HLA-haploidentical stem cell transplantation for chronic granulomatous disease: an EBMT-IEWP retrospective study Free

Chronic granulomatous disease (CGD) is an inborn error of immunity (IEI) affecting phagocyte function due to impaired NAD phosphate oxidase activity. The disease is associated with 6 different genotypes with autosomal recessive [CYBA (p22^phox), CYBC1 (CYBC1), NCF1 (p47^phox), NCF2 (p67^phox), and NCF4 (p40^phox)] or X-linked [CYBB (gp91^phox)] patterns of inheritance.^1,2 The impaired NAD phosphate oxidase leads to a defect of reactive oxygen species, such as superoxide production, resulting in defective microbial killing. CGD manifests as recurrent bacterial and fungal infections.³ Dysregulated inflammatory responses frequently result in granuloma formation, and related inflammatory conditions, such as colitis, frequently occur.⁴ Growth failure related to chronic colitis and/or long-term steroid administration is a common feature of the disease. Although CGD can manifest from infancy to late adulthood, diagnosis typically occurs before the age 5 years. The use of lifelong antibacterial and antifungal prophylaxis, associated, in selected cases, with cautious use of anti-inflammatory drugs to reduce inflammation, has significantly improved overall survival (OS) rates, allowing for patients to reach adulthood. However, progressive organ damage continues to occur over time due to increasing disease burden and medication side effects.^5,6 Patients generally experience both diminished quality of life and premature death.^6-9 

Allogeneic hematopoietic stem cell transplantation (HSCT) as a treatment option for CGD was introduced first in 1977.¹⁰ Over the last 2 decades, HSCT using suitably HLA-matched donors has proven to be an effective curative option for patients with CGD and has been increasingly proposed to patients with high-risk features (such as absence of oxidase activity, as well as serious infectious or inflammatory complications).^11-16 Previous transplant studies have demonstrated OS rates and event-free survival (EFS) exceeding 85% and 75%, respectively, when an HLA-matched donor is available. HSCT has been shown to lead to resolution of persistent infections and inflammation, thereby improving organ dysfunction, growth, and quality of life compared with pretransplant state.^7,17 These very good results, together with the high disease burden, further underline the role of early HSCT as a curative approach in CGD.

In the absence of an HLA-matched donor, the place of alternative transplants with a mismatched related donor (MMRD) in CGD is not well established, and experience is still scarce.^18,19 In the 2 largest series on patients with CGD who underwent transplant, reported by Chiesa et al and Leiding et al (including 712 and 240 patients, respectively), 36 and 21 individuals underwent transplant with an MMRD.^15,20 In both series, OS, EFS, and incidence of acute graft-versus-host disease (GVHD; aGVHD; grade 2-4) were worse than HLA-matched donors in multivariate analysis. Incidence of graft failure was also found to be significantly increased in Chiesa et al.¹⁵ 

However, HLA-haploidentical HSCT has been increasingly used in IEI patients in need of HSCT but lacking HLA-matched donors, thanks to the development of 2 platforms: in vitro TCRαβ/CD19 depletion of the graft and in vivo depletion with posttransplant cyclophosphamide (PTCY). These 2 approaches showed very encouraging results in the field of IEI, but their role in MMRD HSCT for CGD remains poorly defined.^21-26 The aim of our multicenter, retrospective study was to expand the sparse knowledge available, which mainly comes from case reports and 2 small case series.^19,27 

A retrospective multicenter study of genetically and/or functionally confirmed patients with CGD was conducted on behalf of the Inborn Errors Working Party of the European Bone Marrow Transplantation (study number 8427025). Patients who received a first transplant from a MMRD between 2014 and 2022, with either in vitro T-cell depletion using T-cell receptor αβ(TCRαβ)/CD19 depletion (referred to as TCRαβ/CD19) of the graft or in vivo T-cell depletion (referred to as PTCY) were included (number of patients per center in each group is shown in supplemental Table 2, available on the Blood website). Data were retrieved from the EBMT registry, and an additional study-specific questionnaire was filled in by physicians of the contributing centers. Informed consent for registration and data collection within the EBMT database was obtained from all patients, their parents, or legal guardians according to the ethical principles of the Declaration of Helsinki.

