TCRαβ-Depleted Haploidentical Grafts Are a Safe Alternative to HLA-Matched Unrelated Donor Stem Cell Transplants for Infants with Severe Combined Immunodeficiency

Hematopoietic stem cell transplantation and gene therapy are the only curative therapies for severe combined immunodeficiency (SCID). In patients lacking a matched donor, TCRαβ/CD19-depleted haploidentical family donor transplant (TCRαβ-HaploSCT) is a promising strategy. Conditioned transplant in SCID correlates to better myeloid chimerism and reduced immunoglobulin dependency. We studied transplant outcome in SCID infants according to donor type, specifically TCRαβ-HaploSCT, and conditioning, through retrospective cohort analysis of 52 consecutive infants with SCID transplanted between 2013 and 2020. Median age at transplant was 5.1 months (range, 0.8–16.6). Donors were TCRαβ-HaploSCT (n = 16, 31.4%), matched family donor (MFD, n = 15, 29.4%), matched unrelated donor (MUD, n = 9, 17.6%), and matched unrelated cord blood (CB, n = 11, 21.6%). Forty-one (80%) received fludarabine/treosulfan-based conditioning, 3 (6%) had alemtuzumab only, and 7 (14%) received unconditioned infusions. For conditioned transplants (n = 41), 3-year overall survival was 91% (95% confidence interval, 52–99%) for TCRαβ-HaploSCT, 80% (41–98%) for MFD, 87% (36–98%) for MUD, and 89% (43–98%) for CB (p = 0.89). Cumulative incidence of grade II–IV acute graft-versus-host disease was 11% (2–79%) after TCRαβ-HaploSCT, 0 after MFD, 29% (7–100%) after MUD, and 11% (2–79%) after CB (p = 0.10). 9/10 patients who received alemtuzumab-only or unconditioned transplants survived. Myeloid chimerism was higher following conditioning (median 47%, range 0–100%) versus unconditioned transplant (median 3%, 0–9%) (p < 0.001), as was the proportion of immunoglobulin-free long-term survivors (n = 29/36, 81% vs n = 4/9, 54%) (p < 0.001). TCRαβ-HaploSCT has comparable outcome to MUD and is a promising alternative donor strategy for infants with SCID lacking MFD. This study confirms that conditioned transplant offers better myeloid chimerism and immunoglobulin freedom in long-term survivors. Supplementary Information The online version contains supplementary material available at 10.1007/s10875-022-01239-z.


Introduction
Severe combined immunodeficiency (SCID) is a pediatric emergency, and definitive treatment by hematopoietic stem cell transplantation (HSCT) or gene therapy should be performed promptly. This urgency is well-founded, as age > 3.5 months at HSCT confers higher mortality risk compared to those aged <3.5 months, with active or resolved infection further compounding this risk [1][2][3].
While HLA-matched siblings remain the first donor of choice [1,2,4], < 25% of patients have a matched sibling or family donor (MSD, MFD) available, necessitating consideration of alternative donors such as matched unrelated donors (MUD), unrelated cord blood (CB), or mismatched family donors. MUD searches introduce delay in time-to-transplant, increasing the risk of infection-related morbidity and mortality, while use of a cord graft introduces risk of either slower neutrophil and lymphocyte reconstitution if serotherapy is used [5], or enhanced rates of graft-versus-host disease (GvHD) if serotherapy is omitted [6], and precludes harvesting additional cells for various therapies. An alternative is using a mismatched family donor, with ex vivo T-lymphocyte depletion to reduce alloreactivity; while prompt availability of a parental donor may favor their use, this must be balanced against the risks of acute GvHD and delayed T-lymphocyte reconstitution, particularly in the context of viremia [7]. CD3 + TCRαβ/CD19-lymphocyte depletion is increasingly used in SCID and non-SCID inborn errors of immunity (IEI) HSCT [8][9][10] due to increased overall survival and reduced rates of GvHD compared to previous strategies such as CD34 + selection [11], though data on its use in SCID are sparse.
Previous recommendations for conditioning pre-HSCT for SCID vary by subtype and clinical status; the latest European Societies for Immunodeficiencies and Bone Marrow Transplantation (ESID/EBMT) guidance recommends conditioning, for improved rates of myeloid engraftment and immunoglobulin independence [12]. We explored the outcomes of HSCT after TCRαβ/CD19-depleted haploidentical donor (TCRαβ-HaploSCT) compared to other donor types, and the impact of conditioning.

