Structure-Activity Relationship of novel phenylacetic CXCR1 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2009.06.027

We reported recently the Structure-Activity Relationship (SAR) of a class of CXCL8 allosteric modulators. They invariably share a 2-arylpropionic moiety so far considered a key structural determinant of the biological activity. We show the results of rece

Full text of DOI:10.1016/j.bmcl.2009.06.027

Bioorganic & Medicinal Chemistry Letters 19 (2009) 4026–4030
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Structure–Activity Relationship of novel phenylacetic CXCR1 inhibitors
a
a
Manolo Rocco Sablone a, Maria Candida Cesta a, Alessio Moriconi a, , Andrea Aramini , Cinzia Bizzarri ,
Claudia Di Giacinto a, Rosa Di Bitondo a, Isabelle Gloaguen a, Massimiliano Aschi b,
Marcello Crucianelli b, Riccardo Bertini a, Marcello Allegretti a
*
a Research Center, Dompé pha.r.ma s.p.a., via Campo di Pile, 67100, L’Aquila, Italy
b Department of Chemistry, Chemistry Engineering and Materials, University of L’Aquila, 67100, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
We reported recently the Structure–Activity Relationship (SAR) of a class of CXCL8 allosteric modulators.
They invariably share a 2-arylpropionic moiety so far considered a key structural determinant of the bio-
logical activity. We show the results of recent SAR studies on a novel series of phenylacetic derivatives
supported by a combined approach of mutagenesis experiments and conformational analysis. The results
suggest novel insights on the fine role of the propionic/acetic chain in the modulation of CXCL8 receptors.
Ó 2009 Elsevier Ltd. All rights reserved.
Received 27 April 2009
Revised 5 June 2009
Accepted 8 June 2009
Available online 13 June 2009
Keywords:
GPCRs
Allosteric modulators
SAR
Conformational analysis
Chemokine receptors, belonging to the family of 7TM-GPCRs
(seven-transmembrane G-protein-coupled receptors), are differen-
tially expressed by a number of immune and nonimmune cell pop-
ulations and mediate cell activities in response to soluble
chemoattractant molecules called chemokines.1–3 Chemokines
and their receptors control the traffic of leucocytes and lympho-
cytes and a disregulation of the chemokines/chemokine receptors
network has been implicated in a variety of diseases including
RA (rheumatoid arthritis), COPD (chronic obstructive pulmonary
disease), asthma, Alzheimer’s disease, melanoma and psoriasis.4–9
Among chemotactic factors CXCL8/IL-8 is a key mediator of
hPMN (human polymorphonuclear leucocyte) recruitment and
activation.10,11 CXCL8 activates CXCR1 (interleukin-8 receptor type
1) and CXCR2 (interleukin-8 receptor type 2) expressed on hPMN
surface CXCL8 and its receptors are believed to play a pivotal role
in several inflammatory diseases12,13 thus CXCL8/CXCR1-CXCR2
are considered promising pharmacological targets. To date, only a
limited number of SMW (Small Molecular Weight) CXCL8 inhibi-
tors have been described.14–16
As a result, a huge number of 2-phenylpropionic acid deriva-
tives was synthesized, including amide and acylsulfonamide deriv-
atives, leading to potent and selective CXCL8 inhibitors. Within this
class reparixin represents the clinical candidate currently under
development in phase II clinical trials for the prevention of ischae-
mia/reperfusion injury during organ transplantation.18 Specific
mechanistic studies allowed us to demonstrate that reparixin
and its analogues act as neutral allosteric inhibitors binding CXCR1
and CXCR2 in a well-defined pocket in the TM region bordered by
the helices TM 1–3, 6 and 7.
In a previous Letter17 we reported that the phenylacetic acid
diclofenac (1) shares with ketoprofen and ibuprofen the property
to inhibit CXCL8-induced hPMN chemotaxis. By contrast, subse-
quent SAR studies clearly evidenced the crucial role of the propi-
onic chain in determining the biological activity of the class that
is invariably lost in the related phenylacetic analogues.
In this work, we examined a novel class of CXCL8 inhibitors
belonging to the class of diclofenac analogues to clarify the mech-
anism of action and SAR of these atypical phenylacetic CXCL8
inhibitors. SAR studies herein reported provide further insights
into the specific determinants involved in the fine tuning of allo-
steric modulation of CXCL8 receptors.
