3d qsar analysis on arylbenzofuran derivatives as
TRANSCRIPT
-
7/29/2019 3D QSAR Analysis on Arylbenzofuran Derivatives As
1/12
International Journal of PharmTech ResearchCODEN (USA): IJPRIF ISSN : 0974-4304Vol.4, No.4, pp 1691-1702, Oct-Dec 2012
3D QSAR Analysis on Arylbenzofuran derivatives asHistamine H3 Receptor Inhibitors using k Nearest
Neighbor Molecular Field Analysis (KNNMFA)
Sanmati K. Jain* and Sudip Nag
*SLT Institute of Pharmaceutical Sciences, Guru Ghasidas Vishwavidyalaya
(Central University), Bilaspur (Chhattisgarh)-495009, India.
*Corres. author: sanmat i jain72@yahoo .co. inPho ne No . 07752-260027, Mob: +919827088484
Abstract: Three dimensional quantitative structure activity relationship (3D QSAR) analysis using knearest neighbor molecular field analysis (kNN MFA) method was performed on a series of arylbenzofuran
derivatives as histamine h3 receptor inhibitors using molecular design suite (VLifeMDS). This study was
performed with 58 compounds (data set) using sphere exclusion (SE) algorithm and random selection
method for the division of the data set into training and test set. kNN-MFA methodology with step wisevariable selection forward-backward, Simulated Annealing and Genetic Algorithms methods were used for
building the QSAR models. Two predictive models were generated with SW-kNN MFA. The most
predictive model was generated by kNN (stepwise forward backward) using sphere exclusion data selection
method. This model explains good internal (q2
= 0.5445) as well as very good external (Predr2
= 0.9013)
predictive power of the model. The steric descriptors at the grid points, S_2307 and S_2377 plays important
role in imparting activity. This model indicates that two steric descriptors are involved. The kNN-MFA
contour plots provided further understanding of the relationship between structural features of substituted
arylbenzofuran derivatives and their activities which should be applicable to design newer potential
H3 receptor inhibitors.
Keywords: 3D-QSAR, kNN-MFA, h3 receptor inhibitors, arylbenzofuran derivatives.
INTRODUCTION
Histamine is a biogenic amine that influences a
wide range of pathophysiological processes. At
present, four subtypes of histamine G protein-
coupled receptors (GPCRs) are known. H1 and H2receptors are concerned in allergic responses and
gastric acid secretion, respectively. The most
recently discovered H4 receptor is mainly located
on mast cells, eosinophils and lymphoid tissues
and seems to be involved in inflammatory
processes (1).
The histamine H3 receptor was identified in 1983
and was initially described as an autoreceptor,
mainly expressed in the central nervous system
(CNS), regulating histamine biosynthesis and
release from histaminergic neurons (2). Histamine
H3 receptors are expressed in the central nervous
system and to a lesser extent the peripheral
nervous system, where they act as autoreceptors in
presynaptic histaminergic neurons, and also
control histamine turnover by feedback inhibition
of histamine synthesis and release (3).
Subsequently, H3 receptors have also been shown
to act as heteroreceptors on non-histaminergic
neurons, where they inhibit the release of other
neurotransmitters such as acetylcholine, dopamine,
norepinephrine, serotonin and various
neuropeptides (4-9).The high density of H3 receptors in different CNS
areas and their influence on the release of a large
variety of neurotransmitters encouraged wide
pharmacological investigation on their
physiological role and the quest for potential
-
7/29/2019 3D QSAR Analysis on Arylbenzofuran Derivatives As
2/12
Sanmati K. Jain e t a l/Int.J.PharmTech Res.2012,4(4) 1692
therapeutic applications of H3-antagonists in the
treatment of various CNS diseases. Among them,
the most promising ones include attention-deficit
hyperactivity disorders (ADHD), Alzheimersdisease, epilepsy, schizophrenia, obesity and
eating disorders (10-19).Since the discovery of the reference antagonist
thioperamide (20), many classes of potent and
selective H3-antagonists have been reported. The
earliest generation of H3-antagonists was derived
from the endogenous neurotransmitter histamine
and the compounds contained an imidazole ring in
their structures (21,22).
The imidazole-based H3 antagonists such as
ciproxifan, thioperamide are useful tools and
reference standards for pharmacological research,
but they have two drawbacks, which includes:
i) Imidazole-based analogues have poor CNS
penetration (23,24) while low CNS penetration
may be a desirable property in a drug targeted
to the treatment of peripheral diseases.
ii) Inhibit cytochrome-P450 (CYP) drug
metabolizing enzymes, leading to drug-drug
interactions by inhibiting the metabolism of
co-administered drugs (25-29).
iii) Additionally, drugs that inhibit CYP enzymes
have the potential to alter the endogenous
metabolism of important circulating
hormones.For this reason non-imidazoles have been targeted
as potential drug candidates, and several distinct
classes of non-imidazole H3 antagonists (30-36)
have been produced. The increasing interest in the
therapeutic potential of H3 antagonists are driving
current medicinal chemistry efforts to identify
potent, selective, therapeutically efficacious and
safe agents for clinical development.
QSAR analysis is done by using the softwareQSARPlus, a product of VLIfe SciencesTechonologies Pvt. Ltd. Pune (37). The
Quantitative Structure Activity Relationship(QSAR) paradigm is based on the assumption thatthere is an underlying relationship between themolecular structure and biological activity. On thisassumption QSAR attempts to establish acorrelation between various molecular propertiesof a set of molecules with their experimentallyknown biological activity.There are two main objectives for the development
of QSAR:
1) Development of predictive and robust QSAR,
with a specified chemical domain, for
prediction of activity of untested molecules.2) It acts as an informative tool by extracting
significant patterns in descriptors related to the
measured biological activity leading to
understanding of mechanisms of given
biological activity. This could help in
suggesting design of novel molecules with
improved activity profile.
