| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
The Journal of Neurophysiology Vol. 81 No. 1 January 1999, pp. 15-28
Copyright ©1999 by the American Physiological Society
Department of Anatomy and Neurobiology and Program in Neuroscience, University of Maryland School of Medicine, Baltimore, Maryland 21201
 |
ABSTRACT |
|---|
 Abstract
 Introduction
Methods
Results
Discussion
References
|
|---|
Aroniadou-Anderjaska, Vassiliki, Matthew Ennis, and Michael T. Shipley. Current-source density analysis in the rat olfactory bulb: laminar distribution of kainate/AMPA- and NMDA-receptor-mediated currents. J. Neurophysiol. 81: 15-28, 1999. The one-dimensional current-source density method was used to analyze laminar field potential profiles evoked in rat olfactory bulb slices by stimulation in the olfactory nerve (ON) layer or mitral cell layer (MCL) and to identify the field potential generators and the characteristics of synaptic activity in this network. Single pulses to the ON evoked a prolonged (
400 ms) sink (S1ON) in the glomerular layer (GL) with corresponding sources in the external plexiform layer (EPL) and MCL and a relatively brief sink (S2ON) in the EPL, reversing in the internal plexiform and granule cell layers. These sink/source distributions suggested that S1ON and S2ON were generated in the apical dendrites of mitral/tufted cells and granule cells, respectively. The kainate/AMPA-receptor antagonist CNQX (10 µM) reduced the early phase of S1ON, blocked S2ON, and revealed a low amplitude, prolonged sink at the location of S2ON in the EPL. Reduction of Mg2+, in CNQX, enhanced both the CNQX-resistant component of S1ON and the EPL sink. This EPL sink reversed below the MCL, suggesting it was produced in granule cells. The NMDA-receptor antagonist APV (50 µM) reversibly blocked the CNQX-resistant field potentials in all layers. Single pulses were applied to the MCL to antidromically depolarize the dendrites of mitral/tufted cells. In addition to synaptic currents of granule cells, a low-amplitude, prolonged sink (S1mcl) was evoked in the GL. Corresponding sources were in the EPL, suggesting that S1mcl was generated in the glomerular dendritic tufts of mitral/tufted cells. Both S1mcl and the granule cell currents were nearly blocked by CNQX (10 µM) but enhanced by subsequent reduction of Mg2+; these currents were blocked by APV. S1mcl also was enhanced by 
-aminobutyric acid-A-receptor antagonists applied to standard medium; this enhancement was reduced by APV. ON activation produces prolonged excitation in the apical dendrites of mitral/tufted cells, via kainate/AMPA and NMDA receptors, providing the opportunity for modulation and integration of sensory information at the first level of synaptic processing in the olfactory system. Granule cells respond to input from the lateral dendrites of mitral/tufted cells via both kainate/AMPA and NMDA receptors; however, in physiological concentrations of extracellular Mg2+, NMDA-receptor activation does not contribute significantly to the granule cell responses. The glomerular sink evoked by antidromic depolarization of mitral/tufted cell dendrites suggests that glutamate released from the apical dendrites of mitral/tufted cells may excite the same or neighboring mitral/tufted cell dendrites.
The basic anatomic features of the neuronal network in the main olfactory bulb (OB) are well characterized (for a review, see Shipley et al. 1996 Slice preparation
Slices from the olfactory bulbs of 15- to 22-day-old Wistar rats were prepared as described previously (Aroniadou-Anderjaska et al. 1997 Electrophysiological recordings
The laminar structure of the main OB and the placement of stimulating electrodes are shown schematically in Fig. 1. The OB layers are clearly distinguished in slices. When slices are illuminated from above, the glomeruli appear light with darker perimeters, the border between the GL and EPL appears as a dark band, and the MCL also appears as a dark stripe. The border between the internal plexiform layer (IPL) and GCL is not clearly distinct.
CSD analysis
The CSD method of field potential analysis identifies the sites of current flow that generate potential differences in neuronal tissues (for a review, see Mitzdorf 1985
The basic OB circuitry and the placement of stimulating electrodes are shown schematically in Fig. 1. The CSD patterns evoked by stimulation in the ONL or MCL were very consistent from slice to slice. They are described below with reference to representative examples shown in the figures; where small variations existed between slices they are noted. Current sinks evoked by stimulation in the ONL or in the MCL are symbolized by SON and Smcl, respectively.
Stimulation of the olfactory nerve
CSD DISTRIBUTION IN STANDARD MEDIUM.
The laminar field potential profiles and the corresponding CSD distributions evoked by single pulses to the ON are shown in Figs. 2 and 3A1. Similar CSD patterns were obtained with a spatial resolution of 50 µm (Fig. 2) or 100 µm (Fig. 3A1). The shortest latency sink, S1ON, is localized in the GL and at the GL-EPL border. The onset and peak latency of S1ON, measured in the GL, were 1.0 ± 0.21 ms and 5.8 ± 1.43 ms (n = 9), respectively. This sink had a particularly prolonged duration of DISTRIBUTION OF KAINATE/AMPA- AND NMDA-RECEPTOR-MEDIATED CURRENTS.
