Thrombin inhibits NMDA-mediated nociceptive activity in the mouse: possible mediation by endothelin

Abstract

The CNS expresses many components of an extracellular protease signalling system, including the protease-activated receptor-1 (PAR-1) whose tethered ligand is generated by thrombin. Activation of PAR-1 potentiates NMDA receptor activity in hippocampal neurons. Because NMDA activity mediates hyperalgesia, we tested the hypothesis that PAR-1 receptors also regulate pain processing. In contrast to the potentiating effect of thrombin in the hippocampus, NMDA-induced behaviours and the transient mechanical hyperalgesia (von Frey fibres) induced by intrathecally injected NMDA in mice were inhibited by thrombin in a dose-related fashion. This anti-hyperalgesic effect was mimicked by SFLLRN, the natural ligand at PAR-1 binding sites, but not SLIGRL-amide, a PAR-2 agonist. The effects of SFLLRN were less potent and shorter in duration than that of thrombin, consistent with its more transient effect on PAR-1 sites. Both thrombin and SFLLRN inhibited acetic acid-induced abdominal stretch (writhing) behaviours, which were also sensitive to NMDA antagonism, but not hot plate or tail flick latencies, which were insensitive to NMDA antagonists. TFLLR-amide, a selective ligand for PAR-1 sites, mimicked the effects of thrombin while RLLFT-amide, an inactive, reverse peptide sequence, did not. In addition, the effect of TFLLR-amide was prevented by RWJ-56110, a PAR-1 antagonist. Thrombin and TFLLR-amide produced no oedema (Evans Blue extravasation) in the spinal cord that would account for these effects. Based on the reported ability of thrombin to mobilize endothelin-1 from astrocytes, we tested the role of this compound in thrombin's activity. BQ123, an endothelin A receptor antagonist, prevented thrombin's inhibition of writhing and NMDA-induced behaviours while BQ788, an endothelin B receptor antagonist, did not. Thus, activation of PAR-1 sites by thrombin in the CNS appears to inhibit NMDA-mediated nociception by a pathway involving endothelin type A receptors.

Figure 1. Effect of thrombin on NMDA-induced behaviours

A, dose-related increase in the number of biting and scratching behaviours induced by increasing concentrations of NMDA injected i.t.; B, dose-related inhibition of NMDA-induced behaviours (5 μl of a 20 μm solution; i.e. a 100 pmol dose) by i.t. injections of concentrations ranging from 3 pm to 30 nm thrombin (60 min pretreatment), and the ability of a higher dose of NMDA (40 μm) to overcome this inhibitory effect; C, the behavioural response to NMDA (8 μm) or saline i.t., 5, 30, 60 or 120 min after an i.t. injection of thrombin (30 nm) or saline. Caudally directed biting and scratching behaviours were counted for exactly 1 min (A) or 2 min (B and C) beginning immediately after administration of NMDA. Each point represents the mean (± s.e.m.) number of behaviours obtained from an independent group containing four to nine mice as follows: A, n = 5–6; B, n = 4–9; C, n = 6. *, ** and *** indicate P < 0.05, P < 0.01 and P < 0.001, respectively, versus the corresponding saline-injected control group (B) or treatment groups, as reflected by values obtained between 5 and 120 min after injection (C), as analysed by ANOVA followed by Newman-Keuls multiple comparisons. # indicates P < 0.05versus 20 μm NMDA plus 3 pm thrombin group, as analysed by Student's t test.

Figure 2. Inhibition of NMDA-induced behaviours by SFLLRN

NMDA (40 pmol) was either co-administered with different concentrations of SFLLRN or SLIGRL-amide (A) or injected 5 min after pretreatment with TFLLR-amide or RLLFT-amide (B). Each value represents the average number of biting and scratching behaviours, expressed as a percentage of control (± s.e.m.), for an independent group composed of the following number of mice: SLIGRL-amide and RLLFT-amide, n = 5; SFLLRN, n = 4; and TFLLR-amide, n = 5–6. Statistical analysis prior to transformation of data revealed differences as indicated by *P < 0.05 and ***P < 0.001, compared with their corresponding control groups, as analysed by ANOVA, followed by Newman-Keuls multiple comparisons.

Figure 5. TFLLR-amide, but not RLLFT-amide, negates the hyperalgesic effect of NMDA

Von Frey tests with fibres of different sizes (A, #2.83; B, #4.08; C, #5.07) were performed before and 5 min after injections of either NMDA (8 μm) and saline, TFLLR-amide (100 μm) and NMDA (8 μm), or RLLFT-amide (100 μm) and NMDA (8 μm). Values represent the mean difference (± s.e.m.) in the number of responses to von Frey fibres obtained 5 min after injection compared with immediately prior to injection, such that increases reflect increased nociceptive responses or hyperalgesia. Each value represents a group of six mice. Statistically significant differences in these values compared with those obtained in saline-injected control mice (open bars) are indicated by *P < 0.05 and **P < 0.01, as analysed by ANOVA followed by Newman-Keuls multiple comparisons. # indicates a significant difference (P < 0.05) between the groups indicated, as analysed by ANOVA and followed by Newman-Keuls multiple comparisons.

