Sage Journals: Discover world-class research

Abstract

Background

Cluster headache is the most severe primary headache disorder. A genetic basis has long been suggested by family and twin studies; however, little is understood about the genetic variants that contribute to cluster headache susceptibility.

Methods

We conducted a literature search of the MEDLINE database using the PubMed search engine to identify all human genetic studies for cluster headache. In this article we provide a review of those genetic studies, along with an overview of the pathophysiology of cluster headache and a brief review of migraine genetics, which have both been significant drivers of cluster headache candidate gene selection.

Results

The investigation of cluster headache genetic etiology has been dominated by candidate gene studies. Candidate selection has largely been driven by the pathophysiology, such as the striking rhythmic nature of the attacks, which spurred close examination of the circadian rhythm genes CLOCK and HCRTR2. More recently, unbiased genetic approaches such as genome-wide association studies (GWAS) have yielded new genetic avenues of interest including ADCYAP1R1 and MME.

Conclusions

The majority of candidate genes studied for cluster headache suffer from poor reproducibility. Broader genetic interrogation through larger unbiased GWAS, exome, and whole genome studies may provide more robust candidates, and in turn provide a clearer understanding of the causes of cluster headache.

Keywords

Genetics migraine GWAS

Introduction

Cluster headache (CH) is known as the most severe primary headache and is characterized by its striking rhythmic nature in relation to time of day, and time of the year (1). CH is diagnosed according to the International Classification of Headache Disorders (3rd edn) (2) and is classified as episodic or chronic depending on the duration of the attack-free periods separating the bouts. Patients report decreased quality of life and severe limitations in their ability to function (3,4). The impact is substantial for the 20–30% of patients that are non-responsive to currently available medications (5) and developing novel treatments for CH remains a challenge as the pathophysiology is incompletely understood (6). Understanding the underlying genetics can provide insight into the pathophysiology of the disease and facilitate the discovery of novel drug targets. Data from twin, family, and population studies provide support for genetic predisposition for this disorder, with first-degree relatives of CH patients having an estimated 5–18 times higher risk and second degree relatives having a 1–3 fold increased risk of developing CH than the general population (7). Given this evidence, many studies have attempted to identify the genetic components that contribute to the development of CH. This review provides a summary of these studies along with a brief overview of the pathophysiology of CH, which we compare to migraine due to the overlap of symptoms between both diseases.

Pathophysiology

The pathophysiology of CH is complex and still incompletely understood. CH is characterized as a neurovascular disease in which hypothalamic activation and the trigeminal-autonomic reflex are central to understanding the manifestation of CH.

Hypothalamic activation

The characteristic circadian and circannual periodicity of CH attacks and accompanying autonomic symptoms implicate the involvement of the hypothalamus in the disorder (1,8.9). The biological clock, which determines the periodicity of many physiological functions, is located in the suprachiasmatic nucleus in the hypothalamus. Additionally, the hypothalamus controls the autonomic nervous system; therefore, a dysfunction of the hypothalamus may be able to explain both the observed periodicity of CH attacks and the accompanying autonomic symptoms. Furthermore, the anterior hypothalamus, the site of the suprachiasmatic nucleus, is enlarged in CH patients compared to controls (10). Imaging studies have confirmed a role of the hypothalamus in CH attacks, as activation of the ipsilateral hypothalamic grey matter occurs during induced and spontaneous attacks (11,12). Whether the hypothalamus is responsible for initiating attacks, or its activation is a consequence of another mechanism, remains unknown. While activation of the brainstem (instead of the hypothalamus) was initially observed during migraine attacks (13), high-resolution functional magnetic resonance imaging has also confirmed the occurrence of hypothalamic activation during migraine attacks (14).

Trigemino-autonomic reflex

Cranial vascular tone is regulated by the trigeminovascular, parasympathetic, and sympathetic systems. The afferent arm of the trigeminal autonomic reflex contains neurons from the trigeminal nerve, which converge at the trigeminal ganglion (TG) and project to the trigeminocervical complex (15). From the trigeminocervical complex, sensory information is passed to the brain for integration and ultimately perception. Additionally, an efferent reflex activation of the superior salivatory nucleus is produced, which results in the cranial autonomic symptoms associated with CH. Neurons from the superior salivatory nucleus synapse in the sphenopalatine ganglion (SPG) and the facial nerve each have projections that act as parasympathetic effectors (16). Since the SPG has projections to the cranial vasculature, lacrimal glands, nasal glands, palatine glands, and pharyngeal glands, activation of the SPG results in vasodilation and cranial outflow (17), both of which are associated with CH. Like CH, involvement of the trigeminal and parasympathetic pathways and attack-associated vascular changes occur during migraine attacks (15).

Neuropeptide release

Patients undergoing spontaneous CH attacks have increased levels of the cranial neuropeptides calcitonin gene-related peptide (CGRP), vasoactive intestinal polypeptide (VIP) (18), and pituitary adenylate cyclase activating peptide (ADCYAP1/PACAP) (19). VIP, PACAP, and CGRP are all vasoactive and can thereby contribute to the neurovascular component of CH. CGRP and PACAP are present in the TG (20), SPG (21,22), and lamina of the trigeminal nucleus caudalis (TNC) (23), a constituent of the trigeminocervical complex. VIP is also present in the SPG (21), which is consistent with its role as a parasympathetic neurotransmitter. Due to the extreme pain associated with CH, altered nociception has been proposed as a contributing mechanism in CH. In support of this, application of CGRP onto neurons of lamina II–IV of the TNC, such as that which would occur from stimulation of the TG, accelerated firing of the neurons and decreased the current threshold for eliciting spike trains (24). This suggests that a lower stimulus threshold could activate the TNC in the presence of CGRP, resulting in normally innocuous stimuli being propagated and perceived as pain. Furthermore, PACAP elicits significantly increased CGRP release in lamina I/II of the TNC (p < 0.01), but not in the TG or dura mater, thereby contributing to the propagation of pain in CH (23).

Neurogenic inflammation

Another mechanism hypothesized to underlie CH involves neurogenic inflammation. Previously, CH was proposed to occur due to a recurrent inflammation of the cavernous sinus (25). Pathological findings on orbital phlebography are consistent with this; however, these findings are not exclusive for CH and presently there is little evidence to support this hypothesis (26). In rats, electrical stimulation of the TG resulted in CGRP release, mast cell degranulation, plasma protein extravasation and vacuolation in post-capillary venule endothelial cells, all of which are consistent with neurogenic inflammation (27). In addition to CGRP's role as a vasodilator, it also modulates cytokine release from satellite glial cells, which surround the TG, resulting in increased secretion of interleukin 1 beta (IL-1β) for example (28). A positive feedback loop between glia, the TG, and neurogenic inflammation may contribute to the pathophysiology of CH. In vitro, the pre-treatment of trigeminal neurons with IL-1β increased CGRP release from the neurons following stimulation, suggesting a mechanism for this cross-excitation and pronounced pain in CH (29).

Genetics of migraine

Migraine is an inherited disorder with an increased risk in first-degree relatives (30) and higher concordance in monozygotic twins (31). Interestingly, migraine has been described in familial association with CH (32,33). This familial association, in addition to the partial overlapping clinical features, suggests that there may be common genetic variants underlying both disorders. However, it is also important to note that misdiagnosis, or diagnostic delay, results in as many as 32% of female and 21% of male CH sufferers being classified as migraineurs (34). This will undoubtedly result in the inclusion of CH sufferers within the larger cohort-based migraine and meta-analysis studies, and therefore migraine genetic loci may represent CH loci, or loci that impact both migraine and CH (35). Studies implicating genes in migraine may therefore provide important clues for understanding the genetics of CH.

