In the present report, we studied the temporal changes in mechanical and heat pain thresholds in CPCP rats over a 13-week period following cast removal. After cast removal, mechanical hyperalgesia spread beyond the immobilized hindlimb with an increasing temporal delay leading away from the immobilized site. These observations are consistent with our previous results that cast immobilization induced local inflammatory changes and spontaneous pain in the immobilized hindlimb, and long-lasting mechanical hyperalgesia in the immobilized limb, contralateral hindlimb, and the tail
. The present results further demonstrate that cast immobilization can induce heat hyperalgesia in the immobilized side. Birklein et al.
 and Sieweke et al.
 reported in their clinical studies that in patients with acute pain, both heat and mechanical hyperalgesia were observed, whereas in patients with CRPS Type I, only mechanical hyperalgesia was observed. In light of this difference in clinical signs, the authors demonstrated that pain in CRPS Type I is likely related to plastic changes in the central nervous system (CNS). The overlapping features between CRPS and our CPCP rat model suggest that plastic changes in CNS play a role in the pain enhancing mechanisms of the CPCP model.
In the present experiment, we focused on pain enhancing mechanisms through activation of spinal glial cells, which have been reported as one form of plastic abnormality in the CNS
[20, 22, 23, 25, 35–37]. Studies using nerve injury models have described spinal microglia activation at the onset of hyperalgesia, whereas others have indicated that spinal astrocyte activation helps to maintain such pain
[20, 22, 23, 25, 35–37]. Reports from studies using inflammatory models have implicated activated spinal glia in bilateral pain as well
[28–30, 38]. Therefore, we performed a bilateral immunohistological examination of activated spinal microglia and astrocytes in the CPCP model in multiple segments of the lumbar and coccygeal spinal cords at three time points to match the temporal progression of the hyperalgesia after cast removal. Immunohistological analysis with OX42 and GFAP showed activation of spinal microglia 1 day after cast removal, which disappeared 5 weeks after cast removal. However, when microglia activation subsided, astrocyte activation increased. This transition in glial cell activation was observed with a time delay in the coccygeal spinal cord.
It has been shown that the activation of microglia in the spinal cord is accompanied by an increase in the number of microglial cells (proliferation) under several neuropathic pain conditions
[20, 39–41]. On the other hand, astrocyte activation is not accompanied by proliferation
[42–44]. Our present results from CPCP rats showing that spinal microglia were both activated and proliferated, whereas astrocytes were only activated, are consistent with these previous observations in neuropathic pain models. However, Tsuda et al.
 demonstrated that in the careful observation with proliferation markers, such as Ki-67 and phosphorylated-histone H3 (p-HisH3), spinal astrocyte activation was accompanied by definite proliferation in the neuropathic pain model. Considering this result, there is room for further investigation using such proliferation markers in the CPCP rats.
When L-α-AA, an inhibitor of astrocyte activation, was administered intrathecally at week 5, hyperalgesia in all of the body parts was significantly attenuated, and the activation of the lumbar spinal cord astrocytes that had been activated at this time was also attenuated. These results suggest that activated lumbar spinal cord astrocytes are involved in the maintenance of widespread mechanical hyperalgesia in CPCP rats. Using a complete Freund’s adjuvant-induced inflammation model, Gao et al.
 showed that heat hyperalgesia occurs only in the inflamed side, whereas mechanical hyperalgesia occurred bilaterally. They also demonstrated that spinal astrocyte activation was involved in mechanical hyperalgesia bilaterally. These results support our present results that astrocyte activation has an important role in the maintenance of mechanical hyperalgesia in the CPCP model. Future research should examine how astrocytes contribute to the molecular mechanisms of pain enhancement.
Lumbar spinal cord microglia were noted to be activated at the onset of mechanical hyperalgesia in the hindpaws in the CPCP model. Coccygeal cord microglia were also activated when mechanical hyperalgesia was present in the tail. With intrathecally administered L-α-AA, all sites of mechanical hyperalgesia in all of the body parts was attenuated; the extent of attenuation in tail hyperalgesia seems to remain at a lower value than the other sites. Additionally, microglial activation in the coccygeal spinal cord was maintained with intrathecal administration of L-α-AA. These results suggest that activated microglia are involved in the onset of hyperalgesia, as well as in the spread of hyperalgesia.
Several studies using nerve injury models have shown activated astrocytes participate in the maintenance of hyperalgesia
[2, 20, 22]. It has been reported that high-density glial activation was observed only in the spinal segments ipsilateral to the injured nerve, and hyperalgesia arose only in the nerve-injured side
[20, 22, 23, 25, 35–37]. The present findings demonstrate that spinal astrocyte activation in the CPCP model may not accompany clear nerve injury. Our data also indicate that immobilization-induced bilateral glial activation spread beyond the spinal cord segments innervating the immobilized limb, which suggests a different pain inducing mechanism.
The present study showed that hyperalgesia in CPCP rats involved the lumbar and coccygeal spinal segments. This widespread pain is uncharacteristic of neuropathic pain
[20, 22, 25, 35–37]. However, we should address if widespread hyperalgesia in our CPCP model is due to whole-leg casting. In other words, in the neuropathic pain model, only specific branches of the sciatic nerves are injured and only the innervating spinal segments are involved. In the CPCP model, however, the entire leg is inserted in a cast and the affected area is larger. It is possible that a larger number of innervating nerves were affected and an increased number of spinal segments were involved, and in this case, the observed phenomenon could be neuropathic. The present experiment, however, showed that ATF3-positive cell counts were small in L3, L4, and L5 DRG of CPCP rats. These results suggest that 2-week cast immobilization of one hindpaw did not cause clear nerve damage.
A recent study from our laboratory suggested the contribution of reactive oxygen species (ROS) produced by ischemia/reperfusion injury to the development of widespread hyperalgesia in CPCP rats
. It has been reported that ROS receptors expressed on microglia in the CNS contribute to neuropathic pain
[45, 46]. Considering these observations, it is reasonable to assume that ROS produced after cast removal directly activate microglia in the CNS. However, the results of that study also showed that an ipsilateral sciatic nerve block performed at the initial appearance of local inflammation (24 h after cast removal) significantly reduced CWP in the CPCP rats. Based on this observation, the results of the present study suggest that ROS sensitizes the primary sensory neurons in the immobilized hindlimb, and consequently activates spinal microglia. This hypothesis is supported by the present result showing that the pERK immunoreactivity in the ipsilateral small DRG cells was increased 1 day after cast removal, implying that the primary nociceptive neurons were activated.