Antitumorigenic Properties of Omega-3 Endocannabinoid Epoxides

Article abstract of DOI:10.1021/acs.jmedchem.8b00243

Accumulating studies have linked inflammation to tumor progression. Dietary omega-3 fatty acids, such as docosahexaenoic acid (DHA), have been shown to suppress tumor growth through their conversion to epoxide metabolites. Alternatively, DHA is converted enzymatically into docosahexaenoylethanolamide (DHEA), an endocannabinoid with antiproliferative activity. Recently, we reported a novel class of anti-inflammatory DHEA-epoxide derivative called epoxydocospentaenoic-ethanolamide (EDP-EA) that contain both ethanolamide and epoxide moieties. Herein, we study the antitumorigenic properties of EDP-EAs in an osteosarcoma (OS) model. First, we show ～80% increase in EDP-EAs in metastatic versus normal lungs of mice. We found significant differences in the apoptotic and antimigratory potencies of the different EDP-EA regioisomers, which were partially mediated through cannabinoid receptor 1 (CB1). Next, we synthesized derivatives of the most pro-apoptotic regioisomer. These derivatives had reduced hydrolytic susceptibility to fatty acid amide hydrolase (FAAH) and increased CB1-selective binding. Collectively, we report a novel class of EDP-EAs that exhibit antiangiogenic, antitumorigenic, and antimigratory properties in OS.

Full text of DOI:10.1021/acs.jmedchem.8b00243

Subscriber access provided by University of Winnipeg Library
Article
Anti-Tumorigenic Properties of Omega-3 Endocannabinoid Epoxides
Jahnabi Roy, Josephine E Watson, Insup Hong, Timothy M. Fan, and Aditi Das
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b00243 • Publication Date (Web): 01 Jun 2018
Downloaded from http://pubs.acs.org on June 1, 2018
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 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 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.
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 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
AntiꢀTumorigenic Properties of Omegaꢀ3
Endocannabinoid Epoxides
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
Jahnabi Roya, Josephine Watsonb, Insup Hong†, Timothy M. Fan‡ and Aditi Das†,b,c*
†Department of Comparative Biosciences, ‡Department of Veterinary Clinical Medicine a
Department of Chemistry, b Department of Biochemistry; c Beckman Institute for Advanced
Science, Division of Nutritional Sciences, Neuroscience Program and Department of
Bioengineering, University of Illinois Urbana–Champaign, Urbana IL 61802, USA
Correspondence to: Aditi Das, aditidas@illinois.edu
KEYWORDS
Osteosarcoma, Omegaꢀ3 fatty acids, endocannabinoids, cannabinoid receptor
ABSTRACT
Accumulating studies have linked inflammation to tumor progression. Dietary omegaꢀ3 fatty
acids including docosahexaenoic acid (DHA) have been shown to suppress tumor growth
through their conversion to epoxide metabolites. Alternatively, DHA is converted enzymatically
into docosahexaenoylethanolamide (DHEA), an endocannabinoid with antiꢀproliferative activity.
Recently, we reported a novel class of antiꢀinflammatory DHEAꢀepoxides (EDPꢀEAs) that
contain both ethanolamide and epoxide moieties. Herein we evaluate the antiꢀtumorigenic
properties of EDPꢀEAs in an osteosarcoma model. First, we show ~80% increase in EDPꢀEAs in
metastatic lungs versus normal mouse lungs. We found significant differences in the apoptotic
and antiꢀmigratory potency of the different EDPꢀEA regioisomers, which are partly mediated
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 39
1
2
3
4
5
6
7
8
9
through cannabinoid receptor 1 (CB1). Furthermore, we synthesized derivatives of the most proꢀ
apoptotic regioisomer. These derivatives had reduced hydrolytic susceptibility to fatty acidꢀ
amide hydrolase and increased CB1 binding. Collectively, we report a novel class of EDPꢀEAs
that exhibit antiꢀangiogenic, antiꢀtumorigenic and antiꢀmigratory properties in osteosarcoma.
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
INTRODUCTION
Recent studies in cancer have focused on understanding the role of inflammation in tumor
initiation and malignant progression. It has been shown previously that inflammatory conditions
can promote oncogenic transformation. In addition, the carcinogenic milieu can also generate an
1ꢀ3
inflammatory microenvironment, which favors tumor progression.
Dietary omegaꢀ3 fatty
acids such as eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) have been shown to
suppress tumor growth and progression through their conversion to antiꢀinflammatory, proꢀ
4 5 6 7
resolving and antiꢀtumorigenic lipids.
As inflammation can promote tumor initiation,
proliferation, and progression, it is hypothesized that omegaꢀ3 fatty acid derived lipid
5
metabolites prevent tumor formation and metastasis in a multifaceted manner. Hence, there is
biological and clinical justification to discover if potential omegaꢀ3 fatty acid derived lipid
metabolites possess antiꢀtumorigenic and antiꢀinflammatory properties.
Cytochrome P450 (CYP) epoxygenases convert omegaꢀ3 fatty acids such as EPA and DHA into
epoxides that exhibit antiꢀcancer properties. Specifically, the EPAꢀepoxides are antiꢀ
proliferative.8 In addition, the terminal epoxide of EPA, 17,18ꢀepoxyeicosatetraenoic acid
(17,18ꢀEEQ), decreased endothelial cell proliferation, interrupted the cell cycle in Sꢀphase and
downꢀregulated the cyclin D1/cyclinꢀdependent kinase (CDK)ꢀ4 complex. The corresponding
DHA epoxide, epoxydocosapentaenoic (EDP) inhibits metastasis and tumor growth.9
2
ACS Paragon Plus Environment

