The rationale behind grafting haploidentical hematopoietic stem cells

ABSTRACT

The ability to perform hematopoietic cell transplant across major histocompatibility complex barriers can dramatically increase the availability of donors and allow more patients across the world to pursue curative transplant procedures for underlying hematologic disorders. Early attempts at haploidentical transplantation using broadly reactive T-cell depletion approaches were compromised by graft rejection, graft-versus-host disease and prolonged immune deficiency. The evolution of haploidentical transplantation focused on expanding transplanted hematopoietic progenitors as well as using less broadly reactive T-cell depletion. Significant outcome improvements were identified with technology advances allowing selective depletion of donor allospecific T cells, initially ex-vivo with evolution to its current in-vivo approach with the infusion of the highly immunosuppressive chemotherapy agent, cyclophosphamide after transplantation procedure. Current approaches are facile and portable, allowing expansion of allogeneic hematopoietic cell transplantation for patients across the world, including previously underserved populations.

KEYWORDS:

• Haploidentical transplantation
• HLA mismatch
• graft vs host disease
• graft rejection

Introduction

Allogeneic hematopoietic cell transplantation (alloHCT) is a curative procedure for many patients with hematologic abnormalities, both malignant (myeloid and lymphoid) and nonmalignant (bone marrow failure, immune deficiency syndromes and hereditary hemoglobinopathies) [1]. Over the past two decades, advances in molecular medicine have diminished demand for transplantation for disorders such as non-Hodgkin lymphoma, chronic myelogenous leukemia, myelofibrosis and chronic lymphocytic leukemia due to increased availability of less toxic and increasingly effective therapies [2,3]. However, alloHCT remains the mainstay of curative therapy for patients with multiple myeloid malignancies, particularly acute myeloid leukemia (AML) and myelodysplasia (MDS). This ongoing need is highlighted by the relatively recent innovations in reduced intensity and non-myeloablative conditioning that have allowed older patients to undergo successful transplant procedures with curative intent [4]. AML has a median age at diagnosis of age 68 in adults and globally, the annual number of newly diagnosed patients with all leukemias has increased by 46% over the past 30 years [5]. The demand for alloHCT remains significant and continues to increase annually on a worldwide basis.

What is the goal of allogeneic HCT? In its simplest manifestation, it is twofold. The first goal is to eliminate the leukemic stem cell pool. Given the fact that we have no current molecular tools that would allow us to specifically segregate the leukemic stem cells from normal, healthy hematopoietic stem cells (HSC), the foremost goal is to eradicate host hematopoiesis and provide subsequent replacement with a healthy HSC pool. The second major goal is to adoptively transfer a mature donor immune system that can develop immune tolerance to expressed normal cell surface host antigens, but still maintain immune surveillance against leukemia target antigens in the event that complete eradication of the leukemic/malignant stem cell pool was not successful with conditioning therapy.

Although the goals of the effort are quite clear, there are multiple limitations to the application of alloHCT. Throughout this discussion, we will focus on AML and MDS, the two most common malignancies for which alloHCT is offered. Within this model, if we dissect the barriers that would limit alloHCT applications, we start with disease-specific factors. There are constant new molecular analyses that allow us to better risk stratify patients with AML/MDS, identifying those in need of immediate versus delayed versus never alloHCT [6]. There are patient-specific factors, particularly in this older patient age group, with medical comorbidities and pre-existing conditions and exposures that can influence the myeloid malignancy alloHCT outcome [7,8]. Third, there are transplant-specific factors including donor availability, intensity of conditioning, and degree of immunologic matching that are influenced by donor options. On this final issue, a highlight of the last 20 years has been the expansion of the unrelated donor pool (now, over 40 million potential volunteer donors) but also, importantly, the reemergence of haploidentical transplantation as a clinical standard of care [9,10]. In this review, we will focus on the identified barriers to donor hematopoietic engraftment across an HLA disparate barrier.

In the 1980s, prior to the development and adoption of calcineurin inhibitors as the central pharmacologic therapeutic for graft-versus-host disease (GVHD) prophylaxis and prior to the establishment of an unrelated donor registry, there had been successful development of monoclonal antibody technologies capable of purging sibling, HLA-matched donor marrows of mature T cells, either by antibody and complement combination or using immune conjugate depletion sheep RBC rosette technology [11–14]. These technologies had successfully led to alloHCT for older patients (at that time, with these innovations, age 50 became commonly considered the ceiling for alloHCT). In addition, attempts were made to expand donor availability beyond the fully HLA-matched sibling, into mismatched family members including parental into child (P – > F1) haploidentical alloHCT (haploHCT). However, when total T-cell depletion (TCD) was utilized to overcome the HLA barriers in haploHCT, it was fraught with failure, with both disease relapse and graft rejection commonly seen as a consequence [14–16]. These early clinical experiments provided valuable insights regarding the possibilities and the barriers to widespread use of haploHCT as a treatment modality. Specifically, factors known to have limited the success include issues regarding the dose of donor HSC required to overcome host vs graft immunologic barriers and the likelihood that incomplete depletion of haploidentical donor immune cells would impact negatively the transplant procedure with manifestation of severe GVHD in the recipient (Table 1).

