Photobiomodulation Protects the Retina from Light-Induced Photoreceptor Degeneration

Abstract

Purpose.: In this study, the hypothesis that near-infrared (NIR) light treatment (photobiomodulation) attenuates bright-light damage in the albino rat retina was tested.

Methods.: Young adult Sprague-Dawley (SD) albino rats were raised in dim (5 lux), cyclic light and then exposed to bright (1000 lux), continuous light for 24 hours. The animals were treated with 670-nm light (9 J/cm²) in an LED array before, during, or after exposure to light. The retinas were examined for function, structural changes, cell loss, and markers of stress and inflammation at 1 week and 1 month after exposure to damaging white light.

Results.: Bright light caused photoreceptor-specific cell death in control retinas. Significant upregulation of stress and neuroprotective factors and the presence of activated microglia were also noted after light-induced damage. Photobiomodulation profoundly attenuated histopathologic alterations in all three treatment groups. NIR treatment also abolished microglial invasion of the retina and significantly reduced the presence of stress and neuroprotectant molecules. Bright-light–induced reductions in photoreceptor function were significantly ameliorated by photobiomodulation in animals treated before and during exposure to damaging light. Photoreceptor function was initially reduced in animals treated after bright-light–induced damage, but recovered by 1 month after exposure.

Conclusions.: NIR photobiomodulation is protective against bright-light–induced retinal degeneration, even when NIR treatment is applied after exposure to light. This protective effect appears to involve a reduction of cell death and inflammation. Photobiomodulation has the potential to become an important treatment modality for the prevention or treatment of light-induced stress in the retina. More generally, it could be beneficial in the prevention and treatment of retinal conditions involving inflammatory mechanisms.

It has been known since the 1960s that exposure to excessive light causes photoreceptor damage and death. ^1,2 The severity of damage depends on multiple factors, including the intensity or the spectral distribution of the light, the duration of exposure, and the retina's history of light exposure. ^3,4 Lesions produced by light-induced damage (LD) are characterized by the damage or death of photoreceptors, the disorganization or loss of RPE cells, Müller cell gliosis, the disruption of the outer limiting membrane (OLM), and the obstruction of the underlying choroidal vessels. These structural changes are accompanied by the accumulation of microglia and lymphocytes in the retinal and choroidal vessels, typically in areas with the most severe damage. Once the blood–retina barrier (BRB) is disrupted, the invasion of the outer retina by activated microglia becomes apparent. ^{5 –9}  

Photobiomodulation is the process by which specific wavelengths of light are absorbed by cellular photoacceptor molecules, resulting in the activation of signaling pathways that culminate in biological changes in the cell. A typical example would be the absorption of light by chlorophyll in plants, initiating photosynthesis. In animals, light absorbed by opsins, initiates phototransduction in photoreceptors. Apart from these specialized photoreceptor molecules, there are more than 50 nonspecialized photoacceptor molecules in mammalian cells. Water, hemoglobin, oxyhemoglobin, and melanin all absorb light at a wide range of wavelengths. Between the 600- and 1000-nm wavelengths, however, their absorption curve is at its minimum, providing a window for other photosensitive molecules to be activated by light. ¹⁰ Cytochrome c oxidase, the terminal electron acceptor in the mitochondrial electron transport chain, is a key photoacceptor molecule, which has been shown to be activated by far-red (FR) to near-infrared (NIR) light, leading to biological changes. ^11,12  

The cellular responses to light in the range of FR to NIR photon therapy include an increase in metabolic rate, cell migration, and proliferation and the production and secretion of proteins. Zhang et al. ¹³ have demonstrated changes in gene expression profiles in human fibroblasts after exposure to 628-nm light for 3 days, describing upregulation in the mitochondrial respiratory chain and antioxidant genes and downregulation in some apoptosis and stress-response genes. Among the proteins, some neuroprotective factors, including basic fibroblast growth factor (FGF2) and nerve growth factor (NGF), have been shown to be upregulated. ¹⁴ The use of 670-nm light has also been demonstrated to protect neuronal cells after cyanide or TTX poisoning. ¹²  

In vivo experiments showed improved quality of wound healing in rat skin when treated with a low-energy ruby laser ¹⁵ or with a Gallium laser with 685- or 830-nm light. ¹⁶ After spinal cord injury, the use of 810-nm light resulted in increased regeneration and functional recovery in the rat. ¹⁷ Eells et al. ¹⁸ showed that 670-nm LED light was protective of photoreceptors in a methanol-induced degeneration rat model, and others demonstrated that NIR light promoted healing after laser lesions in primate retina. ¹⁹  

Clinically, the use of NIR light has been shown to be beneficial in the treatment of gingival incisions, ²⁰ posttransplantation oral mucositis, ²¹ radiation ulcers of the skin, ²² and acne vulgaris dermatitis. ²³ It promotes wound healing ²⁴ and peripheral nerve repair after trauma ^{25 –27} and in carpal tunnel syndrome. ^28,29 NIR treatment is also used in sports medicine and rehabilitation to treat acute soft tissue injuries. ³⁰ It is worth noting that the presence of inflammation is common to these conditions. 

Inflammatory events occur in the light-stressed retina. ^5,6,9 Moreover, anti-inflammatory measures, such as the administration of the microglial inhibitor naloxone have been shown to reduce LD in the retina, ^31,32 indicating that inflammation is a critical factor in this model. The purpose of this investigation was to test the hypothesis that 670-nm photobiomodulation will ameliorate light-induced retinal damage and modulate cellular immune response in the rat retina. 

