Abstract
Epibatidine is a potent analgetic agent with very high affinity for brain nicotinic acetylcholine receptors (nAChR). We determined the activity profiles of three epibatidine derivatives, RTI-36, RTI-76, and RTI-102, which have affinity for brain nAChR equivalent to that of epibatidine but reduced analgetic activity. RNAs coding for nAChR monomeric subunits and/or concatamers were injected into Xenopus oocytes to obtain receptors of defined subunit composition and stoichiometry. The epibatidine analogs produced protracted activation of high sensitivity (HS) α4- and α2-containing receptors with the stoichiometry of 2alpha:3beta subunits but not low sensitivity (LS) receptors with the reverse ratio of alpha and beta subunits. Although not strongly activated by the epibatidine analogs, LS α4- and α2-containing receptors were potently desensitized by the epibatidine analogs. In general, the responses of α4(2)β2(2)α5 and β3α4β2α6β2 receptors were similar to those of the HS α4β2 receptors. RTI-36, the analog closest in structure to epibatidine, was the most efficacious of the three compounds, also effectively activating α7 and α3β4 receptors, albeit with lower potency and less desensitizing effect. Although not the most efficacious agonist, RTI-76 was the most potent desensitizer of α4- and α2-containing receptors. RTI-102, a strong partial agonist for HS α4β2 receptors, was effectively an antagonist for LS α4β2 receptors. Our results highlight the importance of subunit stoichiometry and the presence or absence of specific accessory subunits for determining the activity of these drugs on brain nAChR, affecting the interpretation of in vivo studies since in most cases these structural details are not known.
SIGNIFICANCE STATEMENT Epibatidine and related compounds are potent ligands for the high-affinity nicotine receptors of the brain, which are therapeutic targets and mediators of nicotine addiction. Far from being a homogeneous population, these receptors are diverse in subunit composition and vary in subunit stoichiometry. We show the importance of these structural details for drug activity profiles, which present a challenge for the interpretation of in vivo experiments since conventional methods, such as in situ hybridization and immunohistochemistry, cannot illuminate these details.
Introduction
Nicotinic acetylcholine receptors (nAChR), first characterized at the neuromuscular junction (Papke, 2014), assemble as pentameric complexes of subunits and function as ligand-gated ion channels, activated by acetylcholine (ACh) or exogenous drugs like nicotine. In total, nine different α subunits (α2–α10) identified by the presence of a pair of vicinal cysteines, and three nonalpha subunits (β2–β4) have been found expressed in vertebrate neuronal tissues.
Functional heteromeric receptors form readily from the coexpression of α2, α3, or α4 with either β2 or β4 (Papke, 2014), with each α−β pair forming ACh binding sites with unique functional and pharmacological properties (Luetje and Patrick, 1991; Papke et al., 1989, 2010, 2013; Papke and Heinemann, 1991). The structurally required fifth subunit can be either an α or β subunit. Beginning with single-channel study of heterologously expressed neuronal nAChR, it was shown that the ratio of α to β subunits was important for determining receptor properties (Papke et al., 1989). Subsequent studies have confirmed the importance of the specific subunit composition (Nelson et al., 2003; Kuryatov et al., 2008; Jain et al., 2016; Lucero et al., 2016). Although not contributing to ACh binding sites, α5 and β3 subunits can be functionally important in receptors, taking the accessory subunit position, with β3 especially important for α6-containing receptors in dopaminergic neurons (Gerzanich et al., 1997, 1998; Kuryatov et al., 2000; Kuryatov and Lindstrom, 2011). Alpha3 subunits will form functional heteromeric receptors with either β2 or β4 subunits (Papke and Heinemann, 1991) and are of primary importance for nAChR function in the autonomic nervous system (David et al., 2010) and the adrenal gland (Sala et al., 2008), but in brain α3 expression is largely restricted to the medial habenula and the interpeduncular nucleus (Wada et al., 1989). The second major subtype of nAChR in brain is composed of homomeric assemblies of α7 subunits. Numerous functional and pharmacological properties distinguish homomeric α7 receptors from the heteromeric receptors in brain (Papke and Lindstrom, 2020).
Efforts to understand both the acute and addictive effects of nicotine focus on β2-containing receptors, especially those formed with α4 and, to a lesser degree, α6 subunits, especially in regard to addiction (Papke et al., 2020). Receptors with two α4 subunits and three β2 subunits (α4(2)β2(3)) are more sensitive to low concentrations of nicotine and more profoundly desensitized by high agonist concentrations than receptors with the reverse ratio (α4(3)β2(2)). Of potential importance to nicotine addictions, it has been observed, at least in vitro, that chronic nicotine favors the expression of the high sensitivity (HS) α4(2)β2(3) form of the receptor (Srinivasan et al., 2011).
The frog toxin epibatidine is a very high affinity ligand for brain nAChR. It binds to the heteromeric nAChR in brain with an affinity 20–50 times higher than nicotine (Anderson et al., 1995; Ondachi et al., 2014). It is an efficacious activator of some nAChR subtypes (Table 1). It has been shown to be an extremely potent analgetic agent (Badio and Daly, 1994), and it has been used extensively as a scaffold to generate numerous novel and probative receptor ligands (Carroll, 2009). In this paper, we report the activity profile of three such derivatives (Fig. 1) on eight different nAChR subtypes, in all but one case controlling the precise receptor subunit composition through the use of linked subunit concatamers (Zhou et al., 2003; Kuryatov and Lindstrom, 2011) (Fig. 2). These compounds have all been previously characterized for receptor binding and their ability to mimic the in vivo systemic effects of injected nicotine regarding analgesia, hypothermia, and spontaneous activity (Table 2). However, since the relationship between these activities and their effects on specific nAChR subtypes is unknown, to some degree these experiments might be considered exploratory. Only one of these compounds, RTI-102, has been previously studied with electrophysiological methods, and despite its in vivo activity it was described as an antagonist (Abdrakhmanova et al., 2006). Interestingly, another high-affinity ligand for α4β2 receptors, sazetidine-A (Xiao et al., 2006), was initially described as an exclusively desensitizing agent but later revealed to be a selective activator for the HS α4β2 subtype (Zwart et al., 2008). Based on the early in vivo data on these compounds and the similarity between the initial characterizations of sazetidine-A and RTI-102 as antagonists, we were interested to determine if these compounds would have a selectivity for HS α4β2 and α4β2 receptors and receptors containing α5 and β3 as structural subunits.
