Structure based drug design and in vitro metabolism study: Discovery of N-(4-methylthiophenyl)-N,2-dimethyl-cyclopenta[d]pyrimidine as a potent microtubule targeting agent

Article abstract of DOI:10.1016/j.bmc.2018.04.010

We report a series of tubulin targeting agents, some of which demonstrate potent antiproliferative activities. These analogs were designed to optimize the antiproliferative activity of 1 by varying the heteroatom substituent at the 4′-position, the basicity of the 4-position amino moiety, and conformational restriction. The potential metabolites of the active compounds were also synthesized. Some compounds demonstrated single digit nanomolar IC₅₀ values for antiproliferative effects in MDA-MB-435 melanoma cells. Particularly, the S-methyl analog 3 was more potent than 1 in MDA-MB-435 cells (IC₅₀ = 4.6 nM). Incubation of 3 with human liver microsomes showed that the primary metabolite of the S-methyl moiety of 3 was the methyl sulfinyl group, as in analog 5. This metabolite was equipotent with the lead compound 1 in MDA-MB-435 cells (IC₅₀ = 7.9 nM). Molecular modeling and electrostatic surface area were determined to explain the activities of the analogs. Most of the potent compounds overcome multiple mechanisms of drug resistance and compound 3 emerged as the lead compound for further SAR and preclinical development.

Full text of DOI:10.1016/j.bmc.2018.04.010

Accepted Manuscript
Structure based drug design and in vitro metabolism study: discovery of N-(4-
methylthiophenyl)-N,2-dimethyl-cyclopenta[d]pyrimidine as a potent microtu-
bule targeting agent
Weiguo Xiang, Shruti Choudhary, Ernest Hamel, Susan L. Mooberry, Aleem
Gangjee
PII:
S0968-0896(17)31952-1
https://doi.org/10.1016/j.bmc.2018.04.010
BMC 14293
DOI:
Reference:
Bioorganic & Medicinal Chemistry
To appear in:
Received Date:
Revised Date:
Accepted Date:
4 October 2017
26 March 2018
3 April 2018
Please cite this article as: Xiang, W., Choudhary, S., Hamel, E., Mooberry, S.L., Gangjee, A., Structure based drug
design and in vitro metabolism study: discovery of N-(4-methylthiophenyl)-N,2-dimethyl-cyclopenta[d]pyrimidine
as a potent microtubule targeting agent, Bioorganic & Medicinal Chemistry (2018), doi: https://doi.org/10.1016/
j.bmc.2018.04.010
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Structure based drug design and in vitro metabolism study:
discovery of N-(4-methylthiophenyl)-N,2-dimethyl-
cyclopenta[d]pyrimidine as a potent microtubule targeting agent
,
Weiguo Xianga,Ɨ, Shruti Choudharya, Ernest Hamelb, Susan L. Mooberryc *, ǁ, Aleem Gangjee a,*, ǁ
a Division of Medicinal Chemistry, Graduate School of Pharmaceutical Sciences, Duquesne
University, 600 Forbes Avenue, Pittsburgh, PA 15282.
b Screening Technologies Branch, Developmental Therapeutics Program, Division of Cancer
Treatment and Diagnosis, Frederick National Laboratory for Cancer Research, National
Institutes of Health, Frederick, MD 21702.
c Department of Pharmacology, Cancer Therapy & Research Center, University of Texas Health
Science Center at San Antonio, 7703 Floyd Curl Drive, San Antonio, TX 78229.
Ɨ Current address: Department of Internal Medicine, University of Michigan, Ann Arbor, MI
48105.
ǁ These authors contributed equally to this work.
*To whom correspondence should be addressed. For A.G.: phone, 412-396-6070; fax, 412-396-
5593; e-mail, gangjee@duq.edu. For S.L.M.: phone, 210-567-4788; fax, 210-567-4300; e-mail,
mooberry@uthscsa.edu.
1

Abstract
We report a series of tubulin targeting agents, some of which demonstrate potent antiproliferative
activities. These analogs were designed to optimize the antiproliferative activity of 1 by varying
the heteroatom substituent at the 4’-position, the basicity of the 4-position amino moiety, and
conformational restriction. The potential metabolites of the active compounds were also
synthesized. Some compounds demonstrated single digit nanomolar IC50 values for
antiproliferative effects in MDA-MB-435 melanoma cells. Particularly, the S-methyl analog 3
was more potent than 1 in MDA-MB-435 cells (IC50 = 4.6 nM). Incubation of 3 with human liver
microsomes showed that the primary metabolite of the S-methyl moiety of 3 was the methyl
sulfinyl group, as in analog 5. This metabolite was equipotent with the lead compound 1 in MDA-
MB-435 cells (IC50 = 7.9 nM). Molecular modeling and electrostatic surface area were
determined to explain the activities of the analogs. Most of the potent compounds overcome
multiple mechanisms of drug resistance and compound 3 emerged as the lead compound for
further SAR and preclinical development.
Keywords: cyclopenta[d]pyrimidine, microtubule targeting agent, metabolism
Abbreviations: TBAs, tubulin binding agents; HB, hydrogen bond; Pgp, P-glycoprotein; ACN,
acetonitrile; Rr, resistance ratio; TLC, thin layer chromatography.
2

