Effects of a conventional photocoagulator and a 3-ns pulse laser onpreconditioning responses and retinal ganglion cell survival after opticnerve crush
O'Sam Shibeeb a, b, John P.M. Wood a, b, Robert J. Casson a, b, Glyn Chidlow a, b, *
a Ophthalmic Research Laboratories, South Australian Institute of Ophthalmology, Hanson Institute Centre for Neurological Diseases, Frome Rd, Adelaide,SA 5000, Australiab Department of Ophthalmology and Visual Sciences, University of Adelaide, Frome Rd, Adelaide, SA 5000, Australia
a r t i c l e i n f o
Article history:Received 28 March 2014Accepted in revised form 10 July 2014Available online 21 July 2014
Previous research has demonstrated that laser photocoagulation treatment of the monkey retina affordsprotection against experimental glaucoma-induced retinal ganglion cell (RGC) loss in areas overlyinglaser spots. The underlying mechanism is unknown, but it is conceivable that the laser acted as a pre-conditioning stimulus, inducing localised, endogenous production of survival factors. The related pur-poses of the current study were firstly to examine whether preconditioning pathways are activated byeither a conventional photocoagulator (CW) laser or a photoreceptor-sparing, short-pulse duration (2RT)laser in the rat retina, and secondly, to examine whether such preconditioning with either laser improvesRGC survival after optic nerve (ON) crush. Pigmented rats were randomly assigned to one of threegroups: sham, CW, 2RT. For the preconditioning study, laser spots were applied randomly to each retinain the posterior hemisphere of the eye taking care to avoid major blood vessels. Animals were killed at6 h, 1d, and 7d after laser treatment, then analysed by qPCR, immunohistochemistry or Westernimmunoblotting. For the neuroprotection study, laser spots were administered to the mid-central retinaof the right eye. The left eye served as a control. In two experiments, rats were lasered either 24 h or 7days before ON crush, then killed a further 7 days later. Wholemount retinas were prepared and doublelabelling immunofluorescence performed. Nestin labelling allowed visualization of laser spots. Brn3alabelling identified viable RGCs. Photomicrographs of Brn3a labelling were taken in areas overlyingnestin-positive laser spots. Quantification of Brn3a RGCs was then performed. Both the CW and 2RTlasers induced local glial cell activation. Moreover, both lasers induced localized upregulations of anumber of well-documented (CNTF, FGF-2 Hsp27, pAKT) or putative (cFOS, ATF-3, IL-6) RGC survivalfactors. However, neither laser caused sustained increases in other factors associated with neuronalpreconditioning, such as BDNF, Hsp70, IGF-1, bcl-2, and nitric oxide synthase. As regards neuro-protection, analysis of the data revealed that ON crush resulted in the loss of approximately 70% ofBrn3a-labelled RGCs after 1 week. Neither the CW nor the 2RT laser augmented Brn3a-positive RGCsurvival in areas overlying and neighbouring laser spots. This was the case irrespective of whetherlasering occurred 1 or 7 days before the ON crush. Our results showed that the CW and 2RT lasers bothstimulated de novo synthesis of certain genes that are well-known RGC survival factors and/or that havebeen implicated in preconditioning-induced neuroprotection studies. Despite these findings, neitherlaser augmented survival of RGCs when delivered prior to ON crush.
oratories, Hanson Centre for, Australia. Tel.: þ61 8 8222
Chidlow).
1. Introduction
Neurodegenerative disease is generally thought to represent asituation whereby cell death signals have overwhelmed cellular/tissue defenses. Glaucoma, a family of neurodegenerative diseasescharacterized by the loss of retinal ganglion cells (RGCs; (Cassonet al., 2012a), currently has a limited range of therapies, all of
O. Shibeeb et al. / Experimental Eye Research 127 (2014) 77e9078
which act to lower intraocular pressure, the highest profile riskfactor for the disease (Maier et al., 2005). A consensus has devel-oped that additional treatment strategies, based on direct neuro-protection of RGCs, are needed (Weinreb and Levin, 1999). Such anapproach may be achieved by augmenting the levels or influence offactors which are protective (“survival factors”) and/or reducing theeffect of signals which are detrimental (“death factors”) to RGCs. Inanimal models of glaucoma, application of exogenous survivalfactors, for example neurotrophins (Chader, 2012), has yieldedpromising results. An alternative strategy is to stimulate endoge-nous production of survival factors. This approach is appealingbecause it harnesses evolutionary biology, taking advantage of thein situ synthesis of multiple survival factors without the need forexogenous delivery of drugs or genetic manipulation of cells. Onemethod of endogenous neuroprotection is “preconditioning”.
Preconditioning refers to a phenomenon whereby a mildnoxious insult protects a tissue from a subsequent severe insult(Gidday, 2006). The phenomenon has been well described in avariety of organs, including the retina (Roth, 2004). Research effortsin the retina have focused on augmenting survival of photorecep-tors and RGCs, since both cell types have proven amenable topreconditioning. This treatment modality can be elicited by a widevariety of stimuli; for example, mechanical injury (Faktorovichet al., 1990, 1992), bright light (Liu et al., 1998) and ischemia(Casson et al., 2003) have all been shown to delay photoreceptordegeneration, while ischemia (Roth et al., 1998), hypothermia(Salido et al., 2013) and lipopolysaccharide (Franco et al., 2008) canenhance RGC survival. The cellular mechanisms responsible forpreconditioning-induced neuroprotection are complex andincompletely understood, but convincing evidence has been pro-vided that activation of genes and de novo protein synthesis ofdiverse families of pro-survival factors are involved. Numerouspotential mediators have been implicated, including heat shockproteins (Hsps), trophic factors, immediate early genes, pro- as wellas anti-inflammatory cytokines, involvement of endothelial andinducible nitric oxide synthase, mitochondrial effectors, and mol-ecules involved in signal transduction (Gidday, 2006; Shpargelet al., 2008).
An intriguing study by Nork and colleagues (Nork et al., 2000)demonstrated that retinal laser photocoagulation in a monkeymodel of glaucoma caused focal loss of photoreceptors, but, affor-ded protection to overlying RGCs when compared to non-laseredareas. The authors of this study concluded that the pathogenesisof glaucomatous RGC loss may occur in an anterograde mannerinitiated at the photoreceptors; however, an alternative interpre-tation is that the retinal laser acted as a preconditioning stimulus,inducing endogenous, localized, synthesis of factors which acted ina protective manner. To date, the spectrum of cellular responsesinduced by retinal laser photocoagulation has not been fully char-acterized, striking increases in certain trophic factors and smallHsps, which are important survival factors for RGCs, have beendemonstrated (Chidlow et al., 2013; Xiao et al., 1998).
Although of interest mechanistically, photocoagulation(continuous wave, CW) lasers would not be clinically applicable toprotect RGCs because of the accompanying thermal destruction ofthe photoreceptors. Of greater interest are newer retinal laserswhose energy is more spatially confined to the retinal pigmentepithelium (RPE). One such laser is the retinal regeneration therapy(2RT) laser, a short pulse duration laser developed for the treatmentof diabetic macular oedema (Casson et al., 2012b). We have shownthat the 2RT laser ablates the RPE, yet causes significantly reducedcollateral damage to overlying photoreceptors in rats (Wood et al.,2013). Interestingly, the 2RT laser, like the CW laser, also induces anupregulation of trophic factors and small Hsps in the retina(Chidlow et al., 2013). Hence, this laser provides an opportunity to
determine whether an RGC preconditioning effect can be achievedwithout causing substantial damage to the sensory retina.
The related purposes of the current study were, firstly, toexamine whether multiple preconditioning pathways are activatedby CW and 2RT lasers, and secondly, to examine whether pre-conditioning with either laser is neuroprotective in a rat model ofRGC degeneration.
2. Materials and methods
2.1. Lasers
Two frequency-doubled Nd:YAG lasers with 532 nm wave-lengths were used for treatment of animals (Ellex R&D Pty Ltd,Adelaide, Australia). The first laser was a continuous wave (CW)laser with a flat top beam profile that produces a 400 mm diameterspot in air and 300 mmdiameter spot on the rat retina. The laser hasa total exposure duration of 100 ms and was used at 90 mW. Thesesettings equate to a delivered energy of 9 mJ per pulse, whichconverts to a radiant exposure level of 12.7 J/cm2/pulse. At thisintensity, the laser produced light laser burns, defined as blanchingof the fundus pigmentation (but without candid whitening) anddevoid of adjacent oedema, subretinal or retinal hemorrhage; set-tings are broadly equivalent to those used in clinical situations. Thesecond laser was a short pulse duration (3 ns; 2RT) laser with a finespeckle beam profile that produces a 380 mm diameter spot size inair and a 285 mmdiameter spot size on the rat retina. The laser has atotal exposure duration of 3ns and was adjusted to deliver an en-ergy of 0.33 mJ per pulse, which converts to a radiant exposurelevel of 0.166 J/cm2/pulse. At this intensity, the laser produced aclear visual effect on the eye at the time of laser application iden-tified as a subtle blanching. A 5.4 mm fundus laser contact lens(Ocular Instruments, Bellevue, WA, USA) was used to focus the lightbeam of each laser onto the retina.
