Patterns of dye coupling involving serotonergic neurons provide insights into the cellular organization of a central complex lineage of the embryonic grasshopper Schistocerca gregaria

Abstract

All eight neuroblasts from the pars intercerebralis of one protocerebral hemisphere whose progeny contribute fibers to the central complex in the embryonic brain of the grasshopper Schistocerca gregaria generate serotonergic cells at stereotypic locations in their lineages. The pattern of dye coupling involving these neuroblasts and their progeny was investigated during embryogenesis by injecting fluorescent dye intracellularly into the neuroblast and/or its progeny in brain slices. The tissue was then processed for anti-serotonin immunohistochemistry. A representative lineage, that of neuroblast 1-3, was selected for detailed study. Stereotypic patterns of dye coupling were observed between progeny of the lineage throughout embryogenesis. Dye injected into the soma of a serotonergic cell consistently spread to a cluster of between five and eight neighboring non-serotonergic cells, but never to other serotonergic cells. Dye injected into a non-serotonergic cell from such a cluster spread to other non-serotonergic cells of the cluster, and to the immediate serotonergic cell, but never to further serotonergic cells. Serotonergic cells tested from different locations within the lineage repeat this pattern of dye coupling. All dye coupling was blocked on addition of an established gap junctional blocker (n-heptanol) to the bathing medium. The lack of coupling among serotonergic cells in the lineage suggests that each, along with its associated cluster of dye-coupled non-serotonergic cells, represents an independent communicating pathway (labeled line) to the developing central complex neuropil. The serotonergic cell may function as the coordinating element in such a projection system.

Introduction

The central complex of the insect brain represents one the most sophisticated neuroarchitectures to be found in the nervous system of arthropods (Strausfeld 1976, 1998, 2009). Its modular structure subserves the fine control of motor behavior (Homberg 1987; Martin et al. 1999; Loesel et al. 2002; Strauss 2002; Hoffmann et al. 2007) and it is the relationship between neuroarchitecture and behavior that is providing insights into both the genetic program for building brain structures (Strauss 2002; Neuser et al. 2008) and the evolutionary origin of insects (Strausfeld 1998, 2009; Strausfeld et al. 1998; Loesel et al. 2002).

From a developmental perspective, the time frame during which the central complex is built also correlates with the animal’s lifestyle. In hemimetabolous insects such as the grasshopper, the central complex is structurally mature at the end of embryogenesis (Boyan and Williams 1997; Williams et al. 2005), while in holometabolous insects such as Drosophila, it only appears during the pupal to adult transition and is therefore essentially an adult structure (Renn et al. 1999; Young and Armstrong 2010). In the grasshopper, 96 neuroblasts (NB) are located in each protocerebral brain hemisphere (Zacharias et al. 1993; Boyan and Williams 1997; Williams et al. 2005) and of these, eight neuroblasts located bilaterally in the pars intercerebralis of the brain have been identified as generating lineages which contribute to the central complex (Boyan and Williams 1997; Williams et al. 2005). A subset of four of these neuroblasts (termed NB W, X, Y, Z) generates lineages whose progeny direct fibers to the central body via the columnar system of w, x, y, z tracts (Boyan and Williams 1997; Williams et al. 2005; Boyan et al. 2010a). Progeny from the other four neuroblasts (termed NB 1-2, 1-3, 1-4, 1-5) contribute to other modules of the central complex such as the lateral accessory lobes via the tangential system (Boyan et al. 2010a). All these neuroblasts undergo apoptosis at around 70–75% of embryogenesis by which time the central complex already has a clear adult-like appearance (Boyan et al. 2008b).

Early in embryogenesis, however, the neuroblasts within a given neuromere of the ventral nerve cord (VNC) have been shown to be electrically coupled to one another to form what is essentially an epithelial network (Goodman and Spitzer 1979, 1981; Goodman et al. 1979, 1980). These neuroblasts are also initially electrically coupled to their progeny, but this coupling progressively breaks down and is all but absent after mid-embryogenesis (Goodman and Spitzer 1979). Such coupling is most likely mediated by gap junctions and indeed gap junctional channel proteins have been shown to be expressed in the grasshopper nervous system both during embryonic development (Ganfornina et al. 1999) and in the adult (Anava et al. 2009). A role for such electrical coupling has been reported in growth cone motility and guidance in axogenesis (Kater et al. 1988; Kandler and Katz 1995), in the generation of rhythmical activity (Anava et al. 2009), and stereotypic patterns of electrical coupling between small groups of developing neurons have been shown to play a role in the establishment of neuronal circuits (Lo Presti et al. 1974; Spitzer 1982; Taghert et al. 1982; Kandler and Katz 1995; Wolszon et al. 1995). Further, coupling among cerebral cortical progenitor cells specifies embryonic neurotransmitter expression in neuroblasts by regulating Ca²⁺ levels which in turn influence proliferation and cell differentiation (Spitzer 2006).

In the grasshopper brain, the progeny of central complex neuroblasts express a range of neuromodulatory substances (Boyan et al. 2008a, 2010a) and their ordered projections generate a layered neuroarchitecture within the modules of this midbrain structure (Vitzthum et al. 1996; Vitzthum and Homberg 1998; Homberg et al. 1999; Homberg 2002; Kurylas et al. 2005; Boyan et al. 2008a, 2010a). One of these molecules is the biogenic amine serotonin, a well-described neuromodulator in the nervous system of vertebrates (Wilkinson and Dourish 1991; Gaspar et al. 2003) and invertebrates (Horvitz et al. 1982; Homberg 1991; Groome et al. 1993). In the insects, serotonin is involved in regulating circadian rhythms (Saifulla and Tomioka 2002; Yuan et al. 2005), plays a role in odor-dependent behaviors (Kloppenburg et al. 1999), contributes to spatial learning and memory (Sitaraman et al. 2008; Zars 2009), and modulates aggression (Edwards and Kravitz 1997; Diereck and Greenspan 2007; Johnson et al. 2009). In the periphery, it plays a role in the control of both skeletal and visceral muscle contraction (see Roeder 2002). Developmentally, serotonin has been considered to function as a general coordinator of biochemical differentiation and neuronal function (Gaspar et al. 2003; Vitalis et al. 2007), and is thought to influence developmental processes such as neurogenesis, axogenesis, and cellular differentiation (Condron 1999; Richards et al. 2003; Sykes and Condron 2005; Stern et al. 2007).

While the adult expression pattern of serotonin in the central complex of the grasshopper has been extensively studied (Homberg 1991), little is known about the pattern according to which individual neuromodulatory substances appear in this structure during development (Boyan et al. 2008a, 2010a; Herbert et al. 2010), something which is likely to be highly relevant for the adaptive functioning of the adult system. The serotonergic system is present in all embryonic lineages of the central complex investigated to date (Boyan et al. 2010a) so that cellular interactions involving serotonergic neurons are likely to be part of this developmental program. To test this possibility, we chose one of the neuroblasts (NB 1-3) of the central complex which generates serotonergic progeny at stereotypic locations within its lineage, reflecting their respective birthdates (Boyan et al. 2010a). Intracellular dye injection into ontogenetically identified cells combined with subsequent immunohistochemical labeling for neurotransmitter identity revealed a stereotypic pattern of dye coupling between serotonergic and other neurons of this lineage. Application of the gap junctional blocker n-heptanol selectively eliminated the dye coupling we observed. The pattern of gap junctional mediated cellular interactions we see in ontogenetically related cells provides insights into the internal organization of a central complex lineage in the brain of the grasshopper Schistocerca gregaria.

