September's ALS Research Roundup

Heat-shock chaperone HSPB1 regulates cytoplasmic TDP-43 phase separation and liquid-to-gel transition

(Lu,?Hu et al.?2022)

One of the few consistent underlying factors in ALS is aggregates of the protein TDP-43 in the cytoplasm of affected cells. However, we still lack understanding of many of the properties of TDP-43 which could help to explain ALS pathology. One of these is liquid-liquid phase separation, a state in which liquids separate into distinct compartments rather than mixing. In the case of TDP-43, this takes the form of a dilute phase as well as a condensed phase which may contribute to pathological protein aggregation. These separate phases can further develop into completely separated solid and gel phases which can induce damaging cellular stress and promote the production of pathological protein structures including solid aggregates known as amyloid fibrils. Phase separation is tightly controlled in normal cells, but often becomes more severe when cells are exposed to various stressors, several of which are observed in ALS. This can result in a harmful cycle, as fibrils themselves have been found to induce these types of stress.

This study confirmed that TDP-43 phase separation can be triggered by stressors including oxidative stress and proteasome inhibition, as well as interactions between TDP-43 and the stress-induced molecular chaperone protein HSP70. However, another chaperone protein which plays an opposite role, HSPB1, was also identified as relevant to phase separation using cultured human cell lines. Similarly to HSP70, HSPB1 is predominantly found in the cytoplasm and binds to TDP-43 in phase-separated droplets. Unlike HSP70, it inhibits the development of solid/gel phases and consequently assembly of TDP-43 into fibrils, as well as contributing to the disassembly of stress-induced solid/gel phases. Accordingly, loss of HSPB1 promotes phase separation and was observed in spinal motor neurons of ALS patients with TDP-43 aggregates. Previous studies have shown that while increased levels of HSPB1 were beneficial in the early stages of ALS in mice, this benefit was lost as the disease progressed. This suggests that while phase separation plays a significant role in the development of ALS, it is less significant in established pathology. This aligns with our understanding of ALS, as aggregates of TDP-43 fibrils are able to catalyse aggregation on their own, without relying on phase separation.

While the loss of HSPB1 had previously been connected to ALS, this study established its connection specifically to the phase separation processes. It is notable that both it and HSP70 are recruited into TDP-43 phase-separated droplets, as disease processes occurring in the absence of HSPB1 suggests that healthy functioning depends on a stable balance of these two proteins. As a result, mutations which result in loss of HSPB1 levels of function are likely to be significant to TDP-43. While cases linked to mutations in this specific gene are rare, they have been identified in ALS as well as several other motor neuron diseases including Charcot-Marie tooth disease and distal hereditary motor dystrophy. In cases linked to HSPB1 dysfunction, gene therapies which redress this balance could be valuable early interventions, helping to prevent the development into irreversible protein cascades.

TAGS: mechanistic, HSP, molecular_chaperones, oxidative_stress, TDP-43, human_cell

Identifying patterns in amyotrophic lateral sclerosis progression from sparse longitudinal data

(Ramamoorthy, Severson?et al., 2022)

Machine learning is an ever-growing technology which is being applied extensively in the study of complex diseases, particularly those involving numerous contributing factors and pathways. While the human mind is extremely good at finding linear, or at least two-dimensional patterns and relationships, machines are capable of identifying multifactorial patterns. These are superior for modelling certain aspects of these diseases, particularly when changes are dependent on many or changing factors. ALS is a prime example of this, as pathology is believed to progress through several stages in which disease-driving proteins build up, aggregate and then enter self-reinforcing cycles, with different physiological effects emerging in each.

This study sought to use machine learning to model patterns of ALS disease progression. The model was based on various measures of ALS-associated functional loss including the ALSFRS-R, forced vital capacity, fine and gross motor control, bulbar and respiratory function. Five large patient datasets were incorporated, covering several thousand patients observed over at least 6-12 months. The models were specifically designed to be flexible, using automatic data-driven patterns which minimised interference by human biases. The models demonstrated changes in the different functional measures over time, and determined different subsets of data which matched certain disease progression patterns. For most of the measures five patterns were identified, although ALSFRS-R only had four. The patterns involved variable periods of stability interspersed with rapid declines, and rarely periods of temporary recovery. Interestingly, while the pattern groups were consistent within particular measure, they did not keep the same groupings between measures. For example, individuals sharing a pattern of fine motor decline may not have the same pattern of respiratory decline. These models were quite accurate as predictive tools, with projected outcomes matching relatively closely to the observed changes over time in the patients’ data.

