Post-transplant cyclophosphamide or cell selection in haploidentical allogeneic hematopoietic cell transplantation?

ABSTRACT

Background

One major limitation for broader applicability of allogeneic hematopoietic cell transplantation (allo-HCT) in the past was the lack of HLA-matched histocompatible donors. Preclinical mouse studies using T-cell depleted haploidentical grafts led to an increased interest in the use of ex vivo T-cell depleted (TCD) haploidentical allo-HCT. TCD grafts through negative (T-cell depletion) or positive (CD34+ cell selection) techniques have been investigated to reduce the risk of graft-versus-host disease (GVHD) given the known implications of alloreactive T cells. A more practical approach to deplete alloreactive T cells in vivo using high doses of cyclophosphamide after allografting has proved to be feasible in overcoming the HLA barrier. Such approach has extended allo-HCT feasibility to patients for whom donors could not be found in the past. Nowadays, haploidentical donors represent a common donor source for patients in need of an allo-HCT. The broad application of haploidentical donors became possible by understanding the importance of depleting alloreactive donor T cells to facilitate engraftment and reduce incidence and severity of GVHD. These techniques involve ex vivo graft manipulation or in vivo utilization of pharmacologic agents, notably post-transplant cyclophosphamide (PTCy).

Discussion

While acknowledging that no randomized controlled prospective studies have been yet conducted comparing TCD versus PTCy in haploidentical allo-HCT recipients, there are two advantages that would favor the PTCy, namely ease of application and lower cost. However, emerging data on adverse events associated with PTCy including, but not limited to cardiac associated toxicities or increased incidence of post-allograft infections, and others, are important to recognize.

KEYWORDS:

• Haploidentical transplantation
• post-transplant cyclophosphamide
• cell selection

Introduction

Despite advances in cancer therapeutics over the past decade, allogeneic hematopoietic cell transplantation (allo-HCT) remains an integral treatment modality for various hematologic malignancies, benign hematologic disorders, and immunologic diseases. A limitation for broader applicability of allo-HCT in the past was the lack of human leukocyte antigen (HLA)-histocompatible donors [1]. For instance, an HLA-matched sibling (MSD) or unrelated donor (MUD) was not available for more than one-third of patients in need of the procedure [2]. This led to performing allo-HCT from partially HLA-mismatched, or HLA-haploidentical donors which was complicated by a high incidence of graft rejection, severe acute and/or chronic graft-versus-host disease (GVHD), and a prohibitive non-relapse mortality (NRM) [3–5].

For more than 30 years, attempts to perform haploidentical allo-HCT were faced with significant graft rejection, severe GVHD and delayed engraftment with resulting fatal infections [3,6]. The attempt of non-T-cell depleted haploidentical allo-HCT using bone marrow (BM) grafts resulted in graft failure in 24% of patients and GVHD in 70% [4]. This was thought to be related to high bidirectional T-cell response to allogeneic HLA haplotype not shared between donor and recipient [4,5]. Moreover, studies elucidating GVHD pathogenesis have shown the effect of the conditioning regimen leading to tissue damage within various organs and succeeding inflammatory reactions. Antigen-presenting cells activate donor T cells, which differentiate into effector T cells, including CD4+/CD8+ T cells and Natural Killer (NK) cells. These activated cells mediate tissue damage [7]. Preclinical mouse studies on use of T-cell depleted (TCD) haploidentical graft led to a heightened interest in use of ex vivo TCD haploidentical allo-HCT [8]. The first successful TCD was reported in 1980s using lectin separated haploidentical allo-HCT from mother in patients with severe combined immunodeficiency (SCID) [9,10]. Engraftment and immune reconstitution were achieved in two patients while one only had transient engraftment [9]. The use of haploidentical transplantation has been rejuvenated with the introduction of post-transplant cyclophosphamide (PTCy) in the setting of BM grafts by the Johns Hopkins group [11]. The rate of grade III–IV acute GVHD was remarkably low at 6%; and the rate of extensive chronic GVHD with two doses of PTCy was only 5% [11]. Use of PTCy later expanded to MUD and MSD in the setting of reduced intensity conditioning (RIC) [12].

In this review, we summarize different modalities of ex vivo TCD or the use of PTCy (in vivo TCD) in haploidentical allo-HCT and discuss their current application.