Previous studies of epibatidine human nAChR expressed in oocytes
Structures of epibatidine [2-(6-chloropyridin-3-yl)-7-azabicyclo[2.2.1]heptane] and test compounds RTI-36 (2′-fluorodeschloroepibatidine), RTI-76 [3′-(3″-dimethylaminophenyl)-epibatidine], and RTI-102 [2′-fluoro-3′-(4-nitrophenyl)deschloro-epibatidine].
nAChR subtypes studied formed from the coexpression of linked subunits and/or subunit monomers. To control the subunit composition of α4β2 and α2β2 receptors, we used concatamers of these subunits coexpressed with subunit monomers as indicated. The dimeric concatamers configure the α and β subunits so that the primary surface of the orthosteric agonist binding site (+) on the α subunit faces the complementary surface (−) on the β subunit. Functional receptors assemble effectively with two concatamers and the subunit coexpressed as a monomer taking the fifth position as the accessory subunit (Zhou et al., 2003). A common form of α6-containing receptors also incorporates α4, β2, and β3 subunits (Gotti et al., 2010). We used a pentameric concatamer (Kuryatov and Lindstrom, 2011) to generate receptors with this subunit composition. Monomer α3 and β4 subunits were coexpressed at equal ratios to most likely yield receptors with both α3(3)β4(2) and α3(2)β4(3) compositions. Functional α7 receptors are homomeric pentamers.
Published data on RTI compounds
Methods and Materials
Commercial Reagents.
ACh chloride, atropine, and other chemicals were purchased from Sigma-Aldrich Chemical Company (St. Louis, MO). Fresh ACh stock solutions were made in Ringer’s solution each day of experimentation.
RTI Compounds.
The synthesis and preliminary characterization of experimental compounds were previously described. RTI-36 (2′-fluorodeschloroepibatidine) was published as compound 3a in Carroll et al. (2005), RTI-76 [3′-(3″-dimethylaminophenyl)-epibatidine] was published as compound 5m in Carroll et al. (2010), and RTI-102 [2′-fluoro-3′-(4-nitrophenyl)deschloro-epibatidine] was published as as compound 5g in Carroll et al. (2010).
Heterologous Expression of nAChRs in Xenopus laevis Oocytes.
Human nAChR clones of monomeric subunits as well as the β2-6-α4 and β3α4β2α6β2 concatamers were obtained from Dr. J. Lindstrom (University of Pennsylvania, Philadelphia, PA). The β2-6-α2 concatamer was obtained from Edwin Johnson (Karolinska Institutet Sweden). The β2-6-α4 concatamer construction is described in Zhou et al. (2003), and the β2-6-α2 was made in similar fashion. Essentially, the N terminus is beta2 with its signal sequence, yet they show that the assembled dimer places the alpha subunit in the primary position. The human resistance-to-cholinesterase 3 clone, obtained from Dr. M. Treinin (Hebrew University, Jerusalem, Israel), was coinjected with α7 to improve the level and speed of α7 receptor expression without affecting the pharmacological properties of the receptors (Halevi et al., 2003). Subsequent to linearization and purification of the plasmid cDNAs, cRNAs were prepared using the mMessage mMachine in vitro RNA transfection kit (Ambion, Austin, TX). The combinations of concatamer and monomeric clones used to generate receptors with defined subunit composition are illustrated in Figure 2. Alpha7:Ric3 was injected 2:1, and concatamer:monomer constructs were injected 1:1.
Oocytes were surgically removed from mature X. laevis frogs (Nasco, Ft. Atkinson, WI) and injected with appropriate nAChR subunit cRNAs as described previously (Papke and Stokes, 2010). Frogs were maintained in the Animal Care Service facility of the University of Florida, and all procedures were approved by the University of Florida Institutional Animal Care and Use Committee. In brief, the frog was first anesthetized for 15–20 minutes in 1.5 l frog tank water containing 1 g of 3-aminobenzoate methanesulfonate buffered with sodium bicarbonate. The harvested oocytes were treated with 1.25 mg/ml collagenase (Worthington Biochemicals, Freehold, NJ) for 2 hours at room temperature in calcium-free Barth’s solution (88 mM NaCl, 1 mM KCl, 2.38 mM NaHCO3, 0.82 mM MgSO4, 15 mM HEPES, and 12 mg/l tetracycline, pH 7.6) to remove the follicular layer. Stage V oocytes were subsequently isolated and injected with 50 nl of 5–20 ng nAChR subunit cRNA. Recordings were carried out 1–7 days after injection, when receptors were expressing well.
Two-Electrode Voltage Clamp Electrophysiology.
