Significant frequency of cytotoxic T lymphocyte precursor cells specific for TNP-modified allogeneic cells in normal lymphocytes

CELLULAR

IMMUNOLOGY

103,27-40 (1986)

Significant Frequency of Cytotoxic T Lymphocyte Precursor Cells Specific for TNP-Modified Allogeneic Cells in Normal Lymphocytes SHIROAIZAWA Division of Physiology and Pathology, National Institute ofRadiological Sciences,Chiba 260 Japan Received December 5,1985; acceptedJune 16.1986 It was tested whether the cytotoxic T-lymphocyte precursor (CLP) repertoire in normal mice is biased toward recognizing foreign antigen in association with self H-2 as opposed to allogeneic H-2. The frequencies of CLPs in normal mice (H-2”‘? specitic for TNP-modified syngeneic and TNP-modified allogeneic cells have been compared by limiting dilution analysis. Normal spleen cells were cultured at a limiting dilution with TNP-modified (TNP-self) or TNP-modified allogeneic (TNF-allo) stimulator cells. Cultures were split into four aliquots and assayedagainst TNP-self, TNP-allo, unmodified syngeneic, and unmodified allogeneic Concanavalin A blast targets and classified for cytotoxic activity directed against TNP-self, TNP-allo, and al10 H-2 determinants. In disagreement with our expectations from the literature, the frequencies of CLPs in H-2* and H-2d responder cells recognizing TN&modified H-2k were higher than the frequencies of CLPs recognizing TNP-self. There was no clear preference for TNP-self in the case of H-2” responder and H-2d allogeneic cells, nor vice versa. Only in the case of H-2’ responder cells was there a distinct preference for TNP-self. The signiftcance of a considerable number of TNP-specitic, allo H-2-restricted CLPs in normal lymphocytes is discussed. o 1986 Academic Press,Inc.

INTRODUCTION Most T cells recognize foreign antigens in association with gene products of the MHC’ gene. Regarding the acquisition of MHC restriction specificity of T cells during ontogeny, numerous studies using bone marrow chimeras have indicated that T cells which undergo differentiation within the H-2-incompatible host/thymus become preferentially restricted to the H-2 type of host/thymus (l-3). This result has then led to the thesis that H-2 restriction specificity of T cells is generated during ontogeny as a result of the positive selection of T cells restricted to the H-2 type of host/thymus. However, the host influence on the expression of H-2 restriction specificity of T cells generated in bone marrow chimeras is not always absolute. For example, the results obtained from experiments with fully allogeneic chimeras (4-6) have shown that T cells generated in such chimeras can manifest restriction specificity for both donor- and host-type H-2. Thus, the precise mechanism for the host influ’ Abbreviations used: MHC, major hi&compatibility complex; CTL, cytotoxic T lymphocyte; CLP, cytotoxic T-lymphocyte precursor, TNP, trinitrophenyl; Con A SN, Concanavalin A-stimulated spleen cell culture supematant. 27 0008-8749186$3.00 Copyright Q 1986 by Academic Press,Inc. All rigbt9 of reproduction in any form merwd.

28

SHIRO

AIZAWA

ence on the expression of H-2 restriction specificity of T cells has yet to be determined. Another potentially rewarding approach for examining this issue would be to find out whether normal T lymphocytes can respond to foreign antigens in association with the allogeneic H-2 determinant. According to the positive selection model (l3), the presence of antigen-specific, allogeneic H-Zrestricted T cells in normal lymphocytes is not indicated. However, distinguishing the response against the foreign antigen from that against the allogeneic H-2 determinant poses a major problem in this type of approach. Most studies to date have used special procedures to remove the alloreactive cells prior to determining the response against antigen-modified allogeneic cells (7-13). However, the results obtained so far have often been contradictory. Even in the experiments concerning H-2 restriction specificity of hapten(TNP)specific cytotoxic T cells, some studies (7, 8) have indicated a strong preference for self H-2 over al10 H-2 as the restriction specificity, whereas others (9, 12) have indicated that normal T cells are significantly or even completely reactive to TNP-modified allogeneic cells. Recently, Good and Nossal(14) described another strategy for looking directly at the response to TNP-modified allogeneic cells in a limiting dilution culture. Microcultures of normal T cells were set up with a TNP-modified allogeneic stimulator, and then split into equal portions and assayedagainst TNP-modified and unmodified syngeneic targets and TNP-modified and unmodified allogeneic targets. They found that CLPs reactive to hapten-modified allogeneic cells do exist, but that their frequency is about one-eighth of that of CLPs reactive to hapten-modified syngeneic cells. However, using a similar limiting dilution culture system, we ( 15) detected that the number of CLPs reactive to TNP-modified ahogeneic cells was not always less than the CLPs reactive to TNP-modified syngeneic cells in the different H-2 haplotype combinations. In addition, Reimann et al. ( 16,17) also observed a large number of CTLs reactive to TNP-modified or virus-infected allogeneic cells to be induced from normal spleen cells in a limiting dilution culture. Thus, a definitive study on the relative frequencies of al10 vs self H-2-restricted T cells should be important for our understanding of the mechanism for the generation of H-2 restriction specificity of T cells. H-2 restriction specificity of TNP-specific CTLs can be definitely studied by discriminating between TNP-specific, H-Zrestricted and unrestricted CTL clones in a split-well limiting dilution assay,though a significant cross-reactivity of TNP-specific CTL populations generated in bulk cultures might make ambiguous the analysis for H-2 restriction specificity of TNP-specific CTLs. In addition, it has been clearly shown that H-2 restriction specificity of TNP-specific CLPs differentiated intrathymitally in some H-Zincompatible bone marrow chimeras can be greatly influenced by the host environment ( 18, 19). The high primary response for the induction of TNPspecific CTLs may allow the quantitative comparison of self H-2-restricted T cells with al10 H-Zrestricted T cells in normal T lymphocytes without the depletion of alloreactive T cells prior to the in vitro assay.Becauseof these reasons,the frequencies of CLPs reactive to TNP-self were extensively compared with that of CLPs reactive to TNP-allo. The results presented here show the first convincing data that there is no consistent self preference for H-2 restriction specificity of TNP-specific CLPs in normal T lymphocytes among six possible combinations with three haplotypes (H-26.“d).