One of the biggest breakthroughs for understanding the genetics of migraine was the identification of genes that cause familial hemiplegic migraine (FHM). FHM is a rare subtype of migraine associated with attacks of migraine with aura (MA), migraine without aura (MO), and hemiparesis. This is an autosomal dominant disorder linked to three ion channel genes: voltage-dependent P/Q-type calcium channel subunit alpha-1A (CACNA1A) (36), sodium/potassium-transporting ATPase subunit alpha-2 (ATP1A2) (37), and sodium channel protein type 1 subunit alpha (SCN1A) (38). This raises the prospect that migraine and CH are consequences of channelopathies, and thus other ion channels, especially neuronally expressed ion channels, and ion channels known to impact pain, such as sodium voltage-gated channel alpha subunits 9 (SCN9A) (39,40) and subunits 11 (SCN11A) (41), could be candidates for migraine and CH susceptibility. The polygenic nature of common migraine (35,42) may also involve less extreme variants in the FHM genes, or variants in proteins that interact with the FHM gene-products (43). Genome-wide association studies (GWAS) have identified variants associated with migraine in three additional ion channels: Transient receptor potential cation channel subfamily M member 8 (TRPM8), solute carrier family 24 member 3 (SLC24A3), and potassium two pore domain channel subfamily K member 5 (KCNK5), which all deserve closer inspection in CH (43,44).

In addition to ion channels, migraine GWAS has revealed links to genes with a wide range of functions including ion homeostasis, vascular function, and pain processing (35). Presently, many of these variants are non-coding and concentrated in regulatory regions that could impact multiple genes (43). As genetic studies advance from using single nucleotide polymorphism (SNP)-based GWAS platforms assessing common variants, to sequence-based platforms where rare and coding variants will be more readily observable, it is expected that these loci will resolve to both causal genes and causal variants. These loci and the surrounding genes should become a focus for CH-targeted gene studies.

Genetics of cluster headache

Mode of inheritance

An early heritability study, which examined 370 probands from Denmark, demonstrated that CH inheritance was most likely multifactorial, with only 7% of cases being familial and that, at best, a single autosomal dominant locus would account for just 5% of cases (45), whereas other studies showed that autosomal recessive inheritance may be more predominant (46). Although the majority of CH cases are sporadic, we do not expect de novo mutations to account for a significant proportion of these cases, as CH presents after vertical gene transmission becomes viable, and fertility is not impacted.

Incomplete penetrance and male-female chronobiology are complicating factors of CH inheritance, and imply strong environmental and likely gene-by-environment interactions. This is most profoundly observed in the male-female ratio where males have an overall higher incidence of CH relative to females (2–4:1) driven in large part by sporadic cases, whereas females predominate within familial cases (7), as well as among sporadic-chronic cases that are early-onset (< 16 years of age) or late-onset (> 49 years of age) (47).

Targeted gene studies

The majority of CH genetic studies have focused on specific candidate genes inferred to be involved from the observed pathophysiology, or because they are known to cause headache or migraine disorders (Table 1). Preventive treatment options remain limited for CH sufferers; however, triptans and oxygen are often effective as abortive therapies to help reduce attack severity and duration (48), and mechanisms related to these therapies have also been drivers of candidate gene studies.

Table 1.

Targeted gene studies investigating genetic variants/markers at the molecular level in cluster headache.

Hypothesis/ mechanism	Gene	Country	Study type¹	Analysis²	Sample size	Genetic variant or marker	MAF	Results	Ref
Implicated in other headache/migraine conditions	mtRNA^Leu^(UUR)	Italy	CS	MS	47 CH cases	nt3243A>G		No mutations found.	50
		Germany	CS	MS	22 CH cases	nt3243A>G		No mutations found.	51
						nt3250T>C		No mutations found.
						nt3260A>G		No mutations found.
						nt3271A>G		No mutations found.
	CACNA1A	The Netherlands	FB	L	One pedigree (three affected)	Haplotype of D19S221, D19S1150, SCA-6, and D19S226		No disease haplotype identified.	52
			CS	MS	One CH patient	All exons		No mutations found.
		Sweden	CC	AA	75 CH cases, 108 controls	D19S1150		No association.	53
						SCA-6		No association.
	MTHFR	Germany	CC	AA	147 CH cases, 599 controls	c.677C>T (p.A222V; rs1801133)	0.31	No association.	55
Implicated in circadian rhythms	CLOCK	Italy	CC	AA	107 CH cases, 210 controls	rs1801260	0.26	No association.	57
		Italy	CC	AA	101 CH cases, 100 controls	rs1801260	0.26	No association.	58
		Italy	CC	AA	54 CH cases, 200 controls	rs1801260	0.26	No association.	59
			FB	S	Eight pedigrees (21 affected)	rs1801260	0.26	No segregation.
		Sweden	CC	AA	449 CH cases, 677 controls	rs1801260	0.26	No association.	63
						rs11932595	0.4	No association.
						rs12649507	0.3	MAF is more common in CH than controls (Bonferroni-corrected P = 0.02; OR = 1.29; 95% CI = 1.08–1.54).
		Chinese	CC	AA	112 CH cases, 192 controls	rs1801260	0.26	No association.	60
	HCRT	Italy	CC	AA	109 CH cases, 211 controls	rs4796717	0.74	No association.	68
						rs8072081	0.0025	No association.
	HCRTR1					c.264C>T (p.R37R; rs1056526)	0.49	No association.
						c.1375T>C (p.I408V; rs2271933)	0.55	No association.
	HCRTR2					c.1246A>G (p.I308V; rs2653349)	0.84	Carriers of the GG genotype vs. the GA/AA genotypes have a five-fold higher risk of developing CH (p = 0.0002, OR = 5.06).
						rs1027650	0.017	No association.
		Germany	CC	AA	226 CH cases, 266 controls	c.1246A>G (p.I308V; rs2653349)	0.84	Carriers of the GG genotype vs the GA/AA genotypes have a two-fold higher risk of developing CH (p = 0.0007, OR = 1.97).	69
		Denmark, Sweden, UK	CC	AA	259 CH cases, 267 controls	c.1246A>G (p.I308V; rs2653349)	0.84	No association.	71
						rs3122169	0.78	No association.
		Denmark, Italy, Sweden, UK	CS	MS	Eight CH cases	Complete protein-coding region and intron/exon boundaries		No mutations found.
		Italy	CC	AA	109 CH cases, 211 controls	rs10498801	0.21	No association.	68
						rs3122156	0.36	MCMC: Allele frequencies (p = 0.0053), genotype frequencies (p = 0.0088).
						rs9357855	0.18	No association.
						rs2653342	0.89	MCMC: Allele frequencies (p = 0.0082), genotype frequencies (P = 0.0136).
						rs3800539	0.19	No association.
						Haplotype		GTAAGG haplotype more frequent in CH patients than controls (OR = 3.68; 95% CI = 1.85–7.67).
			CS	MS	11 CH cases	Complete protein-coding region and intron/exon boundaries		No mutations found.
		Italy	CC	AA	54 CH cases, 200 controls	c.1246A>G (p.I308V; rs2653349)	0.84	No association.	59
			FB	S	Eight pedigrees (21 affected)	c.1246A>G (p.I308V; rs2653349)	0.84	No segregation.
		The Netherlands	CC	AA	575 CH cases; 874 controls	c.1246A>G (p.I308V; rs2653349)	0.84	No association.	70
		Chinese	CC	AA	112 CH cases, 192 controls	6 SNPs	0.3 to 0.8	No association.	60
	PER3	Norway	CC	AA	149 CH cases; 432 controls	VNTR (rs57875989)	Tri-allelic	No association.	64
Implicated in vascular constriction	ADH4	Italy	CC	AA	110 CH patients; 203 controls	rs1800759	0.51	No association.	73
						c.925A>G (p.I309V; rs1126671)	0.75	Carriers of the AA genotype vs. the GG/GA genotypes have a more than two-fold higher risk of developing CH (p = 0.006, OR (95% CI) = 2.33 (1.25–4.37)).
		Italy	CC	AA	54 CH patients; 200 controls	rs1800759	0.51	The WT A allele was more common in CH patients than in controls (p = 0.03) and the WT AA genotype was more common than the CA/CC genotypes in CH patients than controls (p = 0.03).	59
						c.925A>G (p.I309V; rs1126671)	0.75	The A allele was more common in CH patients than controls (p = 0.03).
			FB	S	Eight pedigrees (21 affected)	rs1800759	0.51	The WT AA genotype segregates with CH in two families.
						c.925A>G (p.I309V; rs1126671)	0.75	The homozygous AA genotype segregates with CH in two families.
		Sweden	CC	AA	390 CH patients; 389 controls	rs1800759	0.51	No association.	74
						c.925A>G (p.I309V; rs1126671)	0.75	No association.
		Chinese	CC	AA	112 CH cases, 192 controls	rs1800759	0.51	No association.	60
	NOS1	Sweden	CC	AA	91 CH cases, 111 controls	Dinucleotide repeat (NOS1a)		No association.	77
						Trinucleotide repeat (NOS1b)		No association.
	NOS2A					Bi-allelic AAAT/AAAAT repeat		No association.
						(CCTTT)_n pentanucleotide repeat		194-bp allele of the pentanucleotide repeat is more common in controls (p = 0.01, OR = 0.45); however, the frequency of this allele in CH patients is similar to its frequency in other previously studied populations.
	NOS3					(CA)_n repeat in intron 13		No association.
Implicated in iron metabolism	HFE	Italy	CC	AA	109 CH cases, 211 controls	c.845G>A (p.C282Y; rs1800562)	0.034	No association.	83
						c.187C>G (p.H63D; rs1799945)	0.11	No disease association. Age at onset of CH was later in carriers of the D63D genotype vs. the HH and HD genotypes (p < 0.001).
Implicated in respiratory/ smoking mechanism	SERPINA1	Germany	CC	AA	55 CH cases, 55 controls	F, M, S, and Z variants		No association. CH patients heterozygous for M allele had more attacks/day compared to homozygous M allele carriers (p = 0.02).	89
	CHRNA3 - CHRNA5	Italy	CC	AA	65 CH cases, 263 controls	rs16969968, rs6495306, and rs578776 Also Complete protein-coding region examined	0.26, 0.37, and 0.39	rs578776 showed nominal significance (p = 0.038)	90
Implicated in plasma trace amines	TAR1, TAR3, TAR4, TAR5, PNR, GPR58	Italy	FB	L	Two pedigrees (six affected)	Haplotypes		No disease haplotype identified.	80
Unbiased loci	ADCYAP1R1	Sweden	CC	AA	542 CH cases, 581 controls	rs12668955	0.43	No association	98
	MME					c.674G>C (p.G225A; rs147564881)	0.0015	No association.
	LRFN5 (nearest gene)					rs1006417	0.19	No association.