Experiments were conducted at room temperature (24°C) using OpusXpress 6000A (Molecular Devices, Union City, CA) (Papke and Stokes, 2010). Both the voltage and current electrodes were filled with 3 M KCl. Oocytes were voltage-clamped at −60 mV. The oocytes were bath-perfused with Ringer’s solution (115 mM NaCl, 2.5 mM KCl, 1.8 mM CaCl2, 10 mM HEPES, and 1 μM atropine, pH 7.2) at 2 ml/min (α7) or at 4 ml/min (heteromeric). Drug applications were 12 seconds in duration followed by a 181-second washout period (α7) or 6 seconds in duration followed by a 241-second washout period (heteromeric). A typical recording for each set of oocytes constituted two initial control applications of ACh, the application of an experimental compound, and then follow-up control applications of ACh to determine whether there was desensitization of subsequent ACh-evoked responses. The responses were calculated as both peak current amplitudes and net charge, as previously described (Papke and Porter Papke, 2002). The average holding current for the 30 seconds prior to drug applications was used for baseline adjustment of drug-evoked responses. The calculation of net charge encompassed a standard period of 120 seconds after the initiation of drug application. The average responses of the two initial ACh controls from each cell were used for normalization. For some concentration-response studies, increasing concentrations were tested on the same set of oocytes, but only if the ACh control responses between drug applications were stable (varying less than 25%). Because applications of these compounds at high concentrations produced a degree of residual inhibition/desensitization, in many cases experiments were conducted with single drug applications to each set of cells, along with ACh controls. Data were initially normalized to ACh control responses from the same cells. For the receptor subtypes studied, the ACh control concentrations used were previously determined to give stable responses with repeated applications. They were 10 µM ACh for α4(2)β2(3), α4(2)β2(2)α5, α4β2α6β2β3, and α2(2)β2(3); 30 µM ACh for β3α4β2α6β2; 60 µM ACh for α7; and 100 µM ACh for α4(3)β2(2), α2(3)β2(2), and α3β4. For the determination of efficacy relative to ACh, responses normalized to ACh controls were adjusted by the ratio of the ACh control responses to the ACh maxima previously determined (Papke and Porter Papke, 2002; Stokes and Papke, 2012; Papke et al., 2013) for the respective cell types.
Note that a large sample of the concentration-response data sets (15 drug/receptor combinations, with peak, net charge, and recovery data, 513 distributions total) were tested to determine if the data fit the model for normal distributions. Overall, based on the Shapiro-Wilk test conducted in Prizm (Version 8.4.2; GraphPad Software, San Diego, CA), 87.3% of the distributions passed the normality test (alpha = 0.05).
Data were collected at 50 Hz, filtered at 20 Hz (α7) or at 5 Hz (heteromeric), and analyzed by Clampfit 9.2 or 10.3 (Molecular Devices) and Excel (Microsoft, Redmond, WA). Every experiment began with eight cells (the capacity of the recording system); however, due to the nature of the experiments, not all cells remained viable through entire experiments, and some cells had large responses that could not be adequately voltage clamped. Therefore, n varied from 5 to 8 (averaging 7.3). Experiments were discarded if n fell below 5. Data are expressed as means ± S.D. from at least five oocytes for each experiment (see figure legends or Supplemental Data for the n values of each experiment) and plotted by Kaleidagraph 4.5.2 (Abelbeck Software, Reading, PA). Kaleidagraph 4.5.2 was also used to fit concentration-response functions to the Hill equation:
The values for the curve fits were generated using the Levenberg-Marquardt algorithm to obtain the best chi-square fit to the Hill equation using the Kaleidagraph 4.5.2 plotting program. The errors in the tables are the calculated S.E.s of the fit parameters based on the goodness of fit. The data plotted in the figures are the average response (±S.D.), and the curve fits in the figures are fits to those averages. As an alternative approach that would allow the fitting procedure to respond to the variability of the data at each concentration, we also plotted the concentration-response data with every replicate (Supplemental Figs. 1–16) and generated curve fits to those data. Supplemental Table 1 includes the chi-square and R values for those fits. Note that the curve fit parameters were essentially identical for the two approaches (Supplemental Table 2), except for a few cases with the off-target α7 and α3β4 receptors where the replicate data could not be adequately fit to the Hill equation.
To avoid the potential bias that might come from the selection of “representative” raw data, in some figures we display multicell averages of the raw data for comparisons of responses. The averages of normalized raw data were calculated using an Excel (Microsoft) template for each of the 10,500 points in each of the 210-second traces (acquired at 50 Hz). After subtraction of the basal holding current, data from each cell, including the ACh controls, were normalized by dividing each point by the peak of the ACh control from the same cell. The normalized data were then averaged and S.E.M. for the multicell averages calculated on a point-by-point basis. The dark lines represent the average normalized currents and the shaded areas the range of the S.E.M. Scale bars in the figures of averaged traces reflect the scaling factor relative to the average peak current amplitude of the ACh controls used for the normalization procedures. These plots are effectively augmented versions of typical bar plots of peak currents that additionally illustrate the differences in net charge, the kinetics of the responses, and the variability throughout the entire time course of the responses.
Statistical analyses of pairwise data sets in Figure 3C were conducted based on two-tailed t test comparisons of the normalized net-charge data. Data in other figures may be taken to illuminate the differing activity of these compounds revealed by these exploratory studies.
Averaged raw data traces normalized to the control responses to ACh (see Methods and Materials). The S.E.M.s of the averaged normalized responses are represented by the tan colored areas. (A, B) Averaged responses (n = 7) of cells expressing LS [α4(3)β2(2)] (A) and HS [α4(2)β2(3)] (B) receptors (n = 7) to 1 µM RTI-36 compared with ACh controls. (C) Kinetic comparison of the responses in (A and B). (D) Responses of cells expressing LS [α4(3)β2(2)] receptors (n = 6) to 10 µM RTI-102 compared with ACh controls obtained prior to and after the application of RTI-102. (E) Responses of cells expressing HS [α4(2)β2(3)] receptors (n = 6) to 10 µM RTI-102 compared with ACh controls obtained prior to and after the application of RTI-102. Note that currents had not returned to baseline at the time of the follow-up ACh application, indicating persistent activation after the 6-second application of RTI-102.
Results
Effects of α4:β2 Subunit Stoichiometry on Voltage-Clamp Currents.
