Rat Tissue and Blood Partition Coefficients for n-Alkanes (C 8 to C 12 )

Abstract

Rat tissue:air and blood:air partition coefficients (PCs) for octane, nonane, decane, undecane, and dodecane (n-C₈ to n-C₁₂ n-alkanes) were determined by vial equilibration. The blood:air PC values for n-C₈ to n-C₁₂ were 3.1, 5.8, 8.1, 20.4, and 24.6, respectively. The lipid solubility of n-alkanes increases with carbon length, suggesting that lipid solubility is an important determinant in describing n-alkane blood:air PC values. The muscle:blood, liver: blood, brain:blood, and fat:blood PC values were octane (1.0, 1.9, 1.4, and 247), nonane (0.8, 1.9, 3.8, and 274), decane (0.9, 2.0, 4.8, and 328), undecane (0.7, 1.5, 1.7, and 529), and dodecane (1.2, 1.9, 19.8, and 671), respectively. The tissue:blood PC values were greatest in fat and the least in muscle. The brain:air PC value for undecane was inconsistent with other n-alkane values. Using the measured partition coefficient values of these n-alkanes, linear regression was used to predict tissue (except brain) and blood:air partition coefficient values for larger n-alkanes, tridecane, tetradecane, pentadecane, hexadecane, and heptadecane (n-C₁₃ to n-C₁₇).Good agreement between measured and predicted tissue:air and blood:air partition coefficient values for n-C₈ to n-C₁₂ offer confidence in the partition coefficient predictions for longer chain n-alkanes.

Physiologically based pharmacokinetic (PBPK) models are increasing in use for describing the dosimetry of chemicals (Payne and Kenny 2002) and estimating health risks. The ability of PBPK models to describe the fate of a chemical in the body is, in part, dependent on the solubility of a chemical in biological tissues. The ratio of a chemical concentration in two phases at equilibrium for a given temperature is referred to as a partition coefficient (PC). For chemicals that are volatile and amenable to experimental determinations such as headspace analysis, this means conducting experimental studies and calculating blood:air and tissue:air PC values (Gargas et al. 1989). For implementation in PBPK models to describe the fate of a volatile chemical in tissues, the tissue:air PC value is divided by the blood:air PC value to yield the tissue:blood PC value. Gas or vapor exchange at the lung is simply described in a PBPK model using a blood:air PC value. Recently the use of algorithms to estimate PC values for use in PBPK modeling has received attention because of the lack of adequate experimentally determined PC values (Basak et al. 2002; Fiserova-Bergerova and Diaz 1986; Meulenberg, Wijnker, and Vijverberg 2003; Meulenberg and Vijverberg 2000; Poulin and Krishnan 1996). Thrall et al. (2002) estimated that 60% of all PBPK models have used PC values reported by Gargas et al. (1989).

Normal alkanes (n-alkanes) represent an important class of hydrocarbons found in fuels derived from crude oil such as gasoline (n-C₄ to n-C₁₂), kerosene (n-C₆ to n-C₁₆), JP-4 (n-C₅-n-C₁₄), and JP-8 (n-C₇ to n-C₁₈) (Potter and Simmons 1998). Previous investigators have measured PC values for some of these n-alkanes. Gargas et al. (1989) reported experimentally determined rat tissue:air PC values in muscle, fat, liver and blood:air PC values for hexane (n-C₆) and heptane (n-C₇). Liu et al. (1994) reported experimentally determined in vivo rat blood:air PC values for ethane (n-C₂), butane (n-C₄), hexane, octane, and decane. Most recently, Robinson (2000) reported experimentally determined rat tissue:blood PC values for nonane in liver, fat, muscle, and brain and a blood:air PC value for use in a PBPK model for nonane.

In this paper we report experimentally determined rat tissue:air and blood:air PC values for octane, nonane, decane, undecane, and dodecane using slightly modified vial equilibrium headspace techniques (Gargas et al. 1989). Also we report predicted rat tissue:air and blood:air PC values for tridecane, tetradecane, pentadecane, hexadecane, and heptadecane. These data will then be used to support the development of a PBPK model for jet fuel (JP-8).

