3Protein Aggregation

Publication Details

Key Points Raised by Individual Speakers

  • Protein aggregation is a common characteristic of many neurodegenerative diseases. The aggregates and/or oligomers appear to be toxic, causing injury or death to cells. In general, the greater the degree of aggregation, the greater is the severity of disease.
  • Cells have specific organelles and other cellular components to clear protein aggregates, including proteasomes and lysosomes. Proteasomes are used to degrade smaller aggregates, whereas lysosomes are used for larger ones. The actions of proteasomes and lysosomes are controlled by a range of proteins, including ubiquitinating ligases, deubiquitinating enzymes, and chaperone proteins.
  • Therapies that stimulate the cell's normal clearance mechanisms are likely to show promise for treating neurodegenerative disease, as are therapies that prevent protein aggregation in the first place.

Protein misfolding and other errors in protein generation occur frequently within cells, and the cell has evolved a range of mechanisms to ensure proper folding and to eliminate aggregated or otherwise damaged proteins. A common characteristic of many neurodegenerative diseases is protein aggregation due to a failure of clearance mechanism(s).

The neurodegenerative disorders featured in the workshop share pathological accumulation in the brain of abnormal protein aggregates or inclusions that contain misfolded proteins. These diseases include Alzheimer's disease, Parkinson's disease, amyotrophic lateral sclerosis (ALS), Huntington's disease, dementia with Lewy bodies, frontotemporal diseases, and multiple system atrophy. The same protein can be found in more than one disease; mixed proteinopathies are highly prevalent (see Table 2-1). Proteins within the aggregate possess altered physical properties that are responsible for their misfolding. For example, the mutated huntingtin protein found in Huntington's disease contains excess repeats of the amino acid glutamine (Trottier et al., 1995). The alteration in structure leads to interaction with other proteins and subsequent aggregation that is dependent on age and length of the repeats (Voisine et al., 2010). In other cases, the protein itself may not necessarily be mutated, but it may be produced to excess by disease-related upregulation in protein expression. The sheer amount of additional protein being produced may tilt the balance toward misfolding, said Richard Morimoto of Northwestern University.

An abundance of misfolded proteins appears to be toxic to cells, leading to their injury and death. A disease's severity often correlates with the expression levels of the protein (Voisine et al., 2010; Williams et al., 2006). The toxic accumulation occurs in different parts of the brain and can be in the nucleus, cytoplasm, or extracellular space. Protein aggregation not only has been identified in humans with disease, but also has been replicated in biological model systems, such as in C. elegans, and with pure protein, according to Morimoto. Although not discussed in detail at the workshop, there is debate about whether certain forms of aggregates and/or components of aggregation mechanisms are neutral or even protective rather than toxic (see, for example, Selkoe, 2008; Spires-Jones et al., 2009; Williams and Paulson, 2008; Wolfe and Cyr, 2011).

This chapter summarizes workshop presentations about different mechanisms by which cells clear toxic protein aggregates and, for each, discusses potential therapeutics based on those mechanisms. Because protein aggregation is common across many neurodegenerative diseases, these therapeutic approaches might benefit more than one disease.


Proteostasis, or protein homeostasis, is the collective term used to describe a variety of cellular processes designed to minimize damage from altered, misfolded, and otherwise damaged proteins. Morimoto stressed the importance of proteostasis for ensuring cellular health by proper folding of proteins into native, soluble state instead of improper folding that leads to protein aggregation and cell toxicity. The so-called proteostasis network relies on chaperone proteins that guide protein folding, beginning with protein synthesis. Other chaperone proteins refold denatured proteins. The proteostasis network also relies on clearance mechanisms and detoxifying enzymes to degrade an excess of improperly folded proteins.

Aging, disease-associated mutations, polymorphisms, and energetic deficits place high demands on the proteostasis network, illustrated in Figure 3-1. Once the cell can no longer keep up with the heightened demand, the balance is shifted toward toxic accumulation of misfolded proteins. Morimoto pressed for better understanding of the upstream signaling events within the proteostasis network in order to prevent protein misfolding and aggregation. Preventing protein misfolding and aggregation is an appropriate therapeutic strategy, he said.

