CD44high CD45RBlow CD4+ T cells that induce experimental allergic encephalomyelitis

Journal of Neuroimmunology, 40 (1992) 57-70

57

© 1992 Elsevier Science Publishers B.V. All rights reserved 0165-5728/92/$05.00 JNI 02223

Direct demonstration of the infiltration of murine central nervous system by P g p - 1 / C D 4 4 high CD45RB l°w CD4 + T cells that induce experimental allergic encephalomyelitis R a n a Zeine

a

and Trevor Owens a,b

Departments of Medicine, and b Department of Neurology and Neurosurgery, Montreal Neurological Institute, McGill Unicersity, Montreal, Quebec, Canada

(Received 22 January 1992) (Revised, received 26 March 1992) (Accepted 27 March 1992)

Key words: Central nervous system infiltration; Memory/effector T cells; Experimental allergic encephalomyelitis

Summary In experimental allergic encephalomyelitis (EAE), autoimmune T cells infiltrate the central nervous system (CNS) and initiate demyelinating pathology. We have used flow cytometry to directly analyse the migration to the CNS of MBP-reactive CD4 + T cells labelled with a lipophilic fluorescent dye (PKH2), in S J L / J mice with passively transferred EAE. Labelled cells constituted about 45% of the CNS CD4 ÷ population at the time of E A E onset. Almost all ( > 90%) of the PKH2-1abelled CD4 + T cells from E A E CNS were blasts and were a / ~ T cell receptor (TCR) +, CD44(Pgp-1) high, and the majority were CD45RW °w. By contrast, most PKH2-1abelled CD4 ÷ T cells in lymph nodes, although CD44 high, were CD45RB high cells. The cells that were transferred to induce E A E were essentially similar to antigenprimed lymph node cell populations, containing less than 15% CD44 high cells, and most of them were CD45RB high. The CD44 high CD45RB l°w phenotype is characteristic of m e m o r y / e f f e c t o r T cells that have been activated by antigen recognition. The difference in CD45RB expression between CNS and LN could therefore reflect differential exposure a n d / o r response to antigen. Consistent with this, PKH2labelled CD4 + cells isolated from the CNS were responsive to MBP in vitro, whereas PKH2 + CD4 ÷ cells from lymph nodes showed almost undetectable responses. In control experiments in which ovalbumin (OVA)-reactive T cells were transferred, a small number of fluorescent-labelled CD4 + T cells were also detected in CNS, but there were very few blasts, and these remained CD45RB high. These results argue for induction of the m e m o r y / e f f e c t o r phenotype of CD4 + T cells, and their selective retention in the CNS, as a consequence of antigen recognition.

Correspondence to: T. Owens, Neuroimmunology, Montreal Neurological Institute, 3801 University Street, Montreal, Quebec, Canada H3A 2B4.

5s Introduction

Experimental allergic encephalomyelitis (EAE) is an autoimmune disease of the central nervous system (CNS) which is induced by autoreaetive CD4 + T cells (Raine, 1985). The histopathology of demyelinated lesions and the overall progression of E A E are reminiscent of the human demyelinating disease multiple sclerosis (MS) (McFarlin and McFarland, lg82). The mechanism by which disease-inducing T cells enter the CNS is poorly understood, but the process is probably analogous to the infiltration of other tissues by inflammatory T cells. Studies using radiolabelled MBP-specific T cells have shown the accumulation of CD4* cells in the brains of mice in which EAE was induced by passive transfer (Naparstek et aI., 1982; Trotter and Steinman, 1984; Wekerle ct al., 1986, 1987). The distribution of myelin basic protein (MBP)-speeific T ceils into other organs was similar to that of T ceils with irrelevant specificities (Trotter and Steinman, 1984; Wekerle e t a [ . , 1986, 1987). Occasional radiolabelled ovalbumin (OVA)-specific T ceils were observed in the brains of non-diseased mice, but did not persist there (Wekerle et al., 1986, 1987). The association of neuroautoantigen-specific T cells with the CNS suggests a role for antigen recognition in their retention (Burns et al., 1984; Sgroi et at., 1986; Hailer et al., 1987). A greater understanding of the mechanism by which T cells infiltrate and are retained in the CNS is of both fundamental and clinical interest. One approach to this problem is to characterize infiltrating T cells with regard to their recognition of antigen and induction of specific cellular function. This can be accomplished by analysis of surface phenotype. Two T cell surface antigens whose level of expression correlates with states of activation are CD44 and CD45R. Memory or recall antigen-reactive T ceils are defined by their elevated expression of P g p - 1 / C D 4 4 (Butterfield et al., 1989). Similarly, the differential expression of CD45 isoforms defines stages of T cell activation. Resting or naive T cells express high molecular mass (MM) CD45R isoforms, and there is a gradual transition towards the exclusive expression of low MM CD45R isoforms as a consequence of anti-

gcn recognition. In the human this corresponds to CD45RA (MM 21()-220 kDa) to CD45RO (MM 170-180 kDa) transition, as defined by isoform-specific mAbs (Akbar ct al., 1988: Kristcnson et al., 199(t). In the mouse, monoclonal antibodies (mAbs) that recognize the high MM CD45RB isoform have been used to show that prolonged activation of CD4 ~ T cells in vitro leads to a reduction in the level of high MM CD45R expression (Birkeland et al., 1989). Murinc CD44 higl~ CD4.,~RB h,,~ T cells are potent cytokine producers (Budd et al., 1987; BottomIy ct al., 1989; Lee et a[., 1990; Swain et al., 199(I; Weinberg et al., 1990). This phenotype can thcrefl)re be considered a marker fl)r m e m o r y / c f f c c t o r 1" cells. The T cells in MS lesions and CSF, and isolated from the spinal cords of animals with EAE, have been shown to predominantly express low MM CD45R isoforms (Sobel ct al., 1988; Chofflon et al., 1989; Salonen et al., 1989: Jensen et al., 1990; Zaffaroni et al., 1990; Zeine and Owens, 1991), and T cells in other inflammatory- diseases are similarly biased towards this phenotypc (Potocnik et al., 1990). The accumulation of cells of the memory./effector phenotype at inflammatory, sites could in principle result from selection of that phenotype, or its induction as a consequence of antigen reactivity. Discrimination between these possibilities requires the ability to track infiltrating cells and to simultaneously analyse their antigen reactivity and surface phenotype. The former requires an adoptive transfer system, while the latter is not possible using radiolabelled cells. We have used a lipophilic fluorescent dye (PKH2) to monitor those T cells that migrate to the CNS and to peripheral lymph nodes (LN) in a passive transfer E A E model and to compare their expression of P g p - I / C D 4 4 and CD45RB. Our results show the selective accumulation in the CNS of autoreactivc CD4 + cells that are C D 4 4 high, CD45RB I'''. We find that although non-autoreactivc T cells enter the CNS, they are not retained there and do not express the memo r y / effector phenotype, These results argue for the preferential retention through antigen recognition of m e m o r y / effector T cells in the CNS following infiltration from the periphery-.

