Introduction
Historically, natural products have provided an endless source of medicine, and despite the diversification of drug discovery technology and reduced funding for natural product–based drug discovery, natural products from plants and other biological sources remain an undiminished source of new pharmaceuticals. Indeed, even though industrial funding specifically allocated for natural product–based drug discovery declined from 1984 to 2003, the percentage of natural product–derived, small-molecule patents has remained relatively unchanged1. A comprehensive review of human drugs introduced since 1981 suggests that, out of 847 small-molecule drugs, 5% were natural products, 
27% were derived from natural products (usually semisynthetically) and the remaining 572 were synthetic molecules2. However, 262 of the synthetic molecules had a natural product–inspired pharmacophore or could be considered natural-product analogs. Natural products continue to make the most dramatic impact in the area of cancer. Of 155 anticancer drugs developed since the 1940s only 27% could not be traced to natural products, with 47% being either a natural product or a direct derivation thereof2. Only one drug, the anticancer compound sorafenib, could be traced to completely de novo combinatorial chemistry2. The above analysis does not include biologics and vaccines, which are by definition derived from nature.
Why plants?
It is often noted that 25% of all drugs prescribed today come from plants3, 4. This estimate suggests that plant-derived drugs make up a significant segment of natural product–based pharmaceuticals. Out of many families of secondary metabolites, or compounds on which the growth of a plant is not dependent, nitrogen-containing alkaloids have contributed the largest number of drugs to the modern pharmacopeia, ranging in effects from anticholinergics (atropine) to analgesics (opium alkaloids) and from antiparasitics (quinine) to anticholinesterases (galantamine) to antineoplastics (vinblastine/vincristine)5. Although not as plentiful as alkaloids in the modern pharmacopeia, terpenoids (including steroids) have made an equally important contribution to human health. They range from Na+/K+ pump-inhibiting cardiac glycosides from Digitalis spp. (recognized as a treatment for congestive heart failure by William Withering in 1775)6, to antineoplastic paclitaxel (Taxol, isolated from the Pacific yew in 1967 by Monroe E. Wall and Mansukh C. Wani)7 to antimalarial artemisinin8 (isolated by Tu Youyou in 1972, a component of Artemisia annua extract used for centuries in traditional Chinese medicine), to anti-inflammatory triptolide (isolated in 1972)9, 10. Phenolics (mostly phenylpropanoids) contributed aspirin and podophyllotoxin to modern medicine. Figure 1 summarizes the most important structural classes of pharmacologically active secondary metabolites from plants. It is important to note, however, that the activity of some natural products has yet to be certified by extensive testing or clinical trials; we anticipate that these cases will benefit from the same rigorous attention as those described below for multicomponent botanical therapeutics (MCBTs).
This overrepresentation of natural product–derived drugs begs the question of whether plant secondary metabolites and related synthetic compounds perform better as drugs than randomly synthesized compounds. Plant natural products, enzymes, receptors and regulatory proteins have common evolutionary roots, originating from a small number of parent molecules. These parent molecules were present in primitive life forms and therefore co-evolved to interact with one another. Although their functions and structures have subsequently diverged, some structural kinship still remains that makes natural products, on average, better ligands for human clinical targets than randomly synthesized compounds. As one example of functional conservation between secondary metabolites, the animal eye pigments lutein and zeaxanthin, which protect the eye from harmful short wave radiation, are the same molecules that protect plant photosynthetic machinery from oxidative damage11. These carotenoids probably evolved from a common biochemical ancestor playing a primordial role in cellular photochemistry. However, animals have lost the ability to make these compounds and now need to acquire them from photosynthetic organisms in the form of food.
Secondary metabolites work together
Before the 20th century medicines relied almost exclusively on multicomponent medicines, obtained from natural sources. In contrast, the modern pharmaceutical industry almost exclusively uses single-ingredient drugs, otherwise known as new chemical entities (NCEs). However, the rate of NCE discovery has slowed down significantly during the last decade12, 13. More and more diseases are being treated with combinations of many single-component drugs. These combination therapies are designed to lower the incidence of resistance or target several pathological processes simultaneously. They are particularly important in treating infectious diseases such as HIV14, tuberculosis15, malaria16 and complex chronic diseases like cancer17 and metabolic syndrome18.
Studies have documented the ability of plant secondary metabolites including quercetin, catechins, reseveratrol and curcumin to potentiate the activity of various cancer drugs and/or other phytochemicals19. In addition, some plant secondary metabolites have been shown to overcome multiple drug resistance in tumors19 or in pathogenic bacteria when used in combination with other natural products or antibiotics. In particular, a number of plant extracts and natural products work synergistically with existing antibiotics, restoring antibiotic activity against resistant strains of bacteria such as Staphylococcus aureus (MRSA) and Escherichia coli.
