N-Substituted-2′-O,3′-N-carbonimidoyl bridged macrolides: Novel anti-inflammatory macrolides without antimicrobial activity

Article abstract of DOI:10.1021/jm300356u

Macrolide antibiotics, like erythromycin, clarithromycin, and azithromycin, possess anti-inflammatory properties. These properties are considered fundamental to the efficacy of these three macrolides in the treatment of chronic inflammatory diseases like diffuse panbronchiolitis and cystic fibrosis. However, long-term treatment with macrolide antibiotics presents a considerable risk for promotion of bacterial resistance. We have examined antibacterial and anti-inflammatory effects of a novel macrolide class: N-substituted 2′-O,3′-N-carbonimidoyl bridged erythromycin-derived 14- and 15-membered macrolides. A small focused library was prepared, and compounds without antimicrobial activity, which inhibited IL-6 production, were selected. Data analysis led to a statistical model that could be used for the design of novel anti-inflammatory macrolides. The most promising compound from this library retained the anti-inflammatory activity observed with azithromycin in lipopolysaccharide-induced pulmonary neutrophilia in vivo. Importantly, this study strongly suggests that antimicrobial and anti-inflammatory activities of macrolides are independent and can be separated, which raises development plausibility of novel anti-inflammatory therapeutics.

Full text of DOI:10.1021/jm300356u

Article
pubs.acs.org/jmc
N′-Substituted-2′-O,3′-N-carbonimidoyl Bridged Macrolides: Novel
Anti-inflammatory Macrolides without Antimicrobial Activity
Martina Bosnar,*,†,‡ Goran Kragol,*,†,‡ Sanja Kostrun,†,‡ Ines Vujasinovic,†,‡ Berislav Bosnjak,†,§
̌
́
̌
Vlatka Bencetic Mihaljevic,†,‡ Zorica Marusic Istuk,†,‡ Samra Kapic,†,‡ Boska Hrvacic,†,‡ Karmen Brajsa,†,‡
́
́
̌
́
̌
́
̌
̌
́
̌
†,⊥
Branka Tavcar,†,‡ Dubravko Jelic,†,‡ Ines Glojnaric,†,‡ Donatella Verbanac,†,∥ Ognjen Culic,
̌
̌
́
́
́
Jasna Padovan,†,‡ Sulejman Alihodzic,†,‡ Vesna Erakovic Haber,†,‡ and Radan Spaventi†,‡
̌
́
́
†GlaxoSmithKline Research Centre Zagreb, Prilaz baruna Filipovica
́
29, Zagreb, Croatia
S
* Supporting Information
ABSTRACT: Macrolide antibiotics, like erythromycin, clari-
thromycin, and azithromycin, possess anti-inflammatory
properties. These properties are considered fundamental to
the efficacy of these three macrolides in the treatment of
chronic inflammatory diseases like diffuse panbronchiolitis and
cystic fibrosis. However, long-term treatment with macrolide
antibiotics presents a considerable risk for promotion of
bacterial resistance. We have examined antibacterial and anti-
inflammatory effects of a novel macrolide class: N′-substituted
2′-O,3′-N-carbonimidoyl bridged erythromycin-derived 14- and 15-membered macrolides. A small focused library was prepared,
and compounds without antimicrobial activity, which inhibited IL-6 production, were selected. Data analysis led to a statistical
model that could be used for the design of novel anti-inflammatory macrolides. The most promising compound from this library
retained the anti-inflammatory activity observed with azithromycin in lipopolysaccharide-induced pulmonary neutrophilia in vivo.
Importantly, this study strongly suggests that antimicrobial and anti-inflammatory activities of macrolides are independent and
can be separated, which raises development plausibility of novel anti-inflammatory therapeutics.
interactions.7 Nowadays, macrolides are widely used in the
INTRODUCTION
■
treatment of respiratory and urogenital tract, skin, and soft
Macrolide antibiotics (“macrolides”) are well-established class
of antimicrobial agents characterized by the presence of a highly
substituted macrocyclic lactone ring. Erythromycin, a secondary
metabolite isolated from Saccharopolyspora erythraea, was the
first macrolide to be introduced to clinical use over 50 years
ago. Afterward, several semisynthetic derivatives of erythromy-
cin, such as clarithromycin (6-O-methylerythromycin A)1 and
azithromycin (9-deoxy-9a-aza-9a-methyl-9a-homoerythromycin
A),2 were designed to broaden the antimicrobial spectrum,
reduce gastrointestinal side effects, and increase acid-stability
and bioavailability in this class of antibiotics. Telithromycin, a
representative of a new macrolide subclass called ketolides was
approved for clinical use some 10 years ago.3 This new subclass
of 14-membered macrolides is characterized by a keto group at
the C-3 position instead of the cladinose sugar.4 This change
further increased acid stability and prevented induction of
macrolide−lincosamide−streptogramin B resistance.5
Macrolides are broad-spectrum antibiotics that exhibit
antibacterial activity against aerobic Gram-positive bacteria,
certain Gram-negative bacteria, anaerobic bacteria, and intra-
cellular pathogens such as Mycoplasma, Chlamydia, and
Legionella.6 Macrolides reversibly bind to the 50S subunit of
the bacterial ribosome adjacent to the peptidyl transferase
center, thereby blocking the egress of nascent polypeptides and
consequently destabilizing the ribosome−peptidyl−tRNA
tissue infections.8
In addition to efficacy in treatment of bacterial infections,
many studies over the last 20 years have demonstrated that
certain macrolides are effective in the treatment of various
chronic inflammatory disorders of the respiratory tract.9
Introduction of erythromycin to the treatment of diffuse
panbronchiolitis (DPB) in the 1980s drastically increased 10-
year survival rate, decreased frequency of exacerbations, and
restored lung function.10 Afterward, azithromycin was success-
fully used in the treatment of cystic fibrosis (CF), which shares
a number of similarities in clinical and pathological character-
istics with DPB. In CF patients, azithromycin treatment was
shown to significantly improve lung function and reduce the
frequency of exacerbations. Consequently, certain macrolides,
such as erythromycin, clarithromycin, and azithromycin, are
now the first-line therapy for DPB and recommended for
patients with CF. Clinical efficacy of these macrolides in
diseases such as DPB and CF cannot be directly linked to their
antimicrobial activity.11 They are also being evaluated in the
therapy of chronic obstructive pulmonary disease (COPD),
chronic sinusitis, asthma, bronchiectasis, and bronchiolitis
Received: March 14, 2012
Published: June 14, 2012
© 2012 American Chemical Society
6111
dx.doi.org/10.1021/jm300356u | J. Med. Chem. 2012, 55, 6111−6123

Journal of Medicinal Chemistry
Article
a
Scheme 1. Synthesis of N′-Substituted-2′-O,3′-N-carbonimidoyl Bridged Macrolides
a
Reagents and conditions: (a) Sodium acetate trihydrate, iodine, methanol, 500 W halogen lamp irradiation, 4−5 h, or (b) N-iodosuccinimide,
acetonitrile, 1.5 h. (c) Step 1, R′-NCS, triethylamine, acetonitrile, 60 °C, 1−2 h; step 2, 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC), 60
°C, 24−72 h. (d) 11, glacial acetic acid.18 (e) Byproduct isolated from the reaction: 14, ethyl isothiocyanate, triethylamine, acetonitrile, 60 °C, 1−2
h; step 2, EDC, 60 °C, 44 h. (f) 18, benzyl or isopropyl isothiocyanate, triethylamine, acetonitrile, 60 °C, 1−2 h; step 2, EDC, 60 °C, 24−72 h. (g)
Propargyl bromide, N,N-diisopropylethylamine, DMF, 80 °C, MW, 20 min. (h) H2, Pd/C, ethanol, RT, 40 min. (i) H2, Pd/C, ethanol, 2 bar, 2 h.
obliterans.12 However, the long-term low-dose treatment of
chronic inflammatory disorders of the respiratory tract with
macrolide antibiotics promotes bacterial resistance.13 This risk
limits current use of macrolide antibiotics only to a small
number of inflammatory conditions associated with high
mortality rate despite of existing solid evidence confirming
their efficacy in other numerous inflammatory conditions. The
majority of these diseases exhibit neutrophil-dominated
inflammation and thus do not have adequate therapy coverage.
Therefore, macrolide drugs which would retain, e.g.,
azithromycin type of anti-inflammatory activity, without
antimicrobial effects, could provide therapy for a number of
chronic inflammatory diseases which are currently not
adequately treated. There are examples of 12-, 14-, and 15-
membered macrolides where immunomodulatory and anti-
bacterial activity can be separated by appropriate chemical
derivatization.14
6112
dx.doi.org/10.1021/jm300356u | J. Med. Chem. 2012, 55, 6111−6123

