Third International Dresden Symposium on Immunotherapy of Cancer¹ Free

The Third International Dresden Symposium was held in Dresden, Germany,on June 2 and 3, 2000, and was exclusively devoted to actual aspects of tumor immunotherapy. A major purpose of the Symposium was to bring together leading experts representing various fields of cancer research in an attempt to discuss recent achievements and future avenues in the development of tumor-directed immunotherapy. A primary goal of the Dresden Symposia, which take place every other year, is to provide a forum that encourages discussions between scientists and clinicians to facilitate the transfer of laboratory findings to the bedside. The schedule of the meeting was designed to cover a broad spectrum of current immunotherapy. The main topics discussed were:(a) modulation of tumor growth; (b) newly identified TAAs³ and their use in the tumor-specific activation of T cells;(c) the emerging pivotal role of DCs in immunotherapy;(d) tumor-directed monoclonal antibodies and recombinant constructs thereof and their use in clinical trials; and (e)clinical trials based on the adoptive transfer of allogeneic effector cells.

A thorough understanding of the interplay of tumor cells and the immune system is of particular concern in attempts to engage immune effector mechanisms in the treatment of malignant diseases. The Symposium was opened by Dr. Margaret L. Kripke (University of Texas,Houston, TX), who discussed the modulation of tumor-directed immune reactivity by UV-B irradiation.

The UV-induced skin cancer in mice is a highly immunogenic tumor that,after transplantation into a normal host, is rejected by a T-cell immune response but is tolerated by animals that had been exposed to UV irradiation. In another model, melanomas in the ear skin of irradiated mice were found to grow much better than in nonirradiated animals. Even in mice showing a high level of immunity against melanoma after preimmunization against melanoma cells, the local UV irradiation depressed the immune response, allowing the tumor to grow progressively. This interference with tumor rejection is observed early after exposure to UV irradiation and is suggested to be mediated by suppressor T lymphocytes. Thus, when T cells prepared from UV-irradiated animals were transferred to irradiated hosts, these animals developed skin cancer earlier than mice without transferred suppressor T cells. This immune modulation ultimately can be attributed to the alteration of APCs by UV irradiation, as evidenced by experiments involving sensitization to FITC. When FITC was painted on the skin, it was absorbed by APCs, which could then easily be followed to the draining lymph node. Cells taken from a draining lymph node induced contact hypersensitivity against FITC when injected into the footpad of another mouse. Cells prepared from mice after skin irradiation by UV light generated only a weak immune response. There is experimental evidence that DNA damage induced by UV irradiation impairs the ability of APCs to initiate contact hypersensitivity to FITC. In addition, spleen cells from UV-irradiated mice were able to suppress the immune response in normal hosts, showing that suppressor cells reside in the spleen. These suppressor cells may be descendants of T cells that are found in lymph nodes closely surrounding APCs modified by UV-light. Repairing the DNA damage in APCs restored their ability to induce contact hypersensitivity and eliminated the activity of T suppressor cells.

Dr. Isaiah J. Fidler (University of Texas) discussed angiogenesis as an example of the continuous interactions between metastatic cells and homeostatic mechanisms that tumor cells can manipulate. The extent of angiogenesis is determined by the local balance between proangiogenic molecules, such as bFGF, vascular endothelial growth factor, and IL-8, and antiangiogenic molecules, such as IFN-α or IFN-β. In previous studies, Fidler’s group has shown that IFN-α or -β inhibit transcription and protein production of bFGF and matrix metalloproteinase-2 and -9. In normal tissues under homeostatic control, the proangiogenic and antiangiogenic molecules counterbalance each other. Whether neoplastic angiogenesis is attributable to overexpression by proangiogenic molecules or to the absence of antiangiogenic molecules was examined in a collaborative study with Dr. Judah Folkman and Dr. John Mulliken (Harvard Medical School, Boston, MA). Folkman’s group demonstrated that daily administration of IFN-α can produce involution of fatal infantile hemangiomas. By immunohistochemical analysis of biopsies from normal skin or skin overlaying proliferating, involuting, or involuted hemangiomas, Fidler’s group showed that in proliferating lesions,there is a local absence of IFN-β expression. Once the tumors involuted, IFN-β was reexpressed. To determine whether tumor cells can induce hyperplasia in surrounding tissues, human melanoma cells were injected into the dermis of nude mice. The resulting hyperplastic epidermis expressed high levels of bFGF, VEGF, and IL-8, but not IFN-β. The epidermis distant from the tumor expressed both proangiogenic and antiangiogenic molecules.

