Differential expression of ceruloplasmin isoforms in the cerebrospinal fluid of amyotrophic lateral sclerosis patients

RESEARCH ARTICLE

Differential expression of ceruloplasmin isoforms

in the cerebrospinal fluid of amyotrophic lateral

sclerosis patients

Antonio Conti1, Sandro Iannaccone2, Barbara Sferrazza2, Lucia De Monte1, 3,Stefano Cappa2, 4, Diego Franciotta5, Stefano Olivieri1 and Massimo Alessio1

1 Proteome Biochemistry, San Raffaele Scientific Institute, Milan, Italy2 Department of Neurology, San Raffaele Scientific Institute, Milan, Italy3 Tumor Immunology, San Raffaele Scientific Institute, Milan, Italy4 Neurology, Vita-Salute San Raffaele University, Milan, Italy5 Neuroimmunology, Istituto Neurologico Mondino, Pavia, Italy

Amyotrophic lateral sclerosis (ALS) a fatal degenerative disease that selectively affects motorneurons, likely results from a complex interplay among oxidative injury, excitotoxic stimulation,protein aggregation and genetic factors. Ceruloplasmin (Cp) protein is a ferroxidase that oxidizestoxic ferrous iron to its nontoxic ferric form, protecting the central nervous system (CNS) fromiron deposition. Cp is thus considered as one of the main systems dedicated to the protection ofthe CNS from oxidative stress damage. We investigated Cp protein behaviour in the cere-brospinal fluid (CSF) of ALS patients of recent onset. An increased expression of Cp wasobserved in ALS (n = 16) compared to two control groups (healthy subjects, n = 11 and peripheralneuropathy patients, n = 10). 2-DE analysis revealed a differential expression of Cp isoforms inALS patients compared to controls. ALS samples showed an increase in the relative abundance ofmore basic Cp forms, corresponding to the nonsialylated proteins. Despite the increase in pro-tein expression, ferroxidase activity evaluated in the CSF of ALS patients was comparable to thatof the controls, indicating a Cp functional impairment. Ceruloplasmin isoforms profile may beproposed as disease feature that could provide insight into the molecular mechanisms of ALSpathogenesis.

Received: October 11, 2007Revised: April 22, 2008Accepted: June 9, 2008

Keywords:

Cerebrospinal fluid / Ceruloplasmin / Sialic acid / Two-dimensional gel electrophoresis

1628 Proteomics Clin. Appl. 2008, 2, 1628–1637

1 Introduction

Amyotrophic lateral sclerosis (ALS) [1, 2] is one of the mostcommon and severe neurodegenerative disorders. Its inci-dence varies little worldwide, affecting 1–2.5 cases per

100 000 population [2, 3]. ASL occurs primarily in adult life,with a median survival of 3–5 years. The disease is sporadicin 90% of cases and familial in about the 10%. The two formshave indistinguishable clinical and histopathological fea-tures, suggesting common molecular pathogenic mechan-isms [1, 2].

The 10–15% of familial form presents mutations in thegene coding for the antioxidant enzyme Cu, Zn-superoxidedismutase (SOD1), supporting the concept that oxidativestress plays a role in this disease [4, 5]. Oxidative stress is thecondition arising from an imbalance between the physiolog-ical production of potentially toxic ROS (i.e. superoxide,hydrogen peroxide and hydroxyl radical), and the scavengingactivities. These defence actions include enzymatic activities(as superoxide dismutase), low molecular weight antioxidant

Correspondence: Dr. Massimo Alessio, Proteome Biochemistry,DIBIT, San Raffaele Scientific Institute, Via Olgettina 58, 20132Milan, ItalyE-mail: [email protected]: 139-02-26434153

Abbreviations: ALS, amyotrophic lateral sclerosis; BBB, blood-brain barrier; CNS, central nervous system; Cp, ceruloplasmin;CSF, cerebrospinal fluid; GPI, glycosylphosphatidylinositol; SA,

sialic acid; WB, Western blot

DOI 10.1002/prca.200780081

© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.clinical.proteomics-journal.com

Proteomics Clin. Appl. 2008, 2, 1628–1637 1629

species (as vitamin E) and more complex form of protectionsuch as systems for metal transport [6]. In the last case, thecorrect handling of transition metal ions, as copper and iron,is crucial since these metals are able to undergo alternativeoxidation and reduction, during which toxic ROS are gener-ated.

One of these transporters, which seems to be involved inneurodegenerative disorders, is the copper–protein cer-uloplasmin (Cp), present in cerebrospinal fluid (CSF) [7–9].Cp acts mainly as ferroxidase, oxidizing toxic ferrous iron tonontoxic ferric form [7]. The redox reactions of iron are fun-damental, especially in the case of the central nervous system(CNS) that is particularly sensitive to this kind of damage[10]. The crucial role of Cp is demonstrated by patients withaceruloplasminemia, who show massive iron deposition inbrain [11]. The Cp expressed in the CSF and CNS likely playsan important role in iron homeostasis and antioxidantdefence. It is therefore, possible that an expression levelreduction or alteration of Cp function may contribute to theneurodegenerative process in ALS, by leading to an increasein ferrous iron, which can promote the generation of toxicfree radicals [9].

A definite diagnosis of ALS can be reached on the basis ofa combination of several clinical, physiological and bio-chemical observations. The heterogeneous and nonspecificnature of early symptoms in ALS, as well as the overlappingfeatures with a variety of neurological disorders result in di-agnostic difficulties.

