Abstract
Introduction
Neuropeptides play an important modulatory role in the peripheral and central nervous systems. Many of these are expressed in migraine-relevant locations and have important roles in nociception and neurogenic inflammation, with calcitonin gene-related peptide (CGRP) being a prominent example (1,2). The spotlight has recently shone on the pituitary adenylate cyclase-activating polypeptide (PACAP) and vasoactive intestinal peptide (VIP) peptide family because of accumulating pre-clinical and clinical data linking this family to migraine (3–8). Notably, an anti-PACAP monoclonal antibody was recently reported to reduce migraine frequency in a phase 2 clinical trial (9). Similar to other neuropeptides, PACAP and VIP activate G protein-coupled receptors (GPCRs) with the current classification officially recognizing the PAC1, VPAC1 and VPAC2 receptors (Figure 1) (10,11). Typically, PACAP is more potent than VIP at the PAC1 receptor, and equipotent to VIP at the VPAC receptors (12,13). However, there are many N-terminal or intracellular loop 3 (ICL3) splice variants of the PAC1 receptor that may differ in relative peptide binding affinities or intracellular signaling profiles (Figure 1). Other GPCRs or accessory proteins have also been proposed to contribute to PACAP signaling. Although PACAP and VIP were discovered decades ago, very little is known about how they function at a molecular level, especially in headache, migraine and pain. This narrative review describes this complex system of multiple peptides, several canonical and non-canonical proposed receptors, and downstream signaling pathways to highlight potential therapeutic avenues targeting the PACAP/VIP family for the treatment of headache disorders and beyond.

Pituitary adenylate cyclase-activating polypeptide (PACAP)-responsive receptors and their structure, functional domains and examples of PACAP-responsive receptor splice variants. (A) The canonical PACAP-responsive receptors, which may interact with single transmembrane receptor-activity modifying proteins (RAMP) to alter their function. The different receptor colors represent the different receptor subtypes that belong to a related subfamily of class B G protein-coupled receptors. (B) Proposed/non-canonical human receptors. (C) The overall features of canonical PACAP-responsive receptors represented in the PAC1 receptor, together with the cryo electron microscopy structure of the active PAC1 receptor bound by PACAP-27 (pdb: 8E3X). (D) A selection of human PAC1 receptor splice variants. Created in BioRender. Hay, D. (2025) https://BioRender.com/voxub7e.
PACAP and related peptides
The PACAP/VIP peptide family comprises two biologically active forms of PACAP, VIP and also peptide histidine methionine (PHM) in humans or peptide histidine isoleucine (PHI) in rodents (14–17). The amino acid sequences of PACAP and VIP are highly conserved between species, with complete sequence identity between human, rat, mouse and ovine peptides, suggesting they serve conserved and important biological functions (14). The two forms of PACAP, discovered by Professor Arimura and his team (16,17), are PACAP-38 and PACAP-27. The full-length 38 amino acid peptide sequence, PACAP-38, contains a dibasic cleavage site that allows post-translational processing to form the shorter, C-terminally truncated 27 amino acid peptide, PACAP-27. PACAP-38 and PACAP-27 are generated from the same
VIP and PHM/I are generated from different exons within
Canonical PACAP-responsive receptors and their splice variants
Molecular composition and pharmacology
The biological actions of the PACAP peptide family are largely assumed to occur through their binding to and activation of three class B GPCRs. These are the PACAP 1 receptor (PAC1), VIP receptor 1 (VPAC1) and VIP receptor 2 (VPAC2) (Figure 1) (10,11). These receptors have distinct pharmacological profiles. At PAC1 receptors, originally known as “PACAP type I receptors”, both PACAP-38 and PACAP-27 exhibit high affinity whereas VIP has 10- to 100-fold lower affinity (11,27). PACAP also binds to VPAC1 and VPAC2 receptors with high affinity, equivalent to VIP, originally known as “PACAP type II receptors”. Hence, the actions of PACAP can result from activation of all of these receptors. Accordingly, we refer to them as a family of “PACAP-responsive” receptors when discussing them as a collective. We also use the term PAC1-like when the profile is consistent with this receptor subtype but a specific molecular entity cannot be firmly associated with the effect. A useful pharmacological tool for investigating PAC1 receptors is the unrelated peptide maxadilan. Originally isolated from sandflies by Lerner et al. (28), this peptide exhibits high affinity at the PAC1 receptor but none at 1 µ
These receptors have a typical class B GPCR structure and overall mechanism for binding their peptide ligands (Figure 1). The structured receptor extracellular N-terminal domain binds the peptide C-terminus, with the remainder of the peptide making contact with the receptor extracellular loops/upper transmembrane domains (TMD) (37,38). The ensuing conformational change in receptor structure produces intracellular signaling (Figure 1). As each PACAP-responsive receptor gene encodes multiple exons, there is potential for the generation of receptor variants through alternative splicing events. This is predominantly seen for the PAC1 receptor (Table 1). There is evidence that PAC1 splice and single nucleotide polymorphism variants are associated with an increased risk of post-traumatic stress disorder in females but there is little evidence linking specific variants to headache or migraine to date (39,40).
