DNA binding ligands with improved in vitro and in vivo potency against drug-resistant Staphylococcus aureus

Article abstract of DOI:10.1021/jm049712g

Potent in vivo activity against methicillin-resistant Staphylococcus aureus (MRSA) has been difficult to achieve with previously reported DNA binding antibacterials. Herein, we describe an efficient access to a focused library of new analogues yielding co

Full text of DOI:10.1021/jm049712g

4352
J . Med. Chem. 2004, 47, 4352-4355
Letters
DNA Bin d in g Liga n d s w ith Im p r oved in
Vitr o a n d in Vivo P oten cy a ga in st
Dr u g-Resista n t Sta ph ylococcu s a u r eu s
Wenhao Hu,†,# Roland W. Bu¨rli,†,#
J acob A. Kaizerman,† Kirk W. J ohnson,†
Matthew I. Gross,† Mari Iwamoto,† Peter J ones,†
Denene Lofland,† Stacey Difuntorum,† Hsiu Chen,†
Bu¨lent Bozdogan,§ Peter C. Appelbaum,§ and
Heinz E. Moser*,†
Genesoft Pharmaceuticals, Inc., 7300 Shoreline Court,
South San Francisco, California 94080, and Department of
Physiology, Penn State Milton S. Hershey Medical Center,
500 University Drive, Hershey, Pennsylvania 17033
Received April 14, 2004
Abstr a ct: Potent in vivo activity against methicillin-resistant
Staphylococcus aureus (MRSA) has been difficult to achieve
with previously reported DNA binding antibacterials. Herein,
we describe an efficient access to a focused library of new
analogues yielding compounds with improved activity in a
mouse peritonitis model. The most potent molecules (14 and
19) exhibit efficacy against MRSA at ED50 values of ∼1 and
∼5 mg/kg, respectively, and display excellent in vitro activity
against vancomycin-resistant S. aureus.
F igu r e 1. Structures of distamycin A and lead molecules.
Figure 1. This compound was found to be well tolerated
in rodents at doses that might be required for efficacy.
In a first proof-of-concept study, 1 given intravenously
rescued mice from a lethal infection of methicillin-
susceptible S. aureus (MSSA) with an ED50 of ∼30 mg/
kg. However, at a dose of 50 mg/kg, it was not efficacious
against MRSA in the same model. Further optimization
of 1 has led to an advanced lead molecule (2): the
isoquinoline 2 showed an ED50 of ∼7 mg/kg against
MSSA and 11 mg/kg against MRSA.9,10 In this optimi-
zation process, we have found that structural modifica-
tions of the internal N-methylpyrrole-2-carboxamide
units (Py) are critical for in vivo efficacy against drug-
resistant S. aureus. Interestingly, the in vitro activity
in the presence of human plasma albumin (HPA) has
correlated relatively well with in vivo efficacy following
intravenous administration and was used as a surrogate
for the focused lead optimization presented herein. We
now report our efforts to further optimize lead 2 for
improved in vivo efficacy against MRSA.
