This study validated a rat pharmacokinetic/pharmacodynamic model to rapidly assess drug candidates intended to inhibit tumor necrosis factor-alpha (TNFα) synthesis or release for inflammatory diseases. Lipopolysaccharide was administered to rats to induce TNFα production, and a selective TNFα converting enzyme inhibitor was used as a model compound. The model demonstrated an ability to correlate plasma drug concentrations with inhibition of lipopolysaccharide-induced TNFα levels in vivo. Areas under the concentration-time curves were calculated for the drug and TNFα to determine the overall percentage reduction of TNFα release. This PK/PD model provides integrated information on pharmacokinetics and in vivo potency of test articles.

1. Original article
A rat pharmacokinetic/pharmacodynamic model for assessment of
lipopolysaccharide-induced tumor necrosis factor-alpha production
Qin Wang a,⁎, Yuhua Zhang b,1
, J. Perry Hall b
, Lih-Ling Lin b
, Uma Raut a
, Nevena Mollova a,2
,
Neal Green c
, John Cuozzo b,3
, Shannon Chesley a
, Xin Xu a
, Jeremy I. Levin c
, Vikram S. Patel a
a
Department of Discovery Pharmacokinetics, Wyeth Research, 1 Burtt Road, Andover, MA 01810, USA
b
Department of Inflammation, 200 Cambridge Park Drive, Cambridge, MA 02140, USA
c
Department of Chemical and Screening Science, 200 Cambridge Park Drive, Cambridge, MA 02140, USA
Received 25 October 2006; accepted 14 January 2007
Abstract
Introduction: Tumor necrosis factor-alpha (TNFα) participates in many inflammatory processes. TNFα modulators show beneficial effects for the
treatment of many diseases including rheumatoid arthritis. The purpose of this study was to validate a rat pharmacokinetic/pharmacodynamic (PK/PD)
model for rapid assessment of drug candidates that intended to interrupt TNFα synthesis or release. Methods: Rats received intravenous (IV) or oral
administrations of test article or dose vehicle, followed by LPS challenge. Plasma levels of test article and TNFα were determined. The areas under the
concentration–time curves (AUCdrug and AUCTNFα) were calculated. The overall percentage of inhibition on TNFα release in vivo was calculated by
comparing AUCTNFα of the test article treated group against that for the vehicle control group. Results: The dosing vehicles tested in this study did not
increase plasma TNFα level. At IV dose of up to 100 μg/kg, LPS did not alter the pharmacokinetics of the compound tested. Using a selective TNFα
converting enzyme (TACE) inhibitor as model compound, this PK/PD model demonstrated its ability to correlate plasma test article concentration with
its biological activity of lowering the LPS-induced TNFα plasma levels in vivo. Discussion: A rat PK/PD model for evaluation of the effect of drug
candidates on LPS-induced TNFα synthesis and/or release has been investigated. This model provides integrated information on pharmacokinetics and
in vivo potency of the test articles.
© 2007 Elsevier Inc. All rights reserved.
Keywords: PK/PD model for TNFα release; Rat LPS model
1. Introduction
Recent advances in our understanding of the pathogenesis of
rheumatoid arthritis (RA) reveal that tumor necrosis factor-alpha
(TNFα) plays a pivotal role in progression of the disease. Three
recombinant proteins that neutralize TNFα effectively alleviate
the clinical symptoms of RA. Etanercept, a fusion protein of the
soluble form of the human TNFα receptor p75 and the Fc portion
of human IgG1, binds to circulating TNFα with high affinity and
thereby, blocks the biological functions of TNFα. Infliximab, a
chimeric monoclonal antibody that contains a murine TNFα
binding variable region within human immunoglobulin G1
(IgG1), prevents the interaction of TNFα with its cellular
receptor signaling targets (Nahar, Shojania, Marra, Alamgir, &
Anis, 2003). Adalimumab, a fully humanized anti-TNFα anti-
body, works by a similar mechanism (Olsen & Stein, 2004;
Shanahan & St. Clair, 2002; Braun & Sieper, 2004). Based upon
the success of these biopharmaceutical therapies, it is believed
that reducing TNFα production or release can curtail or even halt
joint damage and the progression of RA (Dayer, 2004), and
perhaps, ameliorate other diseases in which TNFα plays a role
(Stokes & Kremer, 2003). There have been many attempts to
develop new medications that modulate TNFα production or
release, especially orally available small synthetic molecules.
