Structure-activity studies of antitumor agent irofulven (Hydroxymethylacylfulvene) and analogues

Article abstract of DOI:10.1021/jo010458z

Many analogues of the antitumor agent irofulven have been readily prepared by replacing the allylic hydroxyl with a variety of nucleophiles. Analogues of acylfulvene (the precursor to irofulven) were also prepared by Michael reaction with acrolein. The toxicity of the analogues was determined, as well as preclinical antitumor activity. Several analogues exhibited good activity in mouse xenografts. Structural requirements for activity are discussed.

Full text of DOI:10.1021/jo010458z

6158
J . Org. Chem. 2001, 66, 6158-6163
Str u ctu r e-Activity Stu d ies of An titu m or Agen t Ir ofu lven
(Hyd r oxym eth yla cylfu lven e) a n d An a logu es
Trevor C. McMorris,*,† J ian Yu,† Ricardo Lira,† Robin Dawe,‡ J ohn R. MacDonald,‡
Stephen J . Waters,‡ Leita A. Estes,§ and Michael J . Kelner§
Department of Chemistry and Biochemistry, University of California, San Diego,
La J olla, California 92093-0506, MGI Pharma Inc., Bloomington, Minnesota 55438, and Department of
Pathology, University of California, San Diego, Medical Center, San Diego, California 92103-8320
tmcmorris@ucsd.edu
Received May 4, 2001
Many analogues of the antitumor agent irofulven have been readily prepared by replacing the
allylic hydroxyl with a variety of nucleophiles. Analogues of acylfulvene (the precursor to irofulven)
were also prepared by Michael reaction with acrolein. The toxicity of the analogues was determined,
as well as preclinical antitumor activity. Several analogues exhibited good activity in mouse
xenografts. Structural requirements for activity are discussed.
In tr od u ction
Irofulven rapidly enters tumor cells, where it binds to
cellular macromolecules and inhibits DNA synthesis.9-11
Treatment of the human CEM leukemic cell line with
irofulven results in cell-cycle arrest in the S phase and
inhibition of DNA, RNA, and protein synthesis with 50%
inhibitory concentration values of 2, 20, and 70 µmol/L,
respectively.9 Radiolabeled irofulven is localized predomi-
nantly in the nuclear compartment followed by the
cytosol and membrane compartments: more than 60%,
27%, and 11% of drug binds to proteins, DNA, and RNA,
respectively.10 Irofulven induces cytotoxicity principally
by generating DNA strand breakage, as exposure of CEM
cells to the drug does not result in the formation of DNA
intrastrand cross-links or DNA-protein cross-links.10
The most unique aspect of irofulven’s antitumor activ-
ity seems to be its ability to act as a selective inducer of
apoptosis in human tumor cell lines, and in contrast to
conventional antitumor agents, irofulven retains this
activity against tumor cell lines regardless of their p53
or p21 expression.12,13 The illudins and acylfulvenes also
differ from other agents with alkylating activity in their
preferential cytotoxicity toward cell lines deficient in the
excision repair cross-complementing (ERCC) DNA repair
enzymes ERCC2 and ERCC3, which indicates that the
repair of irofulven-induced DNA damage requires a full
The toxic sesquiterpenes illudin S and illudin M (1 and
2) were discovered many years ago at the New York
Botanical Garden during a search for antibiotic sub-
stances in the basidiomycete, Omphalotus illudens
(Clitocybe illudens, Omphalotus olearius).1,2 The com-
pounds are believed to be responsible for poisoning that
occurs when Omphalotus is mistaken for an edible
mushroom.3
Illudins and certain derivatives have been evaluated
for antitumor activity in the National Cancer Institute
Developmental Therapeutics Program. Illudin M signifi-
cantly increased the life span of rats with Dunning
leukemia, but had a low therapeutic index in solid tumor
systems.4 In contrast, we have found derivatives of
illudins, including hydroxymethylacylfulvene (5), now
called irofulven, which produced complete tumor
regression in a variety of xenograft models.5,6 Compound
5, irofulven, in a phase II human clinical trial, demon-
strated efficacy against pancreatic carcinoma,
a
malignancy that is resistant to all other forms of chemo-
therapy. On the basis of these outstanding results, a
large-scale phase III trial was initiated in February
2001.8 In J une 2001 the Food and Drug Administration
granted “fast track” designation for the use of Irofulven
in patients with gemcitabine-refractory pancreatic can-
cer.
(8) Von Hoff, D. D.; Cox, J ohn, V.; Eder, J . P.; Tempero, M. A.;
Eckhardt, G.; Rowinsky, E. K.; Smith, S. L.; Stuart, K. E.; Proper, J .;
Campbell, E. R.; MacDonald, J . R. Activity of Irofulven (MGI 114) in
Patients With Advanced Pancreatic Cancer Refractory to Gemcitabine.
In Proceedings AACR-NCI-EORTC International Conference; Nov
7-10, 2000, Amsterdam; American Association for Cancer Research,
Inc.: Philadelphia, 2000; Abstract No. 365.
† University of California, San Diego.
‡ MGI Pharma Inc.
§ University of California, San Diego, Medical Center.
(1) Anchel, M.; Hervey, A.; Robbins, W. J . Proc. Natl. Acad. Sci.
U.S.A. 1950, 36, 300.
(2) McMorris, T. C.; Anchel, M. J . Am. Chem. Soc. 1963, 85, 831.
(3) Ammirati, J . F.; Traquair, J . A.; Horgen, P. A. Poisonous
Mushrooms of the Northern United States and Canada; University of
Minnesota Press: Minneapolis, 1985; p 290.
(4) Kelner, M. J .; McMorris, T. C.; Taetle, R. Preclinical Evaluation
of Illudins as Anticancer Agents: Basis for Selective Cytotoxicity. J .
Natl. Cancer Instit. 1990, 82, 1562.
