CC BY-NC-ND
0Biology of Blood and Marrow Transplantation 14:323-334 (2008) © 2008 American Society for Blood and Marrow Transplantation 1083-8791/08/1403-0001$32.00/0 doi:10.1016/j.bbmt.2007.12.497
American Society for Blood and Marrow Transplantation
Effects of the NK Cell Recovery on Outcomes of Unmanipulated Haploidentical Blood and Marrow Transplantation for Patients with Hematologic Malignancies
Ying-Jun Chang, Xiang-Yu Zhao, Xiao-Jun Huang
Peking University Institute of Hematology, Peking University People's Hospital, Beijing, People's Republic of China
Correspondence and reprint requests: Xiao-jun. Huang, MD, Peking University, Institute of Hematology, Peking University People's Hospital, No. 11 Xizhimen South Street, Beijing 100044, P.R. China (e-mail: xjhrm@medmail.com.cn).
Received August 16, 2007; accepted December 26, 2007
ABSTRACT
The goal of this study was to investigate the association of natural killer (NK) cell recovery with clinical outcomes after unmanipulated haploidentical blood and marrow transplantation. We sequentially monitored the reconstitution kinetics of circulating NK cells, CD56bright and CD56 m, in 43 patients by flow cytometry, and the functionality recovery of cytokine or cytotoxicity of NK cells by flow cytometry or lactate dehydrogenase release assay after transplantation. Reconstitution of NK cells was rapid but accompanied by skewing of cell subsets mainly in CD56bright, which recovered earlier. Linear regression analysis demonstrated that dose of CD341 cells in the allografts was inversely correlated with the ratio of T/NK cells (b = -0.506, P = .003) and CD56dim/ CD56bright cell (b = -.403, P = .018) by day 14 after hematopoietic stem cell transplantation (HSCT), and the dose of CD31 T cells in the allografts was also inversely correlated with the ratio CD56dim/CD56bright cells by day 14 after HSCT (b = -0.474, P = .005). Moreover, the dose of CD56dim NK cells in the allograft was positively associated with the day 14 CD56brigh NK cells (b = 0.494, P = .032) and inversely correlated with the day 14 ratio of CD56dim/CD56bright cells (b = -0.617, P = .005). Compared with nonacute graft-versus-host disease (GVHD) patients, patients with acute GVHD (aGVHD) had a higher level of NK subsets during week 2 posttransplantation. Cox regression analysis revealed that the patients with more CD56bright NK cells in the recovery stage had a higher survival rate (hazard risk [HR], 0.406; P = .017) and the patients with a higher ratio of T/NK (>1.0) had a higher chance of getting aGVHD (HR, 3.436; P = .059) and chronic GVHD (HR, 3.925; P = .028). Our results suggest that the recovery of NK cells is and can be used as an indicator to predicate the clinical outcomes after unmanipulated haploidentical transplantation. © 2008 American Society for Blood and Marrow Transplantation
KEY WORDS
HLA-mismatched • Haploidentical • HSCT • GVHD • CD56bright NK cells
INTRODUCTION
Natural killer (NK) cells play an important role in the innate host defenses. They specialize in killing virus-infected cells and tumor cells by secreting gran-zyme B similar to those of cytotoxic T lymphocytes. There are 2 distinct subsets of human NK cells— CD56dim and CD56bright NK cells—with distinct functional and phenotypic properties. Approximately 90% of NK cells are CD56 NK cells, expressing intermediate-affinity IL-2R and demonstrating enhanced cytotoxicity activity without proliferation in response to high (nanomolar) concentrations of IL-2,
whereas approximately 10% are CD56bright NK cells, expressing high-affinity IL-2R and proliferating in response to low (picomolar) concentrations of IL-2 [1,2]. Generally, resting CD56dim NK cells are more cytotoxic against NK-sensitive targets than CD56bright; however, by activation with IL-2 or IL-12, CD56bright cells exhibit similar or enhanced cytotoxicity against NK targets compared to CD56dim cells. CD56bright NK cells are identified as the executor of NK cells both in cytokine production and in innate immunoregulatory [3]. Moreover, CD56bright NK cells, expressing high levels of lymph node (LN)
homing molecules L-selectin, CCR7 and CXCR3, are often presented in normal human LN-enriched T cells and dendritic cells (DCs) [4], indicating that interactions between these cells occur in vivo [4,5].
NK cells are the first cells to recover after autologous or allogeneic BM or peripheral blood stem cell (PBSC) transplantation. The effects of NK cells and its alloreactivity on the outcomes of transplantation are controversial in different transplant settings [6-8]. Nguyen et al [9] reported that the increased fraction ofCD56brightNK cells was associated with the impairment of early reconstituted NK cytotoxicity, which accounted for the high relapse rate after purified CD341 haploidentical transplantation. However, no studies were done in the past to show the reconstitution kinetics of NK cells and its contribution to clinical outcomes after unmanipulated haploidentical transplantation, which promotes us to monitor the orderliness of circulating NK and its subsets after haploidentical blood and marrow transplantation in our model. Our studies suggested that the reconstituted CD56bright NK cells was an early indicator for survival, and the ratio of T/NK cells was a useful predictor for acute and chronic GVHD (aGVHD, cGHVD) after transplantation.
MATERIALS AND METHODS
Patients
Patients with hematologic malignancies suitable for allogeneic hematopoietic stem cell transplantation
(HSCT) without HLA-identical related or unrelated donors were candidates for the haploidentical HSCT. Forty-three patients with malignant hematologic disease, who underwent haploidentical allo-HSCT between April 2004 and April 2005, were included in this study. All patients and their donors gave written informed consent, and the institutional review board of Peking University Institute of Hema-tology approved this study. The characteristics of the 43 patients are shown in Table 1.
