Non‐myeloablative stem cell transplants

Despite 30 years of steady progress in allogeneic stem cell transplantation (SCT), the complications following the procedure still largely limit its application to recipients less than 60 years of age with an HLA-identical donor. These hazards restrict allogeneic SCT to situations where the alternative to the transplant is death from progression of the malignancy or (in the case of non-malignant diseases) an unacceptable quality of life. Transplant-related mortality (TRM) is due in part to the side-effects of the myeloablative preparative regimen and in part to the consequence of transplanting a donor immune system, namely graft-versus-host disease (GVHD), and delayed recovery of cell-mediated immunity. Despite its limitations, allogeneic SCT remains a valuable treatment because it can exert a unique and potentially curative graft-versus-leukaemia (GVL) effect (reviewed in Barrett & Malkovska, 1996). In the last decade, there has been a growing realization that the GVL effect, mediated through an alloresponse of donor lymphocytes, makes a major contribution to the cure of haematological malignancies (Horowitz et al, 1990). Indeed, donor lymphocyte infusions (DLI) alone can induce permanent remissions in patients with leukaemia relapsing after allogeneic bone marrow transplants (BMT) (Kolb et al, 1995). To exploit the GVL potential of allogeneic SCT, while improving the safety of the procedure, three strategies have been developed: (i) use of reduced intensity preparative regimens, (ii) induction of tolerance between the host and the donor immune systems by creating a mixed chimaeric state after transplant and (iii) optimizing the GVL effect by increasing the speed and completeness of donor immune recovery. These approaches are collectively grouped under the term ‘non-myeloablative stem cell transplants’ (NSTs), or more colloquially ‘minitransplants’. In the last few years, the number of transplants prepared with low-intensity regimens has grown rapidly. Although the concept is relatively new, it is clear that NST have already revolutionized transplant thinking and promise to reshape the indications for allogeneic SCT (allo-SCT) by greatly extending the range of conditions that can be safely treated. Here, we discuss the rationale for NST, the engraftment kinetics, clinical results, current indications and future developments in this exciting new field. Clinical approaches to NST are shaped by two concepts. First, low-intensity preparative regimens can be expected to have low regimen-related mortality making allo-SCT safer and extending its use to older or debilitated individuals. Second, these preparative regimens coupled with post-transplant immune manipulation can be used to optimize immune recovery and create a controlled mixed chimaeric state, thereby inducing tolerance to prevent graft rejection and GVHD. The principles on which NST are based are summarized in Table I. Decreasing the intensity of the preparative regimen NSTs represent a change of emphasis, arising from the perception that the GVL effect may alone be sufficient to cure some haematological malignancies and that a reduction in preparative regimen intensity could improve the outcome in patients at high risk of dying from transplant-related causes (reviewed by Champlin et al, 1999). At the MD Anderson Cancer Center, Giralt et al (1997) exploited the immunosuppressive and chemotherapeutic properties of purine analogues [2-chlordeoxyadenine (2-CDA) or fludarabine] in conjunction with cytosine arabinoside or idarubicin as low-intensity preparative regimens for older or debilitated patients with advanced or refractory acute myeloid leukaemia (AML) or myelodysplastic syndrome (MDS). Subsequently, the same group reported the use of low doses of fludarabine and cyclophosphamide in a preparative regimen for older and debilitated patients with chronic lymphocytic leukaemia (CLL) and non-Hodgkin's lymphoma (Khouri et al, 1998). These regimens were considered non-myeloablative because they were of equal intensity to standard chemotherapy treatment not requiring stem cell rescue. Slavin et al (1998a) in Jerusalem pioneered low-dose busulphan (8 mg/kg), fludarabine and anti-thymocyte globulin (ATG) in somewhat younger individuals with malignant or non-malignant haematological diseases. Their intention was to create a transplant technique that would ensure engraftment, confer an anti-malignancy effect through GVL and have low overall toxicity. To boost engraftment, they gave DLI to patients who had persisting low levels of donor chimaerism after transplant. Although the use of busulphan renders the regimen more intensive, the protocol can be considered non-myeloablative because thalassaemia transplant recipients who have received similar doses of busulphan have had high rates of autologous recovery (Barrett et al, 1989). Optimizing donor immune recovery We designed a NST regimen to (i) establish