(5) McMorris, T. C.; Kelner, M. J .; Wang, W.; Yu, J .; Estes, L. A.;
Taetle, R. J . Nat. Prod. 1996, 59, 896.
(6) Kelner, M. J .; McMorris, T. C.; Estes, L.; Wang, W.; Samson, K.
M.; Taetle, R. Invest. New Drugs 1996, 14, 161.
(7) Murgo, A.; Cannon, D. J .; Blatner, G.; Cheson, B. D. Oncology
1999, 13, 233.
(9) Kelner, M. J .; McMorris, T. C.; Montoya, M. A.; Estes, L.; Uglik,
S. F.; Rutherford, M.; Samson, K. M.; Bagnell, R. D.; Taetle, R. Cancer
Chemother Pharmacol. 1999, 44, 235.
(10) Herzig, M. C.; Arnett, B.; MacDonald, J . R.; Woynarowski, J .
M. Biochem. Pharmacol. 1999, 58, 217.
(11) Woynarowski, J . M.; Napier, C.; Koester, S.; Chen, S. F.; Troyer,
D.; Chapman, W.; MacDonald, J . R. Biochem. Pharmacol. 1997, 54,
1181.
(12) Woynarowska, B. A.; Woynarowski, J . M.; Herzig, M. C. S.;
Roberts, K.; Higdon, A. L.; MacDonald, J . R. Biochem. Pharmacol.
2000, 59, 1217.
(13) Izbicka, E.; Davidson, K.; Von Hoff, D. D. Anticancer Res. 1999,
19, 1299.
10.1021/jo010458z CCC: $20.00 © 2001 American Chemical Society
Published on Web 08/16/2001

Structure-Activity Studies of Anitumor Agent Irofulven
J . Org. Chem., Vol. 66, No. 18, 2001 6159
Acylfulvene reacts with formaldehyde in dilute sulfuric
acid giving irofulven (5) (NSC 683863). The efficacy of
this compound as an antitumor agent far surpasses that
of acylfulvene.6 The hydroxymethyl group in 5 results in
increased toxicity, the compound being five times more
toxic than 4. Improved efficacy (and higher toxicity) can
be attributed to the reactivity of the allylic primary
hydroxyl. Increased hydrophilic character may also be a
factor. The very facile displacement of the hydroxyl
allowed us to prepare a large number of analogues.
Reactions with various nucleophiles were conveniently
carried out by dissolving the reactants in acetone with
dilute H2SO4 (1 M) at room temperature.
F igu r e 1.
complement of DNA repair mechanisms that may be
absent in sensitive cell lines.14,15
Investigation of reactions of illudins indicate that the
toxicity of these compounds is due to their behavior as
alkylating agents.16 At room temperature, the compounds
react spontaneously with thiol nucleophiles, such as
cysteine or glutathione (GSH), at an optimum pH of
about 6. Toxicity to myeloid leukemia cells (HL60) can
be modulated by altering glutathione levels in the cells.17
The reaction of illudin S with GSH is illustrated in Figure
1. Michael-type addition to the R,â-unsaturated ketone
gives a cyclohexadiene intermediate that is rapidly
converted to a stable aromatic product. Concurrent ring
opening of the cyclopropane results in alkylation of
nucleophiles such as water, DNA and protein.18 Enzy-
matic reduction of illudin S leads to a similar reactive
intermediate and to a stable aromatic product. NADPH
is the coenzyme involved in this bioreductive alkylation
(Figure 1).19
Syn th esis of An a logu es In volvin g Rep la cem en t
of th e P r im a r y Hyd r oxyl of Ir ofu lven . Many ana-
logues of illudin S and M have been investigated with
the aim of finding compounds more selective in their
toxicity toward tumor cells than normal cells. From the
outset, it was evident that key features, viz. R,â-unsatur-
ated ketone and cyclopropylmethyl carbinol, are ncessary
for toxicity and antitumor activity.18 However, certain
modifications could be made that resulted in a greatly
improved therapeutic index. Thus, dehydroilludin M (3)20
and acylfulvene (4)21,22 (the latter prepared by treatment
of illudin S with dilute sulfuric acid) inhibited growth of
tumors in mouse xenografts at doses well tolerated by
the animals. The enhanced efficacy and lower toxicity
correlates with decreased reactivity of dehydroilludin M
and acylfulvene with thiols and also in the enzymatic
reaction with NADPH.23
Products from reactions with thiols have been reported
previously.24,25 When the nucleophile was methanol or
ethanol, corresponding ethers 6 and 7 were obtained. The
latter was also formed with diethyl ether, instead of
ethanol, providing the ethoxy group. A byproduct of this
reaction was the ether 8 formed from two molecules of
irofulven. Ethylene glycol, ethylene bromohydrin, and
glycerol afforded the expected products 9, 10, and 11,
respectively.
Reaction of irofulven with fructose was investigated.
When the compound and a large excess of fructose were
dissolved in acetone-dilute H2SO4 and kept overnight,
the condensed product (from one mole of each reactant)
was obtained. NMR spectral analysis indicated the
product was a mixture. It has been resolved by HPLC
(isocratic elution with CH3CN and H2O) into four com-
(14) Kelner, M. J .; McMorris, T. C.; Estes, L. Biochem. Pharmacol.
1994, 48, 403.
(15) Kelner, M. J .; McMorris, T. C.; Estes, L. Cancer Res. 1995, 55,
4936.
(16) McMorris, T. C.; Kelner, J . J .; Chadha, R. K.; Siegel, J . S.; Moon,
S.-S.; Moya, M. M. Tetrahedron 1989, 45, 5433.
(17) McMorris, T. C.; Kelner, M. J .; Wang, W.; Moon, S.; Taetle, R.
Chem. Res. Toxicol. 1990, 3, 574.
