Identification of a P2X7 receptor in GH(4)C(1) rat pituitary cells: a potential target for a bioactive substance produced by Pfiesteria piscicida.

We examined the pharmacologic activity of a putative toxin (pPfTx) produced by Pfiesteria piscicida by characterizing the signaling pathways that induce the c-fos luciferase construct in GH(4)C(1) rat pituitary cells. Adenosine-5'-triphosphate (ATP) was determined to increase and, at higher concentrations, decrease luciferase activity in GH(4)C(1) rat pituitary cells that stably express c-fos luciferase. The inhibition of luciferase results from cytotoxicity, characteristic of the putative P. piscicida toxin (pPfTx). The actions of both pPfTx and ATP to induce c-fos luciferase were inhibited by the purinogenic receptor antagonist pyridoxalphosphate-6-azophenyl-2',4'-disulfonic acid (PPADS). Further characterization of a P2X receptor on the GH(4)C(1) cell was determined by the analog selectivity of P2X agonists. The P2X1/P2X3 agonist alpha,beta-methylene ATP (alpha,beta-MeATP) failed to increase or decrease c-fos luciferase. However, the P2X7 agonist 2',3'-(4-benzoyl)benzoyl ATP (BzATP), which had a predominant cytotoxic effect, was more potent than ATP. Immunoblot analysis of GH(4)C(1) cell membranes confirmed the presence of a 70-kDa protein that was immunoreactive to an antibody directed against the carboxy-terminal domain unique to the P2X7 receptor. The P2X7 irreversible antagonist oxidized-ATP (oxATP) inhibited the action of ATP, BzATP, and pPfTx. These findings indicate that GH(4)C(1) cells express purinogenic receptors with selectivity consistent with the P2X7 subtype and that this receptor pathway mediates the induction of the c-fos luciferase reporter gene by ATP and the putative Pfiesteria toxin

A biologic activity isolated from toxic Pfiesteria piscicida cultures has been determined to activitate c-fos luciferase in GH 4 C 1 cells (1). P. piscicida is a heterotrophic estuarine dinoflagellate discovered in 1991 by Burkholder and co-workers (2); it has been implicated as the causative agent of major fish kills and fish disease in the two largest U.S. mainland estuaries (the Albemarle-Pamlico of North Carolina and Chesapeake Bay in Maryland and Virginia ) (3). P. piscicida was first implicated as hazardous to human health following accidental exposure of laboratory workers (4). During 1993-1995, environmental exposures were anecdotally reported in North Carolina estuaries (4,5), and in 1997 the first clinical evaluations of people shortly after environmental exposure to P. piscicida blooms were completed in Maryland (6). People who had contact with toxic P. piscicida waters or with potential toxic aerosols reported symptoms including dermal lesions and rashes, a burning sensation on contact with water, fatigue, respiratory irritation, diarrhea, severe headaches, and a neurologic syndrome characterized by learning disabilities manifested as short-term memory dysfunction and other cognitive impairment (4,6). GH 4 C 1 cells are a rat pituitary cell line that has been used to characterize signaling pathways for a variety of first messengers (7).
They have also proven useful for the investigation of algal-derived toxins including maitotoxin (8,9) and a biologic activity produced by Pfiesteria piscicida (1). Each also caused an increase in ionic conductances and an elevation of cytosolic free calcium (9,10). Downstream events include activation of cfos luciferase and cytotoxicity (1,11,12). In vitro methods for characterization of algalderived toxins have relied largely upon functional assays that include receptor-based assays and cell-based toxicity assays (13). Cell based assays can be further modified by changing the end point from the mitochondrial indicator for toxicity (MTT dye-based assay) to specific gene induction (11). These assays, known as reporter gene assays, use responsive cell lines that stably express reporter gene constructs.
In this study we investigated the signaling pathways that elicit the reporter gene response in GH 4 C 1 cells using adenosine-5´triphosphate (ATP) as a model compound. We identified ATP as a novel first messenger for GH 4 C 1 cells that induces c-fos luciferase and cytotoxicity. Using a reporter gene assay, we conducted initial characterization of the ATP receptor through analog specificity studies using P2X receptor agonists and antagonists with differing receptor subtype selectivity. We then used two classes of P2X antagonists to examine whether putative P. piscicida toxin (pPfTx) activates a P2X pathway in these cells.
