BACKGROUND OF THE INVENTION ^\' * The use of living cells in culture for biomedical research was first introduced by Dr. Otto Warburg in the ^ 10 1920 ^, s. Today, this technique enables scientists to model specific biochemical or physiological properties of cells under defined and reproducible conditions. Cell culture technology has made possible the study of many diseases independent of either patient or animal models. The obvious

15 advantage of this methodology is that a variety of experiments can be performed while avoiding the moral dilemma of using human beings or animals as research tools.

Currently, both human and animal cell lines serve as surrogates for living organisms in the screening of efficacy

20 or toxicity of pharmaceuticals, agrichemicals and nearly all chemicals entering the home through consumer products. The rapidly proliferating area of in vitro alternatives to animal testing utilizes cell culture techniques to model such toxic responses as skin and eye irritation.

25 While cell culture technology has given to science the ability to perform research not possible at the beginning of this century, there are significant inherent limitations in the technique. These limitations prevent cell culture methods from superseding research on humans or animals as the ultimate

30 predictor of biological response. Exposure and dosing of cell cultures results dosing of cell culture systems currently lacks a physiologically based foundation. As a result, exposure methods currently are not equivalent to the physiologic pattern of exposure encountered by cells.

35 Normally in living organisms, dose regimens of pharmaceuticals and exposure to environmental chemicals result in ^ concentrations being greatest immediately following exposure and declining until the next exposure to the chemical; - metabolic processes within cells respond to the changes in the

SUBSTITUTE SHEET

test chemical. Pharmacological changes or damage induced by a chemical at peak concentrations may be reversed as the chemical concentration falls and the cell may return to pre-exposure status before a subsequent exposure to the chemical.

Within intact biological systems, concentration and time interact to influence the intensity and duration of a desired pharmacologic response or toxic manifestations. An example of the interaction between concentration and time that influences toxic responses of a cell can be given for a biotransformation reaction occurring in many cell types that functions to protect the cell from chemical damage. Glutathione (GSH) is a tripeptide that is metabolically coupled to reactive (harmful) chemicals by an enzyme known as glutathione transferase. The cell utilizes GSH to protect itself against chemical damage. At high exposure concentrations of toxic chemicals, GSH reserves can be depleted. When subsequent exposures occur before repletion of GSH stores, the cell will manifest a toxic response. If, however, the time between exposures to the chemical is sufficiently long to allow for the replenishment of GSH reserves, no toxicity will develop. Thus the frequency of exposure is as critical a determinant of cellular toxicity as the amount of exposure to a chemical. This characteristic waxing and waning of chemical exposure and resultant cellular responses cannot be simulated by the application of a mathematical function to the results of a cell culture experiment performed with static exposure concentrations.

Previously, the only way to model these fluctuations in exposure conditions and multiple exposures was to use animal models or to expose humans directly to the test chemical. However, the multicompartmental cell culture analog device described herein is capable of modeling the chemical absorption, distribution, metabolism and elimination capacity of any species. The device will allow drug or toxicant and

SUBSTITUTE SHEET

metabolite communication among cell and tissue types through the circulating system. Furthermore, the distribution and elimination of chemicals or metabolites among the cells and tissues of can be based upon physiologic relationships. Naphthalene was selected as a model compound because is provides the opportunity to study numerous features of the ulticompartmental cell culture system. Naphthalene is toxic to pulmonary (lung) cells, but the biotransformation of naphthalene is primarily by the liver. Therefore, the aspect of exchange of putatively toxic metabolites can be studies in the multicompartmental cell mammalian culture system. Naphthalene is extremely toxic to mice and less toxic to rats and hamsters. This species specificity can be examined in the multicompartmental cell mammalian culture system by using combinations of mouse x rat cell types to determine whether the mouse pulmonary cells are as sensitive as rat pulmonary cells when mouse or rat hepatocytes are present. Finally, the specific isozymes responsible for the generation of reactive metabolites of naphthalene can be identified by using cells containing only single isozymes of cytochrome P-450 (e.g. cytochrome P-450IA1) in the hepatocyte compartment. These cells are created by transvection of the gene coding for the cytochrome of interest.

A physiologically based pharmacokinetic model of naphtha gene was constructed from previous experiments describing the biotransformation of the kinetics of the chemical in several species. This model was used to predict the behavior of naphthalene in mice, rats and humans. The mathematical determinants of compartmental volume and flow rates between and among compartments are used to hand set the mechanical pumping devices of the multicompartmental cell mammalian culture system. Additionally, the modeling provides the information relevant to compartmental volumes of distribution for the individual cellular compartments.

SUBSTITUTE SHEET

SUMMARY OF THE INVENTION The present invention is directed to a device that allows the exchange of metabolites or test chemicals among compartments containing cells of differing origin at controlled flow rates, providing a method and means to study the metabolic interactions of several cell types in a multiple compartment cell culturing device. The invention provides for more accurate physiologic and kinetic modeling of the distribution of chemicals among cell types. In accordance with this invention, we have constructed devices in which different cell types may exist in separate chambers while exposed to a common cell culture medium. Cells may be of the same source species or a combination of two or more species. One or more pumping devices can be incorporated into the system to allow for controlled flow rates among and between cellular compartments. Additionally, shunts and reservoirs can be added to further control the kinetics of metabolite or test chemical distribution to individual compartments. Specifically, we have demonstrated that a model for the DNA component uridine (bromodeoxyuridine) can be administered to the device and differential uptake of the agent by normal and transformed cells can be assessed. By modeling several doses of the test chemical the optimal ratio of incorporation by neoplastic cells (H4II-E) with respect to normal cells (lymphocytes) could be determined. ^\'

With a knowledge of the volume of distribution (Vd) in the target organism (human being or example) , the optimal dose may be determined directly from the described invention by means of the relationship:

Dose = CO A Vd where CO is the concentration at time zero in the system and Dose is in g/kg. In this way the invention described allows the determination of dose directly in mg/kg as well as an

estimate of the difference between effects on H4II-E cells (target cells) and normal lymphocytes (nontarget cells) .

The model using uridine, described above, can be adapted for the in vitro screening of anti-neoplastic drugs for efficacy against cancer cells. The system described herein permits tumor cells and normal cells to be exposed to candidate anti-cancer drugs simultaneously, which will permit the determination of a drug\'s ability to discriminate between the two cell types, prior to human or animal testing. Drugs which do not discriminate between cancer cells and normal cells can be eliminated immediately.

Uses of the multicompartmental cell culture system of the invention are numerous in addition to the model described above. For example, the culture system of the invention can determine the efficacy and toxicity simultaneously of a potential pharmaceutical drug, so that the therapeutic index of the drug will be immediately known. Accordingly, a drug that may initially appear to be promising, but whose development would eventually be terminated due to toxicity discovered following administration to humans or animals, can be eliminated immediately using the system of the invention.

Another use for the multicompartmental cell culture system is testing pro-drugs for both efficacy and their ability to be metabolically transformed into the active moiety. In one embodiment of the multicompartmental cell culture system, the system would contain hepatocytes with human biotransforming enzymes, as well as target cell types. In this manner, the multicompartmental cell culture system of the invention will mimic the actual metabolism of a drug in a human or animal system.

The system is also designed to permit determination of multiple dose regimens, since drugs are administered on the same mg/kg basis used for human drug dosages and clearance rates are physiologically based. Such multiple dose

SUBSTITUTE SHEET

experiments would allow for the determination of additive as well as synergistic effects, such as enzyme induction or inhibition.

The multicompartmental cell culture system of the invention, besides use for screening and studying the metabolic profile of a drug, can also be used to identify interactions with other drugs or foods and food additives that effectively halt development of a drug or a biologically active compound in the final stages of clinical testing. Another use for the multicompartmental cell culture system of the above invention is the determination of the effects of cellular metabolites of test chemicals on secondary cells, which would prove useful for the agrichemical and chemical industry. In this embodiment, the multicompartmental cell culture system of the invention would have at least two cultures in separate chambers, namely hepatocytes (for example, human, rat or mouse) and pulmonary Clara cells (for example, human, rat or mouse) . Reactive metabolites produced by the hepatocytes in the first culture may circulate and affect the pulmonary cells in the second culture, provide an in vitro, interactive model which mimics the intact mammalian system.

By modeling the chemical absorption distribution, metabolism and elimination of animals or humans, a physiologically-based pharmokinetic model (PBBK) system can be developed. The system of the invention allows for drug, toxicant and metabolite communication among various cell and tissue types through a circulating system. The distribution and elimination of chemicals and their metabolites among the cells and tissues can be based upon human or animal physiological relationships. This system allows a close resemblance between the dose regimens of pharmaceuticals and exposure to environmental chemicals, as would be experienced in living systems. In the living system, these dose regimens and exposures would result in concentrations being greatest

immediately following exposure and declining until the next exposure to the chemical. Metabolic processes within cells respond to the changes in concentration of the test chemical. Pharmacological changes or damage induced by a chemical at peak concentrations may be reversed as the chemical concentration falls and the cell may return to preexposure status before the subsequent exposure to the chemical. The frequency of exposure is as critical a determinant of cellular toxicity as the amount of the exposure to a chemical. This characteristic waxing and waning pattern of chemical exposure and resultant cellular responses cannot be simulated by the application of mathematical functions to results of a cell culture experiment performed with static exposure concentrations, but this characteristic waxing and waning pattern can be produced in the multicompartmental cell culture system of the invention. The monitoring and adjustments made to the model system of the invention permit a close resemblance to animal or human models without resorting to the use of living organisms. To model the metabolic capacity of humans for any particular drug or chemical, a bank of four to five of the multicompartmental cell culture system units can be set up, each system representing a particular percentile of the population. Due to the genetic variation of humans with respect to xenobiotic metabolism, each of the units is set up to model the metabolism of a particular population segment. Segments can include the elderly, neonates, and pediatric population, as well as account for difference due to gender, ethnicity, or physical condition. Cells from any species of organism, which can be adapted to tissue culture, can be used in the system. These organisms can include human beings, laboratory animals such as rats, mice, and hamster, domestic farm laboratory animals, fish, insects, plants, unicellular organisms, and viruses. The choice of organisms is not meant to be limited by the

preceding list. Particular embodiments can mix complete organisms in one or more compartments with cell, tissue, or organ cultures in one or more of the other compartments.

Cells from an organism with a particular condition can also be selected for inclusion within the system. Age, ethnicity, gender are other aspects that can be considered in cell selection.

