Allogeneic haematopoietic cell transplantation (allo-HCT) has a bi-faceted role in the treatment of haematopoietic malignancies. First, allo-HCT gives high-dose chemotherapy a chance to reduce the leukaemic burden; second, it allows graft-originated natural killer and T cells to initiate an adoptive immunotherapy effect against leukaemic cells via tumour-specific antigens, tumour-associated antigens and minor histocompatibility antigens.1–6 Allo-HCT plays a major role in the graft-versus-leukaemia (GVL) effect on residual leukaemic cells.7–9 The GVL effect can be clinically observed after donor lymphocyte infusions. In chronic myeloid leukaemia, approximately 70–90% of patients can achieve complete remission after donor lymphocyte infusion.10,11 Although the GVL effect is mainly potent in chronic myeloid leukaemia and acute myeloid leukaemia (AML), it is restricted by graft-versus-host disease (GVHD).11–13 While it can be difficult to drive the immune reaction in a GVL direction rather than a GVHD direction, Johnson et al. established a method to promote the GVL effect without triggering GVHD in an animal study.14
The state of art in allo-HCT is based on balancing GVHD and the GVL effect. Selecting the best donor, prudent use of the immunosuppressive drugs for GVHD prevention and treatment, rapid engraftment, minimal toxicity with individualized conditioning regimens and supportive care are the fundamental elements to achieve this objective. In the last decade, with significant progress, allo-HCT has become a rational and curative option.15,16 The purpose of this review is to summarize the recent advances and paradigm shifts in allo-HCT.
Changing from human leucocyte antigen analysis to functional human leucocyte antigen allele matching
The identification of human leucocyte antigens (HLAs) to find a matched donor is the foremost step in allo-HCT. HLA-dependent interactions between lymphocytes and antigen-presenting cells play a fundamental role in bidirectional graft–host interplays.17 This relationship may lead to either graft rejection or GVHD. The traditional low-resolution method to identify HLA proteins was serological analysis using antigen-specific anti-sera. In the last two decades, HLA typing evolved from serological analysis to HLA polymorphism assays. Polymerase chain reaction (PCR) with sequence-specific oligonucleotides and sequence-specific primers was introduced, which revealed further details of amino acid differences.18,19 Currently, the Sanger sequencing method is accepted as the gold standard for HLA polymorphism analysis with the advantage of high-resolution tissue typing.20
Although Sanger sequencing and sequence-specific oligonucleotide methods are the elemental methods for HLA typing, next-generation sequencing (NGS) techniques have emerged in recent years, which provide more sensitive, faster, and more cost-effective information on HLA typing.21–26 HLA typing with nanopore sequencing also offers more accurate throughputs in a faster manner than NGS methods such as Ion Torrent (Thermo Fisher Scientific, Waltham, MA, USA) and MiSeq (Illumina, San Diego, CA, USA).27 Another method that can be used for HLA typing is RNA sequencing.26 This method supplies precise HLA typing results as well as killer immunoglobulin-like receptor mismatch, which may prevent relapse after allo-HCT.28,29 An 8/8 match for HLA-A, -B, -C or -DR loci has been preferred for related donors. Only one-third of the patients may have a fully matched sibling donor;30 therefore, clinicians refer to alternative donor sources, such as matched unrelated donors, haploidentical or umbilical cord blood donors.
The National Bone Marrow Program and the Center for International Blood and Marrow Transplant Research (CIBMTR) recommend DNA-based methods at high-resolution for HLA-A, HLA-B, HLA-C, HLA-DRB1 and HLA-DPB1 loci. HLA-DQB1, HLA-DRB3/4/5, HLA-DQA1 and HLA-DPA1 typing are underlined as a helpful strategy to avoid graft failure risk via determining possible HLA-sensitized patients, but not for routine testing for donor selections even if no impact on survival has been shown.31,32 HLA mismatches trigger T cell alloreactivity in distinct levels determined by T cell-epitope (TCE) mismatching.33 In recent years, in silico modelling has been developed to test functional HLA-DPB1 matching, such as TCE analysis (TCE3 and TCE4) and a scoring system named ‘delta functional distance’ derived from DPB1 TCE algorithm.34 These attempts are the first steps to achieve a significant breakthrough in the HLA typing paradigm.
