Coniine is an optically active toxic piperidine alkaloid and nicotinic acetylcholine receptor (nAChR) agonist found in poison hemlock (Conium maculatum L.). Coniine teratogenicity is hypothesized to be attributable to the binding, activation, and prolonged desensitization of fetal muscle-type nAChR, which results in the complete inhibition of fetal movement. However, pharmacological evidence of coniine actions at fetal muscle-type nAChR is lacking. The present study compared (−)-coniine, (+)-coniine, and nicotine for the ability to inhibit fetal movement in a day 40 pregnant goat model and in TE-671 cells that express fetal muscle-type nAChR. Furthermore, α-conotoxins (CTx) EI and GI were used to antagonize the actions of (+)- and (−)-coniine in TE-671 cells. (−)-Coniine was more effective at eliciting electrical changes in TE-671 cells and inhibiting fetal movement than was (+)-coniine, suggesting stereoselectivity by the receptor. The pyridine alkaloid nicotine did not inhibit fetal movement in a day 40 pregnant goat model, suggesting agonist specificity for the inhibition of fetal movement. Low concentrations of both CTxs potentiated the TE-671 cell response and higher concentrations of CTx EI, and GI antagonized the actions of both coniine enantiomers demonstrating concentration-dependent coagonism and selective antagonism. These results provide pharmacological evidence that the piperidine alkaloid coniine is acting at fetal muscle-type nAChR in a concentration-dependent manner.
Coniine is an optically active piperidine alkaloid and nicotinic acetylcholine receptor (nAChR) agonist found in poison hemlock (Conium maculatum L.) and introduced from Europe (Marion, 1950; Lopez et al., 1999; Vetter, 2004; Lee et al., 2008) (Fig. 1). Acute poisonings by poison hemlock are well documented in both humans and animals (Frank et al., 1995; West et al., 2009; Swerczek 2012). Livestock typically consume poison hemlock plant material or seed, in contaminated feed or forages, and the alkaloids readily cross the placenta and their developing fetuses are exposed to coniine as a result (Edmonds et al., 1972; Panter et al., 1985b; Schep et al., 2009). If animals survive the initial exposure to poison hemlock, they may develop a preference for the plant, and pigs in particular develop an appetite for poison hemlock (Kingsbury 1964; Panter et al., 1985b, 1988). When coniine concentrations in the plant are high enough to cause clinical signs in the pregnant mother, fetal movement is inhibited and multiple congenital contractures (MCC) and cleft palate often develop in the offspring (Panter et al., 1988). The gestational periods of susceptibility vary, but in general, fetal movement must be inhibited for an extended period during days 40 to 100 of gestation. Specific congenital fetal defects that have been directly attributed to fetal coniine intoxication include arthrogyrposis, scoliosis, torticollis, kyphosis, lordosis, and cleft palate (Panter and Keeler, 1992, 1993; Panter et al., 1998). Piperidine alkaloids cause cleft palates by physical obstruction of the palatine shelves by the tongue during programmed palate closure (Panter et al., 1985a; Weinzweig et al., 2008). Signs of acute poisoning include urination, defecation, tachycardia, muscle weakness, muscle fasciculations, ataxia, collapse, and death from respiratory failure (Panter et al., 1988, 1999).
The acute effects of coniine are attributable to its ability to activate and then desensitize nAChRs from the persistent action of the alkaloid at the receptor. This leads to receptor desensitization and inhibition of cation conductances (Buccafusco, 2004). The MCC teratogenic actions of coniine are at fetal muscle-type nAChRs, which are ligand-gated cation channels expressed by the developing fetus. These nAChR have a subunit composition of (α1)2β1γδ, with ligand-binding sites at the αδ and αγ subunits interfaces and activated by the binding of two molecules of acetylcholine. Studies using coniine and other nAChR agonists, such as the pyridine alkaloid nicotine (Fig. 1), suggest that teratogenesis results from the binding, activation, and inhibition of these receptors (Green et al., 2012).
