ava i l ab l e a t www.sc i enced i r ec t . com

C l i n i ca l Immuno logy

www.e l sev i e r . com/ l oca te /yc l im

Clinical Immunology (2010) 137, 422–432

The clinical course of EAE is reflected by the dynamicsof the neuroantigen-specific T cell compartment inthe bloodStefanie Kuerten a,⁎, Andrea Rottlaender a, Michael Rodi a,Virgilio B. Velasco Jr. b, Michael Schroeter b,c, Claudia Kaiser c,Klaus Addicks a, Magdalena Tary-Lehmann d,1, Paul V. Lehmann b,d,1

a Department of Anatomy I, University Hospitals of Cologne, Cologne, Germanyb Cellular Technology Limited, 20521 Chagrin Boulevard, Shaker Heights, OH 44122, USAc Department of Neurology, University Hospitals of Cologne, Cologne, Germanyd Department of Pathology, Case Western Reserve University, Cleveland, OH 44106, USA

Received 2 August 2010; accepted with revision 14 September 2010Available online 8 October 2010

Abbreviations: B6, C57BL/6; drLNsmyelin oligodendrocyte glycoprotein;PLP, proteolipid protein; PLPp, PLP petoxin.⁎ Corresponding author. Department

Cologne, Joseph-Stelzmann-Str. 9,Fax: +49 221 478 6711.

E-mail address: stefanie.kuerten@u1 These authors contributed equally

1521-6616/$ – see front matter © 201doi:10.1016/j.clim.2010.09.004

KEYWORDSCytokines;EAE;ELISPOT;Immune monitoring;MS

Abstract Due to the limited numbers of PBMCs that can be obtained from the blood ofindividual mice, the key question whether central disease parameters such as onset, progressionand severity correlate with the magnitude and cytokine quality of the T cell response inexperimental autoimmune encephalomyelitis (EAE) has remained unanswered.Here we introduce an ELISPOT-based PBMC test system in which as little as 150 μl of murine bloodare sufficient, allowing to bleed mice repeatedly while continuing to observe the clinical courseof EAE. Using this technique, we demonstrate that longitudinal measurements of antigen-specific

IFN-γ and IL-17 production in the blood are a highly suitable approach to predict the diseaseoutcome in remitting–relapsing PLP:139–151- and chronic MOG:35–55-induced EAE of SJL/J andC57BL/6 mice, respectively.Our data propound cytokine monitoring as promising tool in the quest for more efficientdiagnostic and prognostic options in human multiple sclerosis and other autoimmune diseases.© 2010 Elsevier Inc. All rights reserved.

, draining lymph nodes; MOG,MOGp, MOG peptide 35–55;ptide 139–151; PTx, pertussis

of Anatomy I, University of50931 Cologne, Germany.

k-koeln.de (S. Kuerten).to this work.

0 Elsevier Inc. All rights reser

ved

1. Introduction

Experimental autoimmune encephalomyelitis (EAE) is acommon model for multiple sclerosis [1–4]. EAE is inducedby immunization with neuroantigens that trigger an autoim-mune T cell response. In particular, the proteolipid protein(PLP) peptide 139–151-induced EAE in SJL mice [5,6] and themyelin oligodendrocyte glycoprotein (MOG) peptide 35–55-

.

Figure 1 Neuroantigen-specific IFN-γ and IL-17 responses canbe detected in the blood of immunized mice. (A) SJL and B6 micewere either immunized with PBS (n=8 mice per group) or therespective neuroantigen (n=16 SJL mice immunized with PLPp,n=14 B6 mice immunized with MOGp) in CFA. 200 ng of pertussistoxin was given on the day of immunization and 48 h later. Onday 10 after immunization, blood was obtained from the tailvein, then PBMCs were isolated and tested for their respectiveneuroantigen-specific IFN-γ (white circles) and IL-17 (blackcircles) response in double-color ELISPOT assays. Circles refer tothe results obtained from individual mice, while bars denote themean value±SD for each group. Whereas in PBS/CFA-immunizedSJL and B6 mice a neuroantigen-specific response was absent(spot numbers were typicallyb3 spots per 106 cells tested),clear-cut responses were evident in PLPp- and MOGp-immunizedmice for both IFN-γ and IL-17. All results are medium-subtractedand normalized to 106 cells. The results are representative oftwo independent experiments. (B–G) Representative wellimages of single- and double-color IFN-γ/IL-17 ELISPOT assaysperformed with PBMCs isolated from the blood of PLPp-immunized SJL mice 10 days after immunization. PBMCs weretested in the absence (medium) and presence of antigen (PLPp).

423Clinical EAE correlates with antigen-specific IFN-γ/IL-17

triggered disease in C57BL/6 (B6) mice [7,8] have evolved asprevalent murine models. Whereas the former is character-ized by a predictable relapsing–remitting course, the diseaseis chronic in the latter. In both models, the respectiveneuroantigen-specific CD4+ T cells are known to cause theimmunopathology in the absence of antibodies playing asignificant role [9–12].

CD4+ T cells can mediate the immune pathology bysecreting cytokines that activate and attract cells of theinnate immune system, primarily macrophages, into adelayed-type hypersensitivity (DTH)-like reaction. For manyyears, the mediation of DTH in the CNS was attributed to IFN-γ secreting “TH1 type” cells and subsequently, most efforts ofimmune monitoring focused on measuring IFN-γ to identifythe autoreactive effector cells [13–16]. More recentlythough, the pathogenic/encephalitogenic quality of auto-reactive cells was linked to the ability to produce IL-17,accounting for a TH lineage independent of the IFN-γproducing TH1 cell and termed TH17 [17–19]. In line withthis basic understanding of the immune pathology of EAE, onemight predict that the magnitude of the neuroantigen-specific T cell response as reflected by the frequencies ofthe IFN-γ or IL-17 producing T cells is the primary criterionthat defines the severity of the disease.

