Predicting the response to migraine therapy is an urgent possibility for daily clinical practice [1]. Nevertheless, assessment of therapeutic response may be controversial at present time. This is because there are no generally accepted criteria for refractory migraine.

Currently, there are 4 formulations of refractory migraine in the literature:

1. Criteria supported by Goadsby P.J. et al. (2006) [2]. They considered no response to 4 drugs of the following classes: beta-blockers, anticonvulsants, calcium channel blockers, tricyclic antidepressants, other drugs turned out to be effective in at least one controlled study, NSAIDs. The criteria did not indicate whether an episodic or chronic migraine was present, whether there was a drug-induced headache. Refractoriness is applied only to preventive therapy.

2. Criteria supported by Schulman E.A. et al. (2008) [3]. The authors considered refractoriness to drugs for arresting an attack: triptans and dihydroergotamine (injectable form or nasal spray), NSAIDs or combined analgesics. Refractoriness to preventive therapy — 2 out of 4 classes of drugs: beta-blockers, anticonvulsants, calcium channel blockers, tricyclic antidepressants. Episodic and chronic migraine is included, drug-induced headache is possible.

3. Criteria supported by Silberstein S.D. et al. (2010) [4]. The authors reported gradation of refractoriness (mild, moderate, severe). Response to the following drugs and classes is assessed: relief of attacks — NSAIDs, triptans, dopamine antagonists or parenteral NSAIDs, opioids, corticosteroids, parenteral dopamine antagonists. Preventive drugs: beta-blockers, anticonvulsants, calcium channel blockers (verapamil or flunarizine), tricyclic antidepressants, valproic acid, topiramate, combined therapy, gabapentin, other drugs turned out to be effective in at least one controlled study, NSAIDs, metabolic drugs (vitamin B2 , coenzyme Q10). Episodic and chronic migraine is included. No data on drug-induced headache.

4. Criteria supported by Martelletti R. et al. (2014) [5]. Refractory migraine is determined by no response to 3 of the following classes of drugs: beta-blockers, anticonvulsants, tricyclic antidepressants, others (flunarizine, candesartan), onabotulotoxin A. The criteria are applied only for preventive treatment of chronic migraine.

Correlation of certain genetic polymorphisms with the response to migraine therapy was studied in various researches. Christensen A.F. et al. [6] found an association of rs2651899 polymorphism in PRDM16 gene with positive response to triptans. An interesting study was carried out at the Danish Headache Center (2219 patients) [7]. The authors assessed polygenic risk index and response to specific anti-migraine therapy. Genetic risk index was associated with a 2-fold increase in the risk of migraine. Moreover, a significant direct relationship between the polygenic risk index and effectiveness of triptans was obtained. At the same time, there was no relationship between the polygenic risk index and non-specific therapy (NSAIDs) and prevention (beta-blockers, anticonvulsants, calcium channel blockers, angiotensin II receptor antagonists, ACE inhibitors, antidepressants).

Cargnin S. et al [8] analyzed genetic risk index of response to triptans among patients with migraine and no aura. Polymorphisms associated with aura-free migraine (rs9349379, rs2078371, rs6478241, rs11172113, rs1024905, and rs6724624) were selected for analysis. Combination of TRPM8 rs6724624 and FGF6 rs1024905 polymorphisms was associated with the risk of inconsistent response to triptans (OR 0.62; 95% CI 0.43-0.89). This research showed clinical significance of predictors of response to standard therapy.

Thus, the results of previous genetic studies made it possible to clarify the mechanisms of migraine and genetic risk factors. Nevertheless, these parameters did not affect the disease as significant biomarkers.

Identification of biomarkers of non-response to standard therapy may be used to select patients for therapy with monoclonal antibodies to CGRP or receptors, since earlier studies have shown the effectiveness of these drugs in patients with non-response to standard treatment [9].

