Functional characterisation of dura mater αCGRP degradation fragments at vascular, immune, and CGRP receptor targets Authors: Mette N. Jensen, Jens B. Kristensen, Maria Castrillo, Philip V. Reducha, Karin M. L. Nordahl, Debbie L. Hay, Kristian A. Haanes Published: 2026 ABSTRACT ObjectiveCalcitonin gene related peptide (CGRP) is a central mediator in migraine and can trigger attacks in susceptible individuals. Therapies targeting CGRP or its receptor provide clinical benefit, but treatment responses vary. This suggests that local factors regulating CGRP availability and signalling within migraine relevant tissues may influence its biological effects. The extent and functional relevance of CGRP degradation at the dura mater, including possible effects on biomarker detection, remain insufficiently characterised.MethodsProteomic analyses were used to identify human αCGRP (hαCGRP) degradation fragments generated in rat meningeal preparations. Functional properties of full-length hαCGRP and selected degradation fragments were assessed at the dura mater, in isolated arteries, in mast cell degranulation assays, and at CGRP receptors in vitro. Fragment interactions with a CGRP ELISA were examined to assess potential interference with detection of full-length CGRP. Male Sprague Dawley rats were used for the ex vivo experiments.ResultshαCGRP disappeared rapidly in the meningeal environment and reduced transdural resistance in the dura mater. Seven predominant degradation fragments accounted for ~ 81.5% of the detected fragment pool. None of the fragments showed agonist activity at the CGRP receptor or altered dural resistance or mast cell degranulation. In vascular assays, the fragments were likewise inactive, with the exception of the C-terminal fragment hαCGRP11−37, which displayed weak competitive antagonism at the CGRP receptor. This was confirmed at both the human and rat CGRP receptor in vitro. Although the identified fragments were not directly detected by the CGRP ELISA, selected fragments interfered with intact hαCGRP measurements, indicating that degradation products can influence biomarker readout.ConclusionLocal degradation of hαCGRP at the dura mater primarily generates fragments without measurable vascular, immune, or receptor activating effects. The main consequence of this processing appears to be reduced hαCGRP bioavailability rather than formation of secondary bioactive signalling peptides. At the same time, fragment dependent interference in ELISA based detection indicates that local hαCGRP metabolism may affect interpretation of CGRP biomarker measurements in migraine relevant tissues.Supplementary InformationThe online version contains supplementary material available at 10.1186/s10194-026-02390-3. Introduction Migraine is a highly prevalent and disabling neurovascular disorder affecting approximately 15% of the global population [1, 2] and representing one of the leading causes of years lived with disability worldwide [3, 4]. Although its pathophysiology is multifactorial and incompletely understood, activation of the trigeminovascular system is widely recognised as a central component [5–8]. This involves stimulation of trigeminal sensory afferents innervating the meninges, particularly the dura mater, leading to the release of vasoactive and pro-inflammatory neuropeptides that contribute to nociceptive signalling and vascular responses implicated in migraine pathophysiology [8–11]. Within this anatomical compartment, the dura mater represents a key interface between neuronal terminals, vascular elements, and resident immune cells, creating a microenvironment in which released neuropeptides are exposed to extracellular enzymes capable of rapidly modifying or degrading signalling molecules [12]. Among these mediators, calcitonin gene-related peptide (CGRP) has emerged as a key molecule in migraine biology [13, 14]. CGRP is a 37-amino acid neuropeptide, where αCGRP is the most abundantly expressed form in trigeminal neurons [15] and is released upon activation of the trigeminovascular system [16]. It is a potent vasodilator and modulator of nociceptive transmission, acting via the canonical CGRP receptor complex consisting of the calcitonin receptor-like receptor (CLR) and receptor activity-modifying protein 1 (RAMP1) [17–19]. Although specificity of some of the ELISA kits used has been questioned [20], there are previous studies, showing elevated CGRP levels during migraine attacks (reviewed by Garate et al. [21]), and intravenous CGRP administration can trigger migraine-like headaches in susceptible individuals [22, 23]. These observations have driven the development of targeted therapies, including small-molecule CGRP receptor antagonists (gepants) and monoclonal antibodies acting as either CGRP receptor antagonists or CGRP scavengers [18, 24]. The clinical success of these treatments firmly establishes CGRP as a central therapeutic target in migraine. However, not all patients respond to CGRP-targeted therapies [25], and some may experience clinical benefit after switching from one CGRP-directed treatment to another [26]. This variability underscores the complexity of CGRP signalling in migraine and highlights the need for a comprehensive understanding of its underlying biology. One part that is not fully understood is local peptide metabolism, and how it may influence the availability and signalling dynamics of CGRP within migraine relevant tissues. As a peptide, αCGRP is rapidly degraded in vivo and has a half-life of only a few minutes [27]. Extracellular proteases, including endopeptidases and mast cell derived proteases, can cleave αCGRP at multiple sites along the peptide sequence, thereby generating truncated fragments that may alter the duration and nature of CGRP signalling [28, 29]. Although αCGRP degradation patterns have been studied in plasma and spinal cord [30, 31], endogenous αCGRP metabolites generated at the dura mater, a highly relevant site of action in migraine, have not been systematically characterised. This leaves an important gap in understanding how local enzymatic processing may regulate αCGRP bioavailability and fragment formation within the meningeal environment. Within the CGRP field, the N-terminally truncated peptide hαCGRP8−37 is a well characterised competitive antagonist of the CGRP receptor and is widely used as a pharmacological tool to study CGRP signalling pathways [32, 33]. These observations raise the possibility that naturally generated CGRP fragments could retain receptor binding or modulatory properties despite lacking full agonist activity. Another important consideration is that peptide degradation products may influence the biochemical detection of CGRP. Because many analytical assays rely on antibody recognition of short peptide epitopes, truncated CGRP fragments could theoretically interfere with immunoassay-based measurements of CGRP levels and thereby contribute to variability observed in CGRP-biomarker studies [21]. As reported for some endogenous peptide hormones, metabolic processing does not always terminate their activity but can sometimes generate fragments with distinct biological and pharmacological properties [34–38]. Therefore, identifying and characterising endogenous CGRP metabolites generated at relevant sites of action could reveal unrecognised aspects of CGRP biology. Accordingly, the present study aimed to identify CGRP degradation fragments in rat dura mater and evaluate their biological activity. We examined interactions with the CGRP receptor, effects on CGRP-induced vasorelaxation, and potential involvement in inflammatory responses. Methods Buffers Buffers were prepared in advance for the experimental procedures described below. For experiments involving the dura mater, synthetic interstitial fluid (SIF) consisted of 108 mM NaCl, 3.5 mM KCl, 3.5 mM MgSO₄, 26 mM NaHCO₃, 11.7 mM NaH₂PO₄, 1.5 mM CaCl₂, 9.6 mM sodium gluconate, 5.6 mM glucose, and 7.6 mM sucrose at pH 7.4. For wire myography experiments, Krebs buffer consisted of 119 mM NaCl, 4.7 mM KCl, 2.5 mM CaCl₂, 25 mM NaHCO₃, 1.17 mM MgSO₄, 1.18 mM KH₂PO₄, and 5.5 mM glucose. For Liquid chromatography/mass spectrometry (LC-MS) analyses performed by Biogenity, the lysis buffer consisted of 6 M guanidinium hydrochloride, 10 mM tris (2-carboxyethyl)phosphine, 40 mM 2 chloroacetamide, and 50 mM HEPES at pH 8.5. All chemicals were purchased from Sigma Aldrich (Merck). A complete list of peptides, compounds, and their respective suppliers used in the in vitro and ex vivo experiments is provided in the Supplementary Compound List. Animals Fifty-nine male Sprague-Dawley rats (250 to 350 g) purchased from Taconic (Denmark) were used throughout and allowed a minimum of one week acclimatisation in house prior to experiments. The rats were housed in Eurostandard cages (Type VI with 123-Lid) in groups of a maximum of five. Food (standard chow and nuts) and water were provided ad libitum. The temperature was maintained at + 22 °C (± 0.5 °C) and humidity at 55% (± 10%). A 12 h light-dark cycle was maintained. The study complied with Directive 2010/63/EU, and the procedures were approved by the Danish Animal Experiments Inspectorate (license number 2025-25-0201-01469) and reported following ARRIVE 2.0 guidelines. All ex vivo experiments were performed at the Translational Research Centre, Rigshospitalet Glostrup, Denmark, and all in vitro experiments were performed at the Department of Pharmacology and Toxicology, University of Otago, Dunedin, New Zealand. For tissue collection, rats were sedated in 70% CO₂ gas in