Table 1. Overcoming barriers to haploidentical transplantation.

GVHD	Graft rejection	Relapse
Etiology associated with MHC class I & II HLA disparity between donor and recipient. Overcome by effective donor T-cell depletion [14]	Etiology linked with host T-cell expansión followed by the destruction of donor mismatched HLA markers [24] Enhancement of lymphotoxic conditioning regimens as target solution	Etiology attributed to the loss of T-cell mediated ‘graft-vs-malignancy’ effect Mitigation by increasingly sophisticated selective T-cell depletion

Barriers to haploHCT: GVHD and graft rejection

The nature of the immune response was elucidated in 1987 by the efforts of the laboratories of Strominger & Wiley, when the crystal structure of HLA-A2.01 was solved [17,18], confirming the major impact of amino acid variances across various HLA molecules and by demonstrating the critical nature of peptide antigen. Thus clarity was provided regarding the impact of multiple major histocompatibility complex (MHC) antigen disparities encountered in the haploHCT setting. The importance of deep TCD was revealed and demonstrated that GVHD could be limited and even completely abrogated. If suboptimal TCD was performed, death due to GVHD was often encountered. Fortunately some clinical success was identified with these early TCD technologies particularly in patients with nonmalignant disorders [11,19–21]. However, for patients with acute or chronic leukemia, the outcomes were disappointing with high relapse rates identified, approaching 70–80% in some settings [22,23]. In contrast, it also was observed that when donor TCD was used to limit the risk of acute GVHD, graft rejection rates increased [22,23]. Studies of these episodes of graft rejection demonstrated that residual host derived cytotoxic lymphocytes were activated and expanded in-vivo leading to donor graft rejection by targeting donor mismatched HLA molecules [24]. Only in rare circumstances, would clinical tolerance emerge [25]. The challenge then was to increase host immunosuppression to eliminate residual host T cells that were leading to the graft rejection [26–29].

One obvious solution would be more intensive host immunosuppression to prevent donor graft rejection in alloHCT. Total body irradiation with a tight therapeutic window for tissue and organ toxicity was shown to be markedly immunosuppressive. However, further investigations for alternate immunosuppressive conditioning expanded and led to the development of chemotherapy alone protocols including agents such as thiotepa, fludarabine and antithymocyte globulin (ATG) or the additional application of total lymphoid irradiation [15]. These regimens were utilized and demonstrated to have been successful in the prevention of graft rejection (Table 1).

Positive CD34 hematopoietic stem cell selection

Another approach to limiting the risk of graft rejection was identification of the ideal HSC dose. Over the years, one of the major goals of transplantation has been to characterize the HSC. It is likely that the transplanted HSC population is not uniform as the time spent in dormant versus replicating states may be difficult to quantify as characteristics such as telomeric DNA and the age of an HSC are challenging to identify and it is unknown when an HSC reaches a state of exhaustion. Despite extensive attempts, there has never been the ability to identify a unique surface marker that was specific for the HSC itself; rather, the work of the Weisman team suggested that one could identify the purest HSC population by flow cytometry selection techniques, using positive selection with an early marker (Sca-1) followed by negative selection using a panel of lineage specific antibodies [30,31]. In preclinical models, using such purified engraftment approaches, it was found that as few as 100 purified HSC can lead to engraftment [32]. Thus the HSC was felt to be a dormant, noncycling lineage negative cell that has the capacity to differentiate into ‘downstream’ multilineage hematopoietic populations and also to importantly, to have self-renewal capabilities [33]. Subsequent preclinical work addressed issues of dose of HSC required to contribute to engraftment where dose escalated purified HSC populations were demonstrated to provide higher degrees of engraftment [34,35].

