You are on page 1of 17



Departments o f Mathematics and Agricultural Chemistry
Oregon State University
Corvallis, Oregon 97331

The preliminary model of the lipid-phase pharmacokinetics of dieldrin in

mammals is applied to evaluation of the probable role of auto-induction, the
effect(s) of growth, and the role of compartmentalization. Realistic simulations and
comparison to actual data permit the conclusion to be drawn that induction of its
own metabolism plays a relatively minor role in the distribution and level of dieldrin
residues. Although some of the effects of growth are reasonably well simulated, the
model as constructed does not permit compartmentalization. It was concluded that
the model should be revised to include aspects of the as-yet quantitatively unknown
intracellular binding and membrane transport within the flow-limited model.

The first paper of this series (Lindstrom et al. 1974), summarized a set of assumptions
governing the oral ingestion, uptake, and movement of highly lipid-soluble drugs and pes-
ticides, such as HEOD1 (dieldrin), in mammals. These assumptions allowed a preliminary
model for the pharmacodynamics of dieldrin to be constructed. As more quantitative ex-
perimental information becomes available, the model will be refined to include this infor-
mation. For example, quantitation of dieldrin, DDT2, etc., movement across the lumen-
mucosa-blood barrier has yet to be made. Information on the kinetics and thermo-
dynamics of dieldrin binding to proteins, lipo-proteins, and even classes of lipids is
generally still lacking.

The line of reasoning we chose to follow in constructing a working model of the hy-
pothesis of pharmacodynamics of dieldrin (or other highly lipid-soluble drugs as well) in
mammals parallels that of Bischoff and Brown (1966), Dedrick and Bischoff (1968),
Bischoff and Dedrick (1970), Bischoff et al. (1971), and Dedrick et al. (1973). These
workers began with fairly simple hypothetical models for the drugs thiopental and
methotrexate (MTX). As more a priori information became available, it was included in
the models, making them more complete and accurate in their predictive value. Thio-
pental and MTX are quite water-soluble and much less lipid-soluble than dieldrin or DDT.
Dieldrin has a partition coefficient of the order of l0 s for a peanut oil/water system, and
DDT has a partition coefficient of the order of 106 for the same system (Wauchope
1972). In view of this highly important difference, it is conjectured that the excellent and

11,2,3,4,10,10-hexachloro-6,7-epoxy-l,4,4a,5,6,7,8,8a-octahydro-l,4:5,8-endo, exo-dimethanonaph-

Archives of Environmental Contamination 166

and Toxicology, Vol. 3, No. 2, 1975
9 1975 by Springer-Verlag New York Inc.
Applications of Dieldrin Distribution Model 167

valuable kinetic and thermodynamic parameter information presented for water-soluble

drugs by Bischoff et al. (1971) and Dedrick et al. (1973) may not apply for highly
lipophilic drugs in mammals.

Lindstrom et al. (1974) concluded their paper with examples of estimated dieldrin
distributions in a hypothetical rat and a hypothetical man using several modes of dieldrin
administration (e.g., ad lib dietary, a single capsule daffy, intravenous). All blood flow and
compartment parameters (including the dieldrin destruction rate coefficient) were as.
sumed to be constant in time. This paper presents several more examples of estimated
dieldrin distributions in a hypothetical rat when: 1) auto-induction of hepatic microsomes
is allowed, 2) the adipose compartment is allowed to grow linearly from some base value
in time, and 3) the model is compared particularly with experimental studies of dieldrin

Case I. Role of auto-induction. Durham (1969) has posed the question: Do pesticides
acting as hepatic inducers affect their own metabolism? Because dieldrin is a potent
inducer of hepatic mixed function orddases (Ghazal et al. 1964, Gillett and Chart 1968)
which include enzymes involved in dieldrin metabolism (Matthews and Matsumura 1969),
it may, be postulated that dieldrin is capable of altering its own metabolism (e.g., auto.
induction) and that the time course of tissue concentrations and the levels of residues
should reflect effects related to increases in dietary exposure. If auto-induction is strong
(e.g., if/~vl, the coefficient of induction of specific dieldrin-metabolizing enzymes, is
large and positive), residues in tissues would be expected to be markedly lower, relative
to dosage, as dosage is increased. In the initial preliminary model (Lindstrom et al. 1974),
l~pt was assumed to be zero. It is necessary, however, to test the hypothesis that dieldrin
does affect its own metabolism; to that end, application of the model serves a very use.
ful function.
If the level of dieldrin intake influences the relationship of residue to exposure level,
then simulation of residues in rats exposed to several levels of dietary dieldrin at several
arbitrarily set values of/3p t should provide estimates of a value of/~p 1 yielding residues
comparable to those measured in actual feeding experiments. If this value of/3p I is large,
then auto-induction can be assumed to play a role in dieldrin pharmacokinetics; con-
versely, a low value for/~p I would suggest that auto-induction is a negligible factor.

