Anti-MRSA cephems. Part 1: C-3 substituted thiopyridinium derivatives

Article abstract of DOI:10.1016/S0960-894X(01)00060-9

Sixteen novel cephalosporin derivatives with activity against methicillin-resistant Staphylococcus aureus (MRSA) are described. The compounds were synthesized using substituted thiopyridones, generated either by cyclization of functionalized precursors, or by direct alkylation of the enolate of 2-methyl substituted pyrones. The most active compound in vitro against a strain of MRSA (A27223) displayed an MIC of 0.5 μg/mL. The most efficacious compound in vivo had a PD₅₀ of 2.1 mg/kg.

Full text of DOI:10.1016/S0960-894X(01)00060-9

Bioorganic & Medicinal Chemistry Letters 11 (2001) 797–801
Anti-MRSA Cephems. Part 1:
C-3 Substituted Thiopyridinium Derivatives
Dane M. Springer,* Bing-Yu Luh and Joanne J. Bronson
Anti-infective Chemistry, Bristol-Myers Squibb Pharmaceutical Research Institute, 5 Research Parkway, POBox 5100,
Wallingford, CT 06492, USA
Received 15 September 2000; accepted 24 January 2001
Abstract—Sixteen novel cephalosporin derivatives with activity against methicillin-resistant Staphylococcus aureus (MRSA) are
described. The compounds were synthesized using substituted thiopyridones, generated either by cyclization of functionalized pre-
cursors, or by direct alkylation of the enolate of 2-methyl substituted pyrones. The most active compound in vitro against a strain
of MRSA (A27223) displayed an MIC of 0.5 mg/mL. The most efficacious compound in vivo had a PD50 of 2.1 mg/kg. # 2001
Elsevier Science Ltd. All rights reserved.
Nosocomial infections due to methicillin-resistant
Staphylococcus aureus (MRSA) have increased alar-
mingly in the past two decades.1 Vancomycin remains
the most effective treatment of MRSA in the clinic.
Vancomycin resistance has developed in enterococci
precipitating concern that transfer of this resistance
from enterococci to MRSA would produce an extremely
lethal and incurable pathogen.2 Recent clinical isolates
of MRSA with reduced susceptibility to vancomycin
have given credence to this concern.3 Consequently the
search for new antibiotics with anti-MRSA activity
remains of utmost importance to the future manage-
ment of these infections. This search for new anti-
MRSA compounds has recently been extended to
include new derivatives with a cephem core structure.4
Early in our program to discover an injectable anti-
MRSA cephalosporin we found that compounds with a
lipophilic group at C-7, and a C-3 thio-linked pyr-
idinium moiety, had excellent anti-MRSA activity both
in vitro and in vivo.5 One early compound of biological
interest was the C-3 aminopropyl pyridinium derivative
*Corresponding author. Fax: +1-203-677-7702; e-mail: springed@bms.com
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00060-9

798
D. M. Springer et al. / Bioorg. Med. Chem. Lett. 11 (2001) 797–801
Scheme 1. (a) TMSCH2CH2OH, Et3N, CH2Cl2; (b) (COCl)2, PhH, reflux; (c) MeOCH¼CHC(O)CH3, LiHMDS, THF, ꢀ78 ꢁC; (d) TFA, PhH,
reflux; (e) Lawesson’s reagent, toluene, 80 ꢁC; (f) H2N(CH2)3NHBOC, EtOH; (g) TBAF, THF; (h) MeOH, CH2Cl2; (i) TFA, CH2Cl2.
1. The MIC (minimum inhibitory concentration) of 1
against a homo-resistant MRSA strain (A 27223), was
0.125 mg/mL (2 mg/mL when assayed in the presence of
50% calf serum).6 In a mouse systemic infection model,
compound 1 showed a PD50 of 1.4 mg/kg.7 Unfortu-
nately, 1 was found to be acutely toxic to mice upon iv
bolus administration at concentrations above the
therapeutic dose.8
manifold of cephems that were overall neutral in charge,
and attempt to maximize intrinsic antibacterial activity
while improving solubility and reducing toxicity.
One of our approaches to improve the solubility of
compounds such as 2 was to prepare derivatives that
had the ammonium and carboxylate moieties of the
ornithine side chain at C-3 separated by some distance.
