Allogeneic hematopoietic cell transplantation for partial RAG deficiency in children and adults: Excellent outcomes with a reduced-intensity posttransplantation cyclophosphamide–based approach

Abstract

Background:

Partial recombinase activating gene deficiency (pRD) leads to combined immunodeficiency with immune dysregulation. It can be cured by allogeneic hematopoietic cell transplantation (HCT), but optimal referral criteria and approaches remain to be defined.

Objective:

Our study evaluated low-toxicity approaches to HCT for pRD.

Methods:

Thirteen children and adults with pRD received radiation-free, predominantly reduced-intensity conditioning (pentostatin/cyclophosphamide/busulfan) HCT with posttransplantation cyclophosphamide–based graft-versus-host disease (GVHD) prophylaxis at median (range) age 20 (4–46) years.

Results:

With median 2.6 years’ follow-up, overall survival for the entire cohort was estimated at 92% and 83% at 1 and 2 years and 100% and 90% for reduced-intensity conditioning recipients (n = 12), with 2 deaths attributed to sepsis. Reversal of clinical manifestations was associated with immune reconstitution, with minimal de novo autoimmunity, 15% 1-year cumulative incidence of grade III-IV acute GVHD, and no chronic GVHD. V𝛂7.2-positive T-cell proportion increased rapidly after HCT, while mucosa-associated invariant T-cell reconstitution lagged. Dysreactive CD19^hiCD21^lo and 9G4⁺ B cells decreased after HCT, along with clinically relevant autoantibodies. However, baseline elevated anti–type I interferon antibodies, potentially predisposing to severe viral infections, decreased slowly, although neutralizing activity was reduced at last follow-up. Outcomes did not differ by donor carrier status or HLA matching. Bronchiectasis exacerbations incurred rehospitalizations in long-term follow-up of patients who entered HCT with irreversible lung disease.

Conclusion:

Reduced-intensity conditioning HCT with posttransplantation cyclophosphamide–based GVHD prophylaxis is safe and effectively reverses immune dysfunction in patients with pRD.

Capsule summary:

Allogeneic hematopoietic cell transplantation with reduced-intensity pentostatin/cyclophosphamide/busulfan conditioning and posttransplantation cyclophosphamide–based graft-versus-host disease prophylaxis is safe and reverses immune dysfunction in patients with partial RAG deficiency, with greatest benefit before irreversible lung disease has developed.

Graphical Abstract

Hypomorphic, biallelic mutations in the human recombination-activating genes RAG1 and RAG2 result in partial preservation of recombinase activity. This condition, also referred to as partial recombinase activating gene (RAG) deficiency (pRD), is a cause of combined immunodeficiency, often associated with autoimmunity and/or granulomatous disease presenting in childhood or even adulthood.¹ In the setting of RAG loss-of-function mutations leading to severe combined immunodeficiency or Omenn syndrome, newborn screening has facilitated early diagnosis and prompt referral to allogeneic hematopoietic cell transplantation (HCT) affording excellent outcomes.² However, the optimal referral criteria and timing of HCT for patients with pRD has not been as well delineated. In a worldwide cohort of 60 patients with pRD treated by HCT, 1-year and 4-year survival were 77.5% and 67.5%, respectively, and 12-month cumulative risk of acute and chronic graft-versus-host disease (GVHD) were 35% and 22%, respectively. New-onset post-HCT autoimmunity was observed in 7 patients (11.7%). HCT at an earlier age and before development of organ damage was associated with improved immune reconstitution and superior overall survival.³ Thus, effective but low-toxicity approaches are needed for pRD patients who are diagnosed at a young age but are already symptomatic, and for those who remain undiagnosed until significant disease manifestations and organ damage have already occurred.⁴

METHODS

Thirteen children and adults with biallelic RAG1 or RAG2 mutations and clinical manifestations of pRD were enrolled onto natural history study NCT03394053 and received HCT on either of two prospective clinical trials (NCT02579967 and NCT03663933),^5,6 all conducted at the National Institutes of Health (NIH) Clinical Center. Institutional review board approval was obtained for each study, and informed consent and/or assent was obtained from each patient and/or guardian according to institutional standards.

Genetic diagnosis was confirmed via Sanger sequencing, and recombination activity was evaluated to confirm partial recombinase activity (see Table E1 in the Online Repository available at www.jacionline.org).^7–9 Related donors were confirmed to carry no more than one RAG mutant allele.

