Synthesis and antifolate evaluation of the aminopterin analogue with a bicyclo[2.2.2]octane ring in place of the benzene ring

Article abstract of DOI:10.1016/S0223-5234(01)01224-7

N-[4-[[2,4-diamino-6-pteridinyl)methyl]amino]bicyclo[2.2.2]octane-1- carbonyl]-L-glutamic acid (1) was synthesized and tested for antifolate activity. N-(4-Aminobicyclo[2.2.2]octane-1-carbonyl-L-glutamic acid dimethyl ester (6), the side chain precursor to subject compound 1, was synthesized readily via reported bicyclo[2.2.2]octane-1,4-dicarboxylic acid monoethyl ester (2). The side chain precursor 6 was alkylated by 6-(bromomethyl)-2,4-pteridinediamine (7). Subsequent ester hydrolysis then afforded 1. Antifolate and antitumor evaluation of 1 verses L1210 dihydrofolate reductase (DHFR) and three tumor cell lines (L1210, S180, and HL60) showed it to be ineffective. Although compound 1 was very similar to aminopterin structurally, the bicyclo[2.2.2]octane ring system in place of the phenyl ring in the p-aminobenzoate moiety effectively negates the stoichiometric binding displayed by many classical DHFR inhibitors bearing appropriate aromatic ring systems in the side chain.

Full text of DOI:10.1016/S0223-5234(01)01224-7

Eur. J. Med. Chem. 36 (2001) 237–242
´
© 2001 Editions scientifiques et me´dicales Elsevier SAS. All rights reserved
PII: S0223-5234(01)01224-7/SCO
Short communication
Synthesis and antifolate evaluation of the aminopterin analogue with a
bicyclo[2.2.2]octane ring in place of the benzene ring
Robert C. Reynoldsa*, Cheryl A. Johnsona, James R. Pipera, Frances M. Sirotnakb
aOrganic Chemistry Department, Southern Research Institute, Birmingham, AL 35255, USA
bLaboratory for Molecular Therapeutics, Memorial Sloan-Kettering Cancer Center, New York, NY 10021, USA
Received 25 April 2000; revised 29 January 2001; accepted 8 February 2001
Abstract – N-[4-[[2,4-diamino-6-pteridinyl)methyl]amino]bicyclo[2.2.2]octane-1-carbonyl]-
L-glutamic acid (1) was synthesized and
tested for antifolate activity. N-(4-Aminobicyclo[2.2.2]octane-1-carbonyl- -glutamic acid dimethyl ester (6), the side chain precursor
L
to subject compound 1, was synthesized readily via reported bicyclo[2.2.2]octane-1,4-dicarboxylic acid monoethyl ester (2). The side
chain precursor 6 was alkylated by 6-(bromomethyl)-2,4-pteridinediamine (7). Subsequent ester hydrolysis then afforded 1.
Antifolate and antitumor evaluation of 1 verses L1210 dihydrofolate reductase (DHFR) and three tumor cell lines (L1210, S180, and
HL60) showed it to be ineffective. Although compound 1 was very similar to aminopterin structurally, the bicyclo[2.2.2]octane ring
system in place of the phenyl ring in the p-aminobenzoate moiety effectively negates the stoichiometric binding displayed by many
´
classical DHFR inhibitors bearing appropriate aromatic ring systems in the side chain. © 2001 Editions scientifiques et me´dicales
Elsevier SAS
dihydrofolate reductase / Curtius rearrangement / N-(4-aminobicyclo[2.2.2]octane-1-carbonyl)-
momethyl-2,4-pteridinediamine / antifolate evaluation
L-glutamic acid dimethyl ester / 6-bro-
1. Introduction
MTX in antitumor activity, but none has yet proven
to have therapeutic advantages sufficient for it to
displace or join MTX in clinical usage. Agents of such
activity include AMT/MTX analogues in which the
4-aminobenzoyl group is replaced by the 4-amino-1-
naphthoyl group [2]. Despite nearly 50 years of activ-
ity in the area of antifolate chemotherapy of cancer,
new antitumor antifolates continue to be pursued
[3–14].
