CHAPTER 11
`
`The CYP3 Family
`
`DAVID J. GREENBLATT, PING HE , LISA L. VON
`MOLTKE AND MICHAEL H. COURT
`
`Department of Pharmacology and Experimental Therapeutics, Tufts
`University School of Medicine and Tufts-New England Medical Center
`Boston MA, USA
`
`Table of Contents
`
`11.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354
`11.2 Pharmacogenomics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
`11.3 Enzyme Kinetics of CYP3A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360
`11.4 Individual Variability in CYP3A Metabolic Phenotype . . . . . . . . . . 361
`11.5 Age and Gender Effects on CYP3A Phenotype . . . . . . . . . . . . . . . 363
`11.5.1 In vitro and Experimental Studies . . . . . . . . . . . . . . . . . . . . 363
`11.5.2 Clinical Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364
`11.6 Drug Interactions with CYP3A Substrates . . . . . . . . . . . . . . . . . . . 365
`11.6.1 Drug Interactions via Metabolic Inhibition . . . . . . . . . . . . . 365
`11.6.2 Drug Interactions Involving Metabolic Induction. . . . . . . . . 371
`11.7 Comment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372
`Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372
`References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372
`
`11.1 Introduction
`
`The CYP3A enzymes are of major importance in human biology and clinical
`therapeutics.1,2 They are the most of abundant of the CYP enzymes in the
`
`Issues in Toxicology
`Cytochromes P450: Role in the Metabolism and Toxicity of Drugs and other Xenobiotics
`Edited by Costas Ioannides
`r Royal Society of Chemistry, 2008
`
`354
`
`1
`
`TEVA1026
`
`

`

`The CYP3 Family
`
`355
`
`liver,3 and are the only CYPs present in substantive amounts in the enteric
`mucosa of the gastrointestinal tract. Substrate specificity is broad; CYP3A
`mediates the biotransformation of numerous endogenous substances, en-
`vironmental chemicals of potential toxicological relevance, and medications
`used in clinical therapeutics (Table 11.1).
`The significance of CYP3A in terms of human evolution and species preser-
`vation is a topic of logical speculation. The lack of a phenotypic ‘‘null’’ status
`for CYP3A – corresponding for example to CYP2D6 ‘‘poor metabolisers’’ –
`supports the notion that CYP3A is essential. Homeostasis of many endogenous
`steroid hormones, including those required for sexual maturation and repro-
`duction, is dependent on CYP3A.4,5 A number of environmental chemicals,
`both therapeutic and toxic, derived from plants and fungi are in fact substrates
`for metabolism by CYP3A enzymes. Examples include the cinchona alkaloids,
`the taxanes, opiates, aflatoxins, and the immunosuppressants cyclosporine,
`tacrolimus and sirolimus. Further clues derive from the anatomic distribution of
`human CYP3A, and its unique kinetic characteristics. The dual localisation of
`CYP3A in enteric mucosa and liver provides the species with ‘‘two shots’’ at
`protection against potentially toxic environmental chemicals before they reach
`the systemic circulation (Figure 11.1). The property of binding cooperativity,
`
`Table 11.1 Representative CYP3A substrate drugs used in clinical practice.
`
`Clearance completely or nearly
`completely dependent on CYP3A
`
`Clearance partially dependent
`on CYP3A
`
`Amitriptyline
`Citalopram
`Clozapine
`Dextromethorphan
`Diazepam
`Imipramine
`Methadone
`Omeprazole
`Sertraline
`Telithromycin
`Voriconazole
`Zolpidem
`
`Alfentanyl
`Alprazolam
`Atorvastatin
`Buspirone
`Carbamazepine
`Cyclosporine
`Eletriptan
`Erythromycin
`Felodipine
`Midazolam
`Nefazodone
`Nifedipine
`Quetiapine
`Quinidine
`Ritonavir
`Saquinavir
`Sildenafil
`Simvastatin
`Tacrolimus
`Tadalafil
`Trazodone
`Triazolam
`Vardenafil
`Verapamil
`
`2
`
`

