The Influence of Skeletal Muscle on the Electrical Excitability ..., Summaries of Neuroscience

Download The Influence of Skeletal Muscle on the Electrical Excitability ... and more Summaries Neuroscience in PDF only on Docsity! The Journal of Neuroscience, August 1987, 7(8): 2412-2422 The Influence of Skeletal Muscle on the Electrical Excitability of Dorsal Root Ganglion Neurons in Culture Guo-guang Chen,” Alison E. Cole,b Amy B. MacDermott,” G. David Lange, and Jeffery L. Barker Laboratory of Neurophysiology, National Institute of Neurological and Communicative Disorders and Stroke, National Institutes of Health, Bethesda, Maryland 20892 Dorsal root ganglion (DRG) neurons from embryonic mice grown in coculture with dissociated skeletal muscle or in skeletal muscle conditioned medium (CM) showed an in- creased incidence of repetitive firing of action potentials when injected with sustained (So-100 msec) depolarizing current. This is in contrast to DRG neurons grown in mono- culture and normal medium, which exhibit such behavior far less frequently. The first action potential showed less sen- sitivity to block with TTX and more sensitivity to Ca*+ channel blockers than the subsequent action potentials. The in- creased incidence of repetitive firing occurred when CM was added after as few as 2 or as many as 22 d in culture and with as little as l-7 hr exposure to CM. This effect of CM cannot be mimicked by NGF or by coculture with cells from embryonic spinal cord (Peacock et al., 1973), can be elimi- nated by heating the CM at 56°C for 30 min, and partially reversed following short exposure to CM. These results in- dicate that skeletal muscle releases some heat-labile fac- tor(s) that can cause repetitive firing and, in addition, sig- nificant decrease in input resistance in the CM-treated neurons and a depression of the anomalous rectification, neither of which could account for the increase in repetitive firing. Sensory input transduced at specialized receptors and free nerve endings in peripheral tissues is propagated to central targets via all-or-none action potential activity in dorsal root ganglion (DRG) neurons. Variations in action potential properties of DRG neurons can be correlated with conduction velocity (Har- per and Lawson, 1985), which may in turn be associated with particular types of sensory modality. The temporal pattern of action potentials varies with functionally identifiable DRG neu- rons as well. Details of the temporal patterns vary widely among the many sensory modalities, with spike frequency, adaptation rates, sensory thresholds, and conduction velocities ranging over Received Sept. 4, 1986; revised Jan. 15, 1987; accepted Feb. 20, 1987. We wish to thank Drs. Donald E. R. Meyers and Thomas M. Jesse11 for reading an earlier version of this manuscript and making helpful suggestions. We also gratefully acknowledge Veronica Smallwood for the preparation of the cell culture. Correspondence should be addressed to Dr. A. B. MacDermott, Howard Hughes Institute, Columbia University, CPS, 722 W. 168 Street, New York, NY 10032. a On leave from Shanghai Brain Research Institute, Chinese Academy of Sci- ences. Present address: Department of Neuroscience, Albert Einstein College of Medicine of Yeshiva University, 1300 Morris Park Ave., Bronx, NY 1046 1. b Present address: Department of Neurology, Meyer 5- 109, Johns Hopkins Hos- pital, 600 N. Wolfe St., Baltimore, MD 21205. c Present address: Howard Hughes Institute, Columbia University, CPS, 722 W. 168 St., New York, NY 10032. 0270-6474/87/082412-l 1$02.00/O orders of magnitude. The influence of the sensory receptor on the intrinsic properties of the sensory neuron and the relative contributions of the receptor and neuron to the characteristic firing pattern are unclear. Embryonic and early postnatal (DRG) neurons receive both survival and neurite-promoting support from target and non- target tissues through a variety of distinguishable factors (Thoe- non and Edgar, 1986). The roles such factors play in promoting and maintaining appropriate innervation by these neurons or selective survival associated with specific sensory modalities have been widely studied. Tissue culture can be used as a way to observe the action of differentiating factors on neuronal de- velopment. Under such conditions, NGF has been found to promote selective (Yip and Johnson, 1984) or nonselective (Lindsay, 1979) survival of DRG neurons. In addition, extracts from skeletal muscle (Hsu et al., 1984), heart, liver, and brain (Lindsay and Tarbit, 1979), and ocular tissues (Skaper et al., 1982) have been reported to contain survival-promoting factors for DRG neurons in culture. Likewise, conditioned media (CM) from several sources have similar effects in culture. These in- clude CM from cultures ofglioma cells (Barde et al., 1978, 1980) and Schwann cells (Varon et al., 198 1). However, the effects of either extracts or CM from both target and nontarget tissues on the development of DRG cell membrane excitability have re- ceived little