The ability of a drug to produce an effect on an organism is dependent on many of its properties, from its absorption to its stability to its elimination. To briefly summarize these processes, the first factor to be considered is the route of administration, which can determine how rapidly a drug reaches its target organ and which organs it affects. Oral administration typically results in a relatively slow onset of action. Parenteral describes all other routes of administration, including subcutaneous (under the skin), intraperitoneal (into the peritoneal–abdominal cavity), intravenous (into the venous system), intracerebroventricular (into the cerebral ventricular system), intrathecal (into spinal fluid), and intracerebral (into the brain parenchyma) delivery. The bioavailability of a drug determines how much of the drug that is administered actually reaches its target. Bioavailability can be influenced by absorption of the drug from the gut if administered orally. It also can be affected by binding of the drug to plasma proteins, which makes the drug unavailable to bind to its target. Moreover, it can be influenced by a drug’s ability to penetrate the blood–brain barrier if the drug acts on the brain (Chapter 2), or its ability to permeate cell membranes if the drug acts on intracellular proteins.
Drug action also depends on the stability of the drug once it is absorbed, that is, how rapidly it is metabolized to inactive congeners or eliminated from the body through urine, bile, or exhaled air. Some drugs (prodrugs) must be converted into active metabolites before they can exert their biologic effects.
Each of these factors, which can be categorized as pharmacokinetic considerations, is a critical determinant of drug action and influences both the clinical use of drugs and the process of developing new agents. However, these pharmacokinetic properties are not discussed in detail in this book because they are not, strictly speaking, related to the underlying mechanisms of drug action—the pharmacodynamic features that are the primary concern of these chapters. As an introduction to this topic, a brief description of the process by which a drug interacts with its initial protein target follows. Pharmacogenetics, which describes the influence of an individual’s genes in determining the response to a given drug, is also critical, but still in the earliest phases of understanding (see below).
Neuropharmacology is changing rapidly in response to the molecular revolution. In previous decades, neuropharmacology focused on the synapse and, more particularly, on the effects of drugs on neurotransmitters or neurotransmitter receptors. The action of drugs on synaptic targets remains an important field of investigation. The initial target of a drug generally determines the particular cells and neural circuits on which the drug acts and at the same time the potential efficacy and side effects of the pharmacologic agent. However, the molecular revolution has made it clear that the initial binding of a drug to its target—for example, the binding of a drug to a neurotransmitter receptor—is only the beginning of a signaling cascade that affects the behavior of cells and ultimately complex circuits.
When a drug binds to a protein, it affects the functioning of that protein, thereby establishing a form of allosteric regulation. A drug can conceivably bind to any site on a protein. A simple site may involve just a few contiguous amino acid residues in a protein’s primary structure, while a relatively complex site may involve discontinuous residues from the protein’s primary structure that are brought near each other by the protein’s secondary and tertiary structures. Ultimately, the three-dimensional shape, or conformation, of a binding site and the electrostatic charges distributed across the site must complement the shape and charge of the drug. The interaction of a drug with its binding site can influence the intrinsic activity of the target protein, for example, the catalytic activity of an enzyme or the conductance of an ion channel, or it can influence the ability of the protein to interact with some other molecule, such as the ability of a receptor to bind to its neurotransmitter.
In classic studies of drug mechanisms of action, a mechanism is defined by a drug’s ability to bind to an unknown receptor in tissue homogenates or on tissue sections. In these studies the drug, termed the ligand, is radiolabeled and incubated with a tissue preparation, which is washed extensively to remove loosely bound drug. A radioactive atom must be added to the drug without altering its ligand binding properties, a process that can be exceedingly difficult. The resulting ligand binding should be specific; that is, the ligand must bind to its specific target protein, which must be distinguished from binding to other proteins or even to the wall of a plastic test tube. In many cases, binding is stereoselective, or specific for only one stereoisomer of a drug. Binding also should be saturable. A limited amount of ligand binding occurs in the preparation because the amount of the specific target is limited. (A tissue preparation contains a finite amount of an individual receptor protein compared with a test tube wall, which is theoretically infinite.) Additionally, binding should attain a steady state. Time, temperature, and other conditions of incubation should enable the ligand binding to achieve a state of equilibrium.
