- Email: [email protected]
The Effect of KIR Ligand Incompatibility on the Outcome of Unrelated Donor Transplantation: A Report from the Center for International Blood and Marrow Transplant Research, the European Blood and Marrow Transplant Registry, and the Dutch Registry
Biology of Blood and Marrow Transplantation 12:876-884 (2006) 䊚 2006 American Society for Blood and Marrow Transplantation 1083-8791/06/1208-0001$32.00/0 doi:10.1016/j.bbmt.2006.05.007
The Effect of KIR Ligand Incompatibility on the Outcome of Unrelated Donor Transplantation: A Report from the Center for International Blood and Marrow Transplant Research, the European Blood and Marrow Transplant Registry, and the Dutch Registry Sherif S. Farag,1 Andrea Bacigalupo,2 Mary Eapen,3 Carolyn Hurley,4 Bo Dupont,5 Michael A. Caligiuri,1 Christian Boudreau,6 Gene Nelson,7 Machteld Oudshoorn,8 Jon van Rood,8 Andrea Velardi,9 Martin Maiers,7 Michelle Setterholm,7 Dennis Confer,7 Phillip E. Posch,4 Claudio Anasetti,10 Naynesh Kamani,11 Jeffrey S. Miller,12 Daniel Weisdorf,12 Stella M. Davies13 for the KIR Study Group, Center for International Blood and Marrow Transplantation Research 1
The Ohio State University, Columbus, Ohio; 2Ospedale S. Martino, Genova, Italy; 3Statistical Center, Center for International Blood and Marrow Transplant Research, Medical College of Wisconsin, Milwaukee, Wisconsin; 4 Georgetown University School of Medicine, Washington, DC; 5Memorial Sloan-Kettering Cancer Center, New York, New York; 6University of Waterloo, Waterloo, Ontario, Canada; 7National Marrow Donor Program, Minneapolis, Minnesota; 8Leiden University Medical Center, Europdonor Foundation, Leiden, The Netherlands; 9 Policlinico di Monteluce, Perugia, Italy; 10H. Lee Moffitt Cancer Center and Research Institute, Tampa, Florida; 11 Children’s National Medical Center, Washington, DC; 12University of Minnesota, Minneapolis, Minnesota; 13 Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio Correspondence and reprint requests: Dr. Sherif S. Farag, Department of Internal Medicine, Division of Hematology and Oncology and The Comprehensive Cancer Center, The Ohio State University, A433A Starling-Loving Hall, 320 West Tenth Avenue, Columbus, OH 43210 (e-mail: [email protected]). Received February 23, 2006; accepted May 17, 2006
ABSTRACT Matching for HLA class I alleles, including HLA-C, is an important criterion for outcome of unrelated donor transplantation. However, haplotype-mismatched transplantations for myeloid malignancies, mismatched for killer immunoglobulin-like receptor (KIR) ligands in the graft-versus-host (GVH) direction, is associated with lower rates of graft-versus-host disease (GVHD), relapse, and mortality. This study investigated the effect of KIR ligand mismatching on the outcome of unrelated donor transplantation. The outcomes after 1571 unrelated donor transplantations for myeloid malignancies where donor–recipient pairs were HLA-A, -B, -C, and -DRB1 matched (n ⴝ 1004), GVH KIR ligand–mismatched (n ⴝ 137), host-versus-graft (HVG) KIR ligand–mismatched (n ⴝ 170), and HLA-B and/or –C–mismatched but KIR ligand–matched (n ⴝ 260) were compared using Cox regression models. Treatment-related mortality (TRM), treatment failure, and overall mortality were lowest after matched transplantations. Patients who received grafts from donors mismatched at the KIR ligand in the GVH or HVG direction and mismatched at HLA-B and/or C but matched at the KIR ligand had similar rates of TRM, treatment failure, and overall mortality. There were no differences in leukemia recurrence between the 4 groups. These results do not support the choice of an unrelated donor on the basis of KIR ligand mismatch determined from HLA typing. © 2006 American Society for Blood and Marrow Transplantation
KEY WORDS KIR ligand incompatibility 876
●
Leukemia
●
Unrelated donor
●
Bone marrow transplantation
877
KIR Ligand Incompatibility in Unrelated Donor Transplantation
INTRODUCTION Donor and recipient matching of major histocompatibility complex (MHC) class I alleles is an important determinant of clinical outcomes of unrelated donor hematopoietic cell transplantation (HCT), influencing the risk of graft rejection, graft-versus-host disease (GVHD), and relapse. HLA-A and -B antigen mismatches are independent risk factors for graft failure [1,2], and mismatches for HLA-A, -B, and -C are risk factors for GVHD [3]. In 1 study, HLA-C mismatch was an independent determinant of graft rejection after adjusting for other class I antigen mismatches [4]. Although specific effects of mismatch at different MHC class I 1oci on survival remain uncertain, the number of disparate loci appears to be inversely correlated with survival [5,6]. Some epitopes of MHC class I molecules, particularly HLA-C, function as ligands for killer immunoglobulin-like receptors (KIRs), which are important to the recognition of target cells by natural killer (NK) cells [7]. A single KIR recognizes determinants shared between members of a “group” of MHC class I alleles. For example, KIR2DL1 recognizes group 2 (Asn77 Lys80) HLA-C molecules, KIR2DL2/3 recognizes group 1 (Ser77 Asn80) HLA-C molecules, KIR3DL1 recognizes the HLA-Bw4 epitope, and KIR3DL2 recognizes HLA-A3 and ⫺A11 molecules [7,8]. Recognition of the relevant MHC class I molecule by a given inhibitory KIR results in inhibition of the NK cell, whereas nonrecognition leaves activation signals unopposed, promoting target cell lysis [8]. In 1 report of haplotype-mismatched, T-cell– depleted transplantation, mismatching for KIR class I epitopes in the graft-versus-host (GVH) direction was associated with a reduction in relapse, graft rejection, and GVHD in patients with high-risk acute myeloid leukemia (AML) [9,10]. In a mouse model, alloreactive NK cells demonstrated a potent in vivo antileukemic effect and eliminated host T cells and antigen-presenting cells, facilitating engraftment and preventing acute GVHD [10]. These data were not fully confirmed by a subsequent report, however [11]. Complete matching of all HLA class I (and class II) loci, including HLA-C, is generally advocated in unrelated donor HCT. However, the results observed with haplotype-mismatched transplantations, where donor– recipient pairs are mismatched at the KIR ligand epitope [9,10], call this practice into question. It is possible that by mismatching for KIR epitopes in the GVH direction, the outcome of unrelated donor HCT might be similarly improved. A recent analysis has suggested that KIR MHC class I epitope mismatching may indeed improve the outcome of unrelated donor HCT [12]. However, mismatching was associated with a worse outcome in a separate analysis of T-replete unrelated donor transplantations at a single center [13]. The purpose of this anal-
ysis was to investigate the effect of KIR ligand mismatching on the outcome of unrelated donor HCT using data reported to the National Marrow Donor Program (NMDP) and the European Bone Marrow Transplant (EBMT) and Dutch registries to determine whether there might be a benefit from selecting KIR ligandmismatched donors. METHODS Inclusion Criteria
The study population included 1571 patients who received unrelated donor HCT between January 1990 and December 1999 facilitated by the NMDP in the United States (n ⫽ 1424) or reported to the EBMT and Dutch registries (n ⫽ 147). Recipients age ⱕ 65 years with AML, myelodysplasia (MDS) with refractory anemia (RA), RA with excess blasts (RAEB) or excess blasts in transformation (RAEB-T) subtypes using the French-American-British (FAB) criteria [14,15], or chronic myeloid leukemia (CML) were eligible. All recipients received myeloablative regimens. Recipients of previous HCT, peripheral blood or cord blood grafts, or regimens containing Campath-1 antibodies were excluded. HLA Typing and KIR Ligand Matching of Donor–Recipient Pairs
High-resolution typing for HLA-A, -B, -C, and -DRB1 were available for all donor–recipient pairs. All selected donor–recipient pairs were matched for HLA-A and -DRB1 alleles. Cases were divided into those matched for MHC class I alleles (HLA-A, -B, -C) and those mismatched for HLA-B and/or -C alleles. KIR ligand status of donor–recipient pairs was determined from HLA typing; KIR typing was not performed. Donor–recipient pairs mismatched for the HLA-B and/or -C loci were subdivided into KIR ligand–matched, KIR ligand–mismatched in the hostversus-graft (HVG) direction, and KIR ligand–mismatched in the GVH direction. Cases with bidirectional mismatches of KIR ligands were included with those mismatched in the GVH direction. KIR ligand mismatch was defined by the absence of donor KIR ligand class I alleles in the recipient using an algorithm described by Ruggeri et al [10]. The KIR and MHC class I ligands considered included KIR2DL1 with group 2 HLA-C molecules, KIR2DL2/3 with group 1 HLA-C molecules, and KIR3DL1 with HLA-B molecules carrying the Bw4 epitope as defined by their predicted amino acid sequences. Definitions of Clinical Endpoints
Neutrophil recovery was defined as an absolute neutrophil count (ANC) ⱖ 500/L for 3 consecutive days and platelet recovery ⱖ 20,000/L unsupported
878
for 7 days. Grade 2– 4 acute GVHD was assigned to all patients [16], and chronic GVHD was assigned to those patients who survived 90 days or longer [17]. Treatment-related mortality (TRM) was defined as death during continuous remission. Relapse was defined as hematologic leukemia recurrence; patients who failed to achieve remission after transplantation were considered to have had recurrence at day 1. Leukemia-free survival (LFS) was defined as survival in continuous complete remission. Selection of Registry Cases
All cases from the EBMT and Dutch registries who met the eligibility criteria as described above were included. However, only a subset of transplantations facilitated by the NMDP in the United States was included, based on the availability of high-resolution HLA typing and informed consent. Between 1990 and 1999, the NMDP performed retrospective high-resolution typing for 4156 of 7594 transplantations. This subset had to have available donor and recipient pretransplantation samples for retrospective typing and is representative of all transplantations facilitated by the NMDP. The NMDP retrospectively obtained consent for study participation from surviving patients or parents/legal guardians for transplantations that it facilitated in the United States during the study period; the NMDP Institutional Review Board waived consent for patients who died before soliciting consent. To address bias introduced by the inclusion of only a proportion of surviving patients (those consenting) but all deceased patients, a sample of deceased patients was selected using a weighted randomized scheme that adjusts for overrepresentation of deceased patients in the consented cohort. This weighted randomized scheme was developed based on all survivors in the NMDP database. A logistic regression model was fit to identify the factors that predicted whether a patient had consented or not consented to use of data collected by the NMDP. This analysis found that the following factors were associated with the likelihood of a patient consenting: age at transplantation, disease type, race, sex, cytomegalovirus serologic status, and country of transplantation (United States vs non– United States). Using estimated consenting probabilities from this model based on the characteristics of dead patients, the biased coin method of randomization was performed to determine which of the dead patients likely would have consented to participate had they been alive. Thus this procedure includes the preconsented dead patients at the same probability as surviving patients who consented to participate. Approximately 13% of the surviving patients failed to consent, and 12% of the dead patients were deleted by the weighted randomized method. The above-de-
