The Use of Toxicokinetic Data in Preclinical Safety Assessment: A Toxicologic Pathologist Perspective

Abstract

Collection of toxicokinetic data has become a routine practice during the last 15 years in most general toxicity studies on pharma. It enables the correlation of pathological changes with the plasma concentration of drugs and/or their metabolites. This overview summarizes the use of the toxicokinetic data from the perspective of the toxicologic pathologist.

Introduction

What does a toxicologic pathologist seek from a pharmacokineticist? A similar question was raised by Monro to toxicologists (Monro, 1994). Pharmacokinetics may look complicated for some pathologists. Some sophisticated pharmacokinetic approaches exist such as physiologically-based pharmacokinetic modeling and pharmacokinetic-pharmacodynamic modeling. However, in general, toxicologic pathologists only have to understand the basic concepts of pharmacokinetics. The 13th International Symposium of the Society of Toxicologic Pathology has addressed the basic principles, identified critical issues, and demonstrated the value of pharmacokinetic data in the overall assessment and interpretation of drug toxicity (see Toxicologic Pathology, 23, [2], 1995). The purpose of this paper is to highlight concepts that toxicologic pathologists can use in their daily work. It will be made clear that a toxicologic pathologist uses pharmacokinetic data more on an individual animal level in relation to disease or pathological findings rather than the more general approach of the toxicologist who focuses primarily on the mean data in relation to risk assessment.

Definitions, Parameters and Limitations

The first thing to understand is that toxicologic pathologists deal with toxicokinetics instead of pharmacokinetics. So what is toxicokinetics? It may be defined as the description of the concentration of a compound in plasma (or serum or whole blood) with respect to time, based on a limited number of plasma samples—as a measure for exposure—within a toxicologic study. Toxicokinetics deals with what the body does with a drug given at a relatively high dose relative to the therapeutic dose. Ideally, toxicologic pathologists would like to describe the plasma concentration behavior with a curve. It is indeed advisable to look first at the plasma concentration time curve to get a feeling for the kinetic behavior of a compound. However, for obvious reasons, figures are not ideal to compare a vast number of data from different reports.

In practice, toxicologic pathologists inspect primarily the parameters: C_max, T_max, and AUC. It is important to understand that plasma exposure1 is reflected both by AUC and C_max, and that AUC is not per definition superior to C_max to describe relationship between plasma exposure and observed effects (Batra, 1995). Some direct adverse clinical signs correlate better with C_max than with AUC. Moreover, in practice, toxicologic studies with small animals are—in order to minimize the stress to the animals—performed typically with a maximum of 3 to 5 plasma samples (Cayen, 1995; Dahlem et al., 1995). Careful selection of the sample times is essential, since these sample times will directly influence the accuracy of AUC and C_max to reflect the actual mean plasma exposure. In addition the calculated parameters are generally based on noncompartmental analysis that implicitly makes rough assumption about the kinetics. Prediction of in vivo kinetic behavior remains difficult for new lead compounds (Shou, 2005). The calculated parameters are therefore prone to inaccuracies (Smith, 1997), and typically with 3 animals per concentration, variation coefficients in the range of 20–30% and more are obtained.

Here follows a brief description of other basic pharmacokinetic parameters that will be encountered by the toxicologic pathologist. Oral bioavailability (F) is defined as the fraction of the administered oral dose which reaches the systemic circulation. Under linear conditions, F may be calculated by dividing the dose-corrected AUC following oral administration by that following intravenous administration (which represents 100% systemic availability). Two major factors influencing oral bioavailability are absorption and first-pass extraction. In general, absorption will depend on the physicochemical properties of the drug which will influence both its solubility in the intestinal environment and its permeability through lipid membranes. For some drugs, absorption may be reduced by the action of efflux transporters, such as P-glycoprotein, in the gut (Walker et al., 2005). First-pass extraction may be due to metabolism in the gut wall and/or liver as well as elimination in the bile; all of these processes result in a reduction in F. Interspecies variation in oral bioavailability might be very large and may be difficult to predict (Sietsema, 1989).

