Journal of Pharmacokinetics and Biopharmaceutics, Iioi. 23, No. 2, 1995
PERSPECTIVES IN PHARMACOKINETICS Physiologically Based Pharmacokinetic Modeling as a Tool for Drug Development Steven B. Charnick, ~'-~Ryosei Kawai, 2 Jerry R. Nedelman, ! Michel Lemaire, 3 Werner Niederberger, 3 and Hitoshi Sato 3'4 Received August I1, t994--Final September 28, 1994 Since the pioneer#Tg work of Haggard and Teoretl #~ the first half of the 20th centut T, and of Bischoff and Dedrick fl~ the late 1960s, physiologically based pharmacokinetic ( PBPK) modeling has gone through cycles of general acceptance, and of healthy skepticism. Recentlj,, however, the trend in the pharmaceuticals induso T has been away from PBPK models. This is understandable when one considers the time and effort neeessatT to develop, test, and implement a typical PBPK model, and the fact that in the present-day environment for drug development, efficacy and safe O, must be demonstrated and drugs brought to market more rapidly. Although there are many modeliug tools available to the pharmacokineticist today, many of which are preferable to PBPK modeling in most ch'cumstances, there are several situations in which PBPK modeling provides disthwt benefits that outweigh the drawbacks of increased time and effort for implementation. In this Commentary, we draw on our" experience with this modeling technique in an industry setting to provide guidelines on when PBPK modeling techniques couM be applied in an industrial setting to sati,~r the needs of regulatm T customers. We hope these guidelines will assist researchers in deciding when to apply PBPK modeling techniques. It is our contention that PBPK modeling shouM be viewed as one of many modeling tools for drug development.
KEY WORDS: physiologically based pharmacokinetics; modeling; drug development.
INTRODUCTION Progress in biopharmaceutical and biomedical sciences has made greater insight into pharmacokinetics (PK) and pharmacodynamics (PD) possible. tDepartment of Clinical Pharmacology, Drug Safety, Sandoz Research Institute, Sandoz Pharmaceuticals Corporation, East Hanover, New Jersey 07936. 2Tsukuba Research Institute, Sandoz Pharmaceuticals Ltd., 80hkubo, Ibaraki 300-33, Japan. 3Drug Safety, Sandoz Pharmaceuticals Ltd., 4002 Basel, Switzerland. 4Present address: Toyama Medical and Pharmaceutical University, Sugitani 2630, Toyama 930-0 I, Japan. ~To whom correspondence should be addressed. 217 0090-466x/95/0400.0217$07.50/0!~ 1995PlenumPublishingCorporation
218
Charnick et al.
Microscopic and macroscopic experimental approaches have been applied to explore the complex interactions between drugs and the human body. Such experiments generate data about these interactions. Mathematical modeling can contribute to organizing and reducing, and therefore understanding, such data. From existing data, models can predict future experimental observations, or suggest experiments to generate the data needed for a fuller understanding of the system. When a model adequately represents a system, its predictive power can be directed towards controlling the system. When the mechanism underlying a complex system is either beyond our present scope of understanding, or too complex to be completely and practically specified, one is limited to the use of empirical, nonparametric modeling approaches in an effort to organize data, or to predict and control the system. When one has sufficient knowledge of the mechanics of a system, development of a more complex parametric model is feasible. Such a model allows one to comprehensively understand, and therefore to predict and control the system with more insight and confidence. In the context of PK/ PD, extreme examples of the former case include artificial neural networks (1,2,3) and the graph theoretical approach (4,5). Ailometric scaling and classical compartmental modeling techniques incorporate aspects of both approaches. An extreme example of a complex, parametric modeling tool is physiologically based pharmacokinetic (PBPK) modeling. PBPK modeling involves the development of a model, parameterized in physiological, biochemical, and thermodynamic parameters, which organizes much of the knowledge of the drug-body system. Although schematically a PBPK model is a multicompartment model, its compartments represent real organs and other physiological spaces, unlike the empirical compartments of so-called "compartment models." The ultimate goal of PBPK modeling is to develop an "artificial human," for the prediction of human PK and PD, either from first principles, through interspecies extrapolation from preclinical animal data, or through intraspecies extrapolation from other clinical data. Proponents of nonparametric, semiparametric, and parametric methodologies are often at odds with each other's approaches. We feel that they are more complementary than exclusive. One could potentially combine the approaches to develop a hybrid tool for the solution of a particular analytical problem. For example, the combination of PBPK with the population approach (6,7) or with expert systems (8) has been considered. With regard to the practical inplementation of PBPK models, a variety of views have been expressed in the past decade. These views have varied from mostly negative (2,9,10), to moderate (11-14), to mostly positive (15-20). We present a description, from the unique perspective of modelers in the pharmaceuticals industry, of when PBPK modeling could be useful in
