Gallardo et al. The Journal of Headache and Pain (2023) 24:11 The Journal of Headache https://doi.org/10.1186/s10194-023-01542-z and Pain RESEARCH Open Access A study of differential microRNA expression profile in migraine: the microMIG exploratory study V. J. Gallardo1, J. B. Gómez‑Galván1, L. Asskour1, M. Torres‑Ferrús1,2, A. Alpuente1,2, E. Caronna1,2 and P. Pozo‑Rosich1,2* Abstract Background Several studies have described potential microRNA (miRNA) biomarkers associated with migraine, but studies are scarcely reproducible primarily due to the heterogeneous variability of participants. Increasing evidence shows that disease-related intrinsic factors together with lifestyle (environmental factors), influence epigenetic mech‑ anisms and in turn, diseases. Hence, the main objective of this exploratory study was to find differentially expressed miRNAs (DE miRNA) in peripheral blood mononuclear cells (PBMC) of patients with migraine compared to healthy controls in a well-controlled homogeneous cohort of non-menopausal women. Methods Patients diagnosed with migraine according to the International Classification of Headache Disorders (ICHD-3) and healthy controls without familial history of headache disorders were recruited. All participants com‑ pleted a very thorough questionnaire and structured-interview in order to control for environmental factors. RNA was extracted from PBMC and a microarray system (GeneChip miRNA 4.1 Array chip, Affymetrix) was used to determine the miRNA profiles between study groups. Principal components analysis and hierarchical clustering analysis were performed to study samples distribution and random forest (RF) algorithms were computed for the classification task. To evaluate the stability of the results and the prediction error rate, a bootstrap (.632 + rule) was run through all the procedure. Finally, a functional enrichment analysis of selected targets was computed through protein–protein interaction networks. Results After RF classification, three DE miRNA distinguished study groups in a very homogeneous female cohort, controlled by factors such as demographics (age and BMI), life-habits (physical activity, caffeine and alcohol con‑ sumptions), comorbidities and clinical features associated to the disease: miR-342-3p, miR-532-3p and miR-758-5p. Sixty-eight target genes were predicted which were linked mainly to enriched ion channels and signaling pathways, neurotransmitter and hormone homeostasis, infectious diseases and circadian entrainment. Conclusions A 3-miRNA (miR-342-3p, miR-532-3p and miR-758-5p) novel signature has been found differentially expressed between controls and patients with migraine. Enrichment analysis showed that these pathways are closely associated with known migraine pathophysiology, which could lead to the first reliable epigenetic biomarker set. Further studies should be performed to validate these findings in a larger and more heterogeneous sample. Keywords Migraine, Epigenetics, Epigenetic mechanisms, microRNA, miRNA, Biomarkers, Personalized medicine 1 *Correspondence: Headache and Neurological Pain Research Group, Vall d’Hebron Institute P. Pozo‑Rosich of Research (VHIR), Universitat Autònoma de Barcelona, Barcelona, Spain 2 patricia.pozo@vallhebron.cat Neurology Department, Headache Unit, Vall d’Hebron University Hospital, Barcelona, Spain © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. The Creative Commons Public Domain Dedication waiver (http://​creat​iveco​ mmons.​org/​publi​cdoma​in/​zero/1.​0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Gallardo et al. The Journal of Headache and Pain (2023) 24:11 Page 2 of 13 Background sensitization of the peripheral and/or central pain path- Migraine is a genetically-driven, chronic brain disease way [16, 17]. that ranks second on the most disabling neurological Several pilot studies have described miRNAs as poten- disorders of the Global Burden of Disease Study [1, 2]. tial peripheral biomarkers of migraine disease [18–26]. If not adequately diagnosed and treated, migraine can Nevertheless, studies are scarcely comparable and repro- increase in frequency and transform, through a chroni- ducible not only because of the lack of standardized fication process, into chronic migraine (CM), which is protocols, biological matrix or sample size but mainly defined by having 15 or more headache days per month because of the heterogeneous population and clini- [3], where patients’ burden increases exponentially [4]. cal variability included. Increasing evidence shows that Multiple factors, including both genetic and environ- disease-related intrinsic factors altogether with lifestyle mental mechanisms seem to play a role in migraine and factors may influence epigenetic mechanisms and iden- its chronification [5–7] process, although the suscep- tifying them is crucial to find unbiased reliable factors tibility of headache attack recurrence mainly depends associated with a specific disease [27–29]. on the brain-sensorial adaptability of both internal and Hence, the goal of this exploratory study is to find dif- external stimuli [8, 9]. Moreover, migraine stills rep- ferentially expressed miRNAs (DE miRNA) in PBMC of resent a clinical challenge because, up to date, disease patients with migraine compared to healthy controls in diagnosis, evolution and treatment response cannot be a well-controlled homogeneous non-menopausal female predicted due to the lack of reliable biomarkers. cohort. From large-scale genome-wide association studies (GWAS) performed in the last decades, we have learnt Methods that migraine has a polygenic heritability pattern and, Participants and protocol approval up to now, we have identified more than 120 risk loci The study was approved by our Institutional Review associated to migraine [10], mainly codifying for neu- Board (PR(IR)224/2016). All participants provided writ- ronal and vascular functions. However, the effect size ten informed consent before participation. We recruited of each individual reported gene is relatively modest patients treated for migraine in the Headache Outpa- and most of these genes have regulatory effects on gene tient Clinic. Patients had a diagnosis of migraine with or expression rather than protein coding. Environmen- without aura, confirmed by a neurologist according to tal factors may also modulate the expression of DNA. the International Classification of Headache Disorders, Therefore, the study of epigenetics which are the envi- 3rd edition (ICHD-3) [3]. Healthy controls (HC) were ronmental influenced genetic regulatory mechanisms recruited among hospital staff and non-related acquaint- [11] could provide further understanding on migraine ances of patients who denied past or first-degree familial pathophysiology as well as, give insight to the dynamic history of any recurrent primary or secondary headache aspect of the disease specifically contributing to its disorder. To diminish confounding factors, the inclusion chronification [12]. criteria were: female sex, 18–45 years old (non-meno- Among the epigenetic regulatory mechanisms, micro- pausal) and Mediterranean ethnicity for all participants. RNAs (miRNA) are a subtype of noncoding RNA of ∼22 Moreover, patients could not have received prophylac- nucleotides long that act as key regulators of genetic tic treatment for migraine prevention or if they received expression by inhibiting transcription or promoting it, it had been done for a brief period of time more than degradation of selected messenger RNAs (mRNAs). A one year or more before the study and without following single miRNA sequence could regulate multiple mRNA, national guidelines of effective dosage and duration. Due exerting a pleiotropic modulation of cellular processes. to the strict inclusion criteria, participants were included miRNA have been found