To investigate the relative contributions of kainate/AMPA and NMDA receptors to the ON-evoked laminar currents we first applied the kainate/AMPA receptor antagonist CNQX (10 µM). The early component of S1ON was blocked, while most of the slow component remained in CNQX (Fig. 3A2). Sink S2ON was blocked, but a small-amplitude, long-duration sink was revealed at the same location in the EPL. Reducing Mg2+ in the medium (to nominally zero) markedly enhanced both the GL and EPL CNQX-resistant sinks (Fig. 3A3), consistent with the properties of NMDA-receptor-mediated currents (Collingridge and Lester 1989 Stimulation in the MCL
CSD DISTRIBUTION IN STANDARD MEDIUM.
Single pulses were applied to the MCL to antidromically activate the dendrites of M/T cells. Stimulation in the MCL should depolarize directly the somata and proximal dendritic segments of mitral cells, axons of tufted cells, dendrites of granule cells, and fibers of certain centrifugal inputs that ramify above the MCL. Because granule cells are inhibitory (Jahr and Nicoll 1982a
DISTRIBUTION OF KAINATE/AMPA- AND NMDA-RECEPTOR-MEDIATED CURRENTS.
To investigate the types of receptors mediating the currents evoked by antidromic activation of M/T cell dendrites, we bath applied CNQX (10 µM). Sink S1mcl was nearly blocked by CNQX (Fig. 5A2) but subsequently was enhanced in nominally zero concentration of extracellular Mg2+ (Fig. 5A3). The peak latency of this sink (in CNQX and low Mg2+) measured in the GL was 17.7 ± 0.48 ms (n = 4), whereas its duration was
ROLE OF GABAA-RECEPTOR-MEDIATED INHIBITION.
To determine whether inhibitory activity was present in the CSD distribution and/or if it influenced the amount or pattern of excitatory activity, we first applied the GABAA-receptor antagonists BMCl (10 µM; n = 3) or picrotoxin (50 µM; n = 2) to standard medium. Sink S1mcl was enhanced consistently by either antagonist (Fig. 6A2). The EPL sources corresponding to S1mcl also were enhanced, and thus it was not possible to determine if GABAA inhibition also was present in these sources. Sink S3mcl was not affected in a consistent manner. It was evident, however, that this sink was not blocked by the GABAA-receptor antagonists (Fig. 6A2), suggesting that it is an active sink rather than passive currents resulting from inhibition in the EPL. The effects of the GABAA-receptor antagonists on S2mcl were also small (<25% change in amplitude) and inconsistent. Addition of 100 µM APV to BMCl-containing medium (n = 3) reduced but did not eliminate the enhancement of S1mcl caused by BMCl (Fig. 6A3). All currents present in BMCl and APV were blocked by subsequent addition of CNQX (not shown here, see Fig. 7A3). These results suggest that GABAA-receptor-mediated inhibition suppresses both the kainate/ AMPA- and the NMDA-receptor-mediated currents of S1mcl.
Methodological considerations
In the present study, laminar field potential profiles evoked by ON stimulation or by antidromic activation of M/T cell dendrites were analyzed with the one-dimensional CSD method. This method assumes that dipole-producing currents flow mainly in one dimension. In the OB slices, currents were assumed to flow mainly along the axis perpendicular to the laminae because the apical dendrites of the major dipole-producing cells, the M/T and granule cells (Rall and Shepherd 1968 Orthodromic activation of the OB network
Stimulation of the ON produced two current sinks within the OB: 1) a long-duration sink (S1ON) in the GL with corresponding sources in the EPL and MCL. This sink was mediated by both kainate/AMPA and NMDA receptors. And 2) a sink of a relatively short duration (S2ON) in the EPL with corresponding sources in the IPL and GCL. This sink was mediated primarily by kainate/AMPA receptors. However, reduction of extracellular Mg2+ revealed a significant amount of NMDA currents with a sink/source spatial distribution similar to that of S2ON.
SINK IN THE GLOMERULAR LAYER (S1ON).
The distribution of the currents associated with S1ON (inward currents in the GL and outward currents in the EPL and MCL) suggests that this sink should be generated by neurons that receive focal synaptic input in the GL and extend processes into the EPL and MCL. The only type of neurons with this synaptic arrangement and orientation are the M/T cells. Stimulation of the ON also activates certain types of JG cells. However, S1ON does not represent responses of JG cells because these cells do not extend processes into the EPL and MCL (Mori 1987 SINK IN THE EXTERNAL PLEXIFORM LAYER (S2ON).
The second sink (S2ON) evoked by ON stimulation probably was generated by granule cells, as the inward currents were present in the EPL, where the apical dendrites of these cells are located, and the corresponding outward currents were in the IPL/GCL, where the granule cell somata and basal dendrites are located. Similarly, the sink/source distribution of the CNQX-resistant, infraglomerular currents suggested their generation by granule cells. Other cell types also are present in the EPL, and some of them receive input from the lateral dendrites of M/T cells (Toida et al. 1996 Antidromic activation of mitral/tufted cell dendrites
Stimulation in the MCL produced three glutamatergic sinks: a small, prolonged sink in the GL (S1mcl) with corresponding sources in the EPL, a strong, relatively brief sink in the EPL (S2mcl) with corresponding sources extending from the MCL to the deep GCL, and a small, prolonged sink in the IPL (S3mcl) with corresponding sources in the EPL.
GLOMERULAR SINK (S1mcl).
