Bioorganic & Medicinal Chemistry 9 (2001) 2885–2893

Comparative QSAR Studies on Bibenzimidazoles and

Terbenzimidazoles Inhibiting Topoisomerase I
Suresh Babu Mekapati and Corwin Hansch*
Department of Chemistry, Pomona College, Claremont, CA 91711, USA

Received 16 February 2001; accepted 7 May 2001

Abstract—Terbenzimidazoles that inhibit topoisomerase are of interest as anticancer drugs. We have reviewed the literature and
have developed 13 quantitative structure–activity relationships (QSARs) on cleaving DNA or inhibiting the growth of tumor cell
cultures. The results are correlated with octanol/water partition coefficients or molecular refractivity. Suggestions have been made
for the development of improved derivatives. # 2001 Elsevier Science Ltd. All rights reserved.

Introduction The chemotherapeutic action of several anticancer

agents has been linked to their ability to inhibit nuclear
Since its advent,1,2 the subject of quantitative structure– DNA topoisomerases. These are nuclear enzymes
activity relationships (QSARs) has been expanding at involved in generating the necessary topological and
an exponential rate. There are now almost 7000 web conformational changes in DNA critical to many cel-
sites devoted to QSAR. Citations to C. Hansch in the lular processes such as replication and transcription.9
11
period 1979–1989 were only 8788. In the period 1989 Recent studies suggest that topoisomerases are also
through August 2000 the rate expanded greatly to involved in controlling template supercoiling during
102,000. RNA transcription and helicase movement.12
14 Topo-
isomerases have been classified into type I and II
Bioinformatics is clearly called for to get some kind of depending on their ability to produce transient protein-
grasp on the enormous amount of data that is appearing bridged single-strand or double-strand DNA breaks.
at an ever faster rate. On this point, we are building a Inhibitors that interfere with the breaking and rejoining
computerized system of QSAR that now contains reactions of these enzymes by trapping an abortive
16,000 equations, and attendant data, of which 7600 are enzyme–DNA cleavable complex have been termed
for chemical–biological reactions and 8400 are from topoisomerase poisons.2 Such inhibition of mammalian
mechanistic organic chemistry for comparison. In the topoisomerases has been recognized as an effective
course, of this work we find from time to time specific approach for cancer chemotherapy.15,16 It has been
subjects that are worthy of an organized review that we demonstrated that Hoechst 33342 (I) represents a struc-
publish3
8 to illustrate the value of the system. It is most turally unique class of topo I poisons. This agent, that
surprising that with the large number of SAR studies binds to the minor groove of DNA, traps the reversible
being published that very few attempts are made by the cleavable complex formed from DNA and topo I and
authors to formulate any kind of a QSAR. produces a limited number of highly specific single-
stranded DNA breaks.17,18 A limitation of Hoechst
In this report, we consider the anticancer activity of
bibenzimidazoles and terbenzimidazoles. As usual, we
find that of 13 sets of data, only one attempt24 was
made by the authors to formulate QSAR.

*Corresponding author.

0968-0896/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(01)00174-2
2886 S. B. Mekapati, C. Hansch / Bioorg. Med. Chem. 9 (2001) 2885–2893

33342 as an effective anticancer drug is that it is not X=2-tolyl, Y=4-methyl piperazinyl. r2 for ClogP ver-
effective against certain tumor cell lines. Several ana- sus CMR=0.385. The indicator variable I takes the
logues of bibenzimidazoles and terbenzimidazoles have value of 1 for instances where CONH2 is present as
been identified as topo I poisons and have exhibited Y. Its negative coefficient shows the such a group is
good cytotoxicity against RPMI8402 a human lympho- detrimental to activity.
blastoma cell line.19
26 In every instance, we have taken
data from the literature, as noted, and formulated ii. Relative effective concentrations of terbenzimidazoles
QSAR. In all the equations, n is the number of data (Fig. 1) causing 50% cleavage of DNA in the presence
points, r2 is the square of correlation coefficient, q2 is the of Escherichia coli topoisomerase I23 (Table 2).
measure of quality of fit and s is the standard deviation.