Journal of Medicinal Chemistry
Article
Figure 1. Screening cascade. Criteria for progression were: no antimicrobial activity (MIC >64 μg/mL), no influence on cell viability (less than 10%
inhibition at 50 μM concentration), anti-inflammatory activity in vitro (at least 60% inhibition of LPS-stimulated IL-6 production at 50 μM), no
significant potential to induce drug−drug interaction (CYP3A4 IC50 >5 μM), anti-inflammatory activity in vivo (significant inhibition of neutrophil
number in bronchoalveolar lavage fluid following intranasal LPS challenge), pharmacokinetic profile (PK) that would allow once daily oral dosing
(t1/2 >3 h, F > 20%).
Consequently, we have synthesized a focused library of
erythromycin-derived 14- and 15-membered macrolides char-
acterized by N′-substituted 2′-O,3′-N-carbonimidoyl bridged
desosamine sugar and have examined their antibacterial and
anti-inflammatory activities. Because desosamine is directly
involved in the binding of macrolides to ribosomes through a
network of hydrogen bonds and ionic interactions,15 2′-O,3′-N-
carbonimidoyl bridging should diminish the antibacterial
activity of macrolide antibiotics.16 Moreover, various sub-
stituents at the nitrogen of the 2′-O,3′-N-carbonimidoyl bridge
could further affect antibacterial activity but also fine-tune the
macrolide physicochemical properties.
antibiotics as well as pharmacokinetic data found in literature
which suggests that this concentration is above the maximal
concentration achieved after standard and chronic azithromycin
treatment. Namely, Cmax of azithromycin in plasma and blood
of CF patients following long-term treatment with azithromycin
(500 mg daily) was 0.67
0.31 and 2.01
0.74 mg/L,
respectively.20 In addition, the 2 g azithromycin extended
release formulation achieved Cmax of 0.94 and 3.2 mg/L in
serum and epithelial lining fluid, respectively.21 As effective
concentrations are substantially lower, compounds are not
tested above this breakpoint level as it is considered biologically
irrelevant. Physicochemical properties were determined for a
diverse subset from this library in order to establish a model for
the prediction of in vitro activities.
RESULTS AND DISCUSSION
■
In the course of our research focused on novel macrolide
transformations, we described the synthesis of a new macrolide
class marked by N′-substituted-2′-O,3′-N-carbonimidoyl bridged
desosamine sugar.17 Such macrolides can be synthesized in an
easy and straightforward manner in only two steps, 3′-N-
demethylation and subsequent one-pot cyclization via interme-
diary formed thiourea moiety, from commercially available
standard antibacterial macrolides such as azithromycin and
clarithromycin. Accordingly, we expanded the synthesis of 2′-
O,3′-N-carbonimidoyl bridged macrolides to cover a wider
chemical space and fine-tune physicochemical properties,
mostly by introducing a broader range of substituents at N′-
position but also by modifying aglycone rings of standard
macrolide scaffolds (Scheme 1). A small focused library
containing 60 different molecules was prepared. These
compounds were screened for antimicrobial activity against a
panel of common respiratory pathogens (Staphylococcus aureus,
Streptococcus pneumoniae, Streptococcus pyogenes, Moraxella
catarrhalis, and Haemophilus influenzae), cytotoxicity against a
human acute monocytic leukemia cell line (THP-1), and
inhibition of lipopolysaccharide (LPS)-induced interleukin-6
(IL-6) production by murine splenocytes (Figure 1).
Antimicrobially inactive (minimal inhibitory concentration
(MIC) > 64 μg/mL) compounds with anti-inflammatory
activity (at least 60% inhibition of LPS-induced IL-6 production
at 50 μM concentration), which did not influence cell viability
(less than 10% reduction of viability at 50 μM concentration),
were further progressed. Compounds with MIC values higher
than 64 μg/mL were considered antimicrobially inactive based
on microbiological susceptibility breakpoints19 for macrolide
Results from antibacterial activity screening showed that
around 30% of all prepared macrolides still exhibit rather weak
antibacterial activity while the rest had MIC values greater than
64 μg/mL (Table S1, Supporting Information). Complete
suppression of antibacterial activity was best achieved with the
clarithromycin 9a-lactam 1 scaffold because all 2′-O,3′-N-
carbonimidoyl bridged clarithromycin-9a-lactams (compounds
derived from 10) have MIC values >64 μg/mL. The
antibacterial activity of clarithromycin (2) and erythromycin
(5) derivatives can also be successfully suppressed by 2′-O,3′-N-
carbonimidoyl bridging in most (∼90%) cases. On the other
hand, suppression of antibacterial activity on 9a- or 8a-azalide
scaffolds (compounds derived from 12 and 17) was less
successful because ∼45% of such 2′-O,3′-N-carbonimidoyl
bridged compounds had some MIC values below 64 μg/mL,
including three compounds showing an MIC = 1 μg/mL
against S. pneumoniae. Similar results, ∼60% success in
suppression of antibacterial activity, were obtained for
compounds derived from 9-hydroxy scaffolds 15 and 16.
Interestingly, 2′-O,3′-N-carbonimidoyl bridging of rearranged
compounds 77−79 and 20 could not completely suppress their
antibacterial activity (some MICs = 8 μg/mL). The above
results suggest that 2-imino-1,3-oxazolidine condensation to
desosamine is an easy and efficient way of suppressing or at
least diminishing antibacterial activities of standard antibacterial
macrolides.
Less than a half of the synthesized macrolides (23
compounds) had no significant effect on viability of THP-1
cells. Of the 18 macrolides that passed both antibacterial and
6113
dx.doi.org/10.1021/jm300356u | J. Med. Chem. 2012, 55, 6111−6123

Journal of Medicinal Chemistry
Article
Figure 2. PCA analysis of synthesized compounds. A group of 28 diverse compounds were selected for physicochemical profiling (shown in
triangles).
Figure 3. Analysis of THP-1 viability in terms of ChromlogD, pos, and hetrat descriptors.
THP-1 screening, 10 compounds also showed a desired level of
IL-6 inhibition (Table S1, Supporting Information).
pos (number of positively ionizable atoms), and hetrat (ratio of
polar N, O, S vs C atoms in molecule) descriptors (Figure 3).
More polar compounds with a higher hetrat ratio, lower
ChromlogD, and lower pos tended to have weaker effect on cell
viability. As ChromlogD and pos variables are mutually
dependent, compounds with lower lipophilicity can have a
larger number of positively ionizable atoms. To develop
quantitative models that can predict influence of compounds
on THP-1 cell viability, partial least-squares (PLS) analysis was
done using all generated descriptors. A two component model
was developed. The number of descriptors was reduced by
eliminating nonsignificant ones, improving Q2 and stability of
the generated PLS models. Two chromatographic lipophilicity
measures turned out to have the largest impact on the THP-1
cell viability: ChromlogD and HSA %binding (binding to
human serum albumin) as shown in Figure 4. An increase in
In an attempt to establish a correlation between
physicochemical properties of the novel macrolide derivatives,
their cytotoxic effect and IL-6 inhibition, a group of 28 diverse
compounds, out of all compounds synthesized, were selected
for physicochemical profiling (Figure 2). Compounds were
further described applying simple properties and 2D descriptors
as described in the Experimental Section (Table S2, Supporting
Informations).
Experimentally obtained values for the physicochemical
properties, together with calculated descriptors, were used to
analyze and explain the effect of these compounds on THP-1
cell viability and inhibition of IL-6 production. Reliable
separation between compounds which do and do not influence
cell viability of THP-1 cells was achieved by using ChromlogD,
6114
dx.doi.org/10.1021/jm300356u | J. Med. Chem. 2012, 55, 6111−6123

