From Sphingosine Kinase to Dihydroceramide Desaturase: A Structure-Activity Relationship (SAR) Study of the Enzyme Inhibitory and Anticancer Activity of 4-((4-(4-Chlorophenyl)thiazol-2-yl)amino)phenol (SKI-II)

Article abstract of DOI:10.1021/acs.jmedchem.5b01439

The sphingosine kinase (SK) inhibitor, SKI-II, has been employed extensively in biological investigations of the role of SK1 and SK2 in disease and has demonstrated impressive anticancer activity in vitro and in vivo. However, interpretations of results u

Full text of DOI:10.1021/acs.jmedchem.5b01439

Subscriber access provided by UNIV OSNABRUECK
Article
From Sphingosine Kinase to Dihydroceramide Desaturase: A Structure-
Activity Relationship (SAR) Study of the Enzyme Inhibitory and Anticancer
Activity of 4-((4-(4-chlorophenyl)thiazol-2-ylamino)phenol (SKI-II)
Luigi Aurelio, Carmen Scullino, Melissa Pitman, Anna Sexton, Vitoria Oliver, Lorena Davies,
Richard Rebello, Luc Furic, Darren John Creek, Stuart Pitson, and Bernard L. Flynn
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b01439 • Publication Date (Web): 18 Jan 2016
Downloaded from http://pubs.acs.org on January 20, 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 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Title
From Sphingosine Kinase to Dihydroceramide Desaturase: A StructureꢀActivity Relationship
SAR) Study of the Enzyme Inhibitory and Anticancer Activity of 4ꢀ((4ꢀ(4ꢀchlorophenyl)thiazolꢀ2ꢀ
(
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
yl)amino)phenol (SKIꢀII).
Authors
a
a
b
a
a
Luigi Aurelio, Carmen V. Scullino, Melissa R. Pitman, Anna Sexton, Victoria Oliver, Lorena
b
c
c
a
b
a
Davies, Richard J. Rebello, Luc Furic, Darren J. Creek, Stuart M. Pitson, * Bernard L. Flynn. *
a
Monash Institute of Pharmaceutical Science, Monash University, 381 Royal Pde, Parkville, VIC,
3052, Australia
b
Centre for Cancer Biology, University of South Australia and SA Pathology, Frome Road,
Adelaide SA, 5000, Australia
c
Cancer Program, Monash Biomedicine Discovery Institute and Department of Anatomy and
Developmental Biology, Clayton, VIC, 3800, Australia
Abstract
The sphingosine kinase (SK) inhibitor, SKIꢀII, has been employed extensively in biological
investigations of the role of SK1 and SK2 in disease and has demonstrated impressive anticancer
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
activity in vitro and in vivo. However, interpretations of results using this pharmacological agent
are complicated by several factors: poor SK1/2 selectivity; additional activity as an inducer of
SK1ꢀdegradation; and offꢀtarget effects, including its recently identified capacity to inhibit
dihydroceramide desaturaseꢀ1 (Des1). In this study we have delineated the structureꢀactivity
relationship (SAR) for these different targets and correlated them to that required for anticancer
activity and determined that Des1 inhibition is primarily responsible for the antiproliferative
effects of SKIꢀII and its analogues. In the course of these efforts a series of novel SK1, SK2 and
Des1 inhibitors have been generated, including compounds with significantly greater anticancer
activity.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Introduction
Over the last fifteen years there has been considerable interest in identifying the biological
functions of sphingolipid metabolites and the role they play in diseases such cancer, autoimmune
1
disease, inflammation, pain, macular degeneration, asthma and fibrosis. The de novo synthesis of
sphingolipids commences with palmitoyl CoA, which is converted into the central lipid ceramide
(Cer) in four steps, the last step being the introduction of the double bond at C4 by
dihydroceramide desaturaseꢀ1 (Des1) (Scheme 1). Ceramide is regarded as the central lipid in this
biosynthetic pathway from which many other bioactive metabolites are generated, usually in
reversible steps from which ceramide can be regenerated by salvage pathways. Of significant
interest has been the reversible conversion of Cer into sphingosine (Sph) and sphingosineꢀ1ꢀ
phosphate (S1P) (Fig 1). Cer and Sph are proꢀapoptotic signaling lipids whereas S1P promotes cell
1
a,g,k
growth and survival (antiꢀapoptotic and proꢀmitogenic).
Consequently, sphingosine kinases
(SKs), which produce S1P, can play an important role in determining cell fate, providing a
2
“
rheostat” for cancer cell survival and death. There are two isoforms of SK, SK1 and SK2. SK1
1
a,c,g,k
exists in the cytosol and SK2 predominantly in the nucleosome and other organelles.
SK1 is
subject to tight regulation and is upregulated in many cancers where it is thought to play a key role
2
ACS Paragon Plus Environment

Page 3 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
in disease progression through increases in S1P and decreases in Cer and Sph levels, promoting
1a,e,j,o,2
tumour growth and survival.
Thus a number of groups have pursued SK1 inhibitors in order
3,4
to reverse this prosurvival effect of the Cer/Sph/S1P rheostat. SK2 inhibitors are also being
sought and may induce different mechanisms of cancer cell death to those mediated by SK1
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
3
inhibitors. Notwithstanding the compelling rationale for targeting sphingosine kinases, the
emergence of different SK1/2 inhibitors has met with conflicting results regarding their anticancer
3
potential. Early inhibitors included a series of Sph analogues (not shown) with mixed SK1/2
selectivity and IC50 values in range of 5ꢀ30 ꢁM and the nonꢀlipidꢀlike inhibitors, SKIꢀII 1 and
ABC294640 2 (Fig 1). Compound 1 was identified by Smith and coworkers from a high
4
a
throughput screen in 2003, it exhibits some selectivity towards SK2 as a kinase inhibitor (Ki:
SK2 = 7.9 ꢁM, SK1 = 16 ꢁM) but is a more powerful promoter of SK1ꢀdegradation (100% SK1
5
ablation at 10 ꢁM) and has recently been identified as an inhibitor of Des1 activity (Ki = 0.3
6
ꢁ
M). The SK1ꢀdegradation induced by 1 is proposed to arise from its binding to an allosteric site,
distinct from the enzymatic site, which promotes polyubquination and proteasomal degradation of
5
SK1. Compound 2, was also discovered by Smith and coꢀworkers and is a selective inhibitor of
7
SK2 (Ki = 10 ꢁM). Both 1 and 2 have been used in a large number of studies directed at
evaluating the therapeutic potential of targeting SKs in cancer and other indications, both have
7
,8
exhibited favorable efficacy in cancer models and 2 has been progressed to clinical trials. Despite
exhibiting a variety of potentially valuable anticancer properties, 1 and 2 are only moderately
potent and in recent years a number of more potent (IC50 < 1 ꢁM) and more selective SK1/2
4b
4c,d
4e
inhibitors have emerged. These include the SK1 selective inhibitors 3, 4, and PFꢀ543 5 and a
4h,i
dual SK1/2 inhibitor 6, (Fig 1). Like 1, 5 is also a promoter of proteasomal degradation of SK1
4
g
and presumably also has some affinity for the allosteric site on SK1. However, in contrast to 1, 4-
6
do not to show any antiproliferative effects at concentrations that fully suppress S1P production
3,4d,e,i,l
in cells, casting doubt on the use of SK1 inhibitors as proapoptotic agents in cancer therapy.
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
4
h,i,
Similar conflicts arise in the case of SK2, where the moderately potent SK2 inhibitors 6
and
4
f
SLR080811 7 do not exhibit antiproliferative effects, whereas the less potent SK2 inhibitors
4j
4k
K145 8 and MPꢀA08 9 do show increases in Cer concentrations and cancer cell apoptosis (GI50
5ꢀ30 µM). Herein, we report our SAR studies on 1 and attempts to correlate the SAR for SK1/2
=
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
inhibition to the SAR for the antiproliferative effects in androgenꢀindependent PC3 prostate cancer
cells (SK1 expression is raised in PC3 cells relative to androgenꢀdependent LNCap prostate cancer
9
cells). We have found that these two SARs diverge and that inhibition of these kinases cannot
account the anticancer activity of 1 and our broader compound series. We have also studied the
SAR of 1 in relation to SK1ꢀdegradation and inhibition of Des1 activity. The SAR of Des1
inhibition correlates to that required for inhibition of PC3 cell growth, indicating that Des1
inhibition is a key driver of the anticancer effects of 1 and it analogues, which is also supported by
lipidomic studies in PC3 cells. In the course of this effort a number of substantially more potent
inhibitors of some of these enzymes and of PC3 cell growth have been identified.
Scheme 1: Sphingolipid Biosynthesis
4
ACS Paragon Plus Environment

