0306-4492/92$5.00+ 0.00

Cotnp.Biochem.Physiol. Vol. lOlC, No. 1, PP. 163-174,1992 Printed in Great Britain

0

1991 Pergamon Press plc

HELIX ASPERSA NEURONS MAINTAIN VIGOROUS ELECTRICAL ACTIVITY WHEN CO-CULTURED WITH INTACT H. ASPERSA GANGLIA SEEMAK. TIWARI and MICHAEL L. WOODRUFF* Department of Chemistry and Biochemistry, Southern Illinois University, Carbondale, U.S.A. Fax: 618 4536408 (Telephone: 618 453-6438) (Received 7 March 1991) Abstract-l. Identified neurons from the right parietal lobe of the circumoesophageal ganglion of adult land snails, Helix aspersa, were isolated and placed in primary cell culture. 2. The individual neurons were removed from the right parietal lobe by microdissection without the aid of exogenous enzymes and plated on poly-l-lysine coated coverslip in normal Helix Ringer’s solution conditioned with Helix circumoesphageal ganglia. The neurons become firmly attached to the substrate and begin to extend processes within 6 hr. 3. The cells show a dramatic increase in electrical activity when they are co-cultured with intact Helix circumoesophageal ganglia (1 ganglion/ml). Co-cultured neurons have resting potentials between -45 and - 110 mV, comparable to in situ, show spontaneous action potential firing and respond to putative neurotransmitters. Heat-treated conditioned media was inactive. 4. When dopamine or serotonin (5-HT) are added, right parietal beating neurons, Fl4-F16 and F29-F32, (i) show a suppression of action potentials occurrence; (ii) a decrease in action potential amplitude and duration, and (iii) an increase in after-hyperpolarization. 5. Bight parietal bursting neuron Fl in culture fire action potentials only in a beating mode. Dopamine addition to Fl suppresses action potential occurrence and causes an increase in action potential after-hyperpolarization, but there is only a small decrease in duration of action potentials and no significant change in action potential amplitude. S-HT addition to Fl increasesthe occurrence of action

potentials with little or no change in action potential shape. 6. This primary cell culture method is an efficientsystem for doing biochemical and eleetrophysiological studies on individual, identified neurons.

INTRODUCTION Our understanding of nerve cell physiology and the underlying biochemical basis for nerve cell behavior has advanced rapidly because of elegant studies that have used molluscan nerve cells. Recent research with Loligo, Aplysia, Helisoma, Lymnaea and Helix has combined voltage and current measurements with (i) additions of exogenous transmitters and pharmacological agents and (ii) microinjections or internal perfusions of enzymes or other proteins and small effector molecules (Braha et al., 1990; Chad and Eckert, Kramer et al., 1988; Kramer and Levitan, 1988; Kaczmarek et al., 1980; Levitan and Levitan, 1988a,b; Llinas et al., 1985; Piomelli et al., 1987; Sacktor and Schwartz, 1990; Sawada et al., 1989; Sawada et al., 1989: for reviews, see Gerschenfeld et al., 1990; Levitan, 1985; Levitan, 1988). These

experiments have begun to give us insight into the molecular mechanisms underlying transmitter-dependent modulations of excitable membranes. The power of these techniques would be increased if the biochemical effects of adding the exogenous agents could also be measured reliably. The agent-dependent changes in physiology could then be related more directly into induced changes in the level of small molecules, (e.g. cyclic nucleotides, polyphosphoryl*Author to whom correspondence should be sent.

ated inositol, arachidonic acid metabolites) and protein modifications (e.g. phosphorylations). Unfortunately, neurons in situ have complex relationships with other neurons, target cells, input cells and associative cells, such as glia and fibroblasts, and it is difficult to assign measured changes in biochemistry to the cell whose physiology is being monitored. When cells are placed in primary cell culture the synaptic connections between the neurons are broken and the glia and other mobile cells (e.g. fibroblasts) move away from the individual, sessile neurons. In primary cell culture, then, the situation is much more simplified and relating chemistry to physiological changes becomes an easier task. The analytical power of using cell culture techniques depends on the ability to make sensitive biochemical measurements on healthy isolated cells and the extent to which neurons in culture have physiological and biochemical characteristics displayed by the neuron in the intact ganglion. In this report we demonstrate a technique for the primary cell culture of neurons isolated from the central nervous system of the land snail Helix aspersa that results in neurons with high resting membrane potentials and characteristic electrical responses to putative neurotransmitters. We have also been able to measure protein phosphorylations in the primary culture neurons (Tiwari et al., 1989) and this will be the topic of a subsequent report. 163

164

SEEMA K. TIWARIand MICIUEI-L. WOODRUFF

Helix aspersa are easy to maintain in the laboratory and the physiological and biochemical properties of many H. aspersa neurons have been studied by several research groups (Bokisch and Walker, 1986; Brown et al., 1981; Colombaioni et al., 1985; Deterre et a(., 1981; Gerschenfeld et al., 1986; Harris-Warrick et al., 1988; Kerkut et al., 1975; Lee et al., 1978; Paupardin-Tritsch et al., 1985). Specific Helix neurons can be identified and isolated for study in culture (Green et al., 1990; Marom and Dagan, 1987). The culture technique that we have developed uses a mechanical dissection to isolate identified neurons and culture solutions that contain heat-sensitive factors, derived from the HeEix nervous system. These changes from previous methods have allowed us to maintain robust, differentiated cells in culture for 7-10 days.

Helix Ringer’s solution” was prepared fresh for each experiment. After the 1 hr incubation with “conditioned Helix Ringer’s solution” the very thin, tightly adhering transparent connective tissue sheath that encapsulates the neurons was opened using fine forceps. Care was taken not to touch the neurons directly. In this laboratory the removal of the connective tissue sheaths is normally aided by a lOmin, room temperature, treatment of ganglia with trypsin at 2.5 mg/ml. Better results were obtained in the culture experiments by omitting this protease treatment (see Results section). We believe that the 1 hr incubation prior to “desheathing” and the 1 hr incubation before ‘“decapsulating” allow endogenous proteases time to act on the connective tissue. Both the thick and the thin sheaths seem less tough after the incubations. Large (older) animals (shells more than 2.0 cm laterally) have thicker ganglionic connective tissue sheaths and we avoided using them.

