Design of highly active analogues of the pyrrolo[1,2-a]benzimidazole antitumor agents

Article abstract of DOI:10.1021/jm990029h

The pyrrolo[1,2-a]benzimidazole (PBI) reductive alkylating agents have been investigated in this laboratory since their discovery in the late 1980s. Of all the structural modifications of the PBIs investigated so far, the variation of the 3-substituent has the greatest influence on cytotoxicity, toxicity, and in vivo antitumor activity. In the present study, we prepared both the R and S enantiomers of nitrogen-containing 3-substituents possessing hydrogen-bonding capability as well as varying basicity. The rationale was to take advantage of stereoselective DT-diaphorase reductive activation as well as hydrogen bonding in the DNA major groove. As a result of these studies, analogues were discovered possessing among the highest hollow fiber tumor assay scores observed in hundreds of natural and synthetic antitumor agents. Our findings indicate that a relatively basic 3-substituent is required for outstanding PBI cytotoxicity but that the importance of using pure enantiomers is still open to study.

Full text of DOI:10.1021/jm990029h

3324
J . Med. Chem. 1999, 42, 3324-3333
Design of High ly Active An a logu es of th e P yr r olo[1,2-a ]ben zim id a zole
An titu m or Agen ts
William A. Craigo, Benjamin W. LeSueur, and Edward B. Skibo*
Department of Chemistry and Biochemistry, Arizona State University, Tempe, Arizona 85287-1604
Received J anuary 19, 1999
The pyrrolo[1,2-a]benzimidazole (PBI) reductive alkylating agents have been investigated in
this laboratory since their discovery in the late 1980s. Of all the structural modifications of
the PBIs investigated so far, the variation of the 3-substituent has the greatest influence on
cytotoxicity, toxicity, and in vivo antitumor activity. In the present study, we prepared both
the R and S enantiomers of nitrogen-containing 3-substituents possessing hydrogen-bonding
capability as well as varying basicity. The rationale was to take advantage of stereoselective
DT-diaphorase reductive activation as well as hydrogen bonding in the DNA major groove. As
a result of these studies, analogues were discovered possessing among the highest hollow fiber
tumor assay scores observed in hundreds of natural and synthetic antitumor agents. Our
findings indicate that a relatively basic 3-substituent is required for outstanding PBI cytotoxicity
but that the importance of using pure enantiomers is still open to study.
Ch a r t 1
In tr od u ction
Work in this laboratory1-4 has been involved in the
design of a new class of DNA cleaving agent based on
the pyrrolo[1,2-a]benzimidazole (PBI) ring system. These
agents were designed to alkylate the phosphate back-
bone upon DT-diaphorase-mediated reduction resulting
in a hydrolytically labile phosphotriester; the reductive
activation to the hydroquinone form is shown in the
inset of Chart 1. While in the hydroquinone form, these
agents can hydrogen-bond to the major groove and
thereby recognize base pairs.5 As expected, the PBIs
possess outstanding antitumor activity by virtue of their
ability to cleave DNA upon reductive alkylation. In the
past, we have investigated the influence of PBI struc-
tural modifications on cytotoxicity. Thus the fused
pyrrolo ring,6 the benzimidazole ring system,6 the
position of the aziridine ring substituent,6 the 7-methyl
substituent,7 and the 3-substituent8 have all been varied
and correlated with cytotoxicity. The 3-substituent
appears to be the most important determinant of
antitumor activity and cytotoxicity, with LC50 values
varying by orders of magnitude as this substituent is
varied. Indeed, previous studies showed that the chiral-
ity of the 3-substituent is important with respect to both
bioactivation by DT-diaphorase and the interaction with
DNA.5 At the outset of the present study, we postulated
that both the hydrogen-bonding capability of the 3-sub-
stituent and the configuration of the 3-stereocenter
could influence the substrate specificity for DT-diapho-
rase as well as the alkylation of DNA. Previous studies
showed that the presence of a carbamate residue at the
3-position greatly increased the alkylation of DNA,
presumably as a result of hydrogen bonding in the major
groove.1,2,8
compounds are one of the most active among the
hundreds of natural and synthetic agents screened in
this assay. Compound S(-)-2 is both highly cytotoxic
and selective possessing LC50 values < 10-8 against
some cancer lines and LC50 values > 10-4 against other
lines. To follow is a description of the unambiguous
synthesis of the enantiomers of 1-3, their cytotoxicity
and in vivo activity, measurements of their substrate
specificity for DT-diaphorase, and finally their percent
incorporation into DNA.
Syn th esis
Therefore we prepared and studied the enantiomers
of the 3-substituted PBIs 1-3 shown in Chart 1. This
study showed that R(+)- and S(-)-1 and its racemic
form possess outstanding activity in the National Can-
cer Institute’s in vivo hollow fiber assay.9 In fact, these
The preparation of the new PBIs is outlined in
Schemes 1 and 2. Previous synthetic studies led to the
preparation of racemic 4 employing 3-bromo-4-nitro-
toluene and either racemic or S(-)-2,4-diaminobutanoic
10.1021/jm990029h CCC: $18.00 © 1999 American Chemical Society
Published on Web 07/30/1999

Highly Active Analogues of PBI Antitumor Agents
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 17 3325
Sch em e 1
Sch em e 2
acid.4,5,10 Acetylation of S(-)-4 to S(-)-5 was followed
by conversion to S(-)-1, Scheme 1. The previously
unknown R(+) enantiomer was prepared from racemic
4 employing an acyltransferase (ALTUS 20 CLEC
catalyst), which afforded R(+)-5 as the major product.
Altus Inc. provides a kit of 40 acyltransferases to screen
for enantiomeric excesses; only the mentioned acyl-
transferase gave a suitable result. Conversion of R(+)-5
to the R(+)-1 was carried out as previously described.5
To prepare the R(+) or S(-) forms of 2 and 3, the R(+)
or S(-) forms of 4 were trifluoroacetylated to afford 7,
Scheme 2. The trifluoroacetyl group was easily removed
after the bromination and nitration reactions to afford
the common intermediate 9. PBI 2 was prepared from
9 by reduction, Fremy oxidation, and finally aziridina-
tion as previously described for the PBIs.10 The same
procedure was employed to convert the 3-carbamido
derivative 10 to the PBI 2.