METHODS AND MATERIALS

Chemicals

Octane, nonane, decane, undecane, and dodecane (CAS numbers 111-65-9, 111-84-2, 124-18-5, 1120-21-4, and 112-40-3, respectively) were 99%+ pure (Sigma-Aldrich, St. Louis, MO). JP-8 was kindly supplied by Dr. David Mattie, Air Force Research Laboratory, Wright- Patterson AFB, OH.

Animals

Six male Sprague-Dawley rats were obtained from Charles River Breeding Laboratories (Wilmington, MA). Animals were housed two per “shoe-box” style cage. Litter was changed weekly and animals had unrestricted access to PMI 5001 rodent chow (PMI feeds, St. Louis, MO) and water. Animals were kept in a humidity/climate-controlled facility with a 12-h light/dark cycle for a minimum of 14 days prior to euthanasia. Male rats (300 to 500 g) were killed by CO₂ asphyxiation. Heparinized blood was collected via portal vein. Tissues (liver, perirenal fat, thigh muscle, whole brain) and blood were removed and pooled by tissue group in vials. Vials were stored at −20°C or in a few cases used immediately.

Method Development

A significant effort was undertaken to carefully develop an acceptable technique for using the vial-equilibration method with semivolatile n-alkanes, especially for n-C₁₀ and higher, which have vapor pressures near or below 1.0 mm Hg. Frozen instead of fresh tissues and blood were used for PC determinations to help facilitate method development for the semivolatile n-alkanes and to reduce the number of animals killed for the study. The laboratory methods were time-consuming and only one tissue group could be processed each day for each chemical. Preliminary studies were completed indicating modest differences in PC calculations between fresh and frozen tissues and blood. The liver:air and muscle:air PC values were 4% to 13% greater using fresh tissues for octane and dodecane compared to frozen tissues and the blood:air PC values were both 29% greater using fresh blood for octane and dodecane compared to frozen blood. To test the assumption of concentration independence for experimentally determined PC values for n-alkanes, two tedlar bag concentrations were prepared for octane and decane as described below. PC values were evaluated using freshly thawed tissue (muscle and fat) and blood and tedlar bags containing 3009 and 1505 ppm octane or 2980 and 1490 ppm decane. Blood, muscle, and fat PC values for octane were 15% to 17% greater at the lower concentration and for decane the PC values were 11% to 19% lower. The differences in PC values found between the two concentrations of octane or decane were slightly greater than the variability reported for our analytical method (see Results). PC values determined in this study were assumed to be independent of the vapor concentration.

Preliminary studies were also undertaken to ensure that the times to equilibrium were sufficient. Headspace samples were collected from 2 to 4.5 h for blood, liver, and muscle samples and 3 to 7 h for brain and fat samples. Preliminary results indicated that octane, nonane, and decane reached equilibrium by 2 h. Three hours was selected as the experimental time to equilibrium for all tissues using octane, nonane, and decane. For undecane and dodecane a 3-h time to equilibrium was used for blood, liver, and muscle. However, the time to equilibrium for brain and fat required 4 and 6 h, respectively.

Only one tissue (or blood) was analyzed each day per chemical because the semivolatile n-alkanes stick to glass. To ensure the quality of the data, 12 1- or 2.5-ml gas-tight glass syringes were used each day. Seven syringes were used for the sample vials and five syringes for the reference vials. Syringes were cleaned by filling and expressing the syringe three times with hexanes, followed by acetone. The syringe plunger was then removed and the syringe and the syringe barrel blown dry with nitrogen. The next day the cleaned gas tight glass syringes were checked for residual n-alkane by filling the syringe with 0.5 ml of room air and then injecting the sample into the gas chromatography (GC) apparatus. The technique was found to be suitable for all n-alkanes used in this study.

The GC calibration curves for the n-alkanes were produced to include the experimental range of concentrations found in the sample vials. The calibration curves were linear with R ² values for all n-alkanes exceeding .99.

