Content uploaded by Nadine Hövelmeyer
Author content
All content in this area was uploaded by Nadine Hövelmeyer
Content may be subject to copyright.
Experimental autoimmune encephalomyelitis repressed by
microglial paralysis
Frank L Heppner
1
, Melanie Greter
1,2
, Denis Marino
1
, Jeppe Falsig
1
, Gennadij Raivich
3
, Nadine Hövelmeyer
4
,
Ari Waisman
4
, Thomas Rülicke
5
, Marco Prinz
1,7
, Josef Priller
6
, Burkhard Becher
2
& Adriano Aguzzi
1
Although microglial activation occurs in inflammatory, degenerative and neoplastic central nervous system (CNS) disorders, its
role in pathogenesis is unclear. We studied this question by generating CD11b-HSVTK transgenic mice, which express herpes
simplex thymidine kinase in macrophages and microglia. Ganciclovir treatment of organotypic brain slice cultures derived from
CD11b-HSVTK mice abolished microglial release of nitrite, proinflammatory cytokines and chemokines. Systemic ganciclovir
administration to CD11b-HSVTK mice elicited hematopoietic toxicity, which was prevented by transfer of wild-type bone marrow.
In bone marrow chimeras, ganciclovir blocked microglial activation in the facial nucleus upon axotomy and repressed the
development of experimental autoimmune encephalomyelitis. We conclude that microglial paralysis inhibits the development
and maintenance of inflammatory CNS lesions. The microglial compartment thus provides a potential therapeutic target in
inflammatory CNS disorders. These results validate CD11b-HSVTK mice as a tool to study the impact of microglial activation on
CNS diseases in vivo.
Microglial cells are of hematopoietic origin and populate the CNS
early during development to form a regularly spaced network of ram-
ified cells
1
. Microglia become rapidly activated in most pathologi-
cal conditions of the CNS. Although microglial activation has been
described extensively in many CNS diseases
1
, its impact on disease
pathogenesis remains ill defined
1,2
. In autoimmune diseases such as
multiple sclerosis, most data point to a detrimental role of microglia,
for example by producing neurotoxic molecules, proinflammatory
cytokines, chemokines or by presenting self-antigens
3–6
. But there have
been claims that microglial activation in other diseases may counteract
the pathogenic changes by providing neurotrophic or immunosup-
pressive factors
7,8
.
We wished to investigate the role of activated microglia using a
pharmacogenetically inducible in vivo model of microglial paralysis. We
have therefore generated transgenic mice in which the thymidine kinase
of herpes simplex virus (encoded by HSVTK) is driven by the CD11b
promoter
9
. CD11b, encoded by Itgam, is the alpha chain of the Mac-1
integrin and is expressed in cells of myeloid origin, including macro-
phages and microglia. We used a 1.7-kilobase CD11b promoter fragment
which drives sustained expression in macrophages of transgenic mice
at levels similar to those of the endogenous CD11b gene
9
. HSVTK is a
suicide gene that converts antiviral nucleotide analog prodrugs such
as ganciclovir (GCV) to a monophosphorylated form, which is then
transformed into a toxic triphosphate by endogenous cellular kinases
10
.
Expression of HSVTK renders preferentially proliferating cells sensi-
tive to GCV, as the active metabolite competes with thymine for DNA
synthesis. This strategy has been used to selectively ablate defined cell
lineages, for example in transgenic mice
11,12
.
Nondividing HSVTK
+
cells also show susceptibility to GCV, albeit
at reduced levels: in this case, toxicity has been ascribed to interference
with mitochondrial DNA synthesis
13
. Here we report that microglial
cells of CD11b-HSVTK transgenic mice, although mostly resting in the
adult CNS, are efficiently paralyzed by GCV administration. We took
advantage of this phenomenon to study the contribution of micro-
glial activation to two disease models: facial nerve transection and
experimental autoimmune encephalitis.
RESULTS
Characterization of transgenic mice
Before generating transgenic mice, the CD11b-HSVTK transgene was
stably transfected into the BV-2 microglial cell line, and GCV was
added to the culture medium. We detected efficient, dose-dependent
killing of microglial cells with a significant difference between BV2-
TK and controls at 2, 10 and 20 µm GCV (P < 0.001, one-way ANOVA
test), confirming the functionality of the suicide approach (Fig. 1a).
We then established two independent transgenic mouse lines denoted
B6,D2-Tg(CD11b-HSVTK)618Zbz (tg618) and B6,D2-Tg (CD11b-
HSVTK)620Zbz (tg620), and confirmed integration of the transgene by
1
Institute of Neuropathology, University Hospital Zurich, CH-8091 Zurich, Switzerland.
2
Department Neurology, Neuroimmunology Unit, University Hospital Zurich,
CH-8091 Zurich, Switzerland.
3
Perinatal Brain Repair Group, Department of Obstetrics and Gynaecology and Department of Anatomy, University College London,
WC1E 6HX London, UK.
4
Laboratory for Molecular Immunology, Institute for Genetics, University of Cologne, D-50931 Cologne, Germany.
5
Institute of Laboratory
Animal Science, University of Zurich, CH-8091 Zurich, Switzerland.
6
Departments of Psychiatry and Experimental Neurology, Charité, Humboldt-University Berlin,
10117 Berlin, Germany.
7
Present address: Institute of Neuropathology, Georg-August-University Göttingen, D-37075 Göttingen, Germany. Correspondence should be
addressed to A.A. (adriano@pathol.unizh.ch).
Published online 23 January 2005; doi:10.1038/nm1177
146 VOLUME 11
|
NUMBER 2
|
FEBRUARY 2005
NATURE MEDICINE
ARTICLES
© 2005 Nature Publishing Group http://www.nature.com/naturemedicine
NATURE MEDICINE VOLUME 11
|
NUMBER 2
|
FEBRUARY 2005 147
Southern blotting and PCR of genomic DNA (Fig. 1b,c). In the absence
of GCV, transgenic offspring showed no phenotypic alterations except
for male sterility, which is a known consequence of ectopic HSVTK
expression
14
. Line tg620 was backcrossed to C57BL/6 mice for nine
generations and was used for all of the experiments described below.