The functional matching of HLA may contribute significantly to the counterbalance between GVL and GVHD. Consistently using these methods may improve allo-HCT outcomes in terms of overall survival (OS), non-relapse mortality, GVHD-free/relapse-free survival and risk of relapse.34–37
Unrelated matched donor, unrelated cord blood donor or haploidentical related donor: which is better?
Despite the advances in allo-HCT from alternative donor sources, HLA-identical sibling donor transplantation is considered the gold standard.38–41 In the USA, the probability of a 10/10 HLA matched sibling donor ranges from 13–51%.42,43 Sibling match probability is 1.5 times lower in adults younger than 44 years compared with the older ones.42,43 In this regard, finding an alternative donor is indispensable. In a prospective study, Yakoub-Agha et al. showed that unrelated donor HCT has comparable outcomes with matched sibling donor HCT in terms of OS and non-relapse mortality.44 There are also several encouraging studies with retrospective analyses that corroborate the comparable outcomes of unrelated donor HCT and matched sibling donor HCT.45–47 Recently, however, unrelated donor allo-HCT rates have diminished, being replaced by haploidentical donor HCT.48,49 Access to an unrelated donor requires more time and funds.50,51 Donor search costs, transportation expenditure of cell product and increased use of anti-thymocyte globulin and defibrotide increase the cost of allo-HCT. Donor volunteering and the rapid availability of cell products in case of necessity during and after transplantation are the advantages of haploidentical transplantation.
Effective GVHD prophylaxis (post-transplant cyclophosphamide, anti-thymocyte globulin) and successful selection of the best donor have made haploidentical transplantation more preferable in recent years. Although there are no prospective randomized controlled trials, a retrospective cohort study in patients with AML has shown that haploidentical HCT has comparable outcomes with unrelated donor HCT, umbilical cord blood transplantation (UCBT) and even with matched sibling donor HCT.52,53 Even though haploidentical donors are not the first-line donor source for lymphomas, haploidentical transplantation with post-transplantation cyclophosphamide has been shown to result in similar outcomes as unrelated donor HCT and matched sibling donor HCT in Hodgkin lymphoma.54,55 For patients with thalassemia, the availability of a related donor is relatively low, as the donor candidate may suffer the same disease.
Allo-HCT from alternative donors, such as matched unrelated donors and haploidentical donors, is becoming increasingly available with each passing day.56,57 There is also a growing body of evidence to support haploidentical HCT in severe aplastic anaemia.58–60 A meta-analysis performed by the European Society for Blood and Marrow Transplantation (EBMT) Working Party, demonstrates that haploidentical HCT can be a feasible option for patients with severe aplastic anemia with acceptable engraftment success (engraftment rate 97.3%, 95% confidence interval [CI] 95.9–98.7), with relatively reduced complications (annual transplant-related mortality [TRM] rate 6.7%, 95% CI 4.0–9.4).61 Although in recent years, chimeric antigen receptor T (CAR T) cell therapies have provide promising results in non-Hodgkin lymphoma, it is not plausible to say that allo-HCT will be superseded by CAR T-cell therapy.62,63
Data on the role of haploidentical HCT are scarce in patients with non-Hodgkin lymphoma compared to AML and Hodgkin lymphoma. Kanate et al. also showed that post-transplant cyclophosphamide-based haploidentical HCT in non-Hodgkin lymphoma has comparable OS and non-relapse mortality outcomes in addition to lower chronic GVHD rates.64,65 Recently, Dreger et al. conducted a large-scale retrospective analysis to compare haploidentical HCT with unrelated donor HCT and matched sibling donor HCT in diffuse large B cell lymphoma. There was no significant difference between the groups in terms of OS, non-relapse mortality, and progression-free survival (PFS), but a lower incidence of chronic GVHD was seen in the haploidentical HCT group.64
