Pharmaceutical Research, Vol. 22, No. 1, January 2005 (© 2005)
`DOI: 10.1007/s11095-004-9004-4
`
`Commentary
`
`Predicting Drug Disposition via Application of BCS: Transport/Absorption/
`Elimination Interplay and Development of a Biopharmaceutics Drug
`Disposition Classification System
`
`Chi-Yuan Wu1,2 and Leslie Z. Benet1,3
`
`Received September 2, 2004; accepted October 12, 2004
`
`The Biopharmaceutics Classification System (BCS) was developed to allow prediction of in vivo phar-
`macokinetic performance of drug products from measurements of permeability (determined as the
`extent of oral absorption) and solubility. Here, we suggest that a modified version of such a classification
`system may be useful in predicting overall drug disposition, including routes of drug elimination and the
`effects of efflux and absorptive transporters on oral drug absorption; when transporter-enzyme interplay
`will yield clinically significant effects (e.g., low bioavailability and drug-drug interactions); the direction,
`mechanism, and importance of food effects; and transporter effects on postabsorption systemic drug
`concentrations following oral and intravenous dosing. These predictions are supported by a series of
`studies from our laboratory during the past few years investigating the effect of transporter inhibition
`and induction on drug metabolism. We conclude by suggesting that a Biopharmaceutics Drug Dispo-
`sition Classification System (BDDCS) using elimination criteria may expand the number of Class 1
`drugs eligible for a waiver of in vivo bioequivalence studies and provide predictability of drug disposition
`profiles for Classes 2, 3, and 4 compounds.
`
`KEY WORDS: BCS; BDDCS; disposition; drug interactions; food effects; routes of elimination; trans-
`porter-enzyme interplay.
`
`INTRODUCTION
`
`Amidon and co-workers (1) recognized that the funda-
`mental parameters controlling the rate and extent of oral drug
`absorption were the drug’s aqueous solubility and gastroin-
`testinal permeability. They devised a Biopharmaceutics Clas-
`sification System (BCS) that categorized drugs into four
`classes according to their solubility and permeability (ex-
`pressed as the extent of oral drug absorption) as depicted in
`Fig. 1. In 2000, the FDA used the BCS system as a science-
`based approach to allow waiver of in vivo bioavailability and
`bioequivalence testing of immediate-release solid dosage
`forms for Class 1 high-solubility, high-permeability drugs
`when such drug products also exhibit rapid dissolution (2).
`At its core, the BCS is an experimental model, centrally
`embracing permeability and solubility, with qualifications re-
`lated to pH and dissolution. The objective of the BCS is to
`predict in vivo pharmacokinetic performance of drug prod-
`ucts from measurements of permeability and solubility. A
`drug substance is considered “highly soluble” when the high-
`est dose strength is soluble in 250 ml or less of aqueous media
`over a pH range of 1–7.5 at 37°C. A drug substance is con-
`
`1 Department of Biopharmaceutical Sciences, University of Califor-
`nia San Francisco, San Francisco, California 94143, USA.
`2 Current Address: Bristol-Myers Squibb, Pharmaceutical Research
`Institute, Princeton, New Jersey 08543, USA.
`3 To whom correspondence should be addressed. (e-mail: benet@itsa.
`ucsf.edu)
`
`sidered to be “highly permeable” when the extent of the ab-
`sorption (parent drug plus metabolites) in humans is deter-
`mined to be ⱖ90% of an administered dose based on a mass
`balance determination or in comparison to an intravenous
`reference dose. In Table I, we have assembled a list of com-
`pounds in the four BCS classes, predominantly gathered from
`the literature (1,3–18) but judiciously edited. With respect to
`oral bioavailability, it is generally believed that the frame-
`work of the BCS could serve the needs of the earliest stages
`of discovery research. In this manuscript, we demonstrate that
`categorizing drugs into the four classes represented by BCS
`solubility and permeability criteria may provide significant
`new insights to the pharmaceutical scientific community. This
`classification system may be useful in predicting routes of
`elimination, effects of efflux and absorptive transporters on
`oral absorption, when transporter-enzyme interplay will yield
`clinically significant effects such as low bioavailability and
`drug-drug interactions, the direction and importance of food
`effects, and transporter effects on postabsorption systemic
`levels following oral and intravenous dosing. We propose that
`a modest revision of the BCS criteria may result in a classi-
`fication system that yields predictability of in vivo disposition
`for all four classes, as well as increasing the number of Class
`1 drugs eligible for bioequivalence study waivers.
