THE JOURNAL OF COMPARATIVE NEUROLOGY 323:411-422 (1992)

Growth Properties of Larval and Adult Locust Neurons in Primary Cell Culture BARBARA KIRCHHOF AND GERD BICKER Institut fur Neurobiologie, Freie Universitat Berlin, W-1000 Berlin 33, Federal Republic of Germany

ABSTRACT We developed a cell culture system for thoracic neurons of fifth instar or adult locusts (Locusta migratoria) in order to obtain maximum visualization of cellular morphology and direct access to the neurons for electrophysiological analysis, The dissociated neurons regenerated new neurites in a serum-free defined culture medium, in which they remained viable for up to 3 weeks. Viability of the cells was confirmed by intracellular recordings demonstrating active membrane properties and action potentials. While the morphology of the cultured neurons is distinct from their in vivo counterparts, they retained some cellular surface properties and markers related to transmitter metabolism. Two factors influencing cellular morphology in vitro were identified in Locusta: 1) the presence of a primary neurite stump, and 2) membrane contacts between cells. Dissociated neurons of the locust species Schistocerca gregaria grown in a hemolymph-enriched medium showed a marked reduction in branching patterns and a tenfold increase in neurite length compared to neurons growing in a medium without hemolymph. This culture system could prove usehl for identifylngthe action of hemolymph-derivedgrowth factors. o 1992 Wiley-Liss, Inc. Key words: neurite growth, varicosities,neurotransmitter, glia, hemolymph

Cell cultures formed from dissociated neurons provide a useful system for studying the shape, function, and chemistry of neurons in a simplified and controlled environment. In three main invertebrate preparations, the snail Helisoma, the marine slug Aplysia, and the medicinal leech Hirudo, single identified neurons have been placed in culture and their synaptic interactions have been investigated. Work on Helisoma addressed the cellular processes governing neurite outgrowth and synaptogenesis (Wong et al., '81; Kater and Mattson, '88), while work on the leech investigated neurophysiological parameters of synaptic transmission among identified cultured neurons (Ready and Nicholls, '79; Nicholls et al., '90). The techniques of culturing isolated identified cells have also been applied to the large neurons of the mollusc Aplysia (Dagan and Levitan, '81; Schacher and Proshansky, '831, and it has been possible to reconstitute defensive withdrawal circuits comprising three identified neurons in vitro, which are capable of undergoing homosynaptic depression and heterosynaptic facilitation (Rayport and Schacher, '86). Although insect preparations have also been employed as neurobiological model systems, progress in neuronal primary culture has been slow. Early work on neuronal insect cell cultures focused mainly on embryonic growth mechanisms (Chen and Levi-Montalcini, '70; for review, see Beadle and Hicks, '85). Subsequent research on dissociated insect neurons in the cockroach Periplaneta (Beadle et al., 0 1992

WILEY-LISS. INC.

'82), the moth Spodoptera (Hicks et al., '811, the fruit fly Drosophila (Wu et al., '83),and the moth Manduca (Hayashi and Hildebrand, '90; Stengl and Hildebrand, '90) centered on electrophysiologicalmembrane properties, pharmacological ligand binding, and uptake mechanisms for neuroactive compounds (Giles and Ushemood, '85; Beadle et al., '87; Lees and Beadle, '88; Wu, '88). The relatively simple CNS of the grasshopper embryo is an attractive preparation for the investigation of cell lineage and cell recognition during neurodevelopment (Goodman and Bate, '81).Cultures of whole embryos allowed the direct examination of mechanisms of axon guidance by pioneer fibers (Keshishian, '801. Furthermore, locusts have been successfully employed for dissecting complex neural circuits into individual cellular components (for review, see Hoyle, '83). Neuronal cell culture approaches are, therefore, especially promising. A common finding in insect culture systems is that mainly embryonic nervous tissue can support the outgrowth of neural processes in dissociated primary cell culture. Two main approaches have so far been described for the culture of locust neurons. Giles and Usherwood ('85) reported that neurons of third instar locusts could be maintained for up to 4 weeks in culture if cocultured with embryonic locust neurons. The second approach employed Accepted May 13,1992

B. KIKCHHOF AND G. BICKEK

412

explants from embryonic nervous tissue, in which glial survival was also possible (Vanhems and Delbos, '87; Vanhems et al., '90). However, cultures of adult cockroach neurons have in the meantime been established in the presence of insect blood (Howes et al., '911, and adult Musca neurons were grown in a serum-supplemented culture medium (Harrison et al., '90). The purpose of this paper is to describe cell surface- and transmitter-related properties of the relatively large neurons of adult or late instar locusts in a simple primary cell culture system. We will identify several extrinsic factors influencing neurite growth patterns and present evidence that our defined culture system will lend itself to a study of hemolymph-derived factors governing the shape and extension of neurites.

