The use of haploidentical stem cell transplant as an alternative donor source in patients with decreased access to matched unrelated donors

ABSTRACT

Introduction

The likelihood of finding HLA-matched unrelated donors for rare HLA types and non-white European ancestry continues to be a challenge with less than a 70% chance of finding a full match. Mismatched transplants continue to have high rates of transplant-related mortality. With the near-universal ability to find a haploidentical donor in families, haploidentical transplants have become of more critical importance in ethnic minority groups and patients with rare HLA types.

Methods

Data was collected through clinical trials, review articles, and case reports published in the National Library of Medicine.

Results

The use of improved lymphodepleting conditioning regimens, graft versus host disease (GVHD) prophylaxis using regimens such as post-transplant cyclophosphamide, mycophenolate, and tacrolimus have improved engraftment to nearly 100 percent and reduced transplant-related mortality to less than 20 percent. Attention to donor-specific antibodies (DSAs) with interventions using bortezomib, rituximab, and plasmapheresis has decreased graft failure rates.

Conclusion

With improved prevention of GVHD with interventions such as post-transplant cyclophosphamide and management of DSAs, haploidentical transplants continue to improve transplant-related mortality (TRM) compared to patients who received matched-related donor transplants. While TRM continues to improve, ongoing research with haploidentical transplants will focus on improving graft and donor immunosuppression and identifying the best regimens to improve TRM without compromising relapse-free survival.

KEYWORDS:

• Haploidentical transplant
• post-transplant cyclophosphamide
• donor specific antibodies
• HSCT transplant access
• T-cell depleted allograft
• T-cell replete allograft
• umbilical cord blood
• HLA match

Introduction

Allogeneic stem cell transplant (HCT) continues to be a potentially curative option in patients with hematological diseases. The rate of stem cell transplants has continued to increase since its inception in 1957, with more than a million allogeneic stem cell transplants occurring worldwide. The most recent international global survey of autologous and allogeneic stem cell transplants showed an annual increase of 40,000–80,000 from 2006 to 2016, with a projected rate of increase of 5% per year [1]. The rate of allogeneic stem cell transplant had increased from 20,000–38,000 during that time, with more transplants for leukemia, myelodysplastic syndrome, aplastic anemia, and hemoglobinopathies owing to improved transplant-related mortality (TRM) in both younger and older recipients [1]. This improvement can be attributed in part, to better prevention and management of graft-versus-host disease (GVHD) and newer conditioning regimens indicating that projected rates of allogeneic stem cell transplant will likely further increase in the future.

Historically, preference in donor selection relied heavily on having a human leukocyte antigen (HLA) matched sibling donor (MSD) due to the reduced risk of graft-versus-host disease (GVHD) and subsequent lower TRM. With refinements in HLA typing and improvements in the management and prevention of GVHD, donor selection widened to include HLA-matched unrelated donors (MUD). Optimal typing has included matching at the HLA-A, HLA-B, HLA-C, and HLA-DRB1 high-resolution locus (an 8/8 HLA match) and more recently HLA-DPB1 in conjunction with HLA-DQB1 to result in 12/12 matches. To add to the complexity of conventional HLA matching, non-permissive matching at DPB1 alleles has led to increased TRM due to increased rates of GVHD [2]. In addition to this, B-leader sequences on HLA-B exon 1 lead to dimorphic expression and antigen presentation through HLA-E, with mismatching in the HLA-B leader region leading to greater GVHD risk [2]. This has only added to the difficulties in both defining and finding the perfect HLA-matched donor.

Unrelated donor availability based on ethnicity

Based on data published by the United States National Marrow Donor Program (NMDP) in 2014, which classified people with European origin as ‘whites with European (WE) ancestry’ or whites of northern European ancestry (WNE) had a projected 75% chance of finding an adult unrelated donor, and patients with Asian ancestry had a 27–40% chance of finding a donor. Hispanic populations had a 30–40% chance of finding a MUD, and African Americans had a 19% chance of finding a MUD [3]. This was confirmed in European countries such as the United Kingdom where, according to the Anthony Nolan Graft Identification Advisory Service (GIAS), only 21% of patients from a minority ethnic background had a 10/10 HLA MUD compared to 70% of ‘WNE’ ancestry [4]. The GIAS also found that for ‘WNE’, the time for confirmatory typing was faster than for those of non-‘WNE’ ancestry at 27 days vs. 33.5 days. They also found that for non-‘WNE’ ethnic minorities, 61% found at least a 9/10 HLA-matched donor compared to 96% for ‘WNE’ [4].

