Neuropeptide signaling network of Caenorhabditis elegans: from structure to behavior

Abstract

Neuropeptides are abundant signaling molecules that control neuronal activity and behavior in all animals. Owing in part to its well-defined and compact nervous system, Caenorhabditis elegans has been one of the primary model organisms used to investigate how neuropeptide signaling networks are organized and how these neurochemicals regulate behavior. We here review recent work that has expanded our understanding of the neuropeptidergic signaling network in C. elegans by mapping the evolutionary conservation, the molecular expression, the receptor–ligand interactions, and the system-wide organization of neuropeptide pathways in the C. elegans nervous system. We also describe general insights into neuropeptidergic circuit motifs and the spatiotemporal range of peptidergic transmission that have emerged from in vivo studies on neuropeptide signaling. With efforts ongoing to chart peptide signaling networks in other organisms, the C. elegans neuropeptidergic connectome can serve as a prototype to further understand the organization and the signaling dynamics of these networks at organismal level.

In this review, Watteyne et al. focus on functional insights on neuropeptide signaling gained from comprehensive mapping of the neuropeptide signaling network and behavioral studies in the nematode Caenorhabditis elegans. The authors summarize recent discoveries on the evolutionary conservation of neuropeptide systems, the structure of the peptide signaling network, as well as functional knowledge and peptidergic circuit motifs emerging for behavioral studies in C. elegans.

Introduction

Neuropeptides are ubiquitous signaling molecules that have important functions in the regulation of animal physiology and behavior. They constitute one of the largest and most diverse groups of neuronal messengers that mediate neuroendocrine signaling and extrasynaptic communication in nervous systems (van den Pol 2012; Taber and Hurley 2014). Most neurons express and release 1 or multiple peptide messengers through exocytosis of large dense-core vesicles (DCVs; Williams et al. 2017; Deng et al. 2019; Smith et al. 2019; Taylor et al. 2021; Zhong et al. 2022; Thiel et al. 2024). These secreted peptides predominantly bind to G protein-coupled receptors (GPCRs) on neighboring neurons or on more distant target cells to regulate a variety of physiological processes and brain functions, such as feeding, sleep, arousal, reproduction, and learning (Taghert and Nitabach 2012; Powers et al. 2019; Nässel and Zandawala 2020; Shen et al. 2022). While neuropeptides have long been recognized as important neurochemicals, recent studies mapping their brain-wide expression and impact on neuronal activity have highlighted the extent of this vast extrasynaptic signaling network (Smith et al. 2019; Taylor et al. 2021; Zhong et al. 2022; Randi et al. 2023; Ripoll-Sánchez et al. 2023). Gene expression studies in diverse animals show that both neuropeptides and peptide-activated GPCRs are widely expressed throughout nervous systems (Yew et al. 2009; Williams et al. 2017; Smith et al. 2019; Taylor et al. 2021; Styfhals et al. 2022; Thiel et al. 2024). In addition, efforts to comprehensively map neuropeptide networks have provided a framework for system-level studies of peptide signaling and for uncovering organizational features that are potentially conserved across species (Williams et al. 2017; Deng et al. 2019; Smith et al. 2019; Nässel and Zandawala 2020; Smith 2021; Ripoll-Sánchez et al. 2023; Thiel et al. 2024).

In this review, we discuss recent insights into the structure and functional organization of the neuropeptide signaling network in Caenorhabditis elegans. In addition to its synaptic connectome (Albertson et al. 1976; White et al. 1986; Jarrell et al. 2012; Cook et al. 2019; Witvliet et al. 2021), a first brain-wide map of neuropeptide signaling has been reconstructed in the nematode (Taylor et al. 2021; Beets et al. 2023; Ripoll-Sánchez et al. 2023). Moreover, its genetic tractability facilitates investigations into the broad range of neuropeptide actions. We discuss the diversity and evolutionary conservation of peptidergic systems and cover insights into the structure of neuropeptide networks obtained from comprehensive mapping of the C. elegans neuropeptidergic connectome. We also delve into the spatiotemporal scales of neuropeptide signaling and describe peptidergic circuit motifs that underpin the modulation of C. elegans behavior. While we focus on peptide–GPCR signaling between neurons, the principles we discuss may extend to insulin-like peptides that signal via receptors other than GPCRs, as well as to nonneuronal cells, which can both send and receive peptide signals (Cohen et al. 2009; Lee and Mylonakis 2017; Palamiuc et al. 2017; Zheng et al. 2018; De Rosa et al. 2019; Singh and Aballay 2019; Mutlu et al. 2020; Biglou et al. 2021; Florman and Alkema 2022).

Diversity and evolutionary conservation of peptidergic signaling in C. elegans

The repertoire of peptide messengers and receptors is highly diverse but shows significant similarities between animals (Jékely 2013; Mirabeau and Joly 2013; Elphick et al. 2018). In C. elegans, 160 peptide-encoding genes have been identified, by means of bioinformatics and peptidomics, which give rise to more than 300 peptides through enzymatic cleavage and processing of the encoded polypeptide precursors (Nelson, Kim et al. 1998; Nathoo et al. 2001; Husson et al. 2007; Li and Kim 2008; Yamada et al. 2010; Yin et al. 2017; Van Bael et al. 2018; McKay et al. 2022; Cockx et al. 2023). These peptides have been classified into 3 groups based on sequence features: Peptides belonging to the FMRFamide-like peptide family (31 flp genes), peptides related to the insulin family (40 ins genes), and non-RFamide, noninsulin peptides that are named neuropeptide-like proteins (89 nlp genes). This classification, however, does not imply peptides from the same group to be evolutionarily related (Mirabeau and Joly 2013; Fadda et al. 2019; Beets et al. 2023). The flp-34 gene, for example, is considered orthologous to the vertebrate neuropeptide Y (npy) and invertebrate neuropeptide F (npf) genes, as FLP-34 peptides show a conserved NPY/NPF sequence motif and the flp-34 precursor gene contains a conserved exon–intron junction that is characteristic of npy/npf genes (Fadda et al. 2019, 2020). FLP-34 peptides are also the most potent ligands of the NPY/NPF receptor ortholog NPR-11, which clusters closest with invertebrate NPF receptors (NPFRs) as shown by phylogenetic tree analysis (Jékely 2013; Mirabeau and Joly 2013; Gershkovich et al. 2019; Fadda et al. 2020; Beets et al. 2023). Other C. elegans FLP peptides seem to be more closely related to evolutionarily distinct neuropeptide families, such as FLP-18 and FLP-21 that show sequence similarity to the invertebrate short neuropeptide F (sNPF) family and are the most potent ligands of C. elegans neuropeptide receptors (NPR-1 to NPR-6, and NPR-10) that cluster within the sNPF receptor family (Kubiak et al. 2003; Rogers et al. 2003; Cohen et al. 2009; Jékely 2013; Mirabeau and Joly 2013; Fadda et al. 2019; Beets et al. 2023).

