|Home | About | Journals | Submit | Contact Us | Français|
Donor availability is one of the major obstacles to the success of allogeneic hematopoietic stem cell transplantation (HSCT) for the treatment of hematologic malignancies or non-malignant hematologic disorders. Because of historically superior outcomes of human leukocyte antigen (HLA)-matched as compared to partially HLA-mismatched HSCT 1,2, an HLA-matched sibling or unrelated donor is the preferred source of stem cells for transplantation. However, an HLA-matched donor can be identified for only 50–60% of patients referred for HSCT, lower still for patients in ethnic minorities. The ability to cross the HLA boundary safely would increase the availability of a stem cell donor to nearly 100% of patients referred for allogeneic HSCT.
There are two potential sources of grafts for patients lacking HLA-matched donors: 1) unrelated umbilical cord blood (UCB), and 2) partially HLA-mismatched, or HLA-haploidentical, related donors. Results of UCB transplantation in children are encouraging3, and transplantation of 2 UCB units generates cell doses that are sufficient for engraftment in adults4,5. The initial studies of HLA-haploidentical HSCT employed lethal conditioning, infusion of T cell-replete marrow grafts, and graft-versus-host disease (GVHD) prophylaxis with methotrexate, with or without cyclosporine6. These transplants were complicated by excessive bi-directional alloreactivity resulting in high rates of graft failure7, severe GVHD, and non-relapse mortality8. Consequently, event-free survival was poor, especially when donors and recipients were mismatched for two or more HLA antigens8,9. Results of HLA-haploidentical HSCT have improved significantly over the past decade owing to the development of highly immunosuppressive yet nonmyeloablative conditioning, novel graft manipulation, and improved prophylaxis of graft-versus-host disease (GVHD). Further, HLA-haploidentical HSCT harnesses the potential of natural killer (NK) cell alloreactivity to kill tumor cells and reduce the risk of post-transplantation relapse. These recent developments are the subject of this review.
Graft failure is a major complication of HLA-haploidentical HSCT6,10 and is usually a fatal event after myeloablative conditioning. Truly nonmyeloablative conditioning offers the safeguard of reconstitution of autologous hematopoiesis in the event of graft failure. Most nonmyeloablative conditioning regimens incorporate the highly immunosuppressive drug fludabarine. Studies from Tuebingen, Germany and from Duke University in the United States have combined fludarabine-based conditioning with in vivo T cell depletion using OKT311 or CAMPATH12, respectively, to enable the engraftment of HLA-haploidentical stem cells. These regimens were associated with acceptable non-hematologic toxicities and sustained engraftment of donor cells in patients up to the age of 66. Overall survival (OS) at one year after transplantation ranged from 31%–37%12,13, establishing the feasibility of HLA-haploidentical HSCT after nonmyeloablative conditioning.
The groups at Johns Hopkins in Baltimore and the Fred Hutchinson Cancer Research Center in Seattle have been pioneering the use of high-dose, post-transplantation cyclophosphamide (Cy) to achieve the selective depletion of alloreactive cells after nonmyeloablative conditioning and HLA-haploidentical HSCT. In an early report, sixty-eight patients with poor-risk hematologic malignancies were conditioned with fludarabine, Cy, and 2 Gy total body irradiation prior to receiving T cell-replete bone marrow from HLA-haploidentical, first-degree relatives (Figure 1)14. Donors and recipients were mismatched at a median of 4 HLA alleles. GVHD prophylaxis comprised Cy 50 mg/kg IV on day 3 (n=28) or on days 3 and 4 (n=40) after transplantation, followed by tacrolimus and mycophenolate mofetil, each beginning on day 5. Graft failure occurred in 9 patients (13%) but was fatal in only one. Acute grades II–IV and III–IV GVHD occurred in 34% and 6% of patients, respectively, and chronic GVHD developed in 15% of patients. The cumulative incidences of relapse and non-relapse mortality (NRM) at one year after transplantation were 15% and 51%, respectively, and OS and event-free survivals (EFS) at two years after transplantation were 36% and 26%. Only six patients died of infection (n=4) or GVHD (n=2). In this early report, patients with lymphoid diseases had a superior event-free survival compared to patients receiving HSCT for myeloid diseases (p=.02).
