Abstract Metallothionein (MT)-I/II has been shown to beneuroprotective and neuroregenerative in a model of ratcortical brain injury. Here we examine expression patternsof MT-I/II and its putative receptor megalin in rat retina. Atneonatal stages, MT-I/II was present in retinal ganglioncells (RGCs) but not glial or amacrine cells; megalin waspresent throughout the retina. Whilst MT-I/II was absentfrom adult RGC in normal animals and after optic nervetransection, the constitutive megalin expression in RGCswas lost following optic nerve transection. In vitro MT-IIAtreatment stimulated neuritic growth: more RGCs grewneurites longer than 25 �m (P < 0.05) in dissociated retinalcultures and neurite extension increased in retinal explants(P < 0.05). MT-IIA treatment of mixed retinal culturesincreased megalin expression in RGCs, and pre-treatingcells with anti-megalin antibodies prevented MT-IIA-stim-ulated neurite extension. Our results indicate that MT-IIAstimulates neurite outgrowth in RGCs and may do so viathe megalin receptor; we propose that neurite extension istriggered via signal transduction pathways activated by theNPxY motifs of megalin’s cytoplasmic tail.
Strategies to overcome neurotrauma have been vigorouslypursued over several decades but with hitherto disappoint-ing results. A number of recent studies aimed at promotingneuroprotection and neuroregeneration have focused onMetallothioneins (MT), small, cysteine-rich metal bindingproteins. MT-I and -II, usually considered as MT-I/II due totheir structural and functional similarities (Giralt et al.2002), have been demonstrated to be neuroprotective in arange of model systems including focal brain injury (Giraltet al. 2002), focal cerebral ischemia (Trendelenburg et al.2002) and experimental autoimmune encephalomyelitis(Penkowa and Hidalgo 2001). Protective eVects have alsobeen demonstrated for MT-II against retinal neuron damage(Suemori et al. 2006). MT-IIA (the human isoform) hasbeen shown to promote neurite elongation and an increasein reactive axon growth into injured adult rat cortex (Chunget al. 2003).
The response of neurons to injury is thought to involvelocal astrocytic release of MT-I/II (Chung et al. 2004;Chung and West 2004). The identity of the receptor allow-ing MT-I/II uptake has until recently been uncertain. How-ever, a candidate is the multi-ligand endocytotic receptor,megalin, known to mediate in part the uptake of heavymetal MT-I/II complexes in renal cells (Klassen et al. 2004;WolV et al. 2006). Here we examine the neuroregenerativeproperties of MT-I/II and assess associated changes in meg-alin expression using a favoured model of central nerveregeneration, the mammalian visual system. We used twostandard in vitro models. One was dissociated neonatal
M. Fitzgerald (&) · P. Nairn · C. A. Bartlett · L. D. BeazleyExperimental and Regenerative Neurosciences, School of Animal Biology, University of Western Australia, Hackett Drive, Crawley 6009, WA, Australiae-mail: [email protected]
L. D. BeazleyWestern Australian Institute of Medical Research, University of Western Australia, Hackett Drive, Crawley 6009, WA, Australia
R. S. Chung · A. K. WestNeurorepair Group, Menzies Research Institute, University of Tasmania, Private Bag 58, Hobart, TAS 7001, Australia
123
172 Exp Brain Res (2007) 183:171–180
retinal cultures (Geiger et al. 2002) in which retinal gan-glion cells (RGCs) were identiWed by retrograde labellingfrom the brain. The other was retinal explants taken froman essentially mature system, that of 3 week postnatal ani-mals (Avwenagha et al. 2003); the model is considered tomore closely approximate the in vivo cellular context. Thestudies extend the Wnding that MT-I/II mRNA is up-regulatedfollowing axotomy of retinal ganglion cells (RGCs) in vivo(Fischer et al. 2004).
Materials and methods
Animals and surgical procedures
Animals and anaesthesia
PVG rats (supplied by Animal Resources Centre, WA Aus-tralia), were housed under standard animal house condi-tions and had access ‘ad libitum’ to water and pelletedfood. For surgery, pups (male and female) at postnatal day(P) 2 were anaesthetized by subcutaneous injection of6 mg/kg ilium xylazil (Troy Laboratories Pty. Ltd)combined with 50 mg/kg ketamine (Parnell Laboratories,Australia Pty. Ltd). Adult female rats (180 g) were anaes-thetized by intraperitoneal injection of ilium xylazil andketamine as above. Terminal anaesthesia was with pento-barbitone sodium (600 mg/kg bw ip). Studies wereapproved by the appropriate institutional ethics committee.
