Is allogeneic transplantation for sickle cell disease still relevant in the era of gene therapy?

Sickle cell disease (SCD) is the most common inherited blood disease worldwide. In 2021, >500 000 babies were born with SCD, mostly in sub-Saharan Africa, with nearly 8 million people living with the disease globally.¹ In the United States, SCD affects ∼1 in 360 African American newborns and >100 000 individuals.² SCD is an autosomal recessive disorder caused by a point mutation in codon 6 of the β-globin chain, which results in an amino acid substitution of valine for glutamic acid. Red blood cells from individuals with SCD assume sickled forms under hypoxic conditions, leading to chronic hemolysis, inflammation, ischemia-reperfusion injury, and immune system activation, which in turn results in progressive organ disease.² Common complications include acute vaso-occlusive pain events, acute chest syndrome, and stroke. SCD can no longer be referred to as a life-threatening disease in children but rather a chronic disease that has life-threatening events.³ Although the survival of children with SCD has improved in the United States and Europe with the use of hydroxyurea and evidence-based supportive medical care, SCD in adults remains a disease with a high risk of morbidity and early death. Over 50% of adults with SCD have silent cerebral infarcts,⁴ and are at increased risk for new or progressive infarcts.⁵ Other end-organ toxicities such as pulmonary hypertension, cardiac disease, and chronic renal disease are also increased in adults with SCD. The median survival of 48 years in SCD has not changed significantly in the last quarter of a century.^6,7 However, cure, or elimination of sickle cell hematopoiesis, may now be a consideration for many, if not most, individuals with SCD.

The germ line mutation in SCD only affects red blood cells. Thus, transplantation of a new unaffected hematopoietic system should cure the disease. The first report of allogeneic hematopoietic cell transplantation (allo-HCT) for SCD was in 1984, regarding an 8-year-old African American girl with SCD who underwent allo-HCT from her HLA-matched brother for acute myeloid leukemia (AML).⁸ The patient developed fairly severe acute and chronic graft-versus-host disease (GVHD) but at last report was doing well 9 years after the transplant.⁹ The authors offered the following conclusions at the time: “Marrow transplantation was used in our patient primarily to treat acute myeloblastic leukemia, but the treatment also proved effective in eliminating SCD. At present, marrow transplantation can have only a small impact on SCD, since the immunologic complications of the treatment, particularly GVHD, will limit its application to a few selected patients.”⁸ These comments were based on the fact that only ∼10% of patients would have an unaffected matched sibling donor,¹⁰ and at that time even patients with matched sibling donors had mortalities from GVHD approaching 20%. Unfortunately, the development of unrelated registries in the 1990s rarely provided matched donors for individuals with SCD.¹¹ There are several reasons for this, including the number of African Americans in donor registries and the diversity of their HLA system, which often generates uncommon HLA types.^11-13 Another key obstacle was the requirement for myeloablative conditioning, which is too toxic for many individuals with SCD with end-organ disease and will usually cause infertility.^10,14 

Based on these barriers of limited donor pool and the excessive toxicity-related mortality and morbidity, <200 individuals with SCD had undergone allo-HCT before the turn of the millennium.¹⁴ Most were children in the prehydroxyurea and pre–primary stroke prevention era with complications of SCD such as stroke, recurrent episodes of acute chest syndrome, or repetitive vaso-occlusive pain crises. All received myeloablative conditioning with busulfan and cyclophosphamide, and most received GVHD prophylaxis with cyclosporine and methotrexate.¹⁴ The nonrelapse mortality (NRM) was ∼10%, and graft loss with recurrence of SCD occurred in ∼10% of patients.¹⁴ Thus, ∼80% of patients achieved near-normal hemoglobin (Hb) levels and were without acute vaso-occlusive pain events after allo-HCT. A quarter of patients developed severe acute GVHD, and 12% developed chronic GVHD, resulting in the major causes of mortality. Most patients developed gonadal failure, and neurologic complications, particularly seizures, occurred in 25% of patients after allo-HCT.¹⁴ Stable mixed donor–host hematopoietic chimerism was observed in 11% of the individuals with SCD after allo-HCT. Most of these patients with stable mixed chimerism reported resolution of clinical vaso-occlusive episodes, suggesting there may be clinical benefit for such patients.¹⁴ 