LARGE NUMBER OF ALL0 H-2 RESTRICTED T CELLS

29

MATERIALS AND METHODS Mice. C57BL/10J(B10,H-26), BlO.BR/SgSn(BlO.BR,H-29, BlO.D2/Sn(BlO.D2, H-29, C3H/HeJ(C3H,H-29, and DBA/2J(DBA,H-2d) were purchased from the Jackson Laboratory, Bar Harbor, Maine, at the Ontario Cancer Institute, or bred at the Animal Production Facility of this institute. All mice were used at 6 to 12 weeks of age. Mixed-lymphocyte culture. All spleen cells used in the limiting dilution culture were passed through a Sephadex G-10 column. The culture medium was aMEM supplemented with 10% fetal calf serum, 10 mM Hepes buffer, and 5 X lo-’ M 2-mercaptoethanol. Limited numbers (125-2000 cells/well) of responder cells were cultured in the presence of 2 X lo5 irradiated (1500 R) syngeneic cells with either 6 X 10’ irradiated, TNP-modified syngeneic or TNP-modified allogeneic stimulator cells in 0.2 ml round-bottomed wells. Spleen cells were modified with 7.5 miW trinitrobenzene sulfonic acid according to the method of Shearer (20). All microcultures were supplemented with an optimal concentration of the supernatant of Concanavalin A-stimulated rat spleen cell culture (Con A SN). Limiting dilution cultures were incubated at 37°C in a humidified atmosphere of 8% COZ. One week later, 0. I ml was removed and replaced with 0.1 ml of fresh medium containing Con A SN, and irradiated, TNP-modified spleen cells. After a further 4 days of culture, microcultures were assessedfor cytotoxic assay.Microcultures were usually set up at three different concentrations of responder cells with 60 or 120 replicates, respectively, for each stimulator. Peripheral wells were filled with culture medium alone, and were not used for cultures to avoid the possible influence of evaporation. The concentrated Con A-free Con A SN wasprepared from the culture of 5 X 106/ml Lewis rat spleencells stimulated with 2 &ml Con A for 30 hr in aMEM containing 5 X 10e5MZmercaptoethanol(2 1). Cytotoxic assay.Cultures were split into four aliquots and assayedfor cytotoxicity against TNP (7.5 a)-modified syngeneic, TNP-modified allogeneic, unmodified syngeneic, or unmodified allogeneic Con A blast targets. After a 4-hr incubation with 3000 “0-labeled target cells in 0.2 ml, 0.1 ml of supematant was removed for yray counting. Percentage specific lysis was calculated as [(experimental cpm released - spontaneous cpm released)/(total cpm released - spontaneous cpm released)] X 100. Total releasable counts were measured by incubation of target cells with 5% acetic acid. Spontaneous releasable counts were taken in cultures containing only irradiated stimulator cells. a-Methyl-D-mannoside (25 mM, final concentration) was added to the target cell suspensions. The precursor frequency of CLPs was calculated according to the procedure of Porter and Berry (22). Treatment with antibody and complement. Spleen cells (1 X 10’ cells/ml) were incubated for 60 min on ice with either monoclonal anti-Thy- 1.2 (obtained from R. A. Phillips, Ontario Cancer Institute, final concentration of l/2) or monoclonal anti-Lyt-2.2 (Ascites, Cedarlane, final concentration of l/50). They were centrifuged, resuspended in 1 ml of rabbit complement (Low-Tox, Cedarlane) at a final concentration of 1/ 10 and incubated for 60 min at 37°C. The second treatment was performed with a mixture of antibody and complement for 60 min at 37°C. They then were washed three times and cultured in the limiting dilution assay. RESULTS

Properties of TNP-spec$c CTLs generatedin limiting dilution with TNP-mod$ed syngeneicstimulator. Normal spleen cells were cultured in a limiting dilution with

30

SHIRO AIZAWA

0

50 Percent

lysis (610)

0

50 Percent

lysis(TNP-BKIBR)

FIG. 1. Cytotoxic response to TNP-modified syngeneic stimulator in the limiting dilution culture. B 10 spleen cells were cultured at a dilution of 1000 cells/well with 6 X lo5 TNP-modified, irradiated B 10 spleen cells in 60 wells. After 11 days, each microculture was split into four aliquots and assayedagainst TNPB 10, TNP-B lO.BR, unmodified B 10, and BlO.BR targets. The dotted line represents a value of 3 standard deviations (SD) above the mean spontaneous release. Closed circles show cytotoxic responseswith a signifxant cross-reactivity to the TNP-B lO.BR target.

