The Impact of Inflammation on Thermal Hyperpnea: Relevance for Heat Stress and Febrile Seizures

Abstract

Extreme heat caused by climate change is increasing the transmission of infectious diseases, resulting in a sharp rise in heat-related illness and mortality. Understanding the mechanistic link between heat, inflammation, and disease is thus important for public health. Thermal hyperpnea, and consequent respiratory alkalosis, is crucial in febrile seizures and convulsions induced by heat stress in humans. Here, we address what causes thermal hyperpnea in neonates and how it is affected by inflammation. Transient receptor potential cation channel subfamily V member 1 (TRPV1), a heat-activated channel, is sensitized by inflammation and modulates breathing and thus may play a key role. To investigate whether inflammatory sensitization of TRPV1 modifies neonatal ventilatory responses to heat stress, leading to respiratory alkalosis and an increased susceptibility to hyperthermic seizures, we treated neonatal rats with bacterial LPS, and breathing, arterial pH, in vitro vagus nerve activity, and seizure susceptibility were assessed during heat stress in the presence or absence of a TRPV1 antagonist (AMG-9810) or shRNA-mediated TRPV1 suppression. LPS-induced inflammatory preconditioning lowered the threshold temperature and latency of hyperthermic seizures. This was accompanied by increased tidal volume, minute ventilation, expired CO₂, and arterial pH (alkalosis). LPS exposure also elevated vagal spiking and intracellular calcium concentrations in response to hyperthermia. TRPV1 inhibition with AMG-9810 or shRNA reduced the LPS-induced susceptibility to hyperthermic seizures and altered the breathing pattern to fast shallow breaths (tachypnea), making each breath less efficient and restoring arterial pH. These results indicate that inflammation exacerbates thermal hyperpnea-induced respiratory alkalosis associated with increased susceptibility to hyperthermic seizures, primarily mediated by TRPV1 localized to vagus neurons.

Clinical Relevance

Our results provide valuable insight into the impact of inflammation on neonatal respiratory control in the context of heat stress. These findings are timely, given the increased risk of heat-induced convulsions and infections associated with the impending climate catastrophe of global warming. Our findings provide potential anatomical and pharmaceutical targets for therapeutic interventions aimed at reducing subsequent febrile seizure susceptibility and the risk of seizures after heat stress in humans.

Extreme heat caused by climate change is increasing the transmission of infectious diseases and resulting in a sharp rise in heat-related illness and mortality (1, 2). Many of these illnesses involve changes in brain function, including changes in cardiorespiratory control and convulsions, which may be interconnected. Both hyperthermia and fever cause hyperventilation (3–6), but when alveolar ventilation exceeds metabolic requirements, respiratory alkalosis ensues (7, 8). Hyperthermia-induced increases in tidal volume relative to respiratory rate (thermal hyperpnea) form the basis of hyperthermia-induced respiratory alkalosis (8). Hypocapnia, indicative of respiratory alkalosis, is a consistent finding after febrile seizures in children, the severity being more pronounced the closer the analysis is performed to the onset of febrile seizures (9–11). Although likely to stem from fever-induced hyperventilation (9–11), whether hypocapnia is a cause or a consequence of febrile seizures is uncertain. The reason for this is that plethysmography and blood gas measurements have not been done before and during febrile seizures, because most febrile seizures occur at home and bicarbonate concentrations have not been reported (9–11).

In humans, temperatures consistent with febrile seizures (>39.2°C) can result in profound respiratory alkalosis with a 20–30 mm Hg decrease in carbon dioxide tension (Pco₂) and a 0.1–0.2 increase in pH (12). An alkalosis of this magnitude likely increases cerebral excitability, which is proposed to trigger febrile seizures (5). Indeed, studies in animal models indicate that hyperthermia-induced hyperventilation can induce seizures by causing respiratory alkalosis (4, 5), but the mechanisms underlying thermal hyperpnea and subsequent seizures in the context of inflammation are not fully understood. In rodents, experimental febrile seizures are induced by exposure to exogenous heat in the absence of inflammation (13). These rodent models, which concurrently act as models of heat stress, indicate that extreme hyperthermia alone can induce hyperventilation, respiratory alkalosis, and seizures (4, 5, 14). Indeed, elevated temperatures act on the hypothalamus, medulla, and spinal cord (7, 8) and on peripheral thermosensitive neurons within the carotid sinus (15) and vagus nerves (16, 17), any of which might drive an increase in ventilation.

Transient receptor potential cation channel subfamily V member 1 (TRPV1) is a nonselective cation channel that is activated by heat (18). We recently showed that activation of TRPV1 in the vagus nerve, but not in the carotid sinus nerve, exacerbates thermal hyperpnea and convulsions in neonatal rats (4). However, because TRPV1 has a higher activation threshold (∼42°C) than that of most febrile seizures (39–40°C), its possible involvement in febrile seizures is problematic unless it is somehow sensitized before heat activation. Inflammation, which causes fever during an infection and is invariably present during febrile seizures and involved in the pathogenesis of febrile seizures (19–21), is a potent TRPV1 sensitizer (22–25). Interestingly, inflammation also modulates breathing in neonatal animals (26) and may exacerbate hyperthermic seizures. Thus, we hypothesize that inflammation, acting via TRPV1, is critical to the activation of hyperthermia-induced respiratory alkalosis and subsequent hyperthermia-induced convulsions in neonatal rats. Some of the results of these studies were previously reported in the form of an abstract (27).