The CSF, which is in direct contact with the CNS, con-tains proteins deriving from brain metabolism and has animportant diagnostic role in humans [12]. One of the majorchallenges in conducting an analysis of human nervous sys-tem pathologies is to obtain appropriate samples frompatients and controls. For this reason CSF, available as a

consequence of patient monitoring, is the tissue sample ofreference for investigation in humans. CSF analysis has beensuccessfully used for the characterization of different physio/pathological conditions [13–16].

In order to investigate different expression level or spe-cific modifications of Cp occurring in ALS, we comparedCSF samples from patients versus healthy subjects and versuspatients with peripheral neuropathy.

2 Materials and methods

2.1 Patients

Following approval by the hospital’s institutional ethicalreview board and informed patient consent, CSF samples(0.8–1 mL) were collected from patients and controls (fea-tures described in Table 1). We analysed CSF from twogroups: patients with sporadic ALS (n = 16) and healthycontrols (CN, n = 11). The CSF of a third group of patientsaffected by peripheral neuropathy (PN n = 10) has been usedas a further control group of neurological patients affected byan unrelated slow progressive neuronal disease (Table 1).Samples of CSF from PN patients were from a previousstudy [17]. Due to the limited amount of protein obtained insome samples, not all the assays have been performed withall subjects. The exact number of subjects used in the differ-ent experiments are reported in the Section 3 and in the fig-ure legends.

Patient eligibility was based on El Escorial criteria [18].All patients were at first diagnosis and drug free. Quantita-tive evaluation for disability was calculated according to astandardized procedure [19]. All patients underwent the nor-mal procedure for differential diagnosis, including haema-

Table 1. Features of patient and control subjects

ALS CN PN

Sex Age [C] (mg/mL) Sex Age [C] (mg/mL) Sex Age [C] (mg/mL)

1 M 23 0.41 F 19 0.08 M 35 0.512 M 56 0.25 F 50 0.20 M 38 0.473 M 50 0.54 F 30 0.15 M 46 0.454 F 40 0.10 M 71 0.38 M 60 0.335 F 70 0.46 F 30 0.30 M 50 0.586 F 30 0.33 F 53 0.25 F 64 0.327 F 50 0.27 M 70 0.37 F 50 0.238 M 64 0.27 F 30 0.15 F 60 0.359 M 52 0.55 F 25 0.20 F 41 0.47

10 M 50 0.63 F 34 0.15 M 64 0.3911 M 61 0.32 M 45 0.30 – – –12 M 63 0.28 – – – – – –13 F 62 0.27 – – – – – –14 M 71 0.66 – – – – – –15 M 39 0.30 – – – – – –16 F 70 0.40 – – – – – –


1630 A. Conti et al. Proteomics Clin. Appl. 2008, 2, 1628–1637

tological examination (serum immunoelectrophoresis, vita-min B12, folic acid, rheumatoid factor, immunocomplexes,cryoglobulins, extractable nuclear antigen, antimitochon-dria, antismooth muscle and antiDNA antibodies), electro-neurographic (ENG) and electromyographic (EMG) evalua-tions based on standard criteria [20], and lumbar CSF collec-tion and examination. At neurophysiological investigation allpatients showed neurogenic changes and no motor or sen-sory conduction abnormalities (as revealed by EMG andENG studies, respectively). The disease course was chroni-cally progressive in all cases. Patients were submitted toneurological examination using strength evaluation (MRCgrading system) performed at first hospitalisation and 6 and12 months later, by the same neurologist. The clinical diag-nosis of ALS was made on the basis of ‘World Federation ofNeurology criteria for diagnosis of ALS’.

CSF was collected by lumbar puncture as a normal pro-cedure for diagnostic purpose; for some patients (n = 5) andcontrols (n = 6), the peripheral blood was also collected, cen-trifuged in order to eliminate cells and freshly used or storedat 2807C after acetone protein precipitation.

Exclusion criteria included: HIV or HCV positivity; oth-ers neurodegenerative diseases or previous cerebral ischemicevents; severe metabolic disorders (i.e. diabetes). Medicalhistory, neurological examination and magnetic resonanceimaging studies in all patients excluded spinal cord andnerve root injury.

Control CSF was obtained from patients with suspectedneurological disease that resulted normal and free frompathological alterations after testing analysis and clinicalstudies. The control subjects were selected in order to be ho-mogeneous to ALS patients by age and gender distributions.

2.2 2-DE

CSF samples immediately after collection from patients werecentrifuged for 10 min at 5006g at 147C, in order to elim-inate cells and supernatants. Total protein concentration wasdetermined by Bradford method using the Protein Assayreagent (BioRad, Hercules, CA, USA), and sample immedi-ately processed or aliquots of 100 mg were acetone pre-cipitated and stored at 2807C. Proteins extracted from CSFwere resuspended in a buffer containing Urea (8 M),CHAPS (4% w/v), DTT (65 mM), 0.2% v/v IPG-buffer 3–10 NL (GE Healthcare, Milan, Italy) and 0.05% Bromophe-nol blue. Samples (30 mg each) were applied to 7 cm IPGstrips pH 3–10 NL (GE Healthcare) by in-gel rehydration at207C. Focusing was performed with an IPGphor system (GEHealthcare) at 50 mA max per IPG strip with a gradient volt-age (5000 V max) for a total of 30 k?Vh. Strips were equili-brated for 15 min in 50 mM pH 8.8 Tris-HCl buffer contain-ing 6 M urea, 30% glycerol, 2% SDS and 2% DTT, then for15 min in the same buffer replacing DTT by 2.5% iodoacet-amide. The strips were then transferred onto 10% acryl-amide SDS-PAGE gels for the second dimension separation.

