Functional and structural evaluation of lamotrigine treatment in rat models of acute and chronic ocular hypertension

Experimental Eye Research 115 (2013) 47e56

Contents lists available at SciVerse ScienceDirect

Experimental Eye Research journal homepage: www.elsevier.com/locate/yexer

Functional and structural evaluation of lamotrigine treatment in rat models of acute and chronic ocular hypertension Shai Sandalon a,1, Birte Könnecke b,1, Hani Levkovitch-Verbin c, Mikael Simons b, Katharina Hein b, Muriel B. Sättler b, Mathias Bähr b, Ron Ofri a, * a Koret School of Veterinary Medicine, The R.H. Smith Faculty of Agriculture, Food and Environment, The Hebrew University of Jerusalem, P.O. Box 12, Rehovot 76100, Israel b Department of Neurology, University of Göttingen Medical School, Göttingen, Germany c Sam Rothberg Ophthalmic Molecular Biology Laboratory, Goldschleger Eye Institute, Sheba Medical Center, Sackler Faculty of Medicine, Tel-Aviv University, Tel-Hashomer, Israel

a r t i c l e i n f o

a b s t r a c t

Article history: Received 12 March 2013 Accepted in revised form 17 June 2013 Available online 28 June 2013

Voltage gated sodium channels (Nav), are proposed mediators of neuronal damage in ischemic and excitotoxicity disease models. We evaluated the neuroprotective effects of lamotrigine, a Nav blocker, in the acute and chronic rat ocular hypertension models. Additionally, expression of the main Nav subtypes in the optic nerve (ON) was assessed to test whether their upregulation plays a role in the pathogenesis of ocular hypertension induced optic neuropathy. Unilateral intraocular pressure (IOP) elevation was induced for 60 min (80 mmHg) and 14e21 days (670e859 mmHg*day) in the acute and chronic models, respectively. Lamotrigine was administered at dosages of 10 mg/kg twice daily and 12.5 mg/kg once daily in the acute (n ¼ 9) and chronic (n ¼ 11) trials, respectively. Treatment began 2 days prior to IOP elevation until sacrifice. Outer and inner retinal function was evaluated with dark- and light-adapted flash electroretinography and pattern electroretinography, respectively, 6 and 14 days post acute IOP elevation and 13, 28 and 48 days post chronic IOP elevation. Retinal ganglion cell and axon densities and inflammatory reaction were evaluated through Fluorogold, Bielschowsky’s silver impregnation and ED1 labeling respectively. Immunohistochemistry for Nav1.1, 1.2 and 1.6 was performed in ONs of untreated rats 7 and 15 days post IOP elevation in the acute model and after 7, 28 and 50 days in the chronic model. In the acute model, no differences were found in the a-wave amplitudes between lamotrigine-treated and vehicle-treated rats. B-wave amplitudes decreased by 40e66% in both treatment groups 6 days post IOP elevation, with no significant difference between groups (p ¼ 0.38). However, a partial recovery of b-wave amplitudes was found in lamotrigine-treated rats between day 6 and day 14 post procedure (p < 0.05). No differences were found in any other parameter tested in this model. Similarly, lamotrigine treatment did not result in any beneficial effect in structural parameters of the chronic model. Functional evaluation of this model was inconclusive due to super-normal values in the hypertensive eyes. Upregulation of Nav1.1 and 1.2 expression was found in both models, beginning by day 7; an increase of the former continued in a time-dependent manner in the chronic model. Nav1.6 labeling was inconclusive. In conclusion we found lamotrigine treatment to be mostly ineffective in both acute and chronic ocular hypertension models. Ó 2013 Elsevier Ltd. All rights reserved.

Keywords: sodium channels lamotrigine glaucoma electroretinography neuroprotection retinal ganglion cell optic nerve

1. Introduction

Abbreviations: Nav, voltage gated sodium channel; LTG, lamotrigine; IOP, intraocular pressure; RGC, retinal ganglion cells; ON, optic nerve; ERG, electroretinogram; FERG, full field ERG; PERG, pattern ERG; IHC, immunohistochemistry. * Corresponding author. Tel.: þ972 54 8820523. E-mail addresses: [email protected] (H. Levkovitch-Verbin), simons@ em.mpg.de (M. Simons), [email protected] (R. Ofri). 1 These authors made equal contribution. 0014-4835/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.exer.2013.06.018

Glaucoma is a neurodegenerative disease of the inner retina and optic nerve (ON), which leads to progressive visual field loss and possible blindness due to apoptosis and degeneration of retinal ganglion cells (RGC) and their axons. Although mainstream glaucoma therapy is focused on lowering the elevated intraocular pressure (IOP), much research is devoted to developing drugs that would provide direct protection to the neural

48

S. Sandalon et al. / Experimental Eye Research 115 (2013) 47e56

tissue. Numerous substances were suggested as candidates for neuroprotective therapy, based on inhibiting various mechanisms of RGC degeneration and apoptosis, or promoting their survival. Voltage gated sodium channels (Nav), which normally function in the initiation and propagation of action potentials, are proposed to be mediators of neuronal damage by facilitating excessive depolarization, intracellular sodium overload and presynaptic release of glutamate (Bano et al., 2005; Nikolaeva et al., 2005; Stys, 2005; Tekkök et al., 2007). The presence and function of Nav in the ON and retina was localized to RGC soma and axons (Boiko et al., 2003), subsets of amacrine cells, horizontal cells (Mojumder et al., 2007) and 2 types of cone bipolars (Cui and Pan, 2008). Nav are also present in Müller cells where they function in volume regulation (Linnertz et al., 2011). Nav1.6, 1.2 and 1.1 are the main subtypes found in the ON and retina. Lamotrigine (LTG) is a Nav channel blocker approved for the treatment of epilepsy, acting by blocking transient, fast inactivating sodium currents (Spadoni et al., 2002) and decreasing glutamate release (Shuaib et al., 1995). LTG was shown to protect excitable neurons in several models (Bechtold et al., 2006; Crumrine et al., 1997). LTG decreased glutamate-induced excitotoxicity in chick retina ex vivo (Pisani et al., 2001) and protected rabbit retina from apoptotic effects of intravitreal silicone oil injection (Guizzo et al., 2008) in vivo. Guizzo et al. (2005) found LTG to partially protect the rat inner retina from ischemia/reperfusion damage at the histopathological level (vacuolization and cell body densities), though the study did not differentiate RGCs from displaced amacrines. A previous study on the neuroprotective effects of LTG in a chronic ocular hypertension rat model failed to find a beneficial structural effect (Marina et al., 2012). However, due to the great promise shown by the drug in other ischemic and excitotoxic models, we aimed to further this investigation by studying LTG in two ocular hypertension models, using more extensive methodology. Our goal was therefore to study both functional and structural effects of LTG treatment in the acute and chronic ocular hypertension induced optic neuropathy rat models. Furthermore, based on observed changes in Nav channel expression in ON axons, macrophages and microglia of experimental allergic encephalomyelitis (EAE) model mouse (Craner et al., 2003, 2005) we evaluated similar changes in two rat ocular hypertension induced optic neuropathy models. We hypothesized that upregulation of Nav channels in hypertensive eyes would further increase the efficacy of LTG treatment.

