Varying Analgesic Effectiveness of Systemic and Central Intrathecal Administration of Cyclooxygenase-2 Inhibitors in Different Phases of Osteoarthritic Pain in Rats

Abstract

Background: Osteoarthritis (OA) contributes to heightened pain perception by disrupting the normal function of peripheral nerves and spinal nociceptive circuits. Although selective cyclooxygenase-2 (COX-2) inhibitors reduce OA-associated pain, the distinct roles of spinal COX-2 and glial cell activity in this context remain poorly defined.

Methods: The effects of two COX-2 inhibitors, etoricoxib and celecoxib, were examined using an anterior cruciate ligament transection (ACLT) rat model of OA. Mechanical allodynia was assessed behaviorally using the von Frey filament test. COX-2 protein expression and glial cell (astrocyte and microglia) activation in the lumbar spinal cord were analysed via immunohistochemistry.

Results: Spinal COX-2 expression was significantly increased, mainly in neurons and astrocytes, at the 16th week after ACLT (p = 0.026). Microglia and astrocytes were activated from the 2nd to 16th week and from the 6th to 16th week after ACLT, respectively. The intrathecal median effective dose (ED₅₀) of COX-2 inhibitors, etoricoxib and celecoxib, required for reducing mechanical allodynia was lower at the 16th week than at the 2nd week after ACLT surgery (p = 0.0448 and 0.046, respectively). In contrast, the oral ED₅₀ values of etoricoxib and celecoxib for relieving mechanical allodynia were slightly higher at the 16th week than at the 2nd week after ACLT surgery (p = 0.097 and 0.227, respectively).

Conclusions: Our study shows that the efficacy of COX-2 inhibitors in ALCT-induced OA rats depends on the timing and route of administration. In the later phase, spinal glia cells exhibited increased activity and elevated COX-2 expression.

Significance: This study characterises the complex mechanisms underlying OA pain, involving both peripheral and central components, and highlights the stage-specific involvement of COX-2, particularly in the spinal cord. It provides experimental evidence linking central COX-2 activity and glial cell responses to OA pain, offering insights into the temporal dynamics of pain processing and guiding the development of therapeutic strategies.

Keywords: anterior cruciate ligament transection; celecoxib; cyclooxygenase‐2; etoricoxib; mechanical allodynia; osteoarthritis.

FIGURE 1

Time course of anterior cruciate ligament transection (ACLT)‐induced mechanical allodynia in rats. Mechanical allodynia was detected through the paw withdrawal threshold (PWT) every week after ACLT. *p < 0.05, compared with the sham group. Data are presented as mean ± SEM.

FIGURE 2

Anterior cruciate ligament transection (ACLT) activates the spinal microglia in rats. The lumbar spinal cord of rats in the sham and ACLT groups (2, 6, and 16 weeks of age) was harvested after ACLT. (a) A microglia‐specific marker (OX‐42, green) was used to visualise microglia in the spinal cord. High‐magnification images of OX‐42 expression in the ipsilateral and contralateral spinal dorsal horns are presented in the inset. (b) Quantification of OX‐42 in the ipsilateral and contralateral spinal dorsal horns, shown as means ± SEM. Spinal OX‐42 was upregulated by ACLT and expressed primarily in the ipsilateral spinal dorsal horn, especially at Week 6. Scale bars: 100 μm (a); 10 μm in the inset. *p < 0.05, compared with the sham group.

FIGURE 3

Anterior cruciate ligament transection (ACLT) activates the spinal astrocytes in rats. The lumbar spinal cord of rats in the sham and ACLT groups (2, 6, and 16 weeks of age) was harvested after ACLT. (a) An astrocyte‐specific marker (GFAP, green) was used to visualise the astrocytes in the spinal cord. High‐magnification images of GFAP expression in the ipsilateral and contralateral spinal dorsal horns are presented in the inset. (b) Quantification of GFAP in the ipsilateral and contralateral spinal dorsal horns is shown as means ± SEM. Spinal GFAP was upregulated by ACLT and expressed primarily in the ipsilateral spinal dorsal horn, especially at Weeks 6 and 16. Scale bars: 100 μm (a); 10 μm in the inset. *p < 0.05, compared with the sham group.

FIGURE 4

Spinal COX‐2 expression in the rats subjected to anterior cruciate ligament transection (ACLT). The lumbar spinal cord of rats in the sham and ACLT groups (2, 6, and 16 weeks of age) was harvested after ACLT. (a) COX‐2 expression is presented in the ipsilateral and contralateral spinal cord dorsal horns in the sham, 2w, 6w and 16w groups. (b) Quantification of COX‐2 expression in the ipsilateral and contralateral spinal dorsal horns is shown as means ± SEM. ACLT significantly upregulated COX‐2 immunoreactivity on the ipsilateral side at 16 weeks after surgery. Scale bar: 100 μm. *p < 0.05, compared with the sham group.

