Disordered autonomic function during exposure to moderate heat or exercise in a mouse model of Dravet syndrome.

Abstract

Objective:

To examine autonomic regulation of core body temperature, heart rate (HR), and breathing rate (BR) in response to moderately elevated ambient temperature or moderate physical exercise in a mouse model of Dravet syndrome (DS).

Methods:

We studied video-EEG, ECG, respiration, and temperature in mice with global heterozygous Scn1a knockout (KO) (DS mice), interneuron specific Scn1a KO, and wildtype (WT) mice during exposure to increased environmental temperature and moderate treadmill exercise.

Results:

Core body temperatures of WT and DS mice were similar during baseline. After 15 mins of heat exposure, the peak value was lower in DS than WT mice. In the following mins of heat exposure, the temperature slowly returned close to baseline level in WT, whereas it remained elevated in DS mice. KO of Scn1a in GABAergic neurons caused similar thermoregulatory deficits in mice. During exercise, the HR increase was less prominent in DS than WT mice. After exercise, the HR was significantly more suppressed in DS. The heart rate variability (HRV) was lower in DS than WT mice during baseline and higher in DS during exercise-recovery periods.

Significance:

We found novel abnormalities that expand the spectrum of interictal, ictal, and postictal autonomic dysregulation in DS mice. During mild heat stress, there was a significantly blunted correction of body temperature, and a less suppression of both HR and respiration rate in DS than WT mice. These effects were seen in mice with selective KO of Scn1A in GABAergic neurons. During exercise stress, there was diminished increase in HR, followed by an exaggerated HR suppression and HRV elevation during recovery in DS mice compared to controls.

These findings suggest that different environmental stressors can uncover distinct autonomic disturbances in DS mice. Interneurons play an important role in thermoregulation. Understanding the spectrum and mechanisms of autonomic disorders in DS may help develop more effective strategies to prevent seizures and SUDEP.

1. INTRODUCTION

Dravet syndrome (DS) is a rare epileptic encephalopathy that causes high morbidity and mortality. It typically presents in the first year of life with seizures induced by elevated body temperature, often in the context of fever, illness, vaccination, warm bath, warm ambient temperature, or exercise. Subsequently, fever-provoked and spontaneous myoclonic, tonic-clonic, absence, and focal seizures occur. Most children also suffer from comorbid conditions including developmental delays, intellectual disabilities, impaired motor control and coordination, autistic behaviors, and sleep disturbances. Up to 20% die before their 20^th birthday due to sudden death in epilepsy (SUDEP), status epilepticus, and other epilepsy- or treatment-related causes (Cooper et al., 2016; Dravet, 2011).

Heterozygous loss-of-function mutations in SCN1A, the gene that encodes the pore forming subunit of the voltage-gated sodium channel Na_V1.1, are the most common cause of DS, and occur in ~1/16,000 newborns(Claes et al., 2003; Fujiwara, 2006; Verbeek et al., 2015; Wu et al., 2015). Other genes associated with DS-like phenotypes include SCN2A, SCN8A, SCN9A, SCN1B, PCDH19, GABRA1, GABRG2, STXBP1, HCN1, CHD2, and KCNA2 (Steel et al., 2017). Most mutations arise de novo, although some are inherited from mosaic parents. Seizures in DS patients are rarely fully controlled although several new antiseizure medications (e.g., cannabidiol, fenfluramine, and stiripentol) are effective in reducing their frequency. Improving survival, seizure control, and quality of life for children and adults with DS remains a high priority (Devinsky et al., 2018; Myers et al., 2018b; Schoonjans et al., 2017; Wirrell, 2016; Zaccara and Schmidt, 2017).

Disruptions of the autonomic nervous system can occur as sequelae of chronic epilepsy or immediate consequences of focal or generalized seizures (Kalume, 2013; Lee and Devinsky, 2005; Schraeder and Lathers, 1989; Watanabe et al., 2001). Both clinical and pre-clinical studies reveal that seizure-induced autonomic dysfunctions are critical components of the physiological cascade leading to SUDEP (Devinsky et al., 2016; Kalume et al., 2013; Richerson et al., 2016; Ryvlin et al., 2013). Disordered interictal autonomic function can manifest as impaired temperature, cardiac, and respiratory regulations which are prominent comorbidities in DS patients (Delogu et al., 2011; Kim et al., 2018; Skluzacek et al., 2011; Wirrell et al., 2017). However, the full extent of these abnormalities, their associations with the DS-causing SCN1A mutations, and their underlying mechanisms are not fully understood.

