We showed that anti-PACAP mAbs blocked PACAP38 induced hypersensitivity, but not via signaling pathways involved in GTN and levcromakalim pathways. In addition, we showed that sumatriptan had no effect on PACAP38-induced hypersensitivity relevant to migraine. This is the first study to test the effect of a PACAP-inhibiting drug on GTN- and levcromakalim-induced hypersensitivity.
PACAP and GTN
GTN is a nitric oxide (NO) donor and is an established migraine trigger used experimentally. It is blocked by anti-migraine drugs such as ibuprofen, sumatriptan (21, 26) and CGRP-inhibiting drugs (12) in the mouse model of migraine using Von Frey testing. For example, a role for CGRP in the GTN model has been clearly demonstrated in both rats and mice (12, 27–29). The observation that PACAP blockade using mAbs did not work in our GTN model indicates a distinct pathway that does not involve PACAP.
PACAP and some of its receptors have been linked to NO pathways in other experimental models (30, 31). In a mouse model of chronic migraine, repeated GTN administration significantly increased the number of PACAP-R neurons in the trigeminal ganglion but not dorsal root ganglia (30). A study using a peptidomic approach identified PACAP as a mechanistic link between NO-induced chronic migraine and opioid-induced hyperalgesia in mouse models (31) and showed that PAC1 inhibition by a selective PAC1-antagonist (M65) blocked cephalic pain induced by GTN (15). Central neuronal PAC1 receptors mediated delayed activation and sensitization of trigeminocervical neurons in rats (32). In another preclinical study also peripheral blocking of PAC1-receptor was demonstrated to be efficacious in an electrophysiological model (33), yet, a phase II clinical trial of a mAb targeting the PAC1-receptor (AMG 301) failed to show efficacy over placebo (34). This is possibly due to translational challenges and limitations of the models applied in pre-clinical studies of migraine.
In rats, GTN increased PACAP immunoreactivity within the TNC and elevated plasma PACAP concentration (32, 33). The latter was significantly reduced by ghrelin (35, 36). These studies shared the repeated GTN 10 mg/kg i.p. protocol with our experiment but lacked proper vehicle controls. However, similar results would be expected. GTN induced more photophobia, vasodilation, and trigeminal sensitization in wildtype mice compared to PACAP gene-deleted knockout mice (16). These data differ from our present results. One likely explanation may be that PACAP knockout mice have altered pain transmission mechanisms in the central nervous system (37). In contrast, anti-PACAP mAbs and PACAP38 injection may modulate dural or trigeminal nociceptors outside the brain but not cerebral receptors, as they cross the blood brain barrier very poorly (38, 39). The central vs peripheral site of action also explains the discrepancy to the study showing increased PACAP immunoreactivity in in TNC. This is not a site reachable by the anti-PACAP mAbs applied. Interestingly, a recent study based on older data showed that injection of PACAP38 into the paraventricular nucleus of the hypothalamus increased TNC activity, which could be inhibited by an intrathecal PAC1 receptor antagonist (40). Intrathecal injection of PACAP has also been suggested to induce hyperalgesia in mice (41).
PACAP and levcromakalim
PACAP is known to stimulate adenylyl cyclase to increase the formation of intracellular cAMP (42, 43) and there is some evidence that activation of cAMP-dependent pathway results in the opening of KATP-channels (44, 45). These findings have led to the hypothesis that modulation of nociceptive transmission by KATP-channels may be a common pathway in the genesis of a migraine attack (1). We showed that the effect of levcromakalim - acting downstream from cAMP could not be attenuated by anti-PACAP mAbs. In our previous study, using the same mouse model, we showed that glibenclamide (KATP-channel inhibitor) only partially inhibited PACAP38-induced hypersensitivity (10). Also, a recent clinical study showed that glibenclamide did not attenuate PACAP38-induced headache and hemodynamic changes in healthy volunteers (46). Taken together, existing literature and the present findings indicate that KATP channel opening induced hypersensitivity is not mediated by PACAP signaling. How PACAP induces hypersensitivity downstream remains unknown.
