Carbamate analogues of fumagillin as potent, targeted inhibitors of methionine aminopeptidase-2

Article abstract of DOI:10.1021/jm901260k

Inhibition of methionine aminopeptidase-2 (MetAP2) represents a novel approach to antiangiogenic therapy. We describe the synthesis and activity of fumagillin analogues that address the pharmacokinetic and safety liabilities of earlier candidates in this

Full text of DOI:10.1021/jm901260k

J. Med. Chem. 2009, 52, 8047–8056 8047
DOI: 10.1021/jm901260k
Carbamate Analogues of Fumagillin as Potent, Targeted Inhibitors of Methionine Aminopeptidase-2
Christopher C. Arico-Muendel,*,† Dennis R. Benjamin,‡ Teresa M. Caiazzo,§ Paolo A. Centrella,† Brooke D. Contonio,
Charles M. Cook, ,^ Elisabeth G. Doyle,# Gerhard Hannig,3 Matthew T. Labenski,O Lily L. Searle,* Kenneth Lind,†
Barry A. Morgan,† Gary Olson, Christopher L. Paradise,[ Christopher Self,*, Steven R. Skinner,† Barbara Sluboski,
Jennifer L. Svendsen, Charles D. Thompson,z William Westlin,O and Kerry F. Whiteþ
Praecis Pharmaceuticals, Inc., 830 Winter Street, Waltham, Massachusetts 02451-1420, Provid Research, 10 Knightsbridge Road, Piscataway,
New Jersey 08854. †Present address: GlaxoSmithKline, 830 Winter Street, Waltham, MA 02451-1420. ‡Present address: Seattle Genetics, Inc.,
21823 30th Drive SE, Bothell, WA 98021. §Present address: Pfizer Research, One Burtt Road, Andover, MA 01810. Present address: Provid
Pharmaceuticals, Inc., 671 U.S. Route 1, North Brunswick, NJ 089002. ^Deceased August, 2008. #Present address: Vertex Pharmaceuticals,
130 Waverly Street, Cambridge, MA 02139. 3Present address: Ironwood Pharmaceuticals, 320 Bent Street, Cambridge, MA 02141. OPresent
address: Avila Therapeutics, 100 Beaver Street, Waltham, MA 02453. [Present address: The University of Texas;Southwestern Medical Center,
5323 Harry Hines Boulevard, Dallas, TX 75390. zPresent address: Department of Drug Metabolism, Merck & Company Incorporated,
West Point, PA. þPresent address: Infinity Pharmaceuticals, 780 Memorial Drive, Cambridge, MA 02139.
Received August 21, 2009
Inhibition of methionine aminopeptidase-2 (MetAP2) represents a novel approach to antiangiogenic
therapy. We describe the synthesis and activity of fumagillin analogues that address the pharmacokinetic
and safety liabilities of earlier candidates in this compound class. Two-step elaboration of fumagillol with
amines yielded a diverse series of carbamates at C6 of the cyclohexane spiroepoxide. The most potent of
these compounds exhibited subnanomolar inhibition of cell proliferation in HUVEC and BAEC assays.
Although a range of functionalities were tolerated at this position, R-trisubstituted amines possessed
markedly decreased inhibitory activity, and this could be rationalized by modeling based on the known
fumagillin-MetAP2 crystal structure. The lead compound resulting from these studies, (3R,4S,5S,6R)-5-
methoxy-4-((2R,3R)-2-methyl-3-(3-methylbut-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-yl (R)-1-amino-
3-methyl-1-oxobutan-2-ylcarbamate, (PPI-2458), demonstrated an improved pharmacokinetic profile
relative to the earlier clinical candidate TNP-470, and has advanced into phase I clinical studies in
non-Hodgkin’s lymphoma and solid cancers.
Introduction
circumvent the nonspecific toxicity of conventional che-
motherapeutic drugs. Several angiogenesis inhibitors, having
demonstrated efficacy in a number of in vivo models, have
progressed through human clinical trials and gained FDA
approval.2 Two early clinical successes included the VEGFa
receptor antagonists bevacizumab, which gained approval for
nonsmall cell lung cancer and colorectal cancer, and pegapta-
nib, approved for age-related macular degeneration. Both
drugs are macromolecules (an antibody and a pegylated
aptamer, respectively); the first small-molecule approvals
included sorafenib for renal cell carcinoma (RCC) and suni-
tinib for RCC and gastrointestinal stromal tumor. With the
clinical validation of antiangiogenesis as a beneficial che-
motherapeutic strategy in place, the advantages of oral dosing
are driving further development of small molecule angiogen-
esis inhibitors.
The reports of the antiangiogenic activity of fumagillin and
a semisynthetic analogue 3, (TNP-470, AGM-1470), spurred
efforts to determine its molecular target in vivo. In 1997,
employing biotin or photoaffinity labeled conjugates of
fumagillin and the closely related natural product ovalicin,
two groups identified the molecular target of fumagillin as
the enzyme methionine aminopeptidase type II (MetAP2),
which cleaves the N-terminal methioninyl residue of newly
synthesized proteins (Figure 1a).3 Subsequent X-ray crystal-
lographic4 and biochemical studies5 identified a covalent
linkage of fumagillin and MetAP2 between His-231, located
Antineoplastic properties of the sesquiterpenoid natural
product fumagillin ((2E,4E,6E,8E)-10-((3R,4S,5S,6R)-5-
methoxy-4-((2R,3R)-2-methyl-3-(3-methylbut-2-enyl)oxiran-
2-yl)-1-oxaspiro[2.5]octan-6-yloxy)-10-oxodeca-2,4,6,8-tetra-
enoic acid, 1, Figure 1) were reported 50 years ago, but it was
first characterized as an inhibitor of angiogenesis by Folk-
man’s group in 1990.1 Angiogenesis, the process of new blood
vessel formation, has emerged as an attractive strategy for
cancer therapy based on the observation that tumor growth
was significantly limited in its absence. By selectively targeting
proliferating endothelial cells that promote neovasculariza-
tion in tumor tissue, this strategy appears to offer a means to
*To whom correspondence should beaddressed. Phone:781-795-4272.
Fax: 781-890-7015. E-mail: christopher.c.arico-muendel@gsk.com.
(C.A.M.); Christopher.self@providpharma.com (C.S.); lily5719@
yahoo.com (L.L.S.).
aAbbreviations: MetAP2, methionine aminopeptidase-2; VEGF, vas-
cular endothelial growth factor; MTT, methylthiazolyldiphenyl-tetra-
zolium bromide; CNS, central nervous system; EH, epoxide hydrolase;
HPLA, hydroxypropylmethacrylamide; PEG, polyethylene glycol;
NMR, nuclear magnetic resonance; TFA, trifluoroacetic acid; HPLC,
high performance liquid chromatography; DIEA, diisopropylethyla-
mine; LRMS, low resolution mass spectrometry; HEPES, 4-(2-hydro-
xyethyl)-1-piperazineethanesulfonic acid; HUVEC, human vascular
endothelial cell; BAEC, bovine aortic endothelial cell; ADME, absorp-
tion, distribution, metabolism, and excretion; t1/2, elimination half-life;
Cmax, maximum concentration achieved; Tmax, time of maximum con-
centration; AUC, area under curve; F(oral), oral bioavailability.
r
2009 American Chemical Society
Published on Web 11/24/2009
pubs.acs.org/jmc

8048 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 24
Arico-Muendel et al.
Figure 1. (A) Structures of fumagillin, fumagillol, ovalicin, and biotin-fumagillin conjugate.3b (B) Proposed mechanism for TNP-470 metabolism.12
in the enzyme active site, and C2 of the fumagillin spiro-
epoxide group. In particular, the crystal structure demonstrated
that interaction of the C4 side chain, bearing the second, less
reactive epoxide group, with the hydrophobic specificity
pocket of the enzyme served to position the spiroepoxide
precisely for attack by Nε2 of the histidine imidazole ring. In
agreement with the previous data, the polyolefinic side chain of
fumagillin extended out of the active site and was relatively
unconstrained.
Identification of MetAP2 as a potential target for antian-
giogenic therapy prompted widespread efforts to understand
its mechanism of action, employing both chemical and biolo-
gical approaches. The fumagillin analogue 3, disclosed by
Takeda, exhibited enhanced antiproliferative effects in vitro
and demonstrated potent activity against a broad array of
tumors in xenograft models.6 The dose dependence of inhibi-
tion in HUVEC, in particular, was shown to be biphasic, with
the first phase (GI50 = 37 pM) being cytostatic and fully
reversible, whereas the high dose phase was cytotoxic and
irreversible. In endothelial cells, MetAP2 inhibition was
shown to affect amino terminal protein processing and acti-
vate the p53 pathway, leading to accumulation of the CDK
inhibitor p21WAF1/CIP1 and G1 cell cycle arrest.7 The impor-
tance of MetAP-2 as the key regulator of endothelial cell
proliferation was demonstrated by the equivalent and non-
additive effects of MetAP-2 functional downregulation either
by siRNA or fumagillin and its analogues.8 Recently, Liu’s
group correlated MetAP1/2 inhibition with regulation of
c-Src and retardation of cell cycle progression.9
co-workers revealed that 3 was rapidly hydrolyzed in plasma
to an unsubstituted carbamate, which in turn was efficiently
degraded by epoxide hydrolase (EH) (Figure 1b).12 Caplos-
tatin and lodamin,13 comprising HPLA- and PEG-conjugates
of 3, respectively, addressed the PK liabilities and inhibited
tumor formation in animals. In pursuing a small molecule
strategy, we reasoned that replacement of the labile chloroa-
cetyl group could suppress EH mediated degradation.