Cell isolation

MATERIALSAND METHODS Adult Helix aspersa were obtained from Connecticut

Valley Biologicals (Stamford, MA) and kept in a hibernating state until they were used in an experiment. The day before or the morning of the experiments snails were selected and then wetted so that they would become active. Medium size, adult snails were used in all the experiments {shells 1-2 cm laterally). Each animal was sacrificed by severing the head and neck (containing the circumoesophageal ganglion) from the body and shell with a clean, sharp, single-edged razor blade. Each head/neck region was pinned onto paraffin wax in a dissection dish and the entire circumoesophageal ganglion was quickly dissected free of the surrounding tissues and placed in a 35 mm plastic Petri dish with 3 ml of normal Hefix Ringer’s solution: 80 mM NaCl, 4 mM KCl, 7 mM CaCl,, 2.5 mM MgCl,, 0.1 mM NaHCO, , 1OmM glucose and 1OmM HEFES (N-2hydroxyethyl piperazine-N’-2-ethane sulfonic acid), pH 7.2. Under a dissecting microscope, all non-nervous system tissues (e.g. glands) were cut from each ganglion with fine scissors and then the ganglia were transferred to a 35 mm Petri dish containing 3 ml normal Helix Ringer’s solution with 500 units/ml ~n~cillin/streptomy~n (G&co, Grand Island, NY) plus 0.5 mg/ml streptomycin (Sigma Chemical Co., St. Louis, MO). Connectives to the ganglia were not cut excessively close-they were left 0.5-l.Omm in length. All subsequent dissection, incubation and cell culture solutions contained the antibiotics at the above concentrations and were sterile through 0.22 pm syringe filters (Vanex, Vangard International, Neptune, NJ) prior to use. All solutions were made with > 10 MR Mini-Q filtered water (Mi~pore Corp., Bedford, MA). After 1 hr in the antibiotic solution each ganglion was transferred to a Sylgard-covered (Sylgard-184, DowCorning, Midland, MI) 35mm Petri dish containing Ringer’s solution. The ganglia were pinned (0.1 mm, number 00 insect pins, Carolina Biological Supply, Burlington, NC) to the Sylgard and the thick, yellowish connective tissue sheath that covers each ganglion was dissected to expose the neurons of the right parietal lobe (F-Lobe). Sheath covering the other lobes of the ganglia was not removed. The “desheathed” ganglia were then incubated for 1 hr with “Conditioned Helix Ringer’s solution.” “Conditioned Helix Ringer’s solution” was prepared by incubating normal Helix Ringer’s solution with freshly dissected Helix circumoesophagea1 ganglia for 24 hr at a concentration of 1 ganglion/ml. The ganglia used to condition the solution were removed from the animal and trimmed of non-ganglionic tissue as described above. No attempt was made to remove or open the connective tissue sheath of the conditioning ganglia. “Conditioned

Two types of neurons from the right parietal lobe of the ganglia were examined in these experiments: (i) the large (200-300 pm dia.) bursting neuron Fl and (ii) the mediumsized (SO-75 pm dia.) beating neurons just lateral to Fl, cells F14, F15, F16, F29, F30, F31 and F32 (see Kerkut et al., 1975). To remove the neurons just lateral to Fl, the cells were first identified by their size and proximity to Fl and they they were literally sucked from the ganglion using a Pasteur pipet that had been pulled over a flame to obtain an opening of only 1OO~m. This was more difheub than it sounds and required a great deal of patience. The Pasteur pipet is fitted with a conventional two (2) ml soft rubber bulb, and the connection that hold the cells in the right parietal lobe are loosened by a combination of positive and negative pressure through the pipet tip. To remove Fl a slightly different technique was used. FI was first identified by its position and size and then the entire right parietal lobe was cut from the ganglia and teased apart with sharp forceps. Care was taken not to touch the Fl cell. In the final stages of the isolation, Fl was sucked away from neighboring cells using a Pasteur pipet that had been pulled over a flame to obtain an opening of 500600pm. The isolated neurons were transferred to UV-sterilized poly-l-lysine coated coverslips set on the bottom of a 35 mm Petri dish by pipetting them into 1 ml “conditioned Helix Ringer’s solution” bubbled over the coverslin. PolyL-lysine coated coverslips were prepared by a method similar to that described by Wong et al. (1981). The cells were allowed to settle and adhere to the surface for 6 hr. After this period 3 ml of “conditioned Hefix Ringer’s solution” was added to flood the Petri dish and 3 freshlv dissected, antibiotic treated Helix circumoesophageal ganglia (l/ml) were added to the dish just to the side of the coverslip containing the isolated cells. Fresh “conditioned Helix Ringer’s solution” was added every other day in culture. The intact ganglia were replaced with freshly dissected ganglia every 24 hr. Electrical measurements

Coverslips with the cultured neurons were transferred to 35 mm Petri dishes containing normal Helix Ringer’s solution. The dishes were placed on the stage of an inverted microscope (Wilovert II ph, Germany) and the cells were visualized with phase contrast optics. Conventional recording and display techniques were used for intracellular recordings. Thin walled giass filament microcapillaries 1 mm outside diameter (TWlOOF4, World Precision Instruments, Sarasota, FL) were used to make microelectrodes. Eleci trodes were made by heat-pull method in a Horizontal electrode puller (Narishige Scientific Instrument Laboratory, Tokyo, Japan), filled with 3 M KCl, and then mounted in a head stage amplifier~older connected to a unity gain high impedance amplifier (Model 767, World Precision