PACAP-responsive receptor splice variants.
aa = amino acids; ECL = extracellular loop; ICL = intracellular loop; TMD = transmembrane domain.
The generation of splice variants significantly increases the diversity of PACAP-responsive receptors. Alternative splicing of the PAC1 receptor gene (
Two human PAC1 variants are generated through alternative splicing of ICL3. These are named PAC1Hip (originally named SV-1) or PAC1Hop (SV-2), depending on whether the 28 amino acid exon 14 or exon 15 cassette is inserted (Table 1). Additionally, the Hop exon is present as either a 27 amino acid or 28 amino acid cassette due to different splice acceptor sites located three nucleotides apart (44,52). Thus, the addition of the 27 amino acid Hop cassette into ICL3 creates the PAC1Hop2 variant. The PAC1Hip and PAC1Hop1 variants have the same sequence length and are very difficult to differentiate without the use of highly specific primers. The PAC1Hop1 variant exhibits a similar relative order of ligand affinity to the PAC1n receptor, whereas the PAC1Hip variant is more similar to PAC1s due to its higher affinity for VIP (43,44).
Excluding the PAC1vs receptor, all human variants have been identified in rats and mice. Four additional receptor variants are reported in rats with no human counterparts identified to date (Table 1) (62). These appear to exhibit alterations in their signaling and ligand preferences, but additional data are required to fully understand their relevance. In addition to individual splicing events, studies have identified variants with splice combinations, such as the PAC1Hip−Hop1 (SV-3) and PAC1sHop1 receptors, but information regarding their functionality and significance is limited (43,44,59).
The VPAC1 (
Expression profiles in relation to migraine and pain
The trigeminovascular system and many central and peripheral pain processing centers have been linked to migraine. Numerous PAC1 variant transcripts are found in migraine-relevant tissues but it is usually unknown whether the variants are functional. Determining the extent of protein expression is challenging because antibodies cannot yet reliably distinguish between variants. Overall, mRNA suggests that the PAC1 receptor is the most prominent subtype expressed in the trigeminal ganglia (TG) with one study reporting the presence of PAC1n, PAC1s and PAC1 ICL3 variants (63–66). Comparatively, there is little expression of VPAC receptors but generally the VPAC2 receptor is more consistently detected and at higher levels than VPAC1. All three PACAP-responsive receptor subtypes are also expressed in the vasculature, such as the middle meningeal and cerebral arteries (67–70). Within the brain, PAC1 receptor mRNA is the predominant PACAP-responsive receptor, particularly PAC1n and PAC1Hop1 variants, but a role for VPAC receptors cannot be discounted even if their expression overall is comparatively lower (41,52,71,72). PAC1 and VPAC2 mRNA expression is reported in key processing and modulating pain regions such as the thalamus, hypothalamus, amygdala and brainstem, whereas VPAC1 expression appears more restricted, predominantly to the cerebral cortex and hippocampus (73–76). Within the spinal cord, regional differences in PAC1, VPAC1 and VPAC2 mRNA are observed and there is little detectable PACAP-responsive receptor mRNA in dorsal root ganglia (DRG) (64,65,71,77). Interestingly, migraine preclinical models suggest that the protein levels of PAC1, VPAC1 and VPAC2 receptors can be upregulated (78,79). A switch in relative expression may also occur between PAC1 and VPAC2 receptors during neuronal differentiation (80). Immune cells play an important role in neuroinflammation and sensitization and express at least one PACAP-responsive receptor (81,82). VPAC2 receptor mRNA is reported in mast cells and VPAC1 receptor mRNA in lymphocytes, but PAC1 expression is typically not measured and therefore cannot be ruled out. PAC1 and VPAC1 mRNA are found in microglia and macrophages, with VPAC2 receptor expression in macrophages becoming prominent following an immune response trigger (81). Altogether, the expression of PAC1, VPAC1 and VPAC2 receptors at relevant sites indicates that all receptor subtypes could play their own role in migraine. The majority of studies have been performed in rodents and there is limited understanding of receptor subtype expression in human tissues.
Signaling of canonical receptors
GPCR function is tightly controlled, ensuring appropriate responses to stimuli. Once a GPCR has activated G proteins, it is typically phosphorylated, desensitized and removed from the cell surface by internalization. Protein kinase A and GPCR kinases (GRK) are typical kinases associated with these processes. Receptor internalization mediated by the regulatory scaffold protein, β-arrestin, is the most well-established mechanism of internalization; however, β-arrestin-independent mechanisms have also been reported (83,84). Once internalized, GPCRs can be recycled and/or degraded or may continue to signal from intracellular compartments such as endosomes, golgi and nuclear membranes (85). Unlike cell surface receptor signaling, endosomal signaling is proposed to generate unique and sustained signals, which have been linked to central sensitization and chronic pain (86–88). Another important facet in GPCR function is biased signaling, which refers to mechanisms where an individual receptor can produce distinct profiles of signaling in response to different ligands, in different cellular locations, or in different cell types (89,90). Hence, when comparing between studies, it is important to consider that the events reported are likely to be heavily influenced by receptor expression level, cell type, the ligand(s) used and the location of the receptors within cells. Overall, studies from transfected cells provide a useful guide as to which pathways have the capacity to be activated, but these are best complemented by studies in physiologically-relevant systems.