O₂ and euthanised by decapitation. Ussing chamber Dura mater dissection and chamber mounting After euthanasia, fur, mandibles, and muscles were removed, and the head was kept in ice cold SIF until further use. A cranial window was made on the dorsal part of the head from the lambda line to the bregma line while leaving the dura mater untouched. Subsequently, the left parietal bone was removed to the level of the zygomatic arch. An intact piece of dura mater was excised and mounted on the chamber insert with preserved orientation. A rubber ring (Bauhaus, Denmark) was inserted to sandwich the tissue to prevent leakage. The chamber slides were inserted into the Ussing Chamber system, and 4 mL SIF at + 37 °C was gently added simultaneously to each chamber side to avoid dislodging the tissue. The Ussing Chamber setup was based on previous work [39], and a detailed description of the system is provided in the Supplementary Methods. Electrophysiological measurements and sampling Before tissue insertion, recordings were performed for 4 min without a current pulse and for 4 min with a current pulse of − 5 µA applied for 1 s with 3 s intervals in an open circuit system. After tissue insertion, current pulses were applied for at least 20 min to stabilise the tissue. After equilibration, 1 mL hαCGRP (5 ng/mL) from the ELISA kit (Bertin, France) was added to the 4 mL SIF on one chamber side to give a final concentration of 1 ng/mL in the chamber. The hαCGRP from the ELISA kit was selected to ensure accurate and optimal antibody recognition even at low concentrations. In the permeability experiments, hαCGRP was applied to the skull facing side in three preparations and to the brain facing side in three dura-preparations (n = 6 in total). The side to which hαCGRP was added is referred to as the donor side, and the opposite compartment as the acceptor side. In separate fragment experiments, hαCGRP1−17, hαCGRP1−24, or hαCGRP11−37 was applied to the skull facing side at 5 µM (n = 3 per fragment). Compounds remained in the chamber for approximately 60 min. Transdural voltage deflections were recorded throughout the experiment and visualised in real time using LabChart 8 (ADInstruments, UK). Transdural resistance was calculated using Ohm’s law (R = V/I) and expressed as Ω·cm², similar to transepithelial electrical resistance. For peptide quantification, 200 µL was collected from each chamber side at baseline and at 0 min, 15 min, 30 min, 45 min and 60 min after hαCGRP application. Samples were added to 50 µL enzyme immunoassay buffer in low retention microcentrifuge tubes (3451; Thermo Scientific, USA), quickly frozen, and stored at − 80 °C until analysed with ELISA. Hemi-skull preparation Following euthanasia and fur, mandible and muscle removal, heads were divided along the midsagittal plane and the brain halves were removed, leaving one hemi-skull for stimulation and the contralateral hemi-skull for control conditions, as previously described in an established CGRP release model [40]. The hemi-skulls were kept at room temperature, filled with 300 µL SIF. Subsequently, the hemi-skull was washed four times for 10 min each at + 37 °C. After the final wash, 300 µL prewarmed SIF was added and left for 15 min at + 37 °C as baseline. All hemi-skull based experiments described below were initiated directly after this final equilibration step using the same preparation. CGRP degradation experiment for proteomics To examine degradation of CGRP in the hemi-skull preparation (n = 4), hemi-skulls were prepared and equilibrated as described above. 300 µL of SIF was added to the hemi-skulls which included 1 µM capsaicin (to ensure that the degradation was measured under release-like conditions) and 1 µM hαCGRP (to ensure detectable amount of fragments), for 21 min corresponding to approximately 3 half-lives, based on the rapid degradation of CGRP in plasma described by Kraenzlin et al. [27]. Afterwards, 200 µL was transferred from the hemi-skulls to Eppendorf tubes and stored at − 80 °C until LC-MS proteomic analysis at Biogenity (Denmark) following an established procedure [41]. LC-MS proteomics MS sample preparation Samples were processed in two ways, first by maintaining the disulfide bridge (Cystein-Cystein) and second by reducing the disulfide bonds. The reduced condition (referred to as cystB) enables detection of fragments containing the cysteine residues involved in the disulfide bridge, whereas the intact disulfide bridge condition (referred to as cystI) does not permit detection of fragments containing the disulfide bridge, and hence higher sensitivity for other fragments. In the reduction condition the samples were centrifuged at 12,800 g for 10 min and diluted 1:1 with lysis buffer. Samples were boiled at + 95 °C for 5 min and subsequently sonicated on high for five cycles of 60 s using a Bioruptor sonication water bath (Diagenode, Belgium) at + 4 °C to lyse the samples. After this step the samples from both preparation strategies were processed identically. Samples were acidified with hydrochloric acid to pH 2, corresponding to a final HCl concentration of 40 mM and centrifuged at 12,800 g for 20 min at + 4 °C. Samples were desalted using SOLAµ solid phase extraction plates (Thermo Fisher Scientific, Germany). Between each application solvents were passed through by centrifugation at 1500 rpm. SOLAµ plate filters were activated with 200 µL methanol followed by 200 µL 80% acetonitrile containing 0.1% formic acid and equilibrated twice with 200 µL 1% trifluoroacetic acid and 3% acetonitrile before sample loading. After washing twice with 200 µL 0.1% formic acid peptides were eluted using 40% acetonitrile containing 0.1% formic acid. Eluted peptides were concentrated using an Eppendorf SpeedVac and reconstituted with internal retention time peptides (iRT, Biognosys, Switzerland) in 12 µL A*IRT buffer consisting of 2% acetonitrile and 1% trifluoroacetic acid. For each sample, the peptides were eluted into clean 1.5 mL Eppendorf tubes. MS analysis Peptides were loaded onto a 2 cm C18 trap column (Cat# 164946; Thermo Fisher Scientific, Germany) connected in line to a 15 cm C18 reverse phase analytical column (EasySpray ES904; Thermo Fisher Scientific, Germany) using the Thermo EasyLC 1200 HPLC system and loaded using 100% Buffer A (0.1% formic acid in water) at 750 bar with the column oven operating at + 30 °C. Peptides were eluted over a 70 min gradient from 10% to 60% acetonitrile containing 0.1% formic acid at 250 nL per min. The Orbitrap Exploris 480 mass spectrometer was operated in data dependent acquisition mode with a FAIMS Pro interface cycling between compensation voltages of − 50 V and − 70 V every 28 scans. Full MS spectra were acquired at a resolution of 60,000 with an automatic gain control target of 300% and a scan range of 375 to 1500 m/z. Maximum injection time was set to auto. Precursor ions with intensity greater than 5 × 10³ and charge states between 2 and 6 were selected for MS2 analysis. Dynamic exclusion was set to 60 s and precursor ions were isolated using a 1.6 m/z window. Fragmentation was performed using high energy collision dissociation with normalised collision energy of 30%. MS2 spectra were acquired at a resolution of 15,000. The exclusion list was shared between compensation voltage values, Advanced Peak Determination was enabled, ions were collected to a normalised automatic gain control target of 75%, and maximum injection time for MS2 was set to auto. MS performance was verified for consistency by running complex cell lysate quality control standards, and chromatography was monitored to check reproducibility. Proteomic data analysis Biogenity contributed by performing the data analysis, including figures, tables, and reports. The raw files were analysed using MaxQuant 2.4.2.0 [42]. Spectra were matched against Rattus norvegicus database from UniProt with human CALCA (P06881) added to the search. Enzyme was set to unspecific. Variable modifications were set as oxidation (M) and acetyl on protein N-termini. Cysteine carbamidomethyl was set as a static modification. The minimum peptide length was set to 7 and the longest peptide length was set to 40. All results were filtered to a 1% false discovery rate at the peptide spectrum match level. Mast cell degranulation Mast cell degranulation was assessed using the hemi-skull preparation described above (n = 20). After the washing and baseline equilibration steps, hemi-skulls were incubated for 15 min at + 37 °C with 10 µM hαCGRP (n = 5), 10 µM hαCGRP1−17 (n = 3), 10 µM hαCGRP1−24 (n = 3), 10 µM hαCGRP11−37 (n = 4), or 20 µg/mL C48/80 (positive control; n = 5). The 10 µM concentration was selected to ensure a detectable and potentially maximal response if present. For each stimulation, a paired control was exposed to vehicle (SIF or SIF containing 1% DMSO), with DMSO included to ensure solubility of the peptide fragments. The incubation period was chosen based on the CGRP release model [40] combined with the timing used in the CGRP degradation setup. The reaction was terminated, and the dura mater fixed by exposure to 4% paraformaldehyde for 20 min. Mast cells were visualised by incubation with 0.1% acidified toluidine blue. The dura mater was dissected from the skull and whole mounted on glass slides with a coverslip. Images were acquired at 200× magnification using a Nikon Eclipse Ni microscope with Nikon DS Fi2 camera and Nikon digital sight DS U3. Degranulated cells were counted in a blinded manner in five consecutive fields along the stem of the middle meningeal artery combined