Methods

Animals

Male and female Sprague Dawley rats born to breeders within our facility were used. The study was approved by the University of Calgary Animal Care and Use Committee (protocol ID: AC20-0001) and conducted in accordance with the Canadian Council of Animal Care.

Inflammatory Preconditioning

Rat pups from multiple litters received daily i.p. injections of saline or LPS (400 μg/kg; Escherichia coli O111:B4; Sigma-Aldrich) from Postnatal Days 6–10 (P6–P10). Experiments were performed 3 hours after the final injection (Figure 1A).

Figure 1.

Inflammation increases febrile seizure (FS) susceptibility in neonatal rats. The data are presented as mean ± SEM. (A) Schematic representation of the experimental protocol. (B) Body weight of saline- (n = 15) and LPS-exposed (400 μg/kg; n = 22) rat pups (two-way repeated measures ANOVA followed by Holm-Sidak post hoc comparisons; *P < 0.05). (C) Baseline body temperature before induction of FS at Postnatal Day (P)10 (two-tailed Student’s t test). (D–F) FS thresholds (two-tailed Student’s t test). (D) Threshold body temperature. (E) Threshold delta body temperature. (F) Seizure latency (*P < 0.05; **P = 0.01; ***P < 0.001).

Cytokine Analysis

Serum cytokine concentrations were assessed using the Rat 27-Plex Cytokine/Chemokine Array Discovery Assay (Eve Technologies) and Rat Bradykinin ELISA Kit (E-EL-R0124; Elabscience).

Hyperthermic Seizure Induction

Seizures characterized by tonic-clonic movements of all four limbs, with loss of the righting reflex, were induced using heated dry air (4) in pups subjected to inflammatory preconditioning and TRPV1 antagonist AMG-9810 (100 mg/kg i.p.; Abcam) or shRNA-mediated TRPV1 knockdown.

shRNA-mediated Vagal TRPV1 Inhibition

The shRNA sequence targeting rat TRPV1 (Table 1) was designed and packaged in an adeno-associated virus serotype 9 viral vector (28, 29). A Nanoject II microinjector (Drummond Scientific Co.) was used to deliver TRPV1 shRNA (1.0 × 10¹³ genomic copies/μl) or scrambled shRNA (1.6 × 10¹³ genomic copies/μl) bilaterally in the nodose ganglia at P1 (4).

Table 1.

DNA Sequences of shRNA Targeting Rat TRPV1 (accession NM_031982) and Scrambled Control

Construct	97-mer shRNA Sequence
pAAV.U6.shRNA.rTRPV1.CMV. eGFP.WPRE.hGH (TRPV1-shRNA)	5′-TGCTGTTGACAGTGAGCGCGCAGGAAGAAGTTCAACTGAATAGTGAAGCCACAGATGTA TTCAGTTGAACTTCTTCCTGCTTGCCTACTGCCTCGGA-3′
pAAV.U6.shRNA.Scramb.CMV. eGFP.WPRE.hGH (scrambled-shRNA)	5′-TGCTGTTGACAGTGAGCGATTATATTCGTGTAATGATACTTAGTGAAGCCACAGATGTA AGTATCATTACACGAATATAACTGCCTACTGCCTCGGA-3′

Definitions of abbreviations: TRPV1 = Transient receptor potential cation channel subfamily V member 1.

The antisense sequence is underlined.

Plethysmography

The ventilatory response to hyperthermia was assessed using head-out plethysmography (4).

Blood Gas Measurement

Arterial blood gases collected from the descending aorta were measured at baseline and after heat exposure in pups treated with or without AMG-9810, using the Stat Profile pHOx Ultra blood gas analyzer (Nova Biomedical Corporation).

Vagus Nerve Heat Sensitivity

Electrophysiological recording.

We used the superfused in vitro vagus/nodose ganglion preparation (4) to assess the maximum neural spiking activity and temperature index (TI), that is, the temperature and latency at 20% (TI_20%) and 50% (TI_50%) of the maximum neural activity in response to 1-minute 45°C challenges in the presence or absence of 10 μM AMG-9810 or shRNA-mediated vagal TRPV1 inhibition.

Calcium imaging.

Vagus/nodose ganglion was loaded with fluorescent Ca^2 +  indicator Fluo-4 (Thermo Fisher Scientific) (30). Ca²⁺  transients were elicited with 1-minute 45°C heat challenges. Fluorescence signals were captured with an Olympus BX-51 microscope (Warner Instruments) equipped with a charge-coupled device camera (Evolve Delta, Teledyne Photometrics) and analyzed using MetaMorph NX (Olympus America Inc.) and ImageJ (imagej.nih.gov).

