Eur. J. Immunol. 1990. 20: 1265-1272

Michel Seman, Sarah Boudaly, Thierry Roger, Jean Morisset and Ghislaine Pham Laboratoire d’Immunodifferentiation, Institut Jacques Monod, Pans

CD4-CDS-autoreactive Tcells

Autoreactive T cells in normal mice: unrestricted recognition of self peptides on dendritic cell I-A molecules by CDLC-CDS- T cell receptor alp+ T cell clones expressing Vp8.1 gene segments* CD4-CD8- double-negative (DN) and CD4+CD8- T cell clones were derived from splenic precursors resistent to killing by anti-Thy-1, -CD5, -CD4 and -CD8 monoclonal antibodies and complement. Both DN and CD4+ clones express functional T cell receptor (TcR) a/@and exhibit strong autoreactivity in vitro. DN cells can be induced to proliferate by dendritic cells (DC) of all haplotypes tested, although this activation is inhibited by antibodies specific for I-A determinants expressed on the stimulatory DC. In contrast, CD4+ clones only respond to syngeneic or I-Ad-compatible DC. Both DN and CD4+ autoreactive clones do not proliferate when cultured with class 11+H-2d normal or tumor macrophages and B cell lines or with class 11-transfected L cells, suggesting that these cells recognize self peptides only present on the surface of DC. Despite their phenotype resembling that of immature thymocytes and their inability to interact directly with B lymphocytes, DN cloned T cells, like CD4+ T cells, exhibit nonspecific helper functions and can induce polyclonal B cell proliferation and differentiation. DN TcR a/@+peripheral T cells represent, like TcR y/6+ lymphocytes, a new Tcell subset physiological role whose remains to be defined.

1 Introduction The T cell repertoire mostly comprises CD4-CD8+ and CD4+CD8- TcR a/@-bearing lymphocytes recognizing antigenic peptides on self MHC class I or class I1 molecules, respectively, a phenomenon often referred to as MHC restriction [l]. This restriction is the consequence of positive selection during T cell maturation in the thymus [2, 31. Negative selection also occurs in this organ to eliminate self-reactive T lymphocytes [3,4]. However, autoreactive T cells have been described in different experimental situations in peripheral lymphoid tissues from normal individuals [5-71 .These cells proliferate in syngeneic or autologous MLR but their true specificity remains elusive. Like alloreactive T cells, they have often been considered as specific for MHC class I1 determinants [8, 91. This interpretation was supported by evidence suggesting that TcR V gene segments generate a repertoire biased to combine with MHC products [lo, 111 and that self tolerance is acquired, in part, by deletion of self-reactive clones using particular TcR V genes [ l l , 121. However, recent findings indicate that negative selection of putative selfreactive thymocytes involves self peptides and MHC recognition [13, 141.

[I 75921

*

1265

Interestingly, experiments in mice transgenic for an alloreactive TcRcr/P or a TcRa/P-specific for the male H-Y antigen have revealed the accumulation of CD4-CD8cells bearing the transgenic TcR in peripheral lymphoid tissues when expression of this TcR is negatively or not positively selected in the thymus [3,4, 15].The significance of this phenomenon is poorly understood. It suggests that unselected CD4-CD8- double negative (DN) T cells can migrate out of this organ. Whether these cells can maturate in the periphery is not known. Consistent with these observations, we and others have shown that splenocytes from normal mice depleted of mature CD4+ or CD8+ by treatment with appropriate antibodies and complement contain a population of autoreactiveTcel1 precursors which can be activated in vitro [16-181. These precursors seem to have a CD4-CD8- phenotype but some of them, at least, can give rise to CD4+CD8- functional autoreactive T cells. The present report documents CD4-CD8- TcR a/P-bearing autoreactive clones derived from these precursors. These clones, in contrast to the CD4+ cells derived from the same initial population, exhibit an unusual specificity resembling that of ylb-bearing peripheral DN T cells [19, 201. Both DN and CD4+ clones seem to recognize a self antigen distinct from Mls-1 and Mls-2 gene products [21,22]. These cells and their precursors might represent physiological counterparts of DN cells found in MRLlprllpr autoimmune mice and described in TcR a/@transgenic mice.

This work was supported by a grant from the Institut National de la SantC et de la Recherche Medicale.

Correspondence: Michel Seman, Institut Jacques Monod, CNRSUniversite Paris VII, 2 Place Jussieu, F-75251 Paris Cedex 05,. France Abbreviations: DC: Dendritic cell DN: Double-negative (CD4-CD8-) T lymphocytes NMS: Normal mouse serum 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1990

2 Materials and methods 2.1 Mice

DBA/2, B6, BALB.B, BALBlc and BALB.K mice were bred in the animal quarter of the Institut Jacques Monod 0014-2980/90/0606-1265$02.50/0

1266

Eur. J. Immunol. 1990. 20: 1265-1272

M. Seman, S. Boudaly, T. Roger et al.

under standard conditions. All other strains used were kindly suppliedfrom Dr. M. Pla (St. Louis Hospital, Paris). Eight to twelve-week-old males and females were used in most experiments.

