Use of the MolDock re-ranking score also achieved acceptable
results for enrichment of known actives in this study, particularly
in the case of JNK, p38 and PDE4. In the case of COX, exclusion of
scores for arachidonic acid (a substrate) and analogous fatty acids,
all of which showed relatively weak binding, did not result in a
noticeable improvement in EFs.
Docking and pharmacophores are often seen as alternatives,
and their respective characteristics and advantages sometimes
enumerated. LS pharmacophores provide an attractive alternative
to docking, though it is not always clear how they may best be
used. For example, restricting hits to those matching the full
complement of features, though very rapid, can unduly limit
the number of compounds identified, particularly when the number of features is quite large (e.g., above six), since, for obvious
reasons, more complex pharmacophores will pick up a much
more restricted set of compounds than those with fewer
features.
Table 3
Ranking of Chinese formulaic categories according to frequency of target-related herbs fromTable 2in component formulas
Formula category Gp NFP3 MTH Traditional indications Representative herbs (fromTable 2)
Release exterior heat 1 4 5 Fever, headache, stiff neck,
thirst, sore eyes
Mentha haplocalyx, Pueraria lobata, Forsythia suspensa,
Arctium lappa, Schizonopeta tenuifolium
Dispel wind-damp 2 4 5 Rheumatic conditions, pain,
numbness
Ligusticum chuanxiong, Angelica pubescens, Cinnamomum
cassia, Notopterygium incisum, Angelica sinensis
Restrain leakage from
intestines
3 2 5 Diarrhoea, irritable bowel
syndrome, abdominal pain,
lethargy
Myristica fragrans, Psoralea corylifolia, Codonopsis pilosula,
Saussurea lappa, Cinnamomum cassia
Release exterior with head
and neck symptoms
1 3 4 Fever, stiff neck, headache,
nasal discharge
Xanthium sibiricum, Ligusticum chuanxiong, Asarum
sieboldii, Magnolia biondii, Notopterygium incisum
Release exterior–interior
excess
1 2 4 Fever, red sore eyes, nasal
discharge, constipation
Ligusticum chuanxiong, Angelica sinsensis, Magnolia
officinalis, Rheum palmatum
Clear heat, relieve toxicity 4 5 4 Fever, nosebleed, rash,
sores, swelling, pain
Lonicera japonica, Forsythia suspensa, Scrophularia
ningpoensis, Paeonia suffruticosa, Gardenia jasminoides
Invigorate blood, dispel
stasis
5 4 4 Severe pain, abscess, ulcer Ligusticum chuanxiong, Angelica sinensis, Carthamus
tinctorium, Commiphoraspp., Notopterygium incisum
Release wind from skin and
channels
4 1 4 Weeping, itchy red skin
lesions
Schizonopeta tenuifolia, Arctium lappa, Angelica sinensis
Reduce food stagnation 3 3 4 Abdominal pain, nausea,
diarrhoea or constipation
Saussurea lappa, Codonopsis pilosula, Myristica fragrans,
Cyperus rotundus
Release exterior with
interior deficiency
1 4 3.5 Fever, chills, pain and
stiffness, headache, dry
cough
Notopterygium incisum, Angelica pubescens, Cynanchum
atratum, Mentha haplocalyx, Astragalus membranaceus,
Asarum sieboldii
Invigorate blood, treat
traumatic injury
5 3 3.5 Severe pain and swelling
from traumatic injury
Ligusticum chuanxiong, Angelica sinensis, Carthamus
tinctorium, Commiphoraspp., Dracaena cinnabari
Gp = group to which each category (column 1) belongs (see text for details of the five groups to which categories were assigned); NFP3 = no. of formulas with three or more
target-related herbs; MTH = median no. of target-related herbs in these formulas; Traditional indications = signs and symptoms used in Chinese medicine to select a formula
from the relevant category; Representative herbs for each formulaic category are taken fromTable 2.
2216 T. M. Ehrman et al. / Bioorg. Med. Chem. 18 (2010) 2204–2218
Partial pharmacophore search, as used here, can increase both
the number and diversity of compounds selected, but the resulting
hit list, which will contain both strong and weak ligands, cannot be
satisfactorily ranked simply on the basis of LS features. Docking
provides the most obvious solution to this problem, and is likely
to give better results if LS features can be simultaneously incorpo-rated. This will result in identification of compounds which both
satisfy the requirements of LS filtering, and achieve low docking
scores when subject to the same constraints.
