Abstract
Aims: This study was aimed to verify changes in the levels of hypothalamic neuropeptides in migraineurs under preventive treatment with amitryptiline and flunarizine. Thirty-nine migraine patients with a body mass index <25 kg/m2 and without endocrinological or metabolic diseases were assigned to two treatment groups, one receiving amitryptiline, the other flunarizine, for 3 months. Orexin-A, orexin-B and neuropeptide-Y plasma levels were measured at the basal time, at the 1st, 2nd and 3rd months of preventive treatment.
Results: A statistically significant reduction in plasma orexin-A and orexin-B levels emerged in both groups. Conversely, plasma neuropeptide-Y levels were markedly increased, with the highest levels at the 2nd and 3rd months, in both patient groups. Orexin-A levels were also negatively correlated to weight gain in both groups during the treatment period.
Conclusions: These results suggest that changes in the levels of hypothalamic orexinergic peptides may contribute to body weight increase occurring in migraineurs during amitryptiline or flunarizine prophylactic treatment.
Introduction
Weight gain is one of the most common undesirable adverse event of migraine prophylactic treatment and is sometime the cause of its discontinuation (1,2). Amitryptiline (AMT), a tricyclic antidepressant routinely prescribed for migraine prevention, has been associated with the highest incidence of weight gain (3,4). In a recent randomized, double-blind, comparative trial of AMT and topiramate for migraine prophylaxis, the large majority of AMT-treated patients had a significant weight gain over the treatment period (5). Furthermore, it has been reported that migraine patients with major weight gain show a significant elevation of clinical markers of cardiovascular risk (6).
Flunarizine (FLN), a mixed Ca2+ blocker/H1 antagonist which is often used as migraine-preventive medication in Europe but is not available in the US, is also responsible for a significant increase of weight (3,7).
An increase in body mass index (BMI) due to 12-week treatment with either AMT or FLN has been associated with an increase in serum levels of leptin, insulin and C-peptide in a clinical study involving migraine patients with BMI < 25 kg/m2 and without any endocrinological, immunological or chronic diseases (8). However, no relationship between peptide changes and variation in headache frequency was investigated in this study. An increase in leptin levels was also observed in children with migraine treated with cyproheptadine and FLN as antimigraine prophylactic drugs but not with AMT or propranolol (9).
Several lines of evidence support the involvement of the hypothalamus in the regulation of many aspects of energy homeostasis, adjusting both food intake and energy expenditure in response to a wide range of nutritional and non-nutritional signals via the synthesis of various anorectic and orexigenic neuropeptides (2). These latter include the neuropeptides orexin-A (OR-A) and orexin-B (OR-B) that are synthesized by lateral hypothalamic area and the dorsomedial hypothalamic nucleus, and neuropeptide-Y (NP-Y) which is synthesized by the hypothalamic arcuate nucleus (ARC) in the magnum paraventricular nucleus (10). Both orexin and NP-Y pathways play a pivotal role in feeding and appetite regulation. They promote food intake and functionally co-operate resulting in an enhanced feeding behaviour (11,12). NP-Y secretion is also regulated by peripheral signals. In particular, insulin and leptin hormones act as regulators of energy homeostasis, through their inhibitory effect at medial hypothalamic ARC–NPY neurons (13). With the disruption of these inhibitory signals, NP-Y neurons become overactive, contributing to hyperphagia and obesity (10).
Based on the above findings, the present study was aimed at investigating the effects of AMT and FLN prophylactic treatment on peripheral levels of OR-A, OR-B and NP-Y in migraine patients and to evaluate the relationship of these levels with changes in weight and BMI as well as variation in headache frequency during a 3-month treatment period.
Subjects and methods
A total of 39 consecutive Caucasian migraine patients (31 females and 8 males; age range, 22–45 years) with BMI < 25 kg/m2 (range, 22–27 kg/m2) were included in the study. Patients gave their informed consent to participate in the study and the protocol was approved by the institutional review board of our department. They had a diagnosis of migraine without aura according to the current ICHD-II criteria (14) and experienced at least three severe disabling attacks in the last 3 months and, therefore, were eligible to receive a migraine-preventive treatment.