RNA Extraction and qRT-PCR

Total RNA was extracted from pooled nodose ganglia from TRPV1 shRNA– and scrambled shRNA–treated pups using TRIzol reagent (Invitrogen). cDNA was generated using the SuperScript VILO cDNA Synthesis Kit (Invitrogen) and amplified in qRT-PCR using TaqMan Gene Expression Assays for Trpv1 (Rn_00583117_m1) and Gapdh (Rn_01775763_g1) (Thermo Fisher Scientific). Relative TRPV1 expression levels were normalized to GAPDH.

Statistical Analysis

The data were analyzed using SigmaPlot 14.5 (Systat Software). Results are presented as mean ± SEM. A two-tailed Student’s t test or Mann-Whitney rank-sum test, one-way ANOVA, two-way repeated measures ANOVA, linear regressions, and one-way analysis of covariance were used to assess differences between treatment groups as indicated. P values <0.05 were considered significant.

Results

Inflammation Exacerbates Hyperthermic Seizure Susceptibility and the Ventilatory Response to Heat

Neonatal LPS exposure increased serum concentrations of several inflammatory mediators two- to eightfold, including proinflammatory cytokines IL-1β and TNF, antiinflammatory cytokine IL-10, and chemokines MIP-1α and MCP-1, but it resulted in a two- to fourfold decrease (including IL-6) or no change in the concentrations of other inflammatory mediators (Table 2). LPS-exposed pups were ∼3 g smaller than control animals between P8 and P10 (Figure 1B), but this did not affect baseline rectal temperature measured 3 hours after the last LPS injection and immediately before the onset of heat exposure (Figure 1C). The seizure threshold temperature (Figures 1D and 1E) and latency (Figure 1F) were lower in LPS-exposed pups, indicative of increased hyperthermic seizure susceptibility.

Table 2.

Serum Concentrations (pg/ml) of Inflammatory Mediators after Inflammatory Preconditioning

Cytokine	n (Sal/LPS)	Saline	LPS	Response	P Value
IL-1β	8*/9	19.42 ± 5.57	42.58 ± 8.61	Increase	0.008
TNF	9/8*	11.68 ± 1.80	88.60 ± 12.71	Increase	<0.001†
IL-10	8*/9	54.89 ± 7.34	417.99 ± 70.74	Increase	0.0002
IL-5	8*/9	67.08 ± 4.21	138.05 ± 11.59	Increase	<0.0001
MIP-1α	9/9	28.01 ± 1.75	227.59 ± 33.06	Increase	<0.0001
MCP-1	9/9	2,927.41 ± 300.31	22,243.62 ± 4,701.53	Increase	<0.001†
IP-10	9/9	183.25 ± 6.91	649.51 ± 74.98	Increase	<0.001†
Fractalkine	9/9	107.41 ± 6.50	211.37 ± 26.19	Increase	0.001
IL-6	8‡/8‡	780.06 ± 87.77	510.97 ± 84.21	Decrease	0.044
IL-1α	8*/9	115.67 ± 33.10	29.83 ± 6.32	Decrease	0.006†
IL-17A	9/9	25.38 ± 1.70	15.97 ± 0.94	Decrease	0.0002
G-CSF	9/9	39.07 ± 6.37	15.22 ± 2.21	Decrease	0.003
IL-12	8‡/8‡	92.84 ± 16.18	51.13 ± 7.07	Decrease	0.033
LIX	9/8*	4,027.76 ± 90.87	956.30 ± 145.85	Decrease	<0.0001
RANTES	9/9	2,072.75 ± 333.04	790.79 ± 55.96	Decrease	0.002
Leptin	9/8*	8,716.45 ± 1,132.02	3,359.72 ± 722.01	Decrease	<0.001†
IL-2	9/9	13.65 ± 2.21	10.58 ± 0.76	Unchanged	0.166†
IL-4	9/9	15.22 ± 1.14	15.95 ± 1.47	Unchanged	0.701
IL-18	9/9	692.16 ± 119.97	503.61 ± 98.91	Unchanged	0.133†
EGF	9/7‡	9.78 ± 3.04	6.71 ± 1.98	Unchanged	0.442
VEGF	9/9	116.25 ± 4.40	104.92 ± 12.46	Unchanged	0.404
Eotaxin	9/9	6.71 ± 0.60	6.62 ± 0.51	Unchanged	0.914
Bradykinin	9/9	2.94 ± 0.41	3.16 ± 0.20	Unchanged	0.637

Data are presented as mean ± SEM. Five additional inflammatory mediators were removed from the analysis because the concentrations were too low: GM-CSF, IL-13, IFN-γ, GRO/KC, and MIP-2. A two-tailed Student’s t test was used unless otherwise indicated.

^*One outlier removed.

^† Mann-Whitney rank-sum test.

^‡ Concentrations too low to be detected in one or two samples.