Patients enrolled onto the NCT02579967 study (n = 11) received radiation-free, reduced-intensity conditioning (RIC) with pentostatin (4 mg/m², 2 doses), hyperfractionated, low-dose cyclophosphamide (3–5 mg/kg per day for 8 days), and 2 days of pharmacokinetically dosed busulfan with a daily area under the curve target of 4600 μmol × min (18.9 mg × h/L), T-cell–replete bone marrow grafts, and GVHD prophylaxis consisting of posttransplantation cyclophosphamide (PTCy) 50 mg/kg per day on days +3 and +4, followed by sirolimus with or without mycophenolate mofetil. Patients enrolled onto NCT03663933 (n = 2) received distally timed equine antithymocyte globulin as part of their RIC (n = 1) or nonmyeloablative (n = 1) conditioning, along with T-cell–replete peripheral blood stem cell grafts and tacrolimus in lieu of sirolimus (Fig 1). All non–protocol-specified immunomodulatory therapies were stopped no later than day 0.

FIG 1.

HCT platforms for NCT02579967 (A and B) and NCT03663933 (C and D). (A) Patients on original study arm (P1, P2) received mycophenolate mofetil (MMF) through day +35, while subsequent patients participated in MMF duration deescalation study, receiving MMF either through day +18 (P6, P7, P11, P12) or not at all (P3, P4, P5). (B) Patients with matched donors are currently enrolled onto a compressed conditioning schedule using reduced-dose PTCy with the goal of reducing toxicity but preserving low GVHD incidence (P8, P13). (C) This serotherapy-containing platform was geared toward patients with significant immune dysregulation and uncontrolled cytopenias (P9) and/or (D) comorbidities that precluded significant chemotherapy exposure (P10).

HCT recipients were prospectively assessed for engraftment, HCT complications, organ function, virus control and other infections, immune reconstitution, and phenotype reversal (see the Methods section in the Online Repository available at www.jacionline.org).

RESULTS

Thirteen patients referred to the NIH Clinical Center with clinical manifestations suspicious for an inborn error of immunity were found to carry biallelic deleterious RAG variants (Table E1). Median (range) age at diagnosis of pRD was 17 (3–43) years. Twelve patients had a history of recurrent sinopulmonary infections, 9 with bronchiectasis and/or scarring (severe in 5 patients). Nine patients experienced viral infections, mostly represented by Herpesviridae. Clinical manifestations of autoimmunity were present in 9 patients (Table I). Severe granulomatous disease involving the skin and occasionally deep tissues was present in 4 patients. In addition to antimicrobials, 7 patients required immunoglobulin replacement therapy before HCT. Systemic immunosuppression such as corticosteroids, TNF blockers, rituximab, bortezomib, and others had been required in 9 (69%) of 13 patients, and baseline soluble biomarker studies showed hyperactivation of IFN-γ and downstream chemokines (see Fig E1 in the Online Repository available at www.jacionline.org). One patient (P13) had a history of diffuse large B-cell lymphoma, and another (P2) had nonmalignant lymphoproliferation. Chronic kidney disease was present in the two oldest patients. One patient (P8) had a positive newborn screening result at birth; after identification of a biallelic RAG1 variant (p.C176F), referral to HCT occurred after she developed severe influenza, varicella zoster virus infection, and lung scarring.

TABLE I.

Baseline characteristics of 13 patients

Characteristic	Value
Male sex	7 (54%)
Genetic defects leading to partial RAG1/2 deficiency
RAG1 compound heterozygous mutations	8 (62%)
RAG1 homozygous mutations	4 (31%)
RAG2 compound heterozygous mutations	1 (8%)
Age (years) at genetic diagnosis, median (range)	17 (3–43)
Age (years) at HCT, median (range)	20 (4–46)
Years between diagnosis and HCT, median (range)	1.4 (0.1–5.9)
Sino-oto-pulmonary infections	12 (92%)
Bronchiectasis and/or scarring	9 (69%)
Mild	4 (31%)
Severe	5 (38%)
Systemic steroid-dependent lung disease	3 (23%)
History of emergent intubation	2 (15%)
Autoimmunity	12 (92%)
Requiring systemic immunomodulatory therapy	6 (46%)
Cytopenias	3 (23%)
Other*	3 (23%)
Vitiligo and/or alopecia	5 (38%)
Thyroiditis	2 (15%)
Infertility (suspected autoimmune cause)†	2 (15%)
Type I diabetes mellitus	1 (8%)
Serologic, not requiring therapy‡	6 (46%)
Granulomas	4 (31%)
Hematologic malignancy	1 (8%)
Pre-HCT IDDA2.1 score, median (range)	72.6 (30–114)

Data are presented as nos. (%) unless otherwise indicated.

