Trigeminal nerve microstructure is linked with neuroinflammation and brainstem activity in migraine

Abstract

Although the pathophysiology of migraine involves a complex ensemble of peripheral and CNS changes that remain incompletely understood, the activation and sensitization of the trigeminovascular system are believed to play a major role. However, non-invasive, in vivo neuroimaging studies investigating the underlying neural mechanisms of trigeminal system abnormalities in human migraine patients are limited. Here, we studied 60 patients with migraine (55 females, mean ± standard deviation age: 36.28 ± 11.95 years) and 20 age- and sex-matched healthy controls (19 females, age: 35.45 ± 13.30 years) using ultra-high field 7 T diffusion tensor imaging and functional MRI, in addition to PET with the translocator protein ligand 11C-PBR28. We evaluated MRI diffusivity measures and the PET signal at the trigeminal nerve root, in addition to the brainstem functional MRI response to innocuous ophthalmic trigeminal nerve territory stimulation. Patients with migraine demonstrated altered white matter microstructure at the trigeminal nerve root (n = 53), including reduced fractional anisotropy, in comparison to healthy controls (n = 18). Furthermore, in patients, lower fractional anisotropy was accompanied by higher neuroinflammation (i.e. elevated 11C-PBR28 PET signal) at the nerve root (n = 36) and by lower functional MRI activation in an ipsilateral pontine cluster consistent with the spinal trigeminal nucleus (n = 51). These findings were more robust on the right side, which was consistent with the observation that right headache-dominant patients demonstrated higher migraine severity in comparison to left headache-dominant patients in our cohort. Multimodal imaging of the integrated neural mechanisms that characterize migraine underscores the importance of trigeminal system remodelling as both a key aspect of the dynamics underlying migraine pathophysiology and a target for therapeutic interventions.

Keywords: PET; diffusion tensor imaging; functional MRI; migraine; spinal trigeminal nucleus; trigeminal nerve.

Figure 1

White matter microstructural alterations in trigeminal nerve of migraine patients. (A) Representative example of a T₁-weighted MP-RAGE anatomical image in the axial view at the mid-pontine level of the brainstem. The blue box depicts the magnified area of the pons and trigeminal nerves with region-of-interest placement at the right and left trigeminal REZ (shown in red, four voxels in size). (B and C) Bar graphs illustrating average ± standard error of the mean DTI metrics (FA, RD, MD and AD) of migraine patients (red) and healthy controls (blue) for the right (B) and left (C) trigeminal REZ. Migraine patients exhibit significant microstructural alterations in the trigeminal REZ as characterized by lower FA (right and left side), higher RD (right only) and higher MD (right only) compared with healthy controls. *P < 0.05, **P < 0.01 and ****P < 0.0001 (false discovery rate corrected). AD = axial diffusivity; FA = fractional anisotropy; HC = healthy controls; MD = mean diffusivity; MIG = migraine patients; MP-RAGE = magnetization prepared-rapid gradient echo; RD = radial diffusivity.

Figure 2

Combining diffusion tensor imaging and PET: association between trigeminal nerve microstructure and neuroinflammation in migraine patients. (A) Representative example of a trigeminal REZ ¹¹C-PBR28 PET SUVR map and corresponding T₁-weighted image in the axial view in a migraine patient. (B) Scatterplots illustrating the relationship between ¹¹C-PBR28 PET signal and FA in the right and left trigeminal REZ, in addition to the difference (Δ; left minus right) in the migraine patient group only. Significant negative associations were found on the right side and the Δ. SUVR and FA values are adjusted for TSPO polymorphism. (C) Bar graphs illustrating average ± standard error of the mean PET signal of migraine patients with low FA (filled red circles), high FA (open red circles) and healthy controls (blue) for the right and left trigeminal REZ. The right side shows a significant group effect, driven by patients with low FA with elevated PET signal compared to healthy controls. **P < 0.01 (Dunnett’s post hoc test comparing each patient subgroup against a single healthy control group). FA = fractional anisotropy; HC = healthy controls; HFA = high fractional anisotropy; LFA = low fractional anisotropy; MIG = migraine patients; REZ = root entry zone; SUVR = standardized uptake value ratio.

Figure 3

Combining diffusion tensor imaging and functional MRI: association between trigeminal nerve microstructure and brainstem functional response to trigeminal sensory afference in migraine patients. (A) Schematic diagram of the positioning of the non-painful electrical forehead stimulation targeting the right trigeminal nerve territory and the block fMRI design used (stimulus comes on for 8 s and turns off for 14 s for a total of 17 stimulations). (B) Result from the voxelwise analysis showing the right SpV cluster, overlaid on a high-resolution 0.2 mm ex vivo B₀ brainstem, where fMRI response to trigeminal sensory stimulation on the right side demonstrates a significant positive associatiation with FA of the right trigeminal REZ in the migraine patient group. The Duvernoy’s brainstem atlas slice shown aided the anatomical localization of the SpV (pars oralis) fMRI response. FA = fractional anisotropy; fMRI = functional MRI; r = regression coefficient from general linear model analysis in FSL; REZ = root entry zone; SpV = spinal trigeminal nucleus.