Page 3 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
Furthermore, the action of the DHA and EPA epoxides are attenuated in the absence of a soluble
epoxide hydrolase (sEH) inhibitor, indicating that the functional epoxide moiety is key for their
antiꢀtumorigenic properties.
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
Omegaꢀ3 fatty acids such as DHA and EPA are also converted enzymatically into
docosahexaenoyl ethanolamide (DHEA) and eicosapentaenoyl ethanolamide (EPEA) through the
Nꢀacyl ethanolamine synthesis pathway similar to the conversion of arachidonic acid to
10, 11
arachidonoyl ethanolamide (AEA, anandamide).
These ethanolamide derivatives are
collectively termed as endocannabinoids (eCBs) as they elicit similar physiological effects as
cannabinoids found in cannabis. The endocannabinoid system consists of the biosynthetic and
12
degradative enzymes, including diacylglycerol lipase and fatty acid amide hydrolase,
the
eCBs, anandamide and 2ꢀarachidonylglycerol, and the cannabinoids receptors 1 and 2 (CB1 &
13, 14
CB2).
While, CB1 is predominantly found in the central nervous system (CNS), CB2 is
found in immune cells. 15 These receptors interact with various eCBs to exert their physiological
16
activities. Interestingly, these receptors interact with the active components of cannabis, and
are responsible for cannabis’ psychoactive and antiꢀinflammatory properties. Phytocannabinoids,
ꢁꢀ9ꢀTHC and cannabidiol (CBD), are two specific components of cannabis that have been
shown to induce apoptosis in hepatocellular carcinoma, breast cancer and colon cancer.17 As a
whole, the phytocannabinoids and eCBs have been shown to play a powerful role in cancer cell
proliferation, apoptosis, inflammation and immunomodulation.18, 19
The most studied eCBs in the human body are AEA and 2ꢀarachidonylglycerol (2ꢀAG) that are
derived from the omegaꢀ6 fatty acid arachidonic acid (AA). Anandamide was shown to be antiꢀ
20ꢀ23
tumorigenic in breast cancer, colorectal cancer as well as osteosarcoma.
2ꢀAG has been
shown to induce apoptosis in prostate cancer. 24 Omegaꢀ3 derived eCBs, DHEA and EPEA have
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 39
1
2
3
4
5
6
7
8
9
shown greater antiꢀproliferative activity than their parent counterparts in prostate cancer cells.11
Additionally, it was also shown that the activity of eCBs in cancer cells was increased in the
presence of a fatty acid amide hydrolase (FAAH) inhibitor, indicating that the ethanolamide
endocannabinoids, AEA DHEA and EPEA contributes to the increased proꢀapoptotic and antiꢀ
proliferative activity. 10, 11
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 showed that a novel class of eCB epoxides are generated from the crossꢀtalk of the
25
NAPEꢀPLD pathway and CYP epoxygenase pathways that were found in multiple tissues
(Figure 1a). These biꢀfunctional metabolites contain both ethanolamide and epoxide motifs that
function through different signaling pathways leading to several overlapping physiological
outcomes. Both the eCB and epoxyeicosanoid pathway have been implicated in cancer as
8, 9
described above.
Hence, we wanted to evaluate their antiꢀtumorigenic properties of dual
functional eCB epoxides in a cancer model. These lipid metabolites termed as EPEA epoxide
(a.k.a. epoxyeicosatetraenoicꢀethanolamide (EEQꢀEA)) and DHEA epoxide (a.k.a.
epoxydocosapentaenoicꢀethanolamide (EDPꢀEA)) are synthesized from EPEA and DHEA by
cytochrome P450s and are antiꢀinflammatory and vasodilatory in nature. More importantly, the
DHEA epoxides are widely distributed throughout the body. Additionally, herein we show that
there is an ~80% increase in these levels in metastatic lungs (Figure 1b and figure S1) as
compared to lungs of normal mice. This indicates that they might play a physiological role in
tumorigenesis and cancer metastasis.
Herein, we investigate the antiꢀtumorigenic activities of the different regioisomers of DHEA
epoxides on osteosarcoma cells. Osteosarcoma (OS) is the most common primary bone cancer in
humans and presents high rates of invasion and metastasis. The cannabinoid receptor agonist
WINꢀ55212 has been shown to potentiate the antiꢀtumorigenic properties of drugs like
4
ACS Paragon Plus Environment

Page 5 of 39
Journal of Medicinal Chemistry
1
2
3
4
26
adriamycin in OS.
Additionally, the administration of cannabinoid receptor agonists have
5
6
7
8
shown to reduce bone loss in bone tumor models and cannabinoid receptor mediated nociceptive
pain. 27, 28 Taken together, herein we will show that DHEA epoxide regioisomers and their stable
derivatives will exhibit antiꢀtumorigenic activities in osteosarcoma and will potentially reduce
bone cancer pain through receptor selective activation.
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 examine the antiꢀtumorigenic properties of DHEA epoxides, we will first test the
apoptotic potential of the different regioisomers of DHEA epoxides using Annexin V staining
assay in three different OS cell lines ꢀ metastatic OS cell lines MG63 and 143B and the nonꢀ
metastatic cell line HOS, and identified the most potent DHEA epoxide regioisomer. We further
tested the antiꢀmigratory potential of the DHEA epoxides via a wound healing assay in all three
cell lines.29 Interestingly, we observed significant differences between the different DHEA
epoxide regioisomers tested with respect to their proꢀapoptotic and antiꢀmigratory properties that
were uniform across the different OS cell lines chosen for the studies indicating that the specific
isomers of these lipid metabolites triggered common signaling pathways.
Furthermore, we synthesized stable derivatives of the most proꢀapoptotic regioisomer of DHEA
epoxide (10,11ꢀEDPꢀEA) to reduce its hydrolytic susceptibility to fatty acidꢀamide hydrolase
(FAAH) and increase binding to cannabinoid receptor 1 (CB1). The the most potent regioisomer
and its stable derivatives were further tested towards their antiꢀtumorigenic, antiꢀmigratory and
FAAH hydrolysis.
Collectively, we report a novel class of DHEAꢀepoxides and their stable derivatives that exhibit
antiꢀinflammatory,30 antiꢀangiogenic, antiꢀtumorigenic and antiꢀmigratory properties and by
virtue of their ability to bind cannabinoid receptors might reduce nociceptive bone pain in
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 39
1
2
3
4
5
6
osteosarcoma. As osteosarcoma is accompanied by inflammation and pain, these molecules
could aid in alleviating both.
7
8
9
RESULTS
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
DHEA epoxides are endogenously produced in osteosarcoma metastatic lung tissues. We
developed a targeted lipidomics method using LCꢀMS/MS in multiple reaction monitoring
(MRM) mode to identify and quantitate DHEA epoxides (EDPꢀEA) regioisomers (19,20ꢀ, 16,17ꢀ
30
, 13,14ꢀ, 10,11ꢀ, and 7,8ꢀ).
In order to quantify basal levels of endogenous EDPꢀEA
regioisomers in metastatic and nonꢀmetastatic lungs, we measured these metabolites in mice with
osteosarcoma. Next, we used a metastatic tumor model to show. Specifically, 106 K7M2 cells
(murine OS cells) were prepared in HBSS and intravenously injected into the tail vein of 6−8
week old female BALB/c mice in a 200 ꢂL volume. Lungs from ten mice, five mice with lungs
free of tumor and five with lung metastasized osteosarcoma tumors were isolated and the total
endocannabinoids were extracted and quantified by LCꢀMS/MS as described in materials and
methods. As indicated in figure 1b and figure S1, we can observe that both tumorꢀfree lungs in
mice where no OS was induced, and lungs with tumors produced eCB epoxides with several fold
difference compared to DHEA, the parent molecule. There was ~87% increase in the terminal
DHEA epoxide (19,20ꢀEDPꢀEA), 66% increase in 16,17ꢀEDPꢀEA, 72% increase 13,14ꢀEDPꢀ
EA, 49% increase 10,11ꢀEDPꢀEA in lungs with metastatic tumor burden, while DHEA was not
different in lung tissue with metastatic burden as compared to healthy lungs with no tumors. The
high increase of these metabolites in the metastasized lungs warranted further investigation
whether they were promoting tumor or they exhibit antiꢀtumorigenic properties. Previously, we
have shown that these metabolites are antiꢀinflammatory. 30
6
ACS Paragon Plus Environment