CC: case-control; FB: family-based; CS: case study. ²AA: allelic association; S: segregation; L: linkage; MS: mutation screening; MAF: Minor allele frequency.

The first targeted gene investigated for CH, the mitochondrial MT-TL1 (mtRNA^Leu(UUR)) gene, was affiliated with the disorder following a case report of a Japanese CH patient with a 3243A>G point-mutation (49). This mutation is typically found in patients with mitochondrial encephalopathy, myopathy, lactic acidosis, and stroke-like episodes syndrome (MELAS), which often presents with headaches early in life. However, the mutation was not detected in studies that examined 47 Italian CH patients (50) nor in 22 German CH patients (51). CACNA1A was examined due to its causative role in FHM (36); however, only negative results have been reported so far. This includes a linkage analysis of one family that failed to identify linkage or exon-sequence variants in the proband (52), and a case-control study in Sweden for two common alleles that were not enriched when compared to controls (53). While these two studies did not substantiate a link between CACNA1A and CH, this data is insufficient to exclude a role for CACNA1A in CH, and a closer examination of coding variants is warranted in larger populations. The 677C>T (p.A222V) missense variant in methylene tetrahydrofolate reductase (MTHFR) was investigated in a German case-control study because the polymorphism had previously been linked to migraine (54). However, the authors found no association between the MTHFR A222V polymorphism and CH (55).

The largest number of candidate genes investigated so far have focused on the unique circadian and circannual rhythmicity of CH attacks, and involvement of the hypothalamus in CH attacks (11). The circadian locomotor output cycles protein kaput (CLOCK) gene is essential for the function of the circadian system and is highly expressed in the suprachiasmatic nucleus of the hypothalamus. Multiple CH case-control studies and one family-based study investigated the role of the rs1801260 3'UTR SNP that is reported to modify the sleep cycle (56). However, no association or segregation has been identified (57–60). A Swedish population study investigated the role of three CLOCK variants (rs1801260 and two intronic SNPs implicated in sleep duration: rs11932595 and rs12649507) (61) in CH, and found that the minor allele of the rs12649507 variant was more frequent in CH patients than controls (p = 0.02) (62). The authors also demonstrated that CLOCK mRNA expression increased with the addition of each copy of the minor allele in fibroblast cultures established from cutaneous biopsies of patients compared to controls, which may provide a mechanism for how this variant alters susceptibility to CH (62). However, additional studies are needed to confirm these findings in independent cohorts. Period circadian CLOCK 3 (PER3), which also impacts the circadian rhythm and is highly expressed in the hypothalamus, was also investigated in a population-based Norwegian cohort. However, the variant examined, which is reported to correlate with diurnal preference in humans (63), did not associate with CH (64).

The hypocretins (orexins) are hypothalamic neuropeptides thought to have important roles in the regulation of sleep and arousal states (65), as well as roles in pain modulation, autonomic, and neuroendocrine functions (66). Genes coding for hypocretins are strong candidates for CH because hypocretin-containing cells are located exclusively in the posterolateral hypothalamus, which has been implicated in CH by imaging studies (11,12). Furthermore, levels of hypocretin are significantly lower in the cerebrospinal fluid of CH sufferers (67). Consequently, polymorphisms in the genes coding for hypocretin (HCRT), and its two receptors, hypocretin receptor 1 (HCRTR1) and hypocretin receptor 2 (HCRTR2), have been variously studied in CH patients. Investigations of variants in HCRT and HCRTR1 have not yet yielded any significant associations; however, a missense variant in HCRTR2 (1246A>G [p.I308V; rs2653349]) has been reported to be significantly linked to CH in several studies. This includes an Italian study where carriers of the homozygous I308V variant had an over five-fold increased risk of developing CH relative to wildtype or heterozygous carriers of the variant (p = 0.0002) (68). A German-population study likewise showed a significant association, albeit with only a two-fold increased risk in carriers of the homozygous genotype (p = 0.0007) (69). A recent meta-analysis that included 1167 CH patients and 1618 controls from six study populations also found a significant association of the I308V polymorphism with CH (p = 0.006) (70). However, this association has not proven to be universal, and has failed replication in four population-based studies (59,60,70,71) and in one family-based study (66).

Screening the complete protein-coding region and intron/exon boundaries of HCRTR2 did not reveal any further suspected mutations across 19 European CH patients.Given the strong positive and negative genetic results for this gene and the I308V variant in particular, it's likely that simple association results may have reached their limit in determining the importance of this variant. Further work is needed to understand the functional relevance of the isoleucine to valine substitution, and additional coding variants must be observed in CH patients in order to validate HCRTR2 as a true CH gene.