In all of the embodiments described above, besides adding more cell and/or organ culture types, the system can also include cells that have been engineered in various ways. For example, the cells used could be chimeric species (for example, human x mouse combinations) or can be cells transvected with foreign or altered genes.

This invention provides methods and means that allow cells of different species to be exposed simultaneously to chemicals or metabolites under conditions of exchange that may be altered to model one or more than one species.

Furthermore, the multicompartmental cell culture system described above can be automated by attaching a microprocessor to control the circulation of media, metabolites and drugs within the system. Along with the microprocessor, various containers, sensors, probes and measuring devices can be included as part of the overall system to control or vary the cellular environment. Means are included to provide computer control and monitoring of the cell interactions in the multiple compartment cell culturing device. The computer-based system provides substantial flexibility in testing, as by enabling programmably varying the number and types of cell culturing compartments, the configuration of fluid paths, fluid flow rates, parameters to be monitored, and the manner of recording and displaying test results.

BRIEF DESCRIPTION OF THE DRAWINGS The following detailed description, given by way of example, will best be Understood in conjunction with the accompanying drawings, described as follows. Fig. la. A simplified schematic of the multiple cell compartments of the apparatus of Fig. lc.

Fig. lb. A simplified schematic of the multiple cell compartments of the appratus of Fig. lc, with the addition of a compartment for kidney cells and provision for both arterial and venous blood flow.

Fig. lc. A schematic diagram of an apparatus applying the principles of this invention, showing the multiple cell compartments connected to a reservoir.

Fig. 2. Bar graph depicting the viability of human lymphocytes and H4II-E cells following 24 hr exposure to BrdU in the device described in Figure 1 and 30 sec exposure to UV light;

Fig. 3. Biotransformation of naphthalene and naphthalene oxide; Fig. 4. Structure of PBPK for naphthalene and naphthalene oxide;

Fig. 5. GSH resynthesis in mouse hepatocytes after 60 gM 1A, 25-naphthalene oxide is added to a suspension of 1 millioϊ. ^v iepatocytes/ml. Simulation. Data of Buonarati et al., 19B9;

Fig. 6. GSH resynthesis in mouse lung following ip administration of 200 mg/kg naphthalene. Simulation 1,

KLUGSS=0.163 nmole/min/min, KLUGD=0.OO6/min. Data of Warren et al., 1982; Fig. 7. Covalent binding in mouse lung four after ip dosing with naphthalene. Simulation. Data 26,

1-1 of Warren et al., 1982. *Data of Buckpitt and Warren 1983 (100, 200, and 400 mg/kg) and O\'Brien et al., 1989 (300 mg/kg) ;

SUBSTITUTE SHEET

Fig. 8. Covalent binding in mouse liver four hours after ip dosing with napthalene. Simulation. Data of Warren et al., 1982. *Data of Buckpitt and Warren 1983 (100, 200, and 400) mg/kg) and O\'Brien et al. , 1989 (300 mg/kg) ; Fig. 9. GSH levels in mouse lung four hours after ip dosing with naphthalene. Simulation. *Data of Warren et al., 1982;

Fig. 10. GSH levels in mouse liver four hours after ip dosing with naphthalene. Simulation. *Data of Warren et al., 1982;

Fig. 11. Measured NPSH levels and simulated GSH levels in the lung four hours after ip injection of naphthalene into a rat. Simulation. Data of O\'Brien et al., 1985; Fig. 12. Measured NPSH levels and simulated GSH levels in the liver four hours after ip injection of naphthalene into.a rat. Simulation. *Data of O\'Brien et al., 1985;

Fig. 13. Sensitivity of naphthalene disposition to initial GSH level in the mouse liver. 4.1 mM 6.6 mM 8.6 mM; Figs. 14 and 15. Sensitivity of covalent binding levels (four hours after ip injection of naphthalene) to initial GSH level in mouse liver. 4.1 mM 6.6mM 8.6 mM.

Fig. 16. A schematic block diagram of the system of the present invention;

Fig. 17. A diagram illustrating a large number of cell compartments interconnected by a conduit system with reconfigurable fluid paths;

Fig. 18. A flow chart of major components of a software program for a system of the present invention; and Figs. 19a, 19b, 19c, 19d and 19e. Flow charts of subroutines associated with the software of the present invention.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

EXAMPLE 1

Simultaneous Exposure of Normal and Transformed Cells to an Antineoplastic Compound in a Multicompartmental Cell Culture System

Chemicals

All chemicals were purchased from Sigma Chemical Company (St. Louis, MO) unless otherwise stated and were of the highest purity available.

Cell Cultures

Approximately 14 L of freshly drawn human blood were collected into EDTA tubes (Becton Dickinson, Lincoln Park, NJ) . To this blood sample was added 4 L of Hanks balanced salt solution. Two and one-half mL of the blood solution was layered onto 3 mL of ficoll (Pharmacia, Piscataway, NJ) and centrifuged at 1500 rpm for 20 min at ambient temperature. The pellet was washed in an equal volume of Hanks balanced salt solution and recentrifuged. All lymphocytes collected from the 14 mL blood sample were seeded into 100 mL of RPMI 1640 (Gibco, Grand Island, NY) and placed in a Forma Scientific waterjacketed incubator (Fisher Scientific, Philadelphia, PA) set at 370C with 5 percent C02 overnight.

A rat hepatoma cell line, H4II-E obtained from the American Type Culture Collection (Bethesda, MD) , was cultured in RPMI Medium 1640 with 5 percent fetal calf serum, L-glutamine and antibiotics (penicillin G, streptomycin and amphotericin B) (Gibco, Grand Island, NY) . Cells were seeded 24 hr before the experiment. Attachment of H4II-E cells occurs several hours following seeding. Density was approximately 20 percent of confluence at the start of the experiment.

Multicompartmental cell culture system

The multicompartmental cell culture system (MCCS) as described in Fig. la-c generally was used to expose both

SUBSTITUTE SHEET

normal and transformed cells to an antineoplastic compound simultaneously. The volume of distribution of the system was 500 mL. Each compartment (T-flask, Becton Dickinson, Lincoln Park, NJ) contained 75 mL of cell culture medium (RPMI 1640 as previously described) . The flow rate was set at 5 mL per min and maintained by means of a peristaltic pump (Millipore, Bedford, MA) . The entire system was placed inside a Forma Scientific water-jacketed incubator (Fisher Scientific, Philadelphia, PA) set at 370C with 5 percent C02.

Treatment of cells

One hour after connecting the compartments, the MCCS was treated with 5-bromo-2\'-deoxyuridine (BrdU) by injecting 1 mL of dimethylsulfoxide containing 750 mg of BrdU into the reservoir (Fig. lc) . The system was allowed to run for 24 hr. Separate T-f lasks containing untreated lymphocytes and hepatoma cells were incubated concurrently as total viability controls.

Following the 24 hr exposure period, both human lymphocytes and rat hepatoma cells were removed from their compartments, placed on a hemocytometer and exposed to a UV light source for 30 sec. Viable cells were determined by trypan blue dye exclusion immediately following UV exposure.

Results

Eighty-nine percent of the human lymphocytes survived the UV exposure, while only 53 percent of the H4II-E cells were viable following UV exposure (Fig. 2). This enhanced killing was due to greater uptake of the BrdU by the hepatoma cells than the lymphocytes. This was to be expected, since BrdU is incorporated into the DNA of cells during DNA synthesis, and the rate of DNA synthesis of transformed cells such as the H4II-E cells is much greater than the rate of DNA synthesis of nontransformed, slowly dividing cells such as the lymphocytes.

S

These results indicate that it is possible to expose different cell types in the MCCS to a test agent and demonstrate differential uptake of the agent in the device.

Additionally, since the BrdU serves as a model for the common metabolite uridine, the experiment supports the use of the

MCCS to allow different cell types to exchange metabolites.

Furthermore, by altering the flow rates on the pumping device or including individual pumps for each chamber, kinetic modeling of the system is easily performed. A practical use for the MCCS is to provide a device in which several cell types may be exposed to a test chemical simultaneously in order to determine differential toxicity of the test chemical to the cells.

Illustratively, methods such as the following can be employed to identify potential antineoplastic agents that are effective against transformed cells, but have low toxicity to normal.

Additionally, such methods as the following can be used to described the kinetics of a drug or test chemical in the MCCS.

SUBSTITUTE SHEET

EXAMPLE 2

Physiologically Based Pharmacokinetic Model for Naphthalene To model the metabolism of a living organism, the MCCS system can be modified to account for the expected metabolic functions. In order to do this, it is necessary to thoroughly understand the metabolism of certain cell types and compounds. The following example using naphthalene, describes how to work out the equations necessary to set up a physiologically based pharmokinetic (PBPK) model. Once these equations have been derived and testing, then the MCCS system would be set up with the required measuring devices such as probes and spectrophotometer to monitor the metabolism of the system. To control the metabolism of the system, additional reservoirs would be added in order to maintain or adjust the gas mixture, acidity, and to add various compounds or metabolites as required to completely model a physiologically-based system. The entire setup would be controlled by a microprocessor programmed to account for both the metabolic equations and to adjust the various parameters in order to closely maintain a metabolic model of a human or animal system.

Naphthalene is a commercial ly important compound produced from coal tar and petroleum. The toxicology of naphthalene shows unusual species and tissue specificity. The rodent LD50s are 380 mg/kp ip in male Swiss-Webster mice

(Shank et al. , 1984), 533 and 710 mg/kg po for male and female CD-I mice, (Shopp et al., 1984), as compared to 2200 and 2400 mg/kg po for male and female Sherman rats (Gaines, 1969) . Tissue binding is not necessarily indicative of toxicity or metabolism at that site, with similar binding levels occurring in lung and liver, but toxicity is limited to the Clara cells of the lung. For the lung, increases in binding are associated with increasing severity of tissue damage. Sensitivity of the target cell or circulation of reactive metabolites from the liver to the lung have been proposed as

SUBSTITUT

explanations for the toxicity (Buckpitt and Franklin, 1989) . A schematic representation of naphthalene and naphthalene oxide biotransformation are depicted in Fig. 3.

We have developed a physiologically based pharmacokinetic model (PBPK) for naphthalene. Previous PBPKs have been used to explain differences in toxicity for different species and routes of administration and to explore possible mechanisms of toxicity (e.g. Andersen et al., 1987). Sufficient in vitro data for mouse tissue exists to construct such a model for naphthalene. The glutathione (GSH) status of the tissues is important in naphthalene toxicology, and the model for glutathione depletion and resynthesis developed by D\'Souza et al. (1988) has been incorporated into this model.