S. S. Farag et al.
scribed methods were tested several times, and on every occasion the proportion of deleted dead patients was similar. Statistical Analysis
Baseline variables were compared between the groups using the 2 statistic for categorical variables and the Kruskal-Wallis test for continuous variables. Probabilities of overall survival (OS) and LFS were calculated using the Kaplan-Meier estimator [18]. For OS, death from any cause was considered an event, and patients surviving at last follow-up were censored. For LFS, relapse or death (treatment failure) was considered an event, and patients surviving in continuous complete remission were censored at last followup. Probabilities of neutrophil and platelet recovery, acute and chronic GVHD, TRM, and relapse were calculated using the cumulative incidence function [19]. For neutrophil and platelet recovery and GVHD, death without an event was the competing event; for TRM, relapse was the competing event; and, for relapse, treatment-related death was the competing event. Adjusted probabilities of OS and LFS were estimated using the Cox proportional hazards regression model with consideration of the variables in the final multivariate models [20]. Multivariate models were constructed using stepwise forward selection, with a P value ⬍ .01 considered significant. (A significance level of .01 rather than .05 was chosen because of the multiple comparisons performed in this study.) All variables met the proportional hazards assumption. The variable for donor–recipient HLA compatibility was retained in all steps of model building. Other variables considered were recipient age, disease type and disease status at HCT, sex of the recipient and donor, conditioning regimen (irradiation vs nonirradiation-containing), GVHD prophylaxis regimen (ex vivo T-cell depletion vs T-replete), year of transplantation (1990-1995 vs 1996-1999), and the use or nonuse of antithymocyte globulin (ATG). There were no first-order interactions between the variable for donor–recipient HLA compatibility and other variables studied. Because the data were obtained from 3 independent registries, all multivariate analyses were stratified by the variable for registry source. No statistically significant center effects were noted. All reported P values are 2-sided. Analyses were performed using PROC PHREG in SAS version 8.2 (SAS Institute, Cary, NC).
RESULTS Patients
Patient, disease, and transplant characteristics for HLA-matched and mismatched transplantations ac-
879
KIR Ligand Incompatibility in Unrelated Donor Transplantation
Table 1. Patient, Disease and Transplant Characteristics of Donor–Recipient Pairs by HLA and KIR Ligand Matching
n Patient sex Male, n (%) Patient age (years), n (%) <10 10–19 20–29 30–39 40–49 50–65 D/R sex, n (%) M/M M/F F/M F/F Year of BMT, n (%) 1990–1995 1996–1999 HLA mismatches, n (%) HLA-B HLA-C HLA-B and -C Disease, n (%) AML CML MDS Disease status, n (%)‡ Early Advanced Unknown Regimen, n (%) TBI ⴞ other Bu/Cy ⴞ other GVHD prophylaxis, n (%) T-cell depletion No T-cell depletion Follow-up, months Median (range)
GVH KIR Ligand Mismatch*
HVG KIR Ligand Mismatch
HLA-Mismatch, KIR Ligand Match†
Full HLA Match
137
170
260
1004
72 (53)
102 (60)
142 (55)
564 (56)
9 10 17 40 44 17
(7) (7) (13) (29) (32) (12)
3 20 37 41 53 16
(2) (12) (22) (24) (31) (9)
16 26 54 74 65 25
(6) (10) (21) (28) (25) (10)
44 81 160 287 299 133
(4) (8) (16) (29) (30) (13)
44 35 28 30
(32) (26) (20) (22)
60 39 42 29
(35) (23) (25) (17)
85 64 57 54
(32) (24) (22) (21)
398 238 165 201
(40) (24) (16) (20)
P Value
.58 .10
.30
.51 57 (42) 80 (58)
67 (39) 103 (61)
97 (37) 163 (63)
424 (42) 580 (58)
2 (2) 80 (58) 55 (40)
3 (2) 104 (61) 63 (37)
62 (24) 134 (51) 64 (25)
46 (33) 75 (55) 16 (12)
50 (29) 92 (54) 28 (17)
77 (30) 149 (57) 34 (13)
246 (25) 584 (58) 174 (17)
69 (50) 68 (50) 0
90 (53) 80 (47) 0
126 (48) 133 (51) 1 (<1)
536 (53) 467 (47) 1 (<1)
116 (85) 21 (16)
138 (81) 32 (19)
215 (83) 45 (17)
817 (78) 187 (19)
27 (20) 110 (80)
44 (26) 126 (74)
54 (21) 206 (79)
185 (18) 819 (82)
0 0 0 .15
.56
.81
.15
68 (12–143)
69 (19–130)
59 (5–130)
60 (6–154)
R, recipient; D, donor; M, male; F, female; BMT, bone marrow transplantation; Bu, busulfan; Cy, cyclophosphamide. *HLA mismatch refers to mismatching at the HLA-B and/or -C loci. †GVH KIR ligand mismatched group included 13 cases with bidirectional KIR ligand mismatches. ‡AML and MDS patients in first complete remission, and CML patients in chronic phase (CP) at the time of transplantation were considered to have early disease; all others are classified as advanced disease.
cording to KIR ligand status are shown in Table 1. Some 64% of transplants were matched at HLA-A, -B, -C, or -DRB1; 9% were mismatched in the GVH KIR ligand direction, 11% were mismatched in the HVG direction, and 16% were mismatched at HLA-B and/or -C but matched at the KIR ligand. Thirteen of 137 cases in the GVH KIR ligand group were mismatched in both the GVH and HVG directions. Groups were similar in terms of patient, disease, and transplant characteristics. Data on the use of ATG were available for 1424 of 1571 patients. Of these, 14% received ATG. The median follow-up for survivors after HLA-matched, GVH KIR ligand–mismatched, HVG KIR ligand–mismatched, and HLA-B and/or –C–mismatched transplantations was 60, 68, 69 and 59 months, respectively.
Neutrophil and Platelet Recovery
Of the 1571 recipients, 1284 (82%) could be evaluated for neutrophil recovery and 1179 (75%) could be evaluated for platelet recovery. There were no significant differences in the probability of neutrophil or platelet recovery after HLA-matched and HLA-mismatched transplantations. The probability of neutrophil recovery at day 30 was 95% after HLA-matched transplantations, 91% after GVH KIR ligand–mismatched transplantations, 93% after HVG KIR ligand–mismatched trasnsplantations, and 93% after and HLA-mismatched but KIR ligand– matched transplantations. Corresponding probabilities of platelet recovery at day 100 were 69%, 55%, 55%, and 62%.
880
S. S. Farag et al.
Table 2. Results of Multivariate Analysis Comparing Rates of Acute and Chronic GVHD After Matched, GVH KIR-Mismatched, HVG KIR-Mismatched, and HLA-Mismatched Transplantations Hazard Ratio (99% Confidence Interval) *Grade 2–4 acute GVHD* GvH KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. HLA- and KIR-matched HvG KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. KIR-matched, HLA-mismatched Grade 3–4 acute GVHD† GvH KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. HLA- and KIR-matched HvG KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. KIR-matched, HLA-mismatched Chronic GVHD‡ GvH KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. HLA- and KIR-matched HvG KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. KIR-matched, HLA-mismatched
P Value
1.08 1.17 1.28 1.07 1.18 1.10
(0.78–1.51) (0.90–1.50) (0.95–1.73) (0.73–1.57) (0.78–1.79) (0.77–1.57)
.53 .12 .04 .63 .30 .50
1.64 1.41 1.91 0.86 1.16 1.35
(1.11–2.42) (1.01–1.96) (1.33–2.73) (0.55–1.35) (0.72–1.87) (0.88–2.08)
.001 .01 <.001 .39 .42 .07
0.72 0.89 1.01 1.24 1.39 1.13
(0.47–1.11) (0.66–1.20) (0.70–1.44) (0.76–2.01) (0.82–2.36) (0.73–1.73)
.05 .33 .97 .26 .11 .48
*In all groups the risk of grade 2– 4 acute GVHD was higher after transplantation of non–T-cell– depleted grafts relative risk [RR], 1.76; 99% confidence interval [CI], 1.3–2.31; P ⬍ .001) and irradiation-containing conditioning regimen (RR, 1.42; 99% CI; 1.09 –1.83; P ⫽ .001). †In all groups, the risk of grade 3– 4 acute GVHD was lower if transplantations were performed after 1995 (RR, 0.73; 99% CI, 0.57– 0.93; P ⫽ .001). ‡In all groups, the risk was lower if male recipients of grafts from male donors and female recipients of grafts from a male or female donor (RR, 0.65; 99% CI, 0.49 – 8.86; P ⬍ .001) and recipients of non–T-cell– depleted grafts (RR, 1.44; 99% CI, 1.08 –1.92; P ⫽ .001).
