Ribosomal DNA integrating rAAV-rDNA vectors allow for stable transgene expression

Abstract

Although recombinant adeno-associated virus (rAAV) vectors are proving to be efficacious in clinical trials, the episomal character of the delivered transgene restricts their effectiveness to use in quiescent tissues, and may not provide lifelong expression. In contrast, integrating vectors enhance the risk of insertional mutagenesis. In an attempt to overcome both of these limitations, we created new rAAV-rDNA vectors, with an expression cassette flanked by ribosomal DNA (rDNA) sequences capable of homologous recombination into genomic rDNA. We show that after in vivo delivery the rAAV-rDNA vectors integrated into the genomic rDNA locus 8-13 times more frequently than control vectors, providing an estimate that 23-39% of the integrations were specific to the rDNA locus. Moreover, a rAAV-rDNA vector containing a human factor IX (hFIX) expression cassette resulted in sustained therapeutic levels of serum hFIX even after repeated manipulations to induce liver regeneration. Because of the relative safety of integration in the rDNA locus, these vectors expand the usage of rAAV for therapeutics requiring long-term gene transfer into dividing cells.

Figure 1

Experimental design and preliminary data. (a) Schematic representation of the organization of the rRNA genes in eukaryotes. (b) rAAV-rDNA-hFIX and rAAV-stuffer-hFIX vector design. The white boxes represent the 28S ribosomal DNA (rDNA) and the corresponding “stuffer” sequences. The gray boxes represent hFIX expression cassette (TTR promoter, hFIX cDNA+1st intron, and bovine growth hormone pA). (c) hFIX enzyme-linked immunosorbent assay (ELISA) data throughout the course of the study. Solid lines—data for animals that did not undergo liver regeneration. Dotted lines—data for animals that underwent partial hepatectomy and CCl₄ treatment. Time points of partial hepatectomy and CCl₄ treatments are indicated. The expected form of rAAV that expresses transgene at different times during the study is indicated in the lower panel. Error bars represent SD n = 3. (d) Quantification of hFIX levels at the end of the study (week 29). Error bars represent SD n = 3. (e) Linear amplification-mediated-PCR (LAM–PCR) design. The biotinylated primer used for primer extension step is shown as arrow with a star. Long arrows show two possible amplicons—internal noninformative amplicon and external informative amplicon-containing genomic DNA.

Figure 2

In vivo comparison of human factor IX (hFIX) levels in animals treated with 500, 750, or 1,000 bp ribosomal DNA (rDNA) or stuffer vectors. (a) Upper panel: Design of rAAV-rDNA vectors in respect to genomic rDNA sequence; Lower panel: Graphic representation of the six vectors used in the study. The gray box represents hFIX expression cassette (TTR promoter, hFIX cDNA+1st intron, and bovine growth hormone pA). The size of each flank is indicated. (b–d) hFIX levels throughout the study in animals treated with rAAV-rDNA/stuffer-hFIX (b) 500 bp, (c) 750 bp, or (d) 1,000 bp. Data for animals treated with rAAV-rDNA vectors is shown in gray, and data for animals treated with rAAV-stuffer is shown in black. Dotted lines represent data for animals that underwent partial hepatectomy (PH) and CCl₄ treatments. Error bars represent SD n = 5–10. (e) Quantification of data at the last time point shown in (b–d). hFIX levels at the last time point before PH were used for normalization and thus were assigned a value of 100%. Shaded bars represent hFIX levels on day 190 in animals that did not undergo PH/CCl₄. White bars represent hFIX levels on day 190 in animals post PH/CCl₄ normalized to animals followed for the same period of time that did not undergo liver regeneration. Error bars represent SD, n = 5–10.

Figure 3

Vector copy number and vector genome structure analysis. (a) Vector copy number (VCN) obtained from Southern blot analysis of liver samples harvested during partial hepatectomy (PH) on day 86. Black bars indicate the average value for the given group ± SD n = 3–5. (b,c) Southern blot analysis of vector genome structure in liver samples obtained at the end of the study from (b) non–liver-regeneration control or (c) post-PH/CCl₄ animals injected with rAAV-rDNA or rAAV-stuffer vectors. Genomic DNA was digested with zero-, single-, or double-cutter endonucleases. Digested plasmid DNA mixed with digested control genomic DNA was used as standards (1,533 bp). Plasmid DNA was diluted to equal 40, 20, 5, or 1 vector copy per diploid genome in (b) or 20, 5, 1, 05 vector copies per diploid genome in (c). Bands corresponding to head-to-head (H-H) and head-to-tail (H-T) orientations of the rAAV genomes in the concatemers are indicated. Other observed rAAV genome structures are indicated.

Figure 4

Recombinant adeno-associated virus (rAAV)-ribosomal DNA (rDNA) versus rAAV-stuffer comparison in NIH3T3 cells. (a) Graphic representation of vectors used in this part of the study. Vectors encoded hPGK-GFP-pA cassette flanked by 1,000 bp rDNA or “stuffer” control DNA. Fragment of human factor IX (hFIX) cDNA was cloned downstream of pA to keep the overall size of the vectors identical to the 1,000 bp rAAV-rDNA/stuffer vectors used in the in vivo studies (see Figure 2a). The black arrow indicates the biotinylated oligo used during primer extension step of linear amplification-mediated-PCR (LAM–PCR). (b) Comparison of different conditions used to transduce the NIH3T3 cells. Cells were transduced at two different cell numbers (as indicated) and different multiplicities of infection (MOIs). The resulting green fluorescent protein (GFP) levels at the first time point (day 2 post transduction) are indicated. (c) GFP levels in cells transduced with recombinant adeno-associated virus (rAAV)-rDNA and rAAV-stuffer vectors throughout the course of the study. (d) Quantification of GFP levels at the last time point (day 43) of the study. (e) LAM-PCR data. DNA samples extracted from cells on day 43 were analyzed by LAM-PCR. See legend for Table 1 for detailed explanation.