Introduction

Allogeneic hematopoietic stem cell transplantation (HSCT) has completely revolutionized the natural history of several life-threatening or invalidating non-malignant disorders, including primary immune deficiencies (PIDs), bone marrow failure syndromes, and hemoglobinopathies [1, 2]. Over the last decades, the advent of reduced-toxicity conditioning regimens and better infection monitoring and management, more sensitive, molecular-based, tissue typing techniques and advances in supportive care have enormously enhanced the safety and efficacy of HSCT. As the outcome of HSCT improved, the number of non-malignant conditions amenable to definitive treatment by HSCT has continued to grow, placing ever-increasing demands on the pool of stem cell donors [1, 2]. However, only 25% of patients in need of an allograft have an HLA identical sibling, and for less than 70% of the remaining patients, a suitable, HLA-compatible, unrelated donor or umbilical cord blood (UCB) unit can be found [3]. This proportion can be even lower for patients belonging to ethnic groups poorly represented in the registries. Furthermore, the search for an HLA-matched volunteer donor may result in unacceptable delay in certain diseases, such as severe combined immunodeficiency (SCID), for which the goal is to proceed to transplantation as early as possible after diagnosis. In the absence of an HLA-matched donor, HLA-haploidentical relatives are being increasingly used to offer the chance of an allograft to any patient in need of transplantation [3]. Indeed, the majority of patients has a family member, identical for one HLA haplotype and fully mismatched for the other (i.e., HLA-haploidentical), who can immediately serve as HSC donor [4, 5]. Besides availability for almost all patients, transplantation from a full HLA-haplotype-mismatched family member offers other several advantages, among which no delay in graft procurement, the possibility to select the best donor from a panel of candidate members, and easy access to donor-derived cellular therapies whenever required after transplantation. Despite these advantages, widespread use of haploidentical HSCT (haplo-HSCT) has been hampered, for many years, by relevant complications mediated by bidirectional alloreactivity towards mainly incompatible HLA molecules, which was responsible for unacceptably high rates of both graft rejection and severe graft-versus-host disease (GvHD) [6,7,8]. Over the past two decades, an impressive amount of translational immunologic research has resulted in a variety of promising approaches either to regulate GvHD-mediating T cells in vivo or to engineer ex vivo haploidentical grafts [9]. These strategies have yielded encouraging results in haplo-HSCT, with high rates of successful engraftment, effective GvHD control and low transplantation-related mortality (TRM) at the point that, today, this procedure is no longer regarded as a last-resort treatment for otherwise lethal diseases, but is increasingly offered to an ever-widening number of patients with an indication for transplantation who do not have a suitable donor identified within a reasonable time frame. Historically, the great majority of studies regarding haplo-HSCT in non-malignant disorders have focused on T cell depletion (TCD) approaches [10]. Despite that, in recent years, interest in T cell-replete full-haplotype-mismatched HSCT was reawakened by the introduction of new strategies for GvHD prophylaxis [11]. Therefore, in this review of clinical results obtained with haplo-HSCT in non-malignant diseases, we will analyse major advantages and drawbacks of both T cell depleted and unmanipulated HSCT and discuss future challenges for further improving patients’ outcome (Table 1).

Table 1 T cell depleted and T cell-replete haploidentical stem cell transplantation for non malignant disorders: a summary of the most important studies published in the field
Full size table

