Source-linked AI summary
The multilayer connectome of Caenorhabditis elegans
Barry Bentley, Robyn Branicky, Christopher L. Barnes, Edward T. Bullmore, Petra E. Vértes, William R. Schafer
TL;DR
Connectomics must account for extrasynaptic signalling alongside synaptic connections because these routes affect neural activity and behaviour. This study builds draft monoamine and partial neuropeptide connectomes from gene-expression data, finding that synaptic, gap-junction, and neuromodulatory layers form structurally distinct parts of a multiplex neuronal network.
Problem
Connectomics has focused on individual neurons and synaptic connections, although extrasynaptic signalling can profoundly affect neural activity and behaviour.
Method
The study constructs draft monoamine and partial neuropeptide connectomes in C. elegans using new and published gene-expression data.
Results
The extrasynaptic connectomes have topological properties distinct from one another and from the wired connectome.
Takeaways & Limitations
The neuronal connectome can be modelled as a multiplex network with structurally distinct synaptic, gap-junction, and extrasynaptic neuromodulatory layers.
Abstract
from arXiv · showhide
Connectomics has focused primarily on the mapping of synaptic links in the brain; yet it is well established that extrasynaptic volume transmission, especially via monoamines and neuropeptides, is also critical to brain function. Here we present a draft monoamine connectome, along with a partial neuropeptide connectome, for the nematode C. elegans, based on new and published expression data for biosynthetic genes and receptors. Thus, the neuronal connectome can be represented as a multiplex network, with synaptic, gap junction, and neuromodulatory layers representing alternative modes of interneuronal interaction and with distinct network structures. In particular, the monoamine network exhibits novel topological properties, with a highly disassortative star-like structure and a rich-club of interconnected broadcasting hubs. Despite the low degree of overlap between layers, we find highly significant modes of interaction, pinpointing network locations and multilink motifs where aminergic and neuropeptide signalling modulate synaptic activity. The multilayer connectome of C. elegans represents a clear exemplar of a biological multiplex network and provides a prototype for understanding how extrasynaptic signalling can be integrated into the wired circuitry in larger brains.
Author summary
Connectomics has emphasized synaptic links, although extrasynaptic monoamine and peptide signalling also affects neural activity and behaviour. This study adds draft monoamine and partial neuropeptide connectomes for C. elegans and models neuronal communication as a multilayer network.
- Motivation: Extrasynaptic monoamine and peptide signalling are additional modes of interneuronal communication beyond synapses.These pathways can communicate between neurons not connected by synapses.
- Contribution: The study presents draft monoamine and partial neuropeptide connectomes for C. elegans using new and published expression data.The data concern biosynthetic genes and receptors.
- Findings: The extrasynaptic connectomes have topological properties distinct from one another and from the wired connectome.The analysis examines their topology and interactions with synaptic and gap-junction networks.
- Interpretation: The neuronal connectome can be represented as a multiplex network containing synaptic, gap-junction, and neuromodulatory layers.The layers represent interactions with different dynamics and polarity, connected at critical interaction points.
- Implication: The multilayer C. elegans connectome provides a prototype for understanding how neuromodulators interact with wired circuitry in larger nervous systems.The paper specifically connects this prototype to larger nervous systems, including the human brain.
Results
The authors assembled monoamine signalling networks from receptor and biosynthetic-expression data, supplementing incomplete dopamine-receptor maps with reporter profiling. The resulting networks show that monoamine signalling is predominantly extrasynaptic and structurally distinct from wired connectivity.
- Receptor mapping: Additional reporter profiling completed a largely complete map of dopamine receptor expression in the C. elegans hermaphrodite.The three previously incomplete receptor genes were expressed in small, largely distinct neuronal subsets.
- Extrasynaptic signalling: 94% of tyramine-responsive neurons must respond only to extrasynaptic tyramine.Only 7 of 114 receptor-expressing neurons were postsynaptic to a tyraminergic neuron.