1. Introduction
Figure 1. Structures of microtubule targeting agents.
Tubulin binding agents (TBAs) (Figure 1) are a structurally diverse group of compounds and
many have utility in the treatment of cancer.1 Five well established classes of TBAs are classified
on the basis of their interactions within the  heterodimers of microtubules.2 Taxoid and
laulimalide site agents stabilize microtubules, while vinca, maytansine and colchicine site agents
inhibit microtubule formation and cause loss of cellular microtubules. However,
triazolopyrimidines, which bind to vinca site, promote microtubule stability.3 Drugs, including
3

paclitaxel, cabazitaxel, docetaxel , ixabepilone and several other chemotypes bind to β-tubulin at
a site on the interior of the microtubule known as the taxoid site.4 Taxoids are useful in the
treatment of breast, lung, head and neck, ovarian, and prostate cancers.5 Laulimalide and
peloruside A also stabilize microtubules, but they bind to a non-overlapping site on  tubulin on
the exterior of the microtubule.6 The vinca domain agents bind to β-tubulin at the αβ interface
between different tubulin dimers. Vincristine, vinblastine, vinorelbine and eribulin are successful
anticancer agents. Maytansine binds to a site on  tubulin, which is distinct from the vinca site.
Maytansine is used as the payload of an antibody drug conjugate Kadcyla®, a drug used to treat
advanced HER2 positive breast cancer.7 A diverse collection of small molecules along with
colchicine and the combretastatins bind to the colchicine site on β-tubulin at its interface with α-
tubulin within a single tubulin heterodimer.8
The expression of P-glycoproteins (Pgp), a drug efflux pump, and βIII-tubulin, one of many
β-tubulin isotypes, are clinically established mechanisms of drug resistance to microtubule
targeting drugs.9 Most colchicine site binding agents are not Pgp substrates and also possess the
ability to circumvent βIII-tubulin mediated drug resistance.10 Although colchicine itself is not
used as an anticancer agent, several colchicine site agents have been evaluated in clinical trials,
including 2-methoxyestradiol, combretastatin A-4 (CA-4), the phosphorylated prodrug
combretastatin A-4 phosphate, the combretastatin A-1 phosphate prodrug, BNC105P (sodium 6-
methoxy-2-methyl-3-(3,4,5-trimethoxybenzoyl)benzofuran-7-yl phosphate), verubulin (MPC-
6827), ABT-751 (N-[2-[(4-hydroxyphenyl)amino]-3-pyridinyl]-4-methoxybenzenesulfonamide)
and plinabulin (NPT-2358).11 However, to date no colchicine site agent has advanced to Phase III
studies, attesting to the necessity to develop viable colchicine site agents for evaluation as
potential cancer chemotherapeutic drugs.
2. Rationale
Table 1. Structures of lead compound 1 and the designed compounds 2-14.
4

Compd
structure
Compd
structure
Compd
structure
1
6
11
2
3
7
8
12
13
14
4
5
9
10
We previously reported compound 1 (Table 1) as a potent colchicine site antiproliferative
agent (IC50 = 7.0 nM in MDA-MB-435 cells antiproliferative assay).12 In the structure-activity
relationship13 studies of cyclopenta[d]pyrimidine based antitubulin agents, we demonstrated the
5

importance of the para-methoxyphenyl moiety for biological activity. However, as illustrated by
marketed drugs anisindione, adapalene and olodaterol, the major metabolic pathway of the para-
methoxyphenyl moiety is O-demethylation.14,15 The O-demethylated metabolite of 1, compound 2
was synthesized and found to be much less potent (IC50 = 405 nM MDA-MB-435 cells) than 1.
The potential metabolic instability of 1 and poor activity of its probable metabolite 2 propelled us
to pursue cytotoxic compounds with the cyclopenta[d]pyrimidine scaffold and different
substituents at the 4’-anilino site.
The distance and the nature of heteroatom substitution affects hydrogen bond (HB) strength.16
The HB acceptor (HBA) ability of sulfur is slightly weaker than that of oxygen.17 Nitrogen has a
comparable HBA ability as oxygen.18,19 Thus, it was of interest to isosterically replace the oxygen
atom of the 4’-OCH3 of 1 with a sulfur moiety as in 3 or an amino moiety as in 4. The primary
metabolite of drugs containing the methylthiophenyl moiety (e.g., marketed drug thioridazine) is
the methylsulfinylphenyl (e.g., marketed drug mesoridazine).20 To identify the metabolic pathway
of compound 3, an in vitro metabolism study of 3 in human liver microsomes was performed.21
The principal metabolite of the methylthio moiety in 3 was the methylsulfinyl, and one of the
minor metabolites was the methylsulfonyl moiety. Thus, the methylsulfinyl analog 5 and the
methylsulfonyl analog 6 were also synthesized to determine their antiproliferative activities. The
major metabolic pathway of the dimethylaminophenyl moiety (e.g., marketed drug mifepristone)
is N-demethylation.22 Similarly, the 4’-dimethylamino compound 4 is expected to be metabolized
to the monomethylamino 7, and the amino 8 via N-demethylation. Hence, compounds 7 and 8
were also synthesized to evaluate the antiproliferative activity of the potential metabolites of 4.
Compounds 9 and 10 were designed to restrict the conformation of the 4’-methoxy moiety. In
addition, replacement of the aromatic phenyl ring with a saturated cyclohexyl ring, as in
6

compound 11, was also synthesized to determine the importance of the aromatic phenyl moiety to
activity.
We were also interested in the homologation of the para-methoxy phenyl moiety. This was
accomplished by adding a methylene moiety between the phenyl ring and the 4-position amine, as
in 12. This homologation decreases the 4’-methoxy H-bonding distance to water molecule
HOH728 in the x-ray crystal structure of colchicine at the colchicine site (PDB:4O2B23) and
could increase potency. Substitution of a methyl group on the methylene of 12 affords a chiral
center and sterically restricts the conformational preferences of the C4-N bond (bond a, Table 1),
which could affect biological activity. Conformational search about bond a was carried out using
Sybyl-X 2.1.1 with 1° rotation.24 There were 79 low energy conformations generated for
compound 12 within 1 kcal/mol. For compounds 13 and 14, the number of low energy
conformations generated were 46 and 34, respectively, indicating that low energy conformations
of 13 and 14 were indeed decreased compared to 12.
Figure 2. A. Superimposition of the docked poses of 1 (green), 3 (magenta) and colchicine (pink)
in the colchicine site of tubulin (PDB ID: 4O2B). A dash line represented the border of -tubulin
7