Page 5 of 70
Journal of Medicinal Chemistry
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 1: Structures of SK1/2 targeting agents.
SAR Studies: Our SAR studies of 1 have been directed to modifications of the Bꢀring (Table 1)
a
b
and of groups R , R , L and the Cꢀring (Table 2). As part of our SARꢀguided exploration of 1, we
sought to identify useful replacements of the 2ꢀaminothiazole and 4ꢀaminophenol moieties as they
are likely to be subject to rapid metabolism to form reactive metabolites (bioactivation), potentially
1
0
compromising in vivo drug exposure and safety. We have also avoided generating lipidꢀlike,
amphiphilic structures that are apparent in a number of the more potent SK1 inhibitors 3ꢀ6 and
which are known to complicate cellular readouts, including through surfactant associated
4
i
cytotoxicity. SK1/2 inhibition was measured at a single concentration (10 ꢁM) and recorded as
1
1
%
SK1/2 activity relative to vehicle (DMSO) control. PC3 cell growth was measured using an
MTS colorimetric assay and is given as the concentration required to inhibit 50% of cell growth
GI50). Selected compounds were also evaluated for SK1ꢀdegradation, Des1 inhibition, lipidomics
(
analysis and clonogenic potential in PC3 cells (see below). SK1 inhibitors 1, 3 and 5 were used as
comparators in these studies. As expected, 1 gave only a modest level of inhibition of SK1 (%SK1
=
78) and SK2 (%SK2 = 90) and inhibited PC3 cell growth (GI = 11 ꢁM) with a potency that is
50
4a,5a
broadly similar to that seen with other cell lines (Table 1, entry 3).
Consistent with their
4
b
4e
reported high potency towards SK1, 3 (IC50 = 80 nM) and 5 (IC50 = 3 nM) gave low %SK1
values, 6% and 1%, respectively, and were much less potent towards SK2 (Table 1, entries 1 and
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2). Compound 5 inhibited PC3 cell growth at a relatively high concentration (GI50 = 19 ꢁM)
compared to its SK1 inhibition and 3 was inactive (GI50 > 100 ꢁM). The cytotoxicity observed
with 5 (PC3 GI50 = 19 ꢁM) may arise from its amphiphilic nature as has been reported for the
4
i
similarly amphiphilic compound 6. Interestingly, all of our novel bioisosteric Bꢀring analogues of
1, 10ꢀ17 (Table 1, entry 4ꢀ11) gave higher levels of SK1/2 inhibition, with oxazole 17 giving the
best SK1 potency and selectivity (Table 1, entry 11, %SK1 = 17, %SK2 = 53). The most potent
PC3 cell growth inhibitor, 1,3,4ꢀoxadiazole 10 (GI50 = 1.63 µM), is almost ten times more potent
than 1 (Table 1, cf entries 3 and 4). Given the frequent use of the pyrrolidine ring as a polar head
group in 3-7 we also prepared two hybrid structures of 1 and 3, compounds 18 and 19 (Table 1,
entries 12 and 13) but both of these were much less active than 3 towards SK1 and effectively
devoid of any PC3 cell growth inhibition.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 1: B-ring Modifications.
Entry
Bꢀring
ID
%SK1
%SK2
GI50
a
a
b
1
0 ꢁM
6 ± 2
1 ± 2
10 ꢁM
112 ± 6
45 ± 9
PC3
>100
19
1
.
3
5
2.
3
.
.
1
78 ± 2
43 ± 4
90 ± 7
11 ± 2
4
10
44 ± 8 1.6 ± 0.4
5
.
11
55 ± 4
70 ± 3
2.4
6
ACS Paragon Plus Environment

