Design and synthesis of orally active inhibitors of TNF synthesis as anti-rheumatoid arthritis drugs

Article abstract of DOI:10.1016/j.bmcl.2003.08.076

A novel series of TNF inhibitors was identified based on the screening of existing MMP inhibitor libraries. Further SAR optimization led to the discovery of a novel lead compound. Its synthesis, efficacy in experimental animal models, and pharmacokinetic

Full text of DOI:10.1016/j.bmcl.2003.08.076

Bioorganic & Medicinal Chemistry Letters 13 (2003) 3951–3954
Design and Synthesis of Orally Active Inhibitors of TNF
Synthesis as Anti-rheumatoid Arthritis Drugs
Jian Jeffrey Chen,* Nolan Dewdney, Xiaohong Lin, Robert L. Martin,
Keith A. M. Walker, Jane Huang, Frances Chu, Elsie Eugui, Anna Mirkovich,
Yong Kim, Keshab Sarma, Humberto Arzeno and Harold E. Van Wart
Roche Bioscience, 3401 Hillview Ave., Palo Alto, CA 94304, USA
Received 2 December 2002; accepted 31 August 2003
Abstract—A novel series of TNF inhibitors was identified based on the screening of existing MMP inhibitor libraries. Further SAR
optimization led to the discovery of a novel lead compound. Its synthesis, efficacy in experimental animal models, and pharmaco-
kinetic data are discussed.
# 2003 Elsevier Ltd. All rights reserved.
Rheumatoid arthritis (RA)1 is an autoimmune disease
of unknown etiology characterized by chronic inflam-
mation of symmetric, peripheral joints and, in some
cases, by progressive joint erosions. The proin-
flammatory cytokine, tumor necrosis factor-a (TNF) is
a key mediator of inflammatory processes in RA and in
animal models of this disease.2 TNF upregulates the
expression of enzymes called matrix metalloproteinases
(MMPs). Increased levels of several of these MMPs,
including collagenase-3, stromelysin-1, and gelatinases
A and B, have been implicated in joint destruction.3
TNF convertase (TACE) is a metalloproteinase that
releases soluble TNF from the cell surface of macro-
phages. The enzyme cleaves the 26 KD membrane-
bound form of TNF, thereby releasing soluble 17 KD
TNF. TACE is a member of the adamalysin/ADAM
subfamily of the metzincin superfamily, which also
includes astacins, serralysins and MMPs.8,9 Several non-
selective MMP inhibitors were first shown to inhibit
TNF release.10,11 In the last few years a large number of
TACE inhibitors, mostly succinimide, sulfonamide and
phosphonamide based, have been published.12 Here, we
would like to disclose our orally bioavailable sulfamide
based TACE inhibitors.13
The approval and subsequent commercial success of
biologics such as TNF neutralizing antibodies and
soluble receptors (Remicade1 and Enbrel1) for the
treatment of RA and Crohn’s disease have validated the
therapeutic effectiveness of TNF inhibition.4 Although
the current biological agents have ameliorated signs and
symptoms, the delivery of these agents is parenteral and
therapy is complicated by the development of ser-
ological responses against the biological therapies.5 An
orally administered agent that lowers endogenous TNF
and has the same efficacy as biological TNF antagonists
in rheumatoid patients would be an attractive drug for
development. An inhibitor of TNF convertase could
potentially accomplish this goal.6,7
Our starting point for the TACE project was the
screening of our existing MMP libraries. Our primary
screening assay used the inhibition of TNF release from
the human monocytic cell line, Monomac6 (MM6), sti-
mulated with lipopolysaccharide (LPS) and phorbol
myristate acetate (PMA).11 A number of leads were
identified from different chemical classes but most of
them were rejected for lead optimization due to difficult
synthesis and/or poor pharmacokinetic properties.
Compound 1 (Fig. 1), a relatively weak inhibitor of
TNF release (IC50, 5.8 mM), was chosen for further
optimization based on the simplicity of the core struc-
ture. It was quickly learned that introduction of a dou-
ble bond in the piperidine ring improved the potency by
a factor of 5 (see 2, MM6: 1.24 mM). A further
improvement in the potency was observed when the
*Corresponding author at current address: CR&D, M/S 29-1-B,
Amgen Inc., One Amgen Center Drive, Thousand Oaks, CA 91320-
1799, USA. Fax: +1-805-480-1337; e-mail: jianc@amgen.com
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.08.076