Synthesis and testing of the naphthoyl analogues
followed molecular modeling studies that revealed
ample spatial accommodation for the naphthalene
and even larger groups in models based on reported
X-ray crystallographic data describing the multifacted
binding of MTX to human DHFR [2]. These studies
showed the naphthalene ring fitting into the hydro-
phobic pocket while not interfering with the other
critical hydrogen bonds and electrostatic interactions
associated with the extremely tight binding of the
parent antifolates to DHFR. These naphthoyl ana-
logues did in fact show high levels of both DHFR
The potent DHFR inhibitor aminopterin (AMT)
was introduced into the clinic as an antitumor agent
over 50 years ago. It was, however, soon displaced by
its N10-methyl analogue methotrexate (MTX), which
proved to be therapeutically superior. MTX also
came to be used against diseases other than cancer
such as psoriasis and rheumatoid arthritis [1]. During
the five decades since this class of antitumor agent
was introduced medicinal chemists have extensively
investigated structural modifications with the aim of
beneficially altering the therapeutic spectrum [1] of
MTX in terms of selective uptake and activity versus
tumor tissue, activity against a wider panel of tumor
types, and direct clinical efficacy. In fact, numerous
DHFR-inhibiting analogues are at least as potent as
* Correspondence and reprints
E-mail address: reynolds@sri.org (R.C. Reynolds).

238
R.C. Reynolds et al. / European Journal of Medicinal Chemistry 36 (2001) 237–242
described by Shioiri et al. [17]. First, 2 was treated
with DPPA to form its carbonyl azide which under-
went the Curtius rearrangement in boiling toluene
containing benzyl alcohol to give the benzyl carba-
mate 3 in a surprising overall yield (for three steps) of
96%. After mild ester hydrolysis of 3 gave carboxylic
acid 4, DPPA was used again in peptide-type coupling
of 4 with dimethyl L-glutamate to give 5. Hy-
drogenolysis of 5 removed the amine-protecting ben-
zyloxycarbonyl group to give the sidechain 7.
Alkylation of 6 with 6-(bromomethyl)-2,4-pteridinedi-
amine (7) [18] gave the diester 8, which was purified
by column chromatography. Mild ester hydrolysis
gave the target compound 1.
inhibition and antitumor activity that compared fa-
vorably with those of MTX in inhibiting tumor cell
growth (L1210, HL60, S180) in culture and toward
the E0771 mammary adenocarcinoma in mice [2].
While docking various antifolate synthetic candi-
dates (such as the naphthalene analogues) into the
enzyme model [2], we became interested in a structure
in which the benzene ring of AMT is replaced by a
cycloaliphatic group. For example, a 1,4-substituted
bicyclo[2.2.2]octane system is quite similar to the
phenyl ring of the PABA moiety in a classical antifo-
late in terms of orientation of the substituents and
overall length. This ring system is more sterically
demanding, but could potentially occupy the appro-
priate hydrophobic pocket of the enzyme even more
efficiently than a phenyl substituent. Certainly, a bicy-
clo[2.2.2]octane system is more rigid than a simple
1,4-substituted cyclohexane which introduces confor-
mational flexibility into the enzyme binding parame-
ters; the cyclohexane analogue of MTX is
significantly less active than MTX in terms of both
enzyme inhibition and cytotoxicity [20]. Hence, we
prepared the bicyclo[2.2.2]octane analog of the classi-
cal antifolate AMT in order to test whether such an
analogue could bind and inhibit DHFR.
3. Pharmacology
Antifolate evaluation [19] of the bicyclo[2.2.2]-
octane analogue 1 revealed that it was not inhibitory
toward DHFR (L1210) relative to the control drug
aminopterin (AMT) nor did it appreciably inhibit cell
growth in culture of three standard tumor cells lines
(L1210, S180, and HL60; see table I).
