A high-resolution in vitro bioassay to identify neurons containing red pigment concentrating hormone

Color changes in crustaceans are produced by the migration of pigment granules within chromatophores. Erythrophores constitute the predominant type of chromatophore in crayfish. The aggregation of the pigmentary matrix is mediated by the entry of Ca²⁺ induced by red pigment concentrating hormone (RPCH; Lambert and Fingerman, 1979; Fingerman, 1985), whereas its dispersion appears to be mediated by pigment-dispersing hormone (PDH) via cyclic AMP production (Fingerman et al., 1968). The main conglomerate of neurosecretory cells in crustaceans is located in the medulla terminalis of the eyestalk, which is formed by some 120 monopolar neurons whose cell bodies constitute the X organ (XO) and whose axons extend in a tract that terminates in close association with a vascularized region to form the sinus gland (SG). Several peptides have been isolated and sequenced from the XO-SG system; chromatophorotropic effects are exerted by RPCH (Ferlund and Josefsson, 1972; Gauss et al., 1990) and by PDH (Ferlund, 1976; Rao and Riehm, 1989; Löhr et al., 1993). A second group of hormones with metabotropic effects comprises crustacean hyperglycemic hormone (CHH; Kleinholz et al., 1967), molt-inhibiting hormone (MIH; Chung et al., 1996) and vitellogenesis (gonad-inhibiting) hormone (De Kleijn et al., 1994; Soyez et al., 1991). Antibodies raised against some peptides produced by this system indicate that the XO contains subpopulations of neurons (Dircksen et al., 1988; Mangerich et al., 1986). These neuronal subsets have also been identified in culture conditions: CHH-immunopositive neurons have soma diameters between 25 and 50 μm and develop extensive lamellipodia 24 h after plating, MIH-containing neurons have a soma 25–40 μm in diameter and develop branching outgrowth patterns, whereas RPCHergic neurons are small and spherical with a prominent central nucleus (Grau and Cooke, 1992). The correlation between neuronal firing and RPCH release was demonstrated in the isolated XO-SG system of the crab by evaluating the effect of the solution in which neurons have been stimulated on tegumentary erythrophores from isolated meropodite segments of the large walking legs. The results were compared with standard curves obtained using the synthetic peptide (Stuenkel, 1985). The classic bioassay used to demonstrate RPCH activity is based on pigmentary matrix aggregation according to a subjective scale (Hogben and Slome, 1931).

As an alternative to the bioassay described above, we sought to establish a simple and sensitive bioassay based on the co-culture of dispersed XO cells plated with erythrophores to allow the identification of RPCHergic cells and the quantification of the amount of peptide released by a single neuron.

Adult crayfish (Procambarus clarkii) were used throughout the experiments. The XO neurons were isolated from the eyestalks as described previously (García et al., 1990) and plated in sterile dishes precoated with Concanavalin A type III (Sigma; MO, USA). Dissociated neurons were kept at room temperature (16–24 °C) in modified culture medium (Leibovitz’s L-15; Gibco BRL; NY, USA) containing (in mmol l⁻¹): 205 NaCl, 5.4 KCl, 13.5 CaCl₂, 2.5 MgCl₂, 10 Hepes, 5.5 glucose and 2 L-glutamate with 16 g ml⁻¹ gentamycin (Shering Plough, México), 5 g ml⁻¹ streptomycin (Sigma) and 5 units ml⁻¹ penicillin (Sigma).

To isolate tegumentary erythrophores, exoskeleton fragments were obtained from the cephalothorax region. The non-pigmentary epithelium was then removed, and the exoskeleton containing the pigmentary epithelium was incubated for 30 min in filtered crayfish saline solution containing (in mmol l⁻¹): 205 NaCl, 5.4 KCl, 13.5 CaCl₂, 2.5 MgCl₂ and 10 Hepes with 1 pmol l⁻¹ RPCH (Peninsula Labs; CA, USA). The exoskeleton fragments were incubated in modified Leibovitz’s L-15 medium plus protease type I (1 mg ml⁻¹; Sigma) for 35–45 min; the fragments were then washed with culture medium, and the pigmentary epithelium was removed from the exoskeleton and mechanically dispersed to obtain a cell suspension, which was centrifuged over Lymphoprep (Nycodem Pharma AS; Oslo, Norway) at 1000 revs min⁻¹ for 30 min. The red band containing the erythophores was collected and plated in sterile dishes pretreated with Concanavalin A. To obtain co-cultures, the erythrophore cell suspension was added to the Petri dish after plating the XO neurons.