3952
J. J. Chen et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3951–3954
We speculated that the rapid clearance of the drug
might be caused by the metabolic oxidation of the
piperidine ring. We reasoned that we might be able to
modulate the clearance by the introduction of a polar
group on the piperidine ring. Replacement of the pip-
ecolic acid by piperazic acid provides a site for such
modification. Table 2 lists some of the SAR results
obtained.
From the results in Table 2 it is clear that various sub-
stituents are tolerated on the nitrogen. The parent com-
pound (16) is a much weaker inhibitor of TNF release
although alkylation with trifluoroethyl restored the
potency. The compounds were further screened in the
rat LPS-induced TNF inhibition assays. Rat PK data
were obtained to aid the compound selection. In the PK
study the compounds were administrated orally at
30 mg/kg in 90% propylene glycol/10% ethanol to male
Sprague–Dawley rats. Blood samples were collected at
different time points and analyzed by HPLC. The AUC
data (area under the concentration vs time curve calcu-
lated by the linear trapezoid rule) are listed in Table 2.
Base on the MM6 and AUC datad, compound 17 was
selected for further evaluation.
Figure 1. Lead identification.
amino acid was changed from alanine to pipecolic acid
(see 3, MM6: 430 nM).
We then looked at the effect of substitution on the
phenyl ring, see Table 1. It is clear that large groups
such as Ph at the para or meta positions of the phenyl
ring are not tolerated. Relatively speaking the meta
position will accommodate groups such as methoxy and
ethyl better than the para position. But overall, the p-F
or the unsubstituted analogues were the most potent
compounds in the MM6 assay. Since the fluoro group at
the para position might be expected to reduce the poten-
tial metabolic oxidation of the phenyl ring, we chose the
p-F as the optimal substituent on the phenyl ring.
Murine Collagen-Induced Arthritis (MCIA)
Compound 17 was tested in the MCIA assay, a chronic
model of RA.14 In two separate experiments, partial
reduction in the arthritic index for compound 17-treated
animals (dose 2–120 mg/kg po) was evident beginning 1
day after the start of dosing and persisting throughout
the 14-day treatment period (Table 3). The reduction in
arthritic index was 49–96% with a trend toward dose
dependence. Statistically significant reductions in the
arthritic index were achieved. The incidence, obtained
by dividing the number of paws that were inflamed by
the total number of paws at risk, was reduced sig-
nificantly at all doses in both experiments. Compound
17 at 120 mg/kg also significantly reduced severity,
quantitated by averaging the maximum scores of the
Several compounds from Table 1 were tested in the in
vivo rat TNF inhibition assay. The inhibitors were
dosed orally 30 min before LPS was injected intra-
venously. Blood was collected 90 min after LPS injec-
tion. TNF was measured in serum by ELISA. Only
compound 5 gave >50% of inhibition when dosed
orally at 60 mg/kg. PK for compound 5 in rat and dog
indicated that the compound was cleared rapidly, espe-
cially in dog, when dosed orally (see discussion under
Pharmacokinetics section).
Table 1. The effect of substituent on the phenyl ring
Table 2. The effect of substituents on the piperazine ring
Compd
X
MM6 (IC50, mMa)
Compd
Y
MM6
(IC50, nMa)
Rat PK, AUC
(ng h/mLb)
4
5
H
p-F
0.24
0.3
6
7
8
9
10
11
12
13
14
15
p-Cl
p-Br
p-OCH3
m-OCH3
p-Ph
0.48
1.2
15
0.26
19
7.6
3.3
1.2
16
17
18
20
21
22
23
H
CONMe2
3560
200
120
350
250
510
300
na
6180
1790
7940
3270
1540
3260
CONHMe
COmorpholino
SO2NMe2
PO(NMe2)2
CH2CF3
m-Ph
p-Oph
m,p-F2
m-Et
aValues are means of three experiments.
0.32
2.6
bArea under the concentration versus time curve calculated by the
linear trapezoid rule when dosed 30 mg/kg orally. The detection limit
for drug concentration was 50 ng/mL by HPLC/UV method.
p-CH2CH2CH3
aValues are means of three experiments.