Acute and Chronic GVHD
Grade 2– 4 acute GVHD developed in 503 of 1001 recipients of HLA-matched transplantations, 74 of 137 GVH KIR ligand–mismatched transplantations, 96 of 169 HVG KIR ligand–mismatched transplantations, and 140 of 260 HLA-B and/or –C–mismatched but KIR-matched transplantations. Rates of acute grade 2– 4 GVHD were similar in all groups (Table 2). In contrast, the rate of grade 3– 4 acute GVHD was significantly higher after any mismatched transplantation than after HLA-matched transplantation (Table 2). Among recipients of mismatched transplantations, rates of grade 2– 4 or grade 3– 4 acute GVHD did not differ by KIR ligand status. Among those patients surviving for 90 days or longer, chronic GVHD developed in 476 of 956 recipients of HLA-matched transplants, in 40 of 131 recipients of GVH KIR ligand–mismatched transplants, in 62 of 159 recipients of HVG KIR ligand– mismatched transplants, and in 99 of 245 recipients of HLA-B and/or –C–mismatched but KIR-matched transplants. Risks of chronic GVHD were similar in all groups and independent of KIR ligand status (Table 2). Treatment-Related Mortality
Death from treatment-related causes occurred in 433 of 967 recipients of HLA-matched transplants, in
81 of 133 recipients of GVH KIR ligand–mismatched transplants, in 103 of 161 recipients of HVG KIR ligand–mismatched transplants, and in 133 of 251 HLA-B and/or –C–mismatched but KIR-matched transplants. TRM was lowest after HLA-matched transplantation (Table 3; Figure 1) and was significantly higher after any mismatched transplantations, but did not differ by KIR ligand status. Relapse
Leukemia recurred after transplantation in 156 of 967 recipients of HLA-matched transplants, in 24 of 133 recipients of GVH KIR ligand–mismatched transplants, in 16 of 161 recipients of HVG KIR ligand– mismatched transplants, and in 42 of 251 HLA-B and/or –C–mismatched but KIR-matched transplants. Leukemia recurrence was similar in all groups and independent of KIR ligand matching status (Table 3; Figure 2). Leukemia-Free Survival
Treatment failure (relapse or death from any cause) occurred in 589 of 967 recipients of HLAmatched transplants, in 105 of 133 recipients of GVH KIR ligand–mismatched transplants, in 119 of 161 recipients of HVG KIR ligand–mismatched transplants, and in 175 of 251 HLA-B and/or -C–mismatched but KIR-matched transplants. Treatment
881
KIR Ligand Incompatibility in Unrelated Donor Transplantation
Table 3. Results of Multivariate Analysis Comparing Rates of TRM, Relapse, Treatment Failure, and Overall Mortality After HLA- and KIR-Matched, GVH KIR-Mismatched, HvG KIR-Mismatched, and HLA-Mismatched, KIR-Matched Transplantations Hazard Ratio (99% Confidence Interval) Treatment-related mortality* GvH KIR-mismatched vs. HLA and KIR-matched KIR-matched, HLA-mismatched vs. HLA- and KIR-matched HvG KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. KIR-matched, HLA-mismatched Relapse† GvH KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. HLA- and KIR-matched HvG KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. KIR-matched, HLA-mismatched Treatment failure‡ GvH KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. HLA- and KIR-matched HvG KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. KIR-matched, HLA-mismatched Overall mortality§ GvH KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. HLA- and KIR-matched HvG KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. KIR-matched, HLA-mismatched
P Value
1.95 1.51 1.75 0.77 0.89 1.16
(1.43–2.68) (1.16–1.96) (1.32–2.32) (0.54–1.11) (0.61–1.31) (0.82–1.63)
<.001 <.001 <.001 .07 .45 .27
1.56 1.11 0.75 0.71 0.48 0.68
(0.88–2.75) (0.70–1.75) (0.38–1.47) (0.37–1.38) (0.21–1.10) (0.32–1.45)
.04 .56 .27 .19 .02 .18
1.90 1.44 1.49 0.76 0.78 1.04
(1.44–2.51) (1.15–1.80) (1.15–1.93) (0.55–1.04) (0.55–1.11) (0.76–1.41)
<.001 <.001 <.001 .02 .07 .76
1.94 1.45 1.51 0.75 0.78 1.04
(1.47–2.57) (1.17–1.83) (1.16–1.98) (0.54–1.04) (0.55–1.11) (0.76–1.43)
<.001 <.001 <.001 .02 .07 .75
*In all groups, TRM was higher in older recipients (age 20 – 49 years; RR, 2.34; 99% CI, 1.66 –3.32; P ⬍ .001; age 50 – 65 years; RR, 2.90; 99% CI, 1.91– 4.39; P ⬍ .001) and those with advanced disease (not in first CR, first CP, or RA) at transplantation (RR, 1.34; 99% CI, 1.07–1.67; P ⫽ .001). TRM was significantly lower in CML recipients compared to recipients with AML or myelodysplasia (RR, 0.76; 99% CI, 0.61– 0.95; P ⫽ .002). †In all groups, patients with advanced disease (not in first CR, first CP, or RA) at transplantation were more likely to relapse (RR, 2.73; 99% CI, 1.75– 4.26; P ⬍ .001) and those with CML and MDS, less likely to relapse compared to AML patients (RR, 0.45; 99% CI, 0.30 – 0.68; P ⬍ .001 and RR, 0.48; 99% CI, 0.29 – 0.80; P ⬍ .001, respectively). ‡In all groups, treatment failure was higher in older recipients (age 20 – 49 years: RR, 2.06; 99% CI 1.56 –2.73; P ⬍ .001; age 50 – 65 years: RR, 2.55; 99% CI, 1.81–3.60; P ⬍ .001) and those with advanced disease (not in first CR, first CP, or RA) at transplantation (RR, 1.59; 99% CI, 1.31–1.93; P ⬍ .001). Treatment failure was significantly lower in CML recipients (RR 0.70; 99% CI, 0.58 – 0.85; P ⬍ .001). §In all groups, mortality was higher in older recipients age 20 – 49 years: RR, 2.09; 99% CI, 1.57–2.78; P ⬍ .001; age 50 – 65 years; RR, 2.56; 99% CI, 1.80 –3.65; P ⬍ .001) and those with advanced disease (not in first CR, first CP or RA) at transplantation (RR, 1.60; 99% CI, 1.31–1.95; P ⬍ .001), but was significantly lower in CML recipients (RR, 0.69; 99% CI, 0.57– 0.85; P ⬍ .001).
failure was lowest after HLA-matched transplantation (Table 3; Figure 3). Treatment failure was significantly higher after mismatched transplantations and did not differ by KIR ligand status. Overall Survival
Death from any cause occurred in 582 of 1004 recipients of HLA-matched transplants, in 106 of 137 recipients of GVH KIR ligand–mismatched transplants, in 120 of 170 HVG KIR ligand–mismatched transplants, and in 177 of 260 HLA-B and/or -C–mismatched but KIR-matched transplants. Mortality was lowest after HLA-matched transplantation (Table 3; Figure 4). Mortality was significantly higher after mismatched transplantations and did not differ by KIR ligand status.
DISCUSSION The superior results associated with KIR ligand incompatibility in the GVH direction in haplotypemismatched HCT [9,10], and subsequently in 1 small study of unrelated donor HCT [12], suggest a possible clinical benefit from KIR ligand incompatibility in a wider transplantation setting. It is thought that KIR ligand incompatibility in the GVH direction results in early recovery of donor alloreactive NK cells with potent antileukemic activity. In an animal model of T-cell– depleted grafts, such cells are associated with eradication of leukemia and of host antigen-presenting cells, preventing GVHD even in the setting of significant histoincompatibility [10]. The Perugia group has translated these data into the clinical setting
882
S. S. Farag et al.
Figure 1. Cumulative incidence of TRM by donor–recipient HLA and KIR ligand matching. The 5-year rates were 45% after HLAmatched and KIR ligand–matched transplantations (n ⫽ 967), 62% after GVH KIR ligand–mismatched transplantations (n ⫽ 133), 54% after HLA-B and/or –C–mismatched but KIR-matched transplantations (n ⫽ 251), and 65% after HVG KIR ligand–mismatched transplantations (n ⫽ 161) (P ⬍ .001).
Figure 3. Probability of LFS by donor–recipient HLA and KIR ligand matching. The 5-year rates were 39% after HLA-matched and KIR ligand–matched transplantations (n ⫽ 967), 20% after GVH KIR ligand–mismatched transplantations (n ⫽ 133), 29% after HLA-B and/or –C–mismatched but KIR-matched transplantations (n ⫽ 251), and 26% after HVG KIR ligand–mismatched transplantations (n ⫽ 161) (P ⬍ .001).
of haploidentical transplantation [9,10]. In the context of an aggressively T-cell– depleted haploidentical graft with a high stem cell dose and no posttransplantation immune suppression, a marked reduction in relapse and GVHD and excellent engraftment was demonstrated in very– high-risk adults with AML [9,10]. The results of our study contrast with these observations, because we did not observe any beneficial effect of KIR ligand incompatibility in the GVH direction on any of the major clinical endpoints of unrelated donor HCT. Patients receiving HLAmatched transplants had significantly lower TRM, relapse, treatment failure, and overall mortality than patients receiving transplants from GVH KIR ligand– incompatible donors. Furthermore, we could not demonstrate any significant benefit of KIR ligand incompatibility in either the GVH or HVG direction on major clinical endpoints compared with HLA-mis-
matched transplantations without KIR ligand mismatch. Studies investigating the significance of KIR ligand incompatibility in mismatched transplantation have yielded mixed and conflicting results. Although our results using the largest study population to date are consistent with previous reports showing no significant benefit of GVH KIR ligand mismatching in unrelated donor HCT [13,21,22], other studies have reported a variable reduction in relapse [12,23] and improvement in OS [12], although none has reported protection from GVHD. In an additional study, KIR ligand incompatibility was associated with inferior survival [24]. Differences in sample size, transplantation techniques, and methods of determining KIR ligand incompatibility may explain the differences among the current study and previous reports. Thus far, most reports are limited to single institutions, whereas the
Figure 2. Cumulative incidence rates of relapse by donor–recipient HLA and KIR ligand matching. The 5-year rates were 16% after HLA-matched and KIR ligand–matched transplantations (n ⫽ 967), 18% after GVH KIR ligand–mismatched transplantations (n ⫽ 133), 17% after HLA-B and/or –C–mismatched but KIR-matched transplantations (n ⫽ 251), and 10% after HVG KIR ligand– mismatched transplantations (n ⫽ 161) (P ⫽ not significant).
Figure 4. Probability of OS by donor–recipient HLA and KIR ligand matching. The 5-year rates were 42% after HLA-matched and KIR ligand–matched transplantations (n ⫽ 967), 24% after GVH KIR ligand–mismatched transplantations (n ⫽ 133), 31% after HLA-B and/or –C–mismatched but KIR-matched transplantations (n ⫽ 251), and 30% after HVG KIR ligand–mismatched transplantations (n ⫽ 161) (P ⬍ .001).