2.2 mAb The cocktail of anti-Thy-1, anti-CD5 (Ly-1), anti-CD4 (L3T4) and anti-CD8 (Ly-2) mAb used t o eliminate mature splenicT lymphocytes was prepared as described elsewhere [16, 181. KJ16 mAb recognizingTcR Vp8.1 and Vp8.2 gene products [22], MKD6 (anti-I-Ad) and MKS4 (anti-I-AS) were gifts from Dr. P. Marrack (Denver, CO). 34-5-35 (anti-I-Ad.b,p,q), 26-7-118 (anti-I-Ak) and 14-4-4s (antiI-Ek) mAb were purchased from Bionetics (Kensington, MD). F23.2 mAb specific for V08.2 gene products [22] was a gift from Dr. A. Coutinho (Institut Pasteur, Paris, France).

tions were evaluated by measuring the ability of cloned T cells (0.1 x 104-10 x 104/well) to induce B cell proliferation. Splenic B lymphocytes (3 x 105/well)were prepared by cytotoxic treatment with the anti-Tcell mAb cocktail and rabbit C and cultured with irradiated (2500 rad) Tcells in Click's medium containing 1% NMS in 0.2-ml microculture wells. Proliferation was measured after 4 days upon addition of 37 kBq/well of ["IdThd.

2.6 Reverse plaque assay IgM-secreting cells induced by irradiated cloned T cells were detected after 4 days of culture in RPMI 1640, 5% FCS using a reverse PFC assay as already described [16]. IgM-secreting cells were detected using rabbit anti-mouse IgM antibodies (Miles Labs., Naperville, IL). LPS (Difco, Detroit, MI)-activated B cells were used as internal positive controls in all experiments.

2.3 Analysis of cells by FCM

2.7 Northern blot analysis

Clone phenotypes were established by indirect fluorescence using rat anti-Thy-1 (clone 30H12, Becton Dickinson, Mountain View, CA), antLCD5 (clone 53-7-3, Becton Dickinson), anti-CD8 (ADH4/15), anti-CD4 (clone GK1.5 from Dr. €? Marrack, Denver, CO) and anti-Vp8 (KJ16) mAb as previously described [16, 181. CD3 phenotypes were determined by direct fluorescence using FITC-conjugated 145-2C11 hamster mAb [23].

RNA was extracted from cloned cells by the guanidinium isothiocyanate and CsCl gradient centrifugation procedures [25]. Total RNA was electrophoresed through 1% agarose-formaldehyde in 10 mM 4-morpholinepropanesulfonic acid (Mops) and transferred to nitrocellulose. Blots were hybridized to 32P-labeled nick-translated C,, Cp and Vp8 probes [26-281. Following hybridization in 5 x SSC, 5 x Denhardt's, 10% dextran sulfate and 0.1% SDS at 65 "C, blots were washed three times in 1 x SSC, 0.1% SDS at room temperature and once in 0.05 x SSC, 0.1% SDS at 55 "C.

2.4 Cell cloning After two cycles of treatment with a cocktail of anti-Thy-1, -CD5, -CD4 and -CD8 mAb, and rabbit C [16], spleen cells from normal DBA/2 mice were cultured for 5 days with 2.5 pg/ml Con A in Click's medium containing 1% normal mouse serum (NMS). Living cells were then recovered by sedimentation over Ficoll gradient (Lymphoprep, Nyegaard, Norway) and reseeded in Click's medium containing 5% FCS in 25-cm2 tissue culture flasks maintained in an upright position and stimulated by syngeneic irradiated (2500 rad) dendritic cell (DC)-enriched spleen cells prepared as described elsewhere [24]. T cells (1 x 106-2 x lo6 in 5 ml) were then reseeded every week in the same medium with 1 x 106syngeneicDC. After3 months, the cell line was cloned by LD (0.3 cell/well) in 0.2-ml cultures (Falcon 3072 microculture plates, Oxnard, CA) containing Click's medium, 5% heat-inactivated FCS and 1 X lo5 syngeneic DC. Clones were further expanded in flasks (1 X lo6 Tcells in 5 ml) with 5 x lo6 syngeneic irradiated DC. No rIL2 or conditioned medium was ever added.

2.5 Proliferation assay To evaluate clone specificity, 2 x lo4 cells were cultured in Click's medium containing 1% NMS with syngeneic DCenriched irradiated spleen cells in 0.2-ml wells. Proliferation was estimated after 72 h by addition of 37 kBq of [3H]dThd (28 Ci/mmol, CEA, Saclay, France). Cells were harvested 16-18 h after addition of ["H]dThd and radioactivity incorporated in DNA was counted. Helper func-

