Introduction

P2X4 receptors belong to the P2X receptor family of ATP-activated ionotropic membrane receptors comprising seven receptor subunits, P2X1-P2X7 [1]. Crystallization and cryogenic electron microscopy (cryo-EM) structures of the zebrafish P2X4 receptor confirmed its previously proposed homotrimeric structure [2,3,4,5]. Each protein subunit is composed of two transmembrane domains (TM1 and TM2) and a large extracellular loop [2, 3]. The crystal structure of the zebrafish P2X4 receptor revealed the location of three ATP-binding sites in the extracellular domains, one in each subunit [2, 3]. Binding of the physiological agonist ATP induces receptor subunit rearrangement and ion channel opening for Na+, Ca2+ and K+ ions [3, 6]. In addition to homotrimeric receptors, heterotrimeric receptors exist, e.g. P2X2/3 receptors [1, 4, 7].

P2X4 receptors are expressed in the central and peripheral nervous system, especially in astrocytes, neurons, microglia and endothelial cells [1, 8]. A number of studies have shown that P2X4 receptors are involved in neuropathologies [9]; they appear to play a central role in neuropathic pain [10]. After peripheral nerve injury, microglial cells are activated, and P2X4 receptor expression is strongly upregulated [11,12,13,14,15]. This results in the release of brain-derived neurotrophic factor (BDNF) which initiates a cascade of neuroinflammatory signals resulting in pain and hypersensitivity. The pharmacological blockade of P2X4 receptors was reported to reverse tactile allodynia, a prominent symptom of neuropathic pain [11,12,13,14, 16]. P2X4 receptor blockade might also be beneficial for the treatment of other types of chronic pain, e.g. cancer pain [17], inflammatory pain [18, 19], and visceral pain [20]. An upregulation of P2X4 receptors in reactive microglia or in neurons, respectively, was also observed in Alzheimer’s [21] and Parkinson’s disease [22], amyotrophic lateral sclerosis (ALS) [23], epilepsy [24], multiple sclerosis [25] and stroke [26, 27]. The role of P2X4 receptors in all of these neuroinflammatory processes suggests that P2X4 receptor antagonists might be effective drugs for the treatment of associated diseases.

Besides their crucial role in neuroinflammatory diseases, P2X4 receptors have emerged as potential targets for the treatment of cancer, including breast [28], prostate [29], gastric [30], colon [31] and renal cancer [32]. Recently, P2X4 receptors have been reported to play an important role in the fine-tuning of the basal activity and the initial activation of CD4+ T cells [33], indicating that the receptors are involved in immune responses to pathogens and cancer cells.

The discovery of drug-like competitive P2X4 receptor antagonists is hampered by the polar ATP binding site lined with basic amino acid residues [3, 4]. However, several allosteric modulators have been identified and in some cases optimized [34,35,36]. P2X4 receptor antagonists include benzodiazepine derivatives, e.g. 5-BDBD and derived compounds, e.g. NP-1815-PX [37], and its analog MRS4719 [38], phenoxazine derivatives, e.g. PSB-12054 and PSB-12062 [39], sulfonamide derivatives, e.g. BAY-1797 [40], and the urea derivative BX430 and its analogs, e.g. compound 9o (Fig. 1) [41, 42].

PSB-15417 was developed as a potent, highly selective and brain-permeable allosteric P2X4 antagonist by our group showing efficacy in an animal model of neuropathic pain [14, 36]. NC-2600 (structure undisclosed) was the first P2X4 receptor antagonist to enter clinical trials (phase I) [43]. The effects of NP-1815-PX and NC-2600 were evaluated in a murine model of colitis and found to improve several inflammatory parameters [44]. MRS4719 showed neuroprotective and neuro-rehabilitative effects in a murine ischemic stroke model [38].