2.3 Western blot (WB) and Image analysis

CSF or plasma proteins resolved by 2-DE or by 10% acryl-amide SDS-PAGE gel, were electrontransferred onto NCmembranes; after incubation with blocking reagent mem-branes were incubated with sheep polyclonal antihuman Cp(Abcam, Cambridge, UK) or with Rabbit polyclonal antihu-man Cp (Santa Cruz Biotechnology, Santa Cruz, CA, USA).Immunoreactivity was revealed with rabbit polyclonal anti-sheep IgG HRP conjugated (Abcam) or swine antirabbit IgGHRP (DakoCytomation, Glostrup, DK), followed by ECLreaction and film exposure.

Images were acquired using a Personal SI Laser Densi-tometer (Molecular Dynamics, Sunnyvale, CA, USA). Theevaluation of relative abundance of Cp isoforms was done byanalysing OD normalized to percentage using ProgenesisPG240 v.2005 software (Nonlinear Dynamics, Newcastle,UK). Apparent relative molecular mass (Mr) were estimatedby comparison with molecular weight (MW) referencemarkers (Precision, BioRad) and pI values assigned todetected spots by calibration as described in the GE Health-care guide lines.

2.4 Protein deglycosylation

CSF proteins (100 mg) from one individual representativeALS patient, were denatured by adding 0.5% SDS, and 1% b-mercaptoethanol, boiled for 10 min and allowed to coolbefore the addition of either 106digestion buffer (final con-centration 50 mM phosphate buffer, 1% NP40) and 4250Units of PNGase F (New England Biolabs, Boston, MA,USA), or 106digestion buffer (final concentration 50 mMsodium citrate) and 5000 Units of Endo H (New EnglandBiolabs), respectively. The samples were incubated at 377Cfor 1 h and processed as for anti-Cp WB.

2.5 Neuroaminidase treatment

Fifty microgram of CSF proteins either from one individualrepresentative control subject or one ALS patient, wereincubated with 40 mU of Vibrio cholerae Neuraminidase(Roche Diagnostics, Milan, Italy) in 50 mM sodium acetatebuffer (pH 5.5) containing 0.5% SDS, 1% b-mercaptoetha-nol, 1% NP-40 for 18 h at 377C. Control samples were treatedin the same way but without adding the enzymes. Incuba-tions were stopped by adding 100% icecold acetone and thensamples processed as described above to run 2-DE and WB.

2.6 Lectins reactivity

CSF proteins resolved by 2-DE and transferred to NC mem-brane, obtained as in point Section 2.5, were incubated for 1h at room temperature with Sambucus nigra (SNA) andMaackia amurensis (MAA) digoxigenin-labelled agglutininsin combination (Roche), in order to detect sialic acid (SA).After rinsing to remove lectins surplus, membrane was



incubated with polyclonal rabbit anti-DIG/HRP F(ab’)(Da-koCytomation). After further rinsing, reactivity was revealedby ECL (GE Healthcare) reaction and film exposures.

2.7 Analysis of cell surface anchored Cp protein

Cell surface expression of Cp was assessed by cyto-fluorimeter analysis (FACScan™, Becton Dickinson, SanJose, CA, USA) performed as described in ref. [21] with spe-cific antibodies (W6/32, mouse mAb anti-MHC Class I,Abcam; rabbit polyclonal anti-Cp, Santa Cruz Biotechnology)on T98G glioblastoma and U87 MG glioblastoma-astro-cytoma cell lines (America Type Culture Collection, Mana-ssas, VA, USA). Protein expression on whole cell lines lysates(10 and 30 mg loaded) was also tested by 10% acrylamideSDS-PAGE and WB. Cell surface anchored Cp profile fromT98G glioblastoma cells was assessed by 2-DE as describedbefore, and CSF from a representative ALS patient was usedas comparison pattern.

2.8 Bathophenanthroline assay

To analyse ferroxidase activity in CSF we performed a Bath-ophenanthroline assay in which we estimate the oxidation ofFe(II) using a colourimetric method. The CSF samples (20 or60 mg) were incubated with 60 mM of FeCl3 (ferrous form)and analysed at five different times (0, 30, 60, 90, 120 min)using a solution of 1 mM of Bathophenanthroline (Sigma–Aldrich, St. Louis, MO, USA) in acetate buffer (pH 6.2). Thedecrease of absorbance at 535 nm of bathophenanthroline–Fe(II) complex is due to the ferrous iron oxidation in ferricform (FeIII) mainly catalysed by Cp [22].

2.9 Statistical analysis

Anti-Cp reactivity obtained with the Abcam polyclonal anti-body was quantified by densitometric analysis (MolecularDynamics Personal SI Laser Densitometer), normalized byprotein loading and reference proteins staining, and com-pared in patients and controls. The distribution of the differ-ent Cp isoforms was compared in patients and controls bydensitometric analysis of 2-D spots OD (indicated as spotvolume) normalized as percentage of the total anti-Cp iso-forms antibody reactivity. Statistical analysis of categoricaldata (gender distribution) was performed by 262 con-tingency table analysis using Fisher’s exact test and two-tailed p value. Continuous data (age distribution and spot/band volume) were evaluated using unpaired Student’s t-testwith two-tailed p value for the comparison of two means withstandard error if the data passed the normality test for Gaus-sian distribution assessed by Kolmogorov–Smirnov, D’Agos-tino and Pearson and Shapiro–Wilk tests. If data were lim-ited or did not show normal distribution, statistical analysiswas evaluated using the nonparametric Mann–Whitney test.Parametric one-way analysis of variance (ANOVA) and non-parametric Kruskal–Wallis analysis of variance test were also

used to evaluate the statistical difference among three ormore independent groups, including postanalysis performedwith Tukey’s and Dunnetts’s Multiple Comparison Tests forGaussian and nonGaussian distributed data, respectively.