PBPK Modeling as a Drug Development Tool
219
satisfying the needs of regulatory customers. We have developed, in an industrial setting, PBPK models for SDZ IMM 125 (21), and a novel 5HT3 antagonist (22). Now that PBPK modeling has been applied by scientists within a major pharmaceuticals company, we feel that it is the right time to critically reevaluate PBPK, with regard to its applicability in drug development. We begin with a brief, quantitative comparison of published environmental and pharmaceutical applications of PBPK modeling. Then we describe when PBPK modeling can be used for organizing and critically assessing data from multiple sources: in vitro, animals, and humans; and for prediction in pharmacokinetics. Finally we discuss why we feel that PBPK modeling can be applied efficiently, on a case-by-case basis, in an industrial setting. An introduction to the mathematical and experimental procedures for developing PBPK models is presented in chapter 9 of Gibaldi and Perrier (23). HISTORY OF PBPK
The earliest work that could be called PBPK was that of Haggard (24) on the entry of anesthetic agents into the brain. Teorell (25) was the first to suggest the application of mechanistic, physiologically based modeling to describe the kinetics of xenobiotics in the body. Although there were other PBPK studies (26,27) reported in the early 1960s, it was not until Bischoff, Dedrick, et al. published a series of papers (28-30) in the late 1960s and early 1970s that the scientific community was provided with more rigorous methodology for PBPK modeling. Through the 1970s and 1980s, PBPK modeling techniques gradually began to gain some acceptance. The number of publications containing PBPK models for therapeutic compounds peaked in the late 1970s (Fig. l). It was expected that the use of PBPK modeling would become widespread in the late 1980s as in vitro-to-in vivo prediction of metabolism and tissue distribution became more available, and that PBPK modeling would also be applied widely in clinical situations. In fact, the number of PBPK models developed for therapeutic compounds has never again reached such sustained high levels (Fig. 1). One reason for this may be that another, more simple "physiological" approach--allometric interspecies scaling (l 0,31,32)--has been more extensively utilized in pharmacokinetic studies from the late 1980s to the present. Indeed, the view that PBPK may not be a practically useful tool for drug development appeared in several review papers in the early 1990s (10,33,34). In contrast, as Fig. I clearly shows, the use of PBPK in the assessment of the pharmacokinetics of toxic chemicals in the environment has been increasing dramatically since the mid-1980s, and the number of publications presenting PBPK models for toxic substances has far exceeded that for
220
Charnick et ai. 25 co d) "(3 O
v"
2O
i:1. ED
15
"6
10
E Z
5
[3,.
0 1968
1973
1978
1983
1988
1993
C3 Academic/Therapeutic ED Industrial/'rherapeutic l i Academic/Toxic
9 Industrial/Toxic
Fig. !. Trends in physiologically based pharmacokinetic (PBPK) modeling, 1968 1993. Bars represent the number of PBPK models in a particular category published in a particular year. There are four categories, representing each combination of two classes of models. "Academic" models have a first author from academia, "Industrial" models have a first author from industry. "Therapeutic" models are for compounds with some therapeutic benefit, "Toxic" models are for compounds known to have no therapeutic benefit, and some toxic effect
therapeutic compounds since 1990. In fact, several review papers have been published describing the application of PBPK to risk assessment for toxic chemicals (8,35--42), with the ultimate goal being the prediction of organspecific toxicity for these compounds. Since tissue accumulation of such chemicals cannot be studied experimentally in humans, the use of PBPK is especially worthwhile in this field. As for therapeutic compounds, PBPK models have only rarely been proposed since the late ! 980s. Those proposed have been diversified, e.g., for drug-carrier complexes (43,44), drug-containing microspheres (45), metalchelating agents (46,47), tissue-imaging agents (48-50), and polypeptides and proteins (51-53). PBPK models have also been utilized in the analysis of PK/PD relationships for enzyme inhibitors (54,55). Thus, the trend has been to apply PBPK to cases where there is a special need in either drug development or clinical applications, while, prior to the late 1980s, PBPK was applied primarily to conventional drugs that had already been used in clinical situations at the time the model was developed. A glance at Fig. 1 might lead one to conclude that the full potential of PBPK modeling would only be realized when applied to toxic chemicals, not therapeutic compounds. However, we have learned through experience that, in particular cases, PBPK modeling has distinct advantages. We now
