2010;).

The current medical and surgical therapy aimed at lowering IOP and preventing visual loss, but with limited success. Therefore, it is evident that there is a need to develop new therapeutic approaches that can be used as hypotensive drugs, to preserve vision (Levin 2005; Weinreb 2007).

Sildenafil citrate (SC), a selective cyclic guanosine monophosphate (cGMP)‐dependent phosphodiesterase type 5 (PDE‐5) inhibitor, induces vasodilation by enhancing the smooth muscle‐relaxant effects of nitric oxide. cGMP has also been reported to be involved in the regulation of aqueous humour production and outflow, processes that influence aqueous humour dynamics and IOP (Johnston 2005; Ozdegirmenci et al. 2011; Orejana et al. 2012). Sildenafil treatment has resulted in increased neuronal survival, and even neurogenesis, in models of brain, spinal cord and cerebral injury (Serarslan et al. 2010; Ozdegirmenci et al. 2011;).

Sildenafil has powerful cardioprotective effects and could reduce apoptosis and necrosis in cardiac tissues after I/R injury (Salloum et al. 2008). Ebrahimi et al. (2009) observed that SC reduced diabetes‐induced cardiac cell apoptosis at the end of the first and second weeks, when it reached its peak value.

Because of its vasodilative effect (Pache et al. 2002), SC could theoretically have a neuroprotective effect in the retina and RGC layer. We propose that treatment with SC may be protective in a I/R injury model, thereby offering a novel approach for therapeutic intervention in the pathogenesis of retinal I/R injury neuropathy. This study aimed to better understand the mechanism of retinal cell death by apoptosis in retinal I/R induced in rats and to determine whether SC treatment can prevent RGC layer apoptosis.

Methods

Animals

Male Lewis lineage rats (n = 36), 6 months of age with an average weight of 400 g (SD 350–450 g), were used. The animals were clinically and ophthalmically healthy and housed in a ventilated environment in clean and sanitized cages with access to food and water ad libitum, under a 12‐h light: 12‐h dark cycle and controlled temperature. The research was performed according to the guidelines of the Association for Research in Vision and Ophthalmology – ARVO.

Ethical approval statement

All procedures performed in the present study involving animals were in accordance with the ethical standards of the Ethics Committee on Animal Use (CEUA), Universidade Estadual Paulista (UNESP), Botucatu Medical School (Protocol No. 1086‐2014).

Groups and experimental protocols

The animals were divided into five groups: a control group (CG) consisting of six animals without any intervention (standard group), and other four groups with induced retinal I/R injury, treated with SC for 7 days (SG 7, n = 10) and 21 days (SG=21, n = 10) and placebo group, treated with saline for 7 days (PG 7, n = 5), for 21 days (PG 21, n = 5) after I/R injury.

Ischaemia–reperfusion was induced by paracentesis of the anterior chamber in one random eye using a 30‐G needle, coupled with a bag of 0.9% saline, positioned at a height of 150 cm above the eye (Produit‐Zengaffinen et al. 2009; Schmid et al. 2014). The needle was kept in the eye for 60 min, and IOP was measured using a rebound tonometer (TonoVet^® Icare, Vantaa, Finland) before, 3, 30 and 57 min after induction (Table 1). Topical proparacaine hydrochloride (0.5%), and a combination of intramuscular ketamine (75 mg/kg) and xylazine (10 mg/kg) were used for anaesthesia. Afterwards, the animals were treated with dipyrone (200 mg/kg) and morphine (1 mg/kg) by hypodermic injection as analgesics. Sildenafil citrate was prepared in a vehicle of aqueous solution and was diluted to the required strength of 3 mg/ml and was administered at a dose of 1 mg/kg by gavage SID (between 8 h 00 and 8 h 30 am, on the day of I/R injury induction and during every day until euthanasia, while PG were given 0.9% saline solution, with equal criteria). All animals (SG and PG) were euthanized by intramuscular injection with a combination of ketamine (75 mg/kg) and xylazine (10 mg/kg) and a subsequent intracardiac injection of sodium pentobarbital (120 mg/kg) 1 h after the last treatment.The CG was euthanized after clinical and ophthalmological evaluation, using the same method as SG and PG.

Table 1.