- Extrasynaptic signalling: 100% of octopamine-receptor neurons receive no synaptic input from octopamine-releasing neurons, compared with 78% for dopamine and 76% for serotonin receptors.These comparisons support predominantly extrasynaptic monoamine signalling.
- Layer overlap: 96% of 2036 monoamine connections are unique to the monoamine layer rather than overlapping chemical or electrical synapses.Only 89 monoamine connections overlap with wired synapses.
- Network organization: Monoamine networks cluster separately from synaptic and gap-junction networks, whose hubs also have no significant degree correlation with monoamine hubs.The wired and monoamine layers therefore have distinct network organization.
Analysis of monoamine network topology
Monoamine signalling forms a hub-dominated, star-like network that differs from wired circuitry while connecting otherwise segregated regions. Neuropeptide signalling shares some properties but is more clustered and less disassortative.
- Monoamine topology: Monoamine networks contain a few high-degree hubs and exhibit high disassortativity characteristic of star networks.This star-like topology is directly observed across the separate monoamine systems.
- Aggregate network: Adding monoamine connections greatly reduces aggregate path length by linking segregated wired subgraphs.Monoamine signalling directly connects sensory and motor neurons, bypassing prominent premotor interneurons.
- Aggregate network: The aggregate network becomes less modular after monoamine connections are added but remains more modular than random networks.The added links connect functionally segregated units while preserving nonrandom structure.
- Monoamine topology: The monoamine network contains a distinct rich club of dopamine-, serotonin-, and tyramine-releasing neurons.Most serotonergic neurons contain receptors for tyramine and dopamine, and aminergic neurons express receptors for other transmitters.
- Neuropeptide topology: The partial neuropeptide network has shorter paths and lower modularity than wired networks but higher clustering and lower disassortativity than monoamines.Its hubs show no significant degree correlation with hubs in other layers.
Modes of interaction between wired and extrasynaptic layers
Multilink analysis identifies statistically significant ways that wired, monoamine, and neuropeptide connectivity interact despite limited overlap between layers. DVA and RIM occupy central positions across these networks, while specific multilink motifs connect extrasynaptic signalling with synaptic circuitry.
- Multilayer hubs: Although monoamine and wired rich clubs do not overlap, significant links connect them across the multiplex network.The passage reports significant connections between the rich clubs despite their distinct membership.
- Multilayer hubs: DVA receives serotonin, dopamine, and tyramine signalling, while tyraminergic RIM neurons connect strongly with premotor interneurons in the wired rich club.These connections place DVA and RIM at interfaces between neuromodulatory and wired circuitry.
- Multilink motif structure: Seven of 20 multilink motifs were overrepresented and four were underrepresented relative to randomized layers.The analysis compares observed combinations of synaptic, gap-junction, and monoamine links with networks composed from randomized layers.
- Monoamine–synaptic interactions: Monoamine motif 10 combines a unidirectional monoamine link with reciprocal synapses and is significantly overrepresented.The paper suggests this arrangement may support positive or negative feedback and links examples of the motif to pathogen avoidance, escape behaviour, and neuropeptide release.
- Neuropeptide–synaptic interactions: Several motif 20 multilinks belong to the RMG hub-and-spoke network implicated in locomotion, aggregation, and pheromone response.The monoamine counterpart was not overrepresented, perhaps because the monoamine network has low reciprocity.
Discussion
The study expands the C. elegans neuronal connectome with extrasynaptic monoamine and partial neuropeptide networks, revealing distinct topologies and interactions with wired circuitry. These layers are largely separate yet connect through hubs and overrepresented multilink motifs.
- Discussion: The expanded connectome incorporates extrasynaptic monoamine and neuropeptide signalling alongside synaptic and gap-junction networks.The extrasynaptic links were inferred from gene-expression data and add many connections that largely do not overlap with the wired connectome.