Toxin isolation. We used an actively growing, fish-killing culture of P. piscicida for toxin isolation. Using methods described previously (14), we isolated the cultures from a fish kill and toxic P. piscicida outbreak in the Neuse Estuary in North Carolina using fish bioassays and water samples taken from the in-progress kill. P. piscicida from the fishkilling bioassays was cloned and grown with algal prey under axenic conditions (but with bacterial endosymbionts retained in the P. piscicida zoospores) (14,15). Following Koch's postulates modified for toxic rather than infectious agents, the axenic clonal P. piscicida culture (with bacterial endosymbionts) and residual benign algal prey (< 5 cryptomonads/mL) were added to cultures of healthy fish (n = 4). Control fish cultures were treated identically, except that they received similar addition of only residual cryptomonad culture without P. piscicida (n = 4, with each replicate containing three tilapia (Oreochromis mossambica). Fish death occurred and was repeated as additional live fish were added to the cultures of P. piscicida. In contrast, control fish, which had been maintained identically but with addition of benign algal prey and not P. piscicida, remained healthy (14).
We identified P. piscicida to species at three levels of isolation: from the fish bioassays of water collected from the Neuse fish kill, from the clonal isolate grown with algal prey, and from the subsequent mass-culture with fish. Species identification was completed from analysis of suture-swollen zoospore cells by scanning electron microscopy (15). Following standard procedure in the Burkholder/Glasgow laboratory, the species identifications were then cross-confirmed by three independent laboratories. Molecular probe analyses was conducted by D. Oldach [heteroduplex mobility assay to verify both the species identification and uni-dinoflagellate culture status (16)] and P. Rublee  We mass-cultured the toxic P. piscicida isolate with live tilapia in a biohazard III facility (14,15). They were maintained in 15psu (practical salinity unit) sterile-filtered seawater (water source 8 km off Beaufort, NC, diluted using deionized water), or in 15-psu water made using Instant Ocean salts (Aquarium Systems, Mentor, OH) and deionized water. Toxic samples were taken from cultures that were actively killing tilapia at the time of collection. The toxic seawater medium was passed through a preparative C18 column and flushed with fresh water to remove the excess salts. The toxic material was then eluted from the column with 100% methanol. This methanol elutant was concentrated and passed through a silica gel column using an elutropic scheme of increasing polarity. We screened fractions for cytotoxicity and reporter gene activity using GH 4 C 1 cells as previously described (1). The greatest P. piscicida activity was found to elute in the later, more polar fractions. These fractions were evaporated to dryness and then were placed under high vacuum to remove any remaining organic solvents that would interfere with the bioassays. The dry residue was taken up in standard volumes of methanol as the carrier solvent for further analysis. The active fraction was determined not to contain ATP by difference in chromatographic retention and by lack of ATP activity using an ATP-dependent in vitro luciferase assay, described below under "ATP Assay." Because the chemical structure of the bioactive substance has not yet been determined in the absence of a sufficient quantity of toxic culture to enable purification, the bioactive substance is referred to here as putative P. piscicida toxin (pPfTx).
Reporter gene assay. GH 4 C 1 c-fos-luc cells were seeded in a 96-well clear-bottom white plate (Corning Costar, Cambridge, MA) at a density of 30,000 cells/well in 100 mL culture media and allowed to incubate overnight to ensure cell attachment. Cells treated with pPfTx were incubated for 4 hr and those with ATP were incubated for 10 hr. All incubations were performed at 37°C with 5% CO 2 and 95% air. In experiments where oxidized-ATP (oxATP) was used, pretreatment of one group of wells with 400 mM oxATP was initiated 1 hr before cell treatment with increasing concentrations of either ATP, 2´,3´-(4-benzoyl)benzoyl ATP (BzATP) or pPfTx. After incubation, the experimental media was removed and 20 mL cell lysis buffer [1% Triton X-100, 5 mM Tris, 0.4 mM trans-1,2-diaminocyclohexane-N,N,N´,N´-tetraacetic acid (CDTA), 10% glycerol, pH 7.8, and 1 mM dithiothreitol (DTT)] was added to each well. Lysis was allowed to proceed at room temperature for 20 min; we then measured solubilized luciferase protein activity using a luminometer (LumiStar; BMG LabTechnologies, Durham, NC). The luminometer was programmed to inject each well with 20 µL of Luciferase Assay Reagent (Promega, Madison, WI), and read the luminescence generated for 10 sec.