Page 7 of 39
Journal of Medicinal Chemistry
1
2
3
4
In order to investigate the effects of eCB epoxides on osteosarcoma, all five regioisomers were
5
6
25
synthesized as described previously.
Briefly, mCPBA was used to oxidize DHA, and the
7
8
9
epoxide regioisomers were purified using reverse and normal phase HPLC. The individual
regioisomers were coupled to ethanolamine and purified by reverse phase chromatography. The
compounds were quantified using synthesized standards and used for further studies.
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
Effect of DHEA epoxides (EDP-EA) regioisomers on cell viability.
To evaluate the effect of EDPꢀEA isomers on cellular viability of osteosarcoma cells, cell titer
blue (CTB) assay was performed on 143B metastatic OS cells at 5 ꢂM concentration. The cell
titer blue assay measures cellular viability by measuring the conversion of dye reazurin into
fluorescent end product resorufin by living cells. The fluorescence produced is directly
proportional to the viable cells. As seen in Figure 1c only 13,14ꢀ, 10,11ꢀ and 7,8ꢀ EDPꢀEA
reduced total cell viability at 12.5 ꢂM concentration and 19,20ꢀEDPꢀEA and 16,17ꢀEDPꢀEA
were less effective. All regioisomers were tested for their effect on 143B cell’s viability by cell
titer blue (CTB) assay at various concentrations as seen in figure S2. However, at 25 ꢂM
concentration, complete reduction of cell viability was observed for all EDPꢀEA isomers. Since
7,8ꢀ, 10,11ꢀ and 13,14ꢀ were more effective at reducing cell viability than the other two
regioisomers at lower concentrations, their apoptotic and antiꢀmigratory potential was further
evaluated. In order to evaluate if the decrease in cell viability of the three different DHEA
epoxide regioisomers is due to cell death or check in cell migration, we performed apoptosis and
cell migration assay.
Induction of apoptosis in metastatic OS cells by DHEA epoxides (EDP-EA) regioisomers
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 39
1
2
3
4
5
6
7
8
9
The apoptotic potential of 7,8ꢀEDPꢀEA, 10,11ꢀ EDPꢀEA and 13,14ꢀEDPꢀEA were tested in three
different cell lines ꢀ HOS, 143B and MG63 cells using the Annexin V/ PI staining assay at 12.5
ꢂM concentration. This concentration was chosen to reflect a point in the CTB assay where cell
viability is reduced but not completely diminished. The Annexin V/ PI binding assay measures
apoptotic cells by binding labeled annexin V to phosphatidylserine (PS) species. PS is an inner
leaflet lipid that flips to the outer leaflet of the plasma membrane in apoptotic cells. The
fluorescence by annexin V is proportional to apoptotic cells. PI binds to DNA in necrotic cells. A
signal from both annexin V and PI indicate cells in late apoptosis. As seen in Figure 1d, all three
regioisomers induced apoptosis in all the cell lines at a concentration of 12.5 µM. Specifically, in
143B cell line (Table ST3), 7,8ꢀEDPꢀEA induced 8.75 ±0.1% apoptosis (p=0.001). 10,11ꢀEDPꢀ
EA induced 17.92 ± 3.3% (p=0.05) apoptosis and 13,14ꢀ EDPꢀEA induced 9.9 ± 1.9% apoptosis
(p=0.09). Similar to these results, in MG63 cells (Table ST4), 7,8ꢀEDPꢀEA induced 16.4 ±1.4%
apoptosis (p=0.011). 10,11ꢀEDPꢀEA induced 26.6 ± 3.5% (p=0.027) apoptosis and 13,14ꢀ EDPꢀ
EA induced 16.9 ± 2.3% (p=0.04) apoptosis. In HOS cells (Table ST2), the trend followed and
7,8ꢀEDPꢀEA 17.1± 0.5% apoptosis (p=0.023 as compared to control). 10,11ꢀEDPꢀEA induced
31.1 ± 2.3% apoptosis (p=0.007) and 13,14ꢀ EDPꢀEA induced 19.6 ± 3.5% apoptosis (p=0.027).
In contrast to these results, treatment of all three different cell lines with AA, AEA, DHA and
DHEA at 12.5 µM concentration showed no induction of apoptosis as compared to no treatment
control (Table ST1).
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
DHEA epoxides inhibit cell migration of cancer cells determined using wound healing assay
The wound healing assay has been shown to be an effective measure for directional migration of
cells in vivo. 31 This assay measures the migration of cells across a scratch or wound over time as
an indicator of cell migratory potential. A wound healing assay was performed for 7,8ꢀEDPꢀEA,
8
ACS Paragon Plus Environment

Page 9 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10,11ꢀ EDPꢀEA and 13,14ꢀEDPꢀEA at 5 ꢂM. This concentration was chosen to reflect a
concentration where the compounds are not significantly proꢀapoptotic and the reduction in
wound healing would be due to the compound’s antiꢀmigratory properties only. Figure 1e is
representative of the effects of 7,8ꢀEDPꢀEA, 10,11ꢀ EDPꢀEA and 13,14ꢀEDPꢀEA on the wound
closure in 143B cells. Figure 1f represents the ratio of wound remaining in the treatment to the
wound remaining in no treatment control at three time points, 0h, 6h and 10h. Therefore a ratio
of 1 indicates that same percentage area of wound remains in treatment and control and thus
treatment is no different than control. A ratio higher than 1 indicates that the treatment has higher
percentage area of wound remaining and thus a compound that restricts migration. As seen from
the data in Figure 1f, in HOS cells, all three compounds inhibit cell migration to an extent. For
7,8ꢀEDPꢀEA (at 6h p=0.07 and at 10h p=0.35) and 13,14ꢀEDPꢀEA (at 6h p=0.5 and at 10h,
p=0.4) the difference in wound closure at 10 hours is not significant. For 10,11ꢀEDPꢀEA, the
wound closure at 10 hours is 1.5 times better than in control (at 6h, p=0.08 and at 10h, p=0.17).
In 143B cells, Figure 1f, 7,8ꢀEDPꢀEA and 13,14ꢀEDPꢀEA do not inhibit wound healing at 10
hours. However 10,11ꢀEDPꢀEA inhibits wound closure significantly (at 6h p=0.21 and at 10h
p=0.05). In MG63 cell lines, Figure 1f, all three isomers restrict wound healing and at 10 hours
and 10,11ꢀEDPꢀEA inhibited wound closure with 10,11ꢀEDPꢀEA being twice as much as the
control (at 6h p=0.38 and at 10h p=0.06). For 13,14ꢀEDPꢀEA the wound was restricted at 10h
but not at 6h (at 6h p=0.018 and at 10h p=0.015). For 7,8ꢀEDPꢀEA a similar trend was seen with
p= 0.23 at 6h and p=0.23 at 10h. Thus, while all three isomers in most conditions demonstrate
variable capacity to restrict wound healing and thereby show antiꢀmigratory potential, the
directional trend effect is greatest in 10,11ꢀEDPꢀEA.
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
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 39
1
2
3
4
5
6
7
8
9
From our apoptosis and wound healing assays, 10,11ꢀEDPꢀEA was the most potent regioisomer.
This was partly mediated through cannabinoid receptor 1 as the reduction of cell migration by
10,11ꢀEDPꢀEA in the 143B cell line was completely abolished using CB1 receptor antagonist
rimonabant (Supplementary Figure S4). 32
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
10
ACS Paragon Plus Environment