The extent to which a ligand binds to a tissue preparation is a function of the concentration of the ligand 1–1. The total binding comprises two components: (1) specific binding, which is saturable, and (2) nonspecific binding, which is not saturable. In the ideal situation, in which binding to a specific receptor site is competitive and fully reversible in the steady state, specific binding can be defined as the fraction of total binding that can be displaced by incubating the radiolabeled ligand–tissue mixture with a large excess of unlabeled ligand. Conversely, the nondisplaceable radioactive portion of the preparation is considered nonspecific binding.
Radioligand binding assay. In this theoretical representation the amount of radioligand bound to a tissue preparation (eg, homogenate, brain slice) is a function of the concentration of the radioligand. Total binding is the total amount of binding observed. Nonspecific binding represents the nonsaturable portion of binding that is presumably not associated with the specific binding site under investigation; it is often calculated as the binding of radioligand that persists in the presence of a large excess of nonradiolabeled ligand. Specific binding is calculated as the difference between total and nonspecific binding and reflects the amount of radioligand bound to the specific binding site.
There are several discrepancies, however, between ideal and actual conditions. Not all binding to target proteins is truly reversible; the affinity of some ligand–receptor interactions is so high that resulting complexes are not readily dissociable. Moreover, artifactual sites may be present and may show striking apparent specificity. While the ideal situation assumes that the tissue preparation contains just one specific target, in actuality many drugs can bind specifically to many related subtypes of a protein target; for example, serotonin binds to numerous subtypes of serotonin receptors. Consequently, the resulting binding curves can be quite complicated and difficult to interpret.
The specific binding of a ligand to a tissue preparation is quantified according to two properties: the affinity of the binding, which is expressed as a dissociation constant (Kd), and the total amount of the binding (Bmax) 1–2. These terms are analogous to those used in studies of enzyme kinetics—for example, the Michaelis–Menten equation—in which Ka is the activation constant for an enzyme and its cofactor and Vmax is the maximum catalytic activity of the enzyme. The Kd is defined as the concentration of ligand at which half of the specific binding sites are occupied; larger Kd values (eg, 100 nM vs 1 nM) reflect lower affinities of the drug. When ligand binding is plotted as a function of the log of drug concentration, a sigmoidal curve is obtained 1–2B. Ligand binding data are often transformed mathematically to yield a Scatchard plot, in which the ratio of bound ligand to free ligand is plotted as a function of bound ligand 1–3. Because it is difficult to measure the amount of free (unbound) ligand, total ligand minus bound ligand is used. The shape of Scatchard plots provides an indication of the number of binding sites in a tissue preparation, as well as the Kd and Bmax values for each site.
Determination of Kd and Bmax from radioligand binding assays. The amount of specific radioligand binding to a specific site in a tissue preparation (determined in 1–1) is plotted as a function of radioligand concentration, using a normal A or semilogarithmic B plot. The Kd is calculated as the concentration of radioligand that results in 50% of maximal binding (Bmax). The semilogarithmic plot, which better illustrates the effects of low radioligand concentrations, places the Kd near the middle of the graph.
The Scatchard plot. Specific binding data are mathematically transformed to plot the ratio of bound to free radioligand as a function of bound radioligand. A. When one binding site is involved, the data follow a straight line. The slope of the line is the Kd and the x-intercept is the Bmax. B. When more than one binding site is involved, the data follow convex curves, which can be converted into multiple straight lines. The slope of each line and its x-intercept represent the Kd and Bmax, respectively, of each binding site.
Another method for studying ligand–target interactions makes use of competition curves. These curves describe the ability of a drug to compete with a radioligand in binding to a tissue preparation 1–4. The drug concentration at which half of the radioligand binding is displaced (Ki) is a measure of the affinity of the drug for a binding site in the context of a specific radioligand. Historically, such competition studies have played an important role in defining many subtypes of neurotransmitter receptors. In general, such pharmacologic distinctions of receptors accurately predicted broad categories of receptor proteins, which subsequently were identified with greater precision by means of molecular cloning techniques, as discussed later in this chapter. In ideal situations, the competing drug and the radioligand bind to the same site of the target protein; such binding is termed competitive. In more complicated situations, the drug and radioligand bind to different sites on the same protein; in such cases, the binding is noncompetitive and results in far more complicated competition curves. These assessments of ligand binding, which have traditionally been performed on tissue homogenates or membrane fractions, can also be performed on brain sections, a process termed receptor autoradiography.