Expression of HSVTK was ascertained by western blot analysis
of various organs. As expected, all macrophage-containing organs,
including spleen, lung and brain, showed sustained expression of the
HSVTK transgene (Fig. 1d,e). Western blot analysis of purified micro-
glial and astroglial cells derived from CD11b-HSVTK mice established
that transgene expression within the CNS was present in microglial
cells but not in astrocytes (Fig. 1f). Addition of GCV to primary micro-
glial cell cultures (>98% purity, Fig. 1f), or to mixed glial cell cultures,
led to lineage-specific death of microglial cells (Fig. 1g,h).
We then administered GCV to tg620 mice intraperitoneally or in
drinking water. Circulating CD11b
+
cells were substantially reduced
after 3 d and almost entirely ablated after 5–6 d, whereas B- and T-cell
counts were essentially unaltered (Fig. 2a). After 7 d of GCV treatment,
tg620 mice developed fatal aplastic anemia with reduced erythroid and
myeloid cell compartments in both bone marrow and peripheral blood
(Supplementary Fig. 1 and Supplementary Table 1 online). Analysis
of prenatal hematopoiesis in tg620 embryos at embryonic day 14.5
showed GCV-mediated ablation of CD11b
+
AA4.1
+
hematopoietic
precursor cells
15
(Supplementary Fig. 1 online), which therefore seem
to be crucial for normal hematopoietic development.
Microglia-restricted HSVTK expression in tg620
chi
mice
To restrict HSVTK expression to microglial cells and to overcome
GCV-mediated myelotoxicity, we generated bone marrow chimeras.
For donors, we used fetal liver cells (FLCs) or bone marrow derived
from wild-type mice or congenic C57BL/6-β-actin-GFP mice
16
.
Lethally irradiated tg620 transgenic mice were used as recipients. This
strategy is based on the fact that monocytes and extraneural tissue
macrophages are rapidly and efficiently repopulated upon adoptive
bone marrow transfer, whereas recruitment of radioresistant microg-
lia from the peripheral hematopoietic pool to the CNS is slow and
rather inefficient
17–20
. Chimeric tg620 mice hosting wild-type bone
marrow (termed tg620
chi
) showed stable numbers of peripheral
CD11b
+
monocytes and macrophages upon long-term GCV treatment
(Fig. 2b). In reciprocal experiments, tg620 bone marrow was transferred
to wild-type recipients. The latter mice experienced GCV-mediated loss
of peripheral CD11b
+
cells and aplastic anemia (Fig. 2b) similar to tg620
mice. Intactness of the blood-brain barrier in both GCV-treated and
untreated tg620
chi
mice was shown by the lack of increased IgG influx
into the CNS
21
. This excluded the possibility of radiation-induced
blood-brain barrier leakage (Supplementary Fig. 2 online).
Microglial paralysis after GCV treatment
To assess the efficiency and specificity of paralyzing microglia in
tg620
chi
mice, we performed unilateral transections of the facial nerve.
a
b
c
d
e
f
g
h
CD11b
Figure 1 Characterization of CD11b-HSVTK transgene and transgenic mice.
(a) Cell viability assay of microglial BV-2 cell clones stably transfected with
CD11b-HSVTK (BV2-TK; circles, inverted triangles and diamonds).
CMV-GFP-transfected BV-2 (BV2-GFP; triangles) and untransfected
BV-2 controls (BV2; squares) served as negative controls. Ordinate: mean
and s.d. of cell numbers on day 6 divided by cell numbers on day 0.
(b) CD11b-HSVTK transgene depicting the CD11b promoter, the HSVTK gene
and the human growth hormone gene (hGH) providing splice donor/acceptor
sites and a polyadenylation signal
9
. For the Southern blot (below), an
HSVTK-specific 530-bp PCR probe (hatched box) was hybridized to a
fragment of 1169 bp (arrow on the left) in transgene-positive tg618
(lane 2: founder; lane 3: F1 offspring) and tg620 mice (lane 4: founder;
lane 6 and 7: F1 offspring) and not in negative littermates (lanes 1,5). As
positive control, 10-fold dilutions of CD11b-HSVTK plasmid were spiked in
genomic DNA of wild-type mice (ranging from 1,125 pg (lane 8) to 1.125 pg
(lane 11)). Larger bands in lanes 8 and 9 correspond to incompletely digested
plasmid. Right, molecular weight in kb. (c) PCR of genomic DNA resulted in
a transgene-specific band at 599 bp. (d) Western blot of brain tissue: HSVTK
transgene is expressed in CD11b-HSVTK (lane 3) and GFAP-HSVTK brains
12
(lane 2, positive control). No signal in negative littermates (lanes 1, 4). Right,
molecular weight in kDa. (e) Western blot of various organs: strong expression
of HSVTK in testis
14
(lane 1) and in macrophage-containing organs. Up to
four HSVTK proteins were detectable due to cryptic translational initiation
sites in the coding region of HSVTK
46
. (f) Western blot of isolectin IB
4
+
–
microglial (upper right, green color; left, light microscopic (LM) image) and
astrocyte cell culture lysates of tg620 mice shows HSVTK protein expression
in microglial cells (lane 5), but not in astrocytes (lane 3). Tg620 spleen
homogenate (lane 7) and stably CD11b-HSVTK-transfected microglial BV-2
cells (lane 2), positive controls; microglia and astrocytes of non-transgenic
littermates (lane 6 and 4) and GFP-transfected BV-2 cells (lane 1), negative
controls. (g) Strong reduction of transgenic microglial cells in the presence
(filled triangle), but not in the absence of GCV (filled squares), whereas wild-
type microglia in the presence (open squares) or absence (open circles) of
GCV were stable. Number of microglia was normalized to 100% at day 0 and
experiments were done in triplicates. (h) Whereas microglia (IB
4
+
, green)
were strongly reduced in mixed glial cell cultures derived from tg620 mice,
GFAP
+
astrocytes (red) appeared to be unaffected by GCV (lower left).