intracellularly and in all body from August 2016 to September 2017. fluids, including plasma, saliva and urine [13]. Access- As miRNAs can be modulated by different lifestyle ing the central nervous system (CNS) is too invasive habits and disease characteristics, recruited participants and there is an increasing evidence that a relationship completed a very thorough and complete questionnaire between peripheral blood mononuclear cells (PBMC) and and structured-interview to collect demographic infor- brain-based epigenetic activity exists [14]. Thus, PBMC mation including: age, sex, ethnicity, educational level, gene expression may reflect both central and peripheral marital status, gross income; clinical information includ- neurological mechanisms in migraine, induced, for exam- ing: habits of caffeine, tobacco and drugs consumption, ple in migraine by the activation and neuroinflammation physical activity level using the International Physical of the trigeminocervical pain pathway [15]. In fact, previ- Activity Questionnaire (IPAQ-SF) [30]; life-time comor- ous studies in headache disorders have shown that tran- bidities, medications use, menstrual status and migraine scriptional changes of PBMC can reflect alterations in the information including disease duration, acute medication Gallardo et al. The Journal of Headache and Pain (2023) 24:11 Page 3 of 13 use, monthly headache frequency and family history of prior to the study because the sample size was based on headache. We registered headache pain, accompanying the available data for this exploratory analysis. However, symptoms and menstrual cycle the day of the blood sam- to increase the likelihood of finding differential miRNA ple extraction. We only analyzed samples from migraine expression profiles for this study, we selected participants patients with at least 48 h of a headache-free period prior from a well-controlled homogeneous non-menopausal day of the blood extraction. In addition, all patients had female cohort. P-values < 0.05 were considered as statisti- to fill in the following scales: Allodynia Symptom Check- cally significant and are reported for a two-tailed test. list (ASC-12) [31], Headache Impact Test-6 (HIT-6) [32], Migraine Disability Assessment Test (MIDAS) [33], Beck Clinical analysis Depression Inventory ­ 2nd Edition (BDI-II) [34], State- Nominal variables (aura and presence of allodynia) were Trait Anxiety Inventory (STAI) [35] and Perceived stress reported as frequencies (percentages) while median and scale (PSS) [36]. Patients were asked to complete head- interquartile range (IQR) were reported for quantitative ache diary 30 days before extraction to confirm migraine variables: age, BMI, IPAQ-SF, caffeine and alcohol con- diagnosis (screening period). We also excluded patients sumptions, STAI, BDI-II, PSS, disease evolution time, who fulfilled acute medication overuse headache criteria ASC-12, monthly headache frequency (MHD), monthly during the screening period. acute medication intake (MAMI), MIDAS and HIT- 6. Normality assumption of quantitative variables was Blood extraction, PBMC isolation and RNA extraction checked through visual methods (Q-Q plots) and nor- For blood extraction patients were asked to fast for 6 h, mality tests (Shapiro–Wilk test). Statistical significance minimize their exercise/activity and not take analge- between study groups (HC vs. patients with migraine) sic medication 24 h prior to blood extraction. Periph- was assessed by unpaired t-test for BMI, caffeine con- using three, 8 mL, BD Vacutainer® CPT™ Cell Prepara- eral blood was collected by antebrachial vein puncture sumption, STAI and PSS was used and unpaired Wil- coxon rank-sum test was used for the other quantitative tion Tubes with Sodium Citrate (BD, Franklin Lakes, variables that did not follow any normality assumption. NJ, USA) and processed within 2 h following the manu- facturer’s instruction to isolate PBMC pellets that were Bioinformatic analysis immediately stored at -80ºC till miRNA extraction. To The statistical analysis of the Genechip miRNA array minimize variability due to manipulation, the miRNA plate 4.1 was performed using libraries developed for extraction was made by the same laboratory technician the microarray analysis in the Bioconductor Project [39]. in the same facilities. For miRNA extraction, we used the At the moment of the analysis, there were no specific Qiagen miRNeasy Mini Kit (Qiagen, Valencia, CA, USA), packages for miRNA 4.1 arrays available, thus we pro- following the manufacturer’s instruction. RNA concen- grammed our own package. Quality control (QA) was tration was measured using Nanodrop 2000 (NanoDrop assessed with the package arrayQualityMetrics before Technologies, Wilmington, DE). RNA integrity number the normalization process [40]. Additionally, principal (RIN) was determined the kit Agilent RNA 6000 Nano components analysis (PCA) and hierarchical clustering by the Bioanalyzer Agilent 2100 (Agilent Technologies, analysis (HCA) were performed to study samples distri- Santa Clara, CA, USA). We excluded from analysis sam- bution and detect technical problems, including batch ples with RIN under 6. effect. Each array (CEL file) was preprocessed and back- ground corrected, normalized and summarized using Microarray hybridization RMA (Robust Multiarray Average) through Bioconduc- Microarrays service was outsourced and carried out in at tor packages [41, 42]. High Technology Unit at Vall d’Hebron Research Insti- The identification of differentially expressed miRNA tute (HTU-VHIR), Barcelona (Spain). For this study, we (DE miRNAs) was based on adjusting linear model with used the Affymetrix GeneTitan microarray platform and empirical Bayes moderated t-statistic, conducted by the the Genechip miRNA array plate 4.1. Starting material limma package [43] in the comparison between controls was 550 ng of total RNA of each sample. MiRNA in the and patients with migraine. Models were adjusted by sample was labelled using Flash Tag Biotin HSR RNA covariates using surrogate variables estimation [44]. DE Labelling kit following the manufacturer’s instructions. miRNAs that had more than 0.5-fold increase/decrease relative to controls were chosen for further analysis. The Statistical analysis minimum of 0.5-fold expression difference cut-off has All statistical analysis were conducted in R v4.2.0 [37] and been chosen based on the several miRNA profiling stud- figures were produced using the package ggplot2 v.3.3.6 ies using microarrays; confirming that subtle changes in [38]. A statistical power calculation was not conducted miRNA expression, such as a 0.5-fold difference, have a Gallardo et al. The Journal of Headache and Pain (2023) 24:11 Page 4 of 13 significant biological impact [45–48]. Results were fur- to identify distinct active subnetworks and then perform ther corrected for multi-testing using the Benjamini– enrichment analyses on these subnetworks. For assign- Hochberg procedure for control the false discovery rate ing a significance value for each predicted target gene (FDR) [49]. Features that met both criteria (adjusted (pathfindR input), all Homo sapiens miRNA-target gene t-test and ± 0.5 fold change) were combined in a sin- predictions were obtained: 1,954,039 miRNA-gene pre- gle matrix of DE miRNAs that underwent classification dictions from all H. sapiens miRNAs listed in the miRDB analysis. (2,656 miRNAs). Then, the significance for each pre- We used Random Forests (RF) models for the classifi- dicted gene was computed as the probability of observing cation task, a machine learning method widely used in a score greater than or equal to the score of this target the microarray analysis [50–52]. RF algorithm uses an gene over all H. sapiens miRNA-target gene scores [59]. ensemble of classification trees, internally and randomly For genes that