with one peripheral field. Mast cells were classified as degranulated if more than eight extruded granules were localised in the periphery of the cell, or if an extensive loss of staining intensity gave the cell a ghostly appearance. Results are presented as weighted percentile within each group of each experimental run. CGRP release experiment The hemi-skull release model was used to test the effect of hαCGRP13−34 on basal and capsaicin evoked αCGRP release (n = 6). After the baseline incubation described above, hemi-skulls were stimulated for 15 min at + 37 °C with 300 µL prewarmed oxygenated SIF either alone as control or containing 10 µM hαCGRP13−34. The same hemi-skulls were subsequently incubated for a further 15 min with 300 µL prewarmed oxygenated SIF containing either 1 µM capsaicin alone or 1 µM capsaicin together with 10 µM hαCGRP13−34. After baseline and each stimulation step, 200 µL was collected and transferred to tubes containing 50 µL enzyme immunoassay buffer and stored at − 80 °C until ELISA analysis. Wire myography Artery preparation Mesenteric arteries were isolated and transferred to ice cold Krebs buffer continuously gassed with 95% O2 and 5% CO2 (n = 14). Second order mesenteric arterial branches were dissected under a stereomicroscope and cleared of surrounding connective tissue. Segments of 1.0 to 1.6 mm were mounted on two 40 μm steel wires in a small vessel wire myograph system containing Krebs buffer maintained at + 37 °C and continuously aerated. After mounting vessels were equilibrated for 30 min and normalised to determine the internal circumference corresponding to a transmural pressure of 100 mmHg. Arteries were set to 0.9 × L100 and allowed to equilibrate for an additional 30 min. Contractility was verified by two exposures to high potassium Krebs solution containing 60 mM K+. A force-increase of at least 2 mN was used as the inclusion criterion, and isometric tension was continuously recorded using a PowerLab data acquisition system (ADInstruments). Ex vivo pharmacology Arteries were precontracted with 200 nM U46619 (thromboxane A2 analogue) until stable tone was reached. Individual arterial segments were exposed to fixed concentrations of 10 µM hαCGRP1−17, hαCGRP1−24, hαCGRP13−25, hαCGRP11−37, hαCGRP11−24, hαCGRP22−37, hαCGRP26−34, hαCGRP13−34 or vehicle control. The final DMSO concentration was 1% in all conditions. Concentration response curves were generated by cumulative addition of hαCGRP until maximal relaxation was obtained. In parallel, a control hαCGRP concentration response curve was performed. The U46619 induced contractile tone was defined as 0% relaxation, while the resting baseline tension was defined as 100% relaxation. For Schild regression analysis of hαCGRP8−37 and hαCGRP11−37, arterial segments were taken from the same artery, where one segment served as an hαCGRP concentration response curve and parallel segments were incubated with a fixed concentration of antagonist. In vitro receptor pharmacology Cells and transfection Cos-7 cells were obtained in 2003 from Nigel Birch (University of Auckland), having originally been acquired from ATCC (Manassas, VA, USA). Cos-7 cells are a fibroblast-like cell line with low endogenous CGRP receptor expression, allowing receptor-mediated responses to be attributed to the transfected CLR/RAMP1 complex with minimal background signalling [43]. As detailed in previous publications [43–46], these were cultured in complete media of DMEM (high glucose, sodium pyruvate and L-glutamine) supplemented with 8% heat inactivated fetal bovine serum. The cells were maintained at + 37 °C in a humidified incubator supplemented with 5% CO2. Cells used were passaged a maximum of 30 times. They were sub-cultured when reaching 90% confluence and split 1:5 to be ready around 72 h later. Cells were reseeded into 96-well SpectraPlates (6005658, Revvity, MA. USA) at 15,000 to 20,000 cells per well one day before transfection. Cells were transiently transfected using polyethyleneimine (transfection agent) mixed with receptor DNA in a 1:1 ratio. Constructs encoded myc tagged RAMP1 and HA tagged CLR of either rat or human origin [43–46]. The transfection mix was made a minimum of 10 min prior to transfection and left at room temperature. It consisted of 5% filtered sterilised glucose, 25 µg DNA per well and polyethyleneimine at a ratio of 10 to 1 (to total DNA). The transfection mix was diluted 1:9 in complete media. Following removal of the original growth medium from the Cos-7 cells, 100 µL was added to each well, and the plate was returned to the incubator. The use of both rat and human in vitro systems introduces a translational element, where inclusion of human CGRP receptor and hαCGRP data supports direct relevance to human biology. cAMP assay 48 h after transfection, the cAMP assay was initiated as described by Nimick et al. (2025). Cells were incubated in serum free medium containing 0.1% bovine serum albumin and 1 mM IBMX for 30 min at + 37 °C (final DMSO concentration of 0.2%). The medium was replaced with stimulation medium containing either hαCGRP, hαCGRP1−17, hαCGRP1−24, or hαCGRP11−37 acting as agonists (dose-response) or full-length hαCGRP combined with either of the fragments acting as antagonists at 10 µM. In antagonist mode, the full-length hαCGRP was added immediately after the fragments. Forskolin (50 µM) was added to two of the wells per 96-well plate as positive control. Cells were stimulated for 15 min at + 37 °C and reactions terminated with ice cold absolute ethanol and left at -30 °C over night. After complete evaporation of ethanol, 50 µL lysis buffer (containing cAMP detection buffer (50 mM HEPES, 10 mM CaCl) and 0.35% Triton X-100, pH 7.4) was added to each well and left shaking for 15 min at room temperature to resuspend cell extracts and initiate the cAMP detection. Then, from each well 10 µL lysate was transferred to a 384-well plate. A cAMP GS kit (Cat# 62AM4PEJ, Revvity, MA, USA) was used to quantify cAMP levels. In brief, 5 µL d2-labelled cAMP and 5 µL anti-cAMP Cryptate (anti-cAMP antibody linked to a donor fluorophore) were added to each well according to the manufacturer’s instructions. The assay is based on a competitive principle, where binding of d2-labelled cAMP to the anti-cAMP Cryptate brings the d2 acceptor and cryptate donor into proximity, generating a FRET signal. Endogenously produced cAMP, following CGRP receptor activation, competes with the d2-labelled cAMP for binding to the antibody, thereby reducing FRET signalling inversely proportional to intracellular cAMP levels. The 384-well plate was covered with a plate sealer (100sealplt, Excel Scientific Inc, Victorville, CA, USA) and pulse centrifuged for 15 s (400 x g) before a 4 h incubation at room temperature in the dark. The sealer was removed, and cAMP detection was performed using a ClarioStar plate reader (BMG LabTech, Germany) with excitation at 340 nm and emission wavelengths of 665 nm and 620 nm. The concentration of cAMP was determined using the standard curve made specifically for each experiment. For each experiment, a minimum of three technical replicates and three biological replicates was included. ELISA Throughout all ELISA experiments, a previously validated enzyme immunoassay kit (human CGRP ELISA kit, detection limit 0.7 pg/mL; A05481, Bertin, France) was used (Garelja et al. (2025) [20]) and CGRP was quantified according to the manufacturer protocol. CGRP concentrations were calculated from the standard curve. To determine whether the hαCGRP fragments were detected by the ELISA kit, fragments hαCGRP1−17, hαCGRP1−24, hαCGRP11−37, hαCGRP13−25, hαCGRP11−24, hαCGRP22−37 and hαCGRP26−34 were tested at a concentration equivalent to 10,000 pg/mL hαCGRP. To test potential interference with detection, selected fragments (hαCGRP1−17, hαCGRP1−24, hαCGRP11−37 and hαCGRP13−34) were combined with the standard hαCGRP calibration curve in the presence of 10 µM fragment. Concentration dependent interference was examined by adding increasing concentrations of hαCGRP11−37 or hαCGRP13−34 ranging from 1 nM to 10 µM to samples containing 125 pg/mL hαCGRP (chosen as it is the middle concentration of the standard curve) or to samples without hαCGRP. Optical density was measured at 410 nm using a micro plate photometer (Tecan Infinite M200, software SW Magellan v6.3, Switzerland) similar to previous papers [16, 40]. For hαCGRP permeability over the dura mater, the ELISA results are shown as CGRP concentration over time for both the side receiving hαCGRP and the opposite side. In investigating fragment detection by the ELISA kit, the results are shown as concentration detection relative to CGRP detection. For hαCGRP competition in the kit, the data are shown as CGRP concentration as a function of the standard fold change. Lastly, αCGRP release studies using capsaicin and/or hαCGRP13−34 are shown for the different conditions as concentration of CGRP detected. Data analysis Statistical analyses were performed using GraphPad Prism 10.4.1. hαCGRP concentration response curves were analysed by four parameter non-linear regression to determine EC50. For Schild regression analysis in wire myography experiments log [DR − 1] was plotted against log antagonist concentration. The DR (dose ratio) was defined as the EC50 in the presence of antagonist divided by the EC₅₀ in the absence of antagonist. pA₂ values were estimated from the x-intercept obtained by linear regression analysis, and are reported with 95% confidence intervals. One rat represented one experimental unit (n = 1). Where