Plasma half-life (t½) is defined as the time taken for plasma concentrations of drug to decline by one half and is influenced largely by two factors: volume of distribution (V) and clearance (CL). For detailed discussion of volume of distribution and clearance, we would like to refer to our suggested reading (Benet et al., 1995; Caldwell et al., 1995; Rowland et al., 1995). Plasma clearance is the irreversible removal of drug from the plasma and has units of volume of plasma cleared per unit time. Plasma clearance is the sum of all the clearance processes that are in operation for a given drug (e.g., hepatic metabolism, biliary secretion, renal filtration, renal metabolism, etc.). Since the majority of drugs are cleared predominantly by the liver, we tend to think of clearance in relation to hepatic blood flow; arbitrary categories are <25% liver blood flow = “low clearance” and >75% liver blood flow = “high clearance.”

Clearance may be significantly influenced by disease of the clearing organ(s). Distribution is the reversible movement of drug between plasma and tissues. Volume of distribution is a theoretical concept to describe the volume that would be required to accommodate all of the drug in the body at the same concentration present in the plasma. The overall distribution of a drug is a composite of its relative affinities for blood and the various tissues that, in turn, is influenced by the physicochemical properties of the drug, chiefly lipophilicity and ionization. It should be appreciated that drugs with high volumes of distribution (often basic compounds) may have much higher intracellular concentrations compared to plasma.

Interpretation of Toxicokinetics and Pathology Data in Integrated Fashion

So what do toxicologic pathologists want to learn from toxicokinetic data? Toxicokinetics are needed in order to interpret properly the ([histo]pathological) findings in a toxicologic study.

1. Dose-related changes in plasma exposure: linear, sub and supralinear

First of all, toxicologic pathologists want to know how toxicokinetics may explain the dose-response relationship observed. To illustrate this hypothetically: marked pathological changes—not present at the low dose—were observed after increasing the low dose 2-fold. If concomitantly, the plasma exposure was increased 10-fold, the steep dose-response curve is likely due to a supraproportional increase of plasma exposure. Supraproportional increase of plasma exposure with dose is often observed after saturation of a clearance route (for example saturation of elimination capacity of liver or kidney) or where the oral absorption is limited by efflux transporters at low doses.

Subproportional increase of plasma exposure with dose might explain the unexpected absence in increase of severity with increasing dose. Subproportional increase of plasma exposure with dose is for instance seen after saturation of active oral absorption or solubility limitation at higher doses. Saturable binding to protein may lead to higher clearance at high doses due to an increase in fraction unbound available for filtration clearance by the kidney or for metabolism (Welling, 1995). Knowledge of proportionality of plasma exposure to dose is also needed to discriminate toxicodynamic effects that do not relate to the toxicokinetics of the compound.

1. Time course and plasma exposure

What happens with the toxicodynamic response after repeated dosing: do toxicokinetics explain the response? To illustrate: for instance overt changes were observed after 39 weeks of daily dosing concomitant with a dramatically increased plasma exposure as compared to a 2-week study, which had no changes and a relatively low plasma exposure. Knowledge of the relation of plasma exposure to the duration of dosing is also needed to understand whether the dose-response relationship is explained by the toxicokinetics. The finding of dramatic increase in plasma exposure upon repeated dosing should trigger the toxicologic pathologist to look more carefully to clearing organs such as the liver and kidney.

Plasma exposure may increase2 with duration of dosing due to damage to the eliminating organs and subsequent reduction in clearance. Destruction or covalent modification of enzymes involved in metabolism may lead to decreased clearance. Anticipated accumulation may occur as result of the frequency of dosing relative to the half-life. The approach to the plateau solely depends on the drug half-life (Rowland, 1995), thus with a drug with a long half-life, it takes long before the plateau is reached but the accumulation may be extensive if relatively frequent administration is applied.

The plasma exposure may be reduced upon repeated dosing for instance due to induction of an enzyme involved in the metabolism of the compound (Worboys, 2001) or decreased plasma protein concentration leading to higher fraction unbound, which may be encountered under several conditions in a toxicologic study (Tibbitts, 2003). Neutralizing antibody formation to a biological may likewise increase clearance and decrease plasma exposure.

In general, toxicologic studies might be relatively short for the detection of age effects. However, in long-term studies like the rodent 2-year carcinogenicity studies, age is likely to affect the plasma exposure due to changes in, for instance, the body composition. For example the amount of total fat may increase, the total body water content may decrease, and lean muscle mass may decrease upon aging. Thus, the volume of distribution might change with age. Furthermore, physiological changes such as hepatic and renal blood flow may decrease upon aging and consequently clearance may be decreased. Likewise, clearance in juvenile animals may alter significantly upon aging in parallel with maturation of the CYP450 system and serum albumin concentration may also reduce upon aging (to senile animals).

