In Vitro Models for Studying Metabolic Drug Drug Interaction Potential In Vitro Metabolizing Systems

In order to determine the DDI potential accurately the metabolic clearance of the 'victim' compound must be fully characterized.1'14 Once this is performed it should then be a relatively simple matter either experimentally or theoretically to block some of these metabolic pathways, thereby determining the predicted degree of interaction.

Traditionally, human liver microsomes (HLM) have been used for these purposes. HLMs represent a solution of endoplasmic reticular membranes commonly isolated from liver homogenate by differential centrifugation up to 100 000 g. Since these membranes are the major intracellular location of the CYPs, the resultant solution is a highly concentrated reaction environment for studying phase I metabolic reactions, particularly those involving CYPs. The major advantages of using hepatic microsomes include the relative ease of preparation and the ability to store the solutions for long periods of time with little or no reduction in activity.15

Typically, an arbitrary concentration (around 1 mM) of the new chemical entity (NCE) is first incubated with an arbitrary amount of HLM, and then a series of selective CYP inhibitors16,17can be used to examine the reduction in measured CLint to derive an impression as to which CYPs are most important in the clearance of this compound. There are several fundamental assumptions to this approach (such as the substrate concentration being significantly below the lowest metabolite Km and that of negligible nonspecific binding); however, advances in analytical sensitivity have allowed sufficient limits of quantification so that these can often be accounted for in the experimental design. It has been recognized that an NCE must be assessed as both a potential 'victim' and 'perpetrator' of DDIs. In parallel to the above experimental approach, where the 'victim' potential is explored, another approach is often used, whereby fluorescent CYP-selective probes (resorufin, coumarin, and fluorescein derivatives) are used in the presence of varying NCE concentrations to examine the 'perpetrator' potential. Later in the drug development process, as major metabolic pathways are identified and characterized, each particular metabolic route can be explored for liability to DDI.

The advent of recombinant human CYPs (rhCYPs)18 has provided the opportunity to combine several of these steps and obviate some of the reliance upon the 'selective' chemical CYP inhibitors. The enzyme preparations are available from commercial suppliers such as BD BioSciences, Invitrogen, and Cypex. Since these rhCYP systems express only one CYP it is theoretically possible merely to determine the CLint in these and scale them to in vivo both pre- and postmetabolite identification stages (eqn [1]).

where CLint j represents the intrinsic clearance (CLint) from a single CYP enzyme and Aj the abundance of that enzyme in the human liver. The rhCYP systems in common use include those expressed in mammalian (lymphoblast), insect (baculovirus transfected Trichoplusia ni), bacterial (Escherichia colt), and yeast (Saccharomyces cerevisiae) systems.

Discrepancies between CLint derived from HLM and rhCYP have been observed and may be due to several factors, including the extent of nonspecific microsomal binding.19,20 Although differences between the systems in such binding per amount of microsomal protein are minimal,21 experiments with rhCYP systems commonly employ lower protein concentrations than those using liver microsomes, which may cause disparity between the extent of nonspecific binding in the two systems, especially for lipophilic basic drugs. The CLint per unit amount of CYP (intrinsic activity or turnover number) may also vary between HLM and rhCYP enzymes: this has been attributed to differences in the concentrations of accessory proteins (cytochrome b5, NADPH:cytochrome P450 oxidoreductase) and the lipid microenvironment of the enzyme.22-24 Several attempts have been made to correct for this discrepancy.25-27 The initial approach was to use a relative activity factor (RAF) which would account for both this discrepancy and also the relative abundances of CYPs in HLM and rhCYP solutions. Although this approach has been used successfully on several occasions, it fails to address adequately the issues of interindividual variability in CYP expression and the apparent substrate specificity of RAFs. Indeed, the RAF merely demonstrates the amount of rhCYP required to give an equivalent reaction velocity to that of the particular HLM sample used (eqn [2]).

VmaxrhCYPj (nmol min 'pmol 1 rhCYP)

An alternative approach would be merely to compare the intrinsic activities of rhCYP versus HLM and provide CYP abundance scaling by simply mathematical means. This approach is described in detail elsewhere,10 but briefly it is the above RAF approach adjusted for the actual amount of HLM CYP present rather than a theoretical amount (eqn [3]).

VISEF ^_VmaxHLMj (nmol min-1mg-1 HLM)_

j VmaxrhCYPj (nmol min-1pmol-1 rhCYP)x HLM CYPj abundance (pmol mg-1 HLM) [ 1

Since the ISEF uses the actual CYP abundance of that HLM sample, it is readily scalable to in vivo without the added complications of intersample uncertainty. Although maximum rate of metabolism (Vmax) values have been used in eqns [2] and [3], CLint values for the relevant activities can also be used.26'27

In order to provide accurate and reproducible estimates of DDI potential at an early stage in drug development, the above issues must be taken into account when using both HLM and rhCYP. Human hepatocytes may also be used for the determination of DDI potential, but their practical application is hampered by the plethora of biochemical systems in place. HLM remain the experimental system of choice for providing quantitative predictions of DDI, although this can only be representative if the HLM batch in question has been thoroughly characterized with respect to CYP abundance. In Vitro Systems for Assessment of Cytochrome P450 Induction and Transporter Activity