In all analysis, p,0.05 was considered to be statisticallysignificant. Statistical analysis was performed using Graph-Pad Prism 4, V4.03 software (GraphPad, SanDiego, CA,USA).

3 Results

3.1 Ceruloplasmin expression is increased in the CSF

of ALS patients

Table 1 shows the features of patients and control subjects.Age and gender distribution between ALS patients and con-trols did not show differences statistically significant.

To evaluate the Cp expression level in CSF, SDS-PAGEand WB were performed (Fig. 1A). Due to the variability inCSFs protein concentration (see Table 1) we loaded the sameamount of proteins (100 mg) for each sample. Densitometricanalysis showed that Cp expression level in ALS patients wasabout 2.3-fold higher than either CN or PN alone or thangrouped (CN 1 PN) controls (Fig. 1B). Differences were sig-nificant by Mann–Whitney test (ALS versus CN p = 0.0152;ALS versus PN p = 0.0287; ALS versus CN 1 PN p = 0.0097)and by Kruskal–Wallis one-way ANOVA test (p = 0.0207),with postanalysis test statistically significant for ALS versusCN comparison (p,0.05).

Since blood is the reference tissue for diagnostic pur-poses, and Cp is also present in plasma, we investigatedwhether the Cp level was also increased in the plasma of ALSpatients. As shown in Figs. 1C and D, no differences wereobserved in plasma Cp expression between controls and ALSpatients (p = 0.62, ALS n = 5, CN n = 6).

3.2 Cp isoforms show differential expression in ALS

Due to different PTMs several Cp isoforms are present bothin plasma and CSF [7]. In order to investigate whether dif-ferent isoforms occur specifically in the ALS patients, sam-ples were analysed by 2-DE WB with specific antiCp anti-body.

In the CSF, the reactivity of the Cp-specific antibodyidentified a pattern of four major isoforms, showing thesame Mr but different pI (Fig. 2A, spots A–D). Out of a dif-ferent amount of total Cp present in the CSF of controls andALS patients (see Fig. 1), the densitometric analysis of therelative abundance of the Cp isoforms calculated as percent-age of the total anti-Cp reactivity in each experiment, showeda significant increase of the isoform D in the ALS patientscompared to healthy subjects (p = 0.0224). A similar increasewas also observed if compared to peripheral neuropathypatients (p = 0.0426) (Fig. 2B). A statistical significant resultwas obtained also when the three groups were evaluated



Figure 1. Ceruloplasmin espression level in CSF and plasma. (A) WB analysis with anti-Cp antibody performed on SDS-PAGE-resolvedproteins from CSF of ALS patients, healthy controls (CN), peripheral neuropathy patients (PN) and 10 ng of purified Cp (Alexis Biochem-icals, San Diego, CA). (B) Densitometric analysis of the WB reactivity expressed as arbitrary unit of OD normalized by amount of proteinloaded and by the OD of 10 ng of purified Cp (ALS; n = 11) (CN; n = 11) (PN; n = 8). The bar histograms represent the mean of the normalizedbands volumes with standard errors (SE). (C) WB analysis with antiCp antibody performed on SDS-PAGE-resolved proteins from plasma ofpatients (ALS) or controls (CN) and from 10 mg of control CSF. (D) Densitometric analysis of the WB reactivity expressed as arbitrary unit ofOD (ALS; n = 5) (CN; n = 6). The bar histograms represent the mean of the normalized bands volumes with SE. Data were analysed byMann–Whitney test (* = p,0.05; ** = p,0.01), and by Kruskal–Wallis one-way ANOVA (K–W ANOVA) test (* = p,0.0207).

Figure 2. 2-DE profile of Cp in CSF. Proteins from CSF were separated by pI in the first dimension on a pH 3–10 nonlinear range, and byrelative mass in the second dimension on 10% acrylamide SDS-PAGE. Ceruloplasmin spots reactivity was revealed by WB with specificantibody (WB: antiCp). (A) Comparison between 2-DE Cp profile in CSF from representative healthy subject (CN), ALS and PN patients.Enlargements of the 2-DE area of antiCp reactivity are presented. The Cp pattern includes four major spots named A, B, C, D. (B) Densito-metric analysis of the relative abundance of the Cp isoforms calculated as percentage of the total anti-Cp reactivity. Bar histograms repre-sent the mean of the spot volumes normalized as percentage of total reactivity with standard errors (ALS, n = 16; CN, n = 10; PN, n = 6); datawere analysed by Student’s t-test (* = p,0.05), and by one-way ANOVA (p = 0.019). C) Comparison of spot D relative abundance betweengrouped CN and PN (n = 16) versus ALS (n = 16), data were analysed by Student’s t-test (** = p,0.005).

together by one-way ANOVA test (p = 0.0169), with post-analysis test significant for both ALS versus CN and ALS versusPN comparisons (p,0.05) (Fig. 2B). Since the trend of spot Dwas similar in CN and PN, in order to increase the number ofsubjects we considered the two homogeneous groups togeth-er (CN 1 PN versus ALS). Comparing these two groups, the

relative abundance increase of spot D expression in ALSpatients was highly significant (p,0.0043) (Fig. 2C).

The protein spots identity was confirmed both by reac-tivity with a different antiCp antibody (data not shown) andby MALDI-TOF MS analysis (see Table 1 of SupportingInformation).



Figure 3. 2-DE profile of Cp in plasma from controls (CN) or ALSpatients. Proteins were separated by pI in the first dimension on apH 3–10 nonlinear range, and by relative mass in the second di-mension on a 10% acrylamide SDS-PAGE. Ceruloplasmin spotsreactivity was revealed by WB with specific antibody. For com-parison, Cp isoforms (A, B, C, D) profile in the CSF of ALS patientis also reported (upper panel).

