Mechanisms Involved in the Nociception Triggered by the Venom of the Armed Spider Phoneutria nigriventer

Background The frequency of accidental spider bites in Brazil is growing, and poisoning due to bites from the spider genus Phoneutria nigriventer is the second most frequent source of such accidents. Intense local pain is the major symptom reported after bites of P. nigriventer, although the mechanisms involved are still poorly understood. Therefore, the aim of this study was to identify the mechanisms involved in nociception triggered by the venom of Phoneutria nigriventer (PNV). Methodology/Principal Findings Twenty microliters of PNV or PBS was injected into the mouse paw (intraplantar, i.pl.). The time spent licking the injected paw was considered indicative of the level of nociception. I.pl. injection of PNV produced spontaneous nociception, which was reduced by arachnid antivenin (ArAv), local anaesthetics, opioids, acetaminophen and dipyrone, but not indomethacin. Boiling or dialysing the venom reduced the nociception induced by the venom. PNV-induced nociception is not dependent on glutamate or histamine receptors or on mast cell degranulation, but it is mediated by the stimulation of sensory fibres that contain serotonin 4 (5-HT4) and vanilloid receptors (TRPV1). We detected a kallikrein-like kinin-generating enzyme activity in tissue treated with PNV, which also contributes to nociception. Inhibition of enzymatic activity or administration of a receptor antagonist for kinin B2 was able to inhibit the nociception induced by PNV. PNV nociception was also reduced by the blockade of tetrodotoxin-sensitive Na+ channels, acid-sensitive ion channels (ASIC) and TRPV1 receptors. Conclusion/Significance Results suggest that both low- and high-molecular-weight toxins of PNV produce spontaneous nociception through direct or indirect action of kinin B2, TRPV1, 5-HT4 or ASIC receptors and voltage-dependent sodium channels present in sensory neurons but not in mast cells. Understanding the mechanisms involved in nociception caused by PNV are of interest not only for better treating poisoning by P. nigriventer but also appreciating the diversity of targets triggered by PNV toxins.


Introduction
Spiders of the Phoneutria genus, popularly known as the wandering or banana spider, are found in Central and South America, where relevant envenomation cases in humans have been reported [1]. There are four main Phoneutria species, P. fera, P. reidyi, P. keyserlingi and P. nigriventer [2]. The incidence of accidents due to spiders in Brazil has grown in recent years (from 1.9 in 2000 to 13 Thus, Phoneutria nigriventer envenomation is an important publichealth problem in Brazil, which is aggravated by the fact that P. nigriventer is very aggressive and is frequently found in human dwellings as a result of the accumulation of organic waste, which attracts insects, the natural prey of these spiders [1]. Envenomation by Phoneutria nigriventer venom (PNV) in humans may produce systemic symptoms, such as spastic paralysis or tremors, and local symptoms, such as pain and oedema. Intense local pain is the main clinical manifestation following P. nigriventer envenomation, appearing in 96% of envenomed patients [3,1]. The development of local oedema is less frequent than pain and appears in 61% of bitten people [4]. It has been shown that the oedematous effect induced by PNV in rats involves the generation of kinins and the stimulation of tackykinin receptors. This effect also depends on the histamine, serotonin and polypeptides contained in the venom [5,6,7,8]. Because pain may occur after PNV envenomation even in the absence of oedema, the mechanisms involved in these responses could be distinct. Moreover, there are no validated therapies for treating PNV-induced pain, though local anaesthetics, opioids, non-steroidal anti-inflammatory drugs and arachnid antivenin serum have been used empirically [8,2].
PNV contains both low-and high-molecular-weight substances that could contribute to the induction of pain [9,3]. However, the PNV components involved in the nociceptive response as well as the molecular targets stimulated by venom components have not been identified. Thus, the aim of this study was to investigate the mechanisms involved in PNV-induced nociception in mice.

Ethical statements
All experiments were conducted in accordance with the current guidelines for the care of laboratory animals and the ethical guidelines for investigations of experimental pain in conscious animals [10]. All of the protocols employed were approved by the Local Ethics Committee -Comissão de É tica no Uso de Animais (process number 23081.003193/200940). The number of animals and the nociceptive stimulus used were the minimum necessary to demonstrate the consistent effects of drug treatments.

Animals
Male Swiss mice weighing 30-35 g were maintained at 2262uC with free access to water and food under a 12:12 h light:dark cycle. Animals were acclimatized in the laboratory for at least 2 h before testing and were used only once throughout the experiments. Behavioural testing was performed by an experienced observer blinded with respect to drug administration.
The stock solutions of most of the drugs were prepared in phosphate-buffered saline (PBS, 137 mM NaCl and 10 mM phosphate buffer, pH 7.4, Sigma, USA) in siliconised plastic tubes, maintained at 220uC, and diluted to the desired concentration just before use. The SB366791 and resiniferatoxin stock solutions were prepared in absolute ethanol (90%) and Tween 80 (10%). The indomethacin stock solution was prepared in 5% ethanol plus 5% Tween 80. The final concentration of ethanol and Tween 80 did not exceed 0.5% in the administered drug and did not cause any detectable effects (results not shown). The stock tetrodotoxin solution was prepared in acetate buffer (100 mM, pH 4.8), and the pH of the solution was adjusted to 7.4 with Tris buffer (100 mM, pH 9.6). The pH of the glutamate solution was adjusted to 7.4 with 5 M NaOH (to a final sodium concentration of approximately 157 nM).