Mean and standard deviation (SD) of intraocular pressure (iop) from the ischaemia/reperfusion retinal injury, measured with rebound tonometer (mmHg)

Tonovet (IOP mmHg)
	‘Normal’ IOP	3 min	30 min	57 min
SG7	7.4 ± 2.2	39.8 ± 6.2	43.9 ± 3.6	42.1 ± 9.4
PG7	7 ± 2.1	34.4 ± 3.7	35.6 ± 6.7	38.2 ± 3.5
SG21	7.6 ± 2	37.9 ± 5.5	40.3 ± 8	44 ± 6.4
PG21	7.8 ± 1.8	33.4 ± 6.3	36.6 ± 7.2	37.4 ± 4.3

SG7, sildenafil‐treated group – 7 days; PG7, placebo group – 7 days; SG21, sildenafil‐treated group – 21 days; PG21, placebo group – 21 days.

Collection, storage and preparation of samples

After euthanasia, both eyes (together with their optic nerves) were excised and fixed in 10% buffered formaldehyde solution (pH 7.4) for 48 h. The eyes were then washed in running water and kept in 70% ethanol. The entire eye was embedded in paraffin, and the paraffin block was cut transversely in consecutive 3 μm thick sections in an automatic microtome (RM2255) until it reached the centre of the optic nerve. Slides were stained with haematoxylin and eosin (HE) for morphometric and histopathological evaluation or were prepared for immunohistochemistry in charged slides.

Morphometry and histology

The morphometric analysis was performed using an image analysis program QWin v3.013 (Leica) and a conventional optical microscope (Leica DMR) equipped with a digital camera (Leica DFC500). Five images (fields) were obtained from each slide with a 20 ×  objective, covering the full extent of the retina, from ora serrata to ora serrata, including the central portion of the optic nerve. In each image, we evaluated the total thickness of the retina and each of its layers (the ganglion cell, inner plexiform, inner nuclear, outer plexiform and the outer nuclear), and counted the cells from the retinal ganglion cell layer. The mean thickness of each layer and the mean number of cells in the ganglionar layer in each treatment group were compared with the control group.

The histopathological evaluation was performed according to Johnson et al. (2000) and Schmid et al. (2014), including inflammatory changes (oedema and inflammatory infiltrate), cellular changes (necrosis/apoptosis/chromatin) and tissue changes (atrophy). The analysis was descriptive to treatment group.

Immunohistochemistry

We evaluated the tissue distribution and intensity of Caspase‐7, Caspase‐6, Caspase‐9, tumour necrosis factor receptor 2 (Tnf‐r2), Fas ligand (Fas‐l), Bcl‐2 and Bax proteins, focusing on their expression in cells from the ganglionar layer. Table 2 shows the primary antibodies used, their dilution and the antigen retrieval method. Peroxidase blockage was performed, as well as a protein block. A polymer system was used as the secondary antibody conjugated with peroxidase (Histofine Simple Stain MAX PO, Nichirei, Tokyo, Japan) or LSAB (LSAB System HRP K0690, Dako, USA). Staining was performed with 3,3‐diaminobenzidine tetrachloride (DAB, Liquid DAB Cromogen®, Carpinteria, USA). The sections were counterstained with Harris’ haematoxylin and dehydrated in increasing alcohol concentrations and xylene. The slides were mounted with Entellan and overlaid with a coverslip.

Table 2.

Primary and secondary antibodies used, their dilutions, antigen retrieval methods and incubation period

Primary antibody	Manufacturer	Dilution	Antigen retrieval method	Incubation period	Secondary antibody
Caspase‐7	Santa Cruz Biotechnology N‐19/SC‐492	1:200	Citrate pH 6.0/WB	2 h/27°C	LSAB
Caspase‐9	Epotomics Ab32539	1:100	Citrate pH 6.0/WB	2 h/4°C	Histofine
Tnf‐r2	Santa Cruz Biotechnology SC‐1074	1:400	Citrate pH6.0/Pascal	2 h/27°C	LSAB
Fas‐l	Santa Cruz Biotechnology (N‐20): SC‐184	1:2000	Citrate pH6.0/WB	2 h/27°C	Histofine
Bax	Santa Cruz Biotechnology P‐19/SC‐526	1:300	Citrate pH 6.0/Pascal	2 h/27°C	Histofine
Caspase‐6	Santa Cruz Biotechnology SC‐1230	1:100	Citrate pH 6.0/Pascal	2 h/27°C	Histofine
Bcl‐2	Santa Cruz Biotechnology (C19):SC‐649	1:2000	Citrate pH 6.0/Pascal	2 h/27°C	Histofine

For the negative controls, primary antibodies were replaced with Tris buffer, and for the positive controls, rat tissues expressing the proteins were chosen according to the manufacturer's directions for each primary antibody and recommendations at: www.proteinatlas.com.

