Aurintricarboxylic acid promotes survival and regeneration of axotomised retinal ganglion cells in vivo

Neuropharmacology 39 (2000) 889–902www.elsevier.com/locate/neuropharm

Aurintricarboxylic acid promotes survival and regeneration ofaxotomised retinal ganglion cells in vivo

Peter Heiduschka*, Solon ThanosDepartment of Experimental Ophthalmology, University of Mu¨nster, Domagkstraße 15, D-48149 Mu¨nster, Germany

Accepted 22 November 1999

Abstract

Aurintricarboxylic acid (ATA) has been used as an anti-apoptotic drug to counteract ischemic or cytotoxic injury to neurons.We investigated whether ATA has a neuroprotective effect on axotomised, adult retinal ganglion cells (RGC) as a model fortraumatic neuronal cell death. A solution of ATA was injected into the vitreous body of rat eyes whose optic nerves had been cut.In controls, 14% of RGC survived 14 days after axotomy, whereas 44% of RGC survived after a single injection of ATA solution,and 59% survived when the injection was repeated after 7 days. A single injection of ATA 1 day after axotomy rescued 58% ofRGC. However, injection of ATA 4 days after axotomy did not influence the survival of RGC, indicating that crucial, irreversiblecascades of death are initiated prior to this point in time. The TUNEL technique was used to visualise apoptotic ganglion cells andrevealed that 4 days after axotomy their number was significantly less in retinas whose optic nerves were axotomised and treatedwith ATA, than those of controls. As a consequence of neuroprotection, more RGC were recruited to regenerate into a peripheralnerve graft used to replace the cut optic nerve. In this paradigm, ATA-treated RGC extended significantly more axons within thegraft than control RGC. This number could be increased by a second injection of ATA 7 days after axotomy. These data showthat ATA is not only able to delay post-traumatic neuronal death but also enhances the extent of axonal regeneration in vivo.2000 Elsevier Science Ltd. All rights reserved.

Keywords:Rat retina; Ganglion cells; Nerve injury; Neuroprotective drugs; Axonal regeneration; Fluorescent dyes; Aurintricarboxylic acid

1. Introduction

Neurons of the central nervous system (CNS) are vul-nerable to axotomy and usually degenerate thereafter.Being part of the CNS, retinal ganglion cells (RGC) alsorespond to an optic nerve lesion with death and serve asa model system to analyse the cascades initiated byinjury. Under experimental conditions, some RGC canbe rescued if they are provided with an environment thatallows their cut axons to regenerate (Vidal-Sanz et al.,1987). For this, the cut end of the optic nerve is suturedto an autologous peripheral nerve segment, where a cer-tain percentage of RGC are able to survive and regrowtheir axons. In this model, functional recovery can alsobe achieved by positioning the free end of the graft into

* Corresponding author. Tel.:+49-251-83-56917; fax:+49-251-83-56916.

E-mail address:[email protected] (P. Heiduschka).

0028-3908/00/$ - see front matter 2000 Elsevier Science Ltd. All rights reserved.PII: S0028-3908 (99)00245-2

the natural central targets of the RGC axons (Thanos etal., 1997).

Even under growth-permissive conditions and despitethe presence of the peripheral nerve graft, only a smallportion of the RGC regenerate. This highlights theimportance of supplementing peripheral nerve transplan-tation with neuroprotective measures. The latter shouldbe directed at prevention of apoptosis, since injury to theoptic nerve triggers off apoptotic RGC death (Berkelaaret al., 1994; Garcia-Valenzuela et al., 1994; Quigley etal., 1995) after which RGC are rapidly phagocytosed bymicroglial cells.

Aurintricarboxylic acid (ATA) is a well-known inhibi-tor of endonucleases, which split DNA in a key stepduring programmed cell death (apoptosis). In fact, ATAreduced the rate of apoptosis in different experimentalmodels, both in vitro and in vivo. Although its action asan endonuclease inhibitor has been established, ATAalso acts at different sites in the cell.

In most of the models used until now, ATA was

890 P. Heiduschka, S. Thanos / Neuropharmacology 39 (2000) 889–902

applied to prevent cell death caused by ischemia or glut-amate-induced neurotoxicity. Retinal lesions induced byischemia (Lam et al., 1995; Rosenbaum et al., 1997),NMDA or kainate (Zeevalk et al., 1993, 1995) could beattenuated by ATA. However, Zeevalk and colleaguesperformed this study in vitro, with lactate dehydrogenaserelease serving as the criterion of cell toxicity, whereasthe other two studies considered thickness of the retinallayers as a parameter. In order to clearly establish thatRGC are the targets of ATA-neuroprotection, they haveto be identified as such, as has been achieved by retro-grade labelling in this study.

We first studied whether ATA influences the survivalof RGC after a traumatic injury in the adult rat retina.Although we anticipated a protective effect of ATA, adetailed study of survival of RGC was necessary. Sec-ondly, we studied whether ATA interferes with cleavageof DNA using the TUNEL technique. Finally, a segmentof peripheral nerve was grafted to replace the cut opticnerve and the number of RGC whose axons grew intothe graft after treatment with ATA was quantified, inorder to determine if ATA also influences axonal regen-eration, which is the novelty of this study.

2. Materials and methods

Adult male Sprague–Dawley rats (n=144) weighing200–250 g were used throughout this study. Surgery wasperformed under intraperitoneal Ketanest/Xylaxineanaesthesia (Ketanest: Parke–Davis, 50 mg/ml ketaminehydrochloride, 0.2 ml/100 g body weight, Xylazine:Medistar, 2%, 0.05 ml/100 g body weight).

In the axotomy experiments (n=91), the left opticnerve was intraorbitally exposed and completely cutwithout affecting retinal blood supply. Small crystals of4Di–10ASP [N-4-4-(4-didecylaminostyryl)-N-methyl-propidium iodide] (Molecular Probes, Eugene, Oregon)were moistened with incomplete Freund’s adjuvant(SIGMA) and placed en loco at the cross-sectional areaof the ocular stump of the optic nerve in order to retro-gradely label the RGC (Fig. 1A).