full donor T-cell engraftment, (ii) optimize the graft-versus-malignancy effect by using short-term immunosuppression after transplant, (iii) use the graft-versus-marrow effect to achieve full donor marrow engraftment and (iv) reduce TRM and regimen-related morbidity so as to extend transplants to patients too elderly or debilitated to be considered from a standard transplant approach. We chose the lowest dose intensity preparative regimen that could be expected to achieve complete engraftment, relying on the powerful immunosuppressive but modest myelosuppressive properties of the combination of cyclophosphamide and fludarabine. Based on animal data and from our previous studies in T-depleted transplants, high stem cell doses were used to optimize the chance of engraftment, reduce TRM and minimize relapse (Mavroudis et al, 1996; Reisner & Martelli, 1999; Bahçeci et al, 2000). Patients received one or two donations of unmanipulated, granulocyte colony-stimulating factor (G-CSF)-mobilized peripheral blood stem cells (PBSCs) containing a median of 6·5 × 106 CD34+ cells/kg and about 4 × 108 CD3+ cells/kg. To monitor engraftment, blood samples were sorted for T cells (CD3+) and myeloid lineage cells (CD14+, CD15+) with antibody-coated magnetic beads and analysed for lineage-specific chimaerism using a minisatellite polymerase chain reaction (PCR) technique which distinguishes donor from recipient DNA (Childs et al, 1999a). We used the chimaerism results to guide post-transplant management, modifying the immune balance between donor and recipient by raising or lowering the cyclosporin immunosuppression and, if necessary, giving DLI to boost donor engraftment. The transplant strategy is illustrated in Fig 1. Non-myeloablative stem cell transplant protocol. PBSCT, peripheral blood stem cell transplant; CSA., cyclosporin; DLI, up to three donor lymphocyte infusions of increasing doses of 0·1, 1·0 and 5·0 × 107 CD3+ T cells given at 4-week intervals. Establishing mixed chimaerism The idea of creating tolerance by establishing a mixed chimaeric state arose from earlier attempts to render animals tolerant to organ transplants. In rodents, it was observed that under the right conditions a mixed chimaeric state could be established which induced tolerance between donor and recipient, preventing the development of GVHD even across large major histocompatibility complex (MHC)-mismatched barriers (Sykes et al, 1988; Sharibi & Sachs, 1989). Durable marrow engraftment could be achieved in these mixed T-cell and myeloid chimaeras. Remarkably, mixed chimaerism and tolerance could be established across large MHC-mismatched barriers (Sykes, 1996). In proposing a clinical trial using this approach, these investigators argued that the disadvantage of potentially losing the GVL effect through tolerance induction could be offset by additional DLI and the use of mismatched donors to provide a powerful alloresponse against the malignancy (Sykes et al, 1997, 1999). Based on animal experiments, they used cyclophosphamide, ATG and thymic irradiation as a preparative regimen for non-T-cell-depleted mismatched transplants in high-risk leukaemias (Sykes et al, 1999). In Seattle, Storb and colleagues based their NST approach on balancing host-versus-graft against graft-versus-host immunity, aiming for a mixed chimaeric state to serve as a safe platform for subsequent treatment with DLI. This required immunosuppression (but not myeloablation) of the host to accept the graft and post-transplant immunosuppression to prevent both GVHD and graft rejection. They used the dog as a transplant model to explore combinations of different doses of total body irradiation (TBI) followed by post-transplant immunosuppression to find the lowest irradiation dose that permitted engraftment. The transplants were performed between DLA (dog lymphocyte antigen)-identical littermates to serve as a model for HLA-matched transplants in humans. They showed that mixed chimaerism without GVHD could be established with TBI doses as low as 200 cGy, provided immunosuppression with cyclosporin and mycophenolate mofetil was given after transplant (Storb et al, 1997, 1999a; Yu et al, 1998). Their ultimate goal was to avoid all irradiation and chemotherapy in the preparative regimen by substituting biopharmaceutical agents to achieve immunosuppression. They have now successfully treated recipient dogs with concurrent infusions of donor mononuclear cells and CTLA4 immunoglobulin to block the second T-cell co-stimulation signal. This induced specific hyporesponsiveness to the donor, permitting full engraftment with TBI doses as low as 100 cGy (Storb et al, 1999b). Based on these observations, the Seattle group has introduced a low-intensity (200 cGy TBI) transplant regimen for older or debilitated patients with haematological malignancies (Storb, 1998). Table II summarizes recently described NST preparative regimens. Most rely on agents with well-known