Departments of Drug Metabolism and Pharmacokinetics (A.S., C.B.D.,
L.A.P.T., D.C.K., M.G.T.) and of Safety Assessment (E.R.G., D.J.H.,
T.K.H.), SmithKline Beecham Pharmaceuticals, King of Prussia,
Pennsylvania
Keliximab and clenoliximab are monkey/human chimeric CD4
monoclonal antibodies (mAbs) of the IgG1 and IgG4 isotypes,
respectively. The pharmacokinetics (PK) and pharmacodynamics (PD) of
these mAbs were evaluated in transgenic mice bearing human CD4
molecules on their T cells after a single i.v. administration at three
dose levels (5-125 mg/kg). The PK of keliximab and clenoliximab were similar, dose-dependent, and adequately described by a two-compartment model with saturable elimination from both compartments. The
enumeration of circulating CD4+ T cells and density of CD4
on their surface were determined as the PD effects. An indirect
response model was proposed to characterize the PD effects. With the
increase in mAb dose, the maximum intensity (Rmax) of PD effects was increased, and the
time to reach Rmax shifted to later times.
At all three dose levels, keliximab caused a relatively rapid decline
in the number of circulating CD4+ T cells, which then
recovered gradually. In contrast, clenoliximab at the lowest dose (5 mg/kg) did not produce a significant effect on CD4+ T cell
counts compared with the placebo group. At high doses, clenoliximab
caused a significant decrease in the number of CD4+ T
cells. Keliximab appeared to be more potent and efficient in depleting
CD4+ T cells. Both mAbs produced similar down-modulation of
CD4 at corresponding dose levels. The findings of this study are
consistent with the results of a recent clinical trial that emphasize
the importance of this transgenic mouse model for evaluating PK/PD to
support clinical development of anti-human CD4 mAbs.
 |
Introduction |
Monoclonal
antibodies (mAbs) are in development and clinical use for diagnosis and
treatment of numerous diseases. mAbs directed against T cells have been
studied for their potential to modify pathological conditions in which
T cells are involved, including autoimmune disorders such as rheumatoid
arthritis (RA), transplant rejection, or graft versus host disease
(Hafler and Weiner, 1988
; Horneff et al., 1993; Vincenti et al., 1998).
One of the most promising targets for mAbs against T cells has been the
CD4, a membrane-bound glycoprotein expressed on T lineage cells,
including a subset of peripheral T cells. CD4 is critical for
interaction with antigen-presenting cells bearing class II major
histocompatibility complex (MHC II), which results in an enhancement of
the immune response (Veillette et al., 1988). Antibodies to CD4 have
been shown to inhibit T cell proliferation and cytokine production in
vitro (Takeuchi et al., 1987; Schrenzenmeier and Fleisher, 1988).
Keliximab (SB-210396) and clenoliximab (SB-217969) are monkey/human
chimeric anti-human CD4 mAbs. Clenoliximab is an IgG4 derivative of the
IgG1 mAb, keliximab, with heavy- and light-chain-variable regions from
cynomolgus macaque and human lambda light-chain (c
) and 
4
heavy-chain constant domains. Both keliximab and clenoliximab bind to
the same epitope on domain 1 of human CD4, with identical specificity
and affinity, and are very effective in blocking the interaction of CD4
with MHC II on antigen-presenting cells and B lymphocytes. Keliximab
exhibits no binding to complement component C1q and thus does not
mediate complement-dependent cytotoxicity. However, as expected of an
IgG1 mAb, keliximab shows efficient binding to human IgG Fc receptors,
and thus causes depletion of CD4+ T cells. In
contrast, clenoliximab is expected to cause little or no depletion of
CD4+ T cells due to the reduction of the
Fc-dependent effector functions. Recently, clinical studies of these
IgG1 and IgG4 mAbs were conducted in RA patients (Yocum et al., 1998;
Mould et al., 1999).