We used a semi‐quantitative index, with scores of 1–4, to discriminate the percentage of positive cells in the retinal ganglion cell layer, with: 1 = 1–25% positive cells, 2 = 26–50% positive cells, 3 = 51–75% positive cells and 4 = >76% positive cells. Staining intensity was classified as mild (i), moderate (ii) or strong (iii) for each antibody. The distribution and intensity of each antibody was compared to its scores in the CG (e.g. if the CG score for Bax was >3, then scores <3 in the experimental groups were considered to reflect down regulated protein expression).

Quantitative (q) RT‐PCR

We evaluated Caspase‐7, Caspase‐6, Caspase‐9, Tnf‐r2, Fas‐l, Bcl‐2 and Bax expression. mRNA was extracted from ten 5‐μm sections cut from the paraffin blocks using the RecoverAll™ Total Nucleic Acid Kit (Ambion, Life Technologies, Waltham, MA, USA) according to the manufacturer's instructions. mRNA concentration was assessed by spectrophotometer (NanoDrop™, ND‐8000, Thermo Scientific, Waltham, MA, USA), and RNA integrity was evaluated by the Bioanalyzer 2100 and Agilent RNA 6000 Nano kit (Agilent Technologies, CA, USA).

cDNA synthesis was performed using 2 μg of total RNA in a 20 μl reaction volume and an oligo dT primer according to the manufacturer's instructions (Affinity Script; Stratagene, Agilent Technologies SA, Morges, Switzerland). The equivalent of 12.5 ng of original total RNA was used for qPCR amplification using the 2 × brilliant SYBR Green QPCR Master Mix (Stratagene, Agilent Technologies). Forward and reverse primers (1 μM), designed to span an intron of the target gene, were used. qRT‐PCR was performed with the following cycling conditions: 40 cycles of denaturation at 95°C for 30 s, annealing at 59°C for 30 s and extension at 72°C for 30 s. Samples were amplified in triplicate, and the data were normalized to the expression of β‐actin and B2 m by geometric mean.

Statistics

For the histological changes, descriptive analyses were employed. For the thickness values of the total retina and its layers and for the total number of cells count in the retinal ganglion cell layer, the results were evaluated for normality by applying the Kolmogorov–Smirnov test. anova and t‐test were used to evaluate the RT‐qPCR results. We used anova test to evaluate the statistical difference between all groups (CG vs. SG 7 days vs. PG 7 days and CG vs. SG 21 days vs. PG 21 days). The t‐test was used to evaluate statistical difference between two variables (e.g. CG vs. SG 7 days or SG 7 days vs. PG 7 days). The Fisher's exact test was used to evaluate the immunohistochemical results, respectively, with statistical significance set at P < 0.05.

Results

Morphometry and histopathology

In the morphometry analysis, the total retina thickness was significantly decreased in the experimental SG and PG eyes for both the 7‐ and 21‐day treatments comparing with CG eyes. There was a difference in the PG and SG 21‐day retina thicknesses, the latter value being closer to CG than the former (Table 3).

Table 3.

Mean values (±SD) of retinal layers thickness (µm) and the count of cells from the retinal ganglion layer (RGC layer) from normal rats (CG) and rats with induced ischemia/reperfusion retinal injury, treated (SG) or not (PG) with Sildenafil for 7 and 21 days

