Fragment-based discovery of mexiletine derivatives as orally bioavailable inhibitors of urokinase-type plasminogen activator

Article abstract of DOI:10.1021/jm701359z

Fragment-based lead discovery has been applied to urokinase-type plasminogen activator (uPA). The (R)-enantiomer of the orally active drug mexiletine 5 (a fragment hit from X-ray crystallographic screening) was the chemical starting point. Structure-aided

Full text of DOI:10.1021/jm701359z

J. Med. Chem. 2008, 51, 183–186
Chart 2. Compounds 5-16a
183
Fragment-Based Discovery of Mexiletine
Derivatives as Orally Bioavailable Inhibitors
of Urokinase-Type Plasminogen Activator†
Martyn Frederickson,* Owen Callaghan, Gianni Chessari,
Miles Congreve, Suzanna R. Cowan, Julia E. Matthews,
Rachel McMenamin, Donna-Michelle Smith,
Mladen Vinkovic´, and Nicola G. Wallis
Astex Therapeutics Ltd, 436 Cambridge Science Park, Milton Road,
Cambridge, CB4 0QA, United Kingdom
ReceiVed October 31, 2007
Abstract: Fragment-based lead discovery has been applied to uroki-
nase-type plasminogen activator (uPA). The (R)-enantiomer of the orally
active drug mexiletine 5 (a fragment hit from X-ray crystallographic
screening) was the chemical starting point. Structure-aided design led
to elaborated inhibitors that retained the key interactions of (R)-5 while
gaining extra potency by simultaneously occupying neighboring regions
of the active site. Subsequent optimization led to 15, a potent, selective,
and orally bioavailable inhibitor of uPA.
Urokinase-type plasminogen activator1 (uPAa) is a trypsin-
like serine protease2 that, when bound to the extracellular
membrane-associated urokinase-type plasminogen activator re-
ceptor3 (uPAR), catalyzes the conversion of plasminogen to
plasmin through the cleavage of the amide bond of an arginine/
valine motif. Plasmin is responsible for a number of proteolytic
processes that degrade various components of the extracellular
matrix,4 thereby triggering the induction of cell migration. As
a result, uPA has been implicated in the progression of a variety
of disease states associated with abnormal tissue destruction and
cell infiltration including aortic aneurism,5 multiple sclerosis,6
and metastasis processes7,8 in cancer.
a Assay data quoted for compounds with IC50 < 5 µM are averages of
two or more duplicate measurements.
Chart 1. Advanced Inhibitors of uPA
of the key arginine of plasminogen. Unlike other trypsin-like
serine proteases that possess an adjacent hydrophobic alanine
residue (e.g., thrombin and Factor Xa), the main binding cleft
of uPA contains a more hydrophilic serine (Ser190) that
effectively precludes the binding of hydrophobic functionalities
that are commonly found in inhibitors of thrombin10 and Factor
Xa. Advanced inhibitors11 of uPA (in the clinic or in preclinical
studies) are thus dominated by classical arginine mimetics such
as guanidines and amidines (Chart 1); such inhibitors are usually
charged at physiological pH and accordingly have low oral
bioavailability. We were thus interested in the discovery of novel
inhibitors of uPA with less basic S1 moieties that might
overcome this problem.
Herein we disclose a novel series of weakly basic orally
bioavailable inhibitors of uPA derived from mexiletine 5 (Chart
2). We describe briefly the X-ray crystallographic cocomplex
of (R)-5 with uPA, a key fragment hit from an X-ray crystal-
lographic screen12 against the enzyme, and outline the structure-
aided development of the initial hit into more advanced,
selective, and orally bioavailable inhibitors, as exemplified by
compound 15.
A fragment hit from X-ray crystallographic screening
against uPA was the orally active drug mexiletine 5. Despite
having only relatively poor in vitro activity against uPA (IC50
> 1 mM), 5 had a very clear crystallographic binding mode
(Figure 1a); although, being a racemate, only the (R)-
The active site of uPA9 contains an aspartate residue (Asp189)
within the deep S1 pocket, which accommodates the side chain
* To whom correspondence should be addressed. Phone: +44 (0)1223
226239. Fax: +44 (0)1223 226201. E-mail: m.frederickson@
astex-therapeutics.com.