Our model allows the reactive metabolites, the two naphthalene oxide enantiomers, to circulate throughout the body, unlike previous PBPKs where metabolites have been restricted to the tissues in which they are generated. Our model is essentially a system of parallel PBPKs which are bridged by the biotransformation of naphthalene to naphthalene oxide in the lung and liver.

This model is valuable both as a way to compare species and route differences for naphthalene toxicology, and because it can serve as a model for other compounds with circulating metabolites. Circulating metabolites have been offered as explanations for the toxicity of 3-methylindole (Yost et al., 1990), dichlorethylene (Okine and Gram, 1986), and bromobenzene (Casini et al., 1986).

METHODS Model Structure

The model structure is similar to that of D\'Souza et al. (1988), and is depicted in Fig. 4. Equations describing the system appear in Appendix B. A Fortran computer program was written to run on a VAI station using IS 5.4. The differential equations are integrated numerically using LSODA

SUBSTITUTE SHEET

(Livermore Solver for Ordinary Differential Equations, with Automatic Switching Method for Stiff and Non-stiff Problems, Linda R. Petzold and Alan C. Hindmarch, Lawrence Livermore National Laboratory, Livermore, CA) .

Determination of Anatomical Parameters

A model based on a 22-g mouse was developed. The blood volume was calculated from the allometric relation in Kaplan et al. (1983) and divided between venous and arterial blood 2:1 (Gearhart et al., 1990). The weight of the lung, liver, kidney and fat were also calculated from allometric relations (Calder, 1984) . Andersen et al. (1987) calculated the volume of rapidly and slowly perfused tissues as totaling 83 percent of the body weight. The "other tissues" in our model are assumed to weigh 83 percent of the body weight less the weight of the blood and kidney.

The cardiac output was calculated as in Kaplan et al. (1983) and the renal blood flow as in Calder (1984). The liver and fat are assumed to receive 24 and 5 percent of the cardiac output (Andersen et al., 1987). The remainder of the cardiac output flows to the other perfused tissues.

Anatomical parameters for a 220-g rat are the same as those used in Gearhart et al. (1990) with the volumes and flows for rapidly and slowly perfused tissues, brain, and diaphragm summed for our lumped "other tissues" compartment. The parameters used in simulation are summarized in Table 1.

TABLE 1

Anatomical Parameters Used in Naphthalene PBPK

Determination of the Tissue: Blood Partition Coefficients From the known solubility characteristics of naphthalene in air, water (Vargaftik, 1975) and octanol-water (Hansch and Leo, 1979) systems, it is possible to preduct tissue:blood partition coefficients were calculated for lung (0.627), liver (5.41), kidney (3.87) fat (796) and muscle (4.13). The partition coefficient for muscle was used for the other tissues compartment of our model. The same partition coefficients were used for naphthalene and napthalene oxide. The partition coefficient for lung was found to be inconsistent with levels of covalent binding occurring in this

SUBSTITUTE SHEET

tissue. For simulations a value of 4.0 was used instead. This value is more in line with the partition coefficients for the other non-adipose tissues. In PBPKs the same partition coefficient is usually used for all well-perfused tissues, 5 including the lung (e.g. D\'Souza and Andersen, 1988, Ward et al., 1988).

Determination of GSH resynthesis parameters

Steady state values of 2.2 and 6.6 mM were used for

10 lung and liver GSH concentrations in the mouse. The lung value is an average of two literature values (O\'Brien et al., 1985 and Warren et al., 1982). The liver value is an average of several appearing in the literature ranging from 4.1 mM (Buonarati et al., 1989) and 8.6 mM (Richieri and Buckpitt,

151988) .

Initially, the parameter values for the model were calculated as in D\'Souza et al. (1988) , but the different body and organ weights and desired steady state levels lead to discrepancies. The values for synthesis of GSH synthetase 0 (KLUGSS and KLIGSS in the model) (see Appendix A for nomenclature) were adjusted to yield the desired steady state GSH concentrations. This adjustment was subsequently demonstrated to be satisfactory for the liver but not the lung. Adjusting the rate of GSH degradation in the lung 5 (KLUGD) produced better results. For the rat, calculating the GSH resynthesis parameters as in D\'Souza et al. (1988) gave acceptable steady state GSH levels. The kinetic constants used in the GSH model are summarized in Table 2.

0

SUBSTITUTE SHEET

TABLE 2

Kinetic Constants Used In GSH Model

10

15

20

25

Determination of Kinetic Parameters for Naphthalene Biotransformation: Phase I Reactions in the Mouse

30 Biotransformation rates of naphthalene at initial concentrations of 0.05, 0.1 and 1.0 mM by lung and liver microsomes have been determined (O\'Brien et al., 1985, Buckpitt et al., 1984, and Buckpitt et al., 1987). Assuming Michaelis-Menten kinetics, fitting s/v vs s was used to

35 provide Vmax and Km values. Since mouse liver produces the 1- , 2S- and IS, 2--naphthalene oxide enantiomers in equal amounts at all substrate concentrations (Buckpitt et al. , 1987) , the calculated Vmax was divided by two to give the production rates for each enantiomer.

«- ^A For the lung, the biotransformation of naphthalene was modeled as the sum of two separate reactions, each producing a different enantiomer, with the Vmax\'s summing to

the value calculated as described above. The 1R, 2S- naphthalene oxide/lS, 2R- naphthalene oxide ratio (RS/SR) is 30:1 and 10:1 for naphthalene concentrations of 0.015 and 1.0 mM (Buckpitt and Frankliin, 1989, and Buckpitt et al., 1987). 5 Vmax\'s and Km\'s were fit to the reaction rate and ratio data by trial and error.

Determination of Kinetic Parameters for Naphthalene Biotransformation: Phase II reactions in the Mouse

10 The K values used for conversion of 1R,2S- and

IS,2R-naphthalene oxide to l, 2 dihydrodiol by epoxide hydrolase (EH) were those determined for rat liver microsomes

(van Bladeren et al., 1985). Vmax\'s for conversion of each enantiomer in the lung and liver were determined by trial and

15 error fit to data for incubations of naphthalene and microsomes, with no GSH present (O\'Brien et al. , 1985). For determining the EH Vmax\'s, the experimental naphthalene concentration profiles were used rather than the values predicted by fitting a number of experiments (as described in

20 the previous section) .

The rate of reaction of naphthalene oxide with GSH is assumed to be of the form V= Vmax*(lGSH /(GSH +

Kml)}*(N0/(N0+K2)}. Nonenzymatic formation of naphthalene oxide-GSH conjugates is slow or non-existent (Garle and Fry,

251989) . Work with partially purified sheep liver glutathione-S-arene oxide transf rase has provided K\'s for conjugation of GSH and racemic naphthalene oxide at pH 7.4

(Hayakawa et al., 1974) . Individual glutathione-S-transferase isozymes are enantioselective with respect to naphthalene 0 oxide (O\'Brien et al. , 1989), but GSH conjugation proceeds equally fast with 1R, 2S- and IS, 2R-naphthalene oxide

(Buonarati et al., 1990). Therefore Vmax\'s for conjugation with each enantiomer are assumed equal, and the Km for reaction with each enantiomer is assumed to be one-half of the 5 value reported for racemic naphthalene oxide.

The naphthalene oxide-GSH conjugation Vmax\'s were determined by fitting data from incubations of lung and liver microsomes and cytosolic protein with GSH or nepatocyte cultures with concentrations of naphthalene ranging from 0.005 to 1.5 mM (Buckpitt et al., 1987, Buckpitt et al. , 1984, and Richieri and Buckpitt, 1987) .

Determination of Kinetic Parameters for Naphthalene Biotransformation in the Rat Determination of Vmax\'s and K\'s for naphthalene oxide production and Vmax\'s for GSH conjugates and dihydrodiol production in rat microsomes were done in the same way as for the mouse. Sources of data were Buckpitt et al., 1987, Hesse and Mezger, 1979, O\'Brien et al., 1985, van Bladeren et al., 1985 and Yost et al., 1989.

Non-enzymatic Reactions of Naphthalene Oxide

The first order rate constant for non-enzymatic rearrangement of naphthalene oxide to 1-naphthol has been reported (van Bladeren, et al., 1985). The rate of covalent binding was determined from data reported for incubations of hepatocytes with naphthalene oxide (Buonarati et al, 1989) .

Scaling reactions to whole-organ rates To scale reaction rates based on microsomal and cytosolic protein, conversion factors are used. The same values were used for mouse and rat. Levels of 3.67 and 16.4 mg microsomal protein /g of tissue have been reported for rat lung and mouse liver (Rietjens et al., 1988, and Garle and Fry, 1984). The levels of cytosolic protein are 30.5 and 74 mg/g in these tissues (Rietjens et al., 1988, and Disimplico et al., 1989), but Buckpitt et al. (1984) showed that increasing the cytosolic/microsomal protein ratio above 0.5 does not affect the product distribution in the lung, and increasing the ratio above 1.1 has no effect in the liver. Thus reaction rates on the basis of cytosolic protein are

SUBSTIT

scaled as if there were 1.8 and 16.4 mg/g of cytosolic protein in the lung and liver.

The amounts of total protein are 13.8 mg/mouse lung (Kanekal et al., 1990) and 112 mg/mouse liver (Cha and Bueding, 1979) . For the rat, these values are multiplied by the ratio of the organ weights in rat and mouse.

Oral uptake rates

Uptake from an oral dose was modeled as first order with respect to the amount yet to be absorbed. D\'Souza and Andersen (1988) report that oral absorption rates of 1.0 hr are typical for halogenated hydrocarbons administered po in a corn oil vehicle, and used 0.5- 1. 0/hr (0.008-0. 017/min) for vinylidene chloride. An initial estimate in this range was used for naphthalene.

Bioavailability

After oral administration of 2 mg naphthalene to a ra _r 87-91 percent of the dose was recovered as metabolites within 72 hrs (Bakke et al., 1985). Eighty-one percent of a 45 mg/kg dose of 1 naphthol was recovered in 72 hours. Other aromatics have demonstrated bioavailabilities of 57-70 percent (Brouwer and McNamara, 1989, Bakke et al., 1990). Oral bioavailability of 70 and 81 percent were used in simulation.

Simulation of whole animal toxicology

In addition to those equations needed to describe tissue concentrations of naphthalene, the naphthalene oxides and GSH, amounts of naphthalene converted to dihydrodiol and GSH-conjugates or covalently bound were also calculated. See Appendix B for the equations.