The main outcomes of interest were OS, EFS, GVHD-free EFS (GEFS), and cumulative incidence (CI) of GVHD. OS was defined as survival from the first HSCT to the last follow-up or death. EFS was defined as survival without graft failure or second procedures (CD34⁺ stem cell boost and conditioned second transplant). GEFS was defined as survival without graft failure, second procedures, grade 3 to 4 aGVHD, or chronic GVHD (cGVHD), whichever occurred first. Engraftment kinetics (time to neutrophil engraftment was defined as the first day of achieving a neutrophil count ≥0.5 × 10⁹/L for 3 consecutive days; time of platelet engraftment was defined as the first day of achieving a platelet count ≥ 20 × 10⁹/L without platelet transfusion for at least 7 days), graft failure, toxicities, and degree of donor hematopoietic chimerism at engraftment and at the most recent assessment were also collected. Pre-HSCT CGD-related medical problems (infections within 6 months before transplant and inflammation) and their resolution at the last follow-up were also analyzed.

Quantitative values were expressed as the median (interquartile range or range, as indicated), and qualitative values are presented as numbers (percentages). Univariate analysis was performed using the Fisher exact test or the Wilcoxon test, as appropriate, using the CreateTableOne package in R. Survival analysis was performed using the survival package for OS, EFS, and GEFS. A log-rank test was applied to test for significant differences between the TCRαβ/CD19 and PTCY platforms. A Cox proportional hazards model was also applied for OS, EFS, and GEFS, incidence of graft failure, aGVHD, and cGVHD. Forest plots were created using a forest model. Competing risks were taken into account for CI of primary graft failure (PGF; death) or GVHD (death or rejection) using the cmprsk package with the cuminc() function. All tests were 2-sided, and a P value <.05 was considered statistically significant. Data visualization was performed using ggplot2. The analyses were performed using R version 4.2.3 (The R Project For Statistical Computing, Vienna, Austria; http://R.project.org).

Patient characteristics are summarized in Table 1. Sixty-four genetically confirmed patients with CGD from 20 centers in 13 countries were included. The median age at diagnosis was 1.5 years (range, 0-15), and 53 patients (83%) were male. X-linked CGD was the predominant inheritance pattern (Table 1), and the patients’ genetic background is depicted in supplemental Table 1. Twenty-nine patients (45%) had a CGD-related infection within the 6 months before HSCT. Among these infections, 15 were of fungal origin (8 aspergillosis, 6 unspecified fungal infections, and 1 mucormycosis), and 6 were in partial remission at the time of HSCT (Table 1). Forty-four patients (67%) had pre-HSCT inflammation, including 24 (37.5%) receiving immunosuppressive drugs at the time of transplantation. Ten patients had active inflammation despite treatment at the time of HSCT (Table 1). The primary indications for HSCT included infection severity (42.2%), combined severity of infection and inflammation (29.7%), and inflammation alone (17.2%; Table 1). Notably, 7 patients underwent elective transplantation, defined as the absence of any clinical issues related to CGD (infection/inflammation within the 6 months before transplant; Table 1).

Forty patients (62.5%) received haploidentical HSCT with in vitro TCRαβ/CD19 depletion of the graft, and 24 patients (37.5%) received in vivo T-cell depletion with PTCY (Table 1). Mean age at first transplant was 5.8 years (range, 0-33). The stem cell source was primarily peripheral stem cells (79.7%). The median donor age was 35 years (range, 2.8-53). Busulfan (Bu)– and treosulfan-based conditioning regimens (CRs) were administered in 24 (37.5%) and 34 patients (53.1%), respectively, whereas 6 patients (9.4%) received other combinations (Table 2; supplemental Table 1). For Bu, pharmacokinetic drug monitoring data were available in 16 of 24 patients who had a mean area under the curve (AUC) of 70.9 mg × h/L (range, 45-85). Fifty-eight patients (90%) received T-cell–directed serotherapy with rabbit antithymocyte globulin (ATG; n = 21), rabbit antilymphocyte globulin (n = 26), or alemtuzumab (n = 11). Doses and timing within the CR differed (supplemental Table 1). Characteristics of patients and HSCT according to the type of graft depletion methods are shown in Tables 1 and 2. Age at diagnosis, mode of inheritance, frequency of infections within the 6 months before HSCT, and rate of inflammatory complications were similar in both groups. The status of the last infection differed, with more patients presenting with active infection in partial remission at the time of transplant in the PTCY group than the TCRαβ/CD19 group (P = .029). Patients in the PTCY group were more frequently on immunosuppression for inflammatory complications at transplant than those in the TCRαβ/CD19 group, although the difference was not significant (Table 1). The mean age at first transplant differed significantly between the TCRαβ/CD19 group (4.1 years) and the PTCY group (8.8 years; P < .002; Table 2). The source of stem cells also differed, with peripheral blood stem cells alone used in the TCRαβ/CD19 group, whereas peripheral blood stem cells and bone marrow were equally used in the PTCY group (Table 2). CRs differed significantly, with more Bu-based conditioning in the PTCY group (66.7%) and more treosulfan-based conditioning in the TCRαβ/CD19 group (80%; P < .001; Table 2). The donor/recipient cytomegalovirus status was similar between groups (Table 2).