Materials and Methods

Animals

All procedures were in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and with the requirements of The Australian National University Animal Experimentation Ethics Committee. Albino Sprague-Dawley (SD) rats were born and raised in low (5 lux) light levels in a 12-hour light, 12-hour dark cycle. Food and water were available ad libitum. Once the animals reached postnatal days (P)100 to P120, they were divided into six groups. Three groups were designed to serve as baseline comparisons. In the first group, the animals were not exposed to either NIR or damaging white light (control group, n = 16). In the second group, they were exposed to damaging white light only (LD group, n = 16), and in the third group they were exposed to NIR light only (NIR control group, n = 16). In the remaining groups, the animals were exposed to damaging bright white light (BL) and were treated with NIR according to one of three paradigms described below. 

Light Damage

The animals (n = 19) were transferred to individual transparent cages with food placed on the cage floor and water was provided in transparent bottles to avoid shading of the light entering the cage. Fluorescent light tubes (18 W, cool white) were placed 200 mm above the bottom of the cage, so that the light intensity reached 1000 lux at the cage floor. Before exposure to light, the animals were dark adapted overnight. Exposure started at 9 AM in all experimental paradigms. The animals were exposed to bright light for 24 hours and then returned to a low-light environment (5 lux) to recover for 1 week or 1 month. 

NIR Treatment Paradigms

The animals were wrapped in a cloth, to aid manual handling, and were placed under the 670-nm LED array (Quantum Devices, Barneveld, WI). The animals were positioned so that eye level was approximately 2.5 cm away from the light source and were exposed to the light for 3 minutes at 60 mW/cm². This treatment protocol produced an energy fluence of 9 J/cm² at eye level. The first group of animals were treated with NIR one time daily on five consecutive days before BL exposure (preconditioned [Precon] group, n = 16). The second group was treated immediately after the cessation of exposure, once a day for 5 days (post conditioned [Postcon] group, n = 16). The third group was treated 1 day before BL exposure, then two times daily during and immediately after exposure (midconditioned [Midcon] group, n = 16). In the NIR control group, the animals were treated with NIR light once daily for 5 days, but were not exposed to BL (NIR control group, n = 16). Some animals (n = 3) were sham treated before exposure to BL. They were wrapped in a towel and kept under an unswitched LED array box for 3 minutes. 

Electroretinography

The function of the photoreceptors was assessed by full-field, flash-evoked electroretinogram (ERG). The animals were dark-adapted overnight and prepared for recording in dim red illumination, as described previously. ^{33 –35}  

A series of single flashes of sufficient intensity to elicit saturated a- and b-wave responses (44.5 cd/m²) was used to record “mixed” responses, to assess both rod and cone photoreceptor function. As previously reported, the a-wave represents rod function in these mixed responses. ^36,37 All animals were recorded before the commencement of the experimental procedures (baseline recording) and then 1 week (n = 8/group) or 1 month (n = 8/group) after BL exposure, regardless of NIR treatment paradigm. 

Tissue Collection

Tissue was collected 1 week or 1 month after BL exposure, for histologic and immunohistochemical evaluation. The animals were euthanatized with an overdose of pentobarbital sodium (>60 mg/kg, intraperitoneal). The eyes were marked at the superior aspect of the limbus for orientation, enucleated, and immersion-fixed in 4% paraformaldehyde in 0.1 M phosphate-buffered saline (PBS) at pH 7.4 for 2 hours. 

To prepare them for cryosectioning, the eyes were rinsed thrice in 0.1 M PBS and left in a 15% sucrose solution overnight for cryoprotection. The next day, the eyes were embedded in OCT compound (Tissue-Tek; Sakura Finetek, Tokyo, Japan) and snap frozen in liquid nitrogen. They were cryosectioned (CM1850 cryostat; Leica Microsystems, Nussloch, Germany) in the sagittal plane, to allow dorsal-to-ventral observation of the retina at 16 or 60 μm. The sections were mounted on gelatin and poly-l-lysine–coated slides and dried overnight at 50°C before being stored at −20°C. 

Histology and Structural Measurements

Toluidine Blue Staining.

To evaluate structural changes in the retina, we performed toluidine blue staining. Sections were immersed in 70% ethanol for 5 minutes for dehydration. The slides were then washed with H₂O for 5 minutes before 2 drops of 1% toluidine blue were added and left to stain the sections for 5 minutes. Once the sections had assumed a purple-blue color, they were immediately washed with H₂O. The stained sections were coverslipped in glycerol-gelatin medium (Sigma-Aldrich, St. Louis, MO). 

Retinal Thickness Measurements.

To assess retinal thickness, cryosections were labeled with the DNA-specific dye bisbenzamide (Hoechst). The sections were thawed at room temperature before washing in 70% ethanol for 15 minutes, followed by a 5-minute wash in distilled H₂O and two 5-minute washes in 0.1 M PBS. The sections were then incubated for 2 minutes with bisbenzamide (1:10 000), washed in 0.1 M PBS, and coverslipped with a glycerol-gelatin medium. Retinal thickness measurements were made on digital images of stained cryosections. At each location, we recorded the thickness of the outer nuclear layer (ONL), as well as the thickness of the retina, from inner to outer limiting membrane (ILM-OLM). The ratio of the ONL to the ILM-OLM was used for analysis, to account for obliquely cut sections. In at least four sections per animal, we took 10 measurements from each section, approximately 100 μm apart along the retina (a total of 40 measurements per animal). Results from eight animals from each group were averaged and analyzed by the statistical method described below. 

OS Lengths Measurement.

Quantitative analysis of the length of the OS (outer segments) was performed in tissues cryosectioned at 60 μm and labeled with toluidine blue. The sections were examined and measured (LM Axioskop; Carl Zeiss Meditec, Jena, Germany) with a CCD camera (Spot; Diagnostic Instruments, Sterling, MI) with a calibrated 40× objective. Images were taken from both the superior and inferior retina (n = 4). The camera's CCD software was used to trace the length of the OS with a mean of three equally spaced measurements taken from each area. The average OS length from four animals was used for statistical analysis. 

