Novel 5,5-disubstitutedpyrimidine-2,4,6-triones as selective MMP inhibitors

Article abstract of DOI:10.1016/S0960-894X(01)00104-4

The 5,5-disubstitutedpyrimidine-2,4,6-triones represent a new class of MMP inhibitors showing selectivity for the gelatinases A and B, collagenase-3, and human neutrophil collagenase. The SAR presented here is in good agreement with an X-ray structure of compound 5 bound to the catalytic domain of stromelysin-1. While of the barbiturate structural class, compound 5 did not show any toxic or sedative effects.

Full text of DOI:10.1016/S0960-894X(01)00104-4

Bioorganic & Medicinal Chemistry Letters 11 (2001) 969–972
Novel 5,5-Disubstitutedpyrimidine-2,4,6-triones as Selective
MMP Inhibitors
Louise H. Foley,a,* Robert Palermoa,y Pete Duntena and Ping Wanga
aRoche Research Center, Hoffmann-La Roche Inc., 340 Kingsland St., Nutley, NJ07110, USA
Received 18 December 2000;accepted 2 February 2001
Abstract—The 5,5-disubstitutedpyrimidine-2,4,6-triones represent a new class of MMP inhibitors showing selectivity for the gela-
tinases A and B, collagenase-3, and human neutrophil collagenase. The SAR presented here is in good agreement with an X-ray
structure of compound 5 bound to the catalytic domain of stromelysin-1. While of the barbiturate structural class, compound 5 did
not show any toxic or sedative effects. # 2001 Elsevier Science Ltd. All rights reserved.
Matrix metalloproteinases (MMPs) are a family of zinc
dependent endopeptidases involved in the breakdown of
components of the extracellular matrix which facilitates
connective tissue remodeling. This process is important
in embryonic development, pregnancy, growth, and
wound healing. Normally the activity of MMPs is con-
trolled by a tight balance between synthesis of active
MMPs and the presence of endogenous inhibitors, such
as the TIMPs (tissue inhibitor of metalloproteinases). In
a number of disease states, including cancer, this bal-
ance is lost. Overexpressed MMPs are important in at
least three events in malignancy progression: (1) pri-
mary tumor growth, which requires the breakdown of
the surrounding connective tissue;(2) metastasis, invol-
ving the movement of tumor cells across the vascular
basement membrane;and (3) angiogenesis, the growth
and entry of new blood vessels into the growing tumor.
agents, we reinvestigated a compound that had been
shown to be a weak inhibitor of stromelysin in a high
throughput screen. Against stromelysin, compound 1
showed an IC50 of 30 mM and thus was of little interest
in an earlier arthritis project. Interestingly, when com-
pound 1 was assayed against the oncology targets of
gelatinases A and B it was found to be at least 30-fold
more active against these enzymes than against strome-
lysin-1 (IC50 of 1.3 mM and 0.33 mM, respectively,
against gelatinases A and B). This trend of selectivity
for gelatinases over stromelysin-1 was observed with all
tested compounds of the pyrimidine-2,4,6-trione class.
Chemistry
Commercial methyl biphenylacetate and methyl 4-phe-
noxyphenylacetate were converted into the correspond-
ing malonate derivatives by reaction with NaH and
dimethylcarbonate in THF at reflux. The malonate
anion, prepared with NaH in THF, was then alkylated
with the required alkyl halide. Conversion to the 5,5-
The gelatinases A and B (MMPs-2 and -9) appear to be
the MMPs most highly expressed in and around the
widest variety of tumors and have been implicated in
tumor aggressiveness and increased metastatic poten-
tial.1 To date most MMP inhibitors have been strong
zinc binding agents and thus broad spectrum MMP
inhibitors.2 The lack of selectivity many of these com-
pounds exhibit may contribute to the unacceptable
toxicities shown in human trials.
disubstituted-pyrimidine-2,4,6-triones
was
accom-
plished by reaction of the malonate derivatives with
Mg(OCH3)2 and urea (ratio 1:2.8:2) in a minimum
amount of methanol at reflux. The thiocarbonyl deriva-
tive 10 was prepared as above by substituting thiourea
for urea. The N-alkylated derivative 11 was prepared by
reaction of 5 with NaH and methyl iodide in THF and
chromatographic separation of the mono- and dialkyl-
ated products. Reaction of the appropriate mono-
substituted malonate with urea and NaOCH3 (ratio
1:3:2) in methanol at reflux, gave the 5-monosubstituted
pyrimidine trione 12.
During a program to identify new, non-hydroxamate
inhibitors of gelatinases A and B for use as antitumor
*Corresponding author. Fax: +1-973-235-7122;e-mail: louise.foley@
roche.com
yCurrent address: Schering-Plough Research Institute, 2015 Galloping
Hill Road, Kenilworth, NJ 07033, USA
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00104-4