Helix aspersa neurons in co-culture

Instruments). The electrodes were l&15 MG in resistance for intracellular recording of F-Lobe neurons in intact ganglion and 2&25 MG in resistance for cultured neurons. The ground electrode, a silver wire coated with silver chloride by bleaching with Chlorox (Trademark, Chlorox Company, Oakland, CA) solution, was placed in the bath containing the cells or ganglion to be recorded, to complete the electrical circuit. Records were displayed on a digital storage oscilloscope (VG6020, Hitachi Denshi America, LTD, Woodbury, NY) and the data was stored on a magnetic tape VCR system (Unitrade Inc., Philadelphia, PA). The data was also collected directly into a computer (SER-386C, Babtech, Irvine, CA) using an RTl800 A/D board (Analog Devices, Norwood MA) and Codas hardware and software (Dataq Instruments, Akron OH). Data was imported to Quattro Pro (Borland, Scotts Valley, CA) for plotting. For recording cells in intact ganglia, the ganglia were pinned to Sylgard-covered Petri dishes, dissected (without enzymes) to expose the neurons and specific neurons were impaled with microelectrodes and recorded from as described above for the cultured cells. The dissection, culturing of neurons and electrical measurements were all done at room temperature (22°C). RESULTS

Adherence and neurite growth Mechanical microdissection vs. enzyme aided dissection. Our initial efforts to isolate and grow Helix neurons in culture used tryptic digestion to aid in

the removal of the cells from the ganglia. Only 12.2 +,3.2% (mean + SD; see Table 1) of the cells from isolated from each ganglion adhered to the poly-L-lysine substrate and grew processes (Fig. 1A). To test whether this low frequency of adherence and growth was a result of protease treatment we did experiments where we plated out cells that had been isolated using various enzymes, trypsin, chymotrypsin, pronase, and collagenase (Type IV, Sigma). As a control, we plated cells “mechanically dissected” from non-enzyme-treated ganglia. In the non-treated control over half the cells attached to the substrate and extended neurites (18 out of 32) but with enzyme treatments only 0% to 20% of the cells attached to the substrate and grew neurites (Table 1). One of us (ST) developed the suction method described in the Methods section for isolating individual, identified cells with a minimum of damage. Effect of ganglia conditioned media. We also observed significant improvement in adherence and process growth (and electrical activity; see next section) in our cultured cells when a Helix ganglion conditioned solution was used. The adherence and process growth of the neurons increased from 53.1% k 9.5% (mean + SD; N = 5) (non-conditioned media, with mechanical dissection) to 81.2% f 6.2% (mean f SD; N = 25) when conditioned media (with mechanical dissection) was used. Fig. 1B shows a typical neuron plated in “conditioned Helix Ringer’s solution”. Electrical activity of the neurons in culture Normal Helix Ringer’s solution. The right parietal neurons in the intact ganglion generally have resting potentials between - 50 and -70 mV, and show spontaneous action potential activity. The action

165

potentials occur in a beating mode (l-4 Hz) or, as in the case of the giant Fl neuron, in a bursting mode (see Figs 2 and 3). Neurons isolated and plated in normal Helix Ringer’s solution (non-conditioned solution) had low apparent resting potentials, between -10 and -20mV (mean= -18mV, N= 15), and neuer showed spontaneous action potentials. The cells were also unstable: the membrane potential decayed to near zero after only a few minutes of recording. The longest recording was for 15 min. When normal Helix Ringer’s solution was supplemented with Leibowitz’s L-15 (with or without fetal calf serum) or with amino acids the same poor electrical activity was obtained. The best response was obtained from a cell that had been plated with another neuron in close proximity (because of incomplete dissociation during the microdissection). This neuron had a resting potential of -40 mV and fired a volley of action potentials when first penetrated by the microelectrode. Conditioned Helix Ringer’s solution. When we cultured the neurons in “conditioned Helix Ringer’s solution” apparent resting potentials increased to -40 mV + 3.5 mV (mean f SEM, N = 50 cells; in 10 preparations) and the cells had stable resting potentials for over one hour. Spontaneous action potentials were still not observed, but injections of depolarizing current induced sharp, robust regenerative responses. Figure 2B shows the action potentials that are observed in a right parietal beating neuron when 0.2 nA of current is injected through the recording pipet. The beating pattern of a similar right parietal neuron in the intact ganglion is shown in Fig. 2A for comparison. The shape of the action potentials is shown just to the right of the beating records. The prominent shoulder on the falling phase of the action potentials is indicative of action potentials with a voltage-sensitive Ca*+ current component. Co-culturing with Helix ganglia. In order to increase the titer of putative factors, we isolated and plated the neurons in “conditioned Helix Ringer’s solution” and then added intact, freshly dissected ganglia to the culture solution after the cells were allowed to attach to the substrate (co-culture) (i.e. after about 6 hr). The adherence and process growth (and general appearance) of the neurons were not improved by adding ganglia to the cultures over that Helix Ringer’s solobserved with “conditioned ution”; however, co-culturing improved the electrical activity of the neurons dramatically. With co-culturing the resting potentials of the neurons increased to values between -40 and -llOmV(-68mV+lOmV;meanfSEM,N=22, determined 20 min after electrode penetration) and the neurons fired

action potentials spontaneously.

When first penetrated by the microelectrode the resting potential of the co-cultured neurons was between -45 and -60 mV and action potentials fired rapidly. Within 10min the resting potentials usually slowly increased to between -60 and -80 mV. On three occasions we have recorded resting potentials between - 90 and - 110 mV. Tip potentials were never more than 5 mV, so the actual resting potential must have been within 5 mV of this apparent value.

SFEMA K. TIWARIand MICHAEL L. WIIODRLJFF

Fig. 1. Appearance of the cells in primary cell culture. A: a right parietal neuron after 2 days in culture in normal Helix Ringer’s solution with glucose and antibiotics. When a cell is dissected from the ganglia the unipolar process which extends from soma into the neuropil in situ is broken and the stub, 1-2 somal diameters, becomes absorbed into the soma as the cell rounds up (inset). Within 6 hr the soma begins to extend one or more processes and these continue to grow until they have lengthened to 5-10 somal Helix Ringer’s solution”. In diameters. B: a right parietal neuron after 2 days in “conditioned ganglion-conditioned solutions the processes grow more extensively and they are thicker. The electrical activity of the co-cultured cells is improved dramatically compared to the cells grown in normal Helix Ringer’s solution (see text). Fibroblast (f) and glial cells (g) also appear in culture. Scale is 25 pm. The

photos were taken using a Leitz orthoplan photomicroscope slowly hyperpolarized sponAS the membrane tanec urs action potentials became less frequent and belov v about -60 mV the cells stopped firing all

with phase contrast optics.

together. Action potential beating behavior could be elicited by chronic small (0.1 nA) current injectic 3ns that artificially depolarized the membrane.

