AN ATTEMPT TO STUDY THE EFFECTS OF CHEMICAL STRUCTURE ON THE AFFINITY AND EFFICACY OF COMPOUNDS RELATED TO ACETYLCHOLINE

Signalling bias in new drug discovery: detection, quantification and therapeutic impact

Agonists of seven-transmembrane receptors, also known as G protein-coupled receptors (GPCRs), do not uniformly activate all cellular signalling pathways linked to a given seven-transmembrane receptor (a phenomenon termed ligand or agonist bias); this discovery has changed how high-throughput screens are designed and how lead compounds are optimized for therapeutic activity. The ability to experimentally detect ligand bias has necessitated the development of methods for quantifying agonist bias in a way that can be used to guide structure-activity studies and the selection of drug candidates. Here, we provide a viewpoint on which methods are appropriate for quantifying bias, based on knowledge of how cellular and intracellular signalling proteins control the conformation of seven-transmembrane receptors. We also discuss possible predictions of how biased molecules may perform in vivo, and what potential therapeutic advantages they may provide.

Antagonist inhibition curves and the measurement of dissociation constants

1. Experiments carried out on guinea-pig isolated ileum with carbachol as agonist and diphenyl-acetoxyethyl-dimethyl-ethyl-ammonium (DADMEA) bromide as antagonist gave results which fit the theoretical relation between fractional inhibition (Q) of the effects of an agonist ([A]) and the concentration of a competitive antagonist ([B]): this also involves the Hill coefficient (logistic slope factor, P) for the agonist concentration-response curve and the degree of agonist stimulation, [A]/[A]50, where [A]50 produces a half-maximum response. 2. Values of IC50 and an exponent, P', can be obtained by fitting Q to [B] using a logistic approximation to the relation. Both P' and IC50 should be greater with higher agonist stimulation but the increase in P' may be masked by errors in extreme values of Q. Estimates of IC50, however, invariably increased with higher agonist stimulation but with a steep concentration-response curve (P > 1) and low agonist stimulation ([A]/[A]50 < 1, IC50 can be less than KD. 3. KD was calculated from the results in three ways; (i) by a least-squares fit of Q to [B] using the values of P and [A]/[A]50 calculated from the control concentration-response curve; (ii) from the value of IC50 for each line and the values of P and [A]/[A]50 and (iii) by using the agonist concentration-response curve to calculate the dose-ratio and estimate of KD for each in the presence of the antagonist. The methods gave similar results (nM: 11 experiments), 12.4 +/- 1.1 (i), 11.7 +/- 0.9 (ii), 14.8 +/- 1.6 (iii) but there are advantages in using methods (i) or (ii) rather than (iii). 4. The method by which KD is calculated is less important than the experimental design: the plan used in this work, with alternative small and large responses from the tissue, is very suitable for estimating KD with low concentrations of antagonists and small dose-ratios. Although it is not a sensitive test for competitive behaviour because only a small range of concentrations of antagonist is tested, the estimate of affinity should be free from complications involved in the use of higher concentrations of antagonist (and agonist) and the nature of the antagonism can always be checked by doing further experiments in the presence of a known competitive antagonist.

The use of 4-diphenylacetoxy-N-(2-chloroethyl)-piperidine (4-DAMP mustard) for estimating the apparent affinities of some agonists acting at muscarinic receptors in guinea-pig ileum