irradiated, TNP-modified syngeneic spleen cells. After 11 days of culture with a single restimulation, each microculture was split into four aliquots and assayedagainst the following targets: TNP-self, TNP-alIo, unmodified self, and unmodified allo. In Fig. 1, B 10 spleen cells were cultured in 60 wells at a dilution of 1000 cells/well with TNPB 10 stimulator. The killing activity of each microculture against the TNP-self(TNPB 10) target was plotted for those against the unmodified self (B 10) or TNP-alIo (TNPB 1O.BR) targets. Microcultures with lytic activity against the TNP-self but not against the unmodified self target indicated the generation of TNP-specific CTLs in the restimulated culture with a limited number (lessthan 2000 cells/well) of normal spleen cells as responder cells. A microculture was considered to be positive in CTL response to TNP-self if specific percent lysis was greater than 20% against the TNP-self target and less than three standard deviations above the mean spontaneous release against the unmodified self target. No generation of CTL with a killing activity of more than three standard deviations above the spontaneous release against the unmodified self target was observed throughout this study. Furthermore, TNP-specific CTLs were divided into two types of TNP-specific CTLs with/without lytic activity against the TNP-al10 target (Fig. 1B). The proportion of TNP-specific CTL clones which crossreacted with the TNP-allo target was variable, depending on the H-2 haplotype of responder cells (20 to 40% of TNP-specific CTLs of H-2’,’ haplotype strains and 10 to 20% of TNP-specific CTLs of H-2k haplotype strain). Properties of TNP-specijic CTLs generated in limiting dilution culture with TNPmodijed allogeneic stimulator. Normal spleen cells were cultured with a limiting dilution with TNP-al10 stimulator. Cytotoxic activity of each microculture was assayed against TNP-self, TNP-allo, unmodified self, and unmodified allo. Two types of CTLs were expected to be induced clonally in this limiting dilution culture system. One type specific for alloantigen could lyse both unmodified al10 and TNP-allo targets. The other was expected to be specific for TNP-determinant, and lysed TNP-allo

LARGE NUMBER OF ALL0 H-2 RESTRICTED T CELLS

50

0

Percent lysis

(BlO.BR )

0 Percent

31

50 lysis (TNP-El01

FIG. 2. Cytotoxic response to TNP-modified allogeneic stimulator in the limiting dilution culture. BlO spleen cells were cultured at a dilution of 250 cells/well with TNP-BlO.BR stimulator in 60 wells. They were cultured and assayed in the same manner as Fig. 1. Closed circles show cytotoxic responseswith a significant cross-reactivity to the TlW-B 10 target.

but not unmodified allo targets. Figure 2 clearly shows the generation of two such types of CTL clones in the limiting dilution culture with TNP-al10 stimulator. A microculture was considered to be positive in CTL responseto TNP-allo if the specific percentage lysis was greater than 20% against the TNP-allo target but less than three standard deviations above the mean spontaneous releaseagainst the unmodified al10 target. The TNP-specific CTL clones induced with TNP-allo stimulator were also divided into two types of TNP-specific CTLs with/without lytic activity against the TNP-self target (Fig. 2B). The proportion of TNP-specific CTL clones cross-reacting with the TNP-self target was also variable, depending on the H-2 haplotype of stimulator cells (20 to 40% of TNP-specific CTL clones induced by TNP-modified H-2bVd stimulator and 10 to 20% of those induced by TNP-modified H-2k stimulator, irrespective of responder haplotype). Frequencies of TNP-spec$c CLPs. To estimate the frequencies of TNP-specific CLPs, the limiting dilution culture was established with at least three different amounts of responder cells, usually by twofold serial dilution. Although all microcultures were split into four aliquots and assayed against the TNP-self, TNP-allo, unmodified self, and unmodified al10targets in each experiment, microcultures positive for TNP-specific CTLs either restricted or unrestricted to the H-2 haplotype of the stimulator were combined for the most reliable estimation of frequencies of TNPspecific CLPs. Figure 3 demonstrates that CTL responses in the limiting dilution assayused in this study follow single-hit zero-order kinetics, showing that only CLP is limiting. In addition, the relative frequency of CLP reactive to either TNP-self, TNP-allo, or unmodified allo is seen to change according to the combination of responder and allogeneic strains. As shown in Fig. 3A, in the case of BlO(H-2’) responder and BlO.BR(H-29 allogeneic strains, the frequency of CLPs reactive to TNP-allo was significantly higher than that of CLPs reactive to TNP-self. In the combination of B10(H-2b) responder and B lO.D2(H-2d) allogeneic stimulator strains, the frequency of alloreactive CLPs was the highest among the three types of CLPs. The frequency of CLPs reactive to TNP-self was similar to that of CLPs reactive to TNP-

32

SHIRO AIZAWA

BlO anti-TNP-610

610 anti-TNP-61O.ER

810 antkTNP-BlO

B,O anti-BtO.DZ

A .3 '