ARTICLES
© 2005 Nature Publishing Group http://www.nature.com/naturemedicine
148 VOLUME 11
|
NUMBER 2
|
FEBRUARY 2005 NATURE MEDICINE
Subsequent retrograde degeneration strongly
activates microglia within the respective facial
motor nucleus
22
. Here, GCV administration
to tg620
chi
mice considerably repressed acti-
vation of microglia (Fig. 3a), most likely as
a result of activation-induced upregulation
of CD11b-HSVTK which hypersensitizes
microglia to GCV. No differences were seen
in microglial activation within lesioned
facial nuclei of non-GCV-treated tg620 mice
or nontransgenic littermates in the presence
or absence of GCV (data not shown). GCV-
mediated microglial paralysis of tg620
chi
mice
for 7 weeks in the absence of a pathological
stimulus did not result in obvious phenotypic
or histopathological changes except for a
decrease in microglial cell numbers as assessed
by FACS analysis (Supplementary Fig. 3
online). Although it is theoretically possible
that ablation of microglia in nondiseased chi-
meric tg620 mice may result from radiation
toxicity, we consider this unlikely, as others
have found normal numbers of ramified, res-
ting microglia after irradiation at time points
similar to the ones described here
19
.
We then assessed the functional consequen-
ces of microglial paralysis in organotypic hippocampal slice cultures
(OHSCs). OHSCs derived from tg620 mice were stimulated with lipo-
polysaccharide (LPS) and interferon (IFN)-γ, and microglial activation
in the presence or absence of GCV was determined. GCV treatment
of stimulated OHSCs inhibited microglial activation (Supplementary
Fig. 4 online) and abrogated the release of nitrite, tumor necrosis factor
(TNF) and macrophage inflammatory protein (MIP)-1β (Fig. 3b–d).
These factors are thought to interfere with CNS homeostasis either
by damaging tissue directly, or by attracting and/or activating other
immune cells including autoreactive T cells in experimental autoim-
mune encephalitis (EAE)
5
.
Repression of EAE in bone marrow chimeric tg620 mice
Having established that GCV treatment of tg620
chi
mice resulted in
profound paralysis of microglia within the CNS, we studied the impact
of microglial paralysis on the pathogenesis of EAE. Age-matched
female tg620
chi
and wild-type mice were
immunized with an encephalitogenic syn-
thetic myelin oligodendrocyte glycoprotein
(MOG
35–55
) peptide. Untreated tg620
chi
mice
and wild-type mice reconstituted with wild-
ab
Figure 2 In vivo characterization of tg620 mice and FACS analysis of peripheral blood. (a) In tg620
(tk
+
) mice, we observed a drastic GCV-mediated decrease of CD11b
+
monocytes and granulocytes,
whereas B220
+
B cells and Thy1.2
+
T cells were unaffected (right panels). GCV treatment of
nontransgenic littermates (tk
–
), even at doses that were 10 times higher, did not elicit alterations in the
blood leukocyte numbers (left panels). (b) In tg620
chi
mice, the GCV-mediated loss of CD11b
+
cells
was rescued upon adoptive transfer of wild-type bone marrow (tk
–
→ tk
+
; middle panels). Tg620
chi
mice
were indistinguishable from nontransgenic littermates receiving the same treatment (tk
–
→ tk
–
; right
panels). Adoptive transfer of tg620 bone marrow into wild-type mice (tk
+
→ tk
–
; left panels) restored
susceptibility of CD11b
+
cells to GCV. Thy1.2
+
T cells and B220
+
B cells were unaffected in all
experimental groups. Numbers indicate percentage of cells. Arrows indicate transfer of bone marrow.
a
bcd
Overview Lesioned side Unlesioned side
Figure 3 Microglial paralysis in tg620 mice.
(a) GCV-mediated microglial paralysis within the
facial nucleus of tg620
chi
mice 7 d after axotomy
of the left facial nerve. Strong activation and
proliferation of CD11b
+
microglial cells (brown)
within the corresponding facial motor nucleus
of GCV-treated bone marrow chimeric wild-type
mice (tk
–
→ tk
–
; upper left and middle) was
seen versus complete inhibition of microglial
activation within lesioned facial nuclei of
GCV-treated tg620
chi
mice (tk
–
→ tk
+
; lower left
and middle). Activated (insert, upper middle
panel) but not paralyzed microglia (insert, lower
middle panel) was attached to cell bodies of
injured neurons. n = 3–4 mice/group. Circles
indicate the facial motor nuclei. Scale bars: left
panel, 1 mm; middle and right panels, 100 µm.
(b–d) GCV treatment abrogated the LPS-IFN-γ-
induced (black bars) release of microglial
TNF (b), nitrite (c) and MIP-1β (d) in OHSCs from
tg620 mice. White bars, nonstimulated OHSCs.
Statistical significance was assessed by a one-
way ANOVA test. ∗∗∗P < 0.001.
ARTICLES
© 2005 Nature Publishing Group http://www.nature.com/naturemedicine
NATURE MEDICINE VOLUME 11
|
NUMBER 2
|
FEBRUARY 2005 149
type bone marrow showed similar onset and severity of EAE (Fig. 4a),
whereas GCV treatment of tg620
chi
mice resulted in a considerable
delay in disease onset and in repression of clinical EAE signs (Fig. 4b;
P < 0.01–0.05, one-way ANOVA test including Bonferroni’s Multiple
Comparison post test). Conversely, comparison of transgenic tg620
mice with nontransgenic littermates (with or without GCV) showed
indistinguishable clinical disease development (Supplementary Fig. 5
online). GCV itself did not influence EAE, as the clinical course of bone
marrow–reconstituted wild-type littermates of tg620 mice was identi-
cal to that of wild-type bone marrow–reconstituted mice regardless of
GCV treatment (Fig. 4a).
Notably, GCV administration did not alter the cellular composition
and function of the peripheral immune system of tg620
chi
mice.
FACS analysis of peripheral blood showed no abnormalities
in relative numbers of Thy1.2
+
T cells, B220
+
B cells, as well as
CD11b
+
monocytes and granulocytes (data not shown). GCV treat-
ment of tg620
chi
and control mice also did not alter the prolifera-
tive response of T cells from mice immunized with keyhole limpet
hemocyanin (KLH; data not shown) or MOG in in vitro recall assays
(Fig. 4c and Supplementary Fig. 6 online).
