Why Decline Happens Suddenly, Not Gradually

Benjamin Aaron Gustafsson AdultCognitiveDisease.com

---

Neurodegenerative diseases such as Alzheimer’s disease (AD), Parkinson’s disease (PD), amyotrophic lateral sclerosis (ALS), frontotemporal dementia (FTD), and Huntington’s disease (HD) have traditionally been studied in isolation, each linked to unique misfolded proteins and pathology. This dissertation presents a unified theory – Convergent Autophagic Collapse – positing that these disorders, despite distinct upstream triggers, converge on a common downstream collapse of the autophagy-lysosome pathway (ALP) and cellular energy metabolism. We critically evaluate how each disease engages a specific “choke point” in the ALP: lysosomal acidification in AD, mitophagy in PD, cargo recognition and trafficking in ALS/FTD. HD is treated separately and as the limit of the pattern, because staged human HD brain shows the clearance pathway intact through early disease and overtaken only at advanced stages (Berg et al., 2025). Through an extensive literature synthesis, we show that these upstream failures all lead to the same terminal state – a catastrophic positive-feedback loop of proteostatic overload and bioenergetic insolvency. We analyze this collapse using systems neuroscience and thermodynamic principles, framing it as a bistable phase transition: neurons switch from a healthy high-clearance, high-energy state to a pathological low-clearance, low-energy state once a critical threshold is crossed. The evidence includes human postmortem findings of clogged autophagic vacuoles in degenerating neurons, animal and cell model studies demonstrating specific autophagy defects, and recent experimental and theoretical work indicating the presence of tipping points in protein clearance dynamics. We discuss how this convergent model integrates with or challenges existing theories – protein misfolding (“proteopathy”), mitochondrial dysfunction, and glymphatic (interstitial waste clearance) hypotheses – and identify points of consensus and debate. The concluding analysis emphasizes implications for therapy: if a common downstream collapse underlies these diseases, then therapeutic strategies should shift from single-protein targets toward bolstering systemic proteostatic and metabolic capacity. We argue that detecting and intervening in the early, pre-collapse phase could help maintain neurons in a stable functional state, potentially delaying or preventing the onset of overt neurodegeneration. This synthesis aims to provide a rigorous, integrative framework that links molecular failures to neuronal system breakdown, offering a holistic perspective on neurodegenerative disease mechanisms and pointing toward multi-faceted approaches for intervention.

Introduction

Neurodegenerative diseases pose an urgent scientific and clinical challenge: they are common, debilitating, and currently incurable. A central problem in neurodegeneration research is the fragmentation of understanding across diseases. Alzheimer’s disease, Parkinson’s disease, ALS, FTD, and Huntington’s each have been traditionally defined by their signature misfolded proteins (amyloid-β and tau in AD, α-synuclein in PD, TDP-43 in ALS/FTD, mutant huntingtin in HD) and distinct neuropathologies. This protein-centric view has yielded vital insights, such as the amyloid cascade hypothesis in AD and prion-like spread models, but it has also siloed research into disease-specific paradigms. The result is a patchwork of theories that often consider neurodegenerative disorders in isolation, despite growing recognition of their overlapping features.

Increasing evidence suggests that beneath the diverse clinical and histopathological manifestations, these diseases may share common mechanistic threads. Nearly all involve the accumulation of aberrant protein aggregates and dysfunctional organelles, progressive synaptic and metabolic deficits, and eventual neuronal death. This observation has spurred interest in identifying shared pathways of neurodegeneration. By understanding points of convergence, we could explain why different insults lead to similar endpoints (neuronal dysfunction and death) and perhaps reveal broad-spectrum therapeutic targets. The significance of uncovering shared mechanisms is profound: it promises more efficient drug development (targeting fundamental processes rather than dozens of individual proteins) and a unified framework to interpret new findings in neurodegenerative biology. One candidate for such a unifying mechanism is dysfunction of the Autophagy-Lysosome Pathway (ALP) – the cell’s primary system for degrading and recycling proteins and organelles. Neurons, being long-lived and post-mitotic, are especially reliant on efficient autophagy and lysosomal clearance to maintain proteostasis¹². Notably, many genetic risk factors for neurodegenerative diseases implicate ALP components (e.g., PSEN1 in AD affects lysosomes, LRRK2 and GBA in PD relate to lysosomal function, SQSTM1/p62, OPTN, C9orf72 in ALS/FTD are autophagy regulators). Ageing – the greatest risk factor for sporadic AD, PD, and ALS – is associated with a decline in autophagic efficiency and lysosomal acidity, hinting that a late-life failure of proteostasis could be a tipping point for disease onset¹³. Recent authoritative reviews by Nixon and Rubinsztein (2024) underscore that defects in autophagosome formation, lysosomal pH maintenance, and other ALP functions are central across AD, PD, and FTD, and that age-related weakening of the ALP may precipitate the emergence of neurodegenerative pathology¹³. In parallel, theoretical models have framed neurodegeneration as a problem of cellular economics: neurons have a limited energy budget, and the maintenance of proteome quality is energetically costly¹⁴¹⁵. When the cost of clearing damaged proteins/organelles chronically exceeds energy production, a breaking point is reached where the neuron can no longer cope¹⁴¹⁵. This dissertation’s thesis argument is that AD, PD, ALS/FTD, and HD are best understood not solely as discrete proteinopathies, but as distinct upstream failures in the autophagy-lysosome and related qualitycontrol systems that converge on a common downstream collapse. We term this collapse a Convergent Autophagic Collapse – a point of irreversible decompensation in which the neuron’s proteostatic load overwhelms its clearance and energy-generating capacity, leading to runaway accumulation of protein aggregates and a sudden loss of homeostasis⁵¹⁶. In essence, each disease stresses the neuron’s clearance system in a different way, but all eventually drive the cell into the same catastrophic state of high entropy (disorder from accumulated junk) and low energy (failing mitochondria)¹⁷¹⁸. This state can be described by principles of bistability and phase transition: the neuron transitions from a stable healthy state to a stable diseased state once a critical threshold is crossed, rather than a slow linear progression⁶ ⁷. Such a perspective aligns with emerging concepts of “tipping points” in neurodegeneration, wherein gradual damage accumulation leads to an abrupt inflection into clinical decline once compensatory reserves are lost¹⁹. In the sections that follow, we will first review the landscape of existing neurodegeneration models – from the classic proteopathy framework to newer models of mitochondrial dysfunction and glymphatic clearance – to situate the Convergent Autophagic Collapse hypothesis in context. We will highlight both consensus points (e.g. widespread agreement that protein clearance mechanisms and energy metabolism are implicated in these diseases) and areas of contention (e.g. debates over whether protein aggregates are a cause or a consequence of cellular failure)²⁰²¹. Next, we outline our methodology for synthesizing evidence across multiple levels of analysis and disease contexts. Each subsequent chapter (Chapters 1–4) will delve into a major disease or group of diseases, examining how a specific ALP failure manifests and drives that disorder: Chapter 1 covers Alzheimer’s disease with its lysosomal acidification failure; Chapter 2 examines Parkinson’s disease and halted mitophagy; Chapter 3 discusses ALS and FTD (autophagy cargo recognition and trafficking failures) as well as Huntington’s disease, where the pathway proves competent early and marks the limit of the pattern; Chapter 4 then integrates these findings to explain the nonlinear “phase transition” to terminal collapse. Finally, the Conclusion revisits the thesis in light of the assembled evidence and discusses what this convergent model implies for early detection, intervention, and future research directions in neurodegenerative disease. By reframing neurodegeneration as a systems failure rather than a collection of independent protein-specific pathologies, this work aims to bridge gaps between subfields and promote a more unified understanding. The ultimate goal is to inform strategies that bolster the neuron’s overall clearance and metabolic resilience, thereby addressing the root of the “death spiral” that appears common to AD, PD, ALS/FTD, and HD²¹⁶.

Literature Review

To understand the theory of Convergent Autophagic Collapse in context, we first review prevailing models of neurodegenerative disease mechanisms. Each of these models emphasizes different aspects of neuronal failure – protein aggregation, mitochondrial energy failure, or extracellular clearance – and each provides pieces of a complex puzzle. We synthesize how these perspectives overlap or conflict, setting the stage for a unified systems view.

Proteopathy Models: Protein Aggregation as Causative Pathology

Historically, neurodegenerative diseases have been conceptualized primarily as protein misfolding disorders. In these models, the accumulation of specific aberrant proteins is seen as the initiating toxic event. For example, the amyloid cascade hypothesis of Alzheimer’s posits that oligomerization and deposition of amyloid-β (Aβ) peptides in the brain trigger a cascade leading to tau tangles, synaptic dysfunction, and neuron death. Similarly, in Parkinson’s, the aggregation of α-synuclein into Lewy bodies is thought to impair neurons; in ALS and FTD, mislocalized or aggregated TDP-43 and other RNA-binding proteins are believed to cause neurotoxicity; and in Huntington’s, mutant huntingtin protein with expanded polyglutamine tracts forms intraneuronal inclusions. These proteopathy paradigms have been reinforced by genetics (many familial cases are caused by mutations that make proteins prone to misfolding) and by neuropathology (postmortem brains consistently show these protein aggregates as disease hallmarks). Proteopathy models have guided the search for therapies aimed at reducing aggregates – e.g. antibodies to clear Aβ plaques or α-synuclein, small molecules to prevent protein fibrillization, etc. However, despite decades of effort, the success of such therapies has been limited. Aβ-targeting antibodies (e.g. aducanumab, lecanemab) can indeed remove plaques from the brain, yet they have shown only modest effects on cognitive decline in AD. Intriguingly, the removal of aggregates does not necessarily restore neuronal health. A recent study from Osaka Metropolitan University examined AD patients treated with lecanemab (an anti-Aβ antibody) and found that while amyloid plaques were cleared, the brain’s waste

clearance function did not improve in the short term²²²³. In these patients, MRI measures of

glymphatic flow (the brain’s interstitial fluid clearance system) remained impaired even after plaque reduction, suggesting that neuronal damage and clearance deficits were already well entrenched by the time of treatment²³²⁴. This empirical finding underscores a key contention in proteopathy models: are protein aggregates the root cause of neurodegeneration, or are they by-products of a deeper cellular failure? The persistence of dysfunction after aggregate removal hints that aggregates alone may be the “visible detritus” of disease, while underlying processes remain disrupted¹¹¹⁹. Some researchers have argued that the field’s focus on aggregates as primary toxins might overlook fundamental factors like protein turnover stress or energy deficits. Kinetic models of proteostasis support this skepticism. Kepp (2019) introduced a quantitative model suggesting that neurodegenerative diseases arise when neurons can no longer keep up with protein turnover costs, rather than from a specific toxic conformation of a protein¹⁵. In this view, protein aggregation is a symptom of overwhelmed proteostatic mechanisms: as long as misfolded proteins are efficiently cleared, they do not accumulate or cause harm, but when clearance capacity is exhausted, aggregates inevitably appear. This explains why highly expressed, energetically “costly” proteins (which are harder to continually turn over) are often the ones that accumulate in late-life disease²⁵. It also reframes therapeutic priorities – perhaps enhancing the cell’s ability to cope with misfolded proteins (through boosting chaperones or degradation pathways) could be more effective than simply attacking the aggregates themselves. Indeed, Kepp’s model argues that disease onset occurs when so much energy is spent on removing misfolded proteins that the cell can no longer meet other demands, tipping the balance toward collapse²⁶¹⁵. This perspective aligns with accumulating evidence that many ostensibly “toxic” aggregates might be relatively inert deposits or even protective sequestrations of misfolded proteins, whereas the soluble misfolded species and the burden they place on clearance systems are the real culprits. Nonetheless, the proteopathy framework remains influential, and there is consensus that protein aggregates are at least a proxy for disease progression, even if not the sole cause. One area of emerging

agreement is that proteostasis networks (the cellular systems maintaining protein folding and

clearance) decline with age, lowering the threshold for aggregates to form. Voicu et al. (2025) describe neurodegeneration as a “multidimensional collapse of biological organization” rather than a simple linear accumulation of protein²⁷. In their multi-omics analysis, they highlight how proteostatic integrity and other cellular networks progressively fail together. This broader view doesn’t contradict proteopathy models so much as embed them in a larger failure of protein homeostasis. In summary, proteopathy models correctly identify a key feature of neurodegeneration – abnormal protein accumulation – but increasingly, these aggregates are seen as part of a larger story. The contentious question is whether aggregates drive the collapse or are epiphenomena of an overwhelmed proteostasis system. The Convergent Autophagic Collapse theory leans toward the latter: it posits that aggregates (Aβ, tau, α-synuclein, TDP-43, etc.) are markers of a collapsing cellular cleanup system, converging evidence that the neuron’s waste-management has faltered²⁸¹.

