Rigidity versus Flexibility: Is This an Issue in σ1 Receptor Ligand Affinity and Activity?

Article abstract of DOI:10.1021/acs.jmedchem.6b00585

Stereoisomeric 2,5-diazabicyclo[2.2.2]octanes 14 and 15 were prepared in a chiral-pool synthesis starting from (S)- or (R)-aspartate. The key step in the synthesis was a Dieckmann-analogous cyclization of (dioxopiperazinyl)acetates 8, which involved trapping of the intermediate hemiketal anion with Me₃SiCl. The σ₁ affinity was tested using membrane preparations from animal (Guinea pig) and human origin. The binding of bicyclic compounds was analyzed by molecular dynamics simulations based on a 3D homology model of the σ₁ receptor. The good correlation between K_i values observed in the σ₁ assays and calculated free binding energy, coupled with the identification of four crucial ligand/receptor interactions, allowed the formulation of structure-affinity relationships. In an in vitro antitumor assay with seven human tumor cell lines, the bicyclic compounds inhibited selectively the growth of the cell line A427, which is due to induction of apoptosis. In this assay, the compounds behave like the known σ₁ receptor antagonist haloperidol.

Full text of DOI:10.1021/acs.jmedchem.6b00585

Subscriber access provided by University of Sussex Library
Article
Rigidity versus flexibility: is this an issue in #1
(sigma-1) receptor ligand affinity and activity?
Frauke Weber, Stefanie Brune, Frederik Börgel, Katharina Korpis, Carsten Lange, Patrick J
Bednarski, ERIK LAURINI, Maurizio Fermeglia, Sabrina Pricl, Dirk Schepmann, and Bernhard Wünsch
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b00585 • Publication Date (Web): 09 May 2016
Downloaded from http://pubs.acs.org on May 11, 2016
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 48
Journal of Medicinal Chemistry
1
1
2
3
4
Rigidity versus flexibility: is this an issue in σ1 (sigma-1) receptor ligand affinity
5
6
and activity?
7
8
9
Frauke Weber,a Stefanie Brune,a Frederik Börgel,a Katharina Korpis,b Carsten
Lange,b Patrick J. Bednarski,b Erik Laurini,c Maurizio Fermeglia,c Sabrina Pricl,c,d Dirk
Schepmann,a Bernhard Wünscha,e*
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a Institute of Pharmaceutical and Medicinal Chemistry, University of Münster,
Corrensstr. 48, Dꢀ48149 Münster, Germany
Tel.: +49ꢀ251ꢀ83ꢀ33311; Fax: +49ꢀ251ꢀ83ꢀ32144;wuensch@uniꢀmuenster.de
b Institute of Pharmacy, Department of Pharmaceutical and Medicinal Chemistry,
University of Greifswald, FriedrichꢀLudwigꢀJahnꢀStraße 17, 17487 Greifswald,
Germany
c Molecular Simulations Engineering (MOSE) Laboratory, Department of Engineering
and Architecture (DEA), University of Trieste, Via Valerio 6, 34127 Trieste, Italy
d National Interuniversity Consortium for Material Science and Technology (INSTM),
Research Unit MOSEꢀDEA, University of Trieste, Via Valerio 6, 32127 Trieste, Italy
e
CellsꢀinꢀMotion Cluster of Excellence (EXC 1003 – CiM), University Münster,
Germany
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 48
2
1
2
Abstract
3
4
5
6
7
8
A set of stereoisomeric 2,5ꢀdiazabicyclo[2.2.2]octanes 14 and 15 was prepared in a
chiralꢀpool synthesis starting from (S)ꢀ or (R)ꢀaspartate. The key step in the synthesis
was a Dieckmannꢀanalogous cyclization of (dioxopiperazinyl)acetates 8, which
involved trapping of the intermediate hemiketal anion with Me3SiCl. The σ1 affinity
was tested using membrane preparations from animal (guinea pig) and human origin.
The binding of bicyclic compounds was analyzed by molecular dynamics simulations
based on a 3D homology model of the σ1 receptor. The good correlation between Ki
values observed in the σ1 assays and calculated free binding energy, coupled with
the identification of four crucial ligand/receptor interactions allowed the formulation of
structure affinity relationships. In an in vitro antitumor assay with seven human tumor
cell lines, the bicyclic compounds inhibited selectively the growth of the cell line
A427, which is due to induction of apoptosis. In this assay, the compounds behave
like the known σ1 receptor antagonist haloperidol.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Keywords
σ1 Ligands, conformational restriction, Dieckmann analogous cyclization, structure
affinity relationships, tumor cell lines, cytotoxic activity, 3D homology model,
molecular dynamics, docking, ligandꢀreceptor interactions
Introduction
After some misclassification as opioid receptors, σ receptors have now been shown
to represent a receptor class on their own. To date, two subtypes are known, termed
σ1 and σ2 receptor. These subtypes can be differentiated by their molecular weight,
ACS Paragon Plus Environment

Page 3 of 48
Journal of Medicinal Chemistry
3
1
2
tissue distribution, and ligand binding profiles. A particular feature is the different
3
4
5
6
interaction of σ receptor subtypes with dextrorotatory benzomorphans.1,2
7
8
9
After cloning the σ1 receptor from various tissues of animal origin including guinea pig
liver, mouse brain, rat brain and rat kidney,3ꢀ6 the σ1 receptor was also cloned from
the human chorioncarcinoma cell line.7 The identity of σ1 receptors cloned from
different species is around 93%. The σ1 receptor protein encoded by the human gene
consists of 223 amino acids and has a molecular weight of 25.3 kDa. A similarity of
the σ1 receptor protein with other mammalian proteins could not be found, but a 30%
identity and 67% similarity with the yeast enzyme sterol ꢀ8/7ꢀisomerase was
detected.8
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
High density of the σ1 receptor was found in the central nervous system, but also in
peripheral tissues, e.g. heart,9 kidney, and liver.10 Moreover, the σ1 receptor was
identified in endocrine organs,11 immune competent blood cells12 and very
importantly in proliferating tumor cells.13 The σ1 receptor is a membrane bound
protein localized predominantly in the plasma membrane, the membrane of the
endoplasmic reticulum associated with mitochondria (mitochondria associated
membrane) and around the nucleus (perinuclear region of ER).14,15 It has been
reported that the σ1 receptor functions as chaperone interacting with different
neurotransmitter receptors and ion channels, but the exact signal transduction
pathway has not been identified so far.16,17
The σ1 receptor plays an important role in various neurological disorders, including
depression, psychosis, Alzheimer’s disease, and alcohol/drug dependence.18
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 48
4
1
2
3
4
5
6
7
8
9
Furthermore, the antinociceptive system can be modulated by σ1 receptors, i.e. σ1
receptor agonists such as (+)ꢀpentazocine are able to potentiate the analgesic
potential of opioid analgesics.19 Moreover, selective σ1 receptor antagonists, e.g.
S1RA, are able to reduce neuropathic pain.20,21
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In 1990 overexpression of σ receptors in brain tumors was reported.22 Then, high σ1
receptor expression in human breast cancer cell lines and later, in small cell lung and
prostate cancer cell lines was shown by immunocytochemical, immunohistochemical,
and realꢀtimeꢀPCR studies as well Western blotting with a σ1 receptor specific
antibody.23,24 These experiments led to the conclusion that the expression of σ1
receptors in various human tumor cell lines is significantly increased compared with
the σ1 receptor expression level of the corresponding nonꢀtumor cells.25,26 In addition
to the high expression level of σ1 receptors in human tumor cells, it was shown that
they are involved in apoptosis (programmed cell death) and σ1 receptor antagonists
were able to induce caspaseꢀdependent cell death.27 Therefore, selective targeting of
σ1 receptors represents a promising strategy for the therapy of cancer either alone or
as adjuvants in chemotherapy by inducing apoptosis and ultimately cell death. In fact,
treatment of tumor cells with various σ1 ligands, e.g. the σ1 antagonist haloperidol,
led to both cytostatic and cytotoxic effects, although the molecular mechanisms
underlying cell growth inhibition have not yet been clarified.24,28 In addition to blocking
σ1 receptors, activation of σ2 receptors, which are highly expressed in rapidly
proliferating tumor cells, also induced apoptotic processes.26,28ꢀ31
Substantial efforts have been spent in recent years in the design, synthesis and
evaluation of potent and selective σ1 ligands. Many of these wellꢀestablished σ1
ligands contain
a
piperazine ring.32ꢀ35 Monocyclic piperazines
1
with
a
ACS Paragon Plus Environment