Preclinical investigation of the pharmacology of these mAbs has been
complicated by their specificity for human CD4 antigen; these mAbs
cross react only with chimpanzee CD4 antigen. The development of
transgenic mice bearing human CD4 in place of mouse CD4 (HuCD4/Tg) made
it feasible to evaluate preclinical pharmacokinetics (PK) and
pharmacodynamics (PD) of these mAbs. Keliximab and clenoliximab are pharmacologically active in the HuCD4/Tg mice and thus are expected
to produce PD effects similar to those in humans. The main objectives
of this study were to evaluate PK of keliximab and clenoliximab after a
single i.v. administration at three dose levels (5-125 mg/kg) and to
characterize their effects on circulating CD4+ T
cells in HuCD4/Tg mice. A PK/PD model is proposed to describe the time
course of these PD effects.
| |
Materials and Methods |
Chemicals.
Keliximab, clenoliximab, and recombinant
soluble CD4 were expressed in Chinese hamster ovary cells and purified
to homogeneity at SmithKline Beecham Pharmaceuticals. Mouse anti-human
IgG1 (clone HP 6069) and IgG4 (clone HP 6023) mAbs were purchased from
Zymed Laboratories Inc. (San Francisco, CA) and CALTAG (Burlingame, CA), respectively. All other chemicals were of reagent grade or better.
Animal Husbandry.
Male HuCD4/Tg mice were obtained from
Charles River Laboratories (Raleigh, NC). These animals express
physiological levels of human CD4 and the transgene is appropriately
regulated during T cell development. Furthermore, the human CD4 gene
restores normal helper cell functions relative to animals with no
endogenous CD4 gene (Killeen et al., 1993). The mice were approximately
9 to 17 weeks of age (approximately 30-44 g) at the initiation of
dosing. Mice were individually housed in stainless steel cages in a
controlled environment (72 ± 4°F; 40-70% relative humidity)
with a 12-h light/dark cycle. Mice were offered 5002 Certified Rodent
Diet (PMI Feeds, Inc., St Louis, MO) ad libitum. Filtered tap water was
available ad libitum from an automatic watering system. This study was
conducted in accordance with the Guide for the Care and Use of
Laboratory Animals as adopted and promulgated by the National
Institutes of Health.
Administration of Anti-CD4 mAbs.
HuCD4/Tg mice were given a
single i.v. dose of vehicle or drug (5, 25, or 125 mg/kg) via a tail
vein at a dose volume of 10 ml/kg on day 1 (n = 10-13/group). Five mice per group were bled (approximately 50 µl)
from a tail vein into heparinized capillary tubes at approximately the
same time of day on days 
4, 2, 4, 8, and 15 and then every 2 weeks
until the CD4+ T cell count returned to baseline
values (serial bleeding). Each blood sample was divided for lymphocyte
subset analysis (whole blood) and to prepare plasma for the analysis of mAbs.
mAb Assays.
Plasma concentrations of unbound keliximab and
clenoliximab were determined using electrochemiluminescent immunoassays
based on the binding of the anti-CD4 mAbs to recombinant soluble CD4. Plasma was incubated with biotinylated soluble CD4,
streptavidin-conjugated paramagnetic beads, and a ruthenium-labeled
mouse anti-human mAb specific for the CH2 domain
of human IgG1 (for keliximab) or the CH3 domain
of human IgG4 (for clenoliximab) (Hamilton and Morrison, 1993).
Electrochemiluminescent responses were recorded with an Origen
analyzer (Igen Inc., Gaithersburg, MD). The lower limits of
quantification for the immunoassays were 500 and 50 ng/ml using a
50-µl aliquot of sample (plasma diluted 1:100, 0.5-µl equivalents neat plasma) for keliximab and clenoliximab, respectively. For both
assays, inter- and intra-run precision was 
8.0% over the assay
calibration range and average bias was 4.1%. A positive response in
these assays is indicative of a species containing an antigen-combining
site and a remote epitope of the constant region of the heavy chain.
The presence of these two features would suggest that the analyte
measured is primarily active, intact, and unbound monoclonal antibody.
Flow Cytometry.
Peripheral blood lymphocyte populations were
analyzed using a flow cytometer (CytoronAbsolute, Ortho Diagnostics,
Westwood, MA). For characterization of T cell subclasses and anti-CD4
mAb coating, cells were stained with the three-color panels,
CD2/OKT4/CD8 and CD3/OKT4/OKT4A, respectively. After a 40-min
incubation period, the red blood cells were lysed, leaving the
antibody-coated white cells intact. Immunophenotyping analysis included
a determination of the absolute number of CD4+
cells in CD2+ T-lymphocyte population using OKT4
(Ortho Pharmaceuticals, Raritan, NJ) as well as the OKT4 mean
fluorescence intensity (CD4 MFI) on these cells. The latter value (CD4
MFI) is proportional to the number of CD4 epitopes on the surface of
CD4+ T-lymphocytes. The OKT4 antibody used in
these analyses binds to an epitope on the CD4 molecule that is distinct
from the epitope recognized by clenoliximab and keliximab. These
measurements, therefore, were not expected to be affected by treatment
with either drug. This was supported by a relatively constant OKT4 CD4
lymphocyte count before and after drug administration in clinical studies of clenoliximab (data not shown).
