Eur. J. Immunol. 1990. 20: 1435-1443 Forbidden T cell receptors in autoimmune disease 1435

Sharlene Adams, Tracy Zordan, Krishna Sakis and Syamal K. Datta

The Department of Medicine, New England Medical Center and Sackler Immunology Programme, Tufts University School of Medicine, Boston

T cell receptor Vp genes expressed by IgG anti-DNA autoantibody-inducing T cells in lupus nephritis: forbidden receptors and double-negative T cells*

In the (SWR x NZB)F1 (SNF1) model of lupus nephritis, pathogenic variety of IgG anti-DNA autoantibodies are induced by certain T helper (Th) cells that are either CD4+ or CD4-CD8- (double negative; DN) in phenotype. From the spleens of eight SNFl mice with lupus nephritis, 149 Tcell lines were derived and out of these only 25 lines (= 17%) were capable of augmenting the production of pathogenic anti-DNA autoantibodies. Herein, we analyzed the T cell receptor (TcR) Vg genes used by 16 such pathogenic autoantibody-inducing Th cell lines. Twelve of theTh lines were CD4+ and among these five lines expressed Vp8 (8.2 or 8.3). The Vg8 gene family is contributed by the NZB parent to the SNFl mice, since it is absent in the SWR parental strain.Three other CD4+ Th lines expressed Vg4, another was Vp2+ and one line with poor autoantibody-inducing capability expressed Vgl. Four autoantibody-inducing T h lines from the SNFl mice had a DN phenotype and these lines were also autoreactive, proliferating in response to syngeneic spleen cells. Among these DN Th lines, two expressed Vp6 and one expressed Vg8.1 TcR. Both of these are forbidden TcR directed against Mls-la (Mlsa) autoantigens expressed by the SNFl mice and such autoreactive T cells should have been deleted during thymic ontogeny. Thus, the DN Th cells of non-lpr SNFl mice are different from the DN cells or MRL-lpr which lack helper activity and do not express forbidden TcR. The spleens of 6 out of 19 nephritic SNFl animals tested also showed an expansion of forbidden autoreactive TcR+ cells that were mainly DN. Two of these animals expressed high levels of Vp6 (anti-Mlsa) and Vgll (anti-I-E) TcR+ cells, three others had high levels of Vpll+ cells alone and one animal had an expanded population of Vgl7a+ (anti-I-E) cells. The I-E-reactive TcR again should have been eliminated in the SNFl thymus, since they express I-E molecules contributed by the NZB parent. The SWR parents of SNFl, are I-E-; moreover, they lack the Vgll gene but they express Vgl7a in peripheral T cells. Whereas the NZB parents are I-E+, they lack a functional Vgl7a gene and they delete matureVgll+ T cells.Thus, a combination of MHC class I1 and TcR genes inherited from the NZB and SWR parents and abnormalities in thymic deletiodselection processes may generate the auto- antibody-inducing Th cells in the SNFl mice.

1 Introduction

The (NZB x SWR)F1 or (SWR x NZB)Fl (SNF1) progeny of crosses between autoimmune NZB and normal SWR mice uniformly develop lethal and accelerated lupus ne- phritis, in marked contrast to their parents [l]. The SNFl mice produce a selected population of nephritogenic anti-DNA autoantibodies that are qualitatively different from the nonpathogenic anti-DNA autoantibodies pro- duced by their NZB parents [2-41. The pathogenic auto- antibodies of SNFl mice are IgG antibodies, they are

[I 81951

* This work was supported by National Institutes of Health grant

Correspondence: Syamal K. Datta, Department of Medicine, Box 52, New England Medical Center, 750 Washington Street, Boston, MA 021 11, USA

Abbreviations: NZB: New ZealandBlack SNF,: (SWR X NZB)F, or (NZB x SWR)FI DN: Double negative

CA 31789.

predominantly cationic in charge and they constitute two major families sharing a recurrent cross-reactive idiotype called IdLNFl [2-41. The parental NZB or SWR strains do not express these pathogenic autoantibodies, although they contribute the genes for the pathogenic autoantibodies to their SNFl progeny [2-41. Only in the nephritis-prone SNFl progeny is this idiotypically connected population of pathogenic anti-DNA autoantibodies selected to undergo an oligoclonal expansion under the influence of certain Th cell subsets that emerge just before the onset of lupus nephritis [5,6]. One set of these pathogenic autoantibody- inducing Th cells possess the classical CD4+CD8- helper phenotype whereas the other set of T h cells has an unusual CD4-CD8- (double-negative; DN) phenotype. IL 2-de- pendent, pathogenic autoantibody-inducing Th cell lines with either of the above-mentioned phenotypes can be derived from nephritic SNFl mice [5, 61.

Herein, we analyzed the TcR Vg genes expressed by these pathogenic autoantibody-inducing T h cell lines of SNFl mice. Analysis of TcR Vg genes used by these functionally characterized Th cells in the SNFl model is interesting for several reasons [7].The SWR parents of these NZB x SWR crosses, although normal, have deleted about 50% of the

0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1990 0014-2980/90/0707- 1435$3 .50 + .25lO

1436 S. Adams, T. Zordan, K. Sainis and S. K. Datta Eur. J. Immunol. 1990. 20: 1435-1443

TcR Vp gene families, namely Vp5, 8, 9, 11, 12 and 13 [8] . Complementation of those TcR genes inherited from the NZB parent by the SNFl mice may allow the generation of pathogenic autoantibody-inducing Th cells. Indeed, hetero- zygosity of the TcR Vp and MHC class I1 genotypes is associated with the development of accelerated glomeru- lonephritis in the NZB x SWR crosses [9]. Second, recent studies from Kappler and Marrack and several other laboratories have established that products of certain TcR Vp genes are strongly reactive to certain autoantigens. For instance, T cells expressing Vgl7a, 5 or 11 gene products react with I-E; Vp8.1, 6 and 9-encoded TcR respond to Mls-la (Mls") products; and Vp3-encoded TcR are reactive to M I s - ~ ~ (MlsC) products [lo-181. T cells expressing such strongly autoreactive TcR are deleted early in thymic ontogeny. Consequently, T cells expressing these "forbid- den" receptors are found at very low levels in the peripheral lymphoid organs of normal mice that express those auto- antigens [lo-181. The lupus-prone SNFl mice are also interesting in this regard [7] The SWR parents of this cross do not express I-E molecules, hence they have the I- E-reactive Vgl7a+ T cells in the periphery [lo]. The other I-E-reactive TcR genes (Vp5 and Vpll) are lacking in this strain [8]. By contrast, the NZB parents of this cross are I-E+ and, therefore, should not have mature T cells expressing these Vp TcR. Second, in the SWR strain that has the Mls-la and M l ~ - 2 ~ phenotypes, the Mls antigens are non-stimulatory because they have the H-24 haplotype and they are I-E-, therefore mature T cells expressing Mls- reactiveVp6 and Vp3 receptors are present in this strain [ 12, 13,15, 181.The gene for the other anti-MlsTcR,Vp8.1 and Vp9, are lacking in SWR [S]. By contrast, the NZB mice, which are also Mls-la and M l ~ - 2 ~ , are H-2d and I-E+ and therefore they should delete mature T cells with those anti-Mls Vp TcR. The SNFl progeny is I-E+ and have stimulatory Mls antigens due to codominant inheritence of these genes. Moreover, they have the full complement of TcR Vp genes.Therefore, the SNFl progeny should deleteT cells expressing the forbiddenTcR. Herein, we analyzed the pathogenic autoantibody-inducing T h cell lines and peri- pheral Tcells from nephritic SNFl mice for the expression of the forbidden autoreactive TcR. We show that DN (CD4-CD8-) T cells from nephritic SNFl mice that have the functional ability to augment the production of IgG anti-DNA autoantibodies do express forbidden autoreac- tive TcR.

