DNA-PKcs Inhibition Improves Sequential Gene Insertion of the Full-Length CFTR cDNA in Airway Stem Cells

Abstract

Cystic fibrosis (CF) is caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene. Although many people with CF (pwCF) are treated using CFTR modulators, some are non-responsive due to their genotype or other uncharacterized reasons. Autologous airway stem cell therapies, in which the CFTR cDNA has been replaced, may enable a durable therapy for all pwCF. Previously, CRISPR-Cas9 with two AAVs was used to sequentially insert two halves of the CFTR cDNA and an enrichment cassette into the CFTR locus. However, the editing efficiency was <10% and required enrichment to restore CFTR function. Further improvement in gene insertion may enhance cell therapy production. To improve CFTR cDNA insertion in human airway basal stem cells (ABCs), we evaluated the use of the small molecules AZD7648 and ART558 which inhibit non-homologous end joining (NHEJ) and micro-homology mediated end joining (MMEJ). Adding AZD7648 alone improved gene insertion by 2-3-fold. Adding both ART558 and AZD7648 improved gene insertion but induced toxicity. ABCs edited in the presence of AZD7648 produced differentiated airway epithelial sheets with restored CFTR function after enrichment. Adding AZD7648 did not increase off-target editing. Further studies are necessary to validate if AZD7648 treatment enriches cells with oncogenic mutations.

Keywords: CFTR genome editing; CFTR super-exon; Cystic fibrosis; DNA-PKcs inhibition; airway stem cell therapy; sequential gene insertion; universal CFTR correction.

Figure 1.. Insertion of GFP and CFTR cDNA in non-CF ABCs in the presence of AZD7648.

(A) Schematic of the template used for knocking in a GFP expression cassette into exon 1 of CFTR using Cas9 RNP/AAV based gene editing. (B) Representative flow cytometry plots of cells edited in the presence of AZD7648 and AAV only controls. (C) Insertion of GFP in three biological replicates (2 technical replicates per donor). (D) Proliferation of ABCs edited to insert GFP in the presence of AZD7648 was not different from ABCs edited without AZD7648 measured 5 days after editing. (E) Schematic showing insertion of the CFTR cDNA and tCD19 enrichment tag into exon 1 of the CFTR locus using the sequential insertion strategy (universal strategy). The second insertion is initiated by including the same sgRNA sequence from exon 1 of CFTR at the end of the first template. The CFTR cDNA is followed by a BGH polyA tail. The tCD19 cassette is driven by a PGK promoter and has an SV40 polyA tail. (F) Representative flow cytometry plots depicting tCD19 expression in the presence of AZD7648 or AAV only controls. The AAV only plot depicts the minimal episomal expression from the use of AAV. (G) Expression of tCD19 in edited non-CF ABCs in the presence of AZD7648. Presented data is from four biological replicates (2 technical replicates per donor). (H) Proliferation of ABCs after CFTR cDNA and tCD19 insertion in the presence of AZD7648 in comparison to ABCs edited without AZD7648 measured on day 5 after editing. All statistical significance was assessed using Wilcoxon matched-pairs significant rank test. **, * represent p < 0.01 and 0.05 respectively. Cpd = compound.

Figure 2. Insertion of GFP and tCD19 in non-CF ABCs in the presence of ART558 and AZD7648.

(A) Representative flow cytometry plot showing the percent of GFP⁺ cells after genome editing performed in the presence of AZD7648 alone, or AZD7648 and ART558. (B) GFP⁺ cells measured from three different donors with the addition of AZD7648 alone, or AZD7648 and ART558. (C) Cell proliferation assessed when AZD7648 and ART558 compared with the AZD7648 only condition. (D) Representative flow cytometry plots of edited non-CF ABCs edited using the universal strategy in the presence of AZD7648 alone, or AZD7648 and ART558. (E) tCD19 expression in non-CF ABCs with the addition of AZD7648 alone, or AZD7648 and ART558. (F) Cell proliferation of ABCs edited using the universal strategy in the presence of AZD7648 only, or AZD7648 and ART558. (G) Relative edited cell yield between the AZD7648 only, or AZD7648 and ART558 conditions. All statistical significance was assessed using one-way ANOVA followed by Tukey’s test, or by ratio paired T-test. ****, **, * represent p < 0.0005, 0.01and 0.05 respectively. Cpd = compound.

Figure 3. Primary human CF ABCs edited for CFTR cDNA insertion in the presence of AZD7648.

(A) Insertion of tCD19 in CF-donor ABCs edited using the universal strategy. (B) The proliferation of edited CF ABCs was assessed 5 days after editing, with or without AZD7648. (C) The edited cell population was enriched for tCD19⁺ cells. (D) The enriched population was stained and evaluated for the presence of basal stem cell markets, P63 and KRT5, via flow cytometry. There was no significant difference in the percent of cells positive for P63 and KRT5 between the different conditions tested. (E) CF ABCs edited with or without AZD7648 were assessed for allelic frequency of tCD19⁺ alleles. (F) Enriched, edited CF cells were sequenced for off-target activity associated with the one known off-target site for the sgRNA used in editing. Off-target editing was not significantly different between the different conditions by one way ANOVA. Cpd = compound. * represents p < 0.05 in all panels.

Figure 4. Differentiation and expression of CFTR in edited CF airway cells after differentiation on ALI transwells.

WT non-CF ABCs and edited CF ABCS (with or without AZD7648) were differentiated on ALI transwells for at least 4 weeks. (A) After differentiation, transepithelial electrical resistance (TEER) was assessed in unedited CF controls (mock) and edited (No AZD7648 and AZD7648) samples. (B) Differentiated ALI cultures contained ciliated cells expressing acetylated tubulin and secretory cells positive for CD66. (C)CFTR protein expression in differentiated airway cells from non-CF, uncorrected CF and corrected CF samples (with and without AZD7648) was assessed using immunoblotting. (D) Ussing chamber analysis was conducted to evaluate CFTR function of differentiated ALI cultures. Representative traces from fully differentiated ALI cultures derived from non-CF ABCs (WT), unedited CF ABCs (Mock), CF ABCs edited without AZD7648, and CF ABCs edited with AZD7648 are displayed. (E) Percent of CFTR function (assessed from CFTR_inh-172 response) in CF edited and unedited samples relative to non-CF controls is plotted against %tCD19⁺ cells. 20% CFTR function relative to non-CF controls is thought to be therapeutically relevant. (F) Change (ΔIsc) in CFTR inhibition as assessed from CFTR_inh-172 responses. When compared using one way ANOVA followed by Tukey’s test, the average ΔIsc values from CF samples corrected with or without AZD7648 were significantly higher than ΔIsc values uncorrected CF samples (p < 0.005). They were not significantly different from each other or the non-CF controls.