11,12-Epoxyeicosatrienoic acid (11,12-EET): Structural determinants for inhibition of TNF-α-induced VCAM-1 expression

Article abstract of DOI:10.1016/j.bmcl.2003.08.060

A series of 11,12-EET analogues were synthesized and compared using a human endothelial cell based TNF-α-induced VCAM-1 expression assay. The resulting data were used to map a putative recognition/binding domain for 11,12-EET.

Full text of DOI:10.1016/j.bmcl.2003.08.060

Bioorganic & Medicinal Chemistry Letters 13 (2003) 4011–4014
11,12-Epoxyeicosatrienoic Acid (11,12-EET): Structural
Determinants for Inhibition of TNF-ꢀ-Induced VCAM-1
Expression§
J. R. Falck,a,* L. Manmohan Reddy,a Y. Krishna Reddy,a Muralidhar Bondlela,a
U. Murali Krishna,a Yu Ji,a Jianxin Sunb and James K. Liaob,*
aDepartment of Biochemistry, University of Texas Southwestern Medical Center, Dallas, TX 75390-9038, USA
bVascular Medicine and Atherosclerosis Unit, Cardiovascular Division, Brigham & Women’s Hospital and Harvard Medical School,
Cambridge, MA 02139, USA
Received 6May 2003; revised 27 August 2003; accepted 27 August 2003
Abstract—A series of 11,12-EET analogues were synthesized and compared using a human endothelial cell based TNF-a-induced
VCAM-1 expression assay. The resulting data were used to map a putative recognition/binding domain for 11,12-EET.
# 2003 Elsevier Ltd. All rights reserved.
Epoxyeicosatrienoic acids (EETs) are labile autacoids
produced by the cytochrome P450 epoxygenase branch
of the arachidonate cascade.2 Their potent biological
properties, ranging from regulation of ion channels3 and
stimulation of G-protein ADP-ribosylation4 to mod-
ulation of the vasculature5 and gap junctional
communication,6 have attracted wide and continuing
interest. Recently, Liao et al.7 demonstrated that
amongst the four regioisomeric EETs, 11,12-EET is the
most potent inhibitor of tumor necrosis factor-a (TNF-
a) induced vascular cell adhesion molecule-1 (VCAM-1)
expression in cytokine activated human endothelial
cells. In addition, 11,12-EET inhibits VCAM-1 expres-
sion in response to other inflammatory mediators such
as interlukin-1-a (IL-1 a) and bacterial lipopoly-
saccharide (LPS).7 To help elucidate the spatial and
functional determinants of 11,12-EET that contribute to
inhibition as well as expedite the development of meta-
bolically more stable anti-inflammatory agents, a series
of 11,12-EET structural analogues were compared using
a human endothelial cell based TNF-a-induced VCAM-
1 expression assay. These data also provide the first
insights into a putative macromolecular recognition
and/or binding site8 for 11,12-EET.9,10 A representative
sampling of these analogues is presented in Table 1
along with their bioassay results expressed as the per-
cent inhibition of TNF-a-induced VCAM-1 expression
elicited by 100 nM of eicosanoid. Under the assay con-
ditions, rac-11,12-EET (1) and the individual antipodes
2 and 3 suppressed expression by ca. one-third. To
ascertain the contributions of the olefins, a series of
saturated or partially saturated analogues 4–8 were
prepared and found to be comparable or less active than
1. Tetrahydro-analogue 9, in contrast, was nearly twice
as potent as 1. As a consequence of its simpler structure,
9 is easier to prepare and is chemically more stable than
1 since it does not contain the 1,4-diene subunits that
are susceptible to autooxidation. Notably, conversion of
the critical cis-Á8,9-olefin in 9 to a linear acetylene, that
is 10, was detrimental. Alterations to the cis-epoxide
were more or less well tolerated as illustrated by trans-
epoxide 11, furan 12, cyclopropane 13, episulfides 14–
16, and ether 17. The latter is expected to be a useful
alternative to 9 in systems where degradation by epox-
ide hydrolases must be minimized.11 The substantial dip
in activity shown by alcohol 18 suggests the presence of
an oxygen is not sufficient and that this portion of the
molecule should also remain hydrophobic. Alterations
in the chain length at either end (analogues 19–22),
introduction of a heteroatom to obviate b-oxidation
(analogue 23), or the seemingly innocuous shift of the
epoxide by just one carbon toward the o-end (analogue
24) proved undesirable.
§See ref 1.
*Corresponding author. Tel.: +1-214-648-2406; fax: +1-214-648-
6455; e-mail: j.falck@utsouthwestern.edu
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.08.060