Radioligand competition curve. The ability of compounds to bind to a particular site in a tissue preparation can be compared by studying the ability of each to compete with a radioligand for a particular binding site. When the binding data are plotted on a semilogarithmic graph, a sigmoidal curve results. The Ki represents the concentration of drug that results in a reduction of radioligand binding to 50% of maximal values.
As with any technique, the limitations of ligand binding assays must be appreciated. One of the most critical limitations is that binding assays, and the determination of Kd, Ki, and Bmax values, are highly dependent on experimental conditions and therefore must be interpreted with considerable caution. The specific radioligand, the temperature of incubation, the salt and ionic content of the buffer, and the presence of different guanine nucleotides (Chapter 4) are among the many factors that can exert dramatic effects on ligand binding. The cloning of receptors and other target proteins, and the ability to express them on cells without endogenous expression of the target, has made at least some aspects of characterizing the binding properties of drugs more straightforward.
Binding studies describe the physical relationship between a drug and its target but do not directly assess the biologic consequences of this association. Although drug binding and biologic effect are intricately related, they help define two distinct aspects of drug action: potency and efficacy. Potency (affinity, or Kd) describes the strength of the binding between a drug and its target. Efficacy describes the biologic effect exerted on the target by virtue of the drug binding. These properties can be understood by considering the effect of a drug on a neurotransmitter receptor. As previously explained, the drug must physically bind to the receptor, which requires a physical attraction between the two. Subsequently, that binding must elicit a change in the receptor that leads to a biologic response. For a G protein–coupled receptor, drug binding must trigger a conformational change in the receptor that alters its interactions with its G protein α subunit (Chapter 3). For a ligand-gated channel (receptor ionophore), drug binding must trigger a conformational change that opens or closes the pore that is intrinsic to the receptor.
Drugs differ dramatically with respect to their potency and efficacy. Traditionally, two categories of drug have been described: agonist and antagonist. When an agonist binds to a receptor, it mimics the endogenous neurotransmitter by producing the same conformational change and hence the same biologic response. According to this definition, all neurotransmitters are receptor agonists. When an antagonist binds to a receptor, it elicits no such change; thus, an antagonist is inherently inert and exerts a biologic effect only by interfering with an endogenous ligand. For opioid receptors, which are receptors for the endogenous opioid peptides such as the enkephalins (Chapter 7), morphine and naloxone are classic examples of an agonist and antagonist, respectively. The differences in efficacy associated with agonists and antagonists are independent of the affinity with which each binds to its receptor; both can exhibit high or low affinities. How can two molecules that bind to the same receptor site exert such different effects on the receptor? There are two possible explanations. An antagonist may share one moiety that is required for binding to the receptor with an agonist, but may lack a moiety required for efficacy. Alternatively, the antagonist and agonist may bind to overlapping but distinct sites on the receptor.
In addition to the actions of classic agonists and antagonists, an intermediate category of drug efficacy is exemplified by partial agonists. When a drug binds to a receptor and elicits only a partial biologic response, the drug presumably lacks a portion of the molecule required for full biologic effect or binds to a slightly different site on the receptor 1–5. An interesting situation arises when partial agonists possess high potency. At low drug doses, a mild agonist effect is obtained. At high doses, a similarly mild agonist effect is obtained because of limits in the intrinsic efficacy of the molecule. However, at high doses, the drug can antagonize the ability of a full agonist, including the endogenous neurotransmitter, to activate the receptor because its affinity is greater than that of the full agonist. For this reason, partial agonists are sometimes referred to as mixed agonists–antagonists. Partial agonists can be quite useful clinically; for example, buprenorphine is a partial agonist at opioid receptors and is used in the treatment of chronic pain and opiate addiction (Chapters 11 and 16). At low doses, buprenorphine elicits a mild analgesic and rewarding effect. Higher doses not only fail to yield a stronger effect, which limits the abuse liability of this drug, but also antagonize the action of full opioid agonists and thereby discourage abuse of opiates such as morphine.
Drug efficacy versus drug potency. The biologic responses elicited by two drugs that bind to the same site are presented in this representation as a function of drug concentration. Efficacy refers to the maximal biologic response elicited by each drug. In this theoretical representation the partial agonist elicits a smaller maximal response than the full agonist. However, efficacy is independent of the potency (Kd) of the drug; the partial agonist shown is in fact more potent (possesses a higher affinity for the binding site) than the full agonist.