4. Results and discussion
2. Chemistry
These results show that the rigid cycloaliphatic
system, bicyclo[2.2.2]octane, cannot replace the pla-
nar aromatic ring to produce an effective DHFR
inhibitor of the AMT/MTX class. Although com-
pound 2 appeared to meet spatial requirements, the
bicyclo[2.2.2]octane group was no more tolerated
than the modification in the only related example that
we know of, specifically the MTX analogue bearing a
simple cyclohexane ring in place of the 1,4-phenylene
group [20]. That modification was 740-fold less effec-
tive than MTX (IC50 value of 20 mM vs. 0.027 mM for
MTX) in the inhibition of DHFR (pigeon liver), and
it displayed no significant activity against KB cells in
culture or L1210 leukemia in mice.
The accessibility of the 4-aminobicyclo[2.2.2]-
octane-1-carboxylic acid system allowed synthesis of 1
for the testing of this notion. The key precursor, ethyl
hydrogen bicyclo[2.2.2]octane-1,4-dicarboxylate (2),
was prepared by a reported route [16]. Diphenylphos-
phoryl azide (DPPA) was used in two ways as first
Table I. Comparison of properties of aminopterin (1) with
bicyclo[2.2.2]octane analogue (2)
Compound DHFR
inhibition a
Cell growth inhibition IC50 (nM) a
5. Conclusion
Ki (pM)
S180
L1210
HL60
Substitution of the bicyclo[2.2.2]octane ring in
place of the benzene ring in AMT analogue 1 renders
the structure incapable of the tight binding to DHFR
displayed by the parent AMT and analogues bearing
AMT
1
3.8590.3 1.0190.3 0.7290.1 1.0090.20
\1000 14409220 850994 5109380
a Methods described in [19].

R.C. Reynolds et al. / European Journal of Medicinal Chemistry 36 (2001) 237–242
239
Figure 1. Synthetic route to compound 2, the bicyclo[2.2.2]octane analogue of aminopterin.
appropriate aromatic ring systems in the sidechain [4,
5]. The electronic features of an appropriate aromatic
ring may be required for a potential inhibitor of the
AMT class to be fixed in position for the several other
critical binding sites of the structure to form the
hydrophobic, hydrogen, and salt bridge bonds that
collectively produce effective inhibition of the enzyme
[15].
the combination MeCN+0.1 M NaOAc (pH 3.6) chang-
ing from 15% MeCN to 50%. Purifications by prepara-
tive TLC were done on Analtech silica gel G(F) plates (2
mm). Column chromatographic purifications were done
with silica gel (Merck, 60 A, 230–400 mesh for flash
chromatography). Evaporations were performed with a
rotary evaporator, higher boiling solvents (DMF, N,N-
dimethylacetamide or DMAc, DMSO) were removed in
vacuo (<1 mm, bath to 35°C), and more volatile sol-
vents with a water aspirator. Products were dried in
vacuo (<1 mm) at 22–25°C over P2O5 and NaOH
pellets. The final product was dried, and then allowed to
equilibrate with ambient conditions of the laboratory.
Analytical results indicated by element symbols were
within 0.4% of the theoretical values. Spectral determi-
nations and elemental analyses were performed in the
Molecular Spectroscopy Section of Southern Research
Institute under the direction of Dr J.M. Riordan. The
1H-NMR spectra were determined in DMSO-d6 with a
Nicolet NMC 300 NB spectrometer using Me4Si as
internal reference. Chemical shifts (l) listed for multi-
plets were measured from the approximate centers, and
relative integrals of peak areas agreed with those ex-
pected for the assigned structures. Mass spectra were
recorded on a Varian MAT 311A mass spectrometer in
the fast-atom-bombardment (FAB) mode.