Inflammatory preconditioning did not impact respiratory rate compared with control animals either at baseline or in response to heat stress (Figures 2A–2C). However, both tidal volume (Figure 2D) and minute ventilation ($\dot{\mathrm{V}}$e) (Figure 2E) were elevated at baseline in LPS-exposed pups. Despite a transient decrease in breathing variables at the onset of heat exposure (1–3 min) that closed the breathing gap between treatment groups (4–7 min), both tidal volume and $\dot{\mathrm{V}}$e increased significantly thereafter (8–15 min) in LPS-exposed pups, which correlated with their seizure latency. The rate of expired CO₂ ($\dot{\mathrm{V}}$co₂) was greater in LPS-exposed pups independent of ambient temperature (Figure 2F), but the $\dot{\mathrm{V}}$e/$\dot{\mathrm{V}}$co₂ slope was similar to that of control animals (Figure 2G), indicating that the hyperpnea in LPS-exposed pups was metabolically driven.

Figure 2.

The inflammation-induced increased FS susceptibility is associated with thermal hyperpnea and respiratory alkalosis. The data are presented as mean ± SEM. Representative breathing traces for (A) saline- and (B) LPS-exposed P10 rats at baseline and during hyperthermia (vertical and horizontal scale bars, 0.05 V and 0.5 s, respectively). (C–G) Summary graphs of breathing variables at baseline and during hyperthermia in saline- (n = 10) and LPS-exposed (n = 14) pups (two-way repeated measures ANOVA followed by Holm-Sidak post hoc comparisons; *P < 0.05; ^#P < 0.01; ^$P < 0.001). (C) Respiratory rate (RR). (D) Tidal volume (Vt). (E) Minute ventilation ($\dot{\mathrm{V}}$e). (F) Rate of expired CO₂ ($\dot{\mathrm{V}}$co₂; group effect; P = 0.047). (G) $\dot{\mathrm{V}}$e/$\dot{\mathrm{V}}$co₂ slopes for saline- (adjusted R² = 0.828; P < 0.001) and LPS-exposed (adjusted R² = 0.769; P < 0.001) rat pups (one-way analysis of covariance; P = 0.966). (H–J) Arterial blood gas measurements in cohorts different from those in C–E of saline- and LPS-exposed rat pups at baseline and at 9 minutes of hyperthermia before expected FS onset in LPS-exposed pups (one-way ANOVA followed by Holm-Sidak post hoc comparisons; *P < 0.05, **P < 0.01, ***P < 0.001). (H) pH (baseline: saline, n = 11; LPS, n = 12; hyperthermia: saline, n = 10; LPS, n = 11). (I) Pco₂ (baseline: saline, n = 11; LPS, n = 12; hyperthermia: saline, n = 11; LPS, n = 13). (J) HCO₃⁻ (baseline: saline, n = 11; LPS, n = 12; hyperthermia: saline, n = 11; LPS, n = 13).

Blood gas measurements revealed higher arterial pH (Figure 2H) in LPS-exposed pups at baseline, and although arterial pH increased during hyperthermia in both groups, it was alkalotic in LPS-exposed pups compared with control animals after 9 minutes of heat exposure, just before the expected time of seizure onset. Arterial Pco₂ (Figure 2I) and HCO₃⁻ (Figure 2J) were lower in LPS-exposed pups both at baseline and in response to heat, and both variables were lower during hyperthermia than at baseline in both treatment groups. Together these findings indicate that inflammation exacerbates hyperthermia-induced hyperventilation and the associated alkalosis already present before seizure onset and increases susceptibility to hyperthermic seizures.

Inflammation-induced Febrile Seizure Susceptibility and Hyperthermic Ventilatory Response Are TRPV1 Dependent

AMG-9810 administered i.p. did not affect baseline rectal temperature of either LPS-exposed or control pups (Figure 3A). AMG-9810 reversed the impact of LPS on hyperthermic seizure expression such that the threshold temperature (Figures 3B and 3C) and latency (Figure 3D) were no longer different between LPS-exposed and control pups.

Figure 3.

Transient receptor potential cation channel subfamily V member 1 (TRPV1) antagonist AMG-9810 (AMG) reverses the LPS-induced FS susceptibility in neonatal rats. The data are presented as mean ± SEM. (A) Baseline body temperature before FS induction at P10 (one-way ANOVA; P = 0.324). (B–D) Seizure thresholds (one-way ANOVA followed by Holm-Sidak post hoc comparisons; *P < 0.05, **P < 0.01, ***P < 0.001). (B) Threshold body temperature. (C) Threshold delta body temperature. (D) Seizure latency. Saline/vehicle: n = 14; LPS/vehicle: n = 18; saline/AMG: n = 10; LPS/AMG: n = 10.