DiI labeling
Skin overlying the skull of pups was opened along the mid-line. The tip of a scalpel blade was used to cut a Xap in theskull to expose the superior colliculi. A gelfoam pledgetwetted with 2 �l 10% DiI (1,1�-dioctadecyl-3,3,3�,3�-tetramethylindocarbocyanine perchlorate, MolecularProbes, D-282, Eugene, OR, USA) in methanol was appliedto the surface. Two days were allowed for transport to RGCsomas (Dunlop 2003).
Optic nerve transection
The right optic nerve of adult rats was transectioned usingstandard procedures (Rodger et al. 2001). BrieXy the nervewas approached dorsally after incising the skin and con-junctiva and deXecting lachrymal glands, extraocular mus-cles and adipose tissue. A transverse slit was made in thenerve sheath, the nerve parenchyma was then lifted andcompletely transected with iridectomy scissors. The nervesheath was closed with a 10–0 suture thread, lachrymalglands repositioned and the skin sutured (6–0). Rats weremaintained for 1, 2 or 4 weeks.
Cell culture
Dissociated retinal cultures
Retinae from postnatal day (P) 2–4 pups were dissectedinto Hanks Balanced Salt Solution (HBSS) at room tem-perature. Retinae were enzymatically digested using165U papain (Worthington) activated with cysteine(Sigma) at 0.24 mg/ml in HBSS at 37°C for 30 min. Reti-nae were gently agitated in solution every 10 min duringthe incubation. Following digestion tissue was dissociatedby gentle mechanical triturition in Neurobasal + 10%foetal calf serum (FCS) (growth medium) using a sterileplastic Pasteur pipette. Cell suspensions were Wlteredthrough sterile nylon gauze and gently re-triturated. Cellnumbers were adjusted to 2.5 £ 105cells/cm2 and platedonto glass coverslips in 24 well plates or 8 well chamberslides (Falcon) in 0.5 ml of growth medium. All tissueculture surfaces were pre-coated with 10 �g/ml poly-L-lysine in dH2O for 1 h at room temperature, followed by10 �g/ml laminin at 37°C overnight. Tissue culturereagents were from Gibco, Invitrogen unless otherwisespeciWed.
Retinal explants
Retinae from P21 pups were dissected as described aboveand placed RGC side up onto black nitrocellulose Wlterpaper (Millipore) moistened with HBSS. Explants were cutinto 1 mm squares using a McIlwain tissue chopper andplaced RGC side down onto glass coverslips coated asabove. Explants were moistened with growth media andallowed to adhere before addition of further growthmedium.
MT-IIA treatments
Apo-MT-IIA (Bestenbalt) was reconstituted with a 7.5-foldmolar excess of ZnSO4 and pH neutralised with phosphatebuVered saline (PBS). MT-IIA was added to dissociatedand explanted cultures at the time of plating for analysis ofeVects on neurite length. After 24 h incubation withMT-IIA, dissociated retinal cultures were Wxed in fresh4% paraformaldehyde (Merck) in 0.1 M PBS (pH 7.2);explants were similarly Wxed following 5 days in culture.Studies on MT-I/II uptake and megalin expression follow-ing treatment were carried out by treating dissociatedretinal cultures with MT-IIA 24 h after plating. All experi-ments included controls treated with equivalent concentra-tions of ZnSO4 in PBS or PBS alone. No diVerence wasobserved between cultures treated with ZnSO4 or withPBS alone (P · 0.05), and controls presented in this studyare ZnSO4 treated.
123
Exp Brain Res (2007) 183:171–180 173
Immunohistochemistry
Tissue preparation
Terminally anaesthetized adult rats were perfused transcar-dially with 0.9% NaCl followed by freshly prepared 4%paraformaldehyde in 0.1 M PBS (pH 7.2). Right eyes weredissected, the cornea and lens removed and post-Wxed in4% paraformaldehyde for 24 h. Retinae were dissectedfrom the eye-cup and prepared as wholemount; alterna-tively, the whole eye-cup was cryoprotected by immersionin 15% sucrose overnight before being serially cryosec-tioned (16 �m) horizontally, collected onto Superfrost Plusglass slides and stored at ¡80°C.