receptor antagonism was observed in vitro, the pA2 value was derived from a global Gaddum/Schild analysis in Prism, with the Schild slope constrained to one. pA2 refers to the logarithmic concentration of antagonist capable of producing a twofold shift in agonist EC50. Where outliers were removed, this is stated in the Results section. Statistical testing included two tailed unpaired t-test, one-way ANOVA with Tukey multiple comparisons test, one-way ANOVA with Dunnett multiple comparisons test, or two-way ANOVA with Sidak multiple comparisons test, as indicated in the Results section. Data are shown as mean ± SEM and results were considered significant at p < 0.05. Results Reduced transdural resistance and rapid degradation of hαCGRP in the dura mater hαCGRP was examined in the Ussing chamber set-up to assess permeability across dura mater and changes in transdural electrical resistance (TDER) across the tissue. The experimental workflow from dura mater isolation to mounting in the Ussing chamber with continuous recording of transdural voltage deflection and sampling for hαCGRP quantification over time is presented in Fig. 1A. A representative trace of the transdural voltage recorded in real time from empty chamber conditions and approximately 1 h after hαCGRP addition (Fig. 1B). Fig. 1Ussing chamber analysis of hαCGRP permeability and electrical responses across the rat dura mater. (A) Schematic overview of the experimental workflow. The dura mater was isolated from the rat and mounted in an Ussing chamber separating the skull-facing and brain-facing compartments. Transdural voltage deflections were recorded using electrodes while samples were collected from both chamber sides for hαCGRP quantification by ELISA. (B) Representative trace of transdural voltage deflection recorded in real time. The trace shows recordings obtained from empty chamber conditions and after addition of hαCGRP to the mounted dura mater preparation. (C) Quantification of transdural electrical resistance (TDER) calculated from voltage deflections using Ohm’s law. Square symbols indicate experiments where hαCGRP was added to the skull-facing side of the dura mater and round symbols indicate addition to the brain-facing side. (D) Changes in transdural voltage relative to baseline following addition of hαCGRP to either side of the dura mater. (E) hαCGRP concentrations measured by ELISA on the donor and acceptor sides of the chamber over time following peptide application. Data are shown as mean ± SEM, and statistical significance is depicted with asterisk (Student’s t-test). *** represents p < 0.001, n = 6 Ussing chamber analysis of hαCGRP permeability and electrical responses across the rat dura mater. (A) Schematic overview of the experimental workflow. The dura mater was isolated from the rat and mounted in an Ussing chamber separating the skull-facing and brain-facing compartments. Transdural voltage deflections were recorded using electrodes while samples were collected from both chamber sides for hαCGRP quantification by ELISA. (B) Representative trace of transdural voltage deflection recorded in real time. The trace shows recordings obtained from empty chamber conditions and after addition of hαCGRP to the mounted dura mater preparation. (C) Quantification of transdural electrical resistance (TDER) calculated from voltage deflections using Ohm’s law. Square symbols indicate experiments where hαCGRP was added to the skull-facing side of the dura mater and round symbols indicate addition to the brain-facing side. (D) Changes in transdural voltage relative to baseline following addition of hαCGRP to either side of the dura mater. (E) hαCGRP concentrations measured by ELISA on the donor and acceptor sides of the chamber over time following peptide application. Data are shown as mean ± SEM, and statistical significance is depicted with asterisk (Student’s t-test). *** represents p < 0.001, n = 6 From the six different traces obtained, TDER was calculated. No statistically significant difference in TDER was observed between application of hαCGRP to the skull-facing or brain-facing side. Therefore, the data were pooled for comparison between baseline and hαCGRP conditions. After adjustment for the initial baseline recorded in a tissue-free chamber, TDER decreased significantly (Fig. 1C) following hαCGRP exposure (1.44 ± 0.47 Ω·cm2 vs. −6.56 ± 0.60 Ω·cm2; p < 0.001). In contrast to the TDER response, transdural voltage differed depending on the side of application, increasing when hαCGRP was added to the skull-facing side and decreasing when added to the brain-facing side (Fig. 1D). TDER responses to ATP and forskolin tested in the same system were also examined; however, none produced statistically significant changes (Supplementary Fig. 1A). Samples were collected from both chamber compartments every 15 min to estimate hαCGRP permeability across the dura mater (Fig. 1E). On the acceptor side, all measurements were below the detection limit of the ELISA kit, indicating that hαCGRP did not cross the dura mater. In contrast, hαCGRP concentrations on the donor side declined significantly during the first 15 min, decreasing from 769.64 ± 62.21 pg/mL to 305.31 ± 93.95 pg/mL (p < 0.01), with no significant difference between 15 and 60 min (Supplementary Fig. 1B). Degradation patterns of hαCGRP in the rat hemi-skull hαCGRP levels decreased on the donor side in the Ussing chamber without detectable diffusion across the barrier, indicating loss of full-length peptide. This observation is consistent with rapid local degradation near the dura mater. To examine degradation patterns in more detail, a hαCGRP degradation experiment followed by proteomic analysis was performed (Fig. 2A). Fig. 2hαCGRP degradation patterns in the rat hemi-skull. (A) Schematic overview of the experimental workflow used to examine hαCGRP degradation in the hemi-skull preparation. Following preparation and equilibration, hemi-skulls were incubated with capsaicin together with hαCGRP under stimulated conditions, after which samples were collected and analysed by LC-MS to identify degradation fragments. (B) Alignment of hαCGRP fragments selected for functional analysis to the full-length peptide sequence with arrows indicating predicted cleavage sites. (C) Distribution of identified hαCGRP fragments in samples (n = 3) processed with reduction of the disulfide bridge during sample preparation. (D) Distribution of identified hαCGRP fragments in samples (n = 4) processed with the disulfide bridge maintained intact during sample preparation hαCGRP degradation patterns in the rat hemi-skull. (A) Schematic overview of the experimental workflow used to examine hαCGRP degradation in the hemi-skull preparation. Following preparation and equilibration, hemi-skulls were incubated with capsaicin together with hαCGRP under stimulated conditions, after which samples were collected and analysed by LC-MS to identify degradation fragments. (B) Alignment of hαCGRP fragments selected for functional analysis to the full-length peptide sequence with arrows indicating predicted cleavage sites. (C) Distribution of identified hαCGRP fragments in samples (n = 3) processed with reduction of the disulfide bridge during sample preparation. (D) Distribution of identified hαCGRP fragments in samples (n = 4) processed with the disulfide bridge maintained intact during sample preparation In samples in which the disulfide bridge was broken during sample preparation (cystB; n = 3), 31 distinct hαCGRP fragments were identified (Supplementary Fig. 2A). In contrast, samples with an intact disulfide bridge (cystI; n = 4) yielded seven fragments (Supplementary Fig. 2B). Fragments selected for further functional analysis were aligned to the full-length hαCGRP sequence (Fig. 2B), and their distribution across the datasets is summarised in pie charts (Fig. 2C–D). In the cystB dataset, the dominant fragments corresponded to hαCGRP1−17 (61%) and hαCGRP1−24 (18%), followed by hαCGRP13−25 (2%), while hαCGRP26−34 and hαCGRP22−37 were detected at < 1% each. The remaining 26 fragments together represented approximately 19% of the total fragment counts. In the cystI dataset, the most abundant fragments were hαCGRP13−24 (27%), hαCGRP11−37 (20%), hαCGRP22−37 (17%), and hαCGRP26−34 (17%), while the remaining three fragments together accounted for approximately 19% of the total fragment counts. These fragments were selected based on retrieving the highest number of counts and maximising detection across samples. In one of the samples, full-length hαCGRP was still detectable; however, the full-length peptide was excluded from the subsequent analysis, as the aim was to identify degradation fragments and inclusion of intact hαCGRP would confound interpretation of the degradation profile. Effects of hαCGRP degradation fragments on hαCGRP induced relaxation of mesenteric arteries After synthesis by TAGCopenhagen (Denmark) of the seven most abundant hαCGRP fragments identified by proteomic analysis, potential retained biological activity was examined. Because CGRP is a potent vasodilator, the initial analysis focused on vascular responses in isolated mesenteric arteries using wire myography. The experimental setup is illustrated (Fig. 3A). None of the hαCGRP degradation fragments elicited a vasodilatory response when applied at 10 µM to pre-contracted arteries (Supplementary Fig. 3). Hence, the fragments were tested for antagonistic abilities against hαCGRP induced vasorelaxation. Fig. 3Effects of hαCGRP degradation fragments on hαCGRP-induced relaxation of mesenteric arteries. (A) Experimental setup for assessment of vascular responses using wire myography. Mesenteric artery segments were mounted for isometric tension recordings and precontracted before pharmacological stimulation. 