Pathways controlling trigeminal and auditory nerve-evoked abducens eyeblink reflexes in pond turtles.

An in vitro brain stem preparation from turtles exhibits a neural correlate of eyeblink classical conditioning during pairing of auditory (CS) and trigeminal (US) nerve stimulation while recording from the abducens nerve. The premotor neuronal circuits controlling abducens nerve-mediated eyeblinks in turtles have not been previously described, which is a necessary step for understanding cellular mechanisms of conditioning in this preparation. The purpose of the present study was to neuroanatomically define the premotor pathways that underlie the trigeminal and auditory nerve-evoked abducens eyeblink responses. The results show that the principal sensory trigeminal nucleus forms a disynaptic pathway from both the trigeminal and auditory nerves to the principal and accessory abducens motor nuclei. Additionally, the principal abducens nucleus receives vestibular inputs, whereas the accessory nucleus receives input from the cochlear nucleus. The late R2-like component of abducens nerve responses is mediated by the spinal trigeminal nucleus in the medulla. Both the principal sensory trigeminal nucleus and the abducens motor nuclei receive CS-US convergence and therefore both, or either, might be considered potential sites of synapse modification during in vitro abducens conditioning. Further data are required to determine the role of the principal sensory trigeminal nucleus in in vitro conditioning.

Distribution of anterogradely labeled trigeminal and auditory nerve boutons on abducens motor neurons in turtles: implications for in vitro classical conditioning.

A conditioned abducens nerve response is generated in in vitro brainstem preparations from turtles by pairing a weak conditioned stimulus (CS) applied to the auditory nerve that immediately precedes an unconditioned stimulus (US) applied to the trigeminal nerve. Tract-tracing studies showed direct projections from auditory and trigeminal nerves to abducens motor neurons. In light of these findings for convergent CS-US inputs, it is hypothesized that auditory and trigeminal nerve synaptic inputs onto abducens motor neurons are in spatial proximity because the CS is a weak input that may be required to be near the US inputs to have an associative effect, and conditioning occurs only when the CS and US are temporally separated by less than 100 ms. This study examined the spatial relationship of 133 anterogradely labeled synaptic boutons conveying CS or US information on retrogradely labeled abducens motor neurons. The results show that trigeminal and auditory nerve terminal fields occupy primarily the soma and proximal dendrites of abducens motor neurons. Quantitative analysis shows that the majority of labeled boutons (76% and 85% from injections of the trigeminal and auditory nerves, respectively) were apposed to somata or were localized to dendritic segments no more than about 30 microm from the nucleus. There were no quantitative differences between trigeminal and auditory nerve boutons in terms of their localization on dendrites or bouton diameter. Finally, triple labeling experiments demonstrated that individual abducens motor neurons receive inputs from both nerves and that these inputs may be in close spatial proximity to one another. This synaptic arrangement allows for the possibility that in vitro abducens conditioning is generated by coincident CS-US detection mediated by NMDA receptors and may utilize a Hebbian-like plasticity mechanism.

Role for calbindin-D28K in in vitro classical conditioning of abducens nerve responses in turtles.

Intracellular calcium has a pivotal role in synaptic modifications that may underlie learning and memory. The present study examined whether there were changes in immunoreactivity levels of the AMPA receptor subunits GluR2/3 and calcium binding proteins during classical conditioning recorded in the abducens nerve of in vitro brain stem preparations from turtles. The results showed that abducens motor neurons in unconditioned turtle brain stems were immunopositive for GluR2/3, calbindin-D28K, and calmodulin, but were immunonegative for parvalbumin. After classical conditioning, immunoreactivity for calbindin-D28K in the abducens motor nuclei was significantly reduced, whereas there were no significant changes in GluR2/3, calmodulin, or parvalbumin. This reduction in calbindin-D28K immunoreactivity was not observed following conditioning in the NMDA receptor antagonist AP-5, which blocked conditioned responses, suggesting that these changes are NMDA receptor-dependent. Moreover, the degree of the decrease in calbindin-D28K immunoreactivity was negatively correlated with the level of conditioning. Consistent with the immunocytochemical findings, Western blot analysis showed that calbindin-D28K protein levels were reduced after classical conditioning. The results support the hypothesis that in vitro classical conditioning of abducens nerve responses utilizes intracellular calcium-dependent signaling pathways that require NMDA receptor function and suggest a specific role for the calcium binding protein calbindin-D28K.

Immunocytochemical localization of glutamate receptor subunits in the brain stem and cerebellum of the turtle Chrysemys picta.