Page 11 of 39
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: (A) Synthesis of DHEAꢀepoxide regioisomersꢀ 19,20ꢀ(5), 16,17ꢀ(4), 13,14ꢀ(3), 10,11ꢀ
(2) and 7,8ꢀ EDPꢀEA (1). (B) The levels of endocannabinoid epoxides of DHEA and DHEA
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 39
1
2
3
4
5
6
7
8
9
epoxides in normal lungs and lungs of mice with tumor metastasis. Lungs in image were infused
with India ink and white lesions represent metastatic tumor burden (C) Percentage cell viability
as measured by cell titer blue assay by various regioisomers at 5 ꢂM, 12.5 ꢂM and 25 ꢂM. (D)
Apoptosis induced in HOS, 143B and MG63 cells at 12.5 ꢂM concentration of 13,14ꢀEDPꢀEA
(crisscross), 10,11ꢀ EDPꢀEA (checkered) and 7,8ꢀ EDPꢀEA (dots). (E) Wound healing scratch
shown at 0h and 10h in 143B cell lines by 13,14ꢀEDPꢀEA 10,11ꢀ EDPꢀEA and 7,8ꢀ EDPꢀEA at 5
ꢂM concentration. (F) Wound healing represented as a ratio of wound remaining in treatment
and wound remaining in control over 10 hours in three cell lines (i) HOS, (ii) 143B and (iii)
MG63 by 5 ꢂM concentration of 13,14ꢀEDPꢀEA (crisscross), 10,11ꢀ EDPꢀEA (checkered) and
7,8ꢀ EDPꢀEA (dots).
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
Synthesis of amide derivatives of 10,11-DHEA epoxide (10,11-EDP-EA)
The eCB epoxides are dual functional molecules. As the amide bond is susceptible to fatty acid
amide hydrolase (FAAH), we designed more stable analogs of 10,11ꢀEDPꢀEA, which shows
highest efficacy amongst the various epoxide derivatives of DHA. Derivatives were made at the
carboxylate end for different amide modifications, to reduce susceptibility to FAAH as well as
better receptor binding to CB1 and CB2, which are the purported receptors for apoptotic activity.
The current literature has mainly focused on anandamide or arachidonylethanolamine (AEA)
derivatives, since AEA is the biological ligand of CB1 and CB2 receptors and thus is likely to be
the receptor binding to DHEA and its analogs. It has been previously shown that introduction of
a methyl group at the 1’ position of the amide results in increased metabolic stability of AEA,
33
while affecting the CB1 binding ability . It was shown that the R isomer shows 4ꢀfold higher
CB1 binding and the S isomers shows 2ꢀfold lower binding as compared to AEA. Thus the Rꢀ1’ꢀ
methyl isomer was the first modification made to 10,11ꢀEDP as 10,11ꢀEDPꢀIA (figure 2a). Next,
12
ACS Paragon Plus Environment

Page 13 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the cyclopropyl derivative of AEA has been shown to have modest FAAH inhibition with an
IC50 value of 4.1 ±2.0 µM, with significantly increased CB1 binding 34 thus 10,11ꢀEDPꢀCA was
prepared. Finally, the nꢀpropyl derivative (10,11ꢀEDPꢀNA) was made to assess the effect of acyl
chain length on FAAH hydrolysis. It has also been shown to have better affinity for CB1 as
compared to AEA. 35, 36
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 synthesis of these analogs is outlined in the materials and methods section and shown in
Figure 2a similar to the synthesis of 10,11ꢀEDPꢀEA. Briefly, 10,11ꢀEDP was incubated with
NHS and EDC in acetonitrile for 20 mins at 37 °C, followed by additional of the corresponding
amine and stirring overnight at room temperature. The amines were purified by reverse phase
HPLC chromatography and quantified similar to 10,11ꢀEDPꢀEA.
Susceptibility to FAAH hydrolysis of 10,11-DHEA epoxides and its amide derivatives
Fatty acid amide hydrolase (FAAH) is a ubiquitously expressed protein in various mammals and
the sequence of the protein has shown to be highly conserved in humans, 37 mouse 37 and pig 38.
For our studies, we used pig brain membrane preparations to analyze the effect of FAAH on
39
10,11ꢀEDPꢀEA as well as the three amino derived modifications. The products of hydrolysis
were measured by LCꢀMS/MS as indicated in materials and methods section. As observed from
figure 2b, there was significant reduction of hydrolytic potential in all the three derivatives as
compared to 10,11ꢀEDPꢀEA (p<0.01). While the rate of hydrolysis of 10,11ꢀEDPꢀEA to 10,11ꢀ
EDP was 30.0 ± 0.3 pmol minꢀ1 mg proteinꢀ1, the rate of hydrolysis for the derivatives ꢀ 10,11ꢀ
EDPꢀNA, 10,11ꢀEDPꢀIA and 10,11ꢀEDPꢀCA was 0.09 ± 0.01 pmol minꢀ1 mg proteinꢀ1 0.99 ±
0.04 pmol minꢀ1 mg proteinꢀ1and 5.17 ± 0.06 pmol minꢀ1 mg proteinꢀ1 respectively. For 19,20ꢀ
EDPꢀEA, the rate is 150 pmol minꢀ1 mg proteinꢀ1. 25 Since the three derivatives show significantly
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 39
1
2
3
4
5
6
reduced hydrolytic susceptibility, they could have higher bioavailability and thus their apoptotic
and antiꢀmigratory potential was further evaluated.
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
Apoptosis induced by amino modified derivatives of 10,11 DHEA epoxide (10,11-EDP-EA)
All three aminoꢀmodified derivatives of 10,11ꢀEDP were tested for their apoptotic activity by
Annexin V/ PI binding assay. All three compounds showed increased apoptotic potential as
compared to the no treatment vehicle control (Figure 2c). The cyclopropyl analog had the highest
proꢀapoptotic potential among the three derivatives. In HOS, MG63 and 143B cells, it increased
apoptosis to 27.2 ± 4.1% (p= 0.036 as compared to control), 17.1 ± 4.4% (p= 0.078 as compared
to control) and 21.3 ± 1.9% (p=0.01) respectively. The isoꢀpropyl analog was the second most
effective in HOS and MG63 cells with apoptotic potentials of 20.2± 1.8% (p= 0.016 as compared
to control) and 17.5 ± 2.8% (p= 0.033 as compared to control) respectively. In 143B cells it was
only 18.8 ± 0.3% (p=0.0004) for the isoꢀpropyl derivative as compared to 25.7 ± 0.4%
(p=0.0002) for the nꢀpropyl derivative. However, the nꢀpropyl derivative was least effective in
HOS and MG63 cells with only 17.1 ± 1.4% (p= 0.017 as compared to control) and 15.2 ± 3.3%
(p= 0.06 as compared to control) respectively. Overall, all the derivatives showed similar proꢀ
apoptotic potential to the parent compound without discernable differences.
Wound healing assay to determine the potential to inhibit cell migration by amino modified
derivatives of 10,11 DHEA epoxide (10,11-EDP-EA)
A wound healing assay was used to analyze the antiꢀmigratory properties of the three derivatives.
The data in Figure 2d are represented as a ratio of the percentage area remaining in the wound by
compound and the control. As seen in Figure 2d, in all the cell lines, all the 10,11ꢀEDP
14
ACS Paragon Plus Environment

