Oxford Global’s panel discussion ‘Translational Approaches and Biomarkers in Neurodegeneration’, at Discovery Europe 2023 touched on some of the major issues that the world of neuroscience is facing when it comes to biomarkers and preclinical modelling. Leading the discussion was moderator, Giovanna Lalli, a Director of Scientific Affairs at the UK Dementia Research Institute.
Joining the panel was Victoria Demina, Head of R&D and Head of Production at Neuway Pharma. Her team have been working for over 15 years to establish a nanoparticle delivery system. With a background in cell biology and genetics, Demina has worked on the production of nanoparticles since 2007. She has pushed the project from scratch, through to the early stages of development in academia, and beyond, establishing GMP in pharma and trying to find the best targets.
- Challenges in Improving Patient Outcomes: Why Biomarkers Are Difficult to Commercialise
- 3D Cell Model Approaches for Disease Monitoring and Detection
- Investigating Target Identification and Validation Modelling in Neurodegenerative Disease
Rosa Maria Rodriguez Sarmiento is an Expert Scientist in Medicinal Chemistry at F. Hoffmann-La Roche. She works on small molecule compound optimisation mostly within CNS indications. Her focus for 20 years is within the CNS area drug discovery. For many years she has been working on amyloid-β targets and its relation to neurodegenerative disease. In this regard, Roche has investigated gamma secretase inhibitors and modulators for the disease areas Alzheimer’s, in addition she has been involved in projects on the area of Parkinson’s disease or other rare diseases, and many kinds of targets related to neurodegeneration. Rodriguez identified the current challenges of her field as being finding the right diagnoses, identifying the correct patient population, and being more successful with target engagement.
Gergely Toth has a background in computational drug design, biophysics, biochemistry structural biology, all in the interest of small molecule drug discovery. Toth started working on Alzheimer’s and Parkinson’s drug discovery 15 years ago and founded Cantabio Pharmaceuticals in 2009 where he currently serves a CEO & CSO. Cantabio’s aim is to tackle neurodegeneration by targeting proteins such as alpha-synuclein or tau in their monomeric state and reduce their aggregation and misfolding.
Cantabio is also very interested in the role that biochemical stresses such as oxidative and glyoxal stress have to play in the onset of neurodegenerative diseases, with a view towards creating methods to reduce these stresses synergistically in diseases. To do this, the company is targeting a protein called DJ-1 and have developed various therapeutic programs and an ELISA platform to quantify its specific oxidation states. With this, Toth’s company are working towards developing DJ-1 as a biomarker for Parkinson’s and possibly other diseases. Ultimately, the aim is that it may be used in the future as an early diagnostic blood test for Parkinson’s disease.
Key Challenges Stunting Progress in Neurodegenerative Disease Research
The first question asked of the group as what the biggest challenge stunting the progress of neurodegenerative disease research was. Demina said that she thought that this was identifying the right target, not in terms of biomarkers but in terms of disease. She added that having limited resources and a small team makes it hard to mine large datasets for disease targets.
In Rodriguez’s opinion, increasing understanding of the underlying biology is a large hurdle that researchers have to face. “We need to understand more of the biology and the translation between an in vitro system (where you can measure different species of amyloid-β) and the in vivo situation.”
“There are so many; where to start?” said Toth. He said that there was no shortage of good targets in neurodegenerative disease drug discovery: “we have some really good targets such as tau or A-β.” The challenge, however, is developing the technology to engage those targets.
“One option is to use immunotherapy: antibodies which target cellular proteins in the brain, but this has been very difficult. On the other hand, small molecules are also difficult because they have never been used to target these kinds of proteins before. Furthermore, the tractability of good targets which are involved in the pathology presents a problem.”
Translating In Vitro to In Vivo Modelling
Of course, it’s very complicated to find the right system to work with in order to show the deliverability of a drug. Demina said that her approach established BBB modelling for delivery to the brain in the lab before addressing the target afterwards in an in vitro system. Their team now use iPSCs (induced pluripotent stem cells) to create a more robust system for mimicking brain sites. “Before that, it was very challenging. We used isolated cells from mice,” said Demina. She added that the more widespread availability of iPSC to smaller companies has helped a lot in this implementation.