A diverse set of intracellular signaling molecules is involved in the initiation, modulation and amplification of pain and its chronification (91–93). Akin to many other GPCRs, PACAP/VIP binding to its cell surface receptors can lead to a myriad of intracellular signaling responses (Figure 2) (94). The complement of signaling pathways activated by the PACAP/VIP family of peptides is broadly similar to the pro-nociceptive peptides CGRP and substance P (95–97). PACAP-responsive receptors are classically known as Gαs protein-coupled, which leads to the accumulation of cAMP. However, the receptors also have the ability to couple to other G proteins and downstream effector proteins. Ligand binding can promote association with β-arrestin and there are emerging data on internalization and endosomal signaling (Figure 2). Furthermore, signaling can be fully dependent, partially dependent or independent of G proteins, complicating our ability to delineate pathways and understand their physiological significance. However, the diverse range of signaling mediators, in combination with multiple PACAP-responsive receptor ligands, allows the opportunity for biased signaling to occur.

Overview of pituitary adenylate cyclase-activating polypeptide (PACAP)-responsive receptor signaling pathway activation and regulation. Created in BioRender. P, phosphorylation. For signaling molecule definitions, refer to the text. Hay, D. (2025) https://BioRender.com/ct58d1s.
G protein-mediated signaling and downstream proteins
Transfected cells with defined receptor subtypes
A significant proportion of the PACAP-responsive receptor literature describes the ability of peptides to stimulate G protein-mediated signaling pathways, such as to increase the cellular concentration of cAMP (Gαs-mediated), inositol phosphates (IP) or calcium ions (likely Gαq-mediated) (Table 2). At human and rodent PAC1n and PAC1Hop1 receptors, PACAP-38 and PACAP-27 are potent agonists of these pathways; VIP is approximately 100-fold less potent, and PHM is described as a weak activator, consistent with its ligand binding profile (12,13,43,98,99). Until recently, limited signaling data were available for the human and mouse PAC1s receptors, particularly regarding the PHM/I peptides, and no functional information was known about the rat PAC1s receptor, but a more complete picture is now emerging. In human and rat, potent cAMP accumulation is produced by VIP, with a potency comparable to PACAP-38 and PACAP-27. When measuring IP accumulation or calcium flux, VIP has reduced potency but remains more potent compared to the PAC1n receptor (12,42,43,50,100). PHM/I also stimulate G protein-mediated signaling at the PAC1s receptor, but these have consistently lower potency compared to PACAP and VIP (12,50). The PAC1vs receptor can stimulate accumulation of cAMP comparable to the PAC1n receptor but with an overall ~100-fold lower potency, and it is described as being incapable of stimulating IP production (41,43). At the PAC1Hip receptor, PACAP-38 and PACAP-27 potently stimulate cAMP accumulation and a single study that tested VIP found it exhibited a similar potency to PACAP, like the PAC1s receptor (43,44,52). However, PAC1Hip production of IP is more variable with reports of complete inability to stimulate this pathway, differences in peptide potency and potential species-dependent activation, which requires additional investigation to clarify (43,44,52,101).
Summary of PACAP-responsive receptor downstream signaling and regulation.
*Most data are from human receptors with other species needing more detailed study. **For the purposes of this overview, β-arrestin is categorized in its canonical regulatory role, however it is important to note that this protein can also facilitate activation of certain signaling pathways due to its role as a molecular scaffold.
Akt = protein kinase B; CREB = cAMP response-element binding protein; ERK = extracellular signal-regulated kinase; PACAP = pituitary adenylate cyclase-activating polypeptide; VIP = vasoactive intestinal peptide.
PACAP-38, PACAP-27 and VIP are potent agonists of cAMP and IP accumulation at the VPAC1 and VPAC2 receptors, whereas PHM/I is reported to be equipotent or 10-fold less potent compared to VIP (12,13,34,58). Less is known about the ability of rodent receptors to activate IP accumulation, but data indicate that they may behave similarly to human receptors (110,111). Calcium signaling for VPAC receptors may also involve Gαi in addition to Gαq coupling and the contribution of Gαi may differ between the subtypes (111–113).