By coexpressing the linked β2−6−α4 subunits with either α4 or β2 monomers (Zhou et al., 2003) we obtained α4β2 receptors with either α4(3)β(2) or α4(2)β2(3) composition (Fig. 2). Averaged raw data traces are shown in Figure 3. Prior to averaging, each cell’s response was normalized to the peak current of the ACh control responses obtained from the same cells. Note that the α4(3)β(2) and α4(2)β2(3) receptors have previously been identified as having either low sensitivity (LS) or HS to agonists, respectively (Zhou et al., 2003), and so the ACh control concentrations were 100 and 10 µM for the α4(3)β(2) and α4(2)β2(3) receptors, respectively. One striking difference was that the responses of the HS receptors to the RTI epibatidine derivatives were protracted well beyond the period of drug washout compared with the LS receptor responses. This effect is shown for the responses to 1 µM RTI-36 (Fig. 3, A–C). These differences are reflected in the comparisons of peak currents and net charge over the 120-second intervals after the beginning of the RTI-36 applications (see Methods and Materials). For the LS α4(3)β(2) responses, compared with the ACh controls, responses calculated as net charge were only 67% as large as the peak currents, whereas for the HS α4(2)β(3) responses, the normalized net charge values were 267% compared with the peak currents. To systematically study these differences, we conducted concentration-response analyses on both peak currents and net charge measurements in subsequent figures and Table 3.
Curve fit values from plots of the averaged data (see Supplemental Data for plots and curve fits of data with all replicates)
Errors estimated are based on the goodness of fit.
The α4β2 subunit composition also had a large effect on the efficacy of RTI-102 in particular. As shown, the LS α4(3)β(2) receptors were virtually unresponsive to 10 µM RTI-102 (Fig. 3D), whereas the HS α4(2)β(3) receptors gave large and prolonged responses (Fig. 3E).
In addition to measuring the evoked responses, we measured the desensitization of subsequent ACh-evoked responses (calculated as “recovery” in the concentration-response figures). Note that although 10 µM RTI-102 did not activate substantial currents in LS α4(3)β(2) receptors, it did strongly inhibit the response evoked by 100 µM ACh when it was applied 280 seconds after 10 µM RTI-102 (Fig. 3D). Note also that 10 µM ACh produced only a comparably small response when applied to the HS α4(2)β(3) receptors after 10 µM RTI-102; however, this was on top of an elevated baseline due to the protracted response to the 10 µM RTI-102. Such elevated baselines were observed with several receptor subtype/drug combinations and are discussed in later sections. In the case of the HS α4(2)β(3) receptors, all three of the test compounds gave protracted responses of varying duration (Fig. 4), with the RTI-36 responses showing the slowest decay.
Responses of α4(2)β2(3) receptors to 10 µM applications of the test compounds. Control responses to 10 µM ACh were obtained from cells expressing α4(2)β2(3) receptors followed by 6-second applications of either RTI-36 (n = 5), RTI-76 (n = 5), or RTI-102 (n = 6) and then two follow-up applications of Ringer’s solution from the drug application system, basically switching from bulk flow to acute focused 6-second delivery of the same solution. All three compounds evoked responses that failed to return to baseline after 12 minutes of washout. The switch from bulk flow to pipette delivery produced small perturbations in the persisting currents, which may represent changes in the dynamics of drug unbinding and rebinding during sustained responses, perhaps suggesting that the oocyte membrane itself functions as a reservoir for residual drug (Papke et al., 1997).
The complete concentration-response studies for all three compounds on both LS and HS α4β2 receptors are shown in Figure 5, and the curve fit values are given in Table 3. All three compounds were relatively efficacious for HS α4(2)β(3) receptors (RTI-36 ≈ RTI-102 > RTI-76), and for all three compounds the net charge Imax was greater than the Imax for peak currents due to the protracted nature of the responses as illustrated in Figure 3. The only one of the compounds with high efficacy for LS α4(3)β(2) receptors was RTI-36. Note that, consistent with Figure 3, the average net-charge and peak-current values for RTI-36 were similar across the entire range of concentrations. However, the RTI-36 net-charge data were not well fit by the Hill equation, as evidenced by the large error estimate of the EC50.
Concentration-response studies of α4β2 receptors. (A) Activation of HS α4(2)β2(3) receptors by varying concentrations of the epibatidine analogs. Responses were calculated as both peak currents (open circles) and net charge (filled circles) integrated over 120-second periods starting with the drug applications. All data were normalized to the initial 10 µM ACh controls obtained from the same cells. Each point is the average of five to eight cells (±S.D.). Note that 10 µM ACh is the peak current EC90 for this receptor subtype as determined in previous experiments (Papke et al., 2013). Therefore, to display the data relative to ACh maximum, values normalized to ACh controls were multiplied by 0.90. (B) Responses evoked by 10 µM ACh 4.5 minutes after the application of the epibatidine analogs at the indicated concentrations. Note that the ACh responses were calculated relative to the baseline averaged for 30 seconds prior to the ACh application. After the application of the epibatidine analogs at high concentrations, these baselines were significantly elevated compared with the baseline currents prior to drug applications (see Fig. 3). (C) Activation of LS α4(3)β2(2) receptors by varying concentrations of the epibatidine analogs. Responses were calculated as both peak currents (open circles) and net charge (filled circles) integrated over 120-second periods starting with the drug applications. All data were normalized to the initial 100 µM ACh controls obtained from the same cells. Each point is the average of five to eight cells (±S.D.). Note that 100 µM ACh is the peak current EC50 for this receptor subtype as determined in previous experiments (Papke et al., 2013). Therefore, to display the data relative to ACh maximum, values normalized to ACh controls were multiplied by 0.50. (D) Responses evoked by 100 µM ACh 4.5 minutes after the application of the epibatidine analogs at the indicated concentrations. For curve fit values see Table 3. See Supplemental Data for the n values of each point and alternative plots and fits of the data utilizing the replicate measurements at each concentration.