It has been well established that after immunization witha neuroantigen including PLP peptide (PLPp) and MOGpeptide (MOGp), the frequencies of IFN-γ and IL-17producing cells in draining lymph nodes (drLNs) and spleenare highly variable, ranging from very low to rather high [20].It is also well known that the severity of the disease showsconsiderable interindividual fluctuations, ranging from mildto lethal, even when keeping the variables that couldinfluence the disease outcome constant (e.g. age of themice at immunization, gender, genetic background, colony,etc.) [20,21]. Taking the causative relationship betweenCD4+ T cells producing IFN-γ and IL-17 and the developmentof EAE into account, the hypothesis is intuitive that themagnitude of the T cell response and the disease severity willbe in direct correlation to each other. In other words, thosemice that do not develop severe disease — for example, dueto their T cell repertoires — will not display a high frequencyof effector cells, whereas a high effector cell mass willaccumulate in those mice, in which the pathology is severeenough to cause lethality. However, this hypothesis has notbeen tested in mice so far. This is because T cell assaystypically require 5×105 to 1×106 cells per test condition, andif one includes control and experimental sets in replicates,then 3 to 6 ×106 cells may be needed for a singlemeasurement. For this reason, most immune monitoring inmice has been performed with lymphoid tissue such as drLNsand spleen, from which sufficient cell numbers can beobtained for testing. Needless to say, mice have to besacrificed for such tests. Therefore, no information can beobtained as to what clinical disease would have looked like ifthey had been permitted to live longer.

We previously introduced an ELISPOT assay for testingmouse blood in order to move closer to the human standardwhere blood is the primary sampling site for measuring T cellreactivity [20]. For these first generation murine bloodassays, a sufficient amount of blood could be obtained forexamining individual mice, but only if the mice weresacrificed. Thus, these assays were not suited for longitudi-

nal studies either. Here we developed a variant of the assaythat functioned with only 150 μl of blood and therebypermitted repeated bleeding of individual mice, while theclinical course of EAE could be observed over time. To studyboth IFN-γ and IL-17, we performed double-color measure-ments. These technical advances enabled us to experimen-tally address one of the weighty outstanding questions inautoimmune research — that is, how the T cell reactivity in

424 S. Kuerten et al.

the blood is reflective of the clinical course of disease inindividual mice.

2. Materials and methods

2.1. Mice

Female SJL/J and C57BL/6 (B6) mice (6–8 weeks old) werepurchased from the Jackson Laboratory (Bar Harbor, ME) andmaintained at the animal facilities of Case Western ReserveUniversity, Cleveland, OH, USA under specific pathogen-freeconditions. All treatments were performed according to anapproved IACUC protocol and complied with the institutionalguidelines.

2.2. Induction and clinical assessment of EAE

IFA was prepared as a mixture of mannide monooleate(Sigma-Aldrich Corp., St. Louis, MO) and paraffin oil(EMScience, Gibbstown, NJ), CFA was obtained by mixingMycobacterium tuberculosis H37RA (Difco Laboratories,Franklin Lakes, NJ) at 5 mg/ml into IFA. For diseaseinduction, mice were immunized subcutaneously in bothsides of the flank with a total dose of 100 μg PLP peptide139–151 (SJL/J) or 100 μg MOG peptide 35–55 (B6). Bothpeptides were obtained from Princeton Biomolecules,Langhorne, PA. 200 ng pertussis toxin (PTx; List BiologicalLaboratories, Hornby, ONT, Canada) in 500 μl sterile PBS wasgiven on the day of immunization and 48 h later. Control

Figure 2 The double-color IFN-γ/IL-17 ELISPOT assay detectscorresponding single-color assay. SJL and B6 mice were immunized(SJL; panels A and B) or 10, 25 and 40 (B6; panels C and D) after immthen tested for their antigen-specific IFN-γ (A, C) and IL-17 (B, D) prof spots obtained in the single-color assay was correlated to the resulSJL and n=17 B6 mice were tested in four and three independent exnormalized to 106 cells.

mice received both CFA and PTx, however in the absence ofneuroantigen. Clinical assessment of EAE was performeddaily according to the following criteria: (0), no disease; (1),floppy tail; (2), hind limb weakness; (3), full hind limbparalysis; (4), quadriplegia and (5), death. Mice that were inbetween the clear-cut gradations of clinical signs werescored intermediate in increments of 0.5.

2.3. Cell preparations from various organs

Mouse blood (150 μl permouse and timepoint)was collected bytail vein bleeding using heparin as anti-coagulant. The bloodwas diluted at 1:2 with sterile calcium-free PBS. Aftercentrifugation, red blood cells were lysed with ammoniumchloride. Cells were then washed thoroughly with HL-1. Singlecell suspensions from spleen and draining lymph nodes wereprepared as previously described [15]. All cells were countedusing acridine orange/ethidiumbromide (Sigma-Aldrich Corp.).

