Role of AMBRA1 in nervous tissue homeostasis and in neurodegeneration

Abstract

My PhD project focused on the role of AMBRA1 in mouse brain in physiological and pathological conditions. Ambra1 (“Activating Molecule in Beclin 1 Regulated-Autophagy”) is a recently identified gene encoding for a large protein (around 130 kDa), with an N-terminal WD40 domain (Fimia et al., 2007). Ambra1 gene is highly conserved among vertebrates, and is expressed in different splicing isoforms. In the developing mouse, AMBRA1 protein shows abundant expression in the central and peripheral nervous system. At embryonic day 8.5 (E8.5), AMBRA1 is present in the neuroepithelium, while at E11.5 protein expression is localized to the ventral-most part of the spinal cord, the encephalic vesicles, the neural retina, and the dorsal root ganglia. This distribution pattern suggests that AMBRA1 is centrally involved in the proper development of the nervous system. Indeed, Ambra1gt/gt mice, obtained by gene trapping technique, and displaying a deficient expression of the protein, show early and severe neuropathological features, including neuroepithelial hyperplasia and defective neural tube closure, leading to embryonic death (Fimia et al., 2007; Cecconi et al., 2008). These disturbances appear to derive from dysfunctions in the autophagic process, also resulting in an imbalance of apoptotic cell death and cell proliferation. These data, together with in vitro evidence, allowed proving a regulatory role of AMBRA1 in autophagy activation in vertebrates (Fimia et a., 2007). Autophagy is involved in the intracellular turnover of proteins and cell organelles, and has an important role in regulating cell fate in response to stress (Levine, 2005). Three types of autophagy have been described: microautophagy, chaperone-mediated autophagy and macroautophagy. The last one is a bulk degradation pathway and the only intracellular mechanism potentially capable of degrading large protein aggregates or damaged organelles. A cup-shaped isolation membrane forms around cytosolic components, eventually fusing to form a double membrane bound vesicle, the so-called “autophagic vacuole” (AV) (Mizushima et al., 2002). In thi respect AMBRA1 is essential in the induction of AV formation (Di Bartolomeo et al, 2010). Recent studies show that macroautophagy is constitutively active in healthy neurons and is vital to cell survival (Bolland and Nixon, 2006). Mice lacking either Atg5 or Atg7 genes exhibit motor and behavioral deficits as well as degeneration of specific neuronal subtypes (Hara et al., 2006; Komatsu et al., 2006). Diffuse protein aggregates appear in surviving neurons within several brain regions, culminating in the formation of toxic inclusion bodies. This supports an essential role of autophagy is for neuronal health. The evidence so far presented and the severe neuropathological phenotype of Ambra1gt/gt mice prompted us to investigate the expression of AMBRA1 protein in adult mouse CNS, in physiological and pathological conditions. The main results obtained from the research and described in this thesis can be summarized as follows: 1. I provided the first neuroanatomical/histological/ ultrastructural map of the distribution of AMBRA1 expression in mouse brain taking advantage of immunohistochemical, immunofluorescence and immunoelectron microscopy approaches. Wide presence of the protein in the forebrain, midbrain, and hindbrain was observed, demonstrating prevalent expression in neurons, even though astrocytes and microglial cells also contain moderate levels of AMBRA1. This suggests that in physiological conditions AMBRA1 is crucial for neural tissue homeostasis. Ultrastructural analysis revealing association of with the endoplasmic reticulum strongly supports AMBRA1 activity in regulating basal autophagy, essential for neuronal survival. Detailed examination of different brain territories along the rostro-caudal axis allowed me to show that AMBRA1 content varies among brain regions, neuronal populations, and subtypes. The concentration of neuronal AMBRA1 appears at least partially related to cell volume. Neurons featuring a large soma, highly brached dendrites and long axons generally display higher immunoreactivity, compared to smaller cells. Among these, mitral cells in the olfactory bulb, giant pyramidal neurons of the neocortex, motor neurons of the brainstem, Purkinje cells of the cerebellar cortex are paradigmatic. Some of these giant neurons have also been described to contain other pro-autophagic molecules (Tamura et al., 2010), suggesting that their high metabolic and turnover rates are likely to involve large amounts of autophagy regulators, such as AMBRA1. Even more interestingly, we also found some correlation between AMBRA1 content and other parameters, namely