In the transplantation experiments (modified fromVidal-Sanz et al., 1987,n=36), a piece of the right sciaticnerve measuring between 1.5 and 2.0 cm was used asan autologous nerve graft. The proximal end of the graftwas sutured at the ocular stump of the optic nerve. Thegraft was placed into a groove drilled into the skull fromthe orbit roof to the ipsilateral occipital cortex (Fig. 1C).

To reduce intraocular pressure and enable injection oflarger volumes of ATA, 10µl of anterior chamber fluidwas removed by penetrating the cornea with a glasscapillary. The same volume of ATA solution wasinjected through the sclera with another glass capillary(Fig. 1B). Care was taken to avoid damage to the lenswhile injecting because this could cause either inflam-

Fig. 1. Scheme of experimental procedures. A: Transection of opticnerve and application of fluorescent dye. B: Intravitreal injection ofATA solution through the sclera using a glass capillary. C: Blind-ending peripheral graft with regenerating axons which is re-exposedand cut after 14 days for application of the dye. 1 — cornea, 2 —anterior chamber, 3 — lens, 4 — vitreous body, 5 — retina with RGC,6 — stump of the optic nerve, 7 — crystals of fluorescent dye.

matory responses or cataracts. Moreover, ocular injurywas shown to promote survival of RGC (Mansour-Robaey et al., 1994), distorting the evaluation of theinfluence of ATA. Lens injury also occurred in a fewcases (four rats), but these animals were excluded fromthis study. Due to its poor solubility in water, ATA wasfirst dissolved in dimethylsulphoxide (DMSO), and ali-quots of this solution were diluted to yield solutions ofdefined ATA concentrations in phosphate-bufferedsaline (PBS) with 10 vol% DMSO as the vehicle. Ani-

891P. Heiduschka, S. Thanos / Neuropharmacology 39 (2000) 889–902

mals which did not receive an injection or which wereinjected with the vehicle alone served as controls.

To evaluate the neuroprotective effects of ATA in ratswith both axotomy and transplantation, the eyes wereenucleated 2 or 4 weeks after axotomy. Retinal whole-mounts were prepared on black nitro-cellulose filters(Sartorius, Go¨ttingen) and fixed overnight at 4°C in 4%paraformaldehyde in 0.1 M PBS, pH 7.4. After embed-ding the retina in Mowiol (Hoechst) on glass slides, lab-elled RGC were counted using a fluorescent microscope(Axiophot, Carl Zeiss) equipped with a grid spanning anarea of 0.39 mm2. In addition, fluorescent microglialcells were counted which were labelled due to phago-cytosis of RGC (Thanos et al., 1994). Counting was per-formed manually, i.e. without computer assistance. Lab-elled RGC and microglial cells could be easilydistinguished by their distinct morphologies, sizes andstaining patterns (Fig. 2). Taking into account the unevendistribution of RGC in the retina and in order to obtainreproducible values, as many sectors as possible werecounted on each retina (36±15). Calculated values ofaverage cell density, standard deviations and statisticalsignificance (two-tailed Student’s test) were weightedaccording to the number of sectors.

Control rats (n=9) were sacrificed 4 days after

Fig. 2. Typical appearance of RGC (arrows) and microglial cells(arrowheads) which allows clear distinction between these two typesof cells in the whole-mount of a retina, here 14 days after axotomyand after injection of 10% DMSO in PBS. The open arrow indicatesa microglial cell just contacting an RGC. Scale bar: 50µm.

axotomy and labelling, and labelled RGC were countedin order to obtain the control value.

The ApopTag In Situ Apoptosis Detection Kit(Oncor, Gaithersburg, MD) was used to evaluate theextent of apoptotic death. Experimental animals (n=8)where first perfused with PBS and then with 4% parafor-maldehyde solution, 4 or 8 days after axotomy. The eyeswere enucleated and embedded in Tissue-Tek (SakuraFinetek, Torrance, CA), frozen in liquid nitrogen andstored at280°C, after which 15µm thick cryostat-sec-tions were cut. The detection of apoptotic cells consistsof coupling digoxigenin-conjugated dUTP to the 39-OHends of double- or single-stranded DNA, catalysed byterminal deoxynucleotidyl transferase and subsequentdetection of bound digoxigenin by a specific antibodytagged with rhodamine or fluoresceine. RGC were dou-ble-labelled with Fluoro-Gold (Fluorochrome, Engle-wood, CO) to unequivocally identify them as such. Flu-orescent, apoptotic cells in the ganglion cell layerwere counted.

3. Results

Applying 4Di–10ASP to the cross-sectional area ofthe optic nerve enabled 776±90 RGC/mm2 to be lab-elled, which is almost 30% of all RGC (Fig. 3A). Quanti-tative analysis was possible, as this value was consistentin all animals studied.

3.1. Survival of RGC after axotomy

In the first set of experiments, we examined whetherintravitreal injections of ATA solution influenced thesurvival of RGC (Figs. 3 and 4). In controls injectedwith 10% DMSO dissolved in PBS (Fig. 3C), a slightbut hardly significant (P<0.055) increase of RGC sur-vival was observed after 14 days (18±5% survival rate)as compared to the animals without injection (14±2%,Fig. 3B). A low concentration of ATA (6µg/ml, approx.14 µM) had almost no effect compared to controls(22±2%). The number of surviving RGC increased athigher concentrations of ATA, reaching its peak whensaturated ATA solution was injected (44±5%). Increas-ing concentrations of ATA improved RGC survival andreduced the number of fluorescently labelled microglia.Another increase in the survival rate of RGC (59±10%)could be achieved by repeating the injection of a satu-rated solution of ATA 1 week after axotomy (Fig. 3D).