The translation of this preclinical work to humans was limited in the setting of bone marrow allografts. The advent of growth factor mobilized peripheral blood stem cells (PBSC), opened the opportunity to determine if increased numbers of donor progenitor cells, assessed by CD34 expression and available for column selection could be utilized to overcome HLA disparity barriers [36]. The studies from Perugia again reopened interest in family related haploHCT [37–39]. In the original study, patients with high risk acute leukemia received a full haplotype-mismatched alloHCT after receiving intensive conditioning with total body radiation, cyclophosphamide, thiotepa and ATG but notably, had a TCD allograft performed with a combined bone marrow and PBSC product, with a defined target composition of CD34 of 1 × 10⁷/kg and a CD3 of 2 × 10⁵/kg [39]. There were multiple refinements to the approach but significantly, they were able to identify 68 patients who received megadose CD34⁺ transplants to have achieved engraftment with only 3 patients experiencing > Grade 1 GVHD. Most importantly, cure from leukemia was identified with only ∼20% relapse rate with long-term event free survival beyond 5 years at 30%. Similar to cord blood HCT, infection due to incomplete immune reconstitution remained a major issue contributing both morbidity and mortality to the treated individuals with up to and associated 40% risk of non-relapse mortality [39,40].

Selective T-cell depletion strategies

Megadose CD34⁺ alloHCT in the setting of TCD could be used successfully to overcome haploidentical barriers, but immune deficiency as a consequence of the TCD remained a barrier in this setting. The question was raised whether selective allospecific T-cell depletion could be accomplished, while maintaining T-cell specificity against leukemic minor antigens or infectious pathogens (Table 2). The development of selection devices for sorting cell populations marked an advance in the field, as shown by the BMT CTN 0303 study where elderly subjects with AML in CR1 were shown to have improved outcomes with ex-vivo selection of HSC products for CD34+ progenitors, but also with prescribed T-cell dosing [41,42]. Further refinements have been undertaken but one of the earliest successful studies was performed in elderly patients undergoing HLA-matched donor peripheral blood stem cell transplantation [43]. Specifically, recipient peripheral blood mononuclear cells were collected by leukapheresis and CD3⁺ selection performed to generate a population of host cells (irradiated to prevent proliferation) that could be co-cultured with viable donor cells to stimulate donor allospecific T-cell populations. Donor growth factor mobilized PBSCs underwent CD34 positive selection, and the eluate was collected, recognizing it would be enriched for donor CD3 cells that next were co-cultured ex vivo with the previously collected and irradiated host stimulators. Next, the alloreactive, stimulated donor enriched T-cell populations were treated with two cycles of anti-CD25 immunotoxin, targeting the IL-2 receptor, a known T-cell activation marker. After 72 h of co-culture, the experimentally ex vivo depleted, allospecific T cells were cryopreserved for later infusion. After host transplant conditioning was completed, the recipient received both the selected donor CD34⁺ cells as well as the CD3 enriched and activated population that was CD25 depleted. 15 of 16 patient achieved sustained engraftment with only 12% Grade III/IV aGVHD identified. Non-relapse mortality was 6% at day 100 and 25% at 1 year. Five patients experienced relapse related death including three patients with chemo refractory AML and one patient with treatment related MDS with complex cytogenetics. The feasibility of this approach of ex-vivo selected T-cell depletion was confirmed.

Table 2. Examples of historically studied T-cell depletion strategies.

‘Positive’ CD 34+ selection	‘Negative’ CD3+ depletion
Soybean lectin agglutination followed by E-rosetting – using sheep RBCs to isolate and deplete T cells from the donor hematopoietic stem cell product [11]	Allospecific T-cell depletion with add-back of enriched/activated CD3+/CD25-depleted cells [41]
CliniMACS device selection for CD34+ HSC enrichment [41,42]	In-vivo allospecific T-cell depletion with post-transplantation cyclophosphamide exposure [45]