Factors associated with decreased donor availability include lack of donor center availability in some countries, lack of trust and misinformation based on certain ethnic groups’ prior relationships with the medical system, donor deferral due to high risk of communicable disease in the donor’s country of origin, and availability of HLA typing and degree of resolution in HLA typing.

Since 1987, unrelated donor registries have increased from having availability in two countries to 119 countries in 2021. With improved and expanded registries, the number of unrelated donors has increased to 40 million with the inclusion of cord blood units. However, the rates of unrelated donors per 10,000 people vary per continent, with Africa having the lowest at 2.9, followed by 16.6 in Asia, 62 in Oceania, 179 in South America, 199 in North America, and 247 in Europe [1]. In addition, access to unrelated donor registries is affected by macroeconomics through delay in access to unrelated donor registries through limited healthcare resources. Graft-related costs, particularly transplants using cord blood units (CBU), can affect availability in patients who live in impoverished areas of high-income countries and those living in low and middle-income countries. In a publication from 2013, the GIAS in the UK found that utilization of both cord blood transplants and haploidentical transplants were higher in non-‘WNE’ compared to ‘WNE’ with CBU rates at 21.3% for non-‘WNE’ compared to 3.8% for ‘WNE’, and haploidentical transplants for non-‘WNE’ at 10.6% compared to 1.3% for ‘WNE’ [4].

For ethnic minorities and patients with rare HLA allele types, the use of cord-blood registries has improved donor availability for patients who are unable to find a fully matched donor. Matching for cord blood requires high-resolution matching at HLA-A, B, and DRB1 to be considered a 6/6 full match [5]. The rate of complete matching in patients receiving CBUs for adults older than the age of 20 is 17% for ‘WE’, 4% for Asian ancestry, 5% for Hispanic populations, and 2% for African Americans [3]. The likelihood of finding HLA-matched CBUs with adequate cell dose for patients under 20 years old is 38% because single blood units can be used for lower body weight patients compared to adults who need a higher cell dose and thus, may require two CBUs [4]. The low rates for full HLA-matched CBUs are due to decreased donor registry availability, with some donor registries not having CBUs as part of their inventory [1]. Due to this, the cost of CBUs can be high for recipients, as shown by the decreased availability of CBUs in low-income and middle-income countries [1]. Fortunately, the decreased immunogenicity of cord-blood units allows for successful transplantation of less fully matched CBUs with suitable units as low as 4/6 HLA-matched at high resolution at DRB1 with intermediate to high resolution typing at HLA-A and HLA-B as recommended by the Bone Marrow Transplant Clinical Trials Network (BMT-CTN) [6]. In addition to this, per the American Society of Hematology, intermediate resolution testing of HLA-C and high-resolution testing of HLA-A, -B and -DRB1 recommended when only one CBU is used [5].

The likelihood of finding a mismatched unrelated donor (MMUD) is higher with each degree of mismatch (HLA 7/8) but increases the risk of TRM and GVHD [7]. Increased mismatch in the graft versus host (GVH) direction increases GVHD risk because the mismatched antigen(s) is a primary target of donor T-cells [7]. In contrast, a mismatch in the host-versus graft (HVG) increases the probability of graft rejection due to the graft being a target of host T-cells [7]. In a study performed by Kanda et al., patients with HLA-mismatch had a higher risk of GVHD-related death at 16.9% in the MMUD compared to 4.4% in MUD [7]. In particular, the highest associated risk of acute GVHD (aGVHD) was found in HLA Class 1 (HLA-A, -B, and -C) mismatched patients. An additional 8% rate of five-year mortality was seen in recipients mismatched for one HLA allele in the setting of standard calcineurin-based GVHD prophylaxis [3]. For these patients, cord-blood registries have improved donor availability for patients who cannot find a completely matched unrelated donor.

For MMUD donors, the likelihood of finding a 7/8 match is 97% for ‘WNE’, 80% for the Asian population, 70-85% for the Hispanic population, and 70–75% for African Americans [3]. For CBU, the likelihood of finding at least a 4/6 HLA match is 90% for all populations except for African Americans, who have an 80% likelihood of finding a suitable CBU donor [3].