A subsequent report retrospectively compared the outcomes of Hodgkin lymphoma (HL) patients treated with nonmyeloablative conditioning and grafts from HLA-matched related (n=38), unrelated (n=24), or HLA-haploidentical related (n=28) donors15. Recipients of HLA-haploidentical grafts were conditioned as in Figure 1. Patients had received a median of 5 prior regimens, including autologous HSCT in 92%. With a median follow-up of 25 months, 2-year OS, EFS, and incidences of relapsed/progressive disease were 53%, 23%, and 56% (HLA-matched related), 58%, 29%, and 63% (unrelated), and 58%, 51%, and 40% (HLA-haploidentical related), respectively. NRM was significantly lower for HLA-haploidentical related recipients compared to HLA-matched related recipients (p=.02). There were also significantly decreased risks of relapse for HLA-haploidentical related recipients compared to HLA-matched related (P=.01) and unrelated (P=.03) recipients. In a recent report from the CIBMTR, HL patients receiving reduced intensity, unrelated donor HSCT had a 2-year OS and EFS of 37% and 20%, respectively16. HLA-haploidentical HSCT may therefore be uniquely effective for patients with relapsed or refractory HL.
We have recently analyzed, retrospectively, the effect of HLA mismatching on the outcome of 185 hematologic malignancies patients treated with nonmyeloablative, HLA-haploidentical SCT and post-transplantation Cy17. The cumulative incidences of grade II–IV acute GVHD and chronic GVHD were 31% and 15%, respectively. The cumulative incidences of NRM and relapse or progression at one year were 15% and 50%, respectively. Actuarial EFS at one year was 35%. Increasing degrees of HLA mismatch at either class I or class II loci had no significant effect on the cumulative incidence of acute or chronic GVHD or NRM. In contrast, the presence of an HLA-DRB1 antigen mismatch in the GVH direction was associated with a significantly lower cumulative incidence of relapse (Figure 2a; p = 0.04) and improved EFS (Figure 2c; p = 0.009), whereas HLA-DQB1 antigen mismatch status had no effect. Additionally, the presence of two or more class I allele mismatches (composite of HLA-A, -B, and -Cw) in either direction was associated with a significantly lower cumulative incidence of relapse (Figure 2b; p = 0.045 for GVH direction, p = 0.01 for HVG direction) and improved EFS (Figure 2d; p = 0.07 for GVH direction, p = 0.001 for HVG direction). Although the analysis was limited by its retrospective nature and the small numbers of pairs with two or fewer HLA antigen mismatches (n=26), the results raise the possibility that increasing HLA disparity is associated with improved outcomes after nonmyeloablative, HLA-haploidentical HSCT with high-dose, post-transplantation Cy.
Previous studies of myeloablative, HLA-haploidentical HSCT have shown that increasing HLA disparity was associated with a reduced incidence of relapse but an inferior EFS due to increased GVHD and NRM8,18. In contrast, our study of nonmyeloablative, HLA-haploidentical HSCT with post-transplantation Cy showed that increasing HLA disparity was associated with a reduced risk of relapse with no effect on NRM, resulting in an improved EFS. What accounts for the difference? While conceding the pitfalls of retrospective analyses, we raise the possibility that post-transplantation Cy differentially affects the populations of cells mediating GVHD versus those producing graft-versus-leukemia effects. Further clinical and laboratory studies will be required to understand the effects of high-dose, post-transplantation Cy on host tolerance and anti-tumor immunity.
As mentioned previously, the initial trials of HLA-haploidentical bone marrow transplantation for leukemia used lethal conditioning and T cell-replete grafts and were complicated by high rates of severe GVHD and NRM, and 5-year EFS among recipients of grafts mismatched for 2–3 HLA loci was approximately 10%8. Although T cell depletion reduced the risk of GVHD after HLA-haploidentical HSCT, it increased the risk of graft failure and did not improve leukemia-free survival19. Investigators in Perugia, Italy achieved low rates of graft failure and GVHD by conditioning patients intensively and transplanting them with rigorously T cell-depleted grafts containing “megadoses” of CD34+ hematopoietic stem cells20. Event-free survival (+/− standard deviation) rate was 48% +/− 8% and 46% +/− 10%, respectively, for 42 AML and 24 ALL patients receiving transplantation in remission21. These studies also established a potential role for donor natural killer cells in mediating graft-versus-leukemia effects after T cell-depleted, HLA-haploidentical BMT for acute myeloid, but not acute lymphocytic, leukemia22.