It was hoped that the separation of these charged spe-
cies might have a beneficial effect on solubility by
attenuating their intramolecular association, and favor-
ing their interaction with aqueous media. Scheme 1
illustrates the synthesis of our first target based on this
thesis, cephem 5. Key thiopyridone 4 is generated by a
cyclization strategy in a few steps, and allowed to react
with a C-3 chloromethyl cephalosporanic acid inter-
mediate to provide the C-3 pyridinium derivative.11
Deprotection of the silylethyl ester prior to this step,
and the BOC group afterwards, ultimately yielded tar-
get 5. We were gratified to observe that cephem 5
appeared to be more soluble than compound 2 at varied
pH, and in the presence of added saline. For example, at
pH 7 in water the solubility of 5 was found to be 6 mg/
mL, while the solubility of 2 was 1.5 mg/mL.12 At pH 9,
the solubility of 5 rose to 16 mg/mL, while only 2.4 mg/
mL of 2 was in solution. When 5 was dissolved in 0.9%
saline at pH 9, the solubility was found to be 15 mg/mL,
indicating a negligible salt effect on solubility.
Subsequently, the ornithine-substituted C-3 pyridinium
derivative 2 was synthesized and found to have good
activity against MRSA. We were pleased to observe that
this compound was much less acutely toxic in mice than
1.9 However, cephem 2 had poor aqueous solubility at
neutral pH, likely a consequence of the compound’s bis-
zwitterionic nature.
Acute toxicity data from many more of our cephem
derivatives led us to formulate an empirical ‘rule of
thumb’ for predicting toxicity in our series of analogues.
In general, compounds with a net positive charge at
neutral pH (such as 1) produced rapid death in mice
upon iv dosage. Neutral compounds (such as 2) were
usually much safer, while derivatives with a net negative
charge were generally nontoxic when injected into
mice.10 Unfortunately, cephems carrying a net negative
charge usually had weak activity against MRSA. As a
consequence, we were often obliged to stay within the

D. M. Springer et al. / Bioorg. Med. Chem. Lett. 11 (2001) 797–801
799
Scheme 2.
Table 1.
MICa
PD50
No. type
R
MICa
8 (16)
PD50
b
b
No. type
R
5,I
0.5 (4)
5.1
15,I
>20
8,I
1 (4)
2(4)
2.5
16,I
17,I
18,I
19,I
2(4)
>25
9,II
10,I
11,I
4.3
4 (8)
4 (8)
4 (8)
>23
>23
>23
2 (2)
4 (8)
2.1
5.2
12,I
13,I
14,I
2(2)
3.9
20,I
21,I
22,I
8 (8)
4 (8)
2(4)
>23
2(4)
4 (8)
8.4
19.2
>25
>25
aMIC (in m/mL) versus MRSA A27223, value in parentheses is MIC in presence of 50% calf serum. For details see ref 6.
bPD50 (in mg/kg) for activity against MRSA A27223 in a mouse model of systemic infection. For details see ref 7.

800
D. M. Springer et al. / Bioorg. Med. Chem. Lett. 11 (2001) 797–801
While the chemistry of Scheme 1 delivered ample quan-
tities of 5 for initial in vitro and in vivo assay, the over-
all sequence to key 4-thiopyridone 4 suffered from low
yields.13 We therefore desired a more simple and general
route to substituted thiopyridones to facilitate analogue
synthesis in this branched (relative to 1 and 2) pyr-
idinium series of compounds. We found 2,6-dimethyl-
pyran-4-one to be an inexpensive starting material, and
decided to attempt the direct g-alkylation of one of the
methyl groups via the extended enolate ion.14 We were
pleased to observe that dienolate 6 reacted smoothly
with tert-butyl bromoacetate to afford substituted pyr-
one 7 in 84% yield. The conversion of 7 to cephem 8 was
accomplished using chemistry similar to that depicted in
Scheme 1.
A comparison of the data for compound 5 relative to 10
indicates a slight preference in vitro for compounds with
a three-carbon link between the acid group and the
pyridinium ring, rather than a two-carbon link. How-
ever, the effect of this structural change on the in vivo
efficacy of these two compounds is the opposite; a pref-
erence for the two-carbon link is observed. The data for
compounds 8 and 10 indicate that an extra methyl sub-
stituent on the pyridinium ring is well tolerated. Com-
pounds 14–22 were found to be generally inactive in
vivo even though 16 and 22 had good in vitro activity.17
It is clear from the data presented in Table 1 that the
most active compounds in vivo contain an alkylam-
monium group on the pyridinium nitrogen, as com-
pared to
a neutral or anionic group. Further
representative MIC data for the most interesting com-
pounds in this series is given in Table 2. As Table 1
indicates, these compounds are quite active against a
variety of streptococci, staphylococci, and enterococci.