× 2 sites × 2 doses). A score of 2 is given to each test in
which there is a % T/C of 50 or less (tumor mass 50%
or less than the control). Thus the highest possible score
is 96, but the typical score is only 5, and the highest
score achieved so far is 64. The score is broken down
into an intraperitoneal (ip) and a subcutaneaous (sc)
score. A good sc score (g8) indicates that the drug is
able to get to the tumor site (subcutaneous) from a
distant site (intraperitoneal) of injection. Both PBI-A
and PBI-B shown in Table 1 are active in the hollow
fiber assay and are active against melanoma, NSC lung
cancer, breast cancer, and ovarian cancer. The 3-sub-
stituent of the PBI apparently is crucial for cytotoxic
activity but is not required for the actual reductive
alkylation process, which only requires the aziridi-
nylquinone. Thus 3-unsubstituted PBI and the 3-hy-
droxy PBI are devoid of activity, while the lipophilic
3-propanoate derivative showed animal toxicity (100%
drug deaths) at doses as low as 5 mg/kg without any
significant antitumor activity.8,3,4 The acetate and the
carbamate, PBI-A and PBI-B, respectively, appear to
have a suitable balance of polarity and lipophilicity in
the 3-substituent for antitumor activity. These findings
prompted the investigation of 1, which bears a nitrogen
instead of an oxygen bound to the 3-position. The data
in Table 1 shows that both S(-)-1 and RS-1 have a
higher score than PBI-A and -B. The high sc score of 14
for S(-)-1 indicated that this agent can kill cancer cells
at a location distant from the site of injection. However,
In Vivo a n d In Vitr o Scr een in g Resu lts
In vivo screening results of S(-)- and racemic 1,
employing the relatively new hollow fiber assay,11 are
discussed in conjunction with Table 1 and Figure 1.
Hollow fiber assays were carried out in the following
way. Human tumor cells are cultivated in polyvi-
nylidene fluoride hollow fibers, and a sample of each
cell line is implanted into each of two physiological
compartments (intraperitoneally and subcutaneously).
Mice are treated with either a high or a low dose using
a qd × 4 schedule (four daily treatments) administered
intraperitoneally. Altogether, 12 cell lines are studied
resulting in 48 possible test combinations (12 cell lines

3326 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 17
Craigo et al.
Ta ble 1
the S(-) enantiomer. Studies of both enantiomers, side
by side, against a syngeneic tumor implant seem to
suggest equivalent activity, Figure 2.
The National Cancer Institute has tested hundreds
of new synthetic and natural products in the hollow fiber
screen.9 Among these compounds are PBI-A, PBI-B, S(-
)-1, and RS-1 whose rankings are shown in Figure 1
(the R(+) analogue of 1 is currently being screened by
the NCI). The typical antitumor agent only has a hollow
fiber score of ∼5, while S(-)-1 and RS-1 both score at
52 making them among the most active compounds
studied by the National Cancer Institute.
Close inspection of Figure 1 shows that there are a
few compounds with higher scores than S(-)-1 and RS-
1, the highest score being 64 for a natural product. Many
of these compounds are too toxic and do not show
activity against tumor implants. Indeed, the hollow fiber
assay is not equivalent to drug assays against tumor
implants, where angiogenesis metastases can intervene.
Conversely, all compounds showing high activity against
tumor implants are active in the hollow fiber assay.9
F igu r e 1. Plot of the number of compounds studied at the
National Cancer Institute versus their hollow fiber score.
Development of the PBI analogues in this laboratory has
resulted in lower scores in the hollow fiber assay.
Examples of compounds showing higher hollow fiber
scores than the PBIs are the natural products roridin
A and flavopiridol, both of which are shown in Chart 2
along with their National Cancer Institute numbers.
Roridin A did not perform well in human xenograft
the racemic form of 1 shows a lower sc score with a
higher ip score suggesting that R(+)-1 is less active at
distant locations in this assay. Perhaps the R(+) enan-
tiomer is more susceptible to metabolic inactivation than

Highly Active Analogues of PBI Antitumor Agents
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 17 3327
F igu r e 3. Plots of -log LC50 versus cancer type for R(+)-1
and S(-)-1. Each cancer type represents the average of 6-8
different cancer cell lines. The (+) enantiomer is much more
active against melanoma, renal, ovarian, and CNS cancer cell
lines, but this trend is reversed in colon cancer cell lines.
F igu r e 2. Results of S(-)-1 and R(+)-1 in the B-16 melanoma
model in C57/bl mice (run at University of Arizona Cancer
Center). These compounds were studied at two dose levels: 1
and 3 mg/kg ip on days 1, 5, and 9 after tumor implantation
into the front flank muscle. Shown in the bar graph below is
the area under the tumor versus time curve plotted against
dose for each enantiomer. Note that there are no apparent
differences between the enantiomers in these data.
Ch a r t 2
F igu r e 4. Plots of -log LC50 versus melanoma cell line for
R(+)-1 and S(-)-1. Each -log LC50 value is the average of two
assays. Note that the enantiomeric differences reside in only
a few cell lines.
assays, while the cyclin-dependent kinase inhibitor12
flavopiridol has activity and may show clinical prom-
ise.13
a panel consisting of 6-8 human cancer cell lines.
Therefore each -log LC50 value is the average of 12-
16 values. The identity of these cell lines and the assay
methodology have been published.14,15 Inspection of
Figure 3 reveals that the S(-)-1 is less active than the
R(+) enantiomer in melanoma, ovarian, non-small-cell
lung, and renal cancers, but the trend is reversed in
CNS cancers.
The other cancers have nearly the same average -log
LC50 values for each enantiomer. The enantiomeric
differences in cytotoxicity reside in only a few of the
cancers when individual cancer cell lines in the mela-
noma panel are compared, Figure 4. The origin of the
enantiomeric differences may be due to the substrate
specificity for DT-diaphorase or enantioselective reac-
tions with DNA, both of which are discussed in the
following two sections of this report.
To determine if the PBIs R(+)- and S(-)-1 will possess
in vivo activity, assays employing syngeneic tumor
implants (B-16 melanoma in C57/bl mice) were carried
out as described below. The results of these studies with
R(+)- and S(-)-1, Figure 2, failed to show much differ-
ence in activity between the enantiomers. Both S(-)-
and R(+)-1 did show large reductions in tumor mass
with time, as measured from the area under the tumor
mass versus time curve (AUC). In the syngeneic tumor
graft assay, the tumor implant is in the front flank with
ip drug injection, analogous to the sc implants of the
hollow fiber assay. The activity of 1 in this assay is
encouraging since a high hollow fiber assay score often
can translate to high toxicity or inactivity in the in vivo
assay.
In order to assess the influence of the 3-substituent
configuration, the cytotoxicity of R(+)- and S(-)-1
against a variety of cancers is compared in Figure 3.
Each cancer type has a -log LC50 value for each
enantiomer, which is the average of two assays against
Besides studying the influence of stereochemistry, we
also varied the hydrogen-bonding capabilities of the
3-substituent. The properties of the 3-substituent are
crucial for antitumor activity with low toxicity. Figure

3328 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 17
Craigo et al.
F igu r e 6. Km, Vmax, and Vmax/Km parameters obtained for the
enantiomers of 1 and 3 with purified rat liver DT-diaphorase.
The Vmax/Km value is a measure of the relative specificity of
these substrates for the enzyme.
