ToxSci Advance Access originally published online on February 9, 2005
Toxicological Sciences 2005 85(1):429-446; doi:10.1093/toxsci/kfi103
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Derivation of a Human Equivalent Concentration for n-Butanol Using A Physiologically Based Pharmacokinetic Model for n-Butyl Acetate and Metabolites n-Butanol and n-Butyric Acid
* Battelle, Pacific Northwest National Laboratory, Richland, Washington 99352; Health and Environmental Laboratories, Eastman Kodak Company, Rochester, New York 14652-6272; Willem Faber Toxicology Consulting, Victor, New York 14564; NHEERL, USEPA, Research Triangle Park, North Carolina 27709; and ¶ ENVIRON Health Sciences Institute, Ruston, Louisiana 71270
1 To whom correspondence should be addressed at Pacific Northwest National Laboratory, PO Box 999, Mail Stop P7-56, Richland, WA 99352. Fax: (509) 376-9449. E-mail: justin.teeguarden{at}pnl.gov.
Received September 7, 2004; accepted December 20, 2004
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIAL AND METHODSRESULTSDISCUSSIONAPPENDIX 1APPENDIX 2SUPPLEMENTAL DATAREFERENCES |
|---|
The metabolic series approach for risk assessment uses a dosimetry-based analysis to develop toxicity information for a group of metabolically linked compounds using pharmacokinetic (PK) data for each compound and toxicity data for the parent compound. The metabolic series approach for n-butyl acetate and its subsequent metabolites, n-butanol and n-butyric acid (the butyl series), was first demonstrated using a provisional physiologically based pharmacokinetic (PBPK) model for the butyl series. The objective of this work was to complete development of the PBPK model for the butyl series. Rats were administered test compounds by iv bolus dose, iv infusion, or by inhalation in a recirculating closed chamber. Hepatic, vascular, and extravascular metabolic constants for metabolism were estimated by fitting the model to the blood time course data from these experiments. The respiratory bioavailability of n-butyl acetate (100% of alveolar ventilation) and n-butanol (50% of alveolar ventilation) was estimated from closed chamber inhalation studies and measured ventilation rates. The resulting butyl series PBPK model successfully reproduces the blood time course of these compounds following iv administration and inhalation exposure to n-butyl acetate and n-butanol in rats and arterial blood n-butanol kinetics following inhalation exposure to n-butanol in humans. These validated inhalation route models can be used to support species and dose-route extrapolations required for risk assessment of butyl series family of compounds. Human equivalent concentrations of 169 ppm and 1066 ppm n-butanol corresponding to the rat n-butyl acetate NOAELs of 500 and 3000 ppm were derived using the models.
Key Words: PBPK model; pharmacokinetics; acetates; extrapolation; risk assessment; metabolic series approach; n-butyl acetate; n-butanol; n-butyric acid.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTSDISCUSSIONAPPENDIX 1APPENDIX 2SUPPLEMENTAL DATAREFERENCES |
|---|
n-Butyl acetate, n-butanol, n-butyraldehyde, and n-butyric acid (the butyl series) are a family of metabolically related2 compounds also linked by an industrial production method relying on sequential oxidation/reduction reactions. In order to make toxicity testing and the development of acceptable exposure concentrations as efficient as possible, a methodology which takes advantage of the metabolic relationship between the butyl series compounds, called the metabolic series approach (previously, "the family approach"), was developed (Barton et al., 2000
). The metabolic series approach uses a dosimetry-based analysis to develop toxicity information for a group of metabolically linked compounds using pharmacokinetic data for each compound and toxicity data for the parent compound (Barton et al., 2000). In the case of the butyl series, exposure to the parent ester (n-butyl acetate) leads to systemic exposure to each of the three sequential metabolites, which are produced in stoichiometric yield. Exposure to n-butanol leads to systemic exposure to each of the two subsequent metabolites. Results of exposures with a single compound can be extrapolated to subsequent metabolites and from one dose route to another (e.g., oral to inhalation). Dose-route and interspecies extrapolations (e.g., rat inhalation to human inhalation) can be made with a physiologically based pharmacokinetic (PBPK) model. For such an approach to be effective, the PBPK model must be developed, and where possible, verified for exposure routes necessary for interpretation of toxicity studies in test animals and for routes for which there is the expectation of significant human exposure. This includes accurate characterization of absorption, distribution, and metabolic and renal clearance.
For both n-butyl acetate and n-butanol, the primary exposure route of concern for humans is inhalation; exposure to n-butyric acid is of lower concern, in part because the low odor threshold for this compound may be exposure limiting. Rat inhalation toxicity studies are available for butyl acetate (David et al., 1998, 2001). These studies result in co-exposure to n-butanol in rats, raising the possibility of identifying rat n-butanol inhalation exposures that would result in equivalent blood n-butanol concentrations. Consequently, there was a need identified to parameterize and validate a rat model for the inhalation exposure to both n-butyl acetate and n-butanol. The target tissue dose-effect relationships from the rodent toxicity studies could be extrapolated to humans with the use of the PBPK model parameterized for the human.
At the time of publication of the provisional butyl series model (Barton et al., 2000), the limitations of the existing pharmacokinetic (PK) data had been identified, and the initial model was available for assisting in the development and evaluation of additional experimental data required for parameterization and, in some cases, verification of the model. This included PK data to refine the characterization of systemic clearance of the butyl series compounds, and respiratory absorption (and bioavailability) following inhalation of n-butyl acetate and n-butanol in the rat. Clearance in the human would be based on principles of allometric scaling, and literature values for absorption following inhalation would be used for n-butyl acetate and n-butanol.