All of the compounds were effective at reducing subsequent ACh control responses (Fig. 5), although as noted above, for the HS α4(2)β(3) receptors the subsequent ACh response rode on top of the sustained responses that had not returned to the original baselines (see Fig. 3). Interestingly, the least effective activator, RTI-76, was the most potent desensitizer.
Responses of α2β2 Receptors to the Epibatidine Analogs.
Although α4-containing receptors are the most abundant high-affinity nAChR in rodent brain, in primates there are additional high-affinity receptors containing α2 subunits (Han et al., 2000, 2003). By using a β2−6−α2 concatamer similar to the β2−6−α4 used to generate the HS and LS α4β2 subtypes, α2β2 receptors with specific subunit composition were also generated (Fig. 2). We have previously shown that these α2(3)β(2) and α2(2)β2(3) receptors have pharmacological profiles similar to the α4β2 counterparts in regard to ACh, nicotine, and the HS subtype-selective agonist TC-2559 (Papke et al., 2013).
In most regards, the HS and LS α2β2 receptors had similar responses to the epibatidine derivatives as the HS and LS α4β2 subtypes (Fig. 6; Table 3). Only RTI-36 had much efficacy for activating the LS α2β2 receptors, and all three produced protracted responses in the HS α2β2 receptors, as evidenced by the increased net charge compared with peak current Imax values and the consistent increase in baselines after application of the compounds at high concentrations. The most striking difference in the profiles was reduced efficacy of RTI-102 for the HS α2β2 receptors compared with the HS α4β2 receptors. The IC50 values (Table 3) were overall higher for the α2β2 receptors than for the α4β2 receptors, but again RTI-76 was the most potent desensitizer of the test compounds.
Concentration-response studies of α2β2 receptors. (A) Activation of HS α2(2)β2(3) receptors by varying concentrations of the epibatidine analogs. Responses were calculated as both peak currents (open circles) and net charge (filled circles) integrated over 120-second periods starting with the drug applications. All data were normalized to the initial 10 µM ACh controls obtained from the same cells. Each point is the average of five to eight cells (±S.D.). Note that 10 µM ACh is the peak current EC82 for this receptor subtype as determined in previous experiments (Papke et al., 2013). Therefore, to display the data relative to ACh maximum, values normalized to ACh controls were multiplied by 0.82. (B) Responses evoked by 10 µM ACh 4.5 minutes after the application of the epibatidine analogs at the indicated concentrations. Note that the ACh responses were calculated relative to the baseline averaged for 30 seconds prior to the ACh application. After the application of the epibatidine analogs at high concentrations, these baselines were significantly elevated compared with the baseline currents prior to drug applications (see Fig. 3). (C) Activation of LS α2(3)β2(2) receptors by varying concentrations of the epibatidine analogs. Responses were calculated as both peak currents (open circles) and net charge (filled circles) integrated over 120-second periods starting with the drug applications. All data were normalized to the initial 100 µM ACh controls obtained from the same cells. Each point is the average of five to eight cells (±S.D.). Note that 100 µM ACh is the peak current EC34 for this receptor subtype as determined in previous experiments (Papke et al., 2013). Therefore, to display the data relative to ACh maximum, values normalized to ACh controls were multiplied by 0.34. (D) Responses evoked by 100 µM ACh 4.5 minutes after the application of the epibatidine analogs at the indicated concentrations. For curve fit values see Table 3. See Supplemental Data for the n values of each point and alternative plots and fits of the data utilizing the replicate measurements at each concentration.
Responses of Alternative α4β2-Containing Receptors to the Epibatidine Analogs.
Among the nAChR subtypes associated with nicotine self-administration are subtypes containing either α5 (Grady et al., 2010; Jackson et al., 2010; Picciotto and Kenny, 2013) or α6 subunits (Jackson et al., 2009; Liu et al., 2012; Sanjakdar et al., 2015). Although these subunits may incorporate into multiple receptor subtypes, at least some important forms also contain α4 and β2 (Kuryatov et al., 2008; Drenan et al., 2010; Kuryatov and Lindstrom, 2011; Sala et al., 2013). The coexpression of α5 with the β2−6−α4 (Fig. 2) yields a high-sensitivity receptor in some ways similar to α4(2)β2(3) receptors (Zhou et al., 2003; Papke et al., 2013). Although α6-containing receptors are acknowledged as an important target for understanding nicotine’s effects in the brain, they were not an easy receptor to get to function in a heterologous system prior to the development of a pentameric β3α4β2α6β2 concatamer (Kuryatov and Lindstrom, 2011). This construct incorporates β3 as the structural subunit and has ACh binding sites at the α4β2 and α6β2 interfaces (Fig. 2).
Of the three test compounds, RTI-36 was the most efficacious agonist for both α4(2)β2(2)α5 and β3α4β2α6β2 receptors, with net-charge responses having much higher Imax values than the peak-current responses, and baseline shifts consistent with protracted responses like those of the HS α4β2 and HS α2β2 receptors (Fig. 7; Table 3). Although a relatively weak agonist, RTI-76 was a potent desensitizer for both of these receptors.