Journal of Medicinal Chemistry
Article
compounds that decrease cell viability also produce artificial
IL-6 inhibition. We were interested in compounds that show
significant inhibition of IL-6 without effect on cell viability.
There are 12 such compounds in our compound set, with
measured physicochemical properties, which were used to build
up the model for IL-6 inhibition.
A PLS model for IL-6 inhibition based on measured and
calculated descriptors was developed. Experimental vs calcu-
lated IL-6 values and coefficient contribution of most important
descriptors are shown in Figure 6. As in the case of THP-1 cell
Figure 4. Box plot analysis of influence of (a) ChromlogD
(ChromlogD <5, and ChromlogD >5) and (b) HSA %binding data
(HSA %binding <50 and HSA %binding >50) on THP-1 cell viability
for the set of oxazolidine compounds.
lipophilicity clearly decreases THP-1 cell viability. Better
statistics were obtained with HSA %binding as a lipophilicity
measure. It was shown previously that biomimetic chromato-
graphic partition systems may prove to be better models then
standard log P/log D variants22 as it was found in this case as
well.
Figure 6. PLS model for IL-6 inhibition based on calculated and
measured descriptors.
viability, the main parameter governing IL-6 inhibition is
lipophilicity expressed as ChromlogD or HSA %binding, with
HSA giving better correlations. Additional descriptors are CHI
IAM (chromatographic hydrophobicity indices obtained on
immobilized artificial membrane) lipophilicity, number of
hetero atoms in the molecule, and number of positive charges.
CHI IAM lipophilicity and number of positive charges
contributes positively, while number of hetero atoms
contributes negatively to the IL-6 model.
A quantitative PLS model for THP-1 cell viability was
developed with four descriptors (Figure 5): HSA %binding as a
The developed models were externally validated on a set of
10 standard and subsequently synthesized compounds
structurally different from the structures in the training set.
Very good predictions were obtained for IL-6, confirming
stability and prediction power of the developed model (Figure
7). Somewhat less accurate predictions were obtained for THP-
1, with four compounds out of 10 incorrectly classified as toxic
while being nontoxic. However, only one compound had
Figure 5. PLS model for THP-1 cell viability based on calculated and
physicochemical descriptors.
measure of lipophilicity, MV as a measure of molecular size,
number of hydrogen bond acceptors (hba), and hetrat ratio.
The first two descriptors contribute positively to THP-1
cytotoxicity, while the last two describe the lipophilic character
of molecules that decrease effect on THP-1 cell viability of
studied compounds. Whereas this model exhibited significant
accuracy in prediction (R2 = 0.75 Q2 = 0.77), an analogous
model was developed using ChromlogD as a main lipophilicity
measure, with somewhat lower statistical significance (R2 = 0.52
Q2 = 0.45).
Available THP-1 data were further compared with IL-6
inhibition (Figure S1, Supporting Informations) showing that
THP-1 cell viability overlaps with IL-6 inhibition, i.e.,
Figure 7. External validation of the IL-6 model.
6115
dx.doi.org/10.1021/jm300356u | J. Med. Chem. 2012, 55, 6111−6123

Journal of Medicinal Chemistry
Article
Table 1. Combined in Vitro Screening Data for 10 Best Compounds (AZI = Azithromycin)
THP-1 reduction of viability
CYP3A4
(IC50, μM)
antibacterial activity (MIC, μg/mL)
(%)
IL-6 inhibition (%)
S.
S.
S.
M.
H.
compd aureus penumoniae
pyogenes
catarrhalis
influenzae
50 μM
50 μM 25 μM 12 μM
DEF
>10
7BQ
>10
2.9
>10
5.7
AZI
32
33
34
38
39
40
44
54
73
74
0.5
<0.125
<0.125
>64
>64
>64
>64
>64
>64
>64
>64
>64
>64
<0.125
2
0
0
0
0
9
0
0
0
6
0
0
84
73
99
86
95
67
94
86
71
74
67
74
64
96
80
90
55
81
62
44
62
48
61
55
86
70
79
45
61
52
39
55
36
7.9
0.5
7.8
1.0
1.1
1.4
0.7
0.7
1.33
>10
>10
9.8
>10
4.7
8
>10
significant error in prediction while three others are within the
error of the model (<30%). It is also important to say that
validation set was biased toward nontoxic compounds that were
overestimated in the initial model.
Screening results for this focused library and corresponding
developed PLS models for THP-1 cell viability and IL-6
inhibition will be used to design the next generation of anti-
inflammatory macrolides.
of neutrophils (Figure 9). Therefore, inhibition of these
cytokines would be considered beneficial in the suppression
of neutrophil-dominated inflammatory disorders.
Molecular mechanism underlying macrolide anti-inflamma-
tory activity has not been unambiguously elucidated yet.
However, many studies performed in the last two decades
suggest that the activity of a number of molecules belonging to
the 14- and 15-membered ring macrolide class is mediated
through inhibition of nuclear factor kappa B (NFκB), activator
protein 1 (AP-1), and extracellular signal-regulated kinase 1/2
(ERK1/2).25 Moreover, we have suggested that the inhibition
of LPS-induced neutrophil accumulation in lungs by
azithromycin is primarily caused by inhibition of IL-1β and
AP-1 in alveolar macrophages.26 Because azithromycin24 and
compound 38 have similar patterns of cytokine inhibition in
LPS-induced pulmonary neutrophilia, one can assume that the
anti-inflammatory effects of both compounds are mediated
through the same molecular mechanism, shared with other 14-
and 15-membered ring macrolides.
To further estimate suitability of compound 38 for possible
development as a drug candidate, in vivo disposition kinetics
and bioavailability for compound 38 were evaluated in Balb/c
mice and Sprague−Dawley rats. As shown in Table 2,
compound 38 showed an encouraging profile in both species.
It is characterized by a low systemic clearance (ca. 22% and
14% of liver blood flow in mice and rats, respectively), a large
volume of distribution, indicating extensive distribution into
tissues, and a prolonged half-life of ca. 17 h in mice and ca. 20 h
in rats. Following oral administration, compound 38 showed a
good bioavailability in mice (62%) and in rats (33%).
Allometric scaling to humans with these data indicated the
potential to achieve once daily oral dosing. The pharmacoki-
netic profile for compound 38 in rodents closely resembles
profiles observed with the macrolide antibiotic azithromycin in
rodents (internal data) and justifies its progression.
At this point, 10 molecules fulfilled the required criteria for
further progression toward anti-inflammatory candidates (Table
1). Compounds were screened in for their potential to inhibit
the CYP3A4 recombinant human enzyme in a fluorimetry
based high-throughput inhibition assay, in order to filter out
molecules with high risk for drug−drug interactions (DDI). On
the basis of their inhibition potential, within our organization,
we typically classify compounds as follows: very high risk IC50 <
1 μM, high risk IC50 1−5 μM, moderate risk IC50 5−10 μM,
and low risk IC50 > 10 μM for DDI. Results from the
recombinant CYP inhibition screen are used to filter out potent
inhibitors and are crosschecked later on in human liver
microsomes using multiple probes. Within this program, 5 μM
was selected as a cutoff value to filter compounds with high risk
for DDI.23 Screening of these 10 molecules against CYP3A4
inhibition revealed that only two molecules (33 and 38)
fulfilled this additional criterion (Table 1). Furthermore,
activity of the selected macrolides 33 and 38 was examined
in an in vivo model of LPS-induced pulmonary neutrophilia
and compared to standard macrolide antibiotic azithromycin
(Figure 8).
Compound 38 dose dependently reduced total cell and
neutrophil numbers in bronchoalveolar lavage fluid (Figure
8c,d). Compound 33 had lower anti-inflammatory activity in
this animal model (Figure 8a,b). Accordingly, compound 38
was selected as the most promising anti-inflammatory
compound from this novel macrolide class.
Compound 38 was further extensively studied in an in vivo
model of LPS-induced pulmonary neutrophilia. In line with
previously published results for azithromycin,24 inhibition of
LPS-induced pulmonary neutrophilia by compound 38 seems
to be mediated through inhibition of granulocyte-macrophage
colony-stimulating factor (GM-CSF), interleukin-1 β (IL-1β),
chemokine (C-X-C motif) ligand 2 (CXCL2), tumor necrosis
factor α (TNFα), IL-6, and chemokine (C−C motif) ligand 2
(CCL2) production (Figure 9). All of these cytokines are
reported to be involved in activation, recruitment, and survival
CONCLUSION
■
In this study, we have examined antibacterial and anti-
inflammatory effects of a novel macrolide class: N′-substituted
2′-O,3′-N-carbonimidoyl bridged erythromycin-derived 14- and
15-membered macrolides. A small focused library of 60
compounds was prepared and tested in a screening cascade
designed to select compounds that inhibit IL-6 production yet
be without antibacterial activity and effects on THP-1 cell
viability. Analysis of experimental results resulted in PLS
6116
dx.doi.org/10.1021/jm300356u | J. Med. Chem. 2012, 55, 6111−6123

Journal of Medicinal Chemistry
Article
Figure 8. Total cell (a, c, and e) and neutrophil (b, d, and f) number in bronchoalveolar lavage fluid (BALF) 24 h after LPS challenge. Compounds
were administered intraperitoneally 2 h prior LPS challenge at a dose of 200 mg/kg. Data are presented as means SEM of eight animals per group
Asterisk represents p < 0.05, Kruskal−Wallis test followed by Dunn’s multiple comparison test.
models that can be used for the design of next generation anti-
inflammatory macrolides. Compound 38 (2′-O,3′-N-(N′-iso-
propylcarbonimidoyl)-3′-N-demethyl-9-deoxo-9a-methyl-9a-
aza-9a-homoerythromycin A) was selected as the most
promising representative and has been extensively studied in
an in vivo model of LPS-induced pulmonary neutrophilia. In
accordance with the results for azithromycin, inhibition of LPS-
induced pulmonary neutrophilia by compound 38 seems to be
mediated through inhibition of GM-CSF, IL-1β, CXCL2,
TNFα, IL-6, and CCL2. Pharmacokinetic studies in rodents
revealed that compound 38 has good oral bioavailability and a
half-life that supports once-daily dosing in humans. In
summary, the described novel antimicrobially inactive macro-
lide derivative 38 retained the anti-inflammatory activity of
azithromycin, thereby demonstrating that these two types of
activities are independent and can be separated. Such finding
provides basis for the development of novel anti-inflammatory
agents, which will complement existing anti-inflammatory
therapeutics and significantly broaden treatment possibilities.
EXPERIMENTAL SECTION
■
Chemistry. All solvents and reagents were used as supplied, unless
noted otherwise. Mass spectra were recorded on Varian MAT 311
instrument (FAB) and Platform LCZ or LCQ Deca instruments (ESI).
HRMS (ESI) were recorded on Micromass Qtof2. NMR spectra were
recorded at 25 °C in CDCl3 and DMSO-d6 with TMS as the internal
standard on Bruker Avance DRX500 and Bruker Avance III 600
6117
dx.doi.org/10.1021/jm300356u | J. Med. Chem. 2012, 55, 6111−6123