Materials and methods

Animals

S. gregaria eggs from our own crowded culture were incubated at 30°C in moist aerated containers with a 12/12 h light/dark regime. Embryos were staged at time intervals equal to percentage of embryogenesis as described by Bentley et al. (1979). All experiments were performed according to the guidelines for animal welfare as laid down by the Deutsche Forschungsgemeinschaft.

The data presented in this study were obtained from over 100 preparations. Repetitions involving specific dye coupling experiments are detailed in the relevant figure legends (Figs. 4, 5, 6, and 7).

Immunohistochemistry

Preparation of tissues for dye injection and immunohistochemical experiments in brain slices was conducted as follows. Embryos at either the 50% or 100% stage (just at hatching) were dissected from the egg into cold (4°C) embryonic Ringer and freed of surrounding membranes. Two Ringer solutions were tested during immunohistochemical and dye injection experiments. The first solution comprised 125 mM NaCl, 3 mM KCl, 10 mM CaCl₂, 1 mM MgSO₄, 100 mM sucrose, 20 mM glucose, and 5 mM HEPES buffer (pH 7.4); the second, 150 mM NaCl, 3 mM KCl, 2 mM CaCl₂, 1 mM MgSO₄, and 5 mM TES buffer (pH 7.0). No differences either in immunostaining or in the presence or extent of dye coupling were observed on use of these solutions.

Younger embryos (50%) were not handled further but placed directly into fixative (see below). In the case of older embryos (100%), the brain was removed intact from the head capsule by careful dissection. All tissues were then immersion fixed in PIPES-FA (100 mM PIPES, 2 mM EGTA, 1 mM MgSO₄, and 3.7% paraformaledhyde; pH 6.8–7.4) for 90–120 min and subsequently washed in HEPES or TES buffers depending on the fixation time. Preparations were then embedded in 5% agarose/TES or −/HEPES at 55–60°C, the solution allowed to cool, and the resulting block serially sectioned on a Vibratome (Leica VT 1000S) at 40 μm thickness for 50% embryos, and 70–80 μm thickness for older embryos. Sections (always horizontal to the neuroaxis) were collected in HEPES or TES buffer, freed from agarose, and positioned onto Superfrost® Plus (Menzel–Gläser) microscope slides. Selected sections were covered with preincubation solution (1% normal goat serum, 0.25 mg bovine serum albumin in 1 ml 0.5% TES–Tween or HEPES–Tween; pH 7.6) for 2 h at room temperature to block unspecific binding sites.

Primary antibodies

Serotonin (5HT) The anti-5HT monoclonal antibody (Clone 5HT-H209, raised in mouse, immunogen—5-hydroxytryptamine hydrochloride; Dako UK Ltd.) was added to the preincubation medium at a concentration of 1:250 and preparations kept in the dark for 8 h at room temperature, after which they were washed overnight in either TES or HEPES buffer as appropriate.

anti-8B7 The monoclonal antibody 8B7 recognizes the Akt2 isoform of protein kinase B. The Akt2 kinase has an N-terminal (PH–) domain, a central kinase domain, and a hydrophobic C-terminal domain with regulatory function. In grasshopper, the Akt2 kinase is expressed early in development in neuroblasts and their progeny, later in axonal projections (Seeger et al. 1993). The 8B7 primary antibody (gift of M. Bastiani) was diluted 1:200 in preincubation medium.

anti-1C10 The 1C10 (Lachesin) antibody recognizes a cell surface molecule belonging to the Ig superfamily (Karlstrom et al. 1993). In grasshopper, expression of the molecule occurs initially on all differentiating epithelial cells, but only cells involved in neurogenesis such as neuroblasts continue to express the molecule. Lachesin has additionally been found to play roles in neuronal differentiation and axogenesis. Mutant analyses in Drosophila show that the molecule is required for the blood–brain barrier (Strigini et al. 2006) and morphogenesis of the tracheal system (Llimargas et al. 2003). The 1C10 antibody (gift of M. Bastiani) was diluted 1:5,000 in preincubating medium.

anti-phospho-histone H3 (Ser10) The pH3 monoclonal antibody (Millipore GmbH, Germany) recognizes and binds the phosphorylated form of the amine terminal of the histone H3. This binding is only possible when the chromatin lies dissociated from the nucleosome complex, as occurs during mitotic chromosome condensation. Chromosomes in late meta- and early anaphase are labeled best because phosphorylation is then a maximum (see Hendzel et al. 1997; Adams et al. 2001; Boyan et al. 2010b for details). The pH3 antibody was diluted 1:250 in preincubating medium.

Secondary antibodies

After exposure to the primary antibody, sections were washed thoroughly in TES or HEPES and then placed in preincubation medium (see above) to which the relevant secondary antibody was added for 5–6 h at room temperature, in the dark, as follows: for anti-5HT (GAM-Cy3, Dianova, dilution 1:150; DAM-Alexa 633, Invitrogen, dilution 1:200; DAM-Cy5, Dianova, dilution 1:200); for anti-8B7 (GAM-Cy3, Dianova, 1:150 dilution); for anti-1C10 (GAM-Cy3, Dianova, 1:150 dilution); for anti-pH3 (GAR-Alexa 488, Invitrogen, 1:250 dilution). Sections were then washed in either TES or HEPES for 2–3 h with changes every 20–30 min. Specificity of these secondary antibodies was confirmed by their application in the absence of the primary and no case was a staining pattern observed.

Intracellular dye injection

Wholemount, intact brain

Staged embryos were dissected out of the egg into embryonic Ringer solution (125 mM NaCl, 3 mM KCl, 10 mM CaCl₂, 1 mM MgSO₄, 5 mM HEPES buffer; pH 7.4) and freed from embryonic membranes. Embryos were laid dorsal side up on a Silgard® covered glass slide and held in place with fine entomological pins. The slide was transferred to a fixed stage microscope (Zeiss Axioskop 2) equipped with epifluorescence and DIC optics. A fine chlorided silver wire acting as a reference electrode was placed in the Ringer solution covering the preparation. The electrode holder was connected to the head stage of a Getting 5 DC Amplifier. Intracellular penetrations of target cells were monitored optically using a Zeiss ceramic x63 Achromat water immersion objective, a 1.3 MP color CCD camera (Scion Corp.), and Scion Visicapture™ software.

