CC BY-NC-ND
0Biology of Blood and Marrow Transplantation 13:1515-1524 (2007) © 2007 American Society for Blood and Marrow Transplantation l083-879l/07/l3l2-000l$32.00/0 doi:l0.l0l6/j.bbmt.2007.09.007
CD31 Cell Dose and Disease Status Are Important Factors Determining Clinical Outcomes in Patients Undergoing Unmanipulated Haploidentical Blood and Marrow Transplantation after Conditioning Including Antithymocyte Globulin
Lujia Dong,1'2 Tong Wu,1'2 Mei-Jie Zhang,3 Zhi-Yong Gao,1'2 Dao-Pei Lu1'2
1 Peking University, People's Hospital; 2 Dao-Pei Hospital, Beijing, China; and 3 Division of Biostatistics, Medical College of Wisconsin, Milwaukee, WI, USA
Correspondence and reprint requests: Lujia Dong, MD, Peking University, People's Hospital, No.11 Xizhimen South Street, Beijing 100044, China. Tel: 8610 68792784; Fax: 8610 68333439 (e-mail: lujialidong@yahoo.com.cn).
Received June 6, 2007; accepted September 9, 2007
ABSTRACT
Haploidentical transplantation is a feasible alternative for patients with life-threatening hematologic diseases who lack a matched donor. Factors affecting the clinical outcomes of haploidentical transplantation remain under investigation. We analyzed 157 consecutive patients with leukemia who underwent transplantation with nonmanipulated granulocyte colony-stimulating factor (G-CSF)-mobilized marrow and peripheral blood cells (G-BMPBs) from haploidentical donors after receiving myeloablative chemotherapy (Ara-C + BuCy + antithymocyte globulin). Follow up observations after transplantation were made from 48 days to 1191 days (median, 448 days). Multivariate analysis indicated that the cohort given higher doses ofCD3+ cells ($ 177 x 106/kg) in allograft transplantation had a significantly lower treatment-related mortality (TRM) (relative risk [RR] = 0.35; 95% CI = 0.16-0.77; P = .0090), better leukemia-free survival (LFS) (RR = 0.46; 95% CI = 0.26-0.84; P = .0106), and better overall survival (OS) (RR = 0.42; 95% CI = 0.23-0.78; P = .0058). Inversely, advanced-stage disease was a strong predictor of greater posttransplantation relapse (RR = 3.48; 95% CI = 1.26- 9.60; P = .0159), worse LFS (RR = 2.56; 95% CI = 1.33-4.95; P = .0050), and worse OS (RR = 2.77; 95% CI = 1.39-5.53; P = .0038). A high number of CD3+ cells (> 177 x 106/kg) given to patients resulted in statistically less TRM and more intensive graft versus leukemia effect without producing more severe grades of GVHD, all resulting in a significantly better overall clinical outcome from haploidentical transplantation. © 2007 American Society for Blood and Marrow Transplantation
KEY WORDS
Allogeneic transplantation • CD31 cell dose • HLA-mismatched/haploidentical • Leukemia
INTRODUCTION
Since 1956, when a graft-versus-tumor (GVT) effect was first observed in irradiated mice receiving allogeneic but not syngeneic marrow transplants [1], the concept of allogeneic hematopoietic cell transplantation (HCT) as a form of immunotherapy has been well accepted [2]. The dramatic effect of allogeneic he-matopoietic stem cell transplantation (HSCT) demonstrates the ability of donor lymphohematopoietic cells to successfully repopulate treated recipient marrow, making the eradication of certain malignancies possible. Maintaining a stable hematopoietic/immunologic
American Society for Blood and Marrow Transplantation
reconstitution in recipients of HCT and also inducing a considerable GVT effect requires relatively large numbers of donor lymphohematopoietic cells. Previous studies have shown that transplantation of a high dose of marrow cells is associated with reduced treatment-related mortality (TRM) and improved survival and results in improved short-term and long-term graft function [3,4]. The doses of CD31 [5], CD81 [6], and CD341 cells [7-9] may be critical for successful engraftment in a selective CD341 cell transplantation setting. To date, however, a clear association of those T cell subsets with the clinical outcomes
remains uncertain; some results are even controversial [6,7-10]. Currently, no data are available on the proper or optimal cell dose of nonmanipulated granulocyte colony-stimulating factor (G-CSF)-mobilized marrow and peripheral blood cells (G-BMPBs) in transplantation after a conditioning regimen that includes antithy-mocyte globulin (ATG). Such studies of graft constituents are drawing increasing attention, not only because these infused cells play a crucial role in hematopoietic/immune reconstitution after transplantation, but also because determination of cell dose is one of the few manageable (or controllable) pretransplantation factors [11,12]. Family HLA-mismatched/haploidentical transplantation offers a feasible therapeutic approach to patients with lethal malignant hematologic diseases without an HLA-matched donor, a situation that is becoming more common as family size shrinks. An earlier study from our center showed that patients with advanced-stage disease and no family or unrelated HLA-matched donor might achieve nearly comparable therapeutic effects from a family mismatched/haploidentical transplantation [13]. Since then, more patients, including those with early-stage disease with high-risk prognostic features, have been enrolled. The cell subpopulations in the donor grafts and the clinical pretransplantation factors of the patients in this study were all carefully evaluated.
