FimH antagonists: Bioisosteres to improve the in vitro and in vivo PK/PD profile

Article abstract of DOI:10.1021/jm501524q

Urinary tract infections (UTIs), predominantly caused by uropathogenic Escherichia coli (UPEC), belong to the most prevalent infectious diseases worldwide. The attachment of UPEC to host cells is mediated by FimH, a mannose-binding adhesin at the tip of b

Full text of DOI:10.1021/jm501524q

Subscriber access provided by SELCUK UNIV
Article
FimH Antagonists - Bioisosteres to Improve the PK/PD Profile
Beat Ernst, Simon - Kleeb, Lijuan Pang, Katharina Mayer, Deniz Eris, Anja Sigl, Pacal
Zihlmann, Roland Christopher Preston, Timothy Sharpe, Roman Jakob, Daniela Abgottspon,
Aline S. Hutter, Meike Scharenberg, Xiaohua Jiang, Giulio Navarra, Said Rabbani,
Martin Smiesko, Nathalie Lüdin, Jacqueline Bezencon, Oliver Schwardt, and Timm Maier
J. Med. Chem., Just Accepted Manuscript • Publication Date (Web): 10 Feb 2015
Downloaded from http://pubs.acs.org on February 10, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society.
1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the
course of their duties.

Page 1 of 81
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
FimH Antagonists - Bioisosteres to Improve the in vitro and in vivo PK/PD
Profile
a)
a)
a)
a)
a)
Simon Kleeb,  Lijuan Pang,  Katharina Mayer,  Deniz Eris,  Anja Sigl,  Roland C.
1
1
12
a)
a)
c)
b)
a)
Preston, Pascal Zihlmann, Timothy Sharpe, Roman P. Jakob, Daniela Abgottspon, Aline
13
14
a)
a)
a)
a)
a)
S. Hutter, Meike Scharenberg, Xiaohua Jiang, Giulio Navarra, Said Rabbani, Martin
15
16
17
a)
a)
a)
a)
b)
Smiesko, Nathalie Lüdin, Jacqueline Bezençon, Oliver Schwardt, Timm Maier, Beat
1
19
a)
Ernst *
20
2
2
 Equally contributed
2
24
a)
Institute of Molecular Pharmacy, Pharmacenter, University of Basel, Klingelbergstrasse
25
26
5
0, CH-4056 Basel, Switzerland
27
28
29
30
b)
c)
Structural Biology, Biocenter, University of Basel, Klingelbergstrasse 70, CH-4056 Basel
Biophysical Facility, Biocenter, University of Basel, Klingelbergstrasse 70, CH-4056 Basel
31
3
3
34
35
36
37
38
* To whom correspondence should be addressed: Prof. Dr. Beat Ernst, Institute of Molecular
Pharmacy, Pharmacenter, University of Basel, Klingelbergstrasse 50, CH-4056 Basel,
Switzerland; Tel: +41 61 267 15 51, Fax: +41 61 267 15 52; E-mail: beat.ernst@unibas.ch
39
40
41
42
4
ABBREVIATIONS:
4
45
∆H, change in enthalpy; ∆S, change in entropy; AUC, Area under the curve; BSA, bovine
46
47
48
49
50
serum albumin; Cmax, maximal concentration; Caco-2 cells, colorectal adenocarcinoma cells;
CFU, colony forming units; CL , total clearance; CRD, carbohydrate recognition domain; C ,
tot
0
51
52
initial concentration; DL: detection limit; FITC, fluorescein isothiocyanate; FP, fluorescence
53
5
5
polarization; ITC, isothermal titration calorimetry; i.v., intravenous; K , dissociation constant;
D
56
57
MAC , minimal anti-adhesion concentration to inhibit 90% adhesion; PAMPA, parallel
9
0
58
5
60
artificial membrane permeation assay; P , apparent permeability; PD, pharmacodynamics; P ,
app e
effective permeability; PK, pharmacokinetics; p.o., per os; s.c. subcutaneous; UPEC,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 81
1
2
3
4
5
6
7
8
9
uropathogenic Escherichia coli; UTI, urinary tract infection; V , volume of distribution in
z
terminal phase.
1
1
12
13
14
15
16
17
18
19
20
2
2
23
24
25
26
27
28
29
30
31
3
3
34
35
36
37
38
39
40
41
42
4
4
45
46
47
48
49
50
51
52
53
5
5
56
57
58
59
60
ACS Paragon Plus Environment

Page 3 of 81
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
ABSTRACT
Urinary tract infections (UTIs), predominantly caused by uropathogenic Escherichia coli
UPEC) belong to the most prevalent infectious diseases worldwide. The attachment of UPEC
1
1
(
12
1
14
15
16
17
18
19
20
2
2
23
24
25
26
27
28
29
30
31
3
3
34
35
to host cells is mediated by FimH, a mannose-binding adhesin at the tip of bacterial type 1 pili.
To date, UTIs are mainly treated with antibiotics, leading to the ubiquitous problem of
increasing resistance against most of the currently available antimicrobials. Therefore, new
treatment strategies are urgently needed. Here, we describe the development of an orally
available FimH antagonist. Starting from the carboxylate substituted biphenyl -D-mannoside
9 affinity as well as the relevant pharmacokinetic parameters (solubility, permeability, renal
excretion) were substantially improved by a bioisosteric approach. With 3’-chloro-4’-(-D-
mannopyranosyloxy)-biphenyl-4-carbonitrile (10j) a FimH antagonist with an optimal in vitro
PK/PD profile was identified. Orally applied, 10j was effective in a mouse model of UTI by
reducing the bacterial load in the bladder by about 1000-fold.
36
37
38
39
40
41
42
4
4
45
46
47
48
49
50
51
52
53
5
5
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 81
1
2
3
4
5
6
7
8
9
INTRODUCTION
Urinary tract infection (UTI) is one of the most frequent infectious diseases worldwide and
1
1
1
affects millions of people every year. In more than 70% of the reported cases, uropathogenic
12
1
14
2
Escherichia coli (UPEC) is the causal pathogen. Acute, uncomplicated lower urinary tract
15
16
infection, commonly referred to as cystitis, requires an antibiotic treatment for symptom relief
17
18
(i.e. reduction of dysuria, frequent and urgent urination, bacteriuria, pyuria) and for prevention
19
2
2
3
,4
of more devastating or even life threatening complications like pyelonephritis and urosepsis.
2
23
However, the repeated use of antibacterial chemotherapeutics provokes antimicrobial
24
25
5
resistance leading to treatment failure. Hence, a new approach for the prevention and
26
2
28
6
treatment of UTI with orally applicable therapeutics is urgently needed.
29
30
31
3
7
UPEC undergo a well-defined infection cycle within the host. The key step in pathogenesis is
3
34
35
8
bacterial adhesion to the epithelial cells in the lower urinary tract. This interaction prevents
36
37
38
39
40
41
42
UPEC from clearance by the bulk flow of urine and enables the bacteria to colonize the
epithelial cells. The adhesion is mediated by the virulence factor FimH located at the tip of
9
,10
bacterial type 1 pili.
FimH consists of two immunoglobulin-like domains: the N-terminal
4
4
1
1
lectin domain and – connected by a short linker – the C-terminal pilin domain. The lectin
45
4
47
domain encloses the carbohydrate recognition domain (CRD) that binds to the
4
49
12
oligomannosides of the glycoprotein uroplakin Ia on the epithelial cell surface. The pilin
50
51
52
53
domain anchors the adhesin to the pilus and regulates the switch between two conformational
states of the CRD with high and low affinity for mannosides, respectively.
5
5
56
57
5
More than three decades ago, Sharon and co-workers described various oligomannosides and
59
60
1
3,14
aryl -D-mannosides as potential antagonists of the FimH-mediated bacterial adhesion.
ACS Paragon Plus Environment