Helix aspersa neurons Table

in co-culture

167

1. Effect of enzyme treatment on cell adherence and neurite extension

Experiment

Number of ganglia

1 2 3 4 5 6 7 7 7 7 7

4 4 5 6 8 2 3 2 2 2 4

Enzyme treatment’ Trypsin’ Trypsin’ Trypsin* Trypsin’ Trypsin3 Protease IX’ ChymotrypsinS Trypsin6 Pronase’ Trypsin/collagenase* None

Total cells plated

Cells adhering and extending neurites

Percentage

34 36 40 51 70 15 22 15 15 18 32

4 4 6 5 10 0 4 2 3 2 18

11.8 11.1 14.0 9.8 14.3 0 18.2 13.3 20.0 11.1 56.2

‘All incubations at room temperature. ‘2.5 mg/ml 10 min, enzyme removed by washing. ‘2.5 mg/ml 10 min, enzyme removed by washing + soy bean trypsin inhibitor (Type II-S) 42.0 mg/ml 10 min, followed by washing. ‘2.0 mg/ml 10 min, followed by washing. “12.5 mg/ml 2 min, followed by washing. ‘0.5 mg/ml 10 min, followed by washing. *7.5 mg/ml trypsin for 2 min followed by washing and 2.5 mg/ml collagenase for 10 min. F-lobe

(a)

-55mV

beating neuron m situ

-

In culture with conditioned helix solution

(bl

2 set

(cl

II co-culture

20 msec

wkh helix ganglia

i

-45

mV -

1 k

__.-~-

_d

4..___

Fig. 2. Effect of “conditioned Helix Ringer’s solution” on the electrical activity of a right parietal beating neuron in culture. a: Typical beating behavior of a right parietal cell in an intact ganglion. To show the shape of an action potential and expanded view of part of the record appears on the right. b: A record from a neuron in culture for 2 days with “conditioned Helix Ringer’s solution”. Nine action potential were triggered by injecting 0.2 nA of positive current through the recording pipet. The action potential on the right is the third action potential from this series. The resting potential was uncertain in this trace because a bridge circuit was not used. c: A record from a neuron after 2 days in culture with “conditioned Helix Ringer’s solution” plus Helix ganglia (co-culture). Spontaneous action potentials occurred immediately after penetration of the cell with the microelectrode (arrow). The shape of one of these spontaneous action potentials is shown on the right. The voltage and time scales in b apply to a and c as well.

SEEMAK. TIWARI and MICHAELL. WOODRUFF

168

The plasma membrane of these cells was very stable and repeated (i.e. successive) penetrations with microelectrodes did not seem to damage the cells. We could record from the same cell for several days (placed back in culture overnight) without any apparent decrease in its electrical activity. Figure 2c shows a representative recording from a right parietal beating neuron grown in co-culture. The beginning of the record shows the penetration of the electrode into the cell. Part of the record is expanded on the right to show the shape of an action potential. The general shape of the action potentials in culture are the same as the action potentials in siru, but they are diminished in amplitude by about S-10 mV, are slightly longer in duration (about 20%) and have a faster return from the after-hyperpolarization. The giant bursting neurons Fl also showed spontaneous activity when plated in co-culture. (a)

Bursting

exhibits a beating behavior, but we never observed bursting. Fl action potentials were shorter in dur-

ation than the action potentials of the right parietal beating neurons and they had little or no shoulder on the repolarizing, falling phase of the action potential (but see last action potential in a burst, Fig. 3A, right). Heat-sensitivity of the conditioned media. When we co-cultured neurons at 30°C instead of 22°C there was no improvement in their electrical activity compared to that observed in normal Helix Ringer’s solution. Resting potentials of the 30°C cultured cells were between - 14 and - 18 mV and they showed no spontaneous activity (N = 16).

neuron F-i ,in situ

Bursting

-65

Figure 3a shows records from Fl cells in intact ganglia and Fig. 3b shows the activity of Fl in primary co-culture. In situ Fl shows two types of behavior: bursting (top, left of Fig. 3A) and beating (lower, left of Fig. 3A) activity. In culture Fl

behavtor

mV-

___.----.

_ A___ ----.A”-~--

-60 mV -

20 msec

20 mV 2 set

(b)

Burstmg neuron F-l

In co-culture

with helix ganglia

Fig. 3. Record from the giant, bursting neuron Fl in co-culture. a: Recordings from Fl in intact ganglia illustrating the two types of typical behavior for Fl, “bursting” and “beating” activity. The activity observed 90% of the time during intracellular recording is “bursting” and this is probably the behavior of Fl when it is unperturbed by experimental intervention. On three occasions (out of 6) “beating” FI neurons turned into “bursters” spontaneously after 10-20 min of recording. b: Spontaneous activity of Fl in co-culture with Helix ganglia.