Conditioning Regimen, Mobilization, Collection of Stem Cells, and GVHD Prophylaxis
All patients received myeloablative regimens including a combination of cytosine arabinoside (Ara-C) (4 g/m2 x 2 days, on days —10 and —9), busulfan (12 mg/kg administered orally in 12 doses over 3 days, on days —8, — 7, and —6), cyclophosphamide (1.8 g/m2 x 2 days, on days —5 and —4), Simustine (MeCCNU) (250 mg/kg on day —3), and porcine antihuman thymocyte immunoglobulin (Thymoglobulin 2.5mg/kg per day i.v. of the Sangstat product) on days — 5 through —2 [10,11].
Donors received rhG-CSF (Filgrastim) 5 mg/kg daily for 5 to 6 days. On the fourth day, bone marrow cells were harvested. The target total nucleated cell count reached 3.0 x 108 (median, 3.6 x 108, range, 0.24-8.16 x 108) cells/kg recipient weight. On the 5th and 6th days, peripheral blood progenitor cells
Table 1. Patients and Graft Characteristics of the Low, Middle, and High CD56bright NK Patient Group
Characteristics Overall Low CD56bright Group Middle CD56bright Group High CD56bright Group
N 43 14 15 14
Patient age, Mean ± SD 26.7 ± 13.6 27.8 ± 17 25.3 ± 8.7 26.1 ± 13.7
Donor age, Mean ± SD 41 ±8.5 38.8 ± 9.8 44.9 ± 8.2 41.36 ± 5.8
Patent sex, male, n (%) 29(67%) 9(64%) 12(80%) 8(57%)
Diagnosis
AML, n (%) 14(33%) 4(29%) 6(40%) 4(29%)
ALL, n (%) 12(28%) 5(36%) 3(20%) 4(29%)
CML, n (%) 16(37%) 5(36%) 5(33%) 6(43%)
MDS, n (%) 1(2%) 0(0%) 1(7%) 0(0%)
High risk,* n (%) 17(40%) 8(57%) 6(40%) 3(21%)
Patient/donor HLA
compatibility
Single locus (%) 10(23%) 4(29%) 2(13%) 4(29%)
Two locus (%) 17(40%) 5(36%) 8(53%) 4(29%)
Three locus (%) 16(37%) 5(36%) 5(33%) 6(43%)
Mean ± SD
CD34 infused, x l06/kg 2.6 ± 1.4 2.6 ± 1.5 2.1 ± 1.1 3.0 ± 1.4
CD3 infused, x l08/kg 2.2 ± 1.4 1.9 ± 1.4 1.4 ±0.8 2.2 ± 0.6
CD56 infused, x l08/kg 0.25 ±0.16 0.13 ±0.01 0.24 ±0.21 0.29 ± 0.09
CD56dim infused, xl08/kg 0.21 ± 0.11 0.11 ± 0.02 0.19 ± 0.12 0.26 ± 0.09
Patients were subgrouped into ''high,'' "middle," and ''low'' groups based on a 33 and 67 percentage of circulating CD56bright NK cell counts by
day 14 after transplantation. MDS indicates myelodysplastic syndromes.
*Pretransplantation risk category included standard or high risk. Patients with standard risk were defined as those undergoing transplantation during the first complete remission (CR1) of acute leukemia or the first chronic phase (CP1) of chronic myelogenous leukemia. Patients with high risk were those with myelodysplastic syndrome and a more advanced stage of acute leukemia or chronic myelogenous leukemia other than CR1/CP1.
(PBPCs) were collected. The target mononuclear cell count reached 3.0 x 108 (median, 3.65 x 108, range, 2.77-11.7 x 108) cells/kg recipient weight. The fresh and unmanipulated bone marrow and PBPCs were infused into the recipients on the day of collection [10,11].
Prophylaxis for GVHD included cyclosporine A (CSA) and short-term methotrexate (MTX) with my-cophenolate mofetil (MMF) [10,11]. Cyclosporine was started intravenously on day —9, at the dosage of 2.5 mg/kg, and switched to oral formulation as soon as the patient was able to take medication after engraft-ment. The dosage was adjusted to blood levels. MMF was administered orally, 0.5 g every 12 hours, from day 9 before transplantation to day 30 after transplantation, then 0.25 twice a day for 1 to 2 months. The dosage ofMTX was 15 mg/m2, administered i.v. on day 1, and 10 mg/m2 on days 3, 6, and 11 after transplantation. The diagnosis and grading of GVHD was established according to published criteria [12]. Filgrastim (granulocyte-colony stimulating factor [G-CSF]) 5 mg/kg per day was given to all recipients subcutane-ously from day 6 after transplantation until the neutrophil count reached 0.5 x 109 cells/L for 3 consecutive days [10,11]. Bone marrow aspiration and cytogenesis studies were performed at 1, 2, and 3 months after transplantation to assess engraftment. HLA DNA typing and PCR-DNA fingerprinting (short tandem repeat) were used for donor chimerism detection. For each patient, at least 2 methods were used to confirm donor chimerism.
Immunopheotypic Analysis
Blood (10 mL) was drawn from 43 patients at day 14, 21, 30, 60, 120, and 180 after transplantation, and from 15 donors. All the samples were drawn within 4 days ofthe specified days. Peripheral blood (PB) lymphocytes were isolated using Ficoll-Hypaque density centrifugation, and were analyzed for reactivity with a series of mAbs using standard techniques. Monoclonal antibodies (mAbs) CD16-FITC, CD56-PE were used in combination with anti-CD45-Percp and anti-CD3-APC (BD Bioscience, Mountain View, CA). Individual 4-color flow cytometry assays were used to analyze the immunophenotype recovery of the lymphocyte including CD31 T cells, CD561 NK cells and its subsets after transplantation. Samples were available on 43 patients at day 14, 26 patients at day 21, 43 patients at days 30 and 60, 27 patients at day 120, and 26 patients at day 180.
