Hematopoietic stem cell gene editing rescues B-cell development in X-linked agammaglobulinemia

Abstract

Background: X-linked agammaglobulinemia (XLA) is an inborn error of immunity that renders boys susceptible to life-threatening infections due to loss of mature B cells and circulating immunoglobulins. It is caused by defects in the gene encoding the Bruton tyrosine kinase (BTK) that mediates the maturation of B cells in the bone marrow and their activation in the periphery. This paper reports on a gene editing protocol to achieve "knock-in" of a therapeutic BTK cassette in hematopoietic stem and progenitor cells (HSPCs) as a treatment for XLA.

Methods: To rescue BTK expression, this study employed a clustered regularly interspaced short palindromic repeats/CRISPR-associated protein 9 system that creates a DNA double-strand break in an early exon of the BTK locus and an adeno-associated virus 6 virus that carries the donor template for homology-directed repair. The investigators evaluated the efficacy of the gene editing approach in HSPCs from patients with XLA that were cultured in vitro under B-cell differentiation conditions or that were transplanted in immunodeficient mice to study B-cell output in vivo.

Results: A (feeder-free) B-cell differentiation protocol was successfully applied to blood-mobilized HSPCs to reproduce in vitro the defects in B-cell maturation observed in patients with XLA. Using this system, the investigators could show the rescue of B-cell maturation by gene editing. Transplantation of edited XLA HSPCs into immunodeficient mice led to restoration of the human B-cell lineage compartment in the bone marrow and immunoglobulin production in the periphery.

Conclusions: Gene editing efficiencies above 30% could be consistently achieved in human HSPCs. Given the potential selective advantage of corrected cells, as suggested by skewed X-linked inactivation in carrier females and by competitive repopulating experiments in mouse models, this work demonstrates the potential of this strategy as a future definitive therapy for XLA.

Keywords: Agammaglobulinemia; B cell; gene editing; hematopoietic stem and progenitor cells.

Graphical abstract

Fig 1

Optimization of a gene editing platform for XLA. A, Schematic of the gene editing strategy to knock in a BTK coding sequence into exon 2 of the native locus using the CRISPR/Cas9 system and an AAV6 to deliver the HDR donor template. The HDR donor template consists of a codon optimized version of the BTK coding sequence (co-BTK) with a canonical GCCACCATGG Kozak sequence (K), a 300 bp region of intron 18 (I18) and a βgrowth hormone polyadenylation signal (poly A), flanked by the homology arms left (HAL) and right (HAR). B, Western blot analysis of BTK expression in DG75 and Jurkat T-cell clones. DG75 and Jurkat T cells were electroporated with the Cas9/gRNA RNP and transduced with the AAV6 vector containing the co-BTK cassette before clones were derived. The GAPDH housekeeping protein is used as a loading control.

Fig 2

Gene editing efficiency in HSPCs and optimization of a B-cell differentiation protocol. A, Schematic of the donor construct (AAV6-GFP) used to knock in a PGK-GFP cassette in the BTK locus creating BTK KO HSPCs. B, Cutting efficiency of the Cas9/gRNA RNP complex targeting BTK exon 2 in multiple HSPC donors (n = 10) as determined by TIDE analysis. C, Frequency of HDR-mediated targeted insertion of PGK-GFP in the BTK locus. HSPCs (n = 3) were electroporated with Cas9/gRNA RNP, transduced with the AAV6-GFP donor vector, and analyzed by flow cytometry for high GFP expression as a measure of targeted integration. Representative flow cytometry plot showing GFP expression out of live cells (left) with recapitulating plot (right). D, Overview of the HSPC editing and B-cell differentiation protocol adapted from Kraus et al.E, Frequency of immature B cells (defined as CD19⁺/CD10⁺/IgM⁺) emerging from WT HSPCs at different days of the B-cell differentiation protocol when using increasing concentrations of ICAM1-Fc coating agent. F, Development of CD10⁺/CD19⁺/IgM⁺ immature B cells from GFP⁺ BTK KO HSPCs compared with WT HSPCs, which had been placed in B-cell differentiation culture (n = 3; mean ±SD). CLP, Common lymphoid progenitor; SCF, stem cell factor; SR1, StemRegenin 1; UM729, Methyl 4-((3-(piperidin-1-yl) propyl) amino)-9H-pyrimido[4,5-b] indole-7-carboxylate.

