Induction and large-scale expansion of CD8+ tumor specific cytotoxic T lymphocytes from peripheral blood lymphocytes by in vitro stimulation with CD80-transfected autologous melanoma cells.

ARTICLE

INTRODUCTION

The existence of MHC-restricted cytotoxic T lymphocytes (CTL) with specificity for tumor-associated antigens has been demonstrated in melanoma patients. Numerous T cell epitopes of melanoma cells have been identified, and several genes encoding the respective proteins have been cloned [1]. Furthermore, tumor-specific CTL have been directly implicated in spontaneous regression of malignant melanoma lesions [2]. In the majority of patients however, tumor cells are not eliminated by the immune system. One reason for the poor immunostimulatory function of tumor cells may be their inability to provide sufficient costimulatory signals [3]. Only the combination of signals provided by the TCR and costimulatory signals results in full activation of T cells. Stimulation of the TCR in the absence of costimulatory signals leads to unresponsiveness and a state of antigen-specific tolerance [4]. CD80 (B7.1), which is expressed by professional antigen-presenting cells such as dendritic cells, was shown to provide a key signal for the activation of resting T cells and the prevention of anergy [5].

In vitro studies with human lymphocytes have shown that transfection of CD80 into tumor cells enhances proliferation, cytokine release and induction of cytotoxic activity against autologous and allogeneic tumor cells [6-8]. However, the role of costimulatory molecules such as CD80 in long-term expansion of tumor specific CTL has not specifically been addressed.

In this study, we have developed a human CD80 vector system to transduce human melanoma cells and investigated the capacity of CD80 expressed on melanoma cells to induce primary T cell responses and to allow long-term expansion of tumor specific T cells for an adoptive transfer in tumor immunotherapy.

MATERIALS AND METHODS

Tumor cell lines

Fresh tumor cells were obtained from biopsies from 3 melanoma patients (MeR190, MeE384, MeT413), minced mechanically and subjected to digestion with an enzymatic cocktail. After passage through a 0.2 mm mesh, cells were cultured in RPMI 1640 supplemented with 10% fetal calf serum (FCS) (PAN Systems GmbH, Aidenbach, Germany), amino acids, pyruvate, and antibiotics. Tumor cells were maintained in monolayer and passaged with 0.05% trypsin. Early passages were cryopreserved with FCS plus 10% dimethylsulfoxide.

Transfection of melanoma cells

A molecular clone encoding human CD80 was obtained by PCR amplification of total cDNA derived from an EBV-transformed lymphoblastoid cell line with CD80-specific primers. The PCR product was verified by cloning into a TA vector and DNA sequencing. The G-CSF cDNA of pCMV.GCSF.ires.NEO [9] was replaced with the human CD80 cDNA to obtain pCMV.CD80.ires.NEO. The vector expresses a dicistronic mRNA for CD80 and neomycin phosphotransferase under control of the immediate/early CMV promoter. Melanoma cells were transfected with linearized pCMV.CD80.ires.Neo by cationic lipofection with DOSPA/DOPE (Lipofectamin, Gibco BRL, Eggenstein, Germany). Stably transfected melanoma cell clones were isolated under selection with 0.5 mg/ml G418.

Expression of melanoma antigens by reverse transcription-polymerase chain reaction (RT-PCR)

