Hematopoietic Cell Transplantation for Inborn Errors of Immunity Other than Severe Combined Immunodeficiency in Japan: Retrospective Analysis for 1985–2016

Hematopoietic cell transplantation (HCT) is a curative therapy for most patients with inborn errors of immunity (IEI). We conducted a nationwide study on HCT for patients with IEI other than severe combined immunodeficiency (non-SCID) in Japan. Data from the Japanese national database (Transplant Registry Unified Management Program, TRUMP) for 566 patients with non-SCID IEI, who underwent their first HCT between 1985 and 2016, were retrospectively analyzed. The 10-year overall survival (OS) and event-free survival (EFS) were 74% and 64%, respectively. The 10-year OS for HCT from unrelated bone marrow (URBM), accounting for 39% of HCTs, was comparable to that for HCT from matched sibling donor (MSD), 79% and 81%, respectively. HCT from unrelated cord blood (URCB), accounting for 28% of HCTs, was also common, with a 10-year OS of 69% but less robust engraftment. The intensity of conditioning was not associated with OS or neutrophil recovery; however, myeloablative conditioning was more frequently associated with infection-related death. Patients who received myeloablative irradiation showed poor OS. Multivariate analyses revealed that HCT in 1985–1995 (hazard ratio [HR], 2.0; P = 0.03), URCB (HR, 2.0; P = 0.01), and related donor other than MSD (ORD) (HR, 2.9; P < 0.001) were associated with poor OS, and URCB (HR, 3.6; P < 0.001) and ORD (HR, 2.7; P = 0.02) showed a higher incidence of retransplantation. We present the 1985–2016 status of HCT for non-SCID IEI in Japan with sufficient statistical power, highlighting the potential of URBM as an alternative donor and the feasibility of reduced intensity conditioning.


Introduction
Inborn errors of immunity (IEI) comprise heterogeneous hereditary disorders affecting various components of innate and acquired immunity, including T and B lymphocytes, natural killer cells, phagocytes, macrophages, and complement proteins. Clinical manifestations of IEI are broad, including susceptibility to severe or opportunistic infections, autoimmunity, autoinflammation, allergic diseases, lymphoproliferation, and malignancies. They are increasingly being defined owing to recent advances in genetics and molecular sciences. In the most recent classification by the International Union of Immunological Societies (IUIS), 416 diseases have been enrolled as IEI [1]. In the present scenario, the collective prevalence of IEI is estimated to be at least 1 in 1,000 to 5,000 [2].
Hematopoietic cell transplantation (HCT) was first performed for a patient with severe combined immunodeficiency (SCID) in 1968 [3]. Since then, HCT has been widely applied as a curative therapy for patients with IEI, especially those with severe defects or dysregulation in cellular immunity. Unrelated cord blood (URCB) is commonly used in Japan and accounted for 33% of all allogenic HCTs from 2009 to 2018 [4]. We previously performed a nationwide survey in Japan involving 88 patients with IEI who underwent unrelated cord blood transplantation (URCBT) and demonstrated an overall survival (OS) of 69% over 5 years [5]. However, no study has covered all HCTs for patients with IEI in Japan. Recently, we conducted a retrospective analysis of HCT for SCID in Japan. In this study, we conducted a nationwide retrospective analysis of HCT for patients with non-SCID IEI to provide an overview of the status and outcomes of HCTs and develop strategies for HCT in Japan.

Data Collection
This study was approved by the Institutional Ethics Committee of Tokyo Medical and Dental University. The participants (and/or their guardians) provided written informed consents and were registered in the Transplant Registry Unified Management Program (TRUMP), an electronic database of all HCTs performed in Japan established by the Japanese Society for Transplantation and Cellular Therapy (JSTCT) [6]. The patients with non-SCID IEI who underwent their first HCT were included. The diagnoses of patients were collected according to the IUIS 2017 classification [7]. All transplant data were obtained from the TRUMP.