The percentage of T cell coating by mAb (keliximab or clenoliximab) was
assessed by counting the CD4+ cells using OKT4
and a second reagent, OKT4A (Leu3a; Ortho Pharmaceuticals), which binds
to the same epitope as mAbs. Percentage of T cell coating was derived
from the absolute cell counts (OKT4 and OKT4A) using the following
equation:
|
(1)
|
Pharmacokinetics.
A two-compartment model with saturable
elimination from both compartments was proposed to characterize PK of
keliximab and clenoliximab (Fig. 1). A
one-compartment mamillary model was not able to describe the plasma
concentration-time data simultaneously at all three dose levels because
Vmax alone was not sufficient to
characterize a relatively rapid initial fall in concentrations at high
doses when the plasma concentrations were significantly higher than the
estimated Km value. The fitting of
concentration-time data to a two-compartment model with bi-directional
intercompartment rate constants resulted in a negligibly small value
for the intercompartment rate constant
(kTP), which describes the
distribution from tissue to the plasma compartment.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 1.
PK and PD models for keliximab and clenoliximab. The
PD model represents the effect of anti-CD4 antibodies on number of
circulating CD4+ T cells and cell surface CD4. Symbols are
defined in the text.
|
|
The proposed two-compartment PK model is based on the understanding of
distribution and metabolism of these anti-human CD4 mAbs and assumes
that 1) anti-CD4 mAbs permeate into the tissue compartment
unidirectionally from plasma across capillaries; 2) these mAbs bind to
CD4 on T cells in both circulation (plasma compartment) and tissues;
and 3) the major pathway for the disposition of these mAbs is mediated
through their binding to CD4 in both plasma and tissue compartments.
Thus, the disposition of keliximab and clenoliximab after a single i.v.
administration at three dose levels can be described as:
|
(2)
|
|
(3)
|
Initial conditions: CP = dose/VC; CT = 0;
where CP and
CT are concentration of unbound mAb in
the plasma and tissue compartments,
Vmax represents the maximal
elimination rate, Km is the
Michaelis-Menten constant, Vc is the
central volume of distribution, and kPT
describes the distribution of mAb from plasma to tissue compartment.
The plasma concentration versus time data for keliximab and
clenoliximab were sparse around the Km
value and exhibited large between-animal variability. Thus, the mean
concentration-time data were calculated (mean of five animals per
dose). The mean plasma concentration-time data for these two mAbs were
pooled to estimate PK parameters because they were similar. This was also observed in the previous PK studies in both HuCD4 transgenic mice
and in humans (Mould et al., 1999). The mean pooled plasma concentration versus time data for all three dose levels were simultaneously fitted to the proposed PK model using a nonlinear regression program, Adapt II (D'Argenio and Schumitzky, 1997).
Pharmacodynamics
Descriptive Analysis.
PD descriptors [area between the
baseline and effect curve (ABEC),
Rmax, and
TRmax] were calculated to summarize
the effect of treatment on the number of circulating
CD4+ T cells and CD4 density on their surface
(CD4 MFI). The ABEC was determined by the linear trapezoidal method to
characterize the overall effect of treatment. The intensity of the drug
effect was assessed by calculating the maximum inhibitory response
(%Rmax) as:
|
(4)
|
where R0 is the pretreatment
baseline and Rmin is the nadir of the
measured PD determinant after treatment.
PD Modeling.
The number of circulating
CD4+ T cells, density of CD4 molecule on the
surface of T cells (CD4 MFI), and the fraction of CD4 occupied by mAb
(%coating) were determined as the PD effects of keliximab and clenoliximab.
Number of circulating CD4+ T cells.
