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 (28) Citing Articles via Google Scholar Google Scholar Articles by Rothman, J. S. Articles by Manis, P. B. Search for Related Content PubMed PubMed Citation Articles by Rothman, J. S. Articles by Manis, P. B.

Differential Expression of Three Distinct Potassium Currents in the Ventral Cochlear Nucleus

The Center for Hearing Science, ¹Department of Biomedical Engineering, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205; and ²Department of Otolaryngology/Head and Neck Surgery and The Curriculum in Neurobiology, The University of North Carolina, Chapel Hill, North Carolina 27599

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
In the ventral cochlear nucleus (VCN), neurons transform information from auditory nerve fibers into a set of parallel ascending pathways, each emphasizing different aspects of the acoustic environment. Previous studies have shown that VCN neurons differ in their intrinsic electrical properties, including the K⁺ currents they express. In this study, we examine these K⁺ currents in more detail using whole cell voltage-clamp techniques on isolated VCN cells from adult guinea pigs at 22°C. Our results show a differential expression of three distinct K⁺ currents. Whereas some VCN cells express only a high-threshold delayed-rectifier-like current (I_HT), others express I_HT in combination with a fast inactivating current (I_A) and/or a slow-inactivating low-threshold current (I_LT). I_HT, I_LT, and I_A, were partially blocked by 1 mM 4-aminopyridine. In contrast, only I_LT was blocked by 10–100 nM dendrotoxin-I. A surprising finding was the wide range of levels of I_LT, suggesting I_LT is expressed as a continuum across cell types rather than modally in a particular cell type. I_A, on the other hand, appears to be expressed only in cells that show little or no I_LT, the Type I cells. Boltzmann analysis shows I_HT activates with 164 ± 12 (SE) nS peak conductance, -14.3 ± 0.7 mV half-activation, and 7.0 ± 0.5 mV slope factor. Similar analysis shows I_LT activates with 171 ± 22 nS peak conductance, -47.4 ± 1.0 mV half-activation, and 5.8 ± 0.3 mV slope factor.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
The central auditory system has a remarkable ability to extract information from the relatively simple sensory representation provided by the auditory nerve (AN). At the first central processing center, the cochlear nucleus, neurons transform the relatively uniform discharge of AN fibers into a set of parallel ascending pathways that each emphasize different features of the sensory environment. In the ventral cochlear nucleus (VCN), neurons differentially represent stimulus phase, fine temporal periodicity, envelope modulation, and stimulus intensity by their discharge patterns. These transformations are achieved by a variety of mechanisms, including exceptionally rapid synaptic receptor kinetics, dendritic filtering, coincidence detection, and specialized membrane currents in the postsynaptic neurons.

One of the major constituent cell types in the VCN is the bushy cell. Bushy cells typically have one or two apical dendrites that branch profusely within several hundred microns of their cell body. Based on their location in the cochlear nucleus, their AN innervation pattern, and their projections to the superior olivary complex, bushy cells are subclassified as being either spherical or globular. Spherical bushy cells are located in the rostral anterior VCN, receive their AN input via a few (1–3) large calycoidal synapses located on their somata (Brawer and Morest 1975; Brawer et al. 1974; Lorente de Nó 1981; Sento and Ryugo 1989), and project to the lateral superior olive. Globular bushy cells, on the other hand, are located near the AN root region, receive anywhere between 5 and 25 convergent somatic AN calycoidal synapses (Liberman 1991, 1993; Spirou et al. 1990), and project both to the medial nucleus of the trapezoid body (MNTB) and the lateral superior olive (Cant and Casseday 1986).

The second major constituent cell type in the VCN is the multipolar, or stellate cell (Brawer and Morest 1975; Brawer et al. 1974; Cant 1992; Lorente de Nó 1981). In contrast to bushy cells, stellate cells have long, sparsely branched dendrites that receive significant synaptic innervation via small bouton endings; physiological measurements suggest a minimum of five convergent inputs from AN fibers (Ferragamo et al. 1998). Rather than projecting to the superior olive complex, as bushy cells do, stellate cells project to the inferior colliculus and/or the dorsal cochlear nucleus (Adams 1979; Cant 1982; Doucet and Ryugo 1997; Ostapoff et al. 1999).

Given the heterogeneous morphology and innervation pattern of VCN neurons, it has long been postulated that different classes of VCN neurons perform distinct information processing functions (Brownell 1975; Pfeiffer 1966; Rhode and Smith 1986). VCN neurons of different morphology are in fact associated with different responses to acoustic stimuli (Feng et al. 1994; Ostapoff et al. 1994; Rhode et al. 1983; Rouiller and Ryugo 1984; Smith and Rhode 1987, 1989). Spherical and globular bushy cells, for example, respond to tone bursts with a "primary-like" and "primary-like-with-notch" response. Both cell types exhibit excellent phaselocking; in fact, globular bushy cells with convergent inputs can phaselock to low-frequency tones better than AN fibers (Joris et al. 1994). It has been suggested that these two cell types provide distinct information about the fine timing structure of complex stimuli (Shofner 1999). Such information is useful in determining formant structure and pitch. Similarly, octopus cells of the posterior VCN are thought to provide timing information to higher auditory centers, and although their innervation patterns are quite distinct from bushy cells (they receive convergent input from a large array of AN fibers tuned to different frequencies), they appear to use membrane mechanisms similar to those of bushy cells (Golding et al. 1995, 1999). Stellate cells, in contrast, respond to tone bursts with a "chopping" response; these cells presumably report information about the stimulus envelope, or low-frequency amplitude modulation of a sound, an analysis useful in identifying vowels (Frisina et al. 1985, 1990; Kim et al. 1990; Rhode 1998; Shofner 1999; Wang and Sachs 1992, 1994).

