Nociceptor spontaneous activity is responsible for fragmenting non-rapid eye movement sleep in mouse models of neuropathic pain

Abstract

Spontaneous pain, a major complaint of patients with neuropathic pain, has eluded study because there is no reliable marker in either preclinical models or clinical studies. Here, we performed a comprehensive electroencephalogram/electromyogram analysis of sleep in several mouse models of chronic pain: neuropathic (spared nerve injury and chronic constriction injury), inflammatory (Freund's complete adjuvant and carrageenan, plantar incision) and chemical pain (capsaicin). We find that peripheral axonal injury drives fragmentation of sleep by increasing brief arousals from non-rapid eye movement sleep (NREMS) without changing total sleep amount. In contrast to neuropathic pain, inflammatory or chemical pain did not increase brief arousals. NREMS fragmentation was reduced by the analgesics gabapentin and carbamazepine, and it resolved when pain sensitivity returned to normal in a transient neuropathic pain model (sciatic nerve crush). Genetic silencing of peripheral sensory neurons or ablation of CGRP⁺ neurons in the parabrachial nucleus prevented sleep fragmentation, whereas pharmacological blockade of skin sensory fibers was ineffective, indicating that the neural activity driving the arousals originates ectopically in primary nociceptor neurons and is relayed through the lateral parabrachial nucleus. These findings identify NREMS fragmentation by brief arousals as an effective proxy to measure spontaneous neuropathic pain in mice.

Fig. 1.. Acute activation of nociceptors causes brief arousals during NREMS.

(A) Left: Schematic representation of the experimental protocol to trigger transdermal activation of nociceptive fibers during NREMS using a 400-μm-diameter optical fiber. Right: Representative EEG and EMG traces showing an arousal triggered by nociceptive optogenetic activation in a Na(v)1.8::ChR2 mouse. The dashed line represents the 10-ms photostimulation, and the red box highlights the arousal. TTL, transistor-transistor logic. (B) Latency of arousal triggered by optogenetic activation of Na(v)1.8-positive fibers in Na(v)1.8::ChR2 mice (n = 54 trials in four mice). (C) Duration of arousals triggered by nociceptive optogenetic activation in Na(v)1.8::ChR2 mice. Blue dashed line indicates the median; purple dashed lines represent 25th and 75th quartiles. (D) EEG power spectral density during spontaneous (n = 35) and optogenetically induced (n = 38) arousals in Na(v)1.8::ChR2 mice. (E) Schematic representation of the SNI model. (F) Representative EEG and EMG traces showing a brief arousal (BA). The duration of a BA was determined by the duration of the EMG activation. In blue, NREMS; in red, BA. (G) Left: Hourly rate of brief arousals (number per minute of NREMS) across 24-hour period; and right: computed over the light and dark periods before (BSL) and after SNI in C57BL/6J wild-type (WT) mice. (H) Distribution of wake (W) episodes originating from NREMS as a function of their duration, determined by traditional scoring (4-s epoch), before (BSL) and after SNI. (I) Representative EEG and EMG recordings (30 min) under (left) baseline conditions and after (right) SNI procedure. Shown (top to bottom) are EEG spectrogram, raw EEG waveform, filtered and integrated EMG trace, filtered and normalized EEG sigma (10 to 15 Hz) power, and color-coded vigilance states. Brief arousals are in red, NREMS in light blue, REMS in dark blue, and wake in black. (J) Left: EEG power spectral density of the sigma power (10 to 15 Hz) during NREMS and (right) histograms of the phase angle values of the 0.02 Hz-fluctuation at brief arousal onset before (BSL) and after SNI. (K) Normalized EEG power spectral density during brief arousals before (BSL) and after SNI. Shading areas represent SEM. (L) Left: Schematic drawing of a paw lift occurring during brief arousals recorded by video combined with EEG/EMG recordings, and (right) quantification of paw lifts during brief arousals before (BSL) and after SNI. For all panels, data are presented as means ± SEM [n = 4 Na(v)1.8::ChR2 mice; n = 12 WT SNI mice for sleep analyses and n = 5 WT mice for EEG + paw video analyses]. Circles overlaid on the bar in histograms represent data from each individual animal. *P < 0.05 in comparison with baseline (BSL). For analysis of variance (ANOVA) values and post hoc test; please refer to table S2.

Fig. 2.. Peripheral nerve injury fragments NREMS in mice.