Ex vivo TCD to prevent GVHD in haploidentical allo-HCT

CD34+ selection

Investigation into TCD grafts through negative (T-cell depletion) or positive (CD34+ cell selection) has been pivotal in mitigating the risk of GVHD given the recognized implications of alloreactive T cells [13]. CD34+ selection is commonly performed using immunomagnetic techniques, magnetic cell sorter (CliniMACS) or other systems like dynabeads (Isolex system), both leading to high level of purity and yield [14].

After initial success of haploidentical allo-HCT using TCD BM graft in a SCID patient [9] subsequent efforts focused on extending application of this technique to malignant diseases. The experience of TCD of granulocyte colony-stimulating factor (G-CSF) mobilized peripheral blood stem cells (PBSC) via CD34+-cell selection in haploidentical allo-HCT revealed promising outcomes [15]. In this study, patients received mega doses of CD34+ selected cells (mean, 14.0 × 10⁶ CD34+ cells/kg in 12 patients, and 39.8 × 10⁶ CD34+ cells/kg in 28 patients) resulting in robust and durable engraftment in 41 of 43 patients, with 2 patients experiencing primary graft failure. Still, CD4+ counts were low (100–200 cells/mm³) for a long duration after HCT (10–16 months). Notably, no patients developed acute or chronic GVHD post-transplantation, although one developed severe fatal acute GVHD after donor lymphocyte infusion (DLI) [15]. Leukemia relapses occurred in 13 (30.2%) of 43 patients. Infections were the most common cause of death affecting 65% of patients (n = 11 of 17). NRM was 40% [15]. To address the challenges associated with delayed immune reconstitution, infections and relapse reported with CD34+-selected allo-HCT, Dvorak et al. studied the use of megadose of CD34+ selected (median CD34+ cell dose infused was 22 × 10⁶/kg) PBSC haploidentical grafts with a fixed dose CD3+ T cells (fixed dose of 3 × 10⁴/kg CD3(+) cells/kg) in children (n = 21) with hematologic malignancies [16]. All patients received TBI-based MAC regimens. Infusion of DLI at 12 weeks was considered if CD4+ cells were <0.1 × 10⁹/L. Neutrophils engraftment occurred in 90% of patients at a median of 13 days with 2 patients experiencing primary graft failure. Twelve patients received DLI for low CD4 counts, persistent viral infection, or low level of donor chimerism. All cases of acute GVHD were of low grade (47%) with no reported grade III–IV acute GVHD [16]. Chronic GVHD was reported in 35%, including 4/6 with extensive chronic GVHD. The 2-year relapse incidence was 35%. The NRM was 5% at day 100, and an estimate of 17% at 2 years. The 2-year event-free survival (EFS) and overall survival (OS) were both at 62% [16]. It is important to mention that the purity of CD34+ has improved overtime with the use of more sophisticated CD34+ selection methods (CliniMacs). Also, improved conditioning regimens and GVHD prophylaxis favoring engraftment has led to better results with CD34+ selected grafts. Subsequently, several studies have described the use of CD34+ selected cells following RIC or myeloablative conditioning (MAC) in different donor types [17–22].

Collectively, these findings highlighted the feasibility of TCD haploidentical allo-HCT and its ability to mitigate risk of GVHD. However, they also highlight several critical challenges inherent to this approach, including delayed immune recovery, increased susceptibility for infections, and notably, increased relapse risk. This emphasized the necessity for more sophisticated TCD techniques, particularly using selective TCD, which emerged as focal points for optimizing outcomes among haploidentical allo-HCT recipients.

T-cell receptor αβ+ cells depletion

Building upon the success achieved with megadose CD34+ cell selection, novel techniques were developed to enhance engraftment and immune reconstitution post haploidentical allo-HCT, aimed to reduce infection-related mortality. Preclinical data from animal models suggested that allo-HCT using HLA-mismatched NK cells could improve engraftment by eliminating residual host hematopoietic cells [23]. Therefore, an alternative approach to PBSC graft manipulation emerged thorough CD3+ TCD rather than CD34+ stem cell selection. This method aimed to recruit various accessory cell types (NK cells, monocytes, dendritic cells and early myeloid progenitors) together with CD34+ cells potentially improving engraftment and allowing the use of RIC in haploidentical allo-HCT. Additionally, B-cell depletion was performed, aiming at preventing EBV induced post-transplant lymphoproliferative disorder (PTLD) [24]. Despite demonstrating feasibility and promising results in pediatric patients, this approach showed less promising results in adults with hematologic malignancies [25–31]. The reported 2-year NRM rate of 42%, particularly related to infections (26%), was almost comparable to those described with megadose CD34+ selected transplants [15,31]. In addition, the less intense TCD with this technique might have contributed to a higher rate of acute and chronic GVHD [15,31].