`As we were preparing this manuscript, the extensive
`evaluation of the WHO Essential Medicines List in terms of
`BCS classification based on measured solubility and perme-
`ability/absorption data was published (18). We have modified
`the manuscript to include many of the compounds evaluated
`
`11
`
`0724-8741/05/0100-0011/0 © 2005 Springer Science+Business Media, Inc.
`
`Par Pharm., Inc.
`Exhibit 1045
`Page 001
`
`

`

`12
`
`Wu and Benet
`
`and major routes of elimination. Many highly protein bound
`acidic Class 1 and Class 2 compounds exhibit very low vol-
`umes of distribution (e.g., valproic acid, ibuprofen). It would
`be incorrect, however, to conclude that correction for protein
`binding would give a better prediction of the relative size of
`the volume of distribution in comparing Classes 1 and 2 com-
`pounds with Classes 3 and 4 drugs. In fact, our analysis dem-
`onstrates that the generally larger volumes of distribution for
`Class 1 and Class 2 compounds when compared to Class 3 and
`Class 4 compounds is independent of the degree of protein
`binding.
`
`Most New Molecular Entities Are Class 2 Compounds
`
`New molecular entities (NMEs) today are frequently
`large-molecular-weight, lipophilic, poorly water-soluble com-
`pounds that most often fall into BCS Class 2. Lipinski et al.
`(22) pointed out that leads obtained through high-throughput
`screening (HTS) tend to have higher molecular weights and
`greater lipophilicity than leads in the pre-HTS era. Lipinski’s
`Rule of 5 was developed to set “drugability” guidelines for
`NMEs (23). In the drug discovery setting, the Rule of 5 pre-
`dicts that poor absorption or permeation is more likely when
`there are more than 5 H-bond donors, 10 H-bond acceptors,
`the molecular weight is greater than 500, and the calculated
`Log P (CLog P) is greater than 5. However, Lipinski specifi-
`cally states that the Rule of 5 only holds for compounds that
`are not substrates for active transporters (22,23). When the
`Rule of 5 was developed, information about drug transporters
`was very limited. We believe that almost all drugs are sub-
`strates for some transporter. Studies to date have not been
`able to show this because we are just beginning to gain the
`knowledge and tools that allow investigation of substrates for
`uptake transporters. In addition, unless a drug molecule can
`passively gain intracellular access, it is not possible to simply
`investigate whether the molecule is a substrate for efflux
`transporters.
`Lipinski has noted that the Rule of 5 was intended as a
`very crude filter (24). Thus, it is not surprising that predictions
`based only on solubility and Log P or CLog P may frequently
`be in error, often because most drugs may be substrates for
`some transporter. We note that a recent evaluation of the
`provisional biopharmaceutical classification of WHO essen-
`tial drugs (25) reported a generally good correlation between
`in silico parameters and BCS classification; however, some
`obvious misclassifications occurred. For example, acetamino-
`phen (bioavailability ⳱ 88%), dapsone (93%), and theoph-
`ylline (96%), all highly metabolized drugs, are listed as Class
`4 compounds based only on physicochemical criteria (25), as
`opposed to their Classes 1 and 2 listings in Table I.
`
`Cautions
`
`Prior to making further predictions related to trans-
`porter-enzyme interactions, food effects and drug-drug inter-
`actions, we wish to provide the following cautions.
`
`a) There will always be exceptions to the broad general
`rules presented here (e.g., the Class 2 compound digoxin does
`not undergo extensive hepatic metabolism in humans, but it
`does in the rat). As research scientists, we find exceptions to
`predictability (and unexpected events) more intriguing and
`challenging than the expected or predictable events. As in
`
`Fig. 1. The Biopharmaceutics Classification System (BCS) as defined
`by the FDA (2) after Amidon et al. (1).