Exclusion criteria included: a preventive treatment for migraine during the previous 2 months, a concurrent tension-type headache, a concomitant psychiatric disorder (including anxiety and depression, assessed through the administration of respective Beck Inventories) (15), regular drug use including oral contraceptives, alcohol or drug abuse, liver or kidney dysfunction, any metabolic or endocrinological disorder (with particular regard to diabetes), rheumatological or haematological chronic diseases, acute or chronic infectious and/or inflammatory disorders, and other conditions or pathologies that could modify body weight.
Details of patients and controls
Data are expressed as mean ±2 SD.
Each patient had a complete physical and neurological examination before being enrolled in the study. Headache frequency and characteristics (intensity, location) as well as associated symptoms were recorded by the patients using a headache diary at the basal time and every 4 weeks in the 3-month treatment period. Patients were assessed monthly by a clinician and their headache diary was checked for headache frequency and characteristics at the basal time. Their BMI was also calculated monthly according to the standard formula. Being overweight was defined as a BMI ≥ 25 kg/m2 and obesity as a BMI ≥ 30 kg/m2. Details of patients and controls are reported in Table 1.
Plasma OR-A, OR-B and NP-Y concentrations were measured with radioimmunoassay (RIA) methods. Control values were provided by 20 age-matched healthy individuals (4 males, 16 females; mean age, 34.9 ± 7.6 years; mean BMI, 24.08 ± 0.79 kg/m2).
Fasting blood samples were taken from the antecubital veins of patients at 8–9 a.m. to assess plasma OR-A, OR-B and NP-Y. All women in patient and control groups were normally cycling and were assessed at the follicular phase (days 5–8 of the cycle). Blood was collected in the presence of ethylenediaminetetraacetic acid (EDTA; 1 mg/ml), and plasma was separated and stored at –80°C until the assay. The blood samples were immediately centrifuged.
Levels of OR-A and OR-B were determined in duplicate by RIA (Peninsula Laboratories, San Carlos, CA, USA) following an extraction procedure using Sep-Pak C18 cartridges (Waters Associates, Milford, MA, USA). The same extraction procedure was carried out before plasma level determination of NP-Y. Human NP-Y RIA kits was provided by Phoenix Peptide Pharmaceutical (Belmont, CA, USA). OR-A and OR-B RIA kits enable the reliable measurement of OR-A as low as 10 pg/ml. Intra-assay coefficients of variation (CV) for OR-A and OR-B were 5% and 5.3% (10 measurements), and interassay CV 6% and 7% (10 measurements), respectively. Lowest detection limit for NP-Y RIA kit is 26 pg/ml. Intra-assay and inter-assay CVs for NP-Y were 4% (10 measurements) and 7% (10 measurements), respectively.
Statistical analysis
Statistical analysis was carried out using the Software Statistics Release v6.0. Data were tested for normality with Kolmogorov–Smirnov normality test. Comparisons between groups and among times were made using analysis of variance (ANOVA) for repeated measures for non-parametric variables, followed by the Mann–Whitney U-test for independent samples for the intergroup comparison. Spearman correlation test was used for the correlation analysis.
Results
No significant differences were found in the demographic features of migraine patients between the two treatment groups. No difference in the frequency of headache emerged at the baseline between AMT and FLN groups. Compared with controls, migraineurs at the baseline had lower peripheral levels of both OR-A, OR-B and NP-Y. Headache frequency decreased due to both antimigraine preventive drugs and this decrease was significant at the 2nd and 3rd months (Table 1).
All patients in the AMT and FLN groups completed the study. None of the patients discontinued the treatment because of relevant side effects. AMT and FLN treatment induced a significant weight increase from the 1st to the 3rd month of treatment, and this weight gain was greater in the FLN group (ANOVA, P < 0.001 and P < 0.0002, respectively). This corresponded to a slight increase in BMI from the 1st month of treatment with both prophylactic drugs.
Blood concentrations of OR-A, OR-B and NP-Y (all expressed as pg/ml) in the 3-month treatment AMT and FLN period
No correlation emerged among neuropeptide levels variation at each time of the study. They also did not correlate to headache frequency at the basal time and during the 3-month treatment period with AMT and FLN. Interestingly, OR-A levels negatively correlate with weight gain and BMI at the 3rd month in both treatment groups (AMT: R = −0.28, P = 0.05 and R = −0.31, P = 0.05; FLU: R = −0,31, P = 0.04 and R = −0.37, P = 0.03, respectively). A trend toward a negative association between OR-A levels and both weight gain and BMI was also found but without reaching the minimum level of statistical significance. A slight but significant negative correlation between OR-A, but not OR-B measured at basal time and age also emerged in both treatment groups (R = −0.27, P = 0.05 and R = −0.31, P = 0.05, for AMT and FLN groups respectively).