59

Materials and methods

neither did MBP-reactive T cells recognize OVA (not shown).

Mice

Isolation of mononuclear cells from CNS CNS infiltrating lymphocytes were collected by discontinuous density gradient centrifugation (Clatch et al., 1990) once EAE onset within the group had been verified. Mice were anaesthetized with chloral hydrate (3.5 g/kg) and perfused through the heart with PBS. The brains, spinal cords, and lymph nodes were then collected and dissociated by passing through a nylon or stainless steel mesh, respectively. The nervous tissue was centrifuged at 200 x g for 10 min and then resuspended in 4 ml of 70% Percoll (Pharmacia) in RPMI 1640 medium. This was then overlaid by equal volumes of 37% and 30% Percoll and the gradient was centrifuged at 500 x g for 15 min. Mononuclear cells were collected from the 37%:70% interface, washed in medium containing 10% FCS (ICN Biomedicals) and counted.

Female SJL/J mice (5-8 weeks) were obtained from Harlan-Sprague Dawley (Indianapolis, IN). Animal care and experimental protocols were approved by the McGill University Animal Care Committee, and were in accordance with CCAC guidelines.

EAE and PKH2-labelling of MBP-reactive T cells Donor mice were immunized s.c. with 400 Ixg of purified bovine MBP (Sigma, St. Louis, MO) in complete Freund's adjuvant (CFA) (50 ~g Mycobacterium tuberculosis H37RA (Difco, Detroit, MI) per mouse) and boosted 7 days later. Draining LN were collected at day 14, and LN cells (LNC) cultured at 4 × 10 6 cells/ml with MBP (50 ~ g / m l ) for 4 days. Responsiveness to MBP was assessed in parallel microcultures by [3H]thymidine incorporation at 4 days following an overnight pulse (0.5 ~Ci/well (ICN Biomedicals Inc., Mississauga, Ontario)). Stimulation indices of 10-30-fold were routinely observed. Cells were collected by centrifugation on Ficoll-Hypaque (Pharmacia, Montreal, Quebec) and either recultured for another 10 days in vitro with MBP and irradiated (3000 R) LNC as antigen-presenting cells, or labelled with PKH2-GL (Zynaxis, Malvern, PA) (Horan ad Slezak, 1989) and immediately injected i.v. at 10 7 blasts/mouse. Between 70 and 100% of animals developed clinical signs of EAE 10 days later. Mice were monitored daily and assigned clinical scores as follows: 0 (no symptoms), 1 (flaccid tail, clumsiness), 2 (moderate paresis), 3 (severe paresis or unilateral hind limb paralysis), 4 (bilateral hindlimb paralysis), 5 (moribund). Short-term OVA-reactive T cell lines were induced by in vivo priming and boosting with 400 #g OVA (Calbiochem, San Diego, CA) in CFA. LNC were cultured for 4 days with OVA (50 p~g/ml), labelled with PKH2 and transferred by i.v. injection exactly as for MBP-reactive T cells. There was no reactivity of OVA-reactive cells to MBP or to the MBP peptide FFKNIVTPRTPPP (Multiple Peptide Systems, San Diego, CA), corresponding to amino acids 90-102, which is encephalitogenic in SJL/J (Kono et al., 1988),

Flow cytometry mAbs included phycoerythrin-coupled antiCD4 (PE-CD4) (Becton-Dickinson, Mountain View, CA), anti-CD8 (53-6.7) (Ledbetter and Herzenberg, 1979), pan-anti-CD45 (M1/89) (Springer et al., 1978), anti-TcR,/~ (H57-597) (Kubo et al., 1989), anti-Pgp-1/CD44 (IM781) (Trowbridge et al., 1982), and anti-CD45RB (23G2) (Birkeland et al., 1988). Where indicated, mAbs were purified by Protein G-Sepharose affinity chromatography (Pharmacia) and coupled with biotin by incubation with biotinamidocaproate N-hydroxy succinimide ester (Sigma). Cells (5 x 105-106) were incubated with antibody at 4°C for 20 min and then washed and incubated with either FITC-coupled streptavidin (Bio-Can Scientific, Toronto, Ontario), FITC-sheep antimouse Ig (Bio-Can Scientific) or FITC-goat antirat Ig (Southern Biotechnology, Birmingham, AL) after blocking with rat Ig (100 /xg/ml) (Bio-Can Scientific). In some experiments biotinylated mAbs were visualized using Phycoerythrin :Texas Red (Tandem)-streptavidin (Southern Biotechnology). Surface staining was analysed using a FACScan (Becton Dickinson). Dead cells were excluded by propidium iodide staining, or by side

6() s c a t t e r gating. In s o m e e x p e r i m e n t s d e a d ceils w e r e e x c l u d e d by g a t i n g o u t FL3 t'
e q u a l l y b e t w e e n 6 ( L N C ) or 12 ( C N S ) r o u n d - b o t tomed microwells (Falcon). Preliminary counts i n d i c a t e d t h e r e to bc less t h a n 2 × 10 4 o f any w h o l e p o p u l a t i o n . I r r a d i a t e d (3(100 R) s y n g e n c i c s p l e e n cells t h a t h a d b e e n d e p l e t e d of T cells by i n c u b a t i o n with a n t i - T cell m A b s plus c o m p l e m e n t ( O w e n s , 1991) w e r e a d d e d (5 × 1 0 S / w e l l ) , with or without MBP or purified protein derivative ( P P D ) ( C e d a r l a n e , H o r n b y , O n t a r i o ) , b o t h at 5(t # g / m l . A s e n s i t i v e b i o a s s a y for IL-3 p r o d u c lion was u s e d to m e a s u r e T cell a c t i v a t i o n . A f t c r 2 days c u l t u r e , t h e IL-3 c o n t e n t of s u p e r n a t a n t s was m e a s u r e d in a m i c r o - b i o a s s a y in 20 #1 cult u r e s u s i n g t h e R 6 X - E 4 . 8 . 9 I L - 3 - d e p e n d e n t cell line ( O w e n s ct al., 1987). T h e n u m b e r o f viable cells was c o u n t e d u s i n g an i n v e r t e d m i c r o s c o p e a f t e r 24 h.