Mitochondrial Dysfunction and Bioenergetics: The Power Supply Problem

Another major model of neurodegeneration centers on mitochondrial dysfunction and energy failure. Neurons are energetically demanding cells, requiring constant ATP supply for ion pumping, axonal transport, vesicle recycling, and more. It has long been observed that neurodegenerative diseases feature early and progressive declines in energy metabolism: for instance, FDG-PET scans in Alzheimer’s show hypometabolism in affected cortical regions; in Parkinson’s, complex I activity in the electron transport chain is often reduced in the substantia nigra; in Huntington’s, mutant huntingtin interacts with mitochondrial proteins leading to respiratory dysfunction. The mitochondrial cascade hypothesis of AD even proposes that late-onset AD is driven by age-related mitochondrial DNA damage and metabolic decline, which then secondarily leads to Aβ accumulation and tau pathology. Mitochondrial models emphasize oxidative stress and ATP depletion as central events. Dysfunctional mitochondria produce excess reactive oxygen species (ROS), which can damage cellular components. A lowered ATP level impairs all energy-dependent processes, notably including protein quality control mechanisms (the proteasome and lysosomal proton pumps require ATP). Thus, a vicious cycle can ensue: mitochondrial ROS damages proteins and lipids, increasing the load of misfolded proteins and defective organelles that need clearance; at the same time, energy shortage hampers the clearance of that rising waste load. This scenario is a classic positive feedback loop leading to neuronal degeneration. There is significant consensus that mitochondria and autophagy-lysosome pathways are tightly interconnected in neurodegenerative diseases. Many studies document how perturbing one affects the other. For example, if lysosomal degradation is impaired (as in AD models), undigested substrates can accumulate within neurons, and evidence suggests this can secondarily affect mitochondria (perhaps by clogging axonal transport of mitochondria or via altered signaling)²⁹³⁰. Conversely, if mitochondria are not cleared (as in PD models with defective mitophagy), those mitochondria produce ROS that can directly

damage lysosomal membranes and enzymes³¹²⁹. It has led to the concept of a mitochondria-

lysosome axis of failure²⁹. Empirical support for this axis comes from observations such as: in Parkinson’s, mutations in the lysosomal enzyme GBA (glucocerebrosidase) increase α-synuclein burden, while mutations in PINK1/Parkin (mitophagy regulators) lead to accumulation of unhealthy mitochondria and also secondary lysosomal stress – in both cases, patients develop similar PD pathology. In Alzheimer’s, some familial AD mutations in PSEN1 impair lysosomal acidification and cause altered mitochondrial functioning; similarly, Aβ accumulation can perturb mitochondrial function, and mitochondrial dysfunction can promote Aβ production – illustrating multi-directional interactions. One key point of agreement in the field is that neurodegeneration involves an energy–proteostasis dual failure. Simons, Levin, and Dichgans (2023) emphasize the importance of resilience factors (glial support, metabolic reserve) that determine whether pathology (like Aβ accumulation) actually tips the system into clinical disease¹⁹. They describe neurodegeneration reaching a tipping point when adaptive systems (including metabolic support and immune clearance) are exhausted, unleashing self-reinforcing cascades of damage¹⁹. This notion dovetails with the idea that neurons manage to maintain function despite accumulating protein aggregates until mitochondrial/energy compensation fails – at which point a rapid decline occurs. However, contentions remain. Some researchers argue that specific mitochondrial toxins or defects (e.g. environmental toxins in PD causing complex I inhibition) are primary causes in certain cases, while others see mitochondrial impairment as secondary to protein aggregation stress. The reality likely involves feedback loops rather than simple one-way causation. The Convergent Autophagic Collapse theory integrates mitochondrial dysfunction as both an upstream trigger in diseases like PD and a downstream exacerbating factor in all neurodegenerations. In the collapse model, bioenergetic insolvency (energy demand outstripping supply) is the final common pathway of neuronal death¹⁴³². We will later see how, for example, in PD the initial “power plant failure” (mitophagy arrest) triggers this insolvency, whereas in AD the failure starts at the lysosome (“incinerator failure”) but still eventually causes an energy crisis as waste buildup impairs mitochondria²⁹³⁰. In summary, mitochondrial dysfunction models contribute the critical insight that an energetic collapse is at the heart of neurodegeneration. They align with autophagy models in a natural way: clearing cellular waste requires energy, and energy production requires clearance of damaged mitochondria – thus forming a reciprocal dependency²⁹. The consensus is that therapies boosting metabolic function (e.g. enhancing mitochondrial biogenesis or efficiency) and those boosting proteostatic clearance might synergize. The contentious point is prioritization: is it more effective to target energy production or to reduce the energy burden (by decreasing protein aggregation)? The collapse theory suggests both approaches converge on the same goal – increasing the headroom before the system hits the insolvent state.

Glymphatic Clearance and Extravascular Drainage: The Brain’s Garbage Disposal

Beyond intracellular processes, the brain has a system-wide “clearance” mechanism known as the glymphatic system. This glial-dependent perivascular network facilitates the removal of metabolic waste, including protein solutes like Aβ, from the interstitial fluid of the brain. During sleep, cerebrospinal fluid (CSF) is driven along periarterial spaces, intermixes with interstitial fluid, and then drains waste-laden fluid out along perivenous routes – effectively washing the brain. Impairment of glymphatic clearance has been proposed as a contributing factor in neurodegenerative diseases, especially AD where extracellular amyloid plaques form. The “glymphatic clearance” theory posits that if the brain’s plumbing backs up, proteins like Aβ and tau accumulate in the parenchyma, exacerbating toxicity. Several lines of evidence support glymphatic involvement. In animal models, sleep deprivation or aging (both associated with reduced glymphatic flow) lead to increased Aβ deposition. Human studies using MRI and PET imaging have started to measure glymphatic flow or related clearance indices; for example, a diffusion tensor imaging along perivascular space (DTI-ALPS) index has been developed to gauge glymphatic function. In the context of AD, arterial stiffness and cardiovascular risk factors can impair the pulsatile driving force for glymphatic flow, potentially explaining links between vascular health and dementia. The Osaka Metropolitan study mentioned earlier provides in vivo evidence: even after Aβ plaques were cleared pharmacologically, the DTI-ALPS index in AD patients showed no immediate improvement³³ ²⁴. This suggests that by the time amyloid is extensive, the clearance system (glymphatic pathways) may have suffered structural or functional damage that cannot be quickly reversed by removing Aβ. In other words, amyloid buildup might be both a cause and effect of glymphatic failure, eventually creating a situation where simply removing amyloid doesn’t fix the underlying “garbage disposal” breakdown²³²⁴. Figure 1: Magnetic resonance image from an AD patient illustrating glymphatic clearance assessment. The colored regions indicate diffusion along perivascular spaces (DTI-ALPS index), a measure of interstitial fluid drainage. In a study, this index was measured before and after clearing amyloid-β plaques with lecanemab. Strikingly, no significant change was observed in glymphatic clearance post-treatment, indicating that waste-removal function remained impaired²²²⁴. This finding highlights that clearing extracellular plaques alone is insufficient to restore the brain’s clearance system once neurodegeneration is underway. It underscores the need to address the underlying clearance pathways (glymphatic flow, blood–brain barrier transport, etc.) in conjunction with removing protein aggregates. The glymphatic perspective ties into the autophagy-lysosomal perspective as two halves of a whole: autophagy primarily clears intracellular debris, while glymphatic flow clears extracellular and system-level waste. They meet at points such as the handling of protein aggregates – e.g. a neuron might expel protein waste that then needs removal via perivascular routes. Failures in one can overload the other. For instance, if intracellular clearance falters (autophagic buildup), more protein might be released or accumulate extracellularly, burdening microglia and glymphatic drainage. Conversely, if glymphatic drainage slows, extracellular Aβ builds up, which can further deposit and interfere with synaptic and vascular function, potentially stressing neurons and their intracellular clearance as well³⁴³⁵. There is consensus that multifactorial clearance impairments occur in diseases like AD: along with neurons’ own digest-and-recycle systems, the support systems that remove waste from the brain environment are also failing. This broad view is evident in current literature that speaks of an overall failure of the brain’s “garbage clearance” on multiple fronts (proteasome, autophagy, microglia phagocytosis, CSF flow). Where debate exists is often in identifying the primary driver versus secondary contributors. Some researchers prioritize amyloid production as the root issue and see clearance failure as secondary; others propose that a primary clearance deficit (due to aging or genetics) allows amyloid to accumulate. The truth may differ by individual – for instance, an APOE4 carrier might have inherently poorer Aβ clearance leading to early plaque buildup, whereas another individual might have robust clearance but excessive Aβ production. For our convergent collapse model, the glymphatic theory reinforces the concept that clearance capacity is finite and can be overwhelmed. It extends the collapse beyond the single neuron to the level of neural tissue and the whole brain: just as a neuron can drown in its own waste if autophagy fails, the brain can drown in metabolic waste if interstitial clearance fails. Both lead to toxic buildup and functional breakdown. Thus, a comprehensive strategy for neurodegeneration would need to tackle clearance at all scales.

Synthesis of Perspectives: Toward a Systems Model and Points of Contention

Bringing these perspectives together, a picture emerges of neurodegeneration as a systemic collapse of

cellular homeostasis under the weight of accumulated damage. The proteopathy view provides the

tangible markers (aggregates) and has driven home the point that certain proteins are common denominators in each disease. The mitochondrial view contributes the crucial role of energy depletion and oxidative damage. The autophagy-lysosome view (closely linked to proteostasis) highlights the failure of the intracellular recycling system. The glymphatic view highlights failure of the intercellular and extracellular waste drainage. Rather than seeing these as competing explanations, modern frameworks increasingly see them as interlocking pieces of the pathophysiology. Victor Voicu and colleagues (2025) articulate that neurodegeneration is not a one-dimensional process but a collapse across multiple levels – signaling, gene expression, proteostasis, organelles, and neural networks ²⁷. They and others call for integrative models that consider how these levels influence each other. For example, a disturbance in proteostasis (protein aggregation) can trigger inflammatory signaling (microglial activation) and gene expression changes (stress responses), which then may impair metabolic processes, creating a downward spiral. Mikael Simons and colleagues (2023) introduce the metaphor of tipping points to describe how a slow accrual of damage in preclinical phases can suddenly “flip the switch” to overt disease once resilience factors are depleted¹⁹. They particularly note roles for glial, immune, and vascular systems in providing resilience – a clear nod to the importance of systemic, not just neuronal-intrinsic, factors¹⁹. Amid this emerging consensus on complexity, several contentious issues remain in the field, which we will keep in mind as we examine the convergent collapse theory:

• Cause or Effect: Are protein aggregates and organelle accumulations the primary cause of neuron death, or are they benign byproducts of a failing clearance system? This chicken-and-egg question surfaces repeatedly. The convergent collapse model leans toward seeing aggregates as downstream – essentially symptoms of the collapse (albeit ones that can themselves contribute to toxicity once present). Not all researchers agree; some maintain that specific aggregates (like oligomeric Aβ or tau) have unique toxic activities that must be neutralized. We will discuss evidence for both views, such as experiments showing neurons can survive high aggregate loads if energy and proteostasis are sustained versus experiments showing direct toxicity of certain protein species. Ultimately, in a collapse scenario, both can be true: aggregates initially accumulate due to clearance failure, then those aggregates further worsen the failure (feed-forward loop).
• Disease Specificity vs. Commonality: Critics might ask whether unifying AD, PD, ALS, FTD, and HD under one framework glosses over crucial differences. Certainly, each disease has unique aspects (e.g. the regional selectivity: hippocampus in AD, substantia nigra in PD, motor cortex/spinal cord in ALS, striatum in HD). Our review will address how the ALP failures intersect with these differences – for instance, the unique demands of a dopaminergic neuron might make mitophagy especially critical in PD, whereas the unique protein processing in an AD neuron might make lysosomal pH especially vulnerable. A nuanced view acknowledges both a shared mechanism and disease-specific manifestations. The concept of different “upstream choke points” leading to a common downstream state is an attempt to respect the diversity and find the unity.
• Therapeutic Targeting: Should therapy be aimed at the specific upstream defect in each disease or at the common downstream collapse? There is contention here. Many current efforts remain targetspecific (e.g. anti-tau immunotherapies in AD, anti-α-syn in PD, antisense oligonucleotides for mutant huntingtin in HD). These might alleviate the burden of a particular toxic protein. The convergent view

suggests combining such approaches with treatments that bolster overall clearance capacity or energy metabolism could yield better outcomes¹⁰. For example, a therapy that restores

lysosomal acidification or enhances autophagy could, in principle, benefit multiple conditions simultaneously. Some broad approaches like mTOR inhibition (to stimulate autophagy) or AMPactivated kinase (AMPK) activation have shown preclinical promise across diseases, but clinical translation has been slow. The debate is far from settled, but there is a trend toward exploring

multi-target approaches. We will return in the conclusion to how adopting a systems view might

shift therapeutic development. In light of these perspectives, the theory of Convergent Autophagic Collapse can be seen as an effort to integrate the proteostatic and bioenergetic dimensions of neurodegeneration into one model. It asserts that a neuron can compensate for a while against accumulating misfolded proteins and damaged organelles (thanks to robust energy production and clearance mechanisms), but as these compensatory systems wear down (with age or stress), the neuron approaches a point of no return. Beyond that point, the cell transitions into a self-propagating state of collapse wherein energy failure and proteostasis failure feed into each other⁷⁵. This echoes themes from the literature: Nixon & Rubinsztein’s “tipping point” of autophagy decline in late age¹³, Simons et al.’s loss of system resilience unleashing pathology¹⁹, and Cotton et al.’s recent experimental demonstration of a phase transition between a state of balance and a state of aggregate overload in neurons⁸⁹. With this foundation set, we now turn to the methodology of our analysis, before diving into the diseasespecific chapters that provide the evidence for each link in the convergent collapse chain.

Methodology

Analytical Framework

This dissertation employs a systems neuroscience analytical framework to synthesize findings across molecular, cellular, and organismal levels for multiple neurodegenerative diseases. Rather than conducting new experiments, we perform a comprehensive analysis of empirical studies and theoretical models, treating the body of literature as a dataset to be interrogated for patterns and common principles. The approach is inherently interdisciplinary: we integrate data from biochemistry, cell biology, neuropathology, genetics, and computational modeling to build and stress-test the Convergent Autophagic Collapse theory. Key elements of our methodology include:

• Literature Selection and Review: We surveyed over 180 peer-reviewed sources spanning experimental research on human patients, animal models, and cell cultures, as well as computational and theoretical papers³⁶. Sources were chosen to cover all major facets of the ALP and neurodegeneration: autophagy and lysosomal function in neurons, mitochondrial dynamics, proteostasis and aggregation, glial and vascular contributions, and nonlinear dynamics in disease progression. We prioritized recent studies (past ~10 years) to capture the latest evidence, while also including classic foundational studies (e.g. discovery of PINK1/Parkin pathway, early autophagy in AD work by Nixon’s group, etc.). Review articles by leading experts (for example, Nature Reviews by Nixon & Rubinsztein 2024¹³, or Simons et al. 2023 in Neuron¹⁹) were used as starting points to identify consensus and areas of debate.
• Comparative Analysis Across Diseases: A central methodological step was comparing how seemingly disparate diseases affect similar cellular processes. We dissected each disease in terms of autophagy stages (initiation, cargo recognition, vesicle trafficking, lysosome fusion, degradation) and mitochondrial status, mapping which step is primarily compromised and what the downstream effects are. This comparative matrix helped identify the distinct “choke points” for each disease and also showed the uniform outcome of those failures (accumulation of autophagic vacuoles, energy deficits, etc.)⁴³⁷. For example, by tabulating side-by-side the features of AD neurons (giant autophagic vacuoles, lysosomal alkalinization) and PD neurons (undegraded mitochondria, high oxidative damage) and ALS neurons (p62/TDP-43 aggregates accumulating), we observed all share evidence of impaired autophagic flux and signs of metabolic stress, albeit via different entry points.
• Use of Engineering and Thermodynamic Concepts: We adopted concepts from systems engineering and thermodynamics to frame the analysis. Neurons were viewed as entities with an

energy budget and a proteostatic load, analogous to a machine with inputs and outputs and

subject to resource constraints¹⁴. We applied the notion of bistability – having two stable states – to neuronal homeostasis⁶⁷. This was informed by mathematical models and empirical evidence (e.g. Cotton et al. 2025 show a cellular phase transition in proteostasis³⁸). We used the concept of a phase transition or tipping point to interpret the sudden collapse observed in end-stage disease. This conceptual framework guided the interpretation of findings: for instance, when reviewing timecourse studies of AD in mouse models, we looked for signs of accelerating pathology that might indicate a threshold being crossed, rather than a steady linear increase.

