Topographic Quantification of the Transcorneal Electrical Stimulation (TES)–Induced Protective Effects on N-Methyl-N-Nitrosourea–Treated Retinas

Abstract

Purpose: To quantify the transcorneal electrical stimulation (TES)–induced effects on regional photoreceptors and visual signal pathway of N-methyl-N-nitrosourea (MNU)–treated retinas via topographic measurements.

Methods: N-methyl-N-nitrosourea–administered mice received TES or sham stimulations and were subsequently subjected to electroretinography (ERG), multielectrode array (MEA), and histologic and immunohistochemistry examinations. Quantitative reverse transcription–polymerase chain reaction (qRT-PCR) analyses were also performed to determine the mRNA levels of Bax, Bcl-2, Calpain-2, Caspase-3, brain-derived neurotrophic factor (BDNF), and ciliary neurotrophic factor (CNTF).

Results: Amplitudes of ERG b-wave in the TES-treated mice were significantly larger than those in the sham controls (P < 0.01). Microelectrode array examination revealed that the photoreceptors in TES-treated retina were efficiently preserved (P < 0.01). Morphologic measurements showed that the central retina region was more consolidated than the other areas in the TES-treated mice. Together with the disproportionate distribution of immunostaining in retinal flat mounts, these findings indicated that different rescuing kinetics existed among regional photoreceptors. Compared with the sham controls, a significantly increased signal-to-noise ratio was also found in the TES-treated mice (TES100: 2.02 ± 1.12; TES200: 4.42 ± 1.51; sham: 0.25 ± 0.13; P < 0.01). Moreover, qRT-PCR measurements suggested that the altered expression of several apoptotic factors and neurotrophic cytokines was correlated with TES-induced protection.

Conclusions: Regional photoreceptors in the MNU-administered retinas exhibit different sensitivities to TES. Transcorneal electrical stimulation is capable of ameliorating MNU-induced photoreceptor degeneration and rectifying abnormalities in the inner visual signal pathways.

Retinitis pigmentosa (RP) is a heterogeneous group of inherited retinal diseases characterized by initially impaired dark-adapted sight, progressive deterioration of visual fields, and eventual blindness.^1,2 Currently, the pathologic mechanisms of RP remain poorly understood, and there is no satisfactory therapeutic intervention.³ Retinitis pigmentosa models can be used to explore the pathologic mechanisms underlying this disorder, providing insights into the development of effective therapeutics. The N-methyl-N-nitrosourea (MNU) can pharmacologically induce photoreceptor degeneration. After a single systemic administration, active signs of retinal degeneration, such as decreased outer nuclear layer (ONL) thickness, degraded electroretinogram (ERG) response, and hyperexpressed apoptotic labeling, are indistinct in the MNU-treated retinas due to the selective photoreceptor cell loss. The mechanism underlying MNU-induced photoreceptor death is the principal alkylation of DNA, depending on the action of alkyladenine DNA glycosylase (Aag), an enzyme that removes alkylated bases via cleavage of glycosyl bond connecting base to the sugar phosphate backbone, thus generating abasic sites that can be further processed by the base excision repair machinery.⁴ N-methyl-N-nitrosourea could induce 7MeG and 3MeA DNA lesions, both of which are Aag substrates, and photoreceptors would die when the repair process can no longer operate efficiently enough. Generally, the degenerative process in mouse retina occurs within 7 days with an administered dose of 60 mg/kg.^5–7 The electrophysiological and morphologic properties of MNU-treated retinas are partially similar to those of the hereditary RP models. Particularly, damage severity and progression rate can be prospectively planned according to the concentration or the application time of the MNU injection.⁸ These prospective flexibilities could largely circumvent disadvantages such as unalterable time window for pathologic observation and therapeutic intervention in hereditary RP animal models. Therefore, it has been widely used in laboratory explorations for potential treatments.^9–13 

Transcorneal electrical stimulation (TES) is a novel electrical approach to activate the retina, resulting in protective effects on subjective retinal neurons.¹⁴ A series of animal experiments suggests that TES protects the retinal neurons from traumatic- or genetic-induced degeneration, supporting its clinical utilization against various retinal and optical diseases, such as traumatic optic neuropathy, anterior ischemic optic neuropathy (AION), and retinal artery occlusions (RAOs).^15–18 Several pioneering studies sought to clarify the functional mechanisms underlying TES-induced neuroprotective effects, showing that the beneficial effects are not attributed to a single pathway.^19–20 Conversely, multiple mechanisms collectively promote cell survival and contribute to cellular homeostasis of the TES-treated retina. For example, TES simultaneously regulates the expression of apoptosis-associated genes and neurotrophic factors to neutralize the intrinsic survival microenvironment of light-damaged retinas.¹⁵ Meanwhile, TES-induced anti-inflammatory effects are also correlated with this amelioration process.²⁰ The membranes of retinal neurons, such as photoreceptors, retinal ganglion cells (RGCs), and Müller cells, are rich in voltage-gated ion channels, which are reactive to extracellular electric field changes.^21,22 Transcorneal electrical stimulation can change the functional status of these retinal neurons by altering the activity of transmembrane voltage-gated ion channels. A recent study has shown that TES activates the L-type voltage-gated Ca²⁺ channels and triggers neurotrophin exocytosis and related antiapoptotic pathway.²³ 

Thus far, the TES-induced beneficial effects have been reported in the RCS rat²⁴ and the rhodopsin P347L transgenic rabbit.²⁵ However, very little is known about its potency in RP animal models with relatively rapid progressive dynamics. Moreover, most studies have focused on photoreceptor recovery in ONL, while the TES-induced effect on visual signal circuits in inner retina is rarely touched upon. The present study systematically explored the TES-induced effects on MNU-administered retinas via topographic examinations. The TES-induced rescuing effects on regions of the retina were quantified to evaluate therapeutic efficiency. Furthermore, the formerly reported abnormities in the visual signal pathway of MNU-administered retinas were also studied after TES therapy.²⁶ These topographic findings would enrich our knowledge about this electrical therapeutic strategy. It is our hope that the protective effects as evidenced by the present study will provide a better understanding of candidate treatments for human RP. 