Differentiated epithelial cells in tissues that line portals of entry to the body (skin, genitourinary tract, gastrointestinal tract, etc.)express high levels of IFN-β. In contrast, dividing epithelial cells do not. Because many human carcinoma cells have a deletion (loss of heterozygosity) of chromosome 9p, the cells cannot produce IFN-β and, hence, the balance of angiogenic molecules favors angiogenesis.

IFN-α has been widely used alone or in combination with other agents to inhibit cell proliferation and, hence, investigators have used high doses given 2–3 times per week. Recent data from Fidler’s laboratory showed that optimizing dose and schedule of IFN-α is necessary for maximal down-regulation of angiogenesis-related genes, and, hence,inhibition of angiogenesis in neoplasms. Given the experience of hemangiomas with IFN-α therapy, it appears that the maximal response to the antiangiogenic effect of IFN-α requires chronic frequent administration of biologically optimal doses.

The unequivocal demonstration that tumor-associated peptides are recognized by CTLs has stimulated worldwide efforts to identify as many of such antigens as possible and their HLA-restricted presentation by various tumor types. Dr. Hans-Georg Rammensee (University of Tuebingen, Tuebingen, Germany) discussed new promising strategies to predict relevant epitopes presented naturally by tumor cells for recognition by CTLs. A widely used strategy for the identification of T-cell epitopes is to predict HLA-restricted peptides out of the sequence of tumor-associated proteins by the use of appropriate algorithms. These peptides were used for the stimulation of T cells, which were then tested for recognition of tumor cells. The main reason for the frequent failure of this approach is that the predicted peptides are not naturally processed and are therefore not relevant for tumor-cell recognition. Therefore, a biochemical strategy to obviate these failures has been developed. In a first step, the potential peptides are predicted, synthesized, and analyzed by HPLC-linked mass spectrometry. In a second step the naturally produced peptides are eluted from the MHC class I molecules of solid tumors or of tumor cell lines. A mixture of predicted synthetic peptides is then checked against the great number of eluted peptides using a specially adapted computer software. Thus identified, naturally occurring tumor peptides are used for the activation of T cells, and the T cells are tested for tumor cell recognition.

Another promising strategy to improve the prediction of relevant cytotoxic T-cell epitopes comprises the combination of HLA-motif prediction and determination of the cleavage specificity of proteasomes. Proteasomal cleavage is essential for the generation of peptides that are delivered to MHC class I molecules for presentation on the cell surface. To analyze the proteasomal specificity, the protein of interest is degraded in vitro by purified proteasome particles. The cleavage products are purified by HPLC and analyzed further by a combination of Edman degradation and mass spectometry. This approach provides a basis for predicting proteasomal degradation products from which peptides are sampled by MHC class I molecules.

Dr. Ugur Sahin (University of Homburg, Homburg, Germany)outlined that the combination of serological identification of antigens by recombinant expression cloning (SEREX) with T cell-based technologies allowed the identification of many tumor antigens in different tumor tissues. However, for some tumors, the number of such tumor proteins is rather limited. An alternative approach to identify transcripts that are selectively expressed in tumor tissues is based on cDNA microarray analysis. A combination of SEREX or T-cell analysis with cDNA microarray approaches could increase the number of candidates suitable for vaccination trials.

Dr. Peter Rieber (Technical University of Dresden, Dresden,Germany) pointed out that an effective T cell-based immunotherapy depends on the availability of a large collection of proteins that are selectively expressed in tumors and which occur in a number of different tumors. He reported on the newly identified inhibitor of apoptosis protein survivin, which was found to be selectively expressed in a great variety of tumor tissues. In a collaboration with Dr. Hans-Georg Rammensee and Dr. Stefan Stevanovic (University of Tuebingen), two HLA-A*0201-restricted survivin peptides were predicted by the use of a peptide-motif scoring system. Both peptides were shown to induce CD8⁺effector T cells when presented on DCs; one proved to result from the natural intracellular processing of survivin. These findings recommend survivin as a new and widely applicable target for protein- and peptide-based immunotherapy of tumors.