Table I (Lindstrom et al. 1974) lists estimates of the blood lipid flow and compart-
mental lipid mass for each of the respective compartments in a hypothetical 300-g
mature male rat eating 20 g of food per day during a 12-hr interval. The model is assumed
to be strictly flow-limited and all compartments to remain constant except for induced
changes in lipid and enzymes in the liver. The parameters characterizing F, the hepatic
dieldrin metabolism coefficient, are listed in Tables II and III. The absence of some par-
ticular values in Table II arises solely because of the assumptions for this simulation that
/3p2, 3 and S = 0. Table III sets forth the values of/~pl for low (1), intermediate (10), and
strong'(100) induction by dieldrin of dieldrin-metabolizing enzymes, the values of three
dietary levels of exposure (~), and the model simulation of the adipose lipid concentra-
Table I. Compartment mass, blood flow, and other important parameters for a hypothetical 300-g mature male rat U,

Compartment Symbol Mass Lipid fraction Lipid mass Blood flow Blood lipid
x 7]~x fx Mx rate flow rate
(g) ( g lipid ~ (g lipid) qx Qx
\ g tissue / ( g blood
hr ) ( g blood
hr lipid )

Blood pool B 23.4 0.004 0.09 4,751.6 18.21

(blood chambers of
heart, lungs, arteries,
and veins)
Kidney K 2.7 0.05 0.05 698.0 2.79 ~q
Gastrointestinal G.I. 15.8 0.07 1.11 811.1 3.24
(small bowel)
Liver L 11.9 0.06 0.72 989.3 3.96
Brain BR 1.9 0.09 O. 17 680.1 2.72
Depot Fat F 54.0 0.64 34.56 710.1 2.84
Muscle M 120.0 0.07 8.40 533.3 2.13
(includes associated fat)
Other tissue OT 60.0 0.11 6.60 133.3 0.53
(skin, etc.)

Rex = 0.0 (Heath and Vandekar 1964).

~9 = 100.0 [estimated from data of Cole et al. (1971)].
0 = 0.995, O ~ < t < 1 2 h o u r s ~ .
0 0.050, 12 ~< t < 24 hours ~dally periodic function
QBI =" 0.0032 g bile/lipid hr, 0 ~< t < 12 hours I
daily periodic function.
QBI 0.00032 g bile/lipid hr, 12 <~ t < 24 hours
Table lk Liver cell dieldrin destruction rate coefficient parametersa for a hypothetical 300-gin mature male rat

No. Protein masses Induction coefficient Catabolism Proportionality

coefficient coefficient

~'Pi g dieldrin Kcpi (1/hr) prote!,n. _ hr
g lipid - hr KDi \ g lipid

1 0.003 (Given in Table III) 0.05 20

2 0.006 0.0 - 28 t~

3 0.012 0.0 - 1

4 0.300 100.0 0.01 0

5 0.291 0.0 - 0 5"

Lipid Masses (g)

MMs (0) = 0.375

Mcy + M N + MMe = 0.345

aThese estimates were obtained from data supplied by Davis et al. (1973) and by Davis (1973).

170 F.T. Lindstrom et al.

tions (CF) of dieldrin at the end of a 4-day feeding period for each of the combinations
~Pl and ~.

The relationship of the resultant residue to the level of exposure as affected by/~p 1 is
shown in Figure 1, a log-log plot of adipose lipid concentration vs. daily dietary dieldrin
intake. As/~p I increases, the relationship of residue to intake become increasingly curvi-
linear exhibiting slopes progressively less than 1.0, the slope obtained with no auto,
induction or interaction of dietary level with residue. For each value of/~p l' least squares
regression lines can be calculated according to a power law rule (y = k - xn), providing
estimates (Table IV) of the constants for Equation (1):

C F = k(~) n. (1)

Walker et al. (1969) described the "steady state" adipose tissue residues (C~) of dieldrin
in rats by a power law equation:

C~ = 1.91 (CD) 0.96, (2)

where C~ is the whole adipose tissue residue Oag/g tissue) and CD is the dietary dieldrin
concentration 0ag/g food). Based on the assumptions used in our simulation (20 g food
consumed per day, 63.7% of the adipose tissue as lipid), the data of Walker etal. (1969)
can be expressed as:

CF = 0.169~ ~ (3)

Comparison of the parameters of Equation (3) with those for the simulations in Table IV
indicates only a small auto-inductive effect. This view is supported further by examina-
tion of the predicted residues at "steady state" conditions (when dieldrin input equals
dieldrin excretion plus metabolism). Using the parameters of Table IV in Equation (1) it
is shown (Table III) that "steady state" adipose residues in mature male rats (Walker et al.