Methods

Animal Models and Grouping

All experiments were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. All efforts were made to minimize the number of animals used and their suffering. All procedures regarding the use and the handling of the animals were conducted as approved by the Institutional Animal Care and Use Committee of the General Hospital of Chinese PLA. In total 200 C57/BL mice (8–9 weeks old with both sexes) were used in this study. The animals were randomly assigned to four groups with 50 mice in each group: normal control group, TES100-treated group, TES200-treated group, and sham-treated group. All animals were maintained under specified pathogen-free (SPF1) conditions (room temperature 18°–23°, 40%–65% humidity; illumination of 85 lux, 12-hour dark/light cycle with lights on at 07:00 and off at 19:00). All the animals were housed in 590 × 400 × 200-mm cages (Zhongke Animal Tec, ZX210001; Beijing, China) with food and water available ad libitum. The MNU (Sigma-Aldrich Corp., St. Louis, MO, USA) was kept at −20° in the dark. The MNU solution was dissolved in physiological saline containing 0.05% acetic acid (with a concentration of 0.0582 M). The MNU-administered mice received an intraperitoneal injection of MNU at the dose of 60 mg/kg body weight. No death occurred, and no clinical signs or system symptoms were evident in any of MNU-administered animals during the experiment. The normal controls were left untreated. 

TES Treatment

The MNU-administered mice were anesthetized by an intraperitoneal injection of ketamine (80 mg/kg) and chlorpromazine (10 mg/kg). After the cornea was anesthetized with a drop of 0.4% oxybuprocaine HCl, a noninvasive contact lens electrode with a golden wire ring was placed on the cornea with a drop of 2.5% methylcellulose (OmniVision, Puchheim, Germany) to maintain proper electrical contact and prevent corneal drying. A subcutaneously administered needle electrode was inserted dorsally between the eyes to serve as the reference electrode. Biphasic rectangular current pulses were delivered from an isolated constant current stimulator (SDT-1; Dongyuheng, Beijing, China) at a frequency of 20 Hz. A TES treatment lasted for 1 hour and was given to these MNU-administered mice at P1, P3, and P6 (day post MNU injection), respectively. The MNU-administered mice were randomly assigned to three groups based on biphasic rectangular current pulses: The TES (100) eyes received 100-μA stimulating current; the TES (200) eyes received 200-μA stimulating current; the sham eyes received 0-μA stimulating current. After each TES treatment, the mice received one drop of 1% atropine (Hi-Tech Pharmacal Co., Inc., Amityville, NY, USA) and were maintained on a constant temperature panel (22°) for 2 hours until anesthesia wore off. At the time point P9, all the TES-treated eyes were examined by slit lamp, and no side effects such as cataracts or corneal opacity were observed. Subsequent experimental examinations were performed at P9. 

ERG Recordings

The ERG recordings were performed according to the previously described methods.²⁷ Briefly, 10 mice from each animal group were weighed and dark adapted overnight before recording and were then anesthetized by an intraperitoneal injection of ketamine (80 mg/kg) and chlorpromazine (8 mg/kg) under dim red light conditions. Animals were lightly secured to the stage in the UTAS Visual Diagnostic System with a Big Shot Ganzfeld (LKC Technologies, Gaithersburg, MD, USA). Platinum circellus record electrodes were placed on each cornea and a reference electrode was subcutaneously placed between the eyes. White flashes with an intensity of 0.5 log cd-s/m² were applied to stimulate the scotopic ERGs. Photopic ERG measurements were obtained for flash intensities at the intensity 1.48 log cd-s/m². Amplitude of a-wave was measured as the distance from the baseline to a-wave trough, while amplitude of b-wave was defined as the distance between trough and peak of each waveform. The implicit time of the b-wave was measured from the stimulus onset to the peak of the b-wave. Signals were amplified and filtered by the band pass (1–300 Hz). A 50-Hz notch filter was applied to eliminate line noise. In total 60 photopic responses and 10 scotopic responses were recorded and averaged for a- and b-wave analysis. Oscillatory potentials (OPs) were isolated from the averaged recording traces using a 75 to 300-Hz digital filter. The magnitude of OPs was determined as the sum of the major amplitudes. Amplitude of the OPs was measured from the bottom of the trough preceding each OP peak to the top of that peak. 

Multielectrode Array (MEA) Recording

The MEA recording was described in detail in our previous reports.^26,27 Briefly, 10 mice from each animal group were killed under dim red light and their eyes were enucleated. The neural retina was gently removed from the pigment epithelium layer of the eye cup and placed into the recording chamber. The electrode array was composed of 64 electrodes, which were arranged in 8 × 8 layout with 450 μm for space configuration (Alphamed Sciences, Osaka, Japan). During recording, the ONH (optic nerve head) was fixed to the middle of the electrode array. The value of the regional response was averaged across individual responses from each recording channel belonging to that region. Retinal samples were perfused with oxygenated Ringer's solution (95% O₂ and 5% CO₂) during the whole process of recording. The analog extracellular neuronal signals were detected by the MEA system (MED-64; Alpha Med Sciences, Osaka, Japan) and were alternating current (AC) amplified, sampled at 20 kHz, and stored in a compatible computer for subsequent offline software analysis (Neuroexplorer; Nex Technologies, Littleton, MA, USA) software. Light-emitting diodes were driven by a computer stimulator to provide the retina a uniform full-field illumination at a mean photonic intensity of 850 mcd·s/m². Before spike detection, the field potentials were wiped off through a band-pass filter (100–3000 Hz). These candidate spike waveforms were then sorted by the Offline Sorter. The threshold for spike detection was set to four times the standard deviation (SD) of the mean value of the measured signal for each electrode. Those units without visual response were categorized as nonresponsive. Peristimulus time histograms (PSTHs) and raster plots of individual units were used to categorize the RGCs. The PSTHs were smoothed using a Gaussian kernel to analyze the ON and OFF responses. Units without visual response were categorized as nonresponsive. 

