RESEARCH ARTICLE

CSF-1-Activated Macrophages areTarget-Directed and Essential Mediators

of Schwann Cell Dedifferentiation andDysfunction in Cx32-Deficient Mice

Janos Groh, Ines Klein, Claudia Hollmann, Jennifer Wettmarshausen, Dennis Klein,

and Rudolf Martini

We investigated connexin 32 (Cx32)-deficient mice, a model for the X-linked form of Charcot-Marie-Tooth neuropathy(CMT1X), regarding the impact of low-grade inflammation on Schwann cell phenotype. Whereas we previously identified mac-rophages as amplifiers of the neuropathy, we now explicitly focus on the impact of the phagocytes on Schwann cell dediffer-entiation, a so far not-yet addressed disease-related mechanism for CMT1X. Using mice heterozygously deficient for Cx32and displaying both Cx32-positive and -negative Schwann cells in one and the same nerve, we could demonstrate that macro-phage clusters rather than single macrophages precisely associate with mutant but not with Cx32-positive Schwann cells. Sim-ilarly, in an advanced stage of Schwann cell perturbation, macrophage clusters were strongly associated with NCAM- and L1-positive, dedifferentiated Schwann cells. To clarify the role of macrophages regarding Schwann cell dedifferentiation, we gen-erated Cx32-deficient mice additionally deficient for the macrophage-directed cytokine colony-stimulating factor (CSF)21. Inthe absence of CSF-1, Cx32-deficient Schwann cells not only showed the expected amelioration in myelin preservation butalso failed to upregulate the Schwann cell dedifferentiation markers NCAM and L1. Another novel and unexpected finding inthe double mutants was the retained activation of ERK signaling, a pathway which is detrimental for Schwann cell homeostasisin myelin mutant models. Our findings demonstrate that increased ERK signaling can be compatible with the maintenance ofSchwann cell differentiation and homeostasis in vivo and identifies CSF-1-activated macrophages as crucial mediators of detri-mental Schwann cell dedifferentiation in Cx32-deficient mice.

GLIA 2015;63:977–986Key words: dedifferentiation, demyelination, NCAM, macrophage, ERK signaling, CMT1X

Introduction

We and others have demonstrated that in peripheral

nerves of patients suffering from Charcot-Marie-Tooth

type 1 (CMT1) neuropathies and in nerves of the related

mouse models Schwann cell dedifferentiation is a typical fea-

ture of nerve fiber damage (Fledrich et al., 2014; Guenard

et al., 1996; Hanemann et al., 1997; Hutton et al., 2011;

Klein et al., 2014). As a substantial mediator and amplifier of

nerve fiber damage, low-grade inflammation implicating

phagocytosing macrophages has been identified in various

mouse models for CMT1 (Martini et al., 2013). Particularly,

the fibroblast-borne cytokine colony stimulating factor-1

(CSF-1) plays a predominant role as mediator of pathogenic

inflammation in CMT1 models, but the impact of CSF-1-

stimulated macrophages on Schwann cell dedifferentiation

fate has not yet been addressed so far. It is conceivable that

upregulation of dedifferentiation markers is either directly

caused by Schwann cell-intrinsic mechanisms downstream of

the mutation or is secondarily mediated/amplified by low-

grade inflammation. The present study is designed to discrim-

inate between these two possibilities in an established model

for the X-linked dominant form of CMT neuropathies,

CMT1X, caused by mutations in the X-chromosomal gene

for the gap junction tunnel protein connexin 32 (also

View this article online at wileyonlinelibrary.com. DOI: 10.1002/glia.22796

Published online January 27, 2015 in Wiley Online Library (wileyonlinelibrary.com). Received Oct 16, 2014, Accepted for publication Jan 9, 2015.

Address correspondence to Rudolf Martini, Department of Neurology, Developmental Neurobiology, University Hospital Wuerzburg, Josef-Schneider-Str. 11,

D-97080, Wuerzburg. E-mail: rudolf.martini@uni-wuerzburg.de

From the Department of Neurology, Developmental Neurobiology, University Hospital Wuerzburg, Wuerzburg

VC 2015 Wiley Periodicals, Inc. 977

designated GJB1; (Bouhy and Timmerman, 2013; Fledrich

et al., 2012)). The model mice, here named Cx32def mice,

initially show largely normal myelin formation but develop a

late-onset and slowly progressing demyelinating neuropathy

that is accompanied by axonal perturbation, electrophysiologi-

cal alterations and reduced muscle strength (Anzini et al.,

1997; Groh et al., 2010). Here we show that the upregulation

of the established dedifferentiation markers NCAM and L1

by Cx32-deficient Schwann cells is strictly dependent on

CSF-1-mediated macrophage activation. In contrast, increased

ERK-phosphorylation, another immaturity marker and

disease-promotor in nerve pathologies (Fledrich et al., 2014;

Napoli et al., 2012), is inflammation-independent in its ori-

gin and fails to mediate Schwann cell perturbation in Cx32

mutants in the absence of CSF-1-related macrophage activity.