`

`356
`
`Oral dose
`
`Chapter 11
`
`G.I. TRACT
`
`Enteric CYP3A
`
`PORTAL
`CIRCULATION
`
`LIVER
`(hepatic CYP3A)
`
`SYSTEMIC
`CIRCULATION
`
`Figure 11.1
`
`Schematic diagram of the fate of an orally-administered CYP3A substrate
`drug prior to reaching the systemic circulation. Metabolism by CYP3A4
`and CYP3A5 is possible during passage across the enteric mucosa of the
`proximal small bowel as well as during passage through the liver.
`
`yielding homotropic autoactivation, is consistent with a protective effect of in-
`creased CYP3A activity upon exposure to high and potentially toxic quantities
`of exogenous substrate.6 Finally, there is extensive (but not complete) overlap in
`affinity of substrates for metabolism by CYP3A along with efflux transport by
`P-glycoprotein.7–9 Many of the natural substances described above are sub-
`strates for both, consistent with a coincident ‘‘protective’’ purpose.
`A notable observation not supporting the central importance of CYP3A is
`that no life-threatening medical consequences are known to arise from ex-
`tended treatment with highly potent CYP3A inhibitors such as ketoconazole10
`and ritonavir.11 Both of these drugs produce what amounts to a chemically-
`induced ‘‘poor metaboliser’’ phenotype. The antifungal agent ketoconazole has
`been available since the early 1980s, and the viral protease inhibitor ritonavir
`since the mid 1990s. Extended exposure to ketoconazole is associated with
`antiandrogenic effects, and a lipodystrophy syndrome has been linked to ex-
`tended treatment with ritonavir and other protease inhibitors.12 These sequelae
`are possibly attributable to CYP3A inhibition, but this is not proven. In any
`case, the sequelae are significant but not ominous medical consequences, raising
`questions as to how essential it is in adults for CYP3A phenotype to be
`maintained within a ‘‘normal’’ range.
`The unique protective function of hepatic and enteric CYP3A enzymes
`may produce complications in drug development and clinical therapeutics.
`New chemical entities found to be complete or nearly complete substrates for
`clearance by CYP3A may actually be dropped from subsequent develop-
`ment. Such candidates often are seen as facing therapeutic or competitive
`obstacles, largely attributable to the possibility of drug interactions with other
`
`3
`
`

`

`The CYP3 Family
`
`357
`
`agents that are CYP3A inducers or inhibitors, and restrictive labelling that
`could result. Some marketed drugs, such as terfenadine, astemizole and
`cisapride were actually withdrawn for that reason.13–16 Similar concerns
`apply to new drug candidates that themselves are found to be significant
`CYP3A inducers or inhibitors. Some CYP3A substrate drugs that meet a
`pressing medical need, such as the viral protease inhibitors saquinavir and
`lopinavir, have been brought through the development process and approved
`for clinical use despite the drawbacks and obstacles. These two drugs have the
`disadvantage of poor net oral bioavailability, due to some combination of
`hepatic/enteric presystemic extraction together with enteric efflux transport by
`P-glycoprotein (P-gp). So significant is this problem that saquinavir and lopi-
`navir are, with few exceptions, combined with the CYP3A/P-gp inhibitor
`ritonavir for purposes of ‘‘boosting’’ or ‘‘augmentation’’ of oral bioavail-
`ability.17–19
`The topic of this review is the translational pharmacology of CYP3A en-
`zymes in humans. The principal focus will be a number of contemporary and/or
`controversial issues of current importance.
`
`11.2 Pharmacogenomics
`
`The four genes encoding the relevant CYP3A proteins (CYP3A4, 5, 7, and 43)
`are all located in a 231-kb cluster on chromosome band 7q21-q22.1. The re-
`lationship of CYP3A genomic variants to CYP3A protein expression and ac-
`tivity in vitro and in vivo has been extensively investigated over the last
`decade.20–25 The prevailing contemporary interpretation of the existing data
`base is that a great deal of information has been generated, collectively dem-
`onstrating very little in the way of meaningful associations between CYP3A
`genotype and in vivo phenotypic metabolic activity. Furthermore, there is
`substantial disconnect between in vitro studies and human pharmacokinetic
`studies in vivo.
`CYP3A7 is a foetal enzyme, and its expression is silenced after birth. Reports
`exist of persistently detectable CYP3A7 mRNA in adulthood, but there is no
`available evidence that CYP3A7 is of any significance in terms of in vivo drug-
`metabolising activity. CYP3A43 is an adult enzyme, and is known to be lo-
`calised in prostate.26,27 There is no known functional significance of CYP3A43
`identified to date.
`CYP3A4 is the dominant isoform in humans. Numerous studies have dem-
`onstrated that CYP3A4 is the most abundant of the human hepatic cyto-
`chromes P450 and also is (along with CYP3A5) the only cytochrome P450 of
`functional importance in the enteric mucosa of the gastrointestinal tract. Ex-
`pression of hepatic and enteric CYP3A4 is not coordinately regulated; levels
`expressed at the two sites are not intercorrelated.28
`Many single nucleotide polymorphisms (SNPs) have been identified in the
`CYP3A4 locus.29–32 Some of these SNPs are associated with reduced functional
`metabolic capacity in various in vitro systems.33 However,
`there is no
`
`4
`
`

