Site-specific gene correction of a point mutation in human iPS cells derived from an adult patient with sickle cell disease

Abstract

Human induced pluripotent stem cells (iPSCs) bearing monogenic mutations have great potential for modeling disease phenotypes, screening candidate drugs, and cell replacement therapy provided the underlying disease-causing mutation can be corrected. Here, we report a homologous recombination-based approach to precisely correct the sickle cell disease (SCD) mutation in patient-derived iPSCs with 2 mutated β-globin alleles (β(s)/β(s)). Using a gene-targeting plasmid containing a loxP-flanked drug-resistant gene cassette to assist selection of rare targeted clones and zinc finger nucleases engineered to specifically stimulate homologous recombination at the β(s) locus, we achieved precise conversion of 1 mutated β(s) to the wild-type β(A) in SCD iPSCs. However, the resulting co-integration of the selection gene cassette into the first intron suppressed the corrected allele transcription. After Cre recombinase-mediated excision of this loxP-flanked selection gene cassette, we obtained "secondary" gene-corrected β(s)/β(A) heterozygous iPSCs that express at 25% to 40% level of the wild-type transcript when differentiated into erythrocytes. These data demonstrate that single nucleotide substitution in the human genome is feasible using human iPSCs. This study also provides a new strategy for gene therapy of monogenic diseases using patient-specific iPSCs, even if the underlying disease-causing mutation is not expressed in iPSCs.

Figure 1

Activities and specificities of HBB-ZFNs that stimulate gene targeting in a GFP reporter assay. (A) The putatitive recognition sequence of HBB-ZFNs in the HBB gene (5′ to 3′, starting from the first codon). The left (12-bp) and left (19-bp) ZFN sites are underlined. The homologous sequences from other β-locus genes (HBE, HBD, and HBG1/2) and their differences from the HBB gene in L-ZFN, R-ZFN, and spacer regions (no. of mismatches) also are shown. Each was inserted into the GFP reporter as ZFN target sequence to test the specificity of HBB-ZFNs. All the inserts start with a STOP codon (red, taa) and end with a HindIII site (blue, aagctt). A short version of the HBB target sequence (called HBB-Short) also was tested. (B) Schematic of the GFP* reporter rescue assay. An EGIP* mutant was created by inserting a ZFN target sequence including a STOP codon and HindIII site into the GFP gene. Only after gene targeting with a tGFP donor (with or without ZFNs), the EGIP* will be corrected to restore wild-type GFP expression. (C) Flow cytometric analysis of GFP correction after HR in 293T cells stably transfected with EGIP*-HBB reporter. Two days after transient transfection of tGFP donor alone or with HBB-ZFNs, numbers of GFP+ cells were measured by dot plot of 1 million collected cell events. (C) Gene correction efficiency of EGIP* mutants with HBB, HBE, HBD, HBG, or HBB-Short ZFN target sequence using the tGFP donor, with or without HBB-ZFNs. Numbers of GFP+ cells per 10⁶ 293T-EGIP* cells are plotted as mean ± SEM, n = 3.

Figure 2

Site-specific gene correction of the β^s mutation in the HBB gene. (A) A scheme of stepwise gene correction of 1 mutant β^s allele (shown as 3 exons, 2 introns, and flanking sequences), first by HR-mediated gene targeting and followed by Cre-mediated excision. The gene-targeting donor BD2 vector with 2 homology arms (5.9-kb left arm and 2-kb right arm indicated by X) introduces an HR template for T-to-A replacement in the β^s allele and a loxP-flanked drug-selection cassette PGK-Hyg to be inserted into the HBB intron 1. The flanking counter-selection HSV-TK gene (in the form of a TK.GFP fusion) driven by the EF1α promoter (outside the right HR homology arm) is used to reduce the frequency of Hyg resistant clones because of BD2 vector random integration that also allows HSV-TK expression. For validated HR clones, Cre-mediated excision removes only the PGK-Hyg selection gene cassette and leaves 1 copy of the loxP DNA in the middle of HBB intron 1 of the corrected allele, generating “cre” clones with 1 corrected allele β^A (Corrected^ΔloxP). We used 2 PCR primers (red arrows) for initial screening of TI indicative of correct HR. (B) The initial results of gene targeting in 293T cells that were transfected with the 1-μg BD2 donor alone (lane 1) or with HBB-ZFNs at increasing amounts (0.25 μg in lane 2, 1 μg in lane 3, and 2 μg in lane 4). The 3′-TI event (2.5-kb PCR product) was detected when ZFNs were present. Untransfected 293T cells in lane 5 are negative. (C) Similar results of in selected clones after gene targeting in S1 iPSCs and Hyg and GVC selection. Four clones (c36, c64, c68, and c70) showed a positive PCR product. Positive control (p.c.) was DNA from targeted 293T cells (lane 3) in panel B. (D) Southern blot analysis surrounding the HBB locus in the parental S1 and selected targeted iPSC clones. A 3′-probe downstream of 3′-homology arm (top line) is used with genomic DNA digestion by PmeI (P) and EcoRV (E) enzymes to confirm the presence of targeted allele (with EcoRV site inside TI) shown as a green arrow, and the original HBB allele in a red arrow. (E) A Hyg probe is used with genomic DNA digestion by XbaI (X) and SpeI (S) enzymes to confirm the targeted allele (with Hyg insertion, green arrow) or to identify random integration events such as event in clone 65. (F) PCR screening for clones with successful Cre excision, using the 2 primers shown in red. The absence of the Hyg-containing DNA in clones such as cre16 and cre19 indicates the excision of the PGK-Hyg cassette. (G-H) Southern blot analyses of iPSC clones before and after Cre excision. (G) Results with the 3′-probe after P and E digestion as shown in panel D. Red arrows indicated 4.3-kb fragments from β^s allele in S1 and c36 iPSC clones and β^A allele in cre clones (4, 16, and 19). (H) Southern blots with Hyg probe after X and S digestion as shown in panel E. The S1 and 3 cre clones are free of the Hyg gene, which was found in c36 clone as expected (green arrow).