The unexpected finding that some NSAIDs (Non-Steroidal Anti-
Inflammatory Drugs) belonging to the family of 2-phenylpropionic
acids, such as ketoprofen and ibuprofen, inhibited CXCL8-induced
hPMN chemotaxis with
a
COX(cyclooxygenase)-independent
Compounds 1 and 22–27 are commercially available whereas
the synthetic methods for the preparation of the test compounds
10–1919–22 and 2123 were previously published. Compounds 2–9
were prepared according to Scheme 1. Intermediate 1a was
obtained by N-acylation of commercial 2,6-dichlorophenylamine
mechanism,17 encouraged us towards extensive medicinal chemis-
try studies aiming at the identification of novel CXCL8 inhibitors.
* Corresponding author. Tel.: +39 0862 338212; fax: +39 0862 338215.
E-mail address: alessio.moriconi@dompe.it (A. Moriconi).
with 2-chloropropanoyl chloride followed by
a classical
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.06.027

M. R. Sablone et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4026–4030
4027
O
Cl
Cl
H
N
Compds
R
H
a, b
N
3
4
5
6
7
8
9
Cl
Cl
CH3SO2
1a
CH2CH2N(CH3)2
CH2CH2OH
(2,6-diCH3)Ph
C(CH3)3
c
H
N
R
COOH
CH2CH2N+(CH3)3 I-
O
NH
NH
d
Cl
Cl
Cl
Cl
5
9
e
3-8
2
Scheme 1. Synthesis of Compounds 2–9. Reagents and conditions: (a) 2-chloropropanoyl chloride, reflux; (b) AlCl3, 160 °C, 85%; (c) 2 N NaOH, EtOH, reflux, 98%; (d) RNH2,
i
CDI, Pr2NEt, CH2Cl2, room temperature, 45–75%; (e) CH3I, THF, room temperature, 75%.
Friedel–Crafts internal alkylation of the aromatic ring in 85% over-
all yield. The ring opening reaction in ethanolic sodium hydroxide
allowed us to obtain in almost quantitative yield the carboxylic
acid 2. It was transformed into compounds 3–8 by reaction with
the appropriate sulfonamide or amine by classical coupling proce-
dures. Treatment of 5 with iodomethane afforded the quaternary
ammonium iodide 9. Compound 20 was prepared following the
same procedure of 4, but starting from [2-(2-fluorophen-
oxy)phenyl]acetic acid 18.
As previously mentioned, the observation that several 3- and 4-
substituted phenylpropionic acids, as well as their amide deriva-
tives, inhibited the biological activity of CXCL8 whereas the corre-
sponding phenylacetic derivatives were completely inactive (IC50
say. Compounds 1 and 2 exhibited a comparable potency
(IC50 = 8 nM and 10 nM, respectively), thus suggesting a possible
correlation with the class of the phenylpropionic allosteric inhibi-
tors. This hypothesis was further supported by the observation that
identical substitutions of the carboxylic group are tolerated in the
two classes. In fact, primary amide 3, acylmethanesulfonamide 4,
x
-aminoalkylamide 5 as well as ethoxyethanolamide 6 derived
from 2 exhibited comparable activity as the parent carboxylic acid
(Table 1).
In strict analogy with reported SAR14 the introduction of bulky
aliphatic or aromatic amide groups significantly reduced the po-
tency of the inhibitors (7 and 8). Correspondingly, the conversion
of the tertiary amine of the
x-aminoalkylamide 5 in the quaternary
>10
l
M), led to the conclusion that the methyl group is crucial
ammonium salt 9 resulted in a significant loss of potency (IC50
>1000 nM). Thus, in this set of 2-arylamino substituted phenylace-
tic derivatives, the reported SAR are intriguingly superimposable to
those of phenylpropionic acids demonstrating that the methyl
group was not essential for the biological activity.