2 Materials and methods

2.1 Mice and derivation of T cell Lines

NZB, SWR and C57BL/6 mice were obtained from the Jackson Laboratory, Bar Harbor, ME. SNFl mice were bred by us. Female mice were used for all experiments.TheTcel1 lines were derived from different SNFl animals (6-7 months old) with nephritis using previously described procedures [6]. Our strategy was to derive T cell lines that could preferentially augment the production of IgG anti- DNA autoantibodies [6]. Since suchTh cells are activated in vivo in nephritic SNFl mice, they could be isolated in vitro by low doses of rIL2 in culture [6]. The T cell lines were derived by LD cultures, starting with a population of either wholeTor DN (CD4-CD8-) Tcells as described [6]. TheT cell lines were tested within 4 to 6 weeks after initiation of

culture and then retested several times afterwards. The surviving Th cell lines have been in culture for more than a year.

2.2 Helper function of T cell lines

After a week had elapsed since their last feeding regimen [6], the T cell lines were harvested on Ficoll-Hypaque (Pharmacia, Uppsala, Sweden) gradients and repeatedly washed. Serial dilutions of Tcells from the lines were then cultured with syngeneic B cells for 6 days as described [5, 61. Total IgG and IgG class autoantibodies to single- stranded (ssDNA) and double-stranded DNA (dsDNA) produced in the culture SN were then quantified by ELISA [ 5 , 61. Aliquots of the culture SN were affinity purified on DNA-cellulose columns followed by IEF and immunoblot- ting to detect IgG anti-DNA autoantibodies with cationic charge, as described [5, 61.

2.3 Surface phenotype of T cell lines

T cells from the lines were purified by Ficoll gradients and stained for FCM analysis as described [5, 61. Staining for Thy-1.2, CD4, CD8 and CD45R (Ly-5 or B220) was done as described [6]. The following mAb to TcR Vp gene products were also used: €23.1 (IgGz,), and F23.2 (IgG1) were from U. Staerz (Denver, CO) and M. Bevan (Scripps Clinic, La Jolla, CA); the former is specific for Vg8.1, 8.2 and 8.3 whereas the latter recognizesVp8.2 only [ l l , 191. KJ16 (rat IgG) specific for Vg8.1 and 8.2 and KJ23a (IgG2,) specific for Vgl7a were gifts from P. Marrack (Denver, CO) [lo, 20). RR3-15 (rat IgG) specific for Vpll was from 0. Kanagawa (La Jolla, CA) [14] and 44-22-1 (rat IgG) specific for Vp6 was given by J. Payne and B. Huber (Tufts Medical School, Boston, MA) [21].The cells were incubated with saturating concentrations of the TcR-specific mAb at 37 "C or 4 "C for 30 min in sodium azide-containing FCM buffer, followed by three washes and then stained with FITC-conjugated second-step antibodies, namely, goat anti-rat IgG (mouse serum absorbed, from Kirkegaard and Perry, Gaithersburg, MD) in the case of KJ16, 44-22-1 and RR3-15 and rabbit anti-mouse Ig (IgG + IgM + IgA, from Zymed, San Fran- cisco, CA) in the case of F23.1, F23.2 and KJ23a. Back- ground control stainings were done with isotype-matched mouse myeloma Ig (Litton Bionetics, Kensington, MD) followed by the FITC-labeled second-step reagents. Live cells were then analyzed by FACScan (Becton Dickinson, Mountain View, CA) as described [6, 221.

2.4 Northern analysis of RNA from T cell lines

Total RNA was extracted by previously described methods [23], from those T cell lines whose TcR Vp usage could not be assigned by FCM. After electrophoresis in formalde- hyde-containing gels [24], the RNA was transferred to Nytran (Schleicher and Schuell, Keene, NH), and hybrid- ized with "P-labeled probes for Vp genes (families 1 through 16) provided by D. Loh (Washington University, St. Louis, MO) [8] or Cp gene [25] obtained from P. Brodeur.

Eur. J. Immunol. 1990. 20: 1435-1443 Forbidden T cell receptors in autoimmune disease 1437

2.5 Proliferation assays with T cell lines

The assays were done as described [6], except that mito- mycin-treated spleen cells were used as APC. The T cell lines were harvested by Ficoll gradients and washed after 7 days had elapsed since their last feeding regimen. The T cells from the lines, at the optimal concentration of 1 x lo5 cells, were either cultured alone or co-cultured with mito- mycin-treated APC ( 5 x lo5) in 200 p1 medium [6] per well of a 96-well round-bottom plate. The APC were also cultured alone. The cells were cultured in quadruplicate for 72 h with a [3H]dThd pulse-labeling during the last 16 h of the cultures as described [6].