Tumor necrosis factor-alpha converting enzyme (TACE, Black,
2002; Conway et al., 2001; Skotnicki, DiGrandi, & Levin, 2003)
Journal of Pharmacological and Toxicological Methods 56 (2007) 67–71
www.elsevier.com/locate/jpharmtox
⁎ Corresponding author. Tel.: +1 978 247 1364; fax: +1 978 247 2842.
E-mail address: qwang@wyeth.com (Q. Wang).
1
Current address: P&G Pharmaceuticals, Cincinnati, OH 45040, USA.
2
Current address: CV therapeutics, Palo Alto, CA 94304, USA.
3
Current address: Praecis Pharmaceuticals, Waltham, MA 02451, USA.
1056-8719/$ - see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.vascn.2007.02.001

2. and p38 mitogen activated protein kinase (p38 MAPK,
Schieven, 2005) are among the most studied therapeutic targets
in this regard.
The major sources of TNFα are monocytes and macro-
phages. Other cell types that produce TNFα include T
lymphocytes, NK cells, dendritic cells, mast cells, endothelial
cells, smooth muscle cells, osteoblasts, and astrocytes (Aggar-
wal, Samanta, & Feldmann, 2001). Avariety of noxious stimuli,
ranging from physical to chemical and immunological, can
induce TNFα production and release, both in vitro and in vivo.
For example, after administration of bacterial Lipopolysaccha-
ride (LPS) to animals, the blood levels of TNFα increase
rapidly. This, in turn, causes the production of pathophysiolog-
ical changes similar to those observed under inflammatory
conditions (Feldmann & Maini, 2001). Although the relevance
of LPS-stimulated pathways to the pathogenesis of inflammatory
diseases such as RA remains to be determined, it is nevertheless a
common practice to evaluate drug candidates by monitoring the
inhibition of LPS-induced TNFα production (Beck et al., 2002;
Mclay et al., 2001; Zhang, Xu et al., 2004). From drug screening
points of view, it is necessary to link test articles in vitro potency
with its in vivo activities. The information obtained from such
relationship helps to design dosing regimens for chronic disease
models such as collagen-induced arthritis (CIA).
In this report, using a selective TACE inhibitor, TMI-2
(Zhang, Hegen et al., 2004), we have validated a rat PK/PD
model for rapid screening and ranking of drug candidates that
are intended to inhibit TNFα synthesis or release for the
treatment of inflammatory diseases, such as RA.
2. Materials and methods
2.1. Chemicals and test materials
Unless specified, all chemicals, reagents and excipients of
dose formulations, were obtained from Aldrich-Sigma (St
Louis, MO) and used as received. The dose of LPS (E. coli
O111: B4, Lot # B59097, Aldrich-Sigma) was prepared in a
phosphate buffered saline solution (PBS, pH 7.2 to 7.4).
Deionized water, saline and PBS were freshly prepared in the
laboratory. Test article, TMI-2, was synthesized at medicinal
chemistry laboratory of Wyeth Research. Saline and an organic
dose vehicle (N-methyl-pyrrolidone: polyethylene glycol 400:
propylene glycol, 10:40:50, v/v/v) were used as IV dose ve-
hicles for hydrophilic and lipophilic compounds, respectively.
Oral dose formulation was an aqueous suspension containing
2% polysorbate 80 and 0.5% methylcellulose. The dose volume
for the test article was 1 mL/kg for IV and 5 mL/kg for oral
administration. All doses, LPS and test articles, were freshly
prepared on the day of the study.
