CELLIJLARlMMUNOLOGY
141,21-31
(1992)
Expression and Combinatorial Diversity of Germ Line-Encoded Cell Receptor V Genes in Human Peripheral Blood T Cells LUISA IMBERTI, ALESSANDRA
T
SOTTINI, AND DANIELE PRIMI’
Consorzio per le Biotecnologie, Consiglio Nazionale delle Ricerche (CNR), Institute of Chemistry, School of Medicine, University of Brescia, P.le Spedali Civili, I, 25123 Brescia, Italy Received August 8. 1991; accepted November 19, 1991 The potential diversity of the T cell receptor (TcR) is defined by the combinatorial expression of variable segmentsand by mechanismsthat insert or delete nucleotides at the junctional regions. The available repertoire is strongly influenced by negative and positive selection events. To study whether the diversity of the human T cell receptor of peripheral T cells is further restricted by the interaction between the TcR (Yand /3chains, we compared the level of transcription of different Va elements in human T cell blasts expressing either restricted or unrestricted sets of VP genes. Our data establish that in some individuals, but not in others, the transcription of a given Vcu element is independent from the presence of particular VP transcripts. Furthermore, our data also suggestthat, in contrast to mouse, major TcR V gene deletions are absent in humans. Taken collectively, these results indicate that the diversity of the peripheral human TcR repertoire can benefit from the combinatorial expression of all the V elements present in the genome. o 1992 Academic
Press. Inc.
INTRODUCTION The major histocompatibility complex (MHC) recognition component of the T cell receptor (TcR) complex is, in most T cells, formed by an ~$3protein heterodimer in association with the CD3 complex of proteins ( l-3). Over the last 5 years, many genes for the (Yand p chains have been isolated and well characterized [reviewed in Refs. (4,
VI.
The potential diversity of the TcR is defined by the combinatorial expression of the five germ line variable segments (Vo, Ja, VP, Dp, and J/3)and by additional mechanisms that insert or delete nucleotides in the junctional region (4, 6-8). Somatic mutation, which is one of the major mechanisms responsible for the generation of diversity of immunoglobulin genes,does not appear to influence the diversity of TcR V genes (9, 10). In the mouse, however, the mature TcR repertoire is now known to be shaped by both negative and positive selection events occurring within the thymus (11). Thus, TcR with high affinity for constitutively expressedself-antigens are deleted during T cell development, whereasTcR with low affinity for MHC classI and classII molecules are positively selectedto become mature functional T cells. This MHC-restricted repertoire, which is specific for each individual, is also further selected by the presence ’ On leave of absence from the Unite d’Immunochimie Analytique, Institut Pasteur, 28 rue Dr. Roux, 75724 Paris. France. 21 000%8749192 $3.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in my form reserved.
22
IMBERTI, SOTTINI, AND PRIM1
of endogenous superantigens that mediate VP deletions. Another potential source of selection in mice is the polymorphism of TcR V genes(8, 12- 15) that may influence the level to which each Va gene product can productively associate with any V/3 segments. Little is known, however, about the selective mechanisms that influence the diversity of the human TcR repertoire. We have attempted to addressthis problem by using the polymerase chain reaction (PCR) that allows the estimation, in populations of T cell blasts, of TcR mRNA containing any particular Va! or VP elements. By using this strategy we can identify the transcripts of the different VCYelementsin human T cell blast populations expressing either restricted or unrestricted sets of VP genes. Our data suggest that Va or VP deletions do not occur in humans and that in some, but not in all, individuals there is not an apparent restriction in the interaction between TcR Vol and VP segments. These results suggestthat the rules that govern the selection of the mouse peripheral TcR repertoire may not be generalized to all species. MATERIALS AND METHODS Antibodies. Monoclonal antibodies (MoAb) to CD3 (T3 and T3-FITC) were purchased from Coulter Immunology (Hialeah, FL). The anti-TcR-VP8 region MoAb 12G1 (16) was a kind gift of Dr. OresteAcute (Paris, France). The anti-double-stranded DNA 27- 14-D9 hybridoma was obtained as previously described (17). The utilized polyclonal antibody was a goat anti-mouse Ig F(ab’), fragment conjugated with fluorescein (FITC-GaM Ig F(ab’)*; Technogenetics, Milan, Italy). E+-PBMC preparation and cell cultures. Heparinized blood samples from healthy volunteers (20-30 years of age) were separatedby Ficoll-Hypaque gradient centrifugation and E+-PBMC were prepared by the rosetting technique ( 18). Cells were plated at the concentration of 2 X lo6 cells/ml with 100 rig/ml of T3 MoAb or with 1 pg/ ml of 12Gl MoAb. The purified anti-CD3 and anti-VP8 antibodies were crosslinked by immobilization to plastic flask for 2 hr at room temperature. Cells were cultured for 3 days in the antibody-coated plates in RPM1 1640 (GIBCO Laboratories, Grand Island, NY) containing 10% FCS. The lymphocytes were thereafter washed and recultured for 5 days in medium supplemented with 25 U/ml of human recombinant IL-2 (hrIL-2; Hoffman-La Roche, Basel, Switzerland). Forty-eight hours before the cytofluorimetric analysis, part of the cells were repeatedly washed in order to remove the remaining antibody and were recultured in medium containing hrIL-2. Zmmunofluorescenceanalysis. Viable cells (5 X 105)were resuspendedin 100 ~1 of PBS, 2% FCS, and 0.01% sodium azide and incubated for 30 min at 4°C with FITC anti-CD3 MoAb. For the determination of VPS’ lymphocytes, the cells were sequentially incubated with the 12Gl MoAb and FITC-GaM Ig F(ab’), . All samples were analyzed on an EPICS C (Coulter) flow cytometer. Forward angle and 90” light scatter patterns were used to gate the large blastic cells (80-90%) which were easily distinguished from small lymphocytes. Synthesis of specific V~Yand VP oligonucleotides. To analyze human T cell VCYand V/I usage,we synthesized 19 different Va- and 22 VP-specific oligonucleotides for use as 5’ senseprimers for PCR. In addition, 2 &-specific oligonucleotides and 2 Co, one of which was chosen to match a common sequence near the 5’ end of the Cpl and CD2genes,were also prepared and usedasantisenseprimers. All of the oligonucleotides, obtained from previously published reports (19, 20) were prepared using a DNA synthesizer (39 1 DNA Synthesizer; Applied Biosystem, Santa Clara, CA).