3 Results 3.1 Characterization of CD4-CDS- T cell clones DN T cells were prepared from the spleen of normal DBA/2 mice by two cycles of treatment with a cocktail of anti-Thy-1, -CD5, -CD4 and -CD8 mAb to eliminate mature functional T cells [16]. T cell depletion was controlled by FCM. Remaining lymphocytes were then activated by Con A for the first 5 days and subsequently expanded in the sole presence of DC-enriched syngeneic spleen cells as stimulators. A T cell line (TKR101) was derived and then cloned by LD. Twenty-one clones were obtained and their phenotypes determined by immunofluorescence. Among these clones, 16 were CD3+CD4+CD8- as illustrated in Fig. 1for clone 3D4.This phenotype corresponded to that observed on most cells derived from CD4KD8-depleted splenic populations and activated by syngeneic DC [16-181. The five other clones (1G3, 3B5, 1F8, 3D10 and 3D12), although developed under the same conditions, were Thy-l+CD3+ but CD5-CD4-CD8- (see clones 3B5 and 1F8 in Fig. la). Moreover, fluorescence staining with KJ16, an mAb that binds to receptors bearing either Vp8.1 or Vg8.2 gene products, revealed that all five CD4-CD8- clones were KJ16+ (Fig. 1).This indicated the presence of a functional TcR a/P on their surface exclusively using a Vg 8 gene. Of note, only 1of the 16 CD4+ clones was KJ16+. Consistent with this, full length TcRa (1.5 kb) and fi (1.3 kb) trans-

CD4-CD8-autoreactive T cells

Eur. J. Immunol. 1990. 20: 1265-1272

.

385

IF8

304

cripts were detected using C, and Cg probes in both CD4+ and DN clones (Fig. 2) and aVg8-specificprobe identified a 1.3-kb transcript in KJ16+CD4-CD8- clones (data not shown). Finally, none of these clones was stained by the Vg8.2-specific F23.2 mAb indicating that they all use the same Vg8.1 gene segment. These results demonstrated that the CD4-CD8- TcR alpbearing KJ16+ T cells, despite their immature phenotype, can expand in vitro and retain this DN phenotype. Moreover, since no additional exogenous signal such as rIL or conditioned medium was added t o cultures, it could be postulated that DN clones responded to epitopes present on the surface of syngeneic DC.

.

03F9

1267

Figure 1. FCM analysis of the surface phenotype of DBA/2 clones 3B5, 1F8, 3D4 and 03F9. (a) Thy-1 (-); CD5 (---); (b) CD4 (-); CD8 (---); (c) CD3 (-); KJ16 (---); ( a s ) control (. . .).

3.2 Representativity of CD4-CDS- TcR a@+ Vg8+ splenic T cells in normal mice

In previous experiments, we showed by two-color analysis in FCM that CD4-CD8- cells in the TKRlOl cell line were almost exclusively KJ16+ (manuscript submitted). The present DN clones are consistent with this finding. However, it was possible that this functional population of DN TcR a/p-bearing T cells was not representative of lymphocyte subsets present in DBA/2 mice. To address this question, normal and anti-T cell mAb-treated spleen cells from naive DBA/2 animals were cultured with KJ16 mAb immobilized by adherence to tissue culture plates. After 5 days, proliferating cells were recovered and restimulated

Kb

m1.61.00.6-

21)1.6

-

1.0. 0.6.

Figure 2. Northern blot analysis of total cellular RNA from CD4-CD8- (1F8, 1G3) and CD4+ (3A6, 03F9 and 3D4) clones. Hybridization was to 32P-labeled a and fl C region probes.

KJl6 Figure 3. Two-color analysis in direct fluorescence of normal (a) or anti-T cell mAb-treated (b) DBAR spleen cells activated by insoluble KJ16 and restimulated by syngeneic DC. Dotted lines represent the upper limit of autofluorescence.

1268

Eur. J. Irnrnunol. 1990. 20: 1265-1272

M. Seman, S. Boudaly, T. Roger et al.

by irradiated syngeneic DC. Their phenotypes were subsequently determined by two-color analysis in FCM. As shown in Fig. 3, most KJ16+ proliferating T cells from untreated populations were bright CD3+ and CD4+ or C D V , although the number of recovered cells was low. In contrast, cells derived from anti-Tcell mAb-treated spleen were CD3+, KJ16+ but mostly CD4-CD8-, as most KJ16+ clones derived from the TKRlOl cell line. Altogether, this

3

haplotype of stimulating

3.3 Activation of DN clones by DC of different haplotypes To explore the specificityof DN and CD4+ clones these cells were transferred into Click’s medium containing 1%NMS to avoid any possible reactivity to components of the FCS in which they were routinely grown. They were then stimulated by DC from a large panel of H-2 haplotypes and their proliferation was measured after 3 days. Results illustrated in Fig. 4 indicate that cells from 3B5, 1F8 and 1G3 DN clones could be induced to proliferate by syngeneic H-2d but also by allogeneic H-2b-k,s,f.’+r Or P DC. The level of proliferation reached with the various inducer cells was variable but always significant. Similar results were obtained with clones 3D10 and 3D12, suggesting, therefore, that the different CD4-CD8- clones had a very similar but unusual broad pattern of reactivity. As shown in Fig. 5 , the lower level of proliferation observed with C3H.Q or BIO.M allogeneic DC was not improved by increasing the number of stimulatory cells in cultures.

