CELLULAR

IMMUNOLOGY

141,332-341 (1992)

Phenotypic and Functional Characterization of Cytotoxic Cells Derived from Endomyocardial Biopsies in Human Cardiac Allografts LIVIO TRENTIN,* RENATO ZAMBELLO,* GIUSEPPE FAGGIAN,? UGOLINO LIvI,t GAETANO THIENE,$ GIUSEPPEGASPAROTTO,*AND CARLO AGOSTINI* *Department of Clinical Medicine, Clinical Immunology Section, 7Departments of Cardiovascular Surgery, and $Pathology, Padua University School of Medicine, Via Giustiniani 2, 35128 Padova, Italy Received October 3, 1991; accepted December 4, 1991 This study was undertaken to characterize the phenotype and function of lymphocytes derived from endomyocardial biopsies in heart transplant patients. To this aim, tissue infiltrating lymphocytes were derived from seven heart transplant patients and were analyzed for the expression of a panel of markers, including CD3, CD4, CD8, CD16, CD56, CD45RA, CD45R0, a/p and y/6 T cell receptor, and for their ability to lyse a seriesof targets, including NK-sensitive K-562 targets, NK-resistant Raji targets, donor related, and unrelated normal splenocytes.Our data show that the majority of cultured lymphocytes expressed the CD3+ phenotype and the cu//3T cell receptor. The CD4 and CD8 molecules were heterogeneouslyexpressedamong T cell lines tested. Concerning cytotoxic related markers, a significant percentageof cells were CD56+. The evaluation of CD45 isoforms showed that both “naive” and “memory” cells were present among heart TIL. Cytotoxic in vitro studies demonstrated that all our T cell lines showed an efficient cytotoxic machinery when tested against NK-sensitive targets. A marked lysis of donor-related splenocytes was demonstrated in all patients tested. To investigate the role of CD3 and HLA classI molecules in the cytotoxic mechanisms taking place in human heart allograft rejection mechanisms, TIL were assessedfor their lytic activity against different targets in the presenceof anti-CD3 and antiHLA classI monoclonal antibodies (mAbs). Although donor-specific cytotoxicity was considerably inhibited by the anti-CD3 mAb, no inhibitory effect was displayed by this antibody on TILmediated cytotoxicity against donor-unrelated splenocytes. Anti-HLA class I mAb was able to inhibit both allospecitic and nonallospecific cytotoxicity. These data suggestthat different types of cytotoxic cells may be propagated from biopsy specimens of heart transplant patients. 0 1992 Academic

Press, Inc.

INTRODUCTION The diagnosis of cardiac allografi rejection is currently dependent on the microscopic identification of infiltrating lymphocytes with or without myocyte necrosis, edema, and hemorrhage in specimens obtained by endomyocardial biopsies. Several studies have shown that a large proportion of intragraft cells is represented by T lymphocytes, thus indicating that T cells might play a key role in the mechanisms of allograft rejection (l-3). Although the phenotype of intragraft lymphocytes has been partially characterized (4-S), the functional mechanisms by which these cells contribute to transplant rejection are not clearly understood. Recent evidence from animal models suggeststhat allospecific cytotoxic T lymphocytes (CTL) and MHC-unrestricted cytotoxic cells are involved in the phenomena 332 000%8749/92 $5.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form resaved

