RESEARCH ARTICLE

Alterations of Juxtaparanodal Domains in

Two Rodent Models of CNS Demyelination
Lida Zoupi,1,2 Kyriaki Markoullis,3 Kleopas A. Kleopa,3 and Domna Karagogeos1,2

The segregation of myelinated fibers into distinct domains around the node of Ranvier—the perinodal areas—is crucial for
nervous system homeostasis and efficient nerve conduction. Perinodal areas are formed by axo-glial interactions, namely the
interaction of molecules between the axon and the myelinating glia. In a variety of demyelinating pathologies including multiple
sclerosis, the molecular architecture of the myelinated fiber is disrupted, leading to axonal degeneration. In this study we have
analyzed the alterations of TAG-1, Caspr2, and voltage-gated potassium channels (VGKCs), forming the juxtaparanodal tripartite
complex, in relation to adjacent paranodal and nodal molecules, in two different models of CNS demyelination, the experimental
autoimmune encephalomyelitis (EAE) and the cuprizone model of toxic demyelination. We found extensive alterations of the
juxtaparanodal molecular architecture under de- and remyelinating conditions. Inflammation alone was sufficient to disrupt the
borders between the domains leading to the diffusion of juxtaparanodal components to the adjacent paranodal area. EAE
induction and cuprizone-induced demyelination resulted initially in paranodal domain elongation with subsequent diffusion of
the juxtaparanodal components and the reduction of their expression levels. At later stages, with decreasing inflammation and
spontaneous remyelination there was a partial restoration of the paranodal domain but not sufficient re-organization of the
juxtaparanodes. The latter were re-formed only when complete remyelination was allowed in the cuprizone model, indicating
that juxtaparanodal domain reorganization is a later event that may remain incomplete in a hostile inflammatory milieu.

GLIA 2013;61:1236–1249
Key words: nodes of Ranvier; paranodes; EAE; cuprizone; demyelination

Introduction
M ultiple Sclerosis (MS) is a chronic, inflammatory, neurode-
generative disease of the central nervous system (CNS). MS
pathology includes multifocal demyelinated plaques accompanied
by the formation of an active glial scar (Lassmann, 2001). In
actively demyelinating lesions, macrophages containing fragments
of myelin, microglia, and infiltrating T-cells are associated with
lesion formation in the white matter, although recent studies
point out at alterations in gray matter areas as well (Barnett and
Prineas, 2004; Lassmann, 2001, 2008; Lucchinetti et al., 2011;
Prineas and Parratt, 2012). Lesion formation is a multi-factorial
process that includes cytolytic factors like cytokines and proteases,
secreted by inflammatory cells, complement activation, energy de-
privation, oxidative, and nitrative damage. Recently a hypothesis
implicating oligodendrocyte alterations or damage as preceding
macrophage and microglia recruitment in newly formed MS
lesions has been put forward (Barnett et al., 2006; Bradl and Lass-
mann, 2010; Franklin et al., 2012; Haider et al., 2011; Lassmann,
2001, 2003; Lassmann et al., 2001; Lassmann and Lucchinetti,
2008). However, in spite of many advances in the field, the mech-
anisms of disease initiation and progression remain obscure.
A healthy myelinated fiber is periodically interrupted at
the nodes of Ranvier. The anchored sodium channels are re-
sponsible for the generation and efficient propagation of axon
potentials. Next to the node, a physical barrier to the lateral
diffusion of nodal components is formed by the close apposi-
tion of the myelin membrane to the axon and the interaction
of glial NF155 with the axonal immunoglobulin superfamily
(IgSF) molecule Contactin-1 and the Neurexin superfamily
protein Caspr/Paranodin (Charles et al., 2002; Labasque and
Faivre-Sarrailh, 2010; Sherman et al., 2005; Susuki and Ras-
band, 2008). Juxtaparanodes form adjacent to paranodes and

View this article online at wileyonlinelibrary.com. DOI: 10.1002/glia.22511

Published online Jul 5, 2013 in Wiley Online Library (wileyonlinelibrary.com). Received Jan 8, 2013, Accepted for publication Mar 20, 2013.

Address correspondence to Dr. O. Karagogeos, Domna Karagogeos, Department of Basic Science, Faculty of Medicine, University of Crete and Institute of Molecular
Biology and Biotechnology–Foundation for Research and Technology, Vassilika Vouton, 71110, Heraklion, Crete, Greece. E-mail: karagoge@imbb.forth.gr

From the 1Department of Basic Science, Faculty of Medicine, University of Crete, Heraklion, Greece; 2Institute of Molecular Biology and BiotechnologyFoundation for
Research and Technology, Vassilika Vouton, 71110, Heraklion, Crete, Greece; 3Neuroscience Laboratory and Neurology Clinics, The Cyprus Institute of Neurology and
Genetics (CING), P.O. Box 23462, 1683 Nicosia, Cyprus.

Additional Supporting Information may be found in the online version of this article.