Clinical trials based on tumor-associated peptides or proteins were reported by Dr. Pierre van der Bruggen (Ludwig Institute for Cancer Research, Brussels, Belgium). He pointed out that the MAGE gene family is of particular interest for vaccination protocols because it is expressed in a variety of tumor tissues, but it is silent in most normal tissues except testicular germ cells. Among the MAGE genes, MAGE-3 is one of the most frequently expressed antigens in tumors and a number of MAGE-3-derived peptides bound to MHC class I and class II molecules have been identified. In a recent clinical trial, patients with metastatic melanoma received several s.c. injections of a HLA-A1-restricted MAGE-3 peptide. Of 25 patients, 3 complete remissions and 4 partial remissions were observed. However, no increase of peptide-specific CTLs could be detected in the blood of these patients. More recently, in another clinical trial, 36 cancer patients with advanced disease received several injections with MAGE-3 protein. Two partial responses (melanoma and bladder cancer), 3 mixed responses (melanoma) and 1 stable disease(melanoma) were observed for a duration of 4–12 months.

In recent years, DCs became the pivotal players in the activation of T cells against TAAs. Dr. Cornelis J. Melief(Leiden University Medical Center, Leiden, the Netherlands) discussed the role of DCs and CD4+ T helper cells in generating tumor-specific immune responses. In mouse models, the outcome of antigen recognition by naive CD8+ CTLs is orchestrated by CD4+ T cells, leading either to priming or to tolerization. The presence of CD4+ T cells favors the induction of CTL immunity, supposedly through activation of APCs by CD40L expressed on activated CD4+ T cells. CD40-mediated activation results in the up-regulation of costimulatory molecules such as CD80 and CD86, pivotal molecules for the induction of a CTL response. In the absence of CD4+ T cells, however, APCs fail to be activated, and CTL tolerance ensues. In certain mouse models, MHC class I-binding peptides were found to cause tolerance rather than immunity, resulting in enhanced tumor outgrowth. The tolerance induction was explained by the rapid dissemination of some peptides through the mouse body, where the peptides were loaded onto nonactivated APCs, which then tolerized the CTLs. Depending on the activation state of DCs and/or on the way of antigen presentation, the outcome is either immunity or tolerance. Strategies to overcome the state of tolerance were discussed. One possibility is to elongate the peptide with additional flanking sequences. Such peptides can cause excellent protection in most animals. It was presumed that an additional processing step is required for these elongated peptides, which makes sure that only a low amount of the peptide gets into the periphery, which is not sufficient for tolerization. Another approach to obviate tolerization by injected peptides is to inject anti-CD40 antibodies that cause a systemic activation of APCs, including those in the distal lymph nodes. Thus,the combination of a tolerogenic peptide vaccine with an activating antibody against CD40 is able to convert tolerance into strong CTL-mediated immunity. In another in vivo model, Melief’s group has shown that the vaccination of a known peptide together with incomplete Freund’s adjuvant is an excellent preventive vaccine, but,for obvious reasons, it cannot be used for therapy. Once animals have established tumors, this peptide vaccination led only to a delay of tumor growth, and all animals died. Yet, the combination of the peptide vaccine with anti-CD40 antibody could cure 70% of the animals. Thus,the application of anti-CD40 antibodies could convert a preventive tumor-specific peptide vaccine into a therapeutic vaccine. These observations are confirmed by studies, where the immature murine DC line D1 was activated with the anti-CD40 antibody, resulting in a marked increase of MHC-class I and II as well as costimulatory molecules on the cell surface. Priming of CTL immunity in CD4-depleted mice was only observed if these mice were immunized with activated D1 cells. These results support the clinical use of CD40-stimulating agents as components of anti-cancer vaccines.

Dr. Dirk Schadendorf (German Cancer Research Institute,Heidelberg, Germany) presented an update of a clinical trial including metastatic melanoma patients. Monocyte-derived DCs were pulsed with a combination of tumor peptides known to be recognized by CTLs or with tumor lysates. These DC preparations were repeatedly injected into lymph nodes or s.c. in combination with keyhole limpet hemocyanin. Vaccination was well tolerated, and objective clinical responses were evident in 8 of 32 evaluated patients (3 complete responses and 5 partial responses).

