Back

A systematic cross-modal approach identifies astrocytic VCAM1 as a regulator of hippocampal synapse development

Geyskens, C.; Kotoula, E.; Bogaert, K.; Leysen, E.; Vandenbempt, J.; Pestana, F.; Lefever, L.; Magalhaes, P.; Dascenco, D.; de Wit, J.

2026-03-18 neuroscience
10.64898/2026.03.17.712320 bioRxiv
Show abstract

Astrocyte-mediated cell-cell interactions at synapses are essential for circuit formation, yet the molecular mechanisms underlying astrocyte-mediated regulation of hippocampal synapse development remain incompletely characterized. Here, we take a systematic cross-modal approach to identify novel astrocytic cell surface proteins (CSPs) at the tripartite synapse. By using single-cell spatial transcriptomics (scST) targeting CSPs identified in a previous hippocampal synaptic dataset, we find 10 potential candidate astrocytic CSPs. Subsequent systematic protein-level profiling using multiple antibody-based assays establishes GPR37L1, HepaCAM, and VCAM1 as astrocytic peri-synaptic CSPs. To gain insight into the molecular context in which these proteins operate at synapses, we map their synaptic interaction partners using affinity purification-mass spectrometry (AP-MS), revealing distinct hippocampal synaptic interactomes for GPR37L1 and VCAM1. We then develop a custom multi-metric image-based synapse analysis pipeline to assess the roles of GPR37L1 and VCAM1 in synaptic development, using CRISPR/Cas9-mediated gene knockout (KO) in the mouse hippocampus. While loss of GPR37L1 does not substantially affect excitatory and inhibitory synapses, VCAM1 loss impairs excitatory hippocampal synapse development. Conversely, addition of recombinant VCAM1 to cultured hippocampal neurons increases the density of excitatory synapses. Together, these results identify astrocytic VCAM1 as a regulator of hippocampal synapse development.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.