FIGURE 3-1. Proteostasis: balance between function and dysfunction.


Proteostasis: balance between function and dysfunction. (1) In the optimal state, molecular chaperones, clearance mechanisms, and detoxifying enzymes keep in check mutations, biosynthetic errors, energy deficits, and protein damage. (2) Aging intensifies (more...)


One significant component of the proteostasis network is carried out by proteasomes, multisubunit complexes within the nucleus and cytoplasm that degrade soluble protein, according to Alfred Goldberg of Harvard Medical School. The vast majority of damaged proteins are normally degraded by proteasomes, and so too are smaller protein aggregates. The process begins with ubiquitin ligases that attach ubiquitin to damaged proteins or aggregates, marking them for proteasomal destruction. A chain of at least four ubiquitin molecules must be attached for the process to proceed. The bonding between the protein or small aggregate and ubiquitin leads the conjugate to attach to the opening of the proteasome, where the damaged protein is unwound and translocated through a small gate into the proteolytic core of the proteasome. There it is quickly digested into amino acids. If the binding and unfolding are not accomplished within seconds, deubiquitinating enzymes, such as USP14, inhibit proteasomal degradation by stripping away the ubiquitin tags. That releases the substrate from the proteasome and precludes degradation. Selectivity of the degradation process is conferred by a range of highly specific ubiquitin ligases that tack ubiquitin onto the damaged protein. There are at least 30 to 40 ubiquitin ligases of the E2 class and more than 650 of the E3 class. The combination of classes allows enormous opportunities for selectivity in the process of targeting proteins for elimination, Goldberg explained.

Another degradation pathway is via endosome engulfment and translocation to lysosomes, organelles that are larger than proteasomes. Goldberg described his work to identify the enzyme Nedd4, a membrane-associated ubiquitin ligase, which plays a major role in the clearance of α-Synuclein via the endosomal/lysosomal pathway. Nedd4 is found in neurons containing Lewy bodies. Its downregulation increases α-Synuclein content (Tofaris et al., 2011). (α-Synuclein can be degraded by other proteostasis mechanisms as well [Cuervo et al., 2004].) Finally, he pointed out that some protein aggregates tagged by ubiquitin and slated for destruction may be too large for the proteasome, in which case the complex forms a cork-like structure, clogging the proteasome and thereby leading to greater protein accumulation. How protein accumulation in the extracellular space is removed is unknown, Goldberg noted. Another area of insufficient knowledge is of proteasomal function specifically in neurons. Most of the research is done on other cell types, he acknowledged in response to questions.

In terms of therapeutic opportunities, Goldberg argued for stimulating proteasomal degradation by (1) identifying and tapping into the specific ubiquitin enzymes necessary for targeting and degrading proteins by various pathways; and (2) inhibiting deubiquitinating enzymes to prevent dissociation of protein-ubiquitin conjugate from the proteasome. Goldberg referred to the research identifying a small molecule that inhibits the specific deubiquitinating enzyme USP14 (Lee et al., 2010a). In this study, researchers found that the small molecule accelerated in vitro the degradation of neurodegenerative disease-related proteins tau and TDP-43 (Lee et al., 2010a). He observed that deubiquitinating enzymes are more amenable to drug targeting than are the ubiquitin ligases because they are cysteine proteases, which have a defined mechanism of action and highly specific targets.


Autophagy is a dynamic process of bulk degradation of cellular organelles and proteins; this includes proteins that are soluble as well as those that form into oligomers and aggregates. Autophagy clears them from the cell by lysosomes rather than by proteasomes, whose catalytic core is too narrow for bulk material to enter. The most common form of autophagy, known as macroautophagy, involves formation of an isolation membrane appearing around the bulk material to sequester it, then fusion of the edges of the membrane into a double-membrane structure known as an autophagosome. The autophagosome in turn fuses with lysosomes, which destroy the protein with their proteolytic enzymes. At least 35 autophagy-related genes essential for formation of autophagosomes have been identified (Yang and Klionsky, 2010).