2.6 FCM analysis of spleen T cell populations

T cells were enriched from splenocytes by panning on anti-mouse Ig-coated plates as described [5 , 61. Panning was done only once, sinceTh cells that are activated in vivo in nephritic SNFl mice [5 , 61 would be lost by more rigorous purification. DN subset of T cells were obtained from the T cell-enriched population as described [5, 61, except that a one-step treatment with the IgM anti-CD4 mAb RL-172.4 [26] and anti-CD8 mAbTIB-211 (AmericanTissue Culture Collection, Rockville, MD, [27]) plus complement was used to eliminate CD4+ and CD8+ cells. For two-color FCM analysis, staining was done as described [28, 291. The T cell-enriched population were first incubated with 10% normal rabbit serum for 20 min to block FcR, washed and then incubated with the unconjugated, rat mAb, anti-Vp6 (44-22-1) or anti-Vpll (RR3-15) or anti-Vp8.1/8.2 (KJ16). After washing, the cells were incubated with FITC-goat anti-rat IgG previously absorbed on mouse Ig and then washed. Potentially free binding sites on the fluorescei- nated second-step antibody were then blocked by incubat- ing the cells with 10% normal rat serum.The cells were then directly stained by PE-labeled anti-CD4 (anti-mouse L3T4) and biotinylated anti-CD8 (anti-mouse Ly-2) from Becton Dickinson followed by PE-streptavidin (Southern Biotech- nology, Birmingham, AL). Background staining was done in the same way with FITC-anti-rat IgG and PE-strepta- vidin, except normal rat IgG (Zymed) was used in place of the first-step anti-TcR mAb and the third-step anti- CDWanti-CD8 mAb. Two-color staining of aliquots of the same cells was also done with FITC-KJ23a (anti-Vpl7a) or FITC-anti-Thy-1.2 (Becton Dickinson) followed by PE-anti-CD4 and biotin-anti-CD8 staining with PE-streptavidin as described above. FITC-mouse IgG (Cappel, OrganonTeknika,West Chester, PA) was used as a background control for the first step FITC mAb and normal rat IgG followed by PE-streptavidin was used in place of the anti-CD4KD8 mAb. Single-color staining of the T cell- enriched population and subsets were also done with each of the above mentioned mAb and appropriate second-step reagents to verify the results of two-color staining. For additional single-color analysis, aliquots of the T cell- enriched and DN cells were also stained with F23.1 (anti-Vp8) followed by FITC-protein A (Pharmacia, Pisca- taway, NJ) as described [30]. Since theTcells and subsets were not rigorously purified to avoid losing activated T cells, non-Tcells including some B cells contaminated the preparations. Therefore FITC-anti-mouse Ig was not used as a second-step reagent for staining these cell populations from spleen. The DN population obtained from the

nephritic SNFl spleens was also stained by FITC-anti- Th-1.2 (Becton Dickinson) or FITC-anti-CD3 (145-2C11, gift of J. Bluestone and H. Wortis, [31]) to determine the total proportion of T cells. Two-color and single-color analysis of live gated cells was performed as described [6,221.

3 Results

3.1 TcR Vp genes expressed by IgG anti-DNA autoantibody-inducing T cell lines from nephritic SNFl mice

The production of pathogenic, IgG anti-DNA autoanti- bodies isTcell dependent and theseTh cells are activated in vivo in SNFl mice with lupus nephritis [5 , 61. The antigenic specificities of such Th cells are unknown, but activated T cells express IL2R [32], therefore, we derived Tcell lines from nephritic SNFl animals by using low doses of rIL 2 and syngeneic feeder cells in culture. From the spleens of eight

Table 1. TcR Vg genes expressed by IgG anti-DNA autoantibody- inducing T h cell lines from nephritic SNFl mice

Th line") Increase factor in IgG antibody productionb)

Anti-ssDNA Anti-dsDNA Total IgG TcR VB IgG IgG

CD4+ lines 1A 20.7 9w. 7 26.2 8.54 4.9 9.2 8.0 8w.45 6.4 3A 6.0 9.10 6.6 9.37 6.7 9.7 11.0 8.32 3.7 9w.18 48.2 5A 3.7

28.0 7.0 1.9 2.8 2.9

12.7 2.2 4.2 2.9 2.3 6.5 1.0

7.4 8.2 4.9 8.2 4.0 8.2 4.6 8.3 2.7 8f) 4.1 4 4.9 4 5.0 4 5.7 2 2.9 unkd) 5.9 unk 2.9 1

CD4-CD8- lines 4B 4.0 10.0 1.4 6 8.48 10.0 3.8 8.9 6 10w.86 1.9 2.7 1.4 8.1 9.4 4.9 2.5 3.1 unk

Tcell lines were derived from different SNF, females with lupus nephritis. Those lines that augmented the production of IgG anti-DNA autoantibodies when co-cultured with syngeneic B cells are shown. The first number identifying each Th line is also the number of the SNF, animal from which the line was derived. The values represent the increase factor in production of the IgG antibodies when theTcells were co-cultured with syngeneic B cells as compared to B cells cultured alone in the same assay. Serial dilutions of cells from the Th lines were tested several times and the optimum help provided is shown. This line died before further assignment of TcR within theVg8 family could be made. unk, means unknown. These lines are growing poorly, there- fore,TcRVp assignment could not be made but they have been so far negative for Vp8, 6 and 17a expression by FCM analysis.

1438 S. Adams, T. Zordan, K. Sainis and S. K. Datta Eur. J. Immunol. 1990. 20: 1435-1443

F23.1 (8.1.82.8.3) F23.2(8.2) KJ16 (8.1.82) la1

1 L"E1A(CD4*'

Figure 1. (a) Representative FCM histograms of the SNFI- derived, autoantibody-inducing T h cell lines stained with mAb specific for TcR encoded by Vg8 genes are shown (solid lines). Isotype-matched background control stainings are shown by interrupted lines in this and subsequent figures. The Th line 1A is CD4+ (not shown, see [6]), and the line 10w.86 is Thy-l+ but is CD4-CD8-. (b) Representative Northern blots of FWA from the SNF1-derived Th cell lines hybridized with probes specific for TcR genes Vg4 and Vg2. The autoantibody-inducing T h lines, which are all CD4+ (Table l ) , are identified at the top of the respective lanes. The position of the full-length TcRP chain message (1.3 kb) is indicated by an arrow and was confirmed by hybridization of the same blots with a Cg probe. (c) FCM histograms of two autoanti- body-inducing, DN T h cell lines, 4B and 8.48, stained by the mAb specific for Vg6-encoded TcR (solid lines).

Table 2. Proliferative responses of DN autoantibody-inducing T h cell lines

rH]dThd incorporation: mean cpm f SEM and (SI)a)

APC 8.48 (Vg6+) 4B (Vg6+)

None 2 a 4 f 33 4 8 5 f 41 SWI 17928 f 1264 (61.6) 13471 k 623 (26.9) SWR 430 f 36 (0.1) 5506 f 1066 (10.5)

The proliferative responses of the SNFI-derived T h cell lines to APC from the parental (SWR) or syngeneic mice are shown. Stimulation index (SI) values are shown in parentheses.

Mean cpm (Th cells + APC) - mean cpm (APC alone) SI = Mean cpm (Th cells alone)

Mean cpm were calculated from quadruplicate cultures and the results are representative of three separate experiments. The mean c p m f SEM incorporated by the APC cultured alone were as follows: SNFl = 448 f 104 and SWR = 402 * 18.

