Highly Efficient Gene Editing of Cystic Fibrosis Patient-Derived Airway Basal Cells Results in Functional CFTR Correction

Abstract

There is a strong rationale to consider future cell therapeutic approaches for cystic fibrosis (CF) in which autologous proximal airway basal stem cells, corrected for CFTR mutations, are transplanted into the patient's lungs. We assessed the possibility of editing the CFTR locus in these cells using zinc-finger nucleases and have pursued two approaches. The first, mutation-specific correction, is a footprint-free method replacing the CFTR mutation with corrected sequences. We have applied this approach for correction of ΔF508, demonstrating restoration of mature CFTR protein and function in air-liquid interface cultures established from bulk edited basal cells. The second is targeting integration of a partial CFTR cDNA within an intron of the endogenous CFTR gene, providing correction for all CFTR mutations downstream of the integration and exploiting the native CFTR promoter and chromatin architecture for physiologically relevant expression. Without selection, we observed highly efficient, site-specific targeted integration in basal cells carrying various CFTR mutations and demonstrated restored CFTR function at therapeutically relevant levels. Significantly, Omni-ATAC-seq analysis revealed minimal impact on the positions of open chromatin within the native CFTR locus. These results demonstrate efficient functional correction of CFTR and provide a platform for further ex vivo and in vivo editing.

Keywords: basal stem cell; cystic fibrosis; gene editing.

Graphical abstract

Figure 1

Site-Specific Editing of CF Airway Basal Cells (A) Site-specific correction of ΔF508 mediated by targeted nuclease cleavage with wild-type DNA template serving as donor. (B) Integration of CFTR_9-27 cDNA (encoding exons 9–27), preceded by a splice acceptor (SA), followed by polyadenylation (pA) sequences, and flanked by homology sequences, into CFTR intron 8; the spliced mRNA joins endogenous exon 8 with the exon 9–27 transgene.

Figure 2

Sequence-Specific ΔF508 Correction with Restoration of Mature CFTR Protein Expression and Function (A) Schematic of sequence-specific ΔF508 correction strategy. ZFNs were designed to site-specifically recognize and cleave CFTR ΔF508 sequences in exon 11 (ZFN11). Two donors carrying correcting CFTR sequence are shown: a single stranded 200-mer oligo DNA donor (ssDNA) and 2 kb AAV-6 donor (containing 1,963 bp of wild-type CFTR sequence spanning exon 11). Three anticipated outcomes are represented as unmodified, indels, and corrected. Primers f₁ and r₂, located outside the 1,963 bp donor sequence, were used to amplify the targeted region in order to assess genome modification efficiency. (B) Efficiency of genome modification analyzed by NGS. Z, ZFN11 alone; Z+D, ZFN11 plus donor DNA (mean ± SD, n = 3 biological replicates; see also Table S1). (C) H&E staining and immunofluorescence detection of airway epithelium markers in ALI culture. Major epithelial cell types were identified with markers: p63 or keratin 5 (CK5) for basal cells; mucin 5AC (MUC5AC) for secretory cells; acetylated tubulin (ACT) for ciliated cells. Representative 40× images in transverse section staining. Scale bar, 50 μm (panels). (D) Western blotting of CFTR. Sequence-specific ΔF508 correction restores expression of mature, fully glycosylated CFTR protein (band C). Calnexin, loading control. (E) Representative trace of short circuit current (Isc) from 4 samples tested: DMSO-treated, VX-809/VX-770 pre-treated and Z+D-treated CF (ΔF508/ΔF508), and non-CF evaluated by Ussing chamber analysis. (F) Summary of CFTR chloride current stimulated by forskolin and inhibited by CFTR inhibitor 172 (CFTR inh-172). Difference of Isc (Δlsc) (μA/cm²) before and after treatments are calculated from traces represented by (E). Values of Δlsc from samples listed in Table S2 are summarized and shown (mean ± SD, n = 3 experiments). (G) Restored CFTR activity as function of gene-editing frequency. The blue symbols reflect results from individual ΔF508/ΔF508 cell experiments with sequence-specific correction of ΔF508 utilizing either ssDNA or AAV-6 donor. CFTR activity (forskolin-induced) for each experiment is expressed as % of non-CF (e.g., see Table S2A). Frequency of genome modification for each experiment is from Table S1. Shown in open and filled blue circles are results of ΔF508 correction for ΔF508/ΔF508 cells with ssDNA and AAV-6 donors, respectively; the black filled circle is from ΔF508/ΔF508 cells. Linear regression for ΔF508 sequence-specific correction experiments resulted in an R² = 0.92.