883
KIR Ligand Incompatibility in Unrelated Donor Transplantation
current analysis includes more than 1500 patients. In addition, the beneficial effect of GVH KIR ligand mismatching was originally described in haplotypemismatched transplantations performed using a high CD34⫹ cell dose after extensive T-cell depletion of the graft to prevent GVHD without additional posttransplantation immunosuppression [9,10]. In contrast, almost 80% of patients in our study received unmanipulated grafts. Furthermore, although the doses of T cells infused are not available, it is likely that T-cell depletion was only partial, because most methods used (eg, use of T-lymphocyte–specific antibodies and complement, elutriation) are likely to have achieved only a 2–3 log depletion, compared with the more extensive (4 –5 log) depletion of T cells performed in the haplotype-mismatched transplantations. In the current analysis, the use of T-depleted grafts (ex vivo) or ATG (resulting in some in vivo T-cell depletion) failed to demonstrate an effect of KIR ligand mismatching on transplantation outcome. This is consistent with recent data showing that T-cell alloreactivity dominates over NK cell alloreactivity in minimally T-cell– depleted HLA-nonidentical transplantations [19,21]. Furthermore, the presence of significant numbers of T cells in the graft may affect NK cell receptor acquisition after transplantation [25]. A comparison of NK cell receptor expression of baseline recipient and donor-derived engrafting NK cells at 100 days after volunteer unrelated donor (VUD) transplantation showed diminished reconstitution of KIR expression with increased expression of the activating receptor NKG2D in Treplete transplantations compared with T-cell– depleted transplantations [25]. This was also associated with a higher proportion of engrafted NK cells secreting interferon-␥ in response to interleukin (IL)-12 and IL-18 [25]. These observations are consistent with our observation of better outcomes with better-matched grafts, because any effect of KIR ligand incompatibility may have been masked by the effect of alloreactive T cells and/or of posttransplantation immunosuppression. It is important to note that we did not use KIR genotyping to classify our cases, but instead relied entirely on HLA typing to determine KIR ligand incompatibility and hence NK cell alloreactivity [9,13]. Central to this algorithm is the principle that NK cell alloreactivity occurs when the HLA genotype of the donor includes a KIR epitope that is not part of HLA genotype of the recipient cells, with the assumption that the donor’s NK cells express inhibitory KIR for the donor’s own HLA ligands [8]. Recent evidence indicates that this assumption may not always be correct, however. KIR and HLA genotypes segregate independently, with the expressed KIR repertoire regulated by KIR genotype and not by HLA genotype [26-29]. In a significant proportion of cases, therefore,
individuals may lack KIR for their HLA ligands, and vice versa [30]. Indeed, in a recent study of T-cell– depleted haplotype-mismatched transplantations, KIRdriven alloreactivity was better predicted if donor KIR genotype was considered along with the HLA type of the recipient [31]. Relapses were significantly less frequent, and survival improved in cases where the recipient was lacking the HLA ligand for the donorinhibitory KIR as assessed by KIR typing [31]. More recently, this “missing ligand” algorithm was also shown to be strongly predictive of outcome even in T-cell– depleted HLA-matched sibling donor transplantations for AML and MDS, where in this setting ⬎ 60% of HLA-matched recipients were missing 1 or more ligands for their own and their donor’s inhibitory KIR [32]. Finally, a further limitation imposed by the lack of KIR genotyping is the potential confounding effect of activating KIR, which could not be examined in our study but may be important in determining outcome [33-36]. In summary, the results of the current study of unrelated donor HCT show that HLA mismatch is associated with higher TRM, treatment failure, and overall mortality regardless of the presence or absence of KIR ligand mismatch. Furthermore, no advantage of GVH KIR ligand incompatibility for any of the clinical endpoints examined was observed compared with other HLA-mismatched VUD transplantations. Accordingly, our results do not support the choice of unrelated donor HCT on the basis of KIR ligand mismatching using HLA typing alone, even when a HLA-B or -C mismatch is necessary.
ACKNOWLEDGMENTS This work was supported by Public Health Service grant U24-CA76518-08 from the National Cancer Institute, National Institute of Allergy and Infectious Diseases, and National Heart, Lung, and Blood Institute.
REFERENCES 1. Davies SM, Kollman C, Anasetti C, et al. Engraftment and survival after unrelated-donor bone marrow transplantation: a report from the National Marrow Donor Program. Blood. 2000; 96:4096-4102. 2. McGlave PB, Shu XO, Wen W, et al. Unrelated donor marrow transplantation for chronic myelogenous leukemia: 9 years’ experience of the National Marrow Donor Program. Blood. 2000;95:2219-2225. 3. Sasazuki T, Juji T, Morishima Y, et al. Effect of matching of class I HLA alleles on clinical outcome after transplantation of hematopoietic stem cells from an unrelated donor. N Engl J Med. 1998;339:1177-1185. 4. Petersdorf EW, Longton GM, Anasetti C, et al. Association of
884
5.
6.
7. 8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18. 19.
20. 21.
S. S. Farag et al.
HLA-C disparity with graft failure after marrow transplantation from unrelated donors. Blood. 1997;89:1818-1823. Szydlo R, Goldman JM, Klein JP, et al. Results of allogeneic bone marrow transplants for leukemia using donors other than HLA-identical siblings. J Clin Oncol. 1997;15:1767-1777. Flomemberg N, Baxter-Lowe LA, Confer D, et al. Impact of HLA class I and class II high-resolution matching on outcomes of unrelated donor bone marrow transplantation: HLA-C mismatching is associated with a strong adverse effect on transplantation outcome. Blood. 2004;104:1923-1930. Lanier LL. NK cell receptors. Annu Rev Immunol. 1998;16:359393. Farag SS, Fehniger TA, Ruggeri L, et al. Natural killer cell receptors: new biology and insights into the graft-versus-leukemia effect. Blood. 2002;100:1935-1947. Ruggeri L, Capanni M, Casucci M, et al. Role of natural killer cell alloreactivity in HLA-mismatched hematopoietic stem cell transplantation. Blood. 1999;94:333-339. Ruggeri L, Capanni M, Urbani E, et al. Effectiveness of donor natural killer cell alloreactivity in mismatched hematopoietic transplants. Science. 2002;295:2097-2100. Aversa F, Terenzi A, Tabilio A, et al. Full haplotype-mismatched hematopoietic stem-cell transplantation: a phase II study in patients with acute leukemia at high risk of relapse. J Clin Oncol. 2005;23:3447-3457. Geibel S, Locatelli F, Maccario R, et al. Survival advantage with KIR ligand incompatibility in unrelated donor transplantation. Blood. 2003;102:814-819. Davies SM, Ruggeri L, DeFor T, et al. Evaluation of KIR ligand incompatibility in mismatched unrelated donor hematopoietic transplants: killer immunoglobulin-like receptor. Blood. 2002;100:3825-3827. Bennett JM, Catovsky D, Daniel MT, et al. Proposed revised criteria for the classification of acute myeloid leukemia: a report of the French-American-British Cooperative Group. Ann Intern Med. 1985;103:620-625. Bennett JM, Catovsky D, Daniel MT, et al. Proposals for the classification of the myelodysplastic syndromes. Br J Haematol. 1982;51:189-199. Przepiorka D, Weisdorf D, Martin P, et al. 1994 consensus conference on acute GVHD grading chronic graft-versus-host syndrome in man: a long-term clinicopathologic study of 20 Seattle patients. Bone Marrow Transplant. 1995;15:825-828. Shulman HM, Sullivan KM, Weiden PL, et al. Chronic graftversus-host syndrome in man: a long-term clinicopathologic study of 20 Seattle patients. Am J Med. 1980;69:204-217. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1998;16:1141-1154. Gooley TA, Leisenring W, Crowley J, et al. Estimation of failure probabilities in the presence of competing risks: new representations of old estimators. Stat Med. 1999;18:695-706. Cox DR. Regression models and life tables (with discussion). J R Stat Soc B. 1972;34:187-220. Bishara A, De Santis D, Witt CC, et al. The beneficial role of inhibitory KIR genes of HLA class I NK epitopes in haploidentically mismatched stem cell allografts may be masked by residual donor-alloreactive T cells causing GVHD. Tissue Antigens. 2004;63:204-211.
22. Lowe EJ, Turner V, Handgretinger R, et al. T-cell alloreactivity dominates natural killer cell alloreactivity in minimally T-cell– depleted HLA-nonidentical paediatric bone marrow transplantation. Br J Haematol. 2003;123:323-326. 23. Beelen DW, Ottinger HD, Ferencik S, et al. Genotypic inhibitory killer immunoglobulin-like receptor ligand incompatibility enhances the long-term antileukemic effect of unmodified allogeneic hematopoietic stem cell transplantation in patients with myeloid leukemias. Blood. 2005;105:2594-2600. 24. Schaffer M, Malmberg KJ, Ringden O, et al. Increased infection-related mortality in KIR ligand-mismatched unrelated allogeneic hematopoietic stem-cell transplantation. Transplantation. 2004;78:1081-1085. 25. Cooley S, McCullar V, Wangen R, et al. KIR reconstitution is altered by T cells in the graft and correlates with clinical outcomes after unrelated donor transplantation. Blood. 2005;106:4370-4376. 26. Gumperz JE, Valiante NM, Parham P, et al. Heterogeneous phenotypes of expression of the NKB1 natural killer cell class I receptor among individuals of different human histocompatibility leukocyte antigens types appear genetically regulated, but not linked to major histocompatibililty complex haplotype. J Exp Med. 1996;183:1817-1827. 27. Frohn C, Schlenke P, Kirchner H. The repertoire of HLACw–specific NK cell receptors CD158 a/b (EB6 and GL183) in individuals with different HLA phenotypes. Immunology. 1997; 92:567-570. 28. Vilches C, Parham P. KIR: diverse, rapidly evolving receptors of innate and adaptive immunity. Ann Rev Immunol. 2002;20: 217-251. 29. Parham P, McQueen KL. Alloreactive killer cells: hindrance and help for haematopoietic transplants. Nat Rev Immunol. 2003; 3:108-122. 30. Dupont B, Hsu KC. Inhibitory killer Ig-like receptor genes and human leukocyte antigen class I ligands in haematopoietic stem cell transplantation. Curr Opin Immunol. 2004;16:634-643. 31. Leung W, Iyengar R, Turner V, et al. Determinants of antileukemia effects of allogeneic NK cells. J Immunol. 2004;172: 644-650. 32. Hsu KC, Keever-Taylor CA, Wilton A, et al. Improved outcome in HLA-identical sibling hematopoietic stem cell transplantation for acute myelogenous leukemia predicted by KIR and HLA genotypes. Blood. 2005;105:4878-4884. 33. Gagne K. Relevance of KIR gene polymorphisms in bone marrow transplantation outcome. Hum Immunol. 2002;63:271280. 34. Bishara A, De Santis D, Witt CC, et al. The beneficial role of inhibitory KIR genes of HLA class I NK epitopes in haploidentically mismatched stem cell allografts may be masked by residual donor-alloreactive T cells causing GVHD. Tissue Antigens. 2004;63:204-211. 35. De Santis D, Bishara A, Witt CC, et al. Natural killer cell HLA-C epitopes and killer cell immunoglobulin-like receptors both influence outcome of mismatched unrelated donor bone marrow transplants. Tissue Antigens. 2005;65:519-528. 36. Cook MA, Milligan DW, Fegan CD, et al. The impact of donor KIR and patient HLA-C genotypes on outcome following HLA-identical sibling hematopoietic stem cell transplantation for myeloid leukemia. Blood. 2004;103:1521-1526.