In addition to neuropeptide-encoding genes, the C. elegans genome codes for over 150 putative peptide-activated GPCRs (Beets et al. 2023). The majority cluster with peptide GPCR families that are conserved across Protostomia or Bilateria in large-scale phylogenetic analyses (Table 1; Jékely 2013; Mirabeau and Joly 2013; Beets et al. 2023). However, annotations of the first invertebrate genome sequences revealed relatively few vertebrate-like peptides, which initially led to the assumption that vertebrate peptidergic systems are not well conserved in invertebrates (Bargmann 1998; Hewes and Taghert 2001; Nathoo et al. 2001). Later, phylogenetic studies revealed deep conservation of many peptide systems across bilaterian animals, including in C. elegans (Jékely 2013; Mirabeau and Joly 2013; Elphick et al. 2018). Receptor deorphanization studies further confirmed the evolutionary conservation of these peptidergic systems, by experimentally demonstrating biochemical interactions between peptide and receptor orthologs (Beets et al. 2012, 2023; Garrison et al. 2012; Ohno et al. 2017; Van Sinay et al. 2017; Sakai et al. 2021). Based on these phylogenetic and receptor deorphanization studies, C. elegans orthologs of ancestral peptide families have been classified and named according to their evolutionary relationships (Table 1; Beets et al. 2023; Istiban et al. 2024). To date, at least 31 peptide–GPCR signaling systems are known to be widely conserved in bilaterian animals, 17 of which have been identified in C. elegans (Jékely 2013; Mirabeau and Joly 2013; Elphick et al. 2018; Beets et al. 2023). These include orthologs of well-known vertebrate neuropeptides, like vasopressin/oxytocin, tachykinin, cholecystokinin, gonadotropin-releasing hormone, thyrotropin-releasing hormone, and NPY/NPF (Janssen et al. 2008; Lindemans et al. 2009; Beets et al. 2012, 2023; Garrison et al. 2012; Van Sinay et al. 2017; Fadda et al. 2020). Many of these peptidergic systems have conserved functions, such as in the regulation of growth, feeding, and reproductive behaviors (reviewed in Istiban et al. 2024). In addition, the C. elegans genome encodes peptides that are related to protostomian neuropeptide families, such as invertebrate sNPF and myosuppressin-like peptides (Jékely 2013; Mirabeau and Joly 2013; Fadda et al. 2019; Beets et al. 2023). Some peptide and receptor families have expanded in gene number in nematodes, including the RFamide, sNPF, myosuppressin, somatostatin/allatostatin C, and gonadotropin-releasing hormone systems (Jékely 2013; Mirabeau and Joly 2013; Beets et al. 2023).

Table 1.

Evolutionary conserved peptide–receptor systems in C. elegans.

Peptide GPCR familya	C. elegans receptors in familyb	Peptide ligand(s)c
Vasopressin/oxytocin (VP/OT)	NTR-1d, NTR-2	NTC-1 (NLP-75)
Tachykinin (TK)	TKR-1, TKR-2	TACH-1 (NLP-58)
Gonadotropin-releasing hormone/adipokinetic hormone/adipokinetic hormone-corazonin-like peptide (GnRH/AKH/ACP)	GNRR-1	AKHP-1 (NLP-47)
Cholecystokinin/sulfakinin (CCK/SK)	CKR-1d, CKR-2	CCKP-1 (NLP-12)
Neuromedin U/pyrokinin/capability/pheromone biosynthesis activating neuropeptide (NMU/PK/CAPA/PBAN)	NMUR-1, NMUR-2	CAPA-1 (NLP-44)
	NMUR-3	orphan
Neuropeptide Y/neuropeptide F (NPY/NPF)	NPFR-1 (NPR-11)d	NPFP-1 (FLP-34)
	NPFR-2 (NPR-12)d	NPFP-2 (FLP-33)
Calcitonin/diuretic hormone 31 (Calc/DH31)	DHRR-1 (SEB-2)	CTDH-1 (NLP-73)
Orexin/allatotropin (Ox/AT)	NPR-14	orphan
Neuropeptide FF/SIFamide (NPFF/SIFa)	SIFR-1 (NPR-35)	SIFP-1 (NLP-10)
Galanin/allatostatin A (Gal/AstA)	NPR-9	orphan
Thyrotropin-releasing hormone/fulicin (TRH/Ful)	TRHR-1	TRH-1 (NLP-54)
Leucokinin (LK)	TKR-3	orphan
RWamide/RYamide/luqin (RWa/RYa/Luq)	LUQR-1 (NPR-22)d	LURY-1 (NLP-72)
Pigment dispersing factor/cerebellin (PDF/Cer)	PDFR-1	PDF-1 (NLP-74), PDF-2 (NLP-37)
Allatostatin B/myoinhibiting peptide/proctolin (AstB/MIP/Prct)	MIPR-1 (SPRR-1)	MIP-2 (NLP-42)
	MIPR-2 (SPRR-2)	MIP-1 (NLP-38)
	SPRR-3, T21H3.5	orphan
Elevenin (L11)	ELVR-1 (NPR-34)	ELVP-1 (SNET-1)
Pigment dispersing factor receptor-like (PDFR-like)	SEB-3	NLP-49
Short neuropeptide F (sNPF)	NPR-1 to -7, NPR-10, NPR-13	FLP-1, FLP-3, FLP-4, FLP-14, FLP-15, FLP-18, FLP-21, FLP-26, FLP-34
RFamide	FRPR-1 to -19, NPR-37, NPR-39, NPR-43 (+15 orphan GPCRs)	24 FLPs, NLP-17, NLP-23
Myosuppressin-like	DMSR-1 to -16, EGL-6	30 FLPs, NLP-13
Trissin-like	NPR-20, NPR-21	orphan
CNMamide-like	C24B5.1, F16C3.1, H09F14.1	orphan
Somatostatin/opioid/allatostatin C like (SST/Op/AstC-like)	NPR-17	NLP-24
	NPR-24	NLP-62
	NPR-28	RGBA-1
	NPR-32	NLP-64
	NPR-16, NPR-18, NPR-25, NPR-26, NPR-27, NPR-29, NPR-30, NPR-31	orphan
Gonadotropin-releasing hormone-like RPamide (GnRH-like RPamide)	GNRR-3, GNRR-6	NLP-2, NLP-22, NLP-23
	GNRR-2, GNRR-4, GNRR-5, GNRR-7, GNRR-8	orphan

^a Peptidergic systems conserved across Bilateria or Protostomia are named according to the receptor families in Jékely (2013), Mirabeau and Joly (2013), and Elphick et al. (2018).

^b Orthologous C. elegans receptors are divided into receptor families according to their clustering in previous phylogenetic studies (Mirabeau and Joly 2013; Beets et al. 2023).

^c Reviewed in Beets et al. (2023).

^d Receptors are also activated by other peptide ligands (reviewed in Beets et al. 2023).

Mapping receptor–ligand interactions

Predicting peptide–GPCR couples from phylogenetics remains challenging, because peptides typically have short sequences that coevolve with their receptors; therefore, orthologous peptides often share only low levels of sequence similarity (Jékely 2013; Mirabeau and Joly 2013). A frequently used strategy to identify receptor–ligand pairs is reverse pharmacology, in which a receptor is expressed in heterologous cells and tested against compound libraries of possible ligands (Szekeres 2002; Civelli et al. 2013; Foster et al. 2019). To date, ligands of 60 peptide-activated GPCRs have been identified in C. elegans using this approach (Yin et al. 2017; Lezi et al. 2018; Gershkovich et al. 2019; Sakai et al. 2021; Beets et al. 2023). Initial studies focused on only a few C. elegans peptides or peptide GPCR candidates, leaving many predicted peptide receptors orphan and precluding insights into putative crosstalk between ligands and receptors. A recent, large-scale deorphanization study expanded our understanding of C. elegans peptide–receptor interactions with the identification of 461 peptide–GPCR pairs out of over 55,000 possible interactions tested (Beets et al. 2023). Many interactions established in heterologous cells have been validated in vivo and play crucial roles in C. elegans behavior (Marques et al. 2021; Ramachandran et al. 2021; Reilly et al. 2021; Beets et al. 2023; Thapliyal et al. 2023; Marquina-Solis et al. 2024).

Although peptide GPCRs often interact with specific ligands derived from only one peptide-encoding gene, numerous peptide receptors have been found to have more complex ligand interactions (Foster et al. 2019; Abid et al. 2021). Systematic screening of peptide–GPCR pairs revealed 3 main types of signaling motifs in the C. elegans peptide network that show different levels of complexity: one-to-one, one-to-many, and many-to-many ligand–receptor interactions (Beets et al. 2023). Peptidergic crosstalk is particularly prominent for RFamide (FLP) peptides of which the C-terminus is characterized by an RF/RYamide sequence motif. The peptides FLP-1 and FLP-14, for example, each activate multiple closely related as well as phylogenetically more distant peptide GPCRs (Beets et al. 2023). Several C. elegans FLP receptors, such as DMSR-1, DMSR-7, FRPR-8, and EGL-6, also interact with peptides from diverse flp precursor genes (Beets et al. 2023). Crosstalk between FLP ligands and receptors may be linked to the expansion of the FLP neuropeptide family in nematodes (McVeigh et al. 2006; Li and Kim 2008; Mirabeau and Joly 2013; McCoy et al. 2014). However, ligand promiscuity and peptidergic crosstalk have been observed in other animals as well, for example, for the human neuropeptide FF receptors NPFFR1 and NPFFR2 (Kim et al. 2010; Oishi et al. 2011; Elhabazi et al. 2013; Foster et al. 2019), suggesting that this feature of neuropeptide signaling is conserved across diverse species. While the functional implications of these complex interaction patterns require further investigation, promiscuous peptide receptors may regulate different behaviors depending on the interacting ligand. One example is the neuropeptide receptor DMSR-1 that regulates sleep-like behaviors through interaction with FLP-13 and controls food and thermal state-dependent behaviors upon FLP-5 binding (Iannacone et al. 2017; Thapliyal et al. 2023). Conversely, 1 neuropeptide may broadcast onto multiple receptors to establish global states or exert pleiotropic roles through different GPCRs, such as FLP-18 which affects locomotion, feeding, and metabolism through multiple receptors (Kubiak et al. 2008; Cohen et al. 2009; Bhardwaj et al. 2018; Florman and Alkema 2022).

