Adeno-associated virus preferentially transduces human compared to mouse synovium

There is increasing interest in adeno-associated virus (AAV) vectors for a wide variety of gene therapy applications. AAV is a nonpathogenic human parvovirus that can mediate long-term transduction of a number of cell types without provoking a significant immune response. These properties make AAV especially attractive for use in gene therapy of rheumatoid arthritis (RA), a chronic inflammatory disease. To investigate the potential of AAV in gene therapy of arthritis, the ability of AAV to infect synovium in vitro and in vivo was tested. Three human RA synovial fibroblast cell lines and two murine (one DBA/1J and one DBA1J × C3H F1) synovial fibroblast cell lines were used to test AAV transduction in vitro. The cell lines (2 × 10⁵ cells) were infected with 10⁴ particles/cell of a murine IL-10-encoding vector (AAV-mIL-10) alone or with the addition of a low titer (100 particles/cell) of an E1-, E3-deleted recombinant adenovirus to provide E4orf6 activity to enhance second-strand synthesis. The supernatants were harvested from the wells at various time points and assayed for mIL-10 expression by ELISA. Both human synovial cell lines infected with AAV alone demonstrated low-level transgene expression throughout the course of the study. However, by day 10, all human cultures coinfected with adenovirus showed a 16- to 56-fold increase in mIL-10 compared to cultures infected with AAV-mIL10 alone. By day 30, a 31- to 135-fold increase was observed. No such increase was observed in any of the mouse cell lines. To determine the AAV transduction efficiency for synovium in vivo, human RA synovial tissues obtained from patients undergoing joint-replacement surgery were implanted subcutaneously on the backs of NOD.CB17-Prkdc SCID mice. After allowing a 2-week period for engraftment, tissues were injected with 3.4 × 10¹¹ particles of AAV-luciferase alone or in combination with 1.0 × 10¹¹ particles of adenovirus. Two weeks following AAV administration, the tissues were homogenized and assayed for expression of luciferase. Only the tissues coinfected with adenovirus had luciferase levels above background. A similar experiment with AAV-LacZ demonstrated X-gal staining only of synovial tissues coinfected with adenovirus. These findings demonstrate a preferential ability of AAV to transduce human, compared to mouse, synovial tissue and suggest that second strand synthesis may be a limiting factor in gene transduction. Further studies to elucidate the mechanisms limiting gene transduction in human synovium may allow optimization of this vector for the treatment of arthritis.