Small molecules for modulating protein driven liquid-liquid phase separation in treating neurodegenerative disease

Richard J. Wheeler, Hyun O. Lee, Ina Poser, Arun Pal, Thom Doeleman, Satoshi Kishigami, Sukhleen Kour, Eric Nathaniel Anderson, Lara Marrone, Anastasia C. Murthy, Marcus Jahnel, Xiaojie Zhang, Edgar Boczek, Anatol Fritsch, Nicolas L. Fawzi, Jared Sterneckert, Udai Pandey, Della C. David, Benjamin G. Davis, Andrew J. Baldwin, Andreas Hermann, Marc Bickle, Simon Alberti, Anthony A. Hyman

Posted on: 30 August 2019 , updated on: 31 August 2019

Preprint posted on 21 August 2019

Not a big FUS over nothing: small molecules altering phase behavior of stress granules emerge as an effective method to reduce aggregation and reverse ALS-associated phenotypes.

Selected by Madhuja Samaddar

Categories: biophysics, cell biology, molecular biology, neuroscience

Background:

The relatively nascent concept of biomolecular condensates has stirred up increasing interest in recent years, emerging as a fundamental organizational principle in cells¹. Multiple groups have described how these compartments – also described as ‘membraneless organelles’ – are formed by liquid-liquid phase separation and are functionally important for various biological processes². Interestingly, phase separation or demixing also underlies the formation of stress granules: ribonucleoprotein assemblies that sequester mRNAs from the translation machinery under conditions of stress³. These assemblies enable cells to prioritize synthesis of proteins that are required to deal with the emergency, by temporarily halting the translation of non-essential genes³. However, there also appears to be a direct link between stress granules and disease. For example, amyotrophic lateral sclerosis (ALS) is associated with mutations in stress granule proteins⁴. Further, ALS-linked mutations have been shown to alter the material states of condensates to less liquid-like⁵. One might speculate that a productive intervention would target the material properties of stress granules, without a) affecting other phase-separated compartments and b) relying on highly specific interactions with granule components^6,7.

In this exciting new collaborative study, the authors monitor the localization and phase behavior of FUS to screen for compounds capable of modulating these properties. FUS is a protein that normally resides in the nucleus but localizes to stress granules under conditions of stress. Mutations in FUS, many of which shift its localization towards the cytoplasm, are associated with familial forms of ALS. The candidates obtained in the screen are then used to probe the alteration in properties of the phase separated condensates, and to test for reversal of ALS-linked detrimental phenotypes in patient-derived iPSC neurons and in organisms such as worms and flies.

Key findings:

Candidate identification and verification: A cell-based screen identified lipoamide and the closely related compound lipoic acid as having a strong effect on numbers of FUS condensates and intracellular FUS localization in stressed cells. Lipoic acid is a cofactor required for aerobic metabolism and has been previously proposed to be an antioxidant. Although an effect of redox chemistry cannot be ruled out completely, the authors demonstrate that the compounds have additional effects on the phase behavior of stress granules. Further they find that both candidates were able to reduce stress granules induced by diverse stresses, and that the modulatory effect is not restricted to FUS – other stress granule-resident proteins are also similarly affected.
Specificity towards stress granules: Lipoamide did not affect a range of other functional condensates indicating that it preferentially modulates stress granules. This is a leap forward from known stress granule modulators like 1,6-hexanediol which is required in very high concentrations to elicit any effect and alters other membraneless organelles like the nucleolus. Further, lipoamide did not negatively impact the beneficial recruitment of FUS at sites of DNA damage within the nucleus. It is noteworthy that both candidates are non-toxic and lipoic acid also has well-characterized pharmacokinetics.
Proposed mechanism of action: Lipoamide treatment affected key physical properties of the condensates resulting in a) improved liquidity in vitro, and b) delayed pathological ‘aging’ of condensates containing G156E FUS into fibers characteristic of ALS. NMR-based experiments further indicated that the most likely mechanism by which lipoamide acts is the alteration of condensate properties, rather than high-affinity binding with stress granule components.
Demonstration of beneficial effects, in vivo: The candidates were able to reverse protein aggregation and deleterious phenotypes in multiple models. Treatment with lipoic acid reduced the aggregation of PAB-1, a stress granule-related protein in elegans. Further, both lipoic acid and lipoamide treatment reduced aggregation in patient-derived P525L FUS iPSC lines. The compounds also recovered axonal transport and prevented axon die-back in iPSC motor neurons expressing the same ALS-associated FUS mutation. Finally, severe motor neuron defects caused by expressing human FUS mutants in Drosophila were also successfully reversed by dietary lipoamide/lipoic acid supplementation, as estimated by restoration of their ability to climb.