Figure 4. Inhibition of NMDA-induced mechanical hyperalgesia by SFLLRN

Von Frey tests with fibres of different sizes (A, #2.83; B, #4.08; C, #5.07) were performed 5 min after co-administration of NMDA with different concentrations of SFLLRN. Values represent the mean (± s.e.m.) number of responses to von Frey fibres in a group composed of five mice (100 μm SFLLRN) and all others in groups of four mice. The value for SFLLRN (100 μm) alone (□) is indistinguishable from that of SFLLRN + NMDA (•) due to overlapping means. Inhibition of NMDA-induced hyperalgesia by SFLLRN, compared with saline-injected control groups, is indicated by *P < 0.05 as analysed by ANOVA, followed by Newman-Keuls multiple comparisons.

Figure 3. Effect of pretreatment with thrombin on NMDA-induced mechanical hyperalgesia

Von Frey tests with fibres of different sizes (A, #2.83; B, #4.08; C, #5.07) were performed at the times indicated relative to thrombin or saline and 5 min after the i.t. injection of NMDA or saline. Values represent the mean (± s.e.m.) number of positive responses out of a total of 10 trials on the plantar surface of both hind paws. Each group represents a value obtained from a group of five mice. **P < 0.01 and ***P < 0.001 reflect differences between treatment groups (over 5–120 min), as analysed by ANOVA followed by Newman-Keuls multiple comparisons.

Figure 6. Effects of APV, CPP, thrombin or SFLLRN on nociceptive and NMDA-induced behaviours

Bars represent either the mean (± s.e.m.) number of behaviours produced by an i.p. injection of acetic acid (abdominal contractions or ‘stretching’ behaviours) or an i.t. injection of NMDA (biting and scratching behaviours), or the mean (± s.e.m.) latency of response in the hot plate (53 °C) or tail flick assay (53 °C) obtained from independent groups of five to eight mice. A, nociceptive assays were performed 5–15 min after the i.t. injection of either APV or CPP, two competitive antagonists at NMDA receptors, or saline. B, nociception was measured 5–15 min after the i.t. injection of either thrombin or SFLLRN. ** and *** indicate P < 0.01 and P < 0.001, respectively, compared with its corresponding saline control group, as analysed by ANOVA and followed by Newman-Keuls multiple comparisons.

Figure 7. Effects of thrombin and SFLLRN on acetic acid-induced behaviours

Each value represents the mean number of acetic acid-induced stretching behaviours, expressed as a percentage (± s.e.m.) of control from an independent group of mice as follows (in ascending order of concentration): A, n = 6–8; B, n = 4–10; C, n = 4–8; D, n = 6 for TFLLR-amide and RLLFT-amide, n = 8–9 for SFLLRN. A and C, time course of responses following thrombin (3 nm) and SFLLRN (100 μm), respectively. B and D, dose–response relationships produced by the injection of thrombin (1 h), SFLLRN (5 min), TFLLR-amide (5 min) or RLLFT-amide (5 min), respectively. Inhibition of behaviours compared with corresponding saline-injected control groups (100 %), calculated prior to transformation, are indicated by *P < 0.05, **P < 0.01 and ***P < 0.001, respectively, as analysed by ANOVA and followed by Newman-Keuls multiple comparisons.

Figure 8. Prevention of thrombin-induced effects by PPACK and PAR-1-induced effects by RWJ-56110

A, acetic acid- or NMDA-induced behaviours were examined 1 h after pretreatment with PPACK, thrombin or saline (co-administration). B, acetic acid- or NMDA-induced behaviours were also examined after pretreatment with RWJ-56110 (5 min), TFLLR-amide (5 min) or saline. Bars represent the mean (± s.e.m.) intensity of behaviours from groups of five to eight mice (A) and five to six mice (B). * and *** indicate P < 0.05 and P < 0.001, respectively, compared with the corresponding saline control group; #, ## and ### indicate P < 0.05, P < 0.01 and P < 0.001, respectively, between the groups indicated, as analysed by ANOVA and followed by Newman-Keuls multiple comparisons.

Figure 9. BQ123, but not BQ788, inhibits behavioural effects of thrombin

The number of NMDA-induced biting and scratching behaviours (A) or acetic acid-induced writhing behaviours (B) was measured in mice pretreated (10 min) with either saline or thrombin (30 nm). The effect of pretreatment (immediately prior to thrombin) with either BQ123 (10 μl of the solutions indicated), an endothelin receptor A antagonist, or BQ788 (10 μl of a 20 μm solution), an endothelin receptor B antagonist, on the inhibitory effect of thrombin was compared with saline-pretreated control mice. Each value represents the mean number of behaviours from an independent group of mice as follows: A, n = 4–8; B, n = 4–12. Inhibition of behaviours compared with corresponding saline-injected control groups (white bars) are indicated by **P < 0.01 and ***P < 0.001, respectively, as analysed by ANOVA and followed by Newman-Keuls multiple comparisons; # and ### further indicate P < 0.05 and P < 0.001, respectively, between groups as indicated.