T Cell-Depleted HLA-Haploidentical HSCT: the Graft Manipulation Breakthrough

Overcoming major HLA barriers hampering transplantation of mismatched grafts remains a major goal for allogeneic HSCT. Since donor-derived T lymphocytes contained in the graft are the major mediators of severe alloreactions in haplo-HSCT, various attempts have been made to overcome the risk of GvHD by depleting T cells from the graft prior to infusion. In 1983, Reisner and colleagues reported the first successful correction of SCID by T cell-depleted haplo-HSCT using differential agglutination with soybean agglutinin (SBA) and subsequent E-rosette depletion (SBA-E-) [12]. The three infants reported did not receive pre-transplant cytoreduction or post-transplant GvHD prophylaxis. All three patients achieved long-term stable donor T cell chimerism with normal T cell function and did not experience either acute or chronic GvHD. The Duke University reported results in 145 patients with SCID transplanted from HLA-haploidentical related donors using bone marrow cells and TCD through SBA-E-. Of these patients, 109 (75%) survived with T cell and 41% with B cell reconstitution. In this series, the incidence of grades III–IV acute GvHD was low (10%) [13]. Experiences with less rigorously T cell-depleted haplo-HSCT employed together with post-transplant pharmacologic GvHD prophylaxis, such as cyclosporine, has been associated with higher incidence of TRM, mainly due to GvHD and/or infections. The proportion of patients achieving long-term survival with reconstitution of at least T cell function has ranged from 56 to 64% in single-center series [14,15,16]. In the SCETIDE report on children with SCID transplanted from 1983 to 1995, the majority of HLA haplo-HSCT performed in European centers using marrow as graft source were depleted of T cells by using SBA-E-. The overall survival (OS) following such grafts was 52%. The main causes of death were infections (56%) and GvHD (25%) [17]. It can now be estimated that hundreds of SCID patients have been transplanted worldwide using a TCD haplo-HSCT, with a high rate of sustained, either partial or complete, immune reconstitution. Despite this remarkable success, subsequent studies revealed that, outside of the SCID setting, graft failure represented a non-negligible problem in TCD transplantation from donors other than HLA-matched siblings, this complication occurring in more than 20% of patients. Unfortunately, graft failure was a direct adverse effect of TCD, as removing T cells decreased the graft-versus-host response and rendered the donor graft more susceptible to rejection by the host immune system.

The introduction of more effective and standardized approaches for TCD based on immune-adsorption to antibody‐coated paramagnetic beads, allowed rapid purification of CD34+ progenitor from granulocyte colony-stimulating factor (G-CSF)-mobilized peripheral blood stem cells (PBSC). Thanks to this strategy, it was possible to obtain high doses of CD34+ progenitors (>10 × 106/kg) to be infused, while limiting T cell doses to <104 CD3+ T cells/kg. The use of such “megadoses” of hematopoietic progenitors represented another milestone in the field of haplo-HSCT, allowing to overcome the barrier of HLA incompatibility in the donor/recipient pair and to elude the residual anti-donor cytotoxic T lymphocyte activity of the recipient [18].

In the SCETIDE report of 186 patients with SCID given haplo-HSCT T cell depleted through this approach since its introduction in 1996, 68% were surviving with at least T cell reconstitution at a median follow-up of 3 years. This result was superimposable to the 68% probability of survival reported by the same group in 66 recipients of marrow grafts from HLA-matched unrelated donors [19]. In the PIDTC study, of 137 HLA-mismatched grafts, 121 (89%) were depleted of T cells by either the lectin technique (SBA-E-) or CD34+ selection. The 5-year OS for this group was 79% for patients who did not receive conditioning and 66% for those who did. In the same study, Pay et al. collected data retrospectively from 240 infants with SCID who had received HSCT in 25 centers between 2000 and 2009. These data indicate that children with SCID diagnosed either at birth or before the onset of infection, who received transplants from mismatched related donors, unrelated donors, or cord blood units soon after diagnosis have more than a 90% probability of survival with T cell and variable B cell immune reconstitution. Moreover, they found that mortality was increased for patients who had active infection at the time of transplantation. For those patients transplanted from a donor other than an HLA-matched sibling, the survival rate was highest among children who received T cell-depleted grafts from a mismatched, related haploidentical donor [20••]. In 2012, Fernandes et al. in a retrospective analysis compared the outcome of 175 and 74 patients with SCID or Omenn syndrome given either haplo-HSCT or UCBT, showing that haplo-HSCT and UCBT are both valid options in SCID patients lacking an HLA-identical sibling donor. In this study, UCBT recipients had a higher frequency of complete donor chimerism (P = 0.04) and faster total lymphocyte count recovery (P = 0.04) without any statistical significance with the preparative regimen they received. The two groups did not differ in terms of T cell engraftment and, despite the fact that UCBT recipients showed a higher total lymphocyte counts during the first year after transplant, CD3+ and CD4+ T cell numbers did not significantly differ at any time point. By contrast, immunoglobulin replacement therapy was discontinued sooner after UCBT. In UCBT recipients, there was a trend toward a higher cumulative incidence of grades II–IV acute GvHD in comparison with haplo-HSCT recipients. Furthemore, the cumulative incidence of chronic GvHD was significantly higher in the UCBT cohort (P = 0.03) [21]. Recently, feasibility of haplo-HSCT has also been demonstrated in 12 pediatric patients affected by Fanconi anemia (FA) who were given T cell-depleted, CD34(+) positively selected cells from a haploidentical related donor after a reduced-intensity, fludarabine (FLU)-based conditioning regimen. Engraftment was achieved in 75% of patients, and the cumulative incidence of graft rejection was 17%. Cumulative incidences of grades II–IV acute and chronic GvHD were 17 and 35%, respectively. The 5-year OS was 83% [22•].