and -tubulin. B. Superimposition of the docked poses of 1 (green), 5 (orange) and colchicine
(pink) in the colchicine site of tubulin (PDB ID: 4O2B).
Compounds 1-14 were docked in the x-ray crystal structure of the colchicine site in tubulin
(PDB: 4O2B, 2.3 Å)23 using the parameters specified in the experimental section. Multiple low
energy conformations were obtained on docking. As representative examples, Fig. 2A shows the
docked conformation of 3 (magenta), superimposed on the co-crystallized ligand, colchicine
(pink), and lead compound 1 (green). The cyclopenta[d]pyrimidine scaffold of 3 forms
hydrophobic interactions with Alaα180, Valα181, Asnβ258, Metβ259 and Lysβ352. The N4-
methyl in 1 is oriented towards a hydrophobic pocket lined by residues Alaβ250, Lysβ254 and
Leuβ255, but, in 3, the N-methyl group adopts a different conformation and interacts with
Valβ315 and Thrβ353. The N-phenyl group of 3 superimposes on the A ring of colchicine and
interacts with Leuβ248, Alaβ250, Leuβ255 and Alaβ316. The 4'-methoxy group in 1 interacts
with Valβ238, Cysβ241, and Ileβ378, and in 3 it interacts with Cysβ241 and Leuβ242. In addition,
the oxygen atom of the 4'-methoxy group lies within H-bonding distance (3.0 Å for 1 and 2.16 Å
for 3) with a water molecule (HOH728) present in the crystal structure in the vicinity of Cysβ241.
The best docked poses of 1 and 3 had scores of -7.91 kcal/mol and -8.07 kcal/mol, respectively,
indicating that the conformation of 3 was similar to that of 1. Fig. 2B shows the docked pose of 5
(orange), superimposed on the crystallized ligand colchicine (pink) and lead compound 1 (green).
The mode of binding of 5, the principal metabolite of 3, in tubulin was similar to that of 1, except
that for 5 the 4'-sulfoxide group undergoes H-bonding with HOH728. Compound 5 had a docked
score of -8.19 kcal/mol, similar to those of 1 and 3. All the remaining compounds, except 2 and 6,
docked similarly in the colchicine site of tubulin and had scores that ranged between -7.66 and -
8.35 kcal/mol, similar to that of lead compound 1. Compounds 2 and 6 failed to dock in a pose
similar to that shown for 1, suggesting that these analogs may show reduced binding or no
binding at the colchicine site and were included as potential negative controls.
8

3. Chemistry
Scheme 1. Synthesis of intermediate 4-chloro-2-methylcyclopenta[d]pyrimidine, 17.
Reagents and conditions: (a) Acetamidine hydrochloride, KOt-Bu, DMF, 120 oC, 79%; (b) POCl3,
100 oC, 4 h, 73%.
Cyclization of commercially available 15 (Scheme 1) with acetamidine hydrochloride and
potassium tert-butoxide in DMF afforded the cyclopenta[d]pyrimidine 16 in 79% yield.25
Chlorination of 16 with excess POCl3 gave 4-chlorocyclopenta[d]pyrimidine 17 in 73% yield.
Scheme 2. Synthesis of anilines and methylthio derivatives.
Reagents and conditions: (a) NaH, THF, CH3I, 0 oC, 65-75%; (b) mCPBA (1.1 eq.), ACN, 0 oC,
1 h, 67%; (c) mCPBA (1.1 eq.), ACN, 0 oC, 1 h, 59%.
Monomethylation of 18, 22 and 24 (Scheme 2) with CH3I and NaH in THF provided the N-
methyl anilines 19, 23 and 25, respectively, in 65-75% yields. Oxidation of the N-methyl-4-
9

methylthioaniline 19 with meta-chloroperoxybenzoic acid (mCPBA) (1.1 eq.) in acetonitrile
o
(ACN) at 0 C afforded the sulfoxide congener 20 in 67% yield.26 Further oxidation of the
sulfoxide with an additional 1.1 equivalent of mCPBA under similar oxidation condition afforded
the sulfone 21 in 59% yield.
Scheme 3. Synthesis of anilines, derivatives and related intermediates.
Reagents and conditions: (a) HCOOH, Ac2O, DCM, rt, 4-6 h; (b) THF, LiAlH4, 60 oC, 53-72%;
(c) CBzCl, ACN, TEA, 0 oC, 90%; (d) NaH, THF, CH3I, 0 oC, 68%; (e) MeOH, Pd/C, H2 (55
psi), 2 h, 92%.
Formyl protection of the amines in 26a-b and 33a-c (Scheme 3) with formic acid and acetic
anhydride afforded the N-formyl anilines 27a-b and 34a-c. Reduction of the formyl group in 27a-
b and 34a-c with LiAlH4 in THF afforded the corresponding N-methyl anilines in 53-72% yields.
Protection of the amino group in trans-4-hydroxycyclohexylamine 29 with CBzCl and TEA in
10