NK Cell Preparation and Cytotoxicity Assay
PB samples were collected from the donors prior to BM harvest and PBSC collection (n = 8) and from patients at day 30 (n = 11). Patient- (n = 3) or donor-derived (n = 3) NK cells [NK cells were isolated
by depletion of T, B, and myeloid cells (Miltenyi Biotec, Bergish Gladbach, Germany; purity $95%)] and patient (n = 8) or donor (n = 5) PBMNCs were cultured respectively in medium (RPMI with 10% fetal calf serum, 100 U/mL penicillin, 100 U/mL streptomycin, and 1% glutamine) with or without the presence of 1000 IU/mL IL-2 (Xiamen Amoytop Biotech Co. Ltd, Fujian, China) for 16-18 hours for spontaneous and IL-2-stimulated (LAK) NK cytotoxicity assay.
Cytotoxicity of NK cells was examined using major histocompatibility complex (MHC) class I-deficient human erythroleukemia K562 cell line as targets, at effector-target ratios ranging from 20:1 to 2.5:1. Commercial cytotoxicity assays based on lactate dehydrogenate (LDH) detection was used according to the manufacturer's instructions (Cytotox 96; Promega, Madison, WI).
Interferon-g (IFN-g) Production Detected by Intracellular Flow Cytometry
Intracellular staining was performed with the Par-mingen Intracellular Staining kit (BD Parmingen, San Diego, CA). Donor (n = 5) and recipient (n = 9) PBMCs by day 30 (5 x 106/mL) were incubated with medium alone, or 1000 IU/mL IL-2 for 16-18 hours. Then cells were incubated for 5 hours with or without phorbol myristate acetate (PMA) (50 ng) plus 100 ng of inomycin (all reagents from Sigma Chemical, St. Louis, MO) to stimulate maximal IFN-g production. Golgistop (5 mL/mL) was added during the last 4 hours to trap the protein in the cytoplasm. The remaining operations followed the instruction of Parmingen Intracellular Staining kit. Monoclonal antibodies CD3-percp, CD56-PE, and IFN-g-FITC (BD Bioscience) stained for surface markers and intracellular cytokines, respectively.
Statistical Analysis
Nonparametric Mann-Whitney's U test was used in evaluation the difference of NK cell recovery between healthy donors and patients after transplantation as well as between the patients with 0-I aGVHD and II-IV aGVHD. Statistical associations between pretransplantation variables and lymphocyte recovery after transplantation were investigated using linear regression analysis. The associations between NK cells recovery and posttransplantation outcomes were analyzed by the methods of Kaplan-Meier or calculated using cumulative incidence curves to accommodate competing risks [13]. A log-rank test was used in survival, and a Grey test was used in cumulative incidence analyses. To confirm outcomes and adjust for potential confounding factors, multivariate Cox proportional hazards models were assessed for the proportional hazards assumption and for testing interaction terms with covariates. Factors included in the models were recipient and donor ages, sex,
diagnosis, HLA mismatch, pretransplantation risk category, dose of CD31 T cell group, and dose of CD341 cells and the absolute number of day 14 CD56bright NK cells, day 14 CD56dim NK cells, day 14 T cells, as well as the ratio of the day 14 CD56dim/CD56bright NK cells, day 14 T/NK cells. The final multivariate models were built using a forward stepwise model selection approach. Patients' characteristics in "high," "middle," versus "low" CD56bright NK cells groups were compared using chi-square statistics for categoric variables and Mann-Whitney U for continuous variables, respectively. The calculations were carried out with SPSS 13.0 statistical software. R software was used to calculate the cumulative incidence considering the presence of competing risk.
RESULTS
Patient Characteristics
Patient characteristics including age, sex, underlying hematologic disorder, the degree of HLA disparity, the amount of CD31 T cells and CD341 cells, NK cells and its subset CD56dim, were presented in Table 1. All the patients achieved engraftment and complete donor chimerism after transplantation. As of June 1, 2006, there were 29 patients who survived without leukemia, 11 patients died of transplant-related complications after HSCT, and 3 patients relapsed on days 300, 180, and 44, respectively. The median follow-up was 377 days (range: 43-700 days). There were 18 and 25 patients who had 0-I aGVHD and II-IV aGVHD in a series of 43 patients, respectively. Among 37 patients evaluated for the cGVHD who survived longer than 100 days, 19 patients had occurred limited (15 patients) and extensive (4 patients) cGVHD.
Reconstitution Kinetics of Circulating NK Cells after Unmanipulated Haploidentical Blood and Marrow Transplantation
The quantitative evaluation of circulating NK cells after HLA mismatched/haploidentical blood and marrow transplantation without in vitro T cells depletion was performed in a series of 43 patients. For comparison, 15 age-matched healthy donors were used as control. Because the occurrence of aGVHD and its therapy would affect the immune recovery, only 16 of 43 patients survived without leukemia and were exempt from II-IV aGVHD; the reconstitution kinetics of NK cells were analyzed in all 43 patients as well as in those 16 who never developed GVHD.