`
`in that work. We agree with most of the classifications as-
`signed, but not all, as our paper expands the utility of the
`classification to drug disposition. We have added comments
`about some of these differences throughout the manuscript.
`
`Predicting Routes of Drug Elimination
`
`Examining the drug substances listed in the four BCS
`classes in Table I, it becomes obvious that Class 1 and Class
`2 compounds are eliminated primarily via metabolism,
`whereas Class 3 and Class 4 compounds are primarily elimi-
`nated unchanged into the urine and bile (Fig. 2). We are
`unaware that this simple categorization under BCS has pre-
`viously recognized the correlation and fact that the high per-
`meability of the Classes 1 and 2 compounds allows ready
`access to the metabolizing enzymes within hepatocytes, al-
`though Smith (19) has noted that more permeable lipophilic
`compounds make good substrates for cytochrome P450
`(CYP) enzymes. Note that the differential permeability char-
`acteristics defined under BCS do not necessarily reflect dif-
`ferences in permeability into hepatocytes, as a number of
`Class 3 and Class 4 compounds are eliminated into the bile.
`Rather, the high vs. low permeability designation reflects dif-
`ferences in access to the metabolizing enzymes within the
`hepatocytes.
`For the 130 drugs/compounds listed in Table I, only 13 of
`the substances do not have readily accessible, critically evalu-
`ated pharmacokinetic parameters (20,21). Upon reviewing
`the disposition characteristics of the Class 3 and Class 4 drugs
`listed in Table I, all but mebendazole are eliminated predomi-
`nantly in the unchanged form by the renal or biliary route. We
`suspect that mebendazole is misclassified, as it is extensively
`metabolized [note that Lindenberg et al. (18) most recently
`listed mebendazole as either Class 2 or Class 4]. We propose
`that for the purposes of defining the BCS classification for
`predicting drug disposition, the extent of metabolism may be
`a better predictor than the 90% absorption characteristic.
`One might suspect that the high-permeability com-
`pounds (Class 1 and Class 2) should have higher volumes of
`distribution than the low-permeability Class 3 and Class 4
`compounds. When evaluating the published pharmacokinetic
`characteristics (20,21), we observed such a trend, but the con-
`cordance is not even close to that found between BCS class
`
`Par Pharm., Inc.
`Exhibit 1045
`Page 002
`
`

`

`Application of BDDCS to Predict Drug Disposition
`
`13
`
`Table I. Biopharmaceutics Classification System (BCS) Substratesa
`
`High solubility
`
`Low solubility
`
`Class 2
`AmiodaroneI
`AtorvastatinS,I
`AzithromycinS,I
`CarbamazepineS,I
`Carvedilol
`ChlorpromazineI
`CisaprideS
`CiprofloxacinS
`CyclosporineS,I
`Danazol
`Dapsone
`Diclofenac
`Diflunisal
`DigoxinS
`ErythromycinS,I
`Flurbiprofen
`Glipizide
`GlyburideS,I
`Griseofulvin
`Ibuprofen
`IndinavirS
`Indomethacin
`
`ItraconazoleS,I
`KetoconazoleI
`LansoprazoleI
`LovastatinS,I
`Mebendazole
`Naproxen
`NelfinavirS,I
`Ofloxacin
`Oxaprozin
`Phenazopyridine
`PhenytoinS
`Piroxicam
`RaloxifeneS
`RitonavirS,I
`SaquinavirS,I
`SirolimusS
`SpironolactoneI
`TacrolimusS,I
`TalinololS
`TamoxifenI
`TerfenadineI
`Warfarin
`
`Class 4
`Amphotericin B
`Chlorthalidone
`Chlorothiazide
`Colistin
`CiprofloxacinS
`Furosemide
`Hydrochlorothiazide
`Mebendazole
`Methotrexate
`Neomycin
`
`Ketorolac
`Ketoprofen
`Labetolol
`LevodopaS
`LevofloxacinS
`LidocaineI
`Lomefloxacin
`Meperidine
`Metoprolol
`Metronidazole
`MidazolamS,I