Algogen-induced nociceptive and oedematogenic responses
The procedure used was similar to that described previously [11,12,13,14]. The animals were placed individually in chambers (transparent glass cylinders 20 cm in diameter) and were allowed to acclimate to the tube for 20 min before treatment. A volume of 20 ml of PNV (0.3-10 mg/paw) was injected subcutaneously under the surface of the right hind paw (intraplantar, i.pl.). A separate group of animals received an i.pl. injection of the appropriate vehicle (PBS). After treatment, the mice were observed individually for 360 min.
The amount of time spent licking the injected paw was timed with a chronometer for 10 minutes and was defined as the time of nociception.
It has been reported that there are significant advantages to using a combination of several behaviours rather than a single index of pain, in terms of sensitivity and specificity [15]. Therefore, other nociceptive behaviours besides licking the affected limb were observed and categorised based on the position of the injected hind paw, and a score of 0, 1, 2 or 3 was assigned [15]. The behavioural rating criteria were as follows: (0) no painnormal weight bearing on the injected paw; (1) favouring -resting the injected paw lightly on the floor or limping; (2) liftingelevation of the injected paw; and (3) licking -licking or biting the injected paw. Because the mouse moves so quickly, time sampling was used to score other pain-related behaviours. Mice were observed for 5 minutes immediately after PNV or vehicle injection and also for 5 minutes at 15, 30, 60, 120, 240 and 360 min after injection.
The variation in the thickness of the paw was assessed as an oedema index [13]. The thickness was measured with a calliper immediately prior to the injection of PNV and immediately after during nociception measurement. Paw oedema was expressed as

Author Summary
Spiders of the Phoneutria genus live in Central and South America, where relevant envenomation cases have been reported in humans. The incidence of bite by spiders in Brazil has increased in recent years, with Phoneutria nigriventer being the second most important cause of such accidents (approximately 4,000 cases of envenomation in 2011). Pain is the primary local symptom of inoculation with Phoneutria nigriventer venom (PNV), but the mechanisms involved in pain induced by PNV are poorly understood. It is important to find effective treatments to alleviate this pain. This study examined the mechanisms involved in pain caused by PNV in a mouse model as well as the sensitivity of PNV-induced pain to clinically used analgesics. The results show that both the low-and high-molecular-weight components of PNV produce spontaneous nociception action via kinin B 2 , TRPV1, 5-HT 4 or ASIC receptors and the voltage-gated Na + channels present in sensory fibres. Moreover, PNVtriggered nociception could be alleviated by arachnid antivenin, local anaesthetics, opioids and atypical, but not typical, non-steroidal anti-inflammatory drugs. The elucidation of the mechanisms responsible for the nociception induced by PNV is of interest to not only better treat envenomation by P. nigriventer but also understand the diversity of targets triggered by PNV toxins.
the difference (D) in paw thickness (mm) after PNV or vehicle injection relative to the pre-injection value for each animal.

Sensitivity of PNV-induced nociception to clinically used analgesics
To assess the effects of clinically used analgesics on the amount of time spent licking the PNV-injected paw, venom (3 mg/paw, i.pl.) was injected into a mouse paw previously treated with morphine (10 mg/kg, i.p., 30 min prior), dipyrone (500 mg/kg, i.p., 30 min prior), acetaminophen (400 mg/kg, i.p., 30 min prior) or indomethacin (30 mg/kg, i.p., 30 min prior). PNV (3 mg/paw, i.pl.) was also co-administered with lidocaine (0.4 mmol/paw, i.pl.). The control animals received the appropriate vehicle. The dosage choice for each drug was based on previous data described in the literature [12,16,17] or on preliminary experiments conducted in our laboratory (data not shown).
To determine whether the nociceptive behaviour induced by PNV injection is neutralised by arachnid antivenin (ArAv), PNV was co-administered with ArAv (1:30, dissolved in PBS). The control animals received a similar volume of boiled ArAv (1:30, dissolved in PBS and boiled at 90uC for 15 min to inactivate the serum) or PBS (20 ml/paw). The dosage choice for ArAv was based on preliminary experiments conducted in our laboratory.
Boiling and dialysis of Phoneutria nigriventer venom PNV (1 ml of 150 mg/ml diluted in PBS) was boiled for 15 min at 90uC [18]. The procedure used to dialyse PNV was similar to that described previously [5]. The PNV (2 ml of a 1 mg/ml PBS solution) was dialysed (MW cut off of the dialysis membrane was 12-14 kDa, Sigma) for 24 h at 4-6uC against 2 L of PBS. The dialysing solution was changed 7 times, at 2, 4, 6, 8, 10, 12 and 22 h.

Estimation of histamine, serotonin and glutamate levels in the Phoneutria nigriventer venom
The histamine content was evaluated as previously described [19]. First, 300 ml of NaOH (1 M) was added to 1 ml of dialysed PNV or non-dialysed PNV samples and incubated with 40 ml of The serotonin content was evaluated as previously described [20]. Dialysed PNV or non-dialysed PNV samples (0.5 ml) were incubated with 0.5 ml of ninhydrin (0.24%) and 0.5 ml of phosphate buffer (100 mM, pH 7.4). The reaction was incubated at 100uC for 10 min and allowed to rest at room temperature in the dark for 16 h. The fluorescence was read at 380 nm (excitation) and 500 nm (emission) in a spectrofluorophotometer.
The glutamate measurement was performed enzymatically following an increase in fluorescence due to the production of NADPH in the presence of glutamate dehydrogenase and NADP+ [21]. To start the assay, NADP+ (1.0 mM) and glutamate dehydrogenase (50 U) were added to the dialysed or non-dialysed PNV samples 10 min after the emitted fluorescence was measured using a spectrofluorophotometer. The excitation wavelength was set at 360 nm, and the emission wavelength was set to 450 nm.

