cns myeloid dcs presenting endogenous myelin peptides 'preferentially' polarize cd4+ th-17...
TRANSCRIPT
CNS myeloid DCs presenting endogenous myelinpeptides ‘preferentially’ polarize CD4+ TH-17 cellsin relapsing EAE
Samantha L Bailey1, Bettina Schreiner1, Eileen J McMahon2 & Stephen D Miller1
Peripherally derived CD11b+ myeloid dendritic cells (mDCs), plasmacytoid DCs, CD8a+ DCs and macrophages accumulate in
the central nervous system during relapsing experimental autoimmune encephalomyelitis (EAE). During acute relapsing EAE
induced by a proteolipid protein peptide of amino acids 178–191, transgenic T cells (139TCR cells) specific for the relapse
epitope consisting of proteolipid protein peptide amino acids 139–151 clustered with mDCs in the central nervous system, were
activated and differentiated into T helper cells producing interleukin 17 (TH-17 cells). CNS mDCs presented endogenously
acquired peptide, driving the proliferation of and production of interleukin 17 by naive 139TCR cells in vitro and in vivo. The
mDCs uniquely biased TH-17 and not TH1 differentiation, correlating with their enhanced expression of transforming growth
factor-b1 and interleukins 6 and 23. Plasmacytoid DCs and CD8a+ DCs were superior to macrophages but were much less
efficient than mDCs in presenting endogenous peptide to induce TH-17 cells. Our findings indicate a critical function for CNS
mDCs in driving relapses in relapsing EAE.
Epitope spreading is the process by which reactivity to epitopesdistinct from and not cross-reactive with the disease-initiating epitopeis induced during chronic inflammatory processes1. Epitope spreadinghas been described in multiple sclerosis2 and has functional impor-tance in driving the progression of relapsing experimental auto-immune encephalomyelitis (EAE)3–6 and Theiler’s murine encephalitisvirus–induced demyelinating disease7–9. EAE is a T cell–mediatedcentral nervous system (CNS) autoimmune disease widely used as amodel of multiple sclerosis, sharing its clinical, immunological andpathological characteristics10. CD4+ T helper type 1 (TH1) cellsproducing interferon-g (IFN-g) were previously thought to mediateEAE, but data have now shown that neutralizing interleukin 17(IL-17), not IFN-g, abrogates the progression of clinical disease11–13.IL-17-producing effector CD4+ T cells (TH-17 cells) are distinct fromthe TH1 and TH2 subsets and have been associated with many chronicinflammatory and autoimmune diseases, including EAE, arthritis,colitis and asthma11,14–18. Transforming growth factor-b (TGF-b)and IL-6 are required for the generation of TH-17 cells from naiveCD4+ T cells, and IL-23 supports the survival and populationexpansion of differentiated TH-17 cells17,19,20.
In noninflammatory conditions, immune responses to antigens inthe CNS are delayed or absent, and the lack of lymphatic vessels andlocal production of immunosuppressive factors are thought to con-tribute to the ‘immune-privileged’ state21. EAE is induced by primingof the activation and population expansion of CD4+ T cells in
response to CNS myelin antigens normally sequestered behind theblood-brain barrier. Re-presentation of myelin epitopes by CNSantigen-presenting cells (APCs) is required for the initiation ofdisease22–24. In a model in which dendritic cells (DCs) are the onlycells expressing major histocompatibility complex (MHC) class II, DCsalone are sufficient to reactivate myelin specific T cells and initiateEAE25. DCs are sparse in the healthy CNS and are found mainlyin vessel-rich areas such as the meninges and choroid plexus26–29.CNS inflammation induced by autoimmunity or infection is accom-panied by a considerable increase in DC numbers and tissue infiltra-tion24,28–32, but the function of inflammatory CNS DCs in drivingT cell activation is a contentious issue25,30,31. In relapsing EAE andTheiler’s murine encephalitis virus–induced demyelinating disease, thede novo population expansion of naive, CD4+ T cells specific forspread myelin epitopes takes place in the CNS and not in the local orperipheral secondary lymphoid tissues31. Moreover, CNS DCs, notresident microglia or infiltrating macrophages, are the most efficientin driving the activation of those naive myelin-specific CD4+ T cells31.Notably, the CNS DCs are a mixed population of phenotypicallydiscrete types that may have differed in their ability to activate spreadepitope-specific T cells in the CNS31.
Here we identify three discrete populations of DCs in the CNSduring relapsing EAE: CD11b+ myeloid DCs (mDCs), plasmacytoidDCs (pDCs) and CD8a+ DCs (CD8 DCs). Analysis of the function ofthose DC subsets in promoting epitope spreading indicated that CNS
Received 2 August 2006; accepted 30 November 2006; published online 7 January 2007; doi:10.1038/ni1430
1Department of Microbiology-Immunology and the Interdepartmental Immunobiology Center, Northwestern University Medical School, Chicago, Illinois 60611,USA. 2Present address: Biology Department, Westmont College, Santa Barbara, California 93108, USA. Correspondence should be addressed to S.D.M.([email protected]).
172 VOLUME 8 NUMBER 2 FEBRUARY 2007 NATURE IMMUNOLOGY
A R T I C L E S©
2007
Nat
ure
Pub
lishi
ng G
roup
ht
tp://
ww
w.n
atur
e.co
m/n
atur
eim
mun
olog
y
mDCs during the peak acute phase of relapsing EAE induced by aproteolipid peptide of amino acids 178–191 (PLP (178–191)) speci-fically interacted with 5B6 PLP(139–151)-specific T cells (139TCR) inperivascular cuffs in vivo and presented endogenously acquired myelinantigen to efficiently induce naive 139TCR cells to produce largeamounts of IL-17 in vitro and in vivo. The TH-17-inducing biasof CNS mDCs correlated with their high expression of TGF-b1and secretion of IL-6 and IL-23 after CD40 crosslinking. The pDCsand CD8 DCs also infiltrated the inflamed CNS from the blood butaccumulated in distinct locations and were much less efficient thanmDCs in presenting endogenous myelin peptide and in activatingnaive CD4+ T cells to proliferate and produce IL-17. Macrophagesinternalized PLP in the inflamed CNS and induced marginal proli-feration and IFN-g production by a PLP(139–151)-specific CD4+ Tcell line but were poor activators of myelin-specific naive CD4+ T cells.
RESULTS
DCs accumulate in inflamed CNS
Spontaneous remission and relapse and T cell epitope spreading is ahallmark of EAE induced in SJL mice (relapsing EAE)1. As with CD4+
T cells (data not shown) and macrophages, the abundance of DCs inthe CNS correlated with clinical disease severity. CNS DCs appearedduring the acute phase of relapsing EAE, their numbers plummetedduring remission, and they reappeared during disease relapse(Supplementary Fig. 1 online). Macrophage numbers were lowerduring relapse than during the acute phase, whereas DC numbersrecovered to acute numbers. DCs are known to be potent for inducingprimary T cell responses, so their presence in the CNS during naiveCD4+ T cell activation31 warranted further investigation of theirproperties and function. We found three distinct CNS DC populations
during the clinical stages of relapsing EAE: CD45hiCD3e–CD11b+F4/80+CD11c+ mDCs; CD11b–CD11c+B220+PDCA-1+ CD8a+ or CD8a–
pDCs; and CD11b–CD11c+B220–CD8a+ CD205+ or CD205– CD8 DCsubpopulations (Supplementary Fig. 1). The mDCs were the mostabundant during the acute phase of relapsing EAE, comprising 10.8%of the mononuclear cells isolated from the CNS; CD45hiCD11c–
CD11b+ macrophages comprised 8.3%, pDCs comprised 5.4% andCD8 DCs comprised 0.8% of CNS cells (Supplementary Fig. 1). Wefound mDCs, pDCs and CD8 DCs in the CNS of C57BL/6 (B6) micewith chronic EAE induced by a myelin oligodendrocyte glycoproteinpeptide of amino acids 35–55 (MOG(35–55)) and in SJL mice withrelapsing EAE induced by adoptive transfer of PLP(139–151)-specificT cell blasts (data not shown). Thus, DCs accumulated in the inflamedCNS of mice with both chronic and relapsing-remitting EAE.
