Intracellular pathways of calcitonin gene-related peptide-induced relaxation of human coronary arteries: A key role for G β γ subunit instead of cAMP

Background and Purpose: Calcitonin gene-related peptide (CGRP) is a potent vasodilator. While its signalling is assumed to be mediated via increases in cAMP, this study focused on elucidating the actual intracellular signalling pathways involved in CGRP-induced relaxation of human isolated coronary arteries (HCA). Experimental Approach: HCA were obtained from heart valve donors (27 M, 25 F, age 54 ± 2 years). Concentration – response curves to human α -CGRP or forskolin were constructed in HCA segments, incubated with different inhibitors of intracellular

Calcitonin gene-related peptide (CGRP) is a 37-amino acid neuropeptide and potent vasodilator.It exists in two isoforms, α-CGRP and β-CGRP, which differ by three amino acids in humans.In sensory nerves, α-CGRP is predominantly present, whereas in the enteric nervous system β-CGRP is more abundant (Mulderry et al., 1988).The isoform α-CGRP is involved in the regulation of the cardiovascular system (Smillie et al., 2014) and was the main focus of the current study.Under healthy, physiological circumstances, CGRP is not primarily involved in regulation of vascular tone (Smillie & Brain, 2011).
However, in a pathophysiological state, such as during hypertension or ischemia, CGRP can act as a safety molecule by inducing vasodilation and increasing the blood flow to organs and tissues (MaassenVanDenBrink et al., 2016;Smillie & Brain, 2011).
In addition to its effects on the cardiovascular system, CGRP is involved in the pathophysiology of migraine.During an attack, the trigeminovascular system is activated and CGRP is released, causing a relaxation of blood vessels and subsequent activation of nociceptors, resulting in pain (Noseda & Burstein, 2013).Moreover, infusion of CGRP in migraine patients can induce a migraine-like headache (Hansen et al., 2010).Novel antimigraine medication aims to block either CGRP or its receptor, using small-molecule CGRP receptor antagonists (gepants; e.g., atogepant, rimegepant, ubrogepant and zavegepant) or monoclonal antibodies targeting CGRP (eptinezumab, fremanezumab, galcanezumab) or its receptor (erenumab) (de Vries et al., 2020).
CGRP binds to the canonical CGRP receptor, which consists of a seven-transmembrane protein called calcitonin-like receptor (CLR) coupled to receptor activity modifying protein 1 (RAMP1) and receptor component protein (RCP) (Hay et al., 2018).CLR is a G proteincoupled receptor (GPCR), and binding of an agonist to the extracellular domain of the receptor leads to activation of G proteins by exchange of GDP for GTP (Weis & Kobilka, 2018).Subsequently, the α subunit of the G protein bound to GTP can dissociate from the other subunits β and γ and can activate different intracellular signalling pathways.
When Gα s is activated, this leads to subsequent activation of adenylyl cyclase (AC) which converts ATP into cyclic 3 0 -5 0 adenosine monophosphate (cAMP), an important second messenger molecule that is involved in many intracellular processes (Brain & Grant, 2004).
From previous literature, it is known that the binding of CGRP to its receptor leads to an increase in intracellular cAMP in vascular smooth muscle cells, as well as other cell types (e.g., transfected Cos 7 cells, endothelial cells) (Hay et al., 2018;Hirata et al., 1988); accordingly, relaxation of blood vessels is assumed to be mediated by cAMP.In this study, we focused on elucidating the exact signalling pathways downstream of the CGRP receptor that are involved in CGRP-induced relaxation of human isolated coronary arteries, by blocking the intracellular signalling at different levels (Figure 1).The current research aims to enhance the understanding of the exact mechanism of action of the potent vasodilator CGRP, and it could potentially lead to new targets for more effective antimigraine therapy.

What is already known
• CGRP receptor activation results in increases in intracellular cAMP levels and relaxation of human arteries.
• CGRP (receptor) blockade is used as a novel target for the treatment of migraine.