Measurement of kallikrein-like activity in Phoneutria nigriventer venom
The method used for the measurement of kallikrein activity was similar to a previously described method [22], with some modifications. The activity of a kallikrein-like enzyme was measured by the cleavage of a selective peptide-nitroanilide substrate, D-Val-Leu-Arg-paranitroanilide (D-Val-Leu-Arg-pNA) (Sigma, St. Louis, USA), which was dissolved in ultrapure water to a concentration of 1.5 mM and stored at 4uC until use. To determine the kallikrein activity, 25 ml of PNV (500, 150 or 50 mg/ml) or boiled PNV was added to 50 ml of assay buffer (0.2 M Tris-HCl, pH 8.2, containing 0.01 M EDTA) and 25 ml of substrate (D-Val-Leu-Arg-pNA, 0.375 mM) in 96-well microplates. The incubation mixtures were maintained at 37uC, and the reaction was followed by measuring the absorbance at 405 nm (measured in a Plate Reader, Bioteck, USA) from 1 to 16 min. For each experiment, we determined a standard curve using a pure synthetic form of the product generated by the reaction (p-nitroaniline, 0.003-3 mM, Merck, Darmstadt, Germany). The released p-nitroaniline was measured colorimetrically at 420 nm. The protein content of PNV was determined by the Bradford method [23] using a standard curve with known concentrations of bovine serum albumin within the absorbance range. The results of the kallikrein enzyme activity assay are expressed as nmol of the formed product (p-nitroaniline) over time (in minutes) and as the amount of protein (expressed in mg of protein) in each PNV sample [24]. To kinetically characterise the kallikrein-like activity, we incubated PNV (150 mg/ml) with different concentrations of the substrate (D-Val-Leu-Arg-pNA, 0.09-3.00 mM), and the reaction was followed by measuring the absorbance at 405 nm from 1 to 40 min. We also determined the sensitivity of the PNV enzyme to the tissue kallikrein inhibitor aprotinin and the plasma kallikrein inhibitor soybean trypsin inhibitor (SBTI). For this experiment, 25 ml of PNV (500 mg/ml) and 25 ml of the substrate (D-Val-Leu-Arg-pNA, 0.375 mM) were incubated in the presence or absence of aprotinin (10 mg/ml) or SBTI (3 mg/ml) [25]. A final volume of 100 ml was achieved by adding buffer where necessary. The reaction was followed by measuring the absorbance at 405 nm from 1 to 16 min.

Kinin production by PNV
Phoneutria nigriventer venom (3 mg) was incubated with high-or low-molecular-weight human kininogen (200 nM) in 50 mM Tris buffer (pH 7.41, 0.1 M NaCl) in a final volume of 100 mL for 15 min at 37uC, as reported by [26] and [27], with minor modifications. The kinin was extracted in ice-cold ethanol (four times the final volume) and centrifuged for 60 min at 16,0006g at 4uC. The supernatants containing free kinin were collected. The solutions were freeze-dried and resuspended in 100 ml of enzyme immunoassay buffer. A kinin enzyme immunoassay was performed using a high-sensitivity kit for kinin (Bachem) according to the manufacturer's protocol. Briefly, 50 ml of the standard solutions or test samples was added to a 96-well plate. Then, 25 ml of each primary antisera and biotinylated peptide solution were added, and the plates were incubated for 2 h at room temperature with mild agitation. The plates were then washed five times, and 100 ml of a diluted streptavidin-conjugated horseradish peroxidase solution was added to each well. After a 60-min incubation at room temperature, the immunoplates were washed five times, and 100 ml of a 3,39,5,59-tetramethyl benzidine dihydrochloride (TMB) solution was added to each well. After an additional 20-min incubation at room temperature, the reaction was stopped with 100 ml of 2 N HCl. The absorbance was read at 60 min after stopping the reaction at 490 nm. The blank control consisted of 100 ml of TMB solution and 100 ml of 2 N HCl.
To further explore the role of TRPV1-positive sensory fibres in the nociception induced by PNV, animals were subjected to a systemic desensitisation protocol as previously described [33]. The animals were anesthetised with isoflurane and received systemic administration of resiniferatoxin (50 mg/kg) or the vehicle alone (0.5% ethanol, 0.5% Tween-80, PBS). After 7 days, the animals received a subcutaneous injection of PNV (3 mg/paw, i.pl.), capsaicin (1 nmol/paw, used as a positive control) or vehicle (0.5% ethanol, 0.5% Tween-80 and PBS; 20 ml/paw), and the nociceptive behaviour was evaluated.

[ 3 H]Resiniferatoxin binding assay
Binding assays were conducted as described previously [34]. To obtain membranes for the binding studies, the spinal cords of mice were removed and disrupted with a tissue homogeniser in ice-cold buffer A (pH 7.4), which contained 5 mM KCl, 5.8 mM NaCl, 2 mM MgCl 2 , 0.75 mM CaCl 2 , 12 mM glucose, 137 mM sucrose and 10 mM HEPES. The homogenate was first centrifuged for 10 min, 1,0006 g at 4uC. The low-speed pellets were discarded, and the supernatants were further centrifuged for 30 min, 35,0006 g at 4uC. The resulting high-speed pellets were resuspended in buffer A and stored at 220uC until assayed. The binding assays were conducted in duplicate with a final volume of 500 ml, which contained buffer A supplemented with 0.25 mg/ml bovine serum albumin, membranes (100 g/protein) and 50 pM [ 3 H]-resiniferatoxin in the presence or absence of PNV (1.5-150 mg/ml). To measure nonspecific binding, 100 nM of nonradioactive RTX was included. The assay mixtures were set up on ice, and the binding reaction was initiated by transferring the assay tubes to a 37uC water bath. After a 60-min incubation period, the mixtures were cooled on ice to terminate the binding reaction. Then, 100 mg of bovine a1-acid glycoprotein (Sigma, St. Louis, USA) was added to each tube to reduce nonspecific binding. Finally, the bound and free membranes of [ 3 H]-resiniferatoxin were separated by centrifugation for 15 min, 20,0006 g at 4uC. The pellet was quantified using scintillation counting. The specific binding was calculated as the difference between the total and nonspecific binding.