The teratogenicity of coniine was first documented at this laboratory by Keeler (1974). Dosing of 3.3 mg/kg coniine to pregnant cattle at days 55 to 70 of gestation resulted in the formation of terata in the offspring. Sheep required daily doses of 44 mg/kg coniine at 12 to 65 days of gestation to produce terata (Keeler et al., 1980). Later work at this laboratory by Panter et al. (1990a) documented poison hemlock inhibition of fetal movement in pregnant goats at 45 days’ gestation with fetal movement reduced to an average of 0.5 movements per minute, compared with normal fetuses, which show over 15 discrete movements per minute (Panter et al., 1990a). More recently, Lee et al. (2008) showed that coniine enantiomers have differing toxicity, and they ranked (–)-coniine > (±)-coniine > (+)-coniine for both potency at fetal muscle-type nAChR expressed by TE-671 cells and mouse toxicity. The EC50 concentrations for (–)-coniine and (+)-coniine were reported as 115 and 900 µM, respectively, at fetal muscle-type nAChR and 9.6 and 10.2 µM, respectively, at human autonomic type nAChR (predominantly α3β4 nAChR) (Green et al., 2010). The LD50 concentrations for coniine enantiomers in mice were 7.0 and 12.1 mg/kg i.v. for (–)- and (+)-coniine respectively (Lee et al., 2008). These findings indicated that the (−)-coniine enantiomer is more potent. There have been no recent reports of the intravenous administration of purified coniine compounds to a large animal, such as a goat; the effects of intravenous administration of purified coniine enantiomers on fetal movement in livestock; or reports of the competitive antagonism of coniine actions by conotoxins.
We hypothesized that (−)-coniine would be more effective than (+)-coniine at inhibiting fetal movement in a day 40 pregnant goat model and that the actions of (+)- and (−) coniine can be antagonized by α-conotoxin (CTx) EI and GI in TE-671 cells. Over the course of testing the hypothesis, we sought to demonstrate stereoselectivity, agonist specificity, concentration dependency, and antagonism of coniine actions using TE-671 cells and a goat model.
Materials and Methods
All animal work was done under veterinary supervision with the approval of the Utah State University Institutional Animal Care and Use Committee. The estrous cycles of 40 Spanish-type female goats weighing 33.5 ± 1.8 kg were synchronized using Chronogest vaginal sponges containing 45 mg of flugestone acetate (Intervet International B.V., Boxmeer, the Netherlands). Approximately 24 to 48 hours after sponge removal, the does were bred by Spanish-type bucks, which was considered to be day 0 of gestation. The does were maintained ad libitum on water, salt, and alfalfa hay for 40 days from when they were bred by the buck. At day 40, the does were administered intravenously with the alkaloids dissolved in physiologic saline, saline control, or mandelic acid controls. The volume of saline administered (0.8 ml/kg for the coniine mandelic acid solution and 0.4 ml/kg for nicotine) was equal in volume to that of the drug solutions administered on a per kilogram basis. The mandelic acid dose of 4.4 mg/kg was equal on a molar basis to that administered in the form of coniine mandelic acid compounds. Serum concentrations of coniine in the goats were not directly measured; however, assuming an 8% by body weight blood volume and a 40% hematocrit, a 33.5-kg goat would have an approximate serum coniine concentration of 896 μM at the 8mg/kg coniine mandelic acid dose used in this study. The number of fetal movements and movement episodes were measured by ultrasound (Model SSD-3500; Aloka Corporation, Wallingford, CT). If an ultrasound field could be focused on two fetuses simultaneously, both fetuses were treated as individual experimental units. However, in all cases, a minimum of three pregnant does were tested at each time for each compound. For the dose-finding experiments, yearling wethers were given treatment intravenously on a milligram per kilogram body weight basis with either (−)-coniine (+)-mandelic acid, (+)-coniine (−)-mandelic acid, or (−)-nicotine hydrogen tartrate salt (Sigma-Aldrich, St. Louis, MO) to identify the minimum dose of alkaloid needed to produce clinical signs of intoxication, including salivation, urination, defecation, and muscle fasciculations, followed by collapse onto their sternum (data not shown). This dose was then used in the day 40 pregnant goat study.