Candidate genes exploring the links with other mechanisms have been investigated; however, many of these candidates have only been reported in single studies. These mechanisms include vascular constriction and include the genes for alcohol dehydrogenase 4 (ADH4) and nitric oxide synthase (NOS), genes linked with smoking and respiratory mechanisms such as the trace amine receptors (TARs), alpha 1-antitrypsin (SERPIN1A), and the cholinergic receptor, nicotinic, alpha polypeptide 3 and 5 (CHRNA3-CHRNA5) genes, and candidates involved in iron homeostasis such as the hemochromatosis (HFE) gene.

The ADH4 gene is an interesting candidate for CH because of the disputed role of alcohol in this disorder. An Italian population-based study initially investigated the role of two ADH4 SNPs that have been associated with alcohol dependence (rs1800759 in the promoter, and 925A>G [p.I309V; rs1126671] in exon 7) (72) in 110 CH cases and 203 controls. There was no enrichment of the SNP in the promoter region in CH; however, wildtype carriers of the non-synonymous I309V variant had a more than two-fold increased risk of developing CH relative to heterozygous or homozygous carriers of the variant (p = 0.006) (73). This result was supported by a second Italian population-based study (p = 0.03), which also reported enrichment of the minor alleles for rs1800759 and rs1126671 in CH patients versus controls (p = 0.03 for both variants), although genotype frequencies were not significantly different (66). In contrast, a similar-sized study in Chinese CH cases failed to replicate the rs1800759 association (60), and a much larger Swedish study of 390 cases and 389 controls found no association between either of the ADH4 variants (74). Possible explanations for these contradicting results include population and environmental differences between the Italian, Chinese, and Swedish cohorts. However, the smaller Italian studies have an increased likelihood of amplifying the effects of subtle population stratification between the cases and controls, and therefore are more prone to false positive results (75). Moreover, while the rs1800759 variant in the promoter has been shown to alter promoter activity in transfected rat hepatoma cells (76), the functional consequence of the isoleucine to valine substitution induced by the 925A>G variant is unclear (73).

CH attacks can be precipitated in patients who are in-bout when nitroglycerine, a potent vasodilator and NO donor, is administered, pointing towards the NOS genes as possible candidates. Of the variants examined, two in NOS1, one in NOS2A, and one in NOS3 were not associated with CH, while a 194-bp allele of a pentanucleotide repeat in NOS2A appeared more common in controls than in CH patients (p = 0.01) (77). However, the allele frequency of this variant in the CH patients was similar to the frequency observed in other previously studied non-CH populations, suggesting that it is unlikely that the variant contributes greatly to CH susceptibility (77), and may be the result of population stratification between cases and controls (78).

Based on the proposed association between trace amines and primary headaches (79), a linkage analysis was reported for two Italian families using microsatellite markers in the 6q23 region. This region contains a gene-cluster for the G-protein-coupled receptors (TAR1, TAR3, TAR4, TAR5, PNR, and GPR58) that are activated by trace amines; however, the authors were unable to identify a haplotype that segregated with CH (80).

Due to reports linking elevated iron concentration in the pain-mediating structures of the brainstem with primary headaches (81), an Italian population-based study investigated the two missense variants in the homeostatic iron regulator (HFE) gene (p.H63D and p.C282Y) that are responsible for the majority of hereditary hemochromatosis cases (82). While the authors did not find an enrichment of either of the variants in CH, patients homozygous for the minor allele of the H63D variant had a significantly later age of onset of CH compared to both wildtype and heterozygous carriers of the variant (p < 0.001) (83). While this result has not yet been replicated, it provides credibility to the iron-causing hypothesis of CH, and other iron-pathway genes should also be examined as possible modifiers of CH.

A positive correlation between CH and smoking (84), and early reports of a correlation between CH and sleep apnea (85,86) which has not manifested in larger studies (1,87), led to the examination of candidates involved in chronic lung conditions. Common genetic variants in serpin family A member 1 (SERPINA1), a gene which has been associated to chronic lung conditions (88), were investigated in a German CH case-control study. While there was no association of the investigated SERPINA1 variants with CH, patients that were heterozygous for the ‘M’ allele had more frequent CH attacks compared to patients that were homozygous for the M allele (p = 0.02). Variants in the locus coding for the cholinergic receptor, nicotinic, alpha polypeptide 3 and 5 (CHRNA3-CHRNA5) genes, which have been strongly associated with smoking, chronic obstructive pulmonary disease, lung cancer and various other lung diseases, have also been examined. However, complete coding-sequence analysis between cases and controls did not reveal any mutations with a functional effect, and common variants in the untranslated regions were also not significantly enriched in CH sufferers after multiple correction (90).

Unbiased genetic approaches

Few studies have investigated the genetics of CH using unbiased genetic approaches (Table 2). The first study to implement this approach was a genome-wide linkage study of five Danish pedigrees. The authors reported suggestive linkage (NPL/lod ≥ 2.0) at four loci with peaks located on chromosomes (Chr)2 at 164cM and at 227cM, on Chr 8 at 27cM, and on Chr 9 at 14cM. However, none of these putative loci replicated when the analysis was extended to include 33 Danish and Italian pedigrees (71). The inability to identify a reproducible disease locus for CH reinforces the genetic heterogeneity underlying the inherited predisposition to CH (71). The second study to use an unbiased approach used a genome-wide comparative genomic hybridization (CGH) array to search for copy number variants (CNVs) in a small set of 10 sporadic Italian CH patients compared to seven controls. The authors identified two novel CNVs: A 63 bp deletion in intron 2 of the gene coding for thyrotropin-releasing hormone degrading enzyme (TRHDE), and a 200 bp deletion in intron 13 of the neurexin 3 (NRXN3) gene. However, both CNVs were present in controls, and neither segregated well in a follow-up study of five Italian families (59).

Table 2.

Unbiased approaches investigating genetic variants/markers at the molecular level in cluster headache.

Study type	Study design	Country	Sample size	Genetics variants or markers	Results	Ref
Linkage (whole genome)	Family based	Denmark	Five pedigrees (≥three affected/ pedigree)	404 microsatellite markers, average spacing of 9 cM	Potential linkage (NPL/lod ≥ 2.0): D2S1353, D2S1363, GATA151F02, and D9S2169.	71
Denmark, Italy	33 pedigrees (82 affected)	Microsatellite markers on Chr2, Chr8 and Chr9	No significant linkage.
Structural variant association (whole genome)	Case-control	Italy	10 CH patients; seven controls	CGH array: 170,334 60-mer oligonucleotide probes, average spatial resolution of 13 kb	Two novel rearrangements identified. TRHDE: 63 bp deletion (72,739,818-72,739,877) in intron 2 (present in 7/7 controls and in 6/10 CH patients). NRXN3: 200 bp deletion (80,110,451-80,110,510) in intron 13 (present in 0/7 controls and in 3/10 CH patients).	59
GWAS	Case-control	Italy	99 CH patients; 360 controls	566,178 SNPs	No single variant showed statistically significant association at the genome-wide threshold (p < 1.710⁻⁷). Suggestive associations: rs1006417 on chromosome 14 (p = 1.410⁻⁶) and rs12668955 in ADCYAP1R1 (p = 9.1*10⁻⁶).	91
GWAS: gene-based analysis			2568 rare (MAF < 0.05) protein altering SNPs (subset of the 566,178 SNPs) from 745 candidate genes with a plausible role in CH pathogenesis	Most significant gene: MME (p = 2.510⁻⁵). Suggestive associations: NPY1R* (p = 1.210⁻³), BDKRB1* (p = 3.310⁻³), SLC5A3* (p = 4.910⁻³), CCL26* (p = 7.010⁻³), and GRM1* (p = 9.5*10⁻³).