Further preclinical studies in allo-HCT mouse models revealed the implications of donor T-cell receptor (TCR) αβ+ T cells in the pathogenesis of GVHD [32], while γδ+ T cells were found to exhibit minimal alloreactivity [33]. Furthermore, γδ T cells have been shown to correlate with improved immune reconstitution, anti-leukemia effect [34], anti-viral activity [35,36], leading to decreased risk of relapse and improved OS post haploidentical allo-HCT [37]. These reports led to development of a more refined T-cell/B-cell depletion platform using TCR αβ+/CD19+ cells depletion. Compared to CD34+-selection and CD3+/CD19+ cell depletion techniques, TCR αβ+/ CD19+ cells depletion led to similar CD34+ cells yield, yet with a lower T-cell content than CD3+/CD19+ depleted grafts [38]. Early trials have demonstrated feasibility and safety of this approach, mostly in pediatric patients with malignant and non-malignant hematologic disorders [39,40]. Results from the multicenter US phase 2 trial reported encouraging 2-year disease-free survival (DFS) (79%) in pediatric patients with acute leukemia [41]. However, all patients received TCR-αβ/CD19-depleted grafts from haploidentical donors. Rate of graft failure was 7.8%. Immune reconstitution was improved compared to reports with CD34+ selected grafts, with mean CD3 and CD19 levels 554/µl and 339/µl at 6 months. Compared to a matched CIBMTR cohort, there was significant decrease in acute and chronic GVHD in recipients of TCR-αβ/CD19-depleted haploidentical allo-HCT [41]. Less exciting results were reported in a near parallel German trial including children and adults with a 2-year DFS of 50% [42]. Graft failure was reported in 9 (15%) of 60 patients. Acute GVHD was described in 29 of 60 (48%) patients, none being severe grade III–IV GVHD. Chronic GVHD was reported in 15 (31.3%) of 48 patients, 10 being moderate–severe chronic GHVD. CMV reactivation was reported in 46%. CD3+ CD8+ T cells recovered (>100 cells/µl) at a median of 183 days and normalized by 12 months [42]. The perceived advantages of this approach over CD3/CD19 depletion (although direct comparison is lacking) include deeper TCD, potentially contributing to reduced GVHD rates and more robust immune reconstitution, resulting in fewer infection-related mortality. Immunologically, both γδ and NK cells showed early in vivo expansion [43,44]. However, γδ T cells expanded to a greater extent compared to NK cells, possibly due to a specific inhibitory effect by myeloid derived suppressor cells in these grafts [44,45]. Subsequent γδ T-cell subset analysis showed that Vγ1 cells expanded in patients who experienced CMV reactivation and demonstrated an in vitro anti-leukemic effect, while Vγ2 cells recovered earlier and were amenable for phosphoantigen activation (zoledronic acid), with acquired in vitro anti-leukemic effect [44]. Furthermore, patients with higher mature/immature NK cells ratio (CD56 dim/CD56 bright) at day 30 post-HCT were less likely to relapse [43]. Potential disadvantages of this technique include complexities in graft processing (time, labor, and cost), the need for a specialized stem cell lab capable of such depletion, potential need for a second collection (especially in adult recipients) to meet the CD34+ threshold for engraftment, and delayed αβ TCR-cell reconstitution, posing some risk of infections [46–48].

Future directions for this approach may involve confirming its benefit in adults, exploring the use of pharmacologic agents for in vivo innate immune cell activation [49,50] and adding-back of selected donor-derived memory T cells to further prevent viral infections [51].

PTCy in haploidentical allo-HCT

Investigators from Johns Hopkins developed a relatively simple method to selectively deplete alloreactive T cells in vivo by using high doses of cyclophosphamide after allografting. This approach has proved feasibility in overcoming the HLA barrier, facilitating performing the allo-HCT to patients for whom donors could not be found in the past.

