Taspine is a natural product that suppresses P2X4 receptor activity via phosphoinositide 3‐kinase inhibition

P2X4 is a ligand‐gated cation channel activated by extracellular ATP involved in neuropathic pain, inflammation and arterial tone.


| INTRODUCTION
P2X receptors are a family of ligand-gated non-selective cation channels activated by extracellular adenosine 5 0 -triphosphate (ATP).

P2X receptor activation causes membrane depolarisation and
increased cytoplasmic Ca 2+ . The human genome encodes seven P2X receptor subtypes capable of forming homomeric and heteromeric trimer in a subtype-dependent fashion. P2X4 has attracted wide attention as a target for pharmacological modulation (Stokes et al., 2017). It is a modestly desensitising receptor (Fountain & North, 2006) and has an established role in a number of physiological processes. In the cardiovascular system, P2X4 is involved in cardiac contractility (Hu et al., 2002;Yang et al., 2014) and flow-dependent vasodilation and remodelling of arteries (Yamamoto et al., 2000;Yamamoto et al., 2006). In the pulmonary system, P2X4 activity regulates surfactant secretion in type II epithelial cells of alveoli (Miklavc et al., 2011). Activation of P2X4 is associated with inflammation involving peripheral and central inflammatory cell types (Layhadi et al., 2018;Layhadi & Fountain, 2019;Ulmann et al., 2008;Wareham et al., 2009). Up-regulation of P2X4 in spinal microglia contributes to neuropathic pain in preclinical models (Biber et al., 2011;Tsuda et al., 2003;Williams et al., 2019).
More recently compounds including PSB12054, PSB12062 (Hernandez-Olmos et al., 2012), BX430 (Ase et al., 2015) and NP-1815-PX (Matsumura et al., 2016) have been identified as P2X4 inhibitors. Natural products have also been a valuable source of P2X4 modulators. The best known is ivermectin, a bacteria-derived macrocyclic lactone that acts as a positive allosteric modulator of P2X4 receptors (Priel & Silberberg, 2004), but it is also active at other channels (Chen & Kubo, 2018). More recently, protopanaxadiol ginsenosides from the plant Panax ginseng have been identified as positive allosteric modulators of P2X4 (Dhuna et al., 2019). Despite recent advances in the identification of molecules that modulate P2X4 receptor activity (Stokes et al., 2017), a mechanistic description of how such molecules work is often limited.
We have identified that taspine, an alkaloid extract of the plant Croton lechleri, inhibits P2X4 receptor activity. Extracts of Croton lechleri cortex are used in traditional medicine by Amazonian tribes for wound healing (Perdue et al., 1979). The anti-inflammatory and cicatrizant properties of Croton lechleri extract are principally attributed to taspine (Vaisberg et al., 1989). Here, we describe the inhibitory properties of taspine on P2X4 receptor activity and its mechanism of action. were cultured in RPMI 1640 medium containing 10% (v/v) FBS, 50 UÁml À1 penicillin and 50 μgÁml À1 streptomycin. Human P2X2, P2X3 and P2X2/3 1321N1 stable cells were gifted through collaboration with Afferent Pharmaceuticals and described previously . BV-2 cells were exposed to 10 μgÁml À1 LPS for 24 h before experimentation. All cell lines were maintained in a humidified environment at 37 C and 5% CO 2 /95% air.

| Human primary macrophage
The use of human blood samples in this research was approved by the

What is already known
• Taspine is a natural product with anti-inflammatory activity.

What does this study add
• Taspine suppresses P2X4 receptor activity via PI3-kinase inhibition.
• Taspine inhibits pro-inflammatory signalling via inhibition of P2X4 receptors in macrophage.