10 µM hαCGRP1−17 (B), 10 µM hαCGRP1−24 (C), 10 µM hαCGRP13−25 (D), 10 µM hαCGRP11−37 (E), 10 µM hαCGRP11−24 (F), 10 µM hαCGRP22−37 (G), and 10 µM hαCGRP26−34 (H) were added to precontracted mesenteric arteries to test direct vasoactive effects (see Supplementary Fig. 3). Control conditions refer to hαCGRP concentration response curves generated in the absence of degradation fragments. After this incubation, cumulative concentration response curves to hαCGRP were generated (n = 6). Data points are represented as mean ± SEM. Curve fitting was performed using log(agonist) versus response, variable slope (four parameters) Effects of hαCGRP degradation fragments on hαCGRP-induced relaxation of mesenteric arteries. (A) Experimental setup for assessment of vascular responses using wire myography. Mesenteric artery segments were mounted for isometric tension recordings and precontracted before pharmacological stimulation. 10 µM hαCGRP1−17 (B), 10 µM hαCGRP1−24 (C), 10 µM hαCGRP13−25 (D), 10 µM hαCGRP11−37 (E), 10 µM hαCGRP11−24 (F), 10 µM hαCGRP22−37 (G), and 10 µM hαCGRP26−34 (H) were added to precontracted mesenteric arteries to test direct vasoactive effects (see Supplementary Fig. 3). Control conditions refer to hαCGRP concentration response curves generated in the absence of degradation fragments. After this incubation, cumulative concentration response curves to hαCGRP were generated (n = 6). Data points are represented as mean ± SEM. Curve fitting was performed using log(agonist) versus response, variable slope (four parameters) Following fragment incubation, cumulative concentration response curves to hαCGRP were generated (Fig. 3B-H). hαCGRP induced relaxation of mesenteric arteries was not affected by co-incubation with 10 µM hαCGRP derived fragments hαCGRP1−17, hαCGRP1−24, hαCGRP13−25, hαCGRP11−24, hαCGRP22−37, or hαCGRP26−34. However, 10 µM hαCGRP11−37 induced a rightward shift of the hαCGRP concentration response curve while maximal relaxation remained preserved (Fig. 3E). Competitive antagonism of the CGRP receptor by hαCGRP11−37 compared with hαCGRP8−37 Fig. 4Competitive antagonism of hαCGRP₁₁₋₃₇ compared with hαCGRP₈₋₃₇ in mesenteric arteries. (A) hαCGRP concentration response curves obtained in the presence of increasing concentrations of hαCGRP8−37. (B) hαCGRP concentration response curves obtained in the presence of increasing concentrations of hαCGRP11−37. (C) Schild regression analysis comparing antagonist potency of hαCGRP8−37 and hαCGRP11−37. Data points are represented as mean ± SEM, n = 5–6. Curve fitting was performed using log(agonist) versus response, variable slope (four parameters) Competitive antagonism of hαCGRP₁₁₋₃₇ compared with hαCGRP₈₋₃₇ in mesenteric arteries. (A) hαCGRP concentration response curves obtained in the presence of increasing concentrations of hαCGRP8−37. (B) hαCGRP concentration response curves obtained in the presence of increasing concentrations of hαCGRP11−37. (C) Schild regression analysis comparing antagonist potency of hαCGRP8−37 and hαCGRP11−37. Data points are represented as mean ± SEM, n = 5–6. Curve fitting was performed using log(agonist) versus response, variable slope (four parameters) Antagonism by hαCGRP11−37 was further characterised using Schild regression analysis and compared with hαCGRP8−37, which is the most characterised peptide antagonist of the CGRP receptor and widely used in in vitro and ex vivo studies [47–49]. Consistent with competitive antagonism, both hαCGRP11−37 and hαCGRP8−37 produced progressive rightward shifts of the hαCGRP concentration response curves without affecting maximal relaxation (Fig. 4A–B), and Schild plot slopes did not differ significantly from 1 (hαCGRP8−37: slope = 0.85, p = 0.168; hαCGRP11−37: slope = 0.77, p = 0.121; one sample t-test). The pA2 values, derived from the Schild plot, were 7.80 (95% CI 7.61 to 8.11) for hαCGRP8−37 and 6.66 (95% CI 6.29 to 7.52) for hαCGRP11−37, indicating an approximately tenfold lower antagonist potency of hαCGRP11−37, with the non-overlapping confidence intervals supporting a significant difference between the antagonists (Fig. 4C). hαCGRP fragments on cAMP generation in CGRP receptor transfected Cos-7 cells The two most abundant metabolites identified, hαCGRP1−17 and hαCGRP1−24, together with hαCGRP11−37, were further examined in agonist and antagonist mode in Cos-7 cells expressing either the human or rat CGRP receptor. The experimental setup for the human and rat receptor assays is illustrated (Fig. 5A and B). Fig. 5Effects of hαCGRP fragments on cAMP generation in human and rat CGRP receptor transfected Cos-7 cells. Schematic overview of the experimental setup used to assess agonist and antagonist effects at the (A) human CGRP receptor in Cos-7 cells transiently transfected with human CLR and human RAMP1 and at the (B) rat CGRP receptor in Cos-7 cells transiently transfected with rat CLR and rat RAMP1. In agonist mode, cells were stimulated with hαCGRP1−17, hαCGRP1−24, or hαCGRP11−37 alone. In antagonist mode, cells were stimulated with hαCGRP together with fragment, and cAMP accumulation was measured. Concentration response curves for hαCGRP, hαCGRP1−17, hαCGRP1−24, and hαCGRP11−37 in agonist mode at the (C) human CGRP receptor and (D) rat CGRP receptor. (E) hαCGRP concentration response curves in the absence or presence of 10 µM hαCGRP1−17, hαCGRP1−24, or hαCGRP11−37 in antagonist mode at the (E) human CGRP receptor or (F) rat CGRP receptor. Data points are represented as mean ± SEM, n = 3 in triplicate Effects of hαCGRP fragments on cAMP generation in human and rat CGRP receptor transfected Cos-7 cells. Schematic overview of the experimental setup used to assess agonist and antagonist effects at the (A) human CGRP receptor in Cos-7 cells transiently transfected with human CLR and human RAMP1 and at the (B) rat CGRP receptor in Cos-7 cells transiently transfected with rat CLR and rat RAMP1. In agonist mode, cells were stimulated with hαCGRP1−17, hαCGRP1−24, or hαCGRP11−37 alone. In antagonist mode, cells were stimulated with hαCGRP together with fragment, and cAMP accumulation was measured. Concentration response curves for hαCGRP, hαCGRP1−17, hαCGRP1−24, and hαCGRP11−37 in agonist mode at the (C) human CGRP receptor and (D) rat CGRP receptor. (E) hαCGRP concentration response curves in the absence or presence of 10 µM hαCGRP1−17, hαCGRP1−24, or hαCGRP11−37 in antagonist mode at the (E) human CGRP receptor or (F) rat CGRP receptor. Data points are represented as mean ± SEM, n = 3 in triplicate Full-length hαCGRP produced robust agonist responses at both the rat CGRP receptor (pEC50 = 10.29 ± 0.09) and the human CGRP receptor (pEC50 = 9.50 ± 0.21), consistent with previous reports [50, 51]. In contrast, none of the fragments induced a notable increase in cAMP production across the concentration range from 10 pM to 10 µM in agonist mode at neither the rat CGRP receptor nor the human CGRP receptor (Fig. 5C and D; n = 3). In antagonist mode, 10 µM hαCGRP1−17 and hαCGRP1−24 did not affect full-length hαCGRP induced cAMP generation at either receptor (Fig. 5E and F; n = 3). In contrast, 10 µM hαCGRP11−37 produced a 281-fold increase in EC50 at the rat CGRP receptor and a 380-fold increase in EC50 at the human CGRP receptor. At the rat receptor, hαCGRP11−37 produced a pA2 value of 7.55, with a shift in pEC50 from 10.04 ± 0.17 to 6.89 ± 0.28. At the human receptor, hαCGRP11−37 produced a pA2 value of 7.62, with a shift in pEC50 from 9.26 ± 0.22 to 6.80 ± 0.10. Similar to the vascular experiments (Fig. 4), the antagonistic effect of hαCGRP11−37 was compared with a Gaddum-Schild plot made for hαCGRP8−37 (Supplementary Fig. 4). At both the human and rat CGRP receptor, a similar approximately ten-fold lower potency was observed for hαCGRP11−37 compared to hαCGRP8−37. Antagonism by hαCGRP8−37 resulted in pA2 values of 8.32 at the rat receptor (n = 4) and 8.95 at the human receptor (n = 5). The comparable antagonist profile observed in the cell-based assay and in the mesenteric artery myograph experiments supports the functional interpretation that hαCGRP11−37 acts as a weak competitive antagonist at the CGRP receptor and provides cross validation between the receptor system for human and rat supported by the data from the ex vivo vascular model. Effects of hαCGRP fragments on dura mater barrier properties and mast cell degranulation hαCGRP1−17, hαCGRP1−24, and hαCGRP11−37 were further examined in the dura mater using the Ussing chamber preparation. None of the fragments altered transdural resistance compared with baseline conditions (Supplementary Fig. 5A). Fig. 6Effects of hαCGRP and hαCGRP fragments on mast cell degranulation in the rat dura mater. (A) Schematic overview of the experimental workflow used to assess mast cell degranulation in the rat dura mater hemi-skull preparation. After hemi-skull preparation and equilibration, tissues were incubated with hαCGRP (n = 5), hαCGRP1−17 (n = 3), hαCGRP1−24 (n = 3), hαCGRP11−37 (n = 4), or compound 48/80 (C48/80; n = 5) as positive control. Dura mater tissues were subsequently fixed and stained with toluidine blue, and mast cell degranulation was quantified by blinded counting of cells along the middle meningeal artery and peripheral fields. Two control conditions (n = 3–7) were included due to differences in vehicle solvent composition. (B) Quantification