The objective of this work was to complete the development of rat and human PBPK models for butyl series compounds based on the intravenous (iv) kinetics of n-butyl acetate, n-butanol and n-butyric acid in the rat, and inhalation kinetics of n-butyl acetate (rat only) and n-butanol (rats and humans). In addition the PBPK model was used to estimate inhalation-route n-butanol human equivalent concentrations based on rat inhalation route NOAELs.
|
MATERIAL AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTSDISCUSSIONAPPENDIX 1APPENDIX 2SUPPLEMENTAL DATAREFERENCES |
|---|
Experimental Data
Data availability.
Data from six unpublished reports are utilized for model development (Corley et al., 2000
; Deisinger and English, 2000, 2001, 2003; Poet et al., 2003a,b). All studies involving animals were IACUC approved. Experimental details not provided here may be found in the reports, which are available by request from the sponsor (American Chemistry Council, 1300 Wilson Blvd., Arlington, VA 22209).
Rat iv bolus and iv infusion blood kinetics studies (Deisinger and English, 2001).
Young-adult, male Sprague-Dawley rats (280–330 g, Zivic-Miller Laboratories, Zelienople, PA) were surgically prepared by the vendor with femoral and jugular vein cannulae (Deisinger and English, 2001). Pilot studies were conducted to identify acutely nontoxic iv doses for n-butyl acetate, n-butanol, and n-butyric acid, verify the efficacy of the analytical methods, determine background concentrations for each of the analytes, and establish sampling times for the definitive studies. Pilot and definitive iv bolus studies were conducted using the same protocol. Groups of rats (five/group for definitive studies, one/group for pilot studies) were administered a bolus injection of each test chemical dissolved in a saline carrier (0.9% saline; plus 1% Tween 20 vehicle for n-butyl acetate) at 
0.28 and 0.028 (n-butyric acid only) mmol/kg via the femoral vein cannula. Exact administered doses are presented in figure legends. In addition, n-butyric acid was delivered by iv infusion, 0.28 mmol/kg delivered over 3 min, using a syringe pump (Kd Scientific Model 100, Boston, MA). Sample collection and processing was the same for bolus and infusion experiments. Aliquots of blood (100 µl) were obtained by serial sampling from the jugular vein cannula. The blood was immediately spiked with known amounts of hexanoic acid and isoamyl alcohol as internal standards, de-proteinized with 0.4 M H2SO4, extracted with diethyl ether, methylated, and analyzed by GC-MS. Mean data are presented as points in figures with error bars corresponding to two times the SE, equivalent to the 95% confidence limits on the mean value. Background concentrations, collected pre-dosing, were established for each compound. Results from the definitive studies were used for model development while those of the pilot study data were used to verify the model parameterization.
Rat Closed Chamber Exposure and Blood Kinetics (Poet et al., 2003a)
Animals and chamber.
Male Sprague-Dawley rats with indwelling jugular cannulae were purchased from two vendors. Rats used in the n-butanol studies (average body weight 317 g) and one group of two animals (5 and 8, average body weight 305 g) used in the n-butyl acetate studies were obtained from Hilltop Lab (Scottsdale, PA). A second group of four animals (average body weight 354 g) was obtained from Charles River (Raleigh, NC). These animals were significantly less active that the Hilltop Lab animals. The animals were housed in suspended plastic cages with chipped bedding and acclimated to the laboratory for at least three days prior to exposure. Gas uptake methods for several volatile water-soluble compounds, including n-butyl acetate, were developed as described previously (Corley et al., 2000). These methods were used in conjunction with plethysmography to measure both the gas uptake and respiratory parameters of the rats during exposures to target concentrations of 2000 ppm n-butyl acetate (Poet et al., 2003a) or n-butanol (Poet et al., 2003b). n-Butyl acetate was obtained from Sigma (St. Louis, MO, lot no. 13091CN ) and was 99.7% pure; n-butanol (lot no. 952729), obtained from FisherChem (Fairlawn, NJ), was 99.8% pure.
The gas uptake chamber used in previous experiments (Gargas et al., 1986) was modified to allow the introduction of the plethysmograph and its associated connections to a computer monitoring system, and to accommodate a cannula that was accessible for blood draws during the exposure (Supplemental Data, Fig. S1). The standard glass lid was replaced with a stainless steel lid. The volume of the closed chamber and restraint tube was 9.44 l. The restraint tube was included in the empty chamber simulations to capture losses associated with its presence. Oxygen was maintained at 20 ± 1%. Previous studies in this laboratory indicated that significant losses of test materials occurred with the use of a CO2 scrubber and water condensation from an ice-bath normally used to condense water vapor generated by the animals (Corley et al., 2000). Accordingly, the scrubber was removed, no ice bath was included, and chamber runs were limited to a single rat for 2 h.
Exposure and plethysmography.
A whole-body plethysmograph (designed and built at Battelle, Toxicology Northwest) linked to a Buxco Biosystem XA Data Acquisition System (Buxco Electronics, Inc. Sharon, CT) was used for non-invasive measurements of tidal volume (ml), respiratory rate and minute ventilation (ml/min) on conscious rats. The animals were restrained in a constant-pressure plethysmograph with an attached pneumotachograph. A neck seal separated the head chamber from the body chamber and ventilation parameters were computed from flow measurements through the pneumotachograph from the body chamber.
The rats were acclimated to a tube that was similar to the plethysmograph for 30 min/day for three days prior to the exposure. The actual restraint plethysmograph tube was not used for acclimation due to difficulties with the cannulation surgery site and the neck restraint system. Exposures included: (1) empty chamber to determine the loss to the system, (2) an individually exposed dead rat to determine the adsorptive losses of test material to the head of the rat, and (3) individually exposed live rats to determine the respiratory clearance from the gas uptake chamber while concurrently collecting respiratory parameters and blood samples for analysis of compound (n-butyl acetate or n-butanol) and its metabolites. Each rat was weighed and exposed to target concentrations of 2000 ppm of n-butyl acetate (Poet et al., 2003a) or n-butanol (Poet et al., 2003b) for 2 h in the recirculating inhalation chamber. Initial chamber concentrations varied from 1400–2000 ppm.
Chamber monitoring.