What is the clinical significance
• Inhibition of P2X4 receptor activity with taspine has antiinflammatory activity.
University of East Anglia. Monocyte-derived macrophage were prepared as described previously (Layhadi et al., 2018). Briefly, peripheral venous blood was collected from healthy human volunteers through the National Health Service Blood and Transplant (Addenbrooke's Hospital, Cambridge University Hospital, Cambridge, U.K.). Blood was layered on top of Histopaque-1077 (Sigma-Aldrich, Haverhill, U.K.) and centrifuged at 1000Â g for 25 min. Buffy coat layers were collected and peripheral blood mononuclear cells were counted. Cells were adhered onto T75 flasks (Corning, UK) for 2 h and cultured in RPMI 1640 with 2 mM L-glutamine, 2.5% (v/v) heatinactivated autologous serum, 50 UÁml À1 penicillin and 50 μgÁml À1 streptomycin at 37 C for 6 days with 10 ngÁml À1 GM-CSF (PeproTech).

| Patch-clamp electrophysiology
Whole-cell patch-clamp electrophysiology was performed using the port-a-patch planar patch instrument (Nanion) fitted with an eightvalve gravity-fed perfusion panel. Cells in culture were dissociated using TrypLE (Invitrogen) and resuspended at 1 Â 10 6 cellsÁml À1 in extracellular recording solution containing (mM): NaCl, 140; KCl, 4; MgCl 2 , 1; CaCl 2 , 2; D-glucose, 5; HEPES, 10; pH 7.4. 3-5 MΩ NPC-1 chips were filled with internal recording solution contained (mM): NaCl, 10; CsF, 140; EGTA, 2; HEPES, 10; pH 7.4; 10 mM BAPTA replaced EGTA in internal recording solutions for experiments to test the dependency of intracellular Ca 2+ on drug effects. Seals of ≥1 GΩ were formed and allowed to stabilise for 2 min before breakthrough to the whole-cell configuration. Following whole-cell access, slow capacitance was measured using the auto C-slow function (Patchmaster software, HEKA) and cells were held at À80 mV for 2-3 min prior to the initial agonist application. ATP and other agonists were applied for 2 s at 4 min intervals. Taspine, LY294402 or vehicle controls were applied between ATP applications. Data were sampled at 1 KHz and low-pass filtered at 10 kHz (HEKA EPC 10 amplifier).

| CXCL5 secretion assay
Human CXCL5 chemokine was quantified in supernatants of ATPstimulated human primary macrophage cultures by ELISA (Biolegend, Cambridge, UK). Cells were cultured in 96-well plates and treated with taspine or PSB-12062 for 30 min prior to ATP challenge.

| In vitro phosphoinositide 3-kinase activity assay
The enzymatic activity of human recombinant PI3-kinase δ (PI3Kδ, Promega) was quantified using an ADP-Glo lipid kinase system (Promega) and 25 μl reaction volumes. Reactions contained 400 μgÁμl À1 recombinant PI3Kδ and 50 μgÁml À1 phosphoinositol-4,5-biphosphate: phosphatidylserine. Drugs or vehicle control were incubated with the reaction mix for 30 min prior to initiating the kinase reaction by the addition of ATP. All reactions were performed a 25 C. The amount of ADP produced through kinase activity was quantified by luciferase bioluminescence following its conversion to ATP using the ADP-Glo lipid kinase system. Luciferase bioluminescence was measured using a Flexstation III instrument (Molecular Devices) and 500 ms integration time. The relationship between substrate concentration and reaction velocity was fitted using by the

Michaelis-Menten equation
where v is the reaction velocity, V max is the maximum velocity, [S] is the concentration of substrate and K M is the Michaelis constant.
Lineweaver-Burk plots were generated from the linear first-order phases of Michaelis-Menten curves and used to determine K M and V max values.

| Cytotoxicity assay
LDH content of cell supernatants was quantified using colorimetric LDH assay kit (Abcam, Cambridge). 2.5 Â 10 3 cells per well were seeded into 96-well well plates and cultured overnight. Culture medium was replaced with HBPS and experiments performed at 37 C. LDH release was detected by absorbance at 450 nm using a Flexstation III instrument (Molecular Devices).