Study Endpoints
Neutrophil recovery was defined as the achievement of an absolute neutrophil count of ≥ 0.5 × 10 9 /L for 3 consecutive days. Platelet recovery was defined as the achievement of an absolute platelet count of ≥ 50 × 10 9 /L for 3 consecutive days, unsupported by transfusion for 7 days. Primary graft failure (PGF) was defined as the failure to achieve neutrophil recovery, and secondary graft failure (SGF) was defined as the event of an absolute neutrophil count of < 0.5 × 10 9 /L for 3 consecutive days after achieving neutrophil recovery. Because data on PGF, SGF, and chimerism is not available for some patients, and because most cases of PGF or autorecovery of recipient cells resulted in retransplantation or death, retransplantation or death was used as events for the analysis of event-free survival (EFS), to assess survival with sufficient donor cell engraftment. Only patients who achieved neutrophil recovery were included in the analysis for "late retransplantation" that was defined as a retransplantation after achieving neutrophil recovery.
The diseases were classified into the following categories: Wiskott-Aldrich syndrome (WAS), combined immunodeficiency (CID), hemophagocytic lymphohistiocytosis (HLH), chronic granulomatous disease (CGD), non-CGD phagocytic disorder, and primary immune regulatory disorder (PIRD) (see Table 2 for details). Regimens containing one of the following were classified as myeloablative conditioning (MAC) and total body irradiation (TBI) at a total dose of ≥ 800 cGy, busulfan at a total dose of > 8 mg/ kg, or melphalan at a total dose of ≥ 150 mg/m 2 , according to the Center for International Blood and Marrow Transplant Research criteria [8] and previous studies [9,10]. Other regimens were classified as reduced intensity conditioning (RIC), excluding 4 patients (DiGeorge syndrome, n = 3; and CHARGE syndrome, n = 1) who did not receive any chemotherapy or radiation. HLA matching was determined by serology for patients from the initial years and by genotype for those in the more recent years. We used the term "matched" to refer to those 8/8, or 6/6 matched who lacked the HLA-C loci data, especially among patients from the initial years. The donor type was classified as matched sibling donor (MSD), other related donor (ORD, matched or mismatched non-MSD related donor), unrelated cord blood 1 3 (URCB), and unrelated bone marrow (URBM), except for 1 case without evaluable donor data.
We defined active bacterial or fungal infection at HCT as the infection that required systemic antibiotic therapy on the day of HCT. We defined respiratory impairment as conditions that met either of the following: hemoglobin adjusted D LCO < 80%, FEV 1 < 80%, requirement of oxygen, or shortness of breath with/without mild exertion. The whole blood cell chimerism was evaluated from 100 days to 1.5 years after HCT, and the patients who died before achieving neutrophil recovery were excluded. Chimerism was classified as follows: complete (≥ 95% donor chimerism), donor dominant (< 95% and ≥ 80% donor chimerism), mixed (< 80% and ≥ 20% donor chimerism), and low (< 20% donor chimerism or patients who required retransplantation). Posttransplant short stature was defined as a stature at least 2 standard deviations below the mean height for age and sex, and data was requested to be updated annually.

Statistical Analysis
OS and EFS were calculated using the Kaplan-Meier estimates, and the impact of the independent risk factors on OS and EFS was evaluated using Cox proportional hazard models. Retransplantation, neutrophil recovery, platelet recovery, acute graft-versus-host disease (aGVHD), and chronic graftversus-host disease (cGVHD) were analyzed using a cumulative incidence method. Death was considered a competing event for retransplantation. Death and retransplantation were considered competing events for neutrophil recovery, platelet recovery, aGVHD, and cGVHD. The cumulative incidence of cGVHD was calculated and was limited to the patients who survived more than 100 days after HCT. Gray's test was used to compare the cumulative incidence, and the Fine-Gray model was used to evaluate the impact of independent risk factors on the cumulative incidence of retransplantation. For the final multivariate analysis, variables were selected by potential factors that are considered to affect the outcome or according to the results of the respective univariate analyses. The cumulative incidence, OS, EFS, and hazard ratios (HRs) are reported with 95% confidence intervals (CIs). All statistical analyses were performed using the Stata software v16.1 and EZR 1.42 [11]. Two-sided P < 0.05 was considered significant. All values enclosed in brackets represent 95% CI unless otherwise specified.