Page 5 of 48
Journal of Medicinal Chemistry
5
1
2
3
4
5
6
7
8
9
conformationally flexible 3ꢀhydroxypropyl side chain display moderate σ1 affinity.34
(Figure 1) Conformational restriction of flexible ligands is a general strategy in drug
design to increase both binding affinity and selectivity for a particular target.36 As a
result of conformational restriction, the ligand loss of entropy during binding is
reduced and, hence, its free binding energy is increased.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1: Development of ethanoꢀbridged piperazines 4 from the ωꢀhydroxyalkyl
substituted piperazines 1 and 3 and the propanoꢀbridged piperazines 2.
To investigate the influence of conformational restriction on σ1 receptor affinity and
cytotoxicity bridged piperazines 2 were designed by connecting the flexible 3ꢀ
hydroxypropyl side chain of piperazines 1 with the piperazine ring. Receptor binding
studies showed higher σ1 affinity for the bridged piperazines 2 compared with the
monocyclic piperazines 1. For example, a Ki value of 188 nM was found for the
flexible (hydroxypropyl)piperazine 1a bearing pꢀmethoxybenzyl (PMB) and benzyl
(Bn) moieties (R1 = PMB, R2 = Bn) at the Nꢀatoms.34 After construction of the
hydroxypropano bridge of 2 with appropriate configuration and the same
substituents, the σ1 affinity increased 30ꢀfold (2a (R1 = PMB, R2 = Bn, (1R,2R,5S)ꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 48
6
1
2
3
4
5
6
7
8
9
configuration): Ki = 6.5 nM).37 A similar relationship between the structure and the
inhibition of tumor cell growth was observed: the cytotoxic effect against the human
small cell lung cancer (SCLC) A427 cell line of the bridged piperazines 2 (e.g. 2a:
54% inhibition at a concentration of 20 ꢁM) was higher than the cytotoxic effect of the
monocyclic piperazines 1 (e.g. 1a: 23% inhibition at a concentration of 20 ꢁM)34,37 on
the same cell line.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Recently we have shown that the σ1 affinity of (2ꢀhydroxyethyl)piperazines 3 was
higher than that of their 3ꢀhydroxypropyl homologs 1, e.g. the (2ꢀ
hydroxyethyl)piperazine 3a (R1 = PMB, R2 = Bn, Ki = 20 nM) had a 9ꢀfold higher σ1
affinity than the (3ꢀhydroxypropyl)piperazine 1a (Ki = 188 nM) bearing the same
substituents at the Nꢀatoms.34
These observations prompted us to synthesize and evaluate the biological activity of
the bicyclic compounds 4, which are derived from the bicyclic compounds 2 by
removal of one methylene moiety of the propano bridge, and from the 2ꢀ
hydroxyethylꢀsubstituted piperazines 3 by connecting the flexible hydroxyethyl side
chain with the piperazine ring. On condition that conformational restriction should
lead to higher σ1 affinity and tumor cell growth inhibition, the designed 2,5ꢀ
diazabicyclo[2.2.2]octanes 4 were expected to show improved biological activities.
The results of this study were rationalized at the molecular level by atomistic
molecular dynamics simulations of the interactions between ligands 4 and the
recently developed 3D homology model of the σ1 receptor.38,39 These studies should
lead to a deep understanding of the ligand ꢀ σ1 receptor interactions and, moreover,
the contribution of the particular structural elements to the overall interactions.
ACS Paragon Plus Environment

Page 7 of 48
Journal of Medicinal Chemistry
7
1
2
Synthesis
3
4
5
6
7
8
The synthesis of 2,5ꢀdiazabicyclo[2.2.2]octanes of type 4 was planned by bridging
piperazinediones 8ꢀ10 with an appropriate side chain containing two carbon atoms.
The dioxopiperazines 8aꢀd with an acetate side chain were synthesized in a sixꢀstep
reaction sequence starting from (S)ꢀaspartate (5) as described in literature.40 In brief,
the diester 6.HCl was reductively alkylated with different aldehydes and subsequently
acylated to afford chloroacetamide 7. The dioxopiperazines 8aꢀd were obtained by a
Domino reaction (SN2 reaction followed by intramolecular aminolysis) of 7 with
different primary amines. The differently substituted dioxopiperazines 8aꢀd served as
starting material for the exploration of different bridging strategies to obtain
diazabicyclo[2.2.2]octanes (Scheme 1).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
As demonstrated in preliminary investigations, the Dieckmann analogous cyclization
of piperazinylacetates of type 8 provided less than 10% of the desired bicyclic
products 12. Therefore, alternative synthetic routes for the installation of the ethano
bridge were investigated (reaction steps (h) and (i) in Scheme 1).
At first, aldehydes 11a,b should be used as starting material, since the higher
carbonyl activity of aldehydes 11 compared with esters 8 should give higher yields in
the envisaged intramolecular aldol reaction. However, the direct reduction of the
ester 8b with DIBAL in toluene41 did not lead to the aldehyde 11b. Therefore, a twoꢀ
step conversion of the ester 8b into the aldehyde 11b comprising a reduction and
oxidation step was investigated. The selective reduction of the ester moiety of 8b
with LiBH4 afforded the primary alcohol 10b in 41% yield. However, subsequent
oxidation of the primary alcohol 10b with DessꢀMartin periodinane42 gave only very
low yields of aldehyde 11b. Finally, high yields of the aldehyde 11b were obtained by
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 48
8
1
2
3
transformation of the ester 8b into the Weinreb amide 9b43 and its subsequent
4
5
6
reduction with LiAlH4.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
compounds 7ꢀ15
R1
R2
Bn
a
b
c
d
CH2C6H11
Me
CH2C6H11
CH2C6H11
Bn
1ꢀnaphthylmethyl
4ꢀbiphenylylmethyl
Scheme 1: Reagents and reaction conditions: (a) (H3C)3SiCl, H3COH, rt, 16 h;40 (b)
1. R2ꢀCH=O, NEt3, CH2Cl2, rt, 16 h; 2. NaBH4, H3COH, 0 °C, 40 min; 3. ClCH2COCl,
NEt3, CH2Cl2, rt, 2.5 h;40 (c) R1ꢀNH2, NEt3, CH3CN, rt, 16 h – 3 d;40 (d)
.
HN(OCH3)CH3 HCl, Al(CH3)3, CH2Cl2, rt, 5 h; (e) LiAlH4, THF, ꢀ78 °C, 16 h; (f) LiBH4,
THF, ꢀ30 °C, 16 h;21 (g) NaHMDS, THF, ꢀ78 °C, 40 min, then (H3C)3SiCl, ꢀ78 °C, 1 h,
then rt, 2 h; (h) LiHMDS, THF, ꢀ78 °C, 16 h; (i) 1. LiHMDS, THF, ꢀ78 °C, 16 h; 2.
LiAlH4, THF, reflux, 16 h; (j) 0.5 M HCl, THF, rt, 16 h; (k) LiAlH4, THF, reflux, 16 h.
The enantiomers of ent-7 – entꢀ15 were prepared in the same manner.
Reaction of the aldehyde 11b with LiHMDS in THF at ꢀ78 °C induced the
intramolecular aldol reaction affording a bicyclic product. Since the purification of the
ACS Paragon Plus Environment