Journal of Medicinal Chemistry
Article
Figure 9. Concentration of CXCL1 (a), CXCL2 (b), TNF-α (c), IL-6 (d), GM-CSF (e), CCL2 (f), and IL-1β (g) in lung homogenates at different
time points after LPS challenge (solid squares). Compound 38 (open circles) was administered intraperitoneally 2 h prior LPS challenge at a dose of
200 mg/kg. Data are presented as means SEM of five animals per group Asterisk represents p < 0.05, Two-way ANOVA; Bonferroni post test.
spectrometers equipped with 5 mm diameter inverse detection probe
with z-gradient accessory, as well as Bruker Avance DPX300
spectrometer using dual 1H/13C probe. For characterization of
complex organic structures, one-dimensional (1D) (1H and APT)
and two-dimensional (2D) (COSY, HMQC and HMBC) NMR
techniques were used. The purity of all prepared compounds was
determined as ≥95% using HPLC-MS method.
Detailed descriptions of compounds 36−38, 46−48, 55−58, 60−
62, and 64−67 have been given in ref 17. Detailed characterization of
the other 43 compounds can be found in Supporting Information. 3′-
6118
dx.doi.org/10.1021/jm300356u | J. Med. Chem. 2012, 55, 6111−6123

Journal of Medicinal Chemistry
Article
(CH3)2), 3.69 (5′-H), 3.64 (11-H), 3.63 (5-H), 3.61 (2′-H), 3.27 (3″-
OCH3), 3.07 (4″-H), 2.82 (3′-H), 2.80 (2-H), 2.78 (3′-NCH3), 2.70
(10-H), 2.54 (9-Ha), 2.39 (2″-Ha), 2.33 (9a-NCH3), 2.13 (9-Hb), 2.00
(8-H), 1.98 (4′-Ha), 1.96 (4-H), 1.89 (14-Ha), 1.72 (7-Ha), 1.60 (2″-
Hb), 1.47 (14-Hb), 1.42 (4′-Hb), 1.32 (6-CH3), 1.30 (5″-CH3), 1.28
(5′-CH3), 1.28 (7-Hb), 1.23 (3″-CH3), 1.20 (2-CH3), 1.16 (CN-
CH-(CH3)2), 1.11 (12-CH3), 1.08 (10-CH3), 1.06 (CN-CH-
(CH3)2), 0.97 (4-CH3), 0.94 (8-CH3), 0.89 (15-CH3) ppm. 13C
NMR (125 MHz, CDCl3) δ 178.7 (1-C), 155.9 (CN), 99.9 (1′-C),
95.3 (1″-C), 84.9 (5-C), 80.4 (2′-C), 78.8 (3-C), 77.6 (4″-C), 77.5 (13-
C), 74.3 (11-C), 73.3 (6-C), 73.2 (3″-C), 74.2 (12-C), 70.2 (5′-C),
70.1 (9-C), 65.8 (5″-C), 62.9 (3′-C), 62.4 (10-C), 49.4 (3″-OCH3),
46.8 (CN-CH-(CH3)2), 45.1 (2-C), 41.9 (7-C), 41.0 (4-C), 36.5
(9a-NCH3), 36.4 (4′-C), 34.8 (2″-C), 32.3 (3′-NCH3), 27.3 (6-CH3),
26.7 (8-C), 24.3 (CN-CH-(CH3)2), 24.2 (CN−CH-(CH3)2),
21.9 (8-CH3), 21.6 (3″-CH3), 21.1 (14-C), 20.8 (5′-CH3), 18.1 (5″-
CH3), 16.2 (12-CH3), 15.2 (2-CH3), 11.2 (15-CH3), 8.5 (4-CH3), 7.3
(10-CH3) ppm. HRMS (ES) calcd for C41 H75N3O12 (M + H+)
802.5429, found 802.5444.
Physicochemical Measurements. Chromatographic hydropho-
bicity index (CHI) log D and CHI immobilized artificial membrane
(IAM) values were determined by using fast-gradient high-perform-
ance liquid chromatographic (HPLC) method and MS as detector on
the Waters Early Candidate Profiling (ECP) 4-way MUX LC/MS
system (Waters Corp, UK). The CHI28 values were determined using
Luna C18(2) (Phenomenex, UK) HPLC columns. The CHI IAM29
values were measured using an immobilized artificial membrane
column (IAM PC DD column from Regis Analytical, West Lafayette,
IL, USA) with fast acetonitrile gradient at starting mobile phase of pH
= 7.4. CHI values were derived directly from a gradient reversed phase
chromatographic retention time by using a calibration line obtained for
standard compounds, separately for each considered property. The
CHI value approximates to the volume % organic concentration when
the compound elutes. CHI was linearly transformed into ChromlogD
by least-squares fitting experimental CHI values to calculated ClogP
values for over 20K research compounds using the following formula:
ChromlogD = 0.0857CHI-2.00. Chemically bonded human serum
albumin (HSA) and α-1-acidglycoprotein stationary phases were used
in the remaining two channels of the 4-way HPLC/MS system.
Samples were injected onto the four HPLC columns from the same
sample solution at the same time. Isopropyl alcohol gradient was used
on the protein column up to 30% using linear gradient. The run time
was 6 min including the re-equilibration of the stationary phases with
the 50 mM pH 7.4 ammonium acetate buffer. As most of the
macrolides do not contain UV chromophore, it was essential to use the
mass spectrometric detection of the peak retention times. The
obtained gradient retention times were standardized using a calibration
set of mixtures as described in the references (CHI,28 IAM,29 HSA/
AGP30).
Table 2. Pharmacokinetic Properties (CL, Clearance; Vss,
Volume of Distribution; t1/2, Half-Life; F, Bioavailability) of
Compound 38. Data are presented as means S.D. of 3
mice per timepont and 3 rats per group
CL (mL/min/kg) Vss (L/kg)
t1/2 (h)
oral F (%)
mouse
rat
19.8 1.1
12.3 1.9
23 1.8
8.6 1.3
17.3 1.5
20.3 6.2
62 10.5
33
5
N-Demethylated compounds 10−19 were prepared using described
procedures.17,27 A general procedure for 2′-O,3′-N-carboinimidoyl
formation is provided. The procedures for the synthesis of 33 and 38
as well as key intermediate 12 are described in details bellow.
3′-N-Demethyl-9-deoxo-9a-methyl-9a-aza-9a-homoerythromy-
cin A (12). A solution of 3 (10 g, 13.3 mmol) and N-iodosuccinimide
(7.5 g, 33.3 mmol) in acetonitrile (250 mL) was stirred at room
temperature for 1.5 h. Acetonitrile was evaporated and solid residue
dissolved in CH2Cl2 (300 mL), washed with saturated aqueous
Na2SO3 (5 × 70 mL) and saturated aqueous NaHCO3 (5 × 70 mL),
dried over Na2SO4, and evaporated to afford 10.8 g of yellowish solid.
The solid residue was dissolved in the mixture of water (100 mL) and
CH2Cl2 (50 mL), pH was adjusted to 4.1, and layers separated.
Extraction at pH 4.1 was repeated two times with CH2Cl2 (2 × 50
mL). To the aqueous layer, fresh CH2Cl2 (200 mL) was added, pH
was adjusted to 9.3, and then extracted with CH2Cl2 (2 × 100 mL).
The combined organic extracts at pH 9.3 were dried over Na2SO4 and
evaporated to afford title product (8.8 g, purity = 89% according to
LC-MS) as a white solid, which was used in subsequent reactions
without further purification. MS (ES+) m/z: 735.2 [MH]+.
General Procedure for 2′-O,3′-N-Carbonimidoyl bridging:17 To a
solution of 3′-N-demethyl derivatives 10−19 in acetonitrile (c = 0.02
g/mL), corresponding isothiocyanate (3 equiv) and triethylamine (6
equiv) were added. The reaction was stirred at 60 °C for 2 h, and then
EDC (4 equiv) was added. The reaction mixture was further stirred at
60 °C for 24−48 h. Completion of the reaction was checked by LC-
MS. Solvent was evaporated and the residue purified on the
Flashmaster personal solid-phase extraction techniques (eluents 1.5%
MeOH/CH2Cl2−MeOH/CH2Cl2/NH4OH (4.5:90:0.25)) to afford
chromatographically homogeneous fractions of title products.
2′-O,3′-N-(N′-Ethylcarbonimidoyl)-3′-N-demethyl-9-deoxo-9a-
methyl-9a-aza-9a-homoerythromycin A (33). According to the
general procedure for sequential one-pot synthesis, reaction of 12
(0.9 g, 1.23 mmol) and ethyl isothiocyanate afforded 33 (0.31 g, 32%)
as a white foam. 1H NMR (500 MHz, CDCl3) δ 5.07 (1″-H), 4.91 (1′-
H), 4.68 (13-H), 4.22 (3-H), 4.05 (5″-H), 3.70 (5′-H), 3.66 (11-H),
3.66 (5-H), 3.55 (2′-H), 3.36 (CN-CH2-CH3), 3.27 (3″-OCH3),
3.07 (4″-H), 2.77 (3′-H), 2.80 (2-H), 2.72 (10-H), 2.70 (3′-NCH3),
2.53 (9-Ha), 2.35 (2″-Ha), 2.35 (9a-NCH3), 2.09 (9-Hb), 1.97 (8-H),
1.97 (4-H), 1.92 (4′-Ha), 1.88 (14-Ha), 1.74 (7-Ha), 1.61 (2″-Hb), 1.51
(14-Hb), 1.42 (4′-Hb), 1.33 (6-CH3), 1.31 (5″-CH3), 1.30 (5′-CH3),
1.26 (7-Hb), 1.26 (3″-CH3), 1.22 (2-CH3), 1.13 (CN-CH2-
CH3)1.13 (10-CH3), 1.10 (12-CH3), 1.00 (4-CH3), 0.93 (8-CH3),
0.89 (15-H) ppm. 13C NMR (125 MHz, CDCl3) δ 178.5 (1-C), 155.9
(CN), 99.9 (1′-C), 95.6 (1″-C), 84.8 (5-C), 79.9 (2′-C), 79.0 (3-C),
77.9 (4″-C), 77.5 (13-C), 74.5 (11-C), 74.2 (12-C), 73.3 (3″-C), 73.2
(6-C), 70.1 (5′-C), 70.1 (9-C), 65.7 (5″-C), 62.8 (3′-C), 62.1 (10-C),
49.3 (3″-OCH3), 44.9 (2-C), 41.9 (7-C), 40.7 (4-C), 40.6 (CN-
CH2-CH3), 36.4 (9a-NCH3), 36.4 (4′-C), 35.3 (2″-C), 32.1 (3′-
NCH3), 27.1 (6-CH3), 26.6 (8-C), 21.9 (8-CH3), 21.5 (3″-CH3), 20.9
(14-C), 20.7 (5′-CH3), 18.0 (5″-CH3), 16.6 (CN-CH2-CH3)16.1
(12-CH3), 15.3 (2-CH3), 10.1 (15-C), 8.5 (4-CH3), 7.3 (10-CH3)
ppm. HRMS (ES) calcd for C40 H73N3O12 (M + H+) 788.5277, found
788.5301.
Statistical Analysis. Descriptive statistics and graphical analysis
were done by SpotFire DecisionSite software, 8.2.1 (2006, Spotfire,
Inc., US, http://spotfire.tibco.com). Data were analyzed by principal
component analysis (PCA) and partial least-squares (PLS) techniques
implemented within SIMCA-P software 11.0 (Umetrics, Umea,
̊
Sweden). All descriptors were scaled to unit variance. PLS analysis
was done using all generated descriptors. Two component models
were developed. The number of descriptors was reduced by
sequentially eliminating nonsignificant ones, improving Q2 and
stability of the generated PLS models. Elimination of descriptors
was based on the descriptor coefficients and their importance in the
projection (VIP). VIP is defined as a sum over all model dimensions of
the variable contributions (VIN). For a given PLS dimension, a,
2
(VIN)ak is equal to the squared PLS weight (wak)2 of that term,
multiplied by the explained sum-of-squares of that PLS dimension.
Descriptors with large VIP, larger than 1, are the most relevant for
explaining Y. Models were validated by random permutation of
dependent variables and by leave-four-out cross-validation used to
calculate uncertainty measures (standard errors and confidence
intervals) of scores, loadings, PLS-regression coefficients, predicted
Y-values, and important variables.
2′-O,3′-N-(N′-Isopropylcarbonimidoyl)-3′-N-demethyl-9-deoxo-
9a-methyl-9a-aza-9a-homoerythromycin A (38). According to the
general procedure for sequential one-pot synthesis, reaction of 12
(0.47 g, 0.64 mmol) and isopropyl isothiocyanate afforded 38 (0.29 g,
1
57%) as a white foam. H NMR (500 MHz, CDCl3) δ 5.09 (1″-H),
4.92 (1′-H), 4.68 (13-H), 4.19 (3-H), 4.05 (5″-H), 3.81 (CN-CH-
6119
dx.doi.org/10.1021/jm300356u | J. Med. Chem. 2012, 55, 6111−6123