Immunostaining. We performed immunostaining for P2X7 receptors using cell homogenates of GH 4 C 1 cells on Western transfers. GH 4 C 1 cells were removed from 100-mm dishes with PBS-EDTA and resuspended in PBS containing protease inhibitors (4 µM phenylmethylsulfonyl fluoride and 2 µg/mL each of pepstatin, leupeptin, trypsin inihibitor, and aprotinin). Cells were lysed by freeze-thawing and then sonicated at 50 W, three pulses of 20 sec on ice. The lysates were centrifuged at 10,000 × g for 20 min, and supernatants were separated by 7.5% SDS-polyacrylamide gel electrophoresis. Separated proteins were then transferred to nitrocellulose and incubated with 1:200 rabbit anti-rat P2X7 antibody

Effect of ATP and pPfTx on c-fos luciferase.
We examined ATP for its ability to mimic the action of pPfTx in GH 4 C 1 cells. A biphasic luciferase response was generated from the induction of the GH 4 C 1 cells with increasing concentrations of ATP (Figure 1). We observed a half-maximal effect at 30 mM, with a maximal effect occurring at 200 mM. Concentrations of ATP that exceeded 200 mM caused a concentration-dependent inhibition of c-fos luciferase. This decrease is associated with cytotoxicity as determined by MTT cytotoxicity assay. A similar biphasic c-fos luciferase response was generated by the addition of serial dilutions of pPfTx to GH 4 C 1 cells (Figure 2). These results indicate that pPfTx mimics the action of ATP to induce c-fos luciferase and cytoxtoxicity in GH 4 C 1 cells and lead us to conduct preliminary characterization of the ATP receptor on GH 4 C 1 cells.
Analog characterization of the ATP receptor in GH 4 C 1 cells. The preliminary characterization of the ATP receptor was determined by conducting analog selectivity studies. The primary analogs tested that were effective in this study are shown in Figure 3. We first tested the moderately selective P2 antagonist, pyridoxalphosphate-6-azophenyl-2´,4´-disulfonic acid (PPADS). PPADS caused concentration-dependent inhibition of c-fos luciferase in the presence and absence of added ATP (Figure 4). The inhibition of c-fos luciferase by PPADS was not associated with cytotoxicity. We next examined the P2X1 and P2X3 subtype selective agonist α,β-methyleneadenosine 5´-triphosphate (α,β-MeATP). α,β-MeATP failed to increase or decrease c-fos luciferase activity ( Figure 5). Taken together, these results indicate that if P2X receptors mediate the effects of ATP on c-fos luciferase in GH 4 C 1 cells, the receptor is not of the P2X1 or P2X3 subtype.
We next examined a second antagonist, oxATP, which is an irreversible P2X antagonist with moderate selectivity for P2X7 receptors. Pretreatment with 400 µM oxATP inhibited the majority of the effect of ATP to increase c-fos luciferase, and fully inhibited the effects of ATP to decrease luciferase ( Figure  6). oxATP, unlike PPADS, did not decrease c-fos luciferase activity. We next tested the action of an agonist, BzATP, that shows selectivity for P2X7 receptors. BzATP did not increase c-fos luciferase activity, but it caused concentration-dependent inhibition of c-fos luciferase activity (Figure 7). The half-maximal effect of BzATP was nearly 10 times lower than the half-maximal effect of ATP to inhibit c-fos luciferase. This is consistent with an action on P2X7 subtype purinoreceptors. The failure of BzATP to induce c-fos luciferase was unexpected. It is possible that     BzATP only affected the second component of the biphasic response (i.e., cytotoxicty but not induction of the reporter gene). An alternative possibility is that BzATP induces both responses but has greater efficacy for cytotoxicity. We addressed this question by testing BzATP at the shorter incubation period of 4 hr and found that BzATP did cause a biphasic response (data not shown). We also examined the effect of oxATP on the action of BzATP. Pretreatment of 400 mM oxATP fully inhibited the effect of BzATP to decrease luciferase (Figure 7), which is consistent with an effect mediated by purinogenic receptors of the P2X7 class. Identification of the P2X7 receptor by immunoblotting. We examined the presence of the P2X7 receptor by immunostaining Western transfers of GH 4 C 1 cell membranes. GH 4 C 1 cells expressed an approximate 70-kDa band that was immunoreactive to a rabbit antibody directed to the intracellular carboxyl terminal domain, unique to the P2X7 class of purinergic receptors (Figure 8). We examined the specificity of the staining of the 70-kDa band by preabsorption of the primary antiserum with 10 µg/mL of the carboxyl terminal peptide sequence 575-595. No immunostaining of the 70-kDa band was evident under matched conditions (data not shown). This result provides an additional line of evidence for the presence of P2X7 receptors on GH 4 C 1 cells.