SNFl females with lupus nephritis, 149 Tcell lines grew up in culture and out of these only 25 lines (= 17%) had the functional ability to augment the production of IgG anti-DNA autoantibodies when co-cultured with syngeneic B cells. Sixteen of these Th cell lines survived in vitro for further analysis and are shown inTable 1 and Fig. 1.Twelve of the Th lines were CD4+ and four were DN CD4-CD8-. The DN lines were all Thy-l+ and CD3+ (not shown, but results were similar to [6]). Three of these lines (Table 1) were also CD45R+ (B220 or Ly-5), namely the CD4+ line 8.54, and the DN lines 10w.86 and 9.4 (not shown). The CD4+ lines, lA, 3A and 5A and the DN line 4B were described previously [6], and they were retested for helper activity and surface phenotype before analysis of Vp gene usage in this study. All of the lines shown in Table 1 augmented the production of IgG anti-ssDNA autoanti- bodies more than total polyclonal IgG, but the CD4+ lines lA, 9w.7 and 9w.18 were markedly more potent than the others. All of the Th lines, except 5A, also augmented the production of IgG anti-dsDNA autoantibodies, but the CD4+ lines 1A and 3A and the DN Th line 4B were most potent in this helper activity. As described previously, the CD4+ lines 1A and 3A and the DN line 4B are not only strong anti-dsDNA autoantibody inducers but they also augment the production of cationic anti-DNA autoanti- bodies of the IgG class [6]. Indeed, the T h cells that could augment the production of IgG antibodies to dsDNA were also capable of inducing the production of cationic anti- DNA autoantibodies [5, 61.

Five of the twelve CD4+ Th lines were Vp8+ as detected by FCM using the three mAb F23.1, F23.2 and KJ16.Three of these lines, including the strongest autoantibody-inducer lA, were Vp8.2+ because they stained positively with all of the three mAb (Table 1 and Fig. la). Line 9.2 was Vp8.3+ because it stained with F23.1 but not with F23.2 or KJ16. Line 8w.45 stained positively with F23.1.Three other CD4+ Th lines, including the strong autoantibody-inducer 3A, were Vp4+ as detected by Northern analysis (Table 1 and Fig. lb).Three other CD4+ Th lines used different Vp genes genes (Table 1). The T cell line 5A is a poor autoantibody inducer [6], and it used the Vpl gene (Table 1).

Among the four DN Th lines, three used forbidden or strongly autoreactive TcR Vp genes. Lines 4B and 8.48 expressed Vp6, an Mls-la-reactive TcR (Fig. lc) and line 10w.86 also expressed another anti-Mls-la TcR, Vg8.1, because it stained positively with the mAb F23.1 and KJ16 but not with F23.2 (Fig. la). The TcR Vp usage of the remaining DN line has not yet been determined due to poor growth in culture. Thus, in contrast to the CD4+ T h cell lines, almost all of the DN, autoantibody-inducing Th lines expressed strongly autoreactive TcR Vp genes which should have been deleted in these SNFl mice.

3.2 Autoreactive proliferative responses of' DN Th lines expressing the forbidden receptors

The DN Th lines 4B and 8.48, that strongly augmented the production of pathogenic autoantibodies (Table l), also proliferated strongly in response to some endogenous antigen(s) presented by syngeneic SNFl APC (Table 2). TheVp6+ T h line 8.48 did not respond to the parental SWR strain's APC, but the other Vg6+, DN T h line, 4B, did so

Eur. J. Immunol. 1990. 20: 1435-1443 Forbidden T cell receptors in autoimmune disease 1439

significantly. The autoreactive responses of some of the CD4+ Th lines analyzed here have been described previous- ly PI .

3.3 TcR Vg gene expression by the splenic T cell populations of nephritic SNFl mice and the parental strains

Since the autoantibody-inducing, DN Th lines from the SNFl mice expressed forbiddenvp TcR, we analyzed freshly prepared T cell populations from the spleens of other nephritic SNFl animals for any expansion of T cells expressing such autoreactive TcR. Table 3 shows that the total proportion of CD4+ and CD8+ cells was relatively low in the splenic population analyzed as compared to other studies [ 10-181, because here the T cells were enriched by only one cycle of panning on anti-Ig-coated plates to prevent any loss of in vivo activated Tcells.The results from the parental strains were consistent with published work of others as described in Sect. 1 [ 10-18].The SWR mice do not have Vp8 and Vpll genes; however, their peripheral Tcells expressed the autoreactive TcR, Vp6 and Vgl7a, because they lack or cannot present the corresponding autoantig- ens. By contrast, the NZB mice that can present these autoantigens had markedly reduced levels of T cells expressing the forbidden Vp6 and Vpll TcR in their periphery (Table 3). In addition, the NZB strain did not have Vpl7a+ cells since it lacks a functional Vpl7a gene [33], but it expressed Vp8 in the range of other Vp8+ strains (Table 3). Representative examples of FCM histograms from the parental strains are shown in Figs. 2 and 3. In the SNFl progeny, the percentage of Vp8+ cells was about half of that found in the NZB parents due to hemizygosity (Table 3 and Fig. 2). None of the young preautoimmune SNFl had an expansion of Tcells expressing the forbidden autoreactive TcR (Table 3). Of the 19 old SNFl mice with nephritis that were tested 17 had reduced or “deleted” levels of Vp6+ T cells; 14 of these 19 mice also had reduced levels of Vpll+ Tcells and 18 of the 19 mice had reduced levels of Vgl7a+ T cells (Table 3).These results suggest that deletion or a marked reduction of Tcells bearing forbidden autoreactive TcR did occur in these nephritic SNFl mice at the time they were tested. As evident from Table 3 and

consistent with previous work [5, 61, the old nephritic SNFl mice, in contrast to the young preautoimmune SNFl mice, had an expanded population (= 10’30) of DN Tcells, since only ~ 8 9 % of Thy-l+ cells in their spleens were CD4+CD8+.