AMG-9810 changed the breathing pattern of LPS-exposed rats in response to heat, which was characterized by an elevated respiratory rate (Figures 4A–4E) with no effect on tidal volume (Figure 4F), resulting in a further increase in $\dot{\mathrm{V}}$e (Figure 4G). This AMG-9810–dependent breathing pattern was associated with a decrease in the rate of $\dot{\mathrm{V}}$co₂ (Figure 4H) and an increase in the $\dot{\mathrm{V}}$e/$\dot{\mathrm{V}}$co₂ slope (Figure 4I), which indicates that the augmented heat-induced hyperventilation in LPS-exposed pups treated with AMG-9810 was not metabolically driven. We found no statistically significant effect of AMG-9810 injected into naive rats on the respiratory parameters tested (as we previously reported [4]).

Figure 4.

TRPV1 antagonist AMG-9810 (AMG) changes the breathing pattern in response to hyperthermia from hyperpnea to tachypnea and blocks LPS-induced respiratory alkalosis. The data are presented as mean ± SEM. Representative breathing traces at baseline and during hyperthermia for P10 rats exposed to (A) saline/vehicle, (B) AMG/vehicle, (C) LPS/vehicle, (D) LPS/AMG (vertical and horizontal scale bars, 0.05 V and 0.5 s, respectively). (E to I) Summary graphs of breathing variables at baseline and in response to hyperthermia in a different cohort of saline/vehicle- (n = 10), LPS/vehicle- (n = 10), AMG/vehicle- (n = 12), and LPS/AMG-exposed (n = 12) pups (two-way repeated measures ANOVA followed by Holm-Sidak post hoc comparisons). (E) Respiratory rate (RR; *P < 0.05, ^#P < 0.01 [LPS/AMG vs. saline/vehicle]; ^$P < 0.05 [LPS/AMG vs. LPS/vehicle]). (F) Vt. *P < 0.05, ^#P < 0.01 (saline/vehicle and AMG/vehicle vs. both LPS/vehicle and LPS/AMG)]. (G) $\dot{\mathrm{V}}$e. *P < 0.05 (LPS/AMG vs. LPS/vehicle), *P < 0.05, ^#P < 0.01, ^$P < 0.001 (LPS/AMG vs. saline/vehicle); *P < 0.05, ^#P < 0.01 (LPS/vehicle vs. saline/vehicle)]. (H) Rate of $\dot{\mathrm{V}}$co₂. *P < 0.05, ^$P < 0.001 (LPS/AMG vs. saline/vehicle); ^#P < 0.01, ^$P < 0.001 (LPS/AMG vs. LPS/vehicle); *P < 0.05, ^#P < 0.01 (LPS/vehicle vs. saline/vehicle). (I) $\dot{\mathrm{V}}$e/$\dot{\mathrm{V}}$co₂ slopes for LPS/AMG- (adjusted R² = 0.644, P < 0.001), LPS/vehicle- (adjusted R² = 0.659, P < 0.001), AMG/vehicle- (adjusted R² = 0.819, P < 0.001), and saline/vehicle-exposed (adjusted R² = 0.494, P < 0.001) rat pups (one-way analysis of covariance; P < 0.001 [LPS/AMG vs. LPS/vehicle, AMG/vehicle, and saline/vehicle], P = 0.520 (LPS/vehicle vs. saline/vehicle)]. (J–L) Arterial blood (J) pH, (K) Pco₂, and (L) HCO₃⁻ in different cohorts of saline/AMG- and LPS/AMG-exposed rat pups at baseline (saline/AMG, n = 10; LPS/AMG, n = 10) and at 9 minutes of hyperthermia (saline/AMG, n = 10; LPS/AMG, n = 9) before FS onset in LPS-exposed pups (one-way ANOVA followed by Holm-Sidak post hoc comparisons; **P < 0.01 and ***P < 0.001).

AMG-9810 also reversed the effect of LPS on arterial pH (Figure 4J) such that pH was no longer different from that in control animals either at baseline or during hyperthermia, and AMG-9810 blocked the pH increase in response to heat exposure in LPS-exposed pups. AMG-9810, however, did not alter the effect of LPS on arterial Pco₂ (Figure 4K) or HCO₃⁻ (Figure 4L). Together these results demonstrate that during inflammation, hyperthermic seizure susceptibility is dependent on the breathing pattern in response to heat stress, and this effect is mediated in part by TRPV1.

Inflammation-induced Vagus Nerve Heat Sensitivity Is TRPV1 Mediated

We next investigated the effect of inflammation on vagus nerve activity in response to heat to elucidate a mechanism underlying the inflammation-induced hyperthermic ventilatory response and subsequent hyperthermic seizure susceptibility. We selected to study the vagus because we have previously shown that bilateral vagus nerve dissection abrogated the effect of TRPV1 activation to exacerbate hyperthermic seizures (4). Brief heat challenges increased the spiking activity of isolated in vitro vagus nerves from both control (Figure 5A) and LPS-exposed pups (Figure 5B), but whereas subsequent heat challenges had no further effect on control animals, spiking activity almost doubled in the LPS group (Figures 5A–5C). In addition, a comparison of the second heat challenge revealed that the TI_20% and TI_50% were lower (Figures 5D–5F) and the latency to TI_20% and TI_50% was shorter (Figure 5G) in the LPS group, which indicates that inflammatory preconditioning enhances the heat sensitivity of the vagus nerve. Application of AMG-9810 did not significantly impact the spiking activity of control vagus nerves, but it blocked the LPS-enhanced spiking activity (Figures 5H–5K). This demonstrates that the inflammation-induced vagus nerve heat sensitivity is TRPV1 mediated.