Furthermore, the position of the C6 substituent away from
the active site of the enzyme offered an opportunity to
modulate the pharmacokinetic properties of the molecule
without affecting its ability to inhibit MetAP2. We thus
sought to incorporate functionality to minimize brain uptake
and restrict circulation of the drug to the periphery. Here, we
describe this design strategy and present SAR studies around
C6 substituted fumagillin analogues in enzymatic and cell
proliferation assays. The new compounds potently inhibit
MetAP2 with SAR that can be interpreted in terms of inter-
actions with the active site gatekeeper residues. Preliminary
PK and safety data showed significant improvements relative
to 3, validating the design approach.
Results and Discussion
Several groups have described syntheses and biological
activities of fumagillin analogues.14 In designing novel fuma-
gillin-based MetAP2 inhibitors that would improve upon the
in vivo properties of 3, we were guided by the following
interpretations of the structural and biochemical data: (1)
the cyclohexanespiroepoxide, or a reactive analogue, together
with the C4 side chain, constituted a necessary and sufficient
scaffold for potent inhibition, (2) the solvent exposed C6 side
chain offered the best opportunity for modifications that
would not compromise enzyme binding, and (3) the ADME
liabilities of 3, including both its rapid degradation by EH and
its uptake into the CNS, could be remedied by replacement of
the chloroacetyl group with a stable, more polar functionality.
We reasoned that fumagillol, readily accessible from commer-
cial fumagillin, represented a convenient source of the inhi-
bitory core. The crystal structure of MetAP-2 containing a
Several groups have employed fumagillin as starting points
for MetAP-2 targeting therapeutics.6 Among these, 3 pro-
gressed farthest, ultimately to phase 2 clinical studies, where it
showed some benefits in both single and combination dosing
for a number of cancers.2,6 Efficacy of 3 has also been also
demonstrated in rodent models of rheumatoid arthritis, an-
other disease characterized by an angiogenic component.10
However, the development of 3 has been hindered by poor
pharmacokinetics, lack of oral bioavailability, and dose-
limiting CNS side effects.11 Studies by Sommadossi and

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 24 8049
bound fumagillin molecule reveals two hydrophobic side
chains, Leu-328 and Leu-447, which serve as gatekeepers to
the active site.4 As a starting point for design, we chose
carbamates of branched, hydrophobic amino acids that could
interact with the leucine residues.15 Linkage to the fumagillin
core via a carbamate function, as opposed to an ester, was
expected to impart chemical stability and allowed for elabora-
tion of the amino acid carboxyl group as a variety of amides.
Synthesis of the initial series was accomplished in a straight-
forward way by saponifying commercially available (þ)-
fumagillin to fumagillol, which was then activated as a
p-nitrophenyl or N-hydroxysuccinimidyl carbonate ester
and then derivatized with amines. The HOSu carbonate ester
could be prepared on a multigram scale and purified by
precipitation from isopropyl alcohol, allowing a diverse array
of carbamates to be prepared (Scheme 1). Activities of these
constructs against MetAP2 and in endothelial cell prolifera-
tion are summarized in Table 1.
Scheme 1. Synthesis of Fumagillol Derivativesa
Methyl esters of both L- and D-valine (4, 5) were found to
potently inhibit proliferation in endothelial cells, as did struc-
tural homologues such as leucine (6) and phenylglycine (7).
D-Valine amide (8) and alcohol (9) were also highly active.
However, the D-valine free acid exhibited reduced activity in
both the biochemical and cellular assay, as well as reduced
chemical stability due to acid-mediated hydrolysis of the
spiroexpoxide group (data not shown). Collectively, the sim-
ple amino acid based carbamates equaled or exceeded the
activity of 3 in HUVEC growth inhibition.
In addition to antiproliferative activity, we compared the
rate of MetAP2 inhibition by these compounds in a biochem-
ical assay. MetAP-2 activity may be measured by quantita-
tion of methionine release from peptidic substrates such as
a Reagents: (a) NaOH, MeOH, H2O, 0 ꢀC f RT, 18 h; (b) p-nitro-
phenylchloroformate or DSC, TEA, CH3CN, RT; (c) NH-X, DIEA,
CH3CN, RT, 18 h.
Table 1. Inhibitory Activity of R-Amino Acid Analogues in MetAP2a and Cellular Growth Inhibitionb Assays
a Presented as percent activity remaining after 8 h treatment with compound. ND, not determined. b 50% inhibition in HUVEC and BAEC growth assays.

8050 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 24
Arico-Muendel et al.
Figure 2. Time dependence of MetAP2 activity upon treatment with selected compounds. C: ethanol control.
Table 2. Inhibitory Activity of Charged Analogues in MetAP2a and
Table 3. Inhibitory Activity of Dialkylated Amino Acid Analogues and
Truncates in MetAP2a and Cellular Growth Inhibitionb Assays
Cellular Growth Inhibitionb Assays
a Presented as percent activity remaining after 8 h treatment with
compound. b 50% inhibition in HUVEC and BAEC growth assays.
Therefore, we employed residual enzyme activity in the pre-
sence of excess inhibitor as a surrogate for potency. This
approach also allowed for comparisons with end-point-based
determinations of MetAP-2 activity in vivo.18
The scope of the carbamate-based strategy was explored
further by capping the C6 substituent with charged groups
that might interact with exterior solvent. Analogues in which
the D-valinamide moiety was extended with, or replaced by,
aminoethylpyrrolidines (10 and 12, and 11 and 13, re-
spectively) showed nearly complete inhibition of MetAP-2
after 8 h, as well as subnanomolar inhibition of HUVEC
proliferation (Figure 2, Table 2). Acyl hydrazines 14 and 15
also showed good activity in both assays. On the other hand,
quaternary amine compounds 16 and 17, while potent enzyme
inhibitors, were less effective in the HUVEC assay, presum-
ably because of poor cellular permeability, as were derivatives
18, 19, and 20, in which the D-Val moiety was extended or
replaced by carboxylate containing functionalities.
To examine further the scope of allowed hydrophobic
linkers, we replaced the D-Val moiety with two dialkylamino
acids, R,R-dimethyl- and cyclopentylglycine (21 and 22, re-
spectively). In comparison with valine, both of these linkers
resulted in a significant loss of inhibitory activity in both
assays (Table 3). Further simplification of the structures
yielded the isopropyl compound 23, which was highly active,
in contrast to the t-butyl compound 24, which showed only
30% inhibition after 8 h, thus localizing the origins of this
a Presented as percent activity remaining after 8 h treatment with
compound. b 50% inhibition in HUVEC and BAEC growth assays.
Met-Gly-Met-Met.16 However, this assay requires relatively
high enzyme concentrations (>20 nM), limiting its ability to
identify highly potent inhibitors.17 We found that while
MetAP2 activity diminished toward zero in the presence of
excess inhibitor, the rate of inhibition varied considerably
(Figure 2). This behavior is consistent with irreversible (slow
off-rate) inhibition superimposed on a range of on-rates.

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 24 8051
Figure 3. Structure of human MetAP2 (1BOA) in a complex with fumagillin (white), overlaid with docked structures of 8 (green) and 21
(yellow). Oxygen and nitrogen atoms are colored red and blue, respectively.
effect to the substitution of the R carbon. In the MetAP2
crystal structure, Leu-328 and -447 approach the polyolefinic
ester of fumagillin at carbons 3 and 4; presumably, similar
contacts with the carbamates are incompatible with tertiary
substitution of the group closest to the fumagillol scaffold,
whereas a broad variety of more distal substitutions are
tolerated.4 Binding models of several active molecules were
generated to test this theory by constraining their fumagillol
core to that of the crystal structure and minimizing the energy
of the compound plus surrounding residues. In agreement
with the hypothesis, the structure of 8 can be closely super-
imposed upon that of fumagillin, as shown in Figure 3. The
D-chirality at the amino acid may also be inverted to L without
noticeable alterations to the molecular geometry. However,
the dimethyl substituted compound 21 undergoes noticeable
conformational adjustments due to interactions with Asn327,
suggesting that the R,R-dialkyl substitution destabilizes the
bound form of the inhibitor.