were targeted by more than one miRNA, constructed using a bootstrap sample of the data. Then, the lowest significance was kept. the algorithm merges all constructed tree altogether (for- est) in order to get an accurate and stable prediction [53, Results 54]. Once we obtained the forest, the construction of the Clinical characteristics of the patients classifier is performed by a feature reduction algorithm, Participant recruitment is represented in Fig. 1. Since where the less important features are successively elimi- August 2016, a total of 76 (49 patients with migraine and nated and out of the bag (OOB) error is continuously 27 HC) participants were recruited. For this exploratory analyzed. To evaluate the stability of the results and the study, RNA was extracted from 69 samples that fulfilled prediction error rate, a bootstrap (0.632 + rule) was run inclusion/exclusion criteria. Among these, 37 showed a through all the procedure. All classification task and RIN < 6 and were discarded. Therefore, 32 participants bootstrap were performed through the varSelRF R pack- (12 HC, 20 patients) were considered for the analysis age [55, 56]. with a median age of 33.2 [28.0, 39.8] years old. No sta- Target genes of selected DE miRNAs were predicted by tistically significant differences were found neither in using the miRDB database through the multimiR pack- demographical variables nor life-habits measurements age [57] and functional and pathway enrichment analysis between groups. In terms of comorbidities, we only was performed using the pathfindR package [58]. Path- found that patients with migraine presented a statisti- findR identifies gene sets that form active subnetworks cally significantly greater trait-anxiety score (HC: 31.0 in a protein–protein interaction network (PIN) in order [27.0, 38.2] vs. migraine: 39.5 [37.2, 55.8]; p = 0.007). In Fig. 1 Patients’ recruitment flowchart. We recruited participants from August 2016 to September 2017. HC: healthy controls; RIN: RNA integrity number; QA: quality assessment Gallardo et al. The Journal of Headache and Pain (2023) 24:11 Page 5 of 13 patients, median MHD was 10.5 [8.0, 24.2] days/month samples were perfectly segregated along PCA and HCA (d/mo) and median MAMI was 7.5 [4.0, 9.0] d/mo. Main (Fig. 2). demographic and clinical characteristics are shown in Adjusting by STAI-trait values, we found 191 DE Table 1. miRNA between study groups: 59 over-expressed and No statistically significant differences were found in 132 under-expressed transcripts in patients (Fig. 3A). terms of RNA quality (RIN, median [IQR]) between Then, RF analysis combined with a feature selecting study groups: 6.9 [6.5, 7.8] HC vs. 6.7 [6.3, 7.2] patients; algorithm was used to build the classifier and to iden- p = 0.420. tify the best targets that distinguished between controls and migraine. The OOB for the initial model was 0.067 (91.7% sensitivity, 94.4% specificity) and a 3-miRNA sig- MiRNA microarray expression profiling nature was found as best variables without dropping the During microarrays QA, we distinguished two sam- OBB error substantially: miR-342-3p, miR-532-3p and ples (from patients) with higher values of intensity in all miR-758-5p. Then, as RF algorithms are deterministic, their probe sets. Similarly, normalized unscaled stand- performing it on different training samples is the only ard errors values and relative log expression values for way to generate diversity in the selection. To this end, the same two samples diverged from the rest, indicating we performed bootstrapping (+ 0.632) in which the ran- aberrant expression. Hence, we discarded these 2 sam- dom forest constructed for a certain number of variables ples from further analysis (Fig. 1). Control and migraine was subsampled and compared. The prediction error Table 1 Demographics, life habits, comorbidities and migraine characteristics HC Migraine P value (n = 12) (n = 18) Demographics Age, years 30.0 [28.0, 33.5] 36.0 [28.2, 41.0] 0.319‡ 2 Body mass index, kg/m 22.7 [21.7, 23.2] 22.9 [20.4, 25.5] 0.594† Life-habits IPAQ-SF, METs/week 3.1 [2.2, 4.2] 3.4 [1.4, 6.2] 0.750‡ Caffeine consumption, g/week 1.1 [0.9, 1.6] 1.0 [0.2, 1.4] 0.261† Alcohol consumption, g/week 15.0 [0.0, 40.0] 0.0 [0.0, 10.0] 0.062‡ Comorbidities Anxiety (STAI), score State 35.0 [32.0, 41.0] 38.5 [36.2, 52.2] 0.075† Trait 31.0 [27.0, 38.2] 39.5 [37.2, 55.8] 0.007† Depression (BDI-II), score 4.0 [0.8, 6.2] 8.5 [1.2, 14.8] 0.232‡ Perceived stress (PSS), score 16.0 [12.8, 22.5] 19.5 [17.0, 32.0] 0.270† Migraine characteristics Duration of migraine disease, y 19.5 [11.0, 27.0] Aura 38.9% (7) Allodynia 72.2% (13) ASC-12, score 4.5 [2.2, 6.0] Monthly headache frequency, d/mo 10.5 [8.0, 24.2] Monthly acute medication intake, d/mo 7.5 [4.0, 9.0] Migraine-related clinical burden Disability (MIDAS), score 28.5 [17.2, 47.8] Headache-related impact (HIT-6), score 63.5 [61.2, 69.0] Continuous data is represented in median [IQR] and categorical data in % (n). Bold font indicates statistically significant variables IQR: interquartile range, HC healthy controls, y years, IPAQ-SF short-form of the international physical activity questionnaire, MET metabolic equivalents of task, STAI State-Trait Anxiety Inventory, BDI-II Beck depression inventory-second edition, PSS perceived stress scale, ASC-12 12-item allodynia symptom checklist; d/mo days/ month, MIDAS migraine disability assessment, HIT-6 headache impact test † Statistical significance assessed with unpaired t-test ‡ Statistical significance assessed with unpaired Wilcoxon rank-sum test Gallardo et al. The Journal of Headache and Pain (2023) 24:11 Page 6 of 13 Fig. 2 Principal component analysis (A) and unsupervised hierarchical clustering (B) of miRNA normalized expression data. Unsupervised hierarchical clustering (B) was computed through euclidian distances between samples. HC: healthy controls; MIG: patients with migraine, PC: principal component Fig. 3 Volcano plot (A) and heatmap (B) of differentially expressed miRNA between healthy controls and patients with migraine. Scattered points from the volcano plot (A) represent miRNAs: the x-axis is the log twofold-change for the ratio healthy controls vs patients with migraine, whereas the y-axis is the log 10 P-value. Colored dots (blue and green) are differentially miRNA statistically significantly over- (right) and under-expressed (left) in patients compared to controls. Blue dots are differentially expressed miRNA with a log twofold-change > 0.50. A heatmap (B) of the 3-miRNA signature selected from the classification algorithm (RF). Columns represent individual arrays, while rows represent specific DE miRNA. The Z-score depicts a measure of distance, in standard deviations, away from the mean. The relative value for each miRNA is depicted by color intensity, with yellow indicating over-expression and red indicating under-expression. HC: healthy controls; MIG: patients with migraine rate among the bootstrap samples was 0.142. The same selected DE miRNA between groups where the expres- 3-miRNA signature were consistently selected (stability) sion from all of them were under-expressed in patients: in all the sub-samples, with miR-532-3p being selected in miR-342-3p ­(log2FC: -0.551), miR-532-3p ­ (log2FC: the 60% of cases. Figure 3B shows the heatmap from the -1.109) and miR-758-5p (­ log2FC: -0.508). Gallardo et al. The Journal of Headache and Pain (2023) 24:11 Page 7 of 13 Targets prediction and functional enrichment analysis Functional enrichment analysis of miRNA-target genes Predicted target genes associated with the 3-miRNA sig- predicted from the miRDB was performed through nature were retrieved from the miRBD database through active-subnetwork-oriented analysis via pathfindR and the multimiR package. Each target significance was com- a total of 24 enriched KEGG pathways