Use of adenine base editing and homology-independent targeted integration strategies to correct the cystic fibrosis causing variant, W1282X

Abstract

Small molecule drugs known as modulators can treat ~90% of people with cystic fibrosis (CF), but do not work for premature termination codon variants such as W1282X (c.3846G>A). Here we evaluated two gene editing strategies, Adenine Base Editing (ABE) to correct W1282X, and Homology-Independent Targeted Integration (HITI) of a CFTR superexon comprising exons 23-27 (SE23-27) to enable expression of a CFTR mRNA without W1282X. In Flp-In-293 cells stably expressing a CFTR expression minigene bearing W1282X, ABE corrected 24% of W1282X alleles, rescued CFTR mRNA from nonsense mediated decay and restored protein expression. However, bystander editing at the adjacent adenine (c.3847A>G), caused an amino acid change (R1283G) that affects CFTR maturation and ablates ion channel activity. In primary human nasal epithelial cells homozygous for W1282X, ABE corrected 27% of alleles, but with a notably lower level of bystander editing, and CFTR channel function was restored to 16% of wild-type levels. Using the HITI approach, correct integration of a SE23-27 in intron 22 of the CFTR locus in 16HBEge W1282X cells was detected in 5.8% of alleles, resulting in 7.8% of CFTR transcripts containing the SE23-27 sequence. Analysis of a clonal line homozygous for the HITI-SE23-27 produced full-length mature protein and restored CFTR anion channel activity to 10% of wild-type levels, which could be increased three-fold upon treatment with the triple combination of CF modulators. Overall, these data demonstrate two different editing strategies can successfully correct W1282X, the second most common class I variant, with a concomitant restoration of CFTR function.

Keywords: CFTR; CRISPR; HITI; W1282X; adenine base editing.

Figure 1

Representative EditR analyses of genomic DNA samples from A) Flp-In-293 W1282X cells transfected with gRNA-A6/NG-ABEmax and B) for NG-ABEmax (no gRNA) transfected cells as a negative control. C) Analysis from four independent DNA experiments showed A>G editing at the target A at position 6 of 24% ± 3% (n = 4) and A>G editing at the neighbouring A at position 7 of 25% ± 2% (n = 4). Representative EditR analyses of mRNA samples from cells transfected D) with gRNA-A6/NG-ABEmax or E) NG-ABEmax (no gRNA) as a negative control. F) Analysis from three independent mRNA experiments showed A>G editing at the target A at position 6 of 46% ± 2.8% (n = 3) and A>G editing at the neighbouring A at position 7 of 41% ± 1.7% (n = 3).

Figure 2

Protein expression by western blot of edited Flp-In-293 W1282X clones. A) Densitometry analysis of western blot data. Bar graph shows quantification of CFTR protein from the immunoblots using ImageJ. Data are represented as the ratio of full-length mature band C/full-length immature band B (mean± SEM, n = 3; one-way ANOVA followed by Dunnett’s multiple comparison test, **** p < 0.0001). B) Western blot showing CFTR protein expression in representative clones (isolated from a pool of cells transfected with A6 gRNA and ABEmax-NG) which have no editing (clone 2.38), A7 only edited (clone 2.29), both A6 and A7 edited (clones 2.10 and 19), or A2, A6, and A7 edited (clones 2 and 2.45); Na⁺K⁺ATPase was used as a loading control. C) Predicted protein sequence based on clonal DNA sequence—black = wild-type sequence, blue = uncorrected A6 in PTC codon, orange = bystander edits and corresponding amino acid substitutions, boxed amino acids or nucleotides indicate wild-type sequence at codon 1282.

