Incorporation of the Genetic Control of Alcohol Dehydrogenase into a Physiologically Based Pharmacokinetic Model for Ethanol in Humans

Lester G. Sultatos*,1, Gina M. Pastino{dagger}, Clint A. Rosenfeld* and Edward J. Flynn*

* Department of Pharmacology and Physiology, New Jersey Medical School, University of Medicine and Dentistry of New Jersey, Newark, New Jersey 07103, and {dagger} Schering-Plough Research Institute, Lafayette, New Jersey 07843

1 To whom correspondence should be addressed at Department of Pharmacology and Physiology, New Jersey Medical School, UMDNJ, 185 South Orange Avenue, Newark, NJ 07103. Fax: (973) 972-4554. E-mail: sultatle{at}umdnj.edu.

Received September 10, 2003; accepted November 15, 2003

ABSTRACT

The assessment of the variability of human responses to foreign chemicals is an important step in characterizing the public health risks posed by nontherapeutic hazardous chemicals and the risk of encountering adverse reactions with drugs. Of the many sources of interindividual variability in chemical response identified to date, hereditary factors are some of the least understood. Physiologically based pharmacokinetic modeling linked with Monte Carlo sampling has been shown to be a useful tool for the quantification of interindividual variability in chemical disposition and/or response when applied to biological processes that displayed single genetic polymorphisms. The present study has extended this approach by modeling the complex hereditary control of alcohol dehydrogenase, which includes polygenic control and polymorphisms at two allelic sites, and by assessing the functional significance of this hereditary control on ethanol disposition. The physiologically based pharmacokinetic model for ethanol indicated that peak blood ethanol levels and time-to-peak blood ethanol levels were marginally affected by alcohol dehydrogenase genotypes, with simulated subjects possessing the B2 subunit having slightly lower peak blood ethanol levels and shorter times-to-peak blood levels compared to subjects without the B2 subunit. In contrast, the area under the curve (AUC) of the ethanol blood decay curve was very sensitive to alcohol dehydrogenase genotype, with AUCs from any genotype including the ADH1B2 allele considerably smaller than AUCs from any genotype without the ADH1B2 allele. Furthermore, the AUCs in the ADH1C1/C1 genotype were moderately lower than the AUCs from the corresponding ADH1C2/C2 genotype. Moreover, these simulations demonstrated that interindividual variability of ethanol disposition is affected by alcohol dehydrogenase and that the degree of this variability was a function of the ethanol dose.

Key Words: alcohol dehydrogenase; physiologically based pharmacokinetic modeling; ethanol; risk assessment.

Interindividual variability in therapeutic and toxic responses to xenobiotics results from a variety of pharmacodynamic and pharmacokinetic factors (Clewell and Andersen, 1996Go; Nebert, 2000Go). The assessment of the variability of human responses to foreign chemicals is an important step in characterizing the public health risks posed by nontherapeutic hazardous chemicals and the risk of encountering adverse reactions with drugs. Of the many sources of interindividual variability in chemical response identified to date, hereditary factors are some of the least understood. Much uncertainty is often associated with the potential functional significance of both polygenic control and genetic polymorphisms of proteins that are involved in mediating the actions and/or disposition of foreign chemicals. This in turn leads to uncertainty when attempting to quantify risk associated with certain chemical exposures. New approaches are needed to investigate the potential functional significance of polygenic traits and genetic polymorphisms and to use this information for assessing the consequent interindividual variability in chemical response. Physiologically based pharmacokinetic modeling is ideally suited for the consideration of hereditary factors that affect the pharmacokinetic disposition of specific chemicals and for the estimation of the interindividual variability stemming from these genetic factors (Clewell and Andersen, 1996Go; Gentry et al., 2002Go). Our study developed a physiologically based pharmacokinetic model for ethanol that includes the hereditary control of the family of enzymes known as alcohol dehydrogenase (EC 1.1.1.1).

Alcohol dehydrogenases catalyze the reversible oxidation of ethanol to acetaldehyde, with corresponding reduction of NAD+ to NADH (Li et al., 2001Go; Fig. 1). Alcohol dehydrogenase in humans is a dimer and exists as multiple forms, resulting from its well-characterized polygenic control and genetic polymorphisms (Li et al., 2001Go). The nomenclature we used for the various forms of alcohol dehydrogenase and the genes that encode these forms were from Duester et al. (1999)Go. This enzyme in humans is encoded by at least seven separate genes, and multiplicity of the enzyme exists at three different levels (Li et al., 2001Go). Firstly, separate classes of alcohol dehydrogenase are derived from distant gene duplication at early vertebrate times (Jornvall et al., 2000Go). Secondly, more recent gene duplications have led to isozyme multiplicity within certain classes (Jornvall et al., 2000Go). Thirdly, allelic variants exist within several different isozyme forms (Jornvall et al., 2000Go). Those hepatic forms of alcohol dehydrogenase that are primarily involved in the metabolism of ethanol are referred to as class 1. The subunits (monomers) that constitute class 1 alcohol dehydrogenases are encoded by the three genes ADH1A, ADH1B, and ADH1C, giving rise to A, B, and C subunits, which can then combine to form homodimers and heterodimers (Duester et al., 1999Go). The ADH1A, ADH1B, and ADH1C subunits share greater than 93% sequence homology (Plapp et al., 1984Go). Allelic variants, or genetic polymorphisms, occur at ADH1B to give ADH1B1, ADH1B2, and ADH1B3 subunits and at ADH1C to give ADH1C1 and ADH1C2 subunits. Thus, hepatic class 1 alcohol dehydrogenase in humans can exist as 21 different possible forms (homodimers plus heterodimers), with the specific number and type of forms within an individual dependent on that individual's genotype. Additionally, class 4 alcohol dehydrogenase is a single homodimer and is found primarily in the stomach (Li et al., 2001Go).



View larger version (5K):
[in this window]
[in a new window]
 
FIG. 1. The Theorell Chance mechanism for metabolism of ethanol by alcohol dehydrogenase. This kinetic scheme is known as the Theorell Chance mechanism (Segel, 1975Go; Theorell and Chance, 1951Go). Interestingly, the equilibrium of this reaction is very much in favor of ethanol and NAD+ (Backlin, 1958Go; Bosron et al., 1993Go). B, ethanol; E, alcohol dehydrogenase; N, NAD+; P, acetaldehyde; and Q, NADH.

 
While there exists a wealth of information concerning the kinetic characteristics and catalytic differences of various homodimers of alcohol dehydrogenase in vitro (Bosron et al., 1993Go), the functional significance of hereditary control of alcohol dehydrogenase in humans is not well understood due to the paucity of data documenting the pharmacokinetic disposition of ethanol in individuals of known genotype. Therefore, the contribution of the polygenic control and the genetic polymorphisms of alcohol dehydrogenase to the interindividual variability in ethanol disposition within a population is unknown. Using ethanol as a test chemical, the present study has extended the use of physiologically based pharmacokinetic modeling to the assessment of the functional significance of polygenic traits and genetic polymorphisms in modulating chemical disposition and, thereby, to interindividual variability in the pharmacokinetic disposition of foreign chemicals.