The concentration of n-butyl acetate or n-butanol was monitored using a Hewlett Packard 5890 Series II capillary gas chromatograph with flame ionization detection (GC/FID; Hewlett Packard, Palo Alto, CA). A 30 m x 0.32 mm ID DB-624 capilary column with 3 µm film thickness was used for separation (J&W Scientific, Folsom, CA). The concentration was monitored approximately every 7 min via automatic injections of 1 ml of chamber air via a Valco gas sampling valve. The GC/FID was calibrated with a standard curve consisting of standards (n-butanol or n-butyl acetate) prepared in 3-L Tedlar bags (SKC, Fullerton, CA) prior to each exposure.
Respiratory measurements.
The respiratory rates were determined from the peak of the inspiratory portion of the breathing cycle. The respiratory flow was sampled at a minimum frequency of 400 Hz. All calculated respiratory parameters were averaged for each 1-min portion of the measurement period. Data were collected for 10 min prior to the start of the exposure period to be used as the baseline for each animal. Butyl acetate exposed animals were divided into two groups for purposes of simulation: a four-animal group (Charles River, animals 1–4) with body weights higher than those of animals used for parameterizing the model, and a second group (Hilltop Lab, animals 5–8) with body weights in the range of those used for model parameterization. Another distinction between these groups was that the four-animal group had lower ventilation rates and was exposed to higher initial chamber concentrations (1625 ppm vs. 1400 ppm) than the two-animal group (averaged post-experiment).
Blood analysis methods.
Blood was placed in pre-weighed 4-ml screw cap vials containing 2 µg of phenyl methyl sulfonyl fluoride (PMSF) to inhibit any potential further metabolism of n-butylacetate. The vials were quickly sealed, weighed, and placed on dry ice. The blood was stored at –80°C until analysis. A weighed quantity (0.10 g) of blood was extracted with 200 µl of 0.9 M H2SO4, 0.25 g Na2SO4 (to improve extraction efficiencies), and 200 µl ethyl acetate containing known concentrations of internal (matrix) standards (hexyl alcohol and ethoxyacetic acid). Separation and quantitation were achieved using a 60 m x 0.32 mm id x 1.0 film thickness Stabilwax DA capillary column (Restek, Bellefonte, PA) and a Hewlett Packard 6890 GC with flame ionization detection (FID). Splitless injections of 2.0 µl of extract at an injector temperature of 210°C were employed. The initial oven temperature was 60°C for 1.5 min then increased at 15°C/min to 250°C. Hydrogen was used as the carrier gas at 25 psi for 11 min then increased at 50 psi/min to 35 psi. The flame ionization detector (FID) temperature was 275°C. For the preceding conditions, n-butyl acetate, n-butanol, hexyl alcohol, n-butyric acid, and ethoxyacetic acid had retention times of 5.4, 6.1, 8.6, 11.3, and 13.3 min, respectively.
Rat open chamber n-butyl acetate exposure and blood kinetics (Groth and Freundt, 1991).
The intratracheal inhalation experiment of Groth and Freundt (1991) provides estimates of the uptake of n-butyl acetate by the lungs, its metabolism to n-butyl alcohol, and the rate of clearance of n-butyl alcohol. This published study used anesthetized female Sprague-Dawley rats (290–340 g) exposed to air containing an average concentration of 970 ppm n-butyl acetate for 5 h. Arterial blood samples were taken periodically and analyzed for n-butyl acetate and n-butyl alcohol; data were digitized from graphs in the original manuscript, as reported previously (Barton et al., 2000). Since the exposure was by tracheostomy tube, no presystemic nasal metabolic clearance would occur; the concentration presented to the lungs was equal to the measured air concentration. These air concentrations were included in a table function within the modeling software (Advanced Continuous Simulation Language, Aegis, Inc., Huntsville, AL), and interpolated values were used as the instantaneous inhalation concentrations for simulating the 5-h exposure.
Inhalation route blood kinetics of n-butanol in humans (Astrand et al., 1976).
Two groups of six male subjects, average body weight of 78.3 kg and ages between 21 and 34 years, were exposed to n-butanol through a breathing valve and mouthpiece (Astrand et al., 1976). n-Butanol exposure concentrations were either 100 or 200 ppm. Exposures were conducted during rest and at various levels of exercise. Blood samples were taken through catheters placed into a brachial artery and a medial cubital vein (both in the arm). The bioavailability of n-butanol, as a fraction of pulmonary (minute ventilation) ventilation, was estimated by the authors as the difference between the amount given (concentration x ventilation rate) and the amount in expiratory air (total amount collected in the Douglas bag). Defined this way, bioavailability is the fraction of material in the inhaled volume of air reaching the blood exchange region in the lung. Respiratory bioavailability can be reported as fraction of either pulmonary (minute ventilation) or alveolar ventilation. Pulmonary ventilation rates were estimated from the volume of expired air collected in the Douglas bag.
The average bioavailability was 40% of pulmonary ventilation or 59% of alveolar ventilation assuming all absorption into systemic blood occurs in the alveolar region; alveolar ventilation rates are used in the model. Mean pulmonary ventilation rates, blood concentrations and fractional uptakes (inhaled-exhaled n-butanol over a 30 minute period) were reported for each exposure condition (see Supplemental Data, Table S1). Pulmonary ventilation rates were converted to alveolar ventilation rates for use in the PBPK model by assuming alveolar ventilation rates are 67% of pulmonary ventilation rates (Brown et al. 1997). Arterial blood concentrations in a single individual were also reported for a 90-min period following the end of exposure to 100 ppm n-butanol. Arterial and venous blood concentrations of n-butanol did not fall below 1.1 µM after the end of exposure, suggesting this is near background concentrations. This background concentration was subtracted from the reported blood concentrations; only exogenous n-butanol was simulated by the model.
Model structure.
