Neurons often have a complex dendritic tree and an extensive axonal organization, which requires specific properties from the neuronal cytoskeleton to enable structural maintenance and stability. In contrast to the rather rigid nature of postmitotic neurons, during various dynamic processes like neuronal birth, migration and growth, rather a flexible and adaptative cytoskeleton is required. These reorganizations of the cytoskeleton of immature neurons are regulated by binding of a.o. microtubule associated proteins (MAPs) to cytoskeletal elements. However, also plastic changes in mature neurons, like alterations in synaptic

connectivity, might be regulated by MAPs.

MAPs are an important class of structural proteins characterized by the presence of one or more microtubule (MT) binding sequence(s). Upon MAP binding, MTs are stabilized and tubulin polymerization is enhanced. The affinity of a MAP for MTs is further regulated through alternative splicing and phosphorylation. Many MAPs are exclusively expressed in the nervous system where they are involved in growth as well as structural stability of neurons. Examples include tau, but also novel members, like the doublecortin (DCX) and doublecortin-like kinase (DCLK) gene products, which are important a.o., in brain development.

Aim: In the present thesis we have tried to establish a role for the MAPs tau and DCL in structural plasticity during development and adulthood, including mitosis and radial migration but also in synaptic plasticity, LTP and learning and memory.

In this thesis, we have focused on two MAPs: 1) the DCLK splice variant doublecortin-like (DCL) that was studied in relation to early cortical development and radial migration, and 2) on (mutant) tau in relation to structural and synaptic plasticity and memory function. Also, we studied structural plasticity changes in the human brain in presenile Alzheimer cases. Doublecortin (DCX) is a recently discovered MAP that is abundantly expressed in a.o., the embryonic cortex. It is involved in the migration of newborn cells from the ventricular zone (VZ) to their final destination in the cortical plate. Mutations in the DCX gene severely disturb neuronal migration and induce a doublecortex syndrome in humans. Recently, a related gene called doublecortin-camkinase-like (DCLK) has been discovered. One of its splice variants doublecortin-like (DCL) was shown to be involved in mitosis and radial fiber stability of neuronal precursors. These latter studies are presented in the addendum of this thesis.

Although the DCX and DCLK genes share at least partly overlapping

functions, specific differences exist as well, e.g. in their spatiotemporal expression patterns. DCX e.g., in contrast to DCLK, does not appear to be involved in cell birth. To further compare different aspects of DCX and DCL, we performed a detailed spatiotemporal analysis of the expression of both proteins throughout embryonic development. In Chapter 2 we show that DCL is already expressed from E9 onwards and decreases after E13. DCX expression on the other hand, starts modestly at E11 and is still present at E17. Moreover, DCL was expressed in the VZ, the intermediate zone and CP, whereas DCX is not expressed in the VZ. Before E12, DCL and DCX expression do not overlap, indicating clearly different roles in the important early phase of cortical development when precursor expansion is extensive. Finally, DCL is found to be specifially associated with mitotic cells and with vimentin positive radial glia cells and radial fibers in the VZ. This suggests that DCL could be involved in early neurogenesis.

Of the classical MAPs, especially protein tau has been extensively studied because of its involvement in the tangle pathology in dementia. Upon alternative splicing six different tau isoforms can be formed, containing either 3 or 4 repeats (tau-3R or tau-4R). In rodents, tau-3R is primarily expressed during development, whereas expression switches to mainly tau- 4R from the second postnatal week onwards. This period largely coincides with a phase of extensive neurogenesis during the formation of the hippocampal dentate gyrus (DG) that is completed around week 3, after which neurogenesis is strongly reduced.

