Friedreich's ataxia: Oxidative stress and cytoskeletal abnormalities

Journal of the Neurological Sciences 287 (2009) 111–118

Contents lists available at ScienceDirect

Journal of the Neurological Sciences

j ourna l homepage: www.e lsev ie r.com/ locate / jns

Friedreich's ataxia: Oxidative stress and cytoskeletal abnormalities

Marco Sparaco a,1, Laura Maria Gaeta b,1, Filippo Maria Santorelli b, Chiara Passarelli b, Giulia Tozzi b,Enrico Bertini b, Alessandro Simonati c, Francesco Scaravilli d, Franco Taroni e, Charles Duyckaerts f,Michele Feleppa a, Fiorella Piemonte b,⁎a Division of Neurology, Department of Neurosciences, Azienda Ospedaliera “G. Rummo”, 82100 Benevento, Italyb Molecular Medicine Unit, Children's Hospital and Research Institute “Bambino Gesù”, Roma, Italyc Department of Neurological and Visual Sciences, Section of Neurology, University of Verona, Policlinico G.B. Rossi, 37134 Verona, Italyd Division of Neuropathology, Institute of Neurology, University College London, WC1N 3BG London, UKe UO Biochimica e Genetica, Fondazione IRCCS Istituto Neurologico “Carlo Besta”, Milano, Italyf Laboratoire de Neuropathologie Raymond Escourolle, Hôpital de La Salpêtrière, 75651 Paris, France

⁎ Corresponding author. Molecular Medicine Unit, ChInstitute “Bambino Gesù”, P.za S. Onofrio, 4, 00165 Romefax: +39 06 68592024.

E-mail address: [email protected] (F. Piemonte).1 The first two authors contributed equally to this wo

0022-510X/$ – see front matter © 2009 Elsevier B.V. Adoi:10.1016/j.jns.2009.08.052

a b s t r a c t

Article history:Received 2 April 2009Received in revised form 24 July 2009Accepted 13 August 2009Available online 12 September 2009

Keywords:GlutathioneOxidative stressFriedreich's ataxiaCytoskeletal proteins

Friedreich's ataxia (FRDA) is an autosomal recessive disorder caused by mutations in the gene encodingfrataxin, a mitochondrial protein implicated in iron metabolism. Current evidence suggests that loss offrataxin causes iron overload in tissues, and increase in free-radical production leading to oxidation andinactivation of mitochondrial respiratory chain enzymes, particularly Complexes I, II, III and aconitase.Glutathione plays an important role in the detoxification of ROS in the Central Nervous System (CNS), whereit also provides regulation of protein function by glutathionylation. The cytoskeletal proteins are particularlysusceptible to oxidation and appear constitutively glutathionylated in the human CNS. Previously, weshowed loss of cytoskeletal organization in fibroblasts of patients with FRDA found to be associated withincreased levels of glutathione bound to cytoskeletal proteins. In this study, we analysed the glutathionyla-tion of proteins in the spinal cord of patients with FRDA and the distribution of tubulin and neurofilaments inthe same area. We found, for the first time, a significant rise of the dynamic pool of tubulin as well asabnormal distribution of the phosphorylated forms of human neurofilaments in FRDA motor neurons. In thesame cells, the cytoskeletal abnormalities co-localized with an increase in protein glutathionylation and themitochondrial proteins were normally expressed by immunocytochemistry. Our results suggest that in FRDAoxidative stress causes abnormally increased protein glutathionylation leading to prominent abnormalitiesof the neuronal cytoskeleton.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

Friedreich's ataxia (FRDA) is an autosomal recessive disorder,neuropathologically characterized by prominent degeneration of thespinal cord pathways, with predominant neuronal loss in the dorsalroot ganglia and in the Clarke's columns, together with degenerationof their long tracts as well as involvement of the pyramidal tracts [1].

Most FRDA patients are homozygous for expanded GAA triplet-repeat sequences (E alleles) in intron 1 of the FXN gene on chromosome9q13 [2]. This mutation causes the formation of a “sticky” triplex DNAstructure that interferes with correct transcription and reduces the

ildren's Hospital and Research, Italy. Tel.: +39 06 68592105;

rk.

ll rights reserved.

synthesis of frataxin, a mitochondrial protein of 210 aminoacids, whichis normally expressed at high levels in the human spinal cord [3,4].

There is evidence that frataxin is closely involved in the early stepsof the iron–sulphur cluster (ISC) biosynthesis and acts as an iron-storage protein, keeping iron in a non-toxic and bioavailable form[5–7]. Therefore, an excess of unbound iron in mitochondria, resultingfrom frataxin deficiency, might induce the oxidation of cellularcomponents and the inactivation of mitochondrial enzymes, throughits capacity to produce free radicals [8]. Accordingly, yeast lacking thefrataxin homologue gene (YFH1) shows a severe defect of mitochon-drial respiration, intramitochondrial iron accumulation and increasesensitivity to oxidative stress [9]. Moreover, selective knockdown offrataxin in murine heart muscle causes decreased activity particularlyof mitochondrial Fe–S dependent enzymes, i.e. Complexes I, II, III andaconitase, and late-stage iron accumulation [10].

