Successful Preclinical Development of Gene Therapy for Recombinase-Activating Gene-1-Deficient SCID

Abstract

Recombinase-activating gene-1 (RAG1)-deficient severe combined immunodeficiency (SCID) patients lack B and T lymphocytes due to the inability to rearrange immunoglobulin and T cell receptor genes. Gene therapy is an alternative for those RAG1-SCID patients who lack a suitable bone marrow donor. We designed lentiviral vectors with different internal promoters driving codon-optimized RAG1 to ensure optimal expression. We used Rag1 -/- mice as a preclinical model for RAG1-SCID to assess the efficacy of the various vectors. We observed that B and T cell reconstitution directly correlated with RAG1 expression. Mice with low RAG1 expression showed poor immune reconstitution; however, higher expression resulted in phenotypic and functional lymphocyte reconstitution comparable to mice receiving wild-type stem cells. No signs of genotoxicity were found. Additionally, RAG1-SCID patient CD34+ cells transduced with our clinical RAG1 vector and transplanted into NSG mice led to improved human B and T cell development. Considering this efficacy outcome, together with favorable safety data, these results substantiate the need for a clinical trial for RAG1-SCID.

Keywords: B lymphocytes; CD34+ cells; RAG1; SCID; T lymphocytes; gene rearrangement; gene therapy; lentiviral vector.

© 2020 The Authors.

Figures

Graphical abstract

Figure 1

Selecting the Optimal SIN LV Plasmid: Virus Production And In Vitro Efficiency (A) Four different SIN LV plasmids in the CCL backbone carrying different promoters (Cbx3.MND, MND, PGK, and UCOE promoter) were tested to drive expression of a codon-optimized version of RAG1. (B) Production of lentivirus batches with the different constructs. The number of infective particles (infectious genomes/mL) from unconcentrated and concentrated small batches was determined. Three independent lentivirus small batches per plasmid were produced and analyzed (two-way ANOVA test; ∗p < 0.05, ∗∗p < 0.01). (C) Production of lentivirus batches on a large scale with the different constructs. The number of infectious genomes/mL after concentration of large lentiviruses batches was determined. (D) Transduction efficiency of the different SIN lentiviruses in murine lineage-negative cells. VCN was determined by WPRE determination on genomic DNA. Three independent lentivirus batches per plasmid were produced and analyzed (one-way ANOVA test; ∗p < 0.05, ∗∗p < 0.01. (E) Determination of transgene expression in the transduced cells by the different constructs. c.o.RAG1 expression relative to ABL1 was determined by qPCR. Three independent lentivirus batches per plasmid were produced and analyzed (one-way ANOVA test; ∗p < 0.05, ∗∗p < 0.01. (F) Determination of the promoter strength (c.o.RAG1 expression/VCN) of the different plasmids. Three independent lentivirus batches per plasmid were produced and analyzed (one-way ANOVA test; ∗p < 0.05, ∗∗p < 0.01). (G) Total number of B220+ cells (left panel) and total number of B220+IgM+ cells (middle panel) correlated with the expression of c.o.RAG1 in BM. The correlation between VCN and c.o.RAG1 expression in BM of immune reconstituted mice is shown (right panel) (▲, Cbx3.MND; ▼, MND; ▪, PGK; ●, UCOE promoters; gray indicates low-expressing plasmids; black indicates high0expressing plasmids; green circles indicate mice with acceptable immune B and T cell reconstitution). Data shown represent two independent in vivo experiments with in total six or seven mice per group. Each dot represents one mouse. Nonparametric Spearman r correlation, two-tailed; ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. (H) Correlation between total thymocytes (left panel) and DP cells (middle panel) with c.o.RAG1 expression in the thymus. Correlation between VCN and c.o.RAG1 expression in the thymus of immune reconstituted mice (right panel) (▲, Cbx3.MND; ▼, MND; ▪, PGK; ●, UCOE promoters; gray indicates low-expressing plasmids; black indicates high-expressing plasmids; green circles indicate mice with acceptable immune B and T cell reconstitution). Data shown represent two independent in vivo experiments with in total six or seven mice per group. Each dot represents one mouse. Nonparametric Spearman r correlation, two-tailed; ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Figure 2

