Beyond Affinityfunctional Assays Efficacy And Allosteric Agonism

The drawbacks of screening for allosteric modulators using a radioligand binding assay are highlighted in Section 12.2. For these reasons, the most common method to identify such molecules is to run a functional assay to measure a signal transduction event downstream of receptor activation, for example, [35S]-GTPyS binding, cAMP accumulation, inositol phosphate accumulation, intracellular Ca2+ flux, or reporter gene activation. Among the earliest known examples of allosteric modulators of GPCRs were compounds such as gallamine, alcuronium, and brucine, which were shown to display saturable shifts in the potencies or affinities of muscarinic receptor ligands [13, 17, 18]. It was assumed that any effects on potency of agonists in a functional assay (akin to the graphs in Fig. 12.2b,c) simply reflected a change in agonist affinity according to the ATCM. Thus, such data were subjected to analysis according to the ATCM to generate estimates of modulator affinity (KB) and cooperativ-ity with respect to the orthosteric agonist (a).

Exemplifying such a situation is the effect of PD117975 on R-(-)-N6-(2 - phenylisopropyl) - adenosine (R - PIA) - mediated stimulation of extracellular regulated kinase (ERK)1/2 phosphorylation in Chinese hamster ovary (CHO) cells recombinantly expressing the adenosine A1 receptor [19]. PD117975 is a positive allosteric modulator, producing a saturable, leftward shift in the concentration-response curve to R-PIA; the data set is readily amenable to analysis according to the ATCM. However, there are now many examples of allosteric interactions which cannot be adequately explained using the ATCM. For example, the molecule SIB1893 increases both the potency and the maximal response to L-AP4 in a [35S]-GTPyS binding assay at mGluR4 [20], Such an effect on maximal agonist response cannot be accommodated by a model that considers only effects of allosteric ligands on orthosteric agonist affinity. Similarly, the negative allosteric modulator of mGluR1, CPCCOEt, reduces the maximal response to glutamate in an inositol phosphate accumulation assay in a concentration-dependent fashion without affecting equilibrium [3H]-glutamate binding to the receptor [15] , Neither of these observations is consistent with CPCCOEt modulating the affinity of glutamate to binding at mGluR1. It is clear from these data sets that these allosteric modulators must be capable of changing the efficiency of the agonist-receptor complex to produce a stimulus. SIB1893 clearly increases the efficacy of L-AP4 to stimulate [35S]-GTPyS binding at mGluR4, either alone or in addition to effects on L-AP4 affinity, while CPCCOEt reduces the efficacy of glutamate at mGluR1.

Furthermore, there is now increasing evidence that allosteric ligands are capable of altering the equilibrium between inactive and active GPCR in the absence of an orthosteric ligand. For example, the negative allosteric modulator at mGluR5, MPEP, reduces the constitutive activity of the receptor in a concentration-dependent manner, that is, MPEP acts as an inverse agonist [21]. Furthermore, several receptors have now been shown to interact with allosteric ligands, which display positive efficacy, for example, ASLW and RSVM at CXCR4 [22] , PD81723 at adenosine A1 [23] , and AMN082 at mGluR7 [24] , On a practical note, care should be taken to determine whether any apparent "allosteric agonism" is not simply a product of positive modulation of endogenous agonist present in the assay; this phenomenon is common for receptor subtypes for which the endogenous agonist is often found in cell culture media or produced by the cells themselves (e.g., glutamate, adenosine).

However, it is clear that allosteric ligands are capable of modulating affinity and/or efficacy of orthosteric ligands as well as altering the equilibrium between inactive and active receptor in their own right. Such a range of behaviors was predicted by Hall [25], who produced a thermodynamically complete allosteric two- state model that incorporates all of the behaviors described above. However, despite the completeness of the model, it has too many parameters for analysis of any practical data set. More recent efforts have sought to generate semiempirical models to describe the effects of allosteric modulator function. By combining the ATCM with the operational model of agonist action [ 26] , a model has been constructed, which encompasses the modulation of orthosteric agonist efficacy by an allosteric modulator, in addition to any effects on affinity [27] , The operational model of agonism effectively dissects agonist potency into its component parameters, affinity (KA)

Response = Basal +

([A]Kb + KaKb + Ka[B] + a[A][B])n + (ta[A](Kb + aß[B]) + Tb[B]Ka)"

Response = Basal +

(Em - Basal)(([A](Kß + aß[B])) + Tb ■ ECso)n (([A](Kb + aß[B])) + Tb[B] ■ ECso)n + (ECso- (Kß[B])n)

Figure 12.3 (a) Extension of the ATCM to produce an operational model of alloster-ism. This model accounts for the interaction of an orthosteric (A) and allosteric (B) ligand with a receptor (R), governed by equilibrium dissociation constants, KA and KB, respectively. It also incorporates receptor activation mediated by both orthosteric (ta) and allosteric (tb) ligands such that the stimulus can be produced by three species, AR, BR, and ARB. Allosteric modulation of affinity is governed by the affinity cooperativity factor, a; modulation of efficacy is governed by the empirical factor, P, which is the ratio of ta values in the absence and presence of allosteric ligand. Em, Basal, and n represent the maximal system response, the response in the absence of agonist and the transducer function slope, respectively. (b) A simplified version of the operational model in which it is assumed that the orthosteric agonist, A, is a full agonist (maximal response is equal to Em). The model yields a parameter, aP, which represents the net cooperative effect of the allosteric modulator on affinity and efficacy. EC50 represents the midpoint of the orthosteric agonist concentration-response curve.