were identified puted according to the prediction score retrieved from (Table 2). Top pathway was calcium signaling pathway the miRDB (see Methods section). Then, a total of 68 and a total of 9 different target genes were involve in all target genes were identified: 42 genes from miR-342-3p, significantly enriched pathways: CACNA1C, HTR2C, 26 genes from miR-532-3p and 9 genes from miR-758-5p. EP300, PTGER4, ID4, RICTOR and EEA1 from miR- While four common target genes were found between 342-3p; PRKCA and DTX4 from miR-532-3p and ID2 miR-342-3p and miR-532-3p: ATXN7 (ataxin 7), CDKL5 from miR-758-5p. (cyclin dependent kinase like 5), PAPPA (Pappalysin 1) and RRM2 (Ribonucleotide Reductase Regulatory Subu- Discussion nit M2); two common targets were found between miR- In this exploratory study, we found 191 differentially 342-3p, miR-758-5p: DHRSX (Dehydrogenase/Reductase expressed miRNAs (59 over-expressed and 132 under- X-Linked), NUFIP2 (Nuclear FMR1 Interacting Pro- expressed transcripts) between controls and patients tein 2); and 3 common targets between miR-532-3p and with migraine in a very homogeneous female cohort, miR-758-5p: DSG3 (Desmoglein 3), PTP4A1 (Protein controlled by factors as demographics, life-habits and Tyrosine Phosphatase 4A1), TET3 (Tet Methylcytosine comorbidities (Table 1, Fig. 3A). After transcripts classi- Dioxygenase 3). No common target gene was found for fication task, we finally selected three DE miRNAs that all three miRNAs. distinguished study groups: miR-342-3p, miR-532-3p and miR-758-5p. All three DE miRNAs were under-expressed Table 2 Enriched KEGG pathways of predicted target from the 3-miRNA differentially expressed signature ID KEGG Pathway Adj. p-value Target genes hsa04020 Calcium signaling pathway < 0.0001 CACNA1C, HTR2C, PRKCA hsa04330 Notch signaling pathway 0.001 DTX4, EP300 hsa04750 Inflammatory mediator regulation of TRP channels 0.002 HTR2C, PTGER4, PRKCA hsa04720 Long-term potentiation 0.002 CACNA1C, EP300, PRKCA hsa05031 Amphetamine addiction 0.003 CACNA1C, PRKCA hsa04540 Gap junction 0.005 HTR2C, PRKCA hsa04350 TGF-beta signaling pathway 0.006 ID2, ID4, EP300 hsa04912 GnRH signaling pathway 0.006 PRKCA, CACNA1C hsa04713 Circadian entrainment 0.007 CACNA1C, PRKCA hsa04925 Aldosterone synthesis and secretion 0.007 PRKCA, CACNA1C hsa04916 Melanogenesis 0.008 EP300, PRKCA hsa04066 HIF-1 signaling pathway 0.010 EP300, PRKCA hsa04726 Serotonergic synapse 0.011 CACNA1C, HTR2C, PRKCA hsa04935 Growth hormone synthesis, secretion and action 0.013 EP300, PRKCA, CACNA1C hsa04919 Thyroid hormone signaling pathway 0.014 EP300, PRKCA hsa04270 Vascular smooth muscle contraction 0.015 CACNA1C, PRKCA hsa04728 Dopaminergic synapse 0.018 PRKCA, CACNA1C hsa04261 Adrenergic signaling in cardiomyocytes 0.025 CACNA1C, PRKCA hsa04921 Oxytocin signaling pathway 0.027 PRKCA, CACNA1C hsa04150 mTOR signaling pathway 0.028 RICTOR, PRKCA hsa05161 Hepatitis B 0.033 EP300, PRKCA hsa04310 Wnt signaling pathway 0.034 EP300, PRKCA hsa05164 Influenza A 0.039 EP300, PRKCA hsa05152 Tuberculosis 0.046 EP300, EEA1 Adj. p-value are the lowest enrichment p-value obtained through multiple iterations. The target genes are the predicted genes (targets of at least one DE miRNA) involved in the given enriched pathway KEGG Kyoto Encyclopedia of Genes and Genomes, Adj. p-value adjusted p-value corrected for multi-testing using the Benjamini–Hochberg procedure (FDR) Gallardo et al. The Journal of Headache and Pain (2023) 24:11 Page 8 of 13 in patients compared to controls (Fig. 3B). Sixty-eight in PBMC) and miR-96-5p (under-expressed) [26]. Results target genes were predicted from this 3-miRNA signature are heterogenous and study-dependent but not only and in the functional analysis, we found that they mainly because there are differences in lab procedures, biologi- enriched ion channels and signaling pathways (calcium, cal matrix or sample size but mainly because of the large notch, TRP, TGF-beta, GnRH, HIF-1, thyroid, oxy- degree of heterogeneity between study groups. Published tocin, mTOR and Wnt), neurotransmitter and hormone studies only controlled by demographic factors (age and homeostasis (aldosterone, serotonin, growth hormone, sex) but did not consider neither participants selection dopamine and adrenaline), infectious diseases (Hepatitis, nor analysis with an adjustment by covariates other fac- Influenza and tuberculosis) and circadian entrainment tors such as comorbidities and lifestyle habits. Epigenet- (Table 2). ics have been established as a molecular link between MiRNAs may have the potential to accurately diag- genetic function and the environment. Hence, epigenetic nose disorders at an early stage. For instance, it has been marks constitute a malleable language by which internal found that levels of miR-182 in the peripheral blood and external factors can influence the genome and, for of patients with gliomas increase with disease grade, that reason, it is crucial to be able to separate correlative which had allowed to researchers to patent a diagnostic from causal changes in order to understand and quan- kit for early diagnosis of human glioma (patent num- tify the effect of the epigenetic function in diseases [70]. ber: CN101792793A) [60]. This is beneficial for patients Nowadays, we have more evidence that lifestyle factors in order to receive earlier and more individual targeted such as diet [71], smoking [72], alcohol [73, 74], stress treatment therapies, a milestone in migraine since we [75], physical activity [29, 76], treatments [77–79] and already know that if it is not adequately diagnosed and comorbidities [80, 81] have a great impact on the dyna- treated, migraine frequency increases until disease chron- misms of epigenetic marks. Controlling them will allow ification [61]. Moreover, there is increasing evidence us to identify transcripts associated with a disease that that the later we start treating migraine, less effective is are not unmasked for other factors. the treatment and more difficult it becomes to reverse Functional analysis showed that targeted genes associ- the disease [5, 62]. In regards to CNS diseases, specific ated to the 3-miRNA signature enriched pathways have DE miRNAs have also been identified and validated in been previously described and associated with migraine order to diagnose multiple sclerosis (US9187785B2) [63], pathophysiology. One of the most enriched targets is Alzheimer’s disease (EP2733220A1) [64, 65] or major CACNA1C which encodes for the alpha-1 subunit of depressive disorder (EP2529222A4) [66]. MiRNAs have a voltage-dependent calcium channel. Calcium chan- also been given considerable attention as a therapeutic nels mediate the influx of calcium ions into the cell upon targets/agent. In recent years, miRNA-based therapeu- membrane polarization and they are encoded by different tic approaches have been developed in CNS diseases, subunits of the CACNA1 gene (CACNA1A-S) [82, 83]. especially pre-clinical studies in neurodegenerative dis- From migraine pathophysiology, we already know that orders [67], although clinical trials have not started yet the trigeminovascular system (TVS) can be activated by [68]. Hence, these patents and previous studies in other cortical spreading depression (CSD). Mutations in CAC- diseases highlight that investigating miRNA profiles in NA1A can disrupt ionic homeostasis, increasing the sus- migraine will lead us not only to a better understanding ceptibility to CSD and inducing migraine attacks [84]. of their pathogenesis, but also to help develop diagnostic Moreover, a CACNA1A variant has also been associated kits, identify specific cluster of patients, explore disease with migraine with aura in the last migraine association onset, monitor migraine attacks, personalize therapeutic study [10]. So, a down-regulation of CACNA1C gene treatment and