Below is part 2 of my interview with Prof. Hilal Lashuel on the Need to Embrace Complexity and Revisit the Pathology of Neurodegenerative Diseases. Click here to read part 1 on Why a Patient-Centered Approach to Research and Reform of the Academic Incentive and Reward Systems are Essential to Advancing Translational Research.
Are all protein misfolding diseases similar (HD, PD, AD)? How might solving one help us solve the others?
Yes and No. These diseases affect different brain regions and manifest distinct clinical symptoms and pathologies. Moreover, diseases such as HD, PD or AD are not single entity diseases and individuals affected by these diseases experience the disease and respond to therapies differently. That said, it is clear that there are many striking similarities and shared molecular mechanisms and pathways that underpin the pathogenesis of these diseases. For example, protein misfolding and aggregation plays a central role in the pathogenesis of AD, PD, HD, and other neurodegenerative diseases. Despite the large differences in terms of sequence, structure and cellular properties of the proteins associated with each disease, they exhibit similar behavior when it comes to how they aggregate, induce toxicity and/or spread pathology in the brain. Furthermore, increasing evidence suggests that there are common cellular mechanisms and pathways for regulating protein aggregation and responding to cellular stress and toxicity induced by proteins linked to the different neurodegenerative diseases. Therefore, I believe strongly that unraveling the mystery of one of these neurodegenerative diseases will undoubtedly have tremendous impact and open new opportunities to help us understand and treat other neurodegenerative disease. This is why in our group we work on all three diseases (PD, HD and AD) and try to cover as much ground as possible by leveraging the advantages offered by the proteins, pathways and animal models of each of these diseases.
What do you think of the argument that deleterious alpha-synuclein accumulations might be an effect rather than a cause of synucleinopathies like PD?
Personally, I am not married to any specific hypothesis and I am willing to reconsider my position on the basis of the scientific evidence. In the case of alpha-synuclein, the genetics (gene mutations and duplications) and the pathology tell us that changes in alpha-synuclein levels and the process of alpha-synuclein aggregation plays a central role in the disease. While it is also true that different aggregate forms of alpha-synuclein are toxic to the cells, it is also true, and we have recent data from our lab that supports this, that neurons could accumulate large amounts of aggregates without any noticeable or “measurable” deleterious effects. There is evidence for toxic, benign and protective effects of alpha-synuclein aggregation. I believe that part of the problem is the oversimplification of this process and the wishful thinking that there is a single toxic entity or mechanism of toxicity. Of course this is also driven by the fact that the complexity of the disease, and the possibility of multiple mechanisms of toxicity, makes things more complicated for the pharmaceutical industry, which can’t afford to wait for us to reconcile all the different hypotheses. Unfortunately, today it is also much easier to publish data that fits existing hypotheses than data that illustrates the complexity of these diseases or that cannot be explained by the existing hypotheses and working models.
In our laboratory, we approach this problem by focusing on the interplay between the processes driving protein aggregation and neurodegeneration, rather than focusing on one specific species or specific toxic mechanism. We also make a distinction between the two processes of protein aggregation and inclusion formation, as these two processes are very distinct and involve different sets of molecular actors. The first we can easily reproduce in the test tube, but the latter is much more complex and we have yet to succeed in reproducing it, even in animal models. I personally believe that inclusion formation is a protective mechanism, but the dynamic process of protein aggregation and fibril formation, which take place prior to the efficient sequestration of the aggregates in compact inclusion, is where things could go wrong. Obviously, at some point the continued accumulation of inclusions and the failure of neurons to degrade and clear these inclusions will also lead to toxicity. In all cases, we strive to always address our questions in the context of what is happening in the human brain. As Dr. Ismail Kola, former vice president of UCB, once told me, we must remember that the “issue is in the tissue”.
We believe that the mechanisms by which protein aggregation cause neurodegeneration are cell type dependent and vary depending on the environment in which these aggregates are formed and/or exert their effect as this will dramatically influence the structural properties, interactome of these aggregates, and the distribution and stability of the different aggregate species. We have shown previously that in the case of extracellular amyloid-b, as well as alpha-synuclein, toxicity is mediated by the ongoing polymerization and growth of the fibrils, rather than specific oligomeric population. The level of monomers, not oligomers, is the key determinant of aggregation and toxicity, despite the fact that the monomers alone are not toxic. Our studies suggested that it is not the load of the aggregates that matters, but the dynamics of these aggregates and their ability to grow and remodel themselves that make them toxic to neurons. On the basis of these findings, we proposed that targeting protofibrils and fibril growth represent a viable strategy for interfering with all three disease causing processes, including protein aggregation, proteotoxicity and the spreading of pathology. The advantage of this approach is that proteins such as amyloid-b, synuclein and tau, which are normally unfolded and difficult to target with small molecules, become highly ordered in the fibril state and thus much more tractable to target using rational structure-based approaches, especially with the increasing number of fibril structures being available, thanks to advances in Cryo-EM.