1236 V
C 2013 Wiley Periodicals, Inc.
 Zoupi et al.: Juxtaparanodal Alterations in CNS-Demyelinating Models
are comprised of a tripartite complex between the Shaker-
type voltage gated potassium channels (VGKCs), Caspr2
(another member of the Neurexin Superfamily) on the axon
and the IgSF molecule TAG-1/Contactin-2 both on the axon
and the glial cell (Labasque and Faivre-Sarrailh, 2010; Poliak
and Peles, 2003; Poliak et al., 2003; Savvaki et al., 2008;
Traka et al., 2002, 2003). These molecular interactions medi-
ate the axo-glial contact that is essential for the organization
of axonal domains and axonal homeostasis.
Humoral responses were recently found to severely affect
the organization of white matter myelinated fibers in human
MS lesions, initiating the disruption of the paranodal area and
resulting in a progressive axonal denudation, and nerve conduc-
tion impairment (Coman et al., 2006; Howell et al., 2006,
2010; Wolswijk and Balesar, 2003). Moreover, axo-glial contact
components (Neurofascin isoforms and Contactin-2/ TAG-1)
were also identified as auto-antigens in a subset of MS patients,
suggesting that a specific T cell-mediated immune response is
directed against perinodal components, during the course of
the disease (Derfuss et al., 2009; Mathey et al., 2007). These
results highlight the association of MS pathology with altered
molecular organization of the myelinated fiber.
To further clarify the nature and time course of these
perinodal molecular changes in MS-related pathologies we per-
formed a detailed analysis of the tripartite complex of the jux-
taparanodal domain under de-and remyelinating conditions, in
two different rodent models of CNS demyelination: the experi-
mental autoimmune encephalomyelitis (EAE) that is character-
ized as one of the rodent model of MS and the cuprizone
model of toxic demyelination. In these models we analyzed the
effects of myelin disruption on the localization and expression
of juxtaparanodal proteins, with or without the involvement of
a T-cell mediated immune response at different time points.
We identified severe juxtaparanodal alterations under demyeli-
nating conditions in both models. The expression levels of all
three juxtaparanodal proteins were severely decreased compared
with control animals, with disruption of normal localization
and diffusion. In both models, partial paranodal reorganization
was evident correlating with spontaneous remyelination and
decreasing inflammation, while juxtaparanodal reorganization
occurred only with complete remyelination in the cuprizone
model. Our study provides insights into juxtaparanodal protein
complex involvement under demyelinating and remyelinating
conditions in relation to paranodal and nodal proteins in two
different rodent models of CNS demyelination.
Materials and Methods
EAE Induction
For the purpose of the present analysis we used four different groups
of animals that have been previously described by Markoullis et al.,
(2012). Six-week-old female C57BL/6, wild type (WT) mice weight-
August 2013
ing 16 to 18 g were purchased from HarlanTM (San Pietro Al Nati-
sone, Italy) and kept in our mouse facility for 2 weeks before EAE
induction at the age of 8 weeks. Animals were kept under controlled
conditions of temperature (21–23 C), humidity, air-exchange and
light cycle (12/12 h light/dark) and provided with standardized
mouse diet and drinking water ad libitum. For disease induction,
mice were anaesthetized with an intraperitoneal injection of Tribro-
moethanol solution (AvertinV R, Sigma-Aldrich, Germany) and
injected subcutaneously in the dorsal area of the tail base with 160
lL of emulsion containing 200 lg (end volume: 2.5 mg/mL) of
mouse recombinant myelin oligodendrocyte glycoprotein (rMOG)
and an equal volume of Complete Freund’s Adjuvant (CFA) mixed
with 5 mg/mL desiccated M. tuberculosis (Difco Laboratories,
Detroit, USA). rMOG was produced as previously described (Price
et al., 2002). Simultaneously with rMOG as well as 48 h later, each
mouse received intraperitoneally 200 lL Pertussis Toxin (PTX) (1
lg/mL in PBS) (List Biological Laboratories, CA). Control mice (n
¼ 6) were injected with CFA mixed with M. tuberculosis and PTX
in the same way and time points as for EAE. A group of littermate
female mice (n ¼ 6) were used as naive controls. Mice were
weighted and evaluated daily for clinical changes using the following
scale: 0 ¼ no clinical signs, 1 ¼ paralysis of the tail, 2 ¼ parapare-
sis/monoplegia, 3 ¼ paraplegia/hemiparesis, 4 ¼ hemiplegia/quadri-
paresis, 5 ¼ quadriplegia/moribund. Groups of mice (n ¼ 7 each)
were sacrificed at 14 days post immunization (dpi), 2–3 days after
the onset of the disease (MOG-EAE and CFA controls) and at 28
dpi (MOG-EAE and naive controls).
Cuprizone Administration
Eight-week-old male mice (C57B110/CBA) were fed with 0.2%
cuprizone [oxalic bis(cyclohexylidenehydrazide)] (Sigma-Aldrich,
USA) mixed with the animal’s standard chow. As previously
described (Matsushima and Morell, 2001), cuprizone toxin induces
full demyelination of the corpus callosum when administered contin-
uously for 6 weeks, while toxin removal from the chow allows
remyelination to occur. Juxtaparanodal complex formation was tested
after 4.5 (demyelination can be visualized as distinct lesions) and 6
weeks of cuprizone treatment (demyelination and partial remyelina-
tion). Additionally, 6 weeks of treatment with subsequent toxin re-
moval for 3 weeks (referred as 9 weeks) was considered as remyelina-
tion time point. All groups (3–4 animals/group) were compared
with age and sex matched untreated controls.
Immunohistochemistry
Spinal cords and cortices were harvested from mice after transcardial
perfusion with 4% paraformaldehyde (PFA) in 0.1 M phosphate
buffer saline (PBS). Cervical and thoracic spinal cord segments were
isolated from MOG-EAE mice, for the present analysis. Tissues were
then post-fixed in the same fixative for 1 h (h), cryo-protected over-
night in 20% sucrose in 0.1 M PBS, embedded in OCT compound
and stored at 80oC. Quantification of demyelination, inflamma-
tion and axonal loss was performed in transverse sections of the na-
ive, CFA-14dpi and MOG-EAE cervical and thoracic segments as
previously described (Markoullis et al., 2012). Longitudinal white
matter cryosections from the posterior spinal cord (through the
1237
dorsal columns) and corpus callosum (CC) of 10 lm were obtained
and mounted on SuperfrostV R Plus microscope slides (O. Kindler,
Germany). For immunohistochemistry, sections were post-fixed in
ice cold acetone for 10 min at 20oC, blocked and incubated with
primary and secondary antibodies in 5% BSA (Sigma-Aldrich, USA)
0.5 % Triton-X in 0.1 M PBS. Slides were mounted using
MOWIOL mounting medium (Calbiochem, Germany). Samples
were visualized and photographed via confocal microscopy (TCS
SP2, Leica Microsystems, Germany).
The following antibodies were used for immunohistochemistry:
rabbit polyclonal antibodies against TAG-1 (1:1000), against Caspr2
and Caspr (kind gifts by Dr Laurence Goutebroze, Inserm 536, Paris,
both antibodies 1:800); against Kv1.2 (Alomone, APC-010, Jerusalem,
1:200); against IbaI (Biocare Medical CP290A, USA, 1:500), against
Nav1.6 (ASC009, Alomone Labs, Jerusalem, 1:100). Mouse monoclo-
nal PanNav (clone K58/35, S8809, Sigma-Aldrich, Germany, 1:50);
Caspr (kind gift by Dr Elior Peles, Weizmann Institute of Science,
Israel, 1:1000), CC-1 (Merk Millipore OP80, Billerica, MA 01821,
1:100); RT97 (Developmental Studies Hybridoma Bank, 1:1,000) and
rat monoclonal antibody against Myelin Basic Protein (MBP - AbD
Serotec, MCA409S, UK, 1:200) and against Proteolipid Protein (PLP-
Prof. Reynold’ s lab,1:10). Sections were then incubated with fluoro-
chrome-labeled secondary antibodies Alexa Fluor 488 and 555 (all
from Molecular Probes, Eugene, OR, 1:800) and counterstained either
with 40, 60-diamidino-2-phenylindole (DAPI) (Sigma-Aldrich) or To-
Pro 3 iodide (Life Technologies, 642/661) to visualize cell nuclei.
Quantification of Perinodal Pathology
For quantification of pathological changes in perinodal molecular
architecture in MOG-EAE, longitudinal posterior white matter cryo-
sections from cervical and thoracic spinal cord were immunostained
with combinations of antibodies against Caspr, Caspr2, Kv1.2, and
TAG-1, to label paranodes and juxtaparanodes, respectively. We
counted the percentage of paranodes with/without stained juxtapar-
anodes within a square area of 95  95 lm in images obtained at
the confocal microscope. We examined five to six different areas
from each mouse and three different animals per group. In the same
areas, we calculated the total juxtaparanodal length by adding the
length of each juxtaparanodal segment of the myelinated fiber, in
immunostained spinal cord cryosections against Caspr / Caspr2 and
Caspr/TAG-1 (Fig. 2A inset image). ImageJ software was used for
the calculation of the length, width and area of juxtaparanodes. Peri-
nodal pathology was also quantified in respective areas of the corpus
callosum of cuprizone treated animals and depicted as a column
graph of the number of clustered domains versus the total number
of perinodal areas detected in each group. Data are expressed as
means 6 standard deviation of the mean or as scatter plots. Statisti-
cal analysis of measured values between the different animal groups
was performed, using nonparametric, one-way ANOVA with addi-
tional Bonferroni post-test for multiple comparisons between the
animal groups (P < 0.05 was considered statistically significant).
Western Blot Analysis
Tissues were collected from freshly sacrificed adult mice and homoge-
nized directly in ice cold RIPA buffer (10 mM Sodium phosphate pH
1238
7.0, 150 mM NaCl, 2 mM EDTA, 50 mM Sodium fluoride, 1%
NP-40, 1% Sodium deoxycholate and 0.1% SDS) with an addition of
protease inhibitor cocktail (Roche) followed by a brief sonication on
ice. The total protein concentration in each sample was quantified with
the Bradford kit (Bio-Rad Laboratories). Protein extracts were analyzed
by SDS poly-acrylamide gel electrophoresis and transferred to Hybond-
C extra membrane (GE Healthcare Bio-Sciences) for 1 hour using a
wet transfer unit (Bio-Rad Laboratories,). Following 1 hour blocking
(5% powdered skim milk and 0.1% Tween-20 in 0.1 M PBS); the
membrane was incubated overnight with the primary antibodies. After
washing three times for 10 min in 0.1 % Tween-20 in 0.1 M PBS,
samples were incubated for 1 h at room temperature with horseradish
peroxidase–coupled secondary antibodies, and proteins were visualized
by enhanced chemiluminescence (ECL Plus, GE Healthcare Bio-Scien-
ces). The following antibodies were used for Western blot analysis: rab-
bit polyclonal antibodies against TAG-1 (1:4000), rabbit polyclonal
antibodies against Caspr2 and Caspr (kind gifts by Dr Laurence Gou-
tebroze, both antibodies 1:3000), rabbit polyclonal antibody against
Kv1.2 (Alomone, APC-010, Jerusalem, 1:1000), mouse monoclonal
antibody against GAPDH (internal control, Santa Cruz, SC32233,
1:4000) and horseradish peroxidase–coupled secondary antibodies (all
from Jackson ImmunoResearch Laboratories, 1:5000-1:10,000). Band
intensity of comparable animal groups (n ¼ 3 from each group) was
quantified with Tinascan version 2.07d and normalized with the re-
spective GAPDH intensity levels as loading control. Expression values
are shown as % percentage considering naive animal values as 100%.
Data are depicted as means 6 standard deviation of the mean and sta-
tistical analysis was performed, using nonparametric, one-way ANOVA
with additional Bonferroni post-test for multiple comparisons (P <
0.05 was considered statistically significant).
Results
Juxtaparanodal Protein Alterations Upon EAE
Induction
Our analysis of the juxtaparanodal complex alterations in
EAE stages consists of comparisons between MOG-EAE (14
days post induction [dpi] and 28dpi), Complete Freund’s ad-
juvant (CFA-14dpi) and naive control animals. In animals
treated with CFA, an increased CNS inflammation occurs
but without the formation of demyelinating lesions. In our
MOG-EAE model, the disease peaks at 21 days post immuni-
zation (dpi) and exhibits partial remission at 28dpi. We exam-
ined animals at 14dpi, when total demyelination is higher and
at 28dpi, when inflammation remits and demyelinating lesions
are reduced (partial remyelination). The overall axonal loss in
all MOG-EAE time points was 20% in spinal cord white
matter (Markoullis et al., 2012). Details of clinical scores and
quantification of pathology in MOG-EAE are summarized in
Table 1, Supp. Info. Fig. 1 and Supp. Info. Fig. 2.
VGKCs were the first component of the juxtaparanodal
complex we examined via immunohistochemistry on longitudi-
nal cryosections of posterior spinal cord, with antibodies against
paranodal Caspr and the VGKC subtype Kv1.2 (Fig. 1A–D,
Volume 61, No. 8
 Zoupi et al.: Juxtaparanodal Alterations in CNS-Demyelinating Models

