Proteostasis collapse

Proteostasis collapse is the progressive age-related failure of the systems that fold proteins correctly, refold them when they are damaged, and destroy them when they cannot be repaired. Proteostasis — protein homeostasis — is maintained by a network of several hundred chaperones, the ubiquitin–proteasome system, and lysosomal degradation, working under the control of inducible stress responses.¹ Its loss is a primary hallmark in the Hallmarks of aging framework and the mechanistic link between aging and the major neurodegenerative diseases.

§The proteostasis network

1. 1962
   Heat-shock response observed
   Ferruccio Ritossa notices new chromosomal puffing patterns in Drosophila salivary glands after a temperature shift, the first sign of an inducible protein-protection programme.
2. 1972
   Sequence determines structure
   Christian Anfinsen receives the Nobel Prize in Chemistry for showing that a protein's amino acid sequence encodes its folded conformation, framing misfolding as a solvable problem.
3. 2004
   Ubiquitin system recognized
   Aaron Ciechanover, Avram Hershko and Irwin Rose share the Nobel Prize in Chemistry for the discovery of ubiquitin-mediated protein degradation.
4. 2008
   Proteostasis named
   Balch, Morimoto, Dillin and Kelly propose 'proteostasis' for the integrated network and argue it is a druggable target across folding diseases.
5. 2009–2010
   Collapse mapped in a model organism
   Work in C. elegans locates the loss of folding capacity in early adulthood and shows that hundreds of proteins become insoluble with age.

A newly made polypeptide must reach one conformation out of an enormous number of possible ones, in a cytoplasm containing several hundred milligrams of protein per millilitre. Molecular chaperones — the HSP70 and HSP90 families, the small heat-shock proteins, and the chaperonin complex that folds actin and tubulin — bind exposed hydrophobic surfaces and prevent inappropriate association while folding proceeds. Only part of the network does folding work; the rest handles triage and disposal. Predicting a native structure from sequence, the task machine-learning models now perform well, is a different problem from predicting how a chain reaches that structure or how often it fails to, which is what the network exists to manage.

Two degradation systems handle proteins that cannot be rescued. The ubiquitin–proteasome system tags substrates with polyubiquitin chains through a cascade of activating, conjugating and ligase enzymes, and feeds them into the 26S proteasome, which unfolds and cleaves them. Lysosomal degradation handles bulk cargo and organelles by Autophagy, and handles individual proteins bearing a specific pentapeptide motif through chaperone-mediated autophagy, in which HSC70 delivers substrates to the lysosomal receptor LAMP2A for direct translocation.

Three inducible stress responses supervise the network in different compartments: the cytosolic heat-shock response, governed by the transcription factor HSF1; the unfolded protein response of the endoplasmic reticulum, running through IRE1, PERK and ATF6; and the mitochondrial unfolded protein response, which couples proteostasis to mitochondrial state.

§What collapses, and when

The most striking result on timing came from Caenorhabditis elegans, where the capacity to buffer misfolded proteins falls sharply in early adulthood rather than degrading gradually across life. Morimoto's group showed that this collapse coincides with reproductive maturity and precedes any visible aging phenotype, suggesting proteostasis capacity is actively downregulated once reproduction is under way rather than simply worn out.² Long-lived insulin/IGF-1 signalling mutants of the kind Cynthia Kenyon's laboratory characterized delay that collapse, and the FOXO-family transcription factor their longevity depends on drives expression of chaperones and other network components. A companion finding is that hundreds of proteins become detergent-insoluble in aging worms, well beyond the handful associated with named aggregation diseases.³

In mammals the picture is less crisp but consistent in direction. HSF1-driven induction of chaperones in response to heat or oxidative stress weakens with age, and that response is the mechanism usually invoked for sauna bathing, where the human mortality data are observational. Proteasome activity declines in several tissues, though not uniformly. Chaperone-mediated autophagy declines markedly with age in rodent liver, and restoring LAMP2A expression in old mice preserved proteolytic activity and hepatic function.⁴ That experiment is one of the cleaner demonstrations that the decline is not merely correlated with aging but contributes to organ dysfunction.