It is known that axotomised RGC die progressivelyafter optic nerve injury and undergo at least two phasesof death, a rapid one immediately after injury, and adelayed one beginning after the fifth post-axotomy day(Mansour-Robaey et al., 1994). In order to monitor thereversibility of axotomy-induced degeneration and dis-appearance of RGC, single injections of ATA were made


Fig. 3. Sectors of retinal whole-mounts with RGC stained retrogradely with 4Di–10ASP. A: Control animal without injection, retina was prepared4 days after axotomy. Stained RGC are scattered densely across the retina. In B–D, retina was prepared 14 days after axotomy. B: Axotomisedcontrol animal without injection. Most of the RGC have disappeared, and fluorescent microglial cells are visible instead. C: Axotomised controlanimal with injection of vehicle solution. There is no obvious difference compared with B. D: Injection of saturated ATA solution immediatelyafter axotomy and after 1 week. Many RGC have survived, and only a few microglial cells can be seen.

at various time points after injury, ranging from 1 to 4days after transection of the optic nerve (Fig. 5). Itappeared that injection 1 day after axotomy resulted insurvival of 58±14% which is significantly more (P<0.02)than that obtained when injection was performed at thetime of injury. Fewer RGC were rescued when ATA wasinjected 2 or 3 days after injury (37±9% and 31±4%,respectively). No beneficial effect could be observed

when ATA was injected 4 days after axotomy (15±4%,Fig. 5).

3.2. Apoptotic cells in the retina

As mentioned previously, axotomised RGC undergoapoptotic death. If it is assumed that ATA influences theextent of apoptosis, then the number of apoptotic cells


Fig. 3. (continued)

in the retina should vary between controls (vehicleinjected) and eyes treated with ATA solution. For thispurpose, numbers of apoptotic cells were determined 4and 8 days after axotomy (Figs. 6 and 7). We found asignificantly lower number of apoptotic cells in the gang-lion cell layer (GCL) 4 days after axotomy in eyesinjected with a saturated solution of ATA (Figs 6C and7). After 8 days, the total number of apoptotic cells in thevehicle-treated retina decreased, whereas more apoptoticRGC were found after treatment with ATA (Fig. 7).

3.3. Regeneration of RGC axons

The potential of RGC to regenerate an axon under theinfluence of ATA was tested in the neural transplantationmodel where the cut optic nerve was replaced with anautologous sciatic nerve segment (Vidal-Sanz et al.,1987, Fig. 8). Without injection, 72±17 RGC/mm2 couldbe stained from the peripheral nerve graft as depicted inFig. 1C. The same number was achieved when vehiclesolution was injected (73±12 RGC/mm2). With 600


Fig. 4. Dependence of retinal ganglion cell survival and appearance of labelled microglial cells on the concentration of intravitreally administeredATA solution. Numbers of retinal ganglion cells (RGC) are presented as dark grey columns, those of microglial cells (MG) as light grey columns.Error bars indicate standard deviation. On top, significance of differences between the single concentrations of ATA are given (n.s. — not significant).Asterisks were used for the indication of statistical significance as follows: * P<0.05, ** P<0.01, *** P<0.001. Crosses indicate significant differencebetween injection of vehicle and no injection.

Fig. 5. Survival of retinal ganglion cells and appearance of labelled microglial cells when saturated ATA solution was administered at differenttime points after axotomy. Error bars indicate standard deviation. Asterisks indicate significant differences compared to vehicle injection, andcrosses indicate significant differences compared to injection of saturated ATA solution at the time of axotomy.

µg/ml or saturated ATA solution, a significantly highernumber of RGC regenerated their axons (135±34 and192±39 RGC/mm2, respectively, Fig. 8). As in the casewhere only survival rate was checked, an additional

enhancement of the number of regenerating RGC wasfound when ATA injection was repeated after 1 week(250±45 RGC/mm2). If a saturated solution of ATA wasinjected 1 day after surgery, the number of regenerating


Fig. 6. Visualisation of cells with cleaved DNA within the rat retina using the TUNEL technique. A: Positive control created by treatment of anormal retina with DNAse (30 min at 37°C, 1 mg/ml). Cells with cleaved DNA within the ganglion cell layer (g.c.l.) and inner nuclear layer (i.n.l.)are labelled clearly. There are also some labelled cells in the inner plexiform layer (i.p.l.). B–C: Retina 4 days after axotomy with injection ofeither vehicle solution (B) or saturated ATA solution (C). Scale bar: 50µm.

Fig. 7. Number of apoptotic cells per mm length of retina found inthe ganglion cell layer by TUNEL labelling. Error bars indicate stan-dard error of the mean. Asterisks indicate significant differences tovehicle-injected animals of the same time duration after axotomy.

axons was similar to that found after immediate injection(197±46 RGC/mm2). Fewer axons were found to regen-erate if injection was performed 3 days after surgery(127±9 RGC/mm2, Fig. 8).

3.4. Long-term effects of ATA

In the next series of experiments, we studied whetherlong-lasting rescue effects can be achieved with ATA.For this, the survival of RGC was also monitored at 28days after surgery (Fig. 9A). In control animals, 5% ofRGC were still alive in the different control groups irres-

pective of whether vehicle solution was injected or not.After an injection of saturated ATA solution, the numberof surviving RGC was less than the number observedafter 14 days. However, a survival rate of 12±3% wasseen, which is still significantly higher than that of non-injected or vehicle-injected animals (P<0.02).

We also checked the number of regenerating RGCfound after 28 days (Figs 9B and 10). As injection ofvehicle did not have an effect on the observed regener-ation rate after 14 days, we did not perform this experi-ment here. If no injection was performed after transplan-tation, 89±20 RGC/mm2 were found after 28 days,whereas 200±10 RGC/mm2 were counted if saturatedsolution of ATA was injected. Again, there is a signifi-cant increase after injection of ATA, and these numbersare similar to those found after 14 days of regeneration.

4. Discussion

In this study, the optic nerve was axotomised toinduce apoptosis of adult retinal ganglion cells. Ourresults demonstrate that intravitreal injection of ATA cansignificantly enhance the survival of RGC after axotomy.Interestingly, injection of vehicle solution alone alsoleads to a slight, but not significant enhancement of RGCsurvival. This corroborates earlier findings attributing apotential neuroprotective role to DMSO due to its radicalscavenging properties (Shimizu et al., 1997; Phillis et al.,1998) or its influence on membranes and/or ion channels(Hulsmann et al., 1999).

Retinal ganglion cell survival increases with higherconcentrations of ATA and can therefore be attributed toits administration. The survival rate could be increasedfurther when the ATA injection was repeated after 1


Fig. 8. Number of retinal ganglion cells regenerating their axons into the transplanted peripheral nerve graft. Error bars indicate standard deviation.Significance of differences between the single concentrations of ATA are given on top of the figure (n.s. — not significant).