Depending on the organ and species source, the
substrate and inhibition properties of DT-diaphorase
can vary widely.18,20 The structure of the enzyme even
varies by ethnic origin, which may be a factor in the
success of cancer chemotherapy in some patients.21
Since the structure of DT-diaphorase varies by organ
system, the enzyme could also vary by the organ system
cancer. Thus cancers originating from these organs
(lung, renal, ovarian, etc.) will a possess slightly differ-
ent DT-diaphorase, assuming some degree of differen-
tiation. Our previous studies indicated that the bulk of
the 3-substituent and its stereochemistry5 influence the
substrate specificity of DT-diaphorase.
F igu r e 5. Plots of -log LC50 versus cancer type for S(-)-1,
2, and 3. Each cancer type represents the average of 6-8
different cancer cell lines. The 3-amino derivative 2 shows high
cytotoxicity against non-small-cell lung, CNS, melanoma,
ovarian, and renal cancers.
1 shows the increasing activity in the hollow fiber assay
as the 3-substituent is changed from carbamate (PBI-
B), to acetate (PBI-A), and finally to the 3-acetamido
(1) substituent. When the 3-hydroxy group is changed
to 3-amino (2), there is nearly a millionfold change in
cytotoxicity (the 3-hydroxy analogue is noncytotoxic with
LC50 values > 10-4, while the 3-amino analogue is very
cytotoxic with LC50 values in the nanomolar range in
some cell lines). This finding suggests that a basic group
in the 3-position is required for cytotoxicity because the
large difference in cytotoxicity seems to reflect the
relative pKa values of a protonated amine versus a
protonated alcohol.
Figure 5 contrasts the relative cytotoxicities of the
S(-) forms of 1-3 against several types of cancer.
Preliminary studies indicate that the 3-amino substitu-
ent greatly enhances the alkylation of DNA, perhaps
explaining the high cytotoxicity. At the present time we
are carrying out in vivo studies of 2 to determine if the
high cytotoxicity translates to antitumor activity or
toxicity.
In the present study both the stereochemistry and the
hydrogen-bonding capabilities of the 3-substituent (Chart
1) were varied to assess their influence on DT-diapho-
rase substrate specificity and cancer selectivity. The Km,
Vmax, and Vmax/Km parameters for both enantiomers of
1 and 3 are provided in Figure 6. For both 1 and 3, the
DT-diaphorase has a 2-fold higher specificity for the
R(+) enantiomer than the S(-) enantiomer. This ob-
servation could explain the higher cytotoxicity of the
R(+) enantiomer against some cancer panels, Figures
3 and 4. However, the heterogeneity of the enzyme could
either reverse this trend or result in equal cytotoxicity
for the enantiomers in other cancer panels.
Validation of the concept of differential antitumor
activity based on differential DT-diaphorase specificity
will rely on the isolation and characterization of the DT-
diaphorase substrate specificity from a series of human
cancer cell lines. Since only one human cancer DT-
diaphorase has been cloned so far,22 validation of this
concept will be a slow process.
DT-Dia p h or a se Su bstr a te Activity
The enantiomeric differences in cytotoxicity illus-
trated in Figures 3 and 4 could have their origin in the
substrate specificity of DT-diaphorase. This enzyme can
activate antitumor agents by two-electron reduction in
both normal and cancerous tissues.16,17 Some cancer
cells can be sensitive to a reductive alkylating agent by
virtue of high levels of DT-diaphorase18 or a high
substrate specificity. Recent studies have shown that
there is a direct correlation between the substrate
activity for DT-diaphorase and cytotoxicity.19 Our earlier
PBIs, as well as 1-3, show a high correlation with the
concentration of DT-diaphorase in cancer cells.8 The cor-
relation coefficients are in the 0.6-0.7 range, but never
higher. Perhaps the heterogeneous structure of DT-
diaphorase could be responsible for the absence of a high
correlation coefficient. Thus, a low substrate specificity
of the PBI for the cell’s DT-diaphorase would diminish
the correlation with the enzyme’s concentration.
DNA Red u ctive Alk yla tion
Once the PBI is reductively activated in a specific
cancer cell, cytotoxicity would result from the efficient
alkylation of DNA by the reduced species. The work in
this laboratory has validated the major groove binding
model,2,3 wherein the reduced PBI binds to the DNA
major groove in a preequilibrium step followed by
phosphate alkylation,1,2 Chart 3. Although rigorous
evidence of major groove binding is still lacking, the
model has been employed in the design of efficient DNA
cleaving agents. Presumably, the structure and stere-
ochemistry of the 3-substituent greatly influences the
degree of DNA alkylation by hydrogen bonding in the
major groove.2,3,7,8 For example, the reduced carbamate
derivative PBI-B interacts predominately at A-T base

Highly Active Analogues of PBI Antitumor Agents
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 17 3329
Ch a r t 3
protonation of the amino group and better hydrogen
bonding in the major groove as illustrated in Chart 3,
or perhaps an electrostatic interaction between the
positive nitrogen center and the phosphate backbone.
The data in Figure 7 also show enantiomeric differences
in reductive alkylation percent values. Since we used
homopolymeric DNA, rather than alternating polymers,
the 3-substituent will point in the 3′- or 5′-direction
depending on the enantiomer employed. Differing hy-
drogen-bonding interactions will result in a variable
percent alkylation and even alter base pair specificity,
as in the case of R(+)-2. The differential cytotoxicity of
enantiomers (Figures 3 and 4) probably does not origi-
nate with DNA base pair recognition, although we are
investigating this possibility further. The site(s) of DNA
alkylation by reduced 1-3 is currently under investiga-
tion, although phosphate alkylation is likely based on
our experience with the PBIs.
Un iqu en ess of th e P BIs
The differences between the clinically used aziridinyl
compounds and the PBIs were apparent from COM-
PARE correlations8 as well as from previously reported
mechanistic studies.1,2 COMPARE was developed at the
NCI to compare the patterns of cytotoxicity in 60-cell
line cancer screens.14 Antitumor agents with identical
mechanisms of action possess identical or nearly identi-
cal cytotoxicity patterns (correlation coefficient > 0.8).
For example, anthracycline analogues (doxorubicin,
rubidazone, daunamycin) have a high correlation (>0.9)
with each other as do the DNA alkylating agents
(chlorambucil, thiotepa, triethylenemelamine). In con-
trast, the cytotoxicity pattern of the PBIs does not
correlate highly with those of any known antitumor
agent. Shown in Chart 4 are antitumor agents with the
highest correlation coefficients with S(-)-1 and R(+)-1
out of ∼20 000 compounds in the NCI’s archives. The
magnitude of the correlation coefficients in Chart 4
suggests some overlay in cytotoxicity mechanism with
the PBIs. Like the PBIs, triethylenemelamine alkylates
DNA,23 while tetrocarcin A interferes with DNA syn-
thesis.24,25 The overall uniqueness of the PBIs probably
originates with their substrate specificity for DT-dia-
phorase and DNA alkylation mechanism (major groove
binding and phosphate alkylation as opposed to N-7
alkylation).