UCBT can be an alternative donor source in case of unrelated donor unavailability. However, the number of UCBT procedures has decreased in recent years.50,66 The advantages of UCBT are that there is no need for HLA full match donor, the availability of rare HLA types, no need to perform harvest, and lower GVHD rates even in higher HLA mismatch.30,67,68 The disadvantages include an increase in the possibility of graft failure and delay in immune reconstitution as a result of low CD34+ counts, especially for adult patients. Moreover, post-thaw stem-cell counts can be lowered to within the range of 1–25%.69 Novel techniques are being tested for improving engraftment and immune reconstitution in UCBT. UCB ex vivo stem-cell expansion with numerous small molecules (StemRegenin 1, TAT-BMI1, nicotinamide, valproic acid, sitagliptin) have been introduced.70–74 Hypoxia culturing, co-culture with mesenchymal stem cells, co-infusion with mesenchymal stem cell, double UCBT and co-infusion with third-party selected CD34+ haploidentical stem cells are other techniques that can enhance outcomes of UCBT.75–78 However, further confirmatory clinical trials are required for these promising stem cell expansion methods. Due to the lack of randomized controlled trials comparing haploidentical HCT, unrelated donor transplant and UCBT, regional factors and the experience of the transplant centres mostly come to the fore in the preference of donor sources.
Conditioning regimens for allogeneic haematopoietic stem cell transplantation
When choosing a conditioning regimen, the key factors are the type and state of the disease as well as the patient’s comorbidities. Although myeloablative regimens have been preferred in young, fit patients, reduced-intensity conditioning (RIC) regimens have been introduced for elderly, unfit patients with comorbidities. The consideration of the GVL effect of allo-HCT has been the rational basis of RIC regimens.79 The optimal conditioning regimen for patients older than 50 years remains controversial due to the lack of prospective randomized controlled studies.80,81
In young patients with AML or myelodysplastic syndrome, myeloablative regimens have been compared with RIC regimens. A phase III randomized controlled study showed better, but not statistically significant, OS at 18 months with myeloablative regimens compared to RIC regimens (77.5 versus 67.7%, respectively; p=0.07). Higher relapse rates were observed in RIC regimens (67.8% and 47.3%, respectively; p<0.01), and TRM was 15.8% (n=22) in the myeloablative group and 4.4% (n=8) in the RIC group (p=0.02).82 In another study, the RIC and myeloablative regimens showed no significant differences in terms of OS (31% versus 24%, respectively; p=0.10), 10-year non-relapse mortality (16% versus 26%, respectively; p=0.10), and 10-year disease-free survival (55% versus 43%, respectively; p=0.19) in patients with intermediate and high-risk AML between 18 and 60 years old.83
In elderly patients with AML, promising long-term outcomes with more widely used regimens, like fludarabine or busulfan, have been shown. Recent data have shown more favourable outcomes in terms of PFS and GVHD-free/relapse-free survival with fludarabine 160 mg/m2 + melphalan 100 mg/m2 when compared with fludarabine 160 mg/m2 + melphalan 140 mg/m2, fludarabine (with or without clofarabine) 160 mg/m2 + busulfan x 4 days (area under curve [AUC] ≥5,000 µmol/min/day), and fludarabine (with or without clofarabine) 160 mg/m2 + busulfan x 4 days (AUC 4,000 µmol/min/day). Non-relapse mortality at 3 years was 19%, 39%, 35% and 21%, respectively (p=0.06), and 3-year relapse rates were 32%, 32% 30%, and 55%, respectively (p=0.003).84 A study that compared total body irradiation (TBI)-based and busulfan-based conditioning in AML showed no significant difference between the treatment groups in terms of the 2-year OS (64% versus 84%, p=0.73), PFS (52% versus 53%, p=0.96), RFS (76% versus 84%, p=0.95), and non-relapse mortality (27% versus 47%, p=0.56).85 Clift et al. compared busulfan + cyclophosphamide with cyclophosphamide + TBI as a conditioning regimen in AML.86 There was no difference between study arms in terms of OS, PFS and non-relapse mortality.