Table 1 Number of rats in each procedure. LTG ¼ lamotrigine. Number of days indicate day of sacrifice (post IOP elevation).

2. Methods

2.4.1. Experiment design of LTG neuroprotection trial The sequence of procedures is depicted in Fig. 1. LTG and vehicle (VEH) administration began 2 days before induction of IOP elevation and continued thereafter, resulting in both a preventive and therapeutic administration, respectively.

We used two rat ocular hypertension models to evaluate: 1. LTG efficacy as a neuroprotective agent by studying electroretinography (ERG), RGC counts, axon density, and microglia/ macrophage involvement. 2. Changes in ON Nav1.1, 1.2 and 1.6 expression using histopathology and immunohistochemistry (IHC).

IOP elevation LTG efficacy (Study I)

Acute Chronic

Changes in Nav expression (Study II)

Acute Chronic

LTG vehicle LTG vehicle 7 days 15 days 7 days 28 days 50 days

ERG

Fluorogold labeling

Optic nerve structural analysis

8 8 11 8

8 8 11 10

5

5

9 8 11 10 4 9 13 14 6

2.2. Acute IOP elevation The anterior chamber of a randomly chosen eye (the contralateral eye served as normotensive control) was cannulated with 30-gauge needle connected to a 0.9% saline reservoir hanging 150 cm above the eye. IOP elevation lasted 60 min and was assessed with a rebound tonometer before, 3 and 57 min after induction of IOP elevation and by observing blanching of the limbal plexus. Visualization of reperfusion to the limbal plexus within 30 s following cannula removal served as an indication of return to physiological IOP. 2.3. Chronic IOP elevation Limbal plexus laser photocoagulation was used to block venous drainage from the left eye of anesthetized rats as described previously (Levkovitch-Verbin et al., 2002b). In the 1st session, 37e63 cautery pulses of 400 mW were applied. A 2nd photocoagulation session (18e44 cautery pulses, 250e300 mW) was performed 7 days later, except for 4 cases where only one laser session was performed due to marked IOP elevation and blanched appearance of the limbal plexus. IOP was measured with a Tonopen prior to photocoagulation, immediately after the procedure and then on days 1, 7, 14, 21, 27, 33 and 48 post IOP elevation. 2.4. LTG neuroprotection study

2.4.2. LTG administration VEH was 0.5e0.75% methylcellulose in PBS. LTG was suspended in VEH yielding 3.5% and 1.3% suspensions for acute and chronic

2.1. Animals 10e12 weeks old Lewis/SsNHsd and Wistar/HsdBrlHan (Harlan Laboratories Inc, Israel) male rats were used for the acute and chronic IOP elevation procedures, respectively. The number of rats that were used in each procedure is summarized in Table 1. Rats were reared in 12 h light cycle and controlled temperature. All experimental procedures were in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, and the guidelines of the Institutional Animal Care and Use Committee. An intraperitoneal (IP) injection of ketamine and xylazine were used for anesthesia in all procedures.

Fig. 1. Timetable of LTG procedures. A) Acute IOP elevation model. B) Chronic IOP elevation model. ERG ¼ electroretinogram; FG ¼ Fluorogold injection; IOP ¼ intraocular pressure elevation procedure; LTG ¼ beginning of daily lamotrigine administration.

S. Sandalon et al. / Experimental Eye Research 115 (2013) 47e56

trials, respectively (methylcellulose and LTG both from Kemprotec Limited, Middlesbrough, UK). In the acute IOP elevation model 10 mg/kg were administered subcutaneously twice daily. In the chronic IOP elevation model 15 mg/kg were administered IP once daily for two days before the photocoagulation procedure, and subsequently 12.5 mg/kg were administered once daily IP for the rest of the trial. The lower dose in the chronic IOP trial was given in attempt to avoid potential adverse effects associated with drug accumulation. Control animals in both models were treated with VEH only. 2.4.3. Electroretinography Dark adapted intensity-response series of full field ERG (FERG), light-adapted FERG and pattern ERG (PERG) were recorded successively during the same session. 2.4.3.1. Intensity-response series. Rats were dark adapted overnight. Both eyes were recorded in random order while the contralateral eye was covered. Local anesthetic was applied and conductance was maintained with 1.4% hydroxymethylcellulose resulting in impedance <10 KU. The active electrode was a 0.5 mm diameter gold loop wire placed on the cornea. Reference and ground needle electrodes were placed subcutaneously in the lateral cantus and tail base, respectively. Recordings were conducted using the RetiPort 32 system (Roland consult, Brandenburg, Germany). Bandpass filter was 1e 300 Hz. The stimuli consisted of 6 and 3 steps of increasing luminance in the acute and chronic models, respectively, delivered by a miniganzfeld. Flash luminance ranged 0.0096e9.6 cd*sec/m2. Interstep intervals ranged 1e5 min. At each step, 3e10 flashes, presented at intervals of 5e10 s, were averaged. 2.4.3.2. Light adapted ERG. Light adaptation was delivered by the miniganzfeld (8 min, 25 cd/m2). Stimulus was a 3 cd*sec/m2 flash on a 25 cd/m2 background (1.3 Hz, 20 responses averaged). Other settings were identical to the intensity-response series. Only the bwave was evaluated because a-wave was at noise level. 2.4.3.3. PERG. Acute model. Stimuli were dark and light vertical bars with 50% duty cycle that alternated at 2.1 reversals per second (rps) (Viking Quest, Nicolet Biomedical Inc., Madison, Wisconsin, USA). These were displayed on a CRT monitor (NIC 1015, Nicolet Biomedical Inc., Madison, Wisconsin) located 23 cm from the eye, aligned with the globe (w45 ). Bandpass was 1e100 Hz and 200 sweeps were averaged at each spatial frequency. Responses to four spatial frequencies of progressively increasing bar size e 0.250, 0.125, 0.062 and 0.031 cycles per degree (cpd) e were sequentially recorded. Chronic model. Square checkerboard alternating pattern was delivered by a specially modified direct ophthalmoscope (Sandalon and Ofri, 2009) at a frequency of 6.1 rps. Bandpass was 2e250 Hz and 100 sweeps were averaged. Five spatial frequencies e 0.368, 0.184, 0.092, 0.046 and 0.023 cpd e were sequentially tested. PERG sessions were sequentially recorded after FERG sessions. PERG amplitude was measured from 1st positive peak to the following trough. The use of different PERG stimulators in the two models was because the direct ophthalmoscope, which is our preferred stimulator, was not available during the acute hypertension experiment. The use of a different stimulator in each study required slight modifications of the recording protocol. 2.4.4. RGC labeling The procedure is described in detail elsewhere (Blair et al., 2005). Briefly, RGC were retrogradely labeled by bilateral stereotaxic injection of 5% Fluorogold dissolved in 0.9% saline into the superior colliculi. Coordinates were 6.8 mm caudal to the bregma, 1.3 mm