Immunohistochemistry of Retinal Sections

Cryosections were labeled with antibodies against glial fibrillary acidic protein (GFAP; rabbit polyclonal; 1:700; Dako, Carpinteria, CA), basic fibroblast growth factor (bFGF-2; mouse monoclonal; 1:200; Upstate Biotechnology, Lake Placid, NY), ciliary neurotrophic factor (CNTF; goat polyclonal, 1:200; R&D Systems, Minneapolis, MN), ionized calcium binding adaptor molecule 1 (IBA1; mouse monoclonal, 1:500; Abcam, Cambridge, MA), and CD68, also known as ED-1 (mouse anti-rat monoclonal, 1:200; Millipore, Billerica, MA), in a protocol previously described. ³⁵ Briefly, the sections were blocked with 10% normal goat serum for 1 hour before being incubated with primary antibody for 24 hours at 4°C. The sections were treated with an antibody to either rabbit IgG conjugated with AlexaFluor 594 or mouse IgG conjugated with AlexaFluor 488 (1:1,000; Invitrogen-Molecular Probes, Eugene, OR) for 24 hours at 4°C before incubation with the DNA-specific dye bisbenzamide (1:10,000) for 2 minutes. 

RNA Isolation and cDNA Synthesis

The retinas were collected and stored in RNA stabilizer (RNAlater; Ambion-Applied Biosystems, Foster City, CA) overnight at 4°C. RNA isolation was then performed (TRIzol Reagent; Invitrogen, Carlsbad, CA). The retinas were homogenized in 1.5-mL tubes with 200 μL extraction reagent on ice, followed by further addition of 660 μL of the reagent and 160 μL chloroform. The tube was vortexed for 20 seconds, allowed to stand for 7 minutes at room temperature, and then centrifuged at 13,000g for 10 minutes at 4°C. The resulting supernatant was subsequently removed and transferred into a clean 1.5-mL tube with half of its volume in 100% absolute ethanol. The tube with its contents was vortexed briefly before the purification and DNase treatment steps were performed (RNAqueous-micro kit protocol; Ambion–Applied Biosystems). The purified and DNase-treated RNA was quantified on a spectrophotometer (ND-1000; Nanodrop Technologies, Wilmington, DE) followed by analysis of its integrity on a bioanalyzer (model 2100; Agilent Technologies, Santa Clara, CA). The synthesis of cDNA was achieved by reverse transcribing 1 μg of total RNA (Superscript III Reverse Transcriptase First-Strand Synthesis kit; Invitrogen Life Technologies), according to the prescribed manufacturer's protocol. 

Real-Time Quantitative Polymerase Chain Reaction

Validation of the changes in expression of stress-related Gfap and protective Fgf-2 genes were determined by RT-qPCR using gene analysis probes (Taqman Applied Biosystems, Inc. [ABI] Foster City, CA) Rn00566603_m1 for Gfap and Rn00570809_m1 for Fgf-2, combined with master mix (Gene Expression Master-Mix; ABI; on a StepOne Plus qPCR machine with the StepOne software ver. 2.1; ABI). Glyceraldehyde 3-phosphate dehydrogenase (Gapdh) was used as a reference gene. To account for variability, an amplification assay (Taqman; ABI) was performed in duplicate (individual sample variability) with triplicate biological samples to account for individual animal differences. The ratios of change were determined by the comparative cycle threshold method (^ΔΔC_t). 

Microscopy

Images of the immunolabeled retinal sections were obtained with a confocal microscope (LM Zeiss Apotome; Carl Zeiss Meditec). Only samples that were processed and imaged concurrently were used for analysis. During image collection, the photomultiplier settings were kept constant, to allow more accurate comparison of protein levels. 

Statistical Analyses

Data were analyzed with a two-tailed Student's t-test, with P < 0.05 considered to represent a statistically significant difference. All data are presented as the mean ±1 SEM. 

Results

NIR Photobiomodulation Preserved Retinal Structure in all Three Treatment Paradigms

Figure 1A depicts retinas of animals not exposed to BL or treated with NIR (control), animals exposed to BL (LD), animals not exposed to BL but treated with NIR (NIR-treated), and animals exposed to BL and treated with NIR according to one of the three treatment paradigms (Fig. 1; Precon, Midcon, Postcon). Histologic labeling of cryosections with toluidine blue showed that exposure to BL led to structural damage in the outer retina (Figs. 1A, 1B). The severity of the damage was uneven along the retina. One week after light exposure, an area of high-density cell death was detectable 1 to 2 mm superotemporal to the optic disc (depicted in Figs. 1A, 1C) that exhibited significant thinning of the ONL in BL-exposed, nontreated retinas. In this area, both inner (IS) and outer (OS) segments were lost, and only a few rows of photoreceptors remained. The retina was severely distorted, the OLM and the retinal pigment epithelial cell row were disrupted, and the BRB appeared to be compromised (Figs. 1A, 1B; LD). A similar pattern and severity of structural damage was observed in LD-exposed retinas with sham NIR treatment. Quantitative analyses indicated that there were no significant differences in the OS lengths and ONL thickness between the nontreated and sham NIR-treated retinas with LD exposure. In areas outside this hot spot, the retinal changes were minor, limited to the shortening and distortion of the photoreceptor OS (not shown). These images demonstrate that NIR treatment alone did not cause any structural alterations in the retina. In the light-exposed retinas treated with NIR, the preservation of the ONL was evident, whereas the IS and OS were only slightly shortened and mildly disrupted in all three treatment paradigms. 

Figure 1.