Table lII_ Comparison o f hypothetical rat model with data based upon
the experiment o f Walker et at (1969)

CF (/ag/g lipid) Adipose lipid dieldrin residue

Daily dieldrin From model
,~ (/lg]day) From equation 3a ~Pl = 1.0 /Spl = 10.0 ~p: = I00.0

20 3.0 2.8 2.7 2.2

I00 14.1 13.8 12.0 6.9

500 65.9 64.5 43.2 18.2

a From data of Walker et al. (1969).

Applications of Dieldrin Distribution Model 171

1969) are very close to those predicted by the model assuming that auto-induction is
weak (~p 1 = 1.0).

Although true "steady state" conditions in an animal can only be achieved under cir-
cumstances where the compartment lipid mass, blood flow, metabolic rate, etc. are con-
stant (which never actually occurs in a live animal), the residues achieved in the model
simulation approach "steady-state" conditions (Figure 2). Even though the discontinu-
ous administration of dieldrin (12 hr feeding, 12 hr rest) imposes cyclic variations in
blood and adipose residues, by day 15 the differential in residues in these tissues varies
little from one day to the next, at the same hour. If the dieldrin-fed rat is then switched
to a control ration, and the loss from the adipose compartment followed for another 20
days, an approximation of the daffy destruction rate can be calculated. For the case
shown in Figure 2 (tip z = 1.0; ~ = 41.67 #g dieldrin/hr), the apparent first-order rate of
loss is about 12% per day. Values reported in the literature range from 5% to 20% per

100 ...."

O. o,*" &

E o..'~
~z,. . o.."
~5 .:.~
2n o ~176176176176
'~ 10


(~ ~ /3p]. = 1.0

& -- #PZ = 1 0 . 0

X ~ /~p]. = 1 0 0 . 0

n X ....... comparison slope

(slope = 1.0)

1 I ! ! I ! ! ! I ! ! I ! | ! ! ! | I
10 100 1000
Daily dieldrin intake (Mg dieldrin)

Fig. 1. Log plot of adipose tissue lipid phase dieldrin residue (~g dieldrin/gm lipid) vs. the
log of the daily dieldrin intake rate (/lg dieldrin/day). Time t = 40 days on dieldrin.
172 F.T. Lindstrom et al.

day; Walker et el. (1969) found a value of 11% per day. In model simulations of the
"feed-off" period for greater values of #p 1 and ~ (not shown) the decay curves become
increasingly curvilinear and have steeper initial slopes as the value of BpI increases, but
are parallel between various values of ~(Lindstrom et al. 1974).

It is important to point out that with the assumptions used to construct the lipid-phase
model of dieldrin pharmacodynamics, even the assumption of a weak (tip I -- 1.0) auto-
induction is associated with marked changes in both dieldrin-metabolizing enzymes and
hepatic lipids. Figure 3 shows the relative increases in hepatic constituents in a hypotheti-
cal mature male rat (300-g, feeding 12 hr per day,/~Pt = 1.0, ~ = 41.67/ag/hr) during a
30-day exposure to a 25-ppm dieldrin diet followed by a 20-day feed-off period. In addi-
tion to the critical information on the actual mass (or total activity) of the several
enzymes of dieldrin metabolism (Figure 3a) together with the rate constants of each
activity, also unknown in real terms are data on the turn-over rate of these proteins. In-
creases in total and microsomal lipid (Figure 3d and 3e) tend to offset increases in
dieldrin-metabolizing enzymes, by decreasing the concentration of dieldrin available for
reaction, and hence changes in F (Figure 30 are relatively minor. At higher values of#p l'
F is greatly affected bemuse changes in p 1 are relatively large compared to lipid changes.
Since activity changes following the cessation of exposure do not persist as long as do the
rnorphologlc and structural changes (Ortega 1970), increased mierosomal lipid may
persist relative to induced dieldrin-metabolizing enzymes during feed-off (simulated by
setting kcp 4 < kcp 1)" As with several aspects of this preliminary model, detailed data on
hepatic activities and rates of change are vital to a more explicit and sophisticated

Case II. The role of growth (a f'trst look). The growth of an animal during exposure to
a highly lipid-soluble agent such as a chlorinated hydrocarbon introduces numerous
physiologic and biochemical factors which are essential in the development of an a priori
model of residue build-up and distribution. Even in the sexually mature male rat, some
tissues (e.g., muscle, adipose, skin, etc.) are still growing. Not only is the lipid composition
of some tissues undergoing dynamic alteration, but the blood flow (and lipid flow) rates
to the tissues are also changing with growth. However, the assemblage of realistic growth
parameters suitable for incorporation in the preliminary model of lipid-phase dieldrin
pharmacokinetics is a disparate step that the currently sparse literature on lipid corn-

Table IV. Constants f o r equation (1).

~Pl k n

1 0.172 0.97

10 0.180 0.86

100 0.158 0.66

Applications of Dieldrin Distribution Model 173

positions of growing animals does not seem to be able to reconcile. Nevertheless, studies
by Schemmel e t al. (1970a, b) on obesity in rats do provide an opportunity to evaluate
the influence of growth of a single compartment (e.g., adipose) on dieldrin distribution.