1. 4-Diphenylacetoxy-N-(2-chloroethyl)-piperidine (4-DAMP mustard), which is known to block muscarinic M3 receptors in preference to muscarinic M2 receptors, was used to estimate the apparent affinity constants of some agonists acting at muscarinic receptors in guinea-pig ileum. Estimates for carbachol and n-pentyl-trimethyl ammonium iodide were similar to published values obtained in similar conditions: those for n-hexyl-trimethyl ammonium iodide were slightly lower. 2. The results for the agonists, n-pentyl- and n-hexyl-trimethyl ammonium iodides and for the partial agonist, n-heptyl-trimethyl ammonium iodide were not as regular as was suggested by Stephenson, though there is an overall increase in apparent affinity with chain length. 3. Estimates of apparent affinity may be affected by hexamethonium, usually present in experiments on ileum. Its absence had little effect on the results with carbachol but reduced the estimates obtained with n-pentyl trimethyl ammonium, which has strong nicotinic effects compared with its muscarinic effects. On ileum treated with tetrodotoxin the values for n-pentyl trimethyl ammonium were similar to those obtained in the presence of hexamethonium (0.28 mM): slightly higher estimates of affinity were obtained in the presence of indomethacin (2.8 microM). The nicotinic effects of n-pentyl ammonium may involve the release of prostaglandins. 4. The estimates of apparent affinity did not depend on the method used to calculate them as the 'null' method and the 'operational' method give similar answers. Estimates of the transducer-ratio for the partial agonist, n-heptyl-trimethyl ammonium iodide, were numerically the same as those of its efficacy. 5. This work illustrates the use of 4-DAMP mustard as a tool for measuring the apparent affinity of agonists acting at muscarinic M3 receptors.

Characterization of 5-HT3 and 'atypical' 5-HT receptors mediating guinea-pig ileal contractions in vitro

1. Neuronal 5-hydroxytryptamine (5-HT) receptors mediating contraction of guinea-pig ileal segments have been characterized in vitro by the use of methysergide to block 5-HT1-like and 5-HT2 receptors. Concentration-response curves to 5-HT were biphasic (first phase, defined as those responses occurring between 1 nM and 0.32 microM 5-HT, -log EC50 = 7.15 +/- 0.08; second phase, defined as these responses occurring between 0.32 microM and 32 microM 5-HT, -log EC50 = 5.32 +/- 0.03) but monophasic to 5-methoxytryptamine (-log EC50 = 7.0 +/- 0.08) and 2 methyl 5-HT (-log EC50 = 5.2 +/- 0.13). The maximal response of the first phase to 5-HT and the maximal response to 5-methoxytryptamine were 30 +/- 4% and 35 +/- 5% respectively of the maximum response to the second phase of the 5-HT concentration-effect curve (set at 100%). In contrast, the maximal response to 2-methyl-5-HT equalled that obtained with 5-HT (second phase). 2. The responses comprising the second phase of the concentration-effect curve to 5-HT were antagonized by 1 microM ICS 205-930, ondansetron, granisetron, quipazine, N-methyl-quipazine and (R,S)-zacopride and the following pKB values, with 5-HT as the agonist, were obtained at the 5-HT3 receptor: ICS 205-930 7.61 +/- 0.05, ondansetron 6.90 +/- 0.04, granisetron 7.90 +/- 0.04, (S)-zacopride 8.11 +/- 0.06, (R,S)-zacopride 7.64 +/- 0.11, and (R)-zacopride 7.27 +/- 0.06. 3. Under conditions of 5-HT1-like, 5-HT2 and 5-HT3 receptor blockade, the following rank order of agonism was observed: 5-HT > 5-methoxytryptamine = renzapride > (S)-zacopride > (R,S-zacopride > 5-carboxamidotryptamine > BRL 24682 > (R-zacopride > metoclopramide > 2-methyl-5-HT > sulpiride. 8-Dihydroxydiphenylaminotetralin (8-OHDPAT), GR 43175, N,N-dipropyl-5-carboxamidotryptamine, ondansetron, ICS 205-930, granisetron, quipazine and N-methyl-quipazine were inactive as agonists and antagonists. Relative to 5-HT, (R,S)-zacopride acted as a partial agonist (intrinsic activity, alpha = 0.80; -log EC50 = 6.3 + 0.12; -log KA = 6.1 + 0.03) as did (R)-zacopride (alpha = 0.4, -log EC,0 5.7 + 0.08, -log KA = 5.5 + 0.11). (S)-zacopride acted as a full agonist (-log EC,0 = 6.9 + 0.03). ICS 205-930 (3 microM) antagonized competitively responses to 5-HT, 5 methoxytryptamine, (RS)- and (S)- zacopride and 5-carboxamidotryptamine yielding -log KB estimates ranging from 6.1-6.5. 4. It is concluded that two different 5-HT receptors mediate excitatory neuronal responses in the guineapig ileum. 5-HT3 receptors mediate the second phase of the biphasic concentration-response curve, whereas a receptor with properties distinct from the 5-HT1-like, 5-HT2 and 5-HT3 subtypes mediates the initial phase of the concentration-response curve. This receptor, which exhibits a close similarity to the 5-HT4 subtype is: (1) stimulated by 5-methoxytryptamine but not 2-methyl-5-HT; (2) stimulated selectively by certain substituted benzamides; (3) recognizes the optical isomers of zacopride and (4) is blocked by relatively high concentrations ICS 205-930 (pKB = 6.0-6.5) but not ondansetron, granisetron, quipazine or N-methyl-quipazine.