MATERIALS AND METHODS

General model structure.
The structure of the physiologically based pharmacokinetic model, minus the metabolism by hepatic alcohol dehydrogenases, was adapted from a previously presented model (Pastino et al., 2000Go) and is shown in Figure 2. The values for the physiological parameters used, as well as distribution information used for Monte Carlo sampling, are shown in Table 1. The volume of and blood flows to the rapidly perfused and slowly perfused tissue compartments were calculated from the following equations (Pastino et al., 2000Go):

(1)

(2)

(3)

(4)
where BW is body weight; QBR is the blood flow to brain; QC is cardiac output; QF is the blood flow to fat; QL is the blood flow to liver; QR is the blood flow to rapidly perfused tissues; QS is the blood flow to slowly perfused tissues; VBR is the volume of brain; VF is the volume of fat; VL is the volume of liver; VR is the volume of rapidly perfused tissues; and VS is the volume of slowly perfused tissues.



View larger version (20K):
[in this window]
[in a new window]
 
FIG. 2. Schematic diagram of the PBPK model for ethanol in humans, adapted from Pastino et al. (2000)Go.

 

View this table:
[in this window]
[in a new window]
 
TABLE 1 Physiological Parameters and Their Distributions

 
For Monte Carlo sampling, the upper and lower bounds of all distributions (Table 1) were truncated at three standard deviations to eliminate values that were nonphysiological (Clewell et al., 1999, 2000, 2001Go; Gentry et al., 2002Go). Gastric alcohol dehydrogenase content was kept constant since variability estimates were not available. Since the values for the parameters presented in Table 1 were not identical to those values used by Pastino et al. (2000)Go, the absorption constants Ka1, Ka2, and Ka3 (Fig. 2) were reoptimized with the model presented in Pastino et al. (2000)Go, using the mean values in Table 1, to the data sets presented in Wilkinson et al. (1977)Go. This procedure yielded the following values: Ka1 = 0.21 h-1, Ka2 = 3.17 h-1, and Ka3 = 3.28 h-1. For comparison, Pastino et al. (2000)Go reported 0.66 h-1, 1.50 h-1, and 7.00 h-1 for Ka1, Ka2, and Ka3, respectively, while Derr (1993)Go reported 0.66 h-1, 3.12 h-1, and 25.10 h-1 for those same constants. The rat tissue/blood partition coefficients for ethanol used by Pastino et al. (2000)Go were used in our model since human tissue/blood partition coefficients for ethanol were not available.

Modeling alcohol dehydrogenases.
Alcohol dehydrogenase isoforms catalyze reversible oxidation of ethanol to acetaldehyde, with corresponding reduction of NAD+ to NADH, by a kinetic mechanism known as the Theorell Chance mechanism (Segel, 1975Go; Theorell and Chance, 1951Go; Fig. 1). To incorporate the genetic control of alcohol dehydrogenase into the physiologically based pharmacokinetic model, a metabolism submodel for each possible form of alcohol dehydrogenase (a total of 21 possible forms and, therefore, 21 submodels) was developed based on the differential and algebraic equations describing the Theorell Chance kinetic mechanism (Fig. 1). The equations for each submodel were structured as follows:

(5)

(6)

(7)

(8)

(9)
where ET is total active site concentration for a specific form and BT is total ethanol concentration. The term kE represents an operational switch to "turn on" or "turn off" a specific enzyme form, depending on the genotype modeled. All other nomenclature is described in Figure 1. NAD+ levels (N) were kept constant at 0.5 mM since this is in the normal physiological range (Burnell et al., 1989Go Crow et al., 1982Go).

Solving Equations 5-9 required values for the rate constants descriptive of the Theorell Chance mechanism for each enzyme form (Fig. 1). Although these values have not been published, kinetic equilibrium constants (see Segel [1975] for a detailed discussion of these) such as Vmax (maximum velocity), Km (substrate concentration at half of maximum velocity), and Ki (the product inhibition constant) have been published for all homodimers of class 1 alcohol dehydrogenases (Bosron et al., 1993Go; Yin et al., 1984Go), and the appropriate rate constants could be calculated from these equilibrium constants. It should be noted that alcohol dehydrogenase activity runs in both the forward and reverse directions (Bosron et al., 1993Go) and, therefore, has potentially four substrates and four products (Fig. 1). For example, when considering the forward reaction, ethanol and NAD+ are substrates with the products being acetaldehyde and NADH. Conversely, with the reverse reaction, the substrates are acetaldehyde and NADH and the products are ethanol and NAD+. Each substrate has a Vmax and Km, while each product has a Ki (a product inhibitory constant; Segel, 1975Go). Those rate constants required for the Theorell Chance mechanism were calculated from the published kinetic equilibrium constants using a series of equations, derived in Segel (1975)Go, as follows:

(10)

(11)

(12)

(13)

(14)

(15)

(16)

(17)
where Kib is the inhibitory constant for ethanol; Kin is the inhibitory constant for NAD+; Kip is the inhibitory constant for acetaldehyde; Kiq is the inhibitory constant for NADH; Kmb is the concentration of ethanol that gives one-half Vmaxf; Kmn is the concentration of NAD+ that gives half of the maximum velocity when velocity is measured as a function of NAD+; Kmp is the concentration of acetaldehyde that gives one-half Vmaxr; Kmq is the concentration of NADH that gives half of the maximum velocity when velocity is measured as a function of NADH; Vmaxf is the maximum velocity in the forward direction with ethanol as the substrate; and Vmaxr is the maximum velocity in the reverse direction with acetaldehyde as the substrate, as described in detail by Segel (1975)Go.

Rate constants in Equations 10-17 are identical to those described in Figure 1. The Theorell Chance kinetic mechanism was modeled with rate constants (Fig. 1) rather than equilibrium constants due to the extreme complexity of the velocity equations containing the equilibrium constants (Segel, 1975Go), which rendered them impractical for use in our model. Furthermore, it should be noted that the published Vmax values for isoforms could not be used directly since recovery data for the purified forms were not available. All rate constants (Table 2) were calculated directly from published kinetic equilibrium constants reported by Bosron et al. (1983)Go, and Yin et al. (1984)Go, using the above equations, except for k1 for the AA homodimer, k3 for the B1B1 homodimer, and k2 for both the C1C1 and C2C2 homodimers, which were determined as described below.