Fig. 1
figure 1

P2X4 receptor antagonists with inhibitory potencies (IC50) at the human P2X4 receptor. Müller et al. [36]; Matsumara et al. [37]; Toti et al. [38]; Hernandez-Olmos et al. [39]; Werner et al. [40]; Ase et al. [41]; Mahmood et al. [42]

Full size image

Various assays have been developed for studying the potency of P2X receptor ligands. These include calcium influx assays [45], patch-clamp measurements [46], and radioligand binding assays [47, 48]. Radioligands are powerful tools in pharmacology and drug discovery for measuring direct ligand-protein interactions, determining affinities and binding kinetics, including those of unlabeled compounds by performing competition assays. Moreover, radioligands are used for receptor imaging in vitro and in vivo, for autoradiography, scintigraphy (SPECT, single photon emission computer tomography) and positron emission tomography (PET) studies. The orthosteric binding site of the P2X4 receptor can be labeled by the agonist radioligand [35S]ATPγS, an ATP analog with higher metabolic stability than ATP [48, 49].

Potent and selective antagonist radioligands for the labeling of allosteric P2X receptor binding sites have so far only been described for P2X3 and P2X7 receptors [47, 50]. Attempts to develop PET radioligands for the P2X4 receptor were described, based on 5-BDBD, which was radiolabeled with 11C, 18F, or 76Br. However, in vitro binding studies with these radiolabeled 5-BDBD analogs have not been successful [51]. This is not surprising given the moderate potency and low water-solubility of 5-BDBD (Fig. 1).

Here, we present the development of a potent, selective allosteric antagonist radioligand for the P2X4 receptor designated [3H]PSB-OR-2020. This radiotracer has allowed the development of the first antagonist radioligand binding assay for the human P2X4 receptor.

Results

In the search for novel scaffolds with P2X4 receptor-antagonistic activity, we had previously established assays allowing high throughput screening [49]. This led to the discovery of indolylcarboxymethylaniline derivatives which were subsequently optimized resulting in highly potent P2X4 receptor antagonists, e.g., PSB-15417, with IC50 values in the low nanomolar range determined in calcium influx assays (patent application filed in December 2022). Based on the discovery of this new scaffold, we developed a series of related dihydroindole analogs and selected one of the most potent antagonists of this series, (S)-1-(4-(1,2-dihydroxypropane-2yl)indolin-1-yl)-2-((2-methyl-5-(3-methyl-1,2,4-thiadiazol-5-yl)phenyl)amino)ethane-1-one, designated PSB-OR-2020, for the preparation of a radioligand (see Fig. 2). The selection was also based on sufficient polarity of the compound to avoid high non-specific binding.

Characterization of the unlabeled P2X4 receptor antagonist PSB-OR-2020 in calcium influx assays

The unlabeled PSB-OR-2020 was characterized as a highly potent and selective P2X4 receptor antagonist. Calcium influx assays using 1321N1 astrocytoma cells that stably expressed the human P2X4 receptor were performed. Concentration-dependent inhibition of ATP-induced calcium influx was observed. The employed ATP concentration corresponded to its EC80 value (300 nM), and an IC50 value of 6.32 ± 1.52 nM was determined for the antagonist PSB-OR-2020 (see Fig. 2).

Fig. 2
figure 2

Concentration-dependent inhibition of ATP-induced Ca2+ influx by PSB-OR-2020 determined in 1321N1 astrocytoma cells stably transfected with the human P2X4 receptor. An EC80 of ATP was applied (300 nM) after preincubation of the cells with various concentrations of PSB-OR-2020. Data points represent means ± SEM of four independent experiments performed in duplicates. IC50 value: 6.32 ± 1.52 nM

Full size image

The selectivity of PSB-OR-2020 versus other P2X receptor subtypes was determined in the same type of assay using 1321N1 astrocytoma cells stably expressing the respective human P2X receptor subtype (see Table 1).

Table 1 Potency of PSB-OR-2020 as antagonist of human P2X receptor subtypes
Full size table

At the other human P2X receptor subtypes, P2X1, P2X2, P2X3, and P2X7, inhibition of the agonist-induced signal by a high concentration of PSB-OR-2020 (10 µM) was well below 50%, except for the P2X3 receptor. Thus, a concentration-inhibition curve was determined, resulting in an IC50 value of 2,030 ± 480 nM at the P2X3 receptor. This value is about 300-fold higher than the compound’s IC50 value at the P2X4 receptor. Thus, PSB-OR-2020 is not only a potent, but also a highly subtype-selective P2X4 receptor antagonist.