Page 9 of 48
Journal of Medicinal Chemistry
9
1
2
cyclization product turned out to be difficult, the product was directly reduced with
3
4
5
6
1
LiAlH4 to provide the diastereomeric alcohols 14b and 15b. Although the H NMR
7
8
9
spectra showed the desired signals, the yields and the purity of the products were not
sufficient for further investigations.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
During the synthesis of the aldehyde 11a the Weinreb amide 9a had been
synthesized. Weinreb amides can form stable chelates with metal cations after
addition of nucleophiles.44 Thus, after deprotonation of bislactam 9a with LiHMDS at
ꢀ78 °C, a stable Li+ꢀchelate was expected to form by intramolecular aldol reaction.
Hydrolysis of the Li+ꢀchelate should then afford the bicyclic ketone 13a. MS and NMR
spectra confirmed the formation of 13a. However, the yield of 13a was below 5% and
could not be increased although numerous variations of the reaction conditions (type
and amount of base, temperature, reaction time) were investigated.
As a consequence of these results, the Dieckmann analogous cyclization45 of esters
8 (conditions (g) in Scheme 1) was investigated in detail. For this purpose, the ester
8b was treated with LiHMDS at ꢀ78 °C and the anion of the intermediate hemiketal
was trapped after 10 min with (CH3)3SiCl to obtain the mixed methyl silyl ketal 12b in
3% yield. This variation of the Dieckmann condensation (trapping of the hemiketal
anion) allows the formation of small bicyclic systems, which cannot form stabilized
anions of βꢀdicarbonyl compounds at the end of the synthesis due to Bredt`s rule.46,47
Herein, the first cyclization product (i.e. the anion of the hemiketal) was trapped by
(CH3)3SiCl after deprotonation of dioxopiperazine 8b with LiHMDS. Due to the low
yield of the mixed methyl silyl ketal 12b with recovery of large amounts of the educt
8b, this transformation was carefully optimized. In order to improve the yield of 12b,
different counter ions of the base and different time intervals for deprotonation and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 48
10
1
2
trapping with (CH3)3SiCl were evaluated. Systematic variations of the reaction
conditions resulted in an improved yield of 12b of 34%. In particular, the use of
NaHMDS as base, an interval of 40 min for the deprotonation step, and a modified
work up procedure (adsorption of the crude product on silica gel instead of dissolving
the residue before purification by flash chromatography) represent the key features
for achieving this yield. A previous Xꢀray crystal structure analysis45 revealed that the
Dieckmann analogous cyclization provided (7R)ꢀconfigured products 12 with high
diastereoselectivity. Since the transformation of all analogs provided predominantly
one diastereomer with similar signals in the NMR spectra, the (7R)ꢀconfiguration can
be transferred to all mixed methyl silyl ketals 12. The configuration of the chiral
center in 4ꢀposition is defined by the configuration of the starting material (S)ꢀ
aspartate (5) leading to (1S,4S,7R)ꢀconfiguration of the mixed methyl silyl ketals 12.
Hydrolysis with 0.5 M HCl in THF led to the bicyclic ketone 13b, which was reduced
by LiALH4 to yield the diastereomeric bicyclic alcohols 14b and 15b.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The same reaction sequence was used for the synthesis of 14a,c,d and 15a,c,d
starting from esters 8a,c,d. The reaction conditions for the crucial Dieckmann
analogous cyclization of the esters 8a,c,d had to be optimized for each compound
individually. The yields were 26%, 13%, and 22% for 12a, 12c and 12d, respectively.
In order to compare the σ1 and σ2 affinities of enantiomeric alcohols the (R)ꢀ
configured dioxopiperazines ent-8a, ent-8c and ent-8d were prepared from (R)ꢀ
aspartate (ent-5) and transformed into the bicyclic alcohols entꢀ14a, entꢀ15a, entꢀ
14c, entꢀ15c, and entꢀ14d, entꢀ15d. Thus, all four possible stereoisomeric bicyclic
alcohols with a cyclohexylmethyl residue at 2ꢀposition and different arylmethyl
ACS Paragon Plus Environment