Fig 3

Gene editing of XLA HSPCs rescues the B-cell developmental block. A, Rescue of the XLA phenotype after gene editing: experimental overview. XLA HSPCs were electroporated with the Cas9/gRNA RNP (Cas9) and transduced with the AAV6 co-BTK donor (BTK cDNA) vector before undergoing B-cell differentiation. B, Cell viability over time of WT HSPCs, nonedited XLA HSPCs and XLA HSPCs edited with high-dose (XLA EDIT MOI: 25000) and low-dose (XLA EDIT MOI: 2000) AAV6 as assessed by trypan blue analysis. C, Frequency of myeloid (white) versus erythroid (orange) colonies derived from methylcellulose cultures of the samples as in B. D, Frequency of HDR-mediated targeted insertion of the BTK cassette in XLA HSPCs (n = 3) edited using an AAV6 (MOI: 2000), as assessed by ddPCR. E, Western blot analysis of BTK protein expression in WT, XLA nonedited and edited samples at day 25 of B-cell differentiation (upper). WT samples are used as control for protein expression. GAPDH is used as loading control. Densitometric analysis is shown (lower), as a proportion of WT (100%) (data are shown as mean ± SD; n = 3). F, Proportion of immature B cells (CD10⁺/CD19⁺/IgM⁺) emerging over time from the 3 edited XLA HSPCs undergoing B-cell differentiation along with nonedited XLA and WT. G, Representative flow cytometry plot of day 39 (top) with a summarizing histogram (bottom) (n = 3, mean ± SD, 1-way ANOVA followed by Tukey’s multiple comparisons; significance is indicated as ∗P < .05). CFU, Colony-forming unit.

Fig 4

Gene editing restores B cells ability to activate T cells. A, Edited XLA HSPCs along with nonedited XLA and WT control groups were placed in B-cell differentiation culture. After 30 days, 50,000 cells from each condition were cocultured with primary CD4⁺ T cells, +/− SEB. After 7 days, activated follicular T cells were identified as the CXCR5⁺/programmed cell death protein 1 (PD-1) + population of CD4⁺ gated cells. Representative flow plots identifying activated T cells (B) and aggregate data (C). At the same time point, IgM (D) and IgG (E) concentrations in the supernatants were measured by ELISA (n = 3 different donors; shown is mean ± SD; 1-way ANOVA with Tukey’s multiple comparisons; ∗P < .05, ∗∗P < .01, ∗∗∗∗P < .0001).

Fig 5

In vivo evaluation of edited XLA cells in xenotransplantation experiments. A, Edited XLA HSPCs (n = 10 mice for XLA donors 1 and 2; n = 9 mice for XLA donor 3) were transplanted into NSG mice in parallel with nonedited XLA HSPCs (n = 10 mice for XLA donors 1 and 2; n = 6 mice for XLA donor 3) and WT HSPCs (n = 10 and n = 6). Engraftment is shown as percentage of hCD45⁺ cells in the bone marrow (BM). B, Lineage composition of the human graft (CD33⁺ myeloid cells, CD19⁺ B cells and CD3⁺ T cells) in the bone marrow of mice transplanted with WT (n = 16), nonedited (n = 16), and edited (n = 19) XLA HSPCs. The total percentage of B-cell, T-cell, and myeloid lineages is set at 100%. C and D, Analysis of B cells subsets in the BM of transplanted mice. Representative flow cytometry plot (C) and aggregate frequency (D) of CD19⁺/CD34⁻ (cells at the pre–B-cell stage and onward) and CD19⁺/CD34⁺ (pro–B cells) in mice transplanted with WT HSPCs (n = 16) nonedited (n = 16) and edited XLA HSPCs (n = 19); each donor is depicted by a different color. E, Percentage of edited cells out of the total number of human cells (as assessed by ddPCR) in the BM of mice transplanted with XLA HSPCs edited using AAV6 co-BTK with MOI of 2000 for XLA1 and XLA2 and both MOIs of 2000 and 250 for XLA3. The HDR values in the HSPCs pretransplant are shown as a dotted line.F, Average KREC copies per cell determined by ddPCR performed on BM DNA from mice that had been transplanted with WT HSPCs (n = 12), nonedited (n = 15) and edited (n = 16) XLA HSPCs. IgM (G) and IgG (H) concentrations (assessed by ELISA) in plasma samples of mice as in F. Data in A, B, D, F-H (mean ± SD); 1-way ANOVA followed by Tukey’s multiple comparisons: ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, ∗∗∗∗P < .0001.