Under normal conditions, autophagy occurs at a modest basal level. But under conditions of stress and nutrient depletion, autophagy is increased. Normal animals, whose autophagy in the central nervous system is blocked by knocking out essential autophagy genes (e.g., Atg5 or Atg7) needed to assemble the autophagosome membrane, proceed to develop neurodegenerative disease, as evidenced by behavioral deficits and loss of specific nerve cells (Hara et al., 2006; Komatsu et al., 2006).

A deficiency or outright failure of autophagy is thought to permit aggregation of misfolded proteins that lead to neurodegeneration. A key question is what causes the failure of autophagy in disease? The answer to this question could guide the creation of new therapies. There are several possible reasons for failure of autophagy. One is a failure by the autophagosome to recognize aggregated material. This was found to be the case in a study of Huntington's disease, according to Ana Maria Cuervo of Albert Einstein School of Medicine. In mouse models of Huntington's disease and cells from Huntington's patients, autophagosomes failed to efficiently trap protein deposits, organelles, and other cargo (Martinez-Vicente et al., 2010). The rest of the pathway was intact, for autophagosomes formed at a normal rate and fused appropriately with lysosomes. The failure was inefficient engulfment of cytosolic components. Cuervo speculated that the failure of recognition may stem from pathogenic proteins, such as the mutant huntingtin protein, which becomes attached to the inside of the membrane of the autophagosome, interfering with its capacity to recognize bulk cargo. Another cause of failure in autophagy might be that the aggregate itself induces damage to the pathway, rendering the autophagosomes unable to traffic within the cell. More specifically, Warren Hirst of Pfizer pointed out, the problem in one Alzheimer's case was the result of flawed fusion of the autophagosome with the lysosome. However, in another mouse model of Alzheimer's, Cuervo said, the failure occurred because the disease environment changed the pH, which is critical for lysosome's hydrolytic enzymes to work effectively (Lee et al., 2010b).

In the discussion, prompted by several questions, Cuervo said much remains to be known about autophagy and trafficking of autophagosomes in nerve cells. Lysosomes are less likely to be found in nerve cell processes as opposed to the cell body, so autophagosomes forming in the processes may need to traffic to the site of the lysosomes by retrograde transport up the microtubules, although this does not uniformly hold and ongoing studies support lysosomal presence in terminals. She stressed that movement of autophagosomes through retrograde transport is understudied.

One therapeutic opportunity is to enhance autophagy. This has been done successfully with the drug rapamycin, which induces autophagy. Rapamycin slowed the progression of Huntington's disease pathology in experimental models (Ravikumar et al., 2002). But the drug induces autophagy only weakly in physiologically relevant cells, such as cortical neurons. Steven Finkbeiner of the Gladstone Institutes and the University of California, San Francisco, described a small molecule (N10-substituted phenoxazine) that induces autophagy specifically in neurons from the striatum, cortex, and hippocampus. He found that the compound was neuroprotective in an animal model of Huntington's disease (Tsvetkov et al., 2010). These efforts will be difficult to translate into human research unless there are good biomarkers for measuring autophagy in a patient population, noted Finkbeiner.

Finally, therapeutic approaches can be designed to prevent protein aggregation altogether, thus obviating the need for therapies to induce autophagy. John Dunlop of AstraZeneca described a new drug, developed by Pfizer. Already approved in Europe, the drug tafamidis functions to stabilize the correctly folded tetramer form of the transthyretin (TTR) protein. This protein is destabilized in the genetic disease Transthyretin Familial Amyloid Polyneuropathy (TTR-FAP), a rare, progressive, and fatal neurodegenerative disease. In patients with TTR-FAP, the protein dissociates and forms amyloid fibrils, which, in turn, cause failure of the autonomic nervous system and/or the peripheral nervous system, among other bodily sites. The new medication, said Dunlop, counters the pessimistic view that protein–protein interactions are not likely to lend themselves to drug development. It also stands as testimony to the possibilities of drugs designed to combat protein aggregation, he said.