Figure 3

Genomic DNA PCR and sequencing confirm precise monoallelic gene correction and Cre-LoxP excision. (A) Schematic of genomic DNA PCR using primers in 5′-untranslated region of exon 1 and exon 2 of HBB. With a short 72°C extension step (15 seconds), only the mutant allele (291 bp) and the corrected allele after excision (337 bp) can be amplified. (B) DNA gel shows 1 band for the c36 clone representing the mutant allele that can be amplified by the PCR protocol, and 2 bands from every cre clone representing both mutant allele and corrected allele after excision. (C) The mixed PCR products from cre4 iPSCs (after gene correction and Cre-LoxP excision) were cloned into a TOPO vector, and individual clones were sequenced. Among 8 sequenced clones, 4 clones (50%) were shown to contain a mutant allele (top panel), and 4 clones (50%) bore the corrected allele with a remnant loxP site after excision (bottom panel).

Figure 4

Characterization of gene-corrected SCD iPSC clones. Gene corrected SCD iPSC clones (c36 and cre4 shown here) display characteristic pluripotency markers such as alkaline phosphatase (AP), OCT4, NANOG, and TRA-1-60 (A); maintain normal karyotypes (B); and form teratomas bearing cells from all 3 germ layers, that is, ectoderm, mesoderm, and endoderm (C).

Figure 5

Erythroid differentiation of SCD iPSC clones. (A) In vitro hematopoietic differentiation of various iPSC clones to generate HBB-expressing erythroid cells. After 14 days of EB-mediated spontaneous differentiation, iPSC-derived cells were further differentiated and expanded into immature erythroid cells (erythroblasts or EryB) for another 8 days. Then, the erythroid cells were collected for flow cytometry, Giemsa stain of cytospin, and RT-PCR or qRT-PCR. (B) Flow cytometric analysis of S1, cre4, and cre16 using erythroid-specific surface markers CD71 and CD235a (glycophorin A). (C) Giemsa stain of confirming that most of differentiated cells are erythroblasts.

Figure 6

HBB and HBG transcription and translation analyses. (A) HBB and HBG1/2 gene expression (normalized to a house-keeping gene GAPDH) in undifferentiated iPSCs (S1) and differentiated progenies (EB and EryB) of S1, cre4, or cre16 was measured by quantitative RT-PCR (data represent mean ± SEM, n = 3). After the erythroid differentiation from iPSCs, the HBB transcript level increased 10- to 100-fold, although it is still 100- to 1000-fold lower compared with the level in CB-MNCs. (B) Conventional RT-PCR that readily amplifies HBB cDNA in erythroblasts derived from various SCD iPSC clones before (S1) or after gene targeting (c36) and Cre-mediated excision (cre4 and cre16), by 2 primers located at exon 1 and 3 (left illustration). (C) Although sizes of RT-PCR products of the unmodified or corrected alleles are the same, DNA sequencing of the RT-PCR product showed uniform transcript in c36-EryB and mixed transcripts in cre4-EryB and cre16-EryB (bottom chromatographs). Cloning each transcripts into TOPO vector and sequencing at clonal levels will distinguish expression from corrected versus uncorrected alleles. Sequencing of 40 to 60 individual cloned DNA molecules of RT-PCR products from each differentiated iPSC line revealed that the absence of corrected HBB transcript (T, 100% or 40/40 cloned and sequenced) in c36-derived erythroblasts, but in the erythroblasts derived from cre4 and 16 after Cre-mediated excision of the PGK-Hyg gene cassette, expression of the corrected (A) allele was detected. In cre4, both corrected (A, 28% or 17/60) and the unmodified (and mutated T) HBB alleles (72% or 43/60) were expressed. A similar result was obtained in cre16 iPSCs: 12/60 (20%) and 48/60 (80%) of the cloned and sequenced transcripts are from the corrected (A) and the uncorrected (T) alleles, respectively. (D) S1-EryB, cre4-EryB, and cre16-EryB expressed abundant fetal-type hemoglobin HbF, but no detectable adult-type hemoglobin HbA measured by flow cytometry using specific antibodies.

Figure 7

Investigation of reduced expression of the gene-targeted allele. (A) RT-PCR showed the same PGK-Hyg transgene was expressed at very low level in gene-targeted c36-iPSCs (lane 3) and c36-EryB (lane 4) compared with another iPSC line (FPHR, where the PGK-Hyg cassette was targeted into the actively expressed PIG-A gene, lane 1). Nontargeted S1-EryB cells were used as negative control (lane 2). (B) Sequencing of 3.5-kb genomic region of HBB locus in SCD iPSC clones. The 3.5-kb genomic region includes an ∼ 1.1-kb promoter, all the HBB exons and introns, and an ∼ 0.5-kb downstream sequence that were part of BD2 targeting donor (black lines/shapes/names) and also contains an ∼ 200-bp 3′-enhancer sequence (gray lines/shapes/names) downstream of the right homology arm. DNA sequencing of both alleles in early (p24) or late (p56) passage of S1 revealed uniform β^S mutation in exon 1 and wild-type GATA site in 3′-enhancer (underlined with complementary strand sequence underneath). However, in c36, cre4, or cre16, sequencing of mixed alleles showed heterozygous nucleotides (N), including β^A in exon 1, a G-to-T polymorphism near the 3′-end of the right homology arm, and an A-to-G mutation in GATA site all linked on gene-targeted allele.