[CANCER RESEARCH 46, 3969-3978, August 1986]
Pharmacokinetics of Monoclonal Immunoglobulin d, F(ab')2, and Fab' in Mice
David G. Covell, Jacques Barbet, Oscar D. Holton, Christopher D. V. Black, R. J. Parker, and John N. Weinstein
Laboratory of Mathematical Biology [D. C. C., J. B., O. D. H., C. D. V. B., J. N. W.] and Office of the Chief, Division of Cancer Etiology (R. J. P.], National Cancer
Institute, NIH. Bethesda, Maryland 20892
ABSTRACT
The pharmacokinetics of an immunoglobulin Gl (IgGl ) and its F(ab')2
and Fab' fragments following i.v. administration in mice has been studied
by constructing a physiologically based, organ-specific model to describe
antibody biodistribution, catabolism, and excretion. The antibody selected
for study (MOPC-21) has no known binding sites in the body and
therefore is useful for defining antibody metabolism by nontumor tissues.
Whole IgG (a) remains in the body for 8.3 days, the majority of time in
the carcass (53.0% of the total residence time); (b) has a distribution
volume exceeding that of plasma plus interstitial fluid; (c) distributes into
these volumes rapidly for most enterai organs (equilibration time
<2.6 min for liver, spleen, kidney, and lung), slower for the gut (<20
min), and slowest for carcass (<260 min); (il) produces interstitiahplasma
concentration ratios of >0.5 for enterai organs and 0.18 for carcass; (c)
has the greatest percentage of its catabolism due to the gut (72.8%),
followed by the liver (20.5%), then the spleen (3.6%); (/) has the highest
extraction on a single pass by the gut (0.14%) and (K)cycles through the
interstitial spaces of the body at least 2.8 times/g of organ weight before
being metabolized or excreted. When compared with whole IgG, the Fab'
fragment (a) is cleared from the body 35 times faster; (/>)has a larger
total distribution volume; (c) distributes more rapidly into this volume;
(d) produces higher interstitiahplasma concentration ratios; (c) is catabolized principally by the kidney (73.4% of total catabolism), followed
by the gut (22.9%), then the spleen (3.1%); (/) is extracted from the
circulation to the extent of 3.4% on each pass through the kidney, and
less by gut (1.0%) and spleen (0.14%) and (?) cycles through non-kidney
interstitial spaces at least 0.4 cycles/g of tissue weight before metabolism
or excretion. The F(ab')2 fragment has pharmacokinetic characteristics
that fall between those of whole IgG and Fab'. These results (a) provide
pharmacokinetic criteria for selecting whole IgG, I Uili'h. or Fab' for
various in vivo applications; (b) provide a framework for predicting
cumulative tissue exposure to antibody labeled with different isotopes;
and (c) provide a reference metabolic state for the analysis of more
complex systems that do include antibody binding.
INTRODUCTION
Recent advances in immunology and nuclear medicine have
led to improvements in tumor localization with radiolabeled
monoclonal antibodies (1-9). Although many achievements in
the in vivo application of radiolabeled antibodies can be cited
[see reviews by Gallagher (10), Burchiel (11), and Sfakinakis
(12)], the development and clinical availability of monoclonal
antibodies specific for most tumor-associated antigens rank
close to the top of the list (13-20). Nonetheless, the amount of
radiolabeled antibody localized in the tumor relative to nontarget tissues (defined as the uptake ratio) continues to limit
the in vivo utility of this methodology (21, 23).
In an attempt to overcome this limitation, sophisticated
computer-aided techniques have been developed for background
subtraction of radioactivity not associated with tumor (24-26).
This approach often uses coinjection of a nonspecific control
protein, such as radiolabeled albumin, and blood-pool subtrac
tion for the region(s) of interest. Critics have pointed out that
this approach may expose the patient to additional radiation
(24), may add potential sources of error (7, 26), and must be
Received 2/12/86; revised 4/30/86; accepted 5/6/86.
The costs of publication of this article were defrayed in part by the payment
of page charges. This article must therefore be hereby marked advertisement in
accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
conducted by those expert in the technique (27).
Recent research suggests that the antibody fragments F(ab')2
and Fab' may improve the tumor uptake ratio (27-30). Inves
tigations to date, which have been limited but encouraging,
suggest that most of the improvements in tumor uptake ratio
are due to a more rapid plasma clearance of antibody fragments,
producing lower background levels (11, 27).
The levels of background radioactivity, which when elevated
tend to lower the tumor uptake ratio, are determined by the
kinetic factors of antibody biodistribution, catabolism, and ex
cretion (22). Therefore it would be useful to know how these
pharmacokinetic factors compare among whole IgG and its
fragments. Most previous analyses of the metabolism of anti
bodies have lumped various body pools into a single (e.g.,
intravascular or extravascular) compartment (31-33) and thus
have not permitted close examination of regions of the body
that are known to have different antibody concentrations.1
Studies in which individual tissues have been examined for
antibody have pointed to the need for a better understanding of
antibody metabolism by the various organs (34-36).
An alternative approach for investigating how kinetic factors
influence levels of background radioactivity is to develop a
physiologically based, organ-specific model that describes the
pharmacokinetics of radiolabeled antibody. A relatively simple
model, which we will report in detail here, defines the phar
macokinetics of an i.v. injected murine, homologous whole
IgGl (MOPC-21) and its F(ab')2 and Fab' fragments in a mouse
system that has no known antigen binding sites. This system
avoids the additional complexities related to specific binding of
antibody to tumor antigen. The resulting model will be used in
the following three ways: (a) to define quantitatively the phar
macokinetic differences between whole IgG and its fragments;
(b) to examine how these pharmacokinetic differences may be
used for improved in vivo use; and (c) to provide a reference
metabolic state for the analysis of more complex systems that
do include antibody binding.
MATERIALS
AND METHODS
Experimental Background
Our laboratory has been conducting experiments on the fate of i.v.
administered iodine-labeled IgG, I (ah )•
fragments, and Fab' fragments
of an anti-LYB8.2 monoclonal antibody (clone CY34) (37) and of the
MOPC-21 myeloma protein in mice. The anti-LYB8.2 antibody and
its 1 ial>').,and Fab' fragments bind to an allelic determinant on mouse
B-lymphocytes, whereas the MOPC-21 antibody has no known target
antigen in the mouse. Data from these studies have been used to develop
a compartmental model of the pharmacokinetics of MOPC-21 and its
fragments. Complete experimental details can be found in Holton et
al} and will be discussed only briefly here. One-tenth ^g of either whole
l3'I-IgG or its 13ll-labeled fragments, the latter prepared according to
the method described by Parham and coworkers (38), tested by sodium
dodecyl sulfate:polyacrylamide gel electrophoresis (39) and purified by
' B. Rhodes, S. Durcheil, K. Breslow, and D. Laven, unpublished observations
(see Ref. 11).
2O. D. Holton, C. D. V. Black, R. J. Parker, D. G. Covell, J. Barney, S. M.
Sieber. M. J. Talley, and J. N. Weinstein. Biodistribution of monoclonal IgGl,
F(ab'>2, and Fab' in mice after i.v. injection: a comparison between anti-B-cell
(anti-LyB8.2) and irrelevant (MOPC-21) antibodies, submitted for publication.
3969
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1986 American Association for Cancer Research.
IgG KINETICS IN MICE
high-performance liquid chromatography, was injected via the tail vein
into C57BL/6 mice. Groups of 4 mice were sacrificed serially over the
next 24 h, and measurements were made of antibody in plasma, gut,
liver, kidney, spleen, lung, and carcass. The total body recovery and
plasma TCA3-precipitable fraction of label were also determined. Av
eraged (n = 4) data at each time point were used for model development.
Data from longer term studies on plasma clearance were also used. The
present paper will focus only on the data for MOPC-21 IgG and its
fragments.
Model Development. The construction of a useful model of nonspe
cific antibody metabolism must be firmly based on the major physio
logical/pharmacological features of the system (40). Since we can never
expect to have enough details to model the entire system, a simpler
model is proposed that contains those unit processes which appear to
be important globally in antibody pharmacology. For this analysis our
model includes the processes related to (a) circulation of plasma, (/>)
exchange of antibody across the capillary wall, (c) return of antibody
from the interstitial space to the bloodstream via lymph, (</>interaction
of antibody with "cell-associated" material, and (e) formation of meta
bolic products. The "cell-associated" regions of the model represent
antibody interaction with non-plasma, non-interstitial spaces, which
may include direct contact with cells or with matrix within the tissue.
