β-globin gene transfer to human bone marrow for sickle cell disease

Abstract

Autologous hematopoietic stem cell gene therapy is an approach to treating sickle cell disease (SCD) patients that may result in lower morbidity than allogeneic transplantation. We examined the potential of a lentiviral vector (LV) (CCL-βAS3-FB) encoding a human hemoglobin (HBB) gene engineered to impede sickle hemoglobin polymerization (HBBAS3) to transduce human BM CD34+ cells from SCD donors and prevent sickling of red blood cells produced by in vitro differentiation. The CCL-βAS3-FB LV transduced BM CD34+ cells from either healthy or SCD donors at similar levels, based on quantitative PCR and colony-forming unit progenitor analysis. Consistent expression of HBBAS3 mRNA and HbAS3 protein compromised a fourth of the total β-globin-like transcripts and hemoglobin (Hb) tetramers. Upon deoxygenation, a lower percentage of HBBAS3-transduced red blood cells exhibited sickling compared with mock-transduced cells from sickle donors. Transduced BM CD34+ cells were transplanted into immunodeficient mice, and the human cells recovered after 2-3 months were cultured for erythroid differentiation, which showed levels of HBBAS3 mRNA similar to those seen in the CD34+ cells that were directly differentiated in vitro. These results demonstrate that the CCL-βAS3-FB LV is capable of efficient transfer and consistent expression of an effective anti-sickling β-globin gene in human SCD BM CD34+ progenitor cells, improving physiologic parameters of the resulting red blood cells.

Figure 1. The CCL-βAS3-FB LV provirus carrying the HBBAS3 cassette.

(A) The CCL-βAS3-FB LV provirus has the HBBAS3 expression cassette with the human β-globin gene exons (arrowheads) with the 3 substitutions to encode the HbAS3 protein, introns, the 3′ and 5′ flanking regions, and the β-globin mini-locus control region (LCR) with hypersensitive sites 2–4. The 3′ LTR contains the SIN deletion and FB insulator, both transferred during RT to the 5′ LTR of the proviral DNA. (B) To test FB insulator stability, PCR reactions were performed using DNA from cells collected at day 14 of in vitro culture of BM CD34⁺ cells: mock transduced (lane 1), transduced with the CCL-βAS3 LV (lane 2), and transduced with the CCL-βAS3-FB LV (lane 3). Primers amplified either the 5′ LTR (A to B) or the 3′ LTR (C to D) or the FB insertion sites in both LTRs (A to D) of the provirus. The expected sizes of the PCR products with these primer pairs are indicated for the CCL-βAS3 LV and the CCL-βAS3-FB LV. NTC, no template control. (C) CTCF-binding protein ChIP. Chromatin was isolated from K562 cells transduced with the CCL-βAS3-FB LV (FB), the CCL-βAS3-1.2 kb cHS4 LV (cHS4), or the CCL-βAS3 vector lacking the insulator (U3). qPCR amplification was done using primers to the HIV SIN LTR (U3, cHS4, and FB) and to the HIV RRE region of the vector backbone (RRE) as negative control or the cellular c-Myc and H19/ICR sites, known to bind CTCF. *P = 0.006. Values shown are mean ± SD.

Figure 2. Assessment of transduction and hematopoietic potential of BM CD34⁺ cells in CFU assay and under in vitro erythroid differentiation culture.