Page 7 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
6
.
12
50 ± 5
55 ± 4
40
7
8
9
1
.
13
14
15
50 ± 4
34 ± 4
75 ± 4
77 ± 15
64 ± 8
78 ± 8
44
2.6
ND
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
.
.
0.
16
17
18
55 ± 2
17 ± 5
68 ± 1
56 ± 4
ND
3.0
77
11.
1
2.
3.
62 ± 10
106 ± 24
1
19
41 ± 4
118 ± 18 >100
a
The activity of purified recombinant SK1 and SK2 were determined in the presence of 10ꢀM
1
2
compound using an in vitro SK enzyme assay with 10 ꢀM Sph and 100 ꢀM ATP (n = 3).
b
ꢀ8
PC3 cell growth was measured using an MTS colorimetric assay at eight concentrations (10 ꢀ
ꢀ4
1
0 M, n = 3 per concentration) and is given as the concentration required to inhibit 50% of cell
growth (GI50). Where given, the standard deviations (±) is based on 3 independent GI50
determinations.
The 1,3,4ꢀoxadiazole 10 was selected for further SAR studies exploring the effects of modifying
a
b
the linker (L), polar head group R and hydrophobic tail R on SK1/2 activity and PC3 cell growth
(Table 2). Removal of the amino (NH) Lꢀgroup in 10 to give 20, had no effect on the SK1/2
inhibition but completely removed its cell growth inhibitory activity (Table 2, cf entries 1 and 2).
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
Several alternative Lꢀgroups were also evaluated: L = S, C(O)NH, and CH NH (Table 2, entries 3ꢀ
5
). Of these, L = CH
2
NH (23), gave the greatest improvement in SK1/2 inhibitory activity (Table
2
, entry 5) and only 22 and 23 inhibited PC3 cell growth. A key observation in the SAR of
compounds 10 and 20-23 (Table 2, entries 1ꢀ5) is the apparent requirement of a 4ꢀaminophenol for
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
antiproliferative activity, which is not required for SK1/2 inhibition. The 4ꢀaminophenol is
9
commonly linked to redox activity, indicating that the antiproliferative activity of 1, 10, 22 and 23
may arise from a redoxꢀactive target (see below).
a
We also evaluated a series of polar head groups R to replace OH in 10 in the presence and absence
of L = NH (Table 2, entries 6ꢀ18). MacDonald, Webster and Lynch have shown that the amidine
(C(NH)NH
2
) is a useful polar head group in promoting SK1/2 inhibition, in particular the
4
cꢀf
acylpyrrolidineꢀ2ꢀcarboximidamide (APCI) group, as in 4 (Fig 1).
We evaluated the simple
amidine (C(NH)NH
2
) headꢀgroup and APCI as well as a number of other neutral, basic and acidic
a
isosteres of the amidine and OH groups: R = NH
2
, CH
2
NH
2
, CO
2
H, C(O)NH
2
, C(NH)NHOH and
a
tetrazole (Table 2, entries 6ꢀ17). Considering first the effect these variations in R on SK1/2
inhibition for where L is absent (entries 6ꢀ13), the carboxamide 27 (%SK1 = 22) and tetrazole 31
(%SK2 = 14) gave the best SK1/2 inhibitory activity and were selective for SK1 and SK2,
respectively (Table 2, entries 9 and 13). The amidine 28, amidine oxime 29 and APCI 30 head
a
groups (R ) were least effective in SK1/2 inhibition (Table 2, entries 10 ꢀ 12). The introduction of
a
L = NH for these different R groups saw little effect on carboxamide 33 SK1/2 inhibitory activity
(%SK1 = 20), whereas the tetrazole 35 showed reduced activity. By contrast, the introduction of L
=
NH conferred a marked increase in potency of amidine 33 (%SK2 = 14) over 28 towards SK2
(Table 2, cf entries 11 and 15). With respect to PC3 cell growth of the compounds 24-36 (Table 2,
entries 6ꢀ18), only four compounds showed any activity: 28 (GI50 = 45 ꢁM), 34 (GI50 = 0.34 ꢁM)
and 35 (GI50 = 15 ꢁM). Again there is no correlation between SK1/2 inhibition and PC3 cell
growth inhibition in this series.
8
ACS Paragon Plus Environment

Page 9 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a/b
Table 2: Modifications of R , L and C-Ring.
Rb
Ra
a
HN
b
L
R =
R =
O
O
N
O
NH2
Rb
C
Ra
O
N
N
N N
L = absent
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
CyEt
APCI
Cl
H
O
N
O
N
C
N
C
H
N
Cl
N
N
N N
45
44
a
b
a
a
a
Entry
R
ꢀLꢀ
R
ID
%SK1
%SK2
GI50 PC3
1
0 ꢁM
10 ꢁM
(µM)
1.6 ± 0.4
>100
1
2
3
.
.
.
OH
OH
OH
OH
OH
NH
NH
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
10
20
21
22
23
24
25
26
27
28
29
30
31
32
43 ± 4
44 ± 8
40 ± 3
79 ± 22
absent
ꢀSꢀ
42 ± 8
107 ± 27
78 ± 8
>100
4.
ꢀC(O)ꢀNHꢀ
ꢀCH ꢀNHꢀ
139 ± 18 22
5
.
2
22 ± 3
39 ± 3
50 ± 11
57 ± 4
40 ± 8
51
6.
2
absent
absent
absent
absent
absent
53 ± 10
51 ± 8
>100
>100
>100
7
.
CH
CO
2
NH
H
2
8.
2
57 ± 10
22 ± 5
9
1
1
1
1
1
.
C(O)NH2
C(NH)NH
104 ± 12 >100
0.
1.
2.
3.
4.
2
73 ± 5
89 ± 11
102 ± 5
45 ± 21
70
C(NH)NHOH absent
105 ± 13
69 ± 6
APCI
absent
absent
ꢀNHꢀ
135 ± 16 >100
tetrazole
C(O)NH2
133 ± 7
20 ± 2
14 ± 10
43 ± 12
>100
>100
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
5.
6.
C(NH)NH
2
ꢀNHꢀ
Cl
Cl
Cl
Cl
H
33
34
35
36
37
38
39
40
41
42
46 ± 17
93 ± 23
44 ± 8
14 ± 7
>100
1
C(NH)NHOH ꢀNHꢀ
121 ± 8 0.34 ± 0.05
17.
8.
19.
tetrazole
C=NꢀOH
OH
ꢀNHꢀ
ꢀNHꢀ
ꢀNHꢀ
ꢀNHꢀ
ꢀNHꢀ
ꢀNHꢀ
70 ± 6
71 ± 9
23 ± 2
17 ± 2
3 ± 0.1
15
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
41 ± 15
53 ± 27
35 ± 14
53 ± 1
>100
8.1 ± 2.1
1.0 ± 0.8
0.7 ± 0.14
20.
21.
22.
23.
2
OH
CF
I
3
OH
OH
CyEt
56 ± 14
133 ± 21
3 ± 1
144 ± 51 50
Cꢀring = pyr
Cꢀring = pyr
ꢀC(O)NHCH2ꢀ Cl
ꢀNHCH Cl
73 ± 8
>100
>100
4.
2
ꢀ
39 ± 12
a
See Table 1.
b
We also undertook a preliminary examination of the SAR associated with hydrophobic group R
3
(Table 2 entries, 19ꢀ22). Replacement of Cl with hydrophobic groups of increasing size H, CF , I,
b
(37ꢀ39) led to successively increased inhibition of SK2, with 39 (R = I) proving quite potent and
selective towards SK2 (%SK2 = 3) (Table 2, entries 19ꢀ21). On the other hand, the much larger
(cyclohexyl)ethyl (CyEt) group, that proved effective in promoting the SK1 inhibitory activity of
4
c,d
4b
4
and analogues of 3 (not shown) diminished its activity towards SK2 (Table 2, entry 22). Like
other 4ꢀaminophenols, 40ꢀ43 all inhibited PC3 cell growth, with 39 showing the greatest activity,
GI = 0.7 ꢁM (Table 2, entry 21). In this series (37ꢀ40), the SAR for SK2 inhibition does appear to
5
0
coincide with inhibition of PC3 cell growth (Table 2, entries 19ꢀ22).
We next explored the use of the 4ꢀpyridyl ring present in the SK2 inhibitor 2 (Fig 1), as an
2 2
alternative Cꢀring. We employed two linker groups for this: L = ꢀC(O)NHCH ꢀ and ꢀNHCH ꢀ, 41
and 42, respectively (Table 2, entries 23 and 24). Only 42 showed any appreciable SK1/2
10
ACS Paragon Plus Environment