FIGURE 5

Spinal COX‐2 is expressed in specific cells in the late phase after anterior cruciate ligament transection (ACLT). The cellular specificity of COX‐2 expression in the dorsal region of the lumbar spinal cord dorsal horn was ipsilateral to the injury site after ACLT surgery in the 16th week. Double‐immunofluorescence staining of COX‐2 (red) with NeuN (a neuron‐specific marker, green), OX‐42 (a microglia‐specific marker, green), or GFAP (an astrocyte‐specific marker, green) in the spinal cord from rats in the sham and 16w groups. The merged images show that COX‐2 expression was lower in the sham‐operated group, but its was highly expressed in the 16w group and colocalized with the three cell‐specific markers. Spinal COX‐2 was primarily colocalized with astrocytes rather than with neurons and microglia in ACLT rats in the 16th week. Scale bar: 50 μm.

FIGURE 6

Effect of intrathecal administration of etoricoxib on antinociception during the early and late phases in rats subjected to anterior cruciate ligament transection (ACLT). Time course effects of the % maximum possible effect (MPE) of anti‐mechanical allodynia for intrathecal injection of etoricoxib in ACLT‐rats at 2 (a) and 16 (b) weeks after ACLT. In (a, b), the horizontal axis represents the time in minutes after intrathecal injection, and the vertical axis represents the %MPE, calculated as the mean for six animals per dose. Data in (a, b) were transformed to a dose–response curve of etoricoxib for anti‐mechanical allodynia (c). The ED₅₀ values of etoricoxib at 2 and 16 weeks after surgery were 4.36 ± 1.16 and 1.49 ± 0.47 μg, respectively. The anti‐mechanical allodynia effect of intrathecal etoricoxib administration was higher at 16 weeks after surgery than at 2 weeks after surgery. *p < 0.05, compared with the ACLT + vehicle group.

FIGURE 7

Effect of intrathecal administration of celecoxib on antinociception during the early phase and late phases in rats subjected to anterior cruciate ligament transection (ACLT). Time course effects of the % maximum possible effect (MPE) of anti‐mechanical allodynia for intrathecal injection of celecoxib in ACLT‐rats 2 (a) and 16 (b) weeks after ACLT. In (a, b), the horizontal axis represents the time in minutes after intrathecal injection, and the vertical axis represents the %MPE, calculated as the mean for six animals per dose. Data in (a, b) were transformed to a dose–response curve of etoricoxib for anti‐mechanical allodynia (c). The ED₅₀ values of etoricoxib at 2 and 16 weeks after surgery were 9.60 ± 5.04 and 2.55 ± 0.54 μg, respectively. The anti‐mechanical allodynia effect of intrathecal celecoxib administration was higher at 16 weeks after surgery than at 2 weeks after surgery. *p < 0.05, compared with the ACLT + vehicle group.

FIGURE 8

Effect of oral administration of etoricoxib on nociception during the early and late phases in rats subjected to anterior cruciate ligament transection (ACLT). Time course effects of the % maximum possible effect (MPE) of anti‐mechanical allodynia for oral administration of etoricoxib in ACLT‐rats at 2 (a) and 16 (b) weeks after ACLT. In (a, b), the horizontal axis represents the time in minutes after oral administration, and the vertical axis represents the %MPE, calculated as the mean for six animals per dose. Data in (a, b) were transformed to a dose–response curve of etoricoxib for anti‐mechanical allodynia (c). The ED₅₀ values of etoricoxib at 2 and 16 weeks after surgery were 1.55 ± 0.2 and 2.69 ± 0.59 mg, respectively. The anti‐mechanical allodynia effect of oral etoricoxib administration was higher at 16 weeks after surgery than at 2 weeks after surgery. *p < 0.05, compared with the ACLT + vehicle group.

FIGURE 9

Effect of oral administration of celecoxib on nociception during the early phase and the late phase in rats subjected to anterior cruciate ligament transection (ACLT). Time course of the % maximum possible effect (MPE) of anti‐mechanical allodynia for oral administration of celecoxib in ACLT‐rats at 2 (a) and 16 (b) weeks after ACLT. In (a, b), the horizontal axis represents the time in minutes after oral administration, and the vertical axis represents the %MPE, calculated as the mean for six animals per dose. Data in (a, b) were transformed to a dose–response curve of celecoxib for anti‐mechanical allodynia (c). The ED₅₀ values of celecoxib at 2 and 16 weeks after surgery were 2.17 ± 0.82 and 5.17 ± 2.48 mg, respectively. The anti‐mechanical allodynia effect of oral celecoxib administration was higher at 16 weeks after surgery than at 2 weeks after surgery. *p < 0.05, compared with the ACLT + vehicle group.