Arylbenzofuran, nucleus containing drugs are one
of the most important H3-receptor antagonist,
which are non-imidazole containing drugs. In the
present QSAR study a series of arylbenzofuran
analogues were selected, they differ in
substituents the phenyl ring, and benzofurannucleus. Binding affinities (Ki) of substituted
arylbenzofurans at rat cortical H3-receptors and
cloned human H3-receptors are reported, values
reported are in nanomolar units. The research is
going on this class of selective inhibitors of
Histamine H3-receptor, where some novel agents
has to be explored which will having good potency
to selectively inhibits the histamine H3-receptor, as
well as very less side effects and toxicity.
EXPERIMENTAL
The Work Station: Workstations are raster
systems in which a computer with a full operating
system and mass storage facility is integrated with
graphical display. All the computational methods
were performed on HCL PC using QSARPlus
version 1.0.
Data sets: In the present study 58 molecules ofarylbenzofuran derivatives were used which were
reported to have Histamine H3-receptor antagonist
activity (38). The activity data hH3 binding
affinity (M), determined by using human and rat
histamine H3-receptor, activity data hH3 have
been converted to the logarithmic scale pKi
(moles) and then used for subsequent QSAR
analysis as the response variables. pKi data saved
as.txt.file. The various substituents of allcompounds along with their actual biological
activities are shown in Table 1.
Methodology: Structure of the molecule drawn in
the 2DDrawapp option in Tool menu of
QSARPlus. Then 2D structures where exported to
QSARPlus window (2D structure converted to 3D
structure). After the conversion, structures are
saved as .mol2 file in QSARPlus 3D window.
Energy minimization is done by using Merck
molecular force field (MMFF) method which
results in the optimization of the geometry of themolecule using the following criteria.
-
7/29/2019 3D QSAR Analysis on Arylbenzofuran Derivatives As
3/12
Sanmati K. Jain e t a l/Int.J.PharmTech Res.2012,4(4) 1693
Force fieldMMFFChargeMMFFMax. no. of cycles10,000Convergence criteria (rms gradients)0.01
Dielectric properties
*Distance dependent function
*Constant1.0*Gradient type analysis1.0
Non-bonded cut off: Electro static20.00; VdW10.00.
Models were generated by k-NN-MFA in
conjunction with stepwise (SW) forward-
backward [Cross correlation0.5; Term selection q2; Variance cut off 0.0; No. of Max.
Neighbours 5; No. of Min. Neighbours 2;Select prediction method Distance basedweighted average], Simulated Annealing
[Maximum temperature-100.0; Minimum
temperature-0.01; Decrease temperature by-10.0;
Iteration at given temperature-5; Terms in model-
4; Pertubation limit-5; Cross correlation limit-1.0;
Term selection criteria-q2; Seed-0; Scaling-
Autoscaling; No. of Max. Neighbours 5; No. ofMin. Neighbours 2; Select prediction method Distance based weighted average] and Genetic
Algorithms [Cross correlation limit-1.0; Cross
over probability-0.9; Mutation probability-0.1;
Population-10; Number of generations-1000; Print
after iterations-100; Convergence ending criteria-
0; Chromosome length-3; Seed-0; Term selection
criteria-q2; Scaling-Autoscaling; No. of Max.Neighbours 5; No. of Min. Neighbours 2;Select prediction method Distance basedweighted average] variable selection methods.
VLife Molecular Design Suite38
(VLifeMDS),
allows user to choose probe, grid size, and grid
interval for the generation of descriptors. The
variable selection methods along with the
corresponding parameters are allowed to be
chosen, and optimum models are generated by
maximizing q2. k-nearest neighbor molecular field
analysis (kNN-MFA) requires suitable alignment
of given set of molecules. This is followed bygeneration of a common rectangular grid around
the molecules. The steric, hydrophobic and
electrostatic interaction energies are computed at
the lattice points of the grid using a methyl probe
of charge +1. These interaction energy values are
considered for relationship generation and utilized
as descriptors to decide nearness between
molecules. The term descriptor is utilized in the
following discussion to indicate field values at the
lattice points. The optimal training and test sets
were generated using the random selection
method. This algorithm allows the construction oftraining sets covering descriptor space occupied by
representative points. Once the training and test
sets were generated, kNN methodology was
applied to the descriptors generated over the grid.
Molecular Allignment: In 3D QSAR first the
alignment of the optimized molecules is done by
using Template based alignment method.
Model development: The activity data were
subjected to k-nearest neighbor molecular field
analysis method for model building.
Data selection: Biological activity taken as
dependent variable and descriptors as independent
variable. Following methods were used for
creation of training and test set.
Random selection method
Sphere Exclusion method
Random selection: In order to construct and
validate the QSAR models, both internally andexternally, the data sets were divided into training
85%-75% (85%, 80% and 75%) of total data set
and test sets 15%-25% (15%, 20% and 25%) of
total data set in a random manner. 05 trials were
run in each case.
Sphere Exclusion method: In this method initially
data set were divided into training and test set
using sphere exclusion method. In this method
dissimilarity value provides an idea to handle
training and test set size. It needs to be adjusted by
trial and error until a desired division of trainingand test set is achieved. Increase in dissimilarity
value results in increase in number of molecules in
the test set.
After the creation of training and test set, Min and
Max value of the test and training set is checked,
using the QSAR tool, if the values are not
following the Min Max, then the training / testset is again set and procedure is repeated. If the
Min Max is following, then the k nearestneighbor molecular field analysis method was
used for model building.