GRAFT

INFILTRATING

LYMPHOCYTES

IN HEART

TRANSPLANTS

333

leading to allograft rejection (9, 10). In particular, nonspecific cytotoxic lymphocytes, including natural killer (NK) cells, seem to act as effector cells in the early stage of allograft rejection, whereas allospecific CTL are viewed as having a role in the more advanced phasesof the rejection process( 11, 12). In human beings, little information is available on the cytotoxic capabilities of graft infiltrating lymphocytes. Donor-specific CTL have been isolated from renal allograft biopsies in kidney rejection ( 13) but no data are available on the role of cytotoxic cells in human heart transplants. In the present study we investigated whether allospecific and nonspecific cytotoxic cells take part in human heart allografi rejection mechanisms.To this aim, we evaluated the phenotype and cytotoxic in vitro function of tissue-infiltrating lymphocytes (TIL) propagated from endomyocardial biopsies (EMB) in seven heart transplant patients. In particular, the function of graft-derived TIL was assessedby evaluating cell-mediated cytotoxicity against donor-related and unrelated splenocytes, NK-sensitive and NKresistant targets. MATERIALS AND METHODS Patient population. Thirty-two serial endomyocardial biopsies were taken from 12 heart transplant recipients (9 males and 3 females with an age ranging from 36 to 60 years, mean 51 + 4 years) during the first 90 days following the transplant; biopsy specimenswere obtained from the right ventricle using a transjugular technique under fluoroscopic guidance. The causesof heart failure before transplantation were ischemic cardiomiopathy in 9 out of 12 patients and dilated cardiomiopathy in the remaining 3. Patients were subjected to the same immunosuppressive protocol represented by a combination of cyclosporine A and azathioprine. EMB was performed in these two situations: (a) routinely at l-week intervals during the first month, then biweekly during the second month, and once a month for the following 11 months post-transplant or (b) upon suspicion of rejection on the basis of clinical and ecocardiography findings. Three fragments of each EMB were used for histopathologic assessmentof rejection and one for the generation of tissue-infiltrating lymphocytes. Each biopsy slide was reviewed by two pathologists and graded using standardized grading system proposed by the International Society for Heart Transplantation (14). According to these standard criteria, 20 of the 32 EMB were graded as histologically negative for rejection and 12 EMB showed histological evidence of rejection. Propagation qflymphocytes from biopsy specimens. All 32 serial biopsy fragments were cultured in vitro for the generation of TIL according to the following method. The biopsy was minced into two pieces; each was then rinsed extensively in RPM1 1640 and incubated at 37°C in 5% CO1 for 2 hr in tissue culture dishes with 3 ml of RPM1 1640 supplemented with 20% fetal calf serum (FCS) (Flow Laboratories, UK). This preincubation period was performed in order to remove peripheral blood contamination. Each fragment was minced and placed in a 24-well flat-bottomed tissue culture plate (Falcon 3047, Becton-Dickinson, Sunnyvale, CA) in 2 ml of RPM1 1640 medium supplemented with 10% fetal calf serum (FCS), 2 rnJ4 glutamine, 24 mA4 Hepesbuffer, 100 U/ml penicillin, 100 pg/ml streptomycin, 100 U/ml of recombinant IL-2 (Genzyme Corp., Boston, MA). Two to five days later, EMB-derived TIL were restimulated in the presence of IL-2 (100 U/ml) and feeder cells ( 1 X 106/well). A total of six cell lines was generated from histologically positive EMB obtained from different patients; the remaining six EMB with hystological evidence of rejection did

334

TRENTIN

ET AL.