for receptor binding. Nevertheless, receptor binding studies,
molecular modelling techniques and receptor point mutagenesis,
did not provide clear structural information on specific interactions
engaged by the methyl group in the binding pocket. Since diclofe-
nac (1) is the only phenylacetic acid so far described able to inhibit
CXCL8-induced hPMN chemotaxis,17 a set of analogues was syn-
thesized (Table 1) to derive additional information on the SAR of
this class. With the aim of investigating whether a methyl group
is compatible with the biological activity, the phenylpropionic ana-
logue 2 was synthesized and tested in the hPMNs chemotaxis as-
In our previous work14 we highlighted that aliphatic or aro-
matic substituents in the 2-position strongly compromise the
affinity with the receptor pocket. To further investigate SAR, a
series of diclofenac analogues differently substituted on the sec-
ond aromatic ring was synthesized and tested in the chemotaxis
assay (Table 2). Several observations can be derived: an addi-
tional methyl group (10) on the aromatic ring is tolerated in
the binding pocket, although leading to a potency decrease
Table 1
Effect of diclofenac derivatives on CXCL8-induced hPMN chemotaxis
Table 2
Effect of diclofenac derivatives on CXCL8-induced hPMN chemotaxis
Cl
1
R
H
N
NH
R2
O
Cl
R
3
R
R1
R1
O
a
Compds
R
CXCL8 IC50 (nM)
OH
1
2
3
4
5
6
7
8
9
H
OH
OH
NH2
NHSO2CH3
NH(CH2)3N(CH3)2
NH(CH2)2OH
8
10
12
13
25
a
Compds
R1
R2
R3
CXCL8 IC50 (nM)
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
10
11
12
13
14
15
16
Cl
Cl
Cl
F
H
CH3
F
CH3
Cl
H
H
H
Cl
H
H
H
H
H
F
500
25
400
20
9
NHPh[2,6-(CH3)2]
NHC(CH3)3
600
650
>1000
>1000
>1000
>1000
H
H
NH(CH2)3N+(CH3)3 IÀ
a
a
Values are means of three or more experiments, std. dev. are <20% of the IC50
Values are means of three or more experiments, std. dev. are <20% of the IC50
values.
values.

4028
M. R. Sablone et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4026–4030
Table 3
Effect of phenylacetic derivatives on CXCL8-induced hPMN chemotaxis
R1
X
O
R2
R
R1
R2
CXCL8 IC50 (nM)
a
Compds
X
R
17
18
19
20
21
O
O
O
O
OH
OH
OH
NHSO2CH3
OH
H
F
Cl
F
H
H
H
H
Cl
>1000
13
500
50
>1000
CH2
Cl
a
Values are means of three or more experiments, std. dev. are <20% of the IC50
values.
(IC50 = 500 nM), the position of the two chlorine atoms (11) does
not significantly influence the biological activity (IC50 = 25 nM);
while the monochloro derivative 12 shows a lower potency
(IC50 = 400 nM). The marked activity of the 2-fluoro derivative
13 (IC50 = 9 nM) would suggest that a strong EWG is mandatory
for the biological activity. In agreement with this hypothesis,
the unsubstituted and the 2-methyl substituted analogue 14
and 15 did not show significant activity (IC50 >1000 nM for both).
Yet, the lack of activity of the 2,6-difluoro compound 16 (IC50
>1000 nM) would suggest the existence of a threshold for the po-
sitive effect of EW substituents.
Figure 1. Predicted binding mode of 1 in CXCR1. Compound 1 and polar residues
are shown as bold sticks. All the carbons of 1 are shown in magenta whereas both
intramolecular and intermolecular hydrogen bonds are depicted as black dashed
lines. The torsional angles found in the complex
calculation ( = +80°) are close to the conformation extracted by crystallographic
analysis of the 1/COX-2 complex ( = +65°).
(v = +70°) and by ab initio
v
v
A similar trend was observed also within the series of aryle-
thers (17–20) (Table 3), thus excluding a direct involvement of
the amino group in donor hydrogen bond inter or intramolecular
interactions.
These data support the binding model hypothesis but do not yet
provide a clarification for the different role of the methyl group
that is mandatory in the series of 3- and 4-substituted phenylpro-
pionic inhibitors26 and optional for the 2-substituted phenylacetic
acids.
Consistent with the discussed effect of EWGs, the replacement
of the heteroatom by a methylene group (21), resulted in a signif-
icant reduction of the inhibitory potency (IC50 >1000 nM).