Materials and Methods

ANIMALS. Cx32wt, Cx32het, and Cx32def mice (Nelles

et al., 1996) as well as CSF-1op mice (Yoshida et al., 1990)

were on a uniform C57BL/6N genetic background and mice

were crossbred and genotyped according to previously pub-

lished protocols (Carenini et al., 2001; Groh et al., 2012;

Kobsar et al., 2003). Mice were kept in the animal facility of

the Department of Neurology under barrier conditions and

all experiments were approved by the local authority (Regier-

ung von Unterfranken).

Nerve PreparationFor preparation of cryo-sections, animals were killed by

asphyxiation with CO2 (according to guidelines by the State

Office of Health and Social Affairs Berlin), blood was rinsed

with phosphate buffered saline (PBS) containing heparin,

femoral quadriceps nerves were excised, processed as described

(Klein et al., 2014) and cut into 10-lm thick cross-sections

or longitudinal sections on a cryostat (Leica).

For western blot analysis sciatic nerves were quickly dis-

sected, snap frozen in liquid nitrogen and stored at 280�C.

For teased fiber preparations, mice were transcardially

perfused with 2% paraformaldehyde (PFA) in PBS for 10

min. Quadriceps and sciatic nerves were dissected, perineu-

rium was stripped off, and nerve fibers were loosely separated

by forceps on glass slides and air dried.

For immunoelectron microscopy, mice were transcar-

dially perfused with 4% PFA in cacodylate buffer. Femoral

quadriceps nerves and ventral roots were dissected and post-

fixed over night in the same fixative.

ImmunohistochemistryFor immunohistochemistry, fresh frozen nerve sections or

teased fibers were postfixed in acetone (10 min, 220�C) and

incubated with 5% BSA in 0.1M PBS for 30 min at room

temperature to block unspecific binding sites. Afterwards the

respective primary antibodies (mouse anti-Cx32, 1:100, Invi-

trogen; rat anti-F4/80, 1:300, Serotec; rabbit anti-NCAM,

1:200, R&D Systems, rabbit anti-L1, 1:1,000, kindly pro-

vided by Melitta Schachner, chicken anti-P0, 1:500, Acris;

rabbit anti-pERK, 1:100, Cell Signaling, mouse anti-CNPase,

1:500, Sigma-Aldrich, rabbit anti-CCL2, 1:50, Peprotech)

were incubated over night at 4�C in 1% BSA in 0.1M PBS

and detected by corresponding secondary antibodies (goat

anti-chicken Alexa Fluor 488, 1:300, Invitrogen; goat anti-

mouse Cy2, 1:300, Dianova; goat anti-rat Cy3, 1:300, Dia-

nova; goat anti-rabbit Cy3, 1:300, Dianova; goat anti-mouse

Cy5, 1:500, Dianova; goat anti-rat Cy5, 1:500, Dianova). For

teased fiber preparations, all stainings were performed with

the addition of 0.1 to 0.3% TritonX-100. Nuclei were stained

with DAPI (Sigma-Aldrich) and samples were mounted with

Aqua-Poly/Mount (Polysciences) and investigated on an Axio-

phot 2 epifluorescence microscope (Zeiss) or a FluoView

FV1000 confocal microscope (Olympus).

Electron and Immunoelectron MicroscopyImmunoelectron microscopy against NCAM was performed

as previously described (Klein et al., 2014). Nerves were

osmificated and processed for light and electron microscopy

and morphometric quantification of neuropathological altera-

tions was performed as published (Groh et al., 2012)

Western Blot AnalysisSnap frozen sciatic nerves were sonicated (Sonoplus HD60,

Bandelin electronic) in 100 mL RIPA lysis buffer (25 mM

Tris HCl pH 8, 10 mM Hepes, 150 mM NaCl, 145 mM

KCl, 5 mM MgCl2, 2 mM EDTA, 0.1% sodium dodecyl

sulfate, 1% NP-40, 10% glycerol) containing protease inhibi-

tors. Protein concentration was determined by Lowry assay

(Sigma-Aldrich) and proteins were resolved by SDS–poly-

acrylamide gel electrophoresis, transferred to nitrocellulose

membranes and visualized using Ponceau S (Roth). Mem-

branes were blocked with skimmed milk and probed with

antibody solution overnight at 4�C (mouse anti-pERK1/2,

1:1,000, Santa Cruz; rabbit anti-ERK1/2, 1:10,000, Santa

Cruz). Incubation with horseradish peroxidase-conjugated sec-

ondary antibodies was performed for 1 h at room tempera-

ture and detection of the immune reaction was achieved by

use of ECL reagent and ECL hyperfilm (GE Healthcare Bio-

Sciences AB). Sequential stainings were performed after incu-

bating the nitrocellulose membrane with stripping buffer

(0.2 M glycine, 0.1% SDS, 10 mM dithiothreitol, and 1%

Tween) for 60 min.