Antidromic depolarization of M/T cell dendrites produced a sink (S1mcl) in the GL and GL-EPL border, which was mediated by kainate/AMPA and NMDA receptors. This sink probably was generated in the apical dendritic tufts of M/T cells because the corresponding sources were in the EPL. Glomerular interneurons do not extend processes into the EPL, where S1mcl reverses, and thus they cannot be the cells that generate this sink.
GRANULE CELL CURRENTS (S2mcl AND S3mcl).
The dominant synaptic currents (S2mcl) in the laminar response profiles evoked by MCL stimulation were in the EPL. These currents probably were generated in granule cells because the corresponding sources were in the IPL and GCL, where the proximal dendrites and somata of granule cells are located. As discussed above, granule cell currents are mediated via both kainate/AMPA and NMDA receptors. Compared to the NMDA-receptor-mediated granule cell currents evoked by ON stimulation (Fig. 3A3), the granule cell NMDA currents evoked by MCL stimulation were deeper in the EPL and less dispersed temporally (Fig. 5, A3 and C). Thus MCL stimulation appeared to activate mostly lateral M/T cell dendrites that extend in the deep EPL and produced a more synchronous synaptic activation of granule cell dendrites.
RELATION TO THE FUNCTION OF THE OB.
Olfactory information appears to be encoded by the activity patterns of specific groups of glomeruli (Friedrich and Korsching 1997
INTRODUCTION
Abstract
Introduction
Methods
Results
Discussion
References
). Olfactory receptor neurons in the nasal epithelium send their axons, through the olfactory nerve (ON), to the glomeruli of the OB, where they synapse with the apical dendritic arborizations of the projection neurons, the mitral and tufted (M/T) cells, and with juxtaglomerular (JG) interneurons (Hinds 1970; Kosaka et al. 1997; Pinching and Powell 1971). The apical dendrites of M/T cells form dendrodendritic synapses with JG cells (Hinds 1970; Pinching and Powell 1971; White 1973). The lateral dendrites of M/T cells synapse with the dendrites of granule cells in the external plexiform layer (EPL) (Jackowski et al. 1978; Price and Powell 1970b). Both granule cells and certain types of JG cells make feedback, inhibitory synapses with M/T cell dendrites. In addition, the OB is innervated densely by centrifugal fibers from olfactory cortex and subcortical modulatory systems (Carson 1984; Haberly and Behan 1983; Price and Powell 1970a; Shipley et al. 1985). Thus the OB is not simply a "relay station" but rather the site of significant processing of sensory input. Nevertheless, little is known about the physiology and function of the OB neural network (for reviews, see Mori 1987; Nickell and Shipley 1992; Shipley and Ennis 1996).
). Recently, we showed that this field potential consists of a kainate/ 
-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)-receptor-mediated component and a prolonged N-methyl-D-aspartate (NMDA) component, both of which are generated, for the most part, by glomerular neuronal elements (Aroniadou-Anderjaska et al. 1997). We further suggested, based on laminar field potential profiles, that these glomerular synaptic responses are produced in the apical dendrites of M/T cells. In the present study, we used the current source density (CSD) method of field potential analysis to confirm these findings and to identify the field potential generators in other laminae of the OB. We used in vitro OB slices because this preparation allows pharmacological manipulations that can distinguish neurotransmitter receptor types mediating synaptic currents. The basic OB circuitry is preserved in slices, as indicated by the close similarity of the laminar field potential profiles evoked in slices (Aroniadou-Anderjaska et al. 1997) to those recorded in vivo (Nickell and Shipley 1992). Therefore information obtained from field potential and CSD studies in OB slices is relevant to the function of the OB.
METHODS
Abstract
Introduction
Methods
Results
Discussion
References
). Briefly, the rats were anesthetized with chloral hydrate (400 mg/kg body wt) followed by whole-body immersion in ice-cold water. The brain with the two bulbs was gently removed, and a block was cut including the two bulbs and part of the frontal cortex. The ventral surface of the cortex was glued to the stage of a Vibroslicer. Slices, 450-500 µm thick, were cut from the olfactory bulbs in approximately the horizontal plane, and immediately transferred to an interface chamber, maintained at 33°C. The slices were perfused with artificial cerebrospinal fluid [ACSF; composition (in mM): 124 NaCl, 26 NaHCO3, 1.2 NaH2PO4, 3 KCl, 1.3 MgSO4, 2.5 CaCl2, and 10 glucose] at a rate of 1 ml/min.
View larger version (26K):
[in a new window]
FIG. 1.
A schematic representation of the basic olfactory bulb (OB) circuitry. ONL, olfactory nerve layer; GL, glomerular layer; EPL, external plexiform layer; MCL, mitral cell layer; IPL, internal plexiform layer; GCL, granule cell layer; JG, juxtaglomerular cell; ET, external tufted cell; MT, middle tufted cell; MC, mitral cell; GC, granule cell. ON makes excitatory synapses with the apical dendritic tufts of mitral and tufted cells as well as with juxtaglomerular interneurons. Mitral and tufted cells make excitatory, dendrodendritic synapses with glomerular interneurons and granule cells and receive feedback inhibition from these cells. The somata of mitral cells are in the MCL, while somata of tufted cells can be found in any depth of the EPL and in the GL. Recordings were obtained along a vertical axis perpendicular to the layers. Stimulation was applied in the ONL and in the MCL, distal to the recording axis by 300-400 µm and 100-150 µm, respectively.
