Gap junction proteins in the light-damaged albino rat

Animals

All experimental procedures described in this study were approved by the University of Auckland Animal Ethics Committee and comply with the use of animals in eye research. Adult Sprague-Dawley (SD) rats (200–250 g, male or female) were used in this study. The SD rat is an albino strain and ensures the precise and direct effect of light on photoreceptor degeneration [19,20]. Animals were sourced from the Vernon Jansen Unit at the University of Auckland. Light in the animal breeding unit was provided by white fluorescence light (Philips Master TLD 18W/965; Koninklijke Philips Electronics N.V., Shanghai, China). The light source covers broad band fluorescence, from 380 to 760 nm, and the average intensity at the top of the cage is 120 W/m². The animals used in these experiments were housed in normal cyclic light conditions [12 h light (174 lux):12 h dark (<62 lux)] for the duration of the study, except where otherwise stated.

Light damage procedure

Intense light was used to damage the retina in albino rats. Light damage experiments started consistently around 9:00 AM. The luminance was 2,700 lux, created by placing fluorescent light lamps (Philips Master TLD 18W/965) directly above the rat cages. The light source covered broad band fluorescence, from 380 to 760 nm, with no heat generated by the light source. This broad band fluorescence light has been used in a previous study [28]. Continuous exposure was achieved by removing hiding places in the cages. Animals were able to move freely in the cage and had free access to food and water. Non-light-damaged animals were used in the control group. All experimental protocols consisted of 24 h light exposure with variable recovery periods (0 h, 6 h, 24 h, 48 h, or 7 days) in normal 12 h light (174 lux):12 h dark (<62 lux) conditions.

Tissue collection

The animals were deeply anesthetized with intraperitoneal injection (i.p.) with a combination of ketamine (75 mg/kg) and domitor (0.5 mg/kg). The rat’s chest cavity was opened, and a perfusion needle was inserted into the left ventricle. An outlet was made in the right atrium to enable the flush solution and fixative to circulate through the vasculature. The animals were perfused with saline for 2–3 min. For the tissue collected for immunohistochemical procedures, the animals were further perfused with 4% paraformaldehyde (PFA) in a 0.1 M phosphate buffer (PB) for 2–3 min. After transcardial perfusion, the eyes were dissected from the orbit. The posterior eyecup was flattened by making several radial cuts in the margin. A strip of the central retina across the temporal/nasal axis 1 mm above the optic nerve was dissected. In an intense light damage paradigm, the dorsal arc of the retina is affected the most; the molecular and morphological reasons were reviewed by Organisciak and Vaughan [31]. In this study, the central part of the dorsal retina was used. With the sclera side up, the tissue was gently mounted on filter paper (0.8 µm pore size, Gelman Sciences, Ann Arbor, MI) to keep the tissue flat for cryosectioning. The tissue was then further fixed in 4% PFA for 30 min at room temperature and washed in 0.1 M PBS (1X; 154 mM NaCl, 72 mM Na₂HPO₄, 19 mM NaH₂PO₄, pH 7.4). The tissue was cryoprotected using graded sucrose solutions 10% and 20% for 30 min each at room temperature and 30% overnight at 4 °C, and frozen in Tissue-Tek optimum cutting temperature (OCT) compound (Sakura Finetek, Torrance, CA) at −20 °C. The tissue was then cryosectioned in the vertical plane of the retina (16 µm sections) using a cryostat (Leica CM3050S, Heidelberg, Germany). The sections were collected on SuperFrost Plus slides (Labserv, Auckland, New Zealand) and stored at −20 °C until the start of immunolabeling or the staining procedures.

Terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling staining (TUNEL) and cell death analysis

Cell death was investigated in light-damaged and non-light-damaged retina. The TUNEL technique was used to detect cell death. Tissues were processed using a commercial In Situ Cell Death Detection Kit, Fluorescein (Roche Applied Science, Mannheim, Germany). Tissue sections were washed twice with 0.1 M PBS before being permeabilized with 0.1% Triton X-100 and 0.1% trisodium citrate in 0.1 M PBS for 5 min on ice. The sections were washed and incubated with the TUNEL reaction mixture (TdT enzyme solution and fluorescein nucleotide label solution at 1:17 dilution) for 20 min at 37 °C in a dark humid chamber. Negative controls were incubated with the label solution only. The sections were rinsed several times in 0.1 M PBS, and then mounted in Vectashield HardSet Mounting Medium with 4’,6-diamidino-2-phenylindole dihydrochloride (DAPI; Vector Laboratories, Burlingame, CA). The stained sections were visualized using confocal microscopy.