Page 5 of 81
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
However, only weak interactions in the milli- to micromolar range were observed. In recent
years, several high-affinity monovalent mannose-based FimH antagonists with various
1
5
16
17
16
aglycones like n-alkyl,
phenyl,
dioxocyclobutenylaminophenyl,
umbelliferyl,
1
1
18-22
23
24
25
biphenyl,
indol(in)ylphenyl, triazolyl and thiazolylamino have been reported. In
12
1
14
2
6-32
addition, different multivalent presentations of the mannose have been synthesized
and a
15
16
heptavalent presentation of n-heptyl -D-mannoside (1) tethered to -cyclodextin proved to
17
18
be highly effective when applied together with the UTI89 bacterial strain through a catheter
19
20
2
3
2
into the bladder of C3H/HeN mice. Importantly, adverse side effects resulting from non-
2
23
selective binding of FimH antagonists - they are all -D-mannopyranosides - to mannose
24
2
26
3
3
receptors of the human host system have recently been ruled out.
27
28
29
30
31
3
The high affinities of the monovalent -D-mannopyranosides are based on optimal interactions
3
4-37
with the main structural features of the CRD.
First, the mannose binding pocket
3
34
35
36
37
38
39
40
41
42
4
4
45
accommodating the mannose moiety by means of an extended hydrogen bond network and,
second, the entrance to the binding site composed of three hydrophobic amino acids (Tyr48,
Tyr137, and Ile52) and therefore referred to as ‘tyrosine gate’ hosting aliphatic and aromatic
aglycones. As an example, n-heptyl -D-mannopyranoside (1) exhibits nanomolar affinity due
to hydrophobic contacts of the alkyl aglycone with the hydrophobic residues of the tyrosine
46
47
48
49
50
51
52
1
5
gate. Furthermore, aromatic aglycones, such as present in mannosides 2 & 3 (Figure 1),
provide strong - stacking interactions with the tyrosine gate. This interaction is further
favored by the addition of an electron withdrawing substituent on the terminal ring of the
53
5
5
1
8,19
biaryl portion ( 4).
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 81
1
2
3
4
5
6
7
8
9
OH
OH
OH
OH
O
OH
OH
O
O
HO
HO
HO
HO
HO
HO
Cl
O
O
O
O
O
R
N
H
OEt
115
217
3: R = H18
: R = COOMe18
4
1
1
12
1
14
OH
OH
O
OH
OH
O
OH
OH
O
HO
HO
HO
HO
HO
HO
R
R
Cl
O
O
O
O
O
15
16
NHMe
NHMe
17
18
19
6: R = H20
: R = CF320
819
COOMe
5
: R = Me20,38
7
O
NHMe
20
2
2
23
24
25
26
Figure 1. Monovalent FimH antagonists 1-4 acting as reference compounds and 5-8 which have been orally
explored in in vivo disease models.
27
28
29
30
31
Recent in vivo PK studies in mice proved the high potential of the biphenyl -D-mannosides 5-
3
3
8
for an oral treatment, although high doses (≥ 50 mg/kg) were necessary to achieve the
34
35
1
9-21
minimal concentrations required for the anti-adhesive effect in the urinary bladder.
36
3
38
Moreover, the therapeutic effect could only be maintained for a few hours, i.e. four hours for a
39
40
p.o. (per os) single-dose application of 7 (50 mg/kg), because of rapid elimination by
41
42
1
9
glomerular filtration and low reabsorption from the primary urine in the renal tubules.
4
4
45
46
47
48
49
50
51
52
53
To date, the physicochemical properties affecting the rate of renal excretion, i.e. lipophilicity
and plasma protein binding (PPB), or metabolic liabilities promoting non-renal elimination
pathways have been barely investigated for FimH antagonists. The goal of the present study
was to optimize the biphenyl -D-mannoside with respect to oral bioavailability and renal
5
5
5
57
58
59
60
1
9
excretion. Starting from antagonist 9 (Figure 2), we synthesized new biphenyl derivatives,
characterized their affinity to the CRD, structurally investigated their binding mode, and
determined physicochemical and pharmacokinetic parameters predictive for intestinal
ACS Paragon Plus Environment

Page 7 of 81
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
absorption and renal elimination. Furthermore, we determined in vivo PK (pharmacokinetics)
of the most promising new antagonists in a mouse model. After oral administration, the
compound with the best PK profile proofed effective in reducing the bacterial loads upon
bladder infection in a mouse model of UTI.
1
1
12
13
14
15
16
RESULTS AND DISCUSSION
17
18
19
2
2
2
23
As previously reported, the carboxylate substituent present in the biphenyl mannoside 9 – its
electron withdrawing potential being essential for an enhanced drug target interaction –
24
25
1
9
strongly decreases the lipophilicity of the antagonist (log D7.4 < -1.5 ) in comparison to the n-
26
27
28
1
9
22
heptyl ( 1, log P = 1.7 ) or the unsubstituted biphenyl aglycone ( 3, log P = 2.1 ). Since
29
30
low lipophilicity is a major reason for low intestinal absorption and rapid renal excretion of the
31
3
3
1
9,23
systemically available antagonist,
we aspired to improve oral bioavailability as well as
3
35
39
renal excretion by replacing the carboxylate in 9 with various bioisosteric groups (Figure 2).
36
37
38
39
4
41
42
4
4
4
46
4
48
49
OH
OH
NH2
NHMe
10a
10e
10g
S
O
HO
HO
O
O
O O
NHMe
9
O
1
0b
0c
CN
O
N
N
N
OH
1
10h
10i
HN
N
O
S
O
Me
NHCN
10d
OH
OH
O
O
O
O
HO
HO
5
O
10f
51
F
52
5
OH
5
F
5
56
57
Figure 2. Bioisosteric replacement of the carboxylic acid substituent of biphenyl -D-mannopyranoside 9.
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 81
1
2
3
4
5
6
7
8
9
Synthesis. Iodide 11 was prepared from peracetylated mannose and 4-iodophenol in the
2
2
40
presence of BF ·Et O. In a palladium-catalyzed Miyaura-Suzuki coupling with the boronic
3
2
acid or boronate derivatives 12a-g, the biphenyl derivatives 13a-g were obtained in good to
excellent yields. Final deprotection yielded the test compounds 10a-g. Utilizing microwave-
1
1
12
1
14
15
16
17
18
4
1
assisted reaction conditions, the conversion of arylnitrile 13g to tetrazole 14 proceeded
rapidly and with good yield. After deprotection of 14 using Zemplén conditions, the test
compound 10h was obtained (Scheme 1).
19
20
2
OAc
OAc
O
OH
OH
O
a)
AcO
AcO
b)
HO
HO
2
23
24
O
O
13a-f
10a-f
Ar
Ar
O
25
26
2
2
2
3
31
OAc
OAc
O
O
O
a:
d:
b:
c:
Ar
or
B
NH2
NHMe
N
O
AcO
AcO
O
S
O
OH
O
Ar
B
F
OH
2a-g
1
122
e:
f:
g:
1
Me
NHMe
S
OH
CN
I
O
O
O O
F
3
3
34
OR
OR
O
OR
OR
O
a)
RO
RO
c)
RO
RO
35
O
O
36
37
3
3
1
3g, R = Ac
14, R = Ac
10h, R = H
N
b)
b)
0g18, R = H
1
CN
N
HN
N
40
4
Scheme 1. a) Pd(Cl )dppf·CH Cl , K PO , DMF, 80 °C, 4 h (13a-g, 44-99%); b) NaOMe, MeOH, rt, 4 h (10a-h,
2 2 2 3 4
42
4
4
29-86%); c) TMSN , Bu Sn(O), DME, 150 °C,W, 10 min (81%).
3
2
45
46
47
48
49
50
The cyanobenzamide derivative 10i (Scheme 2) was obtained from 9 by peracetylation ( 15)
followed by conversion of the carboxylic acid into its acid chloride with 1-chloro-N,N,2-
51
5
53
42
trimethyl-1-propenylamine. Without isolation, the acid chloride was reacted with sodium
5
5
hydrogen cyanamide in DMF followed by deacetylation under Zemplén conditions to yield the
56
57
58
test compound 10i.
59
60
ACS Paragon Plus Environment