Alternative donor sources in rare HLA allele and ethnic minority populations

Matching requirements using CBUs are less strict owing to the lower immunogenicity of the graft against the donor, with mismatching at one or two loci acceptable in the proper clinical setting of non-HLA factors [1]. Due to the limited hematopoietic elements found in each cryopreserved CBU, additional units might be needed in patients with higher body weight, given delayed neutrophil recovery and graft failure with low cell doses [3]. For patients with mismatching at two or more HLA high-resolution loci, in comparison to matched CBU, neutrophil recovery rates were lower and associated with higher rates of graft failure and mortality [8]. To help reduce the risk of graft failure, the use of double CBUs or a minimum dose of 2.5 × 10⁷/kg with CD34 + cells >1.5X 10⁵/kg in a single CBU has been recommended by the NMDP [8]. Utilizing CBUs does have its benefits, owing to the ease of availability of pre-stored CBUs and decreased graft versus host disease rates allowing for the use of more mismatched CBUs. However, in patients who relapse post-transplant, the lack of availability of donor lymphocyte infusions (DLI) and stem cell boosts in the event of graft failure or poor graft function are a major limitation of cord blood transplants. Additionally, CBUs are considerably more expensive with the acquisition of the units before infusion being more than $80,000 [9]. Obstacles to using CBUs include expense, higher graft failure rates, slow immune reconstitution leading to opportunistic infections, and delayed engraftment leading to increased non-relapse mortality compared to donor-directed matched unrelated donor grafts [9, 10].

As a result of these barriers in transplant for patients of ethnic minorities who can experience high TRM with HLA-mismatched and CBUs, a focus has been on the use and development of haploidentical stem cell transplantation in this population. The utilization of haploidentical stem cell transplants has occurred more frequently in countries without access to umbilical cord banks. Decreased access to unrelated donors has also led to the use of haploidentical stem cell transplants, accounting for up to 39% of all related donor transplants [1]. Haploidentical stem cell transplant (haplo-HCT) uses first-degree relatives, typically either parents or children, but also non-matched siblings and second-degree relatives. The likelihood of finding a sibling haplotype is 50%, with 100% availability from parent to child. Historically, before the advent of improved graft manipulation with T-cell depletion or T-cell suppression in-vivo, TRM with haploidentical stem cell transplants was considerably increased due to bidirectional alloimmunity and immunoreactivity between donor and recipient HLA [11]. Both increased GVH and HVG reactions led to high rates of transplant-related mortality in the 1970s–1980s [12]. One study evaluating aplastic anemia showed 10% graft failure-related death, with 40% of deaths related to GVHD [13]. Among 29 patients with hematological malignancies, TRM was 41% [13]. The first study evaluating haploidentical transplants in unselected first-degree relatives showed 29% graft failure, with 17% death due to acute GVHD and 34% with hyperacute GVHD [14]. The Fred Hutchinson Cancer Research Center published their data using haploidentical transplants and showed a 24% graft failure rate and a 70% rate of aGVHD. Despite using cyclosporine and methotrexate to prevent GVHD, having two mismatch alleles led to an aGVHD Grade III-IV rate of 47% [15].

Donor-specific antibodies effect on haploidentical transplants

An important complication of haploidentical stem cell transplants (which can also be seen with cord blood and mismatched transplants) is the presence of donor HLA-specific antibodies (DSA). Donor HLA-specific antibodies occur secondary to prior alloimmunization to unique HLA antigens. Causes of HLA alloimmunization include pregnancy, blood transfusions, and prior solid organ transplant, with blood transfusions being the most common cause of alloimmunization. Rarely, infections, vaccinations, and trauma can cause passive alloimmunization to HLA. Up to half of all transplant candidates are positive for HLA antibodies, with most being multiparous females [2]. The presence of anti-donor HLA antibodies increases the risk of graft failure and rejection. Historically, the evaluation of HLA antibodies used complement cell-mediated crossmatching with donor and recipient; however, single antigen bead (SAB) testing has become more mainstream to identify the exact antigen and antibody interaction in a semi-quantified manner using mean fluorescent intensity (MFI) by Luminex technology [2].

In a study of 592 patients undergoing unrelated donor transplants and looking at the effect of DSAs on transplant-related mortality, 8 out of 116 (19.6%) patients with HLA antibodies had DPB1 antibodies, with three of the eight patients having primary graft failure [2]. In another trial, 24% of primary graft failure had DSAs against HLA-A, HLA-B, or HLA-DP compared to 1% of the control group [16].

