www.elsevier.com/locate/ynbdi

Neurobiology of Disease 19 (2005) 171–182

PDGF and FGF2 pathways regulate distinct oligodendrocyte

lineage responses in experimental demyelination with

spontaneous remyelination

Joshua C. Murtie,a,1 Yong-Xing Zhou,b Tuan Q. Le,b

Adam C. Vana,c and Regina C. Armstronga,b,c,*

aProgram in Molecular and Cell Biology, Uniformed Services University of the Health Sciences, Bethesda, MD 20814, USAbDepartment of Anatomy, Physiology, Genetics, Uniformed Services University of the Health Sciences, Bethesda, MD 20814, USAcProgram in Neuroscience, Uniformed Services University of the Health Sciences, Bethesda, MD 20814, USA

Received 15 July 2004; revised 28 October 2004; accepted 10 December 2004

Available online 12 February 2005

Repair of myelin damage in the adult CNS requires oligodendrocyte

progenitor (OP) proliferation and subsequent differentiation into

remyelinating oligodendrocytes. Platelet-derived growth factor (PDGF)

and fibroblast growth factor-2 (FGF2) have been predicted to act

individually and/or cooperatively to generate remyelinating oligoden-

drocytes. Analysis of PDGF alpha receptor (PDGFaR) heterozygous

(+/�) mice indicates that PDGFaR expression modulates oligodendro-

cyte density in non-lesioned adult CNS. Analysis of cuprizone

demyelination and recovery in PDGFaR+/� mice, FGF2 knockout

(�/�) mice, and PDGFaR+/� FGF2�/� mice demonstrated that:( 1)

OP proliferation and oligodendrocyte regeneration is impaired in

PDGFaR heterozygotes, (2) PDGFaR+/� and FGF2�/� deletions do

not act cooperatively to impair OP amplification, (3) oligodendrocyte

differentiation is more frequent in FGF2�/� mice, and (4) FGF2

deletion in combination with the PDGFaR+/� genotype rescues

impaired oligodendrocyte regeneration of PDGFaR heterozygotes.

These findings demonstrate distinct roles for PDGF and FGF2 in vivo in

the context of a demyelinating disease with spontaneous remyelination.

Published by Elsevier Inc.

Keywords: Demyelinating disease; Multiple sclerosis; Remyelination;

Demyelination; Oligodendrocyte; Progenitor; Fibroblast growth factor;

Platelet-derived growth factor; Cuprizone; Differentiation

Demyelination in the CNS, such as in multiple sclerosis lesions,

can result in neurological deficits due to impaired nerve conduction

and associated axonal damage. Remyelination can occur in the

0969-9961/$ - see front matter. Published by Elsevier Inc.

doi:10.1016/j.nbd.2004.12.006

* Corresponding author. Department of Anatomy, Physiology, and

Genetics, Uniformed Services University of the Health Sciences, 4301 Jones

Bridge Road, Bethesda, MD 20814-4799, USA. Fax: +1 301 295 1715.

E-mail address: rarmstrong@usuhs.mil (R.C. Armstrong).1 Current address: The Children’s Hospital Boston, Boston, MA 02115,

USA.

Available online on ScienceDirect (www.sciencedirect.com).

adult CNS; however, the capacity to remyelinate becomes limited

with subsequent demyelinating episodes (Bruck et al., 2003).

Immature oligodendrocyte lineage cells (OLCs) persist in the adult

rodent and human CNS and may have the potential to enhance

remyelination if directed to differentiate into oligodendrocytes

(Armstrong et al., 1992; Chang et al., 2002; Goldman, 2003).

Furthermore, OP proliferation appears to be a prerequisite for

extensive remyelination in animal models (Blakemore and

Keirstead, 1999). This observation indicates that transient stim-

ulation of proliferation of persisting OP cells in human conditions

may generate new cells that can more effectively remyelinate

lesions. In vitro and in vivo studies have indicated that OLC

proliferation and differentiation can be regulated by the activity of

specific growth factor signaling pathways. Therefore, the current

study examines OLC responses to growth factor signaling in the

context of a demyelinating disease model that accomplishes

extensive remyelination. Endogenous growth factor pathways that

are employed in spontaneous remyelination may provide insight to

promote remyelination in human demyelinating diseases.

Both PDGF and FGF2 can act individually as mitogens for OPs

from neonatal and adult CNS in vitro (Frost et al., 2003;

McKinnon et al., 1990; Wolswijk and Noble, 1992). PDGF in

combination with FGF2 induces OPs cultured from neonatal and

adult rodents to proliferate as self-renewing stem cells and induces

adult OPs to cycle more rapidly (Bogler et al., 1990; Wolswijk and

Noble, 1992). In addition, FGF2 has been reported to increase OP

expression of PDGFaR in vitro (McKinnon et al., 1990). OPs

express receptors for PDGF and FGF2 during the proliferative

response to demyelination (Redwine and Armstrong, 1998).

Further, corresponding expression of PDGF and FGF2 ligands

increases in demyelinated lesion areas (Armstrong et al., 2002;

Messersmith et al., 2000; Woodruff et al., 2004). It is now

important to address the in vivo function of this endogenous PDGF

signaling in a demyelinated lesion environment and whether FGF2

signaling alters the PDGFaR response in this context.

J.C. Murtie et al. / Neurobiology of Disease 19 (2005) 171–182172

During remyelination, FGF2 may also play an important role

in regulating the timing of terminal differentiation as new

oligodendrocytes are generated. In vitro, FGF2 can inhibit

differentiation of OLCs derived from neonatal rodent brain

(Bansal and Pfeiffer, 1997) or isolated from spinal cords of

mice with demyelinating lesions (Armstrong et al., 2002).

Importantly, when demyelination was induced in FGF2 knock-

out mice, lesion repopulation by oligodendrocytes was signifi-

cantly enhanced without effecting proliferation or survival

(Armstrong et al., 2002). More direct demonstration of the

effect of FGF2 genotype on OLC differentiation in vivo could

determine whether attenuation of FGF2 signaling may improve

the cellular responses to promote remyelination in other

conditions.

The function of endogenous PDGF and FGF2 activity can be

examined in vivo using experimental demyelination of mice with

gene deletions that impair each signaling pathway. PDGF-A

knockout mice have dramatically impaired generation of OPs,

oligodendrocytes, and myelin, yet use in remyelination studies is

precluded due to early postnatal mortality (Fruttiger et al., 1999).

PDGF-A overexpression in experimental demyelination signifi-

cantly increased OP proliferation (Woodruff et al., 2004), but may

not appropriately reflect specific roles of endogenous PDGF

signaling during remyelination in vivo. PDGF ligands directly

regulate OLC responses through activation of PDGFaR, which is

the only PDGFR expressed by OLCs (Hart et al., 1989; McKinnon

et al., 1990). The PDGFaR homozygous null mutation is also

lethal during development; however, PDGFaR heterozygotes

express half the normal amount of PDGFaR without phenotypic

abnormalities (Soriano, 1997). FGF2 knockout mice are also

grossly normal and appropriate for analysis of remyelination

(Armstrong et al., 2002; Zhou et al., 1998).

The current study uses PDGFaR heterozygotes and FGF2

knockout mice to evaluate the contribution of each endogenous

signaling pathway in OLCs in a demyelinating disease model that

undergoes extensive spontaneous remyelination. We find that in

response to demyelination, OP amplification is significantly

reduced and subsequent oligodendroglial repopulation of lesions

is impaired in PDGFaR+/� mice. Using in vivo lineage tracing,

we demonstrate that OLCs differentiate into oligodendrocytes

more frequently in FGF2 knockout mice as compared to wild-type

mice. Our further studies in PDGFaR+/� FGF2�/� mice also

indicate distinct roles of each pathway, and argue against the

predicted cooperation of PDGF and FGF2 in OP amplification in

response to demyelination.