The presence of reactive functionalities on fumagillin and
the covalent basis for its mechanism of action have led to
concerns about the viability of this compound class as a
therapeutic due to potential off-target toxicities. However,
probing cell lysates with immobilized fumagillin analogues
has generally failed to identify alternative targets (data not
shown). The issue may also be addressed by testing the
inherent reactivity of the spiroepoxide core against MetAP2.
We prepared 1-oxaspiro[2.5]octane, 25, which showed only
trace inhibition in the enzyme assay and did not inhibit
HUVEC proliferation (Figure 4). Thus, the potency of fuma-
gillin against MetAP2 results from a more complex
network of noncovalent interactions rather than from a
nonspecific alkylating capability. Taking an orthogonal
approach, Liu’s group recently reported two analogues of
fumagillin, fumagalone and fumarranol, in which the reactive
epoxide is replaced with retention of MetAP2 inhibition.17 In
fumagalone, the spiroepoxide function of fumagillin is re-
placed with a hydroxyaldehyde, and the C6 side chain with a
ketone, converting the scaffold into a reversible inhibitor that
retains its specificity for MetAP2 over MetAP1. In fumarra-
nol, an intramolecular alkylation opens the epoxide into
a hydroxymethyl group while bridging the cyclohexyl ring.
Both examples highlight the contribution to binding of the C4
substituent that mimics the side chain of the methionine
substrate.
Figure 4. Structure and activity of compound 25.
Following in vitro screening, a subset of the more potent
analogues were subjected to initial PK and toxicology studies.
In particular, we found that rats treated daily for 14 con-
secutive days with 8 at 6 and 60 mg/kg (iv) demonstrated a
significantly improved CNS toxicity profile over 3, measured
by incidence of seizures. While several animals treated with
6 mg/kg and all animals treated with 60 mg/kg 3 developed
seizures, none of the animals treated with 8 at either dose
developed seizures or showed other clinical signs of toxicity.18
This outcome supports the premise that the in vivo liabilities
of 3 could be addressed by reengineering a component of the
molecule not required for inhibition. In addition, 8 was
further tested in a cassette PK study along with 10, 11, and
12 (Table 4). All four compounds show a significant increase
in t1/2 relative to 3, whose reported half-life in human studies
was as short as 2 min.11b,d,f The oral bioavailabilities of all
four compounds based on parent molecule plasma concentra-
tion were found to be modest, although they could be some-
what improved by the addition of charged functionality.
Taken together, the results suggest that the carbamate-
based analogues either fail to enter the CNS due to their
higher polarity, or resist modification to brain penetrant
species that cause CNS toxicity. Additional studies of 8 have
led to the discovery of six active cytochrome P450-derived
metabolites that supplement the efficacy of the parent mole-
cule in vivo. A detailed characterization of these species,
including identification, chemical synthesis, and MetAP2
inhibitory activity, has been performed and will be reported
separately.

8052 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 24
Table 4. Pharmacokinetic Data for Selected Compounds in Rats
Arico-Muendel et al.
temperature and treated with another equivalent of NaOH
solution. The mixture was stirred until starting material was
not detected by TLC (∼4 h); the methanol was evaporated and
the residue extracted into ethyl acetate. The mixture was then
extracted with 5% citric acid, brine, bicarbonate, and brine
again, then dried with MgSO4 and concentrated in vacuo. The
crude product was further purified by treatment with activated
charcoal in acetonitrile and filtration through a celite pad.
Concentration yielded a colorless oil, 10.36 g (78%). 1H NMR
(CD3OD): δ 5.23 (t, 1H), 4.36 (m, 1H), 3.54 (dd, 1H), 3.42 (s,
3H), 2.92 (d, 1H), 2.64 (t, 1H), 2.52 (d, 1H), 2.2-2.4 (m, 2H),
2.15 (dd, 1H), 1.94 (d, 1H,), 1.7-1.9 (m, 2H), 1.75 (d, 3H), 1.67
(d, 3H), 1.19 (s, 3H), 0.96 (m, 1H).
dose
Cmax
AUC
(mg/ Tmax (ng/ (min ng/ t1/2 clearance F(oral)
a
route of
administration kg) (min) mL)
3
mL)
(min) (mL/min) (%)
8
8
iv
oral
iv
oral
iv
oral
iv
5
10
5
0
15
0
584
35
8540
1540
22400
7310
6660
2440
38900
8150
13/103
20/178
54
283
5760
220
9.0
16.3
18.3
10.5
10
10
11
11
12
12
456
27
10 120
196
27/120
168
946
742
5
10
5
0
15
0
246
15
3180
128
2640
31
16/67
233
oral
10
60
772
a Where applicable, shown as both initial and terminal half-lives.
2,5-Dioxopyrrolidin-1-yl (3R,4S,5S,6R)-5-methoxy-4-((2R,3R)-
2-methyl-3-(3-methylbut-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-
6-yl Carbonate (Fumagillol N-Hydroxysuccinimidyl Carbonate
Ester, 2c). Fumagillol (10.36 g, 36.68 mmol) was dissolved in
25 mL of acetonitrile and treated with triethylamine (15.34 mL,
110 mmol), and N,N0-disuccinimidyl carbonate (18.80 g,
73.37 mmol). The tan slurry was stirred at room temperature
under nitrogen for 4 h, then treated with celite, activated
charcoal, and anhydrous sodium sulfate and filtered through a
celite pad. The solution was concentrated, taken up in ethyl
acetate, and washed with bicarbonate solution, brine, 5% citric
acid, and brine, then dried over MgSO4 and concentrated to an
amber oil. The residue was precipitated from cold isopropyl
alcohol to yield a white solid, 8.33 g (54%). 1H NMR (CD3OD):
δ 5.54 (m, 1H), 5.23 (t, 1H), 3.78 (dd, 1H), 3.43 (s, 3H), 2.98 (d,
1H), 2.82 (s, 4H), 2.69 (t, 1H), 2.60 (d, 1H), 2.32 (m, 1H),
2.06-2.24 (m, 2H), 1.96 (m, 1H), 1.90 (d, 1H), 1.75 (d, 3H), 1.67
(d, 3H), 1.20 (s, 3H), 1.16 (m, 1H).
Substitution of 2b with Amines (General Procedure A). To a
mixture of carbonic acid-(3R, 4S, 5S, 6R)-5-methoxy-4-[(2R,
3R)-2-methyl-3-(3-methyl-but-2-enyl)-oxiranyl]-1-oxaspiro[2.5]-
oct-6-yl ester 4-nitro-phenyl ester (2b)14 (0.47 mmol) and amine
(2.35 mmol) in EtOH (9 mL) was added dropwise, diisopropyl
ethyl amine (2.35 mmol). After 3-18 h, the ethanol was removed
in vacuo and the crude material was dissolved into EtOAc
(10 mL) and washed with H2O (2 ꢀ 5 mL), and then brine
(5 mL). The organic phasewasdriedover Na2SO4 and the solvent
removed in vacuo. Purification via flash chromatography
(2-5% MeOH/CH2Cl2) afforded product.
Conclusion
We have reported a novel series of fumagillin analogues
containing carbamate-linked amines at C-6 of the oxirane
core. These analogues were designed to preserve the high
affinityof fumagillin for MetAP2while stabilizingthe scaffold
toward metabolism and limiting its CNS uptake. The com-
pounds show rapid inactivation of MetAP2 in a biochemical
assay, as well as potent antiproliferative activity against
HUVEC. As predicted by the fumagillin-MetAP2 crystal
structure, our data confirm that a wide variety of substituents
are tolerated at the C6 side chain; however, we demonstrate
that sterically demanding functions proximate to the cyclo-
hexane ring are not accommodated, and this result may be
explained by a docking study. The first generation lead
compound, 8 (PPI-2458), was found to potently inhibit
MetAP2 enzymatic activity and endothelial cell proliferation.
Notably, 8 also shows efficacy in animal models for melano-
ma and non-Hodgkin’s lymphoma19 as well as for rheumatoid
arthritis, following oral administration.18,20 These outcomes,
together with the lack of CNS toxicity, suggest that the major
pharmacokinetic and toxicological liabilities of 3 have been
remedied. Having demonstrated efficacy in multiple animal
models, 8 was advanced into proof-of-concept studies in man.
The collective in vivo findings for 8 will help to establish the
validity of MetAP2 inhibition as a novel approach for anti-
angiogenic therapy.
Substitution of 2c with Amines (General Procedure B). 2c
(2 mmol) was stirred in acetonitrile with amine salt (4.0 mmol)
in an ice bath. DIEA (4.0 mmol) was added, and the solution
stirred under nitrogen overnight. Extractive workup and pur-
ification was performed as in procedure A. For polar products,
purification was performed by preparative HPLC.