03F9

H-2

confirmed that DN Vg8+ T cells are present in the spleen of normal mice which can differentiate in vitro into autoreactive T lymphocytes.

dendritic cells

Figure 4. Proliferative response of CD4+CD8- (03F9,3D4,3A6) and CD4-CD8- (3B5, 1F8, 1G3) cloned T cells to DC from various origin. T cells (2 x 104) were cultured with 2 X 1OS irradiated DC or L class 11- cells. H-2d,b.k.s,f+r.p.@ DC were from DBA/2 (Mls-la), BALB.B (MIS-l’), BALB.K, BIO.S (MIS-1’). BlO.M, C3H.Q (Mls-lh), BlO.RIII, C3H.NB, D2GD and B 10.A(3R) mice, respectively. Experiments were performed in Click’s medium containing 1% NMS. Proliferation was measured after 3 days with [3H]dThd. Each value represents the mean of triplicates from at least two independent experiments.

385

Since all DN clones responded to all haplotypes tested so far, it was unlikely that they represented a mixture of autoreactive and allospecific T cells. It suggested rather that they were not MHC restricted or that they recognized non-polymorphic MHC determinants. In contrast, all CD4+CD8- clones, as illustrated in Fig. 4 for clones 03F9, 3D4 and 3A6, only responded when stimulated by syngeneic DBAI2 (H-2d) or I-A-identical D2GD DC, indicating that this activation was I-Ad restricted. Hence, analysis of clone activation by DC of various haplotypes allowed their classification into two categories of autoreactive T cells correlated to their CD4+CD8- or CD4-CD8- phenotypes. 3.4 Role of I-A molecules in DN clone activation

Clone specificity was further investigated by testing the inhibitory effect of anti-MHC mAb on the proliferation induced by syngeneic or allogeneic DC. Since none of the

7

3DlO

Dc/weii

i

to’’

Figure 5. Influence of DC concentration on the proliferative response of 1G3,3B5 and 3D10 CD4-CD8- Tcell clones.Tcells (2 x 104) were cultured with DC from DBAR ( O ) ,BALB.B (W), C3H.Q (0) and B1O.M (A) mice. Proliferation was measured after 3days. Each point represents the mean of triplicates from two independent experiments.

Eur. J. Immunol. 1990. 20: 1265-1272

CD4-CD8-autoreactive Tcells

cpm/culture x 10-3 Stimulator

8AL8'8

mAb

, ,~ ,,

F

'IAd 1-p. b.d

103

~

1

1

0

u

m

L

111

cc

1,

F

t t

385

3D4

Figure 6. Inhibition of the proliferative response of DN 1G3 and 3B5 and CD4+ 3D4 T cell clones by mAb. Clones were cultured as described in Fig. 4. Ascites fluid from anti-I-Ad (MKD6), antiI-Ad.h.p.q (34-53S), anti-I-Ak (26-7-113), anti-I-AS (MKS4), antiI-E (14-4-43) and anti-Vp8 (KJ16) were added at a 111OOO final dilution. Proliferation was measured on day 3 by [3H]dThd uptake. Each value represents the mean of triplicates from three independent experiments.

I-A+ cells to induce their proliferation. We first tried to stimulate clones with I-Ak-(CA14.11.14), I-Edlk(CA36.1.3) or I-EWk(CA36.2.1)-transfected fibroblast cell lines [30]. As shown in Fig. 7, these two cell lines did not induce DN or CD4+ T cell clones to proliferate. The same negative results were obtained using WEHI-3, A20-2J and P388-D1 (not shown) BALBk (H-2d) tumor cell lines as stimulators with or without prior IFN-y induction of class I1 expression. In previous experiments [16] we have shown that autoreactiveT cells can respond to total irradiated spleen cells which contain B lymphocytes, although this response requires the addition of larger cell numbers per culture than when DC-enriched spleen cell fractions are used. Recent findings also suggest that activated B lymphocytes have a specific role in the stimulation of alloreactive Tcells [31]. Thus, we tested the clone response to LPS-induced DBA/2 B cell blasts. Whereas all clones proliferated with syngeneic DC, they failed to respond to LPS-activated blasts (Fig. 7). Finally, since DN clones could be activated by DC of all haplotypes, they might be expected to respond to nonpolymorphic class I1 determinants.To explore this hypothesis, we tested their ability to respond to monomorphic elements shared by murine and human class I1 molecules. 1Q3

3A6

clones, as illustrated in Fig. 4, responded when cultured with the class 11- L fibroblast cell line, experiments were first carried out with anti-class I1 mAb. As shown in Fig. 6 for clone 1G3,3B5 and 3D4, the autologous response of all CD4-CD8- and CD4+CD8- clones, was inhibited by mAb recognizing private or public specificities of the I-Ad molecule. None of the 21 clones was inhibited by anti-I-E (anti-Ia.m7) mAb, although I-E molecules bear a monomorphic E, chain [29]. Similarly, the allogeneic response of CD4-CD8- clones to D C of the H-2b,kOr haplotypes was inhibited by mAb to corresponding I-A private specificities expressed on stimulatory cells. Hence, I-A gene products on DC seemed to be involved in DN clone stimulation. In agreement, MKD6 mAb which recognizes a private specificity of I-Ad, did not inhibit stimulation by allogeneic DC. This also ruled out the possibility that clones would react t o themselves as predicted from the lack of class I1 expression on their surface in FCM (data not shown). The role of I-A molecules could also be anticipated from the proliferation observed when DN clones were stimulated by D C from the H-2b,f4and haplotypes which do not express I-E molecules on their surface [29]. Finally, considering the unusual specificity of these clones, we verified that a unique TcR was involved in both syngeneic and allogeneic recognition. As illustrated in Fig. 6, 1G3 and 3B5 (CD4-) clone responses to DBA/2 (H-2d) and to BALB.K (H-2k)D C were completely abrogated by soluble KJ16 mAb, suggesting, therefore, that the same V08+ TcRa/fi was involved in both autologous and allogeneic stimulations.