The model structure and governing equations were as reported by Barton et al. (2000) with some revisions. The model was coded using ACSL (Advanced Continuous Simulation Language, Aegis, Inc., Huntsville, AL) and is available upon request. Briefly, the model is composed of four linked submodels, one each for n-butyl acetate, n-butanol, n-butyraldehyde, and n-butyric acid (Fig. 1). Each submodel is constructed of four or five compartments required for absorption, storage, metabolism, distribution, and clearance. The compartments include lung, arterial and venous blood, fat (n-butyl acetate only), liver, and the remaining perfused tissues (other perfused tissues, OPT) (Fig. 1). All compartments are described as well mixed and flow limited. The upper respiratory tract compartments described in Barton et al. (2000)—nose, trachea, conducting airways—were removed for the simulations reported here. To account for fractional bioavailability of water soluble volatiles such as n-butanol, the incoming air concentration (conc) was multiplied by a factor (fa) representing the fraction of chemical that would be available for absorption (Barton et al., 2000). The model was further revised to run in both closed and open chamber inhalation conditions used in the experiments reported by Poet et al. (2003a,b), and a provision was included to run simulations based on reported study-specific ventilation rates, coded in an ACSL table function. Michaelis-Menten metabolism of n-butyl acetate and n-butyric acid, and first order metabolism of n-butanol was added to the other perfused tissue compartment (Fig. 1), reflecting significant extrahepatic, extravascular metabolism implied by earlier simulations. Simple first order metabolism was assumed for n-butanol in the other perfused tissue compartment in the absence of sufficient data to implement a Michaelis-Menten approach. First order and Michaelis-Menten metabolism are formulated as described in Barton et al. (2000).
|
Development of the rat PBPK model.
The various model parameters were either obtained from the literature, or inferred by visual fitting to in vivo PK data. Values and sources for tissue volumes, blood flows, ventilation rates, partition coefficients, and some metabolic constants were reported previously (Barton et al., 2000
). The sources of new parameters are described here, with a complete list of parameters used in these simulations provided in Tables 1–3.
|
|
|
Intravenous blood time course data for each compound and its subsequent metabolite(s) (Deisinger and English, 2001
) were used to estimate metabolic constants (Vmax and Km or Kmet) for B-carboxylesterase (CE), alcohol dehydrodgenase (ADH), aldehyde dehydrogenase (ALDH), and medium chain acyl-CoA synthetase (mACS), the latter capable of, and presumed responsible for, clearance of n-butyric acid (Knights, 1998). Some general tissue distribution (e.g., the need for activity in tissues other than the liver) information for CE and ADH was also inferred from the blood data. The Km for CE metabolism of n-butyl acetate was assumed to be equivalent to that of ethyl acetate (Deisinger and English, 2003): for liver the Km was set to 1 nM, and for blood and other perfused tissues the value was set to 100 nM. This revision of the Km values initially estimated from fitting a more limited PK data set (Barton et al., 2000) was required to fit peak concentrations of n-butyl acetate measured in the more refined PK studies presented here. Other metabolic constants were fit (as indicated in Table 2) using blood time course data for each compound (Deisinger and English, 2001). Model simulations of blood concentrations were visually fit to the observed data. Butyraldehyde is not observed in blood; metabolic constants for butyraldehyde were fixed to simulate the appearance of n-butyric acid following n-butanol or n-butyl acetate dosing. Differential sensitivities of the distribution- and clearance-dominated phases of the blood concentration time-course to the various model parameters allowed determination of vascular, and extra-vascular metabolic constants for each compound. The definitive iv study was used to parameterize the model, and the pilot studies were used to test the model. The respiratory bioavailability study was not used to parameterize the model, rather, given fitted parameters estimated from the iv data, these data were used solely to estimate bioavailability.
The rate and extent of urinary elimination of n-butanol is unknown. Studies with 2-butanol found urinary clearance of up to 14% in different species, so a value of 30 (h)–1 was used for the n-butanol urinary clearance rate (kfilt2) resulting in about 10% urinary excretion (Dietz et al., 1981). Urinary excretion data were not available for n-butyric acid, so it was assumed to be less than 3.5%, as observed for isobutyric acid in the rat (DiVincenzo and Hamilton, 1979). Rapid metabolic clearance of n-butyl acetate is expected to limit urinary elimination. The value estimated for n-butyric acid was also used for n-butyl acetate.
Development of the human PBPK model.
Physiological parameters such as flow rates and tissue volumes were obtained from the literature and are reported in Table 1. The fractional availability of n-butanol (59% of alveolar ventilation) was used as reported by Astrand et al. (1976). Rate constants for metabolism and renal elimination were allometrically scaled from the rat values, and are reported in Tables 1 and 2. In the absence of better data, the rat partition coefficients were assumed to be a good proxy for human values, and used directly in the human model. The human model was validated only for the inhalation route for n-butanol.
Modeling approach.
Model development for the rat proceeded in two phases and concluded with verification of predictions of blood concentrations of n-butyl acetate, n-butanol, and n-butyric acid following inhalation exposures to n-butyl acetate and n-butanol as reported by Poet et al. (2003a,b). First, the model structure was evaluated and kinetic constants for metabolic clearance (and production) of each metabolic series member was estimated by fitting the model to the corresponding blood kinetics for each compound following iv bolus or iv infusion. The parameterization was verified through simulation of additional iv blood kinetics data. Following the establishment of a convincing parameterization for the systemic fate of the butyl series compounds (based on fitting the model to the iv blood kinetics data), the inhalation route was parameterized. Parameterization was achieved by estimating n-butyl acetate and n-butanol bioavailability, the final parameter, through simulation of closed chamber data. Validation of the inhalation route was conducted by comparison of simulated and observed blood concentrations of butyl series compounds following inhalation of n-butyl acetate and n-butanol. The results of these analyses are presented in the order: parameterization of systemic metabolism using iv studies, verification of systemic metabolism, estimation of bioavailability, and verification of model prediction of blood concentrations of butyl series compounds following inhalation exposure. A list of the experimental studies and their use in developing and validating the model are provided as supplemental data (see Supplemental Data, Table S2).
