- Top of page
- Materials and methods
- Author contributions
- Conflict of interest
Graft-versus-leukaemia (GvL) and graft-versus-host disease (GvHD) are both caused by alloreactive lymphocytes. We previously reported that GvHD correlated with higher numbers of effector CD4 T cells and Natural Killer cells early after allogeneic transplantation using a regimen comprising fludarabine, busulphan and alemtuzumab. Here, we assessed immune cell subset recovery in these patients in the context of early myeloid malignant disease relapse. Despite the close relationship between the GvL and GvHD immune responses, rapid recovery of lymphocyte subsets was not associated with protection from disease relapse. These results indicated that GvL may be weak in this treatment setting for patients with myelodysplastic syndromes and acute myeloid leukaemia. Consistent with low GvL activity, we previously reported that mixed T cell chimaerism had no detrimental effect on relapse in this treatment setting and instead correlated with better outcome because of reduced GvHD incidence. We now report that patients with significantly higher lymphocyte numbers prior to transplantation subsequently maintained the mixed T cell chimaeric state. This pre-transplant profile, together with absence of the early post-transplant signature indicative of GvHD predisposition, could potentially be used to identify patients suitable for early withdrawal of immunosuppression and prophylactic donor leucocyte infusion to boost GvL activity.
Although allogeneic haematopoietic stem cell transplantation (HSCT) is a curative therapy for haematological malignancies, disease relapse and graft-versus-host disease (GvHD) are major causes of morbidity and mortality following treatment. Two aspects of the procedure contribute to the elimination of malignant cells. They are ablative therapy immediately prior to transplant and the post-transplant donor immune response to residual malignant cells known as the graft versus leukaemia (GvL) effect (Horowitz et al, 1990). Both the beneficial GvL and the detrimental immune response that causes GvHD result from donor lymphocyte recognition of allogeneic patient cells. Control of clinical GvHD is achieved by immunosuppression, which curbs all forms of alloreactivity. Therefore the desire to preserve GvL has to be balanced with the need to prevent GvHD (Barber & Madrigal, 2006).
Myeloid malignancies predominantly affect the elderly. Dose-attenuated HSCT protocols, known as reduced intensity conditioning (RIC) regimens, have been developed to enable treatment of these patients (Slavin et al, 1998). The protocols do not aim for complete ablation of the patient’s haematopoietic cells prior to transplant. Instead, donor cell engraftment is dependent on immunosuppression, which also serves to control GvHD. However, there is greater dependence on GvL for eradication of residual malignant cells (McSweeney et al, 2001) so judicious use of immunosuppression is essential.
RIC regimens frequently result in a prolonged period of mixed donor and patient chimaerism after transplantation. In some settings, slow attainment of full donor chimaerism within the T cell population is associated with disease relapse (Childs et al, 1999), (Baron et al, 2005), (Mohty et al, 2007) attributed to inadequate GvL activity. A prolonged state of mixed T cell chimaerism is therefore used to indicate a need for prophylactic donor leucocyte infusion (DLI) in order to boost GvL (Dey et al, 2003), (Peggs et al, 2004) despite the attendant risk of promoting GvHD. However, some studies have not found an association between rapid full donor T cell chimaerism and disease remission (Mattsson et al, 2001), (Montero et al, 2005), (Lim et al, 2007). This ambiguity complicates decisions about use of prophylactic DLI. Variable correlation of mixed T cell chimaerism with disease relapse probably reflects differences in RIC regimens and immunosuppression used and differing susceptibility of malignancy types to GvL activity. The study of homogeneous cohorts of patients is therefore required to understand the relative role of disease type, induction ablative therapy prior to transplantation and the post-transplant GvL effect on disease relapse.
We have recently reported analysis of the relationship between immune cell subset reconstitution and incidence of GvHD in a group of 25 patients presenting with acute myeloid leukaemia (AML) or myelodysplastic syndromes (MDS) and treated with a uniform allogeneic RIC regimen (Matthews et al, 2009). The protocol comprises the alkylating agent busulphan, lympho-depletion with fludarabine plus the anti-CD52 monoclonal antibody alemtuzumab (also known as Campath 1H) and post-transplant GvHD prophylaxis with ciclosporin for 2 months (Ho et al, 2004). Prolonged mixed T cell chimaerism is common after treatment but is not associated with increased relapse (Lim et al, 2007). GvHD break-through was found to correlate with increased numbers of donor-derived effector CD4 T cells and a relative deficit of regulatory CD4 T cells early after transplant (Matthews et al, 2009). We now report analysis of the recovery of immune cell subsets for these patients in the context of disease relapse.
