Specificity of rabies virus as a transneuronal tracer of motor networks: Transfer from hypoglossal motoneurons to connected second-order and higher order central nervous system cell groups

The specificity of transneuronal transfer of rabies virus [challenge virus standard (CVS) strain] was evaluated in a well-characterized neuronal network, i. e., retrograde infection of hypoglossal motoneurons and transneuronal transfer to connected (second-order) brainstem neurons. The distribution of the virus in the central nervous system was studied immunohistochemically at sequential intervals after unilateral inoculation into the hypoglossal nerve. The extent of transneuronal transfer of rabies virus was strictly time dependent and was distinguished in five stages. At 1 day postinoculation, labelling involved only hypoglossal motoneurons (stage 1). Retrograde transneuronal transfer occurred from 2.0–2.5 days postinoculation (stage 2). In stages 2–4, different groups of second-order neurons were labelled sequentially, depending on the strength of their input to the hypoglossal nucleus. In stages 4 and 5, labelling extended to several cortical and subcortical cell groups, which can be regarded as higher order because they are known to control tongue movements and/or to provide input to hypoglossal-projecting cell groups. The pattern of transneuronal transfer of rabies virus resembles that of alpha-herpesviruses with regard to the nonsynchronous labelling of different groups of second-order neurons and the transfer to higher order neurons. In striking contrast to alpha-herpesviruses, the transneuronal transfer of rabies is not accompanied by neuronal degeneration. Moreover, local spread of rabies from infected neurons and axons to adjoining glial cells, neurons, or fibers of passage does not occur. The results show that rabies virus is a very efficient transneuronal tracer. Results also provide a new insight into the organization of cortical and subcortical higher order neurons that mediate descending control of tongue movements indirectly via hypoglossal-projecting neurons. © 1995 Wiley-Liss, Inc.     

Rabies virus is not cytolytic for rat spinal motoneurons <Emphasis Type="Italic">in vitro</Emphasis>

Cultures of purified rat embryonic spinal cord motoneurons were used to investigate the capacity of the neurons to survive rabies virus infection in vitro. In crude primary spinal cord cultures, neurons did not survive more than 2 days after rabies virus infection with the fixed strain Challenge Virus Standard. In contrast, virus-infected purified motoneurons resisted cytolysis for at least 7 days, as also did infected motoneurons treated with conditioned medium sampled from rabies virus-infected crude spinal cord cultures. This survival rate was also observed when motoneurons were grown in the presence of astrocytes or fibroblasts and it was not dependent on the presence of growth factors in the culture medium. Moreover, terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling experiments showed that only 30% of infected motoneurons were apoptotic after 7 days of infection. In vivo, despite the massive infection of the spinal cord in infected rat neonates, the moderate number of apoptotic cells in the ventral horn suggests that only a few motoneurons were affected by this mechanism of cell death. Morphometric analyses showed that motoneurons' axon elongated at a comparable rate in virus-infected and noninfected cultures, a sign of high metabolic activity maintained in rabies virus-infected motoneurons. In contrast, hippocampus neurons were susceptible to rabies virus infection, because 70% of infected neurons were destroyed within 3 days, a large proportion of them being apoptotic. These experiments suggest that spinal cord motoneurons consist in a neuronal population that survive rabies virus infection because the viral induction of apoptosis is delayed in these neurons. They suggest also that paralyses frequently observed in rabid animals could be the consequence of dysfunctions of the locomotor network or of the spinal cord motoneurons themselves, whose parameters could be studied in vitro.     

Inhibition of the transport of rabies virus in the central nervous system

The effect of colchicine, an inhibitor of axonal transport, on the spread of rabies virus in the central nervous system was investigated using Wistar rats. Colchicine was inoculated into the striatum at various times before and after inoculation of rabies virus into the same site. Rats were killed at various times after viral inoculation and the spread of rabies virus was monitored by rabies immunofluorescence of selected areas of brain. The most effective inhibitory effect was obtained by colchicine treatment applied two days before virus inoculation. Under these conditions, no fluorescent foci could be detected until day 3 post-infection whereas control rats exhibited infected cells as soon as two days post-infection. This inhibitory effect is reversible and the general consequence seems to be a delay in the rate of viral spread. However, five days after the virus challenge, some major brain areas were still partially preserved from infection (striatum, frontal cortex, pyriform cortex). Ten days after colchicine treatment, the microtubules have recovered their capacity to transport the virus. At the onset of paralysis, the general pattern of infection in brain sections from colchicine-treated rats was not significantly different from that of control rats. This inhibitory effect on the transport of rabies virus can be prolonged by administration of additional colchicine.     