Concentration-response studies of other α4β2-containing receptors. (A) Activation of α4β2α5 receptors by varying concentration of the epibatidine analogs. Responses were calculated as both peak currents (open circles) and net charge (filled circles) integrated over 120-second periods starting with the drug applications. All data were normalized to the initial 10 µM ACh controls obtained from the same cells. Each point is the average of five to eight cells (±S.D.). Note that 10 µM ACh is the peak current EC69 for this receptor subtype as determined in previous experiments (Papke et al., 2013). Therefore, to display the data relative to ACh maximum, values normalized to ACh controls were multiplied by 0.69. (B) Responses evoked by 10 µM ACh 4.5 minutes after the application of the epibatidine analogs at the indicated concentrations. Note that the ACh responses were calculated relative to the baseline averaged for 30 seconds prior to the ACh application. After the application of the epibatidine analogs at high concentrations, in some cases these baselines were significantly elevated compared with the baseline currents prior to drug applications. (C) Activation of β3α4β2α6β2 receptors by varying concentration of the epibatidine analogs. Responses were calculated as both peak currents (open circles) and net charge (filled circles) integrated over 120-second periods starting with the drug applications. All data were normalized to the initial 30 µM ACh controls obtained from the same cells. Each point is the average of five to eight cells (±S.D.). Note that 30 µM ACh is the peak current EC76 for this receptor subtype as determined in previous experiments (Papke et al., 2013). Therefore, to display the data relative to ACh maximum, values normalized to ACh controls were multiplied by 0.76. (D) Responses evoked by 30 µM ACh 4.5 minutes after the application of the epibatidine analogs at the indicated concentrations. For curve fit values see Table 3. See Supplemental Data for the n values of each point and alternative plots and fits of the data utilizing the replicate measurements at each concentration.
The averaged raw data responses of β3α4β2α6β2 receptors to 10 µM RTI-36 are shown in Fig. 8. Prior to averaging, each single cell’s response was normalized the control 30 µM ACh responses from the same cell. At the time when the follow-up ACh application was made, 4.5 minutes after the 6-second application of RTI-36, the steady-state baseline current was on average 740 nA, 16% the amplitude of the initial ACh control. The estimated net charge during the post–RTI-36 control period (the rightmost trace in Fig. 7) was 106% ± 7% the net charge of the initial ACh control responses.
Protracted currents and baselines with β3α4β2α6β2 receptors stimulated by 100 µM RTI-36. Averaged raw data traces (n = 8), normalized to the control responses to ACh shown (see Methods and Materials). The S.E.M.s of the averaged normalized responses are represented by the tan colored areas. Note that currents had not returned to baseline at time of the second ACh application, indicating persistent activation after the 6-second application of RTI-36.
A summary of all the receptors that showed significant increases in baseline currents prior to the post control ACh applications is shown in Figure 9. Data represent the current 4.5 minutes after drug application normalized to the peak currents of the initial ACh controls. For all three drugs the most sensitive receptors were the HS α4β2 and HS α2β2 subtypes. Smaller shifts were observed for α4(2)β2(2)α5 and β3α4β2α6β2 receptors, although for RTI-36, shifts for these receptors were quite substantial (see Fig. 7).
Baseline shifts with test compounds and sensitive receptor subtypes. (A) Persistent currents stimulated by RTI-36, measured as baseline increases averaged over 30-second periods beginning 4 minutes after the 6-second application of RTI-36 at the indicated concentrations. Baseline increases were calculated relative to the peak current amplitudes of the initial ACh controls from the same cells. (B) Persistent currents stimulated by RTI-76, measured as baseline increases averaged over 30-second periods beginning 4 minutes after the 6-second application of RTI-76 at the indicated concentrations. Baseline increases were calculated relative to the peak current amplitudes of the initial ACh controls from the same cells. (C) Persistent currents stimulated by RTI-102, measured as baseline increases averaged over 30-second periods beginning 4 minutes after the 6-second application of RTI-102 at the indicated concentrations. Baseline increases were calculated relative to the peak current amplitudes of the initial ACh controls from the same cells. (A–C) All points are the average of five to eight cells (±S.E.M.).
Responses of Alternative nAChR Lacking α4 and β2 Subunits to the Epibatidine Analogs.
Homomeric α7 receptors, the second major type of nAChR in brain, which do not bind nicotine or ACh with high affinity, have many features that distinguish them from heteromeric nAChR (Papke and Lindstrom, 2020). The unique rapid concentration-dependent desensitization of α7 receptors makes the measurement of peak currents an almost meaningless measure of the concentration dependence of receptor function, since high concentrations of agonist stimulate peak currents prior to complete application of the drug solutions. This limitation is largely overcome by relying on net charge as a measure of α7 responses (Papke and Porter Papke, 2002; Papke, 2006, 2010, 2014). As with the β2-containing receptors, we saw that RTI-36 was the most efficacious of the three epibatidine derivatives tested and the only one that might be classified as a full agonist (Fig. 10A; Table 3). The potency of RTI-36 for α7 receptors was also relatively high, ranking between the potency for HS and LS subtypes of the α4- and α2-containing receptors. However, the α7 responses to RTI-36 were not protracted, and RTI-36 was not a potent desensitizer. The IC50 was 90-fold higher than the EC50 for net charge (Table 3). Note that the differing curve fit values for α7 peak currents and net charge were as expected for this receptor and represent the artifact associated with α7 desensitization mentioned above (Papke and Porter Papke, 2002). We observed partial agonist activity for RTI-76 and to a lesser degree RTI-102, although they were not very potent and produced relatively little desensitization of the postapplication ACh controls.
Concentration-response studies of α7 and α3β4 receptors. (A) Activation of α3β4 receptors by varying concentration of the epibatidine analogs. Responses were calculated as both peak currents (open circles) and net charge (filled circles) integrated over 120-second periods starting with the drug applications. All data were normalized to the initial 100 µM ACh controls obtained from the same cells. Each point is the average of five to eight cells (±S.D.). Note that 100 µM ACh is the peak current EC39 for this receptor subtype as determined in previous experiments (Papke et al., 2013). Therefore, to display the data relative to ACh maximum, values normalized to ACh controls were multiplied by 0.39. (B) Responses evoked by 100 µM ACh 4.5 minutes after the application of the epibatidine analogs at the indicated concentrations. (C) Activation of α7 receptors by varying concentration of the epibatidine analogs. Responses were calculated as both peak currents (open circles) and net charge (filled circles) integrated over 120-second periods starting with the drug applications. All data were normalized to the initial 60 µM ACh controls obtained from the same cells. Each point is the average of five to eight cells (±S.D.). Note that 60 µM ACh is the peak current EC80 for this receptor subtype as determined in previous experiments (Papke et al., 2013). Therefore, to display the data relative to ACh maximum, values normalized to ACh controls were multiplied by 0.80 on this scale. (D) Responses evoked by 60 µM ACh 4.5 minutes after the application of the epibatidine analogs at the indicated concentrations. For curve fit values see Table 3. See Supplemental Data for the n values of each point and alternative plots and fits of the data utilizing the replicate measurements at each concentration.