F igu r e 7. Bar graph of percent DNA base pair reductive
alkylation of poly(dG)‚poly(dC) and poly(dA)‚poly(dT) by the
enantiomers of 1-3.
pairs presumably due to hydrogen bonding of the
carbamate amine group with the thymine carbonyl,2
Chart 3.
Con clu sion s
Found in Figure 7 is the percent base pair alkylation
of DNA homopolymers obtained with the reduced enan-
tiomers of 1-3. PBI alkylation of DNA followed by air
oxidation affords the blue-colored aminoquinone (blue
DNA), permitting an easy quantitation of alkylation
employing UV-vis spectroscopy (see Experimental Sec-
tion).2 The percent alkylation is based on the number
of alkylations per base pair: all base pairs alkylated
once is 100%, and two alkylations per base pair is 200%.
The data in Figure 7 indicate that the 3-substituents
of 1-3 in most cases favor A-T over G-C alkylation. The
3-amino derivative 2 is a highly efficient reductive
alkylating agent of DNA and it can alkylate up to 2
times per base pair (200%) upon reductive activation.
This alkylated DNA is insoluble in all solvents except
hot DMSO and therefore difficult to work with. The
reason for the overalkylation by reduced 2 may be the
The previously studied PBIs bear an oxygen substitu-
ent at the 3-position, usually attached to an ester or
carbamate group. These PBIs showed good antitumor
activity, which prompted structure-activity studies over
the years.3,4 The present study shows that the place-
ment of a nitrogen at the 3-position, either as an amine,
carbamido, or amide, results in a more active PBI with
respect to both cytotoxicity and antitumor activity. The
reason for the enhanced cytotoxicity/antitumor activity
appears to be the efficient reductive alkylation of DNA
brought about by a relatively basic 3-substituent. For
example the 3-amino analogue 2 can reductively alky-
late DNA up to 2 times per base pair when present in
excess. The importance of the 3-substituent basicity is
well-illustrated by comparing the cytotoxicity of the
3-hydroxy and 3-amino PBIs: the 3-hydroxy substituent
has LC50 values > 10-4, while the 3-amino PBI has LC50

3330 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 17
Craigo et al.
obtained from a 300-MHz spectrometer. All chemical shifts are
reported relative to TMS. Optical rotations were obtained from
a Perkin-Elmer polarimeter employing a 10-cm quartz cell, 0.2
g/dL enantiomer in methanol. Enantiomeric purity was de-
termined using optical rotations and chiral HPLC (see Chiral
High-Pressure Liquid Chromatography). Intermediates lead-
ing to the pure enantiomers of 1-3 were characterized by their
rotation of plane-polarized light (sodium D-line) and/or chiral
HPLC. The final products 1-3 did not show clear rotation of
the sodium D-line, and comfirmation of enantiomeric purity
was made by chiral HPLC; all products were >99% enantio-
merically pure. The syntheses of new compounds are outlined
below.
Ch a r t 4
(R)-(+)-3-Tr iflu or oacetam ido-2,3-dih ydr o-7-m eth yl-1H-
p yr r olo[1,2-a ]ben zim id a zole (R(+)-7) was prepared from
racemic 4 by the following three-step process. To 543 mg (2.9
mmol) of racemic 4 in a 250-mL round-bottom flask was added
310 mg of ALTUS 20 CLEC catalyst along with 104 mL of
ethyl acetate. The resulting mixture was stirred vigorously for
30 min, and solids were filtered off and washed 3× with 10-
mL portions of CHCl3. The extraction solvent was removed
under reduced pressure to afford a residue, which was placed
on a silica gel column employing 3% methanol in chloroform
as the eluent. Pure R(+)-5 eluted from the column as the first
band. The product was recrystallized from ethyl acetate/
hexane: 244 mg (37%) yield; mp 219-220 °C; TLC (chloroform/
methanol [90:10]) Rf ) 0.38; IR (KBr pellet) 3245, 3068, 2990,
1635, 1560, 1420 cm-1 1H NMR (CDCl3) δ 7.75 (1H, s,
;
acetamido proton) 7.47 (1H, d, J ) 8.4 Hz, 5-aromatic proton),
7.02 (1H, d, J ) 8.4 Hz, 6-aromatic proton), 6.88 (1H, s,
8-aromatic proton), 5.41 (1H, m, 3-methine proton), 4.0-3.8
(2H, m, 1-methylene protons), 3.25-3.15 (1H, m, 2-methylene
proton), 2.55-2.39 (1H, m, 2-methylene proton) 2.43 (3H, s,
methyl), 2.09 (3H, s, methyl); MS (EI) m/z 229 (M+), 186 (M+
- COCH3), 171, 158, 145, 133, 116, 104. Rotations: R(+) [R]25
) +104.4° (c ) 0.41, MeOH).
D
The deacetylation of R(+)-5 was carried out by dissolving
210 mg (0.877 mmol) in 12 mL of 1.2 N hydrochloric acid and
refluxing the solution for 9 h. The solvent was removed under
vacuum to afford a white solid, which was dissolved in water
and then buffered to pH 7 with saturated sodium bicarbonate
solution. The neutralized solution was extracted 5× with 30-
mL portions of chloroform, and the extracts were dried over
sodium sulfate. Concentration of the dried extracts and recrys-
tallization from chloroform/hexane afforded 155 mg (95%) of
R(+)-4 as a white solid. Rotations: R(+)-4 [R]25 ) +20.9° (c
D
) 1.1, MeOH) and S(-)-4 [R]25 ) -20.7° (c ) 0.2, MeOH).
D
Trifluoroacetylation of R(+)-4 was carried out by dissolving
155 mg (0.83 mmol) in 3 mL of trifluoroacetic acid followed by
addition of 3 mL of trifluoroacetic anhydride. The reaction
mixture was stirred for 20 min at room temperature and then
poured into 150 mL of 0.1 M pH 7.0 phosphate buffer. The
resulting solution was extracted 3× with 20-mL portions of
ethyl acetate, and the extracts were washed 2× with 10-mL
portions of saturated sodium bicarbonate and then dried over
sodium sulfate. The extracts were concentrated to a residue,
which was recrystallized from ethyl acetate/hexane to afford
values < 10-8 in human cancer cell lines. A conclusion
regarding the better enantiomer for antitumor studies
has not been reached, although the R(+) enantiomer
seems to be more active than the S(-) enantiomer
against some cancer panels. The greater activity of the
R(+) enantiomer may stem from its higher substrate
specificity for DT-diaphorase. Further in vivo antitumor
and toxicity studies with 1-3 should determine which
is the better enantiomer for clinical trials.
152 mg (65%) of R(+)-7. Rotation: R(+) [R]25 ) +101.2° (c )
D
0.55, MeOH).