	7‐day treatment				21‐day treatment
	PG	SG	CG	P	PG	SG	PG	P
Total retinal thickness (μm)	163 ± 26.92a	168 ± 30.55a	231 ± 25	<0.05	149 ± 20a , b	174 ± 38a , b	231 ± 25	<0.05
Ganglion cell layer (μm)	6.6 ± 1.1a , b	10 ± 2.9a , b	16 ± 2.0	<0.05	7.0 ± 1.8a , b	9.2 ± 2.5a , b	16 ± 2.0	<0.05
Inner plexiform layer (μm)	26 ± 8.6a , b	38 ± 12a , b	55 ± 6.7	<0.05	33 ± 13a , b	43 ± 11a , b	55 ± 6.7	<0.05
Inner nuclear layer (μm)	25 ± 4.6a	26 ± 6.0a	35 ± 5.8	<0.05	23 ± 3.7a	24 ± 5.5a	35 ± 5.8	<0.05
Outer plexiform layer (μm)	6.6 ± 1.1a , b	9.1 ± 1.5b	9.8 ± 1.6	<0.05	6.7 ± 1.0a , b	9.2 ± 2.3b	9.8 ± 1.6	<0.05
Outer nuclear layer (μm)	47 ± 11a , b	42 ± 8.8a , b	63 ± 8.4	<0.05	37 ± 7.2a	40 ± 13a	63 ± 8.4	<0.05
RGC layer (no.)	123 ± 23a , b	197 ± 32b	189 ± 10	<0.05	110 ± 38a , b	160 ± 33b	189 ± 10	<0.05

PG, placebo group; SG, sildenafil‐treated group; CG, control group.

P value represents statistical significance of differences between groups by anova test.

^aStatistically significant difference compared to the control (t‐test).

^bStatistically significant difference relative to the treatment groups (t‐test).

Both SG and PG showed decreased thickness compared to the CG, and this applied to both the 7‐ and 21‐day treatments (Table 3). Analysing each retinal layer separately, the thickness of the outer plexiform layer and the number of cells in the RGC layer did not differ statistically between the SG and CG at either 7 or 21 days of treatment. However, both treatment periods, there was a decrease in thickness with significant differences between the PG and SG in the ganglion cell layer, the inner plexiform layer thickness and the number of cells in the RGC layer (Table 3). The SG results were closer to the CG than the PG results.

In the histopathological analysis, the changes were of cellular origin. There was a higher prevalence of lesions in the PG than in CG and SG, mostly related to changes in the cells from the retinal ganglion cell layer (Figures 1, 2, 3 and Table 4).

Figure 1.

Retinas and optic nerves from rats with induced ischaemia–reperfusion; histopathological changes. (a and f) control group. (b) Retinal decrease thickness with the loss of the cells from the RGC layer, mild oedema and atrophy of the inner plexiform layer and the inner nuclear layer. (c) Mild retinal atrophy with discrete loss of the both plexiform and retinal ganglion cell layers. (d) Moderate retinal decrease thickness with the loss of RGC layer and moderate atrophy of the internal plexiform layer. (e) Discrete retinal atrophy with the mild loss of RGCL and the mild atrophy of inner plexiform layer and the inner nuclear layer. (g) Discrete atrophy of optic nerve. (h) Mild intercellular oedema. (i) Extracellular oedema and marked atrophy in the optic nerve. (j) Mild atrophy in the optic nerve. Retina: HE, 40 × ; Optic nerve: HE, 20 × , scale 50 bar. [Colour figure can be viewed at wileyonlinelibrary.com].

Figure 2.

Retinas from rats with induced ischaemia–reperfusion injury, treated with sildenafil for 7 days. (a) Inner nuclear layer with mild‐to‐moderate extracellular oedema (white asterisk) and mild retinal atrophy. (b) Loss of cells from the retinal ganglion cell layer and moderate retinal atrophy. (c) 7‐day sildenafil treatment shows mild extracellular oedema. (d) Ganglion cell layer: nuclear pyknosis (black asterisk), karyolysis (arrowhead) and chromatolysis (arrow). (e) Ganglion cell layer: nuclear pyknosis (black asterisk) and chromatolysis (arrow). a, b, c: HE, 40 ×; d, e: HE, 40 × 1.6, scale 50 bar. [Colour figure can be viewed at wileyonlinelibrary.com].

Figure 3.

Retinas from rats with induced ischaemia–reperfusion injury treated with sildenafil or not (placebo), for 21 days. (a) Placebo, 21 days. Nuclear pyknosis (black asterisk), karyolysis (arrowhead) and chromatolysis (arrow). (b) Sildenafil, 21 days. Nuclear pyknosis (black asterisk) and chromatolysis (arrow). HE, 40 × , scale 50 bar. [Colour figure can be viewed at wileyonlinelibrary.com].