`Minocycline
`Misoprostol
`NifedipineS
`Phenobarbital
`Phenylalanine
`Prednisolone
`PrimaquineS
`Promazine
`PropranololI
`QuinidineS,I
`Rosiglitazone
`Salicylic acid
`Theophylline
`Valproic acid
`VerapamilI
`Zidovudine
`
`FexofenadineS
`Folinic acid
`Furosemide
`Ganciclovir
`Hydrochlorothiazide
`Lisinopril
`Metformin
`Methotrexate
`Nadolol
`PravastatinS
`Penicillins
`RanitidineS
`Tetracycline
`TrimethoprimS
`Valsartan
`Zalcitabine
`
`Class 1
`Abacavir
`Acetaminophen
`Acyclovirb
`AmilorideS,I
`AmitryptylineS,I
`Antipyrine
`Atropine
`Buspironec
`Caffeine
`Captopril
`ChloroquineS,I
`Chlorpheniramine
`Cyclophosphamide
`Desipramine
`Diazepam
`DiltiazemS,I
`Diphenhydramine
`Disopyramide
`Doxepin
`Doxycycline
`Enalapril
`Ephedrine
`Ergonovine
`Ethambutol
`Ethinyl estradiol
`FluoxetineI
`Glucose
`ImipramineI
`
`Class 3
`Acyclovir
`AmilorideS,I
`AmoxicillinS,I
`Atenolol
`Atropine
`Bisphosphonates
`Bidisomide
`Captopril
`Cefazolin
`Cetirizine
`CimetidineS
`CiprofloxacinS
`Cloxacillin
`DicloxacillinS
`ErythromycinS,I
`Famotidine
`
`Highpermeability
`
`Lowpermeability
`
`a The listed compounds are predominantly gathered from the literature (1,3–18).
`b The compounds listed in italic are those falling in more than one category by different authors, which could be a result of the definition of
`the experimental conditions (i.e., acyclovir, amiloride, atropine, and captopril are listed in Classes 1 and 3 but all are highly soluble).
`Furosemide, hydrochlorothiazide, and methotrexate are listed in Classes 3 and 4, but they are all poorly permeable. Mebendazole is listed as
`Classes 2 and 4, but the compound is poorly soluble. Interesting examples are ciprofloxacin and erythromycin, which are listed in Classes 2
`and 3; it could just be that the properties of the compounds are intermediate between Classes 2 and 3. Ciprofloxacin has also been listed as
`Class 4.
`c The compounds listed in bold are primarily CYP3A substrates where metabolism accounts for more than 70% of the elimination; superscript
`I and/or S indicate P-gp inhibitors and/or substrate, respectively.
`
`science as a whole, exceptions are clues to new discoveries
`and new hypotheses.
`b) The BCS classification criteria for bioequivalence
`evaluation will not necessarily be appropriate for predicting
`drug disposition, as mentioned previously for the WHO Es-
`
`sential Medicines List (18), and as will be discussed subse-
`quently.
`c) High-permeability drugs are defined as compounds
`that exhibit 90% absorption in humans following oral dosing
`according to the FDA BCS criteria (2). Some drugs may fulfill
`
`Par Pharm., Inc.
`Exhibit 1045
`Page 003
`
`

`

`14
`
`Wu and Benet
`
`Fig. 2. Predominant routes of drug elimination for drug substances
`by BCS class.
`
`Fig. 3. Transporter effects on drug disposition by BCS class.
`
`these criteria because of the activity in vivo of uptake trans-
`porters in the intestine, rather than just due to high lipid
`passive diffusion permeability as reflected in Log P. Thus,
`some BCS drugs listed in Class 2 (and possibly some Class 1
`drugs) may show marked changes in bioavailability when in-
`testinal uptake transporters are inhibited.
`d) It is probable that some compounds that should be
`considered Class 1 in terms of drug absorption and disposition
`are listed as Class 2 according to the FDA BCS criteria due to
`the requirement of good solubility and rapid dissolution at
`low pH values, which is not limiting for drug disposition. This
`was recently discussed in terms of acidic drugs (26).
