Monoacylglycerol lipase regulates 2-arachidonoylglycerol action and arachidonic acid levels

Article abstract of DOI:10.1016/j.bmcl.2008.08.007

The structure-activity relationships of organophosphorus (OP) and organosulfur compounds were examined in vitro and in vivo as inhibitors of mouse brain monoacylglycerol lipase (MAGL) hydrolysis of 2-arachidonoylglycerol (2-AG) and agonist binding at the CB1 receptor. Several compounds showed exceptional potency toward MAGL activity with IC₅₀ values of 0.1-10 nM in vitro and high inhibition at 10 mg/kg intraperitoneally in mice. We find for the first time that MAGL activity is a major in vivo determinant of 2-AG and arachidonic acid levels not only in brain but also in spleen, lung, and liver. Apparent direct OP inhibition of CB1 agonist binding may be due instead to metabolic stabilization of 2-AG in brain membranes as the actual inhibitor.

Full text of DOI:10.1016/j.bmcl.2008.08.007

Bioorganic & Medicinal Chemistry Letters 18 (2008) 5875–5878
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Monoacylglycerol lipase regulates 2-arachidonoylglycerol action
and arachidonic acid levels
Daniel K. Nomura a, Carolyn S. S. Hudak a, Anna M. Ward a, James J. Burston b, Roger S. Issa a, Karl J. Fisher a,
Mary E. Abood c, Jenny L. Wiley b, Aron H. Lichtman b, John E. Casida a,
*
a Environmental Chemistry and Toxicology Laboratory, Department of Environmental Science, Policy and Management, 115 Wellman Hall, University of California,
Berkeley, CA 94720-3112, USA
b Department of Pharmacology and Toxicology, Virginia Commonwealth University, Richmond, VA 23298-0613, USA
c Department of Anatomy and Cell Biology, Temple University, Philadelphia, PA 19140, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The structure–activity relationships of organophosphorus (OP) and organosulfur compounds were exam-
ined in vitro and in vivo as inhibitors of mouse brain monoacylglycerol lipase (MAGL) hydrolysis of 2-
arachidonoylglycerol (2-AG) and agonist binding at the CB1 receptor. Several compounds showed excep-
tional potency toward MAGL activity with IC50 values of 0.1–10 nM in vitro and high inhibition at 10 mg/
kg intraperitoneally in mice. We find for the first time that MAGL activity is a major in vivo determinant
of 2-AG and arachidonic acid levels not only in brain but also in spleen, lung, and liver. Apparent direct OP
inhibition of CB1 agonist binding may be due instead to metabolic stabilization of 2-AG in brain mem-
branes as the actual inhibitor.
Received 28 May 2008
Revised 30 July 2008
Accepted 1 August 2008
Available online 6 August 2008
Keywords:
Monoacylglycerol lipase inhibitors
2-Arachidonoylglycerol
Arachidonic acid
Published by Elsevier Ltd.
The endocannabinoids (EC) 2-arachidonoylglycerol (2-AG) and
anandamide (AEA) regulate a diverse array of neurological and
metabolic functions and are altered by neuropathic pain, anxiety,
neurodegeneration, obesity, and cardiovascular disorders.1 2-AG
is a full agonist toward the cannabinoid receptor type 1 (CB1)
and its signaling is terminated primarily by monoacylglycerol li-
pase (MAGL). AEA levels are regulated by fatty acid amide hydro-
lase (FAAH).2–4 Augmentation of EC signaling by blockade of 2-
AG or AEA degradation (Scheme 1) is proposed as a therapeutic
strategy. However, characterization of MAGL or 2-AG in brain
and peripheral tissues is hindered by the paucity of systemic MAGL
inhibitors and lack of a MAGL knockout mouse. Discovery of potent
MAGL inhibitors is therefore essential in understanding the bio-
chemical, physiological, and therapeutic roles of this enzyme.
Structural manipulation of organophosphorus (OP) and organo-
sulfur (OS) compounds (Scheme 2) can potentially confer potency
Scheme 2. Organophosphorus (OP 1-8) and organosulfur (OS 9 and 10) compounds
used in this study. In earlier literature OP 1, OP 2, OP 7, and OP 8 are referred to as
IDFP, MAFP, chlorpyrifos oxon, and paraoxon, respectively.4–6
Scheme 1. Endocannabinoids 2-AG and AEA are agonists toward CB1 and are
metabolized by MAGL and FAAH, respectively, to arachidonic acid (AA).
and selectivity for MAGL and FAAH compared to other serine
hydrolases. OP 1 and OP 2 are previously reported as highly potent
MAGL and FAAH inhibitors.3,4 However, some OPs and OSs also dis-
place CB1 agonist binding through an unknown mechanism.5
* Corresponding author. Tel.: +1 510 642 5424; fax: +1 510 642 6497.
