Abstract
A connection between nitric oxide (NO) and headache has been assumed because of the consistent observation that headache occurs as a side-effect of NO-donors (1). The observation that nitroglycerin (NTG), a classic NO-donor, acts as a trigger or provoking agent in cluster headache (CH) dates back to the 1960s (2). In the 1990s NO-donors were also shown to trigger delayed migraine attacks in those prone to the condition (3, 4). The latter investigators have also suggested that NO plays a role in chronic tension-type headache (TTH), as they have shown that nitric oxide synthase (NOS) inhibitors are effective in reducing headache in subjects with chronic TTH (5). Interestingly, this group have also extended this latter observation to migraine headache (6). Despite these clinical observations the role that NO plays in the pathophysiology of headache remains unclear.
Endogenous NO is generated by the conversion of the amino acid arginine to citrulline by three different NOS. Neuronal and endothelial NOS are constitutively expressed and produce relatively small quantities of NO in comparison with the inducible form of the synthase. Inducible NOS is primarily expressed in macrophages and is involved in non-specific immunity. Endothelial-derived NO, initially referred to as endothelial derived relaxation factor (EDRF), is a key player in vasomotor control. Neuronal-derived NO is ubiquitous throughout the nervous system and functions as a (retrograde) messenger and a modulator of neurotransmission. Thus NO is involved in many key physiological functions. Apart from regulating cerebral blood flow, it is thought to play a role in processing sensory information (7, 8) as well as in long-term potentiation (9) and pain sensitization (10).
In this issue of the journal, Costa and co-workers have attempted to address the role of NO in the pathophysiology of CH and its temporal relationship to the onset of pain in patients with CH. They analysed nitrates in the plasma of subjects with CH and healthy control after the ingestion of NTG. Nitrates were analysed at baseline and either at the peak of the headache in CH sufferers or 45 min after ingestion of NTG in control subjects. This latter time point coincided approximately with the peak of the pain in the subjects who developed CH. Finally, nitrates were measured after 120 min in both groups. Although baseline plasma citrulline was measured as a control parameter for the endogenous production of NO, it is unlikely that a small and well-localized increase in endogenous NO production, for example within the CNS, would be reflected by an increase in plasma citrulline as the systemic pool of free citrulline is massive compared with the CNS pool.
Baseline levels of nitrate and nitrite, and citrulline were similar in CH and control subjects. Similarly, the pharmacokinetic profile of plasma nitrates and nitrites in response to ingested NTG does not appear to differ between the two groups. Although this observation appears to be novel, the pharmacokinetic profile was derived from only three obser-vations, baseline, approx. 45 min and 120 min. Therefore subtle differences between the groups, particularly early on after the ingestion of NTG, would be missed. A more detailed temporal profile with more frequent plasma sampling would have been more desirable.
Triggering attacks with a large amount of exogenous NO from a NO-donor makes it difficult to draw any conclusions on changes in endogenous NO production. The large changes in plasma nitrite and nitrate levels as a result of the NO-donor would simply dwarf any increase in plasma levels from endogenous NO production. The baseline findings of Costa et al. have not confirmed the results of a previous study, in which subjects with CH had an increase in plasma nitrites regardless of whether or not they were suffering from a headache at the time (11).
Only limited conclusions can be drawn from this study because of the study design and the inherent difficulties of studying NO metabolism in vivo. NO has a very short half-life: it is either involved in one of many well-documented local nitrosylation reactions (12), converted to the highly reactive species peroxynitrite (13), or to the stable end-products nitrite and nitrate (14). Simply measuring plasma nitrites and nitrates ignores the other pathways involved in NO metabolism. In addition, levels of plasma nitrites and nitrates are also affected by alternative non-NO sources. The intake of nitrite and nitrate in food and water and the production nitrite and nitrate from enteric bacteria equal the amount that comes from endogenous production (15). To their credit, Costa and co-workers controlled for variations in food intake by measuring plasma nitrate and nitrite after an overnight fast while subjects were on a nitrite-poor diet. Despite these precautions, changes in plasma nitrate and nitrite may not directly reflect the amount of endogenously produced NO. The high fractional excretion of nitrate and nitrite via renal ultrafiltration may mask the detection of small physiological fluctuations in endogenous NO production. For this reason urine may prove a more sensitive biological fluid for studying small changes in systemic NO production in patients with headache.
Unfortunately, urine and plasma obtained from peripheral blood as suggested above are both incapable of pinpointing the source of the increase in NO production. Specific experiments using selective venous sampling and/or animal models would be required to answer these questions.
In summary, the only conclusion that can be drawn from this study is that the pharmacokinetics of plasma nitrite and nitrate after the ingestion of the NO-donor GTN appear to be similar in both CH subjects and controls. The questions of whether endogenous NO production is altered in subjects with CH and the role that NO plays in triggering CH remain unanswered. Further studies with larger numbers of subjects and more frequent sampling are required.