3.3 Plasma Cp is similar in controls and ALS patients

2-D analysis of Cp from plasma showed a electrophoreticmigration pattern different from that one observed in CSF,as the spot B was the largest isoform present in plasma(Fig. 3). Noteworthy, the pattern was identical in both controland ALS patients, confirming that Cp differential expressionis a feature of CSF Cp.

3.4 GPI-membrane anchored Cp isoform is not

responsible for differential expression

Then, we asked which might be the differences betweenplasma and CSF Cp. Soluble Cp present in plasma is a pro-tein of prevalent hepatic origin, while the CSF form issecreted by cells of the choroid plexus. These two forms areseparated by blood–brain barrier (BBB) that is not crossed innormal conditions [7, 23]. Moreover, there is a membrane-bound glycosylphosphatidylinositol (GPI) anchored form ofCp that is expressed by astrocytes and leptomeningeal cellsin the CNS [23, 24]. Since GPI-anchored proteins can bereleased in a regulated way [25], the hypothesis that the neu-rodegeneration process could release this protein in the CSFwas taken into consideration. Thus the expression of mem-brane-anchored Cp was investigated on glioma and astro-cytoma cell lines. We found that a fraction of T98G glio-blastoma cell line express at low level the Cp on the cell sur-face, while the U87 MG cell line does not (Figs. 4A and B).2-DE analysis revealed that the electrophoretic migrationpattern of membrane Cp is more acidic than that oneobserved for Cp from the CSF (Fig. 4C). This indicates thatthe Cp isoform differentially expressed in the CSF of ALSpatients has not been originated from the release of the GPI-membrane anchored isoform.

3.5 The differentially expressed isoform corresponds

to nonsialilated Cp

Since Cp is a glycoprotein, we investigated whether differentisoforms corresponded to the presence of different glycosidicresidues. Removal of glycosidic residues showed that Cp

Figure 4. Analysis of membrane-anchored Cp on glioblastoma (T98G) and glioblastoma-astrocytoma (U87 MG) cell lines. (A) Flow cyto-metry analysis of surface membrane Cp expression. Solid lines indicate the staining with W6/32 anti-MHC class I antibody (MHC cl I, leftpanels) or anti-Cp (Cp, right panels) revealed by FITC-conjugated secondary antibodies; dashed lines indicate the control staining obtainedwith irrelevant Ig plus secondary antibodies. Arrow indicates the subpopulation resulted Cp positive. (B) WB detection of Cp in cell linestotal lysates (10–30 mg) and in CSF (10 mg total proteins) from a representative ALS patient. (C) 2-DE profile of membrane Cp in T98G cellline. Proteins were separated by pI in the first dimension on a pH 3–10 nonlinear range, and by relative mass in the second dimension on a10% acrylamide SDS-PAGE. Cp spots reactivity was revealed by WB with specific antibody and compared to the profile of secreted Cp(isoform A, B, C, D) present in the CSF of a representative individual ALS patient.



apparent Mr was decreased when CSF was treated withPNGase-F, but not when treated with Endo-H (Fig. 5A). Thisdemonstrated that Cp containing complex oligosaccharidesis released in the CSF following the secretory pathway. Start-ing from the observation that Cp isoforms are different in pIbut have the same molecular weight (Fig. 2A), we investi-

Figure 5. Ceruloplasmin deglycosylation, SA removal and lectinsreactivity. (A) CSF proteins from a representative individual ALSpatient, were treated either with PNGase-F or Endo-H in order toremove N-linked complex oligosaccharides or high mannoseoligosaccharides, respectively. Proteins were resolved by SDS-PAGE and probed by WB with antiCp antibody. (B) CSF proteinseither from a representative individual control subject (CN) orfrom a representative ALS patient, were resolved by 2-DE on apH 3–10 nonlinear range and in the second dimension, by relativemass on a 10% acrylamide SDS-PAGE. 2-DE analysis of CSF wasalso performed after treatment with neuraminidase in order toremove SA residues (Neu, 1). Cp isoforms profile (brackets andA, B, C, D) was revealed by WB with specific antibody (anti-Cp).Due to the partial digestion of neuraminidase, some intermediateisoforms were generated (A0 and B0 in ALS or streak between Cand D spots in CN). The presence of SA on the Cp isoforms wasevaluate by the specific reactivity of two lectins S. nigra aggluti-nin (SNA) and M. amurensis agglutinin (MAA).

gated whether SA residues at the end of mature glycosidicbranches might be responsible for these differences. SAremoval from both control and ALS CSF proteins resulted ina shift of Cp isoforms toward more basic pH, with anincreased relative abundance of spots C and D (Fig. 5B). Dueto the partial digestion of neuraminidase, some intermediateisoforms were generated (Figs. 5B, A0 and B0 in ALS or streakbetween spots C and D in CN). This suggested that spots Cand D might correspond to nonsialilated proteins. The pres-ence of SA in CSF glycoproteins was evaluated by using thespecific reactivity of two lectins SNA and MAA. A weak lec-tins reactivity was present in both control and ALS CSFs incorrespondence with Cp spots A and B (Fig. 5B). In contrast,no reactivity was observed in the more basic pH Cp isoforms,in spite of protein abundance in particular in the case of ALS(Fig. 5B). As already reported [26], we observed a reduction ofabout 20% of SA contents in CSF-proteins from ALS com-pared to healthy subject as inferred by total spots OD of lec-tins reactivity on the whole CSF proteome resolved by 2-DE(not shown).