Experimental Section
General. Commercially available reagents and solvents were
used as purchased without further purification. Fumagillin
dicyclohexylamine salt was obtained from CEVA. Compound
3 was prepared according to the published procedure.6a Analy-
tical LC-MS was performed on a HP-1100 system interfaced
with a Finnegan LCQ Advantage electrospray mass spectro-
meter, using Phenomenex Luna C8(2) 3 μm columns with
acetonitrile/0.1% aqueous TFA elution. Preparative HPLC
was performed on a Gilson system with a Phenomenex Luna
C8(2), 5 μm, 100 mm ꢀ 21.2 mm column at a flow rate of
24 mL/min and detection at 214 nm. To avoid acid catalyzed
degradation, preparative LC was performed without TFA in the
eluting solvents. Thin layer chromatography was performed
on 250 μm EM Science silica gel 60 plates using UV254 or
iodine visualization. NMR spectra were obtained on a Varian
MercuryPlus 400 MHz spectrometer referenced to TMS at
0.0 ppm. In general, final compound purities were 95% or
greater as determined by HPLC-MS.
Synthesis of Amino Amides (General Procedure C). Boc-
amino acid (2.0 mmol) was dissolved in 3 mL THF, and stirred
in an ice bath under nitrogen. N-Methyl morpholine (220 μL,
2.0 mmol) and isobutylchloroformate (260 μL, 2.0 mmol) were
added, and the mixture was stirred for 5 min. A white precipitate
was removed by filtration, and the filtrate was treated with
amine. The solution was stirred at 0 ꢀC for 20 min and then
warmed to room temperature and stirred 20 min more. The
solution was concentrated in vacuo and reduced to a solid under
high vacuum. It was then treated with 20 mL of 4 M HCl in
dioxane for an hour at room temperature. Concentration and
trituration with ether yielded a white powder of acceptable
purity for acylation with 2c. Alternatively, 2c was used to
N-acylate the free amino acid, and the product was then con-
verted to a substituted amide as a separate step.
(S)-Methyl 2-(((3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-methyl-
3-(3-methylbut-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-yloxy)-
carbonylamino)-3-methylbutanoate (4). General procedure A
was followed using 2b (31 mg, 0.07 mmol), L-valine methyl
ester hydrochloride (58 mg, 0.35 mmol), and DIEA (60 μL,
0.35 mmol) in EtOH (2 mL). Purification via flash chromato-
graphy (1% MeOH/CH2Cl2) afforded the product as a clear oil
(3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-methyl-3-(3-methyl-
but-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-ol, (fumagillol,
2a). Fumagillin dicyclohexylamine salt (30 g, 47.0 mmol) was
suspended in 300 mL of 1:1 methanol:water and treated with
10.7 mL of 35% NaOH solution (94 mmol). The dark-brown
mixture was stirred in an ice bath for 2 h, then warmed to room

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 24 8053
(10 mg, 0.02 mmol, 33% yield); Rf = 0.60 (20% EtOAc/
CH2Cl2). LRMS (m/z) [M þ 1]þ 440.3 (calculated for C23H38-
NO7, 440.3). 1H NMR (CD3OD; conformers due to slow
carbamate cis-trans interconversion indicated as a second
chemical shift): δ 5.47, 5.04 (s, 1H), 5.24 (t, 1H), 4.07, 4.00
(d, 1H), 3.71 (s, 3H), 3.68 (dd, 1H), 3.41 (s, 3H), 2.96 (d, 1H),
2.65 (t, 1H), 2.56 (d, 1H), 2.32 (m, 1H), 2.23 (m, 1H), 2.11 (m,
2H), 2.00 (d, 1H), 1.95 (m, 1H), 1.82 (m, 1H), 1.76 (s, 3H), 1.67
(s, 3H), 1.20 (s, 3H), 1.06 (m, 1H), 1.00, 0.94(m, 6H).
(R)-Methyl 2-(((3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-meth-
yl-3-(3-methylbut-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-
yloxy)carbonylamino)-3-methylbutanoate (5). General proce-
dure A was followed using 2b (41 mg, 0.09 mmol) and D-valine
methyl ester hydrochloride (77 mg, 0.45 mmol) and DIEA
(80 μL, 0.45 mmol) in EtOH (2 mL). Purification via flash
chromatography (1% MeOH/CH2Cl2) afforded the product as
a clear oil (18 mg, 0.04 mmol, 45% yield); Rf = 0.39 (20%
EtOAc/CH2Cl2. LRMS (m/z) [M þ 1]þ 440.3 (calculated for
C23H38NO7, 440.3). 1H NMR (CD3OD): δ 7.24 (d, 1H), 5.42 (m,
1H), 5.24 (m, 1H), 4.07 (m, 1H), 3.71 (s, 3H), 3.67 (dd, 1H), 3.38
(s, 3H), 2.96 (d, 1H), 2.64 (t, 1H), 2.56 (d, 1H), 2.30 (m, 1H),
2.21 (m, 1H). 2.12 (m, 2H), 2.01 (d, 1H), 1.97 (m, 1H), 1.82
(m, 1H), 1.75 (s, 3H), 1.67 (s, 3H), 1.19 (s, 3H), 1.04 (m, 1H),
0.95 (m, 6H).
(3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-methyl-3-(3-methyl-
but-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-yl (R)-1-Hydro-
xy-3-methylbutan-2-ylcarbamate (9). General procedure A was
followed using 2b (290 mg, 0.65 mmol), D-valinol (337 mg, 3.25
mmol), and DIEA (560 μL, 3.25 mmol) in EtOH (5 mL).
Purification via flash chromatography (2% MeOH/CH2Cl2)
afforded the product as a clear oil (200 mg, 0.49 mmol, 75%
yield); Rf = 0.26 (2% MeOH/CH2Cl2). LRMS (m/z) [M þ 1]þ
412.5 (calculated for C22H38NO6, 412.5). 1H NMR (CD3OD): δ
5.42 (m, 1H), 5.23 (t, 1H), 3.67 (dd, 1H), 3.59 (m, 1H), 3.52 (m,
1H), 3.40 (m, 1H), 3.40 (s, 3H), 2.96 (d, 1H), 2.63 (t, 1H), 2.56 (d,
1H), 2.31 (m, 1H), 2.23 (m, 1H), 2.12 (m, 1H), 1.99 (d, 1H), 1.98
(m, 1H), 1.84 (m, 2H), 1.75 (s, 3H), 1.67 (s, 3H), 1.20 (s, 3H), 1.05
(m, 1H), 0.94 (m, 6H).
(3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-methyl-3-(3-methyl-
but-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-yl (2R)-3-Methyl-
1-(2-(1-methylpyrrolidin-2-yl)ethylamino)-1-oxobutan-2-ylcarbamate
(10). General procedure C was followed. LRMS (m/z) [M þ 1]þ
1
536.4 (calculated for C29H49N3O6, 536.4). H NMR (CD3OD):
δ 5.43 (m, 1H), 5.24 (t, 1H), 3.84 (d, 1H), 3.67 (dd, 1H), 3.41
(s, 3H), 3.26 (m, 2H), 3.04 (m, 1H), 2.96 (d, 1H), 2.65 (t, 1H), 2.56 (d,
1H), 2.31 (m, 3H), 2.2 (m, 4H), 2.0 (m, 5H), 1.8 (m, 4H), 1.76
(s, 3H), 1.68 (s, 3H), 1.5 (m, 1H), 1.4 (m, 1H), 1.20 (s, 3H), 1.04
(m, 1H), 0.96 (d, 6H).
(R)-Methyl 2-(((3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-methyl-
3-(3-methylbut-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-
yloxy)carbonylamino)-4-methylpentanoate (6). General proce-
dure A was followed using 2b (23 mg, 0.05 mmol), D-leucine
methyl ester hydrochloride (47 mg, 0.25 mmol), and DIEA
(45 μL, 0.25 mmol) in EtOH (2 mL). Purification via flash
chromatography (1% MeOH/CH2Cl2) afforded the product as
a clear oil (19 mg, 0.04 mmol, 83% yield); Rf = 0.22 (15%
EtOAc/CH2Cl2). LRMS (m/z) [M þ 1]þ 454.3 (calculated for
(3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-methyl-3-(3-methyl-
but-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-yl 2-(1-Methylpyr-
rolidin-2-yl)ethylcarbamate (11). General procedure B was
followed. LRMS (m/z) [M þ 1]þ 437.3 (calculated for C24H40-
N2O5, 437.3). 1H NMR (CD3OD): δ 5.44 (m, 1H), 5.24 (t, 1H),
3.66 (dd, 1H), 3.42 (s, 3H), 3.14 (m, 2H), 3.07 (m, 1H), 2.96
(d, 1H), 2.61 (t, 1H), 2.34 (d, 1H), 2.35 (s, 3H), 2.30 (m, 1H), 2.23
(m, 3H), 2.10 (m, 2H), 1.94 (m, 3H), 1.92 (m, 3H), 1.76
(s, 3H), 1.67 (s, 3H), 1.49 (m, 1H), 1.41 (m, 1H), 1.20 (s, 3H),
1.05 (m, 1H).
