- Email: [email protected]
Dietary and age influence on the pharmacokinetic parameters of 2-acetylaminofluorene in BALBc mice
FUNDAMENTAL
AND APPLIED TOXICOLOGY
4, 164-l
69 (1984)
Dietary and Age Influence on the Pharmacokinetic Parameters of 2-Acetylaminofluorene in BALB/c Mice JOHN F. YOUNG*
AND MICHAEL J. NORVELL~
*Division of Teratogenesis Research, National Center for Toxicological Research, FDA/DHHS. Jefferson, Arkansas 72079. and tToxicology Department, Chemical Research and Development Center, FMC Corporation, Princeton, New Jersey 08540
Dietary and Age Influence on the Pharmacokinetic Parameters of 2-Acetylaminofluorene in BALB/c Mice. YOUNG, J. F., AND NORVELL, M. J. (1984). Fundam. Appl. Toxicol. 4, 164-169. In a chronic study conducted in our laboratories, we found dietary effectson 2-acetylaminofluorene (ZAAF) toxicity. To determine if the absorption, distribution, metabolism, or elimination of 2AAF was altered by the age of the animals, diet, or 2-AAF treatment, pharmacokinetic studies were conducted on young (11 weeks) and old (78 weeks) BALB/c mice that had been fed one of four diets containing 4 or 24% fat and 12 or 24% protein. Blood, urine, and feces samples were obtained over a 3 1-hr period after dosing with 500 ppm 2-[14C]AAF (10 pCi/mouse). Total radioactivity was determined after cornbusting each sample in an oxygen atmosphere and counting in a scintillation counter. The data from each individual mouse was simulated on an analogdigital hybrid computer utilizing a two-compartment model with metabolism. The pharmacokinetic parameters within groups were analyzed to make statistical comparisons of the effects of diet, dose, age, and interactions among the groups. The pharmacokinetic predictions of a shorter elimination half-life, smaller area-under-the-curve value, and therefore a decreased exposure to 2-AAF and metabolites due to an increased elimination rate of the parent compound through the urine were consistent with the decreased pathological effects found from the chronic study for the low protein/low fat diet mice.
A Wweek feeding study with 2-acetylaminofluorene (2-AAF) was conducted to correlate the effect of four diets (high and low levels of fat and protein) on liver and urinary bladder neoplasia in mice (Frith et al., 1980). Mice on the low fat diets when compared to mice on the high fat diets exhibited decreased mortality, decreased incidence of hepatocellular adenomas, and decreased incidence of urinary bladder carcinomas. The purpose of the present study was to determine if these pathological differences might have a pharmacokinetic basis. To determine if the absorption, distribution, metabolism, or elimination of 2-AAF was altered by the age of the animals, diet, or by 2-AAF treatment, pharmacokinetic studies were conducted on young and old BALBJci mice that had been fed one of four combi’ BALB/cStCr/C3Hf/Nctr 0272-0590/84 $3.00
nations of diets containing 4 or 24% fat and 12 or 24% protein. EXPERIMENTAL Pharmacokinetic studies were conducted on 10 groups of BALB/c mice administered 500 ppm 2-[14C]AAF (10 PCi, 3-4 mg/kg) by gavage using 200 ~1 corn oil as the vehicle. Each mouse was maintained on its respective diet until gavaged. The experimental design is outlined in Table 1. Old and young mice were approximately 78 weeks and 11 weeks, respectively, when administered 2-[‘4C]AAF; the older mice were treated as described by Frith et al. (L980). AU had been on their respective diets since weaning. Blood, urine, and feces samples were obtained over a 31-hr period (when blood concentrations were below detectable levels and essentially all radioactivity was accounted for) as described by Bazare et al. (1981). Each 3-~1 whole blood sample was cornbusted in an oxygen atmosphere and the resulting solution was counted on a
2 Oxymat, Model JA 101, Intertechnique, Fairheld, N.J.
female mice. .I, 104
2-ACETYLAMINOFLUORENE
165
IN MICE
TABLE 1 NUMBER
OF MICE IN
EACH
LP/LF” Lifetime 5ClQppm 2-AAFb Lifetime control? Young controls6
4 4 4
EXPERIMENTAL
LP/HF
GROUP HP/LF
HP/HF
-E 4 4
-= 4 4
4 2d 4
a LP/LF = low protein, low fat; LP/HF = low protein, high fat; HP/LF = high protein, low fat; HP/HF = high protein, high fat. b The lifetime animals were approximately 78 weeks old and the young control mice were approximately 11 weeks old at the time of the pharmacokinetic experiments. c Mice in these lifetime groups were not available for the pharmacokinetic experiments since all animals had already developed tumors and had been sacrificed by the time of the scheduled pharmacokinetic experiments. dOnly two lifetime control mice from this diet group were available for the pharmacokinetic experiments since most of the animals had already developed tumors and had been sacrificed. scintillation counter.3 A 0.1~pl urine aliquot and all recovered feces were combusted and counted. Appropriate instrument and procedurai controls were run daily to assure the integrity of the Oxymat and scintillation counter. The identification. of the metaboiic pattern from the individual mouse samples was not attempted due to the low levels of compound in any individual blood, urine, or fecal sampbe. The simulation of the pharmacokinetic data was ao complished utiiizing an analog-digital hybrid computer’ (Pearce and Young, 1981; Young et al., 1981). The data were fit to a two-‘compartment model with a first order metabolic route of eiimination (Fig 1) since 2-AAF had been reported to he extensively metaboiized in the rat (Weisburger and Weisburger, 1973) and mouse (Fullerton and Jackson, 19%). In this model the radioactivity data curves for simulation are the sum ofthe parent compound and the metaboiitle(s) in the blood, mine, and feces; i.e., the blood curve was the sum of B + M, the urine curve was U + u’ and ihe feces curve was F + F (see legend to Fig, 1 for explanation of symbols). An additional elimination pathway was included (X) to account for approximately 1.5% Iof the dose that could not be accounted for in the feces or mine. The area under the curve for the individual ctompartments was obtained by integration of the desired component- by the hybrid computer. All statistical colmparisons were conducted using a Sta* tistical Analysis System program (SAS Users Guide, 197 1) on an IBM 41341system or by simple comparisons of two means using, the Student t distribution (Ostle, 1964).