PBPK Modeling as a Drug Development Tool
221
discuss an example of the application of physiologically based modeling at Sandoz, and describe some advantages provided by this modeling approach. A SANDOZ APPLICATION OF PBPK
Kawai et al. (21) reported the application of a PBPK model to study a new cyclosporine derivative, SDZ IMM 125. Their goal was to systematically assess most of the data obtained in preclinical pharmacokinetic/pharmacodynamic studies. SDZ IMM 125 is a potent immunosuppressant (56) which shows less nephrotoxicity than Sandimmune (cyclosporine A). This derivative, like Sandimmune, is a cyclic polypeptide comprising 11 amino acids, yet it has severalfold lower lipophilicity than Sandimmune. This lipophilicity is reflected in plasma protein binding and transmembrane exchange rates (57). Blood cell distribution was a particular concern in their PBPK model, because blood cells are known to be (i) a specific interaction site of this class of immunosuppressants, and (ii) a major reservoir for drug in the body. Blood cell accumulation of SDZ IMM 125 is saturable in the upper therapeutic range. In addition, SDZ IMM 125 crosses tissue cell membranes slowly; a membrane transport-limited PBPK model was thus applied to several of the compartments in the PBPK model. Both of these concerns were incorporated in the whole-body PBPK model for SDZ IMM 125 by applying local physiologically based models for these processes, and including them in the model. Their whole-body PBPK model for SDZ IMM 125 was capable of reproducing blood concentration-time (C-t) profiles and tissue C-t profiles (which are particularly important for organs which must be monitored from an efficacy/toxicity point of view). In addition, their model was also capable of predicting preclinical data in dogs (blood C-t profiles) and clinical data in humans (blood C-t profiles, nonlinear increase in A U C and Cmax with increasing dose) from preclinical data in rats (blood and tissue C-t profiles, as well as excretion C-t profiles). We next discuss when PBPK modeling could be applied in drug development, and why it would be appropriate in such cases. The PBPK model of Kawai et al. (21) for SDZ IMM 125 is used repeatedly as an example, as it demonstrates many of the advantages of PBPK. WHEN WOULD PBPK BE USEFUL FOR DRUG DEVELOPMENT?
The success and long-term viability of a pharmaceuticals company ultimately depend on the company's ability to get new, effective drugs to market as rapidly as possible. This, in turn, depends on meeting the regulatory
222
Charnick et
aL
customers' needs in order to expedite approval. A sample of how PBPK can help satisfy these requirements follows. Demonstrating an Understanding of a Compound's Unusual Pharmacokinetics
PBPK modeling can help to organize and explain observed pharmacokinetic data in a comprehensive manner. This shows the regulatory agencies that the sponsor has a full understanding of the compound's behavior. PBPK models are especially useful in situations in which drugs exhibit unusual ADME characteristics. For example, nonlinear pharmacokinetics is often due to saturable drug metabolism and/or elimination, which results in a decrease in clearance with increasing dose. Clearance of SDZ IMM 125 in a Phase I trial (21), in contrast, was found to increase with increasing intravenous dose. This was quantitatively explained by a PBPK model in which saturation in blood cell binding of SDZ IMM 125 was taken into account. The increasing number of special formulations (e.g., liposomes and microparticles), prodrugs, and biotech products (mostly peptides and proteins), for which the pharmacokinetics are not simply understood, potentially justifies the use of PBPK in development of such drugs as well. PBPK modeling can also guide the development program by suggesting experiments that elucidate the parameters most important for understanding a compound's pharmacokinetics. This can be accomplished via sensitivity analysis using a PBPK model (58,59). For example, Hetrick et al. (59) showed that the PBPK models they examined were most sensitive to the maximum Michaelis-Menten metabolism rate V,.ax, and the blood/air and the fat/air partition coefficients, while less sensitive to the Michaelis-Menten parameter Km and the other tissue/air partition coefficients. Quantification of Unchanged Drug and/or its Metabolite(s) at Particular Sites
PBPK models can generate C-t profiles in any one of the physiological compartments including in the model. This information is potentially useful not only in toxicity assessments but also in more efficient coupling to pharmacodynamic models for therapeutic efficacy. Kawai et al. (21) showed that animal tissue drug C-t profiles of SDZ IMM 125 could be accurately reproduced by a PBPK model. Furthermore, each organ compartment in their model comprised subcompartments representing plasma, blood cells in the capillaries, interstitial space, and intracellular space. Prediction of C-t profiles in these subcompartments is helpful in developing an understanding of pharmacological and toxicological mechanisms.