Structure of the C. elegans neuropeptidergic connectome

In C. elegans, like in other animals, neuropeptide and peptide GPCR genes are expressed broadly within the nervous system (Kim and Li 2004; Williams et al. 2017; Nässel and Zandawala 2019; Smith et al. 2019; Smith 2021; Taylor et al. 2021; Zhong et al. 2022; Thiel et al. 2024). Interestingly, the reconstruction of a transcriptomic atlas of the worm's nervous system revealed combinatorial expression codes of neuropeptide and peptide receptor genes for every neuron class (Taylor et al. 2021; Smith et al. 2024). Each of the 118 canonical neuron classes in C. elegans expresses distinct combinations of neuropeptide genes and peptide receptors (Taylor et al. 2021). Similarly, different neuron types were found to express distinct combinations of neuropeptide genes in other nervous systems, such as in the zebrafish, the mouse, and the annelid Platynereis dumerilii (Williams et al. 2017; Smith et al. 2019; Smith 2021; Anneser et al. 2024; Thiel et al. 2024), which suggests that a combinatorial code may be used by neurons to decipher the neurochemically complex signals received by many cells (Jékely 2021).

By integrating the expression of peptide and receptor genes with data on their biochemical interactions (Taylor et al. 2021; Beets et al. 2023), putative neuropeptide signaling pathways have been comprehensively mapped onto the neuroanatomy of C. elegans (Fig. 1a; Ripoll-Sánchez et al. 2023). This brain-wide neuropeptide signaling map, inferred from gene expression and receptor–ligand pairing, delineates an extensive wireless signaling network that has several features which differ from those of synaptic and monoamine networks (Fig. 1a; Ripoll-Sánchez et al. 2023). For example, the neuropeptidergic and synaptic connectomes only overlap for 5% of all peptidergic connections, indicating that neuropeptides establish many new, extrasynaptic signaling paths between neurons (Bentley et al. 2016; Ripoll-Sánchez et al. 2023). Extrasynaptic peptidergic signaling is supported by a broad range of behavioral and calcium imaging studies in C. elegans. One example is the neuropeptide FLP-20, secreted from the touch receptor neurons following mechanosensory stimulation (discussed in “Neuropeptidergic signaling cascades”). Release of FLP-20 induces arousal through its receptor FRPR-3 on the interneuron RID, which does not share synapses with the touch-sensing neurons (Chew, Tanizawa et al. 2018). Other peptide–GPCR pairs that signal extrasynaptically include NLP-18/CKR-1 in escape steering, LURY-1/NPR-22 in egg-laying, and CAPA-1/NMUR-1 in gustatory aversive learning, amongst others (Ohno et al. 2017; Watteyne et al. 2020; Chen et al. 2022). A recent study that investigated neural signal propagation in the C. elegans nervous system also identified many instances of DCV-dependent signaling between neurons that are not in direct synaptic contact but express matching neuropeptide ligand and receptor pairs, indicative of extrasynaptic signaling (Randi et al. 2023).

Fig. 1.

The neuropeptide network of the C. elegans nervous system forms a dense signaling network that structurally differs from synaptic and monoamine networks. a) Graphical and connection matrix representations of the signaling networks mediated by chemical synapses (left), monoamines (middle), and neuropeptides (right) in the C. elegans nervous system. Individual neurons in the graph representations are depicted as nodes that are connected by gray edges, indicating neuron-to-neuron interactions either through chemical synapses (left) or through neuromodulator to receptor pathways inferred from receptor–ligand pairing and gene expression data (middle and right). Neuronal nodes are colored based on their anatomical classification, and their size is defined by their relative degree (total sum of number of incoming and outgoing connections). The neuropeptide network is much denser and has a more decentralized topology, as there are many more high-degree neurons that are highly connected in comparison to the synaptic and monoamine networks. The monoamine network has only a limited number of sending neurons, while in the neuropeptide network nearly all neurons both send and receive peptide signals. Raw data from Varshney et al. (2011), Bentley et al. (2016), and Ripoll-Sánchez et al. (2023); visualized using Flourish Studios. b) t-SNE dimensionality reduction shows a mesoscale structure in the neuropeptide network, with neurons of the core network clustering into 3 groups that receive similar peptidergic input connections: One mainly containing sensory neurons, a second containing most of the peptidergic hub neurons, and a third containing many motor neurons. Figure adapted from Ripoll-Sánchez et al. (2023). c) Examples of 4 different topologies identified for individual peptide–receptor networks. Top left: Local network with only a few neurons expressing the neuropeptide capa-1 or its cognate receptor nmur-1. Top right: Pervasive network with many neurons expressing flp-18 and its receptor npr-5. Bottom left: Broadcasting network from a small number of neurons expressing the neuropeptide nlp-49 to many downstream putative partners expressing seb-3, a cognate receptor of NLP-49. Bottom right: Integrative network in which peptide signals from many neurons expressing flp-9 may converge onto a small number of neurons expressing the receptor egl-6. Figure adapted from Ripoll-Sánchez et al. (2023).

The C. elegans neuropeptide signaling network is much denser than the connectomes mediated by chemical synapses and monoamines (Fig. 1a); nevertheless, it has several structural features (Fig. 1b and c). One feature, like many networks, is the presence of a highly interconnected set of core neurons or so-called rich club (Ripoll-Sánchez et al. 2023). Rich club neurons are highly connected among themselves and with the rest of the network and have been assigned crucial roles in the coordination of information flow (Varshney et al. 2011; Towlson et al. 2013; Bentley et al. 2016; Yan et al. 2017; Sabrin et al. 2020; Moyle et al. 2021; Ripoll-Sánchez et al. 2023). In contrast to the synaptic connectome, where the rich club consists of only 11 neurons (Towlson et al. 2013), the peptidergic rich club includes 52% of all neurons (Ripoll-Sánchez et al. 2023). Overall, the peptide network has many more neurons that are highly connected (Fig. 1a), which suggests a more decentralized architecture (Ripoll-Sánchez et al. 2023). Analysis of peptidergic inputs classifies most of the rich club neurons in 3 clusters that primarily group sensory neurons, motor neurons, and the main hub neurons of the peptide network, which are mostly interneurons (Fig. 1b; Ripoll-Sánchez et al. 2023). This organization is hypothesized to help information flow between sensory and motor pathways. All neurons of the synaptic rich club, comprising 11 premotor interneurons (DVA, PVCL/R, AVAL/R, AVBL/R, AVDL/R, and AVEL/R), are also part of the neuropeptide core network (Towlson et al. 2013; Bentley et al. 2016; Ripoll-Sánchez et al. 2023), which implies that their role in driving global brain states is under extensive neuromodulatory control (Kato et al. 2015; Uzel et al. 2022; Atanas et al. 2023). In addition, the neuropeptide network has a number of specialized hub neurons (e.g. AVKL/R, PVQL/R, PVR, and PVT), which are among the network's most highly connected cells but have only a few synaptic contacts (Fig. 2a; Ripoll-Sánchez et al. 2023). These specialized peptidergic hubs are rich in DCVs (Fig. 2a), a hallmark of neurosecretory cells (Hartenstein 2006; van den Pol 2012), and some do not express any classical neurotransmitters or monoamines (e.g. AVKL/R and PVQL/R; Pereira et al. 2015; Wang et al. 2024). Seminal work examining serial-section micrographs of the C. elegans nervous system already described peptidergic hub neurons like PVT and PVR to be filled with “dark-cored vesicles” (White et al. 1986). A more recent report classified around 10% of neurons in the nerve ring neuropil as primarily harboring DCVs (Witvliet et al. 2021). Many of them are highly connected in the peptide network (Fig. 2a) and have long processes that project throughout the worm's body (Fig. 2b), which might enable them to modulate global states such as sleep and arousal (discussed in “Peptidergic circuit motifs in the regulation of behavior”).