Page 15 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
derivatives restrict wound healing at 10 hours. Among the derivatives, the 10,11ꢀcyclopropyl
derivative shows significantly greater potency at the 6 and 10 hour time points inhibiting
migration by 4.5ꢀ 6 times more than vehicle control (at 6 hours p= 0.19 for 143B cells, p= 0.57
for HOS cells and at 10 h, p= 0.11 for 143B cells, p= 0.37 for HOS cells). Similarly as seen in
figure 2d, in the 143B and HOS cells, restriction in wound healing was 1ꢀ 1.9 times by 10,11ꢀ
EDPꢀEA (for HOS cells at 6h, p=0.08 and at 10h, p=0.17 and 143B cells at 6h, p=0.21 and at
10h, p=0.05), 1ꢀ 1.47 times by nꢀpropyl derivative (for HOS cells at 6h, p=0.39 and at 10h,
p=0.44 and 143B cells at 6h, p=0.4 and at 10h, p=0.0002), 1ꢀ 3.9 times by isopropyl derivative
(for HOS cells at 6h, p=0.58 and at 10h, p=0.99 and 143B cells at 6h, p=0.059 and at 10h,
p=0.002). Finally, in the MG63 cell line (figure 2d), the restriction in wound healing was 1ꢀ2
times by 10,11ꢀEDPꢀEA (at 6h p=0.38 and at 10h p=0.06) as compared to the control. The 10,11ꢀ
EDPꢀNA (at 6h, p=0.7 and at 10h, p=0.94) and 10,11ꢀEDPꢀIA (at 6h, p=0.57 and at 10h, p=0.34)
were slightly poorer than the control in restricting cell migration. 10,11ꢀEDPꢀCA (at 6h p= 0.12
and at 10h p=0.09 as compared to vehicle control and p>0.9 as compared to 10,11ꢀEDPꢀEA) was
comparable to 10,11ꢀEDPꢀEA for the MG63 cell line in restricting cell migration. As observed
from the various results, it was concluded that the 10,11ꢀEDPꢀCA was the most potent in
inhibiting wound healing in 143B and HOS cells and comparable to the parent molecule in
MG63 cells.
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
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 39
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 2: (A) Synthetic scheme for EDPꢀEA analogs (6ꢀ8). (B) Fatty Acid Amide Hydrolase
(FAAH) dependent hydrolysis of amide derivatives. (C) Apoptotic potential of 10,11ꢀEDPꢀEA
(checkered), 10,11ꢀEDPꢀNA (7) (dots), 10,11ꢀEDPꢀIA (8) (horizontal stripes) and 10,11EDPꢀCA
(6) (vertical stripes) as compared to control as measured by Annexin V apoptosis assay. (D)
Antiꢀmigratory properties of 10,11ꢀEDPꢀEA, 10,11ꢀEDPꢀNA, 10,11ꢀEDPꢀIA and 10,11EDPꢀCA
16
ACS Paragon Plus Environment

Page 17 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
as compared to control as measured by wound healing scratch assay on HOS, 143B and MG63
cell line
7
8
9
Cell cycle analysis of 10,11ꢀEDPꢀEA and its analogs
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 effects of 10,11ꢀEDPꢀEA and its analogs were tested on cell cycle progression by treating
cells with the compound for 24 hours, staining with propidium iodide and then analyzing with
flow cytometry. Staining by PI reflects the percentage of cells in each stage of the cell cycle G1,
S and G2. As seen in Table 1, in 143B cells there does not appear to be a significant effect on the
cell cycle except a slight suppression in the S phase with 10,11ꢀEDPꢀNA. However, with the
HOS cell lines there is an increase in G1 phase with all compounds but a suppression in the S
phase with all the compounds. Finally, with MG63 there is an increase in the G1 phase and
suppression in S phase with the 10,11ꢀEDPꢀIA.
Cell line untreated
10,11ꢀEDPꢀ
EA
10,11ꢀEDPꢀ
NA
10,11ꢀEDPꢀIA 10,11ꢀEDPꢀCA
143B
56.2 ± 2.7
26.6 ± 0.4
17.2 ± 2.3
50.7 ± 2.03
28.9 ± 3.9
21.1 ± 1.9
52.2 ± 0.1
24.0 ± 0.3*
23.7 ± 0.4
50.1 ± 0.03
41.8 ± 3.7
8.1 ± 3.7
54.2 ± 0.8
22.9 ± 0.3
22.9 ± 0.4
G1
S
G2
HOS
G1
47.8 ± 0.7
37.1 ± 1.6
15.1 ± 0.97
55.8 ± 5.0
24.3 ± 1.4*
20.0 ± 3.7
67.2 ± 0.6**# 62.2 ± 0.4**
54.4 ± 1.4*#
30.5 ± 1.8#
15.1 ± 0.4#
S
20.3 ± 0.9*
20.7 ± 2.1*
12.4 ± 0.3
17.1 ± 1.7
G2
MG63
62.9 ± 1.8
20.2 ± 1.8
67.7 ± 2.2
17.9 ± 2.1
63.5 ± 0.8
19.0 ± 0.4
59.6 ± 0.3
23.1 ± 0.4
G1
S
76.7 ± 0.1*
10.8 ± 0.3*
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 39
1
2
3
4
16.9 ± 0.01
G2
14.4 ± 0.2**
17.5 ± 0.4#
12.5 ± 0.4**# 16.9 ± 0.1##
5
6
7
8
9
Table 1: Table shows percentage of cells in each stage by cell cycle analysis using propidium
iodide, of 10,11ꢀEDPꢀEA, 10,11ꢀEDPꢀNA, 10,11ꢀEDPꢀIA and 10,11ꢀEDPꢀCA as monitored by
propidium iodide staining in all three cell lines. All experiments were done in triplicates. *
p<0.05, **p<0.01 as compared to untreated, # p<0.05, ##p<0.01 as compares to EA.
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
10,11ꢀEDPꢀEA and its analogs prevent angiogenesis in HUVEC cells
Increased angiogenesis is one of the hallmarks of cancer. The development of neoꢀ
40
vascularization is key for the adequate supply of nutrients and oxygen to the site of metastasis.
Endothelial tube formation on matrix has been established as a cellular model for angiogenesis.
41
Endothelial HUVEC cells form a networks of tubes when plated on matrigel and the tube
formation in presence and absence of treatment can be quantified as a measure of restricted
angiogenic potential. Briefly, endothelial HUVEC cells were grown on matrix and treated with
vehicle control, 0.5 ꢂM compound or 1 µM compound. The tube formation was assessed by
Image J analysis. As seen in figure 3a and b, 0.5 ꢂM and 1 ꢂM concentrations restrict tube
formation significantly as compared to the vehicle control. It was also verified that 1 ꢂM
concentration of treatment did not reduce cell viability in HUVEC cells as seen in supplementary
figure S5.
18
ACS Paragon Plus Environment

Page 19 of 39
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 3: (a) Restriction of tube formation in endothelial HUVEC cells on matrigel by 10,11ꢀ
EDPꢀEA, 10,11ꢀEDPꢀNA, 10,11ꢀEDPꢀIA and 10,11EDPꢀCA at 0.5 ꢂM and 1 ꢂM. (b)
Quantification of angiogenesis reduction by all amide derivatives as represented by number of
branches per unit area. (c) PrestoꢀTango assay measurement of the activation of CB1 and CB2 by
amide derivatives. (d) EC50 values of all amide derivatives with CB1 and CB2 by PrestoꢀTango
assay.
Expression of CB1 and CB2 receptors in model cell lines
11
Literature evidence suggests that DHEA is an agonist of CB1 and CB2 receptors. Therefore,
we investigated whether the effect of DHEA is through CB1/ CB2 pathway mediated.
Additionally, CB receptors have been shown to be overexpressed in various cancers including
42
osteosarcoma. We first determined the expression of CB1 and CB2 receptors in all three cell
lines HOS, 143B and MG63 by qRTꢀPCR, which quantifies receptor expression by measuring
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 39
1
2
3
4
5
6
7
8
9
corresponding mRNA levels and comparing to a control marker like beta actin. As seen in figure
S6, all three cell lines expressed both CB1 and CB2 in similar levels. Additionally, the
expression of CB1 and CB2 was also significantly higher in 143B cells as compared to HOS
cells (p= 0.023 and 0.021 for CB1 and CB2 respectively) as well as compared to MG63 cells (p=
0.033 and p=0.027 for CB1 and CB2 respectively).
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
Activation of CB1 and CB2 by 10,11ꢀEDPꢀEA and its amide analogs
To investigate whether 10,11ꢀEDPꢀEA and its amide modified analogs activated the cannabinoid
receptors, a betaꢀarrestin recruitment PRESTOꢀTango assay was performed in a doseꢀdependent
manner for these compounds 43. Prestoꢀtango is an openꢀsource resource developed by Roth
group that allows interrogation of the human GPCRome by drugs and endogenous ligands. In
this assay, upon ligand binding to a specific G proteinꢀcoupled receptor (GPCR), it triggers
desensitization which mediated by the recruitment of proteaseꢀtagged arrestin proteins to the
activated receptor. This proteaseꢀtagged arrestin then cleaves a protease site fused to the Cꢀ
terminus of the GPCR that is being studies thus releasing a nonꢀnative transcription factor (TF).
This TF regulates the transcription of a betaꢀlactamase reporter construct, which is measured
upon addition of the luminogenic substrate. The activation of the reporter gene by the
ligand/drug provides a quantifiable measurement of the degree of interaction between the target
receptor and the proteaseꢀtagged arrestin partner. Briefly, HTLA cells transfected with either
CB1 or CB2 were incubated with appropriate treatment and subsequently luminescence was
measured as an indicator of receptor activation (Figure 3c and d). CP55940 was used as a control
for this experiment and the EC50 values were observed to be 0.9 nM and 1.73 nM, which is in
44
agreement with literature. As observed in figure 3C, a dose response curve was obtained for
10,11ꢀEDPꢀEA and its analogs. 10,11ꢀEDPꢀEA has an EC50 value of 0.43 nM and 22.5 nM for
20
ACS Paragon Plus Environment

