It will be well to restate the main problem at this point and to examine how far the accumulated data can help to elucidate it. The problem is this: Why are old mice generally resistant to all forms of peripheral inoculation of vesicular stomatitis virus when intracerebral injection is equally fatal for mice of all ages? The results of experiments in which the presence of virus was demonstrated by animal passage suggested that the reason can perhaps be found in (a) the different mechanisms of virus progression after intracerebral and peripheral injection, and (b) the development with age of localized barriers capable of halting the spread of virus (1, 2). The present study sought histological evidence for the nature of virus progression and for the changes observed in the older animals. The results clearly demonstrate that after intracerebral injection virus spreads along an open system, the lesions being distributed almost entirely in contiguity with the ventricles and their extensions, while after peripheral inoculations the evidence points to progression of the virus in a closed system of neurons and their processes, at least in the stage preceding neuronal necrosis, the distribution of lesions depending upon the central connections of the primary neurons connected with the inoculated site. Thus, in young mice, nasal instillation of the virus was followed by necrosis of a long chain of neurons, starting with those in the olfactory mucosa and progressing through specific zones of the olfactory pathway, pursuing the same order in which the various regions are known to have their major connections with one another. It is important to note that after nasal instillation the apparent lesions were present where the cell bodies of the neurons are situated, and not along the tracts connecting one group of neurons with another, which accounts for the lack of contiguity between the affected zones and the normal appearing, intervening areas. The assumption that the primary progression of the virus in this case occurs in a closed system is based on the absence of lesions in unrelated areas contiguous to those which are necrotic and to the tracts which connect one affected zone with another.
Additional evidence for the assumption that the initial dissemination of peripherally injected virus is in a closed system is found in the decussating optic nerve pathway primarily pursued by the intraocularly injected virus. The progression of the virus along this decussating pathway was indicated in the experimental data obtained on mice 21 days or older, while in younger animals the spread of virus was so rapid and diffuse that the pathways along which it might have occurred remained obscure (2). In the present study, in which 15 day old mice were used, the lesions in the retinal neurons and the constant involvement of only the contralateral superior colliculus left little doubt that the primary spread of the virus, even in these very young animals, must have occurred within the retinal neuron processes (axons) which decussate in the optic chiasm (in the mouse, as in the rat, very few of these go to the homolateral side) and synapse chiefly with the neurons of the contralateral superior colliculus and also, apparently to a lesser extent, with those of the contralateral external geniculate body, where lesions were also demonstrated. Virus spreading in the optic nerve along the perineural subarachnoid space would be found at the base of the brain at the optic chiasm; virus extending along the interstitial spaces in the optic nerve should involve not only the nuclei of both sides of the optic pathway but also non-optic structures, such as the medial geniculate bodies, posterior colliculi, etc., by means of the commissures of von Gudden and of Meynert, whose fibers course through the chiasm. The highly specific localization observed in the present study is best accounted for by progression along the suggested closed pathway. Hurst (10) observed that poliomyelitis virus, after injection into the left sciatic nerve, may, after invading the lumbar cord, be found first in the contralateral motor cortex or thalamus and he suggested that this was evidence of progression along a decussating pathway and in favor of the axonal hypothesis of virus spread. It was not shown, however, that this particular localization was specifically related to the introduction of virus in the left sciatic nerve, or that it could be reversed by inoculating the sciatic nerve of the opposite side. The hypothesis proposed by Hurst, however, finds support in the present instance for (a) the superior colliculi never showed lesions after intracerebral, intranasal, or intramuscular inoculations, and (b) necrosis was produced in either the right or the left superior colliculus, depending on whether the virus was injected into the left or right eyes.
The localization of lesions after injection of virus into the muscles of one leg indicated that in the young the invasion occurred along the local peripheral nerves, especially the motor fibers (neurons destroyed in the lumbar cord with those in the spinal ganglia intact), after a primary attack on the muscle itself. The only other lesions found at a late stage were in the reticular substance of the medulla, the olfactory portions of the brain appearing entirely normal. In this respect the mechanism of progression of intramuscularly injected vesicular stomatitis virus differs from that of eastern equine encephalomyelitis and pseudorabies viruses similarly injected into mice of the same age and breed: the former (E.E.E.) invades the central nervous system in the majority of instances, by being eliminated on the nasal mucosa and then along the olfactory pathways (18), while the latter appears to employ chiefly the local sensory fibers, attacking primarily the neurons in the spinal ganglia (unpublished observations).
Because the CNS of old mice remain for the most part susceptible to vesicular stomatitis virus (although definite evidence of resistance to necrosis of the neurons was observed), and because after intracerebral injection the virus has been shown to spread in an open (ventricular) system, it is clear why young and old mice are equally susceptible to inoculation by this route. After peripheral inoculation, however, it has been amply demonstrated by experimental and histological methods that the spread of this virus begins and continues, at least until the cells disintegrate, in a closed system within the neurons and their processes and apparently also across the synapses. The halting of the virus somewhere in the anterior rhinencephalon after nasal instillation in resistant mice (1) would appear to be due to an arrest in an insusceptible neuron or an impenetrable synapse somewhere in the chain, and to the failure of the affected neurons to disintegrate (no lesions were found in the CNS of these mice) and thus to liberate the virus into the open system. After intramuscular injection, on the other hand, the virus encounters a different kind of muscle cell in the old mouse, and its inability to invade the nerves may perhaps be bound up with its demonstrated inability to attack and multiply in these changed muscle cells, although the role of a possible alteration in the terminal nerve endings themselves is not yet clear. After intraocular injection, the virus fails to affect visibly the retinal neurons of resistant old mice and the further invasion of the CNS is inhibited. The resistance of old mice to peripheral inoculations of vesicular stomatitis virus thus appears to be the result of (a) changes produced by age not in the whole animal but in certain specific, isolated structures, and (b) the special mode of progression of peripherally injected virus.
It may be of interest to point out two phenomena which may perhaps be related to the one investigated in the present study. Tobacco mosaic virus has been found to produce different types of disease in certain plants of different ages; thus a widespread, systemic necrosis leads to the death of young Nicotiana rustica plants, while in old plants it is possible to produce necrotic foci in many parts of the plant by direct inoculation, although generalization does not occur from an isolated focus as it does in young specimens (19). In other words, age apparently does not change the whole plant, but it does transform something which allows the virus to spread easily from one site to another. MacNider (20) has observed that dogs which survive a severe type of hepatic injury from uranium, repair this injury with a special type of atypical, epithelial cell and become resistant not only to secondary intoxications by uranium but also by chloroform; he has also found that this change in epithelial cell type may be acquired as a product of senility, and that when it develops it imparts to the liver a degree of resistance to chloroform comparable to that induced by a process of repair following a severe hepatic injury from uranium nitrate.