What does this study add
• CGRP-induced vasodilation of human coronary arteries is not inhibited by targeting the cAMP-dependent pathway.
• Inhibition of Gβγ subunits results in inhibition of CGRPinduced relaxation of human isolated coronary arteries.

| Organ bath experiments with human coronary arteries
For functional experiments with the distal coronary arteries, 2-mm segments were mounted in a Mulvany myograph system (Danish Myo Technology, Aarhus, Denmark) using Ø 40-μm stainless-steel wires.In some segments, the endothelium was mechanically removed using a human hair before mounting into the Mulvany myograph system.The organ baths were filled with carbonated and oxygenated Krebs solution and kept at 37 C during the experiment.Vessel segments were allowed to equilibrate for 30 min before normalization, in which the vessel segment was stretched to a tension normalized to 0.9 times the estimated diameter at 100-mmHg transmural pressure (Mulvany & Halpern, 1977).The LabChart data acquisition system (AD instruments Ltd, Oxford, UK) was used to record the tension during the experiment.The vessel segments were contracted using 30-mM KCl, washed with new Krebs solution twice, and subsequently exposed to 100-mM KCl, followed by two additional washing steps.
Functional experiments with the proximal coronary arteries were performed using larger organ baths.The arteries were cut into segments of 4-5 mm and mounted on stainless steel hooks.The vessel segments were stretched to 15 mN (Gupta et al., 2006).Next, the vessels were contracted using 30-mM KCl twice and 100-mM KCl, followed by two washing steps after every contraction.Vessel segments were incubated with AC inhibitors SQ22536 (100 μM) or MDL-12330A (100 nM), the PKA inhibitor H89 (1 μM) or vehicle for 15 min, before precontraction using 30-mM KCl.After another 15 min, increasing concentrations of CGRP (0.1 nM, 10 nM, 1 μM) or forskolin (1 nM, 100 nM, 10 μM) were added to the segments in the presence or absence of PKA or AC inhibitors.Finally, the quality of endothelium was investigated using U46619 and substance P, similar to the distal coronary artery segments.

| Intracellular cAMP measurements
For measurements of intracellular cAMP levels, vessel segments were incubated with or without SQ22536 (100 μM) or DDA (100 μM) for 30 min in Krebs solution containing isobutylmethylxanthine (IBMX, 100 μM), followed by stimulation with CGRP (100 nM) or forskolin (10 μM) for 5 min.Next, vessel segments were snap frozen using liquid nitrogen and stored at À80 C until analysis using the Direct cAMP ELISA kit (Enzo Life Sciences, Farmingdale, USA), according to manufacturer's instructions.

| Data analysis
The functional responses in human isolated coronary arteries were analysed using Prism 8 (GraphPad Software, San Diego, CA, USA).Nonlinear regression analysis was used to construct sigmoidal curves and data are presented as mean ± SEM.Control experiments, and experiments with inhibitors, were performed in segments from the same human donor to obtain paired results.For each inhibitor, experiments were performed in five to eight different donors.The response to KCl, pEC 50 values of individual curves and cAMP levels were compared between conditions with and without inhibitors using a paired t-test for a comparison of two conditions within tissue from a single donor, or by one-way ANOVA with each row representing matched data from the same donor for comparing multiple conditions.If the ANOVA results were significant, groups were compared with the control condition using Dunnett's multiple comparisons test.The significance threshold for all statistical tests was set at P < 0.05; significant results are indicated by a single asterisk (*) in the figures.Data and statistical analyses comply with the recommendations of the British Journal of Pharmacology on experimental design and analysis (Curtis et al., 2022).

| Nomenclature of targets and ligands
Key protein targets and ligands in this article are hyperlinked to corresponding entries in https://www.guidetopharmacology.org and are permanently archived in the Concise Guide to PHARMACOLOGY 2021/22 (Alexander et al., 2021).