Dr. Arie Belldegrun (University of California School of Medicine, Los Angeles, CA) pointed to the great progress in immunotherapy of metastatic renal cell carcinoma in the years from 1985 to 1995, which was, however, followed by stagnation in the last 5 years. New approaches focus on the treatment of residual disease after resection of the tumor. He summarized the results of several clinical trials. When patients with metastatic renal cancer were treated with surgery and IL-2 through the University of California at Los Angeles Medical Center Kidney Cancer Program, a significant improvement of the average survival was observed. The addition of IL-2 treatment expanded the average survival time from 6.5 months to 18.4 months. In a recently finished Phase I study, DCs were loaded with autologous tumor lysate and intradermally injected into 12 patients with metastatic kidney cancer. No toxicity was observed, and one partial clinical response was registered. Combination of tumor lysate-loaded DCs with IL-2 treatment is planned in a forthcoming clinical study. For a more specific approach, a fusion protein consisting of granulocyte macrophage-colony stimulating factor and G250, a recently described renal cell carcinoma-associated protein,was created. DCs were differentiated from monocytes in the presence of this fusion protein and IL-4 and then cocultured with T lymphocytes. G250-specific, HLA-restricted CTLs were induced with the capacity to efficiently lyse G250-expressing tumor targets.

In a special lecture, Dr. Mark M. Davis (Stanford University School of Medicine, Stanford, CA) gave a comprehensive overview on the tetramer technology for the identification of antigen-specific T cells. Although affinity of TCR for MHC/peptide complexes is only in the micromolar range, T cell recognition is highly specific. The fast off-rate of TCR-binding allows for multiple interactions of a few MHC/peptide complexes on an APC with receptors on a T cell. This efficient scanning is made possible by the formation of“immunological synapses” during the contact between T cells and an APC. Videoimaging of fluorochrome-labeled membrane molecules revealed that MHC/peptide complexes are gathered into a core-like structure,whereas other interaction molecules, such as intercellular adhesion molecules, redistribute to form the circumference of the synapse. Increasing the local concentration of signaling molecules such as CD3 and CD4, etc., in an immunological synapse may well be the pivotal mechanism of costimulation. The low affinity of TCR-specific binding of soluble MHC/peptide complexes can be compensated for by multimerization of the ligand, which is achieved by streptavidin-mediated tetramer formation. First shown for CD8+ T cells recognizing viral peptides, this technique also allows recognition of T cells specific for tumor peptides. About 30–40% of patients with metastatic melanoma were found to have T cells specific for tyrosinase or MART1 in their peripheral blood. Quite surprisingly,however, functional analysis of tyrosinase-specific, tetramer-binding T cells in a melanoma patient revealed that these T cells were unable to kill their tumor target cells; whereas virus-specific T cells from the same patient were capable to kill, and other T cell functions in the patient were found to be quite normal. This tyrosinase-specific T-cell anergy was so profound that even stimulation with ionomycin and phorbol myristyl acetate was unable to activate these T cells. On the other hand, limiting dilution analysis and repeated antigenic stimulation could derive antigen-specific T cells, which could kill tumor target cells. This finding shows that monitoring T cell responses by extensive in vitro manipulation may be highly misleading with respect to the actual situation in the patient. Similar caution is also recommended for vaccination trials where the mere monitoring of tetramer-binding T cells without appropriate functional tests may result in erroneous interpretations.

When direct functional tests cannot be performed because of the low frequency of tetramer-positive cells, activation can be monitored by determining the activation marker CD69 on tetramer-positive cells after a short stimulation of 6 h with the respective peptide. A general outcome of these experiments is that, in contrast with virus-specific T cells, tumor-recognizing T cells cannot be activated easily. This can be taken as a strong indication of tumor antigen-specific suppression. Furthermore, in one patient,tumor-specific T cells were rapidly eliminated subsequent to the“Dartmouth” chemotherapy protocol (dacarbazine 220 mg/m² and cisplatin 25 mg/m² daily on days 1 to 3, carmustine 150 mg/m² on day 1, and tamoxifen 10 mg p.o. twice a day; treatment was repeated every 3 weeks), whereas EBV-specific T cells remained untouched. The reason for this particular sensitivity of tumor-specific T cells to cytostatic drugs remains elusive.

In vaccination trials, a clear-cut correlation between the appearance of CD8+ tetramer-positive T cells in response to vaccination and the clinical course is still missing. More encouraging results were obtained in CML patients. Those patients who showed a clinical response to treatment with IFN-α or allogeneic bone marrow transplantation gave rise to tetramer-positive cells recognizing PR1,the dominant HLA-A2-restricted epitope of proteinase 3 (in collaboration with Dr. Jeffrey Molldrem, University of Texas,Houston, TX).

Tetramer technology combined with videoimaging may also be used for the analysis of TCR binding characteristics on individual cells. Overall TCR affinity was found to increase and to be narrowed in the course of an immune response, and it seems to correlate with biological efficacy. For example, low tetramer-staining cells were unable to lyse tumor cells and could only kill peptide-loaded cells, whereas high tetramer-staining cells were found to kill both peptide-loaded cells and tumor cells.