In contrast to the above results, we did detect an expansion of Tcells expressing the strongly autoreactive TcR in some of the nephritic SNFl animals (Table 4 and Figs. 2 and 3). T cell-enriched populations from the spleens of two SNFl animals contained a 9.8-fold and 16.2-fold increase in the percentage of Vp6+ cells as compared to the other nephritic SNFl mice that had “deleted” this autoreactive receptor (Table 4 vs. Table 3). Interestingly, most of the Vp6+ cells in these mice were in the DN population (Table 4 and Figs. 2 and 3). Fig. 2 (single color) and Fig. 3a (two color) show that the density of Vp6 expressed by the SNFl Tcells was variable. Although the very intensely staining subset of Vp6+ cells found in SWR was infrequent in the SNFl mouse, the SNFl T cells clearly stained positive as com- pared to NZB which had deleted Vp6+ cells. Furthermore, unlike the SWR, most of the Vp6+ cells in the whole Tcell population of the SNFl mice were contributed by DN cells as shown by two-color staining (Table 4 and Fig. 3a) and these Vp6+ cells were further enriched by cytotoxic elimi- nation of CD4+ and CD8+ cells from the whole T popula- tion (Figs. 2 and 3a). Again, because we did not rigorously purify the initial Tcell population, the DN cells from the spleens of these SNFl mice that were enriched by removal of CD4+ and CD8+ cells, contained many non-T cells. Indeed, the YO of Thy-l+ or CD3+ cells in these DN populations ranged between 30 to 40% (not shown). Next, five of the nephritic SNFl animals were found to have an increase in Vpll+ cells ranging from 4.3-fold to 6.2-fold above the other nephritic SNFl mice that had “deleted” these autoreactive T cells (Table 4 vs. Table 3). Moreover, two of these Vpll-high SNFl mice were the same animals that also had an increase inVp6+ cells. Unlike the case with Vp6, a significant proportion of the Vpll+ cells in these SNFl mice were either CD4+ or CD8+, although the majority again were DN as shown in Table 4 and in Figs. 2 (single color) and 3b (two color). T cell-enriched spleno- cytes from a C57BL/6 mouse, another I-E- strain, is shown as a positive control, since the SWR strain lacks the Voll

Table 3. Percentages of splenic Thy-1.2+ cells expressing Vg TcR and CD4 or CD8 antigens

Vpll Vgl7a CD4 and CD8 Straina) n Vg8b) VP6

0.06 f 0.05 13.70 f 0.41 99.26 t 2.52 0.95 f 0.20 0.02 k 0.02 98.36 t 1.44

97.84 t 1.95 10 14.19 k 1.73 0.20 k 0.07 1.03 f 0.24 89.14 k 2.13 1.26 f 0.44

SWR 12 0.05 & 0.04 8.27 f 0.41 NZB 9 25.50 & 0.76 0.40 rf: 0.24

SIWl (old) 19c) 14.66 k 1.29 0.31 f 0.08 0.13 f 0.04 0.16 f 0.05

SNFl (young)

a) All mice were females. SWR mice were 6 months old. The NZB were 4 months old when tested because older NZB mice have abnormal spleens due to hemolytic anemia with extramedullary hematopoiesis.Young SNFl mice were 2 months old and they lacked manifestations of autoimmune disease. The old SNFl mice were 6 to 7 months old with 30 mglml proteinuria.

b) The data represent mean % f SEM after subtraction of background staining and applies to the Thy-1.2+ population only. The percentages of Thy-1.2+ cells in the Tcell-enriched splenocytes were 85.09 k 2.24 for SWR, 85.36 t 1.78 for NZB, 86.81 k 2.06 for young SNF, and 89.21 f 2.51 for old SNFl mice.

c) In this table data for Vp8+ and CD4+ and CD8+ cells are from all 19 old SNFl mice tested, but for Vg6, n = 17; for Vgll , n = 14 and for Vg17a, n = 18 because, only those SNFl mice that had deleted or had reduced levels of expression of these forbidden autoreactiveTcR are shown; the rest is shown individually in Table 4.The nephritic SNFl animals from which theTcell lines were derived (Table 1) are not included here or in Table 4, since all of their spleen cells were used for culture.

1440 S. Adams, T. Zordan, K. Sainis and S. K. Datta Eur. J. Immunol. 1990. 20: 1435-1443

Table 4. Expansion of forbidden TcR Vg expressing cells in the spleens of certain nephritic SNF, animals

# I c I . 10"~ # I 2 (1.5s 2.09 .3 .59 #14 7.74 3.64 3.60 # 15 (1. I5 2.24 3.91 # 18 3.03 0.60 2 3 4 5.40 2.3.; j .07 + I 0 5.03 1.07 .3.0(1 7.77 2 . 3 5.49

a) For VB6 and Vp11, the results are calculated from two-color FCM, representative examples of which are shown in Fig. 3.The values are expressed as the percentage of Vp+ cells among the total Thy-1.2+ cells and among the CD4 and CD8-positive and -negative subsets of Thy-1.2+ cells in the Tcell-enriched population from the spleens of each mouse.

b) In this mouse, two-color staining was not done and results are from single-color staining. Upon removal of the CD4+ and CD8+ cells from the splenicTcell-enriched population of this mouse, the resulting DN cells containing 3.62% Vgl7a+ cells (Fig. 2) and 33.16% Thy-l+ cells (not shown).

gene. Among the splenic Thy-l+ cells of this C57BL/6 mouse, 8.2% cells were Vpll+, consistent with previous studies [ 14].The NZB strain had markedly reduced levels of Vpll+ cells, but the deletion was not complete. In the Vpll-high SNFl mice, the density of Vpll expression was in the range of the positive control, C57BL/6, and there was considerable enrichment of the Vgll+ cells in the SNFl DN populations obtained after removal of CD4+ and CD8+ cells. Only one of the 19 nephritic SNFl animals tested had a 6.9-fold increase in Vgl7a+ cells (Table 4). Although

two-color staining was not done in this case, a greater than 3-fold enrichment of the Vp17a+ cells occurred when CD4+ and CD8+ cells were removed from the T cell-enriched population and the intensity of Vgl7a staining in the DN cells of this SNFl mouse was in the range of the positive control, SWR (Table 4 and Fig. 2).There was no increase in the expression of the other forbiddenTcR in thisVgl7a-high SNFl animals. Finally,Vg8+ cells were absent or present at very low levels (0.06% k 0.05%) in the DN population of cells from the nephritic SNFl animals tested, in contrast to their whole T cell-enriched population (Fig. 2a).

LOG FLUORESCENCE INTENSITY

Figure 2. FCM histograms of T cell- enriched populations (WTor wholeT) and DN cells obtained from the spleens of various mouse strains as designated. The age and sex of the mice are described in Tables 3 and 4. The cells were stained by mAb specific for the different TcR Vg gene products as designated on the top of the respec- tive vertical panels. Background con- trols are shown by interrupted lines.