Partition Coefficients

Glass scintillation vials containing frozen pooled tissues and blood from rats were allowed to thaw in a water bath at room temperature for 30 min. Enough tissue or blood was thawed to provide seven duplicate sample vials per day for each chemical tested. Tissues were minced and then smeared along the inside wall of preweighed 10-ml round-bottomed headspace vials (Kimble Glass, Vineland, NJ) with a stainless steel spatula (instead of preparing saline tissue homogenates). Tissue masses used for all experiments were 1.0 g of muscle, 0.5 g of liver, 0.1 g of brain, and 0.05 g of perirenal fat, with the exception of brain and fat for undecane and dodecane. For undecane and dodecane 0.05 g of brain and 0.025 g of fat were used. This smearing technique was used previously for a solvent (Gearhart et al. 1999) and was found suitable for this work. Whole blood (0.75 ml) was pippetted into preweighed vials. Five empty vials were used as reference vials. Vials were crimp-capped with aluminum caps containing teflonlined butyl rubber septa (National Scientific, Duluth, GA). Reference and sample vials were then placed on a vortex evaporator (Labconco, Kansas City, MO) and heated for 15 min at 37°C. Each vial was then vented to room air with a gas tight syringe (Hamilton, Reno, NV) and 26-gauge side port needle (without the plunger assembly). One milliliter of air was removed from each sealed reference and sample vial and replaced with 1 ml of air from the tedlar vapor sampling bag containing the n-alkane.

Tedlar vapor sampling bags for n-alkanes were created as follows. A 3-L tedlar gas-sampling bag (SKC, Eighty Four, PA) was filled to about 80% capacity with filtered room air. With a gas-tight syringe fitted with a side port needle, neat liquid n-alkane was injected into the sampling bag. Each bag was then uniformly and gently warmed with a heat gun to assist evaporation. The tedlar bag concentrations for octane, nonane, decane, undecane, and dodecane were 3009, 3001, 2980, 474, and 570 ppm, respectively. The stability of each n-alkane concentration in the sample bag was checked by measuring the sampling bag concentration immediately before and after loading the vials. Equilibration of the n-alkanes in the tedlar bags was verified by repeated sampling. The low vapor pressures of undecane and do-decane limited the vapor concentration that could be prepared in the tedlar bags. Also, these chemicals are very lipophilic, thus the fat and brain headspace concentrations in the vials at equilibrium were low, but within the calibration curve for our analytical method.

Vials were allowed to incubate at 37°C in a vortex evaporator (Labconco) with moderate shaking until equilibrium was reached. After the incubation period, 0.5 ml of headspace vapor from each reference and sample vials were collected by a gas-tight syringe for injection into a gas chromatograph. The sample and reference vials remained in the heated vortex evaporator block for up to 2 h after the reported time to equilibrium (see Table 1) awaiting manual sampling of headspace in a similar fashion to the technique used by Gargas et al. (1989).

Vial Analysis

Scheduling conflicts required the use of two GC systems in our laboratory. A Hewlett Packard (HP) 5890 Series II GC system was used for nonane and an Agilent 6890 Series II was used for all other n-alkanes. The flame ionization detector responses to the hydrocarbons were similar for both GC systems and reference vials were always used for each GC. The HP 5890 was equipped with a HP-5 column (10 m × 0.53 mm × 2.65 μm). The injector, flame ionization, and oven temperatures were 230°C, 270°C, and 140°C, respectively. The helium, hydrogen, and air flows were 2.3, 23, and 210 ml/min, respectively, with a split of 2.2 ml/min. The Agilent 6890 was equipped with a HP-5 column (15 m × .53 mm × .0015 mm). The retention time for nonane on the column was about 3 min. The injector and flame-ionized detector (FID) temperatures were 200°C and 260°C, respectively. Oven temperatures were 110°C (octane), 120°C (nonane), 130°C (decane), 140°C (undecane) and 160°C (dodecane). The column retention times for octane, nonane, decane, undecane, and dodecane were 3.00, 3.12, 3.66, 4.20, and 3.67 min. The nitrogen, hydrogen, and air flows were 40, 37.5 and 375 ml/min.

JP-8 Analysis

Aerosol and vapor samples were collected from a 12-port nose-only exposure system (In-Tox, Moriarity, NM) at the University of Arizona. The system was operated as previously described (Hays et al. 1995). Aerosol and vapor samples were collected from a sampling port attached at one of the central nose-cone ports with a sampling train consisting of a 47-mm glass fiber filter (GFF; SKC, Eighty Four, PA) housed in a stainless steel filter holder (Gelman Sciences, Ann Arbor, MI) followed by a coconut shell charcoal tube (100/50; SKC). The sample was collected with an AirChek 2000 (SKC) at 100 ml/min. The sampling apparatus was calibrated before each exposure with a Gilibrator 2 (Sensidyne, Clearwater, FL) and checked after sample collection. Samples were collected over a 60-min period or run time. For the 60-min run times, the nebulizer cup and JP-8 were changed every 15 min. Samples were desorbed 1 h in 1 ml of chloroform (ECD tested; Fisher Scientific, Suwanee, GA) for the charcoal tube and 5 ml of chloroform for the glass fiber filter.