-
7/29/2019 3D QSAR Analysis on Arylbenzofuran Derivatives As
4/12
Sanmati K. Jain e t a l/Int.J.PharmTech Res.2012,4(4) 1694
Table 1: General structure of the compounds of Arylbenzofuran derivatives and their biological
activities (data set of 58 molecules)
ON
R1
R
S.
No.
Compound Benzofuran substituent
(R)
Phenyl Substituent
(R1)
hH3(nM) log (1/
hH3)
1 2 H 4-CN 0.45 9.347
2 4a H 3-CN 0.27 9.569
3 4b H 4-F 3.2 8.495
4 4c H 3-F 2.2 8.658
5 4d H 4-Cl 6.3 8.201
6 4e H 3-Cl 5.0 8.3017 4f H 2-Cl 5.3 8.276
8 4g H 4-CF3 5.7 8.244
9 4h H 3-CF3 3.9 8.409
10 4i H 4-Me 7.9 8.102
11 4j H 3-Me 2.7 8.569
12 4k H 2-Me 4.1 8.387
13 4l H 4-OCF3 7.6 8.119
14 4m H 3-OCF3 11 7.959
15 4n H 4-OMe 9.1 8.041
16 4o H 3-OMe 1.2 8.921
17 4p H 2-OMe 15 7.824
18 4q H 3-Cl,4-Cl 3.6 8.44419 4r H 3-Cl,5-Cl 5.0 8.301
20 4s H 3-Me,4-Me 5.9 8.229
21 4t H 3-Me,5-Me 3.6 8.444
22 4u H 4-COOMe 1.9 8.721
23 4v H 3-C(O)Me 0.084 10.076
24 4w H 4-CH2OH 0.88 9.056
25 4x H 3-CH2OH 0.49 9.310
26 8a H 4-Br 7.7 8.114
27 8b H 4-CN,2-Me 2.0 8.699
28 8c H 4-CN,3-Me 0.28 9.553
29 8d H 4-CN,3-F 0.64 9.194
30 8e 7-F 4-CN 0.43 9.36731 8f 7-Me 4-CN 1.1 8.959
32 8g 6-Me 4-CN 0.77 9.114
33 9 H 3-CH(OH)Me 0.44 9.357
34 10 H 3-C(OH)Me2 0.25 9.602
35 11 H 3-COOH 15 7.824
36 12 H 3-C(O)N(Me)OMe 0.62 9.208
37 13a H 3-C(O)Et 0.23 9.638
38 13b H 3-C(O)CH2CHMe2 0.68 9.168
39 13c H 3-C(O)-(3-F)C6H4 2.6 8.585
40 13d H 3-CHO 0.40 9.398
41 14a H 3-C(=NOH)Me 0.49 9.310
42 14b H 3-(=NOMe)Me 0.42 9.37743 14c H 3-C(=NOEt)Me 3.2 8.495
44 14d H 3-C(=NO-t-Bu)Me 4.3 8.367
45 15a H 4-C(O)-c-Pr 0.26 9.585
46 15b H 3-C(O)-c-Pr 0.21 9.678
-
7/29/2019 3D QSAR Analysis on Arylbenzofuran Derivatives As
5/12
Sanmati K. Jain e t a l/Int.J.PharmTech Res.2012,4(4) 1695
47 16 3-I 4-CN 0.51 9.292
48 17a 3-Cl 4-CN 0.39 9.409
49 17b 3-Cl,6-Cl 4-CN 0.83 9.081
50 18a 3-Br 4-CN 0.29 9.538
51 18b 3-Br 4-CN, 3-Me 1.3 8.886
52 19a 3-Ph 4-CN 21 7.67853 19b 3-(3,5-DiMeC6H3) 4-CN 60 7.222
54 19c 3-(3-Pyridyl) 4-CN 13 7.886
55 19d 3-(2-Furyl) 4-CN 2.8 8.553
56 19e 3-(3-Thienyl) 4-CN 2.1 8.678
57 19f 3-(3(2CHO)thienyl) 4-CN 0.73 9.137
58 20 3-(3(2CH2OH)thienyl) 4-CN 1.9 8.721
Total data set of 58 molecules was used for the present study.
Figure 1: Training and test set data fitness plot for Model-1 (trial-1)
RESULTS AND DISCUSSION
Following statistical parameters were used to
correlate biological activity and molecular
descriptors: n = number of molecules, Vn =
number of descriptors, k = number of nearest
neighbor, df = degree of freedom, r2= coefficient
of determination, q2 = cross validated r2 (by the
leave-one out method), pred_ r2 = r2 for externaltest set, pred_ r2se = coefficient of correlation of
predicted data set. Selecting training and test set
by random selection method, the Unicolumn
statics was performed which is shown in Table 2.
The unicolumn statistical analysis can be
interpreted that the mean and standard deviation
for the training and test sets provide insight into
the relative difference in the mean and point
density distribution of the two sets. The minimum
and maximum values in both the training and test
sets are compared such that the maximum of the
test set should be less than that of the training set.
The minimum of the test set should be greater than
that of the training set, suggesting the interpolative
behavior of the test set (i.e., derived within the
minimummaximum range of the training set)which is prerequisite analysis for further QSAR
study.