Sustained by the evident SAR correlations, we attempted to de-
rive a hypothesis for the binding mode of 1 in CXCR1, using the
reparixin binding model as a template. Compound 1 was manually
docked in the CXCR1 binding pocket and subjected to molecular
dynamics calculations.24,25 Interestingly, the results (Fig. 1) show
that 1 binds CXCR1 in a conformation that strictly resembles the
complex COX-2/1 (pdb code: 1PXX; http://www.pdb.org). In the
derived model the carboxylic acid establishes a double ionic inter-
action with Lys992.64 and Glu2917.39 residues, respectively, as ob-
served for 2-phenylpropionic acids. The chlorinated aromatic ring
seems to be involved in a direct electrostatic interaction with the
phenolic moiety of Tyr461.39 that behaves as hydrogen bond donor.
The proposed model nicely fits with the observed SAR inasmuch as
wtCXCR1
K99ACXCR1
100
80
60
**
the OH-
p interaction is clearly favoured by EWG. In addition, 1
40
establishes hydrophobic interactions with the lipophilic side
chains of Val421.35, Ile431.36, Val1133.28, Ile2837.31 and Leu2877.35
residues. As in the complex COX-2/1, the complex is stabilized by
the intramolecular hydrogen bond between the amino and the car-
boxylic acid groups.
To test the computational hypothesis, 1 was tested in the che-
motaxis assay using L1.2 cells transiently transfected with wild-
type CXCR1 or mutated Lys99Ala CXCR1, respectively (Fig. 2).
The CXCR1 mutant did not significantly differ from wild-type
CXCR1 transfectant in terms of receptor expression levels, ligand
binding properties and chemotactic response (data not shown).
Compound 1 significantly inhibited (IC50 = 12 nM) the wild-type
CXCR1/L1.2 transfectant migration induced by CXCL8 (10 nM)
while the efficacy of 1 was completely lost in Lys99Ala CXCR1/
**
**
20
0
-10
-9
-8
-7
-6
10
10
10
10
Diclofenac [M]
10
Figure 2. Migration of L1.2 transfectants expressing wild-type CXCR1 (d) or
Lys99Ala CXCR1 (s) was induced by 10 nM CXCL8 in the presence or absence of
increasing concentrations of 1, as indicated. Results are expressed as percent of
migration considering 100% cell migration in the absence of 1 ( SD) at least in three
**
independent experiments: ( ) P <0.01 versus cell migration in the absence of 1
(Mann–Whitney U-test). Spontaneous migration was 1000 500 cells/well; CXCL8-
induced cell migration was 55,700 2100 cells/well and 53,250 2900 cells/well
for wtCXCR1 and Lys99AlaCXCR1 transfectants, respectively.
L1.2 (IC50 >10 lM).

M. R. Sablone et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4026–4030
4029
As compared to 3- and 4-substituted phenylacetic acids, this
series of diclofenac analogues is characterized by a limited confor-
mational flexibility due to the presence of bulky aromatic groups at
the 2-position. Starting from this consideration, we focused our
attention on the internal flexibility of the inhibitors. In particular,
assuming the bound inhibitor as a conformationally confined sys-
tem, an attractive approach consists in evaluating the reduction
of internal fluctuation upon binding. This approach, previously fol-
lowed in medicinal chemistry studies27,28 allows to relate, within
the assumption of a similar enthalpic behaviour, the affinity of
an inhibitor with the loss of entropy associated to the binding of
the rigid inhibitor to the receptor. Following this line of investiga-
tion, a conformational energy scanning around the C1–C2 torsional
and 24 that exhibit a free rotor behaviour, the related phenylpropi-
onic analogues 22 and 25 show a higher rotation barrier between
the accessible minima. This general trend is confirmed also for
the pair of 2-substituted derivatives 1 and 2 (Fig. 3, panel c), whose
torsional energy profiles significantly differ from those of the
phenylacetic acids 23 and 24, showing deep minima (at
v
=
80°
for 1) separated by a relatively high energy barrier (>40 KJ/mol
for 1).
This limited dataset seems to support the initial hypothesis
that, within the class of phenylacetic acids, a correlation could ex-
ist between the biological activity and the loss of entropy associ-
ated to receptor binding.