The regional distribution of ionotropic (AMPA and NMDA) and metabotropic (mGluR1alpha) glutamate receptor subunits was examined in the brain stem and cerebellum of the pond turtle, Chrysemys picta, by using immunocytochemistry and light microscopy. Subunit-specific antibodies that recognize NMDAR1, GluR1, GluR4, and mGluR1alpha were used to identify immunoreactive nuclei in the brain stem and cerebellum. Considerable immunoreactivity in the turtle brain stem and cerebellum was observed with regional differences occurring primarily in the intensity of staining with the antibodies. The red nucleus, lateral reticular nucleus and cerebellum labeled intensely for NMDAR1 and moderately for GluR1. The cerebellum also labeled strongly for mGluR1alpha. All of the cranial nerve nuclei labeled intensely for NMDAR1 and to varying degrees for GluR1, GluR4, and mGluR1alpha. Counterstaining revealed the presence of neuronal somata where there were no immunoreactive neurons in individual nuclei. This finding suggests that there are subpopulations of immunoreactive neurons within a given nucleus that bear different glutamate receptor subunit compositions. The results suggest that the glutamate receptor subunit distribution in the brain stem and cerebellum of turtles is similar to that reported for rats. Additionally, there is considerable colocalization of NMDA and AMPA receptors as revealed by light microscopy. These results have implications for the organization of neural circuits that control motor behavior in turtles, and, generally, for the function of brain stem and cerebellar neural circuits in vertebrates.

GABA-immunoreactive internuclear neurons in the ocular motor system of lampreys.

The presence of internuclear neurons in the abducens and oculomotor nuclei of lampreys [González, M.J., Pombal, M.A., Rodicio, M.C. and Anadón, R., Internuclear neurons of the ocular motor system of the larval sea lamprey, J. Comp. Neurol. 401 (1998) 1-15] indicates that coordination of eye movements by internuclear neurons appeared early during the evolution of vertebrates. In order to investigate the possible involvement of inhibitory neurotransmitters in internuclear circuits, the distribution of gamma-aminobutyric acid (GABA) in the extraocular motor nuclei of the lamprey was studied using immunocytochemical techniques. Small GABA-immunoreactive (GABAir) neurons were observed in the three ocular motor nuclei. Numerous GABAir neurons were observed in the group of internuclear neurons of the dorsal rectus oculomotor subnucleus. A second group of GABAir neurons was observed among and below the trochlear motoneurons. Two further groups of GABAir interneurons, periventricular and lateral, were located in the abducens nucleus among the cells of the caudal rectus and the ventral rectus motor subnuclei, respectively. In addition to the presence of GABAir neurons, in all the ocular motor nuclei the motoneurons were contacted by numerous GABAir boutons. Taken together, these results suggest that GABA is involved as a neurotransmitter in internuclear pathways of the ocular motor system of lampreys.

Abducens internuclear and ascending tract of Deiters inputs to medial rectus motoneurons in the cat oculomotor nucleus: neurotransmitters.

The abducens internuclear and ascending tract of Deiters (ATD) pathways are the principal excitatory inputs to medial rectus motoneurons in the oculomotor nucleus and are related to the control of conjugate horizontal eye movements. Differences in the morphology and soma-dendritic distribution of abducens internuclear and ATD synaptic endings are correlated with known differences in the physiological properties of these independent inputs. The present study extends these observations to the ultrastructural localization of the excitatory amino acid neurotransmitters, glutamate and aspartate, using a postembedding immunogold procedure combined with the pre-embedding immunoperoxidase localization of anterogradely transported biocytin from the abducens nucleus and the ventral lateral vestibular nucleus. Consistent with their spheroidal synaptic vesicle content and the asymmetric pre/postsynaptic membrane profile, both the abducens internuclear and ATD synaptic endings are labeled with glutamate and aspartate. However, quantitative analysis of the density of colloidal gold particles associated with mitochondria versus synaptic vesicles/axoplasmic matrix reveals significant differences in the metabolic versus neurotransmitter pools of the amino acids in the two populations of synaptic endings. The findings indicate that both aspartate and glutamate, possibly co-localized, are the excitatory neurotransmitters utilized by abducens internuclear synaptic endings whose burst-tonic physiological activity conveys information related to eye position to medial rectus motoneurons. By contrast, glutamate is the excitatory neurotransmitter associated with ATD synaptic endings whose high frequency burst activity is related to head velocity.

Localization of parvalbumin, calretinin, and calbindin D-28k in identified extraocular motoneurons and internuclear neurons of the cat.