FACS P K H 2 - 1 a b e l l e d , M B P - r e a c t i v e T cells w e r e purified by F A C S . L N C a n d C N S m o n o n u c l e a r cells w e r e i s o l a t e d at t h e t i m e o f o n s e t o f p a s s i v e l y - i n duced EAE. Cells were stained with PE-CD4 and t h e n t h e P K H 2 +, C D 4 + p o p u l a t i o n was i s o l a t e d by F A C S ( F A C S t a r , B e c t o n - D i c k i n s o n ) .

Microbioassay for antigen specificity of" sorted cell populations B e t w e e n 10 4 a n d 105 cells w e r e o b t a i n e d w i t h i n e a c h g r o u p , a n d t h e t o t a l yield o f cells w e r e c u l t u r e d in R P M 1 1640 m e d i u m ( G i b c o / B R L , B u r l i n g t o n , O n t a r i o ) s u p p l e m e n t e d w i t h 10% F C S ( I C N B i o m e d i c a l s ) , 50 # M 2 - M E ( S i g m a ) a n d 2 m M L - g l u t a m i n e ( C a l b i o c h e m ) in 1-ml c u l t u r e wells ( F a l c o n , F i s h e r , M o n t r e a l , Q u e b e c ) w i t h IL-2 (10 U / m l ) , for 7 days. A f t e r 7 days, cells were collected, washed once and distributed

Results

Passit,e transfer of in l'itro labelled encephalitogenic T cells In o r d e r to d i r e c t l y e x a m i n e t h e m i g r a t i o n of T ceils to t h e C N S , w e passively t r a n s f e r r e d E A E w i t h M B P - r e a c t i v e T cells t h a t h a d b e e n f l u o r c s -

TABLE 1 ISOLATION OF CD4' T CELLS FROM THE CNS OF MICE FOLLOWING INTRAVENOUS TRANSFER OF PKH2LABELLED LNC EAE was induced as described in Materials and methods. Responsiveness of LN T cells to OVA and MBP were assessed betore transfer. Stimulation indices of 10 30-fold were routinely observed, with no reactivity to irrelevant antigens. CNS infiltrates were collected once EAE onset within the group had been verified, as described in Materials and methods. The number of mice pcr group is indicated in parentheses after the mean EAE score. For each experiment, the day on which symptoms were first observed and the day on which cells were collected are shown. The average number of mononuclear cells isolated per mouse and the percentage that were CD4 + was calculated from direct cell counts and FACS analysis. Treatment

Day of onset

Day of study

Mean EAE score

Number of CD4 + cells/mouse CNS ( × 111 x) PKH2 -

PKH2

Total

Normal

NA ~' NA 7 8 9 6 NA

NA NA 10 11 11 12 10

NA HA 2.1 (7) 2 (10) 1.2 (5) 0.5 (10) (1 (5)

NA NA 11.4 8.0 NC h 2.2 1.5

NA NA 15.1 17.3 NC 1.3 2.5

0.90 0.84 26.5 25.3 8.2 3.5 4.11

NA

10

0

NC

NC

8.4

EAE

OVA

(6)

~' Not applicable. b Not calculated, data not available. In these experiments, only CD4 + PKH2 + events were acquired and analysed. Their proportion within the total CD4 + population was not determined.

61

cently labelled in vitro before transfer. For passive transfer of E A E , T cells must be re-activated in vitro (Pettinelli and McFarlin, 1981). The in vitro manipulation of these T cells before labelling and transfer was limited to a brief culture with antigen, without any other addition to culture or cell fractionation. The lipophilic dye P K H 2 - G L labels all cells, and is retained in membranes over 2-week periods in culture or in vivo

A

10 3 PKH2 10 2

101

A ....

i f,'~'--~*

200 400 Forward

10 3

CD4

"'

,I,

I , ,'

, J ....

600 800 scatter

m

B

102

' ,

!4~ 0~

~

" .

~:~e~'~d:i:.;: ...... .. ,,(~{ .......... :

e,J

E ,.!

...~.~,'~:(:.':'..,'-._.~ :

e,.

i I

' :-.~,.~;~;?.~ !:.. ; ....

.

I

." ",,5~;.~ !".:'."--';'.; " " ' "

.:"..;)i~-~i:,£;!:: :

L

" :~'-;@~:,'':":

!

~'(:t4~~:"V"2:i:;'::-: : " ,

101

0 0 >

10 2

, m

10 3

ee ''''J''''l''''l'''']''''

PKH2 200 400 Forward B E c

tl

o

>

°m o.e

/

n-

101

J

'

L%

% 10 2

,,,

600 800 scatter

Fig. 2. Size and fluorescence of PKH2+-labelled CD4 + T cells from CNS and lymph nodes after passive transfer of EAE. MBP-reactive T cells were labelled in vitro with PKH2GL, and injected i.v. to induce E A E as described in Materials and methods. Following isolation at 10 days post-transfer, cells were stained with PE-CD4. PKH2 fluorescence was analysed in the FL1 channel. Dead cells were excluded by propidium iodide staining. A. CNS CD4 + cells, showing PKH2-fluorescence plotted against forward scatter. B. CD4 +, PKH2-fluorescent (FL1 gated as indicated in (A)) lymph node cells.

10 3

PKH2 Fig. 1. PKH2 fluorescence and CD4 expression by MBP-reactire cells that were transferred to induce EAE. MBP-reactive LNC were collected from culture, labelled with the lipophilic dye PKH2 and stained with PE-CD4 as described in Materials and methods. A. Input cells, showing CD4 expression plotted against PIQ-I2 fluorescence. B. Histogram of fluorescence before and after labelling with PKH2; ( ): labelled cells; ( . . . . . . ), unlabelled cells.

(Horan and Slezak, 1989). Calibration experiments showed that even after a 65-fold increase in cell number, the fluorescence level of T cells in culture was still distinguishable from that of controls, despite reduction of the per cell fluorescence intensity t h r o u g h proliferation. The transferred populations were all P K H 2 positive (Fig. 1), and contained about 55% CD4 + cells (Fig. 1A).