-
7/29/2019 3D QSAR Analysis on Arylbenzofuran Derivatives As
6/12
Sanmati K. Jain e t a l/Int.J.PharmTech Res.2012,4(4) 1696
Data fitness plot for training and test set is shown
in Figure 1. kNN-MFA plot is shown in Figure 2
and 3D alignment of the molecules is shown in
Figure 3. Results of 3D-QSAR kNN-MFA
analysis using sphere exclusion and random data
selection methods are shown in Table-2 andTable 4-6 respectively. Result of the most
predictive model generated by kNN (Trial-1,
stepwise forward backward) method using sphere
exclusion data selection method is as follows:
kNN MethodTraining Set Size = 53
Test Set Size = 5
Selected Descriptors:S_2307
S_2377
Statistics:
k Nearest Neighbour= 2
n = 53
Degree of freedom = 50
q2 = 0.5445
q2_se = 0.4220
Predr2 = 0.9013
pred_r2se = 0.1955
Descriptor Range:S_2307 -0.007384 -0.007189
S_2377 -0.00308 -0.002385
This model explains internal (q2
= 0.5445) as well
as very good external (Predr2
= 0.9013) predictive
power of the model. The steric descriptors at the
grid points S_2307 and S_2377 plays important
role in imparting biological activity. This model
indicates that two steric descriptors are involved
suggesting that activity is dominated by steric
interactions. kNN-MFA result plot in which 3D-
alignment of molecules with the important steric
points contributing with model with ranges of
values shown in parenthesis represented in Figure-
2.
Finally, it is hoped that the work presented here
will play an important role in understanding the
relationship of physiochemical parameters with
structure and biological activity. By studying the
QSAR model one can select the suitable
substituent for active compound with maximum
potency.
Figure 2: kNN-MFA stepwise forward backwards result plot: 3D-alignment of molecules with the
important steric points contributing with model with ranges of values shown in parenthesis.
-
7/29/2019 3D QSAR Analysis on Arylbenzofuran Derivatives As
7/12
Sanmati K. Jain e t a l/Int.J.PharmTech Res.2012,4(4) 1697
Table 2: Results of 3D-QSAR analysis using kNN method (sphere exclusion)
kNN resultTRIA
L
DISSIMILARI
TY VALUE*
TEST SET
Stepwise forward-
backward
Simulated Annealing Genetic Algorithms
1 6.50 4f,4q,8a,4t,4l k Nearest
Neighbour= 2n = 53
Degree of freedom =
50
q2 = 0.5445
q2_se = 0.4220Predr2 = 0.9013
pred_r2se = 0.1955
k Nearest
Neighbour= 4n = 53
Degree of freedom =
48
q2 = 0.0894
q2_se = 0.5967Predr2 = -0.0971
pred_r2se = 0.6520
k Nearest Neighbour=
4n = 53
Degree of freedom =
49
q2 = 0.0816
q2_se = 0.5993Predr2 = 0.4961
pred_r2se = 0.4419
2 7.00 4f,4q,8a,18b,4t,4x,4l
k Nearest
Neighbour= 2
n = 51Degree of freedom =
47q2 = 0.5601
q2_se = 0.4203
Predr2 = 0.4191pred_r2se = 0.4119
k Nearest
Neighbour= 3
n = 51Degree of freedom =
46q2 = 0.1898
q2_se = 0.5704
Predr2 = 0.3218pred_r2se = 0.4451
k Nearest Neighbour=
4
n = 51Degree of freedom =
47q2 = 0.0627
q2_se = 0.6135
Predr2 = 0.5667pred_r2se = 0.3558
3 8.00 4d,4f,4o,4q,8a,4t,4x,4g,8c
k Nearest
Neighbour= 2
n = 49
Degree of freedom =
47q2 = 0.3541
q2_se = 0.5096
Predr2 = -0.3366pred_r2se = 0.6471
k Nearest
Neighbour= 4
n = 49
Degree of freedom =
44q2 = 0.3644
q2_se = 0.5055
Predr2 = 0.4201pred_r2se = 0.4262
k Nearest Neighbour=
5
n = 49
Degree of freedom =
45q2 = 0.0799
q2_se = 0.6083
Predr2 = 0.1339pred_r2se = 0.5209
4 8.50 4e,4d,4o,8a,4n,8f,4x,4g,8d,8c,4t,19a
k Nearest
Neighbour= 2
n = 46
Degree of freedom =
43q2 = 0.4779
q2_se = 0.4496
Predr2 = 0.1840
pred_r2se = 0.5766
k Nearest
Neighbour= 2
n = 46
Degree of freedom =
41q2 = 0.1198
q2_se = 0.5837
Predr2 = -0.1785
pred_r2se = 0.6929
k Nearest Neighbour=
3
n = 46
Degree of freedom =
42q2 = 0.1329
q2_se = 0.5794
Predr2 = 0.3848
pred_r2se = 0.5006
5 9.00 4e,4d,4o,8a,18a,4n
,8f,16,4x,4m,8d,8c,8g,4t,4c,2,41,19c
k Nearest
Neighbour= 2n = 40Degree of freedom =
37
q2 = 0.4292
q2_se = 0.4808
Predr2 = 0.0211
pred_r2se = 0.5876
k Nearest
Neighbour= 4n = 40Degree of freedom =
35
q2 = 0.1759
q2_se = 0.5777
Predr2 = -0.3135
pred_r2se = 0.6806
k Nearest Neighbour=
5n = 40Degree of freedom =
36
q2 = 0.0383
q2_se = 0.6241
Predr2 = 0.2357
pred_r2se = 0.5192
*Sphere exclusion data selection method was used in this study
-
7/29/2019 3D QSAR Analysis on Arylbenzofuran Derivatives As
8/12
Sanmati K. Jain e t a l/Int.J.PharmTech Res.2012,4(4) 1698
Table 3: Min-Max table for model (Trial-1, stepwise forward backward)
Table 4: Results of 3D-QSAR analysis using Random selection method (85%)