) was placed in the center of a glomerulus, and stimulus pulses were applied to the olfactory nerve layer (ONL), to the MCL, or to both layers alternately, using a dipolar stainless steel stimulating electrode (total diameter 100 µm). The intensity of stimulus pulses (5-80 µA, 100 µs duration) was adjusted to evoke a glomerular field potential of 1.5-2.5 mV, which is between 40 and 70% of the maximum peak amplitude. When stimulation is applied in the ONL, this stimulus/response range reliably evokes spiking activity in simultaneously recorded mitral cells (Aroniadou-Anderjaska et al. 1997) and thus should activate uniformly a relatively large area of the OB network. To examine if higher stimulus intensities produce different activation patterns, we compared, in the same slice, the CSD distributions evoked with 30-µA stimulation, which produced 1.5-mV glomerular field potentials in response to either ON or MCL stimulation, with the CSDs evoked with 300-µA stimulation, which produced a 3.5- and 2.8-mV glomerular field potential in response to ON and MCL stimulation, respectively. The CSD distributions evoked with the two stimulus intensities were qualitatively identical. Therefore the activation patterns produced with the stimulus intensities used in this study should be representative of the patterns elicited with a wide range of stimulus strengths that evoke suprathreshold responses in a significant population of principal cells.
). 2) D-2-amino-5-phosphonovalerate (APV), an NMDA-receptor antagonist, at concentrations of 50 or 100 µM, which are known to block NMDA-receptor-mediated responses in brain slices including OB slices (Aroniadou-Anderjaska et al. 1997). 3) Bicuculline methchloride (BMCl) or picrotoxin, both -aminobutyric acid-A (GABAA)-receptor antagonists, at concentrations 10 and 50 µM, respectively; lower concentrations of the GABAA antagonists had little effect on the field potentials, whereas higher concentrations often produced epileptic activity. To prepare stock solutions APV, BMCl, and picrotoxin were dissolved in dH2O, whereas CNQX was dissolved in dimethyl sulfoxide (DMSO; final concentration of DMSO in the slice medium was 0.01%, vol/vol). In some experiments, low Mg2+ medium was used; this was the same as the ACSF except that it did not contain MgSO4.
View larger version (30K):
[in a new window]
FIG. 2.
Field potential depth profile and corresponding current source density (CSD) distribution evoked by stimulation in the ONL. Spatial resolution is 50 µm. Two major current sinks were present: a prolonged sink (S1ON) in the GL and GL-EPL border, with corresponding sources in the EPL and MCL, and a relatively brief sink (S2ON) in the EPL, with corresponding sources extending from the MCL to the deep GCL. Algebraic summation of all the CSD traces resulted in a current of an amplitude close to 0 (bottom of the CSD profile). Thus currents flowing in other dimensions did not affect the laminar CSD distribution (see text).
). This method has been used to reveal physiological and anatomic aspects of neuronal circuits (Aroniadou and Keller 1993; Lambert et al. 1991; Mitzdorf and Singer 1978) on the basis of the spatiotemporal distribution of sinks (inward membrane currents reflected in the extracellular space) and corresponding sources (outward membrane currents) that are generated when a circuit is activated.
View larger version (30K):
[in a new window]
FIG. 3.
Laminar distribution of kainate/ -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)- and N-methyl-D-aspartate (NMDA)-receptor-mediated currents evoked by ON stimulation. A1: field potential depth profile and corresponding CSD distribution in standard medium. Spatial resolution is 100 µm. A2: bath application of 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX; 10 µM) blocked S2ON and a major portion of S1ON and revealed a low-amplitude, prolonged sink in the superficial EPL. A3: CNQX-resistant currents were enhanced in medium with nominally 0 concentration of Mg2+. GL sink reverses in the EPL and MCL. EPL sink reverses in the MCL and in the IPL/GCL. B: S1ON and S2ON at a faster sweep speed for a more clear view of their temporal relationship. Traces are same as in the CSD profile, at 0.4 and 0.5 mm depth. C: CNQX-resistant activity was reversibly blocked by 2-amino-5-phosphonovaleric acid (APV; 50 µM). Glomerular field potential is shown as an example. APV application and APV-wash are in low Mg2+ medium. Traces are averages of 5 sweeps. In this and all subsequent figures, the unequal spacing between traces in the field potential and CSD profiles is only to accommodate all waveforms, while presenting them at a size that all sinks and sources can be clearly visible. Stimulus artifacts have been reduced for the same reason.
), if the resulting current is close to zero, then sinks and sources are balanced. This suggests that the currents present in the CSD distribution are generated by neuronal elements oriented along the CSD axis, with no significant contamination by currents flowing in other directions.
where Im is the net current (a scalar quantity of dimension Amperes/cm3),
z is tissue conductivity along the z axis, and 
is the field potential. The second spatial derivative of the field potential was approximated by the formula (Freeman and Nicholson 1975)
where
z is the sampling interval, and nz is the differentiation grid. Differentiation grids of 100 and 200 µm were tested, and the latter was adopted because it produced more distinct CSD patterns.
View larger version (37K):
[in a new window]
FIG. 4.
Field potential depth profile and corresponding CSD distribution evoked by stimulation in the MCL. Spatial resolution is 50 µm. Three current sinks were present: a low-amplitude, prolonged sink (S1mcl) in the GL, with corresponding sources in the EPL, a strong, relatively brief sink (S2mcl) in the EPL and MCL, with corresponding sources extending from the MCL to the deep GCL, and another low-amplitude, prolonged sink in the MCL and IPL/GCL with corresponding sources in the EPL. Algebraic summation of all the CSD traces resulted in a current close to 0 (bottom of the CSD profile), suggesting that sinks and sources were balanced.