Our data do not permit a distinction between the two; hence, we will
refer to such interactions as "cell associated." Cell-associated antibody
can be irreversibly converted to non-antibody metabolites. The five
processes listed above define the rate of antibody transfer between
capillary plasma, interstitial fluid, and the cell-associated space for each
organ in the model. This analysis is comparable to that of fitting a
three-compartment linear model to the data for each organ. All of the
organs are linked together in a single model, hence permitting analysis
of kinetic interactions among the various organs. Consequently, most
of the complexity of the proposed model is due to having data from a
relatively large number of organs (n = 6) rather than from specification
of a complex model for each organ. A detailed description of the model
can be found in "Appendix li " Definitions of model parameters are
contained in "Appendix A."
Three model parameters for each organ were treated as unknowns
whose values were determined by fitting the model to the data. These
were (a) the PS (ml/min), which describes the movement of antibody
between capillary plasma and interstitial fluid, (In the I, (ml), which
describes the apparent distribution volume of antibody in the nonplasma and non-interstitial fluid volumes in each organ, and (c) the
A.i..-.(ml/min), which defines the conversion of antibody to non-anti
body metabolites for each organ. The remaining model parameter
values were obtained from published literature (see "Appendix B" and
Ref. 41).
The mass balance equations for the proposed antibody model were
simulated and analyzed using the CONSAM computer program (42,
43). Estimates of unknown model parameters were obtained by adjust
ing their values such that the variance-weighted squared error between
model simulations and experimental measurements was minimized.
The variance of each parameter estimate was obtained from the covariance matrix using standard weighted least-squares methods (44).
Calculations. MRT of antibody in each organ defines the average
time that antibody molecules entering the system via i.v. injection
remain in that organ (45, 46). MRT is equivalent to the "area under
the curve" for a plot of fraction of dose in each organ against time (47).
The sum of organ mean residence times equals the average time that
antibody remains in the animal. Consequently, MRT measures the
cumulative exposure to antibody that a tissue site receives following
injection. In terms of interpreting tumor uptake ratios, nontumor
regions with high exposure to antibody will have greater background
levels and thus require greater localization of tumor-specific antigen
for positive immunodetection. Regions with low exposure will produce
low background levels. However, low exposure may also indicate limited
access of specific or nonspecific antibody to that region. MRT will be
3 The abbreviations used arc: TCA, trichloroacetic acid; PS, permeability
surface area product; Vf,volume of the cell-associated compartment; MRT, mean
residence time; /:/'. interstitiakplasma ratio; I,,, time to reach steady state.
used to quantify these effects. Details of MRT calculations can be found
in Ref. 47.
A measure of tissue catabolism is the extraction of antibody from
capillary plasma on a single pass through an organ. Extraction
is defined as
= 1.0 —rau-,..,, nui-
(A)
and will be calculated for each organ by model simulation.
One goal of tumor localization is to maximize uptake by tumorspecific antigen and to minimize nonspecific background. A factor that
affects the uptake of tumor-specific antibody is whether the antibody
has sufficient opportunity to bind to the target antigen. Clearly, getting
to the antigenic site is a prerequisite to binding. From our model it is
possible to calculate the average number of times that an antibody
molecule can be expected to pass through each compartment. This
process is referred to as cycling, and the net number of cycles of antibody
through each region in the model is calculated as the product of the
compartmental mean residence time (min) and the compartmental
turnover time (min"') (47). Regions where net cycling is high will have
the greatest opportunity for specific antibody to locate and presumably
bind to antigen.
Determination of Unknown Model Parameters. Having constructed a
mathematical model to describe antibody metabolism it was necessary
to determine whether the unknown parameters of the model could be
estimated from the available data (48). This issue, commonly referred
to as "the inverse problem," has been discussed in detail by Jacquez
(49). Its solution involves adjusting the values of the unknown param
eters until the model can accurately simulate the experimental data. If
the data are sufficiently rich in information content (SO) and the
parameters are appropriately chosen, the unknown parameter values
can be uniquely determined from the experimental data Theoretical
parameter identifiability for linear systems has been thoroughly docu
mented (51). Unfortunately, the proposed theory for large linear sys
tems is not easily applied.
As an alternative a parameter sensitivity analysis can be performed
(45). This analysis involves determining whether changes of unknown
parameters produce appreciable changes in the model's response to a
standard input. If the parameter change produces a large change in the
simulated response, that parameter is considered to be a sensitive
parameter. Sensitive parameters are more likely to be estimated accu
rately from the data than are less sensitive ones. In this study it was
necessary to determine whether changes in the parameters PS, Vc,and
A.I..,.produce large changes in the model simulations.
Fig. 1 shows the effects on model simulations of changing values for
the three unknown model parameters for a typical organ. From this
analysis it is apparent that an increase in PS results in a corresponding
increase in the peak value of organ radioactivity. An increase in V„
which results in a decrease in the rate constant (R/Vf) for antibody
transfer from the cell-associated compartment, results in a slower fall
of the curve at longer time points. An increase in the catabolic clearance
constant, A'..,,,,,,results in a much more rapid fall off of radioactivity
after the peak value. These simulations suggest that the unknown model
parameters produce increases or decreases in the simulation curves at
A. ORGAN
B. TCA
Fig. 1. Sensitivity plots for IgG in a representative organ over time ( /) and
the corresponding fraction of radiolabel in the plasma that was TCA precipitable
(b). Simulations are for a reference set of parameters (
), and a 5-fold increase
in V, (
), PS (
), or Ä«ii„
(—) for that organ. Parameter values for
kidney IgG pharmacokinetics are used as reference values.
3970
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1986 American Association for Cancer Research.
IgG KINETICS IN MICE
very different temporal locations. Accordingly, the proposed model
appears capable of distinguishing whether changes in PS, KfKm,or V,
occur with different antibody fragments. Although this analysis does
not guarantee parameter identifiability, it does demonstrate that, in the
proposed model, the unknown parameters are sensitive.
RESULTS
The proposed compartmental model was capable of simulat
ing all of the experimental data (see Fig. 2). The model param
eters used for each simulation are listed in Tables 1 and 2. The
model-derived values for the unknown parameters PS, KMm,
and Vccould be estimated from the data with reasonable pre
cision. The coefficient of variation for estimated parameters
ranged from a low of 10% to a high of 120%. However, most
of the parameters had a coefficient of variation around 45%.
The values of A"e|imwerethe most difficult to obtain from fitting
our data. This is presumably the result of not having direct
measurements of non-antibody metabolites in each organ.
The PS determines the transcapillary movement of antibody.
Larger PS values produce greater flux. On a per gram organ
surprisingly high value for lung may be due to its extensive
capillary network. The large values of PS per gram organ weight
for liver and spleen are consistent with the existence of fenestrations in the walls of sinusoids in these organs (52). Values
of PS per gram weight for the other organs decreased in the
following order: kidney (0.09 ml-min~'g~'), gut (0.006 mlmiiT'g"1), and carcass (0.0003 ml-miir'g"1). In a few instances
our parameter estimates for PS can be compared with published
values. Garlick and Renkin (53) studied the transport of large
dextran molecules and albumin from plasma to interstitial fluid
and lymph in the hindlimb of dogs. A very satisfactory agree
ment is observed between our PS values for carcass and their
estimates for molecules with equivalent Stokes' radii [e.g.,
/'•Sdexiraniio
= 0.0023 ml/min versus PSisC = 0.0036 ml/min;
Wdextra„2o
= 0.0061 ml/min versus WF(ab'>2= 0.0041 ml/min;
/'•Sdextranio
= 0.029 ml/min versus PSf^, = 0.050 ml/min.
The general trend of increasing PS with decreasing fragment
size is a consistent feature of our results. The liver was an
exception, with PS being greatest for whole IgG and decreasing
B
008
0.06
weight basis, PS was greatest in the lung (0.53 ml min~'g~'),
liver (0.35 ml-min~'g~'). and spleen (0.20 ml-miir'g"1). The
0.08006(n§
Whole IgG —*
F lab' 12 l
Fab' —i
A—FIgG —
(ab')2
••••••Fab'
Whole IgG —*—
F lab' 12 ••
Fab' —T—
—T—~~.^^^-~--Jp^'•».\
006
5
u
0.04
0.04O
§ 0.04
O
Oû_0020.00
U
O
y 0.02
.'
