CD34+ Stem Cell Selection and CD3+ T Cell Add-Back from Matched Unrelated Adult Donors in Children with Primary Immunodeficiencies and Hematological Diseases

Abstract

Less than 25% of children who require hematopoietic stem cell transplantation (HSCT) for primary immunodeficiencies (PIDs) or genetic hematological diseases have an HLA-identical sibling. For them, a matched unrelated donor (MUD), although baring a greater risk of graft failure, delayed engraftment and immune reconstitution, and severe graft-versus-host disease (GvHD), represents a valid alternative. The stem cell source is also important, as unprocessed peripheral blood stem cells (PBSCs) contain 5 to 10 times more T cells than bone marrow (BM)-derived grafts, a major risk especially for small children with PID. A CD34+ positive selection can mitigate HLA compatibility issues, but the resulting CD3+ T cell depletion hampers engraftment and facilitates infections. To mitigate those problems, we decided to add back a certain number of T cells (30×106 cells/kg body weight [BW]) to the positive CD34+ selection derived from MUD BM or PBSCs and report the results in terms of time to engraftment and immune reconstitution, GvHD incidence, infections, and survival. Our aim was to show not only the feasibility and clinical efficacy of this addback but also that PBSC-derived CD34+ selected grafts with calibrated T cell addback would be equivalent to BM-derived grafts. We analyzed retrospectively our single-center cohort of 76 children (median age, 1.9 years) affected by PID (61) and hematological diseases (15) who received a total of 79 MUD HSCTs with CD34+ selection and addback of 30×106 CD3+ cells/kg BW between 2001 and 2019. We used descriptive and analytic statistics (chi-square, Student's t-test, Mann-Whitney U test, as appropriate) and constructed Kaplan-Meier curves using the log-rank test to compare patients grafted with BM or PBSC-derived inocula. The two groups showed no statistically significant differences in terms of age, sex, HLA-mismatch, or amount of CD3+ cells/kg BW added back to the CD34+ selection. However, the latter being higher in the PBSC group (P=.0001). Overall engraftment rate was 96% (73/76) and occurred faster in the PBSC group than in BM recipients: polymorphonuclear cells, 16 versus 21 days (P=.006); platelets, 15 versus 22 days (P=.001). GvHD incidence was low. No acute GvHD was diagnosed in 24 children, whereas grades I, II, III, and IV occurred in 19, 28, five, and three children, respectively (P not significant). Chronic GvHD was seen in only two children. The CD4+ count at six months after HSCT was higher in PBSC recipients as compared to those receiving BM (184 versus 88 CD4+ cells; P=.003). Overall survival for the whole cohort was 80% at 10 years, with no significant difference between the two stem cell sources (P not significant). Viral infections occurred among five of the PBSC grafted children and 14 in the BM group (P not significant), and no patient suffered from post-transplant lymphoproliferative disorder (PTLD). The results we present show that an addback of 30×106 donor CD3+ cells/kg recipient BW to a MUD BM or PBSC-derived CD34+ selection gives promising results in infants and young children undergoing HSCT for PID or hematological diseases. Furthermore, with this manipulation the inherent limits of PBSC-derived grafts can be overcome, allowing both swift engraftment and immune reconstitution without an increase in GvHD, infections, or PTLD.