HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (10) Citing Articles via Google Scholar Google Scholar Articles by Sugiuchi, Y. Articles by Shinoda, Y. Search for Related Content PubMed PubMed Citation Articles by Sugiuchi, Y. Articles by Shinoda, Y.

Physiological Characterization of Synaptic Inputs to Inhibitory Burst Neurons From the Rostral and Caudal Superior Colliculus

Department of Systems Neurophysiology, Graduate School of Medicine, Tokyo Medical and Dental University, Tokyo, Japan

Submitted 12 May 2004; accepted in final form 6 September 2004

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
The caudal superior colliculus (SC) contains movement neurons that fire during saccades and the rostral SC contains fixation neurons that fire during visual fixation, suggesting potentially different functions for these 2 regions. To study whether these areas might have different projections, we characterized synaptic inputs from the rostral and caudal SC to inhibitory burst neurons (IBNs) in anesthetized cats. We recorded intracellular potentials from neurons in the IBN region and identified them as IBNs based on their antidromic activation from the contralateral abducens nucleus and short-latency excitation from the contralateral caudal SC and/or single-cell morphology. IBNs received disynaptic inhibition from the ipsilateral caudal SC and disynaptic inhibition from the rostral SC on both sides. Stimulation of the contralateral IBN region evoked monosynaptic inhibition in IBNs, which was enhanced by preconditioning stimulation of the ipsilateral caudal SC. A midline section between the IBN regions eliminated inhibition from the ipsilateral caudal SC, but inhibition from the rostral SC remained unaffected, indicating that the latter inhibition was mediated by inhibitory interneurons other than IBNs. A transverse section of the brain stem rostral to the pause neuron (PN) region eliminated inhibition from the rostral SC, suggesting that this inhibition is mediated by PNs. These results indicate that the most rostral SC inhibits bilateral IBNs, most likely via PNs, and the more caudal SC exerts monosynaptic excitation on contralateral IBNs and antagonistic inhibition on ipsilateral IBNs via contralateral IBNs. The most rostral SC may play roles in maintaining fixation by inhibition of burst neurons and facilitating saccadic initiation by releasing their inhibition.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
The superior colliculus (SC) plays an important role in the generation of saccades (for a review see Sparks and Hartwich-Young 1989). In the intermediate and deep layers, the SC contains neurons that show a burst of spikes before and during saccades (Schiller and Körner 1971; Wurtz and Goldberg 1972) and send their axons to the paramedian pontine reticular formation (PPRF) (Grantyn and Grantyn 1982; Izawa et al. 1999; Moschovakis et al. 1998; Scudder et al. 1996). Stimulation of the PPRF produces and a lesion of the PPRF eliminates horizontal conjugate eye movements (Bender and Shanzer 1964; Büttner-Ennever and Büttner 1988; Cohen and Komatsuzaki 1972; Cohen et al. 1968; Henn et al. 1984). Two groups of neurons that show a burst of spikes before and during saccades have been reported in the brain stem: medium-lead burst neurons (MLBNs) and long-lead burst neurons (LLBNs) (Luschei and Fuchs 1972). MLBNs include excitatory (EBNs) and inhibitory burst neurons (IBNs); the former are located in the PPRF (Cohen and Henn 1972; Keller 1974; Luschei and Fuchs 1972) and the latter are in the paramedian pontomedullary reticular formation (PPMRF) (Hikosaka and Kawakami 1977; Yoshida et al. 1982). Scudder et al. (1996) showed that the SC neurons identified as LLBNs projected to the PPRF and the PPMRF.

Precht et al. (1974) and later Grantyn and Grantyn (1976) analyzed the synaptic connections between the SC and abducens motoneurons, using intracellular recording, and showed that abducens motoneurons received excitation from the contralateral SC and inhibition from the ipsilateral SC. This excitation from the contralateral SC was mainly disynaptic or trisynaptic to abducens motoneurons, whereas the inhibition from the ipsilateral SC was mainly trisynaptic. Our previous study reanalyzed these connections and showed that the excitation and inhibition from the SC were mainly disynaptic to abducens motoneurons (Izawa et al. 1999). On the other hand, previous physiological studies reported that MLBNs in the PPRF were activated either monosynaptically (Chimoto et al. 1996) or disynaptically by LLBNs from the SC (Raybourn and Keller 1977). Similarly MLBNs in the PPMRF were monosynaptically activated from the SC (Chimoto et al. 1996). The direct projection from the SC to the PPMRF has been confirmed anatomically (Harting 1977; Olivier et al. 1993).

In the rostral part of the SC, one group of neurons that discharge during fixation and pause during most saccades has been reported (Munoz and Guitton 1989, 1991; Munoz and Wurtz 1993a, b; Peck 1989), and the "fixation zone" that contains such neurons (Guitton 1991; Munoz and Istvan 1998; Munoz and Wurtz 1995b) has been hypothesized to prevent saccades via excitatory projections to pause neurons (PNs) in the nucleus raphe interpositus on the midline within the PPRF (Büttner-Ennever and Horn 1994; Büttner-Ennever et al. 1988; Evinger et al. 1977; Langer and Kaneko 1984, 1990; Ohgaki et al. 1987, 1989; Strassman et al. 1987). The remaining part of the SC is called the "saccade zone" (Gandhi and Keller 1999; Munoz and Istvan 1998), and contains burst and build-up neurons (Anderson et al. 1998; Basso and Wurtz 1998; Munoz and Wurtz 1995a). It has been suggested that 2 regions in the SC have reciprocal functions: the rostral zone maintains fixation, whereas the more caudal region participates in the generation of saccades (Goldberg and Wurtz 1972; Munoz and Guitton 1989, 1991; Munoz and Istvan 1998; Munoz and Wurtz 1993a, b, 1995a, b; Munoz et al. 1991; Paré and Guitton 1994; Paré et al. 1994; Peck 1989). However, the functional independence of the rostral pole of the SC is not necessarily accepted (Gandhi and Keller 1999).

Functionally, when something interesting appears in the visual field and the line of sight is shifted to that object, 2 kinds of different neural mechanisms are thought to be involved: one makes an accurate saccade to the object of concern, and the other suppresses saccades toward other objects that appear in the visual field but are not of interest at the moment. If these 2 different systems of reciprocal functions exist in the SC, the pattern of connections from the rostral and caudal parts of the SC to burst neurons in the brain stem should be different, but detailed information on synaptic inputs from the superior colliculi (SCs) to these burst neurons is not yet available.

The present study was performed to determine neural connections from the rostral and caudal parts of the SC to IBNs projecting to the abducens nucleus, using intracellular recording and staining methods in the anesthetized cat. The results show that the caudal parts of the contralateral and ipsilateral SC exert monosynaptic excitation and disynaptic inhibition by contralateral IBNs on IBNs, respectively. To the contrary, the rostral SC exerts disynaptic inhibition on IBNs on both sides, most likely via PNs in the nucleus raphe interpositus.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Experiments were performed in 17 cats weighing 2.7–4.5 kg. The data on abducens motoneurons were obtained from 7 cats that were also used in a previous report (Izawa et al. 1999). Animal experimentation was conducted in accordance with "Policies on the Use of Animals and Humans in Neuroscience Research," revised and approved by the Society for Neuroscience in 1995, and "Guiding Principles for the Care and Use of Animals in the Field of Physiological Sciences" (The Physiological Society of Japan, revised in 2001). The experimental protocol was approved by the Animal Care Committee of Tokyo Medical and Dental University. The animals were initially anesthetized with ketamine hydrochloride (Ketalar, Parke-Davis; 25 mg/kg, intramuscularly [im]) followed by -chloralose (40–45 mg/kg, intravenously [iv], initial dose, supplemented with additional doses of 10–25 mg/kg, iv throughout the remainder of the experiment). During recording, the animals were paralyzed by the iv administration of pancuronium bromide (Mioblock, Organon, Oss, The Netherlands), and artificially ventilated with end-tidal CO₂ held at 35–40 mmHg. The heart rate was constantly monitored by an electrocardiogram. The body temperature was kept at 37.0–38.5°C by a heating pad. The abducens nerve was detached from the muscle and mounted on a bipolar hook electrode for electrical stimulation. The bone over the parietal and occipital cortex was removed, and the cerebral cortex was removed by suction bilaterally to introduce stimulating electrodes into the SCs under direct visual observation. Usually, 4 concentric bipolar stimulating electrodes (inner and outer diameter, 0.1 and 0.3 mm; interelectrode distance along the longitudinal axis, 0.5 mm) were arranged rostrocaudally at 1.0–1.2 mm intervals along the presumed horizontal meridian of the motor map in the SC on both sides (McIlwain 1986). Their tips were positioned in the intermediate or deep layer (1.5–2.0 mm deep from the surface) of the SC (Izawa et al. 1999; Kawamura and Hashikawa 1978; Moschovakis and Karabelas 1985). The vermis overlying the fourth ventricle was removed by suction to facilitate intracellular recording from neurons in the abducens nucleus and the IBN region.

To identify the abducens nucleus and the IBN region electrophysiologically, antidromic field potentials were mapped in the abducens nucleus region, while stimulating the abducens nerve (Maeda et al. 1971b; Shinoda and Yoshida 1974). For antidromic identification of IBNs and the analysis of inputs from the contralateral IBN region to IBNs, separate electrode arrays were placed in the abducens nucleus and the IBN region contralateral to the recording site. These electrode arrays consisted of 4 monopolar electrodes (100 µm in diameter) insulated except at the tip, which were glued together around a pillar, so that the tips of the 4 electrodes were arranged dorsoventrally at 1-mm intervals. Stimulus currents were delivered between 2 adjacent tips, and the pillar was grounded to reduce stimulus artifacts. To further reduce stimulus artifacts, biphasic pulses were used; a cathodal rectangular pulse of 0.2-µs duration was immediately followed by a smaller anodal pulse of 0.1-µs duration. The amplitude of the latter was adjusted to achieve optimal cancellation of artifacts (Asanuma and Arnold 1975). Negative pulses of 0.2-ms duration were delivered at 100–500 µA for stimulation of the SC, and at <200 µA (usually <100 µA) for stimulation of the abducens nucleus and the IBN region. The positions of the stimulating electrodes in the brain stem were marked by passing negative currents of 20 µA for 20 s after each experiment, and the stimulated sites in the SC, abducens nucleus, and IBN region were histologically confirmed on sections stained with thionine.

Glass microelectrodes for intracellular recording were filled with 0.4 M KCl or 2 M K-citrate and had a resistance of 10–15 M. Glass microelectrodes for both intracellular recording and staining were filled with 7% horseradish peroxidase (HRP, Toyobo, Osaka, Japan) in 0.4 M KCl, and had a resistance of 20–40 M (Shinoda et al. 1986, 1992). To morphologically confirm the identification of electrophysiologically identified IBNs, penetrated neurons were iontophoretically injected with HRP. After a survival time of about 16 h, the animals were deeply anesthetized with pentobarbital sodium (Nembutal, Abbott, Baar, Switzerland; 50 mg/kg, iv) and perfused with 2 L of 10% sucrose phosphate buffer (pH = 7.4) followed by 2 L of a fixative solution containing 2% paraformaldehyde and 1% glutaraldehyde in 4% sucrose phosphate buffer. Serial frozen sections 80 µm thick were cut from the brain stem and treated for HRP by the heavy metal–intensification method (Adams 1981). The details of the intracellular staining method with HRP and the method used for reconstruction of the axonal trajectories of single neurons were previously described (Shinoda et al. 1986, 1992).

The location of IBNs terminating on contralateral abducens motoneurons was determined using the transneuronal labeling method (Sugiuchi et al. 1995). Wheat germ agglutinin–horseradish peroxidase (WGA–HRP; (Toyobo) was injected into the abducens nerve (Izawa et al. 1999). After 4–6 days, the animals were anesthetized with ketamine hydrochloride (25 mg/kg, im) followed by pentobarbital sodium (50 mg/kg, iv), and perfused with 2 L of 10% sucrose phosphate buffer (pH = 7.4) followed by 2 L of a fixative solution containing 4% paraformaldehyde and 0.05% glutaraldehyde with 0.2% picric acid in 4% sucrose phosphate buffer. Serial coronal sections 75 µm thick were cut on a freezing microtome. The method of histological processing for WGA–HRP staining was previously described in detail (Izawa et al. 1999). Labeled neurons were plotted under a microscope using a camera lucida system and a computer-assisted plotting and reconstruction program (Neurolucida, MicroBrightField, Colchester, VT).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Distribution of last-order premotor neurons terminating on abducens motoneurons

To penetrate IBNs efficiently and stimulate the IBN region properly at a low stimulus intensity, we determined the exact location of IBNs using a transneuronal-labeling method in 4 cats. Figure 1 shows a typical example of the distribution of labeled neurons in the midbrain, pons, and medulla after injection of WGA–HRP into the left abducens nerve. Heavily labeled neurons in the left abducens nucleus were regarded as retrogradely labeled abducens motoneurons (Fig. 1C, sections 8 and 9), whereas many lightly labeled neurons in the brain stem were considered to be labeled transneuronally (Fig. 1, A–C). On the side ipsilateral to the injected abducens nerve (Fig. 1C), labeled neurons were distributed in the PPRF from the level of the rostral end of the abducens nucleus to about 2.5 mm rostral, 0.3–2.0 mm lateral from the midline, and 0.5–2.3 mm deep from the floor of the fourth ventricle (small dots in the pontine region in Fig. 1, A and B), the vestibular nuclei [mainly the magnocellular part of the medial vestibular nucleus defined by Gerrits et al. (1985), and the superior vestibular nucleus], the nucleus prepositus hypoglossi, and in and around the oculomotor nucleus (OMN) (Fig. 1C). On the contralateral side (Fig. 1C), transneuronally labeled neurons were found in the vestibular nuclei (mainly the magnocellular part of the medial vestibular nucleus), the nucleus prepositus hypoglossi, its adjacent ventral reticular formation, in and around the OMN, and the PPMRF, the region caudomedial to the caudal part of the abducens nucleus at 0.4–1.3 mm lateral from the midline and 0.5–2.5 mm deep from the floor of the fourth ventricle (large dots in Fig. 1, A and B). This last region corresponds to the region where IBNs were found during the quick phase of vestibular nystagmus (Curthoys et al. 1981; Hikosaka and Kawakami 1977) and during saccades (Yoshida et al. 1982). Similar results were obtained from all 4 cats. Based on these anatomical data, we could accurately determine recording and stimulation sites in the IBN region relative to the abducens nucleus, which was identified by recording typical negative antidromic field potentials (Baker et al. 1975; Maeda et al. 1971b; Shinoda and Yoshida 1974) at the beginning of each experiment.

View larger version (40K): [in this window] [in a new window]   FIG. 1. Distribution of last-order premotor neurons terminating on abducens motoneurons and labeled in the midbrain, pons, and medulla after injection of wheat germ agglutinin–horseradish peroxidase (WGA–HRP) into the left abducens nerve. A: dorsal view of the brain stem showing the distribution of the same labeled neurons as in C but only neurons in the paramedian pontine reticular formation (PPRF) (small dots in the pons), the paramedian pontomedullary reticular formation (PPMRF) (large dots), and in and around the oculomotor nucleus (small dots in the midbrain). Dotted line indicates the lateral border of the floor of the fourth ventricle. B: lateral view of the brain stem showing the distribution of the same labeled neurons as in A. Neurons in each 75-µm-thick section are plotted at intervals of 225 µm (pons and medulla) and 150 µm (midbrain). Labeled neurons are projected onto a parasagittal plane 1.2 mm from the midline. Broken line indicates the dorsal surface of the fourth ventricle in the midline. C: distribution of transneuronally labeled neurons in frontal sections. Labeled neurons observed in 2 consecutive 75-µm-thick sections are plotted in representative frontal sections (sections 4–14) at 900-µm intervals except the frontal sections 1–3 that are arranged at 600-µA intervals. Each dot indicates one neuron. Caud, caudal; CG, central gray; D, descending vestibular nucleus; F, facial nucleus; G, facial genu; IO, inferior olive; L, lateral vestibular nucleus; M, medial vestibular nucleus; MVmc, magnocellular part of the medial vestibular nucleus (Gerrits et al. 1985); PH, prepositus hypoglossi nucleus; RN, red nucleus; Rost, rostral; S, superior vestibular nucleus; SCP, superior cerebellar peduncle; SO, superior olive; TB, trapezoid body; Tmo, motor trigeminal nucleus; Tsp, spinal trigeminal nucleus; Tt, spinal trigeminal tract; III, oculomotor nucleus; III n, oculomotor nerve; VI n, abducens nerve; VI, abducens nucleus; VII, facial nerve.

To analyze synaptic inputs from the SCs to IBNs, we searched for neurons in the IBN region about 0.8 mm lateral from the midline. A previous study showed that tectofugal axons arising from the contralateral SC make contacts on IBNs that terminate on and inhibit contralateral abducens motoneurons (Izawa et al. 1999). Based on these findings, we considered the following criteria for the identification of penetrated neurons as IBNs: 1) Neurons should be located in the PPMRF region (see Fig. 1, A–C) (Chimoto et al. 1996; Curthoys et al. 1981; Hikosaka and Kawakami 1977; Strassman et al. 1986; Yoshida et al. 1982). 2) They should be activated antidromically by stimulation of the contralateral abducens nucleus (Fig. 2 B) (Hikosaka and Kawakami 1977; Hikosaka et al. 1978, 1980; Yoshida et al. 1982). 3) They should receive short-latency (mainly monosynaptic) excitatory input from the contralateral SC (Fig. 2C) (Chimoto et al. 1996; Izawa et al. 1999).

View larger version (36K): [in this window] [in a new window]   FIG. 2. Electrophysiological and morphological identification of an inhibitory burst neuron (IBN). A: schematic drawing of the experimental setup. Intracellular potentials were recorded in IBNs in the left PPMRF. Stimulating electrodes were placed in the bilateral superior colliculi (SCs) and the right abducens nucleus (Abd Nucl). caud SC, caudal part of the superior colliculus (SC). B: antidromic spikes in a left IBN evoked by stimulation of the right abducens nucleus at 150 µA. In this and the following figures the bottom traces are field potentials recorded just outside the penetrated cell. C and D: excitatory postsynaptic potentials (EPSPs) in the left IBN evoked by stimulation of the right caudal SC at 500 µA (C), and inhibitory postsynaptic potentials (IPSPs) in the same IBN evoked by stimulation of the left caudal SC at 500 µA (D) before (a) and after Cl– injection into the cell (b). Bottommost traces (c) are juxtacellular field potentials. Calibration for B also applies to C and D. E: frontal reconstruction of the axonal trajectory of a single IBN on 9 serial sections showing its projection into the abducens nucleus. This neuron labeled by intracellular injection with HRP after electrophysiological identification was located 0.6 mm caudal to the caudal end of the abducens nucleus. Note that a stem axon crossed the midline at almost the same rostrocaudal level as its cell body and terminated extensively in the contralateral abducens nucleus. VN, vestibular nucleus. F: photomicrograph of a stained cell body and dendrites of an IBN in the frontal section. Cell body was located in the left PPMRF. Arrow indicates its stem axon.

Many of these axon terminals made apparent contacts with counterstained cells in the abducens nucleus. Among the 25 stained neurons, 22 had axon branches with terminal boutons in the abducens nucleus and their other axon branches in the contralateral PPRF, PPMRF, vestibular nuclei, and nucleus prepositus hypoglossi. The remaining 3 neurons were poorly stained and axon branches were observed only in the abducens nucleus. The projection sites of these stained neurons were in good accordance with those of IBNs identified physiologically in alert animals (Hikosaka and Kawakami 1977), and their morphological features were consistent with those of IBNs described in previous studies (Strassman et al. 1986; Yoshida et al. 1982). Accordingly, this morphological result confirmed that the neurons that satisfied the above 3 electrophysiological criteria could be regarded as IBNs. In neurons that satisfied these 3 criteria, we usually observed disynaptic inhibition from the ipsilateral caudal SC (Fig. 2D), and therefore we added this property to the above criteria.

Because the firing pattern of these neurons during saccades could not be examined in the present anesthetized preparations, we regarded neurons that satisfied these 4 criteria as IBNs, and we used only the electrophysiological criteria to identify IBNs at later stages of the experiments. However, we could not use these criteria in experiments in which a midline section between the IBN regions was made to interrupt axons of IBNs projecting to the contralateral side. In such experiments, the criterion of antidromic activation was not used. Instead, we selected neurons that received monosynaptic excitation from the contralateral caudal SC because a previous study showed that abducens motoneurons received disynaptic inhibition from the ipsilateral SC by the contralateral IBNs and axon terminals of tectoreticular neurons directly contacted with contralateral IBNs (Izawa et al. 1999).

Synaptic inputs from the superior colliculi on both sides to IBNs

To uncover the properties of synaptic inputs from the rostral and caudal parts of the SC to IBNs, we examined the effects of stimulation on IBNs at 4 rostrocaudal sites along the horizontal meridian of the motor map in each SC. The resting membrane potentials of IBNs ranged from –40 to –75 mV (mean ± SD, –52 ± 15 mV, n = 75). Latencies of antidromic spikes evoked by stimulation of the contralateral abducens nucleus ranged from 0.4 to 1.1 ms (0.6 ± 0.2 ms, n = 62).

Figure 3 shows a typical example of the pattern of synaptic inputs from the bilateral SCs to an IBN. Stimulation of the ipsilateral SC evoked hyperpolarizations (Fig. 3B, 1–4), whereas stimulation of the contralateral SC evoked depolarizations (Fig. 3C, 5–8). The hyperpolarizations and depolarizations usually increased because the stimulation sites were more caudal in the SC. In addition, the most rostral site in the contralateral SC was different from the more caudal sites in that its stimulation evoked depolarization followed by small hyperpolarization (Fig. 3C5). Single stimuli evoked only a small depolarization (Fig. 3C, 7 and 8) and hyperpolarization (Fig. 3B4). On the other hand, double stimuli in the rostral pole could evoke large postsynaptic potentials (PSPs) as a result of temporal facilitation at an interneuronal level because most PSPs, especially hyperpolarizations, were di- or polysynaptic (see following text).

View larger version (40K): [in this window] [in a new window]   FIG. 3. Typical pattern of postsynaptic potentials (PSPs) in an IBN evoked by stimulation of the SC on both sides. A: experimental setup. Four stimulating electrodes (1 to 4 for the left SC and 5 to 8 for the right SC) were placed rostrocaudally along the horizontal meridian of the motor map in each SC. Same arrangement of stimulating electrodes in the SCs and their corresponding records obtained from individual stimulating sites are also used in the following figures. B and C: PSPs evoked by stimulation of the ipsi- (B) and contralateral SC (C), with single- and double-shock stimuli at 500 µA. Numbers 1–4 and 5–8 correspond to individual stimulation sites in the ipsi- and contralateral SC, respectively, as shown in A. Calibration for B also applies to C. Antidromic spikes were evoked in this cell at 0.6 ms by stimulation of the contralateral abducens nucleus at 100 µA (not illustrated).

View larger version (30K): [in this window] [in a new window]   FIG. 4. Comparison of synaptic inputs from the SC on both sides to an abducens motoneuron (B) and an IBN (C). A: experimental setup. LR, lateral rectus muscle; EBN, excitatory burst neuron. B: PSPs evoked in a left abducens motoneuron by stimulation of the ipsilateral (1–4) and the contralateral SC (5–8) at 500 µA. C: PSPs evoked in a left IBN by stimulation of the ipsilateral (1–4) and contralateral SC (5–8) at 500 µA. Calibration for C also applies to B.

View larger version (36K): [in this window] [in a new window]   FIG. 5. Properties of synaptic inputs from the contralateral SC to an IBN (B and C) and an abducens motoneuron (D). A: experimental setup. B and C: intracellular potentials recorded in an IBN evoked by double stimulation (500 µA) of the contralateral SC before (B) and after (C) injection of Cl– into the cell. Numbers 5–8 indicate stimulation sites in the contralateral SC as shown in A. Comparison of B5 and C5 and B6 and C6, respectively, revealed that PSPs in B5 and B6 consisted of EPSPs followed by IPSPs because the polarity of the later parts of the PSPs in B5 and B6 was reversed, but the polarity of their early parts was not in C. Comparison of B7 and C7, and B8 and C8, respectively, revealed that PSPs in B7 and B8 mainly consisted of EPSPs because the polarity of the PSPs did not change. D: intracellular potentials recorded in a left abducens motoneuron evoked by double stimuli (500 µA) of the contralateral SC. These potentials were EPSPs because the polarity of the potentials did not change after Cl– injection. Calibration for B also applies to C and D.

View larger version (25K): [in this window] [in a new window]   FIG. 6. Intraaxonal recording from a tectoreticular axon in the left IBN region (A–C) and latency histograms of PSPs evoked by stimulation of the SC (D–J). A: experimental setup. B: intraaxonal potentials recorded from a left tectoreticular axon arising from the right SC. Stimulation at sites 6 (B6) and 7 (B7) in the SC synaptically evoked spikes with fluctuating latencies at 500 µA. Stimulation at site 8 evoked double spikes at 500 µA (B8), among which the first spikes were directly activated spikes with a fixed latency of 0.4 ms and the second spikes were indirectly activated spikes with fluctuating latencies of 2.1–2.3 ms. C: temporal facilitation of direct spikes by preconditioning stimulation at site 8. At 200 µA, double stimuli given at site 8 evoked indirect spikes (a). At 300 µA, single stimuli evoked no spikes (b), but double stimuli evoked double spikes by temporal facilitation (c). First spikes were direct spikes with a fixed latency of 0.4 ms from the second stimuli. Latencies of the second spikes fluctuated between 2.1 and 2.3 ms, and these were considered to be indirectly activated. D–J: latency histograms of spikes in tectoreticular axons (D) and PSPs in IBNs (E–H) and abducens motoneurons (I, J). Abscissa and ordinate indicate latencies of PSPs and the number of cells, respectively. D: latency histogram of spikes in tectoreticular axons recorded in the IBN region and activated directly from the contralateral SC. E: latency histogram of EPSPs in IBNs evoked from the contralateral caudal SC. F: latency histogram of IPSPs in IBNs evoked from the contralateral rostral SC. G: latency histogram of IPSPs in IBNs evoked from the ipsilateral rostral SC. H: latency histogram of IPSPs in IBNs evoked from the ipsilateral caudal SC. I and J: latency histograms of EPSPs (I) and IPSPs (J) in abducens motoneurons evoked from the contralateral and ipsilateral caudal SC, respectively. These 2 histograms include cells reported in a previous paper (see Fig. 3 in Izawa et al. 1999).

In brief, IBNs received larger monosynaptic excitation from the more caudal parts of the contralateral SC and disynaptic inhibition from the most rostral part of it.

SYNAPTIC INPUTS TO IBNS FROM THE IPSILATERAL SC. Figure 7 shows typical examples of PSPs in an IBN and an abducens motoneuron evoked by stimulation of the ipsilateral SC. Stimulation of both the rostral and caudal parts of the ipsilateral SC evoked hyperpolarizations in the IBN (Fig. 7B, 1–4) and the abducens motoneuron (Fig. 7D, 1–4). The hyperpolarizations in the IBN evoked by both the rostral and caudal stimulation were regarded as IPSPs because the injection of Cl^– into the cell reversed the hyperpolarizations to depolarizations in the IBN (Fig. 7C). In abducens motoneurons, the IPSPs decreased with more rostral SC stimulation and only small IPSPs, if any, were evoked from the most rostral SC (Fig. 7D, 1–4) (Izawa et al. 1999). A similar tendency was also observed in IBNs (Figs. 4C, 2–4 and 7B, 2–4) but, unlike the abducens motoneurons, the IPSPs were always evoked by stimulation of the most rostral site (Figs. 3B1, 4C1, and 7B1) and were often as large as or larger than those evoked by stimulation of the next caudal site in the ipsilateral SC.

View larger version (29K): [in this window] [in a new window]   FIG. 7. Properties of synaptic inputs from the ipsilateral SC to an IBN (B, C) and an abducens motoneuron (D). A: experimental setup. B and C: PSPs in a left IBN evoked by single and double stimuli (500 µA) of the ipsilateral SC before (B) and after Cl– injection (C). Numbers 1–4 indicate stimulation sites in the ipsilateral SC as shown in A. IPSPs evoked by stimulation of 4 sites in the ipsilateral SC showed marked temporal facilitation with double stimuli, suggesting that they are polysynaptically evoked. D: IPSPs in a left abducens motoneuron evoked by stimulation of the ipsilateral SC at 500 µA. IPSPs decreased because the SC stimulation was more rostral and only tiny IPSPs were evoked at a rostral site. Note that in contrast to the abducens motoneuron, large IPSPs were evoked in the IBN even at the most rostral site of the SC.

In brief, ipsilateral stimulation of both the rostral and caudal SC produced disynaptic inhibition in IBNs, whereas contralateral stimulation of the caudal SC produced monosynaptic excitation and that of the most rostral SC produced disynaptic inhibition in IBNs.

Sectioning of unilateral tectoreticular axons

To determine the pathways from the SC to IBNs, we sectioned tectoreticular axons in 4 cats (Fig. 8). A unilateral transverse section was made on the right side at a level about 2 mm rostral to the rostral end of the abducens nucleus (Fig. 8, A and E) to interrupt tectoreticular fibers connecting the SC and the PN region (Curthoys et al. 1981; Evinger et al. 1977, 1982; Ohgaki et al. 1987, 1989). The size of the transections (n = 4) was 1.5–2.0 mm wide from the midline in the transverse plane and 5.0–6.5 mm deep from the surface of the fourth ventricle (Fig. 8, D and E).

View larger version (23K): [in this window] [in a new window]   FIG. 8. Effects of the sectioning of tectoreticular axons on SC-evoked PSPs in IBNs. A: experimental setup. Commissural connections between bilateral SCs are indicated by dotted lines. A transverse section of tectoreticular axons (hatched area in D) was made in the right brain stem just rostral to the pause neuron (PN) region as shown in E (thick bar). B: intracellular records in a right IBN after sectioning. Stimulation of the caudal and rostral parts of the contralateral SC did not evoke monosynaptic EPSPs and disynaptic IPSPs in this IBN, respectively (B, 1–4). In contrast, stimulation of the ipsilateral SC evoked disynaptic IPSPs (B, 5–8) as in the control without sectioning. C: intracellular records in a left IBN after sectioning. Disynaptic IPSPs (C5) and monosynaptic EPSPs (C, 6–8) were evoked by stimulation of the contralateral rostral and caudal SC, respectively. However, disynaptic IPSPs were not evoked by stimulation of the ipsilateral SC (C, 1–4). D: photomicrograph of the brain stem in the transverse section showing sectioning of tectoreticular axons. E: schematic drawing of the dorsal view of the brain stem showing the transverse section level in D.

The effects of the transverse sectioning on SC-evoked PSPs were examined in left IBNs on the side opposite the right transverse section (Fig. 8C). After sectioning, inputs from the caudal and rostral SC on the right side were preserved with evidence of monosynaptic EPSPs (Fig. 8C, 7 and 8) and disynaptic IPSPs (Fig. 8C5), respectively. On the other hand, stimulation of the left rostral (Fig. 8C1) or caudal SC (Fig. 8C, 3 and 4) evoked no disynaptic IPSPs in this IBN and in all of the other 8 left IBNs examined in 3 cats. Therefore the inhibition from the left SC to left IBNs was mediated by right tectoreticular axons arising from left SC neurons. These findings also exclude the possibility that the inhibition from the right rostral SC to left IBNs is attributed to orthodromic commissural activation of, or the antidromic activation of commissural collaterals of, left tectoreticular neurons (dotted lines in Fig. 8A). In contrast, disynaptic inhibition from the contralateral (right) rostral SC remained after sectioning in 7 of the 8 left IBNs. Therefore the disynaptic inhibition from the contralateral (right) rostral SC to left IBNs was mediated by ipsilateral (left) tectoreticular axons arising from right rostral SC neurons.

Midline section between the 2 IBN regions

As the stimulation sites in the SC became progressively more caudal, monosynaptic excitation in the contralateral IBNs increased (Fig. 4C, 6–8), and disynaptic inhibition in ipsilateral IBNs (Fig. 4C, 1–4) and abducens motoneurons (Fig. 4B, 1–4) increased. These findings suggest that the inhibition in IBNs from the ipsilateral caudal SC might be mediated by contralateral IBNs. To explore this possibility, we examined the effect of sectioning the midline between the bilateral IBN regions on SC-evoked IPSPs in IBNs (Fig. 9). The midline section extended for 4–6 mm caudally from the middle of the abducens nucleus, and 5.5–6.5 mm deep from the floor of the fourth ventricle in 3 cats (Fig. 9E). In these preparations, IBNs could not be identified by their antidromic activation from the contralateral abducens nucleus because of the midline section of commissural axons of IBNs. Therefore they were identified by their monosynaptic excitation from the contralateral SC (see METHODS) and their recording site, which was confirmed histologically after the experiments.

View larger version (42K): [in this window] [in a new window]   FIG. 9. Effects of a midline section between the bilateral IBN regions on PSPs in an abducens motoneuron (C) and an IBN (D). A: experimental setup. A midline section was made between the bilateral IBN regions shown in E. B: control records in a left IBN before sectioning. Numbers correspond to stimulation sites in the SCs shown in A. Monosynaptic EPSPs were evoked by stimulation of the contralateral caudal SC stimulation (8) and disynaptic IPSPs were evoked by ipsilateral SC stimulation (1–4). C: PSPs in a left abducens motoneuron after midline sectioning. Disynaptic EPSPs were evoked by stimulation of the contralateral SC stimulation (5, 8), but no PSPs were evoked by stimulation of the ipsilateral SC (1–4). This result confirmed that the midline section between the IBN regions completely transected commissural axons of contralateral IBNs projecting to abducens motoneurons. D: PSPs in a left IBN after midline sectioning. Monosynaptic EPSPs were evoked by stimulation of the contralateral caudal SC (8). Note that the IPSPs were evoked only by stimulation of the ipsilateral rostral SC (1 and 2), but IPSPs from the more caudal parts of the ipsilateral SC (3 and 4) disappeared. Stimulus strength, 500 µA for B–D. Calibration in B applies to C and D. E: photomicrograph of the frontal section of the brain stem showing the midline section between the bilateral IBN regions (a vertical black arrow). A horizontal arrow indicates the most ventral part of the midline section.

Interneurons for disynaptic inhibition from the ipsilateral caudal SC to IBNs

To confirm that the inhibition from the ipsilateral caudal SC to IBNs is mediated by contralateral IBNs, we analyzed PSPs in IBNs evoked by stimulation of the contralateral IBN region (Fig. 10). The IBN region was stimulated with an electrode array consisting of 4 electrodes arranged dorsoventrally (Fig. 10D). Stimulation of all of the 4 sites in and around the contralateral IBN region evoked hyperpolarizations (Fig. 10B, 1–4). These hyperpolarizations were reversed to depolarizations by Cl^– injection (Fig. 10B, dotted traces) or by passing hyperpolarizing current into the cell, indicating that the hyperpolarizations were IPSPs (Eccles 1964). The shortest latencies of IPSPs were 0.9 ms at site 3. Amplitudes of IPSPs evoked from different sites at the same intensity were mapped in Fig. 10D and the effective stimulation sites were located in the IBN region as verified histologically after the experiment (Fig. 10, C and D). The latencies of the IPSPs ranged from 0.7 to 1.8 ms (mean ± SD, 0.9 ± 0.5 ms, n = 25). This finding confirmed that contralateral IBNs exerted monosynaptic inhibition on IBNs.

View larger version (28K): [in this window] [in a new window]   FIG. 10. Direct connection of contralateral IBNs with an IBN. A: experimental setup. An array of 4 stimulating electrodes was placed in the contralateral IBN region as shown in D. B: effects of stimulation of the contralateral IBN region on an IBN. PSPs were evoked in the IBN by stimulation of the 4 sites at different depths in the contralateral IBN region at 100 µA. Number attached to each record corresponds to each stimulation site shown in D. Dashed traces show intracellular potentials after Cl– injection. C: lateral view of the brain stem to indicate the transverse plane (arrows) shown in D including an array of stimulating electrodes in the IBN region. D: distribution of effective stimulation sites in the contralateral IBN region to evoke monosynaptic IPSPs in the IBN in B. Diameter of a circle is proportional to the amplitude of the IPSPs evoked at 100 µA. E: spatial facilitation of IBN-evoked monosynaptic IPSPs in an IBN by preconditioning stimulation of the ipsilateral caudal SC. Stimulation of the ipsilateral caudal SC evoked IPSPs at 500 µA (a) but did not evoke any PSPs at 300 µA (b). Stimulation of the contralateral IBN region evoked small monosynaptic IPSPs at 80 µA (c). Combined stimulation of the IBN region (80 µA) and the ipsilateral caudal SC (300 µA) evoked much larger IPSPs (d) than the algebraic sum (dashed line) of the IPSPs evoked by stimulation of the IBN region (c) and the ipsilateral caudal SC (b). Existence of such facilitation indicated that disynaptic inhibition from the ipsilateral caudal SC was mediated by contralateral IBNs. cSC, caudal SC; T, spinal trigeminal nucleus.

Interneurons for disynaptic inhibition from the rostral SC on both sides

To identify inhibitory interneurons mediating inhibition from the rostral SC, we examined the presence or absence of spatial facilitation of the disynaptic IPSPs evoked by stimulation of the rostral SC on both sides. Stimulation of the rostral parts of the ipsilateral (Fig. 11 Aa) and the contralateral SC (Fig. 11Ab) evoked IPSPs in the same IBN, respectively. Simultaneous stimulation of the ipsilateral and contralateral SCs evoked IPSPs (Fig. 11Ac), which were much larger than the algebraic sum of individual IPSPs (dashed line in Fig. 11Ac). Such spatial facilitation was observed in all of the 9 IBNs tested. The presence of the spatial facilitation indicated that tectoreticular neurons in the rostral part of the SC on both sides converge onto common last-order inhibitory interneurons terminating on an IBN. In 3 IBNs tested, similar spatial facilitation was seen in the preparations in which a midline section, such as that shown in Fig. 9, was made between the bilateral IBN regions. Taken together, these results indicated that tectoreticular axons arising from the rostral part of the SC on both sides converge onto common inhibitory interneurons other than IBNs that are located rostral to the IBN region and terminate on IBNs.

View larger version (31K): [in this window] [in a new window]   FIG. 11. Spatial facilitation of disynaptic IPSPs evoked by the rostral parts of the SC on both sides (A) and a summary diagram of collicular inputs to IBNs (B). A: intracellular potentials evoked in an IBN by stimulation of the left rostral SC at 300 µA (a) and the right rostral SC at 450 µA (b). Combined stimulation of the same rostral parts of the SCs as in a and b evoked much larger IPSPs than the algebraic summation of the IPSPs evoked by separate stimulation of the left and right rostral SC (dashed line). This result indicated the existence of spatial facilitation (c), suggesting that IPSPs evoked by stimulation of the rostral parts of the bilateral SCs were mediated by common inhibitory interneurons other than IBNs. B: summary diagram of neural connections from the SCs to IBNs. Connections indicated by dotted lines are presumed projections and remain undetermined. IN, inhibitory interneuron other than an IBN, most likely a PN.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

Last-order premotor interneurons terminating on abducens motoneurons have been investigated by injecting HRP (Graybiel 1977a, b; Maciewicz et al. 1977) or WGA–HRP (Langer et al. 1986) into the abducens nucleus and mapping retrogradely labeled neurons in the brain stem. However, with this method several points of uncertainty remain for interpretation of the results: difficulty in localizing an injected tracer within the abducens nucleus, possibility of labeling passing fibers within and near the nucleus, difficulty in determining the exact number of neurons terminating in the nucleus, and incapability in determining whether they terminate on abducens motoneurons or internuclear neurons in the abducens nucleus. To overcome some of these methodological problems, a transneuronal-labeling method (Porter et al. 1985) was used for identifying the last-order interneurons terminating on abducens motoneurons, although the sensitivity of this staining method was low (Alstermark and Kümmel 1986; Harrison et al. 1986). When WGA–HRP was first introduced as a retrograde transneuronal tracer, it was suggested that this method could label only inhibitory interneurons such as inhibitory burst neurons and inhibitory vestibular neurons transneuronally (Porter et al. 1985). However, Alstermark and Kümmel (1990) suggested the possibility that both excitatory and inhibitory last-order interneurons could be transneuronally labeled in the spinal cord. By improving the sensitivity of this staining method (Sugiuchi et al. 1995), the present study succeeded in showing that this method could label both inhibitory and excitatory burst neurons and vestibular neurons. However, it was generally true that inhibitory neurons were labeled more heavily and abundantly in each preparation.

In the present study, transneuronally labeled neurons were located in the ipsilateral PPRF and the contralateral PPMRF. Labeled neurons in the PPRF were located from the level of the rostral end of the abducens nucleus to about 2.5 mm rostral, 0.3–2.0 mm lateral from the midline, and 0.5–2.3 mm deep from the floor of the fourth ventricle. Labeled neurons in the PPMRF were located in the region caudomedial to the caudal part of the abducens nucleus at 0.4–1.3 mm lateral from the midline and 0.5–2.5 mm deep from the floor of the fourth ventricle. Therefore the location of the present labeled neurons in the PPMRF is very consistent with the distribution of electrophysiologically identified IBNs in the cat (P 7.5–8.5 mm; Hikosaka and Kawakami 1977; P 7.0–9.0 mm; Yoshida et al. 1982), although the electrophysiological mapping was not so systematic. The information about the exact distribution of IBNs determined with this method helped efficient intracellular recording from IBNs and also effective stimulation of the IBN region.

There has been no previous systematic intracellular analysis of synaptic inputs from the SC to either EBNs or IBNs. The present study has shown that IBNs receive strong disynaptic inhibition from the caudal part of the ipsilateral SC. Stimulation of different sites in the SC evokes characteristic saccadic eye movements in awake cats (Guitton et al. 1980; McIlwain 1986) and monkeys (Robinson 1972; Sparks and Mays 1983; Stanford et al. 1996) and the spatial distribution of the effectiveness of stimulation in the SC for evoking EPSPs in abducens motoneurons is generally in accordance with the "motor map" obtained for horizontal saccadic components in awake animals (Guitton et al. 1980; Robinson 1972); as the distance from the rostral pole to the stimulation site in the SC increases, the amplitude of EPSPs evoked in abducens motoneurons also increases (Izawa et al. 1999). The same tendency was also observed in IBNs. At the onset of a saccade, SC neurons activate contralateral EBNs and IBNs, and these EBNs excite abducens motoneurons on the same side. Because IBNs give rise to axon terminals that ramify in the contralateral abducens nucleus, the PPRF and the PPMRF (Hikosaka et al. 1978; Strassman et al. 1986; Yoshida et al. 1982), they can provide inhibition not only to abducens motoneurons (Hikosaka and Kawakami 1977; Maeda et al. 1971a) but also to IBNs and probably EBNs on the opposite side. Therefore this antagonistic inhibition at supranuclear and motoneuronal levels ensures the suppression of saccades directed toward the side ipsilateral to active SC.

Another important finding in this study is that IBNs are inhibited by the rostral part of the SC on both sides. This inhibition is disynaptic and inhibitory interneurons are not IBNs because the midline section between the bilateral IBN regions did not affect the inhibition from the rostral SC on either side (Fig. 9D, 1 and 5), but eliminated inhibition from the ipsilateral caudal SC on IBNs (Fig. 9D, 3 and 4) and abducens motoneurons (Fig. 9C, 1–4). After transverse sectioning of tectoreticular fibers just rostral to the pause neuron (PN) region, the inhibition from the rostral SC on the nonsectioned side to IBNs on the sectioned side (Fig. 8B, 1 and 2) and to IBNs on the nonsectioned side (Fig. 8C, 1 and 2) would be eliminated. Therefore intervening interneurons for mediating this inhibition may be located at or caudal to the level of the PN region.

Pause neurons are a likely candidate responsible for mediating this inhibition. It has been reported that stimulation of the PN region induced monosynaptic inhibition in IBNs (Nakao et al. 1980). Stimulation of the PN region activates antidromically fixation neurons in the rostral SC (Gandhi and Keller 1997) and rostral SC stimulation activates PNs (Paré and Guitton 1994). The inhibition of IBNs caused by stimulation of the contralateral rostral SC is most likely mediated by inhibitory interneurons on the same side as the recorded IBNs because this inhibition remained after sectioning of the contralateral tectoreticular tract (Fig. 8C5) and disappeared after sectioning of the ipsilateral tectoreticular tract (Fig. 8B1). On the other hand, the inhibition of IBNs caused by stimulation of the ipsilateral rostral SC is most likely mediated by inhibitory interneurons opposite the recorded IBNs because this inhibition was eliminated by sectioning of the contralateral tectoreticular tract at the level just rostral to the PN region (Fig. 8C1) and remained after sectioning the ipsilateral tectoreticular tract (Fig. 8B5). This result also excluded the possibility that commissural neurons in the SC are responsible for the inhibition from the contralateral rostral SC. Specifically, 2 possibilities were eliminated. First, axon collaterals of tectoreticular neurons in the ipsilateral SC could not have been antidromically activated by stimulation of the contralateral SC and, second, ipsilateral SC neurons could not have been synaptically activated by stimulation of the commissural neurons in the contralateral SC (dashed presumed connections in the SCs in Fig. 8A). However, these conclusions cannot exclude the possibility that tri- or more synaptic inhibition might occur by commissural connections between the rostral SC on both sides. Axons of the inhibitory interneurons mediating inhibition from the ipsilateral rostral SC most likely cross the midline more rostrally than the IBN region because a midline section at the level of the IBN region did not eliminate inhibition from the ipsilateral rostral SC (Fig. 9D1). It is known that stem axons of most PNs cross the midline and project to the opposite PPRF and PPMRF (Ohgaki et al. 1987). Therefore if these PNs receive monosynaptic excitation from the rostral SC on the opposite side, this connection can explain the disynaptic inhibition in IBNs from the ipsilateral rostral SC.

In contrast, inhibitory interneurons that mediate the inhibition from the contralateral rostral SC to IBNs are more difficult to identify. Given that single PNs have bilateral projections to IBNs on both sides, stimulation of the rostral SC may evoke inhibition in IBNs on both sides. However, the presence of spatial facilitation between disynaptic inhibition from the rostral SC on both sides (Fig. 11A) indicated that the inhibition from the ipsi- and contralateral rostral SCs was mediated at least partly by common inhibitory interneurons other than IBNs. It