By summing the weight of adipose tissue for the seven selected tissues assayed and
plotting the log of this sum against the log of body weight, the data of Schemmel e t al.
(1970b) provide a good power rule relationship for either a low- or high-fat diet (Figure 4).
Whether or not these curves for the Osborne-Mendel strain of rats are peculiar to that
strain or are applicable to other strains (e.g., Wistar, from which most other parameters
of the rat model have been obtained) obviously limits the validity of simulation. Examina-
tion of the data of Schemmel e t al. (1970b) allows a simple estimate of 0.2 g adipose
tissue/day for an average linear growth rate for rats on the low-fat diet beginning at a
body weight of about 300 g and growing to a body weight of 400 g in about 70 days.
That the sum of the seven fat depots shows a linear log-log relationship raises some
interesting questions concerning the overall mechanics of growth of adipose tissue.

To simulate the effects of adipose tissue growth, a mature male rat weighing 300 g
(at time t = 0) and consuming about 20 g of food per day is assumed to 1) exhibit weak
auto-induction (tip 1 = 1.0), and 2) to maintain a constant fraction of the adipose tissue
as lipid (fir = 0.637). All other parameters were identical to those employed in the pre-
vious simulations. It is necessary to calculate the small increment of blood volume gained
by the addition of the new depot tissue. Assuming that this increment is proportional to
the blood supply of adipose tissue generally [about 2% by volume (Specter 1956)], so
that the incremental volume is about 2% of the daily adipose tissue weight gain, and
assuming that the blood lipid fraction remains constant (fB = 0.004 g lipid/g blood), an
estimated 1.6 • 10 - 5 g of lipid would be added daily to the blood lipid pool. Actually,
similar increments are occurring in muscle, skin, etc. consonant with the general growth,
but are ignored for this illustration.

If the blood lipid flow through the fat mass is constant (QF/MF = Rt), evaluation of

100 jBiood

r O.
"o . _ 1 0
O. v

5 10 15 20 25 30 35 40 45 50 55 60
Time (days)
Fig. 2. Typical semilog plot of blood lipid and adipose tissue lipid phase dieldrin residue
vs. time for a 60-day hypothetical feed-up/feed-off experiment on a 300-g mature male
174 F.T. Lindstrom et al.

R t at t = 0 gives R t = 0.082 L/hr. Because of continuity of flow considerations, the lipid

flow equivalent of cardiac output Qs is given as:

QB(t) = 15.37 + R t 9 MF (g lipid/hr), (4)

where the adipose lipid mass MF(t) is given by the expression:

MF(t) = 34.56 + 0.13 9 t (g lipid), (5)

with t being expressed in days. The blood lipid mass MB(t) is found to be:

MB(t) = 0.09 + 1.6 X 10-s 9 t (g lipid). (6)

I(a) Active inducible
[ mocrosomalprotein mass (P1) Microsomal lipid mass (MMs)

004 e ~
9 1 7 4 9 9

9 o ~ 9

.oo3~ 9 .- "+.6 9 ...

C/ . . . . .

(b) (e)
A Non-active, inducible,
microsomal protein mass (P4) Total liver lipid mass (M L)
E1 .(~ 1.0
g O 9 1 4 9 1 4 9o ~ 9 9 9 9 9 9 9 9 9 9 9 9 9 9
oe o o o o e e 9 9
cq .8
~.6 9 9149 9 9 9 9 9 9 9 9 ~.6
'- .4 3.4
o 99 o
~.2 2
2 ' 6 "1'0 20 30 40 50 o ~ '~'lb ~ ~ ~

(c) Total micros9 5 (f)
protein mass 2; Pi (t)--P4(t) Total destruction rate coefficient (F)
e o e 9 I ~ ~ ~ e ~ 3 'oeeoeeoo
~ . 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9

~I.0 eo
9 9
9 0

~ .8 o oeoe 9 9 9 9 9 9 --.~ . 2
o~ O.
~ .6
"- .4 ~.1
2' 6 'lb 2b 3b ~ ~ ~' ~' 1'0 2b ab ~ s~
Time (days) Time (days)

Fig. 3. Linear plots of some of the important liver morphologic responses to the hypo-
thetical long time chronic dieldrin insult.
Applications of Dieldrin Distribution Model 175

These equations (4-6) are then assumed to def'me the time-dependent relationships of the
growing adipose and blood pool compartments. The simulation format is completed by
specifying that the dietary concentration is 25 ppm for the first 40 days and 0 for 20
days thereafter.