HUMAN
TcR REPERTOIRE
23
Preparation of RNA, cDNA synthesis, and ampliJcation by PCR. Total RNA was prepared from stimulated Ef-PBMC by the guanidinium thiocyanate-phenol-chloroform method as described (21). Two micrograms of the total RNA preparation was used to synthesize the first strand of the complementary DNA (cDNA) using the Riboclone cDNA Synthesis System (Promega Corp., Madison, WI). The amplification was performed with 2.5 U of DNA polymerase (Tth DNA Polymerase from thermus thermophilus HB8; Toyobo, Osaka, Japan) using a Perkin-Elmer thermal cycler (Norwalk, CT) under the following conditions: melting, 93°C for 1 min; annealing, 50°C for 1 min; and extension, 68°C for 2 min. The amplified products were separatedon a 1%agarosegel and the bands, visualized by ethidium bromide staining, were photographed at 302 nm. DNA enzyme immunoassay (DEZA). The analysis of each PCR product was carried out using the DEIA assay,as previously described ( 17). Briefly, amplified Va-Ca and VP-C/3products, denaturated by heating, were added to streptavidin-coated microtiter plates and were preincubated overnight with 5 @well of biotinylated Ca or Cp probes. The wells were washed, and hybridized DNA was revealed by a monoclonal antibody specific for double, but not single-stranded DNA. A sheepanti-mouse IgG conjugated to horseradish peroxidase (ICN Biochemical, High Wycombe, Bucks, England) was used as a second antibody. After the addition of 100 ~1 of the chromogen/substrate solution the reaction was stopped with 1 N H2S04. The absorbance of each sample was read at 450 nm.
RESULTS Selection of T cell populations expressing a specijic VP gene product. The purpose of this work was to determine the levels of genetically or somatically imposed human TcR V gene deletions and to determine if the association between Va and VP segments is a biased phenomenon. Our general strategy was to compare RNA levels, corresponding to different Va genes, in two T cell populations characterized either by the expression of all the available VP segments or by the preferential utilization of one VP gene product only. This strategy was based on the assumption that the expression of the different Va segments should be affected by the availability of different VP segmentsonly if Va-VP pairing is a selective and not a random process. In order to obtain T cell populations characterized by the expression of different sets of VP genes,we isolated Ef-PBMC from the peripheral blood of a single healthy donor. These cells were cultured either in the presence of the T3 MoAb, which drives the polyclonal expansion of all T lymphocytes, or with the 12G 1 MoAb, which selectively interacts with VPS’ lymphocytes and allows the selective expansion of these cells (22). Both blast populations were analyzed by cytofluorimetry for the expression of CD3 and VPS. Figure 1 showsthat lymphocytes expanded with the anti-V@8 MoAb and hrIL-2 are highly enriched in cells expressing the relevant target molecule, while only a small percentageof the blasts activated with anti-CD3 MoAb and hrIG2 reacted with the anti-V/38 MoAb. As expected, virtually all cells, in both blast populations, were CD3+. VP usage among anti- VjS and anti-CD3 selected populations. The complete analysis of the expression of each human TcR V gene product is limited by the lack of a complete set of antibodies specific for each Vcuand VP segment.In addition, in humans, there exists at least 18 Vcu (19) and 20 VP (20) families and therefore the complete
24
IMBERTI, SOTTINI, AND PRIM1 Anti-CD3
mAb selection
Anti-VI38
mAb selectiol
FIG. I. Cytofluorimetric analysis of the expression of V,98 and CD3 molecules on E+-PBMC stimulated with anti-CD3 and anti-V/38 MoAbs. The cut-off value was determined using an irrelevant MoAb.
analysis of the TcR repertoire is often hampered by the available quantity of biological samples. Both of these problems can be overcome by using the PCR, which allows the estimation of the proportion of T cell blasts expressing any particular V gene, starting from a relatively limited number of T cells. This approach assumesthat differencesin the cells expressinga given V/3 gene are reflected on differences in the level of the corresponding transcript in the population. To verify this hypothesis, total cellular RNA from both anti-CD3- and anti-VP8induced blasts were reverSetranscribed into cDNA, which were used as templates for enzymatic amplification by PCR using 22 VP-specific primers that have been reported to match most of the entire human TcR VP repertoire (20). Figure 2 shows the results of this analysis in which the presenceof TcR VP transcripts was operationally defined by the evidence of rearranged VP-Cp bands, visualized by ethidium bromide. The specificity of each amplified product was analyzed by hybridization with a probe internal to the amplified product and the results are expressed as an optical density value obtained by DEIA. The DEIA assay consists of a calorimetric immunoassay recently developed in our laboratory. The test is basedon a monoclonal antibody that recognizes double-stranded, but not single-stranded DNA. The molecule does not react with a specific probe immobilized on microwells through an avidin-biotin bridge nor with nonspecific amplified sequencessince they are removed by washes. The results completely confirm the respective heterogeneity and oligoclonality of the antiCD3 and anti-VP8 blast populations, with respectto the VP expression. Cells selected with anti-V@ MoAb are highly enriched in the VP!3transcript, while the blast population selected by anti-CD3 MoAb expresseda heterogeneous set of VP genes. Interestingly, all of the VP transcripts tested were present in this last T cell population. Since anti-CD3 does not select any particular specificity (23), these results suggest that, in this subject, all genetically available V/3 elements can be transcribed in the periphery.
HUMAN Anti-VI38
mAb
TcR REPERTOIRE
25
selection
vn6+csll*: 69.5% CoO+cella: 95.2% 1.oJ-
Cb Anti-CD3 vnnt 1.5
cells:
mAb 2.8%
selection CDB+cells:
99.2%
r
CD
VP
FIG. 2. Analysis of V@ family expression among anti-V08 and anti-CD3 selected blasts. The amplified
products were revealed by DEIA (top) and ethidium bromide staining (bottom).
Expression of VCYgenes. We next analyzed the presence of Va transcripts in the two selected T cell populations. Reverse-transcribed RNA were amplified with 19 5’ senseVa-specific primers and with a common 3’ antisense Car-specificprimer. The integrity of the cDNA and the efficiency of the amplifications were determined by amplifying the Ccusequence with two &-specific primers. Surprisingly, no clear differencesin the pattern of Va family-specific transcripts could be detected between the two cell populations (Fig. 3). The only major difference in the pattern of Va! transcription between the two samples was the complete absence of the Va7 segment in anti-CD3-enriched blasts, suggesting that the negative selection for this Var gene occurred during differentiation. However, becausethe same Va7 transcript was present in anti-V/38-activated cells, it is likely that, in this donor, Va7 is preferentially transcribed in the VPS’ cells.
26
IMBERTI, SOTTINI, AND PRIM1
FIG. 3. Expressionof TcR Vol family-specific genesin anti-V08 and anti-CD3 selectedblasts.The amplified products were revealed by the DEIA test (top) and ethidium bromide gels (bottom).