PACAP-responsive receptors also stimulate the phosphorylation of downstream proteins such as extracellular signal-regulated kinase (ERK), cAMP response-element binding protein (CREB) and protein kinase B (Akt) (Table 2). However, data are limited and few studies directly compare PACAP and VIP, multiple receptor subtypes and more than a single agonist concentration (12,13). Inhibitors of G protein-mediated signaling indicate there may be no single mechanism of ERK phosphorylation by PAC1 receptors (114,115). In transfected cells, PAC1n, PAC1s, VPAC1 and VPAC2 receptors produce ERK, CREB and Akt phosphorylation in response to PACAP-38, PACAP-27, VIP and PHM, consistent with their ligand binding profiles (12). The PAC1Hop1 variant can also stimulate ERK and Akt phosphorylation in response to PACAP-27 but other ligands have yet to be tested, nor has CREB phosphorylation for any ligand (105,106).
Primary cells and cell lines with endogenous receptors
In cerebral cortex neurons, astrocytes, TG neurons and glia, and spinal cord cultures, PACAP, VIP and PHI all stimulate the accumulation of cAMP with a profile indicative of a PAC1-like receptor (50,66,116–118). A PAC1-like receptor that stimulated cAMP production was also identified in the hypothalamus and cerebellum (118). The PACAP family of peptides are also able to stimulate IP production or calcium (ion) flux in the hypothalamus, cerebellum, TG, spinal cord and astrocyte cultures where a predominately PAC1-like receptor profile is observed (50,118–120). In TG cultures, responses were also observed with a VPAC2 receptor agonist indicating the presence of multiple PACAP-responsive receptors (120).
ERK phosphorylation also occurs in TG glia, spinal cord, cardiac ganglia and myeloma cells, with the ligand and PCR expression profiles suggesting that these responses arise from endogenously expressed PAC1-like receptors (50,66,114,115). Similarly, VPAC1 receptors in neutrophils and VPAC2 receptors in a pituitary cell line also respond to PACAP/VIP ligands to stimulate ERK phosphorylation (121,122). PACAP simulates ERK phosphorylation in cerebellar granule neurons, astrocytes and a neuroblastoma cell line; however, these responses cannot be attributed to a single PACAP/VIP-responsive receptor as multiple subtypes are endogenously expressed and only a subset of ligands were tested (123–126). At endogenously expressed PACAP-responsive receptors, CREB phosphorylation was observed at PAC1-like receptors in spinal cord and myeloma cells, VPAC1-like receptors in microglia and an unknown subset of receptors in cerebellar granule neurons, neuroblastoma and hypothalamic cells (50,115,125,127–130). Less is known about Akt phosphorylation in endogenous systems, but this has been reported for a PACAP-38-responsive receptor in cerebellar granule neurons but not in spinal cord cultures (50,128). p38 mitogen-activated protein kinase is reported to be phosphorylated downstream of VPAC1 receptor activation in microglia and neutrophils but not following PAC1-like receptor activation in TG neurons (66,121,127).
Ligand and signal bias
The ability of PACAP-responsive receptors to activate multiple signaling pathways, combined with several endogenous agonists that these receptors respond to, provides an opportunity for biased behavior not only in relation to synthetic biased ligands, but also potentially as an endogenous mechanism for fine-tuning cellular activity. Reports of bias have been inter-woven into the PACAP peptide family literature for many years. For example, PACAP-38 was reported to be significantly more potent than PACAP-27 at the human PAC1n and rat PAC1n and PAC1Hop receptors when measuring IP production but PACAP-38 was equipotent to PACAP-27 for cAMP production (42,52). A similar profile was observed at endogenously expressing PAC1-like receptors in cerebellar granule and PC12 cells (131,132). However, this difference was not reproducible in later studies at both human and rat receptors, suggesting that differential coupling of PAC1 variants may be highly dependent on the method and cellular background (12,44,50,98,99,133). More recent evidence of biased agonism has been reported in TG glia cultures, where PACAP-38 and PACAP-27 stimulated accumulation of cAMP, but only PACAP-38 was able to phosphorylate ERK (66). Overall, biased behavior is likely present for the PACAP/VIP receptor family but the cellular background and methodology have a large influence on the outcome.
Internalization, β-arrestin recruitment and endosomal signaling
PAC1 receptor
The PAC1 receptor has a cluster of potential phosphorylation sites in ICL3 and the C-terminus that may facilitate β-arrestin interactions and thus internalization (94). A PAC1n and PAC1Hop1 receptor truncation study indicated a contribution of the distal C-terminal region, but additional regions, such as the Hop1 cassette and ICL3, are also likely to play important roles in internalization and β-arrestin recruitment (134). Internalization of the PAC1 receptor has been studied (Table 2); however, the literature largely focuses on the PAC1Hop1 splice variant and less is known about other variants, such as the PAC1n and PAC1s receptors. Furthermore, most studies consider only a single ligand and do not directly compare PACAP-38 and PACAP-27. Despite these limitations, time and concentration-dependent internalization of PAC1n and PAC1Hop1 receptors is reported following PACAP-38 and PACAP-27 stimulation (102,105,107,135–137). The PAC1Hop1 variant insert is within a region where phosphorylation and β-arrestin interactions are likely to occur (Figure 2); therefore, it may exhibit differential behaviors from other variants, although there is no clear consensus from the literature (102,104,134,138). Although both variants internalize and colocalize with β-arrestin in vesicular structures, the PAC1Hop1 receptor internalizes more rapidly, exhibits an increased number of interactions with β-arrestin 1 when modeled and is retained within the cell cytoplasm alongside β-arrestin for longer compared to PAC1n (102,104,139). As such, it is possible that intracellular signaling of this variant could differ from other variants. Although PAC1 can recruit both β-arrestin 1 and 2 proteins, its trafficking and involvement of each β-arrestin in receptor internalization may vary but there is not a complete picture for all variants (107,140). For example, β-arrestin 2 was reported to facilitate PAC1Hop1 internalization, whereas β-arrestin 1 did not appear to be involved (107). Only one study has looked at the internalization of the PAC1s variant (103). Although VIP has a higher binding affinity at the human PAC1s relative to PAC1n, VIP-mediated internalization was comparatively weaker and similar to the PAC1n receptor, suggesting possible pathway bias for this variant.