A phase-1 clinical trial was conducted to assess efficacy of the PTCy platform in 13 patients, median age of 53 (23–61) years, with various high-risk hematologic malignancies [52]. The major objective was to determine the minimum dose of pre-transplantation cyclophosphamide for prevention of allograft rejection that permits the stable engraftment of partially HLA-mismatched marrow (up to 3 HLA antigens) from first-degree relatives [52]. The first cohort of three patients received a nonmyeloablative regimen consisting of fludarabine 30 mg/m²/day (days −6 to −2) and total body irradiation (TBI) of 2 Gy (day −1), without pre-transplantation cyclophosphamide. Post-allograft immune suppression consisted of cyclophosphamide at a dose of 50 mg/kg on day +3, mycophenolate mofetil (MMF) (day +4 to day +35), and tacrolimus (from day +4 to day +50 or beyond) [52]. G-CSF was prescribed at a dose of 5 µg/kg/day subcutaneously starting on day +1 and continuing until neutrophil recovery to >1000/µL for 3 consecutive days. Unfortunately, rejection was reported in 2 patients. As a result, 10 patients in the second treatment cohort received pre-transplantation cyclophosphamide (14.5 mg/kg/day, days –6 and –5) [52]. Sustained donor cell engraftment was reported in 8 of 10 patients, with donor chimerism of >90% on the first reported measurement. Authors reported that subsequent chimerism measurements remained >90% except in the 2 patients who developed BM and hematologic relapse [52]. Median time to a neutrophil count >500/µL was 15 days, suggesting that high-dose PTCy can be administered safely, with G-CSF support, without apparent impairment in hematologic recovery [52].

A subsequent two-center study by the Johns Hopkins and the Fred Hutchinson Cancer Research Center (FHCRC) groups described outcomes of 68 patients, median age of 46 (1–71) years, mostly allografted for hematologic malignancies (acute myeloid leukemia = 40%) using the PTCy platform [11]. All patients were intended to be treated in the outpatient setting with a regimen consisting of intravenous cyclophosphamide 14.5 mg/kg/day (days −6 and −5), intravenous fludarabine 30 mg/m²/day (days −6 to −2) and 200 cGy of TBI (day −1) [11]. Intravenous cyclophosphamide was administered at a dose of 50 mg/kg on day +3 (n = 28 patients from FHCRC) or on days +3 and +4 (n = 40 patients from Johns Hopkins); GVHD prophylaxis with tacrolimus and MMF was initiated the day following PTCy completion [11]. In this study, G-CSF was prescribed starting on day +4 and continuing until the recovery of neutrophils to >1000/µL for 3 consecutive days [11]. The authors reported a median time to neutrophil recovery of 15 days, and a median time to platelet recovery of 24 days. Graft rejection was reported in 9 (13.6%) of 66 evaluable patients. Reported outcomes showed probabilities of grades II–IV and III–IV acute GVHD by day +200 of 34% and 6%, respectively, with no significant difference in the probability of acute GVHD between patients receiving 1 versus 2 doses of PTCy [11]. However, the 1-year incidence of extensive chronic GVHD was significantly lower in those receiving 2 doses of PTCy (5% vs. 25%, hazard ratio [HR] = 0.21 (95% confidence interval (CI) = 0.04–1.01); p = 0.05). The NRM was 4% at day 100 and 15% at 1 year [11]. This study confirmed the feasibility of PTCy in allowing allo-HCT from partially matched-related donors.

Haploidentical transplantation has been compared, albeit non-randomized, to MUD donor allo-HCT in AML patients using data from the Center for International Blood and Marrow Transplant Research (CIBMTR) [53]. Haploidentical allograft recipients received GVHD prophylaxis with a PTCy, calcineurin inhibitor, and MMF while recipients of unrelated donor allo-HCT received a calcineurin inhibitor with MMF and methotrexate [53]. Following MAC regimens, acute grade III–IV GVHD was lower in the haploidentical allo-HCT versus the unrelated donor allo-HCT cohort (7% vs. 13%, p < 0.02) [53]. The 3-year chronic GVHD (30% vs. 53%, p < 0.0001) favored haploidentical transplantation [53]. The authors reported comparable 3-year OS in haploidentical (45%) and unrelated donor allo-HCT recipients (50%), p = 0.38. The 3-year NRM was similar in both groups (14% vs. 20%, p = 0.14) [53].