The Y neuroblast was identified according to its stereotypic location in the dorsal pars intercerebralis of the early embryonic brain (33–34%) using a standard neuroblast map (Zacharias et al. 1993; Boyan and Williams 1997; Williams et al. 2005). Neuroblast, ganglion mother cells, or progeny were impaled with low resistance (30–40 MΩ) thin-walled glass micropipettes containing a solution of 10% Lucifer yellow dye dissolved in 1 M lithium acetate. Cells were stained iontophoretically for 5 min using constant hyperpolarizing current and the injected dye allowed to diffuse passively for a further 10 min. Controls showed that the pattern of dye coupling did not change after this time nor with subsequent histological treatment of the tissue. Following fixation for 15 min in PIPES-FA (100 mM PIPES, 2 mM EGTA, 1 mM MgSO₄, 3.7% paraformaledhyde; pH 6.8–7.4) and washing in 0.2 M PBS, tissues were preincubated for 1 h in a solution of 0.4% PBT (0.2 M PBS plus 0.04% Triton X-100) and 5% FCS and then exposed to peroxidase conjugated anti-Lucifer yellow antibody (Molecular Probes, dilution 1:50) for 12 h at 4°C. After washing for 1 h in 0.2 M PBS, peroxidase-treated embryos were stained with diaminobenzidine (DAB) using Sigma Fast DAB tablets, the staining was intensified with 0.075% ammonium nickel sulfate, and preparations then placed in 90% glycerol in PBS for viewing. Selected preparations were embedded in soft Epon and sectioned at 40 μm thickness using standard histological protocols to allow reconstruction of the lineage.

Brain slices

The intracellular staining of cells in fixed brain slices basically followed the method first described by Wegerhoff and Breidbach (1994) for Tenebrio molitor. The slide with the brain slices (prepared as described in our methods above) was transferred to the same fixed stage microscope as above. A fine chlorided silver wire acting as a reference electrode was placed in the Ringer solution covering the preparation. Neuroblast 1-3 along with its associated lineage was identified optically in the section at 50% of embryogenesis, and in older embryos where the neuroblast is no longer present the lineage was identified according to the stereotypic fiber projection from progeny to the central complex at the level of the protocerebral bridge. This is possible because the neuroblasts of the pars intercerebralis maintain their stereotypic locations from the time of their delamination at 22–25% of embryogenesis to their death at 70–75% (see Williams et al. 2005). Their progeny also remain associated with the neuroblast as a closely packed cell cluster delimited from neighboring clusters by REGA-1 expressing glia-like cells (see Boyan et al. 1995; Seaver et al. 1996). Reconstructions show that the 1-3 lineage is contained within a single horizontal section of 40 μm thickness at 50%, and 70–80 μm thickness at 100% of embryogenesis (see Boyan et al. 2010b).

Cells were impaled with thin-walled glass micropipettes (see above), and staining of cells was accomplished by injection of either 10 mM Alexa Fluor® 568 in 200 mM KCl, 10 mM Alexa Fluor® 488 in 200 mM KCl (both Invitrogen), or a 10% solution of Lucifer yellow CH (Sigma) in 1 M LiAc, each via constant hyperpolarizing current not exceeding 5 nA for ca. 3 min. The Alexa fluorochromes were preferred over Lucifer yellow because their emission spectra are sufficiently narrow so as to prevent spectral overlap with those of the secondary antibodies used in the immunohistochemistry (see above).

Subsequent to dye injection, embryonic Ringer was drained from the slide, and the preparation was covered in Vectashield® (Vector laboratories) and coverslipped for confocal microscopy.

Gap junctional blocker

All the dyes we tested have small molecular weights (see Weber et al. 2004) so that their passage between cells in lineage 1-3 might be mediated by gap junctions. This possibility was tested by adding n-heptanol (a proven gap junctional blocker; see Weingart and Bukauskas 1998; Juszczak and Swiergiel 2009) to the HEPES buffer bathing the brain slice prior to a dye injection experiment. The concentration of n-heptanol in the final bathing solution was varied from 1 × 10⁻⁵ M to 5 × 10⁻⁵ M, and the effect was always the same. The concentrations used were selected following similar experiments involving gap junctional communication in intact, unfixed Hydra (see Chapman et al. 2010 for details). The presence of stained cells showed that the injection of dye itself was not influenced by the presence of n-heptanol.

Imaging

Confocal microscopy

Optical sections of brains from slice preparations were acquired with a Leica TCS SP1 confocal laser scanning microscope equipped with ×25 and ×63 oil immersion objectives. Each fluorochrome was visualized using an argon laser set at the appropriate excitation wavelength. Confocal images were processed using public domain software (ImageJ).

Light and fluorescence microscopy

For light and fluorescence microscopy, preparations were viewed under a Leica DM5000B microscope. Images were captured with a 1.3 MP color CCD camera (Scion Corp.) using Scion Visicapture® software.

Terminology

Neuroarchitecture is described with respect to the neuraxis and not the body axis. The front of the brain (in the head) is defined as being neurally ventral, the top of the brain as neurally anterior, the back of the brain as neurally dorsal, and the base of the brain as neurally posterior. Planes of section are defined such that horizontal to the neuroaxis means parallel to the neuroaxis (anterior is to the top). Location within lineages is with respect to the neuroblast: thus basal implies close to the neuroblast, apical implies distant from the neuroblast.

Results

Identified neuroblasts and lineages of the central complex

Neuroblasts of the embryonic grasshopper brain occupy stereotypic locations throughout their lifetimes which allow them to be mapped precisely at any age (see Zacharias et al. 1993; Williams et al. 2005). Eight neuroblasts located in the so-called pars intercerebralis (PI) of each protocerebral hemisphere generate lineages which build the central complex (see Boyan and Williams 1997; Boyan et al. 2010a, b). Molecular markers (e.g. 1C10) allow these lineages to be visualized during embryogenesis (Fig. 1a) and subsequently reconstructed (see Fig. 7a). These eight neuroblasts (Z, Y, X, W, 1-2, 1-3, 1-4, 1-5) express a second marker (8B7) in common with the axons generated by their progeny (Fig. 1b). This molecular feature allows the fiber projections into the developing neuropil to be unequivocally associated with a given lineage, even after the associated neuroblast has undergone apoptosis. An examination of brain slices at 60% of embryogenesis following 8B7 immunohistochemistry (Fig. 1c), a stage at which modules of the central complex such as the protocerebral bridge (PB) form, reveals that progeny from the most ventral four of the neuroblasts (Z, Y, X, W) direct fibers to the protocerebral bridge, those from the remaining more dorsal four (1-2, 1-3, 1-4, 1-5) to other modules such as the lateral accessory lobes (not visible in section).

Fig. 1

Neuroblasts and lineages that build the central complex of the midbrain. a Confocal image of a horizontal brain slice following 1C10 immunohistochemistry at 50% of embryogenesis. Lineages of identified neuroblasts (Z, Y, X, W, 1-2, 1-3) are located at the dorso-medial border of the pars intercerebralis of each protocerebral hemisphere (PC). The progeny in these lineages contribute to the developing central complex. Dashed line indicates brain midline. Other abbreviations: mOc median ocellus. b Confocal image of a horizontal brain slice at 50% of embryogenesis following 8B7 immunohistochemistry reveals the row of central complex neuroblasts (Z, Y, X, W, 1-2, 1-3) making up the pars intercerebralis of the left protocerebral hemisphere (PC, border dashed white) together with the bundled fiber projections (white arrowheads) of their progeny into the developing neuropil. Central complex has not yet formed at this stage. c At 60% of embryogenesis, confocal image of a brain slice following 8B7 immunohistochemistry reveals the row of central complex neuroblasts in the pars intercerebralis of each protocerebral hemisphere, and the fiber projections of their progeny to the developing protocerebral bridge (PB). Projections from lineage 1-3 (white arrowheads) are directed to a fiber tract (black arrowhead in right hemisphere) which bypasses the PB and targets other modules of the central complex (not in this plane of section). White arrow indicates ventral (v) throughout. Scale bar in c represents 50 μm throughout

Morphology of serotonin immunoreactive cells from the 1-3 lineage

All eight neuroblasts contributing to the central complex have been shown to generate progeny which express the biogenic monoamine serotonin (Boyan et al. 2010a). The cells expressing the transmitter are among the younger (later born) progeny and accordingly the transmitter can only be demonstrated via immunohistochemistry late in embryogenesis (Boyan et al. 2010a; Herbert et al. 2010). Further, this expression commences in lineages after the associated neuroblast has undergone apoptosis (at 70% of embryogenesis; Boyan et al. 2010a, b). Our experiments below therefore focus on embryos just before hatching, which we call 100% of embryogenesis, and where the serotonin expression has been shown to be reliably strong and adult-like in pattern (Herbert et al. 2010).