49

lateral to the midline. In each of 3 depths e 2.7, 3.2 and 3.6 mm e 1.5 ml were injected over 1 min. Fluorogold was injected 7 days prior to IOP elevation in the acute trial (Fig. 1A) to ensure complete RGC labeling before any axon damage had occurred, thus allowing evaluation of true RGC death rather than axonal damage. In the chronic trial, Fluorogold was injected 10 days before sacrifice (Fig. 1B) to avoid time-dependent degradation of the dye had it been administered in the beginning of such a long trial. 2.4.4.1. RGC counting. After euthanasia with an overdose of IP pentobarbitone, eyes were enucleated and fixed in 4% PFA for at least 1 h. The globes were hemisected, and the retinas were flatmounted on glass slides and sealed with a mounting medium. Visualization and imaging were performed with a fluorescent microscope and a matching filter with excitation wavelength of 350  25 nm (11006v3 gold filter, Nikon, Tokyo, Japan) at 100 magnification. Images were processed with PhotoFiltre 6.3.1 software. RGC counting was done by sampling and averaging counts of 12 squares of 62,500 mm2 each, counting 3 regions e central, middle and peripheral e in each retinal quadrant. 2.4.5. Histopathology and IHC Both ONs were dissected, fixed over night in 4% PFA, stored in PBS until processed and paraffin-embedded. Histopathological and IHC examinations were performed on 0.5 mm thick longitudinal sections. All images were evaluated using Axiovision 4.2 software (Zeiss, Oberkochen, Germany) by investigators blinded to the experiments. 2.4.5.1. Axon density. Bielschowsky’s silver impregnation was used to define axonal density. For quantification pictures were taken in 100 magnification using an Axiocam MR (Zeiss, Oberkochen, Germany) and axons crossing a 50 mm long line perpendicular to the course of axons were counted. Three different longitudinal sections were examined in each ON, and axons were counted at nine different locations in each of these sections. 2.4.5.2. ED1-IHC. IHC for ED1 was performed to evaluate infiltration by macrophages and microglia. After blocking with 10% horse serum in PBS for 10 min at room temperature, sections were incubated with 1 mg/ml ED1 antibody (AbD Serotec; MCA341R) in blocking buffer over night at 4  C. Immunoreactivity was visualized using 2.5 mg/ml biotinylated horse anti-mouse antibody (Vector Laboratories, Inc.; BA-2001) and biotineavidin technique. Control sections were incubated without primary antibody. ED1 positive cells were evaluated using a semi-quantitative method, based on a percentage score. Each ON was examined in three different longitudinal sections, and each section was divided schematically into 20 equal parts and screened for ED1 positive cells. A positive staining in one part received a score of 5%. The number of parts (expressed as percentage area) of the entire ON section infiltrated by ED1 positive microglia and macrophages were assigned the following score: 0 e no positive cells, 1 e less than 30% of ON section, 2e30% e 60% of ON section, 3 e more than 60% of the ON section infiltrated with ED1 positive cells. 2.5. Nav expression in optic nerves of acute and chronic hypertensive eyes For Nav1.1, Nav1.2 and Nav1.6 detection paraffin sections were blocked with TBSTNav (0.1% Triton X-100 in TBS) containing 10% donkey serum, 2% bovine serum albumin (BSA) and 0.02% sodium azide for 2 h at RT, followed by an incubation with 0.8 mg/ml Nav1.1 antibody (Alomone Labs; ASC-001), 1.1 mg/ml Nav1.2 antibody (Alomone Labs; ASC-002) and 1.1 mg/ml Nav1.6 antibody (Alomone