View this table:
[in this window]
[in a new window]
 
TABLE 2 Rate Constants for the Metabolism of Ethanol by Isoforms of Hepatic Alcohol Dehydrogenase

 
While in general the activities of the AA, B1B1, C1C1, and C2C2 homodimers have been shown to be consistent with the Theorell Chance kinetic mechanism (Bosron et al., 1983Go; Fig. 1), the ethanol saturation profile revealed substrate inhibition for the AA and B1B1 homodimers and negative cooperativity for the C1C1 and C2C2 homodimers (other homodimers showed no evidence of substrate inhibition or negative cooperativity; Bosron et al., 1983Go). Preliminary modeling with these homodimer submodels indicated that substrate inhibition and negative cooperativity could be mimicked by altering a single rate constant in each submodel as a function of ethanol concentration (data not shown). Therefore, a selected rate constant for each homodimer was fitted (using Tablecurve, SPSS Inc., Chicago, IL) to the homodimer velocity reported by Bosron et al. (1983)Go. For the AA homodimer, the equation describing the relationship between k1 and ethanol concentration was as follows:

(18)
The correlation coefficient was 0.98, while a = 2.96, b = 0.11, c = 0.80, d = 0.001, e = -0.02, f = -5.89e-6, g = 7.42e-5, h = 8.80e-8, and x is the ethanol concentration. This relationship was included in the submodel for the AA homodimer to calculate a value for k1 at any given ethanol concentration.

For the B1B1 homodimer, the equation describing the relationship between k3 and ethanol concentration was as follows:

(19)
The correlation coefficient for the fit of the optimized velocities to the above equation was 0.99, while a = -0.906, b = -0.002, c = 0.025, d = 3.602e-9, and x is the ethanol concentration.

The equation describing the relationship between C1C1 and C2C2 and their k2 values (an inverse of the Hill equation; Segel, 1975Go) was as follows:

(20)
The correlation coefficient for the fit of the C1C1 homodimer optimized velocities to the above equation was 0.98, while a = 0.4008, b = 0.1643, c = 19.2846, d = 0.0807, and x is the ethanol concentration. The correlation coefficient for the fit of the C2C2 homodimer optimized velocities to the above equation was 0.97, while a = 0.4062, b = 0.1737, c = 20.3500, d = 0.08294, and x is the ethanol concentration.

Equations 18–20 were incorporated into the various alcohol dehydrogenase submodels to account for the substrate inhibition and negative cooperativity.

Class 1 alcohol dehydrogenase isoforms can exist as homodimers and heterodimers. The roles of subunits or monomers in homodimer activity have been studies over the past two decades. While early investigations reported that alcohol dehydrogenase monomers were enzymatically inactive (Anderson and Mosbach, 1979Go Briganti et al., 1989Go), Ehrig et al. (1993)Go, using an extremely high ethanol concentration (1.5 M), demonstrated activity with the B1 monomer that exceeded homodimer activity. They noted two important differences between monomers and homodimers: (1) three amino acids of the outer part of the hydrophobic ethanol-binding barrel of a subunit are furnished by the second subunit, leading to a decreased affinity of the monomer for ethanol; (2) an incomplete coenzyme-binding site in a monomer leads to an increased rate of dissociation of NADH, increasing the specific activity of the monomer. Consequently, higher ethanol concentrations were required to drive the reaction with the monomers compared to the homodimers, but the monomer reaction proceeded at a faster rate once it was underway (Ehrig et al., 1993Go). Thus, monomers possess enzymatic activity towards ethanol albeit atypical.

Given that the dimerization of subunits is necessary to obtain typical alcohol dehydrogenase activity, an important issue is whether or not the subunits of homodimers interact differently when present in heterodimers. That is, are the activities of heterodimers predictable based on knowledge of the homodimer activity? Extensive kinetic information has been published for the homodimers but not for the heterodimers. Although Mardh et al. (1986)Go stated that the turnover rates of human AB1, B1C1, and B1C2 isozymes were considerably faster than would be predicted from the values of the relevant homodimers, the evidence to support this statement is not compelling. Wagner et al. (1983)Go, who were quoted by Mardh et al. (1986)Go in support of their conclusion, examined turnover numbers of certain heterodimers and homodimers. Of the heterodimers examined, only the B1C2 form had a turnover number that was not between the two respective homodimers. Furthermore, the authors stated that, in successive liver preparations, turnover values were found to vary no more than 3-fold (Wagner et al., 1983Go), indicating that B1C2 heterodimer activity could have been between that of B1B1 and C2C2.

Yin et al. (1984)Go have reported that the Vmax and Km values for ethanol metabolism by B1B2, AB2, and B2C1 were intermediate between the respective homodimers. Likewise, Fong and Keung (1987) reported that AB2, B2C2, and B2C1 heterodimers had activity that can be predicted from the individual subunits, although the substrate was cyclohexanol and not ethanol. Therefore, for the current ethanol model it was assumed that heterodimer activity could be predicted from the activity of each subunit. In other words, each of the two active sites contained within a heterodimer was assumed to act independently as a monomer. Hence, the concentration of each active site of a heterodimer was equal to one-half of the heterodimer concentration. The differential equations (based on the Theorell Chance kinetic mechanism) for a heterodimer metabolism submodel, using the AB2 heterodimer as an example, are as follows:

(21)

(22)

(23)

(24)
where A is the monomer equivalent to one-half of the AA homodimer; AAk1 is the second-order rate constant for the association of the AA homodimer (or monomer) and NAD+ (N); AAk2 is the second-order rate constant for the association of ethanol (B) with A bound to NAD+ (AN); AAk3 is the first-order rate constant for the dissociation of NADH from the monomer A; AAk10 is the first-order rate constant for the breakdown of A bound to NAD+ (AN) to free A and free NAD+ (N); AAk20 is the second-order rate constant for the association of acetaldehyde with the AA homodimer (or monomer) bound to NADH (Q) to form ethanol (B); AN is the monomer bound to NAD+ (N); and AQ is NADH (Q) bound to the monomer A. The B2 monomeric portion of the AB2 heterodimer is modeled in a manner identical to the A monomer, except that B2B2 rate constants are substituted for the AA rate constants and the B2 monomer replaces the A monomer. Note that all rate constants for each monomeric portion of the heterodimer are equivalent to the same rate constants for the respective homodimer.

The algebraic expressions describing the amount of A monomer and B2 monomer present during the metabolism of ethanol are represented as the following equations:

(25)

(26)
where AB2T represents the total concentration of the AB2 heterodimer and kAB2 represents an operator switch to "turn on" or "turn off" the AB2 heterodimer, depending on which genotype is being simulated. Note that each monomer concentration is one-half of the AB2T concentration.

Solution of equations 21-26 for the AB2 heterodimer resulted in a rate of metabolism in vitro of the AB2 heterodimer that was midway between the rates of metabolism in vitro of the AA and B2B2 homodimers (Fig. 3). All additional heterodimers were modeled in a similar fashion, and activities of heterodimer submodels were midway between that of the respective homodimers (data not shown), indicating that the heterodimer submodels functioned as intended. Finally, it was assumed that the levels of various isoforms within a specific individual were expressed equally. No published information is available regarding expression levels of alcohol dehydrogenase isoforms.



View larger version (17K):
[in this window]
[in a new window]
 
FIG. 3. Simulated in vitro metabolism of ethanol by the AA and B2B2 homodimers compared to simulated in vitro metabolism of ethanol by the AB2 heterodimer (based on Equations 17-22). The dashed line represents the B2B2 submodel output, the dotted line represents the AA submodel output, and the solid line represents the AB2 submodel output. All simulations were done with an ethanol concentration of 10 mM, an active site concentration of 1 µM, and an initial NAD+ concentration of 2.4 mM, thereby simulating the in vitro incubation conditions of Bosron et al. (1983)Go and Yin et al. (1984)Go.