kNN resultTRIA
L
TEST SET
Stepwise forward-
backward
Simulated Annealing Genetic Algorithms
1 14c,17a,4l,4q,4r,4w,8b,8g,9
k Nearest Neighbour= 2n = 49
Degree of freedom = 44
q2 = 0.6321
q2_se = 0.3913Predr2 = 0.4170
pred_r2se = 0.3606
k Nearest Neighbour= 5n = 49
Degree of freedom = 44
q2 = 0.1560
q2_se = 0.5926Predr2 = 0.5958
pred_r2se = 0.3002
k Nearest Neighbour= 4n = 49
Degree of freedom = 45
q2 = -0.1168
q2_se = 0.6817Predr2 = -0.5758
pred_r2se = 0.5928
2 13a,13d,14a,19a,19d,4r,8a,8e,9
k Nearest Neighbour= 2n = 49
Degree of freedom = 45
q2 = 0.4442
q2_se = 0.4529Predr2 = -0.3759
pred_r2se = 0.8353
k Nearest Neighbour= 5n = 49
Degree of freedom = 44
q2 = 0.0829
q2_se = 0.5818Predr2 = 0.0176
pred_r2se = 0.7058
k Nearest Neighbour= 5n = 49
Degree of freedom = 45
q2 = -0.1484
q2_se = 0.6510Predr2 = 0.2088
pred_r2se = 0.6334
3 13d,14c,19d,4d,4j,4
m,4p,4r,9
k Nearest Neighbour= 2
n = 49
Degree of freedom = 43q2 = 0.7018q2_se = 0.3404
Predr2 = -0.2799pred_r2se = 0.7189
k Nearest Neighbour= 5
n = 49
Degree of freedom = 44q2 = 0.1911
q2_se = 0.5606
Predr2 = 0.1789pred_r2se = 0.5758
k Nearest Neighbour= 5
n = 49
Degree of freedom = 45q2 = -0.0692
q2_se = 0.6446
Predr2 = 0.0974pred_r2se = 0.6037
4 12,14d,18a,4c,4g,4i,
4o,4t,9
k Nearest Neighbour= 4
n = 49
Degree of freedom = 44
q2 = 0.4627
q2_se = 0.4685
Predr2 = -0.6210pred_r2se = 0.6606
k Nearest Neighbour= 5
n = 49
Degree of freedom = 44
q2 = 0.0950
q2_se = 0.6080
Predr2 = -0.8115pred_r2se = 0.6983
k Nearest Neighbour= 5
n = 49
Degree of freedom = 45
q2 = -0.1504
q2_se = 0.6855
Predr2 = 0.4306pred_r2se = 0.3915
5 15b,17a,18a,19d,4d,4q,8c,8e,9
k Nearest Neighbour= 4n = 49
Degree of freedom = 44
q2 = 0.4752
q2_se = 0.4429Predr2 = -0.6089
pred_r2se = 0.9035
k Nearest Neighbour= 3n = 49
Degree of freedom = 44
q2 = 0.0775
q2_se = 0.5872Predr2 = 0.3282
pred_r2se = 0.5839
k Nearest Neighbour= 5n = 49
Degree of freedom = 45
q2 = -0.0979
q2_se = 0.6406Predr2 = -0.0840
pred_r2se = 0.7416
Column
Name
Average Max Min StdDev Sum
Training set 8.8166 10.0760 7.2220 0.6253 467.2790
Test set 8.2794 8.4440 8.1140 0.1638 41.3970
-
7/29/2019 3D QSAR Analysis on Arylbenzofuran Derivatives As
9/12
Sanmati K. Jain e t a l/Int.J.PharmTech Res.2012,4(4) 1699
Table 5: Results of 3D-QSAR analysis using Random selection method (80%)
kNN resultTRIA
L
TEST SET
Stepwise forward-
backward
Simulated Annealing Genetic Algorithms
1 12,14a,14d,17a,19a,19f,4f,4j,4k,4r,8d,9
k Nearest Neighbour= 4n = 46
Degree of freedom = 44
q2 = 0.2104
q2_se = 0.5651
Predr2 = -0.6242
pred_r2se = 0.7247
k Nearest Neighbour= 5n = 46
Degree of freedom = 41
q2 = 0.1169
q2_se = 0.5977
Predr2 = -0.6818
pred_r2se = 0.7374
k Nearest Neighbour= 5n = 46
Degree of freedom = 42
q2 = -0.1244
q2_se = 0.6744
Predr2 = 0.2526
pred_r2se = 0.4916
2 14c,17a,19e,4g,4l,4q,4
r,4w,8b,8e,8g,9
k Nearest Neighbour= 4
n = 46
Degree of freedom = 44
q2 = 0.2400
q2_se = 0.5714
Predr2 = -0.2694
pred_r2se = 0.5290
k Nearest Neighbour= 3
n = 46
Degree of freedom = 41
q2 = 0.0198
q2_se = 0.6489
Predr2 = 0.3838
pred_r2se = 0.3686
k Nearest Neighbour= 4
n = 46
Degree of freedom = 42
q2 = -0.1119
q2_se = 0.6911
Predr2 = -0.8464
pred_r2se = 0.6380
3 12,13a,13d,14a,19a,19
d,4r,4x,8a,8c,8e,9
k Nearest Neighbour= 2
n = 46
Degree of freedom = 42q2 = 0.4480
q2_se = 0.4498
Predr2 = -0.0926
pred_r2se = 0.7353
k Nearest Neighbour= 3
n = 46
Degree of freedom = 41q2 = 0.0990
q2_se = 0.5746
Predr2 = -0.2460
pred_r2se = 0.7852
k Nearest Neighbour= 5
n = 46
Degree of freedom = 42q2 = -0.1782
q2_se = 0.6571
Predr2 = 0.2237
pred_r2se = 0.6198
4 12,14d,18a,19c,4c,4g,
4i,4o,4r,4s,4t,9
k Nearest Neighbour= 5
n = 46
Degree of freedom = 40
q2 = 0.6547
q2_se = 0.3739Predr2 = -0.9512pred_r2se = 0.8038
k Nearest Neighbour= 5
n = 46
Degree of freedom = 41
q2 = 0.0646
q2_se = 0.6155Predr2 = 0.2371pred_r2se = 0.5026
k Nearest Neighbour= 5
n = 46
Degree of freedom = 42
q2 = -0.1603
q2_se = 0.6855Predr2 = 0.2633pred_r2se = 0.4939
5 10,14b,17a,18b,4c,4d,
4m,4o,4t,4x,8d,9
k Nearest Neighbour= 2
n = 46
Degree of freedom = 40
q2 = 0.5952
q2_se = 0.4051
Predr2 = -0.9659pred_r2se = 0.8057
k Nearest Neighbour= 5
n = 46
Degree of freedom = 41
q2 = 0.1104
q2_se = 0.6006
Predr2 = -0.0129pred_r2se = 0.5783
k Nearest Neighbour= 4
n = 46
Degree of freedom = 42
q2 = -0.1540
q2_se = 0.6840
Predr2 = 0.1696pred_r2se = 0.5236
Figure 3: 3D-alignment of molecules.