Page 9 of 81
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
OH
OH
OAc
OAc
OH
OH
O
a)
O
b-d)
HO
HO
AcO
AcO
O
HO
HO
O
O
O
919
H
15
10i
N
COOH
COOH
CN
1
O
1
12
13
Scheme 2. a) i) Ac O, DMAP, pyridine, 0 °C to rt, overnight; ii) satd. NaHCO aq., DCM, rt, 2 h (15, 53%); b) 1-
2
3
1
15
chloro-N,N,2-trimethyl-1-propenylamine, toluene, 0 °C to rt, 2 h; c) NaH, NH CN, DMF, 0 °C to rt, overnight; d)
2
1
NaOMe, MeOH, rt, 4 h (10i, 21% for three steps).
17
18
19
2
2
18
Finally, to further improve the pharmacokinetic properties of mannoside 10g (see Table 3), a
2
2
24
25
26
27
28
29
chloride substituent was introduced to the ortho-position of the aromatic ring adjacent to the
anomeric oxygen. For its synthesis, peraceylated -D-mannose (16) was coupled with 2-
chloro-4-iodophenol (17) using BF Et O as promotor ( 18, 76%). After the introduction of
3
2
30
31
3
3
the second aromatic ring by Miyaura-Suzuki coupling ( 19, 75%), deprotection yielded
mannoside 10j (Scheme 3).
34
3
36
3
Cl
HO
OAc
OAc
HO
OAc
OAc
O
B
CN
OR
OR
O
HO
O
17
AcO
AcO
I
AcO
AcO
12g
RO
RO
Cl
Cl
a)
b)
38
OAc
O
O
39
4
16
1823
19, R = Ac
0j, R = H
I
c)
1
41
42
CN
4
4
Scheme 3. a) BF Et O, CH Cl , 40 °C (76%); b) Pd(Cl )dppf·CH Cl , K PO , DMF, 80 °C (75%); c) NaOMe,
3
2
2
2
2
2
2
3
4
45
46
MeOH, rt, 4 h (48%).
47
48
49
50
51
Binding Affinity. The binding affinity of heptyl mannoside 1, the biphenyl mannosides 3, 9,
5
53
18
2
0, and the bioisosteres 10a-j was determined in a competitive fluorescence polarization
5
5
56
57
assay (FP-assay) and with isothermal titration calorimetry (ITC). A protein construct
consisting of the CRD with a C-terminal His-tag with a thrombin cleavage site (FimH-CRD-
58
59
60
43
Th-His ) was used for all experiments.
6
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 81
1
2
3
4
5
6
7
8
9
Competitive Fluorescence Polarization Assay. For the rapid evaluation of binding affinity,
we established a competitive binding assay based on fluorescence polarization (FP). Similar
1
8,44
formats have been applied before for the detection of carbohydrate-lectin interactions.
In
1
1
this assay, the antagonist of interest displaces a fluorescently labeled competitor from the
12
1
14
15
16
17
18
19
20
2
2
23
24
25
4
5
binding site, thereby causing a reduction in fluorescence polarization. To identify the optimal
competitor, fluorescein isothiocyanate (FITC) was connected to the FimH ligand 21 by three
linkers of different lengths ( 22-24, Scheme 4). For optimal sensitivity and signal-to-noise
ratio, three main parameters need to be considered: (i) the affinity of the competitor should not
be impaired by the fluorescent label, (ii) the conformational flexibility of the label upon
binding of the competitor to the CRD should be low and (iii) the fluorescence properties of the
26
27
28
4
6-48
label should not be affected by the connected ligand.
A change in fluorescence properties
29
30
31
3
3
34
35
36
37
38
39
40
was observed for reporter ligand 22 in which the label was linked to the biphenyl agylcone by
an amide bond. The absorption spectrum revealed a lack of the characteristic fluorescein
absorption peak at 494 nm (Scheme 4), likely due to an extension of the conjugated system to
the biphenyl moiety of the ligand. The elongated saturated spacer groups in competitors 23 and
24 ensured that the expected spectral properties of the dye were retained (Scheme 4).
41
42
4
4
45
46
47
48
49
50
51
52
53
5
5
56
57
58
59
60
ACS Paragon Plus Environment