immunosuppressive properties, and many have been used at higher doses in conventional transplants for many years. They can be grouped into regimens combining purine analogues (usually fludarabine) and alkylating agents, regimens using low-dose TBI and those which combine cyclophosphamide with ATG, with or without thymic irradiation. Purine analogues were chosen because of their powerful and prolonged immunosuppressive action and low toxicity profile (Plunkett & Sanders, 1991; Gorski et al, 1993; Goodman et al, 1995). The degree of immunosuppression achieved is so great that patients treated with fludarabine are at risk from transfusion-induced GVHD (Maung et al, 1994; Briz et al, 1995). The dose of fludarabine used varies between 25 and 40 mg/m2, given over 5–6 d. There are no data linking these doses with different degrees of immunosuppression, but higher doses are associated with a greater risk of and and ATG are in the transplant alone is to achieve engraftment in patients with ATG is used as an to agents to engraftment and to GVHD because it donor T-cell by persisting in the after transplant. The alkylating agents and busulphan are used in myeloablative doses in standard transplants, their immunosuppressive at the doses used in NST are not of busulphan toxicity and is because busulphan is et al, 1996). is clear from this that the immunosuppressive intensity of NST in current use varies and is is about the effect of regimen intensity on immunosuppression and transplant outcome is summarized in Fig of immunosuppression, toxicity of regimens. Non-myeloablative with autologous low immunosuppression DLI required for complete donor engraftment low-dose cyclophosphamide in but immunosuppressive DLI required for complete donor engraftment cyclophosphamide, myeloablative with some autologous immunosuppressive DLI required to complete donor engraftment with autologous standard immunosuppressive DLI not required for complete donor engraftment in transplants TBI Table summarizes the and outcome of 106 patients using NST preparative regimens. The data are from three and from our transplant NST regimens can achieve a high of engraftment but rates to the regimen In our of one the transplant. This was a rejection without concurrent and with full autologous haematological recovery. high of complete engraftment has been reported by The combination of cyclophosphamide and ATG was in a high of engraftment in transplants from HLA-identical for chronic the NST approach can be expected to achieve engraftment with or without T-cell recovery NST regimens and post-transplant than conventional allogeneic SCT as they for recovery of both donor and recipient The speed of recovery is by the degree of induced by the preparative In our × at a median of after transplant. and over of the recipients required and are reported in and low-dose TBI regimens et al, 1999). toxicity toxicity from all NST regimens is is and disease is and in regimens. Patients the low-intensity regimens with some to be performed as an it is to be for side-effects with these new we observed a toxicity by and at the of recovery and requiring There are as no data on the of NST regimens. The of malignancies be than with standard transplants, and on development and are to be The low-dose TBI regimen used in Seattle is for after transplant. malignancies from low-dose TBI a The of donor and varies individuals and transplant The contribution of host to recovery can be expected to be less in patients given more stem cell agents as and lineage-specific chimaerism have been used to the of donor and myeloid fludarabine and cyclophosphamide we engraftment for myeloid cells or CD34+ and T cells (CD3+) (Childs et al, 1999a). of T cells were of donor less than of the myeloid cells were donor the T cells were donor in over the myeloid cells of mixed immune manipulation by of cyclosporin or by giving DLI, patients showed a to full donor chimaerism in all In chronic disease recipients given the same regimen but with additional ATG, the recovery of was after full donor myeloid recovery and requiring DLI to achieve donor T cells et al, 1999). of to full donor T-cell chimaerism The change from mixed to full donor T-cell chimaerism was associated with clinical In we have observed acute GVHD in the of full donor after full donor T-cell chimaerism is to donor, by acute GVHD. after the of full donor T-cell The less preparative regimen may have on both and from transplant-related In the post-transplant the and the of reduce the risk of and The of large of in the transplant may the more of cell-mediated immunity. the of after NST was than in standard transplants T-cell and et al, 1999). with these standard transplants, NST recipients had higher lymphocyte in the after transplant, of T and a T-cell by after transplant et al, 1999). GVHD GVHD is the after The of acute GVHD is similar to that after patients to GVHD because GVHD is after full donor T-cell chimaerism is GVHD has