Finally, a genome-wide association study (GWAS) was recently performed on a cohort of 99 CH patients and 360 controls (91). While no variant showed statistical significance at the genome-wide threshold (p < 1.7 × 10⁻⁷), there were several variants with suggestive association. First, the rs1006417 SNP maps to an intergenic region on Chr 14 (p = 1.4*10⁻⁶), with the nearest coding gene being the leucine rich repeat and fibronectin type III domain containing 5 (LRFN5), an interesting neuronally expressed candidate that has been linked to other neurological conditions (92). Secondly, the SNP rs12668955 is an intronic variant of the PACAP receptor gene, pituitary adenylate cyclase activating polypeptide 1 receptor type I (ADCYAP1R1; p = 9.1*10⁻⁶). Finally, when focusing on rarer coding variants on the array, the SNP rs147564881, which causes a missense variant (674G>C; p.G225A) in the membrane metallo-endopeptidase (MME) gene, also called neprilysin, was reported with a probability of p = 2.5 × 10⁻⁵. This variant was found in 7/99 CH patients and 0/359 controls. Sequencing of the complete protein-coding regions of MME in the 99 CH patients did not reveal any additional rare variants except for a synonymous SNP (rs200455903) (91).

Interestingly, both ADCYAP1R1 and MME are known to have a pivotal function in pain mechanisms. As discussed above, PACAP is a peptide involved in pain processing and proposed to contribute to the pathogenesis of migraine (93), and loss of function in mice leads to altered circadian rhythms (94,95) as well as a decreased nociceptive pain response (96). Similarly, MME encodes an endopeptidase that hydrolyses peptides involved in the central regulation of pain, vasomotion, sleep-wake rhythms, and the modulation of trigeminal nociceptive signals (97). This study provides the first evidence that variants in genes involved in pain processing and circadian rhythms may indeed contribute to susceptibility to CH. However, a targeted Swedish study with 542 CH patients and 581 controls failed to confirm the suggestive association of the intergenic variant on Chr 14 (rs1006417), the intronic ADCYAP1R1 variant (rs12668955), or the rare G225A variant in MME (rs147564881) with CH (98).

Several studies have also investigated the genetics of CH using an unbiased approach by studying the gene expression profile of the disorder. A small study with three episodic CH patients and three controls used microarray technology to measure the gene expression profile in peripheral blood. The authors identified 90 differentially expressed genes including the upregulation of several S100 calcium-binding proteins and two human leukocyte antigen genes (HLA-DQA1 and HLA-DQB1). This supports a role for non-infectious inflammation and peripheral immune activation in the disorder (99). A second study with eight CH patients and 10 controls carried out a whole transcriptome analysis in lymphoblastoid derived cell lines. The authors identified 1172 differentially expressed genes, many of which have functions related to endoplasmic reticulum processing and spliceosome pathways. The most significantly altered gene, RNA-binding motif protein 3 (RBM3), interacts with many core circadian genes such as CLOCK and was upregulated in CH patients. Other interesting genes identified include the nuclear receptor subfamily 1 group D member 1 (NR1D1), which interacts with core circadian genes and is downregulated in CH patients, and tryptophan hydroxylase 1 (TPH1), which is important in serotonin metabolism and is upregulated in CH patients (100). Finally, a third study focused on RNA sequencing from peripheral blood to profile the gene expression of 39 CH patients and 20 controls. Contrary to the two previous studies, no single gene expression difference was significant between CH patients and controls (101); however, global tests showed that the overall gene expression profile differed between CH patients and controls. Specifically, at the functional gene set level, associations were observed between brain-related mechanisms such as GABA receptor function and voltage-gated channels. Modules of co-expressed genes also demonstrated a role for intracellular signalling cascades, mitochondria, and inflammation.

The distinct lack of overlap between the three gene expression studies is likely related to differences in study design and tissue sources. It may also reflect that peripheral blood and non-neuronal tissue sources are poor surrogates for tissues implicated in the pathophysiology of CH.

Pharmacogenetics

Pharmacogenetics is a growing field that aims to understand the impact of somatic mutations on therapeutic efficacy and toxicity. Several pharmacogenetic studies have been performed in CH to assess specific genetic loci with response to CH treatments. Of the studies that have been performed, all but one has focused on genes associated with triptan response, which is effective in up to 80% of sufferers. Given that triptans are agonists of the 5-hydroxytryptamine (5-HT)_1B/D receptors, which belong to the G protein coupled receptor superfamily, the role of a variant in the G protein β3 (GNB3) gene (C825T) was studied in a cohort of 231 CH patients (102). No genotype was associated with oxygen, verapamil, or corticosteroid response, and there was no significant difference in allele frequency distribution between triptan responders and non-responders. However, there was a significant difference reported in genotype distribution comparing triptan responders to non-responders: Heterozygous carriers of the GNB3 C825T allele were more likely to be triptan responders compared to carriers of the CC genotype (O.R. = 2.96; 95% C.I. = 1.34–6.56; p = 0.0074). While the C825T polymorphism has been associated with a variety of different diseases, the functional significance of the splice variant resulting from the C825T polymorphism is not yet clear. The splice variant may mediate enhanced signal transduction via G protein coupled receptors (103), which might account for the increased triptan response in C825T carriers if triptan's agonism of 5-HT_1B/D receptors triggers increased intracellular signals via GNB3 mediated processes (102). However, another study suggested the splice variant is biologically inactive, since it did not modulate N-type Ca²⁺ or G protein-gated inwardly rectifying K channels, unlike its wildtype counterpart in rat sympathetic neurons (104). In a genetic association study with 148 CH patients, variants in the serotonin-transporter-linked polymorphic region (5-HTTLPR), a promoter region of the primary serotonin transporter, encoded by the SLC6A4 gene, which clears serotonin from the synaptic cleft, were studied by restriction fragment length polymorphism (RFLP) analysis (105). No association was found between the two variants investigated (a 43-bp indel (rs4795541), and an A > G SNP (rs25531)) with response to triptans.

The one study that did not focus on serotonin pathways investigated the HCRTR2 G1246A polymorphism, which has a disputed link to CH, by RFLP analysis (106). In 184 CH patients, no association was found between response to first line treatments (triptans, oxygen, verapamil, and corticosteroids) and the HCRTR2 G1246A polymorphism.

Conclusions and future directions

Family and twin studies suggest a heritable component for CH; however, candidate gene studies, driven predominantly by the pathophysiology of CH, have failed to identify a consistently reproducible causal gene. Significant deficiencies of past CH genetics studies include small sample sizes, a lack of family-based genetic analyses, and a focus on common variants with no, or poorly understood functional relevance.

While there is solid rationale for selecting certain candidate genes based on pathophysiology, only a small fraction of genes for any mechanism have been examined genetically. For instance, there are more than 30 different genes known to impact circadian rhythms in animal models (107), yet only a handful of genes have been examined in CH patients (Table 1). This applies to all of the proposed mechanisms that contribute to CH.

Many of the issues of past CH genetics studies will be resolved as future studies naturally evolve to include whole-exome and whole-genome screening, collectively known as next-generation sequencing (NGS). This will permit the simultaneous examination of all genes known or suspected to be involved in specific pathways and mechanisms. For instance, all genes known to impact circadian rhythms can be examined simultaneously for their individual impact as well as their collective impact on CH. This will also allow for the examination of other candidate mechanisms and pathways that have thus far been neglected, such as pain pathways, which include SCN9A and SCN11A, which impact familial pain perception (108).