Page 11 of 81
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
12
13
14
15
16
17
18
19
20
2
2
23
24
25
26
27
28
29
30
31
3
3
34
35
36
37
38
39
4
Scheme 4. a)
1-[(1-(Cyano-2-ethoxy-2-oxoethylideneaminooxy)-dimethylamino-morpholinomethylene)]
41
4
methanaminium hexafluorophosphate (COMU), NEt , fluoresceinamine, DMF, rt, 7 h (22, 19%); b) i. DIC, NHS,
3
4
4
45
4
N-Boc-ethylenediamine, DMF, rt, 12 h; ii. TFA, DCM, rt, 10 min (68% over two steps), iii. fluorescein
isothiocyanate (FITC), NEt , DMF, rt, 3 h (23, 48%); c) i. DIC, NHS, N-Boc-PEG2-NH , DMF, rt, 14 h; ii. TFA,
3
2
47
4
DCM, rt, 30 min (62% over two steps); iii. FITC, DMF, rt (24, 65%).
49
50
51
52
53
5
5
56
57
58
59
For the determination of their binding affinity, fixed concentrations of the reporter ligands 23
and 24 were incubated for 24 h with a linear dilution of the FimH-CRD (0-100 nM). FP was
measured using a plate reader, with polarized excitation at 485 nm, and emission at 528 nm
measured through appropriately oriented polarizers. Fitting the single-site binding function of
60
4
9
Cooper to the observed FP data resulted for compound 23 in a dissociation constant (K = 1.7
D
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 81
1
2
3
4
5
6
7
8
9
19
nM, Figure 3A) similar to that of the unlabeled parent compound 21, whereas 24 showed a
five-fold lower affinity (9.9 nM) (Scheme 4). Therefore, the reporter ligand 23 fulfills all
characteristics as an optimal competitor and was used for the FP assay.
1
1
12
13
14
15
16
17
18
19
20
2
2
23
24
Figure 3. A) Direct binding curve of the labeled competitor 23 obtained by adding a linear dilution of FimH-CRD
25
26
(
0-100 nM) and a constant concentration of competitor 23 (5 nM). The KD was determined by fitting the
27
28
experimental data to a single-site binding fit that accounts for ligand depletion. In three FP based direct binding
29
30
experiments the K of competitor 23 was determined to be 1.7 nM. B) Inhibition curve of n-heptyl mannoside (1)
D
31
3
from the competitive FP assay. The IC50 value was determined by nonlinear least-squares fitting to a standard 4-
3
34
45
parameter equation. A modified Cheng-Prusoff equation was used to calculate the corresponding K value (K =
D
D
35
36
2
8.3 nM).
37
38
39
4
41
For the test compounds 1, 3, 9, 20, and 10a-j, a 24 h incubation time was applied before FP
4
4
50
was measured due to the long residence time of FimH antagonists (t1/2 > 3.5 h, Figure 3B ).
4
45
46
47
48
49
50
The 24 h incubation period was empirically determined to be necessary to reach equilibrium
between reporter ligand and compound of interest. IC50 values were obtained by nonlinear
least-squares regression (standard four-parameter dose response curve) and converted to KD
51
5
53
5
5
56
57
58
4
5
values using a modified Cheng-Prusoff equation. This equation accounts for the ligand
depletion effect in competitive titrations involving high-affinity interaction partners present in
similar concentrations. Under these conditions, the free concentration of an interacting species
cannot be assumed to equal the total concentration.
59
60
ACS Paragon Plus Environment

Page 13 of 81
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The K values determined for the test compounds 1, 3, 9, 20, and 10a-j are summarized in
D
Table 1. Against our expectations, the biphenyl mannosides 3 and 9 exhibit similar affinities
(Table 1), despite the presence of an electron withdrawing carboxylate substituent in
1
1
12
13
14
15
16
antagonist 9. According to the crystal structure of FimH co-crystallized with the sulfonamide
derivative 10e (Figure 4A), the outer aromatic ring of the biphenyl aglycone forms π-π
1
5
interactions with the electron rich Tyr48, which is part of the tyrosine gate of FimH.
A
17
18
19
20
2
2
23
24
25
26
27
28
reduction of electron density of the aglycone by the electron withdrawing carboxylate was
expected to enforce these π-π stacking interactions and lead to improved affinity. However,
this beneficial effect might be compensated by an entropic penalty originating from the
improved π-π stacking to Tyr48 that might lead to the reduced flexibility of both protein and
antagonist. Furthermore, a beneficial enthalpy effect might be partially compensated by an
2
30
51
enthalpy penalty originating from the desolvation of the charged carboxylate in 9 (see also
31
3
3
34
experimental part). Although this substituent is solvent exposed, at least a partial desolvation
may be necessary upon antagonist binding. To prove this assumption, we replaced the
35
36
37
19
carboxylate by the corresponding methyl ester ( 20) in order to reduce the desolvation
38
3
40
52
penalty and, as predicted by the Hammett constant σ , to further improve the π-π stacking.
p
41
42
Indeed, a six-fold improvement in affinity was achieved. However, since the methyl ester
4
4
1
9
undergoes rapid enzyme-mediated hydrolysis in vivo, it will not be available at the place of
45
46
47
action in the urinary bladder. The methyl ester was therefore replaced by metabolically stable
4
49
39
52
bioisosteres exhibiting comparable electron withdrawing properties (Table 1, entries 5-13).
50
5
52
The most potent derivatives 10d, 10e and 10g showed affinities in the low nanomolar range.
53
5
5
56
2
2
As previously reported, a chloro substituent in the ortho-position of the aromatic ring
57
5
59
adjacent to the anomeric oxygen is favorable for affinity and improves the physicochemical
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 81
1
2
3
4
5
6
7
8
9
properties relevant for oral bioavailability. Indeed, the corresponding antagonist 10j was the
most potent compound tested in this study.
Table 1. Affinities (K ) of FimH antagonists to FimH-CRD-Th-His ; dissociation constants (K ) were determined
D
6
D
1
in a competitive fluorescence polarization assay.
1
12
13
1
15
16
1
OH
OH
Affinity
O
Entry
Compd
HO
HO
K [nM]
D
OR
18
19
20
2
2
23
24
1
2
1
3
O
O
28.3  5.0
15.1  2.2
25
26
2
28
2
3
31
3
3
34
35
36
37
38
39
40
41
42
4
4
45
46
47
48
49
50
51
5
53
5
5
56
O
O
O
O
O
3
4
5
6
7
8
9
17.9  1.5
3.6  0.9
2.8  0.3
2.9  0.5
3.0  0.1
1.7  0.2
OH
O
20
OMe
NH2
NHMe
N
O
O
O
10a
10b
10c
10d
O
O
O
Me
S
57
O
O
58
59
60
ACS Paragon Plus Environment

Page 15 of 81
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
O
9
10e
2.7  0.4
NHMe
S
O
O
O
1
1
12
13
14
15
16
1
18
19
20
2
2
23
F
1
0
1
10f
10g
10h
3.7  0.2
2.0  0.6
5.7  0.1
OH
F
O
1
CN
O
12
N
N
HN
N
24
25
O
26
27
28
29
30
3
3
1
1
3
4
10i
10j
8.4  0.3
NHCN
O
Cl
O
a)
< 1
3
CN
34
35
36
3
38
a)
The K value of 10j was approximated to be in the subnanomolar range. The IC50 value obtained in the
D
39
40
competitive FP assay was equal to the lowest value that can be resolved by the assay, indicating stoichiometric
41
42
titration of 10j due to its high affinity. Consequently, its K must be below the K of competitor 23.
D
D
4
4
45
46
47
48
49
50
51
52
Isothermal Titration Calorimetry (ITC). To further confirm our hypothesis regarding π-π
stacking and desolvation, we performed ITC experiments with the reference compound 1, the
unsubstituted biphenyl mannoside 3, the carboxylic acid 9, and the bioisosteres 10b-e, g and j
(Table 2). ITC allows the simultaneous determination of the stoichiometry (N), the change in
53
5
5
56
5
3,54
enthalpy (∆H) and the dissociation constant (K ) for ligand-protein binding.
The reliable
D
57
58
determination of these three parameters requires well-defined sigmoidal titration curves
59
6
-1
characterized by the dimensionless Wiseman parameter c (c = Mt(0) K , where Mt(0) is the
D
5
5
initial macromolecule concentration). To be sure that data can be fitted with confidence, the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 81
1
2
3
4
5
6
7
8
9
56
c-value should be between 1 and 1,000 (ideally between 5 and 500), which could be achieved
for the antagonists 3 and 9. For titrations involving low micromolar Mt(0) and interactions in
the low nanomolar or picomolar range, as suggested for the bioisosteres 10b-j, c-values above
1
1
1
,000 were expected. Since these conditions lead to steep titration curves that do not allow the
12
13
14
15
determination of the curve slope representing 1/K , we applied an alternative, competitive
D
5
7,58
format referred to as displacement assay.
First, FimH-CRD-Th-His was pre-incubated
6
16
17
18
19
20
2
2
23
with the low affinity antagonist n-heptyl 2-deoxy--D-mannopyranoside (25, for synthesis see
supporting information). The high-affinity bioisosteres of interest were titrated into the protein-
ligand complex giving well-defined sigmoidal titration curves.
24
25
26
27
Table 2. Thermodynamic parameters from ITC for selected FimH-antagonists binding to FimH-CRD-Th-His ; n,
6
28
29
stoichiometric correction factor; CI, confidence interval from fitting.
30
31
3
3
34
35
OH
[a]
[a]
K
D
G