Peking University researchers developed a novel approach to HLA-mismatched/ haploidentical transplantation without in vitro T cell depletion. This approach, shown in Figure 3, was first reported by Huang et al in a study of 58 patients undergoing HLA-mismatched/haploidentical HSCT23. Since this initial report, 831 additional patients have received HLA-haploidentical stem cell transplants at the Peking University Institute of Hematology.
Huang et al reported 171 patients, including 86 with high-risk disease, receiving grafts from HLA-mismatched/haploidentical family donors24. All patients achieved hematopoietic recovery after transplantation. The median time for myeloid engraftment was 12 days (range: 9–26 days) and median time to platelet recovery was 15 days (range: 8–151 days). There was no significant association between the extent of HLA disparity and the time to myeloid or platelet recovery. On multivariate analysis, a low number of CD34+ cells (<2.19×106/kg) in the graft, and advanced disease stage were independently associated with an increased risk of platelet non-engraftment25. Among children who received HLA-haploidentical grafts, only the dose of infused CD34+ cells/kg of recipient weight was significantly associated with an increased risk of platelet engraftment
Our results suggest that the incidences of grade III–IV acute GVHD (aGVHD) and extensive chronic GVHD (cGVHD) were acceptable in patients after unmanipulated HLA-mismatched/haploidentical transplantation, although the T cell dose in grafts was more than 108/kg. At 100 days after transplantation, the cumulative incidence was 55.0% for grade II–IV aGVHD, and 23.1% for grade III–IV aGVHD. The incidence of cGVHD was 44.67%, with 21.3% for limited and 23.3% for extensive, respectively24. We further reported 42 children below 14 years of age with hematological malignancies treated with HLA-haploidentical HSCT26. The cumulative incidence of aGVHD of grade II–IV was 57.2%, and that of grade III–IV was 13.8%. The cumulative incidence of cGVHD was 56.7% for total and 29.5% for extensive. Apparently, the incidence of grade III–IV aGVHD in pediatric patients was lower than that of adult patients. In contrast to previously published data, there was no significant association of HLA disparity with the incidence or severity of acute or chronic GVHD in this protocol.
These findings may be related to (1) T cell hyporesponsiveness maintained after in vitro mixture of G-PB and G-BM in different proportions27,28; (2) the use of ATG before transplantation, which may induce depletion of infused donor T lymphocytes in vivo and thus lower the incidence of GVHD; (3) a possible effect of the combination of CSP, MTX, and MMF as postgrafting immunosuppression; (4) the application of G-CSF day 5 post transplant, which may further regulate T cell function or (5) the immunomodulatory effect of mesenchymal stem cells (MSCs)/mesenchymal (stroma) progenitor cells (MPCs) from the G-CSF mobilized marrow and PBSC grafts, respectively
Factors correlating the high incidence of aGVHD are KIR ligand mismatch and a higher dose of CD56bright NK cells (41.9×106/kg) in the allografts, while a higher CD56dim/CD56bright NK cell ratio (more than 8.0) in allografts was correlated with a decreased risk of III–IV aGVHD after unmanipulated HLA-mismatched/haploidentical transplantation.
We studied the incidence and management of relapsed malignancy in 250 recipients of HLA-haploidentical transplants at Peking University29. The 3-year probabilities of relapse in the standard-risk group were 11.9% for AML and 24.3% for ALL, and in the high-risk group were 20.2% for AML and 48.5% for ALL. Advanced disease status, a higher CD4/CD8 ratio in G-BM30, and delayed lymphocyte recovery at day 30 post transplantation correlated with an increased relapse rate. Conversely, a higher CD56dim/CD56bright NK cell ratio (more than 8.0) was correlated with a decreased rate of relapse after haploidentical transplantation without in vitro T-cell depletion
Modified DLI was used to treat relapse of patients after unmanipulated HLA-mismatched/haploidentical transplantation31. Twenty patients who underwent T cell-depleted, HLA-haploidentical HSCT between April 1, 2002 and May 1, 2005 were included in this study. After DLI, eleven patients received CsA (blood concentration of 150–250 ng/mL for 2–4 weeks) or a low dose of MTX (10 mg once per week for 2–4 weeks) to prevent GVHD, and nine patients received no GVHD prophylaxis. The incidence of grade III–IV aGVHD was significantly lower in patients with GVHD prophylaxis than those without (9.1% vs. 55.6%, P=0.013). Fifteen patients achieved CR at a median of 289 (40–1388) days after DLI. The 1-year and 2-year LFS were 60% and 40%.