E-mail address: ectl@nature.berkeley.edu (J.E. Casida).
0960-894X/$ - see front matter Published by Elsevier Ltd.
doi:10.1016/j.bmcl.2008.08.007

5876
D. K. Nomura et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5875–5878
This study reports structure–activity relationships of OPs and
OSs with MAGL, FAAH, and CB1, and uses these tools to consider
three interrelationships of the EC system components. The first is
the in vitro potency for inhibiting MAGL, FAAH, and CB1 agonist
binding as a predictor of in vivo behavioral effects and pharmaco-
logical profile. The second is the variation among tissues in their
MAGL activity and differential regulation of 2-AG and AA levels. Fi-
nally, we consider the possibility that OP displacement of CB1 ago-
nist binding is due to 2-AG in membranes which is metabolically
stabilized by MAGL inhibition.
A library of 40 OPs and OSs, mostly prepared and optimized in
the Berkeley laboratory,7 was tested for potency and selectivity as
inhibitors of MAGL, FAAH, and CB1 agonist binding in mouse brain
membranes.8 Five particularly potent OPs for all three targets were
phosphonyl fluorides 1 and 2 and aryl phosphorus compounds 3–
6, all with long alkyl substituents [n-C12H25P, arachidonyl
(C20H33P) or n-C9H19SP] (Table 1 and Supplementary data). One
diethyl phosphate insecticide metabolite (OP 7) was quite potent
and another (OP 8) was only moderately active. Two sulfonyl fluo-
rides (9 and 10) with long alkyl chains were very potent on FAAH,
moderately active on MAGL, and differed greatly in activity on CB1.
Eight potent in vitro inhibitors were administered intraperito-
neally to mice at 10 mg/kg (OPs 1–6) or 100 mg/kg (OS 9 and OS
10) to determine if they were also effective in vivo in modulating
behavior and brain 2-AG and arachidonic acid (AA) levels (Fig. 1).
OP 1 and OP 4 were very effective in vivo in all respects, whereas
OP 2 and OP 3 with similar in vitro potency to 1 and 4 were not
effective in vivo. Thus, in vitro potency is not necessarily a predic-
tor of in vivo activity with metabolic stability a likely contributor.
OS 9 and OS 10 gave the same in vivo effects as OP 1 and OP 4
although at a 10-fold higher dose. Importantly, the OP- and OS-in-
duced increase in brain 2-AG levels was always directly related to
the lowering of brain free AA level.
Figure 1. Modulatory action of OP MAGL inhibitors at 10 mg/kg and OS compounds
9 and 10 at 100 mg/kg on brain 2-AG and AA levels relative to cannabinoid
behavior. Mice with cannabinoid behavior had >10 s latency in the bar test which
assesses catalepsy.10 They also qualitatively had a flattened posture and remained
motionless with their eyes open. Values are means SD, n = 3 mice/treatment
group.
2-AG and AA are important signaling molecules and intermedi-
ates not only in brain but also in other tissues.1,2,11 OP 1 at 10 mg/
kg strongly inhibits brain MAGL activity, elevates 2-AG, and lowers
AA,4 suggesting that it might also do so in other tissues (Fig. 2).
2-AG hydrolase activity was higher in brain than other tissues
examined with 78–83% sensitive to OP 1 in vivo in brain, kidney, tes-
tes, pancreas, and liver and 92–99% OP 1-sensitive in vivo in heart,
spleen, and lung. The apparent coupling of 2-AG and AA levels was
also examined. Among the tissues analyzed, brain, spinal cord, liver,
spleen, and lung, but not kidney, testes, pancreas, or heart showed
the possible co-dependence of 2-AG and AA pools (Fig. 2 and
Supplementary data). Most tissues also had increased levels of other
Table 1
Inhibitory potencies of OPs and OSs for mouse brain MAGL and FAAH activities and
CB1 agonist binding
Compound
IC50 SD (nM)
FAAH
MAGL
CB1
Phosphonyl fluorides
OP 1
OP 2
0.8 0.2a
2.2 0.3a
3
2a
2
1
Figure 2. 2-AG hydrolase activities and 2-AG and AA levels in mice treated with OP