Journal of Medicinal Chemistry
Article
All methods were employed with default settings within
corresponding software. Biological properties of macrolides were
analyzed and modeled in terms of various experimentally determined
physicochemical and calculated descriptors. A number of two-
dimensional (2D) structural descriptors was calculated, like counts
of H-bond donor (hbd) and acceptor (hba) atoms, count of positively
charged atoms (pos), ratio of number of polar atoms (O,N,S), and
number of carbon atoms (hetrat), atom-based E-state descriptors31 as
well as Abraham descriptors (solute excess molar refractivity (R2),
dipolarity/polarizability (π), summation hydrogen bond acidity (α)
and basicity (β), and the McGowan characteristic volume (Vx).32
Because of macrolide complex structure, calculation of 3D descriptors
is time-consuming and not applicable for a larger number of
compounds. Available force fields are not able to reliably reproduce
conformations of macrocyclic ring and flexible substituents. All
employed descriptors were either measured or calculated by the in-
house developed approaches.
Biology. Materials: Chemicals, Antibodies and Drugs. LPS from
Escherichia coli serotype 0111:B4 was obtained from Sigma Chemical
Co. (St Louis, MO, US). Luminex kits were purchased from R&D
Systems (Minneapolis, MN, US). Azithromycin dihydrate was from
PLIVA Inc. (Zagreb, Croatia). All other reagents, if not indicated
otherwise, were from Sigma Chemical Co. (St Louis, MO, US).
Antimicrobial Activity. Whole-cell antimicrobial activity of
compounds against Staphylococcus aureus (ATCC13709), Streptococcus
pneumoniae (ATCC49619), Streptococcus pyogenes (ATCC700294),
Moraxella catarrhalis (ATCC23246), and Haemophilus influenzae
(ATCC49247) was determined by broth microdilution test using
the Clinical and Laboratory Standards (CLSI) recommended
procedure (Document M7-A6A7, Methods for Dilution Susceptibility
Tests for Bacteria that Grow Aerobically). The minimum inhibitory
concentration (MIC) is determined as the lowest concentration of
compound that inhibited visible growth.
%inhibition = [1 − (concentration of IL‐6 in sample
− concentration of IL‐6 in negative control)
/(concentration of IL‐6 in positive control
− concentration of IL‐6 in negative control)]
× 100
The positive control refers to LPS-stimulated samples that were not
preincubated with the compounds. The negative control refers to
unstimulated and untreated samples.
Animals. Studies were performed on 10-week-old male BALB/cJ
mice and 7-week-old CD Sprague−Dawley rats (both from Charles
River, Lyon, France). Mice and rats were kept on wire mesh floors
with irradiated maize granulate bedding (Scobis Due, Mucedola,
Settimo Milanese, Italy) and maintained under standard laboratory
conditions (temperature 23−24 °C, relative humidity 60 5%, 15 air
changes per h, artificial lighting with circadian cycle of 12 h). Pelleted
food (VRF-1, Altromin, Charles River, Isaszag, Hungary) and tap
water were provided ad libitum.
All procedures on animals were approved by the ethics committee
of GlaxoSmithKline Research Centre Zagreb Limited and performed
in accordance with the European Economic Community Council
Directive 86/609.
LPS-Induced Pulmonary Neutrophilia. Experimental pulmonary
neutrophilia was induced as described earlier.34 Briefly, mice (eight
animals per treatment group in the experiments aimed at
determination of cell counts in BALF and five animals per treatment
group in the experiment where cytokine concentration in lung
homogenates was determined), under light anesthesia, were instilled
intranasally with 2 μg of LPS from E. coli/60 μL PBS. Vehicle,
azithromycin, and compounds 33 and 38 were administered
intraperitoneally 2 h before intranasal challenge with LPS at a dose
of 200 mg/kg (body weight). We have previously demonstrated that
600 mg/kg po is an effective dose for inhibition of inflammation in this
animal model.24 On the basis of the oral bioavailability of 35−55%
(internal data), the estimated effective dose of azithromycin after ip
administration would be 200−300 mg/kg. The dosing regimen was
chosen as the lowest effective dose according to the results of
preliminary studies. For administration, macrolides were first dissolved
in dimethlysulfoxide (DMSO) and then diluted with 0.5% (w/v)
methylcellulose (final concentration of DMSO was 5% (v/v)).
Solutions obtained were applied in a volume of 20 mL/kg (body
weight).
Cell Viability. THP-1 cells (TIB-202, ATCC) were cultured in
RPMI 1640 medium supplemented with 10% fetal bovine serum at 37
°C in the presence of test compounds at concentration of 50 μM for
24 h. Afterward, MTS [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxyme-
thoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium] (Promega, US), a
detection reagent, was added and cells were incubated for an
additional 0.5−2 h. The amount of MTS-formazan produced was
determined using a spectrophotometer at 490 nm.33
Percentage of inhibition of cell viability was calculated using the
following formula:
Bronchoalveolar Lavage and Determination of Total and
Relative Cell Number in Bronchoalveolar Lavage Fluid (BALF).
The animals were euthanized by an intraperitoneal overdose of
thiopental (Inresa Arzneimittel GmbH, Freiburg, Germany) 24 h after
LPS application. After preparation and cannulation of tracheas,
bronchoalveolar lavage was performed with phosphate buffered saline
(PBS) in a total volume of 1 mL (0.4, 0.3, and 0.3 mL). The
bronchoalveolar lavage samples were centrifuged (4 °C, 100g, 5 min)
and the pellet of cells resuspended in an equal volume of fresh PBS
and used for total and differential cell counts. Total number of cells in
BALF was counted with a hematological analyzer (Sysmex SF 3000,
Sysmex Co., Kobe, Japan). Percentages of neutrophils were
determined by morphological examination of at least 200 cells on
smears prepared by cytocentrifugation (Cytospin-3, Thermo Fisher
Scientific Inc., Pittsburgh, PA, US) and stained with Kwik-Diff staining
set (Thermo Fisher Scientific Inc., Pittsburgh, PA, USA). Number of
neutrophils (and macrophages) in BALF was calculated for each
sample according to the formula:
%inhibition of cell viability
= (OD490treated cells/OD490untreated cells) × 100
LPS-Induced IL-6 Production by Murine Splenocytes. After
cervical dislocation, mouse spleens were removed by using sterile
dissection tools. Spleens were transferred to a prewetted cell strainer in
a 50 mL sterile conical tube, and cell suspension was made by gentle
puddle. Cells were centrifuged (20 min, 300g) and resuspended in 2
mL of sterile phosphate buffered saline (PBS) (Sigma Chemical Co.,
USA). Red blood cells were lysed by addition of 3 mL of sterile water
and occasionally gentle shaking for 1 min. Afterward, the tube was
filled to 40 mL with DMEM medium and centrifuged (20 min, 300g).
Cells were resuspended in DMEM supplemented with 1% FBS and
seeded in a 24-well plate, 106 cells per mL medium. Cells were
preincubated with the test compounds for 2 h at 37 °C in an
atmosphere of 5% CO2 and 90% humidity. Afterward, cells were
stimulated with 1 μg/mL lipopolysaccharide (LPS, E. coli 0111:B4,
Sigma Chemical Co., USA) and incubated overnight. Concentration of
IL-6 was determined in cell supernatants by sandwich ELISA using
capture and detection antibodies (R&D Systems, USA) according to
the manufacturer’s recommendations.
number of neutrophils
= total number of cells × (neutrophil percentage/100%)
Preparation of Lung Homogenates for Determination of
Cytokines. In separate groups of mice, nonlavaged lungs were
collected immediately prior (0 h) or at various time points after
LPS application for determination of the concentrations of cytokines.
Inhibition (as percentage) was calculated using the following
formula:
6120
dx.doi.org/10.1021/jm300356u | J. Med. Chem. 2012, 55, 6111−6123