| RESULTS
3.1 | CGRP-induced relaxation of human coronary arteries is not affected by inhibitors of AC, PKA, EPAC, Gα q , ODQ, PKG or various potassium channels CGRP potently relaxed human isolated coronary arteries precontracted with KCl (pEC 50 : 8.55 ± 0.08).AC inhibitors SQ22536 + DDA or MDL-12330A did not affect the concentration-response curve to CGRP in these human isolated coronary arteries (Figure 2a,b).When targeting PKA, which is activated downstream of the CGRP receptor by cAMP (Miyoshi & Nakaya, 1995), the concentration-response curve to CGRP was not affected with the PKA inhibitors Rp-8-Br-MB-cAMPs or H89 (Figure 2c,d).Moreover, the EPAC1/2 inhibitor ESI09 had no effect on the relaxation to CGRP at the concentration used in the current study (1 μM) (Figure 2e).When higher concentrations of ESI09 (10 μM, 20 μM) were used, the vessel segments could no longer be precontracted using KCl, and therefore functional experiments were not possible.To determine whether Gα q signalling was involved, the vessel segments were incubated with the Gα q inhibitor YM-254890.However, no shift in the concentration response curve to CGRP could be observed (Figure 2f).Interestingly, the Gα q inhibitor completely abolished the endothelium-dependent vasorelaxation induced by substance P, even after removing the YM-254890 The effect of different inhibitors on concentration-response curves to human αCGRP (10 pM-1 μM) in human isolated coronary arteries.The concentration response curve to CGRP was not affected by the PKA inhibitors Rp-8-Br-MB-cAMPs (Rp-8) (a) or H89 (b), nor the AC inhibitors SQ22536 + DDA (c) or MDL-12330A (d).The EPAC inhibitor ESI09 (e) or Gα q inhibitor YM-254890 (f) also did not affect the relaxation to CGRP.Potassium channel blockers iberiotoxin (g) or glibenclamide (h) did not affect the concentration-response curve to CGRP.Moreover, the combination of two potassium blockers, TRAM-34 plus apamin, together with an inhibitor of endothelial nitric oxide synthase (L-NAME) did not affect the concentration response curve to CGRP (i).Inhibition of GC using ODQ (j), or inhibition of PKG using KT5823 (k) did not affect the concentration-response curve to CGRP, nor did mechanical removal of the endothelium (l).
compound by washing the vessel segment twice (Figure S1A).Next, we investigated the involvement of K + channels in CGRP-induced relaxation.The large-, intermediate-and small-conductance Ca 2+activated K + channels were targeted using iberiotoxin, TRAM-34 and apamin respectively.ATP-sensitive K + channels were blocked using glibenclamide.None of these K + channel blockers affected the relaxation to CGRP (Figure 2g-i).Inhibition of GC using ODQ (Figure 2j) or inhibition of PKG using KT5823 (Figure 2k) did not affect the concentration-response curve to CGRP, nor did mechanical removal of the endothelium (Figure 2l), after which the average response to substance P expressed as percentage of precontraction to U46619 decreased from 83 ± 8 to 6 ± 1, indicating successful denudation of the arteries (Figure S1B).The response to KCl was not affected by removal of endothelium (7 ± 1 mN and 6 ± 1 mN for EC+ and ECÀ, respectively) (Figure S1C).

| Phosphodiesterase inhibitors do not potentiate the relaxation to CGRP in human coronary arteries
Phosphodiesterases catalyse hydrolysis of cAMP to 5 0 AMP and/or cGMP to 5 0 GMP, thereby lowering intracellular cAMP and cGMP levels, respectively (Conti & Beavo, 2007).Here, specific phosphodiesterases were inhibited, to assess whether this could potentiate the relaxation to CGRP.Neither the phosphodiesterase 1 inhibitor vinpocetine, the phosphodiesterase 3A inhibitor cilostazol or the phosphodiesterase 5 inhibitor sildenafil could potentiate the relaxation to CGRP (Figure 3a-c).However, phosphodiesterase inhibitors vinpocetine and cilostazol decreased the contractile response to KCl, which is used for the precontracting the arteries before the concentration-response curve to CGRP.The response is expressed as The effect of phosphodiesterase inhibitors on CGRP-induced vasorelaxation of human isolated coronary arteries.Phosphodiesterase 1 inhibitor vinpocetine (a), phosphodiesterase 3A inhibitor cilostazol (b) and phosphodiesterase 5 inhibitor sildenafil (c) did not affect the concentration response curve to CGRP.Increasing concentrations of phosphodiesterases vinpocetine and cilostazol decrease the preconstriction using 30-mM KCl compared with the KCl 100-mM response of the same vessel segment (d).
the precontraction induced by 30-mM KCl after incubation with phosphodiesterase inhibitors, relative to the response to 100-mM KCl in the same vessel segment before addition of phosphodiesterase inhibitors (Figure 3d).