Spinal and brain circuits to motoneurons of the bulbospongiosus muscle: retrograde transneuronal tracing with rabies virus.

Retrograde transneuronal tracing with rabies virus from the left bulbospongiosus muscle (BS) was used to identify the neural circuits underlying its peripheral and central activation. Rats were killed at 2, 3, 4, and 5 days post-inoculation (p.i.). Rabies immunolabelling was combined with immunohistochemical detection of choline acetyltransferase and oxytocin. Virus uptake was restricted to ipsilateral BS motoneurons (2 days p.i.). The onset of transfer (3 days p.i.) visualized interneurons in the dorsal grey commissure (DGC), intermediate zone, and sacral parasympathetic nucleus (SPN), mainly in DGC at L5-S1, and revealed synaptic connections between BS and external urethral sphincter motoneurons. At 4 and 5 days p.i., higher-order interneurons were labelled in other spinal areas and segments. Supraspinal labelling initially involved only Barrington's nucleus, nucleus reticularis magnocellularis, and paragigantocellularis lateralis (4 days p.i.). Later, labelling extended to other populations traditionally associated with control of sexual activity and micturition (periaqueductal grey, paraventricular nucleus, medial preoptic area, prefrontal cortex), but also indicated the intervention of somatic descending motor pathways (vestibulospinal and reticulospinal neurons, "hindlimb" regions of sensorimotor cortex and red nucleus) and cerebellar nuclei in multisynaptic innervation of the labelled motoneurons. Dual color immunofluorescence disclosed multisynaptic links between these motoneurons and thoracolumbar medial sympathetic (choline acetyltransferase-immunoreactive) neurons. In contrast, preganglionic neurons in SPN and most oxytocinergic neurons in paraventricular hypothalamic nucleus remained unlabelled, suggesting that parasympathetic and somatic outflow to pelvic organs are probably controlled by separate interneuronal populations and that oxytocinergic spinal projections are more likely to influence sacral autonomic rather than somatic outflow.     

Rabies as a transneuronal tracer of circuits in the central nervous system

The ability of selected neurotropic viruses to move transneuronally in the central nervous system makes them particularly well suited for use as tracers in experimental neuroanatomy. Recently, techniques have been developed for using rabies virus as a transneuronal tracer. Several features of rabies infection make the virus particularly useful for this purpose. We examined transneuronal transport of rabies in the central nervous system of primates after intracortical and intramuscular injections. Rabies was transported in a time-dependent manner to infect synaptically-connected chains of neurons. Transport occurred exclusively in the retrograde direction. At the survival times we used, rabies infection was restricted to neurons and did not cause cell lysis. There are several methodological and safety issues that must be considered when designing studies that use rabies as a transneuronal tracer. When appropriate protocols and laboratory practices have been established, transneuronal transport of rabies can be a safe and efficient tool for revealing the organization of multi-synaptic circuits in the central nervous system.     

Locations of neurons with respiratory-related activity in the ferret brainstem

Previous transneuronal tracing studies conducted in the ferret revealed that a large population of neurons that provides inputs to diaphragm and abdominal motoneurons is located in the ventral magnocellular portion of the medial medullary reticular formation. These observations raise the possibility that the neural substrate underlying respiratory rhythmogenesis may be different in the ferret than in other species in which this circuitry has been explored. In the present study, systematic tracking was conducted through the ferret medulla to map the locations of neurons with activity related to the contractions of respiratory muscles. As in the cat, rat, and rabbit, neurons with respiratory-related discharges were distributed either lateral or ventrolateral to the solitary nucleus (dorsal respiratory group) or in the vicinity of nucleus retroambigualis, nucleus ambiguus and the retrofacial nucleus (ventral respiratory group). Although the general organization of respiratory group neurons appeared to be similar in the ferret to that in other mammals, a difference was that few expiratory neurons were located rostrally in the ventral respiratory group. These data suggest that the ventral magnocellular medullary reticular formation is not essential for respiratory rhythm generation, at least during quiet breathing, but may participate in regulating the excitability of respiratory motoneurons or in coordinating the contractions of respiratory muscles during nonrespiratory responses (e.g. coughing or emesis).     