The human model was evaluated by simulating the blood n-butanol kinetics following n-butanol inhalation (Astrand et al., 1976) without changes in parameters. The human model was validated only for the inhalation route for n-butanol.
Sensitivity and uncertainty analysis.
The impact of uncertainty in the various model parameters on simulated arterial blood concentrations was evaluated semi-quantitatively though a combination of sensitivity analysis and a qualitative evaluation of parameter uncertainty. Each parameter was characterized as having low, medium, or high sensitivity and low, medium, or high uncertainty. The sensitivity of the arterial n-butanol concentrations to all model parameters was evaluated and those with sensitivity coefficients greater than 0.1 were also reported. The sensitivity coefficient (SC) is defined as the percent change in the dose metric for a 1% change in the listed parameter. A sensitivity coefficient of 1 indicates that there is a 1 to 1 relationship between the change in the parameter and the internal dose metric. Negative sensitivity coefficients indicate an inverse relationship between the model parameter and dose metric. Based on the sensitivity analysis, each parameter was characterized as having a low, medium or high impact on the selected dose metric using the following criteria:
- Low (L): SC between 0.1 and 0.15
- Medium (M): SC between 0.15 and 0.5
- High (H): SC greater than or equal to 0.5
- Medium (M): SC between 0.15 and 0.5
Parameters with SC less than 0.1 were not reported or considered further in this analysis.
The uncertainty in parameters with SC greater than 0.1 was evaluated qualitatively. The source of the parameter was reviewed and the parameters were characterized as having low, medium, or high uncertainty based on the following criteria:
- Low: Obtained from data in the correct species, coefficient of variation (CV) less than 0.5 or verified through successful use in published PBPK models
- Medium: Correct species, CV greater than 0.5, or scaled value from a different species with a high probability that scaling holds across species
- High: All others
- Medium: Correct species, CV greater than 0.5, or scaled value from a different species with a high probability that scaling holds across species
The sensitivity and uncertainty designations for each parameter were tabulated and used to identify parameters with sufficient uncertainty (uncertainty designation greater than Low) and impact on arterial n-butanol concentrations (sensitivity designation greater than Low) to have some influence on the dose-route extrapolation and risk assessment.
Sensitivity analyses for the rat and human PBPK models, measuring the impact of model parameters on blood n-butanol concentrations, were run for n-butyl acetate (rat only) and n-butanol concentrations appropriate for testing the model in the range it will be exercised for dose-route extrapolations supporting risk assessment. Sensitivity analyses were conducted for two steady state n-butyl acetate concentrations, 500 ppm and 3000 ppm, the NOAELs (not duration adjusted) for reduced weight gain and neurotoxicity, respectively, in rat 90-day subchronic studies (David et al., 1998, 2001). Sensitivity analyses for the rat and human for constant n-butanol exposures were conducted at 700 and 1000 ppm, respectively. These n-butanol concentrations approximate exposures resulting in blood n-butanol concentrations near to those observed for the rat at the NOAEL of 500 ppm n-butyl acetate.
Human equivalent concentrations.
The human n-butanol exposure concentration resulting in a weekly arterial blood n-butanol AUC (µM x h) equivalent to the weekly blood n-butanol AUC (µM x h) in the rat following NOAEL concentration exposure (6 h/day, 5 days/week) is termed the human equivalent concentration (HEC). The HEC was determined for two rat NOAELs, 500 ppm and 3000 ppm, the NOAELs for reduced weight gain and neurotoxicity, respectively, in rat 90-day subchronic studies (David et al., 1998, 2001). The HEC was determined by first using the rat model to determine the a weekly arterial blood n-butanol AUC associated with NOAEL concentration exposures and then using the human model to determine the human n-butanol exposure concentration leading to the same weekly AUC under continuous exposure conditions. The rat and human ventilation rates used in these simulations are found in Table 1.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTSDISCUSSIONAPPENDIX 1APPENDIX 2SUPPLEMENTAL DATAREFERENCES |
|---|
iv Kinetics of the Butyl Series Compounds in Rats
Experimental results (Deisinger and English, 2001).
The background concentrations of the butyl series analytes in rat blood were determined over the course of a day at hourly intervals from three non-dosed, dual cannulated, male SD rats. Blood samples were collected from the jugular vein cannula and were analyzed by GC-MS. Background n-butyric acid concentrations were near or below the limit of detection (0.61 µM), and we chose to represent background levels in figures with a value of 1 µM. Trace amounts of n-butanol (less than 3 µM), below the limit of quantitation (2.15 µM), were detected in two of 21 rat blood samples assayed. N-butyl acetate (limit of detection, 1.43 µM) and n-butyraldehyde (limit of detection, 7.0 µM) were not detected in these samples.
In vivo hydrolysis of the n-butyl acetate ester was very rapid, with a half-life measured in seconds, rather than minutes or hours. n-Butanol, a product of this hydrolysis, has a somewhat slower elimination rate, but is present only in trace amounts in the blood by 30 min post dosing with either n-butyl acetate or n-butanol. The terminal butyl series oxidation product, n-butyric acid, appears to have a half-life of less than 1 min when it is administered in vivo, but its elimination following administration of n-butyl acetate or n-butanol is extended by the concurrent production of the compound from its precursors.
Simulation of iv route blood kinetics.
Parameterization of metabolic constants for downstream metabolites of n-butyl acetate is complicated by blood concentration time course dependence on the metabolism of the upstream metabolite, which controls the production rate. Therefore, metabolic constants were fitted first for n-butyric acid, followed by n-butanol and n-butyl acetate. Blood concentration time course data for metabolites (n-butanol, n-butyric acid) following administration of a parent compound (n-butyl acetate, n-butanol) were also utilized for parameter estimation. All data are from Deisinger and English (2001).