Materials and methods

Animals

Mice were bred and maintained in the USUHS animal housing

facility and all procedures were performed in accordance with

guidelines of the National Institutes of Health, the Society for

Neuroscience, and the USUHS Institutional Animal Care and Use

Committee. PDGFaR-targeted deletion mice on the C57Bl/6

genetic background were obtained from breeding heterozygous

pairs (generously provided by Dr. Soriano, Fred Hutchinson Cancer

Research Center). The PDGFaR targeted deletion replaces a 6.5-kb

fragment corresponding to the signal peptide, first and second Ig

domains (Soriano, 1997). FGF2 knockout mice on the 129 Sv-

Ev:Black Swiss genetic background were obtained from breeding

heterozygous pairs (generously provided by Dr. Doetschman,

University of Cincinnati). This FGF2 knockout was generated by

a targeted deletion replacing a 0.5 kb portion of the FGF2 gene

including 121 bp of the promoter and the entire first exon with an

Hprt mini-gene (Zhou et al., 1998). PDGFaR+/� FGF2�/� mice

and PDGFaR+/+ FGF2+/+ mice were generated by crosses of the

PDGFaR line with FGF2 line and the crosses were maintained on

this mixed genetic background.

Cuprizone experimental demyelination

Cuprizone ingestion results in a reproducible pattern of

extensive corpus callosum demyelination followed by sponta-

neous remyelination, which occurs over a period of weeks after

removal of cuprizone and return to normal chow (Armstrong et

al., 2002; Matsushima and Morell, 2001). For all strains of mice,

cuprizone treatment was started at 8 weeks of age and only male

mice were used. Cuprizone (finely powdered oxalic bis(cyclohex-

ylidenehydrazide); Sigma-Aldrich, St. Louis, MO) doses were

titrated for each of the strains of mice in order to induce a similar

and reproducible extent of demyelination based on histological

analyses (Armstrong et al., 2002). Cuprizone was thoroughly

mixed into milled chow (Harlan Teklad; Madison, WI), which was

available ad libitum. The cuprizone dose used was 0.2% (w/w) for

the PDGFaR line, 0.3% for the FGF2 line, and 0.2% for the

crosses of PDGFaR and FGF2 lines. Mice were maintained on

the cuprizone diet until perfused for analysis or returned to normal

chow pellets after 6 weeks.

Tissue preparation and histopathological analysis

Mice were perfused with 4% paraformaldehyde, and then

brains were dissected prior to overnight post-fixation in 4%

paraformaldehyde (Redwine and Armstrong, 1998). Brain tissue

was cryoprotected overnight at 48C in 30% sucrose and

embedded in OCT compound for immunostaining and in situ

hybridization. For histopathology, tissue was embedded in

paraffin and sections were stained with Luxol fast blue to detect

myelin combined with periodic acid-Schiff reaction for monitor-

ing the macrophage/microglial response (performed by USUHS

histological service).

In situ hybridization

In situ hybridization and preparation of digoxigenin-labeled

riboprobes were performed as previously detailed (Messersmith et

al., 2000; Redwine and Armstrong, 1998). Antisense riboprobes

were used to detect mRNA transcripts for PLP (gift from Dr. Lynn

Hudson; National Institutes of Health; Hudson et al., 1987) and

PDGFaR (gift from Dr. Bill Richardson; University College

London; Fruttiger et al., 1999). The digoxigenin-labeled riboprobes

were hybridized to 15-Am cryosections of brain or spinal cord

tissues. Digoxigenin was detected with an alkaline phosphatase-

conjugated sheep anti-digoxigenin antibody (Boehringer Man-

nheim, Indianapolis, IN), followed by reaction with NBT/BCIP

substrate (DAKO, Carpinteria, CA). For detection of PDGFaR

mRNA, the NBT/BCIP reaction was incubated overnight to

amplify the signal. This protocol allowed grossly similar detection

of PDGFaR mRNA signal from cells in PDGFaR+/� and

PDGFaR+/+ mice.

J.C. Murtie et al. / Neurobiology of Disease 19 (2005) 171–182 173

BrdU incorporation and detection

In situ hybridization combined with bromodeoxyuridine (BrdU)

incorporation was carried out as detailed previously (Redwine and

Armstrong, 1998). Mice were injected intraperitoneally with 200

mg/kg BrdU (Sigma, St. Louis, MO) at 4 h and 2 h prior to sacrifice.

After in situ hybridization detection, sections were treated with HCl

then incubated overnight with a monoclonal anti-BrdU antibody

directly conjugated with horseradish peroxidase (diluted 1:15;

Boehringer Mannheim). Peroxidase activity was detected using

3,3V-diaminobenzidine (DAB; Vector Labs, Burlingame, CA).

Immunohistochemistry

To identify OPs in situ, 15-Am cryosections were immunostained

for NG2 and PDGFaR (Armstrong et al., 2002; Messersmith et al.,

2000). Primary antibodies used were rabbit polyclonal anti-NG2

antibody (1:500; gift from Dr. William Stallcup, La Jolla, CA) and

rat monoclonal anti-PDGFaR antibody (APA5 used at 1:200;

Pharmingen, San Diego, CA). Donkey anti-rabbit IgG F(abV)2conjugated with Cy3 (Jackson Immunoresearch, West Grove, PA)

was used to detect NG2 while the PDGFaR was detected with

biotinylated donkey anti-rat IgG F(abV)2 (Jackson Immunoresearch)

followed by coumarin tyramide amplification (New England

Nuclear, Boston, MA).

Mature oligodendrocytes were identified with CC1, which

immunostains oligodendrocyte cell bodies without labeling myelin

(Fuss et al., 2000). The CC1 antibody (Oncogene Research

Products, Cambridge, MA) was detected with donkey anti-mouse

IgG F(abV)2 conjugated with Cy3 (Jackson Immunoresearch). The

CC1 immunostaining conditions were previously tested to ensure

that CC1 did not label astrocytes or NG2-labeled cells (Messer-

smith et al., 2000).

Myelin was immunostained with monoclonal antibody 8-18C5,

which recognizes MOG (hybridoma cells provided by Dr. Minetta

Gardinier; University of Iowa, Iowa City, IA; Linnington et al.,

1984). MOG immunolabeling was detected with donkey anti-mouse

IgG F(abV)2 conjugated with Cy3 (Jackson Immunoresearch).

Apoptosis

After 6 weeks of cuprizone treatment, adult mouse corpus

callosum was analyzed for cells undergoing apoptosis. Cryosec-

tions (15 Am) were processed using a modified TUNEL assay

(ApopTag Plus peroxidase in situ apoptosis detection kit; Intergen,

Purchase, NY). The 3V-OH DNA ends, generated by DNA

fragmentation typically observed with apoptotic cells, were labeled

with digoxigenin-dUTP using terminal deoxynucleotidyl trans-

ferase (TdT). The digoxigenin tag was then detected with an anti-

digoxigenin antibody conjugated with peroxidase and reacted with

DAB substrate. The sections were lightly counterstained with

methyl green to detect nuclei.

Retrovirus production

A 293-cell line stably transfected with pNIT-GFP, a replication-

incompetent retroviral expression vector encoding GFP, (provided

by Dr. Fred Gage; Salk Institute, La Jolla, CA; Palmer et al., 1999)

was transiently transfected with pMD.G (plasmid containing the

vesicular stomatitis virus glycoprotein) and cultured for 2 days.

Virion-containing supernatants were concentrated 100-fold by

centrifugation at 50,000 � g at 48C for 150 min. Viral pellets

were reconstituted in Hanks balanced salt solution to a final

concentration of 100� the original concentration. Titers (typically

105 cfu/mL) were determined by GFP expression after infection of

NIH 3T3 cells with serial dilutions of concentrated virus.

Replication-incompetent DAP retrovirus, encoding membrane-

associated human placental alkaline phosphatase, was generated

from a stable NIH 3T3 producer cell line (Fields-Berry et al., 1992).

Producer cells were grown to confluence and the supernatant was

collected after 3 days and concentrated 100-fold by centrifugation

at 50,000 � g at 48C for 150 min. Alkaline phosphatase was

detected with NBT/BCIP substrate (DAKO, Carpinteria, CA).

Stereotaxic surgery and retrovirus injection

Three days prior to cuprizone treatment, 8-week-old male mice

were anesthetized with isoflurane and body temperature was

maintained with 378C isothermal heating pads. The head was

stabilized in a stereotaxic apparatus using cushioned ear bars and a

burr hole was drilled into the skull at the following coordinates:

bregma�1.0 mm and 0.25mm lateral to the sagittal suture (Franklin

and Paxinos, 1997). Virus (102 cfu in 1.0 AL) was injected directly

into the corpus callosum at a depth of 1.875 mm and a rate of 0.2 ALper min. After injection, the burr hole was sealed using Gelfoam and

the wound closed with a single wound clip.