Fig. 2.

Hub neurons in the C. elegans neuropeptide signaling network. a) Synaptic and neuropeptide degrees, defined as the respective number of synaptic and peptidergic connections that are incoming and outgoing in each neuron, are positively correlated (r = 0.54, p = 3.1e⁻¹⁴). Synaptic hub neurons with very high synaptic degrees are also highly connected hubs in the neuropeptide network. By contrast, some neurons seem to be specialized peptidergic hubs (e.g. AVKL/R, PVQL/R, PVR, and PVT), as they are highly connected in the neuropeptide network but have relatively few synaptic connections (figure adapted from Ripoll-Sánchez et al. 2023). Dots represent neurons and color codes indicate the relative composition of neuron vesicles (small clear vesicles and DCVs) as determined on electron micrographs of the head's neuropil (data from Witvliet et al. 2021). Neurons that are highly connected in the peptide network tend to have a high prevalence of DCVs. b) Many C. elegans neurons that display a relative excess of DCVs over small clear vesicles extend long processes throughout the body. Anatomical model adapted from the Virtual Worm Project (Sternberg Lab, OpenWorm Project) and visualized with Blender (Blender Foundation).

At the level of individual peptide–GPCR signaling networks, peptidergic systems can feature very different organizations depending on how broad the ligand and receptor are expressed (Fig. 1c; Ripoll-Sánchez et al. 2023). Most peptide–GPCR pairs are expressed in a relatively small number of neurons, forming local networks in the C. elegans nervous system (51/92 neuropeptide–GPCR couples assessed). Others display a broadcasting topology involving a small number of peptidergic neurons and many receptor-expressing cells (23 couples). Such broadcasting peptide–receptor pairs, like NLP-49/SEB-3, have been linked to arousal (Chew, Grundy et al. 2018) and may convey behavioral state information across the nervous system. A smaller number of peptide–GPCR networks show an integrative architecture (8 couples), which potentially integrate neuropeptide signals from many sources at a few receptor-expressing neurons. Finally, a number of pairs show broad expression for both the neuropeptide and its GPCR (10 couples) and form pervasive networks that could carry out both integrating and broadcasting functions.

Besides its high density and mesoscale organization, the mapping of the C. elegans neuropeptidergic connectome highlights other network features, such as putative peptide signaling cascades and autocrine feedback mechanisms (Ripoll-Sánchez et al. 2023). These structural features are not grossly influenced by assumptions of the network models (Ripoll-Sánchez et al. 2023). A dense network architecture likely represents a salient feature of neuropeptidergic connectomes, as most neurons have been found to coexpress specific combinations of neuropeptides and peptide GPCRs in various vertebrate and invertebrate models (Williams et al. 2017; Smith et al. 2019; Smith 2021; Anneser et al. 2024; Thiel et al. 2024). Given its high density, detailed mapping of peptide networks controlling behavior will be central to understanding the functional organization of neuropeptide connectomes. In addition, studies of specific circuits have already provided valuable insights into the circuit motifs and the spatiotemporal dynamics of neuropeptide signaling, which we describe in the next sections.

Peptidergic circuit motifs in the regulation of behavior

Decades of work on neuropeptide signaling have highlighted the important role of neuropeptides in the modulation of behavior. Detailed studies on local microcircuits in the C. elegans nervous system have also shed light onto the broad actions of specific neuropeptides and have started uncovering the functional relevance of interacting neuropeptide pathways. This has led to the emergence of peptide circuit motifs for combinatorial and feedback regulation that underpin behavioral control. Here, we describe a selection of examples from the many studies on neuropeptide functions in C. elegans behavior, which illustrate these signaling motifs [C. elegans neuropeptide functions have been exhaustively reviewed in Li and Kim (2014), Peymen et al. (2014), Bhat et al. (2021), Istiban et al. (2024), and Zhu and Chin-Sang (2024)]. As more neuropeptide systems within the peptidergic network become characterized, we expect additional examples of these motifs to serve important functional roles.

Divergent and convergent signaling

The broad expression of neuropeptides and their receptors, as well as their complex receptor–ligand interactions, suggests that peptidergic signals can integrate or diverge in specific neurons. Indeed, convergent (many-to-one) and divergent (one-to-many) signaling represent widely distributed signaling motifs that are typical of peptidergic systems (Jékely et al. 2018). One neuropeptide can interact with multiple receptors, which in turn can be expressed in different cells and can couple to distinct signaling transduction cascades. Such divergent signaling often underpins the pleiotropic functions mediated by many neuropeptides, as the physiological effect of the peptide will be dependent on its receptor target. A good example is the sNPF-like peptide FLP-18 that is involved in the control of locomotion, feeding, fat metabolism, and entry into the alternative dauer developmental stage (Cohen et al. 2009; Bhardwaj et al. 2018). FLP-18 mediates these diverse effects by acting on at least 3 receptors, which are expressed in different cells and characterized by distinct pharmacology (Fig. 3a and b; Kubiak et al. 2008; Cohen et al. 2009; Bhardwaj et al. 2018; Gershkovich et al. 2019). First, FLP-18 can steer multiple biological processes through activation of a single receptor, NPR-5, in ciliated sensory neurons that control fat metabolism (Fig. 3a), in the chemosensory ASJ neurons that are critical for dauer formation (Fig. 3a), and in the serotonergic ADF neurons that modulate feeding (Fig. 3b; Cohen et al. 2009; Lemieux et al. 2015). Second, FLP-18's modulatory effects are further expanded by it binding 2 additional receptors, NPR-1 and NPR-4. Activation of NPR-4 in the intestine regulates the accumulation of intestinal fat, along with NPR-5 signaling in sensory neurons (Fig. 3a; Cohen et al. 2009). FLP-18 also influences locomotion by controlling reversal frequency, reversal length, and turning behavior through NPR-1, NPR-4, and NPR-5 (Fig. 3b; Cohen et al. 2009; Bhardwaj et al. 2018, 2020; Florman and Alkema 2022). Third, FLP-18 receptors can have different cellular effects, as they activate distinct signaling cascades. In pharmacological assays, NPR-1 and NPR-4 were shown to inhibit cAMP signaling through activation of Gα_i/o proteins, whereas NPR-5 couples to Gα_s and Gα_q proteins that stimulate cAMP and calcium signaling, respectively (Kubiak et al. 2003, 2008; Gershkovich et al. 2019). Finally, while FLP-18's control of fat metabolism and dauer entry mainly require peptide secretion from the AIY interneuron that is postsynaptic to various sensory neurons (Fig. 3a), FLP-18's locomotory functions are largely dependent on release from other sources such as the interneuron AVA (Fig. 3b; Cohen et al. 2009; Bhardwaj et al. 2018; Florman and Alkema 2022). Altogether, this indicates that a neuropeptide can exert distinct functions depending on its receptor, its target circuit, and its cellular source.

Fig. 3.