The control of eye movements by the cerebellar nuclei: polysynaptic projections from the fastigial,interpositus posterior and dentate nuclei to lateral rectus motoneurons in primates

Premotor circuits driving extraocular motoneurons and downstream motor outputs of cerebellar nuclei are well known. However, there is, as yet, no unequivocal account of cerebellar output pathways controlling eye movements in primates. Using retrograde transneuronal transfer of rabies virus from the lateral rectus (LR) eye muscle, we studied polysynaptic pathways to LR motoneurons in primates. Injections were placed either into the central or distal muscle portion, to identify innervation differences of LR motoneurons supplying singly innervated (SIFs) or multiply innervated muscle fibers (MIFs). We found that SIF motoneurons receive major cerebellar ‘output channels’ bilaterally, while oligosynaptic cerebellar innervation of MIF motoneurons is negligible and/or more indirect. Inputs originate from the fastigial nuclei di‐ and trisynaptically, and from a circumscribed rostral portion of the ventrolateral interpositus posterior and from the caudal pole of the dentate nuclei trisynaptically. While disynaptic cerebellar inputs to LR motoneurons stem exclusively from the caudal fastigial region involved in saccades, pursuit and convergence (via its projections to brainstem oculomotor populations), minor trisynaptic inputs from the rostral fastigial nucleus, which contributes to gaze shifts, may reflect access to vestibular and reticular eye‐head control pathways. Trisynaptic inputs to LR motoneurons from the rostral ventrolateral interpositus posterior, involved in divergence (far‐response), is likely mediated by projections to the supraoculomotor area, contributing to LR motoneuron activation during divergence. Trisynaptic inputs to LR motoneurons from the caudal dentate, which also innervates disynaptically the frontal and parietal eye fields, can be explained by its superior colliculus projections, and likely target saccade‐related burst neurons.     

The long incubation period in rabies: delayed progression of infection in muscle at the site of exposure

The striped skunk (Mephitis mephitis) is a host of rabies in large areas of Canada and the United States. In each of two experiments, equal numbers of skunks in two groups were inoculated intramuscularly with low doses of a field strain of rabies virus (street rabies virus). In each experiment, skunks in one group surviving to 2 months were killed at this time and selected tissues were used for examination by the polymerase chain reaction (PCR) method or by immunohistochemistry for rabies antigen. Results of detailed examinations using PCR technology (experiment 1) indicated that muscle at the inoculation site contained viral RNA at 2 months postinoculation, when other relevant tissues on the route of viral migration and early entrance into the central nervous system were negative. The cellular location of virus/antigen, as determined immunohistochemically in experiment 2, was striated muscle fibers and fibrocytes. Our results indicate a major role of muscle (tissue) infection at the inoculation site in the long incubation period of rabies in skunks. These and related findings will be useful in rabies control and, if applicable to other species, will be relevant in postexposure treatment. Received: 31 July 1996 / Revised, accepted: 11 November 1996     

Physiological changes accompanying anatomical remodeling of mammalian motoneurons during postnatal development

The development of respiratory motoneurons provides unique data that may be generalized to other mammalian motoneuron populations. Like other motoneurons, respiratory motoneurons undergo developmental changes in the shape of the action potential and their repetitive firing. The unique observations concern the postnatal change in the recruitment pattern of cat phrenic motoneurons that is correlated with a halving of mean input resistance, a stasis of growth in the cell membrane and a reduction in the complexity of the dendritic tree. A similar pattern of change was observed for hypoglossal motoneurons studied in rat brainstem slices. Without an increase in total membrane surface area, the decreased resistance must result from a reduced specific membrane resistance. Two mechanisms are proposed to explain this decrease in resistance: proliferation and redistribution of either synaptic inputs and/or potassium channels. Although there was a significant contribution of synaptic input in determining input resistance throughout postnatal development, it was the density of cesium- or barium-sensitive potassium conductances that differentiated low resistance from high resistance motoneurons. Low resistance motoneurons had more cesium- and barium-sensitive channels than their high resistance counterparts. Based on the variations in the relative changes observed in input resistance versus membrane time constant with these two potassium channel blockers (cesium and barium), it is proposed that the distribution of these potassium channels change with age. Initially, their distribution is skewed toward the dendrites but as development progresses, the distribution becomes more uniform across the motoneuron membrane. During postnatal development, the rapid decrease in input resistance results from a proliferation of potassium channels in the membrane and of synaptic inputs converging onto developing respiratory motoneurons while the membrane is being spatially redistributed but not expanded.     