RESULTS AND DISCUSSION The purpose of this study was to investigate the effects of diets containing two different levels of Eat and protein on the pharmacokinetic paralneters
of mice administered
%AAi?
3 Model 92, Searle Analytic, Des Plaines, Ill. 4 Pacer 500, Electronics Associates, Inc., West Long Branch, N.J.
In addition, the relationship of these pharmacokinetic parameters to the previuusly reported pathology dati of Frith et al, (1980) was also of interest. Prior to the investigation of these comparisons, the appropriateness of the analysis and modeling was considered. Total radioactivity measurements of pharmacokinetic data are subject to potential misinterpretation. Since this measurement rntiy represent the parent compound, a number of metabolites, or any combination of the parent and metabolites, the choice of a phamlacokinetic model may
f M--R-T
FIG. 1. Two-compartment oral pharmacokinetic model with metabolism. G = 2-AAF in the gastrointestinal tract. B = 2-AAF in the blood/fluid 6r central compartment. T = 2-AAF in a tissue or peripheral compartment. U = 2-AAF excreted itit the urine. F = 2-AFF excreted itito fhe feces. M = Metabolic products of 2-AAF in the blood compartment. U’ = Metabolic products of 2-AAF excreted in the urine. F’ = Metabolic products of 2-AAF excreted in the feces. X = Total radioactivity not accounted for in the excretion products ( I .5% of dose).
166
YOUNG
AND
be somewhat arbitrary. The initial rule is to keep the model as simple as possible; however, physiological reality cannot be ignored in the choice of a model. For these 2-AAF data, a two-compartment model with three elimination pathways (urine and feces to match the data and an “X” to allow for a mass balance within the system) and the addition of metabolism was employed. The hybrid computer system used for this simulation combines the ease and speed of solving differential equations of the analog computer with the ease and accuracy of the digital computer for statistical evaluation, documentation, and control. The model is programmed directly onto the patch panel of the analog computer. All of the components of the model are integrated separately and the pertinent components are added together for the statistical comparison to the experimental data points. The curves for the blood, urine, and feces are simulated simultaneously for each individual mouse. In that way the limited urine and feces data are linked directly with the blood data for analysis to result in a bestfit set of rate constants for each animal (Fig. 2, Table 2). Testing for differences within a pharmacokinetic parameter due to dose, diet, or a dose-diet interaction for the lifetime treatment groups (the first four rows in Table 2) resulted in significant comparison values only for diet and dose-diet interaction for the overall elimination parameter (k,). Neither absorption (/c& nor metabolism (kBM) nor the transfer
FIG. 2. Hybrid F symbols
NORVELL
to and from the tissue compartment (kBT, km) parameters were significantly different among these dose groups. When considering the individual components, the comparison of the dose (2-AAF vs LT control) for the low protein/low fat diet yielded a significant value (a < 0.005) for the overall elimination parameter (t distribution). However, for the low protein/ high fat diet, the comparison was not significant. When combining the dose effect across diet, no significant difference was seen (SAS analysis). Examining the diet effect, a significant difference was seen for the 2-AAF treated animals (p < 0.005) for the overall elimination parameter and for the return from the tissue compartment, no difference for the lifetime control animals for any parameter, and when combined there was a significant effect (p < 0.034) due to diet for the overall elimination parameter only. When combining all of these values, a dose-diet interaction was observed (p < 0.001). The examination of age, diet, and age-diet interaction yielded various significant values among individual pairs of mean values (Table 2, all control data) for the pharmacokinetic parameters for absorption, distribution, metabolism, and elimination. However, the overall analysis indicated that there was a significant age effect (p < 0.001) and age-diet interaction (p < 0.026) only for the overall elimination parameter. Individual comparisons could be made on many different parameters and treatment groups, however, the meaning of any individ-
computer fit of data from mouse No. 5 on a low protein/low fat diet. The arc the total radioactivity data points from blood, urine, and feces, respectively.