METHODS
Data Collection and Patient Selection
A total of 157 consecutive patients who underwent haploidentical transplantation between January 25, 2002, and March 31, 2005, at Peking University Institute of Hematology and Dao-Pei Hospital were enrolled. Only patients with a diagnosis of acute mye-logenous leukemia (AML), acute lymphoblastic leukemia (ALL), chronic myelogenous leukemia (CML), or myelodysplastic syndrome (MDS) and no HLA-matched related or unrelated donors were eligible for the study. The following criteria for disease stage were used:
• Early stage: CML in chronic phase (CP); AML or ALL in first remission (CR1)
• Intermediate stage: CML in accelerated phase (AP); AML or ALL in second remission (CR2); MDS with severe refractory anemia and thrombocytopenia
• Advanced stage: advanced or resistant AML or ALL; MDS-refractory anemia with excess blasts or MDS-AML; CML in blast phase (BP).
The complete immune phenotype was determined for cell populations in each graft. The age range for patients was 6-50 years (median, 25 years), and that for donors was 13-66 years (median, 40 years) (Table 1). A total of 78 patients with early-
Table 1. Clinical data
Number of patients 157
Patient age, median (range), years 25 (6-50)
Donor age, median (range), years 40 (13-66) Donor-patient sex match, n (%)
Female to male 26 (17%)
Others 131 (83%) Donor-patient relationship, n (%)
Sibling 48 (31%)
Mother to child 65(41%)
Others 44 (28%) Disease type, n (%)
CML 52 (33%)
AML 41 (26%)
ALL 54 (35%)
MDS 10(6%) Disease state, n (%)
Early 78 (50%)
Intermediate 37 (23%)
Advanced 42 (27%) Number of HLA-antigen mismatches, n (%)
1 Ag 27(17%)
2 Ag 71 (45%)
3 Ag 59 (38%) Graft type, n (%)
BM 1 PB 156 (99.4%)
PB alone 1 (0.6%)
MNCs, median (range), x l08/kg 7.3 (4.2-16.3)
CD341 count, median (range), x l06/kg 2.3 (0.5-9.7)
CD31 count, median (range), xl06/kg 177(15-991)
CD41 count, median (range), x l06/kg 100 (13-527)
CD81 count, median (range), x l06/kg 74 (15-467)
Follow-up time among living patients, 448 (48-ll9l) median (range), days
stage disease were enrolled because most of them had unfavorable prognostic features. The protocol [13] was reviewed and approved by the Institutional Review Board (IRB) at Peking University Institute of Hematology. All patients or their guardians were required to sign consent forms approved by the IRB. Detailed patient and donor characteristics are given in Table 1.
Donor Source and HLA Disparity
Family members were tested for HLA compatibility using serology and intermediate-resolution DNA typing for HLA-A, -B, and -C antigens and highresolution DNA typing for HLA-DRB1, -DQB1, and -DPB1. Donors were patients' mothers in 65 cases (41.4%), fathers in 25 cases (16%), cousins in 4 cases (2.5%), siblings in 48 cases (30.6%), and children in 15 cases (9.5%), with various degrees of HLA disparity in HLA-A, -B, and -DRB1 antigen levels (Tables 1 and 2).
Graft Collection and Cell Composition Analysis
Donor BM and/or PB cells were collected using modified mobilization protocols; subcutaneous G-CSF (filgrastim) 5 mg/kg/day was given from day —3
Table 2. Patient and donor characteristics according to CD31 cell dose Variable CD31 < I77xI06/kg CD31 $ I77 x I06/kg P
Number of patients 78 79
Recipient age, median (range), years 24 (6-50) 25 (7-50) .95
< 20 26 (33%) 28 (35%) .98
20-35 30 (38%) 3I (39%)
35-50 2I (27%) I9 (24%)
$ 50 I (I%) I (I%)
Donor age, median (range), years 40 (I4-58) 40 (I3-66) .29
< 20 6 (8%) 8 (I0%) .5I
20-35 II (I4%) I4(I8%)
35-50 42 (54%) 45 (57%)
$ 50 I9 (24%) I2 (I5%)
Donor-patient sex matching .08
M to M 20 (26%) 25 (32%)
M to F 35 (45%) 2I (27%)
F to F 9 (I2%) I7(2I%)
F to M I4(I8%) I6 (20%)
Donor-patient relationship .04
Mother to child 38 (49%) 27 (34%)
Father to child 8 (I0%) I7 (22%)
Cousin 4 (5%) 0
Sibling 2I (27%) 27 (34%)
Child to parent 7 (9%) 8 (I0%)
Blood match .I5
Matched 43 (55%) 38 (48%)
Minor ABO incompatibility I8 (23%) I3 (I7%)
Major ABO incompatibility I7 (22%) 28 (35%)
HLA-antigen class mismatches .52
Class I (HLA-A and/or HLA-B) I5 (I9%) I6 (20%)
Class II (HLA-DRBI) 5 (7%) 9 (I2%)
Class I and II 58 (74%) 54 (68%)
Number of HLA-antigen mismatches .4I
I Ag 11 (I4%) I6 (20%)
2 Ag 39 (50%) 32 (4I%)
3 Ag 28 (36%) 3I (39%)
Disease .09
CML 30 (38%) 22 (28%)
AML 24 (3I%) I7(2I%)
ALL 20 (26%) 34 (43%)
MDS 4 (5%) 6 (8%)
Disease status .94
Early 39 (50%) 39 (49%)
Intermediate I9 (24%) I8 (23%)