6. Experimental
6.1. Methods
Examinations by TLC were performed on Analtech
precoated (250 mm) silica gel G(F) plates. High-perfor-
mance liquid chromatography (HPLC) assays were
made with a Waters Associates ALC-242 liquid chro-
matograph equipped with an ultraviolet detector (254
nm) and a M-6000 pump using a 30×0.29 cm C18
mBondapak column. Purity assays were done in the
reversed-phase isocratic mode with a mobile phase con-
sisting of MeCN (10 or 15% by volume) in 0.1 M
NaOAc (pH 3.6). Saponification of 9 to give 2 was
monitored using a 20-min linear gradient system with

240
R.C. Reynolds et al. / European Journal of Medicinal Chemistry 36 (2001) 237–242
6.1.1. Synthesis of sidehain precursor 6 (figure 1, part
1)
pirator) removed residual Et2O. The aqueous solution
was acidified with stirring to pH 3.5 by dropwise addi-
tion of 12 N HCl to cause precipitation of 5 as a white
1
solid; yield 85% (3.56 g), m.p. 189–191°C. H-NMR:
6.1.1.1. Ethyl hydrogen
bicyclo[2.2.2]octane-1,4-dicarboxylate (2)
l=1.74 (s, 12H, CH2 of ring), 4.96 (s 2H, CH2C6H5),
7.00 (s, 1H, NH), 7.3–7.4 (m, 5H, C6H5), 12.04 (s, 1H,
CO2H). Anal. (C17H21NO4) C, H, N.
This compound was prepared essentially as described
by Roberts et al. [16]. Our sample had a m.p. 146–
150°C (lit. [16] m.p. 149.0–150.5°C) and was homoge-
neous by TLC (CHCl3–MeOH–AcOH, 95:5:0.5;
detection by I2 vapor). 1H-NMR: l=1.16 (t, 3H,
CH2CH3), 1.70 (s, 12H, cycloaliphatic CH2), 4.03 (q,
2H, –CH2CH3).
6.1.1.4. N-[4-[[(Benzyloxy)carbonyl]amino]bicyclo[2.2.2]-
octane-1-carbonyl]-
A stirred solution of 4 (2.00 g, 6.59 mmol) and
dimethyl -glutamate·HCl (1.68 g, 7.94 mmol) in DMF
L
-glutamic acid dimethyl ester (5)
L
(50 mL) was chilled to −5–0°C (bath temperature) and
treated with a solution of DPPA (1.7 mL, 2.17 g, 7.89
mmol) in DMF (5 mL) followed by Et3N (1.90 mL, 1.38
g, 13.7 mmol). The reaction mixture was kept at −5–
0°C for 2 h before it was allowed to warm to room
temperature and left for 18 h. DMF was removed in
vacuo (<1 mm, bath to 30°C). The residual mixture of
solid and oil was stirred with CHCl3 (150 mL) and H2O
(100 mL). The H2O layer was extracted again with
CHCl3 (150 mL). After the CHCl3 layers were com-
bined, the solution was washed successively with 5%
NaHCO3, H2O, 0.1 N HCl, and again with H2O before
it was dried (Na2SO4) and evaporated to give crude 5 as
a clear oil. The oil was dissolved in cyclohexane–EtOAc
(1:1), and the solution was applied to a silica gel
column. Elution with the same solvent afforded frac-
tions homogeneous by TLC (Rf 0.3, detection by I2
vapor). The combined and evaporated homogeneous
fractions afforded 5 as clear oil (2.00 g, 67% yield).
Mass: m/z=461 (MH+). 1H-NMR: l=1.74 (s, 12H,
cycloaliphatic ring CH2), 1.85, 1.98 (2m, 2H,
CHCH2CH2, nonequivalent), 2.33 (t, 2H, CHCH2CH2),
3.58, 3.59 (2s, 6H, –CO2CH3), 4.20 (m, 1H,
NHCHCH2), 4.93 (s, 2H, CH2C6H5), 6.99 (s, 1H, –
OCONH), 7.34 (m, 5H, C6H5), 7.64 (d, 1H, CONHCH).