B, U. and
2-ACETYLAMINOFLUORENE
167
IN MICE
TABLE 2 --
HALF-LIFE VALUES(~~~) FORTHERATE
CONSTANTSFROMTHEHYBRID
k BM
bb
2 24k 220 k 101
70 rf: 38 341 + 335
61f 25 344 + 108
97* 54 175 f 207 174 f 64 164 f 85
292 152 69 204
+ 249 f 150 f 14 f 220
151 f 80 64 f Od 184k 0 268 + 192
320 168 249 345
+ + -+ +
90 76 35 96
43 f 23 43 f 23 33+ 11 33 + 23
290 212 278 194
+ f f f
54 f 196 FL 70* 156 f
109 220 138 182
+ * f +
59 100 43 61
k BT
Lifetime 2-AAF LP/LF LP/HF
8.5 f 1.0’ 12.0 + 5.2
85.8 2~ 30.2 25.8 f 9.5
226 f 67+
Lifetime control LP/LF LP/HF HP/LF HP/HF
10.6 9.6 10.2 12.0
+ + + +
2.7 1.5 1.0 1.7
25.2 32.2 34.0 31.8
+ f + f
5.1 15.2 5.5 4.5
15.0 12.2 10.7 11.2
k + + +
1.4 2.2 0 1.3
25.5 29.4 20.2 22.4
f 6.7 t 9.5 * 3.7 + 1.1
Young control LP/LF LP/HF HP/LF HP/HF --
ANALYSIS'
k B”
k3B
~-
COMPUTER
k TB
64 19
119 87 119 25
25 66 11 69
’ The values for kBx, kBF,kMu’, kM; were not included since their contribution was minimal to the overall statistical evaluation. b k, = (kBu + kBF + kBx + kBM)/( 1 + kBT/kTB); t% = 0.693/k, (Garrett, 1974). ’ Mean f standard deviation. d The standard deviation was zero since the same rate constant was used to obtain the best fit of the data with all animals in that group.
ual statistical comparison is debatable. As illustrated by Nelson and Holson (1978) when enough parameters were compared, a certain number of comparisons were significant by chance alone. However, the main reason for any comparisons should be to define their relevance to toxicological findings. Since differences were found between test groups for solme pharmacokinetic parameters, an associatio:n was sought between the pharmacokinetic parameters and the pathological findings of the 7%week study. One conclusion of Frith et al. (1980) was that the mice on the low protein/low fat diet had a decreased mortality, dIecreased incidence of hepatocellular adenomas, and decreased incidence of urinary bladder carcinomas relative to the other diet groups. The significantly shorter halflife for the overall elimination (k,) parameter for the 2-AAF group between the low protein/low fat and low protein/high fat diets (Table 2) was consistent with these pathological findings. This indicated that the low protein/low fat diet anilmals were exposed to 2-AAF for a shorter period of time. The greatly decreased mean hallf-life values ( 1-hr vs nearly 6-hr half-
life for low protein/low fat and low protein/ high fat diets, respectively; Table 2) suggested that the 2-AAF would accumulate to a lesser extent in the low protein/low fat animals than in the low protein/high fat animals. This was also supported by the four- to sixfold decreased area under-the-curve (AUC) data (Table 3) in the low fat diet animals from these same 2AAF lifetime treated mice for the B or T compartments but not for the B + M area. The radioactivity measured total blood level (B + M) indicated no difference between the diet groups, but when the parent and metabolites were separated utilizing the hybrid model analysis, a significant difference between the diet groups was observed. The flux in and out of the peripheral compartment (km and kTrJ was significantly slower for the low protein/ low fat diet animals but the ratio of kB,/kT, = 0.4 was the same for both diets which indicated that each would be classified as shallow compartments (Young, 1983). The major reason for the overall elimination being faster in the low protein/low fat diet animals was the increased elimination rate of 2-AAF to the urine. The mean metabolism
168
YOUNG
AND TABLE
AREA UNDER
Diet LP/LF LP/HF
3
THE CURVE VALUES FOR THE LIFETIME 2-AAF TREATED MICE'
B 0.72 4.04
NORVELL
k 0.08' Ik 2.20
T
@ + Ml 5.77 5.27
IL 4.90 xix2.02
2.27 9.52
+ 1.64 + 2.01
% Metabolism b 82.4 27.0
f f
8.4 15.9
a AUC values (X 10-6) were derived directly from the hybrid computer with units of dpm-minutes. b % Metabolism was calculated from the AUC data using the formula %M = %[(B + M) - (B)]/(B + M). ’ Mean + standard deviation.
value was faster for the low protein/low fat diet animals as well but was not statistically significant due to the high variability of the iow protein/high fat diet animals for this parameter. However, when utilizing the AUC values (Table 3) to estimate the extent of metabolism, the low protein/low fat diet mice appeared to metabolize 2-AAF to a significantly greater extent (82 vs 27%) which aided the overall increased elimination of the total radioactivity. This faster elimination and therefore decreased exposure of 2-AAF in both the central and peripheral compartments in the low protein/low fat diet group could possibly explain the decreased pathological findings for this diet group. Unfortunately, 2-AAF treated animals were not available in the high protein/low fat and high protein/high fat diets for additional confirmation of these pharmacokinetic differences. Another finding (Frith et al., 1980) was that animals on the low fat diets combined (LP/ LF and HP/LF) when compared to the high fat diet animals combined (LP/I-IF and HP/ HF) also exhibited a decreased mortality, decreased incidence of hepatocellular adencrmas, and decreased incidence of urinary bladder carcinomas. The high or low protein levels did not appear to influence the pathological outcome even though the low protein diet animals weighed less throughout the time course of the study when compared to the high protein diet animals. When the young control mice were examined for their metabolic capabilities (kBM, Table 2), a significant difference was seen in all groups compMng high and low fat diets but no difference when the fat content was held constant either high or low and the protein level varied. No differences
were seen in the metabolic capabilities of the lifetime control animal. Examination of the mean metabolism rate constant data (ICBM)for the young control animals in Table 2 indicated that the half-lives for the low fat diets (54 and 70) wese significantly lower than for the high fat diets (196 and 156). These data suggest that the metabolic conversion of 2-AAF was more efficient in the low fat diets and acted to decrease the pathological effects of the chemical. This was consistent with the review by Ioannides and Parke (1979) which stated that the lipid content of the diet affected the composition and function of the liver endoplasmic reticulum. Norred and Wade (1972) depressed the liver cytochrome P-450 concentration and the microsomal metabolism capabilities by feeding rats fat-free diets for 3 weeks when compared to animals receiving a 3% corn oil diet. Hietanen et al. (1975) demonstrated an increased activity of the mixed-function oxidases by the add&ion of 5-10% olive oil to the diet. This same study demonstrated that an increase of dietary fat to 25-55% lowered the hepatic OXygenases, These reports of increased metabolic capabilities of animals on low fat diets (4% in our study) when compared to high fat diets (24% ia our study) were consistent with our data. It must be emphasized that these links between the pharmacokinetic parameters and pathological findings are $tatistical relationships. These correlations may or may not lend insight into the mechanism of action of the 2-AAF induced toxicity. Even thot& the hybrid computer aided mathematical analysis of the pharmacokinetic data allowed for the use of a physiologically realistic model (inclusion
2-ACETYLAMINOFLUORENE
of metabolism), a degree of uncertainty was introduced into the analysis since no metabolic products were specifically quantitated.