| Experimental design, data analysis and statistics
The data and statistical analysis comply with the recommendations of the British Journal of Pharmacology on experimental design and analysis in pharmacology (Curtis et al., 2018). Studies were designed to groups of equal size, using randomisation and blinded analysis where technically or practically feasible. All data analysis was performed using OriginPro software (OriginLab where k = Michaelis constant and n number of cooperative sites. The data sets were then compared with an F test. Pairwise comparison of EC 50 values generated in curve shift experiments was conducted using a paired sample t test. For intracellular Ca 2+ measurements, statistical comparison is made between the peak Ca 2+ response in 1321N1 cells and between Ca 2+ response AUC when using BV-2 cells and human primary macrophage. Data normalisation to control values has been applied to control for inter-experimental variation where applicable. For transformed datasets (e.g. data express as % control), test data points are normalised to control values within the same technical repeat.

| Materials
The initial chemical screen against ATP-evoked Ca 2+ responses in the 1321N1 human P2X4 stable cell line was performed using the National Cancer Institute (NIH, USA) Natural Product Set III. The purity of taspine in this library was >90%. Data in this study were generated using commercially available taspine (>97% purity; Santa Cruz Biotechnology). PSB-12062, BX-430, LY294002, ivermectin, ATP, α,β-methylene ATP, bzATP, staurosporine, thapsigargin and all basic salts were purchased from Sigma Aldrich.  Figure 1c) caused an approximate 80% reduction in response maxima, whilst the ATP EC 50 remained unchanged (Table 1). These effects were mirrored for mouse P2X4 (Table 1). To further explore the mechanism of action of taspine, we examined the rate of onset and reversibility of taspine inhibition in Ca 2+ assays. These experiments revealed that taspine had a slow onset of inhibition, reaching

| Taspine inhibits P2X4 receptor-dependent signalling in inflammatory cells
To investigate the effectiveness of taspine in inhibiting the activity of native P2X4 receptors, we selected inflammatory cell types where a role for P2X4 has been documented previously. Unlike 1321N1 cells that are void of native P2Y receptors (Schachter et al., 1996). We have previously observed that the contribution of P2X receptors to ATP-evoked intracellular Ca 2+ responses in inflammatory cells is often masked by a predominant P2Y response and the P2X4 positive allosteric modulator ivermectin is useful in unmasking P2X4-mediated responses in native cells (Layhadi et al., 2018;Layhadi & Fountain, 2019). In BV-2 mouse microglia cells, pharmacological antagonism of P2X4 has no significant effect on ATP-evoked Ca 2+ signalling (Dhuna et al., 2019), though a P2X4-dependent component can be revealed using ivermectin, a positive allosteric modulator (Priel & Silberberg, 2004). In agreement with previous studies (Dhuna et al., 2019), P2X4 antagonism with PSB12062 (10 μM, 30 min) had no effect on ATP-evoked Ca 2+ signalling in BV-2 cells. Taspine also had no effect. However, ivermectin potentiated ATP-evoked Ca 2+ signalling could be completely reversed by PSB12062 (Figure 3a,b) confirming that the effects of

| Effect of taspine on P2X4 receptor currents
Robust and reproducible ATP-evoked inward currents could be recorded from 1321N1 human P2X4 stable cells in whole-cell configuration ( Figure 4a). The internal and external salt solutions and agonist concentration employed minimise current rundown previously reported for human P2X4 (Fountain & North, 2006). Taspine application had no effect on the magnitude of currents (Figure 4b), though they could be reversibly blocked by BX430 (Figure 4c)