LogRBR ¼ 21:90ð7:90ÞCMR  0:78ð0:28ÞðCMRÞ2

Results  153:55ð55:82Þ
The relative potency of various substituted bibenzimid- n ¼ 10; r2 ¼ 0:869; s ¼ 0:208; q2 ¼ 0:787 ð2Þ
azoles and terbenzimidazoles has been determined by
assessing their ability to induce DNA cleavage in the
presence of the enzyme and from this data we have Opt. CMR=14.1 (13.9–14.2)
derived the following QSARs.
See Table 2 for three outliers.
i. Relative effective concentration of bibenzimidazoles
(II) causing 50% cleavage of DNA in the presence of iii. Relative effective concentrations of benzimidazoles
calf thymus topoisomerase I21 (Table 1). (III) causing 50% cleavage of DNA in the presence of
calf thymus topoisomerase I20 (Table 3)

LogRBR ¼ 10:67ð7:40ÞClogP  1:06ð0:69ÞðClogPÞ2

LogRBR ¼ 7:56ð1:21ÞClogP  0:83ð0:13ÞðClogPÞ2
 1:58ð0:58ÞI  26:70ð19:84Þ
 17:32ð2:74Þ
n ¼ 13; r2 ¼ 0:871; s ¼ 0:321; q2 ¼ 0:724 ð1Þ
n ¼ 9; r2 ¼ 0:976; s ¼ 0:222; q2 ¼ 0:844 ð3Þ

Opt. ClogP=5.03 (4.29–5.27)

Outliers: X=C6H5, Y=C7H4N2. Opt. ClogP=4.44
Outliers: X=2-tolyl, Y=CN; X=phenyl, Y=CONH2; (4.44–4.66).

Table 1. Relative effective concentrations of bibenzimidazoles (II) causing 50% cleavage of DNA in the presence of calf thymus topoisomerase I21

Compound X Y LogRBR ClogP I

Obsd calcd [eq (1)] 

a
1 3-Tolyl CN
2.00 0.03
2.03 5.35 0
2 4-Tolyl CN 0 0.03
0.03 5.35 0
3 1-Naphthyl CN
1.00
0.85
0.15 6.02 0
4 2-Naphthyl CN
1.00
0.85
0.15 6.02 0
5 Phenyl CONH2
1.00a
2.22 1.22 4.50 1
6 2-Tolyl CONH2
2.00
1.91
0.09 4.69 1
7 3-Tolyl CONH2
2.00
1.60
0.40 4.99 1
8 4-Tolyl CONH2
1.00
1.60 0.60 4.99 1
9 1-Naphthyl CONH2
1.30
1.60 0.30 5.67 1
10 2-Naphthyl CONH2
2.00
1.60
0.40 5.67 1
11 Phenyl 4-Methyl-piperazinyl 0
0.10
0.10 4.91 0
12 2-Tolyl 4-Methyl-piperazinyl
1.00a 0.08
1.08 5.11 0
13 3-Tolyl 4-Methyl-piperazinyl 0
0.12 0.12 5.41 0
14 4-Tolyl 4-Methyl-piperazinyl 0
0.12 0.12 5.41 0
15 1-Naphthyl 4-Methyl-piperazinyl
1.30
1.25
0.50 6.08 0
16 2-Naphthyl 4-Methyl-piperazinyl
1.00
1.25 0.25 6.08 0
a
Data points not included in deriving the equation.
 S. B. Mekapati, C. Hansch / Bioorg. Med. Chem. 9 (2001) 2885–2893 2887

Figure 1.

iv. Relative effective concentrations of 5-substituted LogRBR ¼ 3:76ð1:83ÞCMR  0:16ð0:07ÞðCMRÞ2

terbenzimidazoles (IV) causing 50% cleavage of  21:95ð11:33Þ
DNA in the presence of E. coli topoisomerase24
(Table 4) n ¼ 11; r2 ¼ 0:902; s ¼ 0:168; q2 ¼ 0:835 ð4Þ

Outliers: F, OCH3, CN. Opt. CMR=11.81 (11.11–12.13).

A reasonably good correlation is also observed with

ClogP, with an optimum ClogP 5.29 (4.50–5.61)
(r2=0.86). ClogP and CMR are mutually correlated
(r2=0.899).

v. Relative concentrations of bisbenzimidazoles (V)

causing 50% cleavage of DNA in presence of calf thy-
mus topoisomerase19 (Table 5).
2888 S. B. Mekapati, C. Hansch / Bioorg. Med. Chem. 9 (2001) 2885–2893

i. IC50 values (concentration of drug to reduce cell

number to 50% of control) of 4,5-substituted 2-(4-
methoxyphenyl)-1H-benzimidazoles (VI) against human
lymphoblastoma cell line RPMI840222 (Table 6).