These models have a number of potential uses in ALS research. The most direct is in stratification of patients. It may be that the different progression patterns are linked to underlying factors, so clustering could help to identify the key drivers of certain ALS subsets and to more accurately direct therapeutics. Unfortunately, the model currently requires 6-12 months of data to assign a patient to a particular pattern, making it relatively difficult to apply clinically due to the short life expectancy of patients following diagnosis. However, it may have even greater use in determining how effective therapies are. The observed non-linear progression patterns mean that true correlation-causation relationships during treatments can be hard to determine. For example, slowing of patients’ disease progression may be the result of typical fluctuations in their ALS type rather than the effects of the treatment. If we can map patients to a non-linear pattern, then determining true deviation as a result of treatment is likely to be much more accurate. In this way we would hope to identify more truly effective treatments, as well as excluding those which do not cause deviation. It would be interesting to see how closely cell and animal models of ALS map to these patterns, and whether this can lead to a new paradigm of ALS testing and study.

TAGS: model, machine_learning, prediction, human

Brain-restricted mTOR inhibition with binary pharmacology

(Zhan, Fan?et al., 2022)

Currently only two drugs (riluzole and edaravone) have been approved for treatment of ALS, and these only modestly slow disease progression. A multitude of other treatments have been attempted, but all to date have been excluded due to poor performance or excessive side effects. There are many possible reasons for this. It may be that treatments are only effective for certain subsets of patients, or that properly addressing it requires a many-pronged, multi-drug approach. Many have likely failed due to the drug being unable to reach the brain due to the blood-brain barrier (BBB), a physiological barrier which protects the brain by isolating the central nervous system (CNS) from the periphery. Numerous mechanisms have been tested to bypass the BBB, such as binding drugs to molecules capable of crossing the barrier or directly injecting into the CNS (which is difficult and can have significant side effects). However, in almost all cases the amount of drug which reaches the brain is significantly less than that introduced into the periphery. This can result in heightened side effects in the periphery, which may keep these treatments from being clinically viable.

This study proposed a novel avenue of minimising off-target effects from drugs which target the CNS. They administered both a brain-permeable drug which could breach the BBB, along with a non-brain-permeable agent which inactivated the drug in the periphery. The goal was that the drug would spread through the whole body, but everything outside of the CNS would be deactivated. In this case, the brain-permeable drug was the RapaLink-1, an inhibitor of mTOR (a growth regulator which is elevated in ALS). mTOR inhibitors have previously been investigated for use in ALS and other diseases of the CNS, but were hindered by adverse effects including immune suppression and metabolic disorders. The second drug was RapaBlock, a brain-impermeable ligand which binds the protein FKBP12 which RapaLink-1 requires to function. RapaBlock was shown to not affect mTOR signalling by itself, but does counteract RapaLink-1 in a dose-dependent manner. Co-treatment with both drugs in mice resulted in significant inhibition of mTOR activity in the brain comparable to RapaLink-1 alone, while skeletal muscle was completely unaffected.

The authors also considered the expansion of this protocol to other treatments. They presented a strategy for designing cell-permeable, FKBP12-dependent kinase inhibitors from known drug scaffolds which would be sensitive to deactivation by RapaBlock, and so able to be limited to the CNS in the same manner. This was tested with GNE7915, a drug being investigated for treatment of Parkinson’s disease, and they were successful in controlling it in the same way as RapaLink-1.