of mast cell degranulation expressed as percentage of degranulated mast cells for each condition. Data are shown as mean ± SEM, and asterisk indicates statistical significance (One-way ANOVA with Dunnett’s multiple comparisons test, with a single pooled variance). ** represents p < 0.01 Effects of hαCGRP and hαCGRP fragments on mast cell degranulation in the rat dura mater. (A) Schematic overview of the experimental workflow used to assess mast cell degranulation in the rat dura mater hemi-skull preparation. After hemi-skull preparation and equilibration, tissues were incubated with hαCGRP (n = 5), hαCGRP1−17 (n = 3), hαCGRP1−24 (n = 3), hαCGRP11−37 (n = 4), or compound 48/80 (C48/80; n = 5) as positive control. Dura mater tissues were subsequently fixed and stained with toluidine blue, and mast cell degranulation was quantified by blinded counting of cells along the middle meningeal artery and peripheral fields. Two control conditions (n = 3–7) were included due to differences in vehicle solvent composition. (B) Quantification of mast cell degranulation expressed as percentage of degranulated mast cells for each condition. Data are shown as mean ± SEM, and asterisk indicates statistical significance (One-way ANOVA with Dunnett’s multiple comparisons test, with a single pooled variance). ** represents p < 0.01 Although none of the peptide fragments carried a strong positive charge (Supplementary Fig. 5B), which typically is needed for mast cells degranulation, they were tested since CGRP-related signalling and peptide degradation occur in the dural environment where mast cells reside (Fig. 6A). Two control conditions (Fig. 6B) were included due to differences in vehicle solvents, but no statistical difference was observed between these controls (1% DMSO in SIF: 29.9% ± 7.1% vs. SIF: 23.0% ± 2.6%; p = 0.95). The positive control (C48/80) significantly increased the number of degranulated mast cells compared with the vehicle (SIF) conditions (54.2% ± 7.2%; p < 0.01). In contrast, neither full-length hαCGRP nor any of the tested fragments produced a statistically significant increase in mast cell degranulation compared with the vehicle (DMSO) control conditions (Fig. 6B). Degranulation levels were 32.1% ± 6.1% for hαCGRP (p = 0.7953), 25.9% ± 4.2% for hαCGRP1−17 (p > 0.999), 25.3% ± 4.4% for hαCGRP1−24 (p > 0.998), and 33.9% ± 10.3% for hαCGRP11−37 (p > 0.999). Interaction of hαCGRP degradation fragments with the CGRP ELISA Even though no direct association was observed between the degradation fragments identified and the migraine relevant tissues examined, we wanted to investigate the possibility that these fragments could interfere with or be detected in CGRP biomarker assays. Interaction of the fragments with a commonly used, and well validated CGRP ELISA assay was therefore investigated [20]. The specific binding of the ELISA kit is unknown, but we hypothesised that the N-terminal antibody is specific to some of the amino acids in the 1–8 region since the kit does not recognise hαCGRP8−37. Based on this, only the N-terminal (see full-length hαCGRP, Fig. 7A) conserved degradation fragments might interfere with the signal in the assay. Fig. 7Detection and interference of hαCGRP degradation fragments with CGRP ELISA. (A) Schematic representation of the full-length hαCGRP peptide [13] with the disulfide bond. (B) Apparent CGRP signal measured when individual fragments were analysed alone in the ELISA assay. Fragments were tested at a to tested at 2.6 nM n = 3 of each, which corresponds to 10,000 pg/mL full-length hαCGRP based on molar equivalence. Values are compared with non-specific binding (NSB; n = 6) and with the signal obtained for 125 pg/mL hαCGRP measured using the same standard curve (n = 6). (C) hαCGRP standard curve measured in the absence or presence of degradation fragments. The curves were generated in the presence of hαCGRP1−17, hαCGRP1−24, or hαCGRP11−37, each at a concentration of 2.6 nM, corresponding to 10× the maximal CGRP concentration of the ELISA standard range. (D) Competition experiment assessing interference with CGRP detection. Increasing concentrations of hαCGRP11−37 were added to samples containing 125 pg/mL hαCGRP, and the apparent CGRP signal detected by the ELISA assay was measured. Data are depicted as mean ± SEM Detection and interference of hαCGRP degradation fragments with CGRP ELISA. (A) Schematic representation of the full-length hαCGRP peptide [13] with the disulfide bond. (B) Apparent CGRP signal measured when individual fragments were analysed alone in the ELISA assay. Fragments were tested at a to tested at 2.6 nM n = 3 of each, which corresponds to 10,000 pg/mL full-length hαCGRP based on molar equivalence. Values are compared with non-specific binding (NSB; n = 6) and with the signal obtained for 125 pg/mL hαCGRP measured using the same standard curve (n = 6). (C) hαCGRP standard curve measured in the absence or presence of degradation fragments. The curves were generated in the presence of hαCGRP1−17, hαCGRP1−24, or hαCGRP11−37, each at a concentration of 2.6 nM, corresponding to 10× the maximal CGRP concentration of the ELISA standard range. (D) Competition experiment assessing interference with CGRP detection. Increasing concentrations of hαCGRP11−37 were added to samples containing 125 pg/mL hαCGRP, and the apparent CGRP signal detected by the ELISA assay was measured. Data are depicted as mean ± SEM To test this directly, each fragment was analysed individually in the ELISA assay (Fig. 7B). Fragments were tested at a CGRP equivalent of 2.6 nM, which is 10 times the max detection of the kit. None of the fragments produced a detectable signal above the level of non-specific binding. In contrast, 125 pg/mL hαCGRP produced a signal of 101.5 ± 7.5 pg/mL using the same standard curve, which lies within the acceptable detection range of the assay and differed significantly from all fragment conditions. The potential for fragments to interfere with detection of intact hαCGRP was then examined using the same 2.6 nM concentration. Standard curves were generated in the presence of hαCGRP1−17, hαCGRP1−24, or hαCGRP11−37 (Fig. 7C). hαCGRP1−17 and hαCGRP1−24 produced a rightward shift of the CGRP standard curve, indicating reduced apparent detection of hαCGRP, whereas hαCGRP11−37 did not measurably alter the standard curve under these conditions. However, when increasing concentrations of hαCGRP11−37 were added to samples containing 125 pg/mL hαCGRP, the detected CGRP signal progressively decreased (Fig. 7D). Half maximal reduction of the measured CGRP signal occurred at an IC50 of 25.1 nM (95% CI 11.8–53.9 nM). These observations indicate that fragments containing the N-terminal antibody recognition region interfere with ELISA detection at low nanomolar concentrations, whereas fragments containing only C-terminal epitopes require substantially higher concentrations to affect CGRP detection. This suggests that N-terminal degradation fragments represent the primary source of potential interference in CGRP ELISA measurements using this specific kit. Competition with enzymatic degradation using a designed hαCGRP13−34 decoy fragment Previous studies have shown that neuropeptides can compete for enzymatic degradation, for example with substance P competing with CGRP during peptidase mediated breakdown near sensory tissues [28, 29]. This raises the possibility that peptide fragments may influence CGRP stability by competing for enzyme binding sites during degradation. Based on the degradation pattern identified in the hemi-skull proteomic analysis, a new peptide fragment was designed to include residues surrounding the major cleavage sites S17/R18 and K24/N25 while maintaining neighbouring amino acids predicted to compete with enzymatic degradation. This design resulted in the fragment hαCGRP13−34, shown on the top of Fig. 8. Fig. 8Functional assessment of the designed hαCGRP13−34 decoy fragment. The sequence of the designed hαCGRP13−34 fragment is shown above the panels. (A) Vasomotor assessment of hαCGRP13−34 in isolated mesenteric arteries using wire myography (n = 8). Arteries were precontracted and cumulative concentration-response curves to hαCGRP were generated in the absence or presence of 10 µM hαCGRP13−34. (B) Evaluation of potential interference with CGRP ELISA detection. Concentration-response measurements of hαCGRP13−34 alone and in the presence of 125 pg/mL hαCGRP were performed to assess whether the fragment affects assay detection (n = 3 of each). (C) Hemi-skull CGRP release experiment assessing the effect of 10 µM hαCGRP13−34 on basal and capsaicin-evoked CGRP release measured by ELISA (n = 6). (D) CGRP ELISA competition curve showing detection of hαCGRP in the absence or presence of 10 µM hαCGRP13−34 (n = 3). Data are shown as mean ± SEM. Statistical comparisons were made with two-way ANOVA with “Sidak” multiple comparison’s test (A) or one-way ANOVA with Dunnett’s multiple comparison’s test (C) (* p < 0.05, ***p < 0.001) Functional assessment of the designed hαCGRP13−34 decoy fragment. The sequence of the designed hαCGRP13−34 fragment is shown above the panels. (A) Vasomotor assessment of hαCGRP13−34 in isolated mesenteric arteries using wire myography (n = 8). Arteries were precontracted and cumulative concentration-response curves to hαCGRP were generated in the absence or presence of 10 µM hαCGRP13−34. (B) Evaluation of potential interference with CGRP ELISA detection. Concentration-response measurements of hαCGRP13−34 alone and in the presence of 125 pg/mL hαCGRP were performed to