The poly-A mRNA was isolated from melanoma cell lines using the Quick Prep Micro mRNA purification kit (Pharmacia, Dubendorf, Switzerland) according to the manufacturer's instructions and as decribed [10]. First strand cDNA was synthesized from poly-A mRNA by taking advantage of a commercial RT-PCR kit (Perkin Elmer Cetus Instruments, Norwalk, CT, USA). The following primers were used: ß-actin sense, 5'-CACCCACACTGTGCCCATC; ß-actin antisense, 5'-CTAGAAGCATTTGCGGTGGAC, amplifying a 650-bp gene fragment; gp100 sense, 5'-CTGTGCCAGCCTGTGCTAC; gp100 anti-sense, 5'-CACCAATGGGACAAGAGCAG, amplifying a 334-bp fragment; melan-A/MART-1 sense, 5'-AGATGCCAAGAGAAGATGCTC; melan-A/MART-1 anti-sense, 5'-GCTCTTAAGGTGAATAAGGTGG, amplifying a 364-bp gene fragment; tyrosinase sense, 5'-TTGGCAGATTGTCTGTAGCC; tyrosinase anti-sense, 5'-AGGCATTGTGCATGCTGCTT, amplifying a 284-bp gene fragment; TRP-2 sense, 5'-GACTCTGATTAGTCGGAACTC; TRP-2 anti-sense, 5'-GAAATGTGGCAAAGCGTTTGTC, amplifying a 344-bp fragment and MAGE-3 sense, 5'-TGGAGGACCAGAGGCCCCC; MAGE-3 anti-sense, 5'-GGACGATTATCAGGAGGCCTGC, amplifying a 725-bp fragment.

PCR in the case of ß-actin, melan-A/MART-1, gp100, tyrosinase, and TRP-2 was cycled 30 and 35 times using the following profile: 20 sec denaturation at 94° C, 20 sec annealing at 58° C and 40 sec extension at 72° C using "hot-start" technique. For detection of MAGE-3 a different PCR profile was used: 1 min denaturation at 94° C was followed by 2 min annealing at 72° C and 2 min extension at 72° C. PCR products were run on 1.5% agarose gels containing
ethidium bromide and photographed under transillumination.

Isolation of lymphocytes and induction of CTL

Peripheral blood from melanoma patients was collected in heparin, diluted in PBS and peripheral blood mononuclear cells separated by Ficoll. PBMC (5 x 10⁶/well) were cultured with irradiated (10,000 rads) tumor cells at an effector/tumor ratio of 2:1 in 2ml of medium consisting of RPMI 1640 supplemented with 10% human AB serum, amino acids, pyruvate, antibiotics, 50 units/ml human recombinant IL-2 (EuroCetus, Amsterdam, Netherlands), and T cell growth factor (TCGF) (3% final dilution) in 24-well tissue culture plates [11]. The culture medium was replenished every 3 days. The different stimulatory conditions consisted either of autologous melanoma cells, autologous melanoma cells supplemented with anti-CD28 mAb (L293, IgG₁, Becton Dickinson, BD, San Jose, CA, USA) or autologous CD80-transfected melanoma cells. After 7 days of culture, T cells were harvested, counted, and phenotyped by FACS analysis. For further expansion, T cells were restimulated under the same conditions as mentioned above. To analyze the capacity of different cytokine combinations, PBMC were cultured for 7 days with CD80-transfected melanoma cells, along with the indicated cytokines: human recombinant IL-6 (Genzyme, Cambridge, MA, USA) was used at 1,000 U/ml and IL-12 (Sigma-Aldrich Chemie GmbH, Deisenhofen, Germany) at 10 ng/ml. For restimulation, additional stimulator cells were added along with IL-2 (10 IU/ml) and IL-7 (BioConcept GmbH, Umkirch, Germany, 5 ng/ml).

Preparation of TCGF

Preparation of TCGF was described previously [12]. TCGF was produced by stimulating 2.5 x 10⁶/ml PBMC for 2 hours with 5 µg/ml phytohemagglutinin (Murex, Dartford, England), 5 ng/ml phorbol myristate acetate (Sigma-Aldrich) and 5,000 rad-irradiated EBV-transformed B-cells. The cells were then washed to remove the mitogens and resuspended in RPMI 1640 supplemented with 2.5% human AB serum. After 40 hours of incubation, supernatants were harvested, passed through 0.2 µm filters and stored at ­ 70° C.