) as well as in more discretely laminated structures such as the cerebellum (Freeman and Nicholson 1975; Nicholson and Freeman 1975) and the hippocampus (Holsheimer 1987). Similar conclusions have been drawn from measurements in the rabbit OB (Martinez 1982). For these reasons, we assumed that differences in tissue conductivity across the OB layers will not affect significantly the CSD distributions. Thus z was considered constant (and equal to 1).
RESULTS
Abstract
Introduction
Methods
Results
Discussion
References
View larger version (23K):
[in a new window]
View larger version (23K):
[in a new window]
FIG. 5.
Laminar distribution of kainate/AMPA and NMDA receptor-mediated currents evoked by stimulation in the MCL. A1: field potential depth profile and corresponding CSD distribution in standard medium. Data are from the same slice as in Fig. 3. Spatial resolution is 100 µm; unequal spacing between traces is only to accommodate all waveforms. A2: bath application of 10 µM CNQX blocked S2mcl, while a small portion of S1mcl and S3mcl remained in CNQX. A3: In nominally 0 Mg2+ concentration the CNQX-resistant component of S1mcl was enhanced. Sources corresponding to this sink are present in the EPL. Inward currents also appeared in the EPL and at the EPL/MCL border, with corresponding sources below the MCL. These sources mask the CNQX-resistant currents of S3mcl, which can been seen more clearly in a CSD distribution from a different slice shown in C. These currents extend from the MCL-IPL border to the GCL. Corresponding sources are probably in the EPL, overlapping with the sources of the glomerular sink. Spatial resolution in C is also 100 µm. B: CNQX-resistant field potentials were nearly blocked by APV (50 µM). Glomerular field potential is shown as an example (same slice as in A). Traces in B are averages of 5 sweeps.
400 ms. The corresponding sources are in the EPL and MCL.
40 ms). However, because of the overlap of S2ON with the sources corresponding to S1ON, neither the duration nor the onset latency of S2ON could be determined reliably. The sources corresponding to S2ON extend from the MCL to the deep GCL.
). The resulting current was very small relative to the currents in the CSD distribution (Fig. 2, n = 4). This indicates that the inward currents were accounted for by approximately equal outward currents, suggesting that they were produced by neuronal elements oriented along the axis perpendicular to the laminae. Thus the CSD distribution was not significantly affected by lateral dendrite currents.
). The peak latency of the GL sink in low Mg2+ was 23.0 ± 1.1 ms (n = 4), whereas its duration was >400 ms. The sources corresponding to this glomerular sink were in the EPL and MCL (Fig. 3A3). The temporal characteristics of the EPL sink could not be measured accurately because of the overlap with sources associated with the glomerular sink. This EPL sink reversed from the MCL to GCL.
). Thus the generation of both the GL- and the EPL-, CNQX-resistant sinks required NMDA-receptor activation.
; Nowycky et al. 1981; Phillips et al. 1963) and most centrifugal inputs ramify below the MCL (McLean and Shipley 1987; Shipley et al. 1996), most of the excitatory synaptic activity should be generated by the action of neurotransmitter released from the dendrites of M/T cells onto their postsynaptic targets.
200 ms) sink (S1mcl) is present in the GL and at the GL-EPL border. The corresponding sources appear to be in the EPL. In different slices, S1mcl was consistently present in the GL, but there was some variability in regard to the presence or characteristics of this sink at the GL-EPL border, probably due to laminar width differences between slices. For example in Fig. 5A1, the trace at the GL-EPL border has a different time course than the corresponding trace in Fig. 4 or 6A1, probably because the recording site is slightly deeper into the EPL and overlaps with currents of the EPL sink. The distribution of the currents associated with S1mcl is clearer when S1mcl is enhanced in low Mg2+ and the EPL currents are reduced by CNQX (see next section).
40 ms), although overlapping outward currents that follow the S2mcl could shorten its actual duration. These small sources may correspond in part to S1mcl and to a low-amplitude, prolonged sink (S3mcl) extending from the MCL to IPL/GCL. It is also possible that these sources include inhibitory activity, and S3mcl could be the corresponding passive currents rather than an active excitatory sink (see effects of disinhibition in the last section of RESULTS).
View larger version (27K):
[in a new window]
FIG. 6.
Reduction of -aminobutyric acid-A (GABAA)-receptor-mediated inhibition enhances the glomerular sink evoked by MCL stimulation. A1: CSD distribution in standard medium. Spatial resolution 100 µm. A2: bath application of 10 µM bicuculline methchloride (BMCl) enhanced the amplitude and duration of S1mcl. A3: effects of BMCl were reduced by APV (100 µM). B: effects of APV were reversible. Glomerular field potential is shown as an example.
200 ms. The corresponding sources are clearly present in the EPL (Fig. 5, A3 and C).
View larger version (26K):
[in a new window]
FIG. 7.
Inhibitory currents evoked directly by stimulation in the MCL did not contribute significantly to the CSD distributions. A1: CSD profile in standard medium. Spatial resolution 100 µm. A2: CSD profile in the presence of 20 µM CNQX and 100 µM APV. Low-amplitude, outward currents are present in the EPL, with corresponding sources extending from the MCL to the GCL. A3: CNQX/APV-resistant currents were blocked by BMCl (10 µM). There were no currents remaining in the presence of combined CNQX, APV, and BMCl. B: effects of BMCl were reversible. Field potential shown was recorded in the superficial EPL.