3.5 Recognition of self peptides on DC In an attempt to define the role of I-A molecules in the activation of DN clones, we tested the ability of different

1269

0 10

.

.I*

cA.3CZ.I CAQC.I.9

5 10

2 z

3

4

10

\

L 0

10

C e l l / w e l l x 10-4

Figure 7. Proliferative response of CD4+ 3A6 and CD4- l G 3 T cell clones to different class 11+cells.Tcells (2 x lo4) were cultured with irradiated or mitomycin C-treated cells. Proliferation was measured on day 3 by [3H]dThd uptake. Each value represents the mean of triplicates from three independent experiments.

1270

Eur. J. Immunol. 1990. 20: 1265-1272

M. Seman, S. Boudaly,T. Roger et al.

DN clones were cultured with human PBL of various HLA haplotypes. No proliferation was observed under these conditions (data not shown).

Table2. DN (1G3, 3B5) and CD4+ (3D4, 03F9) autoreactive clones can induce polyclonal IgM secretiona)

Altogether, these results demonstrated that I-A molecule are necessary but not sufficient to stimulate both DN and CD4+ autoreactive cloned Tcells. They also suggested that clones do not exclusively react to polymorphic or nonpolymorphic I-A determinants. These clones most likely recognize self peptides with I-A molecules and these peptides are only present on the surface of DC.

Cloned T

IgM-secreting PFclculture X

WllSb)

-

LPS 1G3 3B5 3D4 03F9

DBAR

BALB.B

BALB.K

0.2 f 0.05 15.1 f 2.5 9.6 f 1.7 5.2 f 2.3 3.6f 1.2 22.5 f 2.7

0.5 f 0.03 34.9 f 4.7 17.2 k 5.3 8.3 f 2.7 1.2 0.9 0.4 f 0.1

0.4 f 0.02 27.5 f 5.6 8.5 f 2.8 6.4 f 1.9 0.6 f 0.2 0.7 f 0.3

*

3.6 Helper functions Despite their phenotype resembling that of immature thymocytes, peripheral DN clones displayed strong helper activity and were able to induce B lymphocytes to proliferate and to differentiate into Ig-secreting cells. This was established by testing the ability of the five DN clones to induce syngeneic DBA/2 (H-2d) and allogeneic BALB.B (H-2b)or BALB.K (H-2k) B cells to proliferate. As shown in Table 1, very similar levels of B cell proliferation were achieved with as few as 1 x 103 irradiated cloned T cells under syngeneic and allogeneic conditions. In contrast, 3D4,3A6 and 03F9 CD4+ clones, which were shown to be restricted by I-Ad private determinants, only induced syngeneic DBA/2 B cells to proliferate. This provided further support to the broad I-A specificity of DN clones described above. Similarly, Table 2 shows that all clones could induce syngeneic DBA/2 B lymphocytes to differentiate into IgM-secreting cells. Again, IgM secretion was observed when allogeneic BALB.B and BALB.K B cells were cultured with DN cloned T cells. This was not inconsistent with the above results indicating that clones cannot be activated by B lymphocytes and, therefore, do not directly interact with them. DC are present in the splenic B cell populations which can induce CD4+ and

a) T-depleted spleen cells (1 x lo6) were cultured with 1 X 10" irradiated cloned T cells. Reverse PFC were measured after 4 days. Each value represents the mean of duplicates from two independent experiments 5 SD. b) 1 x l@/culture.

CD4- autoreactive T cells to secrete B cell activating lymphokines. Experiments to be published support this conclusion.

4 Discussion The present results confirm the existence of CD4-CD8TcR a@+peripheral Tcells in normal mice and demonstrate that part of them correspond to fully competent autoreactive T lymphocytes. Since all DN clones described herein use the sameVp8.1 gene segment and derive from the same TKRlOl cell line, it could be argued that they represent the progeny of a single cell with a high multiplication rate. Preliminary results on TcR fi gene rearrangement in the different DN clones do not exclude this possibility (to be published). However, direct activation of anti-T cell mAb-

Table 1. DN (3B5, 1F8, lG3, 3D10, 3D12) and CD4+ (3D4, 3A6, 03FY) autoreactive clones can induce virgin B lymphocytes to proliferate") Clone T celVculture x 10-3