Divergent and convergent signaling in neuropeptidergic circuits. a, b) The neuropeptide FLP-18 has pleiotropic functions, mediated by different receptors, target cells, and source neurons. a) FLP-18 released from AIY interneurons controls olfaction and foraging through its receptor NPR-4 in RIV and AVA neurons. It also regulates fat accumulation by acting on NPR-4 in the intestine and on NPR-5 in ciliated sensory neurons. FLP-18 signaling to NPR-5 in ASJ sensory neurons controls dauer formation. Figure adapted from Cohen et al. (2009). b) FLP-18 secreted from AVA and RIM interneurons regulates locomotion through the activation of multiple receptors on body wall muscles, on ASE sensory neurons, and on AVA. FLP-18 signaling from AVA also regulates feeding through the activation of NPR-5 in ADF sensory neurons. c) The ALA neuron plays a central role in the control of stress-induced sleep by the release of multiple quiescence-promoting neuropeptides such as FLP-13, FLP-24, NLP-8, and NLP-14. These neuropeptides induce distinct components of sleep-like behavior, such as locomotion quiescence, feeding quiescence, defecation quiescence, and a reduction of avoidance responses to aversive stimuli. d) The peptidergic hub neuron HSN controls a suite of temporally distinct egg-laying related behaviors by the release of multiple neuropeptides and monoamines. HSN receives diverse peptidergic inputs that modulate egg-laying according to physiological and environmental context, such as the BAG-secreted neuropeptides FLP-10 and FLP-17 that inhibit HSN in response to aversive cues, and tyramine and neuropeptide signals that feed back to inhibit egg-laying upon mechanical activation of the uv1 neuroendocrine cells during egg release. Serotonin (5-HT) and the neuropeptide NLP-3 coordinately initiate egg-laying upon local release onto the egg-laying machinery. Axonal release of the neuropeptides FLP-2, FLP-26, and FLP-28 is required for locomotory changes associated with egg-laying.

Divergent signaling is also a mechanism for coordinating multiple neuropeptide actions within the modulation of global or behavioral states. In C. elegans males, for example, the neuropeptide FLP-3 drives sex-specific attraction to a hermaphrodite-secreted pheromone, the ascaroside ascr#8, by nonredundantly signaling through 2 distinct receptors, NPR-10 and FRPR-16 (Reilly et al. 2021). Since reporter studies indicate no overlap between npr-10 and frpr-16 expressing cells, FLP-3 signaling is thought to impinge onto multiple targets to promote pheromone attraction in males. Furthermore, behavioral states can be regulated by multiple neuropeptides that are coexpressed in the same neuron, a property that has been ascribed to several peptidergic hub neurons, such as the sleep-promoting neuron ALA. Multiple neuropeptides emanating from ALA act in combination to enable a protective stress-induced sleep state in response to cellular stresses like tissue damage or heat (Fig. 3c; Hill et al. 2014; Nelson et al. 2014; Nath et al. 2016). One neuropeptide involved is FLP-13 (Nelson et al. 2014), which acts at least in part through its receptors FRPR-4 and DMSR-1 on DVA and AIY interneurons, respectively, to regulate feeding quiescence and locomotory quiescence (Nelson et al. 2015; Iannacone et al. 2017). In addition, ALA releases other neuropeptides, like NLP-8, NLP-14, and FLP-24, that induce distinct components of sleep-like behavior upon heat stress, including locomotion quiescence, feeding quiescence, and defecation quiescence (Nath et al. 2016; Honer et al. 2020; Le et al. 2023). Strong global sleep defects are only observed upon the loss of neuropeptide combinations, indicating that these molecules act collectively to control stress-induced sleep. Conversely, overexpression of each of these neuropeptides individually induces unique but overlapping suites of sleep-associated behaviors, with a more overarching role for FLP-13 conform to earlier findings (Nelson et al. 2014). While the targets and downstream pathways through which these neuropeptides control sleep are largely unknown, FLP-13 and FLP-24 do so in part by reducing cAMP/PKA signaling in wake-promoting DVA and RIF interneurons, respectively (Cianciulli et al. 2019).

Divergent neuropeptide transmission is paralleled by convergent signaling, in which multiple neuropeptides target 1 receptor or the same neuron. The hermaphrodite-specific neuron (HSN) is 1 example of a hub neuron where multiple neuropeptides converge to modulate egg-laying behavior in response to diverse physiological and environmental cues (Fig. 3d). Multiple aversive stimuli, such as high CO₂ levels and hyperosmolarity, inhibit HSN activity and egg-laying through FLP-10 and FLP-17 neuropeptide signaling from BAG sensory neurons, acting on their receptor EGL-6 in HSN neurons (Ringstad and Horvitz 2008; Hallem et al. 2011; Huang et al. 2023). Food absence also modulates the temporal pattern of alternating active (∼2 min) and inactive (∼20 min) egg-laying phases and prolongs inactive periods, an effect that requires FLP-1 signaling partially onto HSN (Waggoner et al. 2000; Chang et al. 2015). In addition, HSN synaptic output is shaped by neuropeptidergic feedback from the uv1 (uterine-vulval type 1) cells that sit above the vulval canal and release tyramine along with NLP-7 and FLP-11 neuropeptides to inhibit egg-laying (Collins et al. 2016; Banerjee et al. 2017). Altogether, convergent neuropeptide signaling thus allows fine-tuned regulation of egg-laying behavior in response to the sensory environment. Besides this integrative property, HSN releases multiple neuromodulators to coordinate egg-laying and related locomotory behaviors through divergent signaling (Fig. 3d). These include serotonin and the neuropeptide NLP-3, which induce activity of the ventral type C (VC) motor neurons and contraction of the egg-laying muscles, leading to egg expulsion from the uterus (Trent et al. 1983; Shyn et al. 2003; Zhang et al. 2008; Collins and Koelle 2013; Brewer et al. 2019). HSN also releases other neuropeptides, like FLP-2, FLP-26, and FLP-28, that acutely increase locomotion speed prior to egg-laying (Hardaker et al. 2001; Huang et al. 2023).

Antagonistic and synergistic pathways

Each neuron class in C. elegans expresses a distinct combination of neuropeptides and peptide receptors (Taylor et al. 2021), which suggests that multiple neuropeptide pathways functionally interact in circuit modulation. Indeed, many neural circuits like those for gustatory plasticity and sleep-like behaviors are regulated by multiple neuropeptides (Beets et al. 2012; Hill et al. 2014; Nelson et al. 2014; Nath et al. 2016; Peymen et al. 2019; Van der Auwera et al. 2020; Watteyne et al. 2020). How circuits maintain their stability when challenged with extensive neuromodulation is a long-standing question (Marder 2012). One mechanism that has been suggested to prevent overmodulation or stabilize alternative circuit states is the engagement of neuromodulators with antagonistic actions (Harris-Warrick and Johnson 2010; Bargmann 2012; Marder 2012). This can be achieved through the modulation of opposing mechanisms by the same neuropeptide or through actions of multiple antagonistic neuromodulators (Harris-Warrick and Johnson 2010; Chai et al. 2022).

Antagonistic signaling is a prominent feature of peptidergic circuits in C. elegans. A genetic screen for peptide GPCRs affecting C. elegans entry into the diapause stage, for example, found many neuropeptide receptors with opposing effects on diapause entry to be coexpressed in the same neurons, suggesting that this developmental decision is tightly controlled by antagonistic peptidergic pathways (Chai et al. 2022). Transitions between behavioral states are also controlled by antagonistic neuromodulators, the effects of which can be anatomically distributed (Bargmann 2012). A good example is the modulation of locomotory states in response to food availability or changes in ambient oxygen levels, through the balanced actions of FLP-1 and NLP-12 peptides (Fig. 4a). FLP-1 is the most abundantly expressed neuropeptide in the peptidergic hub neuron AVK, whereas NLP-12 is highly expressed in another hub neuron, DVA (Taylor et al. 2021). In the absence of food, FLP-1 signaling from AVK promotes fast long-distance travel or dispersal by inhibiting reorientations and suppressing deep body bending (Hums et al. 2016; Oranth et al. 2018). FLP-1 promotes dispersal and shallow whole-body waves through activation of its receptors FRPR-7 and NPR-6, the latter on motor neurons in the head (SMB) and in the ventral nerve cord (Oranth et al. 2018; Ji et al. 2023). The action of FLP-1 is functionally opposed by the cholecystokinin ortholog NLP-12, released from the synaptic and peptidergic hub neuron DVA (Table 1 and Fig. 4a). NLP-12 signaling promotes exploration of the local environment or local searching, a behavioral strategy characterized by frequent reorientations (Bhattacharya et al. 2014; Hums et al. 2016). The NLP-12 peptide signals through its receptor CKR-1 on SMD head motor neurons to potentiate reorientations (Ramachandran et al. 2021). FLP-1 and NLP-12 thus have opposing modulatory effects on dispersal and local search through a distributed network of motor neurons. Both peptides also have opposing actions during basal locomotion; for instance, the curvature of body bending is increased in flp-1 mutants, while it is decreased in nlp-12 loss-of-function animals (Nelson, Rosoff et al. 1998; Hu et al. 2011; Ramachandran et al. 2021; Chen et al. 2024). Interestingly, FLP-1 antagonizes other neuromodulators, by inhibiting their biosynthesis or release, in different contexts. During mechanosensory-induced locomotory arousal, FLP-1 counteracts the functions of the neuropeptide NLP-10 that increases locomotion speed trough activation of its receptor NPR-35 in AVB and AIY premotor interneurons (Fig. 4b; Aoki et al. 2023). FLP-1 decreases acceleration by signaling back onto its receptor DMSR-7 in AVK, inhibiting the release of the forward-accelerating peptide NLP-10 from the AVK hub neuron (Aoki et al. 2023). On a longer timescale, FLP-1 antagonizes tyraminergic and octopaminergic signaling in response to infection, to promote pathogen avoidance behavior (Pu et al. 2023; Marquina-Solis et al. 2024). Prolonged exposure of several hours to pathogenic bacteria increases FLP-1 expression and release from AVK, which in turn acts on multiple receptors and cells to drive pathogen avoidance (Fig. 4c; Marquina-Solis et al. 2024). These examples illustrate that neuropeptide release from a single neuron can regulate distinct behaviors by interacting with different antagonistic neuromodulators in a context-dependent manner.