(A) The percentage of plated BM CD34⁺ cells that grew into hematopoietic colonies by in vitro CFU assay is shown. Values presented are the mean ± SD for HD-mock, n = 13; HD-βAS3-FB, n = 16; SCD-mock, n = 18; and SCD-βAS3-FB, n = 24. (B) Distribution of hematopoietic colony types formed by BM CD34⁺ cells. The percentages of the different types of hematopoietic colonies identified are represented, following the same patterns as in A. HD-mock, n = 5 independent experiments; HD-βAS3-FB, n = 7 independent experiments; SCD-mock, n = 6 independent experiments; and SCD-βAS3-FB, n = 8 independent experiments. Values shown are mean ± SD. *P = 0.048, by 2-way ANOVA. (C) In vitro single CFU grown from transduced SCD CD34⁺ BM were analyzed for the presence of CCL-βAS3-FB vector provirus and VC/cell by qPCR (n = 191 colonies, 5 independent experiments). Graph indicates percentages of the CFU that were negative for vector by qPCR (white, n = 134) or that had VC/cell of 1–2 (light gray, n = 50), 3–6 (dark gray, n = 6), and 7–9 (black, n = 1). (D) VC/cell for CCL-βAS3-FB-transduced BM CD34⁺ cells grown under in vitro erythroid differentiation culture. Each point represents an independent transduction and culture. BM CD34⁺ cells were from HD (black circles, n = 11) or SCD donors (white squares, n = 15). Error bars represent mean values ± SD.

Figure 3. In vitro erythroid differentiation of BM CD34⁺ cells.

(A) Fold expansion from BM CD34⁺ cells grown under in vitro erythroid differentiation conditions over time. The growth curves from a representative experiment are shown. HD-mock, black triangles; HD-βAS3-FB transduced, black circles; SCD-mock, white triangles; SCD-βAS3-FB transduced, white squares. (B) Immunophenotypic analysis of CD34⁺ BM SCD–transduced samples during in vitro erythroid culture. Cells were analyzed by flow cytometry for expression of CD34, CD45, CD71, and GpA. Each bar represents the percentage of expression of the indicated surface marker at day 3 (white bars), day 14 (pink bars), and day 21 (red bars). Values shown are mean ± SD of 4 independent experiments. Percentage of enucleated rbc was assessed at day 21 (mean ± SD of 7 independent experiments) by staining with the DNA dye DRAQ5. (C) Flow cytometry analysis of erythroid culture to quantify enucleated rbc. Analysis was made by staining cells with DRAQ5 and antibody to human erythroid marker GpA. Enucleated erythrocytes are present in the left upper quadrant as DRAQ5-negative, GpA-positive cells. (D) Photomicrographs of cytocentrifuge preparations from cultures stained by May-Grunwald-Giemsa showing the progression of erythroid differentiation from erythroblast to normoblast at day 8 and 14 to a mostly uniform population of enucleated reticulocytes and erythrocytes at day 21.

Figure 4. HBBAS3 expression after in vitro erythroid differentiation from CD34⁺ BM samples.

(A) HBBAS3 mRNA expression measured by qRT-PCR from cells transduced to different VC/cell. The percentage of HBBAS3 mRNA achieved from each sample was related to its corresponding VC/cell measured by qPCR. A total of 20 independent transductions are shown. HD, black circles (n = 4); SCD, white squares (n = 16). (B) Representative IEF membrane used to quantify the Hb tetramers present. The left-most lane shows the pI standards of human Hb tetramers from the top down: HbA2, HbS, HbF, and HbA (and the predicted pI for HbAS3). Lanes 1–6 show the IEF of lysates from erythroid cultures initiated with SCD BM CD34⁺ cells, either mock transduced (lane 1) or transduced with the CCL-βAS3-FB LV (lanes 2–6). No HbAS3 protein was detected in the mock-transduced samples (lane 1), while HbAS3 represented of the total Hb the following: 21.78% (lane 2, 1.14 VC), 18.11% (lane 3, 1.08 VC), 19.34% (lane 4, 1.13 VC), 21.34% (lane 5, 0.99 VC), and 20.40% (lane 6, 1.11 VC). Densitometric analyses were used to determine the percentage of HbAS3 of total Hb tetramers, and qPCR was used to measure the VC/cell in the same samples. (C) HbAS3 protein produced from cells transduced to different VC/cell (n = 10). (D) Summary of HBBAS3 expression per VC/cell based on measurement of HBBAS3 mRNA (n = 16) and HbAS3 tetramers (protein, n = 10). Error bars represent mean values ± SD.