assess whether the fragment affects assay detection (n = 3 of each). (C) Hemi-skull CGRP release experiment assessing the effect of 10 µM hαCGRP13−34 on basal and capsaicin-evoked CGRP release measured by ELISA (n = 6). (D) CGRP ELISA competition curve showing detection of hαCGRP in the absence or presence of 10 µM hαCGRP13−34 (n = 3). Data are shown as mean ± SEM. Statistical comparisons were made with two-way ANOVA with “Sidak” multiple comparison’s test (A) or one-way ANOVA with Dunnett’s multiple comparison’s test (C) (* p < 0.05, ***p < 0.001) To examine whether hαCGRP13−34 could compete with CGRP for enzymatic degradation, the fragment was first tested in the wire myograph using the same setup as in Fig. 3. Addition of 10 µM hαCGRP13−34 to precontracted mesenteric arteries did not evoke vasorelaxation (data not shown). We then continued with a dose response curve to hαCGRP, as we hypothesised that the decoy peptide would compete with CGRP for enzymatic degradation, thereby enhancing the vasodilatory response evoked from CGRP. Although no statistically significant difference between groups was observed in pEC50 values (vehicle: 9.68 ± 0.07; decoy: 9.93 ± 0.13; p = 0.13), a small but significant potentiation of the hαCGRP concentration response curve was observed at 300 pM in the presence of the decoy (p < 0.05, Fig. 8A). Before testing whether hαCGRP13−34 could influence CGRP degradation in the hemi-skull model, potential interference with the CGRP ELISA assay was examined. A concentration-response curve of hαCGRP13−34 was therefore generated in the absence and presence of 125 pg/mL hαCGRP (Fig. 8B). The fragment alone did not produce a detectable signal in the assay. A small increase in measured CGRP was observed at 1 nM − 100 nM, whereas no detectable effect occurred at higher concentrations up to 10 µM. Confirming that 10 µM hαCGRP13−34 did not alter the apparent CGRP detected in the ELISA assay, this concentration was used in subsequent experiments examining potential competition during CGRP degradation in the hemi-skull preparation. In the hemi-skull model, basal CGRP release did not differ between control and hαCGRP13−34 conditions (19.4 ± 1.5 pg/mL vs. 14.0 ± 0.7 pg/mL; p = 0.9336). Capsaicin stimulation increased CGRP release markedly in both conditions, from 15.0 ± 1.1 pg/mL to 81.67 ± 7.6 pg/mL in control hemi-skulls (p < 0.001) and from 14.0 ± 0.7 pg/mL to 75.6 ± 8.9 pg/mL in the presence of hαCGRP13−34 (p < 0.001) (Fig. 8C). No statistically significant difference between capsaicin responses with or without the fragment was observed. When compared relative to their own baseline levels, a slightly lower CGRP signal was observed in the hαCGRP13−34 condition compared with control. The ELISA standard curve indicates that hαCGRP13−34 would be expected to increase apparent CGRP detection primarily at low CGRP concentrations (Fig. 8D). If the fragment remained stable in the hemi-skull preparation, this would therefore be expected to increase the baseline CGRP signal. As no such increase was observed, the data instead indicate that rapid degradation of hαCGRP13−34 in the hemi-skull model likely prevents it from protecting hαCGRP against degradation by local peptidases. Discussion The present study examined how hαCGRP is degraded in the rat dura mater and whether the resulting degradation fragments retain biological activity relevant to migraine. The data show that intact hαCGRP disappears rapidly in meningeal preparations, that this loss is associated with a defined degradation pattern dominated by a limited number of fragments, and that most of these metabolites are functionally inactive in the assays examined. Across vascular, receptor activity, mast cell, and barrier related readouts, only hαCGRP11−37 retained measurable activity, where it behaved as a weak competitive antagonist at the CGRP receptor. The dominant consequence of local αCGRP degradation at the dura mater therefore appears to be termination of intact peptide signalling and reduction of peptide availability rather than generation of secondary bioactive fragments. The Ussing chamber experiments provide the first indication of this process. Following administration of hαCGRP to either the skull-facing or brain-facing side, the peptide was not detected in the opposite compartment, whereas the concentration declined markedly on the donor side during the initial 15 min. This pattern indicates that intact hαCGRP does not cross the dura mater under these conditions and instead disappears locally. Non-specific adsorption to the preparation and endosomal processing of ligand–receptor complexes (as demonstrated by Padilla et al. 2007 [52]) cannot be excluded as contributing factors and could explain the initial drop from 1000 pg/mL to approximately 700 pg/mL at the first measurement. Nevertheless, the detection of specific hαCGRP degradation fragments in the subsequent proteomic analysis supports local peptide degradation as the primary explanation for the continued loss within the examined time frame. These observations are consistent with the short half-life of CGRP in biological systems and with earlier work showing that CGRP is metabolised by extracellular peptidases that also process substance P and related neuropeptides [28, 29]. Exposing the dura mater to hαCGRP was associated with a reduction in transdural resistance, likely to reflect local activation of a dural target by hαCGRP. The present experiments do not resolve the molecular basis of this effect, but the observation shows that hαCGRP exposure can induce a measurable electrophysiological response in the dura mater. The results from the Ussing chamber, further indicate that hαCGRP does not readily cross the rat dura mater, likely due to the barrier imposed by the dense, epithelial-like subdural layer [53, 54]. Although transdural passage of CGRP was not detected in this study, partial penetration into the dural tissue cannot be excluded, as CGRP may diffuse into the dura without reaching detectable levels on the opposite side [55]. Proteomic analysis of the hemi-skull preparation provided insight into the degradation pattern. In the reduced samples, hαCGRP1−17 and hαCGRP1−24 were the dominant fragments. Analysis performed without reduction of the disulfide bridge additionally revealed hαCGRP11−24 and hαCGRP11−37. Processing the samples under conditions that either preserve or reduce the disulfide bond enables detection of both N-terminal and C-terminal hαCGRP fragments with potential endogenous relevance. Under reducing conditions, removing the disulfide bridge covering the N-terminal region resulted in hαCGRP1−17 and hαCGRP1−24 being the dominant fragments. In contrast, preservation of the disulfide bond, effectively limited detection of these highly abundant N-terminal fragments. This, in turn, enhanced the sensitivity for detecting less abundant species, including hαCGRP11−37 and hαCGRP22−37. For example, hαCGRP22−37 accounted for < 1% of detected peptides under reducing conditions but approximately 20% when the disulfide bond was preserved. Accordingly, parallel analysis under both conditions increases the likelihood of identifying less abundant degradation fragments. The seven fragments obtained indicate cleavage events at S17-R18 and K24-N25, identifying these positions as major proteolytic sites in the meningeal preparation, similar to plasma [31]. The high abundance of hαCGRP1−17 in the present dataset indicates that cleavage at S17-R18 represents one of the dominant degradation pathways for hαCGRP in the dura mater preparation. This observation also aligns well with previous findings showing that protection of this area by lipidation increases the half-life of hαCGRP8−37 [33]. Although it was beyond the scope of this study to perform a detailed characterisation of responsible proteases it is worth mentioning that insulin-degrading enzyme (IDE) has been identified as the major degradation enzyme for mouse CGRP in plasma and spinal cord at S17-R18. In the present work, hαCGRP1−17 was identified as the dominant fragment which means that IDE could also have been a major involved protease in the degradation of hαCGRP in dura mater [31]. Other in vitro studies have demonstrated that neurons and microglia cleave CGRP at R11 and R18, yielding a fragmentation pattern comparable to that produced by a serine protease, which supports involvement of this protease class in the generation of the CGRP1−17 fragment [56]. Furthermore, mast cell derived tryptase has been shown to cleave CGRP and attenuate its vascular activity [57, 58], with the major cleavage sites being R18 and K24 [58]. Therefore, the high abundance of hαCGRP1−24 may, at least in part, be attributed to tryptase-mediated degradation. The mechanism by which the hαCGRP11−37 fragment arises in the dura mater preparation remains uncertain. The proteomic analysis confirms the presence of this fragment but does not identify the specific cleavage process that generates it. Sequential processing events could produce fragments beginning at residues 8, 9, 10, or 11 before formation of the hαCGRP11−37. Non-enzymatic cleavage mechanisms may also contribute, as it has been shown that spontaneous peptide bond cleavage adjacent to serine or threonine residues has been reported in long lived proteins, and this process can be facilitated when nearby histidine residues promote hydrolysis of the peptide bond through metal ion coordination mechanisms [59] similar to the T9, H10, R11 sites of hαCGRP. Functional experiments showed that the identified degradation fragments