50

S. Sandalon et al. / Experimental Eye Research 115 (2013) 47e56

Labs; ASC-009), respectively, in blocking buffer overnight at 4  C. Immunoreactivity for fluorescence microscopy was visualized using 1.5 mg/ml donkey anti-rabbit Cy3-conjugated antibody (Jackson ImmunoResearch; 711-165-152). Washing steps were performed six times 5 min each in TBSTNav . For double staining of Nav channels with the axonal marker NF200 a sequential protocol was used. Sections were blocked with TBSTNF200 (0.3% Triton X-100 in TBS) containing 10% goat serum and 2% BSA for 1 h at RT. NF200 antibody (dilution 1:500; SIGMA; N0142) was incubated for 72 h at 4  C. For fluorescence microscopy 4 mg/ml goat anti-mouse Alexa Fluor 488-conjugated antibody was used. Nuclei were counterstained with DAPI. The semi-quantitative scale was the same as in the ED1 evaluation. 2.6. Statistics ANOVA with repeated measures, one-way ANOVA followed by Bonferroni’s multiple comparison, Student’s t-test, ManneWhitney and chi-square tests were used as indicated in the text. When ANOVA with repeated measures revealed significant interaction between variables, a t-test with the p-value corrected to the Bonferroni criterion was performed to identify the outstanding variable. Error bars and ‘’ indicate standard deviation (SD). 3. Results 3.1. LTG neuroprotection 3.1.1. Acute model 3.1.1.1. IOP elevation. Mean IOP at baseline, 3 min and 57 min post cannulation are shown in Fig. 2A. Reperfusion to the limbal plexus was observed in all rats after the cannula was withdrawn. No significant differences were found between the LTG and VEH groups at any timepoint (ANOVA with repeated measures, p ¼ 0.68). 3.1.1.2. Functional effects of lamotrigine treatment 3.1.1.2.1. Intensity-response series. A representative intensityresponse series recorded in dark adapted, VEH- and LTG-treated rats 6 and 14 days post IOP elevation is shown in Fig. 3A. A-wave analysis was not performed for the two lowest luminance intensities (0.0096 and 0.03 cd*s/m2) due to near noise amplitude levels. A non significant attenuation (of 17e26%) in awave amplitude of hypertensive vs. normotensive eyes was found in both groups (ANOVA with repeated measures, p ¼ 0.076). A significant day*treatment group interaction (p ¼ 0.005) reflecting a different course of change in amplitude from day 6 to day 14 between treatment groups was seen. More specifically, at a flash intensity of 9.6 cd*s/m2, a significant increase in amplitude ratio was found in LTG-treated rats (t-test p < 0.0083 corrected to Bonferroni criterion, Fig. 3B). However, no differences in amplitude ratios of hypertensive vs. control eyes were found between the treatment groups (ANOVA with repeated measures, p ¼ 0.70). Six days post IOP elevation there was a 40e66% attenuation of bwave amplitude in both groups in response to most flash intensities (Fig. 3C, ANOVA with repeated measures, p < 0.01). This trend increased with flash intensity in both groups. After 14 days the magnitude of attenuation remained similar in VEH-treated rats (39e55% decrease) while milder attenuation was noticed in LTGtreated rats (14e34% decrease). This relative improvement in LTG-treated rats is reflected by a significant day*treatment group interaction (ANOVA with repeated measures, p ¼ 0.02, and a significant p value of the t-test corrected to Bonferroni criterion, p < 0.0083) at 3.0 cd*s/m2 (Fig. 3C). Yet, there were no significant differences in the amplitude ratios between the treatment groups (ANOVA with repeated measures, between groups effects, p ¼ 0.38).

Fig. 2. Mean IOP in acute (A) and chronic (B) hypertension models. No significant differences between the groups were found (p ¼ 0.68 and 0.08 in the acute and chronic models, respectively).

3.1.1.2.2. Light adapted responses. Acute IOP elevation resulted in severe attenuation (89e66%) of light adapted b-wave in both LTG- and VEH-treated rats (Fig. 4, paired t-test, p < 0.004). There were no significant differences between treatment groups 14 days post procedure (ManneWhitney test due to small sample size and non-normal distribution, p ¼ 0.67). 3.1.1.2.3. PERG. In most rats (from both groups) a PERG signal could not be detected in the hypertensive eye at any timepoint. The response to the largest check size stimulus 6 days post procedure is summarized in Fig. 5. The calculated average amplitude for each group takes into account responding rats only; otherwise the average would be near zero for the hypertensive eyes. At both 6 and 14 days, a higher percentage of LTG-treated rats responded, but this difference was not significant (chi square test, p ¼ 0.15; this test was chosen to analyze responders vs. non-responders in each group). 3.1.1.3. Structural effects of lamotrigine treatment 3.1.1.3.1. RGC counting. Mean RGC densities of 2374  260 and 2094  260 cells/mm2 were counted in normotensive control eyes of the LTG- and VEH-treated rats, respectively. Relative RGC survival, measured as hypertensive/normotensive eye RGC count ratio is shown in Fig. 6A. No significant differences were found between the groups (Student’s t-test, p ¼ 0.55). 3.1.1.3.2. Histopathology and IHC. On day 15 after acute IOP elevation, the semi quantitative examination of ED1 positive microglia/macrophages in hypertensive ONs revealed significant increase in the score of both LTG- and VEH-treated groups (oneway ANOVA, both p < 0.01) but no significant difference between the treatment groups (one-way ANOVA, p > 0.05; Fig. 6A). Given

S. Sandalon et al. / Experimental Eye Research 115 (2013) 47e56

51

Fig. 3. FERG in the acute IOP elevation model. A) Representative traces from VEH-(left panel) and LTG (right panel)-treated rats, 6 and 14 days post IOP elevation. Three traces are shown for each rat: left e responses from cannulated eye 6 days post procedure, middle e responses from the same eye 14 days post procedure; right e healthy control eye of the same rat 14 days post procedure. B þ C) Hypertensive/control eye ratios of a- and b-wave amplitudes, 6 and 14 days post acute IOP elevation. B) No differences in a-wave amplitude ratios of hypertensive vs. control eyes were found between the treatment groups (ANOVA with repeated measures, p ¼ 0.70). However, a significant recovery from day 6 to day 14 was found in LTG-treated rats at 9.6 cd*s/m2 (ANOVA with repeated measures, day*treatment interaction, p ¼ 0.005; and t-test with p value corrected to Bonferroni criterion, p < 0.0083). C) No significant differences in the normalized b-wave amplitudes between treatment groups were found (ANOVA with repeated measures, between groups effects, p ¼ 0.38). After 14 days the magnitude of attenuation remained similar in VEH-treated rats (39e55% decrease) while milder attenuation was noticed in LTG-treated rats (14e34% decrease). This relative improvement in LTG-treated rats is reflected by a significant day*treatment group interaction (ANOVA with repeated measures, p ¼ 0.02 and a significant p value of the t-test corrected to Bonferroni criterion, p < 0.0083) at 3.0 cd*s/m2.

