Dermatomal scratching after intramedullary quisqualate injection: correlation with cutaneous denervation

Abstract

Central nervous system lesions cause peripheral dysfunctions currently attributed to central cell death that compromises function of intact peripheral nerves. Injecting quisqualate (QUIS) into the rat spinal cord models spinal cord injury (SCI) and causes at-level scratching and self-injury. Such overgrooming was interpreted to model pain until patients with self-injurious scratching after SCI reported itch motivated scratching that was painless because of sensory loss. Because self-injurious scratching is difficult to explain by central mechanisms alone, we hypothesized that QUIS injections damage peripheral axons of at-level afferents. QUIS was injected into thoracic spinal cords of 18 Long-Evans rats. Animals were killed 3 days after overgrooming began or 14 days after injection. Spinal cord lesions were localized and DRG-immunolabeled for ATF-3. At-level and control skin samples were PGP9.5-immunabeled to quantify axons. Eighty-four percent of QUIS rats overgroomed. Skin in these regions had lost two-thirds of epidermal innervation as compared with controls (P < .001). Rats that overgroomed had 47% less axon-length than nongrooming rats (P = .006). The presence of ATF-3 immunolabeled neurons within diagnosis-related groups of QUIS rats indicated death of afferent cell bodies. Overgrooming after QUIS injections may not be due entirely to central changes. As in humans, self-injurious neuropathic scratching appeared to require loss of protective pain sensations in addition to peripheral denervation.

Perspective: This study suggests that intramedullary injection of quisqualic acid in rats causes death of at-level peripheral as well as central neurons. Self-injurious dermatomal scratching that develops in spinal-injured rats may reflect neuropathic itch and loss of protective pain sensations.

Figure 1

A, Cresyl violet staining of spinal cord 9 days after intramedullary injection of quisqualate (QUIS) into the right dorsal gray matter at approximate spinal level T13. Cell loss within laminae III–V is evident on both sides of the cord. B, Photograph of a QUIS-injected rat 5 days after onset of dermatomal overgrooming.

Figure 2

Photomicrographs showing PGP95⁺ fibers of sampled skin tissues (A–C; arrowhead) and ATF3⁺ cells from an ipsilesional diagnosis-related group (D; arrow) 2 weeks after quisqualate (QUIS) injection. Dense PGP95-immunoreactive nerve fibers were observed in the epidermal and dermal layers from control skin (A). Nongroomed QUIS-injured skin contained fewer PGP95-immunoreactive nerve fibers (C), and groomed QUIS-lesioned skin contained almost no PGP95-immunoreactive nerve fibers (B). Presence of ATF3+ cells were observed from a diagnosis-related group ipsilesional to grooming. Epi, epidermis; Der, dermis.

Figure 3

Mean axonal length in skin taken from unoperated rats (n = 11), rats injected with quisqualate (QUIS) with no signs of gross pathology on the side ipsilateral to skin biopsy (n = 5), rats injected with QUIS with obvious pathology on the side ipsilateral to skin biopsy (n = 20), ungroomed regions after QUIS injection (n = 7), and skin targeted for overgrooming after QUIS injection (n = 13). All skin taken from QUIS-injected animals had reduced neurite density compared with unoperated controls. Skin that was overgroomed after QUIS injection had significantly reduced neurite density compared with ungroomed skin from QUIS-injected animals. *Significant difference from unoperated; ^#Significant difference from ungroomed skin (P < .01).