The effect
of keliximab and clenoliximab on circulating CD4+
T cell count was characterized by an indirect PD response model (Dayneka et al., 1993; Sharma and Jusko, 1996; Sharma and Jusko, 1998).
Recently, similar models were used for PK/PD analysis of anti-CD40
ligand in monkeys (Gobburu et al., 1998) and clenoliximab in RA
patients (Mould et al., 1999).
The proposed model assumes that production of circulating
CD4+ T cells is an apparent zero-order process
(kin) and their loss from circulation
is controlled by a first order rate constant (kout), as shown in Fig. 1. It is
postulated that these mAbs cause stimulation of the loss of
CD4+ T cells
(kout), which fully describes the
decrease in number of circulating CD4+ T
cells. Thus, the effect (R) of keliximab and clenoliximab
can be described as:
|
(5)
|
and
|
(6)
|
Initial condition: R0 = kin/kout;
where S(t) is the stimulation function,
CP is the plasma concentration of
unbound mAb, Smax is the maximum
stimulation of kout attributed to mAb,
SC50 is the plasma concentration of unbound mAb
producing 50% of the maximum stimulation of
kout, and
R0 is the pretreatment baseline. The value
of kout was not estimated and can be
calculated as
kin/R0.
The number of CD4+ T cells versus time data for
the active treatment groups were not adjusted for the placebo effect
because: 1) placebo effect was significant only at the first time
point; 2) the contribution of placebo effect to the PD response of high doses of mAbs may be minimal; and 3) it is difficult to adjust for
inhibitory placebo effects when the active treatment also produces inhibition.
Surface density of CD4.
An indirect PD model similar to the
one described above was used to characterize the down-modulation of CD4
on the surface of circulating T cells after treatment with anti-human
CD4 mAbs. The proposed model assumes that the production and expression of CD4 on the surface of the T cell is an apparent zero-order process
(kin) and the loss of CD4 is
controlled by a first order rate constant
(kout). The model assumes that mAbs
cause stimulation of kout, which fully
describes the down-modulation of CD4 density.
Pharmacodynamic data were analyzed using the nonlinear mixed effects
modeling program NONMEM (version V) (Sheiner and Beal, 1981).
The distribution of random residual errors and interanimal variability
were described by the exponential error models. The first-order
estimation method was applied in NONMEM for the parameter estimates.
The PD data (CD4+ T cell number and CD4 density)
for each drug at all three dose levels were fitted simultaneously,
except for the effect of 5 mg/kg clenoliximab on the number of
circulating CD4+ T cells. This dose group was not
included in modeling because it showed no significant PD effect
compared with placebo. PK parameters estimated from the mean
concentration-time data were used to assign free concentration
(CP) of keliximab and clenoliximab for
eq. 6.
Percent CD4 coating.
The %coating value as a
function of time was described by the Langmuir adsorption isotherm
equation (Sung et al., 1992):
|
(7)
|
where CP is the concentration
of unbound mAb as a function of time, and
KA is the in vitro mAb affinity
constant. PK parameters estimated from mean concentration-time data
were used to assign plasma concentration
(CP) of unbound keliximab and
clenoliximab for eq. 7.
| |
Results |
Pharmacokinetics.
The plasma concentration-time profiles of
keliximab and clenoliximab were described by a mechanism-based PK model
with saturable elimination from plasma and tissue compartments (Fig.
1). The individual plasma concentration-time data along with the fitted curves are shown in Fig. 2. There was
good agreement between the predicted and observed concentrations. The
estimated PK parameters are presented in Table
1. The estimated tissue volume of
distribution was 10-fold higher than the plasma volume of distribution,
which may be explained by the significantly higher number
(approximately 5-fold) of CD4+ T cells in the
extravascular compartments, e.g., spleen, thymus, and lymph nodes,
compared with the same number of components in circulation. Although
the fitted line described the concentration versus time data well, the
estimation of Km may only be an
approximation because limited data were available around the estimated
Km.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 2.
Plasma concentrations of keliximab and clenoliximab
for all animals at three dose levels. Symbols represent individual
experimental data points. The lines represent the results of
least-square regression fitting of the model to the mean pooled data
for both mAbs at all three dose levels simultaneously to a
two-compartment model with a saturable elimination from both
compartments.
|
|
Pharmacodynamics
Descriptive Analysis.