Phenotypic analysis

Tumor cells were incubated with mAbs against HLA class I and II [anti-HLA I (W6/32, Dako, Glostrup, Denmark), anti-HLA-DR (L243, BD), CD80 (PharMingen, Hamburg, Germany), CD54 (ICAM-1, BD), and MEL-1 (R24, Biomedical Diagnostic, Rödermark, Germany), identifying the G₀₃ ganglioside on melanoma cells. T cells were stained with FITC-or PE-conjugated mAbs against CD3 (X35), CD4 (BL4), and CD8 (B9.2), purchased from Immunotech (Marseille, France) and CD56 (MY31) from BD. Analysis was performed on a Beckton-Dickinson FACScan flow cytometer.

Cytotoxicity assays

The cytotoxic activity of the T cell lines was measured by a conventional 4-h ⁵¹Cr release assay using triplicate cultures in V-bottomed plates. Cell lines analyzed as control targets included a human allogeneic melanoma cell line (M10), the NK target K562, and autologous EBV-transformed lymphoblastoid cell lines (EBV-B). E:T ratios were 25:1, 5:1, 1:1, and 0.2:1 on 2,000 target cells/well. The percentage of specific cytotoxicity was calculated conventionally; SD were < 5%. Functional effects of the antibodies on target cells (anti-HLA class-I, control mAb) were tested by incubating each of them for 2 hours at 37° C before performing the assay at the predetermined saturating concentration. The percentage of inhibition of lysis was calculated as 1 ­ % specific lysis in mAb-treated wells/% specific lysis in control wells x 100.

RESULTS

Transduction of human melanoma cell lines with the CD80 gene

Three human melanoma cell lines were transfected by cationic lipofection with the expression construct pCMV.CD80.ires.Neo, and numerous stably transfected clones were isolated. Clones MeR190.15, MeE384.1, and MeT413.1 retained satisfying growth characteristics and expressed high levels of CD80 (Figure 1).

Regulation of surface marker and melanoma-associated antigen expression in CD80-transfected melanoma cells

Phenotypic analysis of melanoma cells was performed simultaneously on CD80-transfected tumor cells and on parental melanoma cells. Cells were stained with fluorescence-labeled mAbs against HLA-class I, class II, CD80, CD54, and MEL-1. After transfection there were no changes in expression of HLA antigens, CD54 or MEL-1 (Figure 1). CD80 expression on parental melanoma cells was negative with the exception of MeT413, showing low expression of CD80. Transfected melanoma cells expressed high levels of the costimulatory molecule CD80 (Figure 1).

Total cellular RNA was extracted from parental and CD80-transfected melanoma cells and reverse transcribed. Expression of gp100, tyrosinase, melan-A, MAGE-3, and TRP-2 genes was analyzed by RT-PCR. The data obtained are shown in Figure 2. The amount of transcripts from the ß-actin house-keeping gene was similar in all samples tested. Parental and CD80-transfected melanoma cells expressed comparable levels of melan-A and tyrosinase in all cell lines tested (Figure 2). However, gp100 and TRP-2 revealed a different expression pattern: identical expression in one CD80-transfected melanoma line (MeR190.15), upregulation of gp100 and TRP-2 in MeE384.1 cells, and downregulation in MeT413.1. Expression of MAGE-3 decreased in two CD80-transfected melanoma lines while it was upregulated after CD80 transfection of line MeT413.1.

CD80-transfected melanoma cells, but not parental or anti-CD28-treated melanoma cells, induce tumor specific T cell responses in patients' PBL

PBL from 3 melanoma patients were stimulated in vitro for 4 weeks either with irradiated parental, parental/anti-CD28 or CD80-transfected autologous melanoma cells. Cells were restimulated every 7 days. After 4 weeks of culture, T cells were tested in a standard ⁵¹Cr release assay against a panel of autologous and allogeneic target cells. As shown in Figure 3A for patient MeR190, PBMC stimulated with CD80-transfected autologous melanoma cells displayed high levels of cytotoxicity against the autologous, parental melanoma cell line in all patients tested, while being incapable of lysing other control targets, including allogeneic melanoma and the NK target K562. Blocking experiments indicated that these T cells have the conventional profile of HLA class I-restricted CTL: cytotoxic activity against autologous melanoma cells was inhibited > 50% by W6/32 (anti-HLA class I monomorphic mAb), while control antibodies had no effect (data not shown).