In recent years, the body of literature has increased concerning minimal/measurable residual disease (MRD), which can be determined with multiparametric flow cytometry, PCR or NGS techniques. However, it is not possible to perform molecular MRD analysis for around 60% of patients with AML, due to a lack of suitable markers for real-time PCR.87 In order to address this limitation, Thol et al. developed an NGS MRD technique that was able to identify a suitable marker in 93% of patients sequenced at the time of diagnosis.88 Patients who have morphologic complete remission after AML treatment, but ongoing MRD positivity before allo-HCT, have lower OS and PFS outcomes compared with patients who achieve MRD negativity.89,90
MRD assessment before transplantation can also help determine which patients may benefit from allo-HCT. The GIMEMA AML1310 study was conducted to examine the clinical outcome of risk-adapted post-remission therapy based on MRD assessment with PCR and multiparametric flow cytometry. In this study, MRD-driven transplantation improved the OS and DFS outcomes of intermediate-risk patients to levels similar to those of favourable risk.91 Shah et al. conducted a study to examine the MRD analysis after allo-HCT in patients with AML. The study showed that detection of MRD positivity in the early post-transplantation phase portends a higher relapse risk.92 Although allo-HCT can be a curative option for AML, the relapse incidence and mortality rates remain high, making it reasonable to seek an optimal post-transplant maintenance approach.
The RELAZA2 study is a phase II, single-arm study set out to clarify the outcomes of azacitidine maintenance in patients with myelodysplastic syndrome or AML with MRD positivity after allo-HCT or conventional chemotherapy. In this study, 58% (95% CI 44–72) of patients who were MRD-positive were relapse-free with azacitidine maintenance (p<0.0001).93 In another major study, SORMAIN, which set out to examine the outcomes of sorafenib maintenance after allo-HCT in FLT3-positive patients with AML with or without MRD positivity, showed that sorafenib maintenance significantly reduced the risk of death and relapse (hazard ratio [HR] 0.39; 95% CI 0.18–0.85).94 Furthermore, Schlenk et al. demonstrated that single-agent maintenance of midostaurin in patients with FLT3-positive AML after allo-HCT is safe and feasible. In this study, starting the midostaurin maintenance at day 100 post-transplant had better event-free survival and OS outcomes compared to within 100 days after transplant.95 An ongoing phase III randomized, placebo-controlled multicentre study is being conducted to test the outcomes of gilteritinib maintenance after allo-HCT in FLT3-positive patients with AML, the results are not yet published (ClinicalTrials.gov Identifier: NCT02997202).96
RIC regimens also can be feasible in older patients with acute lymphoblastic leukaemia (ALL). In a retrospective study, Mohty et al. found that there was no association between the type of conditioning regimen (myeloablative versus RIC) and leukaemia-free survival (p=0.23, HR=0.84).97 Marks et al. compared full-intensity and RIC regimens in older patients with ALL; they showed that the intensity of the conditioning regimen did not affect outcomes of TRM (p=0.92) and relapse risk (p=0.14).98
Although TBI has been accepted as the key component of conditioning regimens in ALL, rather than AML, acute and delayed toxicity of TBI remains as a substantial challenge. CIBMTR data showed a higher relapse rate (38% versus 27%, respectively; p=0.007) but similar survival rates (57% versus 53%, respectively; p=0.35) with busulfan-based conditioning regimens compared with TBI-based regimens in ALL.99 In a Korean study, TBI-based conditioning regimens showed favourable outcomes in young adults independent of acute leukaemia subtype (p=0.005).100 In patients younger than 40 years, TBI provides better 5-year OS (55.1%, p=0.023) and disease-free survival (DFS) (48.6%, p=0.020) outcomes than busulfan-based conditioning regimens.101 The EBMT Acute Leukaemia Working Party showed that the combination of TBI with etoposide is as effective as combination with cyclophosphamide (HR 0.62, p=0.04).102 In a retrospective analysis from Japan, intravenous busulfan + cyclophosphamide had comparable outcomes with TBI-based regimens in ALL.103 RIC regimens in ALL have increased in recent years and mostly formed from chemotherapeutics only.104
Recent studies confirm that allo-HCT provides a clinical benefit in terms of relapse risk and survival in patients with Philadelphia chromosome (Ph)-negative and Ph-positive ALL who have ongoing MRD positivity after induction chemotherapy.105,106 Dhédin et al. demonstrated that allo-HCT improved relapse-free survival in patients with Ph-negative B-cell precursor ALL who had post-induction MRD positivity.107 A retrospective Chinese study, in patients with Ph-positive ALL, found that pre-transplantation MRD positivity resulted in a higher cumulative incidence of relapse compared with MRD negativity (26.1% versus 12.1%, p=0.009); however, non-relapse mortality, OS and leukaemia-free survival outcomes were comparable.108 Although MRD is a broadly accepted key prognostic indicator in B-cell ALL, it is understudied in T-cell ALL. Modvig et al. analyzed the effect of flow cytometry-based MRD assessment on the outcomes of patients with T-cell ALL. They showed that patients who have flow cytometry-based MRD ≥10-3 at the end of the induction have a higher HR for relapse compared to flow cytometry-based MRD ≤10-3.106 There has been increasing convincing evidence to substantiate MRD-guided allo-HCT decision-making.