2.2. Treatment of animals
This study was approved by the Animal Ethics Committees of SAPathology and the University of Adelaide. The study conformed tothe Australian Code of Practice for the Care and Use of Animals forScientific Purposes, 2004, and to the ARVO Statement for the Use ofAnimals in Ophthalmic and Vision Research. Adult Dark Agouti rats(approximately 150 g) were housed in a temperature- andhumidity-controlled room with a 12-h light, 12-h dark cycle andwere provided with food and water ad libitum.
Prior to induction of laser treatment, rats were anaesthetisedwith an intraperitoneal injection of a mixture of 100 mg/kg keta-mine and 10 mg/kg xylazine. When general anaesthesia had beenobtained, the pupils were dilated by topical application of tropi-camide, allowing visualisation of the optimum area of retinathrough the eye. Animals were then placed on a custom-designedplatform attached to the slit lamp laser delivery system. At desig-nated times after lasering, rats in the preconditioning study (seebelow) were killed by transcardial perfusion with physiologicalsaline under deep anaesthesia and the globes enucleated imme-diately for analysis by histology, immunohistochemistry, Westernimmunoblotting or qPCR. Rats in the neuroprotection study (seebelow) subsequently received optic nerve (ON) crush in the righteye. ON crush was performed as previously described (Chidlowet al., 2011b). In brief, the superior muscle complex was dividedand the ON exposed by blunt dissection. The ON was then crushed3 mm posterior to the globe under direct visualisation using self-closing number 5 forceps for 15 s. To avoid confusing retinalischemic changes with the effects of crush, the fundus wasobserved ophthalmologically immediately after nerve crush, in
CST, cell signaling technology.a 2-step immunofluorescence.b Western immunoblotting.
O. Shibeeb et al. / Experimental Eye Research 127 (2014) 77e90 79
order to discard any animals with perfusion dysfunction as a resultof surgery.
2.3. Study design
2.3.1. Preconditioning studyRats were randomly assigned to one of three treatment groups:
sham, CWor 2RT. Laser spots were applied randomly to each retinain the posterior hemisphere of the eye around the optic nerve headtaking care to avoid the optic nerve head and immediately adjacentarea as well as major blood vessels. For animals that were subse-quently used for Western immunoblotting or qPCR, the number ofeyes analysed per treatment group per point was as follows: 6h(n ¼ 6), 1d (n ¼ 6), 7d (n ¼ 6). An additional group of untreated rats(n¼ 6) served as controls. For animals that were subsequently usedfor histology/immunohistochemistry, the number of eyes analysedper treatment group per time point was as follows: 6h (n ¼ 4), 1d(n ¼ 4), 3d (n ¼ 4), 7d (n ¼ 4). A further group of unlasered eyes(n ¼ 4) served as controls.
2.3.2. Neuroprotection studyThe neuroprotection study comprised two experiments.Experiment (1a): Rats were randomly assigned to one of three
treatment groups: sham, CW or 2RT. Sham animals were anaes-thetised, but neither eye was lasered. For the CW and 2RT groups,25 laser spots were applied randomly to the mid-central retina ofthe right eye, as detailed above, while left eyes remained untreated.At 1 day after lasering, rats in all three groups received ON crush inthe right eye. The left ON remained intact. Rats were allowed torecover and killed 1 week after ON crush for quantification of RGCs.Of the 24 rats, there were two deaths related to anaesthesia, whileone rat was excluded due to an abnormal retina. The number ofeyes analysed per treatment group was as follows: sham (n ¼ 6),CW (n ¼ 8), 2RT (n ¼ 7).
Experiment (1b): The design of experiment 1b was identical toexperiment 1a except that rats in all three groups received ONcrush in the right eye at 7 days rather than 1 day after lasering. Ofthe 21 rats, there was one death related to anaesthesia, while onerat was humanely killed prematurely as a result of ocular inflam-mation. The number of eyes analysed per treatment group was asfollows: sham (n ¼ 6), CW (n ¼ 7), 2RT (n ¼ 6).
Experiment (2): Here, rats were randomly assigned to one oftwo treatment groups: sham or CW. Animals in the CWgroup werelasered as in experiment 1a. At 1 day after lasering, rats in bothgroups received ON crush in the right eye. The left ON remainedintact. Rats were allowed to recover and killed 2 weeks after ONcrush for quantification of RGCs. The number of eyes analysed pertreatment group was as follows: sham (n ¼ 6), CW (n ¼ 6).
2.4. Tissue processing and immunohistochemistry
Tissue processing, histology and immunohistochemistry wereperformed as described previously (Chidlow et al., 2011a). In brief,eyes were immersion-fixed in either Davidson's solution or 10%formalin for 24 h, then transferred to 70% ethanol until processing.Globes were embedded sagitally and 4 mm sections were cut. Eyesthat were used for retinal wholemount immunohistochemistrywere dissected into posterior eye cups, post-fixed in 4% para-formaldehyde, then prepared as flattened wholemounts. Forimmunohistochemistry on transverse sections, tissue sections weredeparaffinised, endogenous peroxidase activity was blocked andhigh-temperature antigen retrieval was performed. Subsequently,sections were incubated in primary antibody (see Table 1), followedby consecutive incubations with biotinylated secondary antibodyand streptavidin-peroxidase conjugate. Colour development was
achieved using NovaRed substrate kit. Confirmation of the speci-ficity of antibody labelling was judged by the morphology anddistribution of the labelled cells, by the absence of signal when theprimary antibody was replaced by isotype/serum controls, bycomparison with the expected staining pattern based on our own,and other, previously published results, and, by the detectionwithin retinal samples of a protein at the expected molecularweight by Western blotting.
To perform immunofluorescence on wholemount preparations,samples were permeabilized with PBS containing 0.5% triton andthen incubated with primary antibody (see Table 1), followed byAlexaFluor 488 conjugated secondary antibody. Tissues were sub-sequently mounted and examined under a confocal fluorescencemicroscope. For double labelling immunofluorescence, whole-mounts were treated as described above except that a combinationof primary antibodies was used. Detection of signal was achieved byincubationwith the appropriate combination of AlexaFluor 488 and594 conjugated secondary antibodies.
2.5. Western immunoblotting
Western blotting was performed according to a standardmethodology. In brief, tissue extracts were sonicated in homoge-nisation buffer, diluted with an equal volume of sample buffer, andheated at 70 �C for 3 min. Electrophoresis was performed usingnon-denaturing polyacrylamide gels. After separation, proteinswere transferred to polyvinylidine fluoride membranes for immu-noprobing. Membranes were incubated consecutively with theappropriate primary antibody (as detailed in Table 1), biotinylatedsecondary antibody and streptavidin-peroxidase conjugate. Colourdevelopment was achieved using 3-amino-9-ethylcarbazole andhydrogen peroxide as substrates. Detection of b-actin was assessedin all samples as a positive gel-loading control. Quantification ofdetected proteins was achieved using the program, Adobe Photo-Shop CS2. Statistical analysis was determined by ANOVA, followedby post-hoc Dunnett's Multiple Comparison test (treated vs control.
2.6. qPCR
Real-time polymerisation chain reaction (qPCR) studies werecarried out according to a standard methodology. In brief, retinaswere dissected, total RNA was isolated and first strand cDNA wassynthesised from DNase-treated RNA. qPCR reactions were carried
O. Shibeeb et al. / Experimental Eye Research 127 (2014) 77e9080
2.7. Quantification of RGCs in retinal wholemounts
In order to quantify RGCs in regions of the retina overlying laserspots, it was necessary to perform double labelling immunofluo-rescence using antibodies directed against Brn3a (RGCs) and theintermediate filament, nestin. Nestin is locally and persistentlyupregulated by glial cells at the sites of laser treatment, therebyfacilitating visualization of laser spots. Thus, in CW- and 2RT-treated retinas, a laser spot was identified by nestin immuno-labelling, centredwithin the photomicrographic field, and an imagecaptured of the corresponding Brn3a immunolabelling. In sham-treated retinas, and in control retinas, images were captured ofBrn3a immunolabelling at approximately similar eccentricities tothose of lasered retinas. For each retina, 12 photomicrographicfields were analysed, each of whichmeasured 720� 540 mm in size.The number of labelled RGCs in each of the 12 photomicrographicfields was calculated, averaged, and expressed as mean ± SEM. Allcell counts were conducted by the same investigator, who wasblinded to the identity of the retinas. Comparison of the number ofBrn3a-labelled RGCs in sham, CW- and 2RT-treated groups wasperformed by ANOVA.
