Part Four: Molecular Neuroscience

Molecular neuroscience examines the nervous system at the level of molecules, proteins, genes, and synapses. In molecular science, researchers aim to understand how the biochemical processes within neurons dictate nervous system function, brain development, and cell communication. More specifically, the field zooms in to investigate synaptic transmission, genetics (and epigenetics), and neural development and plasticity [1]. Understanding the nervous system at this microscopic scale provides important insights on brain function and dysfunction.

Through molecular neuroscience, targeted treatments for diseases are possible, alongside the development of precision medicine. By mapping neurotransmitters, ion channels, and receptors, researchers can pinpoint the exact molecular origins of diseases [2]. This allows for the development of highly specific drugs. Additionally, small molecule inhibitors are developed to block abnormal protein folding and aggregation [3].

Techniques like PET (positron Emission Tomography) use radiotracers to bind to amyloid plaques or tau tangles, allowing researchers to visualize disease progression in real time [4]. Finally, molecular neuroscience allows for the identification of biomarkers in the cerebrospinal fluid (CSF) or blood, allowing for early detection of cognitive decline before symptoms appear [5].

With AI, we are able to bridge the gap between genetic architecture and brain function. Below are several examples of how that’s being done.

AlphaFold, an AI system developed by Google DeepMind, predicts the 3D structure of proteins based on their amino acid sequences. AlphaFold is trained on thousands of known protein structures, learning how different amino acid sequences become stable proteins. Additionally, the AI can incorporate physical and chemical constants to output a highly accurate model of the folded protein. With this, researchers can understand the function of individual proteins and how they interact with different molecules [6].

In molecular neuroscience, there is great potential for application. Neurodegenerative diseases, such as Alzheimer’s and Parkinson’s, are caused by misfolded proteins. AlphaFold allows researchers to understand how these proteins misfold and design targeted therapies to prevent it [7].

Single-cell RNA sequencing is technology that helps us understand RNA metabolism at a cellular level. New cell types and states can be identified. This also allows for the discovery of cell-specific differential gene expression changes and gene regulatory networks.

AI is needed to make sense of the large datasets scRNA-seq generates. Large-scale AI models (otherwise known as LCMs, or large cellular models) for single-cell transcriptomics are trained on single-cell RNA-seq data. They have demonstrated superior performance across cell type annotation, data integration, and drug sensitivity [8]

At the molecular level, this has allowed mapping of disease-associated microglia in Alzheimer’s and identification of vulnerable dopaminergic neuron subtypes in Parkinson’s. With scRNA-seq, new perspectives in treating and diagnosing diseases of the central nervous system can be offered. Additionally, new pathophysiological mechanisms, biomarkers, and therapeutic targets can be identified [9].

Despite the potential that AlphaFold and scRNA-seq hold, there is still great limitation. First, the accuracy of predictions made by AlphaFold is an unresolved question [10]. Predictions don’t capture the dynamic conformational changes that proteins undergo in cells, which is necessary to understand how misfolding progresses in disease. While the availability of millions of predicted protein structures provides valuable insights, they may be a barrier for many researchers who may not be familiar with handling macromolecular structure data [11]. Additionally, there are still resolution limitations in the scRNA-seq technology, throughput, and sensitivity [9].

Thus, it’s necessary to provide a comprehensive training platform to make protein structure data more accessible [11]. Additionally, optimizing current protocols of scRNA-seq, alongside through integration with multi-omics, will provide further insight in central nervous system disease progression [9]. Finally, integrating AlphaFold with drug design AI to design novel molecules that target misfolded proteins can accelerate therapeutic devlopment [12].

Play this custom Connections game to check your understanding of molecular neuroscience!