Hypoxia facilitates neurogenic dural plasma protein extravasation in mice: a novel animal model for migraine pathophysiology

Abstract

Migraine animal models generally mimic the onset of attacks and acute treatment processes. A guinea pig model used the application of meta-chlorophenylpiperazine (mCPP) to trigger immediate dural plasma protein extravasation (PPE) mediated by 5-HT2B receptors. This model has predictive value for antimigraine drugs but cannot explain the delayed onset of efficacy of 5-HT2B receptor antagonists when clinically used for migraine prophylaxis. We found that mCPP failed to induce dural PPE in mice. Considering the role 5-HT2B receptors play in hypoxia-induced pulmonary vessel muscularization, we were encouraged to keep mice under hypoxic conditions and tested whether this treatment will render them susceptible to mCPP-induced dural PPE. Following four-week of hypoxia, PPE, associated with increased transendothelial transport, was induced by mCPP. The effect was blocked by sumatriptan. Chronic application of 5-HT2B receptor or nitric oxide synthase blockers during hypoxia prevented the development of susceptibility. Here we present a migraine model that distinguishes between a migraine-like state (hypoxic mice) and normal, normoxic mice and mimics processes that are related to chronic activation of 5-HT2B receptors under hypoxia. It seems striking, that chronic endogenous activation of 5-HT2B receptors is crucial for the sensitization since 5-HT2B receptor antagonists have strong, albeit delayed migraine prophylactic efficacy.

Figure 1. Hypoxia-induced effects in the dura mater of rodents.

(a) The 5-HT_2B/2C agonist mCPP induced PPE in the dura mater of the guinea pig. (b) mCPP dose-dependently induced PPE in the dura mater of hypoxic but not normoxic mice. (c) The more specific 5-HT_2B agonist BW 723C86 (BW) dose-dependently induced PPE in the dura mater of hypoxic but not normoxic mice. Data normalized to negative control (vehicle). All applications i.v. Mean ± SEM. Statistics: *p < 0.01, **p < 0.001; Rank Sum Test (a), multiple pairwise Rank Sum Test comparison with alpha-correction (b,c).

Figure 2. Acute blockage of the PPE induced with mCPP and BW 723C86 in hypoxic mice.

(a) The BW 723C86 (BW)-induced PPE in hypoxic mice was blocked with BF-1. (b) The mCPP-induced PPE in hypoxic mice was blocked with BF-1. (c) The mCPP-induced PPE in hypoxic mice was blocked with L-NAME. (d) The mCPP-induced PPE in hypoxic mice was blocked with sumatriptan (suma). Data normalized to negative control (vehicle). All applications i.v. Mean ± SEM. Statistics: *p < 0.05, One-Way ANOVA and Posthoc test (Bonferroni t-test).

Figure 3. Reversibility of the hypoxia-induced effect by subsequent normoxic treatment.

It was possible to induce dural PPE by intravenous application of mCPP 1 μg/kg after at least four weeks of hypoxia. This effect lasted for two weeks of subsequent normoxic treatment. Data normalized to negative control (vehicle). 1w H = one week hypoxia. 4w H + 1w N = four weeks hypoxia + one week normoxia. White: Vehicle, black: mCPP 1 μg/kg. All applications i.v. Mean ± SEM. n = 5 per group. Statistics: *p < 0.01, Multiple pairwise comparison with alpha-correction (Rank Sum Test).

Figure 4. Hypoxia-induced vascular remodelling occurred in the lung but not in dura mater.

After four weeks of hypoxia treatment, mice displayed a significantly increased number of arterial blood vessels with a diameter 25 μm ≤ Ø ≤ 50 μm in the lung (a). In dura mater, the length of muscularized arterioles with a diameter Ø ≥ 20 μm was equivalent between hypoxic and normoxic mice (b). Data normalized to negative control (normoxia). Mean ± SEM. Statistics: **p < 0.01, Rank Sum Test.

Figure 5. Quantification of HRP-positive endothelial vesicles in dural blood vessels.

The percentage of dural arterioles (A), capillaries (C) and venules (V) that showed HRP-positive vesicles in the endothelium is indicated for individual mice treated with vehicle + HRP (n = 4 normoxic and n = 2 hypoxic mice) and for mice treated with mCPP 1 μg/kg + HRP (n = 4 hypoxic mice). HRP 200 mg/kg. Data points of same colour/ pattern represent data for one individual mouse. All applications i.v.

Figure 6. Detailed depiction of HRP-positive and negative ultrastructures confirmed HRP extravasation via a transcellular pathway.

Higher magnifications of animals that received most extensive treatment (mCPP + HRP) and animals that received fewest treatment (vehicle + vehicle) confirmed that HRP-positive ultrastructures (such as HRP-positive endothelial vesicles (a, arrows) occurred exclusively in animals that received the dye but not in controls (b, arrows). It was further confirmed, that HRP did not penetrate via endothelial fenestrae (c, arrow) and that the tracer stopped at the level of interendothelial tight junctions (e, arrow). There was no damage in fenestrae structure (D, arrow) or tight junction integrity (f, arrow). Scale bars: a, b, e, f: 1 μm; c, d: 0.25 μm.

Figure 7. Chronic blockage of the 5-HT_2B receptor and NOS during hypoxia.

(a) The mCPP-induced PPE in hypoxic mice was blocked by daily BF-1 applications (20 mg/kg i.p.) during the sensitization phase. The application of BF-1 on the last three days of the hypoxic treatment was not sufficient to block the mCPP-induced PPE significantly. (b) The mCPP-induced PPE in hypoxic mice was blocked by daily L-NAME applications (20 mg/kg i.p.) during the sensitization phase. The application of L-NAME on the last three days of the hypoxic treatment was not sufficient to block the mCPP-induced PPE significantly. White: Vehicle, black: mCPP 1 μg/kg. n = 5 per group. Data normalized to negative control (vehicle + vehicle). Mean ± SEM. Statistics: *p < 0.05, **p < 0.01, n.s. - not significant, Multiple pairwise comparison with alpha-correction (Rank Sum Test).