The long-term outcomes of RIC, myeloablative and non-myeloablative conditioning regimens have been compared in large B-cell lymphoma, and results showed that rates of GVHD were similar between the groups. Higher 5-year non-relapse mortality was found in the myeloablative regimen compared to RIC and non-myeloablative (56% versus 47% versus 36%, respectively; p=0.007) and 5-year relapse/progression was lower in the myeloablative regimen compared to RIC and non-myeloablative (26% versus 38% versus 40%, respectively; p=0.031). There were no significant differences in survival at years 1, 3 and 5 between the myeloablative, RIC and non-myeloablative regimens (respective results at 1 year: 38% versus 46% versus 45%; 3 years: 21% versus 27% versus 29%; and 5 years: 18% versus 20% versus 26%).109
In RIC regimens, the importance of lymphoma histology on the outcomes of HCT has been emphasized in studies. Patients with indolent non-Hodgkin lymphoma who underwent allo-HCT with RIC regimens have been found to have favourable outcomes (HR 0.4, p=0.045) compared to those with Hodgkin lymphoma and aggressive non-Hodgkin lymphoma.110,111 A systematic review/meta-analysis in chronic lymphocytic leukaemia, showed that RIC regimens are more commonly used, despite a lack of randomized studies in myeloablative and RIC regimens.112 The Chronic Lymphocytic Leukaemia Working Party of EBMT reported no difference in 5-year OS, event-free survival and non-relapse mortality outcomes of non-myeloablative and RIC regimens (RIC: 46%, 38%, 35%; and non-myeloablative: 52%, 43%, 32%, respectively).113
The incidence of allo-HCT has decreased since the introduction of tyrosine kinase inhibitors for chronic myeloid leukaemia.114 Despite the excellent outcomes of tyrosine kinase inhibitors in chronic myeloid leukaemia, allo-HCT holds its place in the treatment of the multiple tyrosine kinase-inhibitor-resistant conditions. RIC regimens have been preferred due to the strong GVL effect in chronic myeloid leukaemia.115,116 A study based on the CIBMTR database shows that there is no significant difference in adult patients with chronic myeloid leukaemia who underwent allo-HCT with myeloablative and RIC regimens in terms of OS, non-relapse mortality and leukaemia-free survival. However, RIC regimens provided lower chronic GVHD rates (HR 0.77; p=0.02).117
Despite of the fact that allo-HCT is associated with substantial treatment-related morbidity and mortality, it remains the only curative therapeutic option for primary and secondary myelofibrosis.118 Patients with intermediate-2 and high-risk myelofibrosis score according to the International Prognostic Scoring System can be a candidate for allo-HCT. Fludarabine-based RIC regimens have been widely used with promising event-free survival and OS rates.119 In a study that compared busulfan + fludarabine, fludarabine + melphalan and fludarabine + BCNU (bis-chlorethyl-nitroso-urea) + melphalan-based RIC regimens in patients with myelofibrosis, there were no differences between the treatment arms in terms of survival (p=0.65), grade 2–4 acute GVHD (47%, 68% and 68%, respectively; p=0.31), relapse (29%, 14% and 9%, respectively; p=0.21), and non-relapse mortality (29%, 41% and 27%, respectively; p=0.32).120 Kröger et al. analyzed the transplantation outcomes of patients with primary and secondary myelofibrosis who underwent allo-HCT with a busulfan (10 mg/kg) + fludarabine (180 mg/m2)-based RIC regimen.121 In this prospective phase II trial, the RIC regimen showed 51% 5-year event-free survival and 67% OS. The results also showed that age >55 years (HR 2.7; p=0.02) and HLA mismatch (HR 3.04; p=0.006) are risk factors for survival.