At 100% of embryogenesis, extensive 5HT-positive processes are found in the protocerebral bridge (PB), the central body (CB), and throughout the diffuse neuropil of the protocerebrum (PC) (Fig. 2a). The mushroom bodies (MB), by contrast, are completely devoid of serotonin expression. 5HT-positive cell body fibers project from cell clusters c1-2, c1-3 in the pars intercerebralis ventrally towards modules of the central complex (Fig. 2b). Serotonin-positive cells are found at equivalent locations in the eight central complex lineages (Fig. 2c); the largest of these are at the base of the lineage near the former location of the respective neuroblast which has previously undergone apoptosis. Their distribution is symmetrical in the homologous lineage of each brain hemisphere (Fig. 2d). Several smaller serotonergic cells (Fig. 2e, arrowhead) are found at stereotypic spatial intervals elsewhere in the lineage. Closer examination of the large serotonin-positive cells near the base of lineage 1-3 reveals a stereotypic cluster comprising three to four cells in different preparations (Fig. 2e–g). Further, the organization of individual cells within the cluster is almost identical from preparation to preparation and from lineage to lineage (cf. Fig. 2c). Since position in the lineage equates to birth order, their organization suggests that the progeny are born from equivalent cell divisions at the same stage of development in each lineage.

Fig. 2

Serotonin (5HT) immunoreactivity in the brain at 100% of embryogenesis. a Confocal image of a horizontal brain slice at the level of the protocerebral bridge (PB) following 5HT immunohistochemistry. Extensive 5HT-positive processes (white) are found in the PB, the CB, and throughout the diffuse neuropil of the protocerebrum (PC). 5HT-positive cell body fibers (neurites, n1-2, 1-3) run to modules of the central complex from dorsally located cell clusters c1-2, c1-3 (not in section) of the pars intercerebralis. Note the complete absence of 5HT immunoreactivity in the mushroom bodies (MB). Dorsal border of the brain is dashed white. b Confocal image of a horizontal brain slice at the level of the protocerebral bridge (PB) following 5HT immunohistochemistry but from another preparation to that in (a). 5HT-positive cells (c1-2) of lineage 1-2 of the PI direct cell body fibers (n1-2) towards the central complex in the PC. Cell body fibers (n1-3) from cell cluster c1-3 (out of section) are also visible. c 5HT-positive cells from central complex lineages (Z, Y, X, W, 1-2, 1-3, 1-4) montaged from different preparations as seen in horizontal section at the level of the PB (not in view, but see b). Dashed lines indicate both the approximate initial projection axis of cell body fibers from each cluster and the lineage axis. Each lineage stereotypically contains three to four large 5HT-positive cells basally (nearer the neuroblast, already apoptosed at this stage). d Confocal image of a horizontal section of the brain at the level of the PB (not in view) reveals the bilaterally symmetrical cluster of 5HT-positive cells of lineage 1-3 in each hemisphere. Dorsal border of the brain is dashed white. Cell body fibers (n1-3) from the cluster in the right brain hemisphere run ventrally towards the central complex (not in view, but see a). e–g Confocal images of a cluster of 5HT-positive cells (c1-3) located basally in lineage 1-3 of three different preparations as seen in horizontal section at the level of the PB. Each cluster comprises three to four immunoreactive cells. Note the stereotypic organization of cells within the cluster. Serotonin-positive cells are also found at other locations within the lineage (white arrowhead in e). Scale bar represents 100 μm in a; 35 μm in b, e, f, and g; and 65 μm in c and d

Dye coupling

Wholemount, intact brain

In order to establish our brain slice preparation as a viable model for testing the pattern of dye coupling in central complex lineages, we first had to demonstrate that dye coupling is a feature of such lineages in the intact, non-fixed brain, as previously described for lineages of the ventral nerve cord (Goodman and Spitzer 1979, 1981). To test this, we selected the Y neuroblast from the array of eight central complex progenitors because the location of its lineage in the intact brain made it particularly accessible to intracellular and ontogenetic study.

We injected Lucifer yellow dye into various cell types within the Y lineage of the intact, unfixed brain at 33–34% of embryogenesis in repeated preparations (Fig. 3). Note that at this developmental stage there is no serotonin immunoreactivity in the brain. Reconstructions following serial sectioning revealed that dye injected into the neuroblast (Fig. 3a) spread anterogradely first to ganglion mother cells (GMCs) of the lineage, and then to all progeny present. Dye injected into a GMC (Fig. 3b) spread retrogradely to the neuroblast (which was always weakly stained) and anterogradely into all the remaining GMCs and progeny. Dye injected into a single progeny near the apical tip of the lineage (Fig. 3c) spread anterogradely to neighboring progeny such that a small cluster of five to six stained cells resulted, each directing a stained neurite into the developing (y) tract associated with the lineage. These data are consistent with the findings of Goodman and Spitzer (1979, 1981) in the VNC. However, we never observed dye spreading from one neuroblast to another in the intact brain at the stages of embryogenesis we tested, although electrical coupling was reported between such progenitors in the VNC earlier in embryogenesis (Goodman and Spitzer 1979, 1981).

Fig. 3

Intracellular dye injection reveals dye coupling in a central complex lineage of the living, intact brain at 33–34% of embryogenesis. Drawings are reconstructions from serial sections following injection of Lucifer yellow dye into different cells of the Y lineage as indicated by the electrode. Shades of gray correspond to various staining strengths (dark is stronger), unstained cells are drawn empty. Lineages expand such that youngest cells are located basally near the neuroblast (NB), oldest cells apically furthest away from the NB. a Dye injected into the neuroblast (darkest staining) spread anterogradely (red arrow) first to ganglion mother cells (GMCs, black asterisks) of the lineage, and then to all progeny (six) present at the time. b Dye injected into a GMC (black asterisk, darkest staining) spread retrogradely (blue arrow) to the neuroblast (NB, dashed outline, weakly stained) and anterogradely (red arrow) into the remaining GMCs (black asterisks) and progeny. c Dye injected into a single progeny (dark staining) near the apical tip of the lineage spread anterogradely (red arrows) to neighboring progeny such that a small cluster of stained cells resulted, each directing a stained neurite (black) into the developing (y) tract associated with the lineage. Retrograde staining from apical progeny was not observed. Scale bar represents 15 μm

Brain slices

We then repeated the experiments above in treated brain slices at different stages of embryogenesis. Intracellular dye injection into individual central complex neuroblasts (Y, X, W, 1-2, 1-3) in a brain slice at 50% of embryogenesis (Fig. 4a) reveals a bilaterally symmetrical row located at the dorsal border of each protocerebral hemisphere. The organization of these neuroblasts is exactly as revealed above via molecular markers (cf. Fig. 1b, c). Expression of the mitotic marker pH3 by neuroblasts Y and 1-2 of the left hemisphere (Fig. 4a) shows that the neuroblasts were trapped in the metaphase of their cell cycle, and therefore viable, at the time these experiments were being made.