2.4. ELISPOT assays and image analysis

Whatman Unifilter low-volume plates (Whatman Inc., FlorhamPark, NJ)were coated overnightwith the capture antibodies insterile PBS. AN-18 (eBioscience, San Diego, CA) was used at3 μg/ml for capturing IFN-γ and TC-11-18H10 (BD Pharmingen,San Jose, CA) at 4 μg/ml for IL-17. For double-color assays,anti-IFN-γ antibodywas applied to the plate 10 min before theaddition of anti-IL-17. The plates were blocked for 2 h at roomtemperature with sterile PBS containing 1% bovine serumalbumin (Sigma-Aldrich Corp.). The cells were plated in HL-1

the frequencies of neuroantigen-specific cells equal to thewith the corresponding neuroantigen. On days 10, 20, 30 and 40unization, PBMCs were obtained from blood, spleen and drLNs,

oduction in single- and double-color ELISPOT assays. The numberts in the double-color assay for each cell sample. A total of n=30periments, respectively. All results are medium-subtracted and

Figure 3 Absent versus positive correlation between thefrequencies of neuroantigen-specific IFN-γ and IL-17 producingT cells and initial disease severity in PLPp- vs. MOGp-inducedEAE of SJL and B6 mice, respectively. N=16 SJL and n=10 B6mice were immunized with the respective neuroantigen. On day10 after immunization, PBMCs were isolated from the blood(collected from the tail vein) and tested for their neuroantigen-specific IFN-γ and IL-17 response in double-color ELISPOT assays.The magnitude of the T cell response was correlated with themaximal onset severity of the disease in all mice. (A) This graphschematically delineates the two parameters that were corre-lated to each other in both the PLPp/SJL and the MOGp/B6model, that is the antigen-specific IFN-γ and IL-17 responsemeasured on day 10 after immunization and the subsequentmaximal disease severity. (B) The correlation between themagnitude of the PLPp-specific IFN-γ (white circles) and IL-17(black circles) response and the maximal severity of the initialdisease episode in SJL mice is shown. Each circle refers to anindividual mouse. (C) The correlation between the magnitude ofthe MOGp-specific IFN-γ (white circles) and IL-17 (black circles)response and the maximal onset severity in B6 mice is shown.Each circle refers to an individual mouse. Results are represen-tative of two independent experiments performed. All resultsare medium-subtracted and normalized to 106 cells.

425Clinical EAE correlates with antigen-specific IFN-γ/IL-17

supplemented with 1 mM glutamine without (medium) or withantigen (final concentration of 15 μg/ml for both MOG:35–55and PLP:139–151). The plates were cultured at 37 °C and 7%CO2 for 24 h. Subsequently, the detection antibodies wereadded for overnight incubation. FITC-labeled rat anti-mouseR4-6A2 (purified and FITC-labeled in our laboratory) was usedat 0.5 μg/ml for detecting IFN-γ, while rat anti-mouse TC-11-8H4.1-biotin was used for the detection of IL-17 (BD Pharmin-gen; 0.5 μg/ml). For double-color assays, both antibodieswereadded at the same time. As a third reagent for IFN-γ, anti-FITC-AP (DakoNorth America, Inc., Carpinteria, CA)was addedat 1:500 dilution in PBS containing 1% BSA and 0.025% Tween20, and Streptavidin-HRP (Dako)was added at 1:1000 for IL-17.After incubation for 2 h at room temperature, plates weredeveloped using the Vector Blue substrate kit (VectorLaboratories, Inc., Burlingame, CA) for visualizing IFN-γ andAEC substrate solution (Cellular Technology Limited, Cleve-land, OH) for IL-17. For double-color assays, blue spots weredeveloped prior to red spots. The plates were analyzed on anImmunoSpot Series 5 Analyzer. All results were normalized to106 cells, and the difference between stimulated andnonstimulated cells was calculated.

2.5. Statistical analysis

SigmaStat software (Version 7.0; SPSS, Chicago, IL) was usedto assess correlation coefficients and the statistical signif-icance of the differences between two groups using Student'st-test. Statistical significance was set at P≤0.05.

3. Results

3.1. Neuroantigen-specific IFN-γ and IL-17 responsescan be detected in the blood of immunized mice

To determine whether neuroantigen-specific IFN-γ and IL-17responses could be detected in the blood of mice, n=16 SJLmice were immunized with 100 μg PLPp in CFA, and n=14 B6mice with 100 μg MOGp in CFA. In addition, n=8 SJL and n=8B6 mice received PBS in the absence of antigen in CFA. PTxwas given on the day of immunization and 48 h later. On day10 after immunization, mice were bled and PBMCs wereisolated as described in Materials and methods and tested fortheir respective neuroantigen-specific IFN-γ and IL-17response in double-color ELISPOT assays. Results are shownin Figure 1A. PBS/CFA immunization did not trigger PLPp- orMOGp-specific IFN-γ and IL-17 responses in SJL or B6 mice,respectively. In contrast, immunization with PLPp in CFAinduced clear-cut PLPp-specific IFN-γ and IL-17 responses inSJL mice. Likewise, immunization with MOGp in CFA inducedMOGp-specific IFN-γ and IL-17 producing cells in B6 mice.While the difference between IFN-γ and IL-17 producing cellswas statistically significant in the PLPp model (pb0.001),spot numbers were similar in MOGp-induced EAE (p=0.564).Figures 1B–G display representative images from single-colorIFN-γ (Figs. 1B and E) and IL-17 (Figs. 1C and F) ELISPOTassays compared to the double-color variant (Figs. 1D and G)that we have established specifically for this study. Spotcounts in the medium control wells were typically low (b3spots/106 cells) (Figs. 1B–D) and neuroantigen-specific

426 S. Kuerten et al.

responses could clearly be detected over this background(Figs. 1E–G).

3.2. The double-color IFN-γ/IL-17 ELISPOT assaydetects the frequencies of neuroantigen-specificcells equal to the corresponding single-color assay

The sensitivity and specificity of the double-color IFN-γ/IL-17ELISPOT assay was assessed by performing single- and double-color measurements in parallel. Results are shown in Figure 2,correlating the number of spots detected in single- and double-color IFN-γ (Figs. 2A and C) and IL-17 (Figs. 2B and D) ELISPOTassays testing identical cell samples. Results are representa-tive for a total of n=30 SJL mice (Figs. 2A and B) and n=17 B6mice (Figs. 2C and D)mice. For bothmodels and cytokines, the

Figure 4 Recovery from initial EAE is accompanied by a drop in PLN=12 individual SJL mice (A–L) were immunized with PLPp. PBMCs wto the initial episode of the disease and the remission, respectively.assessed in each individual mouse (A–L) on both time points in doucircles to IL-17. The shaded areas denote the clinical disease courscells, and representative of two independent experiments.

correlation coefficients of r2N0.95 delineate that the double-color performance was equal to the individual cytokinemeasurements.

