June 2005
Chem. Pharm. Bull. 53(6) 611—615 (2005)
611
Topochemical Models for Prediction of Anti-tumor Activity of
3-Aminopyrazoles
Sanjay BAJAJ,a Surinder Singh SAMBI,a and Anil Kumar MADAN*,b
a
School of Chemical Technology, GGS Indraprastha University; Delhi-110 006, India: and b Faculty of Pharmaceutical
Sciences, MD University; Rohtak–124 001, India. Received December 20, 2004; accepted February 28, 2005
Relationship between topochemical indices and inhibition of CDK2/cyclin A by 3-aminopyrazoles was investigated using a data set comprising of 42 3-aminopyrazoles. Three topochemical indices—the Wiener’s topochemical index – a distance based topochemical index, atomic molecular connectivity index – an adjacency based
topochemical index and superadjacency topochemical index – an adjacency-cum-distance based topochemical
index were used for the present investigations. The values of Wiener’s topochemical index, atomic molecular connectivity index and superadjacency topochemical index for each of the 42 compounds comprising the data set
were computed using an in-house computer program. Resultant data was subsequently analyzed and suitable
models were developed after identification of the active ranges. Subsequently, a biological activity was assigned to
each of the compounds using these models, which was then compared with the reported CDK2/cyclin A inhibitory activity. High accuracy of prediction ranging from 86 to 89% was observed using these models.
Key words Wiener’s topochemical index; superadjacency topochemical index; CDK2/cyclin A; 3-aminopyrazole; topochemical
model; anti-tumor agents
Graph theory has been used for modeling chemical structures, that is, chemical compounds, intermediates, changes
and mechanisms.1) Over the past decades graph-based molecular structure descriptors have developed into a quite active
field. In these applications a molecular structure is identified
to a graph, with vertices representing non-H atoms and edges
representing chemical bonds.2) During the last decade the
interest in the problem of computer-aided design of organic
structures with prescribed properties has significantly
increased.3) Today, the information feeding the drug design
effort is increasingly quantitative, based upon recent developments in molecular structure description, combinatorial
mathematics, statistics, and computer simulations.4) In chemistry, a graph represents the topology of a molecule in the
sense that it depicts the pattern of connectedness of atoms in
the molecule, being, at the same time, independent of such
metric aspects of molecular structure as equilibrium distance
between nuclei, bond angles, etc.5) Topological indices or
graph invariants are the single numbers for characterization
of molecular structures.6) In recent years a large number of
topological indices have been reported and utilized for chemical documentation, isomer discrimination, study of molecular complexity, chirality, similarity/dissimilarity, QSAR/
QSPR, drug design and database selection, lead optimization, rational combinatorial library design and for deriving
multilinear regression models.7—10) Though a large number
of topological indices have been reported in literature but
only some of them have been successfully employed in SAR
studies. Notable amongst these are Wiener’s index,11,12)
Hosoya’s index,13,14) Randic’s molecular connectivity index,
c ,15) the higher-order connectivity indices, nc , for the paths
of length n defined by Kier and Hall,16) Balaban’s index,
J,17—20) Zagreb group parameters, M1 and M2,21) eccentric
connectivity index,22—26) eccentric adjacency index27,28) and
superpendentic index.29) Topochemical indices are the topological indices derived from the chemical graphs and modified so as to take into consideration the presence and relative
position of heteroatom. Topochemical indices that have been
reported and used for SAR studies include atomic molecular
∗ To whom correspondence should be addressed.
