Allogeneic hematopoietic stem cell transplantation for STAT3 hyper-IgE syndrome: a worldwide study Open Access

Hyperimmunoglobulin E syndrome (HIES) caused by pathogenic germ line variants in signal transduction and activator of transcription 3 (STAT3-HIES, previously Job syndrome) affects <1 per million population.^1-3 The heterogeneous manifestations of this syndrome arise from the many cytokines that transduce through STAT3, and its ubiquitous expression in all cell lines. These manifestations include defects in innate and adaptive immunity, resulting in recurrent staphylococcal and fungal pulmonary and skin infections; eczema; allergy; connective tissue abnormalities, including poor wound healing, facial dysmorphism, and a tendency to minimal-trauma fractures; vascular abnormalities; and increased lymphoma risk.^3-6 Although our understanding of this disorder has grown through dissection of molecular mechanisms and by examination of clinical registries, the optimum strategy to reduce the frequency and sequelae of skin and pulmonary infection is not yet known. Despite antimicrobial prophylaxis directed against Staphylococcus aureus, which commonly causes pulmonary and skin infection, there is a high prevalence of bronchiectasis and chronic respiratory infection.^3,5,7 Immunoglobulin replacement therapy (RT) may also be used to counteract the common finding of suboptimal humoral responses to vaccine-strain antigens, although evidence of a beneficial effect in reducing infection or protecting against progressive pulmonary disease is limited.^3,5,8 Currently, patients with STAT3-HIES may expect to have recurrent pulmonary infection, progressing to parenchymal lung disease such as pneumatoceles and bronchiectasis, colonization with Aspergillus, and frequent skin abscesses complicating eczema despite antimicrobial prophylaxis and immunoglobulin RT.

Allogeneic hematopoietic stem cell transplantation (HSCT) is curative for many inborn errors of immunity (IEIs). Initial reports of HSCT in STAT3-HIES were not favorable. The first patient was successfully underwent transplantation for lymphoma but died of progressive lung disease, whereas the second patient had progression of extraimmune features and persistence of raised IgE.^9,10 However, subsequent publications of HSCT in patients with lymphoma showed an improvement in pulmonary and dermatological symptoms.^11,12 In a series collating the UK experience of HSCT in this disorder, including prolonged follow-up of the initial patient for whom HSCT was reported as “unsuccessful,” we demonstrated 100% survival and minimal peritransplant morbidity across follow-up ranging from 1 to 20 years, with resolution of skin disease, reduction in pulmonary infection, and freedom from antimicrobial prophylaxis and immunoglobulin RT.¹³ Reconstitution of a normal population of interleukin 17 (IL-17)–secreting T-helper 17 (Th17) lymphocytes after HSCT suggested the immune deficit intrinsic to STAT3-HIES may be cured by transplant, though its impact on extraimmune features of the syndrome is less clear, with ongoing minimal-trauma fractures, and 1 patient having a significant vascular event.^13,14 To date, information on 18 patients has been published.^7,15,16 

To determine the impact and limitations of HSCT, we conducted a retrospective study collating the known worldwide experience of transplantation for STAT3-HIES.

An international retrospective clinical study of the outcome of HSCT for patients with STAT3-HIES. Patients and centers were identified through previous publications, and through a wider STAT3-HIES Consortium comprised of clinicians across several international immunological and HSCT societies. Clinical data were collected from participating centers, including updated data for previously published cases (n = 18^11-13,15-18). All patients gave informed consent for sharing of deidentified data relating to HSCT and publication.

Inclusion criteria were HSCT for STAT3-HIES with a confirmed pathogenic, dominant negative STAT3 mutation and post-HSCT follow-up of at least 6 months for surviving patients.