Table 4.

Histopathological changes in the retina from rats with induced ischaemic/reperfusion injury model, treated (SG) or not treated (PG) with sildenafil

Histopathological changes
	Vascular				Inflammatory				Cellular				Tecidual
	0	1	2	3	0	1	2	3	0	1	2	3	0	1	2	3
PG‐7d	–	3/5	2/5	–	4/5	1/5	–	–	–	3/5	2/5	–	–	1/5	4/5	–
SG‐7d	6/10	4/10	–	–	6/10	4/10	–	–	–	8/10	2/10	–	–	8/10	2/10	–
PG‐21d	1/5	4/5	–	–	5/5	–	–	–	–	2/5	3/5	–	–	2/5	3/5	–
SG‐21d	6/10	4/10	–	–	8/10	2/10	–	–	–	9/10	1/10	–	–	9/10	1/10	–

PG, placebo group, SG, sildenafil group; 7, seven‐day treatment/21, twenty‐one‐day treatment.

0, no change; 1, discrete; 2, moderate; 3, severe.

Immunohistochemistry

The immunohistochemistry results are presented in Tables 5 and 6. Statistical analysis of the results showed significant differences for the Caspase‐7 (P < 0.0170) score in SG 7 and PG 7 compared to the control eyes.

Table 5.

Distribution scoresa of immunohistochemical markers by group with respect to cells positive for Tnf‐r2, Fas‐l, Bax, Bcl‐2, Caspase‐9, Caspase‐6 and Caspase‐7 in the RGC layer

Immunolabelling score
Groups	Treatment	Tnf‐r2			Fas‐l			Bax			Bcl‐2			Caspase‐9			Caspase‐6			Caspase‐7
		Number of eyes			Number of eyes			Number of eyes			Number of eyes			Number of eyes			Number of eyes			Number of eyes
	Score	≤3	4	P	≤2	>3	P	≤2	>3	P	≤3	4	P	≤3	4	P	≤2	>3	P	≤3	4	P
7 days	PG	2	3	0.25	1	4	0.08	1	4	1.00	4	1	0.08	1	4	0.08	3	2	0.24	1	4	0.01b
	SG	8	2		8	2		1	9		2	8		8	2		9	1		9	1
21 days	PG	2	3	0.25	2	3	0.25	1	4	1.00	4	1	0.08	2	3	0.32	1	4	0.08	2	3	0.25
	SG	8	2		8	2		1	9		2	8		7	3		8	2		8	2
Control		6	0		6	0		0	6		6	0		6	0		6	0		6	0

^a1 =  <25% positive cells, 2 = 26–50% positive cells, 3 = 51–75% positive cells, 4 =  >76% positive cells.

^bSignificantly different from the treatment groups.

Table 6.

Immunohistochemical staining intensity scorea by group for Tnf‐r2, Fas‐l, Bax, Bcl‐2, Caspase‐9, Caspase‐6 and Caspase‐7 in the RGC layer

Immunostaining intensity
Groups	Treatment	Tnf‐r2			Fas‐l			Bax			Bcl‐2			Caspase‐9			Caspase‐6			Caspase‐7
		Number of eyes			Number of eyes			Number of eyes			Number of eyes			Number of eyes			Number of eyes			Number of eyes
	Score	D+M	I	P	D	M+I	P	D+M	I	P	D+M	I	P	D	M+I	P	D	M+I	P	D	M+I	P
7 Days	Placebo	3	2	0.24	1	4	0.08	1	4	0.08	4	1	0.60	2	3	0.25	3	2	0.24	2	3	0.07
	Sildenafil	9	1		8	2		8	2		6	4		8	2		9	1		9	1
21 Days	Placebo	2	3	0.25	2	3	0.25	1	4	0.11	4	1	0.58	2	3	0.25	4	1	1.00	3	2	0.25
	Sildenafil	8	2		8	2		7	3		5	5		8	2		9	1		8	2
Control		6	0		6	0		6	0		6	0		6	0		6	0		6	0

^aMarker intensity score: D (discrete), M (moderate) and I (severe).