The autonomic nervous system responds to environmental changes. In this study, we used our clinically relevant mouse model of DS, which recapitulates multiple clinical DS features (Catterall et al., 2010; Han et al., 2012a; Han et al., 2012b; Kalume et al., 2015; Kalume et al., 2013; Kalume et al., 2007; Oakley et al., 2009; Yu et al., 2006) to examine autonomic regulation of core body temperature, HR, and respiration rate in response to moderately elevated ambient temperature or moderate physical exercise. When exposed to a moderately elevated ambient temperature, DS mice fail to maintain temperature homeostasis. A similar response was observed in mice with Scn1a KO specifically in GABAergic neurons. In addition, DS mice showed an impaired heart and breathing rate response to this environmental challenge. Following a moderate treadmill exercise, DS mice exhibited a longer latency to peak cardiac and respiratory rate as well as an exaggerated decrease in HR after exercise termination. Our results reveal novel disturbances of autonomic function (unmasked by environmental stressors) which may contribute to the increased seizure and SUDEP susceptibility in DS mice.

2. MATERIALS AND METHODS

2.1. Ethics Statement.

This study was conducted in compliance with the Guide for the Care and Use of Laboratory Animals of the National Institute of Health. All animal procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of the Seattle Children’s Research Institute.

2.2. Animals.

All mice were kept in standard mouse cages, on a 12 h light/12 h dark cycle, and with ad-libitum (Kalume et al., 2015; Yu et al., 2006) access to food and water. Male and female mice were used in this study. Mice carrying the global heterozygous knock-out of Scn1a (Scn1a^+/− or DS mice), the genocopy of DS, were generated by targeted gene deletion(Kalume et al., 2015; Yu et al., 2006). These mice and littermate controls were obtained by breeding Scn1a^+/− males and Scn1a^+/+ females, both of C57BL/6J genetic background, and then genotyped.

To examine a potential link between interneuron specific KO of Scn1a and the thermoregulatory deficits of DS mice, mice carrying the interneuron specific Scn1a KO and their littermate controls were obtained by crossing the floxed Scn1a (Scn1a^flx/flx) (Cheah et al., 2012) with the Dlx5/6Cre mouse (Jackson Laboratories, Bar Harbor, ME). Both were maintained on the C57BL/6J background for more than 25 generations. Mice were genotyped using the following Cre primers and flx primers: Cre-1 (5’-GGTTTCCCGCAGAACCTGAA-3’), Cre-2 (5’-CCATCGCTCGACCAGTTTAGT-3’), flx-311 (5’-CTTGATGTGTTGAAATTCAC-3’), and flx-314 (5’-TATAGAGTGTTTAATCTCAAC-3’). Mutant and control mice used in this study had the following genotypes Scn1a^flx/+:Dlx5/6Cre⁺ and Scn1a^flx/+ :Dlx5/6Cre⁻ respectively.

2.3. Surgery.

Mice underwent survival surgery to implant EEG and ECG electrodes under isoflurane anesthesia with additional subcutaneous bupivacaine (1mg/kg) for analgesia as described previously (Kalume et al., 2015; Roy et al., 2015). Using aseptic technique, a midline incision was made anterior to posterior to expose the cranium. Micro-screw EEG electrodes, connected to fine (diameter: 0.130 mm bare; 0.180 mm coated) silver wires, were inserted through cranial burr holes drilled with a small drill bit. The EEG electrodes were placed at visually identified locations – left and right frontal cortices, approximately 1 mm anterior to the bregma and 3 mm lateral. ECG electrodes were tunneled subcutaneously from the head to the thoracic incision site and sutured in place. A reference electrode was placed at the midline cerebellum and a ground electrode was placed subcutaneously over the back and the skin was closed with sutures. All electrodes were connected to a micro-connector that permitted the interface with the recording system and had a typical impedance < 10 kΩ. The implant was fixed to the cranium using dental cement and the surgical incision was closed with sutures. Mice were allowed to recover for two to three days with carprofen MediGel placed in the animal cages for 72 hours.

2.4. Video-EEG-ECG recording.

Forty-eight to 72 hours after electrode implantation surgery, simultaneous video-EEG-ECG records were collected in conscious mice on a PowerLab 8/35 data acquisition unit using LabChart 8.0 software (AD Instruments, Colorado Spring, Co). All bioelectrical signals were acquired at 1-KHz sampling rate. The EEG signals were processed offline with a 1–70 Hz bandpass filter and the ECG signals with a 3-Hz highpass filter.