As noted in the introduction, the α3 subunit has a relatively restricted pattern of expression in the brain but is essential for synaptic function in the autonomic nervous system, where it can coassemble with β4 subunits. Since all nAChR subunits vary greatly in their intracellular domain sequences (Stokes et al., 2015), the most cogent basis for sequence comparisons of functional domains evaluates just the extracellular and transmembrane portions of the receptors, which are responsible for ligand binding and ion conduction, respectively. In these domains α4 and α2 share 82.98% sequence identity. In contrast, α4 and α3 have only 69.44% sequence identity in these domains. Consistent with its activity on other receptors, we found RTI-36 to be the most efficacious analog of the three tested on α3β4 (Fig. 10). The Imax for RTI-36 peak currents was 9.4 ± 0.8 times the peak currents of the ACh controls. As the ACh control concentration was determined in previous experiments to be the EC39, the estimated Imax for RTI-36 peak currents would be 3.65-fold larger than the ACh Imax. RTI-76 and RTI-102 both effectively activated α3β4, but only in the case of RTI-36 were currents protracted, and only RTI-76 was an effective desensitizer of these receptors.
Discussion
Epibatidine has been an important inspiration to nicotinic drug development and a valuable tool for the characterization of nicotinic receptor binding sites (Houghtling et al., 1995; Carroll, 2009). Like epibatidine, the three analogs used in the present study show high affinity for heteromeric nAChR and very low affinity for α7-type receptors (Table 2). However, they were found to have significantly less activity than epibatidine in measurements of acute analgetic effects. The retention of high-affinity binding, along with reduced activity in at least some functional assays, could be consistent with the conversion of epibatidine from an agonist to a high-affinity antagonist due to the structural differences. However, our current appreciation of the fact that the binding sites present in crude preparations of brain membranes represent a wide variety nAChR subtypes encouraged us to determine the activity profiles of these compounds on a range of structurally defined receptor subunits.
Previously, RT1-102 was characterized as an antagonist using cells expressing α4β2 nAChR (Abdrakhmanova et al., 2006). Our data indicate that this would be consistent with a preferential expression of the LS α4(3)β2(2) subtype in those cells, since RT1-102 was an efficacious agonist for the alternative HS α4(2)β2(3) receptor subtype. RTI-102 then joins sazetidine-A and TC-2559 (Moroni et al., 2006) as an HS α4(2)β2(3) selective agonist.
Our observations regarding the crucial importance of the precise subunit composition on the activity of these compounds for heteromeric α4-containing (and α2-containing) receptors highlights the importance of better identifying the detailed features of brain receptors in vivo. In animals (or people) that have not been chronically exposed to nicotine, are the α4 and α2 receptors primarily in the LS 3α:2β configuration, or are they mixtures of LS and HS subtypes? Do the ratios vary based on neuronal subtypes or locations in the brain? In vitro studies have shown that outside the brain, with tissue-cultured nonneuronal cells (Srinivasan et al., 2011) and Xenopus oocytes (Kuryatov et al., 2005; Zwart et al., 2006), nicotine can act as a molecular chaperone, selectively increasing the surface expression of HS receptors. It is an attractive but unproven hypothesis that this mechanism in part provides the basis for the upregulation of nAChR function in smokers and may relate to nicotine dependence. Although there is evidence that nAChR on striatal synaptosomes consist of both HS and LS types (Marks et al., 2010), it is not clear that this was due strictly to populations with different subunit stoichiometry, rather than different subunit composition (Grady et al., 2010). Furthermore, the composition of the mixed receptor subtypes did not seem significantly different in animals that had received chronic nicotine exposure.
The measurements of macroscopic currents from large populations of receptors provide only limited insights into the underlying molecular processes, and the fact that we are making our observations on time scales of seconds to minutes rather than the millisecond time scale of single-channel transitions is a further limitation. Even with our control ACh responses, we know that it takes several seconds for the full concentration of agonist to be delivered, and what we record is a process in which channels are both activating and desensitizing at the same time. Our peak currents represent the point at which this balance between activation and desensitization is further perturbed as the agonist begins to be washed out of the bath (Papke, 2010). However, in the present studies, the data suggest that ligand-receptor interactions outlast the presence of free ligand in solution, supporting the hypothesis that the ligands stay bound to the receptors long enough to account for the persistent currents of the HS receptors as well as the desensitization of LS receptors to subsequent ACh applications.
A very minimal model for nAChR activation and desensitization is shown in Figure 11. The model assumes that agonist (A) binding promotes conformational changes in the receptor (R) associated with the channel activated state (AR*) or conversion to a threshold desensitized state (AD1) and a stable desensitized state (AD2) observed in the earliest study of nAChR desensitization (Katz and Thesleff, 1957). The rates and conformational equilibria between these states will necessarily depend on the properties of the specific ligands and receptors. Single channels recorded at high (saturating) concentrations of ACh reveal multiple levels of desensitization associated with full occupancy of the agonist binding sites (Sine and Steinbach, 1987; Colquhoun and Ogden, 1988), and D states may be associated with both liganded and unliganded receptors. This model assumes a certain amount of equilibration among the states, which is appropriate given the slow rate of solution exchange in our experimental system (Papke, 2010) or when drugs are delivered in vivo. However, in vivo the delivery of the natural agonist ACh may be very rapid and transient. With rapid application of ACh, α4β2 receptors initially show a very high probability of opening (>80%) but very rapidly begin to equilibrate with desensitized states (Li and Steinbach, 2010).