not reach an adequate number of cells to perform the phenotypic and functional analysis. A lymphocyte cell line was derived from 1 out of the 20 histologically negative biopsies. Interestingly, 2 weeks later this patient showed histological signs of rejection. Feeder cells were represented by irradiated donor specific splenocytes in Patients 1, 2, 3, and 7, and by normal irradiated peripheral blood mononuclear cells in Cases4, 5, and 6 since in theselast patients the spleenwas not available at the time of transplant. Becauseof the limited number of TIL derived from the EMB of patients under study, not all subjects were evaluated for all assays;the identification numbers of individuals studied are indicated in each assaywith the respective data. Flow cytometric analysis. TIL propagated from the EMB were characterized by different groups of monoclonal antibodies (mAbs), most of them belonging to the OK (Ortho Pharmaceuticals, Raritan, NJ) and Leu (Becton-Dickinson, Sunnyvale, CA) series, including those belonging to CD3 (Leu4, OKT3), CD4 (Leu3, OKT4), CD8 (Leu2, OKT8), CD16 (Leu 1la), CD56 (Leu19). The specificity of these reagents has been reported in detail ( 15). The TCR-related TCR 1 (WT3 1, Becton-Dickinson) and TCR-y/G- 1 (Becton-Dickinson) mAbs not classified in CD were also used. WT3 1 recognizes the a//3 TCR while TCR--r/G- 1 mAb reacts with a common epitope of the 6 chain apparently expressedby all y/S cells ( 16). To characterize memory and virgin cells, CD45RA (Leu 17, Becton-Dickinson) and CD45RO (UCHL 1, Dako. Glostrup, Denmark) mAbs were used. The frequency of cells positive for the above reagentswas determined by flow cytometry, as previously reported (17). Cell-mediated cytotoxicity. Cell-mediated cytotoxic assay was assessedby lysis of “Cr-labeled NK-sensitive K-562 targets, NK-resistant Daudi targets, donor-specific and allogenic splenocytes as previously reported ( 18). K-562 and Daudi cell lines were maintained as a continuous culture (mycoplasma-free) in RPM1 1640 containing 10% FCS, 2 mM glutamine, 24 mM Hepes buffer, 100 U/ml penicillin, 100 pg/ml streptomycin in a 95% air and 5% CO2 atmosphere at 37°C. Splenocytes were obtained after mechanical disruption of the spleen and centrifugation on a F/H gradient. The cells obtained were then washed three times with PBS, suspended in RPMI- 1640 containing 60% FCS and 7.5% dimethylsulfoxide, and stored in liquid nitrogen until their use. When used as targets,the cells were thawed and their viability was determined by means of trypan blue exclusion test. Preparations with more than 80% viable cells were used as targets. Cytotoxic assayswere performed using proliferating lymphocytes as effector cells. Briefly, 1 X lo6 targets were labeled for 2 hr at 37°C in 5% CO* with 100 PCi Na(5 lCr)04 (CEA IRE Sorin, Biomedica, Saluggia,Italy) and were extensively washed before use. Target cells (lO’/ml) were suspended in each well of a V-shaped plate (Tirtetek; ICN, Oxnard, CA) and graded concentrations of effector cells were added to wells in triplicate and were incubated at 37°C in 5% CO1 for 4 hr. After this incubation period, supernatants were harvested and counted in a gamma counter. The mean value of triplicate assayswas used to calculate the percentageof cytotoxicity according to the following formula: (cpm effector cells - cpm spontaneous release) X lOO/(cpm maximum release - cpm spontaneous release). Spontaneous release was always less than 8% from the K-562 and Daudi targets, and less than 15% from splenocytes. Blocking experiments with mAbs. To define the role of HLA class I molecules and CD3 antigen in cytotoxicity displayed by TIL, cytotoxic assayswere also performed at a 40: 1 effector:target ratio in the absence and presence of preservative-free anti-

GRAFT

INFILTRATING

LYMPHOCYTES

IN HEART

TRANSPLANTS

335

CD3 (10 Kg/ml) (gently provided by Ortho Pharmaceuticals, Raritan, NJ) and antiHLA class I mAbs (W6/32 clone supernatant, 1:2 supernatant dilution). RESULTS Flow cytometric analysis. Phenotypic analysis of TIL propagated from EMB is reported in Table 1. The large majority of cells were represented by T cells bearing the CD3 antigen and the a/P T cell receptor. The y/S TCR was displayed by a discrete percentage of TIL ( 16%)belonging to one cell line. The distribution of CD4 and CD8 antigens was almost heterogeneous among T cell lines derived from EMB of our patients. The evaluation of cytotoxic-related molecules showed that a discrete number of infiltrating lymphocytes beared the CD56 antigen, but lacked the NK-related CD 16 antigen (data not shown). The study of the distribution of CD25 activation molecule showed that a high percentage of TIL displayed the p55 chain of IL-2 receptor. When T cell lines were evaluated for the expression of memory and naive cell related antigens, CD45ROt cells were particularly represented in four out of seven T cell lines (Nos. 1, 2, 3, and 7) while the other three T cell lines preferentially bear the CD45RA antigen. Cell-mediated cytotoxicity. Cytotoxicity by EMB-derived T cell lines was evaluated against different targets. When TIL were tested for their capability to kill NK-sensitive (K-562) and NK-resistant (Daudi) targets (Fig. 1) high lytic activity was demonstrated against K-562 targets (range from 36 to 64%) and, to a lesser extent, against Daudi targets (range from 18 to 58%). To investigate TIL for the presence of alloreactive-specific cytotoxic T ceils, graftinfiltrating lymphocytes were evaluated for their capability to kill donor specific and nonspecific splenocytes. As shown in Fig. 2, TIL from the three (Nos. 1, 2, and 3) patients tested showed lytic activity against donor-specific splenocytes. When the cytotoxic activity was evaluated against donor nonspecific splenocytes, no high lytic activity was observed, with the exception of two cases(Nos. 1 and 3). In particular TIL from Patients I and 3 showed a cross-reactivity (Fig. 2). The percentage of lysis exhibited from TIL belonging to Patients 2, 5, and 6 was lessthan lo%, although they were able to display high lytic activity against K-562 and Daudi targets. Eflect of anti-CD3 and anti-HLA class I mAbs on cytotoxicity. To define the role of the CD3 and MHC class I molecules in the cytotoxic repertoire displayed by EMBTABLE

