A conformational and structure-activity relationship study of cytotoxic 3,5-bis(arylidene)-4-piperidones and related N-acryloyl analogues

Article abstract of DOI:10.1021/jm0002580

A series of 3,5-bis(arylidene)-4-piperidones 1 and related N-acryloyl analogues 2 were prepared as candidate cytotoxic agents with a view to discerning those structural features which contributed to bioactivity. A number of the compounds were markedly cytotoxic toward murine P388 and L1210 leukemic cells and also to human Molt 4/C8 and CEM neoplasms. Approximately 40% of the IC₅₀ values generated were lower than the figures obtained for melphalan. In virtually all cases, the N-acyl compounds were significantly more bioactive than the analogues 1. In general, structure - activity relationships revealed that the cytotoxicity of series 1 was correlated positively with the size of the aryl substituents, while in series 2, a -σ relationship was established. In particular, various angles and interatomic distances were obtained by molecular modeling, and the presence of an acryloyl group on the piperidyl nitrogen atom in series 2 affected the relative locations of the two aryl rings. This observation, along with some differences in distances between various atoms in series 1 and 2, may have contributed to the disparity in cytotoxicity between 1 and 2. The results obtained by X-ray crystallography of representative compounds were mainly in accordance with the observations noted by molecular modeling. Selected compounds interfered with the biosynthesis of DNA, RNA, and protein in murine L1210 cells, while others were shown to cause apoptosis in the human Jurkat leukemic cell line. This study has revealed the potential of these molecules for development as cytotoxic and anticancer agents.

Full text of DOI:10.1021/jm0002580

586
J . Med. Chem. 2001, 44, 586-593
A Con for m a tion a l a n d Str u ctu r e-Activity Rela tion sh ip Stu d y of Cytotoxic
3,5-Bis(a r ylid en e)-4-p ip er id on es a n d Rela ted N-Acr yloyl An a logu es
J onathan R. Dimmock,*,† Maniyan P. Padmanilayam,† Ramanan N. Puthucode,† Adil J . Nazarali,†
Narasimhan L. Motaganahalli,† Gordon A. Zello,† J . Wilson Quail,‡ Eliud O. Oloo,‡ Heinz-Bernhard Kraatz,‡
J ared S. Prisciak,§ Theresa M. Allen, Cheryl L. Santos, J an Balzarini,| Erik De Clercq,| and
Elias K. Manavathu4
Departments of Chemistry and Biochemistry, College of Pharmacy and Nutrition, University of Saskatchewan, Saskatoon,
Saskatchewan, S7N 5C9 Canada, Department of Pharmacology, University of Alberta, Edmonton, Alberta, T6G 2H7 Canada,
Rega Institute for Medical Research, Katholieke Universiteit Leuven, B-3000 Leuven, Belgium, and Department of Medicine,
Wayne State University, Detroit, Michigan 48201-1908
Received J une 16, 2000
A series of 3,5-bis(arylidene)-4-piperidones 1 and related N-acryloyl analogues 2 were prepared
as candidate cytotoxic agents with a view to discerning those structural features which
contributed to bioactivity. A number of the compounds were markedly cytotoxic toward murine
P388 and L1210 leukemic cells and also to human Molt 4/C8 and CEM neoplasms. Ap-
proximately 40% of the IC50 values generated were lower than the figures obtained for
melphalan. In virtually all cases, the N-acyl compounds were significantly more bioactive than
the analogues 1. In general, structure-activity relationships revealed that the cytotoxicity of
series 1 was correlated positively with the size of the aryl substituents, while in series 2, a -σ
relationship was established. In particular, various angles and interatomic distances were
obtained by molecular modeling, and the presence of an acryloyl group on the piperidyl nitrogen
atom in series 2 affected the relative locations of the two aryl rings. This observation, along
with some differences in distances between various atoms in series 1 and 2, may have
contributed to the disparity in cytotoxicity between 1 and 2. The results obtained by X-ray
crystallography of representative compounds were mainly in accordance with the observations
noted by molecular modeling. Selected compounds interfered with the biosynthesis of DNA,
RNA, and protein in murine L1210 cells, while others were shown to cause apoptosis in the
human J urkat leukemic cell line. This study has revealed the potential of these molecules for
development as cytotoxic and anticancer agents.
In tr od u ction
A number of tumors have lower pH than correspond-
ing normal cells.8 Hence Mannich bases may display
selective toxicity to such neoplasms since there will be
a higher percentage of the molecules in the ionized form
in the tumors (and hence present as the more reactive
species) than is the case with the corresponding R,â-
unsaturated ketones. Such considerations led to the
decision a number of years ago to prepare 1a which may
be considered a Mannich base of a dienone. This
compound showed high activity toward the murine
lymphocytic leukemia P388/MRI cell line and did not
bind to a synthetic DNA, poly [d(AT)].9 Acyclic Mannich
bases of conjugated styryl ketones generally caused
mortality when administered to mice using doses of 200
mg/kg or more.7,10 However, five consecutive daily doses
of 240 mg/kg of 1a did not induce fatalities.9 The
compound therefore is markedly cytotoxic but has low
murine toxicity. Thus the preparation of a number of
analogues of 1a was planned. To discern whether there
was a correlation between the electronic and/or hydro-
phobic properties of the aryl substituents and cytotox-
icity, groups with divergent Hammett σ and Hansch π
values were placed in the aryl rings. In fact, substitu-
ents from three of the four quadrants of a Craig plot11
were utilized.
A number of R,â-unsaturated ketones display cyto-
toxic and anticancer properties.1,2 The rationale for their
synthesis and antineoplastic evaluation is based in part
on their predicted mode of action, i.e., interaction with
cellular thiols with little or no affinity for hydroxy and
amino groups found in nucleic acids.3,4 Thus develop-
ment of these compounds as candidate cytotoxics may
lead to drugs which are bereft of the genotoxic properties
present in certain antineoplastic agents.5 Conversion of
certain conjugated enones into the corresponding Man-
nich bases led to significant increases in both the rates
of thiol alkylation6 and cytotoxicity.7 The increased rate
of thiolation may be due to the protonated Mannich
bases stabilizing the reaction intermediates formed
between thiols and the R,â-unsaturated keto group to a
significantly greater extent than occurs with the precur-
sor enones.