Radiolabeling

Having a P2X4 receptor antagonist with low nanomolar potency and high selectivity in hand, we designed a precursor molecule that would be suitable for catalytic hydrogenation with tritium gas (3H2) to obtain PSB-OR-2020 in 3H-labeled form. Thus, we introduced two Br atoms, one on each aromatic ring of the scaffold, which was expected to result in sufficiently high specific radioactivity after tritiation due to the possible introduction of two tritium atoms. The final tritiation of the brominated precursor was performed by catalytic hydrogenation (custom-labeling), which led to the displacement of the bromine by tritium atoms (see Fig. 3).

[3H]PSB-OR-2020 was obtained in high purity (97.7%). It showed a specific activity of 45 Ci/mmol (1.67 TBq/mmol) indicating a good tritiation efficiency (introduction of ca. 1.5 tritium atoms per molecule, for analytical details see Figs. S1 and S2 of Supporting Information).

Fig. 3
figure 3

Synthesis of [3H]PSB-OR-2020 by catalytic hydrogenation with tritium gas

Full size image

Development of a radioligand binding assay

As a next step, the new radioligand was utilized to establish a P2X4 receptor radioligand binding assay. The initially observed binding of [3H]PSB-OR-2020 to GF/B glass fiber filters was relatively high (up to 2000 cpm at a radioligand concentration of 10 nM) using ice-cold Tris-HCl buffer (50 mM, pH 7.4) in a final volume of 1 mL at 4 °C, and ice-cold Tris-HCl buffer (50 mM, pH 7.4) containing 0.1% bovine serum albumin (BSA) as washing buffer. The GF/B filters were dried for 30 min at room temperature before transferring them into scintillation vials. To reduce filter binding, several experimental parameters were varied, including buffer, additives, filter types, and filter pretreatment. Reduced filter binding (to ca. 500 cpm at a radioligand concentration of 10 nM) was achieved by using an assay buffer composed of 50 mM Tris-HCl, pH 7.4, containing 0.1% BSA, incubation at 4 °C, a washing buffer consisting of ice-cold 50 mM Tris-HCl, pH 7.4, and filtration through GF/C filters, which were dried at 70 °C for 60–90 min after harvesting (see Fig. 4). GF/C glass fiber filters are thinner and have a larger pore size of 1.2 μm, in comparison to GF/B filters (1.0 μm).

Fig. 4
figure 4

Optimization of filter binding. Gray bars: initial GF/B filter binding of [3H]PSB-OR-2020 (assay buffer: ice-cold 50 mM Tris-HCl, pH 7.4; assay volume: 1 mL; washing buffer: ice-cold 50 mM Tris-HCl, pH 7.4 containing 0.1% BSA; incubation on ice for 30 min; filter was dried at room temperature for 30 min). Green bars: optimized assay conditions showing decreased non-specific radioligand binding to GF/C filters (assay buffer: ice-cold 50 mM Tris-HCl, pH 7.4 containing 0.1% BSA; assay volume: 500 µL; washing buffer: ice-cold 50 mM Tris-HCl, pH 7.4; incubation on ice for 30 min; filter was dried at 70 °C for 60–90 min)

Full size image

As a next step, experiments with membrane preparations derived from recombinant, P2X4 receptor-expressing 1321N1 astrocytoma cells were performed. In these cells, the human P2X4 receptor was stably expressed using a retroviral expression system [49]. Non-specific binding was determined in the presence of unlabeled PSB-OR-2020 (either 500 nM or 10 µM). Preliminary experiments using 3 nM or 7 nM of [3H]PSB-OR-2020, respectively, indicated specific binding of [3H]PSB-OR-2020 to the P2X4 receptor-expressing cell membrane preparations, but non-specific binding was still high (> 50% of total binding, data not shown).