The specific complement of K⁺ currents in neurons has been shown to be tremendously important in controlling not only spike shape, but also spike rate, spike adaptation, and regularity of discharge. This is undoubtedly true for VCN neurons, which are already known to possess different complements of K⁺ channels (Manis and Marx 1991). The response of bushy cells to injected current steps, for example, is much different to that of stellate cells: whereas stellate cells respond with a regular train of action potentials, bushy cells respond with a phasic discharge of one to three action potentials (Francis and Manis 2000; Oertel 1983; Schwarz and Puil 1997; White et al. 1994; Wu and Oertel 1984). After this initial discharge of one to three action potentials, bushy cells enter a high conductance state until the applied current is terminated. The high conductance state is generated by a low-threshold, relatively non-inactivating K⁺ current (I_LT) (Manis and Marx 1991). I_LT differs from the conventional delayed rectifier found in the mammalian brain in that it has a particularly low activation voltage (-70 vs. -30 mV) and differs from other K⁺ currents active at subthreshold potentials (e.g., the transient A-type K⁺ current, I_A) in that it shows little inactivation for short test pulses. A detailed model including I_LT replicates many characteristics of VCN bushy cells, including their ability to phaselock at frequencies up to 5 kHz (Rothman et al. 1993). Experimental evidence suggests that blocking I_LT with 4-aminopyridine (4-AP) degrades the ability of these cells to phaselock (Reyes et al. 1994). Other studies suggest I_LT allows cells to act as precise coincidence detectors (Joris et al. 1994; Rathouz and Trussell 1998; Reyes et al. 1994; Rothman and Young 1996; Rothman et al. 1993).

In addition to I_LT, a high-threshold delayed-rectifier-like current (I_HT) has been characterized in both bushy and stellate cells (Manis and Marx 1991). The likely source of I_HT is the KCNC1 channel, which is highly expressed in mammalian VCN cells (Grigg et al. 2000; Perney and Kaczmarek 1997; Perney et al. 1992). This channel is also highly expressed in cells homologous in structure to spherical bushy cells, the MNTB cells (Grigg et al. 2000; Perney et al. 1992; Wang et al. 1998). It has been proposed that I_HT allows neurons to fire at high rates by providing a rapid repolarization of their action potential (Wang et al. 1998).

The present study was undertaken to more thoroughly characterize the K⁺ currents in VCN neurons and especially to provide detailed kinetic data that can be used for modeling. We present this work as a series of three papers. In this, the first paper, we describe the separation of the three different classes of K⁺ currents in VCN neurons. We show that the magnitude of I_LT appears to vary between neurons in a way that suggests a continuum of cell types rather than a collection of discrete classes. We also find a gradation of expression of I_A. In the second paper, we present kinetic analysis of the separated currents and derive equations useful for representing them mathematically. In the final paper, we present a somatic model that incorporates these currents into a single electrical compartment, thereby allowing us to elucidate the role each current plays in controlling the discharge pattern of VCN neurons. Portions of this work have been presented previously in abstract form (Rothman and Manis 1996, 1997).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
Cell isolation and whole cell recordings

Isolated VCN cells were obtained by previously published methods (Harty and Manis 1996). Briefly, young adult guinea pigs (4–12 wks old) weighing 110–340 g were anesthetized with pentobarbital and decapitated. The brain stem was quickly removed from the cranial cavity and immersed in an oxygenated dissection solution [(in mM) 112 NaCl, 5 KCl, 1 CaCl₂, 1 MgCl₂, 30 glucose, 20 PIPES, 1 kynurenic acid, and 0.5 mg/ml bovine serum albumin]. Each cochlear nucleus was blocked and cut into 300-µm-thick slices. From these slices the VCN was cut away and placed in a spinner flask containing oxygenated dissection solution (30°C) supplemented with bovine pancreatic trypsin (0.6 mg/ml). After 30 min of enzymatic treatment, VCN slices were thoroughly rinsed in enzyme-free dissection solution and allowed to incubate at room temperature for 1 h. Dissociated VCN cells were obtained by gently triturating one or two slices through a series of fire-polished Pasteur pipettes with decreasing tip diameter. The resulting cell suspension was then plated onto a petri dish coated with poly-D-lysine. Cells were allowed to settle and attach to the bottom of the dish for 15 min before starting a continuous flow of a HEPES-buffered extracellular solution [(in mM) 130 NaCl, 5 KCl, 2.5 CaCl₂, 1.3 MgCl₂, 30 glucose, and 10 HEPES].

Whole cell voltage-clamp recordings were made at room temperature (22°C) with an Axopatch 200 amplifier (Axon Instruments). Electrodes were pulled from glass capillaries (KG-33, Garner Glass), fire polished, coated with silicone elastomer (Sylgard 184; Dow Corning), and filled with a K-gluconate electrode solution [(in mM) 130 K-gluconate, 4 NaCl, 1 EGTA, 5 sucrose, 10 HEPES, and 4 Mg₂ATP]. Electrode resistance was typically 3–10 M. Access resistance (R_a) was in the range 5–20 M, 75–95% of which was compensated on-line. Cell capacitance (C_m) was also compensated on-line. Recordings were filtered at 1–5 kHz and digitized at 2–10 kHz with a 12-bit A/D converter (Digidata 1200, Axon Instruments). Current records are averages of three records except where noted.

To block specific ionic currents, pharmacological agents were often added to the HEPES-buffered extracellular solution. For example, 1 µM tetrodotoxin (TTX) and 50 µM cadmium (Cd²⁺) were routinely added to block sodium and calcium currents respectively. To this HEPES-TTX/Cd²⁺ solution, various concentrations of 4-AP or dendrotoxin-I (DTX) were sometimes added to block specific K⁺ currents.

Voltage and current corrections

Under whole cell voltage clamp, the voltage command V_c and the transmembrane voltage V_m are not identical. For the most part, this discrepancy arises from two sources: the voltage drop arising from the flow of membrane current I_m across the series resistance R_a and the charging of the membrane capacitance C_m during a step change in V_c. The first source of error was partially compensated online (75–95%) using the "correction" compensation circuitry of the Axopatch amplifier. The remaining error due to the uncompensated portion of R_a was then corrected offline, assuming the following current-voltage relation: V_m = V_c - I_mR_u, where R_u is the residual uncompensated fraction of R_a. In this paper, all figures display the estimated V_m rather than V_c, in which case the voltage traces often show small deviations at the beginning of the command step. The second source of error was also partially compensated on-line (40–95%) using the "prediction" compensation circuitry of the amplifier. There were, however, a number of cells where prediction compensation was not employed (n = 55). With no prediction compensation, V_m will respond to a step change in V_c in the following exponential manner: V_m = V_c [1 - exp(-t/_s)], where _s = R_aC_m. For the average cell in this study, _s = (13 M)(12 p F) = 160 µs, meaning V_m will have a 10–90% rise time of 340 µs and will settle to within 5% of its final value 0.5 ms after the start of the command step. For the analysis of steady-state current-voltage (I-V) relations presented in this paper, the difference in rise times is inconsequential.

Another voltage correction performed offline was the subtraction of the liquid junction potential between the pipette and the extracellular bath solutions, measured to be -12 mV.