In mice retreated with buthionine sulfoximine (BSO) hepatic and pulmonary GSH are depleted. The proposed mechanism is that BSO blocks GSH synthesis Griffith and Meister, 1979) Thus to model the effect of BSO pretreatment,

the GSH synthesis rate is set equal to zero, and the initial levels of GSH set to the experimentally measured values. The natural turnover of GSH by degradation is allowed to continue, in addition to depletion by conjugation with naphthalene oxide.

Forty-eight hours after administration of xylene, GSH levels and metabolism of naphthalene by liver microsomes return to normal, but metabolism by lung microsomes proceeds at 56 percent of the normal rate (Buckpitt and Warren, 1983) . To model the effect of pretreatment with xylene, VLUNRS and VLUNSR were multiplied by 0.56.

Sensitivity of Mouse PBPK

Three test cases were simulated and compared to the base model to test sensitivity of the model to various parameters. These cases were disposition of a lOOmg/kg ip dose of naphthalene and the levels of covalent binding in lung and. liver resulting from 200 and 400 mg/kg ip doses of naphthalene. The parameters tested for sensitivity included steady state levels of GSH, cardiac output, blood flow to the fat, tissue volumes, and partition coefficients.

RESULTS

Kinetic Parameters for GSH Synthesis The time courses of glutathione recovery in hepatocytes and the lung are shown in Fig. 5 and Fig. 6.

Kinetic Parameters for Naphthalene Biotransformation in the Mouse and Rat Biotransformation parameters are summarized in

Table 3. Comparisons between measured in vitro data and model predictions are found in Tables 4-6.

TABLE 3

Parameters For Napirthalene And Naphthalene Oxide Biotransformation

Non-enzymatic reactions:

First order rate constant for rearrangement to 1-napthol: 0.25/min Binding to protein: 0.2 pmole/mg/min/μM NO

TABLE 4

Biotransformation of napthalene by mouse lung microsomes

v: nanomoles of product/mg microsomal protein/minute ratio: ratio of conjugates of 1R, 2S- and IS, 2R- napthalene oxide

\'Data of Buckpitt and Franklin, 1989 ^b Data of O\'Brien et al. , 1985 ^c Data of Buckpitt et al., 1984 ^d Data of Buckpitt et al., 1987

TABLE 5

Mouse microsomes, incubated with 0.1 mM naphthalene, after 6 minutes.

concentrations in nmole/mg ^c Data of Buckpitt et al., 1984

SUBSTITUTE SHEET

Rates in nmole/mg/min

\'Data of Buckpitt et al., 1984

Oral Uptake

Table 7 presents observed levels of mercapturates with predicted amounts of glutathione conjugates. An initial GSH concentration of 6.8 mM was used for rat liver to match the values measured experimentally. Six and one-half hours after administration of a 200 mg/kg oral dose, the liver GSH was 17 percent of the initial concentration (Summer et al. , 1979) . When 1, 2, and 4 hours are used as times for absorbing half the naphthalene, simulation predicts GSH levels 70, 36, and 30 percent of the initial level 6 1/2 hours after administration of naphthalene.

SUBSTITUTE S ET

TABLE 7

GSH conjugations subsequent to po administration of naphthalene in corn oil ^\' vehicle to rats.

Dose Percent Percent GSH conjugates in recovered as time for absorption of 1/2 dose mg/kg mercapturates\' lhr 2hr 4hr

200 26 55.8 58.3 60.2 75 32 61.5 59.7 58.7

30 39 58.9 58.0 57.4

\'Data of Summer et al., 1979.

Simulation of Whole Animal Toxicology Following administration of 100 mg/kg naphthalene to a mouse by intraperitoneal injection, 39 percent of the dose was recovered as mercapturic acids (Stillwell et al., 1982). By dosing directly with naphthalene oxide-GSH conjugates, ranges for recoveries of each of these conjugates as mercapturic acids have been determined for mice (Buonorati et al., 1990). Assuming conjugates 1, 2, and 3 are produced in a 1:2:1 ratio, the high, average, and low recoveries as mercapturates yield estimates of 61.7, 57.3 and 53.8 percent GSH conjugates resulting from a 100 mg/kg dose. Simulation predicts 64 percent.

Simulated and experimental values for covalent binding following ip dosing of a mouse are shown in Figs. 7 and 8. GSH values for these simulations and experiments are shown in Figs. 9 and 10. Figs. 11 and 12 show experimentally determined depletion of nonprotein sulfhydryls (NPSH) and predicted GSH levels following ip administration of naphthalene to a rat.

Treatment of mice with BSO depletes GSH in the liver more than in the lung, in naphthalene dosing experiments, GSH levels in lung and liver were 86 and 35 percent of the control values before administration of a 200 mg/kg dose of naphthalene (Buckpitt and Warren, 1983) . The results of this

SUBSTITUTE SHEET

experiment were compared to the simulation. Simulation predicts an increase in covalent binding from 202 to 510 pmole/mg (factor of 2.52) in the lung and the experiments show an increase from 137 to 411 pmole/mg (factor of 3.0). in the liver, predicted binding levels are 156 and 1387 pmole/mg, and the observed levels are 255 and 1069 pmole/mg.

Simulation of pretreatment with xylene showed no significant changes in covalent binding or GSH depletion following 300 mg/kg naphthalene ip. Similar results have been observed experimentally (Buckpitt and Warren, 1983) .

A comparison of covalent binding levels determined by experiment and simulation at or near the LD50s appears in Table 8.

TABLE 8

Comparison of covalent binding in the lung for LD50s.

Binding in pmole/mg Male mouse, po, bioavailability 70% 335, 544 Female mouse, po, bioavailability 81% 668, 842 Male rat, po, bioavailability 70% 277, 304 Male rat, po, bioavailability 81% 296, 321 Male mouse, ip (simulated) 453

Male Swiss-Webster mouse, ip, 300 mg/kg° 675 Male Swiss-Webster mouse, ip, 400 mg/kg 912 Oral uptake rates of 0.01155 and 0.005775/min \'Data of O\'Brien et al., 1985 ^b Data of Warren et al. , 1982

Sensitivity Analysis Three test cases were chosen for sensitivity analysis of the mouse PBPK. These were the overall disposition of a 100 mg/kg ip dose of naphthalene and the results of administration of 200 and 400 mg/kg naphthalene ip, especially the covalent binding levels resulting form these two doses.

SUBSTITUT

Decreasing the initial concentration of GSH in the lung from 2.2 mM to 1.3 mM, the lowest value reported in the literature (Richieri arid Buckpitt, 1988) did not significantly change the results of the simulation. The effects of changing the initial concentration of liver GSH are shown in Figs. 13-15. The GSH levels used (4.1 and 8.6 mM) are the lowest and highest values found in the literature (Buonarati et al, 1989, and Richieri and Buckpitt, 1987) . Doubling or halving the blood volume did not significantly affect the results.

The highest prediction of cardiac output for a mouse, 27.7 ml/min, results from scaling the cardiac output used by Andersen et al., 1987. Percentages of cardiac output flowing to each tissue group were kept constant. For the 100 mg/kg dose, the concentration in the fat two hours after administration increases from 5.1 mM to 8.1 mM when the cardiac output is increased. Decreases in binding are observed at the 200 and 400 mg/kg doses. At the lower dose, the lung and liver binding levels drop to 39 and 66 percent of the values determined in the base model. For the higher dose, binding levels decrease to 43 and 41 percent of the lung and liver binding in the base model.

Increasing the blood flow to the fat and decreasing blood flow to the "Other Tissues" while keeping cardiac output constant also increases concentrations in the fat and decreases covalent binding.

A change in lung weight does not significantly affect covalent binding, but a shift in the RS/SR ratio for the overall disposition of a 100 mmg/kg dose of naphthalene does occur, if the lung weight is decreased to 0.125 gl the ratio of conjugate 2 to conjugates 1 and 3 is 1.03, but if it is increased to 0.5 g the ratio increases to 1.11.

Increasing the fat: blood partition coefficient from 796 to 7960 had virtually no effect. When decreased to 79.6,

SUBSTITUTE S

the naphthalene concentration in the fat decreased from 5.1 to 3.2 mM two hours after administration of a 100 mg/kg dose.

DISCUSSION

By slightly modifying the GSH resynthesis model of D\'Souza et al. , (1988) we were able to match experimental observations quite well. A good fit was achieved in terms of 5 levels of GSH and the known characteristics of GSH resysnthesis—slight overshoot for hepatic GSH before declining to steady state levels, and a more gradual increase in lung GSH without the over shoot (D\'Souza et al., 1988). One of the modifications, decreasing the degradation rate of

10 lung GSH has a physiological basis—the turnover time for GSH in the lung is longer than in the liver (Griffith and Meister, 1979) .

The parameters for naphthalene biotransformation in the mouse lung give a better fit to the RS/SR ratio than to

15 reaction rates. A compromise between the two was necessary, and it was decided that the ratio was the more important characteristic. The high RS/SR ratio is what is distinctive about metabolism in the mouse lung, compared to other tissues and other species. Also, the simulation of xylene pretreatment

20 demonstrates that the reaction rate in the lung does not have a major impact on the overall process.

Brief incubations with lung microsomes can be somewhat misleading, as can be seen in Table 5. The simulation shows that fairly high levels of naphthalene oxide

25 are present at six minutes, in these incubations, the reaction is "stopped" at six minutes by addition of ice cold methanol (Buckpitt et al., 1984). While this should effectively stop the enzymatic reaction, nonenzymatic reaction to covalently bound adducts and 1- naphthol will continue, albeit at a much

30 slower rate. if assays are not done promptly, the naphthalene oxide will rearrange, giving inappropriately high levels of 1- naphthol and covalently bound adducts.

We found that a relatively rapid oral uptake was needed to see much difference in GSH conjugation between the

35 75 and 200 mg/kg doses (Table 7) . However, if the uptake was

SUBSTITUTE SHEET

too rapid, the recovery of the liver GSH would be well on its way by 6-1/2 hours after po administration. As a compromise, values of 1-2 hours for ^\' absorption half-time were used in other oral simulations. The PBPK simulated whole animal toxicology quite well. The estimate for GSH conjugates of a 100 mg/kg dose was a bit high, but decreasing the initial value of liver GSH from 6.6 to 4.1 mM decreases the percent recovered as GSH conjugates from 64 to 57 percent. This is in the range predicted by combining the results of Stillwell et al. , 1982 and Buonarati et al., 1990.