Patient Characteristics
A total of 566 patients with non-SCID IEI comprising 451 (80%) males and 115 (20%) females who underwent HCT between 1985 and 2016 were included in this study. The median duration of follow-up was 4.2 years (1 day-30.8 years) for all participants and 5.4 years (52 days-30.8 years) for survivors. The characteristics of the participants with comparison according to time periods are shown in Table 1, and the precise diagnoses are shown in Table 2. The patients with CGD (126; 22%), WAS (118; 21%), familial hemophagocytic lymphohistiocytosis (FHL) (101; 18%), and SCN (59; 10%) commonly underwent HCT. The patients received HCT at a median age of 4 years (0-64 years), and the median time from diagnosis to HCT was 1.7 years (0-36.8 years). According to disease category, there was a significant difference in the interval between diagnosis and HCT, showing shorter intervals in patients with HLH, WAS, and longer and more varied intervals in those with CGD (P < 0.001; Fig. S1). Additionally, differences in various factors were observed between the disease categories (Table S1).
Notably, the characteristics of HCT changed significantly over time (Table 1). More patients received RIC, especially melphalan-based RIC in the later period. The HCT from unrelated donors, especially from URBM, has increased in recent years, which may have been associated with increased in vivo T cell depletion. Furthermore, patients with CGD received HCT more frequently in the recent period. However, as a whole non-SCID IEI, there is no significant difference in the age at diagnosis, age at HCT, or time from diagnosis to HCT between each period.

Overall Survival and Event-Free Survival
The summary of the transplant outcomes over the entire period is shown in Table S2. The 10-year OS and EFS were 74% [69-78%] and 64% [60-69%], respectively. We did not observe a significant difference over time in OS or EFS (Fig. 1a, b).

Hematologic Recovery, Graft Failure, and Retransplantation
Among evaluable patients, PGF was observed in 44 (8%, out of 564) cases and SGF was 33 (7%, out of 487) cases (Fig. 2). Retransplantation was performed in 61 (11%) cases with a median interval of 178 days (20 days-9.8 years) from HCT. Only one patient with FHL, one with hyper IgE syndrome, and one with CD40L deficiency survived without retransplantation even though they did not achieve sufficient donor engraftment. The HCT from URCB was associated with slower hematologic recovery than the other donor types (median days  Fig. S4a). However, the difference in the incidence of late retransplantation was not significant between URCB and MSD, or URBM (Fig. S4b). The incidence of HCT from URBM with two or more locus mismatches was associated with frequent retransplantation (10-year: 4% [1-9%] for matched URBM; 5% [1-12%] for 1 mis URBM; 16% [6-31%] for ≥ 2mis URBM; P = 0.04; Fig. S2c). The HLA disparities in URCBT were associated with a tendency of higher incidence of retransplantation (Fig. S2d). Patients with HLH had slower platelet recovery than those with other diseases (Fig. 3d), which is probably due to the high frequency of URCBT (60%) for HLH (Table S1), and URCBT in HLH patients showed a tendency of higher incidence of retransplantation (Fig. S5b). URCBT in CGD patients also showed an increased incidence of   retransplantation or late retransplantation (Fig. S5c,i). However, the association of URCBT and retransplantation was not significant in the other disease categories. Platelet recovery was faster in the patients who received RIC conditioning (Fig. 3f). Although retransplantation was not correlated with the intensity of conditioning (Fig. S4e), late retransplantation was more frequent in RIC than MAC (Fig. S4f). From a specific disease perspective, RIC was associated with a higher incidence of retransplantation in patients with WAS (P = 0.03, Fig. S5m) and HLH (P = 0.06, Fig. S5n). For patients with WAS or HLH, the melphalanbased regimen tended to have a higher incidence of retransplantation than the busulfan-based regimen (Fig. S5s,t). Of note, the incidence of URCBT in the RIC group was common in HLH (72%) but not in WAS patients (29%). The high incidence of retransplantation in HLH patients who received RIC regimens was probably influenced by the frequent use of URCB for these patients.