that we could not detect any significant axonal loss in hypertensive ONs of either treatment group, there was no detectable effect of treatment (one-way ANOVA, p > 0.05; Fig. 6A). 3.1.2. Chronic model 3.1.2.1. IOP elevation. The area under the IOP curves for days 0e21 were 859 and 670 mmHg*day in VEH- and LTG-treated rats, respectively. This difference was not significant (Fig. 2B, Manne Whitney Test due to non-normal distribution, p ¼ 0.08). 3.1.2.2. Functional effects of lamotrigine treatment. Unexpectedly, 48 days after IOP elevation, FERG amplitudes of hypertensive eyes were higher than in control normotensive eyes (ANOVA with repeated measures, hypertensive vs. normotensive eye effect within subjects, p ¼ 0.001). There was no difference between treatment groups in amplitude ratios of the a-wave (ANOVA with repeated measures, p ¼ 0.24) or b-wave (p ¼ 0.074) (Fig. 7). No attenuation of PERG amplitudes in hypertensive eyes of either group was found 48 days post IOP elevation at any spatial frequency tested (ANOVA with repeated measures, p ¼ 0.46).

3.1.2.3. Structural effects of lamotrigine treatment 3.1.2.3.1. RGC counting. Mean RGC density in control eyes of the LTG- and VEH-treated rats was 1850  127 and 1916  180 cells/ mm2, respectively. RGC survival ratio was similar in both groups (Fig. 6B, ManneWhitney Test due to non-normal distribution, p ¼ 0.81). 3.1.2.3.2. Histopathology and IHC. The evaluation of ED1 positive microglia and macrophages at day 50 post chronic IOP elevation showed a significant increase in the score of hypertensive ONs from both treatment groups (one-way ANOVA, p < 0.001), but no difference between the treatment groups (one-way ANOVA, p > 0.05; Fig. 6B). We also found a decrease of axon density in both treatment groups, but once again no differences between groups (one-way ANOVA, p > 0.05; Fig. 6B). 3.2. Nav expression results IOP levels in rats used for evaluation of Nav expression were similar to those of the LTG neuroprotection trial in both acute and chronic models (data not shown).

52

S. Sandalon et al. / Experimental Eye Research 115 (2013) 47e56

Fig. 4. A) Representative traces of light adapted FERG from a LTG-treated rat, 6 days post acute IOP elevation. B) Light adapted b-wave amplitudes 14 days post acute IOP elevation. While strong and significant attenuation of the signal was measured in hypertensive eyes of both groups compared with contralateral eyes (Paired t-test, p < 0.004), there were no differences between responses from hypertensive eyes of LTG- and VEH-treated rats (ManneWhitney test, p ¼ 0.665).

3.2.1. Acute model We detected increased fiber-like Nav1.1 expression 7 days post procedure in ONs of hypertensive eyes from the acute model. Nav1.1 signals were seen in less than 30% of the total area of the hypertensive ON sections. A week later, a further, insignificant increase (one-way ANOVA, p > 0.05) was found (Fig. 8B). We did not observe any IHC staining for Nav1.1 in ONs of control eyes. Nav1.2 and 1.6 stainings were inconsistent and not conclusive. 3.2.2. Chronic model We detected significantly elevated Nav1.1 expression in nearly all ONs of hypertensive eyes at all timepoints compared to contralateral control ONs (one-way ANOVA, all p < 0.001). The positive staining increased significantly from 30% on day 7 to more than 60% on day 28 post IOP elevation (one-way ANOVA, p < 0.001). An additional, insignificant, increase was found on day 50 (Fig. 8C). No colocalization of Nav1.1 and the axonal marker NF200 was detectable (Fig. 8A). Positive IHC staining for Nav1.2 was found only in ONs from hypertensive eyes, while ONs of control animals did not stain for Nav1.2. The staining intensity in ONs of hypertensive eyes was highly variable, possibly due to either the fixation protocol or the polyclonal form of the antibody that leads to reactivity differences between different purchased batches. Therefore, we could not perform a statistical evaluation. Nevertheless, we could detect Nav1.2 expression in ONs from hypertensive eyes at 28 and 50 days

Fig. 5. PERG amplitudes in response to the largest (0.031 cpd) pattern stimulus 6 days post acute IOP elevation. A) Representative traces recorded from a VEH-treated (top 2 traces) and a LTG treated (lower 2 traces) rat. IOP ¼ cannulated eye. B) PERG amplitudes and percent of responders in response to 0.031 cpd stimulus, 6 days post IOP elevation. The calculated average amplitude for each group takes into account the responding rats only; otherwise the average would be near zero for hypertensive eyes. Higher percentage of LTG-treated rats produced a measurable signal in the cannulated eye, compared with control eyes but this finding was not significant (Chi square test, p ¼ 0.15).

post IOP elevation (Fig. 9A and B). This fiber-like Nav1.2 expression resembled the staining pattern of Nav1.1, and once again it was not visible in control ONs. Nav1.2 expression was independent of axon density, evidenced by double staining with axonal marker NF200 (Fig. 9). Co-localization of fiber-like Nav1.2 and axonal marker NF200 was not detected. IHC of polyclonal Nav1.6 antibody showed a highly variable quality of staining, and high intensity, unspecific background staining; therefore, no statistical evaluation was possible. 4. Discussion The main findings of this study are that LTG treatment does not confer protection of retinal function or structure in either the acute, transient IOP elevation model or the chronic intraocular hypertension model. There was a trend of insignificant improvement of retinal function in LTG-treated rats in the acute hypertension model, as evidenced by a partial recovery of FERG responses from days 6e14 (Fig. 3), and a higher ratio of PERG responders (Fig. 5B). Our findings that LTG treatment in the chronic ocular hypertension model has no protective effect are in agreement with the