- Top of page
- Materials and methods
- Author contributions
- Conflict of interest
In our previous study of these patients, we reported that higher numbers of effector CD4 T cells and NK cells in peripheral blood early after transplantation was associated with GvHD (Matthews et al, 2009). Despite the close relationship between GvHD and GvL alloresponses, our current study did not find a correlation between rapid recovery of any lymphocyte subsets and reduced incidence of malignant disease relapse. These results suggest that GvL immunity may be weak or ineffective in the context of patients that present with MDS or the related conditions AML or MDS/MPD and treated with a uniform regimen based on fludarabine and busulphan with alemtuzumab.
Limited GvL efficacy provides an explanation for the lack of correlation between T cell chimaerism status and disease relapse in this treatment regimen that we reported previously (Lim et al, 2007). If donor lymphocytes do not produce an effective GvL response, there would be no benefit associated with rapid progression to full donor chimaerism. Instead, this course becomes disadvantageous because donor lymphocytes cause GvHD. Hence, in this setting, mixed T cell chimaerism correlates with improved outcome (Lim et al, 2007).
By definition, mixed chimaerism after transplantation is dependent upon the contribution of immune cells from the patient. Consequently we found that patients with prolonged mixed T cell chimaerism had significantly higher numbers of lymphocytes immediately prior to conditioning and transplantation. Patient immune competence will be determined by previous treatment and disease type. It is therefore not a factor that can be manipulated to promote the favourable mixed T cell chimaeric state. However, knowledge of patient immune status prior to transplantation may enable prediction of those likely to follow a good clinical course and those at risk of complications.
Some studies in other treatment settings have found early vigorous lymphocyte recovery is associated with reduced incidence of myeloid malignant disease relapse (Powles et al, 1998), (Kumar et al, 2001), (Kim et al, 2004), (Savani et al, 2007). In contrast, we found slower and steady lymphocyte recovery was best for the treatment regimen studied. We noted that slower reconstitution was also exhibited by patients with stable mixed T cell chimaerism. Studies using mouse models show maintenance of mixed haematopoietic chimaerism can occur when mutual donor and recipient immune tolerance exists (Sykes et al, 1997). However, mixed chimaerism may not always reflect a state of tolerance. Slower lymphocyte recovery in patients with mixed T cell chimaerism could be due to low levels of host versus graft and graft versus host activity that mutually restrain expansion. Presence of alloreactivity is consistent with our previous finding that patients with mixed T cell chimaerism can still experience GvHD (Matthews et al, 2009). Others have also reported that mixed T cell chimaerism does not necessarily provide absolute protection from GvHD (Baron et al, 2004). These observations illustrate the importance of understanding how immune reconstitution influences clinical outcome for individual treatment regimens.
Alemtuzumab has a major impact on immune recovery after transplantation. The antibody causes extensive depletion of all lymphocyte subsets which reduces the incidence and severity of GvHD (Chakraverty et al, 2002). This facilitates transplantation of patients previously considered ineligible for treatment due to inability to cope with GvHD. However, the use of alemtuzumab is associated with diminished immunity to infections (reviewed (Chakrabarti et al, 2004)) and our findings suggest that GvL activity is also reduced. This could be due to both the profound lymphopenia early after transplantation and deficiencies within the recovering lymphocyte population. We found that naïve T cells were absent in all patients until at least 6 months after transplantation using alemtuzumab (Matthews et al, 2009). Consequently the alloreactive T cell repertoire early after transplantation is limited to cross-reactivity by memory T cells specific for previously encountered pathogens, which may lack leukaemia specificities. A greater proportion of NK cells express high levels of CD56, consistently comprising 25–35% of the NK cell population during the first year after transplantation compared to 2·9% in healthy volunteers (data not shown). These NK cells are reportedly abundant producers of cytokines but have weak cytolytic activity (Cooper et al, 2001), which is an important potential GvL effector mechanism. Regulatory T cells can suppress alloreactivity and an association between increased frequencies of regulatory CD4 T cells and higher incidence of disease relapse after HSCT for chronic myeloid leukaemia has been reported (Nadal et al, 2007). Regulatory CD4 T cells may therefore cause suppression of GvL. However, we did not detect a correlation between these cells and disease relapse in our patient cohort (data not shown) despite finding a relative deficiency of regulatory CD4 T cells in the context of GvHD (Matthews et al, 2009).