| Forskolin-induced relaxation of human coronary arteries is not inhibited by PKA or AC inhibitors
Forskolin directly activates AC and was able to induce relaxation of human isolated coronary artery segments (pEC 50 : 6.98 ± 0.50).However, the PKA inhibitors Rp-8-Br-MB-cAMPs or H89 (Figure 4a,b), and the AC inhibitors SQ22536 + DDA or MDL-12330A (Figure 4c,d) could not attenuate the relaxation.

| In proximal coronary arteries PKA and AC inhibition has no effect on CGRP-or forskolin-induced relaxation
To investigate whether the responses to CGRP and forskolin, and its intracellular signalling, vary between distal coronary artery segments and more proximal larger coronary arteries, the effect of PKA inhibition and AC inhibition was assessed in proximal coronary arteries as well.
CGRP was significantly less potent in proximal coronary arteries compared with distal coronary arteries (pEC 50 : 7.15 ± 0.38 and 8.55 ± 0.08, respectively) and had a smaller maximum response (E max : 20.96 ± 5.66 and 87.78 ± 3.88 respectively).However, the AC inhibitors SQ22536 and MDL-12330A, and the PKA inhibitor H89, did not attenuate the relaxation to CGRP in proximal coronary arteries (Figure 5a), similar to the observations in distal coronary arteries.Similarly, forskolin was significantly less potent in proximal coronary arteries compared with distal coronary arteries (pEC 50 : 6.07 ± 0.25 and 6.98 ± 0.50, respectively) and showed a smaller maximum response (E max : 56.69 ± 9.74 and 82.22 ± 12.76 respectively), while the AC inhibitors and PKA inhibitor did not attenuate the relaxation to forskolin (Figure 5b).

| AC inhibitors affect intracellular cAMP levels in human coronary arteries
The intracellular cAMP levels were measured after stimulation of human isolated coronary arteries with CGRP or forskolin in the absence or presence of different inhibitors, to verify whether the inhibitors are effective in our system.CGRP (100 nM) and forskolin (10 μM) both significantly increased the intracellular cAMP levels compared with no stimulation (Figure 6a,b), while SQ22536 and DDA significantly inhibited this increase in cAMP after stimulation with forskolin (Figure 6b).For stimulation with 100-nM CGRP, the inhibition caused by SQ22536 (n = 11) did not reach significance, although a clear trend was observed (Figure 6a).For DDA, the same trend could be observed (CGRP 1585 ± 670 fmolÁmg À1 tissue cAMP, CGRP + DDA 1062 ± 296 fmolÁmg À1 tissue in n = 4 paired experiments).However, the increase in cAMP induced by 100-nM CGRP was much smaller compared with the increase after stimulation with 10-μM forskolin (approximately eightfold difference) and increasing the CGRP concentration of the stimulation from 100 nM to 1 μM did not lead to a larger increase in cAMP levels (789 ± 296 and 402 ± 77 fmolÁmg À1 tissue, respectively, n = 11).Moreover, while 10-μM forskolin increased cAMP levels to a great extent (6524 ± 1295 fmolÁmg À1 tissue) compared with no stimulation (196 ± 38 fmolÁmg À1 tissue), a 10-fold lower concentration of forskolin (1 μM) did not induce a significant increase in cAMP levels compared with no stimulation.MDL-12330A at a concentration of 100 nM was not able to decrease cAMP levels (Figure 6c).