DISCUSSION
Abstract
Introduction
Methods
Results
Discussion
References
), are oriented along this axis. However, the lateral dendrites of M/T cells, which extend roughly parallel to the laminae, also could carry a significant amount of current. If these currents produce significant dipoles, they could affect the laminar CSDs, in which case the one-dimensional analysis would be inappropriate. However, summation tests (Figs. 2 and 4) showed that all current sinks were balanced by corresponding sources, suggesting that these currents were produced by neuronal elements oriented along the axis perpendicular to the laminae. Thus the CSD distributions were not contaminated by currents produced by the lateral dendrites.
). 2) The inhibitory currents they generate in response to input from granule cells would have to be large, synchronous, and focal to produce dipoles. MCL stimulation could produce large and synchronous inhibitory currents in M/T cell lateral dendrites because of direct activation of granule cell dendrites. However, inhibitory synapses are present throughout the extent of the lateral dendrites (Mori 1987), an arrangement that does not favor dipole generation. 3) Action potentials generated during the spread of depolarization in the lateral dendrites (Isaacson and Strowbridge 1998) could produce large currents. However, in response to ON stimulation, such dendritic spikes are not likely to be synchronous; lack of synchrony renders action potential currents very susceptible to cancellations due to their biphasic nature and fast time course (Mitzdorf 1985). In response to MCL stimulation, dendritic action potentials could be more synchronous. However, the backpropagation of action potentials from the soma to the dendrites is very fast (see Chen et al. 1997; Isaacson and Strowbridge 1998), and could overlap with the stimulus artifact, which often lasted 4 ms. This probably also explains why the antidromic depolarization of the apical dendrites of M/T cells was not represented in the CSDs (Fig. 7A3).
). Although we do not know how many M/T cells reached spike threshold by MCL stimulation, the synaptic excitatory responses of granule cells and the glutamatergic sink evoked in the apical dendritic tufts of M/T cells (see interpretation of the CSDs in section entitled Antidromic activation of mitral/tufted cell dendrites) indicate that a significant population of both the lateral and apical dendrites of M/T cells were depolarized and released glutamate.
; Shipley et al. 1996), where S1mcl reverses. Furthermore, synaptic currents generated by JG cells are unlikely to contribute significantly to field potentials (Rall and Shepherd 1968) and CSDs because the small size of these neurons does not favor generation of significant dipoles.
). Surgical isolation of the GL from the deeper layers of the OB did not affect the NMDA component, while it caused a small reduction in the amplitude and shortened the time course of the kainate/AMPA component. These results suggested that the NMDA component was generated exclusively by glomerular currents, while the kainate/AMPA component was affected, to some extent, by granule cell currents in the EPL. The CSD analysis in the present study showed that the neuronal elements generating the glomerular synaptic currents are the apical dendritic tufts of M/T cells, as both the kainate/AMPA and the NMDA-receptor-mediated currents of the glomerular sink reversed in the EPL and MCL, where the proximal dendritic sites and the somata of M/T cells are located. A prolonged glomerular sink similar to S1ON also has been detected in the rabbit OB in vivo and has been attributed to M/T cell dendritic activity (Martinez and Freeman 1984).
) of the ON fibers; as a result, small differences in the time to reach spike threshold or in conduction velocity may produce asynchronous depolarization of the ON terminals.
; White 1972), and M/T cells express both kainate/AMPA and NMDA receptors (Gall et al. 1990; Montague et al. 1996; Petralia and Wenthold 1992; Petralia et al. 1994; Watanabe et al. 1993). However, the long duration of S1ON suggests that polysynaptic inputs also may contribute to this sink. ON stimulation excites certain types of JG cells (Bardoni et al. 1996; Heyward et al. 1997; Keller et al. 1998; Shepherd 1971; Wellis and Scott 1990). Many JG cells form synapses with the apical dendrites of M/T cells, and thus they could produce polysynaptic activity. However, most JG cells contain inhibitory or modulatory neurotransmitters (Shipley et al. 1996) and make symmetrical synapses with the dendrites of M/T cells (Pinching and Powell 1971; White 1972). It has been suggested that inhibition of M/T cells by glomerular interneurons is depolarizing (Martinez and Freeman 1984; Siklos et al. 1995). However, the late phase of the glomerular synaptic responses of M/T cells is not depolarizing inhibition because it is enhanced by GABAA antagonists (Aroniadou-Anderjaska et al. 1997). Whether some JG cells release an excitatory transmitter onto the M/T cell dendrites remains to be determined.
; Watanabe et al. 1993), which prolongs the decay of NMDA currents (Monyer et al. 1992); the synaptically depolarized M/T cell apical dendrites may release glutamate, which excites the same or neighboring dendrites of M/T cells, thus increasing the duration of inward currents; and carnosine may be coreleased with glutamate from ON terminals (Rochel and Margolis 1982) and prolong NMDA responses by chelating zinc (Margolis 1980), an antagonist of NMDA receptors (Peters et al. 1987; Westbrook and Mayer 1987).