Despite these encouraging early results, only a small fraction of children were eligible for allo-HCT, because at the time it required an HLA-matched donor. Moreover, NRM, GVHD, and graft failure occurred in significant proportions of individuals, and infertility was universal. Very few adults with SCD disease had undergone allo-HCT by the turn of the millennium, because of the concerns of the toxicities of myeloablative conditioning in older patients with end-organ dysfunction. A survey of adults with SCD found that the majority would be willing to consider a curative option such as allo-HCT even with NRM and graft failure rates of >10%, but chronic GVHD and infertility were unacceptable to the majority.¹⁵ Two different strategies to cure SCD were developed to overcome these intransigent barriers of a limited donor pool and treatment-related toxicities: reduced intensity conditioning (RIC) haploidentical (haplo) transplants with posttransplantation cyclophosphamide (PTCy) and genetic therapies.

The issues with myeloablative conditioning led to attempts to limit this toxicity with the use of nonmyeloablative conditioning or RIC. The first such report in 2003, included 6 patients with SCD who underwent RIC and matched sibling donor allografts.¹⁶ Donor chimerism (range, 25%-85%) was achieved in 5 of 6 patients, with increased HbA and improvement in complications attributable to SCD.¹⁶ However, after posttransplantation immunosuppression was tapered, all patients lost their donor grafts. No severe transplant-related toxicity was observed, as all patients experienced autologous hematopoietic recovery (and SCD recurrence).¹⁶ 

The National Institute of Health published their experience with matched sibling allo-HCT for SCD, using a transplant platform of alemtuzumab, 300 cGy total body irradiation (TBI), and posttransplant sirolimus.^17,18 Of 30 patients, 1 died, and most demonstrated stable mixed chimerism at 1 year. However, the mean level of donor T-cell chimerism was 48%, and only half of the patients could discontinue immunosuppression, raising concerns about the durability of the mixed chimerism. Importantly, suitable HLA-matched sibling donors were found for <10% of individuals with SCD.¹⁸ 

Because of the inability to find HLA-matched donors for individuals with SCD, the investigative team at Hopkins studied their haplo-related donor nonmyeloablative platform using PTCy.¹⁹ They previously showed that with the use of PTCy, haplo transplants could safely be performed in patients with hematologic malignancy with outcomes similar to those seen with matched donor transplants.²⁰ Multiple groups worldwide have now confirmed these findings, and mismatched transplants with PTCy, both related and unrelated, are now considered standard of care.^21,22 The platform used their standard nonmyeloablative haplo donor regimen of fludarabine, cyclophosphamide, and 200 cGy TBI but with the addition of antithymocyte globulin; GVHD prophylaxis used PTCy, mycophenolate mofetil, and tacrolimus or sirolimus. The first such transplant was performed in 2007, and the patient remains disease-free to this day.²³ Of 19 screened patients, 17 underwent haplo-HCT using bone marrow to reduce the risk of GVHD associated with peripheral blood allografts. Although the transplants were well tolerated, with only 1 patient developing skin-only acute GVHD that resolved without any therapy and no deaths, graft failure occurred in 43% of patients. Overall, 11 patients were durably engrafted, and 10 patients were free of any SCD symptoms.¹⁹ 