Page 11 of 48
Journal of Medicinal Chemistry
11
1
2
residues at 5ꢀposition (series a, c, and d) were available for pharmacological
3
4
5
6
evaluation.
7
8
9
In order to determine the enantiomeric purity a chiral HPLC method was developed
to analyze the stereoisomeric benzyl substituted derivatives 14a, entꢀ14a, 15a, and
entꢀ15a. Approximately 10 % of the enantiomers were found in the samples resulting
from base catalyzed partial racemization during the bridging reaction of
piperazinedione 8. However, the contamination with small amounts of the enantiomer
does not affect the biological activity of the compounds.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2: Reagents and reaction conditions: (a) (H3C)3SiCl, H3COH, rt, 16 h;33 (b)
1. PhꢀCH=O, NEt3, CH2Cl2, rt, 16 h; 2. NaBH4, H3COH, 0 °C, 40 min; 3. ClCH2COCl,
NEt3, CH2Cl2, rt, 2.5 h;37 (c) C6H11CH2ꢀNH2, NEt3, CH3CN, rt, 16 h; (d) NaHMDS,
THF, ꢀ78 °C, 40 min, then (H3C)3SiCl, ꢀ78 °C, 1 h, then rt, 2 h; (e) 0.5 M HCl, THF, rt,
16 h; (f) LiAlH4, THF, reflux, 16 h.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 48
12
1
2
Since the pharmacological properties of the diazabicyclo[2.2.2]octanes 14/15 should
be compared with those of the homologous diazabicyclo[3.2.2]nonanes, the
diastereomeric alcohols 22a and 23a were prepared. (Scheme 2) Starting from (S)ꢀ
glutamate (16) the dioxopiperazine 19 was obtained by esterification (17),33
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
benzylation,
chloroacetylation
(18)37
and,
finally,
cyclization
with
cyclohexylmethylamine. Deprotonation of 19 with NaHMDS at ꢀ78 °C and trapping of
the intermediate hemiketal anion after 40 min with (CH3)3SiCl provided the mixed
methyl silyl ketal 20 in 60% yield. This result shows clearly that the moderate yields
obtained during the cyclization of the smaller homologs 8 with an acetate side chain
are due to the shorter bridge increasing the strain of the system. Hydrolysis of the
mixed ketal 20 with diluted HCl led to the bicyclic ketone 21, which was reduced with
LiAlH4 to obtain the diastereomeric alcohols 22a and 23a in 22% and 42% yields,
respectively.
Pharmacological evaluation
Receptor binding studies
The σ affinities of compounds 14/15 and 22/23 were determined in competition
experiments with the appropriate radioligands. All compounds were tested against σ1
and σ2 receptors of animal origin obtained from guinea pig (gp) brain (σ1) and rat liver
(σ2), respectively. Additionally, the interaction of the ligands with human σ1 receptors
was analyzed using membrane preparations obtained from the peripheral blood
human myeloma cell line RPMI 8226.48 These experiments were performed to
investigate the correlation between ligand interactions with human and guinea pig σ1
receptors. [3H]ꢀ(+)ꢀPentazocine served as radioligand for both σ1 assays and [3H]ꢀ
DTG as radioligand in the σ2 assay.49ꢀ51 Compounds with high affinity were tested in
ACS Paragon Plus Environment

Page 13 of 48
Journal of Medicinal Chemistry
13
1
2
triplicate. For compounds with low σ affinity, only the inhibition of the radioligand
binding at a test compound concentration of 1.0 ꢁM is reported.
3
4
5
6
7
8
9
The results of the receptor binding studies of the new compounds are shown in Table
1. The σ1 and σ2 affinity data of various reference ligands are also listed for
comparison. The values in Table 1 demonstrate that Nꢀmethyl substituted bridged
piperazines 14b and 15b do not interact significantly with σ1 and σ2 receptors. This
result correlates nicely with the low affinity observed for (hydroxyethyl)piperazines
3aꢀd (Figure 1), which do not react with σ1 and σ2 receptors when R1 is a small
methyl residue (e.g. 3b).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In the guinea pig assay the σ1 affinity of bicyclic compounds 14 and 15 bearing a 2ꢀ
cyclohexylmethyl substituent is generally in the low nanomolar range, only entꢀ14a
and 15d reveal Ki values higher than 20 nM. Whilst the stereoisomeric bicyclic 5ꢀ
benzyl derivatives 14a, 15a, entꢀ14a, and entꢀ15a show similar σ1 affinity as the
corresponding (hydroxyethyl)piperazine 3a, the σ1 affinity of the bicyclic 5ꢀ
naphthylmethyl (cꢀseries, exception entꢀ15c) and biphenylylmethyl derivatives (dꢀ
series) display slightly reduced σ1 affinity compared to their (hydroxyethyl)piperazine
analogs 3c and 3d.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 48
14
1
2
Table 1 σ1 and σ2 receptor affinity of bicyclic piperazines
3
4
5
6
7
8
9
σ2
(rat)b)
Ki ± SEM
[nM]
σ1
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
σ1
(hum)c)
Ki ± SEM
[nM]
compd.
R1
R2
n
(gp)a)
Ki ± SEM [nM]
3a40
3b40
3c40
3d40
CH2C6H11
CH3
Bn
ꢀ
ꢀ
ꢀ
4.2 ± 1.1
28%d)
116e)
21 ± 4.0
n.d.
Bn
27%d)
CH2C6H11
CH2C6H11
1ꢀNaphꢀCH2
4ꢀPhꢀPhꢀCH2
1.9 ± 0.6
3.5 ± 0.5
26 ± 12
73 ± 45
29 ± 10
34 ± 8.0
14a
15a
CH2C6H11
CH2C6H11
CH2C6H11
CH2C6H11
Me
Bn
Bn
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
4.8 ± 0.7
6.9 ± 1.6
23 ± 13
36 ± 9.0
60 ± 26e)
197 ± 18
501 ± 21
4%d)
3.2 ± 0.4
2.4 ± 0.2
2.8 ± 1.0
1.6 ± 0.4
23%d)
entꢀ14a
entꢀ15a
14b
Bn
Bn
5.7 ± 2.6
13%d)
Bn
15b
Me
Bn
0%d)
7%d)
n.d.
14c
CH2C6H11
CH2C6H11
CH2C6H11
CH2C6H11
CH2C6H11
CH2C6H11
CH2C6H11
CH2C6H11
CH2C6H11
CH2C6H11
1ꢀNaphꢀCH2
1ꢀNaphꢀCH2
1ꢀNaphꢀCH2
1ꢀNaphꢀCH2
4ꢀPhꢀPhꢀCH2
4ꢀPhꢀPhꢀCH2
4ꢀPhꢀPhꢀCH2
4ꢀPhꢀPhꢀCH2
Bn
8.0 ± 2.0
7.1 ± 1.8
14 ± 4.0
0.50 ± 0.1e)
8.7 ± 1.2
23 ± 6.0
11 ± 2.0
11 ± 2.0
6.0 ± 0.2
1.6 ± 0.1
5.4 ± 0.5
6.6 ± 0.9
71 ± 8.0
51 ± 16
157 ± 21
40 ±15
116 ± 33
20 ± 7.0
334 ± 18
202 ± 52
593 ± 53
65 ± 7.0
284 ± 72
ꢀ
13 ± 5.0
7.2 ± 3.9
38 ± 3.0
6.0 ± 2.0
27 ± 9.0
73 ±6.0
27 ± 5.0
24 ± 6.0
6.4 ± 0.9
2.2 ± 1.1
36 ± 5.0
40 ± 5.0
208 ± 26
15c
entꢀ14c
entꢀ15c
14d
15d
entꢀ14d
entꢀ15d
22a
23a
Bn
(+)ꢀpentazocine
Haloperidol
diꢀoꢀtolylguanidine
78 ± 2.0
58 ± 18
a) gp: guinea pig brain; b) rat liver; c) RPMI 8226 cell line; d) Inhibition of radioligand
binding at 1 ꢁM concentration of test compound; e) n = 4; n.d.: not determined.
Ki values represent mean values of three independent experiments (n = 3).
ACS Paragon Plus Environment