As discussed, a limited number of gene expression studies have also been performed to help elucidate the causes of CH attacks. While gene expression differences from hypothalamic tissue from CH patients would be ideal, this is clearly not practical. However, more finessed expression studies from plasma may yet yield crucial insights. CH attacks have characteristic circadian and circannual periodicity, and circadian rhythms are largely believed to be under the control of endogenous cyclic gene expression patterns driven by the activity of a core, highly conserved, transcription\translation feedback loop (TTFL) (109). A closer examination of gene-expression differences between day-night expression of cases and controls may yield more relevant causal candidate genes that interfere with, or are impacted by, an aberrant TTFL.

With larger epidemiological studies we may also begin to elucidate a clearer understanding of the chronobiology, gender imbalances, differences between chronic and episodic CH, and the large impact of environmental factors (1,34,110). Thus far, due in large part to small sample sizes, there has been minimal effort to partition the genetic underpinnings of these variables. For instance, given the expected diverse genetic heterogeneity of CH, treating chronic and episodic CH separately may yield more reproducible results. In addition, replication cohorts may need to be focused on equivalent geographic parallels to better control for the significant impact of daylight and annual rhythmicity on attack triggers, frequency, and severity (1). A focus on family-based genetic analyses may significantly reduce the contribution of these environmental factors that are confounding in CH cohorts.

In the continued absence of the known causal mechanisms of CH, the treatment options have evolved from clinical experience rather than from randomized controlled trials. Successful treatment is thus limited to abortive treatments with inhaled oxygen, or with triptans that are largely administrated subcutaneously or intranasally (48). Novel immunotherapies targeting CGRP are proving effective for migraine (111), and their efficacy on CH are being evaluated in four ongoing clinical trials (112), with early data indicating they may be effective for episodic, but not chronic CH.

Pharmacogenomics, which is the examination of genetic variants across the entire genome of therapeutic responders versus non-responders, could help improve efficacy of triptans and CGRP antagonists, reduce toxicity of these therapies by optimizing dose, shed more light on the pathophysiology and subtypes of CH, and ultimately help guide the identification of novel preventative therapies.

Due to the relative rarity of CH, NGS and pharmacogenomics studies may remain small and underpowered for some time. In the interim, a large number of candidate genes have been identified from very-large GWA studies for migraine sufferers, which undoubtedly contain a significant number of misdiagnosed CH sufferers (34,44). These candidate genes should be examined further in CH cohorts and families as they may represent a valuable source of therapeutic targets.

In conclusion, the insights from current migraine and headache genomics, along with future CH genomics, and revived pharmaceutical interest in addressing these highly debilitating attacks, points towards a brighter future for treating CH attacks.

Footnotes

Article highlights

Lack of association is seen in most candidate genes examined for cluster headache.

HCRTR2 variant I308V is the most replicated association identified, but requires functional characterization, and an understanding of why the association is negative in some well-powered studies.

Unbiased next generation and whole-genome sequencing may help elucidate better candidate genes for cluster headache, as seen in recent migraine GWAS studies.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This review was funded by Xenon Pharmaceuticals Inc. Kate Gibson and Anita Dos Santos received funding from the Natural Sciences and Engineering Research Council of Canada (NSERC) Experience Award.

References

Barloese

Lund

Petersen

, et al. Sleep and chronobiology in cluster headache. Cephalalgia 2015; 35: 969–978.

Headache Classification Committee of the International Headache Society (IHS). The International Classification of Headache Disorders, 3rd edition. Cephalalgia 2018; 38: 1–211.

D'Amico

Usai

Grazzi

, et al. Quality of life and disability in primary chronic daily headaches. Neurol Sci 2003; 24: s97–s100.

Palacios-Cena

Talavera

Lopez-Ruiz

, et al. Living with cluster headache: A qualitative study of patients' perspectives. Headache 2016; 56: 1171–1182.

Schurks

Kurth

de Jesus

, et al. Cluster headache: Clinical presentation, lifestyle features, and medical treatment. Headache 2006; 46: 1246–1254.

Leone

Cecchini

. Advances in the understanding of cluster headache. Expert Rev Neurother 2017; 17: 165–182.

Russell

. Epidemiology and genetics of cluster headache. Lancet Neurol 2004; 3: 279–283.

Barloese

Haddock

Lund

, et al. Chronorisk in cluster headache: A tool for individualised therapy?. Cephalalgia 2018; 38: 2058–2067.

Schürks

. Genetics of cluster headache. Curr Pain Headache Rep 2010; 14: 132–139.

10.

Arkink

Schmitz

Schoonman

, et al. The anterior hypothalamus in cluster headache. Cephalalgia 2017; 37: 1039–1050.

11.

May

Bahra

Büchel

, et al. Hypothalamic activation in cluster headache attacks. Lancet 1998; 352: 275–278.

12.

Sprenger

Boecker

Tolle

, et al. Specific hypothalamic activation during a spontaneous cluster headache attack. Neurology 2004; 62: 516–517.

13.

Weiller

May

Limmroth

, et al. Brain stem activation in spontaneous human migraine attacks. Nat Med 1995; 1: 658–660.

14.

Schulte

Allers

May

. Hypothalamus as a mediator of chronic migraine: Evidence from high-resolution fMRI. Neurology 2017; 88: 2011–2016.

15.

Goadsby

. Pathophysiology of migraine. Ann Indian Acad Neurol 2012; 15: 15–22.

16.

Spencer

Sawyer

Wada

, et al. CNS projections to the pterygopalatine parasympathetic preganglionic neurons in the rat: A retrograde transneuronal viral cell body labeling study. Brain Res 1990; 534: 149–169.

17.

Tepper

Caparso

. Sphenopalatine ganglion (SPG): Stimulation mechanism, safety, and efficacy. Headache 2017; 57: S14–S28.

18.

Goadsby

Edvinsson

. Human in vivo evidence for trigeminovascular activation in cluster headache. Neuropeptide changes and effects of acute attacks therapies. Brain 1994; 117: 427–434.

19.

Tuka

Szabó

Tóth

, et al. Release of PACAP-38 in episodic cluster headache patients – an exploratory study. J Headache Pain 2016; 17: 69.

20.

Eftekhari

Salvatore

Johansson

, et al. Localization of CGRP, CGRP receptor, PACAP and glutamate in trigeminal ganglion. Relation to the blood-brain barrier. Brain Res 2015; 1600: 93–109.

21.

Csati

Tajti

Kuris

, et al. Distribution of vasoactive intestinal peptide, pituitary adenylate cyclase-activating peptide, nitric oxide synthase, and their receptors in human and rat sphenopalatine ganglion. Neuroscience 2012; 202: 158–168.

22.

Csati

Tajti

Tuka

, et al. Calcitonin gene-related peptide and its receptor components in the human sphenopalatine ganglion – interaction with the sensory system. Brain Res 2012; 1435: 29–39.

23.

Jansen-Olesen

Baun

Amrutkar

, et al. PACAP-38 but not VIP induces release of CGRP from trigeminal nucleus caudalis via a receptor distinct from the PAC1 receptor. Neuropeptides 2014; 48: 53–64.

24.

Meng

Ovsepian

Wang

, et al. Activation of TRPV1 mediates calcitonin gene-related peptide release, which excites trigeminal sensory neurons and is attenuated by a retargeted botulinum toxin with anti-nociceptive potential. J Neurosci 2009; 29: 4981–4992.

25.

Hannerz

. Orbital phlebography and signs of inflammation in episodic and chronic cluster headache. Headache 1991; 31: 540–542.

26.

Remahl

Waldenlind

Bratt

, et al. Cluster headache is not associated with signs of a systemic inflammation. Headache 2000; 40: 276–282.

27.

Buzzi

Moskowitz

. Evidence for 5-HT1B/1D receptors mediating the antimigraine effect of sumatriptan and dihydroergotamine. Cephalalgia 1991; 11: 165–168.

28.

