Introduction
Migraine associated with recurrent vestibular symptoms are no strangers to clinicians. Migraine has been estimated to affect about 15% of the general population [
1], and up to 85% patients with migraine suffers from balance problems and dizziness [
2,
3]. Recently, International Headache Society and Bárány Society have prompted efforts to recognize and classify vestibular migraine (VM), which incorporates vestibular and migraine symptoms, as a novel subtype of migraine in International Headache Classification of Headache Disorders, 3rd edition [
4]. Whether vestibular symptoms are associated symptoms as photophobia and phonophobia in migraine, or the distinctive entity that correlates with migraine are still in debate [
5].
Migraine has been proposed as a complex sensory disorder with abnormal sensitization in central trigeminal-vascular system [
6,
7]. Central sensitization of trigeminovascular neurons in trigeminal nucleus caudalis (TNC) has been considered to be the neural basis for sensory hypersensitivity in migraine, such as tactile allodynia [
6]. We previously showed that the development of vertigo was lagged behind migraine by an average of 6 years in VM [
7,
8], and the prevalence of vertigo prominently increased in chronic migraine (CM) compared with episodic migraine [
9]. Whether vestibular symptom is one of clinical manifestations of central sensitization for migraine remains unproven. Clinical data demonstrated that patients with VM exhibited abnormally elevated vestibular-ocular threshold [
10], as well as reduced roll tilt threshold [
11] compared to migraine patients without vestibular symptoms and patients with a peripheral inner-ear disorder, indicating that dysfunction of the vestibular nuclei (VN) might be an underlying mechanism for vestibular symptoms in VM. As an important sensorimotor center in brainstem, VN and its direct connection with TNC had been demonstrated in rodents [
12]. Further, nociceptive trigeminal stimulation could induce vertigo in patients with migraine rather than controls [
13]. Above data suggested that migraine-mediated sensitization of the TNC might affect the sensitivity of VN which receives TNC projections. However, whether vestibular sensitization is the potential mechanism for vestibular symptoms in migraine remains largely unknown.
The neuropeptide calcitonin gene-related peptide (CGRP) plays a crucial role in migraine based on its efficacy of anti-CGRP treatment in clinical trials against headache, and CGRP administration triggers delayed migraine-like headache in patients with migraine [
6,
14]. Pharmacological and electrophysiological studies demonstrated that CGRP might facilitate nociceptive signals in the spinal cord, rather than produce nociception, which contributes to the development of central sensitization [
15]. Our previous study also found a significantly increased expression of CGRP in TNC in a rat model of CM, and that was parallel to the development of central sensitization [
16]. Meanwhile, rodents exposed to repeated bouts of rotary stimulation had an increased number of CGRP expressing neurons in VN, whilst anisodamine could significantly decrease the CGRP expression in VN [
17], pointing that CGRP might be a mediator in trigeminal-mediated sensitization of VN.
Here our aim was to determine whether recurrent nitroglycerine (NTG) administration could induce vestibular dysfunction that is associated with migraine, and subsequently the effectiveness of CGRP knockdown to relieve these symptoms. We also wanted to determine central sensitization in TNC as a precursor in migraine-associated vestibular dysfunction. Using retrograde labeling of central trigeminocervical and vestibular neurons, we analyzed the functional connection changes caused by recurrent NTG administration and whether this manifested as a potential neural basis of VM in a preclinical model.
Discussion
Vestibular symptoms are prevalent among patients with migraine [
2,
3]. The mechanism that relate vestibular symptoms to migraine had not been well elucidated. Thus, current therapies are based on experts’ experience and little clinical progress has been made to develop therapeutic strategies targeting this subpopulation. Central sensitization has been demonstrated as a primary pathophysiological process after CM, in which CGRP has an essential role [
6]. To date, clinical trials witness great success in anti-CGRP treatment on migraine headache [
45]. Therefore, whether sensitization in the vestibular nucleus attribute to the development of vestibular dysfunction after CM, and whether anti-CGRP treatment could attenuate vestibular symptoms in migraine needs further illustrating.