The method presented here is a simple but elegant means of minimising off-target drug effects, with a great deal of utility for diseases of the CNS. While this study was primarily addressing its potential for the treatment of mouse models of glioblastoma (aggressive brain tumours), it also proposes a method of adaptation to other targets. It is unclear from the paper how broadly it could be applied, and whether FKBP12-dependent drugs could be engineered or otherwise induced for a broader range of targets. Targeted inhibition of mTOR alone would have potential in ALS treatment, so the value of this paper in broadening potential drug opportunities should not be undersold.

TAGS: therapeutic, BBB, non-ALS

Transplantation of human neural progenitor cells secreting GDNF into the spinal cord of patients with ALS: a phase 1/2a trial

(Baloh, Johnson?et al., 2022)

While prevention or slowing of disease processes is the major focus of ALS research, there are those who instead attempt to repair the damage done. The introduction of stem cells which can differentiate into and replace damaged motor neurons is among the most direct approaches, but has often failed to produce functional improvements. There are a range of theories as to why they failed, but one is ALS-associated defects in astrocytes, a type of non-neuron cell in the brain which supports neuronal function. This has naturally progressed to research into stem cell-based replacement of astrocytes as well.

In this study, a human neural progenitor cell line was engineered so that it would differentiate into modified astrocytes which produced glial-derived neurotrophic factor (GDNF). GDNF is a protein which promotes the growth and survival of neurons, and has previously been studied as a possible gene therapy for ALS. However, it has a short half-life and difficulty reaching the brain. These modified cells, named CNS10-NPC-GDNF, would be transplanted into and graft to the spinal cord, where they would survive and secrete GDNF for the life of the patient. CNS10-NPC-GDNF cells were produced and transplanted to one side of the spinal cord, such that the non-transplanted leg could be used as an experimental control.

The transplant was found to successfully graft and survive, differentiating into healthy astrocytes which expressed GDNF for at least 42 months after transplantation. Immune suppression was applied for the first year following the treatment, but no immune response was observed even after suppression ceased. While there was some pain following the treatment, there were no functional or long-term side effects. Unfortunately, there was also no significant difference in the strength of the treated leg. However, despite not reaching statistical significance, the rate of decline was lower in the treated group. It is worth noting that all the patients treated were already in late-stage ALS, so it may be that the treatment would be more effective as an early intervention. Unfortunately, finding such early-stage patients is difficult in ALS due to difficulties in diagnosis and the rapid rate of progression.

While unable to produce statistically significant functional improvements here, they were able to validate the methodology of the procedure and confirm a lack of major side effects. As a result, the authors are planning future iterations on the procedure, such as targeting earlier disease stages and injecting into the ventral horn (contains motor neuron cell bodies), rather than the dorsal root which contains nerve projections from these bodies. The dorsal root was targeted due to a higher possibility of causing functional issues in the ventral horn, but may have likewise reduced treatment effectiveness. If this is developed into an effective treatment, its ongoing effect from a single treatment may be of great value in helping to both promote protection and regrowth of affected neurons. Recovering lost neurons may even help to ‘regress’ patients to a state reminiscent of earlier disease stages where treatments are much more effective, so the success of this would be of great value to patients and researchers alike.

TAGS: treatment, astrocyte, GDNF, human

Non-Invasive Transcutaneous Spinal DC Stimulation as a Neurorehabilitation ALS Therapy in Awake G93A Mice: The First Step to Clinical Translation

(Highlander,?2022)

The fundamental basis of signalling in the brain is a series of electrical signals. When the charge within a neuron reaches a certain level it fires, triggering a release of neurotransmitters which increase the charge of the next neuron in the signalling chain. While the threshold at which neurons fire is relatively consistent, the rate of change in charge and the baseline charge level can vary. This is the basis of the technology of transcutaneous spinal direct current stimulation (tsDCS): by passing an electrical field through the brain, the baseline charge level of neurons can be modulated, and so firing made more or less difficult. An associated phenomenon in ALS is the abnormal excitability of neurons. This results in spontaneous motor neuron firing and excitotoxicity, in which overexposure to excitatory stimuli damages cells. Excitotoxicity is seen from the early stages of ALS and continues throughout disease progression, and is a major contributor to ALS-associated cell death. tsDCS was presented as a possible means to counteract this change in a non-invasive and spatially-targeted manner.