CNS DCs are of peripheral origin and interact with T cells
CNS DCs could have arisen from two main sites: from precursors inthe bone marrow that transverse the blood brain barrier, or fromCNS-resident cells32–34. We used bone marrow chimeras to determinethe origin of the CNS DCs during EAE. Cells in the CNS are resistantto radiation35 and therefore retain the host genotype. During the acutephase of relapsing EAE in (B6 � SJL) F1 mice, 84% of mDCs arosefrom the bone marrow, as did 98% of CD11b– DCs (Fig. 1).Forty-eight percent of CD11b+CD11c– microglia or macrophageswere of donor bone marrow origin. Most of those cells wouldbe macrophages, because in chimeras generated by adoptive transferof bone marrow from mice expressing green fluorescent protein-tagged b-actin (GFP–b-actin) into SJL recipient mice, CD45expression was measured accurately and CD45hiCD11b+CD11c–
macrophages were of bone marrow origin (data not shown). Weobtained similar results during the acute phases of EAE in chimericmice using two other strategies: chimeras generated by transfer ofGFP–b-actin bone marrow into SJL mice with PLP(139–151)-inducedrelapsing EAE, and chimeras generated by transfer of B6 CD45.1donor bone marrow into B6 CD45.2 mice with MOG(35–55)-inducedchronic EAE (data not shown).
Immunohistochemical analysis of the spinal cord and the cerebel-lum showed that DCs accumulated in the center of demyelinated areas(Fig. 2a–d). The mDCs dominated in a central position in perivas-cular inflammatory foci, whereas microglia and macrophages weremarginal to those foci (Fig. 2e–h). In the brain, pDCs accumulated incerebellar meningeal foci (Fig. 2i,j), and in the spinal cord, pDCs weresparsely distributed, infiltrating deep into the parenchyma distantfrom inflammatory foci containing nucleated cells and non-pDCs(Fig. 2k). CD8 DCs comprised less than 1% of CNS cells (Supple-mentary Fig. 1) and are not presented here because of the incon-sistency in locating them.
We next sought to determine where CD4+ T cells specific for relapse-associated epitopes accumulated during acute EAE. In the late acutestages (day 16) of relapsing EAE induced by PLP(178–191), epitope
5.1 ± 0.9%
0.6% 0.2%
1.3 ± 0.2% 38.9 ± 8.5%
86.5 ± 12.4%
11 ± 0.8%
1.2 ± 0.1% 0.3%
4.2 ± 0.2% 31.7 ± 6.5%
47.7 ±7.6
36.8 ± 7.67.5 ± 1.484.4 ± 6.6
72 ± 2 91.5 ± 1.5 2.5 ± 2.5 4.1 ± 0.8 84.1 ± 826.7 ± 2.7
1 ± 1
6.3 ± 0.7 3.8 ±3.8
0.3 ± 0.393.2 ± 0.9
CD45.1 PE
CD
45.2
FIT
C
F1
> F
1: O
VA
(323
–339
)F
1 >
F1:
MO
G(3
5–55
)B
6 >
F1:
OV
A(3
23–3
39)
B6
> F
1: M
OG
(35–
55)
88 ± 4.2%
58.2 ± 12.788.5 ± 7
93.7 ± 1 0.1 ± 0.199.9 ± 0.1 63.2 ± 2.9
98 ± 0.6 1.5 ± 0.6
Figure 1 CNS DCs are of peripheral origin. Chimeric (B6 � SJL) F1 mice
engrafted with B6 bone marrow (CD45.2; B6 4 F1) or self bone marrow
(CD45.1 and CD45.2; F1 4 F1) were immunized with MOG(35–55) or
with OVA(323–339) (control). At the acute phase of EAE (day 18 after
immunization), CNS cells were analyzed by flow cytometry. Plots are gated
for mDCs (left), CD11b– DCs (pDCs and CD8 DCs; middle) and macrophages
and/or microglia (CD11c–CD11b+; right). Insets, isotype controls. FITC,
fluorescein isothiocyanate; PE, phycoerythrin. Above plots, percent ofeach CNS population; numbers in quadrants indicate percent cells in
each (mean ± s.e.m of two of experiments with five mice (B6 4 F1) or
three mice (F1 4 F1)).
NATURE IMMUNOLOGY VOLUME 8 NUMBER 2 FEBRUARY 2007 173
A R T I C L E S©
2007
Nat
ure
Pub
lishi
ng G
roup
ht
tp://
ww
w.n
atur
e.co
m/n
atur
eim
mun
olog
y
spreading in response to PLP(139–151) isdetected by delayed-type hypersensitivity31.Using a mixed bone marrow chimera strategy,we found that Thy-1.1+CD4+ PLP(139–151)-specific transgenic T cells (139TCR cells) wereassociated with CNS CD11c+ DCs (Fig. 2l–n), clustering mainly with mDCs (Fig. 2o) insections of the lower lumbar spinal cord.Similar patterns were also evident in thecerebellum, where the 139TCR cells accumu-lated with mDCs in white matter lesions (data not shown). Moreover,most (60.6%) 139TCR cells were activated (CD45RBlo; Fig. 2p) in theCNS during the acute phase of PLP(178–191)-induced relapsing EAE.Thus, most DCs in the inflamed CNS were of peripheral origin, andmDCs accumulated in perivascular spaces, clustering with CD4+
PLP(139–151)-specific transgenic T cells.
CNS mDCs efficiently present endogenous myelin peptides
To determine ability of the DC subpopulations to present endogenousantigens and activate naive myelin-specific T cells in the CNS, wecultured various numbers of mDCs, pDCs, CD8 DCs and macro-phages purified from the CNS of PLP(178–191)-primed mice at thepeak of disease with a fixed number of naive (CD62Lhi) 139TCR cellswith or without (to determine endogenous presentation) addedPLP(139–151). Splenic APCs from mice with EAE failed to activatethe proliferation of naive 139TCR cells in the absence of added
PLP(139–151) (Fig. 3a,b). It has been reported that CNS microgliaalso fail to activate naive PLP(139–151)-specific T cells31. Here,CNS mDCs induced 44-fold and 16-fold population expansion ofthe naive 139TCR cells at APC/T cell ratios of 1:5 (Fig. 3a) and 1:50(Fig. 3b), respectively. CNS CD8 DCs were the next most efficient atendogenous presentation, inducing sixfold proliferation of 139TCRcells versus splenic APCs at an APC/T cell ratio of 1:50 in theabsence of added peptide (Fig. 3b). CNS pDCs were similar to CD8DCs at an APC/T cell ratio of 1:50, inducing fivefold expansion of the139TCR cells.
Macrophages were the least efficient in inducing naive T cellproliferation with endogenously presented peptide, inducing onlytwofold proliferation above background at APC/T cell ratios of1:5 and 1:50. The addition of exogenous PLP(139–151) substantiallyincreased the macrophage-induced activation over endogenouspresentation-induced activation and, notably, increased the efficiency
CD45RB
8.4 ± 3.7
60.6 ± 9.7
Gated Thy-1.1+ CD4+
*
a
e
i
l
p
m n o
j k
f
g
h
b c d
*
Figure 2 Differential localization of DC subsets to
demyelinated areas during peak acute relapsing
EAE, and clustering of mDCs with Thy-1.1+CD4+
T cells specific for relapse-associated epitopes.
(a–k) Sections from the cerebellum (a,b,e,f,i,j) or
lumbar spinal cord (c,d,g,h,k) of mice at the peak
acute phase of relapsing EAE induced with
PLP(178–191). (a–d) Serial sections stained forPLP (green; a,c) or CD11c (red; b,d); arrowheads
indicate demyelinated areas. (e–k) Three-color
overlays of DAPI (blue), CD11c (red), and CD11b
(green; e–h) or B220 (green; i–k) showing
mDCs (yellow; e–h) and pDCs (yellow; i–k).
(f) Enlargement of boxed area in e.
(g) Enlargement of boxed area in h. *, vessel
lumen. (i) Dotted line indicates meninges.
(j) Two-color overlay of i. (k) White arrowheads
indicate double-stained pDCs; circles surround
inflammatory foci containing B220– DCs.
(l–o) Sections from lumbar spinal cords of
PLP(178–191)-primed 139TCR–SJL mixed bone
marrow chimeric mice. (l–n) Three-color overlays
of DAPI (blue), CD11c (red) and Thy-1.1 (green).
(n) Enlargement of boxed area in m. (o) Serial
section overlay of CD11c (red), Thy-1.1 (blue)
and CD11b (green); mDCs (yellow) surround
Thy-1.1+ 139TCR cells (blue). Originalmagnification, �100 (a–e,h), �200 (i,k–m) or
�400 (f,g,j,n,o). Sections are representative of
three mice analyzed individually at day 15
after immunization. (p) Activation status of
Thy-1.1+CD4+ 139TCR cells in the CNS of
mixed bone marrow chimeric mice with
PLP(178–191)-induced EAE, assessed by
CD45RB expression (day 15 after immunization).