(18) McMorris, T. C.; Kelner, M. J .; Wang, W.; Estes, L. A.; Montoya,
M. A.; Taetle, R. J . Org. Chem. 1992, 57, 6876.
(19) Tanaka, K.; Inoue, T.; Tezuka, Y.; Kikuchi, T. Xenobiotica 1996,
26, 347.
(20) Kelner, M. J .; McMorris, T. C.; Taetle, R. Anticancer Res. 1995,
15, 873.
(21) McMorris, T. C.; Kelner, M. J .; Wang, W.; Diaz, M. A.; Estes,
L. A.; Taetle, R. Experientia 1996, 52, 75.
(22) Kelner, M. J .; McMorris, T. C.; Estes, L.; Starr, R. J .; Ruther-
ford, M.; Montoya, M.; Samson, K. M.; Taetle, R. Cancer Res. 1995,
55, 4936.
(23) McMorris, T. C.; Elayadi, A. N.; Yu, J .; Kelner, M. J . Biochem.
Pharmacol. 1999, 57, 83.
(24) McMorris, T. C.; Yu, J .; Estes, L. A.; Kelner, M. J . Tetrahedron
1997, 53, 14579.
(25) McMorris, T. C.; Yu, J .; Ngo, H.-T.; Wang, H.; Kelner, M. J . J .
Med. Chem. 2000, 43, 3577.

6160 J . Org. Chem., Vol. 66, No. 18, 2001
McMorris et al.
Ta ble 2. Assign m en ts for 13C NMR Sp ectr a of th e
Ta ble 1. Assign m en ts for 13C NMR Sp ectr a of Ir ofu lven
(5) a n d Com p ou n d s 12 a n d 14 (DEP T Assign m en ts Ar e
Also Given )
â-F u r a n ose Moiety of Com p ou n d 12
carbon
12 (furanose moiety)
â-fructofuranose
DEPT
carbon
irofulven (5)
12
14
DEPT
2′
5′
3′
4′
1′
6′
103.42
81.58
77.84
77.58
73.06a
63.50R
102.6
81.6
76.4
75.4
63.6
63.2
C
1
7
9
5
8
6
4
2
15
3
10
14
11
13
12
199.82
161.58
143.20
140.28
136.03
134.43
128.86
77.83
56.30
38.78
27.73
16.85
14.65
13.03
9.80
199.58
162.160
144.27
140.12
135.38
131.20
129.09
77.24
64.49
38.90
27.79
17.14
199.57
162.48
144.19
140.13
135.38
131.89
128.92
77.82
64.82
38.90
27.88
17.43
14.84
13.43
9.99
C
C
C
C
CH
C
C
C
CH2
C
CH3
CH3
CH2
CH3
CH2
CH
CH
CH
CH2
CH2
a
Assignments may be reversed.
Ta ble 3. Assign m en ts for 13C NMR Sp ectr a of
â-P yr a n ose Moiety of Com p ou n d 14
carbon
14 (pyranose moiety)
â-fructopyranose
DEPT
2′
5′
4′
3′
6′
1′
99.15
78.78
72.05
69.98
66.00
61.77
99.1
70.0
70.5
68.4
64.1
64.7
C
14.85
13.39
10.01
CH
CH
CH
CH2
CH2
ponents: two major and two minor compounds. The 13C
NMR spectrum and DEPT analysis of the first major
component (12) indicated that the fructose moiety had
replaced the primary allylic hydroxyl (change in chemical
shift of δ 56.3 to δ 64.49 for CH2 and δ 134.43 to 131.20
ppm for a vinyl carbon, Table 1). Signals for the fructose
moiety suggested a â-furanose structure bonded to irof-
ulven via C-1′ (change in chemical shift of δ 63.6 to 73.06
ppm for CH2), though bonding via C-6′ cannot be ex-
cluded. Chemical shift values of the â-furanose moiety
compared to those of â-fructofuranose are given in Table
2.26
Other analogues possessing a primary allylic oxygen
substituent included the acetate 15 prepared from iro-
fulven and sodium acetate in acetic anhydride. Com-
pound 15 was also obtained with acetic anhydride and
BF3‚Et2O. A byproduct was the tertiary acetate 16. With
benzoyl chloride, irofulven formed the benzoate 17, and
with phenylchloroformate, the corresponding carbonate
18 was obtained. The acetal 19 was prepared by reaction
of irofulven and 2-methoxypropene with POCl3 as cata-
lyst.
Rearrangement of 12 to tautomer 13 occurred when
the solution in CD3OD was allowed to stand for several
days. This led to change in the chemical shifts of the
fructose moiety, most notably the change from δ 63.6 to
71.1 for one of the CH2 carbons. Spectral data for the
other major component (14) indicated a pyranose struc-
ture with bonding of the pyranose via C-5 (change in
chemical shift of δ 70.0 to δ 78.78 ppm, Table 3). The
â-furanose and â-pyranose forms (12 and 14) are expected
to be the major tautomers. The two minor components
may possibly be the corresponding R-tautomers.27
An a logu es fr om Acylfu lven e (4). Irofulven was
obtained by nucleophilic addition of acylfuvene 4 to
formaldehyde. Attempts to synthesize analogues by
similar reaction with acetaldehyde, glyoxal, or acrylo-
nitrile were unsuccessful. However, reaction of 4 with
acrolein proceeded smoothly to give the Michael adduct
20. A byproduct 21 was obtained from this reaction.
Reduction of 20 with sodium cyanoborohydride yielded
the alcohol 22 while oxidation of 20 with J ones reagent
gave the corresponding acid 23.
(26) Koerner, T. A. W.; Voll, R. J .; Cary, L. W.; Younathan, S.
Biochem. Biphys. Res. Commun. 1978, 82, 1273.
(27) Pelmore, H.; Eaton, G.; Symons, M. C. R. J . Chem. Soc., Perkin
Trans. 2 1992, 149.