To decrease non-specific radioligand binding to the cell membrane preparations, various compositions of assay and washing buffer were tested. The washing buffer was supplemented with magnesium chloride, copper chloride, ethylenediamine tetraacetic acid (EDTA), ethylene glycol tetraacetic acid (EGTA), acetic acid or sodium phosphate. Since the radioligand is well soluble in ethanol, 5% of ethanol were added to the washing buffer (ice-cold 50 mM Tris-HCl, pH 7.4) to remove unbound radioligand. Preincubation of GF/C filters in polyethylenimine solutions of different concentrations, in trimethylchlorosilane solution, or in washing buffer containing 5% of ethanol, was investigated. However, none of these modifications led to a sufficient reduction in non-specific radioligand binding, and the percentage of specific binding of [3H]PSB-OR-2020 was still too low. The radioligand’s non-specific binding appeared to be mainly attributed to binding non-specifically to the membrane preparation, while filter binding was moderate after it had been optimized. The low percentage of specific radioligand binding may additionally be explained by a rather low expression level of the P2X4 receptor.

Thus, new cell lines were created with the aim to obtain higher receptor expression. Membrane preparations of the resulting cell lines were used for radioligand binding studies to select the most highly expressing, and thus best suitable one. In addition to the previously described 1321N1 astrocytoma cell line, two further P2X4 receptor-expressing cell lines were studied: (i) HEK293 cells transiently expressing the human P2X4 receptor (plasmid: pQCXIN), (ii) HEK293 cells transiently expressing the human P2X4 receptor (plasmid: pcDNA3.1(-)). HEK293 cells transiently expressing the human P2X4 receptor (plasmid: pQCXIN) showed a higher expression level and resulted in a consistent, reproducible, and sufficiently low percentage of non-specific binding.

The optimal conditions for the binding assays were as follows: a final volume of 500 µL of Tris-HCl buffer (50 mM), pH 7.4, containing 0.1% BSA, 10 nM radioligand, 50–150 µg of protein (membrane preparation of HEK293 cells recombinantly expressing P2X4), incubation for 60 min at 4 °C, filtration through GF/C filters, and washing with ice-cold Tris-HCl buffer (50 mM), pH 7.4, containing 5% ethanol. Non-specific binding was determined in the presence of unlabeled PSB-OR-2020 (10 µM).

To characterize the binding kinetics of the radioligand, we performed association and dissociation experiments at 4 °C (see Fig. 5A and B). [3H]PSB-OR-2020 associated with a half-life of 8.44 ± 2.59 min, while the dissociation half-life was 20.4 ± 2.9 min. Three independent association and dissociation experiments were performed to determine the kinetic KD value (koff/kon) for [3H]PSB-OR-2020, which was calculated to be 7.06 ± 3.23 nM.

Fig. 5
figure 5

Representative curves for association (A) and dissociation (B) of [3H]PSB-OR-2020 (10 nM) to HEK293 cell membrane preparations recombinantly expressing the human P2X4 receptor, performed at 4 °C

Full size image
Fig. 6
figure 6

Competition binding of different concentrations of PSB-OR-2020 vs. [3H]PSB-OR-2020 (10 nM) using membrane preparations of HEK293 cells recombinantly expressing the human P2X4 receptor. Data points represent means ± SEM of three independent experiments performed in duplicates. KD value of PSB-OR-2020: 20.4 ± 7.5 nM; Bmax value: 71.7 fmol/mg of protein

Full size image

We subsequently performed homologous competition binding studies, and determined a concentration-inhibition curve for the unlabeled ligand PSB-OR-2020 (see Fig. 6). A KD value of 20.4 ± 7.5 nM was calculated from three independent experiments performed in duplicates. The Bmax value was 71.7 fmol/mg of protein, and thus still not very high. The KD value correlated well with the kinetic KD value (7.06 ± 3.23 nM) and also with the IC50 value of the unlabeled PSB-OR-2020 determined in calcium influx assays (6.32 ± 1.52 nM) versus the EC80 concentration of the agonist ATP (300 nM).

