AAV-mediated genome editing is influenced by the formation of R-loops

Abstract

Recombinant adeno-associated viral vectors (rAAV) hold an intrinsic ability to stimulate homologous recombination (AAV-HR) and are the most used in clinical settings for in vivo gene therapy. However, rAAVs also integrate throughout the genome. Here, we describe DNA-RNA immunoprecipitation sequencing (DRIP-seq) in murine HEPA1-6 hepatoma cells and whole murine liver to establish the similarities and differences in genomic R-loop formation in a transformed cell line and intact tissue. We show enhanced AAV-HR in mice upon genetic and pharmacological upregulation of R-loops. Selecting the highly expressed Albumin gene as a model locus for genome editing in both in vitro and in vivo experiments showed that the R-loop prone, 3' end of Albumin was efficiently edited by AAV-HR, whereas the upstream R-loop-deficient region did not result in detectable vector integration. In addition, we found a positive correlation between previously reported off-target rAAV integration sites and R-loop enriched genomic regions. Thus, we conclude that high levels of R-loops, present in highly transcribed genes, promote rAAV vector genome integration. These findings may shed light on potential mechanisms for improving the safety and efficacy of genome editing by modulating R-loops and may enhance our ability to predict regions most susceptible to off-target insertional mutagenesis by rAAV vectors.

Extended Data Figure 1:. DRIP-seq analysis.

a. DRIP-qPCR analysis in the whole mouse brain (n=3). b. GC (blue) and AT (red) skew acros coding strand of DRIP peaks, including 2kb flanking 5’- and 3’-ends in 3T3 and E24 murine cells. Bands represent 95% CI of mean skew signal c. GC content percentage of DRIP peaks in HEPA1-6, liver, 3T3, and E14 cells. d. As in c, but for AT content. e. Aggregate plots around the center of DRIP peaks showing non-B DNA motif density as percent coverage Statistical analysis: a. Two-way ANOVA with Sidak’s post hoc analysis. Error bars represent standard deviation of the mean.

Extended Data Figure 2:. DRIP seq comparison HEPA1-6 vs. liver

a. Genome browser views of DRIP-seq tracks of representative R-loop forming genes in HEPA1-6 and liver tissue. Tracks from S9.6 immunoprecipitation (IP), Input and IP with RNaseH treatment (IP+RH) are shown. b. Heatmaps showing the DRIP-seq signal around the center of peaks called uniquely in HEPA1-6 Heatmaps are ranked by DRIP-seq signal strength within 3kb of the peak center. Correlation coefficients (Spearman’s Rho) are shown c. Fractions of DRIP peaks from HEPA1-6 or liver overlapping murine genomic features. d. Lengths of genes overlapping DRIP peaks. Gene lengths for liver only peaks are significantly different to other peaksets (P=4.5e-20, Chi-square test). e. As in b, but for DRIP-seq signal around the center of peaks called uniquely in liver tissue. f. AT (top) and GC (bottom) content of DRIP peaks in HEPA1-6 and liver. g. Histogram of peak sizes obtained from peak calling on DRIP versus input samples. h. Table containing descriptive statistics of DRIP peaks in HEPA1-6 and liver.

Extended Data Figure 3:. In vitro AAV-HR and R-loops

a. Luciferase activity in HEPA1-6 upon AAVDJ-mAlb/Fluc transduction at different multiplicity of infection. Non transduced cells (No AAV) were used as negative control (n=3). b. mRNA levels upon transfection of siRNA against R-loops genes normalized against cells treated with Scramble siRNA (n=3). c. Luciferase activity in HEPA1-6 upon transfection of siRNAs against R-loops related proteins and AAVDJ-mAlb/Fluc transduction (n=6). Results were normalized to scramble-treated samples d. Time course of luciferase activity in HEPA1-6 upon treatment with either different doses of totopotecan or DMSO as negative control and transduced with AAVDJ-mAlb/Fluc (n=3). Statistical analysis: a, c. One-way ANOVA with a. Tukey’s and c. Sidak post hoc analysis. b. Multiple t-test using Holm-Sidak method. d. Two-way ANOVA with Sidak’s post hoc analysis. Error bars represent standard deviation of the mean.