Journal of Medicinal Chemistry
Article
Lungs were homogenized on ice in PBS with protease inhibitors (1
μg/mL leupeptin, 2 μg/mL aprotinin, 1 μg/mL pepstatin, and 17 μg/
mL phenylmethylsulfonyl fluoride); 4 mL of PBS with protease
inhibitors were added per gram of lung tissue. Homogenates were
centrifuged (4 °C, 2500g, 15 min) and stored at −80 °C until analysis.
Determination of Protein Concentration. Protein concentration in
lung homogenates was determined by BCA Protein Assay (Thermo
Fisher Scientific Inc., Waltham, MA, US) according to the
manufacturer’s recommendation.
Measurement of Inflammatory Mediators in Lungs. Samples were
analyzed using xMAP technology, which enables simultaneous
measurement of multiple biomarkers. Concentrations of GM-CSF,
IL-1β, IL-6, CCL2 (JE), CXCL1 (KC), CXCL2 (MIP-2), and TNF-α
were determined using Fluorokine MAP multiplex kit (R&D Systems,
Minneapolis, MN, US) according to the manufacturer’s protocol.
Briefly, 50 μL of samples were incubated with antibody-coated
microparticles for 3 h at room temperature. Afterward, washed beads
were incubated with biotinylated detection antibody cocktail for 1 h at
room temperature, washed, and incubated for 30 min with
streptavidin−phycoerythrin conjugate. After the final wash, the
microparticles were resuspended in buffer and analyzed with the
Luminex 200TM (Luminex, Austin, TX, US) and STarStation
software v2.3 (Applied Cytometry Systems, Sacramento, CA, US)
using a five-parameter-logistic-curve fitting.
profiles were generated and noncompartmental PK analysis
(WinNonlin 4.2, Pharsight) used to generate estimates of half-life,
clearance, volume of distribution, and oral bioavailability.
ASSOCIATED CONTENT
■
S
* Supporting Information
In vitro data (antibacterial activity, THP-1 viability, IL-6
inhibition) for all N′-substituted-2′-O,3′-N-carbonimidoyl
bridged macrolides prepared as well as detailed characterization
of 2′-O-3′-N-carbonimidoyl bridged macrolides not reported
previously. This material is available free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*Phone: +385 1 8886319 (M.B.); +385 1 8886357 (G.K.). Fax:
+385 1 8886443. E-mail: martina.bosnar@glpg.com (M.B.);
goran.kragol@glpg.com (G.K.).
Present Addresses
‡Galapagos Research Centre, Prilaz baruna Filipovica
́
29,
Zagreb, Croatia
§Medical University of Vienna, Department of Dermatology,
Concentration of cytokines in lung homogenates was further
normalized to protein concentration in the samples and expressed as
pg of analyte per mg of protein.
Wahringer Gurtel 18−20, Vienna, Austria
̈
̈
∥University of Zagreb School of Medicine, Center for
Statistical Analysis. All values are presented as means SEM. To
define statistically significant differences in cell numbers among
vehicle-treated and macrolide-treated mice 24 h after LPS challenge,
the data were subjected to Kruskal−Wallis test followed by Dunn’s
multiple comparison test using GraphPad Prism version 5.00 for
Windows (GraphPad Software, San Diego, CA, US). To define
statistically significant differences in cytokine concentrations among
vehicle-treated and macrolide-treated mice at different time points
after LPS challenge, the data were subjected to two-way ANOVA
followed by Bonferroni post test using GraphPad Prism version 5.00
for Windows (GraphPad Software, San Diego, CA, US). The level of
significance was set at p < 0.05 in all cases.
In Vitro CYP3A4 Inhibition. Test compounds were incubated using
a concentration range from 0.1 to 10 μM with recombinant human
CYP3A4 (Cypex Bactosomes, CYPEX, Dundee, UK). The compounds
were preincubated (n = 2) at 37 °C for 5 or 10 min with the CYP 3A4
isoform in the presence of 50 mM phosphate buffer (pH 7.4) and two
fluorescent probes (diethoxyfluorescein (DEF) and 7-benzyloxyquino-
line (7-BQ)) in a 96-well plate. The incubations were started by
addition of a cofactor solution (7.65 mg glucose-6-phosphate, 1.7 mg
NADP, 6 units glucose-6-phosphate dehydrogenase per mL of 2%
sodium bicarbonate) prepared freshly the day of the experiment. The
plates (black polypropelene 96-well) were incubated at 37 °C for 10
min, and fluorescent readings were taken on a Tecan Infinite F500
plate reader (excitation wavelength: DEF 485 nm, 7BQ 405 nm;
emission wavelength DEF 535 nm, 7BQ 535 nm; gain DEF 47 nm,
7BQ 50 nm) over a 10 min time-course. The fluorescence data were
then used to calculate the IC50 values for each compound.
Pharmacokinetics. Pharmacokinetic profiles were determined in
male Balb/c mice (n = 3/time point) and CD Sprague−Dawley rats (n
= 3) following single oral and single intravenous administration. In the
mouse, doses of 5 and 25 mg/kg were used of iv and po, respectively.
In the rat, doses of 2 and 10 mg/kg were used for iv and po,
respectively. For oral dosing, the compound was formulated as a
solution in DMSO−PEG 200−PBS (10:20:70) and for intravenous
dosing as a solution in DMSO/acetate buffer (0.1 M)/PBS (4:48:48).
Profiles in the mouse were obtained by taking terminal bleeding over a
range of time points up to 30 h postdose (three animals/time point).
Profiles in the rat (3 animals/group) were obtained by serial blood
sampling over a range of time points up to 30 h postdose. Blood
samples were diluted with water (1:1), prepared by protein
precipitation, and subjected to quantitative analysis by LCMS/MS
using compound-specific mass transitions. Drug concentration−time
̌
Translational and Clinical Research, Salata 2, Zagreb, Croatia
⊥Medvedgradska 70, Zagreb, Croatia
Notes
The authors declare no competing financial interest.
ABBREVIATIONS USED
■
IL-6, interleukin-6; THP-1, human acute monocytic leukemia
cell line; pos, number of positively ionizable atoms; hetrat, ratio
of polar N, O, S vs C atoms in molecule; HSA %binding,
binding to human serum albumin; CHI IAM, chromatographic
hydrophobicity indices obtained on immobilized artificial
membrane; MV, measure of molecular size; hba, number of
hydrogen bond acceptors; drug−drug interactions; GM-CSF,
granulocyte-macrophage colony-stimulating factor; IL-1β, in-
terleukin-1β; CXCL2, chemokine (C-X-C motif) ligand 2;
TNFα, tumor necrosis factor α; CCL2, chemokine (C−C
motif) ligand 2; BALF, bronchoalveolar lavage fluid
REFERENCES
■
(1) Morimoto, S.; Takahashi, Y.; Watanabe, Y.; Omura, S. Chemical
modification of erythromycins. I. Synthesis and antibacterial activity of
6-O-methylerythromycins A. J. Antibiot. (Tokyo) 1984, 37, 187−189.
(2) Schoenfeld, W.; Mutak, S. Azithromycin and novel azalides. In
Macrolide Antibiotics; Schoenfeld, W., Kirst, H. A., Eds.; Birkhauser
Verlag: Basel, Switzerland, 2002; pp 97−140.
(3) Johnson, A. P. Telithromycin. Aventis Pharma. Curr. Opin. Invest.
Drugs 2001, 2, 1691−1701.
(4) Denis, A.; Agouridas, C.; Auger, J. M.; Benedetti, Y.; Bonnefoy,
̀
A.; Bretin, F.; Chantot, J. F.; Dussarat, A.; Fromentin, C.; D’Ambrieres,
S. G.; Lachaud, S.; Laurin, P.; Le Martret, O.; Loyau, V.; Tessot, N.;
Pejac, J. M.; Perron, S. Synthesis and antibacterial activity of HMR
3647 a new ketolide highly potent against erythromycin-resistant and
susceptible pathogens. Bioorg. Med. Chem. Lett. 1999, 9, 3075−3080.
(5) Ackermann, G.; Rodloff, A. C. Drugs of the 21st century:
telithromycin (HMR 3647), the first ketolide. J. Antimicrob. Chemother.
2003, 51, 497−511.
(6) Zuckerman, J. M.; Qamar, F.; Bono, B. R. Macrolides, ketolides,
and glycylcyclines: azithromycin, clarithromycin, telithromycin,
tigecycline. Infect. Dis. Clin. North Am. 2009, 23, 997−1026.
6121
dx.doi.org/10.1021/jm300356u | J. Med. Chem. 2012, 55, 6111−6123