Expression of germ line-encoded TcR V genes in peripheral T cells. To determine whether the expression pattern observed with the first donor was reflected in other individuals, we carried out a complete analysis of TcR V gene expression on antiVP& and anti-CD3-selectedblast populations obtained from two more different donors. Figure 4 showsthe results of this analysis,performed with cells from sample 2. Although VP8 selection was not absolute (48.2%of VPS’ cells, and seeFig. 4A) in this experiment, it is evident that such a selection resulted in a profound skew of the Va repertoire originally available in the T cell population (Figs. 4C and 4D). The difference in Va9 in the two populations is particularly marked and suggeststhat even the 50%enrichment in V/38’ cells is sufficient to point out a strong bias in the association between some Vcuand VP segments.This estimate, however, should be considered as tentative since the assaydoes not take into account variations in the efficiency of amplifications. In donor 3 we found no major differences in the Va repertoire expressedby antiCD3- and anti-VP8-triggered T cells (Figs. 5C and 5D). Thus, the most likely explanation for this pattern of family-specific Va transcripts is that the VP8 segment of these cells can indifferently be transcribed with all the Va-encoded polypeptides tested. The only exception was Va 18, which was not transcribed in anti-CD3 blasts, a finding probably due to the Va-specific negative selection. The asymmetry in the expression of Va-Vp transcripts observed in different individuals suggeststhat genetic or environmental factors may play an important role in determining the size of the available repertoire. The most surprising result of this analysis was the pattern of expression of the Va and VP transcripts analyzed among polyclonally activated peripheral T cells of all donors. Donor 1 was the only one that failed to express a Va gene (Va 18) both in the anti-CD3- and anti-V@3-triggered T cell populations. All other Va and VP gene transcripts were found to be expressed in the T cells studied. Since these cells had already undergone all the steps of negative selections, this finding implies that, in contrast to mouse, genetically or somatically TcR V gene deletions are absent in humans.
HUMAN Anti-VOB
mAb selection
TcR REPERTOIRE Anti-CDS
27
mAb selection
"DB+UIIe: 3.9%cm+cnL: 790%
FIG. 4. Expression of VP (A and B) and VLX(C and D) family-specific genes among E+-PBMC isolated from donor 2 and stimulated with anti-V@8 (A and C) and anti-CD3 (B and D) MoAbs. The amplified products were revealed by the DEIA test (top) and ethidium bromide staining (bottom).
DISCUSSION Analysis of the TcR repertoire. In this study we analyzed the extent to which the human TcR V gene repertoire is affected by genetic and somatic selective processes. We used the PCR technique that allowed us to analyze the expression of Va! and VP genesin two T cell populations, selected on the basis of their differential expression of VP genes.This method is very powerful but, nevertheless,it was necessaryto consider several potential biases inherent to the strategy used. In vitro manipulation of primary cells may pose problems of selective amplifications of particular specificities. This, however, is apparently not the case with selection mediated by anti-CD3 antibody, which results in a random activation of the primary cells (23). One of the major concerns about the experimental model used in our study was that the 12Gl MoAb could recognize VP8 segments only when expressed in a particular configuration, such as when are associatedwith some particular Va chains. Two lines of evidence exclude this possibility. First the antibody reacts with the in vitro-translated product of the relevant gene (0. Acute, personal communication) and second, the results of our experiments demonstrate that VPS’ cells selected by the 12G 1 MoAb can express virtually all Va genestested. It was also important to consider two other sources of errors. The first one was the contribution to the PCR products of genesexpressedby unstimulated cells. This problem, however, was only apparent since the vast majority of the cells obtained by our
B
28
IMBERTI, SOTTINI, AND PRIM1
FIG. 5. Expression of VP (A and B) and Va (C and D) family-specific genes among E+-PBMC isolated from donor 3 and stimulated with anti-V@8 (A and C) and anti-CD3 (B and D) MoAbs. The amplified products were revealed by the DEIA test (top) and ethidium bromide staining (bottom).
culture conditions were blasts. Furthermore, the putative contribution of resting cells to our results would not necessarily interfere with the interpretation of the data. The secondproblem was more relevant for our investigation. Since T cells have the potential to rearrange TcR V gene loci on both chromosomes, transcription from a nonproductively rearranged chromosome might confuse the analysis. However, nonfunctional RNA did not apparently affect our data since we found a strict correlation between the number of cells expressing a particular VP gene and the amount of the corresponding specific RNA (seeFigs. 1 and 2). Our study was greatly facilitated by employing the DEIA test that allows the rapid and specific detection of amplified products. We have previously shown that this assay is highly specific, that its sensitivity is identical to conventional Southern hybridizations with radioactive probes (17), and that there is a perfect correspondence between the optical absorbance values and the intensity of the Southern blot bands (24). Thus, the discrepancies observed in this study between the intensity of the ethidium bromide bands and the optical density values of the DEIA assay stress the requirement of hybridizing PCR products with specific probes in order to avoid false negative or positive results. The only limitation of the test, as it was used in the present study, is that it provides only qualitative and not quantitative results becauseit does not take into account possible variations in the efficiency of amplifications. The internal control used in our experiments consisted of T cells polyclonally activated by an anti-CD3 MoAb. We found that these polyclonally activated cells express
HUMAN
TcR REPERTOIRE
29
all the germ line-encoded VP and Va family-specific transcripts. There were apparent variations in the level of each transcript that probably directly correlates with the number of cells expressing each of the V segments. It cannot be excluded, however, that these differences may result from messagestability or promoter strength that can influence the amount of transcript in a given cell. In the mouse, for instance, VpS.2 is an example of a single VP8 segment with two promoters that could lead to a higher level of transcript per cell (25). Furthermore, assaybias connected with the efficiency of PCR may also influence transcript measurements. All of our primers were specific for different V regions and thus they differed in sequenceand this may have influenced the efficiency of annealing. These considerations warn against measuring absolute numbers of cells by extrapolation from the amounts of specific RNA. For these reasons our estimates of V gene expression were only qualitative and not quantitative. Our results suggestthat the productive pairing of VU and VP segments may be a selective processthat probably depends on the nucleotidic sequenceof the translated genes. We also found that the probabilities of specific (Yand p interactions vary in different individuals, suggesting that, besides the somatic processesoccurring in the thymus, gene polymorphism also plays an important role in this mechanism of selection. It is also important to emphasize that the genes amplified by our primers with cDNA isolated from different individuals may not be true alleles, but may be pseudoalleles that do not necessarilyoccupy homologous positions on paired chromosomes. These genes,therefore, could only be members of closely related structures. It should also be emphasized that we do not have any conclusive genetic evidence for ascribing to gene polymorphisms the asymmetrical expression of Vol-VP combinations observed in different individuals. Our results were obtained with peripheral T cells that had already undergone both the positive and negative selective processesthat control the expression of the various TcR V segments. We cannot therefore exclude that the selective expression of some VCYsegments in association with the V/38 gene product observed in our study reflects the selection of particular clones in the periphery. Selection of germ line-encoded human TcR genes. There are several stagesduring T cell maturation where the antigen repertoire of T cells is influenced. The initial V/3 and Va gene segment recombination events in immature T cells may be biased for particular V segments.TcR with high affinity for constitutive self-antigensin association with appropriate MHC gene products are deleted during T cell development, whereas TcR with low affinity for self-MHC are positively selected (26). Recently a category of antigen has been described for which T cell recognition is essentially based on the presence of the V/3 gene segment, independently from the other variable components of the TcR (20, 23, 27). These antigens, collectively called superantigens, include not only exogenous substancesbut also endogenous molecules. Interestingly, it has been demonstrated that, in mice, a neonatal exposure to such VP selecting elements results in the deletion of nearly all T cells expressingthe specific VP genes(27). The molecular nature of endogenous superantigens has remained elusive for a long time, but recently several groups have demonstrated that, in mice, these ligands are encoded by a retrovirus (28-31). Our analysis of TcR V genes in human peripheral T cells provides new information concerning the shaping of the human TcR repertoire. On the basis of the presence of the corresponding mRNA, tested in this work as well as in several other experiments, no genomic or somatic deletions were detected in our samples. These data also confirm and extend those of Concannon et al. (32) and of Baccala et al. (33) who reported the rarity of genomic TcR V gene deletions in humans. These
30
IMBERTI, SOTTINI, AND PRIM1
findings are strongly contrasted with data obtained in mice where extensive genomic deletions have been identified encompassing a large proportion of the number of total VP genes (34-36). The data presented here suggest that genomic or superantigenimposed deletions are absent in humans and, therefore, that the mechanism of TcR selection may be very different in the two species.Important evolutionary pressures possibly related to susceptibility to retrovirus infections may be responsible for these differences. In conclusion our data suggestthat, in humans, TcR V gene polymorphism or peripheral selection may play an important role in determining the possibility of expression of particular Va-V/3 pairs. This potential mechanism of negative selection, however, is not a general phenomenon and is largely compensated by the absenceof genomic or somatic mechanisms of V gene deletion. The general picture that emerges from this study is that evolution in humans has favored the possibility to exploit the whole potential of the germ line-encoded combinatorial diversity. ACKNOWLEDGMENTS This work was supported by Sorin Biomedica (Saluggia, Italy) and by Consiglio Nazionale delle Ricerche (Target Project Biotechnology and Bioinstrumentation; Grant 89.00250.70).