Page 11 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
inhibitory activity, but in contrast to 2, proved to be a moderately selective SK1 inhibitor (%SK1 =
3
, %SK2 = 39).
We determined the IC50 values of our most potent SK1/2 inhibitors 42 and 39 (Table 3) and
compared these to the most potent SK1 inhibitor yet to be identified 5 (Fig 1). The previously
reported IC50 (SK1) values for 5 of 3 and 28 nM were obtained at 3 and 5 ꢁM of Sph,
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
4e,g
respectively, our value of 387 nM was obtained at 10 ꢁM of Sph. These variations are consistent
4
e
with the substrate (Sph) competitive binding kinetics of 5. While the dual SK inhibitor 42 is not
as potent as 5, it is a more drugꢀlike species (lower MW and cLogP) that exhibits a superior
1
3
lipophilicityꢀcorrected ligand efficiency (LELP): 5 LELP = 16 cf 42 LELP = 10 (Table 3).
Compounds with LELP values of > 10 are considered unlikely to achieve suitable concentrations
12
in vivo to modulate their target.
Table 3: IC50 values of selected compounds
a
a
SK1 IC50
SK2 IC50
(ꢁM)
a
Entry Cmpnd
LELP
16.2
(ꢁM)
1
3
4
5
0.387
>20
3.1
17.0
39
42
1.9
5.5
9.8
10.0
a
Inhibitory constants for purified recombinant enzymes were obtained by an in vitro SK enzyme
assay with 1mM ATP, 10µM Sph, and the compounds at concentrations of 0.005 ꢀ 10 ꢁM. Data
represent mean two independent experiments.
We further evaluated a selected series of our compounds for their capacity to induce SK1ꢀ
degradation (Fig 2). SK1ꢀdegradation was measured using our previously described assay where
TM
expression of FLAG epitopeꢀtagged SK1 is induced in HEK293 cells and the effect of the
compounds on degradation of this existing protein is selectively assessed by Western blot after 16
5a
h. The proteasome inhibitor MG132 was also included in these assays to demonstrate the
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
involvement of the proteasome in the SK1ꢀdegradation induced by these compounds. As
previously demonstrated, 1 completely ablates SK1 levels at 10 µM in a MG132ꢀdependent
5a
manner. In fact, SK1ꢀdegradation is likely to be the primary source of SK inhibition by 1 in cells
as it is only a weak inhibitor of SK1/2 kinase activity. All other compounds used in this study that
induced SK1ꢀdegradation also did so in an MG132ꢀdependent manner (Fig 2). Comparison of 1
with the close structural analog 10, shows 1 to be a much more effective promoter SK1ꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
degradation than 10 despite the fact that 10 is a more potent SK1 kinase inhibitor (Table 1, entries
5
3
and 4). This comparison indicates that, consistent with the allosteric model, SK1ꢀdegradation is
independent SK1 kinase inhibition. Presumably, the exchange of the thiazole ring in 1 for the
oxadiazole ring in 10 has decreased affinity for the allosteric site and increased affinity for either
the ATP or Sph binding sites on SK1 and SK2. A comparison of the close structural analogues 3,
1
8 and 19 further supports the finding that SK1ꢀkinase inhibition and SK1ꢀdegradation are
independent effects. 3 is a potent inhibitor of SK1 kinase activity and modest inducer of SK1ꢀ
degradation, whereas 18 and 19 are much less potent SK1 kinase inhibitors but very effective in
SK1ꢀdegradation. None of these compounds, 3, 18, 19 exhibited significant PC3 cell growth
inhibition (Table 1 entries 1, 12 and 13). The Pfizer compound 5 is both a potent inducer of SK1ꢀ
degradation and inhibitor of SK1 kinase activity. Also, our most potent inhibitor of PC3 cell
growth, amidine oxime 34, is a moderate promoter of SK1ꢀdegradation and has no SK1/2 kinase
inhibitory activity (Table 2, entry 15). These comparisons indicate the promotion of SK1ꢀ
degradation, SK1ꢀkinase inhibition and cell growth inhibition are all independent effects though
some compounds may exhibit more than one of these effects.
Vehicle
1
28
3
18
19
5
10
34
MG132
SK1
−
+
−
+
−
+
−
+
−
+
−
+
−
+
−
+
−
+
tubulin
12
ACS Paragon Plus Environment

Page 13 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 2: Effect of the compounds on SK1 protein degradation. Expression of FLAG epitopeꢀ
tagged SK1 was induced HEK293 cells and the effect of applying the compounds to cells for 16 h
at 10 ꢀM on degradation of this existing protein selectively quantified by Western blot.
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 close correlation of PC3 cell growth inhibition with the presence of redox active groups
indicates that inhibition of a redoxꢀactive enzyme may be responsible for this effect. This overlaps
6
with the recent finding that Des1 activity is potently inhibited by 1 (K
i
= 0.3 ꢁM). To investigate
this further we examined several of our antiproliferative compounds as well as several close
structural analogues that do not exhibit antiproliferative effects for inhibition of Des1 activity (Fig
3
). Des1 activity was obtained by measurement of the fluorescently labeled ceramide analogue,
1
4
C6ꢀNBDꢀdhCer, in Jurkat cells as described previously. The conversion of C6ꢀNBDꢀdhCer to
C6ꢀNBDꢀCer is attributed to Des1 since the alternative Desꢀsubtype, Des2, gives rise to significant
C4ꢀhydroxylation, which was not observed with this study (no detectable C4ꢀhydroxylation of C6ꢀ
1
3
NBDꢀdhCer) and because C4ꢀhydroxysphingolipids are hardly present in Jurkat cells. Since the
assay is performed in whole cells, the direct target leading to the Des1 inhibition cannot be easily
discerned and may include other elements in the electron transport pathway associated with Des1
6
activity. Indeed, Cingolani et. al. propose 1 to be an indirect inhibitor of Des1, possibly targeting
the upstream enzyme, NADHꢀcytochrome b5 reductase (CBR5), due to it being a nonꢀcompetitive
inhibitor of Des1 and it being structurally dissimilar to the endogenous substrate, dhCer. Our
evaluation of 1 and the structurally related phenols 10, 20 and 39 revealed all the 4ꢀaminophenols
1, 10 and 39 to be quite potent inhibitors of Des1, whereas the simpler phenol 20 was inactive (Fig
3). This correlates with the PC3 cell growth inhibition where 1, 10 and 39 inhibit cell growth but
20 is inactive (Tables 1 and 2). It also overlaps with the SAR of the known inhibitor of Des1,
1
5
fenretinide 43 (Fig 4), which also contains a 4ꢀaminophenol and is an inhibitor PC3 cell growth
1
6
(GI ~ 2 ꢁM). 43 is believed to be a direct inhibitor of Des1 that initially shows competitive
5
0
14
inhibition but becomes an irreversible inhibitor over longer incubation times. The 4ꢀaminophenol
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
in 43 group may play an important role in mediating its irreversible inhibition of Des1, where Des1
oxidation of this group generates a reactive iminoquinone intermediate 44 that reacts with
nucleophilic sites on the protein, leading to time dependent irreversible inhibition of Des1 (Fig 4).
A similar mechanism may be operating in the case of 1 and the related 4ꢀaminophenols prepared in
this study, in which case they may also operate as direct inhibitors of Des1. The amidine oxime 34
completely inhibited Des1 activity at 10 ꢁM but showed little activity at 1 ꢁM. Alkyloximes have
been previously shown to undergo oxidation to nitroso groups that form high affinity complexes
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
7
with the iron in the redox enzyme prostaglandin synthase. Amidine oxime 34 may also undergo
oxidative activation to form a nitroso intermediate 45 that forms a high affinity complex with the
Fe O in Des1. Consistent with the redox activation of 34, is the much reduced activity of the
2
2
structurally similar amidine 33 and aldimine oxime 36, which are more resistant to oxidation. If
redox activation is involved in the inhibition of Des1 activity by 34, then it is likely to be a highly
targeted effect requiring quite specific structural interactions as the structurally related amidine
oxime 29 showed no Des1 inhibitory activity at 10 ꢁM. Of these related structures, 29, 33, 34 and
3
6 only the active Des1 inhibitor 34, showed any significant activity against PC3 cell growth,
which is conststant with what was seen for the phenolic compounds (1, 10, 20 and 39).
Notwithstanding the suggested possibilities for the direct inhibition of Des1 by these redox
sensitive compounds, it remains a significant possibility that other enzymes in the redox cycling
pathway are the target of these inhibitors.
Inhibition of Des1 activity and associated increases in dhCer are known to give rise to both
18
apoptotic and autophagic cell death. Lipidomic analysis of 1 and 34 in PC3 cells (10 ꢁM)
indicates that Des1 inhibition dominates their effects on sphingolipid synthesis, with significant
increases in total dhCer levels but little effect on total Cer, Sph and S1P levels (Fig 5). Since 34
inhibited PC3 cell growth at a concentration of < 1 ꢁM we also undertook lipidomic analysis of
this compound in PC3 cells at 1.0 ꢁM, but it showed no effect on dhCer levels (data not shown).
14
ACS Paragon Plus Environment