Numbers in plot indicate percent CD45RBlo cells
(left) or CD45RB+ cells (right). Data are
representative of two separate experiments with
two to three mice analyzed.
174 VOLUME 8 NUMBER 2 FEBRUARY 2007 NATURE IMMUNOLOGY
A R T I C L E S©
2007
Nat
ure
Pub
lishi
ng G
roup
ht
tp://
ww
w.n
atur
e.co
m/n
atur
eim
mun
olog
y
of the three DC subsets to activate 139TCRcells compared with endogenous activation(Fig. 3a,b). Endogenous activation of thenaive 139TCR cells induced by the CNSDC subpopulations was MHC class IIrestricted and dependent on CD80 andCD86 (Fig. 3c,d).
Macrophages were more readily able toinduce the proliferation of an activatedPLP(139–151)-specific CD4+ T cell linewith endogenous acquired peptide thanwere pDCs, but it was far less than theresponse induced by mDCs (Fig. 3e). At thelower APC/T cell ratio (1:50), mDCs inducedconsiderable IFN-g production by thePLP(139–151)-specific T cell line, threefoldhigher than that induced with ten times thenumber of CNS macrophages (Fig. 3f). Thus,CNS mDCs most efficiently presented endo-genous myelin peptides for the activation ofboth naive and activated T cells.
CNS mDCs drive TH-17 cell differentiation
We analyzed the culture supernatants of naive139TCR cells activated with endogenouslyacquired PLP(139–151) presented by the var-ious CNS APC populations for a panel ofinflammatory cytokines and chemokines.IL-2 production correlated with populationexpansion of 139TCR cells for all APC popu-lations tested (Fig. 4a,b). Naive 139TCR cellsactivated with a T cell/splenic APC ratio of5:1 presenting added PLP(139–151) pro-duced 70 pg/ml of IL-17 and 834 pg/ml ofIFN-g (IL-17/IFN-g ratio, 0.09), but whenactivated by an equivalent number of CNSmDCs presenting endogenously acquiredpeptide, the 139TCR cell cytokine profileswitched to 207 pg/ml of IL-17 and 155 pg/ml of IFN-g (IL-17/IFN-g ratio, 1.3; Fig. 4a).At a 139TCR cell/APC ratio of 5:1, pDCsinduced 62 pg/ml of IL-17 and 104 pg/ml ofIFN-g (IL-17/IFN-g ratio, 0.60), whereasmacrophages induced low production ofIL-17 (6.5 pg/ml) and IFN-g (48 pg/ml;IL-17/IFN-g ratio, 0.13; Fig. 4a). IL-17 production in all the culturesat a 139TCR cell/APC ratio of 50:1 was less than 20 pg/ml, but againthe mDCs induced the most IL-17 of the CNS APC populations tested(Fig. 4b). CD8 DCs induced the highest concentrations of IFN-gproduction (100 pg/ml) by the 139TCR cells of the CNS APCsubpopulations tested at the 139TCR cell/APC ratio of 50:1(Fig. 4b). In cultures with CNS APCs alone, IL-17 and IFN-g werebelow the limit of detection (data not shown).
As for the inflammatory cytokines tumor necrosis factor (TNF) andIL-10, CNS mDCs presenting endogenous antigen induced 187 pg/mlof TNF and 125 pg/ml of IL-10, more than fourfold fold higher thanthe concentrations for splenic APCs presenting exogenous PLP(139–151) (Fig. 4a). Macrophages induced relatively high TNF and IL-10compared with proliferation kinetics, and pDCs and CD8 DCsinduced the lowest concentrations of TNF and IL-10 at APC/T cellratios of both 1:5 and 1:50 (Fig. 4a,b). The CNS APC populations also
induced the production of substantial CCL2, CCL3, CCL5 andCXCL10, with mDCs again being the most efficient.
To determine the cytokine profiles of CD4+ T cells in the CNS duringthe acute stages of disease that drive relapsing EAE progression throughreactivity to the spreading myelin epitope PLP(139–151), we transferrednaive Thy-1.1 139TCR cells into Thy-1.2 SJL mice at the onsetof PLP(178–191)-induced clinical disease (days 10–11) and recoveredthe CNS cells in the late acute phase (days 15–17) for analysis of IL-17and IFN-g production. Nonselective stimulation with phorbol12-myristate 13-acetate and ionomycin showed that the transgenic139TCR cells ‘polarized’ to IL-17 production, with an IL-17/IFN-gratio of 2.1, whereas the mixed population of acute-phase associ-ated host Thy-1.2+CD4+ T cells had a TH1-polarized phenotypewith an IL-17/IFN-g ratio of 0.44 (Fig. 4c,d). Thus, mDCs uniquelyinduce a TH-17 profile in naive 139TCR cells both in vitro and in theinflamed CNS.
12
∆c.p
.m. (
×104 )
∆c.p
.m. (
×103 )
IFN
-γ (
pg/m
l)
103 c
.p.m
.
103 c
.p.m
.∆c
.p.m
. (×1
03 )7
244
60
5.5 2 1.1
921 17
No peptide
10 µM PLP(139–151)
1
0
0
1
0 0
100
200
300
1,0002,0003,0004,0005,000
12345
2025303540 39
43
2.31.4
5.0
1.8
2
0
1
2
3
4
5
6
7.5
12.5
15
10
0
2.5
5
10
20
30
mDC pDC mDC
16
13
7.59.6 9.5
7
2 15 6
pDC CD8 DCMϕ
mDC pDC
Media
CTLA-4–lg
5:1
50:1
α-I-As
Mϕ
mDC pDC Mϕ mDC SplpDC Mϕ
Mϕ
mDC pDC CD8 DC Mϕ
SplSpl
a b
c d
e f
Figure 3 Differential ability of CNS DC subpopulations to present endogenous and exogenous
myelin peptide for the activation of naive and activated CD4+ T cells specific for spreading epitopes.
(a,b) Proliferation of naive CD62Lhi PLP(139–151)-specific CD4+ transgenic T cells (139TCR cells)
cultured together with CNS mDCs, pDCs, CD8 DCs, macrophages (Mj) or irradiated CD4– splenocytes
(Spl) isolated from PLP(178–191)-primed SJL mice at the peak of acute EAE, at an 139TCR T cell/
APC ratio of 5:1 (a) or 50:1 (b), in the presence or absence of 10 mM PLP(139–151). Results are
expressed as the change in counts per minute (Dc.p.m.); above bars, stimulation indices, based on
background proliferation with irradiated T cells and splenic cells at a ratio of 5:1, or T cells culturedwith 10 mM PLP(139–151) in the absence of APCs. (c,d) Proliferation analysis as described in a,b,
except that control immunoglobulin (Media), anti–MHC class II (a-I-As) or CTLA-4–immunoglobulin
(CTLA-4–Ig) was added to cultures with an 139TCR T cell/APC ratio of 5:1 (c) or 50:1 (d) without
added PLP(139–151) (results are presented as c.p.m.). Horizontal lines, c.p.m. of 139TCR cells
cultured with irradiated CD4– splenocytes. (e) Endogenous presentation to a PLP(139–151)-specific
CD4+ T helper line at a T cell/APC ratio of 5:1 or 50:1. Results are expressed as the change in c.p.m.;
above bars, stimulation indices. (f) IFN-g in the T helper cell supernatants from e after 92 h. All results
are representative of five similar experiments in which cells from 20 perfused CNS APC donor mice
were pooled and each APC subset was analyzed at least three times.