Transneuronal tracing of neural pathways controlling abdominal musculature in the ferret

Abdominal musculature participates in generating a large number of behaviors and protective reflexes, although each abdominal muscle is frequently activated differentially during particular motor responses. For example, rectus abdominis has been reported to play less of a role in respiration than other abdominal muscles, such as transversus abdominis. In the present study, the inputs to transversus abdominis and rectus abdominis motoneurons were determined and compared using the transneuronal transport of two recombinant isogenic strains of pseudorabies virus. After a 5-day post-inoculation period, infected presumed motoneurons were observed principally in cord levels T10-T15 ipsilateral to the injections. The injection of a monosynaptic tracer, beta-cholera toxin, into transversus abdominis confirmed the distribution of motoneurons innervating this muscle. In the brainstem, neurons transneuronally infected following injection of pseudorabies virus into rectus abdominis or transversus abdominis were located in the same regions, which included the medial medullary reticular formation, the medullary raphe nuclei, and nucleus retroambiguus (the expiration region of the caudal ventral respiratory group). Double-labeled cells providing inputs to both rectus and transversus motoneurons were present in both the medial medullary reticular formation and nucleus retroambiguus. These data show that the medial medullary reticular formation contains neurons influencing the activity of multiple abdominal muscles, and support our hypothesis that this region globally affects the excitability of motoneurons involved in respiration.     

Up-regulation of Fas ligand (FasL) in the central nervous system: A mechanism of immune evasion by rabies virus

Following its injection into the hindlimbs of mice, CVS, a highly pathogenic strain of rabies virus, invades the spinal cord and brain resulting in the death of the animal. In contrast, central nervous system (CNS) invasion by PV, a strain of attenuated pathogenicity, is restricted to the spinal cord and mice infected with this virus survive. Lymphocytes display transient migration into the infected CNS in fatal rabies and sustained migration in nonfatal rabies. The transient migration of T cells in fatal rabies is associated with an increase in T-cell apoptosis. We found that the early production of Fas ligand (FasL) mRNAs was up-regulated only in fatal rabies. FasL is produced by several neuronal cells and mainly in infected neurons. In mice lacking FasL (gld), infection with the neuroinvasive rabies virus strain was less severe, and the number of CD3 T cells undergoing apoptosis was smaller than that in normal mice. These data provide strong evidence that fatal rabies virus infection involves the early triggering of FasL production leading to the destruction of migratory T cells by the Fas/FasL apoptosis pathway. This mechanism could be in part responsible for the fact that T cells cannot control neuroinvasive rabies infection. Thus, rabies virus seems to use an immunosubversive strategy that takes advantage of the immune privilege status of the CNS.     

Infection of cultured rat myotubes and neurons from the spinal cord by rabies virus

Rabies virus multiplication was investigated in cultured primary rat myotubes and neurons. The susceptibility of these two cell types to fixed rabies challenge virus strain (CVS) was monitored by fluorescence and virus titration. Differentiated rat myotubes were susceptible to rabies virus infection, and showed an increasing accumulation of viral material from day one to day four. However, these cells did not release infective viral particles, nor did they accumulate infectious virions in the cytoplasm. In contrast, infected neurons released large amounts of infectious particles. Electron microscopy observation of infected myotubes showed minor alterations and the presence of typical viral inclusions in the cytoplasm without mature virions assembling viral membranes. Competition binding experiments show that alpha-bungarotoxin inhibits rabies virus infection from 10(-5) to 10(-7) M, whereas lower toxin concentrations failed to have any effect. These data do not confirm the hypothesis of a fixed rabies virus amplification step at the site of the viral entry. On the other hand, the high susceptibility of peripheral neurons to rabies virus infection is an argument for the direct uptake of virions by these cells. The restrictive viral multiplication in the myotubes is an alternative explanation for the local persistence of rabies virus at the site of inoculation.     

Motoneurons of the axial swimming muscles in hatchling Xenopus tadpoles: features, distribution, and central synapses.