Selecting the Optimal SIN LV Plasmid to Drive an Immune Reconstitution of Rag1 Deficiency Rag1-deficient mice (experiment with a total of 3 mice/group) were transplanted with 500,000 stem cells: WT cells, mock Rag1 KO cells, Cbx3.MND-c.o.RAG1-treated KO cells (VCN of 0.95), and MND-c.o.RAG1-treated KO cells (VCN of 1.1). (A) Representative FACS plots showing the restoration of B220high+ B cells in the BM. (B) Total number of the different B cell subsets in the BM (left panel) and total number of B cells (B220high+) in the PB (right panel) 16 weeks after SC transplantation. Graphs represent the means and standard deviation of a pilot experiment with two to three mice per group (Mann-Whitney test, one-tailed; ∗p ≤ 0.05; NS, not significant). (C) Representative FACS plots of the thymus reconstitution (CD4 versus CD8) with the different constructs. (D) Total number of the different T cell subsets in the thymus (left panel) and total number of T cells (CD3+TCRαβ+) in PB (right panel) 16 weeks after transplantation. Graphs represent the means and standard deviation of a pilot experiment with two to three mice per group (Mann-Whitney test, one-tailed; ∗p ≤ 0.05; NS, not significant). DN, double negative; ISP, immature single positive; DP, double positive; SP, single positive. (E) Representative samples of GeneScan plots are shown for four different families (x axis indicates CDR3 length; y axis shows the fluorescence intensity of the runoff products). (F) TCR Vβ repertoire analysis by GeneScan. A total of 24 Vβ families were analyzed on spleen cells from three mice per group. Overall score of all of the families was calculated for the different constructs (Mann-Whitney test; p values are represented on the plot; ∗∗∗∗p < 0.0001; NS, not significant). (G) Quantification of total IgG and IgM in mice serum by ELISA (one-way ANOVA test; ∗p < 0.05, ∗∗p < 0.01). (H) IVIM assay was performed on the two constructs to assess their safety (mock cells as negative control; RSF91 γ-retroviral vector as a positive control). Data show results from three complete IVIM assays.

Figure 3

Extensive Immune Reconstitution of Mice Receiving Gene Therapy of Stem Cells with a Clinical-Grade MND-c.o.RAG1 Vector Rag1-deficient mice were transplanted with 250,000 stem cells: WT cells (three mice), mock KO cells (three mice), and MND-c.o.RAG1-treated cells (VCN of 0.2; eight mice). (A) Representative plots of B cell reconstitution in the blood (B220+IgM/IgD cells; top panel) and B cell development in the BM (B220+CD19+ cells; bottom panel) 24 weeks after transplantation. (B) Total number of B cells (B220+CD11b/CD43− cells) in the PB (Mann-Whitney test, one-tailed; ∗p < 0.05, ∗∗p < 0.01). (C) Immature (B220+CD93+ cells; left panel) and mature (B220+CD93- cells; right panel) B cell subsets distribution in spleen. Two-way ANOVA test; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001. (D) Representative plots of T cell reconstitution in the blood (CD3+TCRab+ cells; top panel) and T cell development in the thymus (CD4 versus CD8 cells; bottom panel) 24 weeks after transplantation. (E) Total number of T cells (CD3+TCRab+ cells) in PB at the end of the experiment (24 weeks) (Mann-Whitney test, one-tailed; ∗p < 0.05, ∗∗p < 0.01). (F) Naive, effector memory (EM), and central memory (CM) subset distributions for CD4 (CD3+TCRab+CD4+; left panel) and CD8 (CD3+TCRab+CD8+; right panel). T cell subset distributions in spleen are shown: naive cells (CD44−CD62L+), EM cells (CD44+CD62L−) and CM cells (CD44+CD62L+) 24 weeks after transplantation. (G) Left panel: Hematoxylin and eosin staining of mesenteric lymph nodes (scale bars, 200 μm) and spleen (scale bars, 100 μm; purple indicates germinal centers, and red indicates red pulp). Representative FoxP3 staining in spleen tissue (scale bars, 100 μm) is shown. Arrows indicate positive FoxP3 in germinal centers. Representative images are from WT control, KO control, and MND-c.o.RAG1 gene therapy mice. Right panel: Histological analysis of thymus reconstitution by hematoxylin and eosin staining (scale bars, 50 μm) and cytokeratin 5/6 staining (scale bars, 100 μm) . Representative images from WT control and MND-c.o.RAG1 mice. KO thymus was completely used for phenotyping (FACS, DNA, RNA), but KO thymic histology was previously described by van Til et al.