After 6 weeks of cuprizone treatment and 3 weeks of recovery,

mice were transcardially perfused with 3% paraformaldehyde in 0.1

M phosphate buffer followed by fixation overnight at 48C in 3%

paraformaldehyde. Tissues to be used for immunohistochemistry

were cryoprotected in 30% sucrose overnight at 48C then

embedded in OCT compound and stored at �808C.

Imaging, quantification, and statistical analysis

Images of in situ hybridization results were captured with a Spot

2 digital camera using Spot Advanced image acquisition software

(Diagnostic Instruments, Sterling Heights, MI) on an Olympus IX-

70 microscope. Fluorescent images were captured using a Laser

Scanning PASCAL 5 Zeiss confocal microscope with a 100�objective. A stack of Z-series images was acquired and the

maximum projection of this image stack was generated using Zeiss

Pascal 5 software. Images were prepared as panels using Adobe

Photoshop.

For comparing cell densities, cells expressing PLP mRNAwere

quantified using unbiased stereological morphometric analysis

(Messersmith et al., 2000; Stereologer System Systems Planning

and Analysis, Inc., Alexandria, VA). Analysis was restricted to the

corpus callosum region, from the midline and extending laterally to

below the cingulum in 15-Am-thick coronal sections. Using the

Stereologer System, the thickness is sampled as part of the

definition of each bdissectorQ volume, so that density measure-

ments reflect cells/mm3. The unbiased stereological method could

not be used appropriately for conditions with relatively few cells of

interest in any particular category. Therefore, quantification of

PDGFR/BrdU single- and double-labeled categories in the corpus

callosum required counting all labeled cells and measuring the area

sampled (Armstrong et al., 2002). Without use of the Stereologer

System, section thickness could not be sampled in the mounted

specimen, and so density units are cells/mm2.

Each category analyzed included 3 or more tissue sections per

mouse and 3 or more mice per condition, except where noted in

J.C. Murtie et al. / Neurobiology of Disease 19 (2005) 171–182174

text and/or figure legends. Unpaired Student’s t tests were used to

identify significant differences between genotypes and/or treat-

ments. No statistical comparisons were made between mice with

different genetic backgrounds (i.e., PDGFaR+/� mice were not

compared to PDGFaR+/� FGF2�/� mice). The significance of

proportions generated in lineage tracing was tested using the v2

statistical test.

Results

Oligodendrocyte and OP densities in non-lesioned white matter of

adult PDGFaR heterozygotes

Prior to analysis of OLC responses to demyelination, potential

effects of PDGFaR genotype were examined for the resting OLC

populations of 8-week-old mice maintained on normal chow (Fig.

1). Oligodendrocytes were identified by in situ hybridization for

proteolipid protein (PLP) mRNAwhile OP cells were identified by

in situ hybridization for PDGFaR mRNA or immunostaining for

NG2 chondroitin sulfate proteoglycan. The density of oligoden-

drocytes was significantly reduced in the non-lesioned corpus

callosum of PDGFaR+/� mice compared to PDGFaR+/+ mice

(Fig. 1E). OP density was concurrently decreased in the non-

lesioned corpus callosum of PDGFaR heterozygotes (Figs. 1F, G).

Fig. 1. Reduced OLC populations in the non-lesioned white matter of adult PD

callosum to show oligodendrocytes identified by in situ hybridization for PLP mR

and OP cells identified by PDGFaR mRNA (C from PDGFaR+/+ mouse, D from

chow. Panels C and D are also immunostained to detect BrdU (brown) incor

oligodendrocytes (E) and OPs (F, G) in the corpus callosum of PDGFaR+/+ m

chow. (E) In PDGFaR heterozygotes, the density of oligodendrocytes, identified

0.002). (F, G) The density of OPs was significantly lower in PDGFaR heterozy

hybridization for PDGFaR mRNA (F, P = 0.003) and by NG2 immunostaining

Amplification of the NBT/BCIP reaction allowed similar detection

of cells expressing PDGFaR mRNA transcripts from both

PDGFaR genotypes (Figs. 1 C, D), even though PDGFaR

expression is reduced by half in the heterozygotes (Soriano,

1997). OP proliferation was estimated using a 4-h terminal pulse of

BrdU. Very few cells with incorporated BrdU were detected in the

corpus callosum (28.31 F 4.50 cells/mm2 PDGFaR+/+, n = 3;

20.70 F 1.68 cells/mm2 PDGFaR+/�, n = 3), as expected for the

relatively low rate of cell division typically observed in non-

lesioned adult white matter. In contrast to FGF2 signaling

(Armstrong et al., 2002), these findings indicate that endogenous

PDGFaR signaling has an important role in maintaining OLC

densities in the normal adult CNS.

Oligodendrocyte repopulation of demyelinated lesions in PDGFaRheterozygotes

Our previous study showed that the absence of FGF2 resulted

in a significant increase in oligodendrocyte density during

remyelination (Armstrong et al., 2002). After 3 weeks of recovery

from cuprizone demyelination, significantly enhanced oligoden-

droglial repopulation of lesions was observed in FGF2�/� mice

(325 F 5 cells/mm3 � 103) compared to FGF2+/+ mice (247 F16 cells/mm3 � 103). In contrast to this result, we find that

reduced PDGFaR expression significantly impairs oligodendro-

GFaR heterozygotes. Representative coronal sections through the corpus

NA (A from PDGFaR+/+ mouse, B from PDGFaR+/� mouse; dark blue)

PDGFaR+/� mouse; dark blue) in 8-week-old mice maintained on normal

porated during a 4-hour terminal pulse. Quantification of the density of

ice (black bars) and PDGFaR+/� mice (gray bars) maintained on normal

by PLP mRNA, was significantly lower than in the wild-type mice ( P =

gotes than in the wild-type mice, based upon OP identification by in situ

(G, P = 0.015). Scale bar for A–D = 50 Am, as shown in D.

J.C. Murtie et al. / Neurobiology of Disease 19 (2005) 171–182 175

glial lesion repopulation in the same cuprizone model. Myelin

oligodendrocyte glycoprotein (MOG) immunostaining (Fig. 2) and

histological staining (not shown) demonstrated that PDGFaRheterozygotes undergo the same time course of demyelination and

spontaneous remyelination that has been previously reported for

adult mice fed cuprizone (Armstrong et al., 2002; Matsushima and

Morell, 2001). By 3 weeks into the cuprizone treatment,

demyelination is not yet consistently evident in the corpus

callosum (Fig. 2A), but there is already substantial loss of

oligodendrocytes (Fig. 3C). Demyelination of the corpus callosum

is complete by 6 weeks (Fig. 2B). After a further 3 weeks on

normal chow to allow remyelination to proceed, extensive MOG

immunoreactivity is again present throughout the corpus callosum

(Fig. 2C).

Analysis of oligodendrocyte density over the course of

demyelination using PLP mRNA in situ hybridization revealed a

dramatic decrease in lesion repopulation by oligodendrocytes in

PDGFaR heterozygotes (Fig. 3). In PDGFaR+/+ mice, oligoden-

drocyte regeneration is sufficient to regain normal density

throughout the remyelinating corpus callosum after 3 weeks of

Fig. 2. Similar cuprizone disease course in PDGFaR heterozygotes.

Representative coronal views of MOG immunostaining of myelinated

fibers in the corpus callosum over the course of cuprizone treatment in

PDGFaR+/� (A, B, C) and PDGFaR+/+ (not shown) mice. (A) Myelin

is not yet consistently degenerated and phagocytosed after 3 weeks of

cuprizone ingestion so that MOG immunostaining is still present in the

corpus callosum. (B) Maximal demyelination is observed after 6 weeks

of cuprizone ingestion. (C) Remyelination is indicated by MOG

immunoreactivity throughout the corpus callosum after 6 weeks of

cuprizone ingestion followed by 3 weeks of recovery on normal chow.

(m = midline) Scale bar = 100 Am.