Advanced 20 (26%) 22 (28%)
MNCs, median (range), I08/kg 6.8 (4.2-I2.I) 7.8 (4.2-I6.3) < .000I
< 7.5 55 (7I%) 3I (39%) < .000I
$ 7.5 23 (29%) 48 (6I%)
CD341, median (range), I06/kg I.9 (0.5-7.I) 2.3 (0.5-9.7) .I7
<2 4I (53%) 28 (35%) .03
$2 37 (47%) 5I (65%)
CD41, median (range), I06/kg 69 (I7-373) I3I (I3-527) < .000I
< I00 65 (83%) I3 (I6%) < .000I
$ I00 I3 (I7%) 66 (84%)
CD81, median (range), I06/kg 49 (I5-3I2) I03 (25-467) < .000I
< 75 6I (78%) I9 (24%) < .000I
$ 75 I7 (22%) 60 (76%)
Follow-up among survival patients, median (range), days n = 50 n = 63
4I4 (48-II50) 5I4 (II3-II9I) .03
to day 0. BM was harvested at day 0, after 4 days of G-CSF administration, with a target collection dose of 15 mL/kg (recipient weight), and PB, harvested on day 1, 5 days after the first G-CSF injection. The grafts were analyzed for the total number of mononuclear cells
(MNCs), and, using a standardized Multi-Set kit (Bec-ton-Dickinson, San Jose, CA), the content of CD341 cells [14] and subsets of lymphoid cell populations (CD31, CD41, and CD81 cells) was determined (Tables 1 and 2).
Transplantation Protocol
All patients enrolled in this study were treated under a uniform protocol with the conditioning regimen (Ara-C 1 Bu/Cy +ATG) [13] and GVHD prophylaxis (cyclosporine [CsA] 1 short-term methotrexate [MTX] [15,16] 1 mycophenolate mofetil [MMF]). Pretransplantation conditioning chemotherapy included cytarabine 4 g/m2/day intravenously on days — 10 to -9, Bu 4 mg/kg/d orally on days —8 to -6, Cy 1.8 g/m2/day intravenously on days —5 and —4, Me-CCNU 250 mg/m2 orally once on day —3, and ATG (thymoglobuline; Sang Stat, Lyon, France, now marketed by Genzyme) 2.5 mg/kg/day intravenously on days —5 to —2. First-line therapy for $ grade II GVHD was 1-2 mg/kg/day of prednisolone equivalents and resumption of full-dose cyclosporine (CSP). Patients who developed steroid-refractory acute GVHD (aGVHD) were given tacrolimus (FK506), MMF, CD25 monoclonal antibody (Simu-lect, Novartis), or MTX as second-line immunosup-pressive therapy.
Definition of Endpoints
The study endpoints were engraftment, grade II-IV or III-IV aGVHD, TRM, relapse, leukemia-free survival (LFS), and overall survival (OS).
Neutrophil Engraftment. Neutrophil engraft-ment after transplantation was defined as a recovery of the blood absolute neutrophil count (ANC) to $ 0.5 x 109/L on 3 consecutive days. Platelet recovery was defined as the time after transplantation needed to maintain a blood platelet count $ 20 x 109/L without transfusion support for 7 consecutive days.
GVHD. aGHVD and chronic GVHD (cGVHD) were defined according to Fred Hutchinson Cancer Research Center criteria [17-19]. Patients who survived for > 100 days after transplantation were evaluated for cGVHD.
Relapse. Time to relapse was defined from the date of transplantation to the date of disease recurrence, evaluated by morphological evidence of leukemia in either the BM or any extramedullary site.
TRM. TRM was defined as all causes of death other than those related directly to malignant disease itself, occurring at any time after transplantation.
OS and LFS. OS was defined as days from transplantation to death from any cause. LFS was defined as days from transplantation to disease progression after transplantation.
Statistical Analysis
The characteristics of patient-related, disease-related, and transplantation-related factors of the low and high CD31 cell dose groups were compared using the c2 test for categorical variables and the MannWhitney test for continuous variables. Univariate
Table 3. Multivariate analysis of association with TRM, relapse, LFS, and OS
Outcome RR (95% CI) P
CD31, median (range),I0'/kg
< 177 1.00
$ 177 0.35 (0.16-0.77) .0090
Relapse
CD31: < 177 1.00
$ 177 0.72 (0.29-1.79) .4778
Other covariates
Disease status: Early 1.00 .0263*
Intermediate 1.02(0.25-4.06) .9835
Advanced 3.48 (1.26-9.60) .0159
Treatment failure
(Opposite of LFS)
CD31: < 177 1.00
$ 177 0.46 (0.26-0.84) .0106
Other covariates
Disease status: Early 1.00 .0175*
Intermediate 1.40 (0.67-2.93) .3760
Advanced 2.56 (1.33-4.95) .0050
Overall mortality
(Opposite of OS)
CD31: < 177 1.00
$ 177 0.42 (0.23-0.78) .0058
Other covariates
Disease status: Early 1.00 .0136*
Intermediate 1.46 (0.67-3.18) .3398
Advanced 2.77 (1.39-5.53) .0038
Relapse: P = .0617 for intermediate versus advanced disease. Treatment failure: P = .1025 for intermediate versus advanced
disease.
