[en] BACKGROUND: A major class of axon growth-repulsive molecules associated with CNS scar tissue is the family of chondroitin sulphate proteoglycans (CSPGs). Experimental spinal cord injury (SCI) has demonstrated rapid re-expression of CSPGs at and around the lesion site. The pharmacological digestion of CSPGs in such lesion models results in substantially enhanced axonal regeneration and a significant functional recovery. The potential therapeutic relevance of interfering with CSPG expression or function following experimental injuries seems clear, however, the spatio-temporal pattern of expression of individual members of the CSPG family following human spinal cord injury is only poorly defined. In the present correlative investigation, the expression pattern of CSPG family members NG2, neurocan, versican and phosphacan was studied in the human spinal cord. METHODS: An immunohistochemical investigation in post mortem samples of control and lesioned human spinal cords was performed. All patients with traumatic SCI had been clinically diagnosed as having "complete" injuries and presented lesions of the maceration type. RESULTS: In sections from control spinal cord, NG2 immunoreactivity was restricted to stellate-shaped cells corresponding to oligodendrocyte precursor cells. The distribution patterns of phosphacan, neurocan and versican in control human spinal cord parenchyma were similar, with a fine reticular pattern being observed in white matter (but also located in gray matter for phosphacan). Neurocan staining was also associated with blood vessel walls. Furthermore, phosphacan, neurocan and versican were present in the myelin sheaths of ventral and dorsal nerve roots axons. After human SCI, NG2 and phosphacan were both detected in the evolving astroglial scar. Neurocan and versican were detected exclusively in the lesion epicentre, being associated with infiltrating Schwann cells in the myelin sheaths of invading peripheral nerve fibres from lesioned dorsal roots. CONCLUSION: NG2 and phosphacan were both present in the evolving astroglial scar and, therefore, might play an important role in the blockade of successful CNS regeneration. Neurocan and versican, however, were located at the lesion epicentre, associated with Schwann cell myelin on regenerating peripheral nerve fibres, a distribution that was unlikely to contribute to failed CNS axon regeneration. The present data points to the importance of such correlative investigations for demonstrating the clinical relevance of experimental data.