Chapter 15: Alternative Donor Transplants: Haploidentical Hematopoietic Stem Cell Transplantation

Haploidentical stem cell transplantation (haploSCT) from a first-degree-related haplotype-mismatched donor (siblings, children, parents) could expand allogeneic stem cell transplantation (SCT) to a large proportion of patients with hematologic malignancies without an HLA-matched donor (¹). As the average family size continues to shrink, the likelihood of finding an HLA-matched related sibling donor continues to decrease (²). Moreover, as the population continues to age, finding a young, healthy sibling donor becomes increasingly less likely. The use of matched unrelated donors (MUDs) is limited by the long time to SCT (median 3-4 months), which makes it difficult to treat patients with more advanced disease in rapid need of SCT. The ethnicity/race of the recipient can also limit MUD transplantation as approximately 30% of Caucasians, 70% of Hispanics, and 90% of African Americans do not have a MUD in the worldwide registries (³).

In contrast to unrelated donor stem cells, haploidentical (or “half-matched”) donors can be available immediately, and there are no costs associated with an unrelated donor search, maintaining a registry, or coordinating logistics with distant donor centers. This is an especially valuable option for the non-Caucasian and mixed-race individuals (³). This approach might also be particularly useful in developing countries that may not have the resources to procure unrelated donor transplants or maintain complex unrelated donor registries. Moreover, haploidentical donors offer the possibility to easily collect donor cells for cellular therapy posttransplant. Over the recent decade, significant breakthrough advances in controlling alloreactivity have been made and important steps taken toward graft engineering and posttransplant cellular therapy, approaches that changed dramatically the landscape of haploSCT. Improved haploidentical transplant outcomes represent a major advance in SCT that has practically eliminated the limitation of donor availability for allogeneic SCT.

Historically, unmanipulated T-cell-replete haploSCT grafts with conventional graft-versus-host-disease (GVHD) prophylaxis used in the late 1970s were associated with intense bidirectional alloreactivity and unacceptably high morbidity and mortality rates due to hyperacute GVHD and graft rejection (^4,5,6). This led in the 1980s to the development of complete ex vivo depletion of T cells using CD34-selected grafts. Complete T-cell depletion has been associated with a lower incidence of acute GVHD (aGVHD); however, this caused delayed immune recovery and was associated with a high nonrelapse mortality (NRM) from infections and higher disease relapse rates given the decreased graft-versus-leukemia effect, as well as a higher rate of graft rejection (^7,8,9). While graft rejection was partially overcome with “megadoses” of CD34 cells (typically >10⁷ CD34⁺ cells/kg) and a myeloablative conditioning regimen (including total-body irradiation [TBI], cyclophosphamide, thiotepa) with severe T-cell depletion of the graft, immune recovery remained delayed, leading to high NRM rates in excess of 40% (¹⁰). Improved results with this approach have been reported in some centers with selective depletion of alpha-beta T cells. T-cell depletion strategies have been most successful in children (¹¹). We have used this approach with a different conditioning regimen (fludarabine, melphalan, and thiotepa) and showed that most patients died of NRM related to infectious complications (¹²).

During the last decade, several advances enabled investigators to selectively deplete alloreactive T cells and successfully control GVHD rates while maintaining memory T cells in the graft to accelerate immune recovery and prevent significant infectious complications posttransplant. Table 15-1 summarizes the major contemporaneous approaches to haploSCT. These advances have improved significantly outcomes of patients treated with haploidentical donors, with outcomes now similar to matched transplants (Fig. 15-1) (^12a). Here, we summarize the recent developments with this type of transplant, focusing on advances made at the MD Anderson Cancer Center (MDACC).

The introduction of high-dose posttransplant cyclophosphamide (HDPTCy) for GVHD prevention represented a major turning point for haploSCT. The concept of inducing immune tolerance with posttransplant cyclophosphamide was introduced by Berenbaum and Brown in 1963, showing that the life of a skin allograft can be prolonged with the use of HDPTCy administered 1 to 3 days after the graft (¹³). Mayumi et al demonstrated that microchimerism and robust tolerance to minor histocompatibility antigens can be achieved in mice receiving allogeneic splenic cells by intraperitoneal high-dose cyclophosphamide administered on day 2 or 3 posttransplant (¹⁴). This concept found its best applicability in allogeneic transplantation, particularly in haploidentical transplantation, where HDPTCy induces bidirectional immune tolerance by selectively eliminating the highly dividing alloreactive donor and recipient T cells generated early posttransplant in the setting of a human leukocyte antigen (HLA) mismatched transplant, with decreased rates of GVHD and graft rejection (¹⁵). HDPTCy spares stem cells (due to high levels of aldehyde dehydrogenase present in the cells), which reconstitute the recipient’s hematopoiesis (¹⁶), and nondividing T cells, including memory T cells. This results in a more rapid immune recovery compared to T-cell-depleted approaches, leading to lower NRM (lower rates of infections) compared with T-cell-depleted haploSCT, as shown by our group (¹⁷).