Statistical AnalysisAll quantifications were performed in a blinded manner, with

the investigators unaware of the genotypes of the analyzed

978 Volume 63, No. 6

mice. Statistical analyses were performed using PASW Statis-

tics 18 (SPSS, IBM) software. Data was controlled for normal

distribution by Shapiro–Wilk test and group differences were

either tested by unpaired two-tailed Student’s t-test (paramet-

ric) or by Mann–Whitney U test (nonparametric). For multi-

ple comparisons, one-way ANOVA followed by Tukey’s post

hoc test (parametric) or Bonferroni correction of Mann-

Whitney U tests (nonparametric) was used. Significance levels

were indicated as follows: *P< 0.05, **P< 0.01,

***P< 0.001.

Results

At first, we wanted to score the relationship between macro-

phage activation and Schwann cell dedifferentiation as a

hallmark of Schwann cell perturbation. In order to investi-

gate the spatial association of perturbed nerve fibers and

macrophages in detail, we chose the neatly arranged situa-

tion in which only a subpopulation of Schwann cells is

pathologically affected, as is the case in female mice hetero-

zygously deficient for Cx32 (Cx32het; Fig. 1A,D) (Scherer

et al., 1998). Here we could show that there was a marked

correlation between Cx32-negativity and the close apposition

of macrophage clusters, comprising 3 to 14 F4/80-positive

cells (Fig. 1B,D). No correlation, however, was seen between

the genetic Cx32-status of Schwann cells and the vicinity to

occasionally occurring single macrophages of possibly resi-

dent, nonactivated nature (Fig. 1C). As late stage dedifferen-

tiation markers and molecules faithfully identifying previous

Schwann cell damage, we chose antibodies to the cell adhe-

sion molecules NCAM and L1 that have been demonstrated

to be strongly expressed by Schwann cells which have lost

their myelin and by supernumerary Schwann cells of

Cx32def mice (Klein et al., 2014). At first, these markers

expectedly indicated fewer dedifferentiated Schwann cell

FIGURE 1: Macrophages form clusters along fiber segments associated with Cx32-negative Schwann cells in Cx32het mice. Immunocyto-chemistry against Cx32 in combination with F4/80 and P0 on teased fiber preparations of femoral quadriceps nerves from 8-month-oldCx32het mice (n 5 4). A, Cx32-positive (arrow) and Cx32-negative Schwann cells were identified based on their signal intensity at theparanodal loops. B, Macrophage clusters containing 3 to 14 macrophages were exclusively found along Cx32-negative fiber segments.C, Single macrophages were associated with both Cx32-negative and Cx32-positive fibers with a similar frequency. Macrophages wereconsidered to be associated with fibers when found in direct contact or less than 5 mm distance to the fiber. D, Representative exampleof a macrophage cluster showing preferential association with a Cx32-negative Schwann cell instead of a neighboring Cx32-positive(arrow) Schwann cell. Scale bars: 10 mm.

Groh et al.: Cx32-Deficient Schwann Cells and Macrophages

June 2015 979

profiles in Cx32het mutants than in Cx32def mice (Fig.

2A). Triple-immunofluorescence on longitudinal frozen sec-

tions of Cx32het mutants, using the dedifferentiation

markers and the myelin/differentiation marker P0 on the

one hand and the macrophage marker F4/80 on the other,

confirmed the above-mentioned association of macrophage

clusters with perturbed Schwann cells at a putatively

advanced stage of nerve fiber damage (Fig. 2B). This pre-

dominant association of macrophages with dedifferentiated

Schwann cell structures was confirmed by double-

immunofluorescence with NCAM or L1 and F4/80 anti-

bodies on cross sections (around 60% of all macrophages

were associated with NCAM1 profiles; Fig. 2C) and

immunoelectron microscopy (Fig. 3).

To clarify the putatively important role of macrophages

in the context of Schwann cell perturbation and dedifferentia-

tion, we prevented macrophage activation by cross-breeding

the Cx32 mutants with osteopetrotic (op) mice deficient for

functional CSF-1 (Yoshida et al., 1990). As expected, increase

of macrophage numbers was similarly blocked in Cx32het/

CSF-1op mutants as previously observed in Cx32def/CSF-

1op mutants (Groh et al., 2012) (Fig. 4A) and abnormally

myelinated profiles, identified by electron microscopy, were

also strongly reduced (Fig. 4B,C). Additionally, the concomi-

tant expression of the Schwann cell dedifferentiation markers

NCAM and L1 was strongly attenuated in the absence of

macrophage activation, both in the Cx32het and Cx32def

mutants (Fig. 5A,B).