In addition, levels of IFN-γ and interleukin
(IL)-2 were similar in GCV-treated versus
untreated tg620
chi
mice after recall with KLH
(data not shown) or MOG (P > 0.05, one-way
ANOVA and Tukey post test), except for a diffe-
rence in IFN-γ production between nonchime-
ric and chimeric mice, which can be attributed
to the reconstitution procedure (Fig. 4d). MOG
recall assays allowed us to directly address
the performance of encephalitogenic T cells
in response to their cognate antigen upon
in vivo priming. The results suggest that
T cells of GCV-treated chimeric tg620 mice
are fully capable of inducing expansion
and effector function of MOG-reactive
lymphocytes compared to their non-
GCV-treated counterparts (Fig. 4c,d and
Supplementary Fig. 6 online). These experi-
ments validate that GCV treatment does not
alter a number of critical immune functions of
tg620
chi
mice, including the ability to mount
an encephalitogenic immune response.
Histological examination showed that cli-
nical EAE signs correlated with the extent of
inflammatory CNS infiltrates (Fig. 5a–c).
Immunohistochemical staining for Iba1,
which identifies monocytes, macrophages
and microglia
23
, showed positive infiltrates
mainly within the spinal and cerebellar white
matter (Fig. 5a,b). Accordingly, both myelin
and axons seemed to be damaged (Fig. 5a).
But, in line with the clinical performance,
MOG-immunized tg620
chi
mice treated with
GCV showed few Iba1
+
infiltrates and did not
show major myelin or axonal destruction
(Fig. 5a). We detected very little activated
Iba1
+
microglia in the CNS of GCV-trea-
ted tg620
chi
mice, whereas all control mice
showed strongly activated Iba1
+
microglial
cells (Fig. 5a,b). MOG-immunized bone
marrow chimeric wild-type control ani-
mals, in the presence or absence of GCV, regularly showed strong
inflammatory changes (data not shown) indistinguishable from non-
GCV-treated tg620
chi
mice.
We then quantified and characterized inflammatory white- matter
infiltrates by FACS analysis of mononucleated cells derived from spinal
cord and brain stem tissue 14 d after MOG immunization (i.e., at the
peak of clinical signs). In line with the histological findings, we found
a high percentage of infiltrating cells in CNS tissues of non-GCV-trea-
ted tg620
chi
mice, which consisted of CD45
high
CD11b
–
lymphocytes
as well as of CD45
high
CD11b
+
myeloid cells. Further analysis showed
that approximately two-thirds of the CD45
high
CD11b
–
infiltrating
lymphocytes were CD4
+
cells, whereas only a minority consisted of
CD8
+
cells (Fig. 5c). GCV treatment of tg620
chi
mice substantially
reduced inflammatory infiltrates: lymphocytes were decreased from
36% to 10%, whereas the amount of infiltrating CD45
high
CD11b
+
monocytes was reduced by 40% (Fig. 5c). As expected, there were no
obvious differences in the numbers of CD45
intermediate
CD11b
+
micro-
glial cells in GCV-treated versus untreated tg620
chi
mice. Moreover,
the numbers of GFP
+
microglia recruited from the extraneural pool
Figure 4 Microglial paralysis represses clinical EAE in tg620
chi
mice. (a,b) Clinical EAE score of
bone marrow chimeric mice upon MOG immunization: all control mice showed severe EAE signs (a).
GCV-treated bone marrow chimeric tg620
chi
mice showed considerably repressed EAE (open circles).
Omission of GCV treatment of bone marrow chimeric tg620
chi
mice resulted in severe EAE (closed
circles) (b). Controls consisted of congenic β-actin-GFP (wt
GFP
) bone marrow chimeric wild-type mice (tk
–
)
treated (wt
GFP
→ tk
–
/GCV, blue triangle) or not treated (wt
GFP
→ tk
–
, yellow diamond) with GCV, tg620
chi
mice (wt
GFP
→ tk
+
, black circle) and wild-type mice with transgenic tg620 bone marrow (tk
+
→ tk
–
, red
diamond) and tg620 mice with transgenic tg620 bone marrow (tk
+
→ tk
+
, green square), the latter two
in the absence of GCV. Arrows indicate bone marrow transfer. Experiments were repeated three times for
bone marrow chimeric tg620 and wild-type mice with or without GCV and showed identical results.
n = 9–13 mice/group. (c,d) GCV-treated tg620
chi
mice are fully capable of driving T
H
1 immunity. Recall
proliferation (c) and IFN-γ secretion (d) of encephalitogenic T cells in response to MOG in the presence or
absence of GCV. Experimental groups consisted of tg620
chi
(filled black squares and filled red triangles),
wt
chi
(filled black triangles and filled black diamonds), wild-type (filled black circles and open squares)
and tg620 mice (open triangles) treated with GCV or not, as depicted. Each symbol represents the mean
for triplicate cultures; bars indicate the mean proliferation per experimental group. The proliferation index
is calculated as fold increase in T-cell proliferation upon MOG recall compared to medium only. A detailed
analysis of tg620
chi
T-cell proliferation upon MOG recall is shown in Supplementary Fig. 6 online.
ab
cd
ARTICLES
© 2005 Nature Publishing Group http://www.nature.com/naturemedicine
150 VOLUME 11
|
NUMBER 2
|
FEBRUARY 2005 NATURE MEDICINE
of macrophages in MOG-immunized mice were similar in the absence
or presence of GCV (Supplementary Fig. 7 online).
DISCUSSION
The data described here indicate that GCV administration to tg620
mice brings about efficient conditional paralysis of microglia in vivo.
Treatment with GCV in vivo results in substantial inhibition of microg-
lial activation upon facial axotomy, and abolishes the release of microg-
lia-derived nitrite and proinflammatory cytokines and chemokines
from activated brain-slice cultures. In EAE, microglial paralysis results
in substantial amelioration of the clinical signs and in strong reduction
of CNS inflammation. As facial nerve transection causes no disruption
of the blood-brain barrier
22
, we conclude that GCV diffuses adequately
into the intact CNS.