Coniine Mandelate Crystallization Procedure
(±)-Coniine (502.6 mg, 3.95 mmol) and (−)-mandelic acid (605.5 mg, 3.98 mmol) were added to a 16-ml screw top glass test tube. To the (±)-coniine and (−)-mandelic acid mixture was added 1.6 ml of methanol, and the mixture was warmed until all (±)-coniine and (−)-mandelic acid was dissolved; anhydrous diethyl ether (4.00 ml; Thermo Fisher Scientific, Fair Lawn, NJ) then was added, and the mixture was placed on ice. Crystals of (+)-coniine (−)-mandelate began to form within 30 minutes after adding the diethyl ether. After 16 hours at −20°C, the solvent was filtered from the crystals, and the crystals were collected and dried in vacuo. The crystals were recrystallized three times by dissolving the crystals in methanol (1.6 ml), adding diethyl ether (4 ml), and then allowing them to crystallize at −20°C. The (+)-coniine (−)-mandelate crystals were filtered for a final time, collected, and dried in vacuo (314.7 mg, 1.13 mmol, 57% yield): feathery needles, melting point 115°C; [α]25.1 D = −47.8° (c = 0.52, methanol). This same procedure was used with (±)-coniine (513.7 mg, 4.04 mmol) and (+)-mandelic acid (635.1 mg, 4.17 mmol). The (−)-coniine (+)-mandelate crystals were collected and dried in vacuo (364.0 mg, 1.30 mmol, 65% yield): feathery needles, melting point 117–119°C; [α]24.7 D = +49.0° (c = 0.60, methanol). (+)- and (−)-Coniine-HCl were made as described in Lee et al. (2008).
Cell Culture Assays
Fetal bovine serum was obtained from Hyclone, Inc. (Logan, UT). Penicillin/streptomycin was obtained from Invitrogen (Carlsbad, CA). Dulbecco’s modified Eagle’s medium was from the American Type Culture Collection (Manassas, VA) or Gibco (Grand Island, NY), and the fluorescence dye kits were purchased from Molecular Devices (Sunnyvale, CA). CTx GI obtained was obtained from American Peptide Company (Sunnyvale, CA) and is a α3/5-CTx subfamily conotoxin isolated from Conus geographus that selectively targets the αδ subunit interface ligand-binding site with greater than 10,000-fold affinity over the αγ subunit interface ligand-binding site (Gray et al., 1981; Groebe et al., 1995). We also used CTx EI obtained from American Peptide Company, classified as an α 4/7-CTx isolated from Conus ermineus, which targets both mammalian αδ and αγ subunit interface ligand-binding sites with nearly equal affinity (Martinez et al., 1995; Azam and McIntosh, 2009). The amino acid sequences of the CTxs used in this work are presented in Azam and McIntosh (2009). Acetylcholine chloride (ACh), (±)-coniine, and (−)-nicotine hydrogen tartrate salt were obtained from Sigma-Aldrich.
The human rhabdomyosarcoma cell line TE-671 was obtained from American Type Culture Collection. Membrane depolarization responses from the addition of nAChR agonists were measured by changes in fluorescence of a membrane potential-sensitive dye as previously described (Green et al., 2010). A brief summary of the method follows: TE-671 cells in 96-well black-walled cell culture plates were equilibrated to room temperature for 10 minutes and then loaded with dye for 30 minutes. Serial dilutions of a compound were prepared in 96-well V-bottom plates. Fluid (agonist or KCl) additions and membrane potential measurements were performed using a Flexstation II (Molecular Devices Corporation). Readings were taken every 1.12 seconds for 255 seconds, for a total of 228 readings per well. The first 17 seconds were used as a basal reading. At 18 seconds, 50 μl of a test compound was added to assess agonist activity. At 180 seconds, 25 μl of KCl in saline was added to attain a final concentration of 40 mM KCl in the dye-Hanks’ balanced salt solution bathing the cells. Responses were calculated as equal to: (FMax(Compound) − FBasal)/(FMax(Calibrant) − FBasal).
In all experiments, 10 duplicate wells of a dilution series of epibatidine were included, and each experiment was visually inspected for a stable baseline for the data from that 96-well plate to be included in analysis of individual compound responses. The depolarizing responses to the coniine enantiomers, acetylcholine, and nicotine were normalized to the maximum responses generated by (±)-epibatidine.
