Leukaemic relapse after allogeneic haematopoietic stem cell transplantation and the use of the graft-versus-leukaemia effect

Abstract: Survival after allogeneic haematopoietic stem cell transplantation (HSCT) has improved over the last decades due to improvements in immunosuppressive regimens and supportive care. Today the major risk of treatment failure is recurrent leukaemia, a risk which has stayed relatively constant over the years. This study was undertaken to assess factors that may affect the risk of relapse after HSCT and to explore new potential treatment strategies. Among 306 transplanted patients, significant risk factors for relapse in multivariate analysis cornprised acute leukaemia, as compared to chronic myeloid leukaemia (CML) (P=0.003), total body irradiation (TBI) compared to busulfan treatment (P=0.011), gram-negative prophylaxis with ciprofloxacin during conditioning with cyclophosphamide (CY) (P=0.024), graft-versus-host disease (GVHD) prophylaxis using a combination of methotrexate (MTX) and cyclosporine (CSA), compared to monotherapy (P=0.037) and absence of chronic GVHD (P=0.050). Following this, ciprofloxacin is no longer given during CY conditioning. Chronic GVHD has a major impact on the leukaemia-free survival (LFS). Four hundred and fiftyone bone marrow transplant recipients were analysed for factors predictive of chronic GVHD. Significant factors in multivariate analysis were increasing recipient age (P<0.001), acute GVHD grades (I-IV) (P<0.001), immune female donor to male recipient (P=0.006) and CML, compared with all other diagnoses (P=0.014). In patients with HLA-identical sibling donors and GVHD prophylaxis with MTX and CSA, increasing recipient age (P<0.001) and CML (P=0.007) were significant factors. Furthermore, multivariate analysis of patients with unrelated donors showed recipient age (P=0.006) to be the only significant risk factor for chronic GVHD. To try to stimulate the graft-versus-leukaemia (GVL) effect we evaluated the use of a low-dose CSA regimen (starting at 1 mglkg/day i.v.) of short duration, in combination with 4 doses of MTX CSA was tapered off from 2 months after HSCT and discontinued at a median of 6.6 months. In comparison with retrospective controls the low-dose regimen resulted in a increased risk of mild acute GVHD (78% vs 57 %, P<0.01) and chronic GVHD (60% vs 24%, P<0.001). A high CSA dose was the strongest predictive factor for relapse in multivariate analysis (P=0.03). The regimen appears safe as the risk of moderate to severe acute GVHD, TRM and extensive chronic GVHD did not increase. This treatment may enhance the GVL effect, thus reducing the risk of relapse after HSCT. Donor lymphocyte infusions (DLI) have so far been the most effective way to treat recurrent leukaemia after HSCT. In an evaluation of 44 leukaemic patients receiving DLI at our centre, an initial response was seen in 15 out of 22 patients (68%) with CML. However, 4 patients among the responders (27%) relapsed within two years after DLI, resulting in a three-year current-LFS (cLFS) of 46%. None of the patients with haematological relapse achieved a remission lasting over a year after DLI, compared to a three-year cLFS of 85% for patients in cytogenetic (n=10) or molecular (n=3) relapse. For patients with non-CML diseases the results are less promising, especially for patients relapsing within one year after HSCT. The cLFS at 18 months was merely 13%, suggesting that early and aggressive treatment may be important for these patients. Transduction of T-cells with a suicide gene prior to DLI can potentially allow control of subsequent GVHD. In the optimisation of T-cell expansion for retroviral transduction, we found X-VIVO 15, among other serum-free media, to give the highest rate of serum-free expansion after 21 days in culture (a median of 79-fold expansion, range 20-117). Equal percentages of CD4+ and CD8+ cells were obtained and the cytokines released into the media showed a type 1 cytokine pattern. Transduction with the LN vector, carrying the neomycin resistance gene, and G418 selection resulted in a 14-fold increase in cell numbers. Human serum increased expansion rates for all media. CD3-CD56+ natural killer (NK) cells have been shown to mediate potent anti tumour effects with limited non-specific alloreactivity. Peripheral blood mononuclear cells cultured for 21 days in Cellgro stem cell growth medium with interleukin-2 (500 U/ml) and anti CD3 (OKT3) expanded 193-fold (median, range 21-277) and contained 55% (median, range 7-92) CD3-CD56+ cells. The expanded cell population lysed 26 to 45% of the K562 target cells in a 1:1 effector to target ratio, signifying substantial cytotoxic efficacy. These high-yield CD3-CD56+ cells have been termed cytokine-induced natural killer (CINK) cells and may be prepared and used in a DLI setting.