Internalization of endogenously expressed PAC1 receptors has been investigated in a pancreatic neuroendocrine tumour line (expresses PAC1Hop1 and VPAC1), rat PC12 cells (PAC1) and mouse cortical neurons (PAC1 and VPAC1), which all exhibit internalization of PAC1 receptors in response to PACAP-38 stimulation (137,141,142). This indicates that PAC1 internalization is likely a universal mechanism between different cell systems and at endogenously expressed receptor levels. One study has reported an increase in PAC1 receptor immunoreactivity in cytosolic compared to cytomembrane TG fractions following chronic nitroglycerin or PACAP injections (142). However, the particular PAC1 antibody used detects many non-specific proteins in TG tissue and currently no consistent pattern has emerged from the literature between the same or different antibodies targeting PAC1, complicating interpretation (143–146). Therefore, endogenous receptor internalization is yet to be established in migraine-relevant regions and in response to PACAP-27 or VIP, indicating a major gap in knowledge. The utilization of fluorescently-tagged PACAP/VIP ligands may help elucidate some of these mechanisms (103,135,147,148).
Several studies have implicated PAC1 receptor internalization and subsequent endosomal signaling in regulating neuronal excitability, including modulation of hippocampal dentate gyrus neurons, cardiac neurons and nociceptive responses in the central nucleus of the amygdala (139,149–151). Phosphorylated ERK is the most studied endosomal signaling molecule for PAC1 receptors, because sustained activation is reported in chronic pain models, and PAC1 seems central to these processes (139). However, there is evidence of additional endosomal signaling pathways for GPCRs (85,86,152). Studies using the PAC1Hop1 receptor show that sustained ERK phosphorylation involves β-arrestin 2 recruitment but is only partly dependent on receptor internalization (107,153). By contrast, the involvement of β-arrestin 1 in sustained ERK phosphorylation is unclear because β-arrestin 1 silencing either produces an increase or complete abolishment of ERK phosphorylation (107,154). However, these studies used different cell backgrounds, which may explain this difference. There also appears to be a difference in the degree of endosomal ERK phosphorylation between the PAC1n and PAC1Hop1 variants; mutation of key β-arrestin interacting residues of the PAC1Hop1 ICL3 reduces ERK phosphorylation to levels observed at the PAC1n receptor (104). Endosomal Akt phosphorylation has been investigated for PAC1 receptors and may be an additional signaling molecule activated by this receptor family (12,105).
VPAC1 and VPAC2 receptors
In the context of receptor regulation and endosomal signaling, the VPAC receptors are generally less well-studied compared to the PAC1 receptor. However, the data do suggest that both receptors interact with β-arrestins and internalize but exhibit distinct trafficking profiles. VIP-stimulated VPAC1 and VPAC2 receptor internalization has been repeatedly shown for both human and rodent receptors (147,155–160). Only two studies have probed internalization using PACAP-38 and PACAP-27 (103,108). These generate a similar degree of internalization compared to VIP. Interestingly, although both VPAC1 and VPAC2 receptors are rapidly internalized, only the VPAC2 receptor appears to be recycled back to the cell surface within two hours following agonist removal (159–161). This recycling mechanism is thought to be dependent upon phosphorylation of intracellular residues and β-arrestin. Mutagenesis studies have identified two residues important in preventing recycling of the VPAC1 receptor: one in TMD3 and another in the C-terminus (162,163). However, it has been difficult to identify additional phosphorylated residues involved in receptor internalization as there appears to be no direct correlation between the two processes (147,161). A significant distinction between the VPAC1 receptor and many other class B GPCRs is that, although VPAC1 recruits both β-arrestin 1 and 2, these are not required for receptor internalization (84,156,157). However, the process is dynamin-sensitive. By contrast, β-arrestin 2 is implicated in VPAC2 receptor internalization, but its mechanisms have not been studied in depth, and the role of β-arrestin 1 is currently unknown (108,109). In terms of endosomal signaling, one study has reported that the VPAC1 receptor can produce endosomal cAMP signals after VIP stimulation (84). There is currently no information for the VPAC2 receptor but endosomal signaling may occur because the receptor can internalize. Overall, these data highlight that although PACAP and VIP can activate multiple canonical receptors, variations in the downstream mechanisms of signaling and regulation could underlie biological differences in ligand behavior.