`
`We believe that a different set of criteria, particularly
`those relating to permeability but also to solubility, must be
`developed when using BCS in predicting drug disposition. We
`welcome the opportunity to work with the FDA and phar-
`maceutical manufacturers in setting simple in vitro surrogate
`permeability standards, as we discuss further in the section
`entitled “Biopharmaceutics Drug Disposition Classification
`System.”
`
`PREDICTING THE EFFECTS OF TRANSPORTERS
`
`Oral Dosing and the Predictability of Transporter Effects
`
`Recent work from our laboratory, initially based on cel-
`lular system studies evaluating transporter-enzyme interplay
`(27–29) have led us to the generalizations regarding trans-
`porter effects following oral dosing depicted in Fig. 3. The
`boldface italic items that follow represent the major predic-
`tive generalizations of this section of the current paper.
`Transporter effects will be minimal for Class 1 com-
`pounds. The high permeability/high solubility of such com-
`pounds allows high concentrations in the gut to saturate any
`transporter, both efflux and absorptive. That is, Class 1 com-
`pounds may be substrates for both uptake and efflux trans-
`porters in vitro in cellular systems under the right conditions
`[e.g., midazolam (30) and nifedipine (31) are substrates for
`P-glycoprotein], but transporter effects will not be important
`clinically. As stated above in Caution d, it is probable that
`some compounds that should be considered Class 1 in terms
`of drug absorption and disposition are not Class 1 in BCS due
`
`to the requirement of good solubility and rapid dissolution at
`low pH values. Such pH effects would not be limiting in vivo
`where absorption takes place from the intestine. Examples of
`this from Table I may include the NSAIDs diclofenac, dif-
`lunisal, flurbiprofen, indomethacin, naproxen, and piroxicam,
`as discussed by Yazdanian et al. (26), and warfarin, which is
`almost completely bioavailable (20,21). In contrast, ofloxacin
`is listed as Class 2 because of its low solubility at pH 7.5.
`Efflux transporter effects will predominate for Class 2
`compounds. The high permeability of these compounds will
`allow ready access into the gut membranes and uptake trans-
`porters will have no effect on absorption, but the low solu-
`bility will limit the concentrations coming into the entero-
`cytes, thereby preventing saturation of the efflux transporters.
`Consequently, efflux transporters will affect the extent of oral
`bioavailability (Fextent) and the rate of absorption of Class 2
`compounds.
`Transporter-enzyme interplay in the intestines will be
`important primarily for Class 2 compounds that are sub-
`strates for CYP3A and Phase 2 conjugation enzymes. For
`such compounds, intestinal uptake transporters will generally
`be unimportant due to the rapid permeation of the drug mol-
`ecule into the enterocytes as a function of their high lipid
`solubility. That is, absorption of Class 2 compounds is primar-
`ily passive and a function of lipophilicity. However, due to the
`low solubility of these compounds, there will be little oppor-
`tunity to saturate apical efflux transporters and intestinal en-
`zymes such as CYP 3A4 and UDP-glucuronosyltransferases
`(UGTs). Thus, changes in transporter expression, and inhibi-
`tion or induction of efflux transporters will cause changes in
`intestinal metabolism of drugs that are substrates for the in-
`testinal metabolic enzymes. Note the large number of Class 2
`compounds in Table I that are primarily substrates for
`CYP3A (compounds listed in bold) as well as substrates or
`inhibitors of the efflux transporter P-glycoprotein (indicated
`by superscripts S and I, respectively). Work in our laboratory
`has characterized this interplay in the absorptive process for
`the investigational cysteine protease inhibitor K77 (28,32)
`and sirolimus (29), substrates for CYP3A and P-glycoprotein,
`and more recently for raloxifene (33), a substrate for UGTs
`and P-glycoprotein.
`Absorptive transporter effects will predominate for
`Class 3 compounds. For Class 3 compounds, sufficient drug
`
`Par Pharm., Inc.
`Exhibit 1045
`Page 004
`
`

`

`Application of BDDCS to Predict Drug Disposition
`
`15
`
`will be available in the gut lumen due to good solubility, but
`an absorptive transporter will be necessary to overcome the
`poor permeability characteristics of these compounds. How-
`ever, intestinal apical efflux transporters may also be impor-
`tant for the absorption of such compounds when sufficient
`enterocyte penetration is achieved via an uptake transporter.