 
Cytochrome P450-dependent ethanol metabolism.
Ethanol has been shown to undergo biotransformation by a cytochrome P450-dependent (CYP2E1) pathway (Lieber, 1997Go). As discussed by Matsumoto and Fukui (2002)Go, the contribution of CYP2E1 in the metabolism of ethanol is relatively small at low ethanol concentrations but increases as the ethanol concentration increases. Kinetic constants for metabolism of ethanol by CYP2E1 were taken from Mezey and Tobon (1971)Go, as described by Derr (1993)Go, and were included in the current model in the form of a Michaelis Menten expression.

Conducting simulations.
All simulations were conducted using Advanced Continuous Simulation Language (ACSL, AEgis Technologies Group, Inc., Huntsville, AL) on a PC. Monte Carlo simulations were performed in ACSL Math.

RESULTS

Incorporation of genetic control of alcohol dehydrogenase activity into a previously developed physiologically based pharmacokinetic model for ethanol (Pastino et al., 2000Go) was based on inclusion of metabolism submodels for each possible enzyme isoform, which were then "switched on" or "switched off" as a function of the genotype simulated. The submodels were constructed based on rate constants descriptive of the Theorell Chance mechanism for each enzyme form (Table 1 and Fig. 1). Figure 4 shows examples of four metabolism submodels accurately simulating in vitro metabolism of ethanol by those homodimers reported to display substrate inhibition or negative cooperativity. These four homodimers are the only homodimers for which primary data from in vitro incubations have been published.



View larger version (26K):
[in this window]
[in a new window]
 
FIG. 4. Examples of homodimer submodel simulation of in vitro ethanol metabolism compared to published empirical data. Solid lines represent submodel output while open circles represent empirical data taken from Bosron et al. (1983)Go. The panels represent in vitro metabolism of ethanol by the following homodimers: AA, upper left panel; B1B1, upper right panel; C1C1, lower left panel; and C2C2, lower right panel. The active site concentration was 1 µM with an initial NAD+ concentration of 2.4 mM, thereby simulating the in vitro incubation conditions of Bosron et al. (1983)Go.

 
The equilibrium for alcohol dehydrogenase has been reported to be in favor of ethanol and NAD+ (Backlin, 1958Go; Bosron et al., 1993Go). Stated differently, alcohol dehydrogenase prefers to make ethanol and NAD+ from acetaldehyde and NADH, rather than acetaldehyde and NADH from ethanol and NAD+. The metabolism submodel for each form simulated this observation if allowed to run to equilibrium (data not shown). Consequently, for ethanol to be metabolized under conditions other than initial rate conditions, acetaldehyde and NADH must be removed to prevent their accumulation (Bosron et al., 1993Go). In the intact liver, removal of acetaldehyde occurs through aldehyde dehydrogenase (EC 1.2.1.3, a family of enzymes also displaying genetic polymorphisms), while NADH is oxidized to NAD+ by malate dehydrogenase (EC 1.1.1.37). In our model, for simplicity, both acetaldehyde and NADH were removed through first-order simulated reactions controlled by the first-order rate constants ke and kq, respectively. The values for ke and kq (100,000 h-1) were selected because they eliminated any significant product inhibition of alcohol dehydrogenase (data not shown).

Empirical determination of the total alcohol dehydrogenase active site concentration (ET in Equation 8) within liver has not been previously reported. Therefore, optimization of the model to three data sets from subjects of known genotypes (Thomasson et al., 1995Go) was used to estimate this important parameter (Fig. 5). It should be noted that Thomasson et al. (1995)Go presented only one representative set of blood ethanol decay data. However, Thomasson et al. (1995)Go also presented mean values for pharmacokinetic parameters in each genotype (such as dose, y intercept, and zero-order rate constant for elimination) that allowed the calculation of the mean blood ethanol levels at the time points at which blood was sampled from the subjects. These mean blood ethanol levels are the data points shown in Figure 5. Since all three genotypes had similar active site concentrations (Fig. 5), the mean value (0.03 mM) was selected for use in the model.



View larger version (16K):
[in this window]
[in a new window]
 
FIG. 5. Determination of total hepatic alcohol dehydrogenase active site concentration. The physiologically based pharmacokinetic ethanol model was optimized to three separate data sets derived from subjects with three different genotypes to determine alcohol dehydrogenase active site concentration. The genotypes were as follows: top panel, ADH1B1/B1; middle panel, ADH1B1/B3; and lower panel, ADH1B3B3. All subjects were ADH1C1C1 (Thomasson et al., 1995Go). The open circles in the three panels represent mean ethanol values in 50 subjects (top panel), 50 subjects (middle panel), and 10 subjects (lower panel) with the indicated genotypes (Thomasson et al., 1995Go). The solid lines represent output optimized for total hepatic alcohol dehydrogenase active site concentration (the mean value was 0.03 mM).

 
Application of the ethanol physiologically based pharmacokinetic model to two previously published data sets, describing the pharmacokinetic disposition of ethanol in subjects of known genotype, resulted in reasonably good fits of the simulated data to the empirical data sets (Figs. 6 and 7). While the empirical data from Mizoi et al. (1994)Go seemed to indicate that the parameter ß60 (the slope of the pseudolinear portion of an ethanol blood decay curve) was unaffected by alcohol dehydrogenase genotype (Fig. 7), the model simulations suggested that ß60 was not markedly sensitive to genotype and that slight to moderate differences in simulated ß60 values as a function of genotype might not have been detected in the data of Mizoi et al. (1994)Go because of the variability of the empirical data (Fig. 7). It is worth noting that published pharmacokinetic analyses of ethanol disposition in humans of known genotype are scarce, and those that are available have been presented here (Figs. 5-7).



View larger version (21K):
[in this window]
[in a new window]
 
FIG. 6. Comparison of empirical data with the simulation of intravenous ethanol infusion (200 mg/kg) over 30 min in subjects with different genotypes. The symbols in the graphs in each panel represent data from individual subjects taken from Yoshihara et al. (2000)Go. The subjects in the upper panel were ADH1B2/B2C1/C1 and the subjects in the lower panel were ADH1B1/B2C1/C1. The solid lines represent the model output based on mean values for physiological parameters reported in Table 2. The dotted lines represent model output using physiological values that give the largest and smallest, respectively, AUCs from 1000 Monte Carlo samplings using the parameters presented in Table 2; the dashed lines represent the 95% confidence limits.