H
-TS
Type of
OH
O
Entry Compd
HO
HO
n
[nM]
[kJ/mol]
[kJ/mol]
[kJ/mol]
measurement
OR
36
37
38
39
40
41
4
4
4
2
8.9
-50.3
[
b,c]
1
2
1
O
O
-43.0
-44.2
7.3
0.8
1.00
1.07
direct
direct
(
25.8 – 32.3)
(-50.2 – -50.7)
17.7
-45.0
[
b]
3
(
14.1 – 22.3)
(-44.5 – -45.6)
45
46
4
4
49
50
51
52
53
5
5
56
57
5
5
O
O
O
15.0
-48.7
3
4
5
9
-44.7
-47.8
-47.4
4.0
6.7
7.1
1.05
1.02
0.97
direct
OH
(
13.4 – 16.7)
(-48.4 – -49.0)
O
O
O
4
.3
-54.5
competitive
10b
10c
NHMe
(3.2 – 5.6)
(-54.1 – -54.9)
vs. 25
5.0
-54.5
competitive
O
N
(
3.8 – 6.6)
(-54.1 – -54.8)
vs. 25
60
ACS Paragon Plus Environment

Page 17 of 81
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
O
3
.0
-52.3
competitive
6
7
10d
10e
-48.7
-48.2
3.6
3.9
0.99
1.06
Me
(2.1 – 4.2)
(-51.5 – -53.1)
vs. 25
S
O
O
O
3
.5
-52.2
competitive
1
1
12
NHMe
(
2.9 – 4.3)
(-51.6 – -52.8)
vs. 25
S
O
O
13
14
15
16
1
18
1
20
2
O
2
.8
-58.2
competitive
8
9
10g
10j
-48.8
-50.7
9.4
1.00
1.01
(
2.3 – 3.3)
(-57.8 – -58.6)
vs. 25
CN
Cl
O
1.3
-60.9
competitive
10.1
(
1.1 – 1.6)
(-60.4 – -61.4)
vs. 25
2
CN
23
2
2
26
27
9’386
OH
-19.5
HO
O
1
0
25
HO
(8’555 –
-28.7
-9.1
1.00
direct
O
(
-19.1 – -20.0)
1
0’287)
28
29
30
3
3
[a]
[b]
9
5% confidence interval from fitting in parentheses; Global fit including two direct titration measurements;
3
34
[c]
37
ITC data were previously published with an n-value of 0.82.
35
36
37
3
39
The resulting K
D
values (Table 2) correspond well with the data obtained from the FP assay
40
41
(Table 1). A comparison of the thermodynamic fingerprints of antagonists 3 and 9 reveals that
42
4
4
45
46
47
48
the more favorable enthalpic contribution resulting from facilitated π-π stacking leads to a net
enthalpy gain (ΔΔH: -3.7 kJ/mol). However, an even greater increase in enthalpy is likely
countered by the enthalpy costs for desolvation of the electron withdrawing carboxylate.
49
50
51
5
53
5
5
56
57
58
59
The gain in enthalpy is in turn compensated by an unfavorable entropy (-TΔΔS: 3.2 kJ/mol) as
a result of the reduced flexibility of both the antagonist and the Tyr48 side-chain caused by the
improved interaction. This is not entirely outweighed by the beneficial entropy contribution
related to the partial desolvation of the carboxylate and the related release of water into the
bulk. Added together, the enthalpy and entropy contributions of antagonists 3 and 9 result in
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 81
1
2
3
4
5
6
7
8
9
1
1
12
13
similar affinities (K : 17.7 and 15.0 nM, respectively).
D
In contrast, the replacement of the carboxylate group by various neutral bioisosteres (entries 4-
) reduces the enthalpy costs for desolvation (see calculated free energies of desolvation,
experimental part) and therefore leads to a markedly improved enthalpy (ΔΔH: -3.5 to -5.8
kJ/mol). As a result, an up to fivefold improvement of the K values was achieved. Finally,
7
14
15
16
D
17
18
19
20
2
2
23
24
25
26
27
28
29
30
31
3
with a cyano substituent (entries 8 & 9), the enthalpy term was further improved (ΔΔH: -3.7
kJ/mol) due to a reduced desolvation penalty and improved π-π stacking interactions.
However, this beneficial component is again partially compensated by a decrease in entropy.
This can be attributed, first, to the loss of flexibility of the tightly bound ligand (Figure 4B)
and, second, to the smaller surface area of the cyano substituent compared to amide,
sulfonamide and sulfone, which results in a smaller number of water molecules being released
to bulk upon binding.
3
34
35
36
37
38
39
40
41
42
X-ray Crystallography. To determine the binding poses of the bioisosters, we co-crystallized
the compounds 10e and 10j with the FimH-CRD (Figure 4). Atomic resolution crystal
structures were obtained at 1.07 Å (10e) and 1.10 Å (10j). As observed in previous mannoside
4
4
1
5,18,36
co-crystal structures,
the mannose moiety forms an extensive hydrogen bond network to
45
4
47
48
49
50
51
52
53
5
5
56
57
the well-defined binding site with all of its hydroxyl groups. The biphenyl aglycone is located
between the tyrosine gate residues (Tyr48/Tyr137). The π-π stacking of the second aromatic
ring of the aglycone to the side chain of Tyr48 contributes most to the interaction energy of the
aglycone moiety. Interactions to the Tyr137 side-chain on the other hand are only limited.
Whereas a previously published crystal structure of a biphenyl mannoside in complex with
FimH-CRD suffers from crystal contacts of binding site residues (Tyr48 side-chain to
58
59
60
1
8
backbone oxygen of Val27) possibly causing the distortion of the binding site, the binding
site of our structures are mostly solvent exposed. This revealed the flexibility of the aglycone
ACS Paragon Plus Environment