FIGURE 1: Perinodal complex disruption during the course of MOG-EAE. A–D: Paranodal Caspr (red) and juxtaparanodal VGKCs (green)
immunohistochemistry in posterior spinal cord white matter cryosections from naive (A), CFA-14dpi (B), 14dpi (C), and 28dpi (D) MOG
induced EAE animals. E–H: Paranodal Caspr (red) and juxtaparanodal Caspr2 (green) immunohistochemistry in posterior spinal cord
white matter cryosections from naive (E), CFA-14dpi (F), 14dpi (G) and 28dpi (H) MOG-induced EAE animals. I–L: Nodal sodium channel
subunit 1.6 (red) and juxtaparanodal VGKCs (green) immunohistochemistry in posterior spinal cord white matter cryosections from naive
(I), CFA-14dpi (J), 14dpi (K) and 28dpi (L) MOG induced EAE animals. Juxtaparanodal proteins are normally positioned next to paranodal
Caspr and nodal Nav1.6 in control fibers (A, E, insets below merged images correspond to single fibers, asterisks). CFA inflammation
did not result in any apparent alteration of perinodal protein localization but induced the diffusion of juxtaparanodal Caspr2 to the para-
nodal area; 14 days post injection (B, F, J insets). At the peak of MOG-EAE (14dpi), juxtaparanodal protein clustering is severely
reduced, albeit in a lesser extent than that of paranodal Caspr (C, G, insets). Moreover, the nodal density is strongly affected (K, insets).
At a later MOG-EAE stage (28 dpi) the perinodal protein disorganization is still evident, albeit the focal remyelination efforts observed
at this stage (D, H, L, asterisks in merged images represent properly clustered perinodal domains). [Color figure can be viewed in the
online issue, which is available at wileyonlinelibrary.com.]

insets represent separate channels and merged images of single
fibers). In naive animals (Fig. 1A, insets, asterisks represent
properly clustered perinodal domains), as well as in CFA-14dpi
(Fig. 1B, insets), VGKCs are clustered at juxtaparanodes next to
paranodal Caspr. In contrast, during the peak of demyelination,
a disrupted paranodal phenotype was evident in MOG-EAE
14dpi (Fig. 1C, insets). VGKCs formed reduced overall clusters
at the juxtaparanodal area with apparent diffusion of the
remaining VGKC aggregates to the adjacent domains (Fig. 1C,
insets). At MOG-EAE 28dpi, extensive VGKC diffusion was
still evident (Fig. 1D, insets) in spite of the partial reorganiza-
tion of the paranodal domains at this stage.
Taking these results into account we asked whether the
other two molecules of the juxtaparanodal tripartite complex
display similar phenotypes as VGKCs. Under physiological
conditions (Naive, Fig. 1E), juxtaparanodal Caspr2 was local-
ized next to paranodal Caspr, occupying distinct areas on the
myelinated fiber (insets and asterisks). In CFA-14dpi spinal
cord white matter, inflammatory infiltrates were prominent,
although without the presence of demyelinating lesions (Mar-
koullis et al., 2012). We observed the majority of Caspr2
staining properly aggregated at the juxtaparanodal domain
although occasionally appeared diffused in the Caspr-positive
paranodal region (Fig. 1F, insets). At the same stage we did
not observe any alterations in Caspr localization (Fig. 1F,
insets). In MOG-EAE spinal cord, during the peak of inflam-
mation and demyelination at 14dpi, the paranodal organiza-
tion was disrupted, with the Caspr signal either absent or
diffused along the myelinated fiber (Fig. 1G, insets). Juxtapar-
anodal Caspr2 localization was also affected upon MOG-EAE
induction. At MOG-EAE 28dpi, when the inflammatory
infiltrates decrease and spontaneous re-myelination may

August 2013 1239

 TABLE 1: Clinical and Pathological Features of Animals Examined in the MOG-EAE Study

% white matter covered % of white matter

with inflammatory cells demyelinated axons % axonal loss
(AVG 6 SEM) (AVG 6 SEM) (AVG 6 SEM)

Clinical score*
Animal group (AVG 6 SEM) Cervical Thoracic Cervical Thoracic Cervical Thoracic

Naive (n ¼ 6) 0 0.6 6 0.11 0.64 6 0.02 0 0 0 0

CFA (n ¼ 6) 0 1.63 6 0.31 1.69 6 0.33 0 0 16.97 6 3.3 7.38 6 1.03
EAE 14 dpi (n ¼ 7) 0.5 6 0.28 4.2 6 1.2 5.26 6 1.27 2.83 6 1.45 7.29 6 3.17 17.47 6 6.49 23.17 6 6.33
EAE 28 dpi (n ¼ 7) 1.63 6 0.35 2.44 6 0.89 4.28 6 1.55 1.1 6 0.4 3.88 6 1.91 32.33 6 1.82 20.91 6 3.86
Clinical scores represent average weekly scores.

occur, partial reclustering of paranodal Caspr and juxtapara-
nodal Caspr2 was evident (Fig. 1H, insets).
We similarly analyzed the expression of TAG-1, which
is expressed both by axons and myelinating glia in CNS and
PNS, and plays a key role in the organization and mainte-
nance of the juxtaparanodal complex (Savvaki et al., 2008,
2010; Traka et al., 2002, 2003). Immunohistochemical analy-
sis with specific antibodies for Caspr and TAG-1 showed jux-
taparanodal localization of the latter in naive myelinated
fibers (Supp. Info. Fig. 3A, insets and asterisks). In CFA-
14dpi mice, the non-specific immune response caused a clear
diffusion of TAG-1 protein from the juxtaparanodes to par-
anodes (Supp. Info. Fig. 3B, insets). At MOG-EAE 14dpi
where Caspr staining was disrupted (Supp. Info. Fig. 3C,
insets), TAG-1 juxtaparanodal localization was also reduced
similar to the other two juxtaparanodal proteins. At 28dpi,
juxtaparanodal clustering of TAG-1 was incomplete, since the
detected signal remained diffused (Supp. Info. Fig. 3D,
insets), while clustered localization was observed only in a
small percentage of fibers (asterisks) or in heminodal-like
structures (Supp. Info. Fig. 3D, insets).
The differences in the clustering between paranodal and
juxtaparanodal components prompted us to investigate the
effect on nodal clustering between 14dpi and 28dpi MOG-
EAE stages, in relation to the juxtaparanodal component. To
this end, we immunostained for the mature form of sodium
channels (Nav1.6), a form mainly affected in demyelinating
diseases (Craner et al., 2003; Waxman et al., 2004), along
with juxtaparanodal VGKCs (Fig. 1I–L, insets). CNS inflam-
mation in CFA-14dpi did not have any apparent effect in the
nodal density of the white matter areas (Fig. 1J, insets and
asterisks). In contrast, at MOG-EAE 14dpi, there is an
obvious reduction in the nodal density of the white matter
(Fig. 1K) which is sustained at MOG-EAE 28dpi (Fig. 1L)
and may reflect the percentage of axonal loss of observed in
these stages (Markoullis et al., 2012).
Juxtaparanodal Protein Diffusion in MOG- EAE
Stages
Our following aim concerned the evaluation of the immuno-
histochemical results. More specifically, our interest was
mainly focused to the less well analyzed molecules of the jux-
taparanodal complex, Caspr2 and TAG-1. Protein staining
distribution at the juxtaparanodal area was quantified by
measuring the total length of the stained domain (Caspr2 or
TAG-1- positive) in each myelinated fiber, in relation to the
respective width (Fig. 2A, inset picture). This measurement
was performed in all animal groups and presented as separate
scatter plots (Fig. 2A–H). In addition, the average length and
the total area were also measured and presented as column
graphs in Fig. 2I–L.
Under control conditions, Caspr2-positive areas were
characterized by a specific range of length, width and total
area values (Fig. 2A,I,K). In CFA-14dpi animals, Caspr2 dif-
fusion towards the paranodes did not alter the overall area
length according to the respective scatter plot measurements
(Fig. 2B) while the average length and area (Fig. 2I,K, respec-
tively), showed a nonsignificant increase compared to naive
animals. In contrast, MOG-EAE induction resulted in an
obvious clustering reduction of Caspr2 protein compared to
naive or CFA myelinated fibers (Fig. 1C). Moreover, the
remaining clustered juxtaparanodes diffused, as indicated by
the significant increase in both the length and the area of
Caspr2-positive signal (Fig. 2C,I,K, P < 0.05* and P <
0.01**). Finally, despite the apparent focal perinodal re-orga-
nization (Fig. 1D,H), analysis of the diffusion status of
Caspr2 at 28dpi points at only minor and non-significant dif-
ferences compared with 14dpi animals (Fig. 2D,I,K).