The role of oxidative stress in the pathophysiology of the disease isstill controversial. No evidence of oxidative stress has been found instudies of conditional knockout mouse models [11], whereas

112 M. Sparaco et al. / Journal of the Neurological Sciences 287 (2009) 111–118

“humanized” GAA repeat expansion mouse models of FRDA exhibitedoxidative stress leading to progressive neuronal and cardiac pathology[12].

Moreover, other studies have found increased plasma and urinarylevels of oxidative stress markers in FRDA patients [13–17] and havedemonstrated that some antioxidants, such as idebenone and vitamin E,might have a protective role [18,19].

Several lines of evidence suggest that the disarray of microfilamentand microtubule network represents one of the early events in thedegenerative process of neurons exposed to oxidative stress [20,21]. IndifferentiatedPC-12 cells, for instance, neuronsexposed to100 µMH2O2

present microtubules depolymerization and cell death [22]. Further-more, it has been demonstrated that oxidative stress might negativelyinterfere with the phosphorylation of neurofilaments, leading to animpairment of the axonal transport [23].

We moved from the observation of abnormal cytoskeletal organi-zation infibroblasts of FRDApatients thatwere associated to abnormallyincreased levels of glutathione bound to cytoskeletal proteins [14].

Presently, we studied by immunohistochemistry the spinal cordfrom four autoptic cases of FRDA to better understand the pathogenesisof neuronal degeneration in FRDA. For this purpose, we first tried toconfirm in the human spinal cord of FRDA patients with frataxindeficiency some evidence of mitochondrial dysfunction in the respira-tory chain enzymes. Then, as abnormally increased glutathionylation ofproteins represents a sensitive redox-marker of tissue oxidative stress,we studied the amount and the distributionof glutathionylated proteinsin the same area of the CNS using a monoclonal antibody (ab) thatspecifically reacts with glutathione bound to proteins (GS-Pro). Finally,because cytoskeletal protein thiols are particularly susceptible tooxidation, we analysed the expression of some cytoskeletal proteinsthat are essential for neuronal integrity and function.

2. Materials and methods

2.1. Human samples

Tissues were obtained at autopsy, performed 2 to 10 h after death,fromfourpatientswithFRDA.They included two females (patient2 and4;25 and 33 years old, respectively) and twomales (patient 1 and 3; 47 and24 years old, respectively). All patientsmet the diagnosticHarding criteriafor typical FRDA [24] that include onset before the age of 25, progressivelimb and gait ataxia, absent tendon reflexes in the legs, followed by(within5 yearsofpresentation)dysarthria, areflexiaat all four limbs, signsof pyramidal tract dysfunction in the legs, and distal loss of position andvibration sense [3,24]. The diagnosis was genetically confirmed bydetecting the abnormal GAA expansion in the frataxin gene in cases 1and 2. Technical reasons (i.e. the amount of tissue available) did not allowto perform the molecular genetic analysis in the remaining two cases. Ascontrols, we used comparable specimens from ten age-matchedindividuals with no history of neurological diseases and no neuropath-ological lesions at post-mortemexamination. None of the patients used ascontrols had sufferedmajor respiratory stress/distress or other conditions(i.e. prolonged hypoxia, infective diseases or toxic injury, etc.) that couldlead to tissueoxidative stress. Consentwasobtained fromour InstitutionalEthic Committee for using humanmaterial for research. Samples used forimmunohistochemical studieswerefixed in 10%neutral formalin at roomtemperature for several weeks, sliced, and embedded in paraffin.Immunohistochemical studies were performed on transverse sectionsthrough upper and lower cervical segments, and through lower lumbarsegments of the spinal cord.

2.2. Molecular genetic analyses

Molecular analyses were performed on tissues extracted fromparaffin-embedded material as reported in [25]. Detection of E allelesin tissues and sizing of the “constitutional ormost common” allelewere

estimated by a conventional polymerase chain reaction (PCR)-basedmethodology as outlined in [26]. Briefly, adopting oligonucleotideprimers GAA–104F: (5′–3′) GGCTTAAACTTCCCACACGTGTT and GAA–629R: (5′–3′) AGGACCATCATGGCCACACTT we PCR amplified genomicDNA using the Boehringer-Mannheim Long Template PCR system. PCRconditions were 94 °C for 2 min, 17 cycles at 94 °C for 10 s, 68 °C for2 min 30 s followed by 20 cycles of 94 °C for 10 s, 68 °C for 2 min 30 swith the addition of 20 s to the extension time per cycle, and a finalextension at 68° for 10 min. This generates a PCR product of (500+3n)base pairs, where n is the number of GAA repeats. The size of the PCRproductwas estimated using appropriate size standards (Invitrogen) onagarose 0.8% gel stained with ethidium bromide. The GAA repeatexpansions were 615/681 for case 1 and 650 for case 2.