(The cephems in this class are primarily active against
Gram-positive bacteria, with little activity against the
representative Gram-negative organisms (E. coli, K.
pneumoniae, E. cloacae, P. mirabilis, P. aeruginosa)
included in our screening panel.)
We found that alkylation of 6 was efficient with the
activated halides shown in Scheme 2.15 Using this
approach, and the chemistry of Scheme 1, we synthe-
sized the cephems shown in Table 1.16
Some SARs in this series of cephems can be gleaned
from the data presented in Table 1. In general, we have
found that cephems of Type II, containing a dichloro-
thiopyridyl substituent at C-7, are usually 2–4 times less
active in vitro than the corresponding dichloro-
thiophenyl substituted cephems of Type I . This effect is
observed by the loss of activity for 9 relative to 5 against
our marker strain of MRSA A27223. Interestingly, 9 is
just as active as 5 in vivo. This has often been observed
in our program for cephems of Type II relative to Type
I, and may result from an improved pharmacokinetic or
metabolic profile for compounds containing the
dichlorothiopyridyl substituent at C-7.
In summary, we have synthesized a set of substituted
thiopyridinium cephems via thiopyridones derived from
either acyclic precursors (Scheme 1), or direct alkylation
of 2,6-dimethyl-pyran-4-one (Scheme 2). Many of these
compounds have excellent activity against MRSA in
vitro and in vivo, and remain interesting leads in the
quest for agents to combat these important pathogens.
Additionally, the methods described here should prove
useful to those engaged in the synthesis of compounds
containing substituted thiopyridinium moieties.
Acknowledgements
Table 2. MIC data in mg/mL
MIC of compounds
The authors wish to thank their colleagues in the
Department of Microbiology in Wallingford, CT, for
the biological data reported in this work. We also thank
the Department of Analytical Research in Wallingford,
CT, for analytical data in support of this work.
Organism
A No.
5
8
10
1
2
3
4
5
6
7
8
9
S. pneumoniae/Pen. I
S. pneumoniae/Pen. Resist.
S. pyogenes/Todd Hewitt
E. faecalis
E. faecalis/+50% calf serum A20688
E. faecium A24885
E. faecium/Amp. R, Vanco. S A28156
A27881 0.007
A28272 0.06
0.06
0.25
0.003
0.06
A9604 0.0005 0.0005 0.003
A20688 0.25
0.5
4
2
64
0.5
1
2
64
1
1
32
References and Notes
S. aureus/Pen.ꢀ
A9497 0.001
0.007 0.007
0.125 0.125
0.25
0.25
0.5
1
0.5
1
4
S. aureus/Pen.+
A9606
A9606
0.06
0.25
1. (a) Ayliffe, G. A. J. Clinical Infectious Diseases 1997, 24,
S74. (b) Ehlert, K. Curr. Pharm. Des. 1999, 5, 45.
2. Rotschafer, J. C.; Petersen, M.; Hoang, A. T. D.; Wright,
D. J. Infect. Dis. Pharm. 1998, 3, 2 9.
10 S. aureus/+50% calf serum
11 S. aureus/MS/Pen.+
12 S. aureus/Hetero MR
13 S. aureus/+50% calf serum
14 S. aureus/Hetero MR
15 S. aureus/Homo MR
16 S. aureus/+50% calf serum
17 S. aureus/Homo MR
18 S. aureus/Homo MR
19 S. aureus/Homo MR
20 S. aureus/MR, Pꢀ
0.125
0.125
0.5
1
0.5
2
2
1
4
1
A20241 0.06
A27218
A27218
0.5
0.5
3. (a) Hiramatsu, K.; Hanaki, H.; Ino, T.; Yabuta, K.; Oguri,
T.; Tenover, F. C. J. Antimicrob. Chemother. 1997, 40, 135. (b)
Smith, T. L.; Pearson, M. L.; Wilcox, K. R.; Cruz, C.; Lan-
caster, M. V.; Robinson-Dunn, B.; Tenover, F. C.; Zervos,
M. J.; Band, J. D.; White, E.; Jarvis, W. R. N. Engl. J. Med.
1999, 340, 493.
4. (a) For recent publications concerning the development of
anti-MRSA cephems. see: Hecker, S. J.; Glinka, T. W.; Cho,
A.; Zhang, Z. J.; Price, M. E.; Chamberland, S.; Griffith, D.;
Lee, V. J. J. Antibiot. 2000, 53, 1272. (b) Glinka, T. W.; Cho,
A.; Zhang, Z. J.; Ludwikow, M.; Griffith, D.; Huie, K.;
Hecker, S. J.; Dudley, M. N.; Lee, V. J.; Chamberland, S. J.