attention. One strategy to elucidate the possible role(s) played by various factors in differentiating specific sensory transduction processes involves culturing embryonic DRG cells with and without their peripheral and central targets. In monolayer culture the DRG neurons are readily accessible to detailed electrophysiological study of the effects of cellular contact and diffusible factors on the development of electrically excitable membrane properties. Thus, this preparation is suitable for investigation of the roles played by extrinsic factors on physiological properties of DRG neurons. In this study we have used intracellular recording techniques to evaluate exposure of embryonic DRG neurons to skeletal muscle, one of the peripheral target tissues. The CM from cul- tured skeletal muscle induced dramatic changes in the excit- ability of some cells, generating the capacity to fire repetitive action potentials in response to steady depolarizing currents. Materials and Methods Dissociated cell culture DRG neurons. DRG cells were dissected from 13- to 14-d-old fetal mice, treated with 0.1% trypsin (type VIII, Sigma), mechanically tritu- rated, and plated on collagen-coated 3.5 mm plastic dishes (Falcon) at a density of approximately 3.5 x lo5 neurons/dish. The cells were The Journal of Neuroscience, August 1987, 7(8) 2413 incubated at 37°C in humidified air containing 5% CO,. The plating medium contained 80% Eagle’s medium (DMEM; Gibco), 10% heat- inactivated horse serum (HS; Gibco), 10% heat-inactivated fetal calf serum (FCS; Hazelton), glucose (4.5 g/liter), and 1% glutamine (Gibco). NGF (gift from Drs. G. Guroff, Scholfield, and Rabe) was added at plating at a concentration of 10-30 r&ml. Four days after plating, the cultures were exposed to 10 mg/ml S-fluorodeoxyuridine (FuDr; gift of Hoffmann LaRoche) for 3-4 d to suppress proliferation of non-neu ronal cells. The DRG neurons from each dissection were separated into control and experimental groups. The growth medium for both groups contained the same components as plating medium, except that FCS was deleted and HS reduced to 5%. Culture media of all groups were changed twice a week. Skeletal muscle cells. Cells were obtained from thigh muscles of 19- 2 1 d mouse embryos. The muscle was minced and incubated in a dis- secting medium (99% MEM, 1% HEPES, Gibco) with 0.1% trypsin and 0.1% DNAse (type III, Sigma) at 37°C for 30 min. Soybean trypsin inhibitor (0.01%) was added to stop the trypsin activity. The solution was centrifuged, decanted, and resuspended in plating medium (80% DMEM, 10% HS, 10% FCS). The resulting cell suspension was preplated in a 100 mm dish for 1 O-20 min. Then the muscle cells were plated on collagen-coated 35 mm plastic dishes at a density of 3-5 x lo5 cells per dish. Four days after plating, the cultures were exposed to FuDr (10 mgml). At the first medium change, the cultures were incubated in growth medium (as defined above), which was subsequently changed twice a week. Medium conditioned by the muscle cells was collected, filtered through 0.45 pm filters (Millex), and then diluted 1: 1 with fresh growth medium. Collections were made from cultures from days 5-2 1. In some experiments CM was heat-inactivated at 56°C for 30 min. Feeding schedules were identical for maintenance of all experimental and control plates. Lightly plated muscle cell cultures (2-3 x 1 OS cells/ dish) were used for coculture with DRG neurons. Coculture of DRG neurons and skeletal muscle. The embryonic DRG cell suspension was plated directly onto muscle previously cultured for 7-10 d. The resulting cocultures were grown in medium containing 95% DMEM, 5% HS, and NGF (10-30 @ml). Electrophysiology Solutions. At the time of the experiments, the culture medium was drawn off and replaced by 1.5 ml Hank’s balanced salt solution (HBSS: Gibco) containing (in mM): 142 NaCl, 5.3 KCl, 2 CaCl,, 2 M&l,, 5 glucose, and 10 HEPES adjusted to pH 7.2-7.4. When divalent cations were used to eliminate Ca2+ currents, either Cd2+ or Co2+ was added to HBSS in place of Ca2+. For the Na+-free solutions, choline-Cl was isosmotically substituted for NaCl. All modifications were made keeping the tonicitv constant at 3 10 mOsm. TTX. 1 UM. was added to the HBSS as needed..Experiments were performed at room temperature (22-24°C). Electrodes and electrical recording. Intracellular and whole-cell re- cordings were made from individual DRG neurons under phase-con- trast optics on the modified stage of an inverted microscope (x200- 250). The sharp intracellular microelectrodes were filled with 3 M KC1 and had resistances of 60-100 MO. The whole-cell microelectrodes were filled with solutions containing (in mM): 140 K-gluconate, 2 M&l,. 