2.2. Procedures for animal studies
The study was performed at Wyeth Research Laboratory
(Andover, MA) under the supervision of the Institutional
Animal Care and Use Committee (IACUC). The animals used
in the study were female Lewis rats (Charles River Laboratories,
Wilmington, MA) and weighed between 200 to 350 g. Rats had
jugular vein catheters surgically implanted prior to their arrival
at Wyeth. Briefly, the animals, divided into various treatment
and control groups (n=4 per group), received single doses of
test article or dosing vehicle via either tail vein injection or oral
gavage. The TMI-2 dose were 5 mg/kg for IVand 10 mg/kg for
oral administration, respectively. Blood samples of approxi-
mately 300 μL were collected at various time points through
jugular vein catheter into pediatric plasma separator tubes
containing K2-EDTA. The total blood lost per animal was less
than 15% of rat blood volume. Immediately after the 15 min
time point blood collection, an IV dose of LPS (100 μg/kg) in
PBS was administered via the jugular vein catheter followed by
approximately 200 to 250 μL of blank saline. Blood samples
collection continued after LPS challenge. Plasma samples
were stored at −80 °C until assay for TNFα and test article
concentrations.
2.3. Determination of plasma TNFα concentration
The detailed procedures of in vitro IC50 determination of
TACE inhibitors in rat and human whole blood were described
in a previous report (Zhang, Hegen et al., 2004). Briefly, freshly
collected whole blood was fortified with a test article of various
concentrations, followed by stimulation with LPS (100 ng/mL)
and incubation at 37 °C for about 3 h under gentle rotation. The
blood samples were then centrifuged at 1500 rpm for 15 min.
The plasma samples were collected and frozen at −80 °C until
analysis. The plasma TNFα levels, from both in vitro and in
vivo studies, were determined by a commercial ELISA kit
(BioSource International, Camarillo, CA).
2.4. Liquid Chromatography–Mass Spectrometry (LC–MS/MS)
assay for plasma test article levels
The separation of test articles from plasma samples was
performed on a Perkin Elmer Series 200 HPLC system (Perkin
Elmer, Norwalk, CT) using a XTerra MS C18 column
(2.1×20 mm, 2.5 μm; Waters, Milford, MA). The mobile
phase consisted of Solvent A (0.1% HCOOH in H2O) and
Solvent B (0.1% HCOOH in acetonitrile). The detection of test
article was performed on a PE SCIEX API-3000 triple quad-
rupole mass spectrometer (Applied Biosystems, Concord,
Ontario, L4K4V8) using a TurboIon Spray source. Briefly, an
aliquot of 50 μL plasma was precipitated with 100 μL
68 Q. Wang et al. / Journal of Pharmacological and Toxicological Methods 56 (2007) 67–71

3. acetonitrile containing 500 ng/mL of internal standard, a
compound with similar chemical structure as that for the test
article, in a 0.5 mL 96-well plate. The mixtures were vortexed
and centrifuged at 5700 rpm for 10 min. Supernatants were
directly subjected to LC–MS/MS system for analysis. The
gradient for HPLC elution was isocratic at 0% of Solvent B for
1 min, followed by a linear increase to 100% of Solvent B over
3 min. The column was allowed to equilibrate with 0% Solvent B
for 2 min before the next injection. The flow rate was 0.2 mL/
min and the injection volume was 10 μL. The plasma standard
curves were generated by plotting peak area ratio of test article
and internal standard against nominal concentrations. The limit
of quantitation for TMI-2 in rat plasma was 1 ng/mL.
2.5. Pharmacokinetic–pharmacodynamic (PK/PD) analysis
The pharmacokinetic parameters were computed using
WinNonlin (version 4.1, Pharsight, Mountain View, CA) via
non-compartmental analysis approaches. The estimation of the
area under the plasma concentration versus time curve (AUC) of
TMI-2 (AUCTMI-2) was based upon log–linear trapezoidal rule.
However, the AUC of TNFα (AUCTNFα) was computed by
linear trapezoidal method. The overall percentage of reduction
on TNFα release was calculated by the AUCTNFα ratio of TMI-
2 treated versus dose vehicle treated control groups.
For PK/PD modeling, a first order two compartment model
for IV or a first order no lag time one compartment model for
oral administration of the test article, and a Hill equation
(inhibitory effect Emax sigmoid model) were best fitted for the
plasma TMI-2 and TNFα concentration–time profiles (model
106 in WinNonlin), respectively. No statistical analysis was
conducted.