0.02
1.0
0.00
0.00
0.0
500.0
1000.0
1500.0
0.0
500.0
0.004
Whole IgG —
F lab'12
Fab' —
0.012
1000.0
1500.0TIME 1000.0
0C»/i,-:Whole 500.0
1500.0
(minutes)
TIME (minutes)
TIME (minutes)
0.003
--
E
Whole IgG —*—
F lab' 12 •
Fab' —V--
Ev
0.16
Whole IgG —*—
F lab' )2 •
Fab' —T—
0.12
01 0.008
ui
_j
0.002
o
o
o- 0.004
0.000
008
0.001
0.0
0.04
10000
EIE
000
0.000
500.0
500.0
1500.0
TIME (minutes)
1000.0
0.0
1500.0
500.0
1000.0
1500.0
TIME (minutes)
TIME (minutes)
H
0.4
1.00
<
0.75
0.7S
0.50
O
z 0.50
O
O
Whole IgG —*
F lab'I2
•
Fab' —l
"K.
Whole IgG —AF lab'12 •
•••
Fab' ~<r-
0.25
0.00
500.0
10000
TIME (minutes)
1500.0
0.0
500.0
1000.0
TIME (minutes)
£ 0.25
0.00
1500.0
Whole IgG —*—
F lab'12
•••••
Fab' —T—
0.0
500.0
1000.0
1500.0
TIME (minutes)
Fig. 2. Experimental data and model simulations of IgG, F(ab')2, and Fab' for plasma (A), liver (A), gut (Q, lung (D), spleen (£),kidney (f), carcass (G), TCAprecipitable fraction in plasma (//), and total body recovery (I). The standard deviation for the early time points has been omitted for clarity. Points, mean; bars, SD.
3971
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1986 American Association for Cancer Research.
IgG KINETICS
smaller clearance values for whole IgG were observed for the
remaining organs (liver > carcass > spleen > kidney).
The values of catabolic clearance for F(ab')2 and Fab' frag
Table 1 Model parameters (literature values)
See "Appendix" and Ref. 40.
gValue
tissue0.12
ParameterPlasma
ments changed very little when compared with their values for
whole IgG in liver spleen and carcass. A 2-fold increase in Ke,im
for F(ab')2 and a 7-fold increase in Kenmfor Fab', when com
mly^mv*~v,«*™v,*«~,¡/'
vol,
0.130.10
0.050.01
0.100.05
0.200.28
0.020.03
0.240.400.98
J>,C*TC*U»wVu.y~*Interstitial
pared with whole IgG, was observed for gut. Fits to the gut
data for Fab' were poor, suggesting that a different approach
to the modeling of catabolism by the gut may be necessary. A
greater than 5000-fold increase over the whole IgG clearance
value for kidney was observed for both of the fragments. The
latter result is due to our inability to fit the data for F(ab')2 and
Fab' without introducing an irreversible loss from the cell-
0.050.12
ml>wy¡**Vu»mViAiÃ
vol.
mêyfW-Yi*mPlasma
0.130.11
0.060.01
0.090.05
0.190.61
0.040.03
0.200.49
ml/minÛ*.Q~a*»o**-,ß~-a-.Per
flow,
0.530.10
0.060.06
1.030.44
1.600.40
0.031.33
00.0Table
parametersParameter
2 Unknown
WholeGutPS°
0.011(0.103)***.'
0.273(1.05)
0.167(0.55)LiverPS
Vf
(0.39)*.„„0.324
(0.95)yt
0.039
(0.37)KidneyPS 0.030
(0.35)Anta 0.022
0.004(1.15)K,
(0.47)SpleenPS 0.022
0.012(0.81)A-«,im
0.008(1.03)yf
IN MICE
associated compartment in the kidney. This loss is apparently
related to catabolism of fragments. However, the non-antibody
metabolites produced via this pathway do not return to the
plasma but instead are excreted directly from the body. Con
sequently these non-antibody metabolites cannot contribute to
lowering the TCA-precipitable fraction in plasma. Regardless
of how the comparison is made, kidney clearance values for the
F(ab')2 and Fab' fragments were much greater than those for
10.23%oftotal12.49.70.65.028.23.141.0100.036.87.54.533.130.11
any other organ.
The parameter Vcdetermines the size of the cell-associated
compartment
in each organ. This term represents an apparent
model
volume of distribution of antibody in the regions that do not
include the plasma or interstitial volumes. Larger values of Vc
result in less material returning to the interstitial compartment
per unit time. Increases in Vc may be related to an enhanced
1.40(0.69)0.185(0.21)0.021
0.135(0.78)0.193(0.39)0.020(1.13)0.013(0.67)0.028
diffusibility of the smaller fragments into the interstitial matrix
or to a greater association of antibody with cellular material
(55). On an organ weight basis, Vcvalues for IgG were ordered
(1.10)0.033(0.19)3.08(1.15)38.3
as follows: liver (0.033 ml/g) < lung (0.062 ml/g) < spleen
(0.083 ml/g) < kidney (0.088 ml/g) < gut (0.094 ml/g) <
carcass
(0.31 ml/g). Vcvalues for liver, spleen, carcass, and lung
(0.23)21.0(1.06)0.046(1.04)0.010(0.41)0.021
were comparable between whole IgG, F(ab')2, and Fab'. Sub
(0.93)2.19(0.67)0.042(1.02)0.101
o.oo5(i.oi)CarcassPS
(1.08)0.006(1.05)0.004(0.89)0.011
(0.19)0.050
(0.30)AHI., 0.004
0.015(1.13)V,
(0.46)LungPS 4.67
(0.43)0.009(1.21)3.77 (0.22)0.031
(1.05)4.38(0.13)0.091
(0.34)0.074
(0.24)0.00.007
0.069(0.21)Af««.
(0.25)0.00.023
0.0V,
(0.43)Fab'0.023(0.16)1.98(0.18)
(0.09)
0.008 (0.32)F(ab')i0.010(0.103)0.481(1.05)
" Units of ml/min.
b Numbers in parentheses, one fractional SD and calculated from the covariance matrix using standard weighted least-squares techniques.
' Units of fil/min.
d Units of ml.
by 40% for both fragments. This reduction in PS indicates that
the transcapillary flux of antibody in the liver must depend on
factors other than molecular size. One possible interpretation
of this decrease in PS may be related to loss of Fc receptormediated uptake of whole antibody by Kupffer cells in the liver
(54). Additional investigations of this possibility are required.
The values of catabolic clearance (Kf¡im)
reflect each organ's
ability to degrade antibody that reaches the cell-associated
compartment. The clearance values for all organs were non
zero except for lung. That is, only in the case of lung did the
parameter iteration scheme move Ke[imto zero. The gut had the
largest value of clearance for whole IgG. Order of magnitude
stantial increases in Vcwere observed in the kidney and gut for
Fab' when compared with those for whole IgG and F(ab')2.
This suggested a preferential "trapping" of Fab' fragments in
the cell-associated compartment of kidney and gut.
An important consideration in immunodetection
is the
steady-state ratio of antibody concentration in the interstitial
space to that of circulating plasma. This ratio is related to the
concentration of antibody in the lymph (56). Calculated ratios
(I:P) for all organs of the model and the Ta are listed in Table
3. In all cases the /:/' ratio was less than 1 and indicated that
the steady-state interstitial concentration was less than the
plasma concentration. This observation has also been made for
7-globulins (57-59) and for dextrans (60). In addition, the
regional differences in the interstitial:plasma antibody concen
tration ratios predicted by the model have been observed for
dextrans, with organs such as liver and gut producing the
highest and carcass the lowest concentrations (61). The general
Table 3 Steady-state interstitial.-plasma antibody concentrations (I:P) and time to
reach steady state (T„min)
WholeOrganGutLiverSpleenKidneyCarcassLungI:P°0.540.970.870.660.180.68T0"19.20.71.02.
" Calculated by model simulation.
* Calculated taTa = (y,+ y,)/PS.
3972
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1986 American Association for Cancer Research.
IgG KINETICS
trend for smaller antibody fragments to produce concentrations
closer to that of plasma is consistent with expectations based
on molecular size. The interstitial concentration of Fab' is no
lower than 80% of the plasma concentration for all organs, and
for liver, kidney, and spleen, it is almost at concentration
equilibrium with plasma.
The time to reach steady state for whole IgG and F(ab')2 is
shortest for liver, spleen, and lung (<1.1 min), intermediate for
kidney and gut (<3.4 and <22.0 min, respectively), and longest
for carcass (<270.0 min). Organs reaching steady state in the
shortest times show only slightly smaller values of Tsswith the
Fab' fragment. The remaining organs have at least a 60%
reduction in Tssfor the Fab' fragment. In the case of carcass,
the 7"Hof 251 min for whole IgG was reduced to 17.7 min for
Fab'.