2.3. Immunohistochemistry

Transverse 4 μm-thick formalin fixed and paraffin-embeddedsections of the samples were studied immunohystochemically withthe following antibodies (abs) diluted from 1:5 to 1:500 in phosphate-buffered saline (PBS):

• monoclonal ab against glutathione bound to proteins (GS-Pro)(Virogen, Watertown, MA) [14,27,28];

• monoclonal ab against human frataxin (frataxin) (ImmunologicalSciences, Rome, Italy);

• monoclonal ab against the 30 kDa ISC-containing subunit II ofComplex II (CII Ip) (Molecular Probes, Eugene, OR, USA);

• monoclonal ab against subunit IV of cytochrome c oxidase (COX IV)(Molecular Probes, Eugene, OR, USA);

• monoclonal ab against the human C-terminal β-tubulin (tubulin)(Sigma, St. Louis, MI, USA);

• monoclonal ab against the N-terminal heavy chain of humanneurofilaments (NF-H) (Lab Vision, Fremont, CA, USA);

• monoclonal abagainst thenonphosphorylatedepitope inneurofilamentH (SMI 32) (Covance, Berkeley, CA, USA);

• monoclonal ab against the phosphorylated form in NF-H (SMI 34)(Covance, Berkeley, CA, USA);

• polyclonal ab against tyrosinated tubulin (Tyr-Tub) (Chemicon Interna-tional, Temecula, CA, USA);

• polyclonal ab against detyrosinated tubulin (Glu-Tub) (ChemiconInternational, Temecula, CA, USA).

Sections were immunostained with the avidin–biotin–peroxidasecomplex (ABC) according to previous published techniques [28,29].

Nuclear counter-staining was performed with Mayer hematoxylinsolution. The primary ab was omitted in immunocytochemical controlsections. In order to compare staining intensity, the tissues of patientsand controls were processed together in single runs, exactly in the samelabelling conditions of antibody exposure and background blocking.

3. Results

3.1. Immunohistochemistry

Samples of the spinal cord, which is known to be predominantlyaffected by neuronal degenerative lesions in FRDA [1], were studiedimmunohistochemically.

To confirm whether the reduced expression of frataxin wasassociated with mitochondrial respiratory chain enzymes deficiency inthe spinal cord, cervical and lumbar sections from control individualsand patients were incubated with abs against frataxin and against thesubunit II of Complex II (CII Ip) and the subunit IV of cytochromec oxidase (COX IV) of the mitochondrial respiratory chain. Immuno-reactivity for frataxin was absent or markedly reduced in the graymatter neurons as well as in cells and axons of the white matter of thefour FRDA patients (Fig. 1), while a normal immunostaining was

Fig. 1. Immunostaining of cervical sections of the spinal cord from a control and a patient for the localization of frataxin (A, D, G, L), mitochondrial CII Ip (B, E, H,M) and COX IV (C, F, I, N)protein subunits. The patient shows a negative immunostaining for frataxin and normal for COX IV and CII Ip in the anterior horn cells and in cells and fibers of the fasciculus gracilis. Theimages are representative of all the patients tested. Scale bar=50 μm. For details, see Materials and methods.

113M. Sparaco et al. / Journal of the Neurological Sciences 287 (2009) 111–118

detected both in controls and patients using abs recognizing CII Ip andCOX IV (Fig. 1).

To evaluate the degree and the extent of free-radical cytotoxicity aswell as the role of glutathionylation in the redox regulation of cellularsurvival in FRDA, sections from control individuals and patients wereincubated with abs against GS-Pro, which specifically recognizesglutathionylated proteins. Sections from controls showed a mildimmunoreactivity appearing as a granular pattern in the gray matteras well as in cells and axons of the white matter (Fig. 2). In particular,immunostaining was observed in large somatic motor neurons of theanterior horns (Fig. 2) and in neuronal cells of posteromarginal nucleus(Fig. 2), nucleus proprius dorsalis, dorsal nucleus of Clarke, and nucleusintermediomedialis. In FRDA patients, immunostainingwas significant-ly stronger for GS-Pro (Fig. 2), suggesting a significant increase ofprotein glutathionylation in all gray matter neurons, particularly inmotor neurons of the anterior horns (Fig. 2), as well as in the axonalspinal tracts (mostly in posterior white columns, spinocerebellar andcorticospinal tracts).