Fig. 4.

Antagonistic and synergistic interactions between neuromodulatory pathways. a) AVK and DVA hub neurons release antagonistic neuropeptides to regulate transitions between dispersal and local searching when C. elegans is removed from food. AVK-derived FLP-1 and release of NLP-12 from DVA antagonistically regulate dispersal and local search through a distributed network of receptor-expressing cells. b, c) FLP-1 signaling from AVK antagonizes the actions of other neuromodulators to regulate locomotion speed and pathogen avoidance behavior. b) Autocrine signaling of FLP-1 in AVK neurons reduces locomotion speed by inhibiting the release of the forward-accelerating neuropeptide NLP-10, which increases speed via its receptor NPR-35 on AIY and AVB premotor interneurons. c) Prolonged exposure to pathogenic bacteria increases FLP-1 release from AVK to promote pathogen avoidance. FLP-1 signaling evokes avoidance by antagonizing tyraminergic/octopaminergic signaling from RIM and RIC neurons, through activation of its receptor DMSR-7 in these cells. FLP-1 mediates pathogen avoidance also by acting on other, yet unidentified neurons, through NPR-6 and DMSR-7. d) Locomotion arousal from the quiescence period associated with molts (developmentally timed sleep) is promoted by a synergistic interaction between PDF-1 and FLP-2 neuropeptides. Enhanced secretion of PDF-1 and FLP-2 arouses locomotion and is regulated by reciprocal positive feedback between the PDF-1 and FLP-2 signaling pathways. This may occur in the ASI sensory neurons, which express both peptides along with the FLP-2 receptor FRPR-18, suggesting that FLP-2 secretion is additionally regulated by autocrine feedback. PDF-1 release from sensory neurons, including ASK, mediates locomotory arousal by acting on its receptor PDFR-1 in touch sensory neurons and body wall muscles.

Positive feedback between neuropeptides can be another mechanism for stabilizing a specific behavioral or circuit state, as has been shown for FLP-2 and PDF-1 peptides that promote exit from locomotion quiescence during larval molts (Fig. 4d). The conserved pigment dispersing factor PDF-1 acts in concert with FLP-2 to arouse locomotion and enhances the sensitivity of peripheral mechanosensory receptors by acting on its receptor PDFR-1 in touch-sensing neurons, in addition to activating PDFR-1 on body wall muscles (Choi et al. 2013, 2015; Chen, Taylor et al. 2016). During larval molts, secretion of both PDF-1 and FLP-2 is diminished, while enhanced secretion is linked to aroused locomotion (Choi et al. 2013; Chen, Taylor et al. 2016). PDF-1 and FLP-2 secretion is regulated by reciprocal positive feedback, which has been proposed to stabilize the aroused state (Chen, Taylor et al. 2016). Both neuropeptide pathways are thought to interact in the sensory ASI neurons that control arousal, as they coexpress FLP-2 and PDF-1 peptides (Kim and Li 2004; Janssen et al. 2009; Barrios et al. 2012; Chen, Taylor et al. 2016). These neurons also express the FLP-2 receptor FRPR-18, suggesting that FLP-2 secretion is additionally regulated through autocrine feedback in ASI (Chen, Taylor et al. 2016). Other neurons expressing flp-2 or pdf-1, some of which have been linked to locomotory arousal (e.g. RID and ASK), may be involved in reciprocal feedback as well (Kim and Li 2004; Choi et al. 2013; Lim et al. 2016; Chew, Tanizawa et al. 2018). Autocrine and reciprocal positive feedback are functional motifs that are also conserved in mammalian arousal circuits (Yamanaka et al. 2010; Brown et al. 2012).

Neuropeptidergic signaling cascades

Most neurons are sources and targets of peptidergic modulation, as they express both neuropeptides and peptide-activated receptors. Peptidergic signaling cascades, in which 1 peptide controls the release of another peptide, is a characteristic feature of hormonal signaling axes (Hiller-Sturmhöfel and Bartke 1998; Nässel and Zandawala 2020) and is commonly observed for neuropeptide pathways in both vertebrates and invertebrates (Davis et al. 2007; Taghert and Nitabach 2012; Martelli et al. 2017; Lee et al. 2020). The mapping of the C. elegans neuropeptidergic connectome identified many putative peptidergic signaling cascades, which could potentially interconnect all neurons within the nervous system (Ripoll-Sánchez et al. 2023). Some of these have documented roles in the modulation of behavior, such as the feedforward cascade that controls mechanosensory-induced arousal through FLP-20 signaling (Fig. 5a; Chew, Tanizawa et al. 2018). Mechanosensory stimulation (tapping) evokes a transient minute long period of increased locomotor activity and cross-modal sensitization of the major nociceptors, the ASH neurons, in C. elegans (Chew, Tanizawa et al. 2018). Both effects require the release of FLP-20 neuropeptides from the primary mechanosensory neurons, which signal to their receptor FRPR-3 on the peptidergic hub neuron RID, among other cells. Neuropeptidergic signaling via yet unidentified RID peptides in turn modulates ASH neurons and downstream motor circuits, establishing cross-modal sensitization and locomotory arousal (Chew, Tanizawa et al. 2018). RID thus functions as a central hub in a peptidergic cascade that links sensory-evoked afferent peptidergic signals to efferent neuropeptide signals, which broadly convey behavioral state information. Neuropeptides from AVK are also required for locomotory arousal, as overexpression of nlp-49 in AVK leads to heightened locomotor responses, an effect that is dependent on its receptor SEB-3 (Chew, Grundy et al. 2018). In which cells NLP-49 and SEB-3 are required for this effect remains yet unknown.

Fig. 5.

Neuropeptidergic signaling cascades and autocrine feedback. a) Aversive mechanical stimuli elicit a state of arousal, characterized by locomotory arousal and cross-modal sensitization of the nociceptive ASH neurons. Both effects are mediated by a peptidergic signaling cascade that requires FLP-20 neuropeptide signaling from the touch receptor neurons to its receptor FRPR-3 on the RID interneuron, which in turn evokes arousal through the release of yet unknown neuropeptides. Mechanosensory-induced locomotory arousal is also stimulated by FRPR-3 on the AIY interneurons and by neuropeptides from other cells, such as NLP-49 released from AVK and acting on its receptor SEB-3. b) The response of AWC olfactory neurons upon odor removal is dampened by a peptidergic cascade. NLP-1 released from AWC binds to its receptor NPR-11 in AIA interneurons, triggering the release of the insulin-like peptide INS-1. The insulin signal acts back onto AWC through a yet unknown receptor to dampen AWC olfactory responses. c) Autocrine peptidergic feedback in the RIM interneurons is important for threat-reward decision making between aversive stimuli (high osmolarity) and attractive cues (food odors). Both autocrine PDF-2 signaling and tyramine release from RIM neurons decrease threat tolerance in well-fed C. elegans, while food deprivation increases threat tolerance through suppression of these neuromodulatory pathways. Figure adapted from Li et al. (2016). d) An activity-dependent autocrine insulin signal, INS-1, represses the expression of the BAG neuron fate and its function in CO₂ avoidance, by acting on its receptor DAF-2 and inhibiting expression of the flp-17 neuropeptide gene in BAG.