In Chapter 3 we addressed the role of tau in DG development and more specifically focused on the 3R to 4R switch around the first 2 weeks of life, and on the possible functional consequences for hippocampal development. Therefore, the effect of tau deletion was tested (by others) in primary hippocampal cultures. Deletion of the tau gene increased cell birth, decreased differentiation and decreased neuritic outgrowth. All these effects could be reversed by expression of tau-4R, whereas tau-3R could only partially reverse neuritic outgrowth but failed to affect cell birth or differentiation. To test the relevance of these tau functions for hippocampal development, we subsequently made use of a transgenic mouse tau knockout, human 4R tau knock-in (KOKI) mouse model. These mice lack all mouse tau isoforms, whereas tau-4R is expressed from the second postnatal week onwards at reduced levels in the hippocampus. These mice show a transient increase in cell birth from the second postnatal week. At two months of age, cell birth was again reduced to levels comparable to those in wild types. This increased cell birth was reflected by increased DG neurogenesis and eventually by an increased number and volume of the adult hippocampus. However, the size of the individual dendritic tree of DG granular neurons was significantly reduced at two months of age, suggesting a more immature population of cells. Only in the developing wildtype hippocampus, tau-4R has an inhibitory effect on neurogenesis, whereas in adult mice, tau-4R exerts stimulatory effects on neuritic outgrowth. We subsequently tested the consequences of these structural alterations for hippocampal functioning and found that memory function was significantly improved in KOKI mice in an object recognition task (ORT), but this was not associated with altered LTP. MT affinity of tau can also be influenced by phosphorylation as this reduces MT binding. Phosphorylation is high during development, where it is associated with cell birth, but low in adulthood. However, in Alzheimer's disease (AD) and frontotemporal dementia (FTD), a familial form of dementia often caused by tau mutations, tau is again heavily phosphorylated. Hyperphosphorylation of tau is thought to be relevant for memory impairments in AD and FTD.

To test this, we studied in Chapter 4 memory function in tau transgenic mice bearing the FTD mutation P301L. These mice recapitulate many of the features of FTD including axonopathy and memory impairments that are paralleled by tau hyperphosphorylation at later ages. We hypothesized that memory was not affected before the onset of hyperphosphorylation and therefore studied these mice at young adult ages of 2 months. Surprisingly, ORT performance was improved in P301L mice, which was associated with increased LTP. No changes were noted in any morphological parameter like neurogenesis or individual structure of the dendritic tree. Thus, in mice bearing the P301L tau mutation, hippocampal functioning is not impaired, but rather improved before the onset of tau phosphorylation. These results show that tau can directly affect learning and memory. They also indicate 1) that tau plays an important beneficial role in normal processes underlying hippocampal memory, clearly beyond “merely” the control of cellular morphology, and 2) that not tau mutations per se, but rather the ensuing hyperphosphorylation must be critical for the cognitive decline in tauopathy.

We showed in Chapter 3 that tau affects neurogenesis during development. Albeit at a much lower pace, neurogenesis continues to occur also in the adult DG. Since insults or damage to the hippocampus causes compensatory responses in neurogenesis in rodent models, neurogenesis might be affected by the pathological manifestations occurring in Alzheimer's disease. Therefore, in Chapter 5, we studied proliferative changes in the AD affected hippocampus of a cohort of presenile cases. We showed that the expression of the proliferation marker Ki-67 was increased in AD, which was attributed to changes in glia-rich areas and cells associated with the vasculature but not to changes in neuron rich areas. Based on these results and on additional DCX, GFAP and VWF stainings, we conclude that increased proliferation in the AD hippocampus does not reflect neurogenesis but rather represent glial proliferation and vasculature- associated changes.

Previously, MAPs were always considered to solely act as microtubule stabilizers. Recent data including the results presented in this thesis now show that MAPs are involved in various different functions in the nervous system, including migration, mitosis, structural and synaptic plasticity as well as memory function. Although many of these MAP functions might be conserved amongst different proteins, differences in expression patterns, number of microtubule binding sites, projection domains and phosphorylation sites between MAPs and MAP isoforms are likely to represent functional differences. Indeed in this thesis we have shown a functional difference between tau-3R and tau-4R, matching the switch in their developmental expression. Furthermore, the differential spatiotemporaln expression of DCX and DCL matches their different functions. Together, the different roles of these MAPs provide a better insight in their respective roles during early development.