Overall mortality: P = .0975 for intermediate versus advanced
disease. *2 degree-of-freedom test.
probabilities of achieving ANC and platelet engraftment, developing aGVHD and cGVHD, TRM, and relapse were calculated using cumulative incidence curves to accommodate corresponding competing risks. The Kaplan-Meier estimator was used to evaluate LFS and OS. Comparing the outcomes among CD31 cell dose groups required adjustment for differences in baseline patient characteristics (Table 2). A Cox proportional hazards model was used to adjust for potential imbalance in baseline characteristics between CD31 cell dose groups. Outcome events considered in multivariate analyses were TRM, relapse, treatment failure (opposite of LFS), and overall mortality (opposite of OS). A forward stepwise method was used to identify all significant risk factors for each outcome event. The effect of the dose of CD31 cells transplanted (the main interest in the present study) was included in all steps of model building. The proportionality assumptions were tested by adding a time-dependent covariate; the tests indicated that the proportionality assumptions held for all risk factors. The potential interactions between main treatment groups and significant risk factors were tested, and no interactions were found (Table 3) Adjusted
probabilities of LFS and OS were generated from the final Cox models stratified on treatment; weighted averages of covariate values were computed using the sample proportion as the weight function. These adjusted probabilities estimated the likelihood of outcomes in populations with similar prognostic factors. To examine the aGVHD effect, a time-dependent co-variate was used in the final Cox regression models (Table 4). Finally, the cell dose effects of MNCs, CD341, CD41, and CD81 were evaluated using the Cox regression method (Table 5).
RESULTS
Characteristics of Patients Receiving a Low or High CD31 Cell Dose
The median CD31 cell number in the transplant inoculums was 177 x 106/kg (range, 15-991 x106/ kg). Patients were stratified into 2 cohorts; 1 cohort was given a lower than median dose of CD31 cells (< 177 x 106/kg [n = 78]), and the other was given a greater than median or median dose of CD31 cells ($ 177 x 106/kg [n = 79]). The differences in clinical characteristics between the 2 cohorts were not statistically significant; the patients in the high-dose CD31 cell cohort were infused with significantly higher numbers of MNCs, CD341, CD41, and CD81 cells than those in the low-dose CD31 cell cohort (Table 2).
Neutrophil and Platelet Recovery
All patients in the study achieved complete granulocyte engraftment, as reflected by the ANC attained. The median time after transplant to achieve an ANC of $ 0.5 x 109/L was 12 days (range, 8-25 days); the median time to achieve a blood platelet count of $ 20 x 109/L was 15 days (range, 7-169 days) in all but 5 patients. Three of these 5 patients had died from nonrelapse-related mortality (NRM) at day 3684 posttransplantation. Two were alive; 1 survived to 399 days, the other to 206 days. The cumulative incidence of platelet engraftment was 97% (95% confidence interval [CI] = 94%-100%) in the low CD31
Table 4. Effect of aGVHD on clinical endpoints
Outcome RR (95% CI) P
aGVHD: III-IV versus 0-II Relapse
aGVHD: III-IV versus 0-II Treatment failure
(Opposite of LFS) aGVHD: III-IV versus 0-II Overall mortality (Opposite of OS) aGVHD: III-IV versus 0-II
3.16 (1.43-6.99) 1.34(0.38-4.73)
2.32(1.18-4.57)
2.64(1.29-5.43)
.0046 .6524
Table 5. Effect ofMNC, CD341, CD41, and CDS1 levels on clinical
endpoints
Outcome RR (95% CI) P
MNCs: $ 7.5 vs < 7.5 x l08/kg 0.70(0.33-1.49) .3586
CD341: $ 2 vs < 2 xl06/kg 0.67(0.33-1.40) .2874
CD41: $ 100 vs < l00xl06/kg 0.62(0.30-1.31) .2099
CD81: $ 75 vs < 75 x l06/kg 0.73 (0.35-1.53) .4090
Relapse
MNCs: $ 7.5 vs < 7.5 x l08/kg 2.52 (0.95-4.68) .0632
CD341: $ 2 vs < 2 xl06/kg 2.58 (0.85-7.80) .0941
CD41: $ l00 vs < l00xl06/kg 1.33 (0.53-3.33) .5384
CD81: $ 75 vs < 75 x l06/kg 1.40(0.56-3.51) .4694
Treatment failure
(Opposite of LFS)
MNCs: $ 7.5 vs < 7.5 x l08/kg 1.11 (0.62-1.96) .7316
CD341: $ 2 vs < 2 xl06/kg 1.04(0.58-1.86) .8956
CD41: $ l00 vs < l00 xl06/kg 0.83 (0.47-1.46) .5151
CD81: $ 75 vs < 75 x l06/kg 0.97(0.54-1.73) .9161
Overall mortality
(Opposite of OS)
MNCs: $ 7.5 vs < 7.5 x l08/kg 0.97(0.53-1.76) .9065
CD341: $ 2 vs < 2 xl06/kg 0.96 (0.52-1.75) .8816
CD41: $ l00 vs < l00 xl06/kg 0.80(0.44-1.45) .4537
CD81: $ 75 vs < 75 x l06/kg 1.05 (0.58-1.91) .8762
Effect of aGVHD is treated as a time-dependent covariate.