To further investigate this hypothesis, we tested compounds 26
and 27 (Table 4), methylated analogues of two phenylacetic acids
found inactive in the CXCL8-chemotaxis assay (data not shown).
The insertion of the ortho-methyl group on the heteroaromatic
ring, despite the small size, resulted in a significant increase of
the rotational barrier between the accessible minima (Fig. 3, panel
d) and, was paralleled by a marked increase of the inhibitory po-
tency (Table 4).
angle
v (Fig. 3) was carried out on a small set of active and inactive
phenylacetic and phenylpropionic derivatives (Table 4), by using
ab initio computations at the Hartree-Fock (HF) level of theory
by means of 6–31 g(d) basis set. By comparison of the energy trend
of the two pairs of compounds 22–25 (Fig. 3; panels a–b) it is note-
worthy that the introduction of the methyl group significantly in-
creases the rotational barrier around the C1–C2 bond and that the
energy barrier value is markedly influenced by the nature of the
Based on these speculations, a reasonable interpretation could
be that the methyl group of phenylpropionic acids does not con-
tribute to the receptor binding by direct favourable interactions
substituent of the aromatic ring (DEmax 3-benzoyl >4-isobutyl).
As compared to the biologically inactive phenylacetic acids 23
Figure 3. Conformational energies versus the variation of C1–C2 torsional angle
v
for the couple of compounds 22–23 (A), 24–25 (B), 1–2 (C) and 26–27 (D) shown in Table 4.
The scanning of the potential energy was performed around the torsional angle with an incremental step of 20°. Each point on each curve was obtained by mechanical
minimizations by freezing the investigated dihedral angle.

4030
M. R. Sablone et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4026–4030
Table 4
could be a key factor in the SAR of this class. The data here discussed,
even if not conclusive, put forward the importance of combining dif-
ferent experimental and theoretical approaches (molecular model-
ling, site-directed mutagenesis and specific ab initio
computational studies) for the interpretation of complex SAR
results.
CXCL8-induced hPMN chemotaxis inhibitory activity of phenylacetic and their
corresponding phenylpropionic derivatives
R1
O
R
H
O
Acknowledgements
Compds
R
R1
H
CXCL8 IC50 (nM)
a
This work was supported by the European Union FP6 (INNO-
CHEM, Grant No. LSHB-CT-2005-518167).
Cl
1
8
H
N
Supplementary data
Cl
Cl
Supplementary data (experimental and computational proce-
dures) associated with this article can be found, in the online ver-
sion, at doi:10.1016/j.bmcl.2009.06.027.
Cl
2
22
23
24
CH3
CH3
H
10
50
H
N
References and notes
1. Mackay, C. R. Nat. Immunol. 2001, 2, 95.
2. Murphy, P. M. Annu. Rev. Immunol. 1994, 12, 593.
3. Wells, T. N.; Power, C. A.; Proudfoot, A. E. Trends Pharmacol. Sci. 1998, 19, 376.
4. Gerard, C.; Rollins, B. J. Nat. Immunol. 2001, 2, 108.
5. Shadidi, K. R. BioDrugs 2004, 18, 181.
6. Owen, C. Pulm. Pharmacol. Ther. 2001, 14, 193.
7. Stellato, C.; Brummet, M. E.; Plitt, J. R.; Shahabuddin, S.; Baroody, F. M.; Liu, M.
C.; Ponath, P. D.; Beck, L. A. J. Immunol. 2001, 166, 1457.
8. Xia, M. Q.; Hyman, B. T. J. Neurovirol. 1999, 5, 32.
9. Scheibenbogen, C.; Mohler, T.; Haefele, J.; Hunstein, W.; Keilholz, U. Melanoma
Res. 1995, 5, 179.
10. Van Damme, J. The Cytokine Handbook; Academic press: New York, 1994.
11. Baggiolini, M.; Dewald, B.; Moser, B. Adv. Immunol. 1994, 55, 97.
12. Bizzarri, C.; Beccari, A. R.; Bertini, R.; Cavicchia, M. R.; Giorgini, S.; Allegretti, M.
Pharmacol. Ther. 2006, 112, 139.