Terminology

"Collapse" describes the abrupt loss of buffering capacity seen in invertebrate models. In mammals the change is gradual and tissue-specific, and the word is used loosely. Neither usage implies that all proteostasis functions fail together.

§Aggregation and neurodegeneration

The clinical face of proteostasis failure is protein aggregation in the nervous system. Amyloid-β and tau in Alzheimer's disease, α-synuclein in Parkinson's disease and Lewy body dementia, TDP-43 in most amyotrophic lateral sclerosis and frontotemporal dementia, and polyglutamine-expanded huntingtin in Huntington's disease all form ordered assemblies that spread between cells in a templated, prion-like manner. Age is by a wide margin the largest risk factor for all of them, which is the argument for treating them as manifestations of an aging process rather than as independent diseases — the core claim of the geroscience hypothesis.

Whether the aggregates themselves are the toxic species is unsettled. Soluble oligomers correlate better with neuronal dysfunction than mature fibrils in several systems, and inclusion body formation may sequester damaging material rather than cause damage. The therapeutic record is sobering: decades of amyloid-clearing programmes produced two anti-amyloid antibodies, lecanemab and donanemab, approved in the United States in the 2020s. Both slow cognitive decline modestly, at the cost of amyloid-related imaging abnormalities including brain swelling and microhaemorrhage. The size and clinical meaning of that benefit remain contested among neurologists.

§Long-lived proteins

Not all proteins are replaced. Crystallins in the eye lens are synthesized before birth and never turned over, which is why they accumulate racemized and glycated residues across a lifetime and eventually scatter light as cataract. Nuclear pore complex scaffold proteins, some histones, and elastin and collagen in the extracellular matrix persist for years to decades.⁵ For this fraction of the proteome, no amount of chaperone capacity helps; the damage is chemical, spontaneous, and irreversible without replacement of the whole structure.

This category is a distinctive problem for aging biology because it is not a regulatory failure that can be corrected by restoring a signalling pathway. Advanced glycation end-products cross-link long-lived extracellular matrix proteins and stiffen arteries and skin; breaking those cross-links pharmacologically has been attempted and has not produced a durable clinical result.

§Interventions and evidence

Interventions divide into three groups. Raising folding capacity: HSF1 activators and heat-shock protein co-inducers, of which arimoclomol is the best-known example — it failed a phase 3 trial in ALS before being approved in the United States for a rare lysosomal storage disease. Raising degradation capacity: proteasome activation extends lifespan under proteotoxic stress in nematodes, and enhancing autophagy with Rapamycin, spermidine or dietary restriction improves outcomes in mouse models of aggregation disease. Reducing the load: antisense oligonucleotides and RNA interference that lower production of the aggregating protein, an approach now in clinical use for transthyretin amyloidosis and in trials for Huntington's disease.

Comparative biology offers a suggestive parallel. Naked mole-rats maintain higher proteasome activity and more accurate translation than mice, and their proteins are more resistant to unfolding — one of several traits invoked in explanations of negligible senescence.⁶ Human embryonic and induced pluripotent stem cells likewise sustain unusually high proteasome activity, which falls on differentiation. That observation connects proteostasis to stem cell decline and to reprogramming, since restoring a more pluripotent-like state also restores proteostatic capacity in culture.

§Open problems

There is no validated way to measure proteostatic capacity in a living human. Chaperone transcript levels, proteasome activity in blood cells, and circulating aggregate species have all been proposed and none has been qualified as an aging biomarker. Without such a measure, the hypothesis that boosting proteostasis would slow aging cannot be tested except through disease endpoints in populations already sick.

The direction of causality between proteostasis and the other hallmarks is also unresolved. Failing proteostasis produces damaged proteins that trigger senescence and inflammatory signalling; senescent cells and inflammation in turn impair chaperone induction. Nor is more degradation obviously better: excessive autophagy and excessive proteasome activity both cause pathology, and the network's set points appear to be tuned rather than maximized. Whether an intervention could raise capacity in old tissue without disturbing that tuning is untested in any mammal.

§See also

• Hallmarks of aging
• Autophagy
• Cellular senescence
• Mitochondrial dysfunction in aging
• Geroscience hypothesis
• Negligible senescence

§References