Archive for July, 2010

Abstract— Increasing the odds of hit identification in screening
is of significance for drug discovery. The odds for finding a hit
are closely related either to the diversity of libraries or to the
availability of focused libraries. There are no truly diverse
libraries and it is difficult to design focused libraries without
sufficient information. Hence it is helpful to consider alternative
approaches that can enhance the odds using existing libraries.
Multiple members of a protein family have been considered
collectively in inhibitor design, on the basis of the correlation
between protein families and ligands derived from specific
compound classes. Such a correlation has been exploited in
various drug discovery studies and a general receptor-homologbased
screening scheme may be devised. The feasibility of such a
scheme in enhancing the odds of hit identification is discussed.
Index Terms—Homolgy, inhibitors, screening.
I. INTRODUCTION
High-throughput screening and virtual screening has
been extensively used in drug discovery [1, 2]. The odds
for finding a hit depends on the diversity of compound
libraries used [3]. There is no truly universal set of
representative compounds and the screening has practically
been conducted against subsets of molecules [4, 5x], which
may result in useful hits being missed. This problem was
illustrated recently by Oldenburg [3] in an example of two
highly similar compounds of the steroid family, testosterone
and estrogen, which differ only by a methyl group and a few
double bonds. If only the first is included in a library for
screening against estrogen receptor, estrogen would not be
discovered. The same is true if one starts with estrogen and
tests it against testosterone receptor.
Such a problem is likely of particular concern to
screening for agonist/activator drugs that generally require
more specific structural binding configuration than that of
antagonist/inhibitor drugs. A search of Medline shows that
less than 12% of the publications in drug screening are related
to agonist/activator drugs, which may be partly due to the
difficulty in finding an agonist or activator hit.
Agonist/activator drugs constitute an important drug class. A
search of the therapeutic target database [5] finds 44 targets of
agonist/activator drugs, many of which are important
receptors. Hence, methods for improving the odds of
screening of agonist/activator drugs as well as
antagonist/inhibitor drugs are potentially useful in new drug
discovery.
Increasing the size of libraries in a random fashion may not
always be effective or practical for solving this problem [4].
Drug-like compounds have been found sparsely distributed
through chemistry space [6]. As a result, the design of focused
libraries can be a difficult task without sufficient information.
Therefore it is helpful to consider alternative approaches that
can potentially enhance the odds of hit identification without
relying solely on the simple expansion of existing libraries.
Recent developments in exploiting the correlation between
protein families and ligands from specific compound classes
point to a receptor-homolog-based screening scheme for
improving the odds of hit identification.
II. EXPLOITATION OF THE CORRELATION BETWEEN
PROTEIN FAMILIES AND LIGANDS DERIVED FROM
SPECIFIC COMPUND CLASSES
In the search of inhibitors of specific cyclin-dependent
kinases, multiple members of the kinase family were
considered in a collective manner on the basis of their
common feature of ligand-binding mode [7]. Based on the
known binding mode of purine olomoucine at the ATPbinding
site, compounds were designed from combinatorial
libraries of 2,6,9-trisubstituted purines. Selective inhibitors for
subsets of cyclin-dependent kinases were developed from
these libraries.
The correlation between members of kinase family and
inhibitors derived from specific compound classes has been
shown and exploited in various studies [8]. For instance,
compounds based on quinazoline scaffold were found to
exhibit good structure-activity relationship against EGFR
tyrosine kinase and other related kinases (30-34). Potent ATPbinding
site inhibitors were derived from this scaffold for
EGFR tyrosine kinase (31 or 32), c-erbB2/c-erbB4/EGFR (42,
45, 62), RAF kinase (61), CSF-1R (47, 51), and VEGFR (63-
72), some of which are undergoing clinical trials. Inhibitors
were derived from the phenylamino-pyrimidine class for
PDGFR (80?), PKC-a (75) and EGFR tyrosine kinase (82). A
number of indolocarbazole derivatives were found to be
inhibitors of NGF receptor (210), protein kinase C (215), and
PDGR (213). Pyrazolo[d]pyrimidine derivatives were
designed as inhibitors of LcK (120), v-Src (220), CSF-1R
(47), and EGFR (47).
Increasing The Odds Of Hit Iidentification By
Screening Against Receptor Homologs
Yuzong Chen, Congzhong Cai , Zerong Li , Lianyi Han and Jifeng Wang
H
This correlation also appears in other protein families and
applies to agonists/activators. A few examples are serine
protease and peptide-like inhibitors (x1), nuclear hormone
receptors and steroid agonists (x2), and members of G-protein
coupled receptors and catecholamine agonists (x2).
There have been suggestions of exploiting this relationship
to various drug discovery problems [10]. It was proposed that,
by screening a common and diverse set of small molecule
inhibitors against a set of proteins from a family, specific
structure-activity relationship homology can be derived from
which potential drug discovery targets can be grouped (z1).
Based on the common molecular theme for ligands with a
certain classes of drug targets it was suggested that, in stead of
putting barriers of high-risk targets through expensive screens
of large compound collections, focused libraries of specific
compound classes should be tested first [9]. In order to direct
discovery processes to tractable chemical libraries, potential