Following iv bolus and iv infusion administration of n-butyric acid, modeled blood concentrations of n-butyric acid were in close agreement with observed values, falling within or just outside the 95% confidence limits (Fig. 2). Because the model does not consider endogenous sources of n-butyric acid, the simulation line drops below observed concentrations after blood concentrations decline to background concentrations at 9 min (Fig. 2). The 10 and 15 min data points in the bolus study were close to the background concentrations and assumed not to reflect a terminal elimination phase different than the suggested by the iv infusion data.
|
, five animals, average body weight of 0.3 kg) or i.v. infusion (0.281 mmoles/kg, •, one animal, body weight of 0.33 kg) of n-butyric acid. Close agreement between simulated and observed data implies proper characterization of distribution and metabolic clearance of n-butyric acid. Horizontal line (.......) represents background concentrations. Error bars represent ± 2 SE, equivalent to the 95% confidence limits on the mean value.Blood kinetics following iv administration of n-butanol were also well described by the model, with simulations falling within the 95% confidence limits for most of the time course. Peak concentrations, as well as the distribution- and clearance-dominated phases of the n-butanol blood concentration time course, were in close agreement with observed values (Fig. 3A). Blood concentrations of n-butyric acid were also well described by the model, but peak concentration was modestly under-predicted (Fig. 3B).
|
The time (Tmax) and magnitude (Cmax) of the peak blood concentration for n-butyl acetate were sensitive to blood Vmax/Km and the iv delivery time (2–3 s), while the initial rapid clearance phase was sensitive to blood Vmax/Km, and the slower terminal elimination phase was most sensitive to extra-vascular metabolism. Metabolic clearance of n-butyl acetate is rapid following iv administration (Fig. 4). Sampling times were long relative to Tmax, but modeled and observed blood concentrations for the distribution/elimination phase were in reasonable agreement (Fig. 4A). As with n-butyric acid, significant extrahepatic metabolism of the n-butyl acetate was indicated, and in addition to vascular metabolism, it was necessary to revise the model, attributing extravascular metabolism to the other perfused tissue compartment. Modeled and observed concentrations of n-butanol and n-butyric acid were consistent with observed values following iv administration of n-butyl acetate (Figs. 4B, 4C), although peak concentrations of n-butyric acid were under-predicted by the model.
|
Validation of systemic metabolism.
Parameterization of the model was verified by simulation of the blood kinetics of n-butyl acetate, n-butanol, and n-butyric acid following iv administration of n-butyl acetate to a single animal in a pilot study (Deisinger and English, 2001
). The agreement between modeled and observed blood concentrations of these compounds provides confirmation that their systemic distribution and metabolic clearance are adequately described by the model (Fig. 5).
|
Inhalation Kinetics of the Butyl Series Compounds in Rats
Experimental data.
Average measured minute ventilation rates during inhalation exposure to
2000 ppm n-butyl acetate and n-butanol (Poet et al., 2003a,b) are shown in Figures 6A and 6B. Ventilation rates for animals or groups of animals in the simulations presented here are provided as supplemental data (see Supplemental Data, Ventilation Rates). During both n-butyl acetate and n-butanol exposure, there was an initial respiratory depression followed by an increase in minute ventilation rates over the length of the exposure period. Loss of n-butyl acetate in the empty apparatus with (Fig. 7A, dead rat) and without (data not shown) a nonrespiring rat, was the same. Loss attributable to the chamber apparatus was large relative to that attributable to respiratory uptake (Fig. 7A). Nonetheless, respiratory uptake of n-butyl acetate was both quantifiable and sufficient for measuring respiratory bioavailability. This was also true for n-butanol (Fig. 7B, dead rat data not shown).
|
|
The n-butyl acetate peak concentration in the blood appeared to occur near 20 min, though we would expect peak concentrations to occur earlier for this rapidly metabolized compound, and before chamber concentrations had fallen significantly. Blood n-butanol concentrations peaked 20 min after the start of exposure (Figs. 11 and 12). Butyric acid was not detected in any sample. Blood n-butyl acetate and n-butanol concentrations were greater than the lowest concentration in the standard curve, which was the effective limit of reliable quantitation. These values were 3.7 µM and 9.4 µM for n-butyl acetate and n-butanol, respectively. The detection limit for n-butyric acid was 12.5 µM in the n-butyl acetate experiments. During the n-butanol exposures, peak concentrations of n-butanol (
0.1 µM) were observed approximately 20 min after exposure (Fig. 9). N-butyric acid concentrations were near the limit of quantitation (two times the limit of detection, which was lower in the n-butanol experiments than in n-butyl acetate experiments), reaching peak concentrations near 0.007–0.009 µM after 30–45 min.
|
|
), Rat n-butanol AUC; (x), Human n-butanol AUC. The equations of lines describing the exposure to arterial n-butanol AUC relationship are shown.
|
Inhalation-route bioavailability of n-butyl acetate.
The rate of n-butyl acetate chamber only loss was biphasic (Fig. 7A). Two first order rate constants were fitted to the chamber only (or chamber plus dead rat) data, one describing loss during the first phase and one describing loss during the second phase. The switch between these two rate constants was initiated at the time the rates appeared to change (Fig. 7A). The fitted rate constants described both chamber-only loss and chamber and non-respiratory rat loss (i.e., dead rat) (Fig. 7A), indicating that non-respiratory rat-related losses were minimal.
The rate of n-butyl acetate loss attributable to rat respiratory uptake was well described using the alveolar ventilation rates (estimated from Poet et al., 2003a) and a respiratory bioavailability of 100% (fa = 1.0, Fig. 7A). Defined this way, bioavailability is the fraction of material in the inhaled volume (alveolar in this case) of air reaching the blood exchange region in the lung. For all simulations using measured ventilation rates, the cardiac output was set equal to the ventilation rate.
Inhalation-route bioavailability of n-butanol.