Regional variation in the sensitivity of longitudinal smooth muscle to histamine at H1-receptors in guinea-pig ileum and colon

The sensitivity of the distal ileum, proximal colon, medial colon, and distal colon of the guinea-pig to histamine has been evaluated. The rank order of sensitivity was ileum greater than medial colon greater than proximal colon approximately equal to distal colon. The mean -logEC50 values at receptors in the ileum, medial, proximal, and distal colon were 6.74, 6.18, 5.79, and 5.72, respectively. The apparent dissociation constant for the interaction of histamine with receptors in the various regions was determined. The -log Kd values at receptors in the ileum, proximal colon, medial colon, and distal colon were 4.68, 4.65, 4.62, and 4.44, respectively. The mean apparent -log Kd values for the antagonism of histamine by mepyramine were 9.0, 9.0, 9.1, and 8.9 for receptors on the ileum, proximal, medial, and distal colon, respectively. The results of these experiments provide no evidence that histamine receptors in the colon are distinguishable from H1-receptors as characterized on the ileum. The differences in sensitivity to histamine in the various regions of the intestine may be due to differences in the density of H1-receptors.

The effect of p-chloromercuribenzoate on structure-binding relationships of muscarinic receptors in the rat cerebral cortex

Muscarinic receptors in the rat cerebral cortex, reacted with p-chloromercuribenzoate (PCMB) under different conditions (Phase I and II), have modified binding sites. These exhibit remarkable changes in the structural dependence of the binding of drugs. In Phase I, the structure-binding profile of agonists for both the high and low affinity agonist sites are altered. In Phase II, the structure-binding profile of antagonists is also observed. In Phase II, the ability of potent agonists to discriminate between sub-classes of agonist binding sites is eliminated. There is also a loss of heterogeneity in the binding of the selective antagonist pirenzepine. Of the 16 agonists examined, only pilocarpine has a heterogeneous binding profile in Phase II, the dispersity of binding being increased. The changes in binding properties of the receptors are discussed in terms of general theories of drug-receptor interactions.

Molecular connectivity study of muscarinic receptor affinity of acetylcholine antagonists

A correlation between three molecular connectivity indexes and the muscarinic receptor affinity of 104 acetylcholine antagonists was found. Analysis of structure from these indexes reveals not only the importance of the onium and the bulky portions of the molecule but also their virtual independence of each other on the affinity. Analysis of the onium group portion of the molecules indicates that its contribution to the experimental affinity is virtually constant through a variety of structural variations. The influence of the bulky side chains, in contrast, is quite structure dependent. The equation relating connectivity indexes to muscarinic affinity of antagonists is capable of predicting the affinity of other antagonists as well as a number of agonist molecules.

The effects of replacing ester by amide on the biological properties of compounds related to acetylcholine

1 Replacement of ester by amine in series of derivatives of diphenylacetic acid reduces the affinity for muscarine-sensitive acetylcholine receptors of the guinea-pig ileum from 40- to 100-fold. With similar series of phenylacetic acid the reduction is only 2- to 4-fold. In both series changes in the composition of the onium group produce similar changes in the affinity of amides and esters and it appears that the stiffness of the amide bond reduces the binding of the phenyl groups at the far end of the molecule from the onium atom. 2 Replacement of ester by amide in similar series of acetyl compounds reduces activity on the guinea-pig ileum over 1000-fold and on the frog rectus over 50-fold. Compounds with larger onium groups are antagonists on both preparations with log affinity constant around 3. The amides have similar affinity for electric eel acetylcholinesterase. 3 The amides are slightly bigger than the esters in solution and slightly more hydrophilic. 4 Replacement of ester by amide in acetylcholine reduces the proportion of gauche conformer about the C--C--bond from 100% to 39%. 5 The ability of acetylcholine to activate receptors is thought to depend on some degree of flexibility in the --CO--0--bond, though the hydration of the bond may also be important.