In an attempt to overcome the high graft failure rate, the Hopkins group increased the dose of TBI from 200 to 400 cGy,²⁴ which substantially reduced graft failure while maintaining the safety of haplo-HCT with PTCy; only 1 of 12 patients with SCD on this regimen experienced graft failure, there were no deaths or persistent GVHD, 9 of 11 patients showed full donor T-cell chimerism (the other 2 were 70% and 80% donor), and 10 of 11 were off all immunosuppression at the time of report.²⁴ The Vanderbilt consortium reported similar results by adding thiotepa to the Hopkins haplo backbone instead of increasing the TBI dose.²⁵ The Vanderbilt consortium recently updated their haplo-HCT with PTCy results in 70 patients.²⁶ Graft failure (all with autologous recovery) occurred in 8 of 70 participants, all in children aged <18 years. The median donor chimerism in those with engraftment was >95%, and 57 of 59 were off immunosuppression 1 year after transplant. The 1-year grades 3 to 4 GVHD rate was 10%, and the 2-year moderate-to-severe chronic GVHD rate was 10%. Five participants died from infectious complications for an overall 2-year event-free survival (EFS) of 83% and an overall survival (OS) of 94%.²⁶ 

The Blood and Marrow Transplant Clinical Trials Network recently reported the results of a national multi-institutional trial studying RIC haplo-HCT for young adult patients with SCD, using the Vanderbilt modification of the Hopkins PTCy protocol with the addition of preconditioning with hydroxyurea.²⁷ The results confirmed the outstanding outcomes of haplo-HCT with PTCy for SCD in adults. A pediatric arm has recently been completed, and we are awaiting the results. The median age of the 42 patients with SCD who underwent haplo-HCT was 22.8 (range, 15.5-43.2) years, and the median follow-up was 37.2 (range, 20.4-56.4) months. The estimated 2-year EFS and OS were 88% (95% confidence interval [CI], 73.5-94.8) and 95.0% (95% CI, 81.5-98.7), respectively. Two participants had primary graft failure, whereas the 2-year secondary graft failure estimate was 2.4% (95% CI, 0.2-10.9). The cumulative incidence of grades 3 to 4 acute GVHD at day 100 was 4.8% (95% CI, 0.9-14.4), whereas the 2-year chronic GVHD rate estimate was 22.4% (95% CI, 10.9-36.4). The 2 deaths in the study were related to infectious complications, at least 1 due to COVID-19 infection.

SCD has long been the prototype for the promise of curative gene therapy, because the point mutation in 1 gene is expressed exclusively in 1 cell type, red blood cells. The US Food and Drug Administration has recently approved 2 gene therapy approaches for modifying/curing SCD.^28,29 Both approaches involve ex vivo genetic manipulation of hematopoietic stem cells (HSCs) after removing (or collecting) them from the patient. The genetically modified HSCs are reinfused after myeloablative chemotherapy, which is needed to allow the corrected HSCs to expand. Genetic therapy with lovotibeglogene autotemcel (lovo-cel) uses the BB305 lentiviral vector encoding a modified β-globin gene, producing an antisickling HbA.²⁸ Exagamglogene autotemcel (exa-cel) reactivates fetal Hb synthesis by means of ex vivo CRISPR-Cas9 gene editing at the erythroid-specific enhancer region of BCL11A.²⁹ This approach was based on the fact that individuals with SCD who also inherit the condition of hereditary persistence of fetal Hb have increased levels of fetal Hb and reduced sickling, with limited or no symptoms of the disease.³⁰ Although the median follow-up was short (∼1.5 years in both studies), >85% of patients were free from acute vaso-occlusive pain episodes, most of which were mild.^28,29 Hb levels were generally above 11 g/dL, with some evidence of mild, asymptomatic hemolysis persisting.

Two patients on the lovo-cel trial developed findings consistent with myelodysplastic syndrome (MDS), including anemia, abnormal erythroid precursors on bone marrow examination, and trisomy 8 chromosomal abnormality. However, neither patient was officially given that diagnosis at the time of publication.²⁸ Moreover, 2 additional patients (of 9 patients) in initial studies with lovo-cel developed AML.^31,32 These cases of AML/MDS after SCD gene therapy do not appear to be directly related to prior hydroxyurea therapy,^33,34 myeloablative busulfan conditioning, or insertional mutagenesis³² after ex vivo manipulation.³⁵ Alternatively, the marrow cells themselves from patients with SCD may have an increased propensity to undergo malignant transformation after transplantation.³⁵ 