Additional registry studies from CIBMTR also compared haploidentical transplantation to HLA-matched unrelated donors in allo-HCT recipients with lymphoma [54]. This study evaluated adult patients with lymphoma who received a haploidentical (n = 185) or an HLA-matched unrelated donor (URD) transplantation either with (n = 241) or without anti-thymocyte globulin (ATG; n = 491) following RIC regimen [54]. The 3-year OS was comparable between the three groups (60% vs. 62% vs. 50%) [54]. No difference was reported between the three groups in terms of NRM, relapse/progression, and progression-free survival (PFS). The NRM was higher in URD with ATG group compared to haploidentical group (17% vs. 26%, p = 0.02), but not different between other groups (overall p = 0.08) [54]. RIC haploidentical transplantation with PTCy did not compromise early survival outcomes compared with HLA-matched unrelated allo-HCT. Notably, the haploidentical group had significantly lower 1-year rates of chronic GVHD (13% vs. 51%, vs. 33%, p < 0.001) [54].

PTCy in PBSC allografting

Acknowledging that PBSCs have become the preferred cell source for allo-HCT, owing to easier procurement, it became imperative to demonstrate feasibility of the PTCy platform when using PBSCs. Solomon et al. conducted a trial of haploidentical T-cell replete allografting in 20 adults, median age of 44 (25–56) years, using PBSCs, following a busulfan-based MAC regimen and PTCy [55]. Donor engraftment occurred in all 20 patients, with a median time-to-neutrophil and platelet recovery of 16 (14–21) and 27 (16–56) days, respectively [55]. There were no reported late graft failure; and authors reported fast achievement of full donor chimerism with all evaluable patients achieving durable complete donor T-cell and myeloid chimerism by day 130 [55]. The incidence of grades II–IV and III–IV acute GVHD was 30% and 10%, respectively [55]. After a median follow-up of 20 months, 1-year OS and DFS were 69% and 50%, respectively [55]. This study demonstrated the feasibility of using PBSCs with PTCy in patients without a conventional HLA-matched-related or unrelated donor [55].

Solomon et al., in another phase II trial of PBSCs haploidentical allo-HCT, used TBI (1200 cGy)/fludaranine-based MAC with PTCY, MMF and tacrolimus in 30 patients (median age 46) [56]. The study showed sustained complete donor T cell and myeloid chimerism by day +30 and grades II–IV and III–IV acute GVHD in 43% and 23%, respectively [56]. The authors reported a cumulative incidence of all grade chronic GVHD of 56% (severe = 10%) [56]. With a median follow-up of 24 months, 2-year OS, DFS, NRM, and relapse rates were 78%, 73%, 3%, and 24%, respectively [56].

The Blood and Marrow Transplant Clinical Trials Network (BMT CTN) recently reported results of a randomized, multicenter, phase-3 trial that compared two regimens for GVHD prophylaxis in patients receiving RIC or nonmyeloablative allo-HCT using PBSCs in HLA-matched donors (8/8 and 7/8) [12]. The experimental arm consisted of MMF, tacrolimus and PTCy; the control arm consisted of tacrolimus plus methotrexate [12]. Results showed superior GVHD-free, relapse-free survival, likely resulting from a lower incidence of grade III–IV acute GVHD and chronic GVHD [12].

A recent observational registry study from the Acute Leukemia Working Party (ALWP) of the European Society for Blood and Marrow Transplantation (EBMT) compared haploidentical donor versus HLA-matched unrelated donor allo-HCT in patients in need of a second allo-HCT for the purpose of treating relapsed AML [57]. Median OS was not different between the haploidentical groups and the HLA-matched unrelated donor groups (11 vs. 10 months, p = 0.57) [57]. Another study from the ALWP of EBMT comparing haploidentical donor versus HLA-matched unrelated donor allo-HCT in patients in need of a second allo-HCT for treating relapsed acute lymphoblastic leukemia did not show OS difference when using a haploidentical or an HLA-matched unrelated donor in this setting [58]. In our opinion, the more immediate availability of a haploidentical donor provides an advantage compared to approaching an HLA-matched unrelated donor, through established registries, to perform the second allo-HCT.