 


View larger version (27K):
[in this window]
[in a new window]
 
FIG. 7. Comparison of ß60 values derived from empirical and simulated ethanol blood decay curves following an oral dose (400 mg/kg) of ethanol in subjects with different genotypes. The open bars represent the means and standard deviations of pooled ß60 values reported by Mizoi et al. (1994)Go at the indicated genotypes. The ADH1C genotypes were not reported by Mizoi et al. (1994)Go. A ß60 value is the slope of the pseudo-linear portion of a blood ethanol decay curve following ethanol administration. Mizoi et al. (1994)Go determined ß60 values in subjects from their blood ethanol concentrations at 1, 2, 3, and 4 h following ethanol administration. The filled bars represent the means and standard deviations of ß60 values calculated from simulated blood decay data at those same time points. The three filled bars at each simulated genotype correspond from left to right to the genotypes ADH1C1/C1, ADH1C1/C2, and ADH1C2/C2.

 
Using Monte Carlo sampling of physiological parameters (Table 1) linked to the ethanol, physiologically based pharmacokinetic model, the simulation of an oral ethanol dose of 525 mg/kg demonstrated that an individual's alcohol dehydrogenase genotype played a slight to moderate role in modulating the peak ethanol blood level achieved and the time-to-peak blood ethanol level (Figs. 8 and 9). Simulated individuals that possessed a B2 subunit consistently had lower peak blood ethanol levels and shorter times-to-peak blood ethanol levels than did subjects without a B2 subunit, although there was considerable overlap in the distribution between these populations. Moreover, the simulations presented in Figures 8 and 9 indicate that subjects with the genotype ADH1B2/B2 consistently had the shortest time-to-peak blood ethanol levels and the lowest peak blood ethanol levels. In contrast, the area under the curve (AUC) of the ethanol blood decay curve was very sensitive to alcohol dehydrogenase genotype, with AUCs from any genotype including the ADH1B2 allele considerably smaller than AUCs from any genotype without the ADH1B2 allele (Fig. 10). Moreover, the AUCs in the ADH1C1/C1 and ADH1C1C2 genotypes without a B2 subunit were moderately lower than the AUCs from the corresponding ADH1C2/C2 genotype (Fig. 10). The reduced AUCs documented for the genotypes noted above reflect the more favorable kinetic constants for ethanol metabolism by the B2 and C1 subunits compared to other subunits (Bosron et al., 1983Go; Burnell et al., 1989Go; Yin et al., 1984Go).



View larger version (32K):
[in this window]
[in a new window]
 
FIG. 8. Box plots of peak blood levels generated from Monte Carlo simulations (1000) of the ethanol physiologically based pharmacokinetic model simulating an oral dose of 525 mg/kg in subjects with different alcohol dehydrogenase phenotypes. The solid and dotted lines within each box represent the median and mean, respectively. The upper and lower edges of each box represent the 75th and 25th percentiles, respectively, while the upper and lower ends of the error bars represent the 90th and 10th percentiles, respectively. Finally, the upper and lower filled circles represent the 95th and 5th percentiles, respectively.

 


View larger version (26K):
[in this window]
[in a new window]
 
FIG. 9. Box plots of time-to-peak blood levels generated from Monte Carlo simulations (1000) of the ethanol physiologically based pharmacokinetic model simulating an oral dose of 525 mg/kg in subjects with different alcohol dehydrogenase phenotypes. The solid and dotted lines within each box represent the median and mean, respectively. The upper and lower edges of each box represent the 75th and 25th percentiles, respectively, while the upper and lower ends of the error bars represent the 90th and 10th percentiles, respectively. Finally, the upper and lower filled circles represent the 95th and 5th percentiles, respectively.

 


View larger version (29K):
[in this window]
[in a new window]
 
FIG. 10. Box plots of AUC of the ethanol blood decay curves generated from Monte Carlo simulations (1000) of the ethanol physiologically based pharmacokinetic model simulating an oral dose of 525 mg/kg in subjects with different alcohol dehydrogenase phenotypes. The solid and dotted lines within each box represent the median and mean, respectively. The upper and lower edges of each box represent the 75th and 25th percentiles, respectively, while the upper and lower ends of the error bars represent the 90th and 10th percentiles, respectively. Finally, the upper and lower filled circles represent the 95th and 5th percentiles, respectively.

 
Simulations at three different oral doses of ethanol revealed that AUCs increased disproportionately compared with the dose (Fig. 11). For example, a 3-fold increase in dose (from 263 to 788 mg/kg) led to an 11-fold increase in the median AUC (188 to 2080 mg-h/l) in simulated subjects with an ADH1C1/C1 phenotype but lacking the B2 subunit (Fig. 11). Similarly, in these subjects, the 95th and 5th AUC percentiles increased 9-fold (398 to 3470 mg-h/l) and 12-fold (90 to 1117 mg-h/l), respectively (Fig. 10). Such disproportionate increases are expected with chemicals that undergo mixed-order or pseudo-zero-order elimination.



View larger version (25K):
[in this window]
[in a new window]
 
FIG. 11. Box plots of AUC of the ethanol blood decay curves generated from Monte Carlo simulations (1000) of the ethanol physiologically based pharmacokinetic model simulating three different oral doses of ethanol in subjects with different alcohol dehydrogenase phenotypes. The solid and dotted lines within each box represent the median and mean, respectively. The upper and lower edges of each box represent the 75th and 25th percentiles, respectively, while the upper and lower ends of the error bars represent the 90th and 10th percentiles, respectively. Finally, the upper and lower filled circles represent the 95th and fifth percentiles, respectively. For simplicity, subjects with various ADH1C genotypes were grouped as with and without B2 subunits within each dose level. The letters on the abscissa at each indicated dose level represent the simulated AUCs from the following: A, all ADH1C1C1 genotypes without B2 subunits; B, all ADH1C1C1 genotypes with B2 subunits; C, all ADH1C2C2 genotypes without B2 subunits; D, all ADH1C2C2 genotypes with B2 subunits; E, all ADH1C1C2 genotypes without B2 subunits; and F, all ADH1C1C2 genotypes with B2 subunits.

 
DISCUSSION

The rate of ethanol metabolism in vivo relates to the amount of alcohol dehydrogenase present; the functional characteristics of the isoforms present; and the concentrations of NAD+, NADH, and acetaldehyde (Li et al., 2001Go). Since acetaldehyde is a strong inhibitor of ethanol metabolism, the total amount of aldehyde dehydrogenase and the functional characteristics of the isoforms present might modulate the rate of ethanol metabolism by alcohol dehydrogenase (Bosron et al., 1993Go; Eriksson et al., 2001Go; Li et al., 2001Go). However, hepatic aldehyde dehydrogenase activity is likely very high, with the exception of ALDH2*2 since blood acetaldehyde levels following ethanol administration are usually reported as a very small fraction of the blood ethanol and acetate levels (Li et al., 2001Go). The ALDH2*2 allele encodes for a physiologically inactive aldehyde dehydrogenase that might lead to some acetaldehyde accumulation following ethanol exposure (Borras et al., 2000Go), although there is no direct evidence for this assertion. The ALDH2*2 allele is very rare in non-Asian populations. In our model, significant simulated accumulation of NADH and/or acetaldehyde slowed the rate of ethanol metabolism (data not shown). Therefore, to focus exclusively on the role of alcohol dehydrogenase genetics on ethanol disposition, NAD+ was set at a constant physiological level, while NADH and acetaldehyde were reduced to concentrations that did not result in significant product inhibition. Though minimizing hepatic NADH and acetaldehyde concentrations could have represented an oversimplification, any limited product inhibition that might occur in vivo with alcohol dehydrogenase and that is unaccounted for in the ethanol model would be compensated for by a slightly lower model enzyme active site concentration (since enzyme active site concentration was determined by optimization).