). However, these cells do not extend their processes into the IPL and GCL, where the sources corresponding to S2ON are located. They are unlikely to contribute significantly to the currents of S2ON because the summation tests showed that these currents were balanced by the corresponding sources in the IPL/GCL, suggesting that they were generated almost exclusively by neuronal elements that extended their processes to these layers.
; Rall et al. 1966). Thus the blockade of the granule cell responses to ON stimulation by the kainate/AMPA- and NMDA-receptor antagonists does not necessarily imply that these receptors mediate the responses of granule cells because these antagonists also block the responses of M/T cells to ON stimulation. However, MCL stimulation, which evokes monosynaptic input from M/T cell lateral dendrites to granule cells, also produced kainate/AMPA- and NMDA-receptor-mediated sinks in the granule cell dendrites, indicating that both types of glutamate receptors mediate the granule cell responses to input from M/T cells. These findings are consistent with recent reports on whole cell recordings from granule cells (Isaacson and Strowbridge 1998; Schoppa et al. 1997; Trombley and Shepherd 1992; Wellis and Kauer 1994). It appears, however, that activation of NMDA receptors in granule cells requires reduction of extracellular Mg2+ (Figs. 3A and 5A) (Isaacson and Strowbridge 1998) or membrane depolarization (Isaacson and Strowbridge 1998). In vivo (Jacobson and Hamberger 1986; Wilson et al. 1996) or in slices in physiological concentrations of extracellular Mg2+ (Figs. 3A2 and 5A2), granule cell responses are mediated primarily by kainate/AMPA receptors. This is in contrast to the responses of M/T cells to ON stimulation, which consist of a significant NMDA-receptor-mediated component in physiological concentrations of Mg2+ (late phase of S1ON in Fig. 3) (Aroniadou-Anderjaska et al. 1997; Ennis et al. 1996). One reason for this difference may relate to the NR2C subunit of NMDA receptors, which is present in M/T cells but not in granule cells (Monyer et al. 1992; Watanabe et al. 1993). The presence of this subunit in the NMDA-receptor complex confers a lower sensitivity to Mg2+ (Monyer et al. 1992).
) and could be activated by MCL stimulation producing S1mcl. However, it is not known if these fibers target the apical dendrites of M/T cells, where S1mcl is generated. In addition, recent experiments in our laboratory, in slices that preserve the olfactory bulb-pyriform cortex circuitry (Puche et al. 1998), have shown that selective stimulation of the lateral olfactory tract also produces a sink with current distribution and pharmacological properties similar to those of S1mcl (Aroniadou-Anderjaska, unpublished). Thus although we cannot rule out that centrifugal fibers contribute to the generation of S1mcl when stimulation is applied in the MCL, it is unlikely that these fibers are the major input source for the generation of this sink. 2) S1mcl may be evoked by glutamate release from the apical dendrites of M/T cells. Antidromic stimulation of M/T cells, in the frog, elicits glutamatergic synaptic responses in juxtaglomerular cells (Bardoni et al. 1996), and somatic action potentials of mitral cells, in the rat, produce Ca2+ influx in their glomerular dendritic tufts (Isaacson and Strowbridge 1998), which could evoke glutamate release. Thus it is reasonable to assume that MCL stimulation triggers release of glutamate from the apical dendrites of M/T cells. Because these dendrites have glutamate receptors that mediate their responses to the ON, it is possible that glutamate released from these dendrites activates these receptors on the same or neighboring dendrites producing self-excitation of M/T cells. Self-excitation in the dendrodendritic synapses of the OB initially was proposed by Nicoll and Jahr (1982) in the turtle and recently suggested by observations in the rat (Chen and Shepherd 1997).
; Liu and Shipley 1994; Mori 1987; Orona et al. 1984) or from centrifugal fibers that could be activated by stimulation in the MCL.
; Guthrie et al. 1993; Leveteau and MacLeod 1966; Shepherd 1994; Steward et al. 1979) receiving input from olfactory receptor neurons that express the same olfactory receptor genes (Mombaerts et al. 1996; Ressler et al. 1994; Vassar et al. 1994) and by the firing pattern of the output neurons of the OB (Imamura et al. 1992; Katoh et al. 1993; Laurent 1996; Mori et al. 1992). A key factor determining the firing pattern of M/T cells is the amount and time course of synaptic excitation produced by the ON input. The present study shows that ON activation produces prolonged synaptic excitation in M/T cell apical dendrites, providing the opportunity for modulation and integration of incoming sensory information. Most of the late phase of synaptic excitation in the apical dendrites of M/T cells is mediated by NMDA receptors, a feature that can play an important role in synaptic integration on these dendrites as well as in synaptic plasticity (Ennis et al. 1998). The present results also raise the possibility that synaptic depolarization of M/T cell apical dendrites may trigger self-excitation, contributing to the long duration of excitatory activity in the GL. Whether prolonged excitation in the GL will trigger prolonged firing of M/T cells will depend significantly on the amount and timing of inhibition by granule cells. The time course of S2ON relative to S1ON suggests that, at least in slices, it takes ~7.6 ms from the time the ON input reaches the apical tufts of M/T cells (onset of S1ON) to the time full feedback and lateral inhibition by granule cells is in effect (peak of S2ON). Thus while lateral dendrites and somata of M/T cells are under inhibition, there is still excitation (late phase of S1ON) in the active glomeruli. The balance between the amount of inhibition M/T cells receive and the level and duration of excitation in the apical dendritic tufts will significantly determine the amount and pattern of their firing.