susceptibility to damage. In particular, some neuronal subsets, which are selectively spared by neurodegenerative insults, demonstrably contain high levels of the protein. Representative examples of this concept are cholinergic interneurons in the corpus striatum, virtually unaffected by Huntington’s Disease (HD), and CA3 pyramidal neurons of the hippocampus, resistant to Alzheimer’s Disease (AD) and to ischemic injury. Importantly, this pathology is suggested to involve inefficient regulation of autophagic processes (Jaeger et al., 2009)). However, in some cases, neurons showing relatively high AMBRA1 content are target of specific neurodegenerative disease. For instance, Purkinje and mitral cells are both affected by the so-called “Purkinje cell death” (pcd) pathology. Remarkably, in this syndrome neuronal cell loss which probably occurs through the autophagic, rather than apoptotic, pathway, indicating that dysregulation of autophagy, possibly involving also AMBRA1 expression, may participate to some neuropathologies. 2. I highlighted important age-related variations in AMBRA1 expression in regions prone to neurodegeneration, such as the neocortex and hippocampus during normal and Alzheimer-like ageing. For this part of the project, I used a WT and transgenic strain for AD (Tg2576) at different time points. In the neocortex, the high levels of AMBRA1 in normal young mouse led me to hypothesize that AMBRA1 is essential for a faultless maturation of neocortical neurons. Interestingly, Tg2576 mice show lower AMBRA1 content at the very onset of disease, when most of the histopathological hallmarks are still undetectable, thus suggesting the possible involvement of AMBRA1 down-regulation in the impaired plasticity of neuronal circuits. During normal adulthood, we found an overall decrease of AMBRA1 content which may reflect a relatively balanced cell activity, with a stable regulation of biosynthesis and degradation pathways. The novel increase of the protein in the aged neocortex, which is independent of the genotype, may instead suggest that AMBRA1 is up-regulated in this critical period, since brain ageing likely requires a higher rate of basal autophagy also to counteract oxidative stress and consequent accumulation of damaged organelles. Differences in AMBRA1 expression are not only related to age, but even to the specific brain region, since the the neocortex and the hippocampus show different behavior during ageing and pathological progression. In particular, overall levels in young mouse hippocampal formation are similar in the two genotypes. However, a different distribution in hippocampal subregions of Tg2576 animals compared to WT was detected. In fact, the neurogenic region of dentate gyrus is more intensely AMBRA1 immunoreactive in the transgenic animals than in WT, thus suggesting that in the pathological condition an early response involving enhanced neurogenesis may occur to cope the first deficits. Then we found a decrease in AMBRA1 level during adulthood and ageing in both genotypes, even though, at 18 months, the transgenic hippocampus shows a further drop in AMBRA1 reactivity, possibly contributing to neurodegeneration. 3. I showed that AMBRA1 expression can be induced in dopaminergic neurons by a mild 1-Methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) treatment, producing a pharmacological model of early PD. Importantly, similar effects are also observed in a genetic model of genomic instability (ERCC1 mutants) which is associated with an early PD-like phenotype. These findings support the involvement of AMBRA1 in cellular response against oxidative imbalance in dopaminergic neurons that are closely related to PD onset. Surprisingly, AMBRA1 levels increase in both the substantia nigra (SN) and ventral tegmental area (VTA), the former of which turns to cell death in PD while the latter is spared. One could envision that the activation of the autophagic pathway by AMBRA1 up-regulation, can be defensive in some cases and detrimental in others, strictly depending on the specific neuronal population. To this respect, it is worth noting that the features of neuronal cell death in PD are still debated, because it seems to bring back to autophagic cell death. In conclusion, the results obtained in my PhD project suggest that AMBRA1 is a fundamental molecule in the nervous tissue, given the abundant content in neurons, although to different degree with respect to the brain area and neuronal population. In addition, modulation of AMBRA1 expression may be critical for establishing, promoting or counteracting neurodegenerating processes. Thus, our study opens the way to further investigations aimed to defining the precise contribution of AMBRA1 to nervous tissue development, homeostasis and response to acute or chronic injury.