The specificity of some agonists and antagonists for nicotine-sensitive receptors in ganglia

1 The guinea-pig isolated ileum has been used to estimate the ability of substituted phenylalkylonium salts (related to nicotine) to stimulate or block receptors in ganglia. The effects of hexamethonium were used to indicate which were the most specific ganglion stimulants; these were tested on the blood-pressure of pithed rats and for neuromuscular blocking activity on the rat diaphragm preparation.2m-Hydroxyphenylpropyltrimethylammonium and 3,4-dihydroxyphenethyltrimethylammonium (coryneine, ;quaternary dopamine') were the most active and specific ganglion stimulants but their usefulness in vivo may be limited by their neuromuscular blocking activity. The analogous tertiary compounds are being investigated.3 The affinities of substances which were blocking agents at ganglionic receptors were measured on the isolated ileum with m-hydroxyphenylpropyltrimethylammonium as agonist. The affinities of selected compounds for postganglionic receptors were measured in experiments on the ileum in the presence of hexamethonium and with carbachol as agonist. Some of the compounds were tested for neuromuscular blocking activity on the rat diaphragm.4 Phenylbutyldiethylamine had ganglion-blocking activity greater than pempidine and little postganglionic blocking or neuromuscular blocking activity. Its triethylammonium analogue had higher ganglion-blocking activity but had appreciable neuromuscular blocking activity.5 The aromatic ring system is not essential either for activity or affinity and the effects of substituents are not related to their effects on electron distribution. Stimulant activity is enhanced only by hydroxyl or amino groups in suitable positions; it is not improved by the presence of rigid features (double or triple bonds or a cyclopropane ring) in the side chain. Affinity is slightly increased by chloro or bromo groups in suitable positions but the unsubstituted compounds are among those with the highest affinity. Substituents have similar effects on affinity for postganglionic receptors, though for these receptors the compounds mostly have only about one-tenth of their affinity for ganglionic receptors.

Relationships between chemical structure and affinity for postganglionic acetylcholine receptors of the guinea-pig ileum

1 Some phenylacetyl, diphenylacetyl, benziloyl and (+/-)-cyclohexylphenylglycolloyl esters have been made with 2- and 3-hydroxymethylpyrrolidines, 3-hydroxymethyl-N-methylpiperidine, piperidin-3-ols, piperidin-4-ols, 2,2,6,6-tetramethyl-N-methylpiperidin-4-ol, tropine, pseudotropine and quinuclidin-3-ol, and the affinity of these compounds and of their metho- and etho- derivatives has been measured for postganglionic acetylcholine receptors of the guinea-pig isolated ileum.2 Some of the compounds were very active indeed; the benziloyl esters of N-methylpiperidin-4-ol methiodide, tropine methiodide, and quinculidin-3-ol, and the (+/-)-cyclohexylphenylglycolloyl esters of N-methylpiperidin-4-ol and its methiodide had affinity constants greater than 10(10).3 The effects of inserting an additional methylene group onto the nitrogen were extremely variable, ranging from a decrease in log K of 1.64 units to an increase of 0.97 units. The effects of replacing hydrogen by phenyl in the acid portion ranged from an increase of 1.04 units to an increase of 3.06 units and of replacing hydrogen by hydroxyl from a decrease of 0.09 units to an increase of 1.94 units.4 The extent of the variation in the effects of a particular change in structure on affinity does not appear to be any different in these relatively rigid compounds from that observed with the same changes in open-chain aminoalcohols.5 Reasons for the variable effects of groups on affinity are discussed. If differences in effects on preferred conformations of these particular compounds in solution are of secondary importance, the effect of a group on affinity will be the net result of what it could contribute to binding, offset by the disturbance it causes to existing binding. The maximum effect observed in a large number of comparisons may indicate the contribution in the absence of disturbance and for groups containing only carbon and hydrogen it appears to be related to size, assessed from the increments in apparent molal volume at infinite dilution. The variation in the effects of these groups also appears to be related to size. Changes involving groups containing oxygen can produce bigger contributions to binding, and a bigger variation in contribution, than would be expected from their size.6 It is difficult to predict the extent to which groups may fail to produce their maximum effects. Variation is greatest with groups which could produce the biggest changes and so are of the greatest interest.7 The relevance of the results to the successful prediction of biological activity is discussed.