Several reports now show that donor cell AML/MDS after allo-HCT usually arises from clonal hematopoiesis (CH) present in the donor; this appears to be mostly in patients undergoing allo-HCT using older donors and often with a CH variant allele frequency below the standard level (2%-5%) of detection.^36,37 Similarly, it appears that AML/MDS that occurs after autologous transplantation for lymphomas arises from CH clones present in the patient’s marrow before transplant.^38-40 The cumulative incidence of AML/MDS after autologous transplantation for patients with lymphomas and pretransplant CH is nearly 20%.³⁹ The median time to development of AML/MDS is 3 to 4 years after transplant, but the risk continues for at least 10 years.^39,41,42 CH is a common age-related condition that is unusual before age 50 years but dramatically increases in prevalence after age 60 years.⁴³ Chemotherapy increases the incidence of CH, with a prevalence after chemotherapy estimated to be 25% to 30%.^38,44 Receiving chemotherapy is one of the few circumstances in which CH is seen in individuals aged <40 years.

CH also appears to occur in younger individuals with SCD.⁴⁵ The biological basis for early-onset CH in SCD is likely the unique SCD bone marrow microenvironment with chronic inflammation,⁴⁶ hypoxemia, and expanded hematopoiesis, which contribute to mutations. However, the incidence rate of MDS/AML in SCD is relatively low and similar to that associated with age-related CH.⁴⁷ Similarly, none of the donors in any reported donor leukemia cases have developed AML/MDS, consistent with the low rate of AML/MDS transformation associated with CH.^36,37 We postulate that the pathogeneses are similar for donor cell MDS/AML after allo-HCT, MDS/AML after autologous transplantation for lymphoma, and MDS/AML after autologous transplantation for SCD. Clinically silent premalignant clones that remain dormant under homeostatic conditions may undergo accelerated malignant transformation from the stress associated with the transplant process: a graft containing only 2% to 3% of a person’s normal number of HSCs must dramatically expand to restore lympho-hematopoiesis in the recipient fully.³⁵ The incidence of AML/MDS with gene therapies for SCD remains to be defined; the follow-up of the current studies is only ∼1.5 years,^28,29 whereas the median time to AML/MDS after autologous transplantation for lymphoma is 3 to 4 years and the risk continues for at least 10 years.^39,41,42 AML/MDS has also been seen after allo-HCT for SCD but only in host cells that were present either because of graft failure or mixed chimerism,⁴⁸ again apparently related to CH or perhaps from exposure to low-dose TBI⁴⁹ before allo-HCT.⁵⁰ 

The inability to find HLA-matched donors and the risks of GVHD, graft rejection, and transplant-related death continue to be cited as major limitations of allo-HCT for SCD by many in the field, including the recent publications on lovo-cel and exa-cel.^28,29 However, the recent evidence is that RIC and PTCy allow virtually all interested individuals with SCD (except those with severe heart, lung, or kidney disease) to receive a curative transplant, because almost all will have an appropriately matched or partially mismatched donor, either in the family or the unrelated registries.^24-27 With the use of pretransplant antithymocyte globulin and PTCy along with RIC with either 400 cGy TBI or the addition of thiotepa to 200 cGy, the risks of GVHD, graft failure, and serious GVHD are all <10% in adults.^24-27 Matched allo-HCT with other conditioning regimens and GVHD prophylaxis strategies such as abatacept have also shown promising results in children.^51-53 

Conversely, the current requirement for myeloablative conditioning^28,29 will limit the ability of many adult patients with SCD with end-organ dysfunction to undergo gene therapy approaches. Moreover, myeloablative conditioning will generally causes infertility, a major concern of patients with SCD,¹⁵ although the incidence could be less with single agent, pharmacokinetic-based busulfan.⁵⁴ The incidence of infertility with RIC regimens remains to be fully elucidated but would be expected to be low in young patients (aged <30 years).^55,56 Most patients with SCD undergoing RIC allo-HCT become full donor chimeras,^24-27 so eradication of sickle cell hematopoiesis with allo-HCT should be lifelong. Some degree of sickle cell hematopoiesis persists after busulfan-based conditioning for gene therapy as indicated by the continuing low-level, but clinically insignificant, hemolysis in the treated patients.^28,29 Thus, the durability of the remissions after genetic therapies still needs to be determined. Additionally, CH in patients with SCD may be a risk for AML/MDS after reinfusion of autologous gene-modified HSCs. However, the true incidence of this risk will only become clear with longer follow-up.