* Corresponding author. Address: College of Pharmacy and Nutri-
tion, University of Saskatchewan, 110 Science Place, Thorvaldson
Building, Room 216, Saskatoon, Saskatchewan, S7N 5C9 Canada.
Phone: (306) 966-6331. Fax: (306) 966-6377. E-mail: dimmock@
skyway.usask.ca.
† College of Pharmacy and Nutrition, University of Saskatchewan.
‡ Department of Chemistry, University of Saskatchewan.
§ Department of Biochemistry, University of Saskatchewan.
University of Alberta.
There was, however, a further reason for the prepara-
tion of series 1. The theory of sequential cytotoxicity
| Katholieke Universiteit Leuven.
4 Wayne State University.
10.1021/jm0002580 CCC: $20.00 © 2001 American Chemical Society
Published on Web 01/20/2001

SAR Study of Cytotoxic 3,5-Bis(arylidene)-4-piperidones
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4 587
Sch em e 1. Syntheses of Series 1 and 2a
cell lines have been claimed to be predictors of clinically
useful anticancer drugs.13 In addition, evaluation using
human Molt 4/C8 and CEM T-lymphocytes was under-
taken in order to ascertain whether the compounds were
cytotoxic to human neoplastic cells. Cytotoxicity was
undertaken using concentrations of compounds up to
and including 50 µM in the case of the P388 screen and
500 µM for the remaining cell lines. The results are
summarized in Table 1.
In general, the compounds displayed marked cyto-
toxicity toward the cell lines. However, the IC50 figures
for 2g in the L1210 and CEM screens were not obtained,
and clearly this compound is an outlier with markedly
lower activity than the analogues 2a -f. Consequently,
further discussion pertaining to structure-activity re-
lationships will revolve around 2a -f and the analogues
in series 1, namely 1a -f. The average IC50 figures for
1a -f in the P388, L1210, Molt 4/C8, and CEM screens
were 0.81, 81.9, 59.4, and 33.6 µM, respectively, while
the comparable figures for 2a -f were 0.28, 4.71, 0.96,
and 1.13 µM, respectively. These data reveal that P388
cells are the most sensitive to these compounds in
contrast to L1210 cells which are the most refractory.
The huge differential in the activity of these compounds
toward the two murine cell lines may be due, in part at
least, to the fact that P388 cells grow far more slowly
than L1210 cells. The results for 1a -f and 2a -f reveal
that 44% of the IC50 figures presented in Table 1 were
lower than melphalan, revealing the importance of these
compounds as prototypic molecules from which further
development should take place. The average IC50 figures
for all four cell lines for 1a -f and 2a -f were 44.0 and
1.77 µM, respectively, revealing a 25-fold increase in
cytotoxicity for the N-acryloyl analogues which affords
support for the theory of sequential cytotoxicity. To
obtain an overview of the cytotoxicity of the compounds,
average potency (AP) figures derived from all four cell
lines were calculated. Since these data were derived
from different cell lines of both murine and human
origin, they should be viewed with caution; nevertheless
they enable one to observe rapidly the relative cytotox-
icity of different compounds. These figures are presented
in Table 1.
a
Reagents: (i) HCl/CH3COOH followed by treatment with
K2CO3; (ii) CH2dCHCOCl/K2CO3. The letters a -g indicate the
following substituent pattern: a : R1 ) R2 ) H; b: R1 ) Cl, R2
)
H; c: R1 ) R2 ) Cl; d : R1 ) F, R2 ) H; e: R1 ) NO2, R2 ) H; f:
R1 ) OCH3, R2 ) H; g: R1 ) N(CH2)2, R2 ) H.
states that successive chemical attacks may lead to
greater damage in cancer cells compared to normal
cells.12 The evidence marshalled to support this view-
point has been presented at length.12 Thus, in the case
of the compounds in series 1, initial reaction with a thiol
such as glutathione would lead to the formation of the
corresponding 3-arylglutathionylmethyl-5-arylidene-4-
piperidone which is capable of further reaction with
cellular nucleophiles, i.e., one would predict successive
reactions in vitro and in vivo. To evaluate the hypothesis
further, the formation of the N-acryloyl analogues 2 was
considered. In these molecules, an additional site for
electrophilic attack of biomacromolecules is present. If
the hypothesis of sequential cytotoxicity is valid, the
compounds in series 2 should be appreciably more
cytotoxic than the corresponding secondary amines 1.
The objective of the present investigation therefore was
to prepare two series of prototypic molecules 1 and 2
with a view to discerning the structural features con-
tributing to cytotoxicity and to examining further the
viability of the theory of sequential cytotoxicity.
To discern whether correlations existed between the
Hammett σ, Hansch π, and molecular refractivity (MR)
figures of the aryl substituents and cytotoxicity, the
decision was made to construct linear and semiloga-
rithmic plots between these physicochemical constants
and the IC50 values of 1a -f and 2a -f against each of
the four cell lines as well as the AP figures. These types
of regression analyses have been used by us previously.2
Logarithmic transformations were performed on the
data in an attempt to stabilize the variances and to
linearize the relationships. The correlations were clas-
sified as highly significant (p < 0.05), significant (p <
0.1), or as a trend toward a significant correlation (p <
0.15). In the case of compounds 1a -f, semilogarithmic
plots revealed the following relationships in which the
p values were less than 0.15, namely between the IC50
values and the aryl σ figures for the P388 cells (p < 0.15)
and also between the MR constants using L1210 cells
(p < 0.1), Molt 4/C8 cells (p < 0.15), and CEM cells (p
< 0.15) as well as the AP values (p < 0.15). Both linear
and semilogarithmic plots between the cytotoxicity of
Resu lts a n d Discu ssion
The compounds in series 1 and 2 were prepared by a
procedure outlined in Scheme 1. A Claisen-Schmidt
condensation between 4-piperidone hydrochloride and
the appropriate aryl aldehyde led to the formation of
the 3,5-bis(arylidene)-4-piperidones 1. Acylation of these
enones with acryloyl chloride gave rise to the corre-
1
sponding N-acyl derivatives 2. H NMR spectroscopy
revealed the presence of a single peak for the methine
protons adjacent to the aryl rings, indicating the ste-
reohomogeneity of these molecules. Since X-ray crystal-
lography of 1b,e and 2b,d (vide infra) revealed these
molecules to possess the E configuration, which is
consistent with previous reports for this cluster of
compounds,9 the assumption was made that all mem-
bers of series 1 and 2 had the E stereochemistry.