At this timepoint, there was also the notable emergence of in-vivo selected T-cell depletion providing another approach to decrease catastrophic GVHD without compromising anti-malignancy effect. Historically, since the 1980s, GVHD prophylaxis has been sustained as a calcineurin inhibitor with post-transplant chemotherapy, using the antimetabolite methotrexate to eliminate rapidly proliferating, alloreactive T cells. With this combination, observed rates of Grade II–IV acute GVHD in HLA-matched settings, approximated 30%−50% with related and unrelated procedures [44,45]. In the 7/8, single antigen mismatched setting, it has also been utilized. Results demonstrated increased NRM due to acute GVHD, likely contributing to a GVL effect, with equivalent overall survival as a matched allograft [46]. However, this combination was known to be insufficient for GVHD prophylaxis beyond 7/8 HLA-matched allograft. What remained to be identified is whether a more immunosuppressive cytotoxic agent than methotrexate could be combined with the calcineurin inhibitor and could be more effective for in-vivo allospecific T-cell depletion. This goal was achieved with advent of post-transplant cyclophosphamide (in combination with calcineurin inhibitor and mycophenolate mofetil), based on the work from Johns Hopkins Medical Center, by Luznik et al. [47]. Many of the subsequent contributions within this volume will address aspects of the application of this technology but the rationale for rapid expansion of the unique, and relatively straightforward, GVHD prophylaxis approach includes the fact that there is a high degree of availability of haploidentical donors, with demonstration that both first and second degree relatives can be excellent donors [48]. Increasing identification of patients with germline mutations pre-disposing them to malignancies for which HCT is indicated highlights the importance of having multiple donors to choose from and requires a cautious approach when using family members as donors [49]. Recent studies have demonstrated that stem cell products obtained from younger donors [50] may prove to be a superior product, thus, grandchildren can be used as a stem cell source for alloHCT, recognizing that Mendelian genetics, with predicted independent segregation of the chromosomes, will conserve the integrity of the HLA locus on chromosome 6 through generations. To briefly summarize, the major benefits that are gained by the utilization of post-transplant cyclophosphamide approach include (a) in vivo, allospecific TCD that allows full haplotype-mismatched allograft to be performed as well as fully matched donors, (b) a lower likelihood of developing clinically significant acute GVHD (Grades II–IV), expansion of the donor pool such that nearly all patients in need can identify an available donor, (c) utilization in heavily transfused individuals that are at significant risk of graft rejection such as those with severe aplastic anemia or hemoglobinopathies such as sickle cell anemia or thalassemia and (d) the cost of care of alloHCT can be dramatically reduced, leading to expansion of these technologies in countries with more limited financial resources to offer these haploHCT procedures [51].

AlloHCT remains a viable procedure and is curative for many patients with malignant and nonmalignant hematologic disorders. The cost of care is projected to increase given the multiple applications of adjuvant therapy as well as novel drugs for the management of the associated infectious and noninfectious risks [52]. Approaches that decrease the cost of care and enhance the overall outcome of individuals undergoing SCT remains the goal; the use of the post-transplant cyclophosphamide GVHD prophylaxis strategy meets this goal. However, there is still room for significant improvement and further study.

Remaining problems of post-transplant relapse

Issues that remain to be clarified and improved upon includes the management of relapse. Chromosomal instability is common in patients with acute leukemia and with relapse, clonal evolution has been seen. In an HLA-matched setting, single chromosomal loss (particularly chromosome 6 – where the major histocompatibility complex lies), known as loss of heterozygosity, will not necessarily alter outcomes, as the immune response generated against the leukemia is targeted at peptide antigens, expressed in the peptide binding cleft of the MHC molecule, manifest as a minor histocompatibility antigen recognized by the immune system. In the haploHCT setting, clonal evolution as manifest by chromosomal loss, specifically loss of heterozygosity in MHC gene which are located on chromosome 6 could lead to loss of all the target antigens [53]. If the alloHCT recipient is in need of rescue immunologic interventions, alternative approaches may be needed rather than asking the original donor to provide further mononuclear cell collection to administer a donor leukocyte infusion (DLI) depending on the mechanism of relapse. On that note, in the HLA-matched alloHCT setting, typically a DLI is administered to a patient without active GVHD and with no immune suppression [rev. in 54]. Doses up to 1 × 10⁸ CD3⁺/kg have often been administered as a DLI product. An optimal DLI approach for haploHCT remains to be determined [55]. Finally, recognizing that malignancy typically is a disease of the elderly, and that current research suggests that stem cell products obtained from younger donors may be more optimal, further quality of life and patient reported outcomes studies obtained from very young donors providing product to their parents and grandparents would be welcome to assure that no lasting detrimental emotional burden is generated.

Summary

Haploidentical transplantation is the most rapidly growing transplant procedure in allogeneic HCT [9]. Remarkable advances in biologic understanding has led to the ability to deliver safe and effective transplant procedures to many patients with otherwise fatal malignant and nonmalignant conditions. The development of modalities for ex-vivo and in-vivo TCD have been critical to the success of haplo transplant procedures and have been widely studied both in the haploidentical setting and in other mis-matched and matched donor settings. Ongoing improvements in GVHD prophylaxis, understanding of ideal cell dose and graft modification learned from experience with patients undergoing haploidentical transplants will certainly improve the outcomes for patients undergoing all types of transplant.

Disclosure statement

RTM reports serving as consultant for Autolous, Kite/Gilead and Novartis, research support from Gamida, Allovir, OrcaBio, and Novartis, participating in a DSMB for Athersys, Novartis, Century Therapeutics and VorPharma and a patent with Athersys; no activity has conflict with the material of this article.

RJC reports no conflicts of interest.