A call for a unified and multimodal definition of cellular identity in the enteric nervous system

Abstract

The enteric nervous system (ENS) is a tantalizing frontier in neuroscience. With the recent emergence of single cell transcriptomic technologies, this rare and poorly understood tissue has begun to be better characterized in recent years. A precise functional mapping of enteric neuron diversity is critical for understanding ENS biology and enteric neuropathies. Nonetheless, this pursuit has faced considerable technical challenges. By leveraging different methods to compare available primary mouse and human ENS datasets, we underscore the urgent need for careful identity annotation, achieved through the harmonization and advancements of wet lab and computational techniques. We took different approaches including differential gene expression, module scoring, co-expression and correlation analysis, unbiased biological function hierarchical clustering, data integration and label transfer to compare and contrast functional annotations of several independently reported ENS datasets. These analyses highlight substantial discrepancies stemming from an overreliance on transcriptomics data without adequate validation in tissues. To achieve a comprehensive understanding of enteric neuron identity and their functional context, it is imperative to expand tissue sources and incorporate innovative technologies such as multiplexed imaging, electrophysiology, spatial transcriptomics, as well as comprehensive profiling of epigenome, proteome, and metabolome. Harnessing human pluripotent stem cell (hPSC) models provides unique opportunities for delineating lineage trees of the human ENS, and offers unparalleled advantages, including their scalability and compatibility with genetic manipulation and unbiased screens. We encourage a paradigm shift in our comprehension of cellular complexity and function in the ENS by calling for large-scale collaborative efforts and research investments.

Figure 1:. Primary mouse and human enteric neuron datasets used for cross dataset comparison

A-D) UMAPs of enteric neurons generated from the original datasets of UM-mouse (A), AR-mouse (B), AR-human (C) and ST-human (D). “?” refers to the cluster labeled as “ENC11” or “?” in Morarach et al. E) Total number and distribution of enteric neuron cluster annotations A-D.

Figure 2:. Cross dataset expression of primary enteric neuron cluster specific markers

A) Venn diagram indicating shared markers used by each study. B-E) Violin plot stack of cluster specific markers originally used for UM-mouse, B) in UM-mouse, C) in AR-mouse, D) in ST-human, E) in AR-human. F-I) Violin plot stack of cluster specific markers originally used for ST-human, F) in UM-mouse, G) in AR-mouse, H) in ST-human, I) in AR-human. J-M) Violin plot stack of cluster specific markers originally used for AR-mouse and AR-human, J) in UM-mouse, K) in AR-mouse, L) in ST-human, M) in AR-human. OR: B-J) Violin plot stack of cluster specific markers originally used for UM-mouse (A-D), ST-human (E-H) ST-human and AR-mouse and human (I-L) all the datasets used in this analysis. N-P) Comparison of the distribution of cluster annotations for enteric neurons expressing (N) Nos1 or NOS1, (O) Tac1 or TAC1, and (P) Penk1 or PENK1 transcripts across primary mouse and human ENS datasets used in this study.

Figure 3:. Unbiased cross dataset classification of primary enteric neurons using SingleCellNet

A) Schematics of unbiased label transfer using SCN. B-M) Reference primary enteric neuron scRNA-seq datasets of mouse (UM-mouse (C, L, J), AR-mouse (B, K, M)) and human (ST-human (D, F, H), AR-human (E, G, I)) were used to train SingleCellNet. These models were then used for label transfer and cross annotation in the other datasets.

Figure 4:. Cross-dataset module scoring for transcriptional signatures of primary enteric neuron clusters

A) Heatmap of the average module scores of ST-human and AR-human neuronal subtype transcriptional signatures in UM-mouse. B) Heatmap of the ST-human, AR-human and UM-mouse neuronal subtype transcriptional signatures in AR-mouse. C) Heatmap of the average module scores of AR-human and UM-mouse neuronal subtype transcriptional signatures in ST-human. D) Heatmap of the average module scores of ST-human and UM-mouse neuronal subtype transcriptional signatures in AR-human.

Figure 5:. Comparative analysis of primary ENS neurons using Spearman correlation analysis

A-B) Heatmap matrix of Spearman correlations based on expression of 100 anchor features shared significantly variable genes (or anchor features) between A) primary human (ST-human and AR-human) and B) primary mouse (UM-mouse AR-mouse) enteric neuron subtypes.

Figure 6:. Harmony integration of primary mouse and human enteric neuron datasets.

A) Schematic representation of Harmony integration of UM-mouse and AR-mouse datasets. B, C) Distribution of cells derived from UM-mouse and AR-mouse datasets (B) and their respective broad functional annotations in each Harmony cluster (C). D) Schematic representation of Harmony integration of ST-human and AR-human datasets. E, F) Distribution of cells derived from ST-human and AR-human datasets (E) and their respective broad functional annotations in each Harmony cluster (F).

Figure 7:. Comparative analysis of primary ENS neurons using GOBP hierarchical clustering

Hierarchical clustering of primary enteric neuron clusters based on normalized enrichment scores of biological process gene ontology (GOBP) pathways. Blue boxes indicate closely clustered neuronal subtypes with matching functional annotation from two different datasets.

Figure 8:. Transcriptional identities are not synonymous with functional identities in enteric neurons

Accurate characterization and appropriate annotation of enteric neuron functions require moving beyond relying on a single biological modality, such as transcription (Figure 9). By adopting a multidimensional approach, we can go beyond our current understanding of enteric neuron identity and unravel the complexities of the ENS.

Figure 9:. Enteric neuron identity should be defined based on multiple biological features

These profiling studies should be performed on diverse ENS samples that encompass a broader range of gut regions, animal models, developmental and disease states, gender and ethnic diversities ^,. In parallel, the development of improved computational methods, including dataset integrations and label transfer algorithms, can facilitate the identification and classification of distinct ENS cell types and subtypes. These computational approaches can help uncover the shared and unique features of enteric neurons obtained from various sources. Moreover, integration and side-by-side analysis of tissue datasets from different sources, particularly human patients, requires caution due to potential biological and pathological differences, cell type variations, and diverse cell identities. Overlooking these factors can result in misleading analyses and inaccurate conclusions. In addition, the incorporation of spatial profiling and advanced imaging techniques, such as tissue clearing and organ-level imaging, holds promise for elucidating the spatial organization and connectivity of enteric neurons within the gut. It will ultimately be essential to unify the transcriptomic identities with the functional roles of enteric neurons. One route to this goal would be Patch-seq, which combines whole-cell patch clamp electrophysiological recordings and pharmacological analysis, with transcriptomic analysis. Here, following electrophysiological recording of myenteric and submucosal neurons, the patch electrode would be used to collect the nucleus from the recorded cell for downstream snRNA-seq. Dye labelling of the recorded cell would also enable post-hoc tracing of its projections. While these would be technically challenging and low-throughput studies, together with spatial and imaging techniques, these approaches can reveal the regional diversity and spatial relationships of ENS cells, enhancing our comprehension of their functional roles.