The 4-piperidones in series 1 and 2 were examined
against murine P388 and L1210 neoplasms since these

588 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4
Dimmock et al.
Ta ble 1. Cytotoxicity of 1a -g and 2a -g against Murine P388 and L1210 Cells and Also Human Molt 4/C8 and CEM T-Lymphocytes
IC50 (µM)
P388
cells
L1210
cells
Molt 4/C8
cells
CEM
cells
compd
1a
1b
1c
1d
1e
1f
1g
2a
2b
2c
2d
2e
2f
2g
APa
3.03
16.1
47.8
11.0
11.5
174
157
3.30
0.76
0.92
1.34
0.22
4.36
0.77 ( 0.02
0.74 ( 0.07
1.01 ( 0.06
0.60 ( 0.01
0.20 ( 0.01
1.54 ( 0.3
0.64 ( 0.06
0.50 ( 0.06
0.15 ( 0.02
0.05 ( 0.007
0.27 ( 0.01
0.05 ( 0.001
0.68 ( 0.05
43.29 ( 3.5
0.22 ( 0.01
7.96 ( 0.11
41.5 ( 0.3
129 ( 80
1.67 ( 0.15
13.4 ( 4.0
40.2 ( 1.3
5.00 ( 1.20
8.28 ( 0.75
288 ( 8
1.70 ( 0.02
8.63 ( 0.48
20.8 ( 16.0
2.05 ( 0.36
4.47 ( 2.28
164 ( 104
162 ( 19
36.2 ( 0.1
32.9 ( 4.2
244 ( 43
240 ( 92
226 ( 18
8.69 ( 0.73
1.96 ( 0.08
1.52 ( 0.02
3.89 ( 0.18
0.42 ( 0.07
11.8 ( 0.7
>500
1.42 ( 0.27
0.44 ( 0.06
1.04 ( 0.76
0.41 ( 0.24
0.15 ( 0.06
2.27 ( 0.17
369 ( 185
3.24 ( 0.56
1.48 ( 0.34
0.49 ( 0.23
1.08 ( 0.82
0.80 ( 0.46
0.26 ( 0.02
2.67 ( 0.32
>500
>353
melphalan
2.13 ( 0.02
2.47 ( 0.21
2.02
a
Average potency of all four cell lines.
2a -f and the Hammett σ values revealed the following
results: P388 cells, p < 0.05; L1210 cells, p < 0.05; Molt
4/C8 cells, p < 0.15; CEM cells, (p < 0.1); AP values, p
< 0.05. In all of these plots, the r values were negative
when σ values were employed and positive when the
MR constants were considered. All other plots led to p
values in excess of 0.15.
The following conclusions may be drawn from the
structure-activity relationships (SAR) study. First, in
the case of the N-acyl analogues 2a -f, electronic
parameters are clearly the most important factor influ-
encing cytotoxicity. Thus future expansion of series 2
should bear in mind this correlation. These results are
similar to a recent study with cytotoxic 2-arylideneben-
zosuberones wherein cytotoxicity was negatively cor-
related with the aryl σ values using P388 and L1210
cells and positively correlated with the MR constants
in all of the same four cell lines employed in this report.2
However, the negative correlation between cytotoxicity
and the Hammett σ values was counterintuitive since
one would predict that, by decreasing the fractional
positive charge on the arylidene methine carbon atom,
cytotoxicity would decrease rather than be elevated.
Thus interaction with cellular thiols may not be the
predominant mechanism mediating toxicity. Second, the
lipophilicity of the molecules appears to exert little or
no influence on the antineoplastic activity of the com-
pounds 1a -f and 2a -f.
F igu r e 1. Numbering of the compounds in series 1 and 2 and
designation of the θ1 and θ2 torsion angles and interatomic
distances d1 and d2.
F igu r e 2. Indication of the interatomic distances d3 and d4
and angle ψ.
F igu r e 3. Denotation of the interatomic distances d5-d8.
In addition to evaluating the contributions of the aryl
substituents to bioactivity, the topology of the molecules
was considered. This segment of the work was under-
taken principally to determine those structural features
which contribute to the marked disparity in cytotoxicity
between the compounds in series 1 and 2. The shapes
of 1a -g and 2a -g were generated by molecular model-
ing, and the measurements obtained are illustrated in
Figures 1-3. The torsion angles θ1 and θ2 were formed
between rings A and B and the adjacent olefinic link-
ages, respectively, while the width and length of the
molecules were reflected in the d1 and d2 measurements,
respectively. To obtain the relative positions of the aryl
rings, the distances d3 between the centers of the rings
were determined. Axis 1 was the plane in which ring A
was located. The plane of ring B was noted, and the
distances d4 and angles ψ between the two planes were
obtained. The results of these determinations are sum-
marized in Table 2.
The percentage increases in the torsion angles θ1 and
θ2 in series 2, compared to the analogues in series 1,
were 8 and 24, respectively. Thus a contributing factor
for the greater cytotoxicity of the N-acyl analogues may
be the elevated θ2 figures and to a less extent the θ1
values found in 2. The orientations of the aryl rings in
2 may permit more favorable alignments of the rings
with a critical biomacromolecule; for example, a better
fit into a cleft on a bonding site may occur in the
compounds 2 than is the case with series 1.