3. Results
3.1. Effect of dose vehicles on plasma TNFα concentration
Given that many chemicals stimulate TNFα release, it was
important to examine and select dose vehicles that did not
produce such effects. For the two IV dose formulations, saline
and in particular, the organic vehicle (N-methyl-pyrrolidone:
polyethylene glycol 400: propylene glycol, 10:40:50, v:v:v),
there were no increases in plasma TNFα concentrations. The
average baseline TNFα levels were below 0.6 ng/mL. After
IV bolus injection of LPS at 100 μg/kg, plasma TNFα started
to rise by approximately 30 min and reached its peak
concentration at approximately 90 min. The plasma TNFα
level then underwent rapid decline to baseline by about 3 to
4 h post LPS administration (Fig. 1). The average Cmax of
TNFα ranged between 8 and 10 ng/m with mean plasma
AUCTNFα value of approximately 14–18 ng×hr/mL. As ex-
pected, the aqueous oral dose vehicle did not increase plasma
TNFα levels.
3.2. Effect of LPS on the absorption and disposition of test
article
LPS is a toxic agent that induces a wide array of biological,
biochemical, and immunological reactions (Ray, 1999). There-
fore, it was necessary to validate the effects of concomitant LPS
Fig. 1. Organic dose vehicle of 1 mL/kg (N-methyl-pyrrolidone: polyethylene
glycol 400: propylene glycol, 10:40:50, v:v:v, IV, open circle) did not cause
observable changes in plasma TNFα after 100 μg/kg LPS challenge.
Fig. 2. After IV (2a, 5 mg/kg) and PO (2b, 10 mg/kg) administrations, the
plasma concentration–time profiles of TMI-2 were not altered by concomitant
administration of 100 μg/kg LPS dose.
69Q. Wang et al. / Journal of Pharmacological and Toxicological Methods 56 (2007) 67–71

4. administration on the pharmacokinetics of test articles in
animals. Following IVor oral administration to rats, the plasma
concentration–time profiles of TMI-2 were essentially super-
imposable to those obtained from studies in which LPS was not
dosed (Fig. 2a and b).
3.3. Effect of TNFα modulator on plasma TNFα levels upon
LPS challenge
The plasma concentration–time profiles of TNFα, after IV
or oral administration of TMI-2 followed by LPS challenge,
were plotted in Fig. 3a and b. The overall reduction of TNFα
(AUCTNFα) was 69% after IV treatment at AUCTMI-2 of
2266 ng×h/mL and 81% after oral treatment at AUCTMI-2 of
1450 ng×h/mL. The lesser overall reduction of TNFα
release on higher AUCTMI-2 for IV treatment group could
be due to the timing of the LPS challenge, which occurred at
15 min post TMI-2 dose administration. By then, plasma
concentration had dropped significantly so that a large
portion of the plasma exposure of TMI-2 did not participate
in the inhibition of LPS induced TNFα production–release
process. Since TACE is at the last step of TNFα release and is a
membrane protein of blood cells, test article in plasma directly
interacts with the target. An inhibitory effect sigmoid Emax
model (Eq. (1)) best fitted the PK/PD link model.
E ¼ E0 þ Emax−E0ð Þ Â
Cg
plasma
Cg
plasma þ ECg
50
ð1Þ
where E is the percent of reduction of TNFα, at any given
time, assuming complete prevention of TNFα release (Emax)
can be achieved if plasma test article level (Cplasma) is
sufficiently high. E0 is the baseline from vehicle treated
group. EC50 is 50% reduction in TNFα release (Fig. 4). Here,
γ is the curve shape factor for the Hill equation (Nigrovic &
Amann, 2002).
Based on the curve fitting, the in vivo EC50 for TMI-2 was
approximately 303 ng/mL and 71 ng/mL after IV and oral dose
administration, respectively. The results were in agreement with
in vitro EC50 of 70–100 ng/mL for TMI-2 from rat whole blood
assay.