I

Phenotypic Analysis of TIL Derived from Heart Transplant Endomyocardial

Biopsies

Patient No.

CD3 (%I

CD4

(%I

CD8 (%)

CD56 (%I

CD25 @)

CD45RO (%)

CD45RA (%)

I 2 3 4 5 6 7

99 100 93 91 91 92 91

14 19 41 56 6 22 71

25 80 52 36 83 72 18

20 15 8 4 14 20 6

63 31 78 59 20 18 60

39 73 31 8 4 6 39

24 II 20 60 43 68 24

336

TRENTIN ET AL.

0 1234567

patients H K-562 targets

q

Daudi targets

FIG. 1. Values of cytotoxicity displayed by seven T cell lines derived from seven heart transplant patients against NK-sensitive K-562 targets and NK-resistant Daudi targets in a 4-hr cytotoxic assay.The values of lysis reported are referred to the effector:target ratio of 40: 1.

derived T cell lines we evaluated the lytic function exhibited by TIL against different targets in the presence of antibodies directed against the CD3 and MHC class I molecules (Table 2, Figs. 3 and 4). When K-562 targets are considered (Table 2), anti-CD3 mAb exhibited a blocking effect on TIL-mediated cytotoxicity in Patients 1 and 3, while a marked increase in the cytotoxic function was demonstrated in the other patients (Nos. 2,5, and 6). When anti-HLA class I mAb was added in the test, an inhibitory effect of variable degree was observed in all patients tested. The role of these antibodies on TIL-mediated cytotoxicity against donor-specific and donor-nonspecific splenocytes is reported in Figs. 3 and 4, respectively. In all three patients tested, both anti-CD3 and anti-HLA class I mAbs were able to inhibit the killing of donor-specific splenocytes by EMB-derived TIL (Fig. 3). When donorunrelated splenocytes were used as targets (Fig. 4) anti-HLA class I mAb was consistently able to reduce the lysis of allogeneic splenocytes in all patients tested, whereas a heterogeneous effect was exhibited by anti-CD3 mAb. In fact, an inhibitory effect was observed in Patients 1 and 3, who showed a cross-reactivity, while an increase in the lytic activity was demonstrated in the other three patients. DISCUSSION This study was undertaken to characterize the phenotype and the cytotoxic function of human allograft derived T cell lines obtained from heart transplant subjects. Our study demonstrates that infiltrating lymphocytes were represented by T cells and that a discrete number of them beared the CD56 cytotoxic-related antigen. The evaluation of TCR showed that the majority of the cells were cu/Pand lacked the y/S TCR thus

GRAFT INFILTRATING

1

331

LYMPHOCYTES IN HEART TRANSPLANTS

2

3

5

6

patients

FIG. 2. Cytotoxicity displayed by T cell lines derived from five heart transplant patients against donorspecific (m) and donor-nonspecific (a) splenocytes in a 4-hr cytotoxic assay.The values are referred to the efIector:target ratio of 40: I.

suggestingthat a//3 cells represent the major population involved in rejection mechanisms. Both naive and memory cells were present among TIL, as demonstrated by the expression of CD45RA and CD45RO isoforms. The evaluation of cytotoxic machinery of EMB-derived T cell lines demonstrated that they were functionally able to kill NK-sensitive, NK-resistant targets, and donor-specific splenocytes. By using flow cytometry analysis we established that both CD4 and CD8 antigens were heterogeneously expressed by graft infiltrating T cell lines. Furthermore, the percentage of CD4 and CD8 cells varied randomly from patient to patient and sometimes among T cell lines obtained from the same patient at different times (data not shown). These findings indicate that cells which take part in allograft rejection did

