The effect of different bands of LED exposure on vitreous metabolites and retinal architecture.

The accumulating experimental evidence indicates that exposure to blue and white LED light leads to damage in the visual system against short-term exposure. Chronic exposure, adaptive responses to light and self-protective mechanisms against LED light exposures need to be explored and would be essential to know the repercussions of LED radiations on vitreous metabolites. A total of 24 male Wistar rats were used in our study, which was divided into four groups (n = 6 in each group). Three groups were exposed to either blue, white, or yellow LED light for 90 days (12:12 light-dark cycle routine) with uniform illumination (450–500 lux). Control rats were maintained under standard laboratory conditions. Post-exposure the vitreous was removed for mass spectrometry and retinal tissues for immunouorescence and H&E staining. The thickness of the retina decreased in blue and white light exposure animals compared with controls, whereas the yellow light exposure group showed an increase in thickness (p < 0.001). The number of apoptotic cells was signicantly lower in controls compared to light-exposed groups (p < 0.001) and (p < 0.001). Altered metabolites were observed in light exposure groups particularly in D-alanine, taurine, D-serin (p < 0.05 and lysine (p < 0.001). The self-protective or reworking system in the chronic light exposure could be dazed and drop the ability to compensate for the defending mechanism. This might fail to maintain the metabolomic structural integrity of the vitreous and retina. 0.1% formic in water and mobile phase B: 0.1% formic 90% (2% 98% 25 minutes, 98% B for 10 minutes and equilibrated to 2% B for 10min). spray with MS interface capillary 350°C. 140V. Drying and nebulizer were maintained at litres/minute and pressure of 40 psig respectively. Data was acquired at an acquisition rate of 2Hz in a range of 50-1700 m/z.

The Manipal Academy of Higher Education, Kasturba Medical College animal ethics committee approved the experimental protocol (IAEC/02/2017). Animal handling and investigational procedures were carried out as per IAEC and CPCSEA (No:94/PO/Re Bi/5/99/ CPCSEA) guidelines. The healthy animals (n=24) are divided into three experimental groups (n=6 in each group), one serving as an (unexposed) control group (n=6). We have included healthy rats for the study and the control rat's retinas were selected from the same pool of rats. However, it's safe to assume that the experimental control rats will also not have any degeneration. Both eyes of rats were exposed to blue, white, and yellow LED light with the maintenance of homogeneousness and intensity of light. The light standard was measured and standardized using Asensetek Lighting Passport Pro, New Taipei City, Taiwan [28]. The spectral sensitivity of different wavelengths of light visible spectrum 400-780nm) used for this experiment (450-500 lux) light. The spectral properties of light transmittance of the front surface of light, scotopic sensitivity, icker index and melatonin suppression of each light were measured. The total light exposure per day of (450-500 lux). The height from the source to the animals (cage) was 50 cm and the exposure time (12:12 dark & light cycle routine for 90 days) was matched with the nocturnal time of the rodents. The light exposure group animals were subjected to the blue LED (400 -490nm), white LED (380-780nm) and yellow LED (400-780nm) light exposure (Ack LED Panels, 3W, Epistar, ES-EMBCF22L-A, InGan -series Blue LED chip, ES-SMYLPX42C, UHB-PX, AlGaInP PNseries LED Chip). The light source was tted on the top of the cage (L=100cm, W=70cm and H=50cm) the mounting of the light kept it at 50 cms (100 % system level light output at standard operating voltage range, the total light output of the cage at the testing area (510 lumens). Post-exposure, all animals were sacri ced with a lethal dose of pentobarbital intraperitoneal injection (i.p. 100 mg/kg) (Euthasol®) and xylazine (10 mg/kg) (Proxylaz®). Then eyes were enucleated by using watchmaker forceps (number 5) and Sklar's blunt enucleation scissors. Immediately after enucleation, the vitreous was extracted by aspiration (21-gauge hollow needle) technique and stored in -20 o C for LC-MS and the retinal tissue stored in 4% paraformaldehyde (7 days) for immuno uorescence (IF) and Hematoxylin and eosin (H&E) stain.

Rat whole vitreous:
For the analysis of the chronic light exposure model, the whole vitreous sample was collected by aspiration technique. The samples were sonicated and stored at −20 °C until LC-MS analysis.