Immunohistochemistry

Frozen tissue sections were air-dried at room temperature for 10 min and then washed in 0.1 M PBS. The tissue sections were encircled with a PAP pen (Life Technologies, Camarillo, CA) after the excess moisture around the sections was wiped off. Non-specific binding sites in the tissue were blocked with 6% normal goat serum (NDS from Sigma-Aldrich, Saint Louis, MO) or normal donkey serum (NGS from Invitrogen, Frederick, MD), 1% bovine serum albumin (BSA), and 0.5% Triton X-100 in 0.1 M PBS for 1 h at room temperature. The primary antibodies were diluted in antibody solution containing 3% normal goat serum or normal donkey serum, 1% BSA, and 0.5% Triton X-100 in 0.1 M PBS (primary antibody details; Table 1) and were applied carefully to cover the tissue sections. Tissue sections were incubated with the primary antibody overnight at room temperature. The negative control for each antibody was conducted by incubating sections with antibody solution only following the same protocol as if a primary antibody was applied. After the overnight incubation, the slides were washed three times for 5 min each in 0.1 M PBS and once for 15 min in 0.1 M PBS to remove excess primary antibodies. The secondary antibodies, goat anti-rabbit, goat anti-mouse, or donkey anti-guinea pig, conjugated to either Alexa 488 or Alexa 594 (Life Technologies, Carlsbad, CA), were diluted 1:500 in antibody solution and applied to the tissue for 2 h at room temperature. Slides were washed thoroughly with 0.1 M PBS to remove all excess secondary antibodies. Following the washes, the tissue sections were incubated with DAPI (Sigma) 1 µg/ml in 0.1 M PBS for 15 min to label cell nuclei, followed by another thorough wash with 0.1 M PBS. The slides were then mounted in a non-fluorescence-fading medium (Citifluor, London, UK) and sealed with nail polish to avoid leakage and evaporation.

Imaging and fluorescence quantification

All the images correspond to the superior central retina area. The images were acquired using a high-resolution laser scanning confocal microscope (Olympus FluoView FV1000, Olympus Corporation, Tokyo, Japan) with 405, 473, and/or 559 nm excitation from an argon ion laser. A series of four to eight optical sections with 1 µm interval were collected for each specimen, and image analysis was performed on an image stack intensity projection over the z-axis. All images were acquired with the same gain and brightness. Six retinas obtained from different animals were analyzed for each group, and the representative image for each condition is shown.

Fluorescence quantification was performed using ImageJ 1.45s software (Wayne Rasband, National Institutes of Health, Bethesda, MD). RGB images were converted to binary mode. The background was set on a non-labeled area of the slide. The number of pixels occupied by the marker was measured in a constant 200 µm × 150 µm area of the central-superior retina in at least eight regions per tissue section (n = 6).

Western blotting

The tissues used in the western blot procedure were processed immediately after the animals were deeply anesthetized with a combination of ketamine (75 mg/kg) and domitor (0.5 mg/kg). After transcardial perfusion and dissection of the eye from the orbit, the posterior segment of the eye was isolated. The retina was carefully separated from the choroid and sclera with the use of a dissection microscope. The extracted whole retina tissue was homogenized in a cold, freshly prepared homogenization buffer (10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid/HEPES; 0.25 M sucrose in MilliQ water, pH 7.4) containing 1× protease inhibitor cocktail (Complete Mini Tablet; Roche Diagnostics, Manheim, Germany) using a Teflon homogenizer, according to a protocol for retinal tissue homogenization using a Teflon homogenizer and a 0.9% NaCl solution [32]. Homogenized samples were centrifuged twice at 5000 ×g for 2 min in an Eppendorf centrifuge (5415R; Eppendorf, Hamburg, Germany). The samples were then centrifuged twice (5000 ×g for 2 min at 4 °C) to sediment any unhomogenized tissue, and the supernatant was stored at −20 °C until used in the western blot analysis.

