Migraine is a common episodic brain disorder involving persistent changes in multiple body systems during attacks as well as the interictal period. Two main types of migraine exist: migraine with aura and migraine without aura (Headache Classification Committee of the International Headache Society (IHS), 2013). Familial hemiplegic migraine (FHM) is a rare monogenic type of migraine with aura for which several genes have been identified (van den Maagdenberg et al., 2007) and the underlying neuronal mechanisms are well-characterized (Ferrari et al., 2015).

Traditionally, much attention is given to a neuronal or vascular origin of migraine (Pietrobon and Moskowitz, 2013), however, evidence is accumulating that an aberrant immune response may also play a role in disease pathogenesis (Levy, 2009). Recently, it was suggested that a shifted profile of T-cells in migraine patients would point to a possible autoimmune nature of the disease (Arumugam and Parthasarathy, 2016). Furthermore, studies implicated several pro-inflammatory cytokines, such as IL-1 and IL-6, to be involved in migraine (Uzar et al., 2011; Zhang et al., 2012). These cytokines potentially target a wide range of cells including leukocytes, trigeminal nociceptors, and neuroglia to exert their effect on migraine pathophysiology (Zhang et al., 2012). Also, cytokines can promote the release of the peptide calcitonin gene-related peptide (CGRP) produced in neurons (Neeb et al., 2011), an important migraine neuroimmune mediator that further facilitates a migraine attack. It has been shown that CGRP, can stimulate the secretion of various cytokines by T-cells (Levite, 1998; Cuesta et al., 2002) suggesting a bidirectional crosstalk between cytokines and migraine-relevant neuropeptides. Notably, release of CGRP could contribute to increased cortical excitability (Tozzi et al., 2012) along with disrupted neurovascular coupling in migraine (Jacobs and Dussor, 2016).

Cytokines released by leukocytes are known to affect the function of neurons in the CNS and contribute to the regulation of the pain threshold. For example, headaches were documented after tumor necrosis factor-alpha (TNFα) injection (Schiller et al., 1991). Also, migraine was associated with elevated serum levels of TNFa and IL-1 (Perini et al., 2005). A study of cytokines in cerebrospinal fluid (CSF) of migraine patients revealed an increased level of IL-1, receptor antagonist (IL-1ra), monocyte chemoattractant protein-1 (MCP-1), and transforming growth factor-1 (TGF-1) (Bo et al., 2009). However, in contrast to other neurological diseases, such as multiple sclerosis (Khaibullin et al., 2017) or—more closely related to migraine—stroke (Chamorro et al., 2012), in which a neuroinflammatory component of disease pathology is well-known, our knowledge on the role of cytokines in migraine is limited. Changes in cytokine level in CSF found in neurodegenerative diseases such as multiple sclerosis suggested that cytokines could play role in managing brain damage and determine cortical excitability. Thus, microglia, the immune cells that reside in the brain and are the major source of pro-inflammatory cytokines, co-determine the development of cortical spreading depression (CSD) (Grinberg et al., 2013; Pusic et al., 2014), which is the electrophysiological correlate of the migraine aura. Studies in transgenic mouse models with human pathogenic FHM mutations (Ferrari et al., 2015) have highlighted specific genetic and neurobiological mechanisms that can explain the spectrum of clinical symptoms, such as the aura and the cerebellar ataxia seen with specific mutations, suggesting that FHM may also be an attractive condition to understanding better the contribution of the cytokines in migraine.

In this study, we present the profile of multiple cytokines along with a genetic and neurological characterization of a small FHM family from Russia. Whole-exome sequencing revealed a R583Q missense mutation in the CACNA1A gene in the two most affected persons of this family, i.e., those that had hemiplegic migraine and cerebellar atrophy and ataxia. The other two patients did not carry the mutation and were diagnosed as patients with migraine with aura. Migraine patients of this family have high levels of specific cytokines, some of which are high only in patients with the FHM1 mutation, supporting neuroinflammation in migraine possibly with a specific cytokine signature in those with hemiplegic migraine and cerebellar ataxia.