Human hepatocyte cultures are frequently used as an in vitro system to detect potential CYP-inducing agents in humans.28-32 A number of clinically relevant inducing agents have been shown to induce CYPs known to be affected in vivo at both the mRNA and the catalytically active protein level.33 More recently, immortalized human hepatocytes, which offer the potential of a more useful screening procedure for CYP inducers, have also been used.34 Estimation of hepatic enzyme induction involves incubating the cultured cells for a 48-72-h period with a range of concentrations of the test drug (based on anticipated human plasma drug concentrations). A positive control compound, e.g., a known human clinical enzyme inducer such as rifampicin, is also incubated under the same conditions. A drug that produces a greater than twofold increase in enzyme activity (measured with a probe substrate such as midazolam for CYP3A4 activity) or a fold change in enzyme activity that is 40% of the positive control data is considered to be an enzyme inducer. However, although there are data relating the extrapolation of the in vitro measured effects to the in vivo situation, knowledge of the extent of increased enzyme content in vivo due to induction is scarce and therefore difficult to incorporate in simulations.

In addition, the role of efflux and uptake transporters cannot be ignored, although the level of our understanding of these proteins in terms of their quantitation and interplay lags behind that of the CYPs.10,35,36 Exposure of a drug to enzymes (particularly CYP3A4) during its transit from the gut lumen through the enterocyte and into the hepatic portal system depends on the interplay of transporters, passive membrane permeability, and enterocytic blood flow. A quantitative approach to incorporate these processes in the gut has been described for midazolam, a highly permeable drug.37 It is possible to estimate the effect of efflux transporters for other compounds by comparing the flux of midazolam in an in vitro system (such as Caco-2 cells) with that of other compounds where these data are reported.38 For a drug with a high permeability such as midazolam, the flux value approximates enterocytic blood flow but for substrates of efflux transporters flux from the gut into the hepatic portal vein is reduced. For active uptake transporters, studies using drug in the presence of cells incubated at 4 and 37 °C or with selective uptake inhibitors can be used to assess the degree of uptake into liver cells.39 In Vitro Parameters to Measure

A physiological approach to the simulation of mDDIs requires a range of data on tissue blood flow rates and tissue uptake, plasma protein binding, and nonspecific binding of compounds to the in vitro metabolizing system (hepatocytes, liver microsomes, or rhCYPs, enzyme kinetic data (in the form of CLint for substrates and Ki values for inhibitors). CLint is the enzyme-catalyzed metabolic clearance of a drug which is not influenced by other physiological parameters such as hepatic blood flow. The CLint value is a fundamental link between enzyme kinetics and pharmacokinetics.

Many of the approaches used to calculate the degree of an mDDI rely on estimation of the change of CLint resulting from the interaction of an inhibitor at a particular enzyme (CYP). The determination of this value for a CYP substrate is frequently determined using HLMs by measuring the half-life (in vitro t1/2) of disappearance of the compound at low (p 1 mM) substrate concentrations,40 as shown in eqn [4].

in vitro t1=2 x microsomal protein concentration

The assumption is made that the total CLint is mediated by CYP enzymes. In order to establish which CYPs are responsible for the metabolic process, selective chemical inhibitors are available which can inhibit individual CYP isoforms. Alternatively, rhCYPs can be incubated with the substrate to identify the CYP isoforms involved in metabolism, although adjustment of the rates with recombinant enzymes is necessary, as previously described (see Section

For compounds where metabolites have been identified, the Km and Vmax parameters for the pathways can be generated. In these cases, the value of CLint for compounds which obey Michaelis-Menten kinetics is obtained as the Vmax/Km ratio. The incorporation of CLint for compounds which show atypical (non-Michaelis-Menten) kinetic features are described in Section 5.35.3.

It is important to recognize that the value of CLint generated in an in vitro system may be affected by nonspecific binding of the test compound to the in vitro metabolizing system. This binding is referred to in the literature as/u,mic41 or/u,inc42 and is applied as a correction factor to CLint or Km values obtained in in vitro incubations (eqn [5]).

yu,mic Km xyu,mic

As shown in eqn [5], estimates of Km in vitro are apparent values without appropriate correction for the fraction unbound in the assay system.21'41-44 Knowledge of /ujmic has also been shown to be important for correction of the inhibition constant, Ki, used for the prediction of mDDI.45-48 Substrates and inhibitors which are lipophilic bases are particularly affected by nonspecific binding, whereas acidic and neutral compounds are typically bound to a significant extent only if they are highly lipophilic (logD7.4>3).49

Thus, it is only by incorporating such variability in simulations of CL prediction from CLint (eqn [6]), and of mDDI prediction from reduction in CLint due to inhibition, that the variability of the magnitude of the interaction seen in the clinic can be captured.

Lower Your Cholesterol In Just 33 Days

Lower Your Cholesterol In Just 33 Days

Discover secrets, myths, truths, lies and strategies for dealing effectively with cholesterol, now and forever! Uncover techniques, remedies and alternative for lowering your cholesterol quickly and significantly in just ONE MONTH! Find insights into the screenings, meanings and numbers involved in lowering cholesterol and the implications, consideration it has for your lifestyle and future!

Get My Free Ebook

Post a comment