To verify whether the increased glutathionylation of proteinsdetected in FRDA spinal cord cases might be correlated with theabnormal polymerisation of the cytoskeleton, as previously demon-

strated in fibroblasts of patients with FRDA [14], sections from controlsand patients were incubated with immunological probes recognizingspecific subunits of cytoskeletal proteins. Anormal immunostainingwasdetected in both controls and patients using abs recognizing β-tubulin(tubulin) (Figs. 3 and 4) and the heavy chain of neurofilaments (NF-H)(Figs. 3 and 4). When sections of both controls and patients wereincubated with abs recognizing tyrosinated tubulin (Tyr-Tub) anddetyrosinated tubulin (Glu-Tub) (respectively associated to dynamicand stable microtubules), there was a clear increased staining forTyr-Tub in the anterior hornmotor neurons of all FRDA patients (Figs. 3and 4), whereas Glu-Tub showed normal immunoreactivity in the samecells (Fig. 3). The distribution of phospho-dephospho NF-H was thenanalysed using abs that specifically recognize the de-phosphorylatedform of NF-H (SMI 32) and the phosphorylated one (SMI 34), normallylocated in the axons of neurons. Noteworthy, in the anterior horns ofFRDA cases there was an increase of neuronal staining for SMI 34with acellular distribution of the immunoreactivity preferentially involvingperykaria (Figs. 3 and 4). No significant perikaryal staining occurredwith the same abs in control tissues (Figs. 3 and 4). Conversely, absagainst SMI 32 showed a milder immunoreactivity in spinal motorneurons fromFRDApatients (Fig. 3)when compared to controls (Fig. 3).

Fig. 2. Immunostaining of cervical sections of spinal cord from a control and two patients localizing glutathione bound to proteins (GS-Pro). The patients show a strong immunoreactivityfor GS-Pro, suggesting an increase of protein glutathionylation in cells and fibers of fasciculus gracilis (B, C), posteromarginal nucleus (E, F) and anterior horns (H, I). The images arerepresentative of all the patients tested. Scale bar=50 μm. For details, see Materials and methods.

114 M. Sparaco et al. / Journal of the Neurological Sciences 287 (2009) 111–118

These findings indicate an abnormal distribution of the phosphorylatedform of NF-H in anterior horn cells of FRDA patients, with a pathologicperikarion accumulation.

Finally, to determine whether an alteration of cytoskeletonorganization might be directly related to an abnormally increasedprotein glutathionylation in FRDA, serial sections of the spinal cordfrom controls and patients were stained for the immunolocalization oftubulin, GS-Pro, and SMI 34. Interestingly, in patients with FRDAmotor neurons strongly reacting with GS-Pro had also an increasedimmunostain for Tyr-Tub and SMI 34 (Fig. 5), suggesting that theincrease of protein glutathionylation co-localizes with an abnormalpolymerization of microtubules and a pathological distribution of thephosphorylated form of NF-H in the same cells.

4. Discussion

The precise sequence of pathogenetic events in FRDA remainsuncertain. Current evidence suggests that loss of frataxin impairsmitochondrial iron homeostasis resulting in respiratory chainenzymes dysfunction with excess in free-radical production [3].Reduced activities of mitochondrial holoproteins containing iron–sulphur groups as well as an impairment of tissue energy metabolismhave been demonstrated by biochemical and 31P-MRS studies incardiac and skeletal muscle from FRDA patients [9,30–33]. Further-more, the key role of oxidative stress in the pathophysiology of thedisease has been supported by the finding of increased blood andurinary levels of oxidative stress markers in FRDA patients [13,16] aswell as by in vivo evidence of impairment of antioxidant enzymes andof glutathione homeostasis [15,17].

Following our previous observations which showed loss of thecytoskeletal organization in fibroblasts of patientswith FRDA associatedwith increased levels of glutathione bound to the cytoskeletal proteins

[14], we decided to analyse both the expression and the topographicdistribution of protein-bound glutathione in the spinal cord derivedfrom autopsies of four FRDA patients because the spinal cord isparticularly involved in this condition. We demonstrated in all casesan abnormally increased immunoreactivity for GS-Pro in gray matterneurons as well as in cells and axons of the white matter, suggestingabnormal protein glutathionylation in this condition. Considering thatglutathionylated proteins may be used as a biomarker for oxidativestress [34,35], these findings clearly indicate the presence of anoxidative stress in the spinal cord of our FRDApatients as a consequenceof reduced frataxin expression. Unfortunately, the small number ofpatients examined does not allow us to establish any correlationbetween the degree of abnormal protein glutathionylation and thegenotype.