As noted by Thomasson et al. (1995)Go, Bosron and Li (1986)Go reported that the Vmax for ethanol for the B2B2 and B3B3 homodimers greatly exceeded that of the B1B1 homodimer. For example, the B2B2 homodimer Vmax was about 44 times greater than the B1B1 Vmax (Bosron and Li, 1986Go). In our study, while the presence of a B2 subunit reduced the ethanol AUC by increasing ethanol metabolism (Fig. 10), this AUC reduction was not as great as the Vmax differences documented in Bosron and Li (1986)Go. However, Bosron and Li (1986)Go compared activities of purified homodimers in vitro under optimal conditions, whereas the present study compared AUCs between simulated subjects of different genotypes, where each individual could have six to 15 isoforms of varying activity present depending on that individual's genotype. Therefore, it is not surprising that the comparison of in vitro activities of individual purified homodimers under ideal conditions show greater differences than do the ethanol-metabolizing capacity of simulated subjects of different genotypes.

Many drug-metabolizing enzymes and drug transporters have been shown to exist in multiple forms as a result of genetic polymorphisms and/or polygenic control (Nebert, 1999Go). However, the functional significance of this hereditary control is often difficult to assess without the existence of extremely discordant phenotypes (Nebert, 2000Go). Biologically based modeling can play an important role in assessing the importance of hereditary control of chemical disposition and/or chemical response in the absence of extremely discordant phenotypes. In the case of alcohol dehydrogenase, some differences in phenotype might be expected as a result of the differences in activities between the isoforms encoded by ADH1 genes (Borras et al., 2000Go; Goedde et al., 1992). Yet, as described by Borras et al. (2000)Go, two reports failed to find differences in the ß60 values (slopes of the pseudo-linear portion of the ethanol blood decay curves) of ethanol between east Asian individuals with the ADH1B1 and ADH1B2 alleles (Mizoi et al., 1994Go; Yamamoto et al., 1993). However, the modeling results we have presented indicate that the selection of the pharmacokinetic parameter(s) evaluated is important in the comparison of ethanol disposition in subjects of different genotypes (Figs. 8-10). Neither Yamamoto et al. (1993) nor Mizoi et al. (1994)Go evaluated blood ethanol AUC, the pharmacokinetic parameter that displayed the most dependence on genotype (Figs. 8- 10). Furthermore, simulations of the Mizoi et al. (1994)Go experiments (Fig. 7) suggested that genotype-dependent differences in ethanol pharmacokinetic disposition could have been obscured since the ADH1C genotypes were not reported and since ß60 might not be the optimal pharmacokinetic parameter to analyze. The simulations shown in Figure 7 indicate that ß60, a parameter often used to describe ethanol pharmacokinetic disposition, was only slightly to moderately affected by alcohol dehydrogenase genotype.

As discussed by Haber et al. (2002)Go and Gentry et al. (2002)Go, safe or allowable exposure levels to environmental toxicants are often established by the identification of a critical effect level, which is then lowered by divisions with uncertainty factors to account for certain extrapolations and deficiencies in data (Gentry et al., 2002Go). In recent years, efforts have been made to refine the selection of more appropriate uncertainty factors. For example, separation of intraspecies uncertainty into toxicokinetic and toxicodynamic components to form chemical-specific adjustment factors allows the introduction of adjustment factors that are based on available toxicokinetic and/or toxicodynamic data (Gentry et al., 2002Go). As such, physiologically based pharmacokinetic modeling linked with Monte Carlo sampling has been shown to be a useful tool for quantification of interindividual variability in chemical disposition and/or response when applied to biological processes that displayed single genetic polymorphisms (Barton et al., 1996Go; El-Masri et al., 1999Go; Gentry et al., 2002Go; Haber et al., 2002Go). We have extended this approach by modeling the complex hereditary control of alcohol dehydrogenase, which includes polygenic control and polymorphisms at two alleles, and by assessing the functional significance of this hereditary control on ethanol disposition. Two observations are noteworthy. Firstly, pharmacokinetic variability assessed by peak blood ethanol concentration and time-to-peak blood levels was substantially lower than variability assessed by AUC (Figs. 810). Consequently, considerably different data-derived uncertainty factors or chemical-specific adjustment factors could be calculated, depending on which pharmacokinetic parameter is chosen. Secondly, the variability of blood ethanol AUC increased in a manner disproportionate to dose, probably due to mixed-order or pseudo-zero-order metabolism. Therefore, different chemical-specific adjustment factors could be derived from the pharmacokinetic disposition of ethanol at different doses.

An important objective of the present study was to identify the kinds of data that might assist in the incorporation of complex hereditary control of chemical disposition into physiologically based pharmacokinetic models. Those kinds of data that applied to modeling genetic control of alcohol dehydrogenase and that could be extended to other biotransformation enzymes under complex genetic control are listed in Table 3, including some significant data gaps that were identified. With this report, expression levels of alcohol dehydrogenase forms would have eliminated the need for optimization to obtain enzyme active site concentration (Fig. 5), while additional studies documenting ethanol disposition in subjects of known genotypes would have allowed improved model validation. Undoubtedly, the extensive kinetic analyses of purified human alcohol dehydrogenase homodimers (Bosron et al., 1983Go; Yin et al., 1984Go) were key in the design of the metabolic submodels for each isoform. This is somewhat at odds with Haber et al. (2002)Go, who more strongly endorsed the use of enzyme activities from tissues of individuals having known genotypes rather than activities from purified variant proteins.


View this table:
[in this window]
[in a new window]
 
TABLE 3 Useful Data for the Incorporation of Complex Hereditary Control of a Biotransforming Enzyme into a Physiologically Based Pharmacokinetic Model

 
Whether to use a more molecular approach or a tissue activity approach should be considered on a case-by-case basis. There are certain enzymes, such as alcohol dehydrogenase, where simple Michaelis Menten kinetics applied to tissue homogenates probably should not be used to describe their complex reaction kinetics, since there are multiple substrates and multiple products that require multiple Vmax and Km values. Moreover, for complex enzymatic reaction schemes and complex hereditary control, a detailed knowledge of reaction mechanisms, coupled with an understanding of how these processes are affected by hereditary control, can provide a model with important information that might otherwise be too impractical to obtain. For example, there are 18 possible alcohol dehydrogenase genotypes with respect to class 1, giving rise to a possible 21 different alcohol dehydrogenase forms. The determination of enzyme activities in liver samples from groups of subjects of each possible phenotype would be problematic, particularly in light of the paucity of studies simply documenting blood ethanol levels in subjects of known alcohol dehydrogenase genotype.