`

`358
`
`Chapter 11
`
`substantial evidence that any CYP3A4 SNP is associated with clinically im-
`portant differences in clearance of CYP3A substrates in vivo.31,34–42 Consistent
`with this is the observation that phenotypic distribution of CYP3A metabolic
`activity in vitro and in vivo is unimodal rather than multimodal (bimodal or
`trimodal),43–46 thereby essentially excluding the existence of common ‘‘null’’
`alleles coding for low or zero protein expression or function. Finally, all
`CYP3A4 SNPs with demonstrated functional consequences are of low preva-
`lence in the population. Nonetheless it cannot be fully excluded that one or
`more SNPs (such as CYP3A4*20)33 could account for low CYP3A metabolic
`activity noted in some unusual ‘‘outlying’’ subjects in a number of human
`studies.46–48
`The CYP3A4*1B polymorphism is located in the 50-regulatory region
`of the CYP3A4 gene. This SNP received considerable attention when it was
`first described, due to the statistical association with poor outcome in patients
`with prostate cancer.49 A link of CYP3A4*1B to altered testosterone metabo-
`lism was speculated as possible mechanism, but the study did not evaluate
`plasma testosterone or any other index of metabolic activity. A few subsequent
`reports provided some evidence supporting associations of the CYP3A4*1B
`polymorphism with human disease, but other studies have not.50–57 In care-
`ful clinical phenotype-genotype studies using CYP3A probe substrates
`such as midazolam, a detectable phenotypic importance of CYP3A4*1B
`was not found.31,34–40,58 Thus the mechanism of the linkage of the CYP3A4*1B
`polymorphism to human disease,
`if a link actually exists, remains unex-
`plained. It has been postulated that a partial
`linkage disequilibrium of
`CYP3A4*1B with CYP3A5*3 may explain the association,
`inasmuch as
`CYP3A5 is found in prostate and may play a role in local androgen
`metabolism.59 However this is a speculative explanation without experimental
`support.
`CYP3A5 shares approximately 90% sequence homology with CYP3A4.
`CYP3A5 is expressed in liver and gastrointestinal enteric mucosa, as well as a
`number of other tissues including prostate and kidney.60 In contrast to
`CYP3A4, there is evidence that CYP3A5 is polymorphically expressed.59–63
`The CYP3A5*3 variant, which is the most prevalent form in many human
`populations, actually confers low or zero functional CYP3A phenotypic ac-
`tivity, whereas the less prevalent CYP3A5*1 codes for active protein. Ex-
`pression of immunoactive CYP3A5 protein in human liver samples is consistent
`with genotype. Since CYP3A5*1 has greater prevalence in the African-
`American population compared with Caucasians, the possibility is raised –
`politically alluring to some – that a racial difference in CYP3A phenotype may
`exist. This has been supported by in vitro studies of CYP3A metabolic activity
`in genotyped human liver samples. However, the outcomes of clinical studies
`are not consistent with this scheme. Human CYP3A metabolic phenotype is, at
`most, weakly associated with CYP3A5 genotype, and there is little or no
`evidence of a race-associated difference in metabolic activity.31,34–36,41,64–66
`Again this demonstrates the in vitro-in vivo disconnect described previously.
`At present, the abundance of evidence indicates that CYP3A5, and its
`
`5
`
`