Reducing dose of PTCy

While there is considerable enthusiasm for the use of PTCy, it is important to acknowledge its toxicities, mainly related to increased risk of infections (CMV, fungal, bacterial), hemorrhagic cystitis, cardiotoxicity, delayed count recovery [59–62]. In a retrospective study, Duléry et al. compared outcomes of patients undergoing allo-HCT who received PTCy to those who did not receive PTCy for GVHD prevention [62]. The use of PTCy was associated with a lower incidence of acute grade II–IV GVHD (22% vs. 33%, p = 0.042) but no difference in grade III–IV acute GVHD (12% vs. 7%, p = 0.14). However, patients who received PTCy had increased early cardiac events post-HCT compared to those who did not receive PTCy (p = 0.001), translating into worse OS [62]. This led to an increased interest in decreasing dose of PTCy while maintaining GVHD prevention.

Murine allo-HCT models indicated that an intermediate PTCy dose of 25 mg/kg/day on days +3 and +4 was superior at preventing severe GVHD and reducing mortality compared with 50 mg/kg/day on days +3 and +4 [63]. McAdams et al. reported results of a phase 1/2 study of reduced dosing of PTCy after HLA-haploidentical bone marrow transplantation testing 25 mg/kg/day on days +3 and +4 (experimental dose level 1) and 25 mg/kg on day +4 only (dose level 2), followed by a phase 2 expansion cohort at the better experimental dose level [64]. The first 5 patients were treated with PTCy 50 mg/kg/day on days +3 and +4 (standard dosing) for comparative data (control). All received MAC with intravenous busulfan and fludarabine [64]. The two-day dosing PTCy (25 mg/kg/day) showed more consistent early engraftment and protection against protracted engraftment fevers compared with one-day dosing PTCy (day +4 only) [64]. Moreover, median time-to-neutrophil engraftment was shorter with experimental dose level 1 vis-à-vis standard dosing (14 vs. 19 days, p = 0.0004). Median platelet engraftment was also shorter with experimental dose level 1 versus standard dosing (23 vs. 33 days, p = 0.026). This meant that transfusion requirements were lower with the experimental dose level 1[64]. Ongoing trials are currently evaluating lower doses of PTCy in the allo-HCT setting (NCT05436418, NCT05622318).

Combining PTCy with other agents

To further decrease the incidence of GVHD, PTCy was combined with other agents as a multifaced strategy targeting different aspects of the graft-versus-host immune response. Rabbit anti-thymocyte globulin (rATG) is produced by immunizing rabbits with fresh human thymocytes derived from cardiac surgery donors. The final product is purified and contains immunoglobulins directed at various thymus cellular components, including B and T lymphocytes, stromal and antigen-presenting cells [65]. It is used for the treatment of various diseases and for the prevention of GVHD and graft rejection in allo-HCT through its TCD effect and a paradoxical expansion in regulatory T cells (CD4 + CD25 + Foxp3+) [66,67], among others [68]. Although there was an initial enthusiasm for the use of ATG in allo-HCT, this enthusiasm soon declined with the advent of calcineurin inhibitors in 1980s. Presently, it is used in PBSC allo-HCT in the setting of MSD, MUD, and MMUD [69]. Although several protocols were developed in China using ATG in haploidentical allo-HCT, its utilization has not been widely adopted, owing to concerns of high relapse risk and infections [70]. One study has previously reported outcomes of patients with hematologic malignancies who received haplo-HCT with full dose PTCy and ATG. ATG was prescribed at 5 mg/kg (67%) or 2.5 mg/kg (33%). The rate of bacterial infections was similar to reports of haploidentical allo-HCT with PTCy without ATG (40–62%) [71]. Authors reported high rates of viral reactivation, including CMV (day 100 = 56%) and EBV (day 100 = 53%), and BK virus cystitis (day 100 = 31%). Cumulative incidence of grade III–IV acute GVHD (15%) and extensive chronic GVHD (12%) was low. The addition of ATG did not appear to increase the risk of relapse. Duléry et al. compared outcomes of patients undergoing haploidentical allo-HCT who received low-dose PTCy (70 mg/kg) (n = 33) to PTCy (100 mg/kg) (n = 25) [72]. All patients in both arms received ATG. Authors reported a shorter time-to-neutrophil (16 vs. 19 days, p = 0.006) and platelets recovery (90 days, 91% vs. 64%, p = 0.07), a lower incidence of bacterial infections (38% vs. 72%, p = 0.004) and less frequent cardiac events (12% vs. 44%, p = 0.028) [72]. There was no difference in the incidence of grade III–IV acute GVHD between the groups (0% vs. 5%, p = 0.23) [72]. Wang et al. reported combining ATG to low-dose PTCy in T-cell replete haploidentical allo-HCT [73]. PTCy was given at 14.5 mg/kg on days +3 and +4. Results were compared to ATG alone cohort. All patients received MAC. Low-dose PTCy in combination with ATG significantly decreased risk of grade III–IV acute GVHD (5% vs. 18%; p = 0.003) and 2-year incidence of chronic GVHD (30 vs. 44%; p = 0.07). Median time to myeloid and 100-day platelets recovery were significantly longer in the PTCy arm (15 vs. 12 days; P < 0.001; 90% vs. 97%; p = 0.003). There was improved DFS in the PTCy arm, however, no difference in relapse or OS. These results show that the combination of ATG and PTCy could be promising owing at better control of GVHD, without increasing risk of relapse, and possibly better engraftment.