CD4+ T cells. All
animals had a pronounced decrease in numbers of circulating
CD4+ T cells after a single i.v. dose of
keliximab or clenoliximab as indicated by a significantly lower nadir
for all dose groups than that for placebo (Fig.
3). The placebo group also exhibited a
significant decrease (Rmax ~ 30%)
in CD4+ T cell number at the first sampling time
point (2 day) compared with the pretreatment baseline. The mean
Rmax for CD4+ T
cells increased and the time of occurrence of
Rmax
(TRmax) shifted to later times with
the increase in dose: 2, 4, and 9 days for 5, 25, and 125 mg/kg dose
groups for both mAbs, a signature response pattern for drugs producing
indirect PD responses. With a 25-fold change in the dose of mAbs from 5 to 125 mg/kg, the 1.3-fold increase in
Rmax value was not significant,
indicating that the maximum intensity of the response was achieved at
low doses (Fig. 3).
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 3.
The intensity of the effect of keliximab and
clenoliximab on the number of circulating CD4+ T cells and
cell surface CD4 as measured by the nadir of these PD determinants
after treatment.
|
|
The ABEC for circulating CD4+ T cells increased
with the increase in dose for both mAbs. For keliximab, mean ABEC
values for CD4+ T cells at all three dose levels
were significantly higher than those for placebo (Fig.
4). In contrast, the mean ABEC value at 5 mg/kg clenoliximab was not significantly different from placebo. At
high doses (25 and 125 mg/kg), the effects on
CD4+ T cell numbers were similar for both
clenoliximab and keliximab (Fig. 4). The ABEC value for
CD4+ T cells increased nonproportionately with a
25-fold increase in dose. The degree of nonlinearity was more
pronounced at the higher doses (between 25 and 125 mg/kg) for both
mAbs, especially for keliximab.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 4.
The overall effect of keliximab and clenoliximab on
the number of circulating CD4+ T cells and cell surface
CD4. ABEC refers to the area between baseline and effect curve.
Open columns, placebo; filled columns, keliximab; hatched columns,
clenoliximab.
|
|
The maximum intensity of the effect
(%Rmax) on CD4+
T cell number was similar for keliximab and clenoliximab at
corresponding doses. It is interesting to note that, although the
%Rmax at 5 mg/kg clenoliximab was
significantly higher than that for placebo, the overall response (ABEC)
was not. This is due to the transient and minimal effect of
clenoliximab at 5 mg/kg dose on circulating CD4+
T cells. A similar transient effect on CD4+ T
cell count was observed after treatment with placebo as shown in Fig.
5. In contrast, keliximab at 5 mg/kg dose
produced a relatively prolonged effect on CD4+ T
cell count (Fig. 5).
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 5.
Time course of change in the number of circulating
CD4+ T cells relative to other lymphocytes at the lowest
dose (5 mg/kg) of keliximab and clenoliximab compared with placebo.
Symbols are the mean values (+S.D.) from five animals per dose
group.
|
|
CD4 density.
A significant down-modulation of CD4 density (CD4
MFI) on the surface of circulating T cells was evident in all animals
receiving active treatment as indicated by a lower nadir for all dose
groups compared with that for placebo (Fig. 3). There was no change in CD4 MFI in animals receiving placebo. With the increase in mAbs dose,
mean CD4 MFI Rmax remained almost
constant, indicating that the maximum intensity of the response was
achieved at the lowest dose used in this study (Fig. 3). In contrast,
mean CD4 MFI value for ABEC increased with the increase in dose (Fig.
4). There was an almost proportional increase in CD4 MFI ABEC between
the 5- and 25-mg/kg dose groups for both mAbs: 4- to 5-fold increase in
mean CD4 MFI ABEC value with a 5-fold increase in the dose (Fig. 4).
However, there was a <2.5-fold increase in mean CD4 MFI ABEC with a
further 5-fold increase in the dose from 25 to 125 mg/kg. Overall, both
keliximab and clenoliximab produced similar effects on the CD4 density
at the dose range evaluated.
CD8+ T cells.
Keliximab and clenoliximab have no
effect on the CD8+ T cells (data not shown) as
expected because these mAbs bind to the epitope on CD4 antigen and
exhibit no cross-binding to CD8 antigen.
PD Modeling.