Fig. 3. Impaired oligodendrocyte repopulation of demyelinated corpus

callosum in PDGFaR heterozygotes. At specific time points of cuprizone

treatment and recovery, oligodendrocytes were identified by in situ

hybridization for PLP mRNA. Representative images of PLP mRNA in

situ hybridization of coronal brain sections from PDGFaR+/+ (A) and

PDGFaR+/� (B) mice following 6 weeks of continuous cuprizone

ingestion and 3 weeks of recovery on normal chow (A, B; m = midline;

scale bar = 100 Am). (C) Quantitation of PLP mRNA-labeled cell densities

prior to cuprizone ingestion (8 weeks old; no cup), after 3 weeks or 6 weeks

of continuous cuprizone (3 weeks cup and 6 weeks cup), and after a 3-week

recovery period (3 weeks off) with PDGFaR+/� mice (white bars) as

compared to PDGFaR+/+ mice (black bars). Oligodendrocyte regeneration

was significantly impaired in PDGFaR heterozygotes after 6 weeks of

cuprizone ingestion (*P = 0.008) and after 3 weeks of recovery (**P =

0.003). In addition, oligodendrocyte density in PDGFaR heterozygotes

after 3 weeks of recovery did not return to the level found in PDGFaRheterozygotes prior to demyelination (white bars, 3 weeks off vs. no cup;

~P = 0.015).

recovery from cuprizone treatment. In contrast, the oligodendro-

glial repopulation response in PDGFaR+/� mice was impaired so

that the oligodendrocyte density did not return to the level found in

non-lesioned PDGFaR heterozygotes.

Proliferation in PDGFaR heterozygotes and FGF2 knockout mice

Although PDGFaR expression correlates with the vigorous OP

proliferative response to demyelination (Redwine and Armstrong,

1998), a mitogenic function of endogenous PDGFaR signaling has

not been demonstrated in vivo during remyelination. Relative to

cuprizone disease progression, maximal OP proliferation is

observed in the corpus callosum after 5 weeks of cuprizone

treatment (Armstrong et al., 2002). Therefore, the 5-week time

point was chosen to analyze the OP proliferative response in the

corpus callosum using PDGFaR mRNA in situ hybridization in

combination with detection of BrdU (Fig. 4). A 4-h terminal pulse

of BrdU was selected to detect proliferation while minimizing

subsequent time in which differentiation could proceed. NG2

immunostaining was not used to quantify OP density during

remyelination because the intense NG2 immunoreactivity on OP

cell bodies and processes throughout demyelinated lesions makes

accurate cell counts extremely difficult (Keirstead et al., 1998; our

unpublished observations).

Fig. 4. Decreased OP density and proliferation in PDGFaR heterozygotes

and PDGFaR +/� FGF2 �/� mice. (A) Representative image of PDGFaR

mRNA in situ hybridization (blue cytoplasm) combined with BrdU

immunostaining (brown nuclei) in the corpus callosum of a PDGFaRheterozygote. (B) Quantification of OP populations in PDGFaR wild-type

mice (aR +/+, black bars) and heterozygotes (aR +/�, white bars) in

comparison with PDGFaR+/� FGF2�/� mice (aR +/� FGF2 �/�; dark

gray bars). The density of OP cells that had not incorporated BrdU (black

arrowheads in A; aR+ BrdU� in B) is significantly decreased in PDGFaR

heterozygotes (aR +/+ vs. aR +/�; *P = 0.012). A significant decrease was

also observed in the density of dividing OPs (black arrow in A; aR+ BrdU+

in B) (aR +/+ vs. aR +/�; ** P = 0.026). The density of BrdU-labeled cells

that were not identified as OPs (white arrowhead in A; aR- BrdU+ in B) was

also significantly reduced in PDGFaR heterozygotes (aR +/+ vs. aR +/�;f

P = 0.038). The pattern in PDGFaR+/� mice is similar to PDGFaRheterozygotes crossed with FGF2 knockout mice (aR +/� FGF2 �/�).

Scale bar = 25 Am.

Fig. 5. Retroviral infection labels endogenous cycling cells of the adult

white matter that have the capacity to remyelinate. Example of coronal

section of corpus callosum from a mouse injected with DAP replication-

incompetent retrovirus encoding the alkaline phosphatase reporter. DAP

was injected 3 days prior to initiating cuprizone ingestion so that the virus

infected and heritably labeled the endogenous cycling cells of the normal

adult corpus callosum. After 6 weeks of continuous cuprizone followed by

3 weeks on normal chow, membrane-associated alkaline phosphatase

substrate deposition is found on parallel tracks oriented with axons in the

corpus callosum, indicative of remyelination by cells generated from the

endogenous cycling cells. Scale bar = 50 Am.

J.C. Murtie et al. / Neurobiology of Disease 19 (2005) 171–182176

As compared to PDGFaR+/+ mice, the total OP density in

PDGFaR+/� mice was dramatically reduced during the peak of

OP amplification in response to cuprizone demyelination (Fig. 4).

The density of OPs without detectable BrdU (PDGFaR+, BrdU�cell phenotype) was significantly reduced in PDGFaR+/� mice

relative to PDGFaR+/+ mice. The density of proliferating OP cells

(PDGFaR+, BrdU+ cell phenotype) was also significantly reduced

in PDGFaR+/� mice. The labeling index (i.e., the percentage of

PDGFaR+ OP cells with incorporated BrdU) was similar for

PDGFaR+/� (14.46%) and PDGFaR+/+ (14.68%) genotypes.

Finally, the density of BrdU labeled cells that were not identified as

OPs (PDGFaR�, BrdU+ cell phenotype) was also significantly

reduced in PDGFaR heterozygotes compared to PDGFaR wild-

type mice. This population of PDGFaR-negative cells could

include several cell types with responsive yet undetectable levels

of PDGFaR, such as newly formed oligodendrocytes. In addition,

several reports have shown that reactive astrocytes express low

levels of PDGFaR and may also have the potential to respond to

PDGF (Maeda et al., 2001; Redwine and Armstrong, 1998).

To test the potential cooperative interaction of FGF2 and

PDGFaR signaling in the OP proliferative response to demyeli-

nation, OP density and BrdU incorporation was assessed in mice of

the PDGFaR line crossed with mice of the FGF2 line. Our

previous analysis demonstrated that FGF2 knockout alone did not

significantly alter OP density or proliferation (Armstrong et al.,

2002). OP proliferation was significantly impaired in PDGFaR+/�FGF2�/� mice compared to wild-type mice of the same crossed

background (P = 0.006 for comparison of 75.34 F 9.81 cells/mm2

PDGFaR+/� FGF2�/� mice, n = 5, with 119.38 F 42.16 cells/

mm2 PDGFaR+/+ FGF2+/+ mice, n = 4). Importantly, loss of

FGF2 in PDGFaR+/� FGF2�/� mice did not exacerbate the

already reduced OP density or impaired proliferation observed in

PDGFaR+/� mice (Fig. 4). These findings indicate that PDGFaR

signaling is the significant regulator of OP proliferation in response

to demyelination and does not interact with FGF2 signaling to

promote OP proliferation.

Differentiation in PDGFaR heterozygotes and FGF2 knockout

mice

In vivo lineage tracing with replication-incompetent retrovirus

was used to monitor differentiation of endogenous cycling cells

into oligodendrocytes during the course of demyelination and

remyelination. To determine the impact of endogenous PDGF and

FGF2 pathways on OLC differentiation during remyelination,

lineage tracing was performed in PDGFaR heterozygotes and in

FGF2 knockout mice, with comparison to wild-type mice of each

line. Three days prior to initiation of cuprizone treatment, DAP or

NIT-GFP retrovirus was stereotactically injected into the corpus

callosum of 8-week-old mice to infect the endogenous cycling cells

of the adult white matter. Injecting with retrovirus prior to inducing

demyelination allows infection of endogenous cycling cells but not

reactive astrocytes and microglia (Gensert and Goldman, 1997; our

unpublished observations). Retroviral infection of these endoge-

nous cycling cells with a heritable reporter gene can monitor the

generation of remyelinating oligodendrocytes, as demonstrated by

myelin immunolabeling (Gensert and Goldman, 1997) or by

alkaline phosphatase association with myelin membranes (Fig. 5).

After 6 weeks of cuprizone treatment followed by a 3-week

recovery period for remyelination, cells labeled by NIT-GFP

J.C. Murtie et al. / Neurobiology of Disease 19 (2005) 171–182 177

retrovirus were analyzed. Both OPs and oligodendrocytes were

labeled by retroviral GFP after recovery from demyelination.

Oligodendrocytes were identified morphologically by the presence

of bifurcating processes that form bTQ intersections parallel to axonsas well as by immunostaining for the oligodendrocyte marker CC1

(Figs. 6A–C). OPs were identified by double immunostaining for

the OP markers NG2 and PDGFaR (Figs. 6D–G).