IN MICE
Table 4 Mean residence time per gram wet tissue weight, min/g
ParameterPlasma
(8.6)
Gut
1326.4(10.8)
36.2 (9.0)
31.5(17.2)
Liver
1424.9(11.6)
66.5 (8.7)
19.7(5.6)
Kidney
513.3(4.2)
11 I.I (3.9)
228.8(17.1)
Spleen
1352.4(0.7)
74.2 (0.6)
20.7 (0.3)
Carcass
431.9(53.0)
27.9 (59.7)
10.4 (49.6)
Lung
2320.5 (2.5)
114.9(2.1)
37.8(1.5)
BodyWhole2131.4(17.3)°
612.0(100.0)F(ab')2115.3(15.9)
35.1 (100.0)Fab'28.2 16.1 (100.0)
" Numbers ¡nparentheses, percentage of total body mean residence time for
each organ. The total body mean residence times were 8.3 days for whole IgG,
0.5 days for F(ab')2. and 0.24 days for Fab'. Individual organ residence time is
calculated as the sum of time spent in the organ's capillary plasma, interstitial
fluid, and cell-associated compartment. The average weight of the organs for the
mice used in this study (n = 120 mice) was: gut, 1.8 g (including contents); liver,
0.9 g; kidney, 0.2 g; lung, 0.13 g; spleen, 0.06 g; carcass. 15.3 g; and total, 20.1
g.
whole IgG, its F(ab')2 fragment, and its Fab' fragment relate
DISCUSSION
Our analysis has demonstrated that a simple model of anti
body metabolism, while being neither unique nor comprehen
sive, is useful for examination of the pharmacokinetics of a
nonspecific murine IgGl and its F(ab')2 and Fab' fragments in
mice. This model can be used to establish a set of pharmacokinetic measures for evaluating the in vivo use of radiolabeled
antibody. The results of this analysis apply most directly to the
situation of murine monoclonal used in healthy mice that have
very little tumor burden. Consequently, conclusions about back
ground levels, cumulative tissue exposure, and antibody access
to tumor antigen, as these quantities relate to in vivo immunodetection and therapy, apply to this case. Extrapolation of these
results to heterologous systems is not straightforward and will
require additional experimentation.
We believe that in the
future human monoclonal antibodies will be used for inumi
nodiagnosis and immunotherapy. It remains to be determined,
however, whether the mouse-mouse system is an appropriate
model for the human-human system.
Numerous factors will be important in choosing which anti
body to use for a desired application. The final decision will be
affected by biochemical factors such as antibody binding to
tumor cells, fragmentation, cross-reactivity, labeling isotope;
physiological factors such as blood and lymph circulation to
normal and tumor tissue, and the immune response to challenge
by foreign antibody; and pharmacological factors such as biodistribution, catabolism, and excretion. No attempt will be
made in this discussion to address issues beyond those of
antibody pharmacology. Furthermore, application of these re
sults to antibodies labeled with metal-chelates will require mod
ification of the model to account for cellular trapping of the
chelate.
Two measures will be used to quantify antibody pharmaco
kinetics for comparisons in vivo, (a) Mean residence time per
gram organ weight expresses antibody exposure for various
regions of the body. High exposure to tumor and low exposure
to normal tissues are desirable for both therapy and immunodetection. (b) Access of antibody to regions of the system where
tumor antigen may exist is a prerequisite of antigen-antibody
binding. Whether or not antibody actually binds to antigen
when given the opportunity will depend on the rate-limiting
step in the kinetics of the binding process. However, organs
with the highest access can be expected to have the most
opportunity to bind antigen located within that organ. The
number of cycles of antibody through the various regions of
each organ will be used as a measure of access. These two
measures will be used to examine how the pharmacokinetics of
to their use for immunodetection and therapy. Although we
have no direct evidence that these measures will be different for
different doses of antibody administered i.v., investigations on
the fate of s.c. administered MOPC-214 and s.c. administered
anti-H2K* (62) demonstrate no dose-dependent pharmacoki
netics.
The model-calculated mean residence times in the body for
whole IgG, F(ab')2, and Fab' of 8.5, 0.5, and 0.2 days, respec
tively (see Table 4), are in satisfactory agreement with published
literature (11, 31-33). Plasma, lung, and the enterai organs
liver, gut, and spleen have the highest levels of exposure to
whole IgG on a per gram weight basis. The kidney and carcass
have the lowest levels of exposure to whole IgG, despite the
fact that they account for over 57% of the total body mean
residence time (53.0% from carcass and 4.2% from gut).
The total body residence time for F(ab')2 was '/nth the value
for whole IgG. This is reflected by nearly equivalent reductions
for all organs in mean residence time per gram weight. That is,
the 17-fold decrease in organ exposure to F(ab')2, when com
pared with whole IgG, occurs in all organs, and as a result the
distribution of the total body residence time across all organs
does not change. Taken together these results suggest that
F(ab')2 produces an order of magnitude reduction in organ
exposure when compared with whole IgG.
The total body residence time of Fab' was 2-fold shorter than
that of F(ab')2. This factor of 2 reduction in total body antibody
exposure was found in all organs except the gut, where only a
12% decrease in exposure was observed, and the kidney, where
a 2-fold increase in exposure was observed. This result is
consistent with the shift in the distribution of total body resi
dence time for Fab', such that 34% of the total residence time
could now be accounted for by gut and kidney (17.2% for gut
and 17.1% for kidney).
The distribution of residence times between the capillary,
interstitial, and cell-associated compartments of the model is
summarized in Table 5. For all organs except carcass, the
capillary plasma compartment had the greatest fraction of the
total organ's residence time for whole IgG and F(ab')2. This
suggests that whole nonspecific IgG and F(ab')2 reside for the
most part in the plasma compartment of the body. For Fab'
most of the residence time is due to the cell-associated space
(except for the liver and lung, where most of the residence time
is due to capillary plasma). The general trend for distribution
of antibody within the capillary plasma, interstitial, and cellassociated spaces is consistent with the premise that the smaller
4 R. J. Parker, J. N. Weinstein, A. M. Keenan, S. K. Dower, M. A. Steiler, O.
D. Holton, and S. M. Sieber. Targeting of radiolabeled monoclonal antibodies in
the lymphatics, submitted for publication.
3973
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1986 American Association for Cancer Research.
IgG KINETICS
IN MICE
observed by others (66-68), may be related to the charge of
these proteins (69, 70).
The net number of cycles of antibody per gram tissue through
the interstitial and cell-associated regions of each organ in the
model is listed in Table 7. There is a general tendency for the
number of cycles for each organ to be greatest for whole IgG,
intermediate for F(ab')2, and lowest for Fab'. The best case, in
Table 5 Mean residence time distribution within organ expressed as percentage
of total mean residence time in each organ
Fab'
F(ab')2
Whole
Parameter
GutCapillaryInterstitialCell
associatedLiverCapillaryInterstitialCell
terms of frequent visits by antibody, is for whole IgG in the
liver where 2291 cycles through the interstitial space are pre
dicted. The worst case is for whole IgG in the carcass, where
only 43 visits are predicted. When this result is expressed on a
per gram basis, the difference is even more dramatic (2490
cycles/g in the liver versus 2.8 cycles/g in the carcass). Whether
or not the relatively low access of antibody in the interstitial
space of the carcass explains the difficulties of localization
within tumor when using whole IgG will depend, among other
factors, on how closely tumor vasculature compares with that
of muscle.
The tendency for the number of cycles to decrease for F(ab')2
and Fab' is consistent with the progressively shorter total body
retention for each antibody fragment. F(ab')2 cycles '/nth as
often, and Fab' cycles '/ssth as often as whole IgG for most
associatedKidneyCapillaryInterstitialCell
associatedSpleenCapillaryInterstitialCell
associatedCarcassCapillaryInterstitialCell
associatedLungCapillaryInterstitialCell
organs. There are two notable exceptions to these observations.