Fig 6

Genotoxicity analysis in edited XLA HSPCs. A, Targeted high-throughput sequencing of off-target sites predicted by COSMID in Cas9/gRNA RNP edited (RNP; n = 2) or untreated (UT; n = 2) XLA HSPCs. (∗∗∗P < .001 significant difference between Cas9/gRNA RNP and UT samples when analyzed with 1-tailed, 2-sample Z test of proportion) B, CAST-seq analysis of gross chromosomal aberrations in XLA HSPCs (n = 2). The Circos plot (left) shows a cluster of chromosomal rearrangements at the on-target site (green), at sites of homology to the BTK locus (a mix of yellow and blue), to chromosome X and 18 (blue), and at naturally occurring break sites (NBS; gray). Frequency of events found in the DNA from the 2 patients (right). HMT, Homology mediated translocations; OMT, off target mediated translocations.

Fig E1

Optimization of the AAV6–co-BTK donor cassette in DG75 BTK KO cells. A, A representation of the BTK locus is shown, including the sequence of the therapeutic gRNA targeting exon 2 and the gRNA used to knockout BTK by mutating exon 3. Arrows show the approximate cut site of both gRNAs (left). DG75 cells were edited with a Cas9/gRNA RNP targeting exon 3 of the BTK locus before deriving a clonal KO population (DG75 KO). Flow cytometry analysis of BTK expression in WT and KO DG75 cells (right). B, DG75 BTK KO cells were edited with a Cas9/gRNA RNP complex targeting exon 2 and transduced with an AAV6 HDR donor template containing a codon optimized BTK coding sequence, before deriving clonal populations. Western blot analysis of BTK levels in WT, KO, and a BTK “knocked in” clone. C, Real-time quantitative PCR evaluation of BTK mRNA expression in clones knocked in with a BTK codon-optimized coding sequence (CO; black dots) and a BTK codon-optimized coding sequence containing a canonical Kozak sequence alone (K; each clone shown with different color) or with a 300-bp region from I18 (I18; each clone with different color). The 2^−ΔΔCT values [Delta Ct single clones (Ct coBTK-Ct GAPDH)-deltaCt WT (Ct wtBTK-Ct GAPDH)] are shown. Of note, primers to amplify the WT and codon optimized BTK differ, so conclusions on the relative amount of BTK versus WT cannot be drawn. (Shown is the mean ± SD; ∗P < .005 1-way ANOVA with Tukey’s multiple comparison.) NS, Not significant; SSC, side scatter.

Fig E2

Confirmation of XLA phenotype in XLA HSPCs donors. A, Flow cytometry performed on peripheral blood from 3 XLA HSPCs donors showing staining for CD19 and IgM. B, Table showing the disease-causing mutations of each XLA donor with confirmatory Sanger sequencing traces comparing patient and WT reads.

Fig E3

Similar proportion of CD10⁺ cells among XLA, XLA edit, and WT groups. Proportion of CD10⁺ cells emerging over time from WT, nonedited XLA, and edited XLA HSPCs undergoing B-cell differentiation (n = 3 different donors, mean ± SD, no statistical difference was found with repeated measures 2-way ANOVA followed by Tukey’s multiple comparisons).

Fig E4

Gene editing restores a normal HSPC compartment in the bone marrow and normal levels of B cells in the blood of transplanted mice. A, Flow cytometry analysis of HSCs (CD34⁺/CD38^low/CD45RA⁻/CD90⁺), multipotent progenitors (MPPs: CD34⁺/CD38^low/CD45RA⁻/CD90⁻), and multi-lymphoid progenitors (MLPs: CD34⁺/CD38^low/CD45RA⁺/CD90⁻) out of hCD45⁺ cells in the bone marrow of mice transplanted with WT (n = 16), nonedited (n = 16) and edited (n = 19) XLA HSPCs. B, Percentage of CD19⁺ B cells in the peripheral blood (PB) of mice transplanted with nonedited (n = 8) and edited (n = 8) XLA HSPCs. Mice transplanted with WT HSPCs (n = 8) were used as a control. (Shown is the mean ± SD; 1-way ANOVA with Tukey’s multiple comparisons; ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, ∗∗∗∗P < .0001).

Fig E5

Predicted off-target sites. List of 23 target sites: 1 on-target (On-T) and the top 22 off-target (OT) sites predicted by the COSMID online tool when inputting the sequence of our chosen gRNA and allowing for up to 3 mismatches.