-
7/29/2019 3D QSAR Analysis on Arylbenzofuran Derivatives As
10/12
Sanmati K. Jain e t a l/Int.J.PharmTech Res.2012,4(4) 1700
Table 6: Results of 3D-QSAR analysis using Random selection method (75%)
kNN resultTRIA
L
TEST SET
Stepwise forward-
backward
Simulated Annealing Genetic Algorithms
1 14c,17a,19e,4g,4i,4j,4l,4
q,4r,4s,4w,8b,8e,8g,9
k Nearest Neighbour= 4
n = 43Degree of freedom = 41
q2 = 0.2249
q2_se = 0.5844
Predr2 = -0.0006pred_r2se = 0.4862
k Nearest Neighbour= 5
n = 43Degree of freedom = 38
q2 = 0.0497
q2_se = 0.6471
Predr2 = -0.4015pred_r2se = 0.5755
k Nearest Neighbour= 5
n = 43Degree of freedom = 39
q2 = -0.1349
q2_se = 0.7072
Predr2 = -0.6110pred_r2se = 0.6170
2 12,14d,17b,18a,18b,19c,4c,4g,4i,4o,4r,4s,4t,8e,9
k Nearest Neighbour= 2n = 43
Degree of freedom = 37
q2 = 0.6428
q2_se = 0.3893
Predr2 = -0.9495
pred_r2se = 0.7431
k Nearest Neighbour= 2n = 43
Degree of freedom = 38
q2 = 0.0871q2_se = 0.6224
Predr2 = -0.7086
pred_r2se = 0.6957
k Nearest Neighbour= 5n = 43
Degree of freedom = 39
q2 = -0.1350q2_se = 0.6940
Predr2 = -0.1094
pred_r2se = 0.5605
3 10,14b,15a,17a,18b,4c,4
d,4m,4o,4t,4x,8a,8d,8f,9
k Nearest Neighbour= 2
n = 43
Degree of freedom = 39q2 = 0.4426
q2_se = 0.4760
Predr2 = 0.0867
pred_r2se = 0.5637
k Nearest Neighbour= 3
n = 43
Degree of freedom = 38q2 = 0.1768
q2_se = 0.5784
Predr2 = 0.0557
pred_r2se = 0.5732
k Nearest Neighbour= 5
n = 43
Degree of freedom = 39q2 = -0.2046
q2_se = 0.6997
Predr2 = 0.0663
pred_r2se = 0.5700
4 10,13d,14a,14b,14d,19c,
20,4c,4d,4k,4q,4w,8b,8d,9
k Nearest Neighbour= 2
n = 43Degree of freedom = 38
q2 = 0.6145
q2_se = 0.4046Predr2 = -0.8133
pred_r2se = 0.7161
k Nearest Neighbour= 4
n = 43Degree of freedom = 38
q2 = 0.1529
q2_se = 0.5998Predr2 = -0.4238
pred_r2se = 0.6345
k Nearest Neighbour= 5
n = 43Degree of freedom = 39
q2 = -0.1799
q2_se = 0.7079Predr2 = 0.1903
pred_r2se = 0.4785
5 11,13c,15a,17a,18b,19a,
19d,19e,4b,4d,4f,4j,4n,8
c,9
k Nearest Neighbour= 2
n = 43
Degree of freedom = 40
q2 = 0.5015q2_se = 0.4390
Predr2 = -0.5217
pred_r2se = 0.7820
k Nearest Neighbour= 5
n = 43
Degree of freedom = 38
q2 = 0.0915q2_se = 0.5927
Predr2 = -0.1297
pred_r2se = 0.6738
k Nearest Neighbour= 5
n = 43
Degree of freedom = 39
q2 = -0.0324q2_se = 0.6318
Predr2 = 0.3867
pred_r2se = 0.4965
CONCLUSION
Three dimensional quantitative structure activityrelationship (3D QSAR) analysis using k nearest
neighbor molecular field analysis (kNN MFA)
method was performed on a series of
arylbenzofuran derivatives as histamine
h3 receptor inhibitors using molecular design suite
(VLifeMDS). This study was performed with 58
compounds (data set) using sphere exclusion (SE)
algorithm and random selection method for the
division of the data set into training and test set.
kNN-MFA methodology with step wise variable
selection forward-backward, Simulated Annealing
and Genetic Algorithms methods were used forbuilding the QSAR models. Two predictive
models were generated with SW-kNN MFA. The
most predictive model was generated by kNN
(stepwise forward backward) using sphere
exclusion data selection method. This model
explains good internal (q2
= 0.5445) as well as
very good external (Predr2 = 0.9013) predictive
power of the model. The steric descriptors at the
grid points, S_2307 and S_2377 plays important
role in imparting activity. This model indicates
that two steric descriptors are involved. The kNN-
MFA contour plots provided further understanding
of the relationship between structural features of
substituted arylbenzofuran derivatives and their
activities which should be applicable to design
newer potential H3 receptor inhibitors.