Immunomodulatory compounds, including IFN-β, copolymer I and
statins, can repress CNS autoimmune diseases including EAE
24–26
. But
these drugs exert additional ill-defined effects on many immune cells,
including T cells and peripheral monocytes. Inhibition of peripheral
macrophages has been shown to prevent EAE
27,28
, whereas microglial
activation is thought to be secondary to infiltration of lymphocytes
2
.
Accordingly, the pathogenesis of EAE is widely held to be triggered exclu-
sively within the peripheral immune system
2,29
.
Whereas the importance of autoreactive CD4
+
T cells in EAE has
been extensively documented, particularly by their ability to initiate
disease, the effector mechanism leading to inflammation and demye-
lination within the CNS is likely to be provided by other cell types
such as infiltrating macrophages and resident microglia
29
. But as a
result of the dearth of suitable animal models, a definitive dissection
of potential effector cell types involved, namely infiltrating macropha-
ges and resident microglia, has proven difficult. Early studies pointing
to a more direct role of CNS-derived cells in the pathogenesis of EAE
included the repression of EAE in bone marrow chimeric mice lacking
CD40 or IL-23 predominantly in the CNS
30,31
. But these studies did
not exclude other CNS residents than microglia as a potential source
of these immunomodulators.
What are the mechanisms of microglia-mediated damage in EAE?
Nitric oxide and its adducts may disrupt CNS tissue integrity
26,32
,
whereas microglia-derived cytokines and chemokines such as TNF
and MIP-1β activate and attract blood-derived leukocytes. These may
in turn interfere with CNS homeostasis (e.g., by damaging myelin)
5,26
.
Reactivation of myelin-specific T cells within the CNS upon recogni-
tion of local autoantigens is critical to induce and/or sustain EAE
33
.
But it is still debated whether microglia present myelin-associated
antigens to autoreactive T cells in vivo: it has been reported that the
antigen-presenting capacity of CD45
intermediate
CD11b
+
microglia is
not essential for promoting encephalitogenic myelin-specific CD4
+
T cells in vitro
34
, whereas others regard microglial presentation of
myelin-associated antigens as critical
6,35
. Our data clearly establish
a
bc
Figure 5 Microglial paralysis represses
EAE-associated inflammatory infiltrates in
tg620
chi
mice. (a) Spinal cords of MOG-
immunized and control mice (H&E, hematoxylin
and eosin). Inflammation was largely confined
to the white matter, leading to destruction of
myelin tracts (LN, Luxol-Nissl stain) as well
as axonal integrity (NF, neurofilament). No
inflammation was seen in tg620 mice in the
absence of MOG immunization (upper panel).
Scale bar, 200 µm except for left column,
500 µm. (b) Cerebella of MOG-immunized mice:
tg620
chi
mice (wt
GFP
→ tk
+
) in the absence of
GCV (second panel from top) or of wt
GFP
bone
marrow chimeric wild-type animals (wt
GFP
→ tk
–
)
treated with GCV (lower panel) or not (data not
shown) showed severe inflammation mainly
confined to the white matter. But tg620
chi
mice
treated with GCV (third panel from top) showed
neither activated Iba1
+
microglia nor infiltrating
leucocytes. No inflammation in tg620 mice
in the absence of MOG-immunization (upper
panel). Scale bar, 500 µm. (c) FACS analysis
of spinal cord and brain tissue 14 d after MOG
immunization. Whereas non-MOG-immunized
tg620
chi
mice without (upper row) or with GCV
(data not shown) showed few CD45
high
CD11b
–
lymphocytes in the CNS, MOG immunization
induced a strong inflammatory infiltration
in tg620
chi
mice in the absence of GCV
(middle row) consisting of CD45
high
CD11b
–
lymphocytes and CD45
high
CD11b
+
monocytes.
GCV-mediated microglial paralysis resulted in a
substantial reduction in inflammatory infiltrates
in tg620
chi
mice (lower row). Infiltrating
CD45
high
CD11b
–
lymphocytes consisted mainly
of CD4
+
T cells and few CD8
+
T cells (middle
and right). Numbers indicate percentages of
cells. Each experimental group consisted of
pooled homogenate derived from three mice
representing the average clinical EAE score.
ARTICLES
© 2005 Nature Publishing Group http://www.nature.com/naturemedicine
NATURE MEDICINE VOLUME 11
|
NUMBER 2
|
FEBRUARY 2005 151
that microglia are crucial for the development of EAE, which seems
to be conferred by the release of cytokines and chemokines as well
as reactive oxygen species. This view is in agreement with previous
reports showing that chemokine secretion by CNS residents in vivo
is vital for disease development
30
.
Even in the late neurodegenerative phase of EAE, which goes along
with destruction of neurons, activated microglia may have a crucial
role. Recent reports suggest an inverse relationship between microglial
activation and neurogenesis
36–38
. One might therefore speculate that
reduced neurogenesis, by decreasing the capacity for neural recovery,
may exacerbate EAE.
From a translational viewpoint, the results reported here suggest
that microglial activation may represent a target in the therapy of
multiple sclerosis and other immune-mediated CNS diseases. In
addition to providing insights into the pathogenesis of autoimmune
CNS diseases, tg620 mice are proving a useful tool for the study of
microglial involvement in a broad variety of neuroimmune and
neurodegenerative diseases, including encephalitides, Alzheimer,
Parkinson, prion and motor neuron diseases.
METHODS
Transgenic mice. A 544-base pair BamHI fragment containing the Thy1 cDNA
in pB203 derived from the CD11b-Thy1-human growth hormone (GH) cDNA
9
(provided by D.G. Tenen) was replaced by the BglII to BamHI HSVTK frag-
ment
39
. Correct construction of the transgene was confirmed by sequenc-
ing and by assessing GCV-mediated cell death of BV-2 microglial cells stably
transfected with CD11b-HSVTK using electroporation in accordance with
standard protocols. Following NotI and HindIII excision, we introduced
CD11b-HSVTK into fertilized B6D2F2 hybrid eggs by pronuclear microin-
jection. We analyzed genomic integration by Southern blotting of SmaI-
digested genomic DNA according to standard procedures. An HSVTK-specific
530-bp PCR probe was generated and labeled with
α32
P dCTP. The following
primers were used (Microsynth): 5´-ACAATGGGCATGCCTTATGC-3´ and
5´-GGACATATTGCACAAACGGA-3´. For routine genotyping, we performed
PCR analysis on tail DNA in line with standard protocols using the follow-
ing primers (Microsynth): 5´-GACTTCCGTGGCTTCTTGCTGC-3´ and
5´-GTGCTGGCATTACAGGCGTGAG-3´.