CD4+ T cells. A rapid decline
and slow recovery in the number of circulating
CD4+ T cells was evident at all three dose levels
for keliximab. In contrast, clenoliximab at a 5-mg/kg dose exhibited a
transient effect on CD4+ T cell number similar to
what was observed in placebo-treated animals (Fig. 5). The highest dose
of mAbs used (125 mg/kg) was high enough to produce full inhibition of
the system as evident by almost complete depletion of circulating
CD4+ T cells (Fig.
6). PD data at all three dose levels of
keliximab were fitted simultaneously to the proposed PK/PD model.
However, for clenoliximab, the 5-mg/kg dose was not included in
modeling and data for the 25- and 125-mg/kg dose were fitted
simultaneously. The fitted lines show that the effect versus time data
for both mAbs were adequately described by the indirect response model (Fig. 6). The fitted PD parameters for both keliximab and clenoliximab are listed in Table 2.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 6.
Time course of change in the number of circulating
CD4+ T cells relative to other lymphocytes for indicated
treatment. Symbols are the individual experimental data points for all
three doses (n = 5), and lines represent the
results of nonlinear mixed effect fitting of the indirect response
model to the data from all doses simultaneously.
|
|
The estimated values of zero-order
(kin) and first-order
(kout) rate constants describing the
production and loss of CD4+ T cells were similar
for both mAbs, as expected. Furthermore, the overall
CD4+ T cell turnover calculated from the
estimated parameters of the proposed PD model
[t1/2 = ln(2)/kout = 4.1 weeks] in this study is consistent with the literature values (Sprent and Tough, 1994; Tough
and Sprent, 1994). In general, the naïve T cells have a relatively slow turnover (t1/2 
5 weeks), whereas the memory T cells exhibit rapid turnover
(t1/2 ~ 2 weeks). The HuCD4/Tg mice
used in this study have an approximately 7:3 ratio of naïve to
memory phenotype.
Keliximab appeared to be more potent than clenoliximab in depleting
circulating CD4+ T cells as indicated by 10-fold
lower SC50 values for keliximab (Table 2). These
findings are supported by the results of the in vitro mechanistic
studies that showed that, compared with keliximab, clenoliximab has a
10- to 100-fold lower affinity for Fc receptor, a critical step
involved in the depletion of CD4+ T cells by
anti-CD4 mAbs. Moreover, clenoliximab does not cause any significant
lysis or apoptosis of CD4+ cells in vitro up to a
concentration of 25 µg/ml (Reddy et al., 2000). The joint
fitting of PD data described the CD4+ T cell
number versus time profiles well for keliximab (Fig. 6). However, the
model underpredicted the effect of 25 mg/kg clenoliximab especially at
the early time points. The observed higher suppression of
CD4+ T cell number at the early time points may
be due to the contribution of a placebo effect because clenoliximab has
low potency. For the effect of clenoliximab on
CD4+ T cells, the standard error of the
estimates (S.E.E.) of PD parameters was not estimated by the NONMEM
program because variability in the data was high and the population
size was limited.
CD4 Density.
A down-modulation of CD4 density on the surface
of T cells (CD4 MFI) occurred after a single i.v. administration of
keliximab or clenoliximab at all three dose levels (Fig.
7). The recovery of CD4 density was
relatively rapid after a single i.v. dose for both mAbs. The maximal
down-modulation of CD4 density (CD4 MFI Rmax) was <50% and achieved at 25 mg/kg dose of mAbs. Individual animal CD4 MFI versus time data for all
three doses of each mAb were fitted simultaneously to the proposed PD
model. The estimated PD parameters for keliximab and clenoliximab were
similar (Table 3).
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 7.
Time course of down-modulation of cell surface CD4 on
the surface of circulating T cells for indicated treatment. Symbols are
the individual experimental data points for all three doses
(n = 5), and lines represent the results of
nonlinear mixed effect fitting of the indirect response model to the
data from all doses simultaneously.
|
|
The values of maximum stimulatory factor attributed to mAbs were
estimated [Smax = (R0 Rmax)/Rmax]
from CD4 MFI versus time data obtained at the highest dose (125 mg/kg)
because this dose produced full inhibition of the pharmacological
system (Sharma and Jusko, 1998). The values of
Smax were then fixed to estimate other
PD parameters for keliximab and clenoliximab separately by nonlinear
mixed effect modeling. PD parameters for both mAbs were estimated with
a relatively reasonable precision. The fitted lines show that the CD4
MFI profiles for keliximab were adequately described by the proposed PD
model (Fig. 7). However, the joint fittings were only approximate for
clenoliximab CD4 MFI profiles especially for the 125 mg/kg dose group
which exhibited high variability (Fig. 7).