Inverse agonists achieve efficacy in still another way. When an inverse agonist binds to a receptor, it elicits the biologic response that is the opposite of that associated with an agonist. If an agonist opens an ion channel, an inverse agonist closes the channel. If an agonist facilitates receptor-to-G protein coupling, an inverse agonist attenuates such coupling. The action of an inverse agonist requires some basal activity on the part of the receptor, which means that the receptor is not quiescent in the absence of ligand but instead possesses some level of intrinsic biologic activity, such as channel conductance or G protein coupling. Indeed, most receptors do exhibit such baseline activity.
Very few drugs can be placed in discrete categories—agonist, antagonist, and inverse agonist. Many drugs that are classically described as agonists, such as morphine, are not full agonists but strong partial agonists. Conversely, many drugs that are classically categorized as antagonists—for example, naloxone—are not completely inert and thus can be very weak partial agonists. Moreover, some neurotransmitters show less efficacy than synthetic drugs, which indicates that they also are partial agonists. Consequently, drugs should be thought of as existing on a continuum ranging from full agonist to inert antagonist to full inverse agonist 1–6.
Drug efficacy as a continuum. Ligands for a receptor can be described as agonists (agents that activate the receptor), antagonists (agents that have no intrinsic effect on the receptor but can block the ability of agonists and inverse agonists to regulate the receptor), or inverse agonists (agents that regulate the receptor but produce effects opposite to those produced by agonists). However, ligands rarely can be placed into these discrete categories; instead they are distributed across a continuum. In strict pharmacologic terms, there are very few true antagonists, most being very weak partial agonists or inverse agonists, and very few full agonists or inverse agonists, most being strong partial agonists or inverse agonists.
The complex nature of the interactions between drugs and their target receptors can be illustrated by a discussion of the γ-aminobutyric acid receptor (GABAA)—an important receptor for the neurotransmitter GABA (Chapter 5). This receptor is a heteropentamer that has two main types of subunits, α and β. GABA binds to a site on the β subunit and triggers the opening of a Cl− channel that is intrinsic to the receptor complex. Muscimol (an agonist) and bicuculline (an antagonist) also bind at this site and thus compete with GABA. The α subunit of the GABAA receptor contains a binding site for a class of synthetic molecules known as benzodiazepines. Agonists that bind at this site, such as diazepam, are antianxiety agents that, when bound to the site, allosterically facilitate the ability of GABA to bind to and activate the GABAA receptor (Chapters 5 and 15). Antagonists at this site, such as flumazenil, bind to the site but do not affect receptor function. Because this site lacks endogenous ligands, flumazenil is clinically inactive when bound to the GABAA receptor; however, it can be used to treat diazepam overdose because it displaces diazepam from the binding site. Inverse agonists, such as β-carboline, which intensify anxiety, bind very near the benzodiazepine agonist site and allosterically inhibit the ability of GABA to bind to and activate the receptor. Diazepam and β-carboline are a noncompetitive agonist and inverse agonist, respectively, of the GABAA receptor, in that they interact with a binding site distinct from the GABA site. They are also considered positive and negative allosteric modulators of the GABAA receptor. Although this type of complex receptor pharmacology was first described for the GABAA receptor, a similar level of complexity can characterize drug interaction at virtually any type of receptor.
In more recent years, more complex actions of receptors have contributed to still additional modes of drug action. Specific types of receptors can form complexes with other receptors, with a resulting generation of novel agonist activity for the receptor heteromer 1–7A. For example, heterodimers of μ and δ opioid receptors can be activated or inhibited by drugs that show lower affinities to either receptor type alone. As well, different agonists at a given receptor can direct the receptor to signal via distinct intracellular pathways, a process referred to as ligand-directed or biased signaling 1–7B. Similar ligand–directed differences can be seen with inverse agonists. The discovery of ligand-directed signaling greatly complicates the consideration of the intrinsic efficacy of a ligand, since ligands might vary qualitatively in addition to quantitatively with respect to their downstream actions.
More complex considerations of receptor signaling. A. Several types of receptors have been shown to form dimers, with the receptor dimer responding to different ligands (agonists, antagonists, or inverse agonists) compared with the receptor monomers. B. Other receptors exhibit ligand-directed signaling. In the example shown, one class of agonist at the serotonin 5HT2A receptor activates Gq with downstream signaling via phospholipase C, whereas another class of agonist triggers the receptor to activate Gi/o (and possibly other G proteins) with downstream signaling via arachidonic acid metabolites.