`

`The CYP3 Family
`
`359
`
`Figure 11.2
`
`In vitro intrinsic clearance of four CYP3A substrates by recombinant
`CYP3A4 and recombinant CYP3A5 (see reference 70).
`
`polymorphic variants, have at most a minor effect on human CYP3A pheno-
`type in vivo.61,67 However, there may be a few substrates, such as alprazolam,
`for which CYP3A5 genotype has a more significant role in the modulation of
`drug clearance.68,69
`The value of comparative in vitro studies of CYP3A4 and CYP3A5 has been
`enhanced by the availability of relatively specific, commercially-available anti-
`bodies for separate immunoquantitation of the two proteins, as well as distinct
`recombinant enzymes for evaluation of enzyme function. Using these tools, it has
`been established, for example, that in vitro intrinsic metabolic clearance for a
`number of substrates (midazolam, triazolam, testosterone, nifedipine) via re-
`combinant CYP3A4 exceeds that for CYP3A5 (Figure 11.2).70 In addition, in-
`hibitory potencies of an index inhibitor (ketoconazole), as well as a number of
`other inhibitors, are greater for CYP3A4 compared with CYP3A5.70,71 However,
`a major obstacle for translation of in vitro results to humans is that all known
`CYP3A substrates are metabolised by both CYP3A4 and CYP3A5, and no
`substrate is specific for one or the other. Therefore, the relative in vivo contri-
`butions of the two enzymes to net CYP3A metabolic activity cannot be resolved
`by customary probe substrate methodology.
`The wide individual variability in CYP3A expression and function is
`commonly discussed in the literature, but the metrics used to quantitate vari-
`ation generally have been poorly defined. In many sources the term ‘‘fold
`variation’’ is used. A ‘‘50-fold variation in CYP3A expression’’ may be claimed,
`but the term remains undefined. If it refers to the ratio of maximum value
`divided by minimum value, this is not a useful metric of variation, since it is
`driven by the outlying values at both ends. There is a need for a consensus
`in the field on the meaning of ‘‘variability’’ and how it is best quantitated.
`We suggest the use of the coefficient of variation (CV), calculated as the
`arithmetic standard deviation divided by the arithmetic mean, expressed as
`percent. In any case, it is clear that ‘‘individual variability’’ in CYP3A ex-
`pression and function observed among human liver microsomal samples
`in vitro substantially exceeds variability in CYP3A metabolic phenotype
`in vivo.3,72 This
`is probably attributable to the additional variability
`
`6
`
`

`

`360
`
`Chapter 11
`
`superimposed through acquisition and processing of human liver samples for
`in vitro study: variation in the time between donor death and tissue harvesting/
`preservation; source of the sample (transplant donor vs. autopsy vs. surgical);
`mode of demise of deceased donors (acute illness or trauma vs. chronic disease);
`and the donor’s medication exposure.
`
`11.3 Enzyme Kinetics of CYP3A
`
`Most work on the enzyme kinetic features of CYP3A enzymes has focussed on
`CYP3A4. This enzyme is generally characterised as low-affinity and high-
`capacity (high Km, high Vmax). There are exceptions for some substrates such as
`midazolam, which has a Km value in the low micromolar range for the principal
`metabolic pathway (1-hydroxylation).73,74 However for the majority of CYP3A
`substrates, the apparent Km in vitro greatly exceeds the usual range of con-
`centrations encountered in vivo, with few clinical examples of non-linear
`kinetics.
`For many substrates, in vitro studies assessing reaction velocity in relation to
`substrate concentration yield a sigmoidal pattern, in contrast to Michaelis-
`Menten (hyperbolic) kinetics. The sigmoidal profile is consistent with homo-
`tropic (positive) cooperativity or autoactivation,
`in which binding of one
`substrate molecule facilitates binding of the next molecule, effectively increas-
`ing affinity to that vacant site.75–83 This also implies multiple ligand binding
`sites, and changes in enzyme conformation attributable to ligand binding.
`Affected substrates include amitriptyline, carbamazepine, diazepam, triazolam,
`alprazolam, nifedipine, progesterone and testosterone.
`Sigmoidal kinetics attributable to positive cooperativity complicates the in-
`terpretation of in vitro data, and in vivo extrapolation for purposes of quanti-
`tative reaction phenotyping.84–88 With a sigmoidal kinetic pattern, the substrate
`concentration yielding a reaction velocity equal to 50% of Vmax (S50) is not
`conceptually equivalent to Km in a Michaelis-Menten model. In addition, the
`concept of in vitro intrinsic clearance – calculated as Vmax/Km in the Michaelis-
`Menten model – is not applicable to sigmoidal kinetics. The quantity Vmax/S50
`has been used as a surrogate for intrinsic clearance in a number of studies, but
`this is applied empirically without theoretical support.
`Heterotropic activation is another possible consequence of multiple substrate
`binding sites. With this mechanism, binding of one substrate can increase
`binding affinity for a different substrate. Among the first recognised ‘‘acti-
`vators’’ was 7,8-benzoflavone (a-naphthoflavone),89 and a number of others
`have been subsequently identified.
`Inhibitory allosteric effects are also possible. When substrate binding inhibits
`subsequent binding of other substrate molecules (negative homotropy), the
`observed consequence is substrate inhibition.75–83 Complex situations may arise
`involving concurrent heterotropic activation and inhibition. An example is
`the concurrent effect of testosterone, simultaneously activating triazolam
`a-hydroxylation while inhibiting 4-hydroxylation.70,90
`
`7
`
`