and efficacy (t, tau). The term t is a cell or tissue-dependent expression of the ability of the agonist-receptor complex to produce a stimulus. In the new model described by Kenakin et al., the "efficacy cooperativity" parameter is neither thermodynamic nor bidirectional, but simply represents the ratio of orthosteric agonist efficacies (t and Pt) in the absence and presence of the modulator, respectively; the ratio (P) is the efficacy cooperativity factor. This model was later extended to include the ability of the allosteric ligand to mediate receptor activation in its own right (Fig. 12.3; [28]). Using this empirical efficacy cooperativity factor, these operational models of allosterism have been effectively used to discriminate the mechanism of action of several allosteric ligands in functional assays [29-31].

It is important to recognize a role for radioligand binding studies in defining the mechanism of action of allosteric modulators. As previously discussed, radioligand binding is limited to detection of effects of modulators on orthosteric ligand affinity . but this can be a useful attribute when seeking to dissect out effects on affinity and efficacy. For example, the muscarinic M-receptor positive allosteric modulator, LY2033298, causes a 35-fold left shift in the potency of acetylcholine to stimulate Ca2 + mobilization in CHO cells stably coexpressing the muscarinic M4 receptor and a promiscuous G protein -31] . Three-way radioligand binding studies using [-H]-NMS demonstrated that LY2033298 increased the affinity of acetylcholine by approximately 40fold, suggesting that the positive modulation elicited by this compound was due to changes in agonist affinity, rather than efficacy. Conversely, similar studies have shown that the effects of LY487379 as a positive allosteric modulator of mGluR2 are mediated by changes in glutamate efficacy, rather than affinity [30] .

In the context of determining the mechanism of action of an allosteric modulator, it is important to consider the influence of the biological reagent on the outcome. The most common paradigm for screening is to use a recombinant cell line expressing the receptor of interest. Historically, researchers have sought to develop cell lines that express high levels of receptor (often at much greater levels than would be found in a native environment), with an eye toward increasing the size of the signal-to-noise ratio in assays and potentially decreasing the amount of biological/chemical reagent required to run the assay. This approach works very well for radioligand binding assays and functional assays seeking orthosteric antagonist hits. However, as shown below, the choice of assay end point and receptor expression level can have a marked effect on the outcome when screening for allosteric ligands, particularly positive modulators. Figure 12.4a models the effect of varying concentrations of a positive modulator of an agonist response according to the operational model of allosterism (Fig. 12.3a; [28]) under three different assay conditions, represented by ta (orthosteric agonist efficacy) values of 1, 10, or 100. These different values could correspond to alternative assay formats, successively more distal to receptor activation (and hence subject to progressively greater signal amplification). Alternatively, they could represent the same assay under increasing levels of receptor expression.

Under conditions of either low receptor expression or in an assay paradigm proximal to receptor activation (t = 1), the positive modulator causes both a leftward shift and an increase in the maximal agonist response (Fig. 12.4a). As either the level of receptor expression or the signal amplification increases, the profile of both the agonist and positive modulator changes. At the highest t value (100), the agonist is significantly more potent than at the lower t values. Furthermore, the maximal agonist response is increased to the maximal level of the assay system ("Em" in the terms of the operational model). Under these conditions, the positive modulation of efficacy cannot manifest as an increase in the maximal agonist response and so contributes to a further increase in agonist potency. In the absence of prior knowledge, this profile may incorrectly be taken as evidence of affinity-only modulation. In the early stages

100 X 80

J 40 CD

65 20 0

100 X 80

J 40 CD

65 20 0


a 80

Modulator curve, T = 10 Agonist

Modulator curve, T = 100 Agonist

Figure 12.4 (a) The effect of a positive modulator of affinity and efficacy on agonist concentration-response curves under conditions of varying stimulus-response coupling (t values of 1, 10, and 100). Data are modeled using the operational model of allosterism (Fig. 12.3a) using the following parameters: pKA = 5.0, pKB = 6.0, a = 3, P = 3, tb = 0, Basal = 0, Em = 100, n = 1. (b) The resultant positive modulator titration curves that would be generated using an ECr0 of agonist. The variation in stimulus-response coupling has little or no effect on absolute modulator titration curve potency but becomes apparent when the modulator curves (solid lines) are examined with respect to their respective orthosteric agonist curves (broken lines).

Basal -8

of drug discovery programs, this may not be an issue as the assay is simply being used to compare the activities of different compounds. However, an understanding of the mechanism of action for the allosteric modulator may be important for late- s tage compounds that are being taken forward into efficacy models.

It is interesting to note that in isolation, positive modulator titration curves (see below for detailed description) from each of these models show very similar EC50 values (Fig. 12.4b). This suggests that in a structure-activity relationship (SAR) screening situation, the stimulus-response coupling of the assay (i.e., the "t" value for the orthosteric agonist) is not particularly important, as similar modulator titration curves will be produced irrespective of the efficiency of stimulus-response coupling. It is only when the modulator titration curves are examined in relation to their respective orthosteric agonist concentration-response curve (dashed lines in Fig. 12.4b) that it becomes obvious that there are differences.

0 0

Post a comment