Effects are examined based on final regression models for TRM, relapse, treatment failure, and overall mortality (see Table 3).
cell dose cohort (< 177 x 106/kg) versus 95% (95% CI = 89%-98%) in the high CD31 cell dose cohort ($ 177 x 106/kg)(P = .3626).There was no significant difference in the cumulative incidence of ANC and platelet engraftment between the 2 CD31 cell dose cohorts (Figures 1A and B).
High CD31 cell dose ($ 177 x 106/kg) transplantation correlated with a higher incidence of grade II-IV aGVHD, 51% (95% CI = 40%-61%), compared with low CD31 cell dose (< 177 x 106/kg) transplantations, 39% (95% CI = 28%-49%), butthe difference was not statistically different (P = .1165) (Figure 1C). No significant difference was found in the cumulative incidence of grade III-IV aGVHD between the high CD31 cell cohort, 19% (95% CI = 11%-28%), and the low CD31 cell cohort, 15% (95% CI = 8%-24%), (P = 0.5449) (Figure 1D). In the 157 patients evaluated, the cumulative incidence of cGVHD at 2 years was 44% (95% CI = 32%-55%) in the low CD31 cell dose cohort and 52% (95% CI = 41%-63%), in the high CD31 cell dose cohort (P = .3100).
Relapse. Patients undergoing transplantations containing high CD31 cell doses ($ 177 x 106 CD31 cells/kg) had fewer relapses at 2 years, 12%
Figure 1. Cumulative incidence of outcomes of haploidentical HCT for patients with leukemia. A, ANC (P = 1.0000 at day 100). B, Platelet engraftment (P = .3626 at day 100). C, aGVHD, grade 2-4 (P = .1165 at day 100). D, aGVHD, grade 3-4 (P = .5449 at day 100). E, Relapse (P = .3962 at 2 years). F, TRM (P = .0181 at 2 years).
(95% CI = 6%-21%), than in the cohort receiving low CD31 cell doses (\ 177 x 106 CD31 cells/ kg), 18% (95% CI = 8%-31%), but the difference was not statistically significant (P = .3962) (Figure 1E). Thirteen patients died from disease relapse between 36 and 945 days after transplantation with various coexisting complications: severe infection, 43%; grade III-IV aGVHD, 35.7%; and interstitial
pneumonia (IPn), 50%; along with varying grades of aGVHD.
TRM. A total of 31 patients died from various complications between 72 and 590 days after transplantation: 27.3% from Ipn, 30.3% from lung infection due to mixed pathogens; 57.6% from nonpulmonary infection, and 24.2% from severe (grade 3-4) aGVHD. In patients whose grafts included high CD31 cell
doses ($ 177 x 106/kg), the 2-year TRM of 13% (95% CI = 6%-22%) was statistically lower than the 30% (95% CI = 19%-41%) in those whose grafts contained fewer CD31 cells (\ 177 x 106/kg) (P = .0181) (Figure 1F).
Survival
Follow-up of surviving patients from 48 to 1191 days posttransplantation (median, 448 days) revealed a 53% (95% CI = 39%-66%) probability of 2-year LFS for patients with grafts containing fewer CD3 1 cells (\ 177 x 106/kg) and a 75% (95% CI = 64%-84%) probability of 2-year LFS in those with grafts containing more CD31 cells ($ 177 x 106/kg) (P = .0115) (Figure 2A). OS after haploidentical transplantation in the cohort was 70.2%. Univariate analysis demonstrated that only disease status and CD31 cell dose in the graft were important predictors of survival. The 2-year probability of survival was 77% in those who underwent transplantation in the early stage of disease, 57% in those who did so in the intermediate stage, and 48% in those who did so in an advanced stage. There was no statistical difference in survival probability between those with intermediate-stage disease and advanced disease (log-rank test; P = .3246); however, a significant difference in survival probability was seen between patients who underwent transplantation in early-stage disease (77%; 95% CI = 66%-86%) and those who did so in intermediate-stage or advanced disease (56%; 95% CI = 43%-68%) (P = .0052). In addition, the probability of 2-year survival in the cohort undergoing transplantation with a low CD31 cell dose (\ 177 x 106/kg), 52% (95% CI = 37%-67%), was lower than that seen in the cohort who underwent transplantation with a high CD31 cell dose ($ 177 x 106/kg), 77% (95% CI = 66%-86%) (P = .0075) (Figure 2B).
Months
Multivariate Analysis
Variables considered in the multivariate analyses were patient age group (< 20, 20-34, $ 35 years), donor-patient sex match, donor type (sibling vs others), number of HLA antigen mismatches, disease type, disease stage, graft type, CMV, MNCs ($ 7.5 vs < 7.5 x 108/kg), CD341 ($ 2 vs < 2 x 106/kg), CD31 ($ 177 vs < 177 x 106/kg), CD41 ($ 100 vs < 100 x 106/kg), and CD81 ($ 75 vs < 75 x 106/kg). Multivariate analysis (Table 3) showed that advanced-stage disease at the time of transplantation was associated with a worse LFS (2.56; 95% CI = 1.33-4.95; P = .0050). Inversely, a high dose of CD31 cells in the graft was associated with better LFS (0.46; 95% CI = 0.26-0.84; P = .0106) (Table 3). The disease status before transplantation and CD31 cell dose given also had a significant influence on OS. Advanced-stage disease predicted a worse outcome (relative risk [RR] = 2.77; 95% CI = 1.39-5.53; P = .0038). However, patients had a significantly better survival if the transplant contained a high CD31 cell dose, $ 177 x 106/kg (RR = 0.42; 95% CI = 0.23-0.78; P = .0058) (Table 3). Donor-recipient pair relationships showed no significant effect on outcomes.