A comparison of phenylalkyl- and phenoxyalkyl- trimethylammonium and triethylammonium salts; their apparent molal volumes at infinite dilution and effects on the frog rectus and guinea-pig ileum preparations

1. The effects of phenoxyalkyltrimethylammonium and triethylammonium bromides on the frog rectus preparation and on the isolated guinea-pig ileum in the presence of hexamethonium have been compared with those of analogous phenylalkyltrimethylammonium and triethylammonium bromides. Affinity constants have been measured when possible.2. The apparent molal volumes at infinite dilution of some of the compounds have been measured from estimates of density made with an Anton Paar precision density meter.3. An ether oxygen occupies at infinite dilution in water only about one-third of the volume occupied by a methylene group.4. The replacement of methylene by ether oxygen reduces nicotine-like activity and affinity for nicotine-sensitive receptors. The reduction in affinity may be partly due to the decrease in size.5. The replacement of methylene by ether oxygen reduces activity and affinity at muscarine-sensitive receptors but in some compounds there is a bigger reduction in affinity than would be expected simply from the reduction in size.6. It is suggested that the affinity of these compounds largely depends on hydrophobic bonding and that effects on water structure in the environment of the receptor may also be involved in the actions of agonists.

A comparison of the affinities of antagonists for acetylcholine receptors in the ileum, bronchial muscle and iris of the guinea-pig

1. Isolated preparations of bronchial strip and of intact iris from the guineapig have been adapted for the measurement of affinity constants of substances which block post-ganglionic acetylcholine receptors.2. The affinity constants of 28 compounds on bronchial muscle and of 8 compounds on the iris have been compared with values measured on the guinea-pig ileum.3. Although the compounds differed up to a million-fold in affinity, most of the estimates of log affinity constant for the bronchial muscle and iris differed only slightly from those on the ileum.4. Some of the differences could be attributed to the actions of hexamethonium, used in the tests on the ileum but not, initially, in the tests with the bronchial strip and iris. Hexamethonium reduced most of the estimates of log K for the receptors in the bronchial strip by a variable but significant amount, which could be due, at least in part, to a weak post-ganglionic blocking (atropine-like) action. On average, hexamethonium had little effect on the estimates made with the ileum, appreciably decreasing estimates with some compounds and increasing those with others.5. The results indicate that, although there may be differences between the acetylcholine receptors in the three types of tissue, there is no conclusive evidence, because the differences in affinity which we have observed could have arisen from differences in the experimental conditions. This is illustrated by results obtained with the guinea-pig ileum recorded with the same technique as was used for the bronchial strip, which are presented as an appendix.6. Such differences as may exist between these three types of acetylcholine receptor are likely to be limited to the replacement of one aminoacid in the receptor protein by a homologue.

The significance of differences in the potency of enantiomers of anti-acetylcholine drugs

The anti-acetylcholine potencies of the dimethylaminoethyl and N/-methyl piperidin-4-yl esters of R and S 2-cyclohexyl-2-hydroxy-2-phenylacetic acid and their quaternary derivatives have been measured by in vitro and in vivo procedures. The R-enantiomer of dimethylaminoethyl-2-cyclohexyl-2-hydroxy-2-phenylacetate was approximately 100 times more active than the corresponding S-enantiomer both in vivo and in vitro. In contrast, the differences in potencies of the enantiomers of the other compounds were smaller in vivo than in vitro and moreover, the in vivo differences in potency decreased as the potency of the racemates increased. The relevance of these results to general studies of enantiomeric differences is discussed.