1.1 EGTA (with 2.4 mM NaOH), 5 HEPES. This solution was ad&&d to pH 7.2-7.4 with KOH and to 3 10 mOsm with sucrose. The filled Dinettes had resistances of 5-8 MQ. Seals between the microelectrodes and the DRG cell membranes greater than 1 GQ were routinely established for whole-cell recordings. A bridge circuit was used for current-clamp re- cordings, which allowed simultaneous current injection and voltage measurement. Records of membrane potential and current were either photographed with a Grass oscilloscope camera or digitized and stored on an on-line digital computer (PDP 1 l/23 Plessy Peripheral System) for off-line analysis. Calculations of membrane electrical parameters. The membrane time constant (T) was evaluated by stimulating cells with small (0.1-O. 15 nA) hyperpolarizing current steps 60-100 msec in duration. The amplitude of the voltage response was measured, and the time at which the po- tential reached 63% of the asymptote was taken to be T. Input resistance (R,,) was calculated from the current-voltage (Z-v) relationshin obtained near resting potential; the slope of the linear part of the curve was taken as R,,. Cell diameters were estimated using an ocular micrometer, and an approximate surface area of the soma was calculated assuming spher- ical symmetry (i.e., A = 4rr2). Specific membrane resistance was then R, = AR,,. The estimated specific membrane capacitance, C,,, was de- rived from C,,, = r/R,. Population statistics Conjidence limits. Since we were studying the effects of various growth conditions in culture on physiological parameters, it was not possible to use a cell before treatment as control for that same cell after treatment. It was necessary, therefore, to use population statistics in order to make claims of significant effects. Control and experimental plates (popula- tions) were always sister cultures derived from the same dissection, The criterion for statisticallv sinnificant differences is 95% confidence limits _ I (i.e., p < 0.05). Statistical tests. We were interested in 2 types of data: continuous and enumerative. For the former (membrane notential, spike duration, etc.), we used Student’s t. For the latter (counts of cells with and without repetitive firing, etc.), we used a cz test of contingency. Results DRG neurons grown in culture will produce one or more action potentials in response to suprathreshold current pulses (Fig. 1A). The latency of the first action potential in response to a step stimulus follows the classical pattern of shortening with increas- ing current. Likewise, if there is more than one action potential, the rate of firing or the total number of action potentials in- creases with increasing current. We will take as an operational definition of repetitive firing the generation of more than one action potential during a 60-100 msec suprathreshold depolar- izing current stimulus. The probability with which one encoun- ters cells capable of generating repetitive firing varies with the culture conditions. Figure 1 shows examples of typical voltage responses from DRG neurons under current-clamp in response to the activation protocol used in most of our experiments. For small positive or negative current steps, the transmembrane potential expo- nentially approached a new level with a slow time constant, reflecting the input resistance and capacitance of the cell mem- brane (Peacock et al., 1973). With more intense negative cur- rents, the initial hyperpolarization was followed by a sag back towards resting levels due to activation of inward (anomalous) time- and voltage-dependent rectification (Mayer and West- brook, 1983). Electrical activity of DRG neurons grown in monoculture and coculture Using whole-cell recording techniques, the majority (83%; Table 1) of DRG neurons grown in culture under standard growth conditions generated single action potentials in response to stim- ulation with sustained (60-100 msec), suprathreshold depolar- izing current pulses (Fig. 1A). A minority (17%; Table 1) gen- erated 2 or more action potentials under the same conditions. In contrast, DRG neurons that had been cocultured with skeletal muscle showed high levels of repetitive action potential activity (97%; Table 1). This is a significant increase over cells recorded in the absence of muscle. An example of one such repetitively firing DRG neuron grown in coculture is shown in Figure 1B. In those cells showing repetitive firing, increasing the strength of the depolarizing current stimulus increased the rate of action potential generation. In the hyperpolarizing direction, inward rectification similar to control DRG neurons was routinely de- tected. Excitability of DRG neurons grown in muscle CM The increased excitability of the cocultured neurons could have arisen from direct contact with skeletal muscle fibers or from diffusible factors released by the muscle fibers into the medium. It was also possible that an inhibitory factor was released by 2416 Chen et al. - DRG Excitability in Culture Figure 2. Electrical excitability of it DRG neurons varies with time in cul- ture and treatment. Percentage of re- E petitively firing cells as a function of 9 0 time in culture. Data obtained using whole-cell recording at left; sharp, high- resistance microelectrode intracellular recording, right. In both groups, the percentage ofrepetitively firing cells has been