Grades III-IV aGVHD posttransplantation had a significant effect on the various endpoint outcomes measured (Table 4). However, cGVHD had no obvious impact on these endpoints—a finding that could change with a longer follow-up time. Univariate and multivariate analysis showed that the numbers of MNCs, CD341, CD41, or CD81 cells in the transplants had no significant influence on these clinical endpoints (Table 5).
DISCUSSION
The present study is the largest study to date evaluating the predictors of developing GVHD, TRM,
Months
Figure 2. Adjusted probabilities of LFS and OS after haploidentical HCT for patients with leukemia (adjusted for early, intermediate, and advanced disease status). A, Adjusted probability of LFS (P = .0115 at 2 years). B, Adjusted probability of OS (P = .0075 at 2 years).
relapse, and survival in patients who, after receiving an identical conditioning regimen, (including ATG), underwent transplantation with unmanipulated nucleated cells from haploidentical blood and marrow donors. Stable engraftment and long-term survival after family mismatched/haploidentical HCT likely involves multiple mechanisms such as the conditioning regimen [11,12], inherited maternal/paternal antigen (NIMA/NIPA) mismatch existing in donor-recipient pairs [20,21], disease status [22,23], donorversus recipient natural killer cell alloreactivity [23,24], and the number of CD341 cells infused [8,11,23]. Previous reports were based mainly on the results of CD341 cell selected transplantation, which involved marked T-cell depletion in vitro. Thus, the CD31 cell doses given in those studies were very low, ranging from 103/kg to 105/kg of recipients' body weight [8,11,25,26]. Including ATG in the pretransplantation conditioning regimen provides an effective means of depleting T cells in vivo and preventing severe GVHD. But intensive T cell depletion in vivo by ATG also can delay immune reconstitution, resulting in a high mortality from viral and fungal infections [27,28]. Studies of ATG clearance in HSCT recipients, using ELISA for its detection, showed that significant levels of ATG can remain in the blood for up to 5 weeks posttransplantation [29,30]. Using different commercial preparations and conditioning regimens, as well as different timings and doses of ATG, may contribute to the discrepancies in the reports from different centers.
G-CSF-mobilized bone marrow induces early neutrophil engraftment with a reduced risk of GVHD [31,32]. Ji et al [32] reported that using G-CSF-primed haploidentical marrow grafts along with combined sequential use of CSP, MMF, ATG, and MTX provided an excellent alternative for the treatment of high-risk hematologic malignancy in patients without matched donors. The incidence of aGVHD was surprisingly low, only 6.3% in G-CSF-primed marrow transplants, compared with 33% in steady-state marrow transplants [32]. Recently, the clinical phenomenon has been confirmed by experimental data; in vitro mixtures of G-PB and G-BM in different proportions (G-PB:G-BM = 2:1, 1:1, and 1:2) has successfully introduced T cell hyporesponsiveness and polarization of T cells from Th1 to Th2 [33]. Under the current BM/PB harvest protocol, G-BM was harvested on day 0 after 4 days of G-CSF (target dose, 15 mL/kg recipient weight), and G-PB was harvested in a single leukapheresis on day 1 after 5 days of G-CSF therapy. In general, the number of MNCs obtained from G-BM was 3.87 (1.1-9.52) x 108/kg, and that obtained from G-PB was 3.795 (1.17-12.1) x 108/kg, with a BM:PB ratio of 1.006 (0.2-6.5) (results from analysis of our unpublished data, n = 98).
G-BM also may add BM-derived mesenchymal stromal cells (MSCs) to grafts. In vitro, MSCs inhibit the proliferation of activated T cells and the formation of cytotoxic T cells. In vivo, they appear to have anti-inflammatory effects. Some previous results have shown that MSC therapy can facilitate engraftment [34] and decrease GVHD [35-37]. The affect ofMSCs co-transplanted in allo-HSCT on the effect of graft versus leukemia (GVL) response and the pathophysiologic mechanism for decreasing graft versus host (GVH) response remains under investigation [38-40]. Furthermore, the significance of the BM:PB ratio in the graft, the immune characteristics of the T-cell subsets, dendritic cells, monocytes, natural killer cells (NKC), and the cytokine profile introduced by combining G-mobi-lized BM with G-PB need further investigation.
In our results, overall CD341 doses below those reported by other centers may result from strict gating for CD341 cell assay. Moreover, the dose of G-CSF that we used (5 mg/kg/day) was significantly lower than that reported by others (10-15 mg/kg/day) [31,41,42]. Grafts mobilized by G-CSF have marked tolerogenic properties. The immune characteristics of the G-CSF-mobilized peripheral blood stem cells (PBSCs) and BM may partially explain why the relatively high CD31 cell dose in the transplant did not increase the cumulative incidence of severe GVHD over 100 days or increase the incidence of cGVHD posttransplantation. Our results correspond somewhat to those of a recent report in which the 2 risk factors traditionally associated with aGVHD—age and the infused doses of CD31, CD41, CD81, or CD341 cells with PBSC grafts—did not significantly increase the incidence of GVHD [43].