Figure 3

CFTR functional recovery in W1282X homozygous primary nasal cells electroporated with ABE and W1282X gRNA. A) Editing efficiency at A6 and A7 positions. Individual points are from different replicates. Line & error bars denote mean ± SEM. Statistical significance was determined by one-way ANOVA followed by Sidak’s test for multiple comparisons. ****(p < 0.0001), ***(p < 0.001); B) I_sc traces of primary HNEs^{W1282X/W1282X} electroporated with A6 gRNA and NG-ABEmax plasmids (four independent transfections plus HNEs transfected with ABE in absence of gRNA). R1, R2, R3, R4 indicate technical replicates; C) I_sc trace of non-electroporated HNE^WT/WT; D) Graphical summary of ΔI_sc data from two biological replicates of W1282X/W1282X HNEs. ΔI_sc values from WT/WT control are from a single individual; I_sc trace of non-electroporated HNEs^{W1282X/W1282X} is not shown.

Figure 4

VX-445-VX-661-VX-770 combination therapy partially restores CFTR function in CFBE cells stably expressing R1283G variant. A) Representative short-circuit current (I_sc) recordings in CFBE41o- cells stably expressing R1283G variant. Cell were treated with either DMSO (0.06%) or combination of two correctors VX-445 and VX-661 (3 μM each) for 24 h. VX-770 (10 μM) was added acutely at the time of CFTR function measurement for which cells were mounted on Ussing Chambers. CFTR function was measured under the voltage clamp mode using asymmetric buffer. After the baseline I_sc stabilized, forskolin (10 μM, basolateral), VX-770 (10 μM, apical) and CFTR_inh-172 (Inh-172; 10 μM, apical) were sequentially added at the indicated times. Individual I_sc recordings were acquired with Acquire and Analyze software (Physiologic Instruments) and plotted using GraphPad Prism 7.01 software. B) CFTR-specific function defined as the difference (ΔI_sc) between the sustained phase of the I_sc response after stimulation with forskolin followed by Ivacaftor or DMSO as indicated and the baseline achieved after adding CFTR_inh-172 (mean ± SD). Two different clones of R1283G were tested. Open circle represents mean of two technical replicates from each clone.

Figure 5

A) Design of superexon 23–27 HITI and PCR strategies to assess integration. The superexon construct starts with 165 bp of CFTR intron 22 sequence upstream of the gRNA recognition site, then contains the same gRNA recognition site, but with the PAM and spacer in the reverse complementary orientation, followed by the next 80 bp from intron 22. As detailed in Supplementary Material, Fig. S3, this region contains three PCR primer binding sites (NGS 5′ jct F, NGS R, and int 22 R). To avoid the superexon donor acting as potential template for HDR, the sequence between the NGS 3′ jct F and NGS R primer binding sites has been modified to reduce sequence identity to 66.67% across a 219 bp region; this also prevents the primer NGS 5′ jct F from binding to the superexon construct. The next portion of the superexon is the last 102 bp of intron 22 which comprises the splice acceptor site followed by the 724 bp sequence comprising exons 23–27 and the TAG stop codon. B) Sequence analysis of PCR amplicons to verify correct integration. C) Detection of correct integration at 5′ and 3′ sites in cells enriched for GFP expression. D) Cartoon representation of SE-specific RT-PCR to detect splicing of endogenous exon 21 to ex23* of integrated superexon using primers ex21 F and ex23* R. E) RT-PCR shows detection of splicing of endogenous exon 22 to ex23* only in cells transfected with pSE23-27 and pCas9/gRNA38B. Sequence analysis of RT-PCR from FACS sorted cells confirms correct splice junction of Exon22-Exon23*. F) Cartoon representation of NGS RT-PCR to quantify splicing of endogenous exon 21 to ex23* of integrated superexon using primers ex21 F and ex24^25 R in pool of edited cells. G) Representative sashimi plots of RNA extracted from WT 16HBE14o⁻ (top), unedited 16HBEge W1282X (clone 8—middle) and HITI edited 16HBEge W1282X (clone 19—bottom).

Figure 6

A) Western blot showing CFTR protein expression in the clone 19 with superexon^23–27 inserted in intron 22. Short circuit current measurements in absence of modulators for B) wild-type cells, C) clone 19 with superexon^23–27; D) clone 8 without superexon and E) average values. Short circuit current measurements in presence of single (VX-770) or triple (VX-445, VX-661, and VX-770) combination of modulators for F); clone 8 without superexon G) clone 19 with superexon^23–27; and H) average values.