^*Inflammatory myopathy (n = 1); rheumatoid arthritis and polyarteritis nodosa (n = 1); arthralgias treated with hydroxychloroquine (n = 1).

^†Premature ovarian failure (n = 1), azoospermia (n = 1).

^‡Anti-thyroid antibodies without thyroid disease (n = 4), anti-GAD65 antibodies without evidence of glucose intolerance (n = 1), direct agglutination test positive with mild anemia but no evidence of hemolysis (n = 1).

The 13 patients underwent allogeneic HCT at a median (range) age of 20 (4–46) years (Table I, Table E1). Six of these have been previously reported.^3–5,10 The median (range) time from diagnosis to HCT was 1.4 (0.08–5.9) years. These were high-risk patients, with 9 having an HCT comorbidity index score of 3 or higher.

RIC regimens were used for all but P10, who received a nonmyeloablative, busulfan-free regimen with intermediate-dose PTCy, given the patient’s underlying comorbidities and familial cancer predisposition, which precluded any greater alkylator exposure. Among RIC recipients, all received serotherapy-free HCT with marrow grafts, except P9, who received serotherapy along with a peripheral blood stem cell graft to facilitate lymphodepletion and engraftment in the setting of significant T lymphocytosis in blood and marrow associated with neutropenia (Fig 1). HLA-haploidentical donors were used most frequently (n = 5), followed by HLA-matched sibling donors (n = 4); 78% of related donors were heterozygous carriers.

With a median (range) survivor follow-up of 2.6 (0.5–8.1) years (Table II), 1- and 2-year overall survival were estimated at 92% and 83%, respectively (Fig 2, A). Two deaths occurred, both attributed to postengraftment sepsis. P2 developed fulminant febrile illness with purpura and shock, which was clinically concerning for meningococcal sepsis (unconfirmed), at day +487, despite receiving no immunosuppressive therapy for the previous 10 months, with good immune reconstitution. He had not received post-HCT meningococcal vaccination. P10 died at day +100 in the setting of diabetes mellitus, systemic steroid therapy, and renal replacement therapy, despite broad-spectrum antibiotic therapy; Enterococcus faecium was posthumously detected in the blood. Of note, these were the two patients with the highest baseline Immune Deficiency and Dysregulation Activity v2.1 (IDDA2.1) scores in the cohort.^11,12 When analyzing only patients receiving the preferred RIC approach (n = 12), 1- and 2-year overall survival were estimated at 100% and 90%, respectively (Fig 2, A).

TABLE II.