FIGURE 2: Macrophage clusters are associated with dedifferentiated NCAM-positive Schwann cells in Cx32het mice. A, Immunofluores-cent labelings of NCAM in cross-sections of femoral quadriceps nerves from Cx32wt, Cx32het, and Cx32def mice at eight months ofage showed that clusters of Remak fibers were strongly labeled in all genotypes. There was a prominent upregulation of NCAM expres-sion in nerves of Cx32def mice (right) and a weaker upregulation in nerves of Cx32het mice (middle). Scale bar: 50 mm. B, Labelings ofNCAM in combination with P0 and F4/80 in longitudinal sections showed that NCAM upregulation occurs at fibers associated with mac-rophage clusters. Scale bar: 10 mm. C, Endoneurial macrophages were observed in contact (arrows) with dedifferentiated Schwann cellprofiles in cross-sections of femoral quadriceps nerves from 8-month-old Cx32het mice using immunohistochemistry against F4/80 andNCAM as well as F4/80 and L1 (Scale bar: 20 mm).

980 Volume 63, No. 6

After having shown that macrophage activation by CSF-

1 is essential for Schwann cell damage and dedifferentiation

in Cx32 mutant mice, we investigated the impact of macro-

phage activation on ERK phosphorylation, representing an

intrinsic signaling pathway common to various Schwann cell

reactions. ERK signaling has previously been reported to be

implicated in Schwann cell dedifferentiation and proinflam-

matory cytokine expression in a transgenic mutant with

inducible Raf-kinase activity (Napoli et al., 2012) and in

inhibition of Schwann cell maturation in PMP22 overexpress-

ing rats (Fledrich et al., 2014; Martini, 2014). Our group has

previously demonstrated that ERK signaling is of pathoge-

netic relevance and results in increased expression of the

macrophage-directed cytokine CCL2 in distinct mouse mod-

els of CMT1 (Fischer et al., 2008; Groh et al., 2010; Kohl

et al., 2010; Martini et al., 2013). In Cx32def mice, i.e., in

Cx32 mutants in which all Schwann cells are genetically

devoid of the gap junction protein, we found that in the

absence of CSF-1-dependent macrophage activation, morpho-

logically intact and differentiated Schwann cells retained

increased levels of ERK-phosphorylation as demonstrated by

Western blot analysis (Fig. 6A,B) and immunofluorescence on

teased CNPase-positive (but L1 and NCAM negative)

Schwann cells (Fig. 6C). In addition, increased expression of

CCL2 in Cx32def Schwann cells was not attenuated by CSF-

1 deficiency (Fig. 6D) This is in marked contrast to wt mice

where, expectedly, Schwann cells are morphologically also

intact, but show low levels of ERK activation and CCL2

expression. Thus, macrophage activation is needed for

Schwann cell perturbation and concomitantly increased

NCAM and L1 expression, whereas ERK-phosphorylation in

Schwann cells in the absence of CSF-1 is not sufficient to

mediate perturbation of Schwann cell function and differen-

tiation in Cx32def mice.

Discussion

We show that in mice heterozygously deficient for Cx32 and

containing both Cx32-positive and Cx32-negative Schwann

cells in peripheral nerves, activated macrophages “sense” the

Cx32-deficient Schwann cells and form aggregates in their

close vicinity, whereas neighboring Cx32-positive Schwann

cells are not associated with macrophage clusters, but only

occasionally with single resident cells. Accordingly, in a situa-

tion representing a more advanced stage of Schwann cell per-

turbation, macrophage clusters are in apposition with

dedifferentiated, NCAM- and L1-positive Schwann cells but

not with marker-negative myelinating Schwann cells. Thus,

we could experimentally demonstrate for the first time that

macrophage-related Schwann cell perturbation is the conse-

quence of a spatially focused “attack” by the macrophages

and not the result of a generally gliotoxic milieu in the endo-

neurium of an inflamed nerve. In this context, it is worth-

while to consider the possible underlying communication

pathways that direct the macrophages to their targets, i.e., the

mutant Schwann cells. A candidate molecule is mutant-

Schwann cell-borne CCL2 which is expressed downstream of

FIGURE 3: Macrophages are associated with dedifferentiated NCAM-positive Schwann cells in Cx32def mice. A. Immunoelectron micros-copy of ultrathin longitudinal and B, C, cross-sections of femoral quadriceps nerves from Cx32def mice revealed NCAM-immunoreactivity (electron-dense membrane profiles, indicated by arrowheads) on Schwann cells (SC) deprived of myelin (A, C) or in theprocess of demyelination (B). Note the close vicinity or direct contact of morphologically identified macrophages (MA) containing myelindebris (My) and the proximity of fibroblasts (Fi), the previously identified source of CSF-1. Disrupted basal lamina might indicate macro-phage entry into endoneurial Schwann cell tubes (arrows); ax: axon. Scale bars: 1 mm.