3. Results
3.1. Histology of laser lesions
The principal target of laser irradiation is the retinal pigmentepithelium (RPE), whose intracellular pigmentary melanosomesabsorb up to 50% of incident laser energy (Brinkmann et al., 2000).As shown in Supplementary Fig. 1A, both the CW and 2RT lasersablated RPE cells in the irradiated zone. The CW laser also causedlocalized destruction of the outer nuclear layer and photoreceptorsegments (Supplementary Fig. 1B). The 2RT laser likewise disruptedphotoreceptor outer segments but only produced minor cell loss inthe ONL (Supplementary Fig. 1B). Proliferating cells were present inthe inter-photoreceptor matrix after laser treatment, as evidencedby positive immunolabelling for proliferating cell nuclear antigen(Supplementary Fig. 1C).
3.2. Preconditioning responses to laser treatment: transcriptionfactors
Induction of transcription factors, such as immediate earlygenes, is believed to be an early tissue event following
preconditioning stimuli. We examined the effect of CW and 2RTlaser treatment on expression of two relevant factors, ATF-3 andc-FOS. Both lasers induced a rapid and marked upregulation ofATF-3 mRNA by 6h, which remained significantly elevated at 1d,but had returned to baseline by 7d (Fig. 1A). The increase in ATF-3transcription after laser treatment correlated with an increasein ATF-3 translation, as delineated by Western blotting (Fig. 1B,C). Immunolabelling for ATF-3 in tissue sections localised theprotein to intermittent RGC nuclei overlaying laser spots inlasered, but not sham, retinas (Fig. 1DeF, SupplementaryFig.2AeC).
Immunolabelling for c-FOS revealed occasional, faintly-labelledMüller cells in sham animals, but a rapid upregulation of thetranscription factor within 6h of CW or 2RT laser treatment(Fig. 1GeI). Expression of c-FOS was associated with Muller cellperikarya in the inner nuclear layer overlying and adjacent to theirradiated region (Fig. 1GeI, Supplementary Fig. 2DeF), togetherwith sporadic RGCs (Supplementary Fig. 2GeL), astrocytes andendothelial cells.
3.3. Preconditioning responses to laser treatment:pro-inflammatory mediators
Pro-inflammatory cytokines are traditionally viewed as neuro-toxic; however, when released as components of certain pre-conditioning stimuli (e.g. lipopolysaccharide) prior to a majorinsult, such cytokines can contribute to the neuroprotectivephenotype. Herein, we investigated expression of four classical pro-inflammatory gene products, namely interleukin-1b (IL-1b),interleukin-6 (IL-6), tumour necrosis factor-a (TNF-a) and inter-feron-c (IFN-c) following laser treatment. At 6h after lasering,mRNAs for IL-1b, TNF-a and IL-6 mRNAs (Fig. 2A, B) weredramatically upregulated in the CW group, and significantly, butless markedly, upregulated in the 2RT group. By 24h, mRNA levelswere lower, but still elevated in the CW group, and had essentiallyreturned to baseline in the 2RT group (Fig. 2A, B). No change wasmeasured in expression of IFN-c after laser treatment (Fig. 2B).Immunolabelling experiments (Fig. 2CeE, Supplementary Fig. 3)demonstrated TNF-a expression within microglia in the outer andinner) retina, while IL-6 was observed to be associated withendothelial cells of blood vessels (Fig. 2FeH).
3.4. Preconditioning responses to laser treatment:anti-inflammatory mediators
The production of anti-inflammatory cytokines by astrocytes isalso gradually being recognised to play a part in the neuro-protective response of CNS tissue to preconditioning stimuli. Toassess the effects of laser treatment on the production of suchcytokines, we investigated expression of TGF-b1 and IL-10 usingqPCR and Western blotting (Fig. 3). The overall data displayed noclear-cut pattern of upregulation. There was a trend of slightlyhigher TGF-b1 and IL-10 mRNA levels in the retinas of both lasergroups compared with the sham group (Fig. 3A, B), but inter-animal variability was evident, resulting in largely non-significant effects, whilst evaluation of the Western blottingdata (Fig. 3C, D) revealed neither laser caused a statistically sig-nificant increase in either cytokine at either of the time pointsanalysed.
3.5. Preconditioning responses to laser treatment: heat shockproteins
An increasing body of evidence supports the view that induc-tion of certain Hsps, as a consequence of preconditioning, affords
Fig. 1. Upregulation of activating transcription factor 3 (ATF-3) and c-fos after treatment with CWor 2RT laser. (A) Quantification of the ATF-3 mRNA level at 6h, 1d and 7d followingtreatment with CW or 2RT laser. Values (represented as mean ± SEM, n ¼ 6) are normalised for housekeeping genes and expressed relative to the control group. **P < 0.01, by Pair-wise Fixed Reallocation Randomization Test (treated vs control). (B) Quantification of ATF-3 protein level at different time points following treatment with CW or 2RT laser. Values(represented as mean ± SEM, n ¼ 6) are normalised for actin and expressed relative to the control group. **P < 0.01, by ANOVA followed by post-hoc Dunnett's Multiple ComparisonTest (treated vs control). (C) Representative ATF-3 immunoblots in sham and lasered retinas are shown. (DeF) Representative images of ATF-3 immunolabelling in the normal retina(D) and 1d after CW (E) or 2RT (F) laser treatment. Occassional RGC nuclei overlying the lesion, demarcated by asterisk, are ATF-3-positive (arrows). (GeI) Representative images ofc-fos immunolabelling in the normal retina (G) and 6h after CW (H) or 2-RT (I) laser treatment. Upregulation of c-fos is evident after CW and 2RT treatment in Müller cells and theoccasional RGC (arrows) overlying the lesion, demarcated by asterisk. Scale bar: B, C ¼ 17.5 mm; FeH ¼ 35 mm. GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclearlayer.
O. Shibeeb et al. / Experimental Eye Research 127 (2014) 77e90 81
neuroprotection to vulnerable neurons. Herein, we examined theeffects of CW and 2RT laser treatment on expression of threeinducible Hsps: Hsp27, Hsp70 and Hsp22. At 6h and 1d after lasertreatment, Hsp27 mRNA levels in the CW and 2RT groups weresignificantly upregulated compared to the sham group. Expressionof Hsp27 mRNA remained elevated by the 7d time point after CWlaser treatment, but in the 2RT group the respective increase wasno longer statistically significant (Fig. 4A). In contrast, neither laseraffected synthesis of Hsp22 mRNA (Fig. 4B). Regarding Hsp70, bothlasers had caused a small, but significant elevation in transcriptionof Hsp70 by 6h, but this increase was no longer present at 1d(Fig. 4C). Western immunoblot analyses could not detect signifi-cant elevations in Hsp70 protein content at either 1d or 7d(Fig. 4D).
3.6. Preconditioning responses to laser treatment: trophic factors
One strategy that has been successfully exploited to promoteRGC survival in models of lethal injury is preconditioning-inducedincreases in availability of neurotrophins. In the current study, weexamined whether CW or 2RT laser stimulates de novo expressionof four trophic factors recognised as neuroprotective effectors: FGF-2, BDNF, IGF-1 and CNTF. FGF-2 mRNAwas robustly upregulated forat least 24h after laser treatment, an effect which was markedlygreater in CW animals than 2RT rats (Fig. 5A). As regards BDNF, theCW and to a lesser extent 2RT lasers both caused modest, but sig-nificant, increases in mRNA abundance within the first 24h aftertreatment (Fig. 5B). In contrast, neither laser altered expression ofIGF-1 mRNA (Fig. 5C). We have previously shown an elevation of
Fig. 2. Effect of CW or 2RT laser on expression of pro-inflammatory mediators. (A, B) Quantification of IL-1b, TNF-a, IL-6 and INF-c mRNAs at 6h, 1d and 7d after laser treatment.Values (represented as mean ± SEM, n ¼ 6) are normalised for housekeeping genes and expressed relative to the control group. **P < 0.01, by Pair-wise Fixed ReallocationRandomization Test (treated vs control). Representative images of TNF-a immunolabelling in the normal retina (C) and 6h after CW laser treatment (D, E). After CW laser treatment,TNFa is associated with microglial cells in the outer (D) and inner retina (E). Representative images of IL-6 immunolabelling in the normal retina (F) and 6h after laser treatment (G,H). After CW and 2RT laser treatment, IL-6 is expressed by endothelial cells of blood vessels (E). Scale bar: 17.5 mm. GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outernuclear layer.