3.3. There is no correlation between the frequenciesof PLPp-specific IFN-γ and IL-17 producing T cells andinitial disease severity inPLPp-inducedEAEof SJLmice

Given that the primary variable that defines the immunepathology of EAE is the neuroantigen-specific T cell response,it could be assumed that the magnitude of this response iscorrelated with the severity of the disease. This hypothesiswas assessed in n=16 SJLmice that were immunizedwith PLPpin CFA, with PTx given on the day of immunization and 48 hlater. All mice were bled on day 10 after immunization and the

Pp-specific T cell frequencies in PLPp-induced EAE of SJL mice.ere isolated on days 10 and 20 after immunization, correspondingThe magnitude of the PLPp-specific IFN-γ and IL-17 response wasble-color ELISPOT assays. White circles refer to IFN-γ and blacke over time. Results are medium-subtracted, normalized to 106

427Clinical EAE correlates with antigen-specific IFN-γ/IL-17

magnitude of the PLPp-specific T cell response was deter-mined in double-color IFN-γ/IL-17 ELISPOT assays. Theresulting IFN-γ and IL-17 spot numbers were then correlatedwith the maximal severity of the following initial diseaseepisode that typically occurred on day 13 after immunizationin the SJL model. Figure 3A schematically delineates therelation between both parameters. Figure 3B shows that therewas no correlation between the PLPp-specific IFN-γ and IL-17response measured before EAE onset and the subsequentmaximal disease severity in the PLPp/SJL model (r2=0.129 forIFN-γ, r2=0.097 for IL-17). The maximal disease severityranged from 1 to 4.

3.4. Positive correlations exist between thefrequencies of MOGp-specific IFN-γ and IL-17producing T cells and initial disease severity inMOGp-induced EAE of B6 mice

The correlation between the neuroantigen-specific IFN-γ/IL-17response and the maximal disease severity was also assessed in

Figure 5 Clinical relapse is accompanied by a reincrease of PLPp-legend for Figure 4 applies except that the n=12 individual miccorresponding to clinical remission and relapse, respectively.

MOGp-induced EAE of B6 mice (for a schematic graphdelineating the relation between the two parameters seeFig. 3A). Results are shown in Figure 3C and are representativefor a total of n=10 mice tested. In this model a correlationbetween the magnitude of the MOGp-specific IFN-γ and IL-17response measured prior to EAE onset and the consecutivemaximal disease severity was evident (r2=0.93 for IFN-γ,r2=0.72 for IL-17). In theMOGpmodel the onset severity rangedfrom 0.5 to 3.5 and was typically reached on day 18 afterimmunization.

3.5. Clinical recovery from initial EAE isaccompanied by a drop in PLPp-specific T cellfrequencies in PLPp-induced EAE of SJL mice

The double-color ELISPOT assay that we developed functionedwith as little as 150 μl of murine blood. This allowed us torepeatedlymeasure the neuroantigen-specific T cell responsesin individual mice while following up on the clinical course ofEAE. N=12 SJL mice were immunized with PLP peptide 139–

specific T cell frequencies in PLPp-induced EAE of SJL mice. Thee (A–L) were tested on days 20 and 30 after immunization,

Figure 6 The chronic disease course of MOGp-induced EAE is paralleled by stable frequencies of MOGp-specific T cells. N=10individual B6 mice (A–J) were immunized with MOGp. PBMCs were isolated on days 10, 25 and 40 after immunization, reflecting thepre-onset, peak and chronic stage of the disease, respectively. The magnitude of the MOGp-specific IFN-γ and IL-17 response wasassessed in each individual mouse (A–J) on all time points in double-color ELISPOT assays. White circles refer to IFN-γ and black circlesto IL-17. The shaded areas denote the clinical disease course over time. Results are medium-subtracted, normalized to 106 cells, andrepresentative of three independent experiments.

428 S. Kuerten et al.

151 in CFA,with PTx given on the day of immunization and 48 hlater. On days 10 and 20 after immunization — correspondingto the acute stage of EAE and remission —mice were bled and

the PLPp-specific IFN-γ and IL-17 response was assessed.Results are shown in Figures 4A–L, with each panel referring toan individual mouse. In 10 out of 12 mice tested, the clinical

0

200

600

800

0

1

2

3

4A

0

100

200

300

400

0

100

200

300

400

Mea

n cl

inic

al s

core

Num

ber

of s

pots

0

1

2

3

4

p = 0.98 p = 0.93

IL-17

p = 0.93 p = 0.76

Num

ber

of s

pots

B

D

F

0 10 20 30 40 500 10 20 30 400

200

600

800

IFN-γ IFN-γ

p = 0.003

p = 0.019

IL-17

p = 0.004

p = 0.001

C

E

Days after immunization

MOGp/B6PLPp/SJL

Days after immunization

Figure 7 The antigen-specific IFN-γ and IL-17 response in the blood of PLPp- and MOGp- immunized SJL/J and B6 mice is reflectiveof the clinical course of EAE. Summary of Figures 4–6. (A) SJL/J mice were immunized with PLPp. EAE scores were assessed daily. Ondays 10, 20 and 30 150 μl of blood was obtained from the tail vein, PBMC isolated and tested for their PLPp-specific IFN-γ (C) and IL-17(E) response in double-color ELISPOT assays. (B) B6 mice were immunized with MOGp. EAE scores were assessed daily. On days 10, 25and 40 150 μl of blood was obtained from the tail vein, PBMC isolated and tested for their MOGp-specific IFN-γ (D) and IL-17(F) response as above.