Reconstitution of NK cells was rapid, but accompanied by skewing of cell subset, with a fast increase in CD56bright NK subset and a corresponding reduction in CD56dimNK subset among the NK cells compared with those of healthy control during first 2
month after transplantation (P < .01). Then the ratio of CD56dim/CD 56bright NK subsets in patients eventually reached the level similar to that of healthy control by day 120 in those 16 who never developed GVHD (P = .055) (Figure 1A1), or by day 180 in all 43 patients (P = .191; Figure 1A2). The absolute number of CD56bright NK subset in white blood cells (WBCs) and number of CD56bright NK subset had recovered to the donor's level by day 14, and continuously increased up to their highest levels by day 60 in those 16 who never developed GVHD or by 120 in all 43 patients, which were higher compared with those of healthy controls (P < .05; Figure 1B1 and 1B2). In contrast, those levels of CD56dim NK subset were lower than the donor levels by day 14 (P < .01; Figure 1B1), then recovered to the donor levels by day 30 after transplantation in those 16 who never developed GVHD. However, the absolute number of CD56dim NK subset recovered later, which reached to those of donor levels by day 180 after HSCT in all 43 patients (Figure 1B2). Therefore, the absolute number of overall NK cells recovered to the donors' levels by day 30 in those 16 who never developed GVHD or by day 60 in all 43 patients.
It also should be noted that CD31T cells reconstituted more slowly than CD561 NK cells. Both of the proportion in WBC and the absolute number of the CD31 T cells in patients were significantly lower than those levels of the donors during 1 month after transplantation (P < .001), then recovered to the donors' levels by day 60. Accordingly, the ratio of T/ NK cells was lower in patients by day 14 (median: 0.86, range: 0-3.55, P < .0001) and 30 (median: 1.46, range: 0.14-12.14, P = .001) compared with that of donors (median: 7.43, range: 3.16-15.44), and recovered to the donors' levels by day 60 (Figure 1C1) in those 16 who never developed GVHD. The ratio of T/NK cells recovered to the donors' levels by day 120 in all 43 patients (Figure 1C2).
There were 6 children (age # 18years) among these 16 patients. When we assessed them separately, no differences were found in the NK or T cell immune recovery between the older (age >18 years) and the youger (age #18 years) (data not shown).
Functional Evaluation of NK Cells after Unmanipulated Haploidentical Blood and Marrow Transplantation
The cytotoxicity and cytokine secretion function of NK cells were evaluated in those patients without aGVHD and their donor. The capability of NK cells to produce IFN-g in response to nonspecific stimulation was analyzed by intracellular flow cytometry. The ability of the patient-derived NK cells with or without IL-2 preactivation to produce IFN-g was similar to those of donor-derived NK cells by day 30 after transplantation. Under IL-2 activation, the
Figure 1. Kinetics reconstitution of NK and its subsets after HLA-mismatched/haploidentical HCT without in vitro T cells depletion. Donors (D) or donor-derived patients NK cells and its subsets, CD56bright and CD56dim NK cells, as well as T cells after transplantation (30, 60, 90, 120, and 180 days) were monitored by cytometry. The median, 25 to 75 percentile ranges, and extreme values of the absolute numbers of NK and its subsets (A1 and A2), orratio of CD56dim/CD56brightNK cells (B1 andB2)andT/NK(C1 and C2) were shown in the box plot. Cells/mL indicates the absolute number in WBC of overall NK cells and its subset. The reconstruction kinetics of the 16 patients without II-IV aGVHD were shown in A1, B1, and C1; those of the overall 43 patients were presented in A2, B2, and C2. *Represent P < .05 compared to their donors. **Represent P < .01 compared to their donors.
cytokine production potential in patient- or donor-derived NK cells was increased by 2- to 3-fold (Figure 2A).
The cytotoxicity capability of purified patient- or donor-derived NK cells against K562 cells was assessed by LDH release assay. The spontaneous or
activated (IL-2 activated) killing potential of patient-derived NK cells (either using purified NK cells or peripheral blood mononuclear cells; Figure 2B and C, respectively) were lower than those of the donor derived cells by day 30 after transplantation (P < .01; Figure 2B and C).
Figure 2. The functionality recovery of cytokine and cytotoxicity of NK cells after transplantation. The capability of NK cells to produce IFN-g in response to PMA plus inomycin stimulation was analyzed by intracellular flow cytometry. The ability of the patient-derived NK cells with (□ ) or without IL-2 (■) preactivation to produce IFN-g was similar to those of donor-derived NK cells by day 30 after transplantation (A). The cytotoxicity capability of purified patient- or donor-derived NK cells and patient- or donor-derived PBMNCs against K562 cells was assessed by LDH release assay. The spontaneous or activated (IL-2 activated) killing potential of patient-derived NK cells or PBMNCs were lower than those of donor-derived by day 30 after transplantation (B,C; P < .01).
Interrelationship of Day 14 NK Recovery and Pretransplantation Variables
No correlation were found between the dose of CD341 cell or CD31T cells in the allograft and the absolute counts of NK cells at any time after transplantation. However, we found that the dose of CD341 cells in the allograft was inversely correlated with the ratio of
T/NK cells (b = —0.506, P = .003; Figure 3A) and CD56dim/CD56bri cells (b = —0.403, P = .018; Figure 3B) by day 14 after HSCT. The dose of CD31T cells was also inversely correlated with the ratio of CD56dim/CD56bright cells by day 14 after HSCT (b = —0.474, P = .005; Figrue 3C). The dose of NK cells in the allograft was examined in only 20 donors
Figure 3. Interrelationship of day 14NKrecove ry and doses of CD341, CD31 cells, and CD56dim NK cells in allograft. The dose of CD341 cells in the allograft inversely correlated with the ratio of T/NK cells (A) and CD56dim/CD56bri cells (B) by day 14 after HSCT. The dose of CD31T cells also inversely correlated with the ratio of CD56dim/CD56bright cells by day 14 after HSCT (C). The dose of CD56dim NK cells positively associated with the absolute number of the day 14 CD56brIght NK cell (E), therefore inversely correlating with the ratio of CD56dlm/CD56brIght cells by day 14 after HSCT (F).
and 3 in the low CD56bright group, 8 in the middle CD56bright group, 9 in the high CD56bright group (Table 1). The dose of CD56dimNKcells was positively associated with the absolute number of CD56 right NK cells on day 14 (b = 0.494, P = .032; Figure 3D) and inversely correlated with the day 14 CD56dim/CD56bri ratio (b = —0.617, P = .005; Figure 3E). There was no effect of dose of NK cells and its subsets on the day 14 T/NK cell ratio.