Groh et al.: Cx32-Deficient Schwann Cells and Macrophages

June 2015 981

MEK-ERK signaling (Groh et al., 2010). The observation

that genetically reduced CCL2-expression or the pharmaco-

logical blockade of MEK-ERK activation ameliorates the neu-

ropathy in Cx32def mice supports this hypothesis (Groh

et al., 2010). Thus, with regard of Schwann cell perturbation

in Cx32def mice, we suppose a two-step scenario in that the

Cx32def Schwann cells attract macrophages into their vicinity,

likely due to the down-stream mediators of MEK-ERK acti-

vation (e.g. CCL2). Macrophages are then eventually trig-

gered for damage by fibroblast-borne CSF-1 (Groh et al.,

2012). Possibly, one of the functions of CSF-1 is a focal pro-

liferation of macrophages in the vicinity of genetically abnor-

mal Schwann cells as macrophage clusters rather than single

macrophages are closely associated with Cx32def Schwann

cells or “post-damage” NCAM and L1-positive glial cells.

This view is in line with the observation that such clusters do

not form in the absence of CSF-1 (this study) and that CSF-

1 has been identified as the major promotor of proliferation

of resident macrophages in genetically diseased peripheral

nerves (Muller et al., 2007).

While our present study suggests that ERK activation

and subsequent CCL2 upregulation is involved in the recruit-

ment of macrophages to their prospective targets, it clearly

demonstrates that activation of this pathway is insufficient to

cause Schwann cell dedifferentiation and perturbation in the

absence of CSF-1-mediated macrophage activation. Interest-

ingly, in isolated normal rat Schwann cells, ERK phosphoryla-

tion is compatible with a preserved differentiation state,

provided that cAMP levels are high (Monje et al., 2010).

With regard of these and our present observations the patho-

genic role of ERK-activation in other models of nerve pertur-

bation is worthwhile to consider: in a transgenic mouse

mutant carrying an inducible Raf-kinase transgene in myelin-

ated Schwann cells, ERK activation in otherwise normal

Schwann cells causes macrophage activation and severe demy-

elination (Napoli et al., 2012). It would be interesting to

FIGURE 4: CSF-1 deficiency prevents increase of macrophage numbers and demyelination in Cx32het mice. A, Quantification of F4/801 mac-rophages in femoral quadriceps nerves of 8-month-old Cx32wt/CSF-1wt, Cx32het/CSF-1wt, and Cx32het/CSF-1op mice (n 5 4 per group) byimmunohistochemistry. *P < 0.05. One-way ANOVA followed by Tukey’s post hoc test. B, Quantification of abnormally myelinated axons usingelectron microscopy revealed strongly reduced fiber damage in the absence of CSF-1 (op). C, Representative electron micrographs of femoralquadriceps nerves from 8-month-old Cx32wt/CSF-1wt, Cx32het/CSF-1wt, and Cx32het/CSF-1op mice. Note reduced fiber damage in theabsence of CSF-1 (op) in Cx32het mice. Bonferroni corrected Mann-Whitney U tests. *P < 0.05. Scale bar: 5 mm.

982 Volume 63, No. 6

investigate whether in this model induction of the MEK-ERK

pathway would have the same effect in the absence of CSF-1.

Recent observations by another group show that a pathogenic

consequence of overexpression of PMP22 in rats, an estab-

lished model of CMT1A, leads to a persistent MEK-ERK

activation at the expense of the antagonistic PI3K-Akt matu-

ration pathway (Fledrich et al., 2014; Martini, 2014). Tip-

ping the balance in favor of the maturation pathway and

attenuating MEK-ERK activation by soluble neuregulin-1 led

to an amelioration of pathology (Fledrich et al., 2014). Also

there, the role of CSF-1 is still obscure. One significant dif-

ference between the PMP22 overexpression and Cx32-

deficiency in Schwann cells is that in the first case Schwann

cells show an early and persistent immature phenotype (Fle-

drich et al., 2014; Klein et al., 2014) whereas in the case of

Cx32-deficiency, Schwann cells mature almost normally and

show a later onset dedifferentiation as reflected by upregula-

tion of NCAM and L1 (this study and (Klein et al., 2014)).

It is, therefore, difficult to predict whether in CMT1A mod-

els, CSF-1 plays a similarly essential role as in Cx32def mice

and P0het mutants (Carenini et al., 2001; Groh et al., 2012).