Refinements to T Cell Depletion Strategies: From Positive Selection to Negative Depletion

The main limitation of haplo-HSCT platform based on positive selection of CD34+ cells in the non-malignant setting is represented by delayed immune reconstitution consequent to the physical elimination of T cells from the graft, essential for preventing GvHD occurrence. In fact, with this procedure, recipients cannot benefit from the adoptive transfer of donor memory T lymphocytes, which, through peripheral expansion, are mainly responsible for protection from infections in the first months after transplantation. A profound immune deficiency has been documented to last for at least 4–6 months after haplo-HSCT and to translate into an increased risk of TRM, mainly attributable to severe infections [18, 23, 24].

Over the past two decades, several promising techniques have been proposed to partially or fully overcome the “GvHD-rejection-infection triad” [9]. Some studies have attempted to render donor T cells anergic towards recipient tissues through T cell costimulatory blockade [25]. Other groups proposed regulating GvHD by co-infusing selected or expanded T regulatory cells (T regs) along with controlled doses of conventional T cells (T cons) with the same goal [26]. CD3+/CD19+-depleted haplo-HSCT following reduced-intensity conditioning (RIC) is an alternative option used for children with both malignant and non-malignant diseases. Bader et al. reported very encouraging results in 59 children transplanted with this approach between 2005 and 2011, demonstrating sustained engraftment and a quite rapid immune recovery with low TRM in both malignant and non-malignant patients [27]. Selective depletion of αβ+ T cells and CD19+ B lymphocytes represent a further refinement graft manipulation for haplo-HSCT. Notably, in addition to CD34+ cells, these cell suspensions contain donor mature NK cells and γδ+ T cells, which may provide a first line of defense against different infectious agents, without increasing the risk of both acute and chronic GvHD [28].

In 2014, Bertaina et al. published the first report of children affected by life-threatening non-malignant disorders given an HLA-haploidentical graft manipulated through this innovative approach. More in detail, they treated 23 children with a variety of non-malignant disorders, including SCID, FA, severe aplastic anemia (SAA), osteopetrosis, immunodeficiency with polyendocrinopathy and enteropathy X-linked (IPEX), DOCK8-mutated hyperimmunoglobulin E syndrome, Shwachman-Diamond syndrome, congenital amegakaryocytic thrombocytopenia, primary hemophagocytic lymphohistiocytosis (HLH), and thalassemia major (TM) with autoimmune hemolytic anemia. No patient received post-transplant pharmacologic GvHD prophylaxis. In view of the high OS and disease-free survival (91.1%) coupled with the low incidence of acute GvHD (13.1%) and the absence of chronic GvHD, these pilot data suggest that this novel graft manipulation strategy could represent a suitable treatment option for children with life-threatening diseases lacking an HLA-identical sibling. Although the fatality rate may be lower with this approach compared with earlier methods (the cumulative incidence of TRM being 9.3%), a remarkable rate of rejection (cumulative incidence 16.2%) and a slow recovery of αβ+ T cells still remains [29••]. More recently, Balashov et al. reported on 37 children with primary immunodeficiency transplanted from either a MUD (n = 27) or an halpo-donor (n = 10) after TCR αβ+/CD19+ graft depletion. GvHD prophylaxis consisted of tacrolimus and a short course of methotrexate in 34 patients, tacrolimus and mycophenolate mofetil in 2 patients, and cyclosporine and methotrexate in 1 patient. In the haplo-cohort, the incidence of grades II–IV acute GvHD was 33.3%, while the cumulative incidence (CI) of graft failure was 36%. With a median follow-up period of 392 days (range, 120 to 876), the CI of EFS and OS was 67.7 and 96.7%, respectively. Only one patient died due to transplant-related complications, this leading to a CI of TRM of 3.3% (95% CI, 0.5% to 22%). In this series, CMV reactivation occurred in 16 out of 37 patients, being the CI 46% (95% CI, 32.1 to 66.5%), with no significant difference between haplo and MUD groups [30].

