Over the past decade, tremendous strides have been made in the treatment of major depression due to the ability to rationally develop psychiatric medications.²¹⁵ In the last 10 years, 6 new medications have been marketed as antidepressants in the U.S. (Table 2.1):

• Bupropion
• Fluoxetine
• Sertraline
• Paroxetine
• Venlafaxine
• Nefazodone

Three of these drugs (fluoxetine, paroxetine and sertraline), together with two other drugs (citalopram and fluvoxamine) marketed as antidepressants elsewhere in the world, form the class known as selective serotonin reuptake inhibitors (SSRIs) (Figure 2.1). For many physicians, this class has supplanted tricyclic antidepressants (TCAs) as the antidepressant of first choice due to their greater safety and tolerabilty, coupled with comparable efficacy.

The rapid expansion in the number of antidepressants and the change in what is considered first-line therapy has been accompanied by a substantial amount of commercial claims and counterclaims. This situation can be confusing, even for physicians who specialize in clinical psychopharmacology, and even more so for the general physician who must also contend with developments in other therapeutic areas. Therefore, this book will explain the differences between the SSRIs and the TCAs, which were the mainstay of pharmacotherapy for major depression for many years, and discuss the clinically important similarities and differences between members of the SSRI class.

FIGURE 2.1— Structural Formulas of Several SSRIs

TABLE 2.2 — Criteria for New Drug Development Inclusion criteria: have the desired MOA Exclusion criteria: avoid effects of other MOAs Goals of Such Development Maintain or enhance efficacy Increase therapeutic index Improve tolerability profile Reduce likelihood of pharmacodynamic drug-drug interactions Reference: 217

FIGURE 2.2— Schematic Illustration of Relationship Between Drug Site of Action and Effect

The development of SSRIs occurred over a relatively short interval of time. The first SSRI marketed was zimelidine by Astra. Unfortunately, several cases of Guillain-Barre syndrome were associated with the use of this drug and led to its withdrawal from the market. Nonetheless, five SSRIs were eventually launched successfully in multiple countries around the world. Each was developed by a different company:

Parenthetically, while fluvoxamine is marketed as an antidepressant in many parts of the world, it is marketed only for obsessive-compulsive disorder in the U.S. Citalopram, while marketed in several countries in the world as an antidepressant, is not yet available in the U.S. Fluoxetine was the first SSRI marketed in the United States in 1988. Table 10.2 (in the Appendix) lists the SSRIs that are available in various countries.

The fact that the five SSRIs were produced by five different companies is a testimony to the shift from a discovery process dependent on chance observation to a process of rational drug development. Understanding rational drug development is pivotal to understanding the clinical pharmacology of SSRIs.

How the SSRIs were developed is a scientific success story. The SSRIs are the first rationally designed class of psychotropic medications and, hence, have launched a new era in psychotropic drug development. The strategy behind rational drug development is to design a new drug that is capable of affecting a specific neural site of action (SOA) (eg, uptake pumps, receptors) while avoiding effects on other SOAs. The goal in such development is to produce agents that are more efficacious, safer and better tolerated than older medications (Table 2.2).²¹⁵ This enhanced safety profile includes a reduced likelihood of pharmacodynamically mediated adverse drug-drug interactions by avoiding affects on SOAs that are not essential to the intended outcome (eg, antidepressant efficacy).

A few general comments about what a drug must do to produce a specific clinical effect may be helpful to put this book in perspective. A drug must act on an SOA that is physiologically relevant to the effect (Figure 2.2). That SOA may be, by way of example but not limited to, an uptake pump, an enzyme, or a receptor. The drug "recognizes" and binds to that SOA. The activation or inhibition of a specific site is termed the drug's mechanism of action (MOA). For example, a drug may be an agonist or antagonist at a specific serotonin receptor.

A given drug may affect one or more SOAs over its clinically relevant dosing range and, therefore, may produce multiple and different clinical effects, some desired and some not. Drugs that affect multiple SOAs are more characteristic of drugs developed based on chance discovery, whereas the goal of rational drug development is to produce drugs with a more limited range of effects (Table 2.2).

Prior to the SSRIs, all psychotropic medications were the result of chance observation (Table 2.3). Lithium came from studies looking for putative endogenous psychomimetic substances excreted in the urine of psychotic patients.⁵¹ The phenothiazines came from a search for better preanesthetic agents.¹⁵⁴ The TCAs were the result of an unsuccessful attempt to improve on the antipsychotic effectiveness of phenothiazines.¹⁴⁸ The monoamine oxidase inhibitors (MAOIs) came from a failed attempt to develop effective antitubercular medications.⁶³ The first studies of benzodiazepines were unsuccessful attempts to treat patients with schizophrenia.