| DISCUSSION
The current study investigates the intracellular signalling of CGRPinduced relaxation of human coronary arteries.In accordance with our earlier studies (Gupta et al., 2007(Gupta et al., , 2006)), we found that CGRPinduced vasodilation was mediated via a direct effect on smooth muscle cells in the arterial wall, because mechanical removal of the endothelium did not affect the response to CGRP.Previous research has shown that CGRP-induced relaxation can be inhibited using small molecule CGRP receptor antagonists or a monoclonal antibody against the CGRP receptor in human isolated coronary arteries (de Vries et al., 2023;Rubio-Beltrán, Labastida-Ramírez, et al., 2019), confirming that CGRP receptor activation induces relaxation of blood vessels.
Likewise, it has been shown that binding of CGRP to its receptor leads to an increase in intracellular cAMP levels (Hay et al., 2018;Hirata et al., 1988), which was confirmed in the current study.This increase in cAMP is assumed to be involved in the relaxation of blood vessels (Murray, 1990).Moreover, stimulation with forskolin, which directly activates AC, leads to an increase in cAMP levels and potently relaxes blood vessels.However, in the current study, three different types of AC inhibitors could not attenuate the relaxation induced by CGRP or forskolin, suggesting that the increase in cAMP might not be responsible for the relaxation.In humans, AC exists in nine membrane-bound isoforms or can be present in a soluble form in the cytoplasm (Bitterman et al., 2013).In the current study, three different AC inhibitors were used, of which SQ22536 can most selectively inhibit isoform 5 and 6, DDA specifically binds transmembrane isoforms of AC, while MDL-12,330A does not discriminate between membranebound AC and soluble AC (Bitterman et al., 2013;Brand et al., 2013), yet none of these inhibitors affected the relaxation to CGRP or forskolin.
AC inhibitors SQ22536 and DDA were tested with regard to their ability to reduce cAMP production after stimulation with forskolin and CGRP.Both inhibitors significantly decreased cAMP production after stimulation with 10-μM forskolin, showing that they are effective in these human tissues and in the current experimental setup.For stimulation with 100-nM CGRP, the inhibition caused by SQ22536 and DDA did not reach significance.However, the overall increase in cAMP levels after stimulation with 100-nM CGRP was much smaller compared with stimulation with 10-μM forskolin.Possibly, higher numbers of experiments would have been necessary to reach significant inhibition for this minor increase in cAMP.Further increasing the concentration of CGRP used for stimulation of the tissue to 1 μM did not result in higher levels of cAMP.Moreover, the fact that much higher levels of cAMP can be reached after stimulation with forskolin, while the relaxation is similar as to CGRP, could also gave an indication that cAMP levels might not be the main inducing factor of relaxation in these tissues.It is important to note that in our system, 100 nM of MDL-12330A cannot effectively reduce the cAMP levels in human coronary arteries, while higher concentrations of MDL-12330A do not allow precontraction of the artery.Therefore, this compound cannot be used to study the role of AC inhibitors in CGRPor forskolin-induced relaxation in our myograph system.These findings that AC inhibitors do not affect CGRP-or forskolin-induced relaxation in human coronary arteries, while they do affect cAMP production, are similar to what has been observed previously in ovine pulmonary veins (Gao & Raj, 2001), in which the AC inhibitors SQ22536 and W-7 did not inhibit forskolin-induced relaxation, although they affected cAMP elevation.Furthermore, cAMP levels only seemed to robustly increase at a concentration of forskolin above 300 nM, while at this concentration the vessel tension had already decreased for 75%, suggesting that the relaxation already occurs at a concentration of forskolin that does not increase intracellular cAMP levels substantially (Gao & Raj, 2001).This is similar to the current study, in which 1-μM forskolin did not induce a significant increase in cAMP levels, while an average relaxation of 82% was already reached at this concentration.Moreover, a study in uterine arteries from pregnant women showed that SQ22536 did not attenuate the relaxation to CGRP, although it inhibited relaxation to other peptides (Chauhan et al., 2022).
Inhibition of Epac or PKA, which are both downstream of cAMP in intracellular signalling, did not affect the relaxation to CGRP in human coronary arteries.PKA activates different potassium channels at the plasma membrane, which in turn leads to a hyperpolarization of the cell and subsequent relaxation of the smooth muscle cells in the arterial wall (Jackson, 2017).However, inhibition of different types of potassium channels, including large-, intermediate-and smallconductance Ca 2+ -activated K + channels and ATP-sensitive K + channels, did not affect the relaxation to CGRP.Interestingly, it was shown that this same ATP-sensitive K + blocker glibenclamide was unable to block CGRP-induced headache (Coskun et al., 2021) or levcromakalim-induced headache (Al-Karagholi et al., 2020) in healthy volunteers, although the binding of glibenclamide to the Kir6.
1 / SUR2B subunits of the K ATP channels in the clinical studies might have been suboptimal at the 10-mg dose, with only low target engagement of glibenclamide (Christophersen & Dyhring, 2023).
Moreover, signalling via GC and PKG or Gα q signalling is not involved in the relaxation to CGRP.However, the Gα q inhibitor YM-254890 completely abolished the endothelium-dependent vasorelaxation induced by substance P, indicating that this compound did elicit an effect in the human isolated coronary arteries, even if it did not inhibit CGRP-induced relaxation.Substance P acts on the neurokinin 1 (NK 1 ) receptor in endothelial cells, which subsequently leads to activation of Gα q and Gα s signalling (Thom et al., 2021), which is effectively blocked in our system by the Gα q inhibitor YM-254890.
The relaxation to CGRP was not potentiated using different phosphodiesterase inhibitors, which are involved in the hydrolyses of cAMP to 5 0 AMP, or cGMP to 5 0 GMP, thereby decreasing the intracellular cAMP and cGMP levels.However, the phosphodiesterase inhibitors vinpocetine and cilostazol showed a concentration-dependent effect on precontraction of the vessels with KCl, suggesting that they can elicit an effect in these human isolated coronary arteries, albeit not on CGRP-induced relaxation.Together with the effective inhibition of cAMP production by AC inhibitors and the inhibition of endothelium-dependent relaxation with YM-254890, the phosphodiesterase effect on KCl response serves as a positive control that these inhibitors are effective in our model system.
The results in the current study suggest that increases in cAMP do not necessarily have to result in relaxation of arteries, which is in line with a previous study (Vegesna & Diamond, 1986) that showed a simultaneous increase in cAMP and relaxation in bovine coronary arteries after stimulation with prostacyclin (PGI 2 ) whereas, in rabbit aorta, the increase in cAMP was accompanied by a contraction.
Therefore, the relation between increased cAMP levels and relaxation might not be causal.Over the years, multiple reports have shown that the increase in cAMP does not always coincide with the relaxation (Peters & Michel, 2003).Cholera toxin, iloprost, beraprost and the β-adrenoceptor agonist isoprenaline are all known to increase intracellular cAMP levels and can induce relaxation of arteries (Tanaka et al., 2003;Turcato & Clapp, 1999;Yamaki et al., 2001).However, successful inhibition of cAMP increases using the AC inhibitor SQ22536 does not attenuate relaxation to cholera toxin or prostacyclin analogues iloprost or beraprost in guinea pig aorta or tracheal smooth muscle (Tanaka et al., 2003;Turcato & Clapp, 1999;Yamaki et al., 2001).Moreover, AC inhibition did not affect isoprenalineinduced relaxation of rat abdominal aorta (Matsushita et al., 2006).
