Receptor Binding Assays

The biological identification and pharmacological characterization of receptors have been aided immensely by the advent of radiolabeled agents that retained their biological activity after iodination or tritiation, allowing the development of a method known as the radioligand binding or receptor binding technique. Due to the relative easy way with which radioligand binding assays can be performed, they are still widely used in both academic and industrial laboratories. The quality of data obtained allows the determination of drug affinity, allosteric interactions, the existence of receptor subtypes, and estimates of the receptor numbers. However, the ease of the method does not imply that rigorous criteria are not required for studying binding to functional relevant receptors. These criteria and some practical guidelines are given in the present section while a more detailed information on radioligand binding methodologies, including theoretical aspects, can be found in a variety of other sources [7-9].

The most frequently used assay based on the radioligand binding technique is the membrane filtration receptor assay [9] . A variation of the membrane assay is the intact cell radioligand binding assay which has some advantages in special circumstances such as in studies of receptor internalization [9] . To analyze individual receptor pharmacological properties in the setting of intact tissues, tissue segment binding assays have been also proposed [ 10] . Other additional approaches that are useful in studying receptors in relation to their physiological functions and diseases are receptor autoradiography and in vivo labeling of receptors with positron emission tomography (PET).

The basics of radioligand binding assays with membrane preparations are rather simple: membranes containing the receptor of interest are incubated with a suitable radioligand for an appropriate period of time and then receptor-bound radioactivity is measured. The experiments can be performed according to three major types: saturation, kinetic, and inhibition. A saturation experiment is carried out by holding constant the receptor amount and varying the radioligand concentration, thus allowing to generate a saturation curve. By definition, the radioligand binding must be saturable because there is a finite number of receptors. From this type of experiment, both the receptor density (Bmax) and dissociation constant for the radioligand (KD) can be estimated. If the amount of radioligand and receptor is maintained constant and the time varied, then kinetic data are obtained from which association (kon or k+1) and dissociation (koff or k.) rate constants can be estimated. If the amount of a competing nonradioactive drug included in the incubation is varied while both radioligand and receptor are held constant, then the inhibition constant (Ki ) of the drug for the receptor labeled by the radioligand can be calculated.

Membrane fraction is prepared by standard procedures including homog-enization of tissue or cells in a hypotonic buffer followed by differential cen-trifugation. The need of more than two centrifugation steps on washing purpose is required to remove unwanted substances as endogenous ligands and guanine nucleotides. Details of buffer composition can be found in many papers and together with laboratory experience aid in the choice of the appropriate procedure to prepare the membrane fraction.

Generally for a given receptor, there are now several radiolabeled agonists and antagonists that are commercially available. Some characteristics of the radioligand to be considered include: radioisotope fH or [ 25I); the extent of nonspecific binding; the selectivity and affinity of the radioligand for the receptor; whether the radioligand is an agonist or an antagonist. Tritiated ligands have the advantage that are chemically unaltered and are thus biologically indistinguishable from the unlabeled compound and possess a long half-l ife (12 years). On the other hand, iodinated radioligands have short half. [ives (60 days), but they possess the advantage of having a high specific activity

(~2,000 Ci/mmol). In addition, iodinated radioligands do not need scintillation fluid to count radioactivity, thus eliminating purchasing and disposal costs.

The use of selective radioligands for a given receptor or receptor subtype is required for labeling them and their pharmacological characterization. Usually, high-affinity radioligands (KD, 10pM-10nM) are preferable since lower concentrations need to be used in the assay, and this results in lower levels of nonspecific binding. In order to be useful in binding assays, high-affinity radioligands must also possess high specific activity. Thus, iodinated radioligands have KD values for binding to the receptor in the pmolar range while tritiated radioligands have KD values in the 0.1-10 nM range. When needed, it is always possible to reduce the specific activity by diluting the 125I-radioligand with the unlabeled ligand in order to limit the amount of radioactivity in each assay tube (<106cpm). Another important issue to be considered is whether we are using an agonist or antagonist radioligand to selectively label the GPCR under investigation. Agonist radioligands preferentially label the portion of the total receptor population which is in the high-affinity state, namely those GPCRs tightly coupled to heterotimeric G proteins with no guanine nucleotide bound to the Ga subunit. Antagonist radioligands generally label all available GPCRs and this is certainly the case of neutral antagonists.

