Signature morphoelectric properties of diverse GABAergic interneurons in the human neocortex

INTRODUCTION: Recent studies using single-cell transcriptomic analysis have shed new light on genetically defined cell types in the human brain. However, a deeper understanding of multimodal cellular properties, such as electrical activity and morphology, remains elusive and is central to understanding the role of distinct cell types in cognitive function and disease.RATIONALE: Understanding the fundamental properties of different cell types is important for gaining insights into their role in neural circuits, cognition, and disease. This is challenging for cell types in the human brain because of the limited availability of tissue, the heterogeneity of neurons, and the lack of genetic approaches to effectively target these spatially distributed and in some cases, exceedingly rare types. Rapid viral genetic labeling and patch-clamp electrophysiology combined with RNA sequencing (Patch-seq) facilitates the targeting of specific cell types in tissue from human neurosurgeries and thus allows the characterization of multimodal properties of neurons in the human cortex.RESULTS: Patch-seq sampling facilitated targeted acquisition and analysis of 778 human neurons in cortical layers 2 to 6, across 44 out of 45 gamma-aminobutyric acid-producing (GABAergic) transcriptomic types. Aggregated data from acute and culture paradigms provide new descriptions and direct comparison of the signature morphoelectric properties of the canonical human interneuron subclasses, LAMP5/PAX6, VIP, SST, and PVALB, and a deeper dive into features of select GABAergic transcriptomic types. Detailed characterization of the SST subclass revealed specific multimodal properties of individual transcriptomic types. The SST FRZB transcriptomic type exhibits gene expression signatures along a continuum of PVALB and SST subclasses, but the morphoelectric properties of this type clearly indicate strong alignment with the PVALB subclass. For an additional transcriptomic type, SST CALB1, we found multiple discrete morphological types suggesting that further splitting may be warranted. Double-bouquet cells of the human neocortex belong to two transcriptomic types within the SST subclass with enrichment in temporal versus frontal cortex regions. We compared the morphoelectric properties of homologous mouse and human neocortical GABAergic neuron types and found that human types are more excitable and have a larger spatial extent with less neurite branching.CONCLUSION: An impactful finding of our study is the direct demonstration of how multimodal Patch-seq data is vital to refinement of transcriptomic cell-type taxonomies. Cellular taxonomies built on single-cell transcriptomes and differential gene expression are not static but rather represent a starting foundation to build upon as new data modalities are obtained and aligned at the resolution of transcriptomic types. We also demonstrate the immense potential and utility of viral genetic labeling and Patch-seq for targeted analysis of human neocortical GABAergic neuron subclasses and types in ex vivo brain slices. This work can serve as a roadmap for future functional studies of human brain cell types at the resolution of emerging transcriptomic cell-type taxonomies and provides a rich open-access dataset for exploring gene-function relationships for a wide diversity of human neocortical GABAergic neuron types.