Discussion
The present study investigated the peripheral levels of some hypothalamic neuropeptides (OR-A, OR-B and NP-Y) involved in appetite and weight gain regulation in a small group of migraine without aura patients treated with AMT or FLN as a prophylactic treatment. We found that the 3-month treatment with both preventive drugs induced a significant reduction in plasma levels of the two orexins and this decrease was more relevant, from the 1st month of treatment, for FLN. Conversely, an increase in peripheral levels of NP-Y was observed in the same treatment period for both drugs. According to previous research, we also detected in all patients a small, but significant, increase in body weight due to the two drugs and this was more evident in migraine patients treated with FLN.
To our knowledge, this is the first study aimed at determining the peripheral levels of OR-A and OR-B in episodic migraineurs and to verify changes of these levels due to preventive antimigraine drugs known to induce weight gain. We interpreted the decrease of both orexins as a consequence of the two prophylactic treatments as an hypothalamic response aimed to counteract body weight increase induced by both drugs. Experimental data in animals are contrasting in this regard. Acute injection of orexins seems, in fact, to increase and not decrease short-term feedings in rats and mice and also systemic injection of anti-orexin antibodies resulted in reduced food intake (16). On the other hand, additional lines of evidence support the involvement of orexins in increasing appetite and food intake in animals through the stimulation of the ventral tegmental area, which is the origin of dopamine projections implicated in motivation and reward (17). Furthermore, orexynergic projections to area postrema and subjacent nucleus of the solitary tract are responsible for increasing consummatory (meal size) but not appetitive (meal frequency) responses due to reduced postingestive feedback inhibition induced by OR-A (18). To reconcile these findings with our results, we hypothesized in migraine patients a disturbance of mechanisms controlled by hypothalamic neuropeptides, which can be related to insulin sensitivity and leptin resistance as suggested by experimental data and clinical findings on migraineurs (19–21). They can be further affected by FLN or AMT preventive treatment.
Interestingly, our results concur with some observations obtained in humans, such as the decreased plasma levels of OR-A in adult obese subjects (22) and the negative correlation between OR-A and body weight and BMI demonstrated in obese children (23). As in this latter research, a negative correlation emerged in our study between OR-A, body weight and BMI as well as with age. Conversely, to our knowledge, the reduction in OR-B in relation to the increase in body weight is a novel finding of our study.
Our results apparently contrast with previous evidence from our group of significantly increased levels of OR-A in the cerebrospinal fluid (CSF) of patients with chronic migraine and medication-overuse headache (24). In this study, we did not investigate orexin peripheral levels nor relate them to patient weight. We interpreted the OR-A increase in CSF as a compensatory response to chronic pain and related stress in chronic migraine patients and suggested that it could play a pivotal role in driving drug seeking in medication-overuse headache.
In the present research, we preferred to determine orexins levels in plasma instead that in CSF, where their concentration is higher, for ethical reasons and because plasma samples are easier to obtain at different time points and does not require invasive procedures. The different source may, therefore, account, at least in part, for discrepancies in the results. Moreover, we assessed episodic migraine patients with attacks frequent and severe enough to require a prophylactic treatment during an interictal period but did not investigate peripheral changes due to migraine crises. It cannot be excluded that the orexinergic system may be involved during migraine crises but data about changes in the orexin levels both in blood and CSF in the ictal period are not available at the moment (25). It can also be hypothesised that this system could be affected by preventive treatment with both AMT and FLN.
In our study, OR-A and OR-B levels are lower in migraine patients then in controls at the basal time. This novel finding could be explained as an impairment of orexinergic pathways in migraineurs, also in the episodic form without preventive treatment. Further research involving larger numbers of subjects and longitudinal measures in controls is needed to investigate this aspect better.
In the present study, we did not enrol migraine patients with concomitant anxiety and depression to avoid the possible interference of concomitant psychiatric disorders with food intake; therefore, their contribution to changes of peripheral neuropeptide levels at the baseline and at each times of the study can be excluded. It would be interesting to determine levels of hypothalamic peptides in patients with concurrent anxiety and/or depression which have been reported to have a high prevalence in migraineurs (26). This aspect should be assessed in future studies aimed to verify the contribution of concomitant psychiatric disorders to the changes of hypothalamic neuropeptides levels migraine.