Journal of Medicinal Chemistry
Article
(7) Hansen, J. L.; Ippolito, J. A.; Ban, N.; Nissen, P.; Moore, P. B.;
Steitz, T. A. The structures of four macrolide antibiotics bound to the
large ribosomal subunit. Mol. Cell 2002, 10, 117−128.
(8) (a) Iacoviello, V. R.; Zinner, S. H. Macrolides: A Clinical
Overview. In Macrolide Antibiotics; Schoenfeld, W., Kirst, H. A., Eds.;
Birkhauser Verlag: Basel, Switzerland, 2002; pp 15−24; (b) Bryskier,
A.; Butzler, J. P. Macrolides, In Antibiotic and Chemotherapy: Anti-
infective Agents and Their Use in Therapy; Finch, R. G., Greenwood, D.,
Norrby, S. R., Whitley, R. J., Eds.; Churchill Livingstone: Edinburgh,
2003; pp 310−327.
(9) (a) Zarogoulidis, P.; Papanas, N; Kioumis, I.; Chatzaki, E.;
Maltezos, E.; Zarogoulidis, K. Macrolides: from in vitro anti-
inflammatory and immunomodulatory properties to clinical practice
in respiratory diseases. Eur. J. Clin. Pharmacol. 2012, 68, 479−503.
(b) Kanoh, S.; Rubin, B. K. Mechanisms of action and clinical
application of macrolides as immunomodulatory medications. Clin.
Microbiol. Rev. 2010, 23, 590−615.
activity of acid degradation products of 6-O-methylerythromycins A. J.
Antibiot. (Tokyo) 1990, 43, 570−573.
(19) (a) Morrissey, I.; Robbins, M.; Viljoen, L.; Brown, D. F.
Antimicrobial susceptibility of community-acquired respiratory tract
pathogens in the UK during 2002/3 determined locally and centrally
by BSAC methods. J. Antimicrob. Chemother. 2005, 55, 200−208.
(b) Schmalreck, A. F.; Kottmann, I.; Reiser, A.; Ruffer, U.; Schlenk, R.;
Vanca, E.; Wildfeuer, A. Susceptibility testing of macrolide and
lincosamide antibiotics according to DIN guidelines. Deutsches
Institut fur Normung. J. Antimicrob. Chemother. 1997, 40, 179−187.
̈
(20) Wilms, E. B.; Touw, D. J.; Heijerman, H. G. Pharmacokinetics
and sputum penetration of azithromycin during once weekly dosing in
cystic fibrosis patients. J. Cyst. Fibrosis 2008, 7, 79−84.
(21) Lucchi, M.; Damle, B.; Fang, A.; de Caprariis, P. J.; Mussi, A.;
Sanchez, S. P.; Pasqualetti, G.; Del Tacca, M. Pharmacokinetics of
azithromycin in serum, bronchial washings, alveolar macrophages and
lung tissue following a single oral dose of extended or immediate
release formulations of azithromycin. J. Antimicrob. Chemother. 2008,
61, 884−891.
(10) Kudoh, S.; Azuma, A.; Yamamoto, M.; Izumi, T.; Ando, M.
Improvement of survival in patients with diffuse panbronchiolitis
treated with low-dose erythromycin. Am. J. Respir. Crit. Care Med.
1998, 157, 1829−1832.
́
(22) Valko, K. Application of high-performance liquid chromatog-
raphy based measurements of lipophilicity to model biological
distribution. J. Chromatogr., A 2004, 1037, 299−310. K. Valko, K.
Measurements of lipophilicity and acid/base character using HPLC
methods . In Pharmaceutical Profiling in Drug Discovery for Lead
Selection; Borchardt, R. T., Kerns, E. H., Lipinski, C. A., Thakker, D. R.,
Wang, B., Eds.; AAPS: Arlington VA, 2004; pp 127−182.
(11) Murphy, D. M.; Forrest, I. A.; Curran, D.; Ward, C. Macrolide
antibiotics and the airway: antibiotic or non-antibiotic effects? Expert
Opin. Invest. Drugs 2010, 19, 401−414.
(12) (a) Crosbie, P. A. J.; Woodhead, M. A. Long-term macrolide
therapy in chronic inflammatory airway diseases. Eur. Respir. J. 2009,
̌
33, 171−181. (b) Culic,
́
O.; Erakovic,
́
V.; Parnham, M. J. Anti-
(23) Krippendorff, B. F.; Lienau, P.; Reichel, A.; Huisinga, W.
Optimizing classification of drug−drug interaction potential for
CYP450 isoenzyme inhibition assays in early drug discovery. J. Biomol.
Screening 2007, 12, 92−99.
inflammatory effects of macrolide antibiotics. Eur. J. Pharmacol. 2001,
429, 209−229. (c) Giamarellos-Bourboulis, E. J. Macrolides beyond
the conventional antimicrobials: a class of potent immunomodulators.
Int. J. Antimicrob. Agents 2008, 31, 12−20.
̌
(24) Bosnar, M.; Bosnjak, B.; Cuzic, S.; Hrvacic, B.; Marjanovic, N.;
̌
́
̌
́
́
̌
̌
́ ́ ́
Glojnaric, I.; Culic, O.; Parnham, M. J.; Erakovic Haber, V
(13) Phaff, S. J.; Tiddens, H. A.; Verbrugh, H. A.; Ott, A. Macrolide
resistance of Staphylococcus aureus and Haemophilus species associated
with long-term azithromycin use in cystic fibrosis. J. Antimicrob.
Chemother. 2006, 57, 741−746.
Azithromycin and clarithromycin inhibit lipopolysaccharide-induced
murine pulmonary neutrophilia mainly through effects on macro-
phage-derived granulocyte-macrophage colony-stimulating factor and
interleukin-1beta. J. Pharmacol. Exp. Ther. 2009, 331, 104−113.
(25) (a) Altenburg, J.; de Graaff, C. S.; van der Werf, T. S.; Boersma,
W. G. Immunomodulatory effects of macrolide antibiotics, part 1:
̌
(14) (a) Bauer, J.; Vine, M.; Coric,
́
I.; Bosnar, M.; Pasa
̌ ́
lic, I.; Turkalj,
̌
G.; Lazarevski, G.; Culic,
́
O.; Kragol, G. Impact of stereochemistry on
the biological activity of novel oleandomycin derivatives. Bioorg. Med.
Chem. 2012, 20, 2274−2281. (b) Mencarelli, A.; Distrutti, E.; Renga,
B.; Cipriani, S.; Palladino, G.; Booth, C.; Tudor, G.; Guse, J. H.; Hahn,
U.; Burnet, M.; Fiorucci, S. Development of non-antibiotic macrolide
that corrects inflammation-driven immune dysfunction in models of
inflammatory bowel diseases and arthritis. Eur. J. Pharmacol. 2011, 665,
29−39. (c) Sugawara, A.; Sueki, A.; Hirose, T.; Nagai, K.; Gouda, H.;
Hirono, S.; Shima, H.; Akagawa, K. S.; Omura, S.; Sunazuka, T. Novel
12-membered non-antibiotic macrolides from erythromycin A: EM900
series as novel leads for anti-inflammatory and/or immunomodulatory
agents. Bioorg. Med. Chem. Lett. 2011, 21, 3373−3376. (d) Li, Y. J.;
Azuma, A.; Usuki, J.; Abe, S.; Matsuda, K.; Sunazuka, T.; Shimizu, T.;
Hirata, Y.; Inagaki, H.; Kawada, T.; Takahashi, S.; Kudoh, S.; Omura,
S. EM703 improves bleomycin-induced pulmonary fibrosis in mice by
the inhibition of TGF-beta signaling in lung fibroblasts. Respir. Res.
2006, 7, 16.
́
biological mechanisms. Respiration 2011, 81, 67−74. (b) Lopez-Boado,
Y. S.; Rubin, B. K. Macrolides as immunomodulatory medications for
the therapy of chronic lung diseases. Curr. Opin. Pharmacol. 2008, 8,
286−291. (c) Shinkai, M.; Henke, M. O.; Rubin, B. K. Macrolide
antibiotics as immunomodulatory medications: proposed mechanisms
of action. Pharmacol. Ther. 2008, 117, 393−405.
̌
(26) Bosnar, M.; Cuzic, S.; Bosnjak, B.; Nujic, K.; Ergovic, G.;
́
̌
́
́
̌
Marjanovic, N.; Pasalic, I.; Hrvacic, B.; Polancec, D.; Glojnaric, I.;
Erakovic Haber, V. Azithromycin inhibits macrophage interleukin-1β
́
̌
́
̌
́
̌
́
́
production through inhibition of activator protein-1 in lipopolysac-
charide-induced murine pulmonary neutrophilia. Int. Immunopharma-
col. 2011, 11, 424−434.
(27) Kragol, G.; Marusi
̌ ́ ̌ ́ ̌ ́
c-Istuk, Z.; Hutinec, A.; Bukvic-Krajacic, M.;
Kujundzic, N. Int. Patent WO 2009/130189, 2009.
́
̌
(28) Valko, K.; Bevan, C.; Reynolds, D. Chromatographic hydro-
phobicity index by fast-gradient RP-HPLC: a high-throughput
alternative to logP/logD. Anal. Chem. 1997, 69, 2022−2029.
(29) Valko, K.; Du, C. M.; Bevan, C. D.; Reynolds, D. P.; Abraham,
M. H. Rapid-gradient HPLC method for measuring drug interactions
with immobilized artificial membrane: comparison with other
lipophilicity measures. J. Pharm. Sci. 2000, 89, 1085−1096.
(30) Valko, K.; Nunhuck, S.; Bevan, C.; Abraham, M. H.; Reynolds,
D. P. Fast gradient HPLC method to determine compounds binding
to human serum albumin. Relationships with octanol/water and
immobilized artificial membrane lipophilicity. J. Pharm. Sci. 2003, 92,
2236−2248.
(15) Schluenzen, F.; Zarivach, R.; Harms, J.; Bashan, A.; Tocilj, A.;
Albrecht, R.; Yonath, A.; Franceschi, F. Structural basis for the
interaction of antibiotics with the peptidyl transferase centre in
eubacteria. Nature 2001, 413, 814−821.
̌
(16) Palej Jakopovic,
́
I.; Bukvic
́
Krajaci
I.; Alihodzic,
̌
c,
́
M.; Matanovic
́
Skugor, M.;
̌
̌
Stimac, V.; Pesi
̌
c,
́
D.; Vujasinovic,
́
́
S.; Cipci
̌
c
́
Paljetak, H.;
̌
Kragol, G. Novel desosamine-modified 14- and 15-membered
macrolides without antibacterial activity. Bioorg. Med. Chem. Lett.
2012, 22, 3527−3530.
(17) Vujasinovic,
́
̌
c
́
Ist
̌
uk, Z.; Kapic,
́
́ ̌ ́
I.; Marusi S.; Bukvic Krajacic, M.;
̌
I.; Matkovic-Calogovic, D.; Kragol, G. Novel
́ ́
Hutinec, A.; Đilovic,
́
tandem reaction for the synthesis of N′-substituted-2-imino-1,3-
oxazolidines from vicinal (sec- or tert-)amino alcohols of desosamine.
Eur. J. Org. Chem. 2011, 2507−2511.
(18) Morimoto, S.; Misawa, Y.; Asaka, T.; Kondoh, H.; Watanabe, Y.
Chemical modification of erythromycins VI: structure and antibacterial
(31) Hall, L. H.; Kier, L. B.; Brown, B. B. Molecular Similarity Based
on Novel Atom Type Electrotopological State Indices. J. Chem. Inf.
Comput. Sci. 1995, 35, 1074−1080.
(32) (a) Abraham, M. H.; Chadha, H. S.; Martins, F.; Mitchell, R. C.;
Bradbury, M. W.; Gratton, J. A. Hydrogen Bonding, Part 46. A Review
6122
dx.doi.org/10.1021/jm300356u | J. Med. Chem. 2012, 55, 6111−6123