separated into control, CM-treat- ed, and with whole-cell recording, heat- inactivated conditions. WHOLE CELL 80 60 6-13 14-21 22-34 6-13 14-21 22-34 Possible mechanisms underlying repetitive firing Passive membrane properties It is possible that passive membrane properties changed during treatment with CM and were partly responsible for the differ- ences between repetitive and nonrepetitive firing cells. There- fore, membrane time constant (T), input resistance (R,,), and resting membrane potential (RMP) were measured in control and CM-treated DRG neurons of various ages. The only sig- nificant difference in these parameters was a decrease in input resistance between normal and CM-treated neurons. RMP was near - 50 mV for DRG neurons in all the treatment conditions, as measured by either whole-cell or high-resistance microelec- trodes (Table 3). There were also no significant differences be- tween the 2 treatment groups for 7. There were significant differences in the average R,, and 7 values estimated in whole-cell recordings versus intracellular recordings using high-resistance electrodes. The average Ri, val- ues for control cells using both methods were 2 15 and 39 MQ, respectively, while average T values were 10.7 and 2.7 msec, respectively. Similar measurements were made under whole- cell recording conditions only (Table 1) for neurons grown in coculture and matched controls. Specific membrane capacitance (C,) was derived using the population average of estimated somal surface area and Ri, to obtain a value of specific membrane resistance that was then divided into the average T for each treatment condition. Mem- brane capacitance was normalized in this way to permit com- parison between these and other studies (see Discussion). How- ever, independent of the manner in which membrane capacitance was expressed, it did not change with treatment or with method HIGH RESISTANCE n Control Cond. Med. Heat Inact. TIME (DAYS IN CULTURE) of recording. Thus, essentially all of the difference between val- ues of 7 determined from whole-cell recording and high-resis- tance microelectrode intracellular recording can be accounted for by the difference in R,,. Active membrane properties Active membrane properties were also examined in control and CM-treated cells of all ages. No significant differences were found in action potential amplitude, spike duration, or threshold be- tween control and CM-treated DRG neurons (see Table 2). A brief plateau on the falling phase of the spike was observed in most DRG neurons (Fig. 1). No differences in action potentials were found between groups of neurons grown for different lengths of time in culture. The membrane properties of the DRG neurons grown in co- culture with skeletal muscle were compared with control cells. No differences were found in the passive and active membrane properties monitored compared to the control cells, except for an increased spike duration in the cocultured DRG neurons. The average spike duration of the DRG neurons in the coculture was 5.0 msec, about twice that of the control cells (Table 1). In the majority of these experiments, the increased tendency to fire repetitively was monitored as a shift in the behavior of a population of neurons. Thus, it is important to note that although treatment with CM caused an increase in the per- centage of neurons showing repetitive firing, some cells in the untreated control group fired repetitively while some neurons in the CM-treated group did not. We compared the range of membrane properties expressed under the various treatment conditions and looked for systematic differences between treat- ment groups or between single versus repetitively firing neurons The Journal of Neuroscience, August 1987, i’(8) 2417 B A Control neurons mV CM-treated neurons mV -0.6 10 0.4 -0.6 lo* 0.4 nA nA -Cc Single firing -+-+ Repetitive firing Figure 3. Current-voltage relations of control and CM-treated DRG neurons. Current-voltage relationships of DRG neurons from 4 dissections (15-2 1 d in culture). A, Current-voltage data from 12 control DRG neurons showing a small linear region near the resting potential and membrane rectification on either side. B, Current-voltage data from 13 neurons in matched cultures from the same dissections treated with CM. These current- voltage curves also show rectification; however, the slope of the inward rectification is less flattened for many of the CM-treated cells relative to the controls. In both A and B, jilled circles represent neurons that fired only single action potentials andfilled triangles represent neurons that fired repetitive action potentials. independent of treatment. Data recorded from neurons of the same dissection were compared in order to minimize variability. Figure 3 shows current-voltage relationships of 25 neurons from 4 dissections recorded over several days. Figure 3A has data from 12 control or untreated neurons 15-2 1 d after plating. There is a small linear region of each curve near the resting potential. Regions of both outward and