1 (10 mg/kg, ip 4 h).12 Values are expressed as means SD, n = 3 mice/treatment
group. ND, not determined. Significance expressed as *p < 0.05, **p < 0.01,
***p < 0.001 in unpaired Student’s t-test.
0.10 0.02a
20a
Aryl phosphorus compounds
OP 3
OP 4
OP 5
OP 6
0.14 0.01
42 12
14
4
12
5
1
0.07 0.01
0.28 0.23
0.31 0.03a
12
3
3
2
1
0.15 0.01a
monoacylglycerol species, that is, 1- and 2-palmitoylglycerol and 1-
and 2-oleoylglycerol (Supplementary data). Beyond changes in glyc-
erol esters and AA levels, OP 1 treatment also led to decreases in
other unesterified fatty acid levels (palmitic, oleic, or stearic acid)
in spinal cord, liver, and spleen, indicating off-target effects of OP
1 in these tissues. The heart interestingly showed increases in both
oleic and stearic acids (Supplementary data).
Insecticide metabolites
OP 7
OP 8
10 4a
40 3a
540a
14
3
2300 1100a
1200a
Sulfonyl fluorides
OS 9
OS 10
200 75a
140 2a
2a
7
4
1.9 0.2a
1300 300
a
Data derived from previous studies.3–5,9

D. K. Nomura et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5875–5878
5877
Our results confirm the coordinate regulation of 2-AG and AA
levels by MAGL in brain,4 and show that this regulation also exists
in some peripheral tissues. These findings disfavor the current
model in which AA in many tissues is released primarily through
glycerophospholipid metabolism via multiple phospholipase A2
enzymes, notably cytosolic PLA2 (cPLA2), secretory PLA2 (sPLA2),
and calcium-independent PLA2 (iPLA2). While there are multiple
studies correlating increased PLA2 expression to pro-inflammatory
outcomes, cPLA2À/À mice (also deficient in sPLA2) have identical
levels of plasma and brain nonesterified fatty acid levels and brain
acyl-coenzyme A levels, albeit there were changes in esterified
phospholipid levels.13 Although OP 1 is not completely selective
for MAGL, it does not inhibit iPLA24 and the degree to which OPs
1, 4, 9, and 10 lower AA is equivalent to 2-AG elevation. MAGL
inhibitors may help treat inflammatory diseases not only in brain
but also in multiple peripheral tissues through the dual EC activa-
tion via 2-AG elevation and decreased eicosanoid signaling through
AA reduction.
It was very surprising to find that many OPs are potent inhibi-
tors of CB1 agonist binding in brain membranes. One possible
mechanism is direct OP binding or phosphorylation of CB1 at the
agonist or an allosteric site, and another is indirect by OP inhibition
of MAGL or FAAH to elevate the levels of 2-AG or AEA, or both
which then serve as the inhibitor. Three lines of evidence suggest
that the OPs do not react directly with CB1. Agonist binding is
OP-sensitive in brain membranes but not in recombinant ex-
pressed CB1 (eCB1)14,15 (Fig. 3a), indicating that some factor other
than or in addition to CB1 is required. Covalently derivatized CB1 is
not observed in brain membrane preparations labeled with a bio-
tinylated fluorophosphonate probe under conditions in which
phosphorylated MAGL and FAAH are readily evident.4 Inhibition
by OP derivatization is expected to be essentially irreversible and
noncompetitive with the agonist, whereas inhibition by OP 7 gives
an apparent competitive Scatchard plot (Fig. 3b and Supplemen-
tary data).16
An alternative hypothesis is that the OP inhibits MAGL and/or
FAAH and elevates the 2-AG and/or AEA level which in turn blocks
agonist binding (Schemes 1 and 3). CB1, assayed as agonist binding
with [3H]CP55940, is highly sensitive to many OPs (Table 1, Fig. 3a
and b, and Supplementary data) and OP 1 potentiates the CB1 ago-
Figure 3. Mechanisms of OP action on brain CB1. (a) OP 7 displaces [3H]CP55940
agonist binding in mouse brain membranes, but not in CB1 overexpressed in
nist action of 2-AG in vitro (measured by GTP
OP 1 stimulates GTP S binding at much higher concentration (EC50
0.5 M) (similar to the 0.3
M EC50 of 2-AG for CB1)17 than that re-
c
S binding) (Fig. 3c).17
HEK293 cells (eCB1). The eCB1/brain curve used a mixture of 100
lg eCB1 and
100 g brain membranes. IC50 values ( M) refer to brain (0.01) or components of
l
l
c
eCB1/brain (0.003 and 25). (b) Scatchard plot for apparent competitive OP
7
l
l
(100 nM) displacement of [3H]CP55940 agonist binding. (c) Stimulation of GTP
binding by 2-AG, 2-AG plus OP 1 (150 nM), or OP 1 alone comparing CB1+/+ and
CB1À/À mouse brain membranes. (d) Similar OP sensitivity and specificity profiles
quired to displace agonist binding (IC50 2 nM) (Table 1) in similar
preparations of brain membranes. The choice between MAGL/2-
AG and FAAH/AEA as the target can be approached by OP sensitiv-
ity and specificity considerations, and by analysis for OP-induced
elevations of EC levels. The OP sensitivity and specificity profiles
correlate better for MAGL versus CB1 (r2 = 0.81, n = 27) (Fig. 3d)
than for FAAH versus CB1 (r2 = 0.68, n = 16) (Supplementary data).