S. Sandalon et al. / Experimental Eye Research 115 (2013) 47e56

53

Fig. 6. Effect of LTG treatment on structural parameters. From left to right: representative pictures of ED1 IHC, Bielschowsky silver staining and Fluorogold (FG) RGC labeling in ONs (ED1þBielschowsky) and retinas (FG) of VEH- and LTG-treated animals. A) Acute model. Elevation of IOP led to an increase of ED1 positive cells in the ONs. But, as indicated in the table, neither an anti-inflammatory nor an axon/RGC protective effect of LTG could be shown (p > 0.05). Note that VEH-treated rats had no axon loss, and therefore a protective effect is not expected in this parameter. B) Chronic model. A strong and significant increase in ED1 positive cells and high axonal & RGC loss can be seen in hypertensive retinas and corresponding optic nerves. Nevertheless, neither an anti-inflammatory nor an axon/RGC protective effect of LTG was detectable (p > 0.05).

recent study (Marina et al., 2012). On the other hand, treatment with phenytoin, another registered anti-epileptic Nav blocker, was shown to decrease RGC loss in a similar rat model (Hains and Waxman, 2005). This discrepancy might arise from the ability of phenytoin to block persistent sodium currents, a feature not attributed to LTG. However, Spadoni et al. (2002) found persistent sodium currents to be phenytoin (and LTG) resistant, so it is possible that other mechanisms may account for the differences between the drugs.

One outcome of the chronic model was unexpected and weakened the strength of the model e ERG parameters were similar or even higher in hypertensive eyes, compared with control eyes, regardless of treatment (Fig. 7). This was despite significant decrease in RGC count and axon density (Fig. 6B). We are not aware of any technical fault that might account for this discrepancy. However, translimbal photocoagulation causes a certain degree of uveitis. It is possible that this inflammation may cause a supernormal ERG. Indeed, Stanford et al. concluded that the

Fig. 7. FERG amplitudes, normalized to control eyes, 48 days post chronic IOP elevation. A) A-wave amplitude ratio of both VEH and LTG groups were significantly higher than 1, indicating supernormal amplitudes (ANOVA with repeated measures, p < 0.05). B) B-wave amplitude ratio. Again, hypertensive eyes showed higher amplitudes than controls as reflected by hypertensive/control ratios higher than 1 (ANOVA with repeated measures, p < 0.05). No significant differences of the a- and b-waves amplitude ratios were found between treatment groups (ANOVA with repeated measures, p ¼ 0.24 and 0.074, respectively).

54

S. Sandalon et al. / Experimental Eye Research 115 (2013) 47e56

Fig. 8. Evaluation of changes in Nav1.1 expression over time following acute and chronic IOP elevation. A) Fiber-like Nav1.1 expression (red; arrows) not colocalizing with axonal marker NF200 (green; arrowheads) in ONs of hypertensive eyes 7, 28 and 50 days post chronic IOP elevation. No Nav1.1 signals were visible in a contralateral control ON. B) Acute model. Seven days post IOP elevation, Nav1.1 expression was elevated in hypertensive ONs. Fifteen days after procedure, there was an insignificant increase in expression score (oneway ANOVA, p > 0.05). C) Chronic model. Seven days post IOP elevation, a significant increase in Nav1.1 expression was found in hypertensive ONs (one-way ANOVA, p < 0.001). A further significant increase (one-way ANOVA, p < 0.001) was found on day 28, and similar levels were found on day 50 post IOP elevation (p > 0.05).

supernormal ERG reported in rats in which S-retinal antigen was used to induce autoimmune uveitis was caused by an antibody against this antigen (Stanford and Robbins, 1988; Stanford et al., 1992). It is possible that during the prolonged chronic experiment some autoimmune or other inflammatory response may have developed, thus affecting our ERG results. The acute IOP elevation model, also known as ischemiareperfusion, is a widely used model, in which ischemia is the

major cause of apoptotic (Lagrèze et al., 1998) and/or necrotic (Rosenbaum et al., 2010) damage. Indeed, because LTG was found to be neuroprotective in ischemia (Calabresi et al., 2003) and energy depletion (Crumrine et al., 1997; Lagrue et al., 2007) models, we thought it more likely that of the two models we tested, LTG efficacy will be demonstrated in the acute model. However, only a trend of insignificant functional improvement was found in this model.

Fig. 9. Nav1.2 IHC in chronic hypertensive ONs, 50 days post procedure. Nav1.2 expression in A) relatively preserved ON and B) ON with high axon loss. Expression of Nav1.2 was independent of axon density. No positive staining was noticed in healthy ONs (not shown).

S. Sandalon et al. / Experimental Eye Research 115 (2013) 47e56

Two pathways have been proposed to explain the role of Nav in neuronal death. One proposed cascade includes depletion of ATP due to hypoxic conditions, followed by impaired function of Naþ/ KþATPase, intracellular Naþ overload, and reverse action of the Naþ/Caþþ exchanger (Nikolaeva et al., 2005; Stys, 2005). Naþ/Caþþ exchanger was indeed found in mouse RGC and inner plexiform layers (Inokuchi et al., 2009) and horizontal cells in fish (Hayashida et al., 1998). A second proposed pathway involves excitotoxic release of excessive glutamate to the extracellular medium, which results in toxic Caþþ and Naþ influx (Bano et al., 2005; Tekkök et al., 2007). Sodium dependent glutamate transporters are found mainly in Müller cells, but also in cones and some subsets of cone bipolar cells (Park et al., 2009). The former also express Nav channels that were found to be involved in glutamate release (Linnertz et al., 2011). Lamotrigine treatment showed no significant effect in preserving ERG responses, improving RGC survival or decreasing the extent of inflammatory involvement in either model. This could be due to insufficient levels of LTG in the retina, and/or low efficacy of the drug. Dosages of 10e30 mg/kg/day were previously reported as effective in rats, and plasma levels were in the range of therapeutic window for seizure treatment in humans (Bechtold et al., 2006; Marina et al., 2012; Papazisis et al., 2008). In the present study we used 12.5e20 mg/kg/day and observed mild weight loss during the first days of treatment, and therefore higher dosages were avoided. The extent of RGC loss in the acute model was very mild (Fig. 6A) and thus might have masked any beneficial effect of treatment on preventing RGC death. However, this model shortcoming does not explain the lack of LTG effect on inflammatory cell infiltration. In the chronic model, RGC loss was evident in both RGC count and axon density evaluation, and ED1 positive cells infiltration of the ON was marked (Fig. 6B). Again, LTG treatment was not beneficial. The chronic model is characterized by IOP below systolic pressure therefore the extent of ischemia is most likely reduced. Excitotoxicity might also be less pronounced in this model. For example, no increases in glutamate levels were found in the vitreous of rats after chronic IOP elevation (Levkovitch-Verbin et al., 2002a) or in dogs with spontaneous glaucoma (Källberg et al., 2007). This might account for the ineffectiveness of LTG treatment in the chronic model. Activated macrophages and microglia express Nav. Craner et al., (2005) found increased Nav1.6 expression in activated microglia and macrophages in EAE and multiple sclerosis mouse model. Furthermore, phenytoin, a Nav blocker, decreased the inflammatory cell infiltrate in EAE by 75%. We found no decrease in immune response, as ED1 positive cells counts did not differ between treatment groups in either model (Fig. 6). Again, lack of adequate LTG concentration or its ineffectiveness in blocking persistent sodium currents might explain the differences. We observed Nav1.1 immunoreactivity in ONs from both models. Nav1.2 immunoreactivity was increased in the chronic model only, though we did not quantify it. Axons (Boiko et al., 2001), astrocytes (Black et al., 2010) and reactive microglia/macrophages (Black et al., 2009; Craner et al., 2005) all express Nav in the ON. The positive staining we observed, not colocalized with NF-200, may indicate either astrocytic or microglia/macrophages reactivity. Yet, the fiber-like pattern suggests the former, as astrocytes in glaucomatous ONs in mice undergo thickening of their processes together with ‘appendages growing out of primary astrocyte processes into the axon bundles’ (Lye-Barthel et al., 2013). In conclusion, in our hands LTG treatment failed to protect rat retina and ON in either of the ocular hypertension models evaluated, though changes in the pattern of Nav expression were observed in both models.