Helix aspersa

neurons in co-culture

To test the effect of heat “conditioned Helix Ringer’s solution” we placed freshly conditioned solution in a 100°C water bath for 10 min, cooled it to room temperature and used it to plate the cells onto poly-L-lysine (6 hr, see Methods section) and then to culture neurons (at 22°C) with and without added ganglia in co-culture. The resting potentials of all the cells were between - 9 and - 18 mV and there was no spontaneous activity. Evidently adding ganglia in co-culture is a poor substitute for early exposure of the isolated cells to a heat-sensitive conditioning factor or factors. Non-heated solution from the same preparation of “conditioned Helix Ringer’s solution” gave the improvement in electrical activity described above. Efect of dopamine on the neurons in primary cell culture Right parietal beating neurons. When dopamine is added to right parietal beating neurons in the intact ganglion their membranes hyperpolarize by 5-l 5 mV and spontaneous action potentials are suppressed (Fig. 4a upper trace, and see Kerkut et al. (1975). This is a typical response of these neurons to dopamine and it is due at least partially to a dopaminedependent decrease in Ca2+ current (Kim and Woodruff, 1990). When dopamine is added to the same right parietal neurons isolated in primary cell co-culture action potential firing is suppressed; however, there is no significant immediate change in the membrane resting potential. A typical response to dopamine for cells in culture is also shown in Fig. 4a lower trace. In this experiment 5 PM dopamine was added (arrow) to a co-cultured neuron with a resting potential of near -45 mV that was firing action potentials spontaneously, continuously. One action potential occurred immediately after adding dopamine and then the cell became quiet for the next 12 min. A similar response to dopamine was observed in all the cultured neurons tested for dopaminedependence (N = 10). Between 10 and 15 min following dopamine addition in situ the neurons begin to fire action potentials again. These action potentials (after dopamine) are significantly reduced in amplitude and duration (Fig. 4a, right). Some of the neurons in culture also recovered from the dopamine treatment and they started to fire spontaneous action potentials. The co-cultured neuron in Fig. 4a started to fire spontaneous action potential after 12min of dopamineinduced quiescence. The shape of a “recovered” action potential compared to a control (pre-dopamine) action potential is shown on the right of Fig. 4a. The amplitude was reduced by 10 mV and the duration decreased by about 40%. There was also an increase in the after-hyperpolarization. These shape changes, which reflect changes in the voltagesensitive ionic currents that underlie the action potentials are not caused by a slow deterioration of the dopamine-treated cells (either in culture or in situ). They occur rapidly, and as a direct effect of dopamine addition. In one cultured neuron a single action potential fired 9 set after dopamine addition, just before the cell became quiet for over 1Omin. This action potential after only 9 set showed the same dopamine-dependent decrease in amplitude and loss

169

of shoulder illustrated in Fig. 4a. This experiment also shows that the dramatic changes in the action potential are not caused by differences in resting potential. We have voltage clamped these neurons in the intact ganglion and dopamine causes a decrease in the voltage-dependent Ca2+ current which may contribute to the decrease in action potential duration (manuscript in preparation). Right parietal bursting neuron Fl. When dopamine is added to neuron Fl in situ (in either the bursting or the beating mode of activity) spontaneous action potential firing is inhibited. Figure 4b upper trace shows the inhibitory effect of 5 FM dopamine on an Fl in the beating mode in the intact ganglion. When dopamine was added to Fl in culture action potentials were also inhibited. Figure 4b lower trace shows a typical result. Almost immediately after 5 PM dopamine addition the cell stopped firing action potentials and the membrane oscillated slightly before settling down. These results are qualitatively similar, then, to results obtained with dopamine on the right parietal beating neurons described above. The effect of dopamine on Fl is also reversed both in situ and in culture: within 15 min spontaneous action potentials begin to fire again. The recovered activity both in situ and in culture was in the form of the “beating” activity. The superimposed action potentials to the left of the voltage traces in Fig. 4b show the shape of the action potentials before and after adding dopamine. The only significant change seems to be a small increase in the after-hyperpolarization and the change is similar in situ and in culture. The effect of serotonin (5-HT) on cells in primary cell culture Right parietal beating neurons. In situ, 5-HT causes an inhibition of action potential firing of the right parietal beating neurons. Figure 5a upper trace shows a typical response of one of these neuron to 5 FM 5-HT. In cultured cells 5-HT also causes inhibition of the action potential firing. The record in Fig. 5a lower trace is a typical response of a cell in culture to 5 PM 5-HT. After 10-20 min of 5-HT exposure action potentials begin to reappear both in situ and in culture and it is possible to assay the effect of 5-HT on the shape of spontaneous action potentials. The effect of 5 PM 5-HT on action potentials from the right parietal cells is shown on the right in Fig. 5a. The amplitude and duration of the action potentials was decreased by approximately 30% both in the intact ganglion and in culture. Interestingly, the effect of 5-HT on the action potential shape of neurons in culture was reversed by exposure to the excitatory neurotransmitter glutamate (Fig. 5a, inset, right panel). After 20 min of 5-HT treatment the action potential had not regained original amplitude or duration, however, after only 2 min of glutamate addition (3 PM) the action potential height and duration were close to the control levels and the prominent shoulder on the action potentials reappeared. Glutamate also caused a 5-10 mV depolarization (not shown). We tried to reverse the inhibitory effects of dopamine on the action potentials, shown in Fig. 3a right, with glutamate but the results were more equivocal. Glutamate

SEEMAK. TIWARIand MICHAEL L. WCIODRUFF

170

(a)

F-lobe

-70

beating

neuron N)

sdu

mV mV

2 set

In culture

20 msec

-45mV-4

(b)

Burstmg

-60

neuron F-I

in sifu

mV -

In culture

-40mV

--_--_;

-’ //;“”

Fig. 4. Effect of dopamine on right parietal neurons in the intact ganglion and in co-culture. Five 1M dopamine was added to the bath by pipetting at the arrow in each of the traces on the left. Pipetting control solutions without the dopamine added had no effect on the neurons. The bath included 0.01 mM ascorbic acid to retard depamine oxidation. a: A right parietal (F-lobe) beating neuron recorded from an intact ganglion compared to a dissociated neuron in primary cell co-culture. b: A bursting neuron Fl from an intact ganglion compared to Fl in co-culture. The action potentials on the right were taken from the records before (control) and after exposing the cells to dopamine in the same neuron.

addition to dopamine-treated neurons depolarized the membrane by 5-10 mV and the amplitude of the spontaneous action potentials was increased by about lOmV, but the shoulder of the action potential did not reappear (data not shown). Right parietal bursting neuron Fl. When bursting neuron Fl was exposed to 5 PM 5-HT, in either the bursting mode or the beating mode, it becomes more actiue. Fl depolarizes and fires action potentials more rapidly. A response of an in situ Fl cell to 5-HT addition is shown in Fig. 5b upper trace. This “beating mode” Fl depolarized by 17-20 mV and the action potential firing frequency increased from

2 Hz to about 10 Hz. (The effect of 5-HT on an Fl in the bursting mode is shown in the inset of Fig. 5b.) In culture, we added 5-HT to four quiet Fl cells (i.e. cells not firing action potentials spontaneously, resting potentials - 58 to - 65 mV). Each time the Fl began to fire action potentials. This phenomenon is shown in Fig. 5b lower trace. We added 5-HT to several active (beating) Fl cells in culture expecting to see a depolarization and an increase in action potential firing frequency, but there was no significant change in the resting potential or the firing frequency (data not shown).