Figure 1

 

(A) Representative ERG waveforms of the examined eyes. (B) Both the photopic and scotopic b-wave amplitudes in the sham group were significantly decreased compared with those in normal controls. However, the photopic and scotopic b-wave amplitudes in TES-treated groups were significantly larger compared with those in normal controls. Moreover, the photopic and scotopic b-wave amplitudes in the TES200 group were less impaired than those in the TES100 group. (ANOVA followed by post hoc test; n = 10, *P < 0.05, #P < 0.01 for differences compared between; all values represent mean ± SEM.)

Figure 2

 

(A) Electrodes were classified into three groups according to their distances to ONH: the central channels, midperipheral channels, and peripheral channels. Moreover, the global recording field was divided into four quadrants. (B) Representative field potential waveforms of the examined eyes. (C) The mean amplitude of field potentials in the sham group was significantly reduced compared with that in normal controls. The field potential waveforms in the TES100 and TES200 groups were effectively preserved. Furthermore, the mean amplitude of field potentials in the TES200 group was significantly larger than that in the TES100 group. (ANOVA followed by post hoc test; n = 10, *P < 0.05, #P < 0.01 for differences compared between; all values represent mean ± SEM.)

Morphologic Evaluation by Quantitative Histology

Ten mice from each animal group were killed and their eyes were enucleated. A hole was made in the nasal ora serrata of the eye for orientation purposes. The eye cups were immersed in a fixative solution 4% paraformaldehyde (Dulbecco's PBS; Mediatech, Inc., Herndon, VA, USA) for 6 hours. They were rinsed in PB (phosphate buffer), dehydrated in a graded ethanol series, and embedded in paraffin wax. Five sections (thickness: 4 μm) cut vertically through the ONH of each eye were stained with HE (hematoxylin and eosin) and examined by light microscopy. Image-Pro Plus software (Media Cybernetics, Silver Spring, MD, USA) was used to outline the ONL. The adjacent thickness of the ONL was measured at 250-μm intervals along the vertically superior–inferior axis by a single observer in a masked fashion. Averaged layer thicknesses at each point were calculated and plotted as a function of eccentricity from the ONH, producing a morphometric profile across the vertical meridian. Similar to the MEA classification, the retina was divided into three regions: central (0–750 μm), midperipheral (750–1500 μm), and peripheral (1500–2250 μm) regions. 

Retinal Flat Mounts and Immunohistochemistry

Retinal flat mounts were prepared according to previously described methods.²⁸ Briefly, 10 mice from each animal group were killed and their eyes were enucleated. Optic nerve bud and its surrounding sclera were removed from the back of the eye cup. Soft touching and gentle pressing by forceps on the sclera of the eye cup helped to separate the entire neuroretinal layer from the RPE layer, and the remaining eye cup served for neuroretinal flat mounts. The flat mounts were then rained with PBS and were blocked in 2% normal goat serum, 0.3% Triton X-100 in 1% BSA for 1 hour at room temperature, then incubated overnight in peanut agglutinin conjugated to Alexa Fluor 488 (1:200, L21409; Invitrogen, Carlsbad, CA, USA) at 4°C overnight. After extensive washing with buffer, the flat mounts were incubated in Cy3-conjugated anti-rabbit IgG (1:400, 711-165-152; Jackson ImmunoResearch Laboratories, West Grove, PA, USA). The neuroretinal flat mounts were prepared by four cuts, at 3, 6, 9, and 12 o'clock, and were coverslipped with antifade Vectashield mounting medium (Vector Laboratories, Burlingame, CA, USA) for photographing. Fluorescence in flat mounts was analyzed with the Zeiss LSM 510 META microscope (Zeiss, Thornwood, NY, USA) fitted with Axiovision Rel. 4.6 software. The number of cones present within four 260 × 260-μm squares located 1 mm superior, temporal, inferior, and nasal to the center of the optic nerve was determined. For quantification, all lighting parameters on microscope and camera software were standardized to ensure consistent stable lighting throughout the image capture procedure. A background image of a blank slide was taken for each sample set and was subtracted from the corresponding sample image. Retina was then selected, and the integrated density (sum of pixel values above threshold) of immune staining, as well as the total selected area and its mean labeling intensity (mean value of pixels above threshold), was measured. 

Quantitative Reverse Transcription–Polymerase Chain Reaction (qRT-PCR)

For the qRT-PCR measurement, 10 mice from each animal group received two courses of TES, respectively, at P1 and P3. Then they were killed and their eyes were enucleated at P4. Eyes were marked at 12 o'clock with a hot needle to facilitate orientation. Neuroretinal whole mounts were prepared according to previously described methods and were then divided into quadrant patches with four incisions, at 3, 6, 9, and 12 o'clock.²⁹ Total RNA was extracted from the pooled retinal patches with a commercial reagent (Trizol; Gibco, Inc., Grand Island, NY, USA), followed by cDNA synthesis using the μMACS DNA Synthesis kit (Miltenyi Biotech GmbH, Bergisch-Gladbach, Germany). Primer sequences are depicted in Table 1, and all primers were quality controlled by sequencing the template on a genetic ABI analyzer (Applied Biosystems, Inc., Foster City, CA, USA). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) served as an internal standard of mRNA expression. Reactions were performed with SYBR Green Master Mix (Bio-Rad Laboratories, Reinach, Switzerland) on a real-time CFX96 Touch PCR detection system (Bio-Rad Laboratories). The amplification program consisted of polymerase activation at 95°C for 5 minutes and 50 cycles of denaturation at 95°C for 1 minute, annealing and extension at 59°C for 30 seconds. Duplicate RT-qPCR reactions were performed for each gene to minimize individual tube variability, and an average was taken for each time point. Threshold cycle efficiency corrections were calculated, and melting curves were obtained using cDNA for each individual-gene PCR assay. The relative expression levels were normalized and quantified using comparative threshold cycle (Ct):ΔΔCt = ΔCt (sample)−ΔCt (reference gene) (DATA assist Software v2.2, Applied Biosystems, Inc.). 