~2

Isolation of mononuclear cells .l}'om CNS

fluorescence. Figure 2A shows that the C D 4 T cells in the CNS consisted of two major subpopulations distinguishable by size and fluorescence. The majority of PKH2 cells were small, while most of the PKH2 + cells were blasts (Fig. 2A). By contrast, the majority ( > 75%) of fluorescentlabelled CD4* cells in LN were small cells (Fig. 2B). The fluorescence level in most PKH2 + CD4 ~ CNS T cells was reduced by an order of magnitude from that in thc input, transferred population (compare Fig. IA with Fig. 2A). Similar reductions in PKH2 levels occur in vitro as a consequence of proliferation, so this is consistent with proliferative activation of these cells in CNS. The PKH2 CD4 + T cells in the CNS represent endogenous]y derived migrants from the periphery. Our data show that these make up 40-70c)b of the CNS CD4 + population. This agrees with other estimates of endogenously recruited host T cells, obtained using [~4C]labelled transferred lymphocytes (Trotter and Steinman, 1984; Wekerle et aI., 1986, 1987; Cross et al.. 1990). Correction for the very small numbers of CD4 + T cells that were isolated from unprimed animals does not significantly affect this interpretation. We focussed on the fluorescent-labelled, transferred CD4 + cells in our analysis of the induction of effector phenotypes.

We isolated mononuclear cells by discontinuous density gradient centrifugation (Clatch et al., 1990) from the pooled brains and spinal cords of S J L / J female mice, within 2-3 days of verification of E A E within the group. About 105 mononuclear cells were recovered per mouse (mean 1.3 _+0.6 x 105), compared to less than 5 X 1(14 cells from normal animals (4.5 _+ 0.5 × 104). Most ( > 95%) of the isolated cells were stained with the pan-reactive anti-CD45 mAb M 1 / 8 9 (not shown), and were therefore leukocytes. The number of CD4 ~ cells was significantly higher in E A E compared to control CNS, and correlated well with the severity of disease (3.526.5 × 10 ~ in EAE, compared to 0.9 × 103 in normal mice) (Table 1). All of the CD4 + cells were CD3 +, a//3 TCR + T ceils (not shown). The CD4 CD8 cells were all CD3 . The proportions of CD4 + and CD8 + T cells in these CNS mononuclear populations isolated from early onset E A E are consistent with values from in situ immunopathological studies (Sriram et al., 1982; Traugott et al., 1985) and previous studies of isolated CNS cells (Clatch et al., 1990; Williamson et al., 1991). PKH2-1abelled CD4 + T cells were easily distinguishable from unlabelled cells by their green

i

A

I

;

B .Q

103

31

CD44 -

%!~'i",-'%?

-

102

'

'

! /

•

101

•

[

I "

I

I

200 400 600 800 Forward scatter

i

., o

2

-

-

-

;:G

• -

E

==

C I ]

-

l

3o C¢

200 400 600 800 Forward scatter

Lf I01

i i i I;,lll

102 CD44

103

Fig. 3. Pgp-1/CD44 expression on PKH2+-labelled CD4" cells after passive transfer of EAE. ('ells were isolated from CNS and lymph nodes after onset of EAE, and stained with PE-CD4. Biotinylated IM7.8.1 binding was visualized using Phycoerythrin : Texas red (Tandern)-streptavidin. Dead cells were excluded by side scatter gating. A. CNS. B. Lymph nodes. Profiles show CD44 plotted against forward scatter for PKH2-fluorescent (gated as in Fig. 2) CD4 + cells. C. CD44 distribution on CD4 + cells within the donor. EAE-inducing MBP-reactive population. Input cells were 100% PKH2-fluorescent.

63

C D 4 + T cells that migrated to C N S were C D 4 4 m~h C D 4 5 R B I°~

About 75% (78.3 +_4.2) of the transferred (PKH2 +) CD4 + T cells in the CNS were CD44 high (Fig. 3A). These cells differ from the original input population from which they derive both in their forward scatter profile, and in their CD44 distribution. The MBP-reactive population of CD4 + T ceils that were transferred to induce E A E were essentially similar to antigen-primed LNC populations, containing less than 15% CD44 high ceils (Fig. 3C). Very few (between 1 and

A

.Q

E

G) 0

~)

~) O: l

"i'ii11,I l

'

101

,

' ''''"I

, ,,,i, H

102 CD45RB

i

103

i

B

I I I

10 3

CD45RB

•: L i~ ,i~:i!'i,'~:L'~

10 2

101

•

:~i"/~'~:.I'" '

'

,

•

i

,

]

F

200

.

:

.

"

~

i=

,

,I.

~

i"1

r

l'l'l

T

i

[

,

,

]

,

i

r

z

I

i

400 600 800 Forward scatter

Fig. 4. CD45RB expression on PKH2+-labelled CD4 ÷ cells after passive transfer of EAE. Cell preparation, staining with PE-CD4 and 23G2 and fluorescence analysis was carried out as for other figures. In lymph nodes, about 25% of PKH2 + CD4 + cells were blasts (defined by forward scatter as for Fig. 1). Greater than 90% of CNS CD4 + PKH2 + cells were blasts. A. ( ): CNS CD4 + PKH2 + cells; ( . . . . . . ), lymph node CD4 + PKH2 + blasts. B. CD45RB expression and Forward light scatter by in vitro-cultured MBP-reactive CD4 + T cells.

,=,

200

-

_~ Z >

• IB

I

PKH2 + LNC PKH2+CNS

100

= c

~ --

0 Medium

MBP

PPD

Fig. 5. Antigen-specificity of populations sorted from CNS. EAE was induced by passive transfer with PKH2-1abelled, MBP-reactive T cells, and mononuclear cells isolated from the CNS at the time of onset. Cells were stained with PE-CD4 and PKH2 +, CD4 + cells were isolated by fluorescenceactivated cell sorting. 9 × 104 PKH2 + CD4 + lymph node cells and 3.5 × 104 PKH2 +, CD4 + CNS cells were cultured in IL-2 (10 U / m l ) , 1 ml/culture, for 7 days. After 7 days, cells were distributed equally between 6 (LNC) or 12 (CNS) round-bottomed microwells. Preliminary counts indicated there to be less than 2× 104 of any whole population. Irradiated (3000R) T-depleted syngeneic spleen cells were added (5 × ll)5/well), with or without MBP or PPD. After 2 days culture, the IL-3 content of supernatants was measured in a micro-bioassay as described in Materials and methods. The number of viable cells was counted after 24 h. The assay background was 7 + 2 cells, and 314 + 35 viable cells were counted in the presence of 1 U / m l IL-3.