Chimerism
More robust whole blood donor chimerism was achieved in patients who received HCT from MSD or URBM than those who received HCT from ORD or URCB (Fig. S6a). The donor dominance in chimerism was not significantly different between MAC and RIC regimens (Fig. S6b). The HLA disparity in URBM or URCB and disease category did not show a significant difference in terms of donor chimerism (data not shown).
Other HCT-related complications and the details of the post-transplant infections are shown in Tables S3 and S4, respectively. Bacterial infection (P = 0.005) and gonadal dysfunction (P = 0.006) were more commonly seen in patients who received MAC regimens, whereas the frequency of other complications did not significantly differ between different intensities of conditioning (Table S5). Viral and fungal infection was observed in 24% and 10% of the patients evaluated, respectively. Bacterial infections were observed in 37% of the evaluable patients at a median of 9 days (1-335) after HCT, and 72% of them developed infections before the neutrophil recovery was achieved. URCBT tended to have a slightly higher incidence of post-HCT bacterial infection (P = 0.006), but not of post-HCT viral infection (Table S6). Of note, serotherapy was used in 42% of the URCBT, less frequent than HCT from URBM (74%). Among evaluable patients (n = 541), 14 (3%) patients developed malignancy, most of which were associated with post-transplant lymphoproliferation. Short stature occurred in 18% of the patients. The patients with short stature at HCT were more likely to develop post-transplant short stature than those without (P < 0.001, Fig. S7a). Among patients with normal stature at HCT, 7% of them developed short stature after HCT. On the contrary, among patients with short stature at HCT, 66% of them improved after HCT. We did not observe a significant difference in the incidence of short stature between conditioning; however, MAC regimen containing TBI ≥ 800 cGy tended to cause post-transplant short stature more frequently than other regimens (Fig. S7b).

Cause of Death
Death from infection was the most common, accounting for 43 (33%) cases (Table S7). Among the patients who died, the MAC regimen was commonly associated with death from infection (P = 0.02, Table 6). In contrast, the death from organ dysfunction was relatively common in the patients who received RIC regimens (Tables 6 and S8). Notably, some of these deceased patients, including those with CGD or FHL, presented with poor immunologic reconstitution or failure to control the primary disease, suggesting that the toxicity of conditioning was not directly responsible for deaths in certain cases. The donor type or disease category did not correlate with differences in the cause of death (data not shown).