Figure 5. SCD phenotypic correction.

(A) Phase contrast photomicrographs of deoxygenated erythroid cells. Cells from erythroid differentiation cultures of BM CD34⁺ cells were treated with sodium metabisulfite, and their morphology was assessed using phase contract microscopy. Five examples of srbc are displayed across the top panels, and 5 examples of nrbc are displayed across the bottom panels. (B) Representative field of rbc from mock-transduced SCD CD34⁺ cells (left panel) vs. CCL-βAS3-FB transduced SCD CD34⁺ cells (right panel) upon deoxygenation with sodium metabisulfite. (C) Correlation of the percentage of morphologically “corrected” cells to the VC/cell in each individual culture of CCL-βAS3-FB–transduced SCD BM CD34⁺ cells. The percentage of corrected rbc is defined as the percentage of nonsickled cells in a transduced sample minus the background value of nonsickled cells in the concordant nontransduced sample.

Figure 6. In vivo assessment of CCL-βAS3-FB LV transduction of BM CD34⁺ cells.

(A) Engraftment of human cells in NSG mice. BM cells isolated from mice from each transplant group (nos. 1–6) were analyzed by flow cytometry to measure the percentage of human CD45⁺ cells among all CD45⁺ cells in the marrow (human and murine) as a measurement of engraftment. Mock transduced, white triangles; CCL-βAS3-FB transduced, black triangles. BM samples from HD were used in transplants 3, 4, and 6 and from SCD donors in transplants 1, 2, and 5. (B) Immunophenotypic analysis of human cells isolated from NSG mice transplanted with transduced BM CD34⁺ cells. Flow cytometry was used to enumerate the percentage of the human CD45⁺ cells that were positive for the markers of B-lymphoid cells (CD19, white), myeloid progenitors (CD33, light gray), hematopoietic progenitors (CD34, dark gray), and erythroid cells (CD71, black). Mean ± SD are shown of 3 independent experiments. Mock, n = 4; βAS3-FB, n = 8 mice. (C) VC/cell in human cells cultured from NSG mice transplanted with transduced BM CD34⁺ cells. Black circles represent samples from mice transplanted with HD BM, and white squares represent mice transplanted with SCD BM. All the human cells examined from mock-transduced mice were negative for VC analysis by qPCR. (D) HBBAS3 mRNA expression measured by qRT-PCR from cells transduced to different VC/cell. Five independent transductions are shown. HD, black circles (n = 6); SCD, white squares (n = 4).

Figure 7. Assessment of genotoxicity of the CCL-βAS3-FB LV vector.

(A) Frequency of vector IS in and near cancer-associated genes. The bars represent the frequencies of integrations in transcribed regions or within 50 kb of promoters of cancer-associated genes (in vitro, 32.1%; in vivo, 34.3%), as defined in Higgins et al. (44). (B) Integration frequency around TSS. The frequencies of vector IS in the four 5-kb bins in a 20-kb window centered at gene TSS are plotted. The IS are shown for the following: BM CD34⁺ cells analyzed after 2 weeks growth in vitro (lenti in vitro, n = 2091; gray bars) and 2–3 months in vivo engraftment in NSG mice (lenti in vivo, n = 414; black bars) along with an MLV γ-retroviral vector data set from a clinical gene therapy trial (MLV in vitro, n = 828; white bars) (45) and a random data set generated in silico and analyzed by identical methods (random, n = 12,837; black line). (C) IVIM assay. The replating frequencies for murine lineage-negative cells transduced with the different vectors are shown, calculated based on Poisson statistics using L-Calc software corrected for the bulk VC/cell measured by qPCR on day 8 pTD. The fractions presented across the lower portion of the figure represent the number of negative assays in which no clones were formed divided by the total number of assays performed for that vector. The horizontal bar represents the mean replating frequency of all positive assays. *P = 0.002, by 2-sided Fisher’s exact test.