`

`The CYP3 Family
`
`361
`
`It should be emphasised that all of these findings derive from molecular and
`in vitro models. There are no known clinical consequences or correlates of the
`sigmoidal kinetic profile of a number of CYP3A substrates.
`
`11.4 Individual Variability in CYP3A
`Metabolic Phenotype
`
`Despite implications to the contrary, individual variation in human CYP3A
`metabolic phenotype in vivo is in the same range as that for other CYP enzymes.
`Among relatively homogenous populations, such as groups of healthy, medi-
`cation-free young male volunteers that typically participate in clinical phar-
`macokinetic studies, the between-subject CV for area under the plasma
`concentration curve (AUC) or clearance of a CYP3A probe such as midazolam
`or triazolam generally will not exceed 40–60%.45,47,91–101 Similar variability is
`observed for substrates of CYP2D6,102–105 CYP2C9106–109 and CYP1A2.110–113
`In contrast to between-subject variation, within-subject variation in clearance of
`a CYP3A substrate is in the range of 10–15%.114 This has been determined
`in studies in which the same dose of the same substrate (midazolam) was
`administered to the same subject on multiple occasions.114
`Clearance values of different CYP3A probe substrates across individuals are
`not necessarily highly correlated with or ‘‘predictive’’ of each other.98,115–120
`This is because of the complex relationship of net CYP3A metabolic pheno-
`type to the intrinsic clearance characteristics of the specific substrate, the route
`of administration, and the relative dependence of net clearance on hepatic
`versus enteric CYP3A activity. For low-clearance substrates, such as alprazo-
`lam and erythromycin, net clearance will be dependent primarily upon hepatic
`CYP3A, and essentially independent of route of administration and hepatic
`blood flow. For higher clearance substrates, such as buspirone, midazolam,
`and triazolam, intravenous clearance will depend on a combination of hepatic
`metabolism and hepatic blood flow, whereas clearance after oral dosage
`will depend on a combination of hepatic and enteric extraction, being in-
`dependent of hepatic blood flow (Figure 11.3). Since each individual
`substrate is characterised by a unique combination of these factors,
`it is
`not surprising that phenotypes derived from different substrates are not pre-
`dictive of each other. In any case, CYP3A probes that are also substrates
`for transport by P-gp (such as erythromycin and cyclosporine) are probably not
`the best choices, since their net disposition will reflect both CYP3A and P-gp
`effects.121–123
`The possibility of CYP3A metabolic phenotyping without the use of an ex-
`ogenous probe has received considerable attention. The ratio of urinary ex-
`cretion of 6-b-hydroxycortisol divided by cortisol
`is the most extensively
`evaluated endogenous index;124,125 the outcomes are equivocal at best. The
`principal drawback is that the 6-b-hydroxycortisol/cortisol excretion ratio,
`unlike clearance of other exogenous index substrates, can be quite variable
`
`8
`
`