connectivity index,30) eccentric adjacency topochemical
index,31) Wiener’s topochemical index32,33) and superadjacency topochemical index.34,35)
The importance of cyclin-dependent kinases (CDKs) in
cell cycle regulation, their interaction with oncogenes and
tumor suppressors, and their frequent deregulation in human
tumors, has encouraged an active search for agents capable
of perturbing the function of CDKs.36) Therefore, inhibition
of cyclin-dependent kinases is a theme of major interest in
current anti-cancer agents’ research. Different classes of
chemical inhibitors of these enzymes have been identified
during the past decade and the structural basis of inhibition
has been elucidated by X-ray crystallography studies of one
member of the family, CDK2.37) The CDK2 protein can form
complexes with both cyclins E and A, and it is required for
the G1/S transition and S phase progression and centrosome
duplication.38) One of the key CDK2 substrates is the E2F-1
protein itself, of which the turnover and activity is induced
by phosphorylation by CDK2/cyclin A. This leads to higher
than normal levels of E2F, which have been found to induce
apoptosis.39) The series of compounds used in this study is a
novel class of CDK2/cyclin A inhibitors which was discovered by HTS and have demonstrated tumor growth inhibition
in a xenograft model of human ovarian cancer. Amongst
these, the lead compound 41 has been reported to have
potential to inhibit CDK2/cyclins E and A in an in vivo
setting.39) Therefore this series of compounds have the potential for optimization and development of a compound as drug
for the treatment of human cancers.
In the present study three topochemical indices i.e.
Wiener’s topochemical index – a distance based topochemical
descriptor, atomic molecular connectivity index – an adjacency based topochemical descriptor and superadjacency
topochemical index – an adjacency-cum-distance based
topochemical descriptor have been used for development of
models for prediction of CDK2/cyclin A inhibition by 3aminopyrazoles.
e-mail: madan_ak@yahoo.com
© 2005 Pharmaceutical Society of Japan
612
Vol. 53, No. 6
Methodology
Calculation of Topochemical Indices The Wiener’s Topochemical
Index is defined as the sum of the chemical distances between all the pairs of
vertices in hydrogen suppressed molecular graph, that is
n
1
n
P
2 ∑∑
Wc ⫽
c
ij
i⫽1 j ⫽1
where, Pic is the chemical length of the path that contains the least number
of edges between vertex i and vertex j in the graph G; n is the maximum
possible number of i and j. Wiener’s topochemical index (Wc) can be easily
calculated from the chemical distance matrix of a hydrogen suppressed molecular structure. This matrix is obtained by substituting, row elements corresponding to heteroatom, with relative atomic weight with respect to carbon
atom.32,33)
The Atomic Molecular Connectivity Index (AMCI) is the modification of
path-one molecular connectivity index. It takes into consideration the influence of heteroatom and is defined as the summation of the modified bond
values of adjacent vertices for all edges in the hydrogen suppressed molecular graph. It is denoted by c A and is expressed as
n
χ A⫽
∑ (V V
c
i
c
j )
i⫽1
where, n is the number of vertices, Vic and Vjc are the modified degrees of
adjacent vertices i and j forming the edge {i, j} in a graph G. The modified
degree of a vertex can be obtained from the adjacency matrix by substituting
row element corresponding to heteroatom, with relative atomic weight with
respect to carbon atom.30)
The Superadjacency Topochemical Index is defined as the sum of the
products of the concerned vertex chemical degree and the sum of the adjacent vertex chemical degrees divided by the chemical eccentricity of the
concerned vertex, over all the vertices in the hydrogen suppressed molecular
Ac
graph. It is denoted by ¶ and is expressed as
∫
n
Ac
⫽
S ic ⫽
∑
i⫽1
∑
degVi c* S ic
Eic
degV jc
Where, Sic is the sum of chemical degrees of all vertices (vj), adjacent to vertex i and n is the number of vertices in graph G.
For a molecular graph (G), v1, v2, …, vn are vertices, the number of first
neighbors of a vertex vi is the chemical degree of this vertex and is denoted
by deg(vic). The chemical distance dc (vi, vj | G) between the vertices vi and
vj of G is the length of the shortest path connecting vi to vj. While, chemical
eccentricity Eic of vertex vi, in graph G is the length of the shortest path
from vi to vertex vj that is farthest from vi (Eic⫽max dc (vi, vj); j | G). Superadjacency topochemical index is calculated from the chemical distance matrix (Dc), the chemical adjacency matrix (Ac) and a new matrix, the additive
chemical adjacency matrix (Aa c), obtained by modifying Ac. Chemical distance matrix is utilized for deriving chemical eccentricity while chemical adjacency matrix is utilized for deriving chemical degree of vertices. When
non-zero row elements in chemical adjacency matrix represent the chemical
degree of corresponding vertex in a molecular graph, the matrix may be defined as the additive chemical adjacency matrix. This matrix is utilized for
deriving Sic for the corresponding vertex.34)
Model Development
A dataset38) comprising of 42, 3-aminopyrazoles having
inhibitory activity against CDK2/cyclin A was selected for
the development of model for the present investigation. The
basic structure of these compounds is presented in Fig. 1 and
Fig. 1.