The cost-effectiveness of expensive new therapies always needs to be considered. The estimated cost of lovo-cel treatment is $3.3 million, which includes the 1-time drug acquisition price at a publicly available list price of $3.1 million, as well as the cost of stem cell collection, myeloablative conditioning, and postinfusion supportive care.^57,58 The list price for exa-cel is $2.1 million, so the total cost of treatment would be ∼$2.3 million.⁵⁸ Determining the cost of allo-HCT is a bit more challenging because charges depend on many variables such as type of transplant, state regulations, insurance contracts, and what is considered the transplant episode of care. An analysis of the cost of care for the year after myeloablative allo-HCT for SCD was a median of $467 747 (range, $344 029-$799 219),⁵⁹ although it is likely that RIC transplants with PTCy are nearly 50% less than this.⁶⁰ The average lifetime healthcare costs for an individual with SCD have recently been estimated to be $1.6 to $1.7 million.⁶¹ Importantly, healthcare costs underestimate the economic impact (eg, improved lifespan and productivity, and decreased need for unpaid caregiving). Any treatment that is curative or markedly ameliorates symptoms would be expected to have the same overall economic benefits.

Table 1 shows outcome measures associated with nonmyeloablative allo-HCT and the 2 currently available potentially curative genetic approaches (lovo-cel and exa-cel) for SCD. OS and EFS are quite similar among all 3 approaches, with most patients experiencing resolution of acute vaso-occlusive pain, although follow-up is relatively short. Despite continued references in the literature to limited donor availability and the risks of GVHD, graft rejection, and transplant-related death as constraints to allo-HCT for SCD,^28,29 these issues have been minimized with recent advances.^24-27 It perhaps should not be surprising that many in the field are unaware of these significant advances with allo-HCT for SCD, with analyses suggesting it takes at least 17 years⁶² for physicians to “unlearn” prior established principles that are no longer valid.^63,64 Graft failure remains higher in the pediatric population, and approaches to mitigate this issue are underway.

In contrast, the current need for myeloablation, cost, current insurance company restrictions to its use, and the (to date) poorly defined risk of posttransplant AML/MDS may significantly limit the availability of genetic therapy approaches. It is certainly possible that future advancements, such as the development of noncytotoxic conditioning that will allow engraftment of autologous stem cells, could make gene therapies less toxic. The optimal curative therapy choice for a specific individual with SCD remains to be defined. Moving forward, the curative therapy selection will not likely be based on the 2-year survival or amelioration of acute vaso-occlusive pain events. The clinical decision as to which curative therapy for an individual with SCD is selected will likely be determined based on the individual’s age, the family's choice, the insurance coverage, the ability to tolerate myeloablative therapy, and the yet-to-be-determined late health effects. Late health effects may include infertility; AML/MDS; and progressive heart, lung, and kidney disease, the major causes of death in adults with SCD. Questions also remain regarding which individuals should be considered for curative approaches as well as the optimal timing; for example, most children with SCD currently do well with supportive care.³ Ongoing development of new disease-modifying approaches could also influence decisions regarding therapies with curative potential. Moreover, ongoing advances will continue to lessen the complications and improve the availability of allo-HCT and genetic therapies for SCD. Notably, most individuals with SCD can now envision the potential for cure, with improved, if not average, quality of life and life expectancy as achievable goals.

Contribution: All authors contributed to writing the manuscript and approved the final version.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Richard J. Jones, The Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins, The Bunting Blaustein Cancer Research Building, Room 244, 1650 Orleans St, Baltimore, MD 21287; email: rjjones@jhmi.edu.