**Time dependent elimination of cytoplasmic FUS GFP condensates in HeLa cells.**
(Reproduced from the preprint by Wheeler R.J. et. al., 2019)

Why I chose this preprint:

Currently approved treatments available for ALS either target specific symptoms or serve only to slow down functional deterioration and reduce discomfort. Mutations in core stress granule proteins like TIA-1 have been associated not just with ALS but also in ALS patients with accompanying Frontotemporal Dementia⁸. Further, TIA-1-positive stress granules have been suggested to stimulate tau pathophysiology in Alzheimer’s Disease⁹. Therefore, understanding stress granule biology and identifying interventions to modulate their formation, physico-chemical properties and stability, is key to designing any potential therapeutic approach. This study accomplishes a mammoth task of identifying small molecules from a cell-based screen, demonstrating their effects on FUS condensates in vitro and in cells, and further in reversing FUS-associated defects at the level of individual neurons and organisms. Taken together, it is the first study indicating that a direct link may exist between modulation of phase behavior of subcellular condensates and the outcome on a disease-linked state.

Further, the authors show that lipoamide does not interact via high-affinity binding to FUS. Instead, the proposed mechanism of action is through modulation of physical properties of the condensates, resulting in altered phase behavior. There are likely other compounds/small molecules that can also beneficially modulate biomolecular condensates linked with ALS or with other neurodegenerative diseases.

Finally, this study provides evidence in favor of the argument that modulation of stress granules leading to recovery of normal localization of proteins like FUS, is indeed beneficial.

Questions for authors:

It appears that the dietary supplements in flies leading to reversal of motor defects were supplied chronically throughout life. If so, do you have a sense for whether the effect is preventive and if the timing of introduction of lipoamide/lipoic acid matters? Can an acute intervention in flies that already demonstrate the climbing defects still produce the beneficial effect? In other words, how late is too late?
Lipoamide treatment did not dissolve multiple nuclear compartments either containing FUS or otherwise. In the preprint it is speculated that the interactions and physical chemistry driving different types of condensates could account for this specificity. However, assuming the nucleus and cytoplasm to be already two separate assemblies, could the nuclear envelope somehow represent a stronger permeability barrier to these compounds, thereby protecting the nuclear condensates? How universal is this protection of diverse non-stress granule condensates in the cytoplasm (beyond the example of non-disruption of P-bodies containing DCP1A)?
The formation of stress granules represents a protective cellular mechanism to cope with adverse environmental conditions. Would you expect undesirable side effects when treating patients with any compound that inhibits stress granule formation? In order to be of therapeutic value, would the dose need to be fine-tuned to still allow a sufficient level of protection against cellular stress?

Acknowledgements:

I would like to thank my fellow preLights team members Tessa Sinnige and Monika Magon for very helpful feedback and discussion, preprint author Richard Wheeler for comments on technical accuracy, and Suvrajit Saha for critical comments.

References:

Alberti S. Phase separation in biology (2017). Phase separation in biology. Curr. Biol. 27
Banani SF, Lee HO, Hyman AA and Rosen MK (2017). Biomolecular condensates: organizers of cellular biochemistry. Nat. Rev. Mol. Cell Biol. 18
Decker CJ & Parker R (2012). P‑Bodies and stress granules: possible roles in the control of translation and mRNA degradation. Cold Spring Harb. Perspect. Biol. 4
Bosco DA, Lemay N, Ko HK, Zhou H, Burke C, et. al. (2010). Mutant FUS proteins that cause amyotrophic lateral sclerosis incorporate into stress granules. Hum. Mol. Genet. 19
Patel A, Lee HO, Jawerth L, Maharana S, Jahnel M, al. (2015). A liquid-to-solid phase transition of the ALS protein FUS accelerated by disease mutation. Cell 162
Kroschwald S, Maharana S, Mateju D, Malinovska L, Nuske E, et. al. (2015). Promiscuous interactions and protein disaggregases determine the material state of stress-inducible RNP granules. eLife 4
Wheeler RJ & Hyman AA (2018). Controlling compartmentalization by non-membrane-bound organelles. Philos. Trans. R. Soc. B Biol. Sci. 373
Mackenzie IR, Nicholson AM, Sarkar M, Messing J, Purice MD, et. al. (2017). TIA1 Mutations in Amyotrophic Lateral Sclerosis and Frontotemporal Dementia Promote Phase Separation and Alter Stress Granule Dynamics. Neuron 95
Vanderweyde T, Yu H, Varnum M, Liu-Yesucevitz L, Citro A, et. al. (2012). Contrasting pathology of the stress granule proteins TIA-1 and G3BP in tauopathies. J. Neurosci. 32

