Allogeneic stem cell transplantation (Allo-SCT) is currently the only curative option for a wide range of hematologic malignant diseases and bone marrow failure syndromes. This type of cell therapy involving preparation by myeloablative (MA) chemotherapy conditioning regimens, leads to toxicity and mortality associated with the procedure, which limits its application to a group of patients with specific biological characteristics (age, co-morbidity) [1, 2]. This led to the development of transplantation techniques with reduced intensity conditioning (RIC) with a consequent reduction in terms of toxicity and mortality related to transplantation .
Chimerism provides an important prognostic predictor of relapse or minimal residual disease (MRD) in the post-transplantation period, because relapse remains one of the major problems of transplant and treatment failure in patients with high-risk hematological malignancies like acute myeloblastic leukemia (AML) [4, 5]. The monitoring of chimerism after RIC is standard and is used to assess for graft rejection, relapse, and/or the need for and the timing of donor lymphocyte infusions (DLI).
Quantitative analysis of chimerism by determination of the ratio of donor to recipient's cells is used to monitor the post-transplantation period and could potentially determine the relapse of the underlying disease as early possible [6, 7]. Different methods are available including restriction fragment length polymorphism [8, 9], interphase fluorescence in situ hybridization (FISH) in the case of sex mismatched Allo-SCT [10–13], polymerase chain reaction (PCR) for short-tandem-repeats (STR) or variable number tandem repeats [9, 14], and finally real-time quantitative PCR (RQ-PCR) for donor- and recipient-specific polymorphisms [15–17].
So far, there is no consensus on which technique is most appropriate. The RQ-PCR technique results in higher sensitivity (10–4-10–6) when compared to FISH or the STR-PCR methods , but screening for appropriate gene sequences allowing discrimination of donors and recipients as a precondition for the RQ-PCR is highly laborious. Although a variety of studies have been performed to investigate the potential of chimerism to predict the individual risk of relapse, many studies have investigated other hematologic entities [19, 20].
It has been shown that in some studies, the engraftment of the CD3+ T cell lymphocyte lineage can measure the degree of the graft and assess the risk of GVHD [10–12].
The introduction of RIC regimens is associated with increased rates of mixed chimerism (MC) . In some studies, MC was found to be associated with significantly higher relapse rates compared to complete donor chimerism .
MC was interpreted as a dynamic biological process, reflecting the competition of two co-existing cell populations. Therefore, the increase in the amount of recipient cells rather than their persistence seemed to predict recurrence of disease [23, 24]. There seems to be no doubt that the evaluation of chimerism in the early post-transplantation period (within the first 6 months) is helpful for the prediction of the relapse risk in patients with AML .
The aim of this study was to assess the impact of the acute GVHD on donor CD3+ T cell chimerism (TCC) in our patients at day 120 after RIC Allo-SCT.