largely lacked biological activity in the assays examined. None of the major fragments induced relaxation of mesenteric arteries when applied alone. Similarly, fragments such as hαCGRP1−17 and hαCGRP1−24 did not alter hαCGRP induced relaxation in the wire myograph experiments. The same pattern was observed in the in vitro receptor experiments. hαCGRP1−17 and hαCGRP1−24 did not induce cAMP production in Cos-7 cells expressing the CGRP receptor and did not alter receptor activation by full-length peptide, suggesting that we can expect similar lack of response to these fragments in humans. Current models of CGRP receptor activation at the CLR/RAMP1 receptor complex involve a two-domain interaction in which the C-terminal region contributes primarily to receptor binding while the N-terminal residues are required for receptor activation and signalling [60, 61]. Although the antagonistic data on C-terminal fragments are well established [32], whether N-terminal fragments such as hαCGRP1−17 or hαCGRP1−24 retain the capacity to act as competitive antagonists has not previously been investigated. In the present experiments, neither fragment altered hαCGRP responses in the vascular or receptor assays, indicating that these fragments do not preserve sufficient receptor interaction to produce detectable antagonism. Among the fragments examined, hαCGRP11−37 produced a rightward shift of the hαCGRP concentration response curve without reducing maximal relaxation. Schild-analysis was consistent with weak competitive reversible antagonism. A comparable shift of the hαCGRP concentration response curve was observed in the receptor assays in Cos-7 cells expressing the human or rat CGRP receptor. The antagonist potency observed here was lower than that of hαCGRP8−37 and the difference was similar to what was previously reported between hαCGRP8−37 and hαCGRP11−37 [32]. The effective concentrations of hαCGRP and its degradation fragments at the CGRP receptor remain undefined. We did observe a weak antagonistic effect already at 300 nM hαCGRP11−37, which might have a physiologically relevant effect in the receptor micro-environment. Rodent studies have suggested that CGRP exposure can promote histamine release or interact with meningeal mast cells [62]. Other studies using human mast cells have not observed a clear degranulation response to CGRP [63, 64]. In the present rat hemi-skull preparation neither full-length hαCGRP nor the dominant degradation fragments hαCGRP1−17 and hαCGRP1−24 induced mast cell degranulation when assessed by toluidine blue staining. In a previous study using PACAP-38 to induce mast cell degranulation, it was evident that a high net charge of peptide-fragments is crucial for mast cell degranulation [65]. Despite, hαCGRP11−37, having the highest net charge of our fragments (5.00), it did not produce a significant increase in degranulating mast cells. This suggests that neither hαCGRP nor its degradation fragments evoke mast cell degranulation under these conditions. However, we cannot exclude the possibility of low-level or partial release of mast cell mediators. Previous studies have quantified elevated levels of histamines from rat dural mast cells following CGRP stimulation [66, 67], suggesting that subtle secretory responses may occur that are not captured by the applied assay. Furthermore, other studies have found that hαCGRP degranulation is dependent on IgE activation [68], or that it can boost the immune response initiated by IgE in mast cells, by enhancing microtubule reorganisation [69]. The lack of response observed with hαCGRP, and its fragments could potentially be due to stimulation under isolated conditions or sensitivity of the assay. The ELISA experiments showed that peptide degradation can also influence biochemical detection of CGRP. When tested alone, none of the fragments produced a measurable signal in the assay when tested at 2.6 nM, corresponding to 10,000 pg/mL full-length hαCGRP based on molar equivalence. Fragments containing N-terminal antibody recognition motifs altered detection of intact hαCGRP when present together with the peptide at the same concentrations. This observation is consistent with earlier immunometric assay development, where Forbert et al. (1999) demonstrated that CGRP detection depends on coordinated recognition of distinct N-terminal and C-terminal epitopes, and that peptide truncation differentially affects antibody binding depending on epitope location, with assay characteristics comparable to those observed for the ELISA used in the present study [70]. Addition of hαCGRP1−17 and hαCGRP1−24 at 2.6 nM shifted the CGRP standard curve and reduced apparent CGRP detection. As these are the most abundant fragments (Fig. 2), some care must be taken in analysing ELISA data as they might interfere with the detection of CGRP. In contrast, the C-terminal fragment hαCGRP11−37 affected detection only at higher concentrations, with inhibition of CGRP detection observed with an IC50 of 25.1 nM. This shows that C-terminal fragments are less likely to interfere with the current ELISA assay. However, the concentrations of these fragments in plasma or tissue compartments are not defined, making it difficult to assess their physiological relevance. The present findings therefore highlight a potential source of variability in CGRP biomarker studies and emphasise the need for careful interpretation of ELISA-based measurements in the presence of peptide degradation. Proteolytic processing of hαCGRP may also influence recognition by therapeutic monoclonal antibodies, because these antibodies bind different regions of the peptide. Patent data for fremanezumab describe binding to a C-terminal epitope within residues 25–37, and structural work on eptinezumab supports binding within residues 26–37, whereas documents have described galcanezumab as potentially recognising a more central region of the peptide [71, 72]. In the present dataset, few C-terminal fragments were identified, whereas more abundant fragments such as hαCGRP1−24 retained central sequence elements. This pattern suggests that local degradation may affect the peptide species available for binding different CGRP antibodies, which could contribute to why some patients benefit from switching between antibody treatments [73]. The hαCGRP13−34 decoy experiments examined whether a designed fragment spanning the principal cleavage regions could compete with enzymatic degradation and thereby preserve intact hαCGRP. Earlier studies have shown that neuropeptides such as CGRP and substance P can compete for enzymatic degradation by shared peptidases [28, 29]. In the present experiments, co-incubation with 10 µM hαCGRP13−34 produced a small potentiation of hαCGRP induced relaxation of mesenteric arteries (Fig. 8). The absence of increased CGRP detection together with the ELISA control experiments suggests that hαCGRP13−34 is rapidly degraded in the hemi-skull preparation and therefore cannot maintain a sustained competitive effect on CGRP degradation, limiting the usefulness of this approach as an attempt to increase CGRP levels. The use of human αCGRP was selected to maximise translational relevance and to prioritise use of the human peptide where possible; however, a limitation is that the rat meningeal preparation may introduce species-dependent differences in enzymatic processing and fragment formation. Although such differences cannot be excluded, the dominant cleavage regions identified correspond to previously reported sites, supporting relevance across systems. In addition, the specific proteases responsible for the observed degradation pattern were not directly identified, but the consistency of cleavage sites with known enzymatic targets supports a physiologically relevant process. The endogenous concentrations and temporal dynamics of CGRP degradation fragments in vivo remain undefined, which limits direct extrapolation of ELISA interference to biomarker studies; however, the data demonstrate that interference is possible under controlled conditions and therefore represents a factor that should be considered. Finally, a limitation relates to sample size and study population, as sample sizes were kept to a minimum in vitro when no effect was observed, or in vivo in accordance with reduction principles and only male rats were included. As the study focuses on peptide cleavage patterns rather than regulated biological outputs, major differences in fragment identity would not be expected; however, differences in temporal dynamics or quantitative regulation between sexes cannot be excluded. Overall, the data demonstrate that αCGRP released in the dural environment could undergo rapid proteolytic processing that primarily removes intact peptide from the system. The predominant fragments generated when hαCGRP is added exogenously to the dura, do not activate the CGRP receptor, do not alter vascular responses to intact peptide, and do not induce mast cell degranulation in the rat dura mater. The principal biological outcome of local αCGRP metabolism at the dura mater therefore appears to be degradation of intact peptide, with the potential exception of hαCGRP11−37 which retains weak competitive antagonist activity at the CGRP receptor. These findings indicate that the direct migraine relevance of the identified fragments is likely limited, but this does not exclude that alterations in CGRP degradation and breakdown dynamics may represent an important aspect to investigate in migraine pathophysiology. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1