Autolysosomal acidification failure as a primary driver of Alzheimer disease pathogenesis

ABSTRACT

Genetic evidence has increasingly linked lysosome dysfunction to an impaired autophagy-lysosomal pathway (ALP) flux in Alzheimer disease (AD) although the relationship of these abnormalities to other pathologies is unclear. In our recent investigation on the origin of impaired autophagic flux in AD, we established the critical early role of defective lysosomes in five mouse AD models. To assess in vivo alterations of autophagy and ALP vesicle acidification, we expressed eGFP-mRFP-LC3 specifically in neurons. We discovered that autophagy dysfunction in these models arises from exceptionally early failure of autolysosome/lysosome acidification, which then drives downstream AD pathogenesis. Extreme autophagic stress in compromised but still intact neurons causes autophagic vacuoles (AVs) containing toxic APP metabolites, Aβ/β-CTFs, to pack into huge blebs and protrude from the perikaryon membrane. Most notably, AVs also coalesce with ER tubules and yield fibrillar β-amyloid within these tubules. Collectively, amyloid immunoreactivity within these intact neurons assumes the appearance of amyloid-plaques, and indeed, their eventual death transforms them into extracellular plaque lesions. Quantitative analysis confirms that neurons undergoing this transformation are the principal source of β-amyloid-plaques in APP-AD models. These findings prompt reconsideration of the conventionally accepted sequence of events in plaque formation and may help explain the inefficacy of Aβ/amyloid vaccine therapies.

Autophagic stress, a prominent pathological feature of AD, is characterized by failure of autophagy substrates to progress through the steps required for their complete degradation, resulting in the accumulation of substrate-laden autophagosomes and autolysosomes (collectively termed autophagic vacuoles [AVs]). Autophagic stress is exceptionally severe in AD brain relative to other neurodegenerative diseases and is more reminiscent of certain lysosomal storage disorders.

Considered mainly a protective cellular response, autophagy is mobilized to restore nutrient and energy homeostasis by degrading damaged or obsolete substrates that increasingly arise during aging and disease. Consistent with this concept, we previously reported that lysosomal biogenesis rises in cortical neurons in AD before neuropathological changes, such as extracellular β-amyloid plaques and intracellular MAPT/tau “tangles”, develop in neocortex. Later, transcriptomic analyses of brain and specifically neurons in late-onset AD have confirmed upregulated autophagy induction throughout the course of the disease. Countering this robust induction response is a progressive decline in efficiency of substrate clearance by lysosomes due to brain aging and emerging AD-related factors, leading to the accumulation of substrate-engorged AVs in distended segments of axons (dystrophic neurites) and ultimately in perikarya where AVs are normally rare. These two opposing forces of upregulated autophagy induction and failing substrate clearance, which accelerate as AD advances, propel the extreme autophagic stress/AV buildup in neurons. Genetic evidence underscores this neuropathological scenario by showing that PSEN1 (presenilin 1) harboring a loss-of-function mutation, the most common monogenic cause of AD, directly impedes subunit assembly and activity of the V-ATPase complex responsible for acidifying lysosomes and optimally activating hydrolases. Moreover, APP (amyloid beta precursor protein), arguably the most important gene linked to AD pathogenesis, has also been recently shown to disrupt V-ATPase assembly via a direct inhibitory interaction with APP-βCTF, the APP C-terminal fragment generated by BACE1 cleavage, which is increasingly appreciated to have varied AD-relevant pathogenic properties.

While the foregoing observations strongly implicate dysfunction of the ALP as a primary AD catalyst, it has not been clear what the precise interplay is between the multiple pathogenic metabolites of APP (Aβ, βCTF) and the components of the autophagy/lysosomal network – an interplay proposed to reflect a toxic partnership that is primary to AD etiopathogenesis. To explore this interplay and its downstream impact on extracellular β-amyloid deposition, we developed a tool to assess neuronal autophagy in the brains of mice modeling AD pathology at stages before and throughout the course of the disease [1] . Low-level transgenic introduction of an mRFP-eGFP-tagged LC3 via the Thy1-promoter that drives neuron-specific postnatal expression enables the pH of individual neuronal ALP compartments to be objectively assessed ratiometrically in brain and the numbers, sizes, and distributions of these compartments to be quantified. Analysis of five different transgenic AD models expressing either WT HsAPP, mutant HsAPP, or a combination of mutant forms of APP and PSEN1 yield similar results, uncovering a prominent intraneuronal pathobiology of the autophagy pathway preceding extracellular amyloid deposition (“plaques”) by as much as 5–6 months in the different models.