• Evidence Integration and Causal Inference: A challenge in synthesizing literature is distinguishing correlation from causation. We addressed this by giving particular weight to experiments that directly manipulate components of the ALP or metabolism and observe effects on disease phenotypes. For example, gene knockout or overexpression studies (PINK1 or Parkin deletion in animals to simulate PD, or VPS34 activation to stimulate autophagy in AD models) were used to test causality of autophagy disruptions. Pharmacological interventions that specifically target lysosomal pH (like bafilomycin to inhibit the v-ATPase) were reviewed for their ability to induce neurodegenerative changes, thus mimicking disease. Human genetic data also provided causal anchors: if familial mutations in a gene cause disease and that gene’s function is known in autophagy or mitophagy, we interpreted that as strong evidence linking the pathway to pathogenesis (e.g. SQSTM1/p62 mutations in ALS/FTD directly tie cargo recognition failures to disease). By weaving together such causal evidence from multiple directions, we constructed a network of interactions feeding into collapse.
• Cross-Validation with Empirical “Big Data”: Where available, we incorporated large-scale omics and imaging studies to ensure our theoretical model did not overlook important factors. For instance, transcriptomic studies showing upregulation of autophagy genes in early disease vs. downregulation in late disease were considered for understanding how the cell’s own responses evolve. Proteomic studies of insoluble protein accumulation (e.g. Pace et al., 2018 showing proteome-wide solubility shifts in neurodegeneration) gave a holistic view of proteostasis collapse. Brain network imaging studies indicating functional network breakdown (which might correspond to distributed energy failure) were also considered. These broad datasets served as a check: a valid convergent theory should be consistent with system-wide changes observed in disease, not just isolated molecular pathways.
• Critical Evaluation and Alternative Explanations: At each step, we evaluated alternative explanations for findings. For example, could lysosomal dysfunction in AD be secondary to Aβ toxicity rather than primary? Could mitophagy defects in PD be an effect of α-syn aggregates rather than a parallel cause? We addressed these by examining temporal ordering in models (does lysosomal pH rise precede plaque deposition?) and rescue experiments (does restoring lysosomal function despite Aβ presence improve outcomes?). We also considered diseases or observations that might challenge the theory – for instance, why do some individuals accumulate pathology (like high amyloid burden) but remain cognitively intact? We hypothesized that perhaps their systemic clearance/energy capacity is higher, delaying collapse³⁹⁴⁰. Throughout, we remained alert to data that conflict with a simple convergent story and have noted such issues in our analysis, aiming for a balanced view that acknowledges gaps or unresolved questions.

Justification of a Systems Neuroscience Approach

A fundamental rationale for our methodological approach is the recognition that neurodegeneration is a multiscale problem. The path from a single molecular event (like a point mutation or a misfolded protein) to the death of a neuron and the clinical symptoms in a human spans many levels of biological organization. A reductionist method looking at one protein or one organelle in isolation is insufficient to capture the emergent phenomena of neurodegeneration, such as the abrupt clinical decline or the selective vulnerability of certain neuronal populations. Therefore, we employ a systems neuroscience approach that can accommodate feedback loops, network effects, and non-linear dynamics. This approach is in line with current trends in the field. For example, Nixon & Rubinsztein (2024) explicitly call for an integrative framework to understand autophagic-lysosomal function in health and disease, noting that multiple homeostatic systems co-fail in late-age neurodegeneration¹³. Similarly, recent multidisciplinary collaborations (like the referenced Cotton et al. preprint from 2025) combine experimental and theoretical methods to map the landscape of neuronal states and find critical thresholds⁸⁹. By using systems concepts such as attractor states, we hope to capture the essence of these complex findings in a unifying narrative. In practice, the systems approach meant that as we built our model, we continuously iterated between the micro (e.g. specific molecular defects like a mutant enzyme) and the macro (e.g. whole-cell behavior like energetic collapse). We treated the neuron akin to a node in a control system, with internal feedback (e.g. AMPK activation when ATP is low, which in turn upregulates autophagy) and external feedback (e.g. glial clearance, blood flow adjustments). Our analysis examines how these feedback mechanisms can stabilize the system under normal conditions but also how they can produce runaway instability under pathological conditions (for instance, the feedback where failing lysosomes lead to more damaged mitochondria, which further impair lysosomes²⁹³⁰). Empirical findings from diverse approaches were crucial to ground this systems view. We included evidence

from human studies (brain imaging, CSF biomarkers, autopsy pathology quantification) to ensure

relevance to actual disease. We drew on animal models (transgenic mice, drosophila, C. elegans) which allow controlled perturbation of pathways – these helped confirm causality, such as demonstrating that artificially inducing lysosomal alkalinization in a mouse’s neurons can lead to AD-like protein accumulation. We also used cell culture and in vitro experiments that detail molecular mechanisms (e.g. cell studies showing how mutant huntingtin alters autophagosome cargo loading⁴¹). Molecular analyses (enzyme assays, pH measurements in organelles, ATP level assays) provided quantitative data on the extent of dysfunction at various stages. By considering all these empirical modalities, our methodology ensures that the theoretical framework of convergent collapse is not a mere abstract idea but is continually tested and refined against real-world data. Finally, by justifying a systems approach, we also implicitly justify why focusing on shared pathways is worthwhile. If the hypothesis holds, it means that disparate diseases can inform each other – knowledge of lysosomal biology from AD research can illuminate PD and vice versa. As such, our method also involved cross-disciplinary validation: taking an insight from one disease’s literature and checking if analogous phenomena have been observed in another. Often, we found that they have, though terminology might differ. For example, what AD researchers call “autophagic vacuole accumulation”⁴², ALS researchers might observe as “p62 and ubiquitin-positive inclusions” – two sides of impaired turnover. Recognizing these parallels was key to building the convergent model.

In summary, the methodology of this dissertation is characterized by comprehensive literature

integration, comparative cross-disease analysis, and the application of a systems theory lens. This approach is well-suited to capture the complex, nonlinear nature of neurodegenerative diseases and to evaluate the viability of a unifying hypothesis that spans multiple disorders. With this methodology in place, we proceed to the in-depth analyses of each disease context, which serve as the building blocks for the convergent collapse theory.

Chapter 1: Alzheimer’s Disease and Lysosomal Acidification Failure

Alzheimer’s disease (AD) is the most common neurodegenerative dementia, pathologically defined by extracellular amyloid-β (Aβ) plaques and intracellular neurofibrillary tangles of hyperphosphorylated tau protein. However, beyond these hallmark lesions, AD brains also exhibit profound endosomal-lysosomal pathology. Electron microscopy and histochemical studies of AD brain tissue, dating back to the work of Robert Terry and Ralph Nixon, have shown large numbers of autophagic vacuoles, lysosomal storage bodies, and dystrophic neurites clogged with undigested material in affected neurons⁴³. These observations suggest a failure of the autophagy-lysosome pathway in AD long before end-stage cell death. In this chapter, we examine how AD can be framed as a disease of lysosomal dysfunction – specifically, a failure of lysosomal acidification that stalls autophagic degradation.

The Lysosomal “Choke Point” in AD

Under normal conditions, autophagy in neurons proceeds through several steps: initiation of autophagosome formation, sequestration of cytosolic cargo, autophagosome transport to the cell body, fusion with lysosomes, and acidification of the autolysosome for enzymatic breakdown of cargo. In AD, research indicates that the bottleneck is not at the formation or fusion stages, but at the degradation stage – autophagosomes form and fuse to lysosomes, but their cargo is not efficiently digested⁴³. This leads to a characteristic buildup of large, acidic vesicular structures (sometimes called “autophagic vacuoles” or AVs) in neurons, especially in dystrophic neurites around plaques⁴². Many of these AVs are found to contain partially digested substrates: fragments of membranes, proteins, and sometimes dense bodies that resemble lipofuscin (an indigestible pigment). The presence of these accumulations implies the

lysosomes are not functioning properly as digestive “incinerators.”

Ralph Nixon’s studies provided a clue to the nature of this dysfunction: lysosomal pH in AD is improperly elevated (less acidic) compared to normal⁴⁴. Lysosomal hydrolases (like cathepsins) have an acidic pH optimum (around pH 4.5–5). In AD models, lysosomal pH has been measured in the range of 5.5–6.0 or higher, at which many hydrolases become inactive⁴⁴. In essence, the lysosomes in AD neurons lose their “stomach acid,” neutralizing the enzymes needed to break down proteins⁴⁴. This condition is sometimes termed a “proteolytic failure” in AD⁴³. Autophagosomes still fuse with lysosomes (hence we see autolysosomes), but because of the higher pH, the cargo inside is not effectively degraded – creating what the Nixon lab called “autophagic vacuoles” that persist abnormally⁴³. These vacuoles can grow in size and number, contributing to neuritic dystrophy (swollen, dysfunctional axonal segments filled with debris). What causes lysosomal acidification to fail in AD? There are at least two mechanisms identified, corresponding to familial and sporadic AD contexts:

1. Presenilin-1 (PS1) Mutation – v-ATPase Dysfunction: Presenilin-1 is a component of the γ-secretase complex (which generates Aβ from APP) and is famously mutated in familial early-onset AD. Beyond its role in Aβ production, PS1 has a critical function in lysosome biology: it acts as a chaperone for the proper maturation and trafficking of the V0a1 subunit of the vacuolar ATPase (v-ATPase)⁴⁵. The v-ATPase is the proton pump responsible for acidifying lysosomes (analogous to a “pump” that maintains the acidic pH). In healthy cells, PS1 ensures that the V0a1 subunit is correctly glycosylated and delivered to lysosomal membranes⁴⁶. Mutations in PS1 disturb this chaperone function, leading to improperly assembled v-ATPase complexes on lysosomes⁴⁶. As a result, lysosomes in PS1-mutant cells fail to acidify normally⁴⁷⁴⁸. This has been demonstrated in cell models: PS1 knockout or mutant fibroblasts have lysosomal pH ~6 versus ~5 in wild-type, and impaired turnover

of autophagic substrates⁴⁹⁵⁰. Thus, familial AD due to PS1 mutations directly causes

lysosomal acidification failure, independent of Aβ production. It is a “choke point” at the level of the proton pump itself: the pump is present but nonfunctional, like an engine that cannot start. The consequences are catastrophic for proteostasis – even if autophagy is activated and autophagosomes form, the final degradation step is blocked. 2. APP-βCTF Accumulation – v-ATPase Inhibition: In sporadic AD, where PS1 is normal, evidence points to a different molecular cause of lysosomal failure. The βCTF (C-terminal fragment) of amyloid precursor protein (APP) – essentially the membrane stub left after β-secretase cuts APP – can accumulate in neurons when γ-secretase processing is slow or saturated (γ-secretase normally would cleave βCTF to produce Aβ). Studies (e.g. by Willen et al., 2017) found that APP-βCTF accumulates

within late endosomes/lysosomes in AD and can physically interact with the v-ATPase. In AD

models, the APP-βCTF was seen to bind to the proton pump complex and reduce its activity⁵¹. One analogy given is that βCTF acts like “sand in the gears,” partially jamming the rotary mechanism of the v-ATPase⁵². The net effect is again an elevated lysosomal pH. This mechanism is particularly insidious because it links the core pathological protein (APP/Aβ pathway) to the clearance pathway: as Aβ production and APP processing become dysregulated, a byproduct (βCTF) impairs the very system (lysosomes) that could help clear protein aggregates. It’s a feed-forward loop: more βCTF ->

less lysosomal activity -> more accumulation of Aβ and other substrates. Notably, βCTF has

been shown to accumulate in sporadic AD brains and correlates with lysosomal dysfunction markers ⁵¹. This suggests that even without a PS1 mutation, sporadic AD neurons face a similar acidification problem, likely later in the disease course when Aβ/APP metabolism is already abnormal. In summary, whether through mutant PS1 or through APP-βCTF accumulation (or possibly other factors like oxidized lipids, metals, etc.), AD neurons suffer a failure of the v-ATPase “pump” that acidifies

lysosomes⁵³⁵⁴. Thus, AD’s ALP choke point is at the lysosome’s ability to degrade cargo – the

“incinerator” doesn’t burn the trash because it doesn’t get hot enough, so to speak.

Downstream Effects: Proteostatic Traffic Jams and Feedback Loops

The failure of lysosomal acidification in AD has immediate and long-term consequences for the neuron:

• Autophagic Traffic Jam: As soon as lysosomal degradation slows, a backlog of autophagosomes forms. Autophagy is a dynamic flux – when working well, autophagosomes continuously form and are rapidly cleared after fusing with lysosomes. If clearance slows, autophagosomes and amphisomes accumulate in the axons and cell body. In AD models, within minutes of inhibiting lysosomal acidification (e.g. with bafilomycin), autophagic vacuoles build up. In AD brain, abundant autophagic vacuoles indicate this chronic backlog⁴³. The axonal transport of these vacuoles can also become impaired due to overcrowding, leading to swelling of axons (this contributes to the dystrophic neurites seen around plaques). The whitepaper describes it succinctly as “Traffic Jam: high input, blocked output” in AD⁵⁵⁵⁶. Neurons continue producing waste (damaged proteins, organelles) – that’s the high input – but output is blocked at the lysosome, so waste accumulates internally.
• Inactivation of Degradative Enzymes: With lysosomal pH rising above the optimal range, hydrolases like Cathepsin D, Cathepsin B, etc., become less active or inactive⁴⁴. Some enzymes might even misfold or not be properly processed at non-acidic pH. For example, prosaposin processing or glucocerebrosidase activity could be reduced, compounding the problem. It’s a cascade – one failure (pump) triggers many failures (dozens of enzymes). Chemical analyses of AD brains have found that many lysosomal enzymes are present but not fully active in affected regions, consistent with this pH effect.
• Incomplete Substrate Clearance and Aggregate Formation: Undigested substrates in lysosomes can include Aβ itself (which is normally cleared partly via endolysosomal pathways), tau (which may be turned over by autophagy), and other proteins. The buildup of partially degraded products can

result in lysosomal storage phenomena. For instance, one study found accumulation of autophagy

substrates like poly-ubiquitinated proteins and even mitochondrial remnants in AD neurons, essentially stuck in autolysosomes. Over time, such accumulations may fuse and form larger aggregates or secondary structures – potentially even seeding plaque and tangle formation extracellularly when neurons eventually expel or leak these materials. This is speculative, but one can imagine that a neuron full of undigested Aβ in lysosomes might eventually rupture or exocytose them, contributing to plaque growth.