The simulations of covalent binding over a range of doses are also quite accurate. The sharp increases in covalent binding occur between 200 and 400 mg/kg for both the experiments and simulations (Figs. 7 and 8) and the simulation predicts the binding levels in the liver quite well. The model underpredicts covalent binding and over predicts GSH levels in the lung at higher doses (Figs. 7 and 9) . This suggests that the model assumption of equilibrium between blood and lung for naphthalene oxide concentrations may not be entirely accurate, or that the partition coefficient is too low.

The simulation of pretreatments also matches the available in vivo data. Similar increases in binding after depletion of GSH were observed in the lung for simulation and experiment, in terms of percent increase, the liver binding was way off, but the nature of the increase was the same (binding levels of less than 300 pmole/mg without BSO, over 1000 pmole/mg with BSO) . Detailed whole animal and in vitro data for the rat is not as readily available as for the mouse, but the model does make a fairly accurate prediction of GSH levels in the rat after ip dosing (Figs. 11 and 12) .

Table 8 summarizes binding levels measured and predicted for doses at or near the LD50s. Since all the

simulations have the same weaknesses, it is probably more reliable to compare the simulations to each other than each to the experimentally determined values. All the simulations for male rodents show similar levels of binding in the lung. This suggests that the differences in rodent sensitivity to naphthalene are due to the pharmacokinetics, not to differences between mouse and rat target cells (Clara cells) .

To summarize, we have developed a PBPK which accurately models the toxicology of naphthalene, providing a way to study species and route differences. This was accomplished by incorporating the circulation of reactive metabolites with existing GSH synthesis model. This model can be a paradigm for other compounds which exhibit similar mechanisms of toxicity.

EXAMPLE 3

Automated. Multicompartmental Cell Culture System

Turning to the drawings, Fig. la pictorially illustrates a housing 10 containing various cell culturing compartments including compartments 12, 14, 16 and 18 which simulate, respectively, the lungs, liver, tissue and fat of a species, in certain embodiments, the cell culture compartments may contain cell cultures from more than one species of organism, for example, human lymphocytes in one compartment and rat hepatoma cells in another compartment. The choice of cell type, such as kidney versus pancreas can also be varied as needed. The displays 20 enable internal circuitry (not shown) to output and display various test results as shown. Reference numeral 22 designates pushbuttons, levers or the like through which an operator may control the system in housing 10.

Fig. lb is a schematic illustrating the circulatory system of the present invention that is designed to simulate the flow of arterial blood from a chamber 24, via an arterial blood supply network 26, to the aforementioned cell

SUB

compartments 12, 14, 16 and 18, and also to a kidney cell compartment 28. Media containing a test chemical supplied from the chambers 24 to ^\' the cell compartments 14, 16, 18 and 28 flows into a venous blood compartment 30 via a venous blood supply circulatory system 32. The circulatory path then proceeds to the cell compartment 12 simulating the lungs, finally returning to the chamber 24.

In Fig. 4, fluid paths from a gut chamber 34 and a peritoneal cavity 36 are added to the circulatory system depicted in Fig. lb.

Figs. 2 was previously described in Example 1 and Figure 3 and 5-15 were previously described in Example 2. Fig. 16 is a simplified block diagram showing a system with two cell culturing compartments, and a computer controlled circulation system. This system essentially constitutes a tissue culture incubator with, compartments, optionally plug-in, for various cell cultures utilized for any particular experiment. Flow rates between and among the tissues are established under control of the microprocessor 102 and various settings for the system which are inputted through the keyboard 108 for the species or physiological conditions being modeled. System software (Fig. 18) operating in conjunction with the microprocessor 102 allows for the real time determination of physiologically based flow rates to and from the various compartments. The media circulating through the various compartments consist of standard tissue culture media with horse serum and various surfactants. The media allows the different cell cultures to not only survive but also communicate with each other. Cell cultures may be added by means of convenient snap-in units. Biological containment and sterile conditions are maintained as needed, depending on the particular test system. Test materials may be administered through an injection port. In-line monitors of cell viability, such as albumin production or drug (metabolite) concentration, can provide real time indication

of a desired physiological effect or cell specific drug metabolism profile. These variables are designed to be determined by means of a flow-through multi-channel spectrophotometer. A display may be provided to indicate thereon results pertaining to the various compartments.

The system of Fig. 16 (described in more specificity further on) may be used as a single stand- alone unit or as a bank of four to five systems as desired, for example, to simulate different population segments. This may be beneficial in order to test for the genetic variability of humans with respect to xenobiotic metabolism. In addition, units of the present invention may be used for representing various well- defined segments of the population with known differences with respect to drug metabolism such as neonates or the elderly.

The system of the present invention can be used to determine the biological and toxicological effects of chemicals and pharmaceuticals in human populations. The system of the present invention does not have the attendant uncertainties concerning differences in metabolic profiles between the testing species (such as rats) and humans, because biotransformation reactions are model after human metabolism. The system can be adjusted as needed: to model the metabolism of other organisms, for example, the horse; to model host/parasite relation, for example, malarial parasites and cell cultures from host organisms; or to model an ecosystem, for example, screening pesticides against insects (whole or cell cultures) , plants (whole or cell cultures) and mammalian cell cultures. The system of the present invention is also superior to existing in-vitro systems because of its capacity for intra-cellular communication.

The present invention can provide a superior method of screening early stage drugs for efficacy and toxicity simultaneously. With such information, research efforts can be focused on those pharmaceuticals most likely to be useful

SUBSTI

and to comply with regulatory requirements. Unlike current in-vivo tests, which use live animals and require extrapolation to human results, the system of the present invention is capable of using human cells, therefore 5 eliminating the need to extrapolate data. With the ability to gain directly related information on the effects of considered drugs on humans through the use of the system of the present invention, the dependence on early stage testing in live animals will be greatly lessened. 10 The ability of the system of the present invention to allow for the simultaneous exposure of normal lymphocytes and hepatoma cells to an antineoplastic pharmaceutical analog is described in Example 1. A physiologically-based system of the present invention has been constructed that models the 15 kinetics of the occupationally significant chemical naphthalene in a mouse or a rat, as described in Example 2.

In the system of Fig. 16, a media reservoir 50 holds a test chemical for being supplied via conduit 52 to a pump 54. The pump 54 conducts the test chemical to a first cell 20 culturing compartment 56 holding a first cell culture, for example, mouse pulmonary LL/2 cells obtained from the American Type Culture Collection (Rockville, Md.) . The effluent carrying the test chemical interacts with the LL/2 cells and continues via the conduit 58 to the second cell compartment 25 60, holding a second cell culture, for example, H415-E, urged along by the further pump 62 and passing by needle valve 64 and flow meter 66, as shown. The circulatory path to the compartment 56 is completed by the conduit 68 in which a pump 70 is connected in series. 0 The effluent in media reservoir 50 may be pressurized to a predetermined pressure exerted by gas supplied to the media reservoir 50 from a gas reservoir 78 via the flow meter 80.

Biological, toxological and other effects induced by 5 the test chemical on the cell cultures in the compartments 56

and 60 can be monitored or discerned by means of in-line connected spectrophotometer 82 and 84 to monitor the effluent in the conduits 68 and 58, respectively. in other words, the effects of the test chemical on the culture cells in the compartment 60 can be ascertained with the spectrophotometer 82 and those in the compartment 56 by means of the spectrophotometer 84.

A significant feature of the present invention resides in that the circulatory system in Fig. 16 can be precisely monitored and regulated to simulate actual physiological conditions, in large part due to the use of a central computer system 100. The computer system 100 may comprise a microprocessor 102 having an input/output interface 104 and internal register or cache memory 106. in a typical setup, the microprocessor 102 interfaces to: a keyboard 108 through which operator instructions and test definitions and other information may be entered to the microprocessor 102; a non-volatile storage memory 110, which may comprise a CD writable memory, a magnetic tape memory, or the like; a general purpose memory 112; and look-up tables 114. The look-up tables 114 may physically comprise a portion of the general purpose memory 112 which has been set aside for the storage therein of a set of mass balanced equations applicable to various substances to be modeled in the system of the present invention, in essence, the mass balance equations represent physiologically based pharmacokinetic models for various biological/chemical substances and systems.

The microprocessor 102 may further be interfaced to a display 116 and to a printer/plotter recorder 118 which may provide hard copy of various test parameters and results in printed or graphical format.

In accordance with the present invention, the memories 106 and/or 112 contain a system program in the form of a plurality of program instructions and special data for

SUBSTITUTE S

automatically controlling virtually every function in the system of the present invention.

For example, the temperature of the media in the reservoir 50 is regulated by the microprocessor 102 outputting computer generated commands, via its input/output line 120, to turn the heater coil 122 on and off, responsive to temperature measurements taken with the temperature probe 124 and communicated to the microprocessor 102 via line 126.

Fluid flow monitoring lines 128a, 128b and 128c provide inputs to the microprocessor 102 from flow meters 66, 80 and 76, respectively, enabling precise control over the fluid flow between the various compartments. The fluid flow can be adjusted by program commands transmitted to the pumps 70, 54 and 62 via control lines 130a, 130b and 130c, respectively. For example, the flow rates may be set to 9.5 mL/min in conduit 58, 2.5 mL/min through the flowmeter 66, 7 mL/min through the flowmeter 76, and 2.5 mL/min in the conduit 68.

As already noted, biological and toxological reactions/changes in the compartments 60 and 56 manifest themselves in the conduits 68 and 58, where they are detected/monitored with the spectrophotometer 82 and 84. Software commands to the spectrophotometer 82 and 84 to output discrete or ranges of wavelengths and outputs of the spectrophotometer representing test results are communicated via microprocessor input/output lines 132a and 132b. It is presently contemplated that the spectrophotometer will be control led to output electromagnetic radiation in the range from about 260 to 700 nanometers, as a single, multiple, or a sequentially outputted range of wavelengths.

The vent 134 extending from the media reservoir 50 conducts gaseous by-products to a transducer or biological sensor 136 which then provides a suitable electrical output to the microprocessor 102, enabling further analysis/control over

the chemical/biological activity in the system of the present invention.

Main software blocks and several subroutines associated with the aforementioned control program for the microprocessor 102 are depicted in Fig. 18. Thus, the computer software may initially enter decision block 140 to determine whether the system is configured to proceed in an automatic mode, applicable when the cell compartments 60 and 56 are constructed as electrically "intelligent\', compartments as described further on, or in accordance with a semiautomatic mode, in the semiautomatic mode, the program proceeds via blocks 142, 144, 146 and 148 to receive operational instructions that are entered via the keyboard 108 to define the species which the system is about to model (block 142) , the cell type in each compartment (block 144) , the number and volume of cellular compartments (block 146) , and the temperature and / or wavelength or range of wave lengths at which the experiment is to be carried out.