LPS 3B5 1F8 1G3 3D10 3D12

3D4

-

-

10 1 10 1 10 1 10 1 10 1 10 1

3A6

03F9

10 1 10 1

B cell proliferation ([3H]dThd cpm X lO-%ulture) DBA/2 BALB.B BALB.K

2.3f 0.3 70.4f 5.6 35.0f 5.8 19.8f 3.2 30.9f 6.2 21.7f 4.5 46.2f 9.6 41.5f 7.5 35.4f 7.1 28.7f 3.9 42.9f 9.5 28.7f 4.3 24.4f 3.6 10.3+ 2.4 22.6f 3.1 12.4f 2.3 84.1 f 12.9 36.2f 6.3

2.8f 1.7 101.1 f 9.7 32.2 f 4.1 13.7 f 2.6 38.9 k 7.1 16.8 f 3.4 41.1 f 7.8 27.8 f 3.9 27.2 f 3.4 19.7 f 3.1 24.9 ? 3.7 13.2 f 1.9 2.1 0.7 1.8 f 0.4 4.8 k 0.9 3.0 f 0.5 2.7 f 0.4 1.9 f 0.3

*

1.5 f 0.9 80.4f9.1 46.5 f 5.4 24.3 f 2.1 48.8 8.6 31.1 f 5.8 49.0 f 8.3 36.6 f 5.4 39.7 f 5.7 30.7 4.7 42.2 f 7.8 19.7 f 2.1 2.0 f 0.4 1.5 0.7 4.4 f 0.7 3.6 f 0.8 1.6 f 0.4 1.7 k 0.3

+

*

a) T-depleted spleen cells (3 x los) were cultured with irradiated cloned T cells. Proliferation was measured on day3. Each value represents the mean of triplicates from two independent experiments ? SD.

Eur. J. Immunol. 1990. 20: 1265-1272 treated spleen cells from different DBAI2 mice by insoluble KJ16 mAb, leads to the proliferation of DN Vp8+ autoreactiveTcells (Fig. 3), indicating that their presence in t h e TKRlOl cell line is not fortuitous. Similar DN Vg8+ cells have been described in the thymus of normal and MRL Iprllpr autoimmune mice [32, 331. This phenotype has led to postulate that these DN cells represent an immature thymocyte population which may join the pool of CD4+CD8+cortical thymocytes and subsequently give rise to single-positive CD4+ or CD8+ Tcells [34]. This conclusion was reached from the observation that, in MRL Zprllpr autoimmune mice, cells with a similar phenotype accumulatc in peripheral lymphoid tissues during the lymphadenopathy accompanying the disease [35]. It was, thus, proposed that, in Zpr mice, these precursors leave the thymus prematurely and accumulate in lymphoid tissues as a consequence of their inadequate intrathymic elimination and inability to differentiate normally. In normal mice, however, TcR a@+DN thymocytes, which appear late in ontogeny, may, alternatively, represent a particular T cell lineage of fully competent cells [32]. Consistently, our CD4-CD8- peripheral clones, which bear a functional TcRa/@ and use Vp8 gene segments as their thymic counterparts, are efficient Th cells in vitro. Their autoreactivity also suggests that they may correspond to cells which have not been negatively selected in the thymus, a phenomenon which is believed to occur among CD4+CD8+ thymocytes [2, 3, 361. However, most clones derived from theTKRlO1 cell line, as other autoreactive T cells derived from anti-T cell mAbtreated spleen [16] have a CD4+ phenotype. Despite careful control experiments [17, 181, we cannot formally ascertain whether these clones derive from DN splenic precursors that would mature in vitro into CD4+ autoreactive T cells, as previously suggested, or from a few contaminating CD4+ lymphocytes still present in the initial population. However, this last interpretation does not explain why most contaminating CD4+ T cells that would not be detected after treatment due to CD4 modulation would be highly enriched in autoreactive Tcells. Moreover, using panning and cell sorting methods, Reimann et al. [17] have similarly proposed the existence of DN precursors in the spleen capable of differentiating into CD4+ autoreactive T cells. If so, since most CD4+ clones are Vp8-, it implies that the pool of splenic DN is heterogeneous and contains different subsets with distinct repertoires and differentiation pathways. Alternatively, the population of DN precursors might be of limited heterogeneity but their Vgene usage would evolve in vitro, as already illustrated for V, genes [37].This shift inVgene expression would occur as a consequence of extra-thymic differentiation in cells which subsequently express the CD4 mo1ecule.V gene expression and rearrangement in CD4+ clones is currently under investigation to test this hypothesis. Both CD4+ and DN clones respond to self peptides bound to I-A molecules on the surface of DC. This conclusion is based on the observation that clone activation is inhibited by mAb to I-A molecules on stimulatory cells and cannot be induced by other Ia-bearing cells but DC. Moreover, experiments were performed under conditions in which no known foreign antigen was introduced, establishing clone specificity for self components. Most TcR a/p-bearing