Although there was no statistically significant difference in Tnf‐r2 expression between PG and SG at either 7 or 21 days, there were more animals in the SG with lower Tnf‐r2 scores than in the PG, the latter being more similar to the CG (Figure 4; Tables 5, 6 and Table S1). The same pattern was observed for Fas‐l, Caspase‐6 and Caspase‐9. The same pattern was observed for Bax and Bcl‐2, with higher numbers of animals with scores similar to CG in SG than in the PG, both at 7 and 21 days of treatment.

Figure 4.

Retinas from rats with induced ischaemia–reperfusion injury, illustrating the immunostaining. Note the staining in the cytoplasm from the RGC layer (40x, scale bar 50). (a1–a7) Control group (CG). b1 and d1 have more intensely stained cells and note the increased staining intensity of the placebo group at 21 days (PG 21). c1 and e1 are similar to a1. Fas‐l shows a tendency towards more stained cells in PG 7 and PG 21 (b2 and d2) compared to the sildenafil group (SG) (c2 and e2) and CG (a2). Bax has a tendency towards discrete‐to‐moderate staining intensity in the SG (c3 and e3) compared to the PG (b3 and d3) with a tendency towards high intensity. Bcl‐2 expression is higher in the SG (c4 and e4) than in the PG (b4 and d4) and CG (a4). Caspase‐9 stains more intensely in the PG 7 and PG 21 and displays the increasing marking intensity (b5 and d5), but c5 and e5 are similar to the CG (a5). Caspase‐6 staining is found in a higher number of cells in the PG 21 (d6) compared to the SG (e6) and CG (a6), and there is no difference in the staining intensity in any of the groups compared to the CG. Caspase‐7 staining is found in a higher number of cells in the PG (b7 and d7) compared to the SG (c7 and e7) and CG (a7). [Colour figure can be viewed at wileyonlinelibrary.com].

qRT‐PCR

Results of the qRT‐PCR are presented in Figure 5 (Table S2). Statistical analysis showed significant differences in the expression of most genes studied in SG and PG vs. control eyes. However, for Bax and Caspase‐6 expression, there was no significant difference between the SG and CG in 21‐day treatment. Caspase‐9 transcript was not observed in the CG because it was only amplified in later cycles. Tnf‐r2, Fas‐l, Bax, Bcl‐2, Caspase‐9, Caspase‐6 and Caspase‐7 transcript levels were higher in the PG (7 and 21 days) than in the CG and SG. SG transcript levels for these genes were more similar to the CG transcript levels than to the CG levels, showing the protective action of SC against apoptosis in the retina from rats with induced I/R injury model.

Figure 5.

qRT‐PCR‐ Tnf‐r2, Fas‐l, Bax, Bcl‐2, Caspase‐9, Caspase‐6 and Caspase‐7 expression^a. Tnf‐r2 – There was a significant difference in gene expression between every group. Among the groups, with higher RQs in placebo groups (PG). FAS‐L – There was a significant difference in the treatment groups compared with CG, with down RQs in SG. Bax – There was a higher gene expression in PG than SG with significant difference for RQs between groups, without significant difference between 21 SG and CG. Bcl‐2 – Gene expression is down in the both SG than in the PG. Caspase‐9 – There was more RQs in the PG than SG, with difference significant between PG comparing SG. Note prevalent downregulation of SG and, more gene expression in every PG. The transcript level in control group was below the lower limit of quantification. Caspase‐6, Caspase‐7 – There was more RQs in the PG than SG, with difference significant between PG comparing CG. Note prevalent downregulation of SG and, more gene expression in every PG. ^aa, b, c, d: Significantly different from control. e, f: Significantly different from the treatment groups. * P between 0,1 and 0,05. ** P between 0,1 and 0,0001. *** P low than 0,0001.

Discussion

A number of in vivo and ex vivo mammalian models have been developed to study retinal ischaemia. The ability to extrapolate data from an animal model to the clinical situation requires a model that closely resembles human retinal ischaemia. Small laboratory animals such as rats, rabbits and guinea pigs are commonly used, but all have a different vascular supply. Nevertheless, the rat model is most widely used as a comparison to humans (Lam et al. 1999; Osborne et al. 2004; Kaur et al. 2008; Produit‐Zengaffinen et al. 2009).