In conclusion, our data demonstrate that with the currently available treatments, the dose of donor T cells used in this protocol is rational and acceptable. Relatively high numbers of MNCs and CD341, CD41, and CD81 cells and a significantly high CD31 cell dose contribute to better posttransplantation survival and enhanced LFS, resulting in part from a reduction in treatment-related toxicity without an increase in severe GVHD.
ACKNOWLEDGMENTS
We are indebted to the patients and their families and express our gratitude for their participation in this study. We also acknowledge grants from the national "211 Project'' (92000-242156014) and research financial support from the Peking University EBM group. Special thanks go to all of the staff members at the Division of Hematopoietic Cell Transplantation, the HLA Lab, and the FCM Lab at Peking University and Beijing Dao-Pei Hospital for their diligence and excellent work. We also thank Dr John C. Herion for his critical reading and helpful revisions to the manuscript and Dr Yan-Rong Liu for excellent laboratory support by providing the FACS analysis.
REFERENCES
1. Barnes DWH, Corp MJ, Loutit JF, et al. Treatment of murine leukemia with x-rays and homologous bone marrow: preliminary communication. BMJ. 1956;2:626-627.
2. Appelbaum FR. Hematopoietic cell transplantation as a form of immunotherapy. Int J Hematol. 2002;75:222-227.
3. Bortin MM, Gale RP, Kay HEM, et al. Bone marrow transplantation for acute myelogenous leukemia. JAMA. 1983;249: 1166-1175.
4. Dominietto A, Lamparelli T, Raiola AM, et al. Transplant-related mortality and long-term graft function are significantly influenced by cell dose in patients undergoing allogeneic marrow transplantation. Blood. 2002;100:3930-3934.
5. Urbano-Ispizua A, Rozman C, Pimentel P, et al. The number of donor CD3 1 cells is the most important factor for graft failure after allogeneic transplantation of CD341 selected cells from peripheral blood from HLA-identical siblings. Blood. 2001;97: 383-387.
6. MohtyM, Bagattini S, Chabannon C, et al. CD81 T-cell dose affects development of acute graft-versus-host disease following reduced-intensity conditioning allogeneic peripheral blood stem cell transplantation. Exp Hematol. 2004;32:1097-1102.
7. Schuening F, Miller AD, Torok-Storb B, et al. Study on contribution of genetically marked peripheral blood repopulating cells to hematopoietic reconstitution after transplantation. Hum Gene Ther. 1994;5:1523-1534.
8. Siena S, Schiavo R, Pedrazzoli P, et al. Therapeutic relevance of CD34 cell dose in blood cell transplantation for cancer therapy. J Clin Oncol. 2000;18:1360-1377.
9. Aversa F, Martelli MM, Reisner Y. Use of stem cells from mismatched related donors. Curr Opin Hematol. 1997;4: 419-422.
10. Cao TM, Wong RM, Sheehan K, et al. CD34, CD4, and CD8 cell doses do not influence engraftment, graft-versus-host disease, or survival following myeloablative human leukocyte antigen-identical peripheral blood allografting for hematologic malignancies. Exp Hematol. 2005;33:279-285.
11. Handgretinger R, Schumm M, Lang P, et al. Transplantation of megadoses of purified haploidentical stem cells. Ann N Y Acad Sci. 1999;872:351-362.
12. Tabilio A, Falzetti F, Zei T, et al. Graft engineering for allogeneic haploidentical stem cell transplantation. Blood Cells Mol Dis. 2004;33:274-280.
13. Lu DP, Dong L, Wu T, et al. Conditioning including antithy-mocyte globulin followed by un-manipulated HLA-mis-matched/haploidentical blood and marrow transplantation can achieve comparable outcomes to HLA-identical sibling transplantation. Blood. 2006;107(8):3065-3073.
14. Drach J, Zhao S, Drach D, et al. Expression ofMDR by normal bone marrow cells and its implication for leukemia hematopoi-esis. Leukemia Lymphoma. 1995;16:419-424.
15. Thomas ED, Storb R, Clift RA, et al. Bone marrow transplantation. New Engl J Med. 1975;292:832-843, 895-902.
16. Storb R, Deeg HJ, Fisher L, et al. Cyclosporine versus methotrexate for graft-versus-host disease prevention in patients given marrow grafts for leukemia: long-term follow-up of three controlled trials. Blood. 1988;71:293-298.
17. Glucksberg H, Fefer RS, Buckner CD, et al. Clinical manifestations of graft-versus-host disease in human recipients of marrow from HLA-matched sibling donors. Transplantation. 1974;18: 295-304.
18. Sullivan KM, Agura E, Anasetti C, et al. Chronic graft-versus-host disease and other late complications of bone marrow transplantation. Semin Hematol. 1991;28:250-259.
19. Shulman HM, Sullivan KM, Weiden PL, et al. Chronic graft-versus-host syndrome in man: a long-term clinicopathologic study of 20 Seattle patients. Am J Med. 1980;69:204-217.