In a second run, same scale and procedure as above, the
product crystallized after evaporation following purifica-
tion by chromatography; yield 2.18 g (72%), m.p. 95–
98°C. Mass: m/z=461 (MH+). A portion of the oily
product from the first run crystallized rapidly when
seeded and had identical m.p. with the product from the
second run.
6.1.1.2. Ethyl 4-[[(benzyloxy)carbonyl]amino]-
bicyclo[2.2.2]octane-1-carboxylate (3) from 2 in steps
a–c (figure 1, part 1)
Step a: a solution of 30.0 mmol each of 2 (6.78 g),
DPPA (8.28 g), and Et3N (3.03 g) in dry toluene (120
mL) was stirred at room temperature for 1 h.
Step b: the solution of the resulting azide was then
slowly heated over 45 min to boiling.
Step c: the solution, now containing the isocyanate,
was cooled slightly and treated with an excess of benzyl
alcohol (30 mL, 0.29 mol). The resulting solution was
then refluxed 18 h, cooled, and washed successively with
solutions of 5% citric acid, 5% NaHCO3, and saturated
NaCl. The dried (Na2SO4) and filtered organic phase
was then evaporated in vacuo (finally at <1 mm, both at
80°C) to constant weight to leave 3 as a nearly colorless
oil (9.53 g, 96% overall yield). 1H-NMR: l=1.15 (t, 3H,
CH3), 1.75 (s, 12H, CH2 of ring), 4.00 (q, 2H, CH2CH3),
4.94 (s, 2H, CH2C6H5), 7.02 (s, 1H, NH), 7.3–7.4 (m,
5H, C6H5).
6.1.1.3. 4-[[(Benzyloxy)carbonyl]amino]bicyclo[2.2.2]-
octane-1-carboxylic acid (4)
A stirred solution of 3 (4.56 g, 13.8 mmol) in EtOH
(100 mL) was treated with 1 N NaOH (70 mL). The
resulting solution was kept at room temperature for 20
h before being warmed at 55–65°C (bath temp) for 1 h.
EtOH was then removed by evaporation (H2O aspira-
tor) until the concentrate became cloudy, indicating
incomplete saponification. EtOH (about 30 mL) was
added to clarify, and the solution was kept at room
temperature about 3 h longer. The solution was again
evaporated until nearly free of EtOH; addition of H2O
(250 mL) produced only faint turbidity indicating virtu-
ally complete saponification. Extraction with Et2O
clarified the aqueous phase and evaporation (H2O as-
6.1.1.5. N-(4-Aminobicyclo[2.2.2]octane-1-carbonyl)-
L
-glutamic acid dimethyl ester (6)
Hydrogenolysis of 5 (945 mg, 2.05 mmol) was carried
out in stirred MeOH (20 mL) containing 10% Pd on C

R.C. Reynolds et al. / European Journal of Medicinal Chemistry 36 (2001) 237–242
241
(50 mg) at ambient conditions. Uptake of H2 (observed
by replacement with H2O in a gas buret) had stopped
after 6 h. The filtered solution was evaporated (final
conditions 1 mm, bath 25°C) to give 7 as a clear oil.
Further drying (1 mm over P2O5 at 23–25°C) brought
the oil to constant weight (670 mg, 100% yield), but it
proved to be hygroscopic under ambient conditions.
2H, CH2CH2CO), 3.58, 3.59 (2s, 6H, both CO2CH3),
3.81 (s, 2H, CH2N), 4.22 (q, 1H, NHCHCO), 6.52 (br s,
2H, NH2), 7.65 (d, 1H, CONH) overlapping with 7.0 (s,
2H, NH2), 8.71 (s, 1H C7–H).