Nonlinear regression models (sigmoidal dose-response [variable slope] and sigmoidal dose-response) were compared using F-test. In most cases, the sigmoidal dose-response equation was the preferred model and, therefore, was used to determine the EC50 values and graph the best-fit lines of all data presented in Figs. 3, 4, 5, 6, 7, and 9A with Prism, version 5.04 (GraphPad Software, San Diego, CA). Nicotinic agonists may cause desensitization at high concentrations; thus, the concentration-effect curves were obtained by fitting all concentrations, including those on the plateau of the line. In general, only the 1 mM concentration was excluded from the EC50 calculation. Two-way analysis of variance (ANOVA) comparisons of concentration-effect experiment data were also analyzed using GraphPad Prism. The results from the coniine day 40 pregnant goat experiments were analyzed using SAS software, version 9.2 (SAS Institute Inc., Cary, NC). Fetal movement was examined using a mixed-model repeated-measures analysis. The model included animals, treatments, and time, and their interactions, with animals a random factor in the model. The best-fitting within-subject covariance structure for fetal movement was autoregressive, as determined by Akaike’s Information Criterion. There was a time x treatment interaction, and the least square means of the preplanned comparisons were examined using the PDIFF option in SAS. Results from the nicotine day 40 fetal goat model experiments were analyzed using one-way ANOVA with Tukey’s multiple comparison test as a post-test in Prism. In all cases, the limit for statistical significance was set at P < 0.05.
Concentration-Dependent Effects of CTx EI and GI on the Actions of (+)- and (−)-Coniine HCl at Fetal Muscle-Type nAChR Expressed by TE-671 Cells
The first series of concentration-effect experiments with coniine HCl were in the absence and presence of the nAChR antagonists CTx EI and GI. This was done to identify the optimum concentration of CTxs to use for the blockade of agonist actions in TE-671 cells. Representative tracings of the responses are displayed in Fig. 2. Each well of every 96-well plate in every experiment received a final concentration of 40 mM KCl at 180 seconds that served as a depolarizing calibrant to correct for inter-well differences in dye loading and cell count. The EC50 values of (+)- and (−)-coniine HCl were in the micromolar range, and (−)-coniine HCl was much more effective at eliciting changes in membrane potential sensing dye fluorescence (Tables 1 and 2; Figs. 3 and 4). Thirty minute CTx EI or GI pretreatment of TE-671 cells before the addition of (+)- or (−)-coniine HCl resulted in significant interactions for both enantiomers of coniine HCl and both CTxs (P < 0.0001, P < 0.0001; two-way ANOVA, CTx EI, (+)-, and (−)- coniine HCl respectively; P = 0.0132, P = 0.0001; two-way ANOVA, CTx GI, (+)-, and (−)- coniine HCl respectively).
There were also individual differences between the coniine enantiomer concentration-effect curves pretreated with CTxs, compared with the untreated control curves. Pretreatment with 0.001 or 0.01 μM CTx EI resulted in significant differences from control (P = 0.0001, P < 0.0001; for 0.001 and 0.01 μM CTx EI pretreated (+)-coniine HCl concentration-effect curves vs. the (+)-coniine HCl control concentration-effect curve, respectively; P < 0.0001; for 0.001 μM CTx EI pretreated (−)-coniine HCl concentration-effect curves vs. the (−)-coniine HCl control concentration-effect curve; F-test). Pretreatment with 0.001 and 0.01 μM CTx EI also significantly increased the percentage maximum responses of TE-671 cells to the (+)- and (−)-coniine HCl enantiomers (Fig. 3; Table 1). One micromolar CTx EI depressed the maxima of both enantiomer concentration-effect curves and shifted the EC50 values to the right.
Pretreatment of the cells with 0.001 μM CTx GI also shifted the coniine enantiomer curves (P < 0.0001, P = 0.0109; for 0.001 μM CTx GI pretreated (+)- and (−)-coniine concentration-effect curves vs. the (+)- and (−)-coniine HCl control concentration-effect curves, respectively; F-test). CTx GI at a concentration of 0.001 μM increased the maximum TE-671 cell response to the coniine HCl enantiomers by 35 and 19% for (+)- and (−)-coniine HCl, respectively (Fig. 4; Table 2), but not in a manner that was significantly different from control (P > 0.05, two-way ANOVA, Bonferroni’s multiple comparisons test). At concentrations of 0.1 and 1.0 μM, CTx GI depressed the responses of TE-671 cells to both coniine HCl enantiomers. On the basis of the results of these series of experiments, we chose the 1.0 μM concentrations of CTx EI and GI to block agonist actions at fetal muscle-type nAChR expressed by TE-671 cells.