The data in Table 2 revealed that the distances d1-
d3 in series 1 and 2 were similar; the greatest disparity
of 9% was noted in the case of the d2 values. It is
possible therefore that the lengths of the molecules may

SAR Study of Cytotoxic 3,5-Bis(arylidene)-4-piperidones
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4 589
Ta ble 2. Certain Angles and Distances of Compounds 1a -g and 2a -g Determined by Molecular Modeling
angles (deg)
interatomic distances (Å)
compd
θ1
θ2
ψ
d1
d2
d3
d4
d5
d6
d7
d8
1a -g
-49.5
( 0.05
-53.2
( 0.17
49.4
( 0.11
61.0
37.0
( 0.08
97.3
16.80
( 2.78
16.68
5.03
( 0.70
5.47
10.23
( 0.004
10.22
3.25
( 0.01
3.73
2.39
( 0.005
2.44
2.39
( 0.005
2.48
2a -g
2.15
( 0.005
2.42
( 0.001
( 0.24
( 0.24
( 2.83
( 0.26
( 0.005
( 0.09
( 0.012
( 0.014
F igu r e 4. ORTEP diagram of 2b.
Ta ble 3. Some Torsion Angles and Distances of 1e and 2b,d Determined by X-ray Crystallographya
angles (deg)
interatomic distances (Å)
compd
θ1
θ2
ψ
d1
d2
d3
d4
d5
d6
d7
d8
AP
1b
18.6
(-49.5)
19.6
(-49.5)
11.4
(61.3)
20.0
(-53.2)
-15.6
(49.4)
-23.0
(49.4)
-27.9
(53.2)
-19.3
(-61.1)
7.91
(36.99)
1.59
(36.96)
6.21
(97.59)
0.83
(97.43)
16.46
(15.92)
17.44
(17.15)
16.42
(15.88)
15.63
(15.17)
5.73
(4.99)
4.62
(4.64)
6.13
(5.38)
7.49
(5.39)
10.08
(10.23)
10.32
(10.23)
10.36
(10.21)
10.34
(10.22)
1.39
(3.25)
0.29
(3.24)
1.13
(3.65)
0.15
(3.81)
2.14
(2.39)
2.28
(2.39)
2.14
(2.44)
2.17
(2.44)
2.16
(2.39)
2.18
(2.39)
2.11
(2.48)
2.11
(2.48)
16.1
1e
2b
2d
11.5
0.76
1.34
1.91
(2.15)
1.89
2.29
(2.42)
2.26
(2.16)
(2.43)
a
Figures in parentheses were obtained by molecular modeling.
make only a minor contribution to the variation in
cytotoxicity between 1 and 2. On the other hand, ring
B in series 2 was located further away from axis 1 as
revealed by a 15% increase in the d4 figures in the
N-acyl compounds. This observation was further il-
lustrated in the 163% increase in the ψ figures of series
2 compared to the analogues in series 1.
One may summarize the molecular modeling data by
stating that the acyl group attached to the N1 atom in
series 2 has a significant effect on the spatial orientation
and location of aryl ring B which may well contribute
markedly to the greater cytotoxicity of series 2 compared
to series 1. Thus expansion of series 2, in which groups
of varying sizes are attached to the N1 atom, may shed
further light on the structural features contributing to
the cytotoxicity in this novel series of candidate anti-
neoplastic agents.
X-ray crystallography was undertaken on representa-
tive compounds in series 1 and 2 with a view to
discerning whether the observations pertaining to the
shapes of compounds 1 and 2, determined by molecular
modeling, were corroborated or not using a second
physicochemical approach. Data were obtained for 1b
and 1e,which crystallized in the protonated form with
three molecules of acetic acid, and 2b,d . An ORTEP
diagram of 2b is portrayed in Figure 4. In all four
compounds the E configuration of the chiral axes was
revealed, and the piperidine ring assumed the half chair
conformation.
The distances d5-d8 were measured with a view to
explaining the differences between the θ1, θ2, ψ, and d4
figures of the piperidones 1 and the corresponding
amides 2. The d5 and d7 values in series 2 will indicate
whether nonbonded interactions between the vinyl
proton and oxygen atom, attached to carbon atoms 22
and 21, respectively, interact with the equatorial pro-
tons at positions 2 and 6 of the piperidine ring. Second,
the d6 and d8 distances may reveal whether steric
impedance occurs between the ortho protons of the aryl
rings and the 2 He and 6 He atoms. In general, an
interatomic distance between the centers of two atoms
which is 2.5 Å or less is an indication of a nonbonded
interaction. If the increased figures of θ1, θ2, ψ, and d4
in series 2 were due to nonbonded interactions, as
indicated in Figure 3, then the interatomic distances
d5 and d6 would be predicted to be shorter than d7 and
d8, respectively, i.e., d5 < d7 in series 2 and d6 < d8 in
both series 1 and 2. The figures in Table 2 reveal that
only the first postulate was validated, i.e., d5 < d7, and
thus additional factors account for the differences in
orientation and location of rings A and B.
A review of the data in Table 3 led to the following
observations. First, the θ1 and θ2 angles, although
smaller than those recorded using molecular modeling,
confirmed the lack of coplanarity of the rings A and B
with the adjacent dienone structure. In general, the θ2
> θ1 relationship was noted. Second, in the crystal state,
however, the aryl rings were virtually in the same plane

590 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4
Ta ble 4. Peak Potentials of 1a -d ,f,g and 2b,d ,g
Dimmock et al.
Ta ble 6. Cytotoxicity, Apoptotic Indices, and Inhibition of
RNA and Protein Biosynthesis of 1c and 2c Using Human
J urkat T Cells
Ep
Ep
Ep
compd
(mV)a
compd
(mV)a
compd
(mV)a
% inhibition of synthesis of
IC50 IC80 apoptotic
(µM) (µM)
1a
1b
1c
1208
1230
1345
1d
1f
1g
1323
1270
1064
2b
2d
2g
1134
1180
970
compd
indexa
RNAa
proteina
1c
2c
3.09 4.83
0.58 0.94
48.3
63.5
63.2b
26.4 ( 9.48 31.1 ( 6.09
67.6 ( 1.55 58.1 ( 4.52
85.1b ( 4.71 17.3 ( 4.94
A scan rate of 100 mV s-1 was used.
a
melphalan 2.20b 3.52b
Ta ble 5. Inhibitory Effects of 1e,g and 2e on the Biosyntheses
of DNA, RNA, and Protein in Murine L1210 Cells
a
IC80 concentrations were used in these determinations.
b
Reprinted in part from J . Med. Chem. 1999, 42, 1363. Copyright
1999 American Chemical Society.
IC50 (µM)
compd
1e
1g
2e
DNA
228
g500
2.77
>100
RNA
protein
L1210 cells
cells than 1e. In addition, the greater cytotoxicity of 1e
than that of 1g to L1210 cells was reflected in the lower
IC50 values of 1e toward DNA, RNA, and protein
syntheses. The data in Table 5 reveal that the useful
lead molecule 2e acts in a different way to the estab-
lished anticancer drug melphalan whose principal mode
of action is direct interaction with DNA16 and, from the
results portrayed in Table 5, apparently does not
interfere markedly with the syntheses of DNA, RNA,
and proteins.