In this study, only those compounds satisfying three tests were
identified as potential inhibitors, and considered for further analy-sis. These tests comprised (1) Lipinski ‘rule-of-five’ filtering, to
identify compounds likely to be show favourable absorption, (2)
LS pharmacophore screening, to identify possible mimics of PDB
ligands (which in almost all cases display moderate to strong inhi-bition of their respective targets in vitro), and, finally, (3) docking
(using LS constraints), to further restrict the list of hits to those
with acceptable binding energies. Though docking without
constraints might reveal different binding energies for some
compounds, the interest in this instance was to see how well
compounds could dock in the presence of those pharmacophore-derived constraints which had been responsible for their identifi-cation in the first place.
4.2. Scaffolds and substituents of predicted inhibitors
Table 2indicates that a large number of phytochemical classes
are likely to be involved in inhibition of inflammatory targets. Par-ticularly prominent were phenolics, including lignans and flavo-noids, and the smaller terpenoids, such as monoterpenes, iridoids
and sesquiterpenes.
The larger triterpenes, among the most numerous constituents
in Chinese herbs, were by contrast seldom identified as inhibitors
of any of the targets investigated. As previously mentioned, they
were frequently hit during LS filtering but failed, quite comprehensively, to dock to their respective receptors with appropriate
scores. This may be due to their relatively large size and rigid structure. A number of triterpenes from Chinese herbs are nevertheless
known to inhibit COX (e.g., boswellic acids which are active against
COX-1,
37
and a number of oleanane triterpenes fromPicrorhiza kurroaagainst COX-2
38
). In the case of esculentoside A and its derivatives, however, it has been demonstrated that the introduction of
an aromatic ring greatly enhanced activity,
39
which suggests that
particular patterns of substitution (presumably lacking from the
data used here) may be necessary to ensure good binding. As far
as p38 and JNK are concerned, there is very limited evidence so
far that triterpenes act as inhibitors, though ginsenoside Re has re-cently been reported to inhibit JNK,
40
while for PDE4 a number of
studies indicate that some, including ginsenoside Rg1, inhibit
activity though usually only mildly.
41
There appear to be only lim-ited experimental data, therefore, to indicate that they should have
played a more prominent role in virtual screening.
Concerning the nature of typical substituents that were
observed among high ranking hits, the fact that both prenylated
compounds and small glycosides were moderately enriched is of
some significance. The former have received an increasing amount
of attention in recent years, as prenyl groups increase lipophilicity
of a compound, as well as its affinity both for biological mem-branes, and also protein receptors, which are generally character-ised by hydrophobic pockets.
42
Flavonoids, particularly from the
Fabaceae and Moraceae, show a particularly high diversity of such
compounds, though they are also found in other phenolic classes
and some alkaloids. A number of SARs indicate that prenyl groups
often increase potency, including activity against COX and LOX.
43
The results presented here suggest that they may confer the ability
to bind with greater affinity to other anti-inflammatory targets as
well. Other structurally similar hydrophobic sidechains, such as
butenyl groups, may be equally significant.
The role of small glycosides, which include monoterpenes, iridoids and single ring phenolics, is of further interest. While prenyl
groups may increase the lipophilicity of a molecule and enable it
to bind within a receptor pocket, sugars confer greater capacity
for hydrogen bonding. While less well explored, pharmacologically, than other glycosides such as flavonoid glycosides or triterpene saponins, these compounds characterize a number of very
important herbs in Chinese medicine, such asRehmannia glutinosa,
Paeoniaspp.,Codonopsis pilosula, Eucommia ulmoidesand others.
Many of these are regarded as relatively mild in effect and are
consequently often included in ‘tonic’ formulas.
32
The findings
presented here suggest that it is the possession of a simple sugar,
rather than the nature of the parent aglycone (provided it is of
low molecular weight) which determines the ability of these compounds to bind to multiple protein targets. This may consequently
yield clues as to the mode of action of a variety of Chinese tonic
herbs at the molecular level.
4.3. Multi-target phytochemicals and Chinese medicine
The ‘magic bullet’ paradigm of drug discovery relies on the concept of compounds which bind specifically to a single target and
show a high degree of potency. Of recent interest, however, is
the suggestion that it is non-specific and relatively weak patterns
of activity, by contrast, which may ultimately prove of great importance in drug discovery.
44
Though research on multi-target drugs is
still at an early stage, studies have indicated that certain antipsychotic drugs, for instance, appear to be more effective when several
types of receptor are hit.
45
Further advantages of this approach
may include a reduced risk of adverse effects. An example concerns
memantine, and other low affinity NMDA receptor antagonists,
which are associated with a lower prevalence of side-effects than
high-affinity single target drugs.
46
What of the role that phytochemicals may play in this? Recent
studies point to multi-target activity of a number of well known
plant compounds. These include triterpene gymnemic acids which
đang được dịch, vui lòng đợi..