Antibody staining
RGCs were initially identiWed by retrograde labelling withDiI and were shown to be exclusively positive for �-IIItubulin using the Tuj-1 antibody. Thereafter Tuj-1 was usedto identify RGCs (Yin et al. 2003). Free Xoating Wxed reti-nae permeabilised with PBS containing 1% Triton X-100were incubated overnight in a humid chamber at 4°C withanti-�-III tubulin (Tuj-1, Chemicon, 1:500) in PBS + 1%triton X-100. Following PBS washes antibody binding wasvisualised with anti-mouse Alexa Xuor 488 (MolecularProbes, 1:400). Retinae were wholemounted on glassslides, air-dried and coverslipped using Fluormount G(Southern Biotechnology Associates, Inc.).
Retinal sections were air dried, re-hydrated (PBS + 0.2%Triton X-100, 10 mins) and incubated overnight at 4°Cwith appropriate antibodies to: �-III tubulin for RGCs (Tuj-1,Chemicon, 1:500; Yin et al. 2003); GFAP for astrocytesand Muller cells (Dako, 1:500; Stupp et al. 2005), MT-I/II(Dako, 1:500); megalin (Santa Cruz, 1:400); calbindin foramacrine cells (Chemicon, 1:250; Leaver et al. 2006), ED-1for macrophages (Serotec 1:500; Engelsberg et al. 2004) orwith bisbenzimide trihydrochloride nuclear stain (HoeschtNo. 33342, Sigma, 1 �g/ml). All dilutions were made inPBS containing 0.2% Triton X-100. Triplicate cell culturepreparations were Wxed (4% paraformaldehyde in 0.1 MPBS (pH 7.2), 20 min), permeabilised with 0.2% triton X-100 in PBS (5 min) and incubated with appropriate primaryantibodies as for retinal sections. Following PBS washesantibody binding was visualized with anti-mouse (1:400Alexa Xuor 488, Molecular Probes, Eugene, OR, USA) andanti rabbit (1:400, Alexa Xuor 546 Molecular Probes,Eugene, OR) secondary antibodies. Slides and mountedcoverslips were viewed by conventional Xuorescencemicroscopy (Leitz Diaplan), images captured (CoolSNAP-Pro imaging system, Image Pro software) and analysedusing Image J software where appropriate. For each immu-nohistochemical analysis ten sections from each of three
animals were prepared and viewed. Representative imagesare shown.
Data analysis
Analyses of neurite extension in dissociated retinal cultureswere carried out on a minimum of 200 RGCs per treatmentgroup, and on at least 100 neurites per treatment group inretinal explants. The number of RGCs in dissociated retinalcultures with neurites longer than 25 �m was expressed as aratio of the number of RGCs without neurites. Mean neuritelength of RGCs from explants was determined using ImageJ analysis software. DiVerences in ratios or mean neuritelengths (�m) were detected using ANOVA and diVerencesbetween individual treatments detected using Bonferroni/Dunn post hoc tests.
Results
MT-I/II and megalin expression in vivo
At P5 MT-I/II was present in RGCs (Fig. 1a–c), but wasabsent from both Muller cell processes and displaced ama-crine cells (Fig. 1d, e). In the optic nerve, MT-I/II was pres-ent only in RGC axons whereas megalin was exclusively inastrocytic processes (Fig. 1f). Megalin was presentthroughout the retina (Fig. 1g–i), co-localising with MT-I/IIin RGCs (Fig. 1j–l).
In adult animals, MT-I/II was not signiWcantly expressedin RGC somas (Fig. 2a) and, as for neonatal stages, wasabsent from amacrine and Muller cells (not shown). Bycontrast, megalin expression was high in RGCs and otherretinal layers (Fig. 2b); however, as for MT-I/II, the puta-tive receptor was not expressed by Muller cells. In the opticnerve, astrocytic processes were positive for megalin (notshown).
MT-I/II and megalin expression in vivo following optic nerve transection
In the adult retina MT-I/II was not present at signiWcantlevels in RGC somas at 1, 2 and 4 weeks following transec-tion (Fig. 2d). Unlike Wndings for normal rats, megalinexpression was absent from RGC somas and from other ret-inal layers at each time point (Fig. 2e).
