Long-term outcomes after gene therapy for adenosine deaminase severe combined immune deficiency

Abstract

Patients lacking functional adenosine deaminase activity have severe combined immunodeficiency (ADA SCID), which can be treated with ADA enzyme replacement therapy (ERT), allogeneic hematopoietic stem cell transplantation (HSCT), or autologous HSCT with gene-corrected cells (gene therapy [GT]). A cohort of 10 ADA SCID patients, aged 3 months to 15 years, underwent GT in a phase 2 clinical trial between 2009 and 2012. Autologous bone marrow CD34+ cells were transduced ex vivo with the MND (myeloproliferative sarcoma virus, negative control region deleted, dl587rev primer binding site)-ADA gammaretroviral vector (gRV) and infused following busulfan reduced-intensity conditioning. These patients were monitored in a long-term follow-up protocol over 8 to 11 years. Nine of 10 patients have sufficient immune reconstitution to protect against serious infections and have not needed to resume ERT or proceed to secondary allogeneic HSCT. ERT was restarted 6 months after GT in the oldest patient who had no evidence of benefit from GT. Four of 9 evaluable patients with the highest gene marking and B-cell numbers remain off immunoglobulin replacement therapy and responded to vaccines. There were broad ranges of responses in normalization of ADA enzyme activity and adenine metabolites in blood cells and levels of cellular and humoral immune reconstitution. Outcomes were generally better in younger patients and those receiving higher doses of gene-marked CD34+ cells. No patient experienced a leukoproliferative event after GT, despite persisting prominent clones with vector integrations adjacent to proto-oncogenes. These long-term findings demonstrate enduring efficacy of GT for ADA SCID but also highlight risks of genotoxicity with gRVs. This trial was registered at www.clinicaltrials.gov as #NCT00794508.

Graphical abstract

Figure 1.

VCN in granulocytes and PBMCs. Genomic DNA, isolated from (A-B) granulocyte and (C-D) PBMC fractions isolated from peripheral blood samples obtained at serial times after gene therapy as described in “Materials and methods” was analyzed using ddPCR to quantify average VCN. The test results are displayed chronologically (A,C) or grouped by patient for samples 2 years or later after GT (B,D).

Figure 2.

VISA of PBMCs. Vector integration sites were determined from PBMC samples obtained ∼2 years after gene therapy at the start of LTFU and from the most recently obtained PBMC sample available to assess changes over the time of LTFU. (A) Stacked bars illustrate the 10 most abundant vector integration sites detected at the times indicated for each patient. (B) Word cloud with names of genes most commonly found with vector integrants across all 10 patients. Relative sizes of the gene names indicate their relative abundance. *Within transcription unit (TU); ^∼within 50 kB of oncogene(s); ^!within 50 kB of gene(s) implicated in human lymphomas; N.B., LMO2 gene name appears twice, representing integrants within the LMO2 TU (denoted with an asterisk) and those adjacent to, but not actually within, the LMO2 TU.

Figure 3.

ADA enzyme activity in PBMCs and dAXP in RBCs. (A-B) PBMCs isolated from peripheral blood samples by Ficoll-Hypaque separation were counted, frozen, and assayed for ADA enzyme activity using a radiochemical assay. (C-D) Deoxyadenosine nucleotides (dAXP = dAMP + dADP + dATP) were measured in RBCs, as described. The test results are displayed chronologically (A,C) or group by patient for samples 2 years or later after gene therapy (B,D). The horizontal dashed red line is the upper limit for normal control samples in the testing laboratories (<0.2%); the horizontal dashed blue line is the median value for %dAXP in 22 recipients of allogeneic HSCT for ADA SCID (median = 4.8%; mean = 7.0 ± 11.0% [standard deviation]; M.S.H., unpublished observation).

Figure 4.

ALCs. (A) ALCs determined by clinical complete blood cell counts are shown for each patient over time after GT. Lymphocyte subsets determined by flow cytometry are shown for (B) total CD3⁺ T cells, (C) CD4⁺ T cells, (D) CD4⁺/CD45RA⁺ naïve T cells, (E) CD8⁺ T cells, (F) CD19⁺ B cells, and (G) CD16⁺/CD56⁺ NK cells.

Figure 5.

Serum immunoglobulin levels 2 or more years after gene therapy. Serum IgM (A) and IgA (B) levels were measured in all subjects and IgG (C) in the 4 subjects who discontinued IgRT. Horizontal dashed red line indicates lower limits of normal. Horizontal bars in each patient’s data points represent medians. The numbers in parentheses in panel C represent the time (months) after GT that IgRT was stopped.

Figure 6.

Stability of T-cell repertoires in transplanted patients. (A) TRBV gene use, normalized (heat map indicates z-score) and clustered by row (sample). Each clone is only counted once per sample. TRBV genes that account for <1% of total clones are excluded. Each row is labeled with the sample name and time point and the clone number is given in parentheses. (B) Venn diagrams showing clone numbers that are found at each time point and those that overlap between 2 time points (based on the same samples and time points that are shown in panel A). (C) Ranks of the T-cell clones that overlap between the 2 time points. Numbers to the left indicate the subjects. Columns indicate the time of the sample, with time point 1 (t1) being the earlier time point and t2 the later time point (same samples and time points as in panel A. Numbers in the cells indicate the clone ranks, which are also shown in the heat map. A rank of 1 indicates the clone with the highest sequence copy number fraction in the sample. Only clones with a combined rank of <100 (summed ranks of t1 + t2) are shown.

Figure 7.

Correlations between median granulocyte VCN and biochemical and immunological outcomes. The median granulocyte VCN for each subject (as in Figure 2) were plotted vs the median values calculated for each subject for (A) PBMC ADA enzyme activity, (B) RBC deoxyadenosine nucleotides, (C) ALC, (D) absolute numbers of CD3⁺ T cells, (E) absolute numbers of CD19⁺ B cells, and (F) serum IgM levels. R values shown represent Spearman correlations between median granulocyte VCN and median of each of the outcome parameters.