In vivo OLC differentiation was assessed by determining the

proportion of total GFP-labeled cells that were identified as OPs or

oligodendrocytes in mice perfused after 3 weeks of recovery from

cuprizone treatment. This quantification indicated that differ-

entiation of OLCs was not changed in PDGFaR+/� mice as

compared to PDGFaR+/+ mice (Figs. 6H, I). Similar proportions

of OPs and oligodendrocytes were found in both genotypes. The

remaining proportion of cells that were not identified by cell type-

specific markers had morphologies consistent with OLCs and this

proportion of cells was also not significantly different between

genotypes.

Fig. 6. Retroviral lineage tracing reveals altered in vivo differentiation with FGF2

into the corpus callosum of 8-week old mice, 3 days later the mice began a 6-week

recovery period on normal chow, then the mice were perfused and GFP expressing

(A–C) A GFP-expressing cell immunolabeled with CC1, a marker of oligodendro

is detectable in the cell body and processes while CC1 immunostaining (B, red) is

points to a bifurcated process forming a bTQ intersection with distal process br

immunolabeled with PDGFaR and NG2, markers of OP cells (representative cell

GFP fluorescence (D, green) and are outlined by NG2 (E, red) immunoreactivity w

Quantitation of the proportion of total GFP+ cells that were morphologically iden

not changed by PDGFaR genotype, in contrast to the significant increase (*P = 0.

(n = 3 mice). (I) Quantitation of the proportion of total GFP+ cells that were ide

immunostained with CC1, or as OP cells, in tissue sections immunostained for PD

in mice of each PDGFaR+ genotype. FGF2�/� mice have a significantly lower

proportion of GFP-labeled oligodendrocytes (~P = 0.003) compared to FGF2+/+

mean F standard error of the proportion.

In contrast, FGF2 genotype clearly influenced OLC differ-

entiation in mice that had recovered from cuprizone demyelina-

tion. The proportion of GFP-positive cells that were

morphologically identified as oligodendrocytes was significantly

higher in FGF2�/� mice, as compared to FGF2+/+ mice

(Fig. 6I). This quantification by morphology was confirmed with

CC1 immunohistochemistry in combination with GFP detection

(Fig. 6K). Consistent with this result, the proportion of GFP-

positive cells identified as OPs (i.e., immunolabeled with NG2 and

PDGFaR) was significantly decreased in FGF2�/� mice

(Fig. 6K). These findings indicate that endogenous FGF2 inhibits

OLC differentiation during remyelination.

Survival in PDGFaR heterozygotes and FGF2 knockout mice

TUNEL analysis was completed on PDGFaR+/+ and

PDGFaR+/� mice after 6 weeks of cuprizone treatment to

determine if OLC survival was affected by a decrease in PDGFaR

genotype but not with PDGFaR genotype. NIT-GFP retrovirus was injected

period of continuous cuprizone ingestion which was followed by a 3-week

cells were analyzed by morphology and cell type-specific immunostaining.

cytes (representative cell example; PDGFaR heterozygote). GFP (A, green)

found in the cell body (C, overlap in cell body appears yellow). The arrow

anches aligned parallel to the axon fibers. (D–G) A GFP-expressing cell

example; PDGFaR heterozygote). An OP cell body and processes contain

hile PDGFaR immunoreactivity (F, blue) is within the soma cytoplasm. (H)

tified (see text) as oligodendrocytes. The proportion of oligodendrocytes is

001) observed in FGF2�/� mice (n = 5 mice) compared to FGF2+/+ mice

ntified by cell type-specific markers as oligodendrocytes, in tissue sections

GFaR and/or NG2. The oligodendrocyte and OP populations are the same

proportion of GFP-labeled OP cells (#P = 0.001) and a significantly higher

mice. (A, D) Scale bars = 10 Am. Values shown represent the proportional

J.C. Murtie et al. / Neurobiology of Disease 19 (2005) 171–182178

expression. Cell death from cuprizone toxicity should be minimal at

this time point (Mason et al., 2000). With this in mind, the 6-week

cuprizone time point should be the appropriate stage at which to test

for a survival effect on OPs and newly formed oligodendrocytes.

TUNEL analysis yielded very few TUNEL+ cells in either

genotype; furthermore, there was no significant genotypic effect

on cell survival (50.76F 6.46 cells/mm2 in PDGFaR+/+ mice, n =

3; 27.83F 6.95 cells/mm2 in PDGFaR+/�mice, n = 3; P = 0.063).

These results are similar to results from TUNEL analysis in FGF2

knockout mice (Armstrong et al., 2002). Together, these data

indicate that neither PDGF nor FGF2 have detectable effects on

OLC survival in vivo.

Oligodendroglial lesion repopulation in PDGFaR+/� FGF2�/�mice

The potential interaction of PDGF and FGF2 pathways in the

OLC response to demyelination was further tested by examining

oligodendrocyte repopulation of lesions in crosses of the PDGFaRand FGF2 lines (Fig. 7). As stated previously, the oligodendroglial

repopulation response to demyelination is hindered in PDGFaRheterozygotes and enhanced in FGF2 knockout mice. If these two

pathways interact as mitogens that promote OP amplification in

this context, then oligodendrocyte regeneration in PDGFaR+/�FGF2�/� mice should be reduced even more dramatically than in

PDGFaR heterozygotes. Alternatively, if PDGF and FGF2 signal-

ing pathways have distinct effects on OLCs during remyelination,

then the oligodendroglial repopulation in PDGFaR+/� FGF2�/�mice should be improved relative to PDGFaR heterozygotes to

reflect the combined effect of the individual mutations. Analysis of

non-lesioned mice revealed initial support of the latter interpreta-

tion. Specifically, the absence of FGF2 in PDGFaR+/� FGF2�/�mice enabled the oligodendrocyte density in non-lesioned white

matter to achieve normal levels (Fig. 7; no cup). Further support of

distinct effects was clear when PDGFaR+/� FGF2�/� mice were

demyelinated by cuprizone ingestion. Mice from the cross of

PDGFaR and FGF2 lines underwent the same characteristic

Fig. 7. Absence of FGF2 rescues PDGFaR heterozygote impairment of

oligodendrocyte repopulation of demyelinated lesions. In situ hybridization

for PLP mRNA to assess oligodendrocyte density over the course of

cuprizone treatment in PDGFaR+/� FGF2�/� mice (white bars) as

compared to PDGFaR+/+ FGF2+/+ mice (black bars) showed no

significant differences between genotypes. The oligodendrocyte density

was quantitated prior to the start of cuprizone treatment (no cup; 8 weeks of

age) to compare normal baseline values, during the period of oligoden-

drocyte loss at 3 weeks of continuous cuprizone treatment (3 weeks cup),

and during the period of oligodendrocyte repopulation after 6 weeks of

continuous cuprizone followed by 3 weeks of recovery on normal chow (6

weeks cup, 3 weeks off).

pattern of oligodendrocyte loss after 3 weeks of continuous

cuprizone treatment (Fig. 7; 3 weeks cup). After 6 weeks of

continuous cuprizone ingestion followed by 3 weeks of recovery

on normal chow, significant oligodendrocyte repopulation occurred

and was similar in PDGFaR+/� FGF2�/� mice as compared to

PDGFaR+/+ FGF2+/+ wild-type mice (Fig. 7; 6 weeks cup, 3

weeks off). These findings are consistent with distinct effects of

endogenous signaling through PDGF and FGF2 on OLCs in the

course of experimental demyelination and remyelination. Specif-

ically, these results support a model (Fig. 8) in which PDGF

signaling enhances OP proliferation during demyelination and

FGF2 predominantly inhibits differentiation of OPs into oligoden-

drocytes during remyelination.

Discussion

In response to demyelination, both PDGF and FGF2 ligand and

receptor expression levels are increased (Armstrong et al., 2002;

Hinks and Franklin, 1999; Messersmith et al., 2000; Woodruff et

al., 2004). These findings in combination with functional in vitro

data indicate that FGF2 and PDGF could potentially act

individually or cooperatively to regulate OLC responses in the

lesion environment. In the current study, we show that PDGFaRheterozygotes have a deficit in proliferation and subsequent lesion

repopulation by oligodendrocytes. This impaired proliferative

response to demyelination in PDGFaR heterozygotes was not

exacerbated by the absence of FGF2 in PDGFaR+/� FGF2�/�mice. In contrast to the current results in PDGFaR+/� mice, FGF2

knockout mice have an enhanced oligodendroglial repopulation

response (Armstrong et al., 2002). We now demonstrate that

removal of endogenous FGF2 from the lesion environment

corresponds with more frequent differentiation of immature cells

into mature oligodendrocytes. This result in FGF2�/� mice is not

likely to result from compensation by other FGF family members

since a similar effect was observed in vitro using a short-term

treatment with anti-FGF2 specific neutralizing antibody (Arm-

strong et al., 2002) and initial studies of these FGF2�/� mice

showed that expression of FGF1, FGF4, and FGF5 was not altered

(Zhou et al., 1998).