Cytoskeletal proteins (actin, β-tubulin and NF) are constitutivelyglutathionylated in human CNS [35] and are sensitive to the shifts in theredox state of the cells so that theymay represent a target of free-radicaldamage [20,36]. Accordingly, actin oxidation has been reported in somehuman neurodegenerative diseases [37,38] and actin glutathionylationhas been shown to impair cytoskeletal functions in fibroblasts ofpatients with FRDA [14].

Therefore, to increase our knowledge in the pathogenesis of theneuronal degeneration in FRDA, we analysed the expression of somecytoskeletal proteins considered essential for neuronal function in ourcases. We found abnormally increased immunostaining for tyrosinatedtubulin and normal staining for detyrosinated tubulin (Fig. 3). Becausethese proteins are respectively associated to dynamic and stablemicrotubules in the nervous tissue, our findings seem to indicate asignificant rise of the dynamic pool of tubulin compared to the amountof the stable form, suggesting a different extent of microtubularpolymerization in FRDA. An alteration of microtubules dynamics,caused by an increased expression of Tyr-Tub in FRDA, may have

Fig. 3. Immunostainingof lumbar sectionsof the spinal cord fromacontrol andtwopatients localizing tubulin (A,B, C), tyr-tubulin (D,E, F), glu-tubulin(G,H, I),NF-H(L,M,N), SMI32(O,P,Q),and SMI34 (R, S, T).Motor neuronsof patients showa slightly reduced immunoreactivity for SMI32 (P,Q), and increased for SMI34 (S, T) and tyr-tubulin (E, F),when compared to controls (O,R, D). The images are representative of all the patients tested. Scale bar=50 μm. For details, see Materials and methods.

115M. Sparaco et al. / Journal of the Neurological Sciences 287 (2009) 111–118

profound effects in the cytoskeletal network and axonal flow leading toimpaired trafficking of organelles and eventual collapse of the axonalnetwork [39–41].

In addition,we show for thefirst time in spinal cordneurons of FRDApatients an abnormally increased neuronal staining for the phosphor-ylated form of NF-H (SMI 34) with a cellular distribution of the

Fig. 4. Immunostaining of cervical sections of the spinal cord from a control and two patients localizing tubulin (A, E, I), tyr-tubulin (B, F, L), NF-H (C, G, M) and SMI 34 (D, H, N).Motor neurons of patients show increased immunoreactivity for tyr-tubulin (F, L) and SMI 34 (H, N), when compared to the controls (B, D). The images are representative of all thepatients tested. Scale bar=50 μm. For details, see Materials and methods.

116 M. Sparaco et al. / Journal of the Neurological Sciences 287 (2009) 111–118

immunoreactivity preferentially involving perykaria (Figs. 3 and 4). Asphosphorylation of the heavy NF subunit by several protein kinasesinvolves physiologically more NF located in the axon than those locatedin the perikaryon [42,43], our findings suggest an abnormal distributionof phosphorylated NF-H in FRDA motor neurons. Of note, abnormalphosphorylation and neuronal accumulation of NF, probably related toanactivation of severalNF kinases bymeans of the oxidative stress, havebeen reported in several neurodegenerative diseases such as amyo-trophic lateral sclerosis (ALS), Alzheimer's disease (AD) and Parkinson'sdisease (PD) [44–48].

Overall, our observations suggest alterations of the neuronalcytoskeleton that may contribute to explain the “dying-back” phenom-enon characterizing the pattern of long tract degeneration in FRDA.Particularly, we hypothesize that the abnormal microtubular polymer-ization, found in our cases, could alter the interaction of thesecytoskeletal components with the major microtubule-based molecularmotors, such as kinesin-1 and cytoplasmic dynein, leading to deregu-lation of the axonal transport [40,49].

In the present work, we also show normal expression of the twosubunits of the mitochondrial respiratory chain, namely the nDNA-encoded subunit II of Complex II and the COX IV subunit, in graymatteranterior horn cells (Fig. 2) as well as in areas of the spinal cord that areusually involved in FRDA, such as the spinocerebellar tracts, posterior

columns, and corticospinal tracts (not shown). Our observations are inagreement with previous data [30] reporting biochemically normalactivities of Complexes I, II/III and aconitase in the cerebellumanddorsalroot ganglia from FRDA patients. Normal immunostaining for the ISC-containing protein subunit II of Complex II of our cases suggests thatfrataxin deficiency does not impair biogenesis of ISC in the spinal cordand, consequently, does not alter the stability or the assembly of ISC-containingproteins, such as the succinate-CoQoxidoreductase complex.As shown inYFH1 [50], lack of frataxin could rather impair the activity ofSDH leading to the generation of superoxide radicals and oxidativestress in mitochondria. As in PD, the increased formation of GSSG inmitochondria, caused bymonoamineoxidase-derivedH2O2,might reactwith the thiol groups of proteins of the respiratory chain complexes toform protein mixed disulphides, resulting in the suppression of SH-dependent electron transport [51]. Moreover, inhibition of severalmitochondrial enzymes such as succinate dehydrogenase and isocitratedehydrogenase has been observed in primate intestinal mitochondriaexposed to GSSG [52]. Unfortunately, the lack of frozen samples did notallow us to assay the mitochondrial enzyme activities in the spinal cordhomogenates of our patients.