targets can be screened from all members of a gene family that
have proven records in drug development and chemistry effort
can then be focused on the most intriguing targets (z2).
A. Screening against of receptor-homologs as a general
scheme?
It is of interest to explore the possibility of further
extending the protein-family-based approach into a more
general receptor-homolog-based scheme for drug screening.
In this scheme, as illustrated in Figure 1, screening is
conducted against a receptor and its homologs which are
defined as proteins of similar sequence in the ligand-binding
domain. These homologs likely share common structural
features at ligand-binding sites and structurally similar ligands
may exist for some of these homologs. If one or more of these
ligands is identified as a hit for the entire homolog group, the
rest may be generated by focused library design based on the
identified hits. The subsequent screening of these focused
libraries against the receptor may lead to the identification of
the specific ligands for that receptor.
In the testosterone-estrogen example, if the same set of
compounds is used for screening against estrogen receptor and
all of its sequence homologs (including testosterone receptor),
testosterone would be identified as a hit for the entire receptor
group. A focused library of steroid analogs may be
constructed based on the framework of testosterone, which
likely include estrogen. A subsequent screening of this
focused library against estrogen receptor would discover
estrogen. The odds of finding an agonist hit are thus
significantly increased.
The feasibility of this scheme depends on the existence of
structurally similar ligands, particularly agonists/activators,
for at least some of the homologs. Structurally similar agonists
are searched for two receptors, estrogen receptor and
adrenoceptor, and their sequence homologs. The identified
homologs of these receptors along with structurally similar
agonists are given in Table 1 and 2 respectively. Certain
homologs are excluded from these two groups which include
orphan receptors with unknown ligand and peptide-agonist
receptors or lipid-agonist receptors with binding sites covering
sections different from that of the respective common binding
sites defined by the estrogen receptor or adrenoceptor [11,12].
Structurally similar agonists are found for 7 out of 8 homologs
of estrogen receptor [13-15] and for 4 out of 8 homologs of
adrenoceptor [16-19] respectively. Moreover, agonist
superficially similar to the other agonists is found for a
homolog of estrogen receptor and that of adrenoceptor
respectively. Thus it appears that compounds of common
structural framework can be found for a substantial portion of
receptor homologs.
III. CONCLUSION
The correlation between protein families and ligands of
common structural theme has been exploited in various drug
discovery studies. A general receptor-homolog-based
screening scheme may be devised based on this correlation.
From the study of specific cases, it appears to be feasible to
use this scheme to enhance the odds of hit identification.
References
[1] Ohlstein, E. H. et al. (2000) Drug discovery in the next
millennium. Annu. Rev. Pharmacol. Toxicol. 40, 177-191
[2] Schneider, G. and Bohm, H-J. (2002) Virtual screening
and fast automated docking methods. Drug Discov. Today 7,
64-70
[3] Oldenburg, K.R. (2001) Optimizing screening
technology: how much to invest? – Reply. Drug Discov.
Today 6, 128-129
[4] Hann, M. and Green, R. (1999) Chemoinformatics – a
new name for an old problem? Curr. Opin. Chem. Biol. 3,
379-383
[4] Chen, X. et al. (2002) TTD: Therapeutic Target
Database. Nucleic Acids Res.30, 412-415
[5] Lipinski, C.A. (2000) Drug-like properties and the
causes of poor solubility and poor permeability. J. Pharmacol.
Toxicol. Methods. 44, 235-249
[6] Gray, N. S. et al. (1998) Exploiting chemical libraries,
structure, and genomics in the search for kinase inhibitors.
Science 281, 533-537
[7] Garcia_Echeverria, C. et al. (2000) ATP site-directed
competitive and irreversible inhibitors of protein kinases.
Med. Res. Rev. 20, 28-57
[8] Thorpe, D.S. (2000) Forecasting roles of combinatorial
chemistry in the age of genomically derived drug discovery
targets. Comb. Chem. High Throughput Screen. 3, 421-436
[9] Caron, P. R. et al. (2001) Chemogenomic approaches to
drug discovery. Curr. Opin. Chem. Biol. 5, 464-470
[10] Costantino, G. and Pellicciari, R. (1996) Homology
modeling of metabotropic glutamate receptors. Structural
motifs affecting binding modes and pharmacological profile of
mGluR1 agonists and competitive antagonists. J. Med. Chem.
39, 3998-4006
[11] Hawtin, S.R. and Wheatley, M. (2000) Critical role of
a subdomain of the N-terminus of the V1a vasopressin
receptor for binding agonists but not antagonists; Functional
rescue by the oxytocin receptor N-terminus. Biochemistry 39,
13524-13533
[12] Steinmetz, A.C.U. et al. (2001) Binding of ligands and
activation of transcription by nuclear receptors. Annu. Rev.
Biophys. Biomol. Struct. 30, 329-359
[13] Lala ,D.S. et al. (1997)Activation of the orphon
nuclear receptor steroidogenic factor-1 by oxysteroids. Proc.
Natl. Acad. Sci. USA 94, 4895-4900
[14] Weatherman, R.V. et al. (1999) Nuclear-receptor
ligands and ligand-binding domains. Annu. Rev. Biochem. 68,
559-581
[15] Rang, H.P. et al. (1999) Pharmacology (4th edn)
(Hunter, L. ed) Churchill livingstone
[16] Arvidsson, L. E. et al. (1984) 8-Hydroxy-2-
(alkylamino) tetralins and related compounds as central 5-
hydroxytryptamine receptor agonists. J. Med. Chem. 27,45-51
[17] Van der Goot, H. and Timmerman, H. (2000)
Selective ligands as tools to study histamine receptors. J. Med.
Chem. 35, 5-20
18] Gerhardt, C. C. et al. (1997) Molecular cloning and
pharmacological characterization of a molluscan octopamine
receptor.Molecular Pharmacology. 51, 293-300