CSF-1-activated macrophages mediate dedifferentiation

of Cx32def Schwann cells (this study) resulting in robust

amplification of neuropathy (Groh et al., 2012). This may

appear initially unexpected since CSF-1 is known to polarize

macrophages into an M2 rather than into an M1 phenotype,

the first of which is anti-inflammatory and supposed to pro-

mote repair and reorganization in injured nerves (Ydens et al.,

2012) and in other tissues (Hamilton, 2008; Jones and

Ricardo, 2013). Indeed, detrimental macrophages in Cx32def

mice carry some M2 markers (Klein and Martini, unpub-

lished data) and studies from other groups have shown that

alternatively polarized, M2-like macrophages show much

more qualities shared with M1 cells than previously antici-

pated (Gordon and Martinez, 2010; Qian and Pollard, 2010;

Shechter and Schwartz, 2013). Importantly, once associated

with macrophage clusters, the Schwann cells acquire a dedif-

ferentiated phenotype as reflected by their upregulation of the

cell adhesion molecules NCAM and L1 and possibly others

(Klein et al., 2014). A similar dedifferentiating impact of

CSF-1-activated macrophages might occur in P0het mutants,

as also there CSF-1 is involved in nerve damage (Carenini

FIGURE 5: CSF-1 deficiency prevents upregulation of dedifferentiation markers in peripheral nerves of Cx32het and Cx32def mice.Quantification (excluding Remak clusters) of A, NCAM-positive profiles and B, L1-positive profiles in femoral quadriceps nerves showsthat, both in Cx32het and Cx32def mice, dedifferentiation is dependent on CSF-1 expression (n 5 4 per group). *P < 0.05, **P < 0.01,***P < 0.001. One-way ANOVA followed by Tukey’s post hoc test.

Groh et al.: Cx32-Deficient Schwann Cells and Macrophages

June 2015 983

et al., 2001) and NCAM and L1 are upregulated on demyeli-

nated Schwann cells and supernumerary Schwann cells (Gue-

nard et al., 1996; Klein et al., 2014). According to a

traditional view, dedifferentiation of mutant Schwann cells is

the consequence of demyelination. Our findings may, how-

ever, favor the more provocative idea that the macrophages

first induce a dedifferentiation pathway in Schwann cells

which ends up in demyelination, myelin phagocytosis and

subsequent Schwann cell mitosis. It is conceivable that factors

released by CSF-1-activated macrophages may lead to cAMP

FIGURE 6: CSF-1 deficiency does not affect activation of ERK1/2 and expression of CCL2 in myelinating Schwann cells of Cx32def mice. A,Western blot analysis of sciatic nerve lysates from 6-month-old Cx32wt/CSF-1wt, Cx32def/CSF-1wt and Cx32def/CSF-1op mice using p-ERK1/2- and ERK1/2-specific antibodies. Note similarly increased intensities of pERK signal in Cx32def mice (in comparison to Cx32wtmice), independent of CSF-1 expression. B, Densitometric quantification of ERK1/2 phosphorylation related to total ERK1/2 expressiondemonstrated increased levels in both Cx32def/CSF-1wt and Cx32def/CSF-1op nerves in comparison with Cx32wt nerves (n 5 4 pergroup). *P < 0.05. One-way ANOVA followed by Tukey’s post hoc test. C, Immunocytochemical detection of pERK1/2 in combination withCNPase in single teased fiber preparations of sciatic nerves from 6-month-old Cx32wt/CSF-1wt, Cx32def/CSF-1wt, and Cx32def/CSF-1opmice. Increased pERK immunoreactivity (arrowheads) was detected in CNPase-positive Schmidt-Lanterman incisures (arrows) of myelinatingSchwann cells of Cx32def but not Cx32wt mice, independent of CSF-1 expression. D, Immunocytochemical detection of CCL2 in singleteased fiber preparations of sciatic nerves from 6-month-old Cx32wt/CSF-1wt, Cx32def/CSF-1wt, and Cx32def/CSF-1op mice. IncreasedCCL2 immunoreactivity (arrows) in Schmidt-Lanterman incisures of Cx32def mice was not affected by CSF-1 deficiency. Scale bars: 10 mm.