4. Discussion
LPS-induced TNFα production and release can be simplified
as a series of events that starts with LPS binding to LPS binding
protein (LPB), and that complex then binds to CD14 which
initiates a signaling cascade that culminates in expression of the
tnfα gene, processing of the TNFα precursor protein, insertion
of the TNFα precursor in the cell membrane, cleavage of the
precursor, and finally, release of the active form of TNFα, a
17 kDa soluble fragment, into the extra-cellular spaces
(Anderson, Phillips, Stoecklin, & Kedersha, 2004; Gracie,
Leung, & McInnes, 2002). Alterations in the concentration of
TNFα in fresh blood (or cell culture media) can be expected if
the synthesis or release of TNFα is disrupted in animals (or cell
culture models) during the response to LPS challenge. By
Fig. 3. Plasma concentration–time profiles of TNFα after IV (3a, 5 mg/kg) and
oral (3b, 10 mg/kg) administrations of TMI-2 followed by 100 μg/ml LPS
stimulation.
Fig. 4. An inhibitory effect sigmoid Emax PK/PD link model was best fitted for
the reduction in plasma TNFα levels in the presence of TMI-2. The chart on the
upper right shows the plasma concentration–time profiles of TMI-2 after IV or
oral administration.
70 Q. Wang et al. / Journal of Pharmacological and Toxicological Methods 56 (2007) 67–71

5. choosing an appropriate PK/PD model, one can establish an in
vitro–in vivo correlation for drug candidates that are intended to
modulate this cascade.
At the early phase of drug discovery, drug candidates may
not possess needed potency, selectivity and/or pharmacokinetic
properties. By directly introducing test articles to animals via IV
injection at relatively higher doses, formulated in organic
vehicles if necessary, a PK/PD study can serve as a proof-of-
concept for target validation and a quick assessment of in vitro–
in vivo correlation. Unlike mouse model (Gozzi et al., 1999),
rats provide a quantity of blood volume per animal that is suffi-
cient for complete assessment of plasma (or serum) concentra-
tion–time profiles of both test article and TNFα. This engenders
less interference in the data from inter-animal variations.
In this report, using a selective TACE inhibitor as model
compound, we have studied a rat PK/PD model for validation
for drug candidates that are intended to modulate TNFα levels.
Acknowledgements
We thank Jessica Doherty, Cindy Clark, Amy Ignatowicz,
Terrie Cunliffe-Beamer, and Glen Pedneault for their excellent
technical supports.
References
Aggarwal, B. B., Samanta, A., & Feldmann, M. (2001). TNFα. In J. J.
Oppenheim & M. Feldmann (Eds.), Cytokine reference: A compendium of
cytokines and other mediators of host defense (1st ed.). San Diego:
Academic Press.
Anderson, P., Phillips, K., Stoecklin, G., & Kedersha, N. (2004). Post-
transcriptional regulation of proinflammatory proteins. Journal of Leukocyte
Biology, 76(1), 42−47.
Beck, G., Bottomley, G., Bradshaw, D., Brewster, M., Broadhurst, M., Devos,
R., et al. (2002). (E)-2(R)-[1(S)-(Hydroxycarbamoyl)-4-phenyl-3-butenyl]-
2′-isobutyl-2′-(methanesulfonyl)-4-methylvalerohydrazide (Ro 32–7315), a
selective and orally active inhibitor of tumor necrosis factor-α convertase.
Journal of Pharmacology and Experimental Therapeutics, 302, 390−396.
Black, R. A. (2002). Tumor necrosis factor-alpha converting enzyme. Interna-
tional Journal of Biochemistry & Cell Biology, 34(1), 1−5.
Braun, J., & Sieper, J. (2004). Biological therapies in the spondyloarthritides—
The current state. Rheumatology, 43(9), 1072−1084.
Conway, J. G., Andrews, R. C., Beaudet, B., Bickett, D. M., Boncek, V., Brodie,
T. A., et al. (2001). Inhibition of tumor necrosis factor-α (TNF-α)
production and arthritis in the rat by GW3333, a dual inhibitor of TNFα
converting enzyme and matrix metalloproteinases. Journal of Pharmacol-
ogy and Experimental Therapeutics, 298, 900−908.