When comparing the relaxation to vasoactive intestinal polypeptide (VIP) and isoprenaline in rat aorta, it was discovered that although stimulation with VIP resulted in a much larger increase of cAMP levels compared with isoprenaline, the relaxant effect of isoprenaline was much more pronounced, suggesting that a direct correlation between the total tissue cAMP level and relaxation was absent (Schoeffter & Stoclet, 1985).It is important to note that cAMP levels can vary over different cellular compartments.It does not freely diffuse throughout the cell and can exert distinct effects based on its localization (Johnstone et al., 2018).Therefore, total cAMP levels do not have to directly represent the levels in a certain functional compartment of the cell.
The findings in the current study on CGRP signalling are in accordance with the results of a study using rat mesenteric resistance arteries, in which the authors conclude that CGRP-induced relaxation is mediated via Gβγ signalling instead of activation of Gα s and subsequent increase in cAMP (Meens et al., 2012).They show that inhibition of AC, GC, PKA or phosphodiesterase did not affect CGRPinduced relaxation, while inhibition of Gβγ using gallein could decrease the relaxant response to CGRP.Moreover, gallein increased the cAMP production after stimulation with CGRP, confirming that cAMP levels are not directly coupled to CGRP-induced vasorelaxation.
Based on the findings of this study, the Gβγ inhibitor gallein was tested in the human isolated coronary arteries in our myograph system and was shown to significantly inhibit the relaxation to CGRP, albeit less potently than in rat mesenteric arteries (Meens et al., 2012).Because the relaxation to CGRP was not inhibited completely, the Gβγ inhibitor gallein was combined with the adenylyl cyclase inhibitor SQ22536.However, this did not induce a further shift on top of gallein.The observed inhibition induced by gallein further strengthens the idea that CGRP-induced relaxation is not mediated via cAMP.However, cAMP is often measured as the main second messenger of CGRP receptor activation (Hay et al., 2018), and cAMP production can be inhibited by CGRP receptor blockade (Manoukian et al., 2019).While a significant increase in cAMP after stimulation with CGRP was observed in the current study, the main functional outcome of interest, in this case vasorelaxation, is not related to total cAMP levels.Therefore, the results of the current study emphasize the importance of measuring relevant outcomes.
Moreover, it has been shown that activation of the CGRP receptor results in activation of multiple intracellular pathways in endothelial cells next to an increase in cAMP, including NO production, Ca 2+ mobilization, pERK1 / pERK2 activation, and increased cell proliferation (Clark et al., 2021).
Gβγ subunits likely induce relaxation via an interaction with K V 7.4 channels (Stott et al., 2015).Patch-clamp experiments showed that Gβγ subunits are necessary for K V 7.4 channel responses to membrane voltage and that gallein and other Gβγ subunit inhibitors can disrupt the colocalization of Gβγ and K V 7.4.Moreover, gallein could prevent isoprenaline-induced relaxation in rat renal arteries (Stott et al., 2015).Furthermore, K V 7.4 / Kv7.5 heteromers were shown to be involved in CGRP-induced relaxation of rat middle cerebral arteries (Chadha et al., 2014).Interestingly, K V 7 channels are affected during hypertension in different types of rat arteries, including aorta, coronary and mesenteric arteries, but not in the middle cerebral artery (Chadha et al., 2014;Jepps et al., 2011).Further research should determine whether the K V 7.4 channels are involved in the CGRP-induced relaxation of human coronary arteries, as well as other human vascular beds.
The finding that cAMP levels increase after activation of the CGRP receptor is likely mediated via activation of Gα s while the CGRP-induced relaxation of human coronary arteries now seems to be mediated via Gβγ.The rise in cAMP could be a coincidental occurrence, because CGRP receptor activation leads to both activation of Gα and Gβγ, resulting in a simultaneous rise in cAMP together with relaxation, while the two are not necessarily causal.The current study shows that total cAMP levels do not directly correlate with relaxation, because AC inhibitors can decrease cAMP levels while they do not affect relaxation.However, we cannot exclude that local cAMP levels are involved in inducing relaxation of human coronary arteries.
For the relaxation experiments performed in the current study, human coronary artery segments were precontracted using KCl.
Extracellular KCl results in depolarization of the cell membrane of vascular smooth muscle cells, which results in vasoconstriction via activation of voltage-gated ion channels.In case the vasodilation mechanism of a certain agonist acts via these voltage-gated ion channels, the effect might be affected by the precontraction with KCl (Wenceslau et al., 2021).Moreover, high KCl concentrations could interfere with endothelium-dependent vasodilation by affecting hyperpolarization.Therefore, we cannot categorically exclude that the present experiments are affected by the choice of KCl as contractile agent.However, due to unstable precontraction induced by other agents, we were obliged to use KCl in these human arteries to be able to construct a reliable concentration-response curve to CGRP.However, some preliminary data (not shown) gave us no indication that the inhibitors would respond differently after precontraction with U46619.In the current experiments, removal of endothelium did not affect the relaxation response to CGRP.However, CGRP-induced relaxation was shown to be endothelium-dependent in rodent arteries precontracted with noradrenaline (Wang et al., 1991).Potentially, using KCl as precontraction eliminates the role of endothelium in CGRP-induced relaxation.Therefore, the results in the current manuscript describe endothelium-independent mechanisms of CGRPinduced relaxation.However, even without contribution of endothelium to the vasodilation caused by CGRP, relaxation reached approximately 88% of precontraction, suggesting that the endothelial contribution, if present in these arteries, could only have minor additional effects on top of the relaxation that is currently observed.
The results in the current study show that the response to CGRP is diminished in proximal coronary arteries compared with distal coronary arteries, which in line with previous findings from our lab (Rubio-Beltrán, Chan, et al., 2019;Rubio-Beltrán, Labastida-Ramírez, et al., 2019).The smaller the artery, the more it contributes to overall resistance and to regulating blood flow.Future studies should investigate whether the CGRP-induced relaxation mechanisms are similar in human cranial arteries, which are in the same potency range for CGRP as the distal coronary arteries (Rubio-Beltrán, Chan, et al., 2019).
The observations in the current manuscript are relevant in the context of migraine treatment.Previously, infusion of CGRP was shown to induce migraine-like attacks in migraine patients (Hansen et al., 2010).Interestingly, it was recently shown that migraine can be provoked by targeting cAMP signalling directly, without activation of the CGRP receptor (Do et al., 2023).Combined with the data from the current manuscript, this could suggest that CGRP-mediated cAMP signalling is involved in migraine attacks, whereas the cAMP component is less relevant for the CGRP-induced relaxation of blood vessels.
Considering potential cardiovascular side effects of blocking CGRP (MaassenVanDenBrink et al., 2016), including worsening of ischaemic outcome in a stroke mouse model after blocking CGRP receptors (Mulder et al., 2020), this new knowledge about intracellular signalling pathways that differ between CGRP-induced migraine attacks and CGRP-induced relaxation could be used for development of new and safe antimigraine drugs, absent of any potential cardiovascular side effects.