At the lesion site, however, both MT-I/II and megalinwere clearly up-regulated 1 week after transection com-pared to the surrounding tissue (Fig. 2c, f). Unlike theappearance in normal animals, MT-I/II was associated withmacrophages (Fig. 2g–i) rather than with RGC axons orastrocytic processes. Megalin was also up-regulated inmacrophages (Fig. 2j–l). By 2 and 4 weeks, MT-I/II expres-
123
174 Exp Brain Res (2007) 183:171–180
sion was minimal and megalin was present only in occa-sional macrophages (not shown).
MT-IIA promoted RGC neurite outgrowth in vitro
In dissociated retinal cultures, MT-IIA stimulated neuriteformation rather than neurite extension. Although mostRGCs grew neurites, only a proportion of the RGCs formeda neurite longer than 25 �m, indicating that the process wasaxonal rather than dendritic (Goldberg et al. 2002). Follow-ing treatment with 3 or 10 �g/ml MT-IIA the proportion ofsuch RGCs increased approximately twofold (P · 0.05,Fig. 3a). Correspondingly the proportion of RGCs without
neurites fell (P · 0.05). Smaller but still signiWcantincreases were obtained if the results were expressed as aproportion of total RGCs extending neurites (P > 0.05).Nevertheless, there was no evidence for MT-IIA extendingneurite length.
By contrast, in retinal explants, MT-IIA stimulated neu-rite extension but not neurite formation. Administration of10 �g/ml MT-IIA did not signiWcantly eVect either theproportion of explants that developed neurite outgrowth(P-value = 0.1425) or the numbers of neurites per explant(P-value = 0.1479). However, 10 �g/ml MT-IIA signiW-cantly increased mean neurite length compared to controls(Fig. 3b) (P · 0.01).
Fig. 1 MT-I/II and megalin were present in P5 rat retina. Retinal sec-tions showing Tuj-1 (a green), MT-I/II (b red) and Tuj-1 plus MT-I/II(c MT-I/II positive RGCs are brown, arrow); MT-I/II (green), GFAP(red) and Hoescht (blue) retinal sections contained no brown MT-I/IIand GFAP positive Muller cells (d); MT-I/II (green), calbindin (red)and Hoescht (blue) showed no brown MT-I/II and calbindin positiveamacrine cells (e). Section of P5 optic nerve stained with megalin
(red), GFAP (green) and Hoescht (blue) showed megalin in purple/brown GFAP positive processes (arrow) (f). Retinal sections showingTuj-1 (g green), megalin (h red) and Tuj-1 plus megalin (i megalinpositive RGCs are brown, arrow). Retinal sections demonstrated co-localization of MT-I/II (j green) and megalin (k red) in the RGC layer(l MT-I/II and megalin positive cells are brown, arrow). All images at£25 magniWcation, scale 25 �m
123
Exp Brain Res (2007) 183:171–180 175
MT-I/II was released within 24 h of MT-IIA administration in vitro
Endogenous MT-I/II expression in untreated dissociatedretinal cultures was low and relatively uniform across thecell types present (Fig. 4a). Administration of 3 or 10 �g/mlMT-IIA increased intracellular MT-I/II levels (Fig. 4) inboth neurites and somas. Levels increased uniformly over
3 h (Fig. 4b–d), were maintained at 6 h (not shown) but haddecreased by 24 h, particularly in the cell somas (Fig. 4e).At 24 h, cultures were characterised by brightly stainingparticulate material that did not appear to be associatedwith cells and had adhered suYciently to the substrate towithstand immunocytochemical staining and washing pro-cedures. The deposit could be MT-IIA expelled from cells.By 48 h, endogenous MT-I/II was seen in only a few
Fig. 2 MT-I/II and megalin changed their cellular localizations inresponse to injury. The optic nerves of adult rats were transected andsections were prepared 1 week later. The RGC layer containing RGCsis indicated. Sections from normal adult retinae did not contain signiW-cant levels of MT-I/II (a) but did contain megalin (b arrow) in RGCs.By 1 week after optic nerve transection MT-I/II was unchanged (d) butmegalin was no longer present (e). Sections of the injury site 1 weekafter nerve transection demonstrated upregulated MT-I/II (green)
when compared to surrounding tissue (c). Megalin (red) was alsoup-regulated at the injury site (f). One week after injury MT-I/II (h)co-localised with ED-1 labelled macrophages (g) at the injury site,superimposed images show MT-I/II labeled macrophages (i arrow).Similarly, 1 week after injury megalin (k) co-localised with ED-1 la-belled macrophages (j) at the injury site, superimposed images showmegalin labeled macrophages (l arrow). All images at £25 magniWca-tion, scale 25 �m
123
176 Exp Brain Res (2007) 183:171–180
neurites and particulate matter (Fig. 4f). Cultures were fullyviable at this time point.