PBPK Modelingas a Drug DevelopmentTool
223
Comparison of a Successor Drug to Current Therapy
In this case, the task of PBPK is to help distinguish the drug kinetically from members of its therapeutic class. This was the case when Kawai et al. (21) examined the kinetics of SDZ IMM 125. They found that differences in lipophilicity between the new compound (SDZ IMM 125) and the marketed compound (Sandimmune) could explain differences in PK between the two compounds.
Changes in PK in Different Physiological States
In this case, the task of PBPK is to explain the PK under special conditions in humans, which would facilitate the effective labeling of the compound. Some examples include drugs that change physiological conditions (e.g., blood flow rate), are targeted to special populations (e.g., pediatric, aged, obese), or are targeted to patients in unusual physiological conditions (e.g., organ transplant patients, patients in liver/kidney failure). The application of PBPK modeling to these situations is analogous to the extrapolation of PBPK models from preclinical data to clinical data, in that much of the extrapolation is accomplished via alterations in blood flows and physiological volumes within the model.
Relationship Between Preclinical and Clinical Data
Sometimes, the pharmacokinetics of a compound differ significantly between the species used in toxicity studies and humans. In this case, the aim of PBPK is to give a reasonable explanation for the species differences in PK via a sensitivity analysis using PBPK models (58,59). For example, Hetrick et al. (59) showed that the PBPK models they considered were sensitive to t h e muscle/air partition coefficients only when applied to humans. If there are any sensitive parameters that can be related to the species differences (e.g., metabolic pathways, physiological conditions, receptor sensitivity), the estimation of such parameters should be as accurate as possible.
DISCUSSION In this report, we take the position that PBPK modeling is a useful tool for organizing, understanding data, and for making predictions based on
224
Charnick et al.
the data. There are disadvantages associated with PBPK modeling, however, that have limited its use: 1. Due to their complexity, PBPK models are difficult to implement rapidly. 2. Adequate software for PBPK modeling, capable of rapid solution of complex systems of differential equations, has only recently become available. 3. The lack of human tissue data to verify the model, as well as somewhat "preferential" adjustments to the techniques performed to improve the goodness-of-fit between simulated and experimental values, have made it difficult to objectively evaluate the results of PBPK models. 4. The lack of standard procedures and techniques for the implementation of PBPK models has led to a lower probability of success, less speed, and therefore, a higher cost of development for these models, relative to simpler modeling techniques. 5. Several researchers have expressed reservations over whether PBPK models are useful tools for organizing and understanding data, and for extrapolating beyond the range of the data. D'Souza and Boxenbaum (15) noted that "at their worst, these models provide stale or infertile views of reality and thus frustrate and alienate us with the triviality of their insights." lngs (10) noted that "these models also get more unwieldy as their complexity increases, becoming almost as difficult as understanding the original system in extreme cases." Veng-Pedersen (2) stated that "the predictive usefulness of such models is jeopardized by the fact that the physiological models are highly differentiated complex structures having an excessive number of variables (parameters), many of which are very difficult to accurately determine. This, in combination with a high degree of biological variability (biological errors), identifiability problems, sampling problems, etc., makes the identification of such models very problematic, tedious, and troublesome." As for prediction beyond the range of