Figure 5.

Inflammatory sensitization of TRPV1 increases the temperature sensitivity of the vagus nerves. The data are presented as mean ± SEM. Representative traces of the in vitro neural spiking activity (frequency normalized to baseline) of the vagus nerve from (A) saline- and (B) LPS-exposed rat pups in response to repetitive 1-minute heat (45°C) challenges (indicated by black arrowheads) and K⁺ (15 mM) (also in H and I). Insets are representative raw traces. (C) Summary graph of the vagal activity normalized to baseline for six heat challenges and one K⁺ challenge (two-way repeated measures ANOVA followed by Holm-Sidak post hoc comparisons; **P < 0.01 compared with saline controls (n = 5); ^!P = 0.018; ^!!!P < 0.001 compared with the first heat challenge in LPS-exposed vagi [n = 5]). Representative tracings of the temperature change and vagal response showing the temperature index (TI) and latency to 20% or 50% of the maximum vagal activity during the second heat challenge in (D) saline- and (E) LPS-exposed vagi. (F) Summary graph of the TI_20% and TI_50% (Student’s t test; *P = 0.019, **P = 0.009). (G) Summary graph of the latency to TI_20% and TI_50% (Student’s t test; **P = 0.003, ***P < 0.001). Representative traces of the in vitro temperature response of the vagus nerve from (H) saline- and (I) LPS-injected animals in the presence of AMG. (J) Summary graph of the vagal activity in the presence of AMG normalized to baseline (one-way ANOVA followed by Holm-Sidak post hoc comparisons within each group, saline [n = 5], LPS [n = 5]; *P = 0.025, **P = 0.004, ***P < 0.001). (K) The AMG-induced change in vagal activity normalized to the second heat response without AMG-9810 (two-way repeated measures ANOVA followed by Holm-Sidak post hoc comparisons; *P = 0.025 [group effect]).

Targeted Inhibition of Vagal TRPV1 Attenuates the Effects of Inflammation on Vagal Heat Sensitivity and Hyperthermic Seizure Susceptibility

A brief, 9-day neonatal exposure to shRNA targeting TRPV1 in vagal nodose ganglia (Figures 6A and 6B) resulted in ∼15% knockdown of vagal TRPV1 mRNA (Figure 6C). This small reduction in vagal TRPV1 mRNA blocked the effects of inflammation on vagus spiking activity in response to heat but did not alter spiking activity in the control group (Figure 6D). Ca²⁺ imaging (Figures 6E–6H) revealed a rapid onset (at temperatures as low as 38.2°C) and sustained heat-induced Ca²⁺ release in nodose ganglia cells from the LPS group (Figures 6F and 6I), which was completely blocked by TRPV1 knockdown (Figures 6H and 6I).

Figure 6.

shRNA-mediated targeted inhibition of vagal TRPV1 blocks the LPS-induced increased temperature sensitivity of the vagus nerve. The data are presented as mean ± SEM. (A) Schematic representation of the experimental protocol. (B) Photograph showing microinjection of TRPV1-targeted shRNA into the vagal nodose ganglion at P1 (scale bar, 5 mm; CCA = common carotid artery; ECA = external carotid artery; NG = nodose ganglion; OCA = occipital artery). (C) Real-time (RT)-PCR quantification of TRPV1 mRNA concentrations in vagal nodose ganglia at P10 after in vivo transduction with AAV9 encoding scrambled shRNA (control; n = 19) or TRPV1-targeted shRNA (n = 18) (two-tailed Student’s t test; P = 0.019). (D) The heat response (frequency normalized to baseline) of the vagus from saline- and LPS-exposed pups expressing scrambled shRNA (control) or TRPV1 shRNA (one-way ANOVA followed by Holm-Sidak post hoc comparisons; ***P < 0.001; control/saline [n = 6], control/LPS [n = 7], shRNA/saline [n = 6], shRNA/LPS [n = 6]). (E– H) Representative traces of intracellular Ca²⁺ in response to a 45°C heat challenge in Fluo-4–loaded vagal nodose ganglion cells from saline- and LPS-exposed pups expressing scrambled shRNA or TRPV1 shRNA. Arrow: onset of K⁺ (15 mM) application. Each colored trace depicts a different cell. (E) Control saline. (F) Control LPS group. (G) shRNA saline. (H) shRNA LPS. (I) Summary data of intracellular Ca²⁺ in vagal nodose ganglia from control/saline- (n = 5), control/LPS- (n = 5), shRNA/saline- (n = 5), and shRNA/LPS-treated (n = 6) pups (two-way repeated measures ANOVA followed by Holm-Sidak post hoc comparisons; ^$$$P < 0.001 [control/LPS vs. control/Sal]; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 [control/LPS vs. shRNA/Sal]; *P < 0.05, **P < 0.01 [control/LPS vs. shRNA/LPS]; ^!P < 0.05 [control/Sal vs. shRNA/LPS]).