Fig. 4

Intracellular dye injection into neuroblasts and progeny of central complex lineages in brain slices at mid-embryogenesis. a Confocal image of a brain slice from a 50% embryo following immunohistochemistry (α-pH3, blue) and intracellular dye injection (Alexa 488, green; Alexa 568, red) selectively into identified neuroblasts (Y, X, W, 1-2, 1-3) of the pars intercerebralis (PI). Neuroblasts are symmetrically arrayed along the dorso-medial border of each protocerebral hemisphere (PC, outlined as dashed white). Anti-pH3 staining (blue) reveals distribution of mitotically active progenitors of the PC. Note that neuroblasts Y and 1-2 are mitotically active. Dye injection reveals that none of the neuroblasts are dye coupled to one another, but progeny from selected lineages (white arrowheads) injected with dye confirm that dye coupling is present within the lineage. Experiment repeated in n = 3 preparations. b Confocal image of a brain slice from a different 50% preparation and at higher magnification to that in a above shows identified neuroblasts (X, W, 1-2, 1-3) following dye injection separately into each neuroblast as indicated by the schematic electrode (red, Alexa 568; green, Alexa 488). Lack of color co-localization (yellow) indicates that these progenitor cells are not dye coupled to one another, but neuroblast X is dye coupled to several of its progeny (white stars). Dorsal edge of the PI is dashed white. Experiment repeated in n = 5 preparations. c Dye coupling within lineage 1-3. Alexa 568 dye injected first (electrode 1) into neuroblast 1-3 identifies the neuroblast. Subsequently, Alexa 568 (electrode 2) or Alexa 488 (electrode 3) dye injected into separate progeny were found to spread to immediately adjacent cells [white asterisks for Alexa 568 (red), black open asterisks for Alexa 488 (green)]. In addition, both dyes co-localized (yellow color) within cells of the lineage (black asterisk). Approximate lateral borders of lineage are indicated. d 3D confocal image from a further preparation following dye injection first into the neuroblast (1, Alexa 568, red), then one of its immediate progeny (2, Alexa 568, red), and thirdly more distant progeny (3, Alexa 488, green). Dye coupling among progeny leads to a co-localization (yellow) in three cells (black asterisks). Dashed white arrow indicates lineage axis. Experiments in c and d repeated in n = 14 preparations. Scale bar in d represents 50 μm in a, 20 μm in b, 20 μm in c, and 12 μm in d

Just as in the intact brain (cf. Fig. 3), the central complex neuroblasts in fixed brain slices are not dye coupled to one another (Fig. 4b), but can be dye coupled to their progeny (Fig. 4b). Progeny within the lineage can also by dye coupled to one another (Fig. 4a–c). This latter finding was investigated in more detail by performing multiple dye injections within the same lineage as follows. At 50% of embryogenesis, neuroblast 1-3 was localized in the PI via intracellular injection of Alexa 568 dye (Fig. 4c). One of its progeny was subsequently injected with Alexa 568 dye (red), another with Alexa 488 (green). Both dyes spread to other immediate members of the lineage, and some progeny in physical contact with both injected cells accumulated both dyes (Fig. 4c, yellow). As in the intact brain above (cf. Fig. 3c), both stains reveal neurites projecting from the lineage into the associated tract so that the injected cells can be classified as neurons. The pattern of dye coupling was confirmed in further preparations (e.g., Fig. 4d) suggesting that dye coupling is a normal feature of (ontogenetically related) cells from central complex lineage 1-3 at 50% of embryogenesis. The data are in accord with previous studies from the ventral nerve cord (Goodman and Spitzer 1981), and our results show that the brain slice preparation repeats the essential pattern of dye coupling we see in the intact brain (compare Figs. 3 and 4c, d).

Temporal aspects of dye coupling in lineage 1-3

Having shown that progeny of neuroblast 1-3 exhibit dye coupling at 50% of embryogenesis (see Fig. 4c, d), before serotonin can be immunohistochemically demonstrated, our next question was whether dye coupling is also present later in development once serotonin is expressed, and if so, if it involves serotonergic cells. Serotonin immunoreactivity appears first in this projection system at about 80% of embryogenesis, well after central complex neuroblasts have undergone apoptosis (Herbert et al. 2010). In the absence of the neuroblast, dye-coupling experiments late in embryogenesis (100%) could therefore only be conducted on progeny and whose lineage identity had been established based on their definitive fiber projections in the relevant tract (see Fig. 1b, c).

Alexa 568 dye was injected into one of the large putatively serotonergic cells at the base of the 1-3 lineage near the site previously occupied by the neuroblast. In repeated preparations (Fig. 5a–e), the dye was seen to consistently spread to a small cluster of five to eight neighboring cells in the same lineage. The preparations were subsequently processed for serotonin immunohistochemistry and in each case only the injected cell was found to be 5HT positive. Dye coupling is therefore present in the lineage throughout embryogenesis. These results convinced us that lineage 1-3 could be used as a model system for studying dye coupling involving a serotonergic system of the embryonic brain.

Fig. 5

Dye coupling revealed by injection of Alexa 568 into 5HT-positive cells at the base of lineage 1-3 in horizontal brain slices at 100% of embryogenesis. Slices were subsequently immunohistochemically processed with anti-5HT. a, b, c In each case, dye (Alexa 568, red, white electrode) injected into a serotonergic cell (black star) spreads to other cells (red) in the immediate vicinity (white asterisks), but not to neighboring serotonergic cells (green, white arrowheads). Yellow color reveals co-localization of Alexa 568 dye (red) and 5HT immunoreactivity (green) only in injected cells. d, e Confocal images at higher magnification of repeat experiments to those above confirm Alexa 568 dye coupling from a 5HT-positive cell (yellow, black asterisk) to neighboring 5HT-negative cells (red, white stars), but not to 5HT-positive cells (green, white arrowheads). Further, four 5HT-immunoreactive axons (green, white/open arrowheads) originating from the cluster show no evidence of dye coupling. Dorso-medial border of the brain hemisphere is dashed white. f–h Drawings show reconstructions of the c1-3 cluster in three experiments in which Alexa 568 dye was injected into a single 5HT-positive cell (black asterisk). The dye spread consistently to five to eight neighboring 5HT-negative cells (white asterisks) in each case. 5HT-immunoreactive axons project from the cluster ventrally into the protocerebrum in f and g but are out of section in h. Experiments repeated in n = 8 preparations. Scale bar in e represents 20 μm in a, c, f, g, and h; 40 μm in b; and 15 μm in d and e

Patterns of dye coupling in lineage 1-3

Serotonin immunohistochemistry yielded a further insight into the dye coupling within the 1-3 lineage we describe above. While the cell injected with dye was itself 5HT positive, none of the cells to which it was dye coupled were immunoreactive even though they were in direct contact with one another (Fig. 5c, d). In no case were 5HT-positive cells from the cluster dye coupled to one another. An analysis of the cell population dye coupled to the serotonergic cell revealed a consistent number of between five and eight cells—all smaller, 5HT-negative cells (Fig. 5f–h; Table 1).