Therefore, strategies capable of speeding up immune reconstitution could produce further significant improvements in outcome of TCD haplo-HSCT. In this perspective, Locatelli et al. recently reported the outcome of children with PIDs given αβ+ T cell-depleted HLA-haplo-HSCT followed by adoptive infusion of donor T cells transduced with the new suicide gene-inducible caspase-9 (iC9), which can be activated by using a dimerizing agent, namely rimiducid (AP1903). In this multicenter phase I/II trial (NCT02065869) were enrolled 18 children affected by SCID, chronic granulomatous disease (CGD), HLH, and Wiskott-Aldrich Syndrome (WAS). Five children developed grade I–II skin-only acute GvHD, one of which progressed to grade III involving the gut and, then, resolved with AP1903 infusion, which successfully activated the suicide gene; no patient developed chronic GvHD. Remarkably, none of the 18 patients included in the analysis had graft failure or died [31••].

These impressive results provide a strong clinical basis to extend the access to HSCT to other non-malignant conditions, whose widespread definitive cure is severely hampered by lack of an HLA-matched donor. Within this subgroup, hemoglobinopathies hold a prominent position, since, despite a remarkable improvement in the life expectancy of patients treated with conservative approaches, allogeneic HSCT still represents the only consolidated option for definitive eradication of the disease. In this setting, however, disease-specific features, such as hyperplastic bone marrow and allosensitization due to multiple transfusions, render haplo-HSCT at high risk of graft failure [32•]. It must be emphasized that first attempts at haplo-HSCT in hemoglobinopathies had been already performed in the era of CD34+ cells positive selection. In 2010, Sodani et al., reported on 22 patients with thalassemia major (TM) receiving an HLA-haploidentical HSCT from the mother. In this study, TCD was achieved through of CD34+ positive selection or CD3+/CD19+-negative depletion. Graft source were both G-CSF mobilized PBSC and bone marrow-harvested mononuclear cells, added with the aim of improving immune tolerance and facilitating engraftment [33]. The pre-transplant protocol consisted of an intensive hypertransfusion regimen combined with iron-chelation from day -59 to day -11 coupled with a ‘preconditioning’ regimen with hydroxyurea and azathioprine, and followed by a BU-based conditioning regimen. All patients received cyclosporine for the first 2 months after transplantation. Six patients experienced graft rejection (cumulative incidence 29%), whereas two patients died of transplant-related complications (cumulative incidence 14%), leading to a DFS of 61% [33]. In a recent update including 31 patients, the same group showed cumulative incidences of graft rejection and TRM of 23 and 7%, respectively, whereas DFS survival improved to 70% [34].

As far as sickle cell disease (SCD) is concerned, eight patients were treated with haplo-HSCT at St. Jude Children’s Research Hospital. In this study, donor graft consisted of CD34+ cells selected from G-CSF-mobilized PBSC; cells were cryopreserved and infused on day 0. The donor underwent a second cycle of G-CSF-primed apheresis to provide a fresh CD3+-depleted product to reach the CD34+ cell target dose of 5 × 106 cells/kg. Thus, the final product consisted of a CD34+ cell-selected product infused on day 0 and a CD3+ cell-depleted product infused on day +1 [35]. A fixed CD3+ T cell dose of approximately 1 × 105 cells/kg was targeted for infusion. After a median follow-up of 7.4 years, six out of the eight patients (75%) were alive. Three out of the eight patients (38%) had sustained engraftment and remained disease-free, leading to a Kaplan-Meier estimates of OS and DFS of 75 and 38%, respectively. Two patients (25%) died due to chronic GvHD, at 0.9 and 2.3 years after HSCT, while graft failure occurred in three patients (38%) [35].

As already mentioned, in 2014, Bertaina et al. reported encouraging results in non-malignant disorders using TCR αβ+/CD19+ negative depletion. In a recent update of this cohort, two patients had TM and severe autoimmune hemolytic anemia. Both children were successfully transplanted from their haploidentical mother [36]. Moreover, in an ongoing phase I/II trial, 5 children with TM and 1 with SCD underwent TCR αβ+/CD19+ TCD haplo-HSCT followed by the add-back of donor T lymphocytes transduced with inducible Caspase 9 suicide gene. In all patients, the conditioning regimen consisted of a combination of BU, Thiotepa, and FLU. No post-transplantation GvHD prophylaxis was administered. All patients were engrafted, reaching full donor chimerism; none experienced either acute or chronic GvHD [31••, 37].