The Effect of KIR Ligand Incompatibility on the Outcome of Unrelated Donor Transplantation: A Report from the Center for International Blood and Marrow Transplant Research, the European Blood and Marrow Transplant Registry, and the Dutch Registry Sherif S. Farag,1 Andrea Bacigalupo,2 Mary Eapen,3 Carolyn Hurley,4 Bo Dupont,5 Michael A. Caligiuri,1 Christian Boudreau,6 Gene Nelson,7 Machteld Oudshoorn,8 Jon van Rood,8 Andrea Velardi,9 Martin Maiers,7 Michelle Setterholm,7 Dennis Confer,7 Phillip E. Posch,4 Claudio Anasetti,10 Naynesh Kamani,11 Jeffrey S. Miller,12 Daniel Weisdorf,12 Stella M. Davies13 for the KIR Study Group, Center for International Blood and Marrow Transplantation Research 1
The Ohio State University, Columbus, Ohio; 2Ospedale S. Martino, Genova, Italy; 3Statistical Center, Center for International Blood and Marrow Transplant Research, Medical College of Wisconsin, Milwaukee, Wisconsin; 4 Georgetown University School of Medicine, Washington, DC; 5Memorial Sloan-Kettering Cancer Center, New York, New York; 6University of Waterloo, Waterloo, Ontario, Canada; 7National Marrow Donor Program, Minneapolis, Minnesota; 8Leiden University Medical Center, Europdonor Foundation, Leiden, The Netherlands; 9 Policlinico di Monteluce, Perugia, Italy; 10H. Lee Moffitt Cancer Center and Research Institute, Tampa, Florida; 11 Children’s National Medical Center, Washington, DC; 12University of Minnesota, Minneapolis, Minnesota; 13 Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio Correspondence and reprint requests: Dr. Sherif S. Farag, Department of Internal Medicine, Division of Hematology and Oncology and The Comprehensive Cancer Center, The Ohio State University, A433A Starling-Loving Hall, 320 West Tenth Avenue, Columbus, OH 43210 (e-mail: [email protected]). Received February 23, 2006; accepted May 17, 2006
ABSTRACT Matching for HLA class I alleles, including HLA-C, is an important criterion for outcome of unrelated donor transplantation. However, haplotype-mismatched transplantations for myeloid malignancies, mismatched for killer immunoglobulin-like receptor (KIR) ligands in the graft-versus-host (GVH) direction, is associated with lower rates of graft-versus-host disease (GVHD), relapse, and mortality. This study investigated the effect of KIR ligand mismatching on the outcome of unrelated donor transplantation. The outcomes after 1571 unrelated donor transplantations for myeloid malignancies where donor–recipient pairs were HLA-A, -B, -C, and -DRB1 matched (n ⴝ 1004), GVH KIR ligand–mismatched (n ⴝ 137), host-versus-graft (HVG) KIR ligand–mismatched (n ⴝ 170), and HLA-B and/or –C–mismatched but KIR ligand–matched (n ⴝ 260) were compared using Cox regression models. Treatment-related mortality (TRM), treatment failure, and overall mortality were lowest after matched transplantations. Patients who received grafts from donors mismatched at the KIR ligand in the GVH or HVG direction and mismatched at HLA-B and/or C but matched at the KIR ligand had similar rates of TRM, treatment failure, and overall mortality. There were no differences in leukemia recurrence between the 4 groups. These results do not support the choice of an unrelated donor on the basis of KIR ligand mismatch determined from HLA typing. © 2006 American Society for Blood and Marrow Transplantation
KEY WORDS KIR ligand incompatibility 876
●
Leukemia
●
Unrelated donor
●
Bone marrow transplantation
877
KIR Ligand Incompatibility in Unrelated Donor Transplantation
INTRODUCTION Donor and recipient matching of major histocompatibility complex (MHC) class I alleles is an important determinant of clinical outcomes of unrelated donor hematopoietic cell transplantation (HCT), influencing the risk of graft rejection, graft-versus-host disease (GVHD), and relapse. HLA-A and -B antigen mismatches are independent risk factors for graft failure [1,2], and mismatches for HLA-A, -B, and -C are risk factors for GVHD [3]. In 1 study, HLA-C mismatch was an independent determinant of graft rejection after adjusting for other class I antigen mismatches [4]. Although specific effects of mismatch at different MHC class I 1oci on survival remain uncertain, the number of disparate loci appears to be inversely correlated with survival [5,6]. Some epitopes of MHC class I molecules, particularly HLA-C, function as ligands for killer immunoglobulin-like receptors (KIRs), which are important to the recognition of target cells by natural killer (NK) cells [7]. A single KIR recognizes determinants shared between members of a “group” of MHC class I alleles. For example, KIR2DL1 recognizes group 2 (Asn77 Lys80) HLA-C molecules, KIR2DL2/3 recognizes group 1 (Ser77 Asn80) HLA-C molecules, KIR3DL1 recognizes the HLA-Bw4 epitope, and KIR3DL2 recognizes HLA-A3 and ⫺A11 molecules [7,8]. Recognition of the relevant MHC class I molecule by a given inhibitory KIR results in inhibition of the NK cell, whereas nonrecognition leaves activation signals unopposed, promoting target cell lysis [8]. In 1 report of haplotype-mismatched, T-cell– depleted transplantation, mismatching for KIR class I epitopes in the graft-versus-host (GVH) direction was associated with a reduction in relapse, graft rejection, and GVHD in patients with high-risk acute myeloid leukemia (AML) [9,10]. In a mouse model, alloreactive NK cells demonstrated a potent in vivo antileukemic effect and eliminated host T cells and antigen-presenting cells, facilitating engraftment and preventing acute GVHD [10]. These data were not fully confirmed by a subsequent report, however [11]. Complete matching of all HLA class I (and class II) loci, including HLA-C, is generally advocated in unrelated donor HCT. However, the results observed with haplotype-mismatched transplantations, where donor– recipient pairs are mismatched at the KIR ligand epitope [9,10], call this practice into question. It is possible that by mismatching for KIR epitopes in the GVH direction, the outcome of unrelated donor HCT might be similarly improved. A recent analysis has suggested that KIR MHC class I epitope mismatching may indeed improve the outcome of unrelated donor HCT [12]. However, mismatching was associated with a worse outcome in a separate analysis of T-replete unrelated donor transplantations at a single center [13]. The purpose of this anal-
ysis was to investigate the effect of KIR ligand mismatching on the outcome of unrelated donor HCT using data reported to the National Marrow Donor Program (NMDP) and the European Bone Marrow Transplant (EBMT) and Dutch registries to determine whether there might be a benefit from selecting KIR ligandmismatched donors. METHODS Inclusion Criteria
The study population included 1571 patients who received unrelated donor HCT between January 1990 and December 1999 facilitated by the NMDP in the United States (n ⫽ 1424) or reported to the EBMT and Dutch registries (n ⫽ 147). Recipients age ⱕ 65 years with AML, myelodysplasia (MDS) with refractory anemia (RA), RA with excess blasts (RAEB) or excess blasts in transformation (RAEB-T) subtypes using the French-American-British (FAB) criteria [14,15], or chronic myeloid leukemia (CML) were eligible. All recipients received myeloablative regimens. Recipients of previous HCT, peripheral blood or cord blood grafts, or regimens containing Campath-1 antibodies were excluded. HLA Typing and KIR Ligand Matching of Donor–Recipient Pairs
High-resolution typing for HLA-A, -B, -C, and -DRB1 were available for all donor–recipient pairs. All selected donor–recipient pairs were matched for HLA-A and -DRB1 alleles. Cases were divided into those matched for MHC class I alleles (HLA-A, -B, -C) and those mismatched for HLA-B and/or -C alleles. KIR ligand status of donor–recipient pairs was determined from HLA typing; KIR typing was not performed. Donor–recipient pairs mismatched for the HLA-B and/or -C loci were subdivided into KIR ligand–matched, KIR ligand–mismatched in the hostversus-graft (HVG) direction, and KIR ligand–mismatched in the GVH direction. Cases with bidirectional mismatches of KIR ligands were included with those mismatched in the GVH direction. KIR ligand mismatch was defined by the absence of donor KIR ligand class I alleles in the recipient using an algorithm described by Ruggeri et al [10]. The KIR and MHC class I ligands considered included KIR2DL1 with group 2 HLA-C molecules, KIR2DL2/3 with group 1 HLA-C molecules, and KIR3DL1 with HLA-B molecules carrying the Bw4 epitope as defined by their predicted amino acid sequences. Definitions of Clinical Endpoints
Neutrophil recovery was defined as an absolute neutrophil count (ANC) ⱖ 500/L for 3 consecutive days and platelet recovery ⱖ 20,000/L unsupported
878
for 7 days. Grade 2– 4 acute GVHD was assigned to all patients [16], and chronic GVHD was assigned to those patients who survived 90 days or longer [17]. Treatment-related mortality (TRM) was defined as death during continuous remission. Relapse was defined as hematologic leukemia recurrence; patients who failed to achieve remission after transplantation were considered to have had recurrence at day 1. Leukemia-free survival (LFS) was defined as survival in continuous complete remission. Selection of Registry Cases
All cases from the EBMT and Dutch registries who met the eligibility criteria as described above were included. However, only a subset of transplantations facilitated by the NMDP in the United States was included, based on the availability of high-resolution HLA typing and informed consent. Between 1990 and 1999, the NMDP performed retrospective high-resolution typing for 4156 of 7594 transplantations. This subset had to have available donor and recipient pretransplantation samples for retrospective typing and is representative of all transplantations facilitated by the NMDP. The NMDP retrospectively obtained consent for study participation from surviving patients or parents/legal guardians for transplantations that it facilitated in the United States during the study period; the NMDP Institutional Review Board waived consent for patients who died before soliciting consent. To address bias introduced by the inclusion of only a proportion of surviving patients (those consenting) but all deceased patients, a sample of deceased patients was selected using a weighted randomized scheme that adjusts for overrepresentation of deceased patients in the consented cohort. This weighted randomized scheme was developed based on all survivors in the NMDP database. A logistic regression model was fit to identify the factors that predicted whether a patient had consented or not consented to use of data collected by the NMDP. This analysis found that the following factors were associated with the likelihood of a patient consenting: age at transplantation, disease type, race, sex, cytomegalovirus serologic status, and country of transplantation (United States vs non– United States). Using estimated consenting probabilities from this model based on the characteristics of dead patients, the biased coin method of randomization was performed to determine which of the dead patients likely would have consented to participate had they been alive. Thus this procedure includes the preconsented dead patients at the same probability as surviving patients who consented to participate. Approximately 13% of the surviving patients failed to consent, and 12% of the dead patients were deleted by the weighted randomized method. The above-de-