Eur. J. Immunol. 1990. 20: 1435-1443 Forbidden T cell receptors in autoimmune disease 1441

NZB (W.T) SNFl (W.T) SNFj (DN)

V5 6 FLUORESCENCE

NZB (W.T) SNFl (W.T) SNFl (DN)

t i

V B 11 FLUORESCENCE

4 Discussion

The primary antigen that triggers the autoimmune response in systemic lupus remains elusive due to the enormous complexity of this disease. In spite of this obstacle, we show that T h cell lines that are functionally relevant to the disease can be derived from lupus mice.The antigenic specificity of these Th cells are unknown, but only = 17% of 149 Tcell lines derived from eight SNFl mice with lupus nephritis had the functional capability of augmenting the production of the pathogenic variety of autoantibodies. Thus, autoanti- body-inducing ability is a property of only certain T h cells and not all of the Tcells that are activated in vivo in the lupus mice, a feature that is also found in human SLE [6, 221. These results demonstrate the specificity of our selection criteria for autoantibody-inducingT cell lines, but they are not unexpected, since the lupus mice are respond- ing constantly to all kinds of conventional antigens in their environment in addition to autoantigens. Thus the auto- reactive T cell clones, although significant, would be a minor component in the overall population of activated T cells in the peripheral lymphoid organs of mice with lupus.

Among the two varieties of autoantibody-inducing T h lines described here, the CD4+ set contained some of the lines with the strongest helper activity. Five of the twelve CD4+ Th lines used TcR genes of the Vg (8.2 or 8.3) family. From this sample (16/149 lines), it is not clear whether Vp8 is overrepresented among the anti-DNA autoantibody- inducing Th cells of SNFl mice. Members of the Vp8 family are predominantly expressed in the peripheral T cells of Vp8+ strains and they participate in TcR specific for many other conventional antigens. Nevertheless, what is of interest here is that the TcR Vp8 genes are lacking in the SWR but has been contributed by the NZB parent to the SNFl progeny. These Vp8 genes are then used by the auto- antibody-inducing T cells of SNFl mice. Perhaps TcR encoded by theVg8 gene family are potentially autoreactive in the SWR background (which the SNFl inherits), since SWR mice have deleted the Vg8 genes. To test this hypothesis, transgenic SWR mice with Vp8 genes intro-

Figure 3. Two-color immunofluores- cence dot plots of Tcell-enriched sple- nocytes (WIT) or of DN cells from the spleens of representative mice as designated. The red fluorescence of CD4 and CD8 staining is o n the Y axis and green fluorescence of TcR Vg6 (a) or Vpl l (b) staining is shown on the X axis. Based on background control staining, the dot blots were divided into four quadrants to identify un- stained cells (lower left): cells that stained with only one of the two fluor- ochromes (upper left and lower right) and cells staining with both fluoroch- romes (upper right).

duced into their germ line will have to be derived. Of course, the autoantibody-inducing capability of the Th cells could not be solely attributed to theVg8-encoded segments, the TcR V, gene usage and the V-J or V-D-J junctional sequences [complementarity-determining region (CDR3)] of the TcR expressed by these Th lines remain to be determined. Three of the other CD4+ Th lines, including the strong helper line 3A, were Vp4+, but again whether there is a preferential usage of this Vp gene, cannot be concluded from this sample. These studies demonstrate some heterogeneity of TcR Vp genes used by IgG anti-DNA autoantibody-inducing T h cells from nephritic SNFl mice. The spontaneous autoimmune response in systemic lupus is much more complex than that in organ-specific or experi- mentally induced autoimmune diseases. Therefore, more than one family of TcR may be utilized by the lupus autoantibody-inducing T h cells. Nevertheless, the CDR3 regions of theseTcR that would bind to an autoantigen may be more restricted in structure. Previous studies had shown that some of these CD4+ Th lines (1A and 3A) of SNFl mice preferentially respond to unique F1 hybrid MHC class I1 determinants and the pathogenic autoantibody-inducing help of these Th lines is selective, since they markedly augment the production of cationic, IgG antibodies to ssDNA and dsDNA but not total polyclonal IgG ([6], Table 1). The F1 hybrid Ia molecules could be presenting some endogenous peptides more efficiently to trigger these autoreactive Th cells or they could affect the thymic elimination/selection processes that deletes autoreactive T cell clones early in ontogeny [lo]. Indeed, heterozygosity of MHC class I1 genes is necessary for the development of accelerated lupus nephritis in the NZB X SWR and NZB x NZW crosses [9, 34-36].

In addition to the CD4+ Th cells, a population of DN (CD4-CD8-) Th cells that preferentially augment the production of pathogenic autoantibodies are also expanded in the nephritic SNFl mice and several other lupus strains [5 , 61. Here we show that almost all of the autoantibody- inducing DN T h lines that could be established in vitro from nephritic SNFl mice expressed “forbidden” or strongly autoreactive TcR Vg molecules (Vp6 and Vp8.1). These DN

1442 S. Adams, T. Zordan, K. Sainis and S. K. Datta Eur. J. Immunol. 1990.20: 1435-1443

Th cells are not functionally inactivated or anergic although they express the autoreactiveTcR, since they are capable of proliferating in response to syngeneic APC and augmenting the production of IgG anti-DNA autoantibodies by syn- geneic B cells. The DN Th line 8.48 that expresses the anti-Mls TcR Vg6, did not respond to the SWR parental strain’s APC as expected, since the SWR APC cannot present MIS antigens [ l l , 121. However, the other Vg6+ DN Th line 4B proliferated significantly to SWR APC. The 4B Th cells, in contrast to 8.48 might possess different TcR v, or CDR3 regions that are specific for other SWR antigens. Nevertheless, both of these autoantibody-inducing Th lines proliferated markedly when cultured with syngeneic SNFl APC. Moreover, DN cells expressing such forbiddenTcR Vg gene products (Vg6, 11 or 17a) were also found to be expanded in the spleens of some but not all of the nephritic SNFl animals examined here. Thus the presence of forbid- denTcR-bearing DN cells in the nephritic SNFl mice could be a stochastic phenomenon or these cells might be detectable only at certain time points in these animals. Although one of the DN Th lines expressed Vg8.1, the DN population in most of the SNFl mice tested lacked Vg8+ cells which are predominant among the whole T cell population of these mice. Thus the expanded DN popula- tion of nephritic SNFl mice may mostly consist of other autoreactiveTcR. In a recent study, an expansion of Tcells bearing strongly autoreactive TcR could not be detected in any SNFl mice with nephritis, probably because the Tcells were too rigorously purified in that study by two cycles of panning followed by nylon wool columns [37]. This type of purification would result in a loss of the autoreactiveT cells that are activated in vivo in SLE, since activated T cells tend to adhere to many surfaces [5, 6, 221. Moreover, the DN T cell population that is expanded in SLE [5,6,22] was not examined in that study and functional activity of the T cells was not analyzed [37]. Herein, we found that the frequency of T cells expressing one of the forbidden TcR, namely Vgll, was also increased occasionally among the CD4+ or CD8+ Tcells of nephritic SNFl mice (Fig. 3b and Table 4). This result could be due to the leakiness of the thymic deletion process for Vpll+ T cells [14] and these peripheral Vpll+ Tcells could be mostly CD8+. Although thymic elimination of autoreactive cells probably occurs at the CD4+CD8+ stage [28, 38, 391, there may be a peripheral mechanism for eliminating autoreactive T cells [40]. SinceVpll reacts against class I1 (I-E) molecules, any CD8+ V$l+ Tcells escaping to the periphery would not be selected against as strongly as the highly autoreactive CD4+ Vgll+ cells [28, 411. The expression of forbidden autoreac- tive TcR by the DN T h cells of SNFl mice suggests an abnormality in the thymic deletiodselection processes in this strain with lupus. Indeed, thymuses of mice with lupus manifest severe abnormalities with age [42,43] and neo- natal thymectomy leads to the emergence of such forbidden T cells in the periphery of normal mice, although function- al studies with those T cells have not yet been performed [40]. Moreover, the strongly autoreactive thymocyte pre- cursors of these peripheral DN T h cells could have escaped thymic deletion by their failure to express CD4 (and/or CD8) antigens, as suggested by other studies [28, 38, 391. Alternatively, the DN pathogenic autoantibody-inducing Th cells could arise from down-regulation of CD4 antigens in vivo, an endogenous mechanism to diminish the intensity of autoreactivity. Similar types of DN Tcells expressing an autoreactive TcR gene have been detected in transgenic