Vitreous Metabolomic Analysis
Untargeted quantitative metabolomics was used to analyze the vitreous uid samples using an LC-MS Agilent LC-QTOF system (Agilent Technologies, Santa Clara, CA, USA). It consisted of an Agilent 1200 LC system coupled online with an Agilent 6520 timeof-ight mass spectrometer. The vitreous samples were defrosted on ice, vortexed, and centrifuged at 12,000 rpm for 15 min. An 8 µl aliquot was injected into an Agilent 1290 LC system coupled to an ESI-Q-TOF instrument (Agilent 6520, Agilent Technologies, Santa Clara, CA, USA). HPLC column (Phenomenex, Torrance, CA, USA (P/No:00G-4601-E0; Desc: Kinetex 5μm C18 100A; Size: LC Column 250X 4.6 mm; S/No H18-343854; B/No 5701-0060) was maintained was at 25°C. The injected vitreous data were collected in the positive mode of electrospray ionization (ESI) technique. Basic and neutral metabolites were eluted in positive mode using subsequent gradient at 400 µl/min using mobile phase A: 0.1% formic acid in water and mobile phase B: 0.1% formic acid in 90% acetonitrile (2% to 98% B in 25 minutes, 98% B for 10 minutes and equilibrated to 2% B for 10min). ESI spray voltage was maintained at 3.5kV with MS interface capillary at 350°C. Fragmentor was set at 140V. Drying gas and nebulizer were maintained at 8 litres/minute and pressure of 40 psig respectively. Data was acquired at an acquisition rate of 2Hz in a range of 50-1700 m/z. Based on precise mass value and retention time, an untargeted mass spectrometer approach was employed for con rmation of the most abundant and signi cant metabolites.
Data processing and analysis: The raw data obtained from each run was processed using a molecular feature extraction tool in Qualitative Mass Hunter Analysis Software B.04.00 (Agilent Technologies). The data les containing monoisotopic mass, respective abundance and retention time were used for data alignment and ltering in Mass Pro ler Plus software (MPP) (Agilent Technologies, version B.12.5). Raw data les were segregated aligned, transformed to log10, and baselined to the median of all samples in MPP. Features that were present in at least 75% of individuals in each group were considered for further analysis. Partial least square discriminant analysis (PLS-DA) was performed on metabolome data to discriminate the study groups.Compounds were identi ed in METLIN and HMDB databases based on isotopic pattern distribution and accurate mass within a speci ed tolerance (15 ppm error). Peaks obtained from raw MS data were aligned and subjected to bioinformatics and statistical analysis. The metabolite peaks were attained in NC, BLE, WLE and YLE groups separately (Figure 1:).
Retinal Immunostaining: The twenty-four retinal tissues were embedded in 4% paraformaldehyde immediately after the following enucleation. The retinal tissues were sectioned after calibrating the cryostat for 40µ thickness and the sections were mounted on the gelatin-coated slides serially. Frozen sections (Leica cm3050 s) of 40 microns thickness were made in the sagittal plane and kept at -80°C before further processing. Sections were xed with 2% paraformaldehyde in 0.01M phosphate buffer solution at 24°C for 1 hour, followed by incubation with 1% H 2 O 2 , 2% sodium azide, 0.1% saponin, 10 mM (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) in EBSS -saponin for 36 hours at 24°C in the dark. Sections were then rinsed and incubated in a moist chamber overnight at 24°C with primary rabbit polyclonal to active + pro Caspase 3 antibody directed against cleaved caspase-3 (1:200 aliquoted 4µl of antibody is added to 1.6ml of PBS.TX) with working buffer and approximately 200µl of diluted (Abcam, Cambridge, UK). The retinal tissue sections were washed and incubated with a secondary antibody (1:100, Anti-Rabbit IgG (whole molecule) F(ab)2 fragment-Cy3 antibody, Abcam) for 16 hours at 2°C.