429Clinical EAE correlates with antigen-specific IFN-γ/IL-17

recovery from EAEwas accompanied by a drop in PLPp-specificT cell frequencies (Figs. 4A–K), in one mouse (Fig. 4H) thefrequencies were unchanged, one mouse died of EAE (Fig. 4L).When calculating the mean PLPp-reactive T cell frequenciesfor all mice, there was a significant decrease in IFN-γ spotsfrom a mean of 55.9 to 7.1 spots per 106 cells (p=0.0004). Thesame applied to IL-17. Here, the number of spots decreasedfrom a mean of 89.5 to 23.4 spots per 106 cells (p=0.0001).Both cytokines — IFN-γ and IL-17 — followed a similar trend.

3.6. Clinical relapse is reflected by a drasticreincrease of PLPp-specific T cell frequencies inPLPp-induced EAE of SJL mice

The dynamics of PLPp-specific T cell frequencies relative tothe clinical course of EAE were further assessed by comparingremission and relapse. The same n=12 SJL mice as above wereadditionally bled on day 30 after immunization correspondingto the clinical relapse. PBMCs were isolated and again testedfor their PLPp-specific IFN-γ/IL-17 response in double-colorELISPOT assays. Results are given in Figures 5A–L, with eachpanel referring to an individual mouse. In 9 out of 12 micetested, the re-onset of disease was accompanied by aremarkable reincrease in PLPp-specific T cell frequencies(Figs. 5A–G, J and K). In one mouse (the same as above) thefrequencies remained unchanged (Figs. 5H). Two mice died of

EAE (Figs. 5I and L). Taking all mice together, spot countssignificantly increased from amean of 7.1 to 212.5 IFN-γ spotsper 106 cells (p=0.005) and from a mean of 23.4 to 320.2 IL-17spots per 106 cells (p=0.01). Overall, both antigen-specificIFN-γ and IL-17 responses behaved similarly.

3.7. The chronic course of disease in MOGp-inducedEAE of B6 mice is paralleled by stable frequencies ofMOGp-specific T cells

As the dynamics of clinical disease were reflected byfluctuations in the T cell compartment in PLPp-inducedEAE of SJL mice, we assessed whether the chronic course ofthe disease was equally accompanied by stable frequenciesof MOGp-specific T cells in MOGp-induced EAE of B6 mice.N=10 B6 mice were immunized with MOGp in CFA and PTxwas given on the day of immunization and 48 h later. Micewere bled on days 10, 25 and 40 after immunization. PBMCswere isolated and tested for their MOGp-specific IFN-γ andIL-17 response in double-color ELISPOT assays while wefollowed up on the clinical course of EAE. Results areshown in Figures 6A–J, with each panel representing theresults for an individual mouse. In 10 out of 10 mice, thechronic non-remitting–relapsing course of EAE wasreflected by stable frequencies of MOGp-specific IFN-γproducing T cells (means of 64.7, 60.2 and 72.7 for days 10,25 and 40 with p=0.81 and p=0.45 respectively). The same

430 S. Kuerten et al.

applied to IL-17 (means of 69.2, 60.3 and 67.2 for days 10,25 and 40 with p=0.63 and p=0.77, respectively). Again,both cytokines — IFN-γ and IL-17 — followed a similartrend.

Figure 7 summarizes the relationship between thefrequencies of IFN-γ/IL-17 producing T cells and theclinical disease course in both models: while the dynamiccourse of EAE was reflected by dynamics in the bloodT cell compartment in the PLPp/SJL model (Figs. 7A, C andE), chronic disease was mirrored by stable antigen-specificIFN-γ/IL-17 responses in the MOGp/B6 model (Figs. 7B, Dand F).

4. Discussion

There are two primary options for immune monitoring in MS.One involves the testing of cerebrospinal fluid itself which,however, cannot easily be done repeatedly and longitudi-nally in patients. The other one is testing of peripheral bloodthat, due to its accessibility and the availability ofconsiderable amounts of PBMCs, is the primary site ofsampling. For this reason, most data on neuroantigen-reactive T cells in MS have been obtained using PBMCs fromthe peripheral blood [22–29]. While EAE is generallyconsidered a model for MS, studies of murine blood have sofar been limited by the low numbers of cells that can beobtained without jeopardizing the survival of the mice.

In the blood of MS patients, it has been challenging toreliably measure neuroantigen-specific T cells [30]. There-fore, we first set out to investigate whether such cells couldbe reliably detected in the blood of mice undergoing EAE. Asshown in Figure 1A, all immunized mice displayed a specificand clear-cut recall response to the immunizing antigen.Thus, at least in murine blood, there were no inherentlimitations for detecting the antigen-specific T cells. In thesemice, however, the identity of the neuroantigen is welldefined, being the injected peptide. Moreover, all measure-ments were done within 40 days after immunization. Incontrast, in humans the nature of the primary antigen is notknown, and the immunological onset of the disease/theevents that triggered initial T cell responses may haveoccurred months or years back [30,31]. In mice, it is well-established that with time, the first wave of autoreactive Tcells induced by immunization exhausts and contracts inclonal sizes while determinant spreading engages anddiversifies the autoreactive repertoire [32–36]. Both thesemechanisms may also occur in humans, each complicatingthe detection of autoreactive T cells in the blood.