MT-I/II and megalin were present in MT-IIA treated RGCs
We clariWed which cells in the dissociated retinal culturescontained MT-I/II following administration of MT-IIA andwhether the putative MT-I/II receptor megalin was alsopresent. Following treatment with 3 �g/ml MT-IIA for 1, 3or 24 h, MT-I/II was present predominantly in RGC neu-rites and somas (Fig. 5a–c, 3 h treatment shown) but wasabsent from astrocytes at 1, 3 and 24 h (Fig. 5d, 1 h treat-ment shown). At all time points megalin was found both inRGC somas and astrocytes (Fig. 5e–h) decreasing in inten-sity by 24 h. Taken together, the Wndings indicate thatMT-I/II and megalin were present simultaneously in RGCsomas, compatible with megalin being the MT-I/II receptorin these cells.
MT-IIA treatment up-regulated megalin expression
Endogenous megalin levels were low in all cell types incontrol dissociated retinal cultures (Fig. 6a). Megalinexpression appeared up-regulated by treatment with 3 �g/ml
MT-IIA for 1 or 3 h (Fig. 6b, c), but had decreased by 24 h(Fig. 6d).
Blocking megalin reduced the neurite-stimulating eVects of MT-IIA treatment, despite allowing MT-I/II to enter cells
We used a megalin inhibitor to test whether megalin couldbe a functional receptor for MT-IIA in RGCs. Dissociatedretinal cells were or were not pre-treated with the anti-megalin monoclonal antibody, then cultured in the presenceof 3 �g/ml MT-IIA (Fig. 7). We demonstrated that pretreat-ment abrogated the MT-IIA stimulated increase in neuriteextension (Fig. 7a, P = 0.0013). Representative images ofTuj-1 stained RGCs treated with 3 �g/ml MT-IIA withoutor with anti-megalin antibody pretreatment are shown(Fig. 7b, c). The result indicated that access to megalin wasrequired for RGCs to respond to MT-IIA treatment. Therewas no eVect of the anti-megalin antibody on RGC num-bers (P > 0.05), precluding toxic eVects. Furthermore, pre-treatment with the anti-megalin antibody in the absence ofMT-IIA had no eVect on neurite extension (P > 0.05)implying speciWcity of action of the antibody.
It is important to note that blocking megalin did notresult in lower intracellular MT-I/II levels. There was nodiVerence in MT-I/II staining with or without pre-treatmentwith the anti-megalin antibody (Fig. 7d, e).
Discussion
We show here that MT-I/II is present in RGCs of neonatalrat together with megalin both in vivo and in vitro. Treat-ment of dissociated RGCs with MT-IIA stimulated neuriteextension and megalin expression and pre-treatment of reti-nal cells with anti-megalin antibody prevented RGC neuriteextension. Taken together these results indicate that mega-lin may function as an MT-I/II receptor in RGCs.
MT-I/II and megalin in vivo
The higher MT-I/II levels we report in the retinae ofyounger as compared to adult animals imply a previouslyunsuspected role for MT-I/II in retinal development. More-over, when RGC axons were transected in vivo, MT-I/IIlevels remained insigniWcant in RGC bodies and axons buttransiently increased in macrophages at the lesion site. It isnot yet clear whether the up-regulated MT-I/II was endoge-nous to these cells or was released by RGCs or astrocytes(Chung and West 2004). MT-I/II gene expression has beenshown to increase in RGCs following axotomy (Fischeret al. 2004). Our inability to detect MT-I/II protein in RGCsfollowing optic nerve transection may indicate translationalregulation, sensitivity issues or rapid MT-I/II secretion.