Dieldrin residues in blood and adipose lipid (Figure 5) are compared for constant and
increasing adipose tissue compartments. Substantial and significant differences are appar-
ent. Although residues rise for the f~rst few days to about the same levels, the oscillations
of blood residues in the growth model become increasingly dampened. During this time,
the dilution of growth constitutes less than 3% of the adipose lipid. At the end of 40 days
exposure, the residues in the lipid of the "growing" rat are much lower (55 as compared
to 65/ag/g lipid); however the body burden of the growing rat is correspondingly higher
(Keane and Zavon 1969) than that of the constant Case I. The continuous addition of


Osborne--Mendel strain of rat

A Total adipose (grns) = 467 (B.W')
Dietary fat 60% by wt. ~ 9

\ ./



Total adipose (gms) = 99 (B.W.)
Dietary fat 2.7% by wt.

li t I I I I I | II 1 I I I I LIII I I I ! I II I
0.01 0.1 1.0 10.0
Body Weight (B.W.) Kg

Fig. 4. Log-log plot of Osborne-Mendel strain rat adipose tissue as a function of the body
mass (Schemmel e t al. 1970b).
176 F.T. Lindstrom et al.

lipid mass not only dilutes the adipose residue, but retards redistribution to other tissues.
Even after 40 days a true "steady state" is not achieved in the growing rat; the mean
blood and adipose lipid dieldrin residues are still rising. Residues at day 60 are about
15/~g/g of lipid in the growing model and 5.5 #g/g of lipid in the constant animal (Case I),
further reflecting retention of dieldrin in adipose and prevention of its metabolism in the

The average total body burden at 40 days is about 3.2 mg for Case I and about 3.36 mg
for the growing adipose case. Whereas this difference is only 5% at day 40, by day 60 the
average body burdens have dropped to about 0.26 mg and 0.88 rag; the difference then is
about 240%. One obvious implication of this situation is that attempts to free dieldrin-
contaminated individuals of residues during a period of adipose growth, by metabolism
alone, may not be as effective as during times of no growth.

A second implication is that the effects of growth simulated in this limited fashion by
the preliminary model cannot account for disparities between the observations of Barren
and Walton (1971) and this model. Although the decay curves for the growing compart-
ment model are more curvilinear than those for the constant compartment case (not
easily seen in Figure 5 due to figure reduction), no combination of adipose compartment
growth rate, dieldrin-metabolism rate (KD 1) and induction coefficient (tip 1)' and enzyme
catabolic rate (kCp 1) could be postulated which produced a simulation with: a) a negative
slope to the residues (plotted vs. time) at the terminal portion of the exposure prior to
cessation of exposure; b) the experimentally observed residue level at that point; and
c) the observed residue level 20 days after exposure ceased. It would thus appear that the
hypothesis of a strictly all tissue flow-limited a priori model may be in error. However,
sparse literature data relating parameter~ to specific time portions of the simulation are
the primary limitation. For example, how was the proportion of lipid (fF) changing in
the adipose tissue of the experimental rats? Is the model in error or are subordinate

o 100 rBIood


-~. 10

_.1 1.
5 10 15 20 25 30 35 40 45 50 55 60
Time (days)

Fig. 5. Comparison semilog plots of adipose and blood lipid phase dieldrin residues. Top
curves are for the constant body mass 300-g mature male rat. Bottom curves are for the
growing adipose and blood compartments (linear growth). Growth begins at t = 0 with
initial body mass being 300 g.
Applications of Dieldrin Distribution Model 177

assumptions inadequate? It would appear that this latter situation - the lack of appro-
priate subordinate assumptions and data - prevents validation of this model in every
instance. More probably, the system is neither flow-limited nor membrane-transport
limited, but is in fact a combination of both transport mechanisms (Dedrick e t al. 1973).

Case llI. Comparison of model to experiments in dieldrin pharmacodynamics. Daily

doses of lr given to adult male Osborne-Mendel rats previously fed unlabeled
dieldrin for eight weeks, indicated that there was a significant and progressive increase in
the specific activity (dpm/mmole) of adipose tr for several days post-
exposure, even as the tissue dieldrin level declined (Barton and Walton 1971). Those
authors concluded that some form of intra-organ compartmentalization must be occur.
ring or there was an " . . . orderly active turnover..." (vectorial type) of dieldrin such that
the " . . . material deposited first may be the first mobilized." On the other hand, the
preliminary model for dieldrin pharmacokinetics, because of the assumptions of flow
limitation and behavior of each compartment as a well-stirred chemical reactor, predicts
that the specific activity of the labeled dieldrin would be constant in any given compart-
ment throughout the post-exposure period. Intra-organ compartmentalization could and
probably does occur as a result of membrane transport limitations (in contrast to flow
limitations). It could also result from specific binding to subcellular components (in
contrast to generalized lipid solubilization). Alterations in specific activity of labeled
dieldrin in any compartment might, therefore, arise from the subsequent release and
redistribution of dieldrin compartmentalized elsewhere.