As a next step, the effects of structurally diverse modulators of the human P2X4 receptor were studied (Fig. 7). Based on the IC50 values determined in calcium influx assays, and considering the compounds’ solubilities, high concentrations were employed that were expected to result in a complete blockade of the receptor. PSB-15417 completely displaced the radioligand from its binding site, suggesting that the radioligand and PSB-15417 (both being structurally related compounds) bind to the same allosteric site of the human P2X4 receptor. In contrast, a high concentration of ATP only slightly displaced the radioligand from its binding site, providing evidence for an allosteric binding mode of [3H]PSB-OR-2020 (Fig. 7). The structurally diverse P2X4 receptor antagonists BX430, 5-BDBD, and BAY-1797 (for structures see Fig. 1) also only partially displaced the radioligand from its binding site by less than 40%, even though high concentrations were employed. Thus, we assume that 5-BDBD, BX430 and BAY-1797 bind to a site or sites different from that of [3H]PSB-OR-2020.

Fig. 7
figure 7

Specific [3H]PSB-OR-2020 binding to the human P2X4 receptor in the absence and presence of ATP and various P2X4 receptor antagonists utilizing membrane preparations of HEK293 cells transiently transfected with the human P2X4 receptor. Data represent means of three independent experiments performed in duplicates and triplicates. The final concentrations of the compounds are shown in brackets

Full size image

Discussion

Purinergic signaling is involved in many physiological processes [52]. The balance between extracellular concentrations of pro-inflammatory ATP and anti-inflammatory adenosine is crucial for homeostasis [53, 54]. Under inflammatory conditions, extracellular ATP concentrations may rise leading to the activation of P2X receptors [55]. These receptors are currently in the focus of interest as novel drug targets. The first P2X3 receptor antagonist, gefapixant, has recently been approved for the treatment of chronic cough, characterized by lung inflammation [56, 57]. P2X4 receptor antagonists are potential drug targets for inflammatory and neurodegenerative diseases including neuropathic pain [58], inflammatory pain [58], Parkinson’s disease [22], Alzheimer’s disease [21], multiple sclerosis [59], epilepsy [24], and cancer [28,29,30]. To study P2X4 receptor pharmacology, suitable tool compounds are required. Up to now, no radio- or fluorescence-labeled antagonist suitable for P2X4 receptor binding studies has been developed. This might be attributed to the fact that most of the P2X4 receptor antagonists described so far are only moderately potent and/or display low water-solubility resulting in high non-specific binding to cell membranes. In the present study, we describe the development of the tritium-labeled P2X4 receptor antagonist [3H]PSB-OR-2020 belonging to a novel chemical class of allosteric P2X4 receptor antagonists, and present its preliminary characterization. Major hurdles on the way to develop a binding assay were the compound’s relatively high glass fiber filter binding, and its non-specific binding to cell membranes. Moreover, it was hampered by the low expression of the P2X4 receptor in the available cell lines. Nevertheless, we managed to perform kinetic studies to determine the radioligand’s kinetic KD value of 7.06 nM, and competition binding assays resulting in a KD value of 20.4 nM. These data confirmed the high affinity of [3H]PSB-OR-2020, which is well in agreement with its IC50 value obtained in calcium influx assays.

In future studies, more highly expressing cell lines, or isolated P2X4 receptor protein reconstituted into nanodiscs or amphipols, could be advantageous to further increase specific binding and to reduce non-specific binding.

Our radioligand represents an allosteric P2X4 receptor antagonist with a novel chemical structure. ATP does not compete with its binding to the receptor. Since the radioligand could also not or only partially be displaced by some previously published P2X4 receptor antagonists, including BX430, 5-BDBD, and BAY-1797, we assume that those antagonists bind to different allosteric binding sites.

Different allosteric binding sites, besides the orthosteric site, have been described for P2X receptor antagonists, as revealed by X-ray crystallography, cryo-EM, and mutagenesis studies [5, 36, 60, 61, 62]. Recently, the allosteric binding site of BX430 and BAY-1797 on the zebrafish P2X4 receptor was determined by cryo-EM to be located in the extracellular domain of the receptor, at the inter-subunit interface of the receptor trimer [5]. Our competition binding assays showed that the binding site of the new radioligand [3H]PSB-OR-2020 at the human P2X4 receptor, to which its structurally related antagonist PSB-15417 binds, likely differs from the allosteric binding site of the structurally different P2X4 receptor antagonists BX430 and BAY-1797, and thus appears to occupy a novel allosteric site.