O. Shibeeb et al. / Experimental Eye Research 127 (2014) 77e9082
CNTF after laser treatment, but did not clarify the spatial extent ofexpression. Thus, in the current study, we immunolabelled retinalwholemounts prior to and after laser treatment. An interestingtemporal phenomenon was observed. In the normal retina CNTFexpression was low and restricted to astrocytes (Fig. 5D). By 24hpost-treatment, CNTFwas upregulated inMüller cells located at theboundaries of laser spots (as demarcated by nestin, see section 3.10)in a “halo-like” effect (Fig. 5D). By 7d after laser-treatment, CNTFimmunolabellingwas concentrated in astrocytes andMüller cells inthe centre of the lesion (Fig. 5D; Supplementary Fig. 4). Thesepatterns of labelling reflected the situation in both the CW and 2RTlasers. Western immunoblot data revealed both lasers caused sta-tistically significant increases in CNTF expression that were mostpronounced at 7d after treatment (Fig. 5E).
3.7. Preconditioning responses to laser treatment: nitric oxide andfree radicals
It is now believed that some reactive oxygen species (e.g. nitricoxide) can play key roles as protective mediators in certainneuronal preconditioning responses. We, thus, examined the effectof laser irradiation on expression of endothelial and inducible nitricoxide synthases (eNOS and iNOS, respectively), and on the NOX-2isoform of NADPH oxidase, which is responsible for production ofreactive oxygen species, using qPCR and immunohistochemistry. In
the cases of eNOS and iNOS, there were slight, but non-significanttrends of elevated expression in the treated groups comparedwith the sham group (Fig. 6A, B). This lack of significant change wassupported by analysis of tissue sections immunolabelled for eNOS,which displayed no discernible upregulation (Fig. 6D, E). For NOX-2however, a different pattern emerged: In CW-treated retinas, therewas a clear, statistically significant upregulation in NOX-2 mRNA at1d and 7d compared with sham animals (Fig. 6C). Immunohisto-chemistry revealed localisation of the protein to microglia withinthe vicinity of laser spots in CW animals, whereas in unlaseredretinas no NOX-2 immunoreactivity was detected (Fig 6F, G). In 2RTanimals, the increase in NOX-2 mRNA was substantially moresubdued than after CW treatment (Fig. 6C).
3.8. Preconditioning responses to laser treatment: apoptosis-relatedproteins
Studies in the brain have shown that preconditioning-inducedincreases in expression of members of the anti-apoptotic bcl-2family are associated with greater neuronal survival following le-thal injury. We, therefore, examined the effects of CWand 2RT lasertreatment on levels of the two best characterized family members,bcl-2 (Supplementary Fig. 5A) and bcl-xl (Supplementary Fig. 5B,C), using qPCR and Western blotting, respectively. Our resultsshowed that neither laser treatment induced a significant
Fig. 3. Effect of CW or 2RT laser on expression of anti-inflammatory mediators. (A, B) Quantification of TGF-b1 and IL-10 mRNA levels at 6h, 1d and 7d after laser treatment. Values(represented as mean ± SEM, n ¼ 6) are normalised for housekeeping genes and expressed relative to the control group. *P < 0.05, by Pair-wise Fixed Reallocation RandomizationTest (treated vs control). (C, D) Representative IL-10, TGF-b1 and actin Western immunoblots together with quantification of IL-10 and TGF-b1 protein level at 1d and 7d followinglaser treatment. All values (represented as mean ± SEM, n ¼ 6) are normalised for actin and expressed relative to the control group.
O. Shibeeb et al. / Experimental Eye Research 127 (2014) 77e90 83
upregulation in either family member at any of the time pointsanalysed, when compared to the sham group.
3.9. Preconditioning responses to laser treatment: Akt signalling
Several studies have identified activation of Akt (protein kinaseB) as important for establishing a tolerant cellular phenotype topreconditioning. Accordingly, we examined the effect of CW and2RT laser treatment on the level of phosphorylated (activated) Akt(pAKT). In controls retinas, pAkt protein levels are low; howeverfollowing CW or 2RT laser treatment, there was a statistically sig-nificant, sustained upregulation of pAkt as determined by Westernblotting (Fig. 7A, B).
3.10. Effect of laser treatment on RGC viability in healthy retinas
Before examining the effect of laser treatment on RGC survivalfollowing ON crush, it was important to verify that laser treatmentalone produced no change in the viability of RGCs in regions of theretina overlying and bordering laser lesions. In order to visualizeboth laser spots and RGCs, retinas were wholemounted and doublelabelling immunofluorescence performed using antibodies directedagainst Brn3a and the intermediate filament, nestin (seeSupplementary Fig. 6). We have recently demonstrated that nestin
is locally and persistently upregulated in astrocytes andMüller cellsat the lesion site as a component of the glial scar (Wood et al.,2013); thus, nestin immunolabelling permits reliable visualizationof laser spots. We quantified the number of Brn3a positive neurons7d after laser treatment in defined photomicrographic fieldsfeaturing one (or occasionally two) lesion(s). The data showed thatneither the CW nor the 2RT laser caused any loss of RGCs, whencompared with the sham group (Supplementary Fig. 6).
3.11. Effect of laser treatment on RGC survival after optic nervecrush: experiment 1
To examine whether the CW or 2RT laser augmented RGC sur-vival after ON crush, we again performed double labelling immu-nofluorescence using Brn3a to demarcate RGCs and nestin toidentify photomicrographic fields featuring laser spots. Since pre-conditioning has been shown to involve early and late phases ofgene transcription, with some genes expressed within hours andothers days later, we performed two separate experiments: (1a) ONcrushwas performed 1 day after laser treatment; (1b) ON crushwasperformed 1 week after laser treatment. In both cases, animalswere killed 7d after ON crush. Fig. 8 shows representative images ofBrn3a-nestin immunolabelling in ON crush and contralateral ret-inas from the sham, CW and 2RT groups. Nestin upregulation is
Fig. 4. Effect of CW or 2RT laser on expression of heat shock proteins (Hsps). (AeC) Quantification of Hsp27, Hsp22 and Hsp70 mRNA levels at 6h, 1d and 7d after laser treatment.Values (represented as mean ± SEM, n ¼ 6) are normalised for housekeeping genes and expressed relative to the control group. **P < 0.01, *P < 0.05, by Pair-wise Fixed ReallocationRandomization Test (treated vs control). (D) Representative Hsp70 and actin Western immunoblots together with quantification of Hsp70 protein level at 1d and 7d following lasertreatment. All values (represented as mean ± SEM, n ¼ 6) are normalised for actin and expressed relative to the control group.
O. Shibeeb et al. / Experimental Eye Research 127 (2014) 77e9084
visible in all ON crush retinas, but laser lesions are still readilydistinguishable. In all groups, treated retinas exhibited a markedloss of Brn3a neurons. Analysis revealed that ON crush resulted inthe loss of approximately 70% of Brn3a-labelled RGCs by 1 week(Fig. 8). Neither the CW nor the 2RT laser augmented Brn3a-positive RGC survival in areas overlying and adjacent to laserspots. This was the case irrespective of whether irradiationoccurred 1d (Fig. 8J) or 7d (Fig. 8K) before the ON crush.
3.12. Effect of laser treatment on RGC survival after optic nervecrush: experiment 2
In order to substantiate the lack of neuroprotection afforded bylaser treatment on RGC survival after ON crush, we carried out asecond experiment in which we used an alternative marker ofRGCs, NeuN. In experiment 2, we performed sham or CW lasertreatment 1d prior to ON crush, killed the animals after 14d, andimmunolabelled retinal wholemounts for NeuN (see Materials andMethods). Representative images of NeuN immunolabelling in ONcrush and contralateral retinas, together with analysis of the data,are provided in Fig. 9. It can be seen that ON crush resulted in theloss of 85% of NeuN-positive RGCs after 14d. CW laser treatmentexerted no effect on neuronal survival when compared to shamanimals.