1
C24H40NO7, 454.3). H NMR (CD3OD): δ 5.43 (m, 1H), 5.23
(t, 1H), 4.22 (m, 1H), 3.70 (s, 3H), 3.67 (dd, 1H), 3.40 (s, 3H),
2.96 (d, 1H), 2.63 (t, 1H), 2.56 (d, 1H), 2.32 (m, 1H), 2.23 (m,
1H), 2.15 (m, 1H), 1.98 (d, 1H), 1.93 (m, 1H), 1.82 (m, 1H), 1.76
(s, 3H), 1.72 (m, 1H), 1.67 (s, 3H), 1.57 (m, 2H), 1.19 (s, 3H), 1.06
(s, 1H), 0.94 (m, 6H).
(3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-methyl-3-(3-methyl-
but-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-yl (R)-3-Methyl-
1-oxo-1-(2-(pyrrolidin-1-yl)ethylamino)butan-2-ylcarbamate (12).
General procedure C was followed. LRMS (m/z) [M þ 1]þ 522.6
(calculated for C28H47N3O6, 522.3). 1H NMR (CD3OD): δ 5.43
(s, 1H), 5.24(t, 1H), 3.84 (d, 1H), 3.68 (dd, 1H), 3.53 (m, 1H), 3.40
(m, 1H), 3.40 (s, 3H), 3.03 (m, 6 H), 2.96 (d, 1H), 2.64 (t, 1H), 2.56
(d, 1H), 2.31 (m, 1H), 2.24 (m, 1H), 2.15 (dt, 1H), 2.08 (d, 1H),
2.06 (d, 1H), 1.96 (m, 5H), 1.84 (m, 1H), 1.76 (s, 3H), 1.67 (s, 3H),
1.19 (s, 3H), 1.08 (m, 1H), 0.97 (d, 6H).
(3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-methyl-3-(3-methyl-
but-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-yl 2-(Pyrrolidin-
1-yl)ethylcarbamate (13). General procedure B was followed.
LRMS (m/z) [M þ 1]þ 423.3 (calculated for C23H38N2O5, 423.3).
1H NMR (CD3OD): δ 5.45 (m, 1H), 5.24 (m, 1H), 3.66 (dd, 1H),
3.42 (s, 3H), 3.26 (m, 2H), 2.96 (d, 1 H), 2.60 (m, 7 H), 2.56 (d,
1H), 2.31 (m, 1H), 2.23 (m, 1H), 2.10 (m, 1H), 1.94 (m, 2H), 1.81
(m, 5 H), 1.76 (s, 3H), 1.67 (s, 3H), 1.20 (s, 3H), 1.05 (m, 1H).
(3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-methyl-3-(3-methyl-
but-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-yl (R)-3-Methyl-
1-(morpholinoamino)-1-oxobutan-2-ylcarbamate (14). General
procedure C was followed. LRMS (m/z) [M þ 1]þ 510.5
(calculated for C26H43N3O7, 510.3). 1H NMR (CD3OD): δ
5.42 (m, 1H), 5.24 (m, 1H), 3.74 (m, 4H), 3.66 (m, 1H), 3.40 (s,
3H), 3.30 (m, 1H), 2.96 (d, 1H), 2.78 (t, 4H), 2.64 (t, 1H), 2.56 (d,
1H), 2.30 (m, 1H), 2.23 (m, 1H), 2.13 (m, 1H), 1.99 (m, 2H), 1.92
(m, 1H), 1.80 (m, 1H), 1.75 (s, 3H), 1.67 (s, 3H), 1.19 (s, 3H), 1.03
(m, 1H), 0.96 (m, 6H).
(R)-Methyl 2-(((3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-meth-
yl-3-(3-methylbut-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-
yloxy)carbonylamino)-2-phenylacetate (7). General procedure A
was followed using 2b (37 mg, 0.08 mmol), D-phenyl glycine
methyl ester hydrochloride (83 mg, 0.40 mmol), and DIEA
(72 μL, 0.40 mmol) in EtOH (2 mL). Purification via flash
chromatography (1% MeOH/CH2Cl2) afforded the product as
a clear oil (32 mg, 0.07 mmol, 82% yield); Rf = 0.41 (2%
MeOH/CH2Cl2). LRMS (m/z) [M þ 1]þ 474.3 (calculated for
C26H36NO7, 474.3). 1H NMR (CD3OD; conformers due to slow
carbamate cis-trans interconversion indicated as a second
chemical shift): δ 7.36 (m, 5H), 5.44, 5.37 (m, 1H), 5.28
(s, 1H), 5.22 (t, 1H), 3.70 (s, 3H), 3.66 (dd, 1H), 3.36 (s, 3H),
2.96 (d, 1H), 2.62 (t, 1H), 2.56 (d, 1H), 2.31 (m, 1H), 2.20 (m,
2H), 1.98 (d, 1H), 1.83 (m, 1H), 1.70 (s, 3H), 1.70 (m, 1H), 1.66
(s, 3H), 1.18 (s, 3H), 1.06 (m, 1H).
(3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-methyl-3-(3-methyl-
but-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-yl (R)-1-Amino-
3-methyl-1-oxobutan-2-ylcarbamate (8). General procedure A
was followed using 2b (55 mg, 0.12 mmol), D-valine amide
hydrochloride (93 mg, 0.62 mmol), and DIEA (110 μL, 0.62
mmol) in EtOH (2 mL). Purification via flash chromatography
(2% MeOH/CH2Cl2) afforded the product as a clear oil (42 mg,
0.10 mmol, 80% yield); Rf = 0.19 (2% MeOH/CH2Cl2). LRMS
(3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-methyl-3-(3-methyl-
but-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-yl Morpholinocar-
bamate (15). General procedure B was followed. LRMS (m/z)
1
(m/z) [M þ 1]þ 425.5 (calculated for C22H37N2O6, 425.5). H
NMR (CD3OD; conformers due to slow carbamate cis-trans
interconversion indicated as a second chemical shift): δ 5.43 (m,
1H), 5.24 (t, 1H), 3.94 (d, 1H), 3.67 (dd, 1H), 3.40 (s, 3H), 2.96
(d, 1H), 2.64 (t, 1H), 2.56 (d, 1H), 2.31 (m, 1H), 2.22 (m, 1H),
2.14 (m, 1H), 1.97 (d, 1H), 1.96 (m, 1H), 1.80 (m, 1H), 1.75 (s,
3H), 1.67 (s, 3H), 1.19 (s, 3H), 1.04 (m, 1H), 0.99, 0.96 (m, 6H).
1
[M þ 1]þ 411.1 (calculated for C21H34N2O6, 411.2). H NMR
(CD3OD) δ 5.50 (s, 1H), 5.23 (t, 1H), 3.73 (t, 4H), 3.67 (dd, 1H),
3.42 (s, 3H), 2.96 (1H), 2.77 (s, 4H), 2.62 (t, 1H), 2.57 (d, 1H), 2.31
(m, 1H), 2.24 (m, 1H), 2.09 (m, 1H), 1.92 (m, 2H), 1.84
(m, 1H), 1.75 (s, 3H), 1.67 (s, 3H), 1.19 (s, 3H), 1.06 (m, 1H).

8054 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 24
Arico-Muendel et al.
2-(2-(((3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-methyl-3-
(3-methylbut-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-yloxy)-
carbonylamino)ethyl)-1,1-dimethylpyrrolidinium Iodide (16). To
a solution of 11 (52 mg, 0.119 mmol) in toluene (0.3 mL) was
added iodomethane (0.89 mL, 14.4 mmol), and the mixture was
stirred overnight. The resulting precipitate was washed with
ether and purified by recrystallization from methanol/ether.
LRMS (m/z) [M]þ 451.4 (calculated for C25H43N2O5þ, 451.3).
Elemental analysis: C, 51.81; H, 7.46; I, 21.96; N, 4.65
(calculated for C25H43IN2O5: C, 51.90; H, 7.49; I, 21.94; N,
4.84; O, 13.83).
3-(((3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-methyl-3-(3-methyl-
but-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-yloxy)carbony-
lamino)-N,N,N-trimethylpropan-1-aminium Iodide (17). The
chemistry was conducted as for 16, starting with the correspond-
ing dimethylamino compound, and using 40 equiv of iodo-
methane. LRMS (m/z) [M]þ 425.3 (calculated for C23H41-
N2O5þ, 425.3); Elemental analysis: C, 50.13; H, 7.42; I, 22.80;
N, 4.94 (calculated for C23H41IN2O5: C, 50.00; H, 7.48; I, 22.97;
N, 5.07; O, 14.48).
4-((((3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-methyl-3-(3-methyl-
but-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-yloxy)carbony-
lamino)methyl)benzoic Acid (18). General procedure B was
followed. LRMS (m/z) [M þ 1]þ 460.1 (calculated for C25H33-
NO7, 460.2). Elemental analysis: C, 65.52; H, 7.69; N, 3.88
(calculated for C25H33NO7: C, 65.34; H, 7.24; N, 3.05; O, 24.37).