PBL stimulated with parental melanoma cells either alone or with the addition of anti-CD28 mAb, both demonstrated only limited cytotoxic activity against the autologous melanoma cells (Figure 3B + C).

The phenotype of the cultures was measured by flow cytometry and shown in Table 1. At the beginning of the in vitro stimulation, the percentage of CD3⁺ T cells was 73%, 49% of them were CD4⁺ T cells, 25% CD8⁺ T cells: 13% of the PBL were CD3^­/CD56⁺ NK cells. By the end of the 4-week culture period with autologous CD80-transfected tumor cells, all cultures contained more than 92% CD3⁺/CD8⁺ T cells, while CD4⁺ and CD56⁺ cells decreased to 13% and 4% respectively (see Table 1). Essentially similar results, albeit to a lesser extent, were obtained with parental melanoma cells +/- anti-CD28 mAb (see Table 1). These results demonstrate the exceptional capacity of CD80 stimulation to enable expansion of CD8⁺ T cells in vitro.

CD80-transfected melanoma cells induce large-scale expansion of CD8⁺ T cells suitable for an adoptive transfer

Our goal was to define optimal culture conditions to obtain large numbers of tumor-specific CTL for an adoptive transfer in melanoma immunotherapy. Melanoma PBMC were cultured in RPMI/10% human AB serum supplemented with 50 IU rIL-2/ml and 3% TCGF. Stimulator cells used were either autologous unmodified melanoma cells with/without the addition of anti-CD28 mAb or CD80-transfected autologous melanoma cells. Cell yields of CD8⁺ T cells after 7, 14, 21, and 28 days of culture respectively are shown in Figure 4. It is evident that stimulation with CD80-transfected melanoma cells provided the most efficient conditions for the generation of large numbers of CD8⁺ tumor-specific CTL, demonstrating an 88-119-fold expansion of CD8⁺ T cells after 4 weeks of culture (Figure 4). In contrast, T cells stimulated either with parental tumor cells or parental tumors cells, together with anti-CD28 mAb demonstrated mean increases of 30-fold and 33-fold respectively (Figure 4).

Addition of IL-6 and IL-12 to CD80-transfected melanoma cells failed to induce activation and expansion of autologous tumor-specific CTL

It has been shown by Gajewski et al. that costimulation with CD80, IL-6, and IL-12 are potent inducers of both generation and proliferation of murine antitumor CTL in vitro. These cells can be further expanded by restimulation with IL-2/IL-7 [4]. We therefore analyzed whether the addition of IL-6/IL-12 to CD80-transfected melanoma cells could augment the generation of human melanoma-specific CTL. Autologous PBL were co-cultured with CD80-transfected melanoma cells in the presence of rIL-6 (20 ng/ml) and rIL-12 (0.2 ng/ml). After 7 days, cells were further restimulated weekly with CD80-transfected melanoma cells in the presence of IL-2 (50 U/ml), IL-7 (1 ng/ml). As shown in Figure 5, IL-6/IL-12 failed to induce sufficient cytotoxic activity and proliferation of autologous PBL stimulated with CD80-transfected melanoma cells. In addition, 2 of the 3 patients' PBL revealed no proliferation and died after 7 days of stimulation. In contrast, PBL stimulated with CD80-transfected autologous melanoma cells in the presence of IL-2 (50 U/ml) and TCGF (3%) exhibited potent cytotoxic activity and proliferation of CD8⁺ T cells (Figure 5). Thus, the combination of CD80-transfected melanoma cells together with IL-2 and TCGF appears to stimulate both optimal proliferation and acquisition of tumor-specific activity of autologous CD8⁺ CTL.

