Caffeine exacerbates seizure-induced death via postictal hypoxia

Abstract

Sudden unexpected death in epilepsy (SUDEP) is the leading epilepsy-related cause of premature mortality in people with intractable epilepsy, who are 27 times more likely to die than the general population. Impairment of the central control of breathing following a seizure has been identified as a putative cause of death, but the mechanisms underlying this seizure-induced breathing failure are largely unknown. Our laboratory has advanced a vascular theory of postictal behavioural dysfunction, including SUDEP. We have recently reported that seizure-induced death occurs after seizures invade brainstem breathing centres which then leads to local hypoxia causing breathing failure and death. Here we investigated the effects of caffeine and two adenosine receptors in two models of seizure-induced death. We recorded local oxygen levels in brainstem breathing centres as well as time to cessation of breathing and cardiac activity relative to seizure activity. The administration of the non-selective A₁/A_2A antagonist caffeine or the selective A₁ agonist N6-cyclopentyladenosine reveals a detrimental effect on postictal hypoxia, providing support for caffeine modulating cerebral vasculature leading to brainstem hypoxia and cessation of breathing. Conversely, A_2A activation with CGS-21680 was found to increase the lifespan of mice in both our models of seizure-induced death.

Figure 1

Acute experimental timeline. Male C57BL/6J mice on P50 underwent brain implantation surgery and diaphragmatic surgery (Day 0). Mice recovered for one week before beginning terminal experiment (Day 7). Mice were injected with either caffeine, theophylline, N6-Cyclopentyladenosine (N6) or CGS- 21,680, 30 min prior to infusion of kainic acid. Physiological parameters including heart rate, respiratory rate, brain partial pressure of oxygen and local field potential were recorded until death.

Figure 2

Drug targets of adenosine A₁ and A_2A receptor subtype. (A) Model schematic: following the injection of vehicle or drug, cardiorespiratory (HR: heart rate/ RR respiratory rate), oxygen and LFP signals were measured. electromyograph (EMG) electrode wires record respiratory activity directly from diaphragm and electrodes record cardiac activity. A bipolar electrode in the brainstem records local field potential (LFP) in the pre-Bötzinger complex (PBC) and an oxygen sensing optode in the contralateral PBC records absolute oxygen. Mean time to death in male C57BL6/J mice treated with drugs targeting adenosine receptor subtypes. (B) Vehicle mice (n = 11) mean time to death was 50.44 ± 3.95, Caffeine mice (n = 11) mean time to death was 27.21 ± 2.65 (t₂₀ = 4.87, p =  < 0.0001****). (C) Vehicle mice (n = 10) mean time to death was 51.80 ± 3.17, Theophylline (n = 10) mean time to death was 30.32 ± 3.67 (t₁₈ = 4.42, p = 0.0003***). (D) Vehicle mice (n = 10) mean time to death was 49.12 ± 2.44, N6 (n = 10) mean time to death was 35.95 ± 4.27 (t₁₈ = 2.67, p = 0.0154*). (E) Vehicle mice (n = 10) mean time to death was 19.12 ± 2.44, CGS-21680 (n = 10) 114.0 ± 5.72 (t₁₈ = 10.42, p =  < 0.0001****). Histobars represent mean ± SEM.

Figure 3

Caffeine reduced baseline brainstem oxygen. Mean partial pressure of oxygen (mmHg) pre-injection of drug or vehicle baseline (T-1); 10-min post-injection of drug or vehicle baseline (T-2); Mean 30-min post-injection of drug or vehicle baseline (T-3). (A,B) Hippocampus (HPC) and pre-Bötzinger complex (PBC) oxygen differences were not statistically significant in vehicle mice (n = 9). (C) HPC caffeine: (n = 9), T-1 = 23.51 ± 1.93, T-2 = 19.46 ± 1.72, T-3 = 15.95 ± 1.62, (ANOVA F (2, 24) = 4.567, p = 0.0209). (D) PBC caffeine: T-1 = 50.64 ± 2.72, T-2 = 36.93 ± 3.02, T-3 = 32.85 ± 1.43, (ANOVA F (2, 24) = 13.90, p =  < 0.0001***). (E) HPC CGS-21680: (n = 9), T-1 = 25.68 ± 1.80, T-2 = 27.67 ± 2.84, T-3 = 34.96 ± 2.31, (ANOVA F (2, 24) = 4.281, p = 0.026). PBS CGS-21680: T-1 = 45.03 ± 2.62, T-2 = 39.18 ± 2.98, T-3 = 56.04 ± 2.02, (ANOVA F (2,24) = 11.09, p = 0.0004***). Histobars represent mean ± SEM.

Figure 4

Acute caffeine administration decreases time to terminal apnea hypoxia and death. Representative profile from three C57BL/6J + kainic acid mice. (A,Ai) The mean time to terminal hypoxia of the hippocampus (HPC) was 51.93 ± 2.47 in vehicle (n = 11), 26.22 ± 2.58 in caffeine (n = 11) and 94.45 ± 10.11 in CGS-21680 (n = 11) mice ANOVA F (2,30) = 30.98, p =  < 0001***. (B,Bi) Mean time to terminal hypoxia of the pre-Bötzinger complex (PBC) was 52.32 ± 2.49 in vehicle, 26.37 ± 2.70 in caffeine and 101.2 ± 10.10 in CGS-21680 mice ANOVA F (2,30) = 37.46, p =  < 0.001****. (C,Ci) Mean time to terminal apnea in was 52.80 ± 2.52 in vehicle, 27.21 ± 2.65 in caffeine and 103.4 ± 10.01 in CGS-21680 mice ANOVA F (2,30) = 39.67, p =  < 0.0001****. (D,Di) Mean time to terminal asystole was 58.70 ± 2.69 in vehicle, 32.91 ± 2.61 in caffeine and 109.6 ± 9.92 in CGS-21680 mice ANOVA F (2,30) = 40.49, p =  < 0.0001****. Box plots represent mean ± SEM.

Figure 5

Chronic caffeine accelerates time to death and chronic CGS-21680 delays death. (A,B) (A) Kcna1^−/− mice treated with caffeine (n = 9) lived significantly shorter compared to vehicle (n = 9) treated mice. Median survival day for vehicle mice was 42d compared to median survival of caffeine treated of 37d p = 0.0015** (Log-rank test) (Ai). Mean survival day for vehicle treated mice was 41.89 ± 0.99 compared to mean survival of caffeine treated mice 37 ± 0.72, t₁₆ = 3.97, p = 0.0011**. Chronic CGS-21680 reduces premature mortality and extends life, while chronic caffeine increases premature mortality. (B) Kcna1^−/− mice treated with CGS-21680 (n = 9) lived significantly longer compared to vehicle (n = 9) treated mice. Median survival day for vehicle mice was 44d compared to median survival of CGS-21680 of 51d ****p =  < 0.0001 (Log-rank test), (Bi) Mean survival day for vehicle treated mice was 44 ± 0.52 compared to mean survival of CGS-21680 treated mice 51.78 ± 1.05, t₁₆ = 6.61, p =  < 0.0001****. Histobars represent mean ± SEM.