The median time to neutrophil and platelet recovery was 14 days (range, 9-42) and 13 days (range, 0-77), respectively, and both were significantly later in the PTCY group than TCRαβ/CD19 group (supplemental Figure 1A-B; neutrophils, 16.5 days [range, 10-42] vs 13 days [range, 9.0-24.0]; P < .001; platelets, 17 days [range, 0-77] vs 13 days [range, 0-21]; P = .036; Table 3). PGF occurred in 12 patients, with a CI of 20.6% (Table 3; Figure 1A). This event was independent of the T-cell depletion approach (Figure 1B), age at diagnosis, and patient and transplant characteristics (supplemental Figure 1C). In contrast, CRs differed, with overrepresentation of Bu-based CR in patients with PGF (P = .03; supplemental Figure 1C). Of note, within the TCRαβ/CD19 group, 4 of 8 patients (50%) conditioned with a Bu-based CR had PGF vs 3 of 32 (9.3%) who received a treosulfan-based CR (supplemental Table 1). There was no difference in Bu AUC levels between patients who underwent engraftment and those who experienced graft failure, although data were limited.

Of the 12 patients with PGF, 3 died between 30 and 71 days after transplantation without undergoing retransplantation; 7 underwent retransplantation, at a median of 57 days after the first HSCT (range, 37-109); and 4 successfully engrafted and are alive, whereas 3 died of infection. One patient did not undergo retransplantation, and information is missing for another patient.

Among the 52 patients who underwent engraftment, 49 (94.2%) displayed full donor chimerism (>95% donor) within 100 days after HSCT, whereas 3 (5.8%) had mixed chimerism (>80% donor). At the last evaluation (median, 510 days [range, 115-2533]), chimerism data were available in 49 patients, 43 patients displayed full donor chimerism, and 6 patients had mixed chimerism (with myeloid chimerism >95% donor in 2; 80%-95% donor in 3; and 42% donor in 1; whereas T-cell lineage chimerism was available in 3 and found >95% donor).

The median follow-up after transplantation was 1.9 years (interquartile range, 0.9-3.1; range, 0.08-6.9). The 3-year OS rate in the entire cohort was 75.9% (CI, 63.9%-90.1%; Figure 1C), with 85.6% (CI, 71.4%-100%) in the PTCY group and 67.4% (CI, 49.6%-91.7%) in the TCRαβ/CD19 group (P = .26; Figure 1D). Age at transplantation (with a cutoff of 5 years) and pre-HSCT status (inflammation, infection, and the need for immunosuppressive treatment) were not associated with a worse outcome (supplemental Figure 1D). Among the 12 patients who died, 8 deaths were related to infections (66.7%), 2 to aGVHD (16.7%), and 2 from other causes. Of note, 6 of the 12 patients who experienced PGF died, and the deaths were related to this complication.

The 3-year EFS was 70.2% (CI, 58.6%-84%) in the whole cohort, with 73.1% (CI, 56.4%-94.6%) and 66.7% (CI, 50.5%-88.2%) in the PTCY and TCRαβ/CD19 groups, respectively (P = .74; Figure 1D-E). No patient characteristics were associated with a worse outcome (supplemental Figure 1E). GEFS was 56.1% (CI, 43.7%-71.9%) at 3 years (supplemental Figure 1F), 57.9% (CI, 40.5%-82.8%) in the TCRαβ/CD19 group, and 48.8% (CI, 32%-74.3%) in the PTCY group (P = .2; supplemental Figure 1G). Of note, the use of Bu in the CR was associated with worse GEFS outcomes (65% in the treosulfan-based group vs 45% in the Bu-based group; P = .031; supplemental Figure 1H).