Immuno uorescence imaging and quanti cation: Caspase-3 (active + pro Caspase 3) stained retinae were imaged using a Dmi8-SP8 Confocal Microscope (Leica Microsystems) using a 63 X objective. The laser power offset and gain, and other acquisition parameters were determined by using isotopic control samples. The values of these parameters were then xed across all retinal samples. Retinal mosaics were created in Adobe Photoshop 7.0 (Mountain View, CA, USA). The area of each retinal layer was segmented by a masked observer, and the average intensity was calculated after subtracting the background using ImageJ software (To overcome the background intensity staining we have accounted brightness in each layer, and it was expressed relative intensity to the entire retina, and the total brightness was averaged). A total of 12 images, 2 from each eye were analyzed in each group and the intensity of IF was quanti ed in a masked method.

Measurement of hematoxylin and eosin-
Measurement of hematoxylin and eosin-stained retinal layers was conducted as follows. Images of a sagittal plane containing the optic nerve on para n-embedded blocks of H&E images were taken using a Motic Images Plus 3.0 for Windows 32 (2021-10-29) Microscope (Leica Microsystems, Germany) at 40× magni cation then Retinal mosaics were created in Adobe Photoshop 7.0 (Mountain View, CA, USA). Images were analyzed with the built-in measuring tool in Image-J software. The thicknesses of the whole retina (excluding the outer segments) and the ganglion cell layer (GCL), inner plexiform layer (IPL), inner nuclear layer (INL), outer plexiform layer (OPL) and outer nuclear layer (ONL) were measured for each layer in each picture in three different areas. Data were collected in Excel and then imported to R software (version 3.6.3) for statistical analysis

Data preparation and statistical application
The LC-MS/MS spectral region 5.00 ppm was segmented into bins of 0.05 ppm width Agilent pro software. The region anything above 5.00 ppm error was excluded from the analysis as it might have a remnant noise signal. A total column (bins) of metabolites were obtained and the integrated area within each bin was normalized to a constant sum of 500 for each range to minimize the effects of variable concentration among different samples.

Results
Different band of LED persuaded retinal insult: As we showed in gure 2: after 90 days o ight exposure the retinal layers showed depletion with marked proliferation, increased cellularity and enlarged cells with vacuolation of ganglion cell layer (GCL), and it became thinnerand disorganized, and more hyperchromatic nucleicells were observed in BLE and WLE groups. The GCL of the retinas inYLE was slightly disarrayed with nuclear enlargement and outer nuclear layer (ONL) thickness was increased with a focal enlargement (Figure 2:). The ONL thicknesswas decreased in BLE and WLE groups (Table 1:). We found there were varying notches of decreased thickness in the outer nuclear layer (ONL) in the BLE and WLE groups in contrast with YLE and NC. There was no signi cant difference was found in the INL thickness in all groups (p<0.01). As shown in (Figure 2:) the nucleolar damage, including pyknosis, was distinct with detached IS and OS structures and irregular arrangements of the photoreceptors in groups BLE and WLE when compared with the YLE and NC groups.
The stained retinal sections were obtained in the dorsal-ventral axis for immuno uorescence ( Figure 3A:) (IF) quanti cation. The whole retinal tissue section was selected at different peculiarities, for cell count and the total number of stained nuclei was counted in a 200-μm-long tissue section. For every retinal location, the amount of apoptosis in the NC group of retinae was signi cantly lower than in the light-exposed groups (ANOVA) (Figure 3B:). Besides, statistically signi cant differences were found in light exposure groups of rats (p<0.01) (Figure 3B:).
The retina showed a signi cant change after 90 days across the light exposure groups (Figure 3A:). IF analysis of CASPASE-3 expression showed a reorganization of morphological structures in BL and WL in comparison with NC whereas no obvious morphological changes were noticed in inner retinal layers and signi cant damage occurred in ONL in the YLE group using Cy-3 staining. The percentage of apoptotic cells was increased BLE and YLE in GCL, INL, ONL(p<0.001) and inner and outer segments of photoreceptor and IPL are relatively reduced in the YLE and NC group. However, the apoptosis was 86%, 84%, and 66% for BLE, WLE and YLE -light exposure tissues respectively (p<0.001) ( Figure 3B:). Four regions of interest in every retinal tissue section were selected at different peculiarities, for cell count and the total number of stained nuclei was counted in a 200-μm-long tissue section. The amount of apoptosis in the NC group of retinae was signi cantly lower than in the light-exposed group's rats (p<0.001) ( Figure   3B:). Besides, statistically signi cant differences were found in light exposure groups of rats (p<0.001) and (p<0.001) ( Figure 3B:). However, no self-protective or recovery was observed based on the hypothesis. The acquired images from the posterior pole within 1 mm of the optic nerve and the phototoxicity also were observed in the region of interest (ROI).