data, one valid criticism of PBPK models is that they can lose their general applicability to other animal species, similar compounds, or altered physiological conditions, if the model structure is either too simplified or too refined during its development, in order to fit the experimental data well. The balance between simplicity and complexity is a critical issue in PBPK model development. For organizing and understanding data, simplicity is essential, whereas for accurate prediction and extrapolation beyond the range of data, more complex models are generally required. We have described why we feel the present research environment is ripe for the inclusion of physiologically based pharmacokinetics as a modeling tool, despite these disadvantages. During the past 10-15 years, the hardware and software capable of rapid solution of coupled, ordinary differential
PBPK Modeling as a Drug Development Tool
225
equations has advanced to the point where the solution to such systems (characteristic of PBPK models) can be quickly achieved. Among the many software packages available, ACSL (60), a simulation language, and SimuSolv (61), a simulation program written in ACSL, are powerful. SimuSolv was used in the study of Kawai et al. (21) on SDZ IMM 125. The cost of implementing SimuSolv is not a critical problem for most pharmaceuticals companies, and we thus propose SimuSolv as standard software for PBPK modeling in the pharmaceuticals industry. Tissue data from studies in rodents, dogs, and primates allow PBPK models to be validated in several species for reasonably confident extrapolation to humans. Recently, greater access to individual human data on plasma protein binding (62,63) and physiological parameters such as blood flow in major tissues (64-67), cardiac output (68), and measures of liver function (69-70), by noninvasive methods has been made possible. The potential utility of PBPK in clinical situations has thus been increased, as more accurate measures of these physiological parameters can be included in the model as applied to humans. It is our hope that as researchers within the pharmaceuticals industry apply the techniques of PBPK modeling in appropriate cases, a compendium of standard procedures will evolve. Such a list will also serve as a guideline for what level of complexity is required for the development of a PBPK model for a particular compound. The conceptual advantages of PBPK models are summarized in a comment by D'Souza and Boxenbaum (15): "at their best, they allow us to understand the accumulation of thought in pharmacokinetics and pharmacodynamics, and help with the integration of data and improvement of experimental design." Describing the complex systems involved in new drug research and development is a challenging task. To be successful, the pharmaceuticals industry has to make use of the many modeling tools available, choosing the most appropriate tool for each specific task. This will allow the industry to evaluate the potential and limitations of variable compounds in a comprehensive, efficient manner. Computer-aided molecular modeling and structural analysis have aided rational drug design; computeraided PBPK modeling is an important tool for drug development as well. We propose the limited use of PBPK modeling, on a case-by-case basis, in drug development. We stress that PBPK modeling should be applied only in cases where the potential benefits justify the time and cost associated with its implementation. This must always remain a practical consideration in the pharmaceuticals industry, where efficiency and low cost are paramount. This report reevaluates PBPK modeling with regard to its applicability to drug development and highlights the ability of PBPK modeling to satisfy the needs of regulatory customers.
226
Charnick et al.
REFERENCES I. A. S. Hussain, R. D. Johnson, N. N. Vachharajani, and W. A. Ritschel. Feasibility of
2. 3. 4. 5. 6. 7. 8, 9. 10. 11, 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.