276
336
5.73
>100
214
320
3.21
94.6
32.9
240
0.42
2.13
melphalan
as revealed by the small ψ and d4 figures. Third, the
d5-d8 measurements revealed extensive nonbonded
interactions. In concert with the molecular modeling
data, the distance d5 was shorter than d7 while the d6
and d8 measurements were virtually identical. There-
fore, in general, the evidence of X-ray crystallography
on selected molecules was in accord with the topological
studies conducted by molecular modeling.
A number of anticancer drugs induce apoptosis,17 and
the question posed was whether representative com-
pounds in series 1 and 2 acted by this mechanism. A
previous study revealed that human J urkat T cells were
more sensitive to an R,â-unsaturated ketone than other
cell lines,18 and thus this neoplasm was employed in the
present investigation. The data in Table 6 revealed that
both 1c and 2c were cytotoxic to this cell line. To ensure
that death occurred in the majority of cells, IC80 figures
were generated, and using these concentrations both 1c
and 2c caused apoptosis. The apoptotic index [(number
of cells with apoptotic nuclei/number of cells counted)
× 100] of 2c was only 32% higher than 1c, indicating
that additional factors accounted for the 5-fold greater
cytotoxicity of 2c over 1c toward J urkat cells. In concert
with the results recorded in Table 5, both 1c and 2c
inhibited RNA and protein syntheses and greater activ-
ity was displayed by 2c.
The results from the investigations of representative
compounds summarized in Tables 5 and 6 revealed that
these molecules inhibit macromolecular syntheses and
caused apoptosis. In addition, the compounds in series
2 were more potent in causing these effects than the
analogues in series 1, affording some indication of the
reasons for the greater cytotoxicity of the N-acyl con-
geners. One may speculate from these biodata that there
are multiple sites of action of the 4-piperidones. If
alkylation is the predominant cause of cytotoxicity, then
of relevance is the fact that alkylating agents employed
in cancer chemotherapy also have multiple sites of
action.19
A correlation has been established previously between
the cytotoxicity of various conjugated enones and their
redox potentials.14 In the present investigation, at-
tempts were made to measure this physicochemical
parameter of all of the compounds in series 1 and 2.
However, oxidation of 2a ,c,e was not achieved, and the
scans for 1e and 2f in the region of 1500 and 1700 mV,
respectively, were extremely broad and definitive oxida-
tion potentials could not be ascertained. All other
systems exhibited an irreversible one electron oxidation
at a scan speed of 100 mV s-1. The results for the
remaining compounds are presented in Table 4. Both
linear and semilogarithmic plots between the Ep figures
of 1a -d ,f,g and the Hammett σ values showed positive
correlations (p < 0.05). However, no relationships were
noted between the redox potentials of these compounds
and the IC50 figures using either the four cell lines or
the AP figures (p > 0.15). Thus the cytotoxicity of the
compounds in series 1 was not related to their redox
potentials. In the case of 2b,d ,g, there was no correla-
tion between the Ep and σ values (p >0.15). Since 2g
has markedly divergent activity against the four cell
lines than the analogues and hence eliminated from
further consideration, there was an insufficient number
of compounds in series 2 to evaluate whether the
cytotoxicity of these compounds correlated with their
redox potentials.
An initial investigation into the possible manner
whereby these compounds exerted their cytotoxicity was
undertaken with special reference to discerning why
such a marked disparity in bioactivity was noted
between the compounds in series 1 and 2. A number of
anticancer drugs inhibit the biosyntheses of DNA, RNA,
and proteins.15 The data in Table 5 indicate the far
greater inhibitory effect of 2e on macromolecular syn-
theses than 1e. In fact, the average IC50 figures against
DNA, RNA, and protein syntheses of 1e and 2e are 239
and 3.90 µM, respectively, indicating that 2e is 61 times
more potent than 1e. These results may partially
explain the 78-fold greater activity of 2e toward L1210
Recently the 1,3,4-trisubstituted piperidine 3 pre-
pared in our laboratory was discovered to display
activity toward a wide range of pathogenic fungi,

SAR Study of Cytotoxic 3,5-Bis(arylidene)-4-piperidones
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4 591
CAD-4 diffractometer was used for the collection of X-ray
crystallographic data. TLC was undertaken using silica gel
plastic-backed sheets and a solvent system of 2% methanol in
chloroform. Compounds 2e and 2g were obtained with 0.5 and
0.25 molecules of water of crystallization, respectively.
including a large number of isolates of Aspergillus
fumigatus and Candida albicans.20 Compound 3, which
contains a piperidine ring and two substituted styryl
groups, one of which is conjugated to a carbonyl func-
tion, resembles the structures of the molecules in series
1 and 2. Since there is an urgent need for novel
antifungal agents different in structures and modes of
action from currently available medication, representa-
tive compounds in series 1 and 2 were evaluated against
three isolates of Aspergillus fumigatus and one isolate
of Candida albicans. The average minimum inhibitory
concentration (MIC) figures for 3 and an investigational
azole voriconazole were 3.9 and 1.95 µM, respectively.
However the MIC values of 1b,c,e and 2b,c,e were all
above 500 µM in the four assays. Thus in reference to
the screening procedures used in this investigation, the
compounds displayed significant cytotoxicity toward
both murine and human neoplasms yet were virtually
bereft of antifungal activity. It is conceivable therefore
that the compounds in series 1 and 2 are not general
biocidal agents but target malignant cells, although a
great deal of further study is required in order to
evaluate the extent that these compounds display
toxicity to cancers.