TABLE 2 Cytotoxicity Displayed by TIL Lines from Five Transplant Heart Byopsies against K-562 Targets in the Absence and in the Presenceof Different Monoclonal Antibodies Patient’s TIL

Alone

I

50.1

2 3 5 6

21.4 55.1 6.0 1.3

+ Anti-CD3 mAb

+ Anti-Class I mAb

+ Control IgG mAb

43.0 49.4 39.8 38.1 36.1

24.8 21.9 21.7 2.1 4.8

48.9 29.8 51.3 6.4 7.0

Note. Results are reported at the effector:target ratio of 40: 1.

338

TRENTIN ET AL. 80 II

n. 1

n. 2 patients

n alone q + anti-Class I mAb

n. 3

q + anti-CD3 mAb q + IgG control

FIG. 3. Effect of antibody blocking on donor-specific splenocyte lysis by graft infiltrating T cell lines in three patients. The lytic values showed are referred to the effector:target ratio of 40: 1 and were obtained in the absenceor in the presencein the cytotoxic tests of antibodies directed against CD3 (OKT3), MHC class I molecules (W6/32), and control mouse IgG.

not belong to a unique subpopulation of T cells. Since in our experiments the time of restimulation of graft infiltrating cells changed from 2 to 5 days following the assessment of the biopsy culture, it could be hypothesized that the differences in the growth kinetics of CD4 and CD8 cell subsetsmight account for the phenotypic heterogeneity of TIL. To explain the presence of CD4+ lymphocytes in our T cell lines, another possibility could be that helper T cells might be indirectly involved in the rejection process by promoting the growth and the activation of CD8+ cytotoxic effectors via the production of cytokines which are required for the regulation of the cytotoxic capabilities of CTL, including IL-2 and IFN-7. An alternative, but not mutually exclusive hypothesis could be that both CD8 and CD4 lymphocytes are directly involved in the cytotoxic mechanisms mediating rejection; in this scenario, it is possible that CD8 and CD4 lymphocytes may participate at different times to the allograft rejection process, as recently proposed by others (11, 19-2 1). Further evaluation of the phenotype and function of purified T cell subsets isolated from serial biopsies obtained from a wide group of heart allografts at regular intervals could verify this last possibility. It is generally accepted that human T cells can be divided into two subpopulations on the basis of the expression of the 205 to 220-kDa isoform of CD45 (CD45RA) or the low M, isoforms CD45RO ( 180 kDa). Cells expressingCD45RA antigen are defined as “naive” or “virgin” T cells. After activation, there is a switch, claimed to be uni-

GRAFT

INFILTRATING

LYMPHOCYTES

IN HEART

TRANSPLANTS

339

60

50

2 g

40

g .t+ $ 30 z E 8k

20

a

10

0

1

3

3

5

6

patients m

medium alone + anti-Class I mAb

m [7

+ anti-CD3 mAb + IgG control

FIG. 4. Effect of antibody blocking on TIL-mediated cytotoxicity from five patients against donor-unrelated splenocytes. The values of lysis are referred to the effector:target ratio of 4O:l and were obtained in the absence or in the presence in cytotoxic tests of antibodies directed against CD3 (OKT3), MHC class I molecules (W6/32), and control mouse IgG.

directional, to the expression of lower molecular weight isoforms, one of which can be detected by CD45RO mAbs (22, 23). The CD45RO T cell subset include lymphocytes that have been defined as memory based on their ability of providing a longterm recall proliferative response to antigens (24). In order to obtain information about the events which are involved in the activation of T cells in heart allografts, we evaluated the expression of CD45RA and CD45RO molecules on graft-infiltrating lymphocytes. Our data showed that both CD45 isoforms were distributed among graft infiltrating lymphocytes (Table 1). Interestingly, we observed that T cell lines preferentially expressingCD45RO+ antigens (Nos. 1,2,3, and 7) were derived in the presence of irradiated donor-specific splenocytes; by contrast, TIL propagated in the presence of irradiated normal allogeneic peripheral blood mononuclear cells showed a predominant CD45RA naive phenotype. These findings substantiate the hypothesis that lymphocytes infiltrating the heart allograft might proliferate in responseto recall antigens (represented by the MHC antigens expressed by donor’s splenocytes). Furthermore, the consideration that a discrete number of CD45RA TIL can be propagated by heart allograft suggeststhat rejection process is likely associated to an indiscriminate recruitment of T cells. The recent demonstration of an interconversion of CD45R subsets in experimental models (25) might suggestan alternative explanation for the expression of CD45RA antigens by some T cell lines. Immunohistologic analysis performed on myocardial specimens are needed to investigate the expression of CD45 isoforms by heart graft-infiltrating lymphocytes. To define the functional role of TIL in rejection mechanisms, we tested the cytotoxic activity of propagated TIL against different targets. The evidence that all cell lines