Recipient demographics, donor characteristics, and clinical outcomes

Patient no.	Patient age and sex at HCT	HCT-CI	Donor relatedness and genetics (recombinase activity)	Donor age and sex	GVHD	Other noninfectious complications	Viral infections*	Other infections	Follow-up	Outcome
P1	37M	1	10/10 MUD	22M	Grade 2 acute (gut only)	None	BK-C; shingles, recurrent HSV (acyclovir noncompliance); COVID-19; influenza twice	Demodex folliculitis, pneumonia s/p influenza	8.1 y	• A&W, off IgRT and IS • Functionality improved, negative ENA antibody • Warts improved
P2	4M	2	Haploidentical RAG1 p.R973H, heterozygous (32%)	26M	None	None	BK-C; oral HSV (noncompliance)	Mucocutaneous candidiasis, secondary pneumonia	1.3 y (died)	• AIHA resolved, off IgRT • Died of septic shock
P3	42M	3	10/10 MUD	21M	None	SOS	Severe BK-C; severe orofacial HSV; norovirus; adenoviremia requiring therapy	MRSA pneumonia/sepsis	6.4 y	• A&W, off IgRT and IS • Granulomas, alopecia resolved • +bronchiectasis exacerbations • Eosinophilic cystitis
P4	8F	4	MSD RAG1 p.R112H, heterozygous (41%)	6M	None	Ovarian insufficiency requiring HRT	Influenza twice; BK viremia/viruria	Otitis media → PE tubes, pneumonia	6.3 y	• A&W, not requiring IgRT and IS • +bronchiectasis exacerbations
P5	24M	3	10/10 MUD	29M	None	None	CMV infection (no disease); BK-C; COVID-19		6.3 y	• A&W, off IgRT and IS • Granulomas, alopecia resolved
P6	34F	3	Haploidentical unaffected	31M	None	None	HSV requiring intravenous therapy (suspected BK naive)	Clostridium difficile, Demodex folliculitis	4.3 y	• A&W, off IgRT and IS • Vitiligo, alopecia improved • Negative anti-thyroid antibodies • +recurrent sinusitis and bronchiectasis exacerbations
P7	9M	3	Haploidentical RAG1 p.G393V, heterozygous (0%)	16M	None†	Duodenal hematoma, HA	CMV infection (no disease); BK-C	Streptococcus mitis bacteremia	2.6 y	• A&W, off IgRT and IS
P8	10F	0	MSD RAG1 p.C176F, heterozygous (25.8%)	8F	None	None	No CMV infection on letermovir; (suspected BK naive)	None	1.7 y	• A&W, not requiring IgRT or IS
P9	20M	3	MSD RAG1 p.K86VfsTer33, heterozygous (2.7%)	22F	None	None	CMV infection (no disease); BK-C with mild but prolonged symptoms; influenza	None	2.2 y	• A&W, off IgRT and IS • AIN resolved • Vitiligo improved, CMV controlled • Negative GAD65 antibody
P10	46F	10	8/10 mMUD	27M	Grade 2 acute (skin, gut)	Renal failure. respiratory failure	Adenoviremia not requiring therapy; BK viremia/viruria (no cystitis)	S mitis bacteremia; Scedosporium pneumonia; Giardia (during conditioning)	100 d (died)	• Died of septic shock (suspected E. faecium) • TMA/PRES • Some granuloma healing
P11	7F	3	Haploidentical RAG1 p.R396H, heterozygous (0.5%)	15M	Grade 3 acute (liver, skin)	None	BK viremia/viruria (no cystitis); RSV (mild)	Legionella pneumonia, sinusitis	1.4 y	• A&W, not requiring IgRT • No active GVHD, off IS • Healing granulomas; new poliosis and vitiligo
P12	11M	1	Haploidentical unaffected	25M	None	Bell palsy related to dental procedure	CMV infection (no disease); BK-C (mild); RSV (mild)	None	1.2 y	• A&W, not requiring IgRT or IS; vitiligo repigmentation
P13	26F	4	MSD RAG1 p.C176F, heterozygous (25.8%)	32F	Grade 3 acute (liver, skin)	Seizure related to hypomagnesemia	CMV infection (no disease); BK viremia/viruria (no cystitis)	Staphylococcus aureus disseminated skin infection and pneumonia, suspected fungal pneumonia	6.3 mo	• Alive, outpatient with resolved acute GVHD tolerating systemic IS taper

A&W, Alive and well; AIHA, autoimmune hemolytic anemia; ENA, extractable nuclear antigen; AIN, autoimmune neutropenia; BK-C, BK virus–associated cystitis; COVID-19, coronavirus disease 2019; HA, hemolytic anemia; HCT-CI, HCT comorbidity index; HRT, hormone replacement therapy; HSV, herpes simplex virus; IgRT, immunoglobulin replacement therapy; IS, immunosuppression; mMUD, mismatched unrelated donor; MRSA, methicillin-resistant Staphylococcus aureus; MSD, HLA-matched sibling donor; MUD, HLA-matched unrelated donor; PE, pressure equalization; RSV, respiratory syncytial virus; SOS, sinusoidal obstructive syndrome.

^*Respiratory viral infections are only listed if patient received directed therapy or required hospitalization.

^†Possible acute GVHD of gut; histopathologic findings not definitive and symptoms attributed to adverse effect of foscarnet; resolved without any treatment beyond continued sirolimus prophylaxis.

FIG 2.

Overall survival (A) and CD3⁺ and myeloid donor chimerism over time, presented as box and whisker plot of minimum, first quartile, median, third quartile, and maximum (B). Myeloid chimerism for P10 (sole patient who did not receive busulfan) is presented separately from other patients’ myeloid chimerism outcomes.