†
Coordinates and structure factors for cocomplexes of urokinase-type
plasminogen activator with compounds (R)-5, 6, 9, 11, 14, and 15 have
been deposited in the Protein Data Bank under accession codes 2VIN, 2VIO,
2VIP, 2VIQ, 2VIV, AND 2VIW, respectively.
aAbbreviations: uPA, urokinase-type plasminogen activator; uPAR,
urokinase-type plasminogen activator receptor; LE, ligand efficiency.
10.1021/jm701359z CCC: $40.75
2008 American Chemical Society
Published on Web 12/29/2007

184 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 2
Letters
Figure 1. X-ray crystallographic cocomplexes of uPA with selected inhibitors: (a) with (R)-5 (ligand electron density shown in light blue contoured
at 3σ); (b) with 6 (green) showing 9 (yellow) overlaid for comparison; (c) with 11 (green) showing 14 (yellow) overlaid for comparison; and (d)
with 15. Hydrogen bonds are shown as red dashed lines and water molecules are shown as red spheres. The acetate anion in (b), (c), and (d) has
been removed for clarity.
enantiomer of 5 was found to bind to the enzyme. The
primary amine moiety of (R)-5 binds to the carboxylate
functionality of the aspartate group (Asp189) and to the
backbone carbonyl of the serine residue (Ser190) that lie at
the base of the S1 pocket of the protein, with the pendant
ethanolamine derived spacer and aromatic ring making
substantial hydrophobic contacts with the residues that line
the pocket. The amino functionality is also involved in a
hydrogen bond with the oxygen of a neighboring glycine
residue (Gly219). We simultaneously observed the presence
of an acetate anion (from the crystallization buffer) in the
X-ray structure located in the oxyanion hole of the protein
(formed by the side chain of Ser195 and the three contiguous
backbone NHs of Gly193, Asp194, and Ser195) close to the
catalytic triad (Ser195, His57, and Asp102). Previous crystal-
lographic studies of uPA have similarly highlighted the
propensity of the sulfate anion13 to bind in this location.
Despite its low ligand efficiency14 (LE ∼ 0.30 kcal mol-1
heavy atom-1) when compared to simple benzamidines15 (LE
g 0.50), we chose to develop (R)-5 further as it is a known,
highly water soluble, and orally bioavailable drug [Foral (man)
)
90%]. Compound (R)-5 was also attractive to us due to its
relatively low pKa of 9.2 when compared to many uPA inhibitors
described previously, the basicity of the amine being lowered
significantly by the presence of the ꢀ-oxygen atom in the side
chain. The desire to reduce the pKa of the Asp189 engaging
functionality in uPA inhibitors to increase the likelihood of good
oral bioavailability has been voiced previously.13,15
The clear binding mode of (R)-5, in conjunction with
previously published structural data for a series of naphtha-

Letters
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 2 185
midines16 that simultaneously occupy neighboring binding sites,
strongly suggested that subsequent substitution at the 4-position
of the aromatic ring would lead to derivatives with much
improved potency. In initial experiments, solely for reasons of
immediate commercial accessibility, we initially replaced the
two ortho-methyl groups of 5 with chlorine atoms. We also
chose to remove the angular methyl group in the ethanolamine
side chain; as in the protein–ligand structure, this group was
proximal to the Ser190 side chain (resulting in steric crowding
in the S1 pocket) and so was likely to be detrimental to affinity.
We therefore prepared acid 6 together with a limited number
(guided by known structural data and by virtual screening) of
simple aromatic amides of which 7, 8, and 9 are representative
members.
uPA (Figure 1c) were highly complementary, we prepared the
conjoined derivative 15 (so incorporating the 2-EtO group) as
well as (R)-16, the corresponding specifically substituted angular
methyl derivative of 15 [analogous to (R)-5].
The 2-ethoxy derivative 15 (measured pKa of 8.75) is a potent
inhibitor of uPA (IC50 72 nM; LE ) 0.31) and is comparable
to 1-4 (Chart 1; reported Ki values of 410 nM,17 37 nM,18 12
nM,19 and 20 nM,7 respectively). As expected, (R)-16 (IC50 460
nM; LE ) 0.27) had markedly less in vitro activity than 15.
The maintenance of LE as the series progressed is in accord
with the ideas recently reported20 by Hajduk. The X-ray
crystallographic structure of the cocomplex of 15 with uPA
(Figure 1d) highlighted a binding mode analogous to those of
both 11 and 14.