6.1.2.2. N-[4-[[2,4-Diamino-6-pteridinyl)methyl]amino]-
bicyclo[2.2.2]octane-1-carbonyl]- -glutamic acid (1)
L
1
MS: m/z=327 (MH+). H-NMR: l=1.44, 1.70 (2m,
A stirred suspension of the ester 8 (90 mg, 0.18 mmol)
in H2O (1 mL) was treated with 1N NaOH (0.40 mL, 2.2
molar equiv.). Solution occurred within 20 min. After 20
h at 20–23°C, HPLC analysis showed complete conver-
sion to a single product. The solution was cooled in an
ice-H2O bath and carefully acidified to pH 4.4 using 1 N
HCl (0.4 mL) to cause 1 to precipitate as a pale yellow
solid. The mixture was kept in a refrigerator overnight
before the solid was collected and washed sparingly with
cold H2O; yield 56 mg (56%). MS: m/z=473 (MH+).
1H-NMR: l=1.6–2.0 (overlapping m, 14H; 1.70, 1.78
6H each, CH2 due to bicycloaliphatic ring), 1.85, 1.98
(2m, 2H, CHCH2CH2, nonequivalent), 2.34 (t, 2H,
CHCH2CH2), 3.58, 3.59 (2s, 6H, –CO2CH3), 4.20 (m,
1H, NHCHCH2), 7.62 (d, 1H, CONHCH). Anal.
(C16H26N2O5·H2O) C, H, N.
6.1.2. Alkylation of 6 with 7
6.1.2.1. N-[4-[[(2,4-Diamino-6-pteridinyl)methyl]amino]-
bicyclo[2.2.2]octane-1-carbonyl]-L-glutamic acid dimethyl
ester (8)
cycloaliphatic CH2 over most of
m
due to
A solution of 6 (670 mg, 2.05 mmol) and 6-(bro-
momethyl)-2,4-pteridinediamine hydrobromide (7, 251
mg of 87% purity1, 0.650 mmol) in DMAc (10 mL) was
stirred at room temperature under N2 in a stoppered
flask wrapped in Al foil. After 48 h, silica gel (1.4 g of
60–200 mesh) was added, and the mixture was then
evaporated in vacuo (<1 mm, bath to 35°C) to a dry
yellow dispersion. The dispersion was pulverized and
dried further in vacuo (<1 mm over P2O5) before it was
placed atop a column (200 cm of 3 cm diameter) of
silica gel (230–400 mesh) that had been poured from
CHCl3–MeOH (9:1) containing concentrated NH4OH
(0.5% by volume). Elution by the same solvent system
afforded fractions homogeneous by TLC (Rf ꢀ0.5, 1:1
CHCl3–MeOH). Pure 8 obtained as a yellow solid on
evaporation of the pooled fractions weighed 100 mg
(20% yield). Because the yield was less than expected,
the column was extruded and divided into approximate
thirds. The middle and lower thirds were then extracted
by stirring with elution medium. Evaporation of the
filtered extract from the middle third gave 80 mg of 8
homogeneous by TLC and the lower third gave 25 mg
of essentially homogeneous 8. The total yield was 205
CH2CH2CO2H), 2.22 (m, 2H, CH2CH2CO2H), 3.9–4.2
(overlapping m, CH2NH and CHCH2), 6.64 (s, 2H,
NH2), 7.36 (d, 1H, CONH), 7.68, 7.80 (2 bs, 2H, NH2),
8.73 (s, 1H, C7-H). HPLC: single peak in gradient and
isocratic mode; respective retention times 2.28 and 5.80
min. Anal (C21H28N8O5·4.8H2O) C, H, N.
Acknowledgements
This work was supported by Public Health Service
grant no. CA25236 from the National Cancer Institute,
Department of Health and Human Services.
References
[1] Rosowsky A., in: Ellis G.P., West G.B. (Eds.), Progress in
Medicinal Chemistry, vol. 26, Elsevier, Amsterdam, 1989, pp.
1–252.
[2] Piper J.R., Johnson C.A., Maddry J.A., Malik N.D., McGuire
J.J., Otter G.M., Sirotnak F.M., J. Med. Chem. 36 (1993)
4161–4171.