For haploidentical HCT with anti-donor HLA antibodies, MD Anderson Cancer Center found that three out of four patients with DSAs failed to engraft compared to one out of twenty graft failures without DSAs. The associated MFI with these graft failures was 3000 [17]. In a follow-up study, MD Anderson evaluated 122 patients with 22 having DSA. Seven out of eleven patients with DSA MFI over 5,000 failed to engraft compared to none with DSA MFI less than 5,000. Patients with C1q positivity, associated with complement activation, had increased graft failure rates [18]. Currently, based on this published data, the European Society for Blood and Marrow Transplant’s consensus status is that a positive DSA is classified as a MFI of 1000, with a high DSA considered when more than 5,000 MFI [19].

Therefore, we looked at the current literature and research supporting the use of haploidentical stem cell donor sources along with current GVHD and graft failure prevention strategies.

Methods

Data was collected through clinical trials, review articles, and case reports published in the National Library of Medicine. A haploidentical stem cell donor source was defined by a first-degree, or second degree relative with at least 4/6 allele matching. T-cell depletion was defined by in-vitro graft manipulation by either selecting a predominant T-cell clone, removal of potentially allo-reactive T-cells, or in-vitro lymphodepletion before stem cell infusion. T-cell replete transplantation was defined by in-vivo T-cell depletion using immunosuppressive agents. We evaluated rates of Overall survival (OS) Relapse-free survival (RFS), Transplant-Related Mortality (TRM), graft failure, and incidences of GVHD.

Results

T-cell depleted grafts in haploidentical transplants

Because of high TRM in the 1980s, the prevalence of haploidentical transplants declined. However, this changed with the identification of alloreactive T-cells as the main culprit in TRM for this population. Studies looking at ex-vivo non-selective T-cell depletion showed improved rates of GVHD; however, these transplants were associated with increased rates of graft failure, graft dysfunction with delayed immune reconstitution, and increased morbidity and mortality from opportunistic infections. The lack of donor T-cell elements also led to host anti-donor T-cells eradicating the graft and resulting in rejection. Mega-doses of CD34+ cells were therefore utilized with granulocyte colony-stimulating factor (GCSF) mobilized stem cells in coordination with total body irradiation (TBI) and cyclophosphamide or thiotepa with anti-thymocyte globulin (ATG) before T-cell depleted (TCD) stem cell infusions in an attempt to neutralize the recipient’s anti-donor T-cells. This improved engraftment rates up to 90%; however, TRM remained high, in the range of 40%, with two-thirds of deaths attributed to infectious complications (Table 1) [20, 21]. In pediatric patients, the EBMT found that TCD in the haploidentical transplant setting had similar responses in adults with 37% TRM, and 34% incidence of relapse (Table 1) [22].

Table 1. Comparative studies in T-cell teplete and depleted grafts for haploidentical stem cell transplants.

Study	Patients	Conditioning	GVHD Prophylaxis	Graft failure	Relapse rate	aGVHD 2–4%	TRM	OS
Historical Studies
Powles et al. (1983)	35(AML/ALL)	TBI/MTX	CsA	37%	11%	85%	57%	14%
Beatty et al. (1985)	105(BMF/AML)	Cy/TBI	MTX	20-30%		70%		20-40%
T-cell Depleted Grafts
Aversa et al. (1994)	17 (AML/ALL)	Cy(Thi)/TBI/ATG (G-CSF)	None	10%	12%	6%	40%	35%
Ciceri et al. (2008)	173(AML/ALL)	TBI-based/ATG	None	9%	21-27%	5-11%	25-40%	20-30%
Klingebiel et al. (2010)	127 (ALL)	TBI-based/ATG	None	9%	36%	17%	37%	29%
Federmann et al (2012)	61 (AML/ALL)	Flu/Mel/Thi/OKT-3	None	6%	31%	46%	41%	28%
Lang et al. (2014)	46 (AML/ALL)	Flu/Mel/Thi/ATG	None	13%	60%	27%	20%	26%
Maschan et al. (2016)	13 (AML/ALL)	Treo/Mel/Flu/ATG	Tac/MTX	0%	40%	23%	0%	73%
Giardino et al. (2022)	20 (BMF)	Treo/Flu/Thiotepa	None	25%	N/A	5%	10%	90%
Bethge et al. (2022)	60 (AML/ALL)	Flu/Mel/ATG	MMF	15%	33%	10%	17%	60%
T-cell Replete Grafts
Huang et al. (2006)	171 (AML/ALL)	BuCy/Ara-C, CCNU/ATG	GIAC (MTX,CsA,MMF)	0%	18.7%	55%	19-27%	42-68%
Luznik et al (2008)	67 (AML/Lymphoma)	Flu/Cy/TBI	PTCy/Tac/MMF	1%	51%	34%	15%	46%
Brunstein et al (2011)	50(AML/Lymphoma)	Flu/Cy/TBI	PTCy/Tac/MMF	2%	26%	32%	7%	62%
BMT-CTN 0603
Fuchs et al. (2021)†	182 (AML/ALL)	Flu/Cy/TBI	PTCy/Tac/MMF	1%	48%	28%	11%	57%
BMT-CTN 1101†
Malki et al. (2022)	31 (AML)	Flu/Cy/TMLI	PTCy/Tac/MMF	0%	40%	52%	9%	83%
Al-Homsi et al. (2023)	46 (AML/ALL)	RIC/MAC	PTCy/Tac/Aba	0%	16%	17%	4%	86%
Bejanyan et al (2021)	32 (AML/ALL)	Flu/BU or Flu/Bu/TBI	PTCy/Siro/MMF	0%	22%	19%	19%	72%
Elmariah et al. (2024)	100 (AML/ALL)	RIC/MAC	PTCy/Siro/MMF	0%	22%	45%	27%	63%