inward rectification are apparent for both single and repetitively firing neurons. Data from 13 cells treated with CM in matched cultures are given in Figure 3B. These current-voltage relationships also show both outward and inward rectification bracketing a linear region sim- ilar to the controls. However, the slope of the inward rectifi- cation is less flattened for many of the CM-treated cells relative to the control cells in Figure 3A. This suggests that less inward rectification is activated under current clamp in CM-treated neurons. Cells that fire both repetitively and singly are included in this group. Cocultured neurons as a group showed less anom- alous rectification than controls (not shown), similar to the CM- treated neurons. One hypothesis for the lack of repetitive firing in control neurons is that inward rectification clamps the membrane po- tential such that a second spike cannot be generated. This pos- sibility was tested by applying 1 mM CsCl to the soma of the DRG neurons to abolish inward rectification (Mayer and West- Table 3. Passive membrane properties of DRG neurons recorded under various conditions Control Whole-cell recording Intracellular recording Resting Membrane Estimated Resting Membrane Estimated membrane time Input specific membrane time Input specific potential constant resistance capacitance potential constant resistance capacitance @VI (msec) (Mfl) (pF/cm2) (mv) (msec) (MQ) (pF/cm*) -52.2 k 5.8 10.69 + 5.3 215.38 +- 138.6 2.1?/0.9* -55.31 + 6.3 2.68 -t 1.4 39.20 k 21.3 2.3t/l.O* (89) (89) (65) (121) (114) (89) Cm-treated -49.8 f 8.0 11.14 -t 5.1 179.13 k 117.6 2.8t/1.2* -54.02 + 5.2 3.23 k 1.7 45.69 k 17.4 2.7?/1.2* (90) (90) (73) (51) (44) (34) Heat-inactivated -48.5 zk 8.0 11.49 + 5.5 215.59 f 108.6 3.0t/1.3* CM (22) (22) (21) All values except specific capacitance are means t SD. Values in parentheses are number of cells recorded. Estimated specific capacitance was calculated using the mean area and the mean time constant for each group. Specific capacitance was calculated for rho = 0.0 (t) and rho = 1.3 (*). See Discussion and Brown et al. (198 1) for further details. There are 4 instances of significant difference between whole-cell and intracellular recording. These occur in both measurements of time constant and both measurements of input resistance. The change in time constant is probably due to changes in input resistance and is, therefore, not an independent effect. There is also a significant difference between control and CM-treated cells in input resistance using whole-cell recording. Note that although not statistically significant, the effect is opposite using intracellular recording. Note also that the effect seems to be abolished by heat inactivation. 2418 Chen et al. - DRG Excitability in Culture Normal bath solution 1 mM CsCl f 1 Wash r- -I 50 mV 2.0 nA 25 mS Figure 4. Extracellularly applied 1 rn~ CsCl blocks inward rectification but does not induce repetitive firing. A series of current steps was applied to a DRG neuron before, during, and after 1 mM CsCl was pressure-applied to the soma. Responses to depolarizing current steps were unchanged by the treatment, while the inward rectification was substantially blocked during CsCl application. brook, 1983). If this mechanism were important, a nonrepeti- tively firing cell might be transformed to a repetitively firing cell by applying Cs+ ions. However, as shown in Figure 4, while the inward rectification was totally abolished by this treatment, there was no change in the tendency of the neurons to fire repetitively (n = 6). Threshold properties of the first and subsequent action potentials Comparison of parameters of the initial action potential elicited from control and CM-treated neurons during a sustained de- polarizing current step revealed no significant differences (Table 2). However, the first action potential in cocultured neurons had significantly increased duration relative to controls. The other significant physiological action of coculture and CM on DRG neurons was to lower the threshold for repetitive firing from an essentially infinite value to a finite value; that is, nonrepetitively firing neurons could not be induced to generate a second action potential even when 3 or 4 times the rheobase (threshold cur- rent) for the first action potential was injected. Neurons that did fire repetitively showed increased numbers of action potentials with increased depolarizing current. This observation suggested the possibility of important differences between the first and subsequent action potentials recorded from any given cell. Two approaches were used to evaluate several parameters of the repetitive firing in these cells. We first determined the distribution of current intensities required to elicit repetitive action potential activity. Neurons in both populations that exhibited repetitive firing were selected. The results are illustrated in Figure 5, which is a cumulative distribution function of the percentage of neurons above thresh- old for repetitive firing