The spike of CD8 T cell numbers seen at day 60 in patients that subsequently relapsed may indicate there is an immune response to re-emerging malignant myeloid cells. However, any GvL activity evidently fails to control disease. It has recently been reported that T cell numbers are increased in newly diagnosed patients with AML, but their function is abnormal (Le Dieu et al, 2009). Leukaemia-specific T cells have been detected in patients after allogeneic HSCT for myeloid diseases (Molldrem et al, 2000), (Rezvani et al, 2003), (Melenhorst et al, 2009) but they can be susceptible to rapid replicative senescence (Beatty et al, 2009).
We found a correlation between higher numbers of myeloid dendritic cells in peripheral blood early after transplantation and no disease relapse. The association was probably not due to rapid engraftment of the myeloid lineage because there was no relationship between relapse and CD34 stem cell content of the graft (Table I) or monocyte recovery (Table II). Dendritic cells are heterogeneous antigen presenting cells that are essential for both induction and regulation of immune responses (Ueno et al, 2007). Blood dendritic cells of both myeloid origin and plasmacytoid dendritic cells were increased early after transplantation in patients that did not relapse, which suggests that rapid recovery of the antigen presenting cell population correlates with better outcome. Others have also reported association of higher numbers of dendritic cells early after transplantation with reduced incidence of disease relapse and better clinical course (Reddy et al, 2004), (Talarn et al, 2007). This may result from the promotion of GvL activity by dendritic cells. A mouse model showed this was most effective when the dendritic cells are of host origin (Reddy et al, 2005) which can directly present leukaemia antigens to T cells. However, we found that dendritic cells were predominantly of donor origin. Rapid replacement of dendritic cells in peripheral blood is common after HSCT (Auffermann-Gretzinger et al, 2002), particularly in treatment regimens that incorporate alemtuzumab (Klangsinsirikul et al, 2002), (Buggins et al, 2002). Although donor dendritic cells can cross present leukaemia antigens, murine studies indicate the GvL effect produced is weak and only effective at disease control when malignant cell burden is low (Reddy et al, 2005). An alternative explanation for our detection of an association between higher dendritic cell numbers early after transplantation and no relapse may be that these patients have a healthier immune milieu that is better able to sustain mature dendritic cells transferred with the graft in the period prior to recapitulation of dendritic cell myelopoiesis.
Our study showed no relationship between rapid recovery of donor lymphocytes and protection from disease relapse. The potential existence of qualitative differences in lymphocyte composition between patients that relapse and those that do not relapse needs to be examined. However, we have shown there are deficiencies within the recovering immune cell population, which suggests that patients may benefit from DLI to supplement the immune repertoire. This is particularly pertinent for patients with advanced disease that are at greater risk of relapse (Table I). Therapeutic use of DLI to treat relapsed MDS or AML is not very effective (Kolb, et al 1995), (Collins et al, 1997). This has led to prophylactic use of DLI (Schmid et al, 2006), which aims to boost GvL for disease control when leukaemia burden is low. Administration of prophylactic DLI in the treatment regimen we have studied was associated with reduced incidence of relapse, but at the expense of increased GvHD (Lim et al, 2007). Immune signatures indicative of lower inherent risk of GvHD could help guide clinical decisions. Absence of the early CD4 T cell signature indicative of GvHD predisposition (Matthews et al, 2009) combined with identification of patients likely to undergo slower and steady immune recovery based on immune competence before transplantation could perhaps be used to select patients for early discontinuation of immunosuppression and prophylactic DLI.