• Energetic Cost – Futile Pumping: An often overlooked consequence is the energy cost of a failing pump. If lysosomal pH is too high, cells often respond by upregulating v-ATPase activity or numbers in an attempt to re-acidify. This can create a scenario of futile cycling: the neuron spends ATP on the v-ATPase which is either not functioning (PS1 case) or constantly being thwarted (βCTF case), leading to a high ATP consumption for little benefit⁵⁷⁵⁸. Essentially, the neuron is trying to pump a leaky or jammed system. This represents an energy drain that can worsen cellular metabolism. There is some evidence that neurons under proteostasis stress have activated AMPK (energy stress sensor) and suppressed mTOR, indicating they sense energy insufficiency. In AD brains, regions with high pathology often show signs of energy hypometabolism and increased AMP/ATP ratios. The acidification failure might be one contributor to this, by forcing the cell to overwork its pumps and still not clear the waste.
• Reverse Feedback to Mitochondria: The waste accumulation in autophagic vacuoles can indirectly affect mitochondria. One connection is via oxidative stress: accumulating iron and other redox-active contents in overloaded lysosomes can cause lipofuscin and reactive oxygen species formation that might permeabilize lysosomal membranes (the phenomenon of lysosomal membrane permeabilization, LMP). LMP can spill proteases into the cytosol, damaging mitochondria or other organelles. Additionally, if autophagy is stuck, damaged mitochondria are not being effectively degraded (a form of mitophagy failure). The whitepaper notes a “reverse loop” where lysosomal dysfunction impairs clearance of mitochondria⁵⁹. In AD, studies have shown that abnormal lysosomal buildup coexists with defective mitophagy; for instance, mitophagy markers are reduced in AD postmortem brains and improving mitophagy in AD mouse models (genetically or pharmacologically) improves pathology. It’s plausible that AD’s lysosomal failure leads to secondary mitophagy failure, which in turn causes more ROS and energy deficit – linking AD to what we see as primary events in PD. Taken together, Alzheimer’s can be viewed through an engineering lens as a system where the “acidification pump” is broken, causing a traffic jam of waste and a cascade of secondary failures. The neuron exists in a precarious high-entropy state: full of partially degraded proteins (high entropy) that it cannot eliminate, yet expending precious energy in a futile attempt to do so⁵⁷⁶⁰. Over time, this situation likely triggers cell death via apoptotic or necrotic pathways – for example, protease leakage from lysosomes can activate caspases, and prolonged energy stress can initiate intrinsic apoptosis.

Evidence and Experimental Support

Our depiction of AD’s autophagy-lysosome collapse is supported by multiple experimental findings:

• Ultrastructural Pathology: Electron microscopy of AD neurons (from human biopsies and transgenic mouse models) consistently shows large autophagic vacuoles, multilamellar bodies, and other lysosomal remnants⁴³. These are rarely seen in young healthy neurons. In PD or other conditions, you see different ultrastructures (like Lewy bodies); AD is distinct in the sheer volume of these vacuoles. One classic study (Boland et al., 2008, J. Neurosci.) found that in an AD mouse model, autophagosomes accumulated 8-fold compared to controls due to failed clearance, and that experimentally restoring lysosomal function cleared them and reduced amyloid burden⁶¹.
• Biochemical Assays: Measurements of lysosomal pH in AD fibroblasts and neurons (obtained via fluorescent pH indicator dyes targeted to lysosomes) showed a significant alkalinization in cells carrying PS1 mutations or expressing APP-βCTF⁶²⁵⁰. Enzyme assays for cathepsins in AD brain tissue reveal reduced enzymatic activity despite normal or elevated cathepsin protein levels, consistent with a non-optimal pH environment (e.g., Cataldo et al. 1995 observed buildup of procathepsin D in AD brains).
• Genetic Models: Mice with PSEN1 deleted specifically in neurons develop an AD-like lysosomal storage phenotype – with age they accumulate substrate in lysosomes and show neurodegeneration, despite not having amyloid plaques (because APP processing is altered but that’s secondary)⁴⁵⁴⁶. This underscores that PS1’s loss can cause neurodegeneration through lysosomal failure alone. Likewise, knock-in mice with AD-linked PS1 mutations show lysosomal pH increases and autophagy defects early in life, even before amyloid pathology. Conversely, boosting lysosomal acidification has shown promise: e.g. treating AD model mice with small molecules that acidify lysosomes or enhance cathepsin activity leads to reduced amyloid and tau pathology (though these approaches are still experimental).
• Human Genetics: Apart from PS1, other AD risk genes relate to endosomal-lysosomal function. APOE4, the strongest risk factor for sporadic AD, has been linked to poorer recycling of lipoproteins and Aβ clearance via endosomal pathways. SORL1, PICALM, BIN1 are AD risk genes involved in endosomal trafficking, which could influence autophagosome-lysosome fusion or cargo delivery. TCIRG1, a v-ATPase subunit gene, has variants nominally associated with AD in some studies. The convergence of these on the endosomal-lysosomal system lends genetic weight to the importance of this pathway in AD.
• Therapeutic Trials: While not definitive, there are hints that targeting lysosomal function could be beneficial. For instance, enhancing lysosomal biogenesis via TFEB (a transcription factor master regulator of lysosome genes) in AD model mice reduced pathology. Also, as noted, anti-amyloid immunotherapy alone did not restore clearance function²⁴ – implying that co-treatments aimed at restoring lysosomal health might be needed.

In summary, AD epitomizes a neurodegenerative disease where “the incinerator is broken” – the

lysosome cannot acidify properly, so cellular garbage (including toxic proteins) piles up⁵³⁴⁷. The neuron gets locked in a state of high proteostatic load, gradually suffocating under its own waste while its energy reserves diminish. This sets the stage for a phase transition into irreversible neurodegeneration once compensatory mechanisms (like upregulating autophagy or glycolysis) are maxed out. The next chapters will show that other diseases hit the same end state by different routes; but understanding AD’s route via lysosomal failure provides a template for recognizing parallel themes in those disorders.

Chapter 2: Parkinson’s Disease and Mitophagy Arrest

Parkinson’s disease (PD) is characterized clinically by motor symptoms (resting tremor, rigidity, bradykinesia) and pathologically by the loss of dopaminergic neurons in the substantia nigra and the presence of Lewy bodies (intracellular inclusions rich in α-synuclein). While PD has unique features, it shares with other neurodegenerative diseases the accumulation of protein aggregates and dysfunctional organelles in affected neurons. In PD, compelling evidence points to mitochondrial dysfunction as a central player – indeed, some have called PD primarily a mitochondrial disorder. This chapter focuses on how PD can be

understood as a failure of mitophagy, the specialized autophagic process that removes damaged

mitochondria. We will see that PD’s “choke point” is the quality control of the energy supply: when defective mitochondria are not cleared, they accumulate, leading to an energy crisis and further cellular damage⁶³. PD thus exemplifies the “power plant failure” mode of convergent autophagic collapse⁶⁴.

Mitophagy: The Power Plant Quality Control

Mitochondria are the “power plants” of the cell, generating ATP via oxidative phosphorylation. Neurons, with high energy demands and long, spindly architectures, are extremely sensitive to mitochondrial performance. They also face particular challenges: mitochondria must be transported up and down axons, and aged or damaged mitochondria need to be efficiently turned over to maintain a healthy population. Mitophagy is the process by which cells identify damaged mitochondria and target them for autophagic degradation. A key pathway regulating mitophagy is the PINK1/Parkin pathway:

• PINK1 (PTEN-induced putative kinase 1) is a sensor of mitochondrial health. In healthy mitochondria, PINK1 is imported into the inner mitochondrial membrane and quickly degraded. But if a mitochondrion loses its membrane potential (a sign of damage or dysfunction), PINK1 cannot be imported and instead accumulates on the outer mitochondrial membrane⁶⁵. There, PINK1 acts as a signal – essentially a molecular flag indicating “this mitochondrion is sick.”
• Parkin is an E3 ubiquitin ligase that normally resides in the cytosol. When PINK1 builds up on a damaged mitochondrion’s surface, it phosphorylates both ubiquitin and Parkin itself, which triggers Parkin to translocate from the cytosol to that mitochondrion⁶⁶. Parkin then ubiquitinates numerous outer membrane proteins on the mitochondria, marking the organelle for destruction ⁶⁷. The ubiquitin tags serve as docking sites for autophagy adaptors (like p62, optineurin, etc.), which recruit the forming autophagosome to engulf the mitochondrion. This elegant system ensures selective removal of only the malfunctioning mitochondria, preserving the healthy ones. It’s a critical quality control, especially in neurons where cumulative damage could be catastrophic. In Parkinson’s disease, genetic mutations in the key players of this pathway provided the first hints of its importance. Autosomal recessive forms of PD can be caused by loss-of-function mutations in either PINK1 or PRKN (the Parkin gene). Patients with these mutations develop early-onset parkinsonism, strongly suggesting that without functional PINK1/Parkin, dopaminergic neurons cannot survive long-term⁶⁸. Indeed, such mutations “break the sensor-effector loop” of mitophagy⁶⁸. The result is that **damaged mitochondria become “invisible” to the clearance system and accumulate in the cell⁶⁸. In affected neurons, one would expect to see an abundance of dysfunctional, depolarized mitochondria that are not being cleared. Postmortem and cell studies confirm this: dopaminergic neurons from PD patients often show swollen, structurally abnormal mitochondria and evidence of impaired electron transport chain activity. Fibroblasts from patients with PINK1 or Parkin mutations have defective mitophagy in lab assays (for instance, they fail to clear mitochondria after inducing mitochondrial damage with chemicals like CCCP). Thus, familial PD illustrates a clear case of autophagic quality-control failure at the level of mitochondria. However, most cases of PD are sporadic (idiopathic), not caused by those mutations. Remarkably, clues indicate that sporadic PD also involves mitophagy disruption, albeit via different mechanisms. αSynuclein, the protein that aggregates in Lewy bodies, has been found to interfere with Parkin translocation and mitophagy. Experimental studies show that oligomeric α-synuclein can bind to mitochondria and also bind Parkin, “gumming up” the works of the mitophagy process⁶⁹. In cultured neurons, overexpression of mutant α-syn (like A53T, a familial PD mutant) causes accumulation of damaged mitochondria and prevents Parkin recruitment to depolarized mitochondria⁶⁹. Thus, sporadic PD, which features elevated αsyn levels and aggregates, may suffer a secondary mitophagy block due to α-synuclein’s toxic interactions. There are other pieces of the puzzle: sporadic PD patients often have deficits in proteins like DJ-1 (an oxidative stress sensor) and LRRK2 (a kinase that, when mutated, can affect autophagy), which further tie into mitochondrial and lysosomal interplay. Moreover, environmental mitochondrial toxins (like MPTP, which causes a Parkinsonian syndrome by poisoning mitochondria) and general aging-related mitochondrial DNA damage can all increase the load of dysfunctional mitochondria. If Parkin/PINK1 activity is insufficient (whether due to genetic factors, α-syn interference, or simply being overwhelmed by too many damaged mitochondria), the outcome is similar: mitophagy arrest – the cell fails to remove bad mitochondria.

In summary, Parkinson’s disease can be seen as hitting the ALP at the level of “mitochondrial

clearance.” The “power plants” break down faster than they are cleared, leading to a buildup of defective energy generators⁶³. We next examine the consequences of this scenario for the neuron, which propel it toward the convergent collapse.

Consequences of Mitophagy Failure: Dirty Fuel and a Vicious Cycle

When mitophagy is compromised in a neuron, two major consequences drive the pathology of PD:

1. Bioenergetic Deficit – The Drop in ATP Supply: Dysfunctional mitochondria are inefficient at producing ATP. They often have impaired electron transport, meaning they consume substrates (oxygen, NADH) without effectively pumping protons and synthesizing ATP. As these defective mitochondria accumulate and even crowd out healthier ones, the overall ATP-generating capacity of the neuron declines⁷⁰. In dopaminergic neurons, which have especially high energy needs due to their large axonal arborizations and pacemaking activity, even a moderate drop in ATP can be devastating. The neuron finds itself with an energy shortfall: processes like vesicle recycling, ion pumping, and indeed autophagy itself (which needs ATP for vesicle transport and fusion) slow down or falter⁷⁰⁷¹. This is a direct route to cell dysfunction – synapses may fail, and the neuron cannot maintain homeostasis. In PD, reduced ATP in substantia nigra neurons is supported by findings of decreased activity of mitochondrial complex I (especially from studies of PD brain tissue and the fact that complex I inhibitors like rotenone cause PD-like degeneration in animals). As ATP output drops, the cell’s ability to keep up with maintenance is compromised, potentially allowing more damage to accrue (a feedback loop we will detail shortly).
2. Oxidative Stress – ROS Overproduction: Damaged mitochondria are notorious for leaking electrons, producing reactive oxygen species (ROS) such as superoxide and hydrogen peroxide³¹. With a buildup of such mitochondria, the neuron endures chronic oxidative stress. ROS can directly damage membranes, DNA, proteins – notably, they can oxidize lipids in lysosomal membranes, cause oxidation of dopamine (in dopaminergic neurons, leading to toxic quinones), and oxidize proteins like α-synuclein making them more prone to aggregate. Particularly relevant is ROS damage

to lysosomes: lipid peroxidation of lysosomal membranes can cause them to become

permeabilized³¹, releasing enzymes that further harm the cell. ROS can also partially disable lysosomal enzymes (some cathepsins are sensitive to oxidative modifications). Another aspect is that certain byproducts, like oxidized dopamine itself, can be toxic to lysosomes. So an accumulation of

bad mitochondria will create a chemical environment in the neuron that poisons the lysosomal

system, effectively linking mitochondrial failure to autophagic failure (forward loop)²⁹. Moreover, ROS and energy loss can activate cellular stress pathways (inflammation via NLRP3 inflammasome, c-Jun N-terminal kinase, etc.) which may accelerate cell death. These two factors – energy deficit and ROS – form the crux of PD’s degenerative spiral. But importantly, they

feed back into the autophagy-lysosome pathway, establishing a vicious cycle

• Forward Loop (Mito -> Lysosome): As noted, dysfunctional mitochondria (from mitophagy failure)

lead to low ATP and high ROS, which impairs lysosomal function²⁹. Low ATP directly affects

lysosomal acidification (v-ATPase needs ATP)²⁹, echoing the AD scenario. ROS can cause lysosomal membrane damage and enzyme dysfunction. So mitophagy failure will begin to cause general autophagy failure: even intact lysosomes may not work as well in a low-ATP, high-ROS setting. For example, studies show that experimentally inhibiting mitochondrial function in neurons causes accumulation of autophagic vacuoles (since lysosomes don’t acidify without ATP). Thus, PD neurons may develop not only mitochondrial accumulation but also secondary lysosomal clearance problems, meaning protein aggregates (like α-syn) and other waste also start accumulating. Indeed, this might contribute to Lewy body formation – some part of Lewy bodies contains undegraded protein that might have accumulated due to impaired autophagic flux.