HEK293 cell culture and transfection
HEK293 cells were grown in 89% DMEM, 10% FBS, and 1% of an aqueous solution containing penicillin (5000 units/mL) and streptomycin (5000 mg/mL) (Gibco). The cells were resuspended and passaged every 5-6 days. The plasmid containing rat-TRPV1 was kindly provided by Dr. David Julius (UCSF/CA/USA). After growing to 80-90% confluence, the cells were transfected with rTRPV1using Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. Calcium imaging experiments were performed 12-24 h after the transfection.

Calcium fluorescence imaging
The experiments were performed at room temperature (20-25uC), essentially as described previously [35]. The cells on cover slips were treated with 3 mM Fluo 4-AM (Molecular Probes) for 40 min in HEPES-buffered salt solution (HBSS) containing (in mM) 124 NaCl, 4 KCl, 1 CaCl2, 1.2 MgCl2, 10 glucose, and 25 HEPES, at pH 7.4. The cover slips were washed in dye-free HBSS, transferred to a superfusion chamber system (Bioptechs) and placed on the stage of a microscope. The HBSS solution (alone or with the treatments) was continuously perfused (0.6 ml/min) via a peristaltic pump (Bioptechs) during image acquisition. Calcium imaging was performed with a Leica SP5 laser scanning confocal system using LAS software with a 106 objective lens. Fluo 4-AM was imaged by excitation with a 488-nm argon laser line, and the emitted light was collected at a 510-570 nm emission band. The cells were stimulated for 20 s with HBSS containing capsaicin or PNV at different concentrations. Quantitative measurements were made by re-analysing the stored image sequences using LAS Application software. Regions of interest (ROI) were identified within the soma of individual cells of at least 3 independent plates. Changes in fluorescence (F) were normalised to the initial fluorescence (F0) for each cell and expressed as (F/F0)6100 (% of baseline). For dose-response curves, the average amplitude of the [Ca 2+ ] changes represents the difference between the baseline and the transient peak of fluorescence in response to capsaicin or PNV. A control group, prepared by Lipofectamine transfection without plasmid, is indicated in the text as non-transfected cells.

Statistical analysis
The results are presented as the mean6S.E.M. except for the ED 50 values (i.e., the dose of agonist necessary to produce 50% of the pain response relative to the maximum effect), which are reported as geometric means accompanied by their respective 95% confidence limits, and the spontaneous nociception scores, which were reported as medians and interquartile ranges. The ED 50 values were calculated using linear regression for individual experiments in the GraphPad Prism software. Spontaneous nociception scores were analysed using the Mann-Whitney U test. All other data were analysed using the Student's t-test, one-way ANOVA followed by Dunnett's test or two-way ANOVA followed by Bonferroni's test, as appropriate. P values of less than 0.05 were considered to be significant.

Phoneutria nigriventer venom induced ongoing nociception and oedema
The subcutaneous administration of PNV (0.3-10 mg/paw) in mice produced dose-dependent, continuous nociception, verified as an increase in the amount of time spent licking the injected paw ( Figure 1A-B). This behaviour occurred quickly, peaked from 1 to 5 min and declined 6 min after the PNV injection ( Figure 1A). The calculated ED 50 value of PNV was 1.6 (0.8-2.4) mg/paw, and the maximum licking response time was 147.5617.0 s.
In addition to licking, other nociceptive behaviours (favouring and lifting) were observed and scored according to the position of the injected hind paw. When a combination of behaviours was assessed, PNV (3 mg/paw) was capable of producing pain-related behaviour up to 4 hours after the injection ( Figure 1C). The means of sum of the behaviour scores from 0 to 240 min after PNV injection were 2 and 13 for vehicle and PNV, respectively ( Figure 1D). PNV injection also produced a marked paw oedema that was detected as early as 5 min after the injection and was reduced, but still significant, 4-6 hours after injection ( Figure 1E).
To mimic a clinical setting, PNV (3 mg/paw) was injected, nociception scores were observed up to 30 min, and a clinically used analgesic was administered followed by observation of nociception again for 60 to 240 minutes after PNV injection. Similar to the results of pre-treatment, post-treatment with dipyrone and acetaminophen were capable of reducing nociception scores, though less efficaciously than pre-treatment (55613% and 3969% inhibition, respectively) ( Figure 2G, I). In contrast to the pre-treatment results, post-treatment with neither morphine, lidocaine nor ArAv was unable to alter the nociception scores ( Figure 2H, K and M), though indomethacin inhibited nociception (inhibition of 34610%) ( Figure 2J).
To avoid unnecessary long-lasting animal discomfort and focus on the mechanisms responsible for the development, but not the maintenance, of nociception trigged by PNV, the dose of 3 mg/ paw and the time spent licking the injected paw were chosen to study a subset of mechanisms involved in PNV-induced nociception.

Effect of boiled or dialysed venom in PNV-induced nociception
To determine whether the thermolabile and low-molecularweight substances in PNV are related to venom-induced nociception, animals received a s.c. injection of boiled or dialysed PNV (3 mg/paw). The spontaneous nociception trigged by native PNV was significantly reduced by boiling (55614%) or dialysis (55618%) (Figure 3A and 3B). These results suggest that PNVinduced nociception was induced by the combined action of a thermolabile substance (possibly with a high molecular weight) as well as low-molecular-weight (and thermal-resistant) substances.
Role of glutamate, histamine, serotonin and mast cells in the nociceptive effect triggered by PNV We were able to identify the pro-inflammatory substances histamine, serotonin and glutamate in PNV ( Table 1). The concentrations of these substances were largely reduced by venom dialysis (Table 1). We next verified the effect of serotonin, histamine and glutamate receptor antagonists on PNV-induced nociception.
Regarding the role of glutamate, we observed that the coadministration of either the AMPA/kainate receptor antagonist DNQX or the selective NMDA receptor antagonist MK-801 (both at 1 nmol/paw, Figure 4A and 4B) was not able to alter the nociceptive effect trigged by PNV (3 mg/paw). However, the coadministration of DNQX (1 nmol/paw) or MK-801 (1 nmol/paw) reduced the nociception caused by glutamate (10 mmol/paw), with inhibition values of 5566 and 47612%, respectively (Table S1).
The histamine H 1 receptor antagonist promethazine (3 nmol/paw, Figure 4F) did not alter PNV-induced nociception (3 mg/paw). In addition, we observed that the i.pl. injection of a high dose of histamine (100 nmol/paw) did not produce nociception in mice, but did induce a paw oedema that was reduced by co-administration with promethazine (3 nmol/paw, data not shown).
Moreover, the degranulation of mast cells, which can release histamine and serotonin, was not necessary to induce nociception, given that a previous degranulation of the mast cells did not reduce PNV-induced nociception (3 mg/paw, Figure 4G) but was able to prevent compound 48/80-induced nociception (Table S1).