1H NMR (CD3OD): δ 8.01 (s, 2H), 7.35 (s, 2H), 5.48 (s,1H), 5.23
(t, 1H), 4.34 (m, 2H), 3.68 (d, 1H), 3.43 (s, 3H), 2.96 (d, 1H), 2.62
(t, 1H), 2.55 (d, 1H), 2.32 (m, 1H), 2.17 (m, 1H), 2.12 (m, 1H),
1.98 (m, 2H), 1.84 (m, 1H), 1.75 (s, 3H), 1.66 (s, 3H), 1.19 (s, 3H),
1.06 (m, 1H).
3-((R)-2-(((3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-methyl-
3-(3-methylbut-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-yloxy)-
carbonylamino)-3-methylbutanamido)benzoic Acid (19). General
procedure C was followed. LRMS (m/z) [M þ 1]þ 545.1 (cal-
culated for C29H40N2O8, 545.3). Elemental analysis: C, 63.71; H,
7.28; N, 5.03 (calculated for C29H40N2O8: C, 63.95; H, 7.40; N,
5.14; O, 23.50). 1H NMR (CD3OD): δ 7.93 (s, 1H), 7.83 (d, 1H),
7.70 (d, 1H), 7.32 (t, 1H), 5.45 (s, 1H), 5.24 (t, 1H), 4.07 (d, 1H),
3.68 (d, 1H), 3.42 (s, 1H), 2.96 (d, 1H), 2.66 (t, 1H), 2. 56 (d, 1H),
2.31 (m, 1 H), 2.17 (m, 3H), 2.02 (d, 1H), 1.98 (m, 1H), 1.80 (m,
1H), 1.76 (s, 3H), 1.67 (s, 3H), 1.20 (s, 3 H), 1.02 (m, 7H).
1-(((3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-methyl-3-(3-
methylbut-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-yloxy)carbonyl)-
piperidine-4-carboxylic Acid (20). General procedure B was
followed. LRMS (m/z) [M þ 1]þ 438.2 (calculated for C23H35-
NO7, 438.2). Elemental analysis: C, 62.94; H, 7.97; N, 3.15
(calculated for C23H35NO7: C, 63.14; H, 8.06; N, 3.20; O, 25.60).
1H NMR (CD3OD): δ 5.48 (s, 1H), 5.24 (t, 1H), 4.06 (d, 2H),
3.76 (dd, 1H), 3.41 (s, 3H), 2.98 (d, 1H), 2.90 (m, 2H), 2.68 (t,
1H), 2.56 (d, 1H), 2.30 (m, 2H), 2.22 (m, 1H), 2.09 (m, 1H), 1.90
(m, 5H), 1.75 (s, 3H), 1.67 (s, 3H), 1.59 (m, 2H), 1.19 (s, 3H), 1.08
(d, 1H).
(3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-methyl-3-(3-methyl-
but-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-yl Isopropylcar-
bamate (23). General procedure B was followed. LRMS (m/z)
[M þ 1]þ 368.4 (calculated for C20H33NO5, 368.2). 1H NMR
(CD3OD): δ 5.43 (s, 1H), 5.23 (t, 1H), 3.70 (m, 1H), 3.66 (dd, 1H),
3.43 (s, 3H), 2.96 (d, 1H), 2.61 (t, 1H), 2.56 (d, 1H), 2.31 (m, 1H),
2.19 (m, 1H), 2.10 (dt, 1H), 1.94 (d, 1H), 1.90 (m, 1H), 1.82 (m, 1H),
1.75 (s, 3H), 1.67 (s, 3H), 1.19 (s, 3H), 1.13 (d, 6H), 1.05 (m, 1H).
(3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-methyl-3-(3-methyl-
but-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-yl tert-Butylcar-
bamate (24). General procedure B was followed. LRMS (m/z)
1
[M þ 1]þ 382.3 (calculated for C21H35NO5, 382.2). H NMR
(CD3OD): δ 5.37 (s, 1H), 5.23 (t, 1H), 3.65 (dd, 1H), 3.41 (s, 3H),
2.95 (d, 1H), 2.62 (t, 1H), 2.55 (d, 1H), 2.31 (m, 1H), 2.22 (m,
1H), 2.10 (dt, 1H), 1.96 (d, 1H), 1.92 (m, 1H), 1.80 (m, 1H), 1.75
(s, 3H), 1.67 (s, 3H), 1.29 (s, 9 H), 1.19 (s, 3H), 1.04 (m, 1H).
1-Oxaspiro[2.5]octane (25). To a solution of MCPBA (5.6 g,
32.6 mmol) in CHCl3 was added dropwise methylene cyclohex-
ane (3.75 mL, 31 mmol). The mixture was stirred 16 h at 4 ꢀC,
after which the chloroform was partially evaporated under a
nitrogen stream. The residue was taken up in ether (50 mL) and
washed with sodium bicarbonate. The product was purified by
distillation (bp 70 ꢀC), yield 1.24 g (34%). LRMS (m/z) [M þ 1]þ
1
113.1 (calculated for C7H12O, 113.1). H NMR (CD3OD): δ
2.59 (s, 2H), 1.68 (m, 2H), 1.60 (m, 2H), 1.58 (m, 2H), 1.50 (m,
2H), 1.46 (m, 2H). 13C NMR (CD3OD): δ 54.1 (CH2), 59.1 (C),
33.3 (CH2), 24.6 (CH2), 25.0 (CH2), assigned by HSQC.
Molecular Modeling. Binding poses were generated using the
program MOE (Chemical Computing Group) together with the
cocrystal structure of human MetAP2 with fumagillin (1BOA).4
MetAP-2 Inhibition. Compounds were prepared as 800 nM
solutions in reaction buffer (1% glycerol, 20 mM HEPES,
100 mM KCl, 0.1 mM CoSO4) and combined with equal
volumes of 80 nM MetAP2. Samples were taken in triplicate at
1, 2, 4, and 8 h time points and treated with 2ꢀ Met-AMC
substrate to determine residual activity. Final concentrations of
MetAP2 and compound were 20 and 200 nM, respectively.
HUVEC Assay. Human umbilical vein endothelial cells
(HUVEC) were cultured in HUVEC growth medium
(Clonetics) at 37 ꢀC. For proliferation assays, 2000 cells
(20000 cells/mL) (passage 3-5) were seeded in 96-well plates
(in triplicate). After cell attachment, fresh medium containing 8
was added for 72 h. To each well, 0.25 mg per mL of MTT was
added, and the plates were incubated for 2 h at 37 ꢀC. The MTT
was removed, and the crystals were solubilized with isopropyl
alcohol. After 10 min, the intensity was measured colorimetri-
cally at a wavelength of 570 nm. HUVEC proliferation was
determined by the conversion of MTT to water-insoluble MTT-
formazan by mitochondrial dehydrogenases of living cells.
BAEC Assay. The bovine aortic endothelial growth assay was
performed as described by Turk et al.7a
Cassette PK Study. Two groups of 8 male Sprague-Dawley
rats (aged 8-9 weeks, Taconic) were dosed with multiple
compounds at 10 mg/kg compound (po) or 5 mg/kg (iv), res-
pectively, in 5% HPCD formulation containing 2 mg/mL
compound. Approximately 0.5 mL of blood was drawn by jug
stick from four animals at alternating time points as follows:
2 min, 5 min, 10 min, 15 min, 30 min, 60 min, 2 h, 4 h, 6 h.
Samples were treated with EDTA, centrifuged, and frozen.
Compounds were quantitated from plasma samples by LCMS.
(3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-methyl-3-(3-methyl-
but-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-yl 1-Amino-2-
methyl-1-oxopropan-2-ylcarbamate (21). General procedure B
was followed. LRMS (m/z) [M þ 1]þ 411.6 (calculated for
1
C21H34N2O6, 411.2). H NMR (CD3OD): δ 5.45 (s, 1H), 5.24
(t, 1H), 3.65 (d, 1H), 3.40 (s, 3H), 2.96 (d, 1H), 2.64 (m, 1H), 2.56
(d, 1H), 2.32 (m, 1H), 2.23 (m, 1H), 2.15 (dt, 1H), 1.98 (d, 1H),
1.88 (m, 1H), 1.84 (m, 1H), 1.76 (s, 3H), 1.68 (s, 3H), 1.46 (s, 3H),
1.42 (s, 3H), 1.18 (s, 3H), 1.04 (1H).
Acknowledgment. We thank Dr. Sylvie Bernier for her
thoughtful review of the manuscript.
(3R,4S,5S,6R)-5-Methoxy-4-((2R,3R)-2-methyl-3-(3-methyl-
but-2-enyl)oxiran-2-yl)-1-oxaspiro[2.5]octan-6-yl 1-Carbamoyl-
cyclopentylcarbamate (22). General procedure B was followed.
LRMS (m/z) [M þ 1]þ 437.5 (calculated for C23H36N2O6,
References
1
437.3). H NMR (CD3OD): δ 5.45 (s, 1H), 5.24 (t, 1H), 3.66
(1) (a) DiPaolo, J. A.; Tarbell, D. S.; Moore, G. E. Studies on the
carcinolytic activity of fumagillin and some of its derivatives.