We have used HDPTCy since 2009, soon after the first human trials showed the safety of this approach (¹⁸). Initial studies used a nonmyeloablative conditioning regimen with fludarabine, cyclophosphamide, and 2-Gy TBI, which was associated with a low incidence of grades 2 to 4 aGVHD (35%) and NRM (15%) at 1 year. However, a higher relapse rate was observed (¹⁸). We then hypothesized that more intense conditioning is needed and can be tolerated, especially for patients with leukemia, and used our melphalan-based conditioning regimen (with fludarabine 120 mg/m², melphalan 100 to 140 mg/m² with thiotepa 5 to 10 mg/kg [subsequently changed to 2-Gy TBI]) previously used in T-cell-depleted haploSCT, which had been effective in inducing remission in most patients with leukemia even with advanced disease (^12,17). Updated results for the first 100 patients treated with this regimen showed 3-year PFS rates of 56% to 62% for patients with myeloid malignancies (acute myelocytic leukemia [AML] in complete remission 1/2, myelodysplastic syndrome and chronic myelocytic leukemia in chronic phase) and lymphoma, and 1-year NRM rates of 12% and 22%, respectively (¹⁹).

With haploSCT, outcomes improved significantly; we and others have subsequently compared transplant outcomes of patients treated with a haploidentical versus a matched related donor or a MUD (^20,21,22). These single-institution studies uniformly showed similar outcomes with haploidentical transplant with HDPTCy and HLA-matched donor SCT (^20,21,22). To confirm these findings, we compared outcomes of haploidentical with MUD transplants using the Center for International Blood and Marrow Transplant Research (CIBMTR) database. This retrospective analysis of 2,174 patients with AML showed similar 3-year PFS with haploidentical transplants performed with HDPTCy and MUD transplants (41% vs 42% for myeloablative, P = .87; 35% vs 37% for reduced-intensity conditioning [RIC], P = .89, respectively) (see Fig. 15-1). The incidence of grades 2 to 4 aGVHD was lower for haploidentical compared with MUD transplants (21% vs 42% for myeloablative, 25% vs 35% for RIC) (²³). Our group has proposed a large prospective multicenter study comparing transplant outcomes using haploidentical and MUDs to the Bone and Marrow Transplant Clinical Trials Network group.

Several research strategies are being investigated to optimize the haploidentical graft, maximize immune recovery, and minimize GVHD posttransplant. One promising approach is changing complete T-cell depletion to selective depletion of alpha/beta T cells capable of eliciting GVHD while preserving in the graft memory T cells and gamma-delta T cells (¹¹). It is currently thought that the gamma-delta T cells possess innate and adaptive immune responses and can function without requiring antigen processing or HLA presentation, making them unlikely to generate GVHD (²⁴). Methods to deplete alpha-beta T cells and leaving the gamma-delta subsets intact are being investigated, with encouraging results (²⁵). Other novel approaches involve depletion of naïve T cells (CD45RA⁺) thought to play a major role in the development of GVHD in mouse models (^26,27,28) or administration of T-regulatory cells along with the T-cell-depleted graft. These may further reduce the risk of aGVHD and reduce the rate of relapse (²⁹). Future trials will explore these approaches at MDACC, as they may control GVHD and facilitate rapid immune recovery without posttransplant immunosuppression.

With significant improvements in NRM, disease relapse remains the major cause of death in patients undergoing haploSCT. Several approaches are under investigation at our institution to prevent and treat disease relapse posttransplant (Table 15-2).

Unmodified Donor Lymphocyte Infusion

The readily available haploidentical donors can be sources of posttransplant donor lymphocyte infusion (DLI) administered to prevent or treat early relapse. There is a theoretical higher risk of inducing severe aGVHD with haploidentical DLI; however, the incidence of GVHD was not higher than in matched transplants, possibly due to the tolerizing effect of HDPTCy (³⁰). Among 40 patients with hematological malignancies relapsed after a haploSCT who received unmodified haploidentical DLI (1 × 10⁶/kg CD3⁺ T cells), aGVHD was noted in 25% (grades III-IV aGVHD in 15%). A third of patients achieved a complete response with a median duration of response of 12 months. Most patients received cytoreductive therapy prior to the DLI infusion. Thus, cellular therapy with haploidentical DLI can be effective posttransplant, and future approaches should improve the safety and efficacy of DLI.