Thalakoti

Patil

Damodaram

, et al. Neuron–glia signaling in trigeminal ganglion: Implications for migraine pathology. Headache 2007; 47: 1008–1025.

29.

Capuano

De Corato

Lisi

, et al. Proinflammatory-activated trigeminal satellite cells promote neuronal sensitization: Relevance for migraine pathology. Mol Pain 2009; 5: 43.

30.

Russel

Olesen

. Increased familial risk and evidence of genetic factor in migraine. Br Med J 1995; 311: 541–544.

31.

Ulrich

Gervil

Kyvik

, et al. The inheritance of migraine with aura estimated by means of structural equation modelling. J Med Genet 1999; 36: 225–227.

32.

Kudrow

. Inheritance of cluster headache and its possible link to migraine. Headache 1994; 34: 400–407.

33.

Fusco

Colantoni

Geppetti

, et al. Re-occurrence of cluster headace in two families with significant migraine frequency. Cephalalgia 1999; 19: 423–424.

34.

Lund

Barloese

Petersen

, et al. Chronobiology differs between men and women with cluster headache, clinical phenotype does not. Neurology 2017; 88: 1069–1076.

35.

Gormley

Anttila

Winsvold

, et al. Meta-analysis of 375,000 individuals identifies 38 susceptibility loci for migraine. Nat Genet 2016; 48: 856–866.

36.

Ophoff

Terwindt

Vergouwe

, et al. Familial hemiplegic migraine and episodic ataxia type-2 are caused by mutations in the Ca2 + channel gene CACNL1A4. Cell 1996; 87: 543–552.

37.

De Fusco

Marconi

Silvestri

, et al. Haploinsufficiency of ATP1A2 encoding the Na(+)/K(+) pump alpha2 subunit associated with familial hemiplegic migraine type 2. Nat Genet 2003; 33: 192–196.

38.

Dichgans

Freilinger

Eckstein

, et al. Mutation in the neuronal voltage-gated sodium channel SCN1A in familial hemiplegic migraine. Lancet 2005; 366: 371–377.

39.

Cox

Reimann

Nicholas

, et al. An SCN9A channelopathy causes congenital inability to experience pain. Nature 2006; 444: 894–898.

40.

Goldberg

MacFarlane

MacDonald

, et al. Loss-of-function mutations in the Nav1.7 gene underlie congenital indifference to pain in multiple human populations. Clin Genet 2007; 71: 311–319.

41.

Leipold

Liebmann

Korenke

, et al. A de novo gain-of-function mutation in SCN11A causes loss of pain perception. Nat Genet 2013; 45: 1399–1404.

42.

Russell

Iselius

Olesen

. Inheritance of migraine investigated by complex segregation analysis. Hum Genet 1995; 96: 726–730.

43.

Silberstein

Dodick

. Migraine genetics: Part II. Headache 2013; 53: 1218–1229.

44.

Gormley

Winsvold

Nyholt

, et al. Migraine genetics: From genome-wide association studies to translational insights. Genome Med 2016; 8: 1–3.

45.

Russell

Andersson

Thomsen

, et al. Cluster headache is an autosomal dominantly inherited disorder in some families: A complex segregation analysis. J Med Genet 1995; 32: 954–956.

46.

De Simone

Fiorillo

Bonuso

, et al. A cluster headache family with possible autosomal recessive inheritance. Neurology 2003; 61: 578–579.

47.

Manzoni

Taga

Russo

, et al. Age of onset of episodic and chronic cluster headache – a review of a large case series from a single headache centre. J Headache Pain 2016; 17: 44.

48.

Robbins

Starling

Pringsheim

, et al. Treatment of cluster headache: The American Headache Society evidence-based guidelines. Headache 2016; 56: 1093–1106.

49.

Shimomura

Kitano

Marukawa

, et al. Point mutation in platelet mitochondrial tRNA(LEU(UUR)) in a patient with cluster headache. Lancet 1994; 344: 625.

50.

Cortelli

Zacchini

Barboni

, et al. Lack of association between mitochondrial tRNA(Leu)(UUR) point mutation and cluster headache. Lancet 1995; 345: 1120–1121.

51.

Seibel

Grünewald

Gundolla

, et al. Investigation on the mitochondrial transfer RNA(Leu)(UUR) in blood cells from patients with cluster headache. J Neurol 1996; 243: 305–307.

52.

Haan

van Vliet

Kors

, et al. No involvement of the calcium channel gene (CACNA1A) in a family with cluster headache. Cephalalgia 2001; 19: 959–962.

53.

Sjöstrand

Giedratis

Ekbom

, et al. CACNA1A gene polymorphisms in cluster headache. Cephalalgia 2001; 21: 953–958.

54.

Liu

Geng

, et al. MTHFR C677T polymorphism and migraine risk: A meta-analysis. J Neurol Sci 2014; 336: 68–73.

55.

Schurks

Neumann

Kessler

, et al. MTHFR 677C>T polymorphism and cluster headache. Headache 2011; 51: 201–207.

56.

Katzenberg

Young

Finn

, et al. A CLOCK polymorphism associated with human diurnal preference. Sleep 1998; 21: 569–576.

57.

Rainero

Rivoiro

Gallone

, et al. Lack of association between the 3092 T->C Clock gene polymorphism and cluster headache. Cephalalgia 2005; 25: 1078–1081.

58.

Cevoli

Mochi

Pierangeli

, et al. Investigation of the T3111C CLOCK gene polymorphism in cluster headache. J Neurol 2008; 255: 299–300.

59.

Zarrilli

Tomaiuolo

Ceglia

, et al. Molecular analysis of cluster headache. Clin J Pain 2015; 31: 52–57.

60.

Fan

Hou

Wan

, et al. Genetic association of HCRTR2, ADH4 and CLOCK genes with cluster headache: A Chinese population-based case-control study. J Headache Pain 2018; 19: 1.

61.

Allebrandt

Teder-Laving

Akyol

, et al. CLOCK gene variants associate with sleep duration in two independent populations. Biol Psychiatry 2010; 67: 1040–1047.

62.

Fourier

Ran

Zinnegger

, et al. A genetic CLOCK variant associated with cluster headache causing increased mRNA levels. Cephalalgia 2018; 38: 496–502.

63.

Dijk

D-J

Archer

. PERIOD3, circadian phenotypes, and sleep homeostasis. Sleep Med Rev 2010; 14: 151–160.

64.

Ofte

Tronvik

Alstadhaug

. Lack of association between cluster headache and PER3 clock gene polymorphism. J Headache Pain 2016; 17: 1–6.

65.

Sutcliffe

de Lecea

. The hypocretins: Excitatory neuromodulatory peptides for multiple homeostatic systems, including sleep and feeding. J Neurosci Res 2000; 62: 161–168.

66.

Kukkonen

. Physiology of the orexinergic/hypocretinergic system: A revisit in 2012. Am J Physiol-Cell Physiol 2012; 304: C2–C32.

67.

Barloese

Jennum

Lund

, et al. Reduced CSF hypocretin-1 levels are associated with cluster headache. Cephalalgia 2015; 35: 869–876.

68.

Rainero

Gallone

Valfrè

, et al. A polymorphism of the hypocretin receptor 2 gene is associated with cluster headache. Neurology 2004; 63: 1286–1288.

69.

Schurks

Kurth

Geissler

, et al. Cluster headache is associated with the G1246A polymorphism in the hypocretin receptor 2 gene. Neurology 2006; 66: 1917–1919.

70.

Weller

Wilbrink

Houwing-Duistermaat

, et al. Cluster headache and the hypocretin receptor 2 reconsidered: A genetic association study and meta-analysis. Cephalalgia 2015; 35: 741–747.

71.