Recent findings show a strong correlation between the aggregation of misfolded proteins and the engagement of a stress response of the endoplasmic reticulum (ER), said Claudio Hetz, a professor at the University of Chile. Table 3-1 shows the neurodegenerative diseases for which evidence for ER stress has been documented in cellular/animal models and in postmortem human studies (for a review of this evidence, see Matus et al., 2011). The ER organelle is an “essential compartment for the maturation and processing of proteins” (Matus et al., 2011, p. 239). Hetz noted that ER stress triggers an adaptive response known as the unfolded protein response (UPR), which controls hundreds of genes related to protein quality control and folding. However, if these mechanisms of adaptation are insufficient to recover homeostasis of the ER, irreversible or chronic ER stress can also trigger cell death.

TABLE 3-1. Endoplasmic Reticulum Stress in Neurodegenerative Diseases.


Endoplasmic Reticulum Stress in Neurodegenerative Diseases.

Hetz emphasized that the contribution of the UPR pathway to neurodegenerative diseases is not fully understood, including the circumstances under which it appears to provide an adaptive response that increases survival of neurons, as well as the circumstances under which it represents a pathological mechanism that leads to neuronal dysfunction or cell death when the damage is too high. Hetz and his collaborators have been working to further understand the role of UPR in neurodegenerative disease by generating new mouse models that enable them to manipulate the UPR and see the impact on models of ALS, Huntington's disease, and Parkinson's disease (see, e.g., Hetz et al., 2008, 2009; Vidal et al., 2012; Zuleta et al., 2012). Some recent findings from a different research group indicate that mild ER stress (“preconditioning”) may even inhibit the death of neurons by promoting autophagy, suggesting that preconditioning could have potential value in developing therapies for neurodegenerative diseases (Fouillet et al., 2012). During the discussion period, a workshop participant raised what he termed the “Goldilocks effect”: Because a mild level of stress may be helpful but a high level will be harmful, it will be challenging to develop optimal therapeutic regimes. Hetz clarified that he thinks that decreasing the stress levels will always be good, but that perhaps a mild stress will be sufficient to trigger an endogenous adaptive response.


The speakers at the workshop identified many questions for future research and other opportunities for future action. The research suggestions related to protein aggregation are compiled here to provide a sense of the range of suggestions made. The suggestions are identified with the speaker who made them and should not be construed as reflecting consensus from the workshop or endorsement by the Institute of Medicine.

  • Develop better mechanistic understanding of protein aggregation and clearance mechanisms. (Hirst, Ommaya, Rigo)
  • Use conformation-specific antibodies to recognize specific structures in aggregation intermediates. (Finkbeiner)
  • Develop methods to measure protein homeostasis and quality control. (Cuervo)
  • Develop and standardize an in vitro model of α-Synuclein aggregation. (Kowall)
  • Study propagation of amyloidogenic proteins and ways to prevent or arrest it. Will plasmapheresis, intravenous immunoglobulin, or other manipulations halt or reverse propagation? (Kowall)
  • Use imaging and other methods to measure proteostasis and related quality control mechanisms in humans. (Cuervo, Ommaya)
  • Develop understanding of heterogeneity of protein aggregation in normal and disease populations. (Cuervo)
  • Develop deeper understanding of proteostasis within neurons and within specific neuron subtypes. (Dunlop, Ommaya)
  • Study whether people with neurodegenerative disease have genetic susceptibility to protein overexpression or misfolding. (Ranum)
  • Develop therapies that stimulate proteasomal degradation and autophagy. (Cuervo, Goldberg)
  • Evaluate therapeutic targets or therapies in proteostasis that are shown to be beneficial in one disorder in models of other neurodegenerative diseases. (Finkbeiner)
  • Identify the mechanisms of movement of autophagosomes through retrograde transport. (Cuervo)