developing a neural network for prediction of human pharmacokinetic parameters from animal data. Pharm. Res. 10:466-469 (1993). P. Veng-Pedersen and N. B. Modi. Neural networks in pharmacodynamic modeling. Is current modeling practice of complex kinetic systems at a dead end? J. Pharmacokin. Biopharm. 20:397-412 (1992). A. S. Hussain, X. Yu, and R. D. Johnson. Application of neural computing in pharmaceutical product development. Pharm. Res. 8:1248-1252 (199 I). R. S. Markin, W. J. Murray, and H. Boxenbaum. Quantitative structure-activity study on human pharmacokinetic parameters of benzodiazepines using the graph theoretical approach. Pharm. Res. 5:201-208 (1988). E. Gifford, M. Johnson, and C.-C. Tsai. A graph-theoretic approach to modeling metabolic pathways. J. Comp. Aid. Mol. Dis. 5:(1991), P. O. Droz, M. M. Wu, W. G. Cumberland, and M. Berode. Variability in biological monitoring of solvent exposure. I. Development of a population physiological model. Br. J. Ind. Med. 46:447-460 (1989). T. M. Ludden, S. R. B. Allerheiligen, and R. F. Burk. Application of population analysis to physiological pharmacokinetics. J. Pharmacokin. Biopharm. 19:101S-I13S (1991). D. R. Mattison and F. R. Jelovsek. Pharmacokinetics and expert systems as aids for risk assessment in reproductive toxicology. Envh'on. Health Perspect. 76:107-119 (1987). L. P. Balant, H. Roseboom, and U. M. Gundert-Remy. Pharmacokinetic criteria for drug research and development. In B. Testa (ed.), Advances in Drug Research, Academic Press, London, 1990, pp. 1-137. R. M. J. hags. Interspecies scaling and comparisons in drug development and toxicokinetics. Xenobiotica 20:1201-1231 (1990). W. R. Chappel and J. Mordenti. Extrapolation of toxicological and pharmacological data from animals to humans. In B. Testa (ed.), Advances in Drug Research, Academic Press, New York, 1991, pp. 1-116. W. A. Colburn. Physiologic pharmacokinetic modeling. J. Cfin. Pharmacol. 28:673-677 (1988). W. A. Ritschel and P. S. Banaerjee. Physiological pharmaeokinetic models: Principles, applications, limitations, and outlook. Meth. Find. Exp. Clin. Pharmacol. 8:603-614 (1986). L. E. Gerlowski and R. K. Jain. Physiologically based pharmacokinetic modeling: Principles and applications. J. Pharm. Sci. 72:1103-1127 (1983). R. W. D'Souza and H. Boxenbaum. Physiological pharmacokinetic models: Some aspects of theory, practice, and potential. Toxicol. Ind. Health. 4:151-171 (1988). K. B. Bischoff. Physiological pharmacokinetics. Bull. Math. Biol. 48:309-322 (1986). M. Rowland. Physiological pharmacokinetic models and interanimal species scaling. Pharmacol. Ther. 29:49-68 (1985). M. Rowland. Physiologic pharmacokinetic models: Relevance, experience, and future trends. Drug Metab. Rev. 15:55-74 (1984). A. J. J. Atkinson, T. I. Ruo, and M. C. Frederiksen. Physiological basis of multicompartmental models of drug distribution. Trends Pharmacol. Sci. 12:96-101 (1991). K. J. Himmelstein and R. J. Lutz. A review of the applications of physiologically based pharmacokinetic modeling. J. Pharmacokin. Biopharm. 7:127-145 (1979). R. Kawai, M. Lemaire, J.-L. Steimer, A. Bruelisauer, W. Niederberger, and M. Rowland. Physiologically based pharmacokinetic study on a eyclosporine derivative, SDZ IMM 125. J. Pharmacokin. Biopharm. 22:327-365 (1994). H. Sato, A Bruelisauer, M. Lemaire, and W. Niederberger. Physiological modeling of nonlinear hepatic first-pass of a novel 5-HT3 antagonist, SDZ ICM 567, in rats and dogs, and extrapolation to human. (in preparation). M. Gilbaldi and D. Perrier. Pharmacokinetics, Marcel Dekker, New York, 1982.
PBPK Modeling as a Drug Development Tool
227