NATURE IMMUNOLOGY VOLUME 8 NUMBER 2 FEBRUARY 2007 175
A R T I C L E S©
2007
Nat
ure
Pub
lishi
ng G
roup
ht
tp://
ww
w.n
atur
e.co
m/n
atur
eim
mun
olog
y
CNS DCs express costimulatory molecules and myelin proteins
We next determined if the differences in the activation of naive139TCR cells by mDCs, pDCs, CD8 DCs and macrophages couldbe attributed to differential acquisition of myelin proteins and/orexpression of molecules required for antigen presentation and T cell
activation. Staining of flow cytometry–purified, permeabilized CNSDC subsets and macrophages for PLP showed all had acquired PLP toa similar degree (Supplementary Fig. 2 online), demonstrating thatthe acquisition of myelin peptide was not a chief factor for T cellstimulatory differences. Flow cytometry for CD4+ T cell activation
a
b
c d
1,000 IL-2
IL-2
(pg
/ml)
CC
L3 (
pg/m
l)C
CL3
(pg
/ml) C
CL2 (pg/m
l) CC
L5 (
pg/m
l)
CX
CL10 (pg/m
l)
IL-2
(pg
/ml)
IFN
-γ (
pg/m
l) IL-17 (pg/ml)
IL-10 (pg/ml)
CC
L2 (pg/ml) C
CL5
(pg
/ml)
CX
CL10 (pg/m
l)
CCL3
CCL2
IL-2
∆c.p.m.IFN-γ
CCL3CCL2
CCL5CXCL10
IL-17TNF
IL-10
CCL5
CXCL10
∆c.p.m
. (×104)
∆c.p.m
. (×103)
IFN
-γ (
pg/m
l) IL-17 (pg/ml)
IL-10 (pg/ml)T
NF
(pg
/ml)
TN
F (
pg/m
l)
∆c.p.m.IFN-γIL-17
TNFIL-10
750
500300
200
100
0
1001,000
2,000
0
100
200
300
0
1,000
2,000
0
100
200
0
5
10
15
20
25
30
35
0 0
10
20
30
10
20
00
0
10
20
2.5
5
10
15
20
50
200400500600700800
0
100
100
150
200
250
0
50
mDC
pDC M
ϕSpl
Spl +
PLPm
DCpD
C Mϕ
Spl
Spl +
PLP
mDC
pDC M
ϕSpl
Spl +
PLPm
DCpD
C Mϕ
Spl
Spl +
PLPm
DCpD
C Mϕ
Spl
Spl +
PLP
Thy-1.1
IL-17
IFN
-γ
Thy-1.2104
4.18% 16.65%
65.92%
4.00%
13.43%
1.09%
86.05% 8.67%
103
102
101
100
100 101 102 103 104 100 101 102 103 104
Thy-1.1+ 139TCR cells
IL-1
7/IF
N-γ
0
1
2
3
Thy-1.2+ T cells
P = 0.03
mDC
0 0
25
50
75
10
20
30
40
pDC M
ϕSpl
CD8 DC
Spl +
PLPm
DCpD
C Mϕ
Spl
CD8 DC
Spl +
PLP
mDC
pDC M
ϕSpl
CD8 DC
Spl +
PLPm
DCpD
C Mϕ
Spl
CD8 DC
Spl +
PLPm
DCpD
C Mϕ
Spl
CD8 DC
Spl +
PLP
050100150200250300350400450
0
1
2
3900800700600500
200
100
0
200
100
0 0
10
20
30
75100
125150
0
50
100
150
200
250
Figure 4 CNS mDCs induce a TH-17 inflammatory cytokine profile in naive CD4+ PLP(139–151)-specific T cells both in vitro and in the CNS.
(a,b) LiquiChip assay of secreted cytokines and chemokines in culture supernatants collected at 72 h after culture initiation from various CNS APC
populations presenting endogenously acquired PLP(139–151) to naive 139TCR cells, with a 139TCR cell/APC ratio of 5:1 (a) or 50:1 (b). Spl+PLP,irradiated CD4– splenocytes plus PLP(139–151). (c,d) Production of IL-17 and IFN-g by CD62LhiThy-1.1+ 139TCR cells transferred intravenously into
Thy-1.2+ SJL mice at the onset of clinical disease (days 10–11) and recovered in the late acute phase (days 15–17); CNS cells were stimulated for 3 h
with phorbol 12-myristate 13-acetate and ionomycin in the presence of GolgiStop. (c) Flow cytometry of live gated Thy-1.1 and Thy-1.2 CD4+ T cells
recovered from 16 mice. Approximately 5,000 events are in each plot; quadrants are set relevant to ‘fluorescence minus one’ controls. (d) Data from c
plotted as a scatter diagram of the ratio of IL-17+ cells to IFN-g+ cells for Thy-1.1+ 139TCR cells and Thy-1.2+ host cells in four separate experiments
with CNS cells isolated from 3–16 mice and pooled.
176 VOLUME 8 NUMBER 2 FEBRUARY 2007 NATURE IMMUNOLOGY
A R T I C L E S©
2007
Nat
ure
Pub
lishi
ng G
roup
ht
tp://
ww
w.n
atur
e.co
m/n
atur
eim
mun
olog
y
molecules and for B7-H1 and B7-DC, which may have stimulatory orinhibitory functions for T cell activation36, demonstrated that CNSmDCs, CD8 DCs and macrophages had similar expression of MHCclass II, CD80, CD86 and CD40, as well as B7-DC and B7-H1(Supplementary Fig. 3 online). The pDCs, however, were immature,with absent or low expression of MHC class II and low expression ofCD80, CD86 and CD40. Analysis of the cervical lymph nodes,peripheral lymph nodes and spleens of mice with relapsing EAEsuggested that the inflamed CNS environment determines the imma-ture phenotype of pDCs, as peripheral pDCs had expression of MHCclass II, CD80 and CD86 equivalent to that of mDCs at those sites(data not shown). Thus, the enhanced ability of CNS mDCs relative tothat of macrophages to activate naive and activated T cells was not dueto differential expression of costimulatory molecules or the ability tointernalize myelin debris.
CNS mDCs express TH-17-polarizing cytokines
TGF-b1 and IL-6 are critical cytokines for polarizing TH-17 cells, andIL-23 promotes the survival and population expansion of TH-17cells17,19. Because the mDCs induced naive 139TCR cells to producehigh concentrations of IL-17 relative to IFN-g, we examined the abilityof those cells to produce TGF-b1, IL-6 and IL-23 relative to that of theother CNS DC populations. We assessed TGF-b1 mRNA expression, asCNS DCs perished ex vivo in the absence of serum (unpublishedobservations) and high concentrations of latent TGF-b1 in serum can‘mask’ the measurement of TGF-b1. CNS mDCs expressed the mostTGF-b1 transcripts ex vivo: 3-fold and 14-fold more than CNS macro-phages and pDCs, respectively (Fig. 5a). Ex vivo, CNS macrophagesspontaneously produced more IL-6 than mDCs (Fig. 5b); however,crosslinking CD40 on the CNS populations induced more IL-6from mDCs: 1.7-fold higher than macrophages, 1.8-fold higher thanCD8 DCs and 6.4-fold higher than pDC (Fig 5b). The mDCs, CD8DCs, macrophages and pDCs had equivalent production of IL-23 afterCD40 crosslinking (Fig. 5c). The mRNA for the IL-23 subunits p19 andp40 could be detected in the DC subpopulations and macrophagesex vivo and after CD40 activation (data not shown). Thus, the uniqueability of CNS mDCs to drive IL-17 production is related to their abilityto produce large amounts of cytokines required for the differentiation(IL-6 and TGF-b) and survival (IL-23) of TH-17 cells.
DISCUSSION
Here we have shown that inflamed CNS is not unique and, like othertissues, attracted DCs of bone marrow origin, which acquired myelindebris to present to both naive and activated CD4+ T cells. For
protection from pathogens, that is an advan-tage for the host, but in an inflammatoryCNS demyelinating disease, self proteins canbe picked up and presented to autoreactiveT cells, resulting in the progression of CNSautoimmunity through epitope spread-ing31,37. We tested the relative efficiency ofinflammatory CNS mDCs, pDCs, CD8 DCsand macrophages in activating PLP(139–151)-specific naive CD4+ 139TCR cells withendogenously acquired antigen and foundthat mDCs were highly potent CNS APCsdriving 139TCR cell population expansionand production of pathogenic cytokines(IL-17 and IFN-g). Although CNS mDCs,CD8 DCs and macrophages had similarexpression profiles for accessory molecules,
mDCs were distinct in their ability to activate both the prolifer-ation and differentiation of naive CD4+ T cells. Notably, mDCsclustered with pathogenic CD4+ T cells in perivascular inflammatoryfoci in the CNS, and the CD4+ T cells were activated and producedIL-17. Those findings, along with published findings that epitopespreading occurs directly in the CNS31, suggest that CNS mDCs arethe main ‘drivers’ of epitope spreading. There are no methods forspecific depletion of mDCs at present; future experiments should testthose findings in vivo.