The selectivity of 15 for uPA was determined by reference
to IC50 values (data not shown) measured against a panel of
closely related proteases (plasmin, Factor VIIa, Factor Xa,
tissue-type plasminogen activator, and trypsin). Amine 15 was
found to be selective for uPA (>50-fold) over all but one of
the enzymes but was only moderately so (3- to 4-fold) against
trypsin, the enzyme to which (in terms of the similarity of the
S1 subsite) it is most closely related and over which selectivity
for uPA is most difficult to obtain.21 That such limited selectivity
over trypsin was not an inherent property of all compounds akin
to 15 was shown by reference to compound 14, which showed
somewhat greater selectivity for uPA (around 10-fold).
On the basis of the observed high potency and selectivity,
the pharmacokinetic profile of compound 15 was studied in the
rat. When dosed orally at 4.2 mg/kg, 15 showed well defined
and reproducible PK data (Table 1); importantly, amine 15 had
As we progressed, we observed an increase in in vitro potency
(IC50: 5 > 1 mM; 6, 40 µM; 7, 35 µM; 8, 23 µM; 9, 5.2 µM)
and were able to rationalize this with reference to X-ray
crystallographic structures of cocomplexes of the compounds
with uPA (Figure 1b); in all cases, the amide (or acid) carbonyl
was shown to form a key hydrogen bond with the protein (to
the amidic side chain NH2 of Gln192). The amides also formed
a specific water-mediated hydrogen bond (between the amidic
NH and the backbone carbonyl of Ser214). Additionally, the
binding modes of the anilides (7, 8, and 9) showed the newly
appended aromatic ring to make substantial contacts with a
number of neighboring aromatic residues of the protein while
one of the pair of chlorine atoms was involved in a soft
interaction with a neighboring disulphide bridge (Cys191–Cys220)
that forms part of the S1ꢀ subpocket.9
Having rapidly demonstrated clear SAR around the aniline
ring of 7 and using precise structural data from a number of
uPA crystallographic cocomplexes, we returned our attention
to compounds that, like mexiletine, contained a pair of methyl
groups in place of chlorine atoms. We were keen to fully exploit
the SAR that we had observed, and aided by virtual screening,
we prepared a limited number of ortho- or meta-substituted
anilides, including 10, 11, and 12. Once again, potency against
uPA was shown to increase (IC50: 10, 2.7 µM; 11, 1.3 µM; 12,
720 nM), with the more highly substituted and lipophilic
derivatives being the most effective inhibitors.
high oral bioavailablity [Foral
) 60%]. The moderate
(rat)
clearance level, high volume of distribution, and long half-life
(7½ hours) observed for 15 are consistent with those observed
for mexiletine 5 in man.22 Compound 15 thus displays many
of the highly desirable molecular properties expected of a good
“lead-like” molecule, including high potency and selectivity over
a series of closely related proteases and, importantly, high oral
bioavailability, a molecular property that has so far remained
relatively elusive23 in advanced inhibitors of this heavily studied
therapeutic target.
An additional beneficial effect of the incorporation of an
amide moiety para- to the phenolic oxygen atom became
apparent when the basicities of several of the derivatives were
determined. Thus, 10 had a measured pKa of 8.75, around half
of a log unit lower than that of 5. This effect is presumed to be
due to the amide that, as a mesomeric electron withdrawing
group, subtly enhances the basicity modifying effect of the side
chain ꢀ-oxygen atom.
Also of particular interest was the effect of the 2-ethoxy group
of 11. The structure of the X-ray crystallographic cocomplex
formed between uPA and 11 showed the combined substituents
of the 2,5-disubstituted aniline to be readily accommodated by
the protein. Given the improved potency of 12, we also prepared
a series of 3,5-disubstituted derivatives, with the intention to
subsequently append a 2-EtO substituent onto the most promis-
ing scaffolds that we discovered.
Due to the limited commercial availability of 3,5-disubstituted
anilines, the derivatives that we invested time in preparing were
specifically chosen based upon the structural data that was
available to us and on the results of docking experiments; amides
13 and 14 are key indicative examples. Morpholine derivatives
such as 13 (IC50 1.6 µM) were somewhat less potent than the
analogous piperidines such as 14 (IC50 520 nM). As the
crystallographic structures of both 11 and 14 cocomplexed with
Table 1. Pharmacokinetic Profile for 15 (in Rat)
a
a
Cl (mL/min/kg)a
Vd (L/kg)
t
(hr)
F (%)b
½
43
27
7.5
60
a A 0.95 mg/kg i.v. dose (solution in 100% H2O). b A 4.2 mg/kg oral
dose (solution in 100% H2O).