ACN afforded 30 in 90% yield. N- and O- methylation of 30 with NaH and CH3I afforded 31 in
68% yield. Reductive deprotection of 31 with Pd catalyzed hydrogenation provided 32 in 92%
yield.
Scheme 4. General synthesis of target compounds 3-7, 9-14.
Reagents and conditions: compounds 19-21, 23, 25, 28a-b, 32 and 35a-c, dioxane, HCl (2 N in
dioxane, 1 drop), W, 120-160 oC, 3-6 h, 61-93%.
Nucleophilic displacement of the 4-chlorocyclopenta[d]pyrimidine 17 (Scheme 4) with the
appropriate secondary amines in compounds 19-21, 23, 25, 28a-b, 32 and 35a-c afforded
compounds 3-7 and 9-14 (Table 1) in 61-93% yields.
Scheme 5. Synthesis of target compound 2.
Reagents and conditions: HBr (48% aq.), 80 oC, 3-6 h, 72%.
HBr induced O-demethylation27 of 1 (Scheme 5) afforded 2 in 72% yield.
Scheme 6. Synthesis of target compound 8.
11

Reagents and conditions: (a) N-methyl-4-nitroaniline, dioxane, HCl (2 N in dioxane, 1 drop), W,
130 oC, 3 h, 79%; (b) MeOH, Pd/C, H2 (55 psi), rt, 69%;
Nucleophilic displacement of 4-chlorocyclopenta[d]pyrimidine 17 (Scheme 6) with N-
methyl-4-nitroaniline afforded 36 in 79% yield. Pd catalyzed hydrogenation of 36 gave 8 in 69%
yield.
4. Biological Evaluation and SAR
Table 2. IC50 values for inhibition of proliferation of MDA-MB-435 cells and EC50 values for
microtubule depolymerizing activities in A-10 cells.
IC50 ± SD
(MDA-MB-435)
Cancer cells
(nM)
IC50 ± SD
(MDA-MB-435)
Cancer cells
(nM)
EC50 in A-10
cells
EC50 in A-10
cells
Compd
Compd
(nM)
(nM)
26
5,560
13
7.0 ± 0.7
405 ± 75
4.6 ± 0.5
8.2 ± 0.6
7.9 ± 0.5
ND
70
12.1 ± 0.9
18.4 ± 1.9
ND
1
2
3
4
5
6
9
56
10
11
12
13
14
>10,000
>10,000
>10,000
>10,000
40
450 ± 14
ND
31
>10,000
ND
12

45
20 ± 1.9
CA-4
13.0
3.4 ± 0.6
7
8
5,120
1223 ± 70
Compounds 2-14 were tested for their antiproliferative effects against the drug sensitive
MDA-MB-435 cancer cells in culture using a sulforhodamine B (SRB) assay28,29 with the
antitubulin agents CA-4 and compound 1 as positive controls. The microtubule disrupting effects
of these compounds were evaluated in a cell-based phenotypic screen29 and the EC50 for
microtubule depolymerizing effects determined.
Compound 2 (Table 2), the possible phenolic (4’-hydroxylphenyl) metabolite of 1, had low
activity, and this finding emphasized the necessity to optimize 1 structurally. The S-methyl
analog 3 was more potent in the MDA-MB-435 antiproliferative assay (IC50 = 4.6 nM) and as a
microtubule depolymerization agent (EC50 = 13 nM) than 1. The 4’-dimethylamino analog 4 was
as potent as the lead compound 1, for antiproliferative activity in MDA-MB-435 cells (IC50 = 8.2
nM).
The major metabolite of 3, i.e., the sulfoxide analog 5, was equipotent with lead compound 1
in both the antiproliferative (IC50 = 7.9 nM) and microtubule depolymerization assays (EC50 = 31
nM). One of the minor metabolites of 3, the sulfonyl analog 6, was inactive. The 4’-methylamino
compound 7 (MDA-MB-435, IC50 = 20 nM), one of the probable metabolites of 4, was 2.9-fold
less potent than 1. 4’-Amino compound 8, the other probable metabolite of 4, had low activity.
The conformationally restricted dihydrobenzofuran analog 9 was 1.7-fold less potent than 1
in MDA-MB-435 cells (IC50 = 12.1 nM). Analog 10, in which the conformation of the 4’-
methoxy moiety was partially restricted, was about 2.6-fold less potent than 1 (MDA-MB-435:
IC50 = 18.4 nM). In 9, the methylene group resides at an angle with the phenyl ring that is dictated
13

by the envelope shape of the dihydrofuran ring, thus providing restriction to rotation. In 10, the
3’-methyl group could create steric hindrance and partially restrains the free rotation of the 4’-
methoxy moiety. The lower activities of 9 and 10 compared with 1 indicate that an ortho
substitution to 4’-methoxy group on phenyl ring is not favorable to antiproliferative activity.
A
B
C
14

Figure 3. A. Electrostatic surface of the colchicine site binding pocket in tubulin. B. Electrostatic
surface of colchicine. C. Electrostatic surfaces of compounds 1-8. Red surface indicates electron
rich surface, blue surface indicates electron deficient surface, and white surface represents
hydrophobic surface. Residues Val238, Cys241, Leu248 and Asp251 belong to the β-chain of
tubulin.
The docking of compounds 1-14 in the colchicine site suggested the compounds to be potent
inhibitors of tubulin, except 2 and 6. To further rationalize the low activity of 2, 6, 7, and 8, the
electrostatic surface of the colchicine site pocket (Fig. 3A), of colchicine (Fig. 3B) and of
compounds 1-8 (Fig. 3C) were generated using MOE 2016.08.30 Fig 3A indicates that the surface
of the binding pocket, where the methoxy groups of colchicine bind, to be electron deficient (blue
surface). This suggests that compounds with electron rich groups (red surface) will bind more
15