Basic Structures of 3-Aminopyrazoles
various substituents enlisted in Table 1. The data set comprised of both active and inactive compounds. The values of
Wiener’s topochemical index were computed for all the compounds involved in the data set using an in-house computer
program. For the selection and evaluation of range specific
features, exclusive activity ranges were discovered from the
frequency distribution of response level. Resulting data was
analyzed and a suitable model was developed after identification of the active range by maximization of the moving average with respect to the active compounds (⬍35%⫽inactive,
35–65%⫽transitional, ⱖ65%⫽active).40,41) Subsequently,
each analogue involved in the data set was assigned a biological activity using this model which was then compared with
the reported CDK2/cyclin A inhibitory activity. The activity
of these compounds was reported in terms of IC50 (nM). For
the purpose of this study, the compounds having IC50 less
than 100 nM were considered to be active while those having
IC50 (nM) greater than 100 were considered as inactive. The
percentage degree of prediction of a particular range was derived from the ratio of the number of compounds predicted
correctly to the total number of compounds present in that
range. The overall degree of prediction was derived from the
ratio of the total number of compounds predicted correctly to
that of the total number of compounds present in both the
active and inactive ranges.
The aforementioned procedure was similarly adopted
for atomic molecular connectivity index, c A and superadjacency topochemical index, ¶ Ac. The results are summarized in
Table 2.
Results and Discussion
The recognition of cyclin-dependent kinase (CDK)/cyclin
complexes by various cell-cycle regulatory proteins, including certain tumor suppressors and transcription factors,
occurs at least in part through a protein–protein interaction
with a binding groove on the cyclin subunit. Since CDK
function is generally deregulated in tumor cells, blocking of
this recruitment site prevents recognition and subsequent
phosphorylation of CDK substrates and offers a therapeutic
approach towards restoration of checkpoint control in transformed cells.42) 3-Aminopyrazoles have been recently identified as a class of CDK2/cyclin A/E inhibitors and have been
reported to have potential for optimization. One major advantage of this series of compounds is that they are potent
inhibitors of CDK2/cyclin A since they are effective in
nanomolar quantities.
Topological/topochemical models are now considered as
powerful tools for the prediction of physicochemical and
biological properties of molecules. By using graph theoretic
invariants as descriptors, one can utilize a set of well-understood mathematical properties to describe more complex
physicochemical and biological behavior of molecules.32) In
pharmaceutical chemistry such methods are used for screening compounds to be tested for specific activity,43) lead identification and lead optimization8) mainly because these methods are rapid and cost-effective.
In the present study, models using three topochemical indices, viz. Wiener’s topochemical index, AMCI and superadjacency topochemical index have been developed for the prediction of CDK2/cyclin A inhibitory activity of this series of
compounds. The idea behind choosing these three indices
June 2005
613
Table 1. Relationship of Wiener’s Topochemical Index (Wc), AMCI (c A) and Superadjacency Topochemical Index (¶ Ac) with CDK2/Cyclin A Inhibition by
3-Aminopyrazoles
Substituent
CDK2/cyclin A inhibition
Index value
Comp. No.