Despite these initial failed attempts to use these various drugs therapeutically, astute clinical investigators recognized their therapeutic value in other conditions: lithium for manic-depression, phenothiazines for psychotic disorders, TCAs and MAOIs for major depression, and benzodiazepines for anxiety disorders.

This chance discovery process is not unique to psychiatry, but instead has been the universal first step in any therapeutic area.²¹⁰ Prior to such first steps, too little knowledge of the biology underlying illnesses existed to permit a more rational approach to drug development. However, these first drugs played an important role in providing the first insights into the pathophysiology underlying the illness or, at least, underlying drug responsiveness.

Molecular targeting is the essence of rational drug development.²¹¹ In this approach, the specific target(s) of interest is a fundamental brain mechanism believed to be important in the pathophysiology underlying a specific psychopathologic condition or psychiatric syndrome (eg, major depression). This SOA may be the neuronal uptake pump for a neurotransmitter, a specific neurotransmitter receptor subtype, or a subunit of an ion channel (Figure 2.2).

As was the case with each of the SSRIs, the new molecular entity is developed to stereospecifically interact with the target of interest. At the same time, the molecule is structurally modified so that it does not interact with other targets that mediate unwanted effects (eg, peripheral anticholinergic effects). Through this systematic approach, a new candidate drug is selected for clinical testing to support registration for marketing. This type of rational drug development is now possible in psychiatry because of the improved understanding of central and peripheral mechanisms of action (MOAs) relevant to both desired and undesired central and peripheral effects.

FIGURE 2.3 — Standard and New Generation Antidepressants Mechanisms of Action

Antidepressant pharmacotherapy is the first area in psychopharmacology to have benefitted significantly from such targeted development.²¹⁵ Figure 2.3 illustrates the evolution of antidepressants over the past three decades. TCAs and MAOIs were the first successful antidepressants, but their antidepressant properties were discovered by chance, as discussed above. Nonetheless, this chance discovery was important. First, these drugs provided the first scientifically proven treatments for major depression and demonstrated that major depression was amenable to medical intervention just as other medical conditions, such as hypertension and diabetes. Second, they served as roadmaps to improve our understanding of the MOAs, mediating both their desired antidepressant effects and undesired effects. This information was critical to the rational drug development efforts that followed and led to the SSRIs and other new antidepressants.

In the case of major depression, TCAs and MAOIs implicated the potentiation of neurotransmission in one or more than one central biogenic amine neural system as potential MOAs responsible for their antidepressant efficacy. That finding, coupled with improved means of isolating and studying the effects of drugs on specific neural mechanisms, led to the development of the SSRIs.

The nature of older chance-discovery drugs is that they have many clinical effects either because they affect an SOA with broad implications for organ function (eg, MAOIs that affect an enzyme responsible for the degradation of four major neurotransmitters) or because they affect multiple SOAs (eg, TCAs). Such drugs typically have:

• Narrow therapeutic indices
• Poor tolerability profiles
• Potential for causing multiple types of pharmacodynamic interactions with a wide variety of concomitantly prescribed medications

The SSRIs were developed based on the knowledge gained from studying the effects of the TCAs and the techniques developed in basic neuroscience research to isolate and study the effects of drugs on specific neural SOAs (eg, uptake pumps, receptors). In the case of the SSRIs, each was the product of a similar development strategy in which the goal was to produce a drug capable of inhibiting the neuronal uptake pump for serotonin, a property shared with the TCAs, but without affecting the various other neuroreceptors (ie, histamine, acetylcholine, and a-adrenergic receptors) or fast sodium channels, affected by the TCAs. Actions on these latter sites are responsible for many of the safety and tolerability problems of the TCAs.^220,221 The fact that SSRIs were designed to avoid affecting these other SOAs explains many of the pharmacological differences between the SSRIs and the TCAs (see Section 4) and explains the similarities among the SSRIs (see Section 5). In many ways, the SSRIs are to psychiatry as b-blockers are to internal medicine.

In contrast to rational development, chance discovery is usually dependent on the drug's having a large signal-to-noise ratio (ie, a big clinical effect or multiple clinical effects). Unfortunately, this fact means that chance-discovery drugs typically will produce a number of undesired, as well as desired, effects and will have a narrower therapeutic index in comparison with a drug that was rationally developed to affect only the SOA(s) necessary to produce the desired response.