24. H. W. Haggard. The absorption, distribution, and elimination of ethyl ether. I1. Analysis of the mechanism of the absorption and elimination of such a gas or vapor as ethyl ether. J. Biol. Chem. 59:753-770 (1924). 25. T. Teorell. Kinetics of the distribution of substances administered to the body. Arch. hlt. Pharmaeodyn. Ther. 57:205-240 (1937). 26. W. W. Mapleson. An electric analogue for uptake and exchange of inert gases and other agents. J. Appl. Physiol. 18:197-204 (1963). 27. R. Bellman, R. Kalaba, and J. A. Jacquez. Some mathematical aspects of chemotherapy. Bull. Math. Biophysics. 22:181-190 (1960). 28. K. B. Bischoff, R. L. Dedrick, D. S. Zaharko, and J. A. Longstreth. Methotrexate pharmacokinetics. J. Pharm. Sci. 60:1128-1133 (1971). 29. K. B. Bischoff, R. L. Dedrick, and D. S. Zaharko. Preliminary model for methotrexate pbarmacokinetics. J. Pharm. Sci. 59:149-154 (1970). 30. K. B. Bischoff and R. L. Dedrick. Thiopental pharmacokinetics. J. Pharm. Sei. 57:13461351 (1968). 31. H. Boxenbaum and R. W. D'Souza. lnterspecies pharmacokinetic scaling, biological design, and neoteny. In B. Testa (ed.), Advances in Drug Research, Academic Press, London, 1990, pp. 139-196. 32. J. Mordenti. Man versus beast: Pharmacokinetie scaling in mammals. J. Pharm. Sei. 75:1028-1046 (1986). 33. L. P. Balant and M. Cex-Fabry. Physiological pharmacokinetic modeling. Xenobiotica 20:1241-1257 (1990). 34. E. Voisin, M. Ruthsatz, J. M. Collins, and P. C. Hoyle. Extrapolation of animal toxicity to humans: Interspecies comparisons in drug development. Reg. Toxicol. Pharmacol. 12:107-116 (1990). 35. M. E. Andersen, D. Krewski, and J. R. Withey. Physiological pharmacokinetics and cancer risk assessment. Cancer Letters 69:1-14 (1993). 36. R. B. Conolly and M. E. Andersen. Biologically based pharmacodynamie models: Tools for toxicological research and risk assessment. Ann. Rev. Pharmacol. Toxicol. 31:503523 (1991). 37. H.-W. Leung. Use of physiologically based pharmacokinetic models to establish biological exposure indexes. Am. Ind. Hyg. Assoc. J. 53:369-374 (1992). 38. H.-W. Leung. Development and utilization of physiologically based pharmacokinetie models for toxicological applications. J. Toxicol. Environ. Health 32:247-267 (1991). 39. K. Krishnan and M. E. Andersen. Physiological modeling and cancer risk assessment. New Trends Pharmacokh~. (1991). 40. D. Krewski, J. R. Withey, L. F. Ku, and C. C. Travis. Physiologically based pharmacokinetic models: Applications in carcinogenic risk assessment. New Trends Pharmaeokin. (1991). 41. J. N. Blancato. Physiologically-based pharmacokinetic models in risk and exposure assessment. Ann7 1st. Super. Sanlta. 27:601-608 (199t).42. R. J. Lutz and R. L. Dedrick. Implications of pharmaeokinetie modeling in risk assessment analysis. Environ. Health Perspect. 76:97-106 (1987). 43. A. Boddy, L. Aarons, and K. Petrak. Efficiency of drug targeting: Steady-state considerations using a three-compartment model. Pharm. Res. 6:367-372 (1989). 44. C. A. Hunt, R. D. MacGregor, and R. A. Siegel. Engineering targeting in vioo drug delivery. 1. The physiological and physicochemical principles governing opportunities and limitations. Pharm. Res. 3:333-344 (1986). 45. J. M. Gallo, C. T. Hung, P. K. Gupta, and D. G. Perrier. Physiological pharmacokinetic model of adriamycin delivery via magnetic albumin microspheres in the rat. J. Pharmacokhl. Biopharm. 17:305-326 (1989). 46. F. G. King and R. L. Dedrick. Physiological pharmaeokinetic modeling of cis-dichlorodiamineplatinum(ll) (DDP) in the mouse. J. Pharmaeokin. Biopharm. 20:95-99 (1992). 47. F. G. King and R. L. Dedrick. Physiologic pharmacokinetic modeling of cis-dichlorodiamineplatinum(ll) (DDP) in several species. J. Pharmacokin. Biopharm. 14:131.-157 (1986).
228
Charnlck et al.
48. S. M. Eaton, P. Wedeking, M. F. Tweedle, and W. C. Eckelman. A multi-organ, axially distributed model of capillary permeability for a magnetic resonance imaging contrast agent..1. Pharm. Sci. 82:531-536 (1993). 49. D. W. A. Bourne, J. J. Jacobs, A. Awaluddin, D. J. Maddalena, J. G. Wilson, and R. E. Boyd. Physiological modeling of disposition of potential tumor-imaging radiopharmaceuticals in tumor-bearing mice. J. Pharm. Sci. 81:408-412 (1992). 50. D. C. Maneval, D. Z. D'Argenio, and W. Wolf. A kinetic model for 99mTc-DMSA in the rat. Eur. J. Nucl. Med. 16:29-34 (1990). 