Besides feedforward signaling, neuropeptidergic cascades are often used as feedback mechanism in which peptide release from 1 neuron actuates its target to release a neuropeptide back onto the initial sender. One example is the peptidergic sensorimotor feedback loop that dampens the odor-evoked activity of the olfactory AWC neuron during foraging and odor adaptation (Chalasani et al. 2010). In response to odor cues, AWC releases the buccalin-related NLP-1 neuropeptide, which acts on its receptor NPR-11 on the interneuron AIA to regulate the secretion of the insulin-like peptide INS-1 (Fig. 5b). INS-1 subsequently closes the feedback loop by modulating AWC's responsiveness to sensory stimuli; the receptor involved in this feedback mechanism remains unknown. Of note is that this feedback loop acts on several timescales. On one hand, it acutely tunes AWC presynaptic and AIA postsynaptic activities in response to altering food odors on a timescale of seconds. On the other hand, it participates in olfactory adaptation by blunting AWC activity upon prolonged odorant exposure ranging from several minutes up to 1 h (Chalasani et al. 2010). Altogether, the organization of neuropeptide pathways into signaling cascades enables distinct circuit components to be tightly coordinated through both coherent feedforward as well as elaborate feedback mechanisms.

Autocrine neuropeptide pathways

The mapping of the C. elegans neuropeptide signaling network revealed a high prevalence of putative autocrine signaling pathways. Approximately 58% of C. elegans neurons coexpress neuropeptide receptors with their cognate ligand, which can feed back onto the cell releasing it (Ripoll-Sánchez et al. 2023). Single-cell transcriptome studies suggest that autocrine peptide signaling is also ubiquitous in vertebrate nervous systems, such as in the mouse hypothalamus and cortex (Chen et al. 2017; Smith 2021), where it supports important aspects of neuronal function including feedback modulation of neuron activity (Choi et al. 2012; Mardinly et al. 2016; Henquin 2021). Autocrine signaling peptides mediate similar functions in C. elegans, as is illustrated by the earlier described examples of FLP-18/NPR-4 and FLP-1/DMSR-7 that regulate neuronal activity and peptide release in AVA and AVK neurons, respectively (Figs. 3b and 4b). The reaction to both aversive and attractive stimuli in threat-reward decision making is also modulated by the autocrine action of the pigment dispersing factor peptide PDF-2 on the RIM interneurons (Ghosh et al. 2016; Fig. 5c). Loss-of-function mutants lacking PDF-2 or its receptor PDFR-1 display increased threat tolerance, as they will frequently cross a high osmolarity barrier to reach a spot of attractive food odors. Ghosh et al. (2016) showed that autocrine PDF-2 signaling together with tyramine release from the RIM interneurons decrease threat tolerance in well-fed C. elegans. In this multisensory threat-reward decision paradigm, autocrine PDF-2 signaling and tyraminergic feedback are proposed to set RIM excitability to bias the animal toward threat-sensitive decisions. Interestingly, neuromodulatory feedback onto RIM can be inhibited by changes in the animal's internal state, such as food deprivation, which increases threat tolerance (Ghosh et al. 2016; Li et al. 2016).

Autocrine peptidergic signaling also has important roles during nervous system development. For example, an activity-dependent autocrine insulin signal regulates the development of the chemosensory BAG neurons in C. elegans (Horowitz et al. 2019). This insulin signal antagonizes the expression of a neuropeptide gene, flp-17, that defines the functional identity of BAG neurons and is essential for their role in CO₂ avoidance behavior (Fig. 5d; Guillermin et al. 2017; Horowitz et al. 2019). The developmentally active p38 MAP kinase PMK-3 represses the autocrine insulin signal, which entails this autocrine loop to integrate both activity-dependent and developmentally timed gene programs. Similar developmental functions have been attributed to autocrine peptides in embryogenesis and axon development in vertebrates (Renner et al. 1996; Terauchi et al. 2016).

The number of putative autocrine peptide connections of a neuron strongly correlates with the total number of peptidergic connections that a neuron harbors in the neuropeptide signaling network (Ripoll-Sánchez et al. 2023). This suggests that autocrine signaling may play an important role in regulating the activity of peptidergic hub neurons. Autocrine peptides also serve to coordinate activity across neighboring neurons that share these molecules (Yoshioka et al. 2005; Tse and Wong 2019). This might apply to the excitatory motor neurons of the ventral nerve cord, as they have a number of overlapping autocrine peptidergic pathways (Ripoll-Sánchez et al. 2023). Overall, autocrine neuropeptide feedback may serve broadly in the C. elegans nervous system to tune neuronal activity and peptidergic transmission.

Spatiotemporal scale and plasticity of neuropeptide signaling

In vivo studies underscore that the mapping of peptidergic circuits requires detailed knowledge of when and where neuropeptides are released, as most peptides are expressed in multiple neurons and their release can be shaped by experience and context. Localizing specific sites of peptide secretion may also be important for understanding modulatory effects, as neuropeptides can be released from different neuronal compartments. For example, imaging of fluorescently tagged C. elegans neuropeptides hints at peptide release from both axons and dendrites (Jang et al. 2019; Tao et al. 2019). This is in line with work in other model organisms, which shows that neuropeptides can be secreted from axons, somata, and dendrites of neurons, through independently regulated mechanisms (Ludwig and Leng 2006; Ding et al. 2019; Klose et al. 2021; Qian et al. 2023). A study on the peptidergic hub neuron HSN suggests that different axonal compartments can release distinct neuromodulators in C. elegans (Huang et al. 2023). HSN releases serotonin and the neuropeptide NLP-3 to induce egg-laying (Trent et al. 1983; Shyn et al. 2003; Zhang et al. 2008; Collins and Koelle 2013; Brewer et al. 2019) but also secretes other neuropeptides, such as FLP-2, FLP-26, and FLP-28, that acutely increase locomotion speed before egg-laying (Fig. 3d; Hardaker et al. 2001; Huang et al. 2023). The latter are likely released from HSN's axonal projection to the nerve ring and not from the axon part in the vulval presynaptic region, because axotomy at a point separating these 2 parts attenuates the increased locomotion but keeps egg-laying intact (Huang et al. 2023). While neuropeptide release sites have not yet been investigated in most C. elegans neurons, these studies suggest that compartmentalized release may underpin different modulatory effects.

Neuropeptides can have short-range effects by acting on postsynaptic partners, they can signal extrasynaptically by diffusing over larger distances in the nervous system, and they can signal across different anatomical compartments or tissues. A notable example is the hypoxia-induced FLP-21/NPR-1 peptidergic pathway between neurons in the pharynx and the central neuropil (Pocock and Hobert 2010), the former of which is largely considered as a self-contained autonomously acting unit (Cook et al. 2020). In addition, peptide messengers can mediate potential long-range signaling between the nervous system and other tissues like the gut (e.g. NLP-40, FLP-7, and INS-11) and the body wall muscles (e.g. FLP-18), and can be secreted into the body cavity fluid, or pseudocoelom (Wang et al. 2013; Palamiuc et al. 2017; Florman and Alkema 2022; Shi et al. 2022). The mechanisms that control the spatial scope of neuropeptide signaling, such as the regulation of neuropeptide diffusion by peptidases and the location of neuropeptide receptors, are, however, less well understood.