Dr. Gert Riethmueller (University of Munich, Munich,Germany) pointed out that drugs used in chemotherapy are selected for their antiproliferative effect. Adjuvant immunotherapy should,therefore, concentrate on surviving, nonproliferating cells. Out of the various strategies available, the use of antibodies or antibody constructs is quite common. Riethmueller and collaborators aimed at recruiting additional effector mechanisms in addition to the known Fc-receptor-dependent antibody-dependent cellular cytotoxicity or complement activities. With bispecific antibody constructs(“heterominibodies”) reacting both with CD3 and CD19, CD8+ T cells were redirected to CD19-positive lymphoma cells in vitro and were activated to cytotoxicity during a 4-h culture even at low antibody concentrations. After a 20-h culture, CD4+ cells as well became cytotoxic against lymphoma cells. This antibody construct required experienced memory T cells as effectors. By providing a second stimulatory signal, naive cells, too, could be activated to become cytotoxic effectors. This was achieved by coating lymphoma target cells with an additional antibody construct displaying the B7.1 molecule.

A large prospective trial documented a clear-cut correlation between the presence of micrometastases in the bone marrow and the survival rate of patients with breast carcinoma. Of 199 patients with occult metastatic cells, 49 died of cancer-related causes (25%), whereas, of 353 patients without bone marrow micrometastases, only 22 died of breast cancer (6%). This can be taken as a good reason to focus on these residual cells as preferential targets in adjuvant cancer immunotherapy.

Antibody therapy of breast cancer patients was discussed by Dr. Wolfgang Eiermann (University of Munich, Munich, Germany). HER2-neu overexpression is found in up to 30% of women with breast cancer. This overexpression is associated with negative prognostic factors such as mutated p53, positive lymph nodes, high fraction of S phases among the tumor cells, and lack of estrogen receptor. In clinical trials, the humanized HER2-antibody trastuzumab (Herceptin) at a loading dose of 4 mg/kg with a subsequent weekly maintenance dose of 2 mg/kg showed a response-rate of 15–20% as single agent in progressed high-risk patients, and even some complete remissions were observed. A multicenter, controlled trial in 469 female patients with HER2-neu overexpressing metastatic breast cancer randomly added Herceptin to six cycles of either 60 mg/m²doxorubicin and 600 mg/m² cyclophosphamide(administered every 3 weeks) or 175 mg/m²paclitaxel (administered every 3 weeks), depending on prior chemotherapy. After a follow-up of 25 months, the median overall survival of the patients with the combination therapy was 25.4 months compared with 20.9 months in the chemotherapy alone group(P = 0.045). The main problem induced by the combination therapy was a high rate of grade III/IV cardiac dysfunction in 19% in the doxorubicin/cyclophosphamide group and 4% in the paclitaxel group. The cardiac toxicity may be explained by the expression of HER-2 and HER-4 receptors in the myocardium. The antibody will now be evaluated in the adjuvant setting where immunotherapy combined with conventional chemotherapy seems to be most promising.

Dr. Gerhard Ehninger (University Hopital, Dresden,Germany) reported on the humanized anti-CD33 antibody, which was linked to the antitumor antibiotic calichaemicin and used for treatment of AML. Phase II studies of the anti-CD33-calicheamicin conjugate CMA-676 at 9 mg/m² in AML patients in first relapse showed a response rate of 31%. He suggested an evaluation of the efficacy of CMA-676 in induction therapy and postremission treatment of AML patients who present with cytogenetic aberrations associated with poor prognosis or with overexpression of the multidrug resistance gene.

Dr. Thomas Valerius (University of Erlangen-Nurnberg,Erlangen, Germany) focused his discussion on different ways in which antibodies can affect tumor cells. The chimeric anti-CD20 antibody Rituximab, which has been shown to be effective in relapsed follicular lymphoma, has a direct proapoptotic effect on lymphoma cells. Recently published animal data suggest a principal role of antibody-dependent cellular cytotoxicity. In these experiments, rituximab and trastuzumab did not prevent the growth of inoculated tumor cells in Fc receptor γchain knock-out mice, whereas wild-type mice could be rescued by antibody treatment. In addition, the efficacy of anti-CD20 treatment could be increased by knocking out the inhibitory Fc receptor. New results indicate that the survival of wild-type mice inoculated with CD20 positive lymphoma cells can be increased by 60% with the use of a bispecific anti-CD20/anti-Fc α-R (CD89) antibody.