Among novel regimens using the PTCy platform involved the addition of abatacept to standard dose PTCy in haploidentical allo-HCT aiming at further reducing incidence of GVHD [74]. Abatacept is a cytotoxic T-cell lymphocyte-4-immunoglobulin that blocks T-lymphocyte costimulation (CD80/CD86-CD28) decreasing the risk of GVHD [75]. The rate of grade III–IV acute GVHD was 4.4%; and the 1-year incidence of moderate/severe chronic GVHD was 15.9% without reported increased rates of infections or relapse [74]. Other combinations have been also evaluated in preclinical settings, namely the use of α-galactosylceramide (α-GC) ligand of invariant natural killer T (iNKT) cells showing the enhancement of graft versus leukemia effect without increasing GVHD [76].

Comparing PTCy versus ex vivo TCD

The Blood and Marrow Transplant Clinical Trial Network (BMT CTN) 1301 trial was a phase-3 randomized, multicenter open label trial which compared three approaches among recipients of an HLA-matched graft and using MAC regimens [22]. Patients were randomly assigned to one of three specified interventions: (a) ex vivo CD34 selected T-cell depleted PBSC graft without additional immunosuppression (target CD34+ cell dose of ≥5 × 10⁶ cells/kg), (b) unmanipulated BM graft followed by Cy 50 mg/kg on days +3 and +4 post-allo-HCT, or (c) a control arm of unmanipulated BM graft with tacrolimus (started on days −3) and methotrexate (15 mg/m² intravenously on day +1 and 10 mg/m² IV on days +3, +6, and +11) [22]. Primary endpoint was a composite of moderate to severe chronic GVHD, disease relapse, and survival (GRFS). The authors reported an intent-to-treat rates of 2-year GRFS of 50.6% for CD34+ selection (hazard ratio [HR] vs. control, 0.80 [95%CI = 0.56–1.15; p = 0.24]), 48.1% for PTCy (HR = 0.86 [95%CI = 0.61–1.23; p = 0.41]), and 41.0% for control. Similarly, corresponding OS rates were 60.1% (HR = 1.74 [95%CI = 1.09–2.80; p = 0.02]), 76.2% (HR = 1.02 [95%CI = 0.60–1.72; p = 0.95]), and 76.1%. The authors describe that CD34+ selection was associated with lower incidence of moderate/severe chronic GVHD (HR = 0.25 [95%CI = 0.12–0.52; p = 0.02]) but at the expense of higher NRM (HR = 2.76 [95%CI = 1.26–6.06; p = 0.01]). In the case of PTCy, the authors described comparable chronic GVHD and survival outcomes to control, and lower disease relapse (HR = 0.52 [95%CI = 0.28–0.96; p = 0.037]) [22]. Results of BMT CTN 1301 did not report superiority of calcineurin-free interventions with PTCy or CD34+ selection after HLA-matched allografting. Also, the improvement of chronic GVHD prevention did not result in better survival [22].