4012
J. R. Falck et al. / Bioorg. Med. Chem. Lett. 13(2003) 4011–4014
Table 1. Percent inhibition TNF-a-induced VCAM-1 expression
% (SEM/na)
Eicosanoid (100 nM)
% (SEM/na)
Eicosanoid (100 nM)
19: R=H, 9(S),10(R)
20: R=H, 9(R),10(S)
13 (6/6)
17 (5/6)
1: rac-11,12-EET
2: 11(R),12(S)-EET
3: 11(S),12(R)-EET
39 (2.5/52)
35 (7/6)
31 (7/6)
21: X=CH2, R=H
22: X=CH2, R=(CH2)3CH3
23: X=O, R=CH2CH3
1 (3/6)
À1 (3/6)
3 (2/6)
4: R=R1=R2=H
36 (16/6)
10 (12/6)
30 (16/6)
13 (16/6)
31 (26/6)
69 (4/6)
5: R=H, R1=R2=cis-olefin
6: R=R2=cis-olefin, R1=H
7: R=R1=H, R2=cis-olefin
8: R=cis-olefin, R1=R2=H
9: R=R2=H, R1-cis-olefin
10: R=R2=H, R1-acetylene
13 (4/6)
28 (6/5)
24
25
58 (4/23)
50 (5/13)
31 (8/13)
36(6/13)
11
12
13
26: R=OH
27: R=MeCONH
28: R=4-IC6H4O–
27 (6/6)
63 (5/6)
52 (5/6)
29 (5/5)
29
14: R=R1=cis-olefin
15: R=R1=H
65 (5/13)
49 (7/5)
30: R=Me
31: R=Ph
31 (3/13)
33 (5/13)
26(7/13)
1 (6/6)
16
32
48 (5/13)
32 (5/5)
17
18
33
15 (6/7)
aNumber of incubations