In step 150, one or more balance equations needed to properly model the experiment being run are fetched from the look-up table memory 114, based on the species model, cell type, number and volume of cellular compartments, etc. Thereafter the program proceeds to software block 152 to calculate a physiological based pharmacokinetic model for the chemical being tested.

Software instructions associated with block 154 cause the input/output interface 104 of microprocessor 102 to output physiologically based pump settings determining the fluid flow rates for various cell compartments. Thereafter, upon turning the heater 122 "ON" in step 156, the software proceeds to decision block 158 to monitor the output of the temperature probe 124 to ensure that fluid flow shall begin only after the effluent in media reservoir 50 has reached the correct temperature.

SUBSTITUTE SHEET

Once the proper temperature has been established, the program proceeds to block 160 to actuate the valves 64 and 74 (which valves also have control lines (not shown) connected to them) , thus initiating circulation in the compartments 56 and 60. Biological and toxological effects will slowly begin manifesting themselves. To detect these effects the program proceeds to block 162 to command the spectrophotometer 82 and 84 to output the previously defined wavelengths, enabling monitoring the biological activity of the system being modeled.

After the foregoing preparatory steps have been performed, the program proceeds to the block 164 where general housekeeping tasks, including test data gathering and storing and, optionally, displaying intermediate as well as final data reflecting on the progress of the experiment, are attend to. Periodically, for example responsive to hardware or software generated interrupts, the program proceeds to the decision block 176 to determine whether an end-of-test flag (to be described) has been set. if so, the program proceeds to its defined "test end" at block 177. Otherwise, the program sequentially proceed through the subroutines 166, 168, 170, 172 and 174.

These subroutines are responsible for providing various display options (block 166) , printing of hard copy test results in alphanumerical and/or graphical formats (block 168) , determining whether the experiment has been concluded (block 170) , temperature monitoring (block 172) , and generating alarm conditions (block 174) , as more fully described below by reference Figs. 19a-19e. Fig. 19a depicts the temperature monitoring subroutine which includes the software block 180 responsible for reading a temperature measurement off the probe 124, comparing the measured temperature to an internally stored desired temperature (block 182) , and determining whether the appropriate temperature has been reached (block 184) . if the

temperature is above the desired value, the heater 122 is turned off in block 186. Otherwise, the heater 122 is either turned or kept turned oh in block 188. The software thereafter proceeds, via return block 190, to the main software code.

In the display subrou ^t ine (Fig. 19b) , the program enters the block 192 wherein an internally kept display request table is consulted (block 192) to check for predetermined or operator requested quantities/ parameters/test results that are to be displayed on the display 116. Based on the information selection criteria read in block 192, the program selects (block 194) the appropriate quantities and parameters to be displayed and organizes the same into a form that is suitable for being transmitted to the display 116. After data transmission to the display 116 in block 196, the program returns to the main code via the subroutine return block 198.

Fig. 19c generally follows Fig. 19b, the functions in the blocks 200, 202, 204 and 206 corresponding, respectively, to the functions of blocks 192, 194, 196, 198 in Fig. 19b, enabling producing a hard copy report on the recorder 118.

The end-of-test subroutine (Fig. 19d) includes a first block 208 for testing, if applicable, whether the program has already been running for a predetermined time period. In block 210, the end-of-test condition is determined on the basis of having achieved certain test results, e.g. quantity of cancerous cells, toxicity level, etc. In block 211, it is determined whether an end of test has been requested from the keyboard 108. If the result in any of the blocks 208, 210, or 211 is affirmative, an ^Λ "end of test" flag is set at block 214. In either case, the program proceeds to the main code via the return block 216.

The subroutine in Fig. I9e is responsible for outputting, if desired, certain visual or audible alarms in

SUBSTITUTE SHEET

the event that the program determines that the test experiment appears not to proceed properly. For example, it might be known that the test results profile for a given test should lie within a given range. For a given test, therefore, the program would first proceed to the block 218 to retrieve from memory previously stored alarm conditions. Thereafter, the program proceeds to block 220 at which point dynamically obtained test progress results are compared to the retrieval alarm conditions. In the decisional block 222 it is determined whether the test results are within expected ranges. If not, an alarm is set to actuate visual or audible indications (not shown) . In either case, the program proceeds via the return block 226 to the main program. The system of the present invention is accordingly designed to allow cell forms, of humans or other species, to be exposed to a test chemical in a manner consistent with in vivo exposure. In order to perform this function, the system incorporates several novel features. First, the compartments in which the biological cells reside are constructed to be mathematically equivalent to the volume of distribution (physical size) of the organ or tissue being modeled. For example, the compartment for the hepatocyte culture would represent the volume of distribution for the liver. Second, the flow rates between the various compartments are biologically based to correctly represent the flow rates between and among the corresponding biological organs, tissue, etc. In this way, cells residing in the compartments are exposed to concentration of test materials in a manner consistent with human or animal exposure to the test material. Additionally, metabolites of one cell type, which may be affected by the presence of the test material, or metabolites of the test material, can be exchanged among the biological cells in the compartments.

The system of Fig. 16 constitutes a microprocessor controlled instrument that performs the functions of exposing the biological cells of different tissue origin to test material at a rate modeling a selected species. Species modeling and flow control are performed by the microprocessor which serves as the overall controller so that it carefully controls flow rates, temperature, and other conditions within the system to mimic, as closely as possible, conditions within the human or other species being modeled. The invention permits with appropriate inputs to model, among other things, the numbers of compartments, the volume of compartments, cell type, etc. The microprocessor is able to configure the correct model for a given problem on the basis of a database of physiological organ flow rates obtained from general reference information for each species.

Since the system of the present invention is interactive, in the sense that the computer not only senses but also controls the conditions within the test, corrections can be dynamically instituted into the system and appropriately noted and documented for apprising researchers of the dynamics of the test being run.

Data gathering by the computer consists of the collection of data required for continuous in-Line monitoring of test chemical eflux from each compartment. The spectrophotometer, preferably of the flow-through type, are disposed in-line with the outflow from each compartment, to thus detect, analyze and provide quantitative data regarding the test chemical eflux from each compartment.

The microprocessor may also serve to compute a physiologically-based pharmacokinetic (PBPK) model for a particular test chemical. These calculations may serve as the basis for setting the flow rates among compartments and excretions rates for the test chemical from the system. However, they may also serve as a theoretical estimate for the test chemical.

SUBSTITUTE SHEET

At the conclusion of the experiment, predictions concerning the concentrations of test chemicals and metabolites made by the PBPK determination can be compared to the spectrophotometer data. Hard copy output compares the PBPK model with experimental results. PBPK determinations are calculated as described in Appendix B.

For clarity. Fig. 16 has been drawn to show only a pair of cell compartments 60 arid 56. However, in accordance with the broader concept of the invention, Fig. 17 illustrates a system comprising a complex of conduits interconnecting a larger number of cell compartments, namely cell compartments 250-268. Each of the compartments 250-268 constitutes, in accordance with the embodiment of Fig. 17, an "intelligent" modularly constructed device that can be plug-in fitted into predesignated slots in the housing 10 shown in Fig. la.

Uniformly configured pneumatic and electrical connections may be provided for the compartments 250-268 to enable automatic engagement of electrical and pneumatic connections (not shown) provided in the housing 10. For clarity, the pneumatic connections 259a, 259b and 259c have been drawn only for compartment 258, for which the electrical connection 259d to the electrical bus 270 is also shown.

Some of these compartments can be designed as no more than a piece of tubing to provide flow through to the next compartment, for increased flexibility.

An electronic circuit or module (see module 250\' for compartment 250, 252\' for the compartment 252, etc.) has been designed to hold characterizing information, readable by the microprocessor 102 via the data bus 270. The electronic module 250\', etc. define for the given compartment its cell type, the species to be modeled, etc. The information may be stored in an on-board, non-volatile memory, in a bank of switches, or in any other form used for data storage. The modules 250\', 252\', etc. may optionally comprise intelligence as by including an onboard microprocessor (not shown) and

circuitry for communicating with the system microprocessor 102 via serial data transmission, e.g. an RS 232 serial data bus.

Consequently,\' via the mentioned electrical interfaces and circuits, upon plugging-in of any of the compartments into one of the slots in the housing 10, the microprocessor 102 is enabled to automatically ascertain many aspects pertaining to the test to be run.

The network of fluid carrying conduits including the conduits 272, 274, 276, etc. in Fig. 17 provides direct and indirect fluid connection paths between the various compartments 258-268, each conduit branch including a respective valve (see, for example, the valves 280, 282, 284, etc.) .

Under control of microprocessor 102 and, optionally, based on- inputs received from the operator via the keyboard 108, the system of this invention can be configured to provide a very large number of different test setups. Thus, for example, the media reservoir 150 in Fig. 17 and the various control valves 280, 282, etc. can be configured so that a test chemical injected into the first compartment 250 flows through a circulation path including only the compartments 250, 252, 254, 256 and 258.

In another test setup, the valves 280, 282, etc. may be so set up that the fluid circulation path is through the compartments 250, 252, 266 and 268. It is also possible to establish two parallel and independent circulation paths, which may prove to be helpful for simultaneously analyzing the efficacy of a given test chemical on different cell cultures. For example, one circulation path may involve the compartments 250, 252, 254, 256 and 258 and another compartments 268, 266, 264, 262 and 260. More conduits than shown in Fig. 17 may be provided.

Also, as shown by reference numeral 290, each or several of the control valves 280, 282, etc. may be configured

SUBSTITUTE SHEET

as pairs of anti-parallel connected one-way valves, to provide control over the direction of fluid flow.

In accordance ^\' with yet a further feature of the present invention, instead of providing a dedicated spectrophotometer for each compartment as shown in Fig. 16, a test sample withdrawing conduit network 292 might be set up for withdrawing samples of the effluent from desired compartments and supplying the same to a master spectrophotometer 294 providing a single output 296 to the microprocessor 102. In other words, by appropriate software, the microprocessor 102 may control a network of valves 296 to sequentially select for testing the effluent from the different compartments 250-268. As an alternative, the conduit network 292 may be constructed as a waveguide network for directing electromagnetic radiation to the different locations in the fluid conduit system of Fig. 17 and for receiving a resultant spectra at the spectrophotometer 294.