Effects of 1.0 μM CTx EI and GI on the Actions of ACh
The effects of CTx EI and GI on actions ACh in TE-671 cells were examined next (Fig. 5, Table 3). ACh was a full agonist with a maximum response of 96 ± 2% of epibatidine. Thirty-minute pretreatment with 1μM CTx EI shifted the ACh concentration-effect curve to the right and increased the maximum response of TE-671 cells to ACh by 25%. Although 1μM CTx GI did not shift the ACh concentration-effect curve to the right, the maximum response of TE-671 cells to ACh was increased by 11%. The three concentration-effect curves were compared using the F-test, and there were significant differences among them (P = 0.0014). The data were analyzed by two-way ANOVA, and there was a significant interaction between drug treatment and concentration (P = 0.02). When the datum points were examined using a Bonferroni’s multiple comparisons post-test, 1 μM ACh in the presence of 1 μM CTx EI was significantly different from control (P < 0.05, two-way ANOVA, Bonferroni’s multiple comparison test).
Effects of 1.0 μM CTx EI and GI on the Actions of Coniine Mandelic Acid Combinations in TE-671 Cells
The next series of experiments documented the potencies and efficacies of the coniine mandelic acid combinations in TE-671 cells (Fig. 6; Table 4). Both of the coniine mandelic acid enantiomer combinations acted as weak agonists, initiating a concentration-dependent change in membrane potential sensing dye florescence. The cellular responses to the individual mandelic acid enantiomers alone were not significantly different from zero (P > 0.05, F-test; n = 6 and 7 for the (+) and (−) enantiomer, respectively). Pretreatment of TE-671 cells with 1.0 μM CTx GI shifted and depressed the responses of both (+)- and (−)-coniine mandelic acid combinations as expected. Conversely, 1.0 μM CTx EI pretreatment did not block the actions of the coniine mandelic acid combinations in TE-671 cells. Next, we repeated these experiments with (±)-coniine and free mandelic acid mixtures in place of the individual enantiomers recrystallized with mandelic acid. The actions of (±)-coniine in the absence of mandelic acid in TE-671 cells were blocked by pretreatment with 1.0 μM concentrations of CTx EI and CTx GI (Fig. 7). When either of the free mandelic acid enantiomers was included with serial dilutions of (±)-coniine, CTx EI did not block the (±)-coniine induced change in membrane potential sensing dye fluorescence. In the absence of coniine, the mandelic acid enantiomers and CTx EI were also without effect, except at a 1 mM concentration of (+)-mandelic acid. Of interest, CTx GI blocked all concentrations of (±)-coniine in such a manner that the software package was unable to determine a best-fit line. The EC50 values for the coniine racemate in the absence and presence of the CTxs are displayed in Table 3. The EC50 values for the (±)-coniine, (+)-,(−)-mandelic acid combinations in the presence of 1 μM CTx EI were 209 and 162 μM, respectively (95% confidence intervals: 65–677 and 32–811 μM, and the (±)-coniine, (+)-,(−)-mandelic acid combinations in the presence of 1 μM CTx EI, respectively). The (±)-coniine and the mandelic acid combination concentration-effect curves were compared using the F-test, and there were no significant differences (P = 0.0821; (±)-coniine control, and the (±)-coniine, (+)-,(−)-mandelic acid combinations in the presence of 1 μM CTx EI).
(+)-Coniine (−)-Mandelic Acid and (−)-Coniine (+)-Mandelic Acid Actions in a Day 40 Pregnant Goat Model
After completing the series of TE-671 cell experiments, the coniine mandelic acid enantiomers were tested in the day 40 pregnant goat model. In this experiment, (+)-coniine (−)-mandelic acid and (−)-coniine (+)-mandelic acid compounds were crystallized from racemic liquid coniine purchased from Sigma-Aldrich, as described in the Materials and Methods. The enantiomers were not converted to the hydrochloride salt that we used in previous experiments because of the loss of coniine (low yield) when converting coniine mandalate to coniine HCl.