Effect of aGVHD Occurrence and Its Therapy on NK Cells Recovery after Unmanipulated Haploidentical Blood and Marrow Transplantation
The initial appearance of clinical signs of GVHD occurred at a median of 22.5 days posttransplantation (range: 12 to 33 days). As early as day 14 posttransplantation, significant differences in the distribution of lymphoid subsets were observed. Patients who subsequently developed II-IV aGVHD tended to have a greater proportion in WBC of CD561 NK cells and its subsets (P = .060 for CD56bright subset, Figure 4A; P = .012 for CD56dim subset, Figure 4B; P = .036 for CD561 NK cells, Figure 4C), CD31 T cells (P = .001; Figrue 4D), as well as a higher absolute numbers of CD31 T cells (P = .001; Figure 4E) by day 14 after transplantation than patients who developed 0-I aGVHD. When the ratio of T/NK was examined, significant differences were again noted (P = .004;
Figure 4F). Therefore, in the majority of cases, the striking difference in lymphocytes recovery in patients with grades 0-I and grades II-IV aGVHD was seen at a time before any clinical signs of aGVHD were detected.
However, when aGVHD was clinically evident and already treated, differences could no longer be found in the ratio of T/NK and the levels of T cells by day 60 posttransplantation. Furthermore, patients with II-IV aGVHD showed delayed recovery of NK cells and its subsets proportion in WBC compared with those with 0-I aGVHD by day 30 and/or 60 (P < .05). No difference in the distribution of the absolute number of NK cells and its subsets were observed in patients with 0-I and II-IV aGVHD.
Predictive Value of CD31 T Cells and Ratio of T/NK Cells on aGVHD and cGVHD
The early detection of elevated T cells, NK cells, and T/NK ratios reconstitution in individuals who developed II-IV aGVHD led us to examine the predictive value of specific phenotypic profiles. We again focused on day 14 posttransplant, because clinical signs had not yet appeared at this point in most patients who developed aGVHD.
Factors included in the stepwise multivariate Cox proportional hazards model were recipient and donor ages, sex, diagnosis, HLA mismatch,
Figure 4. The occurrence ofaGVHD delays the recovery of the NK subset. Donors (D) or donor-derived patient NK cells and its subsets, as well as T cells after transplantation (30, 60, 90,120, and 180 days) were monitored by cytometry. Results were presented for donors or patients separated by 0-I and II-IVaGVHD. The median, 25 to 75 percentile ranges, and extreme values of the percentages or absolute number of NK and its subsets, as well as T cells or ratio of T/NK cells (A-F) are shown in the box plot. TNC% indicates the absolute proportion in WBC of overall NK cells and its subset, or of T cells (A-D); cells/mL indicates the absolute number in WBC of T cells (E); black and white column represent the groups of 0-I and II-IV GVHD, respectively. *Represents P < .05 between patients with 0-I and II-IV aGVHD; "represents P < .01 between patients with 0-I and II-IV aGVHD.
pretransplantation risk category, dose of CD31 T cells, and dose of CD341 cells and the absolute number of day 14 NK cells and its subsets, day 14 T cells, as well as the ratio of the day 14 CD56dim/CD56bright NK cells, day 14 T/NK cells. The account of CD31 T cells (T14) and ratio of T/NK cells by day 14 after transplantation emerged as independent factors strongly associated with aGVHD; meanwhile, the ratio of T/NK was the only significant risk factors for cGVHD (Table 2). There was no impact of age, sex, pretransplantation risk category, numbers of HLA-A/B/DR loci mismatch between donor-recipients, or dose of CD341 or CD31 T cells in the allograft and absolute account of NK cells and its subsets by day 14 posttransplantation (NK14, CD56bri14, CD56dim14) on aGVHD and cGVHD. Patients were subgrouped into ''low'' and "high" CD31 T cells group or ''low'' and "high" T/NK ratio group according to the optimal cutoff point of circulating T14 (34.0 cells/mL) or T/NK ratio (1.0) based on their respective ROC curves and sensitivity/specificity curves. Patients in the ''high'' CD31 T cells group had the higher cumulative incidence of II-IV aGVHD (99.4% versus 44.4%, P = .001) and III-IV aGVHD (70.3% versus
Table 2. Multivariate Analysis for LFS, II-IVaGVHD, and cGVHD Outcome and Factors OR (95%CI) P-Value
CD56brli4 group 0.406(0.193-0.852) 0.017
2-4 GVHD
TI4 (>34 cells/mL vs. 2.162(1.023-6.665) 0.045
<34 cells/mL)
Ratio of T/NK (> 1.0 vs. # 1.0) 3.436(0.954-12.378) 0.059
ratio of T/NK (> 1.0 vs. # 1.0) 3.925(1.157-13.315) 0.028
To avoid potential confounding factors, multivariate Cox proportional hazards models were assessed for testing interaction terms with covariates. Factors included in the multivariate models were recipient and donor ages, sex, diagnosis, HLA mismatch, pre-transplantation risk category, dose of CD3 1 T cells, and dose of CD341 cells and the absolute number of day 14 NK cells and its subsets, day 14 T cells, as well as the ratio of the day 14 CD56dim/CD56bright NK cells, and day 14 T/NK cells. The final multivariate models were built using a forward stepwise model selection approach. Patients were subgrouped into a "high," ''middle,'' and "low" group based on a 33 and 67 percentage of cir culating CD56bright NK cell count by day 14 after transplantation (CD56bri14 group). Cutoff points of day 14 T cell count (T14) or day 14 ratio of T/NK was 34 cells/mL or 1.0 by day 14 after transplantation.