Engraftment

Neutrophil and platelet recovery occurred at median (range) day +17 (14–19) and day +31 (17–163); in 3 patients, platelet count never declined below 20,000/mm³. No graft failures occurred, and no patients required unplanned subsequent donor cell infusions. Median donor CD3⁺ chimerism reached 95% by day +28, and remained in the 98% to 100% range throughout the follow-up period (Fig 2, B). For other lymphocyte subsets, median donor chimerism reached ≥95% by day +42 (see Fig E2 in the Online Repository available at www.jacionline.org). In recipients of busulfan-containing regimens (n = 12), median donor myeloid chimerism ranged 98% to 100% throughout follow-up and as early as day +21 (Fig 2, B).

GVHD

Acute GVHD occurred in 4 patients (P1, P10, P11, and P13) for a 1-year overall cumulative incidence of 31% (grade III-IV, 15%), with no significant differences between carrier versus unaffected donors or HLA-mismatched versus HLA-matched donors (see Fig E3 in the Online Repository available at www.jacionline.org). In P1 and P10, GVHD was maximal grade II and steroid responsive; it resolved in P1 but persisted at time of death for P10. P11 and P13 developed maximal grade III GVHD (liver and skin), possibly with an initial component of drug toxicity in P13; disease was steroid refractory but responsive to ruxolitinib in both. By 13 months after HCT, P11 was off all immunosuppression without evidence of active GVHD; P13 remains in early follow-up. Remarkably, no chronic GVHD occurred in any patients.

Infectious complications

The incidence of cytomegalovirus (CMV) requiring preemptive therapy was high. Among known CMV-seropositive patients (n = 7), all of whom had seropositive donors and all of whom received letermovir prophylaxis, 5 patients (P5, P7, P9, P12, and P13) developed CMV infection requiring preemptive therapy after HCT, 4 of them twice, for a median (range) overall preemp tive treatment duration of 27 (21–94) days; no patients developed CMV disease. No CMV infections or disease occurred in two other at-risk patients with seropositive donors but seronegative or unknown recipient status.

Severe or recalcitrant herpes simplex virus infections requiring intravenous therapy occurred in 2 patients. Importantly, no other severe herpesvirus-related disease manifestations, such as Epstein-Barr virus–associated posttransplantation lymphoproliferative disease or human herpesvirus 6–associated encephalitis, were observed.

Adenoviremia without signs of disseminated disease, yet requiring therapy, occurred in 1 patient.

BK viruria and/or viremia was detected in 11 of 13 patients, and BK virus–associated cystitis occurred in 7. Notably, symptoms were particularly severe and/or prolonged in 2 patients: P3, with gross hematuria lasting 174 days, requiring instrumentation, and dysuria for over a year, and P9, who had a mild initial course but continues with residual dysuria ongoing for 2 years, even after BK viruria clearance.

Bacterial lung infections (pneumonia vs bronchiectasis exacerbation) requiring hospitalization beyond the immediate post-HCT period occurred only in patients who already had severe bronchiectasis at HCT. P3 and P6 each had a single, short hospitalization years after HCT, while P4 required recurrent hospitalization along with intermittent bronchoalveolar lavage for clearance of secretions. Other bacterial infections requiring systemic therapy included Staphylococcus aureus (pneumonia/sepsis, n = 1; pneumonia alone, n = 1; skin/soft tissue infection, n = 1), Streptococcus mitis bacteremia (n = 2), Legionella pneumonia (n = 1), and the two deaths attributed to sepsis.

Immune dysregulation

Granulomatous disease universally and remarkably improved within months of HCT even if patients remained on immunosuppression (Fig 3, A and B). Repigmentation of vitiligo patches and, in patients with alopecia, hair regrowth were noted (Fig 3, C-E). No immune cytopenias recurred after HCT, and only one patient developed direct agglutination test–negative hemolytic anemia, which resolved with a single dose of intravenous immunoglobulin at day +46. More severe pre-HCT autoimmune manifestations, such as myositis leading to progressive weakness in P1, stabilized without further damage in the absence of long-term immunosuppression. Accordingly, IDDA2.1 scores decreased in all patients after HCT (see Fig E4 in the Online Repository available at www.jacionline.org).

FIG 3.

Improvement of granulomas (A, P5; B, P10), alopecia (C, P5), and vitiligo (D, P6; E, P12) before and after HCT.