Methods
A retrospective cohort analysis of 52 consecutive patients undergoing first HSCT for SCID between 2013 and 2020 was performed. One patient was excluded (failed lentiviral gene therapy for adenosine deaminase [ADA] deficient-SCID leading to pancytopenia and monosomy 7, followed by successful HSCT). In the study, conditioned transplant included patients who received cytoreductive chemotherapy and serotherapy while unconditioned transplant referred to patients who did not receive any chemotherapy or had alemtuzumab only. Clinical and laboratory data were retrieved from the transplantation database, patients' medical files, and laboratory records. Written informed consent was obtained from the patients and/or parents or legal guardians of the patients as per institutional practice for HSCT.
The primary endpoints were overall survival (OS) and GvHD-free, event-free survival (GEFS). GEFS was defined as survival without graft failure, second procedure, grade III-IV acute GvHD (aGvHD), and chronic GvHD (cGvHD). Secondary endpoints were aGvHD, cGvHD, toxicities, and viremia. In the statistical analysis, quantitative variables were described with median and range while categorical variables were reported with counts and percentages. The association between continuous variables was assessed with the use of Wilcoxon rank-sum test when comparing two groups and Kruskal-Wallis test when comparing more than two groups. Subgroup differences in OS and GEFS were evaluated by log-rank test. Competing risks methods were used for the cumulative incidence of acute and chronic GVHD, with competing events death. Subgroup differences in cGvHD and aGvHD were evaluated by Gray's test. All estimates are reported with 95% confidence intervals. All p-values quoted are two-sided, with a level of significance of 0.05. Statistical analyses were performed using STATA 14.2 and were generated with GraphPad Prism.

Results
Of 51 patients, 48 had genetically or biochemically confirmed etiologies for SCID, 2 had T low B + NK + SCID, and one presented with an Omenn-like syndrome (Supplemental Table S1). Fourteen (27.4%) patients had a newborn diagnosis of SCID due to previous family history. Median age at diagnosis was 2.1 months (range: 0-21.8). The median age at transplant was 6.1 months (1.0-16.6) for non-ADA SCID and 4.0 (0.8-43.3) for ADA SCID (p = 0.42). Interval to transplant was significantly shorter in TCRαβ-HaploSCT (median 2.3 months, range 0.6-9.8 months) compared to adult MUD recipients (median: 3.9 months, range 2.0-14.6 months) (p = 0.039); this analysis excluded patients with ADA-SCID, who were treated with pegylated-ADA while awaiting MUD search, if MFD were unavailable.
There were no statistically significant differences between TCRαβ-HaploSCT and T-replete groups in pre-transplant comorbidities, infection history, or active viremia at HSCT (Supplemental Table S1).
Overall, 6 patients (11.8%) died, five following conditioned transplants. One patient with DNA ligase IV-SCID died following TCRαβ-HaploSCT from disseminated adenovirus; one patient with JAK3-SCID died following an MFD transplant due to disseminated CMV with TMA; three patients with Artemis SCID died after one MFD transplant (respiratory failure from RSV pneumonitis) and two MUD transplants (multi-organ failure with influenza A  One patient with Artemis-SCID who received an unconditioned TCRαβ/CD19-deplete haploidentical transplant Fig. 1 Overall survival (OS, a) and grade II-IV acute or chronic GvHD-and eventfree survival (GEFS, b) at 3 years post-HSCT, cumulative incidence of grade II-IV acute GvHD (c), and cumulative incidence of any viremia in the first year post-HSCT (d) for conditioned transplants according to donor type. Numerical values represent probability % (95% confidence interval). Probabilities calculated using log-rank test for a and b, and Gray's test for c and d. Events were defined as death or second procedure  Table S2 in the Online Repository. *Significant at p < 0.05 while critically unwell with respiratory failure to PCP died 8 days post-HSCT.