Evidence for inter-organ compartmentalization of dieldrin residues has appeared

sporadically. Zabik and Schemmel (1973) found relatively high dieldrin levels in liver,
kidney, and blood (on a lipid basis) in grain-fed Osborne-Mendel male rats receiving
0.8 mg of dieldrin/day for one week. Various adipose tissues (Table V) were about equal
on a lipid residue basis, being lower than liver, etc., but considerably higher than brain.
Obese rats on a high fat ration in the same study showed similar trends, but to a lesser
degree. However, a "steady state" of dieldrin metabolism and excretion equal to intake is
not achieved by this dosing schedule, so that the observed differences in tissue distribu-
tion might arise simply from flow limit considerations and physiological status.

Illustrative of the difficulties in developing good pharmacodynamics and pharma.

cokinetic understanding for pesticide disposition are the contrasting observations of the
apparent half-life of dieldrin in this strain of rats [30 days, based on a single determina-
tion six weeks post-exposure (Zabik and Schemmel 1973) and 4.5 days, based on several
determinations during a 20-day post-exposure period, (Barron and Walton 1971)].
Furthermore, the study by Zabik and Schemmel (1973) is in conflict with the results of
the much longer and larger (but less exhaustive in terms of numbers of tissues examined)
investigations of Walker et al. (1969), whicll serves as a primary basis for the preliminary

The extent of compartmentalization might be affected by specific binding of dieldrin

to tissue or subcellular components. Yadrick et al. (1971) reported a positive correlation
between dieldrin residues and the phospholipid content of specific leg muscles of dieldrin-
178 F.T. Lindstrom et al.

dosed swine. If a specific interaction of dieldrin were occurring with phospholipid, as in

the case of DDT/lecithin (Tinsley et al. 1971), the increased binding of dieldrin in
phospholipid-rich tissues could be responsible for intra- or inter-organ compartmentaliza-
tion. However, a comparison in Table V of the results of Zabik and Schemmel (1973) to
values of total lipid and phospholipid in these tissues gives no indication of any such
correlation. One might therefore ask, what range of flows afford the opportunity, at
least in the short run, for marked differences to occur between similar tissues (e.g.,
specific muscles) in residues expressed on a lipid basis?

One can hypothesize that muscles high in phospholipid are so due to the presence of
proportionately greater amounts of sacrosomal and plasma membranes consistent with
greater metabolic activity, and therefore these muscles could be expected to be more
highly perfused. If the differential perfusion rate were sufficiently great, and if the lipid
composition differences are solely related to perfusion-controlling factors, then under
the regimen used to generate the data of Yadrick et al. (1971) (nine doses in a 13-day
period) it might be possible to simulate residue differences between tissues or muscles in
the preliminary model in some relationship to the phospholipid content. Attempts with
the preliminary model to generate simulated residues that demonstrate dependence on
perfusion rates beyond the first few hours after each dose have been unsuccessful. For
example, a three-fold difference in perfusion rate between two muscles having a similar
three-fold difference in percent phospholipid (and an inverse three-fold difference in

Table V. Relationship o f dieldrin residues to lipid content in various tissues

Tissue Total lipid a Phospholipid a Dieldrin residue b

(mg/g tissue) (mg/g tissue) (/ag/g lipid)
Grain fed Obese

Liver 60 25 33 8

Kidney 50 24-27 28 9

Brain 97 45-47 3 2

Heart 27 (13-15) c 19 11

Adipose 600-900 20 11 4

Blood - - 30 7

Plasma 2.3 0-4.50 0.83 - -

aFrom Long (1961).

b From Zabik and Schemmel (1973); grain-fed rats received 3% (w/w) fat in diet, obese
rats received 60% fat in diet. Both received 0.8 mg/day for 1 week.
c Estimates.
Applications of Dieldrin Distribution Model 179

total lipid) yields residues differing by 133% after one hour, less than 5% by the fifth
hour, and less than 1% after the full 13-day regimen. Thus, it would appear to be neces-
sary for perfusion rates to differ by at least an order of magnitude, and probably more, in
order to be effective in generating residue differences over the three-fold range found by
Yadrick et al. (1971). Since the muscles examined do not differ that greatly in phospho-
lipid concentration, it would appear that the association of high percentage phospholipid
with high residues may be related through some factor other than increased vasculariza-
tion or simple binding of dieldrin to phospholipid.

For example, if the phospholipid-rich muscles had relatively more of a dieldrin-binding

protein or a protein facilitating the movement of dieldrin into the cell, such that this in-
creased protein level served to establish a differential permeability or accumulation, it
would be possible for compartmentalization to be evidenced in relation to phospholipid
concentration. If instead of viewing each tissue as a single well-stirred reactor or tank, we
view the tissues as consisting of two or more tanks separated by a permeable boundary
(membrane), and if the effective forward and backward transfer rate coefficients between
the extracellular compartment and the intracellular compartment(s) are unequal and
both much smaller in magnitude than the respective blood lipid equivalent flow (Q)
through the external compartment, then even at "steady state" conditions the internal
compartment can exist at residue levels significantly higher or lower than the blood
residue level. The specific physiologic and biochemical bases for such a model formula-
tion for dieldrin and similar highly lipo-soluble chemicals have not been established, and
it is quite possible that the various tissues will evidence different rate-limiting mechanisms
(flow limited, membrane-permeability limited, intracellular binding, etc.). A theoretical
consideration of such a model is in preparation, but realistic simulations based on the
same type of a priori evidence as the preliminary model appears to be much in the future.