The limited number of published studies documenting the pharmacokinetic disposition of ethanol in subjects of known genotype represents a significant data gap (Table 3). While the modeling results presented for this study suggest that alcohol dehydrogenase genotype has a significant impact on ethanol pharmacokinetic disposition (Fig. 10), the functional significance of the genetic control of alcohol dehydrogenase remains an unresolved issue.

ACKNOWLEDGMENTS

This work was funded by a research grant from the American Chemistry Council. All model code will be made available upon request to the authors.

REFERENCES

Anderson, L., and Mosbach, K. (1979). The use of solid-phase techniques in the study of alcohol dehydrogenase. 2. Selective carboxymethylation of bioaffinity-bound alcohol dehydrogenase. Eur. J. Biochem. 94, 565–571.[ISI][Medline]

Backlin, K. I. (1958). The equilibrium constant of the system ethanol, aldehyde, DPN+, DPNH, and H+. Acta. Chem. Scand. 12, 1279–1285.[ISI]

Barton, H. A., Flemming, C. D., and Lipscomb, J. C. (1996). Evaluating human variability in chemical risk assessment: Hazard identification and dose-response assessment for noncancer oral toxicity of trichloroethylene. Toxicology 111, 271–297.[CrossRef][ISI][Medline]

Borras, E., Coutelle, C., Rosell, A., Fernandez-Muixi, F., Broch, M., Crosas, B., Hjelmqvist, L., Lorenzo, A., Gutierrez, C., Santos, M., et al. (2000). Genetic polymorphism of alcohol dehydrogenase in Europeans: The ADH2*2 allele decreases the risk for alcoholism and is associated with ADH3*1. Hepatology 31, 984–989.[ISI][Medline]

Bosron, W. F., Ehrig, M. D., and Li, T.-K. (1993). Genetic factors in alcohol metabolism and alcoholism. Sem. Liv. Dis. 13, 126–135.[ISI]

Bosron, W. F., and Li, T.-K. (1986). Genetic polymorphisms of human liver alcohol and aldehyde dehydrogenases, and their relationship to alcohol metabolism and alcoholism. Hepatology 6, 502–510.[ISI][Medline]

Bosron, W. F., Magnes, L. J., and Li, T.-K. (1983). Kinetic and electrophoretic properties of native and recombined isoenzymes of human liver alcohol dehydrogenase. Biochemistry 22, 1852–1857.[ISI][Medline]

Briganti, F., Fong, W. P., Auld, D. S., and Vallee, B. L. (1989). In vitro dissociation and reassociation of human alcohol dehydrogenase class 1 isozymes. Biochemistry 28, 5374–5379.[ISI][Medline]

Brown, R. P., Delp, M. D., Lindstedt, S. L., Rhomberg, L. R., Beliles, R. P. (1997). Physiological parameter values for physiologically based pharmacokinetic models. Toxicol. Ind. Health 13, 407–484.[ISI][Medline]

Burnell, J. C., Li, T.-K., and Bosron, W. F. (1989). Purification and steady-state kinetic characterization of human liver ß3ß3 alcohol dehydrogenase. Biochemistry 28, 6810–6815.[ISI][Medline]

Clewell, H. J., and Andersen, M. E. (1996). Use of physiologically based pharmacokinetic modeling to investigate individual versus population risk. Toxicology 111, 315–329.[CrossRef][ISI][Medline]

Clewell, H. J., Gearhart, J. M., Gentry, P. R., Covington, T. R., VanLandingham, C. B., Crump, K. S., and Shipp, A. M. (1999). Evaluation of the uncertainty in an oral reference dose for methylmercury due to interindividual variability in pharmacokinetics. Risk Anal. 19, 547–558.[CrossRef][ISI][Medline]

Clewell, H. J., Gentry, P. R., Covington, T. R., and Gearhart, J. M. (2000). Development of a physiologically based pharmacokinetic model of trichloroethylene and its metabolites for use in risk assessment. Environ. Health Perspect. 108(Suppl. 2), 283–305.

Clewell, H. J., Gentry, P. R., Gearhart, J. M., Allen, B. C., and Andersen, M. E. (2001). Comparison of cancer risk estimates for vinyl chloride using animal and human data with a PBPK model. Sci. Total Environ. 274, 37–66.[CrossRef][ISI][Medline]

Crow, K. E., Braggins, T. J., Batt, R. D., and Hardman, M. J. (1982). Rat liver cytosolic malate dehydrogenase: Purification, kinetic properties, role of free cytosolic NADH concentration. J. Biol. Chem. 257, 14217–14225.[Free Full Text]

Derr, R. F. (1993). Simulation studies on ethanol metabolism in different human populations with a physiological pharmacokinetic model. J. Pharm. Sci. 82, 677–682.[ISI][Medline]

Dorfler, P., Schliesser, M., Maurer, M., and Becker, G. (2000). Measurement of cerebral blood flow volume by extracranial sonography. J. Cereb. Blood Flow Metab. 20, 269–271.[CrossRef][ISI][Medline]

Duester, G., Farred, J., Felder, M., Holmes, R., Hoog, J.-O., Pares, X., Plapp, B. V., Yin, S.-J., and Jornall, H. (1999). Recommended nomenclature for the vertebrate alcohol dehydrogenase gene family. Biochem. Pharmacol. 58, 389–395.[CrossRef][ISI][Medline]

Ehrig, T., Muhoberac, B. B., Brems, D., and Bosron, W. F. (1993). Monomers of human ß1ß1 alcohol dehydrogenase exhibit activity that differs from the monomer. J. Biol. Chem. 268, 11721–11726.[Abstract/Free Full Text]

El-Masri, H. A., Bell, D. A., and Portier, C. J. (1999). Effects of glutathione transferase {theta} polymorphism on the risk estimates of dichloromethane to humans. Toxicol. Appl. Pharmacol. 158, 221–230.[CrossRef][ISI][Medline]

Eriksson, C., Fukunaga, T., Sarkola, T., Chen, W., Chen, C., Ju, J., Cheng, A., Yamamoto, H., Kohlenberg-Muller, K., Kimura, M., et al. (2001). Functional relevance of human ADH polymorphism. Alcohol. Clin. Exp. Res. 25, 157S–163S.[ISI][Medline]

Fong, W. P., and Keung, W. M. (1989). ß2 (oriental) human liver alcohol dehydrogenase does not exhibit subunit interaction. Oxidation of cyclohexanol by homo- and heterodimers. Biochemistry 26, 5733–5738.