Secondly, the in vivo situation needs to be addressed. Researchers need to be confident that they understand their targets and which animal model is best to use. “A lot of animal models do not work, for example mice are too small to inject nanoparticles,” explained Demina. Furthermore, she stressed the importance of having a predictive and translatable model: “You also need to ensure that your biomarker is good enough so that it can be used with children as young as 1 to 3 years old.”
Rodriguez commented on the use of in vivo animal models vs in vitro iPSCs . “We usually work with Alzheimer’s models that over express Aβ-42, and when we want to check for cognition or other clinical end points, it is not possible, it can’t really be translated into humans,” she said. In addition, due to the fact that in most cases these models need months to be prepared, are transgenic models or need to be used in a six-month test, Rodriguez said that they were therefore not suited to high throughput screening. “That makes it hard to work out even an initial connection between your in vitro and in vivo profile”
Furthermore, it needs be considered that Alzheimer’s is a chronic disease which means that the mandatory safety requirements for its medicines will be much higher. Rodriquez explained that there are some inadmissible aspects when it comes to the toxicity data. Where’s the solution? One possibility is the use of new technologies like organoid models to check the target engagement and safety in addition to suitable preclinical studies.
But this is not only a problem in Alzheimer’s disease; Rodriguez also worked for many years on targets for cognition impairment associated with schizophrenia (or CIAS) that are also challenging to mimic on in vivo settings. For example one of the issues with studying cognition impairment is the translation across species, in addition there are many memory types that need to be considered; these include learning memory, working memory, and associative memory, among others. “So you need to use appropriate cognition models and if you are looking to implement digital biomarkers for example, you need to understand the memory type that you need to be checking,” said Rodriguez.
Problems Modelling Parkinson’s Disease
Applying the conversation to Parkinson’s disease, Toth said that generally in vitro and in vivo models don’t properly recapitulate all of the disease, and that is a main concern. “Just this morning we had a discussion about the uptake of alpha-synuclein fibrils in primary neurons and there was no toxicity or neurodegeneration,” he remarked. Herein, getting a working and translatable cell model is an extremely complicated process and, in the end, will only recapitulate part of the disease.
Toth recounted how he has spent 15 years working on Parkinson’s disease drug discovery, he said that the lack of appropriate cell models is an issue that occurred during his first biology meeting – and is still true today. “Hopefully iPSC derived models will help fill this gap, but it is clearly going to take a long time before these can be applied easily in practice.”
One of the problems for models of in vivo pharmacology is the fact that they only present a very limited part of the disease indication. “They over-express one gene, for example synuclein in Parkinson’s models, but there are probably 500 more proteins that get aggregated in Lewy bodies. So it’s a synuclein model, not a Parkinson’s model.”
The panel discussion shed light on the significant challenges faced by researchers in the field. Identifying the right disease target, understanding the underlying biology, and developing effective technologies for target engagement emerged as key obstacles. The translation from in vitro to in vivo models also presented difficulties, with the need for robust and predictive systems that accurately mimic brain sites.
Furthermore, the complexity of mimicking and safely treating chronic diseases like Alzheimer’s and the difficulties to obtain proper translational models in diseases that involve cognition impairment such as in Alzheimer’s cases or schizophrenia with diverse memory types involved added further complexity to the research process. Despite advancements such as iPSC-derived models, there is still a long road ahead in addressing these challenges and advancing neurodegenerative disease research.
Join and network with over 250 industry leaders at Discovery US, where we will address the latest advancements in target identification, validation and HIT optimisation.
The 2-day programme addresses drug discovery in new modalities such as protein degradation, and case studies on the integration of screening approaches in phenotypic and genomics-based discovery, Organ-on-a-chip and 3D modelling.