[Analysis of multisynaptic neuronal pathways by using rabies virus]

Abstract

Transneuronal transport of neurotropic viruses is a useful tool for morphological analysis of the organization of multisynaptic neuronal pathways. Rabies virus is known to label neurons transsynaptically in a retrograde direction. Here, we examined the input systems of the primary motor cortex with respect to the somatotopic arrangement. Rabies virus was injected into the hindlimb, proximal forelimb, distal forelimb, and orofacial regions of the primary motor cortex in macaques, and the distribution patterns of neuronal labeling were analyzed in the prefrontal cortex, basal ganglia, and cerebellum. Four days after the injections, third-order neuron labeling was observed in various regions of the prefrontal cortex. After the viral injection into the proximal forelimb (shoulder, elbow) region, neuronal labeling was noted primarily in the dorsal region of the dorsolateral prefrontal cortex; on the other hand, after the viral injection into the distal forelimb (wrist, digits) region, neuronal labeling was preferentially distributed in the ventral region, with the highest density in the ventrolateral convexity. In the case of the orofacial injection, prefrontal neuron labeling was predominant not only in the ventrolateral convexity but also in the orbitofrontal cortex. However, the hindlimb injection resulted in relatively sparse neuron labeling as predominantly involving the neurons in the medial prefrontal cortex. With the 4-day postinjection period, neuronal labeling was noted in the striatum retrogradely via the motor cortico-basal ganglia loop. Two distinct sets of striatal neurons were labeled: one in the dorsal putamen and the other in the ventral striatum (ventromedial putamen and nucleus accumbens). The dorsal striatal labeling was somatotopically arranged, indicating that the hindlimb, orofacial, or forelimb region was located in the dorsal, ventral, or intermediate zone of the putamen, respectively. The distribution pattern of the ventral striatal labeling was essentially the same regardless of body part representation. Likewise, Purkinje cells of the cerebellar cortex were also labeled in a somatotopic fashion. Neuronal labeling from the forelimb representation was observed mainly in lobules III-VI and crus I. The proximal forelimb labeling was both rostral and lateral to the distal forelimb labeling. Yet, the hindlimb labeling was located both rostral and lateral to the proximal forelimb labeling.