Bone marrow chimeric mice and GCV administration. We generated bone
marrow chimeric mice as described
30
with bone marrow of adult mice or FLCs,
embryonic day 13.5–14.5. Recipient mice were lethally irradiated with 950 rad
and injected intravenously either with 2 × 10
7
bone marrow cells or 1 × 10
7
FLCs. Engraftment took place over 6–8 weeks. We defined successful reconsti-
tution as >95% engraftment of blood leukocytes by FACS analysis. Congenic
C57BL/6 β-actin-GFP donor mice
16
(Jackson Laboratories) were bred in house
under pathogen-free conditions. We achieved GCV administration by intra-
peritoneal injection of 100 µg GCV (Cymevene; Roche) per gram of body mass
every 2 d or orally by adding 60 µg/ml GCV to the drinking water.
Western blot analysis. We prepared 10% (wt/vol) homogenates of various
organs or cell cultures according to standard procedures
40
. We used a polyclonal
rabbit serum to HSVTK
41
(1:5,000; provided by W.C. Summers) followed by
incubation with a rabbit IgG-HRP-specific antibody (1:2,500; Zymed). Equal
loading of protein (50 µg/lane) was assured by a bicinchoninic acid (BCA) assay
according to standard procedures.
Cell cultures. Cultures of purified microglia and astrocytes were prepared and
maintained as described
42,43
in the presence or absence of 2.5 µg/ml GCV.
Four high-power fields of three wells were counted on days 0, 3 and 6 using an
inverted microscope (Zeiss). We identified microglia by fluorescein isothio-
cyanate–labeled Isolectin-B4 (1:50; Sigma); astrocytes were stained with an
antibody to glial fibrillary acidic protein (1:300; Dako) and visualized by an
Alexa 546-labeled secondary rabbit IgG-specific antibody (1:300; Juro). We per-
formed fluorescence microscopy on a Zeiss microscope (Axioplan 2) equipped
with a digital camera (Axiocam).
Organotypic hippocampal slice cultures. OHSCs were prepared from
12-d-old mice as described
44
, and, where indicated, treated with 5 µg/ml GCV
from day 0 for the duration of the experiment. After 7 d, we initiated activation
of microglia by adding IFN-γ (10 ng/ml) and LPS (1 µg/ml). To assess nitric
oxide adducts, cytokines and chemokines, cell culture supernatant was har-
vested 48 h after LPS and IFN-γ stimulation except for TNF measurements,
for which supernatant was taken after 8 h. Each group contained 3–4 inserts
with 4 OHSCs on each insert.
Induction and evaluation of EAE. We subjected 13–16-week-old
mice to subcutaneous administration of 200 µg of MOG
35–55
peptide
(MEVGWYRSPFSRVVHLYRNGK; Neosystem) emulsified in complete Freund
adjuvant supplemented with 4 mg/ml Mycobacterium tuberculosis (DIFCO),
as described
30
. Mice received intraperitoneal injections with 200 ng pertussis
toxin (Sigma) at the time of immunization and 48 h later. GCV administra-
tion started 7 d before MOG immunization. Mice were scored as described
31
.
Animal experiments were approved by the Swiss Veterinary Office (#203/98,
40/2002, 85/2003 and 136/2004).
Histology. Whole mouse brains or spinal columns were fixed in 4% paraformal-
dehyde in phosphate-buffered saline and embedded in paraffin or snap-frozen
in liquid nitrogen. Antibodies to glial fibrillary acidic protein (1:300; Dako),
synaptophysin (1:50; Zymed), microtubule-associated protein-2 (1:1,000;
Sigma), neurofilament protein (200-kDa subunits; 1:20; Bio-Science) and Iba1
(1:100; Wako Chemicals) were used. We used CD11b-specific antibodies (MCA
711, 1:1,000; Serotec) only on cryosections.
Flow cytometry. We removed spinal cords and processed them for
FACS analysis as described
30,31
. Peripheral blood or spleen and FLC sus-
pensions were analyzed according to standard protocols
40
. Fluorescein
isothiocyanate–, phycoerythrin- or peridinin chlorophyll-a protein–
conjugated monoclonal antibodies to mouse CD11b, B220, CD8, CD4, Thy1.2,
AA4.1, CD45 or biotinylated CD8- or CD11c-specific antibodies were used,
the latter two coupled to a secondary streptavidin-labeled allophycocyanin
antibody (Becton Dickinson). Data were acquired on a FACScalibur
(Becton Dickinson).
Recall responses. Mice were primed by flank injections of 200 µg KLH (Sigma)
or 200 µg MOG
35–55
peptide (Neosystem) emulsified in complete Freund
adjuvant. Mice received intraperitoneal injections of 200 ng pertussis toxin
(Sigma) at the time of immunization. After 7 d, the axillary and inguinal lymph
nodes were removed and homogenized. We placed 2 × 10
5
lymph node cells
as triplicates in a 96-well plate and pulsed them with 100 µg KLH or MOG
or 10 µg ConA (Sigma). After 24 h, cells were pulsed with
3
[H]-thymidine
(NEN-DuPont; final concentration 5 µCi/ml) and incubated for an additional
24 h before they were harvested and thymidine incorporation was assessed
using a Filtermate harvester (Packard Meriden) and TopCount-NXT Packard
Microplate scintillation and luminescence counter. Supernatants of sister cul-
tures were harvested after 48 h and analyzed by ELISA.
Enzyme-linked immunosorbent assay (ELISA) and nitrite measurement.
Supernatants derived from recall assays or OHSCs were analyzed by the use
of ELISA kits for TNF, IFN-γ, IL-2 (Pharmingen) or MIP-1β (R&D systems)
according to the manufacturer’s instructions. Nitrite was measured with the
Griess reagent. We mixed 50 µl supernatant or NaNO
2
standards with 25 µl
N-(1-naphthyl)ethylenediamine (0.1% in water) and 25 µl sulfanilamide (1%
in 1.2 N HCl) in a 96-well plate and the optical density was assessed after 3 min
at 570–690 nm. To include the contribution of the NO metabolite nitrate, we
added 50 µl vanadium(III) chloride (8 mg/ml in 1M HCL) to the Griess reagent
and incubated it at 37 °C for 30 min
45
.