Modified Donor Lymphocyte Infusion Using T Cells With a Safety Switch (Suicide Gene)

One approach to control aGVHD post-DLI would be to insert a suicide gene in the haploidentical donor T cells. If significant aGVHD occurs, a “safety off switch” can “turn off” these T cells and avoid excessive aGVHD. This approach has so far been investigated to boost posttransplant immune recovery after T-cell-depleted haploidentical grafts. Ciceri et al infused DLI engineered T cells to express herpes simplex virus–thymidine kinase suicide gene (can be triggered by ganciclovir to induce apoptosis) (³¹). The aim was to boost posttransplant immune reconstitution by adding back T cells after a complete or partial T-cell-depleted haploSCT. Grades 1 to 4 aGVHD developed in 20% of patients and was successfully terminated by inducing the suicide gene with ganciclovir. However, ganciclovir may not be the ideal drug in this setting given that it is commonly used posttransplantation to control cytomegalovirus reactivation. Another approach is to use DLI engineered to express an inducible caspase-9 transgene (³²). This gene can be induced by a synthetic dimerizing drug, leading to rapid cell death. In a preliminary experience in ten patients, aGVHD developed in five patients and was rapidly reversed with the use of the dimerizing drug.

Chimeric Antigen Receptor T Cells

While the DLI offers nonspecific antitumor activity, the effect is nontargeted. A potential game changer has been the introduction of T cells engineered to express a chimeric receptor, with an extracellular domain that can recognize a specific antigen and an intracellular domain that can activate the cytotoxic T cell. This approach has demonstrated significant activity in tumors expressing CD19, such as acute lymphoblastic leukemia (ALL) or B-cell lymphomas. Maude et al used autologous chimeric antigen receptor (CAR) T cells against CD19 (CTL019) in 30 patients with relapsed-refractory ALL, and complete remission occurred in 90% of patients. They demonstrated that CTL019 cells proliferated in vivo and were detectable in the blood, bone marrow, and cerebrospinal fluid of patients who responded (³³). We are exploring the use of CAR T cells early after haploidentical transplantation to prevent disease relapse, part of a multiarm clinical trial (³⁴). Our center is the only center so far using CAR T cells after haploidentical transplantation. Four patients (three with ALL and one with diffuse large B-cell lymphoma) received CAR T cells generated using the Sleeping Beauty system. The lymphoma patient achieved remission for the first time after transplant and infusion of CAR T cells. Three of these four patients remained in remission at last follow-up. These results are promising and showed that allogeneic CAR T-cell therapy can be safely given without significant GVHD.

Natural Killer Cells and Killer Immunoglobulin-Like Receptor Mismatch

Natural killer (NK) cells are part of the innate immune system and normally are involved in identifying and killing tumor cells or virally infected cells. The NK cells recognize and target “foreign” cells that lack one or more HLA class I alleles specific to the inhibitory receptors (killer immunoglobulin-like receptors, KIRs) (³⁵). The NK cells do not contribute to GVHD as they target hematopoietic cells sparing other body organs, making them ideal in the transplant setting. This was first observed in the T-cell-depleted setting, where patients with a KIR “mismatch” had a lower incidence of relapse (³⁶). There is great interest in this field to identify haploSCT donors with a KIR mismatch to possibly maximize the graft-versus-tumor effects. Several studies suggested a lower risk of relapse with donors who possess specific activating KIR genes, such as KIR2DS1, KIR2DS2, or the KIR “B” haplotype (^37,38,39). We are exploring infusion of ex vivo expanded NK cells using the mb-IL21 method developed at MDACC in haploidentical transplants (protocol 2012-0708) to prevent disease relapse posttransplant in patients with myeloid malignancies (⁴⁰).

Multiple donors may be available to choose from in haploSCT. Several factors are considered when choosing the best donor (⁴¹). One of the most important factors to evaluate in these patients is the presence of anti-HLA donor-specific antibodies (DSAs) (^42,43). Patients may develop antibodies against foreign HLA antigens, particularly parous women or multiply transfused patients. We have shown for the first time that patients with high levels of complement fixing anti-HLA antibodies against donor HLA antigens are at high risk of graft failure (⁴³).

Routine evaluation of all donors with HLA mismatches has now been incorporated into standard practice worldwide. Research is currently ongoing to optimize donor selection to improve outcomes (⁴⁴). In general, younger donors are preferred. There is controversy whether the parent gender is important if a parental donor is needed. Some studies suggested using donors selected for maximal NK cell alloreactivity to maximize the graft-versus-malignancy effects. This includes selecting donors with KIR:KIR ligand mismatch or using KIR B haplotype donors (enriched for activating KIR) to exploit the NK cell alloreactivity and decrease relapse rate posttransplant (³⁷).

The field of haploidentical stem cell transplantation has advanced significantly over the past decade with the introduction of HDPTCy and novel methods of partial T-cell depletion. These newer techniques effectively control strong alloreactive reactions in haploSCT and are associated with robust immune recovery, translating into fewer infections and lower NRM. Data from multiple retrospective studies suggested that outcomes are now comparable to MUD SCT, and this type of transplant is expanding worldwide (²³). Controlled clinical trials are needed to address whether haploidentical transplants are preferred over unrelated donors, at least in some clinical settings. Cellular therapy posttransplant represents a great opportunity to further modulate GVHD and graft-versus-malignancy effects. Future studies will prospectively compare haploSCT to other alternative donor sources and the incorporation of cellular therapy in the treatment of these patients.