S. S. Farag et al.
scribed methods were tested several times, and on every occasion the proportion of deleted dead patients was similar. Statistical Analysis
Baseline variables were compared between the groups using the 2 statistic for categorical variables and the Kruskal-Wallis test for continuous variables. Probabilities of overall survival (OS) and LFS were calculated using the Kaplan-Meier estimator [18]. For OS, death from any cause was considered an event, and patients surviving at last follow-up were censored. For LFS, relapse or death (treatment failure) was considered an event, and patients surviving in continuous complete remission were censored at last followup. Probabilities of neutrophil and platelet recovery, acute and chronic GVHD, TRM, and relapse were calculated using the cumulative incidence function [19]. For neutrophil and platelet recovery and GVHD, death without an event was the competing event; for TRM, relapse was the competing event; and, for relapse, treatment-related death was the competing event. Adjusted probabilities of OS and LFS were estimated using the Cox proportional hazards regression model with consideration of the variables in the final multivariate models [20]. Multivariate models were constructed using stepwise forward selection, with a P value ⬍ .01 considered significant. (A significance level of .01 rather than .05 was chosen because of the multiple comparisons performed in this study.) All variables met the proportional hazards assumption. The variable for donor–recipient HLA compatibility was retained in all steps of model building. Other variables considered were recipient age, disease type and disease status at HCT, sex of the recipient and donor, conditioning regimen (irradiation vs nonirradiation-containing), GVHD prophylaxis regimen (ex vivo T-cell depletion vs T-replete), year of transplantation (1990-1995 vs 1996-1999), and the use or nonuse of antithymocyte globulin (ATG). There were no first-order interactions between the variable for donor–recipient HLA compatibility and other variables studied. Because the data were obtained from 3 independent registries, all multivariate analyses were stratified by the variable for registry source. No statistically significant center effects were noted. All reported P values are 2-sided. Analyses were performed using PROC PHREG in SAS version 8.2 (SAS Institute, Cary, NC).
RESULTS Patients
Patient, disease, and transplant characteristics for HLA-matched and mismatched transplantations ac-
879
KIR Ligand Incompatibility in Unrelated Donor Transplantation
Table 1. Patient, Disease and Transplant Characteristics of Donor–Recipient Pairs by HLA and KIR Ligand Matching
n Patient sex Male, n (%) Patient age (years), n (%) <10 10–19 20–29 30–39 40–49 50–65 D/R sex, n (%) M/M M/F F/M F/F Year of BMT, n (%) 1990–1995 1996–1999 HLA mismatches, n (%) HLA-B HLA-C HLA-B and -C Disease, n (%) AML CML MDS Disease status, n (%)‡ Early Advanced Unknown Regimen, n (%) TBI ⴞ other Bu/Cy ⴞ other GVHD prophylaxis, n (%) T-cell depletion No T-cell depletion Follow-up, months Median (range)
GVH KIR Ligand Mismatch*
HVG KIR Ligand Mismatch
HLA-Mismatch, KIR Ligand Match†
Full HLA Match
137
170
260
1004
72 (53)
102 (60)
142 (55)
564 (56)
9 10 17 40 44 17
(7) (7) (13) (29) (32) (12)
3 20 37 41 53 16
(2) (12) (22) (24) (31) (9)
16 26 54 74 65 25
(6) (10) (21) (28) (25) (10)
44 81 160 287 299 133
(4) (8) (16) (29) (30) (13)
44 35 28 30
(32) (26) (20) (22)
60 39 42 29
(35) (23) (25) (17)
85 64 57 54
(32) (24) (22) (21)
398 238 165 201
(40) (24) (16) (20)
P Value
.58 .10
.30
.51 57 (42) 80 (58)
67 (39) 103 (61)
97 (37) 163 (63)
424 (42) 580 (58)
2 (2) 80 (58) 55 (40)
3 (2) 104 (61) 63 (37)
62 (24) 134 (51) 64 (25)
46 (33) 75 (55) 16 (12)
50 (29) 92 (54) 28 (17)
77 (30) 149 (57) 34 (13)
246 (25) 584 (58) 174 (17)
69 (50) 68 (50) 0
90 (53) 80 (47) 0
126 (48) 133 (51) 1 (<1)
536 (53) 467 (47) 1 (<1)
116 (85) 21 (16)
138 (81) 32 (19)
215 (83) 45 (17)
817 (78) 187 (19)
27 (20) 110 (80)
44 (26) 126 (74)
54 (21) 206 (79)
185 (18) 819 (82)
0 0 0 .15
.56
.81
.15
68 (12–143)
69 (19–130)
59 (5–130)
60 (6–154)
R, recipient; D, donor; M, male; F, female; BMT, bone marrow transplantation; Bu, busulfan; Cy, cyclophosphamide. *HLA mismatch refers to mismatching at the HLA-B and/or -C loci. †GVH KIR ligand mismatched group included 13 cases with bidirectional KIR ligand mismatches. ‡AML and MDS patients in first complete remission, and CML patients in chronic phase (CP) at the time of transplantation were considered to have early disease; all others are classified as advanced disease.
cording to KIR ligand status are shown in Table 1. Some 64% of transplants were matched at HLA-A, -B, -C, or -DRB1; 9% were mismatched in the GVH KIR ligand direction, 11% were mismatched in the HVG direction, and 16% were mismatched at HLA-B and/or -C but matched at the KIR ligand. Thirteen of 137 cases in the GVH KIR ligand group were mismatched in both the GVH and HVG directions. Groups were similar in terms of patient, disease, and transplant characteristics. Data on the use of ATG were available for 1424 of 1571 patients. Of these, 14% received ATG. The median follow-up for survivors after HLA-matched, GVH KIR ligand–mismatched, HVG KIR ligand–mismatched, and HLA-B and/or –C–mismatched transplantations was 60, 68, 69 and 59 months, respectively.
Neutrophil and Platelet Recovery
Of the 1571 recipients, 1284 (82%) could be evaluated for neutrophil recovery and 1179 (75%) could be evaluated for platelet recovery. There were no significant differences in the probability of neutrophil or platelet recovery after HLA-matched and HLA-mismatched transplantations. The probability of neutrophil recovery at day 30 was 95% after HLA-matched transplantations, 91% after GVH KIR ligand–mismatched transplantations, 93% after HVG KIR ligand–mismatched trasnsplantations, and 93% after and HLA-mismatched but KIR ligand– matched transplantations. Corresponding probabilities of platelet recovery at day 100 were 69%, 55%, 55%, and 62%.
880
S. S. Farag et al.
Table 2. Results of Multivariate Analysis Comparing Rates of Acute and Chronic GVHD After Matched, GVH KIR-Mismatched, HVG KIR-Mismatched, and HLA-Mismatched Transplantations Hazard Ratio (99% Confidence Interval) *Grade 2–4 acute GVHD* GvH KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. HLA- and KIR-matched HvG KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. KIR-matched, HLA-mismatched Grade 3–4 acute GVHD† GvH KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. HLA- and KIR-matched HvG KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. KIR-matched, HLA-mismatched Chronic GVHD‡ GvH KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. HLA- and KIR-matched HvG KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. KIR-matched, HLA-mismatched
P Value
1.08 1.17 1.28 1.07 1.18 1.10
(0.78–1.51) (0.90–1.50) (0.95–1.73) (0.73–1.57) (0.78–1.79) (0.77–1.57)
.53 .12 .04 .63 .30 .50
1.64 1.41 1.91 0.86 1.16 1.35
(1.11–2.42) (1.01–1.96) (1.33–2.73) (0.55–1.35) (0.72–1.87) (0.88–2.08)
.001 .01 <.001 .39 .42 .07
0.72 0.89 1.01 1.24 1.39 1.13
(0.47–1.11) (0.66–1.20) (0.70–1.44) (0.76–2.01) (0.82–2.36) (0.73–1.73)
.05 .33 .97 .26 .11 .48
*In all groups the risk of grade 2– 4 acute GVHD was higher after transplantation of non–T-cell– depleted grafts relative risk [RR], 1.76; 99% confidence interval [CI], 1.3–2.31; P ⬍ .001) and irradiation-containing conditioning regimen (RR, 1.42; 99% CI; 1.09 –1.83; P ⫽ .001). †In all groups, the risk of grade 3– 4 acute GVHD was lower if transplantations were performed after 1995 (RR, 0.73; 99% CI, 0.57– 0.93; P ⫽ .001). ‡In all groups, the risk was lower if male recipients of grafts from male donors and female recipients of grafts from a male or female donor (RR, 0.65; 99% CI, 0.49 – 8.86; P ⬍ .001) and recipients of non–T-cell– depleted grafts (RR, 1.44; 99% CI, 1.08 –1.92; P ⫽ .001).