T Cell-Replete Haploidentical Transplantation: a New Old Kid on the Block?

Literature has been largely silent on the use of unmanipulated haplo-HSCT for the treatment of non-malignant disorders. The main reason for this lack of interest must be found in the unacceptable high rates of graft rejection and severe GvHD documented by first attempts with this approach in both the malignant and non-malignant setting [6,7,8]. However, the situation has dramatically changed in the last decade, thanks to the impressive results obtained, mainly, with two different unmanipulated haplo-HSCT strategies in adult patients affected by malignant diseases. Indeed, in adults patients with hematologic malignancies, results of recent retrospective analyses have demonstrated similar survival, GvHD incidence and long-term immune reconstitution between haplo-HSCT and HLA-matched-related or unrelated allogeneic HSCT. On the basis of these extremely encouraging results, T cell-replete haplo-HSCT is increasingly considered as alternative therapeutic strategy for patients with selected non-malignant diseases who do not have a matched sibling or a MUD.

The first unmanipulated approach, pioneered by the Johns Hopkins group, relies on the use of post-transplantation cyclophosphamide (PTCY) [38]. This method takes advantage of the early proliferation of both donor and recipient alloreactive T cells that occurs in the first few days after transplantation. Cyclophosphamide (CY) is given in the window of 72 h after unmanipulated HSCT, causing in vivo depletion of both donor and recipient alloreactive cells, which promotes engraftment and decreases risk of GvHD, while sparing non-alloreactive T cells. Moreover, HSCs are protected from the cytotoxic effects of PTCY thanks to their higher amounts of aldehyde dehydrogenase, an enzyme responsible for metabolizing the drug [39, 40].

In the non-malignant setting, this strategy was employed for the first time for the treatment of two adult patients with hemolytic paroxysmal nocturnal hemoglobinuria (PNH) and 1 with both PNH and SCD, following a non-myeloablative conditioning regimen. Rapid and sustained engraftment without GvHD occurred in two patients, including the one with SCD, while 1 died of fungal sepsis [41]. In 2012, the same group reported the outcome of 14 adult patients with SCD who underwent HSCT from related haploidentical donors. These patients were conditioned with a reduced toxicity regimen of ATG, FLU, CY, and 2 Gy total body irradiation (TBI), while GvHD prophylaxis consisted of PTCY, mycophenolate mofetil (MMF), and tacrolimus or sirolimus. Although this cohort had an OS of 100% at almost 2 years post transplantation, with no documented cases of GvHD, graft rejection was a major problem, being observed in six patients [42].

By contrast, high rates of engraftment were documented, by an English group, in 16 pediatric patients with either SCD (n = 13) or TM (n = 3) who underwent haplo-HSCT using PTCY GvHD prophylaxis. Conditioning regimen was modified with the inclusion of thiotepa in combination with CY and FLU and the source of stem cells was G-CSF-primed BM. All patients engrafted, although one patient experienced graft failure following macrophage activation syndrome and subsequently died. Grades II–IV acute GvHD occurred in two patients and was responsive to treatment with mesenchymal stromal cells (MSCs), while chronic GvHD was observed in one case. After 6 months, all 15 surviving patients but one were off immune suppression with ≥90% donor chimerism both in whole blood and T cells [43].

Recently, a group from Thailand investigated the use of PCTY-based haplo-HSCT in 31 children and young adults (median age 10 years, range 2–20) affected by severe TM and β-thalassemia/hemoglobin E. In this study, all patients received two courses of pre-transplant immunosuppressive therapy (PTIS) with FLU and dexamethasone (Dxm), followed by a conditioning regimen of rabbit ATG, FLU and iv BU. T cell-replete progenitor cells were collected from peripheral blood after G-CSF administration. Twenty-nine patients were engrafted with 100% donor chimerism, while two patients suffered primary graft failure. Nine patients developed grade II acute GvHD, while only five patients developed limited-chronic GvHD. The 2-year OS and EFS were 95 and 94%, respectively [44•].

In 2014, Clay et al. described eight adult patients, four with refractory SAA and four with primary graft failure after unrelated donor HSCT or UCB transplantation. The conditioning regimen consisted of FLU, CY and single-day TBI. Patients underwent HLA-haploidentical, unmanipulated, G-CSF-mobilized PBSCs infusion followed by administration PTCY on days +3 and +4 and post-transplant GvHD prophylaxis with tacrolimus and MMF from day +5. Six patients had sustained donor engraftment, but one of them subsequently died of multiple infectious complications. Graft failure was associated with donor-directed HLA antibodies, despite intensive pre-HSCT desensitization with plasma exchange and rituximab. Only one case of grade II skin GvHD was reported [45].