Facial nerve axotomy. We anesthetized 13–16-week-old mice with ketamin/
rompun according to published protocols. The left facial nerve was transected
at the stylomastoid foramen, and the animals were killed using CO
2
after a
survival time of 7 d
22
. Successful axotomy was assumed on the basis of immo-
bile whiskers on the lesioned side. GCV administration started 1 week before
axotomy. Facial nerve axotomy experiments were approved by the UK Home
Office (PPL 70/5341 to G.R.).
ARTICLES
© 2005 Nature Publishing Group http://www.nature.com/naturemedicine
152 VOLUME 11
|
NUMBER 2
|
FEBRUARY 2005 NATURE MEDICINE
Note: Supplementary information is available on the Nature Medicine website.
ACKNOWLEDGMENTS
We thank D.G. Tenen for providing Itgam-Thy1-hGH cDNA, T. Bush for supplying
GFAP-HSVTK control brains, C. Weissmann for discussion, J. Weber, P. Schwarz,
A. Schifferli, M. König for technical assistance and C. Sigurdson for critical
comments on the manuscript. This work is supported by grants of the Bundesamt
für Bildung und Wissenschaft (EU), the Swiss National Foundation, the US National
Prion Research Program, and the National Center of Competence in Research
(NCCR) on neural plasticity and repair to A.A. F.L.H. was supported by the Human
Frontier Science Program (HFSP), the Stammbach and the Leopoldina foundations.
M.G. is a fellow of the Roche Research Foundation of Switzerland. B.B. is a Harry
Weaver Neuroscience Scholar of the US National Multiple Sclerosis Society (NMSS).
COMPETING INTERESTS STATEMENT
The authors declare that they have no competing financial interests.
Received 25 October; accepted 2 December 2004
Published online at http://www.nature.com/naturemedicine/
1. Kreutzberg, G.W. Microglia: a sensor for pathological events in the CNS. Trends
Neurosci. 19, 312–318 (1996).
2. Weiner, H.L. & Selkoe, D.J. Inflammation and therapeutic vaccination in CNS diseases.
Nature 420, 879–884 (2002).
3. Carson, M.J. Microglia as liaisons between the immune and central nervous systems:
functional implications for multiple sclerosis. Glia 40, 218–231 (2002).
4. Becher, B., Prat, A. & Antel, J.P. Brain-immune connection: immuno-regulatory pro-
perties of CNS-resident cells. Glia 29, 293–304 (2000).
5. Ambrosini, E. & Aloisi, F. Chemokines and glial cells: a complex network in the central
nervous system. Neurochem. Res. 29, 1017–1038 (2004).
6. Ulvestad, E. et al. Human microglial cells have phenotypic and functional charac-
teristics in common with both macrophages and dendritic antigen-presenting cells.
J. Leukoc. Biol. 56, 732–740 (1994).
7. Schwartz, M. & Moalem, G. Beneficial immune activity after CNS injury: prospects for
vaccination. J. Neuroimmunol. 113, 185–192 (2001).
8. Kerschensteiner, M., Stadelmann, C., Dechant, G., Wekerle, H. & Hohlfeld, R.
Neurotrophic cross-talk between the nervous and immune systems: implications for
neurological diseases. Ann. Neurol. 53, 292–304 (2003).
9. Dziennis, S. et al. The CD11b promoter directs high-level expression of reporter genes
in macrophages in transgenic mice [published erratum appears in Blood 1995 Apr
1;85(7):1983]. Blood 85, 319–329 (1995).
10. Fyfe, J.A., Keller, P.M., Furman, P.A., Miller, R.L. & Elion, G.B. Thymidine kinase from
herpes simplex virus phosphorylates the new antiviral compound, 9-(2-hydroxyethoxy
methyl)guanine. J. Biol. Chem. 253, 8721–8727 (1978).
11. Culver, K.W. et al. In vivo gene transfer with retroviral vector-producer cells for treatment
of experimental brain tumors. Science 256, 1550–1552 (1992).
12. Bush, T.G. et al. Fulminant jejuno-ileitis following ablation of enteric glia in adult
transgenic mice. Cell 93, 189–201 (1998).
13. Herraiz, M. et al. Liver failure caused by herpes simplex virus thymidine kinase plus
ganciclovir therapy is associated with mitochondrial dysfunction and mitochondrial
DNA depletion. Hum. Gene Ther. 14, 463–472 (2003).
14. Braun, R.E. et al. Infertility in male transgenic mice: disruption of sperm development
by HSV-tk expression in postmeiotic germ cells. Biol. Reprod. 43, 684–693 (1990).
15. Fisher, R.C., Lovelock, J.D. & Scott, E.W. A critical role for PU.1 in homing and
long-term engraftment by hematopoietic stem cells in the bone marrow. Blood 94
,
1283–1290 (1999).
16. Okabe, M., Ikawa, M., Kominami, K., Nakanishi, T. & Nishimune, Y. ‘Green mice’ as
a source of ubiquitous green cells. FEBS Lett. 407, 313–319 (1997).
17. Hickey, W.F. & Kimura, H. Perivascular microglial cells of the CNS are bone marrow-
derived and present antigen in vivo. Science 239, 290–292 (1988).
18. Eglitis, M.A. & Mezey, E. Hematopoietic cells differentiate into both microglia and
macroglia in the brains of adult mice. Proc. Natl. Acad. Sci. USA 94, 4080–4085
(1997).
19. Priller, J. et al. Targeting gene-modified hematopoietic cells to the central nervous
system: use of green fluorescent protein uncovers microglial engraftment. Nat. Med.
7, 1356–1361 (2001).
20. Asheuer, M. et al. Human CD34+ cells differentiate into microglia and express recom-
binant therapeutic protein. Proc. Natl. Acad. Sci. USA 101, 3557–3562 (2004).