IN MICE
SAS analysis. For always preparing a superb and timely manuscript, the authors also wish to express their appreciation to Rose Huber and Cindy Hartwick.
CONCLUSION The overall design of this pharmacokinetic study allowed for the comparison of the diet, dose, age, and interactions among the groups as well as to assessthe variation in individual pharmacokinetic parameters. Even though a statistic(ally balanced design was not obtained due to the unavailability of animals from two dose groups, the pharmacokinetic evaluation provideid a basis for understanding the conclusions reported by Frith et al. (1980) as to the patlhological effects of 2-AAF. The decreased pathological efkcts of the low protein/ low fat diet animals were consistent with the pharma~cokinetic findings of a shorter elimination half-life, smaller AUC, and therefore a decreased exposure to 2-AAF and metabelites due to an increased elimination rate through the urine. The apparent protection of the low fat diet was consistent with the lower half-life value seen in the young control group metabolic rate constant even though no metabolic differences were demonstrated in the older lifetime animals. The correlations of pharmacokinetic data with patlhological findings were also dependent upon the age of the animal used for comparisons. The pharmacokinetic parameters describing .the 1ilFetimeanimals (both control and 2-AAF treated) were fairly consistent except for the overall elimination parameter. However, a great deal of difference was found when comparing the young and old control animals. Therefore, one must be cautious when associating toxicological findings with pharmacokinetic parameters and be sure to consider environmental and experimental factors which may affect the interpretation of the results. ACPLNOWLEDGMENTS Special thanks go to Johnny Bazare and Mary Lee Leamom for scheduling and performing the pharmacokinetic animal studies and to C. J. Nelson for performing the
169
REFERENCES BAZARE,J. J., LEAMONS,M. L., AND YOUNG, J. F. (198 1). Sampling methods for pharmacokinetic studies in the mouse. J. Pharmacol. Methods 5, 99-120. FRITH, C. H., NORVELL, M. J., UMHOLTZ, R., AND KNAPKA, J. J. (1980). Effect of dietary protein and fat levels on liver and urinary bladder neoplasia in mice fed 2-acetylaminofluorene. J. Food Safety 2, 183- 198. FULLERTON, F. R., AND JACKSON, C. D. (1976). Determination of 2-acetylaminofluorene and its metabolites in urine by high pressure liquid chromatography. Biochem.
Med.
16, 95-103.
GARRETT,E. R ( 1974). Classical pharmacokinetics to the frontier. In Pharmacology and Pharmacokinetics (T. Teorell, R. L. Dedrick, and P. G. Condliffe, eds.), pp. 3-25. Plenum, New York. HELWIG, J. T., AND COUNCIL, K. A. (eds.) (1971). SAS Users Guide. SAS Institute, Raleigh, N.C. HIETANEN, E., LAITINEN, M., VAINIO, H., AND HANNINEN, 0. (1975). Dietary fats and properties of endoplasmic reticulum. II. Dietary lipid induced changes in activities of drug metabolising enzymes in liver and duodenum of rats. Lipids l&467-472. IOANNIDES,C., AND PARKE, D. V. (1979). Effect of diet on the metabolism and toxicology of drugs. J. Human Nutr.
33, 357-366.
NELSON, C. J., AND HOLSON, J. F. (1978). Statistical analysis of teratology data: Problems and advancements. J. Environ. Pathol. Toxicol. 2, 187-199. NORRED, W. P., AND WADE, A. E. (1972). Dietary fatty acid-induced alterations of hepatic microsomal drug metabolism. B&hem. Pharmacol. 21, 2887-2897. OSTLE, B. (1964). Statistics in Research, second ed., p. 119. Iowa State Univ. Press, Ames, Iowa. PEARCE,B. A., AND YOUNG, J. F. (1981). A hybrid computer system for pharmacokinetic modeling. I. Software considerations. Proc. 1981 Summer Computer Simulation Con& 117-121. WEISBURGER, J. H., AND WEISBLJRGER,E. K. (1973). Biochemical formation and pharmacological, toxicological, and pathological properties of hydroxylamines and hydroxamine acids. Pharmacol. Rev. 25, 1-66. YOUNG, J. F. (1983). Pharmacokinetic modeling and the teratologist. In Teratogenesis and Reproductive Toxicology (D. M. Kochhar and E. M. Johnson, eds.), Handbook of Experimental Pharmacology, Vol. 65, pp~ 7-29. Springer-Verlag, New York. YOUNG, J. F., WHITE, C. G., AND PEARCE, B. A. (1981). A hybrid computer system for pharmacokinetic modeling. II. Application. Proc. 1981 Summer Camp. Simulation
Conf
122-129.