20. Ichinohe T, Uchiyama T, Shimazaki C, et al. Feasibility of HLA-haploidentical hematopoietic stem cell transplantation between noninherited maternal antigen (NIMA)-mismatched family members linked with long-term fetomaternal microchi-merism. Blood. 2004;104:3821-3828.
21. Van Rood JJ, Claas F. Noninherited maternal HLA antigens: a proposal to elucidate their role in the immune response. Hum Immunol. 2000;61:1390-1394.
22. Marks DI, Aversa F, Lazarus HM. Alternative donor transplants for adult acute lymphoblastic leukaemia: a comparison of the three major options. Bone Marrow Transplant. 2006;38: 467-475.
23. Aversa F, Terenzi A, Felicini R, et al. Haploidentical stem cell transplantation for acute leukemia. Int J Hematol. 2002;76(Suppl 1):165-168.
24. Ruggeri L, Mancusi A, Capanni M, et al. Donor natural killer cell allorecognition of missing self in haploidentical hematopoi-etic transplantation for acute myeloid leukemia: challenging its predictive value. Blood. 2007;110:433-440.
25. Marks DI, Khattry N, Cummins M, et al. Haploidentical stem cell transplantation for children with acute leukaemia. Br J Hae-matol. 2006;134:196-201.
26. Amrolia PJ, Muccioli-Casadei G, Huls H, et al. Adoptive immu-notherapy with allodepleted donor T-cells improves immune reconstitution after haploidentical stem cell transplantation. Blood. 2006;108:1797-1808.
27. Naujokat C, Berges C, Fuchs D, et al. Antithymocyte globulins suppress dendritic cell function by multiple mechanisms. Transplantation. 2007;83:485-497.
28. Seidel MG, Fritsch G, Matthes-Martin S, et al. Antithymocyte globulin pharmacokinetics in pediatric patients after hematopoi-etic stem cell transplantation. J Pediatr Hematol Oncol. 2005;27: 532-536.
29. Waller EK, Langston AA, Lonial S, et al. Pharmacokinetics and pharmacodynamics of anti-thymocyte globulin in recipients of partially HLA-matched blood hematopoietic progenitor cell transplantation. Biol Blood Marrow Transplant. 2003;9: 460-471.
30. Remberger M, Sundberg B. Rabbit-immunoglobulin G levels in patients receiving thymoglobulin as part of conditioning before unrelated donor stem cell transplantation. Haematologica. 2005; 90:869B.
31. Serody JS, Sparks SD, Lin Y, et al. Comparison of granulocyte colony-stimulating factor (G-CSF)-mobilized peripheral blood progenitor cells and G-CSF-stimulated bone marrow as a source of stem cells in HLA-matched sibling transplantation. Biol Blood Marrow Transplant. 2000;6:434-440.
32. Ji SQ, Chen HR, Wang HX, et al. G-CSF-primed haploiden-tical marrow transplantation without ex vivo T cell depletion: an excellent alternative for high-risk leukemia. Bone Marrow Transplant. 2002;30:861-866.
33. Huang XJ, Chang JC, Zhao XY. Maintaining hyporesponsive-ness and polarization potential of T cells after in vitro mixture of G-CSF-mobilized peripheral blood grafts and G-CSF-primed bone marrow grafts in different proportions. Transplant Immunol. 2007;17:193-197.
34. Le Blanc K, Samuelsson H, Gustafsson B, et al. Transplantation of mesenchymal stem cells to enhance engraftment of hemato-poietic stem cells. Leukemia. 2007;21:1733-1738.
35. Le Blanc K. Mesenchymal stromal cells: tissue repair and immune modulation. Cytotherapy. 2006;8:559-561.
36. Ringdén O, Uzunel M, Rasmusson I, et al. Mesenchymal stem cells for treatment of therapy-resistant graft-versus-host disease. Transplantation. 2006;81:1390-1397.
37. Lazarus HM, Koc ON, Devine SM, et al. Co-transplantation of HLA-identical sibling culture-expanded mesenchymal stem cells and hematopoietic stem cells in hematologic malignancy patients. Biology Blood Marrow Transplant. 2005;11: 389-398.
38. Aggarwal S, Pittenger MF. Human mesenchymal stem cells modulate allogeneic immune cell responses. Blood. 2005;105: 1815-1822.
39. Chung NG, Jeong DC, Park SJ, et al. Co-transplantation of marrow stromal cells may prevent lethal graft-versus-host
disease in major histocompatibility complex mismatched murine hematopoietic stem cell transplantation. Int J Hematol. 2004;80: 370-376.
40. Gieseke F, Schutt B, Viebahn S, et al. Human multipotent mes-enchymal stromal cells inhibit proliferation of PBMCs independently of IFNgR1-signaling and IDO expression. Blood. 2007; 110(6):2197-2200.
41. Isola LM, Scigliano E, Skerrett D, et al. A pilot study of al-logeneic bone marrow transplantation using related donor stimulated with G-CSF. Bone Marrow Transplant. 1997;20: 1033-1037.
42. Couban S, Messner HA, Andreou P, et al. Bone marrow mobilized with granulocyte colony-stimulating factor in related allo-geneic transplant recipients: a study of 29 patients. Biol Blood Marrow Transplant. 2000;6:422-427.
43. Saliba RM, Lima MD, Giralt S, et al. Hyperacute GVHD: risk factors, outcomes, and clinical implications. Blood. 2007;109: 2751-2758.