As a further option, the system may include an additional fluid reservoir 151\' for holding a flushing solution for being pumped through the entire system to remove any residues from prior tests.

The standardized compartments will enable providing to test laboratories pre-prepared cell culture compartments with prepared and pre-identified species, cell type, etc.

Thus, researchers will need do no more than simply order ready made, disposable compartments which they may simply plug into the system of Figs. 16 and 17, greatly facilitating and expediting research. The software of the computer system of Fig. 16 is designed to model different physiological flow rates, and to reflect different pathological conditions, including a person at rest, during exercise, or at sleep.

The versatility of the microprocessor 102 further extends to the display function which might provide on the

display 116 continuously updated readings of spectrophotometric results in the form of a number between 1 and 2 or as an anti-log ^\' number ranging between 0 to 100 or as alphanumeric text providing quantitative and qualitative information.

The microprocessor 102 is also quite easily adaptable to include a program to provide the researcher with interactive control via keyboard 108 enabling, for example, directing the computer to specifically check on the conditions of any of the culture compartments at any given time.

The system of the present invention may also be deployed for the purposes of taking measurements for determining whether healthy cells are in the midst of being transformed to cancerous cells and data relevant thereto. The diversity and flexibility of the present invention also permits carrying out tests using multiple culture compartments of parallel cell types to test the effects of an agent on different cell types simultaneously. The compartments may be filled with cell lines of different species, e.g., mouse, rat, human, of different organs, e.g., lung, heart, and kidneys, or of different populations, e. g. , neonates, middle-aged, and elderly. Many other combinations of parallel cell cultures are feasible and useful.

A further option provided by the present invention is the ability to recall previously stored test results for similar experiments by recalling information from the CD/tape memory 110. Thus, the memory 110 may be preprogrammed to hold historical data taken from published information, data gathered from previously run tests conducted with the system of the present invention or data derived from theoretical calculations.

The provision of the CD/tape memory also affords the possibility of the system being used as an information researching tool to obtain, for example, research data pertaining to a particular test chemical, or to a particular

SUBSTITUTE SHEET

culture line, etc. based on selection* information inputted into the microprocessor 102 via the keyboard 108. By including or developing ^\' a large library of information in the memory 110, researchers will be able to more intelligently 5 configure and plan test runs.

Fig. lc is a schematic of another embodiment applying the principles of this invention, and shows multiple cell compartments including compartments 21, 23 and 25 connected to a reservoir 27. A single pump 29 maintains a 10 constant flow rate through the entire device. A shunt 3, is incorporated to provide a finer control over the effective flow rate to the compartments 21, 23 and 25 and to decrease foaming. Additional pumps and shunts (not shown) could be incorporated to maintain differential flows among the 15 compartments 21, 23 and 25.

In Fig. lc, the circulation path includes conduit 33 feeding effluent to compartment 21, intercompartment conduits 35 and 37, conduit 39 from compartment 25 to reservoir 27, and conduit 41 through which effluent is drawn from the reservoir 20 27 by the pump 29.

Although not shown, it is to be understood that the system of Fig. lc may incorporate the flow meters, valves, computer system, spectrophotometers and other elements and features which have been described above in relation to Fig. 2516.

The terms and expressions which have been employed are used as terms of description and not of limitation, and there is no intention in the use of such terms and expressions of excluding any equivalents of the features shown and 0 described or portions thereof, it being recognized that various modifications are possible within the scope of the invention.

Although the present invention has been described in relation to particular embodiments thereof, many other 5 variations and modifications and other uses will become

SUBSTITUTE SHEET

apparent to those skilled in the art. It is preferred, therefore, that the present invention b limited not by the specific disclosure her\'ein, but only by ;hs appended claims.

SUBSTITUTE SHEET

REFERENCES

Abraham, M.H., Kamlet, M.J., Taft, R.W., Doherty, R.M., and Weathersby, P.K. (1985)-. Solubility properties in polymers 5 and biological media. 2. The correlation and prediction of the solubilities of non-electrolytes in biological tissues and fluids. J.Med.Chem. 28, 865-870.

Anderson, M.E., Clewell III, H.J. Gargas, M.L., Smith, F.A. , 10 and Reitz, R.H. (1987). Physiologically based pharmacokinetiσs and the risk assessment process for methylene chloride. Toxicol.AppI.Pharmacol. 87, 185-205.

Bakke, J.E., Davison, K.L., and Larsen, G.L. (1990). Evidence 15 for the absence of cysteine S-conjugate N-acetyl transferase activity in the metabolism of propachlor, naphthalene, and dichlobanil in calves. Xenobiotica 20, 801-807.

Bakke, J., Struble, C., Gustafsson, J.A., and Gustaffson, B. 20 (1985) . Catabolism of premercapturic acid pathway metabolites of naphthalene to naphthalene to naphthols and methylthio-containing metabolites in rats Proc.Natl.Acad.Sci. USA 82, 668-671.

25 Brouwer, K.R. and McNamara, P.J. (1989) . Influence of pregnancy on the pharmacokinetic disposition of two aromatic retinoids (etrinate and acitretin) in the rat. II Single and multiple oral dosing studies. Drug Metab. Dispos. 17, 625-655.

30

Buckpitt, A.R., Bahnson, L.S., and Franklin, R.B. (1984). Hepatic and pulmonary microsomal metabolism of naphthalene to glutathione adducts: factors affecting the relative rates of conjugate formation. J.Pharmacol.Exp.Ther. 231, 291-300.

35

Buckpitt, A.R. , Castagnoli, Jr., N. , Nelson, S.D., Jones, A.D., and Bahnson, L. S. (1987). Stereoselectivity of naphthalene epoxidation by mouse, rat, and hamster pulmonary, hepatic and renal microsomal enzymes. Drug Metab.Dispos. 15, 40491-498.

Buckpitt, A.R. and Franklin, R.B. (1989) . Relationship of naphtha lene and 2-methyl-naphthalene metabolism to pulmonary bronchiolar epithelial cell necrosis. Pharmac.Ther. 41, 45393-410.

Buckpitt, A.R. and Warren, D.L. (1983) . Evidence for the formation, export, and covalent binding of reactive metabolites in extrahepatic tissues in vivo. 50 J.Pharmacol.Ex .Ther. 225, 8-16.

SUBSTITUTE SHEET

Buonarati, N. , Jones, A.D., and Buckpitt, A. (1990). In vivo metabolism of isomeric naphthalene oxide glutathione conjugates. Drug Metab.Dispos. 18, 183-189. Buonarati, M. , Morin, D., Plopper, C. and Buckpitt, A. (1989). Glutathione depletion and cytotoxicity by naphthalene 1,2-oxide in isolated hepatocytes. Chem.-Biol.Interact. 71, 147-165. Calder III, W.A. (1984) . Size, Function and Life History. Harvard University Press, Cambridge, Massachusetts.

Casini, A.F., Ferrali, M. , Pompella, A., Maellaro, E. , and Comporti, M. (1986) . Lipid peroxidation and cellular damage in extrahepatic tissues of bromobenzene-intoxicated mice. Am. .Pathol. 123, 520-531.

Cha, Y.N. and Bueding, E. (1979). Effect of 2(3) tert-butyl-4- hydroxyanisole administration on the activities of several hepatic microsomal and cytosoplasmic enzymes in mice. Biochem.Pharmacol. 28, 1917-1921.

Chemical Economics Handbook. (1981). Section 300.7600 A-R, SRI International, Menlo Park, California. As referenced in Buckpitt and Franklin (1989) .

Chem, W.H. and Dauterman, W.C. (1983) . Studies on the metabolism and excretion of 1-naphthol-R-D-glucuronide, and 1-naphthol-B-D- glucoside in the mouse. Toxicol.AppI.Pharmacol. 67, 303-309.

Conolly, R.B. and Andersen, M.E. (1991) . Biologically based pharmacodynamic models: tools for toxicological research and risk assessment. Annu.Rev.Pharmacol.Toxicol. 31, 503-23.

DiSimplico, P., Jensson, H. and Mannervik, B. (1989). Effects of inducers of drug metabolism on basic hepatic forms of mouse glutathione transferase. Biochem.J. 263, 679-685. D\'Souza, R.W. , Francis, W.R., and Andersen, M.E. (1988). Physiological model for tissue glutathione depletion and increased resynthesis after ethylene dichloride exposure. J.Pharmacol.Ex .Ther. 245, 563-568. Gaines, T.B. (1969) . Acute toxicity of pesticides. Toxicol.AppI.Pharmacol. 14, 515-534.

Garle, N.J. and Fry, J.R. (1989) . Detection of reactive metabolites in vitro. Toxicology 54, 101-110.

Gearhart, J.M. , Jepson, G.W. , Clewell III, H.J., Andersen, M.E., and Conolly, R.B. (1990). Physiologically based pharmacokinetic model for the inhibition of

SUBSTΓRΠΈ SHEET

acetylcholinesterase by diisopropylfluorophosphate. Toxicol.AppI. harmacol. 106, 295-310.

Griffith, O.W. and Meister, A. (1979) . Glutathione: 5 interorgan translocation, turnover and metabolism. Proc.Natl.Acad.Sci. USA 76, 5606-5610.

Hansch, C. and Leo, A. (1979) . Substituent Constants for Correlation Analysis in Chemistry and Biology. J. Wiley and 10 Sons, New York, New York.

Hayakawa, T., Leilahieu, R.A., and Udenfriend, S. (1974). Studies on glutathione-S-arene oxide transferase—a sensitive assay and partial purification of the enzyme from sheep liver. 15 Arch. Biochem.Biophys. 162, 223-230.

Hesse, S. and Mezger, M. (1979) . Involvement of phenolic metabolites in the irreversible protein-binding of aromatic hydrocarbons: reactive metabolites of (14C)-naphthalene and 20 ( 4C3 1- naphthol formed by rat liver microsomes. Mol.Pharmacol. 16, 667- 675.

Kanekal, S. , Plopper, C. , Morin, D., and Buckpitt, A. (1990). Metabolic activation and bronchiolar Clara cell necrosis from 25 naphthalene in the isolated perfused mouse lung. J. harmacol.Exp.Ther. 252, 428-437.

Kaplan, N.M., Brewer, N.R. and Blair, W.H. (1983). Chapter 11, Physiology. In The House in Biomedical Research, Vol. 30 III, (H.L. Foster, J.D. Small, and J.G. Fox, Eds.) pp. 247-292. Academic Press, New York.

Moron, M.S., DePierre, J.W. , and Mannervik, B. (1979). Levels of glutathione, glutathione reductase and glutathione 35 S-transferase activities in rat lung and liver. Biochim.Biophys.Acta 582, 67-68.