Metabolomic amendment of chronic light exposure and reprogramming.
A total of 24 vitreous samples, withdrawn from light exposure (BLE, WLE and YLE) and control group (Time 10:00 AM-10:30 AM IST) were analyzed. We got 20884 hits across the groups and these molecules were ltered out by frequency out of these 459 molecules which were presented across the samples. These molecules were considered to perform further analysis (one-way ANOVA) to observe the signi cance. 243 molecules were statistically signi cant (p<0.05) for the entity of 2.0-fold change (FC). Further, we did differential expression analysis using a t-test against zero, kept the p-value as asymptotic across the samples with an FC of 3.1. Out of total (20884) molecules 1521 were statistically signi cant to speci c 907 molecules (p<0.05), 348 molecules (p<0.02), 169molecules (p<0.001), 82molecules (p<0.0001) and 20 molecules were (p<0.00001) exhibited de nite statistical alteration.

PCA and PLS-DA
We steered Principal component analysis (PCA) on all detected metabolites. The parting of the NC, BLE, WLE and YLE groups demonstrated their difference in metabolomic expression. Most metabolites with extreme abundance were certainly isolated in light groups. Further, we did a summary of frequency normalization for abundance values (log10) for across the groups out of it 99% of samples were normally distributed and outliers were removed in each group. Only signi cant metabolites into account, the molecules in the same group clustered to each other which stated less difference among the light exposure group.
Through partial least square discriminant analysis, we found the pattern of metabolomics, with a sensitivity of 90% and a speci city of 87%. Notably, the amino acids and small molecule levels were elevated in the BLE and WLE groups. A clear difference in metabolomic expression of the BLE and WLE, YLE groups and the control group is displayed where there is a signi cant cluster of metabolites were observed between the groups.
Further to comprehend we heat map (Figure 4:) representation for the signi cant (243 molecules (p<0.05)) metabolites and we observed a clear upregulation of molecules in BLE, WLE and YLE where these molecules were in normal range/downregulated in NC. Having three groups of light exposure (BLE, WLE and YLE) from age-matched controls (NC). The BLE&WLE exposure groups have been vital to show the speci city of these biomarkers to vitreous. Some metabolites were perturbed in all groups. Untargeted metabolomics by LC-MS/MS revealed 20 metabolites that are biologically relevant to ocular uid. Thus, we next sought to validate our ndings using the targeted analysis to quantify the concentrations of these metabolites. As well as targeting the metabolites dysregulated in BLE, WLE and YLE samples, we expanded the targeted analysis to include metabolites from related metabolic pathways, to determine the biological relevance of our ndings using METLIN, Agilent pro and HMDB databases. The focused analysis was carried out using authentic standards to obtain accurate fold changes of the metabolites in the extracted samples. It was con rmed that 20 metabolites were statistically signi cant.
To further understand the perturbed metabolomic alterations, develop over the light exposure period of 90 days in Wistar rats, we performed one-way ANOVA and these values were compared across the groups. The p-values <0.05 considered as signi cant and signi cantly altered metabolites between the groups are plotted in Table 2: using Tukey HSD post-hoc test.
Pathway enrichment ( gure 5:) analysis was performed for the very signi cant 20 metabolites in the NC, BLE, WLE and YLE groups using the MetaboAnalyst 4.0 program. The analysis revealed that arginine, aspartate glutamate and taurine metabolites were the most perturbed pathways in light exposure conditions (Figure 5:). The analysis revealed that alanine, aspartate, and glutamate (p= 0.0600), Pathway analysis of the signi cant metabolites identi ed for the comparison among BLE revealed signi cant enrichment for aspartate (p = 0.0068), glutathione (p= 0.0263), glycine (p= 0.00436), glutamate (p =0.054), leucine and taurine (p=0.007) and taurine and hypotaurine degradation ( Figure 5:) were the most perturbed pathways in BLE and WLE conditions. This augments a growing frame of evidence to suggest that photo-chemical metabolism is particularly compromised in LE animals that potentially disrupts the taurine and hypotaurine dysregulation pathway.