Dayer, J. M. (2004). The process of identifying and understanding cytokines:
From basic studies to treating rheumatic diseases. Best Practice & Research.
Clinical Rheumatology, 18(1), 31−45.
Feldmann, M., & Maini, R. N. (2001). Anti-TNFα therapy of rheumatoid
arthritis: What have we learned? Annual Review of Immunology, 19,
163−196.
Gozzi, P., Pahlman, I., Palmer, L., Gronberg, A., & Persson, S. (1999).
Pharmacokinetic-pharmacodynamic modeling of the immunomodulating
agent susalimod and experimentally induced tumor necrosis factor-alpha
levels in the mouse. Journal of Pharmacology & Experimental Therapeu-
tics, 291(1), 199−203 (Oct).
Gracie, J. A., Leung, B. P., & McInnes, I. B. (2002). Novel pathways that
regulate tumor necrosis factor-alpha production in rheumatoid arthritis.
Current Opinion in Rheumatology, 14(3), 270−275.
Mclay, L. M., Halley, F., Souness, J. E., McKenna, J., Benning, V., Birrell, M.,
et al. (2001). The discovery of RPR 200765A, a p38 MAP kinase inhibitor
displaying a good oral anti-arthritic efficacy. Bioorganic & Medicinal
Chemistry, 9(2), 537−554.
Nahar, I. K., Shojania, K., Marra, C. A., Alamgir, A. H., & Anis, A. H. (2003).
Inflixmab treatment of rheumatoid arthritis and crohn's disease. Annals of
Pharmacotherapy, 37, 1256−1265.
Nigrovic, V., & Amann, A. (2002). Physiologic–pharmacologic interpretation of
the constants in the hill equation for neuromuscular block: A hypothesis.
Journal of Pharmacokinetics and Pharmacodanamics, 29(2), 189−206.
Olsen, N. J., & Stein, C. M. (2004). New drug for rheumatoid arthritis. New
England Journal of Medicine, 350, 2167−2179.
Ray, P. K. (1999). Stress genes and species survival. Molecular and Cellular
Biochemistry, 196(1–2), 117−123.
Shanahan, J. C., & St. Clair, E. W. (2002). Short analytical review tumor
necrosis factor-α blockage: A novel therapy for rheumatic disease. Clinical
Immunobiology, 103, 231−242.
Schieven, G. L. (2005). The biology of p38 kinase: A central role in
inflammation. Current Topics in Medicinal Chemistry, 5(10), 921−928.
Skotnicki, J. S., DiGrandi, M. J., & Levin, J. I. (2003). Design strategies for the
identification of MMP-13 and Tace inhibitors. Current Opinion in Drug
Discovery & Development, 6(5), 742−759.
Stokes, D. G., & Kremer, J. M. (2003). Potential of tumor necrosis factor
neutralization strategies in rheumatologic disorders other than rheumatoid
arthritis. Seminars in Arthritis and Rheumatism, 33(1), 1−18.
Zhang, Y., Xu, J., Levin, J., Hegen, M., Li, G., Robertshaw, H., et al. (2004).
Identification and characterization of 4-[[4-(2-butynyloxy)phenyl]sulfonyl]-
N-hydroxy-2,2-dimethyl-(3S)thiomorpholinecarboxamide (TMI-1), a novel
dual tumor necrosis factor-alpha-converting enzyme/matrix metalloprotease
inhibitor for the treatment of rheumatoid arthritis. Journal of Pharmacology
and Experimental Therapeutics, 309(1), 348−355.
Zhang, Y., Hegen, M., Xu, J., Keith, J., Jin, G., Du, X., et al. (2004).
Characterization of (2R, 3S)-2-([4-(2-butynyloxy)phenyl]sulfonyl amino)-
N,3-dihydroxybutanamide, a potent and selective inhibitor of TNF-α
converting enzyme. International Immunopharmacology, 4(14), 1845−1857.
71Q. Wang et al. / Journal of Pharmacological and Toxicological Methods 56 (2007) 67–71