The neat JP-8 sample and the aerosol and vapor samples were analyzed with an Agilent 6890N GC system equipped with a 5973N mass spectrometer operated in full scan mode. Separation was performed on a Supelco Petrocol DH 150 column (150 m × 0.25 mm × 1 μm; Bellefonte, PA) and an oven temperature ramp from 90°C to 230°C. The inlet temperature was 230°C and the detector was set at 300°C. The total concentration of JP-8 in the aerosol and vapor phase was estimated by summing the area of the total ions (45 to 260 m/z) collected from 40 to 280 min. Authentic standards of n-alkanes were used to verify the n-alkane identity in the jet fuel.

Partition Coefficient Calculation

PCs were calculated according to Equation 1 found in Gargas et al. (1989). $$P_{\text{i}}=\frac{C_{\text{ref}}(V_{\text{vial}})-C_{\text{i}}(V_{\text{via}}-V_{\text{i}})}{C_{\text{i}}{V}_{\text{i}}}$$ where P _i = partition coefficient, C _ref = concentration of n-alkane vapor contained in headspace of reference vial, V _vial = volume of reference vial (10 ml headspace vial), C _i = concentration of n-alkane in headspace of test vial, and V _i = volume of tissue/blood in test vial.

Estimation of Tissue:Air and Blood:Air Partition Coefficients for n-C₁₃ to n-C₁₇

Water solubility (WS) and log octanol:water partition coefficient (log P) were calculated for octane, nonane, decane, undecane, and dodecane using SPARC, an on-line computational program at http://ibmlc2.chem.uga.edu/sparc/smiles/Smiles.cfm. SPARC (Hilal, Karickhoff, and Carreira 2004) was developed jointly by the U.S. Environmental Protection Agency (EPA) and the University of Georgia. The SPARC predicted log P and the reciprocal of WS (1/WS) for the n-alkanes were each fit to the mean measured n-alkane partition coefficient values for each tissue using simple linear regression as implemented in PROC GLM of SAS V8.2 (SAS Institute, Cary, NC) to predict the C ₁₃ to C ₁₇ n-alkane PC values. Only results using 1/WS are presented in Results because of the superior fit.

RESULTS

The rationale for selecting n-alkane hydrocarbons as a class of hydrocarbons to determine PC values for potential use in a PBPK model for JP-8 is provided in Figure 1. These hydrocarbons are prominent components in JP-8 jet fuel (Figure 1A ) and in an inhalation chamber containing aerosolized JP-8. For the last decade several inhalation toxicology studies have been undertaken with aerosolized JP-8 at the University of Arizona (Dr. Mark Witten) (Harris et al. 1997a, 1997b, 1997c, 2000; Drake et al. 2003; Robledo and Witten 1999; Pfaff et al. 1995, 1996). We collected atmospheric samples of to identify the n-alkane distribution in the aerosol and vapor phases. A relatively high concentration of aerosolized JP-8 (1984 mg/m³) yielded 107 mg/m³ in the aerosol phase and 1877 mg/m³ in the vapor phase. The aerosol phase contained identifiable n-alkanes of n-C₁₁ to n-C₁₇ (Figure 1B ) and the vapor phase, n-C₈ to n-C₁₅ (Figure 1C ).

Partition Coefficients

Mean (± SE) blood:air and tissue:air partition coefficient values for octane, nonane, decane, undecane and dodecane are listed in Table 1. These n-alkanes were the least soluble in muscle and blood and the most soluble in fat. Generally speaking, the solubility of these n-alkanes in muscle, blood, liver, brain, and fat increased as the number of carbon atoms increased from 8 to 12. There was a notable exception with brain (Table 1). The mean brain:air PC value for undecane (35.3) was similar to decane (38.7) and the mean dodecane brain:air PC value was much greater than both decane and undecane brain:air PC values. The reason for this finding is unknown. Experiments were repeated to verify the initial mean brain:air PC value for undecane and dodecane, with similar results.