3.6 Cp from ALS patients shows reduced ferroxidase

activity

Finally, we analysed the ferroxidase activity present in theCSF that mainly corresponds to Cp functionality [22]. De-spite the 2.2-fold increased amount of Cp present in the CSFof ALS patients, no difference in ferroxidase acivity wasdetected between ALS patients (n = 9) and CN group (n = 8)(Fig. 6). This suggests a functional impairment of the Cppresent in the CSF from ALS patients. The functional weak-ness of Cp in ALS is confirmed by the observation that in athree-fold amount of control CSF (60 mg, containing a quan-tity of Cp similar to that present in ALS liquor) a proportionalincrease in ferroxidase activity was recorded (Fig. 6). Inter-estingly, even if performed as single experiment due to thelimited quantity of material, a three-fold amount of CSFfrom a pool of ALS patients, showed an increase in the fer-roxidase activity that was not proportional to the amount ofCp present, thus confirming the functional impairment(Fig. 6).

4 Discussion

ALS is a devastating neurological illness, and the limitedknowledge about the pathogenetic mechanisms prevents toestablish effective therapeutic protocols. We analysed theexpression level and protein profile modification of Cp fer-roxidase in sporadic ALS; being this protein functionallyinvolved in the reduction of oxidative stress damage, serumCp-concentration has been in the past studied as a possiblemarker of ALS, without relevant findings and further inves-tigations [8]. We found an increase of Cp protein expressionlevel in the CSF of ALS patients, confirming the importanceof looking for markers in a biological fluid contiguous to the



Figure 6. Ceruloplasmin ferroxidase activity. Bathophenanthroline assay was performed to analyse the ferroxidase activity in CSF, mainlycatalysed by Cp. The oxidation of Fe21 is evaluated by colourimetric assay in which the decreased absorbance at 535 nm of bath-ophenanthroline-Fe21 complex reflects the oxidation to ferric Fe31 form. Curves reported the concentration reduction of Fe21 along time (0,30, 60, 120, 180 min) in the presence of: buffer alone (CTR); CSF from ALS patients (n = 9); CSF from healthy subjects (CN, n = 8); 36amountof CSF from a pool of healthy subjects (CN 63) containing an amount of Cp similar to that present in the ALS liquor; and a 36amount ofCSF from a pool of ALS patient (ALS 63). B) Linear regression of the bathophenanthroline–Fe21 complex OD at 535 nm and Fe21 micro-molar concentration.

damaged tissue. This might reflect the attempt of the organ-ism to respond to a pathological increase of oxidative stressby inducing proteins with protective functions. Supportingthis hypothesis is the fact that an increase of Cp transcripthas been observed as consequence of experimental oxidativestress in rat hippocampus [27]. This mechanism mightrepresent a redox direct effect on the Cp promoter, or mightbe the results of the gene regulation induced by the interplaybetween regulating elements, like Hypoxia-inducible factorand Activator Protein 1 present in the Cp promoter, thatserve as sensory system for oxygen, iron and copper metab-olisms [28–30].

Interestingly, besides the changes in total expressionlevel, we found a different pattern of Cp isoforms, being therelative abundance of the more basic spot increased in ALSpatients. This may results from one of the different followinghypothetical mechanisms, like (i) direct synthesis of thebasic isoform as Cp precursor, (ii) interconversion of themore acidic isoforms (A, B, C) to D isoform, (iii) release ofthe membrane anchored Cp isoform, (iv) BBB leakage withcrossing of a plasma-specific Cp isoform. The last hypothet-ical mechanism is disclaimed by the results showing a plas-matic Cp 2-DE pattern different from CSF, in which the Disoform is not present. This seems to reflect a CSF specificlocal change, induced by modifications related to pathologi-cal mechanisms. Moreover, the plasma Cp pattern is iden-tical in control subjects and ALS patients, indicating that theBBB does not show gross leakage, at least, in patients at theearly stage of disease.

A different, interesting hypothesis was the possibilitythat the GPI-membrane anchored Cp, which is specificallyexpressed by astrocytes and leptomeningeal cells, wasreleased in the CSF as consequence of the pathology. Thedifferent 2-DE pattern excluded that the D isoform increasein ALS was the results of Cp membrane anchored release.Also the coincidence of the Cp D isoform with an immaturesynthesis precursor can be excluded by the results of glyco-sidase treatments, showing that Cp is released in the CSFfollowing secretory pathway only as glycosidic-mature pro-tein.

Thus, the more likely hypothesis is the isoforms inter-conversion, hence, it was important to identify the biochem-ical features of the D isoform. The evidence that SA removalfrom both control and ALS CSFs results in the accumulationof the signal to the basic D isoform, combined with theabsence of SA-specific lectins reactivity, strongly suggeststhat isoform D of Cp is an asialo-protein.

Interestingly, a general reduction in the content of SA inCSF glycoprotein has been reported in ALS patients [26] andconfirmed in this study by reduction of lectins reactivity inALS patients. Whether the SA release is a regulated phe-nomena or is a side effect is not known; however, it has beenreported that serum Cp is actively desialylated upon bindingto liver endothelial cells in order to be uptaken by hepato-cytes [31, 32]. Moreover, as shown for transferrin [33], trans-port from plasma to CSF upon active desialylation operatedby ependymal cells of choroid plexus, can also be hypothe-sized for Cp.



SA removal from Cp may have a detrimental effect inALS, because changes in carbohydrate structure may havefunctional implications with regards to protein turnover orenzymatic activity of this oxidation-protective protein. In fact,while it is not clear whether ferroxidase activity is affected inasialo-Cp [34], it is clear that the asialo-Cp is quickly removedfrom circulation by liver uptake [35, 36]. This may explain theinduction of Cp synthesis as a balancing effect, aimed tocounteract the increased removal of asialo-Cp.