171

Helix aspersa neurons in co-culture Effect

(a)

F-lobe

-60

beating neuron in

of serotonin

situ

mV-/ __.---W---

--I

t In

-50

20 mV 20 msec

2 set

culture

/‘/ii_,

mV -

1 (b)

Bursting

neuron F-l

in sdu

k -60

mV -

__________---------

------,’

Control After

5-HT

\ ‘--_ C__--.r.l\

In culture

-60

mV -irm

_----

Fig. 5. Effect of serotonin (5-HT) on right parietal neurons in the intact ganglion and in co-culture. Five PM 5-HT was added at the arrows in each of the traces on the left. The action potentials on the right were taken from the corresponding neuron on the left before (control) and after exposing the neurons to 5-HT. a: A right parietal (F-lobe) beating neuron from the intact ganglion compared to a beating neuron in coculture. The insect on the right shows an action potential after adding glutamate to the 5-HT-treated cell. Glutamate (3 PM) was added 20 min after 5-HT addition and the action potential shown was recorded 2 min after glutamate. b: An Fl neuron in beating mode from an intact ganglion compared to an Fl in co-culture. The inset next to the upper trace shows the effect of 5-HT on Fl in a bursting mode. The resting potential of the cultured neuron was approximately -63 mV when 5-HT was added (inset lower trace). The first action potential fired 22 set after 5-HT addition. Individual action potentials in Fl were decrease in amplitude by 5-HT both in situ and in culture (Fig. 5b, right), mainly because they fired from a relatively depolarized resting potential. The general shape of the action potentials was not changed. The slight action potential broadening in situ may be caused by incomplete reversal of K+ inactivation that occurs due to high frequency firing (Aldrich et al., 1979, and see Fig. 3a, right).

DISCUSSION

These experiments describe the primary culturing of Helix neurons using a fairly simplified method. The cells were removed from the ganglion using

no exogenous enzymes to aid in the dissection and then plated onto a poly-l-lysine coated coverslip in Helix Ringer’s solution that had sat overnight with Helix ganglia (conditioned media). Freshly dissected

SEEMAK. TIWARI and MICHAEL L. WOODRUFF

172

Table 2. Summarv of dooamine and serotooin effects

Transmitter

Cell type

Dopamine

FlkFl6, F28-F32 Fl4Fl6, F28-F32 Fl Fl Fl4-F16, F28-F32 FlkFl6, F28-F32 Fl Fl

Serotonin

In sifu in culture

Action potential frequency

Resting potential

Action potential duration

Afterhyperpolarization

Hyperpolarization

Decrease

Decrease

IllCreEW

In culture

No effect

DCCEiS

Decrease

Increase

In sifu In culture In situ

DCCE%C

Decrease

No effect No effect Decrease

Increase Increase No effect

In culture

Hyperpolarization No effect Small Hyperpolarization Hyperpolarization

Decrease

Decrease

No effect

In sifu In culture

Depolarization Slow depolarization

Increase Increase

No effect No effect

No effect No effect

In

situ

ganglia were added to the solution surrounding the cells. The neurons show (i) a high rate of survival (ii) transmembrane potentials comparable to the membrane potentials in neurons in the intact ganglion, (iii) spontaneous action potentials nearly equal in amplitude and duration to those observed cells in the intact ganglion and (iv) responses to dopamine, 5-HT that are similar to the responses of these cells to the neurotransmitters in the intact ganglion. The ionic channels that contribute to the resting potential and that support regenerative activity are active in the cells in primary cell culture. The transmitter-dependent chemistry that leads to a change in the ionic conductances also appears to be maintained. A summary of the effects of dopamine and 5-HT are shown in Table 2. The dopamine-dependent hyperpolarization and decrease in action potential frequency are typical for Helix F-lobe neurons (Kerkut et nl., 1975). The dopamine-dependent decrease in action potential duration in cells Fl4Fl6 and F28-F32 has been observed in many neurons in Helix and has been ascribed in those cells to a dopamine-induced decrease in Ca’+ conductance (Paupardin-Tritsch et al., 1985). The 5-HT-dependent hyperpolarization and decrease in action potential frequency of cells F14-F16 and F28-F32 are typical of about half the Helix F-lobe neurons that respond to 5-HT (Kerkut et al., 1975). The 5-HT-dependent decrease in action potential duration is not well described in the literature. The best-characterized effect of 5-HT on Helix neurons is a 5-HT-dependent increase in action potential duration that is due to a cyclic GMP mediated increase in Ca*+ current (Paupardin-Tritsch et al., 1986). A 5-HT-induced decrease in action potential duration and Ca2+ current has been described for chick dorsal root sensory cells (Dunlap and Fischbach, 1980). Voltage-clamp analysis shows that both dopamine and 5-HT decrease voltage-dependent Ca*+ current of F14-F16 and F28-F32 (Kim and Woodruff, in preparation). The 5-HT-dependent depolarization and increase in action potentials shown here for Fl is typical of the other half of the F-lobe neurons that respond to 5-HT (Kerkut et al., 1975). We are not sure why the enzyme treatments that other investigators use routinely to isolate nerve cells did not work well in our experiments. It is possible that the Helix neurons are unusually sensitive to protease treatment. Additional experiments will have to be performed in order to work out a detailed method of how the cells should be handled with