AND APPLIED TOXICOLOGY
4, 164-l
69 (1984)
Dietary and Age Influence on the Pharmacokinetic Parameters of 2-Acetylaminofluorene in BALB/c Mice JOHN F. YOUNG*
AND MICHAEL J. NORVELL~
*Division of Teratogenesis Research, National Center for Toxicological Research, FDA/DHHS. Jefferson, Arkansas 72079. and tToxicology Department, Chemical Research and Development Center, FMC Corporation, Princeton, New Jersey 08540
Dietary and Age Influence on the Pharmacokinetic Parameters of 2-Acetylaminofluorene in BALB/c Mice. YOUNG, J. F., AND NORVELL, M. J. (1984). Fundam. Appl. Toxicol. 4, 164-169. In a chronic study conducted in our laboratories, we found dietary effectson 2-acetylaminofluorene (ZAAF) toxicity. To determine if the absorption, distribution, metabolism, or elimination of 2AAF was altered by the age of the animals, diet, or 2-AAF treatment, pharmacokinetic studies were conducted on young (11 weeks) and old (78 weeks) BALB/c mice that had been fed one of four diets containing 4 or 24% fat and 12 or 24% protein. Blood, urine, and feces samples were obtained over a 3 1-hr period after dosing with 500 ppm 2-[14C]AAF (10 pCi/mouse). Total radioactivity was determined after cornbusting each sample in an oxygen atmosphere and counting in a scintillation counter. The data from each individual mouse was simulated on an analogdigital hybrid computer utilizing a two-compartment model with metabolism. The pharmacokinetic parameters within groups were analyzed to make statistical comparisons of the effects of diet, dose, age, and interactions among the groups. The pharmacokinetic predictions of a shorter elimination half-life, smaller area-under-the-curve value, and therefore a decreased exposure to 2-AAF and metabolites due to an increased elimination rate of the parent compound through the urine were consistent with the decreased pathological effects found from the chronic study for the low protein/low fat diet mice.
A Wweek feeding study with 2-acetylaminofluorene (2-AAF) was conducted to correlate the effect of four diets (high and low levels of fat and protein) on liver and urinary bladder neoplasia in mice (Frith et al., 1980). Mice on the low fat diets when compared to mice on the high fat diets exhibited decreased mortality, decreased incidence of hepatocellular adenomas, and decreased incidence of urinary bladder carcinomas. The purpose of the present study was to determine if these pathological differences might have a pharmacokinetic basis. To determine if the absorption, distribution, metabolism, or elimination of 2-AAF was altered by the age of the animals, diet, or by 2-AAF treatment, pharmacokinetic studies were conducted on young and old BALBJci mice that had been fed one of four combi’ BALB/cStCr/C3Hf/Nctr 0272-0590/84 $3.00
nations of diets containing 4 or 24% fat and 12 or 24% protein. EXPERIMENTAL Pharmacokinetic studies were conducted on 10 groups of BALB/c mice administered 500 ppm 2-[14C]AAF (10 PCi, 3-4 mg/kg) by gavage using 200 ~1 corn oil as the vehicle. Each mouse was maintained on its respective diet until gavaged. The experimental design is outlined in Table 1. Old and young mice were approximately 78 weeks and 11 weeks, respectively, when administered 2-[‘4C]AAF; the older mice were treated as described by Frith et al. (L980). AU had been on their respective diets since weaning. Blood, urine, and feces samples were obtained over a 31-hr period (when blood concentrations were below detectable levels and essentially all radioactivity was accounted for) as described by Bazare et al. (1981). Each 3-~1 whole blood sample was cornbusted in an oxygen atmosphere and the resulting solution was counted on a
2 Oxymat, Model JA 101, Intertechnique, Fairheld, N.J.
female mice. .I, 104
2-ACETYLAMINOFLUORENE
165
IN MICE
TABLE 1 NUMBER
OF MICE IN
EACH
LP/LF” Lifetime 5ClQppm 2-AAFb Lifetime control? Young controls6
4 4 4
EXPERIMENTAL
LP/HF
GROUP HP/LF
HP/HF
-E 4 4
-= 4 4
4 2d 4
a LP/LF = low protein, low fat; LP/HF = low protein, high fat; HP/LF = high protein, low fat; HP/HF = high protein, high fat. b The lifetime animals were approximately 78 weeks old and the young control mice were approximately 11 weeks old at the time of the pharmacokinetic experiments. c Mice in these lifetime groups were not available for the pharmacokinetic experiments since all animals had already developed tumors and had been sacrificed by the time of the scheduled pharmacokinetic experiments. dOnly two lifetime control mice from this diet group were available for the pharmacokinetic experiments since most of the animals had already developed tumors and had been sacrificed. scintillation counter.3 A 0.1~pl urine aliquot and all recovered feces were combusted and counted. Appropriate instrument and procedurai controls were run daily to assure the integrity of the Oxymat and scintillation counter. The identification. of the metaboiic pattern from the individual mouse samples was not attempted due to the low levels of compound in any individual blood, urine, or fecal sampbe. The simulation of the pharmacokinetic data was ao complished utiiizing an analog-digital hybrid computer’ (Pearce and Young, 1981; Young et al., 1981). The data were fit to a two-‘compartment model with a first order metabolic route of eiimination (Fig 1) since 2-AAF had been reported to he extensively metaboiized in the rat (Weisburger and Weisburger, 1973) and mouse (Fullerton and Jackson, 19%). In this model the radioactivity data curves for simulation are the sum ofthe parent compound and the metaboiitle(s) in the blood, mine, and feces; i.e., the blood curve was the sum of B + M, the urine curve was U + u’ and ihe feces curve was F + F (see legend to Fig, 1 for explanation of symbols). An additional elimination pathway was included (X) to account for approximately 1.5% Iof the dose that could not be accounted for in the feces or mine. The area under the curve for the individual ctompartments was obtained by integration of the desired component- by the hybrid computer. All statistical colmparisons were conducted using a Sta* tistical Analysis System program (SAS Users Guide, 197 1) on an IBM 41341system or by simple comparisons of two means using, the Student t distribution (Ostle, 1964).