Role of tissue kallikrein-like activity in the nociceptive effect trigged by PNV
We detected the tissue kallikrein-like activity of PNV in vitro by measuring the hydrolysis of the selective tissue kallikrein substrate D-Val-Leu-Arg p-nitroaniline. We found that the incubation of PNV with the substrate produced a tissue kallikrein-like activity that increased with time and was linear from 1 to 16 min after incubation ( Figure 5A). Different concentrations of PNV (50-500 mg/ml) produced a kallikrein-like activity that was concentration dependent ( Figure 5B). Furthermore, the substrate kinetic curve (0.09-3 mM) demonstrated that the tissue kallikrein-like enzyme in PNV (150 mg/ml) had a Michaelis-Menten constant (K M ) value of 1.4760.09 mM and a maximal velocity (V max ) of 16.1360.62 nmol/min/mg protein ( Figure 5C). Moreover, the kallikrein-activity of PNV (500 mg/ml) was inhibited by boiling PNV or by the tissue kallikrein inhibitor aprotinin (10 mg/ml) but not by the plasma kallikrein inhibitor SBTI (3 mg/ml) ( Figure 5E).
Because PNV demonstrated kallikrein-like activity, we next determined the ability of the venom to generate kinins from natural kininogen substrates. The results presented in Figure 5D show that the incubation of PNV (150 mg/ml) with low-but not high-molecular-weight kininogens resulted in kinin production.
We next investigated whether the tissue kallikrein-like activity observed in vitro was related to the nociception produced by the venom in vivo. I.pl. co-administration with aprotinin (100 mg/paw, Figure 6A), but not the soybean trypsin inhibitor (3 mg/paw, Figure 6B), reduced the PNV-induced nociception with an inhibition of 70612% compared to the control group. These results indicate that the Phoneutria nigriventer venom has a tissue kallikrein-like activity that is important to its nociceptive action.
Tissue kallikrein generates kinins that may stimulate the B 2 kinin receptor to induce pain [36]. Indeed, the i.pl. coadministration of the selective B 2 receptor antagonist HOE 140 (3 nmol/paw, Figure 6D) with PNV (3 mg/paw) reduced the nociceptive effect induced by PNV with an inhibition of 43614%. These data demonstrate that the tissue kinins generated by PNV play a critical role in the nociceptive effect of PNV.

Additional investigation of the mechanisms involved in PNV-induced nociception
The serotonin or kinin action on the metabotropic 5-HT 4 or B 2 receptors that induces sensory fibre depolarisation and consequently triggers nociception depends on the stimulation of several enzymes and ion channels. Thus, we next investigated some of these targets in the nociceptive behaviour produced by PNV.
To test the possible involvement of the TRPV1 receptor, ASIC receptor, cyclooxygenase and sodium channels, mice received i.pl. co-administration of the ASIC blocker amiloride (100 nmol/paw, Figure 7A), the cyclooxygenase inhibitor indomethacin (30 mmol/ paw, Figure 7B), the sodium channel blocker tetrodotoxin (20 pmol/paw, Figure 7C) or the selective TRPV1 antagonist SB366791 (1 nmol/paw) with PNV (3 mg/paw, Figure 7D). The results demonstrate that amiloride, tetrodotoxin and SB366791, but not indomethacin, reduced PNV-induced nociception with inhibitions of 4365%, 9464% and 4566%, respectively. To confirm its general anti-nociceptive activity at the tested dose, indomethacin (30 mmol/paw) was co-administered with arachidonic acid (100 nmol/paw, Table S1), which reduced nociception by 71619%. Tetrodotoxin was not capable of reducing PGE2-induced nociception, demonstrating that its effect is not unspecific (Table S1). These results show that the ASIC receptor, the tetrodotoxin-sensitive sodium channels and the TRPV1 receptor are some of the mechanisms underlying the nociceptive action of PNV.
It has been recently demonstrated that spider venoms may contain toxins that directly bind to and stimulate the TRPV1 receptor [37,38,39]. It appears that PNV does not possess such toxins because high concentrations of PNV (1.5-150 mg/ml, Figure 7E) were not able to alter the specific binding of the TRPV1 ligand [ 3 H]-resiniferatoxin in vitro. Under the same conditions, capsaicin (10 mM, Figure 7E) inhibited the specific binding of [ 3 H]-resiniferatoxin to membranes (inhibition of 56.060.3% compared to the control group).
We further investigated a possible role of PNV on functional TRPV1 channels. Calcium levels in HEK293 cells transiently expressing TRPV1 were monitored by using the fluorescent calcium probe Fluo 4-AM. Functional expression of TRPV1 was detected by brief Ca 2+ transient signals observed following short pulses (20 s) of capsaicin at different doses ( Figure 8A-B) with an EC 50 of 46 (7-288) nM. Similarly, PNV (20 s pulses) also induced dose-dependent Ca 2+ transient signalling in HEK293-TRPV1 cells (Fig. 8C-D), with an EC 50 value of 0.22 (0.03-1.81 mg/mL). No transient Ca 2+ signal was observed in non-transfected cells after capsaicin or PNV application ( Fig. 8A and C, inserts). Following the first transient Ca 2+ signal, both capsaicin (Fig. 8E) and PNV (Fig. 8G) could elicit a new signal in HEK293-TRPV1 with similar amplitudes to the respective initial stimulus. However, incubation (300 s) with the selective TRPV1 antagonist SB366791 (10 mM) totally inhibited the second response elicited by capsaicin (Fig. 8F) or PNV (Fig. 8H). Taken together, these data show that PNV functionally interacts with TRPV1 to elicit transient calcium signalling in cells expressing these channels.
To investigate the role of TRPV1-expressing sensory fibres in the PNV-induced nociception, animals were subjected to a resiniferatoxin ablation protocol. This protocol largely inhibited the nociception  (Table S1). The sensory fibres that express TRPV1 also produce and release tachykinin after their depolarisation, an effect that could contribute to the development of nociception [40]. Interestingly, we observed that selective antagonists of the tackykinin NK 1 (SR143033, 0.2 nmol/paw) or NK 2 (SR48968, 20 nmol/paw) receptors were able to partially reduce PNV-trigged nociception with inhibitions of 57610% and 4068%, respectively ( Figure 9B). Curiously, the combination of the NK 1 and NK 2 receptor antagonists was only as efficacious as the individual antagonists in reducing PNV-induced nociception (time of nociception of 12464 s and 75611 s for vehicle and SR143033 plus SR48968 groups, respectively; inhibition 4069%).
These data suggest that PNV stimulates TRPV1 receptors and releases tachykinin in sensory fibres to induce a nociceptive effect.