Finally, current traces were routinely corrected for leakage current by computing I-V relations at hyperpolarized potentials, fitting a linear function to the resulting data, then subtracting estimates of the leakage current from the raw current traces at each voltage step (linear interpolation and extrapolation). Because some cells possessed I_LT and/or I_h, whereas others did not, the specified range of hyperpolarizing voltage steps was adjusted on a cell-by-cell basis. Typically, the range fell between -55 and -100 mV for Type I cells, and -75 and -100 mV for Type II cells.

Boltzmann functions

Steady-state activation of a two-state voltage-dependent conductance g can be described by the Boltzmann equation

(1)

where g_max is the maximum conductance, V the membrane potential, V_0.5 the half-activation voltage point, and k the slope parameter. Because the K⁺ currents investigated in this study showed linear instantaneous I-V relations in the form I = g(V - V_r) (from -110 to 0 mV; data not shown), Eq. 1 can be redefined as

(2)

where V_r is the reversal potential of the K⁺ current under investigation. Eq. 2, referred to as the modified Boltzmann function, proved useful in this study because I-V relations could be directly fit without converting current to conductance: an undesirable conversion since it leads to large deviations in the data near V_r. Because K⁺ channels are usually closed near their V_r, in which case V_r is not evident in steady-state I-V relations, V_r had to be predetermined from instantaneous I-V relations (i.e., tail currents), or, if that was not possible, set equal to -70 mV, the average value computed from our tail current analysis (Rothman and Manis 2003a).

Defining threshold of activation

The voltage threshold (V_th) for activation of a current is typically defined as the voltage at which a predefined fraction of a maximum conductance (g_max) has been reached. This definition, however, requires knowing g_max, which is not always possible. For example, whole cell currents in this study sometimes exceeded the upper limit of the recording amplifier (20 nA) at the highest command steps. This current could be achieved before saturation of the conductance was evident, preventing an accurate estimate of g_max via Eq. 2. V_th was therefore defined as the voltage at which the current reached 0.1 nA. This definition is desirable in that it does not require knowing g_max, is small enough to lay within the knee of conductance activation but large enough to lay above the baseline noise, and is small enough that any voltage error due to R_u is negligible. Yet, as Fig. 1 shows, this definition can be problematic. This figure shows three I-V relations of three hypothetical cells sharing the same delayed-rectifier-like conductance (inset) but with varying g_max. Because the cells share the same conductance, they should have the same activation threshold. However, when threshold is defined with respect to current, the activation threshold of each cell is different, and altogether span 5 mV. This range of threshold values could be larger, given a larger range of g_max values; however, the range chosen here, 100–220 nS, is the same range of g_max values observed in this study. Hence it must be kept in mind that a given set of V_th values are only approximations to their "true" threshold values as would be defined in conductance space. Nevertheless, these approximations are capable of delineating between conductances with >10 mV separation in half-activation points, and therefore provide an adequate means of distinguishing the K⁺ currents investigated in this study.

View larger version (19K): [in this window] [in a new window]   FIG. 1. Defining threshold of activation, Vth. Three hypothetical I-V relations computed from Eq. 2 (V0.5 = -14 mV, k = 7 mV, Vr = -70 mV). Values for gmax are shown to the right of each curve. - - -, Vth (0.1 nA). Inset: normalized Boltzmann function (Eq. 1; V0.5 = -14 mV, k = 7 mV); —, 1 and 10% activation. Hence, a 0.1-nA threshold level in current space falls roughly at 5% activation in conductance space. Parameters were chosen to mimic IHT.

Dose-response curves

Dose-response curves for current block by a given pharmacological agent were computed by measuring S as a function of the concentration of agent, where S is either the peak current measured at a given voltage or the slope of an I-V relation within a given voltage range. Dose-response curves were then normalized to control values and fit to the following logistic equation

(3)

where C is the concentration of agent of interest, n the Hill coefficient, IC₅₀ the concentration of agent producing a 50% block, and S_max 1.

Statistics

Statistical significance was assessed using a two-sided t-test for unpaired samples at the significant level (P) indicated. Significant differences are denoted with asterisks as follows: *P < 0.05, **P < 0.01, ***P < 0.001. All results are reported as means ± SE.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
Voltage-clamp recordings were obtained from 173 isolated VCN cells. Due to the mechanically disruptive nature of the dissociation technique, these cells typically had few dendritic or axonal processes. Of those processes that remained, most were shorter than 50 µm in length. Cell bodies were either round, eccentric, or teardrop-shaped, with the shortest diameter ranging 16–30 µm (average 21 ± 3 µm). There was no correlation between cell type, as defined in the following text, and cell body shape or size, except Type II cells were never teardrop shaped. Because cell diameters exceeded 16 µm in this study, it is unlikely that any recordings were made from VCN granule cells.

Given an average cell diameter of 21 µm, and a specific membrane capacitance of 0.9 µF/cm² (Gentet et al. 2000), whole cell capacitance C_m should be on average 12.5 pF, assuming a spherical soma. While this was true for our population of Type I cells, the average C_m our of Type II cells was slightly less (Table 1; 9.8 pF), most likely due to the difficulty of estimating C_m in the presence of active conductances near the resting membrane potential.

VCN neurons exhibit four characteristic voltage-clamp responses

Previously, isolated VCN cells have been categorized as either Type I or Type II, where Type II cells display I_LT at a resting membrane potential near -60 mV and Type I cells do not (Harty and Manis 1996; Manis and Marx 1991). Although the isolated cells in this study could also be categorized as either Type I or Type II, the Type I cells appeared more heterogeneous than originally reported. A number of Type I cells, for example, showed clear signs of a rapidly inactivating A-type current (I_A; 65%, see also Manis et al. 1996), whereas others appeared intermediate in type in that they shared characteristics of both Type I and Type II cells (12%). Hence, the following subclassification of Type I cells was adopted: Type I cells that displayed I_A were subclassified as Type I-t (transient), and those intermediate in character, between Type I and Type II, were subclassified as Type I-i (intermediate). The remaining Type I cells that displayed neither I_A nor I_LT were subclassified as Type I-c (classic). The methods of classification are described in more detail in the next section. Here we give a general description of each cell type.

Figure 2A1 shows a characteristic voltage-clamp response of a Type I-c cell. In HEPES-TTX/Cd²⁺ solution, this cell displayed a single macroscopic K⁺ current: a non-inactivating high-threshold current (I_HT) that activated at V > -40 mV (Fig. 2A2). Here, the name "high threshold" is used to distinguish it from I_LT, whose threshold of activation is somewhere between -70 and -58 mV (see following text). Although I_HT may undergo slow inactivation on the order of seconds, we refer to it as non-inactivating because inactivation was not apparent during the 100- to 150-ms voltage command steps used in this study.