Patients and Methods
We retrospectively analyzed our experience at the Institut Paoli-Calmettes Cancer Centre at Marseille, France, regarding the implementation of allograft chimerism. One-hundred-and-fifteen consecutive patients transplanted between 2001 and 2010 were identified. This group included 57 females and 58 males with a median age of 50 years (range: 26–68). Patients evaluated in this study were adult patients with acute leukemia (myeloblastic or lymphoblastic), chronic myeloid leukemia, or Hodgkin and non-Hodgkin lymphoma who received transplants from HLA-matched related donor (MRD) or matched unrelated donor (MUD) 10/10. One-hundred-and-thirteen patients (98%) received peripheral blood stem cells (PBSC) mobilized with granulocyte colony-stimulating factor (G-CSF), whereas two patients (2%) received bone marrow. All patients received RIC regimen, including fludarabine 90–180 mg/m2 (Fludara ®, Bayer Health, Puteaux, France), the total dose of busulfan (Bu) was 8 mg/kg orally or 3.2 mg/kg/day for 2 or 3 days i.v. (Myleran®, GlaxoSmithKline, Marly-le-Roi, France Busilvex Gold ®, Pierre Fabre, Boulogne-Billancourt, France) and anti-thymoglobulin (ATG) 2.5 or 5 mg/kg (Thymoglobulin®, Genzyme, Saint Germain-en-Lay, France). The baseline characteristics are shown in Table I. Patients affected by hematologic malignancies with the presence of fibrosis in bone marrow, or patients with an unrelated donor HLA mismatch (A, B, C, DQ, or DR), or transplanted from cord blood cells donor, or patients with non-available chimerism between day 30 and day 120 post-transplant were excluded from the study. All patients received a GVHD prophylaxis with cyclosporine-A (CSA) alone at 3 mg/kg; the addition of mycophenolate mofetil (MMF; Cellcept®, Roche, France) with cyclosporine was used in only nine patients (8%). The CSA doses were adjusted to achieve blood levels between 150 and 250 ng/mL and to prevent renal dysfunction. CSA was tapered starting on day 90 if no GVHD appeared. Main characteristics are shown in Table I.
Supportive care has been previously reported and was similar during the whole study period . Donor CD3+ TCC was serially assessed at 30, 60, and 90 days after Allo-SCT. Recipient peripheral blood T lymphocytes were positively sorted by a mix of anti-CD4 and anti-CD8 immuno-magnetic beads (Dynal, Compiègne, France). T-cell purity was controlled by flow cytometry and was always ≥95%. Genomic DNA was amplified using fluorescent PCR primers for polymorphic variable number tandem repeats (VNTR) or short tandem repeats (STR).
One-hundred-and-fifteen patients were identified. Chimerism was monitored in all cases between J30 and J120 post-transplant. The CD3+ TCC count between 5 and 94% is defined as mixed and between 95 and 100% as total. This method has a sensitivity of 1–5%. Clinical manifestations after transplantation were considered when evaluating GVHD to distinguish chronic and late acute GVHD (which includes persistent, recurrent, or late-onset acute GVHD), as shown in Table I.
Data are presented as medians, ranges, and proportions. The chi-square test or Mann–Whitney test was used to compare different variables in both populations with partial and total chimerism. When appropriate, the probability of acute GVHD was estimated using cumulative incidence analysis. The analysis of the time-dependent competing variables for acute and chronic GVHD was performed with a subdistributive hazards ratio model. Logistic regression was used to correlate transplant variables with chimerism in both univariate and multivariate analysis. The cumulative incidence of complete TCC was included with relapse or death as competing events, with the exception of cumulative incidence of overall mortality where no competing events are present. Landmark analysis was performed when TCC on day 120 was used to compare groups. p-value was considered significant when <0.05.
Correlation between chimerism and acute GVHD
In our 115 patients, 58 patients (50.4%) developed an acute GVHD and 57 patients (49.6%) did not. In the population who developed acute GVHD, 18.2, 17.4, 8.7, and 6.1% presented an acute GVHD grade 1, 2, 3, and 4, respectively. Only 37 patients developed acute GVHD of minimum grade 2 (32% of total population). The cumulative incidence of Grade 2-4 GVHD in our population was 25% (95% CI 17–34). The analysis of the 37 patients with acute GVHD grade ≥2 showed that they all have at day 120 post-Allo-SCT a total full donor TCC. On the other hand, all MC were documented in the 78 patients (68%) not presenting grade ≥2 acute GVHD on day 120 post-transplant.
GVHD preceded complete TCC detection in 73% of patients, by a median of 59 days (range: 4–103); GVHD followed complete TCC in the remaining 27%, by a median of 38 days (range: 4–42).
Chimerism and transplant outcome
We evaluated TCC values at days 30, 60, and 90 post-Allo-SCT as a box plot (Fig. 1).