Page 19 of 81
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
12
13
in the FimH-CRD/10e structure, since the electron density towards the solvent-exposed
sulfonamide indicates that there is not one single orientation. Therefore, the aglycone was
modeled in two distinct poses. In contrast, in the FimH-CRD/10j structure the amino acid side
chain of Y48 can be modeled in two distinct rotamers, suggesting flexibility also of the
receptor.
14
15
16
17
18
19
20
2
2
23
24
25
26
27
28
29
30
31
3
3
34
3
36
37
38
39
40
41
42
Figure 4. Ligand binding poses determined by X-ray co-crystallization with compounds 10e resolved to 1.07 Å
(A) and 10j resolved to 1.10 Å (B). The electron density surrounding the aglycone of 10e indicates flexibility of
the aglycone and was modeled in two poses. Both compounds bind in a similar pose with a well-defined hydrogen
network surrounding the mannose moiety and π-π stacking interactions between the second aromatic ring and
Tyr48 side-chain (A). In contrast, in the FimH-CRD/10j structure the amino acid side chain of Y48 can be
modeled in two distinct rotamers, suggesting flexibility also of the receptor (B).
4
4
45
46
47
48
49
50
51
52
53
5
5
56
57
58
59
60
Physicochemical Properties and In Vitro Pharmacokinetics. Intestinal absorption and renal
excretion are prerequisites for a successful oral treatment of UTI with FimH antagonists.
Furthermore, reabsorption of antagonist from the renal ultrafiltrate is desirable for maintaining
the minimal anti-adhesive concentration in the target organ, namely the bladder, over an
extended period of time. To estimate the influence of the bioisostere approach on oral
bioavailability and the rate of renal excretion, we determined lipophilicity by means of the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 81
1
2
3
4
5
6
7
8
9
59
octanol-water distribution coefficient (log D ), aqueous solubility, and membrane
7
.4
6
0
permeability in the artificial membrane permeability assay (PAMPA) and the colorectal
6
1
adenocarcinoma (Caco-2) cell monolayer model.
1
1
12
13
14
15
16
17
18
19
20
2
2
23
24
25
26
27
28
29
30
31
3
3
34
35
36
37
38
39
40
41
42
4
4
45
46
47
48
49
50
51
52
53
5
5
56
57
58
59
60
ACS Paragon Plus Environment

Page 21 of 81
Journal of Medicinal Chemistry
1
2
3
Table 3. Physicochemical and in vitro pharmacokinetic parameters.
4
5
6
7
8
9
Caco-2 Papp,
Metabolic
Solubility
PAMPA log Pe
PPB fb
a)
b)
-6
e)
Compd pK
log D7.4
[10 cm/s]
stability t1/2
a
c)
d)
f)
[µg/mL] / pH
[cm/s] / pH
[%]
g)
ab
ba
[min]
1
1
12
13
14
15
16
1
3
---
---
---
1.65
> 3000
-4.89
-4.7  0.1 / 7.4
-4.7
7.0  0.6
9.4  0.2
81
93  1
93
13
n.d.
1.0
2.1  0.1
2.14
21  1 / 7.4
33.8 / 6.51
10.0  0.9
4.23
19.0  1.2
n.d.
2
0
17
18
19
20
2
2
23
2
25
26
27
28
9
3.88
---
< -1.5
> 3000 / 6.61
no permeation
n.d.
n.d.
73
> 60
n.d.
1
0a
0b
0.5  0.1
12  1 / 7.4
-6.8  0.3 / 7.4 0.12  0.01
0.61  0.03
n.d.
1
---
0.8  0.0
0.2  0.1
0.4  0.0
0.7  0.1
122  13 / 7.4 -9.2  1.4 / 7.4 1.10  0.82
> 250 / 7.4
-7.8  0.3 / 7.4 0.18  0.07
246  17 / 7.4 -7.2  0.0 / 7.4 0.36  0.01
0.87  0.15
1.30  0.03
1.76  0.12
1.82  0.14
n.d.
n.d.
> 60
> 60
> 60
10c
0d
0e
---
48  2
99  1
> 99
1
---
1
---
> 250 / 7.4
-8.6  0.2 / 7.4 0.28  0.23
29
30
31
3
3
34
35
36
37
38
39
40
4
4
4
4
>
>
150 / 3.0
150 / 7.4
-7.7  0.8 / 5.0
0.40  0.02
-8.8  0.1 / 7.4
1
0f
6.5
---
1.1  0.0
1.4  0.0
-1.4  0.1
1.90  0.17
13.2  2.1
0.22  0.01
n.d.
99  0
n.d.
n.d.
> 60
n.d.
1
0g
0h
186  4 / 7.6
1  0 / 3.0
73  2 / 7.4
> 150 / 3.0
150 / 7.4
192  5 / 7.4
-5.7  0.0 / 7.4
2.0  0.1
1
-9.3  1.4 / 5.0
-8.8  1.4 / 7.4
1
3.7
0.17  0.00
2
-6.8  0.2 / 5.0
-7.0  0.1 / 7.4
10i
10j
2.5
---
-1.1  0.1
2.1  0.0
0.22  0.14
2.2  0.4
0.29  0.03
22.1  1.5
n.d.
n.d.
>
4
-5.2  0.0 / 7.4
89  1
> 60
46
47
48
4
a) pK values were determined by NMR spectroscopy; b) Octanol-water distribution coefficients (log D ) were
a 7.4
50
5
52
5
determined by a miniaturized shake-flask procedure at pH 7.4, values represent the mean  SD of sextuplicate
59
measurements; c) Kinetic solubility was measured in a 96-well format using the µSOL Explorer solubility
5
5
analyzer at the indicated pH in triplicate; d) P = effective permeability: passive permeation through an artificial
e
56
5
membrane was determined by the parallel artificial membrane permeation assay (PAMPA), values represent the
58
5
60
mean  SD of quadruplicate measurements performed at the indicated pH; e) Papp = apparent permeability:
60
permeation through a Caco-2 cell monolayer was assessed in the absorptive (ab) and secretory (ba) directions
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 81
1
2
3
4
5
6
7
8
9
61
62
in triplicate; f) Plasma protein binding (PPB) was determined by equilibrium dialysis in triplicate; g)
Metabolic stability was determined by incubating the compounds (2 µM) with pooled rat liver microsomes (RLM,
63
0
.5 mg/mL) in presence of NADPH (1 mM, compounds 1, 9, 10c-e, g, j) or without NADPH (compound 20);
n.d. = not determined.
1
1
12
13
14
Oral Bioavailability. Oral bioavailability of a compound relies on solubility, permeation
15
1
17
64,65
through the membranes lining the intestine, and stability against first pass metabolism.
As
1
66
67
discussed by Lipinski and Curatolo, dose and permeability define the minimum aqueous
solubility required for oral administration. Thus, a dose of 1 mg/kg of a moderately permeable
compound requires a solubility of at least 52 µg/mL. Whereas sufficient aqueous solubility (>
19
20
2
2
23
24
25
26
1
9
3
000 µg/mL) was reported for n-heptyl -mannopyranoside (1), the unsubstituted biphenyl
27
28
29
-D-mannopyranoside 3 as well as the antagonists bearing a methylcarboxylate, carboxamide,
3
31
22
or tetrazole substituent (compounds 20, 10a and 10h) were found to be scarcely soluble. As
3
3
34
35
36
37
38
6
8
proposed by Ishikawa, a possible reason is the apolar and planar aglycone. By contrast, the
polar carboxylic acid moiety present in antagonist 9 or the substituents in the bioisosteres 10b-
j enhance solubility to 122-273 µg/mL, a level sufficient for in vivo PK studies. For in vivo
39
4
41
disease studies, however, dosages of up to 10 mg/kg were foreseen (see below), requiring a
4
4
66,67
solubility of 520 µg/mL.
For this reason, surfactant Tween 80 (1%) had to be added.
4
45
46
47
48
49
50
51
52
6
9
70
Furthermore, permeability data derived from PAMPA and the Caco-2 model suggest
moderate to high permeation of the moderately lipophilic antagonists 1, 3 and 20 (log D7.4
.6) through the intestinal membranes. The bioisosteres 10a-f, h, i, although slightly more
>
1
53
5
5
56
57
58
59
60
permeable than the strongly hydrophilic carboxylic acid derivative 9, show only low values of
permeability compared to n-heptyl -D-mannopyranoside (1) or the unsubstituted biphenyl
mannoside 3. However, the para-cyanobiphenyl derivatives 10g and 10j display elevated log
D7.4 and effective permeability (log P ) in the range for successful intestinal absorption.
e
ACS Paragon Plus Environment