21. Pedchenko, T.V. & LeVine, S.M. IL-6 deficiency causes enhanced pathology in Twitcher
(globoid cell leukodystrophy) mice. Exp. Neurol. 158, 459–468 (1999).
22. Raivich, G. et al. Immune surveillance in the injured nervous system: T-lymphocytes
invade the axotomized mouse facial motor nucleus and aggregate around sites of
neuronal degeneration. J. Neurosci. 18, 5804–5816 (1998).
23. Sasaki, Y., Ohsawa, K., Kanazawa, H., Kohsaka, S. & Imai, Y. Iba1 is an actin-cross-
linking protein in macrophages/microglia. Biochem. Biophys. Res. Commun. 286,
292–297 (2001).
24. Lassmann, H., Bruck, W. & Lucchinetti, C. Heterogeneity of multiple sclerosis pathoge-
nesis: implications for diagnosis and therapy. Trends Mol. Med. 7, 115–121 (2001).
25. Youssef, S. et al. The HMG-CoA reductase inhibitor, atorvastatin, promotes a Th2 bias
and reverses paralysis in central nervous system autoimmune disease. Nature 420,
78–84 (2002).
26. Steinman, L., Martin, R., Bernard, C., Conlon, P. & Oksenberg, J.R. Multiple sclerosis:
deeper understanding of its pathogenesis reveals new targets for therapy. Annu. Rev.
Neurosci. 25, 491–505 (2002).
27. Tran, E.H., Hoekstra, K., van Rooijen, N., Dijkstra, C.D. & Owens, T. Immune invasion
of the central nervous system parenchyma and experimental allergic encephalomyelitis,
but not leukocyte extravasation from blood, are prevented in macrophage-depleted
mice. J. Immunol. 161, 3767–3775 (1998).
28. Martiney, J.A. et al. Prevention and treatment of experimental autoimmune encephalo-
myelitis by CNI-1493, a macrophage-deactivating agent. J. Immunol. 160, 5588–5595
(1998).
29. Benveniste, E.N. Role of macrophages/microglia in multiple sclerosis and experimental
allergic encephalomyelitis. J. Mol. Med.
75, 165–173 (1997).
30. Becher, B., Durell, B.G., Miga, A.V., Hickey, W.F. & Noelle, R.J. The clinical course of
experimental autoimmune encephalomyelitis and inflammation is controlled by the
expression of CD40 within the central nervous system. J. Exp. Med. 193, 967–974
(2001).
31. Becher, B., Durell, B.G. & Noelle, R.J. IL-23 produced by CNS-resident cells controls
T cell encephalitogenicity during the effector phase of experimental autoimmune ence-
phalomyelitis. J. Clin. Invest. 112, 1186–1191 (2003).
32. Willenborg, D.O., Staykova, M.A. & Cowden, W.B. Our shifting understanding of the
role of nitric oxide in autoimmune encephalomyelitis: a review. J. Neuroimmunol. 100,
21–35 (1999).
33. Flugel, A. et al. Migratory activity and functional changes of green fluorescent effector
cells before and during experimental autoimmune encephalomyelitis. Immunity 14,
547–560 (2001).
34. Ford, A.L., Goodsall, A.L., Hickey, W.F. & Sedgwick, J.D. Normal adult ramified micro-
glia separated from other central nervous system macrophages by flow cytometric
sorting. Phenotypic differences defined and direct ex vivo antigen presentation to
myelin basic protein-reactive CD4
+
T cells compared. J. Immunol. 154, 4309–4321
(1995).
35. Matyszak, M.K. et al. Microglia induce myelin basic protein-specific T cell anergy or
T cell activation, according to their state of activation. Eur. J. Immunol. 29, 3063–
3076 (1999).
36. Ekdahl, C.T., Claasen, J.H., Bonde, S., Kokaia, Z. & Lindvall, O. Inflammation is detri-
mental for neurogenesis in adult brain. Proc. Natl. Acad. Sci. USA 100, 13632–13637
(2003).
37. Monje, M.L., Toda, H. & Palmer, T.D. Inflammatory blockade restores adult hippocam-
pal neurogenesis. Science 302, 1760–1765 (2003).
38. Marin-Teva, J.L. et al. Microglia promote the death of developing Purkinje cells. Neuron
41, 535–547 (2004).
39. Mansour, S.L., Thomas, K.R. & Capecchi, M.R. Disruption of the proto-oncogene int-
2 in mouse embryo-derived stem cells: a general strategy for targeting mutations to
non-selectable genes. Nature 336, 348–352 (1988).
40. Heppner, F.L. et al. Prevention of Scrapie Pathogenesis by Transgenic Expression of
Anti-Prion Protein Antibodies. Science 294, 178–182 (2001).
41. Black, M.E., Newcomb, T.G., Wilson, H.M. & Loeb, L.A. Creation of drug-specific herpes
simplex virus type 1 thymidine kinase mutants for gene therapy. Proc. Natl. Acad. Sci.
USA 93, 3525–3529 (1996).
42. Hailer, N.P., Heppner, F.L., Haas, D. & Nitsch, R. Astrocytic factors deactivate antigen
presenting cells that invade the central nervous system. Brain Pathol. 8, 459–474
(1998).
43. Heppner, F.L., Roth, K., Nitsch, R. & Hailer, N.P. Vitamin E induces ramification and
downregulation of adhesion molecules in cultured microglial cells. Glia 22, 180–188
(1998).
44. Heppner, F.L., Skutella, T., Hailer, N.P., Haas, D. & Nitsch, R. Activated microglial cells
migrate towards sites of excitotoxic neuronal injury inside organotypic hippocampal
slice cultures. Eur. J. Neurosci. 10, 3284–3290 (1998).
45. Miranda, K.M., Espey, M.G. & Wink, D.A. A rapid, simple spectrophotometric method
for simultaneous detection of nitrate and nitrite. Nitric Oxide 5, 62–71 (2001).
46. Ellison, A.R. & Bishop, J.O. Herpesvirus thymidine kinase transgenes that do not cause
male sterility are aberrantly transcribed and translated in the testis. Biochim. Biophys.
Acta 1442, 28–38 (1998).
ARTICLES
© 2005 Nature Publishing Group http://www.nature.com/naturemedicine