favorably and hence should afford potent inhibition. This is corroborated from the activities of 1,
3, 4 and 5, all of which, like colchicine, have a red surface in this area (Fig 3B and 3C).
Compound 6 which has a 4'-sulfone group makes the aryl ring highly electron deficient, and,
moreover, the sulfone group would probably lie in an area lined by the carbonyl backbones of
Cysβ241, Valβ248 and Aspβ251 (Fig. 3A), and this results in an unfavorable repulsive interaction.
This can explain, in part, the failure of 6 to dock in the colchicine site and consequently its low
activity. The poor activity of 2 (compared to 1) and of 7 and 8 (compared to 4) can also be
rationalized on the basis of the electrostatic surface of these compounds. The electron
withdrawing effect of the 4'-OH group in 2 and the 4'-NH2 in 8 (encircled blue surfaces in Fig. 3C)
causes an unfavorable interaction in an already electron deficient region of the binding site and
provides an explanation for the low activity of 2 and 8. The 4'-NHMe group in 7 however shows
both blue and red surfaces and shows a slightly reduced activity when compared to 4. Compounds
9 and 10 had electrostatic surfaces similar to that of compound 1 (not shown here). Electrostatic
complementarity has been discussed as an approach to improve compound selectivity,31 and,
based on the electrostatic surfaces observed for compounds 1-8, we propose that these surfaces
can be invoked to explain the differences in the biological activities of these compounds.
However, compounds 11-14 with the slightly different docking result as compound 1 (Supporting
information Figure S2), the electrostatic surfaces generated were similar to that observed for
compound 1 (Supporting information Figure S3), but could not explain the inactivity of these
compounds.
Table 3. Number of conformations generated by compounds 1-14 calculated using MOE2016.08.
Compound No. of conformations Compound No. of conformations
No.
within 7 kcal/mol
No.
within 7 kcal/mol
16

4
2
4
5
4
4
8
2
4
1
2
3
4
5
6
7
8
9
4
10
11
12
13
14
42
34
36
34
We also calculated the number of low energy conformations for 1-14 within 7 kcal/mol
(Table 3). Compounds 1-10 had a very low number of such conformations (2-8). In contrast,
compounds 11-14, owing to their increased flexibility, showed a 5 to 21-times higher number of
conformations. Biological evaluation of these compounds (Table 2) showed that compounds 11,
13 and 14 were devoid of any microtubule depolymerization activity, and 12 had poor MDA-MB-
435 inhibitory activity. The electrostatic surface area calculations were, like the docking studies,
unable to explain the inactivity of compounds 11-14. We therefore conclude that the increased
conformational flexibility at the C4 substitution in 11-14 is detrimental to cytotoxic effects of
these compounds, despite the docking studies and electrostatic surface area calculations that
suggest these compounds should have potent inhibitory activities.
Table 4. Inhibition of tubulin assembly and colchicine binding by CA-4 and 3, 4, 5, 9 and 10 and
their SlogP (log octanol/water partition coefficient calculated using MOE2016.08) values.
Inhibition of colchicine binding
Inhibition of
Compd.
tubulin assembly
% Inhibition  SD
SlogP
17

IC50 (M)  SD
5 µM inhibitor
92 ± 0.7
1 μM inhibitor
3.05
1.6 ± 0.1
70 ± 2
1
3.76
3.11
2.78
2.98
1.3 ± 0.1
1.2 ± 0.01
1.1 ± 0.06
1.2 ± 0.09
92 ± 3
85 ± 3
93 ± 3
85 ± 4
74 ± 5
65 ± 4
79 ± 4
64 ± 5
3
4
5
9
3.36
-
1.1 ± 0.007
1.3 ± 0.1
83 ± 4
59 ± 5
10
CA-4
98 ± 0.6
87 ± 0.2
The effects of 3, 4, 5, 9 and 10 on the polymerization of purified tubulin were evaluated
(Table 4). This allows a study of the direct interaction of these compounds with their intracellular
target. The ability of these compounds to bind to the colchicine site on tubulin was evaluated by
measuring inhibition of [3H]colchicine binding to tubulin (Table 4).32-35 The data show that 3, 4, 5,
9 and 10 are equipotent with 1 and CA-4 as tubulin assembly inhibitors (Table 4). At a 5 M
concentration, these compounds inhibited [3H]colchicine binding from 83 to 93%; while at a 1
concentration, they inhibited colchicine binding from 59 to 79%. Since the tubulin assembly
inhibitory values for these compounds were very similar, the improved activity of 3 in MDA-
MB-435 cell lines (Table 2) compared to 1 can be attributed, in part, to improved passive
diffusion of 3 (SlogP = 3.76) when compared to 1 (SlogP = 3.05). The 4'-SMe in 3 imparts more
hydrophobicity to the molecule and, therefore, better cellular inhibitory activity. Similarly, the
more potent activity of 3 as compared to 5 in the MDA-MB-435 cells (Table 2) can also be
18

ascribed to enhanced passive diffusion of 3 when compared to 5 (SlogP = 2.78), even though 5
has a better inhibitory value for tubulin assembly than 3. These data show that compounds 3, 4, 5,
9 and 10 primarily bind to the colchicine site of tubulin and with potency similar to that of the
standard CA-4.
Table 5. Compound activity in parental and Pgp expressing cell line
IC50 (nM) ± SD
Compd.
Pgp expressing SK-OV-3/ MDR1-
6/-6
Parental SK-OV-3
Rr
11.3 ± 0.2
7.2 ± 0.6
16.4 ± 0.4
9.3 ± 0.4
1.5
1.3
1.7
1.6
1.4
1.6
2.0
1.5
864
1
3
13.0 ± 1.0
12.6 ± 0.3
36.3 ± 3.3
18.7 ± 1.2
23.9 ± 2.5
4.5 ± 0.2
22.6 ± 2.3
19.9 ± 5
4
5
49.7 ± 7.9
29.8 ± 5.6
47.5 ± 2.4
6.6 ± 1.3
7
9
10
CA-4
Paclitaxel
3.0 ± 0.06
2600 ± 270
The abilities of the compounds to overcome Pgp-mediated drug resistance was evaluated in an isogenic cell
line pair using the SRB assay28. Cellular protein was measured using the SRB assay.
The ability to circumvent Pgp-mediated drug resistance was evaluated using an SK-OV-3
isogenic cell line pair (Table 5). Paclitaxel, a known Pgp substrate, was 864-fold less potent in cells
expressing the Pgp drug transporter (Resistance ratio, Rr = 864). As with 1 and CA-4, compounds 3,
4, 5, 7, 9 and 10 were only 1.3 to 2-fold less potent as Pgp-overexpressing cells as compared with
19