Page 23 of 81
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Regarding both, sufficient aqueous solubility and elevated membrane permeability, the para-
cyano substituted bioisosteres 10g and 10j are thus the most promising candidates for oral
absorption. Moreover, combining the bioisosteric replacement with the addition of a chloro
substituent in the ortho-position of the aromatic ring adjacent to the anomeric oxygen (
1
1
12
1
14
2
2
10j) resulted in the most advantageous physicochemical profile for oral bioavailability.
15
16
17
18
Renal Excretion. The rate of renal excretion depends on the rate of glomerular filtration and
19
20
2
7
1
the propensity to tubular secretion and reabsorption of an antagonist. Only the fraction that is
2
23
72
not bound to plasma proteins is expected to enter the glomerular filtrate. Plasma protein
24
2
6
2
binding (PPB) data indicating the fraction bound (f ) are listed in Table 2. The biphenyls 9
b
26
27
28
and 10c were identified as moderate binders to plasma proteins (f ≤ 65%), which suggests a
b
29
30
low impact of PPB on antagonist filtration. The f values of the antagonists 1, 3, 20 and 10j
b
31
3
3
were between 80 and 93%, whereas the bioisosteres 10d, e and g showed particularly high
34
35
protein binding (f ≥ 99%) implying slow compound entry into the primary urine. However,
b
36
37
the kinetic aspects of PPB, that is, association and dissociation rate constants, remain to be
38
3
40
7
3
determined to quantify precisely the influence of PPB on filtration.
41
42
4
4
7
4-76
Furthermore, log D7.4 was identified as key determinant of tubular reabsorption.
45
46
47
48
49
50
51
52
53
5
5
56
57
58
59
60
Accordingly, lipophilic compounds are predominantly reabsorbed from the renal filtrate. Given
that renal clearance is the major route of elimination, this will result in a slow but steady
excretion into the bladder. In contrast, hydrophilic compounds are poorly reabsorbed and thus
quickly renally eliminated, which leads to high initial compound levels in the urine but
narrows the time range where the minimal anti-adhesive concentration is maintained.
Consequently, low log D7.4 as shown for the antagonists 9, 10h and 10i implies low tubular
reabsorption and rapid elimination of the filtered molecules by the urine. Otherwise, log D7.4
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 81
1
2
3
4
5
6
7
8
9
between 0.2 and 0.7, such as determined for the bioisosteres 10a-e, suggests increasing
propensity to tubular reuptake, whereas log D7.4 > 1 as shown for heptyl mannoside 1 and the
biphenyl mannosides 3, 20, 10g, 10f and 10j is optimal for tubular reabsorption from the
glomerular filtrate and thus for slow renal clearance.
1
1
12
13
14
15
16
Metabolic Stability. Increasing lipophilicity is usually paralleled by increasing susceptibility
17
18
19
20
2
2
23
24
25
26
27
28
7
7
to metabolism. Liabilities towards metabolic clearance pathways which prevent the intact
antagonist from reaching the target in the bladder were therefore of interest. To assess their
propensity to cytochrome P450 (CYP450)-mediated metabolism, heptyl mannoside 1, the
carboxylic acid derivative 9, and the bioiosteres 10c-e, g, j were incubated with rat liver
microsomes (RLM, 0.5 mg/mL) in presence of the cofactor -nicotinamide adenine
29
30
31
3
3
34
35
36
37
38
39
40
41
42
4
4
6
3
dinucleotide phosphate (NADPH). To confirm the high propensity of the methyl ester present
in antagonist 20 to carboxylesterase (CES)-mediated hydrolysis, this antagonist was incubated
with RLM only. The profiles of unchanged compound versus time revealed high susceptibility
of heptyl mannoside 1 to CYP450-mediated metabolism (t1/2 = 13 min) and rapid hydrolysis of
the ester 20 by the hepatic CES (t1/2 = 1.0 min). Otherwise, the bioisosteres 10c-e, g & j were
stable against enzyme-mediated bioconversion (t1/2 > 60 min) suggesting lower propensity to
metabolic, non-renal elimination pathways.
45
46
47
48
49
50
51
52
53
5
5
Considering PPB, lipophilicity, and metabolic stability data, we therefore expected (i) a steady
release of compounds 10d, e, g, j into the bladder because of high PPB decelerating glomerular
filtration (10d, e, g) and/or high log D7.4 supporting tubular reabsorption (10g, j), (ii) a fast
excretion of antagonists 9 and 10c via the urine due to low PPB and low log D , and (iii) a
7
.4
56
57
5
59
rapid clearance of heptyl mannoside 1 from the body by renal and metabolic pathways.
60
ACS Paragon Plus Environment

Page 25 of 81
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Compounds featuring high propensity to renal excretion as major route of elimination (10c,
0e and 10j) were selected for in vivo PK studies in a mouse model.
1
1
1
12
13
14
15
16
17
18
19
20
2
2
23
Pharmacokinetic Studies in C3H/HeN Mice. This first part of our study explored the
predicted effects of lipophilicity, PPB, and metabolic stability on antagonist disposition and
elimination upon a single dose i.v. application (50 mg/kg) of compounds 10c and 10e. The PK
parameters of these applications and those of the previously published carboxylate 9 are
summarized in Table 4. The Table also contains the results of the i.v. administration of
compound 10j (0.625 mg/kg).
24
25
26
27
Table 4. Pharmacokinetic parameters determined after a single iv application of compounds 9, 10c, 10e and 10j
28
29
78
in female C3H/HeN mice. Values were calculated using PKSolver. C , initial concentration; V , volume of
0
z
30
31
distribution in terminal phase; AUC, Area under the curve; CL , total clearance; Cmax, maximal concentration.
tot
3
3
Compd
Plasma
Urine
34
3
C (µg/mL)
0
Dose (mg/kg)
V (mL)
z
t1/2 (h)
AUC0-inf (µg x h/mL)
CLtot (mL/h)
Cmax (µg/mL)
36
3
38
39
40
41
42
4
9
40
50
50
25.2
28.3
19.5
52.8
0.33
0.4
23.5
25.3
53.1
49.4
7.1
300
4611
387
10
1
1
0c
0e
109.7
151.6
0.36
50
1.9
175.1
0.07
1
0j
0.625
0.17
218
4
45
46
47
In contrast to the fast plasma clearance of antagonists 9 and 10c (Figure 5A), the methyl
48
4
50
sulfonamide bioisostere 10e attained higher initial concentration in plasma (C ) and lower total
0
51
52
clearance (CL ). Therefore, it could be detected until 6 h post application, resulting in
tot
53
5
markedly higher plasma AUC. The observed high C of compound 10e may be attributed to a
0
5
56
57
7
2
small volume of distribution (V ) resulting from the high PPB (f ≥ 99%). In urine (Figure
z
b
58
59
5
B), the carboxylic acid 9 and the morpholinomethanone 10c displayed high levels
60
immediately following administration and a rapid concentration decrease within the first two
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 81
1
2
3
4
5
6
7
8
9
hours, reflecting the rapid elimination from plasma. Fast renal excretion as major route of
elimination can be rationalized by the physicochemical properties of the antagonists 9 and 10c,
that is, moderate PPB and log D , as well as high metabolic stability. Otherwise, the methyl
7
.4
1
1
12
13
14
15
16
sulfonamide bioisostere 10e showed sustained compound levels in urine over a period of 2 h
and subsequent slow decrease until 6 h post administration. This sustained renal excretion is a
result of the interplay of the antagonist’s elevated PPB and log D .
7
.4
17
18
19
20
2
2
23
24
25
26
27
28
29
30
31
3
3
34
35
36
Figure 5. Antagonist concentrations in (A) plasma and (B) urine after a single i.v. application of 9, 10c, and 10e
50 mg/kg).
(
37
38
39
4
41
42
4
4
45
46
47
48
49
50
51
52
53
5
In a second study, the para-cyano bioisostere 10j, characterized by a high oral absorption
potential, was administered as a single dose i.v. (0.625 mg/kg) and p.o. (1.25 mg/kg). The
plasma concentration curve upon i.v. dosing displays a steep decline within the first hour post
application, while the p.o. curve shows a prolonged period where absorption and elimination
are in equilibrium (Figure 6A). The urine concentration profiles (Figure 6B) parallel the
plasma curves obtained by the two modes of application, i.e. high plasma clearance upon i.v.
bolus injection led to high initial antagonist levels in urine and a rapid concentration decline.
By contrast, sustained plasma concentrations upon p.o. administration resulted in prolonged
urine levels.
5
56
57
58
59
60
ACS Paragon Plus Environment