Structure-Activity Studies of Anitumor Agent Irofulven
J . Org. Chem., Vol. 66, No. 18, 2001 6161
Ta ble 4. IC50 Va lu es for An a logu es Wh en Tested in MV
522 Cellsa
Studies of metabolism of acylfulvene and of irofulven
by rat liver cytosol and NADPH have yielded aromatic
compounds as well as hydroxylated metabolites including
24 and 25.28 The structures of these compounds was
confirmed by partial synthesis from acylfulvene. This
provided enough material for structure-activity studies.
Toxicity of An a logu es P r ep a r ed by Tota l Syn th e-
sis. We recently reported a total synthesis of (()-
hydroxymethylacylfulvene.21 The method makes it pos-
sible to prepare a wider variety of analogues. Thus, two
analogues 26 and 27 have been made. The in vitro
toxicity of these analogues was found to be lower than
that of acylfulvene on short (2 h) or long (48 h) exposure
of MV 522 cells to these compounds. However, analogue
26 was found to have good in vivo activity, similar to that
of mitomycin C, in mice implanted with MV522 cells.30
compd
IC50 (nM)
1 (illudin S)
2 (illudin M)
3 (dehydroilludin M)
4 (acylfulvene)
5 (irofulven)
7
8
9
10
12
15
16
19
20
21
22
24
25
26
4 ( 1
4 ( 1
310 ( 3
350 ( 20
73 ( 8
440 ( 80
320 ( 60
680 ( 180
930 ( 250
18100 ( 5700
1400 ( 200
480 ( 110
170 ( 80
165 ( 55
270 ( 130
850 ( 180
660 ( 200
580 ( 250
4600 ( 200
It is interesting to note the synthesis by Kinder et al.
of even simpler analogues, bicyclic structures containing
spirocyclopropane and unsaturated ketone. Several of
these analogues were reported to have IC50 values less
than 10-6 M against a number of human tumor cell
lines.31
a
For cytotoxicity tests the compounds were dissolved in DMSO
(1 mg/mL stock solution) and the solutions diluted in 20% DMSO/
phosphate buffered saline just prior to addition to cultures of MV
522 cells. Control cells received equal amounts of the DMSO/
phosphate buffered saline. After incubation for 48 h, the cells were
washed, trypan blue was added, and the cells were counted. These
values correlate closely with those determined by colony forming
assay.
Biologica l Activity. There are a number of ways for
determining biological activity of hydroxymethylacylful-
vene analogues. Cytotoxicity tests were used in our initial
studies with illudins and were applied to acylfulvenes
as well. Cells were exposed to the compounds for various
times, and the IC50 values were determined. Thus,
illudins were added to cultures of MV522 (human lung
carcinoma) and HL60 (myeloid leukemia) cells. After 48
h of continuous exposure, cell growth/viability was mea-
sured by trypan blue exclusion. As an alternative to 48
h exposure, cells were exposed to the compounds for 2 h.
This shorter exposure revealed selective toxicity in these
cells due to active uptake of the toxin. For example, the
IC50 values of illudin S in HL60 cells were 10 ( nM and
3 ( 1 nM for 2 h and 48 hours, respectively. For a human
fibroblast (GM637) cell line the corresponding IC50 values
were >1000 nM and 4 ( 1 nM. This indicated rapid
uptake into HL60 cells compared to GM637 cells.4
which does not possess an allylic oxygen function, is on
the same order of magnitude as that of irofulven.
Analogues 24 and 25 obtained from metabolism of
acylfulvene and irofulven were tested and found to be
slightly less toxic than the parent compounds. Thus,
introduction of a second allylic hydroxyl in the acylful-
vene structure appears to partly detoxify the compound.
While in vitro cytotoxicity is a necessary property for
anticancer drugs, selective toxicity of the compounds in
malignant cells compared to normal cells is of much
greater significance. Therefore, a further stage in the
development of new drugs requires in vivo studies,
usually done with mouse xenografts. Such studies have
been carried out with illudin S and M, dehydroilludin
M, acylfulvene, and irofulven. As reported in previous
publications, illudin S and M, although extremely toxic
to malignant cells, had poor efficacy, being extremely
toxic to the host animal. The efficacy of acylfulvene and,
particularly of irofulven, was much greater.
Since the analogues described in this paper all possess
the key structural features found in acylfulvene we
expected them to display comparable efficacy. Several of
the analogues have therefore been tested in vivo using
the MV522 xenograft model. The acetate of irofulven
(analogue 15) was found to be as efficacious as the parent
compound. Likewise, analogue ethers 7, 9, 11, and 19
were comparable to irofulven. Thus, the free allylic
hydroxyl is not essential for good antitumor activity.
An even more definite indication that the allylic
hydroxyl is not necessary was found in analogues 20 and
22, neither of which possesses an allylic oxygen function.
Analogue 20 had an IC50 170 ( 60 nM (cf. irofulven, 75
( nM) and both 20 and 22 showed good activity in the
xenograft model.
The in vitro toxicity of most of the analogues was
determined with MV522 cells for 48 h exposure, and the
results are given in Table 4. None of the analogues was
as toxic as irofulven itself. This suggests that the free
primary allylic hydroxyl might contribute to the toxicity
of irofulven because of increased hydrophilicity. Also, its
inherent reactivity might lead to substitution by a
macromolecular cellular nucleophile at a suitably low pH.
Analogue 15 with an acetoxy rather than hydroxyl group
was expected to be more reactive than irofulven, acetoxy
being a better leaving group. It was actually found to be
slightly less toxic. Thus, there is no simple correlation
between reactivity of the allylic oxygen function and
toxicity. Furthermore the toxicity value found for 20,
(28) McMorris, T. C.; Elayadi, A. N.; Yu, J .; Hu, Y.; Kelner, M. J .
Drug Metab. Dispos. 1999, 27, 983.
(29) McMorris, T. C.; Hu, Y.; Yu, J .; Kelner, M. J . Chem. Commun.