Protein samples were mixed with a loading buffer (10:1) containing 0.1% bromophenol blue, 20% v/v glycerol, and 2% sodium dodecyl sulfate (SDS). Typically, about 9 µg of retinal or choroidal sample protein in a total volume of 20 µl was gently loaded into the leading wells of a 10% Mini-PROTEAN TGX Precast Gel (Bio-Rad Laboratories, Auckland, New Zealand). Proteins were separated according to molecular weight under electrophoresis using a Mini-Protean Tetra Cell (Bio-Rad Laboratories) with an SDS gel running buffer (25 mM Tris base; 19 mM glycine; 0.1% SDS in MilliQ water), which applies a negative charge to each protein in proportion to its molecular weight. A Precision Plus Protein All Blue Standard (Bio-Rad Laboratories) containing proteins with known molecular weights ranging from 10 to 250 kDa was loaded alongside the protein samples in each gel. The separated proteins were transferred from the gel to a polyvinylidene difluoride (PVDF) membrane (Roche Diagnostics) using a semidry transfer system: Trans-Blot Turbo Transfer Starter System (Bio-Rad Laboratories) with a small amount of transfer buffer (25 mM Tris base; 190 mM glycine; 15% methanol in MilliQ water, pH 8.4).

The membrane containing separated and transferred protein samples was incubated with a blocking solution (5% non-fat milk powder; 2% v/v normal goat serum in Tris-buffered saline-tween (TBS-T), containing 20 mM Tris hydrochloride; 137 mM sodium chloride; 0.1% v/v Tween-20, pH 7.2) for 2 h at room temperature. This was followed by incubating the membrane with the primary antibody overnight at 4 °C. The primary antibody (anti-Cx43 antibody, 1:4,000 dilution) was diluted with an antibody solution containing 2% non-fat milk powder and 2% v/v normal goat serum in TBS-T. Following three washes each of 15 min and one wash for 40 min in TBS-T, the membrane was incubated with a horseradish peroxidase-conjugated secondary antibody (donkey anti-rabbit immunoglobulin G, 1:5,000; Amersham Biosciences, Buckinghamshire, UK) for 1 h at room temperature. The membrane was then washed with another series of TBS-T.

The immunoblot was detected using an ECL Plus Western Blot Detection Reagents Kit (Amersham Biosciences). The horseradish peroxidase-conjugated to the secondary antibody catalyzes oxidation of the Lumigen PS-3 Acridan substrate (Blot Detection Reagents Kit, Buckinghamshire, UK), and this reaction produces sustained, high-intensity chemiluminescence with maximum emission at a wavelength of 430 nm, which was subsequently detected on a FujiFilm LAS-4000 imager (GE Healthcare, Buckinghamshire, UK).

A loading control blot was performed following the imaging of the first immunoblot for anti-Cx43. The blot membrane was stripped out from the membrane by incubating it with stripping buffer (2% SDS, 100 mM β-mercaptoethanol, 62.5 mM Tris hydrochloride, pH 6.7) for 30 min at 37 °C. The loading control was then blotted with an anti-tubulin antibody (1:5,000 dilution in antibody solution) on the same membrane.

Quantification of the immunoblot images was conducted using ImageJ 1.45 s software (Wayne Rasband, National Institutes of Health). Histograms indicating the intensity of each band were obtained and converted into numerical values indicating the intensity of the bands with the ImageJ program. The blots for Cx43 were divided by the loading controls for each well, and then were statistically analyzed for the experimental groups.

Statistical analysis

Graphing and statistical analyses were performed using GraphPad Prism 6 (GraphPad Software, San Diego, CA). All data are presented as the mean±standard error of the mean (SEM). The comparison between normal and light-damaged tissue western blots for Cx43 was conducted using a one-way ANOVA followed by Tukey’s test. Unpaired comparison was carried on using the Student t test with alpha = 0.05.

