New Drugs Awaiting Discovery In Chinese Herbs?

Natural products have always represented a significant, though often underappreciated, resource for the development of new medicines. Even now, in an era dominated by combinatorial chemistry, drugs of plant or microbial origin count for more than 30% of worldwide sales, and natural products have been notably successful in the past in opening up new avenues of exploration and in producing entirely new therapeutic classes. Though plant compounds exhibit enormous structural diversity, only a small proportion of that diversity has been seriously explored for its pharmacological potential so far, and there is therefore little reason to believe that this potential has now run dry. By comparison with other areas of pharmaceutical research, however, the screening of natural products has suffered from a lack of data in an appropriate format. In particular, electronic information on chemical structure, pharmacological activity, specificity against known molecular targets, and traditional uses of the herbs in which such compounds are found has been insufficient, though, in the case of Chinese herbs, an increasing amount of information has become available in recent years.

While such information can serve a wide variety of purposes, it is perhaps in the field of virtual screening that it may have its greatest impact. Structures either can be searched for their similarity to known active compounds or can be docked to receptors of particular molecular targets, providing new information on their therapeutic potential. This may serve the needs of drug development and may also provide new insights into Chinese (and other) systems of herbal medicine, an objective which is of much interest in view of the recent growth of traditional Chinese medicine (TCM) beyond its traditional boundaries and its emergence as a significant contributor to healthcare worldwide, a trend reflected in the increasing number of publications on TCM now available in languages other than Chinese. Given the long history of pharmacological prospecting from nature, Chinese herbs are relative newcomers to the scene but have nevertheless shown considerable promise as a source of new drugs in recent years, including artemisinin (qinghaosu), the antimarial compound derived from Artemisia annua, and the acteylcholinesterase inhibitors, huperzines A and B, from Huperzia serrata (Lycopodium serratum), for the treatment of Alzheimer's disease. Chinese herbs are among the best investigated plants from a chemical perspective, and considerably more information is available on their chemical constitution than for herbs from most other parts of the world. Concurrent with work on the elucidation of chemical structures have been investigations into their pharmacology, some of it now summarized in a number of recent volumes, though these contain relatively little data on targets at the molecular level. Information on molecular targets of plant compounds has however also increased over the past few years, though it is only recently that the first such compilation has been published. Again, this is of particular significance for virtual screening in that the information can be used to identify other phytochemicals which may be expected to show similar behaviour and affords the first opportunity to map the ligand-receptor space of plant compounds and their respective molecular targets. Here, we report on the construction of two new databases of use in research relating to Chinese herbs. The first covers details on chemical constituents of many of the major herbs used in TCM, the other provides information on phytochemical compounds with known activity against a wide variety of targets, among them many of proven or suspected therapeutic significance. In the case of the latter, the information is largely restricted to experimental data on in vitro systems of mammalian origin. For all entries in both databases, structures are provided. It is only in recent years that a sufficient volume of information has accumulated to make this possible, and the data are therefore limited. In the case of Chinese herbs, our database contains details on 8264 compounds found in 240 of the most commonly used herbs in Chinese medicine. A number of these, however, are found in more than one herb, with the number of unique compounds being closer to 7000. For the other database, the number of compounds with known targets is 2597. A total of 78 targets is presently covered. Therapeutic Categories and Phytochemical Classes.Aliphatics are notable for the large proportion of compounds active against COX and LOX as well as a smaller proportion active against targets involved in iNOS expression, another inflammation-related process.Simple phenolics also contain a high proportion of compounds active against COX and LOX, in addition to a smaller number active against protein kinases. While coumarins are spread rather evenly between the different categories, polycyclic aromatics are highly uneven and show a strong association with protein kinases, though targets associated with gene expression are also prominent. Flavonoids and isoflavonoids exhibit similar spectra, though aldose reductase inhibitors (diabetes), which are prominent among flavonoids, are lacking among isoflavonoids, as are HIV inhibitors. Like anthraquinones, tannins have a high proportion of protein kinase inhibitors and also inhibit HIV targets and GPCRs to an appreciable extent. Patterns of association for terpenoid compounds should be treated with caution as, with the possible exception of triterpenes, these are currently under-represented in the BPCD. Diterpenes and steroids, however, both have a high proportion of compounds which affect ion channels. The latter are also notable for the large number of compounds which inhibit cAMP phosphodiesterases. In the case of triterpenes, inhibitors of protein kinases, gene expression, and HIV targets (largely HIV protease), are represented more fully than other categories. Finally, alkaloids are high in compounds which affect ion channels, GPCRs, and neurotransmitter converters (mainly acteylcholinesterase and monoamine oxidase in the latter category), reflecting their important effects on the nervous system. A smaller number influence gene expression and inflammation. Further data are required before more detailed statistical analyses can be undertaken, particularly in the case of terpenoid compounds. ConclusionsSources of information on plant compounds which integrate chemical structure, pharmacology, molecular targets, and botanical sources including patterns of traditional usage, are of interest both in terms of drug discovery and in furthering our understanding of herbal medicine.Though there are considerable relevant data available, the great majority are not in a format suitable for data mining and molecular modelling. Recent years have seen improvements in this respect, though much still remains to be done. The work outlined here represents one such contribution to this effort. Information concerning the use and availability of the databases reported here is available from the authors on request. (Source: Journal of Chemical Information and Modelling : King's College London : January 2007.)