Fig. 9. Effects of ATA 28 days after surgery. A: Retinal ganglion cell survival and appearance of labelled microglial cells 28 days after axotomy.B: Regenerating RGC, which could be stained retrogradely 28 days after transplantation without and with injection of a solution of ATA. Errorbars indicate standard deviation. Asterisks indicate significant differences between vehicle and ATA solution, and crosses differences between thetwo transplantation groups.

week. This indicates that the protective effect of ATAis a transient one, and an additional population of RGCwhich otherwise would undergo delayed degenerationcould be rescued by a second injection of ATA. Thetransient nature of neuroprotection is also demonstratedby the numbers of surviving RGC found 28 days afteraxotomy, when partial neuroprotection can still beobserved, but the number of surviving RGC is decreasedcompared to the values obtained after 14 days. Clarke

et al. (1998) administered NT-4/5 intravitreally aftertransection of the optic nerve and also reported that onlya temporary effect could be achieved after this treatment,despite repeated neurotrophin injection. The slight neur-oprotective effect of DMSO discussed above vanishescompletely after 28 days.

An interesting finding is that injection of ATA showeda better effect if it was not injected immediately afteraxotomy but 1 day later. This indicates that ATA takes


Fig. 10. Sectors of retinal whole-mounts with RGC stained retrogradely with 4Di–10ASP. Staining was performed 28 days after transplantation.A: No injection was performed. B: Saturated solution of ATA was injected at the time of surgery. Scale bars: 200µm.

effect approximately 24 h after injury, when the cascadesof cell death are being initiated. This fact may haveimplications for potential medical applications sinceapplication of proper drugs is usually not possibleimmediately after traumatic injury. If ATA is adminis-tered too late, i.e. 4 days after axotomy, the death cas-cade can no longer be influenced. A similar result,though observed in vitro with cortical neurons, has beenreported by Csernansky et al. (1994) who found that

application of ATA 2 h after glutamate exposure had abetter effect than immediate ATA application. Rosen-baum et al. (1997) found that the best effect of ATAafter an ischemic insult was found when it was applied6 h after the lesion. Unfortunately, neither author com-mented on the results of ATA administration at othertime points after the insult.

In the present experiments, microglial cells are alsostained because they ingest fluorescent dye while phago-


cytosing RGC. The number of stained microglial cellsdecreases with increasing survival rate of RGC and viceversa. This finding is in agreement with the well-knownrole of microglial cells as scavengers of dying neurons.

The TUNEL technique is often considered to be appli-cable for the detection of cells undergoing apoptoticdeath. However, it is known that necrotic cells may alsobe stained because there is a DNA fragmentation in thesecells too (Gold et al., 1994). Nevertheless, we appliedtransferase in our assay, which is reported to be morespecific for apoptotic cells (Gold et al., 1994). Moreover,it is well established that the death of RGC after anaxotomy is mainly an apoptotic one (Berkelaar et al.,1994; Garcia-Valenzuela et al., 1994; Quigley et al.,1995). All in all it cannot be completely excluded thatwe also stained some necrotic RGC with the TUNELtechnique.

The results of TUNEL staining clearly show that thenumber of apoptotic RGC 4 days after axotomy isreduced by the application of ATA. This effect could beattributed to the inhibition of endonucleases, althoughother mechanisms may not be excluded. The number ofapoptotic RGC is higher 8 days after axotomy. This alsoindicates the transient nature of the effect of ATA beingable to delay RGC death to a considerable extent. MoreRGC are still present after 8 days if ATA was applied,and with the diminishing intensity of ATA action moreRGC can start apoptosis. Koeberle and Ball (1998)reported a decreased ratio of apoptotic RGC after appli-cation of GDNF, with the maximum being shifted fromday 7 to day 10 after axotomy. More detailed studiesshould be performed in order to establish the time courseof RGC apoptosis after the application of ATA.

The most important result of this study is that ATAsignificantly increases the number of surviving RGC,which regenerate their axons into the peripheral nervegraft. This demonstrates that treatment of RGC withATA not only allows them to survive, but that survivingRGC retain their capacity for active regeneration in vivosuch as extension of axons into the peripheral nervegraft. In contrast to the survival study, administration ofATA solution 1 day after the surgery did not have anoticeable effect on regeneration compared to concomi-tant administration of ATA. Achieving enhanced regen-eration in the CNS of an adult organism by a single phar-macological intervention is of great importance forfurther efforts of restoration of injured nervous path-ways.

Comparison of data obtained 14 or 28 days after sur-gery clearly demonstrates that axotomised RGC arestabilised by giving them the possibility to regeneratetheir axons. After protection with ATA, more RGC canregenerate, and this enhanced number is also preservedafter 28 days. This is in contrast to the survival studywith ATA, where much fewer RGC were found 28 daysafter axotomy.

Trying to protect RGC from axotomy-induced deathhas been a goal of enormous importance for many years.In order to achieve this goal, many in vivo and in vitrostudies have been performed, which aimed to interferewith the cascades of cell death at various levels. Theseinclude localised injection of antioxidants (Castagne andClarke, 1996), scavenging free radicals (Klo¨cker et al.,1998), adding neurotrophic factors (Sievers et al., 1987;Carmignoto et al., 1989; Mey and Thanos, 1993; Rabac-chi et al., 1994a; Mansour-Robaey et al., 1994; Heid-inger et al., 1997; Clarke et al., 1998; Klo¨cker et al.,1998; Koeberle and Ball, 1998; Cuevas et al., 1998).Inhibiting NO synthase, (Klo¨cker et al., 1998), proteinbiosynthesis (Garcia-Valenzuela et al., 1994; Rabacchiet al., 1994b; Castagne and Clarke, 1996; Fagiolini etal., 1997), caspases (Chen et al., 1997; Lucius and Siev-ers, 1997; Kermer et al., 1998), kinases (Maas et al.,1998), and monoamine oxidase (Buys et al., 1995) alsoproved effective, as was blocking of receptors by apply-ing the corresponding antagonists (Lam et al., 1997;Yoles et al., 1997; Chaudhary et al., 1998; Lagre`ze etal., 1998; Kapin et al., 1999) or agonists (Senda et al.,1998; Yoles et al., 1999). Over-expression of the proto-oncogenebcl-2 (Bonfanti et al., 1996; Chen et al., 1997)has also received extensive attention.