The question at hand, however, is whether the preliminary model is sufficiently

accurate to provide predictive and interpretive assistance with studies of dieldrin
pharmacokinetics and effects. Even though the model does not provide mechanisms
whereby intra-cellular compartmentalization is possible, as would be required for simula-
tion of the changed specific activity in the study by Barron and Walton (1971), it is
possible to simulate the adipose tissue residues found in that experiment. Figure 6 shows
two typical curves produced by simulation of the adipose residues in a 300-g male rat
administered dietary dieldrin for only 30 days, then treated according to the protocol of
Barron and Walton. The curves differ only in regard to the assumed value of tip 1" It was
assumed that dietary dieldrin was consumed and absorbed over a 12-hr period and that
encapsulated dieldrin was absorbed in one hr, that no growth was occurring, that bile
flows as is shown in Table I, and that diurnal cycles of hepatic enzyme activity and blood
lipid can be ignored. The data of Barton and Walton (1971) are plotted on the same ex-
perimental time scale by conversion of their" whole adipose tissue residue data to adipose
lipid residues assuming 63.7 g of lipid per 100 g of adipose tissue. Except for the point
seven days post-exposure (day 37 on Figure 6), the data appear to be particularly well
simulated by the model employing weak auto-induction (tip1 = 1.0). One other feature
of the data from Barron and Walton (1971) is that animals continued on dietary dieldrin
developed residues that tended to form a negatively sloped straight line in the log (adipose
180 F.T. Lindstrom e t al.

lipid dieldrin) vs. time plot. No realistic growth rate of adipose tissue alone could afford
simulation of such data. However, if the efficiency of absorption (0) from the ingestion
were to decline with age, then declining residues would occur in spite of constant food
intake. - --

Viewed from the point of the adipose tissue residues, the preliminary model of lipid
phase dieldrin pharmacokinetics behaves as if intracellular and inter-organ compart-
mentalization, whether by differential tissue uptake, selective binding, or " . . . orderly
active turnover...", is a factor of lesser significance than the flow-limiting considerations
inherent in this system. Excellent representations of adipose tissue residue accumulation
with exposure to dieldrin and decay with subsequent cessation of exposure under a
variety of dosage regimens can be simulated for the mature male rat. Auto-induction of
dieldrin metabolism appears to play a relatively minor role in the disposition of ingested
dieldrin. Current physiologic growth data is either lacking or inadequate to simulate
accurately the effects of growth on dieldrin lipid-phase residues, although the preliminary
model behaves generally as predicted from studies such as those of Keane and Zavon
(1969). It would appear also to be useful in evaluating certain hypotheses regarding even
some types of compartmentalization.

It is in the area ofintra-organ and inter-organ lipid-phase residue differences, observed

by some workers cited earlier and in previous and current studies in this laboratory, that
the preliminary model reveals the inadequacies of our understanding both of the intimate
molecular interactions of dieldrin and of the components with which those interactions
occur: membranes, phospholipids, and specific proteins. In addition to requirements for
basic physiological and biochemical parameters of growth and intestinal and cardio-
vascular function with age/sex/nutritional status/species/other chemical exposure, it
would appear that the research needs in this area of disposition of highly lipid-soluble
drugs and chemicals will be served most effectively by a model which in addition to flow





" ' 10 ,5 2'5 I 5o
On dieldrin (500/~g/12 hrs/day) Time (days) 30 Off dieldrin

Fig. 6. Semilog plots for two hypothetical cases of feed-up and feed-off, illustrating the
effect that auto-induction has on long time adipose tissue lipid phase dieldrin residues.
Barton and Walton's (1971) feed-off data (vertical bars) plotted for comparison.
Applications of Dieldrin Distribution Model 181

limits, incorporates features both of lipo-protein binding and of membrane transfer,

such as has been performed for methotrexate pharmacokinetics (Dedrick et al. 1973).
Unfortunately, this iricrease in accuracy and sophistication of the dieldrin model would
be accompanied not only by increased complexity of compartmental design (two or
more compartments per tissue) but also by a geometric increase in the number of un-
known (and some seemingly immeasurable) parameters of the model sub-systems. Over-
coming such difficulties would appear to be critical, and therefore of high priority, in the
continuing evaluation of the human health hazards and environmental implications of
exposure to highly lipophilic materials.


This research was supported by U. S. Public Health Service Grant ES 00040 from the
National Institute of Environmental Health Sciences. This paper is issued as Technical
Paper No. 3787 from the Oregon Agricultural Experiment Station.