2.5. Mouse temperature recording and exposure to warmer ambient temperature.

Mice were implanted with a RET-4 rectal temperature probe (Physitemp Instruments Inc., NJ) which was secured in place with a lab tape against the tail and placed in an experiment chamber using protocols described in our previous work (Oakley et al., 2009). All mice were gently handled to optimally minimize stress. The experiment chamber temperature was controlled using a second temperature probe and a heat lamp attached to a temperature controller in a feedback loop (Physitemp Instruments Inc., NJ). Mouse temperature was continuously recorded using a temperature pod attached to a PowerLab 8/35 data acquisition unit and LabChart 8.0 software (AD Instruments, CO). The temperature signal was acquired at 100 Hz. Recordings of mouse temperature were collected during: 30 minutes at baseline chamber temperature (which corresponded to normal 25 ± 1 °C room temperature), 30 minutes when the chamber temperature was raised and maintained at 30 or 32 °C, and 10 minutes of recovery after the heat lamp was turned off. The levels of heat stress (30 or 32 °C) used in these experiments is within thermoneutral/thermal preferendum zone and was chosen to produce autonomic thermoregulatory response while avoiding behavioral thermoregulatory response (i.e, postural changes, or panting) and thermal seizure occurrence (Gordon, 2012; Hylander and Repasky, 2016). In a pilot study, we examined the response of 5 DS and 5 WT mice exposed to 28, 30, and 32 °C for 30 minutes and identified that at 30 °C all the mice exhibited a change in temperature and none of the DS mice experienced a thermal seizure. Their body core temperature never reached 38.5 °C, the average temperature for thermal seizure occurrence in these mice(Oakley et al., 2009). Exposure to 28 °C did not produce a temperature deflection in 40% of DS mice. In contrast, when exposed to 32 °C ambient temperature, all mice exhibited a spontaneous seizure during the heat challenge. This experimental condition increased the probability to induce a seizure and was adopted for the assessment of seizure influence on thermoregulation.

The line fit analysis module built in Igor pro 6.0 (Wavemetrics, OR) was used to measure the rate of temperature ascent between the baseline temperature at the onset of heat challenge and the peak temperature in the 15 min of heat exposure. The rate of temperature descent was deducted from the linear fit between the peak temperature and the temperature after autonomic adjustment or the temperature at the end of heat challenge if no homeostatic correction was observed. The rate of post-challenge temperature change was deducted from the linear fit between the temperature at end of heat challenge and at 10 min post-challenge.

2.6. Whole-body plethysmography.

Respiratory function was assessed by whole-body plethysmography under unrestrained conditions using a modified chamber that allowed the introduction of EEG cables through the lid for concurrent recording of EEG, ECG and ventilatory activity.

2.7. Treadmill exercise.

A treadmill apparatus for mice (Bioseb, Pinnellas Park, FL) was used for physical exercise of mice. Exercise consisted of 5 min treadmill running at 25 cm/s without incline. To motivate the mice to run the experimenter gently pushed them with a small plastic stick if needed. If a mouse still refused to run it could be motivated by a transient and light electric stimulation from the grid at the beginning of the treadmill platform. Mice that did not complete the task were excluded from the study. We assessed the HR, change in HR from resting to the peak exercise levels, and recovery from the peak exercise level to the level at 5 minutes post exercise.

2.8. ECG and plethysmography data analysis.

HR and HRV calculations were conducted using preset ECG and HRV analysis modules in the Labchart 8 software (AD Instruments, CO). The same analysis protocol was employed for control and mutant groups. Three-minute segments of recording acquired when mice were stationary during the baseline and recovery periods were used. For heat challenge experiments, baseline and recovery data segments were selected 5 minutes before and after heat exposure, respectively. For exercise challenge experiments, data segments were selected one minute before and after exercise. Ectopic activity, sinus pause, atrial and ventricular arrhythmia, and signal artifacts were identified using the beat classifier function of the analysis modules, confirmed by visual inspection, and excluded from the data if present. R waves were detected, and RR intervals and HR were deducted. HRV was assessed using the root mean square of the differences between adjacent RR intervals (RMMSD, in ms), a time-domaine index of short-term variability.

Breathing rate was calculated using the peak analysis module of the Labchart 8 (AD Instruments, CO). Five-minute segments of continuous recording acquired during baseline and recovery periods were used to deduct BR.

2.9. Statistical Analysis.

Data are reported as means ± SD. Comparisons within groups were done by paired two-tailed t-tests and across groups by unpaired two-tailed t-tests. Differences were considered significant at P< 0.05.

