Nitric oxide-induced headache may arise from extracerebral arteries as judged from tolerance to isosorbide-5-mononitrate

This study shows that the development of tolerance to NO is fast in the middle cerebral artery but slow in extracerebral arteries and that tolerance in headache similarly develops slowly.

Nitrate tolerance is a complex phenomenon, which involves neurohormonal counter-regulation, collectively classified as pseudotolerance, a well as intrinsic vascular processes, defined as vascular tolerance [13].

An isosorbide 5-mononitrate (5-ISMN) dose of 30 mg was chosen because a previously performed dose-response study showed that this dose induced significant changes in the radial- and temporal arterial diameters and no major drops in blood pressure or other severe adverse effects [4]. It is probably the maximum tolerable dose when given thrice daily because 5 out of 16 subjects discontinued the study early due to adverse events, see results. 5-ISMN t.i.d was chosen as the NO donor because of its long half life of 4–5 h [14, 15]. In addition, a controlled release formulation was used. Given thrice daily it results in a rather continuous exposure to NO thus enhancing the development of tolerance [16].

Regional cerebral blood flow (rCBF) was not measured in the present study, but in single-dose studies with GTN have shown either an unchanged rCBF in two SPECT (single photon emission computed tomography) studies [17, 18] and in one PET (positron emission tomography) study [19] or a 23% increase in one PET study [20]. Although the Doppler technique does not allow direct measurements of the diameter of the middle cerebral artery (MCA), the relative diameter change can be estimated from the relation: regional cerebral blood flow (rCBF) = mean velocity ×  cross sectional area of the artery [21]. Whether unchanged or slightly increased rCBF is present a decrease in V_mean, as in the present study, will indicate a dilatation of MCA. With an unchanged rCBF it can be estimated that 5-ISMN caused a 10% increase in the diameter of the middle cerebral artery (MCA). If the rCBF is increased the increase of the diameter of the MCA would be more pronounced. The decrease of V_mean of 17% is in the same range as found previously in healthy subjects (20%) after i.v. GTN 0.5 μg/(kg min) in a study where rCBF was measured simultaneously with SPECT and was unchanged [18]. This is somewhat less than the decrease of 28% observed with half the dose of GTN in migraine patients [22]. A comparable decrease (13%) in V_mean was found in another study in healthy volunteers after i.v. GTN 0.5 μg/(kg min) [23].

In one study transdermal GTN (0.6 mg/h) for 7 days was found to induce tolerance to GTN as measured by forearm blood volume [24] whereas pentaerythritol tetranitrate (60 mg three times a day) did not have this effect [24]. The few existing clinical studies, using direct measurements of diameter changes of the radial and temporal arteries, have failed to demonstrate the appearance of nitrate tolerance in these vessels within a study period of 7 [25] and 24–48 h of continuous GTN-infusion [26, 27]. The present study provides long-term data, up to a week, on this phenomenon in the radial and superficial temporal arteries for the first time. It is also the first clinical study of nitrate tolerance in cerebral arteries, the time course of which furthermore is compared to those of the temporal and radial arteries. At 4 h after administration of the first dose of 5-ISMN a significant dilatation was observed in both cephalic and limb arteries. Complete tolerance occurred already within 24 h of the first dose (a total of four doses) in the middle cerebral artery. At 168 h (day seven) partial tolerance was observed in the temporal and radial arteries compared to peak values but the radial was still dilated compared to baseline. The partial tolerance of the temporal and radial arteries may, however, have developed earlier than at 168 h as no measurements were done in the interval between 72 and 168 h. Previous studies [26‐28], investigating the effects of nitrates on the temporal and radial arteries, were not extended until the development of tolerance, but support nevertheless the present finding of a relatively late occurrence of tolerance in these vessels. Thus, during 7 h of GTN-infusion in healthy subjects the mean dilatation of the radial artery was 36% after 5 min of infusion of 0.5 μg/(kg min) and remained stable over time [25]. In one study, GTN 0.5 μg/(kg min) was infused for 24 h in healthy subjects; diameters of the radial and temporal arteries increased significantly after 1 h of infusion followed by an attenuation of this effect but tolerance was not observed [26]. In another study, GTN GTN 0.5 μg/(kg min) was infused for 48 h in healthy subjects with no attenuation of the dilator effect on the radial artery [27].