LogRBR ¼ 1:63ð1:09ÞClogP þ 4:08ð1:25ÞMgVol

 5:24ð5:65Þ
n ¼ 9; r2 ¼ 0:916; s ¼ 0:417; q2 ¼ 0:823 ð5Þ
Log1=C ¼ 0:79ð0:27Þ ClogP þ 1:93ð0:93Þ
n ¼ 8; r2 ¼ 0:893; s ¼ 0:20; q2 ¼ 0:846 ð6Þ
MgVol is the molar volume calculated by the method of
McGowan.27
Outliers X=H, Y=H.
Cytotoxicity studies of benzimidazoles
r2 for ClogP versus CMR=0.028.
Cytotoxicity for QSARs 6–12 of substituted bibenzimid-
azoles and terbenzimidazoles were determined against ii. IC50 values of 2,50 -substituted-bi-1H-benzimidazoles
human lymphoblastoma cell line RPMI8402. (VII) against human lymphoblastoma cell line
RPMI840221 (Table 7).

Table 2. Relative effective concentrations of terbenzimidazoles (Fig. 1)

causing 50% cleavage of DNA in the presence of E. coli topoisomerase
I23

Compound LogRBR CMR

Obsd calcd [eq (2)] 

1 0.00a
0.78 0.78 12.94 Log1=C ¼ 0:88ð0:31ÞClogP  0:55ð0:33ÞI1

2 0.00a
2.74 2.74 12.12
3
0.70
0.78 0.08 12.94  0:36ð0:31ÞI2 þ 1:57ð1:66Þ
4
0.30
0.05
0.25 14.63
5 0.00
0.05 0.05 14.63 n ¼ 15; r2 ¼ 0:909; s ¼ 0:187; q2 ¼ 0:840 ð7Þ
6 0.30
0.00 0.30 13.56
7
0.30
0.11
0.19 13.43
8
0.30
0.13
0.16 13.41 Indicator variable I1 takes the value of 1 for the pres-
9 0.00 0.16
0.17 13.87 ence of 1-naphthyl groups in Y, I2 takes the value of 1
10 0.30 0.13 0.16 14.33 for instances where CONH2 is present in X. The nega-
11 0.30 0.13 0.16 14.33 tive coefficient of I1 and I2 indicates such groups are
12
1.30
1.31 0.01 15.45
13
1.30a
2.94 1.64 16.07 detrimental to cytotoxicity.
a
Data points not included in deriving the equation.
Table 4. Relative effective concentrations of 5-substituted terbenzi-
midazoles (IV) causing 50% cleavage of DNA in the presence of E.
coli topoisomerase I24
Table 3. Relative effective concentrations of benzimidazoles (III)
causing 50% cleavage of DNA in the presence of calf thymus topo- Compound X LogRBR CMR
isomerase I20
Obsd calcd [eq (4)] 
Compound X Y n LogRBR ClogP
1 1-Naphthyl
1.00
0.98
0.02 14.63
Obsd calcd  2 2-Naphthyl
1.00
0.98
0.02 14.63
[eq (3)] 3 Phenyl 0.00 0.08
0.08 12.94
4 Propyl 0.30 0.28 0.02 11.82
1 H H 2
0.04
0.13 0.09 4.61 5 Br 0.00 0.23
0.22 11.21
2 CN H 2 0
0.22 0.22 4.22 6 Piperidnyl 0.30 0.08 0.22 12.94
3 n-C3H7 H 2
2.0
2.30 0.30 6.16 7 Cl 0.00 0.16
0.16 10.92
4 Phenyl H 2
0.30a
3.29 2.99 6.50 8 F 1.30a
0.01 1.32 10.45
5 2-Pyridyl H 2
0.52
0.49
0.03 5.21 9 H 0.00
0.02 0.02 10.43
6 3-Pyridyl H 2
0.30
0.30
0.00 5.00 10 OCH3 1.30a 0.19 1.11 11.05
7 4-Pyridyl H 2
0.30
0.30
0.00 5.00 11 NO2 0.30 0.19 0.11 11.04
8 CN H 1
3.0
2.97
0.03 2.70 12 CN 1.00a 0.15 0.85 10.91
9 CN OCH3 1
0.52
0.19
0.33 4.82 13 OH 0.30 0.04 0.26 10.58
10 CN 4-OCH3–C6H4 2
3.0
2.80
0.20 6.34 14 NH2 0.00 0.12
0.12 10.80
a a
Data point not included in deriving the equation. Data points not included in deriving the equation.
 S. B. Mekapati, C. Hansch / Bioorg. Med. Chem. 9 (2001) 2885–2893 2889