Infections caused by drug-resistant bacteria represent
an increasing threat to society and pose a real challenge
to physicians for selecting the correct antibiotic from the
arsenal of existing therapeutics.1 In addition to methi-
cillin-resistant Staphylococcus aureus (MRSA) and peni-
cillin-resistant Streptococcus pneumoniae (PRSP), two
strains of S. aureus with acquired resistance to vanco-
mycin have recently been isolated.2 The increasing
emergence of resistance to virtually all existing antibi-
otics poses a threat to health care, and novel therapeu-
tics are desperately needed. However, the development
of new antibacterial drugs based on novel targets and
consequently the lack of cross-resistance have been a
major challenge for drug discovery.3 This is underscored
by the fact that despite major efforts within the past
35 years only two novel classes of antibiotics, the
oxazolidinones (linezolid) and lipopeptides (daptomycin),
have reached the market.4,5
The tetrameric isoquinolines 3-5 (Figure 1) were
prepared from the corresponding isoquinoline, imida-
zole, and pyrrole building blocks by standard amide
bond formations; analogous chemistry has been de-
scribed in detail.8 For the preparation of a small library
of analogues of the tetrameric isoquinolines 2 and 5 with
varying C-terminal amino groups, we developed a one-
pot procedure that allows a rapid conversion of the
dimethoxyacetals 10 and 11 to the final compounds:
reductive amination of the acetal in the presence of
solid-phase bound cyanoborohydride and the corre-
sponding amine at ca. 150 °C (microwave) gave the final
compounds (Scheme 1). All materials were purified by
preparative HPLC and characterized by 1H NMR, mass
We have been optimizing the antibacterial potency
and drug-related properties of analogues of distamycin
A, a small DNA minor-groove binding natural product.6
In comparison to distamycin, the first generation of
compounds exhibited a significantly improved potency
against Gram-positive bacteria.7 More recently, we
reported an optimization process in which we identified
compounds bearing novel C- and N-termini.8 A proto-
typic structure, the 3-chlorothiophene 1, is shown in
* To whom correspondence should be addressed. Phone: +1-650-
837-1820. Fax +1-650-827-0476. E-mail: hmoser@genesoft.com.
† Genesoft Pharmaceuticals
#
Authors contributed equally to this work.
§ Penn State Milton S. Hershey Medical Center.
10.1021/jm049712g CCC: $27.50 © 2004 American Chemical Society
Published on Web 07/29/2004

Letters
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 18 4353
Ta ble 1. Antibacterial Activity of Compounds 2-5a
Sch em e 1. Synthesis of 2, 5, and 15-22 (Table 2)a
MRSA
27660
MRSA
MSSA
13709
VRF
51559
PRSP
51422
X
Y
+HPAb
2
3
4
5
CH
N
N
CH
N
CH
N
0.06
0.06
16
1
1
32
0.13
0.06
0.13
2
1
0.13
0.5
0.06
0.25
0.5
CH
0.008
0.008
0.008
0.03
a
MIC values in µg/mL against ATCC strains. MRSA, MSSA:
methicillin-resistant, -susceptible S. aureus. VRF: vancomycin-
resistant E. faecium. PRSP: penicillin-resistant S. pneumoniae.
b
HPA: 2% of human plasma albumin was added to the test
medium. MRSA is ATCC strain 27660.
Ta ble 2. Antibacterial Activity of All NPy-pBz-PyC and
NPy-pBz-ImC Compounds
a
Reagents: (i) for 8: 6, THF, DIEA, 2,2-dimethoxyethylamine,
30 min, 25 °C, then H2, Pd, 25 °C, 2 h, 85%; for 9: 7, 2,2-
dimethoxyethylamine, 20 min, 110 °C, then H2, Pd, 25 °C, 2 h,
68%; (ii) R1OH, BopCl, NMP, DIEA, DMAP, amine, 30 min, 60
°C, 10 (64% from 8), 11 (75% from 9); (iii) 10 or 11, amine, MP-
NEt3+BCNH3-, Cl2CHCOOH, THF, 10 min, 150 °C, then prepara-
tive HPLC.
Sch em e 2. Synthesis of 14a
a
Reagents: (i) 12, 13, HATU, NMP, DIEA, 30 min, 60 °C, 51%.
spectrometry, and analytical HPLC as documented
previously.8 Experimental details and analytical data
are available in the Supporting Information. The purity
of the compounds was generally greater than 95%
(exceptions are specified in Table 2). The reductive
amination protocol allowed a very efficient access to a
small quantity of compound. At larger scales (>50 mg),
compounds were prepared by classical solution-phase
chemistry from their corresponding building blocks.
Similarly, 14 (Scheme 2) and the intermediates 10 and
11 were prepared from the known nitropyrrole 611 and
imidazole 7,12 respectively.