Table 1 Number of dye-coupled 5HT-negative cells present in the rosette at 100% of embryogenesis after injection of Alexa 488 or Alexa 568 into a 5HT-positive cell at various locations in lineage 1-3

Our next question was whether dye coupling within the lineage was bidirectional: that is, whether dye passed equally from a 5HT-positive cell to a 5HT-negative cell, and vice versa. Alexa 568 dye was first injected into a cell which we subsequently immunohistochemically identified as being 5HT positive (Fig. 6a). As described above, the dye spread to neighboring 5HT-negative, but not to neighboring 5HT-positive, cells (Fig. 6b). In the reverse experiment, Alexa 568 dye injected into one of the smaller 5HT-negative cells (Fig. 6c, d) spread to neighboring 5HT-negative cells and also to a neighboring 5HT-positive cell—confirming the bidirectionality of dye coupling. Significantly, the Alexa 568 dye, having passed from the 5HT-negative into the 5HT-positive cell, did not subsequently pass into other 5HT-positive cells of the cluster, confirming our finding that the 5HT-positive cells are not dye coupled to one another.

Fig. 6

Symmetrical dye coupling in cells of the 1-3 lineage at 100% of embryogenesis. a Confocal image of a brain slice following dye injection (white electrode, Alexa 568, red) into a 5HT-positive cell (black asterisk, outlined dashed black) of the cluster. Yellow color confirms co-localization. Red dye has spread to neighboring 5HT-negative cells (white asterisks) but not to neighboring 5HT-positive cells (green, white arrowheads). b Reconstruction (approximately to scale) of the cellular organization seen in a above and superimposed on it the pattern of dye coupling (red arrows) from the injected 5HT-positive cell (yellow) to neighboring cells. Experiment in a and b repeated in n = 7 preparations. c Confocal image of a brain slice following dye injection (white electrode, Alexa 568, red) into a 5HT-negative cell of the lineage. Dye has spread to other 5HT-negative cells (white asterisks) and to a neighboring 5HT-positive cell (black star, yellow) but not to other serotonergic cells (green, white arrowheads). Three 5HT-positive axons (green, white/open arrowheads) project ventrally into the protocerebrum. d Reconstruction (approximately to scale) of the cellular organization in c above and superimposed on it the pattern of dye coupling (red arrows) from the injected 5HT-negative cell (darker red, white asterisk) to neighboring 5HT-negative cells (white asterisks, pastel). Experiments in c and d repeated in n = 2 preparations. Scale bar in c represents 12 μm throughout

Patterns of dye coupling at different locations within lineage 1-3

The neuroblasts of the pars intercerebralis are multipotent in that each generates progeny expressing a range of neuromodulatory substances, and these are found at stereotypic locations within the lineage (Boyan et al. 2010a). Serotonergic cells occur at more than one location within such lineages (see Fig. 2e). We therefore asked whether the pattern of dye coupling we observed basally (see above) was representative of that found elsewhere in the lineage. We selected cells at three further locations in the lineage for analysis (Fig. 7a). Their different locations within the lineage also reflect their different ages. All the cells were subsequently shown via immunohistochemistry to be serotonergic (Fig. 7b–d). As a control, we also compared the various dyes we had employed to stain the basally located cells (Alexa 488, Alexa 586, Lucifer yellow)—the pattern of dye coupling was always the same.

Fig. 7

Patterns of dye coupling at different locations within lineage 1-3. a Schematic (not to scale) of the 1-3 lineage as at 100% of embryogenesis. The lineage contains approximately 50 cells. Neuroblast 1-3 undergoes apoptosis at about 70% of embryogenesis, its former location with respect to its lineage is indicated (dashed black). Arrow within lineage indicates age progression: younger cells are located basally, older cells apically. 5HT-positive cells (green, 1, 2, 3, numbers correspond to those in panels b–d) are found at several locations throughout the lineage in addition to the basal cells (4, see Fig. 2) and were injected with dye. v ventral. b Alexa 488 dye (red) injected into a 5HT-immunoreactive cell at location 1 in the lineage (co-localization yellow) passes into neighboring 5HT-negative cells (red, white asterisks) but not into other 5HT-positive cells (green, black asterisks). c Alexa 488 dye (red) injected into a 5HT immunoreactive cell at location 2 (co-localization yellow) passes into a 5HT-negative neighboring cell (red, white asterisk), but is not present in the processes (white arrowhead) of nearby 5HT-positive cells. d Lucifer yellow dye (LY, red) injected into a 5HT-immunoreactive cell at location 3 (co-localization yellow) passes into neighboring 5HT-negative cells (red, white asterisks) but not into 5HT-positive cells (white arrowheads) nearby. Weakly 5HT-immunoreactive cells are outlined (dashed white). Data above obtained from n = 35 preparations. Scale bar in b represents 20 μm in b, c, and d

Alexa 488 dye injected into a 5HT-positive cell at location 1 in the lineage (Fig. 7b) passed into neighboring 5HT-negative, but not into 5HT-positive, cells. The same result was obtained for an older cell at location 2 (Fig. 7c), although in the instance shown here only one 5HT-negative neighboring cell was dye coupled to the injected cell. Serotonergic processes from nearby cells (themselves out of picture) show no evidence of co-localization. At location 3, Lucifer yellow dye injected into a younger 5HT-positive cell passed into neighboring 5HT-negative, but not 5HT-positive, cells (Fig. 7d).

A statistical analysis (Table 1) reveals that basally in the lineage, at location 4 (see Fig. 6), a mean of 7.6 5HT-negative cells (SD 2.53, n = 15) were dye coupled to each 5HT-positive cell. At locations 1, 2, and 3, the pooled mean number of dye coupled cells was 6.2 (SD 1.71, n = 35). These numbers are not significantly different, suggesting that the pattern of dye coupling is invariant at locations throughout the lineage.

Gap junctional blocker eliminates dye coupling

Our focus then turned to the mechanism of dye transfer between cells in the lineage. The first consideration was whether dye passed indiscriminately as a result of the preparation technique prior to dye injection—such as the presence of detergents (Tween) in the preincubation medium during antibody blocking. The selectivity of dye transfer we observe argues against this possibility, but as a control, intracellular injection of dye was conducted in the absence of Tween from the bathing medium and repeated in the presence of 0.05% Tween. The pattern of dye coupling was identical (data not shown), thereby precluding membrane damage as a mechanism.

The second possibility is that the dyes we employed were transmitted between cells via gap junctions. This was tested by adding a proven gap junctional blocker (n-heptanol; see “Materials and methods”) to the HEPES buffer bathing the brain slice prior to dye injection. We tested two concentrations of n-heptanol based on data obtained in a parallel study involving gap junctional communication in living Hydra (Chapman et al. 2010). In a control experiment (Fig. 8a), Alexa 488 dye was injected into a putatively serotonergic cell in lineage 1-3 at 100% of embryogenesis. Dye spread from the injected cell to a cluster of seven neighboring cells forming a “rosette” like that described for serotonergic cells above (see Fig. 5f, g). When Alexa 488 dye (green) was injected into a 5HT-positive cell (red) of lineage 1-3 but to which 10⁻⁵ M n-heptanol had been added to the bathing medium (Fig. 8b), the dye remained within the cell (yellow color indicates co-localization), and no dye coupling to neighboring cells was observed. The experiment was repeated a further three times with identical results (not shown). Alexa 488 dye (green) injected into a 5HT-negative cell of lineage 1-3 with 10⁻⁵ M n-heptanol in the bathing medium (Fig. 8c, black star) does not spread to neighboring 5HT-positive cells (red, white stars) as it normally would (see Fig. 7 above). The data are consistent with dye coupling being mediated by gap junctional communication channels. Such connections are functional between 5HT-negative cells, between 5HT-negative and 5HT-positive cells, but not between 5HT-positive cells.