4. Discussion
4.1. Effect of retinal lasers on preconditioning pathways
The concept of stimulating endogenous production of survivalfactors as a potential therapy for protecting RGCs in neurodegen-erative conditions such as glaucoma is appealing, firstly, because it
obviates the need for delivery of exogenous molecules, and sec-ondly, because it inherently has the potential to encompass mul-tiple, relevant signalling pathways. The idea is most commonlyassociated with preconditioning strategies, which refer to thenotion that a mild or transient dose of a noxious stimulus increasesresistance to a subsequent, major injury. Research to date hasprovided grounds for hypothesizing that retinal lasers may fulfilsuch a preconditioning role. Clinically, lasers remain importanttools for treatment of visual impairment caused by diabetic mac-ular oedema (Bandello et al., 2012); however, the mechanism(s) bywhich vision is stabilised after treatment is unclear. Any discernedinformation relating to retinal signalling pathways activated bylaser treatment, therefore, not only reinforces their potential usefor neuroprotective preconditioning, but also more generally adds agreater understanding to their effects on the retina.
The cellular pathways responsible for preconditioning-inducedneuroprotection, and RGC preconditioning in particular, remainpoorly understood. Preconditioning responses appear to be multi-faceted. Furthermore, they are tissue-, and stimulus-specific. DNAmicroarray studies have shown that preconditioning stimuli affecthundreds of different genes (Gustavsson et al., 2007; Kamphuiset al., 2007). Yet, published studies typically concentrate on iden-tifying just one or two factors. In attempting to assess whetherretinal lasers stimulate preconditioning pathways, we concentratedon well-known transducers.
Inducible Hsps, in particular Hsp27 (Kretz et al., 2006; O'Reillyet al., 2010; Whitlock et al., 2005) and �70 (Ahn et al., 2008;Biermann et al., 2010), are firmly implicated as neuroprotectiveeffectors in retinal preconditioning modalities. We have previouslydemonstrated sustained upregulation of Hsp27 in response to bothCW and 2RT laser treatment (Chidlow et al., 2013). Results hereinconfirm this finding, but, surprisingly, provide minimal evidence
Fig. 5. Effect of CW or 2RT laser on expression of trophic factors. (AeC) Quantification of FGF-2, BDNF, IGF-1 mRNA levels at 6h, 1d and 7d after laser treatment. Values (representedas mean ± SEM, n ¼ 6) are normalised for housekeeping genes and expressed relative to the control group. **P < 0.01, *P < 0.05, by Pair-wise Fixed Reallocation Randomization Test(treated vs control). (D) Representative images of CNTF immunolabelling in retinal wholemounts in sham rats and at 24h and 7d after laser treatment. In sham rats, CNTF expression(green) is restricted to astrocytes lining blood vessels (arrow). At 24h after laser treatment, CNTF (green) is observed within astrocytes/Müller cell end- feet bordering the lesion,which is demarcated by immunolabelling for nestin (red). At 7d after laser treatment, CNTF (green) is observed within astrocytes/Müller cell end- feet within the lesion, again asdemarcated by nestin (red). Scale bar: 70 mm. (E) Representative CNTF and actin Western immunoblots in sham and lasered retinas together with quantification of CNTF proteinlevel at 1d and 7d after laser treatment. Values (represented as mean ± SEM, n ¼ 6) are normalised for actin and expressed relative to the control group. **P < 0.01, *P < 0.05, byANOVA followed by post-hoc Dunnett's Multiple Comparison Test (treated vs control). (For interpretation of the references to colour in this figure legend, the reader is referred tothe web version of this article.)
O. Shibeeb et al. / Experimental Eye Research 127 (2014) 77e90 85
for upregulation of two other inducible Hsps, namely Hsp70 orHsp22. Endogenous generation of neurotrophic factors is similarlythought to contribute to preconditioning-induced neuroprotection,but only a marginal increase in the level of the neurotrophin mostcommonly associated with neuronal preconditioning, BDNF(Marini et al., 2007), and no change in IGF-1 could be detected.Conversely, CNTF and FGF-2 were both robustly upregulated, withthe CNTF response particularly striking: after laser treatment, CNTF
expression was associated with astrocytes and Müller cells within,and immediately adjacent to, irradiated zones. Substantive evi-dence has highlighted the beneficial effect of supplementing CNTFon RGC survival after ON injury (Leaver et al., 2006b; van Adel et al.,2003; Zhang et al., 2005); thus, glial cell-derived CNTF should be ofassistance to injured, neighbouring RGCs.
Pro-inflammatory cytokines, nitric oxide and other reactiveoxygen species have traditionally been viewed as neurotoxic
Fig. 6. Responses of nitric oxide and NADPH oxidase following CW or 2RT laser treatment. (AeC) Quantification of eNOS, iNOS and NOX-2 mRNA levels at various time points afterlaser treatment. Values (represented as mean ± SEM, n ¼ 6) are normalised for housekeeping genes and expressed relative to the control group. **P < 0.01, *P < 0.05, by Pair-wiseFixed Reallocation Randomization Test (treated vs control). (DeG) Representative images of eNOS and NOX-2 immunolabelling. In sham retinas (D), eNOS is restricted to theendothelium of blood vessels (arrow). No discernible upregulation is evident after CW laser treatment (E). In sham retinas (F), NOX-2 is absent, however, at 3d after CW lasertreatment (G), some microglia express NOX-2 (arrow). Note: the asterisk indicates the position of the laser lesion. Scale bar: D, E ¼ 35 mm; F, G ¼ 17.5 mm. GCL, ganglion cell layer;INL, inner nuclear layer; ONL, outer nuclear layer.
O. Shibeeb et al. / Experimental Eye Research 127 (2014) 77e9086
mediators. Interestingly, however, their controlled release bymicroglia in advance of a major insult, as occurs with certain pre-conditioning stimuli, is associated with increased survival ofvulnerable neurons (Bell et al., 2005; Halder et al., 2013; Kunz et al.,2007; Shpargel et al., 2008). Our results show striking elevations ofIL-1b, TNFa, IL-6, as well as NOX-2, in the retinas of CW-treated rats,andmoderate elevations of the cytokines in rats receiving 2RT laser.
Fig. 7. Effect of CW or 2RT laser on expression of the survival-promoting signalling moleculretinas are shown. (B) Quantification of pAKT protein level after laser treatment. Values (repcontrol group. **P < 0.01, by ANOVA followed by post-hoc Dunnett's Multiple Comparison
Responses were largely confined to lasered areas. It is reasonable tohypothesise, therefore, that these factors may be able to confersome resistance to RGC injury. Neither laser, however, affectedsynthesis of IF-c, iNOS or eNOS, all of which have also beenimplicated in certain neuronal preconditioning regimes. Similar,essentially negative, results were obtained following investigationof the effect of the lasers on the expression of the anti-
e pAKT. (A) Representative pAKT and actin Western immunoblots in sham and laseredresented as mean ± SEM, n ¼ 6) are normalised for actin and expressed relative to thetest (treated vs control).
Fig. 8. Effect of CW or 2RT laser on RGC survival, as delineated by quantification of Brn3a neurons in photomicrographs fields overlying lesions, following ON crush. (AeI)Representative images of contralateral and treated wholemount retinas from each experimental group showing double labelling immunofluorescence of Brn3a (green) with nestin(red). (AeC) Contralateral retinas of sham, 2RT and CW animals, where nestin expression is restricted to blood vessels. (DeI) Treated retinas of sham, CW and 2RT animals in whichlaser was applied 1d (DeF) or 7d (GeI) before ON crush. ON crush causes a marked loss of Brn3a neurons and upregulated nestin expression in all animals; however, focalexpression of nestin overlying CW and 2RT lesions is still clearly visible (arrows). Scale bar: 70 mm. (J, K) Quantification of Brn3a neurons in the contralateral and treated retinas ofeach group. Laser treatment was delivered either 1d (J) or 7d (K) prior to ON crush. All animals were killed 7d after ON crush. Values represent mean ± SEM. For the 1d experiment:sham (n ¼ 6), CW (n ¼ 8), 2RT (n ¼ 7). For the 7d experiment: sham (n ¼ 6), CW (n ¼ 7), 2RT (n ¼ 6). ANOVA showed no statistically significant between the extent of RGC survival inthe sham, CW and 2RT groups (1d experiment, P ¼ 0.54; 7d experiment, P ¼ 0.99).
O. Shibeeb et al. / Experimental Eye Research 127 (2014) 77e90 87
inflammatory cytokines, IL-10 and TGFb which have previouslybeen shown to be secreted by preconditioned astrocytes (Gidday,2006).