Discussion
The Phoneutria nigriventer sting is the most important form of envenomation by spiders in Brazil. Intense local pain is the main clinical manifestation of this spider sting, and the mechanisms involved in pain manifestation are not well described in the literature. In the present study, we studied the mechanisms through which PNV produces spontaneous nociception in mice using different pharmacological tools through blinded behavioural assessment with respect to drug administration. Our results show that an intraplantar injection of PNV venom in mice produced extended nociception that was mediated by the activation of sensory fibres through B 2 , TRPV1, 5-HT 4 or ASIC receptor activation and the stimulation of tetrodotoxin-sensitive voltagegated Na + channels and kallikrein-like activity.
In our study, the PNV injection induced nociception that developed immediately and lasted for 4 hours. Our studies show that the dose that produced nociception in mice (3 mg/paw or 85.7 mg/kg) was close to the dose reported by [41] as the amount of P. nigriventer venom release (2.6-61.7 mg/kg per spider in an adult human being). The inoculation of P. nigriventer venom in humans and dogs promotes a local, spontaneous, unbearable pain that develops quickly after the sting and can last for many hours [1,18,42]. At first, we only recorded the amount of time spent licking the injected paw as the single index of pain because it is the most frequent parameter used to measure on-going pain in mouse models, as in the formalin, capsaicin and bradykinin tests [11,33,14]. Using this parameter, we detected a short nociceptive effect induced by PNV, as opposed to the long-lasting pain observed in humans and dogs. It is not surprising that the task of recognizing pain in rodents is difficult, given their ecological niche as prey animals [43]. Therefore, it has been reported that there are significant advantages to using a combination of several behaviours rather than a single index of pain in terms of sensitivity, specificity and duration [14,15]. When other pain-related behaviours in addition to paw licking (such as paw lifting and favouring) were recorded, we were able to detect long-lasting (up to 4 hours) ongoing nociception. Furthermore, we observed that PNV injection produced paw oedema with a time course similar to the pain-related behaviours. Our data are in accordance with a previous study in which the peripheral injection of PNV in rats evoked a long-lasting (up to 6 h) mechanical hyperalgesia and paw oedema [44]. Thus, the long-lasting nociception seems to be related to the inflammatory process induced by PNV. Moreover, the combination of several behaviours rather than a single index in mice may be useful to measure pain after PNV injection as it better mimicked PNV envenomation in humans, indicating that the mice model may help to manage/optimise acute pain control in humans.
There are no clinically validated therapies for treating PNVinduced pain. The treatment of pain in humans is typically symptomatic. Local anaesthetic infiltration is often used. If the pain persists, a specific serum therapy, analgesics or opiates may be used [2]. It was reported that the pain factor after a PNV injection in dogs may be neutralized by a specific antivenin [8,18]. We detected that the arachnid antivenin produced an antinociceptive effect when it was co-administered with PNV in mice, but not when it was administered 55 min after PNV injection.  These findings indicate that substances presented in PNV and neutralised by the antivenin are important to pain development, but not to pain maintenance, which seems to be dependent on the inflammatory process. In line with this idea, the non-steroidal antiinflammatory drug indomethacin did not prevent nociception development when administered before PNV, but was able to partially reverse nociception ( Figure 2) and oedema (data not shown) when administered 30 min after PNV. We also tested some other clinically used analgesics for the treatment of PNV-induced nociception. We observed that the local anaesthetic lidocaine and the opioid morphine are able to prevent, but not reverse, PNVinduced nociception in mice. Interestingly, PNV-induced nociception was both prevented and reversed by dipyrone and acetaminophen. In accordance with these results, dipyrone is widely used in Brazil to treat pain after spider bites when a local anaesthetic is not sufficient to eliminate pain [45]. Thus, our model could be useful for investigating not only the mechanisms responsible for the development and maintenance of PNV-induced pain but also effective therapies for treating pain related to P. nigriventer in humans. We next focused on the substances and mechanisms involved in the development of PNV-induced pain. Previous studies have indicated that dialysed venom seems to be less painful in dogs [3], which suggests that the nociception may be due to the presence of low-molecular-weight substances in PNV. In fact, the proinflammatory mediators histamine, serotonin and glutamate are present in PNV [9,46]. We were able to detect these mediators in PNV and observed that dialysis reduced the both the amount of these mediators and the nociceptive effect.
Glutamate induces nociception in mice via the peripheral activation of NMDA, AMPA and kainite receptors [47]. We found that neither NMDA nor AMPA/kainate antagonists were able to reduce PNV-induced nociception at doses that were able to reduce the glutamate-trigged nociceptive effect. A possible explanation for this finding is that the low amount of glutamate present in Phoneutria nigriventer venom (0.5 mg/g PNV or 0.00017 mmol/paw in 3 mg/ paw of PNV) was approximately 60 thousand times lower than the concentration used to induce a nociceptive effect (10 mmol/paw).