Abatacept, Aba; AML, Acute Myeloid Leukemia; ALL, Acute Lymphomatous Leukemia; Anti-thymocyte Globulin, ATG; TBI, Bone Marrow Failure, BMF; Busulfan, Bu; CCNU, Lomustine;; Cyclophosphamide, Cy; Cyclosporine, CsA; Cytarabine, ara-C; Fludarabine, Flu; Granulocyte-Colony Stimulating Factor, G-CSF; Myeloablatic Conditioning, MAC; Melphalan, Mel; Methotrexate, MTX; Mycophenolate Mofetil, MMF; Post-transplant Cyclophosphamide, PTCy;Reduced Intensity Conditioning, RIC; Sirolimus, Siro; Tacrolimus, Tac; Thiotepa, Thi; TBI Total Body Irradiation; TBI; Total Marrow Lymphatic Irradiation, TMLI; Treosulfan, Treo.

†: Denotes Haplo-arm of clinical Trial.

Attempts at using CD3 negative/ CD19 negative grafts, in an effort to allow NK cells, antigen-presenting cells, and other immune effector cells to reconstitute the immune system were done in Germany at the University of Tuebingen in patients with hematological malignancies [23]. In this phase II study, 61 patients underwent haplo-HCT using RIC with fludarabine, thiotepa, melphalan, and OKT-3 with CD3/CD19 lymphodepleted grafts. No patients had primary graft failure with 6.5% of patients having secondary graft failure. Out of 61 patients, the 2-yearTRM was 41% with evidence of grade II–IV aGVHD at 42% with 31% of patients relapsing, leading to a 2-year OS of 28% (Table 1) [23]. For pediatric patients, attempts of using CD3 negative/CD19 negative selection showed high relapse rates of 63% at two years, while having improved grade II to grade IV GVHD of 20% and TRM of 20% after myeloablative conditioning (Table 1) [24].

Recently, significant interest has looked at using CD19+ and TCRα/β, naïve T-cell, selected depletion, with sparring of TCR γ/δ. In contrast to CD34+ selection and CD3+ depletion, these grafts restrict only naïve T-cells from being infused into the patient, and keep regulatory T-cells, NK cells, and immune effector cells available for infusion to help with immune reconstitution and graft-versus-leukemia (GVL) effect. One of the first studies looking at patients receiving both MUD and haplo-grafts showed a low risk of TRM (0%) but high risk of relapse (40%) in the haplo-HCT subgroup (Table 1) [25]. A follow-up study looking at 20 pediatric patients receiving haplo-HCT for nonmalignant conditions showed a 15% primary graft failure rate, a 10% secondary graft failure rate with only one patient having grade 4 aGVHD (5%), and 90% of patients alive at two years (Table 1) [26].

More recently, one Phase I/II clinical trial of 60 adults and pediatric patients by the Germany-based group that previously evaluated CD3/CD19 restricted TCD grafts, showed favorable aGVHD rates, only grade II (10%), low TRM rates of 17% at 2 years, though high graft failure rates (15%) and relapse rates (33%). The overall survival was predominantly driven by relapse rates, with 2-year OS at 63% (Table 1) [27].