Page 27 of 81
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
As a result, urine concentrations exceed the minimum level required for the anti-adhesive
7
9
effect as estimated from the in vitro cell infection model (minimal anti-adhesion
2
3
concentration , MAC 0.094 µg/mL) for more than eight hours upon oral single-dose
9
0,
1
1
administration (Figure 6B).
12
13
14
15
16
17
18
19
20
2
2
23
24
25
26
27
28
29
30
31
3
3
Figure 6. Antagonist concentrations in (A) plasma and (B) urine after a single i.v. and p.o. application of
compound 10j (i.v.: 0.625 mg/kg; p.o.: 1.25 mg/kg). MAC , minimal anti-adhesive concentration to inhibit 90%
9
0
34
35
adhesion (0.094 µg/mL).
36
37
38
39
40
41
42
4
4
45
46
47
48
49
50
51
Infection study in C3H/HeN Mice. In a preventive study, six mice were inoculated with
UTI89 following an oral application of 10j (1.25 mg/kg) 40 min prior to infection. Three hours
after inoculation, the animals were sacrificed and bladder and kidneys were removed. Organs
were homogenized and analyzed for bacterial counts. The effect of the FimH antagonist was
compared to a 8 mg/kg dose of ciprofloxacin (CIP), applied subcutaneously (s.c.) ten minutes
before infection. CIP is used as standard antibiotic therapy in humans for the treatment of
52
53
8
0
UTI. In mice, the dose of 8 mg/kg s.c. was shown to mimic the standard human dose
5
5
56
81
regarding peak levels and the AUC in serum. The median reductions in bacterial counts in
24
57
58
mice treated with 10j and CIP compared to the control group three hours after infection are
59
60
displayed in Figure 7.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 81
1
2
3
4
5
6
7
8
9
1
1
12
13
14
15
16
17
18
19
20
2
2
23
24
25
26
27
28
29
30
31
3
3
34
35
36
37
38
Figure 7. Preventive efficacy of 10j in the UTI mouse model 3 h after infection. The bars depict the median
39
40
bacterial load with the interquartile range in the different study groups. Shown are the results of the control group
41
42
(
PBS), control group formulation (5% DMSO in PBS containing 1% Tween 80), and the intervention groups with
4
4
45
46
the preventive applications of either, 1.25 mg/kg or 10 mg/kg 10j p.o. or 8 mg/kg CIP s.c. (representing the
81
murine dose equivalent to a human standard dose). DL, detection limit; CFU, colony forming units.
47
48
49
The median value in the untreated control group showed bacterial counts of 6.6 log colony
1
0
50
5
52
forming units (CFU) in the bladder and 6 log CFU in the kidneys. After oral application of
10
53
5
1
.25 mg/kg of 10j, bacterial loads in the bladder decreased by 1.78 log CFU and 1.07 log
10 10
5
56
57
58
CFU in the kidneys. The lower reduction in the kidneys is most likely due to the differing
adhesion mechanisms between bladder and kidneys (type 1 pili vs P-pili), which is not targeted
59
6
82
by 10j. With CIP (8 mg/kg s.c.) a substantial reduction in both, bladder and kidneys (median
ACS Paragon Plus Environment

Page 29 of 81
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
12
13
14
15
16
17
18
19
20
2
reductions of 2.44 log and 2.47 log , respectively) was observed. Despite the low oral dose
1
0
10
of 10j (1.25 mg/kg), the approximately 100-fold reduction of CFU in the bladder promised an
even higher effect upon dose increase to 10 mg/kg. Since the solubility of 10j for this
increased dose is too low (192 µg/mL), we used 5% DMSO and surfactant Tween 80 (1%) as
solubilizer. To effectively compare the effect of a higher dose of 10j, a control group receiving
the formulation only (5% DMSO in PBS containing 1% Tween 80, termed control group
formulation) was tested in parallel. When 10 mg/kg of 10j were applied, bacterial loads in the
bladder decreased by 2.68 log CFU/mL compared to the control group formulation, clearly
1
0
2
23
exceeding the effect of CIP with a reduction of 2.44 log CFU/mL. However, only a moderate
1
0
24
2
reduction of 1.04 log CFU was achieved in the kidneys.
10
26
27
28
29
30
SUMMARY AND CONCLUSION
31
3
3
34
35
36
37
38
39
40
41
42
4
4
45
46
47
48
49
Recently, numerous monovalent alkyl and aryl -D-mannopyranosides have been described as
potent FimH antagonists. However, most of them suffer from insufficient pharmacokinetic
properties, i.e. modest bioavailability and short duration of the therapeutic effect in the
bladder, their site of action. As a consequence, high doses at short intervals are required to
achieve anti-adhesive effects over an extended period of time. Therefore, the goal of the
present study was an appropriate optimization of the pharmacokinetic profile of biphenyl -D-
mannopyranosides while keeping their high affinity to the CRD of FimH. The starting point
50
5
52
3
9,83
was the biphenyl-carboxylate 9 where the critical carboxylate was replaced by bioisosteres.
53
5
With a series of bioisosteres, a three- to fivefold improvement of affinity was achieved
5
56
compared to 9. Although binding necessitates only partial desolvation of the carboxylate and
57
58
59
60
5
1
its bioisosteric replacements, a reduction of the enthalpy penalty for desolvation was
identified as the source of the improved affinity exhibited by the bioisosteres. Thermodynamic
ACS Paragon Plus Environment