`

`362
`
`Chapter 11
`
`INTRAVENOUS
`CLEARANCE
`
`ORAL
`CLEARANCE
`
`HEPATIC
`BLOOD FLOW
`
`HEPATIC
`CYP3A
`
`ENTERIC
`CYP3A
`
`Figure 11.3 Determinants of metabolic clearance of CYP3A drug substrates when
`given intravenously or orally. I.V. clearance depends on hepatic CYP3A
`and (for high-clearance drugs) on hepatic blood flow. Oral clearance
`depends on a combination of enteric and hepatic CYP3A, but is not
`dependent on hepatic blood flow.
`
`Urinary 6-beta-OH-cortisol/cortisol ratio
`
`r2= 0.34
`Mean CV= 21%
`(range: 1-77%)
`
`0
`
`2
`
`4
`
`6
`8
`First Trial
`
`10
`
`12
`
`14
`
`14
`
`12
`
`10
`
`8
`
`6
`
`4
`
`2
`
`0
`
`Second Trial
`
`Figure 11.4 Within-subject variability in the urinary 6b-hydroxycortisol/cortisol ratio
`determined in the same individuals on two occasions (see reference 128).
`
`from time to time within the same individual (Figure 11.4).126–129 As such, the
`ratio may adequately reflect large differences in CYP3A phenotype as caused,
`for example, via extensive enzyme induction by rifampin. However the ratio
`appears to be relatively insensitive to more subtle individual differences in
`
`9
`
`

`

`The CYP3 Family
`
`363
`
`metabolic activity. Taken together, the evidence does not support the use of the
`6-b-hydroxycortisol/cortisol ratio as a CYP3A phenotypic index.
`
`11.5 Age and Gender Effects on CYP3A Phenotype
`
`The possible modulation of CYP3A metabolic activity by age and gender is of
`both scientific and public health importance. With CYP3A substrate clearance
`values having a CV in the range of 50–60% of the mean, identifying individually
`appropriate dosage levels for therapeutic agents metabolised by CYP3A may
`necessarily involve an iterative process of trial dosage, observation and dosage
`revision as necessary. More precise prediction of clearance based on demo-
`graphic factors such as age and gender could facilitate accuracy of initial dose
`selection, with reduction of the need for revision through trial and error.
`
`11.5.1
`
`In vitro and Experimental Studies
`
`The data base from in vitro studies of human liver samples is large, but the
`findings are inconsistent and inconclusive. Some studies show reduced
`expression of immunoactive CYP3A and/or reduced metabolic activity in liver
`samples from older humans.130–132 Other studies show little or no effect of
`age.3,133,134 The intrinsic limitations of in vitro systems have already been dis-
`cussed (vide supra), and are especially pertinent to assessment of advancing age
`as an independent determinant of protein expression or metabolic function.
`Demise in younger donors is more likely to result from trauma or sudden
`unexpected illness. Preservation of ‘‘normal’’ liver is more likely, and metabolic
`activity could actually be increased (induced) by factors such as high-dose
`exogenous corticosteroids used to treat head trauma. In contrast, elderly
`donors are more likely to have suffered from extended illness (cancer, cardio-
`vascular disease, Alzheimer’s etc.) with associated pharmacologic treatment,
`poor nutrition, inactivity and general debility. Studies of gender effects on
`CYP3A in vitro likewise are inconsistent and inconclusive.131–136 It is important
`to note that should age and gender be independent modulators of CYP3A
`expression/function, outcomes of in vitro ‘‘population’’ studies could be
`confounded, depending on the age/gender composition of the samples available
`to the investigator.
`immunoreactive
`In studies of Fischer-344 rats, reduced expression of
`CYP3A, and reduced CYP3A metabolic activity, has been observed in liver
`samples obtained from aging animals.137 This was not explained by reduced
`testosterone concentrations in the older animals; enteric CYP3A, on the other
`hand, was not influenced by age. The rodent model may not be applicable to
`humans for many reasons, including that the CYP3A enzymes are different
`between the species. Nonetheless, the animal model could provide the oppor-
`tunity to study the regulatory mechanisms through which CYP3A expression
`declines with age.
`
`10
`
`