The primary end points were overall survival (OS) and event-free survival (EFS). Events were death, chronic graft-versus-host disease (GVHD), or graft failure. Graft failure was defined as primary or secondary as per European Society for Blood and Marrow Transplantation (EBMT) criteria.¹⁹ Definitions for acute and chronic GVHD and graft failure followed modified Seattle criteria, National Institutes of Health consensus standards, and the EBMT criteria, respectively.^19-21 Definitions for veno-occlusive disease followed modified Seattle criteria.²² Patients were censored at the latest follow-up visit at which absence of events was confirmed. Patients who received a second HSCT were censored at second HSCT. A secondary end point of immunological remission was defined as survival without graft failure, need for immunoglobulin RT, eczema, or an excess of respiratory infections in patients with >1 year follow-up. Other outcomes of interest were improvement or progression of extraimmune disease manifestations (skeletal disease, scoliosis, vasculopathy, and malignancy).

Categorical data were compared with Fisher exact test on 2 tails. Nonparametrically distributed continuous data were compared with Mann-Whitney U test. Univariate survival data were computed using the Kaplan-Meier method, and compared using the log-rank test and presented with 95% confidence intervals. Multivariable analysis of factors affecting OS/EFS followed a Cox proportional hazards model. Statistical analysis was performed using IBM SPSS Statistics for Windows (version 28.0; Armonk, NY) and GraphPad Prism (version 9.5.0; San Diego, CA). P values <0.05 were considered statistically significant.

Informed consent for participation in retrospective studies was obtained from all individual participants or their parents, in line with institutional policy.

Forty-one patients were identified from 18 centers, receiving 43 transplants in total. The median age at HSCT was 14 years (4.0-45.0; Table 1), and 5 patients received their HSCT before 2010. The median follow-up duration for all patients was 4.5 years (0.8-28.0). Patients underwent HSCT predominantly owing to recurrent skin and pulmonary infection (n = 38 [92.7%]); in addition to these, 3 patients had received transplantation for lymphoma as the primary indication (2 with non-Hodgkin lymphoma, 1 with T-cell lymphoma).

Before HSCT, 38 patients (92.7%) had a history of chronic respiratory infections, with 28 (68.3%) having parenchymal lung disease in the form of bronchiectasis, pneumatoceles, or both. The median forced expiratory volume in 1 second and forced vital capacity were 84.0% (28.0%-122.0%) and 88.0% (26.0%-115.0%) of predicted, respectively, in 29 patients with data available. One-third of patients had received treatment for pulmonary fungal infection, predominantly from Aspergillus. One patient had prior pulmonary tuberculosis, and 1 had Mycobacterium abscessus infection.

Donor and graft characteristics are summarized in Table 2. Patients received peripheral blood stem cells (n = 21 [51.2%]) or marrow (n = 20 [48.8%]) from HLA 10/10–matched unrelated donors (n = 18 [43.9%]), matched family donors (n = 18 [43.9%]), mismatched family donors (including haploidentical, n = 4 [9.8%]), or a 9/10 mismatched unrelated donor (n = 1 [2.4%]). The median CD34⁺ stem cell dose received was 6.2 × 10⁶ cells per kg (range, 0.05 × 10⁶ to 22.0 × 10⁶ cells per kg). Conditioning regimens were predominantly treosulfan-based (n = 24 [58.5%]: dose, 42 g/m²; with thiotepa [n = 14]: dose, 10 mg/kg). Other patients received busulfan-based (n = 10 [24.4%]; dose targeted to area under the curve [AUC]: 37.7 mg∗h/L in 3 patients, to 65.7 mg∗h/L in 3 patients, and nontargeted in 4 patients) or melphalan-based (n = 7 [17.1%]) regimens. Serotherapy was used in 32 of 41 patients (78.1%): 22 (53.7%) received alemtuzumab and 10 (24.4%) received antithymocyte globulin. GVHD prophylaxis was most commonly calcineurin inhibitor-based (cyclosporin A or tacrolimus), plus mycophenolate mofetil or methotrexate. Three grafts underwent ex vivo T-cell receptor αβ/CD19 depletion. Posttransplantation cyclophosphamide was used in 7 transplants (17.1%). Two patients with lymphoma underwent 4 Gy total body irradiation as part of their conditioning regimen.