DISCUSSION

Attempts to generate and expand human tumor-specfic T cells in vitro for an adoptive transfer in tumor immunotherapy have been conducted with limited success. This difficulty may have resulted, in part, from the very low frequency of tumor-specific CTL precursors in patients' blood that may be due to the fact that tumor cells are not potent APC. We therefore investigated in vitro conditions that would improve the antigen presenting function of tumor cells and facilitate the generation and long-term expansion of tumor-specific CTL. The results of our study are consistent with the results of Sule-Suso et al. [13] and Yang et al. [8] which demonstrated that melanoma-specific CTL can be generated in vitro from patients' PBL by stimulation with autologous melanoma cells expressing the costimulatory molecule CD80. Our work extends these findings to include an evaluation of the capacity to expand CTL by repetitive stimulation via the CD80/CD28 pathway. It can be demonstrated that (A) CD8⁺ melanoma-specific CTL stimulated with CD80⁺ but not CD80^­ autologous tumor cells could be expanded by up to 100-fold by weekly restimulation with tumor cells in the absence of professional APC. The results suggest that in our system, CD80 costimulation appears to be essential not only to activate and expand tumor-specific CTL but also to prevent the induction of anergy and apoptosis during a culture period of 4 weeks. (B) Activation and restimulation of T cells with autologous tumor cells and anti-CD28 mAbs binding to monocytes via the Fc receptor proved to be less potent in the induction and expansion of tumor-specific CTL than stimulation with CD80⁺ autologous melanoma cells. The role of the CD28 molecule as a costimulatory pathway has been demonstrated by Harada et al. demonstrating that anti-CD28 mAbs could stimulate CD8⁺ T cells together with anti-CD3 mAbs leading to increased IL-2 production and T cell proliferation [14]. (C) Several factors have been shown to influence the activation and expansion of tumor-specific CTL. We explored the effect of different cytokine combinations on CD80-induced costimulation of melanoma-specific CTL. The combination of CD80, IL-6, and IL-12 has been described to be optimal for inducing murine tumor-specific CTL in vitro [15]. In addition, IL-7 and IL-12 have been reported to augment long-term proliferation of previously activated T lymphocytes [15, 16]. In our study, substitution of IL-6 plus IL-12 to CD80⁺ melanoma cells failed to induce a strong CTL response against autologous melanoma cells. The T cells cultured under these conditions revealed no proliferative activity and died after 1-2 weeks of culture. In contrast, the addition of low levels of exogenous IL-2 to CD80⁺ autologous melanoma cells elicited a strong CTL-reponse against the autologous tumor and allowed optimal expansion during a culture period of 4 weeks. These results suggest that the combination of CD80 and IL-2 is superior in its synergistic capacity to induce human tumor specific CTL when compared to that of CD80 and IL-12, as described by Coughlin et al. [17].

Our findings underscore the importance of CD80/CD28 interactions during the induction of primary antitumor T cell responses as supported by recent data demonstrating that costimulation via CD80 was superior to CD86 in the induction of T cell-mediated antitumor responses [18, 19].

Costimulation induced by an interaction between molecules of the B7 family (CD80, CD86) on APC, and CD28 on T cells plays a role in the induction of both CD4⁺ and CD8⁺ T cell responses against tumor antigens [3, 20]. In the present study, stimulation of resting PBL with CD80-transfected melanoma cells resulted in predominant activation and expansion of CD8⁺ CTL. This may be due to the fact that these melanoma cell lines express MHC class I molecules but not MHC class II molecules as is the case for the majority of primary tumor cells.

Total numbers of CTL generated under these conditions may be adequate for effective treatment when compared to approaches for successful adoptive T cell transfer in the setting of viral or virus-associated diseases in bone marrow chimeras.

Thus, donor leukocytes with a T cell content of only 10⁶-10⁷/kg are capable of inducing remissions in CMV-infections [21] and EBV-induced lymphoproliferative diseases [22].