at a given depolarizing current intensity. The distributions for control and CM-treated neurons are not very different. The second approach was to plot the number of action po- tentials generated as a function of current intensity, as shown in Figure 6. Control data are plotted for 7 neurons and are indicated by dotted lines. These data are scattered over the entire range of current intensities up to 1 nA. The data from 15 CM- treated neurons are indicated by the solid lines. The latter show heavy clustering below 0.5 nA, even for the highest frequencies achieved, except for 2 cells that fall in a much higher threshold range. In general, the slope of the frequency versus current in- tensity curves for the CM-treated cells is steeper than for control cells. The Journal of Neuroscience, August 1987, 7(E) 2421 atively low concentrations of NGF were included in all cultures (controls, CM-treated, and skeletal muscle coculture) as a nec- essary trophic factor. It is possible that the CM from the em- bryonic muscle cultures was adding significantly to the levels of NGF and that this NGF supplement produced the increased excitability. However, when higher doses of NGF were added and compared with the action of CM, no obvious relationship between DRG excitability and higher doses of NGF was ob- served. Therefore, it is unlikely that NGF is the heat-labile factor associated with the expression of repetitive firing. Although the identity of the material(s) responsible for increased repetitive firing is still unknown, the heat lability does seem to point to a protein or peptide. An important issue is whether the increase in repetitive firing with exposure to CM is due to direct or indirect action on DRG neurons. For example, if the CM nonselectively increased neu- ronal survival, the resulting increase in neuronal density could change the rate of development of excitability, as can occur with spinal cord neurons in culture (Westbrook and Brenneman, 1984). Alternatively, CM could promote selective survival of a population of DRG neurons with a greater likelihood of firing repetitively in response to depolarizing current. Another indirect action could be through the background cells, which might re- spond to CM and release substances that, in turn, affect excit- ability of the neurons. The rapid onset of the excitability increase following exposure to CM (l-7 hr) argues against a slow, trophic action of CM, including selection of subpopulations of DRG neurons. Physiological consequences of skeletal muscle CM Passive membrane properties The only significant action of the CM on the passive membrane properties of the DRG neurons was a decrease in input resistance at the soma. This increased leak would tend to prevent firing of action potentials since the threshold for action potential firing is directly related to the leak. Therefore, there was no action of CM on passive membrane properties sufficient to produce the observed changes in firing pattern. However, it is interesting to note that the estimates of the specific membrane capacitance (C,) of the DRG neurons listed in Table 1 are higher than those Yoshida et al., 1978; Fukuda and Kameyama, 1980; F&on, 1986). Thus, the relative contribution of these currents to the depolarized phase of the action potential may be a means of identifying cell type or subpopulation, as well as level of de- velopment. It has been observed that younger neurons (embryonic and early postnatal) have long duration action potentials, with both a TTX-resistant sodium conductance and a calcium conduc- tance contributing to the depolarizing phase. In maturing DRG neurons, TTX sensitivity develops in a subpopulation of the cells with an associated action potential with decreased duration. Such action potentials are associated with A type, fast-con- ducting neurons (Yoshida and Matsuda, 1979; Harper and Law- son, 1985; Fulton 1986). Transition in culture from minimal TTX sensitivity in all cells to some cells showing shorter-du- ration action potentials with TTX sensitivity can also be seen after very long times in culture (> 100 d) with embryonic DRG neurons (Matsuda et al., 1978). In our experiments, there was no correlation between cell size and action potential duration (unpublished observation), suggesting no tendency to show physiological signs of maturation. This is consistent with the embryonic source of neurons and the relatively short times in culture (~60 d). The duration of the action potential of DRG neurons in co- culture with skeletal muscle was longer than in control and CM- treated cells. This might be caused by the changes of the inward CaZ+ current and/or outward K+ currents. A decrease of the voltage- (or calcium-) dependent outward current could underlie both the change of duration of the action potential as well as the increased tendency to fire repetitively. Blockade of outward current by 4-aminopyridine or tetraethylammonium chloride produces repetitive firing and spike broadening in these neurons (unpublished observations). Ionic components