| CONCLUSIONS
While CGRP signalling is generally assumed to act via cAMP, the CGRP-induced vasodilation in HCA could not be inhibited by targeting this intracellular signalling pathway at different levels.Moreover, inhibition of GC, PKG or Gα q also did not affect relaxations to CGRP.In contrast, inhibition of Gβγ subunits did block CGRP-induced relaxation of human coronary arteries, suggesting that, in contrast to common belief, CGRP-induced vasodilation of human coronary arteries is not mediated via cAMP.cAMP levels do increase after activation of the CGRP receptor, but this increase in cAMP is not coupled to the functional relaxation of arteries.It is important to further identify the intracellular signalling cascade after binding of CGRP to its receptor in human arteries of several vascular beds.Further knowledge on the intracellular signalling cascade of CGRP would ultimately allow novel antimigraine medication to target specific parts of the intracellular signalling pathway that reduces migraine, while possibly limiting (cardiovascular) side effects.
relaxation to CGRP, suggesting a different mechanism of CGRP-induced relaxation than generally believed.K E Y W O R D S cAMP, CGRP, GPCR, signalling, vasodilation 1 | INTRODUCTION Human hearts were provided by ETB-BISLIFE (Heart ValveDepartment, Beverwijk, The Netherlands), after removal of the aortic and pulmonary valves for homograft valve transplantation.Donor