Preconditioning-induced neuroprotection obviously dependson the co-ordinated regulation of multiple signalling cascades thatcomprise transcription factors, signal transducers and mitochon-drial and nuclear effectors. ATF-3 and c-FOS, transcription factorsthat bind to a common DNA site, have been implicated in neuronalpreconditioning responses (Zhang et al., 2011). Both lasers inducedsignificant local upregulations of ATF-3 and c-FOS, which wereassociated with RGCs and Müller cells, respectively. Evidence sug-gests ATF-3 promotes neuronal survival via activation of the Aktsignalling pathway (Nakagomi et al., 2003) In fact, a substantial
body of data has implicated Akt as a key player in preconditioning-induced neuroprotection in the retina (Dreixler et al., 2009b; Zhaoet al., 2006). We therefore examined whether laser treatment ac-tivates Akt. Our data showed both lasers induced significant in-creases in the phosphorylation levels of retinal Akt, which is knownto activate this enzyme (Hers et al., 2011). Regarding mitochondrialeffectors, preconditioning strategies can upregulate expression ofmembers of the pro-survival, bcl-2 family. Blocking the induction ofbcl-2 negated the beneficial effect of preconditioning in a model offocal brain ischemia (Shimizu et al., 2001) illustrating its impor-tance in this injury paradigm. Our results showed that neither laserinduced a significant upregulation in Bcl-2 mRNA or bcl-xl proteinat any of the time points analysed.
Fig. 9. Effect of CW laser on RGC survival, as delineated by quantification of NeuN neurons, 14d after ON crush. (AeC) Representative images of contralateral and treatedwholemount retinas from the sham and CW groups. Scale bar: 70 mm. (C) Quantification of NeuN-positive neurons. Sham or CW laser treatment was delivered 1d prior to ON crush.All animals were killed 14d after ON crush. Values represent mean ± SEM, n ¼ 6. Unpaired t-test showed no statistically significant between the extent of RGC survival in the shamand CW groups (P ¼ 0.88).
O. Shibeeb et al. / Experimental Eye Research 127 (2014) 77e9088
4.2. Effect of retinal lasers on RGC survival
As discussed above, both the CW and 2RT lasers induced local-ized upregulations of a number of well-documented (CNTF, FGF-2Hsp27, pAKT) or putative (cFOS, ATF-3) RGC survival factors. Itwas, therefore, unexpected that neither laser displayed any pro-pensity to delay RGC loss following ON crush. There are at leastthree plausible explanations that would account for the lack ofneuroprotection observed in this study. The first possibility is thatthe survival pathways stimulated did afford ameasure of protectionto RGCs, but, the use of Brn3a labelling of RGCs was ill-suited todetect the effect. The second possibility is that the timing of thepreconditioning insult was sub-optimal. The third possibility is thatneither laser induced an endogenous preconditioning response ofeither the necessary magnitude or relevant neurochemicalcomposition to protect cells in such a robust model of RGC loss asthe one employed herein.
Brn3a is an excellent and well-employed marker for identifica-tion of adult rat RGCs inwholemounts as well as transverse sections(Nadal-Nicolas et al., 2012; Nadal-Nicol�as et al., 2009); however,the time-course of Brn3a disappearance in models of ON injuryoccurs prior to the disappearance of the actual RGC soma. It can beargued, therefore, that this characteristic of Brn3a makes it morechallenging for neuroprotection studies than if a more stablemarker of RGCs were used. Alternatives to Brn3a include retrogradetracers, such as fluorogold (Salinas-Navarro et al., 2009) structuralproteins specific to RGCs, for example b3-tubulin (Chen et al., 2011),or histological stains, e.g. Nissl (Li et al., 2007). Nevertheless, thesemarkers have notable disadvantages. Fluorogold is increasinglyassociated with microglia phagocytosing dying RGCs, an effectdetectable as early as 4 days after ON injury (Peinado-Ramon et al.,1996); Nissl labels all retinal cells, not specifically RGCs, making
accurate quantification problematic; b3-Tubulin labels RGC axonswithin the overlying nerve fibre layer as well as somas, necessarilyprecluding quantification within the central retina. Concerns overthe use of Brn3a in neuroprotection experiments were alleviated ina study demonstrating that administration of BDNF, a well-characterized RGC survival factor, preserved Brn3a- andfluorogold-labelled RGCs equally well following ON injury(Sanchez-Migallon et al., 2011). Other neuroprotection studiesinvolving ON crush have similarly made use of Brn3a, (Hellstromet al., 2011; Mead et al., 2013). It is logical to assume, therefore,that Brn3a was an appropriate marker to use. This conclusion isfurther supported by the lack of neuroprotection observed inexperiment 2 when NeuN was utilized instead of Brn3a to docu-ment RGC number.
In order to address the issue of timing, two separate experi-ments were performed in which the retina was subjected to pre-conditioning either 1 day or 7 days prior to ON crush. Thiscombined approach encompassed the time frame of increasedsurvival factor levels delineated in the first part of our study.Neither strategy, however, resulted in any significant delay incrush-induced RGC loss. There is known to be a lag phase, lastingseveral days, between ON crush/transection and actual RGC loss.Thus, it is conceivable that both strategies resulted in maximalneuronal resistance occurring too early; laser treatment shouldperhaps have been deferred until after the crush procedure. Indeed,is known that temporary protective environments can be inducedas early as minutes after preconditioning stimuli (Gidday, 2006).Moreover, pharmacological induction of Hsp27 by simvastatin hasbeen shown to delay RGC loss after ON crush when the drug wasadministered 24 he48 h post-surgery (Kretz et al., 2006). Never-theless, preconditioning strategies have often proven successful inrescuing RGCswhen performed 1e3 days before the insult (Dreixler
O. Shibeeb et al. / Experimental Eye Research 127 (2014) 77e90 89
et al., 2009a; Kamphuis et al., 2007; Roth et al., 1998; Salido et al.,2013) As such, there are no compelling reasons to believe thetimings of the laser therapy in the current study were injudicious.
The third possibility relates to the magnitude, or perhapsneurochemical composition, of the endogenous protectiveresponse induced by laser treatment. Regarding magnitude, laserenergy is principally absorbed by the RPE. At the settings used inthis study, disruption to the retina is essentially restricted to thephotoreceptors, which, particularly in the case of the 2RT laser isrelatively minor (Wood et al., 2013). Although various survivalfactors, typically associated with glial cells, were upregulated afterlaser treatment, RGCs may not have adopted a sufficiently resistantphenotype to overcome ON crush. For example, both lasers causedupregulation of Hsp27 in astrocytes, but were insufficient to induceits expression in the majority of overlying RGCs (Chidlow et al.,2013). Higher energy settings would achieve this goal, but greatercollateral damage to the retina would ensue. As regards the path-ways activated by laser treatment, while various pro-survival fac-tors were undoubtedly upregulated, it was conspicuous that otherkey factors implicated in neuronal preconditioning remainedlargely unchanged (e.g. BDNF, Hsp70, IGF-1 and bcl-2 familymembers). Bcl-2 overexpression promotes the survival of RGCsafter ON crush (Leaver et al., 2006a), as does exogenous adminis-tration of BDNF (Parrilla-Reverter et al., 2009) or IGF-1 (Hommaet al., 2007). The deficit of these factors could well have contrib-uted to the lack of neuroprotection observed. Moreover, CW lasertreatment caused activation of microglia with associated upregu-lation of pro-inflammatory genes, including IL-1b and TNFa,together with induction of NADPH oxidase, which is responsible forproduction of reactive oxygen species. While pro-inflammatorymediators and reactive oxygen species are associated with certainpreconditioning stimuli (Shpargel et al., 2008), it must also berecognised that they are traditionally viewed as neurotoxic to RGCsand their production may have negated the beneficial effects ofother survival factors.
Finally, it must be acknowledged that there are very few reportsof preconditioning-induced neuroprotection of RGCs after ONcrush. Instead, most preconditioning studies have focused onpreservation of RGCs after ischemia-reperfusion. It is likely that thelatter injury paradigm is more amenable to preconditioningstrategies.