 

(A) Representative images of toluidine blue–labeled sections from the superior region of the retina. In control retina, the cells are distinctly arranged in respective layers and OS are neatly organized. In LD and sham NIR LD sections, severe disruption of the outer retina and diminished photoreceptor population were observed. No gross structural changes are seen in any of the NIR-treated groups. White line indicates orientation of ONL thickness measurement. (B) Higher magnification images of retinal sections labeled with toluidine blue, focusing on the OS, the adjacent pigment epithelium, and the choroid. The thick white vertical line indicates orientation of OS length measurement. White horizontal trace: the intact OLM. (C) Quantitative analysis of the photoreceptor population, assessed by measuring the thickness of the ONL and the full retina (n = 8). (D) Quantitative analysis of OS lengths in the different groups (n = 8). To counter for variations in cutting angles, the ratio of the two measures is depicted. (C, D, solid lines) The area of the retina with statistical significance. **Statistically significant differences (P < 0.05) compared with control groups. Scale bar, 25 μm.

The cumulative effect of photoreceptor cell death was assessed by the measurement of the ONL's thickness (Fig. 1C). Toluidine blue staining was used to detect surviving cells 1 week after BL exposure (Fig. 1A). The image of a BL-exposed nontreated retina shows a severe disruption and reduction of the photoreceptor population in the hot-spot area (Figs. 1A, 1B; LD). NIR treatment of BL-exposed retinas profoundly attenuated photoreceptor cell loss in all treatment paradigms. Figure 1C shows the average ONL thickness along the retina in eight animals in each treatment group. The BL-exposed, nontreated retinas showed a significant thinning of the ONL in all areas. In the inferior retina, there was a 25% loss, whereas in the superior retina, there was a 50% reduction of thickness, and in the hot-spot area, ONL thickness was reduced by 75% compared with that of the controls. The NIR-treated, BL-exposed animals exhibited a small decrease in ONL thickness. However, the ONL thickness in the NIR-treated, BL-exposed animals was not significantly different from values measured in the control animals. 

Figure 1D shows the average OS lengths along the retina in eight animals in each group 1 week after BL exposure. In the BL-exposed animals, the OS length was nearly 60% shorter in the inferior retina, when compared with corresponding areas of control retinas, and 70% to 75% shorter in the penumbra of the hot spot in the superior retina than in the controls, with a complete loss of the IS and OS in the hot-spot area. NIR treatment ameliorated OS damage in the BL-exposed animals. In the Precon group, OS lengths were not significantly different from those in the unexposed control animals. The retinas in the Midcon group exhibited an approximate 25% decrease in OS length compared with those of the control. In the Postcon group, the OS length was reduced to 50% of non-BL control values. In the hot-spot region, the length of OS dropped by approximately 60%; however, there was no complete loss of OS in any of the retinas in this group. 

NIR Photobiomodulation Attenuated LD-Induced Retinal Functional Loss

Figure 2 summarizes the changes in photoreceptor function 1 week after exposure to BL. Figure 2A shows typical recordings of selected animals from each group. Dotted lines show the baseline recording, and solid lines show post-BL recordings of the same animals. Gray lines represent nontreated BL-exposed animals, and black lines represent NIR-treated, light-stressed retinas 1 week after exposure. The data also demonstrate that there was a significant loss of retinal function after BL exposure in the Precon and Midcon animals; however, the NIR-treated animals exhibited a more moderate loss than did the nontreated animals, showing significantly better photoreceptor function than in the LD group. In the Postcon group, at 1 week after BL, photoreceptor function in the NIR-treated animals was indistinguishable from that of the nontreated LD animals. 

Figure 2.

 

(A) Representative dark-adapted scotopic ERG recordings showing traces of a-wave responses. Dotted lines: ERG recording before treatment and/or exposure to light; solid lines: ERG recording after light exposure and NIR treatment of the same animals. Photoreceptor function was significantly reduced in nontreated LD retinas (gray solid lines) after exposure to BL. Functional recovery was observed in the Precon and Midcon but not in the Postcon group (black solid lines). (B) Analysis of the relative preservation of the a-wave response of NIR-treated groups against the nontreated, LD retina. Functional changes were measured by taking the ratio of postdamage plus treatment and the baseline values of a-wave amplitude in all animals in each group (n = 8). (C) Analysis of the relative b-wave amplitudes of dark-adapted ERG responses in treated and nontreated SD rat retinas before (baseline recording) and 7 days after exposure to bright light. Changes in b-wave amplitudes were analyzed by taking the ratio of postdamage and baseline values in all animals in each group (n = 8). Bars, mean ± SEM. **Statistically significant differences (P < 0.05) compared with the control group.

Figure 2B shows the average a-wave amplitudes assessed by the dark-adapted, full-field flash ERG in eight animals in all experimental groups. Changes were expressed as the ratio of amplitudes measured 1 week after BL exposure and amplitudes measured before exposure in the same animal; thus, a value of 1 represented no change. The NIR treatment alone did not have a significant effect on photoreceptor function. Retinas exposed to BL but not treated with NIR showed a significant loss of photoreceptor function, decreasing to less than 20% of baseline value 1 week after BL exposure. In the NIR-treated groups the effect of BL on photoreceptor function varied. In the Precon group, the amplitude of the a-wave decreased to 56% of baseline values 1 week after BL. Although the decrease from baseline was significant, the post-LD amplitudes of the a-waves in the NIR Precon group were significantly greater than those measured in the LD group. In the NIR Midcon group, photoreceptor function was reduced to 60% of baseline values indicative of photoreceptor protection compared with the LD group. The NIR Postcon group exhibited significant photoreceptor dysfunction relative to baseline recordings. In this group, there was a 70% reduction in the a-wave amplitude at 7 days after LD. The relative a-wave amplitudes of the NIR Postcon group were not significantly different from those measured in the nontreated BL group. Thus, NIR postconditioning did not offer functional protection for photoreceptors at this early stage. 

