Neuropathic pain activates the endogenous kappa opioid system in mouse spinal cord and induces opioid receptor tolerance

Abstract

Release of endogenous dynorphin opioids within the spinal cord after partial sciatic nerve ligation (pSNL) is known to contribute to the neuropathic pain processes. Using a phosphoselective antibody [kappa opioid receptor (KOR-P)] able to detect the serine 369 phosphorylated form of the KOR, we determined possible sites of dynorphin action within the spinal cord after pSNL. KOR-P immunoreactivity (IR) was markedly increased in the L4-L5 spinal dorsal horn of wild-type C57BL/6 mice (7-21 d) after lesion, but not in mice pretreated with the KOR antagonist nor-binaltorphimine (norBNI). In addition, knock-out mice lacking prodynorphin, KOR, or G-protein receptor kinase 3 (GRK3) did not show significant increases in KOR-P IR after pSNL. KOR-P IR was colocalized in both GABAergic neurons and GFAP-positive astrocytes in both ipsilateral and contralateral spinal dorsal horn. Consistent with sustained opioid release, KOR knock-out mice developed significantly increased tactile allodynia and thermal hyperalgesia in both the early (first week) and late (third week) interval after lesion. Similarly, mice pretreated with norBNI showed enhanced hyperalgesia and allodynia during the 3 weeks after pSNL. Because sustained activation of opioid receptors might induce tolerance, we measured the antinociceptive effect of the kappa agonist U50,488 using radiant heat applied to the ipsilateral hindpaw, and we found that agonist potency was significantly decreased 7 d after pSNL. In contrast, neither prodynorphin nor GRK3 knock-out mice showed U50,488 tolerance after pSNL. These findings suggest that pSNL induced a sustained release of endogenous prodynorphin-derived opioid peptides that activated an anti-nociceptive KOR system in mouse spinal cord. Thus, endogenous dynorphin had both pronociceptive and antinociceptive actions after nerve injury and induced GRK3-mediated opioid tolerance.

Figure 2.

Response thresholds of the ipsilateral (A) and contralateral (B) hindpaw to usually innocuous tactile (von Frey hair) stimuli in wild-type (WT) mice (n = 16), KOR (-/-) mice (n = 6), and prodynorphin (-/-) mice after pSNL (n = 10) during 21 d after partial sciatic nerve ligation. A, After pSNL, both wild-type and prodynorphin (-/-) mice developed significantly (^*p<0.05) greater allodynia on the ipsilateral side as evident from the decreased thresholds to tactile stimulation compared with mice receiving sham-ligation surgeries (baseline). One week after pSNL, prodynorphin (-/-) mice showed significantly less allodynia compared with WT mice after pSNL. KOR (-/-) mice developed significantly more allodynia when compared with WT mice after pSNL throughout the test period. On day 21 after pSNL, norBNI (10 mg/kg, i.p.) administered to wild-type mice significantly decreased response threshold (^#p < 0.05). In contrast, KOR (-/-) mice and prodynorphin (-/-) mice were not significantly affected by norBNI treatment. B, The contralateral hindpaw of wild-type mice showed a significant (p < 0.05) decrease in response threshold during the 21 d after pSNL compared with sham-ligated mice (baseline). The allodynic response of KOR (-/-) mice was even greater than that of wild-type mice as evident from the significantly lower response thresholds. Prodynorphin (-/-) mice after pSNL showed significantly less allodynia compared with WT mice after pSNL.

Figure 1.

Response latencies of the ipsilateral (A) and contralateral (B) hindpaws to noxious thermal (radiant heat applied to the hindpaw) stimuli for wild-type (WT) mice (n = 16), KOR (-/-) mice (n = 6), and prodynorphin (-/-) mice (n = 10) during 21 d after pSNL. A, After pSNL, both WT and knock-out mice developed significant hyperalgesia as evident from the significantly decreased thermal latencies compared with before pSNL. WT mice receiving sham ligation did not significantly change response latencies from baseline when tested on the same repeated measures schedule (data not shown). KOR (-/-) mice developed significantly greater hyperalgesia when compared with WT mice after pSNL. One week after pSNL, prodynorphin (-/-) mice showed significantly less hyperalgesia compared with WT mice. At day 21, administration of norBNI (10 mg/kg, i.p.) significantly reduced the response latency of WT mice (p < 0.05) but did not affect the responses of KOR (-/-) mice or prodynorphin (-/-) mice after pSNL. B, Except on the first day after pSNL, the contralateral hindpaw of neither WT nor prodynorphin (-/-) mice showed any reduction in thermallatency. The contralateral hindpaw of the KOR(-/-) mice showed significantly more hyperalgesia compared with WT mice after pSNL. Data represent means ± SEM; ^*p < 0.05, using one-way ANOVA followed by either Tukey or t test as appropriate.