Xenopus tadpole motoneurons make cholinergic synapses within the spinal cord. This excitation changes with longitudinal position and contributes to the excitation that controls motor activity and its longitudinal spread during swimming. To explore the anatomic constraints on this excitation, backfilling has been used to examine the anatomy and distribution of the whole population of spinal motoneurons, to define the extent of their central axons and to find where they make synapses. Motoneuron features show considerable variation but do not allow their separation into primary and secondary. Most motoneurons have descending central axons and it is likely that central synapses are made from these axons as longitudinal dendritic extent is very limited. Motoneuron density reaches a broad plateau over the mid-trunk region at 12-13 per 100 microm. Soma size does not change with longitudinal position, but the dorsoventral extent of the dendrites decreases caudally, whereas the central axon length increases. Motoneuron distribution data were used to estimate the longitudinal distribution of central motoneuron axons. This has a broad plateau at 12-14 per 100 microm over much of the trunk and only decreases significantly caudal to the anus. This distribution correlates with cholinergic excitation during swimming. Transmission electron microscopy of motoneurons backfilled with horseradish peroxidase was used to show that central motoneuron axons make en passant synapses with motoneuron dendrites and the dendrites of other unstained neurons. By using measures of synapse frequency and total dendrite length, trunk motoneurons are estimated to each receive 100-200 synapses.     

Extra- and intracellular HRP analysis of the organization of extraocular motoneurons and internuclear neurons in the guinea pig and rabbit

The distribution of extraocular motoneurons and abducens and oculomotor internuclear neurons was determined in guinea pigs by injecting horseradish peroxidase (HRP) into individual extraocular muscles, the abducens nucleus, the oculomotor nucleus, and the cerebellum. Motoneurons in the oculomotor nucleus innervated the ipsilateral inferior rectus, inferior oblique, medial rectus, and the contralateral superior rectus and levator palpebrae muscles. Most motoneurons of the trochlear nucleus projected to the contralateral superior oblique muscle although a small number innervated the ipsilateral superior oblique. The abducens and accessory abducens nuclei innervated the ipsilateral rectus and retractor bulbi muscles, respectively. The somata of abducens internuclear neurons formed a cap around the lateral and ventral aspects of the abducens nucleus. The axons of these internuclear neurons terminated in the medial rectus subdivision of the contralateral oculomotor nucleus. At least two classes of guinea pig oculomotor internuclear interneurons exist. One group, located primarily ventral to the oculomotor nucleus, innervated the abducens nucleus and surrounding regions. The second group, lying mainly in the dorsal midline area of the oculomotor nucleus, projected to the cerebellum. Intracellular staining with HRP demonstrated similar soma-dendritic organization for oculomotor and trochlear motoneurons of both guinea pigs and rabbits. Dendrites of oculomotor motoneurons radiated symmetrically from the soma to cover approximately one-third of the entire nucleus, and each motoneuron sent at least one dendrite into the central gray overlying the oculomotor nucleus. In both species, a small percentage of oculomotor motoneurons possessed axon collaterals that terminated both within and outside of the nucleus. The dendrites of trochlear motoneurons extended into the medial longitudinal fasciculus and the reticular formation lateral to the nucleus. Our data on the topography of motoneurons and internuclear neurons in the guinea pig and soma-dendritic organization of motoneurons in the guinea pig and rabbit show that these species share common organizational and morphological features. In addition, comparison of these data with those from other mammals reveals that dendritic complexity (number of dendrites per motoneuron) of extraocular motoneurons exhibits a systematic increase with animal size.     

Antemortem diagnosis and prevention of human rabies

Human rabies still continues to be a significant health problem in India and other developing countries where dogs are the major vectors of transmission. Rabies in humans can present in two clinical forms, i.e., furious and paralytic. While diagnosis of furious rabies can be made based on the typical symptoms and signs, paralytic rabies poses a diagnostic dilemma to the neurologists who may encounter these cases in their practice. Although there are certain clinical features that distinguish this disease from other forms of Guillain-Barre syndromes, confirmation of diagnosis may require laboratory assistance. Conventional techniques such as antigen detection, antibody assays and virus isolation have limited success. The recently introduced molecular techniques show more promise in confirming the cases of paralytic rabies. There has not been much success in the treatment of confirmed rabies cases and recovery from rabies is extremely rare. Therefore, preventive measures of this dreaded disease after an exposure become extremely important. The present article reviews the current status of human rabies with regard to antemortem diagnosis, disease management and post-exposure prophylaxis.     