(a) The Fab' fragment cycles through the interstitial space of
the carcass 2 times more often than F(ab')2. This observation
associated42.423.933.845.443.511.051.433.515.138.934.027.123.68.967.454.829.715.546.525.727.849.045.65.549.628.721.637.329.433.319.227.952.958.132.89.012.89.577.745
may be explained on the basis of the greater transcapillary flux
of the smaller antibody fragment. Consequently, if the Fab' and
F(ab')2 fragments have equal affinity for tumor antigen, the
Fab' fragment would have a 2-fold greater chance of visiting
Table 6 Percentage of extraction per wet weight by catabolizing organs
OrganGutSpleen
(43.3)
0.14(3.1)
0.028 (3.3)
0.013(3.6)
0.013(0.3)
0.004(3.1)
0.008 (20.5)
Liver
3.39 (73.4)
0.059 (50.3)
0.0007(1.7)
Kidney
0.0098 (0.2)
0.0006(1.4)F(ab')20.227
0.0006(0.1)Fab'1.02(22.9)
CarcassWhole0.140(72.8)°
* Numbers in parentheses, percentage of total extraction by all organs.
antibody fragments are more capable of leaving the capillary
plasma to enter interstitial and cell-associated regions.
The percentages of organ extractions are listed in Table 6.
The gut extracts the greatest percentage of antibody on a single
pass (0.14%), followed by the spleen (0.013%), liver (0.008%),
kidney (0.0007%), and carcass (0.0006%). Although the sur
prisingly high extraction by the spleen has also been made for
albumin catabolism (63), it is hard to rule out the possibility
that damage to protein during labeling may affect this result.
The percentage of total body catabolism by each organ is
listed in Table 6. Contrary to suggestions by others (31, 64),
the majority of the total catabolism of whole IgG is accounted
for by the gut (72.8%). The liver is the second most active
catabolic organ and accounts for 20.5% of the total catabolism.
This result is in good agreement with organ perfusion studies
where the liver is estimated to account for 30% of the total
body catabolism of -y-globulin (65). Although it can be argued
that internalization of radiolabeled antibody prior to catabolism
will increase the effective exposure of cell nuclei for some
isotopes, evidence for the importance of this exposure has not
been firmly established.
Organ extraction for F(ab')2 shows little change from that of
the site of tumor (and binding to it). Whole IgG appears to
have the greatest chance of visiting the interstitial space in the
carcass, with cycling occurring 8-fold more often than in the
case of Fab', (b) The number of cycles of Fab' through the
interstitial space of the kidney was nearly 20-fold greater than
that for F(ab')2.This observation is consistent with the greater
role of kidney in Fab' catabolism.
When comparisons of number of cycles are made between
organs, for either whole IgG or its fragments, those organs
receiving the highest plasma flow and having the largest values
of permeability-surface area product also have the largest values
whole IgG for liver and carcass. A doubling of organ extraction,
over the value for whole IgG, was observed for F(ab')2 in gut
and spleen, and a dramatic 100-fold increase for F(ab')2 was
observed for kidney. Clearly, most of the catabolism of F(ab')2
is accounted for by the kidney (50.3%). The shift in catabolism
of F(ab')2 to the kidney is also found for Fab' where an
increased catabolic role of kidney is observed. The increased
role of kidney in antibody fragment catabolism, which has been
Table 7 Number of cycles through the interstitial and cell-associated
compartments before catabolism and excretion of immunoglobulin, calculated as
the product of the compartmental mean residence time and the compartmental
turnover rate (47)
OrganGutInterstitialCell
(41.0)'10.1
(2.0)0.5
associatedSpleenInterstitialCell
(5.7)67.8(1130.0)3.9 (0.3)4.1
(68.3)0.2
associatedLiverInterstitialCell
(65.0)2291.0(2490.2)542.8
(3.3)81.7(88.8)26.6 (1.7)19.9(21.6)6.8
associatedKidneyInterstitialCell
(590.0)194.7(778.8)70.2
(28.9)10.7
associatedCarcassInterstitialCell
(280.8)43.6
(42.2)3.5
(14.0)2.7
(7.3)200.2
(834.4)1.2(4.8)4.9
(2.8)18.8(1.2)401.4(3087.7)10.8(83.1)F(ab')23.6
(0.2)1.2(0.1)22.5(173.1)0.6
(0.4)0.3
associatedLungInterstitialCell
(0.02)8.2
(63.4)0.2
associatedWhole72.2"
(4.6)Fab'2.0(1.1)0.2(0.1)3.1(52.9)0.1
(1.5)
" Cycling values in units of number for total organ.
b Numbers in parentheses, cycling values in units of number/g organ weight.
3974
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1986 American Association for Cancer Research.
IgG KINETICS
for cycling through the interstitial space. The order of ranking
for cycling per organ weight of whole IgG and F(ab')2 through
the interstitial space of each organ is lung > liver > spleen >
kidney > gut > carcass. For Fab' the order of ranking has
shifted so that the kidney has the largest value for cycling,
followed by lung, spleen, liver, gut, and carcass.
The pharmacokinetic results presented above can be used to
make comparisons between whole IgG and its fragments for
their use in immunodetection or therapy. Although explicit
criteria for antibody selection have not yet been specified, the
differential pharmacokinetics of these molecules will clearly be
important.
This analysis has shown that i.v. administered antibody rap
idly distributes into the nonplasma spaces of organs with high
capillary permeability (e.g., liver, spleen, and kidney) and more
slowly into organs with lower capillary permeability (e.g., gut
and carcass). The extreme case is that of whole IgG in the
carcass, for which the model-predicted distribution time was
nearly 5 h. Under no circumstances was the distribution volume
for antibody within an organ restricted to capillary and inter
stitial spaces. The nonplasma, noninterstitial space was smaller
than the sum of plasma and interstitial spaces for each organ
except the carcass (for all fragments) and the gut and kidney
(for Fab' only).
Whole IgG is retained in the body for a considerably longer
time than either fragment. Consequently, tissue exposure per
sists longer for IgG than for either fragment. On a per gram
basis, the liver, spleen, gut, and lung have at least 2-fold more
exposure when compared with kidney or carcass. Long total
body retention of whole IgG may be an advantage in terms of
therapy. In the worst case, each antibody molecule visits the
interstitial space of the carcass (where tumor could be located)
at least 43.6 times or 2.8 times on a per gram tissue weight
basis.
The F(ab')2 fragment is pharmacokinetically different from
whole IgG mostly on the basis of its 17-fold shorter total body
retention time. Other characteristics, such as distribution vol
umes and speed of distribution into these volumes, are compa
rable with those of whole IgG. The shorter length of stay reduces
exposure, when compared with whole IgG. The reduction in
exposure appears to be distributed equally across all compart
ments (except the kidney, where slightly more uptake is ob
served). As a result of the shorter total body retention, F(ab')2
cycles only 2.7 times through the interstitial space of the
carcass, that is, 0.2 times per gram weight. This may be insuf
ficient for good localization of tumor served by capillaries with
continuous endothelia.
The Fab' fragment is pharmacokinetically quite different
from whole IgG or F(ab')2, apparently due to its smaller size.
When compared with whole IgG or F(ab')2, it (a) distributes
more rapidly into a larger total distribution volume, (6) pro
duces higher interstitial:plasma concentration ratios, (c) is more
readily extracted from the circulation and catabolized by kidney,
spleen, and gut, (d) is cleared from the body 35 times faster,
and (e) cycles through non-kidney interstitial spaces less often.
The in vivo differences between F(ab')2 and whole IgG, due to
shorter total body retention, also apply for Fab'. And, as in the
case of F(ab')2, it remains to be determined whether the de
creased number of cycles through the interstitium by Fab' is
IN MICE
lative tissue exposure when using fragments instead of whole
IgG. The less obvious results, those more dependent on formal
modeling, relate to pharmacokinetic interpretation terms of
volume of distribution, permeability-surface area product, or
gan clearance, organ extraction, mean residence time, and
number of cycles through various body regions. These results
provide a basis for quantitative comparisons between antibodies
(and their fragments) for their in vivo use. A physiologically
rational model of antibody kinetics also provides a general
framework for the presentation of kinetic data and may thus
serve as a basis for a wider range of applications. As an example,
this analysis has shown that there may be significant trade-offs
when using rapidly cleared antibody fragments. The smallest
fragment, with the fastest clearance, also cycles the fewest times
through regions in the body where tumor may exist. This
characteristic may prove to offset any advantages in localization
due to lower background exposure. The physiological model
presented here is currently being extended to include binding
to tumor and circulating antigen. It is also being used, in
conjunction with clinical data, in the generation of models for
antibody administration in humans.
ACKNOWLEDGMENTS
The authors would like to thank Dr. Robert Dedrick, Dr. Peter
Bungay, Dr. Steven Larson, and Dr. Andrew Keenan for their helpful
suggestions and stimulating discussions during the preparation of this
manuscript.