CD4-CDS-autoreactive Tcells

1271

antigen-specific CD4+ T cells recognize antigen fragments bound to MHC molecules. In this recognition process, it is assumed that polymorphic amino acid residues within MHC products are responsible in part for the selective binding of antigen to class I1 molecules [38, 391 and may also be recognized by TcRa/P, thus reinforcing MHC restriction. This restriction occurs in the case of CD4+ autoreactive clones which can only be activated by syngeneic or I-Ad-compatible DC. DN clones also seem to recognize self peptides with I-A but can be activated by DC of the various haplotypes tested. Their activation is, therefore, not restricted. Several interpretations can account for this situation. It can be postulated that D N clones recognize a monomorphic self peptide on DC. Presentation of this peptide(s) would be poorly affected by I-A polymorphic residues allowing its recognition by CD4- clones on DC of various haplotypes. It is also conceivable that the TcR of DN clone does not react with I-A residues or reacts with non-polymorphic ones, whereas in CD4+ cells it interacts with polymorphic elements. If so, it is worth noting that in these DN clones TcR permits self peptide recognition but somehow mimics the specificity attributed to the CD4 molecule. Indeed, CD4 has been shown to facilitate antigen recognition by interacting with class I1 monomorphic residues on APC as would the TcR in DN clones [40]. CD4+ cloned cell lines, on the other hand, could recognize either another peptide or the same peptide but with polymorphic I-A residues. These models are compatible with the heterogenous level of proliferation observed when DN clones are act.ivated with D C of different haplotypes. In some respect, this resembles the response to Mls-1 or MIS-2 gene products [21].These responses also show little MHC restriction and Mls-1-reactive Tcells preferentially use V gene segments of theVp8 family [22]. However, these two loci are unlikely to be involved in our system. First, Mls-1-reactive Tcells are found in Mls-lb mice and react to Mls-la expressing leukocytes. Second, a response to Mls-la can be induced by B cells but not by D C or macrophages [41], and our Vp8.1+ DN clones derive from DBA/2 mice (Mls-1”) and react with DC expressing both Mls-la (DBAR, D2.GD) or Mls-lh (B20, BALB/c, C3H) alleles. Similarly, these clones react with D C from both M l ~ - 2(BALB, ~ C3H, DBAR) and M I S - (B6, ~ ~ BlO) mice [13]. One of the most striking results concerns the role of DC in clone activation. Recent reports have established the specific role of B lymphocytes in the activation of Vp17a+ I-E-reactive clones in tolerance induction to the poly GlupolyTyr (GT) copolymer or in the response to Mls-la [ 10, 14, 421. This role has been attributed to particular self peptides presented by Ia molecules on B cell membrane. Similarly, one could assume that other particular self peptides are expressed on D C which are recognized by peripheral autoreactiveT cells. Consistently, DC have been shown to be the best activators in autologous or syngeneic MLR in mice [24]. Moreover, they are excellent APC and are very potent into initiating immunity to alloantigens, H-Yantigens, or contact sensitizers [43]. In some cases too, they seem extremely efficient in inducing autoimmune diseases [43]. Experiments are in progress to define the genetic origin of peptides presented by D C and to which our clones would respond. However, this will not explain why DN autoreactive precursors are present in the periphery

1272

M. Seman, S. Boudaly,T. Roger et al.

and what is their physiological or pathological significance. Our results indicate that DN clones display helper functions but that these functions are mediated by lymphokines ( t o be published), since clones do not directly interact with B cells. I t seems, thus, unlikely t h a t their major role is t o account for polyclonal autoimmune diseases. Considering that cells with similar phenotype also behave as natural suppressor cells of MLR in vitro [44] or that NK-like DN cells activated by DC are involved in the regulation of allogeneic responses [45], it can be assumed that, in normal mice, DN autoreactive T cells, a n d perhaps their CD4+ counterparts, participate in t h e regulation of i m m u n e surveillance and/or immune network. In vivo transfer experiments with these clones are currently u n d e r investigation t o test this hypothesis.

The authors are very grateful to E! Debre' and Yasmine Dutois for providing excellent help with cytofluorograph analysis. They also acknowledge Dr. E. Marche and Dr. A . Traunecker for providing cDNA probes.

Received April 20, 1989: in revised form February 6, 1990.

5 References Marrack, P. and Kappler, J., Immunol. Today 1988. 9: 308. Sha, W. C., Nelson, C. A., Newberry. R. D., Kranz. D. M., Russell, J. H. and Loh, D. Y., Nature 1988. 336: 73. Scott, B., Bluthmann, H.. SiaTeh, H. and Von Boehmer, H., Nature 1989. 338: 591. Zuniga-Pflucker, J. C., Longo. D. L. and Kruisbeek. A. M., Nature 1989. 338: 76. Howe, M. L., J. Immunol. 1973. 100: 1090. Zauderer, M., Campbell, H., Johnson, D. R. and Sernan, M., J. Mol. Cell. Immunol. 1984. I : 65. Smith, J. B. and Pasternak. R.D., J. Immunol. 1978. 121: 1889. Glimcher, L. H., Longo, D. L., Green, I. and Schwartz, R. H., J. Exp. Med. 1981. 154: 1652. Sait0.T. and Raiewskv. K.. Eur. J. Immunol. 1985. 15: 927. 10 Kappler, J.,Wade, T.,'White, J., Kushnir, E., Blackman, M., Bill, J., Roehm, N. and Marrack, P., Cell 1987. 49: 263. 11 Kappler, J., Roehm, N. and Marrack, P., Cell 1987. 49: 273. 12 MacDonald, H . R., Schneider, R.. Lees, R. K., Howe, R. C., Acha-Orbea, H., Festenstein, H., Zinkernagel, R. M. and Hengartner, H., Nature 1987. 332: 40. 13 Pullen, A. M., Marrack, P. and Kappler, J., Nature 1988. 335: 796. 14 Vidovic, D. and Matzinger, P., Nature 1988. 336: 222. 15 Kisielow, P., Bliithmann, H., Staerz, U. D., Steinmetz, M. and Von Boehmer, H., Nature 1988. 333: 742. 16 De Talance, A., Regnier, D., Spinella, S., Morisset, J. and Sernan, M., J. Immunol. 1986. 137: 1101. 17 Reimann, J., Bellan, A. and Conradt, P., Eur. J. Immunol. 1988. 18: 989.