(S)-(-)-3-Tr iflu or oa ceta m id o-2,3-d ih yd r o-7-m eth yl-1H-
p yr r olo[1,2-a ]ben zim id a zole (S(-)-7). A solution of 683 mg
(3.05 mmol) of S(-)-4 was prepared in 20 mL of triflouroacetic
acid. To this solution was added 20 mL of trifluoroacetic
anhydride, and the reaction was stirred for 25 min at room
temperature. The completed reaction was poured into 700 mL
of 0.1 M pH 7.0 phosphate buffer. This solution was extracted
4× with 40-mL portions of ethyl acetate, and then the extracts
were washed 3× with 20-mL portions of saturated aqueous
bicarbonate solution and dried over Na2SO4. The solvent was
removed under reduced pressure and the product recrystal-
lized from ethyl acetate/hexane: 546 mg (63%) yield; mp 227-
228 °C; TLC (chloroform/methanol [90:10]) Rf ) 0.68; IR (KBr
pellet) 3178, 2987, 2912, 2858, 1724, 1572, 1523, 1213, 1186,
Exp er im en ta l Section
All solutions and buffers for kinetic, DNA, and electrophore-
sis studies used doubly distilled water. All analytically pure
compounds were dried under high vacuum in a drying pistol
over refluxing toluene. Elemental analyses were run at
Atlantic Microlab, Inc., Norcross, GA. All TLCs were performed
on silica gel plates using a variety of solvents and a fluorescent
indicator for visualization. IR spectra were taken as thin films
and the strongest absorbances reported. 1H NMR spectra were
1
1153 cm-1; H NMR (CDCl3) δ 9.75 (1H, bs, -NH), 7.41 (1H,

Highly Active Analogues of PBI Antitumor Agents
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 17 3331
d, J ) 8.1 Hz, 5 - aromatic proton) 7.02 (1H, d, J ) 8.1 Hz,
6-aromatic proton), 6.79 (1H, s, 8-aromatic proton), 5.43 (1H,
q, J ) 6.3 Hz, 3-methine proton), 4.08-3.89 (2H, m, 1-meth-
ylene proton), 3.31-3.89 (1H, m, 2-methylene proton), 2.69-
2.58 (1H, m, 2-methylene proton), 2.41 (3H, s, 2-methyl
protons); MS (EI mode) m/z 283 (M+), 186 (M+ - CF3CO), 121,
131, 105. Rotation: S(-)-7 [R]25D ) -108.7° (c ) 0.42, MeOH).
Anal. (C13H12F3N3O) C, H, N.
) -8.1° (c ) 0.42, MeOH); S(+) [R]25 ) 10.1° (c ) 0.16,
D
MeOH). Anal. (C11H11BrN4O2‚0.25H2O) C, H, N.
3-Am in o-6-a zir idin yl-2,3-d ih yd r o-7-m eth yl-1H-p yr r olo-
[1,2-a ]ben zim id a zole-5,8-d ion e (2). A solution consisting of
70 mg (0.225 mmol) of 9, 70 mg of 5% Pd on charcoal, and 70
mL of methanol was shaken under 50 psi H2 for 24 h. The
reaction mixture was filtered through Celite, evaporated to
dryness, and then combined with a solution of 0.333 g KH2-
PO4 in 30 mL of water. The resulting solution was combined
with another solution consisting of 1.0 g of KH2PO4 and 0.700
g of Fremy salt in 50 mL of water. The reaction mixture was
stirred at room temperature for 6 h and then concentrated
under high vacuum to a residue, which was placed on a 25-
mL Bakerbond phenyl reverse-phase column prepared with
100% water. The unstable aminoquinone was eluted from the
column with water, and the combined fractions were concen-
trated to a dry residue, which was dissolved in 10 mL of
methanol and 0.3 mL of aziridine. This reaction mixture was
stirred at room temperature for 5 h. The solvent was evapo-
rated, and the residue was purified by flash column chroma-
tography employing chloroform/methanol [95:5] as the eluent.
The product fractions were evaporated to dryness, and the
residue was recrystallized from chloroform/hexane to afford 2
as a red solid: 6.0 mg (10.3%) overall yield; mp > 240 °C dec;
TLC (chloroform/methanol [80:20]) Rf ) 0.17; IR (KBr pellet)
3387, 2996, 2924, 1674, 1632, 1576, 1518, 1377, 1341, 1312,
1140, 988, cm-1; 1H NMR (DMSO-d6) δ 4.23-4.15 and 4.06-
3.97 (3H, 2m, 3-methine and 1-methylene), 2.89-2.78 and
2.27-2.16 (2H, 2m, 2-methylene), 2.29 (4H, s, aziridinyl
protons), 1.94 (3H, s, 7-methyl); MS (EI) m/z 258 (M+), 240,
214. Anal. (C13H14N4O2‚0.25H2O) C, H; N: calcd, 21.32; found,
20.30.
6-Br om o-3-tr iflu or oa ceta m id o-2,3-d ih yd r o-7-m eth yl-5-
n itr o-1H-p yr r olo[1,2-a ]ben zim id a zole (8) was prepared by
the following two-step procedure starting with 7. To a solution
of 500 mg (1.76 mmol) of 7 in 25 mL of acetic acid was added
0.25 mL of a solution of bromine 0.8 M in acetic acid. The
resulting solution was stirred at room temperature for 20 min
and then quenched by dilution with 500 mL of 0.1 M pH 7.0
phosphate buffer. The resulting mixture was extracted 3× with
80-mL portions of ethyl acetate, and then the extracts were
washed 2× with 80-mL portions of saturated aqueous sodium
bicarbonate solution. The extracts were dried over Na2SO4, and
the solvent was removed under reduced pressure to afford the
bromo derivative which was recrystallized from ethyl acetate/
hexane: 546 mg (85%) yield; mp 215-217 °C; TLC (chloroform/
methanol [90:10]) Rf ) 0.50; IR (KBr pellet) 3178, 2984, 2859,
-1
1730, 1570, 1522, 1445, 1221, 1186, 1148 cm
;
1H NMR
(CDCl3) δ 9.9 (1H, bs, amide NH), 7.61 (1H, s, 5-aromatic
proton), 6.83 (1H, s, 8-aromatic proton), 5.48-5.40 (1H, m,
3-methine proton), 4.17-3.95 (2H, m, 1-methylene protons),
3.33-3.26 (1H, m, 2-methylene proton), 2.77-2.65 (1H m,
2-methylene proton), 2.42 (3H, s, methyl); MS (EI) m/z 363 &
361 (M+, 79Br & 81Br), 264 & 266 (M+ - COCF3), 223, 169,
130, 90. Rotations: R(+) [R]25 ) +92.6° (c ) 0.47, MeOH);
D
S(-) [R]25D ) -97.8° (c ) 0.18, MeOH). Anal. (C13H11BrF3N3O)
6-Br om o-3-ca r b a m id o-2,3-d ih yd r o-7-m et h yl-5-n it r o-
1H-p yr r olo[1,2-a ]ben zim id a zole (10) was prepared by the
following two-step procedure starting with 9. A solution of 150
mg (0.482 mmol) 9 in 9 mL of dry pyridine was cooled to 0 °C
in an ice bath. To the cooled solution was added 0.3 mL of
phenyl chloroformate, and the reaction mixture was stirred
at 0 °C for 15 min. The mixture was then removed from the
ice bath and allowed to warm to room temperature with
stirring over 1 h. The reaction mixture was diluted with 50
mL of ethyl acetate and the resulting mixture washed 3× with
15 mL of 20% aqueous acetic acid, 2× with 10-mL portions of
0.12 N aqueous hydrochloric acid, and finally 2× with 15 mL
of water. The ethyl acetate layer was dried over Na2SO4, and
the solvent was evaporated to a residue, which was recrystal-
lized from chloroform/hexane to afford the phenoxyamido
derivative as a light brown solid: 162 mg (78%) yield; mp 175-
178 °C; TLC (chloroform/methanol [80:20]) Rf ) 0.66; IR (KBr
C, H, N.