In a recent report, 250 acute leukemia (AL) patients received allografts from related donors29. The non-relapse mortality at day 100 after transplantation in the standard- and high-risk groups was 6.8% and 5.9% for AML and 6.9% and 25.9% for ALL, respectively.
An improved leukemia-free survival after unmanipulated HLA-haploidentical blood and marrow transplantation correlated closely with early disease status, higher numbers of CD56bright cells reconstituted day 14 post transplant , lower CD4/CD8 in G-BM, a short time from diagnosis to transplant (≤450d) for CML patients and higher absolute counts of lymphocytes (more than 300/µl) day 30 post transplant. In a large cohort of AL patients, the 3-year probabilities of LFS for standard-risk and high-risk patients were 70.7% and 55.9%, respectively, for patients with AML, and 59.7% and 24.8%, respectively, for patients with ALL29. With respect to CML patients, the probability of 1-year and 4-year LFS was 76.5% and 74.5% for patients in first chronic phase, 85.7% and 85.7% for CP2/CR2 patients, 80% and 66.7% for AP patients, and 53.8% and 53.8% for BC patients.
The most important development in HLA-haploidentical HSCT over the past decade has been the dramatic reduction in transplant-related morbidity and mortality. Highly immunosuppressive conditioning regimens now permit the transplantation of T cell-depleted grafts, resulting in reliable donor cell engraftment without severe GVHD. As a result, the mortality associated with HLA-haploidentical HSCT now approaches that of HLA-matched HSCT32, making partially mismatched related donor transplantation a viable treatment option for patients lacking an HLA-matched donor. Going forward, there is a need to decrease the risk of post-transplant infections by improving immune reconstitution, to harness both T cell and NK cell alloreactivity for improved anti-tumor effects without GVHD, and to define the relative roles of HLA-haploidentical related donor versus unrelated umbilical cord blood SCT for various hematologic malignancies.
The mechanism by which NK cells acquire self-tolerance and alloreactivity has been referred to as NK cell education or licensing. This is one of the most widely debated topics in NK cell biology over the past several years. Several models have been proposed to explain the integration of inhibitory receptor expression with the acquisition of effector function. These concepts differ in their implied mechanisms and whether the process is one of activation or loss of function33,34. What is agreed upon between these and other models is that human NK cells lacking inhibitory receptors are hyporesponsive35,36. Therefore, rather than being autoreactive, they are self-tolerant. Although the exact mechanism remains unknown, self-tolerance may be the result of coordinated developmental pathways whereby mature NK cell function is synchronized with the acquisition of self-inhibitory receptors.
The two main strategies to harness the therapeutic power of alloreactive NK cells are: 1) hematopoietic cell transplantation22 and 2) adoptive transfer of NK cells37. This literature is based on studies from the Perugia group who first proposed the KIR-ligand incompatibility model, which predicts that donor-derived NK cells will be alloreactive when recipients lack C2, C1 or Bw4 alleles that are present in the donor. Many groups, including our own38–40, have tested the clinical efficacy of selecting donors for NK cell therapy or transplantation based on their predicted alloreactivity against the host using one of several models. The potential benefits include: 1) decreased GVHD as host dendritic cells are killed by donor NK cells, 2) better anti-tumor activity via direct cytotoxicity, 3) improved engraftment mediated by NK cell release of hematopoietic cytokines, and 4) enhanced immune reconstitution. Additional clinical trials have supported the finding that KIR ligand mismatch is associated with favorable clinical outcomes in myeloid malignancies41. However, other studies looking at outcomes after KIR ligand mismatched, T-cell replete transplants did not find the same effect, perhaps because T-cells in the graft interfere with NK cell development and KIR reconstitution after allogeneic donor transplant as we have shown42. Taken together, these results suggest that NK cells play a role in allogeneic transplant and cancer therapy; however, the complexities of the KIR system and the presence of other functional receptors on NK cells may explain some of the confusion in interpreting published studies.