I 125

250

500

FOG.3. Limiting dilution analysis for the frequency of CLP reactive to TNP-self, TNP-allo, and unmodified allo. The number of responder cells per well is plotted against the fraction of negative wells for the responseto TNP-self (A) in the culture with the TNP-modified syngeneic stimulator, and for the response to TNP-allo (0) or unmodified allo (0) in the culture with the TNP-modified allogeneic stimulator. (A) B 10 responder and B lO.BR allogeneic strains, (B) B 10 responder and B 10.D2 allogeneic strains, and (C) BlO.BR responder and BlO.D2 allogeneic strains. Frequencies (95% confidence limits) of TNP-specific CLPs:(A)BlOanti-TNP-BlO, l/3432(1/2591-1/4545);BlOantiTNP-BlO.BR, l/1550(1/1113-l/2158), (B) BlO anti-TNP-BlO, l/3241 (l/2447-1/4293); BlO anti-TNPBlO.D2, l/2343 (l/1550-1/3542), (C) BlO.BRanti-TNP-BlO.BR, l/467 (l/375-1/581); BIO.BRanti-TNP-BlO.D2, l/4732 (l/2902-1/7688).

allo. On the other hand, in the combination of BlO.BR(H-29 responder and BlO.D2(H-2d) ahogeneic stimulator strains, there was a profound self H-2 preference for the induction of TNP-specific CTLs (Fig. 3C). Since the relative frequency of CLPs reactive to either TNP-self or TNP-al10 varies according to the strain combinations, these frequencies were extensively studied in the six combinations possible with three H-2 haplotypes (H-2b*kd).As summarized in

LARGE NUMBER OF ALL0 H-2 RESTRICTED T CELLS

33

TABLE 1 Frequency of CLPs Reactive to TNP-Modified Syngeneic or Allogeneic Cells Frequency of CLPs reactive to TNP-modified cells’

Strain combination Responder BlO.BR BlO.BR BlO BlO.D2 BlO BlO.D2 DBA/2 B 10 thymocyte

All0

Self

Al10

Allo/selfb

BlO.D2 BlO BlO.D2 BlO BlO.BR BlO.BR C3H BlO.BR

l/627 l/950 112,531 l/5,848 l/2,952 115,696 1110,466 1123,737

115,643 119,155 111,837 119,446 l/2,007 112,259 l/3,816 l/16,132

0.11 0.10 1.38 0.62 1.43 2.52 2.74 1.47

’ Data from a representative experiment. Experiments were done at least twice in each strain combination. ’ Ratio of the numbers of CLPs reactive to TNP-al10 to those reactive to TNP-self.

Table 1, in the case of BlO.BR(H-27 responder and either BlO(H-2’) or BlO.D2(H2d)allogeneic stimulator strains, the frequency of CLPs reactive to TNP-allo was oneeighth to one-ninth of that of CLPs reactive to TNP-self, showing a profound self H2 preference for the induction of TNP-specific CLPs. In the combination of BlO responder and B 1O.D2 allogeneic stimulator strains and vice versa, no clear preference for CLPs reactive to TNP-self was seen. However, in the caseof B 10 or B 1O.D2 responder and BlO.BR allogeneic stimulator strains, the frequency of CLPs reactive to TNP-al10 was rather higher than that of CLPs reactive to TNP-self. In addition, a similar result was obtained in the case of DBA/2(H-2d) responder and C3H(H-23 allogeneic stimulator strains, showing that the lack of self preference is primarily dependent on the combination with H-2 haplotypes. As mentioned above, the frequencies of TNP-specific CLPs were calculated from the summation of the two types of TNP-specific CTL clones, restricted and unrestricted, to the haplotype of the stimulators. The fraction of TNP-specific clones unrestricted to the H-2 haplotypes of the stimulators was 20 to 40% of TNP-specific CTL clones induced with the TNP-H-2b or TNP-H-2d stimulator and 10 to 20% of those induced with the TNP-H-2k stimulator. Thus, it was clear that if we could estimate the frequencies of TNP-specific CLPs restricted to either self H-2 or allo H-2, the preference for self H-2 observed in the case of H-2k responder cells and the allo H-2 preference seen in the H-2k allogeneic stimulator would become more profound. The results concerning this point will be discussed later. Generation of TNP-spec$c CTLs from thymocytes. The H-2 restriction specificity of TNP-specific CTL clones generated from thymocytes was investigated in order to determine whether allo H-2-restricted T cells can differentiate intrathymically in normal mice. As shown in Table 1, the frequency of TNP-specific CLPs in B 10 thymocytes reactive to TNP-B 10 or TNP-B 1O.BR stimulators was 5- to 1O-fold lower than that of CLPs in normal spleen cells, but the frequency ratio of CLPs reactive to TNP-self over TNP-al10 was quite similar to that seenin spleen cells. In addition, the H-2 restriction specificity of TNP-specific thymic CTL clones showed a pattern sim-

34

SHIRO AIZAWA TABLE 2 Lysis Patterns of TNP-Specific CTLs Generated with TNP-Moditkl

Syngeneic Stimulator

Lysis pattern4

Group

TNP-

TNP-

TNP-

BlO

BlO.BR

BlO.D2

BlO

-

-

+ +

-

-

+

(4 (a-1) (a-2) b-3)

+ + +

-

BlO.BR

Total

Total

+

No. of positive microcultures 100 72 21 7 100

+ +

(b) (4

BlO.D2

+ +

-

+

76 20 196(a+b+c)