Fig. 8

Gap junctional blocker eliminates dye coupling. a Control experiment. Fluorescence micrograph following injection of Alexa 488 dye into a cell in lineage 1-3 at 100% of embryogenesis. Dye has spread from the injected cell (black star) to a cluster of seven neighboring cells (dashed white outline) in the manner reported above for 5HT-positive cells. Intracellular injecting electrode is still in place in all images. b Fluorescence micrograph following injection of Alexa 488 dye (green) into a 5HT-positive cell (red, black star with white outline) with 10⁻⁵ M n-heptanol added to the HEPES buffer. The dye remains within the cell (yellow, co-localization) and does not diffuse to neighboring cells. c Fluorescence micrograph following injection of Alexa 488 dye (green) into a 5HT-negative cell (black star) in the presence of 10⁻⁵ M n-heptanol. The dye does not diffuse into neighboring serotonin immunoreactive cells (red, white stars). Scale bar in c represents 17 μm throughout

Discussion

The insect central complex comprises a set of linked neuropilar modules in the midbrain and functions as a sensory/motor interface for a wide range of behaviors such as locomotion, flight, stridulation, and orientation/navigation (Homberg 1987, 1994; Ilius et al. 1994; Martin et al. 1999; Loesel et al. 2002; Strauss 2002; Heinze and Homberg 2007; Hoffmann et al. 2007). The neuroarchitecture mediating these behaviors in adult Orthoptera and Diptera is very intricate and has been described in great detail (Williams 1975; Strausfeld 1976, 2009; Homberg 1987; Hanesch et al. 1989; Müller et al. 1997; Heinze and Homberg 2008; El Jundi et al. 2010). In the hemimetabolous grasshopper, this neuroarchitecture arises embryonically (Boyan and Williams 1997; Williams et al. 2005; Boyan et al. 2008b); however, in contrast to neuroblasts associated with the mushroom bodies (Cayre et al. 1994), those generating the central complex undergo apoptosis at around 70% of embryogenesis (Boyan et al. 1995; Williams et al. 2005). At this stage, the central complex already possesses a recognizably adult-like scaffold, but is not yet structurally mature (Boyan et al. 2008b). It is therefore essential for any ontogenetic analysis that the lineage remains recognizable even following the disappearance of the neuroblast. The grasshopper central nervous system provides a suitable model system because lineages of the VNC (Goodman et al. 1980; Taghert and Goodman 1984) and brain (Williams et al. 2005; Boyan et al. 2010a, b) have been shown to remain intact, and the birth order of cells to be conserved, throughout embryonic development; lineages therefore retain their embryonic topology up to hatching.

Paralleling its sophisticated, modular neuroarchitecture, the central complex exhibits an equally intricate expression pattern involving a range of neuromodulators (see Homberg 2002). Each neuromodulator tested to date arises from cells located at stereotypic sites, representing different birthdates, within the lineages of identified neuroblasts of the pars intercerebralis (Boyan et al. 2010a). As in the VNC (Goodman et al. 1979, 1980; Taghert and Goodman 1984), the lineages of the central complex possess a temporal topology which is reflected not only in the morphology and projection pattern but also in the biochemical identity of their progeny (Williams et al. 2005; Boyan et al. 2008a, 2010a).

Serotonin and the central complex

Neurons of the pars intercerebralis of the adult (Homberg 1991) and embryonic (Boyan et al. 2010a; Herbert et al. 2010) grasshopper have been shown to express the universal neurotransmitter serotonin. During embryogenesis, serotonin immunoreactivity appears in the central complex only after 75–80% of development, but by 99% the expression pattern is already very similar to that in the adult (Homberg 1991; Herbert et al. 2010). The staining is present initially in the so-called tangential system, with serotonin expression in the columnar system of the pars intercerebralis only commencing shortly before hatching (Herbert et al. 2010). This may have a functional significance in that sensory information (e.g., polarized vision, visual pattern recognition) enters the central complex first via the tangential system and is then passed on to the columnar system for premotor processing (Träger et al 2008). Developing neuronal networks are often active prior to the appearance of the adult behaviors (e.g., flight in the grasshopper; Stevenson and Kutsch 1986), and since serotonin is thought to influence processes such as neurogenesis, apoptosis, and axogenesis (Condron 1999; Richards et al. 2003; Sykes and Condron 2005; Stern et al. 2007; Filla et al. 2008), it may contribute to the development and maturation of the neural circuitry involved.

Pattern of dye coupling

In the grasshopper, all eight lineages of one brain hemisphere whose progeny contribute fibers to the central complex contain a stereotypic population of serotonergic neurons (Fig. 2; Boyan et al. 2010a). Lineages 1-2, 1-3, 1-4, and 1-5 are structurally highly conserved in that serotonergic cells are located at equivalent locations within each (Fig. 2), and contribute to the tangential system of the central complex rather than its columnar system (Boyan et al. 2010a). We have therefore selected one of these lineages (1-3) as our model system, and using intracellular dye injection combined with immunohistochemistry, we have been able to show that serotonergic neurons in our model lineage (Figs. 5a and 9a) are each dye coupled to stereotypic small clusters (“rosettes”) of five to eight neighboring non-serotonergic cells (Table 1), but not to other serotonergic cells (Fig. 9b, d). The biochemical identity of these non-serotonergic cells is not yet known, but once elucidated will provide insights into the regulatory processes mediated by the reciprocal connections with the serotonergic neurons. These connections are likely to involve gap junctions. Gap junctions are the only known cell–cell communication channels which allow a direct exchange of molecules up to around 1 kDa, such as the dyes we employed here (see Weber et al. 2004). Since we were able to prevent dye coupling completely via the gap junctional blocker n-heptanol (Figs. 8 and 9c, e), we suggest that the dye coupling we observe normally is mediated by insect gap junctions (innexins; see Ganfornina et al. 1999; Anava et al. 2009; Juszczak and Swiergiel 2009). Their properties have, for example, been investigated in the lineage of the DUM neuroblast in the VNC (Goodman and Spitzer 1979, 1981), between optic neurons and lamina neuroblasts of the visual system (Lo Presti et al. 1974) as well as in Retzius cells of the leech (see Belardetti et al. 1984).