mice, but those DN cells appear to be functionally inert in contrast to the SNFl T h cells [28,44]. Pathogenic anti-DNA autoantibody-inducing, TcR a/fi expressing DN Th cells are also expanded in humans with lupus nephritis [22]. Those Th cells of lupus patients probably also express forbidden TcR like the SNF1-derived DN cells, although such auto- reactive TcR have not yet been identified in humans.

These functionally competent, autoantibody-inducing DN cells of non-lpr lupus-prone mice described here and in previous studies [5, 61 are probably different from the TcR a@+ DN T cells that are massively expanded in the MRL-lpr lupus-prone strain [45] or those found in the thymuses of normal mice [46, 471. In contrast to the SNF1-derived DN Th, the MRL-lpr DN Tcells are function- ally inert [48], and they predominantly express TcR encoded by the Vg8 family, but none of the forbidden, autoreactive Vp8 genes including Vp8.1 [49-511. Thus in non-lpr lupus-prone mice such as the SNFl the DN Th cells expressing an autoreactive TcR probably play an important role in perpetuation of autoimmune disease, along with the pathogenic autoantibody-inducing CD4+ Th cells found in these mice [5-71. Moreover, the SNFI-derived DN Th lines have maintained a stable DN phenotype for over 1 year in culture, unlike other types of DN Tcells that have been described in normal mice [52].

Although the TcR expressed by these Th cells are somewhat heterogeneous, the pathogenic autoantibodies induced by these Th in SNFl mice are at least idiotypically restricted [3,4]. The pathogenic autoantibodies of (NZB x NZW)F1 mice are also idiotypically restricted, but they may not be as oligoclonal as was considered from the analysis of a random selection of autoantibodies [53-551. Nevertheless, evi- dence of a spectrotypically restricted, clonal expansion of the pathogenic autoantibodies is also seen in human SLE as in the nephritic SNFl mice [22].These observations suggest that a selected population of pathogenic autoantibody- producing B cells are being expanded by the Th cells in nephritic SNFl mice. Indeed the CD4+ and the DN Th cells induce the production of pathogenic anti-DNA autoanti- bodies only by B cells from older pre-nephritic SNFl mice.But they do not induce pathogenic autoantibodies when cultured with B cells from younger preautoimmune SNFl mice or with B cells from the NZB or SWR parents, even though the CD4+ autoreactive Th cells proliferate in response to such B cells [6]. Therefore, the younger SNFl mice and the parental strains may be deficient in the selected B cell population that is committed to produce the pathogenic autoantibodies and due to an intrinsic defect this pathogenic autoantibody-producing B cell population may be in a more susceptible or inducible state in the older SNFl mice.

The results in this report show that functionally character- ized Tcells that are involved in autoimmune disease can express forbidden autoreactive TcR. The evidence is direct, since this work involves autoantibody-inducing Th cell lines unlike the studies with neonatal thymectomy or with cyclosporine [40,56] where the potential of Tcells bearing forbidden receptors to induce autoimmunity has not been demonstrated. These pathogenic autoantibody-inducing lines would become important tools for elucidating the underlying mechanism of lupus.

Received December 18, 1989; in revised form March 12, 1990.

Eur. J. Immunol. 1990. 20: 1435-1443 Forbidden T cell receptors in autoimmune diseasc I443

30 Buerstedde, J.-M., Pease, L. R., Bell, M. I?, Nilson, A. E., Buerstedde, G., Murphy, D. and McKean, D. J., J. Exp. Med. 1988. 167: 473.

31 Bluestone, J. A., Pardoll, D., Sharrow, S. 0. and Fowlkes. B. J., Nature 1987. 326: 82.

32 Cantrell, D. A. and Smith, K., J. Exp. Med. 1983. 158: 1895.

33 Wade, T., Bill, J., Marrack, P. C., Palmer, E. and Kapplcr, J.W.,J. Immunol. 1988. 141: 2165.

34 Hiroshe, S., Nagashawa, R., Sekikawa, I., Hamaoki. M., Ishida, Y., Sato, H. and Shirai, T., J. Exp. Med. 1983. 158: 228.

35 Kotzin, B. L. and Palmer, E., J. Exp. Med. 1987. 165: 1237. 36 Schiffenbauer, J., McCarthy, D. M., Nygard, N. R. , Woulfc, S.

L., Didier, D. K. and Schwartz, B. D., J. Exp. Med. 1989. 170: 971.

37 Kotzin, B. L., Kappler, J.W., Marrack, I? C. and Herron, L. R.. J. Immunol. 1989. 143: 89.

38 Fowlkes, B. J., Schwartz, R. H. and Pardoll, D. M., Nature 1988. 334: 620.

39 Kisielow, P., Bluthman, H., Staerz, U. D., Steinmetz, M. and Von Boehmer, H., Nature 1988. 333: 742.

40 Smith, H., Chen, I.-M., Kubo, R. and Tung, K. S. K., Science 1989. 245: 749.

41 Kanagawa, 0. and Maki, R., J. Exp. Med. 1989. 170: 901. 42 Gershwin, M. E., Ikeda, R. M., Kruse, W. L., Wilson, F.,

Shifrine, M. and Spangler, W., J. Immunol. 1978. 120: 971. 43 Andrews, B. S., Eisenberg, R. A.,Theofilopoulos, A. N., Izui,

S.,Wilson, C. S., McConahey, P. J., Murphy, E. D., Roths, J. B. and Dixon, F. J., J. Exp. Med. 1978. 148: 1198.