MATERIALS AND METHODS The majority of experiments were performed on Locusta inigratoria reared in a crowded colony by H.-J. Pfluger's research group at our institute. Mesothoracic and metathoracic ganglia were dissected out of fifth instar larvae or adult locusts under a sterile Leibovitz L-15 medium (Gibco). The ganglia were incubated for 2 hours in sterile collagenaseidispase (2 mg/ml; Boehringer) and subsequently rinsed twice for 30 minutes in L-15 containing 50 pgiml gentamycin (Gibco). Omission of gentamycin during each of the following steps enhanced the outgrowth of the cells. After the enzymatic treatment and rinses, the ganglia were sucked out of their sheath by means of an Eppendorf tip and transferred into an Eppendorf tube containing 100 ~1 L-15 medium/ganglion. The tissue was triturated by repeated passage ( - 10x1 ) through the tip of an Eppendorf pipette. After a brief spin in a bench-top centrifuge the supernatant containing cellular debris was discarded while the pellet of dispersed cells was resuspended in 150 p1 of medium. The cells were finally plated in Falcon 35 mm Petri dishes containing a volume of 2 ml L-15 medium and cultured in humidified air in an incubator at 28°C. Cell cultures were examined daily and photographed under a Zeiss inverted microscope with phase contrast optics. For vital stainings cultures were incubated according to Peters et al. ('85) with fluorescein diacetate in the dark at room temperature. Then the cultures were counterstained with trypan blue for 2 minutes at 28°C and washed in normal L-15 medium. Subsequently the largest part of the medium was removed, the cultures were covered with a coverslip, and the fluorescence of living cells was observed under a Polyvar fluorescence microscope.

Growth in a hemolymph-conditioned medium Thoracic neurons of a second locust species (Schistocerca gregaria) were grown in L-15 medium in the presence of hemolymph. Experimental animals were reared under crowded conditions in the same room as the Locusta colony. 1 ml During dissection the body cavity was flooded with of L-15 medium per animal, which was repeatedly pipetted up and down. The hemolymph-enriched L-15 medium was withdrawn and passed through a DynaGard 0.2 pm filter for sterilization. Schistocerca neurons were plated in normal L-15 medium as described above. After :3 minutes, the normal medium in the culture dish was replaced by a hemolymph-conditioned L-15 medium for experimental groups, while controls received an additional rinse in normal medium.

-

Neurite measurement The number and length of neurites was quantified by projecting photographic slides of cells onto a screen. The length was measured by a graduated rotational distance gauge. All processes emanating from the soma were regarded as neurites. Somata surrounded by lamellipodia-like extensions were not considered. In addition, the number of varicosities per neurite length was compared for equal numbers of cells grown in isolation (single) and cells making contact with neighbors (contact) within the same culture dish. The assignment of each distinct swelling in a neurite as a varicosity was necessarily somewhat arbitrary. However, when two independent observers counted varicosities, their estimates disagreed by less than 10%. A twotailed t-test was used for statistical evaluation of measurements between two experimental groups. Error bars represent standard errors of the mean.

Immunocytochemistry For immunocytochemical staining against y-aminobutyric acid (GABAIergic neurons, cultures were fixed for 15 minutes in a solution of 5% glutaraldehyde in 0.1 M Tris-buffered saline (TBS) (Bicker et al., '88). Subsequently the cells were washed in TBS (pH = 7.4) 6 times for 10 minutes, incubated in 96% methanol/l% H z 0 2 for 10 minutes in order to inactivate endogenous peroxidase activity, and washed again in TBS. The GABA antiserum was employed at a dilution of 1:1,000 in TBS containing 2% normal goat serum and 0.1% Triton X-100. Incubation in the GABA antiserum occured at 4°C overnight. After several washes in TBS, immunoreactivity was visualized by the indirect peroxidase method with a secondary horseradish peroxidase (HRP)-coupled goat anti-rabbit IgG (Nordic Immunology) diluted 1:200. Either diaminobenzidine (DAB) or chloronaphtol (Bicker et al., '88) served as a substrate for the peroxidase reaction. The commercial supplier (Immunotech) of the GABA antiserum provided the same specifications and purification protocol (Seguela et al., '84) as for an antiserum which has been used in a anatomical description of GABA immunoreactivity (GABA-IR) in the thoracic ganglia of locusts (Watson, '86). A similar staining procedure was employed for neuronal staining with an anti-HRP antiserum (Sigma). The cultures were fixed in 4% paraformaldehyde for 15 minutes, washed in phosphate-buffered saline (PBS; pH 7.4), and incubated with the anti-HRP antiserum at a dilution of 1:200 in PBS containing 2% normal goat serum. Again, immunoreactivity was detected by the indirect peroxidase method using either DAB or chloronaphtol as substrates.

Neutral red staining For neutral red staining, the L-15 medium was replaced by a medium containing filtered neutral red (Sigma) in a concentration of 0.01 mgiml. After incubation for 15 minutes at room temperature, the medium was again exchanged for a normal L-15 medium. Neutral red staining of certain cells in the cultures was observed and photographed under brightfield illumination.

Fig. 1. Time-lapse record of two adult neurons in culture, showing the initial growth of neurites after plating. a-d: Neuron extends lamellipodia-like processes. e-h Neuron extends cylindrical processes with varicosities. Frames were taken approximately 1 day apart. Scale, bar = 25 pin.

