Role of Abnormal Tau Proteolysis in Alzheimer's Disease Neuronal Degeneration

Alzheirner' s disease (AD) is characterized by two main histopathological hallmarks: extracellular plaques made of ~-amyloid protein (Selkoe, 1999) and intraneuronal neurofibrillary lesions (Goedert et al., 1996a). Neurofibrillary lesions are accumulations of non-membrane bound bundles of paired helical filaments (PHFs), sometimes interspersed with straight filaments. The sub-unit protein constitutive of PHFs is the microtubule associated protein tau, modified in such way to form insoluble polymers (Delacourte and Buee, 1997). As it is also shown in a preliminary study introducing this work, AD-tau electrophoretic profile is characterized by a "smear" in the high molecular weight range, corresponding to pathological aggregates. In AD, tau undergoes several post-translational modifications (Mandelkow and Mandelkow, 1998), including hyperphosphorylation, oxidation, ubiquitination, glycation and abnormal proteolysis. This thesis investigates the role of tau abnormal proteolysis in AD neuronal degeneration. The so-called "Truncation hypothesis" represents the conceptual basis of this work. Its original formulation is based on a pioneering work by M. Novak and collaborators. First, a pronase-resistant core of PHFs was shown to be made of a group of 12 Kda fragments of tau (Wischik et al., 1988a and b). These fragments were employed to raise the monoclonal antibody MN423 (Wischik et al., 1988a; Novak et al., 1989), which stains all the main neuropathological hallmarks of AD (Bondareff et al., 1991; Mena et al., 1991; Ugolini et al., 1997). Immunoelectronmicroscopy studies showed that MN423 decorates PHF isolated with or without pronase treatment (Novak et al., 1993). MN423 does not recognize full-length tau, but binds to the 12kDa fragments of core PHFs terminating at Glu-391 (Novak et al., 1993). On this basis, it was hypothesized that tau is endogenously runcated in AD brain and Glu-391, which invariably represents the last amino acid of the 12 Kda fragments, was defined as a "truncation point" (Novak et al., 1993; Novak, 1994). MN423 immunoreactivity is also found intracellularly in AD neurons, even in the absence of neurofibrillary lesions, in the form of pre-tangle intracytoplasmic granular accumulations (Bondareff et al., 1991; Mena et al., 1991; Ugolini et al., 1997). Therefore our working hypothesis was that truncation of tau might precede its assembly into PHFs, thus representing an early intracellular event in the development of neurofibrillary pathology. In a follow-up study our group showed that, although expression of the 12 Kda fragments in a cellular context does not lead to intracellular aggregation, these truncated forms of tau fail to support microtubule assembly in vitro (Fasulo et al., 1996). Moreover, they are unable either to bind to microtubules or to induce their bundling in vivo (Fasulo et al., 1996). A truncated form of tau, terminating at Glu-391 and encompassing the two praline-rich regions, is able to induce apoptosis upon expression in COS cells, raising the possibility that tau fragments present in neuropathological conditions may affect neuron viability more profoundly than it was previously thought (Fasulo et al., 1998). Extensive nyuronal degeneration and loss is observed in many neuropathological conditions, including AD. Evidence has recently accumulated, showing that poptotic cell death is involved in this process (Su et al., 1994; Cotman and Anderson, 1995; Lassmann et al., 1995; Anderson et al., 1996; Cotman and Su, 1996; Gervais et al., 1999). Here it is shown that, in AD cortex, truncated tau, as identified by MN423 immunohistochemistry, is co-localized with neuronal death, as revealed by DNA fragmentation (Ugolini et al., 1997). The second part of this work was devoted to look at the question of tau truncation and cell death from the opposite angle: do truncated forms of tau have any effect on cell viability? A set of tau fragments was identified, displaying apoptotic capacity in different cellular contexts. Together with previous evidence, showing that tau is cleaved during neuronal apoptosis (Canu et al., 1998), these results suggest that tau can be, at the same time, a "substrate" and an "effector of apoptosis". Caspase-3 was identified as candidate protease responsible for the activation of this positive feed-back cycle, ultimately leading to tau-mediated cell death. The final part of this project was to attempt the derivation of an antibody specifically recognizing the cleavage site of caspase-3 on tau. This should allow a direct test of the "tau truncation" hypothesis, as modified through this work