Minding the Cs and Gs

Abstract

e t n a m f r t t Successful gene therapy requires the introduction of foreign DNA into host cells. Until recently, most investigators working in this field did not realize that the DNA sequences themselves contained immune stimulatory sequences, called CpG motifs, that activate immune responses as an unintended consequence of gene delivery. DNA containing CpG motifs activates both the innate and the acquired arms of immune responses (Fig. 1). CpGinduced innate immunity is exemplified by the rapid activation of NK cells, macrophages, and dendritic cells, with the concomitant production of proinflammatory cytokines. This response is a cause of the acute toxicity that has been observed in some gene therapy protocols (1). Further, this activation of dendritic cells, as well as direct effects of CpG DNA on B cells (2), initiates acquired immune responses, including the generation of antibody and cytotoxic T cell responses against delivered antigens (Fig. 1). Although bacterial DNA has been known to have immune stimulatory effects for many years (3), it has only been a few years since the discovery that these immune stimulatory effects depend on the presence in bacterial DNA of unmethylated CpG dinucleotides in particular base contexts and that these effects can be reproduced by synthetic oligonucleotides containing these “CpG motifs” (2). It is now generally accepted that immune activation by CpG motifs evolved as an immune defense against infection. CpG motifs are relatively rare and are heavily methylated in our own DNA. Thus, our DNA has a subtly different molecular pattern from prokaryotic DNAs, and because it lacks these unmethylated CpG “danger signals,” it does not activate this defense. Because gene therapy vectors are unmethylated, they would normally “look” like pathogen DNA to this primitive innate immune defense and will, therefore, trigger it. Although the molecular mechanism of CpG-induced immune activation is still incompletely understood, it appears that the DNA must enter host cells and bind to an intracellular CpG pattern recognition receptor, which then leads to the activation of mitogen-activated protein kinases, NFkB, and other cell-signaling pathways (4). The immune effects of CpG dinucleotides depend critically on the flanking bases (Table 1). If the CG is preceded by a C and followed by a G, immune stimulation is almost abolished, and the activity of nearby stimulatory CpG motifs can be diminished (5). Because viral or nonviral gene therapy vectors contain hundreds of CpG motifs, which are responsible for at least some of the observed immunotoxicity, counter measures are urgently needed if the clinical development of gene therapy is to realize its full promise. Unfortunately, so