The lead optimization, aimed at improvement of
potency of 2, proceeded stepwise. We started by sys-
tematically replacing the internal Py units in 2 by an
N-methylimidazole-2-carboxamide moiety (Im). We specu-
lated that replacement of Py by an Im might improve
the physicochemical properties of the molecules. The Im
building block has been studied extensively in polya-
mides, i.e., larger homologues of distamycin A with
improved DNA binding affinity and sequence specific-
ity;13 it has been shown that in a 2:1 binding motif, an
Im opposite a Py specifically recognizes a G‚C base pair,
whereas in a 1:1 binding mode, it still interacts with
A/T rich sequences but lacks the specificity of Py.14,15
We prepared all three Im analogues of the reference 2
and compared their in vitro potency against drug-
a
b
See footnotes to Table 1 for details. 2% of human plasma
albumin was added to the test medium. c Not determined. MIC
in the presence of 4% HPA. e 94% purity as determined by HPLC-
UV.
d
resistant and -susceptible S. aureus, E. faecalis, E.
faecium, and S. pneumoniae (Table 1 and Supporting
Information).16 Interestingly, the bis-imidazole 3 showed
similar potency compared to the bis-pyrrole 2, whereas
the activity of the Im-Py hybrids 4 and 5 significantly
varied depending on the position of the Im; replacement
of the C-terminal Py by Im resulted in excellent potency
against all Gram-positive strains (MIC ) 0.004-0.03
µg/mL for 5), while its isomer 4 lost potency, in
particular against MRSA (16 µg/mL). More importantly,
5 also exhibited an excellent MIC of 0.13 µg/mL against
MRSA in the presence of HPA.
At this point, we studied the influence of the C-
terminal amino group on in vitro potency. For this, we
prepared a library of analogues (∼100 compounds)
consisting of the NPy-pBz-PyC and the NPy-pBz-ImC core
scaffolds (pBz: 4-aminobenzoate). The in vitro suscep-
tibilities of representative compounds are summarized
in Table 2. Clearly, the NPy-pBz-ImC motif tolerated
more structural diversity at the C-terminus than the
NPy-pBz-PyC system. For instance, the morpholine 2 is

4354 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 18
Letters
(40% survival). Clearly, in vitro activity is not the only
parameter influencing in vivo efficacy, as evidenced by
the identical MIC values of 19 and 20 against S. aureus,
and pharmacokinetics and ADME properties will impact
efficacy as well. The bis(2-methoxyethyl)amine 21 did
not show efficacy at 25 mg/kg.
Compound 22 was not pursued further because the
animals showed marked lethargic behavior in an acute
tolerability test at 50 mg/kg (iv). Instead, the 4-hydroxy-
piperidine 19 was selected for a dose-response study.
It showed complete protection at 10 mg/kg, and two out
of five animals survived when administered at 3 mg/
kg. The compound was inactive at a dose of 1 mg/kg.
From this study, we concluded that 19 has an ap-
proximate ED50 of 5 mg/kg against MRSA, which is
significantly better than the parent compound 2. The
morpholine analogue 5 was directly tested at 10 mg/
kg, since the excellent in vitro profile in the presence
and absence of HPA suggested good efficacy; however,
at 10 mg/kg, only one out of five animals survived.
Last, we investigated the in vivo efficacy of the
benzothiophene 14. Because this compound demon-
strated a great in vitro profile, we tested only low doses
in vivo. Indeed, it resulted in 80% survival at 3 mg/kg
and still 60% protection at 1 mg/kg. Such performance
in this stringent model of MRSA infection is rare, and
the only reference therapeutic we have observed with
comparable efficacy to 14 in this model is vancomycin
(ED50 ≈ 1 mg/kg).
F igu r e 2. VRSA killing kinetics of 14 at various concentra-
tions.
Ta ble 3. Animal Efficacy (Mouse MRSA Peritonitis)a
survival
25 mg/kg
survival
10 mg/kg
survival
3 mg/kg
survival
1 mg/kg
2
5
5/5
ndb
ndb
5/5
2/5
0/5
5/5
3/5
1/5
ndb
5/5
ndb
ndb
ndb
0/5
ndb
4/5
2/5
ndb
ndb
ndb
ndb
ndb
3/5
14
19
20
21
22
0/5
ndb
ndb
ndb
a
The mouse peritionitis model was performed as described
previously.8 Mice were infected intraperitoneally with a 5-10 ×
LD50 inoculum with MRSA (ATCC 27660), treated intravenously
with vehicle, positive control (vancomycin), or test articles with a
single dose 1 h after infection, and monitored for 5 subsequent
b
Earlier lead compounds such as the thiophene 1
protected mice in a peritonitis model against a lethal
infection of MSSA. However, this compound was not
very efficacious against the drug-resistant MRSA strain.