Figure 4

Functional Ig and TCR Rearrangements and Ig Class-Switch after RAG1 Gene Therapy (A) TCR Vβ repertoire analysis by GeneScan from three WT control mice, one KO control mouse, and eight MND-c.o.RAG1 mice. A total of 24 Vβ families were analyzed on spleen cells from three WT control, one KO control, and eight MND-c.o.RAG1 mice (non-immunized and immunized). Overall score of all of the families was calculated by ImSpectR (Mann-Whitney test; p values are represented in the plot; ∗p 

Figure 5

Preclinical Safety Testing of the Clinical-Grade MND-c.o.RAG1 Vector (A) Vector biodistribution in immune and non-immune organs assessed by qPCR on DNA samples from 16 organs in total. Each dot represents a value from one mouse (three mice/control group, eight MND-c.o.RAG1 mice). The horizontal dashed line represents the threshold of the VCN of immune organs versus non-immune organs (starting VCN of transplanted cells of 0.2). (B) LV insertion site analysis by nrLAM-PCR of isolated DNA from BM obtained from Rag1−/− untransduced control mouse (mock) and four MND-c.o.RAG1 mice (male non-immunized/immunized, female non-immunized/immunized). Gels shows results of the linear amplification from the 3′ long terminal repeat (LTR) and 5′ LTR, respectively (L = 1 kb plus marker). (C) Replating frequencies (RFs) of the control samples mock or RSF91 and the test vector MND-c.o.RAG1, in comparison to data of a meta-analysis for control samples (Mock-MA, RSF91-MA, lv-SF-MA [a lentiviral vector with an SFFV promoter]). The data points below the limit of detection (LOD; plates with no wells above the MTT threshold) were manually inserted into the graph (due to the logarithmic scale of the y axis). Above the graph, the ratios of positive (left number) and negative plates (right number) according to the MTT assay are shown. Differences in the incidence of positive and negative assays relative to Mock-MA or RSF91-MA were analyzed by Fisher’s exact test with a Benjamini-Hochberg correction (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; NS, not significant). If above the LOD, bars indicate the mean RF.

Figure 6

Restored T Cell Development in RAG1 SCID Patient Cells 65,000 human CD34+ cells were transplanted intravenously into busulfan pre-conditioned NSG recipient mice (one NSG mouse with untreated cells and one NSG mouse with MND-c.o.RAG1 gene therapy cells with a VCN of 0.1). (A) FACS plots of human B cells (CD13/33−CD19+CD20+ cells; top panel) and total number of B cells (CD13/33−CD19+CD20+IgD/IgM cells; bottom panel) in the spleen at week 24 after transplantation. (B) FACS plots of human T cells (CD3+TCRαβ+; top panel) and total number of T cells, CD4 cells, and CD8 T cells in the PB at week 24 after transplantation (bottom panel). (C) Human T cell development in the thymus: FACS plots (CD4 versus CD8) and distribution of the different T cells subsets in the thymus (24 weeks after transplantation) are shown. (D) Quantification of total human IgM by ELISA of serum from a control NSG mouse transplanted with RAG1-SCID control untreated CD34+ cells (non-hypomorphic), our SCID patient CD34+ cells, and our SCID MND-c.o.RAG1 CD34+ cells. (E) Human TCR Vβ and Vγ repertoire analysis of isolated DNA from NSG thymus (SCID patient and SCID MND-c.o.RAG1) using a TCRB + TCRG T cell clonality assay (x axis indicates fragment sizes; y axis shows the fluorescence intensity of the runoff products). (F) LV insertion site analysis by nrLAM-PCR of isolated DNA from BM obtained from NSG SCID patient untransduced cells (mock) and NSG SCID MND-c.o.RAG1 mice. Gel shows results of the linear amplification from the 5′ LTR (L = 1 kb plus marker). Data are from an independent experiment with n = 1 per condition.