4%) of LN CD4 + T cells were PKH2-1abelled. Of these, only a small proportion were either CD44 high (13.1% +- 2.9) or blasts (16.0% +_ 6.2) (Fig. 3B). Figure 4 shows that CNS CD4 + PKH2 + cells expressed significantly lower levels of CD45RB as a population than PKH2 + CD4 + blasts in LN. The majority of transferred, memory (CD44 high) CD4 + T cells that had infiltrated the CNS were in this way identified as activated effector cells. By contrast, most of the PKH2 + CD4 + blasts in LN were CD45RB high (Fig. 4A), showing a bias towards the activated effector phenotype in the CNS. The CD4 + LN T cells that were transferred did not contain as substantial a proportion of either large blasts or of CD44 high, CD45RB j''w cells (Fig. 4B) as were found in both CNS and LN 10 days post-transfer (see Fig. 3). PKH2-labelled C D 4 + cells in C N S were M B P - r e a c tiue

To verify that the effector CD4 + phenotype in CNS reflected the selective retention of autoreac-

tive T cells, we purified PKH2 + CD4 * cells from CNS and LN by FACS. The antigen specificity of these cells was measured following culture in IL-2 (Whitham et al., 1991). Between 10 4 and 1.5 × 105 cells were recovered after sorting. Because of the low number of cells recovered after sorting and 7 days in vitro culture, we used a sensitive bioassay for IL-3 to amplify responses to antigen. The R6X-E4.8.9 cell line dies quickly in the absence of IL-3, and both survives and proliferates in response to very small IL-3 titers. This microbioassay has been used for analysis of the activation of single, isolated T cells (Kelso and Owens, 1988) and was therefore best suited for the requirements of our experiment. In the experiment shown in Fig. 5, CNS CD4 + T cells responded to MBP. No response was detectable from LN PKH2-1abelled CD4 + cells, although these cells responded to PPD as strongly as did CNS CD4 ~ cells. In other experiments, using greater numbers of PKH2 + LNC per well, response to MBP was detectable, but CNS cells at comparable numbers generated at least five-fold stronger responses.

Phenotype of MBP-reactice LNC following prolonged s'timulation m citro Because of the remarkable contrast in the proportions of memory-effector cells found within the CNS and LNC populations 10 days after transfer, it was of interest to know the phenotype of LNC that had been re-stimulated in vitro for 10 days. We cultured MBP-reactive LNC with MBP in the presence of irradiated syngeneie LNC for a period of 10 days. The cells were then collected, centrifuged on Ficoll and double stained for CD4 and either CD44 or CD45RB. All of these in vitro activated cells were now CD44 high (not shown) and the majority were CD45RB b~' (Fig. 6). It was therefore possible to generate large numbers of mcmory-effectors upon prolonged stimulation of LNC with the appropriate antigen.

Migration of OVA-reactice T cells' to CNS The above results demonstrated the retention in CNS of m e m o r y / e f f e c t o r CD4 + cells that were reactive to antigen in the CNS. They also showed the presence of PPD-reactivc CD4 + T

A

E :3 ¢-

=

iI

m

re -

101

102 CD45RB

103

101

° .

102

103

CD45RB

Fig. 6. CD45RB expression by MBP-reactive C D 4 ~ LNC following in vitro culture for 4 days (A), and for a further 10 days (B). ~ Lymph node cells were isolated from mice that had been primed and boosted with M B P / C F A . The cells were cultured with MBP for 4 days and then collected, centrifuged on Ficoll, ad recultured for a further 10 days with added MBP and irradiated LNC, after which they were again centrifuged on Ficoll and double-stained with PE-CD4 and 23G2 supernatant. 23G2 was visualized using FITC-goat-anti-rat Ig. Dead cells were excluded by side scatter gating.

65

10 3` :

.

CD45RB

..

•

.-,

.. : .,:.: ~; }: (vv.

10 2-

.

=-".: ." :'..:,~r:~.,:,:~:.'.L~,--%d<:.:.,.: '.... :. ?i...y~.]:~i~,i~j'~.~,:~:::~: .1...

.... +

+,

...:;~!,z:,.:.2h~;,:::.".'.~:: ..-

101

Lymph Node

t

"

200

400

600

Discussion

800

Forward scatter

103] CD45RB

PKH2 + CD4 + cells that entered the CNS in O V A transfers, nor was their CD45RB expression reduced (Fig. 7). These data argue against the possibility that T cells with the CD45RB ~°w phenotype were selectively retained in CNS, and suggest instead that their retention was a consequence of antigen recognition.

1

CNS

10 2

..

,-:" • : .

i

::4d," -'

,o,? ',1

:flf . . . .

i

200

. . . .

i

400

,

~

i

i

i

. . . .

600

i

. . . .

800

Forward scatter Fig. 7. C D 4 5 R B expression by PKH2-labelled OVA-reactive C D 4 + cells in CNS. Cells were isolated from CNS and lymph nodes 10 days after transfer of PKH2-1abelled OVA-reactive cells, and C D 4 + cells identified with PE-CD4. Biotinylated 23G2 binding was visualized using P h y c o e r y t h r i n : T e x a s Red (Tandem)-streptavidin. D e a d cells were excluded by side scatter gating.

cells. To further examine the role of CNS-antigen-reactivity we transferred PKH2-1abelled OVA-reactive T cells into S J L / J mice. The mice did not develop EAE. The absolute number of CD4 + T cells recovered from the CNS was about 10-fold less than that obtained from animals with E A E (Table 1). 26% of CD4 + T cells from CNS in the O V A transfer were PKH2 + , but only 6% were blasts (not shown). The proportion of LN CD4 + T cells that were PKH2-1abelled was similar in O V A transfers to that in MBP transfers. However, in O V A transfers there was a greater accumulation of PKH2 + blasts in lymph nodes (about 30%) (Fig. 7). These included cells showing reduced levels of expression of CD45RB (Fig. 7). These were almost no blasts among the