PDGF and FGF2 can each act as OP mitogens in vitro (Frost et

al., 2003; McKinnon et al., 1990; Wolswijk and Noble, 1992).

Indeed, the current findings show that PDGF is a critical

endogenous mitogen for OLCs in the adult CNS in vivo. In

contrast, FGF2 is not a significant OP mitogen in this context

(Armstrong et al., 2002). The current in vivo findings are

consistent with the role of PDGFaR signaling in vitro as

stimulating proliferation of OPs from adult CNS (Shi et al.,

1998; Wolswijk and Noble, 1992). In diverse models of

experimental demyelination, amplification of OPs expressing

PDGFaR has been associated with demyelinated lesions (Penderis

et al., 2003; Redwine and Armstrong, 1998) and increased

expression of endogenous PDGF-A ligand (Redwine and Arm-

strong, 1998; Woodruff et al., 2004). Overexpression of PDGF-A

ligand in experimental demyelination further increased OP cell

densities in normal adult and lesioned GFAP-hPDGF-A mice,

demonstrating a functional response of this signaling pathway in

the context of demyelinating disease (Woodruff et al., 2004). The

present results now extend these studies to demonstrate that

endogenous PDGFaR signaling indeed plays a role in promoting

OP proliferation in the context of a demyelinating lesion environ-

Fig. 8. Proposed model for actions of endogenous PDGF and FGF2 on OLCs in the adult CNS. (A) Previous in vitro studies have predicted multiple potential

roles for PDGF and FGF2 in vivo (see text). PDGF and FGF2 can each stimulate OP proliferation, and cooperative interaction of PDGF and FGF2 can induce

OPs from adult CNS to divide more rapidly as a self-renewing line. PDGF can promote appropriately timed OP differentiation while FGF2 strongly inhibits

differentiation. (B) Based on studies using PDGFaR+/� mice (current study), FGF2�/� mice (Armstrong et al., 2002 and current study), and PDGFaR+/�FGF2�/� mice (current study), these growth factors do not act cooperatively in vivo. Reduced PDGFaR expression led to decreased OP proliferation (thin

circular arrow) and fewer oligodendrocytes. Removal of FGF2 inhibition of differentiation (dotted intersecting bars) led to increased OP differentiation into

oligodendrocytes. (C) These results in mice with genetic deletions led to the proposed model of endogenous PDGF and FGF2 functions. PDGF is an important

OP mitogen (thick circular arrow), particularly during the OLC response to demyelination. In contrast, FGF2 predominantly inhibits differentiation (solid

intersecting bars) of OPs into oligodendrocytes, which is critically important during remyelination. (A, B, C) Thickness of arrows corresponds with amplitude

of cellular response; circular arrow indicates OP proliferation; straight arrow indicates promotion of OP differentiation; intersecting bar indicates inhibition of

OP differentiation.

J.C. Murtie et al. / Neurobiology of Disease 19 (2005) 171–182 179

ment. Both PDGF-A ligand overexpression (Woodruff et al., 2004)

and PDGFaR deficiency (current study) altered OP densities in

demyelinated lesions without altering the BrdU-labeling index, as

compared to wild-type mice. Therefore, in response to demyeli-

nation, PDGF may stimulate conversion of resting OP cells into

rapidly proliferating OP cells without subsequently altering the OP

cell cycle length during the disease progression.

Surprisingly, we also show that PDGFaR genotype plays a

significant role in the maintenance of OPs and oligodendrocytes in

the non-lesioned adult white matter. The current findings in non-

lesioned adult CNS are consistent with previous in vivo

manipulations of PDGF ligand levels that demonstrated the

importance of PDGF in regulation of OP and oligodendrocyte cell

density during development (Barres and Raff, 1993; Calver et al.,

1998; Fruttiger et al., 1999). However, these reports conclude that

axon number is the overriding factor that controls the final

oligodendrocyte density in the non-lesioned white matter at the

end of myelination. Our findings indicate that PDGFaR signaling

continues to contribute to OP and oligodendrocyte cell densities in

the adult CNS. Presumably, PDGFaR signaling then acts in

concert with axonal signals and other growth factor pathways.

Several studies have implicated PDGF and FGF2 as influencing

OLC differentiation (Allamargot et al., 2001; Bansal and Pfeiffer,

1997; Bogler et al., 1990; Wilson et al., 2003). We used retroviral

lineage tracing over the course of demyelination and remyelination

to demonstrate that FGF2 deletion, but not PDGFaR hetero-

zygosity, altered OLC differentiation. This direct analysis of the

progression from cycling cell through differentiated oligodendro-

cyte indicates that endogenous FGF2 inhibits OLC differentiation.

This effect of FGF2 was also predicted from our previous analysis

of OLC population densities in FGF2 knockout mice undergoing

demyelination and remyelination (Armstrong et al., 2002). This

role of FGF2 as an inhibitor of OLC differentiation in vivo during

remyelination contrasts with a report of developmental myelination

in FGFR3 null mice in which endogenous FGFR3 signalling was

predicted to promote OLC differentiation (Oh et al., 2003). These

results may reflect differential expression of FGFR1, FGFR2, and/

or FGFR3 at different OLCs stages in development as compared to

in remyelinating lesions (Bansal et al., 1996; Messersmith et al.,

2000). Alternatively, these findings may be explained by preferred

activation of FGFR3 by FGF family members that may be more

optimal ligands than FGF2 (Chellaiah et al., 1999; Ornitz and

Leder, 1992). Furthermore, future studies will need to address

whether FGF2 acts directly on OLCs in this context, since FGFRs

are expressed by multiple cell types in demyelinated lesions (Liu et

al., 1998; Messersmith et al., 2000).

PDGF can promote survival of OP cells and newly generated

oligodendrocytes (Barres et al., 1992; Ebner et al., 2000), while

FGF2 has been reported to be detrimental for oligodendrocyte

viability (Muir and Compston, 1996). However, an effect on cell

survival was not detected using TUNEL analysis during the

transition from demyelination to remyelination in PDGFaRheterozygotes (current study) or in FGF2 knockout mice (Arm-

strong et al., 2002). TUNEL analysis and other assays of apoptosis

identify cells actively undergoing apoptosis that have not yet been

removed by phagocytosis. Therefore, it is possible that subtle

changes in cell survival are difficult to detect in this interval.

Furthermore, even though PDGFaR expression is reduced to

approximately half the wild-type expression level in PDGFaRheterozygotes (Soriano, 1997), we cannot yet rule out the

possibility that this reduced expression of PDGFaR could still be

sufficient to provide the necessary survival signaling.

J.C. Murtie et al. / Neurobiology of Disease 19 (2005) 171–182180

PDGFaR heterozygotes and FGF2 knockout mice provided an

excellent opportunity to test the potential in vivo interaction

between PDGF and FGF2 signaling pathways. Cooperation

between PDGF and FGF2 has been reported to convert slowly

dividing OPs from adult CNS into rapidly dividing cells in vitro,

indicative of an important potential role in repair of demyelinated

lesions in vivo (Chari and Blakemore, 2002; Wolswijk and Noble,

1992). Accordingly, absence of FGF2 combined with decreased

expression of PDGFaR may have been predicted to exacerbate the

PDGFaR heterozygote effect of impaired OP proliferation and

lesion repopulation by oligodendrocytes. In fact, our results do not

support a role for FGF2 cooperating with PDGFaR signaling for a

mitogenic effect on OPs in vivo. Using PDGFaR+/� FGF2�/�mice, there was no observed interaction between PDGF and FGF2

as OP mitogens (Fig. 4). In addition, oligodendroglial repopulation

in response to demyelination was not further impaired as compared

to PDGFaR+/� mice but was instead improved in PDGFaR+/�FGF2�/� mice (Fig. 7).