Figure 1

B. KIRCHHOF AND G. BICKER

114

Electrophysiology Cultures were mounted on the stage of an Axiovert microscope at room temperature and visualized with phase contrast optics. Neurons were impaled with microelectrodes (resistances 25-50 MS1j filled with 3 M KC1. A conventional DC amplifier allowed current injection through a balanced bridge. Electrical recordings were stored on magnetic tape and displayed on a Gould chart recorder.

RESULTS Attachment and neurite growth The dissociated neurons studied in our cell culture experiments were obtained from the meso- and metathoracic ganglia. The dissociation process was subject to some variability and yielded between 100 and 1,000 single neurons per culture dish. Dissociation of thoracic ganglia from fifth instar larvae was usually more effective than dissociation of adult ganglia, resulting in approximately twice as many neurons per culture dish. Adult neurons showed a slight tendency to grow shorter neurites than larval neurons. Since we did not detect any other systematic differences in cellular phenotypes, the following results apply equally to both larval and adult cell cultures. However, we will indicate whether quantitative data presented in graphs or immunocytocheinical staining were obtained from larval or adult cell cultures. The results are based on the dissociation of 300 thoracic ganglia from fifth instar or adult locust. In order to determine the percentage of cells adhering to the plastic substrate, the resuspended cells were distributed into the culture dishes and after 3 minutes the medium was completely removed and pipetted into a new culture dish. By counting the cells in both dishes we determined that 8 5 9 5 % of the plated cells had adhered to the plastic substrate within the first 3 minutes. The spherical somata of the locust neurons had diameters which ranged from 10 to 100 +m. They attached to the floor of the Petri dish and gradually flattened out in culture. Cells with a primary neurite stump of approximately soma diameter length often shortened and retracted the stump into the soma during the first hours in culture, while longer stumps were not retracted. Within 1 day cells started to sprout new processes. Two types of sprouting neurites are shown in time-lapse records (Fig. 1).The neuron of Figure la-d sprouts lamellipodia-like processes, whereas the neuron of Figure le-h extends more cylindrical processes containing varicosities. After 3 days in culture, 70% of the adhering cells had grown at least two processes which were longer than the soma diameter.

A second external factor contributing to the morphological appearance of neurites seems to be presence or absence of membrane contacts between cells. The processes of cultured neurons often expanded at intervals to form varicosities. The examples of varicosities (arrows) shown in Figure 3a demonstrate the higher density of varicosities on neurites of cells which are in membrane contact with neighboring cells as opposed to the single cell growing in isolation. A quantitative evaluation of 1-week-oldlarval cell cultures indeed showed a significantly higher density of varicosities per neurite length on neurons with membrane contacts to at least one neighboring cell compared to isolated neurons (Fig. 3b).

Viability The percentage of cell survival was initially determined by trypan blue exclusion and fluorescein diacetate staining. Dead cells with fragmented processes and vacuoles in their somata took up trypan blue, while living cells with phase bright somata showed fiuorescein diacetate staining. After 1week in culture, 74% of the attached cells excluded trypan blue and showed fluorescence after fluorescein diacetate incubation. After 3 weeks in optimal culture conditions, which included a weekly exchange of the medium and very brief exposure to light during microscopic examinations, cell survival dropped to 30%. During the main part of the investigation, cell viability was simply judged according to the morphological criteria listed above.

-

Anti-HRP Immunoreactivity Almost all cultured cells expressed strong anti-HRP-IR,a neuronal surface marker of several insect species (Jan and Jan, '82; Snow et al., '87). An example of anti-HRP staining of a network of fine processes which grew within 2 days after plating at a high density is shown in Figure 4a. However, in 1-3-day-old cultures, cells with a very low level of HRP-IR expression were also found (Fig. 4b). These flat multipolar cells, which are presumably glia cells, contained light-refractinggranules and inclusions. Due to their strong adhesiveness the percentage of these cells could be enriched if the ganglia were dissociated in the culture dish instead of the routine trituration process in the Eppendorf tube. Glial cells were not as viable as neurons and died within 1 week. All neurons retained their high level of HRP-IR throughout the entire culture period.