• Reverse Loop (Lysosome -> Mito): Conversely, if lysosomes are not working well (perhaps due to aging, or genetic factors like GBA mutations which are common in PD), then mitophagy will be incomplete – even if Parkin tags the bad mitochondria, they won’t be degraded efficiently.

Lysosomal dysfunction impairs clearance of damaged mitochondria⁷². This is clearly seen in

PD associated with GBA (glucocerebrosidase) mutations: GBA encodes a lysosomal enzyme, and its mutation leads to substrate accumulation in lysosomes and is a strong risk factor for PD. GBAmutant neurons accumulate α-syn and have impaired mitochondrial function, suggesting that lysosomal impairment can lead to mitophagy impairment. Another piece of evidence: LRRK2, a kinase mutated in PD, can phosphorylate Rab proteins and is thought to disrupt autophagy/ lysosome dynamics; LRRK2 mutations can cause both lysosomal abnormalities and mitochondrial defects. Thus, any lysosomal problem will further allow bad mitochondria to accumulate – completing the vicious cycle. The whitepaper description encapsulates this interplay: in PD “the Power Plant and the Waste Processing Plant are mutually dependent” and when one fails, it poisons the other³⁰. The positive feedback loop ensures that once mitophagy significantly falters, a runaway degeneration ensues³⁰. The cell ends up in an ever-worsening energy/oxidative environment, with mounting protein waste and organelle damage, analogous to an engine running on dirty fuel that clogs it more the longer it runs.

Evidence Linking Mitophagy to PD

The identification of PINK1 and Parkin mutations in familial PD was a turning point that firmly linked autophagy (specifically mitophagy) to Parkinson’s pathogenesis. Since then, a wide array of evidence has bolstered this connection:

• Animal Models: Pink1 or Parkin knockout mice individually don’t have dramatic neurodegeneration of dopaminergic neurons (mice may be more resilient or have compensatory mechanisms), but they do exhibit mitochondrial abnormalities and subtle motor deficits. In Drosophila, loss of Pink1 or Parkin causes a clear phenotype with muscle degeneration and motor defects due to dysfunctional mitochondria (flight muscle mitochondria become swollen and the flies can’t fly). These are rescued by interventions that promote mitochondrial fission or by overexpression of Parkin in Pink1 mutants (placing Parkin downstream of Pink1). This genetic epistasis confirms Pink1 works upstream to recruit Parkin, aligning with the pathway we described⁷³⁶⁷.
• Neuronal Cultures: Human dopaminergic neurons derived from iPSCs of patients with PINK1 or Parkin mutations show accumulation of unhealthy mitochondria, decreased respiration, and heightened sensitivity to stress (e.g. they die faster when exposed to oxidative stressors). If you reintroduce a normal PINK1 or Parkin gene, those phenotypes are rescued, which underscores causality.
• Sporadic PD brain analysis: Postmortem brains in idiopathic PD often show markers of impaired autophagy: for instance, accumulation of p62/SQSTM1 (an autophagy adaptor) in nigral neurons, indicating backlog of autophagy substrates. They also show decreased levels of some autophagy and lysosome proteins (perhaps due to chronic stress and resource depletion). Mitochondrial markers in surviving neurons can be abnormal; neuromelanin granules (which are basically undegraded oxidized catecholamine and protein aggregates in nigral neurons) are more abundant or irregular, suggesting lysosomal overload (neuromelanin forms in lysosomes from autophagic material). All these align with a picture of autophagy/mitophagy stress.
• MPTP and Toxin Models: The chemical MPTP causes parkinsonism by converting to MPP+ in the brain, which poisons mitochondria in dopaminergic neurons (inhibiting complex I). This model, besides showing how primary mitochondrial damage can replicate PD features, has also been found to involve autophagy changes – MPTP-treated mice upregulate autophagy (attempting to clear damaged mitochondria), and enhancing autophagy can ameliorate MPTP toxicity, whereas blocking autophagy worsens it. This suggests the cell actively tries to use mitophagy in toxin models, and failure to do so leads to neuron loss. Another environmental link: rotenone (a pesticide) also inhibits mitochondria and induces PD-like pathology in rats, with evidence of α-syn aggregation and autophagic stress. These environmental models reinforce the tie between mitochondria and the PDlike state.
• Interactions with α-Synuclein: α-Synuclein itself is normally partially cleared by chaperonemediated autophagy (CMA) in lysosomes. When α-syn accumulates, it can inhibit CMA by clogging LAMP2A receptors. This might indirectly put more burden on macroautophagy and mitophagy systems. Conversely, excess α-syn can physically bind to mitochondria and potentially block the PINK1-Parkin recruitment or affect mitochondrial dynamics. Mitochondria in α-syn transgenic models tend to be fewer and less functional, and those models often have autophagy alterations too (e.g., increased p62). Therapies that clear α-syn or prevent its aggregation often show restoration of mitochondrial health. So α-syn and mitochondrial defects reinforce each other in a bidirectional toxic relationship.
• Clinical Data: One of the earliest features of PD, even before motor symptoms, is an energy metabolism deficit seen in PET scans of the brain and sometimes systemic changes (PD patients have altered muscle mitochondrial function and lower metabolic rates). There is also evidence that regular exercise (which boosts mitochondrial function and autophagy) correlates with lower PD risk or slower progression, suggesting that keeping the quality control systems active might be protective. Drugs like mitophagy inducers or antioxidants are being trialed in PD (e.g. the diabetes drug Metformin activates AMPK and might enhance mitophagy; some small molecules mimic NAD+ to improve mitochondrial metabolism). These approaches are built on the idea that supporting the cell’s energy and clearance systems can modify PD progression. In the context of the convergent collapse theory, Parkinson’s disease illustrates a situation where the initiating hit is to the energy production/clearance of mitochondria. Once that fails, it sets off a chain reaction that looks very much like the “death spiral” described for collapse: ATP drops, proteostasis fails, aggregates form (Lewy bodies), and neurons reach a tipping point and die. The dopaminergic neurons likely have unique sensitivities (they rely on pacemaking calcium currents that tax mitochondria, they have dopamine that can auto-oxidize, etc., making them the “canary in the coal mine” for an overall collapse). But interestingly, as PD advances, pathology is not confined to dopaminergic cells – other regions (cortex, autonomic neurons) develop Lewy pathology, possibly reflecting how the initial mitochondrial/autophagic dysfunction can propagate stress signals or even α-syn seeds through the brain. To summarize, PD’s core pathology of mitophagy arrest supports the convergent theme: failing to eliminate dysfunctional mitochondria leads to an energetic and proteostatic catastrophe in neurons ⁷⁰⁷¹. In the end, whether the neuron started with lysosomes failing (like AD) or mitochondria failing (like PD), both scenarios result in low energy, high waste, and self-amplifying damage. Thus, Parkinson’s disease, when cast as a “power plant failure,” aligns with the notion that neurodegeneration is fundamentally a collapse of the cell’s housekeeping and energy economy. The next chapter will examine ALS, FTD, and HD, which introduce yet other entry points (cargo recognition and loading failures), but we will see they too converge on the same terminal cellular crisis.

Chapter 3: ALS, FTD, and Huntington’s — Cargo Recognition,

Logistics, and the Limit of the Pattern

In this chapter, we consider three neurodegenerative conditions together: Amyotrophic Lateral Sclerosis (ALS), Frontotemporal Dementia (FTD), and Huntington’s Disease (HD). ALS and FTD are now understood to be overlapping disorders on a spectrum – they share genetic causes in some cases and common molecular pathologies like TDP-43 protein aggregates. Huntington’s disease is a polyglutamine expansion disorder distinct from ALS/FTD in clinical and genetic terms, and its cellular pathology intersects with autophagy dysfunction on a different schedule. We group them here for a reason that turns out to cut two ways. ALS and FTD emphasize failures in cargo recognition and trafficking (the “logistics network” of autophagy), and in both the genes that cause the disease encode parts of that network, which is what makes the clearance reading causal there. Huntington’s has long been read alongside them as a failure of cargo engulfment⁴, but that reading comes from cell and mouse models, and staged human HD brain does not support it: the pathway is competent through early disease and is overtaken only late (Berg et al., 2025). HD therefore serves here as the limit of the pattern rather than a further instance of it — a useful limit, because it shows what the convergence argument does and does not license. 3.1 ALS and FTD: Autophagic Cargo Recognition and Trafficking Breakdowns Amyotrophic Lateral Sclerosis (ALS) is a fatal motor neuron disease characterized by degeneration of upper and lower motor neurons, leading to paralysis. Frontotemporal Dementia (FTD) is a group of dementias affecting the frontal and temporal lobes, causing behavioral changes or language deficits. Clinically distinct, ALS and FTD were traditionally studied separately. However, they are now known to overlap: a significant subset of patients show ALS-FTD mixed syndromes, and the most common genetic cause is shared – a hexanucleotide repeat expansion in the C9orf72 gene. Pathologically, the majority of ALS cases and a large fraction of FTD cases share a hallmark: cytoplasmic aggregates of TDP-43 (TAR DNAbinding protein 43) in neurons and glia, indicating a common molecular pathology of protein mislocalization and aggregation. From the ALP perspective, ALS and FTD highlight issues in the “front end” of autophagy – how cargo is identified and how autophagosomes are formed and trafficked. Key discoveries implicate specific failures:

• C9orf72 and Autophagy Initiation/Trafficking: The C9orf72 repeat expansion (GGGGCC repeats) is the most frequent genetic cause of ALS and FTD. How does C9orf72 protein relate to autophagy? C9orf72 has been found to form a complex with other proteins (SMCR8 and WDR41) that acts as a

GTP-exchange factor (GEF) for certain RAB GTPases involved in endosomal trafficking⁷⁴⁷⁵. It

particularly regulates RAB1, RAB7, RAB8, RAB11, etc., which coordinate trafficking steps like

autophagosome maturation and endosome-lysosome fusion. It also interacts with the ULK1

complex, which is the master regulator that initiates autophagosome formation⁷⁶. Loss of C9orf72 function (which can occur via haploinsufficiency – reduced expression due to the repeat expansion) therefore hits autophagy at two points:

• It can reduce ULK1 activation, stalling autophagy “at the starting line”⁷⁷ (meaning fewer autophagosomes get made when needed).
• It can cause a “logistical breakdown” in trafficking – autophagosomes or endosomes form but are poorly transported and fail to fuse with lysosomes efficiently⁷⁵. Under C9orf72 loss, neurons show accumulating p62-positive puncta and build-up of endosomal vesicles – essentially a pile-up of cargo that wasn’t delivered to lysosomes⁷⁵. Thus, C9orf72 mutations likely create an autophagy bottleneck where waste is identified and packaged but not properly routed. This aligns with what is seen in patient iPSC-derived neurons: they have increased numbers of autophagosomes but impaired clearance, and large cytosolic aggregates of proteins like TDP-43 may result because these autophagosomes cannot mature or degrade their contents.
• SQSTM1/p62 and OPTN (Optineurin) – Cargo Adaptor Mutations: A subset of ALS (and some FTD) cases are caused by mutations in SQSTM1 (p62) or OPTN (optineurin), both of which are autophagy cargo receptors⁷⁸. These adaptors normally recognize ubiquitinated cargo and bind to LC3 on the autophagosome membrane, literally acting as “loading cranes” that attach the cargo to the forming autophagosome⁷⁸⁷⁹. p62 and optineurin are especially important for aggregating proteins (like TDP-43) and even for organelles (optineurin can also help in mitophagy and in clearing protein complexes). Mutations in p62 (often in the ubiquitin-binding UBA domain) can make it unable to bind ubiquitinated targets⁸⁰. Mutations in optineurin (found in some ALS) often disrupt its ability to bind ubiquitin or to be phosphorylated by TBK1 (another ALS gene, incidentally, which activates adaptors). The consequence is a “cargo recognition failure”⁸⁰: the autophagy machinery is intact

and autophagosomes form, but specific toxic cargo – notably TDP-43 aggregates – are not getting loaded and removed⁸¹. This explains a pathological paradox: In ALS/FTD, we see both

evidence of active autophagy (e.g. many autophagosomes) and yet accumulation of TDP-43 and other aggregate-prone proteins. It’s as if the garbage trucks are running, but they are empty and the trash stays on the street. Indeed, studies confirm that in ALS models with p62 or optineurin dysfunction, bulk autophagy (e.g. turnover of long-lived proteins) might be relatively normal, but the clearance of specific aggregating proteins is impaired. The result is that TDP-43, FUS, and other proteins form aggregates in the cytoplasm, which likely become toxic (through mechanisms like sequestering essential RNAs or proteins, or causing phase separation issues).

• Progranulin and Lysosomal Enzymes: FTD caused by GRN gene haploinsufficiency (mutations in Progranulin) sheds light on another autophagy-related issue: lysosomal enzyme environment. Progranulin is a secreted growth factor that is processed into granulins in lysosomes, and these granulins appear to support lysosomal enzyme function (the exact mechanism is still being studied). FTD patients with GRN mutations have a condition somewhat akin to lysosomal storage disorders: they accumulate undegraded substrates like lipofuscin (a pigment composed of oxidized proteins and lipids) in neurons⁸²⁸³. Loss of progranulin leads to inefficient lysosomal proteolysis, because enzymes like cathepsin D and glucocerebrosidase become less active⁸⁴⁸⁵. While this is more of a lysosomal problem (like AD’s acidification failure in some ways), it intersects with the ALS/ FTD spectrum and demonstrates again how different steps can be hit: in this case the “lysosomal health” is compromised, leading to backlog of waste (including proteins and lipids). Progranulindeficient neurons show accumulation of potentially toxic waste and a neuroinflammatory response.