1997, 315.
(30) McMorris, T. C.; Yu, J .; Hu, Y.; Estes, L. A.; Kelner, M. J . J .
Org. Chem. 1997, 62, 3015.
(31) Kinder, F. R.; Wang, R. M.; Bauta, W. E.; Bair, K. W. Bioorg.
Med. Chem. Lett. 1996, 6, 1029.
As mentioned previously, the product of irofulven and
fructose was obtained as a mixture of isomers (12 and
14) that was found to be appreciably less toxic than the
parent compound. This mixture was found to be moder-
(32) McMorris, T. C.; Kelner, M. J .; Dawe, R.; MacDonald, J . R.;
Waters, S. J . Preclinical antitumor activity of a series of acylfulvene
analogues. Proceedings AACR-NCI-EORTC International Conference;
Nov 16-19, 1999, Washington, DC; American Association for Cancer
Research, Inc.: Philadelphia, 1999; Abstract No. 612.

6162 J . Org. Chem., Vol. 66, No. 18, 2001
McMorris et al.
chromatographed on SiO2 with EtOAc-hexanes affording
compound 6 (290 mg, 86%) as a dark orange gum: 1H NMR
(CDCl3) δ 0.62 (ddd, J ) 9.7, 6.7, 4.1 Hz, 1H), 0.98 (ddd, J )
9.9, 6.7, 4.9 Hz, 1H), 1.24 (ddd, J ) 9.7, 6.1, 4.9 Hz, 1H), 1.37
(ddd, J ) 9.9, 6.1, 4.1 Hz, 1H), 1.27 (s, 3H), 2.00 (s, 3H), 2.04
(s, 3H), 3.26 (s, 3H), 3.91 (br s, 1H), 4.29 (dd, 2H), 7.0 (s, 1H);
13C NMR (CDCl3) δ 197.6, 159.7, 142.6, 138.8, 134.3, 129.7,
129.6, 75.9, 65.3, 57.3, 37.3, 27.3, 15.7, 13.9, 12.9, 9.1; EI-
HRMS for C16H20O3 calcd 260.1407, found 260.1408 (M+); UV
λmax (EtOH) 331 nm (ꢀ 5439).
ately active in MV522 xenograft tests. The mixture was
subsequently resolved into two main components that
have similar IC50 values. It is likely that they will have
similar in vivo activity as well, though this has not been
tested.
Biologica l Tests Ca r r ied ou t by NCI: Ca n cer
Dr u g a n d Develop m en t P r ogr a m . Because of the
promising results obtained with many of the irofulven
analogues, they were submitted to the NCI for further
evaluation. They have been found to demonstrate potent
broad-spectrum activity across many of the human cell
lines respresented in the NCI 60-cell panel, with the
majority of GI50 values below 10-6 M. Mouse hollow fiber
test results have demonstrated good activity (total score
ranging from 18 to 46) across a variety of tumor types
and positive net cell kill (9 of 9 analogues tested). In vivo
activity has been demonstrated in the drug-resistant
MV522 mdr 1/gp 170+ human lung carcinoma xenograft
model in which acylfulvene analogues have led to in-
creased life span and tumor growth inhibition or tumor
shrinkage. Three acylfulvene analogues 22 NSC-690179,
20 NSC-690180, and 15 NSC-688985 have displayed
significant activity each producing tumor cures in the
early stage NCI-H123 nonsmall cell lung tumor, the early
stage OVCAR 3 ovarian tumor, and the early stage
RXF 393 renal tumor xenograft models.
Eth er 7 fr om 5 a n d Dieth yl Eth er . To a stirred solution
of 5 (22 mg, 0.89 mmol) in 3 mL of acetone and 1 M H2SO4
solution (1:1) was added 7.5 mL of Et2O. The mixture was
stirred at room temperature for 24 h and was then partitioned
between EtOAc and water. The organic extracts were washed
with saturated NaHCO3 and saline, respectively, until neutral.
After being dried over MgSO4, the solution was concentrated
and chromatographed to give 17 mg of 7 (80%) as a yellow
gum: IR (KBr) 3457, 2968, 1659, 1592, 1502, 1284, 1097 cm-1
;
1H NMR (CDCl3) δ 0.72 (m, 1H), 1.08 (m, 1H), 1.23 (t, J ) 6.9
Hz, 3H), 1.33 (m, 1H), 1.38 (s, 3H), 1.48 (m, 1H), 2.11 (s, 3H),
2.14 (s, 3H), 3.53 (q, J ) 6.9, Hz, 2H), 3.91 (s, 1H), 4.42 (q,
J AB)10.7 Hz, 2H), 7.10 (s, 1H); 13C NMR (CDCl3) δ 197.4,
159.5, 142.2, 138.8, 134.3, 130.0, 126.4, 75.8, 65.0, 63.5, 37.2,
27.2, 15.6, 14.8, 13.8, 12.7, 9.0; EI-HRMS for C17H22O3 calcd
274.1569, found 274.1568; UV λmax 330 nm (ꢀ 7225). The same
ether (7) was obtained with EtOH instead of Et2O.
Syn th esis of Dim er Eth er 8 of Ir ofu lven . To a solution
of 5 (36 mg, 0.146 mmol) in 3 mL of acetone and 1 M H2SO4
solution (1:1) was added 0.5 mL of Et2O. The mixture was
stirred at room temperature for 30 h and was partitioned
between EtOAc and water. The organic extracts were washed
with saturated NaHCO3 and saline, respectively, until neutral.
After being dried over MgSO4, the solution was concentrated
and chromatographed to give 5 mg of 8 (14%), 11 mg of 7, and
13 mg of 5. Compound 8 was a yellow gum: IR (KBr) 3433,
Con clu sion s
Our studies of illudins, acylfulvene, irofulven, and
analogues, led to the following conclusions:
Toxicity and antitumor activity are a result of their
behavior as electrophiles. Bioreductive activation fol-
lowed by cyclopropane ring-opening and alkylation of
protein and DNA, are responsible for biological activity.