Discussion
This study reveals 90 days of (12:12 hours) cyclic LED light exposure to rat's eye resulted in vitreous metabolomic changes and retinal structure alterations. These observations might be related to oxidative stress within the retinal tissue consistent with the aforementioned ndings of short exposure especially a few days to months [6, 13,29,30].
LC-MS showed obvious changes in the vitreous metabolites after 90 days of light exposure (BLE, WLE and YLE) and some of these changes, like upregulated amino acids (AA) and lipids were more intense in BLE and WLE groups while others, like the increase in levels of AA, were observed from YLE. Light exposed vitreous showed several notable differences in the spreading silhouettes of the perceived metabolites between LED light exposure groups and normal controls of the vitreous uid, predominantly, amino acids and proteins. Inadvertently, the role of altered metabolites and excitotoxic potential to light-induced damage in the vitreous is not well documented. It has been involved in a few ischemic induced ocular disease conditions including retinal vessel occlusion, glaucoma, and diabetic retinopathy [31,32]. The glutamate levels were altered in BLE and WLE groups which could also be a triggering factor for retinal apoptosis [33,34]. Glutamate is primarily mediated by overstimulation of the NMDA subtype of glutamate receptor, triggering an increase in intracellular calcium and initiating a cascade of events that nally lead to cell death/apoptosis, depending on glutamate levels [35]. The eccentricities in glutamate metabolism result in elevation of the extracellular concentration of glutamate which might increase the risk of excitotoxicity [33]. This indicates that failure in upholding glutamate homeostasis might lead to ocular toxicity [33].
The ganglion cell layer (GCL) thickness was drastically reduced in BLE and WLE groups, which might be due to the elevation of vitreal glutamate levels [36]. We noticed that metabolites of the vitreous uid were dysregulated in the light exposure (BLE and WLE) groups, which could be due to a shift of glycolysis' products, such as fructose 1,6 bisphosphate (F1,6BP; a product of glycolysis) and citrate, to the inner retina, while glucose 6 phosphate (G6P) was detected at high intensities both in the inner and outer retina [37]. The up-regulated metabolites in BLE and WLE groups could be due to metabolomic requirements to withstand or protect from light-induced damage. Studies have demonstrated the importance of cytokines induced protective mechanism of photoreceptors in light-induced damage against constant exposure [38,39] and stated that this could be due to survival-promoting genes and proteins that might alter during retinal light adaptation to a new bright light environment in the BLE and WLE groups. The identi ed molecules from Alvarez et al., 2001 which can protect against light-induced damage were up-regulated contemporaneously [40].
Interestingly, de la Barca et al., 2017 studied the effect of light stress in the retinal metabolomes and found that the mechanisms of light preconditioning (pre-exposure to moderate light before intense light) remained unknown, even though they seemed to have a self-protective effect. The light-induced changes in vitreous metabolites such as amino acids and lipids could likely involve nitric oxide-related signalling pathways [14,37]. Post-acute bright light exposure, the retinal proteins expressions were up-regulated in particular glial ®brillary acidic protein, heme oxygenase and heat-shock protein as well as cytokines [39,[41][42][43]. Beneath a constant light exposure condition, the endurance of the alteration of vitreous metabolites and its concentration revealed that this could be an endogenous oscillator, it correlates with previous experiments on aqueous proteins [44]. The upregulated metabolites concentration noticed in light exposure groups (BLE and WLE) (Fig. 4:) could be related to light responses in ocular bounded tissues or linked cellular restoration [44] and antioxidant protection to retinal tissues [45]. Abundant proteins remained to play a major role in modifying such light damage [46]. Our current experiment suggests that the degree of susceptibility to light damage in tissues adjoining to the vitreous may also be contingent upon time and light dependence.
