Brigitte Kircher, PhD, Laboratory for Tumour- and Immunobiology, Division of Haematology and Oncology, Department of Internal Medicine, Innsbruck University Hospital, Anichstrasse 35, 6020 Innsbruck, Austria. E-mail: email@example.com
Summary. Donor lymphocyte infusions (DLI) can induce a graft-versus-leukaemia (GvL) reaction in patients with relapsed disease. However, the mechanisms involved in remission induction are not completely known. A patient with chemotherapy-refractory relapse 1 year after human leucocyte antigen (HLA)-identical, unrelated stem cell transplantation (SCT) for bcr/abl-positive common acute lymphoblastic leukaemia (ALL) received a DLI from the original donor, and achieved complete cytogenetic and molecular remission concomitantly with extensive graft-versus-host disease (GvHD). Seven CD8+, donor-derived, alloreactive T-cell clones were generated by stimulating post-DLI remission cells with the patient's pretransplant mature dendritic cells. The minor histocompatibility antigen (mHag) recognized by these T-cell clones was identified as HA-1, a mHag associated with acute GvHD after SCT. Our finding provides evidence of HA-1-associated GvL effects after DLI that paralleled the eradication of full-blown, chemotherapy-refractory ALL relapse after allogeneic SCT.
Donor lymphocyte infusions (DLIs) are an attractive and potentially curative treatment option for patients relapsing after allogeneic stem cell transplantation (SCT) (Dazzi & Goldman, 1999). However, durable remissions after DLI for acute lymphoblastic leukaemia (ALL) are rare, with an estimated overall survival of only 13% at 3 years after transplant (Collins et al, 2000).
One of these mHag, termed HA-1, has been described as a GvL target after SCT (Goulmy, 1997). HA-1 is presented in HLA-A*0201 molecules in 69% of the HLA-A*0201-positive population (Van Els et al, 1992), and its expression is restricted to cells of haematopoietic origin, including leukaemic cells (De Bueger et al, 1992; Van der Harst et al, 1994). Therefore, feasibility studies with HA-1-specific T cells have been developed for cellular therapy of relapsed leukaemia (Mutis et al, 1999).
Patient and methods
Case report. A relapse of the underlying bcr/abl-positive common ALL was diagnosed in a 41-year-old male patient (HLA-A*0201, B*1501, *40012, Cw3, Cw4, DRB1*0404, *1302, DRB3*0301, DRB4*01, DQB1*0604, *0302) 1 year after uncomplicated human leucocyte antigen (HLA)-matched SCT with marrow from an unrelated male volunteer donor (HLA-A*0201, B*1501, *40012, Cw3, Cw4, DRB1*0404, *1302, DRB3*0301, DRB4*01, DQB1*0604, *0302). The relapse was refractory to reinduction with FLAG (fludarabine, cytosine arabinoside) chemotherapy (Montillo et al, 1997) and, in December 1999, the patient received a DLI of 3·5 × 108/kg unmobilized peripheral blood mononuclear cells (PBMCs) with a T-cell dose of 1·5 × 108/kg CD3+ cells from the original bone marrow donor. Thereafter, the patient was lost to follow-up for 2 months. He was readmitted to our hospital in February 2000 because of extensive graft-versus-host disease (GvHD), but in complete haematological and molecular remission. Immunosuppression with cyclosporin A (CsA) and steroids was initiated, resulting in partial remission from GvHD. Extracorporeal photopheresis was started and CsA and steroids were switched to mycophenolate mofetil (MMF, Cellcept®) with almost complete resolution of GvHD. The patient is now well more than 2 years after DLI, without further immunosuppression in sustained remission and with 100% donor chimaerism.
Generation of dendritic cells (DCs). Mature monocyte-derived DCs were generated by culturing the adherent fraction of pretransplant remission PBMCs for 6 d in AIM V medium (Life Technologies Ltd, Paisley, UK) supplemented with 10% human pool male serum (PS), 2 mol/l l-glutamine (Biochrom, Berlin, Germany), 100 U/ml penicillin (Biochemie, Vienna, Austria), 100 µg/ml streptomycin (Fatol Arzneimittel, Schiffweiler, Austria) in the presence of 1000 U/ml granulocyte–macrophage colony-stimulating factor (GM-CSF, Novartis Pharma, Vienna, Austria) and 500 U/ml interleukin 4 (IL-4, R & D Systems, Wiesbaden-Nordenstadt, Germany). This was followed by further incubation for 24 h with tumour necrosis factor-α (TNF-α, 50 ng/ml; Strathmann Biotech, Hannover, Germany). DCs displayed the characteristic morphology and high surface expression of CD83, major histocompatibility complex (MHC) class I and class II molecules, CD86, CD80 and CD40, but negativity for CD14, as determined by flow cytometry. The ALL-specific Philadelphia translocation was detected by fluorescence in situ hybridization in 20·5% of the DCs.