Further reading:

This blog post about the same preprint, by Derek Lowe.
Alberti S. & Dormann D. (2019) Liquid-liquid phase separation in disease Annu. Rev. Genet. 53

Tags: amyotrophic lateral sclerosis, fus, liquid-liquid phase separation, small molecules, stress granules

doi: https://doi.org/10.1242/prelights.13626

Read preprint

(No Ratings Yet)

Author's response

Richard J. Wheeler shared

Thank you for picking our work and summarising it so well! It is wonderful to see our efforts to take a novel approach to tackling a neurodegenerative disease reviewed like this. This has been a challenging multi-group undertaking, from identifying the phase separation behaviour of stress granule proteins as a potential intervention point in ALS, through to testing the resulting small molecules in disease models. However the result – developing and testing the concept of a physicochemical drug – was certainly worth the effort.

We don’t yet know whether lipoamide pre-treatment is necessary to prevent the onset of pathology in animal models, we haven’t tested this in depth. If the pathomechanism of ALS is via toxicity of aggregates then pre-treatment will be needed – lipoamide and lipoic acid didn’t reverse aggregation, only slowed it. However, there are other possible pathomechanisms which can’t be excluded at this stage. We deliberately didn’t search for compounds which reverse FUS fibre formation/aggregation. There are already quite a few well-characterised compounds which dissolve prion-like fibrils but, unfortunately, they haven’t found success in trials. Instead we specifically took a different approach by focusing on compounds which altered liquid-liquid phase separation.

Presence of FUS-containing compartments in both the nucleus and cytoplasm certainly raises some interesting questions. We haven’t been able to directly confirm that lipoamide/lipoic acid enter the nucleus, tracking these small molecules in cells is very challenging, leaving the mechanism for the opposing effects on nuclear and cytoplasmic FUS-containing compartments unclear. However the important concept here is specificity, any physicochemical drug should specifically affect the target compartment – affecting the wrong compartment could be a severe off-target effect. This was why we selected cytoplasmic P-bodies for comparison to stress granules: P-bodies are ribonuclear granules with a fair similarity to stress granules but with distinct biological functions and properties. No effect of lipoamide on P-bodies suggests limited off-target effects. Of course, testing other cytoplasmic compartments would be interesting too though.

Whether promoting stress granule dissolution is a sensible therapeutic strategy certainly isn’t obvious and has been raised as a concern on several occasions. Stress granules presumably evolved for a reason! We hope for a therapeutic window though, either in dose or in time. Any side effect of reduced stress granule formation may well be long-term, such as accumulated tissue damage or increased cancer propensity through inability of cells to properly cope with stresses. The extreme severity of ALS symptoms and high speed of disease progression would provide a large temporal therapeutic window in this case. While the screen was for stress granule dissolution, there’s also the possibility that the primary effect of long-term lipoamide treatment in cells is increased stress granule liquidity which may not cause side effects.

Lipoamide and lipoic acid are certainly not like conventional small molecule therapeutics, but liquid-liquid phase separation is not a conventional drug target. This posed a lot of challenges! Particularly analysing interaction of a small molecule with a protein which remains unstructured, so can’t involve a conventional high-affinity binding site – despite it being some form of direct interaction which affects protein phase separation in a minimal in vitro system. Looking to the future, and despite these challenges, we think modulating liquid-liquid phase separation in the formation of cellular compartments will be very interesting as a therapeutic target.