Table 5. Relative effective concentrations of bisbenzimidazoles (V) causing 50% cleavage of DNA in the presence of calf thymus topoisomerase I19

Compound X Y LogRBR ClogP MgVol

Obsd calcd [eq (5)] 

1 OC2H5 N–(4-CH3)-piperazinyl 0 0.11
0.11 5.36 3.45

2 OCH3 (CH2)3N(CH3)2
0.30
0.75 0.45 5.56 3.31
3 OCH3 (CH2)2N(CH3)2
0.70
0.71 0.01 5.18 3.17
4 OCH3 CH2N(CH3)2
0.40
1.05 0.66 5.04 3.03
5 OCH3 N(CH3)2
3.00
2.84
0.16 5.78 2.89
6 OCH3 NO2
3.00
2.60
0.40 5.12 2.68
7 OCH3 NH2
2.00
2.15 0.15 4.65 2.61
8 OCH3 N–(4-CH3)-piperazinyl 0 0.40
0.40 4.83 3.31
9 OCH3 O–[(4-N–CH3)-piperidinyl]
0.30
0.11
0.19 5.40 3.41

r2 for ClogP versus CMR=0.267. Log1=C ¼ 1:55ð0:51ÞClogP  0:95ð0:49ÞB5X

 0:81ð1:83Þ
iii. IC50 values of 5-substituted terbenzimidazoles (VIII)
against human lymphoblastoma cell line RPMI840224 n ¼ 13; r2 ¼ 0:854; s ¼ 0:448; q2 ¼ 0:773 ð8Þ
(Table 8).

r2 for ClogP versus CMR=0.884. Outliers: OCH3.

iv. IC50 values of 200 -substituted-5-phenyl terbenzimid-

azoles (IX) against human lymphoblastoma cell line
RPMI840225 (Table 9).

Table 6. IC50 values of 4,5-substituted 2-(4-methoxy-phenyl)-1H-benzimidazoles (VI) against human lymphoblastoma cell line RPMI840222

Compound X Y Log1/C ClogP

Obsd calcd [eq (6)] 

1 CN H 4.80 4.54 0.25 3.29

2 CHO H 4.44 4.70
0.26 3.49
3 CH2OH H 4.01 4.03
0.02 2.65
4 CH2NH2 H 3.92 4.02
0.10 2.64
5 CONH2 H 4.18 4.06 0.12 2.68
6 H H 4.35a 4.86
0.51 3.69
7 Br H 5.64 5.60 0.04 4.62
8 NO2 H 4.55 4.78
0.22 3.59
9 H NO2 4.96 4.78 0.18 3.59
a
Data points not included in deriving equation.

Table 7. IC50 values of 2,50 substituted-bi-1H-benzimidazoles (VII) against human lymphoblastoma cell line RPMI840221

Compound X Y Log1/C ClogP I1 I2

Obsd calcd [eq (7)] 

1 2-Tolyl CN 5.92 5.97
0.05 5.05 0 0

2 3-Tolyl CN 6.07 6.23
0.16 5.35 0 0
3 4-Tolyl CN 6.24 6.23 0.01 5.35 0 0
4 1-Naphthyl CN 6.19 6.28
0.09 6.02 1 0
5 2-Naphthyl CN 6.89 6.82 0.06 6.02 0 0
6 2-Tolyl CONH2 5.09 5.08 0.01 4.69 0 1
7 3-Tolyl CONH2 5.09 5.34
0.25 4.99 0 1
8 1-Naphthyl CONH2 5.43 5.39 0.05 5.67 1 1
9 2-Naphthyl CONH2 6.13 5.93 0.20 5.67 0 1
10 Phenyl 4-Methyl-piperazinyl 6.31 5.95 0.36 4.91 0 0
11 2-Tolyl 4-Methyl-piperazinyl 6.23 6.13 0.10 5.11 0 0
12 3-Tolyl 4-Methyl-piperazinyl 6.44 6.39 0.05 5.41 0 0
13 4-Tolyl 4-Methyl-piperazinyl 6.08 6.39
0.31 5.41 0 0
14 1-Naphthyl 4-Methyl-piperazinyl 6.48 6.44 0.04 6.08 1 0
15 2-Naphthyl 4-Methyl-piperazinyl 6.96 6.99
0.03 6.08 0 0
2890 S. B. Mekapati, C. Hansch / Bioorg. Med. Chem. 9 (2001) 2885–2893

Log1=C ¼ 0:94ð0:61ÞClogP þ 1:44ð3:16Þ

n ¼ 5; r2 ¼ 0:889; s ¼ 0:156; q2 ¼ 0:802 ð10Þ
Log1=C ¼ 0:69ð0:23ÞClogP  0:58ð0:19ÞCMR
þ 9:82ð2:94Þ
n ¼ 12; r2 ¼ 0:902; s ¼ 0:281; q2 ¼ 0:790 ð9Þ Outliers: X=OC2H5, Y=N-(4-CH3–piperazine) and
X=OCH3, Y=(CH2)3N(CH3)2.