Baumber

Sjostrand

Leone

, et al. A genome-wide scan and HCRTR2 candidate gene analysis in a European cluster headache cohort. Neurology 2006; 66: 1888–1893.

72.

Luo

Kranzler

Zuo

, et al. ADH4 gene variation is associated with alcohol dependence and drug dependence in European Americans: Results from HWD tests and case-control association studies. Neuropsychopharmacology 2006; 31: 1085–1095.

73.

Rainero

Rubino

Gallone

, et al. Cluster headache is associated with the alcohol dehydrogenase 4 (ADH4) gene. Headache 2010; 50: 92–98.

74.

Fourier

Ran

Steinberg

, et al. Screening of two ADH4 variations in a Swedish cluster headache case-control material. Headache 2016; 56: 835–840.

75.

Button

Ioannidis

JPA

Mokrysz

, et al. Power failure: Why small sample size undermines the reliability of neuroscience. Nat Rev Neurosci 2013; 14: 365–376.

76.

Edenberg

Jerome

. Polymorphism of the human alcohol dehydrogenase 4 (ADH4) promoter affects gene expression. Pharmacogenetics 1999; 9: 25–30.

77.

Sjöstrand

Modin

Masterman

, et al. Analysis of nitric oxide synthase genes in cluster headache. Cephalalgia 2002; 22: 758–764.

78.

Price

Zaitlen

Reich

, et al. New approaches to population stratification in genome-wide association studies. Nat Rev Genet 2010; 11: 459–463.

79.

D'Andrea

Terrazzino

Leon

, et al. Elevated levels of circulating trace amines in primary headaches. Neurology 2004; 62: 1701–1705.

80.

Aridon

D'Andrea

Rigamonti

, et al. Elusive amines and cluster headache: Mutational analysis of trace amine receptor cluster on chromosome 6q23. Neurol Sci 2004; 25: S279–S280.

81.

Welch

KMA

Nagesh

Aurora

, et al. Periaqueductal gray matter dysfunction in migraine: Cause or the burden of illness?. Headache 2001; 41: 629–637.

82.

Merryweather-Clarke

Pointon

Shearman

, et al. Global prevalence of putative haemochromatosis mutations. J Med Genet Genet 1997; 34: 275–278.

83.

Rainero

Rivoiro

Rubino

, et al. Prevalence of HFE (Hemochromatosis) gene mutations in patients with cluster headache. Headache 2005; 45: 1219–1223.

84.

Schurks

Diener

H-C

. Cluster headache and lifestyle habits. Curr Pain Headache Rep 2008; 12: 115–121.

85.

Chervin

Zallek

Lin

, et al. Sleep disordered breathing in patients with cluster headache. Neurology 2000; 54: 2302–2306.

86.

Ludemann

Frese

Happe

, et al. Sleep disordered breathing in patients with cluster headache. Neurology 2001; 56: 984.

87.

Barloese

MCJ

Jennum

Lund

, et al. Sleep in cluster headache – beyond a temporal rapid eye movement relationship?. Eur J Neurol 2015; 22: 656–664.

88.

Greene

Marciniak

Teckman

, et al. α1-Antitrypsin deficiency. Nat Rev Dis Primer 2016; 2: 16051.

89.

Summ

Gregor

Marziniak

, et al. Cluster headache and alpha 1-antitrypsin deficiency. Cephalalgia 2009; 30: 113–117.

90.

Cainazzo

Tiraferri

Ciccarese

, et al. O015. Evaluation of the genetic polymorphism of the α3 (CHRNA3) and α5 (CHRNA5) nicotinic receptor subunits, in patients with cluster headache. J Headache Pain 2015; 16: A88.

91.

Bacchelli

Cainazzo

Cameli

, et al. A genome-wide analysis in cluster headache points to neprilysin and PACAP receptor gene variants. J Headache Pain 2016; 17: 114.

92.

Mikhail

Lose

Robin

, et al. Clinically relevant single gene or intragenic deletions encompassing critical neurodevelopmental genes in patients with developmental delay, mental retardation, and/or autism spectrum disorders. Am J Med Genet A 2011; 155: 2386–2396.

93.

Tajti

Tuka

Botz

, et al. Role of pituitary adenylate cyclase-activating polypeptide in nociception and migraine. CNS Neurol Disord-Drug Targets 2015; 14: 540–553.

94.

Georg

Hannibal

Fahrenkrug

. Lack of the PAC1 receptor alters the circadian expression of VIP mRNA in the suprachiasmatic nucleus of mice. Brain Res 2007; 1135: 52–57.

95.

Hannibal

Georg

Fahrenkrug

. Altered circadian food anticipatory activity rhythms in PACAP Receptor 1 (PAC1) deficient mice. PloS One 2016; 11: e0146981.

96.

Jongsma

Pettersson

Zhang

, et al. Markedly reduced chronic nociceptive response in mice lacking the PAC1 receptor. Neuroreport 2001; 12: 2215–2219.

97.

Turner

Isaac

Coates

. The neprilysin (NEP) family of zinc metalloendopeptidases: Genomics and function. BioEssays 2001; 23: 261–269.

98.

Ran

Fourier

Michalska

, et al. Screening of genetic variants in ADCYAP1R1, MME and 14q21 in a Swedish cluster headache cohort. J Headache Pain 2017; 18: 1–4.

99.

Sjostrand

Duvefelt

Steinberg

, et al. Gene expression profiling in cluster headache: A pilot microarray study. Headache 2006; 46: 1518–1534.

100.

Costa

Squassina

Piras

, et al. Preliminary transcriptome analysis in lymphoblasts from cluster headache and bipolar disorder patients implicates dysregulation of circadian and serotonergic genes. J Mol Neurosci 2015; 56: 688–695.

101.

Eising

Pelzer

Vijfhuizen

, et al. Identifying a gene expression signature of cluster headache in blood. Sci Rep 2017; 7: 1–10.

102.

Schürks

Kurth

Stude

, et al. G Protein beta3 polymorphism and triptan response in cluster headache. Clin Pharmacol Ther 2007; 82: 396–401.

103.

Rosskopf

Koch

Habich

, et al. Interaction of Gbeta3s, a splice variant of the G-protein Gbeta3, with Ggamma- and Galpha-proteins. Cell Signal 2003; 15: 479–488.

104.

Ruiz-Velasco

Ikeda

. A splice variant of the G protein β3-subunit implicated in disease states does not modulate ion channels. Physiol Genomics 2003; 13: 85–95.

105.

Schürks

Frahnow

Diener

, et al. Bi-allelic and tri-allelic 5-HTTLPR polymorphisms and triptan non-response in cluster headache. J Headache Pain 2014; 15: 1–6.

106.

Schürks

Kurth

Geissler

, et al. The G1246A polymorphism in the hypocretin receptor 2 gene is not associated with treatment response in cluster headache. Cephalalgia 2007; 27: 363–367.

107.

Lowrey

Takahashi

. Genetics of circadian rhythms in mammalian model organisms. Adv Genet 2011; 74: 175–230.

108.

Sexton

Cox

Zhao

, et al. The genetics of pain: Implications for therapeutics. Annu Rev Pharmacol Toxicol 2018; 58: 123–142.

109.

Andreani

Itoh

Yildirim

, et al. Genetics of circadian rhythms. Sleep Med Clin 2015; 10: 413–421.

110.

Torelli

Cologno

Cademartiri

, et al. Application of the International Headache Society classification criteria in 652 cluster headache patients. Cephalalgia 2001; 21: 145–150.

111.

Edvinsson

Haanes

Warfvinge

, et al. CGRP as the target of new migraine therapies – successful translation from bench to clinic. Nat Rev Neurol 2018; 14: 338–350.

112.

Doesborg

Haan

. Cluster headache: New targets and options for treatment. F1000Research 2018; 7: 339.