44 Sha, W. C., Nelson, C. A., Newberry, R. D., Kranz, D. M., Russell, J. H. and Loh, D.Y., Nature 1988. 336: 73.

45 Morse,H.C.,III,Davidson,W. E,Yetter, R. A.,Murphy,E. D.. Roths, J. B. and Coffman, R. L., J. Immunol. 1982. 129: 2612.

46 Budd, R. C., Miesche, G. C., Howe, R. C., Lees, R. K., Bron, C. and MacDonald, H. R. , J. Exp. Med. 1987. 166: 577.

47 Fowlkes,B. J.,Kruisbeek, A . M.,Ton-That,H.,Weston, M. A., Colligan, J. E., Schwartz, R. H. and Pardoll, D. M., Nature 1988. 329: 251.

48 Davignon, J. L., Budd, R. C., Ceredig, R., Piguet, I? F., MacDonald, H. R. , Cerottini, J. C., Vassali, P. and Izui, S., J. Immunol. 1985. 135: 2423.

49 Singer, P. A , , McEvilly, R. J., Noonan, D. J., Dixon, F. J. and Theofilopoulos, A . N., Proc. Natl. Acad. Sci. USA 1986. 83: 7018.

50 Kotzin, B. L., Babcock, S. K. and Herron, L. R. , J . Exp. Med. 1988. 168: 2221.

51 Singer, P. A , , Balderas, R. S., McEvilly, R. J., Bobardt, M. and Theofilopoulos, A . N., J. Exp. Med. 1989. 170: 1869.

52 Morisset, J.,Trannoy, E., DeTalance, A., Spinella, S., DebrC, P. Godet, P. and Seman, M., Eur. J. Immunol. 1988. 18: 387.

53 Hahn, B. H. and Ebling, F. M., J. Immunol. 1987. 138: 2110.

54 Schlomchik, M. J., Aucoin, A. H., Pisetsky, D. S. and Weigert, M. G., Proc. Natl. Acad. Sci. USA 1987. 84: 9150.

55 Panosian-Sahakian, N. , Klotz, J. L., Ebling, F., Kroncnberg, M. and Hahn, B., J. Immunol. 1989. 142: 4500.

56 Gao, E.-K., Lo, D., Cheney, R., Kanagawa, 0. and Sprent, J.. Nature 1988. 336: 176.

5 References

1 Eastcott, J.W., Schwartz, R. S. and Datta, S. K., J. Immunol. 1983. 131: 2232.

2 Gavalchin, J., Nicklas, J. A.,Eastcott, J. W., Madaio, M. I?, Stollar, B. D., Schwartz, R. S. and Datta, S. K., J. Immunol. 1985. 134: 885.

3 Gavalchin, J., Seder, R. A. and Datta, S. K., J. Imrnunol. 1987. 138: 128.

4 Gavalchin, J. and Datta, S. K., J. Immunol. 1987. 138: 138. 5 Datta, S. K., Patel, H. and Berry, D., J. Exp. Med. 1987. 165:

6 Sainis, K. and Datta, S. K., J. Immunol. 1988. 140: 2215. 7 Datta. S. K., Clin. Immunol. Immunopathol. 1989. 51: 141. 8 Bchlke, M., Chou, H., Huppi, K. and Loh, D.Y., Proc. Natl.

9 Ghatak, S., Sainis, K., Owen, F. L. and Datta, S. K., Proc. Natl.

10 Kappler, J.W., Roehm, N. and Marrack, P., Cell 1987. 49:

1252.

Acad. Sci. USA 1986. 83: 767.

Acad. Sci. USA 1987. 84: 6850.

-79 L/3.

11 Kappler, J.W., Staerz, U.,White, J. and Marrack, I? C., Nature 1988. .?32: 35.

12 MacDonald, H. R., Schneider, R., Lees, R. K., Howe, R.C., Acha-Orbea, J., Festenstein, H., Zinkernagel, R. M. and Hengartner, H., Nature 1988. 332: 40.

13 Pullen, A. M., Marrack, I? and Kappler, J.W., Nature 1988.335: 796.

14 Bill, J., Kanagawa, O.,Woodland, D. L. and Palmer, E., J. Exp. Med. 1989. 169: 1405.

15 Vacchio, M. S. and Hodes, R. J., J. Exp. Med. 1989. 170: 1335.

16 Happ, M. I?, Woodland, D. L. and Palmer, E., Proc. Natl. Acad. Sci. USA 1989. 86: 6293.

17 Fry, M. A., Cotterman, M. M. and Matis, L. A . , J . Immunol. 1989. 143: 2723.

18 Anderson, G. D., Banerjee, S. and David, C. A., J. Immunol. 1989. 143: 3757.

19 Staerz, U., Rammensee, H.-G., Benedetto, J. and Bevan, M., J. Immunol. 1985. 134: 3994.

20 Haskins, K., Hannum, C., White, J., Roehm, N., Kubo, R., Kappler, J. and Marrack, P., J. Exp. Med. 1984. 160: 452.

21 Payne, J., Huber, B. T., Cannon, N. A., Schneider, R., Schilham, M. W., Acha-Orbea, H., MacDonald, H. R. and Hengartner, H., Proc. Natl. Acad. Sci. USA 1988. 85: 7695.

22 Shivakumar, S., Tsokos. G. C. and Datta, S. K., J. Immunol. 1989. 143: 103.

23 Auffray, C. and Rougeon, F., Eur. J. Biochem. 1980. 107: 303.

24 Maniatis, T., Fritsch, E. F. and Sambrook, J., Molecular Cloning: A laboratory manual, Cold Spring Harbor Laborato- ry, Cold Spring Harbor 1982.

25 Malissen, M., Minard, K., Mijolsness, S., Kronenberg, M., Goverman, J., Hunkapillar,T., Prystowsky, M. B. ,Yoshikai,Y., Fitch, F., Mak,T. W. and Hood, L., Cell 1984. 37: 1101.

26 Ceredig. R., Lowenthal, J. W., Nabholz, M. and MacDonald, H. R., Nature 1985. 314: 98.

27 Sarmicnto, M., Glasebrook, A. L. andFitch, F.W., J. Immunol. 1980. 125: 2665.

28 Teh, H.-S.. Kishi, H., Scott, B. and Von Boehmer, H., J. Exp. Med. 1989. 169: 795.

29 Scott, B., Bluthmann, H., Teh, H.-S. and Von Boehmer, H., Nature 1989. 338: 591.