Interestingly, P11 developed poliosis as early as day +100 after HCT and was later diagnosed with vitiligo. Given the early onset, it is not known if a causative clone was present before HCT and persisted, versus whether this was a donor-derived phenomenon on the GVHD spectrum, albeit one not diagnostic for GVHD.

Immune reconstitution

Patients with pRD have variable degrees of T-cell lymphopenia, but they consistently have a low number of naive CD4⁺ cells.⁹ A similar pattern was present at baseline in the 13 patients reported here. After HCT, progressive increases in the absolute CD3⁺, CD4⁺, CD8⁺, and naive CD4⁺ cell counts, as well as in Treg cells, were observed (Fig 4, A; and see Fig E5 in the Online Repository available at www.jacionline.org).

FIG 4.

Pre- and post-HCT analysis of lymphocyte subsets: absolute CD4⁺naive T cells (A), Vα7.2⁺ T-cell proportion (B), and MAIT cell proportion (C) in peripheral blood; B-cell subset proportions in bone marrow compared to healthy donors (HD) denoting normalized maturation after HCT (D) and at pre- vs post-HCT time points (E) for P6, P7, P8, P11, and P12; absolute class-switched memory B cells (F), CD19^hiCD21^lo B-cell proportion (G), and 9G4⁺ B-cell proportion (H) in peripheral blood. For (A-C) and (E-H), donor types are indicated as matched related donor (dark blue), matched unrelated donor (light blue), haploidentical (purple), and carrier (dashed line) vs unaffected (solid line). Normal control range is designated with gray shading in (B), (C), (G), and (H). Wilcoxon test used for statistical comparison in (D), (E), (G), and (H); *P < .05; ***P < .001.

Furthermore, reduced levels of recombinase activity in patients with pRD affect the process of V(D)J recombination at the T-cell receptor α (TRA) locus in double-positive thymocytes, resulting in a very low proportion of TCR antibody–positive T cells expressing the products of the most 5′ TRAV and the most 3′ TRAJ genes.⁹ Consistent with this, the proportion of T cells stained by the variable alpha 7.2 (Vα7.2) monoclonal antibody (recognizing the product of TRAV1–2, the second most 5′ TRAV gene) were markedly reduced in all 13 patients at baseline. After HCT, a rapid increase in the proportion of Vα7.2⁺ T cells was observed in all patients (Fig 4, B), indicating reconstitution of TCR repertoire diversity. Mucosa-associated invariant T (MAIT) cells are an important subset of immunoregulatory T cells coexpressing an invariant TCR-α chain (Vα7.2-Ja33) and CD161 on the cell surface. As expected, almost all 13 pRD patients had a low proportion at MAIT cells at baseline; however, post-HCT MAIT cell reconstitution lagged behind what was observed for total Vα7.2⁺ T cells (Fig 4, C). Positive selection of MAIT cells in the thymus is promoted by binding of vitamin B₂ ligands to the major histocompatibility complex–like molecule MR1 expressed by MAIT cells.¹³ The microbiota plays a critical role in the production of vitamin B₂ metabolites.¹³ Phylogenetic investigation of communities by reconstruction of unobserved states (PICRUSt) analysis can help identify differences in the metabolic pathways between microbiota of healthy controls and individuals with disease.¹⁴ PICRUSt reanalysis of a recently published series of microbiome data from healthy donors and pRD patients¹⁵ revealed a trend toward low levels of ribA (aka GTP cyclohydrolase II), which catalyzes the first committed step in the biosynthesis of riboflavin, in P4 and P6 after HCT compared to healthy controls, as well as lower levels of flavin mononucleotide, a riboflavin metabolite measured as part of the same report (see Fig E6 in the Online Repository available at www.jacionline.org).