Three patients (4.7%) experienced post-HSCT veno-occlusive disease (Table 3). Two patients were diagnosed with posttransplant microangiopathy and 1 patient with pulmonary hemorrhage (Table 3). The CI of grade 2 to 4 and grade 3 to 4 aGVHD (with death and PGF as competing risk factors) were 28.1% (CI, 17%-39%; Figure 2A; supplemental Figure 2A) and 10.9% (CI, 3.2%-18.6%), respectively, and were significantly higher in the PTCY group than TCRαβ/CD19 group for grade 2 to 4 (P = .04) but not for grade 3 to 4 (P = .07; Figure 2B; Table 3; supplemental Figure 2B). In multivariate analysis, the use of ATG (hazard ratio, 11.73; CI, 1.40-98.12; P = .02) compared with alemtuzumab and the use of PTCY instead of the TCRαβ/CD19 platform (hazard ratio, 3.55; CI, 1.22-10.36; P = .02) were associated with an increased risk of grade 2 to 4 aGVHD (Figure 2C-D). Six patients (10.3%) experienced cGVHD (4 extensive and 2 localized) without a significant difference between groups (CI, 10.4%; supplemental Figure 2C).

Within the first 3 months after HSCT, 37 patients experienced at least 1 infection (57.8%), including 23 (57.5%) in the TCRαβ/CD19 group and 14 (57.3%) in the PTCY group (P = .57). Viral infections were the most frequent (n = 26 events), followed by bacterial (n = 9 events) and fungal infections (n = 7 events). Among these, 5 were de novo infections (no documented or suspected fungi infections in the 6 months before HSCT). Among these 5 patients, 4 suffered from concomitant PGF and died due to this infection. The nature of infections was not different between PTCY and TCRαβ/CD19 groups (Table 3). Nine patients experienced infections between 3 and 12 months after HSCT (14.1%), with 3 in the TCRαβ/CD19 group and 6 in the PTCY group. These infections were of fungal (n = 3 events), viral (n = 2 events), bacterial (n = 2 events), or of unknown origin (n = 2). All patients with late post-HSCT fungal infections experienced aGVHD. Finally, no patients developed a new CGD-related inflammation after HSCT, and patients with pre-HSCT inflammation were in remission after HSCT.

Extensive published experience demonstrates the curative benefits of HSCT with an HLA-matched donor in patients with CGD. Thanks to the development of MMRD transplants with in vitro TCRαβ/CD19 depletion or in vivo depletion with PTCY, alternative donors have been increasingly used in IEI and have led to significant improvements of outcomes.²¹ However, experience with transplant in patients with CGD using these platforms is still scarce, witch motivated us to conduct this retrospective multicenter study.

The 64 patients of our study cohort corresponded to patients with advanced CGD and high disease burden. Nearly half of the patients had experienced an acute CGD-related infection in the 6 months before transplant; 10% of the cohort had an uncontrolled fungal infection, which accelerated the decision to go for a high-risk transplant; and two-third of the patients suffered from autoinflammatory disease, resulting in nearly 40% of the cohort being on immunosuppressive medication at the time of transplant. These results may indicate a selection bias toward the most severely affected comorbid patients, who would be expected to have poorer HSCT outcomes. Because HSCT with MMRD is a fairly new approach, especially for CGD (85% of the transplants were performed after 2017), it is likely that mainly severely affected patients with the highest need for immediate curative treatment have been selected for this approach. It is, therefore, likely that the results of HSCT with MMRD will be even more favorable when performed in younger patients with less active infection and inflammation.

The characteristics of patients between the TCRαβ/CD19 and PTCY groups showed some differences. Patients in the PTCY group were significantly older and showed a trend toward more severe diseases, with significantly more active infections within 6 months and more high-risk transplants done due to uncontrolled fungal infection (5 vs 1). There was also a trend for a higher proportion of patients under immunosuppressive treatment before HSCT in the PTCY group.

The 3-year OS and EFS rates of the entire cohort were 75.9% and 70.2%, respectively. These outcomes are in line with previous reports on transplants with mismatched donors (MMRD and mismatched unrelated donor [MMUD]) but inferior to the OS and EFS reported with HLA-matched donors.^11,15,16,20 Age at transplant, the type of CR, and the type of T-cell depletion of the graft did not significantly affect OS and EFS in this cohort. The lack of correlation with age may be attributed to the young age of this cohort, with a mean age of 5.8 years, and only 6 patients were adolescents or young adults (age >12 years). Infectious status, presence of inflammation, and use of immunosuppression before HSCT did not affect patient outcomes in this series, as previously reported.²⁰ 