Assigned
R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Methyl
Methyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclobutyl
Cyclopentyl
Cyclohexyl
Methyl
Ethyl
Propyl
Isopropyl
sec-Butyl
tert-Butyl
Phenyl
Benzyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
R1
Phenyl
Propyl
Propyl
Phenyl
4-Br–C6H4–
4-Cl–C6H4–
4-OMe–C6H4–
4-COOH–C6H4–
4-CONH2–C6H4–
3-Br–C6H4–
3-Cl–C6H4–
3-OMe–C6H4–
3-CF3–C6H4–
2-Br–C6H4–
2-Cl–C6H4–
2-OMe–C6H4–
2,6-diCl–C6H4–
3,4-diCl–C6H4–
4-Br–C6H4–
4-Br–C6H4–
4-Br–C6H4–
4-Br–C6H4–
4-Br–C6H4–
4-Br–C6H4–
4-Br–C6H4–
4-Br–C6H4–
4-Br–C6H4–
4-Br–C6H4–
4-Br–C6H4–
Phenylacetyl
4-CONH2-phenylacetyl
3-OMe-phenylacetyl
4-(2-Pyrrolidin-1-yl)-ethoxy
4-OCF3-phenylacetyl
4-Biphenylacetyl
4-(3-Fluorobiphenyl)acetyl
4-(3-Methylbiphenyl)acetyl
4-(4-Carboxybiphenyl)acetyl
4-(4-Carboxamidobiphenyl)acetyl
4-(2-Thenyl)-phenylacetyl
4-(2-Naphthyl)acetyl
4-(1-Naphthyl)acetyl
cA
Wc
408.86
231.10
352.94
585.70
740.79
709.26
826.20
950.79
949.21
729.79
698.26
804.20
1061.85
718.79
687.26
782.20
794.78
826.78
861.71
984.63
1127.55
535.95
637.87
756.79
741.79
863.71
847.71
1127.55
1318.47
714.62
1128.38
964.62
837.70
1458.10
1596.14
1756.05
1749.06
2223.56
2221.40
1429.56
1228.30
1180.30
6.99
5.45
6.53
8.07
7.89
8.15
8.84
9.23
9.26
7.89
8.15
8.84
9.37
7.93
8.18
8.86
8.30
8.30
8.39
8.89
9.39
6.81
7.35
7.85
7.72
8.26
8.03
9.39
9.87
8.54
9.74
9.32
8.78
10.21
11.51
11.79
11.90
12.67
12.70
10.52
10.51
10.53
¶
Ac
27.18
22.51
26.13
30.04
26.02
27.76
28.35
29.69
29.54
27.68
29.81
30.27
33.95
31.71
32.76
32.99
37.60
33.20
25.44
25.89
25.35
24.36
23.90
23.35
24.81
24.38
25.96
25.35
24.57
27.63
27.30
28.11
29.11
29.98
29.67
30.32
30.72
29.66
29.55
32.14
30.01
32.53
Ac
Wc
cA
¶
⫺
⫺
⫺
⫺
⫺
⫺
⫾
⫾
⫾
⫺
⫺
⫺
⫾
⫺
⫺
⫺
⫺
⫾
⫾
⫾
⫾
⫺
⫺
⫺
⫺
⫾
⫾
⫾
⫹
⫺
⫹
⫾
⫾
⫹
⫹
⫹
⫹
⫹
⫹
⫹
⫹
⫹
⫺
⫺
⫺
⫺
⫺
⫺
⫾
⫾
⫾
⫺
⫺
⫾
⫾
⫺
⫺
⫾
⫺
⫾
⫾
⫾
⫾
⫺
⫺
⫺
⫺
⫺
⫺
⫾
⫹
⫾
⫹
⫾
⫾
⫹
⫹
⫹
⫹
⫹
⫹
⫹
⫹
⫹
⫾
⫺
⫾
⫾
⫾
⫹
⫹
⫾
⫹
⫹
⫾
⫾
⫺
⫾
⫺
⫺
⫺
⫺
⫾
⫾
⫺
⫺
⫺
⫺
⫺
⫺
⫾
⫺
⫺
⫹
⫹
⫹
⫹
⫾
⫹
⫾
⫾
⫹
⫹
⫾
⫾
⫾
Reported
⫺
⫺
⫺
⫺
⫹
⫹
⫺
⫺
⫹
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫹
⫹
⫹
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫹
⫹
⫹
⫹
⫹
⫹
⫹
⫹
⫹
⫹
⫹
⫹
⫹
⫹, active compound; ⫺, inactive compound; ⫾, compound in the transitional range where activity could not be specifically assigned.
was that these indices provide structural information on three
different concepts: Wiener’s topochemical index is based
upon inter-atomic distances and any increase in linearity and
molecular size results in increase in the value of Wiener’s
topochemical index. The atomic molecular connectivity
index, on the other hand, is based upon adjacency or connectivity of atoms with in a molecule. The value of AMCI
increases not only with linearity and molecular size but also
with cyclization. The superadjacency topochemical index is
both adjacency as well as distance based. As a consequence
of this, the value of this index tends to increase with the molecular size, branching as well as with cyclization.