In the PLPp/SJL model, we did not find a correlationbetween the numbers of neuroantigen-specific IFN-γ or IL-17secreting T cells prior to EAE onset and the severity of theinitial disease episode. This might be explained by the easewith which determinant spreading occurs in these mice: incontrast to MOGp-induced EAE, the PLPp model involveseffector cells that are specific for determinants other thanPLP:139–151 and that possibly need to be considered infuture studies to delineate a correlation with the clinicalseverity [33,35]. Remarkably, however, the subsequentrelapsing–remitting disease course was mirrored by dynamicfrequencies of neuroantigen-specific T cells in the blood.After the acute stage of the disease, the frequencies of the

first wave autoreactive T cells dropped in 10 out of 11 mice(one additional mouse died of EAE), paralleling the clinicalrecovery from EAE. Previous work has demonstrated thatunlike during active primary disease or relapse, in such statesof remission, few if any functional T cells are found in theCNS [20,37]. Therefore, one might postulate that the drop inT cell numbers in the blood reflects an actual exhaustion ofthe first wave of effector cells in SJL mice rather than aredistribution towards the CNS. Why this reduction infrequencies and the recovery itself occurs in SJL mice versusthe B6 model is unclear. Nevertheless, these results areconsistent with the primary effector cell hypothesis,provided this exhaustion of the majority of T cells alsoaffects those cells engaged by determinant spreading duringthe first episode of the disease.

It was striking that in parallel to the EAE relapse, massiverenewed expansions of PLPp-specific T cells were seen in 9out of 10 mice (two additional mice died of EAE). It still needsto be elucidated what drives this renewed expansion, but thisexpansion could explain the occurrence of the clinical diseaseexacerbation well. Neuroantigen expression in peripheraltissues has been observed [38–41] and could boost theautoreactive T cells in addition to neuroantigen release fromthe injured target organ itself [32,33]. The breakdown ordysfunction of immune regulatory mechanisms, includingregulatory T cells, may also be suggested [42–44].

Our studies of MOGp-induced EAE provided results thatwere possibly in line with a simple pathomechanism. Incontrast to the PLPp/SJL model, we found that thefrequencies of both IFN-γ and IL-17 producing T cells priorto disease onset correlated with the subsequent maximaldisease severity. These data correspond to the notion thatthe neuroantigen-specific T cells are the primary mediatorsof the disease. Accordingly, the higher the number of specificT cells, the higher the effector potential and the more severethe resulting disease. This type of direct relationship is alsoknown for passive EAE, in which graded numbers ofneuroantigen-specific T cells are injected into naïve recipi-ents [45]. In the MOGp model, the frequencies of the specificT cells were stable over the observation period and so was theclinical course of the disease. These data might beinterpreted to mean that even in chronic disease, the Tcells engaged by immunization (as opposed to those engagedby determinant spreading) continue to be the primaryeffector cells. This direct relationship does not excludemore complex mechanisms. However, while there is ampleevidence for determinant spreading in the PLPp-induced SJLmodel [34–36], our literature search did not provide evidencefor spreading in the MOGp-induced EAE of B6 mice.

5. Conclusion

Our data clearly show that the limited amount of bloodavailable from mice does not prevent systematic studies ofthe effector cell pool, while offering the possibility ofobserving the clinical disease in parallel. Thereby, we fill animportant gap that has previously hampered immunediagnostic/prognostic research in EAE. This study was notintended to provide detailed insights into the compleximmune mechanisms that underlie neuroantigen-inducedEAE, including comprehensive studies of cytokine signatures

431Clinical EAE correlates with antigen-specific IFN-γ/IL-17

of the effector cells and secondary repertoire dynamicsinvolved in spreading. However, the initial experimentspresented here provide a striking observation, namely, thatneuroantigen-specific T cell frequencies in the blood arereflective of the clinical course of EAE. In addition, it hasbecome increasingly clear that effector cell types are notstable, but tend to convert from TH17 to TΗ1 in vivo [46]. Byoffering the possibility of monitoring such dynamics, ourassay can contribute to the identification of the dominanteffector cell populations causing pathology in each partic-ular disease stage. As more is learned about the neuroanti-gens that are actually targeted in MS, our approach shouldtherefore also serve as valuable tool in the quest for moreefficient diagnostic and prognostic options in patients wherethe blood is the primary material that is available for testing.

Conflict of interest

None of the authors has any potential financial conflict ofinterest related to this manuscript.

Acknowledgments

We would like to thank Villian Naeem and Richard Caspell fortheir valuable technical assistance as well as Ioana Moldovanand LindaMetz for critical reading of themanuscript. Thisworkwas supported by research grants from Cellular TechnologyLimited and the Koeln Fortune Program to S. K. The fundingsources had no involvement in the collection, analysis andinterpretation of data, in the writing of the report and in thedecision to submit the paper for publication.

References

[1] J. Goverman, T. Brabb, Rodent models of experimental allergicencephalomyelitis applied to the study of multiple sclerosis,Lab. Anim. Sci. 46 (1996) 482–492.

[2] M. Sospedra, R. Martin, Immunology of multiple sclerosis,Annu. Rev. Immunol. 23 (2005) 683–747.

[3] L. Steinman, S.S. Zamvil, How to successfully apply animalstudies in experimental allergic encephalomyelitis to researchon multiple sclerosis, Ann. Neurol. 60 (2006) 12–21.

[4] R. Gold, C. Linington, H. Lassmann, Understanding pathogen-esis and therapy of multiple sclerosis via animal models:70 years of merits and culprits in experimental autoimmuneencephalomyelitis research, Brain 129 (2006) 1953–1971.

[5] V.K. Tuohy, Z. Lu, R.A. Sobel, R.A. Laursen, M.B. Lees,Identification of an encephalitogenic determinant of myelinproteolipid protein for SJL mice, J. Immunol. 142 (1989)1523–1527.

[6] R.A. Sobel, V.K. Tuohy, Z.J. Lu, R.A. Laursen, M.B. Lees, Acuteexperimental allergic encephalomyelitis in SJL/Jmice induced bya synthetic peptide ofmyelin proteolipidprotein, J. Neuropathol.Exp. Neurol. 49 (1990) 468–479.