Thirty-seven patients were alive at the latest follow-up (92.5%). The median follow-up for surviving patients was 5.0 years (0.8-28.0). Deaths were transplant-related in all patients: disseminated adenovirus infection (day +86), Stenotrophomonas sepsis with hemoptysis (day +12), fungal pneumonia following high-dose corticosteroids and JAK inhibition for cryptogenic organizing pneumonia (day +134), and grade 4 acute GVHD and aspergillosis after second HSCT for graft failure (day +378 after first HSCT). OS at 2 and 5 years calculated by the Kaplan-Meier method was 92.5% (Figure 1). On both univariate and multivariable analysis, age at HSCT, presence and type of parenchymal lung disease, history of fungal infection, graft and donor choice, and conditioning intensity choice did not significantly associate with mortality (Tables 3 and 4).

EFS at 2 and 5 years was 89.9% (Figure 2). Two patients developed graft failure; 1 had hyperacute graft rejection at day +13 followed by a successful second HSCT, and the other had secondary graft failure with a second HSCT, but ongoing poor graft function in the context of GVHD and aspergillosis.

Overall, the incidence of grade 2 to 4 acute GVHD was 22.0%, and that of grade 3 to 4 acute GVHD was 12.2%; 4 patients developed grade 2 skin GVHD, 3 patients grade 3, and 1 patient grade 4 gastrointestinal GVHD, and 1 patient grade 4 skin, gastrointestinal, and liver GVHD. Four patients developed chronic GVHD: 2 surviving patients have mild cutaneous disease, 1 patient has severe chronic gastrointestinal GVHD, maximum score 3, now controlled on ruxolitinib, and the aforementioned deceased patient had severe acute-chronic overlap GVHD.

Eleven patients (26.8%) had deterioration in their respiratory status peri-HSCT; this was predominantly due to infective causes (fungal infection, cytomegalovirus pneumonitis, Stenotrophomonas infection), but additionally 2 patients developed cryptogenic organizing pneumonia. One patient had eosinophilic pneumonia and a ruptured pneumatocele causing dissemination of Aspergillus and a pneumothorax, requiring urgent lobectomy and later pleurodesis; and 1 patient had veno-occlusive disease.

One patient with significant pre-HSCT bronchiectasis and pulmonary nodules developed disseminated aspergillosis while aplastic. Although the patients survived after dual-agent antifungal therapy and granulocyte infusions until neutrophil engraftment, they developed secondary cutaneous aspergillosis presenting as skin nodules, a fungal mass in 1 kidney, and lumbosacral plexopathy in tandem with T-lymphocyte reconstitution. This manifested as mixed motor and sensory lower limb neuropathy with absence of nerve conduction in bilateral tibial, peroneal, sural, and superficial peroneal nerves. After immunomodulation with IV methylprednisolone and 2 courses of IV IgG (2 g/kg), the patient demonstrated improvement in motor examination, though remains reliant on walking aids.

At the last follow-up, the median donor chimerism was 100% in whole blood (26%-100%), 97.5% in CD3⁺ T cells (31%-100%), and 100% (25%-100%) in myeloid cells (CD15⁺ or CD33⁺; Figure 3C). Mixed donor chimerism was seen in 11 patients, but did not clearly associate with increased infection frequency or severity after HSCT. The incidence of mixed donor chimerism by conditioning regimen was as follows: treosulfan-based, 35%; melphalan-based, 25%; busulfan-based (AUC-guided), 33%; busulfan-based (non–AUC-guided), 0% (P = .70; Figure 3A,D). Immunological remission was seen in 29 of 36 surviving patients (80.5%) with >1 year follow-up. Immunological remission occurred more frequently in patients without pre-HSCT parenchymal lung disease (100% vs 72%), although this did not reach statistical significance (P = .072). There was no significant association in these 2 groups by age at HSCT (P = .350), pre-HSCT fungal infection (P = 1.00), conditioning regimen backbone (P = .270), or degree of donor chimerism (full vs mixed) at the latest follow-up (P = 1.00).