of the first and subsequent action potentials TTX alone was never totally effective in blocking the first action potential. The minimal effect of zero sodium applied by per- fusion pipette can be partially accounted for by the difficulty in producing a true low-sodium condition with this method of application. Zero sodium in the bath, however, was still not sufficient to block the first action potential, suggesting an im- reported by other authors (Peacock et al., 1973; Brown et al., 198 1). We assumed a spherical, net&e-free anatomy (i.e., ratio of neurite input conductance to soma input conductance, r = 0). The electrotonic structure and specific membrane properties of the mouse DRG neurons in tissue culture have been well documented by Brown et al. (1981). According to their calcu- lation, the mean value of r was 1.3; if we use this value, the estimates of C, of our DRG neurons would be in the range of 0.9-l .3 FF/cm*, which agrees with their results. portant calcium component. When recorded in zero sodium Ringer, the remaining action potential was depressed or blocked by Co2+ or Cd2+. This is consistent with other observations of action potential ionic dependence in DRG neurons (Dichter and Fischbach, 1977; Ransom and Holz, 1977; Matsuda et al., 1978; Yoshida et al., 1978). In contrast, the secondary or subsequent action potentials were insensitive to Cd2+ or Co2+ and extremely sensitive to either TTX or zero sodium, independent of the mode of application. This is similar to the observations of Yoshida et al. (1978) and Fulton (1986). TTX and zero sodium were also very effective Action potential duration Correlation between cell size, conduction velocity, and action when applied to the neurites at sites distant from the soma. potential duration could be a physiological means of classifying These data may suggest that the site of initiation of the subse- subpopulations of mature DRG neurons (Yoshida and Matsuda, quent action potentials is different from that of the first action 1979; Harper and Lawson, 1985). Furthermore, these param- potentials. It is also interesting to note the differential sensitivity eters could be developmentally regulated (Yoshida and Matsu- of the first and subsequent action potentials to blockers, even da, 1979; Fulton, 1986). Variation of action potential duration though both have shoulders on the repolarizing phase. Since the in DRGs has been attributed to combinations of 2 or 3 empir- shoulder on the falling phase of the action potential is associated ically separate currents: TTX-sensitive sodium current, TTX- with calcium conductance (Dichter and Fischbach, 1977) its resistant sodium current, and calcium current (Dichter and appearance on the TTX-sensitive secondary action potentials Fischbach, 1977; Ransom and Holz, 1977; Matsuda et al., 1978; may be related to the somal location of the recording electrode. 2422 Chen et al. - DRG Excitability in Culture The site of spike initiation may be neuritic, but as it invades the soma, calcium conductance is activated. The TTX and sodium sensitivity of the subsequent action potentials implies a requirement of a minimum number of so- dium channels to allow the membrane potential to surpass threshold for repetitive firing (Hodgkin and Huxley, 1952). It has been shown for spinal cord neurons that over early times in culture there is an increase in the number of sodium channels, which can account for the increased rate of rise of the action potential and is also consistent with the decreased threshold and increase in repetitive firing (MacDermott and Westbrook, 1986). In our experiments, the increased tendency to fire repetitively with time in culture may partially reflect the increase in neurite outgrowth and associated appearance of sodium channels. An increase in repetitive firing in chick DRG neurons with time in culture has also been noted by Handa (1977). Although a minimum density of sodium channels on the neu- rites is a prerequisite for repetitive firing and thus for the effect of CM, sodium channels are not necessarily the site of action of CM. Indeed, our data require a more complex explanation than simple lowering of the threshold for action potentials, since the threshold for single spikes is widely distributed regardless of treatment and propensity for repetitive firing. Likewise, the threshold for the production of second and subsequent spikes is not a simple monotonic function of the threshold for the production of the first spike. Summary The physiological properties of these embryonic mouse DRG neurons grown in tissue culture, including the absence of cor- relation between cell diameter and action potential duration, are consistent with the immature state of these neurons. We have shown that at this potentially formative time, these cells show sustained physiological alteration in the presence of em- bryonic skeletal muscle or the associated CM. The specificity of the action of skeletal muscle compared to the action of other peripheral target tissues remains to be tested. However, the