Page 15 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
This indicates that while Des1 inhibition alone may explain the anticancer effects of 1 (PC3 GI50
=
1
1 ꢁM) and many of its other 4ꢀaminophenols analogues (Table 1 and 2), it is likely that the cell
growth inhibitory effects of 34 (PC3 GI50 = 0.34 ꢁM) involves another target(s) in addition to
Des1. Alternatively, these different irreversible inhibitors may have different rates of oxidative
activation and inhibition of Des1. The time period of drug exposure in the PC3 cell growth
inhibition assay (72 h) is much longer than that in the Des1 assay (3 h) and in the lipidomics study
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
(5 h). Further evaluation of the inhibition kinetics of 1 and 34 are required in order to verify that
they are irreversible inhibitors of Des1 and if the time course of Des1 inhibition can explain
current apparent differences in the level of Des1 inhibition and PC3 cell growth inhibition for 34.
Figure 3, Inhibition of Des1: Des1 activity was obtained by measurement of C6ꢀNBDꢀCer
conversion in Jurkat cells preꢀlabeled with C6ꢀNBDꢀCer (10 ꢁM). Cells were treated with vehicle
or inhibitor (0.1, 1, 10
ꢀM) for 3 h. Data represent mean ± SD of three independent experiments.
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
OH
O
R
N
H
fenretinide 43
Des1
NH
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4-aminophenol
or
amidine oxime
(Fe O )
2
2
N
O
or
O
N
O2 H O
44
45
2
(
from 4-aminophenol) (from amidine oxime)
Figure 4: Fenretinide and irreversible inhibition of Des1.
Figure 5, Lipidomic analysis: Effects on sphingolipid levels of 1 and 34 at 10 ꢁM in PC3 prostate
cancer cells.
The correlation between redoxꢀactivity and antiproliferative effects seen with 1 and its analogues
does not easily extend to the SK2 inhibitor 2, which does not contain such a moiety. Nonetheless,
4i
consistent with the Amgen group’s studies, our results indicate that the antiproliferative effects
seen with 2 are unlikely to arise from inhibition of SK2 kinase activity alone as the potent SK2
inhibitors 31 and 33 (%SK2 = 14 for both) had no effect on PC3 cell growth even at high
concentrations (100 ꢁM). The estrogen receptor has been identified as a potential off target for 2,
1
9,
which it antagonizes in similar manner to tamoxifen. However, this may not explain all the
anticancer effects seen with 2, which may still involve contributions from SK2 inhibition. Very
2
0
recently, 2 has been reported to be a weak inhibitor of Des1. SK2 inhibition may act
16
ACS Paragon Plus Environment

Page 17 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
cooperatively with Des1 inhibition in the antiproliferative effects seen with 39 (PC3 GI50 = 0.7
20
ꢁM) and 2 (GI50 = 28ꢀ30 µM in several prostate cancer cells line).
To test if the antiproliferative compounds 1, 10 and 34 inhibit the clonogenic potential of PC3
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
21
cells, we performed soft agar colony formation assays (Fig 6). All compounds showed
qualitatively similar potency to inhibit anchorageꢀindependent growth in soft agar compared to that
obtained in the plated assays. For example, the most potent compound, 34, has an IC of 0.5 ꢁM
50
and 0.34 ꢁM in the soft agar and plated assays, respectively.
Figure 6: Inhibition of colony formation in soft agar. PC3 cells were seeded in soft agar and
treated with increasing concentrations of each compound. Data represent mean ± SEM of three
independent experiments performed in duplicate. **P<0.01. ***P<0.001 ****P<0.0001.
Synthetic Chemistry: The synthesis of bioisosteric replacements of the thiazole ring (Table 1)
involved the preparation of a series of oxadiazoles, thiadiazoles, triazoles and oxazoles (Scheme
2). The 1,3,4ꢀoxadiazole 10 and 1,3,4ꢀthiadiazole 11, were both prepared through nucleophilic
substitution of the corresponding heteroaryl bromides 47 and 48. Heteroaryl bromides 47 and 48
were prepared by known methods, involving cyclocondensation of 4ꢀchlorobenzoyl chloride (46)
with hydrazine carboxamide (or thiocarboxamide), followed by Sandmeyer reaction to convert the
2
2
initially formed heteroaryl amines (not shown) into bromides. Bromide 47 underwent direct
substitution by 4ꢀaminophenol to give 10 in a low yield (28%). This direct substitution failed
altogether in the case of the corresponding thiadiazol, necessitating the use of 4ꢀmethoxyaniline in
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
the nucleophilic aromatic substitution followed by demethylation to give 11 (6% from 48). The
triazole 12 and 1,2,4ꢀoxadiazole 13 were also obtained from benzoyl chloride 46, involving initial
conversion to the benzoyl isothiocyanate 49 followed by reaction with 4ꢀaminophenol and MeI to
23
give a Sꢀmethylthiourea 50 and cyclocondensation with hydrazine and hydroxylamine to give
that gives 12 (36%) and 13 (21%), respectively. Preparation of the alternative 1,2,4ꢀoxadiazole 14
involved conversion of 4ꢀchlorobenzonitrile (51) into the amidine oxime 52, followed by
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
24
cyclocondensation with trichlororacetic anhydride to give 53 (71%), which undergoes
nucleophilic aromatic substitution with 4ꢀaminophenol, displacing the trichloromethyl moiety to
give 14 (10%). Thiadiazole 15 was obtained from selective substitution of the
bromochlorothiadiazole 54 (Scheme 2). Chemoselective Pdꢀmediated Stille coupling of 54 with the
aryl stannane 55 selectively substitutes the activated chloro group to give the bromothiadiazole
25
5
6, this is subject to nucleophilic aromatic substitution with 4ꢀmethoxyaniline and demethylation
to yield 15 (21%). The isomeric thiadiazole 16 was obtained by sequential treatment of the 1ꢀ
isothiocyanatoꢀ4ꢀmethoxybenzene (57) with sodium cyanogen and bromine, to yield the 3ꢀbromoꢀ
2
6
1
ꢀ2,4ꢀthiadiazole 58 (70%). This material was Bocꢀprotected, coupled to stannane 55 and then
deprotected and demethylated to give 16 (27%). Oxazole 17 was prepared from αꢀbromoketone 59,
initial displacement of the bromo group with NaN and reaction with PPh gives an azaꢀWittig
3
3
reagent (not shown) that reacts with the isothiocyanate 57 to give oxazole 60 (76%), which is
demethylated to give 17 (65%).
18
ACS Paragon Plus Environment