Retrofit analysis of the data in Tables 1 and 2 reveal the
following information with regard to the corresponding indices.
Model Based upon Wiener’s topochemical index:
of total 42 compounds, 29 were classified correctly in
· Out
both the active and inactive ranges. The overall accuracy
·
·
·
of prediction was found to be 86.2% with regard to
CDK2/cyclin A inhibitory activity.
The active range had Wiener’s topochemical index values
greater than 1128. 10 out of 11 compounds in the active
range exhibited CDK2/cyclin A inhibitory activity. The
average IC50 of the correctly predicted compounds was
30.3 nM, indicating high potency of the active range.
The inactive range had index values of ⬍826. 15 out of 18
compounds (83%) in the inactive range were found to be
inactive. Average IC50 value of the inactive range was
found to be 2972 nM.
Presence of a transitional range having Wiener’s
topochemical index values of 757 to 1128 indicated the
gradual change in biological activity. The average IC50 was
614
Vol. 53, No. 6
Table 2.
Models for Prediction of CDK2/Cyclin A Inhibition by 3-Aminopyrazoles
Model index
Wc
cA
¶
Ac
Nature of range in
proposed model
Inactive
Transitional
Active
Inactive
Transition
Active
Lower inactive
Lower transitional
Active
Upper transitional
Upper inactive
Index value
Number of compounds
falling in the range
Percent
accuracy
of range
Average
IC50 (nM)
Total
Correct
18
13
11
15
N.A.
10
83.3
N.A.
90.9
2486.17
2521.92
936.84
2972.27
N.A.
30.30
86.2
⬍8.27
8.27—9.72
⬎9.72
16
15
11
14
N.A.
10
87.5
N.A.
90.9
2756.44
2228.87
936.64
3141.71
N.A.
30.30
88.9
⬍25.35
25.35—27.29
27.30—29.67
29.68—32.74
⬎32.74
9
6
11
11
5
9
N.A.
9
N.A.
4
100
N.A.
81.8
N.A.
80.0
5386.67
1983.00
196.82
337.54
4316.60
5386.67
N.A.
52.22
N.A.
5375.00
88.0
⬍826.00
826.00—1128.00
⬎1128.00
Total
Correct
Overall
accuracy of
prediction
Ac
Wc, Wiener’s topochemical index; c A, atomic molecular connectivity index; ¶ , superadjacency topochemical index.
found to be 2521.92 nM for the compounds in the transitional
range.
Model Based upon Atomic molecular connectivity index
(AMCI):
Out of total 42 compounds, 27 compounds were classified
correctly in both the active and inactive ranges. The overall accuracy of prediction was found to be 88.9% with regard to CDK2/cyclin A inhibitory activity.
The active range had AMCI values greater than 9.72. 10
out of 11 compounds in the active range exhibited
CDK2/cyclin A inhibitory activity. The average IC50 of the
correctly predicted compounds was 30.3 nM, indicating
high potency of the active range.
The inactive range with index values of ⬍8.27 was observed. 14 out of 16 compounds (87.5%) in this range
were found to be inactive. Average IC50 value of the inactive range was found to be 3142 nM.
Presence of a transitional range having AMCI values of
757 to 1128 indicated the gradual change in biological activity. The average IC50 was found to be 2228.87 nM for the
compounds in transitional range.
Model based upon Superadjacency topochemical index:
Out of total 42 compounds, 25 compounds were classified
correctly in both the active and inactive ranges. The overall accuracy of prediction was found to be 88.0% with regard to CDK2/cyclin A inhibitory activity.
The active range had superadjacency topochemical index
values greater than 27.30 to 29.67. 9 out of 11 compound
in the active range exhibited CDK2/cyclin A inhibitory activity. The average IC50 of the correctly predicted compounds was 52.2 nM, indicating high potency of the active
range.