A Brazilian multicenter study employed a similar strategy, although with higher TBI doses (2–6 Gy), in 16 adult and pediatric patients affected by SAA. The rate of neutrophil and platelet engraftment was 94 and 75%, respectively. Two patients developed secondary graft failure, being successfully rescued with a second HSCT. Grades II–IV acute GvHD was observed in two patients. After a median follow-up period of 355 days, five patients had died because of infectious complications and 1-year OS was 67.1%. Another recent report from India described two children with SAA who underwent successful haploidentical HSCT with PCTY GvHD prophylaxis [46••].

Despite these encouraging results, there are only isolated reports on the successful use of PCTY-based haplo-HSCT in benign disorders other than SAA and hemoglobinopathies, among which FA, HLH, and CGD [47,48,49]. Furthermore, preliminary experience with this approach in malignant infantile osteopetrosis failed to provide consistent engrafted survival [50]. Very recently, the Johns Hopkins group reported the outcome of four pediatric patients with life-threatening non-malignant conditions (Glanzmann’s thrombo-asthenia, IPEX syndrome, XIAP deficiency, and CGD) who underwent HSCT from a haploidentical related donor followed by PTCY in combination with tacrolimus and MMF. The authors reported limited GvHD, no TRM, and successful engraftment sufficient to eliminate manifestations of disease in all patients [50]. One of the major drawbacks of this approach may be related to the limited control of alloreactivity in children below the age of 10 years, which represents a large proportion of HSCT candidates with non-malignant disorders. Indeed, in a recent pilot study PTCY-based haplo-HSCT performed in children with acute leukemia, Jaiswal and colleagues showed that, out of a total of ten grades II–IV acute GvHD cases, severe GvHD occurred exclusively in children below the age of 10 years. This age group also experienced higher incidence of early alloreactivity in the form of hemophagocytic syndrome. Authors hypothesized that the possible cause of intense early alloreactivity in patients under the age of 10 years could be related to the reduced efficacy of PTCY in clearing alloreactive T cells, as a result of the variable metabolism of the drug in this age group [51].

The second unmaninuplated haplo-HSCT strategy, pioneered by the Peking University group, combines myeloablative conditioning, T cell modulation with G-CSF-primed BM and PBSC grafts, ATG, and intensive multiagent GvHD prophylaxis [52]. In the non-malignant setting, this approach has been almost exclusively investigated, with some modifications, for the treatment of refractory SAA. In a first study, the outcome of 19 children and young adults with SAA who received T cell-replete haplo-HSCT was reported. Preparative regimen was based on BU, CY, and ATG, while post‐grafting immunosuppression consisted of the triple regimen of MTX, CsA, and MMF. Two patients (11%) had late graft failure, and three patients did not achieve platelet engraftment. Acute GvHD developed in 42% and chronic GvHD in 56% of evaluable cases. A total of six patients died from transplantation-related causes (GvHD, infection and late graft failure) and the 2‐year OS was 64% [53]. Another study combined such unmanipulated haplo-HSCT approach with growth factor therapy (G-CSF and thrombopoietin mimetics), intensive anti-fungal prophylaxis with micafungin, and high-dose immunoglobulin replacement in 26 adult patients with SAA. Although it remained unclear which of these components were crucial in this study, results were promising, with an engraftment rate of 92.3% and an OS rate was 84.6%. Acute GvHD was observed in 12% patients, while chronic GvHD was reported in 40% of cases (only 1 case with the extensive form) [54]. In the same year, another Chinese group reported T cell-replete haplo-HSCT as salvage therapy in children and adolescents with acquired SAA, after BU, FLU, CY, and ATG conditioning regimen. Sustained neutrophil engraftment was observed in all patients, while a single patient failed to achieve platelet engraftment. Although GvHD prophylaxis with cyclosporine, short-term MTX, and MMF was combined with basiliximab and co-infusion of unrelated umbilical cord blood MSCs, the rate of both acute and chronic GvHD remained high (30.53 ± 11.12%, and 21.25 ± 13.31%, respectively). With a median follow-up of 362 days, 14 patients have survived with an OS rate of 71.60 ± 17% [55]. Very recently, Xu et al. published the results of a prospective, multicenter study performed on 101 adult and pediatric patients with SAA (median age 9 years, range 2–45) who received haplo-HSCT followed by multiagent GvHD prophylaxis with Csa, MMF, and MTX. Ninety-seven patients (96%) achieved myeloid engraftment, while 94 (94.1%) obtained platelet engraftment. With a median follow-up of 18 ± 3 months, haplo-HSCT recipients had a 3-year estimated overall survival (OS) of 89 ± 0% and a failure-free survival (FFS) of 86 ± 8%. These results were similar to those obtained in 48 patients who underwent contemporaneous transplantation from matched related donors (MRD). By contrast, the haplo-HSCT group showed a higher incidence of grades II–IV acute graft-versus-host disease (3 ± 7% vs. 4 ± 2%, P < 0.001) and chronic GVHD (22 ± 4% vs.6 ± 6%, P = 0 ± 014) when compared with the MRD cohort, although there was no difference in the incidence of grades III–IV aGVHD (7 ± 9% vs.2 ± 1%, P = 0.157). In conclusion, although results obtained with this approach in SAA are promising, available data also suggest that this strategy carries a higher incidence of both acute and chronic GvHD when compared with PCTY-based or TCD haplo-HSCT [56].