There are some limitations to the CD80-costimulation approach that have to be borne in mind. Firstly, the most important limiting factor of this method is the difficulty in establishing a proliferating autologous tumor cell line. Secondly, that the tumor is in fact immunogenic, delivering the tumor antigen on MHC class I molecules. It has been shown that CD80 transfection of poorly immunogeneic human tumor cells is not sufficent to induce tumor-specific CTL in vitro [23]. The addition of cytokines such as IL-2 and IL-12, however, overcomes this problem [23]. Thirdly, positive selection of CD80-transduced melanoma cells may generate antigen-loss variants. In the present study, we have observed downregulation of the MAGE-3 gene in 2/3 melanoma cell lines after selection of CD80⁺ tumor cells. The expression of other melanoma antigens tested, melan-A, gp100, tyrosinase, TRP-2, was not affected during selection of CD80+ tumor cells.

In summary, we show that autologous CD80-transfected melanoma cell lines are suitable for inducing and expanding CD8⁺ melanoma-specifc CTL that can be used for an adoptive transfer in tumor immunotherapy. Further studies will show whether professional APC loaded with apoptotic bodies and tumor-specific peptides or these artificial tumor cell APC are the optimal stimuli to generate these effector cells.

CONCLUSION

Acknowledgement.

We gratefully acknowledge the excellent technical assistance of Iris Hentrich, Ursula Möhrle, and Sabine Lilli. This study was supported by the Deutsche Forschungsgemeinschaft (Sonderforschungsbereich 364) and by the Zentrum für Klinische Forschung Freiburg (Project C2).

REFERENCES

1. van den Eynde B J, van der Bruggen P. 1997. T cell defined tumor antigens. (Review) Curr. Opin. Immunol. 9: 684.

2. Mackensen A, Carcelain G, Viel S, Raynal MC, Michalaki H, Triebel F, Bosq J, Hercend T. 1994. Direct evidence to support the immunosurveillance concept in a human regressive melanoma. J. Clin. Invest. 93: 1397.

3. Chen L, Linsley PS, Hellström KE. 1993. Costimulation of T cells for tumor immunity. (Review) Immunol. Today 14: 483.

4. Gimmi C D, Freeman G J, Gribben J G, Gray G, Nadler L M. 1995. Human T-cell clonal anergy is induced by antigen presentation in the absence of B7 costimulation. Proc. Natl. Acad. Sci. USA 90: 6586.

5. Lenschow D J, Walunas T L, Bluestone J A. 1996. CD28/B7 system of T cell costimulation. (Review) Annu. Rev. Immunol. 14: 233.

6. Döhring C, Angman L, Spagnoli G, Lanzavecchia A. 1994.
T-helper- and accessory-cell-independent cytotoxic responses to human tumor cells transfected with a B7 retroviral vector. Int. J. Cancer 57: 754.

7. Imro M A, Dellabona P, Manici S, Heltai S, Consogno G, Bellone M, Rugarli C, Protti M P. 1998. Human melanoma cells transfected with the B7-2 co-stimulatory molecule induce tumor-specific CD8⁺ cytotoxic T lymphocytes in vitro. Hum. Gene Ther. 9: 1335.

8. Yang S, Darrow T L, Seigler H F. 1997. Generation of primary tumor-specific cytotoxic T lymphocytes from autologous and human lymphocyte antigen class I-matched allogeneic peripheral blood lymphocytes by B7 gene-modified melanoma cells. Cancer Res 57: 1561.

9. Veelken H, Re D, Kulmburg P, Rosenthal F M, Mackensen A, Mertelsmann R, Lindemann A. 1996. Systemic evaluation of chimeric marker genes on dicistronic transcription units for regulated expression of transgenes in vitro and in vivo. Hum. Gene. Ther. 7: 1827.

10. Noppen C, Spagnoli G C, Schaefer C. 1996. Isolation of multiple mRNAs from a few eukaryotic cells: a fast method to obtain templates for RT-PCR. Biotechniques 21: 394.