Journal of Medicinal Chemistry
Article
of the Correlation and Prediction of Transport Propertiesby an LFER
Method: Physicochemical Properties, Brain Penetration and Skin
Permeability. Pestic. Sci. 1999, 55, 78−88. (b) Platts, J. A.; Abraham,
M. H.; Butina, D.; Hersey, A. Estimation of Molecular Linear Free
Energy Relation Descriptors Using a Group Contribution Approach. J.
Chem. Inf. Comput. Sci. 1999, 39, 835−845. (c) Zhao, Y. H.; Abraham,
M. H.; Hersey, A.; Luscombe, C. N. Quantitative relationship between
rat intestinal absorption and Abraham descriptors. Eur. J. Med. Chem.
2003, 38, 939−947.
(33) Mosmann, T. Rapid colorimetric assay for cellular growth and
survival: application to proliferation and cytotoxicity assays. J. Immunol.
Methods 1983, 65, 55−63.
(34) Ivetic
Marjanovic,
Erakovic, V. Anti-inflammatory activity of azithromycin attenuates the
́
Tkalce
N.; Ferenci
̌ ́ ̌ ̌ ́
vic, V; Bosnjak, B.; Hrvaci B.; Bosnar, M.;
c,
̌ ̌
̌ ́ ́
c, Z.; Situm, K.; Culic, O.; Parnham, M. J.;
́
́
effects of lipopolysaccharide administration in mice. Eur. J. Pharmacol.
2006, 539, 131−138.
6123
dx.doi.org/10.1021/jm300356u | J. Med. Chem. 2012, 55, 6111−6123