Discovery of 3-OH-3-methylpipecolic hydroxamates: Potent orally active inhibitors of aggrecanase and MMP-13

Article abstract of DOI:10.1016/j.bmcl.2005.05.037

A series of 3-hydroxy-3-methylpipecolic hydroxamate inhibitors of MMP-13 and aggrecanase was designed based on the observation of increased aggrecanase activity with substitution at the 3-position of the piperidine ring. Potency versus aggrecanase was optimized by modification of the benzyloxyarylsulfonamide group that binds in the S1′ pocket. These compounds also possess markedly improved bioavailability and lower metabolic clearance compared to analogous 3,3-dimethyl-5-hydroxypipecolic hydroxamates. These improvements are attributed to lowered lipophilicity proximal to the metabolically labile hydroxamic acid. Synthesis, structure activity relationships, and in vivo efficacy data are described.

Full text of DOI:10.1016/j.bmcl.2005.05.037

Bioorganic & Medicinal Chemistry Letters 15 (2005) 3385–3388
Discovery of 3-OH-3-methylpipecolic hydroxamates: Potent
orally active inhibitors of aggrecanase and MMP-13
Mark C. Noe,* Vijayalakshmi Natarajan, Sheri L. Snow, Lilli A. Wolf-Gouveia,
Peter G. Mitchell, Lori Lopresti-Morrow, Lisa M. Reeves, Sue A. Yocum, Ivan Otterness,
Marcia A. Bliven, Thomas J. Carty, John T. Barberia, Francis J. Sweeney,
Jennifer L. Liras and Marcie Vaughn
Pfizer Global Research and Development, Groton Laboratories, Eastern Point Road, Groton CT 06340, USA
Received 22 March 2005; revised 6 May 2005; accepted 9 May 2005
Available online 13 June 2005
Abstract—A series of 3-hydroxy-3-methylpipecolic hydroxamate inhibitors of MMP-13 and aggrecanase was designed based on the
observation of increased aggrecanase activity with substitution at the 3-position of the piperidine ring. Potency versus aggrecanase
was optimized by modification of the benzyloxyarylsulfonamide group that binds in the S10 pocket. These compounds also possess
markedly improved bioavailability and lower metabolic clearance compared to analogous 3,3-dimethyl-5-hydroxypipecolic
hydroxamates. These improvements are attributed to lowered lipophilicity proximal to the metabolically labile hydroxamic acid.
Synthesis, structure activity relationships, and in vivo efficacy data are described.
ꢀ 2005 Elsevier Ltd. All rights reserved.
Osteoarthritis is a highly debilitating disease character-
ized by progressive cartilage degradation that leads to
pain and reduced motility in affected joints. The degen-
eration of proteoglycan and disruption of the type II
collagen network in cartilage are pathological processes
that reduce joint function.1 Inhibition of the enzymes
that mediate cartilage catabolism (aggrecanases and col-
lagenases) is a strategy for developing disease modifying
therapies for osteoarthritis. Previously, we described the
discovery of potent inhibitors of MMP-13 (collagenase
3) and aggrecanase activity with selectivity versus
MMP-1 and TACE based on the series exemplified by
1.2 Herein we describe another series of pipecolic
hydroxamates with similar functional activity but
greatly improved metabolic stability and bioavailability.
donor and acceptor groups present in the protein.3
However, the presence of a hydroxamic acid presents
significant challenges for the optimization of pharmaco-
kinetic properties necessary for an effective oral therapy.
The hydroxamic acid group is particularly prone to
hydrolysis, reduction, and glucuronidation.4 As part of
the strategy to reduce the metabolic lability of our lead
aggrecanase series 1, we explored the effects of introduc-
ing polar functionality at the 3-position of the piperidine
ring as exemplified by 2.
Appropriately substituted hydroxamic acids are effective
competitive inhibitors of matrix metalloproteases,
ADAMS and ADAM-TS family members. Structural
studies point to a critical bidentate interaction between
the active site zinc atom and the hydroxamic acid ligand
with additional hydrogen bond interactions with the
active site glutamic acid residue and other H-bond
The synthesis of these compounds was carried out as de-
scribed in Scheme 1. Starting with glycine tert-butyl es-
ter, sulfonylation with benzyloxyphenylsulfonyl chloride
afforded the sulfonamide 4 in 81% yield. Alkylation of
*
Corresponding author. Tel.: +1 860 441 8825; fax: +1 860 715
2767; e-mail: mark.c.noe@pfizer.com
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.05.037