Outliers: CH2CH2OCH3; CH2CH2NHCOCH3. r2 for ClogP versus CMR=0.134.

r2 for ClogP versus CMR=0.013. vi. IC50 values of terbenzimidazoles (Fig. 1) against
human lymphoblastoma cell line RPMI840223 (Table 11).
v. IC50 values of bisbenzimidazoles (X) against human
lymphoblastoma cell line RPMI840219 (Table 10).
Log1=C ¼ 13:05ð5:41ÞCMR  0:45ð0:19ÞðCMRÞ2
Table 8. IC50 values of 5-substituted terbenzimidazoles (VIII) against  88:43ð38:18Þ
human lymphoblastoma cell line RPMI840224
n ¼ 11; r2 ¼ 0:855; s ¼ 0:385; q2 ¼ 0:714 ð11Þ
Compound X Log1/C ClogP B5X

Obsd calcd [eq (8)] 

1 1-Naphthyl 5.90 5.85 0.05 7.67 5.50

2 2-Naphthyl 6.80 6.98
0.18 7.67 4.31 Table 10. IC50 values of bis benzimidazoles (X) against human lym-
3 Phenyl 6.72 6.29 0.43 6.50 3.11 phoblastoma cell line RPMI840219
4 Propyl 4.82 5.42
0.61 6.16 3.49
5 Br 5.79 5.92
0.14 5.55 1.95 Compound X Y Log 1/C ClogP
6 Piperidnyl 6.20 5.75 0.45 6.38 3.49 Obsd calcd 
7 Cl 5.89 5.83 0.05 5.40 1.80 [eq (10)]
8 F 5.77 5.38 0.39 4.83 1.35
9 H 5.30 5.37
0.07 4.61 1.00 1 OC2H5 N-(4-CH3)-piperazinyl 8.30a 8.17 0.13 5.36
10 OCH3 6.10a 3.88 2.22 4.91 3.07 2 OCH3 (CH2)3N(CH3)2 5.92a 7.66
1.74 5.56
11 NO2 3.94 3.87 0.08 4.52 2.44 3 OCH3 (CH2)2N(CH3)2 6.22 7.30
1.08 5.18
12 CN 3.88 4.20
0.33 4.22 1.60 4 OCH3 CH2N(CH3)2 6.42 6.94
0.52 5.04
13 OH 3.82 4.56
0.74 4.66 1.93 5 OCH3 N(CH3)2 6.89 6.57 0.31 5.78
14 NH2 4.21 3.60 0.61 4.06 1.97 6 OCH3 NO2 6.19 6.04 0.15 5.12
7 OCH3 NH2 5.77 5.85
0.08 4.65
a
Data point not included in deriving the equation.
a
Data points not included in deriving the equation.
00
Table 9. IC50 values of 2 -substituted-5-phenyl terbenzimidazoles
(IX) against human lymphoblastoma cell line RPMI840225
Table 11. IC50 values of terbenzimidazoles (Fig. 1) against human
Compound X Log1/C ClogP CMR lymphoblastoma cell line RPMI840223
Obsd calcd  Compound Log1/C CMR
[eq (9)]
Obsd Cald [eq (11)] 
1 H 6.57 6.84
0.27 6.50 12.94
2 OH 6.70 6.95
0.25 6.78 13.09 1 7.05a 6.00 1.05 12.94
3 SH 6.92 6.96
0.04 7.34 13.75 2 4.60 4.43 0.17 12.12
4 Cl 7.16 7.12 0.03 7.31 13.43 3 5.15 6.00
0.85 12.94
5 CF3 7.16 7.28
0.13 7.56 13.45 4 5.68a 7.33
1.65 14.63
6 CH3 6.96 6.76 0.20 6.76 13.41 5 7.10 7.33
0.24 14.63
7 CH2OH 6.27 5.77 0.50 5.46 13.56 6 7.05 6.78 0.27 13.56
8 CH2NHCOC6H5 5.02 5.10
0.08 7.18 16.79 7 7.05 6.65 0.40 13.43
9 CH2CH3 6.96 6.86 0.10 7.29 13.87 8 6.96 6.62 0.34 13.41
10 CH2CH2OCH3 6.54a 5.63 0.91 6.03 14.49 9 6.96 7.05
0.09 13.87
11 CH2CH2OH 5.59 5.50 0.09 5.46 14.02 10 7.22 7.28
0.06 14.33
12 CH2CH2NH2 5.03 5.45
0.42 5.56 14.24 11 7.22 7.28
0.06 14.33
13 CH2CH2NHCOCH3 7.00a 4.61 2.39 5.14 15.20 12 7.22 7.06 0.16 15.45
14 CH2CH2CH3 7.22 6.96 0.27 7.82 14.33 13 6.42 6.46
0.04 16.07
a a
Data points not included in deriving equation. Data points not included in deriving the equation.
 S. B. Mekapati, C. Hansch / Bioorg. Med. Chem. 9 (2001) 2885–2893 2891