Photoreceptors were undergoing cell death in light damage

TUNEL labeling was performed on the central retinas in the normal control and light-damaged rats to provide an indication of the amount of photoreceptor cell loss (Figure 1). No TUNEL-positive cells were detected in the retinas of the control animals (Figure 1A). Six hours after light damage, there were a large number of TUNEL-positive cells in the outer nuclear layer, suggesting dramatic cell death and tissue damage (Figure 1B).

Expression of connexin 36 and connexin 45 is changed following light damage

The expression of Cx36 and Cx45 was investigated in the control and light-damaged animals using immunohistochemistry. In the control retinas, Cx36 immunoreactivity was low and limited to the inner plexiform layer and the RPE (Figure 2A). Labeling in the outer plexiform layer was not detectable. The high-magnification confocal image of the inner plexiform layer shows that punctate labeling was confined to the inner part of this layer (Figure 2E). At 24 h after light damage, immunoreactivity of Cx36 in the RPE was present but seemed to be redistributed within the cells. Small randomly distributed puncta labeling was seen in the outer plexiform layer. The Cx36 immunoreactivity in the inner plexiform layer appeared to increase (Figure 2B,F), and quantification indicated significant upregulation (Figure 2K; Student t test, p<0.001). At 7 days after light damage, Cx36 immunoreactivity in all layers was decreased to a level visually comparable to that of the control tissue (Figure 2G).

In the control retinas, Cx45 was mostly expressed in the inner plexiform layer distributed in a specific banding pattern. Labeling in the RPE and the outer segment of the photoreceptors was also evident (Figure 2C). In the nerve fiber layer, there was Cx45 expression possibly on vascular smooth muscle cells and astrocytes (Figure 2H, arrows). At 24 h after light damage, there was no evident change of Cx45 labeling in the RPE and outer segment area or in the nerve fiber layer, but interestingly, Cx45 immunoreactivity in the inner plexiform layer was decreased (Figure 2D,I). Quantification indicated a significant decrease in the inner plexiform layer labeling (Figure 2L, Student t test, p<0.001). However, by 7 days after the onset of light damage, the expression of Cx45 in the inner plexiform layer had returned with a pattern and level similar to that of the control, but there was now no expression in the nerve fiber layer (Figure 2J).

Expression of connexin 43 increases following light damage

Expression of the gap junction protein Cx43 in the non-light-damaged control group was compared to the light-damaged groups with various recovery periods using western blot. Cx43 was present in the choroid and in the retina before and after light damage (Figure 3). Quantification analysis of the western blots showed that the Cx43 level in the choroid and the RPE was significantly increased in the short-term following light damage (Figure 3A,B; one-way ANOVA, p<0.01 for 6 h and p<0.05 for 0 h). However, the expression level of Cx43 in the retina, although there is a trend of decreased expression, was not statistically significant compared to the control tissue (Figure 3C,D, p = 0.22).

Confocal images demonstrated that Cx43 was expressed in the RPE and the choroid (Figure 4A). Following light damage, stronger immunoreactivity of Cx43 was observed mainly in the RPE 0 h post-exposure (Figure 4B) and in the RPE and the choroid 6 h post-exposure (Figure 4C). Cx43 remained high at 24 h (Figure 4D) and 48 h (Figure 4E), but after 7 days appeared to be decreasing again (Figure 4F).

In the normal rat retina, Cx43 immunoreactivity was detected primarily in the ganglion cell layer (Figure 5A), which may be closely associated with the processes of astrocytes and the endothelium of the blood vessels. After light damage and up to 7 days post-exposure, Cx43 immunoreactivity was still detected in the ganglion cell layer and did not seem to be labeled in other cells or layers of the retina (Figure 5B–F).