OR indicates odds ratio; CI, confidence interval; LFS, .
10.5%, P = .003) compared with those in the "low" CD31 T cells group.
The ratio of T/NK cells was also a significantly independent positive predicator for cGVHD (HR, 3.925, 95% confidence interval [CI], 1.157-13.315, P = .028, Table 2). As showed in Figure 5A, B, and C, cumulative incidence curves for II-IV, III-IV aGVHD, and cGVHD were shown with a "low" (#1.0, n = 14) versus "high" (>1.0, n = 29) ratio of T/NK cells. Patients in the "high" ratio of the T/NK cells group had the higher cumulative incidence of II-IV aGVHD (82.4% versus 23.8%, P = .002; Figure 5A) and III-IV aGVHD (73.1% versus 7.1%, P = .017; Figure 5B) compared with those in the "low" ratio of the T/NK cell group. Because 6 patients died before day 100, only 37 of 43 patients enrolled in the analysis of the relationship between the ratio of T/NK cells and the occurrence of cGVHD after transplantation. The cumulative incidence of cGVHD in patients with a "high" T/NK ratio were significantly higher than those patients with a "low" T/NK ratio (73.1% versus 28.4%, P = .013; Figure 5C). However, no differences in the occurrence of cGVHD were found between the patients in the "high" and "low" (64.7% versus 55.4%, P = .554) CD31 T cells groups.
Predictive Value of Early Posttransplantation CD56bright NK Cells Count on Survival
Using the same Cox model, the only independent variable affecting the overall survival (OS) was the number of the day 14 CD56bright NK cell (HR, 0.406, 95% CI, 0.193-0.852, P = .017; Table 2). To assess the possibility of using the number of CD56bright NK cells as a positive predictor for LFS, patients were categorized into the "low" group (<3.59 cells/mL, n = 14), "middle" group (3.59 to 9.27 cells/mL, n = 15), and "high" group (>9.27 cells/mL, n = 14), based on
33% and 67% percentage of circulating CD56bright NK cell count. As showed in Table 1, there were no significant difference in the patients' characteristics including age, sex, diagnosis, pretransplantation risk category, the degree of HLA disparity, and the dose of CD31T cells, CD341 cells, NK cells and its subsets in the grafts among the 3 groups. A highest 2-year cumulative incidence of OS was observed in patients of the "high" CD56bright NK cells group (92.9% versus 42.9% of the "low" group, P = .006), followed by patients in the "middle" (66.7% versus 92.9%, P = .064) and "low" CD56bright NK cell group (42.9%) (Figure 6A). Furthermore, Kaplan-Meier univariate analysis showed the patients with rapid recovery of CD56bright NK cells seemed to be associated with less transplantation-related mortality (TRM) and relapse. The cumulative incidence of TRM and relapse in patient of the "high" group tended to be lower than those of the "low" group (7.1% versus 44.4% of TRM, P = .025; Figure 6B, and 0% versus 20.5% of relapse, P = .090, respectively).
DISCUSSION
In the current study, we investigated the reconstitution kinetics of peripheral NK cells after HLA-mis-matched/haploidentical HCT without in vitro T cells depletion and found that CD56bright subset domi-nanted in the early stage, and this prolonged increasing of CD56bright NK cells could be associated with impairment of NK cytotoxicity during the early recovery of NK cells. This result is consistent with the study in unmodified HLA-identical or purified CD34+ cells HLA-haploidentical HCT [9,14,15]. Furthermore, the early rapid recovery of circulating CD56bright NK cells was correlated with better survival; "high" ratio of T/NK (>1.0) was an independent predictor for
Figure 5. The high ratio of T/NK cells correlated with the increased occurrence of aGVHD and cGVHD after HSCT. Patients were sub-grouped into the "low ratio of T/NK cells (# 1.0)'' and the "high ratio of T/NK cells (>1.0)'' groups based on less or more than the ratio of 1.0, and subgrouped into the ''low'' and ''high'' CD31 T cells group according to the median circulating CD31 T cells counts of 31.36 cells/ mL by day 14 after transplantation. Cumulative incidence estimates of II-IV aGVHD (A), III-IV aGVHD (B), and cGVHD (C) in these 2 and 4 cohorts, respectively.
Figure 6. The high number of cir culating CD56bright NK cells correlated with better OS and less TRM. Patients were subgrouped into the ''high,'' ''middle,'' and ''low'' group based on the 33 and 67 percentage of ci rculating CD56bright NK cell count by day 14 after transplantation. Kaplan-Meier survival estimates OS (A) in these 3 groups and cumulative incidence estimates TRM (B).
posttransplantation aGVHD and cGVHD in this transplantation protocol.
Compared with result demonstrated by Nguyen et al [9], which showed that 7 out of 10 relapsed and 3 out of 10 died, the clinical outcomes of our study were encouraging; only 3 patients out of a total of 43 relapsed and 11 died of transplantation-related complication after haploidentical transplant without in vitro T cells depletion. In our transplantation model the number of NK cells and its subsets' reconstitution kinetics are similar to that of the Nguyen et al report. In Nguyen's study, CD31 T cells reconstituted slowly and could not be detected until the fourth month after purified CD341 haploidentical transplantation. The loss of T cells results in the deficiency of T cell antitumor function. The rapid reconstituted NK cells may play an antitumor role in the early stage after purified CD341 HSCT. The number of CD31T cells reached that of the donor's level 2 months after transplantation in our model. Both numbers of CD56dim and CD56bright NK reached the number of those in donor's body 1 month after transplantation. The rapid recovery of NK and T cells after HSCT may enable the interaction between the NK and T cells, which may enhance the individual antitumor and anti-infection effect of reconstituted NK and T cells after transplantation. Previous studies have demonstrated that the CD56bright NK cells are enriched in the T cell region of resting LNs [4,5], and the communication among CD56bright NK cells, DC cells, and T cells is crucial for the interaction between adaptive and innate immunity [4,5]. As far as the cytotoxicity of NK cells is concerned, we found that the spontaneous or activated (IL-2 activated) killing potential of patient-derived NK cells by day 30 after transplantation were lower than those of the donor-derived cells. In vitro experiments demonstrated that NK cell cytotoxicity could be retained or decreased after being treated with CSA or MTX [16,17]. Therefore, the effects of CSA and MTX on this altered cytotoxicity could not be ex-
cluded completely and should be clarified in vivo, particularly in our transplant settings.