Barron, R. L., and M. S. Walton: Dynamics of HEOD (dieldrin) in adipose tissue of the
rat. Toxicol. Appl. Pharmacol. I8, 958 (1971).
Bischoff, K. B., and R. G. Brown: Drug distribution in mammals. Chem. Eng. Prog.
Symp. Ser. A. 66, 32 (1966).
Bischoff, K. B., and R. L. Dedrick: Generalized solution to linear two-compartment,
open model for drug distribution. J. Theor. Biol. 29, 63 (1970).
Bischoff, IC B., t L L Dedrick, D. S. Zaharko, and J. A. Longstreth: Methotrexate pharma-
cokinetics. J. Pharm. Sei. 60, 1128 (1971).
Cole, J. F., L. M. Klevay, and M. R. Zavon: Endrin and dieldrin: A comparison of hepatic
excretion in the rat. Yoxicol. Appl. Pharmacol. 16, 547 (1971).
Davis, J. L., R. O. Morris, and I. J. Tinsley: Quantitative measurement of microsomal
subfractions isolated from livers of rats fed with 1,1,1.trichloro-2,2-bis(p-chloro-
phenyl) ethane (DDT). Biochem. Pharmacol. 22, 869 (1973).
Davis, J. L.: Personal communication (1973).
Deddck, R. L., and K. B. Bischoff: Pharmacokinetics in applications of the artificial
kidney. Chem. Eng. Prog. Syrup. Ser. A, 68, 32 (1968).
Dedrick, R. L., D. S. Zaharko, and R. J. Lutz: Transport and binding of methotrexate
in vivo. J. Pharm. Sei. 62, 882 (1973).
Durham, W. B.: Body burden of pesticides in man. Ann. N. Y. Acad. Sei. 160, 183
182 F.T. Lindstrom et al.

Ghazal, A., W. Koransky, J. Portig, H. W. Vohland, and I. Klempan: Bescheuhigung yon

Ent~ftungsreaktionen durch verschiedene Insecticide. Arch. Exp. Pathol. Pharma-
kol. 249, 1 (1964).
Gillett, J. W., and T. M. Chan: Cyclodiene insecticides as inducers, substrates, and inhibi-
tors of microsomal epoxidation. J. Agr. Food Chem. 16, 590 (1968).
Heath, D. F., and M. Vandekar: Toxicity and metabolism of dieldrin in rats. Brit. J.
Industr. bled. 21, 269 (1964).
Keane, W. T., and M. R. Zavon: The total body burden of dieldrin. Bull. Environ. Contain.
Toxicol. 4, 1 (1969).
Lindstrom, F. T., J. W. Gillett, and S. E. Rodecap: Distribution of HEOD (dieldrin) in
mammals: I. Preliminary model. Arch. Environ. Contam. Toxicol. 2, 9 (1974).
Long, C. (ed.): "Biochemist Handbook," D. Van Hostrand, Princeton, NA. 1192 pp.
Matthews, H. B., and F. Matsumura: Metabolic fate of dieldrin in the rat. J. Agr. Food
Chem. 17, 845 (1969).
Ortega, P.: Tissue changes caused by DDT and dieldrin, in "The Biological Impact of
Pesticides in the Environment," Environmental Health Sciences Series No. 1, J. W.
Gillett, ed., Oregon State University, Corvallis, OR, p. 111 (1970).
Schemmel, R., O. Miekdsen, arid L L. Gill: Dietary obesity in rats: Body weight and
body fat accretion in seven strains of rats. L Nutr. 100, 1041 (1970a).
Sehemmd, R., O. Mickelsen, and V. Mostosky: Influence of body weight, age, diet, and
sex on fat depots in rats. Anat. Rec. 66,437 (1970b).
Spector, W. S. (ed.): Handbook of Biological Data. Fed. Am. Soc. Exp. Biol. Saunders,
Philadelphia, 584 pp. (1956).
Tinsley, I. J., R. Haque, and D. Sehmedding: Binding of DDT to lecithin. Science 174,
145 (1971).
Walker, A. I. T., D. E. Stevenson, J. Robinson, E. Thorpe, and M. Roberts: The toxicology
and pharmacodynamics of dieldrin (HEOD): Two-year oral exposure of rats and
dogs. Toxicol. Appl. Pharmacol. 15, 345 (1969).
Wauchope, D. R.: Personal Communication (1972).
Yadrick, M. K., M. E. Zabik, and F. Funk: Dieldrin levels in relation to total, neutral, and
phospholipid composition in selected pork muscles. Bull. Environ. Contam. Toxicol.
8, 289 (1971).
Zabik, M. E., and R. Schemmel: Dieldrin storage of obese, normal, and semistarved rats.
Arch. Environ. Health 27, 25 (1973).

Manuscript received March 20, 1974; accepted May 20, 1974