1240 Volume 61, No. 8

 Zoupi et al.: Juxtaparanodal Alterations in CNS-Demyelinating Models

FIGURE 2: Diffusion of juxtaparanodal components at different MOG-EAE stages. Scatter plot analysis of the distribution of Caspr2 (A–
D) and TAG-1 staining (E–H) at juxtaparanodes as the length in relation to the corresponding width of the juxtaparanodal protein1 area
(A: the lines in the inset picture show the measured area). Quantification of the average length and area of Caspr2 and TAG-1 protein
staining is depicted in graphs I and J, respectively. CNS inflammation (CFA-14dpi) transiently disrupted the borders between the
domains leading both Caspr2 and TAG-1 proteins to the paranodal domain, whilst with no significant alteration of the length and area
of the stained segment (B, F, I–K). At MOG-EAE 14dpi, the fiber organization was impaired. Caspr2 and TAG-1 aggregates that were
still detected at this stage were diffused towards the adjacent areas (C, G, I–K). At EAE-28dpi, there is an obvious reorginazation of par-
anodal domains and not of juxtaparanodal proteins as indicated bythe unchanged values in the length and area of Caspr2 and TAG-1
staining, compared to MOG-EAE 14dpi (D, H, I–K)–; P < 0.05 *, P < 0.01 **).

For the TAG-1 protein, a similar measurement was per-
formed (Fig. 2E–H,J,L). In CFA-14dpi, the respective scatter
plot (Fig. 2F) showed a shift to higher length values indicat-
ing TAG-1 diffusion but without alterations in the average
juxtaparanodal length and occupied area compared with naive
animals (Fig. 2E,F,J,L). At the peak of demyelination (MOG-
EAE 14dpi), TAG-1 diffusion was obvious with a subsequent
increase in the total length and area (Fig. 2G,J,L, P < 0.05)
that was not altered at the 28dpi stage (Fig. 2H,J,L). Finally,
we also observed a subsequent increase of the juxtaparanodal
area width in both MOG-EAE stages compared to control
animals (Fig. 2A–H).
EAE Effects on Protein Expression Levels of the
Perinodal Components
The activation of the immune system and the subsequent de-
myelination of the white matter fibers caused altered localization
of the perinodal proteins. We next asked whether the domain
disorganization is coupled to alterations of protein expression
levels (Fig. 3). Caspr protein levels were insignificantly reduced
in CFA-14dpi spinal cord, compared to naive animals. In con-
trast, MOG-EAE 14dpi mice showed significantly reduced lev-
els (P < 0.05), while a robust increase of Caspr expression was
found at MOG-EAE 28dpi coinciding with the remission of
inflammation (Fig. 3, panel A, P < 0.001).

August 2013 1241

FIGURE 3: MOG-EAE effects on protein expression levels of the perinodal components. Western blot analysis of spinal cord lysates
from naive, CFA-14dpi, MOG-EAE 14dpi and 28dpi animals. Band intensity quantification of the percentage of protein expression levels
normalized to the corresponding GAPDH levels A: Paranodal Caspr displayed reduced although not statistically significant protein levels
in both CFA and MOG-EAE 14dpi, compared with naive animals. At 28 dpi the protein levels were significantly increased compared with
all the groups analyzed. B: Western blot analysis for the potassium channel subunit 1.2 (VGKCs) revealed a significant decrease in CFA
and MOG-EAE 14dpi and 28dpi spinal cords. C: Caspr2 protein levels were significantly reduced in both CFA and MOG-EAE 14dpi ani-
mals. At 28dpi Caspr2 levels slightly increased but not significantly, compared to the other animal groups. D: TAG-1 expression levels
significantly decreased in MOG-EAE 14dpi- but not in CFA animals- a reduction that persisted in MOG-EAE 28dpi. (P < 0.05*, P <
0.01**, P < 0.001 ***).

Similar to Caspr, juxtaparanodal VGKC, Caspr2 and
TAG-1 protein levels were significantly reduced in MOG-
EAE 14dpi samples. However, in contrast to Caspr, there was
no alteration of the juxtaparanodal protein levels at MOG-
EAE 28dpi spinal cords (Fig. 3, panels B–D). Specifically,
VGKCs protein levels were strongly affected by the immune
response initiation and EAE induction, with protein levels
dropping to 40% of the naive levels in CFA (P < 0.01) and
20% at the peak of EAE demyelination (P < 0.01), while
remaining at similar levels at 28dpi (Fig. 3, panel B). Caspr2
protein was decreased in both CFA and MOG-EAE 14dpi (P
< 0.01) but the subsequent increase at 28dpi was not signifi-
cant (Fig. 3, panel C, P > 0.05). Finally, TAG-1 levels were
also affected, although to a lesser extent, exhibiting a 40%
decrease at MOG-EAE14dpi (P < 0.05) and an additional
20% reduction at 28dpi (Fig. 3, panel D, P < 0.001).

1242 Volume 61, No. 8

 Zoupi et al.: Juxtaparanodal Alterations in CNS-Demyelinating Models

FIGURE 4: Quantification of the perinodal domain clustering at different MOG-EAE stages. A: EAE induction resulted in a decrease of
the perinodal protein clustering at 14dpi, compared with naive and CFA animals that did not recover until 28dpi (P < 0.01 **, P <
0.001***) B: Nodal density measurement of white matter areas in the posterior spinal cord indicated a significant reduction of mature
nodes in both MOG-EAE stages (P < 0.05 *, P < 0.01 **). C: Quantification of heminodal formations between naive, 14dpi and 28dpi
MOG-EAE animals (P < 0.01 **, P < 0.001***). D: Total diagram of paranodal and juxtaparanodal distribution in the EAE stages. E: Peri-
nodal protein redistribution at 14 and 28dpi EAE animals. At 14dpi, paranodes were more susceptible to damage than the juxtaparano-
des (Caspr1 paranodes, light gray). On the contrary, juxtaparanodes were less affected since the three proteins of the juxtaparanodal
tripartite complex displayed higher levels of clustering with no significant differences between them. At 28dpi (dark gray), paranodal
clustering increased while juxtaparanodal staining decreased.