Note.B 10 spleen cells were cultured at 1000 cells/well with TNI-B 10 stimulator. Data were obtained from 798 micrccultures from a total of seven experiments. Bach microculture was split into six aliquots and assayedagainst TNP-modified and unmodified B 10, B 1O.BR, and B 10.D2 targets. ’ Groups (a) and (a-l) were identilied as the cultures with lytic activities of more than 20% against TNPB 10 but less than 3 SD above mean spontaneous release against the other targets. In the other groups, the lytic activities of more than 20% against TNP-B 10 and more than 3 SD above mean spontaneous release against the other targets were considered positive.

ilar to TNP-specific splenic CTL clones in respect to the proportion of highly H-2restricted CTLs against the unrestricted CTLs (data not shown). This result indicated that TNP-specific, allo H-Zrestricted CLPs can efficiently differentiate within the thymus of normal mice. Thefrequencies of TNP-speciJic CLPs restricted to selfH-2 or allo H-2. In Table 1, the frequencies of CLPs reactive to TNP-self or TNP-al10 were determined from the summation of numbers of TNP-specific CTLs, restricted and unrestricted, to the H2 haplotype of the stimulator. To confirm the fact that there was no self preference for the H-2 restriction specificity of TNP-specific CTLs observed in some strain combinations, the relative frequencies of TNP-specific CLPs highly restricted to self H-2 or allo H-2 were studied in the caseof B 10 responder and B lO.BR allogeneic cells. For this purpose, B 10 responder spleen cells were cultured at a constant 1000 cells/ well with the TNP-self (TNP-BlO) stimulator and 250 cells/well with the TNP-al10 (TNP-BlO.BR) stimulator, respectively. Microcultures raised in the usual manner were then split into “six” ahquots and assayedagainst the TNP-modified and unmodified BlO, BlO.BR (relevant), and BlO.D2 (irrelevant) targets. TNP-specific CTLs induced in the microcultures with the TNP-self stimulator (Table 2) were first classified into three groups from their lytic patterns against TNP-B 10, TNP-B lO.BR, unmodified B 10, and unmodified B lO.BR targets. TNP-specific CTLs of group a were identified asbeing restricted to self H-2(H-2b), but those of group b were cross-reactive to the TNP-al10 (TNP-B 1O.BR) targets. TNP-specific CTLs of group c were shown to cross-react with the al10 determinant of the BlO.BR(H-29 target. TNP-specific, self H-Zrestricted CTLs of group a could be divided into a further three groups from their lytic patterns against TNP-modified or unmodified BlO.D2(H-2”) allogeneic

35

LARGE NUMBER OF ALL0 H-2 RESTRICTED T CELLS TABLE 3 Lysis Patterns of TNP-Specific CTLs Generated with TNP-Modified Allogeneic Stimulator Lysis pattern”

Group

(4 (d-1) (d-2) (d-3)

TNPBlO

TNPBlO.BR

-

+

-

+ + +

TNPBlO.D2 + +

BlO

BlO.BR

-

-

-

-

+

t

-

-

83 70 7 6

36 119 (d t e)

Total

(0

+

83

Total

(4

BlO.D2

No. of positive microcultures

-

+

-

t

157

Note. B 10 spleen cells were cultured at 250 cells/well with TNP-BlO.BR stimulator. Data were obtained from a total of 840 microcultures from seven experiments. Each microculture was assayed in the same manner as Table 2. ’ Groups (d) and (d- 1) were identified as the cultures with lytic activities of more than 20% against TNPBlO.BR but lessthan 3 SD above mean spontaneous release against the other targets. In the other groups, the lytic activities of more than 20% against TNP-B 1O.BR and more than 3 SD above mean spontaneous release against the other targets were considered positive.

targets. In this classification, TNP-specific CTLs of group a-1 were still classified as those restricted to self H-2, but those of group u-2 were shown to cross-react with the TNP-modified B lO.D2(H-2d) target. TNP-specific CTLs of group u-3 were identified as those cross-reactive with the allo determinant of the H-2d haplotype. In the microcultures with the TNP-allo (TNP-BlO.BR) stimulator, three types of CTLs were first classified from their lytic patterns against TNP-BlO, TNP-BlO.BR, unmodified BlO, and unmodified B lO.BR targets (Table 3). They consist of TNPspecific, allo H-Zrestricted CTLs (group d), TNP-specific, H-2-unrestricted CTLs (group e), and alloantigen-specific CTLs (groups). TNP-specific, al10 H-Zrestricted CTLs of group d could be divided into a further three groups according to their lytic patterns against TNP-modified and unmodified B 10.D2(H-2d) targets. TNP-specific CTLs of group d-l were still classified as those restricted to the al10 H-2(H-27 haplotype but those of group d-2 were shown to cross-react with the TNP-BlO.D2 target. TNP-specific CTLs of group d-3 were identified as those cross-reactive with the allo determinant of the H-2d haplotype. As shown in Tables 2 and 3, 100 microcultures were positive for the induction of CTLs specific for TNP-self in 798 microcultures with the TNP-self (TNP-BlO) stimulator at a dilution of 1000 cells/well responder cells, and 83 microcultures were positive for the induction of CTLs specific for TNP-allo in 840 microcultures with the TNP-aho (TNP-B lO.BR) stimulator at a dilution of 250 cells/well responder cells. From the number of microcultures positive for the induction of TNP-specific, H-2restricted CTLs (groups a and d) and the number of responder cells and microcultures established, the relative frequency of CLPs specific for TNP-al10 was about