Histamine derived from venom or mast cell degranulation has been implicated in several of the biological effects of animal venoms [48,49]. In the present study, we found that mast cell depletion, which we previously demonstrated reduced the amount of tissue histamine by 88% [12], is ineffective for reducing the algesic effect produced by PNV. Based on the finding that the mast cell degranulator compound 48/80 (derived from bee venom) was capable of inducing nociception, it is concluded that mast cell mediators may release algogenic substances, but it is not the mechanism responsible for PNV-induced nociception. Although we found that PNV contains histamine, an H 1 receptor antagonist was not able to reduce the PNV-induced nociception. Of note, several studies have shown that histamine is important for the induction of oedema and pruritus, but not nociception [50]. In accordance with these findings, we observed that histamine induced oedema but not nociception when injected into the mice paw. Thus, neither the histamine derived from venom nor mast cell degranulation is related to the nociceptive effect of PNV.
The stimulation of serotonin receptors seems to be relevant to the production of pain [51]. Although we found serotonin in PNV, the amount (0.23 mg/ml or 0.23 mg/g PNV or 4 pmol/paw in 3 mg/paw of PNV) was approximately 25 thousand times lower than the concentration needed to induce a nociceptive effect in mice (100 nmol/paw), suggesting that serotonin is not relevant to the PNV-induced nociception. Moreover, we observed that serotonin caused nociception primarily through the 5-HT 3 and not the 5-HT 4 receptors, based on our finding that the co- administration of GR113908 did not reduce the nociceptive effect induced by serotonin. In contrast, PNV-trigged nociception is mediated by 5-HT 4 but not 5-HT 3 receptors. Our findings are in agreement with the electrophysiological study conducted by Costa and collaborators (2003) [52], which found that PNV induced rat vagus nerve depolarisation via the 5-HT 4 but not the 5-HT 3 receptor. Because this study used dialysed PNV, this effect does not appear to be mediated by a low-molecular-weight substance such as serotonin. Thus, PNV-induced nociception is dependent on 5-HT 4 stimulation by a still unknown and most likely high-molecularweight molecule. However, more studies on this topic are needed.
It was expected that voltage-gated Na + channels would contribute to PNV-induced nociception, given that local anaesthetics reduce PNV-induced pain in humans [2] and nociception in mice (current study). Furthermore, we observed that TTX was very efficacious in reducing PNV-induced nociception in mice without altering the nociception induced by PGE 2 , indicating an important role for tetrodotoxin-sensitive voltage-gated sodium channels in the pain caused by PNV, but not by PGE 2 . This is in agreement with a report that PGE 2 -induced nociception is mediated by tetrodotoxinresistant sodium channels in rodents [53]. Moreover, our findings suggest that the inhibitory effect of TTX in PNV-induced nociception was not simply due to conduction block in sensory neurons. Of note, families of small (3-9 kDa) peptide toxins have been detected in the venom of P. nigriventer [51]. Fraction 2 of PNV venom (PhTx2) contains at least two peptide toxins, namely PnTx2-5 and PnTx2-6, that directly bound and decreased the inactivation of tetrodotoxin-sensitive Na + channels and facilitated their activation, thus inducing membrane depolarisation in vitro [54,55,56,57]. However, more studies are needed to elucidate whether PNV components directly or indirectly activate voltagesensitive sodium channels to induce nociception in vivo.
In addition to peptide toxins, some larger (.10 kDa) protein components of PNV have also been described [54]. Our results also show that boiling PNV to induce protein denaturation reduced the nociceptive effect. Moreover, the PNV dialysed in membrane with a MWCO of 12-14 KDa presented a lower, but still significant, nociceptive effect. Thus, we suggest that both low-and high-molecular-weight components contribute to PNV-induced nociception. The presence of proteases in the PNV has been previously demonstrated in vitro [54], and another report showed that PNV contains a non-dialysable component with tissue   kallikrein-like activity in vivo that is involved in venom-induced oedema in rabbits [5]. Thus, we tested the possible involvement of the kallikrein-kinin system in PNV-induced nociception using biochemical and pharmacological tools. First, we detected a potent kallikrein-like activity in the venom in vitro that was active against a synthetic substrate and cleaved a low-molecular-weight kininogen to generate kinins. Second, based on the substrate preference (lowversus high-molecular-weight kininogen) and the sensitivity to inhibitors (aprotinin, but not SBTI) [58,25], we suggest that this activity resembles a tissue kallikrein-like enzyme. Third, we observed that inactivating (by boiling) or inhibiting the kallikreinlike activity reduced PNV-trigged nociception. Consistent with our data, the PNV-induced oedema formation in rabbit skin involves local kinin synthesis in response to tissue (but not plasma) kallikreinkininogen-kinin system activation [6]. Thus, the kallikrein-like activity plays a critical role in PNV-induced nociception, but further studies are needed to identify this component. Given that the tissue kallikrein of PNV cleaves kininogen, it could promote the generation of kinins and the stimulation of B 2 receptors present in sensory fibres to induce nociceptor depolarisation and a consequent pain response [36]. In fact, we found that the tissue protease inhibitor aprotinin and the B 2 receptor antagonist HOE 140 inhibited PNV-induced nociception. B2 receptor antagonists prevented PNV-induced plasma extravasation in rabbits [6] and nociception in mice (current study), but not plasma extravasation in rats [7]. This discrepancy may be related to the inter-species variability of the kinin system [59] and by the 10-to 100-fold lower dose of PNV (and consequently low kallikrein-like activity) used in the rat study (0.3 mg/site) compared to the rabbit (30 mg/site) or mice (3 mg/paw) studies.