The effects of NIR treatment on the functional response downstream from the photoreceptors was also assessed by measuring the b-wave amplitudes in the NIR-treated retinas 7 days after LD exposure. The relative b-wave amplitude was analyzed and presented in the same manner as the relative a-wave values were assessed (please refer to the previous section). As shown in Figure 2C, the b-wave response in the nontreated LD retinas was reduced by 73% from the baseline value. There was a slight variation observed in responses in b-wave amplitudes from the retinas in the NIR-treated groups. The reduction of b-wave response in the Midcon retina was only 20%, whereas the Precon group showed a 30% loss of b-wave amplitude. Conversely, the Postcon retinas showed the most severe reduction, with the b-wave response reaching 30% of control values, a level similar to that of the nontreated LD retina. There was no change observed in the NIR-treated only control. 

NIR Photobiomodulation Mitigated Retinal Stress and Modulated Neuroprotectant Regulation

We observed the levels of two known retinoprotectant proteins, CNTF and FGF-2, and that of GFAP, a known stress marker, after exposure to light and assessed the effects of NIR on these protein levels. CNTF was not detectable in the normal retina. Treatment with NIR alone did not change the pattern of protein expression and did not increase its concentration. After the exposure, there was an upregulation of protein levels in the astrocytes and in the Müller cells, which exhibited a strong, punctate labeling along the cell processes. In the penumbra of the hot spot, CNTF labeling in the retinal tissue was limited to the OS of photoreceptors. In animals exposed to BL and treated with NIR, CNTF protein upregulation varied. The least upregulation was found in the Precon group where CNTF was barely detectable in the retinal tissue. Retinas of the Midcon group showed a stronger presence of the protein, localized in the OS. Retinas of the Postcon group demonstrated the strongest upregulation in the photoreceptor OS layer and the astrocytes, where the labeling showed a pattern and intensity similar to that found in the penumbra region of the control retina (Fig. 3D). 

Figure 3.

 

(A, B) Validation of the differential expression of Fgf-2 and Gfap genes using quantitative RT-PCR. cDNAs from NIR-treated and nontreated, BL-exposed LD retinas were used for quantitative expression analysis and Gapdh was the reference standard. The level of expression of these genes in all groups was compared with that in the control group, and the value was generated for a change in expression. A change of 1-fold indicates no change (dashed line), a change >1-fold shows upregulation, and <1-fold indicates downregulation. Bars, mean ± SEM. **Statistically significant differences (P < 0.05) compared to LD group. (C) Retinal stress status of the treated and nontreated animals after 7 days of exposure to damaging levels of light through double immunolabeling with the stress marker GFAP (red) and the neuroprotective protein FGF-2 (green). Treatment with NIR mitigated the upregulation of both proteins in all three treatment groups. There was a gradient in the immunoreactivity for both proteins in the NIR treatment groups, with the strongest upregulation in the Postcon group. (D) Retinal sections labeled with CNTF antibody (red), showing a strong upregulation in the inner retina (astrocytes) and in the remnants of the ONL in the hot spot of LD retinas. In the penumbra, a few micrometers away, CNTF protein presence was more moderate and limited to the OS layer. A gradient of immunoreactivity in OS and GCL is demonstrated in the three NIR-treated groups. The blue labeling is bisbenzamide, a DNA-specific stain, indicating the nuclear layers of the retina. Scale bar, 25 μm.

GFAP labeling was limited to astrocytes in the innermost layer of the healthy retinas. In control animals and in animals treated with NIR alone, GFAP labeling showed a normal pattern, as demonstrated in Figure 3C in red. In light-stressed retinas, labeling of Müller cell processes became apparent. In nontreated, BL-exposed retinas a strong upregulation of GFAP was detectable in astrocytes and Müller cells 1 week after LD. Labeling was most prominent in the hot spot, where the entire Müller cell became GFAP⁺ from the ILM to the outer retina, and in the subretinal space, where activated Müller cells started to form a glial scar (Fig. 3C, red). In regions outside the hot spot, labeling was present in the Müller cells but was more muted (not shown). 

In the NIR-treated animals, in all three treatment paradigms, GFAP labeling was upregulated and was present in the Müller cells, but localized to the inner layers of the retina predominantly, not reaching the outer processes of Müller cells. However, there was a notable gradient between treatment paradigms, with a more moderate change in the Precon and Midcon animals and a more significant upregulation in the Postcon animals, although, even there, the levels of GFAP regulation did not reach the levels found in nontreated BL-exposed animals. 

FGF-2 is normally present in retinal macroglia (Müller cells and astrocytes), ganglion cells, and RPE, but not in photoreceptors. ^{38 –40} In control retinas, FGF-2 labeling was prominent in the Müller cell bodies in the inner nuclear layer (INL; Fig. 3C, green). After exposure to damaging light, FGF-2 protein was upregulated in the ONL. This upregulation was noted along the retina, but was most prominent in the hot spot. 

NIR treatment alone did not cause a change in FGF-2 protein expression. In animals from the Precon group, FGF-2 upregulation was not detected in the ONL. In the Midcon group, only a moderate increase in FGF-2 labeling was detectable in the INL and ONL of the hot-spot region, but not outside this area. There was a significant upregulation of FGF-2 protein in the ONL in the Postcon group, similar to that in control light-exposed retinas. 