Your lab tries to develop new tools to help us better understand our biology. Could you give some insight into the methodology of tool making? Is it a form of engineered serendipity?
My group works at the interface of chemistry, biochemistry/biophysics and neurobiology. Our goal is to leverage advances in chemistry and develop novel tools that will help us address specific questions that cannot be addressed using the existing knowledge, methods and tools. The best example to illustrate this is our work on elucidating the role of post-translational modifications (PTMs) in regulating protein aggregation and toxicity in neurodegenerative diseases. In AD and PD, and most neurodegenerative diseases, the aggregating proteins found in the pathological aggregates associated with each disease are always heavily decorated with a large number of different types of post-translational modifications, usually multiple modifications. In fact, antibodies that were developed to detect these modifications (e.g. ubiquitination and phosphorylation) remain the only tools we have today to characterize, classify and monitor the formation of the pathological aggregates in post-mortem brains and animal models. Interestingly, despite the strong association between PTMs and the pathology of these diseases, you rarely find PTMs incorporated in the general schemes of protein aggregation and proteotoxicity presented in most review articles and publications. We know that these modifications are always there, but we don’t know why they are there.
One of the problems of investigating PTMs is that most of the time we do not know, 1) what are the enzymes that are responsible for introducing the PTMs of interest, 2) whether they occur before or after protein aggregation, and 3) what is their role in regulating protein aggregation, toxicity and clearance. This is further complicated by the fact that usually several of these modifications occur on the same molecule or aggregates, suggesting that PTMs represent a complex combinatorial code rather than a single letter code. This complexity, and the lack of tools to deal with it, led to further oversimplification of the problem, or simply ignoring the PTM code and relying on existing tools, which are based on introducing natural mutations to mimic or block the effect of these mutations. The most common and classical example is the use of serine/threonine to glutamate mutations to mimic the charge effects of phosphorylation. The problem is that this approach does not always work and is only possible for one of the many PTMs.
To address these challenges, we decided that the best way to is to develop strategies where we are able build these proteins from scratch, one amino acid at a time. By doing so, we would have the flexibility to introduce the desired modifications at single or multiple sites in the protein, and produce proteins that are homogeneously modified. These advances enabled us to develop strategies that give us full control in introducing simple or complex modification into proteins, including, phosphorylation, nitration, acetylation, SUMOylation and mono, di- and poly ubiquitination. In parallel, we have ongoing efforts to discover the enzymes that regulate these modifications, with an emphasis on phosphorylation. The combination of synthetic strategies and knowledge about the enzymes that regulate some of the PTMs have enabled us to begin to understand the crosstalk between the different PTMs and pave the way for cracking the PTM code in PD, AD and HD.
We were the first to develop synthetic strategies that enable site-specific modifications of alpha-synuclein, exon1 of the Huntington protein and more recently Tau. Our studies have shown that phosphomimetic mutations do not reproduce all aspects of bona fide phosphorylation and thus the use of these mutations in cellular and animal models could be misleading, or at best, offer only an approximation of the true biological effects of this modification.
In the case of synuclein, we developed several modular strategies that enabled the generation of a comprehensive library of pure proteins that includes every form of synuclein that has been detected in the brain and biological fluids (CSF and blood). When I told colleagues that we plan to work on alpha-synuclein PTMs, they always told me which ones, there are too many. We started by focusing on disease-associated PTMs, but ultimately decided to study them all. We have taken the same approach with Huntingtin and more recently with Tau.
One of the main surprising findings from our work on alpha-synuclein, Huntingtin and Tau, is that the great majority of the PTMs associated with the disease pathology did not influence or inhibit the process of alpha-synuclein aggregation. These findings suggest that these PTMs occur after protein aggregation, thus we recently shifted our focus to developing tools and methods that would allow us to dissect the role of post-fibrillization PTMs in regulating the formation and clearance of inclusion formation. Our recent and most exciting data reveal that these PTMs are not simply markers of pathology, as they have always been viewed and used, but act as master regulators in the processing of aggregates and their packaging into the inclusions that we see in the diseased brains. Our findings also suggest that the major contribution of these modifications may be related to their role in regulating the normal function(s) of these proteins. Towards this goal, we have started a systematic program to map the PTM-dependent interactome code of these proteins.
How complex is the ‘post translational modifications code of proteins’ that your lab works to decipher?