Generation of a polyclonal T-cell line (CTL) and T-cell clones. A CTL was established according to standard protocols (Goulmy, 1988; Eibl et al, 1996, 1997). Briefly, non-adherent cells isolated 3 months post DLI (during extensive, untreated GvHD) were incubated for 7 d with mature DCs at a responder:stimulator ratio of 25:1. Responder cells were restimulated weekly with irradiated (30 Gy) mature pretransplant DCs (3:1) and pretransplant remission PBMCs (1:10) in AIM V medium supplemented with 10% PS, glutamine, antibiotics and 1·5 mg/ml amphotericin (Grünenthal, Stolberg, Austria) and, after 3 weeks, with irradiated (50 Gy) pretransplant patient Epstein–Barr virus (EBV)-transformed B-lymphoid cell lines (EBV-LCL, 1:1). Lymphokult-T (50%; Biotest, Dreieich, Germany) was used as the source of IL-2 and refreshed three times a week by replacing half the medium. After a total culture period of 8 weeks, CD8+ cells were positively immunoselected using magnetic activated cell sorting (MACS) technology (from 12% to 76%; Miltenyi Biotec, Bergisch Gladbach, Germany) and cloned by limiting dilution to 0·3 cells/well in AIM V medium supplemented with 20 U/ml IL-2 (Proleukin; Fresenius AG, Bad Homburg, Germany) and 1% phytohaemagglutinin (PHA; Difco, Detroit, MI, USA). Seven stable growing, CD8+ (93–99%) cytotoxic T-cell clones were further expanded in AIM V medium supplemented as for CTL and were restimulated at a responder:stimulator ratio of 1:2 with irradiated EBV-LCL and at a ratio of 1:5 to 1:10 with a pool of seven HLA-mismatched, irradiated PBMCs.
Isolation and culturing of distinct cell types. EBV-LCL and PHA blasts were generated as previously described (Eibl et al, 1996). The antigen-processing mutant cell line T2 (Salter et al, 1985) was grown in Roswell Park Memorial Institute (RPMI) medium (PAA Laboratories, Linz, Austria) supplemented with 10% fetal calf serum (FCS, Life Technologies), glutamine and antibiotics and fed twice weekly.
Fibroblasts were allowed to grow out of 1 mm-explant sections of shaved skin biopsies. Confluent fibroblasts were trypsinized and subcultured in Dulbecco's modified Eagle's medium (DMEM) (Life Technologies) supplemented with 10% FCS, glutamine and antibiotics.
Chromium-release assay. Cytolytic activity was determined using a classic chromium-release assay as described (Zier et al, 1977).
Peptides. Potential HLA-A2 ligands from the ALL-specific bcr/abl fusion sequence were predicted using the database SYFPEITHI (Rammensee et al, 1999).
The peptides HA-1 (VLHDDLLEA; Den Haan et al, 1998) and HA-2 (YIGEVLVSV; Den Haan et al, 1995) and seven predicted ALL-specific p190 bcr/abl peptides nos. 397 (ALQRPVASD), 399 (GAFHGDAEA), 401 (QIWPNDGEG), 404 (GAFHGDAEAL), 405 (AFHGDAEAL), 409 (ALQRPVASDF) and 410 (DAEALQRPV) were synthesized in an automated peptide synthesizer 432 A (Applied Biosystems, Weiterstadt, Germany) following the Fmoc/tBu strategy. After removal from the resin by treating with trifluoroacetic acid/phenol/ethanedithiol/thioanisole/water (90/3·75/1·25/ 2·5/2·5 by volume) for 1 h or 3 h (arginine-containing peptides), peptides were precipitated from methyl tertiary-butyl ether, washed once with methyl tertiary-butyl ether and twice with diethyl ether and resuspended in water prior to lyophilization. Synthesis products were analysed by high-performance liquid chromatography (HPLC, Varian Star, Zinsser, Munich, Germany) and matrix-associated laser desorption/ionization-time of flight (MALDI-TOF) mass spectrometry (G2025A, Hewlett-Packard, Waldbronn, Germany).
For peptide recognition assays target cells were pulsed with peptides for 1 h at 37°C at concentrations ranging from 1 µg/ml to 1 ng/ml.
HA-1 polymerase chain reaction (PCR). HA-1H and HA-1R alleles of the patient and the donor were determined by HA-1 allele-specific PCR as previously described (Wilke et al, 1998).
Seven CD8+ cytotoxic, donor-derived, alloreactive T-cell clones were established from a polyclonal CTL, which was generated by stimulating remission cells after DLI with pretransplant patient DCs. The recognition of the T-cell clones was restricted to patient's pretransplant EBV-LCL and PHA blasts (Fig 1). Donor cells and cells from the patient during remission post DLI (which were 100% donor cells) were never lysed by the T-cell clones (Fig 1).