3386
M. C. Noe et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3385–3388
Scheme 1. Synthesis of 3-hydroxy-3-alkylpipecolic-3-hydroxamic acids from glycine tert-butyl ester.
the sulfonamide nitrogen using the appropriate c-
hydroxyketone under Mitsunobu conditions afforded
the cyclization precursor 5 in moderate yield. Intramo-
lecular aldol condensation proceeded in somewhat vari-
able yield that was dependent on the size of the ketone
substituent R0 and the type of base. For substrate 5
where R0 = methyl, potassium tert-butoxide was the
optimum base, affording the cyclization product in ca.
70% yield. However, compounds with larger R0 groups
failed to cyclize under these conditions, and a much
cleaner reaction was observed using LDA in THF
initially at ꢀ78 ꢁC. In either case, a 1:1 to 8:1 mixture
of diastereomers was obtained, depending on the size
of the R0 group. The resulting pipecolate ester diastereo-
mers could be separated by silica gel chromatography.
Preparative chiral separation using a Chiralpak AD
column gave each of the four stereoisomers of 6. Depro-
tection of the phenol, followed by alkylation with the
appropriate alkyl halide and acidic removal of the
tert-butyl ester afforded the pipecolic acid 8. Formation
of the hydroxamate was best accomplished by coupling
with O-allylhydroxylamine–HCl in the presence
of EDCI and HOBT with sufficient HunigÕs base to
neutralize the hydrochloride salt. Palladium catalyzed
deallylation afforded the test compounds in 41% overall
yield from the pipecolic acid.5
The effect of stereochemistry and substitution at the
3-position of the piperidine ring was examined first. Bio-
logical data for several of these analogues appear in
Scheme 2.6 Both aggrecanase and MMP-1 potency are
slightly influenced by the pendant alkyl group and ste-
reochemistry of the 3-position, and compound 11 exhib-
ited the most potent inhibition of aggrecanase. Using
this compound as the lead, we next explored SAR at
the aryl sulfonamide. Biological data for these ana-
logues are shown in Table 1.
As observed for the 3,3-dimethyl-5-hydroxypipecolic
hydroxamates, incorporation of a lipophilic ortho-sub-
stituent in the benzyl ether group results in potent agg-
recanase activity.2 This observation is highlighted by
the dramatic effect of placement of a fluoro substituent
on activity in substituted benzyl ether analogues
15–17. The reduced potency observed for the ortho cya-
no substituted analogue 22 and the ortho methoxy
Scheme 2. Aggerecanase and MMP inhibitory activity of 3-hydroxy-3-alkyl pipecolic hydroxamates.