Fig. 9

Diagrams summarizing the pattern of dye coupling in lineage 1-3 resulting from intracellular injection of Alexa 488, Alexa 568, or Lucifer yellow (all show consistent patterns) into the soma of 5HT-positive and/or 5HT-negative cells. The pattern of dye coupling is independent of location in the lineage. Injected dye is always shown as red, 5HT immunoreactivity as “+/green”, 5HT non-immunoreactivity as “−”, co-localization as yellow. a Schematic lineage of neuroblast (NB) 1-3 at 100% of embryogenesis (former position of the NB—apoptotic at 70%—is outlined). Only the basal third of the lineage is shown (dashed lines indicate that the lineage continues apically). Dashed arrow indicates age profile of the lineage (y younger, o older). Serotonin (5HT) immunoreactive cells (green, white cross) and their neighbors at four locations throughout the lineage were studied. b Dye (red) injected via the electrode into a 5HT-positive cell (+, hence co-localization as yellow) only spreads (blue arrows) to neighboring 5HT-negative cells (red, −), but not to neighboring 5HT-positive cells (+, green). Functional gap junctions mediating dye coupling are shown white. c Same experiment as in b above but with dye (red) injected into a 5HT-positive cell (+, yellow) following addition of 10⁻⁵ M n-heptanol to the bathing medium. Gap junctions are blocked (black) preventing the dye from passing to neighboring cells. d Dye coupling is symmetrical. Dye (red) injected into a 5HT negative cell (−, red) spreads (blue arrows) to neighboring 5HT-negative cells (−, pastel red indicates weaker staining than in injected cell), and to a 5HT-positive cell (+, yellow) via various putative paths (blue arrows, dashed arrows). Further 5HT-positive cells (+, green) are not stained confirming a lack of dye coupling among such cells. e Same experiment as in d above but with dye (red) injected into a 5HT-negative cell (−, red) following application of 10⁻⁵ M n-heptanol to the bathing medium. Gap junctions are blocked (black) preventing the dye from spreading either to the 5HT-positive cell (+, green) or neighboring 5HT-negative cells. f Data in this study suggest that serotonergic cells (+, green), and the cluster of non-serotonergic cells to which they are dye coupled (red), represent individual functional units within the 1-3 lineage (drawn schematically, not to scale). Each of these clusters is proposed to derive from a transit amplifying progenitor (blue) known to be present as a serial array in the lineage. All cells project fibers to modules of the central complex (CX) via the associated tract. Other progeny of the neuroblast (NB) are ganglion mother cells (oblong) or unidentified postmitotic neurons (open circles)

During early embryonic development, neuroblasts of the grasshopper ventral nerve cord have been shown to be electrically coupled to one another and to their progeny via such gap junctions (Goodman and Spitzer 1981). Gap junctional communication between neuroblasts breaks down rapidly and is no longer present at 50% of embryogenesis (as is the case in the brain; Fig. 3), but is maintained for longer within the lineages. In the case of central complex lineages, gap junctional communication (as evidenced by dye coupling) is functional throughout embryogenesis (Figs. 4 and 5). This extended time span may be associated with the fact that the central complex neuropil itself only forms after 65% of embryogenesis (Boyan et al. 2008b) and requires the coordinated activity of its participating neurons.

Labeled lines in central complex lineages

Serotonergic cells occur as clusters at several seemingly equivalent locations in the brain of Drosophila (Valles and White 1988), bee (Seidel and Bicker 1996), and grasshopper (Homberg 1991). Lineage analysis in the embryonic grasshopper shows that each of the eight central complex lineages contains serotonergic cells (Boyan et al. 2010a), and their location within the lineage (Figs. 2, 7, and 9a) suggests that these tend to be the younger (later born) cells of the neuroblast. Further, central complex neuroblasts of the grasshopper are multipotent with respect to the biochemical identities of their progeny, and a range of neuromodulators (allatostatin, leucokinin, locustatachykinin) has been shown to be expressed by progeny at stereotypic locations within a lineage (Boyan et al. 2010a). A feature of the biochemical expression pattern in all embryonic central complex lineages examined thus far is the lack of co-expression involving serotonin on the one hand and neuromodulators such as locustatachykinin, allatostatin, or leucokinin on the other, consistent with their representing different neurochemical projection systems in the central complex (Boyan et al. 2010a). This concept of “labeled lines” projecting into central complex modules is expanded by our current data set.

The finding that dye coupling occurs exclusively between serotonergic and non-serotonergic neurons, even though the serotonergic neurons may be clustered together within the lineage, shows that it is selective (Fig. 9b, d). Our data suggest that the lineage comprises putatively functional units, essentially “labeled lines”, each coordinated via gap junctions from a focal serotonergic cell (Fig. 9f). The bundling of fibers we see originating from a given lineage (Fig. 1c; Williams et al. 2005) suggests that all the cells belonging to such a functional unit direct axons in parallel to modules of the tangential system of the central complex. Serotonergic neurons have been shown to project into discrete layers within central complex neuropils of the embryo (Boyan et al. 2010a) and adult (Homberg 1991). Whether each labeled line also forms a topographic projection from its lineage in the pars intercerebralis to the central complex as occurs within the columnar system (Williams 1975; Williams et al. 2005) remains to be answered.

Even though we were unable to individually identify the non-serotonergic cells dye coupled to a given serotonergic cell, the fact that dye injected into a non-serotonergic cell only dye coupled to one immediate serotonergic cell (Fig. 6b) suggests that each serotonergic cell is associated with a discrete cluster of non-serotonergic cells (Fig. 9b–d). Such an organization is consistent with a role for serotonin as a coordinating neuromodulator during development (Gaspar et al. 2003; Vitalis et al. 2007). The concept of serotonin as a coordinator in development has recently received support from studies in locusts which show the phase change from solitary to gregarious to involve raised serotonergic levels in the nervous system (Rogers et al. 2004). Serotonin release is triggered by mechanical stimulation of identified leg mechanoreceptors such as would occur when individuals make physical contact with one another (Anstey et al. 2009). As a major organizing center for motor behavior (cf. Strauss, 2002 for Drosophila), the associated behavioral changes are likely to involve the central complex and therefore the lineages which we are investigating here.

Cellular networks involving small clusters of dye coupled cells (as revealed by Lucifer yellow) have been shown to organize information flow in the developing grasshopper nervous system (Taghert et al. 1982). Such coupled cells also appear to communicate via gap junctions, if only transiently (Lo Presti et al. 1974; Kater et al. 1988; Wolszon et al. 1995). With respect to brain circuitry, the question arises as to how cells are selected to participate in such clusters. We show that clustering of dye-coupled cells within a lineage is present prior to mid-embryogenesis (Fig. 3c). One mechanism which presents itself may involve an organizational feature particular to central complex lineages. We have recently shown that central complex lineages contain so-called transit amplifying progenitors (TAPs) through which they generate additional numbers of progeny (Boyan et al. 2010b). These TAPs are organized as a linear array of five to eight cells spread throughout the lineage, and each TAP generates a subset of progeny via self-renewal. We have shown in this present study that serotonin-positive cells within a central complex lineage are each linked via dye coupling to a small cluster of non-serotonergic cells. It is conceivable that these clusters equate to the subsets of cells generated via the TAPs. Cell clusters would therefore represent an organizational feature of the central complex lineage itself with serotonin-positive cells as the focusing elements. The resulting connectivity patterns might then form a substrate for the topological projection pattern from these lineages into the midbrain and the adaptative behaviors the central complex mediates. Further studies in which we take advantage of our knowledge of neuroblast identity, lineage topology, and biochemical expression patterns aim to provide insights into the connectivity patterns involving the serotonergic system of the central complex.