Even with improved selection of T-cell ex-vivo depletion, graft failures continue to remain prominent with high relapse rates following transplant.

Use of T-cell replete donor grafts in haploidentical setting

Due to issues with TCD-haploidentical transplants with associated graft failure and TRM, T-cell replete (TCR) haploidentical protocols have been pursued using immunosuppressive therapy to target the alloreactive T-cells while sparing the non-alloreactive T-cell population, such as regulatory T-cells, in the graft.

The first trials looking at this approach at Peking University in Beijing utilized the ‘GIAC’ protocol: combining G-CSF stimulated donor cells, post-transplant cyclosporine, and mycophenolate along with short-course methotrexate and ATG. The rates of GVHD did not increase with the utilization of these GCSF-mobilized stem cells due to increased formulation of T-helper 2 (Th2) T-cells causing reduced T-cell responsiveness and decreased alloreactivity when combined with immunosuppression. All patients achieved engraftment with Grade II-IV aGVHD rates of 55% and Grade III-IV GVHD rates of 23% [28]. The one-year TRM for the standard-risk disease group was 17.4% and 29.4% for the high-risk group (Table 1) [28].

Another approach to TCR-replete transplants utilized post-transplant cyclophosphamide to decrease alloreactive T-cells while not affecting non-alloreactive T-cells. Post-transplant cyclophosphamide was first identified as a potential source based on research at Kyushu University in which murine skin allografts had prolonged survival after cyclophosphamide was given up to four days post-transplantation [12]. Post-transplant cyclophosphamide (PTCy) created drug-induced immunotolerance by alkylating the DNA of mitotically active or proliferating T lymphocytes while avoiding senescent T-cells. This discrimination was possible because senescent T-cells and bone marrow stem cells had high expression of aldehyde dehydrogenase, which protected against alkylator damage. The use of PTCy as potential GVHD prophylaxis was published by Johns Hopkins in 2001 when they used PTCy on Days three and four after HLA-mismatched mouse models with conditioning using fludarabine, cyclophosphamide and total body irradiation (TBI) [29]. The Johns Hopkins group found that spleen cells survived longer in those receiving PTCy and had attenuated GVH reactions compared to controls [29]. This study corroborated the results found at Kyushu University and advanced PTCy’s use in living models. Furthermore, Johns Hopkins conducted a Phase Two trial in conjunction with Fred Hutchinson Cancer Center of 67 patients receiving PTCy 50 mg/kg on Days 3 and 4 after transplant. Graft failure occurred in 1.4% of patients, with grade II-IV GVHD rates of 34% but only 6% had grade III–IV. Relapses occurred in half of the patients (51%) with NRM at 15%: this led to an overall survival of 36% at one year, predominantly driven by relapse [30].

With the use of post-transplant cyclophosphamide, results showed improved GVHD control, engraftment rates, and less profound immunosuppression compared to TCD-haploidentical transplants. However, rates of relapse, particularly in combination with reduced-intensity conditioning regimens, continued to be a barrier. A study performed by the Blood and Marrow Transplant Clinical Trials Network (BMT-CTN) (BMT-CTN 0603) of 50 patients utilized fludarabine, cyclophosphamide, and TBI conditioning with post-transplant cyclophosphamide, tacrolimus, and mycophenolate mofetil (MMF). The engraftment rate was 96%, with only one patient having graft failure [31]. The rate of Grade II acute GVHD was 32% with no Grade III or Grade IV acute GVHD, with 13% chronic GVHD (cGVHD) rates, and a TRM rate of 7% (Table 1) [31]. Relapse rates, however, increased to 45% at one year. This trial was in direct comparison to umbilical cord blood (UCB) transplants and ran as a parallel trial due to similar patient characteristics [31]. The GVHD prophylaxis used for these transplants consisted of cyclosporine and mycophenolate mofetil (MMF). Given similar patient outcomes in these 2 phase 2 trials of UCB and haploidentical transplants, the BMT-CTN conducted a phase 3 trial, BMT-CTN 1101, randomizing patients to receive either a UCB or a haploidentical transplant [32]. Between the two groups, the rates of aGVHD and cGVHD were similar with overall similar relapse rates; however, the rate of TRM at two years was lower for the haploidentical transplant group at 11% compared to 18% for UCB transplants leading to an improved OS rate for haploidentical transplants at 57% compared to 48%, respectively (Table 1) [32]. The cause of TRM in the UCB transplant group primarily involved infectious and bleeding complications related to slow engraftment rates [32]. Follow-up research looking at health-related quality of life between the two groups did not show any difference between the two groups [33]. In comparison to UCB transplants, haploidentical transplants had lower direct medical care costs [9].