J. J. Chen et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3951–3954
3953
Table 3. Compound 17 inhibits MCIA
Dose (mg/kg
qd po)
Max. score
(meanÆSD)
Inhibition Positive paws Severity
(%)
day 13 (%)
(mean)
Vehicle
8.7 (Æ5.1)
2.9 (Æ4.1)
2.7 (Æ4.1)
4.5 (Æ4.0)
3.8 (Æ4.6)
0.3 (Æ0.8)
70
3.2
2.7
2.4*
2.8
2.9
1.0**
2
5
15
45
120
67*
69*
49
56
96**
27**
30**
41**
33**
4**
*0.01<p <0.05; ** p <0.01.
Results from a single experiment are shown. Comparable results were
obtained in a separate experiment.
inflamed paws. Additionally, following administration
of 120 mg/kg orally complete protection of stifle, hock,
and interphalangeal joints were observed histologically.
Pharmacokinetics
As mentioned earlier, the objective of introducing a
polar functional group in 5 was to reduce its metabolic
clearance. Table 4 lists the comparison of plasma con-
centrations for compound 5 and 17 in rat and dog. It is
clear that the clearance of 17 is indeed much slower than
that of compound 5.
Scheme 1. Preparation of compounds 4–15.
Syntheses
Compounds mentioned in this paper were synthesized
according to the reactions shown in Schemes 1 and 2. The
key step in Scheme 1 was the direct coupling of the sulfa-
moyl chloride 27 with the acid 28 which had been silylated
with TMSCN first. The acid 29 was then converted to the
acid chloride and coupled with bis-(trimethylsilyl)hydrox-
amine to give the final hydroxamate after aqueous
workup. In Scheme 2, the acid 30 was first persilylated with
HMDS and reacted sequentially with BocON and the sul-
famoyl chloride 27. The intermediate 32 underwent reac-
tion with the appropriate electrophiles to give 33.
Scheme 3 shows the process route for large-scale (kg)
preparation of compound 17. It was made from the acid
30 in two steps. The key step involves the dissolution of
the dihydrochloride salt of acid 30 in HMDS and
coupling the persilylated 30 sequentially with dimethyl-
carbamoyl chloride and 27 to produce acid 34 in 62%
isolated yield. Acid 34 is converted to hydroxamic acid
17 via the acid chloride. The overall yield is 52%.
Scheme 2. Preparation of compounds 16–23.
Table 4. Comparison of the PK parameters for compound 5 and 17a
Concn (ng/mL)
5b
17
Rat Cmax
Rat C (@ 7 h)
3060
0
1050
233
Dog Cmax
Dog C (@ 2 h)
117
0
182
110
aDosed orally at 30 mg/kg. The detection limit for the drug was 50 ng/
mL using HPLC/UV method. The standard deviation for the HPLC
assay was 20% and among animals was 40–50%.
bRacemic material was used.
Scheme 3. Preparation of compound 17.