Table 1

 

Primer Sequences for mRNAs Amplified in qRT-PCR

 

Statistical Analysis

In MEA recording of firing spikes, clusters were first identified by a K-mean cluster algorithm. Analysis of variance followed by Bonferroni's post hoc analysis was performed to examine statistical differences between normal controls, TES (100) group, TES (200) group, and sham group. P < 0.05 was considered significant. The values are presented as means ± standard error of mean (SEM) unless otherwise specified. 

Results

TES Is Harmless to the Healthy Mouse Retina

In preliminary experiments, eight healthy C57/BL mice received five courses of TES200 every other day, and another eight healthy C57/BL mice received five courses of TES100. Subsequently, these mice were examined under a slit lamp. No corneal or lens damage was found in any of these subjects. Two days after the last course of TES treatment, these mice were subjected to functional and morphologic evaluations. The amplitudes of ERG b-waves in TES-treated mice were not significantly different from those in normal controls (P > 0.05). Meanwhile, the mean amplitudes of field potential in TES-treated mice were not significantly different from those in normal controls (P > 0.05). Furthermore, the morphologic analysis suggested that the mean ONL thicknesses of TES-treated mice were not significantly different from those of normal controls (P > 0.05). These results indicated that the TES did not affect the photoreceptor function or morphology in healthy C57/BL mice, and it could serve as a safe treatment for healthy subjects. 

TES Induced Protective Effects on the ERG of MNU-Administered Mice

Representative ERG waveforms of the examined eyes are shown in Figure 1A. The results showed that amplitudes of both the photopic and scotopic b-waves in the sham group (scotopic: 10.2 ± 6.1 μV; photopic: 5.2 ± 4.1 μV) were substantially decreased compared with normal controls (scotopic: 425.3 ± 87.7 μV; photopic: 95.4 ± 23.6 μV, P < 0.01; n = 10, Fig. 1B). However, amplitudes of the photopic and scotopic b-waves in the TES100 group (scotopic: 186.1 ± 43.6 μV; photopic: 29.0 ± 11.5 μV; n = 10) and TES200 group (scotopic: 241.5 ± 70.7 μV; photopic: 47.6 ± 16.4 μV, n = 10) were significantly larger compared with the sham group (P < 0.01). It is notable that the amplitudes of photopic and scotopic b-waves in the TES200 group were less impaired than in the TES100 group (P < 0.01). Moreover, the amplitude of a wave in the TES200 group (61.5 ± 15.1 μV, n = 10) was significantly larger than that in the TES100 group (41.0 ± 13.2 μV, n = 10, P < 0.01). The sum of the OP amplitudes in the TES100 group (125.5 ± 33.6 μV, n = 10) was significantly larger than that in the sham group (10.2 ± 9.1 μV, n = 10, P < 0.01). However, it was significantly smaller than that in the TES100 group (162.6 ± 39.1 μV, n = 10, P < 0.05). These OP results suggested that the TES treatment would be beneficial for the inner retinal function of MNU-administered mice. To test the sensitivity of the rod and cone system in treated eyes, we measured the implicit time of the photopic and scotopic b-waves. We found no significant difference in the implicit times of the b-waves between the TES100 (scotopic: 66.8 ± 4.26 ms; photopic: 45.0 ± 2.71 ms) and the TES200 group (scotopic: 65.1 ± 4.71 ms; photopic: 46.1 ± 2.53 ms, P > 0.05; n = 10). Furthermore, they were not significantly different from those of the normal controls (scotopic: 65.1 ± 4.71 ms; photopic: 47.1 ± 2.53 ms, P > 0.05; n = 10) or the sham group (scotopic: 64.8 ± 4.25 ms; photopic: 45.5 ± 2.90 ms, P > 0.05; n = 10). 