The present study revealed that the development of vestibular dysfunction coincide with that of hyperalgesia in a rat model of CM. Meanwhile, the structure of vestibular afferent terminals was preserved after CM. CM induced neuronal activation in the TNC and VN, and CM-activated neurons in the TNC were primarily TNC-projecting VN neurons. In vivo knockdown of CGRP in the trigeminal ganglion could alleviate neuronal activation and upregulation of CGRP in the VN, further attenuating vestibular dysfunction in CM rats. Collectively, these studies suggested the possibility of vestibular sensitization to impair vestibular function after CM, and anti-CGRP treatment to restore vestibular dysfunction in patients with CM.
Systemic recurrent NTG administration has been proven to be a reliable method to produce the preclinical CM model: inducing acute and sustained hypersensitivity, which mimics the core clinical characteristics of CM; causing distinctively associated features, like photophobia, facial grimace behaviors, and upregulation of CGRP in the TNC and dura mater [
15,
45]. However, little is known about the changes of vestibular-mediated behaviors in this CM model. Further, since up to one-fourth percentage of peripheral vestibular deficits, like hearing loss and endolymphatic hydrops, has been detected among migraine patients with vestibular symptoms and the development of vestibular symptoms often lagged several years behind headache [
8,
46], some researchers deduce that the vestibular dysfunction in migraine patients may result from the damage of peripheral vestibular apparatus [
40‐
42,
47]. In the present study, we first revealed that recurrent NTG administration produced allodynia and significant vestibular dysfunction, and the severity of vestibular dysfunction was comparable with kainic acid treated group. The development of vestibular dysfunction in our study is in line with clinical-based studies showing that most migraine patients are vestibular symptoms-free during the initial years of their migraine experience, but gradually develop motion intolerance or vestibular symptoms with increasing frequency of migraine attacks [
9]. Meanwhile, morphological examination of the afferent terminal impairments showed that the swelling and structure disorders were much less pronounced in the CM group when compared with kainic acid treated group. These data indicated that it might be more likely a central, rather peripheral, component contribute to the vestibular dysfunction in migraine.
Migraine has been more generally considered as a neurological disorder of sensitization, with heightened sensitivity to light, sound, smell and motion [
48]. Clinical and preclinical studies show that the central sensory systems are sensitized in migraine [
49]. Thus, a reasonable hypothesis is that sensitization of vestibular pathways may contribute to the enhanced motion sickness susceptibility and episodic attacks of vestibular symptoms in migraine [
50]. C-Fos has been widely used as an early indicator of neuronal activation that normally responds within 2 h [
51]. Long-term c-Fos expression after NTG administration suggests continuous activation of secondary sensory neurons [
37,
52]. In our study, significant fos-protein labeling in the superficial lamina of the TNC and VN after chronic NTG administration were observed. It is noteworthy that anatomic pathways giving extensive projections between brainstem regions associated with the TNC and the VN have been traced in the physiological condition [
12]. We then asked whether CM-activated neurons belonged to the population of trigeminovestibular neurons, and whether persistent activation of TNC neurons facilitated subsequent activation of VN neurons. To gain a further understanding, FluoroGold and CTB-555 were selected as retrograde tracers in vehicle and CM groups, mainly based on published reports of their sensitivity and uni-directional transportation [
22,
53]. Our immunofluorescent results showed that FluoroGold labeling neurons were predominantly located in the superlayer of the TNC, and CM-activated neurons in the TNC were primarily TNC-projecting VN neurons. Furthermore, to investigate whether changes observed in the VN were due to the hypersensitivity of TNC neurons or directly caused by NTG, viral vectors containing CGRP RNAi were injected into the trigeminal ganglion to knockdown CGRP synthesis in NTG-treated animals. We observed that c-fos expression was significantly decreased in the TNC among CM group, and similar pattern was also observed in the VN. Above data indicated the possibility of trigeminal-mediated sensitization of vestibular nucleus neurons in migraine.