It is well known that kinin activity at B 2 receptors excites sensory nerve endings by both the activation and sensitisation of TRPV1 receptors in vivo [11]. We observed that the antagonism of TRPV1 or the ablation of TRPV1-positive sensory fibres reduced the PNV-induced nociception. In addition to the indirect effects, it has been recently demonstrated that spider venoms may contain toxins that directly bind to and stimulate the TRPV1 receptor [37,38,39]. Accordingly, we observed that low concentrations of PNV are capable of activating HEK cells expressing TRPV1 receptors, but not control cells. Moreover, our binding assay demonstrated that PNV components do not bind to the TRPV1 receptor at the intracellular vanilloid site, similar to other spider venom toxins that interact at extracellular sites of TRPV1 [37,38,39]. However, the exact components presented in PNV that stimulate TRPV1 and their site of interaction must be further identified. Our data indicate that, besides indirect mechanisms, PNV may contain components that directly bind and stimulate the TRPV1 receptor.
It has been well established that the treatment of animals with TRPV1 receptor agonists, such as RTX or capsaicin, produces a selective ablation of TRPV1-positive sensory fibres and largely reduced the amount of the neuropetides substance P and neurokinin A [11,60,61,62]. We observed that the ablation of such peptidergic fibres significantly inhibited PNV-induced nociception. The reduced time that the animals spent licking the paws could theoretically be caused by motor impairments or sedation induced by the treatment RTX, rather than reflecting an analgesic effect. However, the number of crossings in the open-field test and the per cent of sedation in the platform test were similar in vehicle and RTX groups (data not shown), excluding this possibility.
In addition to the ablation of peptidergic fibres, the antagonism of the tachykinin NK 1 and NK 2 receptors also significantly inhibited PNV-induced nociception. Consistent with this, it was reported that peripheral tachykinin receptor antagonists reduced PNV-induced nociception in rats [44]. TRPV1-positive sensory fibres express NK 1 and NK 2 receptors as well as produce the selective ligands substance P and neurokinin A, which are released in response to nociceptive stimulation and may produce paracrine or autocrine effects that amplify the nociceptive process [63]. In fact, NK 1 and NK 2 receptors expressed by sensory nerves cause curves for capsaicin (B) or PNV (D) on transient calcium signals. When applied to cells twice, both capsaicin (200 nM) (E) and PNV (1 mg/mL) (G) elicited Ca 2+ transient signals with similar amplitudes. Such response was totally abolished by pre-incubation (300 s) with 10 mM SB366791 (F and H). Points denote mean6S.E.M. of representative plates (10-30 cells, 3 independent experiments). doi:10.1371/journal.pntd.0002198.g008 sensitisation of TRPV1 receptors, increase neuronal excitability and induce nociception in rodents [63,64,65]. We observed that a combination of NK 1 and NK 2 receptor antagonists was as efficacious as the antagonists alone to reduce PNV-induced nociception, but was less efficacious than sensory nerve ablation. These data indicate that NK 1 and NK 2 receptors use a convergent pathway to induce nociception (such as an indirect stimulation of the TRPV1 receptor) and that other sensory nerve-dependent mechanisms (such as a direct stimulation of TRPV1) are related to PNV-induced nociception. Thus, the nociceptive stimulation produced by PNV seems to release neuropeptides from sensory fibres, which sensitise the TRPV1 receptor by acting on NK 1 and at NK 2 , and might amplify the nociceptive response. Of note, though NK1 receptor antagonists are usually able to inhibit pain in sensitised states and thus reverse hyperalgesia, we observed that tachykinin receptor antagonists prevented PNV-induced nociception. In accordance, subcutaneous injection of substance P has been shown to induce continuous pain in humans and nociception in rodents [64,66], and peripheral NK1 and NK2 antagonists may prevent nociception induced by several algogens such as capsaicin and formalin [67].
Finally, we also studied the involvement of acid-sensing ion channels in PNV-induced nociception because these ion channels are an important target of several animal venoms [68]. Our results indicate that the ASIC receptor is involved in PNV-induced nociception because the co-administration of the ASIC blocker amiloride inhibited the PNV nociceptive effect. PNV could produce ASIC activation by both direct and indirect mechanisms. It has been demonstrated that the stimulation of Gq proteincoupled receptors, such as the B 2 or 5-HT 4 receptors that mediate PNV-induced nociception, may stimulate ASICs through protein kinase C [69]. Furthermore, it has been shown that arachidonic acid, in a manner independent of its metabolisation by cyclooxygenase, markedly sensitizes ASICs [70]. Accordingly, we observed that the cyclooxygenase inhibitor indomethacin did not reduce PNV-induced nociception. Of note, phospholipase A 2 , the enzyme responsible for arachidonic acid production, was detected in PNV [71]. Finally, we cannot exclude the possibility that PNV components could directly activate ASICs, as it was recently described that snake toxins isolated from Micrurus tener venom induce nociception through the direct stimulation of ASICs in sensory neurons [72]. However, more specific studies are needed to better understand this mechanism. Thus, PNV stimulates ASICs to produce nociception, an effect that could be mediated by both direct and indirect mechanisms.
In summary, the current results show that both the low-and high-molecular-weight substances in PNV produce spontaneous nociception action through direct or indirect activation of B 2 , TRPV1, 5-HT 4 or ASIC receptors and the tetrodotoxin-sensitive voltage-gated Na + channels present in primary afferents neurons, but not in mast cells. The elucidation of the mechanisms responsible for the nociception induced by PNV is of interest to better treat envenomation by P. nigriventer and understand the diversity of targets triggered by PNV toxins.