This issue can be readily understood by examining the pharmacology of TCAs that has served as the cornerstone of antidepressant pharmacotherapy for almost 30 years. TCAs affect multiple SOAs over a relatively narrow concentration range so that patients are likely to experience multiple effects while taking these medications.^{34,66,221,225} Some MOAs of TCAs (ie, the inhibition of the fast sodium channels) can cause potentially serious effects on cardiac conduction and occur at concentrations only an order of magnitude higher than the concentration needed to inhibit the neuronal uptake pumps for norepinephrine and serotonin, the putative MOAs mediating the antidepressant effects of TCAs. This fact explains why an overdose of TCAs of only 5 to 10 times their therapeutic dose can cause serious toxicity and why patients who have a slow clearance rate for these drugs can develop serious adverse effects on routine doses due to the accumulation of toxic concentrations.²²⁸

To put this issue with TCAs in perspective, Table 2.4 illustrates the cocktail of drugs, each having only one predominant MOA, that would have to be given to a patient to reproduce the effects that occur in a patient receiving a tertiary amine TCA, such as amitriptyline. Obviously, the problem with amitriptyline is that the patient has to experience a large number of effects to receive the benefit of the mechanism that mediates antidepressant response.

The issue of multiple MOAs over a narrow concentration range is further complicated by the fact that there is a large interindividual variability in the clearance rates of TCAs, even in physically healthy individuals.²¹³ The variability is even larger when dealing with the elderly, the medically ill, and patients on concomitant medications that can either induce or inhibit the clearance of these drugs. With TCAs, patients can have numerous types of adverse effects ranging from nuisance problems (eg, dry mouth) to serious toxicity (eg, seizures, cardiac arrhythmias). Patients who clear the drugs slowly may experience the latter due to the accumulation of excessive concentrations despite being on conventional doses.

This situation is made even more complicated because the early signs of TCA-induced toxicity can mimic worsening of major depression so that the physician may unfortunately respond by increasing rather than reducing the dose.²²⁷ These facts considered together have made therapeutic drug monitoring (TDM), at least once during early treatment (at the end of the first week of treatment with a stable dose), a standard aspect when prescribing TCAs.^221,228 Using the TDM results, rational dose adjustment can then be made to compensate for the intraindividual differences in clearance rate, and thus ensure that the patient will be treated with a dose that will achieve a concentration that is optimal for most patients with regard to efficacy, safety and cost effectiveness.

FIGURE 2.4 — In Vitro Potency of Amitriptyline as a Representative Tricyclic Antidepressant for Different Sites of Action and Related Mechanisms of Action

Adapted from references: 34, 66

Unfortunately, TDM-driven dose adjustment does not substantially improve the tolerability of TCAs because MOAs for producing adverse effects (eg, those mediated by histamine or muscarinic receptor blockade) are more potent and hence occur at lower concentrations than their presumed MOAs underlying their antidepressant efficacy (ie, inhibiting the neuronal uptake for norepinephrine and serotonin) (Figure 2.4). Hence, patients who are sensitive to a given MOA may experience discomforting adverse effects even at concentrations that are subtherapeutic for treating major depression. That problem has been addressed by rational drug development of new drugs (eg, SSRIs) with a much wider gap between their potency for an effect on the desired versus undesired targets. As a result, we now have seven major classes of antidepressants based on putative MOAs mediating antidepressant response (Table 2.1).

The SSRIs were all developed to have a similar MOA: the potentiation of serotonin (5-HT) by the inhibition of its neuronal uptake pump. As such, all SSRIs have common 5-HT agonistic effects that appear to mediate both their desired (eg, antidepressant efficacy) and undesired (eg, sexual dysfunction) reactions. As a class, SSRIs are considerably more selective in comparison to TCAs in terms of their central nervous system MOAs, but differ in other clinically important ways, as will be discussed in detail in this book (see Sections 6 through 8).

The reason to choose serotonin uptake inhibition as the desired MOA is based on the emerging understanding of the role of serotonin in the brain as well as on the pharmacology of TCAs and MAOIs. From a phylogenetic standpoint, serotonin is one of the oldest neurotransmitters.²⁵⁵ It is found in such relatively simple organisms as jellyfish. In the human brain, serotonin-containing neurons are highly localized in specific clusters in the brainstem and spinal cord.²⁷¹ From these sites, the cells send out axons that end in serotonin-containing terminals innervating the diverse areas throughout the brain. These regions include:

• Spinothalamic pain fibers
• Brainstem
• Cerebellum
• Hypothalamus
• Basal ganglia
• Neocortex

This anatomy explains why serotonin is implicated in so many brain functions including:

• Pain perception
• Sleep
• Thermal regulation
• Appetite
• Gut regulation
• Balance
• Reproductive function
• Motor function
• Higher cognitive function
• Sensory interpretation