%coating.
Coating of the CD4 on the surface of circulating T
cells by keliximab or clenoliximab was evident in all animals receiving active treatment (Fig. 8). The duration
of saturated (100%) coating increased significantly with the increase
in dose for both mAbs. The coating persisted for at least 2, 4, and 9 days in all animals after a single i.v. administration of 5, 25, and
125 mg/kg dose of mAb, respectively (Fig. 8). Coating of CD4 by the
mAbs was not apparent when plasma concentrations of keliximab and
clenoliximab became nonquantifiable. As expected from the high affinity
of these anti-CD4 mAbs and the relatively small amount of CD4 antigen in circulation, relatively low blood concentrations of mAbs
resulted in complete antigen coating. The %coating profiles for all
three dose levels were simulated by the Langmuir adsorption isotherm (eq. 7) using in vitro affinity constant
(KA = 109
M
1 = 0.0067 liter/µg) and concentrations
predicted by the PK model. These simulated profiles based on the in
vitro affinity constant adequately describe the observed time course of
%coating in vivo (Fig. 8).
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 8.
Time course of CD4 occupancy on the surface of
circulating T cells for indicated treatment. Symbols are the individual
experimental data points for all three doses (n = 5), and lines represent the results of simulation using in vitro
affinity constant (KA) and plasma
concentrations of keliximab and clenoliximab predicted from typical
values of the PK parameters.
|
|
| |
Discussion |
Pharmacokinetics.
The disposition profiles of both keliximab
and clenoliximab in transgenic mice bearing human CD4 in place of
murine CD4 were similar. The plasma concentrations of both mAbs
declined rapidly after a duration that was dependent on the dose. The
marked nonlinearity of the PK of keliximab and clenoliximab may be
attributed to saturable binding of these mAbs to the cell surface CD4.
The steady-state volume of distribution estimated in this study was
higher (Vss = VC + VT = 800 ml/kg) than that reported for
other antibodies. This relatively high
Vss may be due to the specific binding
of these mAbs to CD4 in highly perfused organs such as spleen as well
as nonspecific uptake in the liver (Davis et al., 1996). These findings
are supported by the previous studies in which it was observed that the
disposition of keliximab was highly dependent on the presence and
distribution of the CD4 molecule (Davis et al., 1996; Davis and
Bugelski, 1998). It is plausible that the main pathway for the
disposition of these mAbs is mediated through binding to the human CD4
because murine CD4 knockout mice had substantially prolonged and higher
plasma concentrations of these mAb(s) relative to those in transgenic
mice expressing human CD4 (Davis et al., 1996).
Pharmacodynamics.
Keliximab and clenoliximab mediate their in
vivo immunomodulatory effects via indirect response mechanisms such as:
1) removal of CD4+ T cells from the peripheral
lymphoid organs, systemic circulation, and/or sites of inflammation; 2)
down-modulation of cell surface CD4; and/or 3) inhibition of CD4-MHC II
interactions resulting in the inhibition of T cell activation. The
removal of CD4+ T cells can occur via antibody
effector mechanisms such as induction of antigen-specific lysis or
apoptosis of CD4+ cells, and down-modulation of
cell surface CD4 may be due to internalization or stripping of CD4 from
the surface of T cell as has been shown in humans (Hepburn et al.,
1998).
In this study, the PD endpoints such as number of circulating
CD4+ T cells, density of CD4, and %coating of
CD4 on T cells were determined as surrogate markers for the
immunomodulatory effect of keliximab and clenoliximab in transgenic
mice expressing human CD4. The design of this study was suitable for
estimating parameters of the indirect PD model due to the following
features: 1) i.v. bolus doses were used and 2) the highest dose used
(125 mg/kg) was sufficient to produce full inhibition of the
pharmacological system as indicated by near complete depletion of
circulating CD4+ T cells and maximal
down-modulation of CD4 density (CD4 MFI
Rmax) at the 125-mg/kg dose.