Page 19 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
H
Y
X
N
O
X
N
a, b, c
d, e, f
Cl
Cl
Cl
Cl
Cl
N
N
N
OH
Cl
4
4
6
47 X = O, Y = Br 70% (from 46)
8 X = S, Y = Br 53% (from 46)
d) gives 10 28%
e), f) gives 11 6% (from 48)
4
g)
H
HN
S
OH
N
N
O
h, i
j, k
N
O
Cl
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
N
X
OH
NCS
9
j) gives 12 X = NH 36%
f) gives 13 X = O 21%
5
0 45% (from 46)
H
CCl3
N
NH
N
N
N
N
l
n
m
Cl
CN
Cl
Cl
Cl
O
O
OH
NHOH
5
1
52
53 (48% from 51)
14 (10%)
H
Br
N
Br
N
S
N
N
o
p, f
Cl
Cl
Cl
N
N
OH
N
S
54
S
Cl
SnMe3
56 69%
15 21% (from 56)
5
5
H
H
N
N
N
N
q, r
s, t, u, f
SCN
Cl
OMe
O
Br
Cl
S
OMe
S
N
OH
5
5
N
5
7
9
58 70% (from 57)
16 27% (from 58)
H
H
N
N
N
N
v, w
f
Cl
Cl
5
7
O
OMe
O
OH
Br
5
60 76% (from 59)
17 65%
Scheme 2: a) H NC(X)NHNH Cl (X = O or S), NaHCO , AcONa, H O. b) H SO , rt (X = S);
2
3
3
2
2
4
POCl , toluene, 140°C (X = O). c) isopentylnitrite, CuBr , MeCN, rt. d) 4ꢀAminophenol, DIPEA,
3
2
DMF, 70 °C. e) 4ꢀMethoxyaniline, DIPEA, DMF, 105 °C. f) BBr
acetone, rt. h) 4ꢀAminophenol, acetone, reflux. i) MeI, K CO , THF, 0 °C. j) N
reflux. k) NH OH.HCl, DBU, EtOH, reflux. l) NH OH.HCl, Et N, EtOH, reflux. m) trichloroacetic
anhydride, 110 °C. n) 4ꢀAminophenol, DBU, DMSO, rt. o) Pd(PPh ) 5 mol%, CuTC 10 mol%,
3
, CH
2
Cl
2
. g) NH
4
.SCN,
2
3
2
H
4
.H
2
O, EtOH,
2
2
3
3
4
5
5, 1,4ꢀdioxane, reflux. p) 4ꢀmethoxyphenol, DIPEA, NMP, microwave: 160 °C. q) NaNHCN,
acetone, rt. r) Br , EtAc, rt. s) Boc O, DMAP, 1,4ꢀdioxane, rt. t) Pd(PPh) 5 mol%, 55 1,4ꢀ
dioxane, reflux. u) TFA, CH Cl , rt. v) NaN , acetone, rt. w) 57, PPh , dioxane, 100 °C.
2
2
4
2
2
3
3
The two hybrid structures of 1 and the Genzyme SK1 inhibitor 3, compounds 18 and 19, were
prepared as described in Scheme 3. Preparation of 18 commenced with the 4ꢀ(bromomethyl)ꢀ
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
3
benzonitrile, the bromo group was substituted for an azide (NaN ) and the nitrile converted into an
amidine oxime 62 (65%) which was cyclocondensed with 4ꢀchlorobenzoyl chloride (46) to give
27
the 1,2,4ꢀoxadiaxole 63 (98%). Reduction of azide 63 into an amine and coupling with the Bocꢀ
protected hydroxyproline 64 and deprotection gave 18 (44%). Its oxadiazole isomer 19 was
prepared by cyclocondensation of the acid 65 (activated as a mixed anhydride) with the amidine
oxime 52 followed by deprotection to give 66 (36%), which is then coupled to the Bocꢀprotected
hydroxyproline 64 and deprotected to give 19 (87%).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 3: a) NaN
3 2 3
, DMF, rt. b) NH OH.HCl, Et N, EtOH, reflux. c) 46, pyridine, xylene, reflux.
d) NH .HCO , Zn, MeOH, rt. e) DIPEA, HATU, DMF, rt. f) EtOAc, HCl in dioxane. g) Nꢀ
4
2
methymorpholine, isobutylchloroformate, dioxane, rt, then 52 and reflux. h) HATU, DIPEA,
EtOAc.
SAR studies directed to finding alternatives to the 4ꢀaminophenol group were all based on
analogues of 1,3,4ꢀoxadiazole ring system 10. Direct access to systems in which the amino linker
unit (L) was absent (diarylꢀ1,3,4ꢀoxadiazoles) could be attained through cyclocondensation of the
tetrazole 67 with benzoyl chlorides 68 and 69 to give 70 (87%) and 71 (28%) followed by
demethylation and deprotection to provide test compounds 20 (91%) and 25 (96%), respectively
(Scheme 4). The bromoꢀoxadiazole 47 also served as useful substrate for Suzuki coupling to give
20
ACS Paragon Plus Environment