Among these studies, only a few addressed the prob-lem of regeneration of axons. One example where neuro-protective measures directed at the RGC lead to a betterregeneration is the proto-oncogenebcl-2 which not onlyrescued RGC from apoptosis, but also enhanced regener-ation of axons in transgenic mice over-expressingbcl-2(Chen et al., 1997). The effect of promoting the regener-ation was found in vitro and also in vivo, if post-natalmice were used. Enhanced regeneration did not seem tobe an indirect consequence of the well-known anti-apop-totic activity of bcl-2. Other RGC-directed neuroprotec-tive agents did not lead to an improved regeneration.Mansour-Robaey et al. (1994) achieved a better survivalafter injection of brain-derived neurotrophic factor(BDNF), but this was not reflected in a significantlyenhanced number of regenerating RGC. The caspaseinhibiting peptide ZVAD, which supported neuronal sur-vival in vitro, did not promote axon regeneration (Chenet al., 1997), whereas the peptide YVAD seemed to havethe additional effect of a better axonal regrowth in vitro(Lucius and Sievers, 1997). Kinase inhibitors, whichprevented apoptosis, however, inhibited neurite out-growth (Maas et al., 1998).

The application of ATA in various biological studiesbegan in the early 1970s, and interest in ATA increaseddramatically when it was discovered that ATA couldinhibit the initiation of protein synthesis (Stewart et al.,1971; Siegelman and Apirion, 1971). Later, ATA wasfound to prevent DNA fragmentation by inhibitingendonucleases, which was utilised to rescue immaturethymocytes exposed to glucocorticoid methylpredniso-


lone (McConkey et al., 1989). Subsequently, it wasestablished that ATA could prevent cell death in manysystems, and it was used as a criterion to determine ifcell death, in the system under study, was due toapoptosis. There are many reports where ATA was usedas an inhibitor of apoptosis in vitro and in vivo. Recentexamples of ATA’s ability to attenuate the severity ofapoptosis in different models are listed in Table 1.

ATA acts primarily by inhibiting endonucleases, butis an agent, which has a different mechanism of actionin different targets. ATA inhibits endonucleases, DNApolymerases, initiation of transcription, formation of theribosomal complex, nuclease binding to DNA, RNA andcalpain. It has also been reported that ATA can protectneurons from glutamate excitotoxicity, e.g. by blockingNMDA and non-NMDA glutamate receptors (Zeevalk etal., 1995) and down-regulating glutamate receptor sub-units at the level of transcription (Aronica et al., 1998).ATA is known to decrease neuronalβ-amyloid precursorprotein gene expression (Joseph et al., 1993) andincrease phosphorylation of tyrosine in proteins involvedin signal transduction pathways (Okada and Koizumi,1995, 1997). These examples clearly demonstrate thatATA does not mediate its anti-apoptotic effects by a sin-gle mechanism, but that different modes of action areinvolved in neuronal survival.

Table 1Recent examples of inhibition of apoptosis by the application of ATA

Cell type Apoptosis induced by

In vitro:Chick retina tissue NMDA, kainate (Zeevalk et al. 1993, 1995)PC12 cells NGF deprivation (Batistatou and Greene, 1991)

EGTA (Lindenboim et al., 1995)6-hydroxydopamine (Walkinshaw and Waters, 1994)peroxynitrite (Estevez et al., 1995)

Sympathetic neurons NGF deprivation (Batistatou and Greene, 1991)hypoxia (Rosenbaum et al., 1994)

Murine cortical neurons glutamate (Regan et al., 1995)Cerebellar granule cells carbamazepine (Gao et al., 1995)

kainate (Simonian et al., 1996)cytosine arabinonucleoside (Ishitani and Chuang, 1996)tunicamycin (Chang et al., 1997)cholesterol oxides (Chang and Liu, 1998)

Neurons from rat cerebral cortex age (Sunaga et al., 1995)NMDA (Posner et al., 1995)glutamate (Mandavilli and Rao, 1996)

Rat cortical neurons camptothecin (Morris and Geller, 1996)age (Ishitani et al., 1996)

Mice neocortical neurons glutamate and NMDA (Csernansky et al., 1994)Human-derived neurotypic cell line SH-S454 β-amyloid peptide (β-A4) (Li et al., 1996)Rat hippocampal neurons macrophage-conditioned medium (Flavin et al., 1997)

nitric oxide (Vincent and Maiese, 1999)In vivo:Rat retinal neurons ischemia/reperfusion (Lam et al., 1995)

ischemia (Rosenbaum et al., 1997)Mouse olfactory receptor neurons bulbectomy (Holcomb et al., 1995)Sympathetic neurons in gerbils ischemia (Rosenbaum et al., 1998)Rat hippocampal neurons NMDA and ischemia (Roberts-Lewis et al., 1993)

In this study, we administered ATA by intravitrealinjection. Intravitreal surgery is a standard ophthalmo-logical technique. However, it requires experience andcare in order to avoid infections, and the concentrationand volume of the drugs injected must be accurate toprevent damage to the retina and lens. These factorsshould be considered for potential therapeutic appli-cations, especially considering the discomfort associatedwith multiple injections. Several studies have involvedrepeated intravitreal injections with different percentagesof surviving RGC. Mansour-Robaey et al. (1994)injected BDNF solution four times and achieved 67%survival. However, with four injections of vehicle sol-ution they obtained 51% of surviving RGC compared to25% with a single injection or 12% without an injection,which could be a sign of the influence of lens damaging.Clarke et al. (1998) injected a solution of NT-4/5 twotimes and could rescue 36% of RGC, whereas Kermeret al. (1998) achieved 45% after four injections of a sol-ution of ZDEVD-cmk, an inhibitor of caspase-3. Klo¨ckeret al. (1998) could rescue 37% after three injections ofBDNF, and this number could be enhanced up to 73%by intraperitoneal injections of the anti-oxidantN-tert-butyl-(2-sulfophenyl)-nitrone twice a day. We confinedthe number of injections to one or two only. Interest-ingly, the survival rate of RGC after single adminis-


tration of ATA 1 day after axotomy (58%) is similar tothat if ATA was administered twice (59%). This impliesthat multiple injections of ATA can be spared if the timepoint of the single injection is optimised.