Considering the binding assay conditions, the choice of the buffer type, the addition of cations or protease inhibitors depends on both the radioligand and the GPCR under investigation. Generally, the pH should be in the physiological range (pH 7-8) while tris(hydroxymethyl)amino-methane (Tris) buffer is often used as buffer for binding reactions. Some GPCRs such as the A2a adenosine receptor [11] have the absolute requirement of Mg2+ ions for agonist binding. The addition to the incubation buffer of GTP or a nonhydrolyzable analog such as GTPyS decreases the affinity of agonists for the GPCR and is able to convert a biphasic or shallow inhibition curve of an unlabeled agonist for radiolabeled antagonist binding into a monophasic inhibition curve (Fig. 9.1). Protease inhibitors may be added to the incubation buffer to prevent degradation of peptide radioligands.

Since for saturation and inhibition experiments it is required to work at equilibrium, the incubation time needs to be sufficient to ensure reaction equilibrium or at least steady state. The time to reach steady state is dependent on the radioligand concentration, but at radioligand concentrations near to their KD value, most of them reach steady state at room temperature within 20-60 min. The demonstration that the amount of specific binding is constant over a period of time indicates that a steady state has been reached. Indeed, most assays are performed at steady state rather than at equilibrium.

The radioligand concentration in the assay is dependent on the type of experiment being performed. In kinetic experiments, the concentration should be reasonably low but high enough to obtain a reasonable level of specific binding. In inhibition assays, the general rule is to use the KD concentration or lower. In saturation experiments, the range of radioligand concentrations

Figure 9.1 Competition curve of a radiolabeled neutral antagonist bound to a GPCR by an unlabeled agonist. In the absence of GTP, the curve is clearly biphasic and better fitted by a two-site competion model of the nonlinear regression analysis (GraphPad Prism Version 4.0). The addition of 100 |M GTP in the reaction mixture causes a right shift of the curve, which is now monophasic, and it is better fitted by one-site model. This is a typical example of GPCR two affinity states for the agonist ligand. The agonist has higher affinity for GPCRs tightly coupled to G proteins than for those receptors uncoupled to G proteins. On the other hand, the neutral antagonist binds with the same affinity to coupled and uncoupled receptors.

Figure 9.1 Competition curve of a radiolabeled neutral antagonist bound to a GPCR by an unlabeled agonist. In the absence of GTP, the curve is clearly biphasic and better fitted by a two-site competion model of the nonlinear regression analysis (GraphPad Prism Version 4.0). The addition of 100 |M GTP in the reaction mixture causes a right shift of the curve, which is now monophasic, and it is better fitted by one-site model. This is a typical example of GPCR two affinity states for the agonist ligand. The agonist has higher affinity for GPCRs tightly coupled to G proteins than for those receptors uncoupled to G proteins. On the other hand, the neutral antagonist binds with the same affinity to coupled and uncoupled receptors.

should be from approximately 0.1 x to 10 x KD value. Of course, this is in ideal conditions that are not always possible, but some alternative options are available to study binding saturability, such as homologous competition experiments.

Usually, the higher the membrane protein concentration, the better is the binding. This is especially true for membranes prepared from tissues and untransfected cells. In this case, a membrane protein concentration range of 100-500 |g/mL is adequate. For transfected cells overexpressing the receptor, the protein concentration required in the assay is much lower. Increasing receptor concentration causes an increase of the ratio between specific binding to nonspecific binding. The amount of specific binding should be linearly related to the membrane protein concentration. However, the rule is that less than 10% of the added radioligand must be bound.

In inhibition experiments, it is important to consider the concentration range of the inhibiting or competing drug. When the inhibition of radioligand binding by a drug follows a single site model, then about 10 inhibitor concentrations spanning the range of at least 100-fold on both sides of the 50% inhibitory concentration (IC5 0) are adequate. In the case of multiple binding sites or affinity states, at least 20 concentrations of the inhibitor over a larger range are required.

In any radioligand binding assay, the most important consideration is the definition of specific binding. By definition, specific binding is the binding to the receptor of interest while nonspecific binding is any other binding. Nonspecific binding is measured in the presence of an appropriate excess of an unlabeled drug (e.g., 10033old its IC30 value) to completely block the receptor of interest. Nonspecific binding includes radioligand binding to other receptor sites, to glass fiber filters, adsorption, and dissolution in the membrane lipids. Specific binding is calculated as the difference between total and nonspecific binding. Nonspecific binding reaches steady state more rapidly than specific binding, but it does not saturate in the same ways as total and specific binding do. To determine nonspecific binding, it is better to use a drug that is chemically dissimilar from the radioligand. For an adequate binding assay, specific binding must be at least 50% of total binding, but the best conditions are at 70-90% of total binding. Peptide radioligands frequently give problems of high nonspecific binding. In this case, some strategies can be adopted to reduce nonspecific binding [ 9], but unfortunately, they are not always successful.