DCCW3SC

enzymes. We will be trying patch clamp experiments with our cells in culture and it might be useful to be able to “clean” the surface of the cells without affecting their viability. The conditioning of the Helix Ringer’s solution and the addition of freshly dissected Helix ganglia to the cells in culture (co-culturing) had a dramatic effect on the electrical activity of the isolated neurons: resting potentials jumped from an average of about - 18 mV to greater than -45 mV; and action potentials increased from zero occurrence to common occurrence. Wong et al. (1981) observed no increase in electrical activity with ganglia conditioned media in their experiments with Helisoma. Their observation that process growth in Helisoma neurons in culture required conditioned media also is in apparent contradiction to the results that we have obtained in our Helix neurons. Our poly-l-lysine plated Helix neurons do grow processes in non-conditioned solutions (Fig. 1A) (even in normal Helix Ringer’s salt solution with or without glucose added!). However, we can not rule out that a small amount of conditioning factor(s) may have been isolated with the cells as they were sucked from the ganglion during the dissection. We can conclude that a higher titer of the conditioning factor is not necessary for process growth. Conditioning factors may be species specific and have different important properties for each species. Wong et al. (1983) showed that the factor(s) from different gastropod species, Lymnaea, Biomphalaria could improve the neurite growth of Helisoma but that Helisoma worked less well in promoting the neurite growth of either Lymnaea or Biomphalaria. Recently, Kits et al. (1990) discovered that an insulinrelated neuropeptide produced by endocrine cells in the snail Lymnaea can promote neurite outgrowth in Lymnaea neurons in primary cell culture. The principal drawback to using conditioned media is that it is not “defined.” We attempted to use Leibowitz L-15 or normal Helix Ringer’s solution supplemented with amino acids, pyruvate and sugars, media used by other investigators to culture molluscan neurons (Cohan et al., 1987; Dagan and Levitan, 1981; Green et al., 1990; Kaczmarek et al., 1979; Marom and Dagan, 1987; Parsons and Chow, 1989; Schacher and Proshansky, 1983; Wong et al., 1981), but this did not seem to be better than “culturing” the cells in normal (non-conditioned) Helix Ringer’s solution. We also used these solutions with 5 and 10% fetal bovine serum, but no dramatic effect was observed. We tried placing the cells in media

Helix aspersa neurons in co-culture

containing methyl cellulose as described by Dagan and Levitan (1981) and Bodmer et al. (1984) but this, too, did not improve the electrical activity of the neurons. It is interesting that co-culturing improves the electrical activity to a greater extent than using ganglia conditioned Helix Ringer’s solution alone. The principal differences between the two methods, as far as we can see, is that in co-culturing, substances released from the ganglia are fairly immediately and directly available to the neurons by diffusion over a short distance (1 cm). Labile chemicals released from the ganglia would be constantly presented to the cells in co-culture, but would only be transiently presented to the cells when ganglia-conditioned solutions are added (without co-culturing). It seems reasonable to assume that the factor or factors are fairly short-lived and that amounts of them are included with the cells when the cells are plated in normal Helix Ringer’s solution; relatively more of them are included when the cells are plated with “conditioned Helix Ringer’s solution”, and a maximum amount of them are added when the cells are plated in culture with intact Helix ganglia. The success of our mechanica dissection technique may reside in the fact that important conditioning factors are not hydrolyzed by exogenous proteases; however, the chemical nature putative factors needs to be established. It seems worthwhile trying to identify the factors since they seem to improve the electrical activity of neurons. Ac~owledgeme~ts-us

research was supported by Grant 2-11633 from the Office of Research Administration and Development at Southern Illinois University. We thank John Bozzola for assistance in the photomicroscopy, Young-Kee Kim for helpful research discussions and Debra Passmore for technical assistance in preparing the manuscript. REFERENCES

Aldrich R. W., Getting P. A. and Thompson S. H. (1979) Mechanism of frequency-dependent broadening of molluscan neuron soma spikes. J. Physioi. 291, 531-544. Bodmer R., Dagan D. and Levitan I. B. (1984) Chemical and electronic connections between Aplysia neurons in primary cell culture. J. Neurosei 4, 228-233. Bokish A. J. and Walker R. J. (1986) The ionic mechanism associated with the action of putative transmitters on identified neurons of the snail, Helix aspersa. Camp. Biochem. Physiol. 84C, 23 l-241. Braha O., Dale N., Hochner B., Klein M., Abrams T. W. and Kandel E. R. (1990) Second messengers involved in the two processes of p~synaptic facilitation that contribute to sensitization and dishabituation in Apiysiu sensory neurons. Proc. Natn Acad. Sci. U.S.A. 87, 204&2044.

Brown A. M., Morimoto K., Tsuda Y. and Wilson D. L. (1981) Calcium current-dependent and voltage dependent inactivation of calcium channels in He&x usuersa. J. Physiof., Land. 320, 193-218. Chad J. E. and Eckert R. (19861 An enzvmatic mechanism for calcium current ihactiiation ii dialysed Helix neurones. J. Physiof., Lond. 378, 31-U. Cohan C. S., Conner J. A. and Kater S. B. (1987) Electrically and chemically mediated increases in intracellular calcium in neuronal growth cones. J. Neurosci. 7, 3588-3599.

173

Colombaioni L., Paupardin-Tritsch D., Vidal P. P. and Gerschenfeld H. M. (1985) The neuropeptide FMRFamide decreases both the Ca*+ conductance and a cyclic 3’,5’-adenosine monophosphate-de~ndent K+conductance in identifi~ moll~n neurons. J. Neurosci. 5, 2533-2538.

Dagan D. and Levitan I. B. (1981) Isolated identified Apfvsia neurons in cell culture. J. Neurosci. 1, 736-740. De&e P., Paupardin-Tritsch D., Bochaert J. and Gerschenfeld H. M. (1981) Role of cvclic AMP in a serotonin-evoked slow‘inw&d current in snail neurones. Nature 290, 783-785.

Dunlap K. and Fischbach G. D. (1980) Neurotran~t~rs decrease the calcium conductance activated by depolarization of embryonic chick sensory neurones. J. Physiof., Lond. 341, 519-535.