J. R. Falck et al. / Bioorg. Med. Chem. Lett. 13(2003) 4011–4014
4013
Hydroxylation of C(20) in 9 and 17 gave 25 and 26,
respectively, but abolished the gains in efficacy achieved
in the parents. However, this could be largely restored
using less polar derivatives as seen in acetamide 27 and
iodophenoxy 28, despite the latter’s increased steric sig-
nature. Replacement of the carboxylic acid in 9, 11, and
17 with N-acylsulfonamides provided the ionizable iso-
steres12 29–31, but failed to improve the level of inhibi-
tion and, in the latter case, was actually
counterproductive. Likewise, reduction of 9 to alcohol
33 resulted in diminished activity compared to the free
acid form.
modest agonist activity. However, this terminus is steri-
cally constrained. (ii) The Á5,6-olefin spans a mostly lipo-
philic region. While this section makes only a minor
contribution to recognition, the replusion of a heteroatom
in the chain can significantly disrupt binding. (iii) The
Á
8,9-olefin seems to occupy a shallow pocket that con-
forms to the bent configuration of this cis-olefin, but
poorly to a linear acetylene. The presence of an electron
lone pair or p-bond in this vicinity is vital for inhibition.
(iv) An oxygen or sulfur positioned between C(11) and
C(12), irrespective of its nature (cis- or trans-epoxide/
thiirane, ether, or furan, but not an alcohol), is also
essential for maximum activity. The strong positional
dependency suggests a single site of coordination. And
(v), there is a lipophilic pocket that can accommodate
the terminal methyl or other lipophilic group at the o-
terminus.
Because of their anticipated chemical and/or metabolic
stability and comparative potencies, tetrahydro analo-
gue 9 and ether 17 were selected for further analysis.
Their IC50 values (Fig. 1) were determined to be 60 nM
and 45 nM, respectively.
Analogue Syntheses
Recognition/Binding Domain Map for 11,12-EET
The chemical syntheses of analogues 913 and 17 are
outlined in Scheme 1 and are representative of the
approach used for the other analogues. Protected
acetylene 34, prepared from 8-nonynol14 (TBDPS-Cl/
Im, CH2Cl2, rt, 4 h, 90%), was metalated using n-BuLi
and then alkylated with either 35a or b15,16 to give 36a
and b, respectively. Desilylation and partial hydrogena-
tion afforded 37a which was oxidized to 9 using PDC
whereas 37b was transformed to 17 by Jones reagent.
The foregoing structure–activity relationships were used
to map the recognition/binding domain responsible for
the inhibition induced by 11,12-EET and its analogues.
The representation presented in Figure 2 identifies five
general regions and incorporates different binding
motifs: (i) the ionic attraction of the carboxylate is
important, although hydrogen bonding may also make
a contribution since the corresponding alcohol retains
Bioassay
Human saphenous vein endothelial cells (HSVEC)
grown to confluence were replated on low pyrogen
fibronectin (1.5 mg/cm2) at 2Â104 cells/cm2. After wash-
ing with PBS, the HSVEC monolayer was incubated
with 11,12-EET (1) or an analogue (100 nM) for 1 h,
and then stimulated with TNF-a for an additional 16h.
The cells were fixed in 96-well microtiter plates using
1% paraformaldehyde for 45 min, washed thrice with
PBS, and then incubated with VCAM-1 antibody for 2
h (1:100 dilution in PBS). Following a PBS rinse, the
cells were sequentially incubated with biotinylated horse
anti-mouse IgG antibody (1:1000 dilution, Vector Labs,
Inc., Burlingame, CA, USA) for 1 h, streptavidin-alkaline
phosphatase (Zymed, South San Francisco, CA, USA)
for 30 min, and p-nitrophenyl phosphate disodium (1.5 mg
/mL) for 30 min at 22 ꢀC. The final absorbance was mea-
sured at 410 nm to determine the cell surface expression of
Figure 1. Inhibitory effects of analogues 9 (closed circles) and 17 (open
circles) on TNF-a-induced VCAM-1 expression in human endothelial
cells.7 Data represent the meanÆstandard deviation (n=6).
Scheme 1. Synthesis of analogues 9 and 17: (a) n-BuLi, THF/HMPA
(4:1), À40 ꢀC, 2 h; bromide, À40 ꢀC to rt, 2 h; (b) n-Bu4NF, THF,
0 ꢀC, 4 h; (c) P-2 Ni/H2, EtOH, rt, 2 h; (d) PDC, DMF, rt, 16h;
(e) Jones reagent, acetone, À20 ꢀC, 4 h.
Figure 2. Recognition/binding domain map.