Structure-activity relations for anticholinergic 2(1-aryl(or cyclohexyl)-1-hydroxy-1-phenyl)methyl-1,3-dioxolans

The syntheses and configurational assignments of some 4-dimethyl aminomethyl-2[1-aryl (or cyclohexyl)-1 -hydroxy-1-phenyl]methyl-1,3-dioxolans are described. The anticholinergic potency of the 4-dimethylaminomethyl-2[l-cyclohexyl-l-hydroxy-l-phenyl]methyl-1,3-dioxolans, both in tertiary and quaternary form, depends principally on the configuration of the benzylic carbon atom, secondly on the C-2 configuration and thirdly, and to a much lesser extent, if at all, on the C-4 configuration. The dioxolans, which are derived formally from 4-dimethylaminomethyl-2-methyl-1,3-dioxolan methiodide (or its tertiary analogue) by replacement of the 2-methyl substituent by a 2[l-aryl (or cyclohexyl)-l-hydroxy-l-phenyl]methyl group and the glycollates which are derived formally from acetylcholine (or its tertiary analogue) by corresponding substitution of the acetoxymethyl group have closely similar anticholinergic potencies.

Structure-activity relations for anticholinergic dioxolans

4-Dimethylaminomethyl 1, 3-dioxolan derivatives have been examined for anticholinergic potency. Quaternary derivatives in which C-2 was substituted with two bulky substituents were found to have anticholinergic potencies similar to those of atropine in the peripheral nervous system. A comparison of the enantiomeric cis and trans 2-cyclohexyl-2-phenyl-4-dimethylaminomethyl-1, 3-dioxolan methiodides showed that the observed configurational selectivity depended mainly on the configuration at C-2 and not on the geometrical relation between C-2 and C-4.

Relationships between chemical structure and affinity for acetylcholine receptors

1. Series of analogues of acetylcholine have been prepared in which the acetyl group was replaced by phenylacetyl, cyclohexylacetyl, diphenylacetyl, dicyclohexylacetyl, (+/-)-phenylcyclohexylacetyl, benziloyl and (+/-)-phenylcyclohexylhydroxyacetyl groups and the trimethylammonium group was replaced by Me(2)EtN(+), MeEt(2)N(+), Et(3)N(+), [Formula: see text] Further series were prepared in which the acetoxyethyl group was replaced by ethoxyethyl, phenylethoxyethyl, cyclohexylethoxyethyl, diphenylethoxyethyl, and dicyclohexylethoxyethyl groups, and by n-pentyl, 5-phenylpentyl, 5-cyclohexylpentyl and 5:5-diphenylpentyl groups.2. The ethoxyethyl and n-pentyl series contain some compounds which are agonists or partial agonists when tested on the isolated guinea-pig ileum, but all the other compounds are antagonists.3. The affinity of the compounds for the postganglionic ("muscarinesensitive") acetylcholine receptors has been measured in conditions in which the antagonists have been shown to be acting competitively. There were considerable differences between their affinities, the most active (log K, 9.8) having one million times the affinity of the least active (log K, 3.7).4. The changes in affinity as the onium group was modified were not entirely independent of changes in the rest of the molecule. Increasing the size of the onium group, as judged from conductivity measurements on simpler onium salts, increased affinity in the series containing one large group (phenyl or cyclohexyl) but, in the series with two large groups, affinity declined when the size was increased beyond -(+)NMeEt(2).5. In general, the effects of changes in the rest of the molecule on affinity were bigger than the effects of changes in the onium group and there were bigger interactions. Affinity was increased to a greater extent by introducing one phenyl and one cyclohexyl group together than by introducing either two phenyl or two cyclohexyl groups; the increment was greater than the separate contributions made by one phenyl and one cyclohexyl group.6. The factors which influence the binding of molecules to receptors are discussed. There is no evidence that the separation between the onium group and the group in the receptor with which it interacts is greater in compounds with high affinity nor is there any evidence, from the study of the series which contain agonists and partial agonists, that ability to activate receptors depends upon the onium group being able to come close to this charged group in the receptors.