ACKNOWLEDGEMENTThe authors would like to thanks Head, SLT
Institute of Pharmaceutical sciences, Guru
Ghasidas University, Bilaspur (CG) for providing
the necessary facilities.
-
7/29/2019 3D QSAR Analysis on Arylbenzofuran Derivatives As
11/12
Sanmati K. Jain e t a l/Int.J.PharmTech Res.2012,4(4) 1701
REFERENCES
1. Parsons M.E., Ganellin C.R., Histamine and
its receptors, Br. J. Pharmacol., 2006, 147,
S127-35.
2. Arrang J.M., Garbarg M., Schwartz J.C.,Auto-inhibition of brain histamine release
mediated by a novel class (H3) of histamine
receptor, Nature, 1983, 302, 832-37.
3. Arrang J.M., Garbarg M., Schwartz J.C.,
Autoregulation of histamine release in brain
by presynaptic H3-receptors, Neuroscience,
1985, 15, 553-62.
4. Arrang J.M., Drutel G., Schwartz J.C.,
Characterization of histamine H3 receptors
regulating acetylcholine release in rat
entorhinal cortex, Br. J. Pharmacol., 1995, 114
(7), 1518-22.5. Schlicker E., Betz R., Gthert M., Histamine
H3 receptor-mediated inhibition of serotonin
release in the rat brain cortex, Arch.
Pharmacol., 1988, 337 (5), 588-90.
6. Schlicker E., Fink K., Hinterthaner M.,
Gthert M., Inhibition of noradrenaline release
in the rat brain cortex via presynaptic H3receptors, Arch. Pharmacol., 1989, 340 (6),
633-38.
7. Schlicker E., Fink K., Detzner M., Gthert M.,
Histamine inhibits dopamine release in the
mouse striatum via presynaptic H3 receptors,J. Neural. Transm., 1993, 93(1), 1-10.
8. Clapham J., Kilpatrick G.J., Histamine H3
receptors modulate the release of [3H]-
acetylcholine from slices of rat entorhinal
cortex: evidence for the possible existence of
H3 receptor subtypes, Br. J. Pharmacol., 1992,
107 (4), 919-23.
9. Burgaud J.L., Oudart N., Effect of an
Histaminergic H3 Agonist on the Non-
adrenergic Non-cholinergic Contraction in
Guinea-pig Perfused Bronchioles, J. Pharm.
Pharmacol., 1993, 45(11), 955-58.
10. Schwartz J.C., Arrang J.M., Garbarg M.,
Traicort E., In Psychopharmacology: The
Fourth Generation of Progress; Bloom F.E.,
Kupfer D.J., Eds; Raven: New York, 1995; pp
397-406.
11. Stark H., Schlicker E., Schunack W.,
Developments of histamine H3-receptor
antagonists, Drugs Future, 1996, 21(5), 507.
12. Leurs R., Blandina P., Tedford C.,
Timmerman H., Therapeutic potential of
histamine H3 receptor agonists andantagonists, Trends Pharmacol. Sci., 1998,
19(5), 177-84.
13. Leurs R., Bakker R.A., Timmerman H., de
Esch I., The histamine H3 receptor: from gene
cloning to H3 receptor drugs, Nat. Rev. Drug
Discov., 2005, 4(2), 107-20.
14. Wijtmans M., Leurs R., de Esch I., Histamine
H3 receptor ligands break ground in a
remarkable plethora of therapeutic areas,
Expert Opin. Invest. Drugs, 2007, 16(5), 967-85.
15. Stark H., Arrang J.M., Ligneau X., Garbarg
M., Ganellin C.R., Schwartz J.C., Schunack
W., In King F.D., Oxford A.W., Eds.; Progress
in Medicinal Chemistry; Elsevier: Amsterdam,
2001, Vol. 38, pp 279308.16. Witkin J.M., Nelson D.L., Selective histamine
H3 receptor antagonists for treatment of
cognitive deficiencies and other disorders of
the central nervous system, Pharmacol. Ther.,
2004, 103(1), 1-20.
17. Glase S.A., Robertson D.W., Wise L.D.,
Attention deficit hyperactivity disorder:
Pathophysiology and design of new
treatments, Annu. Rep. Med. Chem., 2002, 37,
11-20.
18. Vohora D., Pal S.N., Pillai K.K., Histamine
and selective H3-receptor ligands: a possible
role in the mechanism and management of
epilepsy, Pharmacol. Biochem. Behav., 2001,
68(4), 735-41.
19. Yoshimatsu H., Itateyama E., Kondou S.,
Tajima D., Himeno K., Hidaka S., KurokawaM., Sakata T., Hypothalamic neuronal
histamine as a target of leptin in feeding
behavior, Diabetes, 1999, 48(12), 2286-91.