APPENDIX
A: GLOSSARY
OF TERMS
CFL
labeled antibody concentration in plasma (PM)
d,j
tissue (PM)
labeled antibody concentration in interstitium of they'th tissue
Cc.j
cell-associated labeled antibody concentration for they'th tis
Qj
sue (PM)
plasma flow toy'th tissue (ml/min)
VPL
Vp.j
Vi,j
Vc.j
volume of plasma not within organs (ml)
capillary plasma volume ofy'th tissue (ml)
interstitial volume ofy'th tissue (ml)
cell-associated volume ofy'th tissue (ml)
R/VC
transfer rate of antibody from cell-associated to interstitial
compartment (min"1)
transfer rate of antibody from interstitial to cell-associated
compartment (min"1)
catabolic clearance of antibody from cell-associated compart
ment (ml/min)
lymph flow rate (ml/min)
Staverman reflection coefficient
organ (gut, liver, spleen, kidney, carcass, lung)
Cpj
R/'Vi
Äeiim
L
a
j
labeled antibody concentration in capillary plasma of they'th
(PM)
APPENDIX
B
In this section the mathematical model for antibody pharmacokinet
ics used to fit the data is developed. A physiological flow model, similar
to that proposed by Bischoff and Dedrick for the anticancer agent
methotrexate (40), is proposed to describe the circulation of plasma
throughout the body. The convective and diffusive components of
transcapillary exchange are modeled as proposed by Landis and Pap
sufficient for acceptable localization.
penheimer (71). Catabolism of antibody is assumed to occur only when
In conclusion, the modeling approach presented here has antibody interacts with cell-associated material of the organs.
yielded some results that are obvious and others that are not.
The circulation portion of the model consists of plasma flowing
The more obvious results, which have also been found by others
through highly perfused organs (e.g., liver, lung, kidney, spleen, and
who have not used formal modeling, relate to the lower cumu
gut) and more slowly perfused organs (e.g., muscle, skin, etc.). A
3975
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1986 American Association for Cancer Research.
IgG KINETICS
IN MICE
schematic of the model is shown in Fig. 3. Plasma antibody is assumed
fluid is (72)
to distribute to each organ at a rate determined by the plasma flow to
J, = 0.5yv(l - «)(CP+ Q + PS(Cf (B)
that organ. The amount of antibody delivered to each organ per unit
time (pmol/min) is determined as the product of the plasma flow to
the organ (ö„rgan.
ml/min) and the plasma antibody concentration (CfL, where Cp and C, are the capillary plasma and interstitial concentrations
of antibody, J, is the net rate of volume flow through the capillary
pmol/liter). The amount of antibody returning to plasma from each
organ per minute is the product of the plasma flow to the organ (ßorgan) membrane (ml/min), a is the reflection coefficient for restricted move
ment of molecules by convective flux, and PS has been defined as DA/
and the antibody concentration in the capillary plasma of the organ
d (ml/min), where D is the free diffusion coefficient (cm2/min), A is the
(Q.
total cross-sectional area of the aggregate of transcapillary pores (cm2),
The transcapillary exchange component of the model describes the
and d is the average distance across the capillary wall. This formulation
movement of antibody between plasma and interstitial fluid. A sche
assumes that an "ideal" membrane containing a uniform distribution
matic of this system is shown in Fig. 4. This exchange is assumed to
of pores of equal dimensions separates capillary plasma and interstitial
occur via diffusive and convective pathways through the capillary enfluid. The value of J, for each organ will be taken as equivalent to the
dothelium and basement membrane. The transport equation describing
net rate of lymph flow, L (73).
the net flux of antibody (/„pmol/min) between plasma and interstitial
Antibody in interstitial fluid of each organ may exchange with the
capillary plasma (as described above) or enter the lymphatic system and
return directly to circulating plasma. The rate at which antibody leaves
the interstitium in the lymph and returns directly to plasma is <",/,.
Total
Alternatively, antibody may associate nonspecific-ally with either cells
or matrix within the tissue. Our data do not permit a distinction between
the two; hence, we will refer to such interactions as cell associated. In
the model, cell-associated antibody can be irreversibly converted to nonantibody metabolites, which enter the plasma and are subsequently
excreted from the body. In the absence of catabolism the concentration
of antibody in the interstitium would eventually equilibrate with the
concentration in the cell-associated compartment.
In summary, a linear model is proposed to describe the kinetics of
nonspecific antibody pharmacokinetics in normal mice. The salient
features of the model include plasma flow to each organ, plasma
antibody distribution to the interstitial fluid of each organ, the forma
tion and flow of lymph, and the uptake and catabolism of antibody by
cell-associated material of each organ. Mass balance equations for the
model are as follows. For plasma:
VPLdCPL/dt
=
T
-
(ßuver
iÃ-gut^-Lgut
+
okidney
•*-spleen^/,sple«n
+
öc.rcaJCW
'
+
^Uver^-i.liver
ölUngQlu„g
'
^-kidney*-
(C)
/.kidney
"T ¿«carcass^'i.carcass '
For lung:
^J>.lling</C,,liing/</'
=
3liverCj>.liver +
öcarcassCp.carcass +
ökidncyQ.kidnty
(D)
p,]a„s
—Cj.lung) —£|ung(l ~ o
Fig. 3. Compartmental model for IgG pharmacokinetics. Antibody is distributed lo each organ according to the arterial plasma flow to that organ (Q,). Total
organ
measurements
of antibody were made for liver, spleen, gut, kidney, carcass,
lune, and
olasma.
lung, and plasma.
an(j cor ^
-tn organ.
_ .. = QfCM
_ _ ~ (Qi
.,, ~ Lj)<~f.j
. .„
»*;</C,j/</f
(E)
- PS(Cp,j - C,,i) - L¡(\ddj/dt
Js=PS(C1-Cp)+0.5L(l-o)(Cp+Cj)
CAPILLARY
PLASMA
= PS(Cp.j - C,,-)
+ L,(l - a)C - LA;
(F)
- R(Ci.j - Cc,j)
ycjdCc.j/dt = Rj(C¡j—Ce.]) —KtHm.jCcj
(G)
A set of equations analogous to those proposed for the lung describes
the liver, with the modification that liver receives venous effluent from
the gut and spleen. A definition for the terms in this model can be
found in "Appendix A."
To permit interpretation of data on plasma TCA-precipitable frac
tion, a model for non-antibody metabolites has been included in the
CELL-ASSOCIATED
analysis (see Fig. 5). Irreversible losses of antibody from liver, kidney,
gut, spleen, and carcass due to catabolism constitute inputs directly
into the non-antibody metabolite model. Non-antibody metabolites are
assumed to distribute between plasma and the rest of the total body
Fig. 4. Organ model for exchange of antibody between capillary plasma,
water and be excreted from the body (74).
interstitial space, and cell-associated compartment. Q is plasma flow (ml/min), L
Not all the parameters of the proposed model can be determined
is lymph flow, and C and V are the antibody concentrations and distribution
from information contained in our experiments. However, values for
volumes for capillary plasma (C,, yp), interstitial (C,, V,), and cell-associated (C,
certain parameters can be predicted from data in the literature, and
V,) compartments.
3976
INTERSTITIAL
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1986 American Association for Cancer Research.
IgG KINETICS IN MICE
ANTIBODY
USA, 78: 3199-3203. 1981.
MODEL
18.19.Larson, S. M., Brown, J. P., Wright. J. A., Carrasquillo. J. A., Helstrom, I.,
"out
20.
Fig. 5. Non-antibody metabolite model. Catabolic losses from the antibody
model (shown in Fig. 4) constitute inputs into the plasma compartment of the
non-antibody metabolite model. Non-antibody metabolites distribute throughout
body water and are excreted.
realistic model constraints can be imposed to reduce further the number
of unknown parameters. In this analysis, literature values for plasma
flows to each organ, (->,.and the plasma and interstitial distribution
volumes, Vpand V¡,
for each organ were used (40, 75). These values are
listed in Table 1 and remained constant throughout the modeling
analysis. The flow of lymph for visceral and nonvisceral organs was
assumed to be 2 and 4% of the plasma flow to each organ, respectively.
The reflection coefficient, a, for each organ was fixed at 0.95 for the
whole antibody, 0.90 for the F(ab')2 fragment, and 0.5 for the Fab'
fragment. These values are consistent with estimates of reflection
coefficients for proteins of equivalent Stokes' radii (56, 76, 77). The
rate constant, R, which determines the time for antibody in the inter
stitial and cell-associated compartments to reach concentration equilib
rium (under conditions of no catabolism) was assumed to be 0.02 mm.