340

TRENTIN ET AL.

tested were able to kill both NK-sensitive and NK-resistant targets suggeststhat the majority of cells derived from rejecting patients are represented by cytolytic T cells. Again, we tested T cell lines against donor-specific and donor-nonspecific splenocytes to investigate whether the lytic activity was donor specific (Fig. 2). Our findings showed that all T cell lines tested exhibited donor-specific cytotoxicity, indicating that allospecific CTL are the major cytotoxic cell population involved in the rejection process. The observation that mAbs against HLA classI molecules block allospecific cytotoxicity is a further demonstration that these T cell lines kill donor-specific targets with specificity for MHC determinants. In this regard, the evidence of a cross-reactivecytotoxicity by two cell lines (Nos. 1 and 3, Fig. 2) might be interpreted as the consequence of the recognition of an unknown MHC determinant that is coexpressedby the two target splenocytes. Another observation that deservessome comment is represented by the role of the CD3 molecule on the cytotoxic activity of graft-infiltrating T cells. Since anti-CD3 mAb seemsto mimic the effect of antigen on T cells by either determining an inhibition of cytotoxic T lymphocytes (CTL) or by inducing an activation of cytotoxic machinery (26-28), we carried out a seriesof experiments using anti-CD3 mAb. We demonstrated that anti-CD3 mAb was able to reduce the cytotoxicity against donor related splenocytes; in contrast, CD3 also induced an increase in the lysis of donor-nonspecific splenocytes and other target cell lines tested. These findings are in accordance with the results reported by Van Seventer et al. (27) in which anti-CD3 mAb treatment of effector cells resulted in the inhibition of specific recognition of the target cell by cytotoxic T lymphocytes. In addition, we suggestthat the increased killing observed against donor unrelated splenocytes and K-562 targets following incubation with antiCD3 mAb could be related to the well-known inductive effect exhibited by this mAb on the nonspecific reactivity of CTL (26, 29). The functional evaluation of heart TIL at clonal level and the study of the frequency of donor-specific CTL and precursors in the peripheral blood might provide supplementary data on the role of MHC-restricted and MHC-unrestricted cytotoxic cells in the heart rejection mechanisms. ACKNOWLEDGMENTS The authors thank the Ortho Pharmaceutical (Raritan, NJ) for providing the anti-CD3 mAb, Dr. Antonella Milani for expert technical assistance,Dr. Gianpietro Semenzato for helpful discussions,and Mr. Martin Donach for his help in the preparation of the manuscript. Supported in part by the National Council for Research, target project “BBS,” Rome, Italy. L. Trentin is a recipient of a fellowship from the National Council for Research,Rome. R. Zambello is recipient of a fellowship from Istituto Superiore di Sanid, Rome.

REFERENCES I. 2. 3. 4.

Dallman, M. J., and Mason, D. W., Transplantation 33, 221, 1982. Hancock, W., Immunol. Rev. 17, 61, 1984. Hall, M. H., and Dorsch, S. E., Immunol. Rev. 77, 31, 1984. Mayer, T. G., Fuller, A. A., Fuller, T. C., Lazarovits, A. I., Boyle, L. A., and Kurnick, J. T., J. Immunol. 134,258, 1985.