Another single center study compared PTCy and TCD (through CD3+ negative selection by CliniMacs) [77]. GVHD prophylaxis was cyclosporine (CSA) in TCD and PTCy/CsA/MMF in the PTCy groups. The median time to neutrophil engraftment was 15 days in the PTCy group and 10 days in the TCD group (p = 0.0009) [77]. No graft failure was reported in the PTCy group but two patients failed to engraft in the TCD group and two other had secondary graft failure. In the PTCy group, three patients died before day +30, due to hepatic sinusoidal obstructive syndrome. CD4+ immune reconstitution was greater in the PTCy group compared to the TCD group (p = 0.041). Rates of grade III–IV acute GVHD were 5% and 7% and severe cGVHD were 12% and 9% in the PTCy and TCD groups, respectively [77]. The cumulative incidence of NRM at day 100 and 1 year were 18% and 24% in the TCD group and 22% and 26% in the PTCy group, respectively. Comparable 2-year OS and cumulative incidence of relapse were reported in the two groups (p = 0.917 and p = 0.485) [77]. This study shows that both modalities are comparable in relation to survival and relapse outcomes, however, TCD was associated with poor engraftment and decreased immune reconstitution.

Another study by the MD Anderson group included 65 patients who received TCD graft (n = 33) or T-cell replete (TCR) graft with PTCy, tacrolimus, MMF (n = 32) [47]. The stem cell source was mainly bone marrow in the TCR group and PBSCs in the TCD group. There was no difference in the rate of neutrophil engraftment between the two groups (p = 0.10) [47]. One patient in the TCD group developed graft failure. The cumulative incidence of grade III–IV acute GVHD was 5% and 9% for the TCR and TCD groups (p = 0.59), respectively. The 1-year chronic GVHD rate was 7% in the TCR group compared to 18% in the TCD group (p = 0.03), none being extensive chronic GVHD [47]. The NRM was lower in the TCR compared to the TCD group (1-year 0% vs. 67%, p = 0.001). Similar rates of relapse were reported in both groups (1-year: 18% TCR, 8% TCD, p = 0.9). The 1-year OS was improved in the TCR group (92%) compared to the TCD group (33%) (p = 0.02). Contrary to the previous study, this analysis reported worse overall survival in the TCD deplete group mostly resulting from a significantly higher NRM. It is difficult to draw a solid conclusion regarding the rates of chronic GVHD between TCD and TCR grafts given difference in stem cell source between in each of the groups.

To our knowledge, there is no randomized trial addressing these questions in the setting of haploidentical allo-HCT.

Discussion

Both TCD (negative depletion of TCR αβ+ T-cell depletion or positive selection of CD34+ cells) and PTCy platforms have broadened the applicability of allo-HCT to patients who do not have a suitable MSD or MUD donors available. These approaches have allowed performing the procedure in patients for whom a haploidentical donor is the only available option. This has benefited minorities for whom donors are less likely to found in major registries, hence narrowing the allo-HCT disparity gap.

Comparing the different TCD techniques is challenging due the heterogeneity among studies, including different patient populations of malignant and non-malignant diseases, variability across patient age, and different conditioning regimen intensities. While acknowledging that no randomized controlled prospective studies have been yet conducted comparing ex vivo TCD versus PTCy in haploidentical allo-HCT recipients, there is no difference in survival after haploidentical allo-HCT using these various techniques and extent of GVHD prevention. While CD34+ selected grafts are associated with graft failure and poor immune reconstitution, αβ+ T-cell depletion seems to offer acceptable rate of GVHD prevention with no or shorter immune suppression post-transplant, shorter time-to-neutrophil and platelet recovery [77,78], and more robust early innate immune cell recovery [46,79,80]. There is no definitive answer pertaining to comparative risk of infection between the different procedures. Yet, there are several aspects of the allo-HCT procedure that would favor the PTCy. For instance, PTCy represents a simpler approach from the practical standpoint and is less expensive when compared to costs associated with ex vivo TCD [81]. These two advantages of using the PTCy approach are contributing to widespread the availability of haploidentical transplantation, particularly in developing countries where financial resources are more limited; and sophisticated technology required for cell separation might not be readily available. Yet, emerging data on adverse events associated with PTCy including, but not limited to, cardiac toxicities or increased incidence of post-allograft infections, and others, are important to recognize. Efforts to reduce the dose of PTCy have demonstrated to be feasible. A longer follow-up will help decipher whether dose reduction(s) would translate into a lower incidence of toxicities, cardiac or others.

Authors contributions

RM, ZA-K, MAK-D designed, wrote the first draft, edited and approved the final version of the review.

Disclosure statement

RM and ZAK have no competing conflicts of interest to declare. MAK-D declares grant/research from Novartis, Bristol Myers Squibb and Pharmacyclics and Consultancy from Kite Pharma.