Antibiot. Annu. 1958-1959, 6, 541–546. (b) Ingber, D.; Fujita, T.;
Kishimoto, S.; Sudo, K.; Kanamaru, T.; Brem, H.; Folkman, J.
(dd, 1H), 3.40 (s, 3H), 2.96 (d, 1H), 2.63 (t, 1H), 2.56 (d, 1H),
2.31 (m, 2H), 2.18 (m, 2H), 1.99 (m, 3H), 1.87 (m, 3H), 1.76 (m,
4H), 1.76 (s, 3H), 1.68 (s, 3H), 1.18 (s, 3H), 1.04 (d, 1H).

Article
Synthetic analogues of fumagillin that inhibit angiogenesis and sup-
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 24 8055
Sosman, J.; Clark, J.; Vogelzang, N. J. Multi-Institutional Study of the
Angiogenesis Inhibitor TNP-470 in Metastatic Renal Carcinoma.
J. Clin. Oncol. 1999, 17, 2541–2545. (f) Moore, J. D.; Dezube, B. J.;
Gill, P.; Zhou, X.-J.; Acosta, E. P.; Sommadossi, J.-P. Phase I dose
escalation pharmacokinetics of O-(chloroacetylcarbamoyl)-fumagillin
(TNP-470) and its metabolites in AIDS patients with Kaposi's sarcoma.
Cancer Chemother. Pharmacol. 2000, 46, 173–179. (g) Logothetis, C.
J.; Wu, K. K.; Finn, L. D.; Daliani, D.; Figg, W.; Ghaddar, H.;
Gutterman, J. U. Phase I trial of the angiogenesis inhibitor TNP-470
for progressive androgen-independent prostate cancer. Clin. Cancer
Res. 2001, 7, 1198–1203.
press tumor growth. Nature 1990, 348, 555–557.
(2) (a) Folkman, J. Angiogenesis: an organizing principle for drug
discovery? Nature Rev. Drug Discovery 2007, 6, 273–286. (b) Ribatti,
D. The discovery of antiangiogenic molecules: a historical review.
Curr. Pharm. Des. 2009, 15, 345–352.
(3) (a) Sin, N.; Meng, L.; Wang, M. Q. W.; Wen, J. J.; Bornmann, W.
G.; Crews, C. M. The anti-angiogenic agent fumagillin covalently
binds and inhibits the methionine aminopeptidase, MetAP-2. Proc.
Natl. Acad. Sci. U.S.A. 1997, 94, 6099–6103. (b) Griffith, E. C.; Su,
Z.; Turk, B. E.; Chen, S.; Chang, Y.-H.; Wu, Z.; Biemann, K.; Liu, J. O.
Methionine aminopeptidase (type 2) is the common target for
angiogenesis inhibitors AGM-1470 and ovalicin. Chem. Biol. 1997, 4,
461–471.
(4) Liu, S.; Widom, J.; Kemp, C. W.; Crews, C. M.; Clardy, J.
Structure of human methionine aminopeptidase-2 complexed with
fumagillin. Science 1998, 282, 1324–1327.
(5) Griffith, E. C.; Su, Z.; Niwayama, S.; Ramsay, C. A.; Chang, Y.-
H.; Liu, J. O. Molecular recognition of angiogenesis inhibitors
fumagillin and ovalicin by methionine aminopeptidase 2. Proc.
Natl. Acad. Sci. U.S.A. 1998, 95, 15183–15188.
(12) (a) Moore, J. D.; Sommadossi, J.-P. ACTG 215 Study Team.
Determination of O-(chloroacetylcarbamoyl)fumagillol (TNP-
470; AGM-1470) and two metabolites in plasma by high-perfor-
mance liquid chromatography/mass spectrometry with atmo-
spheric pressure chemical ionization. J. Mass Spectrom. 1995, 30,
1707–1715. (b) Placidi, L.; Cretton-Scott, E.; De Sousa, G.; Rahmani,
R.; Placidi, M.; Sommadossi, J.-P. Disposition and metabolism
of the angiogenic moderator O-(chloroacetyl-carbamoyl) fumagillol
(TNP-470; AGM-1470) in human hepatocytes and tissue microsomes.
Cancer Res. 1995, 55, 3036–3042.
(6) (a) Marui, S.; Itoh, F.; Kozai, Y.; Sudo, K.; Kishimoto, S.
Chemical Modification of Fumagillin. I. 6-O-Acyl, 6-O-Alkyl,
and 6-O-(N-Substituted-carbamoyl)fumagillols. Chem. Pharm.
Bull. 1992, 40, 96–101. (b) Bernier, S. G.; Westlin, W. F.; Hannig, G.
Fumagillin class inhibitors of methionine aminopeptidase-2. Drugs
Future 2005, 30, 497–508. (c) Lee, H.; Cho, C.; Kang, S.; Yoo, Y.; Shin,
J.; Ahn, S. Design, synthesis, and antioangiogenic effects of a series of
potent novel fumagillin analogues. Chem. Pharm. Bull. 2007, 55,
1024–1029. (d) Zhong, H.; Bowen, J. P. Antiangiogenesis drug design:
multiple pathways targeting tumor vasculature. Curr. Med. Chem.
2006, 13, 849–862. (e) Folkman, J. Tumor angiogenesis. In Accom-
plishments in Cancer Research 1997; Wells, S. A., Sharp, P. A., Eds.;
Lippincott Williams & Wilkins: Philadelphia, 1998; pp 32-44.
(f) Lefkove, B.; Govindarajan, B.; Arbiser, J. Fumagillin: an anti-
infective as a parent molecule for novel angiogenesis inhibitors. Expert
Rev. Anti-Infect. Ther. 2007, 5, 573–579.
(7) (a) Turk, B.; Griffith, E.; Wolf, S.; Biemann, K.; Chang, Y.-H.; Liu,
J. Selective inhibition of amino-terminal methionine processing by
TNP-470 and ovalicin in endothelial cells. Chem. Biol. 1999, 11,
823–833. (b) Zhang, Y.; Griffith, E. C.; Sage, J.; Jacks, T.; Liu, J. O. Cell
cycle inhibition by the anti-angiogenic agent TNP-470 is mediated
by p53 and p21WAF1/CIP1. Proc. Natl. Acad. Sci. U.S.A. 2000, 97,
6427–6432.
(8) Bernier, S. G.; Taghizadeh, N.; Thompson, C. D.; Westlin, W. F.;
Hannig, G. Methionine aminopeptidases type I and type II are
essential to control cell proliferation. J. Cell. Biochem. 2005, 95,
1191–1203.
(9) Hu, X.; Dang, Y.; Tenney, K.; Crews, P.; Tsai, C.; Sixt, K.; Cole, P.;
Liu, J. Regulation of c-Src nonpeptide tyrosine kinase activity by
bengamide A through inhibition of methionine aminopeptidases.
Chem. Biol. 2007, 14, 764–774.
(10) (a) Peacock, D. J.; Banquerigo, M. L.; Brahn, E. Angiogenesis
inhibition suppresses collagen arthritis. J. Exp. Med. 1992, 175,
1135–1138. (b) Oliver, S. J.; Banquerigo, M. L.; Brahn, E. Suppression
of collagen induced arthritis using an angiogenesis inhibitor, AGM-
1470, and a microtubule stabilizer, Taxol. Cell. Immunol. 1994, 157,
291–299. (c) Peacock, D. J.; Banquerigo, M. L.; Brahn, E. A novel
angiogenesis inhibitor suppresses rat adjuvant arthritis. Cell. Immunol.
1995, 160, 178–184. (d) Oliver, S. J.; Cheng, T. P.; Banquerigo, M. L.;
Brahn, E. Suppression of collagen-induced arthritis by an angiogenesis
inhibitor, AGM-1470, in combination with cyclosporin: reduction of
vascular endothelial growth factor (VEGF). Cell. Immunol. 1995, 166,
196–206.
(13) (a) Satchi-Fainaro, R.; Puder, M.; Davies, J. W.; Tran, H. T.;
Sampson, D. A.; Greene, A. K.; Corfas, G.; Folkman, J. Targeting
angiogenesis with a conjugate of HPMA copolymer and TNP-470.
Nature Med. 2004, 10, 255–261. (b) Satchi-Fainaro, R.; Mamluk, R.;
Wang, L.; Short, S. M.; Nagy, J. A.; Feng, D.; Dvorak, A. M.; Dvorak,
H. F.; Puder, M.; Mukhopadhyay, D.; Folkman, J. Inhibition of vessel
permeability by TNP-470 and its polymer conjugate, caplostatin.
Cancer Cell 2005, 7, 251–261. (c) Benny, O.; Fainaru, O.; Adini, A.;
Cassiola, F.; Bazinet, L.; Adini, I.; Pravda, E.; Nahmias, Y.; Koirala, S.;
Corfas, G.; D'Amato, R. J.; Folkman, J. An orally delivered small-
molecule formulation with antiangiogenic and anticancer activity. Nat.