An integrated PK/PD model was proposed to characterize the time course
of the effect of keliximab and clenoliximab on the surrogate markers
based on the mechanism of action of these mAbs. The model used is based
on the indirect-response concept that accounts for mechanistic delay in
production of responses and shift of the time of occurrence of maximum
response to the later times with the increase in dose (Dayneka et al.,
1993; Sharma and Jusko, 1996, 1998). Because of the indirect nature of
these responses, even as plasma concentrations of keliximab and
clenoliximab approached nonquantifiable levels and CD4 coating returned
to the baseline value, effects on the circulating
CD4+ T cells (depletion of T cells and
down-modulation of CD4 density) apparently persisted. The gradual
return to baseline can be explained by the magnitude of the rate
controlling input of new CD4+ T cells into the
blood and the expression of CD4 molecule on the surface of T cells.
Clenoliximab was designed to attenuate the CD4+
cell-depleting potential of keliximab but to retain immunomodulatory
effects. The depletion of CD4+ cells, a concern
in treatment of diseases requiring chronic administration, may occur
through binding of the Fc domain of keliximab to Fc receptors on the
effector cells and the F(ab) domain to CD4+
cells. This study showed that, unlike keliximab, clenoliximab at low
doses (5 mg/kg) did not exhibit a significant effect on circulating
CD4+ T cell count compared with placebo.
Furthermore, clenoliximab appeared to be significantly less potent and
efficient than keliximab in causing depletion of circulating
CD4+ T cells. These finding are consistent with
the results of an antibody-dependent cellular cytotoxicity assay using
chromium-labeled CD4+ SUPT1.18 cells in which
clenoliximab and keliximab were compared for their in vitro
cytotoxicity (Reddy et al., 2000). The percentage-specific lysis for
clenoliximab (<10%) was significantly lower than that for keliximab
(~30%) in the concentration range evaluated (0-25 µg/ml).
Furthermore, clenoliximab has 10- to 100-fold lower affinity for Fc
receptor in vitro than does keliximab (Reddy et al., 2000). However, in this PK/PD study, clenoliximab at doses of 25 and 125 mg/kg
caused significant depletion of CD4+ T cells.
This may be explained by the differences in the concentrations achieved
in vivo in this study and those used in the in vitro study. The plasma
concentrations of clenoliximab at 25 and 125 mg/kg were >20-fold
higher (>500 µg/ml) than the maximum in vitro concentration
evaluated in the antibody-dependent cellular cytotoxicity assay.
Although human effector cells were used in the in vitro study, it is
possible that at these high concentrations clenoliximab can cause
antigen-specific lysis of CD4+ cells via binding
to Fc receptors on the mouse effector cells in this transgenic model in
a manner similar to keliximab.
Down-modulation of cell surface CD4, evident in all active treatment
groups after a single i.v. dose of keliximab or clenoliximab, may be
caused in part by antibody-mediated stripping as observed in humans. It
is important to recognize that these mAbs cannot strip the majority of
CD4 from the surface of T cells even at very high doses, and thus the
estimated value of the maximum stimulatory factor was low (0.40-0.54)
for both mAbs. It is plausible that receptor-receptor interactions are
involved in down-modulation and, below 60 to 70% residual density, the
frequency of these interactions is reduced. Similar observations were
made in the clinical data from RA patients (Mould et al., 1999).
In conclusion, keliximab and clenoliximab cause a dose-dependent
decrease in the number of circulating CD4+ T
cells and down-modulation of CD4 on the surface of T cells. We have
presented an indirect response model to characterize the time course of
effects of two anti-CD4 mAbs on CD4+ T cells in
transgenic mice expressing human CD4. The findings of this study are
similar to the results from clinical trials at comparable doses. The
results of this study emphasize the importance of this transgenic model
for preclinical studies evaluating PK/PD of anti-human CD4 mAbs before
clinical development.
mAb, monoclonal antibody;
ABEC, area between
the baseline and the effect curve;
%coating, fraction of CD4 occupied
by the antibody;
HuCD4/Tg, transgenic mice bearing human CD4;
CD4 MFI, fraction mean fluorescence intensity;
SC50, plasma
concentration of mAb producing 50% of maximum stimulation;
RA, rheumatoid arthritis;
MHC, major histocompatibility complex;
PK, pharmacokinetic(s);
PD, pharmacodynamic(s).
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Top
Abstract
Introduction
