Peripheral nerve injury induces cannabinoid receptor 2 protein expression in rat sensory neurons.

We have localized cannabinoid receptor 2 protein in rat and mouse somatic sensory nervous system, using an antibody that recognizes mouse cannabinoid receptor 2. Little or no cannabinoid receptor 2 immunoreactivity was found in sections of naive rat or mouse dorsal root ganglia or spinal cord. This was in accord with the lack of detectable cannabinoid receptor 2 mRNA in (dorsal root ganglion) neurons by in situ hybridization experiments described in the literature. However, we could detect cannabinoid receptor 2 immunoreactivity following unilateral nerve damage-either by sciatic nerve section, or by spinal nerve ligation. It was localized to the superficial laminae of the dorsal horn of the spinal cord, ipsilateral to the nerve damage, coincident with the area of termination of damaged afferents which was marked by loss of isolectin B4 binding. This upregulation was not seen in cannabinoid receptor 2 null mice. The cannabinoid receptor 2 protein in spinal cord appeared to be expressed on sensory neuron afferent terminals as it colocalized with two markers of damaged afferents, namely growth associated protein-43 and the neuropeptide galanin. Moreover, it did not colocalize with markers of activated microglial cells (OX-42) or astroglial cells (glial fibrillary acidic protein) in rat spinal cord. In the peripheral nerve, accumulation of cannabinoid receptor 2 immunoreactivity was seen in nerve sections proximal, but not distal, to the ligation site, suggesting transport down the nerve from the cell bodies. Although convincing cannabinoid receptor 2 immunoreactivity was seen in neither uninjured nor injured dorsal root ganglion neuron cell bodies in tissue sections, expression was detectable in isolated, cultured neurons that had received a prior axotomy in vivo. This clear demonstration of CB(2) receptors on sensory neurons suggests an additional cellular target for CB(2) agonist induced analgesia, at least in neuropathic models.