Principal cells of the medial nucleus of the trapezoid body (MNTB) play a pivotal role in the processing of binaural informationby brain stem auditory circuitry. As a result of their glycinergicnature (Bledsoe et al. 1990; Sanes et al. 1987; Wenthold et al.1987; Zarbin et al. 1981), these cells convert an excitatory signal,received from globular bushy cells (GBCs) in the contralateralcochlear nucleus, to an inhibitory signal. The inhibition is projectedchiefly to the ipsilateral lateral superior olive (LSO), whichalso receives excitatory input from spherical bushy cells (SBCs)of the ipsilateral cochlear nucleus. Many LSO cells, then, areexcited by stimulation of the ipsilateral ear and inhibited bystimulation of the contralateral ear. As a consequence, they becomethe initial point at which sensitivity to interaural level disparitiesis processed.
The excitatory GBC input to the MNTB is in the form of large somatic terminals known as the calyces of Held (Banks and Smith1992; Friauf and Ostwald 1988; Glendenning et al. 1985; Held 1893;Lenn and Reese 1966; Morest 1968; Smith et al. 1991; Spirou etal. 1990; Tolbert et al. 1982; Warr 1972). The size and somaticlocation of the calyx of Held recently has made it the focus ofa considerable amount of interest in the cellular/biophysicalneuroscience community where, for the first time in the CNS, wholecell patch-clamp recordings have been made from a presynapticterminal, sometimes while simultaneously recording from the postsynapticcell as well (Borst and Sakmann 1996; Borst et al. 1995; Forsythe1994; Takahashi et al. 1996). The calyceal recordings showed thepresence of specialized potassium and calcium conductances thatwould be appropriate for a synapse that has to accurately processhigh rates of spike activity (Borst et al. 1995; Forsythe 1994)but also indicated the existence of presynaptic metabotropic glutamatereceptors that may act to modify the output of this terminal (Takahashiet al. 1996). Both in vitro sharp and patch electrode recordingsfrom rodent MNTB cells (Banks and Smith 1992; Banks et al. 1993;Borst et al. 1995; Brew and Forsythe 1995; Forsythe and Barnes-Davis1993a,b; Wu and Kelly 1991) indicates that the mature calycealinput acts primarily on non-N-methyl-D-aspartate glutamatergicreceptors to generate a large, fast suprathreshold synaptic response.The rapid repolarization of the synaptic response, allowing thesecells to follow their inputs at high rates, and the tendency ofthese cells to fire once to sustained depolarizing current are,in part, due to a dendrotoxin sensitive, low-threshold potassiumconductance. A similar conductance has been reported for the bushycells that provide the calyceal input to MNTB (Manis and Marx1991; Oertel 1983; Wu and Oertel 1984). A second fast, high-thresholdpotassium conductance also is present in MNTB cells serving torapidly repolarize the action potential (Brew and Forsythe 1995).Thus MNTB is part of an afferent chain with morphological andphysiological specializations for temporally precise transmissionof signals.
