The Impact Of Functional Assay Format On Allosteric Modulator Screening

In the sections above, functional assays have been spoken about as a homogenous group of assays in which a whole range of allosteric effects can be detected. It is true that all of the functional assays described can detect allosteric modulation of agonist affinity, efficacy, and allosteric agonism. However, the specific assay format chosen can greatly impact the results that are obtained. This section will discuss the issues associated with the use of transient read systems (such as intracellular Ca2+ flux) versus accumulation assay formats (exemplified by [35S]-GTPyS binding) to screen for positive allosteric modulators of GPCRs.

Measurement of intracellular Ca2+ flux is a very popular screening platform due to the relatively inexpensive cost of reagents and its applicability to most GPCRs via either native coupling (for Gq/11-coupled receptors) or coupling through chimeric or promiscuous G proteins [33]. A screen in the modulator titration curve format for a positive allosteric modulator may be run as follows: In brief, cells would be exposed to varying concentrations of test compound while being monitored for 1-2 min for fluorescence levels (the "agonist" read). This read would detect any agonist activity intrinsic to the compound. After an incubation period (typically between 2 and 30 min), the cells would be exposed to a fixed EC20 concentration of orthosteric agonist and monitored for fluorescence again (the "modulator" read). This read will detect any allosteric modulatory activity of the compounds, as evidenced by an increase in the activity of the EC20 concentration relative to control. This "dual-read" protocol is common for screening for positive allosteric modulators.

Contrast this method with a [35S]-GTPyS assay in which the agonist, modulator, and receptor (in the form of a membrane preparation) are all allowed to pre-equilibrate for up to 1 h prior to the start of the reaction with a radiolabeled, nonhydrolyzable form of GTP ([35S]-GTPyS). This incubation is allowed to continue for generally anywhere between 30 min and 2h before being terminated, and the amount of bound radioligand is determined using scintillation spectroscopy. Both of these formats will give the user information about the modulatory and agonist activity of a test compound. However, whereas the full range of activity of the compound will be apparent in the single output of the [35S]-GTPyS assay, in the calcium assay, any agonist activity of the compound will be highlighted only in the first read, and any modulatory activity will be apparent only in the second read.

Already this has made the holistic determination of a compound's activity more difficult as there are two separate data sets that need to be considered. More importantly, it leads to an important question—as the two activities have now been separated into two reads, does the first read impact on the output of the second? Figure 12.7 addresses this question with a data set from a GPCR positive allosteric modulator program screened in a Ca2+ flux assay on the FLIPR platform (Molecular Devices, Sunnyvale, CA) using a similar protocol to the one described above. Figure 12.7a-c shows the data for 10 positive allosteric modulators in the first "agonist" read, triaged according to their level of agonist activity. Figure 12.7a highlights two compounds with little or no agonist activity, Fig. 12.7b shows compounds with partial agonist activity, and Fig. 12.7c shows compounds with full agonist activity.

Figure 12.7d then shows the effects of these compounds in the second, "modulator," read. Compound 11 (one of the compounds with no agonist activity in the first read) produces a concentration-dependent leftward shift in the agonist concentration-response curve, with a small increase in the maximal agonist response, consistent with positive allosteric modulation. Figure 12.7e shows the effect of a different modulator on an agonist response. Compound 10, which displayed a full agonist response in the first read, also

Low/no agonism

Compound 7 Compound 11

Partial agonism

Compound 1 Compound 4 Compound 5 Compound 9

Compound 11 - no agonism

160 140

Compound 11 - no agonism

40 jaM

Control (Compound 11)

160 nM

40 jaM

160 140

Control (Compound 11)

Full agonism o)100-

Compound 1 Compound 4 Compound 5 Compound 9

Full agonism

Compound 2 Compound 6 Compound 8 Compound 10

Compound 2 Compound 6 Compound 8 Compound 10

Compound 10 - full agonism

Compound 10 - full agonism

40 jaM

Control (Compound 10)

160 nM

40 jaM

Control (Compound 10)

Compound 11

120 10080 60 40 20

Compound 4

Compound 10

Compound 4

Compound 10

-9 -8 -7 -6 -5 -4 -10 -9-8-7-6-5-4 -10 -9 -8 -7 -6 -5 -4 Log [modulator] (M) Log [modulator] (M) Log [modulator] (M)

Figure 12.7 Effect of 10 positive allosteric modulators on intracellular calcium mobilization measured on a FLIPR in a CHO cell line stably expressing a GPCR and a promiscuous Ga16 protein in the absence of orthosteric agonist. Modulator responses are shown, triaged according to their level of agonist response in the first read: (a) low/ no response, (b) partial agonist, and (c) full agonist. The effect of a range of concentrations of Compound 11 (d) and Compound 10 (e) on agonist-stimulated calcium mobilization as determined using FLIPR technology. The results are shown from the second, "modulator," read as described in the text. Both compounds produce leftward shifts and increases in the maximal agonist response, consistent with positive allosteric modulation. However, at the highest modulator concentrations (which produced a response in the first, "agonist" read), Compound 10 ablates the orthosteric agonist response. Resultant modulator titration curves are shown for Compound 11 (f), Compound 4 (g), and Compound 10 (h).