3. RESULTS

3.1. Impaired thermoregulation in DS mice.

We examined the ability of DS mice (N=8, 5 females, 3 males) to maintain temperature homeostasis when exposed to a slightly elevated environmental temperature in comparison with WT mice (N=7, 4 females, 3 males) at an advanced stage of the disease, post-natal day (P) 30–35. The core body temperatures of WT (36.8 ± 1.0 °C) and mutant mice (36.5 ± 1.0 °C) were similar during baseline conditions (p>0.05; ambient temperature 25°C). In the first 15 mins of exposure to warmer ambient temperature, two out of eight DS mice displayed seizures which lead to core body temperature instability. They were excluded from this analysis and were separately studied for the influence of seizure activity on thermoregulation as described below. In the remaining DS mice (N=6, 5 females, 1 male) and the WT mice (N=7, 4 females, 3 males), a smaller temperature increase was registered in DS than WT mice (DS: 1.6 ± 0.3 vs. WT: 2.5 ± 0.5 °C, p< 0.01, Figure 1 A, B1). The rate of temperature ascent (calculated by linear fit) was lower in DS than WT control mice (DS: 0.3 ± 0.05 vs. WT: 0.5 ± 0.1 °C/min, p<0.01, Figure 1 A, B3). After 15 min of heat exposure, the initial response was followed by a slow return of temperature to baseline in WT mice, but not in DS mice. The rate of temperature descent from peak was blunted in DS compared to WT mice (DS: −0.001 ± 0.02 vs. WT: −0.3 ± 0.05 °C/min, p<0.001, Figure 1 A, B4). This led to a greater temperature deviation from baseline in DS than in WT mice (DS: 1.45 ± 0.4 vs. WT: −0.5 ± 0.5 °C, p<0.001, Figure 1 A, B2) at the end of the 30 minutes of heat exposure. It was only after the heat lamp was turned off that the temperature of DS mice slowly receded toward the baseline level. A higher rate of temperature correction was observed in DS than WT mice during this period (DS rate: −0.085 ± 0.01 vs. WT rate: −0.03 ± 0.01 °C/min, p<0.001, Figure 1 A, B5). Unlike WT mice which returned to baseline temperature during heat exposure, DS mice did not reach this temperature range until the after the end (17:39 ± 2:05 (min:sec)) of the challenge. Together, these findings reveal that Scn1a KO mice have impaired temperature regulation in a moderately warmer environment.

Figure 1. Impaired temperature regulation in the global and interneuron-specific Scn1a knockout mice exposed to moderate elevation of ambient temperature.

A. (Top) Representative temperature changes of the experimental chamber during ambient heat challenge. The chamber temperature was elevated from baseline (25 °C) and sustained at 30 °C for 30 minutes using a heat lamp. After this period, the heat lamp was turned off. (Middle) Changes of core body temperature showing a peak value within the first 15 min of heat exposure of global Scn1a KO (DS) and WT mice, as well as failure of core temperature return to baseline in the DS mice in the last 15 min of het exposure. (Bottom) Changes of core body temperature showing that the core body temperature of interneuron-specific Scn1a KO mice showed similar pattern of change compared to the control mice as the DS mice. B1 and B2. Changes in temperature at peak time and after 30 min of heat exposure in DS and WT mice. B3, B4, B5. Slopes of temperature ascent to peak, descent from peak, and recovery post heat challenge for DS and WT mice. C1 and C2. Changes in temperature at peak time and after 30 min of heat exposure for interneuron KO and control mice. C3, C4, C5. Slopes of temperature ascent to peak, descent from peak, and recovery post heat-challenge for interneuron KO and control mice. Abbreviations: DS: Dravet syndrome, KO: knock out, WT: wild type.

3.2. KO of Scn1a in GABAergic interneurons impairs thermoregulation

Selective KO of Scn1a in GABAergic neurons, while sparing the excitatory neurons, is sufficient to reproduce seizures and several comorbid conditions of DS in mice(Cheah et al., 2012; Kalume et al., 2013; Kalume et al., 2007). We investigated whether such conditional KO of Scn1a is also sufficient to cause DS-like thermoregulatory impairment, using Dlx56 Scn1a KO mice in which the mutation is restricted to GABAergic interneurons (Methods). Conditional KO mice (Scn1a^flx/+:Dlx5/6Cre⁺, N=8, 3 females, 5 males) and littermate control mice (Scn1a^flx/+:Dlx5/6Cre⁻, N=8, 4 females, 4 males) were used for this experiment at P20–25, an age range of advanced disease stage in this line. The Dlx56 specific KO mice exhibited a more severe phenotype than the global KO counterparts; they had more frequent seizures and all died by P30. During the first 15 minutes of exposure to warmer environmental temperature, the body core temperature of conditional KO and control mice increased. The maximal response was similar in Dlx56 specific KO and control mice (Dlx56-Cre KO: 1.9 ± 0.6 vs Control: 1.2 ± 0.3 °C, p>0.05, Figure 1 A, C1), but the rate of temperature change was reduced in conditional KO compared to control mice (Dlx56-Cre KO: 0.22 ± 0.07 vs Control: 0.4 ± 0.1 °C/min, p<0.01, Figure 1 A, C3). During the subsequent 15 minutes of exposure to warmer ambient temperature, the core body temperature of the conditional KO mice did not return to baseline level, like in control mice. The change in body core temperature (at 30 minutes), relative to baseline level, was greater in conditional KO mice than in control mice (Dlx56-Cre KO: 1.4 ± 0.5 vs. Control: 0.18 ± 0.3 °C, p<0.001, Figure 1 A, C2). Thus, the rate of temperature descent from peak value was significantly reduced in mutant relative to control mice (Dlx56-Cre KO: −0.03 ± 0.001 °C/min vs Control: −0.07 ± 0.002, p<0.01, Figure 1 A, C4). After the heat lamp was turned off, the temperature of conditional KO mice slowly deviated toward baseline level and the temperature of control mice, which was already at baseline, adjusted slightly. However, the rates of these two changes were not statistically different (Dlx56-Cre KO: −0.13 ± 0.04 vs. Control: −0.1 ± 0.02 °C/min, p>0.05, Figure 1 A, C5). Conditional KO mice returned to baseline temperature at 14:36 ± 4:36 minutes: seconds after the heat stress. Together, these findings indicate that targeted KO of Scn1a in GABAergic neurons causes similar thermoregulatory deficits as global KO of the same gene in mice.