PACAP38 and sumatriptan
It is comforting when an experimental model of migraine responds to specific anti-migraine drugs, such as the triptans (5-HT1B/1D/1F receptor agonist). On the other hand, demanding such effect would make it impossible to find drugs with a novel mechanism of action. Here, we showed that sumatriptan had no effect on PACAP38-induced hypersensitivity. Sumatriptan does, however, alleviate GTN-induced hypersensitivity in mice (21). Our finding may therefore question the validity of this PACAP38 model of migraine in mice or may be interpreted as additional evidence that PACAP mediated hypersensitivity is distinct from known pathways. Our results contradict the finding of a recent randomized clinical trial showing that sumatriptan prevented PACAP38-induced migraine attacks if administered intravenously and early in 37 migraine patients (47). Moreover, administration of sumatriptan reduced levels of PACAP measured in the external jugular vein during spontaneous migraine attacks (48), and in rodents, prolonged administration of triptan reduced brain mRNA transcription of PACAP (49). Noteworthy, not all patients respond to sumatriptan and targeting the PACAP signaling pathways may possibly be a relevant therapeutic target in such patients.
PACAP38 signaling pathways are distinct
We recently showed that repeated injections of PACAP38 in wildtype mice resulted in hind paw hypersensitivity that was independent of CGRP because mice lacking Ramp1, a crucial part of the CGRP receptor, and mice pre-treated with mAbs against CGRP could still be sensitized by PACAP38 (10). This contrasts with previous observations in mice sensitized by GTN and levcromakalim where CGRP antagonism was highly effective (22). Light aversion in mice as a surrogate for migraine-like photophobia was used to compare CGRP and PACAP38 (11). It showed that one-third of mice did not respond to PACAP38, which was not seen with CGRP. In the same study anti-PACAP38 mAbs blocked PACAP38-induced light aversion but not CGRP-induced light aversion. Conversely, anti-CGRP mAbs could not block PACAP38-induced light aversion. Thus, our present results are in keeping with a sizable previous literature suggesting that PACAP antagonism acts independently from other migraine triggers. This further suggests that PACAP mAbs may be effective in a small subgroup of migraine patients who do not respond to CGRP antagonist or sumatriptan. It also suggests that PACAP mAbs may advantageously be combined with CGRP mAbs.
Strengths and limitations
The strength of the study is the use of a well-validated mouse model for dissecting signaling pathways (12, 14, 17–20, 22). Our data were not biased by sedation or movement inability as we saw normal motor functions in all experimental groups.
The current study uses the classical routes of administration for mice (i.p. and s.c.), while human provocation studies primarily have used intravenous administration (infusions) (7, 50, 51). Furthermore, species differences in metabolism, pharmacokinetics, and -dynamics also need to be considered. The focus of the present study was to induce tactile hypersensitivity, not to mirror human dosing.
Another important methodological issue is the measurement of plantar versus cephalic sensitivity. In this study, only cutaneous mechanical sensitivity of the plantar area is used. The periorbital response is argued by some to be more fitting for this type of disorder than peripheral measurements (22). However, both methods are relevant in migraine research as hypersensitivity can be induced by GTN (19, 52), levcromakalim, cilostazol (22) and PACAP38 (10) both in the periphery and in trigeminal innervated dermatomes in mice. Furthermore, both plantar and periorbital response can be inhibited by migraine-specific drugs without general analgesic effects (53, 54) such as sumatriptan and olcegepant (12, 17, 19). In addition, the sensitivity to mechanical stimulation is different in the periorbital area and the vehicle-treated animals have lower 50% thresholds as compared to the hind paw thereby leaving a narrow effect window (10, 15). This also provides the challenge that subtle differences can be more difficult to detect, and the test therefore requires larger groups due to more variability (22, 52). Still, 168 mice were used for the present study. We employed only one behavioral endpoint, tactile sensitivity. Other endpoints such as light sensitivity or grimacing could be relevant but they do not always report equally over time and would require a much higher number of mice because of their higher variability (55).
We have not been able to replicate previous studies showing that PACAP and/or the PAC1 receptor is involved in mediating the effects of NO-donors. Therefore, the present study adds significantly to expand the preclinical portfolio on PACAP involvement in migraine models. Our interpretation is that if PACAP is involved in NO-induced signaling, it is not in the periphery, but centrally.