; Schoppa et al. 1997), may be determined by the level of granule cell depolarization produced by the pattern of M/T cell input and by centrifugal inputs.
| |
ACKNOWLEDGEMENTS |
|---|
We thank Drs. T. J. Teyler and A. Keller for critical review of the manuscript.
This study was supported by National Institutes of Health Grants DC-03195, DC-00347, DC-02588, and NS-36940.
| |
FOOTNOTES |
|---|
Address for reprint requests: V. Aroniadou-Anderjaska, Dept. of Anatomy and Neurobiology, HSF, Univ. of Maryland, School of Medicine, 685 W. Baltimore St., Baltimore, MD 21201.
Received 2 July 1998; accepted in final form 25 September 1998.
| |
REFERENCES |
|---|
|
Abstract
Introduction
Methods
Results
Discussion
References
|
|---|
This article has been cited by other articles:
|
|
 |
|
  E. Chaigneau, P. Tiret, J. Lecoq, M. Ducros, T. Knopfel, and S. Charpak The Relationship between Blood Flow and Neuronal Activity in the Rodent Olfactory Bulb J. Neurosci., June 13, 2007; 27(24): 6452 - 6460. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Heinbockel, K. A. Hamilton, and M. Ennis Group I Metabotropic Glutamate Receptors Are Differentially Expressed by Two Populations of Olfactory Bulb Granule Cells J Neurophysiol, April 1, 2007; 97(4): 3136 - 3141. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Laaris, A. Puche, and M. Ennis Complementary Postsynaptic Activity Patterns Elicited in Olfactory Bulb by Stimulation of Mitral/Tufted and Centrifugal Fiber Inputs to Granule Cells J Neurophysiol, January 1, 2007; 97(1): 296 - 306. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Ennis, M. Zhu, T. Heinbockel, and A. Hayar Olfactory Nerve-Evoked, Metabotropic Glutamate Receptor-Mediated Synaptic Responses in Rat Olfactory Bulb Mitral Cells J Neurophysiol, April 1, 2006; 95(4): 2233 - 2241. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Heinbockel, P. Heyward, F. Conquet, and M. Ennis Regulation of Main Olfactory Bulb Mitral Cell Excitability by Metabotropic Glutamate Receptor mGluR1 J Neurophysiol, November 1, 2004; 92(5): 3085 - 3096. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. R. Neville and L. B. Haberly Beta and Gamma Oscillations in the Olfactory System of the Urethane-Anesthetized Rat J Neurophysiol, December 1, 2003; 90(6): 3921 - 3930. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Friedman and B. W. Strowbridge Both Electrical and Chemical Synapses Mediate Fast Network Oscillations in the Olfactory Bulb J Neurophysiol, May 1, 2003; 89(5): 2601 - 2610. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. S. Isaacson and H. Vitten GABAB Receptors Inhibit Dendrodendritic Transmission in the Rat Olfactory Bulb J. Neurosci., March 15, 2003; 23(6): 2032 - 2039. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Ennis, F.-M. Zhou, K. J. Ciombor, V. Aroniadou-Anderjaska, A. Hayar, E. Borrelli, L. A. Zimmer, F. Margolis, and M. T. Shipley Dopamine D2 Receptor-Mediated Presynaptic Inhibition of Olfactory Nerve Terminals J Neurophysiol, December 1, 2001; 86(6): 2986 - 2997. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Heyward, M. Ennis, A. Keller, and M. T. Shipley Membrane Bistability in Olfactory Bulb Mitral Cells J. Neurosci., July 15, 2001; 21(14): 5311 - 5320. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. W. Margrie, B. Sakmann, and N. N. Urban Action potential propagation in mitral cell lateral dendrites is decremental and controls recurrent and lateral inhibition in the mammalian olfactory bulb PNAS, December 14, 2000; (2000) 11523098. [Abstract] [Full Text]
|
|
|
|
|
|
J. S. Isaacson Mechanisms governing dendritic gamma -aminobutyric acid (GABA) release in the rat olfactory bulb PNAS, December 14, 2000; (2000) 21445798. [Abstract] [Full Text]
|
|
|
|
|
|
V. Aroniadou-Anderjaska, F.-M. Zhou, C. A. Priest, M. Ennis, and M. T. Shipley Tonic and Synaptically Evoked Presynaptic Inhibition of Sensory Input to the Rat Olfactory Bulb Via GABAB Heteroreceptors J Neurophysiol, September 1, 2000; 84(3): 1194 - 1203. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. C. Carlson, M. T. Shipley, and A. Keller Long-Lasting Depolarizations in Mitral Cells of the Rat Olfactory Bulb J. Neurosci., March 1, 2000; 20(5): 2011 - 2021. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. Aroniadou-Anderjaska, M. Ennis, and M. T. Shipley Dendrodendritic Recurrent Excitation in Mitral Cells of the Rat Olfactory Bulb J Neurophysiol, July 1, 1999; 82(1): 489 - 494. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. W. Margrie, B. Sakmann, and N. N. Urban Action potential propagation in mitral cell lateral dendrites is decremental and controls recurrent and lateral inhibition in the mammalian olfactory bulb PNAS, January 2, 2001; 98(1): 319 - 324. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. S. Isaacson Mechanisms governing dendritic gamma -aminobutyric acid (GABA) release in the rat olfactory bulb PNAS, January 2, 2001; 98(1): 337 - 342. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