IL-17 is crucial for chronic inflammation in many disease models,including prototypic TH1 diseases such as EAE and colitis11–14 andairway hypersensitivity that is TH2 polarized11, as well as for hostprotection against bacterial pathogens18,38 and allograft rejection39.Stimulated naive CD4+ T cells can be polarized to either IFN-g-producing TH1 cells or IL-17-producing TH-17 cells, and thosecytokines are not produced together17,19. We have shown that TH1and TH-17 cells accumulated in the CNS at the peak of the acute phaseof relapsing EAE in SJL mice. IL-17-polarized CNS T cells wereactivated by the relapse-associated spreading epitope PLP(139–151)in PLP(178–191)-induced relapsing EAE, whereas T cells participatingin the acute-phase response were TH1 biased after mitogen reactiva-tion. Those data are notable, given the pathological function of IL-17in EAE pathogenesis. Neutralizing IL-17 or inhibiting T cell TGF-bresponsiveness and TH-17 induction after EAE induction amelioratesrelapse incidence and severity11,13,20, whereas increasing IL-17-produ-cing CD4+ CNS T cells by enhanced TGF-b production in vivoincreases EAE severity and morbidity17. The findings that relapse-associated 139TCR cells clustered with mDCs in the CNS and thatthose T cells had a pathogenic TH-17 profile indicate that mDCs areprobably central to the pathogenic progression of EAE.
PLP(139–151)-pulsed, unpurified splenic APCs induced naive139TCR cells to differentiate into TH1 cells, as demonstrated by lowIL-17/IFN-g ratios, low concentrations of IL-10 and undetectableproduction of IL-4 and IL-5. However, endogenous presentation ofPLP(139–151) by mDCs isolated from inflamed CNS induced moreIL-17 than IFN-g from naive 139TCR cells. We found that mDCs werethe only CNS APC population tested that ‘biased’ the 139TCR cells to aTH-17 profile. Discrete DC subsets have been described as having anintrinsic ability to polarize T helper cells40. The ability of CD8 DCs to‘preferentially’ induce TH1 differentiation correlates with high IL-12production41. Our data suggest that mDCs specifically induce thegeneration of TH-17 cells ex vivo and in vivo. Moreover, in the contextof CNS autoimmune inflammation, mDCs were unique in their abilityto ‘drive’ TH-17 cells with endogenously collected antigen. High
a b c2 3,5003,0002,5002,0001,5001,000
500100
350
300
250
200
150
100
50
0
50
0
1
0mDC mDCpDC pDC
ND
Tgfb1
(×1
04 )
IL-6
(pg
/ml)
IL-2
3 (p
g/m
l)
CD8 DC CD8 DC
Media
CD40L
Mϕ Mϕ mDC pDC CD8 DC Mϕ
Figure 5 CNS mDCs express TGF-b1 and produce IL-23 and high concentrations of IL-6 after CD40
activation. (a) Real-time PCR analysis of TGF-b1 expression by isolated CNS populations. ND, not
detectable. (b) IL-6 production in 4-day cultures of 1 � 104 purified CNS populations (Media) or
2 � 105 purified CNS populations incubated for 2 d with soluble CD40L (CD40L). (c) IL-23 production
in 2-day cultures of 2 � 105 purified CNS populations incubated with soluble CD40L. Results are
representative of two experiments in which cells from 20 mice were pooled (open bars); error bars,
s.e.m. of three experiments in which cells from 20 mice were pooled (filled bars).
NATURE IMMUNOLOGY VOLUME 8 NUMBER 2 FEBRUARY 2007 177
A R T I C L E S©
2007
Nat
ure
Pub
lishi
ng G
roup
ht
tp://
ww
w.n
atur
e.co
m/n
atur
eim
mun
olog
y
TGF-b1 expression and IL-6 production by mDCs may be an intrinsicproperty of that DC subset or the response of that subset to the CNSenvironment. Comparison of the TH-17-polarizing abilities of DCsubsets from different anatomical sites should elucidate that issue.
The survival of TH-17 cells is dependent on IL-23, and after CD40triggering, mDCs, pDCs, CD8 DCs and macrophages isolated frominflamed CNS secreted IL-23. Through its effects on TH-17 cells, IL-23has been noted for its ability to ‘drive’ relapsing EAE, as neutralizingIL-23 with a specific antibody during the acute phase of EAE and alsoduring remission suppresses clinical relapses12. Ubiquitous secretionof IL-23 by infiltrating CNS DCs and macrophages may be regulatedby crosstalk with T cells, through CD40-CD154 interactions or by anas-yet-unidentified receptor-ligand complex, as has been described forthe IL-18-independent engagement of IL-18 receptor-a that isrequired for the production of IL-12–IL-23 p40 by APCs42.
An associative function of DCs in chronic inflammatory conditionshas been noted for multiple sclerosis25,43,44, psoriasis45, arthritis46 andlupus47. In the CNS of SJL mice at the peak acute phase of relapsingEAE, the organization of CNS mDCs in the center of inflammatoryfoci, interacting with CD4+ T cells with macrophages marginalizedaround the inflammatory cells, is reminiscent of secondary lymphoidorgan organization. New formation of lymphoid tissue structures ininflamed tissues has been demonstrated in the pancreas in thenonobese diabetic model of type 1 diabetes48,49 and in the CNSduring late-stage relapsing EAE50. Chemokine production in theneolymphoid organ is crucial for attracting infiltrating cells49. Wehave demonstrated here the production of CCL3, CCL5 and CCL2 incocultures of 139TCR cells and CNS DCs. Those chemokines arenoted for their ability to attract immature DCs and CD4+ T cells andmay promote DC–T cell clusters in the CNS. The DC populationsseemed to have differential homing and/or accumulation in the CNS.The pDCs accumulated in the cerebellar meningeal area, whereasmDCs and macrophages accumulated throughout the spinal cord andcerebellum. That may have been due to intrinsic differences in homingabilities by the DC subsets, defined by chemokine responsiveness, and/or mechanisms of tissue infiltration. For example, pDCs do not entersystemic tissues during inflammatory responses but migrate intoparticipating lymph nodes through the high endothelial venules51
and likewise may enter the autoimmune CNS at sites containinghigh endothelial venules49. Immature and circulatory mDCs respondto a variety of chemokines expressed in the CNS during clinical EAE,including CCL3, CCL5 and CCL2 (refs. 51,52). That ubiquitousinflammatory homing capacity may contribute to the perivasculardistribution of mDCs throughout the spinal cord and cerebellum atthe peak acute phase of EAE.
CNS CD11b–F4/80– DCs (pooled pDCs and CD8 DCs) are moreefficient at presenting endogenous peptide and activating naive139TCR cell proliferation than are pooled CD11b+CD11c+ DCs andmacrophages31. Purifying the discrete CNS APC subsets, however,showed the mDCs were vastly superior in the activation of naive CD4+
T cells with endogenous peptide. Notably, macrophages in theinflamed CNS actively suppressed endogenous presentation andmDC-driven naive 139TCR cell proliferation in vitro by about 70%(unpublished observations), accounting for our findings when thosecells were pooled as APCs. Preliminary data suggest macrophage-expressed B7-H1 is not important in that suppression. However, wedo not believe that macrophage suppression of mDC-mediated naiveT cell activation is dominant in the CNS at the peak acute phase ofEAE. First, naive 139TCR cells transferred into PLP(178–191)-primedmice at disease onset undergo five rounds of proliferation in 3 d andall have proliferated by 6 d after transfer, indicating naive T cell
proliferation is highly active in the CNS31. Second, at the peak acutephase of PLP(178–191)-induced relapsing EAE, 139TCR cells asso-ciated mainly with mDCs, not CD11b+CD11c– macrophages, andbecame activated. Third, naive CD4+ T cells cultured together withCNS macrophages were fully responsive to TCR reactivation (unpub-lished observations) and therefore were still potentially pathogenic. Inaddition, our data suggest that during an acute inflammatory responsein the CNS, infiltrating macrophages would support the limitedpopulation expansion and cytokine production of previously activatedCD4+ T cells. Thus, in the inflamed CNS, macrophages may induceeffector functions of myelin-specific CD4+ T cells that have beenactivated by mDCs.
There are many correlations between DCs in EAE and multiplesclerosis. It has been shown that mDCs positive for the marker C-typelectin receptor DC-SIGN accumulate in multiple sclerosis perivascularcuffs, associating with T cells25,44, as did CD11b+ mDCs in EAE. Incerebral spinal fluid from patients with multiple sclerosis, mDCs aremature, with high expression of accessory molecules53, and circulatorymDCs produce higher concentrations of IL-6 than do those fromhealthy controls54. The mDCs that infiltrated the CNS from thecirculation in mice with EAE had high expression of accessorymolecules and produced IL-6. The observations that treatment ofpatients with multiple sclerosis with IFN-b and glatiramer acetatesuppresses DC maturation emphasizes the idea that DC-targetedimmunotherapy could be beneficial for the treatment of autoimmunediseases of the CNS55,56. However, DCs are not a homogenouspopulation and, as shown here, different populations have discretefunctions in the adaptive immune response in the CNS.