The rate of n-butanol chamber-only loss was biphasic (Fig. 7B). Two rate constants were fitted to the chamber-only data, one describing loss during the first phase and one describing loss during the second phase. The fitted rate constants described both chamber only loss and chamber and non-respiratory rat loss (i.e., dead rat) (data not shown), indicating that non-respiratory, rat-related losses were minimal.
The rate of n-butanol loss attributable to rat respiratory uptake was well described using the measured ventilation rates (adjusted to alveolar ventilation rates; Poet et al., 2003b) and an average respiratory bioavailability of 50% (FA = 0.5, Fig. 7B). Some variability was apparent: respiratory bioavailability was estimated to be 40% for two rats and 60% for two others.
Validation of Model Predicted Inhalation-Route Blood Kinetics
Blood kinetics following inhalation of n-butyl acetate.
Model predicted blood kinetics of n-butyl acetate were in excellent agreement with the observed blood concentration time-course (Fig. 8A). Predictions of blood n-butanol concentrations were also in relatively good agreement with observed concentrations, though the Tmax was early and Cmax was under predicted by approximately 50%. Simulation of a separate (from the same laboratory, same protocol, Poet et al., 2003a) four animal data set showed a poorer fit to blood n-butyl acetate concentrations, with simulations falling outside the 95% confidence limits for all but two points (Fig. 8B). These animals were obtained from a different vendor (see Methods). Model predicted n-butanol blood concentrations were in close agreement with observed data for the first 30 min, thereafter over predicting blood concentrations 50–100%.
|
The difference between the fits for the first and second data sets is likely attributable to experimental variability or observed differences in the two groups of animals. Given two of the known differences between the experiments, ventilation rates somewhat lower and initial chamber concentrations of n-butyl acetate slightly higher (1625 ppm vs. 1400 ppm) in the four animal group (Fig. 8B), the model prediction of blood n-butyl acetate concentrations similar to those in the two animal group (Fig. 8A) is expected because the dose rates would be similar. The experimental data in the four animal group, in contrast, shows n-butyl acetate concentrations that are roughly half of those in the two animal group. Such large differences in experimentally measured blood concentrations would not be expected, and are not attributable to known differences analytical techniques or handling of the blood samples. Differences in body weights alone between these groups are also not likely to be responsible for large differences in measured blood concentrations. Although it cannot be ruled out, a
15% increase in body weight is unlikely to result in increases in metabolism necessary to reduced blood concentrations by a factor of two. In addition, to the degree principles of allometry hold, the model accounts for body weight related differences in metabolism. However, the difference could be related differences in the physiology of this second group of animals since they appeared to be lethargic.
Blood kinetics following inhalation of n-butanol in rats.
Without adjustment, the butyl series model predicted n-butanol concentrations that were modestly (40%) less than those observed experimentally (Fig. 9) following inhalation of n-butanol. The simulated and observed Tmaxs were similar, but the former was slightly earlier than the latter. The general shape of the blood n-butanol concentration time-course predicted by the model approximated the observed time-course fairly well. N-butyric acid concentrations were modestly over-predicted, although concentrations were generally low, on the order of the limit of quantitation, for the majority of the study period.
Blood kinetics following intratracheal inhalation of n-butyl acetate in rats.
Blood kinetics of n-butyl acetate and n-butanol are strongly ventilation-rate dependent, but Groth and Fruendt (1991) did not report ventilation rates. Consequently, this data set is of limited usefulness for verifying model predictions. Nonetheless, evaluation of the model for general consistency with this data set seemed prudent.
In the absence of measured ventilation rates, simulations were conducted at ventilation rates that varied between 7 and 10 l/h/bw0.75. These ventilation rates were chosen to result in predicted blood n-butyl acetate concentrations in the observed range while remaining in the range of biologically plausible values (Fig. 10A). The presumption here is that n-butyl acetate metabolism is well characterized relative to ventilation rates for this study, and fitting should be accomplished by changing ventilation rates. The resulting ventilation rates were low, as would be expected for anesthetized, intubated rats. Predicted blood n-butanol concentrations were similar to those observed experimentally, but tended towards modest under-prediction (Fig. 10B). Despite the major uncertainties in simulating these data (unknown ventilation rates, experimental details poorly described), the model is reasonably consistent with these data.
|
Inhalation route blood kinetics of n-butanol in humans.
Model-predicted arterial n-butanol concentrations in male humans following inhalation of 100 ppm n-butanol at various exercise levels (Astrand, 1976
, Series II) agreed well with observed concentrations (Fig. 11A). Simulations were within 2 SE of the data during exposure. Moreover, the simulated post exposure arterial blood n-butanol time course was not inconsistent with data collected from a single individual (Fig. 11A, post exposure data points). Model simulated arterial blood n-butanol concentrations were approximately 50% higher than observed values at steady state, but generally within 2 SE of the mean values (95% confidence limits on the mean; Fig. 11B) following exposure to 200 ppm n-butanol (Astrand, 1976, Series I).
Sensitivity and Uncertainty Analysis
The results of the sensitivity and uncertainty analyses for rat and human PBPK models are presented in Tables 4 and 5, respectively. Analyses supporting the uncertainty designations are presented in Appendix 1. Sensitivity coefficients were determined for all model parameters. Only parameters with sensitivity coefficients greater than 0.1 are reported and evaluated for uncertainty. Parameters with sensitivity designations of H or M and uncertainty designations of H or M are assumed to have sufficient potential to add uncertainty to model simulated arterial n-butanol concentrations to warrant further discussion.
|
|
For n-butyl acetate exposures in the rat, the key parameters are ventilation rate, other perfused tissues ADH activity, urinary n-butanol elimination rate, and hepatic ADH Vmax (Table 4). The sensitivity coefficients for these parameters following exposure to 3000 ppm n-butyl acetate were 1.1, –0.54, –0.18, and –0.2, respectively. Modeled arterial n-butanol concentrations were only sensitive to hepatic ADH Vmax at high (3000 ppm) n-butyl acetate exposures. Hepatic elimination was blood flow limited at lower concentrations. Although the other parameters (ventilation rate, other perfused tissue ADH activity and urinary n-butanol elimination rate) have uncertainty designations of H or M, they have been used successfully to simulate a large pool of rat toxicokinetic data and are therefore expected to be characterized fairly well.
