The natural and molecular history of familial or sporadic Alzheimer's disease (AD) shows that APP (amyloid protein precursor) dysfunction is a consensual central etiological factor in Alzheimer's disease (AD). This is demonstrated by 1) genetic defects involving APP gene or APP dysfunction (such as PS1 or PS2), leading to the formation of neocortical amyloid plaques in familial AD; 2) transgenic mice with these mutated genes that develop plaques; 3) both sporadic and familial AD develop plaques. But two alternatives to explain the physiopathology can be proposed: a gain of toxic function of AB peptide (reflected by the amyloid cascade hypothesis) or a loss of function of APP, a ubiquitous and well conserved protein with numerous possible neurotrophic activities. On the other hand, AD is also characterized by another inescapable degenerating process: tauopathy, an intraneuronal aggregation of tau proteins into neurofibrillary tangles. Remarkably enough, progression of tauopathy in neocortical areas fully explains the progressive clinical deficits of AD, from memory loss to aphasia, apraxia, agnosia. Also one has to bare in mind that most demented patients and most dementing neurodegenerative disorders have a tauopathy. From that, it is concluded that APP an Tau are solid therapeutic targets. But if we know that APP and Tau dysfunctions interact to boost neurodegeneration in AD, we still do no know what are the intraneuronal signaling pathways to activate or to inhibit to stop the degenerating process. There are many hypotheses and many possible approaches: the inhibition of toxicity of plaque, of AB protofibrils, or of AB oligomers inside or outside the neuron, using vaccination or ligands (Alzhemed). On the other hand, modulation of secretases that cleave APP by inhibiting those involved in the amyloidogenic pathway or by stimulating those of the non-amyloidogenic pathway, is a major route of research. Also modulation of kinases or phosphatases possibly involved in the aggregation of tau is also investigated. Because animal models are not perfectly relevant, at the end of the long and costly pathway of drug discovery, therapeutic trials are the only way to test these different hypotheses.