RAMPs and PACAP-responsive receptor dimerization
To add further complexity to the PACAP/VIP receptor family, interactions with accessory proteins, such as receptor activity-modifying proteins (RAMPs) or the formation of receptor dimeric or oligomeric complexes have been reported. Such receptor:receptor or receptor:RAMP complexes are reported to alter peptide affinity, signaling and cellular localization (164–167). Unlike the canonical CGRP receptor, which is an obligate dimer between the calcitonin receptor-like receptor and RAMP1, the PACAP-responsive receptors do not require a RAMP to be expressed at the cell surface or for functionality. The PAC1, VPAC1 and VPAC2 receptors are predicted to interact with all three RAMPs (112,168–171). However, detecting physiologically-relevant receptor:RAMP interactions is challenging, and the existing data for PACAP-responsive receptors provide no clear consensus even from transfected cell systems. One key issue is that, if co-transfection of RAMP and receptor produces a functional effect, this could potentially be explained by indirect effects on endogenously expressed receptors and may be difficult to discern from a direct interaction. Association of the PAC1 receptor with all three RAMPs is inferred but the functional consequences of such interactions are currently unknown (170,171). However, although no direct interaction was reported, a preprint has reported that PAC1 promoted the cell surface expression of RAMP1, suggesting some sort of association between the pair (171). RAMP2 is reported to enhance IP but not cAMP accumulation for the VPAC1 receptor and RAMP2 association with the VPAC2 receptor may alter G protein activity and increase PHM affinity (112,169).
Studies of PAC1 dimerization are limited but the N-terminal HSDCIF motif has been identified as key in dimerization of the PAC1Hop1 receptor (172). PAC1 dimers may also exhibit ligand-independent basal activity (173). Furthermore, it was found that PAC1 dimers on nuclear membranes produced higher cAMP concentrations compared to those on the cell surface (174). As well as homodimerization of the PAC1 receptor, additional receptors may form heterodimers with PAC1 through its TMD (175). These potential interacting partners are the glucose-dependent insulinotropic polypeptide (GIP) and secretin receptors. VPAC1 or VPAC2 receptor interactions were not observed with these same receptors in this study. Interestingly, compared to the PAC1 receptor alone, dimerization with the GIP receptor but not secretin receptor suppressed PACAP-stimulated signaling and neuronal differentiation, meaning that these dimerization events could have functional consequences (175). Formation of VPAC1 and VPAC2 receptor homodimers has been reported with no change in receptor pharmacology (176–178). However, VPAC2 dimerization may be required for Gαi-mediated signaling (178). Interestingly, VPAC1 and VPAC2 can also interact to form a functional heterodimer, in addition to forming heterodimers with the secretin receptor that become intracellularly trapped, resulting in complete inhibition of signaling (176,177). Hence, although dimerization of the PAC1, VPAC1 and VPAC2 receptors is reported, caution is warranted because results are not always consistent, and many of the studies have used modified receptors that may not accurately reflect naturally-occurring receptors in endogenously expressed systems.
Non-canonical receptors
In addition to the canonical receptors described above, PACAP has also been proposed to bind and activate other unrelated receptors. These may serve to mediate some of the functional effects of PACAP, including those relevant to migraine. These receptors are GPR55, and a mas-related (mrg) receptor, MRGPRX2. However, neither of these receptors have been de-orphanized, which means that these receptors are not officially recognized as PACAP-responsive receptors. Therefore, their gene names will be used when referring to both receptor gene and protein. In addition to PACAP ligands, multiple unrelated ligands have been reported to modulate the activity of these receptors.
GPR55 is widely expressed including in DRG, in mast cells and in regions of the brain (179–182). This receptor has been implicated in metabolic disorders, neuropathic pain and inflammation (183–186). GPR55 was initially identified as a cannabinoid-like receptor but can also be activated by lysophospholipids and peptides including PACAP (181,187,188). In response to cannabinoid ligands and lysophosphatidylinositol, GPR55 can couple to Gαi and Gαq proteins, promoting calcium flux and phosphorylation of ERK and CREB (187,189). These pathways have not been investigated for PACAP and GPR55, although a single study has reported that PACAP-38 and PACAP-27 stimulation could promote GPR55 internalization but not β-arrestin recruitment (188). Additional studies are required to confirm these findings and elucidate any further role of GPR55 in the physiological actions of PACAP.