Connexin 43 expression is detected in macrophages after light damage

Closer analysis of the labeling pattern of Cx43 expression in the choroid using double-labeling immunohistochemistry showed an association between Cx43 and inflammatory cells and oxidative stress markers. In the non-light-damaged choroid, there were CD45 immunolabeled macrophages, but they did not colocalize with Cx43 (Figure 6A–C). Cx43 was not only increased in amount but also coexpressed on some CD45 immunolabeled macrophages 24 h after light damage (Figure 6D-F). Interestingly, expression of Cx43 was also detected on nitrotyrosine immunolabeled cells in the choroid post-exposure (Figure 7D-F), a feature not seen in the non-light-damaged control (Figure 7A-C).

We further analyzed nitrotyrosine-labeled cells in the light-damaged choroid by investigating the expression of the ionized calcium-binding adaptor molecule-1 (Iba-1). Iba-1 recognizes all dendritic-derived macrophages and has been proven to be a reliable marker for microglia in the nervous system [33]. Colocalization of nitrotyrosine and Iba-1 demonstrated that the nitrotyrosine-labeled cells are Iba-1-positive macrophages seen at 6 h and at 24 h after light damage (Figure 8).

Alterations in connexin 36 and connexin 45 expression levels may contribute to light adaptation in the light-damaged animal model

Previous studies on the rat retina have demonstrated that Cx36 is expressed on AII amacrine cells, ON cone bipolar cells, and alpha retinal ganglion cells that are the key elements of the light signal transduction pathways [7]. The current study showed that Cx36 is increased in the inner region of the retinal inner plexiform layer within 24 h following light damage. This finding is consistent with previous reports on Cx36 expression levels in dark- or light-exposed animals [34-36]. Although we found increased expression of Cx36 in the light-damaged retina, the transcript and protein levels of Cx36 were significantly decreased after dark-adaptation in the mouse retina [36]. Increased Cx36 protein in the inner plexiform layer has been demonstrated to be associated with secondary cell loss in laser-created retinal lesions, and Cx36-deficient (Cx36^−/−) mice exhibit a significant increase in cell death after trauma [35]. These data must be interpreted with some caution since, as noted, Cx45 and Cx36 are known to change expression with light-dark adaptation. In this study, returning animals to normal light after intense light for 24 h could either appear as a light reduction (dark adaptation), or their circadian rhythms may have been disrupted. Some of the change in these connexins might then simply reflect disruption of light-dark adaptation pathways.

In addition to Cx36, our results for the immunohistochemical labeling showed a significant decrease in Cx45 immunoreactivity in the light-damaged retina compared to the control tissue. Cx45 is expressed on AII amacrine cells, ON cone bipolar cells, and alpha retinal ganglion cells in the rat retina [7]. Interestingly, another study reported that these cells participate in enhanced electrical coupling between alpha ganglion cells in the albino rabbit retina from night to day [34]. To the best of our knowledge, changes in Cx45 expression in the light-damaged animal have not been reported. Findings of connexin protein levels in the mouse retina following prolonged dark-exposure are not parallel with our results but have shown that the transcript of Cx45 is significantly decreased in response to dark-adaptation [36].

Heterotypic electrical synapses are formed between Cx36 located in AII amacrine cells and Cx45 located in ON cone bipolar cells in the inner plexiform layer, and Cx36^−/− mice and Cx45^−/− mice show decreased activity of the retinal interneurons in the rod and cone pathways [37]. The current study has demonstrated that the changes in Cx36 and Cx45 are restricted to the inner plexiform layer in the light-damaged retina, where bipolar cells, amacrine cells, and ganglion cells interact. This indicates that in addition to the RPE and photoreceptors in the outer retina, intense light exposure also leads to neuron reactions in the inner retina. The Cx36 and Cx45 expression changes suggest that these two connexin proteins react to photoreceptor damage and provide further evidence that Cx36 is associated with Cx45 in intercellular communications between inner retinal neurons in the light transduction pathway.