TES Induced Protective Effects on the Field Potential of MNU-Administered Mice

Multielectrode array recording detected light-induced field potentials and provided topographic information about regional photoreceptor function. The electrodes were classified into three groups according to their distances to ONH: the central channels, midperipheral channels, and peripheral channels (Fig. 2A). Moreover, the global recording field was divided into four quadrants: superior temporal (ST), superior nasal (SN), inferior temporal (IT), and inferior nasal (IN). Representative field potential waveforms of the examined eyes are shown in Figure 2B. The mean amplitude of field potentials in the sham group significantly decreased compared with normal controls (0.228 ± 0.033 vs. 0.012 ± 0.010 mV, P < 0.01; n = 10, Fig. 2C). Conversely, the field potential waveforms in the TES100 and TES200 groups were effectively preserved (0.088 ± 0.023 vs. 0.012 ± 0.010 mV, P < 0.01; 0.119 ± 0.029 vs. 0.012 ± 0.010 mV, P < 0.01; n = 10). Moreover, the mean amplitude of field potentials in the TES200 group was significantly larger than in the TES100 group (0.088 ± 0.023 vs. 0.119 ± 0.029 mV, P < 0.05; n = 10). Intriguingly, the field potential responses in the TES200 group were not uniformly equal and formed a topographic gradient across retina: The field potentials in the central region were retained with larger amplitudes than in the other two regions (0.148 ± 0.025 vs. 0.123 ± 0.020 mV, P < 0.05; 0.148 ± 0.025 vs. 0.100 ± 0.021 mV, P < 0.01; n = 10, Fig. 3A). Additionally, the mean amplitude of field potentials in the midperipheral region was significantly larger than in the peripheral region (0.123 ± 0.020 vs. 0.100 ± 0.021 mV, P < 0.05; n = 10). Similar disproportions among positional regions were also found in the TES100 group (0.122 ± 0.021 vs. 0.094 ± 0.019 mV, P < 0.01; 0.122 ± 0.021 vs. 0.073 ± 0.017 mV, P < 0.01; 0.094 ± 0.019 vs. 0.073 ± 0.017 mV, P < 0.05; n = 10). Topographic photoreceptor function of TES-treated retinas was quantified and compared with the normal controls: 61.3%, 50.1%, and 41.8% of the photoreceptor function, respectively, was retained in the central, midperipheral, and peripheral regions of the TES200 group. Meanwhile, 50.8%, 39.8%, and 31.5% of photoreceptor function, respectively, was retained in the central, midperipheral, and peripheral regions of the TES100 group. 

Figure 3

 

(A) Field potential responses in the TES200 group were not uniformly equal and formed a topographic gradient across the retina: The waveforms in the central region were retained with larger relative amplitudes than those in the other two regions. Moreover, amplitude of the waveforms in the midperipheral region was significantly larger than that in the peripheral region. Similar disproportions among positional regions were also found in the TES100 group. (B) Field potential of the ST quadrant was most efficiently preserved in TES-treated groups. Conversely, field potential of the IN quadrant was the smallest. The amplitude of the field potential in TES-treated eyes conformed to the following rule: ST > SN > IT > IN. However, the quadrant asymmetry was not found in the sham group or normal controls. (ANOVA followed by post hoc test; n = 10, *P < 0.05, #P < 0.01 for differences compared between; all values represent mean ± SEM.)

Waveforms of the retinal quadrants were also examined in greater detail, revealing that the field potential of the ST quadrant was most efficiently preserved in the TES-treated groups (TES200 and TES100: ST versus SN, P < 0.05; ST versus IT, P < 0.01; ST versus IN, P < 0.01; n = 10, Fig. 3B). Conversely, the field potential of the IN quadrant was smallest (TES200: IN versus SN, P < 0.05, IN versus IT, P < 0.05; TES100: IN versus SN, P < 0.01; n = 10). Field potential amplitude in the TES-treated mice conformed to the following rule: ST > SN > IT > IN. In sharp contrast, quadrant asymmetry was not found in the sham group or normal controls, indicating that regional photoreceptors were comparatively sensitive to TES treatment. 

TES Induced Protective Effects on the Retinal Morphology of MNU-Administered Mice

To measure the topographic retinal thickness, sections were taken along the superior–inferior axis to access the vertical meridian of each hemisphere. The mean ONL thickness of the sham group significantly decreased compared with normal controls (38.3 ± 4.33 vs. 1.6 ± 1.12 μm, P < 0.01; n = 10, Fig. 4). Moreover, mean ONL thickness of the TES200 group was significantly larger than in the TES100 group (18.6 ± 3.61 vs. 13.1 ± 2.58 μm, P < 0.01; n = 10). Closer examination of ONL thickness of the central, midperipheral, and peripheral regions revealed that photoreceptors in the central region were more sensitive to TES treatment: The ONL thickness in the central region was significantly larger than in the peripheral and midperipheral regions (TES100: 16.3 ± 3.16 vs. 8.3 ± 2.21 μm, P < 0.01; 16.3 ± 3.16 vs. 12.5 ± 2.08 μm, P < 0.01; TES200: 24.2 ± 3.75 vs. 10.9 ± 2.80 μm, P < 0.01; 24.2 ± 3.75 vs. 18.5 ± 3.01 μm, P < 0.01; n = 10). Compared with the normal controls, 57.1%, 46.6%, and 31.7% of the ONL thickness, respectively, was retained in the central, midperipheral, and peripheral regions of the TES200 group, with 38.1%, 33.0%, and 23.5%, respectively, in the TES100 group. 

Figure 4

 

(A) Sections stained with HE suggested that the ONL in TES-treated eyes was efficiently preserved. (B) The adjacent thickness of the ONL was measured along the vertically superior–inferior axis. Averaged layer thicknesses at each point were calculated to produce a morphometric profile across the vertical meridian. (C) Mean ONL thickness of the sham group significantly decreased compared with normal controls. Moreover, mean ONL thickness of the TES200 group was significantly larger than that of the TES100 group. In greater detail, we separately examined the ONL thickness of the central, midperipheral, and peripheral regions and found that photoreceptors in the central region were more sensitive to TES treatment. (ANOVA followed by post hoc test; n = 10, *P < 0.05, #P < 0.01 for differences compared between; all values represent mean ± SEM.)