O\'Brien, K.A.F., Smith, L.L., and Cohen, G.M. (1985). Differences in naphthalene-induced toxicity in the mouse and 40 rat. Chem. - Biol.Interact. 55, 109-122.

O\'Brien, K.A.F., Suverkropp, C. , Kanekal, S., Plopper, C.G., and Buckpitt, A.R. (1989) . Tolerance to multiple doses of the pulmonary toxicant, naphthalene. Toxicol.Ap I.Pharmacol. 99, 45487- 500.

Okine, L.K. and Gram, T.E. (1986). In vitro studies on the metabolism and covalent binding of (14C)1, l-dichloroethylene by mouse liver, kidney and lung. Biochem.Pharmacol. 35, 50 2789-2795.

SUBSTITUTE SHEET

Richieri, P.R. and Buckpitt, A.R. (1988) . Gluthathione depletion by naphthalene in isolated hepatocytes and by naphthalene oxide in vivo. Biochem.Pharmacol. 37, 2473-2478. Richieri, P.R. and Buckpitt, A.R. (1987) . Efflux of naphthalene oxide and reactive naphthalene metabolites from isolated hepatocytes. J.Pharmacol.Exp.Ther. 242, 485-492.

Rietjens, I.M.C.M., Dormans, J.A.M.A. , Rombout, P.J.A., and Van Bree, L. (1988) . Qualitative and quantitative changes in cytochrome P-450-dependent xenobiotic metabolism in pulmonary microsomes and isolated Clara cell populations derived from ozone exposed rats. J.Toxicol.Environ.Health 24, 515-531. Shank, R.C., Barrows, L.P. and Buckpitt, A.R. (1980).

Comparative metabolism of hydrazine and naphthalene. Annual Report to Air Force Aerospace Medical Research Laboratory, Report No. AFARMAL-TR80-103. Aerospace Medical Division, Air Force System, WrightPatterson Air Force Base, Ohio.

Shopp, G.M. , White, Jr., K.L. Holsapple, M.P. , Barnes, D.W. , Duke, S.S., Anderson, A.C., Condie, Jr., L.W. Hayes, J.R., and Borzelleca, J.F. (1984). Naphthalene toxicity in CD-I mice: general toxicity and immunotoxicology. Fundam.AppI.Toxicol. 4,406- 419.

Stillwell, W.G., Horning, M.G., Griffin, G.W., and Tsang, W.S. (1982) . Identification of new sulfur-containing metabolites of naphthalene in mouse urine. Drug Metab.Dispos. 10, 624-631.

Suga, T. , Ohata, I., and Akagi, M. (1966). Studies on mercapturic acids: effect of some aromatic compounds on the level of glutathione and the activity of glutathionase in the rat. - Biochem. 59, 209-215.

Summer, K.H., Rozman, K. , Coulston, F., and Greim, H. (1979). Urinary excretion of mercapturic acids in chimpanzees and rats. Toxicol.AppI.Pharmacol. 50, 207-212. van Bladeren, P.J., Sayer, J.M. , Ryan, D.E., Thomas, P.E., Levin, W. , and Jerina, D.M. (1985). Differential stereoselectivity of cytochrome P-450b and P-450c in the formation of naphthalene and anthracene 1,2-oxides. J.Biol.Chem. 260, 10226-10235.

Vargartik, N.B. (Ed.) (1975). Handbook of physical properties of liguids and gases. Second edition, Hemisphere Publishing, Washington, D.C.

Ward, R.C., Travis, CC. , Hetrick, D.M., Andersen, M.E., and Gargas, M.L. (1988) . Pharmacokinetics of tetrachloroethylene. Toxicol.APPI. harmacol. 93, 108-117.

SUBSTITUTE SHEET

Warren, D.L. Brown, Jr., D.L., and Buckpitt, A.R. (1982). Evidence for cytochrome P-450 mediated metabolism in the bronchiolar damage by naphthalene. Chem.Biol.Interact. 40, 287-303.

Yost, G.S., Buckpitt, A.R. , Roth, R.A. , and McLemore, T.L. (1989) . Mechanisms of lung injury by systemically administered chemicals. Toxicol.AppI.Pharmacol. 101, 179-195. Yost, G.S., Kuntz, D.J., and McGill, L.D. (1990). Organ selective switching of 3-methylindole toxicity by glutathione depletion. Toxico1.App1.Pharmaco1. 103, 40-51.

SUBSTITUTE SHEET

Appendix A

Nomenclature Decoding abbreviations

First letter(s) : parameter type

C: Concentration

K: Reaction rate KM: Michaelis constant

P: Partition coefficient

Q: Flow rate

V: Maximum reaction rate

W: Weight

Next letter(s) : anatomical location

If these letters do not appear, the parameter is not tissue specific.

Next group of letters: Compound(s)

C: Naphthalene oxide-glutathione conjugate

D, DI: Naphthalene l, 2 dihydrodiol G: Glutathione

GS: Glutatione synthetase

N: Naphthalene

NO: 1R, 2S, 2-R-naphthalene oxide

NOH: 1-Naphthol P: Protein

RS: 1R, 2S-naphthalene oxide

SR: IS, 2-R-naphthalene oxide

If two compound abbreviations are part of one parameter name, it involves reaction of the first compound to form the second.

Final letters:

SUBSTITUTE SHEET

For rates:

D: Degradation S: Synthesis

For concentrations:

SS: Steady state PT: A prior time, T-TD

Examples: VLURSD = maximum velocity (v) for reaction in the lung (LU) of II?, 2S-napthalene oxide (RS) to dihydrodiol (D) . KLIGGS: rate (K) in the liver (LI) of glutathione (G) at steady state (SS) .

Other nomenclature: AOR: Amount of oral dose which is bioavailable

KUP: Rate of oral uptake

T: Time

TD: Time Delay for glutathione synthetase synthesis

UNIT: Unit conversions for scaling up reaction rates from mg protein to whole organ levels

KIP: Rate of ip uptake

Appendix B

Mass Balance Equations for Naphthalene and Naphthalene oxide

d CLUN/dt = QC*(CVN-CAN)/WLU - (VLUNRS*CLUN/ (CLUN+KMLUNRS)

+VLUNSR*CLUN/ (CLUN+KMLUNSR) ) *UNIT1 d CLIN/dt = QLI*(CPVN-CVLIN)/WLI- VLINRS*CLIN/ (CLIN+KMLINRS)

+VLINSR*CLIN/ (CLIN+KMLINSR) ) *UNIT2 d CKN/dt = QK*(CAN-CVKN)/WK d CFN/dt = QF*(CAN-CVFN)/WF d CON/dt = QO*(CAN-CVON)/WO d CLURS/dt = QC*(CVRS-CARS)/WLU

+VLUNRS*CLUN/ (CLUN+KMLUNRS) *UNIT1 -VLURSD*CLURSD/ (CLURSD+KMRSDI) *UNIT1 -VLUNOC*CLUN*CLUG/ (CLUN+KMNOC) /

(CLUG +KMGSS) *UNIT5 -KNOH*CLURS-KP*CLURS/WLU) *UNIT3 d CLIRS/dt = QLI*(CARS-CVLIRS)/WLI

+VLINRS*CLIN/(CLIN+KMLINRS) *UNIT2 -VLIRSD*CLIRSD/ (CLIRSD+KMRSDI) *UNIT2 -VLUNOC*CLIN*CLIG/ (CLIN+KMNOC) /

(CLIG+KMGSS) *UNIT6 -KNOH*CLIRS-(KP*CLIRS/WLI) *UNIT4 d CKRS/dt = QK*(CARS-CVKRS) /WK-KNOH*CKRS d CFRS/dt = QK*(CARS-CVFRS) /WK-KNOH*CFRS d CORS/dt = QK*(CARS-CVORS)/WK-KNOH*CORS

The mass balance equations for IS, 2-R-naphthalene oxide are the same as for 1R, 2S-naphthalene oxide, with "SR" substituted for "RS."

Partitioning

CAN = CLUN/PLUB CVLIN = CLIN/PLIB

The equations for lung, kidney, fat and other tissues are the same as for the liver, with LU, K, F and O replacing LI.

SUBSTITUTE SHEET

Mixing

CVN = ((2VLIN*QLI+CCVKN*QK+CVFN*QF+CVON*QO/QC For CVRS and CVSR, replace N in the above equation with RS or SR.

CPVN = CAN+KIP/QLI (ip dose)

CPVN = CAN+KUP*AOR*EXP(-KUP*T)/QLI (oral dose)

Accounting for non-enzymatic rearrangement in the blood.

CARS ₂ =CARS _! ( l+KNOH*WBL*2 / 3 / QC)

CVRS ₂ =CVRS ₁ (l+KNOH*WBL/3/QC)

For CASR and CVSR replace RS with SR.

GSH mass balance and GSH synthetase synthesis d CLUG/dt = KLUG/WLU-

KLUGS*CLUG+CLUSR/(CLUSR+KMNOC) )*UNITS d KLUG/dt = (KLUGSS+KMGSS)/(CLIGPT+KMGSS)-KGSD*KLUG

For liver, replace LU with LI and UNIT5 with UNIT6.

Conversion to dihydrodiol, GSH conjugates, and total metabolites d DlOL/dt = (VLURSD*CLURSD/(CLURSD+KMRSDI) +VLUSRD*CLUSRD/(CLUSRD+KMSRDI) ) *UNIT1*WLU

+(VLIRSD*CLIRSD/ (CLIRSD+KMRSDI) +VLISRD*CLISRD/(CLISRD+KMSRDI) )*UNIT2*WLI d CONJUGATE2/dt = VLUNOC*(CLUG/ (CLUG+KMGC) ) *(CLURS/ (CLURS+KMNOC) *UNIT5*WLU

+VLINOC*(CLIG/ (CLIG+KMGC) )*(CLIRS/ (CLIRSIKMNOC* *UNIT6*WLI

For conjugates 1 and 3, replace RS with SR in the above equation. d N/dt = (VLUNRS*CLUN/(CLUN+KMLUNRS)

+VLUNSR*CLUN*(CLUN+KMLUNSR) ) *UNIT1 +(VLINRS*CLIN/ (CLIN+KMLINRS) =VLINSR*CLIN/(CLIN+KMLINSR))*UNIT2

Covalent binding d (Binding of RS in LU)/dt=KP*CLURS

For liver, replace LU with LI, for SR, replace RS with SR.

SUBSTITUT