HCN channels reveal conserved and divergent physiology in supragranular pyramidal neurons in primate species

Abstract

The physiological properties of human and rodent neurons differ, yet the extent to which these differences reflect human specializations is often unclear. Compared with their rodent counterparts, human supragranular pyramidal neurons possess enriched HCN-channel-dependent intrinsic membrane properties and a related sensitivity to synaptic inputs containing delta/theta band frequencies. We tested whether other primate species possess enriched HCN-channel dependent membrane properties. We found ubiquitous HCN1 subunit gene expression in supragranular glutamatergic neurons across New World Monkeys, Old-World Monkeys, and great apes in single nucleus RNA-sequencing datasets. Using Patch-seq recordings from acute and cultured brain slices, we found robust HCN-dependent physiological properties in supragranular pyramidal neurons in a species of New-World monkey (Saimiri sciureus) and two species of Old-World Monkey (Macaca mulatta, Macaca nemestrina). In both human and macaque neocortex, HCN-related intrinsic properties increased in magnitude with increasing laminar depth, especially in one transcriptomic cell type. Within this type, HCN dependent properties were more pronounced in macaque than human neurons. These findings indicate that HCN-governed membrane properties and sensitivity to delta/theta band frequencies are roughly conserved in supragranular pyramidal neurons across at least 36 million years of primate evolution.

Figure 1 -. HCN channel-related gene expression is enriched in supragranular glutamatergic neurons in various primate species.

a) Dot plots of HCN1, HCN2, HCN4 and PEX5L expression in six species: mouse, marmoset, macaque, gorilla, chimpanzee and human. Colors and size of the dot plot correspond to the mean gene expression and the fraction of cells expressing the gene for the group, respectively. b) Violin plot of HCN1 expression in layer 2/3 IT and L5 ET neurons for each species. Colors denote mean group expression.

Figure 2 -. Supragranular pyramidal neurons in Old World and New World monkeys display subthreshold membrane resonance.

a) Primate phylogenic tree showing the evolutionary distance among species, expressed in millions of years (MYA). b) Schematic of the experimental workflow. Tissue was resected from both temporal and motor neocortical areas for ex vivo slice preparation and patch-clamp recordings. From a subset of experiments, the nucleus was extracted for RNA sequencing. c) Example voltage response to a chirp current injection for each species. The same chirp stimulus applies to M.mulatta and S.sciureus examples. d) Impedance amplitude profile for these same cells. Arrows denote the resonance frequency for each neuron. e) Number of neurons in each species and neocortical area that had a resonance frequency greater than 2Hz. f) Resonance frequency of recorded cells across areas (temporal and motor) and species (M.nemestrina, S.sciureus, M.mulatta). Resonance frequency was higher in motor cortex for all species (denoted by **, p = 5.82e-07) and did not differ between brain slice preparation (p = 2.48e-01) or species (p = 6.25e-01); 3-Way ANOVA with species, preparation and area as the factors (Supplementary Table 2).

Figure 3 -. HCN-related intrinsic membrane properties are dependent on somatic depth from pia in the temporal cortex of the pig-tailed macaque.

a) We quantified the subthreshold features of supragranular pyramidal neurons throughout the radial depth of the temporal cortex in Macaca nemestrina. b) example voltage response to a chirp stimulus for a superficial and deep neuron and c) corresponding impedance amplitude profiles. d) impedance amplitude profiles averaged across the superficial most one third (green) and deepest one third of neurons (red). e) 3dB cutoff and resonance frequency plotted as a function of distance from pial surface. f) Example membrane response to a series of hyperpolarizing current injections for a superficial and deep neuron. g) Sag ratio, steady-state input resistance and resting potential plotted as a function of distance from the pial surface. All p and R-squared values reflect the depth from pia effect from a 2-way ANOVA (Effect on 3db cutoff: p=4.20e-03, on resonance frequency: p=7.00e-05, on SSRn: p= 1.70e-04, on resting membrane potential p= 4.20e-03, on sag ratio: p=9.26e-07; Supplementary Table 3).

Figure 4 -. Blocking HCN conductance eliminates subthreshold resonance and decreases membrane conductance in supragranular neurons in temporal cortex.

a) We measured the effect of ZD7288 on HCN channel-related properties of supragranular neurons in the temporal cortex of M.nemestrina. b) Left - Hyperpolarizing current injection step and example membrane response of a cell before and after ZD7288 application. middle - Chirp current injection and example membrane response of a cell before and after the bath application of ZD7288. Right - average impedance amplitude profile of all cells (n = 19) before and after ZD7288 application. c) Input resistance for each neuron (dashed lines) before and after ZD7288 application. Mean values are connected by the thick line. d) Percent change in input resistance plotted as a function of distance from the pial surface. e) Resonance frequency for each cell before and after ZD7288 application. f) Change in resonance frequency plotted as a function of percent change in input resistance.

Figure 5 –. Patch-seq reveals divergence and similarities in HCN-dependent properties in temporal cortex of macaque versus human.

a) Patch-seq samples were collected from macaque (M. nemestrina) and human temporal cortex. b) Dendrogram showing that samples predominately mapped to two clusters in the cross-species consensus taxonomy. The bar graphs denote the number of samples mapping to each cluster. c) Depth from pia of each Patch-seq sample by transcriptomic type. d) Spearman’s rho correlations between HCN channel-dependent membrane properties, depth from pia and HCN channel-gene expression for human and macaque samples mapping to L2/3 IT_1. e) Depth from pia and species effect sizes for HCN channel dependent features. f) HCN channel-dependent properties plotted as a function of distance from pia for each species. p values denote significance for depth from pia effect (Supplementary Table 5).