A27217 0.25
A27223
A27223
A27621
A27295
0.5
2
0.5
1
0.5
2
A27226 0.25
A27225 0.5
A24548 0.015
A25783 0.03
A27368 0.125
A21638 0.03
0.5
0.5
0.03
0.25
0.5
0.125
2
1
21 S. epidermidis
22 S. epidermidis/MR
23 S. epidermidis/Imipenem R
24 S. haemolyticus
25 S. haemolyticus/Homo MR
0.03
0.25
0.25
0.06
2
A27229
1

D. M. Springer et al. / Bioorg. Med. Chem. Lett. 11 (2001) 797–801
801
Antibiot. 2000, 53, 1045. (c) Ishikawa, T.; Iizawa, Y.; Okonogi,
K.; Miyake, A. J. Antibiot. 2000, 53, 1053. (d) Ishikawa, T.;
Kamiyama, K.; Matsunaga, N.; Tawada, H.; Iizawa, Y.;
Okonogi, K.; Miyake, A. J. Antibiot. 2000, 53, 1071. (e)
Tsushima, M.; Iwamatsu, K.; Umemura, E.; Kudo, T.; Sato,
Y.; Shiokawa, S.; Takizawa, H.; Kano, Y.; Kobayashi, K.;
Ida, T.; Tamura, A.; Atsumi, K. Bioorg. Med. Chem. 2000, 8,
2781. (f) Pae, A. N.; Lee, J. E.; Kim, B. H.; Cha, J. H.; Kim,
H. Y.; Cho, Y. S.; Choi, K. I.; Koh, H. Y.; Lee, E.; Kim, J. H.
Tetrahedron 2000, 56, 5657. (g) Yamazaki, H.; Tsuchida, Y.;
Satoh, H.; Kawashima, S.; Hanaki, H.; Hiramatsu, K. J.
Antibiot. 2000, 53, 546. (h) Yamazaki, H.; Tsuchida, Y.;
Satoh, H.; Kawashima, S.; Hanaki, H.; Hiramatsu, K. J.
Antibiot. 2000, 53, 551. (i) For prior publications from the
laboratories of Bristol-Myers Squibb: D’Andrea, S. V. et al.
Tetrahedron 2000, 56, 5687. (j) Singh, J. et al. Org. Process
Res. Dev. 2000, 4, 488. (k) Kim, O. K.; Hudyma, T. W.;
Matiskella, J. D.; Ueda, Y.; Bronson, J. J.; Mansuri, M. M.
Bioorg. Med. Chem. Lett. 1997, 7, 2753. (l) Kim, O. K.; Ueda,
Y.; Mansuri, M. M.; Russell, J. W.; Bidwell, V. W. Bioorg.
Med. Chem. Lett. 1997, 7, 1945.
scattering spectrometry. These aggregates often passed
through the 0.2 m filter employed prior to the dosing of the
compound in the toxicity assay. Many safe compounds carry-
ing a net zero or net negative charge were found to be inso-
luble and aggregated at neutral pH; therefore we cannot
definitively conclude that insolubility and toxicity are directly
linked for our compounds. The most direct correlate for toxi-
city was that of overall charge of the molecule, as stated in the
text. (We do not have an explanation for this correlation, thus
the use of the term ‘empirical’. We cannot rule out that toxic
cholinergic responses may be involved; however, we have not
performed any experiments to determine the cholinergic
properties of our positively charged derivatives. We also can-
not rule out that a strong interaction with blood constituents
(such as cells, serum proteins and platelets) is operative for our
positively charged compounds, followed by formation of emboli
in the capillaries of the lung. For an example of these latter effects
regarding the toxicity of positively charged macromolecular spe-
cies, see: Moreau, E.; Ferrari, I.; Drochon, A.; Chapon, P.; Vert,
M.; Domurado, D. J. Contr. Rel. 2000, 64, 115.)
11. For a synthesis of the C-3 chloromethyl cephalosporanic
acid intermediate see ref 4i.
5. Kim, C. U.; Misco, P. F.; Wichtowski, J. A.; Ueda, Y.;
Hudyma, T. W.; Matiskella, J. D.; D’Andrea, S. V.; Hoeft, S.
E.; Miller, R. F. U.S. Patent 5,567,698, 1996; Chem. Abstr.
1996, 126, 7903.
12. For solubility determinations, a known quantity of test
compound is dissolved in water or 0.9% saline and the pH
adjusted to the desired point by the addition of minute
amounts of 1 N NaOH. The solutions are stirred for 2h, and
then filtered through a 0.2 m filter. The filtrate is then assayed
by HPLC and the concentration of dissolved compound
determined by comparison to a standard curve.