In summary, on the basis of the data available from our patientsamples, this study highlights the abnormal glutathionylation ofproteins and alteration of the cytoskeletal organization in several

Fig. 5. Serial sections of the lumbar spinal cord from a patient and a control stained for the immunolocalization of tyr-tubulin (A, B), GS-Pro (C, D), and SMI 34 (E, F). Motor neurons ofthe patient show a strong immunostain for tyr-tubulin (B), GS-Pro (D) and SMI 34 (F), compared to the control (A, C, E). These findings indicate that the increase of proteinglutathionylation co-localizes with an alteration of microtubular polymerization, and with an abnormal distribution of the phosphorylated form of NF-H. High magnification: scalebar=50 μm; low magnification: scale bar=200 μm. For details, see Materials and methods.

117M. Sparaco et al. / Journal of the Neurological Sciences 287 (2009) 111–118

structures of the spinal cord, consisting in abnormal microtubularpolymerisation and irregular distribution of phosphorylated NF-H inFRDA. However, further studies are needed to extend our investigationsto additional patients and to other areas of the CNS.

Acknowledgement

We are very grateful to Roberto Virgili (Department of PathologicAnatomy, Children's Hospital “Bambino Gesù” IRCSS) for his experttechnical assistance.

References

[1] Oppenheimer D, EsiriM. Disease of the basal ganglia, cerebellumandmotor neurons.In: Adams JH, Corsellis J, Duchen LW, editors. Greenfield's neuropathology. Ed 5.London: Arnold; 1992. p. 988–1054.

[2] Fogel BL, Perlman S. Clinical features and molecular genetics of autosomalrecessive cerebellar ataxias. Lancet Neurol 2007;6(3):245–57.

[3] Alper G, Narayanan V. Friedreich's ataxia. Pediatr Neurol 2003;28(5):335–41.[4] Sakamato N, Chastain PD, Parniewski P, Ohshima K, Pandolfo M, Griffith JD, et al.

Sticky DNA: self-association properties of long GAA TTC repeats in RR Y triplexstructures from Friedreich's ataxia. Mol Cell 1999;3(4):465–75.

[5] Adamec J, Rusnak F, Owen WG, Naylor S, Benson LM, Gacy AM, et al. Iron-dependent self-assembly of recombinant yest frataxin: implications for Frie-dreich's ataxia. Am J Hum Gen 2000;67(3):549–62.

[6] Gerber J, Muhlenhoff U, Lill R. An interaction between frataxin and Isu1/Nfs1 thatis crucial for Fe/S cluster synthesis on Isu1. Embo Rep 2003;4(9):906–11.

[7] Muhlenhoff U, Richhardt N, RistowM, Kispal G, Lill R. The yeast frataxin homologueYfh1pplays a specific role in thematuration of cellular of Fe/S proteins. HumMol Gen2002;11(17):2025–36.

[8] Schon EA, Manfredi G. Neuronal degeneration and mitochondrial dysfunction. J ClinInvest 2003;11(3):303–12.

[9] Calabrese V, Lodi R, Tonon C, D'Agata V, SapienzaM, Scapagnini G, et al. Oxidativestress, mitochondrial dysfunction and cellular stress response in Friedreich'sataxia. J Neurol Sci 2005;233(1–2):145–62.

[10] Puccio H, SimonD, CosseeM, Criqui-Filipe P, Tiziano F, Melki J, et al. Mousemodels forFriedreich ataxia exhibit cardiomyopathy, sensory nerve defect and Fe–S enzymedeficiency followedby intramitochondrial irondeposits. NatGenet 2001;27(2):181–6.

[11] Seznec H, Simon D, Bouton C, Reutenauer L, Hertzog A, Golik P, et al. Fiedreichataxia: the oxidative stress paradox. Hum Mol Genet 2005;14:463–74.

[12] Al-Mahdawi S, Pinto RM, Varshney D, Lawrence L, Lowrie MB, Hughes S, et al. GAArepeat expansionmutationmousemodels of Friedreich ataxia exhibit oxidative stressleading to progressive neuronal and cardiac pathology. Genomics 2006;88:580–90.

[13] Emond M, Lepage G, Vanasse M, Pandolfo M. Increased levels of plasmamalondialdehyde in Friedreich ataxia. Neurology 2000;55(11):1752–3.