GABA immunoreactivity

Amino acid neurotransmitters can be directly visualized by immunocytochemistry with specific antisera (StormMathisen et al., '83; Seguela et al., '84). Throughout an observation period of 2 weeks, immunocytochemical stainings with a GABA antiserum showed immunoreactivity in Morphology of neurites somata, processes, and varicosities of both larval and adult Unlike the neurons in the ganglion, which have one neurons. A quantitative evaluation was performed on days primary neurite, many cells in vitro are multipolar (Fig. 1). 7 and 8 after plating, resulting in 14% GABA-IR cells A factor which contributes to the final shape of the cultured (n = 95 GABA-IR cells). Figure 5a-d shows neurons which neurons seems to be the length of the primary neurite were processed for GABA-IR in the same Petri dish. A which remains attached to the cell body after the dissocia- comparison of GABA-IR in neurons photographed under tion process (Fig. 2a). In a randomly chosen population of brightfield (Fig. 5a) and phase contrast optics (Fig. 5c) larval cells cultured for 1 week, cell bodies with a single shows the extent of immunoreactivity in the neuritic short primary neurite (Fig. 2b) exhibited more processes branching. Neurons from the same Petri dish but without than cell bodies with a long primary neurite (Fig. 2c). The immunoreactivity were photographed under corresponding number of neurites originating from the soma was plotted conditions (Fig. 5b,dj. GABA-IR was never expressed in the against the length of the broadest process (Fig. 2a). The flattened glia cells. We occasionally found neurons with a result shows that the probability of outgrown neurites lower level of immunoreactivity; however, it is uncertain decreases with the increasing length of the primary neurite whether this indicates a change in transmitter status or simply in viability. stump.

length of p r i m a r y n e u r i t e / d i a m e t e r

Fig. 2 . Presence of primary neurite stump influences sprouting from larval somata. a: Diagram, showing the number of processes sprouting from the soma versus the corrected length of primary neurite stump. The length of the primary neurite stump was divided by the soma diameter to correct for different cell sizes and was, therefore, plotted in arbitrary units. The assignment of the major process as part of the former primary neurite was based on close microscopic inspection of the 30 cells which were evaluated (see Fig. 2c). Data were fitted by a logarithmic function. Letters b and c correspond to neurons shown below. b: Neuron with a short primary neurite stump (arrow) showing

of s c m a

many processes originating from the soma. For better visibility of processes, the neuron was stained with anti-HRP antibody. c: Neuron with a long primary neurite stump showing only few processes originating from the soma. The soma tapers off gradually into a long cylindrical process. Arrow delineatesthe position where the cylindricalprocess abruptly flattens out into an area which generated many new neurites. The cylindrical process was conceivably part of the primary neurite which remained attached to the soma during dissociation and subsequently regenerated the flattened area in culture. For better visibility of processes the neuron was stained with anti-HRP antibody. Scale bars = 25 km,

B. KIRCHHOF AND G. BICKER

416

***

+

-4

P .-0 E

lo

0

._ v) 0

.-L0 o '

0

single

contact

b Fig. 3. Cell contact and expression of varicosities in larval neurons. a: Varicosities (arrows indicate examples) on neurites of two neurons, which have grown connections to each other. The single neuron expresses only two (arrowhead) varicosities. Scale bar = 25 p m . b: Mean number of varicosities/lOO pm neurite in cells grown as singles

and in cells grown in membrane contact to neighbors. Cells grown in contact show a significantly higher ( * * * equivalent to P < 0.001; two-tailed t-test) density of varicosities. The calculation of the densities is based on a total neurite length of 4,129 pm in the single group and 5,635 p m in the contact group. Each group contained n = 17 cells.

Neutral red staining

ing neurites and soma diameters in the range of 20-40 km, the mean membrane potential was -43.3 2 2.3 mV when measured directly in the L-15 medium. Spontaneous spiking activity was found in four neurons (Fig. 6). One neuron which was outgrown in isolation showed endogeneous fluctuations in membrane potential and spiking during the recording period of 30 minutes, (Fig. 6a). Since the action potential in the soma was non-overshooting ( - 13 mV peak amplitude), the spike-generating region of the neuron was most likely located in the regenerating neurites. Nine neurons showed overshooting action potentials (Fig. 6b,, while the remainder displayed oscillatory membrane responses (Fig. 6c) during depolarizing current injection. The traces of Figure 6c and d were obtained 3% hours apart, demonstrating the feasibility of monitoring electrical excitability in neurons during growth in our culture system.

We found two different neutral red staining patterns in both larval and adult cell cultures. The first type is a rather homogeneous distribution of red staining in the cytoplasm of 1.6% of the adult neurons cultured for 1 week (n = 17 stained cells), for example, Figure 5e. A neuron with a T-shaped neurite was stained with the dye, whereas the two surrounding neurons remained unstained. The T-shaped neurite can be seen more clearly under phase optics in Figure 5f, showing the ensemble of neurons before application of the dye. The neutral red uptake and the T-branch close to the soma with symmetrical axons running into the two hemiganglia are characteristics of the DUM-cell phenotype (Evans and O'Shea, '78; Adams et al., '831, suggesting that the stained cell could be a DUM neuron. The second type of neutral red staining was found to be most strongly concentrated in vacuoles of glia (Fig. 5g) and also of disintegrating neurons.