Gerschenfeld H. M., Hammond C. and Paupardin-Tritsch D. (1986) Modulation of the calcium current of molluscan neurones by neurotransmitters. J. exp. Bfof. 124, 73-91. Gerschenfeld H. M., Paupardin-T~t~h D., Hammond C. and Harris-Warrick R. (1990) Intracellular mechanism of neurotransmitter-induced modulations of voltagedependent Ca current in snail neurons. Cell Eiol. Int. Rep. 13, 1141--l 154. Green K. A., Powell B. and Cottrell G. A. (1990) Unitary K+ currents in growth cones and perikaryon of identified Helix neurones in culture. J. exp. Biof. 149, 79-94. Harris-Warrick R. M., Hammond C., PaupardinTritsch D., Homburger V., Rouot B., Bockaert J. and Gerschenfeld H. M. (1988) An aqo subunit of a GTPbinding protein immunologically related to G, mediates a dopamine-induced decrease of Ca’+ current in snail neurons. Neuron 1, 27-32. Kaczmarek L. K., Jennings K. R., Strumwasser F., Nairn A. C., Walter U., Wilson F. D. and Greengard P. (1980) Microinjection of catalytic subunit of cyclic AMP-dependent protein kinase enhances calcium action potentials of bag cell neurons in cell culture. Proc. Natn Acad. Sci. U.S.A. 77, 7487-749 1.

Kaczmarek L. K., Finbow M., Revel J. P. and Strumwasser F. (1979) The morphology and coupling of Aplysiu bag ceils within the abdominal ganglion and in eel1 culture. J. Neurobiof. 10, 535-550. Kerkut G. A., Lambert J. D. C., Gayton R. J., Loker J. E. and Walker R. J. (1975) Mapping of nerve cells in the suboesophageal ganglion of Helix uspersa. Comp. Biochem. Physiof. JOA, l-25. Kim Y-K. and Woodruff M. L. (1990) Two different calcium currents in Helix uspersu neurons are differentially effected by dopamine. Biophys. J. 57, Abstr. W-Pos 407. Kits K. S., deVries N. J. and Ebberink R. H. M. (1990) Molluscan insulin-related neuropeptide promotes neurite ougrowth in dissociated neuronal cell cultures. Neurosci. L.&t. 109, 253-258.

Kramer R. H. and Levitan I. B. (1988) Calcium-dependent inactivation of a potassium current in the Apfysianeuron R15. J. Neurosci. 8. 17961803. Kramer R. H., Levitan E. S., Wilson M. P, and Levitan I. B. (1988) Mechanism of calcium-dependent inactivation of a potassium current in Apfysia neuron R15: interaction between calcium and cyclic AMP. J. Neurosci. 8, 1804-1813. Lee K. S., Akaike N. and Brown A. M. (1978) Properties of internaliy perfused, voltage-clam~d, isolated nerve cell bodies. .I. gen. Physiol. 71, 489-507. Levitan I. B. (1985) Phosphorylation of ion channels. J. Membr. Biol. 87, 177-190. Levitan I. B. (1988) Modulation of ion channels in neurons and other &lls. 2. Rev. Neurosci. 11, 119-136. Levitan E. S. and Levitan I. B. (1988a) Serotonin acting via cyclic AMP enhances both the hyperpolarizing and

174

SEEMAK.

TIWAIUand MICHAELL.

depolarizing phases of bursting pacemaker activity in the Ap/ysia neuron R15. J. Neurosci. 8, 1152-1161. Levitan E. S. and Levitan I. B. (1988b) A cyclic GMP analog decreases the currents underlying bursting activity in the Aplysiu neuron R15. J. Neurosci. 8, 1162-1171. LlinLs R., McGuinness T. L., Leonard C. S., Sugimori M. and Greengard P. (1985) Intraterminal injection of synapsin I or calcium-calmodulin-dependent protein kinase II alters neurotransmitter release at the squid giant synapse. Proc. Natn Acad. Sci. U.S.A. 82, 3035-3039. Marom S. and Dagan D. (1987) Calcium current in growth balls from isolated Helix aspersa neuronal growth cones. Ppiigers Arch. geo. Physiol. 409, 578-581.

Parsons T. D. and Chow R. H. (1989) Neurite outgrowth in primary cell culture of neurons from the squid, Loligo Pealei. Neurosci. Lett. 97, 23-28.

Paupardin-Tritsch D., Colambioni L., Deterre P. and Gerschenfeld H. M. (1985) Two different mechanisms of calcium spike modulation by dopamine. J. Neurosci. 5, 2522-2532. Paupardin-Tritsch D., Hammond C., Gerschenfeld H. M., Nairn A. C. and Greengard P. (1986) cGMP-dependent protein kinase enhances Ca2+ current and potentiates the serotonin-induced Ca2+ c-urrent increase in snail neurones. Nature 323, 812-814. Piomelli D., Voterra A., Dale N., Siegelbaum S. A., Schwartz J. H. and Belardetti F. (1987) Lipoxygenase

WOODRUFF

metabolites of arachidonic acid as second messengers for presynaptic inhibition of Aplysi sensory cells. Nature 328, 38-43.

Sacktor T. C. and Schwartz J. H. (1990) Sensitizing stimuli cause translocation of protein kinase C in Aplysia sensory neurons. Proc. Natn Acad. Sci. U.S.A. 87, 2036-2039. Sawada M., Ichinose M. and Maeno T. (1989) Protein kinase C activators reduce the inositol triphosphateinduced outward current and the Ca2+-activated outward current in identified neurons of Aplysia. J. Neurosci. Res. 22, 158-166.

Schacher S. and Proshansky E. (1983) Neurite regeneration by Aplysia neurons in dissociated cell culture: modulation by Aplysia hemolymph and the presence of the initial axonal segment. J. Neurosci. 3, 2403-2413. Tiwari S., Kim Y-K. and Woodruff M. L. (1989) Dopaminedependent modulation of ionic conductances and protein phosphorylations in Helix aspersa neurons. Sot. Neurosci. Abstr. 211, 12.

Wong R. G., Hadley R. D., Kater S. B. and Hauser G. C. (1981) Neurite outgrowth in molluscan organ and cell cultures: the role of conditioning factor(s). J. Neurosci. 1, 1008-1021. Wong R. G., Martel E. C. and Kater S. B. (1983) Conditioning factor(s) produced by several molluscan species promote neurite outgrowth in cell culture. J. exp. Biol. 105, 389-393.