Two inactive ranges with index values of ⬍25.35 for lower
inactive range and ⬎32.74 for upper inactive range were
observed. Though all the nine compounds in the lower inactive range were found to be inactive but only one compound in the upper inactive range exhibited biological activity. Average IC50 values of the order of 5387 and
5375 nM were observed for the lower inactive and upper
inactive ranges respectively.
Active range was ideally bracketed by two transitional
ranges having superadjacency topochemical index values
·
·
·
·
·
·
·
·
of 25.35 to 27.29 and 29.68 to 32.74. Existence of these
transitional ranges indicates gradual change of biological
activity. The average IC50 was found to be 1983 nM for the
compounds in lower transitional range and 337.54 nM for
the compounds in upper transitional range.
All the three topochemical models exhibited high accuracy
of prediction ranging from 86% in case of Wiener’s topochemical index to a maximum of 89% in case of atomic molecular
connectivity index. This percentage has been derived from
the ratio of the total number of compounds predicted correctly with respect to the assigned biological activity, to that
of the total number of compounds present in both the active
and inactive ranges. In the models based upon Wiener’s
topochemical index and AMCI, the active and inactive
ranges are separated by a transitional range. Existence of
such a transitional range is ideal because it simply indicates a
gradual change in biological activity as one proceeds from
active to inactive range and vice versa. The model based
upon the superadjacency topochemical index is different
from the other two models since this model has two inactive
ranges and each one is separated from the active range by a
transitional range. One of the appreciative features of all
these models is the exceptionally high potency of the active
ranges. Further, analysis of the structures of compounds in
the active ranges reveals that the active ranges comprise
mainly of 3-phenylacetamide derivatives. This is in line with
the reported38) structure–activity relationship of these compounds. When compared to the 3-propylamido and 3-benzamido aminopyrazoles (compounds 1—29), 3-phenylacetamido derivatives (compounds 30—42) have been reported
to have potent cellular activities, because the former compounds displayed weak activity in tumor anti-proliferation
test despite good activity in biochemical assay for CDK2/cyclin A.
Conclusion
The results indicate that the proposed models, based upon
topochemical indices, have good predictability. These models
may prove to be highly useful in prediction of activities of
this series of compounds prior to synthesis. These models
offer vast potential for lead optimization and may prove to be
highly beneficial in the development of potent anti-tumor
June 2005
agents.
References
1) Milan R., J. Math. Chem., 9, 97—146 (1992).
2) Klein D. J., Indian J. Chem., 42A, 1264—1269 (2003).
3) Kier L. B., Hall L. H., Med. Chem. Res., 7, 335—339 (1997).
4) Tratch S. S., Lomova O. A., Sukhachev D. V., Palyulin V. A., Zefirov
N. S., J. Chem. Inf. Comput. Sci., 32, 130—139 (1992).
5) Devillers A., Balaban A. T. (eds.), “Topological Indices and Related
Descriptors in QSAR and QSPR,” Gordon & Breach, Reading, 1999,
p. 563.
6) Chang-Yu H., Lu X., J. Chem. Inf. Comput. Sci., 37, 311—315 (1997).
7) Nikolic S., Kovacevic G., Milicevic A., Trinajstic N., Croat. Chem.
Acta, 76, 113—124 (2003).
8) Estrada E., Uriarte E., Curr. Med. Chem., 8, 1573—1588 (2001).
9) Gozalbes R., Doucet J., Derouin F., Curr. Drug Targets Infect. Disord.,
2, 93—102 (2002).
10) Estrada E., Patlewicz G., Uriate E., Ind. J. Chem., 42A, 1315—1329
(2003).
11) Wiener H., J. Am. Chem. Soc., 69, 2636—2638 (1947).
12) Wiener H., J. Am. Chem. Soc., 69, 17—20 (1947).
13) Hosoya H., Bull. Chem. Soc. Jpn., 44, 2332—2337 (1971).
14) Hosoya H., J. Chem. Doc., 12, 181—183 (1972).
15) Randic M. J., Am. Chem. Soc., 97, 6609—6615 (1975).