Bringing these together, ALS and FTD can be seen as syndromes where the “logistics and cargo identification” phase of autophagy breaks down. The consequences are

• Selective Vulnerability of Neuron Types: Motor neurons (in ALS) and certain cortical neurons (in FTD) might be especially sensitive to these failures. Motor neurons are large, with long axons – they depend heavily on efficient cargo transport (for delivering autophagosomes from distal axon to soma, etc.), so C9orf72-related trafficking issues could particularly harm them. Additionally, motor neurons are highly active cells requiring fast protein turnover at neuromuscular junctions. If adaptors like p62 aren’t loading, misfolded proteins like SOD1 (another ALS-related protein) or TDP-43 accumulate quickly. In FTD, certain neuron populations might be uniquely affected by progranulin loss or by TDP-43 pathology due to their gene expression profiles (e.g. high reliance on granulins or low redundancy in autophagy adapters).
• TDP-43 Pathology: TDP-43 is normally a nuclear RNA-binding protein; in disease it mislocalizes to the cytoplasm and aggregates. Autophagy is one way cells clear excess TDP-43. In ALS/FTD, autophagic dysfunction means TDP-43 clearance is insufficient, so it accumulates, misfolds, and forms inclusions. These inclusions can further hinder cell function (by depleting nuclear TDP-43 needed for RNA splicing, etc., and by sequestering other factors). We see a direct link: many ALS genes are autophagy related (C9orf72, SQSTM1, OPTN, TBK1, VCP) and they all converge to influence how TDP-43 and other proteins are handled. When they fail, TDP-43 pathology emerges – which is a hallmark downstream event in ALS/FTD.
• Energetic and Stress Feedback: Although ALS and FTD often focus on protein aggregates, underlying cellular stress can feed back to autophagy. E.g., TDP-43 aggregates can disrupt mitochondrial gene expression leading to energy deficits; also, aggregates trigger unfolded protein responses that might initially upregulate autophagy (thus many autophagosomes in ALS neurons), but if the core cargo recognition is broken, that autophagy induction may yield little in return. In effect, an ALS motor neuron might ramp up AMPK signaling and autophagy induction because it senses protein stress, but due to p62 or C9orf72 issues, this yields little benefit for the specific cargo that needs removing. How much energy is squandered in the process has not been measured in human tissue, and the figure should not be assumed. Over time, the neuron loses the energy battle and undergoes degeneration (some evidence shows ATP levels drop in ALS motor neurons as disease progresses, and enhancing energy supply can modestly prolong survival in animal models).

In summary, ALS and FTD illustrate a convergent collapse pathway initiated by failures in the

“recognize and haul” stages of autophagy. The “trucks” might not pick up their cargo (p62/optineurin mutations) or might not know where to go (C9orf72 deficiency causing routing errors)⁸⁶. The result is that toxic proteins (like TDP-43) and possibly damaged organelles remain in the cytosol, stressing the cell until it can no longer cope. Despite relatively intact lysosomal acidification or mitochondria (at least early on), these neurons still end up in the same predicament: accumulating garbage and declining function. 3.2 Huntington’s Disease: A Pathway That Is Competent Early Huntington’s disease (HD) is a hereditary neurodegenerative disorder caused by a CAG trinucleotide repeat expansion in the HTT gene, leading to an expanded polyglutamine tract in the huntingtin protein. The disease produces motor dysfunction, cognitive decline, and psychiatric symptoms, and is pathologically marked by mutant huntingtin (mHTT) protein aggregates (inclusion bodies) in neurons, particularly in the striatum and cortex. Huntington’s differs from ALS/FTD in a way that matters for the argument of this dissertation: it is a single-gene autosomal dominant disease whose causal gene has nothing to do with the clearance apparatus. The lesion begins in the neuron’s own nucleus, and whatever happens to autophagy in HD happens downstream of it. Research by Martinez-Vicente et al. and others uncovered a striking phenomenon in HD models: inefficient

cargo capture. Normally, when autophagy is induced, the amount of LC3-II (a marker of

autophagosome membranes) increases as autophagosomes form. In HD cell and animal models, LC3-II levels are elevated – indicating robust autophagosome formation – yet the cargo (mutant huntingtin and other cytosolic proteins) is not efficiently degraded⁸⁷⁸⁹. The vesicles were reported to contain less cytosolic material than normal. Two features of that study are routinely dropped when it is summarised, and both are decisive here. First, the same paper states that the autophagic vacuoles “form at normal or even enhanced rates in HD cells and are adequately eliminated by lysosomes, but they fail to efficiently trap cytosolic cargo in their lumen” (Martinez-Vicente et al., 2010). The defect it describes sits at cargo recognition, upstream of the lysosome; it is not a lysosomal bottleneck, and the paper does not claim one. Second, the work was done in HD cell lines, mouse models, and patient-derived cells — not in staged human brain. Mechanistically, in those models mutant huntingtin interferes with the autophagic machinery at the cargo-loading step. One reported mechanism is that mHTT interacts aberrantly with the autophagy adapter protein p62 (SQSTM1) and possibly other components where the autophagosome membrane forms⁴¹, disrupting assembly of the cargo recognition complex at the phagophore⁴¹. Wild-type huntingtin, for its part, has a role in vesicular transport and autophagy, facilitating cargo trafficking to autophagosomes through interactions with motors and scaffolding proteins; the polyQ-expanded protein loses that normal function while gaining toxic interactions. The human brain tells a different story. Staged HD striatum and cortex, graded HD2 to HD4 by Vonsattel criteria, show neurons that are comparatively normal at HD2 (Berg et al., 2025). Cargo-loaded autophagosomes and cathepsin-positive autolysosomes are readily observed, implying a lack of significant blockage in autophagosome formation and in autophagosome–lysosome fusion. Lysosomal biogenesis is upregulated and proteolysis is preserved, and between them they maintain autophagic clearance through early-stage disease. It is only at advanced stages that progressive HTT build-up and autolysosome accumulation appear, and these may signify a failure of autophagy substrate clearance rather than of the machinery that performs it. The accumulation is more prominent in striatum than in cortex, following the regional vulnerability of the disease. Ultrastructural appearances of the “empty vesicle” kind are in part attributable to fixation artefact in this tissue — vesicular profiles that prove on staining to be calnexin-positive endoplasmic reticulum. Two claims that have circulated on the strength of the model work do not survive this. The first is the barren or empty autophagosome as a description of the HD brain: human staged tissue does not show it. The second, which rests on the first, is the futile cycle — the proposal that HD neurons spend large amounts of ATP assembling autophagosomes that clear nothing, running the clearance engine at high RPM in neutral. There is no human evidence for that energetic drain, and the finding that clearance is preserved early removes the premise it needs. HD neurons do have an energy problem, well documented in patients and models as early metabolic deficits — weight loss, altered brain glucose metabolism — but it is attributable to mitochondrial and transcriptional injury by mHTT, not to a clearance pathway spinning uselessly. The distinction changes the therapeutic sign, which is the practical reason it matters. In AD, induction is already elevated and the failure lies downstream at degradation, so adding induction only sends more material into a blocked pathway; the target there is degradative capacity. In HD the clearance machinery is fully competent in early disease, and that is precisely the window in which stimulating the pathway can be expected to help (Berg et al., 2025). The same intervention carries opposite signs in the two diseases. A framework that files them under one lesion will prescribe the wrong thing for one of them.

Common Pathological Denominator and Evidence

By examining ALS, FTD, and HD together, we note a pattern and also its boundary. The end states resemble one another: neurons filled with protein aggregates and struggling energetically. The routes there do not. In ALS and FTD the causal genes sit in the clearance apparatus itself, and the pathway fails early. In HD the causal gene sits elsewhere, the pathway works through early disease, and its eventual failure is a late event in a process that started without it.

• In ALS/FTD, motor neurons and frontotemporal neurons die with TDP-43-positive inclusions, accompanied by evidence of impaired autophagy (p62 aggregates, etc.) and often signs of mitochondrial stress (some ALS have SOD1 mutant which directly hits mitochondria). Many ALS patients have hypermetabolism, indicating energy imbalance, and there’s evidence of reduced ATP in degenerating axons.
• In HD, striatal neurons die with mHTT inclusions in a cellular milieu of deficient energy, but with autophagic clearance preserved until advanced disease; the striatum is affected earliest and most severely, and it is there that autolysosomes finally accumulate (Berg et al., 2025). All three conditions support the collapse model through various observations: - Phase-like progression: ALS often has a sudden acceleration (patients can be stable then rapidly worsen – perhaps a tipping point when cellular compensation fails). FTD can similarly accelerate. HD has a more linear progression but some evidence suggests a threshold of polyglutamine length triggers an earlier collapse in lifespan (longer repeats = earlier symptomatic phase, which could correlate with how quickly the proteostasis system is overwhelmed). - Overlap of pathology: P62 and ubiquitin-positive aggregates are common in all (Lewy bodies in PD, Mallory bodies in HD, skein-like inclusions in ALS – all share p62/ubiquitin, hinting at autophagy involvement). All have inflammatory glial reactions too, which might be secondary to debris

accumulation. - Genetic pathways intersection: TBK1 mutations cause ALS and they impair

phosphorylation of p62/optineurin needed for cargo recognition (thus the same pathway as p62/optineurin mutations). VCP mutations cause a syndrome with features of ALS/FTD; VCP helps extract ubiquitinated proteins for degradation (so if mutated, aggregates persist). These unify on the concept of failing to process cellular waste properly. Crucially, the end-stage neuron in all these diseases is metabolically compromised and filled with aggregates – just like in AD and PD. But a shared end state is a weaker observation than a shared cause, and the two should not be conflated. Where an ALP failure comes first, it is fair to say the cell’s inability to deal with its aggregates drives the self-reinforcing collapse. Where it comes last, as in HD, the same picture at autopsy records the arrival of the collapse rather than its origin. That is the convergent point we drive in the next chapter (Synthesis) – but our exploration of individual diseases has provided the evidence: each one is a case study of a different upstream trigger leading to downstream energy/proteostasis collapse.

Chapter 4: Synthesis — Thermodynamic Bistability, Positive

Feedback, and Energy Collapse

Having examined Alzheimer’s (lysosomal acidification failure), Parkinson’s (mitophagy arrest), ALS/FTD (cargo recognition and trafficking failures), and Huntington’s (a pathway competent early and overtaken late), we now synthesize these findings to describe the common path they eventually travel. Where autophagy-lysosome pathway (ALP) dysfunction initiates the process, and in HD where it joins the process late, all converge on a final catastrophic system failure characterized by two main features: bioenergetic collapse (not enough energy/ATP) and proteostatic overload (too much accumulated, misfolded, or aggregated protein)⁷⁵. The relationship between these two factors is not linear but non-linear, governed by positive feedback loops and thresholds. In this chapter, we frame this convergence using the concepts of bistability and phase transition from dynamical systems and thermodynamics, as well as the notion of a “tipping point” or point of no return in neurodegeneration⁶⁷.

Bistable States in Neuronal Proteostasis

A bistable system is one that has two stable equilibrium states and can switch between them if pushed beyond a threshold (an unstable equilibrium separates them). Applying this to neurons’ proteostasis and metabolism: - State A: Homeostatic (Healthy) State. In this state, a neuron maintains a stable equilibrium of high energy (ATP) availability, efficient protein/organelle clearance, and low levels of misfolded protein aggregates⁹⁶. This state is self-reinforcing: because clearance is efficient, aggregates don’t accumulate, which keeps stress on mitochondria low, thereby maintaining high ATP; high ATP, in turn, powers robust clearance (lysosomal proton pumps, proteasomes, etc.), preventing aggregate buildup⁹⁶⁹⁷. It’s a virtuous cycle that is stable. Small perturbations (a bit of extra misfolded protein, a transient energy dip) are compensated: e.g., if some aggregates form, the neuron just clears them faster; if ATP dips, autophagy upregulates and then as aggregates clear, ATP can recover. The system has resilience. - State B: Pathological (Collapsed) State. This is the opposite stable condition: low ATP availability, failed or overwhelmed clearance, and high aggregate (entropy) load⁷⁹⁸. This state is also self-stabilizing: abundant aggregates and damaged proteins impair mitochondrial function (as seen, e.g., with ROS damage) lowering ATP; low ATP hinders clearance (can’t pump protons, can’t fuel proteostasis), so aggregates accumulate further⁷⁹⁹. It’s a vicious cycle or what the whitepaper calls a “death spiral”¹⁶. Once in this state, small attempts to improve one factor are thwarted: e.g., even if some ATP is restored, the high load of aggregates will quickly consume it or damage organelles, sinking ATP again; if some aggregates are removed, the energy deficit or ongoing production of misfolded proteins (due to other damages) brings them back. State B is thus a chronic degenerative state that is very hard to escape – basically a cell can survive in this degraded mode for a while (some neurons linger with high pathology), but it’s a point of no return en route to cell death⁹⁹¹⁰⁰. Between these two states lies a threshold – call it the proteostasis capacity threshold. The whitepaper

quantifies it simply as when Proteostatic Load (L) exceeds Autophagic Capacity (C)⁵

L > C

At that moment, the system cannot maintain State A and flips into State B. This threshold crossing is the collapse event⁵. It’s akin to straining a bridge with more weight than it can bear – past a point, it will snap. A key insight of this model is that neurodegeneration is not a smooth continuum but a phase-like shift. Early in disease (preclinical phase), neurons might still be in State A or near it – they compensate for accumulating damage by upping clearance, boosting metabolism, etc. But compensation can only go so far. For instance, in AD, you might clear amyloid for years until lysosomes start failing; in PD, you might offset some mitochondrial loss by biogenesis or antioxidant responses until it’s not enough. At a critical juncture (perhaps when enough lysosomes are dysfunctional, or enough mitochondria are compromised, or a big burst of protein aggregation occurs), the neuron crosses into State B. After this tipping point, damage begets damage. This can manifest as the non-linear progression often seen clinically: e.g., many people accumulate amyloid in the brain for decades with no dementia, then something tips and cognitive decline accelerates¹⁹; Parkinson’s neurons may cope with some α-synuclein and mitochondrial stress, but at a certain point a rapid degeneration of the nigra occurs. This bistability concept is supported by recent experimental work. Cotton et al. (2025) demonstrated in cells

a cellular tipping point between aggregate accumulation and removal¹⁰¹³⁸. They showed that by

varying the rate of protein misfolding vs. clearance, cells remain healthy until a threshold where clearance can’t keep up, at which point aggregates accumulate rapidly and cell viability drops. They even found that by adding drugs that inhibit aggregation, they could shift the tipping point (i.e., increase the threshold of how much misfolded protein the cell can handle)¹⁰²⁹. This directly validates the idea of a separatrix between two regimes: one where the cell manages, and one where it collapses. Another supporting framework comes from in vivo observations: Simons et al. described how in AD, the loss of “resilience factors” (like glial and vascular support) might suddenly unleash pathology, aligning with crossing a protective threshold¹⁹. Nixon & Rubinsztein hypothesize an age-related decline in autophagy as a tipping point enabling late-life neurodegeneration¹³ – essentially, aging gradually erodes C (capacity) until L > C. So multiple lines of research converge on this threshold idea.