Perinodal Domain Clustering at Different EAE
Stages
So far we have reported alterations in the expression levels,
localization, length, and area of clustering of the perinodal
proteins, focusing on the juxtaparanodal tripartite complex
components. These differences led us to further examine the
differences on protein clustering between paranodes and jux-
taparanodes, as the disease progresses.
First we wanted to evaluate the extent of damage in the
perinodal clustering by quantifying the percentage of physiolog-
ically clustered perinodal domains (Supp. Info. Fig 2a) in spinal
cord white matter cryosections, at the two MOG-EAE stages
compared to control animal areas (Fig. 4A and Supp. Info. Fig.
4a). The overall clustering of perinodal proteins was markedly
reduced during the peak of demyelination (MOG-EAE 14dpi)
by 50% compared with naive and CFA-14dpi animals (Fig. 4A,
P < 0.001, P < 0.01 respectively). Similarly, our quantification
at 28dpi did not show any significant difference when com-
pared to 14dpi indicating a sustained clustering defect of peri-
nodal components in both MOG-EAE stages (Fig. 4A).
We additionally quantified the number of nodal sodium
channel clusters per microscopic area in CFA-14dpi, MOG-
EAE 14dpi, MOG-EAE 28dpi and naive animals by perform-
ing multiple comparisons between groups (Fig. 4B). CNS
inflammation in CFA-14dpi was not able to alter the white
matter nodal density (Fig. 4B). At MOG-EAE 14dpi, quanti-
fication of the axonal loss ranged between 15 and 20% com-
pared with naive animals (Supp. Info. Fig. 1K). Surprisingly,
the number of nodal clusters was reduced approximately by
50% at this stage and it was significantly reduced compared
to naive (P < 0.05) and to CFA-14dpi animals (P < 0.01)
(Fig. 4B). These results indicate extensive nodal damage upon
demyelination that is not restored following remission of
inflammation at 28dpi.
In parallel with the nodal and perinodal clustering during
the course of MOG-EAE, we have also quantified the amount
of heminodal structure formation in naive, 14 and 28 dpi
MOG-EAE. Heminodes are structures found at the borders of
each myelin segment that fuse to give rise to mature nodes dur-
ing the process of myelination (Feinberg et al., 2010). We
found a significant increase in the number of heminodes in spi-
nal cord white matter at MOG-EAE 28dpi, indicating an
ongoing remyelination attempt (Fig. 4C, P ¼ 0.004).
The immunohistochemical analysis of spinal cord white
matter revealed specific phenotypes of perinodal clustering in
the two MOG-EAE stages (Supp. Info. Fig. 4b and c). In
particular, we encountered paranodes without the presence of
juxtaparanodes and vice versa. Taking into account these

August 2013 1243

clustering defects, we have quantified the ratio of each
abnormality to the total number of detected perinodal
domains for each protein of interest (Fig. 4D). At MOG-
EAE 14dpi paranodal Caspr was either diffused along myelin-
ated fibers or disrupted. In addition, all three juxtaparanodal
proteins were affected in terms of their localization. Consider-
ing that paranodal area elongation is thought to be an early
event before demyelination (Coman et al., 2006; Howell
et al., 2006; Howell et al., 2010) we examined separately par-
anodal and juxtaparanodal integrity at the peak of demyelin-
ation. Indeed, at MOG-EAE 14dpi, paranodes were more
affected than juxtaparanodes (increased numbers of clustered
juxtaparanodes with disrupted paranodes), confirming the
increased paranodal susceptibility to damage, upon demyelin-
ation (Fig. 4D, E). At 28dpi, remission of inflammation and
spontaneous remyelination permitted the re-clustering of the
paranodal components (increased numbers of clustered paran-
odes with diffused juxtaparanodes), as indicated by the
increased levels of Caspr protein expression (Fig. 3, panel A)
and clustering (Fig. 4D,E). In contrast, juxtaparanodal pro-
teins remained diffused, similar to 14dpi (Fig. 4D,E).
Perinodal Protein Clustering is Affected in the
Cuprizone Model of Toxic Demyelination
The results from the MOG-EAE model highlight the perino-
dal disorganization during the peak of EAE and at a subse-
quent stage when the immune T cell infiltrates are decreasing
(28dpi). Although reorganization of the paranodal domain
was observed upon decreased inflammation, it was not
adequate to restore the aggregation of the juxtaparanodal
components. We wished to examine whether juxtaparanodal
proteins were affected upon demyelination and remyelination
without the implication of the immune system. For this rea-
son, we used the cuprizone model of toxic demyelination
(Matsushima and Morell, 2001).
Corpus callosum cryosections were initially immuno-
stained for MBP, for the discrimination of well-defined
lesions (L) (Fig. 5A–D) followed by immunohistochemical
analysis using antibodies against mature oligodendrocytes
(CC-1) and activated microglia (Iba1) for the detection of
innate immunity (Supp. Info. Fig. 5). Antibodies against
nodal sodium channels (Nav, Fig. 5E–H), paranodal Caspr
and juxtaparanodal proteins Caspr2 (Fig. 5I–L) and VGKCs
(Fig. 5M–P) were subsequently used.
Untreated animals showed uniform MBP staining of the
corpus callosum (Fig. 5A), normal numbers of mature oligoden-
drocyte and microglial populations (Supp. Info. Fig. 5A) and
proper myelinated fiber organization with the juxtaparanodal
protein staining to be adjacent to the paranodal Caspr and to
nodal Nav (Fig. 5A,E,I,M). After 4.5 weeks of cuprizone treat-
ment MBP staining was disrupted and distinct lesions in the
1244
corpus callosum were evident (Fig. 5B). Moreover, the nodal
density was obviously reduced after 4.5 and 6 weeks of cupri-
zone administration and nodal domain elongation of denuded
axons was also observed (Fig. 5F,G arrows indicate disrupted
phenotypes). In this model, we also quantified the overall my-
elinated fiber clustering in the corpus callosum, performed as in
the case of the MOG-EAE model (Fig 4A). During demyelin-
ation the number of properly organized fibers was severely
reduced (Fig. 5Q,R). The remaining Caspr localization was
characterized by disruptions, shrinkage or diffusion along the
myelinated fiber (Fig. 5F,J,N, arrows), phenotypes that were also
sustained after 6 weeks of cuprizone treatment (Fig. 5G,K,O,
arrows). By quantifying Caspr-positive paranodal numbers in the
different groups, we observed a significant reduction of parano-
des after 4.5 weeks of cuprizone administration, compared to
untreated animals (Fig. 5Q). Slightly increased paranodal forma-
tion was observed at 6 weeks of treatment, characterized by
extensive demyelination followed by spontaneous remyelination
(Fig. 5G,K,O). Further nodal and paranodal recovery occurs at
9 weeks with full remyelination (Fig. 5H,L,P,Q).
Juxtaparanodal proteins displayed various alterations in the
cuprizone model similar to the ones observed in the MOG-EAE
model. Caspr2 protein signal was progressively diffused (at 4.5
weeks) and lost (at 6 weeks) (Fig. 5J,K, arrows). A similar pheno-
type was observed also for TAG-1 protein (data not shown).
Finally, VGKCs appeared diffused along the fiber at both treat-
ment stages (Fig. 5N,O, arrows). However, toxin removal fol-
lowed by remyelination restored the uniform MBP staining (Fig.
5D), increased the presence of mature oligodendrocytes in the
white matter and reduced the occurrence of activated microglia
in the area (Supp. Info. Fig. 5D), in contrast to the EAE situa-
tion where activated microglia persisted at 28dpi (Supp. Info.
Fig. 2). In addition, both paranodal and juxtaparanodal proteins
reclustered, leading to restoration of axonal domains (Fig. 5L,P).
After 4.5 weeks of treatment, the clustering of the
domains was severely reduced compared with untreated ani-
mals and persisted also after 6 weeks (Fig. 5R, P < 0.05 and P
< 0.01, respectively). In contrast, when the toxin was removed
and remyelination was allowed, axonal domains were fully
restored and were indistinguishable from the untreated ani-
mals. Thus, although our results indicate a similar pathological
pattern of domain disruption in both models during demyelin-
ation, remyelination in the cuprizone model resulted in resto-
ration of both paranodal and juxtaparanodal areas, while in the
MOG-EAE model juxtaparanodal domain clustering remained
incomplete even after paranodal reorganization.
Discussion
In this paper we provide a detailed analysis of perinodal pro-
tein alterations in two different models of CNS
Volume 61, No. 8
 Zoupi et al.: Juxtaparanodal Alterations in CNS-Demyelinating Models

FIGURE 5: Perinodal protein clustering is affected in the cuprizone model of toxic demyelination. Immunohistochemical analysis of cor-
pus callosum with antibodies against MBP (Myelin Basic Protein- A–D), nodal sodium channels (Nav) and perinodal proteins Caspr,
Caspr2 and VGKCs (E–P). Four different groups of animals were analyzed: untreated controls, animals treated with cuprizone toxin for
4.5 weeks, 6 weeks, and 6 weeks followed by 3 weeks of toxin removal for remyelination to occur (9 weeks). Untreated animals showed
normal MBP staining (A) They also displayed proper clustering of nodal Nav, paranodal Caspr and juxtaparanodal proteins (E, L, M). Af-
ter 4.5 and 6 weeks of Cuprizone treatment MBP staining was significantly reduced, an obvious lesion (L) was formed (B, C). Perinodal
protein clustering was strongly affected in demyelinated areas with reduced nodal density and the appearance of elongated nodes (F,G)
and Caspr staining appearing severely reduced, diffused or completely absent (F,J,N,G,K,O,Q,R, arrows indicate disrupted phenotypes).
In addition, both juxtaparanodal protein clustering was strongly affected exhibiting Caspr2 loss and VGKC diffusion towards the interno-
des (J, N, K, O, R, arrows). Toxin removal led to remyelination (D). Remyelination resulted in the restoration of perinodal protein local-
ization with reformed nodes, paranodes and re-clustered juxtaparanodes (H, L, P, Q). Quantification of the clustering status between
the groups is shown in as the ratio of fibers with clustered domains per total number of perinodal domains (R) while the average number
of paranodes per area is presented in Q (P < 0.05*, P < 0.01 **). [Color figure can be viewed in the online issue, which is available at
wileyonlinelibrary.com.]