Page 15 of 48
Journal of Medicinal Chemistry
15
1
2
The similar σ1 receptor affinities of the four stereoisomeric benzyl substituted
derivatives 14a, entꢀ14a, 15a and entꢀ15a indicate that the configuration has a
negligible effect on the interaction with σ1 receptors. Replacement of the benzyl
residue (aꢀseries) by the voluminous biphenylylmethyl moiety (dꢀseries) results in a
slight reduction of σ1 affinity as shown for the stereoisomers 14d, entꢀ14d, 15d and
entꢀ15d. As observed for the benzyl derivatives (aꢀseries) the stereochemistry of the
biphenylylmethyl derivatives (dꢀseries) does not influence the σ1 affinity,
considerably.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Expansion of the ethano bridge by a methylene moiety does not reflect into a
considerable change in the σ1 affinity of the corresponding derivatives. Indeed, the
propano bridged homologs 22a and 23a show almost the same σ1 affinity as the
ethano bridged ligands 14a and 15a with the same stereochemistry and the same
substitution pattern.
The σ2 receptor affinity of all bicyclic compounds is lower than their σ1 affinity (gp
assay, RPMI 8226 assay) varying from slight preference up to a high selectivity for
the σ1 receptor. The range of the σ1:σ2 selectivity is demonstrated by the
naphthylmethyl derivatives (cꢀseries), which show σ1:σ2 selectivity of 6, 3, 22, and
230ꢀfold for 14c, entꢀ14c, 15c and entꢀ15c, respectively. The particular high σ1:σ2
selectivity of the (1S,4R,7S)ꢀconfigured ligands entꢀ15a (90ꢀfold), entꢀ15c (230ꢀfold),
and entꢀ15d (55ꢀfold) should be emphasized.
The affinity of the bicyclic compounds towards human σ1 receptors (RPMI 8226 cell
line) shows a good correlation to the affinity recorded in the guinea pig brain assay.
In general the Kiꢀvalues for the naphthylmethyl (cꢀseries), biphenylylmethyl
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 48
16
1
2
derivatives (d-series) and propano bridged homologs 22a and 23a are slightly higher
in the RPMI 8226 assay than in the guinea pig brain assay. In contrast, the benzyl
derivatives (aꢀseries) show slightly stronger interactions with the human σ1 receptor
in the RPMI 8226 assay than with the guinea pig σ1 receptors. However, most of the
measured differences are due to the variability of the assays, thus lacking
significance. Interestingly, the most potent ligand in the guinea pig assay (entꢀ15c, Ki
= 0.50 nM) shows also very high affinity in the RPMI 8226 assay (Ki = 6.0 nM)
rendering it to one of the most affine ligands in this assay as well.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In conclusion, reduction of the conformational flexibility of (hydroxyethyl)piperazines
3 by incorporation of the pharmacophoric elements in a diazabicyclo[2.2.2]octane
framework led to the same or slightly reduced σ1 affinity. Ki values recorded in the
guinea pig assay are in good accordance with Ki values recorded in the RPMI 8226
assay. (1S,4R,7S)ꢀConfigured bicyclic ligands display high σ1:σ2 selectivity. The
ligand entꢀ15c represents the most promising σ1 ligand of this series of bicyclic
compounds with Ki values of 0.50 nM (guinea pig assay), 6.0 nM (RPMI 8226 assay)
and 230ꢀfold respective 20ꢀfold selectivity over the σ2 subtype.
Cytotoxicity
The ability of the new σ1 ligands to inhibit the growth of seven human tumor cell lines
was investigated in vitro by using two microtiter plateꢀbased assays: the growth of the
adherent cell lines A427 (small cell lung cancer), LCLCꢀ103H (large cell lung cancer),
5637 and RTꢀ4 (bladder cancer), DANꢀG (pancreatic cancer) and MCFꢀ7 (breast
cancer) was determined by a crystal violet staining assay described previously,52
whilst for the cell line HL60 (leukemia) growing in suspension the MTT assay was
ACS Paragon Plus Environment

Page 17 of 48
Journal of Medicinal Chemistry
17
1
2
3
4
used.52 The known σ1 ligands (+)ꢀpentazocine and haloperidol were included in these
5
6
investigations and served as reference compounds.
7
8
9
Table 2 displays the 50% growth inhibition concentrations (IC50) of the synthesized
bicyclic σ ligands 14, 15, 22a, and 23a together with the effects of the reference
compounds (+)ꢀpentazocine and haloperidol. As expected, 15b bearing a small
methyl moiety at Nꢀ2 did not inhibit the growth of any of the tumor cell lines up to a
concentration of 20 µM. This effect correlates well with its negligible affinity towards
both σ receptor subtypes.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The naphthylmethyl substituted derivatives 14c and 15c reveal rather unselective
inhibition of tumor cell growth based on very similar IC50 values over all cell lines,
thus indicating unspecific cytotoxicity rather than a precise mechanism of action.
Compounds 15a, entꢀ14a, entꢀ15c, and 14d slightly reduced the growth of the
bladder cancer cell line 5637. However, the most striking result is the selective
growth inhibition of the small cell lung cancer cell line A427 by the cyclohexylmethyl
substituted bicyclic compounds (exception made for 14c and 15c, which are not
selective). The growth of the other five cell lines was not influenced up to a test
compound concentration of 10 µM or 20 µM. Therefore, the following discussion will
focus on the growth inhibition of tumor cell line A427, which expresses high levels of
σ1 receptors37 and is the most sensitive cell line towards these bicyclic ligands.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 48
18
1
2
Table 2: Growth inhibition of human tumor cell lines, average IC50 ± SD [µM] of three
3
4
5
independent determinations (except where noted)
6
7
8
9
human tumor cell line
compd.
LCLCꢀ
103Ha
A427a
16.5 ± 6.2
9.8 ± 4.3
2.8 ± 1.7
11.2 ± 4.8
n.d.
5637a
> 20
RTꢀ4a
> 20
DANꢀGa
> 20
MCFꢀ7a
> 20
> 20
> 20
> 20
n.d.
HL60b
> 20
> 20
> 20
> 20
n.d.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
14a
15a
> 20
> 20
9.2 ± 6.3
4.8 ± 3.1
> 20
> 20
> 20
entꢀ14a
entꢀ15a
14b
> 20
> 20
> 20
> 20
> 20
> 20
n.d.
n.d.
n.d.
n.d.
15b
> 20
> 20
> 20
> 20
> 20
> 20
n.d.c
14c
2.3 ± 0.9
6.0 ± 3.8
1.8e)
10.4 ± 0.9
8.8 ± 2.0
> 10
4.3 ± 1.8
4.9 ± 1.2
> 10
12.9 ± 7.3
11.3 ± 5.9
n.d.
10.2 ± 3.1
16.1 ± 2.1
9.3e)
7.0 ± 4.1 11.2 ± 1.6
6.8 ± 1.5 14.7 ± 1.4
15c
entꢀ14c
entꢀ15c
14d
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
> 20
> 20
> 20
> 20
4.3 ± 2.3
1.6 ± 1.2
4.5 ± 5.7
3.7 ± 3.6
1.9 ± 1.5
7.6 ± 4.7
10.3 ± 2.9
> 20
> 10
2.3 ± 0.9
4.9 ± 4.1
> 10
n.d.
> 10
3.2e)
n.d.
4.9 ± 2.0
> 10
n.d.
15d
> 10
n.d.
n.d.
ent-14d
ent-15d
22a
> 10
> 10
n.d.
9.1 ± 1.0
> 10
n.d.
> 10
> 10
n.d.
n.d.
> 20
> 20
> 20
> 20
14 ± 2.8
16 ± 2.8
> 20d
> 20d
23a
> 20
> 20
> 20
> 20
+)ꢀpentazocine
Haloperidol
> 20
3.5 ± 0.9
2.3 ± 1.4
> 20d
16 ± 5d
> 20
9.6 ± 3.7
10.9 ± 1.9
> 20
a)determined with the crystal violet assay after a 96 h exposure to test compounds;
b)determined with the MTT assay after a 48 h exposure of the HL60 cell line to test
compounds; c)n.d.: not determined, d)values from ref.37, e)n = 2
As discussed for the σ1 receptor affinity, the four stereoisomeric cyclohexylmethyl
derivatives 14a, entꢀ14a, 15a and entꢀ15a display very similar antiproliferative activity
against A427 cell line, indicating a low influence of the stereochemistry on cell growth
inhibition. Similar observations were made for the growth inhibition of the
ACS Paragon Plus Environment