Acknowledgements

This work was supported by the University of Mu¨nsterSchool of Medicine, grant IMF HE 129826 to P.H., Bun-desministerium fu¨r Bildung und Forschung (BMBF),grant no. 01 Ko 9805/7 and the Deutsche Forschungsge-meinschaft, (grant No Th 3868-1 and 10-1 to S.T.). MrsIlka Romann is acknowledged for her help with theTUNEL staining and Dr. Rita Naskar for linguistic help.

References

Aronica, E.M., Gorter, J.A., Grooms, S., Kessler, J.A., Bennett,M.V.L., Zukin, R.S., Rosenbaum, D.M., 1998. Aurintricarboxylicacid prevents GluR2 mRNA downregulation and delayed neurod-egeneration in hippocampal CA1 neurons of gerbil after globalischemia. Proceedings of the National Academy of Science USA95, 7115–7120.

Batistatou, A., Greene, L.A., 1991. Aurintricarboxylic acid rescuesPC12 cells and sympathetic neurons from cell death caused bynerve growth factor deprivation: correlation with suppression ofendonuclease activity. Journal of Cellular Biology 115, 461–471.

Berkelaar, M., Clarke, D.B., Wang, Y.C., Bray, G.M., Aguayo, A.J.,1994. Axotomy results in delayed death and apoptosis of retinalganglion cells in adult rats. Journal of Neuroscience 14, 4368–4374.

Bonfanti, L., Strettoi, E., Chierzi, S., Cenni, M.C., Liu, X.H., Martinou,J.-C., Maffei, L., Rabacchi, S.A., 1996. Protection of retinal gang-lion cells from natural and axotomy-induced cell death in neonataltransgenic mice overexpressing bcl-2. Journal of Neuroscience 16,4186–4194.

Buys, Y.M., Trope, G.E., Tatton, W.G., 1995. (-)-Deprenyl increasesthe survival of rat retinal ganglion cells after optic nerve crush.Current Eye Research 14, 119–126.

Carmignoto, G., Maffei, L., Candeo, P., Canella, R., Comelli, C., 1989.Effect of NGF on the survival of rat retinal ganglion cells followingoptic nerve section. Journal of Neuroscience 9, 1263–1272.

Castagne, V., Clarke, P.G., 1996. Axotomy-induced retinal ganglioncell death in development: its time-course and its diminution byantioxidants. Proceedings of the Royal Society London B Biologi-cal Sciences 263, 1193–1197.

Chang, J.Y., Korolev, V.V., Wang, J.Z., 1997. Neurotoxicity oftunicamycin on cultured cerebellar granule cells. Neurotoxicology18, 129–135.

Chang, J.Y., Liu, L.Z., 1998. Neurotoxicity of cholesterol oxides oncultured cerebellar granule cells. Neurochemistry International 32,317–323.

Chaudhary, P., Ahmed, F., Sharma, S.C., 1998. MK801 – a neuropro-tectant in rat hypertensive eyes. Brain Research 792, 154–158.

Chen, D.F., Schneider, G.E., Martinou, J.C., Tonegawa, S., 1997. Bcl-2 promotes regeneration of severed axons in mammalian CNS. Nat-ure 385, 434–439.

Clarke, D.B., Bray, G.M., Aguayo, A.J., 1998. Prolonged adminis-tration of NT-4/5 fails to rescue most axotomized retinal ganglioncells in adult rats. Vision Research 38, 1517–1524.

Csernansky, C.A., Canzoniero, L.M., Sensi, S.L., Yu, S.P., Choi,

D.W., 1994. Delayed application of aurintricarboxylic acid reducesglutamate-induced cortical neuronal injury. Journal of Neurosci-ence Research 38, 101–108.

Cuevas, P., Carceller, F., Redondo-Horcajo, M., Lozano, R.M.,Gimenez-Gallego, G., 1998. Systemic administration of acidicfibroblast growth factor ameliorates the ischemic injury of the ret-ina in rats. Neuroscience Letters 255, 1–4.

Estevez, A.G., Radi, R., Barbeito, L., Shin, J.T., Thompson, J.A.,Beckman, J.S., 1995. Peroxynitrite-induced cytotoxicity in PC12cells: Evidence for an apoptotic mechanism differentially modu-lated by neurotrophic factors. Journal of Neurochemistry 65,1543–1550.

Fagiolini, M., Caleo, M., Strettoi, E., Maffei, L., 1997. Axonal trans-port blockade in the neonatal rat optic nerve induces limited retinalganglion cell death. Journal of Neuroscience 17, 7045–7052.

Flavin, M.P., Coughlin, K., Ho, L.T., 1997. Soluble macrophage fac-tors trigger apoptosis in cultured hippocampal neurons. Neurosci-ence 80, 437–448.

Gao, X.M., Margolis, R.L., Leeds, P., Hough, C., Post, R.M., Chuang,D.M., 1995. Carbamazepine induction of apoptosis in cultured cer-ebellar neurons: Effects ofN-methyl-d-aspartate, aurintricarbocylicacid and cycloheximide. Brain Research 703, 63–71.

Garcia-Valenzuela, E., Gorczyca, W., Darzynkiewicz, Z., Sharma,S.C., 1994. Apoptosis in adult retinal ganglion cells after axotomy.Journal of Neurobiology 25, 431–438.

Gold, R., Schmied, M., Giegerich, G., Breitschopf, H., Hartung, H.P.,Toyka, K.V., Lassmann, H., 1994. Differentiation between cellularapoptosis and necrosis by the combined use of in situ tailing andnick translation techniques. Laboratory Investigations 71, 219–225.

Heidinger, V., Hicks, D., Sahel, J., Dreyfus, H., 1997. Peptide growthfactors but not ganglioside protect against excitotoxicity in rat reti-nal neurons in vitro. Brain Research 767, 279–288.