Consistent with previous observations,^16,17 analysis of the bone marrow B-cell compartment at baseline revealed a partial block at the pro-B/pre-B stage, which gradually resolved after HCT (Fig 4, D and E). This was associated with an increase in the proportion of switched memory B cells in the periphery (Fig 4, F; and see Fig E7 in the Online Repository available at www.jacionline.org). All 7 evaluable patients who required immunoglobulin replacement therapy before HCT were able to discontinue this treatment after HCT, consistent with improved endogenous immunoglobulin production (see Fig E8 in the Online Repository). As previously reported,⁹ all patients had an increased proportion of dysreactive CD19^hiCD21^lo and 9G4⁺ B cells at baseline, which decreased after HCT (P = .002 and P = .027, respectively, for patients with at least 1 year’s follow-up) (Fig 4, G and H). Autoantibodies against clinically relevant antigens (thyroid peroxidase and/or thyroglobulin, n = 6; extractable nuclear antigen in the context of myositis, n = 1; GAD65 in the absence of diabetes mellitus, n = 1) were no longer detectable by 1 to 2 years after HCT (n = 5) and/or were not associated with new clinical disease (n = 6) in all evaluable patients. However, anti–type I interferon antibodies, predominantly anti–IFN-α and anti–IFN-ω, which were elevated at baseline in all but one agammaglobulinemic patient, decreased slowly, sometimes remaining elevated even 5 or 6 years after HCT, although neutralizing activity was confirmed to be significantly reduced for each patient at last follow-up (P = .002 and P = .0005, respectively) (Fig 5).

FIG 5.

Trajectory of levels and neutralizing activity of anti–IFN-α (A and B), anti–IFN-ω (C and D), and anti–IFN-β (E) antibodies over time after HCT. For all panels, donor type is indicated as matched related (dark blue), matched unrelated (light blue), haploidentical (purple), and carrier (dashed line) vs unaffected (solid line). Normal control range is designated with gray shading. Wilcoxon test used for statistical comparison in (B) and (D); **P < .01; ***P < .001.

DISCUSSION

Patients with pRD, depending on their disease manifestations and age, represent a clinical dilemma regarding the optimal HCT timing, donor, and approach to HCT.

The largest cohort of patients with pRD published to date included a significantly younger population, with a median age of 3.5 years compared to 20 years in our cohort, with 77.5% 1-year overall survival.³ The majority of those patients received myeloablative or reduced-toxicity conditioning along with serotherapy. In our cohort of 13 patients, over half of whom were adults with active disease manifestations, our approach afforded 100% engraftment with a 92% 1-year survival rate—an overall improvement over historical outcomes. Our experience shows that despite the significant immune dysregulation seen in these patients when undergoing HCT, engraftment is not compromised with the use of a serotherapy-free, reduced-intensity approach in the setting of T-cell–replete grafts, although particular disease manifestations may require a tailoring of the approach, as was done in our patient with immune-mediated neutropenia and another patient with comorbidities limiting the receipt of alkylators. Potentially, the specific use of PTCy contributed to control of underlying immune dysregulation, thus promoting a more hospitable milieu for engraftment despite the risk factor of hyperactivated IFN-γ pathways, as was noted in these patients at baseline.¹⁸ Furthermore, although recipient natural killer–mediated rejection has been proposed as a mechanism for increased graft failure incidence in the RAG1/2 severe combined immunodeficiency setting when myeloablative conditioning is not used, we did not observe any graft failure in our non–severe combined immunodeficiency cohort, and donor chimerism was robust in not only the T-cell and myeloid compartments but also in natural killer and B cells.¹⁹

As in the larger study which found that organ damage was associated with poorer outcomes, we noted that the patients who had developed severe lung disease before HCT had particularly complicated courses and/or experienced ongoing exacerbations over subsequent years. Despite the high comorbidity indexes in the majority of patients, transplant-related mortality was low and was no worse for adults as compared to children.

Consistent with previous studies,⁹ the majority of patients in our study had manifested pre-HCT autoimmunity. Reconstitution of the Treg cell compartment and rapid disappearance of dysreactive B cells after HCT may explain why none of them experienced disease recurrence after HCT. A high incidence of post-HCT new-onset autoimmunity was previously reported in a larger cohort of pRD patients and was possibly attributed to persistent defects in negative selection mediated by autoimmune regulator–positive medullary thymic epithelial cells and/or impaired generation of Treg cells.³ We did not note any definite de novo post-HCT autoimmune complications requiring systemic treatment, which may relate to the universal use of T-cell–replete grafts with the potential for a broad repertoire originating from proliferating donor-derived T cells infused as part of the graft and therefore less reliance on a functional thymic compartment.²⁰ Additionally, the use of PTCy and sirolimus, which both promote an increased regulatory T-cell proportion in the early post-HCT setting, may have attenuated the risk of post-HCT autoimmunity in our patients.^21–23