Further, we correlated TCC at a fixed time point with distinct transplant outcomes: overall mortality (OM) according to TCC on day 120; GVHD according to TCC at day 30; and relapse according to TCC at day 30, 60, and 120. As shown in Fig. 2, OM was not different according to TCC at day 120 (p = 0.96); analysis of GVHD according to TCC at day 30 was not informative because only two patients obtained complete TCC at day 30. Relapse incidence according to TCC at day 30 and day 60 was not significant but, interestingly, we found a significant inferior relapse incidence among patients with complete TCC in comparison with patients with MC at day 120 (p = 0.01; Fig. 3).
Moreover, we compared cumulative incidence of complete TCC among patients developing acute GVHD vs. those who did not and as expected, patients with GVHD obtained faster and more complete TCC compared with the other group (p = 0.007; Fig. 4).
Factors associated with complete chimerism
Full donor TCC was achieved in 93 patients (81%) at a median of 77 days (range: 30–120) post-transplant. Of those, 77 patients (83%) received 2 days of Bu in the conditioning regimen, and 16 patients (17%) received 3 days of Bu. In the second group of 22 patients with MC, 21 patients (95%) received 2 days of Bu and one patient (5%) received 3 days of Bu in the conditioning regimen, respectively, (p = 0.27; Table II).
We calculated the kinetics of complete TCC among the two groups of Bu (2 or 3 days) and observed a comparable cumulative incidence (p = 0.09; Fig. 5).
Factors found to be significantly associated with complete chimerism in univariate analysis (Table II) were: donor origin (MUD vs. MRD, HR = 6.51 (1.10–51.17)), days of ATG (two vs. one, HR =2.97 (1.07–8.26)), and acute GVHD (yes vs. no, HR = 13.26 (1.71–102.93)).
Correlation between chimerism and acute GVHD was confirmed after multivariate analysis, HR = 15.72 (1.96–126.10; p = 0.01).
Interestingly patients who received ATG 2.5 mg/kg for 2 days obtained a higher probability of complete TCC compared with those receiving ATG 2.5 mg/kg for 1 day (p = 0.03; Table II).
Chimerism is well-established for surveillance of patients after Allo-SCT, and is considered as an important prognostic predictor in the post-transplantation period but interpretation of the results and techniques is not yet standardized. There seems to be no doubt that the evaluation of chimerism in the early post-transplantation period (within the first 6 months) is helpful for the prediction of the relapse risk in patients with AML [25, 27].
However, it remains controversial whether a finding of MC alone justifies initiating therapeutic interventions in the post-transplantation period, considering, for example, the risk of acute or chronic GVHD in case of DLI application or the hematological toxicity of demethylating agents . Thresholds for such intervention have not yet been defined.
The major causes of treatment failure are disease relapse, graft rejection and GVHD . One of the main goals of post-transplantation monitoring is to predict these negative events to set up the relevant preventive therapeutics. In this context, chimerism quantification has been proposed as an important method for monitoring post-Allo-SCT outcome. In fact, several studies have suggested that an accurate quantitative analysis of chimerism kinetics would permit early differentiation between the absence of engraftment and a delay in engraftment as well as early detection of patients with a high risk of GVHD or those liable to relapse [23, 24].
The simple demonstration of the existence of MC by qualitative methods is of little interest as regards the clinical consequences for the patient. Conversely, MC was proposed to be associated with a lower risk of acute GVHD, and was suggested to be not always predictive, but may be a consequence of reduced conditioning or of early graft rejection [8, 12, 27]. The kinetics of chimerism might allow a more accurate prediction of relapse and we must also consider the potential role of residual disease [13, 23]. However, many years ago our group showed that lymphoid chimerism at day 30 has a significant relationship with the development of acute GVHD and the day 100 chimerism is related to relapse of the disease [12, 26].
