One of the most important hallmarks of cancer is that the DNA of the cancerous cell has been altered from its 'normal' state. Although a number of limited models have been put forth, in most oncogenic processes it is still not clear what the key event is that causes a cell to become oncogenic, nor how the DNA alteration has arisen (the Where and How). Our work on a set of enzymes, known as DNA deaminases, has provided some novel insights into these processes. In the past, we have been able to show that the mutagenic potential of DNA deaminases (especially AID), can be directly linked to the hormone oestrogen. Enhanced amount of oestrogen in a cell leads to enhanced oncogenic translocations. Because DNA deaminases act directly on DNA and mutate cytosine bases to uracil, the How of AID dependent cancer induction is understood. Our current work now focuses on the Where.
AID (activation induced deaminase) is an enzyme that induces the somatic alterations of the antibody gene loci. Once cytosine is deaminated to uracil in single stranded DNA, various DNA repair proteins are 'hijacked' to the uracil lesion, leading to DNA point mutations or DNA recombination. Recent results have shown that AID is not limited in its targeting to the antibody loci, but its activity can also be found near a number of proto-oncogenes as well as tumour suppressor genes. When AID is targeted to an artificial target, it can induce DNA instability at that site, but most often does not cause DNA point mutations. This is indicating that although DNA deaminases have a high potential of being oncogenic, the majority of their activity is controlled by proper DNA repair.
AID and transcription elongation
In an attempt to understand how AID may be targeted to the antibody gene, we have undertaken a novel approach in tagging AID protein in vertebrate cell lines. By using homologous recombination, we have tagged endogenous AID with various short peptide tags. Using highly specific antibodies for those tags, we can isolate native AID from cell lines in larger quantities, without over-expressing the protein. Using Mass-Spec analysis we have been able to identify 32 unique protein-peptides that were found in a complex with AID, but only when AID was bound to the chromatin fraction of a cell. Using a 'complex association algorithm', we were able to place almost all peptides within a network of interacting proteins. Importantly, 8 of the peptides (and the 6 of the top 8) were part of the transcription elongation complex (PAF/Spt/FACT). Since it has been known for over a decade that transcription plays an important role during AID induced antibody diversification, identifying this particular complex is reassuring that we have identified a genuine complex. Furthermore, the mutation profile at the antibody loci is such that mutations due to AID start 100-200 bp after the promoter initiation. This distance is equivalent to the distance of where an RNA pol II complex switches from an initiator to elongator complex. Our current work is to establish the validity of each of these members in the AID/chromatin complex.
AID and cellular transcripts
In a parallel study on understanding how AID can initiate cancer inducing mutations, we have discovered that AID expression induces changes in the transcriptome of a cell. Two identical cell lines, one expressing AID one not, were analysed for their total RNA expression profile (transcriptome) on day 0. The same cell lines were grown for 6 weeks and then reanalysed for their transcriptome profile. Plotting Week 0 vs. Week 6 for either cell line showed that RNA transcripts that are either turned off or on as being off the diagonal (Figure 1A and B). It became clear that the cell line expressing AID induced more RNA transcript alterations (more off-diagonal events) then the one expressing no AID. This data was verified by RNAseq (deep-sequencing the total RNA of each cell line at each time point). Importantly, bioinformatic analysis showed that the global effect on transcription was predominantly due to changes at the 5¿ end of genes (near the transition of initiation to elongation - Figure 1C). This would be in accordance of AID influencing the RNA pol II elongation machinery, which could be due to its presence at RNA poll elongation sites on chromatin.
The parallel discovery of AID being part of a RNA pol II transcription complex as well as AID's ability to influence RNA transcription on a more global scale points to a mechanism were AID gains access to various sites of the genome to induce its DNA deamination events. As mentioned above in the majority of cases these events are properly repaired and do not lead to any genetic alterations. On the other hand, most DNA damage will induce alterations in chromatin configuration, as DNA repair pathways need access to the lesion. These alterations usually manifest themselves as histone marks, while alterations in histone marks are known to change transcriptional profiles. This would lead us to conclude that AID has the potential to (and possibly a function in) epigenetic alterations of the genome. AID's ability to alter the transcriptome may have a direct effect on the overall plasticity and maybe even pluripotency of ES cells (as reported previously, AID is highly expressed in pluripotency cells such as oocytes and ES cells).
Outlook
Our current and future work is to identify the direct interaction partners of AID on chromatin and how AID can influence the activity of that complex. For the biochemical and enzymological analysis we will collaborate with Jesper Svjestrup's lab, while using our unique cellular tools to identify the cellular and genetic interactions. Further deep sequence analysis from our cell lines will confirm AID's role in transcription and possibly point towards particular sequence elements within the genome that can help us identify more precisely AID targets and the target mechanism.
Figure 1. Transcriptome analysis. A. Comparing total RNA microarray data of week 0 to week 6 of AID expressing cell lines. B. Same as A but the cell line is not expressing AID. C. RNA deep sequence analysis of differential expression samples A and B, identifying the 5¿ end of genes as the region of differences between the two sets of data..
For a list of refereed research papers, see Publications (in navigation on left).