1. Gender effect and plasma exposure

Is there a gender effect on the plasma exposure? There are several instances of pathologies that are specific to one sex. In some of these cases toxicokinetic events play a role. For example: gender specific preferential expressed enzyme involved in biotransformation are known. It is well known that male rodents often have higher expression levels of specific CYP P450s (Lin et al.,1995; Ghosal et al., 1996). The consequence of increased metabolism in the male rat could be a reduction in pathological findings due to a lower plasma exposure to the drug candidate or, conversely, more severe findings due to increased formation of a toxic metabolite or pharmacological metabolite that contribute to the overall response (Baldrick, 2003).

1. Individual response and plasma exposure

Is there an aberrant animal for which the plasma exposure explains the different toxicodynamic response? It is of special value for the toxicologic pathologist to study the sensitive animal(s) with relatively major ”aberrant” lesions in more detail. A holistic view of clinical data, clinical chemistry, toxicokinetic data and pathology is warranted. A few common causes to be considered more closely are summarized here:

• Check for liver or kidney disease, which may hamper removal of the compound from the body. If available, the individual clinical chemistry data are very helpful to interpret the study data if correlated with the individual toxicokinetics and pathology.

• Small changes in fraction of binding to plasma protein may significantly influence the free plasma concentration for highly plasma protein bound compounds. Indirectly the plasma protein concentration may be influenced by liver or kidney disease, leading to different plasma exposure (see above). For example: a decrease of a compound from 99% to 96% plasma protein binding will result in a 4-fold increase in ’free’ plasma drug concentration, which may lead to large interanimal variation. However, it should be noted that under many circumstances plasma exposure is independent of free fraction (Benet et al., 2002).

• Clinically relevant phenomena as vomiting. A dog, which was vomiting immediately after oral dosing, may not absorb any compound at all.

• Age of the animals: more especially with non-rodent studies (e.g., Cynomolgus monkeys) animals of different ages may be used within one study, hence physiological changes related to aging may contribute to aberrant animal toxicokinetic responses (see also discussion of age under time effects).

• Oral absorption may likewise be increased or decreased by gastro-intestinal disease (with or without associated diarrhea), thus affecting the plasma exposure directly. Also the fasted or fed status might contribute to marked toxicokinetic differences (Mayer, 1995; Keenan, 1996).

• Losses at injection site, for example by precipitation of the compound when administered with a massive dose subcutaneously, leads also to aberrant plasma concentration profiles. This crystallization process may be highly variable from animal to animal leading to large interindividual differences.

1. Miscellaneous

Toxicokinetics helps to support dose-selection for subsequent studies (Baldrick, 2003). Toxicokinetic data are just as helpful to interpret differences in the outcome of repeated studies (interstudy variability) outcome. It is not unlikely that the toxicokinetic behavior of animals from different supplier may be different but, in the same way, strain differences may also lead to differences. Often, different drug batches (particle size, crystal polymorphism) or formulations may have been used that may affect the plasma exposure. Surfactants or emulsifying agents may improve drug solubility and absorption (Mayer, 1995).

Discussion

It is important to realize that plasma concentration is still a surrogate for the target tissue concentration, and it may not be representative of its concentration in the target tissue (Monro, 1994; Batra, 1995; Lin, 2006). Moreover, ideally, toxicologic pathologists would also like to correlate animal plasma levels of metabolites to human exposure to these metabolites to calculate safety margins (especially for highly metabolized drugs, exposure to metabolites may be significant greater than exposure to the parent drug). However, in practice this is often not feasible3 due to the complicated laborious and time consuming bioanalysis needed. Notwithstanding, measures of plasma exposure are more appropriate to correlate with toxicologic findings than relationships on a dose basis (Mayer, 1995). Toxicokinetic data are essential for toxicity testing, and help the toxicologic pathologist to interpret the findings in an appropriate manner. While evaluating histopathological findings, toxicologic pathologists should have an awareness of the meaning of toxicokinetic changes in order to be able to better interpret and evaluate the findings together with the toxicologist, clinical chemist, and pharmacokineticist.