Our method of acute IOP elevation, known as ‘I/R’, led to the stress response and resulted, by morphometric and morphology analysis, in the decrease in thickness of all retinal layers in PG and SG 7‐ and 21‐day treatment, comparing with CG. In this study, we examined the potential involvement of apoptosis in the loss of retinal tissue after retinal I/R insult. The I/R induced a significant loss in the number of cells from the RGC layer in PG 7 and 21 days. This finding is supported by the increase in apoptotic pathway in this group (Bax, Caspase 9, 6 and 7). This result was similar to those reported in the literature, especially regarding retinal layer lesions and a decrease in the total retinal thickness and in the RGC layer count (Katai & Yoshimura 1999; Produit‐Zengaffinen et al. 2009; Giedrius et al. 2012; Dvoriantchikova et al. 2014; Schmid et al. 2014).

The most prevalent histopathological findings were intercellular oedema in the inner and outer plexiform layers in PG 7 days, chromatolysis, pyknosis, karyolysis of the cells from the RGC layer, apoptosis of those cells and the decrease in retinal layers in both treatment groups. This corroborates with the course of I/R development (Lam et al. 1999; Dvoriantchikova et al. 2014; Schmid et al. 2014). An interesting finding, related to the beneficial results of SC treatment, was that there were less histopathological changes in SG 7 and SG 21 days than in their PG counterparts.

The loss of RGCs occurs in some ophthalmic conditions such as glaucoma and diabetes. RGC death has been reported to occur in many experimental studies using different methods to induce retinal ischaemia (Quigley 2007; Kaur et al. 2008; Stuart et al. 2009), as we also observed in our I/R model. The neurodegeneration in I/R results in apoptosis of RGCs, and in the inner nuclear layer and occasionally in the outer nuclear layer, mainly generated by the activation of the caspase apoptotic cascade, and an additional decrease in the thickness of outer and inner plexiform layers (Lam et al. 1999; Reichstein et al. 2007). We also found a reduced number of cells in the RGC layer, by necrosis (picnotic cells) and apoptosis. The inner retinal layer from PG was thinner than the control group, showing the neurodegeneration reported by other authors (Lam et al. 1999; Reichstein et al. 2007). The outer nuclear layer was also reduced in thickness.

The lack of statistically significant differences in the immunohistochemical analysis could be due to the low number of samples, but there was a tendency for the animals treated with SC to have scores that were more similar to the CG than those of the PG. The extrinsic apoptotic pathway is initiated by the Fas ligand (Fas‐l) or tumour necrosis factor (Tnf) linked to death receptors (Fas, Tnf‐r1 and Tnf‐r2) located on the cytoplasmic membrane. There is strong evidence for this pathway's involvement in the apoptosis of cells from the RGC layer in I/R, with elevated IOP and the upregulation of Fas‐l and Tnf‐r2 causing the loss of optic nerve, oligodendrocytes and RGCs (Tezel 2008; Qu et al. 2010; Park et al. 2014). Sildenafil treatment prevented the higher Tnf‐r2 and Fas‐l protein and gene expression in retinal I/R induced injury with more animals having scores and RQs similar to those of CG, probably due to SC treatment's prevention of cell apoptosis.

There are also many intrinsic factors that can evoke apoptosis without the involvement of death receptors, such as neurotrophin deprivation and oxidative stress (Kaur et al. 2008; Qu et al. 2010; Park et al. 2014). Although the initial steps of the intrinsic pathways are not clear, they all eventually converge to the mitochondria and release of cytochrome C, regulated by the Bcl‐2 family members in the mitochondrial outer membrane (Kaur et al. 2008; Qu et al. 2010; Park et al. 2014). Bcl‐2 RQs from SG 7‐ and 21‐day treatment were closer to control than PG, showing an anti‐apoptotic role. Although there was no significant statistical difference in Bcl2 expression (p = 0.08) between PG and SG, there were a greater number of samples with scores similar to CG in SG than in PG in 7‐ and 21‐day treatment.

In our study, the distribution of the pro‐apoptotic protein Bax was similar in the SG and PG comparing with CG, but when we considered the qRT‐PCR expression, there were more animals with higher RQs in the PG than in SG. Although immunohistochemistry is not the best method to quantify a protein, all the reactions were conducted simultaneously; then, the higher intensity of Bax in the PG could indicate a higher concentration of this protein in the non‐treated animals.