Electrophysiological properties of neurons In order to assay the electrical activity in the cell cultures and to demonstrate the viability of the neurons, we performed intracellular recordings from the somata of dissociated adult Locusta neurons which had been cultured for 7-8 days. In a sample of 36 recorded neurons with sprout-

Neuronal shape in a conditioned medium A common strategy to improve neurite outgrowth in primary culture is to condition or supplement the growth medium with nutritional or growth-promoting additives. When we explored whether hemolymph conditioning would affect neurite growth in our culture system, we found that neurite-promotingeffects were counterbalanced by a higher probability of cells detaching from the surface of the Petri

LOCUST NEURONS IN CELL CULTURE

417

Fig. 4. Anti-HRP staining of larval cultures. a: A network of neurites and fine filopodial extensions (arrows) grown in culture. All parts of the neurons were stained by the anti-HRP antibody. b:

Comparison of strong anti-HRP immunoreactivity of neurons and rather weak reactivity of glia (arrow). Glia cells contain granular inclusions. Scale bars = 25 pm.

dish during incubation. Thus we abandoned this strategy for Locusta neurons, which grew reasonably well in an unconditioned medium. Dissociated neurons of Schistocerca gregaria, a species which is also frequently used in neurobiological research, showed less neuritic outgrowth than Locusta neurons when cultured under the same conditions. While Locusta neurons grew single neurites up to a length of 210 p.m, we never observed neurites longer than 70 p.m on Schistocerca neurons. However when Schistocerca neurons were grown in a hemolymph-conditioned medium (Fig. 7a), we observed a striking and reliable influence on the shape and outgrowth of neurites. Compared to the multipolar neurons grown in a normal medium (Fig. 7b), 50% percent of neurons in the hemolymphconditioned medium attained a bipolar or almost unipolar morphology (Fig. 7a), with neurites displaying long segments without any branching. Hemolymph conditioning affected not only the branching pattern, but also increased the total length and, as evident from Figure 7, the diameter of neurites. The effect of the hemolymph-conditioned medium during neurite outgrowth is quantified for larval cultures in Table 1, showing an approximately tenfold increase in the average total length of neurites. Neurons continued to extend their neurites throughout an observation period of 2 weeks until the processes reached confluency. When we evaluated the processes of cells which had

grown without contact, we observed in some cases single neurites ranging between 1and 2 mm in length.

DISCUSSION We found that outgrowth of locust neurons in vitro was not only restricted to larval organotypic cultures (Vanhems and Delbos, '87) or larval neurons cocultured with embryonic tissue (Giles and Usherwood, '851, but that adult neurons are also capable of growing new neurites (Fig. 1). Some of the sprouted neurons exhibited regenerative action potentials (Fig. 61, demonstrating the feasibility of an electrophysiological analysis of the regenerating neurites. The average resting membrane potential of -43 mV in L-15 medium compares with values in the range of -50 to -60 mV, obtained from neurons of intact locust thoracic ganglia under standard Ringer's solution perfusion.

Cellular phenotype We tried to assess the potential of phenotypic changes in primary culture by comparing surface properties, neurotransmitter-related markers, and the morphology of thoracic neurons in situ and in vitro. In Drosophila embryos anti-HRP antibodies recognize an epitope expressed by at least 17 membrane glycoproteins

Figure 5

LOCUSTNEURONSINCELLCULTURE a

t

419

u

b

Fig. 6. Electrophysiology of adult neurons in culture medium. a: Non-overshooting spikes recorded in a spontaneously active neuron. b: Single overshooting spikes evoked by depolarizing current injection. c , d Depolarizing current injection evokes spike-like oscillatory mem-

brane responses typical of most of the recorded neurons. Trace d was obtained after 3%hours of continuous recording. Calibration: vertical V, I: a, b, 60 mV, c, 30 mV; b, c, 1 nA, horizontal: a, 400 ms; b, c, 50 ms.

(Snow et al., '87). Because the anti-HRP epitope is such a powerful neuronal marker in studies of insect neurodevelopment (Jan and Jan, '82; Snow et al., '871, we applied the antiserum to our culture system, in which it stained somata and sprouting processes of neurons (Fig. 4).The anti-HRP antiserum labels cells of a flat appearance containing translucent grana in the cytoplasm with a weaker intensity than the neuronal cells. A common feature of some types of insect glia is the deposition of materials like glycogen or lipids in storage bodies (Wigglesworth, '60; Hoyle, '86). Since during dissection any adhering tissue was carefully removed from the ganglion, we propose that these flat multipolar cells derive from the central nervous system and are of glial origin. A different carbohydrate composition of neurites and glial processes has also been demonstrated by differential binding of plant lectins in locust embryonic explant cultures (Vanhems and Delbos, '89). Neurotransmitter phenotype is a fundamental cellular property of neurons. Cultures of Periplaneta (Beadle et al., '87) and Drosophila (Wu et al., '83; Wu, '88) synthesize and accumulate neurotransmitters or their precursors. We examined the expression of the neurotransmitter GABA in distinct neurons by means of immunocytochemistry. According to Goodman and Bate ('811, each thoracic ganglion contains approximately 3,000 somata. Watson ('86) reported approximately 250 GABA-immunoreactive somata in each thoracic ganglion which, with the exception of the common inhibitory motoneurons, belong to interneurons. It follows that about 9% of the thoracic cells are GABA-IR whereas we noted strong GABA-IR in 14%of the cultured