11. Mackensen A, Veelken H, Lahn M, Wittnebel S, Becker D, Köhler G, Kulmburg P, Brennscheidt U, Rosenthal F, Franke B, Mertelsmann R, Lindemann A. 1997. Induction of tumor-specific cytotoxic T lymphocytes by immunization with autologous tumor cells and interleukin-2 gene transfected fibroblasts. J. Mol. Med. 75: 290.

12. Hercend T, Meuer S, Brennan A, Edson M A, Acuto O, Reinherz E L, Schlossman S F, Ritz J. 1983. Identification of a clonally restricted 90 kD heterodimer on two human cloned natural killer cell lines. J. Exp. Med. 158: 1547.

13. Sule-Suso J, Arienti F, Melani C, Colombo M P, Parmiani G. 1995. A B7-1-transfected human melanoma line stimulates proliferation and cytotoxicity of autologous and allogeneic lymphocytes. Eur. J. Immunol. 25: 2737.

14. Harada M, Okamoto T, Omoto K, Tamada K, Takenoyama M, Hirashima C, Ito O, Kimura G, Nomoto K. 1996. Specific immunotherapy with tumor-draining lymph node cells cultured with both anti-CD3 and anti-CD28 monoclonal antibodies. Immunology 87: 447.

15. Gajewski T F, Renauld J C, Van Pel A, Boon T. 1995. Costimulation with B7-1, IL-6, and IL-12 is sufficient for primary generation of murine antitumor cytolytic T lymphocytes in vitro. J. Immunol. 154: 5637.

16. Lynch D H, Miller R E. 1994. Interleukin-7 promotes long-term in vitro growth of antitumor cytotoxic T lymphocytes with immunotherapeutic efficacy in vivo. J. Exp. Med. 179: 31.

17. Coughlin C M, Wysocka M, Kurzawa H L, Lee W M, Trinchieri G, Eck S L. 1995. B7-1 and interleukin 12 synergistically induce effective antitumor immunity. Cancer Res. 55: 4980.

18. Gajewski T F, Fallarino F, Uyttenhove C, Boon T. 1996. Tumor rejection requires a CTLA4 ligand provided by the host or expressed on the tumor: superiority of B7-1 over B7-2 for active tumor immunization. J. Immunol. 156: 2909.

19. Matulonis U, Dosiou C, Freeman G, Lamont C, Mauch P, Nadler L M, Griffin J D. 1996. B7-1 is superior to B7-2 costimulation in the induction and maintenance of T-cell mediated anti-leukemia immunity. Further evidence that B7-1 and B7-2 are functionally distinct. J. Immunol. 156: 1126.

20. Li Y, McGowan P, Hellström I, Helström K E, Chen L. 1994. Costimulation of tumor-reactive CD4⁺ and CD8⁺ T lymphocytes by B7, a natural ligand for CD28, can be utilized to treat established mouse melanoma. J. Immunol. 153: 421.

21. Walter E A, Greenberg P D, Gilbert M J, Finch R J, Waranabe K S, Thomas E D, Riddell S R. 1995. Reconstitution of cellular immunity against cytomegalovirus in recipients of allogeneic bone marrow by transfer of T cell clones from the donor. N. Engl. J. Med. 333: 1038.

22. Papadopoulos E B, Ladanyi M, Emanuel D, Mackinnon S, Boulard F, Carabasi M H, Castro-Malaspina H, Childs B H, Gillio A P, Small T N, Young J W, Kernan N A, O'Reilly R J. 1994. Infusions of donor leukocytes to treat Epstein-Barr virus-associated lymphoproliferative disorders after allogeneic bone marrow transplantation. N. Engl. J. Med. 330: 1185.

23. Mogi S, Ebata T, Setoguchi Y, Fujime M, Heike Y, Kohsaka T, Yagita H, Okumura K, Azuma M. 1998. Efficient generation of autologous peripheral blood-derived cytotoxic T lymphocytes against poorly immunogenic human tumors using recombinant CD80-adenovirus together with interleukin 12 and interleukin 2. Clin. Cancer Res. 4: 713.