This study involved two main stages. The first stage aimed to assess the clinical potential of tsDCS-induced changes in motor neuron excitability in genetically modified mouse models of ALS. This stage was intended as a general assessment of potential, as well as identifying any technical challenges. The second stage was to determine how best to use tsDCS, testing the effects of each orientation of stimulation in the same type of ALS mice. Stimulation can be either anodal (excites neuronal activity) or cathodal (inhibits neuronal activity), and it was presumed that one would be beneficial while the other was harmful. Unfortunately, they were unable to produce statistically significant changes in primary disease outcomes. However, anodal stimulation was shown to reduce survival times and several known correlations among disease symptoms were disrupted (particularly correlations between functional loss and survival), suggesting some level of effect. As the intensity of tsDCS was kept low in order to prevent skin burns, it may simply be that more power is required to achieve significant effect. This deficiency is even stronger in the possible protective cathodal tdDCS, which has a lower amplitude and short-lasting functional effects than anodal stimulation. Additionally, tsDCS preferentially effects larger neurons which degenerate first in ALS, so the treatment may have been applied too late in disease progression to achieve the best effects. However, as with so many other treatments, identifying presymptomatic cases in order to start treatment early is easier said than done.?

The attempt here to produce a fully non-invasive electrical ALS therapy in awake animals was largely ineffective, but hints at possible potential in the future. However, it seems likely that this will come at the cost of greater pain as part of the treatment. We can hope that more intensive treatment under anaesthesia would have lasting benefits, but it remains to be seen whether that will be the case. Another possibility is the use of subcutaneously implanted electrodes to more directly deliver the current, but this would no longer be the fully non-invasive intervention which the authors sought. Ultimately, this study ruled out the simplest and easiest method of electrical treatment, so we will see to what extent the technology will reach.

TAGS: therapeutics, device, hyperexcitability, SOD1, mouse

Stress induced TDP-43 mobility loss independent of stress granules

(Streit, Kuhn?et?al., 2022)

While the aggregation of various misfolded proteins is a common disease-driving factor in ALS, it does not occur in isolation. Various factors can contribute to aggregation, either through altering the proteins’ properties or production, or changing dynamics within cells. The latter is the case with stress granules, naturally occurring clusters of RNAs and proteins which develop when cells fail to make proteins properly during stress. They have also been shown to accumulate disease-related proteins, but the effect that this has on ALS has been hotly debated. Some claim that they act as ‘aggregation crucibles’, densely packing misfolded proteins and forging them into higher order aggregates far more rapidly than they would be otherwise. Others consider them to play a protective role, sequestering the proteins and preventing them from exerting toxic effects in other parts of the cells.

This study sought to confirm the relationship between the ALS-associated protein TDP-43 and stress granules in cell models of ALS, tracking the movement of TDP-43 when cells were exposed to various stressors. Two methods were used; ‘stimulated-emission depletion microscopy’ (STED) and ‘tracking and localization microscopy’ (TALM), which together showed TDP-43 substructures within stress granules as well as the location of DP-43 throughout the cytoplasm. Stress resulted in decreased mobility of ALS throughout the cell, indicating aggregation. This effect was heightened inside stress granules, supporting their role as ‘aggregation crucibles’. The removal of short-stress allowed them to recover, while long-term stress resulted in a permanent reduction in mobility. This may indicate that the formation of irreversible mature aggregates, which have been observed to catalyse further protein aggregation. TDP-43 structures were observed within stress granules but they also formed independently throughout the cytoplasm, providing evidence that TDP-43 aggregates can form independently of stress granules. However, aggregates within stress granules formed distinct ‘hotspots’ of activity, while cytoplasmic aggregates were more homogenous.

This study provided fairly conclusive evidence that TDP-43 can aggregate both within stress granules and independently of them, although the mechanics differed somewhat between them. The more focussed aggregation within stress granules may be a stronger disease driver, but that remains to be seen. While not groundbreaking, a proper understanding of what structures underly pathological aggregation can help with the targeting of therapeutics or further mechanistic research.