been as a syndrome of and and by but an of the risk of chronic GVHD is not because of the In all and in our GVHD is the of transplant-related death and GVHD mortality to be by recipient or GVHD was to patients over years in our et al reported one death from acute GVHD and two from chronic GVHD in a of Slavin et al (1998a) reported a mortality from acute GVHD in a of patients with a median age of years. Giralt et al (1997) no GVHD greater than II and no GVHD mortality in older GVL in myeloid malignancies its to cure by allogeneic lymphocytes, chronic myeloid leukaemia an model to the GVL potential of We and colleagues have reported and and remissions of after NST et al, Slavin et al, et al, et al, 1999). the of a with after an NST a delayed and GVL In the 30 after transplant, marrow and peripheral blood showed persisting with a to × At this of the myeloid cells were of donor Subsequently, the T-cell from a mixed to donor, with the development of II GVHD requiring treatment on there was a in the and a of × on followed by a recovery and a in donor myeloid the marrow donor and no cells the were the was for The remains in of on a low dose of cyclosporin to GVHD after transplant. Slavin et al (1998a) reported NST for patients with in chronic disease one from GVHD and one had a described NST in patients with in chronic and in et al, 1998). Patients received a low-intensity preparative regimen of cyclophosphamide and dose cytosine patients and a of years, there was an that NSTs confer a GVL effect at as powerful as that observed after standard transplants for There so no for a GVL effect of NST in Slavin et al (1998a) reported no in in but two patients from GVHD. Giralt et al (1997) reported regimen in or refractory There have been few NSTs for myelodysplastic syndrome but there is some for a GVL Giralt reported one complete and in a We have observed complete remissions donor myeloid in three patients for refractory with of refractory in to and chronic but is have been in Clinical following an HLA-identical NST in a with in chronic was the blood by full donor T-cell engraftment, subsequent complete and from by after transplant. GVL in malignancies et al reported NST in patients with chronic remissions were in three of patients with chronic lymphocytic leukaemia (CLL) and of patients with non-Hodgkin's lymphoma Subsequently, the same group reported NST using a regimen in patients with At a median of all patients have achieved a complete at the with no transplant-related mortality (Khouri et al, 1999). We observed prolonged remissions in of malignancies three These in and given low doses of cyclophosphamide and the of a GVL effect from NSTs have been used to In 1997, we a of NST for cell and These to to be to and can achieve rates of but no treatment The combination of and a potential to to these malignancies to the effect of and patients with cell of had treatment with received NST from HLA-matched or of patients from transplant have a with of and the two are now in complete with a or a (Childs et al, 2000). 4 donor engraftment and subsequent of cell in the treated with an The effect may in malignant and but too few transplants have been with too a to establish NST have a of in a with cell given an HLA-identical 30 after transplant, progression of after transplant, of all after transplant, of NST promise to have an in the treatment of non-malignant haematological The use of the graft to recipient marrow through a the to and haematological Slavin has the of donor in establishing donor and that NST followed by DLI to boost donor chimaerism was of establishing donor marrow in of that of the host in and et al, We have a between the of donor T-cell chimaerism and the of myeloid cells to those of the donor. donor marrow engraftment is by an GVHD which as a with a to low doses of a disease to the effect of NST because patients debilitated by chronic are at risk from transplant-related mortality from a standard transplant. in treatment and have the in making standard transplantation with its an used case of a NST for has recently been described et al, 1999). In an et al used NST to patients with The preparative regimen was cyclophosphamide, fludarabine and ATG followed by a transplant. was after transplant. patients after transplant, with full or myeloid has chronic GVHD the The low in this debilitated group the case for using a NST technique to provide sufficient for of haematological the of establishing bone marrow without development and and with risk of regimen-related malignancy makes the use of NST in with as thalassaemia and cell an exciting Transplant-related mortality (TRM) The causes of death in and in our are in Table TRM is in the of and disease we a TRM in patients over years, but no TRM in patients under years (Childs et al, The of death was with regimen-related The higher mortality reported by Giralt et al (1997) is to the and older age of their transplant the of NST the mortality from