View larger version (31K): [in this window] [in a new window]   FIG. 2. Four voltage-clamp response types of ventral cochlear nucleus (VCN) neurons. Outward currents elicited by the voltage-clamp protocol shown in A1–D1, bottom, recorded in HEPES-TTX/Cd2+ solution. Electrode capacity transients at the beginning of each step have been deleted. A1: outward currents from a Type I classic (Type I-c) cell. Only IHT is apparent. The returning command step was set to -52 mV instead of -62 mV to increase the size of the tail currents for later kinetic analysis. A2: I-V relation of the steady-state current in A1. Dashed line denotes Vth at -40 mV. B1: outward currents recorded from a Type I transient (Type I-t) cell, consisting of IHT and IA (arrow). Arrowhead denotes Ih that activated at V < -80 mV. B2: I-V relation of the non-inactivating current in B1 (Vth = -48 mV). C1: outward currents recorded from a Type II cell. Here, 5 clear signs of ILT are apparent (numbered arrows; see text for description). Inset: comparison of the activation time course of the outward current from this Type II cell (solid line) to the Type I-c cell in A1 (dotted line; step from -106 to -30 mV). The activation time course is clearly faster in the Type II cell. C2: I-V relation of the current in C1 (Vth = -61 mV). D1: outward currents recorded from a Type I intermediate (Type I-i) cell, in which case only subtle signs of ILT are apparent (arrows 2, 3, and 5). Current traces are not averaged. D2: I-V relation of the current in D1 (Vth = -51 mV).

Figure 2B1 shows a characteristic voltage-clamp response of a Type I-t cell. In HEPES-TTX/Cd²⁺ solution, this cell displayed not only I_HT, but also I_A (arrow). Using prepulse protocols, we were able to isolate I_A in Type I-t cells (Rothman and Manis 2003a). This analysis shows that I_A activates 15 mV below I_HT (see Table 2, V_0.5 values) and inactivates significantly faster than all other currents. Also apparent in the current traces in Fig. 2B1 is a hyperpolarization-activated inward current (I_h, arrowhead) that showed slow time-dependent activation at V < -80 mV. Although I_h was often found in Type I-t cells, it was found in the other cell types as well. Unfortunately, because I_h was only weakly expressed in our dissociated cells (possibly due to the enzymatic treatment), it was not possible to analyze it systematically.

View larger version (33K): [in this window] [in a new window]   FIG. 6. Comparison of I-DTX and I+DTX. A: I-V relations of IDTX in Type I cells () and Type II cells (). All currents were recorded in either 10 or 100 nM DTX. Type I-c, Type I-t, and Type I-i cells are all denoted as Type I. , the boundary of data points. Omitted from this figure is the Type II cell in Fig. 4, in which case a small amount of ILT remained in 100 nM DTX. Inset: mean and SE as a function of voltage after I-V relations were interpolated onto a common voltage scale. The 1 voltage point where a statistical difference was found is denoted with * (P = 0.04). B: Vth of I-DTX in A (—, ) and I+DTX in C (, ). - - - at -48 mV, a natural boundary between the 2 currents. C: I-V relations of I+DTX in Type I cells () and Type II cells () obtained by subtracting I-V relations of the DTX-insensitive current from I-V relations of the whole cell current recorded under control conditions. , the same as that in A. D: Boltzmann functions derived from the data in A and C (see Table 2 for parameters). Average Boltzmann functions for Type I cells are denoted as - - - and those for Type II cells as — (DTX-insensitive curves appear above -40 mV). I-V relations were interpolated to a common voltage scale before averaging. Symbols are the same as in A–C. *, significant differences between I+DTX in Type I and Type II cells. No statistical differences were found between IDTX in Type I and Type II cells. Note that the deviation between the DTX-insensitive Boltzmann functions and the data points >0 mV is due to a decrease in the number of data points at these potentials.

Figure 2D1 shows a characteristic voltage-clamp response of a Type I-i cell. This response is similar to the Type II cell in Fig. 2C1 in that the whole cell current shows fast activation kinetics (5) and slow inactivation (3). There are, however, several differences from the Type II response, including almost no steady outward current at a holding potential near -60 mV (1); barely discernable deactivating inward currents in response to voltage steps below the K⁺ reversal potential (2); tail currents that are small, showing no signs of removal of inactivation (4; no "crossing tail currents"); and a high activation threshold (Fig. 2D2, solid line; V_th = -51 mV). These similarities and differences are best explained by the presence of I_LT, but one whose magnitude is significantly less than that shown in Fig. 2C1 and, perhaps, whose activation is shifted more positive. Evidence in support of this conclusion is given in the following text, where we show dendrotoxin specifically blocks a small I_LT in this Type I-i cell, as well as in 4 other Type I-i cells.

Classification of isolated VCN neurons

As previously mentioned, isolated VCN cells have been classified as Type II if they displayed I_LT near rest, or Type I if they did not (Harty and Manis 1996; Manis and Marx 1991). The basis of these classifications was that Type II cells (putative bushy cells), which fired only one or two action potentials in response to depolarizing current pulses, possessed both I_LT and I_HT, whereas Type I cells (putative stellate cells), which fired regular trains of action potentials in response to depolarizing current pulses, possessed only I_HT. Hence, when studying K⁺ currents under voltage clamp, I_LT could be used as the delineating factor between Type I and Type II cells. In the present study, the same "low-threshold" criterion was used initially to classify isolated VCN cells: cells with visible signs of I_LT at a holding potential near -60 mV were classified as Type II, and all other cells were classified as Type I. As shown in Fig. 2C1, visible signs of I_LT at -60 mV include: a small steady outward current at the holding potential (1) and small deactivating inward currents (tail currents) in response to voltage steps below the K⁺ reversal potential (2). After visually classifying cells as Type I or Type II, two subsequent criteria were used to further subclassify Type I cells. The first criterion was the presence of I_A. In 84% of the Type I cells, I_A was readily apparent when the membrane potential was stepped positive from a holding potential near -60 mV (e.g., Fig. 2B1). In the remaining instances (16%), I_A was not apparent until the positive command steps were preceded by a hyperpolarizing prepulse to potentials below -100 mV, in which case inactivation of I_A was removed [data not shown, but see Fig. 1 of Rothman and Manis (2003a) for an example of the prepulse protocol used to isolate I_A]. Hence, in this study, prepulses were routinely used to verify the presence of I_A. If a Type I cell showed visible signs of I_A with or without a prepulse, it was classified as Type I-t. The second criterion for subclassifying Type I cells was the presence of fast activation and slow inactivation in the outward current traces (putative signs of I_LT; see Fig. 2D1); such cells were classified as Type I-i. All remaining Type I cells, which presumably possess only I_HT, were classified as Type I-c.