Conclusions

Available data suggest that T cell depleted haplo-HSCT is a suitable option for the definitive treatment of an ever-widening spectrum of non-malignant disorders, in the absence of an HLA-identical donor. Moreover, while the use of haploidentical donors can extend safe transplantation to virtually all patients in need, the immediate availability of the haploidentical donor allows performing such procedure without undue delay, anticipating the development of life-threatening infections or severe disease-specific organ complications.

The excellent results obtained with TCD haplo-HSCT could challenge, in the near future, the current hierarchical algorithm in which MUD and UCBT are preferred to haploidentical donors. This is particularly true for children undergoing TCR αβ+/CD19+ TCD haplo-HSCT, for which success rates exceeding 90% have been reported [29••]. Moreover, the platform of TCR αβ+/CD19+ TCD haplo-HSCT is amenable of further refinements by the adoptive transfer of donor T lymphocytes transduced with suicide genes, thereby paving the way for even better results.

Alternatively, an unmanipulated haploidentical graft could represent an option in selected, life-threatening conditions. However, T cell-replete approaches employed so far have non-negligible limitations, since preliminary data show limited control of alloreactivity with PTCY in patients under the age of 10 years [51], and the Chinese experience with G-CSF-primed grafts and intensive post-HSCT immunosuppression carries a high incidence of GvHD despite intensive prophylaxis [54,55,56] (Fig. 1). Indeed, in the comparison between TCD- and T cell-replete approaches, it is critical to consider that any risk of GvHD is unacceptable in non-malignant disorders, since it cannot be balanced by a stronger graft-versus-tumor effect as seen in malignant diseases. Moreover, since a large proportion of HSCT candidates with non-malignant disorders is represented by children, the impact of acute and, especially, chronic GvHD may be particularly detrimental.

Fig. 1
figure 1

T cell-depleted and T cell-replete haploidentical HSCT strategies employed in non-malignant diseases. Main advantages and disadvantages of each strategy are reported

Full size image

On the other hand, it may be argued that T cell depletion in conjunction with adoptive transfer of selected T cell populations for accelerating immune reconstitution requires adequate graft manipulation facilities and specialized personnel, and might therefore be more costly. However, although no thorough cost-benefit analysis has been done to date, it would seem that financial issues related to T cell depletion are unfounded since this approach does not envisage GvHD prophylaxis or treatment with expensive immunosuppressive drugs, compared to T cell-replete transplants, which instead require prolonged immune suppression and extended hospitalization [10].

It is important to note that randomized studies have never been conducted to compare T-replete and T-depleted haplo-HSCT, and the majority of clinical data currently gathered for haploidentical transplants come from non-randomized trials with retrospective analysis, making difficult to prove the superiority of one specific method. Despite that, given the impressive results observed in newer approaches of TCD haplo-HSCT, it would not be unwise to postulate that, in the near future, this strategy might become the preferred alternative option for patients with benign disorders without an HLA-identical sibling. Prospective studies comparing TCD haplo-HSCT to other alternative donor sources, such as MUD and UCB, are warranted in the next few years to support more definitive reccomendations.