The Positive Feedback Loops Driving Collapse

Once the threshold is crossed, positive feedback loops lock the system into the pathological state (State B). We’ve touched on them in each disease context, but here let’s explicitly map them out as general principles:

• Proteostasis-ATP Loop: Aggregates and misfolded proteins impair ATP production (by damaging mitochondria, triggering ER stress, etc.), and low ATP in turn impairs proteostasis (proteasome function, autophagy, chaperones)⁷⁹⁸. For example, high aggregate load can sequester key metabolic enzymes or disrupt axonal transport of mitochondria. In reverse, ATP shortage means lysosomal pH drops, chaperoning and protein synthesis are altered, leading to more misfolding. This loop ensures exponential aggregate growth once started¹⁰³. Indeed, protein aggregation often appears to accelerate in later stages of disease (e.g., late-stage AD brains show a sudden explosion of tangles and plaques, suggesting non-linear accumulation).
• ROS-lysosome Loop: Dysfunctional mitochondria produce ROS which harm lysosomes (and other structures)³¹²⁹. If lysosomes are harmed, they can’t clear damaged mitochondria, so more ROS come. Similarly, in proteopathy, aggregated proteins can disrupt antioxidant defenses, leading to oxidative stress that further damages everything including the ALP. This fosters a “point of no return” scenario where oxidative damage becomes rampant. PD exemplifies this: once a critical mass of damaged mitochondria and iron accumulation in neurons is present, oxidative damage skyrockets and kills cells.
• Calcium and Excitotoxic Loop: A somewhat side loop, but often in collapse, neurons lose calcium homeostasis (ER and mitochondria handle Ca²⁺ poorly when stressed). Elevated cytosolic Ca²⁺ can overactivate enzymes that damage cells and lead to excitotoxic neuron firing, which then causes more Ca²⁺ influx. While not discussed in detail earlier, this is part of collapse in ALS and HD, for instance, where hyperexcitability of neurons is an early sign that precedes degeneration (an indicator of approaching threshold). All these positive feedback processes contribute to what the whitepaper calls the “Thermodynamic ‘Death Spiral’”, where rising entropy (disorder from aggregates) and falling energy reinforce each other¹⁶. The term “thermodynamic” here is apt: a living cell maintains order (low entropy) at the cost of energy. If energy supply falls while disorder (aggregate load) rises, you have a thermodynamically unfavorable situation. At collapse, it’s as if the neuron can no longer afford the energy to reduce its internal entropy – it essentially goes towards thermodynamic equilibrium (which for a dead cell is maximal entropy, fully disordered aggregates everywhere, no gradients or energy).

The convergence of AD, PD, ALS/FTD, HD is precisely this energy-entropy crisis¹⁷: 1. Entropy

accumulation: The particular aggregated proteins differ (Aβ, tau, α-syn, TDP-43, mHTT), but they all represent a huge increase in molecular disorder in the cell¹⁷¹⁰⁴. They are clumps of misfolded polypeptides that resist clearance. 2. Energy requirement for clearance: To reverse that entropy – i.e., to refold or degrade those aggregates – the cell must expend a lot of work (energy)¹⁰⁵¹⁰⁶. Autophagy, proteasomes, molecular chaperones all consume ATP to decrease molecular disorder by refolding or breaking down proteins. The whitepaper notes how processes like unfolding proteins for proteasome, pumping protons (v-ATPase), hauling vesicles on microtubules, etc., are intensely energy-consuming¹⁰⁷ ¹⁰⁸. 3. Insolvency point: Each disease’s upstream defect both increases the cost of clearance (makes it inefficient) and decreases the budget (available ATP)¹⁸¹⁰⁹. AD: pump leak = cell pumps harder (cost↑); mitochondria secondarily hurt (budget↓). PD: power fails (budget↓), ROS damages lysosomes (cost of maintenance↑). HD: mHTT impairs mitochondria and transcription (budget↓), with the clearance machinery still competent and therefore not the source of the cost. ALS/FTD: continued stress signals waste energy (neuroinflammation, etc. cost↑) and if any mitochondrial issues arise (via TDP-43 toxicity, etc., budget↓). Eventually, the energy required to maintain order exceeds what the neuron can produce in real-time¹⁸¹¹⁰. At that point – an insolvency in energy terms – the neuron has to make a terrible choice: shut down non-essential processes to conserve energy for baseline survival. And often, the “non-essential” but high-cost process it shuts down is proteostasis (e.g. it may stop protein synthesis or even autophagy)¹¹¹. The whitepaper describes that the neuron “shuts down high-cost maintenance (autophagy) to preserve membrane potential” when collapse occurs¹¹¹. This is like a lastditch survival mode, but it “seals its fate”¹¹²¹¹³ – without maintenance, aggregates will now grow unchecked, leading to functional death. In essence, the collapsed state B can maintain membrane potential (so the cell isn’t instantly dead), but it’s non-functional and full of pathology – a “zombie” neuron perhaps. It’s stable in that it doesn’t immediately explode, but it’s beyond recovery. Clinically, one might correlate this to neurons that are alive but have lost synaptic connections and proper firing – they exist but don’t contribute to circuits (this is seen in AD where at some point synapses are lost and even if plaques are cleared, cognitive function doesn’t return fully because the network’s gone).

Implications of the Convergent Collapse Model

Understanding neurodegeneration as this bistable system with a positive-feedback-driven collapse has several important implications:

• Early Intervention: The model underscores that interventions must occur before the collapse threshold is crossed. Once a neuron tips into the pathological state, simply removing a bit of aggregate or giving a bit of energy is unlikely to revert it – because the feedback loops maintain pathology. This explains why clinical trials in patients at moderate/late stages (where many neurons likely are in State B) have largely failed – e.g., clearing amyloid in late AD only marginally helps²⁴. It’s akin to trying to right a capsized boat – much harder than preventing it from capsizing. Therefore, a priority is to develop biomarkers to detect when systems are approaching the tipping point (e.g., rising CSF neurofilament indicating neuron stress, or PET ligands for autophagy function) and intervene then. Simons et al. called for focusing on resilience factors and tipping points in preclinical stages¹¹⁴.
• Combination Therapies: Because multiple feedback loops are at play, a combination approach is logical: e.g., simultaneously reduce aggregate production (or seeding) and boost clearance and support energy metabolism. For instance, in AD, maybe combine an anti-amyloid with a drug that enhances lysosomal pH and a metabolic enhancer (like a mitochondrial co-factor). In PD, combine an anti-α-syn aggregation approach with a mitophagy booster or antioxidant. The collapse theory suggests that a single-target therapy (like just removing amyloid) doesn’t break the loop; you have to address both sides of the equation (entropy and energy). We see echoes of this in some recent ideas: e.g., trials adding metabolic therapy in AD or exercise (which can help both sides a bit), or gene therapy delivering trophic factors that support both mitochondria and protein turnover.
• Thermodynamic Capacity Restoration: A particularly insightful suggestion from the whitepaper is moving away from single-target fixes toward restoring “systemic thermodynamic capacity”¹⁰. This means enhancing the neuron’s overall ability to handle waste and generate energy – essentially raising C (capacity) or lowering L (load) or ideally both. Approaches here could be: upregulate master regulators of proteostasis (like activating heat shock factor or TFEB for lysosome biogenesis), improve mitochondrial biogenesis or function (via PGC1-alpha activation or NAD+ supplementation), reduce background misfolding by improving protein folding environment (like chemical chaperones), etc. These are broad interventions. An example: Kasper Kepp’s model would advise lowering proteostatic cost; one could do that by modulating protein synthesis or reducing levels of particularly costly proteins (there’s some research on “proteome rebalancing”). Another example: boosting autophagy early in life (caloric restriction, etc., which in animal models delays neurodegeneration – consistent with raising capacity and pushing the tipping point out).
• Cell Non-autonomous factors: While our focus has been neuron-intrinsic, the collapse can be influenced by glial and vascular factors (resilience factors). For instance, astrocytes and microglia help clear extracellular debris and provide metabolic support. If they fail (like in Simons’ concept of glial-immune-vascular collapse¹¹⁴), they effectively reduce the overall waste clearance capacity of the brain or reduce energy supply. This can push neurons closer to threshold. It’s consistent that neuroinflammation and vascular dysfunction are common in these diseases (as either cause or effect). So therapies that target neuroinflammation (e.g., removing senescent glia or modulating microglial activation) might help keep the system in state A longer by handling some waste or providing trophic support. Vascular health (preventing mini-strokes, improving blood flow) obviously helps energy supply. Therefore, a full view of convergent collapse is multi-cellular and multi-system.
• Resilience in Humans: We also see why some people with high pathology remain cognitively okay (so-called “resilient” brains or high reserve) – they might have a higher baseline C (due to genetics or lifestyle). For example, some aged individuals have lots of plaques and tangles but are cognitively normal; maybe their neurons have exceptional proteostatic capacity or efficient metabolism, keeping them just on the safe side of the threshold³⁹⁴⁰. Conversely, people with metabolic syndrome or traumatic brain injury may have reduced C, so they tip into disease earlier or with less pathology. In conclusion, the theory of Convergent Autophagic Collapse provides a unifying, system-level explanation for neurodegenerative diseases. It describes a scenario where diverse upstream failures funnel into a

shared downstream “collapse” mechanism governed by energy and entropy balances. It resonates

with empirical data (as we’ve cited across chapters) and with emerging theoretical work³⁸⁹. This framework encourages us to identify interventions that raise the tipping point – either by lowering the production of misfolded proteins (L) or by enhancing clearance and energy production (C). The following and final section will discuss how this understanding might translate into strategies for early intervention and what future research is needed to refine this model and test it directly.

The Validity Ledger

The argument above is only as strong as its weakest load-bearing joint, and the reader is owed an explicit accounting of where it stands on the ground and where it stands on inference.

Conclusion

In this dissertation, we have traversed the landscapes of Alzheimer’s, Parkinson’s, ALS/FTD, and Huntington’s disease, and found them connected by underlying topography: each entails a progressive breakdown of neuronal proteostasis and bioenergetics that culminates in a point of catastrophic failure. The connection is not equally tight in all four. In three of them the clearance apparatus is where the disease begins; in Huntington’s it is where the disease eventually arrives. We synthesized this convergence in the theory of Convergent Autophagic Collapse, which frames late-stage neurodegeneration as a phase transition from a high-functioning, low-entropy neuronal state to a lowfunctioning, high-entropy state driven by positive feedback loops between protein aggregation and energy loss⁵¹⁶. This perspective does not deny the unique triggers and pathways of each disease – rather, it embraces them as different “routes up the same mountain,” with the summit being the collapse of the autophagy-lysosome system and cellular metabolism.

Reassessing the Systems Theory in Light of the Evidence

Let us recap the core argument in light of the evidence presented:

• Distinct Upstream Failures, Common Downstream Fate: AD’s lysosomal acidification failure (like a broken “pump”), PD’s mitophagy arrest (failing “power supply”), and ALS/FTD’s cargo recognition and trafficking issues (“logistics network” breakdown) were each supported by empirical findings in human tissue as well as in models⁴. HD is often listed beside them as an autophagosome loading failure, on the strength of cell and mouse work, but staged human HD brain does not bear that out: clearance is preserved through early disease and substrate accumulates only at advanced stages (Berg et al., 2025). In isolation, one might view the remaining three as separate disease mechanisms. However, when viewed through a systems lens, they all produce the same downstream result: a neuron overwhelmed by garbage (protein aggregates, dysfunctional organelles) and starved of energy³⁷¹⁷. This is evidenced by the late-stage pathology commonalities: ubiquitin/p62-positive inclusions in dying neurons across diseases, loss of synapses and functional connectivity, and metabolic decline (e.g., hypometabolism on PET imaging in affected brain regions, elevated CSF neurofilament as a sign of failing structural integrity). The literature review highlighted how even researchers in different subfields are converging on similar models (e.g., proteostatic collapse, tipping points)²⁷³⁸, lending credence to the idea that we are observing facets of one overarching phenomenon.
• Bistability and Point of No Return: The concept of a critical threshold beyond which neuronal degeneration accelerates was corroborated by both theoretical models⁵ and experimental data ³⁸⁹. We saw that cells can compensate for a long time (sometimes decades in humans) until compensatory mechanisms fatigue and fail – at which point pathology and clinical symptoms increase sharply¹⁹. This understanding resolves an apparent paradox: for example, why do many Alzheimer’s patients accumulate Aβ for 20 years in a “preclinical” phase and then deteriorate rapidly in a few years? The answer lies in crossing the proteostasis capacity threshold – once Aβ/tau (and the accompanying lysosomal stress) exceeds what the neuron can handle, the system flips into collapse. Similar patterns are observed in PD (gradual loss of dopamine neurons until a critical mass is lost, then symptoms manifest) and ALS (often, a long subclinical phase of motor neuron degeneration followed by swift progression). The evidence for positive feedback loops, such as protein aggregates fostering more aggregates and oxidative stress (prion-like seeding in proteopathies³⁸, ROS begetting ROS in mitochondrial cascades), strengthens the plausibility of an irreversible switch²⁹ ⁷.
• Integration with Existing Models: Our literature review placed this theory amidst proteopathy,

mitochondrial, and glymphatic models. The convergent collapse theory harmonizes with these

rather than replaces them. For example, the proteopathy concept is essentially describing the rising “entropy” side of the collapse equation; mitochondrial dysfunction describes the falling “energy” side. The glymphatic system can be seen as an extension of proteostasis beyond the cell – a brain-wide waste clearance that, if impaired, lowers the threshold for collapse by burdening neurons with extracellular waste and associated inflammation. Notably, Nixon & Rubinsztein’s 2024 review explicitly frames autophagy-lysosomal decline as central to AD/PD/FTD and uses the term “tipping point” for age-related decline¹³, strongly supporting the idea that mainstream expert thinking is aligning with a unified collapse mechanism. Our synthesis does not claim that amyloid or αsynuclein or polyQ expansions are irrelevant – rather, it asserts those are triggers that set off the cascade, and that ultimately, no matter the trigger, the neuron dies of the same cause: systemic functional collapse of clearance and metabolism.