See Table 11 for two outliers. Discussion

r2 for ClogP versus CMR=0.919. CMR (calculated MR) is defined as molecular

refractivity=(n2
1/n2+2)MW/d, where n=refractive
vii. IC50 values bis and terbenzimidazoles (XI) against index, Mr=molecular weight and d=density. ‘n’
human lymphoblastoma cell line RPMI840220 (Table accounts for the polarizability. Unless there is a sig-
12) nificant variation in ‘n’, CMR is a measure of molecular
volume. Inspection of the 13 QSARs shows that the two
parameters of greatest significance are ClogP (calculated
octanol/water partition coefficient) and CMR. If care is
not taken in preparation of the data sets collinearity can
be a serious problem. Hence, in each example we have
listed the degree of collinearity for these two variables.
Serious collinearity occurs in eqs (2), (3), (4), (8), and
(11). This information must be kept in mind in the data
analysis. There is a clear need to test a few more well
designed congeners for these sets to be certain of the
Log1=C ¼ 1:11ð0:52ÞCMR  7:09ð 0:30Þ role of hydrophobicity. From QSARs (1) and (3), we
find optimum ClogP of 5.3 and 4.5, respectively. The
n ¼ 7; r2 ¼ 0:855; s ¼ 0:534; q2 ¼ 0:741 ð12Þ similarity is a bit surprising considering the difference in
the parent structure. This is also seen in eq (3), where n
can be either 1 or 2 but no account of this was needed to
r2 for ClogP versus CMR is 0.232. form the QSAR.

viii. IC50 values of 5-X, 6-Y, 200 -Z-substituted terbenzi- However gross bulk and polarizability are important if
midazoles (XII) against human lymphoblastoma cell the variation is large enough as seen in QSAR (2).
line RPMI840226 (Table 13).
Although we have used CMR in eq (4), the collinearity
between ClogP and CMR is so high that ClogP might
well be the significant parameter. QSAR (5) also shows
the importantance of bulk in terms of volume but with
negative coefficent of ClogP.

For QSARs (1)–(5), activity is defined as the relative

concentration required to cleave 50% of DNA. QSARs
(6)–(13) are based on cell inhibition.