transplant is number of and the outcome after NST At there are data to which transplant approach the of the to be are regimen intensity already there are in the regimens that investigators as The of NST is not The regimens complete recipient haematological recovery d. The of a marrow may have as after an NST a of the cells from cells those of the This may the GVL if the are from a and In regimens containing stem cell agents as busulphan are to induce and full donor myeloid establishing a of GVHD The use of GVHD T-cell can reduce but may the risk of rejection and full donor T-cell the graft-versus-malignancy and graft-versus-marrow a graft-versus-malignancy effect is not it may be more to reduce the risk from GVHD by the transplant of T cells and the of post-transplant immunosuppression, aiming for mixed data on the clinical outcome of specific NST be required it be to specific approaches to The of of the malignancy for it is to rely on the delayed effect in the case of or in an malignancy to use a preparative regimen which some disease et al, 1999). alternative strategy for malignancies is to a of followed by the NST as a of et al, 1998). Clinical indications for NST of than 200 and a of treatment used and for transplant, it is to the of NST in specific disease NSTs improve in older or debilitated individuals is not data are too to the of NST from donors HLA-identical the of NSTs are in situations where standard transplants would not be no studies of transplants have or are to performed in the NST with a two or three donor using a non-myeloablative approach the of establishing a graft across an without GVHD. Although these are NST attempts are because they may provide a more powerful graft-versus-malignancy effect across a mismatched coupled with their reduced regimen-related toxicity in this and for patients with advanced malignancy and are to establish tolerance by creating mixed using followed by DLI to provide a graft-versus-malignancy effect or to a transplant in a but et al reported results with the mixed chimaerism patients with advanced were given cyclophosphamide, ATG and thymic followed by an one or two patients two showed disease but GVHD was a major The second approach, using a graft the risk from GVHD but the of the immunosuppression to achieve engraftment. investigators are induction with to and which block T-cell co-stimulation et al, or of donor that have been in by host (Mavroudis et al, 1998). clinical for NST In the few years, an in the of NST can be these transplants their promise of low immune recovery and powerful graft-versus-malignancy they to be not in the of older and debilitated patients but standard transplants for younger a has already been in a at our for patients with in chronic with an HLA-identical donor. standard transplants with NST be required to the of These studies at a of to the of second malignancies between NST and standard SCT et al, The for on and development would be of NST and standard transplants in non-malignant as thalassaemia and cell in which quality of than of mortality from a malignant are the in transplant is the potential of NST to minimize the and improve quality of in the while NST as a platform for of malignant disease the of a and powerful effect in cell the is to explore after NST in many of refractory the platform of a donor immune provided by the NST the of from the donor, without their rejection. In and a growing number of are described and to T cells are The graft-versus-malignancy effect of the NST could be by the of T cells from the donor, or on the malignant The of NSTs to a new immune into a recipient, without major new for refractory to standard There is a growing in the use of SCT to refractory patients with immune and the of patients have received immunosuppression followed by an autologous The idea of using an allogeneic SCT to from the state to the immune of the donor is because of the that the autologous transplant approach can the of T cells and for the 2000). the of NST, few investigators have used standard to because of the high that NST in the treatment of diseases. transplants data that NST could be used in conjunction with organ transplants to establish tolerance to the the for immunosuppression to prevent organ rejection. the of creating tolerance to the organ by establishing mixed chimaerism with bone marrow transplants. low of donor cells can create a state of stem cells may be of establishing et al, et al, 1996). The cell tolerance to be it be to achieve tolerance by an NST from the organ donor followed by the organ transplant. These clinical of NST some in clinical marrow First, it is no to that myeloablative is required to for donor myeloid cells to or to ensure leukaemia Second, in cell the that are for because and render the of stem cell We are on the of a new and that could the clinical of bone marrow stem cell transplants. more than is a to to new and developments and from clinical studies to these SCT