REFERENCES 1. Yague, J., White, J., Coleclough, C., Kappler, J., Palmer, E., and Marrack, P., Ce//42, 81, 1985. 2. Ohashi, P. S., Mak, T. W., Van den Elsen, P., Yanagi, Y., Yoshikai, Y., Calman, A. F., Terhorst, C., Stobo, J. D., and Weiss,A., Nature 316, 606, 1985. 3. Dembic, Z., Haas, W., Weiss,S., McCubrey, J., Kiefer, H., von Boehmer, H., and Steinmetz, M., Nature 320,232, 1986. 4. Kronenberg, M., Siu, G., Hood, L., and Shastri, N., Annu. Rev. Immunol. 4, 529, 1986. 5. Davis, M. M., and Bjorkman, P. J., Nature 334, 395, 1988. 6. Davis, M. M., Annu. Rev. Immunol. 3, 537, 1985. 7. Toyonaga, B., and Mak, T. M., Annu. Rev. Immunol. 5, 585, 1987. 8. Wilson, R. K., Lai, E., Concannon, P., Barth, R. K., and Hood, L. E., Immunol. Rev. 101, 149, 1988. 9. Patten, P., Yokota, T., Rothbard, J., Chien, Y. H., Arai, K. I., and Davis, M. M., Nature 312,40, 1984. 10. Behlke, M. A., Spinella, D. G., Chou, H. S., Sha, W., Hartl, D. L., and Loh, D. Y., Science 229, 566, 1985. 11. Berg, L. J., Curr. Opin. Immunol. 2, 87, 1989. 12. Yoshikai, Y., Clark, S. P., Taylor, S., Sohn, V., Wilson, B., Minden, M., and Mak, T. W., Nature 316, 837, 1985. 13. Klein, M. H., Concannon, P., Everett, M., Kim, L. D. H., Hunkapiller, T., and Hood, L., Proc. Natl. Acad. Sci. USA 84,6884, 1987. 14. Kimura, N., Toyonaga, B., Yoshikai, Y., Du, R. P., and Mak, T. W., Eur. J. Immunol. 17, 375, 1987. 15. Primi, D., Drapier, A. M., and Cazenave, P. A., J. Immunol. 138, 1607, 1987. 16. Acute, O., Meuer, S. C., Hodgdon, J., Schlossman,S. F., and Reinherz, E. L., J. Exp. Med. 158, 1368, 1983. 17. Mantero, G., Zonaro, A., Bertolo, P., Albertini, A., and Primi, D., Clin. Chem. 37, 422, 1991. 18. Mayer, L., Fu, S. M., and Kunkel, G. H., J. Exp. Med. 156, 1860, 1982. 19. Oksenberg, J. R., Stuart, S., Begovich, A. B., Bell, R. B., Erlich, H. A., Steinman, L., and Bernard, C. C. A., Nature 345, 344, 1990. 20. Choi, Y., Kotzin, B., Herron, L., Callahan, J., Marrack, P., and Kappler, J., Proc. Natl. Acad. Sci. USA 86,8941, 1989. 21. Chomczynski, P., and Sacchi, N., Anal. Biochem. 162, 156, 1987. 22. Imbcrti, L., Sottini, A., Spagnoli, G., and Primi, D., Eur. J. Immunol. 20, 2817, 1990. 23. Kappler, J., Kotzin, B., Herron, L., Gelfand, E. W., Bigler, R. B., Boylston, A., Carrel, S., Posnett, D. N., Choi, Y., and Marrack, P., Science 244, 811, 1989.
HUMAN
TcR REPERTOIRE
31
24. Bettinardi, A., Imberti, L., Sottini, A., and Primi, D., J. Zmmunol. Methods, 1992, in press. 25. Chou, H., Anderson, S. J., Louie, M. C., Godambe, S. A., Pozzi, M. R., Behlke, M. A., Huppi, K., and Loh, D. Y., Proc. Natl. Acad. Sci. USA 84, 1992, 1987. 26. Davis, M. M., In “Molecular Immunology” (B. D. Hames and D. M. Glover, Eds.), pp. 61-79. IRL Press,Oxford/Washington, 1988. 27. Marrack, P., and Kappler, J., Science 248, 705, 1990. 28. Dyason, P. J., Knight, A. M., Fairchild, S., Simpson, E., and Tomonari, K., Nature 349, 531, 199I. 29. Frankel, W. N., Rudy, C., Coffin, J. M., and Huber, B. T., Nature 349, 526, 1991. 30. Marrack, P., Kushnir, E., and Kappler, J., Nature 349, 524, 1991. 31. Woodland, D. L., Happ, M. P., Gollob, K. J., and Palmer, E., Nature 349, 529, 1991. 32. Concanon, P., Gatti, R. A., and Hood, L. E., J. Exp. Med. 165, 1130, 1987. 33. Baccala, R., Kono, D. H., Balderas, R., and Theofilopoulos, A. N., Proc. Natl. Acad. Sci. USA 88,2908, 1991. 34. Behlke, M. A., Chou, H. S., Huppi, K., and Loh, D., Proc. Natl. Acad. Sci. USA 83, 767, 1986. 35. Haqqi, T. M., Banejee, S., Anderson, G. D., and David, C. S., .Z.Exp. Med. 169, 1903, 1989. 36. Haqqi, T. M., Banejee, S., Jones, W. L., Anderson, G. D., Behlke, M. A., Loh, D. Y., Luthra, H. S., and David, C. S., Immunogenetics 29, 180, 1989.