Log1=C ¼ 0:75ð0:38ÞZ þ 0:86ð0:47ÞI þ 5:86ð0:34Þ QSAR (6) is clearly ClogP-dependent and this would
suggest the presence of a hydrophobic site in the recep-
n ¼ 10; r2 ¼ 0:851; s ¼ 0:309; q2 ¼ 0:669 ð13Þ tor that would not accommodate larger molecules.
QSAR (7) has a ClogP term but needs help from two
indicator variables. I1 accounts for a bulky naphthyl
Outliers: X=4-Cl-C6H4, Y=H, Z=CH3; X=Br, group in the X-position and its negative coefficient sug-
Y=Br, Z=H and X=Y=C6H5, Z=H.
Table 13. IC50 values of 5,6,200 -substituted terbenzimidazoles (XII)
The indicator variable I takes the value of 1 for instan- against human lymphoblastoma cell line RPMI840226
ces where C6H5 is present in X. Its positive coeffi-
cient indicates that presence of phenyl group is Compound X Y Z Log 1/C pZ I
more beneficial than Br. Obsd calcd 
[eq (13)]
Table 12. IC50 values of bis and ter-benzimidazoles (XI) against
1 C6H5 H H 7.05 6.72 0.33 0 1
human lymphoblastoma cell line RPMI840220 2 C6H5 H Cl 7.16 7.25
0.09 0.711
3 C6H5 H CF3 7.40 7.38 0.02 0.881
Compound X Y n Log 1/C ClogP
4 Br H H 5.80 5.86
0.06 0 0
Obsd calcd  5 Br H OH 5.19 5.36
0.17
0.670
[eq (12)] 6 Br H CH2CH2CH3 6.68 7.03
0.35 1.550
7 4-Cl–C6H4 H CF3 6.48a 7.38
0.90 0.881
1 H H 2 4.85 4.45 0.41 4.61 8 Br Br H 6.59a 5.86 0.73 0 0
2 n-C3H7 H 2 5.12 5.99
0.87 6.16 9 Br Br Cl 6.96 6.39 0.57 0.710
3 Phenyl H 2 7.05 7.23
0.18 6.50 10 Br Br CF3 6.60 6.52 0.08 0.880
4 2-Pyridyl H 2 6.80 6.99
0.20 5.21 11 C6H5 C6H5 H 5.80a 6.72
0.92 0 1
5 3-Pyridyl H 2 7.46 6.99 0.46 5.00 12 Br OCH3 H 5.80 5.86
0.06 0 0
6 4-Pyridyl H 2 7.46 6.99 0.46 5.00 13 C6H5 OCH3 H 6.46 6.72
0.26 0 1
7 CN OCH3 1 4.57 4.66
0.09 4.82
a
Data points not included in deriving the equation.
2892 S. B. Mekapati, C. Hansch / Bioorg. Med. Chem. 9 (2001) 2885–2893

gests a steric effect. It is of interest that only 1-naphthyl that could metabolize molecules.29 A place to start
is so parameterized, 2-naphthyl is accommodated by looking for increased potency would with set (7). Com-
ClogP. pound 5 is not covered by the indicator variable. Hence
replacing CN group (p=
0.57) with SO2NHNH2
Eq (8) is also highly ClogP-dependent, however, bulky (p=
2.04) would lower Log P 1.5 units. Many other
variations of X have a deleterious effect brought out by possibilities are available.30 Over 1000 p values have
B5X. been published.

QSAR (9) is derived from a very large parent structure; There are not many obvious points where one might
nevertheless, with a small CMR correction, ClogP is start to make more potent compounds. In QSARs (1),
quite significant. In this instance, substitution is only at (2), (3), (4), (9), and (11), where the optimums have been
one site on the molecule. established for either ClogP or CMR, there are no
obvious paths to more potent compounds. In none of
QSAR (10) is clearly ClogP determined after removing the QSARs could we find an electronic role for sub-
two of most basic congeners. There is essentially no stituents. The one general opportunity would be in the
variation in X. structures for QSARs (6), (10), (12), and (13) where
optimum values have not been established for CMR or
In the case of QSAR (11), very large terbenzimidazoles hydrophobic terms. Hence, increases in these para-
with rather large substituents must be considered. The meters should yield more potent compounds.
collinearity between ClogP and CMR is extreme.
Despite this, r2 using ClogP is only 0.690. Although
CMR is considerably better, one can not be sure that References and Notes
both effects are to some degree present.