(A) Left: NREMS episode mean duration; center: number; and right: amount, per 2-hour bins before (BSL) and 1 week after SNI. (B) NREMS episode duration in % of BSL after SNI in male and female WT mice (n = 7 males and 11 females). (C) Distribution of wake episodes originating from NREMS during the light period as a function of duration, determined by traditional scoring (10-s epoch), before (BSL) and after SNI. (D and E) NREMS episode average duration (D) and number (E) during the light and dark periods at BSL and at different time points after SNI. Each row of the heat maps shows data from one mouse. Average absolute values represented on top and individual data shown as heatmap below (in % of own BSL). (F and G) REMS episode average duration (F) and number (G) during the light and dark periods at BSL and at various times after SNI. Average absolute values represented on top, and individual data shown as heatmap below (in % of own BSL). Data are presented as means ± SEM (n = 8 C57BL/6J WT mice for time-course analyses, n = 7 males and n = 11 females for sex analysis). Circles overlaid on the bar in histograms represent data from each individual animal. *P < 0.05 in comparison with baseline. For ANOVA values and post hoc tests, refer to table S2. w, weeks.

Fig. 3.. CCI and sciatic nerve crush fragment NREMS.

(A) Far left: Mice were instrumented for EEG/EMG monitoring and assessed for baseline sleep and pain behaviors, two chromic catgut ligatures were then loosely tied around the common sciatic nerve without blocking epineural vascularization (CCI) and assessed for 11 weeks. Left: NREMS episode mean duration; center: number; and right: amount, per 2-hour bins before (BSL) and 5 weeks after CCI. (B) Distribution of wake episodes originating from NREMS during the light period as a function of duration, determined by traditional scoring (5-s epoch), before (BSL) and after CCI. (C and D) NREMS episode average duration (C) and number (D) during the light and dark periods at BSL and at different time points after CCI. Each heatmap row shows data from one mouse (n = 7 mice). Averaged absolute values represented on top and individual data shown as heatmap below (in % of own BSL). (E) Far left: Mice were instrumented for EEG/EMG monitoring and assessed for baseline sleep and pain behaviors, then the common sciatic nerve was crushed (SNCrush), and animals were assessed for 7 weeks. Left: NREMS episode mean duration; center: number; and right: amount, per 2-hour bins before (BSL) and 2 weeks after SNCrush. (F) Distribution of wake episodes (W) originating from NREMS during the light period as a function of duration, determined by traditional scoring (10-s window), before (BSL) and (top) 2 and 3 weeks or (bottom) 5 and 7 weeks after SNCrush. (G) Left: Punctate mechanical allodynia thresholds assessed by von Frey filaments; and right: cold hyperalgesia assessed by acetone paw test after SNCrush. (H) Left: Brief arousal rate. Right: NREMS episode average duration before (BSL) and at different time points after SNCrush. Time points where pain hypersensitivity is present are highlighted in red. Data presented as means ± SEM (all mice are C57BL/6J WT; n = 7 CCI, n = 7 SNCrush for sleep analyses, and n = 10 mice for SNCrush pain behaviors). Circles overlaid on the bar in histograms represent data from each individual animal. *P < 0.05 in comparison with baseline. For ANOVA values and post hoc test, refer to table S2.

Fig. 4.. Inflammatory and neurogenic pain does not fragment NREMS.

(A to D) Left: Mechanical pain hypersensitivity; center: NREMS episode average duration; and right: distribution of wake episodes originating from NREMS as a function of duration during the light period (FCA, CARR, and INC) or for 3 hours (after cap injection) in mice after: (A) intraplantar FCA injection (20 μg in 10 μl); (B) intraplantar carrageenan (CARR) injection (1% in 10 μl); (C) intraplantar injection of capsaicin (1 μg in 10 μl); (D) plantar skin incision (INC) of hindpaw. Data presented as means ± SEM (all mice are C57BL/6J WT, n = 6 FCA, n = 6 CARR, n = 5 capsaicin, and n = 4 plantar incision). *P < 0.05 in comparison with baseline (BSL) for inflammatory models. *P < 0.05 in comparison with sham animals for the incisional model. For complete statistical analyses (ANOVA values and post hoc test), refer to table S2. d, days.

Fig. 5.. The neural activity responsible for NREMS fragmentation by brief arousals originates ectopically in sensory neurons.