CGRP is a 37-amino acid neuropeptide, and widely expressed in central and peripheral nervous system, with prominent localization in the outer laminae of the spinal cord dorsal horn and TNC [
15,
54]. The majority of CGRP in the TNC is synthesized in small- and medium-diameter neurons of trigeminal ganglion, and transported to central terminals [
15]. CGRP is believed to be a key regulator in central sensitization of trigeminovascular neurons, attributing to migraine headache and associated hypersensitivity [
6,
15]. Contrast to a nociceptive role, CGRP enhances the abnormal pain sensitivity rather than normal acute nociceptive signals, as evidenced by the fact that intrathecal administration of CGRP did not alter response thresholds to noxious thermal stimuli in normal rats, while anti-CGRP antiserum significantly blocked hyperalgesia under inflammatory condition [
55]. Previous study demonstrates that the CGRP protein is expressed in afferent terminals in the TNC [
15], and in the vestibular nuclei and cerebellum, CGRP immunoreactivity can be detected on neuronal somas [
54,
56]. In consistent with previous study, we found the extensive expression of CGRP immunoreactive fibres in the superlayer of the TNC and CGRP-positive neurons in the VN [
37,
56]. To date, little evidence shows the change of CGRP expression in the VN after CM. In this study, we noted that CGRP expression was significantly elevated in the VN after CM, suggesting that the endogenous CGRP expression change in the VN was in response to CM. We also found that CGRP positively stained neurons could be detected in four major vestibular nuclei after CM, with a slight more expression in medial vestibular nucleus according to the rat brain atlas of Paxinos and Watson [
56]. The medial vestibular nucleus is the largest nucleus within the vestibular complex and has various types of neurons [
57]. Medial vestibular nucleus neurons are essential to the maintenance of vestibulo-ocular reflex, especially in stabilizing the images during head movement [
58]. Clinical-based studies provide supporting evidence that ocular and perceptual thresholds are significantly increased in migraine patients with vestibular symptoms relative to those without vestibular symptoms, implying that sensitization of the medial vestibular nucleus might be a primary pathophysiological process underlying the vestibular hypersensitivity in migraine patients with vestibular symptoms [
10,
11].
Clinical trials investigating CGRP antibodies or CGRP receptor antagonists have shown statistically significantly efficacy for migraine treatment [
45]. Both of CGRP antibodies or CGRP receptor antagonists have limited permeability of blood-brain barrier [
59], indicating that trigeminal ganglion may be the target region for anti-CGRP therapy in migraine headache [
59]. Present study used lentiviral vectors comprising CGRP short hairpin RNA (LV-CGRP) to knockdown CGRP production in the trigeminal ganglion. The neuronal activation and CGRP expression in the TNC were decreased, meanwhile, the migraine-associated hyperalgesia and spontaneous facial pain were alleviated after CM, suggesting knockdown of CGRP in the trigeminal ganglion could efficiently block development of sensitization of the TNC and relieve migraine headache after CM. We also found that knockdown of CGRP in the trigeminal ganglion had similar effects on the VN, and restored vestibular dysfunction after CM. Currently, the relationship between CGRP and sensitization of VN neurons remains obscure. Indirect evidence shows that CGRP expression is significantly increased among VN neurons in a preclinical model of motion sickness, whilst anisodamine could decrease CGRP expression in the VN and relieve motion sickness [
17]. Collectively, these data pointed a role of CGRP in the VN to facilitate sensitization of VN neurons, and blocking CGRP in the trigeminal ganglion might relieve vestibular symptoms in migraine.
In this study, we focused on c-fos and CGRP expression changes in the TNC and VN, as well as the behavior changes after CM or CGRP knockdown. Previous studies showed that the indicators of central sensitization also contained electrophysiologic, like spontaneous and evoked activity of wide-dynamic-range neurons or response thresholds [
15]. Hence, whether CM could induce electrophysiological changes in the VN neurons remains to be determined. Moreover, immunohistochemistry study reported that CGRP could be detected in various brain regions, including cerebral cortex and thalamic nuclei [
54], thus, the effects of other brain regions on the regulation of neuronal activation and CGRP expression in the VN warrant further investigation.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.