Neurogenic inflammation plays an important role in migraine and chronic pain. One of the cell types involved in neurogenic inflammation is mast cells. These cells can be found in close proximity to peripheral nerve endings (190,191). In humans, the proposed novel PACAP-responsive mast cell receptor is named MRGPRX2 and many naturally occurring missense variants and single nucleotide polymorphisms are reported (192,193). MRGPRX2 is mainly found in mast cells but also has limited expression in the periphery and central nervous system, notably the highest being in DRG (194–196). Rats and mice exhibit significant expansion of their mrg coding genes compared to humans, and no obvious rodent gene ortholog for MRGPRX2 has been identified thus far (194,197,198). However, pharmacological and expression data suggest that

Amino acid sequence alignment of human MRGPRX2, with its proposed mouse (Mrgprb2) and rat (Mrgprb3) orthologs. Black = 100% similarity, dark grey = 80–100%, light grey = 60–80% and white = < 60% similarity using the in-built ClustalW sequence alignment with Blosum62 matrix in Geneious Prime. TMD = transmembrane domain.
Ligands for mrg receptors are predominately identified based on their mast cell degranulation ability. Because this response could be mediated by a range of endogenous receptors in mast cells, this only provides indirect evidence that specific receptors mediate these effects. PACAP-38 and PACAP-27 were initially identified as agonists of mast cell receptors based on this ability and because mast cells do not endogenously express a PAC1 receptor (200,203–205). Interestingly, PACAP6-38, which acts as an antagonist at the canonical PACAP-responsive receptors, acted as an agonist for mast cell degranulation (201,205). VIP can promote mast cell degranulation, but this is presumed to occur through the endogenously expressed VPAC2 receptor (201,205,206). PACAP is also a potent agonist at VPAC2 receptors, and therefore the degranulation response that is presumed to occur through an mrg receptor, could potentially be explained in some contexts by the VPAC2 receptor. Despite this, a role of mast cell receptors in the actions of PACAP-38, especially its ability to sensitize TG neurons and dilate middle meningeal arteries, is still likely because these responses were not observed in Mrgprb2 or mast cell-deficient mice (200,207).
Signaling of the MRGPRX2 receptor is predominantly measured using calcium flux, although changes in cAMP accumulation and β-arrestin recruitment have been reported with some proposed ligands (208,209). In cell models transfected with the human MRGPRX2, rat Mrgprb3 or mouse Mrgprb2 receptors, calcium flux or IP accumulation are reported in response to PACAP and VIP. In addition to calcium signaling, PACAP/VIP ligands are reported to stimulate the recruitment of β-arrestin to the MRGPRX2 receptor but further G protein-dependent and independent pathways have yet to be elucidated (210,211). Overall, it appears that PACAP and VIP can act as ligands for mrg receptors, although typically 100-fold higher concentrations are required compared to the canonical PACAP-responsive receptors. One consideration that the field should be aware of is that many commonly transfected cell lines, such as HEK293 cells, endogenously express PACAP-responsive receptors (84,157,212). Therefore, functional responses could be mediated by this endogenous receptor as well as or instead of the receptor introduced by transfection. It is important that researchers check for background receptor responses, such as by using a vector only control transfection condition, to avoid potential misinterpretation of the data (66).
Designer drugs
As neuropeptides, there are several ways in which PACAP and VIP signaling could theoretically be modulated, some of which are more amenable to pharmacological approaches than others (Figure 4). In the context of migraine, the hypothesis is that PACAP-responsive receptor signaling needs to be reduced, and, below, we therefore consider the potential approaches in this light (3).

Potential ways to therapeutically target neuropeptide signaling systems. To impact the quantity of neuropeptide available for receptor signaling, the biosynthetic neuropeptide processing pathway that produces mature bioactive neuropeptide (1), the release mechanism from dense core vesicles (2) and post-release proteolytic neuropeptide processing (3) could each be modulated. The mature bioactive neuropeptide could be targeted with an antibody to prevent it from activating its receptor(s) (4) and, lastly, several strategies to target the receptor(s) itself could be taken, such as different modalities of ligand (e.g. antagonist, allosteric modulator, biased ligand) to target distinct pools of receptor, distinct signaling pathways, or receptors on distinct cell types (5). GPCR = G protein-coupled receptor. Created in BioRender. Hay, D. (2025) https://BioRender.com/a2tfo7g.
Neuropeptides are synthesized from larger precursor peptides called prepro and propeptides, and, as such, there is the potential to disrupt their processing or maturation, and thus how much functional peptide is released (1,11). The release mechanism itself could also be targeted, although this would need to be specific to the peptide of interest, and there is still much to understand about how many neuropeptides may be stored together in individual dense core vesicles, and about the release mechanisms. There is no transport-mediated reuptake mechanism for neuropeptides, as there is with neurotransmitters. Therefore, enzymatic mechanisms usually control the amount of bioactive peptide present after release. For example, PACAP can be cleaved by dipeptidyl peptidase-IV (213). To reduce bioactive neuropeptide concentrations, there would be the need to identify a way to enhance enzyme activity or metabolize the peptides in other ways.