Given these diverse responsibilities, dysfunction of serotonin neurons have been implicated in a wide variety of diseases, including major depression. For the same reason, serotonin-active drugs can have many different clinical effects by virtue of their physiological effects on diverse brain regions. This anatomy explains why even "selective" drugs such as SSRIs can produce so many diverse clinical effects (eg, nausea, a feeling of incoordination, suppression of REM sleep, decreased libido, akathisia) as well as being useful in such apparently disparate disorders as major depression, anxiety disorders, pain disorders, and premature ejaculation. While SSRIs are "selective" in terms of affecting the neuronal uptake pump for serotonin, this action affects a multitude of specific postsynaptic serotonin receptors (eg, 5-HT1A, 5-HT1D, 5-HT2A, 5-HT2C, and 5-HT3) which, in turn, affects a multitude of neural systems.¹²⁸

Chirality

Although all SSRIs are products of rational drug development, one of the major goals of such a development was not realized with two of the SSRIs due to the phenomenon of chirality: that goal was to produce a drug that is a single molecule with a precise, limited (or focused) range of pharmacological actions. If the molecule has an asymmetrical carbon, then it exists in enantiomeric forms (ie, chirality). As can be seen in Figure 2.1, all of the SSRIs except fluvoxamine have an asymmetrical carbon. However, only one enantiomer of paroxetine and sertraline, respectively, is contained in the marketed formulation of these two drugs. In contrast, citalopram and fluoxetine are marketed as the racemates of their two enantiomers. Hence, patients on these two SSRIs achieve plasma and tissue levels of each enantiomer and their respective metabolites, which are also enantiomers.

This fact raises the question of whether there are substantial differences in the pharmacodynamics and pharmacokinetics of these enantiomers and whether such differences contribute in a meaningful way to the variance in drug response among different patients.¹² If the different enantiomers have meaningful differences in their therapeutic ratios, one enantiomer can contribute disproportionately to adverse consequences relative to therapeutic benefit. The presence of enantiomers complicates the use of TDM in both research and clinical practice since many assays will not distinguish between the two enantiomeric forms of a drug. If there are meaningful differences in their pharmacodynamics and pharmacokinetics, that fact can add substantial "noise" to such results and thus confound their interpretation.

As is the case with the enantiomers of citalopram and fluoxetine, there is often limited data on their relative pharmacodynamics and pharmacokinetics to answer these questions. A summary of that data for these two SSRIs follows.

The racemic mixture of citalopram produces racemic desmethylcitalopram and didesmethylcitalopram. The S-enantiomers are potent and selective inhibitors of serotonin uptake in contrast to the relatively inactive corresponding R-enantiomers.¹³⁰ The active S-enantiomer of citalopram is generally only one-third of the total citalopram plasma level under steady-state conditions.²⁴² However, there is variability in this ratio among different patients that may be characteristic of patients genetically deficient in cytochrome P450 (CYP) 2C19.²⁴² CYP 2C19 enzyme is the principal enzyme responsible for the metabolism of citalopram.²⁵⁴

This variability in the ratio of the active to the relatively inactive enantiomer can contribute to variability in response to the drug among different patients. Given the relative levels and activity of the enantiomers of citalopram and its metabolites, studies attempting to correlate plasma levels of citalopram with serotonin mediated effects should report on the levels of each enantiomer or should focus on the levels of S-citalopram.

Racemic fluoxetine produces racemic norfluoxetine. While S-fluoxetine, R-fluoxetine, and S-norfluoxetine are potent and selective inhibitors of serotonin uptake in vitro and in vivo, that is not true for R-norfluoxetine.^98,290,291 Under steady-state conditions, the plasma levels of racemic fluoxetine and norfluoxetine are comparable.¹⁸⁹ Thus, studies attempting to correlate the plasma levels of fluoxetine and norfluoxetine should ideally take into account the relative inactivity of the R-norfluoxetine in terms of the inhibition of serotonin uptake.

The R-enantiomers of fluoxetine and norfluoxetine are also weaker inhibitors of CYP 2D6 than are the S-enantiomers.²⁶⁷ Thus, failure to distinguish between these enantiomers in studies attempting to correlate plasma levels of fluoxetine and norfluoxetine with the inhibition of the metabolism of CYP 2D6-dependent substrates will hamper the ability to establish such a relationship.

Tables summarizing the above data are found in the sections dealing with the effects of the SSRIs on neural mechanisms (Section 3, Table 3.4) and on CYP enzymes (Section 8, Table 8.11), respectively. There may be other important differences in the pharmacodynamics and pharmacokinetics of the enantiomers of these two SSRIs which are not known at this time. There is little active research ongoing in this area; therefore, knowledge of these enantiomers may not expand appreciably in the near future. This discussion should be kept in mind as a caveat when reading the rest of this book. Unless specified otherwise, the data in this book on in vitro and in vivo studies with citalopram and fluoxetine were done with the racemic mixtures, and the plasma and tissue levels reported of the parent compound and the metabolites are the combined levels of their enantiomeric forms.