Conclusion

In conclusion, we have successfully designed and synthesized a P2X4 receptor radioligand which allowed for the first time to establish a binding assay for this therapeutically relevant receptor. [3H]PSB-OR-2020 may become a useful tool in P2X4 receptor pharmacology and drug development.

Methods

Chemicals

Chemicals were obtained from Roth (Karlsruhe, Germany), AppliChem (Darmstadt, Germany) or Thermo Fisher Scientific (Darmstadt, Germany). The radioligand [3H]PSB-OR-2020 (45 Ci/mmol; 1.67 TBq/mmol) was produced by custom-labeling (Pharmaron UK Ltd., Cardiff, UK), see Supporting Information.

Retroviral transfection

The 1321N1 astrocytoma cell line stably expressing the human P2X4 receptor was generated using GP + envAM-12 fibroblast-type mouse packaging cells to produce a recombinant helper virus as described before [63]. To produce recombinant helper viruses, 6.25 µg of the plasmid DNA (human P2X4 receptor DNA in pQCXIN) and 3.75 µg of vesicular stomatitis virus G (VSV-G) protein DNA (in pcDNA3.1) were combined. Subsequently, packaging cells were transfected using lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) as a transfection reagent. After harvesting the virus particles, the target cells (non-transfected 1321N1 astrocytoma cells) were infected. 48–72 h after infection, the selection was started using the selection marker G418 (800 µg/mL). Cells were cultured in Dulbecco’s Modified Eagle medium (DMEM) (Thermo Fisher Scientific, Darmstadt, Germany), 10% fetal calf serum (FCS) (PAN Biotech, Aidenbach, Germany), a mixture of penicillin/streptomycin (100 U/mL penicillin, 100 µg/mL streptomycin) (PAN Biotech, Aidenbach, Germany), and the selection antibiotic geneticin G418 (800 µg/mL). Cells were grown at 37 °C and 10% CO2.

Transient transfection

HEK293 cells were transiently transfected with the human P2X4 DNA in pQCXIN or pcDNA3.1(-). Transfection was conducted with 10 µg of the plasmid DNA and a mixture of Opti-MEM™ medium (Thermo Fisher Scientific, Darmstadt, Germany) and lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) that served as the transfection reagent. The selection was started after 24–72 h using DMEM, 10% FCS, a mixture of penicillin/streptomycin (100 U/mL penicillin, 100 µg/mL streptomycin) and G418 (800 µg/mL). Cells were grown at 37 °C and 5% CO2.

Membrane preparation

Cells expressing the human P2X4 receptor were grown to 70–90% confluence in cell culture dishes at 37 °C and 10% CO2. After removing the medium, the cells were washed with phosphate buffer (phosphate-buffered saline (PBS)), and frozen at -20 °C. The next day, the cells were scraped off the dishes after adding 1 mL of cold hypotonic buffer (50 mM Tris-HCl, 1 mM EDTA, pH 7.4) per dish. All steps were carried out on ice. After homogenizing the cell suspension using an UltraTurrax® (IKA Labortechnik, Staufen, Germany), the homogenate was centrifuged for 10 min at 1,000 g and 4 °C. The supernatant was centrifuged again for 60 min at 48,000 g and 4 °C. The remaining pellet was resuspended in 50 mM Tris-HCl buffer, pH 7.4, and stored in aliquots at -80 °C.

Protein concentration determination with the Bradford assay [64]

To estimate the protein concentration of the membrane preparations, a Bradford assay was performed. The Bradford stock solution contained 0.1 g of Coomassie-Brillant Blue G dissolved in 50 mL ethanol (v/v) with 100 mL of 85% phosphoric acid, diluted to 1000 mL with water. A BSA calibration curve (100 µg/mL − 500 µg/mL) was prepared in 50 mM Tris-HCl buffer, pH 7.4. The membrane preparations were diluted 1:10 and 1:20 with 50 mM Tris-HCl buffer, pH 7.4. Then, 1000 µL of the Bradford stock solution and 50 µL of the sample were mixed in a cuvette. The samples were measured at 595 nm and the final protein concentration was determined based on a BSA calibration curve.