Gentry, P. R., Hack, C. E., Haber, L., Maier, A., Clewell, H. J. (2002). An approach for the quantitative consideration of genetic polymorphism data in chemical risk assessment: Examples with warfarin and parathion. Toxicol. Sci. 70, 120–139.[Abstract/Free Full Text]

Goedde, H. W., Argarwal, D. P., Fritze, G., Meier-Tackmann, D., Singh, S., Beckman, G., Bhatia, K., et al. (1992). Distribution of ALDH2 and ADH2 genotypes in different populations. Hum. Genet. 88, 344–346.[ISI][Medline]

Haber, L. T., Maier, A., Gentry, P. R., Clewell, H. J., and Dourson, M. L. (2002). Genetic polymorphisms in assessing interindividual variability in delivered dose. Reg. Toxicol. Pharmacol. 35, 177–197.[CrossRef][ISI][Medline]

Jornvall, H., Hoog, J.-O., Persson, B., and Pares, X. (2000). Pharmacogenetics of the alcohol dehydrogenase system. Pharmacology 61, 184–191.[CrossRef][ISI][Medline]

Kuiken, S. D., Samson, M., Camilleri, M., Mullan, B. P., Nurton, D. D., Kost, L. J., Hardyman, T. J., Brinkmann, B. H., and O'Connor, M. K. (1999). Development of a test to measure gastric accommondation in humans. Am. J. Physiol. 277, G1217–G1221.[ISI][Medline]

Li, T.-K., Yin, S.-J., Crabb, D. W., O'Connor, S., and Ramchandani, V. A. (2001). Genetic and environmental influences on alcohol metabolism in humans. Alcohol. Clin. Exp. Res. 25, 134–144.

Lieber, C. (1997). Cytochrome P-4502E1: Its physiological and pathological role. Physiol. Rev. 77, 517–539.[Abstract/Free Full Text]

Mardh, G., Falchuk, K. H., Auld, D. S., and Vallee, B. L. (1986). Testosterone allosterically regulates ethanol oxidation by homo- and heterodimeric {gamma}-subunit-containing isozymes of human alcohol dehydrogenase. Proc. Natl. Acad. Sci. USA 83, 2836–2840.[Abstract]

Matsumoto, H., and Fukui, Y. (2002). Pharmacokinetics of ethanol: A review of the methodology. Addict. Biol. 7, 5–14.[CrossRef][ISI][Medline]

Mezey, E., and Tobon, F. (1971). Rates of ethanol clearance and activities of the ethanol-oxidizing enzymes in chronic alcoholic patients. Gastroenterology 61, 707–715.[ISI][Medline]

Mizoi, Y., Yamamoto, K., Ueno, Y., Fukunaga, T., and Harada, S. (1994). Involvement of genetic polymorphism of alcohol and aldehyde dehydrogenases in individual variation of alcohol metabolism. Alcohol Alcohol. 29, 707–710.[Abstract]

Nebert, D. W. (1999). Pharmacogenetics and pharmacogenomics: Why is this relevant to the clinical geneticist? Clin. Genet. 56, 247–258.[CrossRef][ISI][Medline]

Nebert, D. W. (2000). Extreme discordant phenotype methodology: An intuitive approach to clinical pharmacogenetics. Eur. J. Pharmacol. 410, 107–120.[CrossRef][ISI][Medline]

Pastino, G. M., Flynn, E. J., and Sultatos, L. G. (2000). Genetic polymorphisms in ethanol metabolism: Issues and goals for physiologically based pharmacokinetic modeling. Drug Chem. Toxicol. 23, 179–201.[CrossRef][ISI][Medline]

Plapp, B. V., Leidal, K. G., Smith, R. K., and Murch, B. P. (1984). Kinetics of inhibition of ethanol metabolism in rats and the rate-limiting role of alcohol dehydrogenase. Arch. Biochem. Biophys. 230, 30–38.[ISI][Medline]

Portier, C. J., and Kaplan, N. L. (1989). Variability of safe dose estimates when using complicated models of the carcinogenic process. Fund. Appl. Toxicol. 13, 533–544.[ISI][Medline]

Segel, I. H. (1975). Enzyme Kinetics, pp. 593–605. John Wiley & Sons, New York.

Theorell, H., and Chance, B. (1951). Studies on liver alcohol dehydrogenase. 2. The kinetics of the compound of horse liver alcohol dehydrogenase and reduced diphosphopyridine nucleotide. Acta Chem. Scand. 5, 1127–1144.[ISI]

Thomas, R. S., Bigelow, P. L., Keefe, T. J., and Yang, R. S. H. (1996a). Variability in biological exposure indices using physiologically based pharmacokinetic modeling and Monte Carlo simulation. Am. Ind. Hyg. Assoc. J. 57, 23–32.[ISI][Medline]

Thomas, R. S., Lytle, W. E., Keefe, T. J., Constan, A. A., and Yang, R. S. H. (1996b). Incorporating Monte Carlo simulation into physiologically based pharmacokinetic models using Advanced Continuous Simulation Language (ACSL): A computational method. Toxicol. Appl. Pharmacol. 31, 19–28.[CrossRef]

Thomasson, H. R., Beard, J. D., and Li, T.-K. (1995). ADH2 gene polymorphisms are determinants of alcohol pharmacokinetics. Alcohol. Clin. Exp. Res. 19, 1491–1499.

Wagner, F. W., Burger, A. R., and Vallee, B. L. (1983). Kinetic properties of human liver alcohol dehydrogenase: Oxidation of alcohols by class 1 isoenzymes. Biochemistry 22, 1857–1863.[ISI][Medline]

Whitwell, J. L., Crum, W. R., Watt, H. C., and Fox, N. C. (2001). Normalization of cerebral volumes by use of intracranial volume: Implications for longitudinal quantitative MR imaging. Am. J. Neuroradiol. 22, 1483–1489.[Abstract/Free Full Text]

Wilkinson, P. K., Sedman, A. J., Sakmar, E., Kay, D. R., and Wagner, J. G. (1977). Pharmacokinetics of ethanol after oral administration in the fasting state. J. Pharmacokin. Biopharm. 5, 207–224.[ISI][Medline]

Yamamoto, K., Ueno, Y., Mizoi, Y., and Tatsung, Y. (1993). Genetic polymorphism of alcohol and aldehyde dehydrogenase and the effects on alcohol metabolism. Jpn. J. Alcohol and Drug Dependence 28, 13–25.

Yin, S.-J., Bosron, W. F., Magnes, L. J., and Li, T.-K. (1984). Human liver alcohol dehydrogenase: Purification and kinetic characterization of the ß2ß2, {alpha}ß2, and ß2{gamma} "Oriental" isoenzymes. Biochemistry 23, 5847–5853.[ISI][Medline]

Yoshihara, E., Ameno, K., Nakamura, K., Ameno, M., Itoh, S., Ijiri, I., and Iwahashi, K. (2000). The effects of the ALDH*1/2, CYP2E1 C1/C2 and C/D genotypes on blood ethanol elimination. Drug Chem. Toxicol. 23, 371–379.[CrossRef][ISI][Medline]





This Article
Abstract
FREE Full Text (PDF)
All Versions of this Article:
78/1/20    most recent
kfh057v1
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Search for citing articles in:
ISI Web of Science (1)
Disclaimer
Request Permissions
Google Scholar
Articles by Sultatos, L. G.
Articles by Flynn, E. J.
PubMed
PubMed Citation
Articles by Sultatos, L. G.
Articles by Flynn, E. J.