Fig. 3 MT-IIA stimulated RGCs to extend neurites. MT-IIA atincreasing concentrations from 0.01 to 10 �g/ml, or ZnSO4 control,were added to dissociated retinal cultures. Following 24 h, cells wereimmunocytochemically stained with Tuj-1. The number of RGCs withneurites longer than 25 �m was expressed as a ratio of the number ofRGCs without neurites (a). Retinal explant cultures were treated with1 or 10 �g/ml MT-IIA or ZnSO4 control and Wxed after 5 days in cul-ture. The average length of neurites (�m) from all explants with neuriteoutgrowth (§SEM) was determined using Image J software (b). Thenumbers of neurites measured per treatment group were 144, 134 and99, respectively. Statistical analyses were performed using ANOVAand post hoc tests (*P · 0.01 signiWcantly diVerent from controls)
(a)
(b) *
oitaR
0.40.2
0.6
0.8
1.21.4
1.0
MT-IIA Concentration (micrograms/ml)
0 0.01
MT-IIA Concentration (micrograms/ml)0
htgneL
etirueN
naeM
)sertemorci
m( 50100150200250300
*
*
0.1 1.0
1.0
3.0 10.0
10.0
123
Exp Brain Res (2007) 183:171–180 177
Further studies assessing Xuorescently labelled endogenousMT-I/II in the Wrst days following injury would help toaddress the issue. The MT-I/II up-regulation observed inastrocytes surrounding neural injury to the cerebral cortex(Chung et al. 2004) was not replicated in our study of injuryto the visual system. Up-regulation of MT-I/II in the lesionsite may lead to reduced inXammation as well as neuropro-tection, as indicated by studies in MT-I/II knockout mice(Giralt et al. 2002; Suemori et al. 2006). It is possible thatthe up-regulated MT-I/II may, as in the cortex (Chung et al.2003), also induce greater neuroregeneration. Assessmentof neuroregeneration following optic nerve injury in MT-I/II knockout mice would shed further light on the impor-tance of endogenous MT-I/II in macrophages followinginjury. Further to this we have added exogenous MT-IIA toa range of lesions of the CNS in vivo and demonstrated sig-niWcant improvements in neuroregeneration (manuscriptsubmitted).
EVects of MT-I/II in vitro
Our Wnding that exogenous MT-IIA stimulated RGC neu-rite extension/formation in two in vitro models accords
with Wndings for cortical brain injury following MT-IIAtreatment (Chung et al. 2003). Dissociated retinal culturesand retinal explants incorporate the many diVerent celltypes present in the retina and as such are realistic modelsof exogenous MT-IIA treatment. These two models demon-strated similar but slightly diVerent eVects of MT-IIA;treatment of dissociated cells leading to more RGCs withneurites and treatment of RGCs in retinal explants leadingto longer neurites. The cellular architecture maintained inretinal explants may play a role in deWning the responses ofRGCs to this agent.
Megalin as a MT-I/II receptor in RGCs
Megalin had previously been identiWed as a receptor forMT-I/II in renal cells but its role in the neuroprotectiveand neuroregenerative eVects of MT-I/II had yet to beexplored. Our observation that megalin is present in RGCsomas and astrocytic processes of young and adult rats,both in vivo and in vitro, is compatible with it being a MT-I/II receptor also in neurons. Given that RGCs can containboth MT-I/II and megalin, it is possible that MT-I/II mayfunction in an autocrine fashion, although such a loop
Fig. 4 MT-I/II presence follow-ing exogenous MT-IIA adminis-tration in dissociated retinal cultures. MT-IIA (10 �g/ml) was added to cultures 24 h after plating. MT-I/II presence in cells was assessed using an antibody to MT-I/II immediately (untreat-ed control a), 30 min (b), 60 min (c), 3 h (d), 24 h (e) or 48 h (f) after treatment. All images are shown at £25 magniWcation, scale 50 �m
123
178 Exp Brain Res (2007) 183:171–180
would rely on release of endogenous MT-I/II by thesecells. Our demonstration of a rapid MT-IIA-mediatedmegalin up-regulation in dissociated RGCs implies thatexpression of the receptor might be triggered during MT-I/IItherapy. The up-regulation would presumably be neces-sary for eVectivity since, as we reported, megalin expres-sion in RGCs is reduced in vivo 1 week after nervetransection. The observed accumulation of megalin,together with MT-I/II, in macrophages at the lesion sitefurther supports the relationship between MT-I/II andmegalin. The lack of MT-I/II in astrocytes expressingmegalin implies that astrocytes process MT-I/II diVerentlyto RGCs.