Only ventilation rate, with a sensitivity coefficient of 1.0, had sufficient sensitivity (H) and uncertainty (M) to influence uncertainty in arterial blood n-butanol in rats following inhalation exposure to n-butanol. Human blood n-butanol concentrations were not sensitive to any parameters with uncertainty designations of M or H (Table 5). It should be pointed out, however, that ventilation rates vary considerably with activity level. Nonetheless, the predictions of arterial n-butanol concentrations during inhalation exposure to n-butanol made with the human PBPK model are generally insensitive to parameters with high or medium levels of uncertainty. Rat blood:air partition coefficients were used as a proxy for human values, though, for other volatile organics, human blood:air partition coefficients are approximately 1.5–2 times lower. Because modeled blood concentrations are insensitive to blood:air partition coefficients (SC < 0.01), this choice has no impact on the results presented here.
Rat and Human Tissue Dosimetry
The rat and human models were used to simulate steady state arterial blood concentrations of butyl series compounds following n-butyl acetate exposure (rats only) and n-butanol exposure. The results are presented in Tables 6 (rat), and 7 (human). Arterial blood concentrations of n-butyl acetate, n-butanol, and n-butyric acid at the rat NOAEL of 500 ppm n-butyl acetate were 20.4 µM, 37 µM, and 40.1 µM, respectively. The corresponding values for the 3000 ppm rat n-butyl acetate NOAEL are 122 µM, 235 µM, and 59 µM.
TABLE 7 Modeled Steady State Human Arterial Blood Concentrations During Inhalation Exposure to n-Butanol
| |||||||||||||||||||||||||||||||||||||||||||||||||||
|
The use of blood AUC, rather than blood concentrations, as the relevant dose metric for use in calculating HECs has the benefit of direct incorporation of an adjustment for exposure duration. Arterial blood n-butanol AUCs (µM x h, for a week of exposure) were calculated in rats following inhalation exposure to a range of n-butyl acetate and n-butanol exposure concentrations and in humans for a range of n-butanol exposure concentrations (Fig. 12). Rat AUCs were calculated for a study week: 6 h/day, 5 days exposure per week. Human AUCs were calculated assuming constant exposure for a 7-day week. Average weekly AUCs for both rats and humans are the same when calculated for longer exposure durations.
Human Equivalent Concentrations
The HECs for rat NOAELs of 500 ppm (reduced weight gain) and 3000 ppm (neurotoxicity) n-butyl acetate are 169 ppm and 1066 ppm, respectively. While the PBPK models can be used in the future to derive HECs for other rodent n-butyl acetate or n-butanol exposures, an alternative approach was developed that allows non-PBPK model users to derive HECs. Equations describing the relationship between n-butyl acetate (rat only) or n-butanol (human and rat) exposure (Fig. 12) and weekly arterial blood n-butanol AUC were combined and solved for the relationship between rat external exposures and human exposures producing the same weekly arterial blood n-butanol AUC (the HEC). The rats AUCs reflect the 6 h/day, 5 days/week exposure protocol of the toxicity study and the human AUCs are calculated for continuous exposures, thus accounting for the required exposure duration adjustment. The derivation is presented in Appendix 2. These algebraic equations are equivalent to using the PBPK model, and like the model reflect our best understanding of the kinetics of these compounds. The equations can be used with highest confidence in the range of the rat (2000 ppm n-butyl acetate/butanol) and human (100–200 ppm n-butanol) inhalation exposures used to parameterize the model. In light of the consistency in the kinetics of these compounds at high and low iv doses and the scalability of the model from rat to human, we might expect the approach to be valid across a wider dose range. Given a rat n-butyl acetate NOAEL, the HEC may be calculated using the formula:
 |
 |
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTSDISCUSSIONAPPENDIX 1APPENDIX 2SUPPLEMENTAL DATAREFERENCES |
|---|
Toxicity is expected to be a more direct function of target tissue exposure than external exposure. Estimating risks associated with human exposure to volatile organic compounds by regression of the exposure-response relationship has specific limitations. Awareness of these limitations has promoted the development of dosimetry based approaches to risk assessment.
In the case of the butyl series compounds, the exposure route of concern for humans is inhalation. Such exposure is expected to result in systemic exposures to both the inhaled compound and its metabolites, though systemic exposure to n-butyraldehyde is believed to be very low because it is not experimentally measurable in the blood. The model addresses this by assuming rapid and almost complete intra-hepatic metabolism of n-butyraldehyde, which appears consistent with the rapid appearance of n-butyric acid in the blood. Inhalation-route subchronic toxicity studies have been conducted for n-butyl acetate in rats (David et al., 1998, 2001), providing toxicity data attributable to n-butyl acetate, n-butanol, and/or n-butyric acid exposure. The dosimetry-based approach proposed for interpreting the toxicity studies and extrapolating the results to humans involved the development of a PBPK model. Here we have developed and validated a rat PBPK model for inhalation pharmacokinetics of the butyl series compounds n-butyl acetate and n-butanol. The resulting rat model can be used directly to derive RfCs by determining the NOEL or benchmark dose for one butyl series member given toxicity data from a different family member. Standard default procedures can be applied to arrive at the RfC, or the human model can be utilized to conduct the species extrapolation.
Particularly important to the validation of the PBPK model for the inhalation route was a novel combination of experimental approaches—typically used independently—to simultaneously measure ventilation rates, closed chamber loss (respiratory uptake), and blood concentrations of the inhaled volatile and its metabolites (Poet et al., 2003a