TES Induced Protective Effects on the Cone Photoreceptors of MNU-Administered Mice

Because rods account for at least 96% of total photoreceptors in the mouse retina, ONL thickness mainly reveals rod integrity and should be considered as an indicator of the rod number and vitality.^30,31 However, it remained to be shown whether the cones were effectively rescued by TES treatment. Therefore, we performed peanut agglutinin lectin (PNA) immunostaining to verify the TES-induced effects on cones (Fig. 5A). Compared with normal controls, there was significantly less PNA immunostaining in the sham group following MNU treatment (613.4 ± 40.1 vs. 15.2 ± 8.8, P < 0.01; n = 10, Fig. 5B). Cone density of the TES-treated retinas was significantly greater compared with the sham group (TES200: 335.1 ± 31.4 vs. 15.2 ± 8.8, P < 0.01; TES100: 231.1 ± 26.0 vs. 15.2 ± 8.8, P < 0.01; n = 10). Moreover, cones in the TES200 group were more efficiently preserved than in the TES100 group (335.1 ± 31.4 vs. 231.1 ± 26.0, P < 0.01; n = 10). Compared with normal controls, 55.6% of cone density was retained in the TES200 group, with only 36.6% in the TES100 group. Additionally, cone densities of the retinal quadrants were examined, suggesting that cone density in the ST quadrant was significantly greater than in the other three quadrants of TES-treated retinas (ST versus SN, P < 0.05; ST versus IT, P < 0.01; ST versus IN, P < 0.01; n = 10, Fig. 5C). Meanwhile, cone density in the IN quadrant was smallest (IN versus SN P < 0.01; IN versus IT, P < 0.01; n = 10). Similar quadrant asymmetry was not found in the sham or the normal control groups, indicating that cones in the ST quadrant were more preferentially rescued by TES treatment. 

Figure 5

 

(A) PNA immunostaining was performed to flat mounts to verify the TES-induced effects on cones. (B) PNA immunostaining of the sham group was remarkably eliminated by MNU administration at P9. The cone density in TES-treated retinas was significantly higher compared with that in the sham group. Moreover, the cones in the TES200 group were more efficiently preserved than those in the TES100 group. (C) The cone density of the ST quadrant was significantly higher than that of the other three quadrants in TES-treated retinas. Meanwhile, the cone density of the IN quadrant was the smallest among the four quadrants. Similar quadrant asymmetry was not found in the sham group or normal controls, indicating that cones in the ST quadrant were more preferentially rescued by the TES treatment. (ANOVA followed by post hoc test; n = 10, *P < 0.05, #P < 0.01 for differences compared between; all values represent mean ± SEM.)

TES Induced Protective Effects on the Visual Signal Pathways of MNU-Administered Mice

Firing spikes of RGCs were harvested by the MEA recording system (Fig. 6A), suggesting that the spontaneous firing rate in the sham group was significantly higher compared with normal controls (2.63 ± 1.48 vs. 11.05 ± 3.23 spikes/s, P < 0.01; n = 10, Fig. 6B). In the TES200 group, the spontaneous firing rate significantly increased (2.63 ± 1.48 vs. 4.40 ± 2.08 spikes/s, P < 0.05; n = 10, Fig. 6B), but was significantly lower than in the TES100 group (6.48 ± 2.32 spikes/s, P < 0.05; n = 10) and sham group (11.05 ± 3.23 spikes/s, P < 0.01; n = 10). 

Figure 6

 

(A) Firing spikes of RGCs were recorded by MEA system. (B) The spontaneous firing rate in the sham group was significantly higher compared with that in normal controls. The spontaneous firing rate in the TES200 group also increased significantly, while it was significantly lower compared with those in the TES100 group and the sham group. (ANOVA analysis followed by post hoc test; n = 10, *P < 0.05, #P < 0.01 for differences compared between; All values represent mean ± SEM.)

Light-induced RGC responses were evoked by full-field stimulus and the firing spikes were extracted by offline software analysis (Neuroexplorer). Six categories of RGC populations were distinguished by their responsive characteristics to light stimulus (Fig. 7A). The total firing rate significantly decreased in the sham group compared with the normal controls (29.57 ± 4.85 vs. 2.48 ± 1.64 spikes/s, P < 0.01; n = 10). Moreover, the total firing rate in the TES200 group and TES100 group was less impaired (TES200: 16.54 ± 3.41 vs. 2.48 ± 1.64 spikes/s, P < 0.01; TES200: 12.20 ± 2.97 vs. 2.48 ± 1.64 spikes/s, P < 0.01; n = 10, Fig. 7B). We separately examined the ON and OFF response, and found that the OFF pathway was more efficiently preserved than the ON pathway: In the TES200 group, the ON firing rate was significantly smaller than in the normal controls (22.72 ± 3.65 vs. 9.63 ± 2.53 spikes/s, P < 0.01; n = 10). However, the difference in OFF firing rate between the TES200 group and normal controls was not statistically significant (15.09 ± 3.38 vs. 12.91 ± 3.06 spikes/s, P > 0.05; n = 10). In the TES100 group, the rescuing effects were less effective because both ON and OFF responses were significantly impaired compared with the normal controls (ON response: 22.72 ± 3.65 vs. 6.59 ± 2.62 spikes/s, P < 0.01; OFF response: 15.09 ± 3.38 vs. 9.01 ± 2.73, spikes/s, P < 0.01; n = 10). The spatial configuration of firing spikes was also monitored in parallel with field potentials. However, no topographic change in firing spikes was found in any of recorded retinas. 

Figure 7

 

(A) Raster plot (up) of RGC populations with corresponding PSTHs (down). Mainly six categories of RGC populations were distinguished by their responsive characteristics to light stimulus. (B) The total firing rate in the sham group decreased significantly compared with that in normal controls. However, the total firing rate in the TES200 and TES100 groups was less impaired. The ON firing rate in the TES200 group was significantly smaller compared with that in normal controls; meanwhile, no significant difference was found between the TES200 group and normal controls in terms of the OFF firing rate. In the TES100 group, the rescuing effects were less effective because both ON and OFF responses were significantly impaired compared with normal controls. (ANOVA analysis followed by post hoc test; n = 10, *P < 0.05, #P < 0.01 for differences compared between; All values represent mean ± SEM.)