| The pharmacological properties of taspine are mimicked by LY294002
To further explore the mechanism of action, we sought to test the hypothesis that taspine suppresses P2X4 activity indirectly through inhibition of PI3-kinase. We hypothesised this could be a rational explanation for the following reasons. Firstly, the IC 50 for taspine at mouse and human P2X4 orthologues is similar (Figure 1) and it acts in an apparent orthosteric fashion (Figure 2), possibly suggestive of a common target other than the receptor itself. Secondly, the onset of taspine action is very slow and irreversible (Figure 2), and these are unlikely characteristics of a direct receptor effect and suggestive of depletion of a cellular component or pathway that positively regulates P2X4. Thirdly, phosphoinositides positively regulate rat P2X4 (Bernier et al., 2008). Finally, taspine inhibits VEGF signalling in endothelial cells limiting phosphorylation of Akt and Erk1/2 . As for taspine, the known PI3-kinase inhibitor LY294002 (Vlahos et al., 1994) inhibited ATP-evoked Ca 2+ responses in 1321N1 human P2X4 cells (Figure 5a) ( Table 1). The concentration of half-maximal inhibition was 25.8 ± 4.2 μM (n = 6) ( Figure 5b). In patch-clamp electrophysiological studies, LY294002 displayed the same properties as taspine (Figure 4). LY294002 had no effect on P2X4 currents under control conditions (Figure 5c) but was able to fully reverse  (Figure 5d). For completeness, we also tested the effect of LY294002 on the activity of other P2X receptors (Table 1). LY294002 showed no inhibitory effect on human P2X1, P2X3 or P2X7 but displayed activity at human P2X2, human P2X2/3 and mouse P2X4 in calcium assays (Table 1).