TAGS: mechanistic, stress_granule, TDP-43, human_cell

BL-918, a small-molecule activator of ULK1, induces cytoprotective autophagy for amyotrophic lateral sclerosis therapy

(Liu, Zhu?et?al., 2022)

The accumulation and aggregation of misfolded proteins is a key aspect of many neurodegenerative diseases, including ALS. However, these proteins are not invasive pathogens or unique mutations, but instead relatively common defects which emerge in the process of producing proteins. However, they are normally destroyed by cellular quality control processes long before they can accumulate to the level seen in disease. It has been theorised that this imbalance between the production and destruction of misfolded proteins is the key driver of ALS, with the various other factors simply altering the balance in various ways. One of the most significant protein quality control processes in autophagy, in which membranous structures (autophagosomes) capture target proteins before delivering them to lysosomes which contain digestive enzymes.

Several prior studies have identified dysfunction in autophagy in ALS, although it is uncertain whether this is a cause or consequence of pathological processes. One key protein involved in ALS-associated pathways is the ULK1 complex, which initiates autophagy by contributing to the production of the autophagosomes and recruitment of downstream proteins. ULK1 is prevented from activating by mTOR, which is elevated in ALS, and so recovery of ULK1 function and consequently autophagy may be a viable therapeutic pathway. Previous studies into enhancing autophagy produced some functional improvement, although many had unwanted side effects as well.

This study investigated the effectiveness of BL-918, a small molecule which activates ULK1, as a therapeutic autophagy enhancer in SOD1 cell and mouse models of ALS. In cells, treatment with BL-918 was able to induce ULK1-dependent autophagy in a dose-dependent manner, increasing levels of autophagic proteins and accelerating the formation of autophagosomes. This led to the elimination of toxic SOD1 aggregates. Autophagy was not induced in cells without ULK1, confirming that this improvement occurs through the ULK1 pathway. Treatment with BL-918 was also effective in the ALS mice, extending the lives and improving motor function. It was specifically shown to enhance autophagic clearance of pathological aggregates in the spinal cord and cerebral cortex, which are both affected in ALS.

While the disease model used here was SOD1 for both cells and animals, other variants of ALS have also been linked to ULK1 dysfunction and so could benefit from this treatment. These included C9orf72, the most common driver of familial ALS, which regulates autophagy through the ULK1 complex. No previous autophagy treatments have been tested in C9orf72-ALS, so this has potential for future studies. The main benefits of this treatment were a lack of off-target side effects, and that it enhanced existing biological processes rather than inducing changes with possible unknown downstream consequences. The dose-dependent response also indicates both a direct connection between the treatment and improvement, and that the response rate can be modulated and controlled. If advanced into clinical treatment, autophagy enhancement therapies could be used to both slow the spread of ALS protein pathology and, if identified early, potentially even prevent it from developing into a full disease state.

TAGS: therapeutic, autophagy, SOD1, human_cell, mouse

References

Baloh, R. H., J. P. Johnson, P. Avalos, P. Allred, S. Svendsen, G. Gowing, K. Roxas, A. Wu, B. Donahue, S. Osborne, G. Lawless, B. Shelley, K. Wheeler, C. Prina, D. Fine, T. Kendra-Romito, H. Stokes, V. Manoukian, A. Muthukumaran, L. Garcia, M. G. Ba?uelos, M. Godoy, C. Bresee, H. Yu, D. Drazin, L. Ross, R. Naruse, H. Babu, E. A. Macklin, A. Vo, A. Elsayegh, W. Tourtellotte, M. Maya, M. Burford, F. Diaz, C. G. Patil, R. A. Lewis and C. N. Svendsen (2022). "Transplantation of human neural progenitor cells secreting GDNF into the spinal cord of patients with ALS: a phase 1/2a trial." Nature Medicine.

Highlander, M. M. A. U. E. S. M. T. I. N.-I. T. S. D. C. S. a. a. N. A. L. S. T. i. A. (2022) "Non-Invasive Transcutaneous Spinal DC Stimulation as a Neurorehabilitation ALS Therapy in Awake G93A Mice: The First Step to Clinical Translation." Bioengineering 9 DOI: 10.3390/bioengineering9090441.