A rabbit polyclonal antiserum to modified synthetic RPCH was used to identify RPCH neurons (RodrÍguez-Sosa et al., 1994). In previous work, the specificity of the antiserum was tested by comparing the immunoadsorption reaction for RPCH and for other peptides of similar structure such as the adipokinetic hormones. The RPCH antiserum appears to recognize the 3–5 residues close to the carboxy terminal (RodrÍguez-Sosa et al., 1994).

The identification of RPCHergic neurons in both isolated eyestalks and cultured cells was carried out according to the following procedure. Samples were fixed in 2 % paraformaldehyde (5 h for eyestalks and 20 min for cultured cells), rinsed three times and incubated in normal crayfish saline solution containing 0.2 % Triton X-100 (Merck) and 20 % normal goat serum (5 h for eyestalks and 1 h for cultured cells). The preparations were then incubated overnight with the antiRPCH serum (diluted 1:2000 with crayfish saline solution) at room temperature (16–24 °C). After washing, the samples were incubated with a goat anti-rabbit biotinylated antibody at a concentration of 5 g ml⁻¹ for 60 min (Vector Labs; CA, USA), followed by fluorescein–streptavidin complex at 5 g ml⁻¹ for 60 min (Vector). To reduce background fluorescence, the cultured cells were counterstained with 0.1 % Evans Blue dissolved in crayfish saline solution. Finally, the eyestalks were dehydrated through an alcohol series and cleared with methylsalicylate (Merck). The preparations were observed and photographed with a Bio-Rad MRC-600 confocal microscope (Hemel Hempstead, UK).

To evaluate the ability of dispersed erythrophores to aggregate their pigmentary matrix, they were incubated with synthetic RPCH diluted in crayfish saline solution at concentrations from 0.001 fmol l⁻¹ to 1 pmol l⁻¹. The time course of the effect of peptide was quantified by measuring the distance from the tip of selected filipodia to the center of the erythrophore on photomicrographs taken at intervals from 0 to 70 min. The fully dispersed state was expressed as zero and maximal aggregation as 100 %.

In co-culture conditions, presumptive RPCHergic neurons were stimulated by injection of depolarizing current, and the effect was evaluated by comparing the time course of the pigmentary matrix retraction of neighboring erythrophores with that obtained using the synthetic peptide. The identity of these neurons was confirmed immunocytochemically.

Whole-cell current-and voltage-clamp experiments were performed in XO neurons with an Axoclamp-200A amplifier (Axon Instruments; CA, USA). Pipettes were constructed from borosilicate capillaries (Sutter Instruments; CA, USA) using a horizontal puller (P-87 Flaming Brown; Sutter Instruments). Pipettes were filled with a solution containing (in mmol l⁻¹): 215 KCl or CsCl; 2.86 CaCl₂; 2 MgATP; 5.25 EGTA and 10 Hepes, adjusted to pH 7.4 with KOH or CsOH. After being filled, the pipette tip resistance was 2.5 MΩ. Series resistance was electronically compensated (70–80 %). Ca²⁺ currents were filtered with a corner frequency of 5 kHz and stored on hard disk using the DigiData 1200 hardware acquisition system and its software (pClamp6, Axon Instruments). Transient capacitative and leak currents were subtracted using the P/4 protocol. To isolate the Ca²⁺ current, the co-cultures were incubated in a solution containing (in mmol l⁻¹): 195 NaCl, 20 tetraethylammonium chloride, 13.5 CaCl₂, 10 Hepes and 1 mol l⁻¹ tetrodotoxin (Sigma) at pH 7.4.