984 Volume 63, No. 6

reduction in Schwann cells (Monje et al., 2010), possibly as

downstream reaction of p38 MAPK elevation in Schwann

cells (Yang et al., 2012). Additionally, myelin-phagocytosing

macrophages have been shown to release mitogenic factors for

cultured Schwann cells (Baichwal et al., 1988). In this con-

text, it is tempting to speculate that in the mutant nerve,

macrophages polarized by CSF-1 into an M2 phenotype may

partially acquire characteristics of tumor-associated macro-

phages (Chanmee et al., 2014; Mantovani and Sica, 2010) or

tumor-associated microglial cells (Charles et al., 2012). These

tumor-associated macrophage variants not only share their

dependency on CSF-1 and their M2 polarization with nerve

macrophages but also express a plethora of growth factors and

other molecules that foster proliferation of tumor cells (Man-

tovani and Sica, 2010; Qian and Pollard, 2010). We speculate

that the nerve macrophages are, in analogy to the tumor mac-

rophages, directly involved in driving dedifferentiation and

mitosis of mutant Schwann cells. While the mutation primar-

ily results in an increased activation of ERK signaling, this

pathway alone is insufficient to cause dedifferentiation in the

absence of secondary CSF-1-stimulated macrophage reactions.

Further studies will be necessary to determine the exact mech-

anisms of how CSF-1-triggered M2-macrophages mediate

dedifferentiation of mutant Schwann cells and thereby con-

tribute to damage of nerve structure and function.

Acknowledgment

Grant sponsor: German Research Council; Grant number:

MA 1053/6-1; Grant sponsor: Interdisciplinary Centre for

Clinical Research (IZKF) of the University of Wuerzburg;

Grant number: A-122

The authors are grateful to Heinrich Blazyca, Silke Loserth,

and Bettina Meyer for expert technical assistance and Helga

Br€unner, Jacqueline Schreiber, Anja Weidner, and Jennifer

Bauer for attentive care of mice. The authors thank Steve S.

Scherer (Philadelphia) for valuable discussions.

ReferencesAnzini P, Neuberg DHH, Schachner M, Nelles E, Willecke K, Zielasek J, ToykaKV, Suter U, Martini R. 1997. Structural abnormalities and deficient mainte-nance of peripheral nerve myelin in mice lacking the gap junction proteinconnexin 32. J Neurosci 17:4545–4551.

Baichwal RR, Bigbee JW, DeVries GH. 1988. Macrophage-mediated myelin-related mitogenic factor for cultured Schwann cells. Proc Natl Acad Sci USA85:1701–1705.

Bouhy D, Timmerman V. 2013. Animal models and therapeutic prospects forCharcot-Marie-Tooth disease. Ann Neurol 74:391–396.

Carenini S, Maurer M, Werner A, Blazyca H, Toyka KV, Schmid CD, RaivichG, Martini R. 2001. The role of macrophages in demyelinating peripheralnervous system of mice heterozygously deficient in p0. J Cell Biol 152:301–308.

Chanmee T, Ontong P, Konno K, Itano N. 2014. Tumor-associated macro-phages as major players in the tumor microenvironment. Cancer 6:1670–1690.

Charles NA, Holland EC, Gilbertson R, Glass R, Kettenmann H. 2012. Thebrain tumor microenvironment. Glia 60:502–514.

Fischer S, Weishaupt A, Troppmair J, Martini R. 2008. Increase of MCP-1(CCL2) in myelin mutant Schwann cells is mediated by MEK-ERK signalingpathway. Glia 56:836–843.

Fledrich R, Stassart RM, Klink A, Rasch LM, Prukop T, Haag L, Czesnik D,Kungl T, Abdelaal TA, Keric N, Stadelmann C, Br€uck W, Nave KA, SeredaMW. 2014. Soluble neuregulin-1 modulates disease pathogenesis in rodentmodels of Charcot-Marie-Tooth disease 1A. Nat Med 20:1055–1061.

Fledrich R, Stassart RM, Sereda MW. 2012. Murine therapeutic models forCharcot-Marie-Tooth (CMT) disease. Br Med Bull 102:89–113.

Gordon S, Martinez FO. 2010. Alternative activation of macrophages: Mecha-nism and functions. Immunity 32:593–604.

Groh J, Heinl K, Kohl B, Wessig C, Greeske J, Fischer S, Martini R. 2010.Attenuation of MCP-1/CCL2 expression ameliorates neuropathy in a mousemodel for Charcot-Marie-Tooth 1X. Hum Mol Genet 19:3530–3543.

Groh J, Weis J, Zieger H, Stanley ER, Heuer H, Martini R. 2012. Colony-stim-ulating factor-1 mediates macrophage-related neural damage in a model forCharcot-Marie-Tooth disease type 1X. Brain 135:88–104.

Guenard V, Montag D, Schachner M, Martini R. 1996. Onion bulb cells inmice deficient for myelin genes share molecular properties with immature,differentiated non-myelinating, and denervated Schwann cells. Glia 18:27–38.

Hamilton JA. 2008. Colony-stimulating factors in inflammation and autoim-munity. Nat Rev Immunol 8:533–544.

Hanemann CO, Gabreels-Festen AA, Stoll G, Muller HW. 1997. Schwann celldifferentiation in Charcot-Marie-Tooth disease type 1A (CMT1A): Normal num-ber of myelinating Schwann cells in young CMT1A patients and neural celladhesion molecule expression in onion bulbs. Acta Neuropathol 94:310–315.