970
L. H. Foley et al. / Bioorg. Med. Chem. Lett. 11 (2001) 969–972
Table 1.
Assay
Enzyme assays were carried out in a microtiter-format
using the fluorogenic substrate described by Bickett et
al.3 Full length forms of gelatinases A4 and B5 and
stromelysin-16 were prepared as described. Assays with
gelatinases were run at pH 7.5, 50 mM TrisHCl, 200
mM NaCl, 5 mM CaCl2, 20 mM ZnCl2, and 0.05% (wt/
v) Brij-35 with a final substrate concentration of 10 mM
and concentration for gelatinase A 0.08 nM;gelatinase
B 0.05 nM. For stromelysin-1 the assay buffer was pH
6.5, 50 mM Mes, 2.5 mM EDTA, 5 mM CaCl2, and
0.05% (wt/v) Brij-35. Substrate concentration was 10
mM, and enzyme concentration was 10 nM.
IC50 (mM)
Compd
R1
R2
Stromelysin Gelatinase A Gelatinase B
1
Hexyl
30
1.3
0.33
Early modeling work with the pyrimidine-2,4,6-trione 1
suggested that the phenyl group at C-5 would fit into
the S10 pocket and that since this pocket is a very deep
cleft, replacing the phenyl with a longer aromatic unit
should improve potency. However, compound 3, with
the longer biphenyl moiety showed only modest
improvement in gelatinase A inhibitory activity and had
a decreased potency against gelatinase B (compare
compounds 1 and 3 in Table 1). Modeling of 3 sug-
gested that the biphenyl was too rigid and that the sec-
ond phenyl group would be forced into the S10 pocket
wall.
2
3
Methyl
Hexyl
NDa
50
11
ND
0.868
0.898
0.081
0.021
0.863
0.696
0.052
0.018
4
5
6
Ethyl
Methyl
Hexyl
9
To improve the flexibility of the C-5 aromatic unit the 4-
phenoxyphenyl derivative 5 was prepared and, in sup-
port of modeling, showed a >10-fold improvement in
potency. An X-ray structure of 5 complexed with the
catalytic domain of stromelysin-17 was obtained and
the interactions with the protein are shown in Figures 1
and 2.
2
The effect of the size of the alkyl group at C-5 was
explored in the phenyl, biphenyl, and 4-phenoxyphenyl
2
7
8
0.93
ND
0.019
0.059
0.017
0.118
9
Methyl
1
0.53
0.164
Figure 1. X-ray structure of compound 5 (blue) in the active site of
stromelysin-1.
aND=Not determined.