Page 21 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
CB1 and CB2 respectively. This indicates that it shows higher activation of CB1 than CB2. A
similar trend was observed for all 10,11ꢀEDPꢀEA analogs. 10,11ꢀEDPꢀNA had EC50 values of
0.19 nM and 7.3 nM respectively. 10,11ꢀEDPꢀCA had EC50 values of 0.03 nM and 0.46 nM for
CB1 and CB2, respectively. Finally, 10,11ꢀEDPꢀIA showed the smallest differences in activation
of CB1 and CB2 and showed EC50 values of 0.07 nM and 0.08 nM, respectively. Therefore, all
amide analogs have increased activation of both CB1 and CB2 as compared to the parent
compound 10,11ꢀEDPꢀEA and all activate CB1 better than CB2. As observed in this data, 10,11ꢀ
EDPꢀCA is the best activator of CB1. The cannabinoid receptor mediated reduction of cancer
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
19
cell migration has been reported in literature, and it can be hypothesized that 10,11ꢀEDPꢀCA
has increased antiꢀmigratory potential due to increased activation of CB1.
Apoptotic effect of DHEA epoxides in the presence of CB1 and CB2 antagonists
As an initial investigation into the mechanism of action for these endocannabinoids, we assessed
their apoptotic activity in the presence of CB1 and CB2 antagonists, rimonabant and AM 630
respectively. As seen from supplementary tables 2ꢀ4, rimonabant and AM 630 exhibit very
limited apoptotic activity as compared to control. However, the blocking of CB1 and CB2
receptors has very limited effect on the apoptotic potential of the endocannabinoid epoxides. In
HOS cell line (Table ST2), the decrease in apoptosis in presence of CB1 or CB2 antagonist or
both is modest with all three regioisomers, indicating this is not the primary pathway of action by
endocannabinoid epoxides. In 143B cells (Table ST3), when CB1 and CB2 antagonist were
incubated with 7,8ꢀEDPꢀEA, a slight increase in apoptotic activity with 14.6 ± 1.8% and 10.1 ±
1.4% respectively. A similar result was seen for 13,14ꢀEDPꢀEA where the treatment plus CB1
antagonist showed 13.0 ± 3.6%. The treatment plus CB2 antagonist showed no difference in
apoptosis as compared to treatment alone at 9.2 ± 0.8%. For 10,11ꢀEDPꢀEA there was slight
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 39
1
2
3
4
5
6
7
8
9
suppression of apoptotic activity with both antagonists. The slight increase in apoptosis might be
as a combinatorial effect of the treatment with the antagonist. Finally, as seen in table ST4, the
effect of CB1 and CB2 antagonists on MG63 cells was similar to that observed in 143B cells.
Alternatively, since the molecules have two different reactive centers, the epoxide as well as the
amide, they likely act through two different receptor classesꢀ the cannabinoid receptor and the
unknown receptor for epoxy fatty acids. It is possible that the blocking of the CB1 or CB2
receptors shuttle the molecule to the second epoxide receptor pathway thereby increasing the
apoptotic activity. Finally, 10,11ꢀEDPꢀEA showed a decrease in apoptotic potential with CB1
and CB2 antagonist at 7.4 ± 2.2% and 10.9 ± 0.1% indicating that this molecule mainly goes
through the CB1 and CB2 receptors.
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
DISCUSSION:
Lipid metabolites that are generated from omegaꢀ3 fatty acids such as DHA have been shown to
4
exert antiꢀinflammatory, proꢀresolving
and antiꢀtumorigenic effects through various
5 6 7
mechanisms.
In one mechanism, DHA is converted to DHA epoxides (EDPs) by CYP
epoxygenase that have shown to exhibit have antiꢀtumorigenic properties. Alternatively, DHA is
converted to DHEA, an endocannabinoid that have shown antiꢀproliferative efficiency in prostate
11
cancer cells.
Previously, we have shown that CYPs convert DHEA into DHEA epoxides
(EDPꢀEA) that are dual functional molecules containing both ethanolamide and epoxide moiety
25
(Figure 1a) as both the endocannabinoid and epoxyeicosanoid pathway have been separately
implicated in cancer 8, 9 and the terminal DHEA epoxide was shown to exhibit antiꢀinflammatory
25
properties.
Herein we evaluated the antiꢀtumorigenic properties of all the different
regioisomers of DHEA epoxides.
22
ACS Paragon Plus Environment