As described above, a rather extensive investigation of both the anatomic and physiological features of the calyceal inputand the MNTB cell has been made in brain stem slices. However,several major gaps remain in our knowledge of the function ofthese cells in vivo. First, the response features of MNTB principalcells to simple auditory stimuli have not been unequivocally established.The only published study of recordings from positively identifiedandsubsequently labeledprincipal cells comes from the rat MNTB (Sommeret al. 1993). In this paper, the authors recorded intraaxonallyfrom 11 MNTB principal cells and successfully injected them withhorseradish peroxidase (HRP), labeling cell body and dendritictree as well as much of the axonal field. However, the brief recordingtimes allowed only a very limited measurement of the responsesof these cells to auditory stimuli: the characteristic frequency(CF, the frequency at which the threshold intensity was the lowest)of only six cells was determined, and only 10 stimulus trialswere used to generate short tone responses to the CF (STCF responses),making it impossible to determine their physiological responsetype, i.e., primarylike (PL; a response resembling auditory nervefibers), primarylike-with-notch (PLN: a response resembling theglobular bushy cell) or phase-locked (cells of low CF with spikesoccurring at a particular phase of a low-frequency stimulus cycle).The remaining in vivo studies of MNTB have used extracellularmetal electrodes (Guinan and Li 1990; Guinan et al. 1972a,b; Liand Guinan 1971; Tsuchitani 1994, 1997) and indirect evidence,the presence of a large prepotential, to identify the recordedcells. With metal electrode recordings, many cells in the vicinityof the MNTB exhibit spikes with complex waveforms, so-called prepotentialcells. Based on similar recordings from the cochlear nucleus thatwere proposed to arise from the large auditory nerve endbulb ofHeld terminal/bushy cell complex (Pfeiffer 1966), Guinan and Li(1990) postulated that these waveforms arose from the calyx/MNTBprincipal cell complex and consisted of a prepotential in thepresynaptic calyx followed by a postsynaptic spike. Responsesto short tones from cells displaying such spike waveform wereeither primarylike, primarylike-with-notch or phase-locked. SubsequentlyTsuchitani, (1994, 1997) categorized cells in the vicinity ofMNTB as being MNTB principal cells based on the presence of aprepotential and/or a PL or PLN short tone response. Given thelarge number of fibers in the trapezoid body coursing directlythrough the MNTB with PL and PLN responses (Smith et al. 1991,1993b), we believe the presence of a prepotential is essentialfor positive identification of extracellular recordings from MNTBcells. Furthermore, while it is generally thought that MNTB cellsare monaurally driven by the contralateral ear, there are no documentedinteraural level difference functions in the literature for MNTBcells documenting this characteristic.
A second gap exists regarding the surprising observation in several rodents, bats, and in cat (Adams and Mugnaini 1990; Banksand Smith 1992; Kuwabura and Zook 1992; Kuwabura et al. 1991;Smith 1995; Sommers et al. 1993) that MNTB cells project to themedial superior olive (MSO), a nucleus containing cells that arethought to be comparing the time of arrival of the excitatoryinputs from the two ears (Yin et al. 1997). All but one of thesestudies were in vitro so the auditory response features (CF, STCFresponse, spontaneous rate) of these MSO-projecting MNTB cellscould not be assessed. In the one in vivo study (Sommers et al.1993), no auditory responses were recorded from the one labeledMNTB cell that projected to MSO. The excitatory input to the MSOcomes bilaterally from the spherical bushy cells of the anteroventralcochlear nucleus (AVCN), cells that exhibit enhanced synchronizationas compared with their auditory nerve input to low CF tones (Joriset al. 1994a). It is essential to know the character of the inhibitoryinput to these cells from the MNTB if we are to understand theprocessing of interaural time disparities in the MSO and interaurallevel differences in the LSO.
A third gap in our knowledge of the characteristics of these cells is in the anatomy of their synaptic input and output. Forthe input, early electron microscopic studies (Jean-Baptiste andMorest 1975; Lenn and Reese 1966; Morest 1968, 1973) describedthe calyceal ending and mentioned that there were other "noncalycealterminals" on the cell body, but no measures were made of theextent of these terminals, and, because of the unlabeled natureof the cells, a description of the synaptic input to only themost proximal dendrites could be given. For the MNTB output, severallight microscopic studies have described the projection patternof MNTB axons in various species (Banks and Smith 1992; Kuwaburaand Zook 1991, 1992; Schofield and Cant 1992; Sommer et al. 1992),but no electron microscopy was done to confirm or describe theterminal morphology. Cant (1984) reported that in the LSO (a majorrecipient of MNTB axon collaterals), almost three-fourths of thesurface of the LSO principal cell body and proximal dendritesare covered with synaptic terminals that are almost exclusivelythose containing small vesicles many of which are flattened orcylindrical. She proposed that these terminals arose from MNTBaxons but admitted that it remained to be demonstrated experimentally(Cant 1984). Thus the ultrastructural identification of the axonterminals of identified MNTB principal cells has never been unequivocallymade.
Our goal in these in vivo experiments was to answer the questions described above by characterizing the basic response featuresof cat MNTB cells, with glass electrodes, and to subsequentlylabel the cell with either HRP or neurobiotin. Labeled cells couldbe studied at the light and electron microscopic level to examinethe morphology of the axonal and dendritic tree as well as thedistribution of synaptic inputs and features of the output terminals.