In this paper, we emphasized the importance of the early rapid recovery of CD56bright NK cells. For the first time, we showed that the early rapid recovery of CD56brightNK cells was a good predictive sign for better survival and less occurrence of TRM after this HLA-mismatched/haploidentical HSCT without in vitro T cells depletion. Although it was originally suggested that CD56bright NK cells represent the precur-
subpopulation, and the
of the CD56
disproportion of CD56brightand CD56dimNKcells after transplantation implies an immature NK pheno-type [9,14,15,18]. Recent cumulative data indicated that CD56bright and CD56dim NK cells are phenotyp-ically, functionally, and developmentally different from NK cell subsets. Cooper et al [3] provided evidence to support the idea that CD56bright NK cells are the major cytokine-producing subset of human NK cells with unique immunoregulatory roles in vivo. Therefore, the rapid expansion of donor-derived CD56bright NK cells may have particular implication for the immune recovery owing to their immunoregulator roles and the ability to proliferate and IFN-g production upon interaction with DC and T cells.
Although the rapid recovery of T and NK cells could benefit from antitumor or anti-infection, we also noticed that the rapidly recovered T cells could became a risk factor for aGVHD after transplantation. Furthermore, the "high" ratio of T/NK cells (>1.0) increased the incidence of aGVHD. It is similar to Soiffer's report [19] that patients with high counts of CD81 T cells or low counts of CD561 NK cells during week 2 post-BMT developed high risk for aGVHD occurrence, and none of patients with low levels of CD81 cells and high levels of CD561 cells developed grades II-IV GVHD after CD6-depleted allogeneic bone marrow transplantation. Previous studies suggested that NK cells in the allograft are associated with less aGVHD occurrence after transplantation
[20-22]; therefore, the roles of reconstituted NK cells or its interaction with T cells in aGVHD still need to be further explored.
To test the possibility that the ratio of early reconstituted T and NK cells may be more significant in predicating the occurrence of cGVHD, we also evaluated the stratified ratio of T/NK cells on the incidence of cGVHD, and our result confirmed that the "high" ratio of T/NK cells (>1.0) was an independent risk factor for the high incidence of cGVHD, whereas the high number of CD31 T cells had no direct correlation. Chronic GVHD pathogenesis may be involved in autoreactive T cells that escape negative selection in the thymus, which is damaged by preconditioning or aGVHD. Th2 immune response of donor CD41 T cells released from negative selection facilitate host B cells to synthesize autoantibodies [23,24]. A previous report illustrated that NK cells could generate the suppressive effect of CD81 T cells on autoantibody synthesis of B cells through producing transforming growth factor-b (TGF-b) after optimal interaction between the NK and CD81 T cells [25]. Therefore, the relative low number of NK cells over T cells (T/NK >1.0) cannot optimally develop a suppressive effect on T cell-mediated autoantibody synthesis of B cells. Abrahamsen et al [26] demonstrated that patients with extensive cGVHD showed low NK cell counts, which suggest that a low number of NK cell counts might be a possible pathogenesis of cGVHD.
In conclusion, our results show that although a prolonged high proportional CD56bnght NK subset could be associated with the impairment of NK cytotoxity after transplantation, patients with a "high" number of CD56bright NK cells have an encouraging OS after haploidentical HSCT without ex vivo T cells depletion, which may be because of its potential immuno-regulatory roles and its ability to interaction with DC, monocyte, and T cells. Furthermore, the ratio of T/NK cells may be involved in initiating and prevention of GVHD. Therefore, further investigation on the interaction between the immune cells (NK, DC, and T cells) may help decrease the occurrence of GVHD, TRM, relapse, and increase overall survival rate.
ACKNOWLEDGMENTS
We would like to acknowledge the staff of the Peking University Institute of Hematology for their help in organizing and facilitating the collection of patient samples and thank San Francisco Edit (www.sefedit. net) for their assistance in editing this manuscript. This work was supported by the National Outstanding Young Scientist's Foundation of China (Grant No. 30725038), Hi-tech Research and Development Program of China (No. 2006AA02Z4A0), and Program for Innovative Research Team in University (IRT0702). Ying-Jun Chang performed research,
analysis, and interpretation of data, drafting the article, and final approval of the version to be published; Xiang-Yu Zhao performed research, analysis and interpretation of data, drafting the article, and final approval of the version to be published; Xiao-Jun Huang was responsible for conception and design, revising the article critically, and final approval of the version to be published. The authors reported no potential conflicts of interest.
REFERENCES
1. Cooper MA, Fehniger T, Caligiuri M. The biology of human natural killer-cell subsets. Trends Immunol. 2001;22:633-640.
2. Baume DM, Robertson MJ, Levine H, Manley TJ, Schow PW, Ritz J. Differential responses to interleukin 2 define functionally distinct subsets of human natural killer cells. Eur J Immunol. 1992;22:1-6.