Acute and Chronic GVHD
Grade 2– 4 acute GVHD developed in 503 of 1001 recipients of HLA-matched transplantations, 74 of 137 GVH KIR ligand–mismatched transplantations, 96 of 169 HVG KIR ligand–mismatched transplantations, and 140 of 260 HLA-B and/or –C–mismatched but KIR-matched transplantations. Rates of acute grade 2– 4 GVHD were similar in all groups (Table 2). In contrast, the rate of grade 3– 4 acute GVHD was significantly higher after any mismatched transplantation than after HLA-matched transplantation (Table 2). Among recipients of mismatched transplantations, rates of grade 2– 4 or grade 3– 4 acute GVHD did not differ by KIR ligand status. Among those patients surviving for 90 days or longer, chronic GVHD developed in 476 of 956 recipients of HLA-matched transplants, in 40 of 131 recipients of GVH KIR ligand–mismatched transplants, in 62 of 159 recipients of HVG KIR ligand– mismatched transplants, and in 99 of 245 recipients of HLA-B and/or –C–mismatched but KIR-matched transplants. Risks of chronic GVHD were similar in all groups and independent of KIR ligand status (Table 2). Treatment-Related Mortality
Death from treatment-related causes occurred in 433 of 967 recipients of HLA-matched transplants, in
81 of 133 recipients of GVH KIR ligand–mismatched transplants, in 103 of 161 recipients of HVG KIR ligand–mismatched transplants, and in 133 of 251 HLA-B and/or –C–mismatched but KIR-matched transplants. TRM was lowest after HLA-matched transplantation (Table 3; Figure 1) and was significantly higher after any mismatched transplantations, but did not differ by KIR ligand status. Relapse
Leukemia recurred after transplantation in 156 of 967 recipients of HLA-matched transplants, in 24 of 133 recipients of GVH KIR ligand–mismatched transplants, in 16 of 161 recipients of HVG KIR ligand– mismatched transplants, and in 42 of 251 HLA-B and/or –C–mismatched but KIR-matched transplants. Leukemia recurrence was similar in all groups and independent of KIR ligand matching status (Table 3; Figure 2). Leukemia-Free Survival
Treatment failure (relapse or death from any cause) occurred in 589 of 967 recipients of HLAmatched transplants, in 105 of 133 recipients of GVH KIR ligand–mismatched transplants, in 119 of 161 recipients of HVG KIR ligand–mismatched transplants, and in 175 of 251 HLA-B and/or -C–mismatched but KIR-matched transplants. Treatment
881
KIR Ligand Incompatibility in Unrelated Donor Transplantation
Table 3. Results of Multivariate Analysis Comparing Rates of TRM, Relapse, Treatment Failure, and Overall Mortality After HLA- and KIR-Matched, GVH KIR-Mismatched, HvG KIR-Mismatched, and HLA-Mismatched, KIR-Matched Transplantations Hazard Ratio (99% Confidence Interval) Treatment-related mortality* GvH KIR-mismatched vs. HLA and KIR-matched KIR-matched, HLA-mismatched vs. HLA- and KIR-matched HvG KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. KIR-matched, HLA-mismatched Relapse† GvH KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. HLA- and KIR-matched HvG KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. KIR-matched, HLA-mismatched Treatment failure‡ GvH KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. HLA- and KIR-matched HvG KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. KIR-matched, HLA-mismatched Overall mortality§ GvH KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. HLA- and KIR-matched HvG KIR-mismatched vs. HLA- and KIR-matched KIR-matched, HLA-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. GvH KIR-mismatched HvG KIR-mismatched vs. KIR-matched, HLA-mismatched
P Value
1.95 1.51 1.75 0.77 0.89 1.16
(1.43–2.68) (1.16–1.96) (1.32–2.32) (0.54–1.11) (0.61–1.31) (0.82–1.63)
<.001 <.001 <.001 .07 .45 .27
1.56 1.11 0.75 0.71 0.48 0.68
(0.88–2.75) (0.70–1.75) (0.38–1.47) (0.37–1.38) (0.21–1.10) (0.32–1.45)
.04 .56 .27 .19 .02 .18
1.90 1.44 1.49 0.76 0.78 1.04
(1.44–2.51) (1.15–1.80) (1.15–1.93) (0.55–1.04) (0.55–1.11) (0.76–1.41)
<.001 <.001 <.001 .02 .07 .76
1.94 1.45 1.51 0.75 0.78 1.04
(1.47–2.57) (1.17–1.83) (1.16–1.98) (0.54–1.04) (0.55–1.11) (0.76–1.43)
<.001 <.001 <.001 .02 .07 .75
*In all groups, TRM was higher in older recipients (age 20 – 49 years; RR, 2.34; 99% CI, 1.66 –3.32; P ⬍ .001; age 50 – 65 years; RR, 2.90; 99% CI, 1.91– 4.39; P ⬍ .001) and those with advanced disease (not in first CR, first CP, or RA) at transplantation (RR, 1.34; 99% CI, 1.07–1.67; P ⫽ .001). TRM was significantly lower in CML recipients compared to recipients with AML or myelodysplasia (RR, 0.76; 99% CI, 0.61– 0.95; P ⫽ .002). †In all groups, patients with advanced disease (not in first CR, first CP, or RA) at transplantation were more likely to relapse (RR, 2.73; 99% CI, 1.75– 4.26; P ⬍ .001) and those with CML and MDS, less likely to relapse compared to AML patients (RR, 0.45; 99% CI, 0.30 – 0.68; P ⬍ .001 and RR, 0.48; 99% CI, 0.29 – 0.80; P ⬍ .001, respectively). ‡In all groups, treatment failure was higher in older recipients (age 20 – 49 years: RR, 2.06; 99% CI 1.56 –2.73; P ⬍ .001; age 50 – 65 years: RR, 2.55; 99% CI, 1.81–3.60; P ⬍ .001) and those with advanced disease (not in first CR, first CP, or RA) at transplantation (RR, 1.59; 99% CI, 1.31–1.93; P ⬍ .001). Treatment failure was significantly lower in CML recipients (RR 0.70; 99% CI, 0.58 – 0.85; P ⬍ .001). §In all groups, mortality was higher in older recipients age 20 – 49 years: RR, 2.09; 99% CI, 1.57–2.78; P ⬍ .001; age 50 – 65 years; RR, 2.56; 99% CI, 1.80 –3.65; P ⬍ .001) and those with advanced disease (not in first CR, first CP or RA) at transplantation (RR, 1.60; 99% CI, 1.31–1.95; P ⬍ .001), but was significantly lower in CML recipients (RR, 0.69; 99% CI, 0.57– 0.85; P ⬍ .001).
failure was lowest after HLA-matched transplantation (Table 3; Figure 3). Treatment failure was significantly higher after mismatched transplantations and did not differ by KIR ligand status. Overall Survival
Death from any cause occurred in 582 of 1004 recipients of HLA-matched transplants, in 106 of 137 recipients of GVH KIR ligand–mismatched transplants, in 120 of 170 HVG KIR ligand–mismatched transplants, and in 177 of 260 HLA-B and/or -C–mismatched but KIR-matched transplants. Mortality was lowest after HLA-matched transplantation (Table 3; Figure 4). Mortality was significantly higher after mismatched transplantations and did not differ by KIR ligand status.
DISCUSSION The superior results associated with KIR ligand incompatibility in the GVH direction in haplotypemismatched HCT [9,10], and subsequently in 1 small study of unrelated donor HCT [12], suggest a possible clinical benefit from KIR ligand incompatibility in a wider transplantation setting. It is thought that KIR ligand incompatibility in the GVH direction results in early recovery of donor alloreactive NK cells with potent antileukemic activity. In an animal model of T-cell– depleted grafts, such cells are associated with eradication of leukemia and of host antigen-presenting cells, preventing GVHD even in the setting of significant histoincompatibility [10]. The Perugia group has translated these data into the clinical setting
882
S. S. Farag et al.
Figure 1. Cumulative incidence of TRM by donor–recipient HLA and KIR ligand matching. The 5-year rates were 45% after HLAmatched and KIR ligand–matched transplantations (n ⫽ 967), 62% after GVH KIR ligand–mismatched transplantations (n ⫽ 133), 54% after HLA-B and/or –C–mismatched but KIR-matched transplantations (n ⫽ 251), and 65% after HVG KIR ligand–mismatched transplantations (n ⫽ 161) (P ⬍ .001).
Figure 3. Probability of LFS by donor–recipient HLA and KIR ligand matching. The 5-year rates were 39% after HLA-matched and KIR ligand–matched transplantations (n ⫽ 967), 20% after GVH KIR ligand–mismatched transplantations (n ⫽ 133), 29% after HLA-B and/or –C–mismatched but KIR-matched transplantations (n ⫽ 251), and 26% after HVG KIR ligand–mismatched transplantations (n ⫽ 161) (P ⬍ .001).
of haploidentical transplantation [9,10]. In the context of an aggressively T-cell– depleted haploidentical graft with a high stem cell dose and no posttransplantation immune suppression, a marked reduction in relapse and GVHD and excellent engraftment was demonstrated in very– high-risk adults with AML [9,10]. The results of our study contrast with these observations, because we did not observe any beneficial effect of KIR ligand incompatibility in the GVH direction on any of the major clinical endpoints of unrelated donor HCT. Patients receiving HLAmatched transplants had significantly lower TRM, relapse, treatment failure, and overall mortality than patients receiving transplants from GVH KIR ligand– incompatible donors. Furthermore, we could not demonstrate any significant benefit of KIR ligand incompatibility in either the GVH or HVG direction on major clinical endpoints compared with HLA-mis-
matched transplantations without KIR ligand mismatch. Studies investigating the significance of KIR ligand incompatibility in mismatched transplantation have yielded mixed and conflicting results. Although our results using the largest study population to date are consistent with previous reports showing no significant benefit of GVH KIR ligand mismatching in unrelated donor HCT [13,21,22], other studies have reported a variable reduction in relapse [12,23] and improvement in OS [12], although none has reported protection from GVHD. In an additional study, KIR ligand incompatibility was associated with inferior survival [24]. Differences in sample size, transplantation techniques, and methods of determining KIR ligand incompatibility may explain the differences among the current study and previous reports. Thus far, most reports are limited to single institutions, whereas the
Figure 2. Cumulative incidence rates of relapse by donor–recipient HLA and KIR ligand matching. The 5-year rates were 16% after HLA-matched and KIR ligand–matched transplantations (n ⫽ 967), 18% after GVH KIR ligand–mismatched transplantations (n ⫽ 133), 17% after HLA-B and/or –C–mismatched but KIR-matched transplantations (n ⫽ 251), and 10% after HVG KIR ligand– mismatched transplantations (n ⫽ 161) (P ⫽ not significant).
Figure 4. Probability of OS by donor–recipient HLA and KIR ligand matching. The 5-year rates were 42% after HLA-matched and KIR ligand–matched transplantations (n ⫽ 967), 24% after GVH KIR ligand–mismatched transplantations (n ⫽ 133), 31% after HLA-B and/or –C–mismatched but KIR-matched transplantations (n ⫽ 251), and 30% after HVG KIR ligand–mismatched transplantations (n ⫽ 161) (P ⬍ .001).
