Electrophysiological and molecular diversity of forebrain interneurons

Abstract

Despite their low abundance, telencephalic interneurons demonstrate a great amount of molecular and electrophysiological heterogeneity, suggesting that they exhibit different functions within the circuit. Interneurons are commonly classified into distinct cell types based on their morphology, intrinsic electrophysiological properties, connectivity or molecular profile. However, it is only when combining knowledge about the aforementioned parameters that appropriate classification of cellular and circuit function can be achieved.

In Paper I, we characterized the molecular and electrophysiological diversity of striatal interneurons, either using single-cell-RNA-seq (scRNA-seq) alone or in combination with whole-cell ex-vivo electrophysiological recordings (PatchSeq). Interestingly, unlike in other regions of the brain, striatal Pvalb-expressing cells did not constitute a discrete cluster, but was instead part of the larger group labeled with the gene Pthlh. Using PatchSeq, we were able to show that gradient-like differences in gene expression found within the Pthlh group, correlate with a continuum of electrophysiological properties. In Paper IV, we show that more stable readouts such as anatomical location, morphology, and long-range inputs also gradually differ across this molecular and electrophysiological continuum. Hence, we suggest that separate parts of this gradient exhibit a distinct circuit function. Upon long-term activation of the Pthlh population, we detected an increase in Pvalb expression within the most ventral part of the striatum. Thus, the gradient-like differences in gene expression that we observed within the Pthlh population could be caused by long-term changes in activity upon altered input.

In Paper II, the molecular diversity of hippocampal CA1 interneurons was characterized using scRNA-seq. While the main clusters were clearly separated, similar to the findings of Paper I, many sub-clusters exhibited more of a continuum within the continents. To further study the biological significance of gradient-like differences in gene expression across and within cell types, latent factor analysis was run across all clusters, revealing a common mode of variation across all cell types. Interestingly, the latent factor also seemed to correlate with the axon target location of the corresponding cell type. This suggests that the gradient-like differences in gene expression are most likely reflected in distinct functions within the CA1 circuit. In Paper III, we used a novel approach to bridge transcriptional data to neuronal phenotype and function. By using publicly available datasets that characterize distinct neuronal populations, based on gene expression, electrophysiology, and morphology, we identified cross-cell type correlations between these data modalities. Using multiple PatchSeq datasets, we showed that the gene-property correlations observed across cell types were further predictive of within-cell type heterogeneity.

Taken together, the molecular diversity of interneurons across the telencephalon was observed as discrete clusters or as a continuum within and across cell types. Linking molecular profiles with additional parameters suggests that both continuous and discrete diversity likely reflects distinct functions within the circuit.