36

SHIRO AIZAWA 610 anti-TNP-El0

0’ 0

” ld

BIO anti-TNP-BIOBR

’

0’

3xlo4105 Number

0 of cold

’ lo4

” 3x104 lot

targets

FIG. 4. Cold target competition in cytotoxic responsesagainst TNP-modified syngeneic and TNP-modified allogeneic targets. Cytotoxic T-cell clones specific for either TNP-self (BlO anti-TNP-B10; A, B) or TNP-al10 (BlO anti-TNP-BlO.BR, C, D) were induced in the limiting dilution and then expanded. Cytotoxic activity of each clone was assayedagainst 10005’Cr-labeled TNP-B 10or TNP-B lO.BR targets, respectively, in the presence of unlabeled TNP-BlO (0), TNP-BlO.BR (A), BlO (O), and BlO.BR (A) targets. Percentagesof specific lysis of TNP-BlO, TNP-BlO.BR, BlO, and BlO.BR targets in the absence of unlabeled targets were respectively (A) 69,6,5, and 1%; (B) 77,2,5, and 3%; (C) 1,37, -4, and -1%; (D) 2, 31,O,andO%.

threefold higher than that of CLPs specific for TNP-self in this strain combination. Thus, this result confirms the notion that there is no preference for CLPs specific for TNP-self in the combination of B 10 responder and B 1O.BR allogeneic strains. Similar data could also be obtained from a comparison of the frequencies of TNP-specific, self H-2- or allo H-Zrestricted CLPs strictly classified from their lytic patterns against six different targets including TNP-BlO.D2 and unmodified BlO.D2 targets (groups a-1 and d-1). Furthermore, it should be noted that a significant fraction of TNPspecific CTL clones induced with TNP-self and TNP-allo stimulators cross-reacted with other allo determinants (groups a-3, c, and d-3). Specificity of CTL clones specijic for TNP-self or TNP-allo for target cell recognition. A competitive experiment with cold target cells was performed to confkm the specificity of TNP-specific CTL clones at the stageof target cell recognition. Figure 4 shows the results of this experiment with CT’L clones specific for either TNP-self or TNP-allo. The lysis of the TNP-self target by CTL clones specitk for TNP-self was inhibited by the appropriate cold target (TNP-self) but not by unrelated cold targets.

LARGE NUMBER OF ALL0 H-2 RESTRICTED T CELLS

Percent

Percent

lysis

lysis

37

( BlO 1

( BlO.BR

1

FIG. 5. Sensitivity of TNP-specific cytotoxic T-cell precursors in B 10 spleen cells to treatment with antiThy-l .2 or anti-Lyt-2.2 antibody plus complement. BlO spleen cells treated with complement alone were cultured at a dilution of 2000 cells/well (54 wells) with TNP-BlO stimulator (A), or at 500 cells/well (60 wells) with TNF’-BlO.BR stimulator (D). Spleen cells treated with anti-Thy-l.2 (B, E) or anti-Lyt-2.2 (C, F) antibody plus complement were cultured at 4000 cells/well (60 wells) with TNP-BlO stimulator (B, C), or 1000 cells/well (60 wells) with TIW-BlO.BR stimulator (E, F).

In a similar manner, the lysis of the TNP-alIo target by CTL clones specific for TNPallo was specifically inhibited by the unlabeled TNP-allo target.

Sensitivity of CLPs to treatments with anti-Thy-l.2 and anti-Lyt-2.2 antibodies pZuscomplement.An experiment was performed to determine whether CLPs reactive to either TNP-self or TNP-al10 share the common surface markers characteristic of cytotoxic T-cell precursors (CLPs). BlO spleen cells were treated with monoclonal anti-Thy-l.2 or anti-Lyt-2.2 antibody plus complement and then assayedfor TNPspecific CTLs in the culture with either the TNP-self or the TNP-al10 stimulator. As shown in Fig. 5, CLPs reactive to either the TNP-self or the TNP-allo stimulator and CLPs reactive to alloantigens were distinctly sensitive to these treatments. DISCUSSION The limiting dilution assaydescribed in this paper allows us to study the frequencies of CLPs in normal lymphocytes reactive to TNP-modified allogeneic cells without the depletion of ahoantigen-reactive T cells prior to the culture. The data revealed