Blocking pro‐apoptotic factors and promoting anti‐apoptotic factors has a neuroprotective function in RGC layer injury (Qu et al. 2010; Park et al. 2014). Knockout mice lacking functional Bax gene exhibit several supernumerary populations of neurons, including RGCs, indicating the importance of Bax in the neuronal regulation of programmed cell death during development (Kirkland et al. 2002). Our results indicate Bax down expression in SG, probably interfering with retinal cell death by apoptosis, as observed in the retinal total thickness, more similar to CG in SG 21 days.

In our study, Caspase 9, 6 and 7 had higher expression levels in the PG 7 and PG 21 days, as validated by qRT‐PCR, indicating a probable neuroprotective action in SG. Moreover, our results in the PG showed greater activation of apoptotic pathways, while the SG presented higher number of cells in the RGC layer, in agreement with the literature (Lam et al. 1999; Pache et al. 2002; Luchs & Pantaleão 2010).

In general, gene overexpression has been detected in many experimental glaucoma models associated with RGC apoptosis. Caspase‐9 mRNA and procaspase‐9 are upregulated in axotomized and hypertensive rat retinas (Quigley et al. 1995; Lu et al. 2003; Nie et al. 2009; Qu et al. 2010; Park et al. 2014). Caspase‐9 is only detected in damaged retinas (Qu et al. 2010), which explains why we did not observe its transcripts in the CG. The extrinsic and intrinsic apoptotic pathways are not entirely independent. There is crosstalk, and together they amplify and accelerate the apoptotic cascades. As described, activated Caspase‐9 can activate the downstream extrinsic pathway. At the same time, a pro‐apoptotic member of the Bcl‐2 family might be activated (Quigley et al. 1995; Lu et al. 2003; Nie et al. 2009; Qu et al. 2010; Almasieh et al. 2012,). Our results showed deregulation between pro‐ and anti‐apoptotic genes resulting in a more intense pro‐apoptotic signalling cascade in the PG.

Based on the chronology of animals treated, both treatment groups showed beneficial effects on daily use of SC in every analysis. However, based only on the morphometric analysis, the SG 21 days showed more protection of all retina thickness and RGC layer. Ezra‐Elia et al. (2016) revealed a tendency of less necrosis in retinas of sildenafil‐treated animals, suggesting that SC prevents apoptosis in treated animals. Thus, there is a trend towards a positive effect of SC that could implicate the possible anti‐apoptotic effect, demonstrated by decreased apoptosis and increased anti‐apoptosis expression. However, there are some conflicting results: other studies have reported visual adverse effects of SC in human patients or laboratory animals, even after long periods of daily drug use (Santaella & Fraunfelder 2007).

Based on these results, we suggest that the mechanism of retinal cell death was mainly via apoptosis, as there was an increase in pro‐apoptotic factors and decrease in the anti‐apoptotic. The treatment with SC in this retinal I/R injury model can help in preventing apoptosis of retinal cells, as shown by histology, morphometry, and analysis of pro‐ and anti‐apoptotic gene and protein expression. Further studies should be made to confirm the efficacy of SC in the treatment of glaucoma, but it seems to be a good potential therapeutic option.

Conflict of interest

We declare that there is no conflict of interest.

Funding source

This work was supported by a grant from Fapesp (Fundacao de Amparo a Pesquisa do Estado de Sao Paulo), Grant number FAPESP (No 2011/51269‐4; 2013/01926‐4).

Supporting information

Table S1. Relative quantification (RQ) of the target genes, immunohistochemistry distribution and intensity scores for individual results from control group.

Table S2. Relative quantification (RQ) of the target genes, immunohistochemistry distribution and intensity scores for individual results from placebo group – 7 days.

Table S3. Relative quantification (RQ) of the target genes, immunohistochemistry distribution and intensity scores for individual results from sildenafil group – 7 days.

Table S4. Relative quantification (RQ) of the target genes, immunohistochemistry distribution and intensity scores for individual results from placebo group – 21 days.

Table S5. Relative quantification (RQ) of the target genes, immunohistochemistry distribution and intensity scores for individual results from sildenafil group – 21 days.