The neuron has a complex, branching 3D structure, and so capturing SNSA presents a number of technical challenges in imaging and image analysis that we have had to address in order to lay a firm foundation for the project. The key is to be able to achieve the greatest, most representative coverage of individual neurons and their synapses while maintaining best image quality. This is impacted by several factors, including the thickness of the tissue section being examined and the distance between the individual images captured through its depth. The individual images collected need to be stitched together in a manner that accurately reconstructs the whole neuron and sufficiently reports its SNSA in 3D. We have made substantial progress towards identifying, evaluating and optimising each of these technical parameters and steps that will faithfully output the SNSA.
In parallel, in order to be able to identify synapse diversity on individual neurons, and for different functional types of neurons, we have been creating new genetic tools that will facilitate expression of molecular markers in specific neuron types (‘conditional labelling’). These new research tools tag proteins that are key in synapse structure and function, including PSD93, PSD95, SAP102 and GluR1. They will allow us to align SNSA with neuron function in the brain and will be of substantial utility to the wider neuroscience community.
Fig. 1 provides an example of how we address the challenge of capturing and characterising the SNSA of individual neurons in and among the tangled multitude in the mammalian brain. This example is from a region of the hippocampus called CA1. We first need to ensure that we can see the whole neuron, and for this we use the bright red tdTomato dye, which fills the neuron, dendritic structures and axonal projections. We can visualise the synapses, which we can see as a cloud surrounding the dendrites, by labelling the synapse protein PSD95 with eGFP (green fluorescent protein). The bottom half of Fig. 1 illustrates how we can identify the synapses that belong to an individual dendrite (using clustering analysis), and then look at various features or ‘parameters’ (e.g. their size, intensity, roundness, solidity) of each of these synapses to see how similar or different they are not only from each other, but also from those on other dendrites. Here, the synapses are labelled with PSD95eGFP, but it’s important to note that we will be labelling several different proteins because this is our key measure of synapse molecular diversity (i.e. some synapses will contain PSD95, others PSD93 or SAP102, or any combination of these proteins). Fig. 2 provides examples of diversity in synapse (blue; PSD95 top, SAP102 bottom) morphology and in their distribution along dendrites (red) in cortex layers, hippocampus and further regions of the mouse brain including striatum and thalamus, as well as an impression of molecular diversity (PSD95 (green) + SAP102 (yellow)) in the dentate gyrus (right).
In summary, we have established a method that combines viral cell-specific Cre recombinase delivery with synaptomics as a means to visualise, capture and analyse synapse diversity (molecular composition, morphology, protein lifetime) in specific types of neurons in different regions of the mouse brain. We are now moving forward from stretches of dendrites, developing methods for 3D detection and reconstruction of synaptic puncta and their placement within whole-neuron architecture reconstructed from 2D z-plane images, a key step to capturing SNSA. These new developments in image capture and analysis - collectively, the SYNEURON pipeline - will constitute an important new research tool for the neuroscience community.