QSAR (12) clearly demonstrates the role of CMR. Only 1. Hansch, C.; Maloney, P. P.; Fujita, T.; Muir, R. M. Nature
one example where n=1 was tested and it is well fit 1962, 194, 178.
2. Hansch, C. Acc. Chem. Res. 1969, 2, 232.
without special parameterization.
3. Hansch, C.; Hoekman, D.; Gao, H. Chem. Rev. 1996, 96,
1045.
In example (13), the use of p constants (hydrophobicity 4. Gao, H.; Hansch, C. Drug Metab. Rev. 1996, 18, 513.
of the substituent) clearly points to a lipophilic pocket. 5. Hansch, C.; Gao, H. Chem. Rev. 1997, 97, 2995.
6. Gao, H.; Katzenellenbogen, J. A.; Garg, R.; Hansch, C.
It is of course of great interest to see which are the most Chem. Rev. 1999, 99, 723.
potent compounds. Most of the end points are relative 7. Garg, R.; Gupta, S. P.; Gao, H.; Babu, M. S.; Debnath,
biological responses and are not good for comparison. A. K.; Hansch, C. Chem. Rev. 1999, 99, 3525.
In six examples [(6), (7), (8), (9), (11), and (13)] we can 8. Kurup, A.; Garg, R.; Hansch, C. Chem. Rev. 2000, 100, 909.
determine the most potent congener as shown in Table 9. Wang, J. C. Annu. Rev. Biochem. 1985, 54, 665.
10. Liu, L. F. Annu. Rev. Biochem. 1989, 58, 351.
14. 11. D’Arpa, P.; Liu, L. F. Biochem. Biophys. Acta 1989, 989, 163.
12. Liu, L. F.; Wang, J. C. Proc. Natl. Acad. Sci. U.S.A. 1987,
The agreement between sets (7)–(13) is remarkable and 84, 7024.
would suggest that we are at a dead end. Of course, we 13. Tsao, Y.-P.; Wu, H.-Y.; Liu, L. F. Cell 1989, 56, 111.
must keep in mind that some of the correlations are not 14. Koo, H. S.; Classen, L.; Grossman, L.; Liu, L. F. Proc.
as sharp as one would like. Only sets (7) and (12) have Natl. Acad. Sci. U.S.A. 1991, 88, 1212.
no outliers. Set (8) has 1, others have 2 or 3. In a way, 15. Schneider, E.; Hsiang, Y.-H.; Liu, L. F. Adv. Pharmacol.
this is amazing when one considers the complexity of 1990, 21, 149.
the structures involved. The receptor site would seem to 16. Chen, A. Y.; Liu, L. F. Annu. Rev. Pharmacol. Toxicol.
be relatively flexible. Clearly, there is considerable 1994, 34, 191.
17. Chen, A. Y.; Yu, C.; Bodley, A. L.; Peng, L. F.; Liu, L. F.
hydrophobic space. There is almost a one to one corre- Cancer Res. 1993, 53, 1332.
spondence between Log 1/C and ClogP. However, these 18. Chen, A.; Yu, C.; Gatto, B.; Liu, L. F. Proc. Natl. Acad.
results are from cell studies. It is quite likely that in Sci. U.S.A. 1993, 90, 8131.
animal or human studies, a lower LogP would be desir- 19. Sun, Q.; Gatto, B.; Yu, C.; Liu, A.; Liu, L. F.; LaVoie,
able for better bioavailability.28 Very hydrophobic E. J. Bioorg. Med. Chem. Lett. 1994, 4, 2871.
compounds tend to be retained by lipophilic compart- 20. Sun, Q.; Gatto, B.; Yu, C.; Liu, A.; Liu, L. F.; LaVoie,
ment of the body. They may also induce P450 enzymes E. J. J. Med. Chem. 1995, 38, 3638.
21. Kim, J. S.; Gatto, B.; Yu, C.; Liu, A.; Liu, L. F.; LaVoie,
E. J. J. Med. Chem. 1996, 39, 992.
Table 14.
22. Kim, J. S.; Sun, Q.; Gatto, B.; Yu, C.; Liu, A.; Liu, L. F.;
Compound Highest Log 1/C ClogP LaVoie, E. J. Bioorg. Med. Chem. 1996, 4, 621.
23. Kim, J. S.; Yu, C.; Liu, A.; Liu, L. F.; LaVoie, E. J. J.
6 4.8 3.29 Med. Chem. 1997, 40, 2818.
7 7.0 6.08 24. Kim, J. S.; Sun, Q.; Yu, C.; Liu, A.; Liu, L. F.; LaVoie,
8 6.8 7.67 E. J. Bioorg. Med. Chem. 1998, 6, 163.
9 7.2 7.56
25. Rangarajan, M.; Kim, J. S.; Jin, S.; Sim, S.-P.; Liu, A.;
11 7.2 7.43
13 7.4 7.20 Pilch, D. S.; Liu, L. F.; LaVoie, E. J. Bioorg. Med. Chem.
2000, 8, 1371.
 S. B. Mekapati, C. Hansch / Bioorg. Med. Chem. 9 (2001) 2885–2893 2893

26. Rangarajan, M.; Kim, J. S.; Sim, S.-P.; Liu, A.; Liu, L. F.; 29. Hansch, C.; Sinclair, J. F.; Sinclair, P. R. Quant. Struct.-
LaVoie, E. J. Bioorg. Med. Chem. 2000, 8, 2591. Act. Relat. 1990, 9, 223.
27. Abraham, M.; McGowan, J. J. Chromatographica 1987, 30. Hansch, C., Leo, A., Hoekman, D., Exploring QSAR.
23, 243. Hydrophobic, Electronic and Steric Constants; American Chem-
28. Hansch, C.; Björkroth, J. P.; Leo, A. J. Pharm. Sci. 1987, ical Society, Washington D.C., 1995.
76, 663.