While the incidence of acute GVHD in our cohort was comparable to published rates, the incidence of acute GVHD among the pRD patients was notably higher than what was seen overall in other patients with inborn errors of immunity enrolled onto the same clinical trials.⁵ Previous studies have shown that reconstitution of microbiome diversity early after HCT allows a rapid increase in the number of MAIT cells, which in turn is associated with reduced incidence of acute GVHD.²⁴ In this study, the number of circulating MAIT cells remained low in most patients after HCT. PICRUSt analysis of the microbiome of 2 patients after HCT showed low levels of ribA and associated metabolites. While these data suggest that persistence of microbiome abnormalities after HCT may play a role in causing delayed reconstitution of the MAIT cell compartment and increased risk of acute GVHD, additional microbiome and metabolomic studies in a larger cohort of pRD patients after HCT are needed to confirm this hypothesis. Remarkably, no pRD patients in long-term follow-up developed chronic GVHD.

Neutralizing anti–type I interferon antibodies has been frequently reported in pRD.⁹ Type I interferons play an important role in virus control, and in the non-HCT setting, neutralizing anti–type I interferon autoantibodies have been implicated in susceptibility to life-threatening viral infections.^25–28 In our study, severe or recurrent herpes simplex virus infections requiring parenteral therapy occurred with greater frequency among pRD patients compared to patients with other diagnoses enrolled onto the same clinical trials, as did CMV requiring preemptive therapy, although later use of letermovir in at-risk patients confounds direct comparison (38% compared to 29% pre-letermovir 100-day cumulative incidence of CMV overall in patients with inborn errors of immunity receiving PTCy at our center).⁵ The incidence of BK virus–associated cystitis was similar to that seen in a recent large study at our center (64% vs 72% in known at-risk patients in the larger cohort),²⁹ but prolonged symptom duration was noted in two of the pRD patients. Of note, reduced-dose PTCy in P8 and P13 may have contributed to improved viral control, but the use of reduced-dose PTCy in patients with inborn errors of immunity remains an area of ongoing study in NCT02579967.³⁰ The slow decline of anti–type I interferon antibodies after HCT may possibly be attributable to slower plasma cell turnover to donor chimerism as opposed to other compartments; clarifying marrow subset chimerism is a future direction of investigation.

Finally, we have recently shown that heterozygous carriers of deleterious RAG variants may manifest subtle signs of immune dysregulation, including increased proportion of dysreactive B cells and elevated levels of anti–type I IFN antibodies (albeit non-neutralizing), in the absence of clinical manifestations.⁹ These observations could raise the question of whether heterozygous carriers of RAG deficiency can be safely used as HCT donors. In this study, no differences were seen in cellular or humoral immune reconstitution, nor in reduction of dysreactive B cells or autoantibodies over time, between recipients of grafts from genetically unaffected versus carrier donors.

In conclusion, here we present an effective, low-toxicity, radiation-free, reduced-intensity conditioning approach to HCT for children and adults with pRD. The availability of safe HCT approaches facilitates recommending sooner HCT in affected patients—before organ damage has occurred. Outside the clinical trial setting, post-HCT GVHD prophylaxis dosing (PTCy) or duration (mycophenolate mofetil/sirolimus) may be more closely tailored to individual patient scenarios as more data become available on reducing post-HCT immunosuppression in order to balance infectious complications while maintaining low GVHD incidence. Longer-term follow-up of more HCT recipients is needed to better characterize the lifelong potential risks of autoimmunity in the setting of underlying thymic dysfunction and/or use of carrier donors, and whether these risks may be modulated by HCT timing or strategy.

Clinical implication:

pRD can be cured by RIC-based HCT with PTCy, with greatest benefit prior to irreversible lung disease development.

Abbreviations used

CMV

Cytomegalovirus

GVHD

Graft-versus-host disease

HCT

Allogenic hematopoietic cell transplantation

IDDA2.1

Immune Deficiency and Dysregulation Activity v2.1

MAIT

Mucosa-associated invariant T

NIH

National Institutes of Health

PICRUSt

Phylogenetic investigation of communities by reconstruction of unobserved states

pRD

Partial RAG deficiency

PTCy

Posttransplantation cyclophosphamide

RAG

Recombinase activating gene

RIC

Reduced-intensity conditioning

TRA

T-cell receptor α

Vα7.2:

Variable alpha 7.2