Mother and child reunion

While the graft-versus-leukemia effect of T-cell–depleted haploidentical family stem-cell transplantation is considered to be largely based on NK-cell alloreactivity, T-cell alloreactivity might still play an important role. In this issue of Blood, Stern and colleagues show that recipients of haploidentical stem-cell grafts may experience better outcome if the mother rather than the father serves as stem-cell donor, an effect that occurred independent of NK-cell alloreactivity.

Although the probability of finding at least one HLA-A, -B, and -DR–matched donor for a Caucasian patient is currently high (70%-80%), finding a match for HLA-A, -B, -C, -DR, and -DQ is considerably less (35%-40%) and even lower for non-white patients (www.bmdw.org). Recently, the selection of donors with preferably no more than 1 mismatched allele out of 8 was advocated.¹  If that criterion is adhered to and if the number of patients who may progress during the search process is taken into account, only approximately 50% of patients with an indication for unrelated donor allogeneic stem cell transplantation (alloSCT) may ultimately receive the intended allograft.²  That percentage may be even lower for patients with a nonwhite background. Therefore, alternative donors and stem-cell sources are urgently needed, and currently, both haploidentical family donors as well as unrelated cord blood (UCB) stem cell sources are being applied with increasing frequency.

Family mismatched donor alloSCT, as has been developed by the Perugia team,³  is associated with a high rate of engraftment, low incidence of graft-versus-host disease (GVHD), and an event-free survival of approximately 45% to 50% for patients receiving their transplant in remission. The antileukemic activity is based on both an intensified preparatory regimen and the possibility of a donor-versus-recipient natural killer (NK)–cell alloreactivity,⁴  whereby a strong graft-versus-leukemia (GVL) effect may be exerted by NK-mismatched family donors. As with cord blood, haploidentical family donor alloSCT offers the advantage of immediate availability, which compares favorably to the 2 to 3 months needed to identify and prepare for a matched unrelated adult donor.

While the antileukemic effect of haploidentical family donor SCT is predominantly based on an NK-alloreactive effect, the residual low number of donor T cells present in the T cell–depleted allograft may also be involved in the GVL effect. In this issue of Blood, Stern and coauthors present results from a retrospective study evaluating whether donor gender might influence outcome after parental donor haploidentical alloSCT. While reproducing the observation of a lower relapse rate associated with KIR-ligand mismatch, they also observed a lower relapse rate significantly associated with recipients of a maternal graft, independent of NK alloreactivity. This lower relapse rate may result from minor-antigen–specific memory T cells that were induced long ago during pregnancy by transplacental leukocyte trafficking. Minor rather than major HLA-antigen specificity is suggested by absence of more GVHD. These results compare well with earlier observations by Kolb et al, who observed the same phenomenon in a series of recipients of haploidentical stem cell grafts, which were less severely depleted of T cells using anti-CD6 antibodies.⁵  Meanwhile, investigators from the European Group of Blood and Marrow Transplantation (EBMT) have very recently shown that the GVL effect associated with UCB alloSCT may not be restricted to alloreactive T cells, but may also involve KIR-ligand–mismatched NK cells.⁶  Collectively, these results highlight the stronger GVL effect that may be associated with alternative donor alloSCT, but also evoke the urgent question of which alternative donor is to be preferred.

First, based on the results by Stern and Kolb, it may seem reasonable that an NK-cell nonalloreactive haploidentical transplant from a father, brother, or sister should be avoided for recipients with acute leukemia in remission and in need of an alternative donor. Second, if a nonmyeloablative regimen is indicated, such as in older patients, 1 or 2 adequately sized UCB transplantations may be preferred,⁷  due to the higher risk of nonengraftment associated with a full haplotype-mismatched donor. Third, the benefits and risks of UCB alloSCT on the one hand and those from NK-alloreactive or maternal haploidentical alloSCT on the other should now be explored and weighed prospectively, taking into account the newly discovered parameters that affect GVL and nonrelapse mortality (NRM). Thereby, we might be able to better select the preferred alternative donor or stem-cell source, optimally profit from allogeneic immunotherapy, and reduce NRM.