neurons. The percentage of GABA-IR cells observed in vivo and in vitro is consistent with the retention of neurotransmitter phenotype in GABAergic interneurons. This interpretation is supported by the observation that the large motoneurons, which express glutamate-IRbut not GABA-IR in vivo (Bicker et al., '881, did not stain with CABA antiserum in culture. A correspondence with the frequency of GABA-IR on sections and in primary cultures of honeybee brains has also been reported (Kreissl and Bicker, '92). In the metathoracic ganglion of the locust 30 neurons of the DUM cell group stain with neutral red (Evans and O'Shea, '781, an indicator of aminergic and/or peptidergic neurons in several invertebrates (Stuart et al., '74; Adams et al., '83). This represents 1%of the estimated 3,000 neurons in the metathoracic ganglion. Again, we counted a 1.6% stained neurons in culture. In one relationship of case we found a neutral red stained neuron which displayed a T-branch characteristic of DUM-cell morphology. The appearance of bifurcating cell processes in vitro was also described in a culture system for adult cockroach neurons (Howes et al., '91). One of the factors which contributes to the neuronal shape ofAplysia neurons in vitro is the presence of an axon stump which restricts the formation of multipolar processes (Schacher and Proshansky, '83).Even though we did not experimentally manipulate the length of the primary neurite stump, we observed a similar correlation in cells with a single major process (Fig. 2). If the dissociation procedure removed the primary neurites, somata readily sprouted processes in all directions. Little if any sprouting occurred, however, when a long primary neurite remained attached (Fig. 2). This might be explained most simply by an axonal transport system which preferentially targets building material for regenerating processes from the soma into the remaining primary neurite stump. A second factor contributing to neuronal morphology was the presence of connections between cells, which led to an almost threefold increase in the density of varicosities on neurites. It is most probable that the difference in the density of varicosities is due to membrane contacts rather than a conditioning of the medium by a diffusible factor, since there was only little expression of varicosities in the isolated neurons growing in the same culture dishes. How-

Fig. 5. Immunocytochemistry and neutral red staining of adult cells. a: GABA-IR in two neurons photographed under brightfield illumination. Note the immunoreactivity in the processes of the right neuron. Immunoreactivity was visualized with chloronaphtol as substrate. b Unlabeled neuron processed for GABA-IR in the same Petri dish as the neuron in a. c: Neurons of a photographed under phase contrast optics after immunocytochemical processing, revealing fine details of the arborization. d Neurons of b photographed under phase contrast optics after immunocytochemical processing. The unlabeled processes are clearly visible. e: Neutral red uptake in neuron with T-shaped neurite (arrow) photographed under brightfield illumination. f: Phase contrast micrograph of neuron shown in e but before neutral red staining. g: Neutral red staining in vacuoles (arrows) of glia. Note granular inclusions. Scale bar = 25 pm.

-

-

Fig. 7. Growth pattern of hemolymph conditioned larval cell cultures. a: Typical growth pattern of a Schzsfocerce neuron after 8 days in a heinolymph enriched medium photographed under phase (contrast.Rlopo-

dia explore the regions surrounding the two growth cones at the bottom and right margin of the figure. b Typical growth pattern of a Schistocerca neuron after 8 days in a normal L-15 medium. Scale bars = 25 fim.

LOCUST NEURONS IN CELL CULTURE

421

TABLE 1. Growth of Larval Neurites in a Normal and in a HemolymphConditioned Medium Days

Control Hemolymphconditioned

4

7

13

29.4 + 7.4

52.6 i 11.9

95.5 ? 21.4

298.6 + 46.6

537.5 i 76.6

906.1 T 178.4

The total length of neurites was measured from randomly chosen larval neurons in each goup on days 4, 7 and 13. Results are presented a s means 2 SEM (n = 20). Since many neurites of different cells had already reached confluency on day 13, which introduced ambiguities in measuring the length of single processes, only neurons growingin isolation from neighboring cells were evaluated. Neurons in the hemolymph conditioned medium show significantly longer neurites than the controls (p < 0.001; two-tailed t-tests performed between experimentals and controls on days 4,7,13).

ever, a conditioning factor acting over short distances cannot be excluded if dilution rendered it ineffective for cells growing a t a distance from the site of its release. It is tempting to relate the varicosities which we observed in vitro to vesicle-filled varicosities which have been described as synaptic output regions on certain neurites of thoracic neurons (Watson and Burrows, '85). However, without examination of the ultrastructure of the varicosities such a comparison would be premature. An electron microscopical study of cultured neurons of Spodoptera showed that varicosities contained large numbers of mitochondria and 50 nm clear and 100 nm dense-core vesicles (Hicks et al.,

'81).