Page 21 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
24 (96%) and through nucleophilic substitution with amines to give 35 (29%) and 42 (43%)
(Scheme 5).
Many of the test compounds were achieved through cyclocondensation of the hydrazine amide 75
with different electrophiles (Scheme 6). Cyclocondensation of 75 with chloroacetic acid or ethyl
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
2
8
29
oxalyl chloride in POCl
which could be converted into test compounds 22, 23 and 41 by nucleophilic substitution (Scheme
). Hydrolysis and decarboxylation of the ester in 77 afforded the C5ꢀunsustituted oxadiazole 82,
which was coupled to 4ꢀiodobenzonitrile to give 83 (75%) using a previously described Cuꢀ
3
delivered the functionalised oxadiazoles 76 and 77 , respectively,
6
3
0
mediated CꢀH activation process. Cyclocondensation of the hydrazine amide 75 with CS gave
2
3
1
thiol 78, which was used in a Cuꢀmediated coupling with 4ꢀiodoanisole to give sulfide 21 (60%).
75 was also cyclocondensed with benzoyl chloride 79 and 4ꢀisothiocyanatobenzonitrile (80) to
give the carboxylic acid 26 (83%) (after ester hydrolysis) and nitrile 81 (67%), respectively.
The carboxylic acid 26 and the two nitriles 81 and 83 were used in the synthesis of other test
compounds (Schemes 7 and 8). The carboxylic acid 26 was converted to the simple amide 27
(
43%) and the proline derivative, amidine 30 (10% over 3 steps) (Scheme 7). Nitrile 83 was
converted to the tetrazole 31 (71%), by cycloaddition of NaN (Scheme 8). Both nitriles 83 and 81
were converted into amidine oximes 29 (52%) and 34 (66%) and amidines 28 (66%) and 33 (27%)
upon addition of LiHMDS and NH OH, respectively (Scheme 8). Hydration of 81 gave the simple
3
2
amide 32 (94%) and DIBAL reduction afforded an aldehyde (not shown) that gave the aldimine
oxime 36 (20%) upon Schiff’s condensation with NH OH.
2
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
3 2 2
Scheme 4: a) Pyridine, toluene, reflux. b) BBr , CH Cl . c) HCl in dioxane, EtOAc
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3 4 2 3
Scheme 5: a) Pd(PPh ) 10 mol%, Na CO , DMF, 80 °C. b) DIPEA, DMF, heat. c) HCl in
dioxane, EtOAc
22
ACS Paragon Plus Environment

Page 23 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
R = 4-chlorophenyl
Cl
O
R
O
N
O
R
OEt
N
N
N
77
7
6
b
a
R
O
N
O
c
SH
Cl
N
HN NH2
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
78
75
O
SCN
CN
g
CO2Me
Cl
79
8
0, 49%
d, e, f
H
N
O
R
O
N
O
R
N
CN
N
N
OH
81
2
6 83% (from 75)
OH
ClH N
2
h, i, j
k, l
O
R
7
6
N
N
23 25% (from 76)
O
OH
O
R
7
7
N
H
N
N
22 21% (from 77)
O
m, j
O
7
7
N
R
H
NHCl
N
N
41 19% (from 77)
n
O
o
CN
77
R
O
R
N
N
N
N
82 100%
83 75%
p, i
S
O
R
7
8
N
N
OH
1 60% (from 78)
2
Scheme 6: a) Chloroacetic acid POCl
3
3 2
, reflux. b) Ethyl oxalyl chloride, then POC , 100 °C. c) CS ,
3 2
KOH, MeOH, reflux. d) 79, pyridine, rt.e) POCl , reflux. f) LiOH, THF/H O, then HCl(aq). g)
THF, rt, then tosyl chloride, pyridine, reflux. h) 4ꢀMethoxyaniline, DMSO, K CO , rt. i) BBr ,
2
3
3
CH
rt. m) 74, EtOH, reflux. n) LiOH, THF/H
CuI 20 mol%, 1,10ꢀphenathroline 40 mol%, Cs
phenathroline 10 mol%, K CO , DMF, 120 °C.
2
Cl
2
. j) EtOAc, HCl dioxane. k) LiOH, THF/H
2
O, rt. l) 4ꢀAminophenol, HATU, Et
3
N, DMF,
2
O, rt, then HCl(aq) pH = 1ꢀ2, rt. o) 4ꢀIodobenzonirile,
CO , DMSO, 100 °C. p) CuI 5 mol%, 1,10ꢀ
2
3
2
3
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 7: a) Oxalyl chloride, CH
2 2 4 3
Cl , DMF cat., NH OH, rt. b) 84, HATU, Et N, DMF, rt. c)
LiHMDS, THF, rt. d) EtOAc, HCl dioxane.
24
ACS Paragon Plus Environment

Page 25 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 8: a) NH
4
Cl, NaN
3
, DMF, 100 °C. b) NH
2
OH.HCl, Et
3
N, EtOH, reflux. c) NaOMe,
O, DMAP 2 mol%,
OH.Cl, Et N,
NH Cl, MeOH, 50 °C. d) EtOAc, HCl dioxane. e) H
4
2
SO , H O, rt. f) Boc
4
2
2
dioxane. g) LiHMDS, THF, HCl (aq). h) DIBAL, toluene, then HCl (aq). i) NH
3
3
MeOH, rt.
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 small number of analogues of 1 wherein the chloro substituent was removed or replaced was
also generated for this study (Scheme 9). These were prepared by reaction of the benzoyl
hydrazine amides 85 and 86 with the isothiocyanate 57 to give 87 (89%) and 88 (19%),
respectively. The iodo group in 87 was removed by hydrogenation to give 89 (73%) and coupled to
ethynylcyclohexene 91 followed by hydrogenation to give 90 (61%). Demethylation of 87ꢀ90
yielded the test compounds 37, 38, 39 and 40.
Scheme 9: a) THF, rt, then TosCl, pyridine, reflux. b) Pd/C, H
2
(g), 1 atm, MeOH / EtOAc, rt. c)
Cl
9
1, Pd(PPh Cl 5 mol%, CuI 3 mol%, DMF, Et N, rt. d) BBr , CH
3
)
2
2
3
3
2
2
8
Conclusion: This study delineates the SAR of the extensively utilized SK1/2 inhibitor 1 (SKIꢀII)
5
6
that is also known to induce SK1ꢀdegradation and inhibit Des1. SK1ꢀkinase inhibition and
induction of SK1ꢀdegradation appear to be independent effects, where different compounds give
rise to either of these effects though some compounds do both. This is consistent with the previous
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
proposal that compounds may either bind to the substrate (Sph) domain to inhibit SK1ꢀkinase
activity or to an allosteric site to promote proteasomal degradation of SK1 and that some
5c
compounds, such as 1 and 5, do both. However, while our data are consistent with this allosteric
model, other explanations for small molecule promoted SK1ꢀproteasomal degradation, including
offꢀtarget effects, remain a possibility. Importantly, our data do show that PC3 cell growth is not
inhibited by inhibition of SK1ꢀkinase activity or by SK1ꢀdegradation and that inhibition of Des1 is
the most likely source of the anticancer activity of 1 and it congeners. While SK2 inhibition in
isolation does not appear to exert a cell growth inhibitory effects (e.g. 31), the combination of Des1
and SK2 inhibition may have cooperative effects (e.g. 39), which also may be of relevance to
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
20
clinical candidate 2, but this cooperativity needs to be further verified. These findings should be
7
,8
borne in mind when interpreting findings made using 1 and 2 as probes for SK1/2 activity.
In the course of this study new lead compounds have been identified exhibiting different activities:
selective SK2ꢀinhibition by 31 (Table 2, entry 13), dual SK2/Des1 inhibition by 39 (Table 2, entry
2
1 and Fig 3); SK1/2 kinase inhibition by 42 (Table 2, entry 24); SK1ꢀdegradation by 19 (Fig 2);
and Des1 and PC3 cell growth inhibition by 34 (Table 2, entry 16 and Fig 3). These new leads
provide a basis for future studies which will focus on better understanding the source of each of
these activities and further optimizing them to create valuable probes and potential therapeutics.
Experimental
PC3 cell viability assays
Routine cell culture: PC3 prostate cancer cell lines were cultured in DMEM (containing 10% fetal
2
calf serum and penicillinꢀstreptomycin). Cells were grown at 37 °C with 5 % CO and passaged
when 80ꢀ90% confluent 4 times before use. Cells were harvested by trypsin treatment (5 min) then
quenched with an equal volume of serum containing media and the cell suspension then
centrifuged at 200 xg for 5 min and the pellet resuspended in 5 mL of media. Cells were exposed
26
ACS Paragon Plus Environment