The antiserum against RPCH stained a subset of XO neurons in the isolated eyestalks (N=12) varying in number from three to seven cell bodies (Fig. 1A). This subpopulation was located in the periphery of the XO close to the medulla interna. A one-to-one correlation between the number of immunoreactive cell bodies in the XO and the stained axons that project to the SG was observed. Immunoreactive neurons and axons were detected in other structures of the eyestalk, as reported in crayfish (RodrÍguez-Sosa et al., 1994) and crabs (Mangerich et al., 1986). As we expected, in culture conditions, only a few neurons in a dispersed XO were immunopositive (5±2, mean ± S.D., N=16 experiments). This subpopulation had a small spherical soma less than 20 μm in diameter, while large neurons with extensive lamellipodia (veiler neurons) or branching outgrowth patterns were immunonegative (Fig. 1B–D).

Isolated erythrophores developed extensive filipodia, maintained their stellate morphology after 24 h in culture and appeared similar to those in fresh isolated exoskeleton fragments (Fig. 2A). The size of cultured erythrophores varied greatly; cell body diameters ranged from 10 to 150 μm and filipodium length from 20 to 200 μm. When they were incubated in crayfish saline solution with RPCH (10 fmol l⁻¹), a progressive retraction of the pigmentary matrix occurred (Fig. 2A–C, photographs were taken 0, 5 and 10 min after exposure to RPCH). The standard aggregation curves (Fig. 2D) were constructed by testing synthetic RPCH (0.001 fmol l⁻¹ to 1 pmol l⁻¹) on dispersed erythrophores after 24 h in culture. The lowest concentration tested (0.001 fmol l⁻¹⁾ did not induce significant changes after 60 min of incubation (open circles), whereas 0.01 fmol l⁻¹ induced only partial retraction over the same incubation period (Fig. 2D; filled circles). In contrast, 0.1 fmol l⁻¹ was effective in causing complete retraction of the pigmentary matrix after 60 min of incubation (Fig. 2D; open triangles). When the concentration tested was 10 fmol l⁻¹, aggregation was complete in 30 min (Fig. 2D; filled triangles). As summarized in Fig. 3, RPCH acted on the target cells in a dose-dependent manner, and the minimal effective concentration was estimated to be 0.01 fmol l⁻¹.

As an additional test of the specificity of the RPCH antiserum, we explored its ability to block the effect of the synthetic peptide on pigmentary matrix aggregation. Preincubation with immune serum (1:50) was effective in blocking the aggregation induced by RPCH (1 fmol l⁻¹), while preincubation with preimmune serum (1:50) did not significantly modify the RPCH response (Fig. 4).

It is not possible to determine from previous studies of RPCH-induced pigmentary matrix retraction in erythrophores within the epithelium whether the effect at the cellular level is a local response. However, isolated erythrophores in culture are a suitable model in which to explore this question. Fig. 5A shows a dispersed erythrophore after 48 h in culture; local application of a pulse of RPCH resulted in a progressive retraction of the pigmentary matrix of the filipodia tested over a 10 min period (Fig. 5B,C).

To identify presumptive RPCH-containing cells in co-culture, we selected neurons with soma diameters of between 15 and 20 μm plated close to erythrophores in which the pigmentary matrix within filipodia adjacent to the neurons was partially retracted (Fig. 6). Neuronal firing induced by intracellular current injection resulted in a progressive aggregation of the neighboring erythrophore (Fig. 6A–C). Continuous perfusion overnight with fresh medium was effective in redispersing the pigmentary matrix (Fig. 6D). In this experiment, electrical stimulation of a second neuron resulted in the same effect on the erythrophore (Fig. 6E,F), and its pigmentary matrix again recovered after perfusion (Fig. 6G). Traces H and I in Fig. 6 show the electrical activity recorded during each experiment. Similar results were obtained from nine other neurons. In three cases, the RPCH content of such neurons was confirmed immunocytochemically, as illustrated in Fig. 7. The aggregation curve from these experiments is plotted in Fig. 2D (open squares). When the Ca²⁺ current was evoked in presumptive RPCHergic neurons under voltage-clamp conditions, using depolarizing command pulses to 20 mV from a holding potential of −60 mV at 0.5 Hz for 5 min, a progressive aggregation of the pigmentary matrix within erythrophores was observed (Fig. 8; N=5). The time course of the aggregation observed during these experiments is plotted in Fig. 2D (filled squares) and was similar to that obtained by neuronal firing induced by depolarizing current injection (Fig. 2D; open squares).