[7] I. Mendel, N.K. de Rosbo, A. Ben Nun, A myelin oligodendrocyteglycoprotein peptide induces typical chronic experimentalautoimmune encephalomyelitis in H-2b mice: fine specificityand T cell receptor V beta expression of encephalitogenicT cells, Eur. J. Immunol. 25 (1995) 1951–1959.

[8] I. Mendel, N.K. de Rosbo, A. Ben Nun, Delineation of theminimal encephalitogenic epitope within the immunodominantregion of myelin oligodendrocyte glycoprotein: diverse V betausage by T cells recognizing the core epitope encephalitogenic

for T cell receptor V beta b and T cell receptor V beta in H-2bmice, Eur. J. Immunol. 26 (1996) 2470–2479.

[9] N.T. Potter, T.S. Stephens, Humoral immune recognition ofproteolipid protein (PLP)-specific epitopes in the SJL/J mouse,J. Neurosci. Res. 37 (1994) 15–22.

[10] B.L. McRae, S.D. Miller, Fine specificity of CD4+ T cell responsesto the dominant encephalitogenic PLP 139–151 peptide in SJL/Jmice, Neurochem. Res. 19 (1994) 997–1004.

[11] P. Hjelmstrom, A.E. Juedes, J. Fjell, N.H. Ruddle, B-celldeficient mice develop experimental autoimmune encephalo-myelitis with demyelination after myelin oligodendrocyteglycoprotein sensitization, J. Immunol. 161 (1998) 4480–4483.

[12] A.R. Oliver, G.M. Lyon, N.H. Ruddle, Rat and human myelinoligodendrocyte glycoproteins induce experimental autoimmuneencephalomyelitis by different mechanisms in C57BL/6 mice,J. Immunol. 171 (2003) 462–468.

[13] J.E. Merrill, D.H. Kono, J. Clayton, D.G. Ando, D.R. Hinton,F.M. Hofman, Inflammatory leukocytes and cytokines in thepeptide-induced disease of experimental allergic encepha-lomyelitis in SJL and B10.PL mice, Proc. Natl Acad. Sci. USA89 (1992) 574–578.

[14] V. Navikas, H. Link, Review: cytokines and the pathogenesis ofmultiple sclerosis, J. Neurosci. Res. 45 (1996) 322–333.

[15] H.H. Hofstetter, A.Y. Karulin, T.G. Forsthuber, P.A. Ott, M.Tary-Lehmann, P.V. Lehmann, The cytokine signature of MOG-specific CD4 cells in the EAE of C57BL/6 mice, J. Neuroimmu-nol. 170 (2005) 105–114.

[16] H.H. Hofstetter, T.G. Forsthuber, Kinetics of IL-17- andinterferon-gamma producing PLPp-specific CD4 T cells in EAEinduced by coinjection of PLP/IFA with pertussis toxin in SJLmice, Neurosci. Lett. 476 (2010) 150–155.

[17] H.H. Hofstetter, S.M. Ibrahim, D. Koczan, N. Kruse, A.Weishaupt, K.V. Toyka, R. Gold, Therapeutic efficacy of IL-17neutralization in murine experimental autoimmune encepha-lomyelitis, Cell. Immunol. 237 (2005) 123–130.

[18] I.I. Ivanov, B.S. McKenzie, L. Zhou, C.E. Tadokoro, A. Lepelley,J.J. Lafaille, D.J. Cua, D.R. Littman, The orphan nuclearreceptor RORgammat directs the differentiation program ofproinflammatory IL-17+ T helper cells, Cell 126 (2006)1121–1133.

[19] Y. Komiyama, S. Nakae, T. Matsuki, A. Nambu, H. Ishigame, S.Kakuta, K. Sudo, Y. Iwakura, IL-17 plays an important role inthe development of experimental autoimmune encephalomy-elitis, J. Immunol. 177 (2006) 566–573.

[20] H.H. Hofstetter, O.S. Targoni, A.Y. Karulin, T.G. Forsthuber, M.Tary-Lehmann, P.V. Lehmann, Does the frequency and theavidity spectrum of the neuroantigen-specific T cells in theblood mirror the autoimmune process in the central nervoussystem of mice undergoing experimental allergic encephalo-myelitis? J. Immunol. 174 (2005) 4598–4605.

[21] S. Kuerten, D.A. Kostova-Bales, L.P. Frenzel, J.T. Tigno, M.Tary-Lehmann, D.N. Angelov, P.V. Lehmann, MP4- and MOG:35–55-induced EAE in C57BL/6 mice differentially targetsbrain, spinal cord and cerebellum, J. Neuroimmunol. 189 (2007)31–40.

[22] D.J. Mahad, J. Lawry, S.J.L. Howell, M.N. Woodroofe,Longitudinal study of chemokine receptor expression onperipheral lymphocytes in multiple sclerosis: CXCR3 upregu-lation is associated with relapse, Mult. Scler. 9 (2003)189–198.

[23] B. Bielekova, R. Martin, Development of biomarkers in multiplesclerosis, Brain 127 (2004) 1463–1478.

[24] L. Rinaldi, P. Gallo, Immunological markers in multiplesclerosis: tackling the missing elements, Neurol. Sci. 26 (2005)S215–S217.

[25] M. Reindl, M. Khalil, T. Berger, Antibodies as biological markersfor pathophysiological processes in multiple sclerosis, J. Neu-roimmunol. 180 (2006) 50–62.

432 S. Kuerten et al.

[26] A. Lutterotti, T. Berger, M. Reindl, Biological markers formultiple sclerosis, Curr. Med. Chem. 14 (2007) 1956–1965.

[27] R.J. Fox, P. Kivisakk, E. Fisher, B. Tucky, J.C. Lee, R.A. Rudick,R.M. Ransohoff, Multiple sclerosis: chemokine receptor expres-sion on circulating lymphocytes in correlation with radiographicmeasures of tissue injury, Mult. Scler. 14 (2008) 1036–1043.