We next assessed the impact of vagal TRPV1 knockdown on hyperthermic seizure expression. We noted that baseline temperatures (Figure 7A) in these pups were ∼1°C higher than those in pups that did not have surgery. Seizure threshold temperatures were also much higher than previously observed and were not significantly impacted by inflammatory preconditioning, regardless of whether vagal TRPV1 was inhibited (Figures 7B and 7C). However, TRPV1 inhibition significantly reversed the effect of inflammatory preconditioning on seizure latency (Figure 7D). Together, these findings confirm that inflammatory preconditioning sensitizes vagal TRPV1 channels to heat, which in turn contributes to increased hyperthermic seizure susceptibility.

Figure 7.

shRNA-mediated targeted inhibition of vagal TRPV1 attenuates the proconvulsant effects of LPS. The data are presented as mean ± SEM. (A) Baseline body temperature of saline- and LPS-pretreated pups injected with scrambled shRNA (control) or TRPV1 shRNA before the induction of FS at P10 (one-way ANOVA; P = 0.422; control/saline [n = 9], control/LPS [n = 8], shRNA/saline [n = 8], shRNA/LPS [n = 7]). (B–D) The effect of shRNA-mediated targeted inhibition of vagal TRPV1 receptors on FS thresholds (one-way ANOVA followed by Holm-Sidak post hoc comparisons; **P < 0.01, ***P < 0.001). (B) Threshold body temperature. (C) Threshold delta body temperature. (D) Seizure latency.

Discussion

This study investigated the interaction between inflammation and TRPV1 in the breathing response to heat stress and its relevance to hyperthermic seizures in neonatal rats. We found that LPS-induced inflammation exacerbated thermal hyperpnea, leading to respiratory alkalosis and an increased propensity for hyperthermic seizures. Pharmacological inhibition of TRPV1 with AMG-9810 changed the breathing pattern in hyperthermic, LPS-treated animals from hyperpnea to tachypnea (panting). This was associated with a significant reduction in the inflammation-induced respiratory alkalosis and associated hyperthermic seizure susceptibility, suggesting that TRPV1 activation increases the depth and efficiency of breathing in neonates. We also found that the heat sensitivity of the vagus nerve was enhanced in LPS-treated rats, which was blocked by AMG-9810. shRNA-mediated knockdown of vagal TRPV1 had a similar effect in vitro and had antiseizure effects in vivo. To our knowledge, this is the first study to demonstrate the effects of chronic inflammation on thermal hyperpnea and subsequent hyperthermic seizures in neonatal animals and to show that this interaction is mediated by TRPV1.

We studied the effects of chronic exposure to inflammation on breathing and hyperthermic seizures by exposing neonatal rats to LPS repeatedly (31). This chronic exposure better models children with febrile seizures who are often ill for days before seizure onset (32) and the many people who have chronic inflammatory conditions who are exposed to extremely high temperatures because of global warming (1). We observed that chronic LPS exposure resulted in changes in the concentrations of inflammatory cytokines and mediators, similar to a single acute LPS injection (33). One notable exception is a reduction in IL-6, which is important for fever genesis (31, 34). Interestingly, we did not observe a fever after the LPS injections, most likely because of the low ambient temperature or because of endotoxin tolerance leading to lower IL-6.

The effects of acute inflammation on neonatal respiratory control in the context of hypoxia and hypercapnia have been reported previously. Acute inflammation attenuated the ventilatory response to hypoxia but had no effect on the hypercapnic ventilatory response (33, 35, 36). With regard to thermal hyperpnea, the single study that assessed only two time points (baseline and 38°C) reported a decrease in breathing during hyperthermia with no effect of an acute single LPS injection (33). In contrast, we found that chronic inflammation increased baseline breathing and exacerbated thermal hyperpnea, mostly because of an increase in tidal volume. The enhanced ventilation in the LPS-exposed pups was metabolically driven (Q₁₀ effect), as indicated by the increased rate of $\dot{\mathrm{V}}$co₂, which resulted in lower arterial Pco₂ (hypocapnia) and higher arterial pH (alkalosis) than baseline and in saline-treated control animals. These data suggest that inflammatory preconditioning exacerbates thermal hyperpnea and accelerates the development of hyperpnea-induced respiratory alkalosis during heat stress.

Our present study found that TRPV1 antagonist AMG-9810 reversed the effects of inflammation to exacerbate thermal hyperpnea in neonatal animals. This is likely due to a direct effect of AMG-9810 on breathing as opposed to an effect on metabolism. Supportively, we found in LPS-exposed pups that the TRPV1 antagonist caused a shift in the breathing pattern from thermal hyperpnea (increased tidal volume) to panting (thermal tachypnea; increased respiratory rate) in response to heat stress. Panting increases ventilation of the upper airways but has minimal effects on alveoli ventilation and thus rarely results in respiratory alkalosis (8). Importantly, our data demonstrate that LPS-induced respiratory alkalosis is abolished with AMG-9810 treatment. These results indicate that inflammatory sensitization of TRPV1 facilitates a breathing pattern that contributes to the development of thermal hyperpnea and consequent respiratory alkalosis at a clinically relevant temperature for febrile seizures.