the parental cells. These data indicate that 3, 4, 5, 7, 9 and 10 are poor substrates for Pgp and
therefore would be better than some clinically used microtubule targeting drugs that are substrates
for transport by Pgp.
Table 6. Effects of compounds in parental and overexpressing βIII cells.
IC50 (nM) ± SD
Compd.
Wild type βIII expressing
Parental HeLa
Rr
HeLa
12.0 ± 1.2
7.0 ± 0.5
11.3 ± 0.7
6.8 ± 0.4
31.0 ± 1.6
16.0 ± 1.2
22.8 ± 1.0
4.7 ± 0.2
1.6 ± 0.2
9.6 ± 0.9
5.3 ± 0.8
9.8 ± 0.9
10.0 ± 1.3
19.1 ± 3.9
15.1 ± 0.7
22.6 ± 1.6
5.7 ± 0.4
7.7 ± 0.2
0.8
0.8
0.9
1.5
0.6
0.9
0.8
1.2
4.7
1
3
4
5
7
9
10
CA-4
Paclitaxel
Another clinically relevant drug resistance mechanism of tubulin-targeting agents is the
overexpressing of the III subtype of tubulin. The ability of compounds 3, 4, 5, 7, 9 and 10 to
overcome III-tubulin mediated drug resistance was examined using an isogenic HeLa cell line
pair (Table 6). All these compounds had similar resistance ratio (Rr) values that ranged from 0.6
to 1.5, similar to the value obtained with CA-4 (Rr = 1.2). These data suggested that these
20

compounds overcome III-tubulin mediated drug resistance compared with paclitaxel with a Rr
value of 4.7.
Table 7. Cancer cell inhibitory activity GI50 of 3.
Panel/Cell
line
GI50
Panel/
GI50
Panel/ Cell
line
GI50
Panel/ Cell
line
GI50
(nM)
Cell line
(nM)
(nM)
(nM)
Renal
Ovarian
cancer
Prostate
Cancer
NSCLC
Cancer
A549/ATC
C
<10
786 – 0
<10
IGROV1
<10
PC-3
<10
<10
EKVX
HOP-62
HOP-92
<10
<10
<10
A498
ACHN
CAKI-1
<10
<10
<10
OVCAR-3
OVCAR-4
<10
<10
DU-145
Breast Cancer
MCF7
OVCAR-5 >10000
<10
<10
MDA-MB-
231/ATCC
NCI-H226
NCI-H23
<10
<10
RXF 393
SN 12C
TK-10
<10
<10
OVCAR-8
<10
<10
<10
NCI/ADR-
RES
HS 578T
BT-549
<10
<10
<10
NCI-
>10000
<10
SK-OV-3
H322M
MDA-MB-
468
NCI-H460
UO-31
Melanoma
Colon
Cancer
COLO
NCI-H522
CNS
<10
LOX IMVI
MALME-
<10
Leukemia
<10
CCRF-CEM
<10
21

Cancer
SF-268
205
HCC-
2998
3M
<10
<10
M14
<10
HL-60
<10
MDA-MB-
435
SF-295
SF-539
SNB-19
<10
<10
<10
HCT-116
HCT-15
HT29
<10
<10
<10
<10
<10
K-562
MOLT-4
RPMI-8226
SR
<10
<10
<10
<10
SK-MEL-2
SK-MEL-
28
>10000
SNB-75
U251
<10
<10
KM12
<10
<10
SK-MEL-5
UACC-62
<10
<10
SW-620
Compound 3, the most potent microtubule depolymerizing agent in the current study, was
selected by the National Cancer Institute for evaluation in the NCI 60 cell line panel. It showed
potent GI50s against most of the NCI 60 cancer cell lines (Table 7). Compound 3 is equipotent
with paclitaxel across the panel and thus has the potential for further preclinical development and
would have advantages over paclitaxel in resistant tumors.
5. Metabolite Identification Study of 3
22

Figure 4. Identification the metabolites of 3 by LC/MS/MS. HPLC spectra of 3 incubated with
human liver microsomes at 0 min and 60 min. Metabolites M1-M5 and parent 3 were illustrated
on the spectra. Some peaks (retention times larger than compound 3) in 60 min sample have
different absorbance compared to that in 0 min sample, which may due to metabolic changes of
the microsomes initiated by compound 3.
Table 8. Summary of the metabolites identified by LC/MS/MS.
MS Peak Area in 60 min sample
RT
(min)
17.4
7.9
Metabolite
[M+H]+ (m/z)
286
Metabolite
Met ID
HPLC area
2,061,000
98,900
Percentage (%)
Parent
46.4
2.2
3
M1 (6)
318
sulfone
23