20. Arrang J.M., Garbag M., Lancelot J.C.,
Lecomte J.M., Pollard H., Robba M.,
Schunack W., Schwartz J.C., Highly potent
and selective ligands for histamine H3-
receptors, Nature, 1987, 327, 117-23.
21. Barocelli E., Ballabeni V., Caretta A., Bordi
F., Silva C., Morini G., Impicciatore M.,
Pharmacological profile of new thioperamide
derivatives at histamine peripheral H1-, H2-,H3-receptors in guinea-pig, Agents Actions,
1993, 38(3-4), 158-64.
22. Leurs R., Vollinga R.C., Timmerman H., The
medicinal chemistry and therapeutic potentials
of ligands of the histamine H3 receptor, Prog.
Drug Res., 1995, 45, 107-165.
23. Ganellin C.R., Leurquin F., Piripitsi A.,
Arrang J.M, Garbarg M., Ligneau X.,
Schunack W., Schwartz J.C., Synthesis of
Potent Non-imidazole Histamine H3-Receptor
Antagonists, Arch. Pharm. Weinheim, 1998,
331(12), 395-404.24. Liedtke S., Flau K., Kathmann M., Schlicker
E., Stark H., Meier G., Schunack W.,
Replacement of imidazole by a piperidine
moiety differentially affects the potency of
-
7/29/2019 3D QSAR Analysis on Arylbenzofuran Derivatives As
12/12
Sanmati K. Jain e t a l/Int.J.PharmTech Res.2012,4(4) 1702
histamine H3-receptor antagonists, Arch.
Pharmacol., 2003, 367 (1), 43-50.
25. La Bella F.S., Queen G., Glavin G., Durant G.,
Stein D., Brandes L.J., H3 receptor antagonist,
thioperamide, inhibits adrenal steroidogenesis
and histamine binding to adrenocorticalmicrosomes and binds to cytochrome P450,
Br. J. Pharmacol., 1992, 107 (1), 161-64.
26. Yang R., Hey J.A., Aslanian R., Rizzo C.A.,
Coordination of Histamine H3 Receptor
Antagonists with Human Adrenal Cytochrome
P450 Enzymes, Pharmacology, 2002, 66(3),
128-35.
27. Harper E.A., Shankley N.P., Black J.W.,
Characterization of the binding of [3H]-
clobenpropit to histamine H3-receptors in
guinea-pig cerebral cortex membranes, Br. J.
Pharmacol., 1999, 128 (4), 881-90.
28. Alves-Rodrigues A., Leurs R., Wu T., Prell
G., Foged C., Timmerman H., [3H]-thio
peramide as a radioligand for the histamine H3
receptor in rat cerebral cortex, Br. J.
Pharmacol., 1996, 118(8), 2045-52.
29. Ballabeni V., Impicciatore M., Bertoni S.,
Magnanini F., Zuliani V., Vacondio F.,
Barocelli E., CNS access of selected H3-
antagonists: ex vivo binding study in rats,
Inflamm Res., 2002, 51(Suppl.1), S55-6.
30. Cowart M., Altenbach R., Black L., Faghih R.,Zhao C., Hancock A.A., Medicinal Chemistry
and Biological Properties of Non-Imidazole
Histamine H3 Antagonists, Mini Rev. Med.
Chem., 2004;, 4(9), 979-92.
31. Chai W., Breitenbucher J.G., Kwok A., Li X.,
Wong V., Carruthers N.I., Lovenberg T.W.,
Mazur C., Wilson S.J., Axe F.U., Jones T.K.,
Non-imidazole heterocyclic histamine H3
receptor antagonists, Bioorg. Med. Chem.
Lett., 2003, 13(10), 1767-70.
32. Zaragoza F., Stephensen H., Knudsen S.N.,
Pridal L., Wulff B.S., Rimvall K., 1-Alkyl-4-
acylpiperazines as a New Class of Imidazole-
Free Histamine H3 Receptor Antagonists, J.Med. Chem., 2004; 47, 2833-38.
33. Miko T., Ligneau X., Pertz H.H, Ganellin R.,
Arrang J. M., Schwartz J. C., Schunack W.,
Stark H., Novel Nonimidazole Histamine H3Receptor Antagonists: 1-(4-(Phenoxy methyl)benzyl)piperidines and Related Compounds, J.
Med. Chem., 2003, 46, 1523-30.
34. Faghih R., Dwight W., Pan J.B., Fox G.B.,
Krueger K.M., Esbenshade T.A., McVey J.M.,
Marsh K., Bennani Y.L., Hancock A.A.,
Synthesis and SAR of aminoalkoxy-biaryl-4-
carboxamides: novel and selective histamine
H3 receptor antagonists, Bioorg. Med. Chem.
Lett., 2003, 13(7), 1325-28.
35. Apodaca R., Dvorack C.A., Xiao W., Barbier
A.J., Boggs J.D., Wilson S.J., Lovenberg T.,
Carruthers N.I., A New Class of Diamine-
Based Human Histamine H3 Receptor
Antagonists: 4-(Aminoalkoxy)benzylamines,J. Med. Chem., 2003, 46, 3938-44.
36. Stark H., Recent advances in histamine H3/H4receptor ligands, Expert Opin. Ther. Patent,
2003, 13, 851-65.37. VLife QSARPlus version 1.0, Vlife Sciences
Technologies Pvt. Ltd., Pune, India (2004),
www.vlifesciences.com.
38. Gfesser G.A., Faghih R., Bennani Y.L., Curtis
M.P., Esbenshade T.A., Hancock A.A., Cowar
D., Structureactivity relationships ofarylbenzofuran H3 receptor antagonists,
Bioorg. Med. Chem. Lett., 2005, 15(10),
2559-63.
*****