The concept of allogeneic immunotherapy after nonmyeloablative conditioning was introduced by Dr. Richard Childs (NIH,Bethesda, MD). Clinical data were presented on the initial experience with allogeneic cell therapy in 19 patients with metastatic renal cell carcinoma who failed to respond to conventional treatment. They received 120 mg/kg Cyclophosphamide and 125 mg/m²Fludarabine before infusion of granulocyte-colony stimulating factor-mobilized peripheral blood stem cells from matched sibling donors. When cyclosporine A was given to prevent GvHD,the incidence of acute GvHD was 53%, and 3 of 19 patients experienced grade III/IV acute GvHD. No graft-rejection was observed, and complete donor chimerism occurred both in the T-cell and in the myeloid compartment. The overall response rate was 53%, including three complete and seven partial responses. Two patients died from treatment-related toxicity and seven patients from progressive disease. An interesting observation was that it took an average of 150 days until objective tumor regression was seen in responding patients. Patients with acute GvHD had a significantly higher response rate compared with patients without GvHD (P = 0.005), suggesting that a functional donor immune system is a prerequisite for graft-versus-tumor effects. In support of the specificity of the observed antitumor effect are in vitro data showing that, in some patients, activated T cells that express antigen receptors of one Vβ family specifically lysed the original tumor cells, and that this cytotoxicity could be inhibited by anti-HLA class I antibodies.

The concept of out-patient allogeneic blood stem cell transplantation using a conditioning regimen of very low intensity was presented by Dr. Brenda M. Sandmaier (Fred Hutchinson Cancer Research Center, Seattle, WA). An immunosuppressive conditioning regimen consisting of total-body irradiation with 2 Gy, cyclosporine A,and mycophenolate mofetil was followed by infusion of peripheral blood stem cells from matched sibling donors. The report included 56 patients treated in Seattle, Stanford, and Leipzig who were not eligible for standard allo-transplant protocols. The rate of acute GvHD was 56%,and the actuarial overall survival was 73%; 46% of the patients remained in remission. In addition, molecular remissions were reported in 4 of 10 patients with chronic lymphocytic leukemia and in 5 of 6 patients with CML who had stable engraftment. Remissions were also seen in patients with myelodysplastic syndrome, refractory AML, and Hodgkins’s disease. When high-dose chemotherapy with autologous stem cell transplantation was followed immediately by a mini-allograft, 5 of 6 patients with multiple myeloma achieved partial or complete remission.

Dr. Tuna Mutis (Leiden University Medical Center, Leiden,the Netherlands) summarized the Leiden experience in using mHags to induce leukemia-specific CTL responses in vitro. Having reviewed the polymorphic peptides described as mHags thus far (HA-1, HA-2, HY), he presented data on the expression of mHags HA-1 and HA-2 in hematopoietic tissues. HA-1/2-specific CTLs can be generated either with the help of peptide-pulsed DCs or by DCs transduced with a 312-kb HA-1 gene segment included in a retroviral vector. Thus far, six patients with mHag-negative donors have been enrolled in a clinical protocol where mHag-specific CTLs are infused in patients with hematological or cytogenetic relapses of AML,ALL, and CML, respectively.

Another promising application of antigen-specific allogeneic CTLs was presented by Dr. Helen E. Heslop (Baylor College of Medicine, Houston, TX). The incidence of EBV-lymphoma after T cell-depleted allogeneic bone marrow transplantation could be reduced from 11% to 0% by the prophylactic infusion of EBV-specific CTLs activated ex vivo against EBV-transfected donor B cells. The infused CTLs were labeled with the neomycin resistance gene and could thus be detected in resected tissue specimens from patients who were treated for overt lymphoma. Additional data concerned the use of EBV-specific CTLs in patients with Hodgkin’s disease. Although the follow-up is still limited, some clinical responses were observed.

In conclusion, this Symposium was highly successful in focusing in a comprehensive way on recent achievements and future perspectives in basic research and in clinical testing covering the principal areas of current tumor immunotherapy. The increasing number of reports convincingly showing successful immunotherapy of tumor patients and the growing insight into the molecular and cellular mechanisms governing immune reactions against tumor cells warrant a continuous and dedicated experimental effort in this area. Vaccination strategies may be improved by unraveling the modulation of tumor growth, identifying new TAAs and by elucidating further the functional characteristics of DCs and lymphocyte subpopulations. Additional therapeutic benefit is expected from the development of antibody constructs that efficiently redirect immune effector cells to the tumor target in patients.