revert disease chronification. could be also associated with these ionic disturbances in Several authors have made an effort in previous studies neuronal synapses; indicating that miR-342-3p has a reli- to identify potential miRNAs as peripheral biomarkers of able effect on migraine pathophysiology. migraine [12]. Numerous miRNAs have been described Other enriched pathways codifying for vascular func- associated with pain-free patients with migraine in tions which may have and important role in migraine comparison to healthy controls: miR-382-5p (over- etiopathogenesis were the notch signaling pathway, the expressed in serum and PBMC) [21, 23]; miR-34a-5p inflammatory mediator regulation of TRP channels and (over-expressed in serum, saliva and PBMC) [21, 24]; the HIF-1 signaling pathway. Notch receptors (Notch1- miR-27b (over-expressed in plasma) [25]; miR-181a, let- 4) are a family of type-1 transmembrane proteins which 7b, and miR-22 (under-expressed in PBMC) [25]; miR- are essential for embryonic development. Mutations 155, miR-126, and let-7 g (over-expressed in plasma) [18]; in the Notch pathway genes, like NOTCH2 and JAG1, miR-30 (over-expressed in PBMC) [69]; miR-5189-3p, are associated with vascular abnormalities [85]. In miR-3613-5p, miR-99a-3p, miR-542-3p (over-expressed migraine, there is evidence that distinct polymorphisms Gallardo et al. The Journal of Headache and Pain (2023) 24:11 Page 9 of 13 of NOTCH4 gene significantly modify clinical char- been used to relieve migraine attacks through intrana- acteristics of migraine phenotype [86]. The Transient sal or intravenous administration [105] since OT recep- Receptor Potential (TRP) channels are a superfamily of tors included in the trigeminal ganglia and neurons can integral membrane proteins well-described in pain con- inhibit the release of CGRP [106]. ditions that functions as cation channels: TRPA, TRPC, Altogether, the 3-miRNA signature found could be TRPM, TRPP, TRPL, and TRPV [87–89]. These chan- a reliable biomarker for migraine diagnosis not only nels contribute to a several physiological processes rang- because they are differentially expressed between female ing from thermosensation and pain to the regulation of patients and controls, but they are also involved in Ca2 + levels in the endoplasmic reticulum. TRP chan- enriched biological processes which have been previously nels are also expressed in the meningeal nociceptors described in migraine pathophysiology. So, the under- and it has been proved that their activation, promotes expression of these transcripts may cause a dysregulated the release of the neuropeptide calcitonin gene related expression of genes associated with the vascular dysfunc- peptide (CGRP) [90]. The role of CGRP in migraine has tion and neurohormonal imbalances characterized by been exhaustively described in both preclinical and clini- patients with migraine. cal studies during the last decades [91] and the CGRP This study has limitations. First, RNA-seq is the most antagonism has shown to be efficacious for the treat- commonly used NGS technique to explore differences ment of migraine [92–94]. However, the mechanism by at transcript level to date. RNA-seq is able to detect which CGRP is released during migraine is unclear, and novel transcripts with higher specificity and sensitiv- for that reason, TRP channels remain a focus of interest ity. However, at the moment of designing and planning for their potential contribution to attacks [95]. Moreover, the current study, arrays where the more standardized from our experience in clinical trials and real-world stud- technology. In addition, the RNA-Seq procedure has a ies there are patients with migraine who do not respond few disadvantages compared to arrays: a lack of opti- to treatments, even target-driven one such as anti-CGRP mized and standardized protocols for analysis [107, therapies [96], suggesting that CGRP is not the only 108]; the size of RNA-Seq files is considerably larger; mechanism driving migraine attacks [97, 98]. In regards and RNA-seq requires highly intensive and expensive to hypoxia-inducible factor 1 (HIF-1), it is a transcrip- computation infrastructure and analytics, as well as, tion factor that functions as a master regulator of oxygen longer analysis times [108, 109]. Nevertheless, arrays homeostasis. Since HIF-1α is implicated in neuropro- technologies are still a reliable technology to study tection and inflammation inhibition, a recent study has gene expression with results comparable to RNA-seq demonstrated that roxadustat administration, an anti- [110]. Second, we did not evaluate the hypothesis that anemia medication, ameliorated migraine-like behaviors the miRNA profile could be related to the phenotype of and inhibited central pain sensitization in nitroglycerin- circulating PBMC phenotype, rather than a differential injected mice, which was mainly mediated by HIF-1α/ miRNA expression. Several changes in PBMC profile NF-κB/inflammation pathway [99]. related to migraine have been reported but, as differ- The relationship between hormones and neurotrans- ent methodologies were used, the overall results are mitters with migraine is not entirely clear, but we do inconclusive. For instance, a meta-analysis of studies know it exists and plays a fundamental role in its patho- from 1966 to 1999 addressing immunological changes physiology. On one hand, thyroid hormones, along with related to migraine did not find clear alterations in cir- PTX3 and, insulin resistance share common mechanisms culating leukocytes or other immunological parameters in the development of endothelial dysfunction [100], a [111]. More recently, a relative lymphocytosis has been vascular pathophysiological genetically driven mecha- reported in chronic migraine associated with medi- nism of migraine [101]. Moreover, differentially pattern cation overuse headache, that was considered prob- levels have been found between controls and patients ably related to a chronic inflammatory state and stress with migraine of the hypothalamic-tuberoinfundibu- response rather than specifically to migraine [112, lar system (prolactin and growth hormone) [102]. On 113]. Finally, the sample was small, and highly selected the other hand, the gonadotropin releasing hormone and homogeneous, only considering non-menopausal (GnRH) controls the release of sexual hormones although women in order to reduce sample size and confounding estrogens can negatively regulate the GnRH from the factors. Moreover, the group of individuals diagnosed hypothalamus at the time of the preovulatory surge, sug- with migraine was not totally homogeneous in terms of gesting that estrogen could be related to the increase in disease pathophysiology (e.g., either episodic or chronic migraine prevalence in relation with puberty and ovu- migraine and either the presence or absence of aura). lation in women [103, 104]. Other hypothalamic hor- Hence, this 3-miRNA signature should be validated in mone linked to migraine is the Oxytocin (OT). OT has a larger sample size and more heterogeneous cohort. Gallardo et al. The Journal of Headache and Pain (2023) 24:11 Page 10 of 13 However, data gathered in this exploratory study will be draft of the paper. All authors have critically revised and finally approved the version of the paper to be published. used to calculate the required sample size for the sub- sequent discovery and validation (RT-qPCR) studies. Funding Due to lack of power in small sample size experiments, This study was solely funded by the Migraine Research Foundation. true effects are often missed and many of the detected effects will probably not be further validated. Therefore, Declarations the objective of this study was to to perform an explor- Ethics approval and consent to participate atory study using the same in- and exclusion