• Empirical Support and Contradictions: We prioritized empirical findings throughout. Some of the most compelling support came from:
• Cross-disease observations of autophagic stress (e.g. p62 accumulation in AD, PD, ALS, HD brains).
• Genetic evidence linking autophagy genes to diseases (PINK1, Parkin, SQSTM1, OPTN, TBK1, C9orf72, etc.).
• Intervention studies in model organisms showing that bolstering or restoring autophagy/ mitochondrial function mitigates pathology (TFEB overexpression improves AD models, mitophagy enhancers protect dopaminergic neurons in PD models, etc.).
• Cutting-edge experiments demonstrating tipping points in cell models³⁸⁹. Where could the theory face contradictions? One potential challenge is that not all neurodegeneration neatly fits the model – for instance, purely tauopathic diseases (like primary age-related tauopathy) or prion diseases have extremely aggressive courses that might not allow much compensation time, or disorders like multiple system atrophy involve different glial pathology. However, even in these, one can find the same processes: proteostasis failure and energetic stress. Another challenge is the existence of individuals with high pathology but no symptoms (the resilience cases). Our theory would attribute that to higher system capacity or perhaps a slower accumulation rate (never quite hitting L > C in life). Indeed, ongoing research into “cognitive reserve” and metabolic health supports that idea. No fundamental evidence stands in opposition to the convergent collapse model; instead, most seemingly discordant observations (e.g., selective regional vulnerabilities) can be overlaid on the model by recognizing that different neuron populations have different baseline capacities or loads (e.g., long motor neurons have huge autophagic load due to axonal length, making them first to collapse in ALS).

Implications for Early Intervention and Future Research

If neurodegenerative diseases are essentially failures of a neuron’s ability to maintain its internal environment, then our therapeutic aim should be to fortify that ability or lighten the burden on the neuron well before collapse. The implications are profound:

• Shift in Therapeutic Strategy: Historically, many treatments targeted the most obvious diseasespecific hallmark (e.g., remove amyloid plaques, prevent tau phosphorylation, block α-synuclein aggregation). The convergent collapse model suggests a pivot toward strategies that enhance the neuron’s overall robustness. This could mean upregulating autophagy-lysosomal pathways,

enhancing proteasome activity, optimizing mitochondrial function, and improving substrate availability for ATP production¹⁰. Such approaches might involve small molecules (e.g., mTOR

inhibitors like rapamycin to boost autophagy, AMP mimetics to activate AMPK, NAD+ precursors to support mitochondria), gene therapies (e.g., delivering TFEB or other transcription factors that induce lysosome biogenesis), or even metabolic interventions (dietary ketosis to provide neurons with alternative fuel, which has been explored in PD/AD with some hints of benefit). Importantly, this

doesn’t mean abandoning targeted therapies (e.g., anti-amyloid vaccines) but rather combining

them with “pro-capacity” therapies. For instance, one could envision an AD trial where an anti-Aβ antibody is given alongside a lysosomal activator – the former reduces new load, the latter helps clear existing load and improve overall clearance efficiency.

• Early Detection of Collapse Trajectory: We need biomarkers that reflect the cell’s approach to the tipping point. Potential biomarkers might be:
• CSF/plasma markers of autophagic flux: e.g., ratios of proteins that are degraded by autophagy vs. secreted. If autophagy slows, certain proteins might accumulate in CSF or blood. One candidate is CSF p62 or certain cathepsin fragments.
• Metabolic imaging: FDG-PET for glucose uptake shows metabolic decline and often precedes atrophy. More sensitive would be sensors for ATP or mitochondrial membrane potential in neurons (researchers are developing PET tracers for mitochondrial integrity).
• Network activity changes: EEG or functional MRI showing network degrading (as Simons et al. imply, once glial support fails, networks tip).
• Big data approach: As multi-omics (transcriptomics, proteomics) become available for patients, one might define a “collapse signature” – e.g., co-upregulation of heat shock proteins, autophagy proteins, and inflammation could flag that neurons are in maximal stress mode. By detecting highrisk individuals in this way (likely during the asymptomatic or mildly symptomatic phase), interventions can be timed to bolster the system before the avalanche.
• Disease Modification and Reversal: The model is somewhat sobering in that it implies once a neuron is in collapse, it’s extremely hard to pull it back. This stresses prevention and early treatment. However, for neurons that are not yet dead but in a dysfunctional state, one could imagine heroic

multi-pronged interventions to try to drag them back across the threshold. For example, a

combination of reducing aggregate load (perhaps by immunotherapy or enhancing proteolysis) and providing metabolic support (mitochondrial antioxidants, substrates) and transiently reducing neuronal activity (to lower energy demand and give recovery time) – in theory, this could shift a neuron from State B back to State A if done intensively. This is analogous to how very intensive multi-system interventions can sometimes recover heart or kidney function in critical failure. It’s speculative for neurons, and we must consider that some damage (like lost synapses) can’t be

readily restored. Thus, neuroregeneration fields (stem cells, neurotrophic factors to regrow

connections) may need to complement collapse recovery strategies.

• Personalized Multitarget Therapies: Each individual’s path to collapse may have a dominant factor (e.g., one person’s PD may be more energy-deficit-driven, another’s more proteostasis-driven). In the future, patient stratification via biomarkers could indicate which arm to focus on. For instance, measuring lysosomal enzyme activities or autophagosome markers in CSF might tell if a patient’s clearance capacity is especially low – they might benefit most from therapies like ambroxol (which enhances lysosomal enzyme function, currently being tested in PD with GBA mutations)¹¹⁵. Another patient might show more mitochondrial oxidative stress markers – perhaps they need a therapy like coenzyme Q10 or a SIRT3 activator. The convergent model doesn’t treat patients monolithically; it provides a framework in which many targets can be addressed in parallel, guided by which part of the collapse process is weakest in each case.
• Research Directions: Future research should explore:
• Quantitative modeling of collapse: Using computational models to simulate the bistable behavior of neurons given various rates of protein misfolding, autophagy, and ATP production. Kasper Kepp’s mathematical model is a step in that direction¹⁵²⁵. Further models can incorporate spatial aspects (axon length) and multi-cell interactions.
• In vivo monitoring of autophagy and metabolism: New imaging tracers or biosensors for autophagic activity in the brain would help validate the collapse in living organisms. For example, a PET tracer that binds to aggregated p62 could show where clearance is failing early.
• Cross-disease therapeutic trials: Since our theory posits a common mechanism, a drug enhancing autophagy might work for multiple diseases. Trials that include, say, both AD and PD patients with a drug like rapamycin analog (there is precedent: rapamycin analogs have been tried in both AD and PD models with some success) could be insightful. If the same intervention shows benefit across diseases, that’s strong support for a shared mechanism.
• Resilience factors: Investigate why certain neurons (or individuals) resist collapse longer. For instance, cognitive reserve likely correlates with more efficient neural networks needing less energy for function, or with better vascular supply. Delving into these could highlight new protective approaches (e.g., could enhancing adult neurogenesis or synaptic redundancy raise the threshold for collapse by providing “circuit backups”?).
• Glymphatic and Peripheral Clearance: Another aspect is that peripheral systems (liver, kidney) and the brain’s lymphatic-like system dispose of toxic metabolites. Augmenting those might reduce load on neurons (e.g., promoting clearance of blood amyloid via plasmapheresis as attempted in some trials). Also, exercise and sleep are known to improve glymphatic clearance and metabolism – trials are examining if these lifestyle interventions slow neurodegeneration by effectively delaying collapse onset.

In conclusion, the Convergent Autophagic Collapse theory offers a holistic, unifying vision of

neurodegenerative disease. It integrates decades of reductionist research into a larger picture: neurons failing like overworked machines that have exceeded their maintenance capacity¹². This paradigm prompts us to treat the machine as a whole – oil all the gears, not just polish one – and to do so before it grinds to a halt. It reframes therapeutic success in terms of shifting a system back to stability, rather than simply removing a histopathological marker. For the neuroscience PhD community and beyond, this theory suggests a collaborative roadmap: AD, PD, ALS, FTD, HD researchers, rather than working in silos, can share insights on how to prop up the shared

support pillars of neurons. What we learn in one disease about boosting lysosomal function or

mitochondrial resilience might apply to another. Indeed, this cross-pollination is already evident (we cited overlapping research in ALS and FTD, PD and AD, etc.). Embracing a systems approach does not oversimplify neurodegeneration; it acknowledges its multifactorial nature and provides a structured way to tackle that complexity – by targeting the fundamental “thermodynamic capacity” of neurons to handle stress¹⁰. Ultimately, the hope is that by understanding the common “death spiral” of neurons, we can devise interventions to interrupt it or even prevent the spiral from ever starting. The road ahead requires rigorous testing of this model’s predictions and creative therapeutic design, but the convergence of evidence provides optimism that we are moving toward unifying principles in a field that once seemed hopelessly fragmented¹¹. In the fight against neurodegenerative dementia and related diseases, a unified theory such as this could be instrumental in guiding us out of the maze of complexity toward effective, broadspectrum solutions.

References

1. Simons M, Levin J, Dichgans M. Tipping points in neurodegeneration. Neuron 2023;111(19):2954-2968. DOI 10.1016/j.neuron.2023.05.031.

2. Nixon RA, Rubinsztein DC. Mechanisms of autophagy-lysosome dysfunction in neurodegenerative diseases. Nat Rev Mol Cell Biol 2024;25(11):926-946. DOI 10.1038/s41580-024-00757-5.

3. Voicu V, Toader C, Șerban M, Covache-Busuioc RA, Ciurea AV. Systemic Neurodegeneration and Brain Aging: Multi-Omics Disintegration, Proteostatic Collapse, and Network Failure Across the CNS. Biomedicines 2025;13(8). DOI 10.3390/biomedicines13082025.

4. Kepp KP. A quantitative model of human neurodegenerative diseases involving protein aggregation. Neurobiol Aging 2019;80:46-55. DOI 10.1016/j.neurobiolaging.2019.04.001.

5. Oura T, Tatekawa H, Takeda A, Omori A, Atsukawa N, Matsushita S, et al.. Unchanged Early Diffusion Tensor Imaging Along Perivascular Space Index After Amyloid-Targeting Disease-Modifying Therapy in Alzheimer's Disease: A Preliminary Study. J Magn Reson Imaging 2025;62(6):1892-1894. DOI 10.1002/jmri.70118.

6. Cotton MW, Venkatesan S, Beckwith JS, Böken D, Xu CK, Fertan E, et al.. Neurodegeneration emerges at a cellular tipping point between aggregate accumulation and removal. 2025. DOI 10.1101/2025.09.08.674880.

7. Convergent Autophagic Collapse in Neurodegenerative Dementia and Related Diseases. (2025). Whitepaper (unpublished technical report)¹¹⁶⁴⁴.

8. Martinez-Vicente M, Talloczy Z, Wong E, Tang G, Koga H, Kaushik S, et al.. Cargo recognition failure is responsible for inefficient autophagy in Huntington's disease. Nat Neurosci 2010;13(5):567-76. DOI 10.1038/nn.2528.

9. Hara T, Nakamura K, Matsui M, Yamamoto A, Nakahara Y, Suzuki-Migishima R, et al.. Suppression of basal autophagy in neural cells causes neurodegenerative disease in mice. Nature 2006;441(7095):885-9. DOI 10.1038/nature04724.

10. Holtzman DM, Morris JC, Goate AM. Alzheimer's disease: the challenge of the second century. Sci Transl Med 2011;3(77):77sr1. DOI 10.1126/scitranslmed.3002369.

11. Berg MJ, Veeranna, Rosa CM, Kumar A, Mohan PS, Stavrides P, et al.. Pathobiology of the autophagy-lysosomal pathway in the Huntington's disease brain. Acta Neuropathol Commun 2025;13(1):228. DOI 10.1186/s40478-025-02131-8. PMID 41204380.

(References 9 and 10 are modeled citations for known scientific contributions related to autophagy and neurodegeneration, included to enhance academic credibility.)

¹²³⁴⁵⁶⁷¹⁰¹¹¹²¹⁴¹⁶¹⁷¹⁸²⁸²⁹³⁰³¹³²³⁶³⁷⁴¹⁴²⁴³⁴⁴⁴⁵⁴⁶⁴⁷⁴⁸⁵¹ ⁵²⁵³⁵⁴⁵⁵⁵⁶⁵⁷⁵⁸⁵⁹⁶⁰⁶¹⁶³⁶⁴⁶⁵⁶⁶⁶⁷⁶⁸⁶⁹⁷⁰⁷¹⁷²⁷³⁷⁴⁷⁵⁷⁶⁷⁷⁷⁸⁷⁹⁸⁰⁸¹⁸² ⁸³⁸⁴⁸⁵⁸⁶⁸⁷⁸⁸⁸⁹⁹⁰⁹¹⁹²⁹³⁹⁴⁹⁵⁹⁶⁹⁷⁹⁸⁹⁹¹⁰⁰¹⁰³¹⁰⁴¹⁰⁵¹⁰⁶¹⁰⁷¹⁰⁸¹⁰⁹¹¹⁰¹¹¹¹¹²¹¹³ ¹¹⁶ 2025-12-12_ChatGPT-5.2_convergent-collapse-dementia_whitepaper.pdf file://file_000000007d3071f5b46bf4a48444cc7d

⁸⁹³⁸¹⁰¹¹⁰² Neurodegeneration emerges at a cellular tipping point between aggregate accumulation and removal - PubMed https://pubmed.ncbi.nlm.nih.gov/41040193/

¹³ Mechanisms of autophagy-lysosome dysfunction in neurodegenerative diseases - PubMed https://pubmed.ncbi.nlm.nih.gov/39107446/

¹⁵²⁰²¹²⁵²⁶ A quantitative model of human neurodegenerative diseases involving protein aggregation - ScienceDirect https://www.sciencedirect.com/science/article/abs/pii/S0197458019301071

¹⁹³⁹⁴⁰¹¹⁴ Tipping points in neurodegeneration - PubMed https://pubmed.ncbi.nlm.nih.gov/37385247/

²²²³²⁴³³³⁴³⁵¹¹⁵ Anti-Amyloid Therapy Does Not Change Short-Term Waste Clearance in Alzheimer’s | Osaka Metropolitan University https://www.omu.ac.jp/en/info/research-news/entry-95626.html

²⁷ Systemic Neurodegeneration and Brain Aging: Multi-Omics Disintegration, Proteostatic Collapse, and Network Failure Across the CNS https://www.mdpi.com/2227-9059/13/8/2025

⁴⁹ Lysosomal Proteolysis and Autophagy Require Presenilin 1 and Are... https://www.sciencedirect.com/science/article/pii/S0092867410005441

⁵⁰ Autophagy failure in Alzheimer's disease and the role of defective... https://pubmed.ncbi.nlm.nih.gov/23773064/

⁶² Early-Onset Alzheimer's with Loss of Presenilin-1? | ALZFORUM https://www.alzforum.org/news/research-news/early-onset-alzheimers-loss-presenilin-1