883
KIR Ligand Incompatibility in Unrelated Donor Transplantation
current analysis includes more than 1500 patients. In addition, the beneficial effect of GVH KIR ligand mismatching was originally described in haplotypemismatched transplantations performed using a high CD34⫹ cell dose after extensive T-cell depletion of the graft to prevent GVHD without additional posttransplantation immunosuppression [9,10]. In contrast, almost 80% of patients in our study received unmanipulated grafts. Furthermore, although the doses of T cells infused are not available, it is likely that T-cell depletion was only partial, because most methods used (eg, use of T-lymphocyte–specific antibodies and complement, elutriation) are likely to have achieved only a 2–3 log depletion, compared with the more extensive (4 –5 log) depletion of T cells performed in the haplotype-mismatched transplantations. In the current analysis, the use of T-depleted grafts (ex vivo) or ATG (resulting in some in vivo T-cell depletion) failed to demonstrate an effect of KIR ligand mismatching on transplantation outcome. This is consistent with recent data showing that T-cell alloreactivity dominates over NK cell alloreactivity in minimally T-cell– depleted HLA-nonidentical transplantations [19,21]. Furthermore, the presence of significant numbers of T cells in the graft may affect NK cell receptor acquisition after transplantation [25]. A comparison of NK cell receptor expression of baseline recipient and donor-derived engrafting NK cells at 100 days after volunteer unrelated donor (VUD) transplantation showed diminished reconstitution of KIR expression with increased expression of the activating receptor NKG2D in Treplete transplantations compared with T-cell– depleted transplantations [25]. This was also associated with a higher proportion of engrafted NK cells secreting interferon-␥ in response to interleukin (IL)-12 and IL-18 [25]. These observations are consistent with our observation of better outcomes with better-matched grafts, because any effect of KIR ligand incompatibility may have been masked by the effect of alloreactive T cells and/or of posttransplantation immunosuppression. It is important to note that we did not use KIR genotyping to classify our cases, but instead relied entirely on HLA typing to determine KIR ligand incompatibility and hence NK cell alloreactivity [9,13]. Central to this algorithm is the principle that NK cell alloreactivity occurs when the HLA genotype of the donor includes a KIR epitope that is not part of HLA genotype of the recipient cells, with the assumption that the donor’s NK cells express inhibitory KIR for the donor’s own HLA ligands [8]. Recent evidence indicates that this assumption may not always be correct, however. KIR and HLA genotypes segregate independently, with the expressed KIR repertoire regulated by KIR genotype and not by HLA genotype [26-29]. In a significant proportion of cases, therefore,
individuals may lack KIR for their HLA ligands, and vice versa [30]. Indeed, in a recent study of T-cell– depleted haplotype-mismatched transplantations, KIRdriven alloreactivity was better predicted if donor KIR genotype was considered along with the HLA type of the recipient [31]. Relapses were significantly less frequent, and survival improved in cases where the recipient was lacking the HLA ligand for the donorinhibitory KIR as assessed by KIR typing [31]. More recently, this “missing ligand” algorithm was also shown to be strongly predictive of outcome even in T-cell– depleted HLA-matched sibling donor transplantations for AML and MDS, where in this setting ⬎ 60% of HLA-matched recipients were missing 1 or more ligands for their own and their donor’s inhibitory KIR [32]. Finally, a further limitation imposed by the lack of KIR genotyping is the potential confounding effect of activating KIR, which could not be examined in our study but may be important in determining outcome [33-36]. In summary, the results of the current study of unrelated donor HCT show that HLA mismatch is associated with higher TRM, treatment failure, and overall mortality regardless of the presence or absence of KIR ligand mismatch. Furthermore, no advantage of GVH KIR ligand incompatibility for any of the clinical endpoints examined was observed compared with other HLA-mismatched VUD transplantations. Accordingly, our results do not support the choice of unrelated donor HCT on the basis of KIR ligand mismatching using HLA typing alone, even when a HLA-B or -C mismatch is necessary.
ACKNOWLEDGMENTS This work was supported by Public Health Service grant U24-CA76518-08 from the National Cancer Institute, National Institute of Allergy and Infectious Diseases, and National Heart, Lung, and Blood Institute.
REFERENCES 1. Davies SM, Kollman C, Anasetti C, et al. Engraftment and survival after unrelated-donor bone marrow transplantation: a report from the National Marrow Donor Program. Blood. 2000; 96:4096-4102. 2. McGlave PB, Shu XO, Wen W, et al. Unrelated donor marrow transplantation for chronic myelogenous leukemia: 9 years’ experience of the National Marrow Donor Program. Blood. 2000;95:2219-2225. 3. Sasazuki T, Juji T, Morishima Y, et al. Effect of matching of class I HLA alleles on clinical outcome after transplantation of hematopoietic stem cells from an unrelated donor. N Engl J Med. 1998;339:1177-1185. 4. Petersdorf EW, Longton GM, Anasetti C, et al. Association of
884
5.
6.
7. 8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18. 19.
20. 21.
S. S. Farag et al.
HLA-C disparity with graft failure after marrow transplantation from unrelated donors. Blood. 1997;89:1818-1823. Szydlo R, Goldman JM, Klein JP, et al. Results of allogeneic bone marrow transplants for leukemia using donors other than HLA-identical siblings. J Clin Oncol. 1997;15:1767-1777. Flomemberg N, Baxter-Lowe LA, Confer D, et al. Impact of HLA class I and class II high-resolution matching on outcomes of unrelated donor bone marrow transplantation: HLA-C mismatching is associated with a strong adverse effect on transplantation outcome. Blood. 2004;104:1923-1930. Lanier LL. NK cell receptors. Annu Rev Immunol. 1998;16:359393. Farag SS, Fehniger TA, Ruggeri L, et al. Natural killer cell receptors: new biology and insights into the graft-versus-leukemia effect. Blood. 2002;100:1935-1947. Ruggeri L, Capanni M, Casucci M, et al. Role of natural killer cell alloreactivity in HLA-mismatched hematopoietic stem cell transplantation. Blood. 1999;94:333-339. Ruggeri L, Capanni M, Urbani E, et al. Effectiveness of donor natural killer cell alloreactivity in mismatched hematopoietic transplants. Science. 2002;295:2097-2100. Aversa F, Terenzi A, Tabilio A, et al. Full haplotype-mismatched hematopoietic stem-cell transplantation: a phase II study in patients with acute leukemia at high risk of relapse. J Clin Oncol. 2005;23:3447-3457. Geibel S, Locatelli F, Maccario R, et al. Survival advantage with KIR ligand incompatibility in unrelated donor transplantation. Blood. 2003;102:814-819. Davies SM, Ruggeri L, DeFor T, et al. Evaluation of KIR ligand incompatibility in mismatched unrelated donor hematopoietic transplants: killer immunoglobulin-like receptor. Blood. 2002;100:3825-3827. Bennett JM, Catovsky D, Daniel MT, et al. Proposed revised criteria for the classification of acute myeloid leukemia: a report of the French-American-British Cooperative Group. Ann Intern Med. 1985;103:620-625. Bennett JM, Catovsky D, Daniel MT, et al. Proposals for the classification of the myelodysplastic syndromes. Br J Haematol. 1982;51:189-199. Przepiorka D, Weisdorf D, Martin P, et al. 1994 consensus conference on acute GVHD grading chronic graft-versus-host syndrome in man: a long-term clinicopathologic study of 20 Seattle patients. Bone Marrow Transplant. 1995;15:825-828. Shulman HM, Sullivan KM, Weiden PL, et al. Chronic graftversus-host syndrome in man: a long-term clinicopathologic study of 20 Seattle patients. Am J Med. 1980;69:204-217. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1998;16:1141-1154. Gooley TA, Leisenring W, Crowley J, et al. Estimation of failure probabilities in the presence of competing risks: new representations of old estimators. Stat Med. 1999;18:695-706. Cox DR. Regression models and life tables (with discussion). J R Stat Soc B. 1972;34:187-220. Bishara A, De Santis D, Witt CC, et al. The beneficial role of inhibitory KIR genes of HLA class I NK epitopes in haploidentically mismatched stem cell allografts may be masked by residual donor-alloreactive T cells causing GVHD. Tissue Antigens. 2004;63:204-211.
22. Lowe EJ, Turner V, Handgretinger R, et al. T-cell alloreactivity dominates natural killer cell alloreactivity in minimally T-cell– depleted HLA-nonidentical paediatric bone marrow transplantation. Br J Haematol. 2003;123:323-326. 23. Beelen DW, Ottinger HD, Ferencik S, et al. Genotypic inhibitory killer immunoglobulin-like receptor ligand incompatibility enhances the long-term antileukemic effect of unmodified allogeneic hematopoietic stem cell transplantation in patients with myeloid leukemias. Blood. 2005;105:2594-2600. 24. Schaffer M, Malmberg KJ, Ringden O, et al. Increased infection-related mortality in KIR ligand-mismatched unrelated allogeneic hematopoietic stem-cell transplantation. Transplantation. 2004;78:1081-1085. 25. Cooley S, McCullar V, Wangen R, et al. KIR reconstitution is altered by T cells in the graft and correlates with clinical outcomes after unrelated donor transplantation. Blood. 2005;106:4370-4376. 26. Gumperz JE, Valiante NM, Parham P, et al. Heterogeneous phenotypes of expression of the NKB1 natural killer cell class I receptor among individuals of different human histocompatibility leukocyte antigens types appear genetically regulated, but not linked to major histocompatibililty complex haplotype. J Exp Med. 1996;183:1817-1827. 27. Frohn C, Schlenke P, Kirchner H. The repertoire of HLACw–specific NK cell receptors CD158 a/b (EB6 and GL183) in individuals with different HLA phenotypes. Immunology. 1997; 92:567-570. 28. Vilches C, Parham P. KIR: diverse, rapidly evolving receptors of innate and adaptive immunity. Ann Rev Immunol. 2002;20: 217-251. 29. Parham P, McQueen KL. Alloreactive killer cells: hindrance and help for haematopoietic transplants. Nat Rev Immunol. 2003; 3:108-122. 30. Dupont B, Hsu KC. Inhibitory killer Ig-like receptor genes and human leukocyte antigen class I ligands in haematopoietic stem cell transplantation. Curr Opin Immunol. 2004;16:634-643. 31. Leung W, Iyengar R, Turner V, et al. Determinants of antileukemia effects of allogeneic NK cells. J Immunol. 2004;172: 644-650. 32. Hsu KC, Keever-Taylor CA, Wilton A, et al. Improved outcome in HLA-identical sibling hematopoietic stem cell transplantation for acute myelogenous leukemia predicted by KIR and HLA genotypes. Blood. 2005;105:4878-4884. 33. Gagne K. Relevance of KIR gene polymorphisms in bone marrow transplantation outcome. Hum Immunol. 2002;63:271280. 34. Bishara A, De Santis D, Witt CC, et al. The beneficial role of inhibitory KIR genes of HLA class I NK epitopes in haploidentically mismatched stem cell allografts may be masked by residual donor-alloreactive T cells causing GVHD. Tissue Antigens. 2004;63:204-211. 35. De Santis D, Bishara A, Witt CC, et al. Natural killer cell HLA-C epitopes and killer cell immunoglobulin-like receptors both influence outcome of mismatched unrelated donor bone marrow transplants. Tissue Antigens. 2005;65:519-528. 36. Cook MA, Milligan DW, Fegan CD, et al. The impact of donor KIR and patient HLA-C genotypes on outcome following HLA-identical sibling hematopoietic stem cell transplantation for myeloid leukemia. Blood. 2004;103:1521-1526.