Here we present a small Russian FHM1 family with two patients (D and F) who carry the R583Q missense mutation in the CACNA1A gene. Both patients suffered from hemiplegic migraine with aura and showed neurological cerebellar changes with cerebellar atrophy, as supported by neuroimaging.

The R583Q mutation is a previously described pathogenic mutation, often (Ducros et al., 2001), but not in all cases (Terwindt et al., 2002), associated with cerebellar ataxia, and therefore the likely cause of disease in patients D and F. The mutation affects the voltage sensor of the S4 helix of the α_1A subunit of voltage-gated Ca_V2.1 (P/Q-type) Ca²⁺ channels and results in a shift of the voltage dependence of activation and inactivation to more negative potentials with a slower recovery from inactivation (Kraus et al., 2000). Given the highest level of expression of Ca_V2.1 channels in cerebellum (Fletcher et al., 1996) it is long recognized that the gain-of-function effect of R583Q-mutated channels can explain cerebellar atrophy and cerebellar ataxia in patients. The link with various forms of ataxia and mutations in CACNA1A is rather complex, suggesting that neuronal calcium influx through Ca_V2.1 channels is tightly controlled. Notably, some, but not all, gain-of-function FHM1 mutations are associated with chronic cerebellar ataxia (Ferrari et al., 2015), whereas loss-of-function mutations in the same gene, exerting dominant-negative effects (Jeng et al., 2006; Dorgans et al., 2017), cause episodic ataxia type 2 (EA2) (Ophoff et al., 1996). Finally, specific C-terminal CAG expansions, with complex functional effects involving e.g., the aggregation of alternative spliced C-terminal fragments (Du et al., 2013; Mark et al., 2015), cause late-onset cerebellar ataxia, namely spinocerebellar ataxia type 6 (SCA6) (Zhuchenko et al., 1997; Giunti et al., 2015).

A novelty of the study was the extensive profiling of multiple cytokines in migraine patients, including those with an FHM1 gene mutation. One of key questions is whether dysregulation of cytokines could directly or indirectly contribute to the main pathological changes in our patients including the hemiplegic migraine and the cerebellar atrophy that is linked with the ataxia.

We found an increased serum levels or IL-6 in both FHM1 patients (D and F), as well as in the two migraine with aura patients without the genetic mutation (B and S). IL-6 is an abundant prototype inflammatory cytokine, shown to be elevated in patients with the common forms of migraine (Uzar et al., 2011; Zhang et al., 2012). In animal experiments, this cytokine can enhance excitability of dural afferent nerves and produce migraine-related allodynia (Yan et al., 2012). Along with IL-6, pro-inflammatory cytokine IL-18 is also up-regulated in serum of all migraine patients, albeit stronger in patients with the gene mutation. Consistent with the higher level of IL-6, which is a strong activator of CXCL1 (Roy et al., 2012), we found high levels of CXCL1 in all but one (patient S) migraine patients. CXCL1, like IL-6 is a pro-inflammatory and pro-nociceptive cytokine (Wang et al., 2008; Yang et al., 2016) that is co-expressed in sensory neurons together with neuropeptide CGRP (Cao et al., 2016), which is a major migraine mediator (Tfelt-Hansen and Olesen, 2011).

High serum levels of LIF and MIF were consistently found in FHM1 patients (but not in patients without the gene mutation). LIF is a member of the IL-6 family of cytokines, which can act directly on small-size nociceptive trigeminal neurons (Tamura et al., 2003). LIF also promotes the growth of mast cells (Hiragun et al., 2000; Gyotoku et al., 2001), which are considered to play a triggering role in pathogenesis of migraine (Levy et al., 2006; Kilinc et al., 2017). MIF is a pro-inflammatory cytokine expressed in the nociceptive system in neurons and glial cells that may maintain inflammatory states (Vera and Meyer-Siegler, 2003). Thus, combined pro-nociceptive and pro-inflammatory action of cytokines IL-6, IL-18, CXCL1, LIF, and MIF could occur in the trigeminovascular system, which comprises of peripheral trigeminal neurons and vessels in meninges located outside the brain blood barrier, that might explain specific migraine related changes in our patients. Previous studies in experimental animal studies suggested a possible role of enhanced neuronal excitability in the brain of carriers of an FHM1 gene mutation (van den Maagdenberg et al., 2004; Nair et al., 2010; Pietrobon and Moskowitz, 2013). Our current study extends this view suggesting the influence of pro-inflammatory cytokines on enhanced neuronal excitability and, possibly, aura features in patients.