RESULTS AND DISCUSSION The purpose of this study was to investigate the effects of diets containing two different levels of Eat and protein on the pharmacokinetic paralneters
of mice administered
%AAi?
3 Model 92, Searle Analytic, Des Plaines, Ill. 4 Pacer 500, Electronics Associates, Inc., West Long Branch, N.J.
In addition, the relationship of these pharmacokinetic parameters to the previuusly reported pathology dati of Frith et al, (1980) was also of interest. Prior to the investigation of these comparisons, the appropriateness of the analysis and modeling was considered. Total radioactivity measurements of pharmacokinetic data are subject to potential misinterpretation. Since this measurement rntiy represent the parent compound, a number of metabolites, or any combination of the parent and metabolites, the choice of a phamlacokinetic model may
f M--R-T
FIG. 1. Two-compartment oral pharmacokinetic model with metabolism. G = 2-AAF in the gastrointestinal tract. B = 2-AAF in the blood/fluid 6r central compartment. T = 2-AAF in a tissue or peripheral compartment. U = 2-AAF excreted itit the urine. F = 2-AFF excreted itito fhe feces. M = Metabolic products of 2-AAF in the blood compartment. U’ = Metabolic products of 2-AAF excreted in the urine. F’ = Metabolic products of 2-AAF excreted in the feces. X = Total radioactivity not accounted for in the excretion products ( I .5% of dose).
166
YOUNG
AND
be somewhat arbitrary. The initial rule is to keep the model as simple as possible; however, physiological reality cannot be ignored in the choice of a model. For these 2-AAF data, a two-compartment model with three elimination pathways (urine and feces to match the data and an “X” to allow for a mass balance within the system) and the addition of metabolism was employed. The hybrid computer system used for this simulation combines the ease and speed of solving differential equations of the analog computer with the ease and accuracy of the digital computer for statistical evaluation, documentation, and control. The model is programmed directly onto the patch panel of the analog computer. All of the components of the model are integrated separately and the pertinent components are added together for the statistical comparison to the experimental data points. The curves for the blood, urine, and feces are simulated simultaneously for each individual mouse. In that way the limited urine and feces data are linked directly with the blood data for analysis to result in a bestfit set of rate constants for each animal (Fig. 2, Table 2). Testing for differences within a pharmacokinetic parameter due to dose, diet, or a dose-diet interaction for the lifetime treatment groups (the first four rows in Table 2) resulted in significant comparison values only for diet and dose-diet interaction for the overall elimination parameter (k,). Neither absorption (/c& nor metabolism (kBM) nor the transfer
FIG. 2. Hybrid F symbols
NORVELL
to and from the tissue compartment (kBT, km) parameters were significantly different among these dose groups. When considering the individual components, the comparison of the dose (2-AAF vs LT control) for the low protein/low fat diet yielded a significant value (a < 0.005) for the overall elimination parameter (t distribution). However, for the low protein/ high fat diet, the comparison was not significant. When combining the dose effect across diet, no significant difference was seen (SAS analysis). Examining the diet effect, a significant difference was seen for the 2-AAF treated animals (p < 0.005) for the overall elimination parameter and for the return from the tissue compartment, no difference for the lifetime control animals for any parameter, and when combined there was a significant effect (p < 0.034) due to diet for the overall elimination parameter only. When combining all of these values, a dose-diet interaction was observed (p < 0.001). The examination of age, diet, and age-diet interaction yielded various significant values among individual pairs of mean values (Table 2, all control data) for the pharmacokinetic parameters for absorption, distribution, metabolism, and elimination. However, the overall analysis indicated that there was a significant age effect (p < 0.001) and age-diet interaction (p < 0.026) only for the overall elimination parameter. Individual comparisons could be made on many different parameters and treatment groups, however, the meaning of any individ-
computer fit of data from mouse No. 5 on a low protein/low fat diet. The arc the total radioactivity data points from blood, urine, and feces, respectively.
B, U. and
2-ACETYLAMINOFLUORENE
167
IN MICE
TABLE 2 --
HALF-LIFE VALUES(~~~) FORTHERATE
CONSTANTSFROMTHEHYBRID
k BM
bb
2 24k 220 k 101
70 rf: 38 341 + 335
61f 25 344 + 108
97* 54 175 f 207 174 f 64 164 f 85
292 152 69 204
+ 249 f 150 f 14 f 220
151 f 80 64 f Od 184k 0 268 + 192
320 168 249 345
+ + -+ +
90 76 35 96
43 f 23 43 f 23 33+ 11 33 + 23
290 212 278 194
+ f f f
54 f 196 FL 70* 156 f
109 220 138 182
+ * f +
59 100 43 61
k BT
Lifetime 2-AAF LP/LF LP/HF
8.5 f 1.0’ 12.0 + 5.2
85.8 2~ 30.2 25.8 f 9.5
226 f 67+
Lifetime control LP/LF LP/HF HP/LF HP/HF
10.6 9.6 10.2 12.0
+ + + +
2.7 1.5 1.0 1.7
25.2 32.2 34.0 31.8
+ f + f
5.1 15.2 5.5 4.5
15.0 12.2 10.7 11.2
k + + +
1.4 2.2 0 1.3
25.5 29.4 20.2 22.4
f 6.7 t 9.5 * 3.7 + 1.1
Young control LP/LF LP/HF HP/LF HP/HF --
ANALYSIS'
k B”
k3B
~-
COMPUTER
k TB
64 19
119 87 119 25
25 66 11 69
’ The values for kBx, kBF,kMu’, kM; were not included since their contribution was minimal to the overall statistical evaluation. b k, = (kBu + kBF + kBx + kBM)/( 1 + kBT/kTB); t% = 0.693/k, (Garrett, 1974). ’ Mean f standard deviation. d The standard deviation was zero since the same rate constant was used to obtain the best fit of the data with all animals in that group.