Liu, W., S. O. Zhu, Y. L. Guo, L. F. Tu, Y. Q. Zhen, R. Y. Zhao, L. Ou-Yang, H. Kurihara, R. R. He and B. Liu (2022). "BL-918, a small-molecule activator of ULK1, induces cytoprotective autophagy for amyotrophic lateral sclerosis therapy." Acta Pharmacol Sin.

Lu, S., J. Hu, O. A. Arogundade, A. Goginashvili, S. Vazquez-Sanchez, J. K. Diedrich, J. Gu, J. Blum, S. Oung, Q. Ye, H. Yu, J. Ravits, C. Liu, J. R. Yates and D. W. Cleveland (2022). "Heat-shock chaperone HSPB1 regulates cytoplasmic TDP-43 phase separation and liquid-to-gel transition." Nature Cell Biology.

Ramamoorthy, D., K. Severson, S. Ghosh, K. Sachs, E. G. Baxi, A. N. Coyne, E. Mosmiller, L. Hayes, A. Cerezo, O. Ahmad, P. Roy, S. Zeiler, J. W. Krakauer, J. Li, A. Donde, N. Huynh, M. Adam, B. T. Wassie, A. Lenail, N. L. Patel-Murray, Y. Raghav, K. Sachs, V. Kozareva, S. Tsitkov, T. Ehrenberger, J. A. Kaye, L. Lima, S. Wyman, E. Vertudes, N. Amirani, K. Raja, R. Thomas, R. G. Lim, R. Miramontes, J. Wu, V. Vaibhav, A. Matlock, V. Venkatraman, R. Holewenski, N. Sundararaman, R. Pandey, D.-M. Manalo, A. Frank, L. Ornelas, L. Panther, E. Gomez, E. Galvez, D. Perez, I. Meepe, S. Lei, L. Pinedo, C. Liu, R. Moran, D. Sareen, B. Landin, C. Agurto, G. Cecchi, R. Norel, S. Thrower, S. Luppino, A. Farrar, L. Pothier, H. Yu, E. Sinani, P. Vigneswaran, A. V. Sherman, S. M. Farr, B. Mandefro, H. Trost, M. G. Banuelos, V. Garcia, M. Workman, R. Ho, R. Baloh, J. Roggenbuck, M. B. Harms, C. Prina, S. Heintzman, S. Kolb, J. Stocksdale, K. Wang, T. Morgan, D. Heitzman, A. Jamil, J. Jockel-Balsarotti, E. Karanja, J. Markway, M. McCallum, T. Miller, B. Joslin, D. Alibazoglu, S. Ajroud-Driss, J. C. Beavers, M. Bellard, E. Bruce, N. Maragakis, M. E. Cudkowicz, J. Berry, T. Thompson, S. Finkbeiner, L. M. Thompson, J. E. Van Eyk, C. N. Svendsen, J. D. Rothstein, J. D. Glass, C. N. Fournier, A. Sherman, C. Lunetta, D. Walk, G. Hayat, J. Wymer, K. Gwathmey, N. Olney, S. Ajroud-Driss, T. Heiman-Patterson, X. Arcila-Londono, K. Faulconer, E. Sanani, A. Berger, J. Mirochnick, T. M. Herrington, J. D. Berry, K. Ng, E. Fraenkel, A. L. S. Answer, A. L. S. C. T. C. Pooled Resource Open-Access and A. M. N. H. Consortium (2022). "Identifying patterns in amyotrophic lateral sclerosis progression from sparse longitudinal data." Nature Computational Science 2(9): 605-616.

Streit, L., T. Kuhn, T. Vomhof, V. Bopp, A. C. Ludolph, J. H. Weishaupt, J. C. M. Gebhardt, J. Michaelis and K. M. Danzer (2022). "Stress induced TDP-43 mobility loss independent of stress granules." Nature Communications 13(1): 5480.

Zhang, Z., Q. Fan, X. Luo, K. Lou, W. A. Weiss and K. M. Shokat (2022). "Brain-restricted mTOR inhibition with binary pharmacology." Nature.