`It has been suggested (Refs. 5, 15, and others in meeting
`presentations) that products containing Class 3 drug sub-
`stances should qualify for a waiver of in vivo bioequivalence
`studies on the basis of dissolution studies alone, as for drug
`products containing Class 1 drugs. This is inappropriate, as it
`is now obvious that components of a Class 3 drug formulation
`can affect uptake transporters and modify bioavailability. Un-
`til more is known about the importance of intestinal trans-
`porters and validated methodology to predict the effects of
`formulation components on these transporters has been de-
`veloped, any expansion of in vivo bioequivalence study waiv-
`ers beyond Class 1 compounds is unwise policy. However, our
`proposal, presented below, could increase the number of
`drugs that qualify for Class 1 bioequivalence study waivers.
`It would be expected that Class 4 compounds could be
`substrates for both absorptive and efflux transporters. On
`first principles, we might expect that no Class 4 compounds
`would become effective drugs due to their solubility and per-
`meability deficiencies. However, it is probable that a number
`of Class 4 compounds are misclassified in terms of in vivo
`characteristics, as solubility in aqueous solutions may not re-
`flect solubility in gut contents. For example, the FDA gener-
`ated publication (15) and others have suggested that solubility
`measurements in surfactant containing solution may be a
`more appropriate basis for the solubility criteria. For true
`Class 4 compounds, oral bioavailability is minimal and trans-
`porter effects could be relevant, for example, where a change
`from 2% to 3% bioavailability could make a significant dif-
`ference.
`
`Food Effects (High-Fat Meals)
`It is well-known that food can influence drug bioavail-
`ability, both increasing and decreasing the extent of availabil-
`ity (Fextent) and the rate of availability. In December 2002, the
`FDA issued a guidance entitled “Food-Effect Bioavailability
`and Fed Bioequivalence Studies” (34). Fleisher et al. (6)
`noted that food effects on the extent of bioavailability could
`generally be predicted based on BCS class, as depicted in Fig.
`4. We have added the time to peak exposure (Tmax) designa-
`tions to the figure. High-fat meal studies are recommended by
`the FDA, as such meal conditions are expected to provide the
`greatest effects on gastrointestinal physiology so that systemic
`drug availability is maximally affected (34). It is generally
`believed that food effects result from changes in drug solu-
`bility and other factors as listed by the FDA (34), such as food
`may: “delay gastric emptying; stimulate bile flow; change gas-
`trointestinal pH; increase splanchnic blood flow; change lu-
`minal metabolism of a drug substance; and physically or
`chemically interact with a dosage form or a drug substance.”
`We hypothesize that although these other factors may be im-
`portant, drug-transporter interactions could often be the pri-
`mary mechanism for the food effect. We suspect that high-fat
`meals may inhibit drug transporters, both influx and efflux,
`and we have carried out preliminary studies that suggest that
`a high fat meal will inhibit P-glycoprotein (J. M. Custodio and
`L. Z. Benet, unpublished data).
`
`Fig. 4. Predictability of high-fat meal effects by BCS class after
`Fleischer et al. (6).
`
`High-fat meals will have no significant effect on Fextent
`for Class 1 compounds because complete absorption may be
`expected for high solubility/high permeability compounds,
`and as noted previously, no transporter drug interactions
`would be expected for Class 1 compounds.
`However, high-fat meals may delay stomach emptying
`and therefore cause an increase in peak time.
`High-fat meals will increase Fextent for Class 2 com-
`pounds due to inhibition of efflux transporters in the intestine
`and additional solubilization of drug in the intestinal lumen
`(e.g., micelle formation). Peak time could decrease due to
`inhibition of efflux cycling or increase due to slowing of stom-
`ach emptying; a combination of the two will usually be domi-
`nated by the delayed emptying. This will be true in cases
`where membrane permeation is passive, such as for the im-
`munosuppressants cyclosporine, tacrolimus, and sirolimus.