ual statistical comparison is debatable. As illustrated by Nelson and Holson (1978) when enough parameters were compared, a certain number of comparisons were significant by chance alone. However, the main reason for any comparisons should be to define their relevance to toxicological findings. Since differences were found between test groups for solme pharmacokinetic parameters, an associatio:n was sought between the pharmacokinetic parameters and the pathological findings of the 7%week study. One conclusion of Frith et al. (1980) was that the mice on the low protein/low fat diet had a decreased mortality, dIecreased incidence of hepatocellular adenomas, and decreased incidence of urinary bladder carcinomas relative to the other diet groups. The significantly shorter halflife for the overall elimination (k,) parameter for the 2-AAF group between the low protein/low fat and low protein/high fat diets (Table 2) was consistent with these pathological findings. This indicated that the low protein/low fat diet anilmals were exposed to 2-AAF for a shorter period of time. The greatly decreased mean hallf-life values ( 1-hr vs nearly 6-hr half-
life for low protein/low fat and low protein/ high fat diets, respectively; Table 2) suggested that the 2-AAF would accumulate to a lesser extent in the low protein/low fat animals than in the low protein/high fat animals. This was also supported by the four- to sixfold decreased area under-the-curve (AUC) data (Table 3) in the low fat diet animals from these same 2AAF lifetime treated mice for the B or T compartments but not for the B + M area. The radioactivity measured total blood level (B + M) indicated no difference between the diet groups, but when the parent and metabolites were separated utilizing the hybrid model analysis, a significant difference between the diet groups was observed. The flux in and out of the peripheral compartment (km and kTrJ was significantly slower for the low protein/ low fat diet animals but the ratio of kB,/kT, = 0.4 was the same for both diets which indicated that each would be classified as shallow compartments (Young, 1983). The major reason for the overall elimination being faster in the low protein/low fat diet animals was the increased elimination rate of 2-AAF to the urine. The mean metabolism
168
YOUNG
AND TABLE
AREA UNDER
Diet LP/LF LP/HF
3
THE CURVE VALUES FOR THE LIFETIME 2-AAF TREATED MICE'
B 0.72 4.04
NORVELL
k 0.08' Ik 2.20
T
@ + Ml 5.77 5.27
IL 4.90 xix2.02
2.27 9.52
+ 1.64 + 2.01
% Metabolism b 82.4 27.0
f f
8.4 15.9
a AUC values (X 10-6) were derived directly from the hybrid computer with units of dpm-minutes. b % Metabolism was calculated from the AUC data using the formula %M = %[(B + M) - (B)]/(B + M). ’ Mean + standard deviation.
value was faster for the low protein/low fat diet animals as well but was not statistically significant due to the high variability of the iow protein/high fat diet animals for this parameter. However, when utilizing the AUC values (Table 3) to estimate the extent of metabolism, the low protein/low fat diet mice appeared to metabolize 2-AAF to a significantly greater extent (82 vs 27%) which aided the overall increased elimination of the total radioactivity. This faster elimination and therefore decreased exposure of 2-AAF in both the central and peripheral compartments in the low protein/low fat diet group could possibly explain the decreased pathological findings for this diet group. Unfortunately, 2-AAF treated animals were not available in the high protein/low fat and high protein/high fat diets for additional confirmation of these pharmacokinetic differences. Another finding (Frith et al., 1980) was that animals on the low fat diets combined (LP/ LF and HP/LF) when compared to the high fat diet animals combined (LP/I-IF and HP/ HF) also exhibited a decreased mortality, decreased incidence of hepatocellular adencrmas, and decreased incidence of urinary bladder carcinomas. The high or low protein levels did not appear to influence the pathological outcome even though the low protein diet animals weighed less throughout the time course of the study when compared to the high protein diet animals. When the young control mice were examined for their metabolic capabilities (kBM, Table 2), a significant difference was seen in all groups compMng high and low fat diets but no difference when the fat content was held constant either high or low and the protein level varied. No differences
were seen in the metabolic capabilities of the lifetime control animal. Examination of the mean metabolism rate constant data (ICBM)for the young control animals in Table 2 indicated that the half-lives for the low fat diets (54 and 70) wese significantly lower than for the high fat diets (196 and 156). These data suggest that the metabolic conversion of 2-AAF was more efficient in the low fat diets and acted to decrease the pathological effects of the chemical. This was consistent with the review by Ioannides and Parke (1979) which stated that the lipid content of the diet affected the composition and function of the liver endoplasmic reticulum. Norred and Wade (1972) depressed the liver cytochrome P-450 concentration and the microsomal metabolism capabilities by feeding rats fat-free diets for 3 weeks when compared to animals receiving a 3% corn oil diet. Hietanen et al. (1975) demonstrated an increased activity of the mixed-function oxidases by the add&ion of 5-10% olive oil to the diet. This same study demonstrated that an increase of dietary fat to 25-55% lowered the hepatic OXygenases, These reports of increased metabolic capabilities of animals on low fat diets (4% in our study) when compared to high fat diets (24% ia our study) were consistent with our data. It must be emphasized that these links between the pharmacokinetic parameters and pathological findings are $tatistical relationships. These correlations may or may not lend insight into the mechanism of action of the 2-AAF induced toxicity. Even thot& the hybrid computer aided mathematical analysis of the pharmacokinetic data allowed for the use of a physiologically realistic model (inclusion
2-ACETYLAMINOFLUORENE
of metabolism), a degree of uncertainty was introduced into the analysis since no metabolic products were specifically quantitated.
