Selective irradiation of the vasculature of the rat spinal cord was used in this study, which was designed specifically to address the question as to whether it is the endothelial cell or the glial progenitor cell that is the target responsible for late white matter necrosis in the CNS. Selective irradiation of the vascular endothelium was achieved by the intraperitoneal (ip) administration of a boron compound known as BSH (Na(2)B(12)H(11)SH), followed by local irradiation with thermal neutrons. The blood-brain barrier is known to exclude BSH from the CNS parenchyma. Thirty minutes after the ip injection of BSH, the boron concentration in blood was 100 microg (10)B/ g, while that in the CNS parenchyma was below the detection limit of the boron analysis system, <1 microg (10)B/g. An ex vivo clonogenic assay of the O2A (oligodendrocyte-type 2 astrocyte) glial progenitor cell survival was performed 1 week after irradiation and at various times during the latent period before white matter necrosis in the spinal cord resulted in myelopathy. One week after 4.5 Gy of thermal neutron irradiation alone (approximately one-third of the dose required to produce a 50% incidence of radiation myelopathy), the average glial progenitor cell surviving fraction was 0.03. The surviving fraction of glial progenitor cells after a thermal neutron irradiation with BSH for a comparable effect was 0.46. The high level of glial progenitor cell survival after irradiation in the presence of BSH clearly reflects the lower dose delivered to the parenchyma due to the complete exclusion of BSH by the blood-brain barrier. The intermediate response of glial progenitor cells after irradiation with thermal neutrons in the presence of a boron compound known as BPA (p-dihydroxyboryl-phenylalanine), again for a dose that represents one-third the ED(50) for radiation-induced myelopathy, reflects the differential partition of boron-10 between blood and CNS parenchyma for this compound, which crosses the blood-brain barrier, at the time of irradiation. The large differences in glial progenitor survival seen 1 week after irradiation were also maintained during the 4-5-month latent period before the development of radiation myelopathy, due to selective white matter necrosis, after irradiation with doses that would produce a high incidence of radiation myelopathy. Glial progenitor survival was similar to control values at 100 days after irradiation with a dose of thermal neutrons in the presence of BSH, significantly greater than the ED(100), shortly before the normal time of onset of myelopathy. In contrast, glial progenitor survival was less than 1% of control levels after irradiation with 15 Gy of thermal neutrons alone. This dose of thermal neutrons represents the approximate ED(90-100) for myelopathy. The response to irradiation with an equivalent dose of X rays (ED(90): 23 Gy) was intermediate between these extremes as it was to thermal neutrons in the presence of BPA at a slightly lower dose equivalent to the approximate ED(60) for radiation myelopathy. The conclusions from these studies, performed at dose levels approximately iso-effective for radiation-induced myelopathy as a consequence of white matter necrosis, were that the large differences observed in glial progenitor survival were directly related to the dose distribution in the parenchyma. These observations clearly indicate the relative importance of the dose to the vascular endothelium as the primary event leading to white matter necrosis.