Discussion
Our results show a 10-year OS of 74% for patients with non-SCID IEI, who underwent their first HCT between 1985 and 2016, which is comparable to that from multicenter studies in other countries (Europe, 69% over 10 years [12]; Australia and New Zealand, 72% over 5 years [13]; Brazil, 72% over 5 years [14]; Colombia, 62% over 5 years [15] We demonstrated the effect of URBM and URCB on the outcome of HCT for non-SCID IEI in Japan. The HCT from URBM was the most frequently performed, showing comparable 10-year OS to that for HCT from MSD (79% vs 81%, respectively). The equivalent outcome for HCT from URBM and MSD has also been reported from other countries [12,13]. Although the incidence of aGVHD was high with HCT from URBM, the excellent survival was partly due to robust hematologic recovery and sufficient donor engraftment. The preparation for HCT from URBM takes several months in Japan, and our analysis reconfirms that URBM can be considered a useful alternative donor source for stable patients who have enough time to prepare for HCT.
The OS for URCBT over 10 years was 69%. Although the OS was inferior to that for HCT from MSD, this might be acceptable for patients who require urgent transplantation and do not have MSD. A similar incidence of GVHD in URCBT and HCT from MSD also suggested its utility in Japan. However, the engraftment after URCBT was not robust, as evident from a slow hematologic recovery and less sufficient donor engraftment. While the risk for OS in URCBT in the recent period may have been reduced, multivariate analyses consistently demonstrated that URCB was an independent risk for poor EFS and retransplantation. URCBT for SCID patients in Japan showed excellent outcomes, including OS and engraftment [16]. However, the disadvantage for engraftment is well known in the HCT for hematologic disorders other than SCID [17][18][19][20][21]. Despite the ready availability and feasibility of URCBT, we recognize the risk for poor engraftment for non-SCID IEI as a whole.
For patients who received HCT from ORD, most of which are mismatched relatives, we observed a poor OS/EFS, as well as poor engraftment and a high incidence of cGVHD. In our cohort, post-transplant cyclophosphamide or TCRαβ + / CD19 + depletion, which are beginning to be adopted in haploidentical HCTs for IEIs worldwide [22][23][24][25][26][27] as well as in Japan [28], was not used in most of the cases. The introduction of these novel techniques would be expected to expand donor options and improve the outcome of HCT from ORD in the coming decades. Furthermore, gene therapy for Table 6 Association between cause of death and intensity of conditioning in the deceased patients Patients who received no conditioning (n = 3) or unspecified intensity of conditioning (n = 9) and patients with unknown cause of death (n = 5) were excluded from this analysis; a Fisher's exact test; b among them, death by organ dysfunction occurred in 16 patients (multi-organ failure, n = 5; liver failure, n = 5; cardiac failure, n = 2; central nervous system dysfunction, n = 2; renal failure, n = 2; also see Table S8) numerous IEIs, including SCID, WAS, CGD, and leukocyte adhesion deficiency, is being developed [29]. Promising results for these novel approaches should improve the prognosis of IEI patients without suitable donors. We analyzed the association of conditioning regimens and the outcomes of HCT. In the recent decade, RIC regimens have been commonly chosen. Although RIC regimens may have been associated with late retransplantation, the OS and donor chimerism for RIC regimens were not significantly different from those for MAC regimens, indicating sufficient efficacy of RIC regimens. In our cohort, MAC regimens were more commonly associated with death from infection. Considering the higher incidence of bacterial infection in patients who received MAC regimens, we speculate that strong tissue injury associated with MAC, such as mucosal damage, probably contributed to infection-related deaths. Furthermore, TBI-based MAC regimens specifically showed poor OS in all disease categories, as well as a higher tendency toward post-transplant short stature. RIC regimens potentially reduce short-and/or long-term conditioningrelated toxicities and are considered suitable in HCT for IEI.
We showed the risk of respiratory impairment at HCT on OS. The strong association between respiratory impairment and infection implied that the infection was responsible for dyspnea in most patients. The presence of infection alone was not associated with poor survival, but infection and subsequent pulmonary damage could be a risk. The pre-HCT management for non-infectious manifestations, as well as infectious events, is equally important. For instance, it is well known that the remission status of HLH is associated with good survival after HCT [30,31]. Several targeted therapies have been developed for IEI in recent years, such as anti-interferon-γ antibody for HLH [32], JAK inhibitor for HLH [33], or STAT1 or STAT3 gain-of-function [34], CTLA4-Fc fusion protein for CTLA4 haploinsufficiency [35] or LRBA deficiency [36], and PI3K inhibitor for activated PI3Kδ syndrome [37]. Those novel pharmacological treatments are expected to control the disease activity as bridging therapies before HCT.