Increased connexin 43 expression is primarily in the choroid and the RPE in the light-damaged animal model

This study has demonstrated that Cx43 protein expression is significantly increased in the choroid and the RPE following light damage but little change was detected in the retina itself, which may be because the whole retina was used for western blotting analysis instead of selective use of the dorsal retina, which was the most stressed/damaged area. Cx43 upregulation in the light-damaged choroid and RPE is parallel with many studies on animal models of central nervous system (CNS) injuries [38,39]. Increased Cx43 protein and transcript have also been found in many human CNS injuries and diseases, including stroke, epilepsy, and Huntington’s disease [16-18]. The immunoreactivity of Cx43 is upregulated in the human glaucomatous retina compared to normal retinal tissue [15]. Increased Cx43 expression is mainly seen in the astrocytes located in the retinal ganglion cell layer of the peripapillary and mid-peripheral regions. In the peripapillary region, upregulated Cx43 is also expressed along glial fibrillary acidic protein (GFAP)-labeled Müller cell processes throughout the entire retina. Retinal endothelial cells retain the same expression level of Cx43 in the glaucomatous retina as the controls [15]. Elevated Cx43 expression was also detected in the rat retina, parallel with changes in GFAP immunoreactivity and ganglion cell loss following optic nerve injury [40]. It is interesting, therefore, to compare the changes seen in Cx43 following light damage in the albino rat model where it appeared to be restricted primarily to the choroid and the RPE.

Changes in connexin 43 expression is associated with oxidative stress and inflammation in light-damaged tissue

The present study found that the upregulated expression of Cx43 was throughout the choroid but notably colocalized with CD45-positive macrophages in the choroid and the retina following light damage. A previous in vitro study showed that upregulated Cx43 protein was detected in the primary culture of microglia after treatment with bacterial lipopolysaccharide (LPS) plus interferon-γ (INF-γ) or tumor necrosis factor-α (TNF-α) plus INF-γ [41]. Monocytes and macrophages express Cx43 only after being stimulated by exposure to IFN-γ and TNF-α but not under physiologic conditions [42]. Hemichannels constituted of Cx43 are located in the astrocytes of the CNS, and electrophysiological studies have indicated that these hemichannels are likely to be closed under physiologic conditions and to be opened under pathological conditions, such as oxidative stress [43]. A more recent study demonstrated that expression of Cx43 by activated phagocytes is critical for regulating phagocytosis and rearranging the F-actin cytoskeleton [44]. Cx43 is also expressed on many other residential and circulating leukocytes, including neutrophils, dendritic cells, and lymphocytes [45]. However, thus far, relatively little has been reported on Cx43 expression by inflammatory cells in tissue from in vivo investigations. Results from the present study suggest that Cx43 is expressed on activated macrophages in the choroid following intense light exposure, which may be critical for the role of these inflammatory cells. Many studies have suggested the involvement of the complement system in the choroid that may contribute to late AMD [23,46]. Results from this study using the light-damaged animal model indicate that Cx43 expression in macrophages may also be an important factor in pathogenesis and disease progression of early AMD. Tuning down the function of Cx43 may result in downregulation of the inflammatory response in the tissue as is seen after downregulation of Cx43 in skin wounds [47,48], liver failure [49], arthritis [50], and spinal cord injury [39,51].

The present study has shown that the expression of Cx43 was also detected in the nitrotyrosine-labeled macrophages in the choroid following light damage. Nitrotyrosine as a marker of protein nitration is increased in the photoreceptor outer segment and the RPE in the light-damaged animal model, when compared to the animals maintained in the dark [52]. Colocalization of nitrotyrosine and Iba-1 in the light-damaged choroid indicates a close association between oxidative stress and inflammatory responses. The lipid whisker model suggests that oxidized phospholipids from senescent and apoptotic cells can be recognized by receptors on macrophages and are then cleared up by the innate immune system [53]. In this study, nitrotyrosine immunoreactivity detected in the Iba-1 immunolabeled cells has provided a good indication of where the oxidized phospholipids end up. These findings suggest that oxidative stress is a preamble for the inflammatory responses that follow.

In summary, the present study has revealed changes in Cx36 and Cx45 expression in the retina in the light-damaged albino rat model. Increased Cx43 expression has been detected in the choroid following light damage, and this increase is closely associated with tissue damage, including oxidative stress and inflammation. Findings from the present study contribute to better understanding of tissue dysfunction and injury process occurring in the light-damaged albino rat model. This study also suggests that gap junction channels, particularly Cx43 channels, may be potential sites that contribute to reduced visual loss in some ocular diseases and injuries.