Of the 173 cells investigated in this study, 49 were classified as Type II, 23 as Type I-c, 80 as Type I-t, and 19 as Type I-i. Only two cells failed to fit into any category, and were therefore classified as unusual. These two cells were similar to each other in that, besides I_HT, they possessed a slowly activating, slowly inactivating K⁺ current that was quite distinct from the rapidly activating, rapidly inactivating I_A, or the rapidly activating, slowly inactivating I_LT.

4-AP does not specifically block I_LT

To quantify the expression of I_LT in VCN neurons, it was first necessary to isolate the current pharmacologically. We first attempted to use 4-AP because this drug has previously been shown to block I_LT (Manis and Marx 1991; Rathouz and Trussell 1998; Reyes et al. 1994; Zhang and Trussell 1994). However, as the results in this section indicate, 4-AP not only blocks I_LT, but I_HT and I_A as well. Thus 4-AP is not an effective tool to isolate I_LT in VCN neurons (see also Rathouz and Trussell 1998).

Figure 3A shows five I-V relations of a Type II cell when exposed to various concentrations of extracellular 4-AP (0 – 4 mM). Inspection of these I-V relations above and below -48 mV (the dividing line between I_LT and I_HT, as shown in the following text) shows that at no concentration of 4-AP was there a specific block of I_LT in this Type II cell; that is, either I_LT appeared only partially blocked (0.1 mM) or both I_LT and I_HT appeared to be significantly blocked by 4-AP (1 and 4 mM).

View larger version (23K): [in this window] [in a new window]   FIG. 3. Sensitivity of ILT and IHT to 4-aminopyridine (4-AP). A: effects of 4-AP on a Type II cell. I-V relations computed from a voltage-clamp protocol similar to that in Fig. 2C. B: dose-response curves for ILT (—; IC50 = 79 µM) and IHT (- - -; IC50 = 54 µM). For ILT (—; n = 5 Type II cells), slope conductances were computed from I-V relations, as in A, from -70 to -50 mV (linear regression), and normalized to control conditions. For IHT (; n = 14 Type I cells), slope conductances were computed from I-V relations from -40 to -20 mV and normalized to control conditions.

Because I_HT activates at V > -50 mV, we were able to quantify the effects of 4-AP on I_LT alone in five Type II cells by measuring the slope of their I-V relations, such as those in Fig. 3A, from -70 to -50 mV (S_-50/-70) at various concentrations of 4-AP (0.01, 0.1, 1, and 4 mM). These S_-50/-70 values were then normalized to control conditions and fit to Eq. 3 (Fig. 3B), yielding an IC₅₀ = 79 µM.

Unfortunately, it was not possible to quantify the effects of 4-AP on I_HT in Type II cells because I_LT coactivates with I_HT at V > -50 mV. However, it was possible to quantify the effects of 4-AP on I_HT in Type I-c and Type I-t cells because the primary K⁺ current in Type I-c cells was I_HT, and the presence of I_A in Type I-t cells was effectively eliminated by computing I-V relations after most of I_A was inactivated (t > 100 ms after the command step onset). To further ensure our analysis pertained to I_HT, and not some small I_LT, only cells with high activation threshold (V_th > -48 mV) were included in the analysis. Similar to I_LT, we computed I-V relations and slope conductance values from -40 to -20 mV as a function of 4-AP concentration and then fit these data to Eq. 3 (Fig. 3B). Results of this analysis yielded an IC₅₀ = 54 µM (n = 14). These results then suggest that it is not possible to pharmacologically separate I_LT and I_HT in VCN neurons with 4-AP, due to their similar IC₅₀ values.

Using prepulse protocols (see Rothman and Manis 2003a), we were also able to compute the sensitivity of I_A to 4-AP. Results of this analysis revealed an IC₅₀ near 1 mM (n = 5 Type I-t cells), which is considerably higher than the IC₅₀ for I_HT and I_LT.

Dendrotoxin specifically blocks I_LT

As a second attempt to isolate I_LT pharmacologically, we used Toxin-I, a dendrotoxin (DTX) from black mamba snake venom. This drug was chosen because several studies have already demonstrated its ability to specifically block I_LT (Robertson et al. 1996; Stansfeld et al. 1986), including I_LT in MNTB neurons (Brew and Forsythe 1995) and chick n. magnocellularis neurons (Rathouz and Trussell 1998). As Fig. 4 demonstrates, this is also true for I_LT in VCN neurons. In Fig. 4A, whole cell currents recorded from a Type II cell show the five characteristic signs of I_LT (numbered arrows). Computing peak current as a function of voltage resulted in an I-V relation with low activation threshold (Fig. 4D, open circles; V_th = -69 mV), consistent with the presence of I_LT. In Fig. 4B, whole cell currents recorded from the same Type II cell after bath application of 100 nM DTX no longer shows signs of I_LT. Moreover, the remaining current, referred to as I^-_DTX, resembles I_HT in both kinetics of activation (compare to Fig. 2A1) and its I-V relation (Fig. 4D, closed circles; V_th = -50 mV). In Fig. 4C, the current obtained by subtracting B from A (see legend for details), referred to as I⁺_DTX, shows all characteristic signs of I_LT. Thus in this Type II cell, I_LT has been selectively blocked by 100 nM DTX.

View larger version (29K): [in this window] [in a new window]   FIG. 4. Pharmacological separation of ILT and IHT with DTX. A: outward currents from a Type II cell bathed in HEPES-TTX/Cd2+ solution. Voltage-clamp protocol is shown above C. Numbered arrows denote 5 distinct signs of ILT, as explained in text. B: outward currents from the same Type II cell now in 100 nM DTX. Only a small ILT remained. The bulk of the remaining DTX-insensitive current (I-DTX) activated at V > -50 mV. C: DTX-sensitive (I+DTX) obtained by subtracting traces in B from A. Prior to subtraction, traces in B were linearly interpolated to the command voltages in A at each point in time. D: I-V relations of the currents in A (open circles), B (closed circles), and C (closed triangles). Y axis denotes maximum current during the command step. Solid line denotes Vth of the control current. Dotted line denotes Vth of IDTX

The sensitivity of I_LT to DTX was investigated in 12 other Type II cells. Whereas a concentration of 1 nM DTX produced a partial block of I_LT (n = 5; not shown), a concentration of 10 nM DTX (n = 2) or 100 nM DTX (n = 8) produced a near complete block of I_LT. Computing S_-50/-70 values from the I-V relations of these 12 cells resulted in the following estimates for the percent conductance block of I_LT at concentrations of 1, 10, and 100 nM DTX: 34, 95, and 97%. Fitting these data to Eq. 3 gave an IC₅₀ 1 nM.