The Cp expression increase we observed in the ALSpatients may be the necessary response of the patients tomaintain a ferroxidase activity comparable to that recordedin healthy subjects, in which Cp protein is about two-foldlower. This seems to be sufficient in the early stage of ALSdisease to maintain the protective activity of Cp at physio-logical level. It would be interesting to investigate whetherthe ferroxidase activity in the CSF is further reduced alongthe disease progression. However, if the reduced Cp func-tionality is due to the asialo-Cp needs to be demonstrated, infact no direct correlation has been observed at the momentbetween Cp functionality and spot D expression level. Cpfunction impairment and asialo-Cp isoform increase mayresult from the difficulties that CNS cells, including theependymal of corhoid plexus, have to counteract oxidativestress. Such CSF abnormalities could mirror the putativederangement of anti-oxidative stress response occurring atmotor neuronal level. However, the reduction of ferroxidaseactivity, might directly contribute to pathology progression.

In humans the only proven therapy of ALS is Riluzole,which extends survival by approximately 3 months [1, 37]. Wehave had the chance to analyse also Cp isoforms pattern in theCSF from three ALS patients that were under Riluzole treat-ment (3–6 months). The observation that the Cp pattern ofpatients under therapy was similar to that of the untreatedones (data not shown), suggest that the drug does not affect thepathological mechanisms underlined by Cp modifications.This intriguing observation needs further investigations.

That Cp pattern is a feature of ALS not present in anunrelated neurological disease is promising for differentialdiagnosis. In fact, clinically, the motor peripheral neuropathy(resulting from degeneration of the motor nerves) and ALS(in which the spinal motor neurons are involved) are bothassociated with progressive weakness, wasting and fascicu-lations, but in the absence of clinical pyramidal signs, thetwo may be difficult to differentiate at the early stages.

Whether Cp pathological features found in ALS (i)expression level increase, (ii) modified electrophoretic pat-tern and (iii) SA contents reduction, are the result of a dis-ease-specific or process-specific phenomena shared by dif-ferent neurodegenerative diseases (i.e. Parkinson, Alzheimeror Huntington diseases) will be the topic of further study.

This work was supported by Fondazione CARIPLO(NOBEL GuARD Project; 2006.0537/105411 project; and2006.0538 project), Regione Piemonte-Ricerca sanitaria fina-lizzata 2006 and RF-F5R-2007-637144.

The authors have declared no conflict of interest.

5 References

[1] Rowland, L. P., Shneider, N. A., Amyotrophic lateral sclero-sis. N. Engl. J. Med. 2001, 344, 1688–1700.

[2] Mitchell, J. D., Borasio, G. D., Amyotrophic lateral sclerosis.Lancet 2007, 369, 2031–2041.

[3] Jackson, C. E., Bryan, W. W., Amyotrophic lateral sclerosis.Semin. Neurol. 1998, 18, 27–39.

[4] Cleveland, D. W., Rothstein, J. D., From Charcot to Lou Geh-rig: Deciphering selective motor neuron death in ALS. Nat.Rev. Neurosci. 2001, 2, 806–819.

[5] Pasinelli, P., Brown, R. H., Molecular biology of amyotrophiclateral sclerosis: Insights from genetics. Nat. Rev. Neurosci.2006, 7, 710–723.

[6] Carri, M. T., Ferri, A., Cozzolino, M., Calabrese, L., Rotilio, G.,Neurodegeneration in amyotrophic lateral sclerosis: Therole of oxidative stress and altered homeostasis of metals.Brain. Res. Bull. 2003, 61, 365–374.

[7] Hellman, N. E., Gitlin, J. D., Ceruloplasmin metabolism andfunction. Annu. Rev. Nutr. 2002, 22, 439–458.

[8] Domzal, T., Radzikowska, B., Ceruloplasmin and copper inthe serum of patients with amyotrophic lateral sclerosis(ALS). Neurol. Neurochir. Pol. 1983, 17, 343–346.

[9] Patel, B. N., Dunn, R. J., Jeong, S. Y., Zhu, Q. et al., Cer-uloplasmin regulates iron levels in the CNS and preventsfree radical injury. J. Neurosci. 2002, 22, 6578–6586.

[10] Ke, Y., Ming Qian, Z., Iron misregulation in the brain: A pri-mary cause of neurodegenerative disorders. Lancet Neurol.2003, 2, 246–253.

[11] Harris, Z. L., Klomp, L. W., Gitlin, J. D., Aceruloplasminemia:An inherited neurodegenerative disease with impairment ofiron homeostasis. Am. J. Clin. Nutr. 1998, 67, 972S–977S.

[12] Ranganathan, S., Williams, E., Ganchev, P., Gopalakrishnan,V. et al., Proteomic profiling of cerebrospinal fluid identifiesbiomarkers for amyotrophic lateral sclerosis. J. Neurochem.2005, 95, 1461–1471.

[13] Hochstrasser, D. F., Sanchez, J. C., Appel, R. D., Proteomicsand its trends facing nature’s complexity. Proteomics 2002,2, 807–812.

[14] Rohlff, C., Southan, C., Proteomic approaches to centralnervous system disorders. Curr. Opin. Mol. Ther. 2002, 4,251–258.

[15] Davidsson, P., Folkesson, S., Christiansson, M., Lindbjer, M.et al., Identification of proteins in human cerebrospinal fluidusing liquid-phase isoelectric focusing as a prefractionationstep followed by two-dimensional gel electrophoresis andmatrix-assisted laser desorption/ionisation mass spectrom-etry. Rapid Commun. Mass Spectrom. 2002, 16, 2083–2088.