The incidence of PGF was 20.6%, higher than the 13% reported by Chiesa et al but close to the range reported by Leiding et al (17.6%).^15,20 Engraftment remains challenging in HSCT for CGD. The presence of HLA mismatches in the context of MMRD transplants may exacerbate this obstacle. It is possible that the inflammatory process present in CGD, with an interferon gamma type II signature shown in hematopoietic niches by Sobrino et al,²⁸ predisposes to rejection; the deleterious role of interferon gamma in engraftment has been previously shown in IEI, although in other contexts.²⁹ 

The choice of the T-cell depletion method, either in vitro TCRαβ/CD19 depletion or in vivo depletion with PTCY, did not significantly affect OS, EFS, CI of graft failure, or GEFS, but it did influence the CI of aGVHD. The overall CI of grade 2 to 4 and grade 3 to 4 aGVHD (28.1% and 10.9%, respectively) were comparable with those reported by Chiesa et al and Leiding et al, who showed a CI of 18% to 20% for grade 2 to 4 and 6% to 9% for grade 3 to 4 aGVHD. In univariate and multivariate analyses, PTCY was associated with an increased risk of grade 2 to 4 aGVHD, possibly in line with differences in the quality of alloreactive T-cell depletion. In a previous report by Parta et al on 7 patients with CGD who received haplotransplant with PTCY conditioned with a reduced intensity conditioning (RIC) containing total body irradiation (TBI) of 2 Gy, aGVHD was also an issue contributing to 2 deaths.¹⁹ Importantly, in this report, we did not observe a significant difference in grade 3 to 4 aGVHD, which has more significant impact on morbidity and mortality. The use of ATG or the absence of serotherapy was also associated with an increased risk of grade 2 to 4 aGVHD compared with alemtuzumab or rabbit antilymphocyte globulin. It has been shown that the magnitude of graft exposure to serotherapy targeting T cells influences the incidence of graft failure and GVHD and kinetics of immune reconstitution in T-cell–replete transplants.^30-32 This highlights the importance of optimizing T-cell–directed serotherapy in the setting of haploidentical transplantation.

CR varied in this retrospective series, with as many as 7 different regimens. The choice of myeloablative drugs was primarily driven by the method of depletion, as previously reported.²¹ Bu-based CR was significantly associated with an increased incidence of PGF and lower GEFS. Pharmacokinetic data for Bu, although not reported in all patients, showed a median total AUC in the expected range (median, 73.5 mg × h/L), not suggesting underexposure to Bu.^16,33 These differences were not observed in the report of Chiesa et al, but the data for Bu AUC were insufficient in that study, and most transplants were HLA matched.¹⁵ A higher rate of primary and secondary graft failure was observed with treosulfan-based CR than Bu-based CR in a retrospective study on Wiskott-Aldrich syndrome, a prospective randomized trial in nonmalignant diseases and a single-center study in CGD.^34-37 At this point, given the limited available data, no definitive explanation can be provided to account for this difference. The higher incidence of PGF when Bu is combined with TCRαβ/CD19 depletion suggests that this combination should be avoided.

In conclusion, MMRD HSCT represents a justified and feasible option for patients with CGD with severe disease burden who lack matched donors. Despite challenges, such as graft rejection and aGVHD, survival outcomes are promising, underscoring the importance of early consideration and timely intervention with MMRD transplantation for this high-risk patient population. Future research should focus on refining transplant protocols to minimize the risk of graft failure and prevent aGVHD and on ensuring stable engraftment. This will further improve outcomes and broaden access to this life-saving therapy for individuals with CGD who lack HLA-identical donors.

The authors thank patients and families, and all physicians and staff of hematopoietic stem cell transplantation units who took care of the patients.

Contribution: B.N., T.G., M.S., and M.H.A. designed the study; Q.R. and B.N. analyzed and interpreted data and wrote the manuscript; J.K., T.S., K.H., B.B., M.Z., C.S., M.F., F.L., D.M., A.R.G., M.H., and T.G. reviewed and revised the manuscript; and M.H., S.H.L., A.L., K.H., H.H., M.K., H.S., F.P., B.B., M.Z., S.A., V.B., C.S., J.B., M.F., C.W., W.H., K.M., F.L., A.P.M., F.S., D.M., A.R.G., D.B., M.H.A., M.S., T.G., and B.N. recruited patients and provided clinical care.

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

Correspondence: Bénédicte Neven, Hospital Necker-Enfants Malades, Assistance Publique–Hospitaux de Paris, INSERM, 149 Rue de Sèvres, 75015 Paris, France; email: benedicte.neven@aphp.fr.