We also investigated the sensitivity of Type I cells to DTX. As previously mentioned, cells classified as Type I show no visible signs of I_LT at -60 mV; hence, we would expect DTX to have little effect on Type I cells. However, as shown in Fig. 5, a DTX-sensitive current was evident in five of the six Type I cells shown here (cells 1–5). The first indication of the effects of DTX is the decrease in the steady outward current (insets, — and - - -), as well as the resulting I-V relations ( and , V = -60 to -20 mV). The second indication is the horizontal shift in V_th (— and - - - beneath I-V relations). For the Type I-i cells, a third line of evidence is the absence of slow inactivation in the outward current traces in DTX (compare — and - - -). These results are consistent with the presence of a small I_LT in these five Type I cells. For the one Type I cell that showed little sensitivity to DTX (cell 6; Type I-c), there was little to no change in the steady outward current from -60 to -20 mV nor any significant shift in V_th. Whether DTX was responsible for the slight change in activation kinetics () is not known.

View larger version (33K): [in this window] [in a new window]   FIG. 5. Sensitivity of Type I cells to DTX. Representative current traces and I-V relations of 6 Type I cells before and after bath application of DTX. I-V relations of the Type I-t cells represent the mean current during the last 5 ms of the command step, rather than maximum current for the other cells. I-V relations are denoted as follows: , control; , 10 nM DTX; and —, 100 nM DTX. Inset: —, control currents; - - -, currents recorded in the presence of DTX; the voltages at which these currents were recorded are denoted on the x axis ().|, Vth of the control current; , Vth of I-DTX. Scale bars apply to all current traces. , denote signs of IA. The extra current trace in cell 4 (pp x 10; multiplied by a factor of 10) is from a prepulse protocol recorded in the presence of 100 nM DTX, to show the presence of IA in this cell because it was not readily apparent in the standard protocol. , in cell 6 points to a change in kinetics of IHT that may or may not be due to DTX (traces were taken 15 min apart).

The sensitivity of I_A to 100 nM DTX was investigated in six Type I-t cells. In none of these cells was I_A blocked by 10 or 100 nM DTX, although it was not possible to properly assess changes in magnitude or kinetics. Two of these Type I-t cells are shown in Fig. 5 (cells 4 and 5). For both cells, I_A is apparent in both control and DTX conditions (). Hence, neither I_A nor I_HT are blocked by nanomolar concentrations of DTX, as I_LT is. We conclude, therefore, that DTX can be used to specifically block I_LT in VCN neurons.

Comparison of DTX-sensitive and insensitive currents

First, we address the question of whether I^-_DTX is the same in all cell types. This question is addressed in Fig. 6A where I-V relations of I^-_DTX from 9 Type I cells () and 13 Type II cells () are plotted together. As this figure shows, I-V relations of both Type I and Type II cells overlap within the region highlighted in gray. To directly compare the I-V relations, the data were interpolated onto a common voltage scale and averaged as a function of voltage. The relations for Type I cells again overlap the Type II cells (inset). Except at -15 mV, the interpolated data points are not statistically different. Similarly, V_th values of the Type I cells overlap those of Type II cells (Fig. 6B) and are not statistically different (P = 0.54). Hence these results suggest Type I and Type II cells posses similar, if not identical DTX-insensitive high-threshold currents. To further support this claim, we fit modified Boltzmann functions (Eq. 2) to the I-V relations in Fig. 6A and compared g_max, V_0.5, and k values. Results of this analysis show very similar values (Fig. 6D and Table 2, I^-_DTX), suggesting the two currents are the same.

Next, we address the question of whether I⁺_DTX is the same in all cell types. This question is addressed in Fig. 6C where I-V relations of I⁺_DTX from 9 Type I cells () and 13 Type II cells () are plotted together (see legend for methods); cell classifications were determined as described in the preceding text on the basis of visual cues alone. As this figure shows, there are now clear differences between Type I and Type II cells: I-V relations of I⁺_DTX in Type II cells are larger than those of Type I cells, and appear to activate at lower potentials. Similarly, V_th values of Type II cells are lower than those of Type I cells (Fig. 6B; < -58 mV). These differences may be due to a difference in g_max values, a difference in V_0.5 values, or both. Unfortunately, the explanation of these differences cannot be determined with the given data because several of the I-V relations of the Type I cells showed non-monotonic behavior at V > -25 mV, and therefore did not fit well to a Boltzmann function. We believe the non-monotonic I-V relations are probably due to a change in magnitude of I_HT measured between control and DTX conditions, a time that could take 15 min. Nevertheless, the I-V relations in Fig. 6C indicate at a minimum that g_max is smaller in Type I cells.

Finally, we address the question of whether I^-_DTX and I⁺_DTX can be statistically separated using V_th values. This question is addressed in Fig. 6B, where V_th values of I^-_DTX in Fig. 6A (, ) are plotted with V_th values of I⁺_DTX in Fig. 6C (, ). As this figure shows, V_th < -48 mV for I⁺_DTX, and V_th > -48 mV for I^-_DTX. Because I⁺_DTX in Type I cells is small, the line of separation is less dramatic for Type I cells ( and ) than it is for Type II cells ( and ). These results then indicate that a cell with V_th < -48 mV is likely to possess I⁺_DTX (i.e., I_LT).

Quantitative analysis reveals a differential expression of I_LT

Using the preceding DTX results, we were able to compare the expression of I_LT in our population of Type I and Type II cells. In this analysis, two measures were computed from each cell's I-V relation: V_th and S_-50/-70 (Fig. 7, inset). V_th was used because, as the above results indicate, a cell with V_th < -48 mV is likely to posses I_LT. S_-50/-70 was used because it provided an estimate of the magnitude of I_LT from -70 to -50 mV. Also, in a previous study, S_-50/-70 was shown to be larger in Type II cells than in Type I cells (23.8 ± 35.8 nS as compared to 2.0 ± 1.6 nS) (Harty and Manis 1996), suggesting a possible means of separating the two cell types quantitatively.