demyelination, highlighting for the first time the molecular nodal protein organization is closely associated to myelinated
pathology of the juxtaparanodal domain of the myelinated fiber homeostasis and efficient signal transmission along the
fibers. During the last decade it has become evident that peri- axons. Analysis of different mouse mutants has offered

August 2013 1245

valuable information about the importance of axo-glial inter-
actions in maintaining perinodal domain integrity. Paranodal
protein mutants (Boyle et al., 2001; Rios et al., 2003; Sher-
man et al., 2005; Tait et al., 2000; Zonta et al., 2008) display
disrupted paranodal junctions, perinodal protein diffusion
and domain elongation. Disrupted paranodes are also related
to axonal pathology, since elimination of the paranodal bar-
rier and loss of axo-glial contact led to neurofilament disorga-
nization and swelling of the nodes and paranodes (Griffiths
et al., 1998; Mathis et al., 2001; Pillai et al., 2009). In addi-
tion to paranodal mutants, mice lacking either Caspr2 or
TAG-1 protein exhibit specific disorganization of the juxta-
paranodal domain with subsequent diffusion of potassium
channels towards the internode (Poliak et al., 2003; Traka
et al., 2003). Moreover, Tag-1-/- animals display significant
hypomyelination of the optic nerve and impaired gait, motor
co-ordination and memory deficits (Chatzopoulou et al.,
2008; Savvaki et al., 2008). These studies emphasize the
necessity of intact axo-glial organization for the establishment
of efficient nerve conduction and CNS function.
Apart from studies in mutant animals, recent reports
show the involvement of perinodal domains in human MS
pathology (Coman et al., 2006; Derfuss et al., 2009; Howell
et al., 2006; Howell et al., 2010; Mathey et al., 2007). More-
over, specific perinodal proteins (Neurofascin and TAG-1)
were identified as auto-antigens in MS patients (Derfuss
et al., 2009; Mathey et al., 2007). Nodal sodium channels
and NF186 redistribution was evident in active demyelinating
and chronic lesions in post-mortem MS brain (Craner et al.,
2004a,b; Howell et al., 2006). In addition, paranodal Caspr
(expressed by the axon) and NF155 (expressed by glia) were
shown to be affected in demyelinating MS lesions and in
rodent models of MS (EAE). Domain border disruption and
subsequent area elongation is now thought to be an early
event before the initiation of demyelination (Coman et al.,
2006; Howell et al., 2006, 2010). While nodal and paranodal
areas have been previously analyzed, less attention has been
given to the adjacent juxtaparanodal components. This is the
first study that analyzes all proteins of the juxtaparanodal tri-
partite complex in relation to paranodal and nodal compo-
nents in two different rodent models of CNS demyelination,
the EAE and the cuprizone model.
Our results showing juxtaparanodal alterations in these
two models add new information to the working model of
perinodal domain distribution under demyelinating condi-
tions (summarized in Fig. 6). We have shown that CNS
inflammation alone (CFA animals) is sufficient to disrupt the
barriers between the perinodal domains, resulting in the diffu-
sion of the juxtaparanodal components into the paranodal
area. This phenomenon has been previously reported by
Howell and colleagues (Howell et al., 2010), who noticed the
1246
transient overlap of paranodal NF155 with VGKCs in CFA-
10dpi animals. In addition, VGKC diffusion is evident in
EAE animals and post-mortem human MS lesions (Coman
et al., 2006; Howell et al., 2010). Our analysis supports and
extends these studies since in CFA-14dpi animals the overlap
of Caspr2 or TAG-1 with paranodal Caspr is evident.
Domain border disruption may be induced by the
inflammatory milieu alone as an early and direct event of
inflammation; this phenomenon may be restored by perinodal
protein recycling and border restoration as time passes. CNS
inflammation is also sufficient to affect juxtaparanodal protein
expression levels, causing a significant reduction of Caspr2
and VGKC levels. In contrast to the other two members of
the complex, TAG-1 protein expression was not significantly
reduced. Since TAG-1 is a key molecule of the juxtaparanodal
complex, interacting with both Caspr2 and VGKCs, one pos-
sibility for this discrepancy may be that it is more stable than
its partners and thus less susceptible to degradation upon
mild or temporary changes in the domain architecture. For
the above mentioned reasons it would be interesting to test
perinodal protein expression levels in MS patients, while jux-
taparanodal protein recycling should be further examined.
MOG-EAE induction results in altered distribution and
expression of juxtaparanodal proteins (Fig. 6A). During the
peak of inflammation/demyelination, a severe reduction of the
total clustering of perinodal domains is observed in spinal cord
white matter. Paranodes are primarily affected consistent with
their localization on the myelinated fiber that makes them
more susceptible to damage than the less exposed juxtaparano-
des. However, juxtaparanodes are also disorganized during
MOG-EAE. All three components of the juxtaparanodal com-
plex display reduced protein levels and disrupted clustering
(Fig. 6A). Previous studies (Howell et al., 2010) have shown a
profound diffusion of VGKCs even at 10dpi in MOG-EAE
spinal cords. The observed increase of the length occupied by
the signal and the respective area for both Caspr2 and TAG-1
during MOG-EAE, indicates diffusion of the proteins and
elongation of the juxtaparanodal domain. Domain width is also
an important factor as width augmentation has been related to
axonal pathology that also takes place under demyelinating
conditions (Soulika et al., 2009; Wolswijk and Balesar, 2003).
Having demonstrated that severe CNS inflammation
and subsequent demyelination leads to disruption of the orga-
nization of myelinated fibers, we analyzed 28dpi animals,
when inflammation is diminishing. At this stage, paranodal
Caspr protein levels increase and a marked reclustering of the
respective domain is evident. In addition, increased numbers
of heminodal-like structures are detected. In contrast, juxta-
paranodal protein levels and localization are similar to 14dpi
(Fig. 6A). Our results confirm the notion of sequential and
progressive disorganization of myelinated fibers under
Volume 61, No. 8
 Zoupi et al.: Juxtaparanodal Alterations in CNS-Demyelinating Models

FIGURE 6: Perinodal protein distribution in the two rodent models of CNS demyelination. Schematic representation of the nodal, para-
nodal and juxtaparanodal component distribution in the different MOG-EAE stages (A) and in the cuprizone model of toxic demyelin-
ation (B). A: CFA immunization alone disrupts the boundaries between paranodal and juxtaparanodal domains leading to the diffusion
of the juxtaparanodal proteins into the paranodal area. At MOG-EAE 14dpi, destruction of the myelin sheath causes perinodal domain
disorganization and the diffusion of perinodal proteins and the reduction of the nodal density. At MOG-EAE 28dpi, paranodes begin to
reform without the subsequent reclustering of the juxtaparanodal components. B: After 4.5 weeks of cuprizone treatment there is for-
mation of demyelinating lesions and disruption of perinodal architecture. This phenotype persists after 6 weeks of treatment although
an increase in the number of paranodal clustering is observed. Juxtaparanodal protein re-organization is evident only when extensive
remyelination takes place.

pathological conditions, even though the observed paranodal
protein aggregation and heminodal formation reflect the
ongoing remyelination process. A putative functional signifi-
cance of delayed reorganization of the juxtaparanodes may be
the partial or incomplete level of remyelination that is
observed in MOG-EAE 28dpi and possibly in shadow pla-
ques in the human MS brain. Additionally, our MOG-EAE
model is characterized by significant axonal loss that persists
and even increases at 28dpi, likely contributing to persistently
reduced axonal protein levels despite remyelination.
Adaptive immunity with invasion of activated T lym-
phocytes against myelin components in the CNS is a key
player in the induction and progression of MS pathology,
while the role of innate immunity is being increasingly appre-
ciated (Howell et al., 2010). In order to assess the changes of
juxtaparanodal organization in the absence of induced inflam-
mation, we studied another rodent model of CNS demyelin-
ation, the cuprizone model of toxic demyelination (Groebe
et al., 2009; Kipp et al., 2009; Matsushima and Morell
2001). At demyelinating stages, total perinodal clustering is