3954
J. J. Chen et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3951–3954
pion, B.; Connolly, K.; Conway, J.; Furdon, P.; Karp, S.;
Summary
Kidao, S.; McElroy, A.; Nichols, J.; Pryzwansky, K.; Schoe-
nen, F.; Sekut, L.; Truesdale, A.; Verghese, M.; Warner, J.;
Ways, J. Nature 1994, 370, 558.
From the initial weak TNF inhibitor lead (5.8 mM) we
were able to identify a series of potent and orally active
TNF inhibitors. Compound 17 has been shown to inhi-
bit the release of TNF in cells and in animals. It is active
in a chronic RA model (MCIA) when administrated
orally. The compound was advanced for further pre-
clinical evaluation.
12. (a) Only a few latest references are given: Duan, J. J.-W.;
Lu, Z.; Xue, C.; He, X.; Seng, J. L.; Roderick, J. J.; Wasser-
man, Z. R.; Liu, R.; Covington, M. B.; Magolda, R. L.;
Newton, R. C.; Trzaskos, J. M.; Decicco, C. P. Bioorg. Med.
Chem. Lett. 2003, 13, 2035. (b) Zask, A.; Gu, Y.; Albright,
J. D.; Du, X.; Hogan, M.; Levin, J. I.; Chen, J. M.; Killar,
L. M.; Sung, A.; DiJoseph, J. F.; Sharr, M. A.; Roth, C. E.;
Skala, S.; Jin, G.; Cowling, R.; Mohler, K. M.; Barone, D.;
Black, R.; March, C.; Skotnicki, J. S. Bioorg. Med. Chem.
Lett. 2003, 13, 1487. (c) Sawa, M.; Kurokawa, K.; Inoue, Y.;
Kondo, H.; Yoshino, K. Bioorg. Med. Chem. Lett. 2003, 13,
2021.
References and Notes
1. For a general review on RA, see: Smolen, J. S.; Seiner, G.
Nat. Rev. Drug Discov. 2003, 2, 473.
2. For a general review see: Newton, R. C.; Decicco, C. P. J.
Med. Chem. 1999, 42, 2295.
3. For the latest review on MMP, see: Clark, I. M.; Parker,
A. E. Exp. Opin. Ther. Targets 2003, 7, 19.
13. Broka, C. A.; Campbell, J. A.; Castelhano, A. L.; Chen, J.
J.; Hendricks, R. T.; Melnick, M. J.; Walker, A. M. US Patent
6143744, 2000; Chem. Abstr. 2002, 136, 340996.
14. Induction and assessment of collagen-induced arthritis:
male DBA/1J (6–8 weeks old) mice were immunized intrader-
mally in the base of the tail with 0.1 mL of an emulsion of
100 mg of bovine type II collagen (Chondrex, Seattle, WA,
USA) in Freund’s complete adjuvant on day 0. Four weeks
after the immunization, when some animals began to show
signs of inflammation, mice were challenged with an intraper-
itoneal injection of 0.5 mL (50 mg) of lip polysaccharide (LPS,
E. coli, Sigma, Cat.# L3129). Mice were dosed orally once a
day with either vehicle or compound 17 at the doses indicated
in Table 3, beginning 24 h after the LPS boost and lasting 14
days. The extremities were evaluated three times each week
and the severity of the arthritis on each paw was scored on a
scale of 0–4. The score 0 represented a normal paw and a score
of 4 represented severe arthritis on the digits and paws with
the mouse avoiding weight-bearing on the paw. Groups of 10–
12 mice were used for each dose group and for the vehicle
controls.
4. For the latest review on anti-TNF biologics, see: Machold,
K. P.; Smolen, J. S. Exp. Opin. Biol. Ther. 2003, 3, 351.
5. Day, R. Lancet 2002, 359, 540.
6. Nelson, F. C.; Zask, A. Exp. Opin. Invest. Drugs 1999, 8,
383.
7. Moss, M. L.; White, J. M.; Lambert, M. H.; Andrew, R. C.
Drug Discov. Today 2001, 6, 417.
8. Stocker, W.; Grams, F.; Baumann, U.; Reinemer, P.;
Gomis-Ruth, F.-X.; Mckay, D. B.; Bode, W. Protein Sci.
1995, 4, 823.
9. Hooper, N. M. FEBS Lett. 1994, 354, 1.
10. Gearing, A.; Beckett, M.; Christodoulou, M.; Churchill,
M.; Clements, J.; Davidson, A.; Drummond, A.; Galloway,
W.; Gilbert, R.; Gordon, J.; Leber, T.; Mangan, M.; Miller,
K.; Nayee, P.; Owens, K.; Patel, S.; Thomas, W.; Wells, G.;
Wood, L.; Wooley, K. Nature 1994, 370, 555.
11. McGeehan, G.; Becherer, J.; Bast, R.; Boyer, C.; Cham-