Although AEA has higher CB1 affinity than 2-AG18, the ꢀ1000-fold
greater level of 2-AG may override the affinity difference. Impor-
tantly, there is sufficient accumulation of 2-AG on OP treatment
to strongly inhibit the CB1 site (Fig. 3e).19 These results support li-
pid rafts20 as an important compartment for 2-AG in its interac-
tions with CB1 and MAGL. The weight of evidence favors OP
action on CB1 initiated by MAGL inhibition rather than FAAH inhi-
bition or direct on the receptor.
In conclusion, we report the discovery of several OP MAGL
inhibitors with unprecedented in vitro potency (IC50 < 1 nM), a
subset of which is effective in vivo in dramatically raising brain
2-AG levels, leading to cannabinoid behavior. These inhibitors are
attractive probes to uncover specific functions of MAGL and 2-AG
in EC signaling in vivo both centrally and peripherally, and to
investigate MAGL as a therapeutic target. The findings establish
that MAGL and 2-AG, and not phospholipases and phospholipids,
for MAGL and CB1. (e) OP
1 (10 mg/kg ip, 4 h) significantly elevated brain
membrane 2-AG levels. Values are means SD, n = 3–6. Significance expressed as
**p = 0.01.
Scheme 3. Several lines of evidence are presented for indirect OP inhibition of CB1
agonist binding in brain membranes by inhibiting MAGL to elevate 2-AG that binds
CB1 rather than direct binding or phosphorylation of CB1.
regulate brain levels of free AA in multiple tissues. Finally, we pro-
pose a mechanism for OPs and other MAGL inhibitors to indirectly
displace exogenous CB1 agonist binding in which elevated 2-AG
levels, metabolically stabilized in brain membranes by MAGL inhi-
bition, serve as the actual inhibitor.

5878
D. K. Nomura et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5875–5878
hexanes as effluent to give OP 4 (0.38 mmol, 83%). Analogous procedures were
Acknowledgments
used for 12 other O-aryl alkylphosphonates (Supplementary Table 1).
8. MAGL activity was determined with either unlabeled 2-AG or
[
14C]
1-oleoylglycerol, FAAH activity with [3H]anandamide and CB1 agonist
binding with [3H]CP55940 as described previously.3,4 The same IC50 value
This work was supported by Grant ES008762 (J.E.C.) and
DA003672 (A.H.L., grant to Billy R. Martin) from the National
Institutes of Health and the University of California Toxic
Substances Research & Teaching Program (D.K.N.). We thank Laura
J. Sim-Selley and Dana E. Selley of Virginia Commonwealth
University (Richmond, VA) for assistance with experimental design
and data analysis of the GTP binding experiments. We acknowl-
edge our University of California, Berkeley colleagues Rita
Nichiporuk and Ulla Andersen for advice in the mass spectrometry
studies. This work is dedicated to Benjamin Cravatt for exciting
collaborative research.
(0.07 l
M) was found for OP 4 in assays with 2-AG and [14C]1-oleoylglyerol
(Supplementary data).
9. Quistad, G. B.; Liang, S. N.; Fisher, K. J.; Nomura, D. K.; Casida, J. E. Toxicol. Sci.
2006, 91, 166.
10. Catalepsy bar test was performed as previously described.4
11. (a) Kunos, G.; Osei-Hyiaman, D. Am. J. Physiol. Gastrointest. Liver Physiol. 2008,
294, G1101; (b) Serhan, C. N.; Chiang, N.; Van Dyke, T. E. Nat. Rev. Immun. 2008,
8, 349; (c) Cipollone, F.; Cicolini, G.; Bucci, M. Pharmacol. Ther. 2008, 118, 161;
(d) Wang, D.; DuBois, R. N. Cancer Lett. 2008, 3, 4. doi:10.1016/
j.canlet.2008.03.004.
12. 2-AG hydrolase activity and 2-AG and AA levels were determined as described
previously.4
13. (a) Rosenberger, T. A.; Villacreses, N. E.; Contreras, M. A.; Bonventre, J. V.;
Rapoport, S. I. J. Lipid Res. 2003, 44, 109; (b) Farooqui, A. A.; Horrocks, L. A.;
Farooqui, T. J. Neurochem. 2007, 101.