Eur. J. Immunol. 1990. 20: 1265-1272 18 Morisset, J.,Trannoy, E., DeTalance, A., Spinella, S., Debrt5, P., Godet, P. and Seman, M., Eur. J. Immunol. 1988. 18: 387. 19 Brenner, M. B., McLean, J., Dialynas, D. P., Strominger, J. L., Smith, J. A., Owen. F. L., Seidman, J. G. and Krangel, M. S., Nature 1986. 322: 145. 20 Raulet, D. H., Immunol. Rev. 1989. 7: 175. 21 Janeway, C. A., J. Immunogenetics 1988. 15: 1.19. 22 Kappler, J. W., Staerz, U., White, J. and Marrack, P.. Nature 1988. 332: 35. 23 Leo, D.. Foo, M., Sachs, D. H., Samerlson, L. E. and Bluestone, J. A., Proc. Natl. Acad. Sci. USA 1987. 84: 1374. 24 Nussenzweig, M. C. and Steinman, R. M., J. Exp. Med. 1980. 151: 1196. 25 Chirgwin, T. M., Przybyla, A. E., MacDonald, R. J. and Rutter, W. J., Biochemistry 1979. 18: 5294. 26 Chien,Y., Becker, D. M., Lindsten, T., Okarnura, M., Cohen, D. I. and Davis, M. M., Nature 1984. 312: 31. 27 Hedrick, S. M., Nielsen, E. M., Kavaler, J., Cohen, D. I. and Davis, M., Nature 1984. 308: 153. 28 Chien, Y., Gascoigne, N. R. J., Kavaler, J., Lee, N. E. and Davis, M. M., Nature 1984. 309: 322. 29 Klein, J., Figueroa, F. and David, C. S., Immunogenetics 1983. 17: 553. 30 Shastri, N., Malissen, B. and Hood, L., Proc. Natl. Acad. Sci. USA 1985. 82: 5885. 31 Marrack, P. and Kappler, J., Nature 1988. 332: 840. 32 Fowlkes, B. J., Kruisbeek, A. M..Ton-That, H.,Weston. M. A., Coligan, J. E., Schwartz, R. H. and Pardoll, D. M., Nature 1987. 329: 251. 33 Budd, R. C., Miesher, G., Howe, R. C., Lees, R. K., Bron, C. and MacDonald, H. R.. J. Exp. Med. 1987. 166: 577. 34 Lanier, L. L., Federspie1,W. A,, Ruitenberg, J. J., Phillips, J. H., Allison, J. l?, Littman, D. and Weiss, A., J. Exp. Med. 1987. 165: 1076. 35 Budd, R. C., Schreyer, M., Miesher, G. C. and MacDonald, H. R., J. Imrnunol. 1987. 139: 2200. 36 Sia Teh, H . , Kisielow, F!, Scott, B., Kishi, H., Uematsu, Y., Bluthmann, H. and Von Boehmer, H., Nature 1988. 335: 229. 37 Marolleau, J. P., Fondell, J. D., Malissen, M.,Taucy, J., Barbier, E., Marcu, K. B., Cazenave, P.-A. and Primi, D., Cell 1988.55: 201. 38 Gay, D., Buus, S., Pasternak, J., Kappler, J. and Marrack, P., Proc. Natl. Acad. Sci. USA 1988. 85: 5629. 39 Kourilsky, P. and Claverie, J. M., Annu. Immunol. Inst. Pasteur (Paris) 1986. 1370: 3. 40 Clavcrie, J. M. and Kourilsky, P., Annu. Immunol. Inst. Pasteur (Paris) 1986. 1370: 425. 41 Webb, S. R., Okamatu, A., Ron,Y. and Sprent, J., J. Exp. Med. 1989. 169: 1. 42 Hamilos, D. L., Mascali, J. J., Chesnut, R. W.,Young, R. M., Ishioka, G. and Grey, H. M., J. Immunol. 1989. 142: 1069. 43 Knight, S. C., Farrant, J., Chan, J., Bryant, A., Bedford, €? A. and Bateman, C., Clin. Immunol. Immunopathol. 1988. 48: 277. 44 Hertel-Wulf, B., Lindsten,T., Schwadron, R., Gilbert, D. M., Davis, M. M. and Strober, S., J. Exp. Med. 1987. 266: 1168. 45 Shah, P. D., Cell. Immunol. 1987. 104: 440.