`However, if high permeability for a Class 2 compound results
`from uptake transporters, rather than ready partition into the
`intestinal membranes (see Caution c above), high-fat meals
`could inhibit both uptake and efflux transporters. Then, de-
`pending upon the relative magnitude of inhibition of uptake
`and efflux transporters, meal effects may be confounding,
`more likely having little effect on Fextent, but still increasing
`peak time due to delayed gastric emptying.
`Formulation changes that markedly increase the solu-
`bility of Class 2 compounds will decrease or eliminate the
`high-fat meal effects for these drugs. We believe that this is
`the reason that the newer cyclosporine microemulsion formu-
`lation (Neoral) eliminates the food effects associated with the
`older olive oil formulation (Sandimmune). In practice, drug
`formulators attempt to enable a Class 2 compound to function
`as a Class 1 compound, thereby eliminating food effects on
`Fextent and other transporter-drug interactions, as explained
`earlier for Class 1 drugs.
`High-fat meals will decrease Fextent for Class 3 com-
`pounds due to inhibition of uptake transporters in the intes-
`tine. Recent evidence suggests that intestinal drug uptake can
`be decreased by inhibiting organic anion transporting poly-
`petides, as shown by the effect of fruit juices on fexofenadine
`(35). As noted above, some Class 3 compounds can be sub-
`strates for intestinal efflux transporters. Depending upon
`whether the meal effects are more pronounced on efflux or
`
`Par Pharm., Inc.
`Exhibit 1045
`Page 005
`
`

`

`16
`
`Wu and Benet
`
`influx transporters for a Class 3 drug that is a substrate for
`both, an unexpected increase in the extent of bioavailability
`or no meal effect may be observed. We hypothesize that this
`may be the explanation for the lack of a high-fat meal effect
`on acyclovir. For Class 3 drugs, peak time would be expected
`to increase by a high-fat meal due to the combination of
`delayed stomach emptying and slower absorption.
`For Class 4 compounds, it is difficult to predict what will
`occur, as all of the interacting effects mentioned for Class 2
`and Class 3 compounds can be seen here. However, although
`not shown in Fig. 4, we believe that if high-fat meal effects are
`to occur, an increase of Fextent is more likely, resulting from
`the combination of increased solubilization of drug in the
`intestine and inhibition of efflux transporters.
`
`Postabsorption Effects and Intravenous Dosing
`
`For intravenous dosing, drug concentrations at the elimi-
`nating organ will always be relatively low due to the diluting
`effects of volume of distribution, as compared to concentra-
`tions of drug in the intestine. Therefore, saturation of trans-
`porters (and enzymes) will be minimal, if at all, and solubility
`considerations will be unimportant when measurable systemic
`concentrations of the drug are achieved.
`High extraction ratio drugs, where clearance approaches
`blood flow, are mainly limited to Class 1 compounds (and
`possibly a few Class 2 compounds so designated because
`of poor solubility at low pH; see Caution d). This will be true
`because the metabolism of such drugs is not rate limited
`either by dissolution or permeability. Examining Table I
`compounds with respect to pharmacokinetic data (20,21)
`reveals that 16 Class 1 compounds exhibit total clearance
`greater than half of liver blood flow (>10.7 ml min−1 kg−1)
`(abacavir, amitryptyline, buspirone, captopril, diltiazem,
`doxepin, imipramine, isosorbid dinitrate, labetolol, levo-
`dopa, metoprolol, meperidine, misoprostol, propranolol, ve-
`rapamil, and zidovudine), whereas only four Class 2 com-
`pounds (haloperidol, indinavir, itraconazole, and raloxifene),
`one Class 3 (pravastatin), and one Class 4 (mebendazole—
`probably misclassified) compound meet this criterion (Cau-
`tion a).