Page 19 of 48
Journal of Medicinal Chemistry
19
1
2
stereoisomeric naphthylmethyl (cꢀseries) and biphenylylmethyl (dꢀseries) substituted
derivatives as well as for enantiomeric monocyclic piperazine derivatives.54 The most
potent compounds are the naphthylmethyl substituted compounds 14c (IC50 = 2.3
µM) and entꢀ14c (IC50 = 1.8 µM) as well as the biphenylylmethyl substituted
derivatives 14d (IC50 = 2.3 µM) and entꢀ15d (IC50 = 1.9 µM). With Ki values of 13 nM
(14c) and 27 nM (14d) in the human RPMI8226 assay, the (1R,4S,7S)ꢀconfigured
compounds belong to the group of very high affinity σ1 ligands. Although a precise
correlation between the antiproliferative activity against the A427 cell line and the σ1
affinity is not given, the high affinity σ1 ligands entꢀ14a (Ki(human) = 2.8 nM) and entꢀ
15c (Ki(human) = 6.0 nM) inhibit the growth of the A427 tumor cell line also with high
activity (IC50 = 2.8 µM (entꢀ14a), IC50 = 4.3 µM (entꢀ15c)).
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The size of the bridge does not influence considerably the growth inhibition of the
A427 cell line, since the homologs 22a and 23a with an additional CH2 moiety in the
bridge show similar antiproliferative activity as their smaller homologs 14a and 15a.
It can be concluded that a clear correlation between the σ1 affinity of the test
compounds and their antiproliferative activities in the A427 cell line could not be
detected, but some trends were observed. For the interpretation of these results it
has to be considered that some physicoꢀchemical properties of the test compounds,
such as lipophilicity, which determine penetration of drugs through the cytoplasmic
membrane to enter the cells and interact with σ1 receptors located in the membrane
of the endoplasmic reticulum, influence the overall effect on tumor cell growth. These
aspects are not relevant in receptor binding studies with membrane preparations.
However, all bicyclic compounds bearing a cyclohexylmethyl moiety behave like the
σ1 receptor antagonist haloperidol in the inhibition of the growth of the A427 cell line.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 48
20
1
2
Therefore, the bicyclic compounds are likely to also be acting as σ1 receptor
antagonists.
3
4
5
6
7
8
9
Induction of apoptosis
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Based on their σ1 affinity and tumor cell growth inhibition, the bicyclic 5ꢀbenzyl
derivative entꢀ14a (aꢀseries), the naphthylmethyl derivatives entꢀ14c and entꢀ15c (cꢀ
series), and the biphenylylmethyl derivative entꢀ14d (dꢀseries) were selected for
further investigation of apoptosis induction in A427 cells (small cell lung cancer). This
human tumor cell line displays a high expression of σ1 receptors37 and a high
sensitivity towards cytotoxic effects of the compounds (Table 2). Cells were treated
for 24 h and 48 h with a 2ꢀfold higher concentration of the compounds than the
corresponding IC50 value, established by the crystal violet proliferation assay (96 h).
Subsequently the cells were doubleꢀstained with annexin VꢀFITC and propidium
iodide (PI) to distinguish between early apoptotic and late apoptotic/necrotic cells.
The stained cells were analyzed by flow cytometry. The anticancer agent doxorubicin
(0.5 µM for 24 h, 0.1 µM for 48 h), a wellꢀknown inducer of apoptosis55,56, was
included as a positive control.
The results of the annexin V / PI double staining experiments are shown in Figure 2.
Whilst after 24,h significant increases in the population of early apoptotic cells
(annexin Vꢀpositive, PIꢀnegative) could only be observed for the biphenylylmethyl
derivative entꢀ14d (39.7 ± 1.3 %), after 48 h the fraction of early apoptotic cells
significantly increased for all four of the tested compounds (entꢀ14a: 32.6 ± 1.5 % ,
entꢀ14c: 33.3 ± 5.4 %, entꢀ15c: 48.9 ± 3.8 %, entꢀ14d: 56.4 ± 4.3 %) compared to a
0.1 % (v/v) DMSOꢀcontaining solvent control (19.0 ± 2.5 % after 24 h, 17.3 ± 1.8 %
after 48 h). Untreated cells (medium only) displayed similar fractions of early
ACS Paragon Plus Environment