(A) Top: Genetic construct to produce Cre recombinase-dependent expression of tet-tox and (bottom) CGRP protein amount quantified by ELISA from supernatant of cultured sensory neurons from Na(v)1.8::tet-tox mice (Tg) and littermates (Lm) exposed to capsaicin or KCl (n = 4 mice per group). HBSS, Hanks’ balanced salt solution. (B) Left to right: Behavioral responses of Na(v)1.8::tet-tox mice and littermates to calibrated forceps (n = 7 Lm and n = 8 Tg), pinprick (n = 9 Lm and n = 6 Tg), laser heat (n = 13 Lm and n = 12 Tg), and intraplantar injection of capsaicin (1 μg in 20 μl; n = 7 Lm and n = 7 Tg), respectively. (C) NREMS amount per 2-hour bins over 24 hours in naïve (baseline conditions) Na(v)1.8::tet-tox and littermate mice (n = 6 Lm and n = 4 Tg). (D) Far left: Na(v)1.8::tet-tox mice (silenced Aδ and C-fibers) were subjected to SNI surgery. Left, center: Distribution of wake episodes originating from NREMS during the light period as a function of duration before (BSL) and after SNI in Na(v)1.8::tet-tox mice and littermates. Right: Brief arousal rate (per min of NREMS) expressed in % of baseline (dashed line) after SNI in Na(v)1.8::tet-tox mice and littermates (n = 6 Lm, n = 4 Tg). (E) Left: NREMS episode duration; and right: amount in % of baseline (dashed line) after SNI in Na(v)1.8::tet-tox mice and littermates (n = 6 Lm and n = 4 Tg). (F) Left: TRPV1-Cre mice were bred with fl-STOP-fl-DTA mice to produce animals with TRPV1-lineage ablation, which underwent SNI. Center: Distribution of wake episodes originating from NREMS during the light period as a function of duration before (BSL) and after SNI in TRPV1::DTA mice; and right: NREMS episode duration in % of baseline (dashed line) after SNI in TRPV1::DTA mice (n = 5). (G) Intraplantar injection of 2% lidocaine and 0.5% QX-314 (10 μl was performed to block nerve terminals in the skin of SNI mice instrumented for EEG/EMG monitoring, and vehicle control. Left: Brief arousal rate and NREMS episode average duration in SNI mice after injection of 2% lidocaine/0.5% QX-314 or vehicle. Insets: Results for 6 hour after injection (n = 5 veh and 5 lido-QX). Dashed line represents the NREMS episode duration before injury (BSL). Data are presented as means ± SEM. Circles overlaid on the bar in histograms represent data from each individual animal. *P < 0.05 in comparison with baseline. #P < 0.05 in comparison with littermates. For complete statistical analyses (ANOVA values and post hoc test), refer to table S2.

Fig. 6.. Pharmacologic and genetic restoration of sleep continuity in SNI mice without causing sedation.

(A and B) (A) Brief arousal rate and (B) NREMS episode duration in 2-hour bins in SNI mice treated with vehicle or gabapentin (3 to 30 mg/kg ip). Dashed line represents the brief arousal rate before injury (BSL). (C) Left: NREMS latency; and right: NREMS amount averaged over 4 hours after injection in SNI mice of vehicle or gabapentin (3 to 30 mg/kg ip). (D) Left: Distribution of wake episodes originating from NREMS as a function of duration; center: rate of brief arousals during the first 4 hours; and right: NREMS episode average duration in 2-hour bins after injection of carbamazepine (5 mg/kg sc) in SNI mice. Dashed line represents the brief arousal rate at baseline. (E) Left: NREMS onset latency; and right: total NREMS amount during the first 4 hours after injection in SNI mice treated with vehicle or carbamazepine (5 mg/kg sc). (F) Left: Brief arousal rate; and right: NREMS episode duration for 2 hours after injection of morphine (1 mg/kg ip) in SNI mice. Data presented as means ± SEM (all mice are C57BL/6J WT; n = 7 for gabapentin, n = 5 for carbamazepine, and n = 5 for morphine). All pharmacological injections were performed at 12:00 p.m. Circles overlaid on the bar in histograms represent data from each individual animal. *P < 0.05 in comparison with vehicle. For complete statistical analyses (ANOVA values and post hoc test), refer to table S2.

Fig. 7.. Ablation of CGRP-positive neurons in the lateral PB prevents NREMS fragmentation after nerve injury.

(A) Left: Genetic construct of the AAV-Flex-DTA vector used for bilateral microinjections into the PB of CGRP-CreER mice or WT littermates; center: histological validation of CGRP-positive neuron deletion; and right: quantification of CGRP-positive neurons within the lateral PB. Scale bar, 50 μm (n = 3 mice per group). BF, bright field; Scp, superior cerebellar peduncle. (B) Left, center: Distribution of wake episodes originating from NREMS during the light period as function of duration; and right: rate of brief arousal 2 weeks after SNI as a percentage from baseline in WT littermates and PB^CGRP-DTA. (C) Left: NREMS episode average duration; and right: amount expressed in % of baseline in PB^CGRP-DTA and Lm mice 2 weeks after SNI. (D) NREMS episode number expressed as % of baseline in PB^CGRP-DTA and control mice 2 weeks after SNI. Data presented as means ± SEM (n = 7 PB^CGRP-DTA and n = 7 control Lm). Circles overlaid on the bar in histograms represent data from each individual animal. *P < 0.05 in comparison with vehicle. For complete statistical analyses (ANOVA values and post hoc test), refer to table S2.