141,21-31
(1992)
Expression and Combinatorial Diversity of Germ Line-Encoded Cell Receptor V Genes in Human Peripheral Blood T Cells LUISA IMBERTI, ALESSANDRA
T
SOTTINI, AND DANIELE PRIMI’
Consorzio per le Biotecnologie, Consiglio Nazionale delle Ricerche (CNR), Institute of Chemistry, School of Medicine, University of Brescia, P.le Spedali Civili, I, 25123 Brescia, Italy Received August 8. 1991; accepted November 19, 1991 The potential diversity of the T cell receptor (TcR) is defined by the combinatorial expression of variable segmentsand by mechanismsthat insert or delete nucleotides at the junctional regions. The available repertoire is strongly influenced by negative and positive selection events. To study whether the diversity of the human T cell receptor of peripheral T cells is further restricted by the interaction between the TcR (Yand /3chains, we compared the level of transcription of different Va elements in human T cell blasts expressing either restricted or unrestricted sets of VP genes. Our data establish that in some individuals, but not in others, the transcription of a given Vcu element is independent from the presence of particular VP transcripts. Furthermore, our data also suggestthat, in contrast to mouse, major TcR V gene deletions are absent in humans. Taken collectively, these results indicate that the diversity of the peripheral human TcR repertoire can benefit from the combinatorial expression of all the V elements present in the genome. o 1992 Academic
Press. Inc.
INTRODUCTION The major histocompatibility complex (MHC) recognition component of the T cell receptor (TcR) complex is, in most T cells, formed by an ~$3protein heterodimer in association with the CD3 complex of proteins ( l-3). Over the last 5 years, many genes for the (Yand p chains have been isolated and well characterized [reviewed in Refs. (4,
VI.
The potential diversity of the TcR is defined by the combinatorial expression of the five germ line variable segments (Vo, Ja, VP, Dp, and J/3)and by additional mechanisms that insert or delete nucleotides in the junctional region (4, 6-8). Somatic mutation, which is one of the major mechanisms responsible for the generation of diversity of immunoglobulin genes,does not appear to influence the diversity of TcR V genes (9, 10). In the mouse, however, the mature TcR repertoire is now known to be shaped by both negative and positive selection events occurring within the thymus (11). Thus, TcR with high affinity for constitutively expressedself-antigens are deleted during T cell development, whereasTcR with low affinity for MHC classI and classII molecules are positively selectedto become mature functional T cells. This MHC-restricted repertoire, which is specific for each individual, is also further selected by the presence ’ On leave of absence from the Unite d’Immunochimie Analytique, Institut Pasteur, 28 rue Dr. Roux, 75724 Paris. France. 21 000%8749192 $3.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in my form reserved.
22
IMBERTI, SOTTINI, AND PRIM1
of endogenous superantigens that mediate VP deletions. Another potential source of selection in mice is the polymorphism of TcR V genes(8, 12- 15) that may influence the level to which each Va gene product can productively associate with any V/3 segments. Little is known, however, about the selective mechanisms that influence the diversity of the human TcR repertoire. We have attempted to addressthis problem by using the polymerase chain reaction (PCR) that allows the estimation, in populations of T cell blasts, of TcR mRNA containing any particular Va! or VP elements. By using this strategy we can identify the transcripts of the different VCYelementsin human T cell blast populations expressing either restricted or unrestricted sets of VP genes. Our data suggest that Va or VP deletions do not occur in humans and that in some, but not in all, individuals there is not an apparent restriction in the interaction between TcR Vol and VP segments. These results suggestthat the rules that govern the selection of the mouse peripheral TcR repertoire may not be generalized to all species. MATERIALS AND METHODS Antibodies. Monoclonal antibodies (MoAb) to CD3 (T3 and T3-FITC) were purchased from Coulter Immunology (Hialeah, FL). The anti-TcR-VP8 region MoAb 12G1 (16) was a kind gift of Dr. OresteAcute (Paris, France). The anti-double-stranded DNA 27- 14-D9 hybridoma was obtained as previously described (17). The utilized polyclonal antibody was a goat anti-mouse Ig F(ab’), fragment conjugated with fluorescein (FITC-GaM Ig F(ab’)*; Technogenetics, Milan, Italy). E+-PBMC preparation and cell cultures. Heparinized blood samples from healthy volunteers (20-30 years of age) were separatedby Ficoll-Hypaque gradient centrifugation and E+-PBMC were prepared by the rosetting technique ( 18). Cells were plated at the concentration of 2 X lo6 cells/ml with 100 rig/ml of T3 MoAb or with 1 pg/ ml of 12Gl MoAb. The purified anti-CD3 and anti-VP8 antibodies were crosslinked by immobilization to plastic flask for 2 hr at room temperature. Cells were cultured for 3 days in the antibody-coated plates in RPM1 1640 (GIBCO Laboratories, Grand Island, NY) containing 10% FCS. The lymphocytes were thereafter washed and recultured for 5 days in medium supplemented with 25 U/ml of human recombinant IL-2 (hrIL-2; Hoffman-La Roche, Basel, Switzerland). Forty-eight hours before the cytofluorimetric analysis, part of the cells were repeatedly washed in order to remove the remaining antibody and were recultured in medium containing hrIL-2. Zmmunofluorescenceanalysis. Viable cells (5 X 105)were resuspendedin 100 ~1 of PBS, 2% FCS, and 0.01% sodium azide and incubated for 30 min at 4°C with FITC anti-CD3 MoAb. For the determination of VPS’ lymphocytes, the cells were sequentially incubated with the 12Gl MoAb and FITC-GaM Ig F(ab’), . All samples were analyzed on an EPICS C (Coulter) flow cytometer. Forward angle and 90” light scatter patterns were used to gate the large blastic cells (80-90%) which were easily distinguished from small lymphocytes. Synthesis of specific V~Yand VP oligonucleotides. To analyze human T cell VCYand V/I usage,we synthesized 19 different Va- and 22 VP-specific oligonucleotides for use as 5’ senseprimers for PCR. In addition, 2 &-specific oligonucleotides and 2 Co, one of which was chosen to match a common sequence near the 5’ end of the Cpl and CD2genes,were also prepared and usedasantisenseprimers. All of the oligonucleotides, obtained from previously published reports (19, 20) were prepared using a DNA synthesizer (39 1 DNA Synthesizer; Applied Biosystem, Santa Clara, CA).