16) Kier L. B., Hall L. H., “Molecular Connectivity in Structure–Activity
Analysis,” Research Study Press, Letchworth, 1986.
17) Balaban A. T., Chiriac A., Motoc I., Simon Z., Lect. Notes Chem., 15,
22—27 (1980).
18) Balaban A. T., Chem. Phys. Lett., 89, 399—404 (1982).
19) Balaban A. T., J. Chem. Inf. Comput. Sci., 25, 334—343 (1985).
20) Balaban A. T., Filip P., MATCH-Commun. Math. Comput. Chem., 16,
163—190 (1984).
21) Gutman I., Trinajstic N., Chem. Phys. Lett., 17, 535—538 (1972).
22) Sharma V., Goswami R., Madan A. K., J. Chem. Inf. Comput. Sci., 37,
273—282 (1997).
23) Sardana S., Madan A. K., MATCH-Commun. Math. Comput. Chem.,
43, 85—98 (2001).
24) Sardana S., Madan A. K., MATCH-Commun. Math. Comput. Chem.,
45, 35—53 (2002).
25) Sardana S., Madan A. K., J. Comput. Aid. Mol. Des., 16, 545—550
615
(2002).
26) Sardana S., Madan A. K., J. Mol. Model, 8, 258—265 (2002).
27) Gupta S., Singh M., Madan A. K., J. Comput. Aid. Mol. Des., 15,
671—678 (2001).
28) Quigley J. M., Naughton S. M., J. Chem. Inf. Comput. Sci., 42, 976—
982 (2002).
29) Gupta S., Singh M., Madan A. K., J. Chem. Inf. Comput. Sci., 39,
272—277 (1999).
30) Gupta S., Singh M., Madan A. K., J. Mol. Str. (Theochem), 571, 147—
153 (2001).
31) Gupta S., Singh M., Madan A. K., Ind. J. Chem., 42A, 1414—1425
(2003).
32) Bajaj S., Sambi S. S., Madan A. K., J. Mol. Str. (Theochem), 684,
197—203 (2004).
33) Bajaj S., Sambi S. S., Madan A. K., MATCH-Commun. Math. Comput.
Chem., 2004, in press.
34) Bajaj S., Sambi S. S., Madan A. K., QSAR & Comb. Sci., 23, 506—
514 (2004).
35) Bajaj S., Sambi S. S., Madan A. K., Bioorg. Med. Chem. Lett., 15,
467—469 (2005).
36) Ruetz S., Fabbro D., Zimmermann J., Meyer T., Gray N., Curr. Med.
Chem. Anti-Cancer Agents, 3, 1—14 (2003).
37) Furet P., Curr. Med. Chem. Anti-Cancer Agents, 3, 15—23 (2003).
38) Pevarello P., Gabriella M., Amici R., Orsini P., Traquandi G., Corti L.,
Piutti C., Sansonna P., Villa M., Pierce B. S., Pulici M., Giordano P.,
Martina K., Fritzen E. L., Nugent R. A., Casale E., Cameron A.,
Ciomei M., Roletto F., Isacchi A., Fogliatto G. P., Pesenti E., Pastori
W., Marsiglio A., Leach K. L., Clare P. M., Fiorentini F., Varasi M.,
Valupetti A., Warpehosli M. A., J. Med. Chem., 47, 3367—3380
(2004).
39) DeGregori J., Leone G., Miron A., Jakoi L., Nevins J. R., Proc. Natl.
Acad. Sci. U.S.A., 94, 14, 7245—7250 (1997).
40) Bajaj S., Sambi S. S., Madan A. K., Bioorg. Med. Chem., 12, 3695—
3701 (2004).
41) Gupta S., Singh M., Madan A. K., J. Math. Anal. Appl., 275, 386—
401 (2002).
42) McInnes C., Andrews M. J. I., Zheleva D. I., Lane D. P., Fischer P. M.,
Curr. Med. Chem. Anti-Cancer Agents, 3, 57—69 (2003).
43) Natarajan R., Kamalakanan P., Nirdosh I., Ind. J. Chem., 42A, 1330—
1346 (2003).