The involvement of vagal afferents in regulating the hyperthermic ventilatory response has been established previously (4, 16, 37). Vagal afferents are heat sensitive (38), express TRPV1 (4, 39), and terminate on neurons in the nucleus of the solitary tract, which project to and modulate respiratory networks (26). Specifically in the context of inflammation, vagal afferents mediate the effects of inflammation on the ventilatory response to hypoxia but not hypercapnia (35). We show that vagal afferents are involved in mediating the effects of inflammation on the ventilatory response to heat stress. Using in vitro electrophysiology and calcium imaging, we found that the vagal response to heat stress was more rapid, robust, and sustained and had lower temperature indices in vagi from LPS-exposed than in control pups. We acknowledge that high temperatures in the range of 42.8–43.9°C were required to elicit a vagal response in the electrophysiological studies, likely because the isolated vagus nerve preparation lacks afferent input and other modulators of nodose neurons. However, in the calcium imaging studies, temperatures as low as 38°C produced a robust response in vagi from LPS-exposed animals, likely because calcium imaging in this context is a more sensitive measure of early changes in excitability.

The stimulatory effect of inflammation on the vagal response to heat stress was also TRPV1 mediated, because AMG-9810 attenuated the LPS-induced enhanced vagal heat responses. shRNA targeting TRPV1 knockdown also suppressed the LPS-induced vagal heat response. Surprisingly, although shRNA-mediated knockdown of vagal TRPV1 mRNA was low (15%), the physiological effect on vagi from LPS-exposed rats was robust. It appears that shRNA exposure had an inhibitory effect on the responsiveness of TRPV1-containing cells to heat stress, which is independent of the effect of shRNA on TRPV1 expression, similar to other studies using antisense oligonucleotides (40). The lower temperature requirement to elicit the pronounced response in vagi from LPS-exposed pups is indicative of inflammatory sensitization of vagal TRPV1 to heat after LPS exposure. This TRPV1 sensitization may occur via direct binding of LPS to Toll-like receptor 4 (41) on TRPV1-containing neurons or indirectly via proinflammatory mediators, including IL-1β, TNF, and bradykinin (23, 25, 42, 43). Whatever the inflammatory mediator, this work on hyperthermic seizures in rodents now directly implicates TRPV1 in febrile seizures and heat-induced convulsions in humans by demonstrating its involvement in precipitating the alkalosis-induced seizures at physiologically relevant temperatures. A previous criticism of hyperthermia-induced seizures as a model of febrile seizures was the extremely high temperature often needed to precipitate the seizures. The present work provides important validation of the model if used in the context of an inflammatory background.

Previous studies showed that hyperthermia-induced respiratory alkalosis triggers hyperthermic seizures (4, 5, 14). Here, we show that inflammation accelerates initiation of hyperthermic seizures, likely because of the early onset of the breathing pattern that predisposes neonates to respiratory alkalosis. Treatment with AMG-9810 or exposure to shRNA targeting vagal TRPV1 had antiseizure effects in LPS-exposed animals but not control pups. These findings indicate that the effect of inflammation on hyperthermic seizures is, at least in part, TRPV1 mediated, with a peripheral initiating site in the vagus. We note that in the shRNA study, there was no difference in the temperature threshold for hyperthermic seizures, although there was a robust effect on the seizure latency. This might be due to the surgical manipulations, because all animals developed hyperthermia after surgery. We acknowledge that inflammation influences the development and function of neurons in brainstem respiratory networks (26, 37, 44) and has known widespread effects to elevate brain cytokines and increase brain excitability (21, 45). These effects, which most likely occur in tandem with those reported here, may act independently or synergistically to cause hyperthermic seizures. The present findings are important because they highlight an additional mechanism by which inflammation may predispose humans to febrile seizures or heat-induced convulsions (Figure 8).

Figure 8.

Schematic showing the interaction between inflammation and TRPV1 in the breathing mechanism underlying FS genesis. Adapted by permission from Reference (49). PGE = prostaglandin E.

Conclusions

Our results provide valuable insight into the impact of inflammation on neonatal respiratory control in the context of heat stress. These findings are timely, given the high rates of febrile seizures in children with COVID-19 (46–48) and the increased risk of heat-induced convulsions and infections associated with global warming (2). In the context of inflammation, our data suggest that the initiating site of hyperthermic seizures is at least in part peripherally based and due to an effect of inflammation to exacerbate thermal hyperpnea-induced respiratory alkalosis in a TRPV1-dependent manner. Our findings provide potential anatomical and pharmaceutical targets for therapeutic interventions aimed at reducing subsequent febrile seizure susceptibility and the risk of seizures after heat stress in humans.