HUMAN
TcR REPERTOIRE
23
Preparation of RNA, cDNA synthesis, and ampliJcation by PCR. Total RNA was prepared from stimulated Ef-PBMC by the guanidinium thiocyanate-phenol-chloroform method as described (21). Two micrograms of the total RNA preparation was used to synthesize the first strand of the complementary DNA (cDNA) using the Riboclone cDNA Synthesis System (Promega Corp., Madison, WI). The amplification was performed with 2.5 U of DNA polymerase (Tth DNA Polymerase from thermus thermophilus HB8; Toyobo, Osaka, Japan) using a Perkin-Elmer thermal cycler (Norwalk, CT) under the following conditions: melting, 93°C for 1 min; annealing, 50°C for 1 min; and extension, 68°C for 2 min. The amplified products were separatedon a 1%agarosegel and the bands, visualized by ethidium bromide staining, were photographed at 302 nm. DNA enzyme immunoassay (DEZA). The analysis of each PCR product was carried out using the DEIA assay,as previously described ( 17). Briefly, amplified Va-Ca and VP-C/3products, denaturated by heating, were added to streptavidin-coated microtiter plates and were preincubated overnight with 5 @well of biotinylated Ca or Cp probes. The wells were washed, and hybridized DNA was revealed by a monoclonal antibody specific for double, but not single-stranded DNA. A sheepanti-mouse IgG conjugated to horseradish peroxidase (ICN Biochemical, High Wycombe, Bucks, England) was used as a second antibody. After the addition of 100 ~1 of the chromogen/substrate solution the reaction was stopped with 1 N H2S04. The absorbance of each sample was read at 450 nm.
RESULTS Selection of T cell populations expressing a specijic VP gene product. The purpose of this work was to determine the levels of genetically or somatically imposed human TcR V gene deletions and to determine if the association between Va and VP segments is a biased phenomenon. Our general strategy was to compare RNA levels, corresponding to different Va genes, in two T cell populations characterized either by the expression of all the available VP segments or by the preferential utilization of one VP gene product only. This strategy was based on the assumption that the expression of the different Va segments should be affected by the availability of different VP segmentsonly if Va-VP pairing is a selective and not a random process. In order to obtain T cell populations characterized by the expression of different sets of VP genes,we isolated Ef-PBMC from the peripheral blood of a single healthy donor. These cells were cultured either in the presence of the T3 MoAb, which drives the polyclonal expansion of all T lymphocytes, or with the 12G 1 MoAb, which selectively interacts with VPS’ lymphocytes and allows the selective expansion of these cells (22). Both blast populations were analyzed by cytofluorimetry for the expression of CD3 and VPS. Figure 1 showsthat lymphocytes expanded with the anti-V@8 MoAb and hrIL-2 are highly enriched in cells expressing the relevant target molecule, while only a small percentageof the blasts activated with anti-CD3 MoAb and hrIG2 reacted with the anti-V/38 MoAb. As expected, virtually all cells, in both blast populations, were CD3+. VP usage among anti- VjS and anti-CD3 selected populations. The complete analysis of the expression of each human TcR V gene product is limited by the lack of a complete set of antibodies specific for each Vcuand VP segment.In addition, in humans, there exists at least 18 Vcu (19) and 20 VP (20) families and therefore the complete
24
IMBERTI, SOTTINI, AND PRIM1 Anti-CD3
mAb selection
Anti-VI38
mAb selectiol
FIG. I. Cytofluorimetric analysis of the expression of V,98 and CD3 molecules on E+-PBMC stimulated with anti-CD3 and anti-V/38 MoAbs. The cut-off value was determined using an irrelevant MoAb.
analysis of the TcR repertoire is often hampered by the available quantity of biological samples. Both of these problems can be overcome by using the PCR, which allows the estimation of the proportion of T cell blasts expressing any particular V gene, starting from a relatively limited number of T cells. This approach assumesthat differencesin the cells expressinga given V/3 gene are reflected on differences in the level of the corresponding transcript in the population. To verify this hypothesis, total cellular RNA from both anti-CD3- and anti-VP8induced blasts were reverSetranscribed into cDNA, which were used as templates for enzymatic amplification by PCR using 22 VP-specific primers that have been reported to match most of the entire human TcR VP repertoire (20). Figure 2 shows the results of this analysis in which the presenceof TcR VP transcripts was operationally defined by the evidence of rearranged VP-Cp bands, visualized by ethidium bromide. The specificity of each amplified product was analyzed by hybridization with a probe internal to the amplified product and the results are expressed as an optical density value obtained by DEIA. The DEIA assay consists of a calorimetric immunoassay recently developed in our laboratory. The test is basedon a monoclonal antibody that recognizes double-stranded, but not single-stranded DNA. The molecule does not react with a specific probe immobilized on microwells through an avidin-biotin bridge nor with nonspecific amplified sequencessince they are removed by washes. The results completely confirm the respective heterogeneity and oligoclonality of the antiCD3 and anti-VP8 blast populations, with respectto the VP expression. Cells selected with anti-V@ MoAb are highly enriched in the VP!3transcript, while the blast population selected by anti-CD3 MoAb expresseda heterogeneous set of VP genes. Interestingly, all of the VP transcripts tested were present in this last T cell population. Since anti-CD3 does not select any particular specificity (23), these results suggest that, in this subject, all genetically available V/3 elements can be transcribed in the periphery.
HUMAN Anti-VI38
mAb
TcR REPERTOIRE
25
selection
vn6+csll*: 69.5% CoO+cella: 95.2% 1.oJ-
Cb Anti-CD3 vnnt 1.5
cells:
mAb 2.8%
selection CDB+cells:
99.2%
r
CD
VP
FIG. 2. Analysis of V@ family expression among anti-V08 and anti-CD3 selected blasts. The amplified
products were revealed by DEIA (top) and ethidium bromide staining (bottom).
Expression of VCYgenes. We next analyzed the presence of Va transcripts in the two selected T cell populations. Reverse-transcribed RNA were amplified with 19 5’ senseVa-specific primers and with a common 3’ antisense Car-specificprimer. The integrity of the cDNA and the efficiency of the amplifications were determined by amplifying the Ccusequence with two &-specific primers. Surprisingly, no clear differencesin the pattern of Va family-specific transcripts could be detected between the two cell populations (Fig. 3). The only major difference in the pattern of Va! transcription between the two samples was the complete absence of the Va7 segment in anti-CD3-enriched blasts, suggesting that the negative selection for this Var gene occurred during differentiation. However, becausethe same Va7 transcript was present in anti-V/38-activated cells, it is likely that, in this donor, Va7 is preferentially transcribed in the VPS’ cells.
26
IMBERTI, SOTTINI, AND PRIM1
FIG. 3. Expressionof TcR Vol family-specific genesin anti-V08 and anti-CD3 selectedblasts.The amplified products were revealed by the DEIA test (top) and ethidium bromide gels (bottom).