Calcium-binding proteins have been shown to be excellent markers of specific neuronal populations. We aimed to characterize the expression of calcium-binding proteins in identified populations of the cat extraocular motor nuclei by means of immunohistochemistry against parvalbumin, calretinin, and calbindin D-28k. Abducens, medial rectus, and trochlear motoneurons were retrogradely labeled with horseradish peroxidase from their corresponding muscles. Oculomotor and abducens internuclear neurons were retrogradely labeled after horseradish peroxidase injection into either the abducens or the oculomotor nucleus, respectively. Parvalbumin staining produced the highest density of immunoreactive terminals in all extraocular motor nuclei and was distributed uniformly. Around 15-20% of the motoneurons were moderately stained with antibody against parvalbumin, but their axons were heavily stained, indicating an intracellular segregation of parvalbumin. Colchicine administration increased the number of parvalbumin-immunoreactive motoneurons to approximately 85%. Except for a few calbindin-immunoreactive trochlear motoneurons (1%), parvalbumin was the only marker of extraocular motoneurons. Oculomotor internuclear neurons identified from the abducens nucleus constituted a nonuniform population, because low percentages of the three types of immunostaining were observed, calbindin being the most abundant (28.5%). Other interneurons located within the boundaries of the oculomotor nucleus were mainly calbindin-immunoreactive. The medial longitudinal fascicle contained numerous parvalbumin- and calretinin-immunoreactive but few calbindin-immunoreactive axons. The majority of abducens internuclear neurons projecting to the oculomotor nucleus (80.7%) contained calretinin. Moreover, the distribution of calretinin-immunoreactive terminals in the oculomotor nucleus overlapped that of the medial rectus motoneurons and matched the anterogradely labeled terminal field of the abducens internuclear neurons. Parvalbumin immunostained 42% of the abducens internuclear neurons. Colocalization of parvalbumin and calretinin was demonstrated in adjacent semithin sections, although single-labeled neurons were also observed. Therefore, calretinin is proven to be a good marker of abducens internuclear neurons. From all of these data, it is concluded that parvalbumin, calretinin, and calbindin D-28k selectively delineate certain neuronal populations in the oculomotor system and constitute valuable tools for further analysis of oculomotor function under normal and experimental conditions.

Evidence for colocalization of GABA and glycine in afferents to retrogradely labelled rat abducens motoneurones.

The coexistence of gamma-aminobutyric acid (GABA) and glycine in axon terminals impinging on rat abducens motoneurones was investigated using a double staining procedure combining retrograde labelling of the motoneurones with HRP and post-embedding immunocytochemical staining of axon terminals. Adjacent ultrathin sections of cell bodies of identified motoneurones were individually treated with GABA or glycine antibodies. The terminals single labelled for GABA represented 11.4% of the terminals analyzed, while 8% of them were glycine immunoreactive and 9% were both GABA and glycine immunoreactive. All the labelled terminals contained pleomorphic vesicles. The mean length of apposition of the double labelled terminals was statistically larger (2.20 +/- 0.97 microns) than the GABA (1.65 +/- 0.57 microns) or glycine immunoreactive ones (1.37 +/- 0.35 microns).

GABAergic innervation of rat abducens motoneurons retrogradely labelled with HRP: quantitative ultrastuctural analysis of cell bodies and proximal dendrites.

In this quantitative electron microscopic study we investigated the distribution of GABA axon terminals on rat abducens motoneurons by combining retrograde labelling of montoneurons with post-embedding immunodetection of GABA. We analysed the synapses on 13 cell bodies and 60 proximal dendritic profiles distributed along the entire rostro-caudal extent of the nucleus. For each of these two compartments, we analysed 1754 and 1176 axon terminals in contact with 6042 and 3299 microns of postsynaptic membrane. The axon terminals were classified as Sv-type (containing spherical vesicles) or Pv-type (containing pleomorphic vesicles). The GABAergic terminals contained pleomorphic vesicles and established mainly symmetrical synaptic contacts. Their apposition lengths were greater than those of unlabelled terminals. On cell bodies, the percentage of GABAergic synaptic covering varied from 2.5% to 14.1% and the synaptic frequency of GABAergic axon terminals varied from 0.6% to 8.9%. These two parameters were significantly correlated with the diameter of the motoneurons. The percentage of synaptic covering and synaptic frequency were smaller on dendrites of small motoneurons than on those of large ones. The proximal dendrites of small motoneurons had a lesser GABAergic innervation than large ones. The total synaptic covering and frequency were smaller on somata than on dendrites. However, the percentage of synaptic covering by GABA terminals was higher on cell bodies than on proximal dendrites.