Estimation of partition coefficient values (blood:air, fat:air, liver:air, and muscle:air) for larger n-alkanes (C₁₃ to C₁₇) were completed by linear regression (Table 2) using the measured partition coefficients for C₈ to C₁₂. The brain:air PC value was excluded because of the unexplained experimental results that was mentioned previously and linear regression could not be used if this data set was included in the analysis. Simple linear regression using the inverse of the predicted water solubility provided reasonable fits to our measured data for C₈ to C₁₂ (Table 3, Figure 2). As expected, the predicted solubility of the larger n-alkanes increased in tissues and blood with increasing number of carbon atoms.

DISCUSSION

Vial equilibration was successfully used for determining C₈- C₁₂ n-alkane partition coefficient values in blood, brain, fat, liver, and muscle. C₉ to C₁₂ n-alkanes are semivolatile chemicals with low vapor pressures, whereas the C₈ n-alkane has physical properties similar to other volatile organics used with vial equilibration. Special considerations were required when handling C₉ to C₁₂ alkanes, which included care in cleaning syringes, cleaning tedlar bags, and calculating the atmospheric saturation for these compounds. Attempts to measure tridecane (C₁₃) and tetradecane (C₁₄) in the headspace of the equilibration vials failed. The headspace concentrations were below the limits of detection for our analytical method. Thus, these chemicals were not evaluated for tissue solubility.

The n-alkane blood:air PC values increased with number of carbon atoms suggesting that lipophilicity of the blood is important for describing blood:air PC values. The water solubility of the n-alkanes decreases with increasing number of carbon atoms while the log P values increase (Table 4). We compared linear fits of our measured PC values with SPARC-calculated log P and 1/WS to predict the PC values for longer chain n-alkanes. The use of 1/WS provided the best fits and was therefore used to predict the PC values for C₁₃ to C₁₇. By inference, the inverse of water solubility (1/WS) is lipid solubility, although log P also reflects lipophilicity. The reason why 1/WS provided the better fit is unknown.

The only experimentally determined rodent PC values found in the literature for chemicals that we evaluated were in vivo blood:air PC values for n-octane and n-decane (Liu et al. 1994) and in vitro tissue:air and blood:air for n-nonane (Robinson 2000). Our in vitro blood:air PC values for octane and decane (3.1 and 8.1) were less than the in vivo PC values reported by Liu et al. (1994) (7.5 and 17.3). Liu et al. (1994) reported difficulties with decane which may be related to the decane sticking to glass (Perleberg, Keys, and Fisher 2004) and yielding an artificially high partition blood:air PC value. Our blood:air, liver:air, fat:air, muscle:air, and brain:air PC values of nonane (5.8, 11.3, 1588, 4.7, and 22.3, respectively) using thawed frozen tissue and blood compared favorably to those reported by Robinson (2000) (5.1, 6.6, 1254, 7.1, and 25.9, respectively) using fresh tissue and blood from rats.

The brain:air PC value for undecane appears inconsistent with other n-alkanes (Table 1). That is, the mean undecane brain:air PC value (38.73) was only slightly greater than decane (35.25) and the mean brain:air PC for dodecane was much greater (485.9). The expectation, based on other tissues, was that the undecane brain:air PC value would be greater than we measured. We did repeat the experiments with decane, unde-cane, and dodecane (seven replicate vials) and obtained similar findings to our first experiments using seven replicate vials. The reason for the discrepancy is unknown, but may be related to regional composition differences in the brain.

The goal of this research is to eventually develop a PBPK model for JP-8, including the n-alkane fraction (Figure 1) and aromatic fraction of the fuel. To that end we will rely on measured PC values in the literature for some hydrocarbon components (Dennison, Andersen, and Yang 2003), PC values reported in this paper, and predicted PC values for n-alkanes larger than C₁₂ that are reported in this paper. However, a recent phar- macokinetic analysis of the behavior of decane in our laboratory using a PBPK model (Pereleberg, Keys, and Fisher 2004) suggests that the tissue distribution and availability of decane for exchange between tissue and blood and perhaps lung blood and alveolar air is complex. The kinetic behavior of decane could not be simulated by assuming that the model compartments are uniformly mixed and that PC values reflected the degree of exchange between tissue and blood. The reason for this kinetic behavior with decane remains to be elucidated. Further research is needed to determine if n-alkanes larger than decane exhibit this complex kinetic behavior.