Figure 3.

κ Opioid receptor activation increased the intensity of KOR-P Ab staining in the spinal cord dorsal horn. Injection of 50 mg/kg U50,488 30 min before perfusion caused a significant increase in KOR-P IR (A, D) compared with that of saline injection (B, E) in the dorsal horn of the mouse lumbar spinal cord. This increase in KOR-P Ab staining was blocked by 10 mg/kg norBNI pretreatment (C,F) in wild-type mice and was not evident in mice lacking κ opioid receptors (KOR-/-) (G) or GRK3 (GRK-/-) (H). Negative controls showed the lack of specific staining in the absence of primary Ab (I). Scale bars: A-C, G-I, 400 μm; D-F, 200 μm. The results shown are representative images taken from more than three independent replications.

Figure 4.

pSNL resulted in an increased intensity of KOR-P Ab staining in the lumbar spinal cord dorsal horn in wild-type mice. All mice were perfused and fixed on day 7 after pSNL or sham-ligation surgery. A substantial increase in immunofluorescence intensity of KOR-P Ab staining was evident in wild-type pSNL mice (A, B) but not in sham-ligated mice (C, D) in the lumbar spinal cord, mainly in laminas I-III and X. No apparent differences were evident between the ipsilateral and contralateral sides (A). B shows a higher-power image of the same dorsal horn as in A. The increase in KOR-P staining was blocked by 10 mg/kg norBNI pretreatment before pSNL (E, F) and was not evident in mice lacking prodynorphin (proDYN-/-) (G) or GRK 3 (GRK-/-) (H). Scale bars: A, C, E, G, H, 400 μm; B, D, F, 200 μm. The results shown are representative images taken from more than three independent replications.

Figure 5.

Comparison of immunolocalization of KOR-P and GABA IR in the dorsal horn of wild-type mouse lumbar spinal cord on day 7 after pSNL shows partial overlap. Representative sections show that KOR-P and GABA Ab staining were mainly in laminas I-III (A, B). KOR-P (A, D, green) and GABA immunoreactivity (B, E, red) colocalized in neurons evident in the dorsal horn (C, F, yellow). These results suggest GABAergic neurons expressed KOR that was phosphorylated after pSNL. Bottom panels (D-F) were higher magnification of the same confocal images shown in the top panels (A-C).

Figure 6.

Comparison of immunolocalization of KOR-P and GFAP IR in the dorsal horn of wild-type mouse lumbar spinal cord on day 7 after pSNL shows partial overlap. Representative images show that KOR-P and GFAP Ab staining were mainly in laminas I-III (A, B). KOR-P (A, D, green) and GFAP immunoreactivity (B, E, red) were coexpressed in the dorsal horn (C, F, yellow). Bottom panel is higher-magnification confocal images of the same images shown in the top panel.

Figure 7.

Dose-response curves showing the antinociceptive effect of theκ agonist U50,488 in wild type (WT), prodynorphin (-/-), and GRK3 (-/-) mice after sham-ligation surgery or after pSNL in the Hargreave's radiant heat test. Different doses of U50,488 (5, 10, 25, 50, or 75 mg/kg) were administered intraperitoneally to mice 30 min before their paw withdrawal response latencies were measured. Maximum possible effect (%MPE) was calculated as 100 × (latency-baseline)/(cutoff-baseline); cutoff was 15 sec in this experiment. A, U50,488 was significantly less effective in wild-type mice after pSNL (p < 0.05) than in WT mice after sham-ligation surgery (n = 10 for each group). B, pSNL did not affect the U50,488 dose-response curves for either the prodynorphin (-/-) or the GRK3 (-/-) mice (p>0.05) (n=5 for each group);however, the prodynorphin (-/-) mice were significantly less sensitive than the GRK3 (-/-) mice to U50,488. C, Response latencies of the ipsilateral hindpaw to noxious thermal (radiant heat applied to the hindpaw) stimuli in wild-type (WT) mice after pSNL (n = 16) and in GRK3 (-/-) mice after pSNL (n = 5). GRK3 (-/-) mice after pSNL developed partially significantly less hyperalgesia compared with wild-type mice after pSNL. At day 21, norBNI(10 mg/kg, i.p.) was injected in WT and GRK3 (-/-) mice after pSNL. GRK3 (-/-) mice after pSNL developed a significant decrease in latency (^#p < 0.05). Consistent with results shown in Figure 1 A, wild-type mice showed a significant decrease in response latency as well. Results are shown to illustrate that GRK3 (-/-) and prodynorphin (-/-) mice both show a decrease in response latencies after pSNL. Thus, the shift in U50,488 EC₅₀ shown in A cannot result simply from a stronger nociceptive stimulus.