13. Bizzarri, C.; Allegretti, M.; Di Bitondo, R.; Cervellera, M. N.; Colotta, F.; Bertini,
R. Curr. Med. Chem. & Anti-Inflammatory Anti-Allergy Agents 2003, 2, 67.
14. Allegretti, M.; Bertini, R.; Cesta, M. C.; Bizzarri, C.; Di Bitondo, R.; Di Cioccio, V.;
Galliera, E.; Berdini, V.; Topai, A.; Zampella, G.; Russo, V.; Di Bello, N.; Nano, G.;
Nicolini, L.; Locati, M.; Fantucci, P.; Florio, S.; Colotta, F. J. Med. Chem. 2005, 48,
4312.
15. Gonsiorek, W.; Fan, X.; Hesk, D.; Fossetta, J.; Qiu, H.; Jakway, J.; Billah, M.;
Dwyer, M.; Chao, J.; Deno, G.; Taveras, A.; Lundell, D. J.; Hipkin, R. W. J.
Pharmacol. Exp. Ther. 2007, 322, 477.
16. Busch-Petersen, J. Curr. Top. Med. Chem. 2006, 6, 1345.
17. Bizzarri, C.; Pagliei, S.; Brandolini, L.; Mascagni, P.; Caselli, G.; Transidico, P.;
Sozzani, S.; Bertini, R. Biochem. Pharm. 2001, 61, 1429.
18. Bertini, R.; Allegretti, M.; Bizzarri, C.; Moriconi, A.; Locati, M.; Zampella, G.;
Cervellera, M. N.; Di Cioccio, V.; Cesta, M. C.; Galliera, E.; Martinez, F. O.; Di
Bitondo, R.; Troiani, G.; Sabbatini, V.; D’Anniballe, G.; Anacardio, R.; Cutrin, J. C.;
Cavalieri, B.; Mainiero, F.; Strippoli, R.; Villa, P.; Di Girolamo, M.; Martin, F.;
Gentile, M.; Santoni, A.; Corda, D.; Poli, G.; Mantovani, A.; Ghezzi, P.; Colotta, F.
Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 11791.
19. Moser, P.; Sallmann, A.; Wiesenberg, I. J. Med. Chem. 1990, 33, 2358.
20. Blobaum, A. L.; Marnett, L. J. J. Biol. Chem. 2007, 282, 16379.
21. Cohen, M. P. (Exocell, Inc.) US Patent 6,552,077, 2003.
22. Atkinson, D. C.; Godfrey, K. E.; Meek, B.; Saville, J. F.; Stillings, M. R. J. Med.
Chem. 1983, 26, 1353.
C
C
>1000
>1000
O
O
H
25
26
CH3
20
50
90
O
O
N
H
Cl
R
O
N
27
H
23. Allegretti, M.; Cesta, M. C.; Bertini, R.; Bizzarri, C.; Colotta, F. (Dompé S.p.A.)
WO 03043625, 2003.
24. MacroModel,
www.schrodinger.com).
25. GAUSSIAN 03, Revision C.02; Gaussian: Wallingford, CT, 2004.
26. Moriconi, A.; Cesta, M. C.; Cervellera, M. N.; Aramini, A.; Coniglio, S.; Colagioia,
S.; Beccari, A. R.; Bizzarri, C.; Cavicchia, M. R.; Locati, M.; Galliera, E.; Di
Benedetto, P.; Vigilante, P.; Bertini, R.; Allegretti, M. J. Med. Chem. 2007, 50,
3984.
version
8.6;
Schrödinger:
Portland,
OR
(http://
a
Values are means of three or more experiments, std. dev. are <20% of the IC50
values.
but plays a determinant role in stabilizing a rotational conformer
constraining the inhibitors in a minimum close to the receptor
binding conformation.
In conclusion, the above results support the concept that the
unfavourable entropy variation associated with receptor binding
27. Franchetti, P.; Marchetti, S.; Cappellacci, L.; Javaram, H. N.; Yalowitz, J. A.;
Goldstein, B. M.; Barascut, J. L.; Dukhan, D.; Imbach, J. L.; Grifantini, M. J. Med.
Chem. 2000, 43, 1264.
28. Bianchini, G.; Aschi, M.; Cavicchio, G.; Crucianelli, M.; Preziuso, S.; Gallina, C.;
Nastari, A.; Gavuzzo, E.; Mazza, F. Bioorg. Med. Chem. 2005, 13, 4740.