August 2013 1247

significantly reduced compared to naive animals. Phenotypi-
cally, Caspr is progressively diffused and disrupted. Moreover
the nodal density is severely reduced and elongated nodal
areas are obvious. Finally, juxtaparanodal Caspr2 is gradually
lost while VGKC clustering appears severely diffused towards
both paranodal and internodal regions. However, by the 6th
week of cuprizone treatment, paranodal re-formation is evi-
dent devoid of juxtaparanodal re-organization, which occurs
only when efficient remyelination is allowed and innate im-
munity is withdrawn (Fig. 6B). In contrast to MOG-EAE
animals, in the cuprizone model, paranodal regions re-organ-
ize first, followed by juxtaparanodal ones.
Our findings are relevant for the effects of dalfampridine
(4-AP), a known potassium channel blocker, in MS. Myelin
damage may expose VGKCs at juxtaparanodal regions which
are covered by myelin in healthy axons (Bostock et al., 1981;
Chiu and Ritchie, 1980, 1981; Haghighi et al., 1995; Karimi-
Abdolrezaee et al., 2004; Nashmi et al., 2000; Sherratt et al.,
1980). The potassium efflux through fast potassium channels
may cause axonal conduction block by preventing sufficient
depolarization and initiation of the action potential at the node
of Ranvier. 4-AP has been shown to restore action potential
conduction in demyelination models (Bostock et al., 1981;
Leung, 2012). In addition to myelin-dependent alterations of
potassium channel expression, as described here, myelin-inde-
pendent somatodendritic Kv2.1 are also altered in EAE (Juk-
kola et al., 2012), indicating that pharmacological actions of 4-
AP may occur at different levels and may be difficult to corre-
late with the sequential changes at the juxtaparanodes alone.
Our study provides a detailed analysis and therefore
highlights juxtaparanodal dis-and re-organization under path-
ological conditions. Demyelination leads to the gradual disor-
ganization initially of the paranodal area, followed by disrup-
tion of juxtaparanodal and nodal domains. Expression levels
of perinodal proteins are also affected by demyelination. As
inflammation diminishes, paranodal expression levels increase
and paranodes are reformed in order to restore the disrupted
nerve conduction and protect the axon from further damage.
Juxtaparanodal restoration does not occur in parallel with par-
anodal reorganization. This implies that restoration of the
juxtaparanodal complex is either a later event or a process
that remains incomplete possibly due to the persistence of a
hostile inflammatory environment as the disease progresses.
Acknowledgment
Grant sponsor: ELA Foundation; Grant number: 2007-
02512C; Grant sponsor: MS UK; Grant number: 847/07;
Grant sponsor: Cyprus Research Promotion Foundation;
Grant numbers: HEALTH/BIOS/0308/01, HEALTH/BIOS/
0609/BIE/09; Grant sponsor: Cyprus Telethon; Grant num-
1248
ber: 2010-14; Grant sponsors: IMBB\-FoRTH Scholarship
and University of Crete, Manasaki Scholarship.
The authors would like to thank Irene Sargiannidou
and Natasha Schiza for help with western blots and immuno-
staining and Dr Richard Reynolds for providing recombinant
MOG and advice on EAE induction. They also would like to
thank Maria Savvaki and Maura Strigini for comments on
the manuscript and Dr C. Linington for helpful discussion.
References
Barnett MH, Henderson AP, Prineas JW. 2006. The macrophage in MS: Just
a scavenger after all? Pathology and pathogenesis of the acute MS lesion.
Mult Scler 12:121–132.
Barnett MH, Prineas JW. 2004. Relapsing and remitting multiple sclerosis:
Pathology of the newly forming lesion. Ann Neurol 55:458–468.
Bostock H, Sears TA, Sherratt RM. 1981. The effects of 4-aminopyridine and
tetraethylammonium ions on normal and demyelinated mammalian nerve
fibres. J Physiol 313:301–315.
Boyle ME, Berglund EO, Murai KK, Weber L, Peles E, Ranscht B. 2001.
Contactin orchestrates assembly of the septate-like junctions at the paranode
in myelinated peripheral nerve. Neuron 30:385–397.
Bradl M, Lassmann H. 2010. Oligodendrocytes: Biology and pathology. Acta
Neuropathol 119:37–53.
Charles P, Tait S, Faivre-Sarrailh C, Barbin G, Gunn-Moore F, Denisenko-
Nehrbass N, Guennoc AM, Girault JA, Brophy PJ, Lubetzki C. 2002.
Neurofascin is a glial receptor for the paranodin/Caspr-contactin axonal
complex at the axoglial junction. Curr Biol 12:217–220.
Chatzopoulou E, Miguez A, Savvaki M, Levasseur G, Muzerelle A, Muriel MP,
Goureau O, Watanabe K, Goutebroze L, Gaspar P, et al. 2008. Structural
requirement of TAG-1 for retinal ganglion cell axons and myelin in the
mouse optic nerve. J Neurosci 28:7624–7636.
Chiu SY, Ritchie JM. 1980. Potassium channels in nodal and internodal axonal
membrane of mammalian myelinated fibres. Nature 284:170–171.
Chiu SY, Ritchie JM. 1981. Evidence for the presence of potassium channels
in the paranodal region of acutely demyelinated mammalian single nerve
fibres. J Physiol 313:415–437.
Coman I, Aigrot MS, Seilhean D, Reynolds R, Girault JA, Zalc B, Lubetzki C. 2006.
Nodal, paranodal and juxtaparanodal axonal proteins during demyelination and
remyelination in multiple sclerosis. Brain 129(Part 12):3186–3195.
Craner MJ, Hains BC, Lo AC, Black JA, Waxman SG. 2004a. Co-localization
of sodium channel Nav1.6 and the sodium-calcium exchanger at sites of
axonal injury in the spinal cord in EAE. Brain 127(Part 2):294–303.
Craner MJ, Lo AC, Black JA, Waxman SG. 2003. Abnormal sodium channel
distribution in optic nerve axons in a model of inflammatory demyelination.
Brain 126(Part 7):1552–1561.
Craner MJ, Newcombe J, Black JA, Hartle C, Cuzner ML, Waxman SG.
2004b. Molecular changes in neurons in multiple sclerosis: altered axonal
expression of Nav1.2 and Nav1.6 sodium channels and Naþ/Ca2þ
exchanger. Proc Natl Acad Sci USA 101:8168–8173.
Derfuss T, Parikh K, Velhin S, Braun M, Mathey E, Krumbholz M, Kumpfel T,
Moldenhauer A, Rader C, Sonderegger P, et al. 2009. Contactin-2/TAG-1-
directed autoimmunity is identified in multiple sclerosis patients and mediates
gray matter pathology in animals. Proc Natl Acad Sci USA 106:8302–8307.
Feinberg K, Eshed-Eisenbach Y, Frechter S, Amor V, Salomon D, Sabanay H,
Dupree JL, Grumet M, Brophy PJ, Shrager P, et al. 2010. A glial signal
consisting of gliomedin and NrCAM clusters axonal Naþ channels during the
formation of nodes of Ranvier. Neuron 65:490–502.
Franklin RJ, Ffrench-Constant C, Edgar JM, Smith KJ. 2012. Neuroprotection
and repair in multiple sclerosis. Nat Rev Neurol 8:624–634.
Volume 61, No. 8
 Zoupi et al.: Juxtaparanodal Alterations in CNS-Demyelinating Models
Griffiths I, Klugmann M, Anderson T, Yool D, Thomson C, Schwab MH,
Schneider A, Zimmermann F, McCulloch M, Nadon N, et al. 1998. Axonal
swellings and degeneration in mice lacking the major proteolipid of myelin.
Science 280:1610–1613.
Groebe A, Clarner T, Baumgartner W, Dang J, Beyer C, Kipp M. 2009.
Cuprizone treatment induces distinct demyelination, astrocytosis, and microglia
cell invasion or proliferation in the mouse cerebellum. Cerebellum 8:163–174.
Haghighi SS, Pugh SL, Perez-Espejo MA, Oro JJ. 1995. Effect of 4-
aminopyridine in acute spinal cord injury. Surg Neurol 43:443–447.