Extended Data Figure 4:. In vitro AAV-HR and R-loops

a. Mice body weight upon treatment with different doses of topotecan and AAV-HR-mAlb/hFIX (n=3). b. AAV vector genome copy number in liver of mice treated with AAV-HR-mAlb/hFIX and different doses of topotecan (n=3). c. mRNA levels in HEPA1-6 transduced with AAV-HR-mAlb/Fluc and AAV-HR_neg-mAlb/Fluc (n=3). d. Luciferase activity in HEPA1-6 upon the cell were treated with either totopotecan or pRNaseH1 plasmid or their combination, and AAV-HR-mAlb/Fluc or AAV-HR_neg-mAlb/Fluc (n=3). Statistical analysis: b. One-way ANOVA with Tukey’s post hoc analysis. Error bars represent standard deviation of the mean.

Figure 1:. DRIP-seq analysis.

a. Schematic of the DRIP-seq workflow for HEPA 1-6 and murine liver. Two biological replicates for each sample were sequenced. b. Schematic of whole tissue preparation for DRIP-seq. c. Fractions of DRIP peaks overlapping murine genomic features from HEPA1-6 (Hepa), liver, E14 embryonic stem cells (E14) and 3T3 fibroblast cells (3T3). d. Left, aggregate plots of mean GC (blue) and AT (red) skew across the coding strand of DRIP peaks, including 2 kb flanking the 5’- and 3’-ends. Bands represent 95% CI of mean skew signal. Right, heatmaps showing skew across all individual peaks plotted on the left, ranked by magnitude of skew. Scale bar shows skew values. e. Aggregate plots around the center of DRIP peaks showing the G-quadruplex motif density. f. Aggregate plot of DRIP-seq signal around the TSS of mouse genes, showing signal from IP, inputs (In) and RNaseH-treated (+RH) samples. g. As for f, but around the TES of mouse genes. h. Venn diagram of genome areas (in megabases; Mb) occupied by peaks called from DRIP IP vs input in HEPA1-6 and liver tissue. i. Venn diagram of R-loop positive genes in HEPA1-6 and liver tissue. j) Aggregate plots showing DRIP-seq signal around the center of different sets of DRIP peaks from HEPA1-6 and liver. IP and input samples are shown. h. Heatmaps showing the DRIP-seq signal around the center of peaks called in HEPA1-6 (left) or liver (right). Heatmaps are ranked by DRIP-seq signal strength within 3kb of the peak center. Correlation coefficients (Spearman’s Rho) are shown.

Figure 2:. RNA-seq analysis and comparison with DRIP-seq in HEPA1-6 and liver.

a. Volcano plot showing downregulated (blue) and upregulated (red) genes in HEPA1-6 vs. Liver, gene numbers a shown. b. Gene expression levels in HEPA1-6 and liver. Expression level cutoffs are from EMBL-EBI Expression Atlas (https://www.ebi.ac.uk/gxa/): off (<0.5 FPKM), low (between 0.5 to 10 FPKM), medium (between 11 to 1000 FPKM), high (above 1000 FPKM). c. Scatter plot showing RNA-seq FPKM values for 65,440 mouse genes, Pearson’s correlation coefficient is shown. d. Heatmap graph showing differential expression analysis between HEPA1-6 and murine liver RNA-seq data (HEPA1-6 vs. Liver; n=3). e. Scatter plot showing DRIP-seq signal at 65,440 mouse genes, Pearson’s correlation coefficient is shown. f. Gene expression level (FPKM) of genes overlapping (R-loop positive) or not (R-loop negative) with DRIP-seq peaks in HEPA1-6 and liver. Box: 25th and 75th percentiles; central line: median. g. Scatter plot showing DRIP-seq signal vs RNA-seq signal at 65,440 mouse genes for HEPA1-6 (left) and liver (right), Pearson’s correlation coefficients are shown. h. Aggregate plots of DRIP-seq signal around the TSS of mouse genes for HEPA1-6 (left) and liver (right). Mouse genes were sorted by gene expression levels and distributed in octiles of increasing FPKM.