13. (a) Our synthesis of pyrone 3 was based on the chemistry
reported by Koreeda and Ganem. The presence of the pendant
ester group led to reduced yields through this sequence. For
syntheses of pyrones such as 3, see: Koreeda, M.; Akagi, H.
Tetrahedron Lett. 1980, 21, 1197. (b) Morgan, T. A.; Ganem,
B. Tetrahedron Lett. 1980, 21, 2773.
14. (a) For previous reports of alkylation of the dienolate of
2,6-dimethyl-pyran-4-one, see: Yamamoto, M.; Sugiyama, N.
Bull. Chem. Soc. Jpn. 1975, 48, 508. (b) Smith, A. B., III;
Scarborough, R. M., Jr. Tetrahedron Lett. 1978, 19, 4193. (c)
West, F. G.; Fisher, P. V.; Arif, A. M. J. Am. Chem. Soc.
1993, 115, 1595. (d) West, F. G.; Amann, C. M.; Fisher, P. V.
Tetrahedron Lett. 1994, 35, 9653.
15. Isolated yields after chromatography were in the range of
20–80% for the products illustrated in Scheme 2 and are
unoptimized. The use of chloromethyl methyl sulfide and 1-
iodopropane as alkylating agents for this reaction did not
provide any detectable product. Attempted Michael addition
of the dienolate to t-butyl acrylate was similarly unsuccessful.
16. The thiopyridone required for the synthesis of cephem 10
was made via alkylation of the dienolate anion of the known
2-methyl-pyran-4-one. For a synthesis of this pyrone, see:
Dorman, L. C. J. Org. Chem. 1967, 32, 4105.
6. Antibacterial MICs were determined in broth according to
the standard conditions recommended by the National Com-
mittee for Clinical Laboratory Standards (NCCLS). MIC
assays against MRSA utilized Mueller-Hinton broth+2%
NaCl, a bacterial inoculum size ꢂ 5ꢃ105 CFU/mL, and were
incubated at 35 ꢁC for 24 h. MIC was defined as the lowest
drug concentration inhibiting all visible growth. MIC90s were
determined against a panel of 58 strains of MRSA. Com-
pounds bearing the dichlorothiophenyl amide at C-7 often
have good activity against penicillin-binding protein 2a
(PBP2a). IC50s for inhibition of PBP2a were determined for
three of the compounds discussed here: 1=1.2 mM; 2=2.8
mM; 5=3.4 mM.
7. The PD50 values reported represent the concentration of
the compound that protects 50% of the infected animals from
death in a mouse model of systemic infection. Mice were given
(ip) a lethal dose of MRSA A27233 (homogeneous strain,
2.4ꢃ108 CFU) on day 1, and the test compound was adminis-
tered (im) twice on day 1 at 0.2h and 2h post infection. Ani-
mals were monitored until the end of day five for survival.
8. Compounds were assayed for acute toxicity in mice as fol-
lows: The test compound was dissolved in 5% aqueous dex-
trose at a concentration between 5 and 25 mg/mL. The
solution was filtered through a 0.2 m filter, and between 0.1
and 1 mL of the solution was injected (in usually less than a
min) into the tail vein of three mice. The usual toxic reaction
observed after injection of these types of compounds was red
coloration of the feet, ears, tail and muzzle of the mice, fol-
lowed by respiratory distress and death. Death, when
observed, usually occurred within min of injection. Mice sur-
viving 1 h post injection usually recovered with no obvious
adverse effects. Compound 1, given at a dose of ꢂ200 mg/kg,
resulted in death to all three mice tested.
17. In general, compounds that were found to be inactive in
vivo were not evaluated in mouse pharmacokinetic or phar-
macodynamic studies. Thus, there remain many possible (but
unevaluated) explanations for the inactivity of compounds 14–
22. Each individual compound may have different metabolism
and distribution properties due to the nature of the pendant
groups on the C-3 thiopyridinium ring. A particularly simple
explanation could be that inactive compounds remain loca-
lized at the site of intramuscular injection limiting the expo-
sure of the animal to the test drug. This could either be a
consequence of low solubility, or poor absorption into the
vascular system from the muscle tissue.
9. Compound 2, given at a dose of ꢂ300 mg/kg, was safe to
all three mice tested.
10. We did not observe a direct correlation between a com-
pound’s solubility profile and its acute toxicity. Most of our
compounds appeared to form microparticulate aggregates that
were observed by polarized light microscopy and/or laser light