[14] Pastore A, Tozzi G, Gaeta LM, Bertini E, Serafini V, Di Cesare S, et al. Actinglutathionylation increases in fibroblasts of patients with Friedreich's ataxia: apotential role in the pathogenesis of the disease. J Biol Chem2003;278(43):42588–95.

[15] Piemonte F, PastoreA, TozziG, TagliacozziD, Santorelli FM,CarrozzoR, et al. Glutathionein blood of patients with Friedreich's ataxia. Eur J Clin Invest 2001;31(11):1007–11.

118 M. Sparaco et al. / Journal of the Neurological Sciences 287 (2009) 111–118

[16] Schulz JB, Dehmer T, Schols L, Mende H, Hardt C, Vorgerd M, et al. Oxidative stressin patients with Friedreich ataxia. Neurology 2000;55(11):719–21.

[17] Tozzi G, Nuccetelli M, Lo Bello M, Bernardini S, Bellincampi L, Ballerini S, et al.Antioxidant enzymes in blood of patients with Friedreich's ataxia. Arch Dis Child2002;86(5):376–9.

[18] JauslinML,Vertuani S,Durini E, Buzzoni L, Ciliberti N,Verdecchia S, Palozza P,Meier T,Manfredini S. Protective effects of Fe-Aox29, a novel antioxidant derived from amolecular combination od idebenone and vitamin E, in immortalized fibroblasts andfibroblasts from patients with Friedreich ataxia. Mol Cell Biochem 2007;302(1-2):79–85.

[19] SeznecH, SimonD,Monassier L, Criqui-Filipe P, Gansmuller A, Rustin P, et al. Idebenonedelays the onset of cardiac functional alteration without correction of Fe–S enzymesdeficit in amousemodel for Friedreich's ataxia. HumMol Genet 2004;13(10):1017–24.

[20] Allani PK, Sum T, Bhansali SG, Mukherjee SK, Sonee M. A comparative study of theeffect of oxidative stress on the cytoskeleton in human cortical neurons. ToxicolAppl Pharmacol 2004;196(1):29–36.

[21] Neely MD, Sidell KR, Graham DG, Montine TJ. The lipid peroxidation product 4-hydroxynonenal inhibits neurite outgrowth, disrupts neuronal microtubules andmodifies cellular tubulin. J Neurochem 1999;72(6):2323–33.

[22] Hinshaw DB, Miller MT, Omann GM, Beals TF, Hyslop PA. A cellular model ofoxidant-mediated neuronal injury. Brain Res 1993;615(1):13–26.

[23] Shea TB, Zheng Y, Ortiz D, Pant HC. Cyclin-dependent kinase 5 increases perikaryalneurofilament phosphorylation and inhibits neurofilament axonal transport inresponse to oxidative stress. J Neurosci Res 2004;76(6):795–800.

[24] Harding AE. Friedreich's ataxia: a clinical and genetic study of 90 families with ananalysis of early diagnostic criteria and intrafamilial clustering of clinical features.Brain 1981;104(3):589–620.

[25] Santorelli FM, Tanji K, Shanske S, Krishna S, Schmidt RE, Greenwood RS, et al. Themitochondrial DNA A8344G mutation in Leigh syndrome revealed by analysis inparaffin-embedded sections: revisiting the past. Ann Neurol 1998;44(6):962–4.

[26] Filla A, De Michele G, Cavalcanti PianeseL, Monticelli A, Campanella G, Cocozza S.The relationship between trinucleotide (GAA) repeat length and clinical featuresin Friedreich ataxia. Am J Hum Genet 1996;59(3):554–60.

[27] Wang J, Tung YC, Wang Y, Li XT, Iqbal K, Grundke-Iqbal I. Hyperphosphorylationand accumulation of neurofilament proteins in Alzheimer disease brain and inokadaic acid-treated SY5Y cells. FEBS Lett 2001;507(1):81–7.

[28] SparacoM, Simonati A, Cavallaro T, Bartolomei L, GrausoM, Piscioli F, et al. MELAS:clinical phenotype and morphological brain abnormalities. Acta Neuropathol2003;106(3):202–12.

[29] Hsu SM, Raine L, Fanger H. A comparative study of the peroxidase–antiperoxidasemethod and an avidin–biotin complex method for studying polypeptide hormoneswith radioimmunoassay antibodies. Am J Clin Pathol 1981;75(5):734–8.

[30] Bradley JL, Blake JC, Chamberlain S, Thomas PK, Cooper JM, Schapira AH. Clinical,biochemical and molecular genetic correlations in Friedreich's ataxia. Hum MolGenet 2000;9(2):275–82.