3.3. Seizure events disrupt temperature stability in DS mice.

Although, most DS mice used in the experiments described above exhibited no seizures during exposure to warmer (30 °C) temperature, a small portion (2/8, 2 males) exhibited mild myoclonic (MC) or generalized tonic-clonic (GTC) seizures. To examine the impact of ictal activity on thermoregulation, peri-ictal changes in body core temperature were evaluated. MC seizures caused a smaller drop in temperature (−0.25 ± 0.25 °C, N of MC=5) than GTC seizures (−2.16 ± 0.25 °C, N of seizures=3, p<0.001, Figure 2 A, B). The rate of temperature decay was greater for MC than did GTC seizures (0.35 ± 0.15 vs 0.1 ± 0.01 °C/min, respectively, p=0.001, Figure 2 B). Given the smaller sample size in this analysis, further examination of ictal activity’s influence on thermoregulation was conducted on a separate cohort of DS mice (N=9, 6 females, 3 males) exposed to a slightly higher (32 °C) temperature to promote seizure occurrence in all tested mice (Methods). Similar to findings in the previous two mice, the magnitude of temperature decline was smaller when induced by MC (−0.3 ± 0.2 °C, N of MC=18) than GTC seizures (−1.8 ± 0.27 °C, N of seizures=9, p< 0.001). But the rate of temperature change was greater for MC than GTC seizures (0.5 ± 0.18 vs 0.3 ± 0.01 °C/min, respectively, p< 0.001, Figure 2 C) and there was a poor correlation between temperature drop and seizure duration. Together, these findings demonstrate that seizures can directly cause a temperature correction toward normal level in DS.

Figure 2. Seizure induced alterations of core body temperature in DS mice.

A. Representative trace of instantaneous core body temperature of a DS mouse which experienced a myoclonic seizure and a generalized tonic clonic seizure at time points marked by vertical lines during exposure to moderate ambient temperature showing a larger change temperature decline associated with the GTC than the MC seizure. B. (Left) Changes in core body temperature following a MC and GTC seizure and (Right) Slope of temperature change associated with MC and GTC seizure. When exposed to 30 °C ambient temperature. Only 2 out of 8 had seizure events in this experimental condition. C. (Left) Change in core body temperature following a MC and GTC seizure and (Right) Slope of temperature change associated with MC and GTC seizure when exposed to a slightly higher (32 °C) ambient temperature to promote seizure occurrence in most mice. All (9/9) mice displayed seizure events in this experimental condition. Abbreviations: DS: Dravet syndrome, GTC: Generalized tonic clonic, MC: Myoclonic. D. Relationship between seizure duration and seizure-related body core temperature decline showing a poor correlation between the two measurements for mice exposed to 32 °C ambient temperature. Abbreviations: DS: Dravet syndrome, GTC: generalized tonic clonic, MC: Myoclonic.

3.4. Impaired cardiac and respiratory autonomic response to moderate elevation in ambient temperature in DS mice.

We assessed changes in resting HR when DS mice were exposed to moderately warmer ambient temperature. A different cohort of DS (N=8, 6 females, 2 males) and WT (N=7, 3 females, 4 males) mice was implanted with ECG electrodes and studied. We observed a significantly less prominent suppression of HR in DS mice (−85 ± 24 BPM) than in WT control littermates (−140 ± 19 BPM, p<0.01, Figure 3 A, B) during the heat stress. In addition, a blunted recovery of the HR was observed in DS mice (2.3 ± 9 BPM) compared to WT controls (54.2 ± 7.0 BPM, p<0.001, Figure 3 B) after removal of the heat stressor. These findings reveal a novel interictal cardiac autonomic impairment caused by a moderate increase in ambient temperature in DS mice.

Figure 3. Changes in resting heart rate and breathing rate associated with moderate heat exposure.

