Signal transduction pathways activated by CpG-DNA.

While more and more attention has been paid to CpG-DNA with respect to its usefulness as an adjuvant, its molecular mechanism of action is less well defined. Over the last few years, at least two major signalling pathways have been shown to be utilized by CpG-DNA: the NF-kappa B activation pathway and the stress-kinase pathway. Direct downstream events of these pathways are induction of transcriptional activity of NF-kappa B and transcriptional activity of AP-1. As far as investigated, CpG-DNA uses signal transduction pathways originally described for other stimuli, such as LPS, IL-1 or TNF. Therefore, to us, the prime question is: where does CpG-DNA-induced signalling enter these known pathways? This raises questions about the existence of a CpG-DNA-sequence-specific receptor. Several points of evidence support the probability of the existence of a cellular receptor: There is a strong motif (unmethylated CpG) dependency for CpG-DNA-induced signalling. There is cell-type specificity. Dendritic cells, macrophages and B cells respond to CpG-DNA, but other cell types, such as fibroblasts and T cells, do not. In addition, classic signal-transduction pathways are rapidly switched on in a parallel manner, as is known for other receptors. Using competing non-CpG ODNs and inhibitors of endosomal acidification, some evidence has been obtained that CpG ODNs are taken up into endosomes by a CpG-independent receptor, followed by a pH-dependent step before signalling starts. A model based on these findings is proposed in Fig. 4. Nevertheless, other receptor-independent activities of CpG-DNA cannot yet be ruled out. Although unlikely, we should consider the possibility that CpG-DNA directly interacts with cellular nucleic acids either by direct hybridization with complementary nucleotides or by formation of DNA triplexes (VASQUEZ and WILSON 1998). While these possibilities have been explored by antisense technology, using a huge variety of ODNs, there is no experimental evidence that such interactions are important for the activity of CpG-DNA. In this context, it is important to note that DNA, especially phosphothioate-stabilized ODNs with poly-G stretches, have substantial CpG-independent activities, although these activities seem not to depend on specific, antisense-like DNA-DNA interactions (PISETSKY 1996). One good example comes from experiments using ODNs on primary T cells. Co-stimulation of CD3-primed T cells with CpG ODN leads to a significant increase of IL-2 secretion and proliferation; however, these effects are CpG independent (K. Heeg, personal communication). Remarkably, these poly-G stretches seem to be inactive when transferred to double-stranded DNAs, such as plasmid DNA (WLOCH et al. 1998). In contrast, to my knowledge, no immune-stimulatory effect of bacterial DNA has been described that can not be abolished by CpG-specific methylation. Taken together, CpG-dependent and CpG-independent activities must be distinguished from one another. Among these effects, CpG-dependent signalling is better defined. Much effort is going into the investigation of the pharmacological applications of CpG-DNA. Once CpG-receptor-like structures are known, the question of the physiological role of CpG-DNA can be tackled.