- Discussion: Monoamine signalling forms a highly disassortative, star-like network with a small number of high-degree broadcasting hubs and an interconnected rich-club core.These hubs are distinct from, but linked to, the wired connectome’s rich club.
- Discussion: The neuropeptide layer is highly clustered and more reciprocal than the monoamine network, despite being only partially characterized.These properties are associated with cohesion and functional segregation among nervous-system modules.
- Discussion: The inferred neuromodulatory links have largely unknown valence and strength, limiting direct interpretation of their functional effects.The paper notes that comparable information is also lacking for much of the synaptic connectome.
- Discussion: The identified multilink motifs occur in circuits implicated in learning, memory, aggregation, and arousal, while monoamine rich-club structure can coordinate global neural and behavioural states.Additional examples may emerge as more neuropeptide systems are characterized.
A prototype for multiplex network analysis
The C. elegans multiplex connectome combines synaptic, gap-junction, and extrasynaptic layers aligned to the same 302 neurons. Their distinct structures and limited overlap make it a useful prototype for multilayer network analysis.
- A prototype for multiplex network analysis: Existing multiplex datasets are often limited in size or edge-type coverage, whereas this connectome combines multiple interaction layers in one system.This contrast motivates using C. elegans to develop and validate multilayer network concepts.
- A prototype for multiplex network analysis: The multiplex connectome may become a gold standard for multilayer networks, paralleling the wired connectome’s role for monoplex network studies.Its complete node alignment and multiple interaction types support this proposed benchmark role.
- A prototype for multiplex network analysis: The multiplex connectome integrates synaptic, gap-junction, and monoamine layers as distinct communication channels.The layers show limited degree–degree correlation, suggesting they are not merely facets of one underlying network.
- A prototype for multiplex network analysis: The layers operate on different time scales, potentially producing distinct dynamical phenomena.The paper identifies temporal scale as an additional source of multilayer network behavior.
- A prototype for multiplex network analysis: The monoamine and neuropeptide layers add extrasynaptic interactions to the wired connectome.These interactions can be represented across a common set of 302 neuronal nodes.
Prospects for complete mapping of multilayer connectomes
A complete multiplex neuronal connectome is feasible in C. elegans but remains constrained by incomplete expression data and uncharacterized signalling systems. Improved markers and broader receptor mapping could expand the current draft toward a more complete functional network.
- Prospects for complete mapping of multilayer connectomes: The current monoamine network is a reasonable draft, but receptor-expression data indicate that non-synaptic edges are probably undercounted.False negatives appear more likely than false positives, so high-degree monoamine hubs may be understated.
- Prospects for complete mapping of multilayer connectomes: Reporter transgenes can underreport functional expression domains, although recently developed marker strains could help fill missing cell-identification gaps.The limitation is illustrated by incomplete identification of cells expressing receptors such as ser-5.
- Prospects for complete mapping of multilayer connectomes: Additional neuromodulatory layers remain unmapped, including insulin-like peptides, purines, acetylcholine, and GABA.Other monoamines and orphan receptors may also contribute to the signalling network.
- Prospects for complete mapping of multilayer connectomes: Extrasynaptic connectomes for larger brains will be substantially more difficult because larger systems add structural and dynamical influences on extracellular diffusion.Examples include glial barriers, cellular swelling, and arterial pulsations.
- Prospects for complete mapping of multilayer connectomes: C. elegans is unusually suitable for expanding multilayer mapping because its nervous system is small and its cells are precisely characterized.A comprehensive connectome could serve as a prototype for understanding interactions among signalling modes.
Synaptic & gap junction networks
The study combines the full 302-neuron wired connectome with monoamine and partial neuropeptide signalling networks. These neuromodulatory layers are constructed from literature-based gene, receptor, peptide, and cell-expression evidence.
- Synaptic & gap junction networks: The synaptic and gap-junction networks are based on the full hermaphrodite connectome containing all 302 neurons.The wired network combines published somatic and pharyngeal connectome sources.