In vitro studies confirm a faster development of nitrate tolerance in the intracranial arteries compared to peripheral arteries [29, 30]. Thus, in smooth muscle cells cultured from peripheral vessels, pre-treatment with GTN was requested for 60 min in order to induce tolerance [29]. In contrast, in smooth muscle cells from cerebral arteries a 5-minute pre-treatment with GTN reduced the response to rechallenge by >50% [30]. Also, time to recovery from tolerance was shorter in cerebral artery smooth muscle cells than in peripherally derived cells [30]. Collectively these and other in vitro data [31‐33] demonstrate that vascular tolerance to NO-producing agents can be accounted for by actions at the cell level [31‐33]. Such intrinsic abnormalities include decreased biotransformation of organic nitrates to NO, desensitisation of the target enzyme guanylyl cyclase, enhanced vascular superoxide anion production, which degrades NO, and enhanced propensity for vasoconstriction due to increased endothelin expression within the vascular smooth muscle [31‐34]. Inhibition of mitochondrial aldehyde dehydrogenase (ALDH-2) has in several studies been shown to be involved in the development of tolerance to GTN [35‐38] and GTN tolerance was described as a syndrome characterized by mitochondrial dysfunction [37]. Bioactivation of 5-ISMN, however, presumably involves an ALDH-2-independent pathway [36] and alternative explanations for the tolerance to 5-ISMN should be investigated, see [13]. In the clinical setting other mechanisms of tolerance than those at a cellular level are considered to play a major role including a neurohormonal counter regulation with response of the sympathoadrenal axis and of the renin-angiotensin system that counteracts the effects of the drug [13, 34]. For a recent review on nitrate tolerance, see [13]. Cellular and systemic mechanisms have most likely a differentiated impact on arteries of different origin resulting in the different time-effect curves as observed in the present study.

A possible mechanism whereby NO could cause headache is by dilatation of cerebral and extracerebral blood vessels. This concept is supported by studies in which arterial dilatation during and after infusion of GTN or oral administration of 5-ISMN was consistently observed [4, 17, 18]. The present study demonstrated a significant dilatation of both MCA, temporal and radial arteries at 4 h after administration of the first dose of 5-ISMN. The 5-ISMN-induced headache was most severe during the first couple of days after 5-ISMN administration, followed by a gradual attenuation, until a state of tolerance to headache had developed on the 6th day [7]. A close temporal association between the disappearance of headache and attenuation of the 5-ISMN induced dilatation of the superficial temporal artery was observed (Fig. 1). In contrast, tolerance in the middle cerebral artery was complete and appeared already after 24 h, i.e. earlier than the development of tolerance to headache (Fig. 1). The fact that the temporal artery was still dilated to a minor degree after 7 days, whereas headache had disappeared indicates that there is most likely a threshold for NO-induced headache. These findings suggest that the locus of nociception in nitrate-induced headache could be related to dilatation of extracerebral arteries including possibly the middle meningeal artery, also a branch: of the external carotid artery [39].

The present study of NO-induced headache does not exclude the involvement of intracranial arteries in migraine that may be caused by dilatation independent pain-inducing effects of NO. Thus, migraine pain may possibly arise from normally calibrated arteries due to sensitization of perivascular sensory nerve terminals [40].

In conclusion, the development of nitrate tolerance is faster and more complete in cerebral arteries than in small-sized extracranial and middle-sized limb arteries. This may reflect differentiated impact of cellular and systemic mechanisms of tolerance in arteries of different size and origin. If vasodilatation is the cause of NO-induced headache, our results point to extracerebral arteries as the locus of nociception.