criteria for The study was approved by our Institutional Review Board (PR(IR)224/2016) a miRNA high-throughput screen [114]. and all participants consented to participated. Competing interests VJ.G., JB.G.G. and A.L. have nothing to declare. M.T.F. has received honoraria as Conclusion a consultant or speaker for Allergan-Abbvie, Almirall, Chiesi, Eli Lilly, Novartis and Teva. A.A. has received honoraria from Allergan-Abbvie, Novartis, Chiesi. A 3-miRNA signature (miR-342-3p, miR-532-3p and E.C. reports no diclosures. P.P.R. has received honoraria as a consultant and miR-758-5p) has been found differentially expressed speaker for: AbbVie, Amgen, Biohaven, Chiesi, Eli Lilly, Lundbeck, Medscape, between controls and patients with migraine in a very Novartis, Pfizer and Teva. Her research group has received research grants from AbbVie, Novartis and Teva; and, has received funding for clinical trials from homogeneous cohort of non-menopausal women, con- Alder, AbbVie, Amgen, Biohaven, Electrocore, Eli Lilly, Lundbeck, Novartis and trolled by factors such as demographics, life-habits and Teva. She is the Honorary Secretary of the International Headache Society. comorbidities. Enrichment analysis show that these She is in the editorial board of Revista de Neurologia, associate editor for Cephalalgia, Headache, Neurologia, Frontiers of Neurology. She is a member three miRNAs are closely associated with migraine of the Clinical Trials Guidelines Committee and Scientific Committee of the pathophysiology, maybe becoming the first reliable bio- International Headache Society. She has edited the Guidelines for the Diag‑ marker of disease. Further studies will be performed nosis and Treatment of Headache of the Spanish Neurological Society. She is the founder of www.​midol​ordec​abeza.​org. Pozo-Rosich P does not own stocks to validate them in a larger and more heterogeneous from any pharmaceutical company. sample. Received: 30 November 2022 Accepted: 27 January 2023 Abbreviations ASC-12 Allodynia Symptom Checklist BDI-II Beck Depression Inventory 2nd Edition CM Chronic Migraine CNS Central Nervous System References DE Differentially Expressed 1. Stovner LJ, Nichols E, Steiner TJ, Abd-Allah F, Abdelalim A, Al-Raddadi FDR False Discovery Rate RM et al (2018) Global, regional, and national burden of migraine and GWAS Genome-Wide Association Studies tension-type headache, 1990–2016: a systematic analysis for the Global HC Healthy Controls Burden of Disease Study 2016. Lancet Neurol 17(11):954–976 HCA Hierarchical Clustering Analysis 2. Stovner LJ, Hagen K, Linde M, Steiner TJ (2022) The global prevalence of HIT-6 Headache Impact Test-6 headache: an update, with analysis of the influences of methodological ICHD-3 International Classification of Headache Disorders (3rd edition) factors on prevalence estimates. J Headache Pain 23(1):1–17 IQR Interquartile Range 3. Headache Classification Committee of the International Headache IPAQ-SF International Physical Activity Questionnaire (Short Form) Society (2018) The International Classification of Headache Disorders, MAMI Monthly Acute Medication Intake ­3rd edition. Cephalalgia 38(1):1–211 MHD Monthly headache frequency 4. Lanteri-Minet M (2014) Economic burden and costs of chronic MIDAS Migraine Disability Assessment Test migraine. Curr Pain Headache Rep 18(1):1–6 miRNA MicroRNA 5. May A, Schulte LH (2016) Chronic migraine: risk factors, mechanisms mRNA Messenger RNA and treatment. Nat Rev Neurol 12(8):455–464 OOB Out Of the Bag 6. Goadsby PJ, Charbit AR, Andreou AP, Akerman S, Holland PR (2009) PBMC Peripheral Blood Mononuclear Cells Neurobiology of migraine. Neuroscience 161(2):327–341 PCA Principal Components Analysis 7. Torres-Ferrús M, Ursitti F, Alpuente A, Brunello F, Chiappino D, de Vries PIN Protein-protein Interaction Network T et al (2020) From transformation to chronification of migraine: patho‑ PSS Perceived Stress Scale physiological and clinical aspects. J Headache Pain 21(1):1–12 QA Quality Assessment 8. Bernstein C, Burstein R (2012) Sensitization of the trigeminovascular RF Random Forest pathway: perspective and implications to migraine pathophysiology. J RIN RNA Integrity Number Clin Neurol 8(2):89–99 RMA Robust Multiarray Average 9. Schwedt TJ (2013) Multisensory integration in migraine. Curr Opin STAI State-Trait Anxiety Inventory Neurol 26(3):248 VHIR Vall d’Hebron Research Institute 10. Hautakangas H, Winsvold BS, Ruotsalainen SE, Bjornsdottir G, Harder AVE, Kogelman LJA, et al (2022) Genome-wide analysis of 102,084 Authors’ contributions migraine cases identifies 123 risk loci and subtype-specific risk alleles. VJ.G. and P.P.R. made substantial contributions to the conception and Nat Genet 54(2):152–60 design of the study. L.A. made substantial contribution to the lab methodol‑ 11. Tammen SA, Friso S, Choi SW (2013) Epigenetics: The link between ogy (blood extraction, PBMC isolation and RNA extraction) and HTU-VHIR nature and nurture. Mol Aspects Med 34(4):753–64. https://​doi.​org/​10.​ performed microarray hybridization. JB.G.G., M.T.F., A.A. and E.C. contributed 1016/j.​mam.​2012.​07.​018 to patient recruitment and data collection. VJ.G. performed all clinical and bioinformatic analysis. He also designed all figures, tables and wrote the first Gallardo et al. The Journal of Headache and Pain (2023) 24:11 Page 11 of 13 12. Gallardo VJ, Vila‐Pueyo M, Pozo‐Rosich P (2023) The impact of epige‑ 35. Oei TPS, Evans L, Crook GM (1990) Utility and validity of the STAI with netic mechanisms in migraine : Current knowledge and future direc‑ anxiety disorder patients. Br J Clin Psychol 29(4):429–432 tions. Cephalalgia43(2):3331024221145916 36. Sanz-Carrillo C, Garcıa-Campayo J, Rubio A, Santed MA, Montoro M 13. Ha M, Kim VN (2014) Regulation of microRNA biogenesis. Nat Rev Mol (2002) Validation of the Spanish version of the Perceived Stress Ques‑ Cell Biol 15(8):509–524 tionnaire. J Psychosom Res 52(3):167–172 14. Kos MZ, Puppala S, Cruz D, Neary JL, Kumar A, Dalan E et al (2022) 37. Team RC (2013) A language and environment for statistical computing. Blood-Based miRNA Biomarkers as Correlates of Brain-Based miRNA R Foundation for Statistical Computing, Vienna, Austria (http://​www.R-​ Expression. Front Mol Neurosci 15(March):1–13 proje​ct.​org/) 15. Buzzi MG, Moskowitz MA (2005) The pathophysiology of migraine: Year 38. Wickham H (2016) ggplot2: Elegant Graphics for Data Analysis. 2005. J Headache Pain 6(3):105–111 Springer-Verlag, New York. ISBN 978-3-319-24277-4. https://​ggplo​t2.​ 16. Gerring Z, Rodriguez-Acevedo AJ, Powell JE, Griffiths LR, Montgomery tidyv​erse.​org GW, Nyholt DR (2016) Blood gene expression studies in migraine: 39. Gentleman RC, et al. Bioconductor: open software development for potential and caveats. Cephalalgia 36(7):669–678 computational biology and bioinformatics. Genome Biol. 2010;11:202. 17. Aczél T, Kun J, Szőke É, Rauch T, Junttila S, Gyenesei A et al (2018) Tran‑ 2004;5(10):R80. scriptional alterations in the trigeminal ganglia, nucleus and peripheral 40. Kauffmann A, Gentleman R, Huber W (2009) arrayQualityMetrics - A blood mononuclear cells in a rat orofacial pain model. Front Mol bioconductor package for quality assessment of microarray data. Bioin‑ Neurosci 11(June):1–14 formatics 25(3):415–416 18. Cheng CY, Chen SP, Liao YC, Fuh JL, Wang YF, Wang SJ (2018) Elevated 41. Irizarry RA, Bolstad BM, Collin F, Cope LM, Hobbs B, Speed TP (2003) circulating endothelial-specific microRNAs in migraine patients: a pilot Summaries of Affymetrix GeneChip probe level data. Nucleic Acids Res study. Cephalalgia 38(9):1585–1591 31(4):e15–e15 19. Zhai Y, Zhu YY (2018) MiR-30a relieves migraine by degrading CALCA. 