Signal-to-Noise Ratio and Synaptic Efficacy

The signal-to-noise ratio (SNR = average driven firing rate/spontaneous background firing rate^26,31,32) was calculated to analyze the efficiency of visual signal transmission. Compared with normal controls, the impaired light-induced response and spontaneous hyperactivity collectively contributed to decreased SNR values in the sham group (sham versus normal controls, P < 0.01, n = 10; Table 2). Nevertheless, the SNR value in the TES100 group was significantly larger than that in the sham group (sham versus TES100; P < 0.01, n = 10). Moreover, the SNR value in the TES200 group was at least 2-fold larger than that in the TES100 group (TES200 versus TES100, P < 0.01; n = 10), and 16-fold larger than that in the sham group (TES200 versus sham, P < 0.01; n = 10), indicating that the RGCs in the TES200 groups could transmit visual signals much more reliably and economically. 

Table 2

 

The Signal-to-Noise Ratio (SNR) of Recorded Retinas (Spikes/s: Mean ± SE; n = 10 per Group)

 

The Mechanisms Underlying the TES-Induced Protective Effects

The mRNA expression levels of four apoptotic-associated genes, including Bax, Bcl-2, Calpain-2, and Caspase-3, were assessed by qRT-PCR. After MNU administration, the expression of these apoptotic-associated genes was upregulated compared with normal controls (sham versus normal control, P < 0.01; n = 10, Fig. 8). In TES-treated groups, the expression levels of Bax and Calpain-2 were significantly lower than those in the sham group (TES100 versus sham, P < 0.01; TES200 versus sham, P < 0.01, n = 10). Conversely, the expression level of Bcl-2 in TES-treated groups was higher than that in the sham group (TES100 versus sham, P < 0.01; TES200 versus sham, P < 0.01, n = 10). It was noteworthy that the expression level of Caspase-3 in the TES-treated groups was not significantly different from that in the sham group (TES100 versus sham, P > 0.05; TES200 versus sham, P > 0.05, n = 10). These findings suggested that Bax, Bcl-2, and Calpain-2, rather than Caspase-3, were involved in the TES-induced protective effects against MNU toxicity. 

Figure 8

 

The mRNA expression levels of Bax, Bcl-2, Calpain-2, and Caspase-3 were assessed by qRT-PCR. After MNU administration, the mRNA levels of these apoptotic-associated genes were all upregulated. In TES-treated groups, the mRNA levels of Bax and Calpain-2 were significantly lower compared with those in the sham group; meanwhile, the mRNA levels of Bcl-2 in TES-treated groups was higher than in the sham group. It was noteworthy that the mRNA levels of Caspase-3 in TES-treated groups were not significantly different from those in the sham group. Moreover, the mRNA levels of both BDNF and CNTF in TES-treated groups were significantly upregulated compared with normal controls. Particularly, the mRNA levels of BDNF and CNTF in the TES200 group were significantly higher than those in the TES100 group. Meanwhile, the mRNA levels of BDNF and CNTF in the sham group did not change significantly compared with normal controls. (ANOVA analysis followed by post hoc test; n = 10, *P < 0.05, #P < 0.01 for differences compared between; all values represent mean ± SEM.)

Additionally, the mRNA level of two neurotrophic factors, BDNF (brain-derived neurotrophic factor) and CNTF (ciliary neurotrophic factor), was quantified by qRT-PCR. The mRNA levels of both BDNF and CNTF in TES-treated groups were significantly upregulated compared with normal controls (TES100 versus normal controls, P < 0.01; TES200 versus normal controls, P < 0.01, n = 10; Fig. 8). In particular, the mRNA levels of BDNF and CNTF in the TES200 group were significantly higher than those in the TES100 group (TES100 versus TES200, P < 0.01; n = 10). Moreover, the mRNA levels of BDNF and CNTF did not change significantly in the sham group compared with normal controls. These findings suggested that neurotrophic factors BDNF and CNTF collectively contributed to the TES-induced protective effects. 

Discussion

Electrical stimulation alters the electrical charge balance of photoreceptors, thereby exerting beneficial effects on the retina.^23,33 Experimental and clinical studies have demonstrated protective effects of TES against traumatic- or genetic-induced retinal degeneration.^34–37 Retinitis pigmentosa is a heterogeneous group of retinal diseases characterized by progressive photoreceptor degeneration and visual disturbances. Results from the present study indicate that TES is capable of ameliorating the photoreceptor degeneration in MNU-administered mice, a pharmacologically induced RP model with rapid progressive dynamics. It is possible that repetitive and mild electrical stimulation is needed to induce a sustained neuroprotective effect on the microenvironment of degenerative retinas. Therefore, in the present study, the MNU-administered mice received three courses of successive TES treatments. The results showed that 200-μA TES was more effective than 100-μA TES. Beneficial effects on the regional retina were evaluated by topographic measurements to quantify the therapeutic efficiency: Central retinas were more effectively preserved by the TES therapy. 

The present study provides a novel example of integrating topographic technologies for evaluating therapeutic efficiency. Photoreceptor degeneration always progresses disproportionately in hereditary or pharmacologically induced RP models.^5,38 Therefore, the topographic characteristics of photoreceptor degeneration are crucial for therapeutic evaluations: The anticipated protective effect is directly affected by the retinal position selected for examination. Without the supports of topographic technologies, the effectively preserved zone might be readily overlooked, resulting in an underestimated therapeutic effect, or in contrast, a regional effect might be mistaken for global efficiency, giving rise to positive errors.²⁹ In the present study, TES-induced effects on the regional retina were systemically measured using topographic methods. Both the MEA and morphologic measurement suggested that the photoreceptors in the central retina were more efficiently preserved compared to the peripheral and midperipheral regions. Intriguingly, another study using light-damaged retinas also reported that the central photoreceptors were better rescued by TES treatment. It is highly possible that TES stimulates one portion of the retina more effectively rather than being uniform. The comparative sensitivities of regional photoreceptors to TES might be due to the asymmetric distribution of low-density current because it tends to go through the vitreous via a relatively low impedance path such as the optic nerve, which is located in the central retina.¹⁶ Moreover, the photoreceptors in mouse retina are not evenly distributed, and each subtype forms a gradient across the retinal regions.^30,39 The relative number and distribution of photoreceptor populations across retinal regions might be correlated with the regional differences in protective effects. Therefore, disproportional neuronal counts across retinal regions might mask the effects seen in different regions and act as an influencing factor. Further experiments are needed to confirm this possibility. 