- Synaptic & gap junction networks: Monoamine networks were mapped by identifying monoamine receptors, transporters, and synthetic enzymes, then collecting cell-level expression data.Expression searches used published literature and curated WormBase and WormWeb databases.
- Synaptic & gap junction networks: Neurons expressing multiple receptors for one monoamine receive a single edge from each sending neuron.The network representation treats reciprocal connections as separate unidirectional connections.
- Synaptic & gap junction networks: Receptor-expressing neurons were identified primarily from cell-body position, shape, and co-labeling with established markers.Additional reporter crosses, dye filling, and marker strains were used to resolve ambiguous identities.
Topological network measures
The analysis represents network layers as directed graphs, compares them with degree-preserving randomized null models, and applies measures of participation, clustering, paths, modularity, assortativity, reducibility, and reciprocity.
- Topological network measures: Edge counts, adjacency matrices, and reducibility clusters were computed from binary directed network layers, excluding self-connections for other measures.Independent layer randomization was used in the analysis.
- Topological network measures: Network measures were compared with 100 degree-preserving edge-swap null networks.Each edge was swapped 10 times to promote randomization while preserving degrees.
- Topological network measures: The normalized degree-rank product identifies neurons with high participation across all network layers.Ranks are scaled to [0, 1] before multiplication, giving a value of 1 to a neuron ranked highest in every layer.
- Topological network measures: Global clustering, or transitivity, measures the ratio of triangles to triples in a network.For directed networks, the calculation uses adjacency structure, in-degree, out-degree, and node-level triangle counts.
- Topological network measures: Characteristic path length is obtained from geodesic distances between node pairs.The geodesic is the minimum-distance path connecting each pair of nodes.
- Topological network measures: Modularity partitions a network into non-overlapping modules to maximize within-module and minimize between-module connectivity.The resulting value is the proportion of edges connecting nodes within the same module.
- Topological network measures: Assortativity measures the correlation between the degrees of nodes at opposite ends of a link.Structural reducibility instead compares relative Von Neumann entropies to quantify layer distinguishability.
- Topological network measures: Reciprocity is the fraction of network edges that are reciprocal.It is defined using the total edge count and the number of reciprocal edges.
Rich-‐club coefficient
The rich-club analysis identifies highly interconnected communities among high-degree nodes by examining progressively filtered subnetworks and comparing their connectivity with randomized networks. The multilink analysis extends this framework across synaptic, gap-junction, and monoamine layers.
- Rich-club coefficient: High-degree nodes are tested for rich-club structure by removing nodes with degree ≤ k and computing the remaining subnetwork’s rich-club coefficient Φ(k).Φ(k) is the ratio of remaining connections to the maximum possible number of connections.
- Multilink motif analysis: Multilink motifs enumerate the possible combinations of links between node pairs across network layers.The analysis follows the multilink concept and examines cross-layer link combinations rather than treating layers only in isolation.
- Multilink motif analysis: The motif analysis uses three layers—synaptic, gap junction, and monoamine—yielding 20 possible multilink motifs.Monoamine layers were restricted because of their conceptual and structural similarity; neuropeptides were handled separately in the supplementary information.
- Multilink motif analysis: Motif instances are counted by traversing all three layers simultaneously and compared with 100 randomized three-layer networks generated by rewiring real-network edges while preserving degree distributions.The randomized networks provide motif z-scores and p-values for comparison with the actual network.
ntr-1 Expr11372 ASH, RIC, ADL, ADF, PVW, PVR,
The receptor-expression records list neuronal locations for several neuromodulatory receptors and note a binding-data caveat for NPR-11/NLP-1.
- The ntr-1 expression record lists ASH, RIC, ADL, ADF, PVW, and PVR among its expressing neurons.
- Additional receptor records include egl-6 expression in HSN, DVA, and SDQ and pdfr-1 expression across multiple sensory and interneuron classes.
- For NPR-11/NLP-1, no EC50 value was reported, although strong biological activity was observed in the micromolar range.