The distribution of challenge virus standard rabies virus versus skunk street rabies virus in the brains of experimentally infected rabid skunks

Summary The proposal that the bizarre behavioral changes which occur during rabies infection are due to selective infection of limbic system neurons was further studied in skunks (a species important in naturally occurring disease). A detailed immunohistochemical study of brains of skunks experimentally infected with either Challenge virus standard (CVS) or street rabies virus revealed only trace amounts of viral antigen in many limbic system neurons and marked differences in viral distribution between street and CVS virus. These data were collected during early stage rabies when behavioral changes occur. Areas which contained heavy accumulations of street rabies virus but low amounts of CVS rabies virus were the neuronal perikarya and processes of the dorsal motor nucleus of the vagus, midbrain raphe, hypoglossal and red nuclei. In contrast, large accumulations of CVS virus were found in the Purkinje cells of the cerebellum, the habenular nuclei and in pyramidal cells throughout the cerebral cortex, while corresponding areas in all street virus-infected skunks contained minimal antigen. These findings were very consistent for animals of the same experimental group and between skunks inoculated both intramuscularly and intranasally with skunk street virus. Skunks inoculated intramuscularly with CVS rabies virus failed to develop rabies. Since, in this model, street virus infection generally produces furious rabies and CVS infection results in dumb rabies, we speculate that the behavioral changes which occur in these two different clinical syndromes are due to the heavy and specific accumulation of virus in different regions of the CNS. These results show that regions other than those of the limbic system may also be involved in the pathogenesis of behavior changes in rabid animals.Supported by an MRC fellowship (NLS)     

Rabies virus infection: An update

There are still many unanswered questions in the pathogenesis of rabies, but recent progress has been made. During most of the long incubation period of rabies, the virus likely remains close the site of viral entry. Centripetal spread to the central nervous system and spread within the central nervous system occur by fast axonal transport. Neuronal dysfunction, rather than neuronal death, is responsible for the clinical features and fatal outcome in natural rabies. Recent work has changed our perspective on the ecology of rabies virus under particular circumstances in certain species. Hopefully, advances in our understanding of rabies pathogenesis will lead to advances in the treatment of this dreaded disease.     

Motor outflow to cervical motoneurons from raccoon motorsensory cortex

Motor outflow from forelimb motorsensory cortex (MsI) to forelimb muscle motoneurons in raccoon has been investigated using three approaches: 1) determination of latencies for cortically evoked efferent discharge in forelimb nerves; 2) determination of latencies for cortical facilitation of forelimb monosynaptic reflexes; and 3) intracellular recording of cortically evoked synaptic potentials. All three approaches indicated a major polysynaptic pathway (minimally disynaptic) for corticofugal facilitation or inhibition of cervical motoneurons. Suggestive evidence for a monosynaptic connection between forelimb MsI and cervical motoneurons was found for only one motoneurons. Nevertheless, the motor pathway between MsI and cervical motoneurons was shown to be more efficacious (defined on the basis of central delays) than in the cat under similar experimental conditions. The results are discussed in terms of organization changes in forelimb MsI which appear to be related to the extent to which certain mammals use their forelimbs for manipulating and exploring objects.     

Cervical prephrenic interneurons in the normal and lesioned spinal cord of the adult rat

Although monosynaptic bulbospinal projections to phrenic motoneurons have been extensively described, little is known about the organization of phrenic premotor neurons in the adult rat spinal cord. Because interneurons may play an important role in normal breathing and recovery following spinal cord injury, the present study has used anterograde and transneuronal retrograde tracing to study their distribution and synaptic relations. Exclusive unilateral, first-order labeling of the phrenic motoneuron pool with pseudorabies virus demonstrated a substantial number of second-order, bilaterally distributed cervical interneurons predominantly in the dorsal horn and around the central canal. Combined transneuronal and anterograde tracing revealed ventral respiratory column projections to prephrenic interneurons, suggesting that some propriospinal relays exist between medullary neurons and the phrenic nucleus. Dual-labeling studies with pseudorabies virus recombinants also showed prephrenic interneurons integrated with either contralateral phrenic or intercostal motoneuron pools. The stability of interneuronal pseudorabies virus labeling patterns following lateral cervical hemisection was then addressed. Except for fewer infected contralateral interneurons at the level of the central canal, the number and distribution of phrenic-associated interneurons was not significantly altered 2 weeks posthemisection (i.e., the point at which the earliest postinjury recovery of phrenic activity has been reported). These results demonstrate a heterogeneous population of phrenic-related interneurons. Their connectivity and relative stability after cervical hemisection raise speculation for potentially diverse roles in modulating phrenic function normally and postinjury.