Inclusion criteria, study design, patients and methods

We have analyzed the patients who appealed to the "University’s Headache Clinic" LLC in 2012-2016. Cephalgologist diagnosed migraine, assessed the patient's condition and response to therapy at initial visit (visit 1), after 6 months (visit 2) and after 12 months (visit 3).

Inclusion criteria:

1. Informed consent of patients.

2. Age 18-75 years.

3. Diagnosis of migraine (migraine without/with aura, chronic migraine) in accordance with International Classification of Headache III (2018); ICD-10 G43.0-2.

4. Duration of migraine over 6 months before inclusion in the study.

5. Treatment with at least 1 class of drugs for prevention of migraine (beta-blockers (metoprolol and propranolol), angiotensin II receptor blockers (candesartan), anticonvulsants (topiramate, valproic acid), antidepressants (amitriptyline and venlafaxine), onabotulotoxin A).

6. Relief of migraine with at least 1 specialized class of the drugs (triptans, ergotamine drugs, simple or combined analgesics).

Non-inclusion/exclusion criteria:

1. Diagnosis of probable migraine.

2. Other forms of primary and secondary headache.

3. Patient's inability to characterize the headache, cognitive impairment or other circumstances limiting the patient's ability to complete the Headache Diary and follow the physician’s recommendations.

Evaluation of therapeutic response

Patients were divided into groups of responders and non-responders depending on response to preventive therapy. We applied the method by Kogelman L.J.A. [7] as the most practical for assessing the response to migraine therapy. Non-responders (resistant) did not respond to both seizure therapy and preventive medication. Response to seizure therapy was considered to be at least 50% reduction in pain intensity within 2 hours after drug intake. We analyzed at least two drugs from different groups: triptans, ergotamine drugs, simple or combined analgesics.

Preventive therapy was effective if the number of days with headache was decreased by more than 50%. At least two drugs from different groups were evaluated: beta-blockers (metoprolol and propranolol), angiotensin II receptor blockers (candesartan), anticonvulsants (topiramate, valproic acid), antidepressants (amitriptyline and venlafaxine), onabotulotoxin A, calcium channel blockers (flunarizine, cinnarizine), and hormonal therapy.

Genetic survey

Venous blood samples were used for genetic testing (2 ml). Blood sampling was carried out at the treatment room of the "University’s Headache Clinic" LLC. We used 3% EDTA 1:10 as an anticoagulant. All samples were frozen and stored at –20°C. Transportation to the laboratory was carried out in thermoboxes. DNA was extracted in accordance with the manufacturer's instructions (“Isogen Laboratory” LLC, MagnaTM DNAPrep 200 set). In case of PCR, PCR-RFLP and AS-PCR, primers were selected manually without specialized programs or taken from publications. In case of RT-PCR, probes were selected by “DNA-Synthesis” LLC (Moscow). All primers and fluorescent-labeled probes were synthesized at the “DNA-Synthesis” LLC (Moscow). PCR, PCR-RFLP and AS-PCR were performed using a commercial HSTaq DNA polymerase reagent set (“Eurogen” CJSC, Moscow). PCR was carried out in a T100 amplifier (Bio-Rad, USA) or CFX96 RT-PCR amplifier (Bio-Rad, USA) according to the following scheme: preliminary denaturation (3 min, 92°C); 35—40 cycles: denaturation (30 sec, 94°C), primer annealing (30 sec), elongation (30 sec, 72°C). In case of PCR-RFLP, restriction of PCR products was performed using endonucleases produced by “SibEnzyme” company (Novosibirsk). PCR, AS-PCR and PCR-RFLP products were separated in a 2.5% agarose gel.

