Cluster Headache and Hypoxia: Breathing New Life into an Old Theory, with Novel Implications

Abstract

Cluster headache is a severe, poorly understood disorder for which there are as yet virtually no rationally derived treatments. Here, Lee Kudrow's 1983 theory, that cluster headache is an overly zealous response to hypoxia, is updated according to current understandings of hypoxia detection, signaling, and sensitization. It is shown that the distinctive clinical characteristics of cluster headache (circadian timing of attacks and circannual patterning of bouts, autonomic symptoms, and agitation), risk factors (cigarette smoking; male gender), triggers (alcohol; nitroglycerin), genetic findings (GWAS studies), anatomical substrate (paraventricular nucleus of the hypothalamus, solitary tract nucleus/NTS, and trigeminal nucleus caudalis), neurochemical features (elevated levels of galectin-3, nitric oxide, tyramine, and tryptamine), and responsiveness to treatments (verapamil, lithium, melatonin, prednisone, oxygen, and histamine desensitization) can all be understood in terms of hypoxic signaling. Novel treatment directions are hypothesized, including repurposing pharmacological antagonists of hypoxic signaling molecules (HIF-2; P2X3) for cluster headache, breath training, physical exercise, high-dose thiamine, carnosine, and the flavonoid kaempferol. The limits of current knowledge are described, and a program of basic and translational research is proposed.

Keywords: cluster headache; genes; hypoxia inducible factor; neurochemistry; pharmacology; prevention; trigeminovascular.

Figure 1

Structures involved in hypoxia detection and response, and in cluster headache. Abbreviations: CA—carotid artery; CB—carotid body; CSN—carotid sinus nerve; CVLM—caudal ventrolateral medulla; GPN—glossopharyngeal nerve; NTS—nucleus tractus solitarius; PVN—paraventricular nucleus of the hypothalamus; RVLM—rostral ventrolateral medulla; SPG—sphenopalatine ganglion; SSN—superior salivatory nucleus.

Figure 2

The tricarboxylic acid cycle and electron transport chain in normoxia (A) and hypoxia (B). In hypoxia (1), the electron transport chain functions in reverse, generating reactive oxygen species at Complex 1 and Complex 3, and allowing NADH to accumulate at Complex 1; (2) Complex 2/succinate dehydrogenase is impaired in the forward direction, causing succinate to accumulate; (3) α-ketoglutarate is nonenzymatically converted to succinate; and (4) pyruvate is preferentially routed through pyruvate carboxylase to produce oxaloacetate, causing some reverse flux through the latter part of the tricarboxylic acid cycle. As a result of (2)–(4), succinate builds up. Reactive oxygen species, NADH, and succinate signal the presence of hypoxia. Abbreviations: C1 to C5—Complex 1 to Complex 5; ETC—electron transport chain; NAD—nicotinamide adenine dinucleotide; NADH—nicotinamide adenine dinucleotide (reduced form); ROS—reactive oxygen species; SDH—succinate dehydrogenase; TCA—tricarboxylic acid cycle.

Figure 2

The tricarboxylic acid cycle and electron transport chain in normoxia (A) and hypoxia (B). In hypoxia (1), the electron transport chain functions in reverse, generating reactive oxygen species at Complex 1 and Complex 3, and allowing NADH to accumulate at Complex 1; (2) Complex 2/succinate dehydrogenase is impaired in the forward direction, causing succinate to accumulate; (3) α-ketoglutarate is nonenzymatically converted to succinate; and (4) pyruvate is preferentially routed through pyruvate carboxylase to produce oxaloacetate, causing some reverse flux through the latter part of the tricarboxylic acid cycle. As a result of (2)–(4), succinate builds up. Reactive oxygen species, NADH, and succinate signal the presence of hypoxia. Abbreviations: C1 to C5—Complex 1 to Complex 5; ETC—electron transport chain; NAD—nicotinamide adenine dinucleotide; NADH—nicotinamide adenine dinucleotide (reduced form); ROS—reactive oxygen species; SDH—succinate dehydrogenase; TCA—tricarboxylic acid cycle.