Neurite growth in a hemolymphconditioned medium The in vitro analysis of an embryonic insect nervous system by Levi-Montalcini et al. ('73) and the successful growth of isolated molluscan neurons in conditioned media (Wong et al., '81; Schacher and Proshansky, '83) suggested that, similar to vertebrate nerve cells, invertebrate neurons depend on growth factors for survival and sprouting. Recent evidence indicates that the hemolymph of cockroaches contains growth-promoting activity, because longterm cultures of adult cockroach neurons could be established on hemolymph-treated culture dishes (Howes et al., '91). Neuropeptides have also been implicated to act as diffusible growth factors, since insulin and neuroparsin promote neurite outgrowth in organotypic cultures of the locust embryo CNS (Vanhems et al., '90). Clearly, the identification of diffusible or substrate-attached neurite extension factors in the insect nervous system will depend on further progress in neuronal cell culture. Our cell culture system has revealed the action of yet unidentified neurite extension factors in locust hemolymph. Neurons of the locust species Schistocerca showed little outgrowth in unconditioned L-15 medium (Fig. 7b). However, neurons which grew out in the hemolymph-enriched medium expressed strikingly elongated unbranched neurites (Fig. 7a). Hemolymph factors seem to restrict branching but promote elongation, two important parameters contributing to the shape of fully differentiated neurons. A subsequent biochemical characterization of neurite extension factors in hemolymph may thus contribute to an understanding of developmental and regenerative processes i n an accessible insect nervous system.

ACKNOWLEDGMENTS The authors thank H.-J. Pfluger and P. Stevenson for providing us with experimental animals and helpful discus-

sions. We wish to thank G. Braun, S. Eichmuller, M. Hammer, and S. Kreissl for encouragement and exchange of ideas, and A. Klawitter and S.Schaare for photographic assistance. The help of N. Bidwell and A. Carney on the manuscript is gratefully acknowledged. This work was supported by a grant of the Deutsche Forschungsgemeinschaft (Bi 262/3-3).

LITERATURE CITED Adams, M.E., C.A. Bishop, and M. O'Shea (1983) Strategies for the identification of amine- and peptide-containing neurons. In N.J. Strausfeld (edk Functional Neuroanatomy. Berlin: Springer-Verlag, pp, 239249. Beadle, D.J., and D. Hicks (1985) Insect nerve cell culture. In G. Kerkut and L. Gilbert (eds): Comprehensive Insect Biochemistry, Physiology and Pharmacology, Vol. 5. Oxford: Pergamon Press, pp. 181-212. Beadle, D.J., D. Hicks, and C. Middleton (1982) Fine structure ofPeriplaneta arnerzcana neurones in long-term culture. J. Neurocytol. 11:611-626. Beadle, C.A., I. Bermudez, and D.J. Beadle (1987) Amino acid uptake by neurones and glial cells from embryonic cockroach brain growing in uitro. J. Insect. Physiol. 33:761-768. Bicker, G., S. Schafer, O.P. Ottersen, and J. Storm-Mathisen (1988) Glutamate-like immunoreactivity in identified neuronal populations of insect nervous systems. J. Neurosci. 82108-2122. Chen, J.S., and R. Levi-Montalcini (1970) Axonal growth from insect neurons in glia-free cultures. Proc. Natl. Acad. Sci. USA. 66:32-39. Dagan, D., and I.B. Levitan (1981) Isolated identified Aplysia neurons in cell culture. J. Neurosci. I:736-740. Evans, P.D., and M. O'Shea (1978) The identification of an octopaminergic neurone and the modulation of a myogenic rhythm in the locust. J. Exp. Biol. 73:235-260. Giles, D.P., and P.N.R. Usherwood (1985) Locust nymphal neurones in culture: A new technique for studying the physiology and pharmacology of insect central neurones. Comp. Biochem. Physiol. 801253-59. Goodman, C.S., and M. Bate (1981)Neuronal development in the grasshopper. Trends Neurosci. 4:163-172. Harrison, J.B., C.A. Leech, J. Katz, and D.B. Sattelle (1990) Embryonic and adult neurones of the housefly (Musca dornestica) in culture. Tissue Cell 22:337-347. Howes, E.A., T.R. Cheek, and P.J.S. Smith (1991) Long-term growth in uctro of isolated, fully differentiated neurones from the central nervous system of an adult insect. J. Exp. Biol. 156:591-605. Hayashi, J.H., and J.G. Hildebrand (1990) Insect olfactory neurons in uttro: Morphological and physiological characterisation of cells from the developing antenna1 lobes of Manducu sexta. J. Neurosci. 10:848-859. Hicks, D., D.J. Beadle, and D.P. Giles (1981) Ultrastructure of dissociated neurones from pupae of Spodoptera lcttoralis growing in culture. Int. J. Insect. Morphol. Embryol. I0:399-407. Hoyle, G. (1983) On the way to neuroethology: The identified neuron approach. In F. Huber and H. Mark1 (eds): Neuroethology and Behavioral Physiology. Berlin: Springer-Verlag, pp. 9-25. Hoyle, G. (1986) Glial cells of an insect ganglion. J. Comp. Neurol. Z46:85-103.