To 12 mL of fuming nitric acid, cooled to 0 °C was slowly
added 500 mg (1.4 mmol) of the bromo derivative with stirring.
To the resulting solution was added 1.14 mL of acetic anhy-
dride with continued stirring and cooling. After stirring at 0
°C for 5 min, the reaction mixture was allowed to come to room
temperature and then stirred for 1.5 h. The completed reaction
was poured into 450 mL of ice water and the mixture buffered
to pH 7 with NaHCO3. The product was extracted 3× with
50-mL portions of ethyl acetate. The extracts were then dried
over Na2SO4, and solvent was removed to afford a yellow
residue. Recrystallization from a minimum amount of ethyl
acetate, facilitated by the addition of hexane, afforded 8 as
light yellow crystals: 466 mg (89%) yield; mp 245-247 °C; TLC
(chloroform/methanol [90:10]) Rf ) 0.60; IR (KBr pellet) 3246,
3044, 1741, 1657, 1570, 1538, 1370, 1229, 1159 cm-1; 1H NMR
(CDCl3) δ 10.21 (1H, s, amide NH), 7.38 (1H, s, 8-aromatic
proton), 4.75-4.70 (1H, m, 3-methine proton), 4.52-4.42 (1H,
m, 1-methylene proton), 4.20-4.12 (1H, m, 1-methylene
proton), 3.50-3.40 (1H, m, 2-methylene proton), 2.95-2.85
(1H, m, 2-methylene proton), 2.58 (3H, s, methyl); MS (EI) m/z
406 (M+), 361 (M+ - NO2), 309, 293, 280, 263, 189. Rotations:
R(+) [R]25D ) +70.12° (c ) 0.42, MeOH); S(-) [R]25D ) - 64.0°
(c ) 0.39, MeOH). Anal. (C13H10BrF3N4O3) C, H, N.
1
pellet) 3306, 1736, 1537, 1491, 1208 cm-1; H NMR (DMSO-
d6) δ 8.55 (1H, d, J ) 9 Hz, amide proton), 7.87 (1H, s, C(8)
proton), 7.4-6.7 (5H, 3m, aromatic protons), 5.28 (1H, q, J )
6.6 Hz, 3-methine), 4.36-4.26 and 4.18-4.09 (2H, 2m, 1-
methylene), 3.13-3.01 and 2.61-2.49 (2H, 2m, 2-methylene),
2.54 (3H, s, 7-methyl); MS (EI) m/z 430 and 432 (M+, 79Br &
81Br). Rotations: R(-) [R]25D ) +13.6° (c ) 0.26, MeOH); S(+)
[R]25 ) -14.0° (c ) 0.33, MeOH). Anal. (C18H15BrN4O4‚0.25
D
3-Am in o-6-br om o-2,3-dih ydr o-7-m eth yl-5-n itr o-1H-pyr -
r olo[1,2-a ]ben zim id a zole (9). To a solution of 45 mL of
methanol cooled to -70 °C in an 2-propanol/dry ice bath was
bubbled gaseous ammonia until the volume of liquid doubled.
To the resulting solution was added 300 mg (0.737 mmol) of
8, and the mixture was then removed from the dry ice bath
and stirred for 48 h at room temperature. The solvent was
evaporated, and the residue was recrystallized from chloroform/
hexane to afford light brown crystals: 151 mg (66%) yield; mp
H2O) C, H, N.
Gaseous ammonia was passed into a dry flask cooled to -70
°C with 2-propanol/dry ice bath until 20 mL of liquid ammonia
was obtained. A solution of 150 mg (0.348 mmol) of the
phenoxyamido derivative in 10 mL of dry methylene chloride
was added to the liquid ammonia, and the resulting mixture
was stirred for 30 min at -70 °C. The reaction mixture was
then removed from the dry ice bath and stirred for 3 h at room
temperature. The solvent was evaporated, and the solid
residue was recrystallized from chloroform/hexane to afford
10 as an off-white solid: 104 mg (84.5%) yield; mp 240-241
°C; TLC (chloroform/methanol [80:20]) Rf ) 0.34; IR (KBr
pellet) 3410, 3293, 2922, 1657, 1588, 1532, 1371 cm-1; 1H NMR
(DMSO-d6) δ 7.83 (1H, d, J ) 1.2 Hz, aromatic proton), 6.69
(1H, d, J ) 8.4 Hz, amide protons), 5.69 (2H, s, amide protons),
5.22 (1H, m, 3-methine), 4.30-4.21 and 4.11-4.03 (2H, 2m,
1-methylene), 2.99-2.88 and 2.41-2.30 (2H, 2m, 2-methylene),
145-148 °C; TLC (chloroform/methanol [80:20]) Rf ) 0.22; IR
1
(KBr pellet) 3389, 2924, 1532, 1458, 1381, 1294, 878 cm-1
;
H NMR (CDCl3) δ 7.36 (1H, s, aromatic proton), 4.58 (1H, dd,
J ) 7.9 Hz, J ) 6.6 Hz, 3-methine proton), 4.29-4.21 and
4.09-4.01 (2H, 2m, 1-methylene protons), 3.15-3.03 and
2.51-2.39 (2H, 2m, 2- methylene), 2.58 (3H, s, 7-methyl); MS
(EI) m/z 310 and 312 (M+, 79Br & 81Br), 293 and 295 (M+
-
D
NH3), 263 and 265 (M+ - NH3 - NO). Rotations: R(-) [R]25

3332 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 17
Craigo et al.
2.54 (3H, s, 7-methyl); MS (EI) m/z 353 and 355 (M+, 79Br &
81Br). Anal. (C12H12BrN5O3‚0.5 H2O) C, H, N.
was calculated from the initial and final absorbance values
for complete quinone reduction; usual value for ꢀ is 6000-8000
M-1 cm-1. The value of ∆ꢀ was used to calculate Vmax in M
s-1. The results were fitted to a Lineweaver-Burke plot from
which Vmax/Km values were calculated.
Ch ir a l High -P r essu r e Liqu id Ch r om a togr a p h y. All
HPLC separations were performed on Chirex columns (50 ×
3.2 mm) from Phenomenex. The detector was set at 287 nm
and the sensitivity at 0.5. The flow rate was 0.1 mL/min, and
the mobile phase consisted of hexane/1,2-dichloroethane/
ethanol in a ratio of 60/30/10.