The PTM code is very complex. This is clearly illustrated by examining the number of modifications that we have seen in proteins such as alpha-synuclein and Tau and the fact that several of these modifications usually occur simultaneously in the same molecule and within short sequence motifs. In short, I believe that the PTM code is not a single letter code but rather is a combinatorial code. More like a bar code we see today in products we purchase. This adds a layer of complexity that has implications for not only how we study these proteins, but also how we target, detect and quantify these proteins in biological samples.
Furthermore, our work shows that the aggregation, and potentially the functions of these proteins, are regulated by a complex cross-talk between different neighboring PTMs or PTMs in different regions of the proteins. This further indicates that the PTM code is a combinatorial code. Therefore, the prevailing reductionist approach based on investigating these PTMs individually is unlikely to help in unraveling this complex code and assessing its potential in developing novel diagnostic and therapies. Despite this, there is data to show that modulating single PTMs could have dramatic effects in inhibiting the aggregation and toxicity of amyloid proteins and/or enhancing their clearance. This suggests that the enzymes involved in regulating some of these PTMs, which are durable, represent viable targets for the treatment of neurodegenerative diseases.
The alpha-synculein protein library we generated has proven to be instrumental in our ability to dissect the role of PTMs in modulating alpha-synuclein structure and aggregation and is now used extensively to advance several Michael J Fox funded international collaboration on the discovery and validation of biomarkers and to facilitate the development of assays to accurately detect and quantify alpha-synuclein in biological fluids.
Today, there are many assays used in research and clinical settings to measure total alpha-synuclein and Tau, yet all of these assays were developed using a single form of these proteins and still rely on a single form as a calibrant. In simple words, we do not know if they indeed capture all the forms of these proteins and truly measure total synuclein or Tau. Similarly, with respect to assays that measure a specific form of these proteins, such as phosphorylated alpha-synuclein or Tau. We do not know whether the presence of neighboring modifications influence our measurements. Indeed, recent data from our group suggest that this is the case.
In the case of alpha-synuclein, the great majority of these are immunoassays based on capturing antibodies targeting the regions of the protein that are known to be heavily modified. We believe that this is one of the primary reasons for the discrepancy in existing data regarding the correlation between total alpha-synuclein levels and PD or PD progression. We have recently received funding from the Michael J Fox foundation to use our alpha-synuclein library to compare these assays and determine what they actually measure. This knowledge will then be used to develop more sensitive assays that capture all forms of alpha-synuclein and provide a more accurate measure of total alpha-synuclein species.
If computers become capable of accurately and quickly modeling the 3D structure of proteins and simulating their interactions, as some claim quantum computers might enable us to do in the near future, will that solve these diseases that are thought to be protein folding problems?
I wish that this were the case. Again, even in computational biology where one would like to believe that the sky is the limit, our computational colleagues continue to model and engineer proteins without accounting for or exploiting the role of post-translational modifications in regulating the structure, dynamics and function of proteins. Given the complexity and combinatorial nature of the PTM code, it is almost impossible to explore and decipher this code using a purely experimental approach. I would even go further and argue that achieving a complete understanding of the genetic code is not possible without understanding the PTM code, since PTMs are the main biological switches for regulating the dynamics of protein structure and function in health and disease.
Today, we do not have the tools to model and predict the effect of even the simplest and most common post-translational modification, phosphorylation, on protein structure and dynamics, especially for intrinsically disordered proteins such as alpha-syuclein. A couple years ago, I asked a colleague to help us model the effect of phosphorylation at one residue on the structure of a simple 17 amino acid peptide. All attempts failed to reproduce the experimental data we obtained on this peptide. This is in large part due to the lack of experimental data on site-specific and homogeneously modified proteins, which could be used to develop and optimize computational methods. Everyone in the field of computational biology acknowledges that this is a major problem and a limitation to advancing the field, but very little is done to address this issue today.
We would like to think that there are no longer any excuses that should prevent us from embracing the complexity of the PTM code. We have the methods to generate the proteins and to qualitatively and quantitatively characterize their properties and obtain the data we need to guide the development of new computational methods and approaches.
We are already reaching out to computational chemists and physicists, bioinformatics and structural biology groups, to leverage the experimental structural, biophysical and cellular data generated using our comprehensive libraries of site-specifically modified proteins to develop novel computational tools to 1) predict the effect of specific PTMs on protein structure and aggregation; 2) decipher the roles governing cross-talk between different and competing PTMs; 3) determine the general rules of when natural mutations can be used to mimic post-translational mutations when investigating biological and/or disease mechanisms and pathways in cells and in vivo.
I would like to end this interview by emphasizing that embracing complexity, collaborations and a patient-centered approach to research are essential to making transformative discoveries that will pave the way for understanding and findings cures for neurodegenerative diseases.