neuritic site of the direct or indirect enhancement of TTX- sensitive sodium channel activation is particularly interesting since changes in neuritic properties might be more relevant to the activity of the maturing neurons in vivo. The CM-induced presence of neuritic action potentials in response to somal stim- ulation suggests that the response of distal neurites to normal sensory input in vivo could also be considerably modified. References Barde, Y. A., R. M. Lindsay, D. Monard, and H. Thoenon (1978) New factor released by cultured glioma cells supporting survival and growth of sensory neurons. Nature 274: 8 18. Barde, Y. A., D. Edgar, and H. Thoenon (1980) Sensory neurons in culture: Changing requirements for survival factors during embryonic development. Proc. Natl. Acad. Sci. USA 77: 1199-1203. Brown, T. H., D. H. Perkel, J. C. Norris, and J. H. Peacock (198 1) Electrotonic structure and specific membrane properties of mouse dorsal root ganglion neurons. J. Neurophysiol. 45: l-l 5. Dichter, M. A., and G. D. Fischbach (1977) The action potential of chick dorsal root ganglion neurons maintained in cell culture. J. Phys- iol. (Lond.) 267: 281-298. Fukuda, J., and M. Kameyama (1980) Tetrodotoxin-sensitive and tetrodotoxin-resistant sodium channels in tissue-cultured spinal gan- glion neurons from adult mammals. Brain Res. 182: 19 1-197. Fulton, B. P. (1986) Postnatal changes in active membrane properties of rat dorsal root ganglion neurons. J. Physiol. (Lond.) 377: 80~. Handa, Y. (1977) Morphological and electrophysiological changes in cultured spinal ganglion cells during development. Tohoku J. Exp. Med. 121: 13-25. Harper, A. A., and S. N. Lawson (1985) Electrical properties of rat dorsal root ganglion neurons with different peripheral nerve conduc- tion velocities. J. Physiol. (Lond.) 359: 47-63. Hodgkin, A. L., and A. F. Huxley (1952) Currents carried by sodium and potassium ions through the membrane of giant axon of Loligo. J. Physiol. (Lond.) 116: 449-472. Hsu, L., D. Natyzak, and G. L. Trupin (1984) Neurotrophic effects of skeletal muscle fractions on neurite development. Muscle Nerve 7: 211-217. Hulst, J. R., and M. Bennett (1986) Motoneurone survival factor produced by muscle increases in parallel with messenger RNA se- quences homologous to NGF-cDNA. Dev. Brain Res. 25: 153-l 56. Lindsay, R. M. (1979) Adult rat brain astrocytes support survival of both NGF-dependent and NGF-insensitive neurons. Nature 282: 80- 82. Lindsay, R. M., and J. Tarbit (1979) Developmentally regulated in- duction of neurite outgrowth from immature chick sensory neurones (DRG) by homogenates of avian or mammalian heart, liver and brain. Neurosci. Lett. 12: 195-200. MacDermott, A. B., and G. L. Westbrook (1986) Early development of voltage-dependent sodium currents in cultured mouse spinal cord neurons. Dev. Biol. 113: 3 17-326. Matsuda, Y., S. Yoshida, and T. Yonezawa (1978) Tetrodotoxin sen- sitivity and Ca component of action potentials of mouse dorsal root ganglion cells cultured in vitro. Brain Res. 154: 69-82. Mayer, M. L., and G. L. Westbrook (1983) A voltage-clamp analysis of inward (anomalous) rectification in mouse spinal sensory ganglion neurones. J. Physiol. (Lond.) 340: 19-45. Peacock, J. H., P. G. Nelson, and M. W. Goldstone (1973) Electro- physiologic study of cultured neurons dissociated from spinal cord and dorsal root ganglia of fetal mice. Dev. Biol. 30: 137-l 52. Ransom, B. R., and R. W. Holz (1977) Ionic determinants of excit- ability in cultured mouse dorsal ganglion and spinal cord cells. Brain Res. 136: 445-453. Skaper, S. D., I. Selak, and S. Varon (1982) Molecular requirements for survival of cultured avian and rodent dorsal root ganglionic neu- rons responding to different trophic factors. J. Neurosci. Res. 8: 251-261. Thoenon, H., and D. Edgar (1986) Neurotrophic factors. Science 229: 238-242. Varon, S., S. D. Skaper, and M. Manthorpe (198 1) Trophic activities for dorsal root and sympathetic ganglionic neurons in media condi- tioned by Schwann and other peripheral cells. Dev. Brain Res. 1: 73- 87. Westbrook, G. L., and D. E. Brenneman (1984) The development of spontaneous electrical activity in spinal cord cultures. In Develop- mental Neuroscience: Physiological, Pharmacologicaland ClinicalAs- pects, F. Caciagli, E. Giacobini, and R. Paoletti, eds., pp. 11-17, Elsevier, Amsterdam. Yip, H. K., and E. M. Johnson, Jr. (1984) Developing dorsal root ganglion neurons require trophic support from their central processes: Evidence for a role of retrogradely transported nerve growth factor from central nervous system to the periphery. Proc. Natl. Acad. Sci. USA 81: 6245-6249. Yoshida, S., and Y. Matsuda (1979) Studies on sensory neurons of the mouse with intracellular-recording and horseradish peroxidase injection techniques. J. Neurophysiol. 42: 1134-l 145. Yoshida, S., Y. Matsuda, and A. Samejima (1978) Tetrodotoxin-re- sistant sodium and calcium components of action potentials in dorsal root ganglion cells of the adult mouse. J. Neurophysiol. 41: 1096- 1106.