[16] Conti, A., Sanchez-Ruiz, Y., Bachi, A., Beretta, L. et al., Pro-teome study of human cerebrospinal fluid following trau-matic brain injury indicates fibrin(ogen) degradation prod-ucts as trauma-associated markers. J. Neurotrauma 2004,21, 854–863.

[17] Conti, A., Ricchiuto, P., Iannaccone, S., Sferrazza, B. et al.,Pigment epithelium-derived factor is differentially expres-sed in peripheral neuropathies. Proteomics 2005, 5, 4558–4567.



[18] Brooks, B. R., El Escorial World Federation of Neurology cri-teria for the diagnosis of amyotrophic lateral sclerosis. Sub-committee on Motor Neuron Diseases/Amyotrophic LateralSclerosis of the World Federation of Neurology ResearchGroup on Neuromuscular Diseases and the El Escorial“Clinical limits of amyotrophic lateral sclerosis” workshopcontributors. J. Neurol. Sci. 1994, 124, 96–107.

[19] Appel, V., Stewart, S. S., Smith, G., Appel, S. H., A ratingscale for amyotrophic lateral sclerosis: description and pre-liminary experience. Ann. Neurol. 1987, 22, 328–333.

[20] Kimura, J., in: Kimura, J. (Ed.), Electrodiagnosis in Diseasesof Nerve and Muscles, Davids FA, Philadelphia, PA 1989, pp.227–249.

[21] Alessio, M., De Monte, L., Scirea, A., Gruarin, P. et al., Syn-thesis, processing, and intracellular transport of CD36 dur-ing monocytic differentiation. J. Biol. Chem. 1996, 271,1770–1775.

[22] Vassiliev, V., Harris, Z. L., Zatta, P., Ceruloplasmin in neuro-degenerative diseases. Brain Res. Brain Res. Rev. 2005, 49,633–640.

[23] Mittal, B., Doroudchi, M. M., Jeong, S. Y., Patel, B. N., David,S., Expression of a membrane-bound form of the ferrox-idase ceruloplasmin by leptomeningeal cells. Glia 2003, 41,337–346.

[24] Patel, B. N., David, S., A novel glycosylphosphatidylinositol-anchored form of ceruloplasmin is expressed by mamma-lian astrocytes. J. Biol. Chem. 1997, 272, 20185–20190.

[25] Kondoh, G., Tojo, H., Nakatani, Y., Komazawa, N. et al.,Angiotensin-converting enzyme is a GPI-anchored proteinreleasing factor crucial for fertilization. Nat. Med. 2005, 11,160–166.

[26] Niebroj-Dobosz, I., Janik, P., Jamrozik, Z., Kwiecinski, H.,Immunochemical quantification of glycoconjugates inserum and cerebrospinal fluid of amyotrophic lateralsclerosis patients. Eur. J. Neurol. 1999, 6, 335–340.

[27] Wang, X., Pal, R., Chen, X. W., Kumar, K. N. et al., Genome-wide transcriptome profiling of region-specific vulnerabilityto oxidative stress in the hippocampus. Genomics 2007, 90,201–212.

[28] Mukhopadhyay, C. K., Mazumder, B., Fox, P. L., Role ofhypoxia-inducible factor-1 in transcriptional activation ofceruloplasmin by iron deficiency. J. Biol. Chem. 2000, 275,21048–21054.

[29] Martin, F., Linden, T., Katschinski, D. M., Oehme, F. et al.,Copper-dependent activation of hypoxia-inducible factor(HIF)-1: Implications for ceruloplasmin regulation. Blood2005, 105, 4613–4619.

[30] Das, D., Tapryal, N., Goswami, S. K., Fox, P. L., Mukho-padhyay, C. K., Regulation of ceruloplasmin in humanhepatic cells by redox active copper: Identification of a novelAP-1 site in the ceruloplasmin gene. Biochem. J. 2007, 402,135–141.

[31] Tavassoli, M., Kishimoto, T., Kataoka, M., Liver endotheliummediates the hepatocyte’s uptake of ceruloplasmin. J. Cell.Biol. 1986, 102, 1298–1303.

[32] Irie, S., Tavassoli, M., Liver endothelium desialates cer-uloplasmin. Biochem. Biophys. Res. Commun. 1986, 140,94–100.

[33] Thompson, E. J., The CSF Proteins: A BiochemicalApproach, Elsevier, Amsterdam 1988.

[34] Morell, A. G., Van den Hamer, C. J., Scheinberg, I. H., Ash-well, G., Physical and chemical studies on ceruloplasmin. IV.Preparation of radioactive, sialic acid-free ceruloplasminlabeled with tritium on terminal D-galactose residues. J.Biol. Chem. 1966, 241, 3745–3749.

[35] Morell, A. G., Gregoriadis, G., Scheinberg, I. H., Hickman, J.,Ashwell, G., The role of sialic acid in determining the survi-val of glycoproteins in the circulation. J. Biol. Chem. 1971,246, 1461–1467.

[36] Morell, A. G., Irvine, R. A., Sternlieb, I., Scheinberg, I. H.,Ashwell, G., Physical and chemical studies on cer-uloplasmin. V. Metabolic studies on sialic acid-free cer-uloplasmin in vivo. J. Biol. Chem. 1968, 243, 155–159.

[37] Miller, R. G., Mitchell, J. D., Lyon, M., Moore, D. H., Riluzolefor amyotrophic lateral sclerosis (ALS)/motor neuron dis-ease (MND). Amyotroph. Lateral Scler. Other Motor NeuronDisord. 2003, 4, 191–206.


Documents

Differential expression of ceruloplasmin isoforms in the cerebrospinal fluid of amyotrophic lateral sclerosis patients