LINE-1 copy number variation in Alzheimer's disease

Transposable Elements (TEs) are a class of repetitive DNA sequences able to mobilize and change location in the genome. They make up almost 50% of the human genome and slightly less in the mouse (Lander et al., 2001; Waterston et al., 2002). The autonomous non-LTR retrotransposon Long Interspersed Nuclear Element-1 (L1) is for sure the most impactful and still active TE in the human and mouse genome (Richardson et al., 2014). Recently, it was shown that L1s are active in the mouse, rat and human neural progenitor cells (NPCs), and are more abundant in the genome of cells of the human hippocampus, than in other non-nervous tissues (Thomas and Muotri, 2012; Coufal et al., 2009; Reilly et al., 2013). Besides human diseases caused by the direct effect of L1 insertion and few neurological diseases that were demonstrated to misregulate L1 retrotransposition, it is still unknown whether L1 retrotransposition can directly cause neurological disorders (Muotri et al., 2010; Coufal et al., 2011; Thomas et al., 2012; Erwin et al., 2014). Alzheimer’s disease is the main cause of dementia in the elderly, affecting almost two third of the world population over 65 (Bettens et al., 2013). It has been recently demonstrated that AD patients suffer from vitamin B12 deficiency and high homocystein content in blood, which contribute to the disregulation of S-adenosylmethionine synthesis and DNA methylation (Scarpa et al., 2006). Since these DNA epigenetic alterations observed in AD patients could have an impact on L1 retrotransposition, we decided to investigate the Copy Number Variation (CNV) of L1 sequences in the genome of different tissues from AD patients and healthy controls (from three different cohorts), and in a mouse model of the disease. By using the qPCR with Taqman probes, we observed in some of the tissues that we analyzed, a decreased amount of full length L1 sequences in AD patients compared to controls. Assuming that in AD patients there is a lower degree of DNA methylation, we can speculate that a higher retrotransposition of L1 elements occurred in certain neurons, may have caused the death of these cells, eventually leading to the detection of a lower amount of L1 sequences in AD patients. We then developed new Taqman assays to study L1 CNV in the TgCRND8 mouse at P0 and at two stages of the adulthood (3 and 8 months). We observed in transgenic mice a higher amount of L1 sequences in the cortex at P0, in the hippocampus at 3 months, while no difference were detectable at 8 months of age, supporting the hypothesis that in AD there is a higher L1 retrotransposition that causes cell death. We also set up a technique called SPAM (SPlinkerette Analysis of Mobile elements), aimed at identifying the insertion sites of L1 sequences in the human genome. After a first test of the protocol, we used this technique to compare L1 insertion profile of AD patients and controls. Interestingly, we found that AD susceptibility genes present several novel insertion sites that would warrant future investigation. Finally we adapted the SPAM technique to a different repetitive element, the roo element, in a different organism: the Drosophila melanogaster, demonstrating that SPAM technique is a scalable approach, suitable for the integration sites discovery of different kinds of repetitive elements in different organisms.