CNS pDCs in EAE were immature, with low expression of accessorymolecules, as are circulatory pDCs in patients with multiple sclerosis,which have defects in maturation compared with those from healthycontrols57,58. Although pDCs from the CNS of EAE mice inducednaive T cell proliferation and cytokine and chemokine productionwith endogenous peptide, it was minimal compared with that of CNSmDCs. The reported attributes of circulatory pDCs in mediatingtolerance to allografts59 warrants further investigation into a potentialregulatory function of that DC population in EAE and multiplesclerosis and may demand DC population–specific immunotargetingof CNS populations.
In conclusion, our data indicate that mDCs attracted to theinflamed CNS from the bloodstream acquire myelin antigens andpresent them to naive autoreactive CD4+ T cells to effectively activatea pathogenic T cell population that drives epitope spreading andthus clinical relapses in EAE. Therefore, mDCs and their TH-17-inducing mediators are prime targets for immunotherapy to haltthe progression of EAE or multiple sclerosis. The involvement ofmDCs and pDCs in the pathogenesis of relapsing EAE makes thismodel highly relevant for studying the driving and regulatoryimmune forces underlying multiple sclerosis and other chronicautoimmune diseases.
METHODSMice. Female SJL/J mice were purchased from Harlan Sprague Dawley or
Taconic. Female B6 mice were obtained from Jackson. (B6 � SJL) F1 hybrids
were purchased from Taconic. SJL Thy-1.1+ 5B6 PLP(139–151)-specific TCR-
transgenic mice (139TCR mice) have been described31,60. All mice were housed
and cared for according to approved protocols of the Northwestern University
Animal Care and Use Committee.
Antibodies. Directly conjugated antibodies specific for mouse antigens CD3x(145-2C11), CD4 (RM4-5), CD8a (53-67), CD11b (M1/70), CD11c (N418),
CD40, (HM40-3), B220 (RA3-6B), CD45 (30-F11), CD45.1 (A20), CD45.2
178 VOLUME 8 NUMBER 2 FEBRUARY 2007 NATURE IMMUNOLOGY
A R T I C L E S©
2007
Nat
ure
Pub
lishi
ng G
roup
ht
tp://
ww
w.n
atur
e.co
m/n
atur
eim
mun
olog
y
(104), CD62L (MEL-14), CD80 (16-10A), CD86 (GL1), Thy-1.1 (HIS51),
B7-H1 (M1H5), B7-DC (TY25), Thy-1.2 (53-2.1) and mouse IFN-g (554412)
and IL-17 (18H10.1) were purchased from BD Pharmingen or eBioscience.
Purified I-As/q (KH118) and CTLA-4–immunoglobulin (552133) were pur-
chased from BD. Antibody to CD205 (anti-CD205; DEC-205; MCA949) and
anti-PLP (MCA839G) were purchased from Serotec, and anti-F4/80 (RM2901)
was purchased from Caltag. Biotinylated anti-I-As (clone MKS4) was detected
with phycoerythrin-conjugated anti-biotin (120-001-086; Miltenyi).
Peptides. PLP(139–151) (HSLGKWLGHPDKF), PLP(178–191) (NTWTTCQ
SIAFPSK), ovalbumin amino acids 323–339 (OVA(323–339); ISQAVHAA
HAEINEAGR) and MOG(35–55) (MEVGWYRSPFSRVVHLYRNGK) were
synthesized by Genemed Synthesis.
Induction of EAE. SJL mice were immunized subcutaneously with 100 ml of
emulsified incomplete Freund’s adjuvant (Difco) supplemented with 200 mg
Mycobacterium tuberculosis H37Ra (Difco) and 100 mg of either PLP(178–191)
or 50 mg PLP(139–151). Female (B6 � SJL) F1 mice 12–15 weeks of age were
immunized subcutaneously with 100 mg MOG(35–55) in complete Freund’s
adjuvant and received intraperitoneal injections of 200 ng pertussis toxin
(Sigma) at the time of immunization and 48 h later. Chronic-progressive
EAE (‘chronic EAE’) was induced in female B6 mice by subcutaneous
immunization with complete Freund’s adjuvant containing 200 mg MOG(35–
55). Mice received intraperitoneal injections of 200 ng pertussis toxin (Sigma)
at the time of immunization and 48 h later.
Isolation of spleen or lymph node cells. Spleens and lymph nodes were
digested for 45 min at 37 1C in 0.2 units Liberase R1 (Roche), 50 mg/ml of DNase I
(Sigma) and 25 mM HEPES. Cell suspensions were filtered through nylon with
pores of 100 mm. For spleens, red blood cells were lysed with Tris-NH4Cl.
Enrichment of DCs from peripheral lymphoid organs. Spleen and lymph
node cell suspensions were resuspended at a density of 1 � 108 cells per ml or
less in 0.8% NaCl, 10 mM HEPES, pH 7.4 (HEPES-buffered saline), 0.5% BSA
and 1 mM EDTA, mixed with a 40% volume of Optiprep (Nycomed) and
overlaid with an equal volume of 25% Optiprep and twice the volume of 20%
Optiprep (made in the same buffer). Cells were fractionated at 600g for 30 min
at 24 1C. The lower-density layer contained 25–40% CD11c+ DCs.
Isolation of CNS leukocytes. CNS cells were isolated from perfused mice as
described31. Alternatively, CNS cells were fractionated from myelin in 40%
(volume/volume) isotonic Percoll centrifuged for 10 min at 350g.
Flow cytometry and gating. Cells for analysis were blocked for at least 10 min
on ice with 10% (volume/volume) rat serum, 10% (volume/volume) hamster
serum and anti-CD16/32 before being stained with five- to six-color antibody
‘cocktails’. Data were acquired on an LSR II cytometer (BD) and were analyzed
with FACSDiva software (BD) or FlowJo software (Treestar). The change in
mean fluorescence intensity (DMFI) for any antigen on a gated population was
determined by subtraction the MFI of the isotype control from the MFI of the
stained population.
CNS antigen-presentation assay. Antibody-stained CNS leukocytes resus-
pended in R10 media (RPMI medium with 10% (volume/volume) fetal calf
serum, 2 mM L-glutamine, 100 U/ml of penicillin, 100 mg/ml of streptomycin,
0.1 mM nonessential amino acids, 50 mM 2-mercaptoethanol, 1 mM amino-
guanidine and 20 mM indomethacin) were sorted on a MoFlo (DakoCytoma-
tion). Cell populations were gated as described in Supplementary Figure 1.
Sorted populations were more than 95% pure. APCs were cultured for 96 h at
various ratios together with 1 � 105 CD4+ T cells from a PLP(139–151)-specific
line or CD62LhiCD4+Thy-1.1+ cells from 5B6 transgenic mice (139TCR).
Incorporation of [3H]thymidine (1 mCi/well; ICN) at 16 h was analyzed
with a TopCount-NXT (Packard). Cytokines in the supernatants after 72 h
were assessed by cytokine bead array for measurement of IL-2, IL-4, IL-6,
IL-10, IL-17, IFN-g, TNF, granulocyte-macrophage colony-stimulating
factor, CCL2, CCL3, CCL5, CXCL10 according to the manufacturer’s
instruction (Upstate).
Culture and CD40 activation of CNS APCs. Purified populations of CNS
DCs and macrophages were cultured in R10 media plus 3 ng/ml of
granulocyte-macrophage colony-stimulating factor (R&D). Soluble CD40L
(1 mg/ml; R&D) was added to some samples. After 48 h, IL-23 in the
supernatant was assessed by enzyme-linked immunosorbent assay according
to the manufacturer’s instruction (eBioscience) and IL-6 protein was measured
by cytokine bead array (Upstate).
Nucleic acid preparation and real-time PCR. Total RNA was purified from
sorted cells with the SV total RNA kit (Promega) according to the manufac-
turer’s instructions. The RNA from pooled CNS cells of 20 mice was
concentrated and cDNA synthesis and real-time PCR was done in triplicate
as described61. The results for b-actin were normalized to those of an internal
control (neonatal microglia activated for 48 h with 500 U IFN-g and TNF) and
then transcripts in each sample were normalized to b-actin.