38

SHIRO AIZAWA

that the self preference for the H-2 restriction specificity of TNP-specific CLPs in normal lymphocytes was not consistently observed, rather surprisingly, a significant allo preference was seenin some H-2 haplotype combinations. In the limiting dilution culture system used in this study, TNP-specific CTLs were induced with a higher killing activity and higher frequency than those reported previously (12, 23). The effect of restimulation on the efficient generation of antigenspecific CTLs in limiting dilution cultures has also been shown by Christensen et al. (24). The higher killing activities of the CTL clones and the use of panels of four different targets allowed us to classify TNP-specific and alloantigen-specific CTL clones after TNP-modified allogeneic stimulation. As a result, this method permits us to study CTLs in normal lymphocytes specific for TNP-determinants expressed on allogeneic cells without the depletion of alloantigen-reactive T cells prior to the culture. The advantage of this method is that it minimizes the possible artifacts which might be caused by the manipulation involved in the depletion of alloreactive T cells in normal lymphocyte populations. Although the development of nonspecific killer cells has been reported in a limiting dilution culture (25), such cells were not observed in the present assay.Teh and Yu (26) reported that the induction of nonspecific killer cells could be greatly reduced by using a responder cell population depleted of nylon wool-adherent cells. Sephadex G-10 column-passed spleen cells were used in this study as a source for both responder and stimulator cells to avoid the generation of nonspecific killer cells. Among the studies on the existence of antigen-specific, allo H-2-restricted T cells in normal lymphocytes (7- 14), studies on their frequencies have been quite limited. Good and No&( 14) have shown that CLPs reactive to TNP-allo do exist, but their frequency is about one-eighth of that of CLPs reactive to TNP-self. However, they analyzed only one combination of CBA(H-29 responder and Balb/c(H-2d) allogeneic strains. A similar result was also observed in this study, showing the self preference of H-2 restriction specificity of TNP-specific CLPs in the same H-2 haplotype combination. Stockinger et al. (12) have reported that the frequency of TNP-specific, al10 H-2-restricted CLPs in CBA(H-27 and Balb/c(H-27 thymocytes is three- to ninefold lower than that of self H-Zrestricted CLPs. Thus, they observed significant numbers of al10 H-2-restricted CLPs differentiated intrathymically, but there was generally a self preference for H-2 restriction specificity. However, in the experiment (Table 1) there was no self preference in the combination of B 1O(H-2’) responder thymocytes and the BlO.BR(H-23 allogeneic strain. As shown in Table 1, since the H-2“ haplotype is the strongest of the three haplotypes (H-2bVk.d) used in this study for the induction of TNP-specific CTLs, self preference or the loss of self preference for the induction of TNP-specific CTLs may be primarily due to the particular haplotype combination of responder and allogeneic strains. The results reported here show some critical discrepancies in comparison to the numerous studies showing thymus/host influence on the generation of H-2 restriction specificity of T cells differentiated in bone marrow chimeras (l-3). For example, Kruisbeek et al. (18, 19) clearly showed the host influence on the expression of H-2 restriction specificity of TNP-specific CLPs differentiated intrathymically. They also observed that TNP-specific CTLs restricted to donor-type H-2 could be induced from spleen cells of bone marrow chimeras, suggesting that these TNPspecific CLPs are able to expand to detectable numbers in extra/post-thymic environments by encountering H-2 determinants expressedon donor-derived accessorycells which had repop-

LARGE NUMBER OF ALL0 H-2 RESTRICTED T CELLS

39

ulated the peripheral lymphohemopoietic tissues of chimeras ( 19). In the present paper, it has been shown that TNP-specific, allo H-2-restricted T cells can differentiate in normal mice without encountering aIlo H-2 determinants. The reason for this discrepancy is not yet clear, but it must be of considerable importance for our total understanding of the mechanism influencing the expression of H-2 restriction specificity of T cells. However, there are several possible explanations for the development of a significant number of allo H-Zrestricted CTLs. First, the likelihood exists that the restimulated cultures used in this study involve not only the continued clonal expansion of stimulated T-cell precursors but also the emergence of new clones of TNP-specific CTLs; i.e., allo H-Zrestricted CTLs might emerge from self H-2-restricted or unrestricted, TNP-specific CLPs, or alloreactive CLPs through a change of specificity during the restimulated culture in vitro. Reimann et al. (27,28) have suggestedthat the lytic patterns expressed by individual microcultures set up under limiting dilution conditions are not stable. Recently, they (16, 17) also observed a large number of TNP- or virus-specific, allo H-2-restricted CTLs arising from the restimulated limiting dilution culture of normal spleen cells, and emphasized the possibility that such CTLs might be generated during in vitro culture from alloreactive CLPs by somatic diversification. However, their results (16) obtained from the assay at an early stage (Day 7) of incubation and the sequential reassay of microcultures at different times during in vitro incubation did not show any significant instability of T-cell specificity. The instability of lytic patterns of individual cultures under limiting dilution conditions was also examined in this study by comparing the proportion of TNP-specific CTL clones and alloreactive CTL clones generated in short-term (7-day) (data not shown) and restimulated (1 l-day) cultures, but there was no significant difference between the two patterns. The instability of T-cell specificity is under further investigation. Second, the existence of a significant number of TNP-specific, allo H-Zrestricted T cells might be discussed from the facts that TNP-specific, self H-2- or allo H-2restricted T cells cross-react with the third type of allo determinant (BlO.D2) (Tables 2 and 3). In addition, various kinds of antigen-specific, self H-2-restricted T cells have been shown to cross-react with certain ahoantigens or foreign antigens in association with different haplotypes of H-2 determinants (29-36). These significant cross-reactivities could lead to the postulate that TNP-specific, allo H-Zrestricted T cells might cross-reactwith some other foreign antigens in association with self H-2. This concept was originally proposed by Doherty and Bennink asthe concept of “aberrant recognition” ( 10). However, the questions are not yet answered whether TNP-specific, alloH-Zrestricted CTLs recognize “TNP-aho” as an neoalloantigen or TNP plus al10 H-2, as well as whether antigen-specific, self H-Zrestricted T cells recognize “antigenself’ as an alloantigen or an antigen plus self H-2. The most notable finding among the present results is that there is no consistent self preference for H-2 restriction specificity of TNP-specific CLPs in normal T lymphocytes among some strain combinations; that is, the self preference for H-2 restriction specificity of normal T cells is not a general rule. This observation is not expected from the data showing that H-2 restriction specificity of TNP-specific CLPs is profoundly influenced by the host environment in bone marrow chimeras and, therefore, should be important for our consideration of the mechanism for the generation of H-2 restriction specificity of T cells.