Radioligand binding studies

Kinetic and competition binding studies using membrane preparations from HEK293 cells recombinantly expressing the human P2X4 receptor were performed in 50 mM Tris-HCl, pH 7.4, and 0.1% BSA at 4 °C, in a final assay volume of 500 µL. The assay vials contained either 5 µL of the test compound (dissolved in dimethyl sulfoxide (DMSO)) or 5 µL of DMSO, 50 µL of [3H]PSB-OR-2020 in assay buffer (10 nM), and 50 µL of protein in 50 mM Tris-HCl buffer, pH 7.4. Non-specific and total binding of [3H]PSB-OR-2020 were determined using 10 µM of unlabeled PSB-OR-2020 dissolved in DMSO, and 5 µL of DMSO, respectively. To adjust a suitable amount of protein, the specific binding of the radioligand was measured for each newly generated membrane preparation. After adding the cell membrane suspension (50–150 µg of protein), the incubation was started and lasted for 60 min at 4 °C with gentle shaking. For association experiments, 50 µL of the cell membrane suspension were added at specific time points over a 90 min period. For dissociation experiments, the radioligand was first incubated with the cell membrane suspension for 1 h. Then, 5 µL of unlabeled PSB-OR-2020 (10 µM) were added at specific time points (ranging from 180 min to 0 min) to initiate dissociation. The incubation was terminated by rapid vacuum filtration using a Brandel 24-well harvester (Brandel, Gaithersburg, MD, USA), and washing with 3 × 3 mL of ice-cold washing buffer (50 mM Tris-HCl buffer, pH 7.4, containing 5% ethanol) through Whatman GF/C glass fiber filters (GE Healthcare Life Sciences Whatman™, USA). The filters were preincubated in ice-cold washing buffer for 30 min. After filtration, the filters were punched out and soaked in 2.5 mL of scintillation cocktail (ProSafe FC+, Meridian Biotechnologies Ltd., Surrey, UK) for at least 6 h before being counted using a liquid scintillation counter (Tri-Carb® 2900 TR, Perkin-Elmer, Rodgau, Germany, 53% efficiency).

Calcium influx assay

To determine the effects of the compound on P2X4, P2X2, and P2X7 receptors, the agonist-mediated increases in cytosolic Ca2+ concentrations were measured using the fluorescent Ca2+ chelating dye Fluo-4 acetoxymethyl ester (AM) (Thermo Fisher Scientific). On the first day, 1321N1 astrocytoma cells (45.000 cells/well) stably expressing the respective P2X receptor were seeded into a 96-well plate (No. 3340, Corning, Kennebunk, Maine, USA), and incubated at 37 °C and 10% of CO2. The next day, the cells were incubated with Fluo-4 AM (3 µM) in Hanks’ balanced salt solution and 1% of Pluronic® F127 (Sigma Aldrich, St. Louis, MO, USA) for 1 h at room temperature with slight shaking. For human P2X7 receptors, a different assay buffer was used (150 mM Na-glutamate, 5 mM KCl, 0.5 mM CaCl2, 0.1 mM MgCl2, 10 mM D-glucose, 25 mM HEPES, pH 7.4). Calcium 5 (Molecular Devices, San Jose, CA, USA) was used as a fluorescent dye to investigate recombinantly expressed human P2X1 and P2X3 receptors. The cells were loaded with Calcium 5 dye in HBSS buffer, and incubated for 1 h at 37 °C and 10% CO2 for human P2X1, and 5% CO2 for human P2X3 receptors.

Following the incubation, the dye solution was carefully removed, and the antagonist (dissolved in DMSO) and HBSS buffer were added to the cells. The final DMSO concentration was 1% for human P2X4, P2X2, and P2X7 receptors, and 0.5% for human P2X1 and P2X3 receptors. To activate the receptors, transparent 96-well reagent plates (Boettger, Bodenmais, Germany) containing the agonists dissolved in HBSS buffer (10-fold concentration of previously determined EC80) were prepared. The plates were measured after 30 min of preincubation with antagonist with the imaging plate reader NOVOstar (BMG Labtech GmbH, Offenburg, Germany) at 520 nm for 30 s at 0.4 s intervals.

Data analysis

Data were analyzed using Microsoft Excel and PRISM Version 8.0 (GraphPad Prism 8, San Diego, CA).