Freedom from, or reduced frequency of bacterial and fungal respiratory infection was reported by 32 of 37 surviving patients (86.5%), primarily among those with preexisting Aspergillus colonization and parenchymal disease causing ongoing cough. One patient required an emergency lobectomy due to life-threatening hemoptysis from fungal infection. No patients have developed new pneumatoceles after HSCT, and no patients required further thoracic surgery. Supplemental Figure 1 provides an example of pre-HSCT and post-HSCT thoracic imaging for patient 29. In 2 patients, worsening scoliosis has contributed to poor respiratory function. One patient with recrudescence of respiratory infection had mixed donor chimerism; the remainder of the patients with mixed chimerism have not had further infections.

One patient had recrudescence of eczema requiring phototherapy with complete donor CD3⁺ and myeloid chimerism; all other patients had resolution of clinically significant dermatitis predominantly within a year after HSCT. Two patients are on long-term immunoglobulin RT: patient 26, with protein-losing enteropathy; and patient 38, with chronic GVHD on ruxolitinib (supplemental Table 1).

Median serum IgE after HSCT was 457 kU/mL (10-11 813). In 6 patients for whom paired pre-HSCT and post-HSCT recordings were available, IgE levels after HSCT was significantly lower (Figure 3B; P = .03), typically by a factor of 10, though frequently remained above the upper limits of normal.

Two patients (4.9%) developed autoimmune cytopenias, both in the context of full donor chimerism; there was no other post-HSCT autoimmunity.

Eight surviving patients (19.5%) had new fractures after HSCT, predominantly of long bones and not related to significant trauma. One patient, who received steroids as treatment for immune thrombocytopenia after HSCT, developed left knee avascular necrosis. Their clinical characteristics and changes in bone densitometry, as available, are summarized in supplemental Table 2. Six patients (14.6%) had new or worsening of existing scoliosis.

Regarding vasculopathy, 1 patient (patient 28) who developed coronary ectasia leading to an anterior ST-elevation myocardial infarction and subsequent impaired left ventricular systolic function over a decade after HSCT was previously reported.¹⁴ This patient did not have any pre-HSCT evaluation for vasculopathy. At the latest follow-up, he has stable impaired cardiac function and has had no further vascular events. Two further patients developed new coronary artery tortuosity, not present at pre-HSCT evaluation, and only detected on screening magnetic resonance imaging, and 1 patient developed aortic dilatation that resolved spontaneously on follow-up. There were no additional vascular events after HSCT. Evaluation for any vasculopathy before HSCT occurred in a minority of patients.

One patient developed chronic protein-losing enteropathy of unclear etiology 10 years after HSCT and has significant pulmonary and peripheral edema affecting quality of life. This is associated with progressive mixed donor chimerism (latest whole blood chimerism, 33%) and T lymphopenia. The patient has not been fit to tolerate anesthesia for invasive investigations such as liver biopsy or endoscopy.

None of the patients with pre-HSCT malignancy had disease recurrence. One patient developed thyroid carcinoma, treated with resection, and a further patient had a meningioma resected.

This study includes data on all published cases of HSCT for STAT3-HIES, with the exception of the initial patient who received transplantation reported by Nester et al,¹⁰ and expands the experience of this treatment to 41 patients, many of whom received transplantation within the past 5 years. Given that HSCT and its results have been reported only infrequently in this disease, it is unsurprising that the study cohort reflects the severe end of the spectrum of STAT3-HIES, with very high rates of parenchymal lung disease and pulmonary fungal infection including multiple patients who have undergone lobectomy or pneumonectomy. With a median age at HSCT of 14 years, patients reported in this study were older than those in recent international studies of other non–severe combined immunodeficiency IEIs, such as chronic granulomatous disease (median [range], 7.0 [0.2-48.6] years) and CD40L deficiency (median [range], 4.0 [0.5-38.3] years), which may partially explain the high rates of organ damage accrued before transplantation.^23,24 Furthermore, the oldest patient in our series (patient 40) was 45 years old at HSCT with significant pretransplant morbidity including pneumonectomy. The presence of organ damage influences the outcome of HSCT for various diseases, with composite scores such as the Hematopoietic Cell Transplant– specific-Comorbidity Index (HCT-CI) and the Pediatric Adapted Risk Index (PARI) predicting how comorbidities influence transplant-related mortality risk.^25-27 Although the HCT-CI and PARI were not calculated for patients in this study, the presence of parenchymal lung disease and reduced forced expiratory volume in 1 second percentage alone would predict an increased risk of mortality in both the PARI (hazard ratio, 2.15) and HCT-CI (hazard ratio of 1.37 in the context of IEI HSCT) models.^25,26 Thus, OS >90% is promising, and comparable to the results of larger, international studies of HSCT in other individual IEIs such as chronic granulomatous disease or CD40L deficiency.^23,24 The fact that high survival was maintained at 2 and 5 years suggests that in these patients, as in many other HSCT indications, risk of adverse outcome is highest after transplant, and then stabilizes as shown by no late events even in the minority of cases with >10-year follow-up. With greater understanding of how to optimize the degree of myeloablation to permit engraftment and minimize infective and other end-organ toxicities, as well as changes in patient selection, we may thus be able to further improve outcomes by reduction in transplant-related mortality.