Immunohistochemistry. Frozen cerebellar and lumbar spinal cord sections
from PBS-perfused mice were cut 6 mm in thickness and were analyzed
by immunohistochemistry as described22. Staining was analyzed with a
Leica DM5000B fluorescent microscope and Advanced SPOT software
(Diagnostic Instruments).
Bone marrow chimeras. Mice were irradiated with two doses of 350 rads, 4 h
apart, and were reconstituted intravenously with 2 � 107 to 3 � 107 bone
marrow cells from donor mice. After 6–8 weeks, over 95% of peripheral blood
mononuclear cells and over 99% of CD11b+ peripheral blood mononuclear
cells were of donor origin. Mixed bone marrow chimeras in which SJL mice
were reconstituted with a 1:1 mix of Thy-1.2+ SJL and Thy-1.1+ 139TCR bone
marrow have been described31.
Transgenic T cell transfer and cytokine staining. At the onset of PLP(178–
191)-induced relapsing EAE (days 10–11), mice received intravenous injection
of 1 � 107 naive CD4+ CD62L+ lymph node cells isolated from 139TCR mice
with negative and positive magnetic selection according to the manufacturer’s
instructions (Miltenyi Biotec). In late acute disease (days 15–17), mice were
perfused and CNS mononuclear cells were isolated as described above. Cells
were activated for 3 h with 5 ng/ml of phorbol 12-myristate 13-acetate and
500 ng/ml of ionomycin in the presence of GolgiStop (Sigma) before being
stained for extracellular ligands and intracellular cytokines. When mitogen-
stimulated cells were stained with LIVE/DEAD Fixable dye (Molecular Probes;
Invitrogen), the ratio of IL-17-producing cells to IFN-g-producing cells did
not change in the live, gated Thy-1.1 and Thy-1.2 populations. In some
experiments, CNS cells were incubated with 1 mg/ml of anti-CD3 for 16 h
before activation, with no differences in the IL-17/IFN-g ratio detected in the
T cell populations.
Statistical analysis. Differences between groups were determined with an
unpaired Mann-Whitney test.
Note: Supplementary information is available on the Nature Immunology website.
ACKNOWLEDGMENTSWe thank A. Dzionek (Miltenyi Biotec) for providing anti-PDCA-1; J. Marvin(Northwestern University) for cell sorting; M. Degutes for technical assistance;and colleagues at the Myelin Repair Foundation for discussions. Supported bythe National Institutes of Health (NS-30871 and NS-26543; and AI-07476 toE.J.M.), the National Multiple Sclerosis Society (RG 3793-A-7; and FG 1563A-1 to S.L.B.), the Myelin Repair Foundation, and DeutscheForschungsgemeinschaft (B.S.).
AUTHOR CONTRIBUTIONSS.L.B. did most of the experiments; B.S. did the intracellular cytokine stainingof 139TCR cells recovered from the CNS and helped with purification andphenotypic analysis of the CNS APC subsets; E.J.M. assisted with theimmunohistochemistry; and S.D.M. provided intellectual input, secured thefunding and guided the preparation of the manuscript.
COMPETING INTERESTS STATEMENTThe authors declare that they have no competing financial interests.
NATURE IMMUNOLOGY VOLUME 8 NUMBER 2 FEBRUARY 2007 179
A R T I C L E S©
2007
Nat
ure
Pub
lishi
ng G
roup
ht
tp://
ww
w.n
atur
e.co
m/n
atur
eim
mun
olog
y
Published online at http://www.nature.com/natureimmunology/
Reprints and permissions information is available online at http://npg.nature.com/
reprintsandpermissions
1. Vanderlugt, C.L. & Miller, S.D. Epitope spreading in immune-mediated diseases:implications for immunotherapy. Nat. Rev. Immunol. 2, 85–95 (2002).
2. Tuohy, V.K., Yu, M., Yin, L., Kawczak, J.A. & Kinkel, R.P. Spontaneous regression ofprimary autoreactivity during chronic progression of experimental autoimmune ence-phalomyelitis and multiple sclerosis. J. Exp. Med. 189, 1033–1042 (1999).
3. Lehmann, P.V., Forsthuber, T., Miller, A. & Sercarz, E.E. Spreading of T-cell auto-immunity to cryptic determinants of an autoantigen. Nature 358, 155–157 (1992).
4. McRae, B.L., Vanderlugt, C.L., Dal Canto, M.C. & Miller, S.D. Functional evidence forepitope spreading in the relapsing pathology of experimental autoimmune encephalo-myelitis. J. Exp. Med. 182, 75–85 (1995).
5. Vanderlugt, C.L. et al. Pathologic role and temporal appearance of newly emergingautoepitopes in relapsing experimental autoimmune encephalomyelitis. J. Immunol.164, 670–678 (2000).
6. Yu, M., Johnson, J.M. & Tuohy, V.K. A predictable sequential determinant spreadingcascade invariably accompanies progression of experimental autoimmune encephalo-myelitis: A basis for peptide-specific therapy after onset of clinical disease. J. Exp.Med. 183, 1777–1788 (1996).
7. Miller, S.D. et al. Persistent infection with Theiler’s virus leads to CNS autoimmunityvia epitope spreading. Nat. Med. 3, 1133–1136 (1997).
8. Katz-Levy, Y. et al. Temporal development of autoreactive Th1 responses and endogen-ous antigen presentation of self myelin epitopes by CNS-resident APCs in Theiler’svirus-infected mice. J. Immunol. 165, 5304–5314 (2000).
9. Neville, K.L., Padilla, J. & Miller, S.D. Myelin-specific tolerance attenuates theprogression of a virus-induced demyelinating disease: Implications for the treatmentof MS. J. Neuroimmunol. 123, 18–29 (2002).
10. Hohlfeld, R. & Wekerle, H. Autoimmune concepts of multiple sclerosis as a basis forselective immunotherapy: from pipe dreams to (therapeutic) pipelines. Proc. Natl.Acad. Sci. USA 101, 14599–14606 (2004).
11. Park, H. et al. A distinct lineage of CD4 T cells regulates tissue inflammation byproducing interleukin 17. Nat. Immunol. 6, 1133–1141 (2005).
12. Chen, Y. et al. Anti-IL-23 therapy inhibits multiple inflammatory pathways andameliorates autoimmune encephalomyelitis. J. Clin. Invest. 116, 1317–1326 (2006).
13. Langrish, C.L. et al. IL-23 drives a pathogenic T cell population that inducesautoimmune inflammation. J. Exp. Med. 201, 233–240 (2005).
14. Yen, D. et al. IL-23 is essential for T cell-mediated colitis and promotes inflammationvia IL-17 and IL-6. J. Clin. Invest. 116, 1310–1316 (2006).
15. Nakae, S. et al. Antigen-specific Tcell sensitization is impaired in IL-17-deficient mice,causing suppression of allergic cellular and humoral responses. Immunity 17,375–387 (2002).
16. Nakae, S., Nambu, A., Sudo, K. & Iwakura, Y. Suppression of immune induction ofcollagen-induced arthritis in IL-17-deficient mice. J. Immunol. 171, 6173–6177(2003).
17. Bettelli, E. et al. Reciprocal developmental pathways for the generation of pathogeniceffector TH17 and regulatory T cells. Nature 441, 235–238 (2006).
18. Mangan, P.R. et al. Transforming growth factor-b induces development of the TH17lineage. Nature 441, 231–234 (2006).
19. Veldhoen, M., Hocking, R.J., Atkins, C.J., Locksley, R.M. & Stockinger, B. TGFb in thecontext of an inflammatory cytokine milieu supports de novo differentiation of IL-17-producing T cells. Immunity 24, 179–189 (2006).
20. Veldhoen, M., Hocking, R.J., Flavell, R.A. & Stockinger, B. Signals mediated bytransforming growth factor-b initiate autoimmune encephalomyelitis, but chronicinflammation is needed to sustain disease. Nat. Immunol. 7, 1151–1156 (2006).
21. Bailey, S.L., Carpentier, P.A., McMahon, E.J., Begolka, W.S. & Miller, S.D. Innate andadaptive immune responses of the central nervous system. Crit. Rev. Immunol. 26,149–188 (2006).
22. Tompkins, S.M. et al. De novo central nervous system processing of myelin antigen isrequired for the initiation of experimental autoimmune encephalomyelitis. J. Immunol.168, 4173–4183 (2002).