Selective optimization of side activities of biologically active
compounds, the so-called SOSA approach,24 has gained favor
in recent years as a viable alternative to more traditional
screening methods. The idea that screening small collections
of structurally diverse drug molecules against new therapeutic
targets can aid the discovery of new drugs has appealed to many,
chiefly because the safety, pharmacology, and bioavailability
profile of any hit so obtained has implicitly already been well
studied in man. Mexiletine 5 is a Class Ib antiarrythmic licensed
for use in patients with life threatening irregular heartbeat and
may also be of use in the control of refractory pain. The findings
outlined herein clearly demonstrate that the functionalities
present in (R)-5 are also pertinent in the development of orally
bioavailable inhibitors of the clinically relevant serine protease
uPA.
Fragment-based screening experiments against uPA have been
reported previously, using both X-ray crystallography13 and
NMR techniques.15 Our studies have now shown that X-ray

186 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 2
Letters
crystallographic screening12 is an appropriate technique for the
discovery of novel, low basicity fragment hits against uPA that
can be successfully developed into more advanced lead-like
compounds (such as 15) with highly desirable physicochemical
properties, including high oral bioavailability.
at high resolution and reveals a novel structural subsite. J. Biol. Chem.
2000, 10, 7239–7248.
(10) Schweinhorst, A. Direct thrombin inhibitors-a survey of recent
developments. Cell. Mol. Life Sci. 2006, 63, 2773–2791.
(11) Abbenante, G.; Fairlie, D. P. Protease inhibitors in the clinic. Med.
Chem. 2005, 1, 71–104.
(12) Hartshorn, M. J.; Murray, C. W.; Cleasby, A.; Frederickson, M.; Tickle,
I. J.; Jhoti, H. Fragment-based lead discovery using X-ray crystal-
lography. J. Med. Chem. 2005, 48, 403–413.
(13) Nienaber, V. L.; Richardson, P. L.; Klighofer, V.; Bouska, J.; Giranda,
V. L.; Greer, J. Discovering novel ligands for macromolecules using
X-ray crystallographic screening. Nat. Biotechnol. 2000, 18, 1105–
1108.
(14) Hopkins, A. L.; Groom, C. R.; Alex, A. Ligand efficiency: A useful
metric for lead selection. Drug DiscoVery Today 2004, 9, 430–431.
(15) Hajduk, P. J.; Boyd, S.; Nettesheim, D.; Nienaber, V.; Severin, J.;
Smith, R.; Davidson, D.; Rockway, T.; Fesik, S. W. Identification of
novel inhibitors of urokinase via NMR-based screening. J. Med. Chem.
2000, 43, 3862–3866.
(16) Wendt, M. D.; Rockway, T. W.; Geyer, A.; McClellan, W.; Weitzberg,
M.; Zhao, X.; Mantei, R.; Nienaber, V. L.; Stewart, K.; Klinghofer,
V.; Giranda, V. Identification of novel binding interactions in the
development of potent inhibitors of urokinase. Synthesis, structural
analysis, and SAR of N-phenyl amide 6-substitution. J. Med. Chem.
2004, 47, 303–324.
Acknowledgment. The authors thank Joe Coyle for as-
sistance with protein production, Jacqueline A. Higgins, E. Jon
Lewis, and Douglas Ross for DMPK support and Robin A. E.
Carr, David M. Cross, Christopher W. Murray, and Glyn
Williams for useful discussions. pKa values were determined
by Pharmorphix Ltd, 250 Cambridge Science Park, Milton Road,
Cambridge, CB4 0WE, U.K.
Supporting Information Available: Experimental procedures
for the synthesis of compounds 6-16, in vitro assay protocols for
determination of IC50 values against uPA, and X-ray crystal-
lographic details for uPA cocomplexed with (R)-5, 6, 9, 11, 14,
and 15. This material is available free of charge via the Internet at
http://pubs.acs.org.
(17) Stürzebecher, J.; Vieweg, H.; Steinmetzer, T.; Schweinitz, A.; Stubbs,
M. T.; Renatus, M.; Wikström, P. 3-Amidinophenylalanine-based
inhibitors of urokinase. Bioorg. Med. Chem. Lett. 1999, 9, 3147–3152.
(18) Barber, C. G.; Dickinson, R. P.; Fish, P. V. Selective urokinase-type
plasminogen activator (uPA) inhibitors. Part 3: 1-Isoquinolinylguanidines.