`

`364
`
`11.5.2 Clinical Studies
`
`Chapter 11
`
`A large number of clinical pharmacokinetic studies evaluated the influence of
`age and gender on human CYP3A phenotype, and on age-related changes in
`drug disposition in general.138–145 Numerous CYP3A substrates have been
`studied, and trial designs vary widely in terms of sample size, method of subject
`selection and approach to statistical analysis. Reviews of the topic generally are
`not fully comprehensive, largely because the scope of the published literature is
`so large. The conclusions of a review may be dependent on which specific
`studies are included in the evaluation and which are omitted.
`Based on a literature review that is as complete as possible, there is extensive,
`but not completely consistent, evidence that weight-normalised metabolic
`clearance of CYP3A substrates declines with age (Table 11.2).138 There is also
`evidence that effects of age and gender are not independent; in many studies the
`decrease in clearance with increasing age among male subjects is far greater
`than the clearance decrease in women. For this reason, studies that combine
`data for men and women may obscure an age effect. Enhanced pharmacody-
`namic effects of CYP3A substrate drugs in the elderly have also been demon-
`strated in many studies. This may be attributable to the higher plasma drug
`concentrations in elderly subjects, but also from an independent effect of age on
`intrinsic drug sensitivity.146 The net clinical implication is that lower doses of
`most CYP3A substrate drugs are recommended for the elderly.
`
`for
`Table 11.2 CYP3A substrate drugs
`which reduced clearance in the
`elderly has been demonstrated
`in clinical studies.a
`
`Adinazolam
`Alfentanyl
`Alprazolam
`Amlodipine
`Bromazepam
`Clarithromycin
`Diltiazem
`Erythromycin
`Felodipine
`Midazolam
`Nefazodone
`Nifedipine
`Prednisolone
`Quinidine
`Tirilizad
`Trazodone
`Triazolam
`Verapamil
`Zolpidem
`Zopiclone
`
`aSee Reference 138.
`
`11
`
`

`

`The CYP3 Family
`
`365
`
`The data on gender effects on CYP3A phenotype are murky.138,145,147–154 We
`have concluded that most studies comparing young men and young women
`show minimal or no difference in weight-normalised clearance. Some studies
`show higher clearance in women, but generally the gender difference is small in
`magnitude and unlikely to be of clinical importance. There is little or no evi-
`dence to suggest that use of hormonal contraceptive preparations or oestrogen
`supplementation has a significant effect on CYP3A phenotype.
`
`11.6 Drug Interactions with CYP3A Substrates
`
`Pharmacokinetic drug interactions involving drugs metabolised by CYP3A are
`of major contemporary scientific and public health importance.155–157 The
`visibility of the topic increased substantially in the mid to late 1990s, in con-
`nection with adverse reactions to the nonsedating antihistamine terfenadine
`(Seldane).13–16 Under usual circumstances, terfenadine itself is a prodrug, being
`nearly completely biotransformed to fexofenadine by CYP3A enzymes through
`presystemic extraction. However, in rare patients in whom CYP3A metabolic
`activity was depressed due to co-treatment with ketoconazole or erythromycin,
`significant concentrations of intact terfenadine did appear in the circulation,
`producing serious and even fatal cardiac arrhythmias due to its arrhythmogenic
`properties. Terfenadine was withdrawn from the market, as were astemizole
`and cisapride, two other drugs with similar properties. The calcium channel
`antagonist mibefradil was likewise withdrawn from clinical use, being a strong
`CYP3A inhibitor. While epidemiologic studies demonstrate that clinically im-
`portant or hazardous drug interactions involving CYP3A actually are unusual
`relative to the prevalence of treatment with multiple drugs,158 the sensitivity of
`the public and the scientific community to the drug interaction issue has
`nonetheless been heightened. In the process of drug discovery and develop-
`ment, drug candidates found to be predominant CYP3A substrates, and/or
`significant CYP3A inhibitors/inducers, frequently are dropped from further
`development due to actual or perceived liabilities or hazards attributable to
`drug interactions.
`Drug interactions with CYP3A substrates may occur via metabolic inhibition
`or metabolic induction. These are mechanistically different events with different
`clinical consequences (Table 11.3).
`
`11.6.1 Drug Interactions via Metabolic Inhibition
`
`A number of drugs used in clinical practice are inhibitors of CYP3A metabolic
`activity (Table 11.4). Some of these, including the azole antifungal agents
`ketoconazole and itraconazole, and the viral protease inhibitor ritonavir are
`highly potent inhibitors, such that coadministration with CYP3A substrate
`drugs could produce important and possibly hazardous impairment of clearance
`and elevation of plasma levels.10,11 New drug candidates that are CYP3A
`
`12
`
`