M. C. Noe et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3385–3388
3387
Table 1. Aggrecanase and MMP-13 inhibitory activity of benzyl ether analogues of 9 (R0 = (3R) Me)
Compound
R
Aggrecanase
IC50 (nM)
MMP-13
IC50 (nM)
MMP-1
IC50 (nM)
rTACE
IC50 (nM)
TNF HWB
IC50 (uM)
15
16
17
18
19
20
21
22
23
24
25
26
27
2-Fluorophenyl
3-Fluorophenyl
14
85
0.49
1.5
180
340
—
—
14
—
—
—
—
—
10
15
—
—
—
33
15
4-Fluorophenyl
2-Chlorophenyl
120
3.7
78
3.7
1.0
420
270
7.2
8.3
17
3-Chlorophenyl
2-Methylphenyl
2.5
860
310
6.9
73%
55%
11
0.65
18
9.7
2.2
14
2-Methoxyphenyl
2-Cyanophenyl
2600
2700
1300
220
100
1.6
2.5
1.9
10.4
73
2-Methyl-3-fluorophenyl
2-Methyl-4-fluorophenyl
2-Methyl-5-fluorophenyl
2-Chloro-4-fluorophenyl
2-Fluoro-4-chlorophenyl
3.2
2.9
4.7
15
2.5
0.9
600
310
1.5
7.9
0.87
120
Values expressed as a percentage represent percent inhibition of enzyme activity at 500 nM.
substituted analogue 21 relative to the ortho methyl (20),
chloro (18) or fluoro (15) analogues is consistent with
the notion that a lipophilic residue at this position im-
parts potent aggrecanase activity. MMP-1 inhibition is
also sensitive to the presence of a lipophilic ortho substi-
tuent, and a ꢁ8-fold reduction in MMP-1 activity is ob-
served for 21 relative to 20. The addition of a second
halogen substituent in the para position results in com-
parable potency for ortho alkyl or halo substituted ana-
logues (compare 24 with 20 and 26 with 18). Selectivity
for aggrecanase over MMP-1 inhibition improves for
2,3-disubstituted benzyl ethers relative to the monosub-
stituted analogues (compare 23 with 20). The 2,4-disub-
stituted compounds provide the best aggrecanase
potency, with the 2-chloro-4-fluoro analogue 26 provid-
ing low nanomolar inhibitory activity for aggrecanase
and MMP-13, and > 100· selectivity over MMP-1.
versus MMP-1 and poor potency against TNF-a release
in LPS treated whole blood. The complete MMP-inhibi-
tion profile for this compound appears below:
MVP
1
2
4
3
5
8
1
9
13
14
83
IC50 (nM)
310
11
0.9
Compound 26 showed potent inhibition of IL-1 induced
aggrecan (IC50 = 10.3 nM) and collagen (IC50 = 100 nM)
degradation in bovine nasal cartilage explants.7 In human
osteoarthritic cartilage, 26 inhibited IL-1 induced aggre-
can degradation with an IC50 of 6.0 nM (n = 2).8
Compound 26 showed moderate clearance in rat
(Clp = 13 mL minꢀ1 kgꢀ1, t1/2 = 2 h) and dog (Clp =
11 mL minꢀ1kgꢀ1, t1/2 = 4.6 h) that was consistent with
predicted clearance based on incubations with rat and
dog hepatocytes (ER = 0.20 and 0.28, respectively).9,10
The markedly lower metabolic clearance with this com-
pound compared to the original lead 1 (R = 2,4-di-
chloro; rat Clp = 44 mL minꢀ1 kgꢀ1, t1/2 = 2 h; dog
Clp = 25 mL minꢀ1 kgꢀ1, t1/2 = 2 h) is attributed to the
presence of the polar hydroxyl group as well as equiva-
lent steric hindrance proximal to the metabolically labile
hydroxamic acid. It is also possible that intramolecular
hydrogen bonding between the hydroxyl group and
the hydroxamic acid influences absorption and clear-
ance. When dosed orally in hamsters, 26 exhibited
potent inhibition of IL-1 induced aggrecan degradation
(ED50 = 4.6 mg kgꢀ1) and MMP-13 induced collagen
degradation (ED50 = 13 mg/kg) in hamster knee joints.11
Several substituted heterocyclic analogues were pre-
pared, and some of these compounds provide good
MMP-13 and aggrecanase activity with reduced lipo-
philicity compared to the substituted phenyl analogues.
Biological data for these analogues are shown in Table
2. Consistent with SAR observed with simple substi-
tuted benzyl ethers, aggrecanase potency is strongly
influenced by the presence of a lipophilic substituent
ortho to the benzylic ether linkage (compare 32 with
28). Only the 4-isoquinolinyl derivative 30 shows
> 100· selectivity for aggrecanase inhibition over
MMP-1.
We chose compound 26 for further profiling based on its
excellent aggrecanase and MMP-13 potency, selectivity
Table 2. Aggrecanase and MMP-13 inhibitory activity of heteroarylmethyl ether analogues of 9 (R0 = (3R) Me)
Compound
R
Aggrecanase
IC50 (nM)
MMP-13
IC50 nM
MMP-1
IC50 (nM)
rTACE
IC50 (nM)
TNF HWB
IC50 (uM)
28
29
30
31
32
33
Pyridin-4-yl
Pyrazinyl
8.1%
8.0%
17
2.0
7.4
18
190
1600
7800
120
—
—
—
—
—
—
7.1
60
4-Isoquinolinyl
4-Quinolinyl
2.5
0.48
6.5
3.3
19
15
4.1
0.43
1.5
2-Chloro-4-pyridinyl
2-Methyl-3-pyridinyl
520
1600
51
Values expressed as a percentage represent percent inhibition of enzyme activity at 500 nM.