We have been able to track infiltrating CD4 + T cells from the periphery into the CNS in a passive transfer E A E model. By using a fluorescent dye to label cells, we could examine the phenotype of the CD4 + T cells that accumulated in the CNS. Our results show that the CNS-infiltrating CD4 + T cells in E A E are of an activated or m e m o r y / e f f e c t o r phenotype that is relatively poorly represented in LN or in the in vitro-activated autoreactive populations that transfer EAE. The accumulation of cells of a defined phenotype in a organ or tissue could result from selective migration, or selective retention. The fact that PPD-reactive CD4 + T cells are found in the CNS argues against there being an absolute prohibition of entry of non-CNS-reactive T cells. This is consistent with other reports of entry of non-CNS-reactive T cells to the CNS (Trotter and Steinman, 1984; Wekerle et al., 1986, 1987), and with the lack of antigen specificity of either the expression or function of adhesion molecules that are implicated in the interaction of T cells with endothelia (Picker et al., 1991; Shimizu et al., 1991). However, the number of CD4 + T cells in the CNS was very much less following the transfer of OVA-reactive LNC than that of MBP-reactive T cells. This points to a role for CNS-antigen reactivity in the accumulation of CD4 + T cells in the CNS, and this could result from one of two processes, which are not mutually exclusive. Either MBP-reactive T cells are favored for entry to the CNS, or they are preferentially retained without there being any specific barrier to T cell entry. A role for antigen recognition is strongly favoured by the dramatic bias among infiltrating MBP-reactive cells towards the CD44 high CD45RW °w phenotype. The homogeneous expression of high CD44 levels by CD4 +

MBP-reactive cells in the CNS contrasts with activated LN populations where at most 15% of CD4 + T cells are CD44 "igh (Fig. 3). Recall antigen reactivity is associated with this phenotype (Butterfield et al., 1989), and prolonged activation of T cell populations in vitro (e.g. to generate cell lines) induces homogeneously high levels of CD44. This argues that thc CNS-infiltrating CD4 + T cells were all activated. Because CD45RB big" cells were also found in the CNS, the CD45RB I°'~ phenotype cannot bc a requirement for extravasation. The fact that C D 4 5 R B " ' ' blasts were not found in the CNS following transfer of OVA-reactive cells but were present in LN shows that cells with this phenotype are not committed to migrate into the CNS. The CD45RB "'w phenotype is induced in vitro by activation through T C R / C D 3 ligation (Birkeland et al., 1989). The relative absence of blasts and the high level of CD45RB expression among the few CNS infiltrating T cells in O V A transfers, in contrast to the CD45RB h'' blasts in MBP-induced E A E argues that retention of MBP-reactive T cells in CNS was a consequence of their activation through antigen recognition. This conclusion is further supported by our data showing the generation of CD45RB >w cells by antigenic stimulation for 10 days in vitro. The remarkable correlation between sequestration and the effector phenotype seems to us compatible with the following model for the induction of E A E in an antigen-primed mouse. T cell activation in primed LN induces a m e m o r y / effector (CD44 high CD45RB I°~') phenotype, which predisposes T cells to migrate from LN and to infiltrate tissues (Mackay et al., 1990; Mackay, 1990). Antigen-specificity may influence entry to certain tissues but the effect is not absolute. The major influence on accumulation is at the level of retention as a consequence of antigen recognition. The selective accumulation in the CNS of CD45RB ~'''', CD4 + T cells then reflects the secondary activation of CNS-reactive cells through antigen recognition. Whether non-CNS-reactivc cells that have co-infiltrated are also induced to express the CD45RB l°w phenotype (e.g. through bystander cytokine activation) remains to be determincd. For instance, we do not know the CD45RB phenotype of the PPD-reactive CD4 + T

cclls dcscribed in Fig. 5. Secondary activation in thc CNS may be accompanied by other phenotypic changes that lead to T cell retention in thc tissue. Prolonged activation of T cells in vitro induces down-regulation of CD45RB expression (Fig. 6; Birkcland et al., 1989). Our data arc consistent with proliferative activation in the CNS that induces the CD45RB I°'~ phenotypc. Activated T cells have been shown to preferentially traffic in the blood (Mackay et al., 1990), and this accounts in part for the dominance of the CD45RB high phenotype even in primed LN populations. The significance of our observations is further enhanced by the fact that these phenotype changes are dynamic. Reverse transitions from low to high MW CD45R expression can occur and have been described for CD45RB in rat (Bell et al., 1991)) and for CD45RO to CD45RA in human T cell lines (Lasalle and Hafler, 1991; Rothstein et al., 1991), Such reverse transitions must contribute to the heterogeneity of LN populations. In this regard, the near-homogeneity of CD45R expression by MBP-reactive CD4 ~ T cells further supports our view that these cells are actively responding to antigen. In fact, it was possible to generate homogeneous populations of CD45RB l°' cells following prolonged in vitro culture of LNC although freshly isolated cells from LN were always heterogeneous. The implication would be that cells with the memory-effector phenotype that are generated within the LN migrate to the CNS where they accumulate within that antigenically rich environment. The CD44 high, CD45RB "~w phenotype has been implicated as a m e m o r y / e f f e c t o r subset largely on the basis of in vitro analysis. That this phenotype is predominantly represented at a site of autoimmune pathology supports its designation as an effector T cell subset. In a separate study, we have shown that the T cells in the CNS of animals with E A E induced by immunization with spinal cord in adjuvant were also of a m e m o r y / effector phenotype. The frequency of cytokinesecreting CD4 + T cells in perivascular infiltrates in the brains of those mice were 10-fold higher than in primed LN or in in vitro-activated T cell populations (Renno et al., submitted). This is entirely consistent with the results presented here.

67

By studying the migration of PKH2-1abelled antigen-primed T cells, we have been able to assess the dynamics of effector selection and retention in vivo, with a minimum of perturbation of the cells being analysed. Our experimental approach has therefore allowed us to identify activated effectors within the CNS, and by use of cell sorted fluorescent-labelled populations from the CNS, to ascertain their antigen specificity. These results confirm a role for antigen recognition in the CNS in EAE, and point to its mediation of effector cell retention at this site.