nAChR activation and desensitization. Shown on top is a simplified scheme for agonist (A) activation and desensitization of nAChR. In this reduced model (Papke and Lindstrom, 2020) the binding of a single agonist (AR state) promotes conformation change to either the open channel form (AR*) or a desensitized state (AD1) that can, with a certain probability, revert back to the activatible AR state or a more stable desensitized (AD2) state. The D states are known to bind agonists with higher affinity than the resting closed (R) states of the receptors. Once agonist dissociates from receptors in the AD2 state (D2), the receptors retain that high affinity and therefore may rebind agonist and return to the AD2 state, or if the agonist concentration is low for a long enough period of time, the receptors may revert back to the low-affinity R state. Shown in the middle, for the LS α4(3)β2(2)–RTI-102 drug-receptor combination, there is little activation from the AR to the AR* state, and equilibrium favors the D2 state over the D1 state. Shown at the bottom is a possible way this model might explain the persistent currents of HS α4(2)β2(3) receptors stimulated by RTI-102 (see Fig. 4). Relatively increased rate constants are represented by thicker arrows. For this drug/receptor combination, there may be rapid conversion back and forth between AR and AD1 states and relatively slow conversion from AD1 to AD2 states.
Consider the contrasting responses of LS α4β2 and HS α4β2 to 10 µM RTI-102 shown in Figure 3, C and D, respectively. RTI-102 is ineffective at activating this receptor, so that the rate for conversion from AR to AR* is very low (as represented by line thinness) especially as compared with the rate from AR to AD1. The rates between AD1 and AD2 favor receptors in AD2, accounting from the decreased response to subsequent applications of ACh. An alternative interpretation, that LS α4β2 receptors simply remain in the AR state with RTI-102 bound, is not likely since the AR state is a low-affinity state, which is not consistent with the prolonged effect of RTI-102 applications.
For RTI-102 and HS α4β2 receptors, the activation rate from AR to AR* is relatively high, and receptors readily return to the AR state from the D1 state and only slowly convert to the D2 state. Having intermittent bouts of nAChR activation under conditions when receptors are predominantly desensitized has been referred to as “smoldering” (Campling et al., 2013). In the case of these HS receptor responses to the epibatidine analogs, the condition is perhaps closer to “wildfire.” The fact that current is sustained after drug washout for HS receptors but not for LS, combined with the observation that the LS receptors are reduced in responses to subsequent ACh applications, indicates that the drugs do stay bound to the orthosteric sites of both receptors but that the equilibrium between activation and desensitization is different.
In considering why the HS α4β2 and α2β2 receptors, and to a lesser degree the α5- and α6-containing receptors, manifest these persistent currents, it is tempting to speculate that it may be as much about the presence of the putative low-affinity α−α binding site on the LS forms of the receptors as about specific effects of the alternative accessory subunits. Perhaps binding of the analogs to the α−α site somehow puts the brakes on channel activation or acts to stabilize receptors in the D2 state.
These compounds were initially found to have relatively low efficacy in models of acute analgesia, hypolocomotion, and hypothermia (Carroll et al., 2004, 2005, 2010) (Table 2). Subsequent studies have found them all to exert agonist activity (Rodriguez et al., 2014), with rank order apparent efficacy RT-36 > RTI-76 > RTI-102 in a model of nicotine’s subjective effects and potential abuse-related effects. It may be that the different apparent efficacies exhibited by these epibatidine analogs in mice reflect the relative contribution of the LS α4β2 and HS α4β2 forms of nAChRs to the effects of each compound. Some of these compounds, RTI-36 in particular, have a relatively long duration of in vivo action, and it remains to be determined whether nAChR channel kinetics, in addition to drug metabolism, may be responsible for these prolonged actions. Moreover, it remains to be determined what the activity of these and other epibatidine analogs in the RTI series will be in other assays of nAChR function, including preclinical assessments of nicotine addiction and dependence. The activity profiles we report here may serve to shed light on the molecular basis for those behaviors, especially when considered in the context of other published studies (Table 4) that have provided basic associations between receptor subtypes and specific behavioral and physiologic effects. Our work, however, highlights the need for finer analyses of receptor subtypes that take into account not only receptor subunits but also subunit stoichiometry.
Major proposed roles of nAChR subtypes in mediating some well known behavioral and physiologic effects of nicotine in preclinical studies
Acknowledgments
We thank Marina Picciotto (Yale University) for helpful comments and Fernando Valenzuela, (Department of Neurosciences, University of New Mexico, Albuquerque, NM) for conducting additional statistical tests.
Authorship Contributions
Participated in research design: Corrie, Stokes, Wilkerson, McMahon, Papke.
Conducted experiments: Corrie, Stokes.
Contributed new reagents or analytic tools: Carroll.
Performed data analysis: Corrie, Stokes, Papke.
Wrote or contributed to the writing of the manuscript: Stokes, McMahon, Papke.
Footnotes
- Received April 10, 2020.
- Accepted July 2, 2020.
This work was support by the National Institutes of Health National Institute of General Medical Sciences [Grant GM57481] and National Institute on Drug Abuse [Grants DA47855 and DA48353].
↵
This article has supplemental material available at molpharm.aspetjournals.org.
Abbreviations
- ACh
- acetylcholine
- HS
- high sensitivity
- LS
- low sensitivity
- nAChR
- nicotinic acetylcholine receptor
- RTI-36
- 2-(6-chloropyridin-3-yl)-7-azabicyclo[2.2.1]heptane
- (epibatidine)
- 2′-fluorodeschloroepibatidine
- RTI-76
- 3′-(3″-dimethylaminophenyl)-epibatidine
- RTI-102
- 2′-fluoro-3′-(4-nitrophenyl)deschloro-epibatidine
- Copyright © 2020 by The American Society for Pharmacology and Experimental Therapeutics
