Influence of white-light-emitting diodes on primary visual cortex layer 5 pyramidal neurons (V1L5PNs) and remodeling by blue-light-blocking lenses

This study was approved by the Institutional Animal Ethics Committee (IAEC/KMC/35/2020) of Kasturba Medical College, Manipal Academy of Higher Education (MAHE), Manipal. Following the approval, healthy adult male 8-week-old Wistar rats (n = 24) were procured from the Central Animal Research facility, MAHE. Animal handling and investigational procedures were carried out in accordance with the prescribed guidelines from CPCSEA (94/PO/Re Bi/5/99/CPCSEA).

The procured animals were housed in a controlled laboratory setup in sterile polypropylene cages (L = 100 cms, W = 70 cms, H = 50 cms) with paddy husk bedding, including water and a standard pellet diet available ad-libitum [40]. The healthy rats were randomly divided into four groups: (1) control group (NC, n = 6) (2), white-light exposure group (LE, n = 6), (3) BBL-I—(Crizal Prevencia (CP, n = 6)) and (4) BBL-II—(Duravision Blue (DB, n = 6)). Animals in the NC group were maintained under a normal laboratory environment, whereas the LE group was exposed to white LEDs (450–500 lx) directly for 28 days (12:12-h dark/light cycle) to match the nocturnal time of the rodents with 100% light output. The light properties were standardized using a spectrometer (“Asensetek Lighting Passport Pro Spectrometer | Ushio America, Inc.,”) [41]. The white LEDs were fixed at a height of 52 cm on the top of the cage with a uniform illumination of 450–500 lx uniform exposure to light was maintained throughout the cage. For the treatment groups, the BBLs (CP and DB) were fixed to the LEDs and sealed to prevent direct light exposure.

The Morris water maze test was conducted in an open circular pool with a 150 cm diameter and a 40 cm depth with four imaginary quadrants filled approximately halfway with water and equilibrated the water temperature to room temperature of 18–22 °C [42]. The concealed platform was immersed 1 cm below the water surface in one of the four quadrants considered the target quadrant. This hidden platform was (4″ × 4″) camouflaged with non-toxic white tempera paint, rendering it indistinguishable due to the low visual aspect ratio to the water seen by the animal while swimming [40, 43‐45]. The orientation was facilitated by maintaining a definite visible cue (symbol ‘ + ’; 10 cm H = 10 cm, W = 10 cm, 100% contrast black target on white background) in the target zone [45]. The four quadrants in the pool are considered four zones, and the target zone is the fifth zone. A video camera (Logitech B525 HD Webcam) connected to a computer system (HANNS-G) was placed above the center of the pool to record the video and capture images (640 × 480 pixels). This computer system had special tracking software (ANY-maze version 4.82) to track the animals' movement and assess the recorded video [46].

After 28 days of LED exposure, all animal groups underwent four consecutive days of training, with each day consisting of four trials. During the prep training, an animal was placed on a platform kept at the center of a pool of water at 26 °C. The platform was exposed one inch above the water's surface to make the animal aware of its presence. Each animal underwent four trials, placing it on the platform for twenty seconds. The water maze had four starting positions, exploring different directions (anterior, posterior, right and left lateral) and the animal was taken to one of these positions. To avoid accentuating the animal, it was lowered into the water tail-end first with the support of a hand rather than dunking headfirst. The animal was allowed to find the platform within 60 s. If the animal failed to find the platform within the given time, it was trained to swim to the platform by being guided gently. This training procedure was repeated until the animal learned to find the platform. This process was repeated for four trials, each starting from a different position. After completing all four trials, the animal was warmed with a cotton towel. Post-training, the water maze test was performed.

The rats were placed facing the pool’s sidewalls from different starting positions, and the time taken to reach the hidden platform was recorded. Following the last training session, the animals were subjected to one session of memory retention test, during which the hidden platform was removed. The time to reach the target quadrant during the four consecutive training days and the memory retention test was calculated in seconds. Each animal’s total time to reach the target quadrant was measured in seconds [47]. On the fifth day, the hidden platform was removed from sight before conducting a memory retention test (60 s), which lasted for each animal. Latency (> 60 s) reaching the target quadrant suggested memory impairment. The time taken to reach the target quadrant was compared across all the groups.