A. Representative traces of heart rate changes before, during, and after heat exposure showing a less pronounced suppression of HR in DS than WT mice. B. Changes in HR during heat exposure and recovery periods for DS and WT mice. C. Representative traces of breathing rate changes before, during, and after heat exposure showing the suppression of BR in DS than WT mice. D. Changes in BR during heat exposure and recovery periods for DS and WT mice. Abbreviations: DS: Dravet syndrome, WT: Wild type, ECG: electrocardiogram, HR: heart rate.

To examine the integrity of respiratory function in DS, we evaluated changes in respiratory rate during exposure to moderately warmer ambient temperature. BR decreased in both DS and WT control littermate mice during exposure to warmer environmental temperature. However, this BR change was slightly smaller in DS (−36 ± 8.0 BPM) than WT (−46.6 ± 13.6 BPM, p<0.01, Figure 3 C, D) mice. In addition, when the environmental heat stressor was removed, WT mice showed a partial recovery response (21 ± 5.6 BPM) whereas DS mice had no significant recovery (−5.3 ± 5.3 BPM, p<0.001, Figure 3 C, D). These findings illustrate a novel aspect of interictal impairment of breathing in DS mice.

3.5. Impaired cardiac autonomic response to exercise in DS mice.

Another separate cohort of DS (N=9, 6 females, 3 males) and WT (N=6, 3 females, 3 males) mice implanted with ECG electrodes was used to assess the integrity of the cardiac autonomic response to exercise in DS mice. During exercise on a treadmill, the HR increase was less prominent in DS mice (38 ± 5.0 BPM) than in WT mice (65.8 ± 10 BPM, p<0.001, Figure 4 A, B). The latency to peak HR was greater in DS (3.8 ± 0.6 BMP) than WT (2.2 ± 0.3 BPM, p<0.001, Figure 4 C) mice. After exercise, the HR was significantly more suppressed in DS (HR change was −147.4 ± 6.7 BPM in DS compared to −72.6 ± 11.8 BPM in WT mice) (p<0.001, Figure 4F). Baseline heart rate variability, estimated by RMSSD index, was lower (1.3 ± 0.2 ms) in DS than in WT mice (8.6 ± 0.5 ms, p<0.05, Figure 4D). During recovery, the RMSSD values were higher (12.3 ± 3.0 ms) in DS than in WT (8.6 ± 0.4 ms, p<0.001, Figure 4E). These findings suggest that DS mice have 1) a delayed parasympathetic withdrawal during exercise, 2) an excessive vagal response during recovery, and 3) abnormal cardiac autonomic responses to exercise resulting from the Scn1a mutation.

Figure 4. Impaired changes in heart rate during and after exercise in DS mice.

A. Representative traces of heart rate before, during, and after exercise for DS and WT mice. Note the late peak HR in DS than in WT mice. Inserts a1-a3 illustrating examples of ECG traces recorded during periods marked in A. B-D. Changes in HR, Time to peak HR, and Heart rate variability expressed in Racine mean square for DS and WT mice during exercise and baseline. E and F. Change in HRV and HR for DS and WT mice during recovery from exercise. Abbreviations: DS: Dravet syndrome, WT: Wild type, ECG: Electrocardiogram, HR: Heart rate.

4. DISCUSSION

Autonomic system disorders occur in many neurological diseases and can cause severe disabilities, impair quality of life, or increase morbidity and mortality in affected individuals. We identified several novel abnormalities in the autonomic regulation of body temperature, HR, and BR in DS mice exposed to moderate heat stress or physical exercise. Similar thermoregulatory abnormalities were observed in both global and GABAergic interneuron specific Scn1a KO mice, suggesting that the Scn1a deficit in GABAergic neurons is sufficient to cause these impairments. When the environmental temperature was slightly elevated, the homeostatic response of DS mice, unlike wildtype littermates, failed to return the body temperature to normal, showing that heterozygous Scn1a mutation KO impairs thermal regulation. DS mice had a larger decrease in body temperature immediately after a GTC than an MC seizure. Furthermore, we found that Scn1a KO mice have lower baseline HRV, delayed parasympathetic withdrawal during exercise, and excessive vagal response during recovery. Since SUDEP occurs after GTC seizure, marked by severe hypermotor activity, and postictal autonomic dysregulations contribute to SUDEP pathogenesis in DS mice, the post-exercise parasympathetic hyperactivity that we discovered may be an important contributor to the SUDEP mechanisms in these mice.