5. Hoshinaga, K., Mohanakumar, T., Goldman, M. H., Wilfgang, T. C., Szentpetery, S., Lee, H. M., and Lower, R. R., Transplantation 38, 634, 1985. 6. Zeevi, A., Zerbe, T. R., Kaufman, C., Rabin, B. S., Griffith, B. P., Hardesty, R. L., and Dusquesnoy, R. J., Transplantation 41, 620, 1986. 7. Orosz, C. G., Zinn, N. E., Sirinek, L. P., and Ferguson, R. M., Transplantation 41,84, 1986.

GRAFT INFILTRATING

LYMPHOCYTES IN HEART TRANSPLANTS

341

8. Krensky, A. M., Weiss, A., Crabtree, G., Davis, M. M., and Parham, P., N. Engl. J. Med. 322, 510, 1990. 9. Orosz, C. G., Zinn, N. E., Sirinek, L., and Ferguson, R. M., Transplantation 41, 75, 1986. 10. Orosoz, C. Cl., Bishop, D. K., and Ferguson, R. M., Transplantation 48, 818, 1989. 11. Totterman, T. H., Hanas, E., Bergstrom, R., Larsson, E., and Tufveson, G., Transplantation 47, 8 17, 1989. 12. Nakamura, H., and Gress, R. E., Transplantation 49,453, 1990. 13. Miltenburg, A. M., Meijer-Paape, M. E., Daha, M. R., van Bockel, J. H., Weening, J. J., van Es, L. A., and van der Woude, F. J., Transplantation 48, 296, 1989. 14. Ratliff, N. B., Myles, J. L., McMahon, J. T., Golding, L., Hobbs, R., Rincon, G., Sterba, R., and Stewart, R., Tranplant. Proc. 19, 2568, 1987. 15. Knapp, W., Dorken, B., Rieber, P., Schmidt, R. E., Stein, H., and von dem Borne, A. E. G. Kr., “Leucocyte Typing IV” Oxford Univ. Press,Oxford, 1990. 16. Band, H., Hochstenbach, F., McLean, J., Hata, S., Krangel, M. S., and Brenner, M. B., Science 238, 682,1987.

17. Trentin, L., Migone, N., Zambello, R., Francis di Celle, P., Aina, F., Feruglio, C., Bulian, P., Masciarelli, M., Agostini, C., Cipriani, A., Marcer, G., Foa, R., Pizzolo, G., and Semenzato, G., J. Immunol. 145,2147, 1990. 18. Trentin, L., Zambello, R., Agostini, C., Ambrosetti, A., Chisesi, T., Raimondi, R., Bulian, P., Pizzolo, G., and Semenzato, G., Blood 75, 1525, 1990. 19. Tufveson, G., Forsum, U., Claesson, K., Klareskog, L., Larsson, E., Karlsson-Parra, A., and Frodin, L., &and. J. Immunol. 18, 37, 1983. 20. Hall, B. M., Bishop, G. A., Farnsworth, A., Duggin, G. C., Horwath, D. S., Sheil, A. G. R., and Tiller, D. J., Transplantation 37, 564, 1984. 21. Hancock, W. H., Gee, D., De Moerloose, A., Rickles. F. R., Ewan, V. A., and Atkins, R. C., Transplantation 41, 116, 1986. 22. Budd, R. C., Cerottini, J. C., Horvath, C., Bron, C., Pedrazzini, T., Howe. R. C., and MacDonald, H. R., J. Immunol. 138, 3120, 1987. 23. Sanders,M. E., Makgoba, M. W., and Shaw, S., Immunol. Today9, 195, 1988. 24. Akbar, A. N., Terry, L., Timms, A., Beverly, P. C., and Janossy,G., J. Immunol. 140, 2171, 1988. 25. Bell, E. B., and Sparshott, S. M., Nature348, 163, 1990. 26. Mentzer, S. J., Barbosa, J. A., and Burakoff, S. J., J. Immunol. 134, 34, 1985. 27. Van Seventer, G. A., Kuijpers, K. C., Van Lier, R. A. W., De Groot, E. R., Aarden, L. A., and Melief, C. J. M., J. Immunol. 139, 2545, 1987. 28. Jung, G., Martin, D. E., and Muller-Eberhard, H. J., J. Immunol. 139, 639, 1987. 29. Spits, H., Yssel, H., Leeuwenberg, J., and de Vries, J. E., Eur. J. Immunol. 15, 88, 1985.