Figure 3

Efficient SA-CFTR_9-27-pA TI into CFTR Intron 8 of ΔF508/ΔF508 Airway Basal Cells (A) Schematic of site-specific targeted editing of CFTR intron 8. Unmodified and indel diagrams highlight CFTR genomic sequences between exons 8 and 9 (black boxes; not to scale). TI-8 diagram shows intron 8 TI of human codon optimized CFTR_9-27 cDNA preceded by a splice acceptor, followed by bovine growth hormone (bGH) pA sequence, and flanked by 313 bp homology left (H_L) and 351 bp homology right (H_R) intron 8 sequences. Arrows indicate oligos amplifying unmodified, indel, or TI-8 events and used to quantify frequency of each by NGS. (B) The percentage of genome modification (mean ± SD) determined by NGS (Table S3A) from five independent TI-8 experiments (utilizing the AAV-6 donor at either 2 × 10⁶ or 6 × 10⁶ viral genomes [vg]/cell). There was no enhancement of TI efficiency as one increased the amount of donor. Efficiency was measured 4 days after the delivery of ZFNs targeting intron 8 (ZFN8) followed immediately by AAV-6 CFTR_9-27 cDNA donor. Z, ZFN8 alone; Z+D, ZFN8 and AAV-6 donor. (C) Southern blot analysis of TI-8. The schematic shows the expected genomic organization of non-targeted (no TI-8) and targeted (TI-8) alleles with the expected sizes resulting from EcoRI digestion. The non-targeted allele yields a 9.8 kb band while the TI-8 allele yields a 6.9 kb band due to the new EcoRI site in the CFTR_9-27 cDNA donor construct. (D) H&E staining and immunostaining of well-differentiated airway epithelium in ALI culture. Representative images of cross sections from ΔF508/ΔF508 or TI-8 ΔF508/ΔF508. Immunostaining shows major epithelial cell types: basal cell (CK5 and P63), secretory cells (MUC5AC), and ciliated cells (ACT). 40× images, scale bar, 50 μm. (E) Detection of transgene CFTR_9-27 mRNA. Schematic of endogenous and transgene CFTR mRNA is shown here. Primer T9 recognizes codon-optimized transgene exon 9 sequence only (blue) while Primer E9 recognizes endogenous exon 9. A 250 bp E8-T9 RT-PCR amplicon, evidence of the chimeric endogenous-transgene CFTR mRNA, was present only in the TI-8 ΔF508/ΔF508 sample. (F) Restoration of fully glycosylated CFTR protein via TI-8. Western blots of protein lysates harvested at 4 weeks of ALI culture. Band C, representing the mature, fully glycosylated form of CFTR, is present in non-CF, absent in ΔF508/ΔF508 cells, and restored in TI-8 ΔF508/ΔF508. Calnexin, loading control. (G) CFTR functional assay in ALI cultures at 4 weeks. TI-8 ΔF508/ΔF508 cells show the restoration of CFTR function measured as Δlsc (μA/cm²) ± SD and compared with ΔF508/ΔF508 treated with VX-809/VX-770. (H) Restored CFTR activity as function of gene-editing frequency. The blue and black symbols and blue linear regression line are from individual ΔF508/ΔF508 cell experiments with sequence-specific correction of ΔF508 (from Figure 2G). Shown in filled square and open square orange symbols are the results of ΔF508/ΔF508 cell TI-8 and ΔF508/ΔF508 cell TI-7 experiments, respectively.

Figure 4

Targeted Integration of a Partial cDNA at CFTR Intron 8 Does Not Disrupt the Open Chromatin Profile of the ΔF508/ΔF508 CFTR Locus Integrative Genomics Viewer (IGV) browser graphics of CFTR TAD regions showing Omni-ATAC profiles of non-CF, ΔF508/ΔF508, and TI-8 ΔF508/ΔF508 ALI cells. The number of sequencing reads is shown in Table S7. Key cis-regulatory elements are marked above the ATAC profiles by arrows. DHS denotes all DNase I-hypersensitive sites across the locus. Below is a magnified panel of the intron 8 region showing appearance of a peak of open chromatin at the TI-8 cDNA insertion site.

Figure 5

Efficient TI-8 Correction in CF Airway Basal Cells Carrying CFTR PTC Mutations (A) Frequency of genome modification determined by NGS in CF basal cells carrying ΔF508/R553X (mean ± SD, n = 3 experiments) or G542X/R785X (mean ± SD, n = 4 experiments) variants (from Tables S3C and S3D). Highly efficient TI-8 was confirmed in CF basal cells treated with ZFN8 and AAV-6 CFTR_9-27 cDNA donor (Z+D) 4 days after genome editing. (B) Southern blot analysis of TI-8. A 6.9 kb fragment confirmed the expected TI-8 genomic organization in CF basal cells of either genotype. (C) Restoration of fully glycosylated CFTR protein via TI-8. Western blot of protein lysates harvested at 4 weeks of ALI culture shows the presence of band C in TI-8 and the absence in non-targeted cells. Calnexin, loading control. (D) CFTR functional assay in ALI cultures. Shown are short-circuit current measurements for bulk TI-8 ΔF508/R553X and TI-8 G542X/R785X cells. Restored CFTR currents are shown as Δlsc (μA/cm²) ± SD and are in excess of the VX-809/VX-770 treated cells. No VX-809/VX-770 response is expected for G542X/R785X.