Supplementary data
14. Abood, M. E.; Ditto, K. E.; Noel, M. A.; Showalter, V. M.; Tao, Q. Biochem.
Pharmacol. 1997, 53, 207.
15. eCB1 overexpressed in HEK293 cells is not sensitive to OP 1 or OP 7 (up to
100,000 nM) although it displays appropriate [3H]CP55940 binding and
sensitivity to WIN55212-2. Upon addition of brain membranes to eCB1, OP
7-sensitivity was partially restored.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2008.08.007.
References and notes
16. Kinetic experiments were performed as binding isotherms for OP
7
displacement of [3H]CP55940 agonist binding (see Supplementary data). Kd
(nM) 0.86 for control and 2.0 for OP 7. Bmax (pmol/mg) 0.27 for control and 0.26
for OP 7.
1. (a) Piomelli, D. Nat. Rev. Neurosci. 2003, 4, 873; (b) Di Marzo, V.; Petrosino, S.
Curr. Opin. Lipidol. 2007, 18, 129.
2. (a) Ahn, K.; McKinney, M. K.; Cravatt, B. F. Chem. Rev. 2008, 108, 1687; (b)
Blankman, J. L.; Simon, G. M.; Cravatt, B. F. Chem. Biol. 2007, 14, 1347; (c)
Cravatt, B. F.; Demarest, K.; Patricelli, M. P.; Bracey, M. H.; Giang, D. K.; Martin,
B. R.; Lichtman, A. H. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 9371.
3. Quistad, G. B.; Klintenberg, R.; Caboni, P.; Liang, S. N.; Casida, J. E. Toxicol. Appl.
Pharmacol. 2006, 211, 78.
4. Nomura, D. K.; Blankman, J. L.; Simon, G. M.; Fujioka, K.; Issa, R. S.; Ward, A. M.;
Cravatt, B. F.; Casida, J. E. Nat. Chem. Biol. 2008, 4, 373.
5. Segall, Y.; Quistad, G. B.; Sparks, S. E.; Nomura, D. K.; Casida, J. E. Toxicol. Sci.
2003, 76, 131.
6. Martin, B. R.; Beletskaya, I.; Patrick, G.; Jefferson, R.; Winckler, R.; Deutsch, D.
G.; Di Marzo, V.; Dasse, O.; Mahadevan, A.; Razdan, R. K. J. Pharmacol. Exp. Ther.
2000, 294, 1209.
7. In a representative reaction, ethyl n-dodecylchlorophosphonate (Björkling, F.;
Dahl, A.; Patkar, S.; Zundel, M. Bioorg. Med. Chem. 1994, 2, 697–705) (0.63
mmol) in methylene chloride (4 ml) was added to 4-cyanophenol (0.46 mmol)
17. Guanosine-5-O-(
previously described.4 Stimulation of GTP
by preincubation with OP 1 (150 nM) shifting the EC50 of 2-AG from 1.0 to
c
-thio)-triphosphate (GTP
c
S) binding was determined as
c
S binding by 2-AG is potentiated
0.3 lM. Interestingly, there is significant 2-AG-mediated stimulation of GTP
binding in CB1À/À mouse brain, also potentiated by OP 1, indicating the
possible existence of another cannabinoid receptor. OP 1 alone at higher
concentrations stimulates GTP
but not in CB1À/À membranes, suggesting a possible direct stimulatory action
of OP on CB1, but not on the other 2-AG-responsive receptor. This
c lM),
S binding in CB1+/+ membranes (EC50 0.5
1
concentration is 600-fold higher than the IC50 of MAGL and ꢀ250-fold higher
than the IC50 of CB1 agonist binding.
18. Sugiura, T.; Kondo, S.; Sukagawa, A.; Nakane, S.; Shinoda, A.; Itoh, K.;
Yamashita, A.; Waku, K. Biochem. Biophys. Res. Commun. 1995, 215, 89.
19. Upon consideration of the 2-AG concentration in a typical CB1 binding assay,
controls would have 16 nM compared to 360 nM in brain membranes from OP
1-treated mice, consistent with the 2-AG levels required to induce stimulation
of GTP binding.
and triethylamine (100
ll) in methylene chloride (2 ml). The mixture was
stirred for 20 h at room temperature, diluted with ethyl acetate (20 ml), filtered
through silica gel, and purified on a silica gel column using 30% ethyl acetate/
20. Dainese, E.; Oddi, S.; Bari, M.; Maccarrone, M. Curr. Med. Chem. 2007, 14,
2702.