The increase in period of cumulative and constant light exposure for 90 days on vitreous uid, the self-protective or reworking system in light (12:12 light-dark cycle) period could be dazed and might drop the ability to compensate for the defending mechanism, and thus might fail to maintain the metabolomic structural integrity. Such mechanisms increase post LED exposure if the compensatory mechanism is failed due to light damage and might lead to lethal consequences to the biochemical composition of vitreous uid.
With the effects of blue, white, and yellow LEDs on the whole-mount retina, the signi cant structural alterations and damage leading to the disruption of the retina, could be due to disruption of blood-retinal barriers by increased production of ROS [47]. We found that the BLE and WLE exposed retinal tissues exhibited higher levels of CASPASE-3 immunostaining mostly in the INL, ONL and GCL of the retina (Fig. 5:) and this is consistent with previous ndings [6,30]. Furthermore, few studies found that post light exposure with higher intensity of blue and white LEDs inhibited the increased proliferation of retinal cells with apoptosis compared with yellow LED [6]. Unfortunately, studies have not been conducted for chronic light exposure (Beyond 28 days). The relationship between long-term light exposure and retinal damage was unidenti ed. Few studies have indicated that lipofuscin could be a possible mediator of the risk associated with long-term exposure to blue light-induced retinal damage [48,49]. When lipofuscin absorbs blue light, it produces ROS that eventually leads to oxidative stress and retinal damage [27]. This mechanism (lipofuscin) is directly related to the type of light and its spectral composition [50]. Thomas et al reported that chronic blue light exposure could accelerate photoreceptor degeneration in an animal model in a retinal degeneration study [51]. The lipofuscin levels in retinal pigmented epithelium may alter with age, with levels being low in young and high in old animals [52]. This could also be one of the confounding factors to accelerate the retinal damage along with chronic exposure to the blue and white component of the light spectrum. The presence of a self-protective mechanism diminishes with age [53]. Light exposure leads to oxidative stress eventually resulting in the visual system no longer being able to protect the retinal architectures against chronic light, thus triggering irreversible vagaries in the visual system. Overall, the attained results can suggest the following sequences of events the ROS produced by the high-energy light from a blue and white LED [54] and the altered vitreous metabolites by inducing protein and AA modi cation affecting the hypotaurine pathway. This observation suggests that photo-chemical metabolism is compromised in light exposure rats potentially disrupting taurine and hypotaurine metabolism that might impact neural signals in the retina Fig. 5. This study has some limitations as we used only (CASPASE-3) immunohistological markers, which make little difference to the blue and white light delivered to the animals used in this study. A more detailed study, with more time point sampling and speci c cell death assay (e.g. western blot), further reveal understated anatomical changes to use OCT (real-time imaging) in chronic light exposure. The results shown here were obtained in rats with cumulative and continuous LED light exposure, their eyes are different from humans because they are nocturnal animals and do not have a macula this cannot be directly applicable to humans.

Conclusion
Our ndings state that chronic exposure to blue and white LED light damages the retina and alters the vitreous metabolites. The prolonged acute cumulative blue and white LED light exposure causes indemnities to the structural integrity of the vitreous and retinae. The self-protective or reworking system in the light-exposed visual system is dazed and deteriorate the ability to compensate for the self-protective mechanism.  Tables   Table 1: The different wavelength of LED light exposure of individual retinal layer thicknesses stained with hematoxylin and eosin were measured using ImageJ software (RPE to RGC) layer thickness were measured at quarterly points for each retinal crosssection and averaged. Averaged retinal thickness was converted to the percentage of the thicknesses from keeping standard against controls. Also, the layers are examined with different parameters (Atrophy, vacuolation, pyknosis and morphological alteration) by histopathologist.   Figure 1 Represents the spectra features of metabolite peaks were obtained in NC, BLE, WLE and YLE groups separately. The X-axis represents the count versus acquisition of metabolites time in minutes and Y-axis represents the generated distinct peaks on the total ion chromatogram (TIC) traces (*10 6 )    Metabolite set enrichment analysis: The Aspartate, Glycine, Serine, and glutathione metabolic pathways are the most signi cantly affected in the upper panel; 4 Upregulated FDR = 2.99E-01 and 4/20; FDR = 1.00E+00) and in the lower panel; 4 dysregulated features out of 20; False discovery rate (FDR) = 1.00E+00)