Several studies using myeloablative conditioning regimens such as thiotepa, busulfan, and fludarabine in the haploidentical transplant setting have shown decreased relapse rates. In one study looking at 50 patients with high-risk disease with this conditioning regimen, the rate of Grade II–IV aGVHD was 12%, TRM was 18%, and the relapse rate was 26% (Table 1) [34]. Another strategy to reduce relapse rates includes targeted radiation to the bone marrow and lymph nodes while sparing many organs and therefore reducing total organ toxicity. In one Phase 1 clinical trial of 31 patients using Total Marrow and Lymph Node Irradiation (TMLI), patients received haplo-HCT using PBSC grafts with fludarabine/cyclophosphamide and incremental TMLI (1200–2000cGy) dose with GVHD prophylaxis using PTCy, MMF and Tacrolimus. Cumulative incidence of acute GVHD grade 2–4 was 52% and grade 3–4 at 6% with relapse rates of 40% at 1 year for the entire population, and 17% at the higher TMLI dose of 2000cGy (Table 1) [35].

With the advent of PTCy’s use as GVHD prophylaxis in haplo-HCT settings, other immunosuppressive therapies have been added to the PTCy GVHD prophylaxis backbone. Abatacept, a soluble fusion protein of Immunoglobulin IG1 (IgG1)’s Fc region and Cytotoxic T-lymphocyte associated antigen 4 (CTLA4), blocks the costimulatory domain of T-cells, preventing activation [36]. In a study performed at New York University (NYU) Langone, 46 patients receiving haplo-HCT with either MAC or RIC received the CAST protocol: PTCy, Abatacept at fixed dosing schedule on days 5, 14, 28, and 56, and short-course tacrolimus with taper initiated on Day 60 if no aGVHD. Out of the 46 patients, the incidence of grade II-IV aGVHD was 17.4%, TRM at 4.4%, relapse rate 15.9% with overall survival of 86% (Table 1) [36].

Sirolimus, another immunosuppressive that had previously been used in combination with either tacrolimus or methotrexate as GVHD prophylaxis, was evaluated in combination with PTCy for immunosuppression in patients receiving haplo-HCT. Sirolimus is an mTOR (mammalian target of rapamycin) inhibitor that affects T-cell expansion of regulatory T-cells and is thought to improve GVHD prevention [37]. In a study performed at Moffitt Cancer Center, 32 patients received MAC or RIC/NMA conditioning therapy for hematological malignancies, followed by PTCy, sirolimus, and MMF for GVHD prophylaxis. At one year follow-up, no patients had graft failure with 72% of patients, 22% with relapse of their primary disease, and 19% of patients dying of complications due to infection or GVHD (Table 1) [37].

Following the pivotal BMT-CTN 1101 and this trial by Moffitt, the CIBMTR retrospectively evaluated 423 patients receiving haplo-HCT using either tacrolimus or sirolimus with PTCy and MMF [38]. The rates of aGVHD were comparable between both groups (45% vs. 47%) with no graft failure, and comparable TRM (27% versus 22%, P = 0.36) and relapse rates (23% vs 25%) (Table 1) [38].

Donor specific antibody removal strategies to reduce graft failure

Due to high graft failure rates associated with alloimmunization, multiple attempts have been made to reduce graft failure by removing antibodies through apheresis techniques or targeting the cells making DSAs. Therapeutic plasma exchange (TPE) is a way to mechanically separate blood components to reduce plasma and replace the plasma with colloidal fluid. Therapeutic plasma exchange has become a mainstay treatment for removing DSAs and has been widely used in the setting of solid organ transplants. Other transfusion medicine-based treatments include donor platelet or buffy coat infusion to adsorb the HLA antibodies before stem cell transfusion. Polyclonal intravenous immunoglobulin (IVIG) has been used to provide passive immunotolerance through Fc-receptor downregulation of innate and adaptive immune activation [2].

Depletion of CD20-positive circulating B-cells can also play a significant role in desensitization. Rituximab, an anti-CD20 monoclonal antibody, has been widely used to improve immunotolerance and remove alloreactive B cells. Plasma cells responsible for anti-HLA IgG production are another target for treatment, focusing on using therapies classically used to treat multiple myeloma. Bortezomib, a proteasome inhibitor, has been used to reduce circulating plasma cells in managing anti-HLA antibodies [39].