View larger version (26K): [in this window] [in a new window]   FIG. 7. Slope conductance vs. activation threshold. Data pooled from 171 isolated VCN cells. Long vertical line at Vth = -58 mV denotes the defined boundary between Type I and Type II cells. Short vertical line at Vth = -48 mV denotes an experimentally derived boundary between I+DTX and I-DTX (see Fig. 6B). Dashed horizontal line at S-50/-70 = 8 nS denotes a natural boundary between Type II and Type I-c cells. Encircled Type I-i cells are probably Type II cells with smaller than average ILT. Inset: methodology of analysis. Activation threshold is defined as the voltage at which point the whole cell current reaches 0.1 nA (computed by linear interpolation). Slope conductance, S-50/-70, is defined as the slope of the I-V relation from -70 to -50 mV, computed by linear regression. Cells classified as described in the text.

Results of the analysis are shown in Fig. 7, where cell types, as denoted in the figure legend, are plotted with respect to V_th and S_-50/-70. Three important points are to be noted about Fig. 7. First, there is a significant number of Type I cells with V_th < -48 mV (44%), suggesting the presence of I_LT. Ten of these Type I cells were exposed to DTX (4 shown in Fig. 5: cells 1– 4), and all 10 showed the presence of a small DTX-sensitive current that activated at V < -48 mV. The second point is that there appears to be a gradual change in both V_th and S_-50/-70 values between Type I cells (circles) and Type II cells (triangles). Only at V_th = -58 mV does there appear to be a small separation between cell types. These results suggest a gradation in magnitude of I_LT from zero (V_th > -48 mV; Type I-c and Type I-t cells) to modest (-58 < V_th < -48 mV; Type I-t and Type I-i cells) to significant (-58 mV < V_th; Type II cells). Third, Fig. 7 confirms our visual cell classification; that is, Type II cells show significant I_LT (V_th < -58 mV and S_-50/-70 > 8 nS), whereas Type I-c cells do not (V_th > -52 mV and S_-50/-70 < 8 nS). Note that 7 of the 23 Type I-c cells have V_th < -48 mV; hence, these cells may possess a small I_LT that was difficult to detect. It is also possible that these cells possess only I_HT, but one whose V_th is more negative than usual. In either case, most if not all of the K⁺ current in these seven Type I-c cells represents I_HT. Finally, the Type I-i cells appear truly intermediate in type, in that they fall between Type II and Type I-c cells with respect to both V_th (-58 to -48 mV) and S_-50/-70 values (2–17 nS). This finding again suggests the presence of a small I_LT in the Type I-i cells.

Results from this section are further summarized in Table 1, where average values of S_-50/-70 and V_th are tabulated with respect to cell type. Statistical differences in comparison to the Type I-c cells are denoted with asterisks. In summary, Type I-t cells were not significantly different from Type I-c cells for either measure, whereas Type I-i and Type II cells were significantly different from Type I-c cells.

Quantitative analysis reveals a differential expression of I_HT

In this section, we compare the steady-state activation of I_HT across cell types. This analysis shows two things. First, the voltage dependence of I_HT is similar across cell types, suggesting the same "high-threshold" K⁺ channels in VCN neurons. Second, the magnitude of I_HT is similar in Type II, Type I-i, and Type I-c cells but somewhat smaller in Type I-t cells.

First, we compare I_HT in Type I-c cells to I^-_DTX in Type I and Type II cells. Here, I^-_DTX is used since I_HT coactivates with I_LT at V > -50 mV and therefore cannot be independently examined under control conditions. Figure 8A shows the first comparison, where I-V relations of 14 Type I-c cells are plotted (). These I-V relations reflect the steady-state activation of I_HT because V_th > -48 mV for this sample of Type I-c cells, suggesting the absence of I_LT (Fig. 6B). In the background of Fig. 8A is the shaded region from Fig. 6A, which defines the boundaries of I^-_DTX in Type I and Type II cells. Hence, steady-state activation of I_HT is similar to that of I^-_DTX. Of the 14 I-V relations of Type I-c cells in Fig. 8A, 10 could be satisfactorily fit to a modified Boltzmann function (Fig. 8B), resulting in the following parameter estimates for steady-state activation of I_HT in Type I-c cells: g_max = 164.2 ± 11.5 nS, V_0.5 = -14.3 ± 0.7 mV, k = 7.0 ± 0.5 mV. Although these results resemble those of I^-_DTX in Type I and Type II cells (see Table 2), they are significantly different (Fig. 8B, inset), largely due to a difference in V_0.5: V_0.5 of I_HT sits 3 mV negative to that of I^-_DTX. The difference in V_0.5 values could be due to small approximation errors, small effects of DTX on I_HT, or the presence of small amounts of I_LT in the Type I-c cells. Regardless, the difference in V_0.5 values is small. As reported in our next paper (Rothman and Manis 2003a), there are enough kinetic similarities between I_HT and I^-_DTX to suggest they constitute the same current.

View larger version (29K): [in this window] [in a new window]   FIG. 8. Steady-state activation of IHT. A: I-V relations of the steady-state current (t > 100 ms after command step onset) for 14 Type I-c cells () and 40 Type I-t cells (). For all cells, Vth > -48 mV. , the boundaries of I-DTX in Fig. 6A. Inset: mean and SE as a function of voltage after I-V relations in A were interpolated onto a common voltage scale. Significant differences between Type I-c and Type I-t cells are denoted (*). —, the results of a simultaneous fit to a modified Boltzmann function when V0.5 and k were shared between data sets (Type I-c data: gmax = 157.2 nS, V0.5 = -15.9 mV, k = 7.6 mV; Type I-t data: gmax = 77.9 nS, V0.5 = -15.9 mV, k = 7.6 mV). B: modified Boltzmann fits (- - -) to the Type I-c data in A (). Both data and fits were normalized to Imax values, where Imax = gmax(V - Vr). Inset: mean and SE as a function of voltage after normalized I-V relations in B were interpolated onto a common voltage scale. —, average Boltzmann function of the data (V0.5 = -14.3 mV, k = 7.0 mV). For comparison, means and SE are shown for I-DTX () computed from the data in Fig. 6D (Type I and Type II cells averaged together). - - -, average Boltzmann function of I-DTX in Type I and Type II cells (V0.5 = -11.4 mV, k = 7.0 mV). *, significant differences between IHT and I-DTX. C: modified Boltzmann fits (- - -) to the data in A (, ). As in B, data and fits were normalized to Imax values. Inset: mean and SE as a function of voltage after normalized I-V relations in C were interpolated to a common voltage scale. —: average Boltzmann function for Type I-c cells in B. - - -: average Boltzmann function for Type I-t cells (V0.5 = -11.3 mV, k = 9.7 mV). *, significant differences between Type I-c and Type I-t cells.