4014
J. R. Falck et al. / Bioorg. Med. Chem. Lett. 13(2003) 4011–4014
VCAM-1. Nonbinding control antibodies (OX 6,
against MHC class II antigen) were used in each
experiment.
relevant discussion of the distinctions, see: Laduron, P. M.
Biochem. Pharmacol. 1984, 33, 833.
9. A stereospecific, membrane binding site for 14,15-EET has
been reported: Wong, P. Y. K.; Lin, K. T.; Yan, Y. T.; Ahern,
D.; Iles, J.; Shen, S. Y.; Bhatt, R. K.; Falck, J. R. J. Lipid
Mediators 1993, 6, 199.
10. For an EET/fatty acid binding protein, see: Widstrom, R. L.;
Norris, A. W.; Spector, A. A. Biochemistry 2001, 40, 1070.
11. (a) Fang, X.; Kaduce, T. L.; Weintraub, N. L.; Harmon,
S.; Teesch, L. M.; Morisseau, C.; Thompson, D. A.; Ham-
mock, B. D.; Spector, A. A. J. Biol. Chem. 2001, 276, 14867.
(b) Zeldin, D. C.; Wei, S.; Falck, J. R.; Hammock, B. D.;
Snapper, J. R.; Capdevila, J. H. Arch. Biochem. Biophys. 1995,
316, 443.
Acknowledgements
The anti-VCAM-1 antibody (mouse IgG, E1/6) was
kindly provided by Michael A. Gimbrone, Jr., Brigham
& Woman’s Hospital, Boston, MA, USA. Financial
support provided in part by the Robert A. Welch
Foundation and NIH (GM31278, DK38226, and
HL48743).
12. Drummond, J. T.; Johnson, G. Tetrahedron Lett. 1988,
29, 1653.
References and Notes
13. 1H NMR (400 MHz) of analogue 9: d 0.88 (t, J=6.8 Hz,
3H), 1.20–1.63 (m, 22H), 1.99 (q, J=6.8 Hz, 2H), 2.14–2.22
(m, 1H), 2.27–2.38 (m, 3H), 2.64–2.70 (m, 2H), 5.31–5.40 (m,
1H), 5.42–5.51 (m, 1H). 11: d 0.88 (t, J=6.7 Hz, 3H), 1.20–
1.44 (m, 17H), 1.45–1.58 (m, 3H), 1.59–1.70 (m, 2H), 2.05 (q,
J=6.7 Hz, 2H), 2.13–2.24 (m, 1H), 2.29–2.42 (m, 3H), 2.89–
2.97 (m, 2H), 5.38–5.47 (m, 1H), 5.48–5.56(m, 1H). 14: d 0.88
(t, J=7.0 Hz, 3H), 1.20–1.41 (m, 6H), 1.72 (quintet, J=7.6
Hz, 2H), 2.05 (q, J=6.4 Hz, 2H), 2.13 (q, J=6.7 Hz, 2H), 2.20–
2.45 (m, 4H), 2.48–2.61 (m, 2H), 2.80 (apparent t, J=6.4 Hz, 2H)
2.92–3.00 (m, 2H), 5.33–5.58 (m, 6H). 15: d 0.88 (t, J=6.4 Hz,
3H), 1.20–1.42 (m, 15H), 1.41–1.58 (m, 3H), 1.59–1.70 (m, 2H),
1.82–1.92 (m, 2H), 2.0–2.10 (m, 2H), 2.32–2.40 (m, 3H), 2.41–
2.52 (m, 1H), 2.90–3.00 (m, 2H), 5.43–5.52 (m, 2H). 17: d 0.88 (t,
J=6.4 Hz, 3H), 1.22–1.40 (m, 18H), 1.53–1.65 (m, 4H), 1.98–2.08
(m, 2H), 2.33 (quintet, J=7 Hz, 4H), 3.38–3.43 (m, 4H), 5.33–
5.40 (m, 1H), 5.41–5.47 (m, 1H).
1. Presented in part at the 225th ACS National Meeting, New
Orleans, LA, March 23–27, 2003: MEDI 258.
2. Capdevila, J. H.; Falck, J. R. Prostaglandins and Other
Lipid Mediators 2002, 68–69, 325.
3. (a) Zou, A.-P.; Fleming, J. T.; Falck, J. R.; Jacobs, E. R.;
Gebremedhin, D.; Harder, D. R.; Roman, R. J. Am. J. Phy-
siol. 1996, 270, F822. (b) Salvail, D.; Dumoulin, M.; Rous-
seau, E. Am. J. Physiol. 1998, 275, L432. (c) Archer, S. L.;
Gragasin, F. S.; Wu, X.; Wang, S.; McMurtry, S.; Kim, D. H.;
Platonov, M.; Koshal, A.; Hashimoto, K.; Campbell, W. B.;
Falck, J. R.; Michelakis, E. D. Circulation 2003, 107, 769.
4. Li, P.-L.; Chen, C.-L.; Bortell, R.; Campbell, W. B. Cir.
Res. 1999, 85, 349.
5. Roman, R. J.; Maier, K. G.; Sun, C. W.; Harder, D. R.;
Alonso-Galicia, M. Clin. Exp. Pharm. Physiol. 2000, 27, 855.
6. Popp, R.; Brandes, R. P.; Ott, G.; Busse, R.; Fleming, I.
Cir. Res. 2002, 90, 800.
14. Hopf, H.; Kruger, A. Chem. Eur J. 2001, 7, 4378.
15. Prepared by bromination (CBr4/TPP, CH2Cl2, 0 ꢀC, 1 h)
of the corresponding alcohol, Kotsuki, H.; Ushio, Y.; Yoshi-
mura, N.; Ochi, M. J. Org. Chem. 1987, 52, 2594.
16. As in ref 15: using the alcohol described by, Takeuchi, R.;
Kashio, M. J. Am. Chem. Soc. 1998, 120, 8647.
7. Node, K.; Huo, Y.; Ruan, Z. X.; Yang, B.; Spiecker, M.;
Ley, K.; Zeldin, D. C.; Liao, J. K. Science 1999, 285, 1276.
8. Recognition and/or binding site is the preferred terminol-
ogy at this time since insufficient data are available to satisfy
acceptable criteria for the existence of a receptor. For a