We have shown that adoptive transfer of haploidentical natural killer (NK) cells can induce remissions in 27% of patients with refractory or relapsed acute myeloid (AML) 6. The remissions induced by adoptive NK cell transfer were not durable. We hypothesized that this may be in part related to the lack of in vivo expansion of NK cells on all patients. Since lymphocyte homeostasis is determined by factors resulting from lymphodepletion, we increased our preparative regimen and added a CD34+ stem cell infusion to create a non-myeloablative haploidentical transplantation protocol. Radiation (200 cGy twice a day on day -13) was added to a preparative regimen used in non-transplant patients that included fludarabine 25 mg/m2 × 5 (day -18 through day -14) and cyclophosphamide 60 mg/kg × 2 (days -16 and -15). The NK cell product was activated with 1000 U/ml IL-2 and infused on day -12 followed by 6 doses subcutaneous IL-2 (10 million units) given every other day to promote in vivo NK cell expansion. The mean NK cell dose was 1.85 × 107 cells/kg. A CD34-selected peripheral blood graft from the same donor was given with Thymoglobulin 3 mg/kg days 0, +1 and +2 as the only additional immunosuppression. In the 13 patients a significantly higher rate of NK cell expansion (75% [9/12 evaluable]; mean 607±184 NK cells/ml) was achieved compared to the adoptive NK cell transfer regimen, which did not include radiation. This adoptive NK cell plus allograft protocol led to 66% of relapsed or refractory AML patients (8/12 evaluable) clearing leukemia by day -1. Patients who did not clear leukemia (N=4) did not engraft. All others (N=6) engrafted promptly at a median 17 days [range 11–31]). None developed graft vs. host disease (GVHD), but infectious complications were common, not unexpected in a high-risk cohort where subjects typically had prolonged neutropenia prior to transplantation. In summary, in patients with refractory AML, addition of haploidentical NK cells to a non-myeloablative haploidentical transplantation yields NK cell expansion in a majority of patients, achievement of complete remission, and quick engraftment without GVHD. This is a promising platform upon which to add other strategies aimed at improving disease-free survival in patients with refractory AML. Additional strategies to sensitize NK cells to leukemia, to target leukemic stem cells, to improve in vivo expansion, to interrupt inhibitory receptor interactions with class I MHC and to pick donors are among future strategies to improve this therapy.
The importance of killer immunoglobulin-like receptors (KIR) in determining clinical outcome after hematopoietic cell transplantation (HCT) remains controversial. We genotyped donors and recipients from 209 HLA-matched and 239 mismatched T-replete URD transplantations for AML43. Three-year overall survival was significantly higher after transplantation from a KIR B/x donor (31% [95% CI: 26–36] vs 20% [95% CI: 13–27]; P=.007). Multivariate analysis demonstrated a 30% improvement in the relative risk of relapse-free survival with B/x donors compared with A/A donors (RR: 0.70 [95% CI: 0.55–0.88]; P=.002). This demonstrates that unrelated donors with KIR B haplotypes confer significant survival benefit to patients undergoing T-replete HCT for AML. KIR genotyping should be added to donor selection criteria in addition to HLA typing, to identify donors with B KIR haplotypes. Future investigators are aimed at subsetting the KIR B haplotype for a more refined donor selection strategy.
NK cells have been of therapeutic interest for decades as they kill tumor targets in vitro and in animal models. Strategies to activate autologous NK cells dominated the early literature but were found to limited efficacy. This was explained by the discovery of inhibitory receptors on NK cells that recognize “self” MHC molecules. Current strategies using allogeneic NK cells are based on the premise that they will result in a higher frequency of donor cells that will be reactive against the recipient. The promising finding in AML strongly support a role for the therapeutic use of NK cells and offers the opportunity to further manipulate these cells to exploit their full potential when combined with allogeneic transplantation.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.