[31] Rotig A, De Lonlay P, Chretien D, Foury F, Koenig M, Sidi D, et al. Aconitase andmitochondrial iron–sulphur protein deficiency in Friedreich ataxia. Nat Genet1997;17(2):215–7.

[32] Lynch DR, Lech G, Farmer JM, Balcer LJ, BankW, Chance B, et al. Near infraredmusclespectroscopy in patients with Friedreich's ataxia. Muscle Nerve 2002;25(5):664–73.

[33] VorgerdM, Schols L, Hardt C, RistowM, Epplen JT, Zange J. Mitochondrial impairmentof humanmuscle in Friedreich ataxia in vivo. Neuromuscul Disord 2000;10(6):430–5.

[34] Giustarini D, Rossi R, Milzani A, Colombo R, Dalle-Donne I. S-glutathionylation:from redox regulation of protein functions to human diseases. J Cell Mol Med2004;8(2):201–12.

[35] Sparaco M, Gaeta LM, Tozzi G, Bertini E, Pastore A, Simonati A, et al. Proteinglutathionylation inhumancentralnervous system:potential role in redoxregulationofneuronal defense against free radicals. J Neurosci Res 2006;83(2):256–63.

[36] Banan A, Zhang Y, Losurdo J, Keshavarzian A. Carbonylation and disassembly of theF-actin cytoskeleton in oxidant induced barrier dysfunction and its prevention byepidermal growth factor and transforming growth factor in human colonic cellline. Gut 2000;46(6):830–7.

[37] Collard JF, Cote F, Julien JP. Defective axonal transport in a transgenic mousemodelof amyotrophic lateral sclerosis. Nature 1995;375(6526):61–4.

[38] Aksenov MY, Aksenova MV, Butterfield DA, Geddes JW, Markesbery WR. Proteinoxidation in the brain in Alzheimer's disease. Neuroscience 2001;103(2):373–83.

[39] Contin MA, Arce CA. Tubulin carboxypeptidase/microtubules association can bedetected in the distal region of neural processes. Neurochem Res 2000;25(1):27–36.

[40] Morfini G, Pigino G, Mizuno N, Kikkawa M, Brady ST. Tau binding to microtubulesdoes not directly affect microtubule-based vesicle motility. J Neurosci Res2007;85:2620–30.

[41] Francis F, Roy S, Brady ST, Black MM. Transport of neurofilaments in growing axonsrequires microtubules but not actin filaments. J Neurosci Res 2005;79(4):442–50.

[42] Grant P, PantHC. Neurofilament protein synthesis and phosphorylation. J Neurocytol2000;29(11–12):843–72.

[43] Sternberger LA, Sternberger NH. Monoclonal antibodies distinguish phosphory-lated and nonphosphorylated forms of neurofilaments in situ. Proc Natl Acad SciUSA 1983;80(19):6126–30.

[44] Nguyen MD, Lariviere RC, Julien JP. Deregulation of Cdk5 in a mouse model of ALS:toxicity alleviated by perikaryal neurofilament inclusions. Neuron 2001;30(1):135–47.

[45] Sobue G, Hashizume Y, Yasuda T, Mukai E, Kumagai T, Mitsuma T, et al.Phosphorylated high molecular weight neurofilament protein in lower motorneurons in amyotrophic lateral sclerosis and other neurodegenerative diseasesinvolving ventral horn cells. Acta Neuropathol 1990;79(4):402–8.

[46] Sternberger NH, Sternberger LA, Ulrich J. Aberrant neurofilament phosphorylationin Alzheimer disease. Proc Natl Acad Sci USA 1985;82(12):4274–6.

[47] Wang J, BojaES, TanW,TekleE, FalesHM, English S, et al. Reversible glutathionylationregulates actin polymerization in A431 cells. J Biol Chem 2001;276(51):47763–6.

[48] Pollanen MS, Bergeron C, Weyer L. Characterization of a shared epitope in corticalLewy body fibrils and Alzheimer paired helical filaments. Acta Neuropathol1994;88(1):1–6.

[49] Roy S, Zhang B, Lee VM, Trojanowski JQ. Axonal transport defect: a common themein neurodegenerative diseases. Acta Neuropathol 2005;109(1):5–13.

[50] Gonzalez-Cabo P, Vazquez-Manrique RP, Garcia-Gimeno M, Sanz P, Palau F.Frataxin interacts functionally with mitochondrial electron transport chainproteins. Hum Mol Genet 2005;14(15):2091–8.

[51] Cohen G, Farooqui R, Kesler N. Parkinson disease: a new link between monoamineoxidase and mitochondrial electron flow. Proc Natl Acad Sci U S A 1997;94:4890–4.

[52] Bernard O, Balasubramanian KA. Effect of oxidized glutathione on intestinalmitochondria and brush border membrane. Int J Biochem Cell Biol 1995;27:589–95.