Biotechnol. 2008, 26, 799–807.
(14) (a) Han, C. K.; Ahn, S. K.; Choi, N. S.; Hong, R. K.; Moon, S. K.;
Chun, H. S.; Lee, S. J.; Kim, J. W.; Hong, C. I.; Kim, D.; Yoon, J.
H.; No, K. T. Design and synthesis of highly potent fumagillin
analogues from homology modeling for a human MetAP-2.
Bioorg. Med. Chem. Lett. 2000, 10, 39–43. (b) Jeong, B.-S.; Choi,
N. S.; Ahn, S. K.; Bae, H.; Kim, H. S.; Kim, D. Total synthesis and
antiangiogenic activity of cyclopentane analogues of fumagillol.
Bioorg. Med. Chem. Lett. 2005, 15, 3580–3583. (c) Kim, E.-J.; Shin,
W.-H. General pharmacology of CKD-732, a new anticancer agent:
effects on central nervous, cardiovascular, and respiratory system. Biol.
Pharm. Bull. 2005, 28, 217–223. (d) Lee, H. W.; Cho, C. S.; Kang, S.
K.; Yoo, Y. S.; Shin, J. S.; Ahn, S. K. Design, synthesis, and
antiangiogenic effects of a series of potent novel fumagillin analogues.
Chem. Pharm. Bull. 2007, 55, 1024–1029. (e) Rodeschini, V.; Boiteau,
J.-G.; Van de Weghe, P.; Tarnus, C.; Eustache, J. MetAP-2 inhibitors
based on the fumagillin structure. Side-chain modification and ring-
substituted analogues. J. Org. Chem. 2004, 69, 357–373. (f) Rodeschini,
V.; Van de Weghe, P.; Salomon, E.; Tarnus, C.; Eustache, J. Enantiose-
lective approaches to potential MetAP-2 reversible inhibitors. J. Org.
Chem. 2005, 70, 2409–2412. (g) Rodeschini, V.; Van de Weghe, P.;
Tarnus, C.; Eustache, J. A Simple spiroepoxide as methionine amino-
peptidase-2 inhibitor: synthetic problems and solutions. Tetrahedron
Lett. 2005, 46, 6691–6695. (h) Baldwin, J. E.; Bulger, P. G.; Marquez, R.
Fast and efficient synthesis of novel fumagillin analogues. Tetrahedron
2002, 58, 5441–5452. (i) Pyun, H.-J.; Fardis, M.; Tario, J.; Yang, C. Y.;
Ruckman, J.; Henninger, D.; Jin, H.; Kim, C. U. Investigation of novel
fumagillin analogues as angiogenesis inhibitors. Bioorg. Med. Chem.
Lett. 2004, 14, 91–94. (j) Mazitschek, R.; Huwe, A.; Giannis, A.
Synthesis and biological evaluation of novel fumagillin and ovalicin
analogues. Org. Biomol. Chem. 2005, 3, 2150–2154.
(15) Olson, G. L.; Self, C.; Lee, L.; Cook, C. M.; Birktoft, J. J. U.S.
(11) (a) Kudelka, A. P.; Levy, T.; Verschraegen, C. F.; Edwards, S. P.;
Termrungruanglert, W.; Freedman, R. S.; Kaplan, A. L.; Kieback,
D. G.; Meyers, C. A.; Jaeckle, K. A.; Loyer, E.; Steger, M.; Mante,
R.; Mavligit, G.; Killian, A.; Tang, R. A.; Gutterman, J. U.;
Kavanagh, J. J. A phase I study of TNP-470 administered to
patients with advanced squamous cell cancer of the cervix. Clin.
Cancer Res. 1997, 3, 1501–1505. (b) Figg, W. D.; Pluda, J. M.; Lush, R.
M.; Saville, M. W.; Wyvill, K.; Reed, E.; Yarchoan, R. The pharmaco-
kinetics of TNP-470, a new angiogenesis inhibitor. Pharmacotherapy
1997, 17, 91–97. (c) Dezube, B. J.; Van Roenn, J. H.; Holden-Wiltse, J.;
Cheung, T. W.; Remick, S. C.; Cooley, T. P.; Moore, J.; Sommadossi, J.-
P.; Shriver, S. L.; Suckow, C. W.; Gill, P. S. Fumagillin analog in the
treatment of Kaposi's sarcoma: a phase I AIDS clinical trial group study.
J. Clin. Oncol. 1998, 16, 1444–1449. (d) Bhargava, P.; Marshall, J. L;
Rizvi, N.; Dahut, W.; Yoe, J.; Figuera, M.; Phipps, K.; Ong, V. S.; Kato,
A.; Hawkins, M. J. A phase I and pharmacokinetic study of TNP-470
administered weekly to patients with advanced cancer. Clin. Cancer
Res. 1999, 5, 1989–1995. (e) Stadler, W. M.; Kuzel, T.; Shapiro, C.;
Patent 6,548,477, April 15, 2003.
(16) (a) Kendall, R. L.; Bradshaw, R. A. Isolation and characterization
of the methionine aminopeptidase from porcine liver responsible
for the co-translational processing of proteins. J. Biol. Chem. 1992,
267, 20667–20673. (b) Zuo, S.; Guo, O.; Ling, C.; Chang, Y.-H.
Evidence that two zinc fingers in the methionine aminopeptidase from
Saccharomyces cerevisiae are important for normal growth. Mol. Gen.
Genet. 1995, 246, 247–253.
(17) (a) Zhou, G.; Tsai, C. W.; Liu, J. O. Fumagalone, a reversible
inhibitor of type 2 methionine aminopeptidase and angiogenesis.
J. Med. Chem. 2003, 46, 3452–3454. (b) Lu, J.; Chong, C. R.; Hu, X.;
Liu, J. O. Fumarranol, a rearranged fumagillin analogue that inhibits
angiogenesis in vivo. J. Med. Chem. 2006, 49, 5645–5648.
(18) Bernier, S. G.; Lazarus, D. D; Clark, E.; Doyle, B.; Labenski, M.
T.; Thompson, C. D.; Westlin, W. F.; Hannig, G. A methionine
aminopeptidase-2 inhibitor, PPI-2458, for the treatment of rheu-
matoid arthritis. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 10768–
10773.

8056 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 24
Arico-Muendel et al.
(19) (a) Cooper, A. C.; Karp, R. M.; Clark, E. J.; Taghizadeh, N. R.;
Hoyt, J. G.; Labenski, M. T.; Murray, M. J.; Hannig, G.; Westlin,
W. F.; Thompson, C. D. A novel methionine aminopeptidase-
2 inhibitor, PPI-2458, inhibits Non-Hodgkin’s lymphoma cell pro-
liferation in vitro and in vivo. Clin. Cancer Res. 2006, 12, 2583–2590.
(b) Hannig, G.; Lazarus, D. D.; Bernier, S. G.; Karp, R. M.; Lorusso, J.;
Qiu, D.; Labenski, M. T.; Wakefield, J. D.; Thompson, C. D.; Westlin, W.
F. Inhibition of melanoma tumor growth by a pharmacological inhibitor
of MetAP-2, PPI-2458. Int. J. Oncol. 2006, 28, 955–963.
(20) (a) Brahn, E.; Banquerigo, M. L.; Schoettler, N. R.; Lee, S. A new
antiangiogenesis inhibitor, PPI-2458: regression of collagen-in-
duced arthritis and prevention of erosions. Arthritis Rheum. 2001,
44, S271. (b) Brahn, E.; Schoettler, N. R.; Banquerigo, M. L.; Thomp-
son, C. D. Involution of collagen-induced arthritis with angiogenesis
inhibitor PPI-2458. Arthritis Rheum. 2003, 48, S347. (c) Hannig, G.;
Bernier, S. G.; Hoyt, J. G.; Doyle, B.; Clark, E.; Karp, R. M.; Lorusso,
J.; Westlin, W. F. Suppression of inflammation and structural damage in
experimental arthritis through molecular targeted therapy with PPI-
2458. Arthritis Rheum. 2007, 56, 850–860. (d) Lazarus, D. D.; Doyle,
E. G.; Bernier, S. G.; Rogers, A. B.; Labenski, M. T.; Wakefield, J. D.;
Karp, R. M.; Clark, E. J.; Lorusso, J.; Hoyt, J. G.; Thompson, C. D.;
Hannig, G.; Westlin, W. F. An inhibitor of methionine aminopeptidase
type-2, PPI-2458, ameliorates the pathophysiological disease processes
of rheumatoid arthritis. Inflamm. Res. 2008, 57, 18–27. (e) Bainbridge,
J.; Madden, L.; Essex, D.; Binks, M.; Malhotra, R.; Paleolog, E. M.
Methionine aminopeptidase-2 blockade reduces chronic collagen-in-
duced arthritis: potential role for angiogenesis inhibition. Arthritis Res.
Ther. 2007, 9, R127.