Post-behavioral assessment, the animals were placed in the laboratory environment for two days before scarifying. The animals were killed with an overdose of Diethyl Ether 98% (LOBA CHEMIE PVT.LTD.), and the brain tissue was harvested and impregnated into freshly prepared Golgi-Cox stain (brain was separated into two hemispheres for better impregnation) for 21 days. Hemispheres are immersed in each tissue sample container of 10 ml Golgi-Cox solution and stored in a dark room at 24 °C. The Golgi-Cox solution was prepared 24 h prior and replaced once every five days for 21 days to ensure uniform staining and better penetration. Mercury chloride (Medilise Chemicals, KRL/KNR/00087/2003), potassium chromate, potassium dichromate, and distilled water were dissolved using a magnetic stirrer (ROTEK magnetic stirrer) to prepare Golgi-Cox solution. Post-Golgi-Cox staining (21 days of fixation), the brain tissue was fixed to the sledge microtome (Size 250 mm × 210 mm, H-325 × W-260 × D-610 mm) (Radical scientific equipment, Pvt. Ltd) plate with one drop of quick (PELCO^® Pro CA44 Tissue Adhesive) fix glue and sections of visual cortex with a thickness of 150 μm were obtained. The tissue sections were collected with a thick brush (Camel Hair Brushes, 3.18 mm wide) and transferred to tissue cassette (Leica-lp-c-biopsy-cassettes) and soaked in a 5% sodium carbonate solution for 20 min to enable clear visibility of tissues and neurons. The tissue sections were then dehydrated in ethyl alcohol (99.9% ethanol, UN No. 1170) in grades of 70% for 10 min (thrice), 90% for 10 min (thrice), and pure alcohol (99.9%) for 10 min (thrice) before being cleared with sulfur-free xylene (Spectrum) rendering the tissues transparent. Finally, those tissues were mounted on microscopic slides (BIOCRAFT, 26 × 76MM, CAT NO. 7101) using dibutyl phthalate Polystyrene Xylene (DPX, Sigma-Aldrich).

Motic Images Plus 2.0 ML software was used to capture microphotographs (every 5 µm thickness), well-stained neurons were captured using a light microscope with a digital microscope camera (Moticam 580, 5.0 MP; Model no. 12000425). A total of 846 neurons (36 neurons each animal) were selected from layer five pyramidal visual cortex neurons (V1L5PNs) in each group and 30 images were obtained from each neuron in the Z-axis using microscopy. The neurons were selected based on the following parameters: complete staining of an individual neuron, background staining, uniformity of neuronal staining and clarity in staining with dendritic spines. Any artifacts in staining were not taken into consideration during quantification. To minimize bias, two investigators (EOA and MS) manually traced dendrites from the coded slides, and the mean values were tabulated by a third investigator (RP). The neurons were traced from the soma (cell body) to assess apical and basal arborization for every 0 µm up to 140 µm and intersections from 20 up to 140 µm (5 µm precision). To obtain a precise measurement of dendrite quantification and branching, while avoiding any potential errors due to missed pruning or arborization, we took 30 images (each with a 5 µm resolution) for every neuron. These images were captured up to 150 µm from the soma (Z plane). This method has been demonstrated to be highly effective in minimizing errors, and the analysis technique was adopted [48, 49].

Functional and structural changes of the visual cortex post-white LED exposure were analyzed using an R programming software environment for statistical computing (version 3.6.3, Massachusetts Institute of Technology (MIT), Cambridge, MA, USA). The data obtained after performing the MWM test (training and retention) and structural data for neuronal quantification were analyzed across all the groups using two-way analysis of variance (Two-way ANOVA) followed by Tukey’s post hoc analysis to report the significance between the groups, if any. To mitigate any potential biases, two investigators (EOA and MS) manually traced dendrites from the coded slides. The mean values were then tabulated by a third investigator (RP), and statistical analysis was carried out by SBG.

Post 28 days of white LED exposure, the mean latencies (in seconds) to reach the hidden platform were compared using two-way ANOVA across all the groups for all four consecutive trials and memory retention test (Fig. 1A). Two-way ANOVA revealed a statistically significant difference (F _{3, 392} = 0.001, P < 0.001). Post hoc analysis between the groups showed significance only on day 1 (p < 0.05) and day 2 (p < 0.01) of the trial, respectively. On day 3 and day 4, the animals in all the groups demonstrated a similar latency trend.

The track plots (Fig. 1B) represent the path taken by the animals in the circular pool to reach the hidden platform. Longer path lengths taken by the animals in the light exposure group represent detrimental effects on spatial learning and memory.

The average time it took to find the hidden platform over four consecutive days was 16.38 s for the control group (Fig. 1B-a), 12.57 s for the light exposure group (Fig 1B-b), 15.04 s for the CP group (Fig. 1B-c), and 11.52 s for the DB group (Fig. 1B-d). On the memory retention day, the control group took an average of 19.76 s to find the platform, while the light exposure group took only 8.6 s, the CP group took 15.67 s, and the DB group took 9.53 s.

Apical and basal branching points and intersections of Golgi-stained V1L5PNs were compared across all the groups. There was a significant reduction in apical branching and intersection points at all the distances from the soma in the LE group and BBL groups compared to the NC group with a statistically significant difference (F_{18, 140} = 0.001, p < 0.001) (Fig. 2). A similar trend was seen in basal branching points and intersections with significance (F_{18, 140} = 0.019, p < 0.001) when other groups showed reduced basal branching and intersection points. This illustrates that the DB group had more branching points than the NC, LE, and CP groups (Fig. 3). These findings showed that white LED exposure has caused degenerative alterations in the neurons of the visual cortex region.