| DISCUSSION
In this study, we demonstrate that taspine inhibits the P2X4 receptor over other P2X subtypes tested. Rather than a direct antagonism of the receptor, our data suggest that the effect of taspine on P2X4 is indirect and via inhibition of PI3-kinase and depletion of phosphoinositides. The identification of taspine as a PI3-kinase inhibitor is novel and our work supports a role of PI3-kinase and phosphoinositides in regulating P2X4 activity (Bernier et al., 2008(Bernier et al., , 2012. The activity of taspine is physiologically relevant as it can suppress P2X4-dependent activity in human primary macrophage including the secretion of the CXCL5 (Layhadi et al., 2018), supporting a proinflammatory role of PI3-kinase signalling (Guiducci et al., 2008;Koorella et al., 2014). This and our previous work Average current density data (right panel; n = 5; *P < 0.05) (Layhadi et al., 2018;Layhadi & Fountain, 2019) suggest that direct antagonism of P2X4 or inhibition of pathways that enhance P2X4 activity are both routes to anti-inflammatory effects. In addition, P2X4-mediated ATP-induced chemotaxis in microglia is inhibited by PI3-kinase inhibitors (Ohsawa et al., 2007). Taspine has previously been reported to possess anti-inflammatory properties (Vaisberg et al., 1989), though this is likely due to activity at multiple targets, possibly including P2X4.
Membrane permeability is a required property of taspine if it is to access and inhibit PI3-kinase. Taspine (calculated logP 2.7) and LY294002 (calculated logP 3.6) are both hydrophobic molecules and that are predicted to be membrane permeable. In addition, pharmacokinetic and tissue distribution of taspine after intravenous administration in mice supports cell penetrance and tissue accumulation (Lu et al., 2008). Other reported intracellular targets of taspine that support its cell penetrance include topoisomerases (Fayad et al., 2009). The slow onset and irreversibility of taspine action at P2X4 are consistent with phosphoinositide depletion . The PI3-kinase pathway converts cellular PIP 2 to PIP 3 and can be activated by tyrosine kinase receptors and GPCRs. Its inhibition depletes cellular PIP 3 and downstream phosphoinositides that are dependent on PIP 3 for their synthesis (Jean & Kiger, 2014).
Previous studies using wortmannin, a non-specific PI3-kinase inhibitor, have demonstrated that depletion of phosphoinositides reduces recombinant and native rodent P2X4 receptor activity, suggesting that the C-terminus of P2X4 harbours a binding domain for several biologically active phosphoinositides (Bernier et al., 2008). In contrast to this study, some previous studies investigating P2X3 and P2X2/3 (Mo et al., 2009) and P2X7  have observed either direct effects of phosphoinositide application or PI3-kinase inhibitors.
In this study, we have focused on human P2X receptor subtypes and some differences observed here may be due to differences in the orthologue of P2X receptors investigated and the expression systems  -Fish et al., 2016). The effects of BAPTA may therefore suggest taspine inhibits the activity of Ca 2+ -dependent PI3-kinase though whether Ca 2+ entry via P2X4 directly regulates PI3-kinase remains to be determined.
In agreement with our own data, Bernier et al. (2008) demonstrate that wortmannin alone has no effect on P2X4 currents, though the effectiveness of ivermectin as a positive allosteric modulator of P2X4 is reduced by wortmannin. This is an interesting observation and suggests ivermectin activity during electrophysiological recordings is inhibited following PI3-kinase inhibition with either wortmannin (Bernier et al., 2008), LY294002 or taspine. It is unlikely that these three chemically distinct molecules directly prohibit ivermectin binding to the P2X4 receptor (Gao et al., 2015;Latapiat et al., 2017). Ivermectin has direct effects on P2X4, reducing channel desensitisation and stabilising the channel in the open conformation (Priel & Silberberg, 2004). However, several studies have also suggested that ivermectin facilitates P2X4 activity by increasing the number of channels at the cell surface (Stokes, 2013;Toulme et al., 2006). As PI3-kinase regulates cellular trafficking (Lindmo & Stenmark, 2006), LY24002 and taspine may limit the trafficking activity of ivermectin and therefore reverse its effect on P2X4. Differences in trafficking pathways available for P2X4 (Royle et al., 2005) in intact cells (Ca 2+ imaging experiments) versus dialysed cells (wholecell patch clamp experiments) may also account for the difference effects of taspine and LY294002 observed in this study. Our data shows that the inhibitory action of taspine is restricted to P2X4, though the known PI3-kinase inhibitor LY294002 did cause significant inhibition of the human P2X2 and P2X2/3 response. The effect of LY294002 is consistent with the findings of Fujiwara and Kubo (2006) who observed acceleration of rat P2X2 desensitisation kinetics in the presence of LY294002. The effects of taspine and LY294002 are consistent for all P2X receptors tested except for human P2X2 and P2X2/3 receptors. This discrepancy is difficult to reconcile as both compounds inhibit PI3-kinase. Possible reasons include the observed less efficacious action of taspine on PI3-kinase compared with LY294002, or an off-target direct effect of LY294002 on the P2X2 receptor as observed for other kinase inhibitors and P2X receptors (Dayel et al., 2019).
Using human recombinant PI3-kinase, we demonstrated that taspine, like LY294002, directly inhibits the enzyme in a competitive manner. LY294002 is a known competitive PI3-kinase inhibitor (Vlahos et al., 1994), with recent structural information suggesting this mechanism is achieved through partial overlap of the binding site for LY294002 and the adenine group of ATP in the ATP/enzyme complex (Walker et al., 2000). Likewise, taspine must also exert its effect by competition with ATP binding in the enzyme, though the site is currently elusive. However, taspine shares a fused heterocyclic structure with wortmannin, another known competitive inhibitor of PI3-kinase (Walker et al., 2000).
In summary, taspine inhibits the activity of recombinant P2X4 and native P2X4 during physiologically relevant processes. We suggest that the apparent non-competitive inhibition of P2X4 by taspine is observed because of inhibiting a positive modulatory role of PI3-kinase, rather than any direct binding of taspine to P2X4 ( Figure S3). We suggest the mechanism of taspine effect is via competitive inhibition of PI3-kinase. These data support a role of PI3-kinase in regulating P2X4 receptor function and the antiinflammatory action of taspine.