Acknowledgements We thank Diane Heath for technical assistance and preparation of reagents, and Claude Cantin (IRCM, Montreal) for FACS. We are grateful to Dr. R. Kubo (Denver) for provision of the H57597 hybridoma and to Dr. Michael Julius (McGill University) for M1/89 antibody. We thank Dr. K. Bottomly (Yale University) for generously providing purified anti-CD45RB mAb (16A) which was used for the initiation of these studies. We thank Drs. Michael Ratcliffe, Neil Cashman, and Jack Antel for discussions throughout this study, and for review of the manuscript. This work was supported by the Multiple Sclerosis Society of Canada, and by personal support awards from MRC-Canada (T.O.) and Fonds pour la Formation de Chercheurs et l'Aide ?t Ia Recherche (FCAR)-Qudbec and the Multiple Sclerosis Society of Canada (R.Z.).

References Akbar, A.N., Terry, L., Timms, A., Beverley, P.C.L. and Janossy, G. (1988) Loss of CD45R and gain of UCHL1 reactivity is a feature of primed T cells. J. Immunol. 140, 2171-2178. Bell, E.B. and Sparshott, S.M. (1990) Interconversion of CD45R subsets of CD4 T cells in vivo. Nature 348, 163165. Birkeland, M.L., Metlay, J., Saunders, V., Fernandez-Botran, R., Vitetta, E.S., Steinman, R.M. and Pure, E. (1988) Epitope on CD45R (T200) molecules define differentiation antigens on murine B and T lymphocytes. J. Mol. Cell. Immuno]. 4, 71-85. Birkeland, M.L., Johnson, P., Trowbridge, I.S. and Pur6, E.

(1989) Changes in CD45 isoform expression accompany antigen-induced murine T-cell activation. Proc. Natl. Acad. Sci. USA 86, 6734-6738. Bottomly, K., Luqman, M., Greenbaum, L., Carding, S., West, J., Pasqualini, T. and Murphy, D.B. (1989) A monoclonal antibody to murine CD45R distinguishes CD4 T cell populations that produce different cytokines. Eur. J. Immunol. 19, 617-623. Budd, R.C., Cerottini, J.-C. and MacDonald, H.R. (1987) Selectively increased production of interferon-y by subsets of Lyt-2 + and L3T4 + T cells identified by expression of Pgp-1. J. Immunol. 138, 3583-3586. Burns, J., Rosenzweig, A., Zweiman, B., Moskovitz, A. and Lisak, R. (1984) Recovery of myelin basic protein reactive T cells from spinal cords of Lewis rats with autoimmune encephalomyelitis. J. Immunol. 132, 2690-2692. Butterfield, K., Fathman, C.G. and Budd, R.C. (1989) A subset of memory CD4 + helper T lymphocytes identified by expression of Pgp-1. J. Exp. Med. 169, 1461-1466. Chofilon, M., Weiner, H.L., Morimoto, C. and Hailer, D.A. (1989) Decrease of suppressor inducer (CD4 + 2 H 4 + ) T cells in multiple sclerosis cerebrospinal fluid. Ann. Neurol. 25, 494-499. Clatch, R.J., Miller, S.D., Metzner, R., Dal Canto, M.C. and Lipton, H.L. (1990) Monocytes/macrophages isolated from the mouse central nervous system contain infectious Theiler's Murine Encephalomyelitis Virus (TMEV). Virology 176, 244-254. Cross, A.H., Cannella, B., Brosnan, C.F. and Raine, C.S. (1990) Homing to central nervous system vasculature by antigen-specific lymphocytes. I. Localization of 14C-labeled cells during acute, chronic, and relapsing experimental allergic encephalomyelitis. Lab. Invest. 63, 162-170. Hailer, D.A., Benjamin, D.S., Burks, J. and Weiner, H. (1987) Myelin basic protein and proteolipid protein reactivity of brain and cerebrospinal fluid-derived T cell clones in multiple sclerosis and postinfectious encephalomyelitis. J. Immunol. 139, 68-72. Horan, P.K. and Slezak, S.E. (1989) Stable cell membrane labelling. Nature 340, 167-168. Jensen, M.A., Noronha, A., Toscas, A. and Arnason, B.G.W. (1990) Activated T cells in the spinal cord in EAE differ from activated T cells in draining nodes (Abstract). Neurology 40 (Suppl. 1), 394. Kelso, A. and Owens, T. (1988) Production of two hemopoieric growth factors is differentially regulated in single T lymphocytes activated with an anti-T cell receptor antibody. J. Immunol. 140, 1159-1167. Kono, D.H., Urban, J.L., Hovrath, S.J., Ando, D.G., Saavedra, R.A. and Hood, L. (1988) Two minor determinants of myelin basic protein induce experimental allergic encephalomyelitis in SJL/J mice. J. Exp. Med. 168, 213-227. Kristensson, K., Dohlsten, M., Fischer, H., Ericsson, P.O., Hedlund, G., Sjogren, H.-O. and Carlsson, R. (1990) Phenotypical and functional differentiation of CD4 + CD45 RA + human T cells following polyclonal activation. Scand. J. lmmunol. 32, 243-253. Kubo, R.T., Born, W., Kappler, J.W., Marrack, P. and Pigeon,