The observed pronounced up-regulation seen with certain cytokines in the two FHM1 patients may reveal a contribution to the enhanced neuronal hyperexcitably and cerebellar atrophy in FHM1 (Ferrari et al., 2015). For instance, pro-apoptotic TRAIL (TNF-related apoptosis-inducing ligand) can induce degeneration of neurons (Kichev et al., 2014) whereas down-regulation of TRAIL itself and of its death receptors with specific antibodies ameliorates neuronal damage (Cantarella et al., 2014). However, it cannot be excluded that abnormal cytokine levels may in fact reflect some secondary effect, for instance they may occur as consequence of long-lasting multiple pathological changes and/or recurrent pain attacks in patients D and F.

A major limitation of the study is the very small number of investigated patients; too small even to apply statistics. Large FHM families are difficult to obtain—certainly in a country the size of Russia with patients spread over a large geographical distance—and this study has its merits as hypothesis-generating. Our study suggests previously not recognized associations of cytokines in migraine pathophysiology. Whether these cytokines actually play a role should be assessed in larger studies. Any neuroinflammatory role of such abnormal cytokines may be investigated in experimental animal models.

Genome-wide association studies that systematically test for association between single nucleotide polymorphisms (SNPs) (as they also occur in cytokine-encoding genes) and a phenotype have not surfaced cytokine genes among the top migraine hits in the largest migraine GWAS (that consisted of over 59,000 patients) (Gormley et al., 2016). Still, a genome-wide analysis of blood gene expression in 83 migraine patients (and a similar number of controls) for which data of close to 18,000 expression probes were measured, although they did not detect a genome-wide significant hit for a cytokine at the individual gene level, strongly suggests the relevance of cytokines in migraine at the pathway level, as highlighted by significant hits with relevant canonical pathways that include “immune system,” “interferon signaling,” and “cytokine signaling in immune system” (Gerring et al., 2017). Lack of finding a significant association for a cytokine at the SNP level is perhaps not surprising given that a GWAS study that investigated the genetic underpinning of concentrations of 41 circulating cytokines (of those only SCGF-β, IL-18, TRAIL, HGF, Il-2ra, IL-16, MIF were investigated in our study) in whole blood only 27 genome-wide significant loci were identified, which for all loci explained only very little of the variance (Ahola-Olli et al., 2017), so many (non-)genetic factors that co-determine circulating levels of these cytokines remain unknown.

Collectively, the serum cytokine profile identified in two patients with genetically confirmed FHM1 suggests novel contributors to main disturbances implicated in the pathogenesis of this type of migraine. High levels of certain pro-apoptotic cytokines may have contributed to the cerebellar degeneration seen in the mutation carriers, whereas a concerted up-regulation of pro-nociceptive cytokines may reflect more migraine as phenotype. These results suggest that combined genetic and non-genetic neuroimmune mechanisms could be involved in the complex clinical phenotype of monogenic and other types of migraine.

SK: writing the manuscript, specifically the cytokine data analysis, and discussion; intellectual contribution into the discussion; EGM: Selection and recruitment of patients; collection and interpretation of clinical data, and writing the manuscript; LS and GG: whole-exome sequencing, data analysis; ES: study design, exome data analysis; AN: collecting and preliminary analysis of whole exome data; EM: Luminex analysis of cytokine levels; YD: isolation of genomic DNA from patient samples; Quality control and preparation for molecular genetics analysis; EB: collection and interpretation of clinical data; OG: analysis of whole-exome data, wrote the manuscript; AV and RG: design of the study, writing the manuscript; AR: Design of molecular genetics and cytokine profiling experiments. Writing the manuscript.