Expression of germ line-encoded TcR V genes in peripheral T cells. To determine whether the expression pattern observed with the first donor was reflected in other individuals, we carried out a complete analysis of TcR V gene expression on antiVP& and anti-CD3-selectedblast populations obtained from two more different donors. Figure 4 showsthe results of this analysis,performed with cells from sample 2. Although VP8 selection was not absolute (48.2%of VPS’ cells, and seeFig. 4A) in this experiment, it is evident that such a selection resulted in a profound skew of the Va repertoire originally available in the T cell population (Figs. 4C and 4D). The difference in Va9 in the two populations is particularly marked and suggeststhat even the 50%enrichment in V/38’ cells is sufficient to point out a strong bias in the association between some Vcuand VP segments.This estimate, however, should be considered as tentative since the assaydoes not take into account variations in the efficiency of amplifications. In donor 3 we found no major differences in the Va repertoire expressedby antiCD3- and anti-VP8-triggered T cells (Figs. 5C and 5D). Thus, the most likely explanation for this pattern of family-specific Va transcripts is that the VP8 segment of these cells can indifferently be transcribed with all the Va-encoded polypeptides tested. The only exception was Va 18, which was not transcribed in anti-CD3 blasts, a finding probably due to the Va-specific negative selection. The asymmetry in the expression of Va-Vp transcripts observed in different individuals suggeststhat genetic or environmental factors may play an important role in determining the size of the available repertoire. The most surprising result of this analysis was the pattern of expression of the Va and VP transcripts analyzed among polyclonally activated peripheral T cells of all donors. Donor 1 was the only one that failed to express a Va gene (Va 18) both in the anti-CD3- and anti-V@3-triggered T cell populations. All other Va and VP gene transcripts were found to be expressed in the T cells studied. Since these cells had already undergone all the steps of negative selections, this finding implies that, in contrast to mouse, genetically or somatically TcR V gene deletions are absent in humans.
HUMAN Anti-VOB
mAb selection
TcR REPERTOIRE Anti-CDS
27
mAb selection
"DB+UIIe: 3.9%cm+cnL: 790%
FIG. 4. Expression of VP (A and B) and VLX(C and D) family-specific genes among E+-PBMC isolated from donor 2 and stimulated with anti-V@8 (A and C) and anti-CD3 (B and D) MoAbs. The amplified products were revealed by the DEIA test (top) and ethidium bromide staining (bottom).
DISCUSSION Analysis of the TcR repertoire. In this study we analyzed the extent to which the human TcR V gene repertoire is affected by genetic and somatic selective processes. We used the PCR technique that allowed us to analyze the expression of Va! and VP genesin two T cell populations, selected on the basis of their differential expression of VP genes.This method is very powerful but, nevertheless,it was necessaryto consider several potential biases inherent to the strategy used. In vitro manipulation of primary cells may pose problems of selective amplifications of particular specificities. This, however, is apparently not the case with selection mediated by anti-CD3 antibody, which results in a random activation of the primary cells (23). One of the major concerns about the experimental model used in our study was that the 12Gl MoAb could recognize VP8 segments only when expressed in a particular configuration, such as when are associatedwith some particular Va chains. Two lines of evidence exclude this possibility. First the antibody reacts with the in vitro-translated product of the relevant gene (0. Acute, personal communication) and second, the results of our experiments demonstrate that VPS’ cells selected by the 12G 1 MoAb can express virtually all Va genestested. It was also important to consider two other sources of errors. The first one was the contribution to the PCR products of genesexpressedby unstimulated cells. This problem, however, was only apparent since the vast majority of the cells obtained by our
B
28
IMBERTI, SOTTINI, AND PRIM1
FIG. 5. Expression of VP (A and B) and Va (C and D) family-specific genes among E+-PBMC isolated from donor 3 and stimulated with anti-V@8 (A and C) and anti-CD3 (B and D) MoAbs. The amplified products were revealed by the DEIA test (top) and ethidium bromide staining (bottom).
culture conditions were blasts. Furthermore, the putative contribution of resting cells to our results would not necessarily interfere with the interpretation of the data. The secondproblem was more relevant for our investigation. Since T cells have the potential to rearrange TcR V gene loci on both chromosomes, transcription from a nonproductively rearranged chromosome might confuse the analysis. However, nonfunctional RNA did not apparently affect our data since we found a strict correlation between the number of cells expressing a particular VP gene and the amount of the corresponding specific RNA (seeFigs. 1 and 2). Our study was greatly facilitated by employing the DEIA test that allows the rapid and specific detection of amplified products. We have previously shown that this assay is highly specific, that its sensitivity is identical to conventional Southern hybridizations with radioactive probes (17), and that there is a perfect correspondence between the optical absorbance values and the intensity of the Southern blot bands (24). Thus, the discrepancies observed in this study between the intensity of the ethidium bromide bands and the optical density values of the DEIA assay stress the requirement of hybridizing PCR products with specific probes in order to avoid false negative or positive results. The only limitation of the test, as it was used in the present study, is that it provides only qualitative and not quantitative results becauseit does not take into account possible variations in the efficiency of amplifications. The internal control used in our experiments consisted of T cells polyclonally activated by an anti-CD3 MoAb. We found that these polyclonally activated cells express
HUMAN
TcR REPERTOIRE
29
all the germ line-encoded VP and Va family-specific transcripts. There were apparent variations in the level of each transcript that probably directly correlates with the number of cells expressing each of the V segments. It cannot be excluded, however, that these differences may result from messagestability or promoter strength that can influence the amount of transcript in a given cell. In the mouse, for instance, VpS.2 is an example of a single VP8 segment with two promoters that could lead to a higher level of transcript per cell (25). Furthermore, assaybias connected with the efficiency of PCR may also influence transcript measurements. All of our primers were specific for different V regions and thus they differed in sequenceand this may have influenced the efficiency of annealing. These considerations warn against measuring absolute numbers of cells by extrapolation from the amounts of specific RNA. For these reasons our estimates of V gene expression were only qualitative and not quantitative. Our results suggestthat the productive pairing of VU and VP segments may be a selective processthat probably depends on the nucleotidic sequenceof the translated genes. We also found that the probabilities of specific (Yand p interactions vary in different individuals, suggesting that, besides the somatic processesoccurring in the thymus, gene polymorphism also plays an important role in this mechanism of selection. It is also important to emphasize that the genes amplified by our primers with cDNA isolated from different individuals may not be true alleles, but may be pseudoalleles that do not necessarilyoccupy homologous positions on paired chromosomes. These genes,therefore, could only be members of closely related structures. It should also be emphasized that we do not have any conclusive genetic evidence for ascribing to gene polymorphisms the asymmetrical expression of Vol-VP combinations observed in different individuals. Our results were obtained with peripheral T cells that had already undergone both the positive and negative selective processesthat control the expression of the various TcR V segments. We cannot therefore exclude that the selective expression of some VCYsegments in association with the V/38 gene product observed in our study reflects the selection of particular clones in the periphery. Selection of germ line-encoded human TcR genes. There are several stagesduring T cell maturation where the antigen repertoire of T cells is influenced. The initial V/3 and Va gene segment recombination events in immature T cells may be biased for particular V segments.TcR with high affinity for constitutive self-antigensin association with appropriate MHC gene products are deleted during T cell development, whereas TcR with low affinity for self-MHC are positively selected (26). Recently a category of antigen has been described for which T cell recognition is essentially based on the presence of the V/3 gene segment, independently from the other variable components of the TcR (20, 23, 27). These antigens, collectively called superantigens, include not only exogenous substancesbut also endogenous molecules. Interestingly, it has been demonstrated that, in mice, a neonatal exposure to such VP selecting elements results in the deletion of nearly all T cells expressingthe specific VP genes(27). The molecular nature of endogenous superantigens has remained elusive for a long time, but recently several groups have demonstrated that, in mice, these ligands are encoded by a retrovirus (28-31). Our analysis of TcR V genes in human peripheral T cells provides new information concerning the shaping of the human TcR repertoire. On the basis of the presence of the corresponding mRNA, tested in this work as well as in several other experiments, no genomic or somatic deletions were detected in our samples. These data also confirm and extend those of Concannon et al. (32) and of Baccala et al. (33) who reported the rarity of genomic TcR V gene deletions in humans. These
30
IMBERTI, SOTTINI, AND PRIM1
findings are strongly contrasted with data obtained in mice where extensive genomic deletions have been identified encompassing a large proportion of the number of total VP genes (34-36). The data presented here suggest that genomic or superantigenimposed deletions are absent in humans and, therefore, that the mechanism of TcR selection may be very different in the two species.Important evolutionary pressures possibly related to susceptibility to retrovirus infections may be responsible for these differences. In conclusion our data suggestthat, in humans, TcR V gene polymorphism or peripheral selection may play an important role in determining the possibility of expression of particular Va-V/3 pairs. This potential mechanism of negative selection, however, is not a general phenomenon and is largely compensated by the absenceof genomic or somatic mechanisms of V gene deletion. The general picture that emerges from this study is that evolution in humans has favored the possibility to exploit the whole potential of the germ line-encoded combinatorial diversity. ACKNOWLEDGMENTS This work was supported by Sorin Biomedica (Saluggia, Italy) and by Consiglio Nazionale delle Ricerche (Target Project Biotechnology and Bioinstrumentation; Grant 89.00250.70).
