Intrathymic AAV delivery results in therapeutic site-specific integration at TCR loci in mice

Abstract

Adeno-associated virus (AAV) vectors have been successfully exploited in gene therapy applications for the treatment of several genetic disorders. AAV is considered an episomal vector, but it has been shown to integrate within the host cell genome after the generation of double-strand DNA breaks or nicks. Although AAV integration raises some safety concerns, it can also provide therapeutic benefit; the direct intrathymic injection of an AAV harboring a therapeutic transgene results in integration in T-cell progenitors and long-term T-cell immunity. To assess the mechanisms of AAV integration, we retrieved and analyzed hundreds of AAV integration sites from lymph node-derived mature T cells and compared these with liver and brain tissue from treated mice. Notably, we found that although AAV integrations in the liver and brain were distributed across the entire mouse genome, >90% of the integrations in T cells were clustered within the T-cell receptor α, β, and γ genes. More precisely, the insertion mapped to DNA breaks created by the enzymatic activity of recombination activating genes (RAGs) during variable, diversity, and joining recombination. Our data indicate that RAG activity during T-cell receptor maturation induces a site-specific integration of AAV genomes and opens new therapeutic avenues for achieving long-term AAV-mediated gene transfer in dividing cells.

Graphical abstract

Figure 1.

Retrieval of AAV ISs. (A) Schema of the experimental strategy. LN cells, SPL, and LIV tissues were collected from Zap70-deficient mice treated with AAV8. BR and LIV tissues were collected from Mecp2-deficient mice systemically injected with AAV-PHP. DNA was purified from the collected tissues and SLiM-PCR was performed to amplify vector/host-genome junction. PCR amplicons were assembled into libraries, sequenced, and reads were analyzed using RAAVIoli, a bioinformatics pipeline tailored to identify AAV ISs. (B) The numbers of AAV ISs retrieved from LN-derived mature T cells from LN, SPL, LIV, and BR tissues from the different groups of treated mice are presented with the median indicated as a colored line. (C) Relative frequency (indicated as percentage) of AAV ISs characterized by either precise homology breakpoints (Exact) and microhomology regions (MHoRs) between the vector and the host chromosomal sequences or random nucleotide insertions (insert). (D) Integration site distribution within gene bodies and their surrounding genomic regions. Each gene interval was quantified from the TSS up to the end of its coding region; this interval is considered as 100% and then normalized in bins of 20%. The surrounding genomic regions are divided in intervals of 20 kb. In panel C, statistical analyses were performed using a Fisher exact test ∗∗∗∗ P < .0001; see supplemental Table 2 for detailed statistical comparison.

Figure 2.

Cluster of AAV ISs within TCR genes. Genome-wide distribution of AAV ISs in the different tissues of ZAP70-deficient (ZAP70), WT mice, and Mecp2-deficient mice (Mecp2) intrathymically injected with AAV: LN, SPL, LIV, and BR are shown as indicated. Most of AAV integrations in the LN and SPL data sets from intrathymically injected mice were detected at chr6, 13, and 14.

Figure 3.

Distribution of AAV ISs within Tcr loci. (A-C) Genomic distribution of AAV IS in the LN and SPL data sets of ZAP70-treated mice and targeting TCR genes: TCRα (A), TCRβ (B), and TCRγ (C), as indicated. Genomic coordinates and scale are indicated in each panel. Black and blue lines indicate the position of the AAV IS from the LN (black) and SPL (blue) data sets. V, D, and J segments are indicated by yellow, blue, and purple rectangles, respectively. (C) Constant regions are represented as brown rectangles. Clusters of AAV ISs are identified in TCR genes, especially in the J-segment region, whose genomic area is enlarged below the Tcrα and Tcrβ loci; gene segment genomic coordinates were retrieved from IMGT, the international ImMunoGeneTics information system. (D) Gene targeting frequency (indicated as percentage) of Tcrα, Tcrβ, and Tcrγ in the different AAV ISs data sets as indicated.

Figure 4.

AAV integrations occurred at RAG-induced DSBs. (A) Schematic representation of the organization of the murine Tcrα/Tcrδ, Tcrγ, and Tcrβ loci. Genomic proportions are not preserved. V, D, and J segments are indicated by yellow, blue, and purple rectangles, respectively. RSS are represented as triangles close to the V, D, and J gene segments: RSS with a 12-bp spacer are indicated as grey triangles, RSS with a 23-bp spacer are indicated as white triangles. TCR constant regions are indicated as Ca, Cb, Cg and Cd and represented as pink rectangles. The T early exon (TEA) in the TCRα locus is represented by a green rectangle. Trypsinogen genes in the TCRβ locus (T) are represented by brown rectangles. Enhancers are represented by grey circles and indicated as E. Germ line transcription is indicated by black arrows above each segment. (B) Frequency (indicated as percentage) of AAV ISs surrounded by candidate RSSs in a genomic window of 100 bp (± 50 bp). (C) Significant motifs identified by MEME algorithm within the LN and SPL data set of ZAP70-treated mice. (D) Schema of the conserved heptamer and nonamer sequence of RSS. (E, F) Distribution of AAV ISs at the 3′ region of V segments (E) and at the 5′ region of J segments (F). Statistical analyses in panel B were performed using the Fisher exact test. ∗∗∗∗P < .0001. chr, chromosome; MEME, multiple Em for motif elicitation.

Figure 5.

AAV genomic features at the vector/host-genome junction. (A-E) Pie charts indicating the frequency of AAV features found at the junction breakpoint between the integrated vector and the host genomic region for the different IS data sets as indicated. (F-J) Heatmap of the AAV 3′ITR secondary structure with the red scale indicating the frequency of AAV insertions occurring at the indicated nucleotide position for the different IS data sets. bGlobinSA, beta globin splice acceptor; CDS, coding sequence; hGH, human growth hormone; mPGK, murine phosphoglycerate kinase.

Figure 6.

Absolute and relative abundance of AAV integration sites. (A) Cellular genomes observed for each IS (Genomes/IS) in the different IS data sets are presented as indicated. (B-E) Stacked bar plots showing the relative abundance of AAV IS retrieved in ZAP70-treated mice (B,C), in WT mice intrathymically injected with a GFP-expressing AAV (D), and in Mecp2-deficient mice (E). In each column, every integration is represented by a different color whose height is proportional to the number of genomes retrieved for that specific IS (percentage of IS abundance, y-axis). The number of unique IS identified in each mouse is indicated in blue above each column. Ribbons connect AAV ISs tracked between different tissues of the same mouse. In A and B, statistical analyses were performed by a two-tailed Mann-Whitney test. ∗∗∗∗P < .0001.