A proportion of diabetic patients experience chronic pain that is frequently refractory to treatment with conventional analgesics. Experimentally diabetic animals exhibit a range of behavioral abnormalities indicative of pain, and can be used to model human painful diabetic neuropathy. Studies suggest that pain during diabetes is not associated with structural damage to the peripheral nerves, and biochemical amplification of nociceptive processing may be of importance. The formalin test is used to measure spinal sensitization in animals. Diabetic rats exhibit exaggerated flinching during this test, in addition to increased constitutive spinal expression of cyclooxygenase-2 (COX-2) protein. Studies were designed to establish causality between increased spinal COX-2 activity and formalin hyperalgesia in diabetic rats. Rats with one, two or four weeks of streptozotocin-induced diabetes exhibited significantly increased levels of spinal COX-2 protein and activity, along with exaggerated paw flinching in response to 0.5% paw formalin injection. Increased flinching of diabetic rats was attenuated by intrathecal pre-treatment with a selective COX-2 inhibitor immediately prior to formalin injection, confirming the involvement of COX-2 activity in formalin hyperalgesia observed in diabetic rats. We hypothesized that glucose flux through the polyol pathway in the central nervous system (CNS) causes elevated spinal COX-2 expression and formalin hyperalgesia in diabetic rats. Chronic treatment with insulin or ICI222155, an aldose reductase inhibitor (ARI) previously shown to cross the blood-brain barrier (BBB) and prevent formalin-evoked hyperalgesia in diabetic rats, prevented elevated spinal COX-2 protein and activity in diabetic rats. In contrast, the ARI IDD676 showed efficacy against AR activity only in the peripheral nervous system, and had no effect on elevated spinal COX-2 or hyperalgesia to paw formalin injection, despite demonstrating efficacy against indices of peripheral nerve dysfunction. In the spinal cord, aldose reductase (AR) expression was observed solely in oligodendrocytes, which also expressed COX-2. These data suggest that activity of AR in spinal oligodendrocytes may be a primary initiating factor linking hyperglycemia with hyperalgesia in diabetic rats. Further, CNS targeting of ARIs may be important in treating painful diabetic neuropathy in patients