Holcomb, J.D., Mumm, J.S., Calof, A.L., 1995. Apoptosis in the neu-ronal lineage of the mouse olfactory epithelium: Regulation in vivoand in vitro. Developmental Biology 172, 307–323.

Hulsmann, S., Greiner, C., Ko¨hling, R., Wolfer, J., Moskopp, D., Rie-mann, B., Lucke, A., Wassmann, H., Speckmann, E.-J., 1999.Dimethyl sulfoxide increases latency of anoxic terminal negativityin hippocampal slices of guinea pig in vitro. Neuroscience Letters261, 1–4.

Ishitani, R., Chuang, D.M., 1996. Glyceraldehyde-3-phosphatedehydrogenase antisense oligodeoxynucleotides protect againstcytosine arabinonucleoside-induced apoptosis in cultured cerebellarneurons. Proceedings of the National Academy of Science USA93, 9937–9941.

Ishitani, R., Kimura, M., Sunaga, K., Katsube, N., Tanaka, M., Chu-ang, D.M., 1996. An antisense oligodeoxynucleotide to glyceral-dehyde-3-phosphate dehydrogenase blocks age-induced apoptosisof mature cerebrocortical neurons in culture. Journal of Pharma-cology Experimental Therapy 278, 447–454.

Joseph, R., Tsang, W., Han, E., Saed, G.M., 1993. Neuronalβ-amyloidprecursor protein gene expression: regulation by aurintricarboxylicacid. Brain Research 625, 244–255.

Kapin, M.A., Doshi, R., Scatton, B., De Santis, L., Chandler, M.L.,1999. Neuroprotective effects of eliprodil in retinal excitotoxicityand ischemia. Investigative Ophthalmology Visual Science 40,1177–1182.

Kermer, P., Klocker, N., Labes, M., Ba¨hr, M., 1998. Inhibition ofCPP32-like proteases rescues axotomized retinal ganglion cellsfrom secondary cell death in vivo. Journal of Neuroscience 18,4656–4662.

Klocker, N., Cellerino, A., Ba¨hr, M., 1998. Free radical scavengingand inhibition of nitric oxide synthase potentiates the neurotrophiceffects of brain-derived neurotrophic factor on axotomized retinalganglion cells in vivo. Journal of Neuroscience 18, 1038–1046.

Koeberle, P.D., Ball, A.K., 1998. Effects of GDNF on retinal ganglioncell survival following axotomy. Vision Research 38, 1505–1515.


Lagreze, W.A., Knorle, R., Bach, M., Feuerstein, T.J., 1998. Meman-tine is neuroprotective in a rat model of pressure-induced retinalischemia. Investigative Ophthalmology Visual Science 39, 1063–1066.

Lam, T.T., Fu, J., Hrynewycz, M., Tso, M.O.M., 1995. The effect ofaurintricarboxylic acid, an endonuclease inhibitor, onischemia/reperfusion damage in rat retina. Journal of Ocular Phar-macology Therapy 11, 253–259.

Lam, T.T., Siew, E., Chu, R., Tso, M.O., 1997. Ameliorative effectof MK-801 on retinal ischemia. Journal of Ocular PharmacologyTherapy 13, 129–137.

Li, Y.P., Bushnell, A.F., Lee, C.M., Perlmutter, L.S., Wong, S.K.F.,1996. β-Amyloid induces apoptosis in human-derived neurotypicSH-SY5Y cells. Brain Research 738, 196–204.

Lindenboim, L., Haviv, R., Stein, R., 1995. Inhibition of drug-inducedapoptosis by survival factors in PC12 cells. Journal of Neurochemi-stry 64, 1054–1063.

Lucius, R., Sievers, J., 1997. YVAD protect post-natal retinal ganglioncells against axotomy-induced but not free radical-induced axonaldegeneration in vitro. Brain Research Molecular Brain Research48, 181–184.

Maas, J.W. Jr., Horstmann, S., Borasio, G.D., Anneser, J.M., Shooter,E.M., Kahle, P.J., 1998. Apoptosis of central and peripheral neu-rons can be prevented with cyclin-dependent kinase/mitogen-acti-vated protein kinase inhibitors. Journal of Neurochemistry 70,1401–1410.

Mandavilli, B.S., Rao, K.S., 1996. Accumulation of DNA damage inaging neurons occurs through a mechanism other than apoptosis.Journal of Neurochemistry 67, 1559–1565.

Mansour-Robaey, S., Clarke, D.B., Wang, Y.-C., Bray, G.M., Aguayo,A.J., 1994. Effects of ocular injury and administration of brain-derived neurotrophic factor on survival and regrowth of axotomizedretinal ganglion cells. Proceedings of the National Academy ofScience USA 91, 1632–1636.

McConkey, D.J., Hartzell, P., Nicotera, P., Orrenius, S., 1989. Cal-cium-activated DNA fragmentation kills immature thymocytes.FASEB Journal 3, 1843–1849.

Morris, E.J., Geller, H.M., 1996. Induction of neuronal apoptosis bycamptothecin, an inhibitor of DNA topoisomerase – I: Evidencefor cell cycle-independent toxicity. Journal of Cellular Biology 134,757–770.

Mey, J., Thanos, S., 1993. Intravitreal injections of neurotrophic fac-tors support the survival of axotomized retinal ganglion cells inadult rats in vivo. Brain Research 602, 304–317.

Okada, N., Koizumi, S., 1995. A neuroprotective compound, aurintricarboxylic acid, stimulates the tyrosine phosphorylation cascadein PC12 cells. Journal of Biological Chemistry 270, 16464–16469.

Okada, N., Koizumi, S., 1997. Tyrosine phosphorylation of ErbB4 isstimulated by aurintricarboxylic acid in human neuroblastoma SH-SY5Y cells. Biochemical Biophysical Research Communications230, 266–269.

Phillis, J.W., Estevez, A.Y., O’Regan, M.H., 1998. Protective effectsof the free radical scavengers, dimethyl sulfoxide and ethanol, incerebral ischemia in gerbils. Neuroscience Letters 244, 109–111.