To validate the expression of the FGF-2 and GFAP proteins described in the previous sections, we used qRT-PCR to assess the regulatory effects of NIR treatment on Gfap and Fgf-2 genes 1 week after BL exposure. We observed an increase in gene expression of Fgf-2 and Gfap in both light-exposed and NIR-treated retinas. Figure 3A shows that in the nontreated LD retina, there was a 5.8-fold increase in the expression of the Fgf-2 gene. However, the NIR-treated groups exhibited various levels of upregulation. There was a 3-fold increase in the Precon group, 6.2-fold in the Midcon group, and 8.7-fold in the Postcon group. When compared with the nontreated LD retina, the Precon group showed a significant difference. There was no change observed in the level of expression of the Fgf-2 gene in the NIR-treated control. A similar pattern of gene expression was observed in the stress-related Gfap gene after LD exposure and NIR treatment (Fig. 3B). An increase of 12.2-fold was seen in the nontreated, LD retina, although the upregulation in the Precon group was slightly down to 8.9-fold, and there was a more modest increase of 12-fold in the Midcon-treated retinas and 21.8-fold in the Postcon group (Fig. 3B). No change was observed in the NIR-treated control group. 

NIR Light Treatment Mitigated the Light-Induced Inflammatory Reaction

Antibodies against ED-1, a macrophage and monocyte marker, and IBA1, a microglia marker were used to assess the presence of inflammatory cells in the light-stressed retina (Fig. 4). In control non–BL-exposed, nontreated animals, very few monocytes were seen, and those that were present were limited to the choroidal vessels. In BL-exposed retinas immediately after BL exposure, monocytes were detectable in both the choroidal and retinal vessels in the area of the incipient hot spot. By 1 week after exposure, a breakdown of the BRB was apparent in the hot spot area, where monocytes invaded the retinal tissue and were present in the severely damaged ONL (Fig. 4A). NIR treatment alone did not change the number of monocytes in the choroidal or retinal vessels or in retinal tissue. In BL-exposed, NIR-treated animals, there was a slight increase in the number of monocytes both in retinal and choroidal vasculature; however, this increase was significantly less than the level in the light-exposed control animals. In addition, retinal invasion of macrophages/monocytes was not detectable in any of the treated animals (Fig. 4A). Quantitative assessment of the monocyte population was performed by counting ED-1⁺ profiles in different layers of the retina and in the choroid of four animals in each experimental group and then summarized (Fig. 4B). Light exposure caused a significant increase in the presence of monocytes in the choroid and inner retina, as well as in the severely damaged ONL of the hot spot. Although, there was a slight increase in the number of monocytes in both the choroidal and retinal vessels in the NIR treated, light-stressed retinas, this increase was significantly less than in the nontreated, BL-exposed retinas. Moreover, there were no ED-1⁺ profiles present in the retinal tissue in any of the NIR-treated retinas. 

Figure 4.

 

(A) Representative images of retinal sections, from the superior retina, labeled with the marker for macrophages, ED-1 (red). White arrows: ED-1⁺ cells in the NIR-treated retinas. (B) Quantification of the macrophages present in the choroidal or retinal vasculature and the retina. ED-1⁺ cells were sparsely present in the choroid of the control normal retina but absent in the ONL and inner retina. One week after exposure to LD, a significant increase in the number of ED-1⁺ cells was observed in the choroidal and inner retinal vasculatures, including an apparent tissue invasion. After NIR treatment, the increase in the number of ED-1⁺ cells was more muted and localized only to the choroid and retinal vasculatures in all treated groups (white arrows). (C) Representative images of retinal sections labeled with IBA1 (green), showing the resting and activated retinal microglial cells. In control retina, the resting microglia, identified by the ramified morphology (inset, C, left) were present along the inner retinal vasculature only. Seven days after BL exposure, there was an increase in the number of microglia that showed amoeboid morphology (inset, C, middle) in the ONL and retinal vasculatures suggestive of its activated form. After NIR treatment, activated cells were still present, but they were limited to the inner retinal vessels of all treated paradigms and not in the tissue. (D) The number of microglial cells present, distinguishing between resting and activated cells. **Statistically significant differences (P < 0.05) compared with control groups. Scale bar, 25 μm.

To establish the source of the activated microglia present in the retina, we performed IBA1 labeling (Fig. 4C). This antibody can detect a protein specifically expressed in microglia and macrophages and discriminate resting and activated microglia. Figure 4C shows the distribution of microglia in the retina. In control retinas, microglia present in choroidal and retinal vessels exhibited a ramified appearance, typical of resting microglial cells. After exposure to BL, most of the microglial cells that were detected in the retina showed an amoeboid configuration, typical of cells in an activated phase. When we quantified resting and activated microglia in the retina (Fig. 4D), we found that NIR treatment alone did not change the number or the activity status of the microglia when compared with the control. After exposure to BL, the number of microglia did not increase significantly; however, the ratio of activated cells increased significantly, reaching 85% of all microglia present. In the Precon and Midcon groups the ratio of activated cells increased significantly, although it only reached 66% and 57%, respectively, of the total microglial population detected. In the Postcon group, the number of microglial cells and the ratio of activated cells within their population were similar to those in the nontreated, BL-exposed group. 

Long-Term Effects of NIR Treatment on the Light-Stressed Retina

To gauge the extent of the protective effect of NIR treatment, we assessed retinas 1 month after BL exposure. Figure 5A summarizes photoreceptor function in all six experimental groups, expressed as the average of the ratio of post- to preexposure amplitudes in each animal; thus, the value of 1 means no change. Again, NIR treatment alone showed no significant effect on photoreceptor function. In the light-exposed but nontreated group, the average a-wave amplitude (Fig. 5A, top) fell to less than 15% of the preexposure value, which was not significantly different from values found at the 1-week time point. The Precon group showed a mild improvement from the 1-week average of 56% to 65% by 1 month after BL exposure. The Midcon group showed no further improvement from the values measured at the 1-week postexposure time, it remained close to 60%. However, when Postcon animals were allowed to recover for 1 month after BL exposure, there was a significant recovery in their photoreceptor function. Although the average a-wave amplitude was 30% of baseline values at the 1-week time point, by 1 month after BL, the average value rose to 58%, reaching the levels of the other two NIR treatment paradigm groups, becoming significantly greater than in the nontreated, BL-damaged group. 

