Mammalian artificial chromosomes as vectors: progress and prospects

Artificial chromosomes have long been touted as the ideal vector for gene therapy and biotechnology purposes based on the idea that such a chromosome would mimic the natural state of DNA in the cell. This, it is argued, would mean that essentially unlimited amounts of DNA could be incorporated into such a vector enabling either large genes or whole metabolic pathways to be provided to the recipient cell or organism. Additionally, such a vector would not integrate into the genome of the host cell and so would not cause mutagenesis by insertion and could perhaps be withdrawn from the cell or organism when no longer required. A number of preconditions are implicit in these claims. First, the chromosome should have a segregation efficiency approaching 100% in order to be useful in a cell population undergoing multiple rounds of cell divisions. Second, the chromosome should have a defined structure for regulatory and practical reasons. A defined structure is needed to maximize the control of expression of the genes that it contains. Third, the chromosome should not be so large that delivery becomes a problem. Finally, chromosomal effects such as centromeric or telomeric silencing should not dominate the expression of genes contained in an artificial chromosome. In this article, we discuss our own and others' efforts to achieve these aims using a variety of nonviral approaches to the problem.