Modulating the functional bioactivity of neuropeptides is a relatively achievable therapeutic approach. One strategy is to develop antibodies that bind to the peptide and modify its pharmacological activity (anti-ligand antibodies). In the context of PACAP, this strategy has already met with some success. Two monoclonal antibodies against the PACAP peptide reached clinical development. Lu AG09222 (ALD 1910) showed superiority over placebo in reducing monthly migraine days and has progressed to phase III clinical trials (9). Clinical development of LY3451838 appears to have been terminated, although the reasons for this are unclear. In a similar manner to anti-CGRP monoclonal antibodies, PACAP-targeted antibodies bind PACAP to prevent it from binding to and activating its receptors (214). Given that PACAP exists in two forms, it is important to consider which forms of PACAP these antibodies recognize.
Antibodies can also be developed against the receptor that can act as antagonists (anti-receptor antibodies). One such antibody, AMG301, which was developed to target the PAC1 receptor is an example of this (215). Although this antibody reached clinical development, there was no therapeutic benefit in a phase II trial of this antibody for migraine prevention. Given the complexity of this receptor system highlighted elsewhere in this review, there could be a range of factors that contributed to this. More commonly, peptide or small molecules are designed to target the receptor, with the modalities to reduce neuropeptide signaling being an antagonist or negative allosteric modulator. There are several examples of peptide antagonists and of small molecules that have been used in
In the case of any receptor-centric approach, there are a number of important considerations. Some of these are wider considerations for targeting GPCRs in light of contemporary understanding. Some may be unique to or exemplified by PACAP-responsive receptors. Approaches to target GPCRs are becoming more sophisticated, going beyond simple affinity or selectivity-driven strategies. Based on increased understanding of how, and from where receptors signal, different pools of receptors can potentially be targeted, such as cell surface or endosomal. The cellular location of receptors is important in several ways. For example, a receptor antibody would likely preferentially inhibit cell surface receptors. A small molecule, depending on its physicochemical properties, and features that could be specifically engineered, could have accessibility to different pools of receptors. Therefore, the location of receptors is an important consideration both for the design of novel entities, and for interpreting how existing agents may act.
Biased ligands could be developed that favor specific signaling profiles. However, it is important to determine, in physiologically-relevant systems, which pathways are beneficial or detrimental in a given context. For example, although PACAP-27 exhibits bias for cAMP production over ERK phosphorylation in TG glia, whether this should be targeted or avoided is not known. Whether endogenous PACAP-27 is important in the context of migraine remains to be demonstrated. However, a critical aspect is its relative sensitivity to antagonism compared to PACAP-38. Although the mechanism is not known, PACAP-38 is relatively resistant to antagonism compared to PACAP-27. This has been observed in primary cell cultures, in human arteries and in cells transfected with PACAP-responsive receptors (12,50,66,70). A major consequence of this behavior is that the use of PACAP-38 to screen for antagonists could result in a low sensitivity to detect hits because this peptide is more difficult to antagonize.
A key remaining question that extends beyond modality is which receptor or combination of receptor(s) should be targeted for migraine or other conditions. This remains an open question given the complexity of the receptor system. A within-receptor family multi-target approach may need to be considered if multiple PACAP-responsive receptors contribute to migraine. However, it is not only a matter of which receptor(s) (i.e. PAC1, VPAC), but also which splice variant(s)? Unfortunately, it is not yet known which of these may, or may not, be important but each of these potentially presents a unique opportunity for developing designer drugs. For example, structural differences as a consequence of variations in the length of the PAC1 N-terminus could create unique binding pockets for small molecules (Figure 1 and Table 1). As non-canonical receptors have emerged, there may be more targets than initially realized. Higher order complexes, such as dimers or RAMP:receptor complexes, may also need considering as an emerging strategy because there could be a more diverse receptor binding interface or unique functional properties to target. A recent study has considered the potential for targeting both PACAP and CGRP systems at once with single peptides that are capable of antagonizing receptors from both families (216).
Thus, the opportunities for “designer” drugs against the PACAP signaling system seem extensive but not all approaches will be pharmacologically tractable, and safety remains an important question. Differences between species need to be borne in mind as target validation studies are pursued to ensure translatability of mechanisms and molecules to the clinic.
Conclusions
This narrative review has provided a summary of current understanding of PACAP-responsive receptor ligands, as well as canonical and candidate non-canonical receptors. This system exhibits significant complexity in the number of receptors and signaling pathways that can be activated. There are many open questions in the field, including which receptor(s), ligands(s), cell signaling pathway(s) or location(s) would provide an optimal therapeutic in migraine or other conditions. In part, this is because of gaps in understanding how this system functions in a pathophysiological setting. Despite this complexity, as a neuropeptide:GPCR signaling system, it is likely to be pharmacologically tractable as already evidenced by several investigational drugs reaching clinical trials.
Article highlights
PACAP and VIP are neuropeptides that are implicated in migraine. They act through a sub-family of canonical GPCRs (PAC1, VPAC1 and VPAC2), with functional diversity played out through receptor splice variation and a variety of signaling mechanisms. Non-canonical receptors are proposed that may further increase the spectrum of cell signaling processes. Opportunities for drug design include targeting these receptors individually, or in combination, and targeting the peptide ligands.