Figure 5.

 

(A) Relative a-wave (top) and b-wave (bottom) amplitudes analyses before and 1 month after NIR treatment and BL exposure. (B) Representative images of retinal sections stained with toluidine blue at 1-month's recovery. White vertical line: the thickness of the ONL; black slanted line: orientation of OS length measurement. (C, D) Quantitative analyses of the OS length and ONL thickness sampled across the retina from the inferior to the superior edge (n = 8). Bars, ± SEM. **Statistically significant differences (P < 0.05) compared to control groups. Scale bar, 25 μm.

The long-term effect of NIR on the function of inner retinal cells was evaluated by measuring the b-wave amplitudes of NIR-treated animals 1 month after BL exposure. As demonstrated in Figure 5A (bottom), the b-wave amplitude response from nontreated, LD-exposed animals showed further reduction from the values detected at 1 week after LD, reaching 10% of baseline values. Conversely, the b-wave responses in the NIR-treated groups showed no further loss of function. In the Precon and Midcon groups, the values of the responses remained stable from 1 week after LD. However, there was a significant improvement observed in the Postcon group at 1 month after LD exposure. The b-wave amplitudes in this group reached 82% of baseline values, representing a more than 50% recovery from the levels 1 week after LD. Treatment with NIR alone did not cause any significant change in the photoreceptor function. 

Figure 5B demonstrates the retinal structure in the hot-spot area, after 1 month of recovery. In the nontreated, light-exposed retina, the ONL was obliterated and the RPE disrupted. The RPE and ONL were both present in retinas treated with NIR, albeit the photoreceptor layer was thinner when compared with that in the nonchallenged control retinas. The OS showed an organized appearance, compared with the disturbances observed in the Midcon and Postcon groups 1 week after BL. 

OS length measurements (Fig. 5C) showed that there was a total loss of OS in a wide area of the BL-exposed retina, spanning over 5 mm in length and centered by the original hot-spot area. In the NIR-treated groups, the length of the OS was reduced compared to those in the control retinas, but they were significantly longer than that of the nontreated, light-exposed retina. Moreover, the OS were present along the retina in all three treatment groups. 

To assess the cumulative effect of light and NIR treatment on the photoreceptor population, we measured the ONL thickness in all groups (Fig. 5D). In the LD group, photoreceptor cells were lost in a large, 4-mm long area in the superior retina extending from the original hot-spot area, whereas there was a more moderate thinning of the ONL in the rest of the retina. In the NIR-treated retinas the ONL thickness did not change greatly from that at the 1-week time point. 

Discussion

BL-Induced Retinal Damage

Our findings in retinas exposed to bright, continuous light for 24 hours are consistent with those reported in the literature. Three major events, characteristic of light-induced retinal damage are (1) wide-spread death of photoreceptors, (2) degeneration of pigment epithelial cells, and (3) reduction or total extinction of the ERG. ^2,41 Rapp and Williams ⁴² reported on the variable sensitivity of retinal regions to light. Typically, the superior central portion of the retina shows a higher sensitivity than other areas. ^{9,42 –44} This region has been shown to contain the highest concentration of ganglion cells, ⁴⁵ and photoreceptors in this area possess longer OS, ⁴⁶ which renders this part of the retina the functional area centralis in the rat. Recently, we reported that in the LD paradigm used in this study, the most severe, irreversible damage occurred in this central area of the retina. Moreover, severe, acute damage became the center of progressive degeneration. ⁹  

Photobiomodulation-Induced Retinoprotection

This study demonstrated that 670-nm photobiomodulation ameliorates the damaging effects of bright, continuous light on the retina. Treatment with 670-nm light before, during, or even after exposure to BL led to a significant reduction in photoreceptor cell death and prevented the severe disruption of the outer retina and the RPE. Photobiomodulation prevented the obliteration of the choroidal vascular network, thus assuring the maintenance of the blood–retina barrier. Photoreceptor structure was maintained in the treated groups, albeit the OS was shortened and misaligned, and some small vacuoles were present in the hot-spot area. These changes, however, were reversible, because after a period of recovery in dim light, the OS became more organized and longer than at the early stages after light-induced stress. Our findings in the Precon group somewhat agreed with those reported by Qu et al. ⁴⁷ Although they showed a reduced rate of photoreceptor loss and better retinal function in the NIR-treated animals in their LD paradigm, they were not able to avoid the development of a hot spot that could be a source of long-term degeneration. 

Photobiomodulation also reduced cell stress and inflammatory reaction in the retina. NIR treatment mitigated the upregulation of GFAP in Müller cells, a known stress marker. It also reduced the number of microglial cells in the retinal and choroidal vessels. Although it did not prevent the activation of microglia, it significantly reduced the number of activated cells and prevented their invasion of the outer retina. 

In conjunction with the maintenance of outer retinal structure, photoreceptor cell function was also preserved. In animals treated before (Precon) or during (Midcon) BL, photoreceptor function was significantly better than in the nontreated controls at 1 week after exposure, and function was maintained at the 1-month time point. In animals treated after BL (Postcon), photoreceptor function was significantly reduced 1 week after exposure. However, by 1 month after exposure, photoreceptor function recovered in this group as well, indicative of a protective effect similar to the other two treatment paradigms. In contrast, photoreceptor function remained diminished in the nontreated BL-exposed animals. 