Haider L, Fischer MT, Frischer JM, Bauer J, Hoftberger R, Botond G,
Esterbauer H, Binder CJ, Witztum JL, Lassmann H. 2011. Oxidative damage
in multiple sclerosis lesions. Brain 134(Part 7):1914–1924.
Howell OW, Palser A, Polito A, Melrose S, Zonta B, Scheiermann C, Vora AJ,
Brophy PJ, Reynolds R. 2006. Disruption of neurofascin localization reveals
early changes preceding demyelination and remyelination in multiple
sclerosis. Brain 129(Part 12):3173–3185.
Howell OW, Rundle JL, Garg A, Komada M, Brophy PJ, Reynolds R. 2010.
Activated microglia mediate axoglial disruption that contributes to axonal
injury in multiple sclerosis. J Neuropathol Exp Neurol 69:1017–1033.
Jukkola PI, Lovett-Racke AE, Zamvil SS, Gu C. 2012. Kþ channel alterations
in the progression of experimental autoimmune encephalomyelitis. Neurobiol
Dis 47:280–293.
Karimi-Abdolrezaee S, Eftekharpour E, Fehlings MG. 2004. Temporal and
spatial patterns of Kv1.1 and Kv1.2 protein and gene expression in spinal
cord white matter after acute and chronic spinal cord injury in rats:
Implications for axonal pathophysiology after neurotrauma. Eur J Neurosci
19:577–589.
Kipp M, Clarner T, Dang J, Copray S, Beyer C. 2009. The cuprizone animal
model: New insights into an old story. Acta Neuropathol 118:723–736.
Labasque M, Faivre-Sarrailh C. 2010. GPI-anchored proteins at the node of
Ranvier. FEBS Lett 584:1787–1792.
Lassmann H. 2001. Classification of demyelinating diseases at the interface
between etiology and pathogenesis. Curr Opin Neurol 14:253–258.
Lassmann H. 2003. Axonal injury in multiple sclerosis. J Neurol Neurosurg
Psychiatry 74:695–697.
Lassmann H. 2008. Models of multiple sclerosis: New insights into
pathophysiology and repair. Curr Opin Neurol 21:242–247.
Lassmann H, Bruck W, Lucchinetti C. 2001. Heterogeneity of multiple
sclerosis pathogenesis: Implications for diagnosis and therapy. Trends Mol
Med 7:115–121.
Lassmann H, Lucchinetti CF. 2008. Cortical demyelination in CNS
inflammatory demyelinating diseases. Neurology 70:332–333.
Leung YM. 2012. Involvement of C-type inactivation gating in the actions of
voltage-gated Kþ channel inhibitors. Pharmacol Ther 133:151–158.
Lucchinetti CF, Popescu BF, Bunyan RF, Moll NM, Roemer SF, Lassmann
H, Bruck W, Parisi JE, Scheithauer BW, Giannini C, et al. Inflammatory
cortical demyelination in early multiple sclerosis. N Engl J Med 2011;365:
2188–2197.
Markoullis K, Sargiannidou I, Gardner C, Hadjisavvas A, Reynolds R, Kleopa
KA. 2012. Disruption of oligodendrocyte gap junctions in experimental
autoimmune encephalomyelitis. Glia 60:1053–1066.
Mathey EK, Derfuss T, Storch MK, Williams KR, Hales K, Woolley DR, Al-Hayani
A, Davies SN, Rasband MN, Olsson T, et al. 2007. Neurofascin as a novel target
for autoantibody-mediated axonal injury. J Exp Med 204:2363–2372.
Mathis C, Denisenko-Nehrbass N, Girault JA, Borrelli E. 2001. Essential role
of oligodendrocytes in the formation and maintenance of central nervous
system nodal regions. Development 128:4881–4890.
Matsushima GK, Morell P. 2001. The neurotoxicant, cuprizone, as a model to
study demyelination and remyelination in the central nervous system. Brain
Pathol 11:107–116.
August 2013
Nashmi R, Jones OT, Fehlings MG. 2000. Abnormal axonal physiology is
associated with altered expression and distribution of Kv1.1 and Kv1.2 Kþ
channels after chronic spinal cord injury. Eur J Neurosci 12:491–506.
Pillai AM, Thaxton C, Pribisko AL, Cheng JG, Dupree JL, Bhat MA. 2009.
Spatiotemporal ablation of myelinating glia-specific neurofascin (Nfasc NF155)
in mice reveals gradual loss of paranodal axoglial junctions and concomitant
disorganization of axonal domains. J Neurosci Res 87:1773–1793.
Poliak S, Peles E. 2003. The local differentiation of myelinated axons at
nodes of Ranvier. Nat Rev Neurosci 4:968–980.
Poliak S, Salomon D, Elhanany H, Sabanay H, Kiernan B, Pevny L, Stewart CL,
Xu X, Chiu SY, Shrager P, et al. 2003. Juxtaparanodal clustering of Shaker-
like Kþ channels in myelinated axons depends on Caspr2 and TAG-1. J Cell
Biol 162:1149–1160.
Price JC, Kelley DE, Ryan CM, Meltzer CC, Drevets WC, Mathis CA, Mazumdar
S, Reynolds CFIII. 2002. Evidence of increased serotonin-1A receptor binding in
type 2 diabetes: A positron emission tomography study. Brain Res 927:97–103.
Prineas JW, Parratt JD. 2012. Oligodendrocytes and the early multiple
sclerosis lesion. Ann Neurol 72:18–31.
Rios JC, Rubin M, St Martin M, Downey RT, Einheber S, Rosenbluth J,
Levinson SR, Bhat M, Salzer JL. 2003. Paranodal interactions regulate
expression of sodium channel subtypes and provide a diffusion barrier for the
node of Ranvier. J Neurosci 23:7001–7011.
Savvaki M, Panagiotaropoulos T, Stamatakis A, Sargiannidou I, Karatzioula P,
Watanabe K, Stylianopoulou F, Karagogeos D, Kleopa KA. 2008. Impairment
of learning and memory in TAG-1 deficient mice associated with shorter CNS
internodes and disrupted juxtaparanodes. Mol Cell Neurosci 39:478–490.
Savvaki M, Theodorakis K, Zoupi L, Stamatakis A, Tivodar S, Kyriacou K,
Stylianopoulou F, Karagogeos D. 2010. The expression of TAG-1 in glial cells is
sufficient for the formation of the juxtaparanodal complex and the phenotypic
rescue of tag-1 homozygous mutants in the CNS. J Neurosci 30:13943–13954.
Sherman DL, Tait S, Melrose S, Johnson R, Zonta B, Court FA, Macklin WB,
Meek S, Smith AJ, Cottrell DF, et al. 2005. Neurofascins are required to
establish axonal domains for saltatory conduction. Neuron 48:737–742.
Sherratt RM, Bostock H, Sears TA. 1980. Effects of 4-aminopyridine on
normal and demyelinated mammalian nerve fibres. Nature 283:570–572.
Soulika AM, Lee E, McCauley E, Miers L, Bannerman P, Pleasure D. 2009.
Initiation and progression of axonopathy in experimental autoimmune
encephalomyelitis. J Neurosci 29:14965–14979.
Susuki K, Rasband MN. 2008. Molecular mechanisms of node of Ranvier
formation. Curr Opin Cell Biol 20:616–623.
Tait S, Gunn-Moore F, Collinson JM, Huang J, Lubetzki C, Pedraza L,
Sherman DL, Colman DR, Brophy PJ. 2000. An oligodendrocyte cell adhesion
molecule at the site of assembly of the paranodal axo-glial junction. J Cell
Biol 150:657–666.
Traka M, Dupree JL, Popko B, Karagogeos D. 2002. The neuronal adhesion
protein TAG-1 is expressed by Schwann cells and oligodendrocytes and is
localized to the juxtaparanodal region of myelinated fibers. J Neurosci 22:
3016–3024.
Traka M, Goutebroze L, Denisenko N, Bessa M, Nifli A, Havaki S, Iwakura Y,
Fukamauchi F, Watanabe K, Soliven B, et al. 2003. Association of TAG-1 with
Caspr2 is essential for the molecular organization of juxtaparanodal regions
of myelinated fibers. J Cell Biol 162:1161–1172.
Waxman SG, Craner MJ, Black JA. 2004. Naþ channel expression along
axons in multiple sclerosis and its models. Trends Pharmacol Sci 25:584–591.
Wolswijk G, Balesar R. 2003. Changes in the expression and localization of
the paranodal protein Caspr on axons in chronic multiple sclerosis. Brain
126(Part 7):1638–1649.
Zonta B, Tait S, Melrose S, Anderson H, Harroch S, Higginson J, Sherman
DL, Brophy PJ. 2008. Glial and neuronal isoforms of Neurofascin have
distinct roles in the assembly of nodes of Ranvier in the central nervous
system. J Cell Biol 181:1169–1177.
1249