These results suggest that the partial aggregation of the pigmentary matrix observed in erythrophores plated close to RPCHergic neurons could be due to spontaneous release of hormone at a lower rate than that evoked during electrical stimulation. However, cultured erythrophores retain the ability to redisperse their pigmentary matrix after aggregation.

The mean resting membrane potential for neurons able to induce pigmentary matrix aggregation was −52±5 mV, and the mean input resistance was 860±320 MΩ (means ± S.D., N=14). These values are similar to those measured in larger-diameter neurons with veiling or branched outgrowth patterns that failed to induce aggregation of the pigmentary matrix even during prolonged stimulation (30 min, data not shown).

Previous immunocytochemical studies suggest that the XO-SG system in several crustacean species is divided into subpopulations (Keller et al., 1985; Bellon-Humbert et al., 1986; Mangerich et al., 1986; RodrÍguez-Sosa et al., 1994). In the present study, we have identified a variable number of immunoreactive cell bodies in the XO of the medulla terminalis using an antiRPCH serum. From each dispersed XO maintained in culture, between three and seven immunopositive neurons with a small spherical soma (diameter 15–20 μm) and a prominent nucleus were found. Large neurons in XO cell cultures were immunonegative. These results are in agreement with previously reported observations (Grau and Cooke, 1992).

The ability of cultured erythrophores to keep their stellate morphology, as well as their ability to aggregate the pigmentary matrix, dependent upon on the RPCH concentration, provided a suitable model in which to test the specificity of the antiRPCH serum. The dose–response aggregation curves obtained from isolated erythophores revealed an increase in sensitivity to synthetic RPCH in comparison with curves obtained from exoskeleton segments of freshwater shrimp and crayfish, in which the lowest concentration that induced retraction was 1 pmol l⁻¹ (Britto et al., 1990; RodrÍguez-Sosa et al., 1994). Our bioassay was 100 times more sensitive than the meropodite preparation from the large walking legs of a crab (Cooke and Haylett, 1984). The increase in sensitivity could be explained by the absence of other cellular elements and connective tissue acting as barriers to diffusion or by substrate-induced redistribution of RPCH receptors in the target cell, as has been reported in neurons from Aplysia californica and Helix aspersa, in which Concanavalin A induces a change in glutamate sensitivity (Kehoe, 1978). These two mechanisms acting together could explain the increase in the time course of aggregation.

At least two results suggest that pigmentary matrix retraction is a local response: (a) micropulses of RPCH onto long filipodia induced aggregation at the application site, and (b) erythrophores plated close to presumptive RPCHergic neurons showed partial retraction of the pole adjacent to the neuron, suggesting that such neurons could be releasing the peptide at a low rate. This view is supported by microfluorometric experiments on Fura-2-loaded neurons in culture, which generate spontaneous firing at low frequencies (0.1–0.3 Hz) and are associated with transient elevations in the intracellular free Ca²⁺ concentration (Murbartián et al., 1998). The increase in neuronal firing rate (8–10 Hz) induced by depolarizing current injection enhanced the secretion rate in RPCH-containing cells. Since the voltage-dependent Ca²⁺ current in cultured crayfish OX neurons is due to the activation of P-type Ca²⁺ channels (García-Colunga et al., 1999), it seems likely that these channels represent the functional substrate for this secretory activity.

In summary, the bioassay system described here enables the detection of RPCH release from a single neuron and also simplifies the identification of RPCHergic cells. Using this technique, it should be possible to study the modulation of secretory activity by neurotransmitters in the absence of other cell types or humoral factors.

We thank to Dr Meredith Gould for critical reading of the manuscript. This work was supported by grant (26400-N) from CONACyT.