REFERENCES 1. Yague, J., White, J., Coleclough, C., Kappler, J., Palmer, E., and Marrack, P., Ce//42, 81, 1985. 2. Ohashi, P. S., Mak, T. W., Van den Elsen, P., Yanagi, Y., Yoshikai, Y., Calman, A. F., Terhorst, C., Stobo, J. D., and Weiss,A., Nature 316, 606, 1985. 3. Dembic, Z., Haas, W., Weiss,S., McCubrey, J., Kiefer, H., von Boehmer, H., and Steinmetz, M., Nature 320,232, 1986. 4. Kronenberg, M., Siu, G., Hood, L., and Shastri, N., Annu. Rev. Immunol. 4, 529, 1986. 5. Davis, M. M., and Bjorkman, P. J., Nature 334, 395, 1988. 6. Davis, M. M., Annu. Rev. Immunol. 3, 537, 1985. 7. Toyonaga, B., and Mak, T. M., Annu. Rev. Immunol. 5, 585, 1987. 8. Wilson, R. K., Lai, E., Concannon, P., Barth, R. K., and Hood, L. E., Immunol. Rev. 101, 149, 1988. 9. Patten, P., Yokota, T., Rothbard, J., Chien, Y. H., Arai, K. I., and Davis, M. M., Nature 312,40, 1984. 10. Behlke, M. A., Spinella, D. G., Chou, H. S., Sha, W., Hartl, D. L., and Loh, D. Y., Science 229, 566, 1985. 11. Berg, L. J., Curr. Opin. Immunol. 2, 87, 1989. 12. Yoshikai, Y., Clark, S. P., Taylor, S., Sohn, V., Wilson, B., Minden, M., and Mak, T. W., Nature 316, 837, 1985. 13. Klein, M. H., Concannon, P., Everett, M., Kim, L. D. H., Hunkapiller, T., and Hood, L., Proc. Natl. Acad. Sci. USA 84,6884, 1987. 14. Kimura, N., Toyonaga, B., Yoshikai, Y., Du, R. P., and Mak, T. W., Eur. J. Immunol. 17, 375, 1987. 15. Primi, D., Drapier, A. M., and Cazenave, P. A., J. Immunol. 138, 1607, 1987. 16. Acute, O., Meuer, S. C., Hodgdon, J., Schlossman,S. F., and Reinherz, E. L., J. Exp. Med. 158, 1368, 1983. 17. Mantero, G., Zonaro, A., Bertolo, P., Albertini, A., and Primi, D., Clin. Chem. 37, 422, 1991. 18. Mayer, L., Fu, S. M., and Kunkel, G. H., J. Exp. Med. 156, 1860, 1982. 19. Oksenberg, J. R., Stuart, S., Begovich, A. B., Bell, R. B., Erlich, H. A., Steinman, L., and Bernard, C. C. A., Nature 345, 344, 1990. 20. Choi, Y., Kotzin, B., Herron, L., Callahan, J., Marrack, P., and Kappler, J., Proc. Natl. Acad. Sci. USA 86,8941, 1989. 21. Chomczynski, P., and Sacchi, N., Anal. Biochem. 162, 156, 1987. 22. Imbcrti, L., Sottini, A., Spagnoli, G., and Primi, D., Eur. J. Immunol. 20, 2817, 1990. 23. Kappler, J., Kotzin, B., Herron, L., Gelfand, E. W., Bigler, R. B., Boylston, A., Carrel, S., Posnett, D. N., Choi, Y., and Marrack, P., Science 244, 811, 1989.
HUMAN
TcR REPERTOIRE
31
24. Bettinardi, A., Imberti, L., Sottini, A., and Primi, D., J. Zmmunol. Methods, 1992, in press. 25. Chou, H., Anderson, S. J., Louie, M. C., Godambe, S. A., Pozzi, M. R., Behlke, M. A., Huppi, K., and Loh, D. Y., Proc. Natl. Acad. Sci. USA 84, 1992, 1987. 26. Davis, M. M., In “Molecular Immunology” (B. D. Hames and D. M. Glover, Eds.), pp. 61-79. IRL Press,Oxford/Washington, 1988. 27. Marrack, P., and Kappler, J., Science 248, 705, 1990. 28. Dyason, P. J., Knight, A. M., Fairchild, S., Simpson, E., and Tomonari, K., Nature 349, 531, 199I. 29. Frankel, W. N., Rudy, C., Coffin, J. M., and Huber, B. T., Nature 349, 526, 1991. 30. Marrack, P., Kushnir, E., and Kappler, J., Nature 349, 524, 1991. 31. Woodland, D. L., Happ, M. P., Gollob, K. J., and Palmer, E., Nature 349, 529, 1991. 32. Concanon, P., Gatti, R. A., and Hood, L. E., J. Exp. Med. 165, 1130, 1987. 33. Baccala, R., Kono, D. H., Balderas, R., and Theofilopoulos, A. N., Proc. Natl. Acad. Sci. USA 88,2908, 1991. 34. Behlke, M. A., Chou, H. S., Huppi, K., and Loh, D., Proc. Natl. Acad. Sci. USA 83, 767, 1986. 35. Haqqi, T. M., Banejee, S., Anderson, G. D., and David, C. S., .Z.Exp. Med. 169, 1903, 1989. 36. Haqqi, T. M., Banejee, S., Jones, W. L., Anderson, G. D., Behlke, M. A., Loh, D. Y., Luthra, H. S., and David, C. S., Immunogenetics 29, 180, 1989.
