To assess the distribution of insulin-like growth-factor-related proteins during autoimmune CNS demyelination and remyelination, experimental autoimmune encephalomyelitis was produced by injecting Lewis rats with an emulsion containing guinea pig spinal cord and complete Freund's adjuvant. Tail weakness appeared at 10-12 days and was followed by hind and forelimb weakness. Paraplegia and incontinence were observed in some animals. From 8-40 days postinoculation (dpi), spinal cord sections were used to correlate lesion location and severity with mRNA distributions of insulin-like growth factor I (IGF-I), IGF-binding protein 2 (IGFBP-2), IGF-I-receptor (IGFR-I), glial fibrillary acidic protein (GFAP), and myelin basic protein (MBP). These were determined semiquantitatively by in situ hybridization. Fourteen dpi, there were inflammatory infiltrates and demyelination in both white matter (WM) and grey matter (GM). IGF-I and GFAP mRNAs were increased in these lesions and transcripts encoding myelin basic protein (MBP) were greatly reduced. Large lesions with extensive demyelination were evident in both WM and GM when mRNA levels of GFAP and IGF-I peaked 26 dpi. MBP mRNA levels began increasing 21 dpi and peaked 26 dpi, when a few thin regenerating myelin sheaths were found morphologically. Astrocytes, identified by their morphology and GFAP immunoreactivity, expressed very low levels of IGFBP-2 mRNA and peptide in normal controls; their levels were significantly higher 14 dpi, peaked 26 dpi, and then gradually decreased. Some neurons, as well as oligodendroglia in areas undergoing remyelination, expressed IGFR-I. Although levels of IGF-I, IGFBP-2, and GFAP mRNAs were highest in lesion areas, levels were also elevated around lesions and in some normal-appearing areas of WM and GM 14-40 dpi. The gene expression of both IGF-I and IGFBP-2 by hypertrophic GFAP-positive astrocytes was demonstrated 14-40 dpi by combined in situ hybridization and immunocytochemistry as well as by double immunostaining. Coexpression of IGF-I and IGFBP-2 in the same astrocyte was a frequent finding. Relative increases in both IGF-I, GFAP, IGFBP-2, IGFR-I, and MBP mRNAs peaked at about the same time. This suggests that during lesion progression and recovery, astrocytic expression of IGF-I-related peptides may reduce immune-mediated myelin injury. We also suggest that astrocytic IGFBP-2 in lesions may help target IGF-I to IGFR-I-expressing oligodendrocytes and promote remyelination of demyelinated axons.