Our report of MT-IIA-mediated RGC neurite extensionbeing prevented by an anti-megalin antibody provides evi-dence of a functional role for megalin as an MT-I/II recep-tor in RGCs, while not precluding additional MT-I/IIreceptors in these cells. It is intriguing that the inhibition ofMT-IIA-mediated neurite extension by the anti-megalinantibody occurred despite increased levels of MT-I/II inMT-IIA treated RGCs. Possibly the observed high levels ofintracellular MT-I/II were not exogenously derived but rep-resented endogenous up-regulation in response to bindingof megalin with the antibody. However, this possibilitywould not explain the failure to stimulate neurite extension.We consider it more likely that MT-IIA generates its eVects
Fig. 5 Cellular localization of MT-I/II and megalin in dissoci-ated retinal cultures. Cultures were treated with 3 �g/ml MT-IIA then Wxed and stained immu-nocytochemically. MT-I/II (red a) was present predominantly in RGCs (green b), indicated by the brown/green colour follow-ing 3 h treatment with MT-IIA (c). Red staining separate from RGCs was not above back-ground. MT-I/II (green) was not present in GFAP positive cells (red) for 24 h following treat-ment (1 h treatment shown, d). Direct comparison of images (e) and (f) illustrate that megalin (red) was present in GFAP posi-tive (green) astrocytes (arrow) following 3 h treatment with MT-IIA. Similarly, direct com-parison of images (g) and (h) illustrate that megalin (red) was present in RGCs (green, arrows) following 1 hour treatment with MT-IIA. All images at £40 magniWcation, scale 50 �m
123
Exp Brain Res (2007) 183:171–180 179
Fig. 6 Treatment with MT-IIA up-regulated megalin expression in mixed retinal cultures. Fixed cultures were stained with anti-megalin antibody following treatment with ZnSO4 for 1 h (a), 3 �g/ml MT-IIA for 1 (b), 3 (c) or 24 h (d). All images at £40 magniWcation, scale 50 �m
Fig. 7 Pretreatment of dissoci-ated retinal cultures with anti-megalin antibody prevented MT-IIA stimulation of neurite extension but did not abolish MT-I/II in RGCs. Dissociated retinal cultures were (a, c, e) or were not (a, b, d) pre-treated with 10 �l of anti-megalin anti-body for 30 min followed by treatment with 3 �g/ml MT-IIA. Following 24 h, cells were immunocytochemically stained with Tuj-1 (b, c) or anti-MT-I/II (d, e). The number of RGCs with neurites longer than 25 �m was expressed as a ratio of the num-ber of RGCs without neurites (a). The ratio of neurites greater than 25 �m in untreated control cultures is shown as a line across the graph at 0.41 § 0.01. Statis-tical analyses were performed using ANOVA and post hoc tests (*P < 0.05 signiWcantly lower than MT-IIA treated RGCs without anti-megalin anti-body). Images at £25 magniWca-tion, scale 50 �m
(a)
*Rat
io
tnemtaerT
AII-TM + AII-TMnilagem-itna
2.0
4.0
6.0
8.0
0.1
(d) ydobitna nilagem-itna oN (e) ydobitna nilagem-itna htiW
(b) ydobitna nilagem-itna oN (c) ydobitna nilagem-itna htiW
II/I-TM-itna
sCGR
123
180 Exp Brain Res (2007) 183:171–180
via signal transduction pathways activated by the NPxYmotifs of megalin’s cytoplasmic tail (Qui et al. 2006).Obstruction of megalin by the anti-megalin antibody mayprevent MT-IIA binding to megalin and activating thesepathways. Co-immunoprecipitation of MT-I/II and megalinmay contribute to evidence of this interaction. Neverthe-less, irrespective of the mechanism involved, our Wndingssupport the potential of MT-I/II as a neuroregenerativeagent and implicate megalin as its receptor.
Acknowledgments The authors would like to thank: Lauren Evilland Sherralee Lukehurst, School of Animal Biology, for technicalwork and advice with image analysis; Dr Peter Arthur, BiochemistryDepartment, UWA; and Dr Jennifer Rodger and Prof. Sarah Dunlop,Experimental and Regenerative Neurosciences, School of AnimalBiology, UWA. This work was supported by the NeurotraumaResearch Program, Western Australia.