160 nM

160 nM

p100

st 80

causes a leftward shift in the agonist concentration-response curve and a small increase in the maximal agonist response. However, at the highest modulator concentrations, the maximal agonist response is markedly reduced or even abolished. Closer inspection reveals that it is the agonist responses in the presence of the concentrations of modulator which evoked a significant agonist response in the first read that are affected (Fig. 12.7c,e). This provides clear evidence that activity in the first read has negatively impacted the second read. This effect could be due to a number of reasons, including receptor desensiti-zation, but it is most likely due to the depletion of intracellular calcium stores during the "agonist" read; the stores simply do not have sufficient time to recover prior to the second challenge. This is even more clearly shown when the results of the second read are visualized in a modulator titration curve format. Figure 12.7f-h shows the modulator titration curves (based on a fixed agonist concentration of 4 |M) for Compounds 11, 4, and 10, which display no, partial, and full agonism in the first read, respectively.

Compound 11 potentiates the response to 4 |M agonist in a concentration-dependent fashion (Fig. 12.7f). Compounds 4 and 10 also potentiate the agonist response at low modulator concentrations, but at higher modulator concentrations, the response is progressively diminished, yielding bell-shaped concen-tration-.esponse curves (Fig. 12.7g,h). For these curves, it is impossible to accurately quantify their modulatory effect. If the agonist activity did not adversely affect the second read response, what would the effect have been at modulator concentrations greater than 1 |M? Would the modulator titration curve have reached a plateau or continued to increase? Unfortunately, it is impossible to say, and such uncertainty adversely affects the assessment of these compounds in an SAR program. Furthermore, on a practical level, it leaves a question of how best to analyze the modulator titration curves. Should points be excluded or should the experimenter try to fit a biphasic concentration-response curve? Both are likely to require manual intervention, and neither are attractive options. Unless an experimenter can argue that a transient response system more faithfully reflects the physiological signaling of the receptor under test, then the obvious answer is to avoid transient read assay formats and use accumulation assay formats, such as p5S]-GTPyS binding or cAMP. However, as long as platforms such as Ca)+ flux or aequorin remain popular, these questions will remain.

It is also important to highlight the pleiotropic nature of allosteric effects depending on the pathway monitored. The phenomenon of "biased-signalling," "agonist-trafficking," or "functional-selectivity" is well described in the literature [34], and allosteric modulators have been shown to exhibit such effects. For example, the allosteric modulator of chemoattractant receptor-homologous molecule expressed on Th2 cells (CRTH2), 1-(4-ethoxyphenyl)-5-methoxy-2-methylindole-3-carboxylic acid (Compound 1), displays neutral cooperativity with respect to the agonist prostaglandin-D2 (PDG2) in assays of G protein activation (i.e., it does not affect the agonist concentration-response curve in assays of PDG2-stimulated [35S]-GTPyS binding). However,

Compound 1 is a potent negative allosteric modulator of PDG25 stimulated recruitment of P-arrestin, a G protein-independent functional end point [35]. Unless the disease-relevant signal transduction pathway for a given receptor is known, then there is no way to know which is the most appropriate end point to measure; rather, the assay decision is generally driven by other factors such as cost, assay robustness, and availability of platforms and reagents.

Differential signaling through divergent effector systems is worth considering when using chimeric or promiscuous G proteins. As discussed above, this is a common practice as it allows non-Gq/11-coupled GPCRs to couple to the PLC-IP3-Ca2+ pathway for detection of intracellular calcium using platforms such as FLIPR or aequorin. However, this introduces a non-native coupling system, which may be subject to differential regulation. For example, both the orthosteric agonist, L-AP4, and the allosteric agonist, AMN082, inhibit for-skolin-stimulated cAMP accumulation in CHO cells stably expressing rat mGluR7 and the promiscuous G protein, Ga15 [36] . In contrast, only L-AP4 stimulates Ca2+ mobilization in the same cell line, whereas AMN082 is without effect [36] . If a HTS had been run using the Ga15 calcium assay alone, then AMN082 would have not been identified. In a similar vein, several studies have recently shown that allosteric agonists may not cause GPCR desensitiza-tion in the same manner as orthosteric agonists [22, 37]. This array of novel pharmacologies must be considered when selecting a screening assay—for example, the recruitment of P-arrestin is now often used as screening paradigm for GPCRs [38]. Clearly, it is not going to be possible to cover all bases by running multiple assay formats for screening compounds against a single target, but researchers should be aware of these caveats when developing assays for the routine assessment of their compounds.

0 0

Post a comment