Lead optimization and internal replacement of selected
units of the Py3 element led to 2, which protected mice
in the same model against MRSA with an approximate
ED50 of ∼11 mg/kg. A focused optimization of the
isoquinoline 3 resulted in molecules with remarkable
in vitro potency. In particular the MIC value in the
presence of human plasma albumin was greatly im-
proved compared to previously reported DNA binding
antibacterials. Some of these new compounds, especially
the 4-hydroxypiperidines 14 and 19, exhibited in vivo
efficacy at low doses (ED50 of approximately 1 and 5 mg/
kg, respectively) comparable to the best of reference
drugs such as vancomycin. In vitro, these compounds
showed excellent potency against various drug-resistant
and -susceptible Gram-positive bacteria (MIC g 0.004
µg/mL). This study clearly demonstrates the potential
of DNA binding molecules as antibacterials, not only
in vitro but also in a mouse model of MRSA infection.
days with a primary endpoint of survival. Not determined.
potent against all Gram-positve strains, while all de-
rivatives (15-18) lost activity to various degrees. In
contrast, all compounds of the NPy-pBz-ImC series
showed excellent in vitro potency, irrespective of the
nature of their C-terminal amino group (19-22).
We then tested a few previously identified N-termini
in the NPy-pBz-ImC scaffold, using the 4-hydroxypiperi-
dine C-terminus. The MIC profile of the benzothiophene
14 (Scheme 2) turned out to be among the most
promising and is shown as a representative example.
Consistent with the isoquinoline 19, 14 exhibited re-
markable potency against all Gram-positive bacteria
and relatively good activity against E. coli.
The most potent compounds, 5 and 14, were tested
against one of the existing two clinical isolates of
vancomycin-resistant S. aureus (VRSA),17 which was
also resistant to methicillin. Both compounds demon-
strated excellent potency with identical MIC values of
0.008 µg/mL and were 30-fold more active than the
parent compound 2 (Table 2). Time kill kinetics of both
compounds were determined against VRSA. Whereas
5 was bacteristatic at the concentrations examined, 14
was bactericidal at 2-, 4-, and 8-fold MICs after 12 h
(Figure 2). Similar killing kinetics were observed for
both compounds against MRSA (ATCC 27660), however,
without the occurrence of regrowth for 14 between 12
and 24 h (data not shown).
Ack n ow led gm en t. The authors thank Steve Ley
and Heinz Gschwend for helpful discussions and Vernon
J iang and Zhijun Ye for analytical support. Excellent
technical support by Martin Garcia is acknowledged,
This work was supported in part by the Defense
Advanced Research Projects Agency (DARPA Grant No.
N65236-99-1-5427).
Selected compounds were tested in a mouse peritoneal
sepsis model against MRSA as documented previ-
ously.8,9 The hydroxy compounds 19 and 22 successfully
rescued all animals, similar to the parent compound 2
(Table 3). In contrast, the use of methoxy ether 20, the
direct analogue of 19, only partially protected animals
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
1
cedures, H NMR spectra, mass spectra, and purity analysis
data (HPLC-UV). This material is available free of charge
via the Internet at http://pubs.acs.org.

Letters
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 18 4355
(10) Bu¨rli, R. W.; Kaizerman, J . A.; Duan, J .-X.; J ones, P.; J ohnson,
K. W.; Iwamoto, M.; Truong, K.; Hu, W.; Stanton, T.; Chen, A.;
Touami, S.; Gross, M.; J iang, V.; Ge, Y.; Moser, H. E. DNA
binding ligands with in vivo efficacy in murine models of
bacterial infection: optimization of internal aromatic amino
acids. Bioorg. Med. Chem. Lett. 2004, 14, 2067-2072.