L. H. Foley et al. / Bioorg. Med. Chem. Lett. 11 (2001) 969–972
971
the thiol tautomer8 and further, should Glu202 be pro-
tonated, it is also reported that the thiocarbonyl of
thioamides is a poor hydrogen-bond acceptor.9 The
importance of the amide hydrogens at N-1 and N-3 was
shown by the loss of activity exhibited by N-methyl
derivative 11. The dialkylated derivative was also
prepared and was similarly inactive.
Figure 2. Hydrogen bonds and zinc coordination between protein and
compound 5.
The requirement for the tetrahedral carbon at C-5 was
shown by preparation of compound 12. Replacing the
alkyl group at C-5 with a hydrogen, as shown in 12,
allowed enolization10 and places C-4, C-5, and C-6 of
the pyrimidine trione ring and the phenyl moiety in the
same plane, thus the phenoxyphenyl unit is unable to
enter the S10 pocket and results in a considerable loss of
activity. Each of these results also agrees with the X-ray
structure of a different class of pyrimidine triones with
human neutrophil collagenase (MMP-8).11
classes. In contrast to the phenyl series, where potency
was lost with the shorter alkyl substituent at C-5 (com-
pound 1 vs 2, Table 1), in the biphenyl series shortening
of the C-5 hexyl to an ethyl group did not change
potency (compare compounds 3 and 4 in Table 1).
However in the phenoxyphenyl series a moderate
improvement in activity was observed when the C-5
methyl in 5 was replaced by hexyl (compound 6, Table 1).
These observations suggest that there is some difference
in how the alkyl group at C-5 interacts with the protein
in the three different compound classes and further
suggests that in the phenyl series it is the hexyl group of
1 which is occupying the S10 pocket.
None of the pyrimidine triones showed activity against
matrilysin (MMP-7) (IC50 >50 mM) and were only
weakly active against collagenase-1 (MMP-1) (IC50 >2
mM) and stomelysin-1 (MMP-3) (see Table 1). In addi-
tion to inhibiting gelatinases A and B (MMP-2 and -9),
compounds 5 and 6 were assayed against human neu-
trophil collagenase (MMP-8) (IC50 for compounds 5
and 6, 0.081 and 0.13 mM, respectively), and col-
lagenase-3 (MMP-13) (IC50 for 5 and 6, 0.065 and 0.078
mM, respectively).
Replacing the alkyl at C-5 with the methyl benzyl ether
group, as shown in compound 7, resulted in similar
potency to the hexyl derivative 6. However, replacing
the alkyl with the more hydrophilic ethanol moiety in
compound 8 resulted in some loss in activity toward
gelatinase A and a 7-fold loss against gelatinase B. The
effect of additional flexibility and the requirement for
the second aryl group at C-5 was explored by replace-
ment of the phenoxyphenyl group in 5 with the benzyl
alkoxy unit shown in 9. This modification resulted in
some loss in inhibitory activity against both gelatinase
A and gelatinase B (>6-fold loss against gelatinase A
and about a 3-fold loss against gelatinase B).
The pyrimidine triones belong to the barbiturate com-
pound class and therefore might be expected to exhibit
barbiturate-like activities. This was explored with a tol-
erability study using compound 5 at an ascending single
oral dose (10, 30, 100, and 1000 mg/kg) and an ascend-
ing 5 day oral dose (100, 300, and 1000 mg/kg/day) in
CD-1 mice. In neither study was there any indication of
toxicity or sedative effects, however some increase in the
hours the mice were awake and feeding was observed in
the treated animals (C. Slater and L. Hall, Hoffmann-La
Roche, personal communication).
Confirmation of the interactions between 5 and the zinc
atom and Glu202 shown in the X-ray structure (Figs. 1
and 2), was obtained by replacing the C-2 carbonyl with
a thiocarbonyl, compound 10. This modification caused
a >13-fold loss in potency against both gelatinases.
This result supports structural information suggesting
that enolization of the N-3, C-2 amide unit of the pyr-
imidine trione provides N-3 as the fourth zinc ligand
and that the resultant hydroxyl at C-2 provides the
hydrogen for a bidentate hydrogen bond to Glu202. In
thioamides it is reported that the thione is favored over
In summary, the 5,5-disubstitutedpyrimidine-2,4,6-
triones have been shown to be a novel and nontoxic
class of MMP inhibitors. The selectivity shown for
MMPs-2, -9, -8, and -13 make them very attractive as
antitumor agents. The SAR presented here is in excel-
lent agreement with the information available from
the X-ray structures of pyrimidine-2,4,6-triones with
stromelysin-17 and human neutrophil collagenase.11