Temporally, neuropeptides have been shown to modulate behaviors over broad timescales, ranging from several seconds to minutes or hours (Flavell et al. 2013; Chew, Tanizawa et al. 2018; Marquina-Solis et al. 2024). They typically elicit relatively slow but profound cellular changes through intracellular signaling cascades downstream of their cognate GPCR (Nadim and Bucher 2014). Antagonism between the neuropeptide PDF-1 and serotonin, for example, stabilizes opposing roaming and dwelling states on a timescale of minutes (Flavell et al. 2013; Ji et al. 2021). Other neuropeptides like FLP-1 can regulate behaviors at longer timescales. Besides its effects on locomotion (Fig. 4a and b), FLP-1 signaling promotes pathogen avoidance over the course of several hours upon prolonged exposure to pathogenic bacteria (Fig. 4c; Hums et al. 2016; Oranth et al. 2018; Pu et al. 2023; Marquina-Solis et al. 2024). Although peptidergic signaling is typically considered as a relatively slow mode of transmission, some neuropeptides can mediate rapid effects in the range of seconds. Such signaling properties have been attributed to NLP-40 in the control of the defecation motor program. Following release from the intestine, NLP-40 activates the motor neuron DVB within seconds, which in turn controls enteric muscle contraction (Wang et al. 2013). Furthermore, measurements of functional connectivity in the C. elegans nervous system suggest that extrasynaptic peptidergic signaling broadly impacts neural dynamics on short timescales (Randi et al. 2023). By combining cell-specific optogenetic activation with whole-brain imaging, this study identified many second-scale functional relationships between neurons that were dependent on DCV release. Most of these fast, DCV-dependent connections were between neurons that do not share synapses but express matching neuropeptide and receptor pairs (Randi et al. 2023). This is the case for example for the pharyngeal neuron M3L that may communicate with the sensory neuron URYVL through second-scale peptide transmission, as these neurons do not share synapses, but URYVL responds rapidly to optogenetic activation of M3L in a DCV-dependent manner (Randi et al. 2023).

Temporal dynamics of neuropeptide action is further diversified by context- and experience-dependent plasticity of peptidergic signaling. Some neuropeptides like NLP-45 show developmentally regulated expression and are required for distinct temporal transitions in exploratory behavior (Sun and Hobert 2021). The expression of others is shaped by experience, including FLP-21 that mediates the hypoxia-induced enhancement of gustatory perception (Pocock and Hobert 2010), FLP-1 that promotes pathogen avoidance behavior upon infection (Marquina-Solis et al. 2024), and FLP-5 and FLP-6 that context-specifically regulate behavioral states (Chen, Chen et al. 2016; Thapliyal et al. 2023). The regulation of neuropeptide expression likely plays a general role in the long-term plasticity of circuit function. In addition, neuropeptide release is often regulated in a context-dependent manner. Indeed, many neuropeptides are only required under specific conditions, as is for instance the case for those released by ALA to induce sleep specifically upon cellular stress (Fig. 3b). Similarly, the Neuromedin U-like neuropeptide CAPA-1 is specifically required during the performance of learned salt aversion and not during the acquisition of the learned response (Watteyne et al. 2020). While there are still many questions considering the spatiotemporal regulation of neuropeptide signaling, recently developed tools for visualizing peptide release and receptor interaction could improve our understanding of the spatial scope and timescales at which neuropeptides signal throughout the nervous system (Klose et al. 2021; Duffet et al. 2022; Wang et al. 2023; Kim et al. 2024).

Conclusion

The molecular diversity and widespread expression of neuropeptide signaling molecules in nervous systems are nothing short of astonishing. The comprehensive mapping of these signaling pathways in C. elegans has revealed an extensive extrasynaptic signaling network with structural features that differ in several ways from those of synaptic and monoamine networks (Taylor et al. 2021; Beets et al. 2023; Ripoll-Sánchez et al. 2023). In vivo studies have pointed out the crucial role of neuropeptide signaling in generating flexible behaviors across broad timescales, while system-wide analysis of peptide–receptor interactions and functional connectivity in the C. elegans nervous system have been important to improve our understanding of neuropeptide actions and have highlighted the wide impact of extrasynaptic signaling on neural dynamics (Beets et al. 2023; Randi et al. 2023). One central theme that has emerged from all these studies is the combinatorial nature of peptidergic neuromodulation, for example, through convergent, antagonistic, feedforward, and feedback signaling cascades. Neuropeptide release from 1 neuron can modulate different behaviors through distinct receptors, target cells, or interacting pathways in different contexts. Combinatorial peptide signaling can also coordinate the modulation of multiple circuits involved in global behavioral states, such as sleep and arousal. Likewise, parallel work in other invertebrate and vertebrate nervous systems has uncovered broad coexpression of neuropeptides and peptide receptors, suggesting widespread interactions between neuropeptide systems (Williams et al. 2017; Smith et al. 2019; Smith 2021; Zhong et al. 2022). While the exact effects of these coexpression fingerprints on brain function are yet to be understood, combinatorial peptide signaling may help to address signals from 1 neuron to another within pervasive signaling networks that stretch over most of the nervous system (Jékely 2021).

Besides interactions between neuromodulators, the context-dependent regulation of neuropeptide expression and release shapes the organization of the neuropeptide network. A growing body of work, with recent examples discussed in this review, illustrates how the timing and location of neuropeptide release are influenced by context, such as internal state, current and previous experience, and life stage. Peptidergic signaling cascades and autocrine feedback play a prominent role in this context-dependent regulation. These features are also characteristic of hormonal and neuroendocrine signaling axes found, for example, in Drosophila and in vertebrates (Hiller-Sturmhöfel and Bartke 1998; Nässel and Zandawala 2020).

The various resources for synaptic connectivity, whole-brain neural activity, molecular expression, and receptor–ligand interactions in the C. elegans nervous system provide a powerful framework for studying how peptide signaling networks control behavior and physiology (White et al. 1986; Jarrell et al. 2012; Cook et al. 2019; Taylor et al. 2021; Witvliet et al. 2021; Atanas et al. 2023; Beets et al. 2023; Randi et al. 2023). Subsequent work at the molecular, circuit, and network levels is required to improve our understanding of how these networks function. At the molecular level, filling in gaps on the ligand interactions and intracellular signaling pathways of neuropeptide receptors will further complete the peptide network and help understand the cellular effects of neuropeptide transmission. Likewise, a more detailed understanding of the subcellular localization of neuropeptide receptors, and the regulation of neuropeptide release and diffusion, will likely be important for decoding the network's function, considering that both neuronal and nonneuronal targets can contain distinct physiological compartments (Schroeder and McGhee 1998; Hendricks et al. 2012; Steuer Costa et al. 2019). Whole-brain calcium imaging and detailed behavioral analysis will also be crucial to understand the impact of neuropeptide signaling on network activity and behavior. Recently generated GPCR and neuropeptide mutant libraries can be valuable resources to expedite these studies and uncover possible functional redundancies (Pu et al. 2023). In addition, genetically encoded sensors for neuropeptide release and receptor interaction can help address several of these questions by providing tools to visualize neuropeptide signaling in vivo (Kim et al. 2017; Ding et al. 2019; Klose et al. 2021; Qian et al. 2023; Wang et al. 2023).

Some neurons within the C. elegans neuropeptide network distinguish themselves as specialized neuropeptidergic cells based on anatomical and network properties. While several have well-documented roles in conveying behavioral state information using neuropeptides, future work on other, lesser-studied, hub neurons within this network might uncover similar central roles in influencing neural circuit activity and behavior. Finally, incorporating nonneuronal cells into the peptide signaling network will be important to understand its impact at organismal scale. For example, neuropeptides also mediate neuro-intestinal signaling to regulate intestinal function, metabolism, and immunity, as well as neuronal signaling and behavior (Cohen et al. 2009; Lee and Mylonakis 2017; Palamiuc et al. 2017; De Rosa et al. 2019; Singh and Aballay 2019; Mutlu et al. 2020). Future work integrating these pathways into the peptide network might highlight neurons and nonneuronal cells playing prominent roles in intertissue peptidergic communication.

Data availability

No new data were generated or analysed in this article.

Funding

We acknowledge grant support from the European Research Council (ERC 950328), the Research Foundation Flanders (FWO G0B5322N, G079521N, and 12AKR24N), the Medical Research Council (MRC MC-A023-5PB91), and the Baillet Latour Fund.