Finally, it must be emphasized that binding sites with a high affinity for drugs do not necessarily have an endogenous ligand. No evidence, for example, supports the existence of an endogenous ligand for the benzodiazepine binding site on the GABAA receptor. Rather, the discovery of this class of drugs and their binding site is testimony to the power and promise of medicinal chemistry to target distinctive features of proteins that are not exploited by nature. Indeed, our growing knowledge of the many complexities of receptor function and drug action discussed above is contributing to the development of a host of agents with novel pharmacologic and hence clinical activity.
Dose-Dependent Drug Response
That the effect of a drug on a target protein is dependent on the concentration of a drug is implicit in the discussions of drug binding and efficacy presented in the preceding sections. This dose dependency of drug action is one of the principal tenets of neuropharmacology and illustrates the importance of studying the effects of a wide range of drug doses.
One application of dose–response curves is in determining whether a form of treatment—for example, chronic exposure to an antidepressant—increases or decreases the responsiveness of a particular receptor system. Hypothetical cases are illustrated in 1–8, which shows that a reduction in receptor sensitivity in response to treatment is characterized by a rightward or downward shift in the dose–response curve, whereas an increase in receptor sensitivity is characterized by a leftward or upward shift in the dose–response curve.
Rightward and leftward shifts in dose–response curves. A rightward, or downward, shift indicates a reduction in drug sensitivity: more drug is needed at all concentrations to elicit the same level of biologic response. A leftward, or upward, shift indicates an increase in drug sensitivity: less drug is needed at all concentrations to elicit the same level of biologic response.
Dose–response curves also can reveal that the biologic effects of a specific drug may not be a simple (monotonic) function of drug dose. When the effects of a drug are more complex, nonmonotonic, for example, they are represented by an inverted U-shaped curve 1–9. Such drugs elicit a progressively greater biologic response with greater drug dose up to a point, after which higher drug doses begin to produce smaller effects. This shift in effect most likely occurs because the drug begins to act on a different target protein at higher drug doses, and action on the second target opposes the effects of the first. Alternatively, high doses may cause receptor desensitization.
Inverted U-shaped dose–response curve. Dose–response curves that are placed on a semilogarithmic plot often are not sigmoidal, such as those in previous figures, but instead form an inverted U shape. Such curves contain an ascending limb at lower drug doses and a descending limb at higher drug doses. These curves indicate that the biologic response elicited by a drug progressively increases as the drug dose increases and subsequently peaks at a moderate dose; higher doses elicit progressively smaller responses.
The analysis of full dose–response curves is necessary to determine reliably whether a particular treatment causes an increase or decrease in drug responsiveness. 1–10 shows a leftward shift in the dose–response curve for a drug whose biologic effects are an inverted U-shaped function of drug concentration. Without an analysis of the full dose–response curve, an investigator may incorrectly interpret effects of the drug; for example, depending on the concentration of drug used to activate the receptor, a shift in the curve may indicate a reduction, an increase, or a lack of change in drug response.
Analysis of a full dose–response curve. Because the biologic effects of many drugs are described by an inverted U-shaped dose–response curve, the effects of a drug should be analyzed over a wide range of doses. The graph shows a leftward shift in an inverted U-shaped dose–response curve occurring after an experimental treatment. With analysis of a single drug dose, it might be determined that the treatment causes (1) an increase in drug sensitivity (dose a); (2) no change in drug sensitivity (dose b); or (3) a decrease in drug sensitivity (dose c). The leftward shift in the dose–response curve becomes apparent only after a wide range of doses are analyzed.
Drug Interaction With Nonreceptor Proteins
Although most principles of drug action have been ascertained from studies of neurotransmitter and hormone receptors, the same general principles apply to interactions between drugs and nonreceptor proteins. A drug binds to a specific site on a protein, which can be determined by means of ligand binding assays. Drug binding influences the function of a protein by either facilitating or inhibiting that protein’s normal functioning, including its interactions with other macromolecules. Some drugs create a new function for the protein to which they bind; examples include FK506 and related drugs that bind immunophilins (Chapter 4). When such drugs bind to immunophilin proteins, the proteins become potent inhibitors of calcineurin, a protein phosphatase.
The conditions under which two proteins interact are conceptually similar to those for drug–target interactions. Protein–protein interactions have emerged as a central theme of cell regulation (Chapter 4). The binding of proteins such as transcription factors to specific sequences of DNA, which is a key mechanism of gene regulation in development and in neural plasticity in the adult, also operates according to principles like those of drug–target interactions.