Jan, L.Y., and Y.N. J a n (1982) Antibodies to horseradish peroxidase as specific neuronal markers in Drosophila and in grasshopper embryos. Proc. Natl. Acad. Sci. USA. 79:2700-2704. Kater, S.B., and M.P. Mattson (1988) Extrinsic and intrisic regulators of neurite outgrowth and synaptogenesis in isolated, identified Helisorna neurons in culture. In D.J. Beadle, G. Lees, and S.B. Kater (eds): Cell Culture Approaches to Invertebrate Neuroscience. London: Academic Press, pp. 1-31. Keshishian, H. (1980)The origin and morphogenesis of pioneer neurons in the grasshopper metathoracic leg. Dev. Biol. 80:388-397. Kreissl, S., and G. Bicker (1992) Dissociated neurons of the pupal honeybee brain in cell culture. J. Neurocytol. in press. Lees, G., and G . Beadle (1988) Cockroach neuronal cultures as models for investigating the neuropharmacology of the insect nervous system. In D.J. Beadle, G. Lees, and S.B. Kater (eds): Cell Culture Approaches to Invertebrate Neuroscience. London: Academic Press, pp. 109-147. Levi-Montalcini, R., J.S. Chen, K.R. Seshan, and L. Aloe (1973) An in uitro approach to the insect nervous system. In D. Young (ed): Developmental Neurobiology of Arthropods. Cambridge: Cambridge University Press, pp. 5-36.

422 Nicholls, J.G., Y . Liu, B.W. Paynton, and D.P. Kuffler (1990) The specificity of synapse formation by identified leech neurones in culture. J. Exp. Biol. 153:141-154. Peters, J.H., H. Baumgarten, and M. Schulze (1985) Monoklonale Antikorper. Berlin: Springer-Verlag. Rayport, S.G., and S. Schacher (19861 Synaptic plasticity in uztm: Cell culture of identified Aplysia neurons mediating short-term habituation and sensitization. J. Neurosci. 65'59-763. Ready, D.F., and J . Nicholls (1979) Identified neurones isolated from leech CNS make selective connections in culture. Nature 281:67-69. Schacher, S., and E. Proshansky (19831 Neurite regeneration by Aplysia neurons in dissociated cell culture: Modulation by Aplysza hemolymph and the presence oftlie initial axonal segment. J. Neurosci. 35'404-2413. Seguela, P., M. Geffard, R. Buijs, and M. Le Moal (1984) Antibodiee against gamma-aminobutyric acid: Specificity studies and immunocytocnemical results. Proc. Natl. Acad. Sci. USA 81.38883892. Snow, P.M., N.H. Patel, A.L. Harrelson, and C.S. Goodman (1987) Neuralspecific carbohydrate moiety shared by many surface glycoproteins in Drosophila and grasshopper embryos. J. Neurosci. 7:4132-4144 Stengl, M., and J.G. Hildebrand (1990) Insect olfactory neurons on uztro: Morphological and immunocytochemical characterisation of malespecific antenna1 receptor cells from the developing antennae of male Manduca sexta.J. Neurosci. 10:837-847. Storm-Mathisen, J., A.K. Leknes, A.T. Bore, J.L. Vaaland, P. E h i n s o n , F.M.S. Haug, and O.P. Ottersen (1983) First visualisation of ghtarnate and GABA in neurons by immunocytochemistry. Nature 301517-520. Stuart, A.E., A.J. Hudspeth, and Z.W. Hall (1974) Vital staining of specific monoamme-containing cells in the leech nervous system. Cell Tissue Res. 153.55-61.

B. KIRCHHOF AND G. BICKER Vanhems, E., and M. Delbos (1987) Differentiation of glial cells and neuritc outgrowth obtained from embryonic locust central nervous system explants. Brain Res. 411:129-138. Vanhems, E., and M. Delbos (1989)Ultrastructural localization of lectinbinding sites and laminin-like immunoreactivity in glial cells and neurites growing out from explant cultures of the central nervous system of the embryonic locust. Cell Tissue Res. 258:429-436. Vanhems, E., M. Delbos, and J. Girardie (1990) Insulin and neuroparsin promote neurite outgrowth in cultured locust CNS. Eur. J . Neurosci. 2776-782. Watson, A.D.H. (1986) The distribution of GABA-like irnmunoreactivity in the thoracic nervous system of the locust Schistocerca gregaria. Cell Tissue Res. 246:331--341. Watson, A.D.H., and M. Burrows (19851 l'he distribution of synapses on the two fields of neurites of spiking local interneurones in the locust. J. Comp. Neurol. 240:219-232. Wigglesworth, V.B. (1960) The nutrition ofthe central nervous systcm in tlic cockroach Peripplaneta americana L. J. Exp. Biol. 37t500-512. Wong, R.G., K.D. Hadley, S.B. Kater, and G.C. Hauser (1981) Neurite outgrowth in molluscan organ and cell cultures: The role of conditioning factor(s1. J. Neurosci. 1;1008-1021. Wu, C.-F. (1988) Neurogenetic studies of Drosophzta central nervous system neurons in culture. In D.J. Beadle, G. Lees, and S.B. Kater (eds): Cdl Culture Approaches to Invertebrate Neuroscience. London: Academic Press, pp. 149-187. Wu, C.-F., N. Suzuki, and M.-M. Po0 (1983) Dissociated neurons from normal and mutant Drosuph~lolarval central nervous system in cell culture. J. Neurosci. 3t1888-1899.