[28] J. Drulovic, E. Savic, T. Pekmezovic, S. Mesaros, N. Stosjavl-jevic, I. Dujmovic-Basuroski, J. Kostic, V. Vasic, M. MostaricaStojkovic, D. Popadic, Expression of TH1 and TH17 cytokines andtranscription factors in multiple sclerosis patients: doesbaseline T-Bet mRNA predict the response to interferon-betatreatment? J. Neuroimmunol. 215 (2009) 90–95.

[29] S. Sawai, H. Umemura, M. Mori, M. Satoh, S. Hayakawa, Y. Kodera,T. Tomonaga, S. Kuwabara, F. Nomura, Serum levels of comple-ment C4 fragments correlate with disease activity in multiplesclerosis: proteomic analysis, J. Neuroimmunol. 218 (2010)112–115.

[30] C.M. Pelfrey, R.A. Rudick, A.C. Cotleur, J.C. Lee, M. Tary-Lehmann, P.V. Lehmann, Quantification of self-recognition inmultiple sclerosis by single-cell analysis of cytokine production,J. Immunol. 165 (2000) 1641–1651.

[31] R. Hohlfeld, H. Wekerle, Autoimmune concepts of multiplesclerosis as a basis for selective immunotherapy: from pipedreams to (therapeutic) pipelines, Proc. Natl Acad. Sci. USA101 (2004) 14599–14606.

[32] P.V. Lehmann, T. Forsthuber, A. Miller, E.E. Sercarz, Spreadingof T-cell autoimmunity to cryptic determinants of an autoanti-gen, Nature 358 (1992) 155–157.

[33] B.L. McRae, C.L. Vanderlugt, M.C. Dal Canto, S.D. Miller,Functional evidence for epitope spreading in the relapsingpathology of experimental autoimmune encephalomyelitis,J. Exp. Med. 182 (1995) 75–85.

[34] C.L. Vanderlugt, W.S. Begolka, K.L. Neville, Y. Katz-Lewy, L.M.Howard, T.N. Eagar, J.A. Bluestone, S.D. Miller, The functionalsignificance of epitope spreading and its regulation by co-stimulatory molecules, Immunol. Rev. 164 (1998) 63–72.

[35] C.L. Vanderlugt, K.L. Neville, K.M. Nikcevich, T.N. Eagar, J.A.Bluestone, S.D. Miller, Pathologic role and temporal appearanceof newly emerging autoepitopes in relapsing remitting experi-mental autoimmune encephalomyelitis, J. Immunol. 164 (2000)670–678.

[36] V.K. Tuohy, M. Yu, L. Yin, J.A. Kawczak, J.M. Johnson, P.M.Mathisen, B. Weinstock-Guttman, R.P. Kinkel, The epitope

spreading cascade during progression of experimental autoim-mune encephalomyelitis and multiple sclerosis, Immunol. Rev.164 (1998) 93–100.

[37] W.S. Begolka, C.L. Vanderlugt, S.M. Rahbe, S.D. Miller,Differential expression of inflammatory cytokines parallelsprogression of central nervous system pathology in twoclinically distinct models of multiple sclerosis, J. Immunol.161 (1998) 4437–4446.

[38] T.M. Pribyl, C.W. Campagnoni, K. Kampf, T. Kashima, V.W.Handley, J. McMahon, A.T. Campagnoni, The human myelinbasic protein gene is included within a 179-kilobase transcrip-tion unit: expression in the immune and central nervoussystems, Proc. Natl Acad. Sci. USA 90 (1993) 10695–10699.

[39] T.M. Pribyl, C. Campagnoni, K. Kampf, V.W. Handley, A.T.Campagnoni, The major myelin protein genes are expressed inthe human thymus, J. Neurosci. Res. 45 (1996) 812–819.

[40] R.B. Fritz, I. Kalvakolanu, Thymic expression of the golli-myelin basic protein gene in the SJL/J mouse, J. Neuroimmu-nol. 57 (1995) 93–99.

[41] N. Kawakami, S. Lassmann, Z. Li, F. Odoardi, T. Ritter, T.Ziemssen, W.E. Klinkert, J.W. Ellwart, M. Bradl, K. Krivacic, H.Lassmann, R.M. Ransohoff, H.D. Volk, H. Wekerle, C. Linington,A. Flügel, The activation status of neuroantigen-specific T cellsin the target organ determines the clinical outcome ofautoimmune encephalomyelitis, J. Exp. Med. 1999 (2004)185–197.

[42] A.L. Astier, G. Meiffren, S. Freeman, D.A. Hafler, Alterations inCD46-mediated Tr1 regulatory T cells in patients with multiplesclerosis, J. Clin. Invest. 116 (2006) 3252–3257.

[43] A.L. Zozulya, H. Wiendl, The role of regulatory T cells inmultiple sclerosis, Nat. Clin. Pract. Neurol. 4 (2008) 384–398.

[44] A. Ma, Z. Xiong, Y. Hu, S. Qi, L. Song, H. Dun, L. Zhang, D. Lou,P. Yang, Z. Zhao, X. Wang, D. Zhang, P. Daloze, H. Chen,Dysfunction of IL-10-producing type 1 regulatory T cells andCD4(+)CD25(+) regulatory T cells in a mimic model of humanmultiple sclerosis in Cynomolgus monkey, Int. Immunopharma-col. 9 (2009) 599–608.

[45] I.M. Stromnes, J.M. Goverman, Passive induction of experi-mental allergic encephalomyelitis, Nat. Prot. 1 (2006)1952–1960.

[46] G. Shi, C.A. Cox, B.P. Vistica, E.F. Wawrousek, I. Gery,Phenotype switching by inflammation-inducing polarized Th17cells, but not by Th1 cells, J. Immunol. 181 (2008) 7205–7213.