Also in Fig. 8A are the I-V relations of 40 Type I-t cells (). Again, the I-V relations reflect the steady-state activation of I_HT because V_th > -48 mV for all cells; furthermore, the I-V relations were computed for t > 100 ms after the command step onset, at which time I_A in these cells had been inactivated. As Fig. 8A shows, data points from both Type I-c and Type I-t cells predominantly fall within the region defined by I^-_DTX. However, data from the Type I-c cells tend to fall toward the upper boundary, whereas data from the Type I-t cells tend to fall toward the lower boundary. When these I-V relations were interpolated onto a common voltage scale and averaged as a function of voltage, averages of the Type I-t cells were smaller than those of the Type I-c cells at V > -30 mV (Fig. 8A, inset); these differences were significant (P < 0.05). The difference in current magnitudes suggest Type I-c and Type I-t cells either possess two distinct I_HT, or possess the same I_HT but with different g_max values. Although the former scenario cannot be ruled out as a possibility, there are several reasons to suspect the latter scenario is true. First, average I-V relations of Type I-c and Type I-t cells are well described by the same modified Boltzmann function but with different g_max values (Fig. 8A, inset, —). Second, modified Boltzmann fits to individual I-V relations result in similar Boltzmann functions for Type I-c and Type I-t cells (Fig. 8C). Although there are differences between these Boltzmann functions, the differences do not appear large enough to suggest two distinct currents. Third, tail current analysis shows that I_HT in Type I-c cells is kinetically similar to I_HT in Type I-t cells (Rothman and Manis 2003a). Hence, I_HT in Type I-t cells is probably the same as that in Type I-c cells, but expressed in smaller magnitude.

Differential expression of I_A

Because the isolation of I_A required the use of prepulse protocols, which worked better on cells that expressed relatively large I_A (e.g., the Type I-t cell in Fig. 2B), we were not able to quantify the magnitude of I_A in our entire population of VCN cells. However, we do note here two observable instances of a differential expression of I_A. First, within our population of Type I-t cells, the magnitude of I_A was clearly different from cell to cell. In some instances (16%), I_A was so small, it was not apparent until a prepulse was used to remove its inactivation. Second, I_A was not apparent in any of our Type II cells. This was true even after I_LT was blocked with DTX (n = 13 Type II cells; DTX-insensitive currents observed with and without prepulses).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
Our results suggest that VCN neurons can be subdivided according to their K⁺ currents further than previously indicated. In total, we identified four separate cell groups: the Type I-c cells, which express a high-threshold delayed-rectifier-like current (I_HT) with neither a DTX-sensitive low-threshold current (I_LT) nor a rapidly inactivating transient A-type current (I_A); the Type I-t cells, which express both I_HT and I_A, and sometimes a weak I_LT; the Type I-i cells, which express I_HT and a weak I_LT, but no I_A; and the Type II cells, which express I_HT and a strong I_LT, but no I_A. The latter two groups appear to belong to a continuum of cells with varying levels of I_LT. We also found that only I_LT was sensitive to the peptide blocker DTX, whereas all three K⁺ currents were sensitive to 4-AP.

Limitations of the cell classification

Two obvious limitations of our cell classification are the absence of correlation to morphology and the absence of correlation to current-clamp responses. Previous studies correlated current-clamp responses with cell morphology (Wu and Oertel 1984) or voltage-clamp responses with current-clamp responses (Manis and Marx 1991). However, because we wished to obtain detailed kinetic descriptions of I_A, I_LT, and I_HT, these limitations were unavoidable for the following reasons. First, cells are far better voltage clamped in an isolated-cell preparation than in a slice preparation because most of their dendritic and axonal processes are removed. Second, obtaining voltage-clamp recordings from mature VCN neurons is more feasible in an isolated-cell preparation than in a slice preparation. Third, most amplifiers with good voltage-clamp performance have only fair current-clamp performance due to the headstage design (Magistretti et al. 1998). Nevertheless, based on results from previous in vitro studies, it can be inferred that the Type I cells reported in this study would have a Type I current-clamp response (i.e., repetitive firing in response to depolarization) (Manis and Marx 1991) and a stellate morphology (Wu and Oertel 1984), whereas the Type II cells would have a Type II current-clamp response (i.e., a phasic discharge in response to depolarization) and a bushy morphology. In the future it will be of significant interest to determine how our four physiologically defined categories parse against cell morphology.

DTX reveals I_LT in Type I cells

One interpretation of our DTX results is that as many as 47% of the Type I cells possess small amounts of I_LT (Fig. 7; -58 < V_th < -48 mV). Of these Type I cells, 54% were classified as Type I-t and 33% were classified as Type I-i. The magnitude of I_LT in Type I cells was on average six times smaller than that in Type II cells. Consequently, the difference between Type I and Type II cells is not simply the absence or presence of I_LT, as was previously supposed, but rather a difference in magnitude of I_LT. However, there is no clear boundary between cells with large and small I_LT, as Fig. 7 illustrates. The line drawn at -58 mV in this figure represents an arbitrary definition: Type II cells were defined as those cells possessing visual signs of I_LT near -60 mV, the average resting potential of VCN neurons. If the line at -58 mV is ignored, then the data reflect a near continuum of responses between -32 to -68 mV. Models of the cells using our measured kinetics (Rothman and Manis 2003b) show that this continuum of responses with increasing slope conductance values (S_-50/-70) can be accounted for by a continuum of I_LT expression. Furthermore, this continuum of I_LT can in turn generate a range of discrete current-clamp responses, including current-clamp responses intermediate in character between Type I and Type II. Cells with intermediate discharge patterns and action potential shapes have in fact been observed experimentally in the rat and gerbil VCN (Francis and Manis 2000; Schwarz and Puil 1997).

Although the data in Fig. 7 suggest VCN neurons might be viewed as a continuum of cells according to their expression of I_LT, this conclusion should be tempered by the experimental conditions under which the data was obtained. The cells are isolated, so they possess variable (but limited) amounts of dendritic membrane, which might distort the apparent current density if channels are not uniformly distributed on the cell surface and if some fraction of the dendritic membrane collapses into the somatic membrane as a result of the cell isolation. The cells were also treated with a proteolytic enzyme during the dissociation procedure, whereby channel function might have been variably compromised. Finally, adult cells in a dissociated situation are metabolically fragile, so it is possible that the variability of current density results from differences