This value is sufficiently large that it never becomes rate limiting. The
remaining parameters in the model were treated as unknown with
values to be determined from the data. The unknown model parameters
are PS, A"C|im,
and Ve.
REFERENCES
1. Bale, W. F., and Spar, I. L. Studies directed toward the use of antibodies as
carriers of radioactivity for therapy. Adv. Biol. Med. Phys., 5: 285-356,
1957.
2. Pressman, D. the development and use of radiolabeled antitumor antibodies.
Cancer Res.. 40: 2960-2964, 1980.
3. Goldberg. D. M. An introduction to the radioimmunodetection of cancer.
Cancer Res., 40: 2957-2959, 1980.
4. Borowitz, M. J. Laboratory diagnosis of human cancer using monoclonal
antibodies. In: D. S. Gordon (ed.). Monoclonal Antibodies in Clinical Diag
nostic Medicine, Chap. 7. New York: Igaku-Shoin Medical Publishers, 1985.
5. Milstein, C. Monoclonal antibodies. Sei. Am.. 00: 66-74, 1980.
6. Rollo, F. D., and Patton. J. A. Imaging techniques for the radioimmunode
tection of cancer. Cancer Res.. 40: 3050-3053, 1980.
7. DeLand, F. H., Kim. E. E., Simmons, G., and Goldenberg, D. M. Imaging
approach in radioimmunodetection. Cancer Res., 40: 3046-3049, 1980.
8. Order, S. E., Klein, J. L., Ettinger. D., Alderson. P., Seigleman, S., and
Leichner, P. Use of isotopie immunoglobulin in therapy. Cancer Res., 40'
3001-3007. 1980.
9. Chose. T.. Tai, J., Guclu, A., Noevell, S.. Bodurtha. A.. Aquino. J., and
MacDonald, A. J. Antibodies as carriers or radionuclides and cytotoxic drugs
in the treatment and diagnosis of cancer. Ann. NY Acad. Sci., 277:671-689.
1976.
10. Gallagher. B. M. Monoclonal antibodies: the design of appropriate carrier
and evaluation systems. In: R. M. Lambrecht and W. Eckelman (eds.),
Animal Models in Radiotracer Design, pp. 61-105. Basel: Springer-Verlag
1983.
11. Burchiel, S. W., Khaw, B. A.. Rhodes, B. A.. Smith. T. W., and Haber, E.
Immunopharmacokinetics of radiolabeled antibodies and their fragments. In:
S. W. Burchiel and B. A. Rhodes (eds.). Tumor Imaging, pp. 125-139. New
York: Masson Publishing, 1982.
12. Sfakianakis, G. V. and DeLand, F. H. Radioimmunodiagnosis and radioimmunotherapy. J. NucÃ-.Med., 23: 840-850. 1982.
13. Gitlin, D., Landing, B. H., and Wipple, A. The localization of homologous
plasma proteins in the tissues of young human beings as demonstrated with
fluorescent antibodies. J. Exp. Med., 97: 163-174, 1952.
14. Gold. P., and Shuster. J. Historical development and potential uses of tumor
antigens as markers of human cancer growth. Cancer Res.. 40: 2973-2976,
1980.
15. DeLand, F. H., Kim, E. E., and Goldenberg, D. M. Lymphoscintigraphy
with radionuclide-labeled antibodies to carcinoembryonic antigen. Cancer
Res., 40: 2997-3000, 1980.
16. Chose, T., Norvell, S. T., Aquino, J., Belitsky, P., Tai, J., Guclu, A., and
Blair, A. H. Localization of 13ll-labeled antibodies in human renal cell
carcinomas and in mouse hepatoma and correlation with tumor detection by
photoscanning. Cancer Res., 40:3018-3031, 1980.
17. Colcher, D., Horan, P.. Nuli, M., and Schlom, J. A spectrum of monoclonal
antibodies reactive with human mammary tumor cells. Proc. Nati. Acad. Sci.
21.
22.
23.
24.
25.
26.
27.
and Helstrom, K. E. Imaging of melanoma with 1-131 labeled monoclonal
antibodies. J. NucÃ-.Med., 24: 123-129, 1983.
Colcher. D., Zalutsky, M., Kaplan, W., Kufe, D., Austin, F., and Schlom, J.
Radiolocalization of human mammary tumors in athymic mice by monoclo
nal antibody. Cancer Res., 4: 736-742, 1983.
von Kleist, S. U. Diagnostic significance of tumor markers. Cancer Res., 40:
2977-2978. 1980.
Rockoff, S. D., Goodenough, D. J., and Mclntire, K. R. Theoretical limita
tions in the immunodiagnostic imaging of cancer with computed tomography
and nuclear scanning. Cancer Res., 40: 3054-3058. 1980.
Bradwell, A. R., Fairweather, D. S., Dykes, P. W., Keeling, A., Vaughn, A.,
and Taylor, J. Limiting factors in localization of tumours with radiolabeled
antibodies. Immunol. Today, 6: 163-170. 1985.
Bale, W. F., Contreras, M. A., and Grady, E. D. Factors influencing local
ization of labeled antibodies in tumors. Cancer Res., 4: 2965-2972, 1980.
Goldenberg, D. M., DeLand, F. H.. Kim, E., Bennett, S., Primus, J., van
Nagell, J. R., Estes, N., DeSimone, P., and Rayburn, P. Use of radiolabeled
antibodies to carcinoembryonic antigens for the detection and localization of
diverse cancers by external photoscanning. N. Engl. J. Med., 298: 13841388. 1978.
Dykes, P. W., Hiñe,K. R., and Bradwell, A. R. Localization of tumor
deposits by external scanning after injection of radiolabeled anti-carcinoembryonic antigen. Br. Med. J., 280: 220-222. 1980.
Mach, J. P., Carrel, S., Forni, M., Ritschard, J., Donath, A., and Alberto, P.
Tumour localization of radiolabelled antibodies against carcinoembryonic
antigen in patients with carcinoma. N. Engl. J. Med., 303: 5-10, 1980.
Wahl, R. L., Parker, C. W., and Philpott, G. W. Improved imaging and
tumor localization with monoclonal F(ab')2. J. NucÃ-.Med., 24: 316-325,
1983.
28. Powe, J., Pak, K. Y., Paik, C. H., Steplewski, Z., Ebbert, M. A., Herlyn, D.,
Ernst, C., Alavi, A., Eckelman, W. C.. Reba. R. C.. and Koprowski, H.
Labeling monoclonal antibodies and F(ab')2 fragments with ['"Injindium
using cyclic DPTA anhydride and their in vivo behaviour in mice bearing
human tumor xenografts. Cancer Drug Deliv., I: 125-135, 1984.
29. Mach, J., Chatal, J., Lumbroso, J., Buchegger, F., Forni, M., Ritchard, J.,
Berche, C., Douillard, J., Carrel, S., Herlyn, M.. Steplewski, Z., and Ko
prowski, H. Tumor localization in patients by radiolabeled monoclonal
antibodies against colon carcinoma. Cancer Res., 43: 5593-5600, 1983.
30. Chatal, J., Saccavini, J., Fumoleau, P., Douillard, J., Curtet, C., Kremer, M.,
Le Mevel, B.. and Koprowski, H. Immunoscintigraphy of colon carcinoma.
J. NucÃ-.Med., 25: 307-314, 1984.
31. Waldmann, T. A., and Straber, W. Metabolism of immunoglobulins. Prog.
Allergy, 13: 1-110, 1969.
32. Smith, T. W., Lloyd, B. L.. Spicer, N., and Haber, E. Immunogenicity and
kinetics of distribution and elimination of sheep diogoxin-specific IgG and
Fab fragments in the rabbit and baboon. Clin. Exp. Immunol. 36: 384-396,
1979.
33. Spiegelberg. H. L., and Weigle, W. O. The catabolism of homogenous and
heterologous 7S gammaglobulin fragments. J. Exp. Med., 121: 323-338,
1965.
34. Houston, L. L., Nowinski, R. C., and Bernstein, I. D. Specific in vivo
localization of monoclonal antibodies directed against the Thy 1.1 antigen.
J. Immunol., 125: 837-843, 1980.
35. Bernstein, J. D., Tarn, M. R., and Nowinski, R. C. Mouse leukemia: therapy
with monoclonal antibodies against a thymus differentiation antigen. Science
(Wash. DC). 207: 68-71, 1980.