Testing pPfTx for ATP activity. We examined the role of P2X7 receptors in the action of the pPfTx. We sought to determine whether the pPfTx contained any activity attributable to ATP. We used an ATPdependent luciferase assay to quantify ATP. The sensitivity of the assay was 40 nM; 4 µM ATP generated a 100-fold increase in response ( Figure 9). pPfTx, given in an amount that caused a maximal induction of c-fos luciferase in the reporter gene assay, failed to mimic any effect of ATP to activate the luciferase enzyme directly. The pPfTx contained < 40 nM ATP, but by the reporter gene assay, it contained 200 µM ATP equivalents, indicating that the effect of pPfTx in the reporter gene assay is not attributable to ATP.
Effect of P2X antagonists on pPfTx induction of c-fos luciferase. We used two P2 antagonists of differing selectivity to examine the role of P2X receptors in the action of pPfTx to induce c-fos luciferase and cytotoxicity. PPADS, a P2 agonist, given at 200 µM inhibited both the activity of pPfTx and ATP ( Figure 10). oxATP, an irreversible P2X antagonist that has selectivity for P2X7 receptors, was added at a concentration of 400 µM as a pretreatment 1 hr before the addition of increasing concentrations of both ATP and pPfTx ( Figure 11). oxATP inhibited the luciferase induction of both ATP and pPfTx, suggesting that both substances induce c-fos luciferase and cytotoxicity by a common mechanism involving a P2X7 subtype receptor.

Discussion
ATP has a dual role as both an energy source for enzymatic reactions and as a first messenger for several classes of G protein-coupled receptors and ligand-gated ion channels. ATP was found to induce c-fos luciferase in GH 4 C 1 cells with a characteristic biphasic response. pPfTx also induced c-fos luciferase in a similar manner. These results suggest that the pPfTx is either ATP or an ATP agonist. We examined whether pPfTx was ATP using two lines of evidence. The first was that ATP and pPfTx do not share common      chromatographic retention properties (data not shown). The second was that pPfTx could not mimic the action of ATP to catalyze isolated luciferase enzyme in the presence of cofactors. Taken together, these results indicate that pPfTx is not ATP, but rather an ATP agonist. ATP activates receptors of the purinogenic P2 class (19). P2 receptors are divided into two classes-P2X and P2Y-based on molecular structure. P2X receptors are ATPactivated ion channels, and P2Y receptors are ATP-activated G protein-coupled receptors. We began by examining analog selectivity for induction of c-fos luciferase by ATP using P2 agonists and antagonists of differing selectivity. PPADS, a P2 antagonist of moderate selectivity for P2X receptors (19), caused concentration-dependent inhibition of ATP induction of c-fos luciferase in GH 4 C 1 cells with a half-maximal effect (100 µM), which is consistent with an action on P2 receptors. Further examination with α,β-methylene ATP (α,β-MeATP), both a P2X1-and P2X3-selective ATP agonist (19,20), indicated that α,β-MeATP did not affect c-fos luciferase, providing additional evidence that the pathway of activation was neither a P2X1 or P2X3 receptor. A second antagonist, oxATP, which is irreversible and more selective for P2X receptors (21), inhibited nearly fully the actions of ATP on c-fos luciferase in GH 4 C 1 cells. Because oxATP has been reported to have some degree of selectivity for P2X7 receptors, we tested an agonist, BzATP, which has selectivity for P2X7 receptors (21)(22)(23). BzATP failed to increase c-fos luciferase in GH 4 C 1 cells, but it did cause a concentration-dependent inhibition of luciferase activity. BzATP was nearly 10 times more potent at inhibiting cfos luciferase than ATP. BzATP induces greater maximal ion conductance than ATP in cells expressing P2X7 receptors; this may be the basis for the greater efficacy of BzATP for cytotoxicity (21). The action of BzATP was fully inhibited by oxATP. Taken together, these analog selectivity experiments with BzATP and oxATP are consistent with the presence of P2X7 receptors on GH 4 C 1 cells. The presence of P2X7 receptors on GH 4 C 1 cells was additionally supported by the immunostaining of Western transfers of GH 4 C 1 cell membranes using an antibody specific to the unique carboxy terminal domain of the P2X7 receptor.