42. Bolstad BM, Irizarry RA, Astrand M, Speed TP (2003) A comparison of Eur Rev Med Pharmacol Sci 22(7):2022–2028 normalization methods for high density oligonucleotide array data 20. Chen S, Chang Y, Chou C, Juan C, Lee H, Chen L et al (2021) Circulat‑ based on variance and bias. Bioinformatics 19(2):185–193 ing microRNAs associated with reversible cerebral vasoconstriction 43. Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W et al (2015) limma syndrome. Ann Neurol 89(3):459–473 powers differential expression analyses for RNA-sequencing and micro‑ 21. Greco R, de Icco R, Demartini C, Zanaboni AM, Tumelero E, Sances G array studies. Nucleic Acids Res 43(7):e47 et al (2020) Plasma levels of CGRP and expression of specific microRNAs 44. Leek JT, Johnson WE, Parker HS, Fertig EJ, Jaffe AE, Storey JD et al (2019) in blood cells of episodic and chronic migraine subjects: towards the sva: Surrogate variable analysis. R package version 3:882–883 identification of a panel of peripheral biomarkers of migraine? J Head‑ 45. Lei B, Zhou J, Xuan X, Tian Z, Zhang M, Gao W et al (2019) Circular ache Pain 21(1):1–12 RNA expression profiles of peripheral blood mononuclear cells in 22. Chen YH, Wang H (2021) The association between migraine and hepatocellular carcinoma patients by sequence analysis. Cancer Med depression based on miRNA biomarkers and cohort studies. Curr Med 8(4):1423–1433 Chem 28(27):5648–5656 46. Wang M, Liang L, Li L, Han K, Li Q, Peng Y et al (2017) Increased miR- 23. Andersen HH, Duroux M, Gazerani P (2016) Serum microRNA signatures 424-5p expression in peripheral blood mononuclear cells from patients in migraineurs during attacks and in pain-free periods. Mol Neurobiol with pemphigus. Mol Med Rep 15(6):3479–3484 53(3):1494–1500 47. Smith-Vikos T, Liu Z, Parsons C, Gorospe M, Ferrucci L, Gill TM et al (2016) 24. Gallelli L, Cione E, Peltrone F, Siviglia S, Verano A, Chirchiglia D et al A serum miRNA profile of human longevity: findings from the Baltimore (2019) Hsa-miR-34a-5p and hsa-miR-375 as biomarkers for monitoring Longitudinal Study of Aging (BLSA). Aging (Albany NY) 8(11):2971 the effects of drug treatment for migraine pain in children and adoles‑ 48. Freedman JE, Ercan B, Morin KM, Liu CT, Tamer L, Ayaz L et al (2012) cents: a pilot study. J Clin Med 8(7):928 The distribution of circulating microRNA and their relation to coronary 25. Tafuri E, Santovito D, de Nardis V, Marcantonio P, Paganelli C, Affaitati G disease. F1000Res 1:50 et al (2015) MicroRNA profiling in migraine without aura: pilot study. 49. Benjamini Y, Yekutieli D (2001) The control of the false discovery rate in Ann Med 47(6):468–473 multiple testing under dependency. Ann Statist 29(4):1165–88 26. Aczél T, Benczik B, Ágg B, Körtési T, Urbán P, Bauer W et al (2022) Disease- 50. Pérez LO, González-José R, García PP (2016) Prediction of non-genotoxic and headache-specific microRNA signatures and their predicted mRNA carcinogenicity based on genetic profiles of short term exposure targets in peripheral blood mononuclear cells in migraineurs: role of assays. Toxicol Res 32(4):289–300 inflammatory signalling and oxidative stress. J Headache Pain 23(1):1–20 51. Gunther EC, Stone DJ, Gerwien RW, Bento P, Heyes MP (2003) Prediction 27. Pérez RF, Santamarina P, Fernández AF, Fraga MF (2019) Epigenetics and of clinical drug efficacy by classification of drug-induced genomic lifestyle: the impact of stress, diet, and social habits on tissue homeosta‑ expression profiles in vitro. Proc Natl Acad Sci 100(16):9608–9613 sis. Epigenetics Regen 11:461-89 52. Man MZ, Dyson G, Johnson K, Liao B (2004) Evaluating methods for 28. Manokaran S, Binoy A, Bhat D, Babu S, Bhat JG, AH MR (2021) Nutrig‑ classifying expression data. J Biopharm Stat 14(4):1065–1084 enomics in lifestyle disorders: a review. ECS Transactions 107(1):9249 53. Breiman L, Friedman JH, Olshen RA, Stone CJ (2017) Classification and 29. Alegría-Torres JA, Baccarelli A, Bollati V (2011) Epigenetics and lifestyle. regression trees. Routledge. Chapman & Hall/CRC, Boca Raton. https://​ Epigenomics 3(3):267–277 doi.​org/​10.​1201/​97813​15139​470 30. Lee PH, Macfarlane DJ, Lam TH, Stewart SM (2011) Validity of the 54. Breiman L (1996) Bagging predictors. Mach Learn 24(2):123–140 international physical activity questionnaire short form (IPAQ-SF): A 55. Díaz-Uriarte R, Alvarez de Andrés S (2006) Gene selection and classifica‑ systematic review. Int J Behav Nutr Phys Act 8(1):1–11 tion of microarray data using random forest. BMC Bioinform 7(1):1–13 31. Melhado EM, Thiers Rister HL, Galego DR, de Oliveira AB, Buttarello IA, 56. Diaz-Uriarte R (2007) GeneSrF and varSelRF: a web-based tool and R Belucio IS et al (2020) Allodynia in menstrually related migraine: score package for gene selection and classification using random forest. BMC assessment by Allodynia Symptom Checklist (ASC-12). Headache Bioinform 8(1):1–7 60(1):162–70 57 Ru Y, Kechris KJ, Tabakoff B, Hoffman P, Radcliffe RA, Bowler R et al (2014) 32. Kosinski M, Bayliss MS, Bjorner JB, Ware JEJ, Garber WH, Batenhorst The multiMiR R package and database: Integration of microRNA-target A et al (2003) A six-item short-form survey for measuring headache interactions along with their disease and drug associations. Nucleic impact: the HIT-6. Qual Life Res 12(8):963–974 Acids Res 42(17):e133 33. Stewart WF, Lipton RB, Dowson AJ, Sawyer J (2001) Development and 58. Ulgen E, Ozisik O, Sezerman OU (2019) pathfindR: an R package for testing of the Migraine Disability Assessment (MIDAS) Questionnaire to comprehensive identification of enriched pathways in omics data assess headache-related disability. Neurology 56(6 Suppl 1):S20–S28 through active subnetworks. Front Genet 10:858 34. Beck A, Steer R, Brown G (1996) Manual for the Beck Depression 59 Dogan B, Gumusoglu E, Ulgen E, Sezerman OU, Gunel T (2022) Inte‑ Inventory-II. In: Gellman MD, Turner JR (eds) San Antonio Psychol Corp. grated bioinformatics analysis of validated and circulating miRNAs in Springer New York, New York, NY. https://​doi.​org/​10.​1007/​978-1-​4419-​ ovarian cancer. Genomics Inform 20(2):e20 1005-9_​441 Gallardo et al. The Journal of Headache and Pain (2023) 24:11 Page 12 of 13 60 Mondal I, Kulshreshtha R (2021) Potential of microRNA based diagnos‑ 85. Turnpenny PD, Ellard S (2012) Alagille syndrome: pathogenesis, diagno‑ tics and therapeutics in glioma: a patent review. Expert Opin Ther Pat sis and management. Eur J Hum Genet 20(3):251–257 31(1):91–106. https://​doi.​org/​10.​1080/​13543​776.​2021.​18377​75 86. Rubino E, Fenoglio P, Gallone S, Govone F, Vacca A, de Martino P et al 61. Manack AN, Buse DC, Lipton RB (2011) Chronic migraine: Epidemiology (2013) Genetic variants in the NOTCH4 gene influence the clinical and disease burden. Curr Pain Headache Rep 15(1):70–78 features of migraine. J Headache Pain 14(1):1–5 62. Alpuente A, Gallardo VJ, Torres-Ferrús M, Álvarez-Sabin J, Pozo-Rosich P 87. Vriens J, Appendino G, Nilius B (2009) Pharmacology of vanil‑ (2020) Short and mid-term predictors of response to Onabotulinumtox‑ loid transient receptor potential cation channels. Mol Pharmacol inA: real-life experience observational study. Headache 60(4):677–685 75(6):1262–1279 63. Keller A, Leidinger P, Lange J, Borries A, Schroers H, Scheffler M et al 88. Julius D (2013) TRP channels and pain. Annu Rev Cell Dev Biol (2009) Multiple sclerosis: microRNA expression profiles accurately differ‑ 29:355–384 entiate patients with relapsing-remitting disease from healthy controls. 