Syn th esis of Ser ies 1. The appropriate aryl aldehyde
(26.71 mmol) was added to a suspension of 4-piperidone
hydrochloride monohydrate (13.03 mmol) in acetic acid (35
mL). Dry hydrogen chloride was passed through this mixture
for 0.5 h during which time a clear solution was obtained. After
stirring at room temperature for 24 h, the precipitate was
collected and added to a mixture of a saturated aqueous
potassium carbonate solution (25% w/v, 25 mL) and acetone
(25 mL); the resultant mixture was stirred for 0.5 h. The free
base was collected, washed with water (50 mL), and dried. The
compounds were recrystallized from 95% ethanol (1a ,e),
methanol (1b-d ,f), or dimethylformamide (1g). The melting
points (°C) and yields (%) of 1a -g were as follows: 1a : 177-
178, 87; 1b: 201, 82; 1c: 196-197, 84; 1d : 212-213, 75; 1e:
205-206, 68; 1f: 210-211, 100; 1g: 238-240, 72. The 1H
NMR spectrum (DMSO-d6) of a representative compound 1a
was as follows: δ 4.21-4.28 (s, 4H, piperidyl H), 7.42-7.52
(m, 10H, aryl H), 7.74 (s, 2H, arylidene H). The NH absorptions
were not observed for most of the compounds in series 1.
Syn th esis of Ser ies 2. Acryloyl chloride (8.84 mmol) was
added to a suspension of the appropriate 3,5-bis(arylidene)-
4-piperidone (5.75 mmol), acetone (10 mL), and a solution of
potassium carbonate (29 mmol) in water (10 mL) which was
cooled externally with an ice bath. After being stirred at room
temperature for 12 h, the mixture was diluted with water (20
mL), and the precipitate was collected, washed with water (20
mL), dried, and recrystallized from 95% ethanol (2a -f) or
dimethylformamide (2g). The melting points (°C) and yields
(%) of 2a -g were as follows: 2a : 126-127, 79; 2b: 154-156,
90; 2c: 151, 78; 2d : 213, 85: 2e: 169-170, 68; 2f: 180-182,
80; 2g: 234, 58. The 1H NMR spectrum (CDCl3) of a repre-
sentative compound 2d was as follows: δ 4.70-4.82 (br s, 2H,
piperidyl H), 4.82-4.96 (br s, 2H, piperidyl H), 5.60-5.65 (dd,
1H, J ) 12.34 Hz, COCHdCH2), 6.15-6.25 (m, 2H, COCHd
CH2), 7.10-7.20 (t, 4H, 3,5-aryl H), 7.32-7.50 (br s, 4H, 2,6-
aryl H), 7.80 (s, 2H, arylidene H).
Str u ctu r e-Activity Rela tion sh ip s. The σ, π, and MR
values were taken from the literature.24 Since the MR value
of hydrogen is 1.03 and not 0.00, the sizes of the groups in
both the 3- and 4-positions of the aryl rings were calculated.
For example, in the case of 1a and 2a , the MR figure employed
was 2.06 (2 × 1.03) while for the 4-chloro analogues 1b and
2b the MR value chosen was 7.06, i.e., 6.03 + 1.03 for the
chloro and hydrogen atoms, respectively. In this way, differ-
ences in the sizes of the substituents were obtained. The
statistical analysis was undertaken using commercial soft-
ware.25
Molecu la r Mod elin g. Models were built using a Macro-
Model 4.5 program26,27 which was followed by a conformational
search using the Monte Carlo method and an amber force field
to obtain the shapes of the compounds possessing minimal
energy. The distances and angles in Tables 2 and 3 were
obtained directly from the models using standard analytical
geometrical procedures.
Con clu sion s
Many of the compounds prepared in this study were
shown to demonstrate pronounced cytotoxicity toward
four different tumor cell lines. N-Acryloylation of the
secondary amino group in series 1 leading to the
congeners 2 was clearly an important molecular change
leading to a significant increase in bioactivity. This
result is in accord with the theory of sequential cyto-
toxicity. Development of series 2 should take into
consideration the -σ relationship established by the SAR
investigations pertaining to the aryl substituents. Fur-
thermore, altering the size of the substituent at position
1 may lead to considerable variations in the relative
locations of the aryl rings which in turn may correlate
with changes in cytotoxicity. These molecules have the
potential for development into even more potent bioac-
tive agents, and 2e, possessing approximately nine
times the potency of melphalan, is a particularly
promising lead molecule. In addition, a number of
questions have been raised in this study that may be
addressed in subsequent investigations using this clus-
ter of compounds. For example, experimentation could
be undertaken with a view to discerning why cytotox-
icity was influenced by the shapes of the molecules as
well as identifying the biochemical mechanisms which
cause these compounds to be pro-apoptotic agents.
Finally, recent studies have revealed the promising
in vivo activities of both an R,â-unsaturated ketone21
and a Mannich base of an enone,22 thereby confirming
the value of vigorously investigating these compounds.
X-r a y Cr ysta llogr a p h y of 1e a n d 2b,d . The compounds
were recrystallized from benzene by slow evaporation (1b),
ethanol-2-propanol by vapor diffusion (1e), ethanol by slow
evaporation (2b), or methanol-2-propanol by vapor diffusion
(2d ). Compound 1e crystallized as a trisolvate of acetic acid.
Two of the three acetic acid molecules were involved in
hydrogen bonding with N1. A Nonius CAD-4 diffractometer
with a ω scan was used for data collection, and the structures
were solved using NRCVAX28 and refined using SHELXL97.29
Atomic scattering factors were taken from the literature.30 All
non-hydrogen atoms were found on the E-map and refined
anisotropically. Hydrogen atom positions were calculated and
not refined.
Exp er im en ta l Section
A. Ch em istr y. Melting points are uncorrected. Compounds
1c,g and 2a ,c have been reported in the literature and, in
general, had melting points similar to those recorded previ-
ously.9,23 Elemental analyses (C, H, N) were undertaken by
Mr. K. Thoms, Department of Chemistry, University of
Saskatchewan, and are within 0.4% of the calculated values.