How can our results be reconciled with those of Nork and col-leagues (Nork et al., 2000), who showed that laser photocoagula-tion effected localised RGC survival? The answer probably lieswithin one or more of the many methodological differences be-tween the two studies. It is tempting to speculate that theconsiderably greater radiant exposure level delivered by Norkinstigated a correspondingly greater preconditioning response.This viewpoint is partially supported by the results herein andthose of our previous work (Chidlow et al., 2013). Specifically, theCW laser has a higher radiant exposure than the 2RT laser and itsapplication typically resulted in more efficacious upregulations ofpreconditioning-related molecules. The radiant exposures used inthe current study approximate those used in the clinic for treat-ment of DME and would potentially, at least in the case of the 2RTlaser, be of therapeutic interest, whereas those of Nork et al. (2000)are not, as was acknowledged in the publication itself. Yet, it mustalso be recognised that Nork did perform ON transection in place ofglaucoma in a single animal and found no protection from laserphotocoagulation. While it would be unwise to infer much from asingle animal, the possibility cannot be dismissed that laser treat-ment augments RGC survival during experimental glaucoma butnot after mechanical injury to the ON, despite the great similaritiesbetween the pathologies of the twomodels of injury (Chidlow et al.,2011b; Salinas-Navarro et al., 2010). In fact, Nork put forward an
appealing hypothesis to account for their findings of laser-inducedprotection of RGCs from glaucomatous damage, suggesting that thestresses associated with glaucoma compromise the efficient func-tioning of the glutamateeglutamine cycle between photoreceptorsand Müller cells, resulting in the gradual accumulation of excito-toxic glutamate. Laser photocoagulation destroys photoreceptors,hence removing the source of the toxic glutamate, and therebyprotecting overlying RGCs from excitotoxicity. It would have beenpreferable to have additionally performed the current study in amodel of experimental glaucoma to test this hypothesis; however,this was impractical for the following reasons: (1) laser treatmentof the retina requires the use of pigmented rats but our well-characterized laser model of glaucoma is calibrated for use withnon-pigmented rats; (2) ON crush leads to a uniform loss of RGCsthroughout the retina, which enables straightforward, reproduciblecomparison between RGC survival in lasered and sham-laseredretinas, whereas during experimental glaucoma RGC death issectorial (Vidal-Sanz et al., 2012), and also displays notable inter-animal variability, which would result in a very high signal-to-noise ratio.
Funding and role of funding source
This study was supported in part by Ellex R&D Pty Ltd, Adelaide,Australia. The funding source had no influence on study design,data analyses, interpretation of data, or writing of the manuscript.
Acknowledgement
The authors are grateful to Mark Daymon for expert technicalassistance and Malcolm Plunkett for helpful advice.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.exer.2014.07.008.
References
Ahn, J., Piri, N., Caprioli, J., Munemasa, Y., Kim, S.H., Kwong, J.M.K., 2008. Expressionof heat shock transcription factors and heat shock protein 72 in rat retina afterintravitreal injection of low dose N-methyl-d-aspartate. Neurosci. Lett. 433,11e16.
Bandello, F., Cunha-Vaz, J., Chong, N.V., Lang, G.E., Massin, P., Mitchell, P., Porta, M.,Prunte, C., Schlingemann, R., Schmidt-Erfurth, U., 2012. New approaches for thetreatment of diabetic macular oedema: recommendations by an expert panel.Eye 26, 485e493.
Bell, R.M., Cave, A.C., Johar, S., Hearse, D.J., Shah, A.M., Shattock, M.J., 2005. Pivotalrole of NOX-2-containing NADPH oxidase in early ischemic preconditioning.FASEB J. 19, 2037e2039.
Biermann, J., Lagr�eze, W.A., Dimitriu, C., Stoykow, C., Goebel, U., 2010. Pre-conditioning with inhalative carbon monoxide protects rat retinal ganglioncells from Ischemia/Reperfusion injury. Invest Ophthalmol. Vis. Sci. 51,3784e3791.
Brinkmann, R., Hüttmann, G., R€ogener, J., Roider, J., Birngruber, R., Lin, C.P., 2000.Origin of retinal pigment epithelium cell damage by pulsed laser irradiance inthe nanosecond to microsecond time regimen. Lasers Surg. Med. 27, 451e464.
O. Shibeeb et al. / Experimental Eye Research 127 (2014) 77e9090
Chidlow, G., Daymon, M., Wood, J.P., Casson, R.J., 2011a. Localization of a wide-ranging panel of antigens in the rat retina by immunohistochemistry: com-parison of Davidson's solution and formalin as fixatives. J. Histochem Cytochem.59, 884e898.
Chidlow, G., Ebneter, A., Wood, J.M., Casson, R., 2011b. The optic nerve head is thesite of axonal transport disruption, axonal cytoskeleton damage and putativeaxonal regeneration failure in a rat model of glaucoma. Acta Neuropathol. 121,737e751.
Chidlow, G., Shibeeb, O., Plunkett, M., Casson, R.J., Wood, J.P., 2013. Glial cell andinflammatory responses to retinal laser treatment: comparison of a conven-tional photocoagulator and a novel, 3-nanosecond pulse laser. Invest. Oph-thalmol. Vis. Sci. 54, 2319e2332.
Dreixler, J.C., Hemmert, J.W., Shenoy, S.K., Shen, Y., Lee, H.T., Shaikh, A.R.,Rosenbaum, D.M., Roth, S., 2009b. The role of Akt/protein kinase B subtypes inretinal ischemic preconditioning. Exp. Eye Res. 88, 512e521.
Faktorovich, E.G., Steinberg, R.H., Yasumura, D., Matthes, M.T., LaVail, M.M., 1990.Photoreceptor degeneration in inherited retinal dystrophy delayed by basicfibroblast growth factor. Nature 347, 83e86.
Faktorovich, E.G., Steinberg, R.H., Yasumura, D., Matthes, M.T., LaVail, M.M., 1992.Basic fibroblast growth factor and local injury protect photoreceptors from lightdamage in the rat. J. Neurosci. 12, 3554e3567.
Franco, P.J., Fernandez, D.C., Sande, P.H., Keller Sarmiento, M.I., Chianelli, M.,S�aenz, D.A., Rosenstein, R.E., 2008. Effect of bacterial lipopolysaccharide onischemic damage in the rat retina. Invest. Ophthalmol. Vis. Sci. 49, 4604e4612.
Gustavsson, M., Wilson, M.A., Mallard, C., Rousset, C., Johnston, M.V., Hagberg, H.,2007. Global gene expression in the developing rat brain after hypoxic pre-conditioning: involvement of apoptotic mechanisms? Pediatr. Res. 61, 444e450.
Halder, S.K., Matsunaga, H., Yamaguchi, H., Ueda, H., 2013. Novel neuroprotectiveaction of prothymosin alpha-derived peptide against retinal and brain ischemicdamages. J. Neurochem. 125, 713e723.
Hellstrom, M., Pollett, M.A., Harvey, A.R., 2011. Post-injury delivery of rAAV2-CNTFcombined with short-term pharmacotherapy is neuroprotective and promotesextensive axonal regeneration after optic nerve trauma. J. Neurotrauma 28,2475e2483.
Hers, I., Vincent, E.E., Tavar�e, J.M., 2011. Akt signalling in health and disease. Cell.Signal. 23, 1515e1527.
Homma, K., Koriyama, Y., Mawatari, K., Higuchi, Y., Kosaka, J., Kato, S., 2007. Earlydownregulation of IGF-I decides the fate of rat retinal ganglion cells after opticnerve injury. Neurochem. Int. 50, 741e748.
Kamphuis, W., Dijk, F., Kraan, W., Bergen, A.A., 2007. Transfer of lens-specifictranscripts to retinal RNA samples may underlie observed changes incrystallin-gene transcript levels after ischemia. Mol. Vis. 13, 220e228.
Kretz, A., Schmeer, C., Tausch, S., Isenmann, S., 2006. Simvastatin promotes heatshock protein 27 expression and Akt activation in the rat retina and protectsaxotomized retinal ganglion cells in vivo. Neurobiol. Dis. 21, 421e430.
Kunz, A., Park, L., Abe, T., Gallo, E.F., Anrather, J., Zhou, P., Iadecola, C., 2007. Neu-rovascular protection by ischemic tolerance: role of nitric oxide and reactiveoxygen species. J. Neurosci. 27, 7083e7093.
Leaver, S.G., Cui, Q., Bernard, O., Harvey, A.R., 2006a. Cooperative effects of bcl-2 andAAV-mediated expression of CNTF on retinal ganglion cell survival and axonalregeneration in adult transgenic mice. Eur. J. Neurosci. 24, 3323e3332.
Leaver, S.G., Cui, Q., Plant, G.W., Arulpragasam, A., Hisheh, S., Verhaagen, J.,Harvey, A.R., 2006b. AAV-mediated expression of CNTF promotes long-termsurvival and regeneration of adult rat retinal ganglion cells. Gene Ther. 13,1328e1341.