3388
M. C. Noe et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3385–3388
degradation was stimulated by the addition of IL-1a
(5 ng/ml). Ten hours later, aggrecan degradation was
quantified by scintillation counting of 35S in both the
conditioned medium and cell layers
We have discovered a novel series of MMP inhibitors
structured around 3-alkyl-3-hydroxy pipecolic
a
hydroxamate template. Many of these compounds pos-
sess potent inhibitory activity versus aggrecanase and
MMP-13 with selectivity versus MMP-1 and TACE
(as measured by TNF-a release in LPS-stimulated hu-
man blood). The pharmacokinetic properties and in vivo
activity observed with compound 26 suggest that it
could provide a useful starting point for the discovery
of agents for the treatment of osteoarthritis. Further
studies are in progress and will be reported in due
course.
7. Plugs (4 mm diameter and length) of bovine nasal cartilage
(BNC) were cultured in serum-free DMEM + IL-1a and +
26] Samples were taken at 48 h for analysis of GAG
content using the DMMB assay (Farndale, R., et al.,
Biochim. Biophys. Acta 1986, 883, 173). Medium was
replaced every 3–4 days with fresh IL-1 and inhibitor.
Assays were terminated after 2–3 weeks of culture and
collagen degradation was determined via hydroxyproline
analysis of hydrolyzed conditioned medium samples
(Otterness, I. et al., Arthritis Rheum. 1998, 41, 2068)
8. Human articular cartilage was obtained at total knee
replacement. Pieces (1 · 2 mm) were cultured in DMEM
containing 0.5% fetal bovine serum. Aggrecan degradation
was stimulated by the addition of IL-1a (50 ng/
ml) + oncostatin M (50 ng/ml). After 48 h, media samples
were assayed for GAG content using the DMMB assay
9. Male Sprague–Dawley rats were obtained with jugular
vein catheters from Charles River Labs, Wilmington, MA.
iv bolus doses were administered in glycerol formal (5 mg/
kg). Blood samples (600 mL) were collected from the
jugular vein (t = 5, 15, 30 min, 1, 2, 4, 6, 8, and 24 h).
Plasma was diluted with an equal volume of water and
extracted with MTBE. Drug levels were determined by
LC–MS/MS analysis. Pharmacokinetic parameters were
calculated using the non-compartmental method in Win-
Nonlin v2.1 (Pharsight, Mountain View, CA)
10. Predicted clearances were obtained from the disappear-
ance of drug in cryopreserved rat and dog hepatocytes
according to the half-life method of Shibata, Y.; Takah-
ashi, H.; Ishii, Y. Drug Metab. Dispos. 2000, 28, 1518.
11. Aggrecan degradation was initiated in hamster knees by
the intraarticular injection of 40 ng of murine IL-1a.
Animals were orally dosed with 26 4 h after IL-1. At 7 h
post IL-1, the animals were sacrificed and the GAG
content in lavaged synovial fluid was determined via the
DMMB assay. Collagen degradation was initiated by the
intraarticular injection of 2 lg of recombinant human
MMP-13. 26 was administered orally 1 h prior to MMP-
13 injection. Two hours after MMP-13 injection, the
animals were sacrificed and type II collagen fragments
were measured in the joint lavage using an ELISA specific
for collagenase cleaved type II collagen (Downs, J., et al.,
J. Immunol. Methods 2001, 247, 25).
References and notes
1. Hedbom, E.; Hauselmann, H. J. Cell. Mol. Life Sci. 2002,
59, 45.
2. Noe, M. C.; Natarajan, V.; Snow, S. L.; Mitchell, P. G.;
Lopresti-Morrow, L. L.; Reeves, L. M.; Yocum, S. A.;
Carty, T. J.; Barberia, J. T.; Sweeney, F. J.; Liras, J. L.;
Vaughn, M.; Hardink, J.; Hawkins, J. M.; Tokar, C.
Bioorg. Med. Chem. Lett. 2005, 15, 2808.
3. Letavic, M. A.; Axt, M. Z.; Barberia, J. T.; Carty, T. J.;
Danley, D. E.; Geoghegan, K. F.; Halim, N. S.; Hoth, L.
R.; Kamath, A. V.; Laird, E. R.; Lopresti-Morrow, L. L.;
McClure, K. F.; Mitchell, P. G.; Natarajan, V.; Noe, M.
C.; Pandit, J.; Reeves, L. M.; Schulte, G. K.; Snow, S. L.;
Sweeney, F. J.; Tan, D. H.; Yu, C. H. Bioorg. Med. Chem.
Lett. 2002, 12, 1387.
4. Aranapakam, V.; Davis, J. M.; Grosu, G. T.; Baker, J.;
Ellingboe, J.; Zask, A.; Levin, J. I.; Sananayaka, V. P.;
Du, M.; Skotnicki, J. S.; DiJoseph, J. F.; Sung, A.; Sharr,
M. A.; Killar, L. M.; Walter, T.; Jin, G.; Cowling, R.;
Tillett, J.; Zhao, W.; McDevitt, J.; Xu, Z. B. J. Med.
Chem. 2003, 46, 2376.
1
5. All new compounds exhibited satisfactory H NMR and
MS analysis
6. TACE, MMP, and whole blood TNF-a release assays
were conducted as described in Ref. 3. Aggrecanase
activity was assessed using a cell based assay: porcine
chondrocytes were plated in 48-well tissue culture plates.
Glycosaminoglycan (GAG) chains were labeled by includ-
ing [35S]sulfate (5 mCi/ml) in the culture medium. Unin-
corporated label was washed out and aggrecan