Restoration of STAT3 signaling within hematopoietic cells post-HSCT cures the immunodeficiency associated with STAT3-HIES, as previously reported, including post-HSCT detection of IL-17 producing Th17 lymphocytes.¹³ Most patients had reduced or no further respiratory infection, and eczema and skin disease almost universally resolved, likely related to restitution of the Th17/IL-17 axis and reversal of the Th2 response imbalance. The flares of infection seen in 5 patients are likely to reflect the significant degree of preexisting parenchymal lung disease. The lower proportion of patients in immunological remission (defined as freedom from immunoglobulin RT, eczema, and an excess of infections in alive patients with successful engraftment) among the subgroup with pre-HSCT parenchymal lung disease compared with those without pretransplant bronchiectasis or pneumatoceles, while only approaching statistical significance, suggests that earlier selection and referral for HSCT before development of parenchymal lung disease may offer improved outcomes in this disease. This is reinforced by the notable development of disseminated aspergillosis in 1 patient during the aplastic phase of HSCT, resulting in hemoptysis, skin nodules, a renal aspergilloma, and a subsequent immune reconstitution inflammatory syndrome affecting their sacral plexus.

The persistence of elevated serum IgE in many patients after HSCT raises further questions about how the hyper-IgE state is maintained despite restored STAT3 transcriptional activity, particularly given the short half-life of IgE in humans.^28,29 One putative explanation for this might be persistence of a population of IgE-producing plasma or tissue-resident memory B cells, as suggested for similar findings in DOCK8 deficiency after HSCT.³⁰ 

The failure to reverse existing connective tissue manifestations such as scoliosis and bone disease, which progressed in some patients, is consistent with the constitutional expression of the mutated STAT3 gene and the limitations of HSCT to correct the defect only in hematopoietic stem cell lines. There is insufficient understanding of the role of hematopoietic stem cell–derived cells, such as osteoclasts in the development of STAT3-HIES–related bone disease, or vasculopathy to help predict how their restored ability to signal through STAT3 might affect at a tissue or organ level. Minimal-trauma fragility fractures are described after pediatric HSCT, with incidence from one single-center study of 10.5%, though these were most commonly fractured vertebrae, rather than long bones, typically as from the detrimental impact of conditioning chemotherapy and corticosteroids on bone health.³¹ In this study, patients with post-HSCT fractures did not have clear differences in donor chimerism compared with those without fractures. Bone mineral density in STAT3-HIES and how it relates to fracture risk has proven difficult in the limited literature on this topic, and whether bisphosphonates may alter fracture risk remains unclear.³² Exploring whether restoration of STAT3 expression in hematopoietic cells results in any changes to mesenchymal stem cells may help us better understand the interplay of immune and extraimmune features of STAT3, such as how abnormal tissue remodeling influences recovery of lung disease. It is interesting that no further pneumatoceles have been reported in patients after HSCT, despite reasonable follow-up and, in some patients, subsequent lung infection. This may suggest that frequency and severity of lung infection is a greater driver of pneumatocele formation than the connective tissue defect, though this remains an observation and should be explored in greater detail, for example in murine models. The limitations of HSCT to correct only the immune phenotype of STAT3-HIES may be further supported by this study, in which development of new coronary tortuosity occurred after transplant, albeit in 2 patients who received transplant before adulthood. Longitudinal data on the natural history and progression of these lesions in the overall STAT3-HIES population will help contextualize whether HSCT has a neutral, beneficial, or detrimental impact on vasculopathy.