To further specify the definitive target antigen, donor EBV-LCL and T2 cells were pulsed with peptides for the two HLA-A2-restricted mHag, HA-1 and HA-2, and the seven ALL-specific bcr/abl peptides. All T-cell clones were highly specific for HA-1, as neither HA-2 nor any of the p190 bcr/abl peptides were recognized (Fig 2).
HA-1 (HA-1H) represents a nonapeptide with an allelic polymorphism differing from its HA-1-negative allelic counterpart in one single amino acid (HA-1R). Whereas patient pretransplant EBV-LCL expressed both alleles (HA-1H and HA-1R), donor EBV-LCL were negative for HA-1H (data not shown).
We determined a population frequency of 71% in HLA-A*0201-positive individuals and confirmed the haematopoietic-restricted tissue distribution by testing fibroblasts of HA-1-positive individuals which were not recognized by the T-cell clones (data not shown).
To analyse the GvL activity of these HA-1-specific T-cell clones we tested the recognition of freshly isolated, HA-1-positive myeloid and lymphoid leukaemic cells. As shown in Fig 3, HA-1-positive cALL, acute myeloid leukaemia (AML) and chronic myeloid leukaemia (CML) blasts were all lysed by the T-cell clones, whereas HA-1-negative cells were never recognized (data not shown).
We established several cytotoxic T-cell clones from an ALL patient at the time of active, untreated GvHD following DLI for leukaemic relapse after HLA fully matched unrelated SCT. These findings are of great importance for several reasons. As yet the exact mechanisms and target antigens responsible for remission induction and/or the development of GvHD after DLI are still unknown. Although mHag appear to be the favourite target structures for both reactions, this has definitely not been proven to date in vivo.
From our findings it must be strongly assumed that HA-1 is indeed the target antigen for both reactions. It seems very unlikely that other mechanisms mediated alone, for instance by GvHD-driven cytokines or other non-T cell-mediated mechanisms (e.g. natural killer activity, humoral immunity), would have the strength to eradicate chemorefractory relapsing leukaemia (Ferrara, 2000).
HA-1 is the only mHag which has been shown to be significantly associated with acute GvHD after SCT (Goulmy et al, 1996; Tseng et al, 1999). In this context it is of interest that our patient experienced typical signs of extensive, chronic and not acute GvHD after DLI. Although the presence of HA-1-specific T cells during chronic GvHD has already been reported (Van Els et al, 1990; Mutis et al, 1999), a significant association with chronic GvHD is so far lacking (Gallardo et al, 2001).
The discrepancy between the haematopoietic-restricted tissue distribution of HA-1 and the occurrence of GvHD in tissues of non-haematopoietic origin can be explained at least partly by the pathophysiological interactions in GvHD. The patients' HA-1-positive antigen-presenting cells (dendritic cells or Langerhans cells; Van Lochem et al, 1996) were lysed by donor-derived T cells. In a second step, T cells with other specificities were attracted by cross-priming or cytokine release and subsequently destroyed the GvHD target tissues.
All in vitro-generated HA-1-specific, CD8+ cytotoxic T-cell lines/clones published to date in the literature showed activity against leukaemic cells, although little information about their in vivo anti-leukaemic activity was provided (Van Els et al, 1990). In accordance with our findings, the increase in the ‘anti-leukaemic’ cytotoxic T-lymphocyte precursor frequency (Falkenburg et al, 1997) and in HA-1-specific CD8+ T cells after DLI (Marijt et al, 2001) supports the hypothesis that HA-1 is indeed a possible target of the GvL reaction. Whether in this context the inability to detect any ‘leukaemia-specific’ p190 bcr/abl-specific T-cell clones is due to the immunodominance of HA-1 remains to be determined (Rufer et al, 1998). The binding activity of the ALL-specific p190 bcr/abl peptide 399 is 64%, peptide 404 binds to 25% and peptide 409 to 20%. All the other peptides tested for recognition by the T-cell clones had a binding activity lower than 10% (as determined in comparison to the high-affinity peptide YLLPAIVHI in a competition assay, data not shown). It is currently being investigated whether this binding capacity is strong enough to induce a primary T-cell response.
In conclusion, our data are in line with previous in vitro reports and provide strong evidence that HA-1 is definitively an in vivo target antigen for GvL activity after DLI. Whether HA-1 genomic typing might be useful not only for identifying patients at risk for GvHD (Goulmy et al, 1996; Tseng et al, 1999), but also for identifying patients who are most likely to benefit from DLI, remains to be shown.
The authors thank A. Petzer for providing cellular material, Hans-Christoph Duba for performing the fluorescence in situ hybridization, Margot Haun and Klaus Eisendle for excellent technical assistance, and the nursing staff of the BMT unit for excellent care of the patient.