Monitoring of donor chimerism following Allo-SCT with myeloablative regimens (MA), is not considered useful because engraftment is thought to occur rapidly and consistently. The 2001 workshop of the American Society of Blood and Marrow Transplantation and the International Bone Marrow Transplant Registry recommended against the routine assessment of chimerism in this group of patients . Some recent studies have confirmed these findings and recommendations; however, a recent report found that 28% of patients showed MC at day 90 following MA Allo-SCT, and these patients had worse outcomes than those with complete chimerism. The authors concluded that the detection of MC following MA Allo-SCT should prompt clinical intervention .
There is no consensus on when to perform chimerism monitoring or which type of chimerism quantification (myeloid or lymphoid or both) should be used after RIC Allo-SCT. Keil et al.  have shown that the rapid establishment of lymphoid but not myeloid donor chimerism is a prognostic factor for long-term stable donor engraftment after non-myeloablative (NMA) Allo-SCT . This confirms previously published data by Childs et al., who demonstrated that rapid full donor chimerism was essential in patients undergoing NMA Allo-SCT .
Recently, Lange et al.  demonstrated that monitoring of WT1 expression in peripheral blood and myeloid chimerism (CD34+) in bone marrow may prove to be useful to predict relapse of acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS) after RIC Allo-SCT.
In this study, they did not monitor the lymphoid CD3+ chimerism . Unfortunately, in our study, we do not have data on myeloid chimerism since our institutional policy is to evaluate only lymphoid chimerism, with the exception of rare, specific cases.
Mickelson et al. recently compared the timing of donor lymphoid cell engraftment in patients undergoing RIC regimen with Fludarabine + TBI with those receiving MA conditioning with Bu- or TBI-based regimens . Achievement of >90% donor leukocyte chimerism occurred rapidly and consistently in all three groups and time to achievement of >90% donor T cells was similar among the three groups (p = 0.57). Achievement of >90% donor leukocyte chimerism was not associated with risk of acute or chronic GVHD, graft rejection, relapse or all-cause mortality in multivariate analyses. Donor TCC of >90% was significantly associated with development of extensive chronic GVHD.
Interestingly in our retrospective study, we found a significant association between acute GVHD and the presence of a complete TCC on day 120. The existence of this biological element (GVHD grades 2–4) means in the analyzed population, the presence of full donor TCC in 100% of cases. Given these considerations, one may question the interest and the benefit of TCC analysis in patients presenting an acute GVHD.
Regarding the role of the dose intensity of conditioning or T-cell depletion, in the present study, we observed a comparable cumulative incidence of complete TCC among the two groups of Bu (2 or 3 days; p = 0.09; Fig. 5).
Interestingly, we observed that the use of ATG 2.5 mg/kg for 2 days was associated with a higher probability of complete TCC compared to the use of ATG 2.5 mg/kg for 1 day (p = 0.03). The reason behind this observation is unclear. Furthermore, it appears to be difficult to compare our results with those from other studies, because the published data regarding the role of different dose intensity conditioning regimens on the achievement of full TCC were reported in NMA regimens using a low-dose TBI and not ATG .
In conclusion, our study demonstrates that acute GVHD was predictive of full donor TCC after RIC Allo-SCT. Therefore, one could question the frequent or close monitoring of donor chimerism in some patients with ongoing acute GVHD. However, chimerism testing could represent an attractive modality for MRD detection or impeding relapse warranting further prospective studies.
Jean El-Cheikh and Alberto Vazquez collected and analyzed data, performed statistical analysis, provided clinical care, wrote and revised the manuscript; Roberto Crocchiolo, Sabine Fürst, Luca Castagna, Catherine Faucher, Claire Oudin, and Angela Granata collected data, provided clinical care, and commented on the manuscript; Roberto Crocchiolo performed statistical analysis. Didier Blaise recruited patients, provided clinical care, and commented on the manuscript.
The authors thank the nursing staff for providing excellent care for our patients and the physicians of the Haematology Department at the Institut Paoli-Calmettes for their important study contributions and dedicated patient care.