A crucial step in receptor binding assays is the separation of bound radio-ligand from free radioligand. During separation, it is important to prevent dissociation of receptor radioligand complex. This is obtained by reducing the temperature and performing separation as rapidly as possible. On this purpose, in membrane binding assays, the most frequently used technique is vacuum filtration through glass fiber filters which retain membranes. Membranes and filters are also washed with large amount of cold buffer which reduce nonspecific binding. In addition, to decrease nonspecific binding, filters can be pre-soaked with a 0.1% aqueous solution of polyethylenimine or bovine serum albumin (0.2% BSA). Filter presoaking with a BSA solution is used in the case of peptide or protein radioligands. Filtration under reduced pressure can be used only when the radioligand KD for binding to the receptor is lower than 10nM. In the case of KD values in the 10nM-1|M range bound from free radioligand is separated by centrifugation which is also able to reduce nonspecific binding. An aliquot of the radioligand as that added to each assay tube should be also counted.

Analysis of data derived from radioligand binding experiments is crucial for studying receptor pharmacology and identification of new ligands. The best way to fit saturation data and thus derive KD and Bmax values is to use nonlinear regression analyses such as that provided by GraphPad Prism (GraphPad Software Inc., San Diego, CA) or a variety of other software packages. Data are visualized as hyperbolic (bound vs. free radioligand concentration) or sigmoidal (bound vs. logarithm of free radioligand concentration) curves. The GraphPad Prism software also possesses excellent graphic capabilities which aid in data visualization. It is frequently desirable to present data transformed into a linear form, such as bound/free vs. bound or Rosenthal plot (also known as Scatchard plot) [12] . In this case, the intercept on the x-axis gives the Bmax value while KD corresponds to the negative reciprocal of the slope. To identify the presence of two binding sites or affinity states, nonlinear regression analysis of bound vs. free radioligand or eventually the Rosenthal plot is preferable over the semilogarithmic plot that can mask the complexity of the data. However, a nonlinear regression analysis of saturation data is always the best choice since linear transformation distorts experimental errors.

Data from inhibition experiments are also better analyzed using nonlinear regression techniques. Data are visualized as sigmoidal inhibition curves of bound radioligand expressed as percent of maximum vs. the logarithm of the inhibitor concentrations. The concentration of the inhibitor that reduces bound radioligand by half is the IC50 (inhibitory concentration 50%) or EC50 (effective concentration 50%) and can be estimated by the inhibition curve. The GraphPad Prism software fits inhibition curves using a nonlinear regression analysis, directly calculates Log EC-0 , and derives K; using the equation of Cheng and Prusoff - 13] . If the radioligand and the inhibitor compete for a single class of binding sites, the curve is really sigmoidal descending from 90% to 10% bound over an 81-fold increase in inhibitor concentration and the slope factor (also called Hill slope) is -1. If the inhibition binding curve is shallow with the slope factor less than -1, we may consider the existence of a heterogeneous receptor population, more than one receptor affinity state (Fig. 9.1 ) or negative cooperativity. The GraphPad Prism software allows to fit data to two equations according to one-site or two-site competitive binding models and compare the two fits -14] (more detailed information are available at: http://www.curvefit.com). If the existence of multiple binding sites is suspected, data can also be visualized using a plot of bound vs. bound x inhibitor concentration [15]. The plot is linear for a single class of binding sites but is markedly nonlinear for two or more classes of binding sites. The intercept on the y-axis is the amount of binding in the absence of the inhibitor while the slope is the negative reciprocal of the IC50.

A particular case of competition binding experiments is homologous competition. Sometimes, this type of experiment is performed to study binding saturability without varying the radioligand concentration. However, to obtain appropriate data, four conditions must be respected: (1) receptors must have identical affinity for the labeled and unlabeled ligand; (2) there is no coopera-tivity; (3) there is no ligand depletion; and (4) nonspecific binding is proportional to the concentration of the labeled ligand. Analysis of homologous competition binding data can be tricky, but the use of the nonlinear regression technique is quite useful [14] (detailed information are at: http://www.curvefit. com).

Kinetic binding data can be also analyzed using a nonlinear regression analysis (GraphPad Prism). From dissociation binding data, koff (min-1) and t1/2 (min) are obtained while Kon (molar-1 min-1) is derived from association binding data. Thus, from kinetic experiments, the calculation of KD is possible since it is equal to the ratio koff/kon.

Radioligand binding assays have been the mainstay of drug discovery and development. Even in the era of high-throughout screening (HTS), this versatile technique has retained its fundamental role. Indeed, incorporation of scintillation proximity technology together with automation and ratiometric counting instrumentation have served to maintain receptor binding assays as one of the primer tools of drug discovery (reviewed in Reference 16).

0 0

Post a comment