Interictal autonomic dysfunction occurs in DS patients (Delogu et al., 2011; Kim et al., 2018; Skluzacek et al., 2011; Wirrell et al., 2017) and DS mice (Kalume et al., 2013). Core body temperature regulation is essential for normal functioning of mammalian cells, tissues, and organ systems. Our system-level studies revealed that DS mice have impaired thermoregulatory capacity and GABAergic interneurons likely mediate this physiological abnormality. We previously demonstrated that epilepsy as well as ataxia, circadian rhythm defects, cognitive impairment, autistic-like behaviors, sleep disturbances, and premature death result from impaired Scn1a action potential firing in the GABAergic interneurons of different brain regions in DS mice (Cheah et al., 2012; Han et al., 2012b; Kalume et al., 2015; Kalume et al., 2013; Yu et al., 2006). Therefore, impaired thermoregulation, uncovered in this study, likely results from hypoexcitability of hypothalamic (arcuate, medial preoptic, and anterior nuclei) GABAergic neural networks that control core body temperature (Morrison, 2016a; Morrison, 2016b; Osaka, 2012; Zhao et al., 2017). These neurons can be warm-sensitive, cold-sensitive, or temperature insensitive (Morrison, 2016b).

Although our thermal challenge protocol was designed to maintain the mouse core body temperature below the thermal seizure threshold range (Oakley et al., 2011; Oakley et al., 2009), a few mice developed MC and GTC seizures after several minutes of moderate hyperthermia, indicating that both intensity and duration of temperature increase can affect seizure threshold. The GTC seizures were immediately followed by a larger decrease in body temperature than the MC seizures. These findings indicate that seizure activity may propagate to hypothalamic thermoregulatory nuclei and influence their function. It appears that seizures help reset the temperature to normal level during heat challenge. Therefore, while seizure impact on temperature regulation is unlikely to contribute to SUDEP, related aberrant postictal dysregulation of other autonomic functions such as respiration and sympathetic:parasympathetic balance may be more critical.

Body warming by environmental temperature influences cardiac and respiratory systems (Liu et al., 2015; Rubini and Bosco, 2013; Rubini et al., 2012; Wilson and Crandall, 2011). Prior studies have reported abnormal interictal or peri-ictal cardiorespiratory functions at normal ambient temperature in DS patients and mouse models. Decreased resting-state HRV is the most common interictal cardiac autonomic dysfunction observed; it was identified in comparisons of DS patients with either control subjects or other epilepsy patients (Delogu et al., 2011; Ergul et al., 2013; Lyu et al., 2019; Myers et al., 2018a; Shmuely et al., 2020). Other cardiac autonomic abnormalities in DS patients include interictal QT and P wave dispersion, albeit these were observed in most (Ergul et al., 2013; Lyu et al., 2019) but not all studies (Delogu et al., 2011). In animal models, reduced resting-state interictal HRV was observed in a study using the Scn1a KO mice whereas opposite finding (increased HRV) was noted in another study conducted on mice carrying a Scn1a R1407X allele (Kalume et al., 2013; Kim et al., 2018). Additionally, irregular breathing patterns were also observed in mouse models carrying the Scn1a R1407X or A1783V (Kim et al., 2018; Kuo et al., 2019). Peri-ictal cardiac and respiratory abnormalities identified in cohorts of DS patients include QTc lengthening, severe breathing irregularities, and blunted ventilatory response to CO2 (Kim et al., 2018; Shmuely et al., 2020). In mouse models, postictal bradycardia was noted in initial studies that did not simultaneously assess ventilatory function. Recently, severe apnea either preceding or concurrent with bradycardia was found when both cardiac and respiratory functions were monitored (Bard et al., 2016; Kim et al., 2018). It is worth noting that though common in mouse models, postictal bradycardia were seen less frequently in DS than in a non-DS epilepsy patient cohort in the Shmuely and colleagues’ study (Shmuely et al., 2020). The reason for this unexpected finding is unknown. The reduced responses of HR, HRV, and ventilatory rate to slight elevations in ambient temperature observed in this study indicate that moderate heat stress aggravates the interictal cardiorespiratory autonomic dysfunctions. We postulate that such dysfunctions will cause a compromising physiological condition of the subject and increase the vulnerability to SUDEP in the advent of a subsequent seizure. This prediction is in line with the recently proposed autonomic “second hit hypothesis” for SUDEP mechanism which proposes that a seizure can trigger a fatal autonomic deterioration in people with epilepsy and already severe autonomic dysfunctions (Myers et al., 2018a). However, further studies are warranted to confirm this hypothesis.

Autonomic adaptation during and after exercise is mediated by coordinated changes in sympathetic and parasympathetic activity: marked parasympathetic withdrawal early during exercise, followed by sympathetic activation as exercise duration and intensity increases, and a parasympathetic re-engagement immediately after exercise terminates (Javorka et al., 2002). Indices of autonomic activity during and after clinical exercise stress testing are used in diagnosing and prognosticating several medical conditions. For instance, HR recovery (the decline in heart rate toward baseline level after cessation of exercise) is a strong and independent predictor of cardiovascular disease and all-cause mortality (Cole et al., 1999; Cole et al., 2000; Nishime et al., 2000; Qiu et al., 2017; Watanabe et al., 2001). Attenuated HR recovery strongly correlates with reduced vagal reactivation post-exercise and high risk for adverse health outcomes (Coote, 2010; Gourine and Ackland, 2019; Imai et al., 1994). Exercise testing-induced ventricular tachycardia may also predict cardiovascular morbidity and mortality, but this may be limited to cases associated with structural heart disease (Jouven et al., 2000; Marine et al., 2013; Yang et al., 1991).