972
L. H. Foley et al. / Bioorg. Med. Chem. Lett. 11 (2001) 969–972
However, analysis of these X-ray structures and a pub-
lished X-ray structure of gelatinase A12 does not pro-
vide a rationale for the selectivity shown by the
pyrimidine triones toward the gelatinases over strome-
lysin-1. The explanation for this will require a co-crystal
structure of gelatinase A or B with a pyrimidine trione.
(c) Itoh, T.;Tanioka, M.;Yoshida, H.;Yoshioka, T.;Nishi-
moto, H.;Itohara, S. Cancer Res. 1998, 58, 1048. (d) Liotta,
L. A.;Tryggvason, K.;Garbisa, S.;Hart, I.;Foltz, C. M.;
Shafie, S. Nature 1980, 284, 67.
2. Skiles, J. W.;Monovish, L. G.;Jeng, A. Y. Annu. Rep.
Med. Chem. 2000, 35, 167.
3. Bickett, D. M.;Green, M. D.;Berman, J.;Dezube, M.;
Howe, A. S.;Roth, P. J.;McGeehan, G. M. Anal. Biochem.
1993, 212, 58.
4. Sang, Q. A.;Bodden, M. K.;Windsor, L. J.
Chem. 1996, 15, 243.
J. Protein
Acknowledgements
5. Murphy, G.;Willenbrock, F.;Ward, R. V.;Cockett, M. I.;
Eaton, D.;Docherty, A. J. P. Biochem. J. 1992, 283, 637.
6. Marcy, A. I.;Eiberger, L. L.;Harrison, R.;Chan, H. K.;
Hutchinson, N. I.;Hagmann, W. K.;Cameron, P. M.;Boul-
ton, D. A.;Hermes, J. D. Biochemistry 1991, 30, 6476.
7. Dunten, P.;Kammlott, U.;Crowther, R.;Levin, W.;
The authors thank Vincent Madison for assistance in
modeling, Kwei Lan Tsao and Alejandro Lugo for
running the IC50 assays, and Stephen Ferguson for
assistance in the early stages of this program. Following
the merger of Hoffmann-La Roche with Boehringer
Mannheim we became aware that scientists at Boehrin-
ger Mannheim had also found the pyrimidine trione
class to be excellent gelatinase inhibitors. We thank
Hans-Willi Krell and Frank Grams for keeping us
informed of their work and publication plans.
Foley, L. H.;Wang, P.;Palermo, R.
press.
Protein Sci. 2001, in
8. Kjellin, B.;Sandstro m, J. Acta Chem. Scand. 1973, 27, 209.
¨
9. Artis, D. R.;Lipton, M. A. J. Am. Chem. Soc. 1998, 120,
12200.
10. A H NMR suggests 12 is a 3:2 mixture of keto and enol
tautomers in DMSO-d6.
1
11. Brandstetter, H.;Grams, F.;Glitz, D.;Lang, A.;Huber,
R.;Bode, W.;Krell, H.-W.;Engh, R. A. J. Biol. Chem. 2001,
in press.
References and Notes
12. Dhanaraj, V.;Williams, M. G.;Ye, Q.-Z.;Molina, F.;
Johnson, L. L.;Ortwine, D. F.;Pavlovsky, A.;Rubin, J. R.;
Skeean, R. W.;White, A. D.;Humblet, C.;Hupe, D. J.;
Blundell, T. L. Croatica Chem. Acta 1999, 72, 575.
1. (a) Garbett, E. A.;Reed, M. W. R.;Brown, N. J. Br. J.
Cancer 1999, 81, 287. (b) Stetler-Stevenson, W. G.;Azna-
voorian, S.;Liotta, L. A. Annu. Rev. Cell. Biol. 1993, 9, 541.