In severe aplastic anaemia, allo-HCT from a matched sibling donor is the first-line treatment choice in patients younger than 50 years old. The standard conditioning regimen for matched sibling donor transplantation is cyclophosphamide (200 mg/kg) and anti-thymocyte globulin.122 In immunosuppressive therapy-refractory patients who are older than 30 years, fludarabine can be a less toxic alternative option to cyclophosphamide.123 Maury et al. demonstrated survival benefit with a fludarabine-based conditioning regimen compared with cyclophosphamide-based conditioning (p=0.04).124 The survival benefit might be related to lower graft failure rates (0% versus 11%, p=0.09) in the fludarabine arm.
In thalassemia major, busulfan + cyclophosphamide-based conditioning regimens have been accepted as the gold standard for matched sibling donor allo-HCT, for decades.125 Disease-specific conditions, such as haemochromatosis secondary to multiple transfusions, smooth the way for the hepatotoxicity of the busulfan + cyclophosphamide regimen. Mathews et al. demonstrated that a treosulfan-based regimen (treosulfan + thiotepa + fludarabine) reduced the incidence of sinusoidal obstruction syndrome (78% to 30%; p=0.000) and TRM (46% to 13%; p=0.005) in high-risk patients with thalassemia major.126
In primary immunodeficiency diseases, conditioning regimens may not be needed if matched sibling bone marrow sources are preferred.127,128 Dvorak et al. demonstrated that unrelated donor transplantation can be performed without a conditioning regimen, and showed comparable OS (71% versus 92%, p<0.01), acute GVHD (50% versus 39%, p<0.01), chronic GVHD (22% versus 5%, p<0.01), and event-free survival outcomes with matched sibling donor HCT.129 The Inborn Errors Working Party recommends fludarabine + busulfan or fludarabine + treosulfan-based conditioning regimens.130
Despite its effectiveness, physicians search for alternatives to TBI because of secondary malignancies and organ damage, such as pneumonitis, infertility and veno-occlusive disease.131 In this regard, targeted radiotherapy via radiolabeled monoclonal antibodies can spare the vital organs and give the advantage of potentiating the effective radiotherapy dosing. In a recent phase I study, Vo et al. reported the efficacy of allo-HCT with yttrium-90-labeled CD45-targeted radiotherapy-based RIC in elderly patients with active leukaemia or myelodysplastic syndrome.132 The yttrium-90-labeled CD45-targeted radiotherapy-based combined conditioning regimen resulted in complete remission for 87% of the patients. One-year relapse rate and 2-year OS were 41% and 46%, respectively. Bouabdallahet al. conducted a phase II clinical trial to investigate the effects of yttrium-90 ibritumomab tiuxetan as part of RIC allo-HCT in high-risk patients with non-Hodgkin lymphoma. In this study, both 2-year OS and event-free survival were 80%.133
Graft-versus-host disease prophylaxis
GVHD is one of the major complications of allo-HCT. Traditional GVHD prophylaxis with calcineurin inhibitors has been insufficient, especially since the introduction of new allo-HCT modalities.134 Approximately 13 years ago, Storb et al. showed that cyclosporine and methotrexate are associated with lower acute GVHD rates (33% versus 54%, p=0.014) compared with cyclosporine alone.135 To date, calcineurin-based GVHD prophylaxis is widely accepted as the gold standard.