1H NMR spectra were recorded for 1a -g and 2a -g using a
Brucker AM 500 FT NMR machine (500 MHz). A Nonius
Red ox P oten tia l Deter m in a tion s. The electrochemical
experiments were carried out using solutions of the compounds

592 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4
Dimmock et al.
(5) Benvenuto, J . A.; Connor, T. A.; Monteith, D. K.; Laidlaw, J .
W.; Adams, S. C.; Matney, T. S.; Theiss, J . C. Degradation and
inactivation of antitumor drugs. J . Pharm. Sci. 1993, 82, 988-
991.
(6) Dimmock, J . R.; Smith, L. M.; Smith, P. J . The reaction of some
nuclear substituted acyclic conjugated styryl ketones and related
Mannich bases with ethanethiol. Can. J . Chem. 1980, 58, 984-
991.
(7) Dimmock, J . R.; Taylor, W. G. Evaluation of nuclear-substituted
styryl ketones and related compounds for antitumor and cyto-
toxic properties. J . Pharm. Sci 1975, 64, 241-249.
(8) Wike-Hooley, J . L.; van den Berg, A. P.; van der Zee, J .;
Reinhold: H. S. Human tumour pH and its variation. Eur. J .
Cancer Clin. Oncol. 1985, 21, 785-791.
(9) Dimmock, J . R.; Arora, V. K.; Wonko, S. L.; Hamon, N. W.; Quail,
J . W.; J ia, Z.; Warrington, R. C.; Fang, W. D.; Lee, J . S. 3,5-bis-
Benzylidene-4-piperidones and related compounds with high
activity towards P388 leukemia cells. Drug Des. Delivery 1990,
6, 183-194.
(10) Dimmock, J . R.; Patil, S. A.; Leek, D. M.; Warrington, R. C.;
Fang, W. D. Evaluation of acrylophenones and related bis-
Mannich bases against murine P388 leukemia. Eur. J . Med.
Chem. 1987, 22, 545-551.
(11) Craig, P. N. Interdependence between physical parameters and
selection of substituent groups for correlation studies. J . Med.
Chem. 1971, 14, 680-684.
(12) Dimmock, J . R.; Sidhu, K. K.; Chen, M.; Reid, R. S.; Allen, T.
M.; Kao, G. Y.; Truitt, G. A. Evaluation of some Mannich bases
of cycloalkanones and related compound for cytotoxic activity.
Eur. J . Med. Chem. 1993, 28, 313-322.
(13) Suffness, M.; Douros, J . In Methods in Cancer Research, Volume
XVI, Part A; De Vita, V. T., J r., Busch, H., Eds; Academic
Press: New York, 1979; p 84.
(14) Dimmock, J . R.; Kandepu, N. M.; Hetherington, M.; Quail, J .
W.; Pugazhenthi, U.; Sudom, A. M.; Chamankhah, M.; Rose, P.;
Pass, E.; Allen, T. M.; Halleran, S.; Szydlowski, J .; Mutus, B.;
Tannous, M.; Manavathu, E. K.; Myers, T. G.; De Clercq, E.;
Balzarini, J . Cytotoxic activities of Mannich bases of chalcones
and related compounds. J . Med. Chem. 1998, 41, 1014-1026.
(15) Foye, W. O.; Sengupta, S. K. Cancer chemotherapy. In Principles
of Medicinal Chemistry, 4th ed.; Foye, W. O., Lemke, T. L.,
Williams, D. A., Eds.; Williams and Wilkins: Media, PA, 1995;
p 829.
(16) Silverman, R. B. The Organic Chemistry of Drug Design and
Drug Action, Academic Press: San Diego, CA, 1992; pp 247-
248.
(17) Tsurusawa, M.; Saeki, K.; Fujimoto, T. Differential induction of
apoptosis on human lymphoblastic leukemia Nalm-6 and Molt-4
cells by various antitumor drugs. Int. J . Hematol. 1997, 66, 79-
88.
(18) Vashishtha, S. C.; Nazarali, A. J .; Dimmock, J . R. Application
of fluorescence microscopy to measure apoptosis in J urkat T cells
after treatment with a new investigational anticancer agent (NC
1213). Cell. Mol. Neurobiol. 1998, 18, 437-445.
(19) Patrick, G. L. An Introduction to Medicinal Chemistry; Oxford
University Press: Oxford, 1995; p 74.
(20) Manavathu, E. K.; Vashishtha, S. C.; Alangaden, G. J .; Dim-
mock, J . R. In vitro antifungal activity of some Mannich bases
of conjugated styryl ketones. Can. J . Microbiol. 1998, 44, 74-
79.
(21) McMorris, T. C.; Yu, J .; Ngo, H.-T.; Wang, H.; Kelner, M. J .
Preparation and biological activity of amino acid and peptide
conjugates of antitumor hydroxymethylacyfulvene. J . Med.
Chem. 2000, 43, 3577-3580.
(0.1 mM) in freshly distilled acetone which had been dried over
anhydrous sodium sulfate. The supporting electrolyte was
tetrabutylammonium perchlorate (0.1 M). Determinations
were made using a Bioanalytical Systems Inc. (BAS) glassy
carbon working electrode (diameter of 2 mm) and a platinum
wire counter electrode. The reference electrode was a BAS
silver/silver chloride electrode. A scan rate of 100 mV s-1 was
employed. IR compensation was applied, and background scans
of the supporting electrolyte were collected and subtracted
from the spectra.
Biologica l Eva lu a tion s. The cytotoxic evaluation of the
compounds in series 1 and 2 against P388D1 cells was
undertaken using a literature procedure,31 while the evalua-
tion using murine L1210 cells as well as human Molt 4/C8
and CEM cells was carried out by a previously reported
methodology.32 Measurement of the inhibition of DNA, RNA,
and protein syntheses in murine L1210 cells by selected
compounds was accomplished using tritiated thymidine, uri-
dine, and leucine, respectively. The cytotoxicity and apoptotic
indices of 1c and 2c and melphalan using human J urkat T
cells was undertaken by a literature procedure18 except that
nigrosin dye was used instead of trypan blue in the cytotoxicity
experiments and a solution of ethidium bromide (0.1% w/v)
and acridine orange (0.1% w/v) was employed in measuring
apoptosis. The inhibition of RNA and protein syntheses using
tritiated uridine and 35S-methionine, respectively, was mea-
sured by a literature procedure.33 The antifungal evaluation
of 1b,c,e and 2b,c,e,3 and voriconazole was accomplished by
a broth microdilution method described previously20 using
three isolates of Aspergillus fumigatus (ATCC 208995, ATCC
208996, ATCC 208967) and one of Candida albicans (ATCC
90028).