When day 40 pregnant does were given a 8.0 mg/kg concentration of either (+)- or (−)-coniine mandelic acid, there were immediate signs of cholinergic toxicity that included salivation, urination, defecation, and muscle fasiculations, followed by collapse onto their sternum. The does recovered and were standing by 15 to 30 min after administration. Ultrasound data from the day 40 pregnant goat experiments were analyzed by repeated measure analysis, which identified a time by treatment interaction for fetal movement (P = 0.0003). Fetal movement was reduced to zero at 30 minutes after injection for both the (+)- and (−)-coniine mandelic acid enantiomers (Fig. 8; Table 5). At 1 hour after injection, the fetal movement began to recover in (+)-coniine (−)-mandelic acid–treated animals, whereas fetal movement in the (−)-coniine (+)-mandelic acid–treated animals was still zero. It was not until 2 hours after intravenous administration that fetal movement in the (−)-coniine (+)-mandelic acid–treated animals started to recover. However, these movements were still significantly different from the saline control at that time (P = 0.0239, mixed-model repeated-measures analysis, (−)-coniine (+)-mandelic acid vs. saline control). Fetal movement in (−)-coniine (+)-mandelic acid–treated animals returned to a similar value as the saline control at 8 hours after injection (P = 0.3865, mixed-model repeated-measures analysis (−)-coniine (+)-mandelic acid vs. saline control). Both (+)-mandelic acid and (−)-mandelic acid were not significantly different from the corresponding saline controls at any of the time points examined (P > 0.05 mixed-model repeated-measures analysis).
Effects of 1.0 μM CTx EI and GI on the Actions of Nicotine and the Effect of Nicotine on Fetal Movement in a Day 40 Pregnant Goat Model
In a separate series of experiments, as a piperidine to pyridine alkaloid structural control, we tested the actions of nicotine in both TE-671 cells and the day 40 pregnant goat model. A representative tracing of the raw data for a nicotine response is displayed in Fig. 2. In the TE-671 cell experiments, the nicotine concentration-effect curves were shifted to the right in the presence of CTx EI or GI, and the three curves were significantly different, as determined by F-test (P < 0.0001) (Fig. 9A). CTx GI decreased the maximum response of 1 mM nicotine by 69%, compared with control (Table 3). When the data were analyzed by two-way ANOVA, there was significant variation between treatment and concentration (P = 0.004, P < 0.0001; for treatment and concentration respectively, two-way ANOVA). The 100 μM and 1 mM nicotine concentrations in the presence of 1 μM CTx GI were significantly different from their respective nicotine control values (P < 0.001, P < 0.01 for nicotine concentrations of 100 μM, and 1 mM, respectively, two-way ANOVA, Bonferroni’s multiple comparison test).
In the day 40 pregnant goat model, nicotine at an intravenous dose of 0.6 mg/kg caused immediate clinical signs of cholinergic toxicity in the does, although they recovered similarly to those given coniine. When the ultrasound fetal movement data were analyzed, there were no significant differences detected (P > 0.05, ANOVA, nicotine vs. saline for each time point) between the saline controls and the nicotine-treated fetuses (Fig. 9B).
Numerous reports have described the teratogenicity of coniine (Green et al., 2012). However, they typically used orally administered poison hemlock plant material or included experiments that did not monitor the effects of coniine on fetal movement (e.g., Keeler and Balls, 1978; Panter et al., 1990a, b). In a chick model, Forsyth et al. (1996) documented that nicotine and coniine can cause deformities and inhibit movement, but they were unable to block coniine actions with nAChR antagonists and, thereby, unable to conclude nAChR-mediated effects of the agonists. This research attempts to overcome these shortcomings by providing pharmacological evidence for coniine actions at fetal muscle-type nAChR. We used TE-671 cells, a human rhabdomyosarcoma cell line that expresses fetal muscle-type nAChR for in vitro experiments. Second, we used a fetal goat movement model for in vivo radio-ultrasound experiments to measure the inhibition of fetal movement, the putative mechanism behind MCC-type defect formation.