55

Acknowledgments This study was funded by Grant G2008-010 from The American Health Assistance Foundation (BrightFocus Foundation). Shai Sandalon was funded by the Joseph Alexander Foundation and Birte Könnecke was funded by the European Union (FP6 program LSHMCT-2005-018637, Neuropromise). We thank Britta Kaltwasser and Irina Graf for excellent technical assistance and Tali Bdolah-Abram for her invaluable assistance with statistical analysis of our results. References Bano, D., Young, K.W., Guerin, C.J., Lefeuvre, R., Rothwell, N.J., Naldini, L., Rizzuto, R., Carafoli, E., Nicotera, P., 2005. Cleavage of the plasma membrane Naþ/Ca2þ exchanger in excitotoxicity. Cell 120, 275e285. Bechtold, D.A., Miller, S.J., Dawson, A.C., Sun, Y., Kapoor, R., Berry, D., Smith, K.J., 2006. Axonal protection achieved in a model of multiple sclerosis using lamotrigine. J. Neurol. 253, 1542e1551. Black, J.A., Liu, S., Waxman, S.G., 2009. Sodium channel activity modulates multiple functions in microglia. Glia 57, 1072e1081. Black, J.A., Newcombe, J., Waxman, S.G., 2010. Astrocytes within multiple sclerosis lesions upregulate sodium channel Nav1.5. Brain 133, 835e846. Blair, M., Pease, M.E., Hammond, J., Valenta, D., Kielczewski, J., LevkovitchVerbin, H., Quigley, H., 2005. Effect of glatiramer acetate on primary and secondary degeneration of retinal ganglion cells in the rat. Invest Ophthalmol. Vis. Sci. 46, 884e890. Boiko, T., Rasband, M.N., Levinson, S.R., Caldwell, J.H., Mandel, G., Trimmer, J.S., Matthews, G., 2001. Compact myelin dictates the differential targeting of two sodium channel isoforms in the same axon. Neuron, 91e104. United States. Boiko, T., Van Wart, A., Caldwell, J.H., Levinson, S.R., Trimmer, J.S., Matthews, G., 2003. Functional specialization of the axon initial segment by isoform-specific sodium channel targeting. J. Neurosci., 2306e2313. United States. Calabresi, P., Marti, M., Picconi, B., Saulle, E., Costa, C., Centonze, D., Pisani, F., Bernardi, G., 2003. Lamotrigine and remacemide protect striatal neurons against in vitro ischemia: an electrophysiological study. Exp. Neurol. 182, 461e469. Craner, M.J., Damarjian, T.G., Liu, S., Hains, B.C., Lo, A.C., Black, J.A., Newcombe, J., Cuzner, M.L., Waxman, S.G., 2005. Sodium channels contribute to microglia/ macrophage activation and function in EAE and MS. Glia 49, 220e229. Craner, M.J., Lo, A.C., Black, J.A., Waxman, S.G., 2003. Abnormal sodium channel distribution in optic nerve axons in a model of inflammatory demyelination. Brain 126, 1552e1561. Crumrine, R.C., Bergstrand, K., Cooper, A.T., Faison, W.L., Cooper, B.R., 1997. Lamotrigine protects hippocampal CA1 neurons from ischemic damage after cardiac arrest. Stroke 28, 2230e2236. discussion 2237. Cui, J., Pan, Z.H., 2008. Two types of cone bipolar cells express voltage-gated Naþ channels in the rat retina. Vis. Neurosci., 635e645. England. Guizzo, R., Cairrao, M.A.R., Coutinho-Neto, J., Silva, A.R.M.e., Coimbra, N.C., dos Santos, W.F., 2005. Neuroprotection in acute ischemia and ischemia/reperfusion in rat retina. Int. J. Pharmacol. 1, 369e375. Guizzo, R., Paques, M.W., Anhezini, L., Simon, C.R., Scott, I.U., Jorge, R., Santos, W.F., 2008. Neuroprotective effects of oral lamotrigine administration on rabbit retinas after pars plana vitrectomy and silicone oil injection. Retina 28, 638e644. Hains, B.C., Waxman, S.G., 2005. Neuroprotection by sodium channel blockade with phenytoin in an experimental model of glaucoma. Invest. Ophthalmol. Vis. Sci., 4164e4169. United States. Hayashida, Y., Yagi, T., Yasui, S., 1998. Ca2þ regulation by the Na(þ)eCa2þ exchanger in retinal horizontal cells depolarized by L-glutamate. Neurosci. Res. 31, 189e199. Inokuchi, Y., Shimazawa, M., Nakajima, Y., Komuro, I., Matsuda, T., Baba, A., Araie, M., Kita, S., Iwamoto, T., Hara, H., 2009. A Naþ/Ca2þ exchanger isoform, NCX1, is involved in retinal cell death after N-methyl-D-aspartate injection and ischemia-reperfusion. J. Neurosci. Res. 87, 906e917. Källberg, M.E., Brooks, D.E., Gelatt, K.N., Garcia-Sanchez, G.A., Szabo, N.J., Lambrou, G.N., 2007. Endothelin-1, nitric oxide, and glutamate in the normal and glaucomatous dog eye. Vet. Ophthalmol. 10 (Suppl. 1), 46e52. Lagrèze, W.A., Knörle, R., Bach, M., Feuerstein, T.J., 1998. Memantine is neuroprotective in a rat model of pressure-induced retinal ischemia. Invest. Ophthalmol. Vis. Sci. 39, 1063e1066. Lagrue, E., Chalon, S., Bodard, S., Saliba, E., Gressens, P., Castelnau, P., 2007. Lamotrigine is neuroprotective in the energy deficiency model of MPTP intoxicated mice. Pediatr. Res. 62, 14e19. Levkovitch-Verbin, H., Martin, K.R., Quigley, H.A., Baumrind, L.A., Pease, M.E., Valenta, D., 2002a. Measurement of amino acid levels in the vitreous humor of rats after chronic intraocular pressure elevation or optic nerve transection. J. Glaucoma 11, 396e405. Levkovitch-Verbin, H., Quigley, H.A., Martin, K.R., Valenta, D., Baumrind, L.A., Pease, M.E., 2002b. Translimbal laser photocoagulation to the trabecular meshwork as a model of glaucoma in rats. Invest Ophthalmol. Vis. Sci. 43, 402e410. Linnertz, R., Wurm, A., Pannicke, T., Krügel, K., Hollborn, M., Härtig, W., Iandiev, I., Wiedemann, P., Reichenbach, A., Bringmann, A., 2011. Activation of voltage-