One study of 20 haplo-HCTs performed by the Madrid group from 2012 to 2020 showed that in 19 patients with MFI >5,000 who received plasma exchange, IVIG, and Rituximab, 90% of patients had neutrophil engraftment with only one having graft failure and another dying before engraftment due to infection [40]. Another study performed by Ciurea and colleagues at MD Anderson evaluated 37 patients with elevated DSAs using plasma exchange, Rituximab, IVIG, and irradiated buffy coat before haploidentical stem cell transplant [41]. Patients treated with this combination of therapy had a mean MFI of 10,000 with a reduction to 5931 after desensitization. The rate of graft failure remained high for patients who had MFI greater than 20,000 before desensitization with the presence of C1q. However, patients with MFI <20,000 before desensitization with clearance of C1q after desensitization had graft failure rates comparable to patients with low-normal DSA MFIs [41].

Another monoclonal antibody, daratumumab, that targets CD38 on plasma cells, has recently been evaluated in a solid organ transplant population and shown response in a small group of patients through case reports; however, no clinical trial has evaluated this as an option of therapy at this time [42–44].

Conclusion

For patients undergoing stem cell transplants, HLA-matching continues to be an important approach to improve transplant-related mortality and graft failure. Though patients with northern European ancestry can have up to 70% likelihood of finding a full HLA match, patients of other ethnic groups have lower rates of finding a matched donor. Umbilical cord blood donors have been evaluated to help improve donor availability. However, due to increased cost and decreased availability of multiple units, combined with increased risk of graft failure and infectious complications, only select institutions use this as a donor option for patients without matched unrelated donors. With the near-universal ability to find a haploidentical donor in families, haploidentical transplants have become more critical in ethnic minority groups and patients with rare HLA allele types. However, historically, haploidentical transplants had high graft failure rates due to DSAs, host immunity factors, and GVHD. The results of multiple clinical trials have shown that haploidentical stem cell transplants continue to be a viable option for patients of ethnic minorities who do not have an available matched related or unrelated donor. Benefits of haploidentical stem cell transplants over umbilical cord blood include easier availability and cost compared to umbilical cord blood along with availability of additional cells from the donor for DLI and/or stem cell boosts without compromising patient quality of life [9, 33].

With improved prevention of GVHD using TCD and TCR-haploidentical transplants and management of DSAs, haploidentical transplants continue to improve TRM compared to patients who received matched-related donor transplants. Strategies to reduce graft-failure rates in patients receiving TCD transplants have focused on improving T-cell selection, with the current focus on using TCR α/β and CD19+ depleted approach. In patients receiving TCR transplants, the backbone of PTCy has helped reduce aGVHD rates, however, relapse has continued to be a problem in high-risk patients with hematological malignancies. Current modalities using additional GVHD prevention such as abatacept and short-course tacrolimus to reduce total tacrolimus duration have sought to combat this rising risk of relapse after transplant. While TRM continues to improve, ongoing research with haploidentical transplants will focus on improving graft and donor immunosuppression and identifying the best TCD and TCR regimen to improve TRM without compromising relapse-free survival. With improved TCR-haploidentical transplants, infectious complications and aGVHD rates have significantly improved. However, rates of relapse continue to be high; therefore, selecting the best T-cell suppressive regimen continues to be of paramount need.

Authorship contributions

Christopher Graham: responsible for writing, conceptualizing article. Mark Litzow: responsible for conceptualization of article, editing and writing the article.

Acknowledgements

No additional acknowledgements.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data supporting this article can be found published through the National Library of Medicine. No new data was generated in this article.

Additional information

Notes on contributors

Christopher Graham

Christopher Graham: Dr. Graham is a Blood and Marrow Transplant Fellow at Mayo Clinic. His research focuses on increasing access of allogeneic stem cell transplant to patients.

Mark Litzow

Mark Litzow: Dr. Litzow is Professor of Medicine in the Division of Hematology at Mayo Clinic. He was the Director of the BMT Program at Mayo from 1992 to 2008 and continues to remain active in the BMT program. He was the Director of the Myeloid Disease Oriented Group in the Division of Hematology from 2002 to 2016. He was appointed chair of the Leukemia Committee of the Eastern Cooperative Oncology Group-American College of Radiology Imaging Network (ECOG-ACRIN) in September 2013 after having served as co-chair of the committee since 2001.