[3] Stromhaug A., Warren D.J., Cancer Chemother. Pharmacol. 45
(6) (2000) 450–456.
[4] Adjei A.A., Ann. Oncol. 11 (10) (2000) 1335–1341.
1
mg (41%). MS: m/z=501 (MH+ for C23H32N8O5). H-
NMR: l=1.55, 1.74 (2m, 6H each, cycloaliphatic CH2),
1.85, 1.98 (2m, 2H, CHCH2CH2, nonequivalent), 2.34 (t,
[5] Gangjee A., Yu J., McGuire J.J., Cody V., Galitsky N., Kisliuk
R.L., Queener S.F., J. Med. Chem. 43 (21) (2000) 3837–3851.
[6] Wright J.E., Vaidya C.M., chen Y., Rosowsky A., Biochem.
1 Compound 8, prepared as described in [9], was estimated
from its integrated 1H-NMR spectrum to be of 87% purity;
other matter present was the 6-methyl analogue (2%) and
2-PrOH (11%).
Pharmacol. 60 (1) (2000) 41–46.
[7] Pignatello R., Spampinato G., Sorrenti V., Di Giacomo C.,
Vicari L., McGuire J.J., Russell C.A., Puglisi G., Toth I., Eur.
J. Pharm. Sci. 10 (3) (2000) 237–245.

242
R.C. Reynolds et al. / European Journal of Medicinal Chemistry 36 (2001) 237–242
[8] O’Dwyer P.J., Nelson K., Thornton D.E., Semin. Oncol. 26
[15] Freisheim J.H., Mathews D.A., in: Sirotnak F.M., Burchall
J.J., Ensminger W.B., Momtgomery J.A. (Eds.), Folate Antag-
onists as Therapeutic Agents, vol. 1, Academic, Amsterdam,
1984, pp. 69–131.
(2:6) (1999) 99–104.
[9] Faessel H.M., Slocum H.K., Rustum Y.M., Greco W.R.,
Biochem. Pharmacol. 57 (5) (1999) 567–577.
[10] Rosowsky A., Wright J.E., Vaidya C.M., Bader H., Forsch
R.A., Mota C.E., Pardo J., Chen C.S., Chen Y.N., J. Med.
Chem. 41 (26) (1998) 5310–5319.
[16] Roberts J.D., Moreland W.T. Jr, Frazer W., J. Am. Chem. Soc.
75 (1953) 637–640.
[17] (a) T. Shioiri, K. Ninomiya, S. Yamada, J. Am. Chem. Soc. 94
(1972) 6203–5. (b) K. Ninomiya, T. Shiori, S. Yamada, Tetra-
hedron 30 (1977) 2151–7.
[11] Corona P., Vitale G., Loriga M., Paglietti G., Costi M.P.,
Farmaco 53 (7) (1998) 480–493.
[12] Li W.W., Tong W.P., Bertino J.R., Clin. Cancer Res. 1 (6)
(1995) 631–636.
[13] Mauritz R., Bekkenk M.W., Rots M.G., Pieters R., Mini E.,
van Zantwijk C.H., Veerman A.J., Peters G.J., Jansen G., Clin.
Cancer Res. 4 (10) (1998) 2399–2410.
[18] Piper J.R., Montgomery J.A., J. Org. Chem. 42 (1977) 208–
211.
[19] Sirotnak F.M., DeGraw J.I., Schmid F.A., Goutas L.J., Moc-
cio D.M., Cancer Chemother. Pharmacol. 12 (1984) 26–30.
[20] Montgomery J.A., Piper J.R., Elliott R.D., Temple C., Roberts
[14] Lee H.K., Chui W.K., Bioorg. Med. Chem. 7 (6) (1999) 1255–
1262.
E.C., Shealy Y.F., J. Med. Chem. 22 (1979) 862–868.
.