TREK-1, a K+ channel involved in polymodal pain perception

Abstract

The TREK-1 channel is a temperature-sensitive, osmosensitive and mechano-gated K+ channel with a regulation by Gs and Gq coupled receptors. This paper demonstrates that TREK-1 qualifies as one of the molecular sensors involved in pain perception. TREK-1 is highly expressed in small sensory neurons, is present in both peptidergic and nonpeptidergic neurons and is extensively colocalized with TRPV1, the capsaicin-activated nonselective ion channel. Mice with a disrupted TREK-1 gene are more sensitive to painful heat sensations near the threshold between anoxious warmth and painful heat. This phenotype is associated with the primary sensory neuron, as polymodal C-fibers were found to be more sensitive to heat in single fiber experiments. Knockout animals are more sensitive to low threshold mechanical stimuli and display an increased thermal and mechanical hyperalgesia in conditions of inflammation. They display a largely decreased pain response induced by osmotic changes particularly in prostaglandin E2-sensitized animals. TREK-1 appears as an important ion channel for polymodal pain perception and as an attractive target for the development of new analgesics.

Figure 1

TREK-1 mRNA expression in sensory neurons. (A) TREK-1 mRNA localization in DRG of wild-type mice with an antisense probe (TREK^+/+) and a control sense probe (sense), and in DRG of knockout mice with an antisense probe (TREK^−/−) showing labeled neurons in green (as indicated by an arrowhead). (B) TREK-1 mRNA colocalization with SP, isolectin B4 (IB4) and TRPV1. TREK-1 mRNA labeling appears in green and other markers in red. Coexpressing neurons appear in yellow. A downward pointing arrow shows a double-labeled neuron, an arrowhead a TREK-1 expressing neuron and an upward pointing arrow indicates an SP-, IB4- or TRPV1-labeled neuron. The diagram on the right side indicates the percentage of coexpressing cells in each neuronal population in normal conditions. (C) The representative picture of TREK-1 (green) identified with a previously described antibody (Maingret et al, 2000) and TRPV1 (red) channels expression shows a colocalization (yellow) in small neuron bodies in DRG neurons in culture. (D) Distribution of TREK-1 mRNA positive neurons relative to their cross-sectional area and diameter in both normal and inflammatory conditions; n=3–4 mice, with at least 400–600 cells counted per condition. (E) Semiquantitative RT–PCR results of TREK-1 mRNA levels in normal (nor) and inflammatory (inf) conditions expressed as signal intensity normalized to actin control (n=3 mice).

Figure 2

TREK-1 like currents in mouse small diameter DRG neurons. (A) Activation of TREK-1 like currents by 10 μM AA in DRG neurons from TREK-1^+/+ mice. Currents were elicited by voltage ramps of 1 s duration, from −100 to +50 mV. Holding potential was −80 mV. Results were obtained in the whole-cell configuration and at room temperature (20–22°C). Inset: Capsaicin-sensitive currents recorded at −50 mV (elicited by 10 μM capsaicin, cap) in the same neuron. (B) The absence of AA-activated currents in DRG neurons from TREK-1^−/− mice (same experiments as in (a)). (C) Mean density of AA-sensitive currents (measured at 0 mV) in capsaicin-sensitive DRG neurons from TREK-1^+/+ (n=30) and TREK-1^−/− mice (n=21). (D) Percentage of small diameter DRG neurons expressing capsaicin- and AA-sensitive currents from TREK-1^+/+ mice (n=63). (E) Activation of TREK-1 like potassium currents by 10 μM AA and inhibition by 500 μM CPT-cAMP in TREK-1^+/+ DRG neurons. Membrane potential was 0 mV, the zero current is indicated by an arrow, results shown are typical of 12 experiments. (F) Activation of TREK-1 like potassium currents by 10 μM AA and inhibition by 5 μM PGE₂ in TREK-1^+/+ DRG neurons. Membrane potential was 0 mV, results shown are typical of 15 experiments. (G) Activation of TREK-1 like currents by 10 μM AA and pH 5.5 in inside-out patches from TREK-1^+/+ mouse DRG neurons. Membrane potential was 0 mV, results shown are typical of 18 experiments. (H) Dose–response curve for TREK-1 like current activation (inside-out configuration) by membrane stretch (NPo was calculated assuming a single channel current of 5 pA at 0 mV). Inset: Reversible activation of the native current by a membrane stretch to −55 mmHg. Membrane potential was 0 mV, results shown are typical of five experiments.

Figure 3

Heat responses of saphenous nerve C-fibers with receptive fields in isolated hairy skin of TREK-1^−/− mice are increased. (A) A representative original recording showing action potentials in response to noxious heat in a single fiber from a TREK-1^−/− mouse. Top inset shows an average of 643 action potentials recorded from a CMH unit (conduction velocity 0.46 m/s, von Frey threshold 16 mN). The top of the trace illustrates action potentials (Events), depicted as vertical bars, in response to heat stimulus (lower trace) applied on the chorion side of the skin. (B) Averaged histograms of heat responses of mechano-heat sensitive C-fibers of TREK-1^+/+ (n=28, dark bars) and TREK-1^−/− (n=24; white bars). Action potentials are binned per °C. (C) Total number of action potentials per heat stimulus across n=24 experiments for TREK-1^−/− (white circles, left column) and n=28 experiments for TREK-1^+/+ mice (white circles, right column). Means are dark circles.

Figure 4

Comparison of nociceptive behavior of TREK-1^+/+ (black bars) and TREK-1^−/− (white bars) mice to noxious (a, b, d) and osmotic (c) stimuli. (A) Withdrawal and paw licking latencies to noxious thermal stimulus (tail immersion test, hot plate test, respectively) are given as mean±s.e.m. (B) Paw withdrawal thresholds (mean±s.e.m) to mechanical stimulation were determined with the von Frey test (performed in healthy mice) (left) and the paw pressure assay in TREK-1^+/+ and TREK-1^−/− mice. The paw pressure test was performed before (0 day) and after (14 days) sciatic nerve ligation. (C) In the flinch test (Alessandri-Haber et al, 2003, 2005), Iso corresponds to injection of 10 μl of isotonic solution (0.9% NaCl) in the hindpaw, Hypo to an injection of 10 μl of hypotonic solution (deionized water), Hyper to an injection of 10 μl of hypertonic solution (2 or 10% NaCl). PGE₂ (100 ng/paw) was injected alone or 30 min prior to the hypotonic or hypertonic stimulus. (D) Response (mean±s.e.m.) to chemical stimuli; duration of licking and biting the hindpaw induced by formalin, number of abdominal writhings induced by acetic acid and duration of acetone-induced hindpaw stretches. Acetone experiments were carried out on non-neuropathic animals (0 day) as well as on mononeuropathic mice, 14 days after sciatic nerve ligation. P<0.05 versus wild-type animal on TREK-1^+/+ animal scores, n=10–20 per group.

Figure 5

Compared mechanical and thermal inflammatory hyperalgesia of TREK-1^−/− (white circle) and TREK-1^+/+ (black circle) mice. Withdrawal thresholds to tactile stimulation (von Frey filaments) of the right hindpaw (A) and withdrawal latencies to noxious thermal stimulus (right hindpaw immersion, 46°C) (B) were determined before and after injection of carrageenan in the right hindpaw. P<0.05 compared to TREK-1^+/+ mice, n=12 per group.