40

SHIRO AIZAWA

ACKNOWLEDGMENTS I thank Dr. Richard G. Miller (Ontario Cancer Institute) for his active support and encouragement for this study, a great proportion of which was carried out under his tutelage. This study was set up at DC1 and completed at NIRS. I also express my gratitude to Dr. S. Muraoka, Dr. H. Ishikawa, Dr. G. Suzuki, and Dr. T. Sado for their assistanceand critical advice.

REFERENCES 1. Bevan, M. J., Nature (London) 269,4 17, 1977. 2. Zinkemagel, R. M., Gallahan, G. N., Althage, A., Cooper, S., Klein, P. A., and Klein, J., J. Exp. Med. 147,882,1978. 3. Zinkemagel, R. M., and Doherty, P. C., In “Advances in Immunology,” Vol. 27, p. 51, 1979. 4. Matzinger, P., and Mirkwood, G., J. Exp. Med. 14?3,84,1978. 5. Wagner, H., RUinghoff, M., Rodt, H., and Thierfelder, S., Eur. J. Zmmunol. 10,521, 1980. 6. Aixawa, S., Sado, T., Muto, M., and Kubo, E., J. Zmmunol. 127,2426, 1981. 7. Janeway, C. A., Murphy, P. D., Kemp, J., and Wigzell, H., J. Exp. Med. 147,1065,1978. 8. Schmitt-Verhulst, A. M., and Shearer, G. M., J. Supramol. Struct. 6,206,1977. 9. Wilson, D. B., Lindahl, K. F., Wilson, D. H., and Sprent, J., J. Exp. Med. 146,36 1,1977. 10. Doherty, P. C., and Bennink, J. R., J. Exp. Med. 149,150,1979. 11. Stockinger, H., Plirenmair, K., Hardt, C., RUinghoff, M., and Wagner, H., Proc. Natl. Acad. Sci. USA 77,7390,1980. 12. Stockinger, H., Bartlett, R., Phzenmair, K., RUinghoff, M., and Wagner, H., J. Exp. Med. 153,1629, 1981. 13. Teh, H. S., Bennink, J., and vonBoehmer, H., Eur. J. Zmmunol. 12,887, 1982. 14. Good, M. F., andNossal, G. J. V., Proc. Natl. Acad. Sci. USA88,1693,1983. 15. Aizawa, S., and Miller, R. G., J Cell. B&hem. Suppl. SA, 155,1984. 16. Reimann, J., Heeg, K., Miller, R. G., and Wagner, H., Eur. J. Zmmunol. 15,387, 1985. 17. Reimann, J., Kabelitx, D., Heeg, K., and Wagner, H., J. Exp. Med. 162,592, 1985. 18. Kruisbeek, A. M., Hodes, R. J., and Singer, A., J. Exp. Med. 153,13, 1981. 19. Bradley, S. M., Kruisbeek, A. M., and Singer, A., J. Exp. Med. 156,1650,1982. 20. Shearer, G. M., Eur. J. Zmmunol. 4,527, 1974. 2 1. Aizawa, S., Sado, T., and Kubo, E., Cell. Zmmunol. 83,360, 1984. 22. Porter, E. H., and Berry, R. J., Brit. J. Cancer 17,583,1963. 23. Teh, H. S., Phillips, R. A., and Miller, R. G., J. Zmmunol. 121,1711,1978. 24. Christensen, N., Skinner, M., and Marbrook, J., Eur. J. Zmmunol. 13,701,1983. 25. Good, M. F., Boyd, A. W., and Nossal, G. J. V., J. Zmmunol. 130,2046,1983. 26. Teh, H. S., andYu, M., J. Zmmunol. 131,1827, 1983. 27. Reimann, J., and Miller, R. G., J. Zmmunol. 131,2128, 1983. 28. Reimann, J., and Miier, R. G., CelZ40,57 1, 1985. 29. Bevan, M. J., Proc. Natl. Acad. Sci. USA 74,2094, 1977. 30. Hiinig, T. R., and Bevan, M. J., J. Exp, Med. 155,111,1982. 31. Lemonnier, F., Burakoff, S. J., Germain, R. N., and Benacerraf, B., Z’roc. Nat/. Acad. Sci. USA 74, 1229,1977. 32. vonBoehmer, H., Hengarter, H., Nabholtz, M., Lehnhardt, W., .S~hreier,M., and Haas, W., Eur. J. Zmmunol. 9,592,1979. 33. Sredni, B., and Schwartz, R. H., Nature (London) 287,855,1980. 34. Braciale,T. J.,Andrew,M. E.,andBraciale,V.L., J. Exp. Med. 153,1371,1981. 35. Guimexanes, A., Davigon, J. L., and Schmitt-Verhulst, A. M., Immunogenetics 153,1371,1982. 36. Matxinger, P., Nature (London) 292,497, 1981.