Transport of the pituitary adenylate cyclase-activating polypeptide across the blood-brain barrier: implications for migraine

The peptide pituitary adenylate cyclase-activating polypeptide (PACAP) is widely distributed in the nervous system and is found in two major isoforms, PACAP38 and PACAP27, which exert various effects via activation of the VPAC1, VPAC2 and PAC1 receptors [1]. A growing body of evidence suggests that PACAP plays an important role in migraine pathogenesis. For example, intravenous infusion of PACAP38 provokes migraine attacks in migraine patients without aura [2, 3] and induces a marked sustained dilation of extra-cerebral but not intracerebral arteries in both healthy volunteers [4] and migraine patients [5]. The exact mechanisms of PACAP38 induced migraine are unknown. However, given the abundant distribution of PACAP in the CNS [6], central pain mechanisms caused by exogenous PACAP infusion could hypothetically be important. This raises the important question to which degree PACAP may cross the blood-brain barrier (BBB). Preclinical studies have reported an active transport system of PACAP38 across the BBB in the rat [7]. However, investigations of rat and human cerebral arteries in vitro, showed that PACAP38 only had effect after abluminal but not luminal application [8]. Understanding the function of the BBB in relation to PACAP is important to the understand the mechanisms behind PACAP-induced migraine attacks triggering, but also to develop antimigraine drugs targeting the PACAP receptors [9]. Here, we aimed to present a systematic review on studies investigating the transport of PACAP across the BBB.

We performed a systematic literature search in July 2017 via PubMed to identify studies reporting original data on PACAP in relation to the BBB. The following search terms were used: pituitary adenylate cyclase activating polypeptide or PACAP and blood brain barrier or BBB. No filters were used in the search. Only studies reporting original data published in English-written peer-reviewed journals were included. The first author assessed all titles and abstracts to identify articles containing relevant data. Subsequently, the entire articles were read and additional studies known to be relevant were also included.

Our search strategy resulted in 96 hits out of which 11 studies were included in the final review. In addition, two studies known to be relevant by the authors were also included ending up with a total of 13 studies (Tables 1 and 2). Two out of 13 studies exclusively examined human arteries [4, 5], while one study reported data on rat and human arteries [8]. The remaining ten studies investigated mice [10‐18] or rat models [19]. Only one of the studies were performed as migraine models [5].

The main finding in this review is that PACAP can cross the BBB in both directions (i.e. from blood to brain and from brain to blood). Preclinical studies carried out in mice or rats consistently report that PACAP38 is transported actively from the blood to the brain (influx) and from the brain to blood (efflux) by the protein transport system-6 (PTS-6), which is located in the endothelium. In contrast, studies investigating effect of PACAP38 on the cerebral arteries (i.e. middle cerebral artery [MCA]) suggest that PACAP38 is not able to cross the endothelium in sufficient amount to activate receptors in the smooth muscle cells in the arterial walls [4, 5, 8], which will be discussed in the following. Historically, the BBB was considered as a structural barrier between the blood and brain consisting of capillary endothelial cells glued together with tight junctions and surrounded by glial cell projections. We now know that the BBB is a functional rather than just a solid structural barrier, allowing passage in from blood to brain and from brain to blood via different mechanisms. Peptides that cross the BBB are either actively transported by saturable carrier- or receptor-mediated mechanisms or by non-saturable transmembrane or intercellular diffusion. Banks and colleagues [10] first demonstrated that both PACAP27 and PACAP38 could cross the BBB. Interestingly, the permeability for PACAP38 decreased as the concentration of PACAP38 increased in the blood, which was not true for PACAP27. Based on these initial observations authors suggested that PACAP38 traversed the BBB via a saturable mechanism, whereas the PACAP27 uptake must be via a non-saturable mechanism [10]. Ki (influx constant) for PACAP38 (2.86 μL/g/min) was slightly higher than and PACAP27 (2.13 μL/g/min) although PACAP38 is a larger and less lipid soluble molecule. However, the percentage of intact iodine-labelled PACAP27 (I-PACAP27) in the brain was 76.9% versus 58.9% for the PACAP38 [10]. Taken together, these observations suggest efflux of PACAP27 from the brain to blood via the saturable beta-F1 ATPase, which is highly specific for PACAP27 [16]. Interestingly, the passage of PACAP38 was not affected by a general disruption of the BBB by lipopolysaccharide [15], but after occlusion of the MCA [19].

Two human MR-angiography studies found no functional effect of intravenously infused PACAP38 on the MCA but a 17–19% and 2 h sustained dilation of the middle meningeal artery (MMA) in healthy volunteers [4] and migraine patients [5]. Interestingly in vitro studies reported presence of PACAP receptors in the meningeal as well as cerebral arteries [20, 21]. Thus, the selective dilation of the MMA but not the MCA may suggest that intravenously administrated PACAP38 most likely did not cross the BBB. Another human study has shown that infusion of PACAP38 can induce premonitory symptoms in 48% of patients with induced attacks [22], which may be a central effect via the hypothalamus [23]. However, PACAP38 infusion did not induce more premonitory symptoms in patients who developed an attack compared with those who did not develop an attack, which suggest that induction of premonitory symptoms is not a necessary mechanism for the induction of migraine. Moreover, PACAP38-induced migraine attacks are accompanied by change in the intrinsic brain connectivity [24]. These observations suggest either that there is a difference between species or that the efflux mechanisms for PACAP38 work even faster in man to transport PACAP38 from the endothelial cells back to the blood. Thus, PACAP cannot exert its effect although it can cross the BBB or enter the endothelial cells. In contrast, Nonaka et al. [14] reported the highest amount of iodine-labelled PACAP38 (I-PACAP38) in for instance the hypothalamus. Alternatively, the PTS-6 system may be limited to the capillary endothelium and not the large arteries of the brain, where the endothelium may work as a structural barrier denying access to PACAP38 from the blood to the smooth muscle cell where the PACAP-receptors are located. PACAP-receptors are found in the entire brain and one study reported uptake of I-PACAP38 in all brain regions after intranasal administration [17]. It would be interesting to directly compare effect on the central nervous system and cerebral arteries after intravenous and intranasal administration.

It has previously been speculated and suggested that the BBB might be more permeable during migraine attacks [25]. However, recent advanced MRI using intravenous gadolinium contrast performed during and outside of attacks of migraine with [26] and without aura [27] reported no significant change in the BBB permeability. The clinical impact of determining whether intravenous PACAP has a CNS effect is related to the potential of developing PACAP or PACAP receptor antibodies as treatment. Currently, there is an ongoing PAC1 receptor antibody trial for migraine treatment (ClinicalTrials.gov Identifier: NCT03238781).

PACAP38 transport from the blood to brain and from brain to blood is saturable carrier-mediated, whereas PACAP27 crosses the BBB from the blood to brain by transmembrane diffusion. The transport of PACAP27 the brain to blood is mediated by a saturable and peptide-carrier. There is no solid evidence yet of PACAP38 passing the BBB following exogenous infusion in migraine models that can lead to brain function and brain vessel changes, so it is possible that the PACAP migraine inducing effect is caused by peripheral mechanisms, which could be via inducing changes in the meninges or affecting nociceptors in extracranial vessels both devoid of the BBB. Future advanced in vivo human studies and exploration of PAC1-receptor antibodies using radio-labelled tracers are highly needed to clarify how and where PACAP may lead to the development of migraine pain.