IN MICE
SAS analysis. For always preparing a superb and timely manuscript, the authors also wish to express their appreciation to Rose Huber and Cindy Hartwick.
CONCLUSION The overall design of this pharmacokinetic study allowed for the comparison of the diet, dose, age, and interactions among the groups as well as to assessthe variation in individual pharmacokinetic parameters. Even though a statistic(ally balanced design was not obtained due to the unavailability of animals from two dose groups, the pharmacokinetic evaluation provideid a basis for understanding the conclusions reported by Frith et al. (1980) as to the patlhological effects of 2-AAF. The decreased pathological efkcts of the low protein/ low fat diet animals were consistent with the pharma~cokinetic findings of a shorter elimination half-life, smaller AUC, and therefore a decreased exposure to 2-AAF and metabelites due to an increased elimination rate through the urine. The apparent protection of the low fat diet was consistent with the lower half-life value seen in the young control group metabolic rate constant even though no metabolic differences were demonstrated in the older lifetime animals. The correlations of pharmacokinetic data with patlhological findings were also dependent upon the age of the animal used for comparisons. The pharmacokinetic parameters describing .the 1ilFetimeanimals (both control and 2-AAF treated) were fairly consistent except for the overall elimination parameter. However, a great deal of difference was found when comparing the young and old control animals. Therefore, one must be cautious when associating toxicological findings with pharmacokinetic parameters and be sure to consider environmental and experimental factors which may affect the interpretation of the results. ACPLNOWLEDGMENTS Special thanks go to Johnny Bazare and Mary Lee Leamom for scheduling and performing the pharmacokinetic animal studies and to C. J. Nelson for performing the
169
REFERENCES BAZARE,J. J., LEAMONS,M. L., AND YOUNG, J. F. (198 1). Sampling methods for pharmacokinetic studies in the mouse. J. Pharmacol. Methods 5, 99-120. FRITH, C. H., NORVELL, M. J., UMHOLTZ, R., AND KNAPKA, J. J. (1980). Effect of dietary protein and fat levels on liver and urinary bladder neoplasia in mice fed 2-acetylaminofluorene. J. Food Safety 2, 183- 198. FULLERTON, F. R., AND JACKSON, C. D. (1976). Determination of 2-acetylaminofluorene and its metabolites in urine by high pressure liquid chromatography. Biochem.
Med.
16, 95-103.
GARRETT,E. R ( 1974). Classical pharmacokinetics to the frontier. In Pharmacology and Pharmacokinetics (T. Teorell, R. L. Dedrick, and P. G. Condliffe, eds.), pp. 3-25. Plenum, New York. HELWIG, J. T., AND COUNCIL, K. A. (eds.) (1971). SAS Users Guide. SAS Institute, Raleigh, N.C. HIETANEN, E., LAITINEN, M., VAINIO, H., AND HANNINEN, 0. (1975). Dietary fats and properties of endoplasmic reticulum. II. Dietary lipid induced changes in activities of drug metabolising enzymes in liver and duodenum of rats. Lipids l&467-472. IOANNIDES,C., AND PARKE, D. V. (1979). Effect of diet on the metabolism and toxicology of drugs. J. Human Nutr.
33, 357-366.
NELSON, C. J., AND HOLSON, J. F. (1978). Statistical analysis of teratology data: Problems and advancements. J. Environ. Pathol. Toxicol. 2, 187-199. NORRED, W. P., AND WADE, A. E. (1972). Dietary fatty acid-induced alterations of hepatic microsomal drug metabolism. B&hem. Pharmacol. 21, 2887-2897. OSTLE, B. (1964). Statistics in Research, second ed., p. 119. Iowa State Univ. Press, Ames, Iowa. PEARCE,B. A., AND YOUNG, J. F. (1981). A hybrid computer system for pharmacokinetic modeling. I. Software considerations. Proc. 1981 Summer Computer Simulation Con& 117-121. WEISBURGER, J. H., AND WEISBLJRGER,E. K. (1973). Biochemical formation and pharmacological, toxicological, and pathological properties of hydroxylamines and hydroxamine acids. Pharmacol. Rev. 25, 1-66. YOUNG, J. F. (1983). Pharmacokinetic modeling and the teratologist. In Teratogenesis and Reproductive Toxicology (D. M. Kochhar and E. M. Johnson, eds.), Handbook of Experimental Pharmacology, Vol. 65, pp~ 7-29. Springer-Verlag, New York. YOUNG, J. F., WHITE, C. G., AND PEARCE, B. A. (1981). A hybrid computer system for pharmacokinetic modeling. II. Application. Proc. 1981 Summer Camp. Simulation
Conf
122-129.