Page 21 of 48
Journal of Medicinal Chemistry
21
1
2
apoptotic cells (18.9 ± 2.8 % after 24 h, 16.6 ± 1.1 % after 48 h). Thus, for these four
compounds, the biphenylylmethyl derivative entꢀ14d with high σ1 affinity (Ki = 11 nM)
and growth inhibition (IC50 (A427) = 3.7 µM) is the most effective and fastest inducer
of apoptosis in A427 cells. Comparable timeꢀdependent induction of apoptosis by σ1
ligands with hydroxyethyl framework (see 3 in Figure 1) has been observed
before.40,54
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
****
****
****
****
**
***
***
Figure 2: Analysis of apoptotic effects of entꢀ14a, entꢀ14c, entꢀ15c and entꢀ14d by
annexin V / PI double staining. Annexin Vꢀpositive and PIꢀnegative A427 cells after
treatment with entꢀ14a (5.6 µM), entꢀ14c (3.4 µM), entꢀ15c (8.7 µM), and entꢀ14d
(7.4 µM) for 24 h and 48 h, respectively. Apoptosis was evaluated by flow cytometry
by determining the percentage of annexin Vꢀpositive, PIꢀnegative cells. Results
expressed as mean ± SD [µM] of at least three independent experiments.
Doxorubicin (0.5 µM for 24 h, 0.1 µM for 48 h) as positive control, 0.1 % (v/v) DMSO
as solvent negative control. * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001, **** p < 0.0001 (twoꢀ
way ANOVA followed by Dunnett’s multiple comparisons test by using GraphPad
Prism, GraphPad Software)
Molecular simulations
With the purpose of explaining the interactions between this new series of bridged
piperazines and the σ1 receptor at the molecular level, all derivatives were docked in
the binding site of our 3D homology model38,40,57 and the corresponding
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 48
22
1
2
3
4
58ꢀ61
ligand/protein free energies of binding (ꢀGbind
)
were evaluated by applying a
5
6
7
8
molecular dynamics (MD)ꢀbased scoring procedure in the framework of the soꢀcalled
Molecular Mechanics/PoissonꢀBoltzannn Surface Area (MM/PBSA) approach.62
9
Further, we performed a perꢀresidue decomposition of the enthalpic component of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
58ꢀ61
ꢀGbind
in order to quantitatively identify contribution afforded by the σ1 amino
acids mainly involved in binding these bicyclic compounds.
The analysis of the MD trajectories reveals that the new derivatives can establish a
series of intermolecular interactions quite similar to those previously detected for the
more flexible (ωꢀhydroxyalkyl)piperazines 1 and 3.40 Actually, all new synthesized σ1
ligands can bind their target receptor by exploiting four highly specific molecular
determinants, schematically represented in Figure 3A. In details, the
cyclohexylmethyl substituent at Nꢀ2 of the diazabicyclic system is encased in the
hydrophobic pocket generated by the σ1 residues Ile128, Phe133, Tyr173, and
Leu186, thereby establishing favorable, hydrophobic interactions with their side
chains. The basic Nꢀarylmethyl nitrogen atom (Nꢀ5) is engaged in a permanent salt
bridge with the carboxylic group of Asp126 whilst the different arylmethyl moieties,
common to all these new derivatives, are anchored in place by stabilizing πꢀtype
interactions. Specifically, residues Arg119 and Tyr120 are involved in receptor/ligand
πꢀcation and πꢀπ interactions, respectively (see Figure 3A). Finally, the hydroxy
substituent present on one of the three chiral carbon atoms of the diazabicyclic
scaffold plays an important functional group in the structure of these new molecules.
Indeed, it serves as a hydrogen bond acceptor, the donor counterpart being the –OH
group of the Thr181 side chain. The general binding mode of the new ligands
ACS Paragon Plus Environment

Page 23 of 48
Journal of Medicinal Chemistry
23
1
2
described above is portrayed in details in Figure 3B, taking compound 14d as a
3
4
5
6
proofꢀofꢀconcept.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. (A) 2D schematic representation of the identified interactions between the
3D homology model of the σ1 receptor and the bicyclooctane(ꢀnonane) compounds
synthesized in this work. The lines/arrows indicate key interactions between the
receptor and its ligand. (B) Equilibrated MD snapshot of the complex of the σ1
receptor with compound 14d. The main protein residues involved in these
interactions are Arg119 and Tyr120 (πꢀinteractions, cyan), Asp126 (salt bridge, red),
Ile128, Phe133, Tyr173, and Leu186 (hydrophobic interactions, purple), and Thr181
(hydrogen bond, green). Compound 14d is shown in atomꢀcolored sticksꢀandꢀballs
(C, gray; N, blue; and O, red). H atoms are omitted, but the salt bridge and the Hꢀ
bond are indicated as black dotted lines. Water molecules, ions, and counterions are
not shown for clarity.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 48
24
1
2
3
4
5
Table 3. MM/PBSA calculated binding enthalpy (ꢀHbind), binding entropy (ꢀTꢀSbind),
6
7
binding free energy ( Gbind), and the calculated Ki values for all compounds
ꢀ
8
9
considered in this work. The corresponding experimental values (Table 1) are also
shown in the last column for comparison.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
σ1
σ1 (hum)
i ± SEM [nM]
ꢀHbind
ꢀTꢀSbind
ꢀGbind
(calcd)
K
[kcal/mol]
[kcal/mol]
[kcal/mol]
K
i [nM]a)
21
3a
3b
3c
3d
ꢀ22.19 (0.16) ꢀ12.01 (0.28)
ꢀ17.44 (0.15) ꢀ10.40 (0.31)
ꢀ23.34 (0.17) ꢀ12.62 (0.27)
ꢀ23.31 (0.18) ꢀ13.02 (0.29)
ꢀ10.18 (0.32)
ꢀ7.04 (0.34)
ꢀ10.72 (0.43)
ꢀ10.29 (0.34)
21 ± 4.0
n.d.
6900
14
29 ± 10
34 ± 8.0
29
14a
15a
ꢀ20.68 (0.21) ꢀ10.15 (0.28)
ꢀ20.63 (0.16) ꢀ10.18 (0.29)
ꢀ20.90 (0.17) ꢀ10.09 (0.31)
ꢀ21.14 (0.21) ꢀ10.21 (0.27)
ꢀ10.53 (0.35)
ꢀ10.45 (0.33)
ꢀ10.81 (0.35)
ꢀ10.93 (0.34)
ꢀ7.46 (0.38)
ꢀ7.41 (0.39)
ꢀ11.19 (0.32)
ꢀ11.01 (0.34)
ꢀ11.38 (0.34)
ꢀ11.46 (0.36)
ꢀ10.68 (0.35)
ꢀ10.56 (0.35)
ꢀ10.33 (0.31)
ꢀ10.24 (0.35)
ꢀ10.50 (0.37)
ꢀ10.60 (0.33)
19
22
3.2 ± 0.4
2.4 ± 0.2
2.8 ± 1.0
1.6 ± 0.4
23%d)
entꢀ14a
entꢀ15a
14b
12
9.7
3400
3700
6.3
8.5
4.6
4.0
15
ꢀ17.45 (0.20)
ꢀ17.27 (0.19)
ꢀ9.99 (0.30)
ꢀ9.86 (0.34)
15b
n.d.
14c
ꢀ21.52 (0.16) ꢀ10.33 (0.28)
ꢀ21.25 (0.18) ꢀ10.24 (0.29)
ꢀ21.80 (0.22) ꢀ10.42 (0.26)
ꢀ21.85 (0.21) ꢀ10.39 (0.29)
ꢀ21. 21 (0.18) ꢀ10.53 (0.30)
ꢀ21.02 (0.21) ꢀ10.46 (0.28)
ꢀ20.54 (0.15) ꢀ10.21 (0.28)
ꢀ21.02 (0.17) ꢀ10.78 (0.31)
ꢀ20.51 (0.23) ꢀ10.01 (0.29)
13 ± 5.0
7.2 ± 3.9
38 ± 3.0
6.0 ± 2.0
27 ± 9.0
73 ± 6.0
27 ± 5.0
24 ± 6.0
6.4 ± 0.9
2.2 ± 1.1
15c
entꢀ14c
entꢀ15c
14d
15d
18
entꢀ14d
entꢀ15d
22a
27
31
20
23a
ꢀ20.58 (0.19)
ꢀ9.98 (0.27)
17
a)The σ1 Ki values were obtained from the corresponding
relationship: Gbind = ꢀRT ln Ki
ꢀGbind values using the
ꢀ
.
ACS Paragon Plus Environment