Although there are no prospective randomized studies, several researchers have attempted modalities of calcineurin-free GVHD prophylaxis.136–138 A phase II prospective study, conducted by Bejanyan et al., compared cyclosporine + mycophenolate mofetil combination with sirolimus + mycophenolate mofetil combination.138 There were no significant influences on the risk of acute GVHD, OS and PFS with calcineurin-free prophylaxis; however, there were significant reductions in thrombotic microangiopathy and infection rates. Recently, a meta-analysis that evaluated cyclosporine + methotrexate versus tacrolimus + methotrexate showed that tacrolimus + methotrexate resulted in lower grade 2–4 acute GVHD (odds ratio 0.42; p<0.0001), chronic GVHD (odds ratio 0.79; p=0.015), non-relapse mortality (odds ratio 0.62; p=0.03), and improved OS (odds ratio 1.3; p<0.0001) outcomes.139
Mycophenolate mofetil can also be combined with calcineurin inhibitors instead of methotrexate, especially in non-myeloablative conditioning regimens.140 Abatacept, which is an inhibitory protein of T cell co-stimulation, was tested in a phase I trial for GVHD prophylaxis in combination with tacrolimus and methotrexate. Combination of abatacept for GVHD prophylaxis was found safe and the rate of grade 2–4 acute GVHD was 28.6% at day 100.141 Similarly, a retrospective study set out to determine whether abatacept + cyclosporine + methylprednisolone is effective compared to standard GVHD prophylaxis with tacrolimus + mycophenolate mofetil in patients with beta-thalassemia. The study showed that none of the patients in the abatacept combination group experienced grade 3/4 acute GVHD (0% versus 50%, p=0.001) or graft failure.142
To investigate the role of CXCR5 blocking with maraviroc in GHVD prophylaxis, Reshef et al. carried out a phase II trial which demonstrated that grade 3/4 acute GVHD at day 180 and 1-year moderate to severe chronic GVHD were 5 ± 4% and 8 ± 5%, respectively.143 A meta-analysis indicated that mesenchymal stromal cells in the GVHD prophylaxis setting improved OS rate and decreased the grade 4 acute GVHD.144 A prospective study evaluated the combination of ruxolitinib and post-transplant cyclophosphamide for GVHD prophylaxis in patients with primary or secondary myelofibrosis. The rate of acute GVHD grade 2–4 and grade 3/4 were 25% and 15%, respectively. From a total of 20 patients, 11 (55%) experienced severe poor graft function and one patient primary graft failure.145 Luznik et al. studied the bidirectional cytotoxic T-cell tolerance effect of post-transplant cyclophosphamide. They found that it is efficient in preventing GVHD by selectively eliminating T cells, and can result in durable engraftment.146
Consideration of post-transplant cyclophosphamide on ameliorating the detrimental effect of HLA disparity has been highly encouraged in the haploidentical allo-HCT setting. It has also been found as a rational option in matched unrelated donors and matched sibling donor allo-HCT in recently published data.53,147–149 One of the unanswered questions is whether post-transplant cyclophosphamide or anti-thymocyte globulin will be superior over others. Battipaglia et al. compared anti-thymocyte globulin with post-transplant cyclophosphamide in patients who had undergone HLA-mismatched unrelated donor transplantation. Results showed that post-transplant cyclophosphamide may have more favourable GVHD-free/relapse-free survival (37% versus 21%, p<0.03) outcomes without a survival advantage (56% versus 38%, p=0.07).150
In vivo T-cell depletion with anti-thymocyte globulin has been widely used in unrelated allo-HCT. Indications, type of anti-thymocyte globulin and dosing have been discussed in different studies. Donor type, conditioning regimen preferences, absolute lymphocyte count just before anti-thymocyte globulin infusion, ethnicity and GVHD prophylaxis preference appear to be good indicators of optimal anti-thymocyte globulin dose.151
Graft-versus-host disease treatment
