Morpho-electric and transcriptomic divergence of the layer 1 interneuron repertoire in human versus mouse neocortex

Abstract

Neocortical layer 1 (L1) is a site of convergence between pyramidal neuron dendrites and feedback axons where local inhibitory signaling can profoundly shape cortical processing. Evolutionary expansion of human neocortex is marked by distinctive pyramidal neuron types with extensive branching in L1, but whether L1 interneurons are similarly diverse is underexplored. Using patch-seq recordings from human neurosurgically resected tissues, we identified four transcriptomically defined subclasses, unique subtypes within those subclasses and additional types with no mouse L1 homologue. Compared with mouse, human subclasses were more strongly distinct from each other across all modalities. Accompanied by higher neuron density and more variable cell sizes compared with mouse, these findings suggest L1 is an evolutionary hotspot, reflecting the increasing demands of regulating the expanding human neocortical circuit.

Signature morpho-electric properties of diverse GABAergic interneurons in the human neocortex

“…In contrast, we identified a notable gene expression signature in a subset of our dataset corresponding to a module of microglia related genes. Surprisingly, this identical signature was observed for both neurons recorded in acute and cultured slices (but more pronounced in culture) and has been similarly observed in other human and mouse acute brain slice Patch-seq datasets ( 15 – 17 , 35 , 36 ). Further work is warranted to understand the cause and consequences of this gene expression signature in brain slice experiments.…”

“…Human samples were obtained from neurosurgically resected neocortical tissues and processed with standardized protocols across three experimental sites in the U.S., Netherlands and Hungary ( 14 , 15 , 18 , 35 ); most samples originated from the temporal and frontal lobes, along with smaller fractions in parietal, occipital and cingulate areas (Data S1). Human Patch-seq neurons were mapped to a middle temporal gyrus (MTG) single nucleus transcriptomics-based reference taxonomy ( 1 ) and assigned to a transcriptomic subclass and type using a “tree mapping” classifier (see Methods) ( 16 ).…”

“…Although they are found in all cortical layers, we report predominately on neurons in L2 and L3. An in-depth investigation of L1 cells can be found in Chartrand et al 2022. LAMP5/PAX6 neurons are distinguished by their stellate dendrites, numerous primary dendrites, and extensive axon branching.…”

“…This taxonomy consists of a hierarchical dendrogram of cell types, along with a set of marker genes defined to distinguish types at each split in the tree. The Patch-seq transcriptomes were mapped to the reference taxonomy following the ‘tree mapping’ method (map_dend_membership in the scrattch.hicat package) as described previously ( 35 , 45 ). Briefly, at each branch point of the taxonomy we computed the correlation of the mapped cell’s gene expression with that of the reference cells on each branch, using the markers associated with that branch point (i.e., the genes that best distinguished those groups in the reference), and chose the most correlated branch.…”

Transcriptomic cytoarchitecture reveals principles of human neocortex organization

“…These analyses provide a new map for the field to understand organizational principles and place a cellular lens on thinking about cortical functional variation as variation in the proportions and properties of the component cell types that define the input-output properties of those areas. Recent studies have shown that morphological and anatomical characteristics are correlated with transcriptomic identity and gradient properties ( 16 , 24 , 66 ), indicating that the transcriptomic maps are also highly predictive for cell phenotype variation. Challenges for the future will be to map the entire human neocortex, understand graded features versus discrete boundaries, and directly measure the relationship between transcriptomically defined cell types, cellular phenotypes and functional architecture.…”