In TE-671 cells, the effects of coniine were concentration dependent (Figs. 3, 4, and 6) and were antagonized by both CTxs. The results from the first cell-based experiments did not have the expected order of potency for the coniine HCl enantiomers, as observed in previous work (Lee et al., 2008; Green et al., 2010). This result could have been attributable to the sequential nature of the experiments, in which concentration-effect curves for each enantiomer were tested in the presence and absence of both CTxs separately over an extended period. Later experiments with the coniine mandelic acid combinations during the same two-week period had the expected order of potency in TE-671 cells of (−)-coniine (+)-mandelic acid > (+)-coniine (−) mandelic acid. We also confirmed the coniine stereoselective inhibition of fetal movement in goats hypothesis. Specifically, in goats given (−)-coniine (+)-mandelic acid, there were no observed fetal movements for one hour and significant reductions in movement for up to 4 hours. Movement in the (+)-enantiomer–treated animals was observed only at 30 minutes after injection (Fig. 8). These observations confirm our hypothesis that (−)-coniine would be more effective than (+)-coniine at inhibiting fetal movement in a day 40 pregnant goat model and that the actions of coniine can be antagonized by CTxs in TE-671 cells to provide pharmacological evidence for coniine actions at fetal muscle-type nAChR.
Next, we compared the actions of coniine, a piperidine alkaloid, to nicotine, a pyridine alkaloid, in cell culture and goats. There are reports of sows that consumed waste stalks of tobacco produced deformed piglets, and the teratogen was assumed to be nicotine (Crowe, 1969; Menges et al., 1970; Crowe and Pike, 1973; Crowe and Swerczek, 1974). However, when nicotine was given to pregnant sows as an isolated alkaloid, it did not produce MCC deformities in the sow's piglets; thus, the deformities were attributed to anabasine, which is found at high concentrations in the pith of tobacco stalks (Crowe, 1978; Keeler, 1979).
In this study, nicotine-like coniine acted as an agonist in TE-671 cells (Fig. 9A), and its actions were antagonized by 1 μM CTx GI. CTx GI has a 10,000-fold greater affinity for the αδ ligand-binding site of the fetal muscle-type nAChR than the αγ ligand-binding site (Azam and Mcintosh, 2009). This result suggests that nicotine has a preference for the αδ site in this model system. Conversely, the actions of coniine were sensitive to both CTxs, although there was a complete block of coniine actions by 1 μM CTx GI (Fig. 4), suggesting a slight preference for the αδ site. In the day 40 pregnant goat model, nicotine was without significant effect on fetal movement. These results suggest that, although both alkaloids are agonists at fetal muscle-type nAChR, there are agonist-specific differences in their interactions with the receptor in vivo and in vitro. Moreover, the inhibition of fetal movement and, thus, teratogenicity is restricted to the piperidine alkaloid coniine, which has the required structural features for teratogenicity (Keeler and Balls, 1978). This observation provides evidence for agonist-specific inhibition of fetal movement by piperidine alkaloids at fetal muscle-type nAChR.
In this study, we used two CTxs from different subfamilies to control for any differences in agonist-antagonist interactions at the ligand-binding site. In previous studies with TE-671 cells, we have documented agonist-specific effects with methyllycaconitine, an nAChR-selective antagonist (Green et al., 2011). In this work, each of the test agonists were antagonized by one or both CTxs, suggesting that there are differential agonist-antagonist–specific interactions at this receptor. It is noteworthy that CTxs at low concentrations (typically 1 and 10 nM) potentiated the responses of TE-671 cells to coniine HCl (Figs. 3 and 4). Potentiation of agonist responses by low concentrations of CTx EI has been observed by other investigators. For example, preincubation of TE-671 cells with 10 nM CTx EI potentiates the responses of the cells to nicotine (Lopez-Vera et al., 2007). The potentiation of the coniine response was likely to have been attributable to coagonism. Coagonism occurs at hetero-pentamer nAChR, which possess two ligand-binding sites when the receptor is activated by the binding of one molecule of antagonist and one molecule of agonist (Cachelin and Rust, 1994; Steinbach and Chen, 1995). We have previously documented that other nAChR antagonists, such as (d)-tubocurarine and methyllycaconitine, can act as coagonists in this model system (Green et al., 2011). Moreover, CTx GI has a 10,000-fold greater affinity for the αδ ligand-binding site, compared with αγ site, providing additional support for coagonism as the mechanism of CTx potentiation.