As for orexins, migraine patients at the basal time showed lower levels of NP-Y compared to controls. This evidence agrees with previous findings of our group in patients with juvenile migraine assessed interictally (27), in which the reduction of NP-Y levels was attributed to an impairment of the sympathetic system. It would be important to verify, in further research, the hypothesis that NP-Y is particularly low in patients with a high frequency of attacks, and that NP-Y increases according to the effectiveness of preventive treatment. This could be confirmed by measuring NP-Y levels in a group of patients treated unsuccessfully with AMT and FLN and in patients with sporadic attacks.
An additional novel finding of our study is the demonstration of an increase in the plasma levels of NP-Y induced by AMT and FLN treatment. Increased NP-Y activity appears to promote feeding and fat deposition and increases body mass (28–30). This increase can, therefore, contribute to the weight gain of migraineurs treated with either AMT or FLN. As suggested in a previous study (8), the α1-adrenergic receptor blocking of AMT may contribute to the ‘leptin resistance’ in migraineurs by reducing transport in blood–brain barrier (BBB). Likewise, due to its high lipophilic nature, FLN might interact with leptin transport system in the BBB and this may also contribute to the leptin resistance in migraineurs treated with this drug. A further mechanism involved in weight gain due to AMT is the blockage of H1 histamine receptors. This is also supported by the demonstration of a regulatory action of body weight and adiposity by postsynaptic H1 receptors in hypothalamus (31). The action of FLN on high voltage-gated Ca2+ channels can play a role in modulating Ca2+ contents in NP-Y and pro-opiomelanocortin neurons (12). The stimulatory effects of orexins are, in fact, controlled by the increase in cytosolic Ca2+ levels of NP-Y neurons in ARC and Ca2+ increase may also play a pivotal role in inhibiting leptin effects (32). The Ca2+ channel blocking effect of FLN may, therefore, influence these systems resulting in the modification of the peripheral levels of these neuropeptides. Effect on D2-like receptors on the excitatory presynaptic terminals impinging onto orexin neurons, and indirectly influencing NP-Y neurons can also be advocated as an additional mechanism involved in weight gain due to FLN treatment in migraine (33). Conversely, inhibition of re-uptake of both norepinephrine and 5-HT antagonism of AMT, which seems to mediate the anti-migraine effect of this drug, is unlikely to contribute to weight gain due to prophylactic treatment of migraine (34).
Independently of the putative effects of AMT and FLN involving appetite increase and weight gain in migraineurs, it can be excluded that reduction of migraine attacks render patients more prone to eat by improving mood and quality of life. Migraine patients, in fact, often describe loss of appetite during attacks, believed to be due to the nausea and the activation of anorexic neurons in the ventromedial hypothalamus during attacks. It is possible that a low frequency of migraine attacks allows patients to spend more time in normal appetite and this option cannot be excluded in interpreting our results. It would be interesting to investigate in future research changes in diet, caloric intake (in relation to daily exercise or level of physical activity) and waist circumference and as well as mood changes after preventive treatment which can affect appetite and feeding behaviour.
Obesity has been reported to be a risk factor for transformed migraine or chronic migraine (6,35). Insulin resistance has also been demonstrated in migraine and has been considered to play a putative role in the co-morbidity between migraine and metabolic syndrome and vascular disorders (36,37). In our study, the short-lasting treatment with AMT or FLN did not induce a significant increase in BMI which was not in the range of obesity (30–34.9 kg/m2) or morbidly obesity (>35 kg/m2) in any of our patients at each time of the study. Therefore, the risk of a paradoxical increase in the attack frequency due to the weight gain induced by both drugs is unlikely (6). Furthermore, a larger sample of migraine patients should be studied to differentiate those that are prone to increase weight from those that tolerate drugs without weight gain.
Conclusions
The present study furnishes novel information on the pathophysiological and neurobiological basis of weight gain due to prophylactic treatment in migraine. The demonstration of the involvement of OR-A, OR-B and NP-Y in this increase underlies the relevance of hypothalamus in integrating these mechanisms. To investigate these aspects better, further research is needed, involving larger samples of patients, longitudinal determinations in controls and correlation between migraine features and body weight.
Footnotes
Acknowledgements
The authors thank John A. Toomey for editing the English and Marisa M. Morson for technical assistance. This study had no sponsorship. The authors report no disclosures nor any conflict of interest.