3-Car bam ido-6-azir idin yl-2,3-dih ydr o-7-m eth yl-1H-pyr -
r olo[1,2-a ]ben zim id a zole-5,8-d ion e (3) was prepared from
10 by the following two-step procedure. A solution of 120 mg
(0.339 mmol) of 10, 80 mg of 5% Pd on charcoal, and 200 mL
of methanol was shaken under 50 psi H2 for 15 h. The reaction
mixture was filtered through Celite, and the solvent was
removed by evaporation. The residue was dissolved into a
solution of 296 mg of KH2PO4 in 120 mL of water. To this
solution was added a Fremy salt solution consisting of 1.2 g
of Fremy salt and 1.8 g of KH2PO4 in 174 mL of water. The
reaction was stirred at room temperature for 1.5 h and then
concentrated with high vacuum. The residue was placed on a
25-mL Bakerbond phenyl reverse-phase column prepared with
100% water. The yellow product fractions were collected from
the column employing water as the eluent. The fractions were
evaporated to dryness, and the residue was recrystallized from
ethyl acetate/hexane: 46 mg (52%) yield; mp 180-183 °C; TLC
(chloroform/methanol [80:20]) Rf ) 0.48; IR (KBr pellet) 3393,
3297, 3212, 1659, 1593, 1570, 1516, 1329, 1152 cm-1; 1H NMR
(DMSO-d6) δ 6.64 (1H, d, J ) 8.4 Hz, amide proton), 6.57 (1H,
d, J ) 2.1 Hz, aromatic proton), 5.67 (2H, s, amine protons),
5.09-5.01 (1H, m, 3-methine), 4.29-4.21 and 4.14-4.06 (2H, 2m,
1-methylene), 2.99-2.88 and 2.41-230 (2H, 2m, 2-methylene),
2.01 (3H, d, J ) 1.5 Hz, 7-methyl); MS (EI) m/z 260 (M+),
243 (M+ - NH3), 217 (M+ - H-NdCdO). Anal. (C12 H12N4O3‚
0.35H2O) C, H, N.
A solution of 20 mg (0.077 mmol) of the carbamide quinone,
10 mL of dry methanol, and 0.3 mL of aziridine was stirred at
room temperature for 3 h. The reaction solvent was evapo-
rated, and the red residue was dissolved into a 9:1 chloroform/
methanol mixture and purified by flash column chromatog-
raphy using 9:1 chloroform/methanol as the eluent. Evaporation
of the solvent gave a red solid which was recrystallized from
chloroform/hexane: 8.1 mg (35.0%) yield; mp 200-205 °C; TLC
(chloroform/methanol [80:20]) Rf ) 0.37; IR (KBr pellet) 3351,
1678, 1647, 1522, 1316, cm-1; 1H NMR (DMSO-d6) δ 6.62 (1H,
d, J )7.8 Hz, amide proton), 5.65 (2H, s, amide protons), 5.04-
4.96 (1H, m, 3-methine), 4.25-4.17 and 4.10-4.01 (2H, 2m,
1-methylene), 2.97-2.86 and 2.39-2.30 (2H, 2m, 2-methyl-
ene), 2.30 (4H, s, aziridinyl protons), 1.95 (3H, s, 7-methyl);
MS (EI) m/z 301 (M+). Anal. (C14H15N5O3‚1.0H2O) C, H; N:
calcd, 21.90; found, 21.19.
Alk yla tion of DNA by Red u ced P BIs. To a mixture of
1-2 mg of DNA (poly[dA]‚poly[dT] or poly[dG]‚poly[dC]) in 2.0
mL of 0.05 M pH 7.4 Tris buffer and 2 mg of Pd on carbon
was added a 5:1 base-pair equivalent amount of the PBI
dissolved in 0.5 mL of dimethyl sulfoxide. The resulting
solution was degassed under argon for 30 min, after which
the mixture was purged with H2 for 10 min. The solution was
then purged with argon for 10 min and placed in a 30 °C bath
for 24 h. The reaction was opened to the air, and the catalyst
was removed with a Millex-PF 0.8-µM syringe filter. The
filtrate was adjusted to 0.3 M acetate with a 3 M stock solution
of pH 5.1 acetate and the diluted with two volumes of ethanol.
The mixture was chilled at -20 °C for 12 h and the DNA pellet
collected by centrifuging at 12 000g for 20 min. The pellet was
redissolved in water and then precipitated and centrifuged
again. The resulting pellet was suspended in ethanol, centri-
fuged, and dried.
DT-Dia p h or a se Red u ction Kin etics Stu d ies. Rat liver
DT-diaphorase was isolated as previously described.5,26 Kinetic
studies were carried out in 0.05 M pH 7.4 Tris‚HCl buffer,
under anaerobic conditions, employing Thunberg cuvettes. A
2 mM stock solution of the enantiomers of 1 and 2 was
prepared in dimethyl sulfoxide. To the top port was added the
quinone stock and to the bottom port was added DT-diaphorase
and NADH in the Tris buffer. The top and bottom ports were
purged with argon for 20 min and equilibrated to 30 °C. The
ports were then mixed, and the reaction was followed at 296
nm for 25 min to obtain initial rates. The concentrations after
mixing were 0.3 mM NADH, 1-20 × 10-5 M quinone, and 14.5
nM (based on flavin) of enzyme active sites. The value of ∆ꢀ
Ack n ow led gm en t. We thank the American Cancer
Society, the National Institutes of Health, the National
Science Foundation, and the Arizona Disease Control
Commission for their generous support. We also thank
the National Cancer Institute for supplying hollow fiber
data.
Refer en ces
(1) Skibo, E. B.; Schulz, W. G. Pyrrolo[1,2-a]benzimidazole-Based
Aziridinyl Quinones. A New Class of DNA Cleaving Agent
Exhibiting G and A Base Specificity. J . Med. Chem. 1993, 36,
3050-3055.
(2) Schulz, W. G.; Nieman, R. A.; Skibo, E. B. Evidence for DNA
Phosphate Backbone Alkylation and Cleavage by Pyrrolo[1,2-
a]benzimidazoles, Small Molecules Capable of Causing Sequence
Specific Phosphodiester Bond Hydrolysis. Proc. Natl. Acad. Sci.
U.S.A. 1995, 92, 11854-11858.
(3) Skibo, E. B. The Discovery of the Pyrrolo[1,2-a]benzimidazole
Antitumor Agents -The Design of Selective Antitumor Agents.
Curr. Med. Chem. 1996, 2, 900-931.
(4) Skibo, E. B. Pyrrolobenzimidazoles in cancer treatment. Exp.
Opin. Ther. Patents. 1998, 8, 673-701.
(5) Skibo, E. S.; Gordon, S.; Bess, L.; Boruah, R.; Heileman, J .