36. Badger, C. C., Krohn, K. A., Peterson, A. V.. Shulman. H., and Bernstein, I.
D. Experimental radiotherapy of murine lymphoma with "'I-labeled antiThy 1.1 monoclonal antibody. Cancer Res., 45: 1536-1544, 1985.
37. Symington, F. W., Subbarao, B.. Moiser, D. E., and Sprent, J. Lyb-8.2: a
new B-cell antigen defined and characterized with a monoclonal antibody.
Immunogenesis, 16:381-391. 1982.
38. Parham. P.. Androlewicz. M. J.. Brodsky. F. M., Holmes, N. J.. and Ways,
J. P. Monoclonal antibodies: purification, fragmentation, and application to
structural and functional studies of Class I MHC antigens. J. Immunol.
Methods. 53: 133-173, 1982.
39. O'Farrell, P. H. High resolution two-dimensional electrophoresis of proteins.
J. Biol. Chem., 250: 4007-0000. 1975.
40. Bischoff. K. B., Dedrick. R. L.. and Zaharko, D. S. Methotrexate pharmacokinetics. J. Pharm. Sci., 60: 1128-1133, 1971.
41. Altman, L. A. Biological Handbooks: Blood and Other Body Fluids, pp. 911. Washington. DC: Federation of American Societies for Experimental
Biology, 1961.
42. Berman, M., and Weiss, M. SAAM Users Manual. DHEW Publication
(NIH)79-I80, 1978.
43. Boston, R. C., Greif, P. C., and Berman, M. Conversational SAAM—an
interactive program for kinetic analysis of biological systems. Comput. Pro
grams Biomed.. 13: 111-119, 1981.
44. Bard, Y. Nonlinear Parameter Estimation. New York: Academic Press, 1974.
45. Berman, M. Compartmental modeling. In: J. Laughlin and E. Webster (eds.).
Advances in Medical Physics, pp. 279-296. Boston: Second International
Conference on Medical Physics, 1971.
46. DiStefano, J. J. Concepts, properties, measurement, and computation of
clearance rates of hormones and other substances in biological systems. Ann.
Biomed. Eng., 4: 302-319, 1976.
47. Covell, D. G., Berman, M., and DeLisi, C. Mean residence time—theoretical
3977
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1986 American Association for Cancer Research.
IgG KINETICS
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
development, experimental determination, and practical use in tracer analy
sis. Math. Biosci. 72: 213-244, 1984.
Herman. M.. and Schoenfìeld,R. Invariants in experimental data of linear
kinetics and the formulation of models. J. Appi. Phys., 27:1361-1370,1956.
Jacquez, J. A. Compartmental Analysis in Biology and Medicine. Amster
dam: Elsevier, 1972.
Berman, M. Information content of data with respect to models. Am. J.
Physiol., 245: 620-623, 1983.
Cobelli, C, and DiStefano, J. J. Parameter and structural identifìability,
concepts, and ambiguities: a critical review. Am. J. Physiol., 239: R7-R24,
1980.
Renkin, E. M. Multiple pathways of capillary permeability. Circ. Res., 41:
735-743, 1977.
Garlick, D. G., and Renkin, E. M. Transport of large molecules from plasma
to interstitial fluid and lymph in dogs. Am. J. Physiol., 219: 1595-1605,
1970.
Hopf, U., zumBushenfelde, K., and Dierich, M. P. Demonstration of binding
sites for IgG Fc and the third complement component (C3) on isolated
hepatocytes. J. Immunol., 117: 639-645, 1976.
Watson, P. D., and Grodins, F. S. An analysis of the effects of the interstitial
matrix on plasma-lymph transport. Microvasc. Res., 16: 19-41, 1978.
Taylor, A. E., and Granger. E. Exchange of macromolecules in the microcir
culation. In: E. M. Renkin and C. C. Michel (eds.), The Handbook of
Physiology: The Cardiovascular System, Vol. 4, Part 1, pp. 477, 486, 495.
Bethesda, MD: American Physiological Society, 1984.
Gitlin, D., and Janeway, C. Studies on the plasma proteins in the interstitial
Huid. Science (Wash. DC), 120: 461-463, 1954.
Dewey, W. C. Vascular-extravascular exchange of 1-131 plasma proteins in
the rat. Am. J. Physiol., 192:423-431, 1959.
Bill, A. Plasma protein dynamics: albumin and IgG capillary permeability,
extravascular movement, and regional blood flow in unanesthetized rabbits.
ACTA Physiol. Scand., 101: 28-42, 1977.
Fox, J. R., and Wayland, H. Interstitial diffusion of macromolecules in the
rat mesentary. Microvasc. Res., 18: 255-276, 1979.
Mayerson, H. S., Wolfram, C. G., Shirley, H. H., and Wasserman, K.
Regional differences in capillary permeability. Am. J. Physiol., 198: 155160, I960.
Steller, M. A., Parker, R. J., Covell, D. G., Holton, O. D., Ill, Keenan, A.
M., Sieber, S. M., and Weinstein, J. N. Optimization of monoclonal antibody
delivery via the lymphatics: the dose dependence. Cancer Res., 46: 1830-
IN MICE
1834, 1986.
63. Yedgar, S., Carew, T. E., Pittman, R. C., Beltz, E. F., and Steinberg, D.
Tissue sites of catabolism of albumin in rabbits. Am. J. Physiol., 244: E101107, 1983.
64. Cohen, S., Gorden, A. H., and Matthews, C. Catabolism of -y-globulin by
isolated perfused rat liver. Biochem. J., 82: 197-205, 1962.
65. Fukumotio, T., and Brandon, M. R. The site of IgG2a catabolism in the rat.
Mol. Immunol., 18: 741-750, 1981.
66. Scheinberg, D. A., and Strand, M. Kinetic and catabolic considerations of
monoclonal antibody targeting in erythroleukemic mice. Cancer Res., 43:
265-272, 1983.
67. Solomon, A., Waldmann, T. A., Fahey, J. L., and McFarlane, A. S. Metab
olism of Bence-Jones proteins. J. Clin. Invest., 48: 103-117, 1964.
68. Wochner, R. D., Straber, W., and Waldmann, T. A. The role of the kidney
in the catabolism of Bence Jones proteins and immunoglobulin fragments.
J. Exp. Med., 126: 207-221, 1967.
69. Deen. W. M., and Satvat, B. Determinants of the glomerular filtration of
proteins. Am. J. Physiol. (Renal and Fluid Electrolyte Physiol.), 241: F162F170. 1981.
70. Christensen, E. I., Rennke, H. G., and Carone, F. A. Renal tubular uptake
of protein: effect of molecular charge. Am. J. Physiol., 244: F436-F441,
1983.
71. Landis, E. M., and Pappenheimer, J. R. Exchange of substances through the
capillary walls. In: W. F. Hamilton and P. Dow (eds.), The Handbook of
Physiology: Circulation II. Washington, DC: American Physiological Soci
ety, 1963.
72. Bressler, E. H., and Groóme, L. J. On equations for combined convective
and diffusive transport of neutral solute across porous membranes. Am. J.
Physiol., 241: F469-F476, 1981.
73. Perl, W. Convection and permeation of albumin between plasma and interstitium. Microvasc. Res., 10: 83-94, 1975.
74. Singh, B., Sharma, S. M., Patel, M. C., Raghavendran, K. V., and Berman,
M. Kinetics of large therapy doses of 131-1 in patients with thyroid cancer.
J. NucÃ-.Med., 75:674-678, 1974.
75. O'Connor, S. W., and Bale, W. F. Accessibility of circulating immunoglobulin
G to the extravascular compartment of solid rat tumors. Cancer Res., 44:
3719-3723, 1984.
76. Curry, F. E., Huxley, V. H., and Adamson, R. H. Permeability of single
capillaries to intermediate-sized colored solutes. Am. J. Physiol., 245: H495H505, 1983.
77. Curry, F. E. A hydrodynamic description of the osmotic reflection coefficient
with applications to the pore theory of transcapillary exchange. Microvasc.
Res., A: 236-252, 1974.
3978
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1986 American Association for Cancer Research.
Pharmacokinetics of Monoclonal Immunoglobulin G1, F(ab′)2,
and Fab ′ in Mice
David G. Covell, Jacques Barbet, Oscar D. Holton, et al.
Cancer Res 1986;46:3969-3978.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/46/8/3969
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1986 American Association for Cancer Research.