Liu, C., Peng, M., Laties, A.M., Wen, R., 1998. Preconditioning with bright light evokesa protective response against light damage in the rat retina. J. Neurosci. 18,1337e1344.
Maier, P.C., Funk, J., Schwarzer, G., Antes, G., Falck-Ytter, Y.T., 2005. Treatment ofocular hypertension and open angle glaucoma: meta-analysis of randomisedcontrolled trials. BMJ 331, 134.
Marini, A.M., Jiang, X., Wu, X., Pan, H., Guo, Z., Mattson, M.P., Blondeau, N.,Novelli, A., Lipsky, R.H., 2007. Preconditioning and neurotrophins: a model forbrain adaptation to seizures, ischemia and other stressful stimuli. Amino Acids32, 299e304.
Mead, B., Logan, A., Berry, M., Leadbeater, W., Scheven, B.A., 2013. Intravitreallytransplanted dental pulp stem cells promote neuroprotection and axonregeneration of retinal ganglion cells after optic nerve injury. Invest. Oph-thalmol. Vis. Sci. 54, 7544e7556.
Nadal-Nicolas, F.M., Jimenez-Lopez, M., Salinas-Navarro, M., Sobrado-Calvo, P.,Alburquerque-Bejar, J.J., Vidal-Sanz, M., Agudo-Barriuso, M., 2012. Wholenumber, distribution and co-expression of brn3 transcription factors in retinalganglion cells of adult albino and pigmented rats. PLoS ONE 7, e49830.
Nadal-Nicol�as, F.M., Jim�enez-L�opez, M., Sobrado-Calvo, P., Nieto-L�opez, L., C�anovas-Martínez, I., Salinas-Navarro, M., Vidal-Sanz, M., Agudo, M., 2009. Brn3a as a
marker of retinal ganglion cells: qualitative and quantitative time coursestudies in naïve and optic nerveeinjured retinas. Invest. Ophthalmol. Vis. Sci.50, 3860e3868.
Nakagomi, S., Suzuki, Y., Namikawa, K., Kiryu-Seo, S., Kiyama, H., 2003. Expressionof the activating transcription factor 3 prevents c-Jun N-terminal kinase-induced neuronal death by promoting heat shock protein 27 expression andAkt activation. J. Neurosci. 23, 5187e5196.
Nork, T.M., Poulsen, G.L., Nickells, R.W., Ver Hoeve, J.N., Cho, N.C., Levin, L.A.,Lucarelli, M.J., 2000. Protection of ganglion cells in experimental glaucoma byretinal laser photocoagulation. Arch. Ophthalmol. 118, 1242e1250.
O'Reilly, A.M., Currie, R.W., Clarke, D.B., 2010. HspB1 (Hsp 27) expression andneuroprotection in the retina. Mol. Neurobiol. 42, 124e132.
Parrilla-Reverter, G., Agudo, M., Sobrado-Calvo, P., Salinas-Navarro, M., Villegas-Perez, M.P., Vidal-Sanz, M., 2009. Effects of different neurotrophic factors on thesurvival of retinal ganglion cells after a complete intraorbital nerve crushinjury: a quantitative in vivo study. Exp. Eye Res. 89, 32e41.
Peinado-Ramon, P., Salvador, M., Villegas-Perez, M.P., Vidal-Sanz, M., 1996. Effects ofaxotomy and intraocular administration of NT-4, NT-3, and brain-derivedneurotrophic factor on the survival of adult rat retinal ganglion cells. A quan-titative in vivo study. Invest. Ophthalmol. Vis. Sci. 37, 489e500.
Pfaffl, M.W., Horgan, G.W., Dempfle, L., 2002. Relative expression software tool(REST) for group-wise comparison and statistical analysis of relative expressionresults in real-time PCR. Nucleic Acids Res. 30, e36.
Roth, S., 2004. Endogenous neuroprotection in the retina. Brain Res. Bull. 62,461e466.
Salido, E.M., Dorfman, D., Bordone, M., Chianelli, M., Gonzalez Fleitas, M.F.,Rosenstein, R.E., 2013. Global and ocular hypothermic preconditioning protectthe rat retina from ischemic damage. PLoS ONE 8, e61656.
Salinas-Navarro, M., Alarcon-Martinez, L., Valiente-Soriano, F.J., Jimenez-Lopez, M.,Mayor-Torroglosa, S., Aviles-Trigueros, M., Villegas-Perez, M.P., Vidal-Sanz, M.,2010. Ocular hypertension impairs optic nerve axonal transport leading toprogressive retinal ganglion cell degeneration. Exp. Eye Res. 90, 168e183.
Salinas-Navarro, M., Jim�enez-L�opez, M., Valiente-Soriano, F.J., Alarc�on-Martínez, L.,Avil�es-Trigueros, M., Mayor, S., Holmes, T., Lund, R.D., Villegas-P�erez, M.P., Vidal-Sanz, M., 2009. Retinal ganglion cell population in adult albino and pigmentedmice: a computerized analysis of the entire population and its spatial distri-bution. Vis. Res. 49, 637e647.
Sanchez-Migallon, M.C., Nadal-Nicolas, F.M., Jimenez-Lopez, M., Sobrado-Calvo, P.,Vidal-Sanz, M., Agudo-Barriuso, M., 2011. Brain derived neurotrophic factormaintains Brn3a expression in axotomized rat retinal ganglion cells. Exp. EyeRes. 92, 260e267.
Shimizu, S., Nagayama, T., Jin, K.L., Zhu, L., Loeffert, J.E., Watkins, S.C., Graham, S.H.,Simon, R.P., 2001. bcl-2 Antisense treatment prevents induction of tolerance tofocal ischemia in the rat brain. J. Cereb. Blood Flow Metab. 21, 233e243.
Shpargel, K.B., Jalabi, W., Jin, Y., Dadabayev, A., Penn, M.S., Trapp, B.D., 2008. Pre-conditioning paradigms and pathways in the brain. Clevel. Clin. J. Med. 75, S77.
van Adel, B.A., Kostic, C., Deglon, N., Ball, A.K., Arsenijevic, Y., 2003. Delivery ofciliary neurotrophic factor via lentiviral-mediated transfer protects axotomizedretinal ganglion cells for an extended period of time. Hum. Gene Ther. 14,103e115.
Vidal-Sanz, M., Salinas-Navarro, M., Nadal-Nicolas, F.M., Alarcon-Martinez, L.,Valiente-Soriano, F.J., de Imperial, J.M., Aviles-Trigueros, M., Agudo-Barriuso, M.,Villegas-Perez, M.P., 2012. Understanding glaucomatous damage: anatomicaland functional data from ocular hypertensive rodent retinas. Prog. Retin. EyeRes. 31, 1e27.
Weinreb, R.N., Levin, L.A., 1999. IS neuroprotection a viable therapy for glaucoma?Arch. Ophthalmol. 117, 1540e1544.
Wood, J.P., Shibeeb, O., Plunkett, M., Casson, R.J., Chidlow, G., 2013. Retinal damageprofiles and neuronal effects of laser treatment: comparison of a conventionalphotocoagulator and a novel 3-nanosecond pulse laser. Invest. Ophthalmol. Vis.Sci. 54, 2305e2318.
Xiao, M., Sastry, S.M., Li, Z.Y., Possin, D.E., Chang, J.H., Klock, I.B., Milam, A.H., 1998.Effects of retinal laser photocoagulation on photoreceptor basic fibroblastgrowth factor and survival. Invest. Ophthalmol. Vis. Sci. 39, 618e630.
Zhang, C.W., Lu, Q., You, S.W., Zhi, Y., Yip, H.K., Wu, W., So, K.F., Cui, Q., 2005. CNTFand BDNF have similar effects on retinal ganglion cell survival but differentialeffects on nitric oxide synthase expression soon after optic nerve injury. Invest.Ophthalmol. Vis. Sci. 46, 1497e1503.
Zhang, S.J., Buchthal, B., Lau, D., Hayer, S., Dick, O., Schwaninger, M., Veltkamp, R.,Zou, M., Weiss, U., Bading, H., 2011. A signaling cascade of nuclear calcium-CREB-ATF3 activated by synaptic NMDA receptors defines a gene repressionmodule that protects against extrasynaptic NMDA receptor-induced neuronalcell death and ischemic brain damage. J. Neurosci. 31, 4978e4990.
Zhao, H., Sapolsky, R.M., Steinberg, G.K., 2006. Phosphoinositide-3-kinase/akt sur-vival signal pathways are implicated in neuronal survival after stroke. Mol.Neurobiol. 34, 249e270.
