AAV vectors containing rDNA homology display increased chromosomal integration and transgene persistence

Abstract

Although recombinant adeno-associated viral (rAAV) vectors are promising tools for gene therapy of genetic disorders, they remain mostly episomal and hence are lost during cell replication. For this reason, rAAV vectors capable of chromosomal integration would be desirable. Ribosomal DNA (rDNA) repeat sequences are overrepresented during random integration of rAAV. We therefore sought to enhance AAV integration frequency by including 28S rDNA homology arms into our vector design. A vector containing ~1 kb of homology on each side of a cDNA expression cassette for human fumarylacetoacetate hydrolase (FAH) was constructed. rAAV of serotypes 2 and 8 were injected into Fah-deficient mice, a model for human tyrosinemia type 1. Integrated FAH transgenes are positively selected in this model and rDNA-containing AAV vectors had a ~30× higher integration frequency than controls. Integration by homologous recombination (HR) into the 28S rDNA locus was seen in multiple tissues. Furthermore, rDNA-containing AAV vectors for human factor IX (hFIX) demonstrated increased transgene persistence after liver regeneration. We conclude that rDNA containing AAV vectors may be superior to conventional vector design for the treatment of genetic diseases, especially those associated with increased hepatocyte replication.

Figure 1

AAV-rDNA-hFah yields about 30 times higher integration frequency. (a) AAV-rDNA-hFah structure and homologous recombination into 28S rDNA. The overall structure of the mouse 45S rDNA unit and the location of the region used in the rDNA vector are shown. (b) Schematic of Fah^−/− liver function rescue experiment. (c) Comparison of the dose responses of AAV8-rDNA-hFah and AAV-stuffer-hFah. The NTBC cycle number needed to rescue the Fah^−/− mice is given. At each virus genome (vg) dose, AAV-rDNA-hFah required fewer NTBC cycles to rescue liver function. (d) Experimental design for the comparison of AAV-rDNA-hFah and AAV-stuffer-hFah without transgene selection. (e) Copy number determination by qPCR. The δCt for hFAH versus Rosa26, (donor hepatocyte marker) is shown for AAV-rDNA-hFah and AAV-stuffer-hFah. All hFAH and Rosa26 PCR Ct values were normalized to the single copy gene Maai. (f) Relative total integration frequencies at the indicated dose. All integration frequencies were normalized to that of AAV-stuffer-hFAH at 1 × 10¹⁰ (3rd bar), which was arbitrarily set to 1. AAV, adeno-associated viral vector; NA, not applicable; NTBC, 2-(2-nitro-4-trifluoro-methylbenzoyl)-1,3-cyclohexanedione; qPCR, quantitative PCR; RSV, Rous sarcoma virus; rDNA, ribosomal DNA.

Figure 2

Assessment of integration frequency by nodule analysis. (a) Experimental design. (b) Representative FAH staining (brown color) of liver specimens after AAV infusion after 25 days of selection. Left two panels: ×50 magnification. Right two panels: ×200 magnification. Sections from three separate animals are shown in each column. The vector used is indicated above. The increased nodule frequency obtained with the rDNA vector is clear. (c) Quantitative analysis of nodule frequency. AAV, adeno-associated viral vector; FAH, fumarylacetoacetate hydrolase; rDNA, ribosomal DNA.

Figure 3

Site-specific integration of AAV-rDNA-hFah. (a) AAV-rDNA-hFah–mediated site-specific rDNA integration was detected by different sets of junction PCRs (J-PCR1, J-PCR2, J-PCR3, J-PCR4) in primary infected and secondary transplanted liver as well as neonatal mouse liver. One primer was located within the Fah transgene (junction PCR reverse in Figure 1a) and one within rDNA adjacent to the homology (junction PCR forward, Figure 1a). A PCR for the human Fah transgene (hFah PCR) served as positive control for presence of the vector. J-PCRs 1, 2, a nd 4 were positive in multiple samples of AAV-rDNA-hFah–injected liver. The far left panel shows that junction PCRs were negative when genomic DNA was mixed with unintegrated pAAV-rDNA-hFah plasmid. (b) Site-specific integration was detected in liver, kidney, lung, heart, and muscle, as early as 3 days postinjection. (c) Alignment of rescued junction fragment sequences with genomic rDNA sequence. AAV, adeno-associated viral vector; gDNA, genomic DNA; rDNA, ribosomal DNA.

Figure 4

Analysis of integrated AAV-rDNA-hFah. (a) hFAH southern blot of liver DNA after selection for AAV-rDNA-hFah integration and full repopulation (>90%). Each lane is from one mouse from either primary repopulation after virus injection (lanes 6–16) or from serial transplant recipients of virally rescued livers (lanes 17–20). The Fah+ hepatocytes from serial transplantation have undergone at least 16 rounds of replication. Both AAV2 and AAV8 were used. A dilution series of linear 7.0 kb plasmid pAAV-rDNA-hFah (0.1–1.5 copies per diploid genome) was used for quantitation (left lanes).Location of probe and EcoR1 restriction sites shown in Figure 1a. (b) Quantification of integrated vector copy numbers by phosphorimager readings. (c) Nested junction PCR on 0.3 ng of genomic DNA/sample from liver fully repopulated with AAV-rDNA-hFah–transduced hepatocytes. Lanes 2–33: individual 0.3 ng samples. Lane 34: water control. Lanes 1 and 35: size markers. AAV, adeno-associated viral vector; FAH, fumarylacetoacetate hydrolase; gDNA, genomic DNA; rDNA, ribosomal DNA; vg, vector genome.

Figure 5

Analysis of AAV-rDNA integration sites in the rDNA region. (a) Schematic of four integration breakpoints (from Supplementary Table S1) on the AAV-rDNA vector. The AAV-rDNA vector genome is 4,354 nucleotides in length. The terminal 145 nucleotides (positions 1–145 and 4110–4254) represent inverted terminal repeats (ITRs). (b) Location of integration breakpoints on rDNA relative to the homologous sequence (position on rDNA: 13412–15451) present within the AAV-rDNA vector. Position is based on rDNA sequence GenBank: X82564.1. (c) Junction sites sequence alignments with the AAV-rDNA vector and rDNA locus. AAV, adeno-associated viral vector; rDNA, ribosomal DNA.

Figure 6

AAV-rDNA-hFIX demonstrates improved transgene persistence. (a) Structure of the AAV-rDNA-hFIX vector. (b) Copy number changes of the hFIX transgene after liver regeneration by qPCR. The Maai copy number served as an internal control. (c) hFIX serum levels after partial hepatectomy and two sublethal CCl₄ treatments. About 30% of the expression persisted. AAV, adeno-associated viral vector; hFIX, human factor IX; qPCR, quantitative PCR; rDNA, ribosomal DNA.