Our DS mice exhibited diminished exercise-induced tachycardia and exaggerated HR reduction during exercise recovery. These observations suggest a delayed parasympathetic withdrawal during exercise and excessive vagal tone during recovery. However, these autonomic responses to exercise run counter to the traditional markers of death risk. Since motor activity-related abnormalities can occur during GTC seizures in DS mice, we hypothesize that excessive postictal suppression of HR can facilitate SUDEP while a diminished ictal tachycardia will not. This prediction is consistent with our and others’ finding that SUDEP follows seizure-induced severe bradycardia in DS mice (Kalume et al., 2013; Kim et al., 2018). Future studies will directly test the hypothesis that HR overcorrection after exercise may predict SUDEP risk in DS mice.

Cardiac rhythm abnormalities may reflect SCN1A defects in brainstem autonomic or higher autonomic regulatory centers. However, isolated ventricular myocytes from DS mice exhibit increased transient and persistent sodium current and ECG studies reveal QT prolongation, ventricular ectopic foci, idioventricular rhythms, beat-to-beat variability, ventricular fibrillation, and focal bradycardia (Auerbach et al., 2013). Thus, autonomic and cardiac dysfunctions in DS may result from abnormalities in central or peripheral nervous system or cardiac effects of SCN1A mutations.

One limitation of this study is that mice carrying a floxed-Scn1a allele in the absence of a cre allele (Scn1a^flx/+:Dlx5/6Cre⁻ mice) constituted the only control group in the assessments of thermoregulatory function in the conditional KOs. Studies have shown that cre expression in the absence of a floxed allele can cause an unexpected phenotype in mice (Janbandhu et al., 2014). This effect is attributed to DNA damage caused by cre-recombinase activity at sites that resemble LoxP in the genome. It has been documented in cells such as neurons, gastric cells, spermacides, and cardiomyocytes of different cre-driver lines and could serve as an experimental confounder (Harno et al., 2013; Huh et al., 2010; Pugach et al., 2015; Rehmani et al., 2019; Schmidt et al., 2000). The possibility that cre alone (independent of the floxed allele) contributes to the development of the thermoregulatory phenotype in the conditional KO (Scn1a^flx/+:Dlx5/6Cre⁺) mice cannot be ruled out in our study. Including, a WT control group and a Dlx5/6Cre control group to our experiments could have helped elucidate this point. However, the similarity of thermoregulatory defect in the global and conditional KOs suggests that the phenotype may be more dependent on the KO of Scn1a than an independent cre activity. Future studies on whether cre alone causes a phenotype in the Dlx5/6Cre driver line will certainly shed more light on this issue.

Another limitation of this study is the lack of assessments for cardiac electric conductions, including QTc and P wave dispersions. As discussed above, an increased dispersion of these parameters is strongly linked to elevated risk of fatal ventricular and atrial dysrhythmia (Algra et al., 1991; Dogan et al., 2008; Li et al., 2002). These abnormalities have been found in patients with DS syndrome and are believed to contribute to the elevated risk for SUDEP in this epilepsy population (Lyu et al., 2019; Shmuely et al., 2020). Studies of the impact of environmental challenges on the integrity of cardiac conduction in our mouse model of DS will provide further insight into the contribution of environmental factors in the mechanisms of SUDEP.

5. CONCLUSIONS

Our findings extend the spectrum of ictal, interictal and postictal disorders of autonomic homeostasis in DS. Sympathetic and parasympathetic imbalances can manifest differentially with sympathetic hyperactivity during interictal and ictal periods and parasympathetic hyperactivity postictally. Understanding the spectrum and mechanisms of autonomic disorders in DS, as well as the role of age, temperature, seizure features and other factors that modify phenotypic expression of autonomic dysfunction may help develop more effective strategies to prevent seizures and SUDEP. Future studies will investigate the cellular and network basis of the reported autonomic derangements in DS as well as their significance in potentially predicting SUDEP.

KEY POINTS.

• Autonomic regulation of body temperature, HR, and BR was studied in Dravet syndrome mice exposed to moderate heat or exercise stress.

• When exposed to moderate heat stress, these mice failed to maintain temperature homeostasis.

• During exercise, they exhibited a delayed parasympathetic withdrawal on electrocardiogram.

• During recovery from exercise, they exhibited an excessive vagal response.

• We uncovered novel abnormalities of autonomic regulation in these mice that may help develop more effective approaches against seizures or SUDEP in Dravet syndrome.