•
Unravelling exact intracellular mechanisms allows development of novel antimigraine medication devoid of cardiovascular side effects.screening and acceptance was performed by the Dutch Transplant Foundation (Leiden, The Netherlands).After circulatory arrest, the hearts were harvested and stored in a sterile organ protecting solution kept at 4 C. Human coronary arteries were isolated from 27 male (age 54 ± 2 years) and 25 female (age 55 ± 3 years) post-mortem heart valve donors, with permission to use for research.Both distal human coronary arteries (0.3-to 1.5-mm internal diameter) and proximal coronary arteries (3-to 4-mm internal diameter) were obtained and stored at 4 C in carbonated and oxygenated Krebs solution composed of 118-mM NaCl, 4.7-mM KCl, 2.5-mM CaCl 2 , 1.2-mM MgSO 4 , 1.2-mM KH 2 PO 4 , 25-mM NaHCO 3 and 8.3-mM glucose (pH = 7.4).

4
The effect of AC or PKA inhibitors on forskolin-induced relaxation of human isolated coronary arteries.The concentration response curve to forskolin was not affected by PKA inhibitors Rp-8-Br-MB-cAMPs (Rp-8) (a) or H89 (b), nor the AC inhibitors SQ22536 (SQ) + DDA (c) or MDL-12330A (MDL) (d).

F
I G U R E 7 Inhibition of CGRP-induced relaxation of human coronary arteries using the Gβγ subunit inhibitor gallein.(a) Gallein significantly shifts the concentration-response curve to CGRP (n = 7).(b) pEC 50 for the concentration-response curve to CGRP in the presence or absence of gallein (n = 7).