Source: dspace.mit.edu

In its search for new molecules against malaria, MMV has so far supported the screening of more than 5 million compounds for their potential activity against the malaria parasite. Three partners of the MMV-supported early discovery projects have gone a step further and released the data pertaining to the active molecules into the public domain. This bold move will enable scientists the world over to access these data ensuring their antimalarial potential be used to the full.

GlaxoSmithKline (GSK), have worked to screen more than 2 million compounds from their in-house library for activity against Plasmodium falciparum – the most prevalent strain of malaria parasite in Africa. The screen resulted in the identification of more than 13,500 compounds with activity against the parasite. The largest group of compounds with a known mode of action to be identified was the kinase inhibitors, which are currently being explored in two other MMV-supported projects: GSK miniportfolio and Monash University’s Kinase Platform.

The Genomics Institute of the Novartis Research Foundation has screened more than 800,000 compounds, from a range of sources, resulting in more than 5,600 molecules active against the parasite. Four of these chemical series have been selected for further development at Novartis Institute of Tropical Diseases (NITD). One series is now undergoing candidate selection for clinical trials, while the other three are in lead optimization.

The team from St Jude Children’s Research Hospital has screened more than 300,000 unique chemical structures also against P. falciparum, yielding over 1,100 potent and selective drug discovery starting points. In collaboration with Rutgers University in New Jersey the team at St Jude is working to take these starting points to the next level.

These projects bring the total number of promising compounds to over 20,000.

Data from GSK and St Jude’s screening work has been published in Nature, while all data, including those of Novartis, can be found online through the European Bioinformatics Institute (EMBL-EBI). Sharing these data will not only expedite the discovery and development of future antimalarials it has also set a precedent for other actors in neglected disease drug discovery to follow the open access model.

source: mmv.org