References
Avwenagha O, Campbell G, Bird MM (2003) Distribution of GAP-43,�-III tubulin and F-actin in developing and regenerating axonsand their growth cones in vitro, following neurotrophin treatment.J Neurocytol 32:1077–1089
Chung RS, West AK (2004) A role for extracellular metallothioneinsin CNS injury and repair. Neuroscience 123:595–599
Chung RS, Vickers JC, Chuah MI, West AK (2003) Metallothionein-IIA promotes initial neurite elongation and postinjury reactiveneurite growth and facilitates healing after focal cortical brain in-jury. J Neurosci 23:3336–3342
Chung RS, Adlard PA, Dittmann J, Vickers JC, Chuah MI, West AK(2004) Neuron-glia communication: metallothionein expressionis speciWcally up-regulated by astrocytes in response to neuronalinjury. J Neurochem 88:454–461
Dunlop SA (2003) Axonal sprouting in the optic nerve is not a prereq-uisite for successful regeneration. J Comp Neurology 465:319–334
Engelsberg K, Ehinger B, Wasselius J, Johansson K (2004) Apoptoticcell death and microglial cell responses in cultured rat retina. Gra-efe’s Arch Clin Exp Ophthalmol 242:229–239
Fischer D, Petkova V, Thanos S, Benowitz LI (2004) Switching matureretinal ganglion cells to a robust growth state in vivo: geneexpression and synergy with RhoA inactivation. J Neurosci24:8726–8740
Geiger LK, Kortuem KR, Alexejun C, Levin LA (2002) Reduced re-dox state allows prolonged survival of axotomised neonatal reti-nal ganglion cells. Neuroscience 109:635–642
Giralt M, Penkowa M, Lago N, Molinero A, Hidalgo J (2002) Metal-lothionein-1 + 2 protect the CNS after a focal brain injury. ExpNeurol 173:114–128
Goldberg JL, Klassen MP, Hua Y, Barres BA (2002) Amacrine-sig-naled loss of intrinsic axon growth ability by retinal ganglioncells. Science 296:1860–1864
Klassen RB, Crenshaw K, Kozyraki R, Verroust PJ, Tio L, Atrian S,Allen PL, Hammond TG (2004) Megalin mediates renal uptake ofheavy metal metallothionein complexes. Am J Physiol RenalPhysiol 287:F393-F403
Leaver SG, Cui Q, Plant GW, Arulpragasam A, Hisheh S, VerhaagenJ, Harvey AR (2006) AAV-mediated expression of CNTF pro-motes long-term survival and regeneration of adult rat retinal gan-glion cells. Gene Ther 13:1328–1341
Penkowa M, Hidalgo J (2001) Metallothionein treatment reduces pro-inXammatory cytokines IL-6 and TNF-� and apoptotic cell deathduring experimental autoimmune encephalitis (EAE). Exp Neurol170:1–14
Qiu S, Korwek KM, Weeber E.J (2006) A fresh look at an ancientreceptor family: emerging roles for low density lipoprotein recep-tors in synaptic plasticity and memory formation. NeurobiolLearn Mem 85:16–29
Rodger J, Lindsey KA, Leaver SG, King CE, Dunlop SA, Beazley LD(2001) Expression of ephrin-A2 in the superior colliculus andEphA5 in the retina following optic nerve section in adult rat. EurJ Neurosci 14:1929–1936
Stupp T, Pavlidis M, Busse H, Thanos S (2005) Lens epithelium sup-ports axonal regeneration of retinal ganglion cells in a coculturemodel in vitro. Exp Eye Res 81:530–538
Suemori S, Shimazawa M, Kawase K, Satoh M, Nagase H, YamamotoT, Hara H (2006) Metallothionein, an endogenous antioxidantprotects against retinal neuron damage in mice. Invest Ophthal-mol Vis Sci 47:3975–3982
Trendelenburg G, Prass K, Priller J, Kapinya K, Polley A, MuselmannC, Ruscher K, Kannbley U, Schmitt AO, Castell S, Wiegand F,Meisel A, Rosenthal A, Dirnagl U (2002) Serial analysis of geneexpression identiWes metallothionein-II as major neuroprotectivegene in mouse focal cerebral ischemia. J Neurosci 22:5879–5888
WolV NA, Abouhamed M, Verroust PJ, Thevenod F (2006) Megalin-dependent internalization of cadmium–metallothionein and cyto-toxicity in cultured renal proximal tubule cells. J Pharmacol ExpTher 318:782–791
Yin Y, Cui Q, Li Y, Irwin N, Fischer D, Harvey AR, Benowitz LI(2003) Macrophage derived factors stimulate optic nerve regener-ation. J Neurosci 23:2284–2293