`

`366
`
`Chapter 11
`
`Table 11.3 Comparison of inhibition and induction.
`
`Mechanism:
`
`Inhibition
`
`Induction
`
`Direct chemical
`effect on enzyme
`
`Increased protein
`synthesis
`
`Time-course of onset
`Time-course of offset
`In vitro model
`Metric for inhibition or induction potency
`
`Slow
`Rapid
`Slow
`Rapid
`Cell homogenates Cultured hepatocytes
`Not established
`Ki or IC50
`
`Clinical consequences:
`
`Effect on clearance of victim
`Effect on plasma levels of victim
`Principal therapeutic concern
`
`Decreased
`Elevated
`Toxicity
`
`Increased
`Reduced
`Reduced efficacy
`
`Table 11.4 Representative CYP3A inhibitors.
`
`Strong
`
`Clarithromycin
`Erythromycin
`Itraconazole
`Ketoconazole
`Nefazodone
`Ritonavir
`Telithromycin
`Voriconazole
`
`Moderate or weak
`
`Atazanavir
`Delavirdine
`Fluconazole
`Fluvoxamine
`Grapefruit juice
`Lopinavir
`Nelfinavir
`Saquinavir
`
`substrates generally are subjected to ketoconazole or ritonavir drug interaction
`studies, designed to delineate the ‘‘worst case’’ drug interaction scenario. A
`number of other inhibitors are less potent, and will produce inhibitory drug
`interactions that are quantitatively smaller, and may or may not be of clinical
`importance. Ultimately, the magnitude and clinical importance of an inhibitory
`drug interaction will depend on the properties of the substrate as well as the
`potency of and systemic exposure to the inhibitor. For substrates (such as
`midazolam) that ordinarily undergo high presystemic extraction after oral
`dosage, CYP3A inhibition may profoundly increase oral bioavailability through
`impaired presystemic extraction, resulting in very large increases in substrate
`plasma concentrations. These interactions may be clinically hazardous or – in
`the case of deliberate ‘‘boosting’’ – elevate plasma concentrations of a drug with
`very low bioavailability, such as saquinavir or lopinavir, into a theoretically
`effective range.17–19 In contrast, CYP3A substrate drugs with low presystemic
`extraction and high oral bioavailability under control conditions will also have
`impaired clearance with coadministration of a potent CYP3A inhibitor, but the
`effect will be evident mainly as a prolongation of half-life rather than a change in
`bioavailability. Comparative studies of midazolam, triazolam, and alprazolam
`coadministered with ketoconazole illustrate this point (Table 11.5).73,159–162
`
`13
`
`

`

`The CYP3 Family
`
`367
`
`Table 11.5 Comparative pharmacokinetics of oral alprazolam, triazolam, and
`midazolam, with and without coadministration of ketoconazole.a
`
`Principal metabolites
`Ketoconazole in vitro Ki (mM)
`vs. a-OH pathway
`vs. 4-OH pathway
`
`Kinetic properties without inhibitor
`(control condition)
`Cmax (ng/ml)
`Elimination half-life (hr)
`Oral clearance (ml/min)
`
`Kinetic properties with ketoconazole
`coadministration
`Cmax (ng/ml)
`Elimination half-life (hr)
`Oral clearance (ml/min)
`
`Ratio of total AUC (with ketoconazole
`divided by control)
`
`aSee References 73, 159–162.
`
`Alprazolam
`(1.0 mg)
`
`Triazolam
`(0.25 mg)
`
`Midazolam
`(6.0 mg)
`
`a-OH,