M2 (5)
M3
9.6
9.9
302
318
302
302
sulfoxide
Di-Hydroxy
Hydroxy
2,199,500
57,156
49.5
1.3
1.8
2.1
M4
15.5
16.0
81,795
M5
Hydroxy
92,546
In order to study its metabolic stability, 3 (Table 1) was incubated with human liver
microsomes (from both genders) for 60 min. Aliquots were examined by LC/MS/MS and the
results show that 49.5% of 3 was oxidized to 5 (M2) (Figure 4 and Table 8). The structure was
validated by MS/MS (Supporting information Figure S1). The sulfone (M1) and presumptive
hydroxyl (M4 and M5) and dihydroxyl (M3) metabolites are minor, each of which consist of less
than 2.2% of the overall metabolite pool. The MS 302.1306 ([M+H]+ of 5) (Supporting
information Figure S1) and MS/MS 239.1340 (mass of the demethylsulfinyl fragmentation
product of 5) clearly illustrated that M2 was 5. Similarly, MS 318.1272 ([M+H]+ of 6) and
MS/MS 239.1340 (mass of the demethylsulfonyl fragmentation product of 6) clearly proved that
M1 was 6. Hence, 5 was the primary metabolite, and 6 was a minor metabolite of 3.
24

6. Conformational study of 3
1
Figure 5. H NMR of N4-desmethyl compound 3a and compound 3. In 3a, the 5-position
hydrogens have a chemical shift 2.76 ppm. In compound 3, the 5-position hydrogens have a
chemical shift 1.84 ppm. This is the result of the shielding effect of the aniline moiety. (400 MHz,
solvent: DMSO-d6).
We36 reported that potent furo[2,3-d]pyrimidine based microtubule targeting agents have (in
DMSO-d6 solution) a stable conformation, in which the aniline ring resides above the furo ring. It
was of interest to determine the stable conformation of compound 3. We used N4-desmethyl
compound 3a as a comparison. The chemical shifts of the 5-position hydrogens are 2.76 ppm for
3a, and 1.84 ppm for 3 (Figure 5). The shielding effect in 1H NMR of the 5-position hydrogens in
3 demonstrated that in DMSO solution the aniline ring does reside above the cyclopenta ring.
25

This stable conformation of compound 3 was further validated by NOE signal between phenyl
protons and 5-CH2 in NOESY experiment (Supporting information Figure S4).
7. Summary
We bioisoterically replaced the oxygen in 4’-methoxy group in lead compound 1 (O-methyl),
which could be used to evaluate its hydrogen bonding effect. We designed and synthesized the S-
methyl analog 3 and the N-dimethyl analog 4, both of which demonstrated high antiproliferative
potency (MDA-MB-435 cells: IC50 = 4.6 nM for 3; IC50 = 8.2 nM for 4), strong antitubulin
activities and ability to overcome drug resistance mechanisms. Interestingly, in the cellular
microtubule depolymerization assays 3 was more potent than 4 and it was a more effective
inhibitor of colchicine binding at both 1 and 5 µM. A possible metabolite of lead compound 1, the
4’-hydroxyl analog 2 had low potency (MDA-MB-435 cells: IC50 = 405 nM). The liver
microsomal metabolite of 3, the methyl sulfinyl analog 5 however, also exhibited high
antiproliferative potency (MDA-MB-435 cells: IC50 = 7.9 nM), strong activities against tubulin,
and ability to overcome drug resistance mechanisms. Compound 3 emerged as the lead compound
for further preclinical study due to its high potency and its active metabolite. Probable
metabolites of compound 4, the 4’-mono methylamino analog 7 and the 4’-amino analog 8, were
less active (MDA-MB-435 cells: IC50 = 20 nM for 7; IC50 = 1223 nM for 8). Conformational
restriction of 4’-methoxy with a dihydrofuran in 9 or an ortho-methyl in 10 decreased their
potency (MDA-MB-435 cells: IC50 = 12.1 nM for 9; IC50 = 18.4 nM for 10). Homologation of the
4-anilino moiety, as in 12-14, resulted in decrease or loss of activities. In the absence of an x-ray
crystal structure, and assuming that the binding of these compounds did not induce any major
changes to the tubulin protein, the activities can be explained in part by molecular docking studies,
comparing the complementarity of electrostatic surface area of compounds to the colchicine site
on tubulin as well as conformational analysis.
8. Experimental Section
26

8.1 Synthesis
Analytical samples were dried in vacuum (0.2 mm Hg) in a CHEM-DRY drying apparatus
over P2O5 at 50 °C. Melting points were determined on a digital MEL-TEMP II melting point
apparatus with a FLUKE 51K/J electronic thermometer and are uncorrected. Nuclear magnetic
resonance spectra for protons (1H NMR) were recorded on a Bruker Avance II 400 (400 MHz) or
on a 500 (500 MHz) NMR system. The chemical shift values are expressed in ppm (parts per
million) relative to tetramethylsilane as an internal standard: s, singlet; d, doublet; t, triplet; q,
quartet; quint, quintet; m, multiplet; br, broad singlet. The coupling constants are measured by the
software MestreC. Thin-layer chromatography (TLC) was performed on Whatman Sil G/UV254
silica gel plates with a fluorescent indicator, and the spots were visualized under 254 and 365 nm
illumination. Proportions of solvents used for TLC are by volume. Column chromatography was
performed on a 230−400 mesh silica gel (Fisher Scientific) column. Elemental analyses were
performed by Atlantic Microlab, Inc., Norcross, GA. Elemental compositions are within ± 0.4%
of the calculated values and indicate > 95% purity of the compounds. Fractional moles of water
or organic solvents found in some analytical samples could not be prevented despite 24 - 48 h of
1
drying in vacuo and were confirmed where possible by their presence in the H NMR spectra.
HPLC-MS was analyzed with an Acquity system. A linear gradient of 90% of 0.1% formic acid
in water, 10% of 0.1% formic acid in ACN over 10 min, and then 100% ACN for 5 min. All final
compounds were > 95% pure as established by elemental analysis, HPLC, or both. Optical
rotation was measured on a Rudolph polarimeter. All solvents and chemicals were purchased
from Sigma-Aldrich or Fisher Scientific and were used as received.
8.1.1 4-Methylthio-N-methylphenylamine (19)
Compound 18 (0.695 g, 5 mmol) and NaH (60% dispersed in mineral oil, 0.12 g, 7.5 mmol)
o
were dissolved in THF at 0-5 C. To this solution CH3I (1.06 g, 7.5 mmol) was added, and the
27