In our recent work, most importantly, we observed that autolysosome acidification fails early in neurons and well before β-amyloid deposits extracellularly even though lysosomes have fused with autophagosomes. This deficit is associated with substantially lowered V-ATPase activity in isolated brain lysosomes. Equally important, we observed that Aβ and APP-βCTF build up selectively within substrate-laden poorly acidified autolysosomes. The onset of these changes, which precedes neuropathological AD hallmarks, ranks with the earliest known neuronal functional changes (e.g., hyperactivity and endosomal anomalies) reported in AD models. The pH and V-ATPase defects and the appearance of APP-βCTF in lysosomes support our additional evidence that APP-βCTF, but not Aβ, directly inhibits lysosomal V-ATPase. These findings coupled with mounting genetic and epidemiological evidence imply that ALP disruption is a primary event in the etiology of AD, an implication bolstered by the genetic evidence used to support the amyloid cascade hypothesis.

Significant additional findings revealed major pathogenic implications of this advancing autophagic stress in the most affected neurons. In compromised but still structurally intact neurons, Aβ-containing AVs are profusely packed into huge membrane blebs that bulge out from the perikaryal membrane surface. By confocal microscopy, these brightly fluorescent blebs surrounding a perikaryon, especially when the nucleus is DAPI-stained blue, resemble petals of a flower; hence, the name PANTHOS (Poisonous + ANTHOS: flower) given to this unique pattern. These numerous perikaryal blebs strongly resemble, but are not, the dystrophic neurites that have long been described in plaques but only by light microscopy analysis and 2-D conventional EM. Our additional 3-D analyses by confocal tomography and serial 3-D reconstruction of PANTHOS neurons clearly identify blebs as continuous with the cytoplasm of the perikaryon and projecting from its plasma membrane via a tapered “neck”.

A second distinctive autophagic phenomenon revealed for the first time in AD model mice features AVs that amass around the nucleus, some coalescing with membrane tubules of the endoplasmic reticulum (ER), suggesting reticulophagy and possibly also membrane recruitment for new autophagosomes. Most surprisingly, we identified bundles of β-amyloid fibrils within many ER tubules. When immunostained by Aβ antibodies alone, this perinuclear concentration of β-amyloid and, more peripherally, the accumulations of Aβ in AVs within an intact PANTHOS neuron are identical to the appearance of extracellular amyloid plaques. Quantitative analyses show that upon delayed lysosomal cell death and microglial invasion, the β-amyloid lesion within an intact PANTHOS neuron is, in fact, the principal origin of extracellular amyloid plaques – akin to MAPT/tau “ghost tangles” that form in neurons and resist clearance after the neuron dies. While our dual fluorescence autophagy reporter facilitated identification of PANTHOS in mouse transgenic AD models, we have also detected PANTHOS neurons in late-onset AD brain with multiplex immunofluorescence labeling with appropriate autophagy, lysosome, and amyloid markers and DAPI staining.

Apart from the mechanistic insights into the etiological significance of lysosome dysfunction in AD afforded by our analyses, the overarching finding of clinical importance is the severity of ALP pathobiology detected prior to the neuron’s demise which yields the extracellular plaque lesion. For many years, the extracellular lesions have been considered the main therapeutic target in AD brain. On the contrary, our data imply that AD therapy is more likely to be successful if it targets intraneuronal ALP dysfunction, leading to neuron cell death and subsequent formation of extracellular plaques.

Funding Statement

This work was supported by the NIH [P01AG017617 and R01AG062376] to R.A.N.

Disclosure statement

No potential conflict of interest was reported by the authors.