Based on our in vivo analyses of genetic deletions in

experimental demyelination (current study and Armstrong et al.,

2002), we propose a model in which endogenous PDGF and FGF2

predominantly mediate distinct effects on OLCs in demyelinated

white matter (Fig. 8). Specifically, in vivo PDGFaR signaling is a

predominant regulator of OP proliferation in response to demyeli-

nation while FGF2 is a predominant inhibitor of OP differentiation

during remyelination. These effects have been observed within the

CNS white matter and in the course of experimental demyelination

with spontaneous remyelination. This model does not address

potential PDGF and/or FGF2 regulation of neural stem cell

responses that could occur in distinct sites. Furthermore, the

genetic deletions utilized here address removal of endogenous

PDGFaR and FGF2 signaling. Additional potential neural stem

cell and OLC responses may be elicited by genetic overexpression

or exogenous administration to elevate PDGF and FGF2 levels

(Lachapelle et al., 2002; Ruffini et al., 2001; Woodruff et al.,

2004). This model contrasts with PDGF and FGF2 roles predicted

from mainly in vitro studies that indicated a more significant

contribution of endogenous FGF2 in the OP proliferative response

to demyelination (Chari and Blakemore, 2002; Wolswijk and

Noble, 1995). Our previous in vitro analysis of OLC cells derived

from demyelinated lesions of wild type mice used pharmacological

reagents to transiently impair PDGFaR and/or FGF2 activity, with

comparison to administration of exogenous PDGF-AA and/or

FGF2 (Frost et al., 2003). This in vitro work also indicated an

FGF2 effect on both OLC differentiation and OP proliferation as

well as a cooperative effect of PDGF and FGF2 signaling (Frost et

al., 2003). Therefore, we speculate that the current difference

between previous in vitro studies and the current in vivo model

reflects a difference of signaling in the complexity of the in vivo

context. However, we cannot rule out the possibility that other

differences in the PDGFaR+/� or FGF2�/� mice could have

occurred during development and play a role in the current results.

We show that PDGFaR signaling is critical for normal OP

amplification in response to demyelination. However, generation

of oligodendrocytes from this reduced OP pool was improved by

removal of FGF2 inhibition of OLC differentiation. This finding is

extremely important with respect to potential treatment of diseases

in which oligodendrocyte regeneration is needed for remyelination.

To promote remyelination, transient attenuation of FGF2 signaling

might be sufficient for overcoming reduced OP density and/or a

lesion environment that lacks adequate support for fully exploiting

the potential for oligodendrocyte regeneration. In the current

disease model, as in most rodent experimental models, the

oligodendrocyte density is sufficient and the environment is

favorable for remyelination so that these cellular effects do not

appear to markedly alter remyelination (data not shown for

PDGFaR heterozygotes, current study; Armstrong et al., 2002).

Further studies will be needed to determine the extent to which the

cellular responses identified in this study can be exploited to

improve remyelination and recovery of function in less favorable

lesion contexts that may mimic human disease conditions.

With direct significance to the human context, OPs from human

fetal tissue are responsive to PDGF and express PDGFaR (Wilson

et al., 2003; Zhang et al., 2000). Furthermore, OPs that express

PDGFaR can be abundant in MS lesions (Maeda et al., 2001),

indicating the potential to respond to changes in PDGF expression.

FGF2 may also be present in MS lesions (Holley et al., 2003) and

may limit successful remyelination through inhibition of endoge-

nous OLC differentiation. PDGF and FGF2 are likely to regulate

OLC responses in vivo in the context of multiple signals in the

lesion environment, such as other growth factors, cytokine, and cell

adhesion molecules. Understanding the interactions between

various signals and the resulting OLC responses will be necessary

for the success of growth factor-based treatments to promote repair

in demyelinating diseases.

Acknowledgments

We thank Dr. Philippe Soriano for providing breeding pairs of

the PDGFaR targeted deletion mice, Dr. Thomas Doetschman for

providing breeding pairs of the FGF2 targeted deletion mice as

heterozygotes, Drs. William Stallcup and Minetta Gardinier for

antibodies, Dr. Fred Gage for providing pNIT-GFP 293 cells and

pMD.G plasmid, Dr. Steve Levison for advice on the retroviral

lineage tracing, and Drs. Lynn Hudson and William Richardson for

plasmids. We appreciate the comments of Dr. Nicole Dobson and

Dr. Joe Nielsen. This work was supported by NIH grant NS39293.

References

Allamargot, C., Pouplard-Barthelaix, A., Fressinaud, C., 2001. A single

intracerebral microinjection of platelet-derived growth factor (PDGF)

accelerates the rate of remyelination in vivo. Brain Res. 918 (1–2),

28–39.

Armstrong, R.C., Dorn, H.H., Kufta, C.V., Friedman, E., Dubois-Dalcq,

M.E., 1992. Pre-oligodendrocytes from adult human CNS. J. Neurosci.

12 (4), 1538–1547.

Armstrong, R.C., Le, T.Q., Frost, E.E., Borke, R.C., Vana, A.C., 2002.

Absence of fibroblast growth factor 2 promotes oligodendroglial

repopulation of demyelinated white matter. J. Neurosci. 22 (19),

8574–8585.

Bansal, R., Pfeiffer, S.E., 1997. Regulation of oligodendrocyte differentiation

by fibroblast growth factors. Adv. Exp. Med. Biol. 429, 69–77.

Bansal, R., Kumar, M., Murray, K., Morrison, R.S., Pfeiffer, S.E., 1996.

Regulation of FGF receptors in the oligodendrocyte lineage. Mol. Cell.

Neurosci. 7 (4), 263–275.

Barres, B.A., Raff, M.C., 1993. Proliferation of oligodendrocyte

precursor cells depends on electrical activity in axons. Nature 361

(6409), 258–260.

Barres, B.A., Hart, I.K., Coles, H.S., Burne, J.F., Voyvodic, J.T.,

Richardson, W.D., Raff, M.C., 1992. Cell death and control of cell

survival in the oligodendrocyte lineage. Cell 70 (1), 31–46.

J.C. Murtie et al. / Neurobiology of Disease 19 (2005) 171–182 181

Blakemore, W.F., Keirstead, H.S., 1999. The origin of remyelinating cells

in the central nervous system. J. Neuroimmunol. 98 (1), 69–76.

Bogler, O., Wren, D., Barnett, S.C., Land, H., Noble, M., 1990.

Cooperation between two growth factors promotes extended self-

renewal and inhibits differentiation of oligodendrocyte-type-2 astro-

cyte (O-2A) progenitor cells. Proc. Natl. Acad. Sci. U. S. A. 87 (16),

6368–6372.

Bruck, W., Kuhlmann, T., Stadelmann, C., 2003. Remyelination in multiple

sclerosis. J. Neurol. Sci. 206 (2), 181–185.

Calver, A.R., Hall, A.C., Yu, W.P., Walsh, F.S., Heath, J.K., Betsholtz, C.,

Richardson, W.D., 1998. Oligodendrocyte population dynamics and the

role of PDGF in vivo. Neuron 20 (5), 869–882.

Chang, A., Tourtellotte, W.W., Rudick, R., Trapp, B.D., 2002. Premyeli-

nating oligodendrocytes in chronic lesions of multiple sclerosis.

N. Engl. J. Med. 346 (3), 165–173.

Chari, D.M., Blakemore, W.F., 2002. New insights into remyelination

failure in multiple sclerosis: implications for glial cell transplantation.

Mult. Scler. 8 (4), 271–277.

Chellaiah, A., Yuan, W., Chellaiah, M., Ornitz, D.M., 1999. Mapping

ligand binding domains in chimeric fibroblast growth factor receptor

molecules. Multiple regions determine ligand binding specificity.

J. Biol. Chem. 274 (49), 34785–34794.

Ebner, S., Dunbar, M., McKinnon, R.D., 2000. Distinct roles for PI3K in

proliferation and survival of oligodendrocyte progenitor cells. J. Neuro-

sci. Res. 62 (3), 336–345.

Fields-Berry, S.C., Halliday, A.L., Cepko, C.L., 1992. A recombinant

retrovirus encoding alkaline phosphatase confirms clonal boundary

assignment in lineage analysis of murine retina. Proc. Natl. Acad. Sci.

U. S. A. 89 (2), 693–697.

Franklin, K.B.J., Paxinos, G., 1997. The Mouse Brain in Stereotaxic

Coordinates. Academic Press, San Diego.

Frost, E.E., Nielsen, J.A., Le, T.Q., Armstrong, R.C., 2003. PDGF and

FGF2 regulate oligodendrocyte progenitor responses to demyelination.