Discussion
The interval from diagnosis to transplant was shorter for TCRαβ-HaploSCT vs T-replete MUD when patients with ADA-SCID were excluded (Table 1), confirming they are more rapidly available. TCRαβ/CD19-deplete PBSC grafts were significantly enriched with nucleated and CD34 + stem cells, and consequently myeloid engraftment occurred comparably with T-replete MSD. Reconstitution of T-lymphocytes in conditioned patients followed previous experience of TCRαβ-HaploSCT in other IEI [8,9], with thymic output of naïve CD4 + and CD8 + lymphocytes at 3 months post-HSCT in 40.5% and 45.9% of patients, respectively. The comparatively brisk reconstitution of B-lymphocytes in recipients of CB units may reflect reduced serotherapy doses, or relatively higher proportions of B-lymphocyte precursors [5]; these cells, which do not express CD19 [13], would not account for the comparatively lower CD19 + cell dose in CB compared to MFD/MUD grafts ( Table 1). Both T-depleted and T-replete conditioned grafts led to similar proportions of patients with a month-6 post-HSCT CD4 + count > 500 cells/μL, with counts < 500 cells/μL previously demonstrated to be a risk factor for mortality in SCID transplant recipients [1]. Long-term, myeloid chimerism and rates of immunoglobulin use are similar between conditioned donor types. In line with the recently updated EBMT/ESID guidance [12], our data support the role of conditioning in generating higher myeloid chimerism and freedom from Fig. 3 Kinetics of CD3 + lymphocyte reconstitution for unconditioned and serotherapyonly transplants over the first year post-HSCT by donor source Fig. 4 Donor myeloid chimerism (a) and requirement for immunoglobulin (Ig) replacement at latest-follow-up (b) by conditioned or unconditioned transplant, and donor myeloid chimerism by immuno-globulin dependence (c) for all patients at latest follow-up. Patients receiving alemtuzumab-only have been classified as "unconditioned." *Significant at p < 0.05. **Significant at p < 0.01 immunoglobulin replacement, and at our center conditioning is the default in the absence of life-threatening infection due to enhanced long-term outcomes.
Across patients undergoing HSCT for IEI, previous data from our center demonstrate improved overall and event-free survival with the advent of TCRαβ-HaploSCT compared to other methods of ex vivo T-lymphocyte depletion such as CD34 + selection, but with increased rates of viremia [11], which remain a concern due to associated morbidity and mortality [3,9,11]. While options for enhancing viral immunity in these patients include CD45RA + -depleted lymphocyte infusions [14], our study of SCID patients demonstrated similar cumulative incidence of any viremia post-HSCT between TCRαβ-HaploSCT and other donor types. Rates of CMV viremia were lower in our cohort compared to comparable studies looking at other inborn errors of immunity [7][8][9]. This may relate to our cohort being comprised solely of infants with SCID, where infection prevention measures including antiviral prophylaxis, immunoglobulin replacement, and consideration of breastfeeding cessation are universal compared to other diagnoses. Our cohort did not share the increased risk of acute GvHD demonstrated in other series [7,8] despite comparable doses of TCRαβ + T-lymphocytes, with the cumulative incidence of clinically significant acute GvHD being comparable to other unrelated donor types. Our survival data demonstrate favorable outcomes for TCRαβ-HaploSCT recipients compared to other donor types, while the over-representation of Artemisdeficient SCID in our mortality data echoes the published experience of HSCT for this challenging genotype [1,15].

Conclusion
The role of TCRαβ/CD19 depletion for haploidentical donor transplant in IEI is evolving, with several series demonstrating similar survival and morbidity compared to unrelated donors [8][9][10], though patients with SCID have been underrepresented. Our data demonstrate that conditioned TCRαβ/ CD19-depleted haploidentical grafts are a safe and effective alternative for infants with SCID who lack HLA-matched family donors and support the use of conditioning pre-HSCT to improve long-term immunological recovery. We have thus amended our donor hierarchy to: 1) Matched family donor; 2) Parental TCRαβ-HaploSCT or matched unrelated donor,thereby removing a barrier to definitive treatment of SCID.