56

S. Sandalon et al. / Experimental Eye Research 115 (2013) 47e56

gated Naþ and Ca2þ channels is required for glutamate release from retinal glial cells implicated in cell volume regulation. Neuroscience 188, 23e34. Lye-Barthel, M., Sun, D., Jakobs, T.C., 2013. Morphology of astrocytes in a glaucomatous optic nerve. Invest. Ophthalmol. Vis. Sci. 54, 909e917. Marina, N., Sajic, M., Bull, N.D., Hyatt, A.J., Berry, D., Smith, K.J., Martin, K.R., 2012. Lamotrigine monotherapy does not provide protection against the loss of optic nerve axons in a rat model of ocular hypertension. Exp. Eye Res. 104, 1e6. Mojumder, D.K., Frishman, L.J., Otteson, D.C., Sherry, D.M., 2007. Voltage-gated sodium channel alpha-subunits Na(v)1.1, Na(v)1.2, and Na(v)1.6 in the distal mammalian retina. Mol. Vis. 13, 2163e2182. Nikolaeva, M.A., Mukherjee, B., Stys, P.K., 2005. Naþ-dependent sources of intraaxonal Ca2þ release in rat optic nerve during in vitro chemical ischemia. J. Neurosci. 25, 9960e9967. Papazisis, G., Kallaras, K., Kaiki-Astara, A., Pourzitaki, C., Tzachanis, D., Dagklis, T., Kouvelas, D., 2008. Neuroprotection by lamotrigine in a rat model of neonatal hypoxic-ischaemic encephalopathy. Int. J. Neuropsychopharmacol. 11, 321e329. Park, C.K., Cha, J., Park, S.C., Lee, P.Y., Kim, J.H., Kim, H.S., Kim, S.A., Kim, I.B., Chun, M.H., 2009. Differential expression of two glutamate transporters, GLAST and GLT-1, in an experimental rat model of glaucoma. Exp. Brain Res. 197, 101e109. Pisani, F., Pedale, S., Macaione, V., Torre, V., Oteri, G., Avanzini, G., Ientile, R., 2001. Neuroprotective effects of lamotrigine and remacemide on excitotoxicity induced by glutamate agonists in isolated chick retina. Exp. Neurol. 170, 162e170.

Rosenbaum, D.M., Degterev, A., David, J., Rosenbaum, P.S., Roth, S., Grotta, J.C., Cuny, G.D., Yuan, J., Savitz, S.I., 2010. Necroptosis, a novel form of caspase-independent cell death, contributes to neuronal damage in a retinal ischemia-reperfusion injury model. J. Neurosci. Res. 88, 1569e1576. Sandalon, S., Ofri, R., 2009. The effect of topical anesthesia on the rat electroretinogram. Doc. Ophthalmol. 118, 101e108. Shuaib, A., Mahmood, R.H., Wishart, T., Kanthan, R., Murabit, M.A., Ijaz, S., Miyashita, H., Howlett, W., 1995. Neuroprotective effects of lamotrigine in global ischemia in gerbils. A histological, in vivo microdialysis and behavioral study. Brain Res., 199e206. Netherlands. Spadoni, F., Hainsworth, A.H., Mercuri, N.B., Caputi, L., Martella, G., Lavaroni, F., Bernardi, G., Stefani, A., 2002. Lamotrigine derivatives and riluzole inhibit INa, P in cortical neurons. Neuroreport 13, 1167e1170. Stanford, M.R., Robbins, J., 1988. Experimental posterior uveitis. II. Electroretinographic studies. Br. J. Ophthalmol. 72, 88e96. Stanford, M.R., Robbins, J., Kasp, E., Dumonde, D.C., 1992. Passive administration of antibody against retinal S-antigen induces electroretinographic supernormality. Invest. Ophthalmol. Vis. Sci. 33, 30e35. Stys, P.K., 2005. General mechanisms of axonal damage and its prevention. J. Neurol. Sci. 233, 3e13. Tekkök, S.B., Ye, Z., Ransom, B.R., 2007. Excitotoxic mechanisms of ischemic injury in myelinated white matter. J. Cereb. Blood Flow Metab. 27, 1540e1552.