51. Y. Sugiyarna, D. C. Kim, H. Sato, S. Yanai, H. Satoh, T. Iga, and M. Hanano. Receptor-mediated disposition of polypeptides: Kinetic analysis of the transport of epidermal growth factor as a model peptide using in vitro isolated perfused organs and in vivo systems. J. Control. Rel. 13:157-174 (1990). 52. H. Sato, Y. Sugiyama, Y. Sawada, T. Iga, and M. Hanano. Physiologically based pharmacokinetics of radioiodinated human-13-endorphin in rats. An application of the capillary membrane-limited model. Drug. Metab. Dispos. 15:540-550 (1987). 53. D. G. Covell, J. Barbet, O. D. Holton, C. D. V. Black, R. J. Parker, and J. N. Winstein. Pharmacokinetics of monoclonal immunoglobulin GI, F(ab')2, and Fab' in mice. Cancer Res. 46:3969-3978 (1986). 54. S. M. Somani, S. K. Gupta, A. Khalique, and L. K. Unni. Physiological pharmacokinetic and pharmacodynamic model of physostigmine in the rat. Drug Metab. Disp. 19:655660 (1991). 55. J. M. Gearhart, G. W. Jepson, H. J. Clewell, M. E. Andersen, and R. B. Conolly. Physiologically based pharmacokinetic and pharmacodynamic model for the inhibition of acetylcholinesterase by diisopropylfluorophosphate. Toxicol. Appl. Pharmacol. 106:295-310 (1990). 56. P. C. Hiestanel, M. Graeber, V. Hurtenbach, P. Herrmann, S. Caunisuli, B. P. Richardson. M. K. Ebecle, and S. F. Bire]. The new cyclosporine derivative, SDZ IMM 125; hi vitro and in vivo pharmacologic effects. Transphmt Proc. 24:31-38 (1992). 57. R. Kawai and M. Lemaire. Role of blood cell uptake on cyclosporine pharmacokinetics. In Proc. Int'l. Syrup. on Blood Binding and Drug Transfer, J. P. Tillement and H. Eckert (eds.), Fort et Clair. Paris, 1993, pp. 89-108. 58. F. Y. Bois. L. Zeise, and T. N. Tozer. Precision and sensitivity of pharmacokinetic models for cancer risk assessment: Tetrachloroethylene in mice, rats, and humans. Toxieol. Appl. Pharmaeol. 102:300-315 (1990). 59. D. M. Hetrick, A. M. Jarabek, and C. C. Travis. Sensitivity analysis for physiologically based pharmacokinetic models. J. Pharmacokht. Biopharm. 19:1-20 (1991). 60. Advanced Ctmtinuous Simttlation Language (ACSL), Mitchell and Gauthier Associates, Concord, MA, 1987. 61. E. C. Steiner, P. S. McCroskey, and T. D. Rev. ShmtSolv: Modeling and Sbnulation S~ftware, Reference Guide, Dow Chemical, Midland, MI, 1990. 62. L. J. Notarianni. Plasma protein binding of dri~gs in pregnancy and in neonates. Clin. Pharmacokin. 18:20-36 (1990). 63. S. M. Wallace and R. K. Verbeeck. Plasma protein binding of drugs in the elderly. Clin. Pharmacokin. 12:41-72 (1987). 64. D. Alvarez, R. Mastai, A. Lennie, G. Soifer, D. Levi, and R. Terg. Noninvasive measurement of portal venous blood flow in patients with cirrhosis: Effect of physiological and pharmacological response. Digest. Dis. Sci. 36:82 86 (1991). 65. B. C. Chert, S.-C. Huang, G. Germano, W. Kuhle, R. A. Hawkins, D. Buxton, R. C. Brunken, H. R. Schelbert, and M. E. Phelps. Noninvasive quantification of hepatic arterial blood flow with nitrogen-13-ammonia and dynamic positron emission tomography. J. NucL Med. 32:2219-2228 (1991). 66. J. R. Horn, B. Zierler, L. A. Bauer, W. Reiss, and J. E. Strandness. Estimation of hepatic blood flow in branches of hepatic vessels utilizing a noninvasive, duplex doppler method. J. Clin. Pharmacol. 30:922-929 (1990). 67. F. Weber, M. Antauf, and M. Serdarevic. Noninvasive, quantitative determination of muscle blood flow in man by a combination of venous-occlusion plethsmography and computed tomography. Basic Res. Cardiol. 83:327-341 (1988).
PBPK Modeling as a Drug Development Tool
229
68. A. L. Hinderliter, M. A. Fitzpatrick, N. Schork, and S. Julius. Research utility of noninvasive methods for measurement of cardiac output. Clin. Pharmacol. Ther. 42:419425 (1987). 69. W. Weber, M. Looby, and J. Brockmoeller. Evaluation of hepatic function using the pharmacokinetics of a therapeutically administered drug. Application to the immunosuppresant cyclosporin. Clin. Pharmacokin. 23:69-83 (1992). 70. W. R. Crom, S. L. Webster, L. Bobo, M. E. Teresi, M. V. Relling, and W. E. Evans. Simultaneous administration of multiple model substrates to assess hepatic drug clearance. Clht. Pharmacol. Ther. 41:645-650 (1987).