Page 23 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
To ascertain the biological relevance of endocannabinoid epoxides in the tumor
microenvironment, we induced tumors in mice and extracted total endocannabinoids from the
lung metastasized tumor that showed that the DHEA epoxides were increased 20ꢀ80% depending
on the regioisomers (Figure 1b) in the metastatic mice with osteosarcoma tumors as compared to
the normal lungs. The bioactive lipids such as DHEA epoxides are generated by various cell
types including immune cells. We hypothesize that the upregulation of cytochrome P450s in
these tumor tissues and surrounding immune cells leads to the rapid synthesis of these DHEA
epoxides that are antiꢀinflammatory,25 antiꢀangiogenic, antiꢀmetastatic and potentially proꢀ
resolving. It has been shown before that proꢀresolving mediators help in reduction of
inflammation and modify the tumor micro environment.45 We further show that the EDPꢀEAs
show a reduction of migratory potential of the cancer cells and moderately increased apoptosis in
osteosarcoma cell lines. We tested all the five epoxide regioisomers of EDPꢀEAs towards their
ability to induce apoptosis in three human osteosarcoma cell linesꢀ 143B, MG63 and HOS.
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 all the three cell lines, the various regioisomers show an apoptotic effect and an antiꢀmigratory
effect and this is significantly higher than the parent molecules DHA and DHEA. However, this
effect is differential among the various EDPꢀEA regioisomers. For instance, 10,11ꢀEDPꢀEA
emerges as the most potent regioisomer in reducing cell migration. Interestingly, this effect was
mediated via the CB1 receptor as using a CB1 antagonist, the antiꢀmigratory potential of 10,11ꢀ
EDPꢀEA was significantly reduced (Supplementary figure 4).
Since 10,11ꢀEDPꢀEA showed the greatest potency in our assays, we wanted to characterize the
molecule and explore its therapeutic potential in vivo. However, EDPꢀEAs has two hydrolytically
susceptible groups. The amide group is hydrolyzed by fatty acid amide hydrolase (FAAH) and
the epoxide group is hydrolyzed by soluble epoxide hydrolases. Thus, we prepared synthetic
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 39
1
2
3
4
5
6
7
8
analogs that demonstrated similar apoptotic and migratory potentials (Figure 2c and 2d) as
compared to parent 10,11ꢀEDPꢀEA, but markedly reduced susceptibility to FAAH hydrolysis
(Figure 2b).
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 cannabinoid receptor mediated apoptosis has been characterized in literature and is thought
46
to be mediated through ceramide synthesis. Additionally, the CB1 mediated restriction of cell
47
migration has also been studied.
In our studies, we show that while all endocannabinoid
epoxides and their derivatives activate both cannabinoid receptors, their activity is not
completely mediated through these receptors. Therefore, their activity could also be mediated
through other receptors including the uncharacterized receptor for epoxy fatty acids, PPAR and
TRPV. 48
CONCLUSIONS
Taken all the findings together, we demonstrate a novel class of eCB epoxides and their stable
30
derivatives exhibit antiꢀinflammatory,
antiꢀangiogenic, antiꢀtumorigenic and antiꢀmigratory
properties. These molecules particularly reduced the migratory ability of OS cells and therefore
may be of greater therapeutic significance to prevent metastasis of tumor. By virtue of their
ability to bind CB receptors, these molecules might also exhibit antiꢀnociceptive bone pain
activities. As osteosarcoma and other cancers are accompanied by inflammation and pain, these
molecules might aid in alleviating both conditions. These molecules are naturally occurring and
their levels increase ~80% in osteosarcoma metastatic lungs, increasing their levels in vivo using
FAAH and sEH dual inhibition will be effective adjuvant therapy for osteosarcoma. 49, 50
EXPERIMENTAL SECTION
24
ACS Paragon Plus Environment

Page 25 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
Materials: Annexin V/ PI kit was purchased from BD biosciences (556547). HUVEC cells were
procured from Thermo Fisher (C0035C) and endothelial cell media was obtained from Lonza
(EGM2 bullet kit CCꢀ3162). Angiogenesis kit was purchased from Abcam (ab204726).
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
General synthesis of EDPꢀEAs (1ꢀ5): EDPꢀEA synthesis was performed in a twoꢀstep
approach. First a nonꢀspecific epoxidation reaction with 50 mg of or DHA dissolved in 2 mL of
dichloromethane (DCM) with 2 molar equivalents of metachloroperoxybenzoic acid (mCPBA)
and was reacted for 1 h at room temperature. The reaction was stopped with an equal volume of
10% aqueous NaHCO3 to remove mCPBA from the organic layer, and the aqueous layer was reꢀ
extracted thrice with equal volumes of DCM and the combined organic layer was dried in vacuo.
Purification of regioisomers was achieved using normalꢀphase HPLC (NPꢀHPLC) using a
ZorbaxꢀNH2 5 ꢂm × 9.4 mm × 250 mm semipreparative column (Agilent, PN880952ꢀ208) with
an isocratic gradient (hexane/isopropanol/acetic acid; 90:10:0.1) coupled to a HPLC system. For
coeluting regioisomers of DHA epoxides (19,20ꢀEDP and 13,14ꢀEDP), the mixtures were further
purified on the same system using reversedꢀphase HPLC (RPꢀHPLC), a Sun Fire Prep C18 5 ꢂm
× 19 mm × 50 mm column (PN 186002566; Waters), and a mobile system composed of solvent
A (H2O/acetonitrile/acetic acid, 95:5:0.1), solvent B (H2O/acetonitrile/acetic acid, 5:95:0.1), and
a linear gradient from 50% A to 0% A in 50 min.
In the second step, the coupling of the purified EDP regioisomer to ethanolamine was achieved
by dissolving the epoxide in acetonitrile and incubating at 37 °C for 20 min with 1ꢀethylꢀ3ꢀ(3ꢀ
dimethylaminopropyl)carbodiimide (EDC) and Nꢀhydroxysuccinimide (NHS). Ethanolamine
was subsequently added and the mixture was stirred overnight at room temperature. The
resulting EDPꢀEA regioisomers were purified by RPꢀHPLC using a Sun Fire Prep C18 5 ꢂm ×
19 mm × 50 mm column (PN 186002566; Waters) and a mobile system composed of solvent A
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 39
1
2
3
4
5
6
(H20/acetonitrile/acetic acid 95:5:0.1) and solvent B (H20/acetonitrile/acetic acid 5:95:0.1) and a
linear gradient from 50% A to 0% A in 50 minutes.
7
8
9
Isolation of endocannabinoids from tissue samples: K7M2 Experimental Metastasis Tumor
Models of Osteosarcoma. All animal experimental procedures were reviewed and approved by
the University of Illinois Institutional Animal Care and Use Committee (Protocol 15120).
1,000,000 K7M2 cells were prepared in HBSS and intravenously injected into the tail vein of
6−8 week old female BALB/c mice (day 0) in a 200 ꢂL volume. After 14 days the mice were
sacrificed and lung tissue was isolated, weighted and total endocannabinoids were extracted. The
tissue was mechanically homogenized on ice with a BioHomogenizer (BioSpec) in a 1:1 ratio
with PBS supplemented with sEH inhibitor 12ꢀ(3ꢀadamantanꢀ1ꢀylꢀureido)ꢀdodecanoic acid
(AUDA) (Cayman Chemical) (30 ꢂm) and FAAH inhibitor PMSF (1 mM) till a homogenous
slurry was formed. 1 g of tissue was homogenized with 40 mL of ethyl acetate/hexane (9:1)
using a BioHomogenizer (BioSpec) at room temperature and then was sonicated for 1 min. The
resulting solution was washed with a 30% volume of water and centrifuged for the layers to
separate. The organic layer was removed, and the aqueous layer was extracted with an equal
volume of ethyl acetate/hexane (9:1) twice. The total organic layer was concentrated in vacuo
using a Buchi 120 rotary evaporator and was reconstituted in 1 mL of chloroform. The
metabolites of interest were isolated by solidꢀphase extraction (SPE) using a 1ꢀmL silica gel
column (no. 214477; Sigma). The reconstituted samples were added to the silica column and
washed in three column volumes of chloroform, and consequently eluted with four column
volumes of methanol/chloroform (1:9). The eluent was dried, and samples were reconstituted in
ethanol for analysis via LCꢀMS/MS.
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
26
ACS Paragon Plus Environment