Cumulative incidences of grade 3–4 acute GVHD range between 39% and 59% according to the stem cell source, donor type and conditioning regimen.152 Although first-line treatment of GVHD is corticosteroids, second-line treatment remains challenging. In a recently published randomized phase III trial, ruxolitinib showed durable overall response rates (ORR) in corticosteroid-resistant acute GVHD (40% versus 22%, p<0.001) with a tolerable safety profile compared with best available therapy.153 In a phase I study of pacritinib, a Janus kinase 1 inhibitor, 75% of the treatment-naive patients with acute GVHD had at least partial response at day 28. In steroid-refractory patients with acute GVHD, 28-day ORR was 70.6%.154
Magenau et al. showed, with a prospective phase II study, that α1-antitrypsin, a serine protease inhibitor, may also have a role in the treatment of steroid-resistant acute GVHD. Treatment responses were sustainable in around two-thirds of the patients without any need for additional immunosuppression.155 The phase II/III, multicentre, randomized MODULAATE study of α1-antitrypsin in patients with acute GVHD is currently ongoing (ClinicalTrails.gov Identifier: NCT03805789). In a phase Ib trial, vedolizumab, as an anti-α4β7 integrin monoclonal antibody, showed an ORR in 79% (n=23) of patients; 28% (n=8) of patients had a complete remission.156 Recently, in a phase III trial, human mesenchymal stromal cells induced a more favourable ORR at day 100 in steroid-refractory patients with acute GVHD compared with a prespecified control group (70.4% versus 45%, respectively; p=0.0003).157
Chronic GVHD is the major long-term consequence for allo-HCT survivors.158 In 2017, ibrutinib became the first US Food and Drug Administration (FDA)-approved drug approved for the treatment of chronic GVHD, in which patients failed with one or more lines of treatment. Ibrutinib is a selective inhibitor of Bruton’s tyrosine kinase proteins and interleukin-2 inducible kinase, which has a role in T-cell activation.159 Miklos et al. conducted a phase Ib/II study for testing the safety and efficacy of ibrutinib in patients with chronic GVHD who were non-responders to corticosteroid therapy and at least one line of therapy. The ORR was 67%, with a complete remission rate of 21% and a partial response rate of 45%.160
Graft failure is a substantial handicap in allo-HCT. Growth factors for tri-lineage, such as eltrombopag, recombinant human erythropoietin, and granulocyte stimulating agents, can be used successfully in the treatment of graft failure.161,162 A second allo-HCT can also be a solution.163
Supportive care
Infection is the leading cause of mortality during allo-HCT, especially in the early post-transplantation period.164 Not only is there an increased risk of bacterial infections, but also fungal and viral infections.165 Cytomegalovirus reactivation has been associated with poor allo-HCT outcomes and can induce graft failure or non-relapse mortality.156,166 Cytomegalovirus treatment may also result in myelosuppression as an adverse event.
Letermovir, an antiviral drug that inhibits the subunit of UL56 in the terminase complex that is essential for viral replication, has been approved for prophylaxis of cytomegalovirus by the FDA.167 In a multicentre phase III trial, prophylactic letermovir resulted in a lower reactivation rate and all-cause mortality compared with placebo without prominent myelosuppression.168 Maribavir is another antiviral drug active against cytomegalovirus infections, even in ganciclovir-, foscarnet- or cidofovir-resistant strains.169 Maertens et al. investigated the efficacy of maribavir 400 mg, 800 mg and 1,200 mg (by mouth, twice daily) on cytomegalovirus reactivation compared with valganciclovir in recipients of solid organ transplantation or allo-HCT in a phase II, randomized, dose-ranging study. Results showed similar response rates between maribavir treatment and the valganciclovir, and between the three maribavir dosing groups. Severe adverse event rates were higher in the maribavir groups (44% versus 32%).170
In conclusion, in recent years, there has been promising advancement in the state of art for allo-HCT regarding GVHD prophylaxis and treatment. Improvements have also been seen in optimal conditioning regimens, novel HLA typing methods, more available alternative donor sources and supportive care. For the vast majority of the haematologic diseases, allo-HCT seems to remain the only curative therapeutic option.