For the first time, TES-induced effects on the visual signal transmission of inner retinal circuits were explored. As the ultimate grade of retinal neurons, RGCs collect visual signals from noisy presynaptic inputs and convert them into spikes, which are considered the standard phenotype of visual information that is eventually conveyed to occipital cortex.^31,32 The success of any therapeutic strategy ultimately depends on the functional integrity of RGC outputs.^40,41 Retinal ganglion cell hyperactivity exists in a variety of retinal degenerations, including RP, and results in significantly altered electrophysiological RGC properties.^42–45 Retinal ganglion cell hyperactivity is hazardous because it could add undesirable noise to the communication between eye and brain and eventually affect the authentic effectiveness of rescued visual function in animal models or patients. Our MEA results suggest that RGC hyperactivity in MNU-administered retinas is significantly restrained by TES therapy, suggesting minimized background noise and optimized dynamic properties for RGCs. It has been proposed that RGC hyperactivity could be attributed to a negative feedback loop between a bipolar cell and an amacrine cell that exhibit oscillations in membrane potential.⁴⁶ Whether the oscillations are critically tuned in the TES-treated retina remains to be shown by further tests with whole-cell recordings. Meanwhile, a substantial proportion of the light-induced RGC signal is also rescued by TES therapy: Both the ON and OFF visual signal pathway are retained in these treated retinas. These results verify that the basic configurations of signal transmission pathways are efficiently preserved by TES therapy. Moreover, a lower SNR reflects a decrease in the synaptic signaling efficacy and loss of retinal responsiveness to discriminate the system signal from background noise.^31,32 Compared with the sham controls, the significantly increased SNR in TES-treated retinas could be attributed to the improved robustness of light-induced response and the simultaneously attenuated spontaneous response, indicating that TES treatment partly restores deteriorated signaling efficiency and enhances the ability of RGCs to reliably and economically encode visual signals. 

Apoptosis is recognized as the final common death pathway in all RP phenotypes, although tremendous genetic heterogeneity exists in this disorder.^47–49 Bax, Bcl-2, Calpain-2, and Caspase-3 are all critical factors in the photoreceptor apoptotic cascades.^50–52 Our study suggests that TES rectifies abnormalities in the apoptotic cascade by altering Bcl-2 and Bax expression, indicating that TES could be developed into a promising and general therapeutic strategy against multiple RP phenotypes. It is especially noteworthy that the expression of Caspase-3, a key executioner of the classical caspase-dependent apoptotic pathway, remains unaffected after TES treatment. Conversely, the downregulation of Calpain-2, a calcium-dependent cysteine protease, suggests that TES alleviates the calcium overload that is considered as an initiator of photoreceptor apoptosis.⁵¹ Transcorneal electrical stimulation could also promote the endogenous release of neurotrophic factors, simultaneously enhancing the intrinsic sensitivity of retinal neurons to these factors.^23,24 Our results show that TES regulates the expression of BDNF and CNTF mRNA, which are crucial for activating the cellular survival system and neutralizing the retinal microenvironment.^16,53 

Although the safety and efficacy of TES might be easily verified and more readily acceptable in RP animal models, it remains challenging to prove these virtues in RP patients. The natural course of disease progression in RP patients is highly variable as the tremendous heterogeneity implies in the initiating mutation.^1,3 These heterogeneously genetically determined degenerative processes make RP inherently difficult to treat. Consequently, the TES-induced benefits in RP retinas should be validated in various RP retinas to evaluate efficacy before clinical application. Transcorneal electrical stimulation could promote photoreceptor survival and preserve the retinal function in the RP models with a relatively more slowly progressing time course, and it has been hypothesized that TES is futile with respect to delaying photoreceptor death in degenerative retinas with rapid progressive dynamics.²⁴ However, our work verifies that MNU-administered photoreceptors that would otherwise die within 1 week are efficiently rescued by TES. Admittedly, the etiology of MNU-induced photoreceptor degeneration is far from what is present in models of RP, although the pathologies are somewhat similar. The mechanism underlying MNU-induced photoreceptor death is the principal alkylation of DNA, depending on the action of Aag.⁴ On the other hand, most forms of RP involve mutations in retina-specific genes such as those involved in phototransduction, protein trafficking, and phagocytosis leading to gradual demise of rods. The novel and surprising observation that cones are also killed by MNU injection does not correspond to cone degeneration in RP either, since demise of this population occurs in a secondary wave of rod death.^5,9 Therefore, although the outcome of MNU administration partially resembles RP (i.e., widespread rod death), the mechanistic underpinnings and kinetics are very different. These considerations are crucial in any study that proposes clinical testing of TES treatment. 

In conclusion, regional photoreceptors in MNU-administered retinas exhibit comparative TES sensitivities. The present topographic study indicates that TES counteracts MNU-induced photoreceptor degeneration and rectifies abnormalities in inner visual signal pathways. Further studies are needed to determine the optimal parameters and long-term stability of TES. 

Acknowledgments

Supported by the National Key Development Program for Basic Research (973 Program; No. 2013CB967001), National Natural Science Foundation of China (No. 81600767), and China Postdoctoral Science Foundation (No. 2015M582852). 

Disclosure: Y. Tao, None; T. Chen, None; Z.-Y. Liu, None; L.-Q. Wang, None; W.-W. Xu, None; L.-M. Qin, None; G.-H. Peng, None; Y.-F. Huang, None