Page 27 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Liquid chromatography mass spectrometry analysis. The LC separation was performed on an
Agilent Eclipse XDBꢀC18 (4.6 x 150mm, 5ꢂm) with mobile phase A (0.1% formic acid in water)
and mobile phase B (0.1% formic acid in acetonitrile). The flow rate was 0.4 mL/min. The linear
gradient was as follows: 0ꢀ2min, 90%A; 8min, 55%A; 13ꢀ25min, 40%A; 30min, 30%A; 35min,
25%A; 40ꢀ47min, 20%A; 47.5ꢀ54min, 90%A. The autosampler was set at 5°C. The injection
volume was 10 ꢂL. Positive mass spectra were acquired with the ion spray voltage of 5500 V
under electrospray ionization (ESI). The source temperature was 450°C. The curtain gas, ion
source gas 1, and ion source gas 2 were 32 psi, 60 psi, and 60 psi, respectively. Multiple reaction
monitoring (MRM) was used for quantitation: 19,20ꢀEDPꢀEA and 7,8ꢀEDPꢀEA m/z 388.1 →
m/z 62.1; 16,17ꢀEDPꢀEA, 13,14ꢀEDPꢀEA, and 10,11ꢀEDPꢀEA m/z 388.1 → m/z 370.1; DHEA
372.4 à m/z 62.1; Anandamide m/z 348.3 à m/z 62.1. Internal standards were monitored at: m/z
376.4 → m/z 66.1 for DHEAꢀd4.
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
Cell culture: Human cell lines HOS, MG63 and 143B were cultured in DMEM (Gibco,
Invitrogen, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS) and 1mg mlꢀ1
penicillin–streptomycin (Gibco, Invitrogen) at 37 °C and 5% CO2 in a humidified incubator.
Cell titer blue assay: 20,000 cells were plated per well in 96 well plates and allowed to adhere
overnight. In the morning, the cell media was aspirated and replaced with fresh 100 ꢂL of cell
media and incubated with compound or DMSO vehicle control (1 ꢂL) and allowed to incubate at
37 °C and 5% CO2 in a humidified incubator for 4 hours. 20 ꢂL of cell titer blue reagent was
added and the plate was incubated for 1 hour before the fluorescence was read using a plate
reader with excitation at 534 nm and emission at 590 nm.
Annexin V assay: Annexin V kit from BD biosciences) was used. Briefly, cells were scraped
and collected with the dead cells. They were washed with ice cold 2X PBS twice and
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 39
1
2
3
4
5
6
7
8
9
resuspended in 1X binding buffer and incubated with annexin VꢀFITC and propidium iodide for
15 mins, followed by flow cytometry by BD Accuri C6 instrument. Annexin V/PI assay was
performed as per manufacturer’s instructions. Briefly, cells were trypsinized and collected for
analysis. Dead cells in the media were collected for testing along with the rest of the sample. The
total cell sample was washed twice with cold PBS and then resuspended in 100 ꢂL of binding
buffer. 5 ꢂL of each annexin V and PI were added and the cells were vortexed and incubated in
the dark for 15 mins. The cells were then diluted with 400 ꢂL of binding buffer and analyzed by
BD Accuri flow cytometer within an hour. The treated samples were compared with negative
and positive stained samples both with vehicle control.
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
Wound healing assay: Cells were grown to a confluent monolayer in 12 well plates and a
scratch was made across the center of the plate with a pipette tip and images were taken. The
media was replaced with fresh media and treatment or vehicle control was added to the wells.
The cells were maintained at 37 °C and 5% CO2 in a humidified incubator and images were
taken every two hours. The images were processed using ImageJ via the wound healing plugin
and normalized to 100% area at time 0.
General synthesis of amine EDPꢀEAs and analogs (1ꢀ8): To a solution of epoxide in
acetonitrile, a solution of 1ꢀethylꢀ3ꢀ(3ꢀdimethylaminopropyl) carbodiimide (EDC) (10 mg/mL in
acetonitrile) and Nꢀhydroxysuccinimide (NHS) (10 mg/mL in acetonitrile) in the ratio of
(1:50:50) was added and incubated at 37 °C for 20 mins. To this we added the amine solution (2
mg mLꢀ1 in acetonitrile) in the ratio of 1:50 with epoxide and this was stirred at room
temperature overnight. The acetonitrile was removed under reduced pressure. The sample was
reconstituted in ethanol and purified via HPLC. Purification of 10,11ꢀEDP derivatives was
performed using a reversed phase HPLC (RPꢀHPLC), Sun Fire Prep C18 5 µm 19 x 50 mm
28
ACS Paragon Plus Environment

Page 29 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
(Waters, PN 186002566) and a mobile system composed of solvent A (H20/acetonitrile/acetic
acid 95:5:0.1) and solvent B (H20/acetonitrile/acetic acid 5:95:0.1) and a linear gradient from
50% A to 0% A in 50 minutes. Compounds 1ꢀ5 have been previously synthesized and
9
1
1
characterized.30 HRꢀESI and Hꢀ NMR characterization performed is detailed below and H
NMR spectra are shown in Supplementary figures 7ꢀ10. All compounds showed >95% purity by
1HꢀNMR analysis.
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
(4Z,7Z)ꢀNꢀ(2ꢀhydroxyethyl)ꢀ9ꢀ(3ꢀ((2Z,5Z,8Z)ꢀundecaꢀ2,5,8ꢀtrienꢀ1ꢀyl)oxiranꢀ2ꢀyl)nonaꢀ4,7ꢀ
dienamide Or 10,11ꢀ EDPꢀEA (2) : Synthesis was performed as mentioned in previous section
using 10,11ꢀEDP and ethanolamine. HRꢀESIꢀMS (m/z) 387.2821, [M+H]+
(C24H38NO3): 387.28
calculated
1H NMR: 5.58ꢀ5.27 (m, 10H), 3.77ꢀ3.62 (m, 3H), 3.45ꢀ3.32 (m, 2H), 3.29ꢀ 3.22 (t, 2H), 3.21ꢀ
3.15 (d, 1H), 3.03ꢀ2.94 (m, 2H), 2.88ꢀ2.75 (m, 4H), 2.75ꢀ2.65 (t, 1H), 2.32ꢀ2.22 (m, 5H), 1.15ꢀ
1.09 (t, 3H), 1.00ꢀ0.94 (t, 3H)
(4Z,7Z)ꢀNꢀcyclopropylꢀ9ꢀ(3ꢀ((2Z,5Z,8Z)ꢀundecaꢀ2,5,8ꢀtrienꢀ1ꢀyl)oxiranꢀ2ꢀyl)nonaꢀ4,7ꢀ
dienamide or 10,11ꢀEDPꢀCA (6): Synthesis was performed as mentioned in previous section
using 10,11ꢀEDP and cyclopropylamine. HRꢀESIꢀMS (m/z) found: 384.2910; [M+H]+
calculated C25H38NO3: 383.58
1H NMR: 5.57ꢀ5.22 (m, 10H), 3.30ꢀ3.24 (t, 2H), 3.24ꢀ 3.14 (q, 2H), 2.99ꢀ2.93 (m, 1H), 2.87ꢀ2.77
(m, 6H), 2.73ꢀ2.66 (m, 1H), 2.42ꢀ 2.36 (m, 3H), 2.20ꢀ2.14 (t, 2H), 2.12ꢀ2.05 (t, 2H), 1.87ꢀ79 (t,
2H), 1.00ꢀ0.92 (t, 3H), 0.88ꢀ0.84 (t, 3H)
(4Z,7Z)ꢀNꢀ(3ꢀhydroxypropyl)ꢀ9ꢀ(3ꢀ((2Z,5Z,8Z)ꢀundecaꢀ2,5,8ꢀtrienꢀ1ꢀyl)oxiranꢀ2ꢀyl)nonaꢀ
4,7ꢀdienamide or 10,11ꢀEDPꢀNA (7) : Synthesis was performed as mentioned in previous
29
ACS Paragon Plus Environment