These results face similar limitations to other retrospective multicenter studies of HSCT for rare immunodeficiencies, namely a small and heterogeneous sample over a study period encompassing many changes to management of STAT3-HIES and HSCT for IEIs, which precluded identification of patient or HSCT factors influencing OS and EFS on multivariate analysis. Despite this, some observations may be made regarding the approach to HSCT: although conditioning regimens varied, treosulfan-based reduced-toxicity conditioning was the most common choice, with serotherapy choice depending on the type of donor used. This aligns with the European Society for Immunodeficiency and EBMT Inborn Errors Working Party guidance for allogeneic HSCT, which favors treosulfan or pharmacokinetically guided busulfan for a balance of sufficient myeloablation to allow engraftment without excessive toxicity.³³ The degree of donor engraftment required to immunologically cure STAT3-HIES is not clear; however, a number of patients have moderately low multilineage donor chimerism without obvious detriment, suggesting that a degree of mixed donor chimerism may be tolerated. Furthermore, rates of immunological remission were not significantly different between full and mixed donor chimerism groups. This may thus permit use of reduced toxicity, rather than fully myeloablative conditioning, which may be important to reduce the risk of transplant-related morbidity and mortality particularly from fungal infection and pulmonary disease. In the future, genome editing techniques may offer a promising therapeutic option by restoring STAT3 signaling without the attendant risks of allogeneic HSCT. A study using allele disruption techniques to functionally “knock-out” the disease-causing allele in cells was effective in rendering cells functionally haploinsufficient, though this was not sufficient to restore STAT3 signaling.³⁴ 

This study reports, to our knowledge, the largest cohort of patients receiving allogeneic HSCT for STAT3-HIES immunodeficiency or related lymphoma. These results help guide selection of patients for allogeneic HSCT: transplant-related morbidity in this cohort was predominately related to respiratory disease, and therefore transplantation before the development of significant parenchymal or fungal lung disease may improve outcome. Although the subset of patients at risk of parenchymal lung disease who have yet to develop it will garner the most benefit from HSCT, their identification relies on further research to characterize the natural history of STAT3-HIES. However, patients should be counseled regarding the lack of benefit and possible detrimental effect of HSCT on skeletal disease and other extraimmune features.

The authors acknowledge the patients, their families, and clinical and nonclinical staff who have contributed to the this study.

This work was supported by the Job Research Foundation (C.T.); the Germany Society for Allergology and Clinical Immunology and the “Foerderprogramm fuer Forschung und Lehre,” Faculty of Medicine, Ludwig Maximillian Univeristy Munich, Munich, Germany (J. Raedler); the National Institute for Health and Care Research University College Londo Hospital Biomedical Research Centre (E.C.M. and B.C.); the Medical Research Council (award MR/W01677X/1 [V.B.]); and in part by the Division of Intramural Research of the National Institutes of Health, National Institute of Allergy and Infectious Diseases, and National Cancer Institute, Centre for Cancer Research, the Warren Grant Magnuson Clinical Centre (A.F.F., D.D., C.G., D.A., and S.-Y.P.).

The content of this publication does not necessarily reflect the views of policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US government.

Contribution: C.T., A.W., A.F.F., and A.R.G. conceived and planned the study; C.T. collated data, performed analyses, and drafted the manuscript; and all authors contributed patient data, and reviewed, edited, and approved the final manuscript.

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

Correspondence: Christo Tsilifis, Paediatric Haematopoietic Stem Cell Transplant Unit, Great North Children’s Hospital, Newcastle Hospitals NHS Foundation Trust, Newcastle upon Tyne NE2 3QX, United Kingdom; email: c.tsilifis@nhs.net.