68 M. (I 989)Characterization of a monochmal antibody which detects all murine r r / ~ T cell receptors. J. lmmunol. 142. 2736 2742. LaSallc, J. and Hailer, D.A. (1001) The coexprcssion of C D 4 5 R A and C D 4 5 R O isoforms on T cells during the S / G _ , / M stages of cell cycle. Cell. lmmunol. 138, 197 208. Ledbctter, J.A. and Herzenberg, L.A. (19791 Xenogcneic monoclonal antibodies to mouse lymphoid differentiation antigens, lmmunol. Rev. 47, 63 g0. Lee, W.T., Yin, X.-M. and Vitetta, E.S. (1990) Functional and ontogenic analysis of murine CD45R hi and CD45 I'' C D 4 T cells. J. Immunol. 144, 3288 3205. Mackay, C.R. (1990) T cell memory: the connection between function, phenotype and migration pathways, lmmunol. Today 12, 189-192. Mackay. C.R., Marston, W.L. and Dudler, L. (199l)) Naive and memory T cells show distinct pathways of lymphocyte recirculation. J. Exp. Med. 171,801-817. McFarlin, D.E. and McFarhmd, tt.F. (1982) Multiple Sclerosis, N. Engh J. Med. 307, 1183-1188. Naparstek, Y., Holoshitz, J., Eisenstein, S.. Rcshcf, T., Rappaport, S., Chemke, J., Ben-Nun, A. and Cohen. I.R. (1982) Effector T lymphocyte line cells migrate to the thymus and persist there. Nature 300, 262-264. Owens, T. (1991) A role for adhesion molecules in contact-dependent T help for B cells. Eur. J. lmmunol. 21,979-983. Owens, T. and Fazekas de St. Groth, B. (1087) Participation of L3T4 in T cell actiwltion in the absence of class ll MHC: inhibition by anti-L3T4 antibodies is a function both of epitope density and mode of presentation of anti-receptor antibody. J. Immunol. 138, 2402 2409. Pettinelli, C.B. and McFarlin, D.E. (19811 Adoptive transfer of experimental allergic encephalomyelitis in S J L / J mice after in vitro activation of lymph node cells by myelin basic protein: Requirement for Lyt I + 2 T lymphocytes. J. lmmunol. 127, 1420-1423. Picker, L.J., Kishimoto, T.K., Smith, C.W., Warnock, R.A. and Butcher, E.C. (19911ELAM-1 is an adhesion molecule for skin-homing T cells. Nature 349, 796-799. Potocnik, A.J., Kinnc. R., Menninger, K., Zacher, J.. Emmrich, F. and Kroczek, R.A. (1990) Expression of activation antigens on T cells in Rheumatoid Arthritis Patients. Scand. J. Immunol. 31,213 224. Raine, C.S. (1985) Experimental allergic encephalomyelitis. In: J.C. Koetsier (Ed.), Handbook of Clinical Neurology, Vol. 3(47), Elsevier, Amsterdam, pp. 429 466. Rothstein, D.M., Yamada, A., Schlossman, S.F. and Morimoto, C. (1991) Cyclic regulation of CD45 isoform expression in a long term h u m a n CD4 " C D 4 5 R A ~ T cell line. J. lmmunol. 146, 1175-1183. Salonen, R.. llonen, J., Jfigerroos, K., Syrjfilh. H., Nurmi, T. and Reunaen, M. (1989) Lymphocyte subsets in the cerebrospinal fluid in active mtdtiple sclerosis, Ann. Neurol. 25, 500-502. Sedgwick, J.D., M6ssner, R.. Schwender, S. and ter Meulen, V. (1991) Major histocompatibility complex-expressing

nonhematopoietic astroglial cells prime only CD8 * T lymphocytes: astroglial cells as perpetuators but not initiators of CD4 ' T cell responses in the central nervous system. J. Kxp. Mcd. 173, 1235 1246. Sgroi, D., (?(then, R.N., kingcnheld, E.G., Strong, M.K., Binder, T., Ooldschneider, 1., Orcincr, D., OrnnneI, M. and Clark, R.B. (1086)T cell lines derived from the spinal cords of mice with Experimental Allcrgic Encephalomycli tis are self reactive. J. lmmunol. 137, i850 1854. Shimizu, Y., Shaw, S., Garber, N., Gopal. T.V., l torgan, K.,I., Van Seventer, G.A. and Newman, W. (1991) Activation-independent binding of human m c m o u T cells to adhesion molecule ELAM-I. Nature 349,709 802. Sobel, R.A., Hafler, D.A., Castro, E.A., Morimoto, C. and Weiner. H . L (lt188) The 2tt4 (CD45R) antigen is selectively decreased in multiple sclerosis leskms. ,I, Immunol. 140, 2210-2213. Springer, T.. Galfre, G., Secher, D.S. and Milstein, C. (1978) Monoclomd xenogeneic antibodies to murine cell surface antigens: identification of novel leukocyte differentiation antigens. Eur. J. Immunol. 8,539 551. Sriram, S., Solomon, D., Rouse, R.V. and Steinman, L. (1982) Identification of T cell subsets and B lymphocytes in mouse brain experimental allergic encephalitis lesions. J. lmmunol. 129, 1649 1651. Swain, S.L., Weinbe,g, A.D. and English, M. (19901 CD4 ~ T cell subsets: Lymphokine secretion of memory cells and of cffector cells that develop from precursors in vitro. J. Immunol. 144, 1788 1799. Traugott, U., Raine, C.S. and McFarlin, D.E. (1985) Acute experimental allergic encephalomyelitis in the mouse: lmmunopathology of the developing lesion. Cell. lmmunol, 91,240 254. Trotter, J. and Steinman, L. (1984) ltoming of Lyt-2 + and Lyt 2 T cell subsets and B lymphocytes to the central nervous system of mice with acute experimental allergic encephalomyelitis. J. lmmunol. 132, 2919-2023. Trowbridge, I.S., Lesley, J., Schulte, R., Hyman, R. and Trotter, J. (19821 Biochemical characterization and cellular distribution of a polymorphic, murine cell-surface glycoprotein expressed on lymphoid cclls. Immunogenetics 15. 229 312. Weinberg, A.D., English, M. and Swain, S.L. (19901 Distinct regulation of lymphokinc production is found in fresh versus in vitro primed murine helper T cells. J. lmmunol. 144, 1800 1807. Wekerle, H., Linington, C.. Lassmann, H. and Meyermann, R. (19861 Cellular immune reactivity within the CNS. Trends Neurosci. 9, 271-277. Wekerle, H., Sun, D., Opereza-Wekerle, R.L. and Meyermann, R. (1987) I m m u n e reactivity in the nervous system: modulation of T-lymphocyte activation by glial cells. J. Exp. Biol. 132, 43-57. Whitham. R.H., Bourdette, D.N., Hashim, G,A., Herndon, R.M., fig, R.C., Vandenbark, A.A. and Offner, H. (19911 Lymphocytes from S J E / J mice immunized with spinal

69 cord respond selectively to a peptide of proteolipid protein and transfer relapsing demyelinating experimental autoimmune encephalomyelitis. J. Immunol. 146, 101-107. Williamson, J.S.P., Sykes, K.C. and Stahlman, S.A. (1991) Characterization of brain-infiltrating mononuclear cells during infection with mouse hepatitis virus strain JHM. J. Neuroimmunol. 32, 199-207.

Zaffaroni, M., Rossini, S., Ghezzi, A., Parma, R. and Cazzulo, C.L. (1990) Decrease of CD4 + CD45 + T-cells in chronic progressive multiple sclerosis. J. Neurol. 237, 1-4. Zeine, R. and Owens, T. (1991) Selective migration of CD44 hi CD45RB L°wCD4 + T cells into the central nervous system in Experimental Allergic Encephalomyelitis. (Abstract) Ann. Neurol. 30, 269.