Significant determinants of mouse pain behaviour

Abstract

Transgenic mouse behavioural analysis has furthered our understanding of the molecular and cellular mechanisms underlying damage sensing and pain. However, it is not unusual for conflicting data on the pain phenotypes of knockout mice to be generated by reputable groups. Here we focus on some technical aspects of measuring mouse pain behaviour that are often overlooked, which may help explain discrepancies in the pain literature. We examined touch perception using von Frey hairs and mechanical pain thresholds using the Randall-Selitto test. Thermal pain thresholds were measured using the Hargreaves apparatus and a thermal place preference test. Sodium channel Nav1.7 knockout mice show a mechanical deficit in the hairy skin, but not the paw, whilst shaving the abdominal hair abolished this phenotype. Nav1.7, Nav1.8 and Nav1.9 knockout mice show deficits in noxious mechanosensation in the tail, but not the paw. TRPA1 knockout mice, however, have a loss of noxious mechanosensation in the paw but not the tail. Studies of heat and cold sensitivity also show variability depending on the intensity of the stimulus. Deleting Nav1.7, Nav1.8 or Nav1.9 in Nav1.8-positive sensory neurons attenuates responses to slow noxious heat ramps, whilst responses to fast noxious heat ramps are only reduced when Nav1.7 is lost in large diameter sensory neurons. Deleting Nav1.7 from all sensory neurons attenuates responses to noxious cooling but not extreme cold. Finally, circadian rhythms dramatically influence behavioural outcome measures such as von Frey responses, which change by 80% over the day. These observations demonstrate that fully characterising the phenotype of a transgenic mouse strain requires a range of behavioural pain models. Failure to conduct behavioural tests at different anatomical locations, stimulus intensities, and at different points in the circadian cycle may lead to a pain behavioural phenotype being misinterpreted, or missed altogether.

Figure 1. Comparison of different transgenic mice reveals test-site and stimulus-intensity specific mechanosensory responses.

Nav1.7^Nav1.8 mice (blue columns, n = 7), Nav1.7^Advill mice (red column, n = 9) and Nav1.7^Wnt1 mice (green column, n = 9) mice show normal responses to von Frey hairs applied using either the up-down method (a) or the repeated measures method in comparison to littermate mice (white columns, n = 36). (b). Both Nav1.7^Advill mice (n = 9) and Nav1.7^Wnt1 mice (n = 9) show a behavioural deficit in response to the abdominal von Frey test in comparison to Nav1.7^Nav1.8 mice (n = 7) and littermate mice (n = 36) (c). The abdomens of C57BL/6 (n = 12) mice are significantly more sensitive than the plantar surface of the hindpaw (d), which is loss if the abdomen is shaved (e). Shaving the abdominal hair attenuates the sensitivity to von Frey hair stimulation of Nav1.7^Nav1.8 (n = 10) and littermate mice (n = 21) but has no effect of Nav1.7^Advill (n = 7) or Nav1.7^Wnt1 mice (n = 11) (f). Nav1.7^Nav1.8 (n = 14), Nav1.7^Advill (n = 8) Nav1.7^Wnt1 (n = 9) show a significant increase withdrawal threshold in response to the Randall-Siletto test when applied to the tail but not the paw when compared to littermate (n = 26) mice (g). Nav1.8KO (light blue column, n = 11) and Nav1.9KO (turquoise column, n = 8) but not Nav1.3KO (yellow column, n = 6) show a significant increase withdrawal threshold in response to the Randall-Siletto test when applied to the tail when compared to littermate (n = 27) mice, however no difference is seen when applied to the paw (h). TRPA1 KO mice (pink columns, n = 8) show a behavioural deficit to Randall-Selitto test applied to the paw but not tail in comparison to littermate mice (white columns, n = 8) (i). Data analysed by two-way analysis of variance followed by a Bonferroni post-hoc test. Results are presented as mean ± S.E.M. ** P<0.01 and *** P<0.001 (individual points).

Figure 2. The DRG innervating the hindpaw and tail consist of different ratios of neuronal subpopulations.

Example section of an L4 (a) and an S1 (b) DRG (N52: green, Nav1.8: red, scale bar = 250 µm). Overall percentage of estimated number of N52, Nav1.8 and double stained cells within L4 (n = 52), L5 (n = 43), L6 (n = 32), S1 (n = 18) and S2 (n = 17) DRG (c). Total estimated number of N52, Nav1.8 and double-stained cells within L4 (n = 52), L5 (n = 43), L6 (n = 32), S1 (n = 18) and S2 (n = 17) DRG (d). All data analysed by two-way analysis of variance followed by a Bonferroni post-hoc test. Results are presented as mean ± S.E.M. ** P<0.01 and *** P<0.001 (individual points).

Figure 3. Comparison of different transgenic mice reveals stimulus-intensity specific responses to noxious thermal stimuli.

Behavioural responses of different Nav1.7 tissue-specific knockouts to the Hargreaves test applied to the hindpaw. (a) Nav1.7^Nav1.8 mice (blue columns, n = 14), Nav1.7^Advill mice (red column, n = 7) and Nav1.7^Wnt1 mice (green column, n = 12) all show a behavioural deficit in response to the Hargreaves test at a heat ramp of 0.6°C.s⁻¹ in comparison to littermate mice (white columns, n = 27), however only Nav1.7^Advill and Nav1.7^Wnt1 mice show a behavioural deficit in response to the Hargreaves test at a heat ramp of 2.0°C.s⁻¹. (b) Nav1.8KO mice (light blue column, n = 6) and Nav1.9KO mice (turquoise column, n = 10) but not Nav1.3KO mice (orange column, n = 6) show a significantly increased withdrawal latency to the Hargreaves test at a heat ramp of 0.6°C.s⁻¹ in comparison to littermate mice (white columns, n = 18), however this significant increase is lost the when the Hargreaves test is conducted using a heat ramp of 2.0°C.s⁻¹. Data analysed by two-way analysis of variance followed by a Bonferroni post-hoc test. Results are presented as mean ± S.E.M. * P<0.05 and *** P<0.001 (individual points).

Figure 4. Comparison of different transgenic mice reveals distinct mechanisms underlie responses to cooling and noxious cold stimuli.

(a) Nav1.7^Advill mice (red columns, n = 8) avoid the 0°C test plate to the same extent as littermate controls (white columns, n = 8) in the thermal place preference test but show a behavioural deficit in the noxious cooling range (14–12°C). (b) Nav1.8-DTA mice (black columns, n = 6) avoid the cooling stimuli to the same extent as littermate controls (white columns, n = 6) in the thermal place preference test but show a trend indicating a behavioural deficit in response to 0°C. Data analysed by two-way analysis of variance followed by a Bonferroni post-hoc test. Results are presented as mean ± S.E.M. * P<0.05 (individual points).

Figure 5. The affect of circadian rhythm on von Frey responses over a 24-hour period.

(a) Behavioural responses of C57BL/6 mice to the von Frey hairs applied to the hindpaw over a 24 h period. Measurements were taken every 4 hours starting at 07:00. (b) Behavioural responses of Nav1.8-DTA mice to the von Frey hairs applied to the hindpaw over a 24 h period. (a) Data analysed by two-way analysis of variance followed by a Bonferroni post-hoc test and (b) t-test. Results are presented as mean ± S.E.M. * P<0.05, *** P<0.001.