Figure 3:. DRIP-qPCR in liver and HEPA1-6 correlates with DRIP-peaks.

a. Genome browser tracks showing HEPA1-6 and liver DRIP-seq signal at the albumin gene (Alb). Tracks from S9.6 immunoprecipitation (IP), Input and IP with RNaseH treatment (IP+RH) are shown. b. Schematic of Alb locus and design of primers used for DRIP-qPCR. c-d. DRIP-qPCR data in HEPA-1–6 (n=2) (c) and liver (n=6) (d). RNH is in vitro RNaseH treatment prior to immunoprecipitation. e. Western blot showing expression of human RNaseH1 in HEPA1-6 upon transient transfection of the ppy-CAG-RNaseH1 (pRNAseH1) plasmid. f. Luciferase activity in HEPA1-6 upon transient transfection of pRNAseH1 plasmid and AAVDJ-mAlb/Fluc transduction (n=12). g. Luciferase activity in HEPA1-6 upon transfection of siRNAs against R-loop targeting proteins (siFancm, siSrsf1, or siScramble) and AAVDJ-mAlb/Fluc transduction. Cells were transiently transfected with or without pRNAseH1 plasmid (n=3). h. Luciferase activity in HEPA1-6 following treatment with fludarabine, totopotecan or DMSO (negative control) and transduced with AAVDJ-mAlb/Fluc. Cells were transiently transfected with or without pRNAseH1 plasmid (n=3). Statistical analysis: f. Unpaired t-test. d, g, h. Two-way ANOVA with Sidak’s post hoc analysis. Error bars represent standard deviation of the mean.

Figure 4:. R-loops determine in vitro and in vivo AAV-HR at Alb locus.

a. Schematic representation of topotecan treatment in vivo. b. DRIP-qPCR analysis of murine livers upon treatment with 5mg/kg of topotecan (n=2). c. Schematic representation of the combination of topotecan treatment and AAV-HR in vivo. d. Time course of hFIX circulating levels in mice treated with the combination of AAV-HR-mAlb/hFIX and topotecan (n=3). e. Schematic representation of AAV-HR and AAV-HR_neg design for genome targeting at the Alb locus. f. Time course of luciferase activity and AAV vector genome copy number in HEPA1-6 transduced with AAV-HR-mAlb/Fluc and AAV-HR_neg-mAlb/Fluc (n=3). g. Time course of hFIX circulating levels in mice treated with either AAV-HR-mAlb/hFIX or AAV-HR_neg-mAlb/hFIX (n=4). h. AAV vector genome copy number in liver of mice treated with either AAV-HR-mAlb/hFIX or AAV-HR_neg-mAlb/hFIX (n=4). Statistical analysis: d. Two-way ANOVA with Sidak’s post hoc analysis. Error bars represent standard deviation of the mean.

Figure 5:. Correlation between off-target genes and R-loops

a. IGV visualization of previously reported genes where AAV vectors were found integrated upon in vivo liver gene transfer. b. Heatmap showing DRIP-seq and RNA-seq values of genes represented in a.

Figure 6:. Mechanism of AAV-mediated integration at R-loop-forming loci

Systemic treatment with AAV for liver-mediated gene therapy. The single-stranded AAV genome is released into the nucleus of hepatocytes, undergoing homologous recombination (HR) with the single-stranded genomic DNA displaced by R-loop formation in transcriptionally active regions.