Page 25 of 48
Journal of Medicinal Chemistry
25
1
2
3
4
The MM/PBSA estimated values of the free energy of binding ꢀGbind shown in Table
5
6
3 confirm that all bicyclic derivatives – with the notable exception of the Nꢀmethyl
7
8
substituted molecules 3b, 14b and 15b (vide infra) ꢀ are endowed with high affinity
9
toward the σ1 receptor, since the extrapolated σ1 Ki values are in nanomolar range.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The per residue deconvolution of the enthalpic contribution to ligand binding, ꢀHbind,
allows to derive two further, important structural considerations about this new series
of compounds within the receptor binding site: the role of a bulky cycloalkyl
substituent and the absolute configuration related to the specific stereochemistry of
these molecules.
Concerning the first point, the replacement of the cyclohexylmethyl moiety with the
considerably smaller methyl group in compounds 14b and 15b leads to a dramatic
decrease in the relevant σ1 affinity, quantified by a three orders of magnitude
plummet in the corresponding σ1 Ki values (Tables 1 and 3). The reasons for this
behavior can be directly attributed to the reduced efficiency of the methyl substituent,
with respect to the bulkier cyclohexyl moiety, in generating substantial hydrophobic
connection with the residues lining the receptor binding pocket. This is clearly
supported by the relevant interaction spectra shown in Figure 4. De facto, the
favorable interactions between the cyclohexyl group and the side chains of Ile128,
Phe133, Tyr173, and Leu186 amount to ~ 3.5 kcal/mol for the corresponding
derivatives while, in the presence of the methyl group, the same interactions barely
afford an energetic stabilization to the receptor/ligand complex of 0.75 kcal/mol. This,
in turn, exerts a negative effect on global binding conformation of derivatives 14b and
15b, ultimately resulting in a general decrement of all enthalpic contributions to ligand
binding (Figure 4). In addition, the same analysis confirms that the addition of one
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 48
26
1
2
methylene moiety in the cyclic structure (i.e., the diazabicyclononane derivatives 22a
and 23a) does not result in any significant advantage in the binding of these
compounds with the σ1 receptor, as they exhibit an interaction profile utterly similar to
those characterizing the diazabicyclooctane counterparts.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. Per residue binding enthalpy decomposition (interaction spectra) for
compounds 14a-d and 22a in complex with the σ1 receptor. Only those σ1 amino
acids involved in major intermolecular interactions (see Figure 3) are displayed for
simplicity. Legend abbreviations: π = πꢀtype interactions; SB = salt bridge; HB =
hydrogen bond; HI = hydrophobic interactions.
Our MD simulation results confirm all present and previous experimental
observations regarding the stereochemistry issue: indeed, the flexible nature of the
σ1 binding site enables the receptor to easily and efficiently accommodate each
configuration of enantiomeric ligands when these result in small modifications in the
orientation of the molecular pharmacophore requirements. Figure 5A clearly shows
the obvious similarity in the equilibrated binding poses of the diastereomeric
compounds 14d and 15d, according to which the relative position of the hydroxy
group is practically irrelevant.
ACS Paragon Plus Environment

Page 27 of 48
Journal of Medicinal Chemistry
27
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. (A) Overlay of binding modes of diastereomers 14d (orange) and 15d (blue)
in the binding pocket of the σ1 receptor (colored transparent ribbons). The two
ligands are shown as colored sticksꢀandꢀballs, whereas the main interacting residues
are shown as colored sticks and labelled. Hydrogen atoms, ions, counterions and
water molecules are omitted for clarity. (B) Per residue binding enthalpy
decomposition (interaction spectra) for compounds 14d, ent-14d, 15d, and ent-15d in
complex with the σ1 receptor. Only those σ1 amino acids involved in major
intermolecular interactions (see Figures 3 and 4) are shown for simplicity. Legend
abbreviations: π = πꢀtype interactions; SB = salt bridge; HB = hydrogen bond; HI =
hydrophobic interactions.
Quantitatively speaking, also the corresponding enantiomers ent-14d and ent-15d do
not display significant differences upon binding to the receptor. In fact, according to
the corresponding interaction spectra shown in Figure 5B, all four specific molecular
determinants required for stabilizing the relevant σ1 receptor/ligand complexes are
practically not affected as concerns their enthalpic contribution to binding.
Notably, however, even if the new, conformationally more constrained piperazine
compounds overall seem to be as potent as the more flexible, monocyclic
compounds with respect to σ1 receptor affinity, this similar behavior is the result of an
underlying enthalpyꢀentropy compensation effect. In fact, as shown in Figure 6, the
presence of a substantially more rigid scaffold in the bicyclic derivatives reflects in a
less negative (i.e., more favorable) entropic binding component (ꢀTꢀSbind), of ~ 2
kcal/mol compared to the monocyclic compounds 3a, 3c and 3d. On the other hand,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 48
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 29 of 48
Journal of Medicinal Chemistry
29
1
2
diazabicyclooctanes 14a,c,d and 15a,c,d binding in the low nanomolar range (Ki ≤ 23
nM) at σ1 receptors. The high σ1 affinity of the cyclohexylmethyl derivatives is
explained by favorable interactions of the cyclohexyl group with the side chains of
Ile128, Phe133, Tyr173, and Leu186, which amount to ~ 3.5 kcal/mol of binding
enthalpy. Interaction of these residues with the small methyl moiety affords only a
binding enthalpy of 0.75 kcal/mol, reflecting the reduced affinity. The stereochemistry
of the bicyclic compounds has only limited influence on σ1 receptor binding.
Molecular dynamics calculations confirm the adaptation of the flexible σ1 receptor
binding site to stereoisomeric bicyclic ligands resulting in similar binding poses. In
particular the orientation of the ꢀOH moiety in the stereoisomers is practically
irrelevant. The conformationally restricted derivatives 14a,c,d and 15a,c,d reveal the
same or slightly reduced σ1 affinity as their flexible monocyclic counterparts 3a,c,d.
The similar σ1 affinities are the result of an enthalpyꢀentropy compensation effect.
Whereas the entropic binding component of the bicyclic compounds is increased (~ 2
kcal/mol) the enthalpic component is reduced by approx. the same amount, resulting
in comparable binding free enthalpies for both series of ligands.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In order to analyze species differences of σ1 receptors, membrane preparations
obtained from peripheral blood human myeloma cell line RPMI 8226 were used as
receptor material in an additional assay and the data were compared with data
recorded in the standard guinea pig brain σ1 assay. In general the affinity data
recorded in both assays are well comparable, which reflects the 93% sequence
identity of human and guinea pig σ1 receptors (see Figure S2). The small differences
between the values recorded in the assays could be due to the assay conditions. The
data recorded in the RPMI 8226 assay are of particular impact since all the molecular
ACS Paragon Plus Environment