Page 27 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
to Trypan blue (excludes dead cells) and counted with a haemocytometer. Before treatment with
drug compounds, cells were plated at 2,500 cells/well in 96 well plates and incubated at 37 °C with
5
1
% CO
2
in a humidified incubator for 24 h. Drug stock solutions (50 or 10 mM) were diluted x
000 in media to a final concentration of either 50 µM or 10 µM with a DMSO vehicle
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
concentration of 0.1%. Compounds were then serially diluted in media (containing 0.1% DMSO)
to give 8 final concentrations, all at 0.1% DMSO. Cell culture supernatants were aspirated and
replaced with drug containing media. Drug treatments were performed in duplicate wells, while
potential plate layoutꢀspecific variation in cell growth was accounted for by addition of a vehicle
control (0.1% DMSO). An untreated control (media only) and active compound control (50 µM 1)
was included in each assay. Cells were then incubated with drug compounds at 37 °C with 5 %
2
CO in a humidified incubator for 72 h prior to assay. Cell media were diluted with CellTitre
AQueous One Solution to produce a final concentration of 317 µg/mL. Cell culture supernatants
were then aspirated from wells and replaced with 100 µL of CellTitre solution. Triplicate cellꢀfree
control wells containing only CellTitre solution were also included in each assay. Cells were then
incubated at 37 °C with 5 % CO in a humidified incubator for 1 h at which time absorbance was
2
read at 490 nm with a microplate reader. When analysing data, background absorbance (taken from
cellꢀfree control wells) was subtracted from each reading. To determine percentage inhibition of
cell viability, absorbance readings for each drug treatment were expressed as a fraction of the
vehicle control (0.1% DMSO) readings. For each drug concentration the mean (± SEM) is
calculated and a sigmoidal curved is fitted to the data and used to calculate the IC50 of each
compound.
Sphingosine kinase-1/2 Activity Assay
32
The SphK assays employed, measures SK activity through the production of Pꢀlabelled S1P
32
11
following the addition of exogenous Sph and [
γ P] ATP as described previously.
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Activities of SK1 and SK2 were determined using identical assay conditions with the exception
that SK1 assays used 3ng/assay recombinant Hisꢀtagged human SK1, while SK2 assays used
12
3
0ng/assay recombinant human SK2.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Compounds were dissolved in DMSO at 10 mM, the mixture was vortexed (and sonicated if
necessary). Stock solutions were kept at 4 °C until used. Aliquots were made containing 100 µM
drug compound. These aliquots were then diluted in assay buffer: 100mM Tris/HCl buffer (pH 7.4)
containing, 100 mM NaCl, 1 mM sodium orthovanadate, 10 mM NaF to make up a 10 µM
solution.
Measurement of dihyrdoceramide desaturase-1 (Des1) activity:
Measurement of Des1 activity was performed by HPLC using intact Jurkat cells labeled with
14
DhCerꢀC6ꢀNBD as described previously with modifications to enhance sensitivity and
reproducibility. These modifications included the use of parental Jurkat cells, 0.5% serum in the
culture media, and cell harvesting via centrifugation at 500 xg to maximise ceramide extraction.
Extracted samples (50 ꢁl) were analysed on a Waters HPLC coupled to a fluorescence detector
2
using a 30cm C18 reverseꢀphase column eluted with 1 ml/min 20% H O and 80% acetonitrile with
0.1% trifluoroacetic acid. NBDꢀlabelled substrate and product were quantitated with excitation and
emission wavelengths of 465 nm and 530 nm, respectively.
Lipidomics method
In vitro cultured PC3 cells were exposed to test compounds (10 µM) or vehicle control (0.1%
DMSO) for five hours. Cells were lifted with trypsin, quenched and washed with cold saline (4
6
°C), counted and adjusted to 10 cells per sample. Lipids were extracted from cell pellets by
addition of 200 µL chloroform/methanol/water (1:3:1) and three freezeꢀthaw cycles. Protein and
cellular debris was removed by centrifugation and samples stored at ꢀ20 °C for <1 week until LCꢀ
MS analysis.
28
ACS Paragon Plus Environment

Page 29 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Lipid analysis utilized LCꢀMS with reversed phase chromatography (Dionex Ultimate RSLC3000;
Thermo) and high resolution mass spectrometry (QꢀExactive; Thermo). 10 µL sample was injected
onto an Ascentis Express C8 column (5cm x2.1mm, 2.7ꢁm; Supelco Anaytical), with 30 minute
gradient elution. Mobile phase A consisted of 40% Isopropanol, 60% H O, 2 mM formic acid, 8
2
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
mM ammonium formate. Mobile phase B: 2% H O in isopropanol, 2 mM formic acid, 8 mM
2
ammonium formate. The gradient ran from 0% B to 20% B in 1.5min, to 28% B (7 min) to 35% B
(
8 min) to 65% B (24 min) at a flow rate of 200 ꢁL/min, and increased to 100% B at 500ꢁL/min
25ꢀ27 min), and returned to 100% A (29 min). The mass spectrometer was operated in positive
(
and negative mode (rapid switching) electrospray ionization with a mass resolution of 140,000 and
mass accuracy < 2ppm. Dataꢀdependent high resolution (17,500) MSMS was performed on pooled
extracts to confirm lipid identifications.
3
2
Lipidomics data was analysed using the IDEOM software to provide relative quantification of all
detected features. Lipids and other metabolites were identified according to accurate mass, with
additional manual curation of significant lipids based on retention time and MSMS fragmentation
spectra. Ceramides and other sphingolipids were quantified by targeted extraction of peak areas for
extracted ion chromatograms using the Tracefinder software (Thermo).
Soft agar assay:
To test clonogenic potential and anchorage independent growth of PC3 cells, we performed colony
21
formation assay as previously described. Briefly, in a 6ꢀwell plate, 5,000 filtered PC3 cells were
added in the presence of 0.35% low melting point agarose onto a bed of 0.7% low melting point
agarose with RPMI media containing 10% Fetal Bovine Serum (FBS) final concentration and
allowed to solidify as a single cell suspension at 4 °C for 4 min before incubation at 37 °C in a
2
high CO cell culture incubator. Compounds were diluted in RPMI media and added to cells 24 h
after incubation. Plates were incubated at 37 °C for 21 days. Subsequently, media were removed
29
ACS Paragon Plus Environment