(11) Krowicki, K.; Lown, J . W. Synthesis of novel imidazole-contain-
ing DNA minor-groove binding oligopeptides related to the
antiviral antibiotic netropsin. J . Org. Chem. 1987, 52, 3493-
3501.
Refer en ces
(1) Projan, S. J .; Youngman, P. J . Antimicrobials: new solutions
badly needed. Curr. Opin. Microbiol. 2002, 5, 463-465.
(2) Mongodin, E.; Finan, J .; Climo, M. W.; Rosato, A.; Gill, S.;
Archer, G. L. Microarray transcription analysis of clinical
Staphylococcus aureus isolates resistant to vancomycin. J .
Bacteriol. 2003, 185, 4638-4643.
(3) Projan, S. J . New (and not so new) antibacterial agentssfrom
where and when will the novel drugs come? Curr. Opin.
Pharmacol. 2002, 2, 513-522.
(4) Clemett, D.; Markham, A. Linezolid. Drugs 2000, 59, 815-827.
(5) Abbanat, D.; Macielag, M.; Bush, K. Novel antibacterial agents
for the treatment of serious Gram-positive infections. Expert
Opin. Invest. Drugs 2003, 12, 379-99.
(6) Arcamone, F.; Penco, S.; Orezzi, P.; Nicolella, V.; Pirelli, A.
Structure and synthesis of distamycin A. Nature 1964, 203,
1064-1065.
(7) Bu¨rli, R. W.; Ge, Y.; White, S.; Baird, E. E.; Touami, S. M.;
Taylor, M.; Kaizerman, J . A.; Moser, H. E. DNA binding ligands
with excellent antibiotic potency against drug-resistant Gram-
positive bacteria. Bioorg. Med. Chem. Lett. 2002, 12, 2591-2594.
(8) Kaizerman, J . A.; Gross, M. I.; Ge, Y.; White, S.; Hu, W.; Duan,
J . X.; Baird, E. E.; J ohnson, K. W.; Tanaka, R. D.; Moser, H. E.;
Bu¨rli, R. W. DNA Binding Ligands Targeting Drug-Resistant
Bacteria: Structure, Activity, and Pharmacology. J . Med. Chem.
2003, 46, 3914-3929.
(9) Gross, M.; Bu¨rli, R.; J ones, P.; Garcia, M.; Batiste, B.; Kaizer-
man, J .; Moser, H.; J iang, V.; Hoch, U.; Duan, J .-X.; Tanaka,
R.; J ohnson, K. W. Pharmacology of novel heteroaromatic
polycycle antibacterials. Antimicrob. Agents Chemother. 2003,
47, 3448-3457.
(12) Baird, E. E.; Dervan, P. B. Solid-phase synthesis of polyamides
containing imidazole and pyrrole amino acids. J . Am. Chem. Soc.
1996, 118, 6141-6146.
(13) Dervan, P. B.; Edelson, B. S. Recognition of the DNA minor-
groove by pyrrole-imidazole polyamides. Curr. Opin. Struct. Biol.
2003, 13, 1-16.
(14) Urbach, A. R.; Dervan, P. B. Toward rules for 1:1 polyamide:
DNA recognition Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 4343-
4348.
(15) Maeshima, K.; J anssen, S.; La¨mmli, U. K. Specific targeting of
insect and vertebrate telomeres with pyrrole and imidazole
polyamides. EMBO J . 2001, 20, 3218-3228.
(16) National Committee for Clinical Laboratory Standards 2000.
Methods for Dilution Antimicrobial Susceptibility Tests for
Bacteria That Grow Aerobically, 5th ed.; Approved Standard
(M7-A5); National Committee for Clinical Laboratory Stan-
dards: Wayne, PA, 2000.
(17) Bozdogan, B.; Esel, D.; Whitener, C.; Browne, F. A.; Appelbaum,
P. C. Antibacterial susceptibility of a vancomycin-resistant
Staphylococcus aureus strain isolated at the Hershey Medical
Center. J . Antimicrob. Chemother. 2003, 52, 864-868.
J M049712G