89. Jardín I, López JJ, Diez R, Sánchez-Collado J, Cantonero C, Albarrán L PLoS ONE 4(10):e7440 et al (2017) TRPs in pain sensation. Front Physiol 8:392 64. Geekiyanage H, Jicha GA, Nelson PT, Chan C (2012) Blood serum 90. Russell FA, King R, Smillie SJ, Kodji X, Brain SD (2014) Calcitonin miRNA: non-invasive biomarkers for Alzheimer’s disease. Exp Neurol gene-related peptide: physiology and pathophysiology. Physiol Rev 235(2):491–496 94(4):1099–1142 65 Schipper HM, Maes OC, Chertkow HM, Wang E (2007) MicroRNA 91. Goadsby PJ, Edvinsson L (1993) The trigeminovascular system and expression in Alzheimer blood mononuclear cells. Gene Regul Syst Bio. migraine: studies characterizing cerebrovascular and neuropeptide 1:GRSB-S361 changes seen in humans and cats. Ann Neurol 33(1):48–56 66. Schroeter ML, Abdul-Khaliq H, Krebs M, Diefenbacher A, Blasig IE (2008) 92. Tso AR, Goadsby PJ (2017) Anti-CGRP monoclonal antibodies: Serum markers support disease-specific glial pathology in major the next era of migraine prevention? Curr Treat Options Neurol depression. J Affect Disord 111(2–3):271–280 19(8):1–11 67. Paul S, Vázquez LAB, Uribe SP, Reyes-Pérez PR, Sharma A (2020) Current 93. Edvinsson L, Haanes KA, Warfvinge K, Krause DN (2018) CGRP as the status of microrna-based therapeutic approaches in neurodegenerative target of new migraine therapies—successful translation from bench disorders. Cells 9(7):1–26 to clinic. Nat Rev Neurol 14(6):338–350 68. Chakraborty C, Sharma AR, Sharma G, Lee SS (2021) Therapeutic 94. Charles A, Pozo-Rosich P (2019) Targeting calcitonin gene- advances of miRNAs: a preclinical and clinical update. J Adv Res related peptide: a new era in migraine therapy. Lancet 28:127–138 394(10210):1765–1774 69. Zhai Y, Zhu YY (2018) MiR-30a relieves migraine by degrading CALCA. 95. Benemei S, Dussor G (2019) TRP channels and migraine: recent devel‑ Eur Rev Med Pharmacol Sci 22(7):2022–2028 opments and new therapeutic opportunities. Pharmaceuticals 12(2):54 70. Feil R, Fraga MF (2012) Epigenetics and the environment: emerging 96. Torres-Ferrús M, Gallardo VJ, Alpuente A, Caronna E, Gine-Cipres E, patterns and implications. Nat Rev Genet 13(2):97–109 Pozo-Rosich P (2021) The impact of anti-CGRP monoclonal antibod‑ 71. Waterland RA, Jirtle RL (2003) Transposable elements: targets for ies in resistant migraine patients: a real-world evidence observational early nutritional effects on epigenetic gene regulation. Mol Cell Biol study. J Neurol 268(10):3789–98 23(15):5293–5300 97. Alpuente A, Gallardo VJ, Asskour L, Caronna E, Torres-Ferrus M, Pozo- 72 Lee KW, Pausova Z (2013) Cigarette smoking and DNA methylation. Rosich P (2022) Salivary CGRP can monitor the different migraine Front Genet 4:132 (PubMed; 2013) phases: CGRP (in) dependent attacks. Cephalalgia 42(3):186–96 73. Ponomarev I, Wang S, Zhang L, Harris RA, Mayfield RD (2012) Gene 98. Alpuente A, Gallardo VJ, Asskour L, Caronna E, Torres-Ferrus M, Pozo- coexpression networks in human brain identify epigenetic modifica‑ Rosich P (2022) Salivary CGRP and Erenumab Treatment Response: tions in alcohol dependence. J Neurosci 32(5):1884–1897 Towards Precision Medicine in Migraine. Ann Neurol 92(5):846–59 74. Kamat PK, Mallonee CJ, George AK, Tyagi SC, Tyagi N (2016) Homocyst‑ 99. Yang D gang, Gao Y yao, Yin Z qun, Wang X rui, Meng X she, Zou T feng, eine, alcoholism, and its potential epigenetic mechanism. Alcohol Clin et al (2022) Roxadustat alleviates nitroglycerin-induced migraine in Exp Res 40(12):2474–2481 mice by regulating HIF-1α/NF-κB/inflammation pathway. Acta Pharma‑ 75. Franklin TB, Russig H, Weiss IC, Gräff J, Linder N, Michalon A et al (2010) col Sin (2):308–20 Epigenetic transmission of the impact of early stress across generations. 100. Mirouliaei M, Fallah R, Bashardoost N, Partovee M, Ordooei M (2012) Biol Psychiatry 68(5):408–415 Efficacy of levothyroxine in migraine headaches in children with sub‑ 76. Wallace RG, Twomey LC, Custaud MA, Turner JD, Moyna N, Cummins clinical hypothyroidism. Iran J Child Neurol 6(4):23 PM et al (2018) The role of epigenetics in cardiovascular health and 101. Paolucci M, Altamura C, Vernieri F (2021) The role of endothelial ageing: A focus on physical activity and nutrition. Mech Ageing Dev dysfunction in the pathophysiology and cerebrovascular effects of 174:76–85 migraine: a narrative review. J Clin Neurol 17(2):164 77. Menke A, Binder EB (2022) Epigenetic alterations in depression and 102. Peres MFP, del Rio MS, Seabra ML, Tufik S, Abucham J, Cipolla-Neto J antidepressant treatment. Dialogues Clin Neurosci 16(3):395–404 et al (2001) Hypothalamic involvement in chronic migraine. J Neurol 78. Terlizzi R, Bacalini MG, Pirazzini C, Giannini G, Pierangeli G, Garagnani Neurosurg Psychiatry. 71(6):747–51 P et al (2018) Epigenetic DNA methylation changes in episodic and 103. Borsook D, Erpelding N, Lebel A, Linnman C, Veggeberg R, Grant PE et al chronic migraine. Neurol Sci 39(1):67–68 (2014) Sex and the migraine brain. Neurobiol Dis 68:200–214 79. Bocchio-Chiavetto L, Maffioletti E, Bettinsoli P, Giovannini C, Big‑ 104. MacGregor EA, Frith A, Ellis J, Aspinall L, Hackshaw A (2006) Incidence notti S, Tardito D et al (2013) Blood microRNA changes in depressed of migraine relative to menstrual cycle phases of rising and falling patients during antidepressant treatment. Eur Neuropsychopharmacol estrogen. Neurology 67(12):2154–2158 23(7):602–611 105. Phillips WJ, Ostrovsky O, Galli RL, Dickey S (2006) Relief of acute 80. Blažeković A, Borovečki F (2021) Psychiatric Comorbidities in Parkinson’s migraine headache with intravenous oxytocin: report of two cases. J Disease Seen through the Prism of Genomics and Epigenetics. Psychiatr Pain Palliat Care Pharmacother 20(3):25–28 Danub 33(suppl 4):456–462 106 Tzabazis A, Kori S, Mechanic J, Miller J, Pascual C, Manering N et al 81. Citraro R, Leo A, de Caro C, Nesci V, Gallo Cantafio ME, Amodio N et al (2017) Oxytocin and migraine headache. Headache 57:64–75 (2020) Effects of histone deacetylase inhibitors on the development 107. Chandramohan R, Wu PY, Phan JH, Wang MD (2013) Benchmark‑ of epilepsy and psychiatric comorbidity in WAG/Rij rats. Mol Neurobiol ing RNA-Seq quantification tools. Annu Int Conf IEEE Eng Med Biol 57(1):408–421 Soc 2013:647–50. https://​doi.​org/​10.​1109/​EMBC.​2013.​66095​83 82. Cain SM, Snutch TP (2011) Voltage-gated calcium channels and disease. 108. Hayer KE, Pizarro A, Lahens NF, Hogenesch JB, Grant GR (2015) BioFactors 37(3):197–205 Benchmark analysis of algorithms for determining and quantifying 83. Pietrobon D (2013) Calcium channels and migraine. Biochimica et full-length mRNA splice forms from RNA-seq data. Bioinformatics Biophysica Acta (BBA)-Biomembranes 1828(7):1655–65 31(24):3938–3945 84. Anne D, Elisabeth TL, Marie-Germaine B (2002) The genetics of 109. Robinson MD, Oshlack A (2010) A scaling normalization method for Migraine. Lancet Neurol 1:285–293 differential expression analysis of RNA-seq data. Genome Biol 11(3):1–9 Gallardo et al. The Journal of Headache and Pain (2023) 24:11 Page 13 of 13 110. Rao MS, van Vleet TR, Ciurlionis R, Buck WR, Mittelstadt SW, Blomme EAG et al (2019) Comparison of RNA-Seq and microarray gene expres‑ sion platforms for the toxicogenomic evaluation of liver from short- term rat toxicity studies. Front Genet 9:636 111. Kemper RHA, Meijler WJ, Korf J, ter Horst GJ (2001) Migraine and function of the immune system: a meta-analysis of clinical literature published between 1966 and 1999. Cephalalgia 21(5):549–557 112. Forcelini CM, Dantas DCM, Luz C, Santin R, Stein AT, Barros HMT et al (2011) Analysis of leukocytes in medication-overuse headache, chronic migraine, and episodic migraine. Headache 51(8):1228–1238 113. Grazzi L, Corsini E, Ciusani E, Usai S, Vasco C, Bussone G (2014) Evalu‑ ation of immune parameters in chronic migraine with medication overuse. Neurol Sci 35(SUPPL. 1):171–173 114. Kok MGM, de Ronde MWJ, Moerland PD, Ruijter JM, Creemers EE, Pinto-Sietsma SJ (2018) Small sample sizes in high-throughput miRNA screens: a common pitfall for the identification of miRNA biomarkers. Biomol Detect Quantif 15:1–5 Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional affiliations. Ready to submit your research ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations • maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions