P2-g02 Properties of Renshaw cells excited by recurrent collaterals of pudendal motoneurons in the cat

s S167 evoked by stimulation to the mesencephalic locomotor region (MLR). In the animals with acute SC hemisection, MLR stimulation evoked hindlimb locomotion only on the intact SC side, while in the animals at the 4–6th day after SC hemisection, the stimulation evoked bilateral hindlimb locomotion. These indicate that locomotor driving signals arising from the MLR and descending in the intact side of the SC primarily activated lumbar CPGs for hindlimb locomotion on the same SC side, but few days after the SC hemisection, such MLR signals could activate lumbar CPGs on the bilateral SC, possibly due to some neural plastic changes that might occur within the SC. doi:10.1016/j.neures.2009.09.868 P2-g02 Properties of Renshaw cells excited by recurrent collaterals of pudendal motoneurons in the cat Ken Muramatsu1, Masatoshi Niwa2, Kenji Sato3, Sei-Ichi Sasaki4 1 Dept Phys Ther, Health Science Univ, Yamanashi; 2 Dept Occup Ther, Health Science Univ, Yamanashi; 3 Anatmy and Physiological Science, Tokyo Medical and Dental Univ, Tokyo; 4 Center for Medical Sciences, Ibaraki Prefectural Univ of Health Sciences, Ibaraki, Japan In order to investigate the existence of Renshaw cells (RCs) excited by pudendal motoneurons (MNs) and their recurrent inhibition (RI) back to pudendal MNs, we used twelve cats with the L5–S3 vertebrae laminectomy and bilateral section of dorsal roots of L6 to S3 under anesthesia. The pudendal nerve was mounted on a bipolar silver-hook electrode for stimulation. To test for existence of RCs and examination of RIs, extra-cellular and intra-cellular recordings were made from ventral horn of S1 spinal level and pudendal MNs, respectively. Although we found RCs excited by recurrent collateral of pudendal MNs were present around the Onuf’s nucleus, only 1 out of 10 MNs received very weak RI. These results suggest that RCs excited by pudendal MNs were different in nature from RCs excited by hindlimb MNs. doi:10.1016/j.neures.2009.09.869 P2-g03 Superior colliculus lesion affects air righting reflex in the rat Xinping Yan, Kazuyoshi Okito, Takasi Yamaguchi Yamagata University, Japan Previously we reported that striatal rats exhibited abnormal air righting reflex (ARR), and supposed that the basal ganglia interfered ARR movement. The interfering output might be relayed through the superior colliculus (SC). To test the hypothesis, we investigated effects of SC ablation on ARR movements in rats. After the posterior cortex, and the dorsal hippocampus were removed, the SC was aspirated under deep anesthesia, and ARR was examined usually p.o.1 to 3d. Most prominent features of movement disorder were related to the immediate action, and the order of rotation; the ventriflexion usually observed in control was sometimes replaced by the dorsiflexion, and the first action of the forequarter rotation (in control) was frequently changed to simultaneous rotation of the forequarter and the hindquarter. Nevertheless, these effects were so moderate that normal ARR was also evoked in some trials of experiments in each animal. This presumably implied that the SC influences on the center of ARR, but the action was not indispensable for its activation. doi:10.1016/j.neures.2009.09.870 P2-g04 Electrical stimulation-evoked rhythmic burst activity in the trigeminal nerve of the neonatal mice in vitro Kiyomi Nakayama, Yoshiaki Ihara, Tomio Inoue Dept Physiol, Showa Univ School of Dent, Tokyo, Japan Sucking is rhythmical jaw movement, which can be induced by labial or intraoral tactile stimulation. In the present study, we examined whether electrical stimulation of the trigeminal nerve could elicit oromotor rhythmical activity in an isolated brainstem preparation from neonatal mice at postnatal day 0–2. Application of a train of stimulation (10 Hz, 100 pulses) to the trigeminal nerve in one side evoked rhythmic burst activity in the contralateral trigeminal nerve. The frequency of the burst activity became higher with an increase in stimulus intensity. Such burst activity was abolished by a removal of the Ca2+ ions from perfusate, suggesting that the production of the burst activity depends upon chemical synaptic transmission. In a hemilateral jaw-attached preparation, electrical stimulation of the contralateral trigeminal nerve evoked rhythmical jaw movements, of which frequency was similar to the rhythmic burst activity observed in the trigeminal nerve. This will provide a new preparation to explore the rhythm generator of suckling. doi:10.1016/j.neures.2009.09.871 P2-g05 Modulation of jaw position by cortical stimulation Isoko Ishihara, Tadafumi Adachi, Shoko Toi, Toshifumi Morimoto, Yuji Masuda Div Oral and Maxillofac Biol, Matsumoto Dent Univ, Shiojiri, Japan Electrical stimulation to cortex produces jaw movements. Its relation to motor control remains unclear. Derived from the report indicating that sustained train stimulation of cortex drove the arm toward the same position, we hypothesized that stimulation to cortex drives the jaw toward a focal position. To identify this hypothesis, jaw movements evoked by 3 types of stimulation were recorded in guinea pigs. 3 types of stimulation were repetitive 6 s train, 0.2 ms pulses at 30 Hz, sustained 0.5 s train, 0.2 ms pulses at 200 Hz and short 16 ms train, 0.3 ms pulses at 500 Hz. We specified the site to which sustained train stimulation drove the jaw toward a focal position. Short train stimulation to this site produced muscle twitch around lip. We also found another site to which repetitive stimulation induced rhythmic jaw open-close movements. And sustained train stimulation to this site drove the jaw toward a focal position, following several jaw open-close movements. This indicates that cortex has the specific site to which sustained train stimulation drives the jaw toward a focal position. doi:10.1016/j.neures.2009.09.872 P2-g06 Developmental changes of synapse related proteins in fetal rat trigeminal neurons Hidefumi Yamada, Kohji Ishihama, Kouichi Yasuda, Yoko Nakayama, Tetsuhiro Umemura, Shigehiro Shimoji, Masaki Okayama, Takahisa Yamada, Minoru Yamaoka, Kiyofumi Furusawa Matsumoto Dental University, Japan Chemical synapses consist of a number of diverse proteins, which form as postsynaptic density. PSD-95, SAP102, and GRIP implicate in formation and maturation of excitatory synapses. PSD-95 regulates the localization of the NMDA receptor by means of binding with NMDA receptor subunit 2 (NR2), which plays a critical role for rhythmical oro-maxillofacial activities, such as sucking and chewing. Neural networks for rhythm generation of sucking activities as well as respiratory activities should be prepared before birth. It is, however, unclear that the developmental changes of neural networks for sucking behavior. Here we examined the temporal distributions of synapse related proteins using with immunohistochemical study in developing rat, E14 to E21, brainstem during prenatal development. Expression of SAP102 protein exhibited early onset in the trigeminal motoneurons at E16. It was suggested that the neural networks involving trigeminal activities began to form from this pregnant period. doi:10.1016/j.neures.2009.09.873 P2-g07 Two components of phasic activity in the locus coeruleus Naohiro Saito1, Janusz Rajkowski2, Gary Aston-Jones2 1 Dept Clinic Neurosci, Sch of Med Sci, Yamagata Univ, Yamagata, Japan; 2 Dept of Neurosci, MUSC, Charleston, South Carolina, USA We investigated whether LC phasic activity influences behavioral outcome. Two monkeys were trained to perform a color-reward association task. In this task, two differently colored cues were displayed on the screen, and one of the colors was associated with reward. The animal was required to release either the left or right lever to obtain reward based on the position of the target. The color-reward association was arbitrary across blocks of trials, and the monkey was required to find the target color by trial and error in each block. During task performance, LC neurons showed both sensory and premotor-related activities. We found that the amplitude of the early component of LC phasic responses was related to behavioral response time, such that larger early LC responses were associated with faster behavioral responses. No significant relationship with behavioral responses was found for the premotor component of LC phasic activity. These results support the view that LC phasic activity facilitates behavioral responding and execution of decisions. doi:10.1016/j.neures.2009.09.874 P2-g08 Multisynaptic inputs from the internal segment of the globus pallidus (GPi) to the dorsal premotor area (PMd) of macaques Eiji Hoshi1, Yosuke Saga1, Daisuke Takahara2, Yoshihiro Hirata2, Kenichi Inoue2, Shigehiro Miyachi3, Jun Tanji 1, Masahiko Takada2 1 Tamagawa University Brain Science Institute, Tokyo, Japan; 2 Department Syst. Neurosci., Tokyo Met. Inst. Neurosci., Tokyo, Japan; 3 Sec. Brain Res., Primate Res Inst, Kyoto University, Inuyama, Japan


Background
The pudendal motoneurons (PMNs) are located in Onuf 's nucleus, innervating perineal striated muscles, such as external anal and urethral sphincters, and the ischiocavernosus and bulbospongiosus muscles [1,2]. These muscles contribute to urinary continence, fecal continence, and sexual function by the interplay of autonomic nervous system [3,4].
The present study focused on the connection of recurrent collaterals of PMNs. Recurrent collaterals of MNs are known to make synapses on inhibitory interneurons named Renshaw cells (RCs) [5]. Recurrent inhibitions (RIs) of MNs are mediated by these RCs. RIs modulate not only MN activity but also suppress certain excitatory synaptic input to MNs [6,7]. Thereby, the connection of recurrent collaterals of MNs is important to understand motor control of pelvic floor muscles. Although anatomical studies had indicated PMNs to also give off recurrent collaterals, they were not considered to make synapses on RCs, and rather terminated the signals themselves [8]. Previous studies had reported recurrent inhibitions (RIs) mediated by RCs to be absent in PMNs [9,10]. There, however, could be a possibility of weak RI associated with PMNs, since small recurrent inhibitory post-synaptic potentials (IPSPs) were observable when averaging techniques were used. Because previous studies have indicated that about 100-μV weak RIs can be detected using averaging techniques [5,11], in this study, we investigated whether (i) RCs driven by PMNs exist, and (ii) RIs occur in PMNs. of Health Sciences and were in accordance with the guiding principles for care and use of animals in the field of physiological sciences outlined by the Physiological Society of Japan. Fourteen Male and female adult cats (Shiraishi Animals, Koshigaya, Japan) weighing 2.8-4.6 kg were anesthetized with an i.p. injection of sodium pentobarbital (35-40 mg/kg). The femoral artery and forearm vein were cannulated, and arterial blood pressure was maintained at 100-130 mmHg via intravenous administration of appropriate amounts of pressor agents (Noradrenalin, Daiichi-Sankyo, Tokyo, Japan). Deep aesthesia (narrow pupil size and stable arterial blood pressure) was subsequently maintained using supplemental doses of sodium pentobarbital throughout the experiments (4-7 mg/kg/h, i.v.). The animals were then immobilized with pancuronium bromide and ventilated artificially. The end-tidal CO 2 and blood pressure were continuously monitored and body temperature maintained at 37 °C with a heating sheet. The animals were placed in a holder, laminectomy was performed from the L5 to S3 vertebrae, and the dorsal roots of L6 to S3 were bilaterally sectioned. Subsequently, the pudendal nerve, muscle nerve of the external anal sphincter muscle (EAS nerve), muscle nerve of the external urethral sphincter muscle (EUS nerve), and perineal nerve that innervate the bulbospongiosus and ischiocavernosus muscle were mounted on a bipolar silver-hook electrode for stimulation. Nerves were stimulated with rectangular pulses of 150 μs in duration. The preparation described above allows only the antidromic volley in the motor nerve fibers to enter the spinal cord when the central stump of the pudendal nerve is stimulated. All exposed tissues were covered with a pool of paraffin oil kept at 37 °C ( Fig. 1). Seven cats were used for mapping of RCs and others were used for intracellular recording of PMNs.

Mapping of interneurons
For recording spikes of interneurons that were driven by recurrent collaterals of PMNs, extracellular recordings from the ventral horn of the S1 to S2 region were made using glass-capillary microelectrode filled with Fast Green FCF dye in 3 M NaCl. The minimum strength of stimulation evoking field potentials of PMNs was defined as threshold. Discharge of interneurons was recorded during gradual increase of the strength of stimulation come down to evoke maximal amplitude of antidromically field potential of PMNs. When each branch of pudendal nerve was able to stimulate separately, in addition to the above manipulation, we stimulated each branch of the pudendal nerve to test their convergence to interneurons from other muscle nerves. Single unit's discharge was recorded using a data recorder (PC208AXx, SONY, Tokyo, Japan) and analyzed using PowerLab/8 s (ADInstruments, Dunedin, New Zealand). After a cell was identified and its physiological characteristics examined, a negative current of 20 µA was passed through the electrode for 15 min and recording site marked by Fast Green FCF dye.
After the experiment, the animals were deeply anesthetized and perfused transcardially with 10% formalin solution. Next, the spinal cord from L7 to S3 was removed and was serially and transversally sectioned at 100 µm on a freezing microtome. Subsequently, sections were mounted on gelatinized glass slides, stained with Cresyl violet, and examined using bright field light microscopy. In case of successful staining, a small intense green spot could be seen in the ventral horn of spinal cord.

Intracellular recording of PMNs
Intracellular recording from antidromically identified MNs was obtained with glass-capillary micro-electrodes filled with 2 M K-citrate. To examine recurrent IPSPs, the corresponding pudendal nerve was stimulated at an intensity just sub-threshold for antidromic excitation of the impaled cell, and the record was averaged over 200 times. Subsequently, the electrode was slightly moved to the outside of the impaled cell and extracellular recording was performed in the same manner as intracellular recording. Records were stored in a data recorder (PC208AX, SONY) and analyzed using PowerLab/8 s (AD Instruments). Data are presented as mean ± S.D.

Results
As described below, we found that the interneurons are driven by recurrent collateral of PMNs. Regardless of firing properties and the fact that these interneurons resemble RCs, we could not record recurrent IPSPs from PMNs. Therefore, it cannot be concluded that these interneurons are RCs. For convenience, we call these interneurons Renshaw-like cells (RLCs) below.

Physiological properties of RLCs
We recorded the discharge of 13 RLCs which were activated by axon collaterals of PMNs from 7 animals. Discharges from these RLCs were recorded from ventral horn of S1 vertebral level. Four of the RLCs were excited by stimulation of the whole pudendal nerve, three were excited by stimulation of the EAS nerve, and six were excited by stimulation of the EUS nerve. The thresholds of orthodromic activation of RLCs were approximately same or just above the threshold of antidromic field potential, and these RLCs did not show tonic background firing (Fig. 2a, c).
The latencies of onset of the first spike of RLC from the onset of antidromic field potentials showed a tendency to shorten according to increasing stimulus strength (Fig. 2b). At maximal stimulation, the value of the latencies ranged around 0.5-2.62 ms. The initial frequency of discharge of RLCs, measured from the interval between first and second spikes, was highest for the first impulses and gradually declined thereafter (Fig. 2a) RLCs that received recurrent excitatory collateral from EAS or EUS MNs were not fired when other muscle nerves were stimulated at supra-maximal intensity of antidromic volley (Fig. 2c).

Recording sites of RLCs
We recorded the discharge of 13 RLCs activated by axon collaterals of PMNs from 7 animals. The photograph in Fig. 2d shows the typical appearance of green spots. These spots were distinctly different from that in other tissues, and hence, easily identified. A synthetograph showing the anatomical relationship between Onuf 's nucleus and RLCs is shown in Fig. 2d. All the RLCs were found very close to Onuf 's nucleus while no RLC was found inside the nucleus. Eleven of 11 RLCs were found in the medial sites of Onuf 's nucleus and only 2 were found in the lateral sites of Onuf 's nucleus.

Intracellular recording of PMNs
Obtaining stable intracellular recording from PMNs has been more difficult than from hindlimb MNs because of the smaller size of cell bodies [12]. Therefore, we could obtain intracellular recordings without obvious damages, at membrane potential more negative than 40 mV, with no spontaneous firing of action potentials, and summed signals over 200 times in only 10 out of 22 MNs. These 10 MNs (EAS MNs: 6, EUS MNs: 4) which complied with the standards as mentioned above were used for analysis. The membrane potential of these MNs was 46.2 ± 5.12 mV. Following stimulation of the EAS or EUS nerve at an intensity just below the threshold for activation of antidromic spikes in the impaled MNs, only the noise level fluctuation of the membrane potential, a few micro-volts or less, was observed. We could not find any synaptic potentials from these MNs. Even if we included the 12 discarded MNs for data analysis, there was no synaptic potential (Fig. 2e).

Discussion
In the present study, RLCs that were synaptically activated by recurrent collaterals of PMNs were found around Onuf 's nucleus. Hindlimb RCs are well known to deliver a high-frequency burst with an initial rate of approximately 1000/s and a duration of about 40 ms [5].
In the same manner, RLCs in the present study fired with an initial rate of approximately 900/s. While the duration of spike discharge was much shorter than in hindlimb RCs, it was similar to that in RCs driven by abdominal and intercostal MNs [5,13,14]. The latencies of the onset of the first spike of RLC from the onset of antidromic activations ranged over 0.5-2.62 ms, suggesting the spikes to have been activated mono-synaptically in the shortest pathway, same as in hindlimb RCs [5]. The recording site of RLCs was frequently seen in a medial part of Onuf 's nucleus and less commonly found in a lateral part. Results were in accordance with the fact that the buttons of axon collaterals of pudendal MNs are rich in the medial sites of Onuf 's nucleus and scarce in the lateral sites [8]. It is still not clear whether RLCs were located in the inside of Onuf 's nucleus, since field potentials of PMNs might have hidden the spikes of RLCs. The number of recorded RLCs in each animal was very small compared to the buttons of axon collaterals previously reported [8]. This suggested only a small number of collaterals to terminate the RLCs, most terminating on PMNs themselves, as predicted previously [8]. Interestingly, there was no convergence to RLCs from other muscle nerves, thereby suggesting the activities of these RLCs to be highly reflecting the motor output of MNs innervating a particular muscle; the functional significance of such connection, however, still remains unclear.
Surprisingly, RIs were absent in PMNs, although we averaged the membrane potentials over 200 times. The results were in accordance with previous studies, although the observations did not use averaging technique [9,10]; small recurrent IPSPs are known to be observable only when averaging techniques were used [11]. Hence, our observation indicated most PMNs to not receive even weak RIs from RLCs. However, whether PMNs completely lack RIs could not be concluded fully, since a small number of PMNs might receive weak RIs, as reported recently in abdominal MNs [13]. Collectively, existence of RIs in PMNs remains debatable.
MNs innervating the limb muscles in cats are known to exhibit strong RI pathways, which weaken in the trunk muscle [5,13,14]. Especially respiratory MNs and abdominal MNs exhibit weak recurrent inhibition [11,13,14]. It is possible that RIs in trunk muscle become weaker as one moves from the outside to the center of the muscles, and disappear eventually in pelvic floor muscle. In other words, target cells of these RLCs may be different from typical RCs. If so, it becomes significant to know where the RLCs connect. However, we could not address this point in the present study, and would highlight the need for further studies in future.
The present study possesses some notable limitations. First, there is a possibility of bias in the selection of MNs for recordings; the MNs from which we obtained intracellular recordings may have been larger PMNs because of the difficulty in obtaining stable intracellular recordings from smaller PMNs. We might not be able to test recurrent IPSPs in smaller PMNs. Second, all experiments were performed with the subjects under anesthesia, and it remains unknown how RCs behave during wakefulness. Analysis using electrophysiological techniques, however, is nearing its limits because of the difficulty discussed above. Therefore, in the future, it is needed to proceed with research not only using electrophysiological methods but also by combining histological methods such as immunohistochemical analysis [15].

Conclusions
In summary, the present study indicated recurrent collaterals of PMNs to project RLCs. This presented a new aspect of the termination of recurrent collaterals of PMNs, since motor axon collaterals of PMNs had been previously predicted to make synaptic connection with PMNs only [6]. However, we could not identify the synaptic connection of these neurons in this study. Further studies are warranted to clarify the target neuron and synaptic nature of these RLCs and understand the functional significance of these neurons.