Besides the results for non-SCID IEI as a whole, IEI comprises heterogeneous diseases. Each disorder is associated with different backgrounds of the patients (Fig. S1 and Table S1) or outcomes of HCT (Fig. S3, S5). In patients with WAS, similar outcomes for URBM, URCB, and MSD confirmed that URBM and URCB were preferable alternative donors, in agreement with the finding from other studies [38,39]. RIC regimens showed equivalent OS but increased incidence of retransplantation in this study. In contrast, busulfan-based regimens were associated with less incidence of retransplantation than melphalan-based regimens, which is also consistent with the findings of a previous report [38].
The interval between diagnosis and HCT was the shortest for patients with HLH compared with that for patients with other diseases, indicating the urgency for HCT. URCB was the most commonly chosen for these diseases probably owing to rapid availability. The 10-year OS for URCBT was 58%, which was similar to that reported from Europe [40] and Japan [41]; however, OS was not satisfactory compared to that of MSD (10-year OS: 79%), and we also observed a higher incidence of retransplantation after URCBT. Further approaches, including optimal conditioning regimen, exploring indication of haplo-HCT with post-transplant cyclophosphamide [23], or better pre-HCT disease control using molecular-targeted therapies [32,33], would be necessary for improving the management of HCT in the coming decades.
In patients with CGD, the outcome for HCT from URBM and MSD was equivalent. Multivariate analysis showed that CGD was a risk for retransplantation as well as poor EFS, and URCBT for CGD showed a high incidence of retransplantation. The patients were more commonly complicated with infection or respiratory impairment at HCT (Table S1), which may also pose a risk for infection, concerning poor engraftment in URCBT. As this study has shown, URCBT for CGD patients is also reported to have poor engraftment [42,43]. Because the time between diagnosis and HCT was relatively long, URCB may be used for these diseases only on limited occasions. The intensity of conditioning showed no difference in outcomes in this study, and prospective clinical trials have also shown that a fludarabine/busulfanbased RIC regimen is effective in CGD patients [44,45]. Thus, RIC is recommended for these diseases to reduce regimen-related toxicity, especially in recipients with concurrent infection.
Our study has several limitations. First, some important information, such as the precise genotype of the diseases, was not available in the TRUMP registry for the patients in the earlier period, which might have reduced the sample size and affected the analyses. The data such as genotype of the HLA or donor chimerism are also partially unavailable, especially in some early patients. Second, the TRUMP registry was not oriented for the HCT for IEI; active viral infection at HCT or disease-specific complications that might affect the outcome of HCT were missing. The data of immunologic reconstitution after HCT, such as lineage-specific chimerism or discontinuation of immunoglobulin, as well as sequential data for chimerism were also unavailable. Third, a precise analysis of each disease was not performed. We provided some insights for the preferred management of HCT for some disease categories. However, to establish better disease-specific management, it is important to conduct a precise evaluation for each disease through retrospective analyses, prospective studies, or trials for novel therapeutic modalities. For further detailed analysis, we have already published retrospective studies for each IEI from Japan [42,46,47] and plan to perform such studies for other diseases in the future on behalf of the Hereditary Disorder Working Group of the JSTCT, collaborating with the Primary Immunodeficiency Database in Japan [48] and the TRUMP.
In conclusion, we present an overview of the backgrounds and outcomes of HCT for non-SCID IEIs in Japan with a large number of patients for sufficient statistical power. We demonstrate that the OS for HCT from URBM and MSD was almost equivalent in Japan, confirming URBM as an alternative donor source in HCT for non-SCID IEI. URCBT, which was also commonly performed in Japan, showed substantial applicability for some diseases but posed a high risk for poor engraftment. We also demonstrate the efficacy of RIC regimens and highlight the importance of disease control before HCT. These results should contribute to the development of future management strategies for IEIs in Japan. Furthermore, detailed evaluation for individual IEI, along with recent advances in novel therapeutic approaches, needs to be addressed for establishing an optimal HCT strategy for each disease.