`Post intestinal absorption and following intravenous
`dosing, both uptake and efflux transporters can be important
`determinants of the disposition for Classes 2, 3, and 4 com-
`pounds. They will also be important for Class 1 compounds
`where high permeability results from uptake transporters
`(Caution c). Recent work in our laboratory has evaluated the
`importance of the rat hepatic uptake transporter oatp2 for
`digoxin (36,37), erythromycin (38), and atorvastatin (39). Us-
`ing the rat isolated perfused liver, we were not able to dem-
`onstrate a significant role for this transporter in the hepatic
`uptake of cyclosporine, dantrolene, nelfinavir, saquinavir,
`simvastatin, and talinolol (Y. Y. Lau, H. Okochi, N. Watan-
`abe, and L. Z. Benet, unpublished results). These studies em-
`phasize the difficulty in presenting a simple generalized con-
`clusion (Caution a) about the importance of uptake transport-
`ers in determining the disposition of the highly permeable
`Class 2 compounds. Obviously, there will be gradations within
`each broad BCS class for the permeability and solubility pa-
`rameters (16,40). The difference observed here between the
`importance of an uptake transporter for atorvastatin vs. sim-
`vastatin, two HMG-CoA reductase inhibitors, is most likely
`
`related to the differences in lipophilicity (often reflective of
`pKa) at the biologically relevant pH [simvastatin Log P ⳱
`4.42, Log D (pH 7.0) ⳱ 4.41 vs. atorvastatin Log P ⳱ 4.23,
`Log D (pH 7.0) ⳱1.54]. That is, at a pH close to physiologic,
`simvastatin is much more permeable than atorvastatin. It is
`more obvious that both uptake and efflux transporters will be
`involved in determining the disposition characteristics for
`Class 3 and Class 4 compounds as demonstrated by the recent
`double transfected cellular system studies reported by the
`Sugiyama and Kim groups investigating the importance of
`both uptake and efflux transporters on pravastatin (41) and
`fexofenadine (42).
`Biliary secretion of parent drug can be an important
`component of disposition for Classes 3 and 4 compounds.
`Biliary secretion of most Classes 1 and 2 parent drugs will be
`negligible due to extensive metabolism, although biliary ex-
`cretion of metabolites can be important.
`Renal elimination of Classes 3 and 4 compounds can be
`affected by both uptake and efflux transporters. Furthermore,
`metabolism of Classes 3 and 4 compounds in the kidney, and
`transporter-enzyme interplay, may be important for drugs
`where a kidney (vs. liver) specific uptake transporter is in-
`volved (e.g., furosemide). Metabolism of Class 2 (and possibly
`Class 1) compounds can be important in the kidney, when a
`kidney-specific enzyme such as CYP3A5 is identified (e.g.,
`tacrolimus and cyclosporine).
`Attempts to use markers of enzymatic processes (e.g.,
`midazolam vs. erythromycin breath test) to predict metabo-
`lism of another substrate cannot be expected to work when
`the test drugs are in different BCS classes. Even when two
`enzymatic substrates are in the same class, there is little
`chance to detect a potential correlation when the two test
`compounds are substrates for different uptake and efflux
`transporters. Many, many papers have investigated the po-
`tential for one substrate to predict the metabolism of other
`substrates by the same enzyme. Almost all of these attempts
`have failed, and we believe that the reason for the lack of
`correlation is due to differences in transporter susceptibilities.
`For example, erythromycin is a substrate for both uptake and
`efflux transporters as well as of CYP3A4. It is obvious that
`the ability of erythromycin metabolism to predict the metabo-
`lism of other CYP3A4 compounds will be compromised if
`differences in transport are not identified and fully taken into
`account. We believe, at this time, administration of “cock-
`tails” of substrates (i.e., a mixture of small quantities of drugs
`that are specific substrates for particular metabolic enzymes)
`to characterize a patient’s metabolic potential will be of little
`use, except for the most obvious pharmacogenetic differences
`in enzyme capacity.
`
`Drug-Drug Interactions
`
`Drug-drug interactions are not limited to enzymatic pro-
`cesses but can frequently be mediated by transporter interac-
`tions and often involve transporter-enzyme interplay for
`Class 2 compounds. Using our CYP3A4 transfected Caco-2
`cellular system (27), we demonstrated that for flux in the
`apical to basolateral direction, inhibition of P-glycoprotein
`caused a decrease in the extraction ratio of K77 (28) and
`sirolimus (29), both substrates for CYP3A4 and P-
`glycoprotein, although under the same conditions there was
`
`Par Pharm., Inc.
`Exhibit 1045
`Page 006
`
`

`

`Application of BDDCS to Predict Drug Disposition
`
`17
`
`no change i