o
mixture was stirred at 0-5 C for an additional 3 h. The reaction mixture was carefully quenched
with water and was diluted with EtOAc and water. The organic layer was separated and washed
with water (2 × 10 mL). After drying with anhydrous Na2SO4, the organic solvent was
concentrated in a rotary evaporator to less than 1 mL. This residue was placed on top of a silica
gel column and eluted with hexane and EtOAc. Fractions containing the product (TLC) were
pooled and evaporated to afford a yellow oil, compound 19, in 75% yield. TLC Rf 0.52 (hexanes:
1
EtOAc, 3:1); H NMR (CDCl3) 2.43 (s, 3H, NCH3), 2.85 (s, 3H, SCH3), 3.75 (br, 1H, exch,
NH), 6.58 (d, J = 8.51 Hz, 2H, ArH), 7.28 (d, J = 8.55 Hz, 2H, ArH); HPLC > 95%; ESIMS m/z
[M+H]+, calcd 154.06, found 154.06.
8.1.2 4-Methylaminophenyl methyl sulfoxide (20)
Compound 19 (0.507 g, 3 mmol) and mCPBA (0.564 g, 3.3 mmol) were dissolved in ACN at
o
0-5 C and stirred for 1 h. The reaction mixture was diluted with EtOAc and water. The organic
layer was separated and washed with water (2 × 10 mL). After drying with anhydrous Na2SO4,
the organic solvent was concentrated by a rotary evaporator to less than 1 mL. This residue was
placed on top of a silica gel column and eluted with CHCl3 and MeOH. Fractions containing the
product (TLC) were pooled and evaporated to afford compound 20 as a yellow oil in 67% yield.
TLC Rf 0.35 (CHCl3: MeOH, 20:1); 1H NMR (CDCl3) 2.71 (s, 3H, NCH3), 2.90 (d, J = 4.79 Hz,
3H, SCH3), 4.15 (m, 1H, exch, NH), 6.69 (d, J = 8.66 Hz, 2H, ArH), 7.50 (d, J = 8.64 Hz, 2H,
ArH); ESIMS m/z [M+H]+ , calcd 170.06, found 170.00.
8.1.3 4-Methylaminophenyl methyl sulfone (21)
Compound 20 (0.169 g, 1 mmol) and mCPBA (0.376 g, 2.2 mmol) were dissolved in ACN at
o
0-5 C and stirred for 1 h. The reaction mixture was diluted with EtOAc and water. The organic
layer was separated and washed with water (2 × 10 mL). After drying with anhydrous Na2SO4,
28

the organic solvent was concentrated by a rotary evaporator to less than 1 mL. This residue was
placed on top of a silica gel column and eluted with hexane and EtOAc. Fractions containing the
product (TLC) were pooled and evaporated to afford compound 21 as an off-white solid in 59%
o
1
yield. mp 127.7-129.0 C; TLC Rf 0.37 (hexanes: EtOAc, 3:1); H NMR (CDCl3) 2.93 (d, J =
4.79 Hz, 3H, SCH3), 3.03 (s, 3H, NCH3), 4.34 (m, 1H, exch, NH), 6.64 (d, J = 8.78 Hz, 2H, ArH),
7.74 (d, J = 8.77 Hz, 2H, ArH); HPLC > 95%; ESIMS m/z [M+H]+, calcd 186.05, found 186.13.
8.1.4 N-Methyl-2,3-dihydrobenzofuro-5-amine (23)
Compound 23 was synthesized from 22, following the procedure of 19, and it was obtained as
an off-white solid in 68% yield. The color of this compound slowly turned dark when exposed to
air, and it was directly used for the next step. mp 118.3-121.3 oC; TLC Rf 0.47 (hexanes: EtOAc,
3:1); 1H NMR (CDCl3) 2.82 (s, 3H, NCH3), 3.18 (t, J = 8.56 Hz, 2H, CH2), 3.43 (br, 1H, exch,
NH), 4.50 (t, J = 8.58 Hz, 2H, CH2), 6.42 (dd, J = 2.47 Hz, 8.43 Hz, 1H, ArH), 6.58 (d, J = 2.25
Hz, 1H, ArH), 6.67 (d, J = 8.43 Hz, 1H, ArH); ESIMS m/z [M+H]+, calcd 150.08, found 150.09.
8.1.5 N,3-Dimethyl-4-methoxyaniline (25)
Compound 25 was synthesized from 24, following the procedure of 19, and it was obtained as
an off-white solid in 65% yield. The color of this compound slowly turned dark when exposed to
air, and it was directly used for the next step. mp 109.6-111.3 oC; TLC Rf 0.44 (hexanes: EtOAc,
3:1); 1H NMR (CDCl3) 2.22 (s, 3H, CH3), 2.83 (s, 3H, NCH3), 3.39 (br, 1H, exch, NH), 3.79 (s,
3H, OCH3), 6.47 (dd, J = 8.50 Hz, 2.43 Hz, 1H, ArH), 6.52 (d, J = 2.13 Hz, 1H, ArH), 6.74 (d, J
= 8.51 Hz, 1H, ArH); ESIMS m/z [M+H]+, calcd 152.11, found 152.17.
8.1.6 N,N'-(1,4-Diphenylene)diformamide (27a)
29