ATP-activated ion channel receptors were originally identified in mast cells (24) and later found in other myeloid-derived cells, including macrophages and microglia (25,26). These channels, designated P2Z, were subsequently found to be of the P2X class and renamed P2X7 (27). Although the P2X7 receptor subtype is found largely in cells of immune origin, it has also been identified in cell lines and primary cell cultures of nonimmune origin (18,28,29). The role for a P2X7 receptor in GH 4 C 1 cells has not been determined; however, it is well known that this cell lineage (growth hormone-producing cells) mediates local inflammatory responses in the pituitary gland and may modulate the hypothalamic pituitary axis during systemic inflammatory reactions (30).
The antagonists PPADS and oxATP, which we used to characterize the P2X receptor in GH 4 C 1 cells, were also useful in examining the action of pPfTx. The pathway(s) leading to the biphasic effect on c-fos luciferase of pPfTx in GH 4 C 1 cells appears to be mediated by the same receptor that mediates the response to ATP. Both PPADS and oxATP inhibited pPfTx induction of c-fos luciferase. These results are consistent with pPfTx acting as a P2X7 receptor agonist. Although these results do not prove that P2X7 is the initial cellular target for the putative toxin, they do indicate that this receptor is necessary in the signal transduction pathway. At this point it is not possible to exclude effects of pPfTx on additional receptors, including other P2X receptor subtypes. This may be most readily determined using expression systems for various cloned receptors.
It remains to be determined whether the P2X7 agonist activity isolated from P. piscicida is responsible for the wildlife effects associated with this organism. Macrophages and mast cells express P2X7 receptors, which have been suggested to have a role in inflammation. The entry of monocytes into peripherial tissues precedes their differentiation into activated macrophages, a process that involves the action of interferon-γ, which in turn leads to expression of P2X7 receptors (31). In activated macrophages, P2X7 receptors mediate chronic inflammatory responses normally driven by ATP. The responses include fusion of macrophages into multinucleated giant cells and several inflammatory responses that result from production of interleukin 1-β, including release of prostaglandins, production of matrix, and chemoattraction of neutophils (32)(33)(34). These responses are characteristic of granulatomatous lesions found in fish that are associated with toxic P. piscicida (9,35). Because P. piscicida has the capacity to phagocytize blood cells and cause tissue injury (2), it may initiate an acute inflammation that is potentiated to a chronic response by pPfTx, behaving as a potent ATP mimic at P2X7 receptors on activated macrophages.
Whether the P2X7 agonist activity isolated from P. piscicida contributes to the human neurocognitive effects associated with this organism is less obvious. P2X7 receptors in the central nervous system have been best characterized in microglia (31,36). Microglia are the central nervous system counterpart to tissue macrophages and they normally provide a defensive inflammatory response to infections and tissue damage (37). However, inappropriate activation of microglia can elicit neurotoxic effects that may include release of excitotoxic amino acids and cytolytic and inflammatory agents (37). One approach used to study the effects of Pfiesteria on neurocognitive impairment is a rat model using radial arm-maze testing (38).
Our results indicate that the cytotoxic effect originally described for a putative P. piscicida toxin is mediated by a P2X7 receptor. Based on the current understanding of the role of P2X7 receptors in disease and the observed effects directly attributable to exposure to P. piscicida toxins, P2X7 receptor-mediated Articles • Purinergic receptors, a target for putative Pfiesteria toxins Environmental Health Perspectives • VOLUME 109 | NUMBER 5 | May 2001 461 Figure 11. oxATP and PPADS inhibition of ATP and pPfTx induction of c-fos luciferase in GH 4 C 1 cells pretreated with 400 mM oxATP for 1 hr then treated with increasing concentrations of ATP and pPfTx. RLU, relative light units. See "Materials and Methods" for details. The oxATP treatment did not differ from the vehicle [Milli-Q-treated water (Millipore, Bedford, MA), final concentration 2.5%] but did cause a significant (Tukey multiple comparison test, p < 0.05) difference in pPfTx induction of c-fos luciferase. Each point represents the mean ± SE of three wells for an experiment repeated once with similar results. Figure 10. PPADS inhibition of pPfTx induction of c-fos luciferase in GH 4 C 1 cells treated with ATP or pPfTx. RLU, relative light units. See "Materials and Methods" for details. The ATP and pPfTx treatments differed (Tukey multiple comparison test, p < 0.05) from the vehicle [Milli-Q-treated water (Millipore, Bedford, MA), final concentration 2.5%] for the control group but not for the PPADS group. Each point represents the mean ± SE of three wells for an experiment repeated once with similar results. 16