Posner, A., Raser, K.J., Hajimohammadreza, I., Yuen, P.W., Wang,K.K.W., 1995. Aurintricarboxylic acid is an inhibitor ofµ- and m-calpain. Biochemistry Molecular Biology International 36, 291–299.

Quigley, H.A., Nickells, R.W., Kerrigan, L.A., Pease, M.E., Thibault,D.J., Zack, D.J., 1995. Retinal ganglion cell death in experimentalglaucoma and after axotomy occurs by apoptosis. InvestigativeOphthalmology Visual Science 36, 774–786.

Rabacchi, S.A., Ensini, M., Bonfanti, L., Gravina, A., Maffei, L.,1994a. Nerve growth factor reduces apoptosis of axotomized retinalganglion cells in the neonatal rat. Neuroscience 63, 969–973.

Rabacchi, S.A., Bonfanti, L., Liu, X.H., Maffei, L., 1994b. Apoptotic

cell death induced by optic nerve lesion in the neonatal rat. Journalof Neuroscience 14, 5292–5301.

Regan, R.F., Panter, S.S., Witz, A., Tilly, J.L., Giffard, R.G., 1995.Ultrastructure of excitotoxic neuronal death in murine cortical cul-ture. Brain Research 705, 188–198.

Roberts-Lewis, J.M., Marcy, V.R., Zhao, Y., Vaught, J.L., Siman, R.,Lewis, M.E., 1993. Aurintricarboxylic acid protects hippocampalneurons from NMDA- and ischemia-induced toxicity in vivo. Jour-nal of Neurochemistry 61, 378–381.

Rosenbaum, D.M., Michaelson, M., Batter, D.K., Doshi, P., Kessler,J.A., 1994. Evidence for hypoxia-induced, programmed cell deathof cultured neurons. Annals of Neurology 36, 864–870.

Rosenbaum, D.M., Rosenbaum, P.S., Gupta, A., Michaelson, M.D.,Hall, D.H., Kessler, J.A., 1997. Retinal ischemia leads to apoptosiswhich is ameliorated by aurintricarboxylic acid. Vision Research37, 3445–3451.

Rosenbaum, D.M., D’Amore, J., Llena, J., Rybak, S., Balkany, A.,Kessler, J.A., 1998. Pretreatment with intraventricular aurintricar-boxylic acid decreases infarct size by inhibiting apoptosis followingtransient global ischemia in gerbils. Annals of Neurology 43,654–660.

Senda, T., Mita, S., Kaneda, K., Kikuchi, M., Akaike, A., 1998. Effectof SA4503, a novelσ1 receptor agonist, against glutamate neuro-toxicity in cultured rat retinal neurons. European Journal of Phar-macology 342, 105–111.

Shimizu, S., Simon, R.P., Graham, S.H., 1997. Dimethylsulfoxide(DMSO) treatment reduces infarction volume after permanent focalcerebral ischemia in rats. Neuroscience Letters 239, 125–127.

Siegelman, F., Apirion, D., 1971. Aurintricarboxylic acid, a preferen-tial inhibitor of initiation of protein synthesis. Journal of Bacteri-ology 105, 902–907.

Sievers, J., Hausmann, B., Unsicker, K., Berry, M., 1987. Fibroblastgrowth factors promote the survival of adult rat retinal ganglioncells after transection of the optic nerve. Neuroscience Letters 76,157–162.

Simonian, N.A., Getz, R.L., Leveque, J.C., Konradi, C., Coyle, J.T.,1996. Kainic acid induces apoptosis in neurons. Neuroscience 75,1047–1055.

Stewart, M.L., Grollman, A.P., Huang, M.T., 1971. Aurintricarboxylicacid: inhibitor of initiation of protein synthesis. Proceedings of theNational Academy of Science USA 68, 97–101.

Sunaga, K., Takahashi, H., Chuang, D.M., Ishitani, R., 1995. Glyceral-dehyde-3-phosphate dehydrogenase is over-expressed during apop-totic death of neuronal cultures and is recognized by a monoclonalantibody against amyloid plaques from Alzheimer’s brain. Neuros-cience Letters 200, 133–136.

Thanos, S., Naskar, R., Heiduschka, P., 1997. Regenerating ganglioncell axons in the adult rat establish retinofugal topography andrestore visual function. Experimental Brain Research 114, 483–491.

Thanos, S., Seeger, J., Kacza, J., Mey, J., 1994. Old dyes for newscopes: The phagocytosis-dependent long-term fluorescence label-ling of microglial cells in vivo. Trends in Neurosciences 17,177–181.

Vidal-Sanz, M., Bray, G.M., Villegas-Perez, M.P., Thanos, S.,Aguayo, A.J., 1987. Axonal regeneration and synapse formation inthe superior colliculus by retinal ganglion cells in the adult rat.Journal of Neuroscience 7, 2894–2909.

Vincent, A.M., Maiese, K., 1999. Nitric oxide induction of neuronalendonuclease activity in programmed cell death. Experimental CellResearch 246, 290–300.

Walkinshaw, G., Waters, C.M., 1994. Neurotoxin-induced cell deathin neuronal PC12 cells is mediated by induction of apoptosis. Neu-roscience 63, 975–987.

Yoles, E., Muller, S., Schwartz, M., 1997. NMDA-receptor antagonistprotects neurons from secondary degeneration after partial opticnerve crush. Journal of Neurotrauma 14, 665–675.

Yoles, E., Wheeler, L.A., Schwartz, M., 1999.α2-adrenoreceptor


agonists are neuroprotective in a rat model of optic nerve degener-ation. Investigative Ophthalmology Visual Science 40, 65–73.

Zeevalk, G.D., Schoepp, D., Nicklas, W.J., 1993. Aurintricarboxylicacid prevents NMDA-mediated excitotoxicity: evidence for itsaction as an NMDA receptor antagonist. Journal of Neurochemistry61, 386–389.

Zeevalk, G.D., Schoepp, D., Nicklas, W.J., 1995. Excitotoxicity at bothNMDA and non-NMDA glutamate receptors is antagonized by aur-intricarboxylic acid: Evidence for differing mechanisms of action.Journal of Neurochemistry 64, 1749–1758.

Aurintricarboxylic acid promotes survival and regeneration of axotomised retinal ganglion cells in vivo

Documents