J. Neurobiol. 54 (3), 457–472.

Fruttiger, M., Karlsson, L., Hall, A.C., Abramsson, A., Calver, A.R.,

Bostrom, H., Willetts, K., Bertold, C.H., Heath, J.K., Betsholtz, C.,

et al., 1999. Defective oligodendrocyte development and severe

hypomyelination in PDGF-A knockout mice. Development 126 (3),

457–467.

Fuss, B., Mallon, B., Phan, T., Ohlemeyer, C., Kirchhoff, F., Nishiyama, A.,

Macklin, W.B., 2000. Purification and analysis of in vivo-differentiated

oligodendrocytes expressing the green fluorescent protein. Dev. Biol.

218 (2), 259–274.

Gensert, J.M., Goldman, J.E., 1997. Endogenous progenitors remyelinate

demyelinated axons in the adult CNS. Neuron 19 (1), 197–203.

Goldman, J.E., 2003. What are the characteristics of cycling cells in the

adult central nervous system? J. Cell. Biochem. 88 (1), 20–23.

Hart, I.K., Richardson, W.D., Heldin, C.H., Westermark, B., Raff, M.C.,

1989. PDGF receptors on cells of the oligodendrocyte-type-2 astrocyte

(O-2A) cell lineage. Development 105 (3), 595–603.

Hinks, G.L., Franklin, R.J., 1999. Distinctive patterns of PDGF-A, FGF-2,

IGF-I, and TGF-beta1 gene expression during remyelination of

experimentally-induced spinal cord demyelination. Mol. Cell. Neurosci.

14 (2), 153–168.

Holley, J.E., Gveric, D., Newcombe, J., Cuzner, M.L., Gutowski, N.J.,

2003. Astrocyte characterization in the multiple sclerosis glial scar.

Neuropathol. Appl. Neurobiol. 29 (5), 434–444.

Hudson, L.D., Berndt, J.A., Puckett, C., Kozak, C.A., Lazzarini, R.A.,

1987. Aberrant splicing of proteolipid protein mRNA in the dysmyeli-

nating jimpy mutant mouse. Proc. Natl. Acad. Sci. U. S. A. 84 (5),

1454–1458.

Keirstead, H.S., Levine, J.M., Blakemore, W.F., 1998. Response of the

oligodendrocyte progenitor cell population (defined by NG2 labelling)

to demyelination of the adult spinal cord. Glia 22 (2), 161–170.

Lachapelle, F., Avellana-Adalid, V., Nait-Oumesmar, B., Baron-Van

Evercooren, A., 2002. Fibroblast growth factor-2 (FGF-2) and

platelet-derived growth factor AB (PDGF AB) promote adult SVZ-

derived oligodendrogenesis in vivo. Mol. Cell. Neurosci. 20 (3),

390–403.

Linnington, C., Webb, M., Woodhams, P.L., 1984. A novel myelin-

associated glycoprotein defined by a mouse monoclonal antibody.

J. Neuroimmunol. 6 (6), 387–396.

Liu, X., Mashour, G.A., Webster, H.F., Kurtz, A., 1998. Basic FGF and

FGF receptor 1 are expressed in microglia during experimental

autoimmune encephalomyelitis: temporally distinct expression of

midkine and pleiotrophin. Glia 24 (4), 390–397.

Maeda, Y., Solanky, M., Menonna, J., Chapin, J., Li, W., Dowling, P., 2001.

Platelet-derived growth factor-alpha receptor-positive oligodendroglia

are frequent in multiple sclerosis lesions. Ann. Neurol. 49 (6), 776–785.

Mason, J.L., Ye, P., Suzuki, K., D’Ercole, A.J., Matsushima, G.K., 2000.

Insulin-like growth factor-1 inhibits mature oligodendrocyte apoptosis

during primary demyelination. J. Neurosci. 20 (15), 5703–5708.

Matsushima, G.K., Morell, P., 2001. The neurotoxicant, cuprizone, as a

model to study demyelination and remyelination in the central nervous

system. Brain Pathol. 11 (1), 107–116.

McKinnon, R.D., Matsui, T., Dubois-Dalcq, M., Aaronson, S.A., 1990.

FGF modulates the PDGF-driven pathway of oligodendrocyte develop-

ment. Neuron 5 (5), 603–614.

Messersmith, D.J., Murtie, J.C., Le, T.Q., Frost, E.E., Armstrong, R.C.,

2000. Fibroblast growth factor 2 (FGF2) and FGF receptor expression

in an experimental demyelinating disease with extensive remyelination.

J. Neurosci. Res. 62 (2), 241–256.

Muir, D.A., Compston, D.A., 1996. Growth factor stimulation triggers

apoptotic cell death in mature oligodendrocytes. J. Neurosci. Res. 44

(1), 1–11.

Oh, L.Y., Denninger, A., Colvin, J.S., Vyas, A., Tole, S., Ornitz, D.M.,

Bansal, R., 2003. Fibroblast growth factor receptor 3 signaling regulates

the onset of oligodendrocyte terminal differentiation. J. Neurosci. 23

(3), 883–894.

Ornitz, D.M., Leder, P., 1992. Ligand specificity and heparin dependence of

fibroblast growth factor receptors 1 and 3. J. Biol. Chem. 267 (23),

16305–16311.

Palmer, T.D., Markakis, E.A., Willhoite, A.R., Safar, F., Gage, F.H., 1999.

Fibroblast growth factor-2 activates a latent neurogenic program in

neural stem cells from diverse regions of the adult CNS. J. Neurosci. 19

(19), 8487–8497.

Penderis, J., Shields, S.A., Franklin, R.J., 2003. Impaired remyelination and

depletion of oligodendrocyte progenitors does not occur following

repeated episodes of focal demyelination in the rat central nervous

system. Brain 126 (Pt. 6), 1382–1391.

Redwine, J.M., Armstrong, R.C., 1998. In vivo proliferation of oligoden-

drocyte progenitors expressing PDGFalphaR during early remyelina-

tion. J. Neurobiol. 37 (3), 413–428.

Ruffini, F., Furlan, R., Poliani, P.L., Brambilla, E., Marconi, P.C., Bergami,

A., Desina, G., Glorioso, J.C., Comi, G., Martino, G., 2001. Fibroblast

growth factor-II gene therapy reverts the clinical course and the

pathological signs of chronic experimental autoimmune encephalomye-

litis in C57BL/6 mice. Gene Ther. 8 (16), 1207–1213.

Shi, J., Marinovich, A., Barres, B.A., 1998. Purification and character-

ization of adult oligodendrocyte precursor cells from the rat optic nerve.

J. Neurosci. 18 (12), 4627–4636.

Soriano, P., 1997. The PDGF alpha receptor is required for neural crest cell

development and for normal patterning of the somites. Development

124 (14), 2691–2700.

Wilson, H.C., Onischke, C., Raine, C.S., 2003. Human oligodendrocyte

precursor cells in vitro: phenotypic analysis and differential response to

growth factors. Glia 44 (2), 153–165.

Wolswijk, G., Noble, M., 1992. Cooperation between PDGF and FGF

converts slowly dividing O-2Aadult progenitor cells to rapidly dividing

cells with characteristics of O-2Aperinatal progenitor cells. J. Cell Biol.

118 (4), 889–900.

Wolswijk, G., Noble, M., 1995. In vitro studies of the development,

maintenance and regeneration of the oligodendrocyte-type-2 astrocyte

(O-2A) lineage in the adult central nervous system. In: Kettenmann, H.,

J.C. Murtie et al. / Neurobiology of Disease 19 (2005) 171–182182

Sutliff, R.L. (Eds.), Neuroglia. Oxford Univ. Press, New York,

pp. 149–161.

Woodruff, R.H., Fruttiger, M., Richardson, W.D., Franklin, R.J., 2004.

Platelet-derived growth factor regulates oligodendrocyte progenitor

numbers in adult CNS and their response following CNS demyelina-

tion. Mol. Cell. Neurosci. 25 (2), 252–262.

Zhang, S.C., Ge, B., Duncan, I.D., 2000. Tracing human oligodendroglial

development in vitro. J. Neurosci. Res. 59 (3), 421–429.

Zhou, M., Sutliff, R.L., Paul, R.J., Lorenz, J.N., Hoying, J.B.,

Haudenschild, C.C., Yin, M., Coffin, J.D., Kong, L., Kranias, E.G.,

et al., 1998. Fibroblast growth factor 2 control of vascular tone. Nat.

Med. 4 (2), 201–207.