Hutton EJ, Carty L, Laura M, Houlden H, Lunn MP, Brandner S, Mirsky R,Jessen K, Reilly MM. 2011. c-Jun expression in human neuropathies: a pilotstudy. J Peripheral Nerv Syst 16:295–303.

Jones CV, Ricardo SD. 2013. Macrophages and CSF-1: implications for devel-opment and beyond. Organogenesis 9:249–260.

Klein D, Groh J, Wettmarshausen J, Martini R. 2014. Nonuniform molecularfeatures of myelinating Schwann cells in models for CMT1: Distinct diseasepatterns are associated with NCAM and c-Jun upregulation. Glia 62:736–750.

Kobsar I, Berghoff M, Samsam M, Wessig C, Maurer M, Toyka KV, Martini R.2003. Preserved myelin integrity and reduced axonopathy in connexin32-deficient mice lacking the recombination activating gene-1. Brain 126:804–813.

Kohl B, Fischer S, Groh J, Wessig C, Martini R. 2010. MCP-1/CCL2 modifiesaxon properties in a PMP22-overexpressing mouse model for Charcot-Marie-Tooth 1A neuropathy. Am J Pathol 176:1390–1399.

Mantovani A, Sica A. 2010. Macrophages, innate immunity and cancer: Bal-ance, tolerance, and diversity. Curr Opin Immunol 22:231–237.

Martini R. 2014. Neuregulin-1 alleviates Charcot-Marie-Tooth disease in rats.Nat Med 20:984–985.

Martini R, Klein D, Groh J. 2013. Similarities between inherited demyelinatingneuropathies and Wallerian degeneration: An old repair program may causemyelin and axon perturbation under nonlesion conditions. Am J Pathol 183:655–660.

Monje PV, Soto J, Bacallao K, Wood PM. 2010. Schwann cell dedifferentia-tion is independent of mitogenic signaling and uncoupled to proliferation:role of cAMP and JNK in the maintenance of the differentiated state. J BiolChem 285:31024–31036.

Muller M, Berghoff M, Kobsar I, Kiefer R, Martini R. 2007. Macrophage col-ony stimulating factor is a crucial factor for the intrinsic macrophageresponse in mice heterozygously deficient for the myelin protein P0. ExpNeurol 203:55–62.

Napoli I, Noon LA, Ribeiro S, Kerai AP, Parrinello S, Rosenberg LH, CollinsMJ, Harrisingh MC, White IJ, Woodhoo A, Lloyd AC. 2012. A central role forthe ERK-signaling pathway in controlling Schwann cell plasticity and periph-eral nerve regeneration in vivo. Neuron 73:729–742.

Groh et al.: Cx32-Deficient Schwann Cells and Macrophages

June 2015 985

Nelles E, B€utzler C, Jung D, Temme A, Gabriel HD, Dahl U, Traub O,St€umpel F, Jungermann K, Zielasek J, Toyka KV, Dermietzel R, Willecke K.1996. Defective propagation of signals generated by sympathetic nerve stim-ulation in the liver of connexin32-deficient mice. Proc Natl Acad Sci USA 93:9565–9570.

Qian BZ, Pollard JW. 2010. Macrophage diversity enhances tumor progres-sion and metastasis. Cell 141:39–51.

Scherer SS, Xu YT, Nelles E, Fischbeck K, Willecke K, Bone LJ. 1998. Con-nexin32-null mice develop demyelinating peripheral neuropathy. Glia 24:8–20.

Shechter R, Schwartz M. 2013. Harnessing monocyte-derived macrophagesto control central nervous system pathologies: no longer ’if’ but ’how’.J Pathol 229:332–346.

Yang DP, Kim J, Syed N, Tung YJ, Bhaskaran A, Mindos T, Mirsky R, JessenKR, Maurel P, Parkinson DB, Kim HA. 2012. p38 MAPK activation promotesdenervated Schwann cell phenotype and functions as a negative regulator ofSchwann cell differentiation and myelination. J Neurosci 32:7158–7168.

Ydens E, Cauwels A, Asselbergh B, Goethals S, Peeraer L, Lornet G,Almeida-Souza L, Van Ginderachter JA, Timmerman V, Janssens S. 2012.Acute injury in the peripheral nervous system triggers an alternative macro-phage response. J Neuroinflamm 9:176.

Yoshida H, Hayashi S-I, Kunisada T, Ogawa M, Nishikawa S, Okamura H,Sudo T, Shultz LD, Nishikawa S-I. 1990. The murine mutation osteopetrosis isin the coding region of the macrophage colony stimulating factor gene.Nature 345:442–444.

986 Volume 63, No. 6

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具