23. Kawakami, N. et al. The activation status of neuroantigen-specific T cells in the targetorgan determines the clinical outcome of autoimmune encephalomyelitis. J. Exp. Med.199, 185–197 (2004).
24. Archambault, A.S., Sim, J., Gimenez, M.A. & Russell, J.H. Defining antigen-dependentstages of T cell migration from the blood to the central nervous system parenchyma.Eur. J. Immunol. 35, 1076–1085 (2005).
25. Greter, M. et al. Dendritic cells permit immune invasion of the CNS in an animal modelof multiple sclerosis. Nat. Med. 11, 328–334 (2005).
26. Matyszak, M.K. & Perry, V.H. The potential role of dendritic cells in immune-mediatedinflammatory diseases in the central nervous system. Neuroscience 74, 599–608(1996).
27. Serot, J.M., Foliguet, B., Bene, M.C. & Faure, G.C. Ultrastructural and immunohisto-logical evidence for dendritic-like cells within human choroid plexus epithelium.Neuroreport 8, 1995–1998 (1997).
28. Hanly, A. & Petito, C.K. HLA-DR-positive dendritic cells of the normal human choroidplexus: a potential reservoir of HIV in the central nervous system. Hum. Pathol. 29,88–93 (1998).
29. McMenamin, P.G. Distribution and phenotype of dendritic cells and resident tissuemacrophages in the dura mater, leptomeninges, and choroid plexus of the rat brainas demonstrated in wholemount preparations. J. Comp. Neurol. 405, 553–562(1999).
30. Suter, T. et al. The brain as an immune privileged site: dendritic cells of the centralnervous system inhibit T cell activation. Eur. J. Immunol. 33, 2998–3006 (2003).
31. McMahon, E.J., Bailey, S.L., Castenada, C.V., Waldner, H. & Miller, S.D. Epitopespreading initiates in the CNS in two mouse models of multiple sclerosis. Nat. Med.11, 335–339 (2005).
32. Newman, T.A., Galea, I., van Rooijen, N. & Perry, V.H. Blood-derived dendritic cells inan acute brain injury. J. Neuroimmunol. 166, 167–172 (2005).
33. Fischer, H.G. & Bielinsky, A.K. Antigen presentation function of brain-derived dendri-form cells depends on astrocyte help. Int. Immunol. 11, 1265–1274 (1999).
34. Santambrogio, L. et al. Developmental plasticity of CNS microglia. Proc. Natl. Acad.Sci. USA 98, 6295–6300 (2001).
35. 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).
36. Greenwald, R.J., Freeman, G.J. & Sharpe, A.H. The B7 family revisited. Annu. Rev.Immunol. 23, 515–548 (2005).
37. McMahon, E.J., Bailey, S.L. & Miller, S.D. CNS dendritic cells: critical participants inCNS inflammation. Neurochem. Int. 49, 195–203 (2006).
38. Ye, P. et al. Requirement of interleukin 17 receptor signaling for lung CXC chemokineand granulocyte colony-stimulating factor expression, neutrophil recruitment, and hostdefense. J. Exp. Med. 194, 519–527 (2001).
39. Antonysamy, M.A. et al. Evidence for a role of IL-17 in organ allograft rejection: IL-17promotes the functional differentiation of dendritic cell progenitors. J. Immunol. 162,577–584 (1999).
40. Shortman, K. & Liu, Y.J. Mouse and human dendritic cell subtypes. Nat. Rev. Immunol.2, 151–161 (2002).
41. Maldonado-Lopez, R. et al. CD8a+ and CD8 a– subclasses of dendritic cells direct thedevelopment of distinct T helper cells in vivo. J. Exp. Med. 189, 587–592 (1999).
42. Gutcher, I. et al. Interleukin 18–independent engagement of interleukin 18 receptor-ais required for autoimmune inflammation. Nat. Immunol. 7, 946–953 (2006).
43. Plumb, J., Armstrong, M.A., Duddy, M., Mirakhur, M. & McQuaid, S. CD83-positivedendritic cells are present in occasional perivascular cuffs in multiple sclerosis lesions.Mult. Scler. 9, 142–147 (2003).
44. Serafini, B. et al. Dendritic cells in multiple sclerosis lesions: maturation stage, myelinuptake, and interaction with proliferating T cells. J. Neuropathol. Exp. Neurol. 65,124–141 (2006).
45. Nestle, F.O. et al. Plasmacytoid predendritic cells initiate psoriasis through interferon-alpha production. J. Exp. Med. 202, 135–143 (2005).
46. Jongbloed, S.L. et al. Enumeration and phenotypical analysis of distinct dendriticcell subsets in psoriatic arthritis and rheumatoid arthritis. Arthritis Res. Ther. 8, R15(2006).
47. Farkas, L., Beiske, K., Lund-Johansen, F., Brandtzaeg, P. & Jahnsen, F.L. Plasmacytoiddendritic cells (natural interferon-a/b-producing cells) accumulate in cutaneous lupuserythematosus lesions. Am. J. Pathol. 159, 237–243 (2001).
48. Ludewig, B., Odermatt, B., Landmann, S., Hengartner, H. & Zinkernagel, R.M.Dendritic cells induce autoimmune diabetes and maintain disease via de novoformation of local lymphoid tissue. J. Exp. Med. 188, 1493–1501 (1998).
49. Drayton, D.L., Ying, X., Lee, J., Lesslauer, W. & Ruddle, N.H. Ectopic LT ab directslymphoid organ neogenesis with concomitant expression of peripheral node addressinand a HEV-restricted sulfotransferase. J. Exp. Med. 197, 1153–1163 (2003).
50. Magliozzi, R., Columba-Cabezas, S., Serafini, B. & Aloisi, F. Intracerebral expression ofCXCL13 and BAFF is accompanied by formation of lymphoid follicle-like structures inthe meninges of mice with relapsing experimental autoimmune encephalomyelitis.J. Neuroimmunol. 148, 11–23 (2004).
51. Yoneyama, H. et al. Evidence for recruitment of plasmacytoid dendritic cell precursorsto inflamed lymph nodes through high endothelial venules. Int. Immunol. 16, 915–928(2004).
52. Vecchi, A. et al. Differential responsiveness to constitutive vs. inducible chemokines ofimmature and mature mouse dendritic cells. J. Leukoc. Biol. 66, 489–494 (1999).
53. Pashenkov, M. et al. Two subsets of dendritic cells are present in human cerebrospinalfluid. Brain 124, 480–492 (2001).
54. Huang, Y.M. et al. Multiple sclerosis is associated with high levels of circulating dendriticcells secreting pro-inflammatory cytokines. J. Neuroimmunol. 99, 82–90 (1999).
55. El Behi, M. et al. New insights into cell responses involved in experimentalautoimmune encephalomyelitis and multiple sclerosis. Immunol. Lett. 96, 11–26(2005).
56. Sanna, A. et al. Glatiramer acetate reduces lymphocyte proliferation and enhances IL-5and IL-13 production through modulation of monocyte-derived dendritic cells inmultiple sclerosis. Clin. Exp. Immunol. 143, 357–362 (2006).
57. Stasiolek, M. et al. Impaired maturation and altered regulatory function of plasma-cytoid dendritic cells in multiple sclerosis. Brain 129, 1293–1305 (2006).
58. Lopez, C., Comabella, M., Al-Zayat, H., Tintore, M. & Montalban, X. Altered maturationof circulating dendritic cells in primary progressive MS patients. J. Neuroimmunol.175, 183–191 (2006).
59. Ochando, J.C. et al. Alloantigen-presenting plasmacytoid dendritic cells mediatetolerance to vascularized grafts. Nat. Immunol. 7, 652–662 (2006).
60. Waldner, H., Whitters, M.J., Sobel, R.A., Collins, M. & Kuchroo, V.K. Fulminantspontaneous autoimmunity of the central nervous system in mice transgenic for themyelin proteolipid protein-specific T cell receptor. Proc. Natl. Acad. Sci. USA 97,3412–3417 (2000).
61. Begolka, W.S., Vanderlugt, C.L., Rahbe, S.M. & Miller, S.D. Differential expressionof inflammatory cytokines parallels progression of central nervous system pathology intwo clinically distinct models of multiple sclerosis. J. Immunol. 161, 4437–4446(1998).
180 VOLUME 8 NUMBER 2 FEBRUARY 2007 NATURE IMMUNOLOGY
A R T I C L E S©
2007
Nat
ure
Pub
lishi
ng G
roup
ht
tp://
ww
w.n
atur
e.co
m/n
atur
eim
mun
olog
y