A new model of traumatic axonal injury has been developed by causing a single, rapid, controlled elongation (tensile strain) in the optic nerve of the albino guinea pig. Electron microscopy demonstrates axonal swelling, axolemmal blebs, and accumulation of organelles identical to those seen in human and experimental brain injury. Quantitative morphometric studies confirm that 17% of the optic nerve axons are injured without vascular disruption, and horseradish peroxidase (HRP) studies confirm alterations in rapid axoplasmic transport at the sites of injury. Since 95% to 98% of the optic nerve fibers are crossed, studies of the cell bodies and terminal fields of injured axons can be performed in this model. Glucose utilization was increased in the retina following injury, confirming electron microscopic changes of central chromatolysis in the ganglion cells and increased metabolic activity in reaction to axonal injury. Decreased activity at the superior colliculus was demonstrated by delayed HRP arrival after injury. The model is unique because it produces axonal damage that is morphologically identical to that seen in human brain injury and does so by delivering tissue strains of the same type and magnitude that cause axonal damage in the human. The model offers the possibility of improving the understanding of traumatic damage of central nervous system (CNS) axons because it creates reproducible axonal injury in a well-defined anatomical system that obviates many of the difficulties associated with studying the complex morphology of the brain.

The development of a model for axonal injury in the optic nerve of the guinea pig has allowed analysis of early morphological changes within damaged axons. We provide evidence that the initial site of damage after stretch is the nodes of Ranvier, some of which develop 'nodal blebs'. The development of nodel blebs is correlated with the loss of subaxolemmal density, disruption of the neurofilament cytoskeleton and aggregation of membranous profiles of smooth endoplasmic reticulum. Nodal blebs are numerous 15 min after injury but less so at later survivals. The glial-axonal junction is intact at early survivals in damaged nodes. Marked accumulation of membranous organelles occurs in the paranodal and internodal regions adjacent to damaged nodes between two and six hours and is correlated with disruption of the myelin sheath. Axotomy and the formation of degeneration bulbs occurs between 24 and 72 h. The area of axonal injury is invaded by phagocytic cells by 72 h and large numbers of myelin figures occur within the neuropil until 14 days. The results are compared with those of other studies of diffuse axonal injury and other neuropathies. The time course of axonal changes is more rapid than during Wallerian degeneration. Our data from longer surviving animals is exactly comparable with published data. We are confident that the principal site of axonal injury is the node of Ranvier. We suggest that damage at the node results in disruption of axonal transport, which in turn leads to a cascade of events, culminating in axotomy between 24 and 72 h after the initial insult.

There is increasing evidence that there is a direct response of the cerebral microvasculature to head injury. We have investigated using SEM and TEM the response of microvessels within the white matter of the baboon brain to lateral head acceleration. There is rapid endothelial disruption and swelling of perivascular astrocytes near the sites of petechial haemorrhage. The formation of microvilli in all vessels reaches a peak at 6 h and extends at least 5 mm from the site of haemorrhage. The astrocyte response suggests a partial recovery by 6 h. The endothelial response is most marked in arterioles and venules and is maintained for 6 days after injury. We suggest there is a biphasic cerebrovascular response to brain injury. First there is rapid astrocytic swelling possibly correlated with transient disruption of the blood-brain barrier. This is followed by morphological changes in the endothelium of all vessels which are most marked in arterioles and venules and extend considerable distances throughout the neuropile. This response is discussed in the light of disruption of the blood-brain barrier.

Thin-section cytochemistry has been used to demonstrate the formation of glycogen deposits within axons after stretch injury to the optic nerve of adult guinea pigs in a model of focal axonal injury. Glycogen deposits occurred within 17% of structurally normal but, we suggest, damaged fibres within the stretched optic nerve. Adjacent fibres did not stain for glycogen. Small numbers of beta glycogen particles were present 15 min after injury within damaged axons and increasing numbers of particles occurred until 72 h. Degeneration bulbs formed by 72 h, but beta glycogen particles were sparse within these. By 7-14 days after injury there was a marked reduction in the numbers of glycogen particles within axons. Alpha rosettes of glycogen were infrequent within damaged axons. Deposition of glycogen particles within astrocytes after nerve injury was confirmed. Alpha rosettes of glycogen occurred within astrocytes by 6 h and remained until 14 days after injury. Possible mechanisms for the development of glycogen deposits within damaged axons are discussed in relation to a hypothesized influx of Ca2+ at the time of injury into damaged axons. We suggest that glycogen deposition within reactive axons reflects Ca2+ mediated alteration of glycogen synthase activity and compromized axonal transport.

No abstract