3. Cooper MA, Fehniger TA, Turner SC, et al. Human natural killer cells: a unique innate immunoregulatory role for the CD56bright subset. Blood. 2001;97:3146-3151.
4. FehnigerTA, Cooper MA, Nuovo GJ, et al. CD56bright natural killer cells are present in human lymph nodes and are activated by T cell-derived IL-2: a potential new link between adaptive and innate immunity. Blood. 2003;101:3052-3057.
5. Massimo Vitale MDC, Simona Carlomagno, Chiara Romagnani, Andreas Thiel, Lorenzo Moretta, Alessandro Moretta. The small subset of CD56brightCD162 natural killer cells is selectively responsible for both cell proliferation and interferon-r production upon interaction with dendritic cells. Eur J Immunol. 2004;34: 1715-1722.
6. Kim DH, Sohn SK, Lee NY, et al. Transplantation with higher dose of natural killer cells associated with better outcomes in terms of non-relapse mortality and infectious events after allogeneic peripheral blood stem cell transplantation from HLA-matched sibling donors. Eur J Haematol. 2005;75:299-308.
7. Larghero J, Rocha V, Porcher R, et al. Association of bone marrow natural killer cell dose with neutrophil recovery and chronic graft-versus-host disease after HLA identical sibling bone marrow transplants. Br J Haematol. 2007;138:101-109.
8. Zhao XY, Huang XJ, Liu KY, Xu LP, Liu DH. Prognosis after unmanipulated HLA-haploidentical blood and marrow transplantation is correlated to the numbers of KIR ligands in recipients. Eur J Haematol. 2007;78:338-346.
9. Nguyen S, Dhedin N, Vernant J-P, et al. NK-cell reconstitution after haploidentical hematopoietic stem-cell transplantations: immaturity of NK cells and inhibitory effect of NKG2A override GvL effect. Blood. 2005;105:4135-4142.
10. Lu D-P, Dong L, Wu T, et al. Conditioning including antithy-mocyte globulin followed by unmanipulated HLA-mismatched/ haploidentical blood and marrow transplantation can achieve comparable outcomes with HLA-identical sibling transplantation. Blood. 2006;107:3065-3073.
11. HuangX-J, Han W, Xu L-P, ChenY-H, Liu D-H, LuJ. Anovel approach to human leukocyte antigen-mismatched transplantation in patients with malignant hematological disease. Chin Med J. 2004;117:1778-1785.
12. Glucksberg H, Storb R, Fefer Aea. Clinical manifestations of graft-versus-host disease in human recipients of marrow from HLA-matched sibling donors. Transplantation. 1974;18. 295-230.
13. Gooley TA, Leisenring W, Crowley J, Storer BE. Estimation of failure probabilities in the presence of competing risks: new representations of old estimators. Stat Med. 1999;18:695-706.
14. Shilling HG, McQueen KL, Cheng NW, Shizuru JA, Negrin RS, Parham P. Reconstitution of NK cell receptor repertoire following HLA-matched hematopoietic cell transplantation. Blood. 2003;101:3730-3740.
15. Chklocskaia E, Nowbakht P, Nissen C, Gratwohl A, Bargetzi M, Wodnar-Filipowicz A. Reconstitution of dendritic and natural killer-cell subsets after allogeneic stem cell transplantation: effects of endogenous flt3-ligand. Blood. 2004;103:3860-3868.
16. Nakarai T, Ueno Y, Ueno Y, Koizumi S. Paradoxical enhancement of interleukin-2-mediated cytotoxicity against K562 cells by addition of a low dose of methotrexate. Cancer Immunol Immunother. 1990;32:8-12.
17. Wang H, Grzywacz B, Sukovich D, et al. The unexpected effect of cyclosporin A on CD56+CD16- and CD56+CD16+ natural killer cell subpopulations. Blood. 2007;110:1530-1539.
18. Nagler A, Lanier L, Cwirla S, Phillips J-H. Comparative studies of human FcRIII-positive ans negative natural killer cells. JImmunol. 1989;143:3183-3191.
19. Soiffer R, Gonin R, Murray C, et al. Prediction of graft-versus-host disease by phenotypic analysis ofearlyimmune reconstitution after CD6-depleted allogeneic bone marrow transplantation. Blood. 1993;82:2216-2223.
20. Asai O, Longo D, Tian Zea. Surpression of graft-versus-host disease and amplification of graft-versus-tumor effect by activated natural killer cells after allogeneic bone marrow transplantation. J Clin Invest. 1998;101:1835-1842.
21. Ruggeri L, Capanni M, Casucci M, Volpi I, Tosti A, Perruccio Kea. Role of natural killer cell alloreactivity in HLA-mis-matched hematopoietic stem cell transplantation. Blood. 1999; 94:333-339.
22. Yamasaki S, Henzan J, Ohno Y, et al. Influence of transplanted dose of CD561 cells on development of graft-versus-host disease in patients receiving G-CSF-mobilized peripheral blood progenitor cells from HLA-identical sibling donors. Bone Marrow Transplant. 2003;32:505-510.
23. Weinberg K, Blazar B, Wagner J, et al. Factors affecting thymic function after allogeneic hematopoietic stem cell transplantation. Blood. 2001;97:1458-1466.
24. Iwasaki T. Recent advances in the treatment ofgraft-versus-host disease. Clin Med Res. 2004;2:243-252.
25. GrayJ-D, Hirokawa M, Horwitz D-A. The role of transforming growth factor b in the generation of suppression: an interaction between CD8 1 TandNKcells. JExpMed. 1994;180:1937-1942.
26. Abrahamsen I, Somme S, Heldal D, Egeland T, Kvale D, Jjennfjord G. Immune reconstitution after allogeneic stem cell transplantation: the impact of stem cell source and graft-ver-sus-host disease. Haematologica. 2005;90:86-93.