Statistical analysis

SPSS 10.0 software package was used for statistical analysis. We applied descriptive statistics (analysis of incidences and mean values). All variables were tested for normal distribution using Kolmogorov-Smirnov test. Parametric methods were used for data with normal distribution. Chi-square test was used to compare the percentage of alleles in all groups. Searching for gene-candidates was carried out using the PathwayStudio10 program (Elsivier). The combined migraine-associated genotypes were identified using APSampler v3.6 polygenic data analysis program. This program is designed to search for compound genetic biomarkers using the Markov Chain Monte Carlo (MCMC) method. This method allows analyzing the samples that did not pass the Hardy-Weinberg equilibrium test, as it has more power due to the use of Bayesian statistics. We used a permutation test (by Westfall-Young). Certain combined genotype was considered a significant biomarker of disease in case of odds ratio (OR) over 20.

Results

The study included 125 patients (81 ones with episodic migraine (76 — migraine without aura, 5 — migraine with aura), 44 patients with chronic migraine). Mean age of patients was 41.7 ± 12.5 years. There were 18 men and 128 women. All patients were Caucasian and lived in the Moscow region.

There were 115 (92%) responders to therapy and 20 (8%) non-responders. Both groups were comparable by gender and age.

Polymorphisms MTDH_rs1835740: CC, MTR_rs1805087: GG, CCKAR_rs1800857: TT, MIR22_rs6502892: CC were significantly more common in non-responders to migraine therapy (Table 1).

Table 1. Differences in polymorphisms in responders and non-responders to migraine therapy

We studied the role of complex genotypes on sensitivity to migraine therapy since the combination of genes with polymorphic changes has a greater impact on the risk of disease or symptom.

We found 2 complex genotypes significantly associated with unresponsiveness to standard migraine therapy: CCKBR_rs1805000: T; eNOS3_rs2070744: T; MAOA_VNT: 350 and MTDH_rs1835740: C; CCKBR_rs1805000: T; MAOA_VNT: 350 (Table 2).

Table 2. Significant complex genotypes and odds ratio of refractoriness to therapy (p<0.01 for all identified complex genotypes after 100 permutations)

Discussion

An association of some genes with unresponsiveness to migraine therapy was revealed. Correlation of migraine with MTDH gene was obtained in the first genomic study of migraine. This research was conducted in 2010 and included about 6000 patients. It was found that the T allele rs1835740 (NC_000008.11: g.97154685T> C) on chromosome 8q22.1 increases the risk of migraine, especially migraine with aura [10]. Single nucleotide polymorphism rs1835740 is located between the MTDH gene (encodes metadherin) and the PGCP gene regulating glutamate metabolism.

MTR gene encodes methionine synthase (folate cycle enzyme). Methionine synthase ensures reverse conversion of homocysteine to methionine. Polymorphism rs1805087 (Asp919Gly) reduces enzyme activity and results homocysteine accumulation.

MIR22 gene encodes micro-RNA 22 structure. This short non-coding RNA is a significant negative post-transcriptional modulator of gene expression. Micro-RNA dysregulation results local inflammation and sensitization [11].

MAOA gene encodes the structure of monoamine oxidase A. This enzyme breaks down triptans. Polymorphism MAOA_VNT: 350 can lead to accelerated metabolism of triptans and their insufficient effect.

eNOS3 gene encodes endothelial NO-oxidase 3. Polymorphism can lead to accumulation of nitric oxide that stimulates CGRP release.

CCKBR gene encodes cholecystokinin receptors type B. Cholecystokinin activates CGRP-containing neurons via CCKBR receptors at the level of parabrachial nucleus. Normally, this mechanism ensures a feeling of satiety. In case of migraine, it causes such symptoms as nausea and anorexia [12]. Enhanced CGRP release may be one of the mechanisms of drug-resistant migraine. This does not contradict the fact that anti-CGRP therapy is highly effective for drug-resistant migraine.

Thus, we have identified certain gene polymorphism associated with refractory migraine and biomarkers relating to unresponsiveness to standard migraine therapy.

Acknowledgment. The authors are grateful to the patients who took part in the study, as well as to Kondratyeva N.S., Zaitseva A.I., Anuchina A.A. for assistance in molecular genetic survey.

The authors declare no conflicts of interest.