Bioorg. Med. Chem. Lett. 2004, 14, 3227–3230.
(19) Fish, P. V.; Barber, C. G.; Brown, D. G.; Butt, R.; Collis, M. G.;
Dickinson, R. P.; Henry, B. T.; Horne, V. A.; Huggins, J. P.; King,
E.; O’Gara, M.; McCleverty, D.; McIntosh, F.; Phillips, C.; Webster,
R. Selective urokinase-type plasminogen activator inhibitors. 4. 1-(7-
Sulfonamidoisoquinolinyl)guanidines. J. Med. Chem. 2007, 50, 2341–
2351.
(20) Hajduk, P. Fragment-based drug design: How big is too big? J. Med.
Chem. 2006, 49, 6972–6976.
(21) Mackman, R. L.; Katz, B. A.; Breitenbucher, J. G.; Hui, H. C.; Verner,
E.; Luong, C.; Liu, L.; Sprengeler, P. A. Exploiting subsite S1 of
trypsin-like serine proteases for selectivity: Potent and selective
inhibitors of urokinase-type plasminogen activator. J. Med. Chem.
2001, 44, 3856–3871.
References
(1) Duffy, M. J. The urokinase plasminogen activator system: Role in
malignancy. Curr. Pharm. Des. 2004, 10, 39–49.
(2) Ilies, M. A.; Scozzafava, A.; Supuran, C. T. Therapeutic applications
of serine protease inhibitors. Expert Opin. Ther. Pat. 2002, 12, 1181–
1214.
(3) Ge, Y.; Elghetany, T. Urokinase plasminogen activator receptor
(CD87): Something old, something new. Lab. Hematol. 2003, 9, 67–
71.
(4) Wong, A. P.; Cortez, S. L.; Baricos, W. H. Role of plasmin and
gelatinase in extracellular matrix degradation by cultured rat mesangial
cells. Am. J. Physiol. Renal Physiol. 1992, 263, 1112–1118.
(5) Jean-Claude, J.; Newman, K. M.; Li, H.; Gregory, A. K.; Tilson, M. D.
Possible key role for plasmin in the pathogenesis of abdominal aortic
aneurysms. Surgery 1994, 116, 472–478.
(6) Gveric, D.; Hanemaaijer, R.; Newcombe, J.; van Lent, N. A.; Sier,
C. F. M.; Cuzner, M. L. Plasminogen activators in multiple sclerosis
legions. Brain 2001, 124, 1978–1988.
(7) Schweinitz, A.; Steinmetzer, T.; Banke, I. J.; Arlt, M. J. E.;
Stürzebecher, A.; Schuster, O.; Geissler, A.; Giersiefen, H.; Zeslawska,
E.; Jacob, U.; Krüger, A.; Stürzebecher, J. Design of novel and
selective inhibitors of urokinase-type plasminogen activator with
improved pharmacokinetic properties for use as antimetastatic agents.
J. Biol. Chem. 2004, 279, 33613–33622.
(8) Almholt, K.; Lund, L. R.; Rygaard, J.; Nielsen, B. S.; Danø, K.; Rømer,
J.; Johnsen, M. Reduced metastasis of transgenic mammary cancer in
urokinase-deficient mice. Int. J. Cancer 2005, 113, 522–532.
(9) Nienaber, V. L.; Wang, J.; Davidson, D.; Henkin, J. Re-engineering
of human urokinase provides a system for structure-based drug design
(22) Monk, J. P.; Brogden, R. N. Mexiletine. A review of its pharmaco-
dynamic and pharmacokinetic properties, and therapeutic use in the
treatment of arrhythmias. Drugs 1990, 40, 374–411.
(23) Bruncko, M.; McClellan, W. J.; Wendt, M. D.; Sauer, D. R.; Geyer,
A.; Dalton, C. R.; Kaminski, M. A.; Weitzberg, M.; Gong, J.; Dellaria,
J. F.; Mantei, R.; Zhao, X.; Neinaber, V. L.; Stewart, K.; Klinghofer,
V.; Bouska, J.; Rockway, T. W.; Giranda, V. L. Naphthamidine
urokinase plasminogen activator inhibitors with improved pharmaco-
kinetic properties. Bioorg. Med. Chem. Lett. 2005, 15, 93–98.
(24) Wermuth, C. G. Selective optimization of side activites: Another way
for drug discovery. J. Med. Chem. 2004, 47, 1303–1314.
JM701359Z