Attenuation of reactivity in acylfulvenes and dehydro-
illudin M (compared to illudin S) presumably leads to
lower toxicity and greater selectivity toward malignant
cells versus normal cells.
The hydroxymethyl group in irofulven confers remark-
able efficacy on this compound. However, a free allylic
hydroxyl is probably not the sole reason for this efficacy
since analogues with ether or thioether groups retain
antitumor activity. Moreover, analogues with longer
chain substituents, e.g., hydroxypropyl derivative (22),
have produced tumor cures as mentioned above. There-
fore, an acylfulvene with a side chain at C-6 containing
an oxygen or other heteroatom substituent results in
enhanced efficacy.
1
2920, 1659, 1592, 1502, 1350, 1163 cm-1; H NMR (CDCl3) δ
0.67 (m, 1H), 1.08 (m, 1H), 1.31 (m, 1H), 1.37 (s, 3H), 1.48 (m,
1H), 2.07 (s, 3H), 2.11 (s, 3H), 4.48 (s, 2H), 7.10 (s, 1H); 13C
NMR (CDCl3) δ 197.9, 159.9, 143.3, 139.1, 134.6, 129.6, 126.8,
76.1, 63.2, 37.6, 27.5, 15.9, 14.2, 13.1, 9.4; EI-HRMS for
C
30H35O5 (M + H) calcd 475.2535, found 475.2467; UV λmax
330 nm (ꢀ 12 905).
Rea ction of 5 w ith Eth ylen e Glycol. To a stirred solution
of 5 (9 mg, 0.037 mmol) in 9 mL of acetone and 1 M H2SO4
solution (1:1) was added 4.5 mL of ethylene glycol. The mixture
was stirred at room temperature for 2 h and was partitioned
between EtOAc and water. The organic extracts were washed
with saturated NaHCO3 and saline, respectively, until neutral.
After being dried over MgSO4, the solution was concentrated
and chromatographed to give 11 mg of 9 (100%) as a yellow
gum: IR (KBr) 3439, 2914, 1665, 1598, 1508, 1344, 1103 cm-1
;
1H NMR (CDCl3) δ 0.71 (m, 1H), 1.06 (m, 1H), 1.32 (m, 1H),
1.36 (s, 3H), 1.47 (m, 1H), 2.11 (s, 3H), 2.14 (s, 3H), 2.55 (s,
1H), 3.57 (t, J ) 4.5 Hz, 2H), 3.73 (t, J ) 4.5 Hz, 2H), 3.98 (s,
1H), 4.50 (q, J AB ) 12 Hz, 2H), 7.09 (s, 1H); 13C NMR (CDCl3)
δ 197.9, 160.0, 142.9, 138.9, 134.5, 129.6, 126.8, 76.1, 70.9, 64.2,
61.6, 37.5, 27.4, 16.0, 14.1, 13.1, 9.3; EI-HRMS for C17H22 O4
calcd 290.1518, found 290.1515; UV λmax 331 nm (ꢀ 9404).
Rea ction of 5 w ith 2-Br om oeth a n ol. To a stirred solution
of 5 (188 mg, 0.764 mmol) in 10 mL of acetone and 1 M H2SO4
solution (1:1) was added 5 mL of 2-bromoethanol. The mixture
was stirred at room temperature for 4.5 h and was partitioned
between EtOAc and water. The organic extracts were washed
with saturated NaHCO3 and saline, respectively, until neutral.
After being dried over MgSO4, the solution was concentrated
and chromatographed to give 179 mg of 10 (66%) and a trace
amount of a diether (C15H17O2-OCH2CH2OCH2CH2Br). Com-
pound 10 formed yellow crystals: mp 92-94 °C; IR (KBr) 3445,
Further structure activity studies are planned with
preparation of analogues having side chains of varying
length and with heteroatoms including nitrogen. It is
possible that there is interaction of the side chain with a
receptor and there is a side chain that possesses optinum
binding, which will give the best therapeutic activity.
Exp er im en ta l Section
General directions can be found in previous papers; for
example, see ref 25.
Eth er 6 fr om Ir ofu lven 5 a n d Meth a n ol. To a stirred
solution of 5 (320 mg, 1.3 mmol) in 3 mL of Me2CO were added
MeOH (3 mL) and dilute H2SO4 (1 M, 3 mL). The reaction
mixture was stirred at room temperature for 24 h. It was then
extracted with Et2O, and the combined extracts were washed
with saturated NaHCO3, followed by saturated brine, and
dried over MgSO4. After concentration, the residue was
1
2914, 1650, 1592, 1502, 1097 cm-1; H NMR (CDCl3) δ 0.71
(m, 1H), 1.07 (m, 1H), 1.35 (m, 1H), 1.38 (s, 3H), 1.48 (m, 1H),
2.15 (s, 3H), 3.47 (t, J ) 6.0 Hz, 2H), 3.77 (t, J ) 6.0 Hz, 2H),
3.91 (s, 1H), 4.54 (q, J AB ) 12 Hz, 2H), 7.09 (s, 1H); 13C NMR
(CDCl3) δ 198.1, 160.6, 143.2, 138.9, 134.4, 129.3, 127.0, 76.3,
69.4, 64.1, 37.7, 30.6, 27.6, 16.4, 14.3, 13.2, 9.5; EI-HRMS

Structure-Activity Studies of Anitumor Agent Irofulven
J . Org. Chem., Vol. 66, No. 18, 2001 6163
for C17H20BrO3 calcd 352.0674, found 352.0671; UV λmax 332
nm (ꢀ 7777).
phenyl chloroformate at 0 °C under argon. The mixture was
stirred for 3 h and was partitioned between EtOAc and water.