One unexpected result from these experiments was the rightward curve-shift and depression of the maximum coniine response of TE-671 cells in the presence of 1.0 μM CTx EI and 0.1 and 1.0 μM CTx GI (Figs. 3 and 4). CTxs have been classified as competitive antagonists, which we did observe in the ACh experiments (Fig. 5), suggesting that noncompetitive blockade was unlikely (Arias and Blanton, 2000; Lopez-Vera, et al., 2007). An alternative explanation of these results is that there was the formation of a hemi-equilibrium state between the CTxs and between nicotine and coniine, but not ACh. The formation of a hemi-equilibrium state because of slow dissociation of the antagonist from the receptor and incomplete re-equilibration after agonist addition would decrease the maxima of concentration-effect curves (Kenakin, 2009). CTxs are known to have slow dissociations rates from nAChR ligand-binding sites (Prince and Sine, 1999). We therefore speculate that these differences in antagonism between ACh and the other agonists are 2-fold. First, there may be specific differences in the ligand-binding site agonist-antagonist interactions that could affect the rate of offset of the antagonist from the receptor during re-equilibration after agonist addition. Second, the short period of the membrane potential sensing dye response could preclude the ability of the system to re-equilibrate after the addition of coniine or nicotine in the presence of the CTxs.
In this study, coniine mandelic acid compounds were tested in both TE-671 cells and goats. In TE-671 cells, the presence of mandelic acid prevented the antagonism of the coniine response by CTx EI but not GI (Fig. 6). To confirm this observation, we ran a separate series of concentration-effect experiments with mixtures of (±)-coniine (the starting material for the isolation of the enantiomers from Sigma-Aldrich) and the mandelic acids in the presence of CTx EI (Fig. 7). In these experiments, (±)-coniine alone was antagonized by CTx EI and GI. However, in the presence of mandelic acid, CTx EI was without effect. The lack of CTx EI antagonism could have been attributable to mandelic acid interactions with the receptor or the alkaloid itself. The exact mechanism of mandelic acid inhibition of CTx nAChR antagonism is beyond the scope of this manuscript. We speculate that this observation may have important clinical implications for human medicine, because CTxs have been identified as promising drug development leads (Azam and McIntosh; 2009), and mandelic acid is found in urine after occupational exposure to styrene monomer (Rueff et al., 2009).
Over the course of testing our hypothesis that the teratogenicity of piperidine alkaloids is attributable to their actions at fetal muscle-type nAChR, we used both in vitro and in vivo models. In this study as in previous reports, we have documented the concentration dependency of coniine enantiomers in TE-671 cells (Lee et al., 2008; Green et al., 2010). We demonstrated the stereoselectivity of fetal muscle-type nAChR for (−)-coniine (+)-mandelic acid in TE-671 cells and in the goat model. Agonist specificity was documented by the inhibition of fetal movement by coniine, a piperidine alkaloid, but not by nicotine, a pyridine alkaloid. These differences were most likely attributable to structure-activity relationships between the ligand and the receptor, because these alkaloids readily cross the placenta (Schep et al., 2009). Finally, we compared the actions of the coniine HCl enantiomers and the coniine mandelic acid combinations in the presence and absence of two CTxs, both of which are potent and selective muscle-type nAChR antagonists. In TE-671 cells at a concentration of 1 μM, both CTxs antagonized the actions of coniine. Taken together, these results provide evidence that the piperidine alkaloid coniine is acting at fetal muscle-type nAChR to inhibit fetal movement.
The authors thank Clint Stonecipher, Edward L. Knoppel, Scott Larsen, Isabelle McCollum, Kermit Price, Terrie Wierenga, and Rex Probst for technical assistance and Bryan Stegelmeier and Mark G. Collett for the critical reading of this article.
Participated in research design: Green, Lee, Welch, Pfister, Panter.
Conducted experiments: Green, Lee, Welch, Pfister, Panter.
Contributed new reagents or analytical tools: Lee.
Performed data analysis: Green, Pfister.
Wrote or contributed to the writing of the manuscript: Green, Lee, Welch, Pfister, Panter.
This work was supported by the US Department of Agriculture, Agricultural Research Service.
- multiple congenital contractures
- nicotinic acetylcholine receptor
- Received August 24, 2012.
- Accepted October 17, 2012.
- Copyright © 2013 by The American Society for Pharmacology and Experimental Therapeutics