Studies of Pyrrolo[1,2-a]benzimidazole Quinone DT-Diaphorase
Substrate Activity, Topoisomerase II Inhibition Activity, and
DNA Reductive Alkylation. J . Med. Chem. 1997, 40, 1327-1339.
(6) Skibo, E. B.; Islam, I.; Heileman, M. J .; Schulz, W. G. Structure-
Activity Studies of Benzimidazole-Based DNA-Cleaving Agents.
Comparison of Benzimidazole, Pyrrolobenzimidazole and Tet-
rahydropyridobenzimidazole Analogues. J . Med. Chem. 1994, 37,
78-92.
(7) Zhou, R.; Skibo, E. B. Chemistry of the Pyrrolo[1,2-a]benzimi-
dazole Antitumor Agents: Influence of the 7-Substituent on the
Ability to Alkylate DNA and Inhibit Topoisomerase II. J . Med.
Chem. 1996, 39, 4321-4331.
(8) Schulz, W. G.; Islam, E.; Skibo, E. B. Pyrrolo[1,2-a]benzimida-
zole-Based Quinones and Iminoquinones. The Role of the 3-Sub-
stituent on Cytotoxicity. J . Med. Chem. 1995, 38, 109-118.
(9) Hollingshead, M.; Plowman, J .; Alley, M.; Mayo, J .; Sausville,
E. Relevance of Tumor Models in Anticancer Drug Development.
Contributions to Oncology; Krager Verlag: Germany, 1999.
(10) Skibo, E. B.; Islam, I.; Schulz, W. G.; Zhou, R.; Bess, L.; Boruah,
R. The Organic Chemistry of the Pyrrolo[1,2-a]benzimidazole
Antitumor Agents. An Example of Rational Drug Design. Synlett
1996, 297-309.
(11) Hollingshead, M. G.; Alley, M. C.; Camalier, R. F.; Abbott, B.
J .; Mayo, J . G.; Malspeis, L.; Grever, M. R. Life Sci. 1995, 27,
131-141.
(12) Sedlacek, H. H.; Czech, J .; Naik, R.; Kaur, G.; Worland, P.;
Losiewicz, M.; Parker, B.; Carlson, B.; Smith, A.; Senderowicz,
A.; Sausville, E. Flavopiridol (L86 8275; NSC 649890), a new
kinase inhibitor for tumor therapy. Int. J . Oncol. 1996, 9, 1143-
1168.
(13) Schrump, D. S.; Matthews, W.; Chen, G. A.; Mixon, A.; Altorki,
N. K. Flavopiridol Mediates Cell Cycle Arrest and Apoptosis in
Esophageal Cancer Cells. Clin. Cancer Res. 1998, 4, 2885-2890.
(14) Paull, D. K.; Shoemaker, R. H.; Hodes, L.; Monks, A.; Scudiero,
D. A.; Rubinstein, L.; Plowman, J .; Boyd, M. R. Display and
Analysis of Differential Activity of Drugs Against Human Tumor
Cell Lines: Development of Mean Graph and COMPARE
Algorithm. J . Natl. Cancer Inst. 1989, 81, 1088-1092.
(15) Alvarez, M.; Robey, R.; Sandor, V.; Nishiyama, K.; Matsumoto,
Y.; Paull, K.; Bates, S.; Fojo, T. Using the national cancer
institute anticancer drug screen to assess the effect of mrp
expression on drug sensitivity profiles. Mol. Pharmacol. 1998,
54, 802-814.
(16) Riley, R. J .; Workman, P. DT-diaphorase and Cancer Chemo-
therapy. Biochem. Pharmacol. 1992, 43, 1657-1669.
(17) Marin, A.; deCerain, A. L.; Hamilton, E.; Lewis, A. D.; Martin-
ezPenuela, J . M.; Idoate, M. A.; Bello, J . DT-diaphorase and
cytochrome B-5 reductase in human lung and breast tumours.
Br. J . Cancer 1997, 76, 923-929.

Highly Active Analogues of PBI Antitumor Agents
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 17 3333
(18) Rauth, A. M.; Goldberg, Z.; Misra, V. DT-diaphorase: Possible
roles in cancer chemotherapy and carcinogenesis. Oncol. Res.
1997, 9, 339-349.
(19) Beall, H. D.; Winski, S.; Swann, E.; Hudnott, A. R.; Cotterill, A.
S.; OSullivan, N.; Green, S. J .; Bien, R.; Siegel, D.; Ross, D.;
Moody, C. J . Indolequinone antitumor agents: Correlation
between quinone structure, rate of metabolism by recombinant
human NAD(P)H: quinone oxidoreductase, and in vitro cyto-
toxicity. J . Med. Chem. 1998, 41, 4755-4766.
(20) Wu, K. B.; Knox, R.; Sun, X. Z.; J oseph, P.; J aiswal, A. K.; Zhang,
D.; Deng, P. S. K.; Chen, S. Catalytic properties of NAD(P)H:
quinone oxidoreductase-2 (NQO2), a dihydronicotinamide ribo-
side dependent oxidoreductase. Arch. Biochem. Biophys. 1997,
347, 221-228.
(21) Kelsey, K. T.; Ross, D.; Traver, R. D.; Christiani, D. C.; Zuo, Z.
F.; Spitz, M. R.; Wang, M.; Xu, X.; Lee, B. K.; Schwartz, B. S.;
Wiencke, J . K. Ethnic variation in the prevalence of a common
NAD(P)H quinone oxidoreductase polymorphism and its impli-
cations for anti-cancer chemotherapy. Br. J . Cancer 1997, 76,
852-854.
(22) Beall, H. D.; Mulcahy, R. T.; Siegel, D.; Traver, R. D.; Gibson,
N. W.; Ross, D. Metabolism of Bioreductive Antitumor Com-
pounds by Purified Rat and Human DT-Diaphorase. Cancer Res.
1994, 54, 3196-3201.
(23) Reynolds, R. C. Aziridines. In Cancer Chemotherapeutic Agents;
Foye, W. O., Ed.; American Chemical Society: Washington, DC,
1995; pp 187-197.
(24) Morimoto, M.; Fukui, M.; Ohkubo, S.; Tamaoki, T.; Tomita, F.
Tetrocarcins, New Antitumor Antibiotics. 3. Antitumor Activity
of Tetrocarcin A. J . Antibiot. (Tokyo) 1982, 35, 1033-1037.
(25) Tamaoki, T.; Kasai, M.; Shirahata, K.; Ohkubo, S.; Morimoto,
M.; Mineura, K.; Ishii, S.; Tomita, F. Tetrocarcins, Novel
Antitumor Antibiotics. II. Isolation, Characterization and An-
titumor Activity. J . Antibiot. (Tokyo) 1980, 33, 946-950.
(26) Ho¨jeberg, B.; Blomberg, K.; Stenberg, S.; Lind, C. Biospecific
Adsorption of Hepatic DT-Diaphorase on Immobilized Dicou-
marol. Arch. Biochem. Biophys. 1981, 207, 205-216.
J M990029H