The organic extracts were washed with saturated saline. After
being dried over MgS04, the solution was concentrated and
chromatographed to give 20 mg of 18 as a yellow gum: 1H
NMR (CDCl3) δ 0.85 (m, 1H), 1.18 (m, 1H), 1.43 (m, 1H), 1.52
(s, 3H), 1.61 (m, 1H), 2.12 (s, 3H), 2.28 (s, 3H), 4.04 (s, 1H),
5.06 (q, J AB ) 11.1 Hz, 2H), 6.93-7.47 (m, 6H).
The diether was a yellow gum: 1H NMR (CDCl3) δ 0.72 (m,
1H), 1.05 (m, 1H), 1.32 (m, 1H), 1.37 (s, 3H), 1.50 (m, 1H),
2.13 (s, 3H), 2.15 (s, 3H), 3.46 (t, J ) 6.3 Hz, 2H), 3.65 (m,
4H), 3.79 (t, J ) 6.3 Hz, 2H), 3.90 (s, 1H), 4.51 (q, J AB ) 12
Hz, 2H), 7.09 (s, 1H); EI-HRMS for C19H26BrO4 calcd 397.1015,
found 397.0996.
Rea ction of Ir ofu lven w ith Glycer ol. To a stirred solu-
tion of 5 (110 mg, 0.447 mmol) in 15 mL of acetone and 1 M
H2SO4 solution (1:1) was added 5 mL of glycerol. The mixture
was stirred at room temperature for 22 h and was partitioned
between EtOAc and water. The organic extracts were washed
with saturated NaHCO3 and saline, respectively, until neutral.
After being dried over MgSO4, the solution was concentrated
and chromatographed (adding 5% methanol to the solvent
system) to give 79 mg of 11 (55% mixture of isomers) as a
yellow gum (with 40 mg of 5 recovered): IR (KBr) 3415, 2926,
Rea ction of 5 w ith 2-Meth oxyp r op en e. To a stirred
solution of 5 (260 mg, 1.057 mmol) in 6 mL of 2-methoxy-
propene was added two drops POCl3. The mixture was stirred
at room temperature for 6 days and was then partitioned
between EtOAc and water. The organic extracts were washed
with saturated NaHCO3 and saline, respectively, until neutral.
After being dried over MgSO4, the solution was concentrated
and chromatographed to give 133 mg of 19 (40%) as yellow
crystals (with 87 mg of 5 recovered): mp 93-95 0 °C; IR (KBr)
1
3457, 2980, 1665, 1598, 1502, 1091 cm-1; H NMR (CDCl3) δ
1
1659, 1586, 1103 cm-1; H NMR (CDCl3) δ 0.72 (m, 1H), 1.08
0.72 (m, 1H), 1.06 (m, 1H), 1.25 (m, 1H), 1.38 (s, 3H), 1.41 (s,
3H), 1.42 (s, 3H), 1.49 (m, 1H), 2.15 (s, 3H), 3.25 (s, 6H), 3.95
(s, 1H), 4.43 (s, 2H), 7.11 (s, 1H); 13C NMR (CDCl3) δ 197.7,
159.5, 142.2, 134.9, 134.8, 130.5, 126.7, 100.3, 76.1, 54.4, 48.6,
37.4, 27.5, 24.4, 24.3, 15.9, 14.0, 13.0, 9.3; EI-HRMS for
(m, 1H), 1.26 (m, 1H), 1.37 (s, 3H), 1.50 (m, 1H), 2.10 (s, 3H),
2.15 (s, 3H), 2.57 (s, 1H), 3.58 (m, 4H), 3.86 (m, 1H), 3.91 (s,
1H), 4.51 (q, J AB ) 12.9 Hz, 2H), 7.10 (s, 1H); 13C NMR (CDCl3)
δ 198.0, 160.1, 143.2, 138.8, 134.6, 129.4, 126.9, 76.2, 70.9, 70.6,
64.4, 63.8, 37.6, 27.4, 16.1, 14.2, 13.1,9.4; EI-HRMS for
C
19H26O4 calcd 318.1831, found 318.1823; UV λmax 330 nm (ꢀ
8728).
Rea ction of Acylfu lven e w ith Acr olein . To a stirred
C
18H24O5 calcd 320.1623, found 320.1616; UV λmax 331 nm (ꢀ
7920).
Rea ction of 5 w ith F r u ctose. To a stirred solution of 5
solution of acylfulvene (4, 1 g, 4.63 mmol) in 5 mL of acetone
and 2.5 mL of 1 M H2SO4 solution was added 2.5 mL of
acrolein. The mixture was stirred at room temperature for 7
h and was partitioned between EtOAc and water. The organic
extracts were washed with saturated NaHCO3 and saline,
respectively, until neutral. After being dried over MgSO4, the
solution was concentrated and chromatographed to give 378
mg of 20 (30%) and 241 mg of 21 (14%).
(1.5 g, 6.098 mmol) in 66 mL of acetone and 40 mL of 1 M
H2SO4 solution was added 20 g of fructose. The mixture was
stirred at room temperature overnight and worked up the
usual way employing methylene chloride and methanol as
solvents. Chromatography afforded 350 mg of 12 and 14 (14%,
mixture) as a yellow gum (701 mg HMAF recovered): EI-
HRMS for C21H29O8 (M + H)+ calcd 409.1863, found 409.1869;
UV λmax 332 nm (ꢀ 4745).
The mixture was resolved by HPLC (isocratic elution with
CH3CN and H2O) into two major compounds 12 and 14 and
two minor compounds (structures undetermined). See Tables
1-3 for spectral data.