Ack n ow led gm en t. The authors thank the following
agencies for financial support of the work described in
this study (initials of investigator funded in parenthe-
ses): CoCensys Inc. (J .R.D.), the Natural Sciences and
Engineering Research Council of Canada (A.J.N., J.W.Q.,
H.-B.K.), the National Cancer Institute of Canada
(T.M.A.), and the Fonds voor Wetenschappelijk Onder-
zoek-Vlaanderen (J .B., E.D.C.). Dr. L. Prasad is thanked
for her assistance in the execution and interpretation
of the molecular modeling and X-ray crystallography
experiments. Thanks are also extended to Mrs. C.
J amont who typed various drafts of the manuscript.
Su p p or tin g In for m a tion Ava ila ble: Crystal data, atomic
coordinates, equivalent isotropic displacement parameters,
bond lengths and angles, anisotropic displacement parameters,
hydrogen coordinates, isotropic displacement parameters, tor-
sion angles, and ORTEP diagrams34 for 1e, 2b,d as well as
elemental analysis data for 1a -g, 2a -2g. This information
is available free of charge via the Internet at http://pubs.ac-
s.org.
(22) Dimmock, J . R.; Vashishtha, S. C., Patil, S. A.; Udupa, N.;
Dinesh, S. B.; Devi, P. U.; Kamath, R. Cytotoxic and anticancer
activities of some 1-aryl-2-dimethylaminomethyl-2-propen-1-one
hydrochlorides. Pharmazie 1998, 53, 702-706.
(23) Dimmock, J . R.; Arora, V. K.; Duffy, M. J .; Reid, R. S.; Allen, T.
M.; Kao, G. Y. Evaluation of some N-acyl analogues of 3, 5-bis-
(arylidene)-4-piperidones for cytotoxic activity. Drug Des. Dis-
covery. 1992, 8, 291-299.
(24) Hansch, C.; Leo, A. J . Substituent Constants for Correlation
Analysis in Chemistry and Biology; J ohn Wiley and Sons: New
York, 1979; p 49.
(25) Statistical Package for the Social Sciences, SPSS for Windows,
Standard Version, Release 9.0.0; SPSS Inc.: Chicago, IL, 1998.
(26) MacroModel, Version 4.5; Department of Chemistry, Columbia
University: New York, August 1994.
Refer en ces
(1) Dimmock, J . R.; Elias, D. W.; Beazely, M. A.; Kandepu, N. M.
Bioactivities of chalcones. Curr. Med. Chem. 1999, 6, 1125-1149.
(2) Dimmock, J . R.; Kandepu, N. M.; Nazarali, A. J .; Kowalchuk,
T. P.; Motaganahalli, N.; Quail, J . W.; Mykytiuk, P. A.; Audette,
G. F.; Prasad, L.; Perje´si, P.; Allen, T. M.; Santos, C. L.;
Szydlowski, J .; De Clercq, E.; Balzarini, J . Conformational and
quantitative structure-activity relationship study of cytotoxic
2-arylidenebenzocycloalkanones. J . Med. Chem. 1999, 42, 1358-
1366.
(3) Mutus, B.; Wagner, J . D.; Talpas, C. J .; Dimmock, J . R.; Phillips,
O. A.; Reid, R. S. 1-p-Chlorophenyl-4,4-dimethyl-5-diethylamino-
1-penten-3-one hydrobromide, a sulfhydryl-specific compound
which reacts irreversibly with protein thiols but reversibly with
small molecular weight thiols. Anal. Biochem. 1989, 177, 237-
243.
(27) Mohamadi, F.; Richards, N. G. J .; Guide, W. C.; Liskamp, M.;
Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. MacroModel
- An integrated software system for modeling organic and
bioorganic molecules using molecular mechanics. J . Comput.
Chem. 1990, 11, 440-467.
(4) Baluja, G.; Municio, A. M.; Vega, S. Reactivity of some R,â-
unsaturated ketones towards sulfhydryl compounds and their
antifungal activity. Chem. Ind. 1964, 2053-2054.

SAR Study of Cytotoxic 3,5-Bis(arylidene)-4-piperidones
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4 593
(28) Gabe, E. J .; LePage, Y.; Charland, J . P.; Lee, F. L.; White, P. S.
An interactive program system for structure analyses. J . Appl.
Crystallogr. 1989, 22, 384-387.
(29) Sheldrick, G. M. SHELXL-97. Program for the Refinement of
Crystal Structures; University of Gottingen: Germany, 1997.
(30) International Tables for X-ray crystallography; Kynoch Press:
Birmingham, 1974; Vol. IV.
(31) Phillips, O. A.; Nelson, L. A.; Knaus, E. E.; Allen, T. M.; Fathi-
Afshar, R. Synthesis and cytotoxic activity of pyridylthio, py-
ridylsulfinyl and pyridosulfonyl methyl acrylates. Drug Des.
Delivery 1989, 4, 121-127.
(32) Balzarini, J .; De Clercq, E.; Mertes, M. P.; Shugar, D.; Torrence,
P. F. 5-Substituted 2-deoxyuridines; correlation between inhibi-
tion of tumor cell growth and inhibition of thymidine kinase and
thymidylate synthetase. Biochem. Pharmacol. 1982, 31, 3673-
3682.
(33) Martin, S. J .; Lennon, S. V.; Bonham, A. M.; Cotter, T. G.
Induction of apoptosis (programmed cell death) in human
leukemic HL-60 cells by inhibition of RNA and protein synthesis.
J . Immunol. 1990, 145, 1859-1867.
(34) J ohnson, C. K. ORTEP II. Report ORNL - 5138; Oak Ridge
National Laboratory: Tennesse, 1976.
J M0002580