Publication
The causes of sporadic Alzheimer's Disease (AD), the most common form of dementia at old age, are still unknown. Familial early onset AD (FAD) is mainly caused by mutations of Presenilins (PS), the active subunits of γ-secretase. Processing of its substrate proteins by γ-secretase generates shorter fragments. In the case of the Amyloid Precursor Protein (APP) these are Aβ species, of which the ratio of Aβ40, which is 40 amino acids (aa) long and Aβ42, 42 aa in length respectively, has been connected with their aggregation and AD plaque formation. Another cleavage product of many γ-secretase substrate proteins are intracellular domains (ICDs), which enter the nucleus and alter gene transcription. Be it via statins, the levels of cAMP, the generation of which is also catalyzed by cilia specific Adenylyl-cyclase III (ACIII) or Wnt pathways, cilia can be connected with various AD relevant proteins, pathways and functions, but their physiological and thus pathological cerebral roles are not understood. We hypothesized that γ-secretase activity alterations impact transcription levels of specific genes and potentially signaling and functional pathways as well as cilia genes in ways relevant to AD. We determined the transcriptional impact of enhanced γ-secretase activity in Chinese hamster ovary (CHO) cells with a cDNA microarray experiment. Gene ontology cluster analysis of the 1981 genes reported by our microarray gave insights into functional groups represented. Matching of known physical protein-protein interactions as reported by the string 8.0 database for 21 genes, which were confirmed to be transcriptionally altered with γ-secretase activity, registered various members of three different Wnt pathways. Consistent with an hypothesized role of cilia, the array also reported two dozen of cilia-related genes. The most decreased transcript levels were of Anti-mitotic exit network homolog 1 (AMN1), a gene of predicted expression in the cilia of sensory olfactory neurons and for which protein data was available in yeast only. Our immunohistochemistry results confirm Amn1 protein expression in mammalian cells, specifically in murine and human brains. We showed that Amn1 protein is localized in cortical neurons and in sharply outlined intracellular structures close to motile cilia and the opposing cell pole of monocilia. Aged APPPS1 transgenic mice, which over express human APP and PS1 with FAD mutations, did neither show differences in lateral ventricular cilia range, cilia density, the percentage of ciliated hippocampal neurons nor ACIII protein levels. Consistently with previous reports of increased cAMP levels in the cerebrospinal fluid of AD patients, we detected significantly increased levels of ACIII protein in human AD cortex samples (AD n=11, non demented n=12). These results support a new focus on the role of cilia-related proteins in AD. They also suggest some limitations of the APPPS1 transgenic animal model. In recent years the impact and role