Compound 20 was a yellow gum: 1H NMR (CDCl3) δ 0.68
(m, 1H), 1.07 (m, 1H), 1.32 (m, 1H), 1.36 (s, 3H), 1.46 (m, 1H),
2.01 (s, 3H), 2.06 (s, 3H), 2.65 (t, J ) 7.8 Hz, 2H), 3.00 (m,
2H), 3.93 (s, 1H), 7.12 (s, 1H), 9.83 (s, 1H); 13C NMR (CDCl3)
δ 200.4, 196.3, 157.3, 139.4, 138.3, 135.4, 133.7, 125.3, 75.4,
43.5, 36.9, 27.0, 19.5, 15.4, 13.4, 12.4, 8.6; EI-HRMS for
Rea ction of 5 w ith Acetic An h yd r id e. To a stirred
solution of 5 (300 mg, 1.22 mmol) in acetic anhydride (4 mL)
was added sodium acetate (287 mg, 3.5 mmol). The mixture
was stirred overnight at room temperature and then filtered,
and the filtrate was concentrated under reduced pressure.
Chromatography of the residue afforded the acetate 15 (261
mg, 74%) as a yellow gum: 1H NMR (CDCl3) δ 0.740 (m, 1H),
1.103 (m, 1H), 1.324 (m, 1H), 1.385 (s, 3H), 1.507 (m, 1H), 2.037
(s, 3H), 2.086 (s, 3H), 2.171 (s, 3H), 3.898 (s, 1H), 5.100 (s,
2H), 7.109 (s, 1H); 13C NMR (CDCl3) δ 197.56, 170.46, 159.39,
144.25, 138.54, 134.34, 127.23, 126.93, 76.68, 58.21, 37.78,
27.55, 20.99, 16.16, 14.42, 13.27, 9.62. EI-HRMS for C17H20O4
calcd 288.1362, found 288.1364; UV λmax 330 nm (ꢀ 5904).
When 5 was treated with BF3‚Et2O in acetic anhydride at
-78 °C, 15 was the major product and 16 was obtained in low
yield as a yellow gum: 1H NMR (CDCl3) δ 0.97 (m, 1H), 1.16
(m, 2H), 1.46 (m, 1H), 1.51 (s, 3H), 2.10 (s, 3H), 2.14 (s, 3H),
2.19 (s, 3H), 4.60 (s, 1H), 4.65 (s, 2H), 7.18 (s, 1H); EI-HRMS
for C17H20O4 calcd 288.1362, found 288.1364
Rea ction of 5 w ith Ben zoyl Ch lor id e. To a stirred
solution of irofulven (116 mg, 0.447 mmol) in 10 mL of
methylene chloride was added 0.10 mL of pyridine and 0.25
mL of benzoyl chloride at room temperature under argon. The
mixture was stirred for 2 h and was partitioned between
EtOAc and water. The organic extracts were washed with
saturated saline. After being dried over MgSO4, the solution
was concentrated and chromatographed to give 20 mg of 17
(92%) as a yellow gum (with 13 mg of 5 recovered): 1H NMR
(CDCl3) δ 0.65 (m, 1H), 1.02 (m, 1H), 1.18 (m, 1H), 1.32 (s,
3H), 1.44 (m, 1H), 2.03 (s, 3H), 2.16 (s, 3H), 3.86 (s, 1H), 5.28
(q, J AB ) 13.2 Hz, 2H), 7.06 (s, 1H).
C
17H20O3 (M+) calcd 272.1413, found 272.1416; UV λmax 331
nm (ꢀ 8500).
Compound 21 was also a yellow gum (isomeric mixture):
EI-HRMS for C23H28O5 (M+) calcd 384.1937, found 384.1947;
UV λmax 329 nm (ꢀ 6000).
Red u ction of Ald eh yd e 20. To a stirred solution of 20 (30
mg, 0.110 mmol) in 5 mL of THF was added 5 drops of HOAc
and excess sodium cyanoborohydride. The mixture was stirred
at room temperature for 1 h and was partitioned between
EtOAc and water. The organic extracts were washed with
saturated NH4Cl and saline, respectively, until neutral. After
being dried over MgSO4, the solution was concentrated and
chromatographed to give 21 mg of 22 (69%) as a yellow gum:
1H NMR (CDCl3) δ 0.67 (m, 1H), 1.06 (m, 1H), 1.26 (m, 1H),
1.36 (s, 3H), 1.46 (?,1H), 1.73 (m, 2H), 2.06 (s, 3H), 2.07 (s,
3H), 2.74 (m, 2H), 3.70 (t, J ) 6.3 Hz, 2H), 3.96 (s, 1H), 7.14
(s, 1H); 13C NMR (CDCl3) δ 197.0, 157.7, 139.6, 139.0, 136.6,
136.5, 128.2, 75.9, 62.0, 37.3, 33.0, 27.5, 24.0, 15.9, 13.8, 12.8,
9.0; EI-HRMS for C17H22O3 calcd 274.1569, found 274.1557;
UV λmax 330 nm (ꢀ 6700).
Oxid a tion of Ald eh yd e 20. To a stirred solution of
aldehyde 20 (44 mg, 0.162 mmol) in acetone (6 mL) in an ice
bath was added 6 drops of J ones reagent. The mixture was
stirred for 20 min at 0 °C and was quenched by addition of
methanol. The mixture was then partitioned between EtOAc
and saline. The organic extracts were washed twice with saline
and dried over MgSO4. After concentration, the crude product
was chromagraphed to give the acid 23 (29 mg, 62% yield) as
a yellow gum: 1H NMR (CDCl3) δ 0.69 (m, 1H), 0.88 (m, 1H),
1.05 (m, 1H), 1.36 (s, 3H), 1.47 (m, 1H), 2.06 (s, 3H), 2.07 (s,
3H), 2.52 (m, 2H), 3.03 (m, 2H), 7.13 (s, 1H).
Rea ction of 5 w ith P h en yl Ch lor ofor m a te. To a stirred
solution of 5 (163 mg, 0.663 mmol) in 10 mL of methylene
chhloride was added 0.18 mL of pyridine and 0.34 mL of
J O010458Z