The apparent discrepancy between the state of retinal structure and its function in the Postcon group prompted us to investigate the possible mechanisms responsible for the reduction of photoreceptor function without apparent major structural damage. One such mechanism may be through the upregulation of the neuroprotectants CNTF and FGF-2. Earlier studies have shown that both of these proteins are effective retinoprotectants. ^{48 –51} It has also been demonstrated, that both of these factors have a direct effect on retinal function. ^{49,52 –55} In our case, both CNTF and FGF-2 were upregulated in the light-exposed nontreated group, similar to earlier published data. ⁵⁶ The upregulation of both factors has been related to the reduction in retinal function, ^52,53 and thus they could be factors in the functional changes observed in our animals. In the NIR-treated animals, we were able to detect the upregulation of CNTF in the photoreceptor OS layer and FGF-2 protein in the ONL to various degrees, which could explain the difference in functional response in the three treatment groups. The most severe loss of function correlated with the highest level of CNTF and FGF-2 protein upregulation. It has been shown that the neuroprotectant upregulation after an insult is only temporary, ⁵² and once the stress is removed, protein levels are gradually reduced. It is possible, that the transient loss of function in the Postcon retinas are due, at least partly, to the increased presence of neuroprotectants in the outer retina, which allowed the survival and recovery of photoreceptors. Once the crisis was over and their presence was no longer necessary, the level of these factors reduced, allowing the recovery of retinal function. In retinas exposed to damaging light, but not treated by 670-nm photobiomodulation, the white-light–induced damage of the photoreceptors and RPE may have overwhelmed the retinoprotective system, and despite the presence of these factors, a large number of photoreceptors were lost and retinal function could not recover. 

Potential Mechanism of Photobiomodulation

Photobiomodulation, or low-energy photon irradiation by FR/NIR light using low-energy lasers or LED arrays, collectively termed photobiomodulation, has been applied clinically in the treatment of soft tissue injuries and acceleration of wound healing for more than 30 years. ⁵⁷ Recent studies have demonstrated that low-energy laser and LED NIR penetrates diseased tissues including the heart, spinal cord, and brain. ^58,59 NIR treatment has been documented to improve recovery from ischemic heart injury, ⁵⁸ attenuate degeneration in the injured retina and optic nerve, ^58,60 and improve recovery in experimental and clinical stroke. ^59,61,62 Clinical trials in the treatment of radiation-induced mucositis and stroke have demonstrated therapeutic efficacy of 670- and 830-nm light administered at doses ranging from 3 to 6 J/cm². Thus, existing data demonstrate the therapeutic potential of NIR light. 

The cytoprotective action spectrum of FR/NIR light corresponds with the cytochrome oxidase absorption spectrum. ^57,63 Recent studies have shown that NIR irradiation produces redox alterations in the cytochrome oxidase molecule, resulting in the activation of intracellular signaling cascades that culminate in improved mitochondrial function and increased synthesis of cytoprotective factors. ^{12,57,59,60,64 –68} Others showed increased tissue concentrations of antioxidants (glutathione, mitochondrial superoxide dismutase [SOD2]) and cytoprotective growth factors. ⁶⁹ Recently, we reported on the gene regulatory effects of NIR treatment in the normal and light-stressed retina. ⁷⁰ Our findings showed that NIR had an effect on many intracellular pathways, among others a direct effect on antioxidant protection and the downregulation of a chemokine, ccl2, that has been shown to induce leukocyte recruitment and activation. ⁷¹ The present study provided further evidence of the direct effect of NIR light on inflammation, by showing a reduction of monocyte recruitment and microglial activation in treated eyes. 

Our earlier gene chip study suggests that NIR has no direct effect on the regulation of neuroprotectants, such as Fgf-2 or Cntf genes. This study confirmed the negative correlation between structural protection and FGF-2 regulation that suggests that photobiomodulation does not act through neuroprotective pathways directly, but possibly at a point more upstream, thereby preventing, rather than mitigating cell damage. 

More interestingly, our previous wide-scale gene analysis showed that NIR upregulates many noncoding RNAs (ncRNA). ⁷⁰ That could explain the lack of any apparent effect of NIR treatment on the healthy, nonchallenged retina. Further research is necessary to understand the exact role these noncoding sequences play in the retina, but one possible hypothesis presents itself, that although they do not have any apparent effect on healthy tissue, they may act as a preconditioning event that allows the modification of the regulation of potentially damaging genes in the presence of stress stimuli. 

Clinical Implications

The use of photobiomodulation remains controversial, as a consequence of an incomplete understanding of its mechanisms of action, differences in treatment paradigms and wavelengths, and mixed results in both the laboratory and clinic. ⁷² In addition, it is hard to assess the efficacy of the treatment in the clinic, because of the lack of appropriately controlled and blinded clinical trials. Photobiomodulation is used in many types of tissues targeting a variety of conditions using different wavelengths, dose, and light sources (laser versus noncoherent light), which makes standardization near impossible. Despite the controversies, many successes have been achieved in a few areas clinically, specifically in three areas: (1) wound healing, (2) the relief of inflammation, and (3) the reduction of neurogenic pain. 

Present data suggest that treatment with 670-nm red light can lead to significant protection of the retina from LD. This treatment has the potential to reduce the adverse effects of bright light exposure; moreover, this noninvasive therapeutic modality has considerable promise for the treatment of retinal degenerative disorders and ocular inflammatory disease conditions. 

Footnotes

 Supported by the Australian Research Council through Grant CE0561903 from the ARC Centre of Excellence in Vision Science and the Foundation for Fighting Blindness Grant FFB TA-NP-1107-0435-UWI, TA-NP-0709-0465-UWI and The International Retinal Research Foundation.

Footnotes

 Disclosure: R. Albarracin, None; J. Eells, None; K. Valter, None