The ideology of wanting to assign a specific biological activity to a specific compound has slowed acceptance of the idea of multicomponent therapies in Western medicine. Yet, perhaps taking a cue from plants, which have always relied on mixtures of biologically active molecules to defend themselves against diseases and predation, drug combination therapies are gaining popularity. Eastern medical systems, including traditional Chinese and Ayurvedic medicine, are more reliant on natural product mixtures, with all their complex interactions, for disease prevention and treatment. In these systems, multicomponent/multifunctional botanicals are considered better suited to prevent or control complex, pleiotropic diseases than single active ingredient pharmaceuticals. Indeed, knowledge of the healing properties of plants existed before many medicinal and food plants were domesticated. For example, Tripterygium wilfordii (Celastraceae), known as Thundergod vine, has a long history of diverse use in traditional Chinese medicine. For centuries it has been collected in the mountains of southern China. Recently, a root extract from this plant was successfully tested in a US National Institutes of Health–supervised phase 2 clinical study for rheumatoid arthritis9 (Box 1 and Fig. 2). Further evidence for the exciting future of this approach comes from another recent study of Russian tarragon, which was successfully tested against diabetes22 (Box 2 and Fig. 3).
Figure 2: Thundergod vine and its natural products.
(a) Thungergod vine Tripterygium wilfordii grown hydroponically. (b) Structures of the diterpenoids triptolide and tripdiolide, two active components in Thungergod vine.
Full size image (30 KB)Figure 3: A. dracunculus is being grown under strict environmental conditions for the production of medically active compounds being tested for antidiabetic activity.
(a) Structures of six active compounds isolated from A. dracunculus: 4,5-di-O-caffeolquinic acid (1), davidigenin (2), 6-demethoxycapillarison (3), 2',4'-dihydroxy-4-methoxydihydrochalcone (4), 2',4-dihydroxy-4-methoxydihydrochalcone (5) and sakuranetin (6). (b) A. dracunculus growing hydroponically.
Full size image (86 KB)Botanical therapeutics: definitions and regulations
Like the extracts of Russian tarragon and the Thundergod vine, botanical therapeutics are plant-produced compounds used to treat and prevent diseases or maintain health and wellness (Table 1). Although NCE from plants are regulated by the US Food and Drug Administration (FDA) in the same way as synthetic drugs, the regulatory pathways to the approval of MCBTs are less tested and more undefined. A relatively recent addition to the set of regulations on botanical mixtures is the "Guidance for Industry: Botanical Drug Products," published by the FDA Center for Drug Evaluation and Research in 2004 (http://www.fda.gov/cder/Guidance/4592fnl.htm). For the first time, requirements for getting new drug application (NDA) or over-the-counter approval for plant extracts are outlined. According to the guidance, botanical drugs include extracts obtained from plants, algae, or macroscopic fungi. The guidance provides an exemption for botanical drugs with some prior history of human use, allowing them to advance through phase 2 clinical trials with fewer preclinical and toxicological studies than would be required for NCEs. In addition, the guidance does not require full characterization of all extract components or full elucidation of their interactions, and may tolerate some variation in the final composition of the botanical drug. Although botanical drugs are not likely to be protected by the composition of matter patents used to protect NCEs, they can be well protected by method of use or preparation patents. Botanical drugs are considered relatively safe from generic competition because a generic is required to be bioequivalent to the approved botanical drug. This requirement is difficult for a third party manufacturing facility to achieve, as it necessitates additional clinical and toxicological testing. This assumption, however, is yet to be tested, as the first botanical drug, a topical antiviral cream, Polyphenone E, was just approved by the FDA in the fall of 2006.
Botanical dietary supplements are primarily regulated by two sets of guidelines published by the FDA: the "Dietary Supplement and Health Education Act of 1994" and the "Qualified Health Claims Guidance of 2004" (http://www.cfsan.fda.gov/~dms/lab-qhc.html). After the FDA removed ephedra (Ephedra sinica extract) from the supplement market in 2004, concerns of safety, consistency and efficacy of many botanical dietary supplements rose considerably. Unscrupulous and misleading labeling and advertising has further undermined consumer confidence in many botanical supplements. As a result, many companies are choosing the more expensive and time-consuming, generally-regarded-as-safe (GRAS) regulatory path23. Food additives, which require premarket FDA approval, is a category generally reserved for synthetic, semisynthetic or highly purified products such as sweeteners.
The medical or special dietary use food category contains GRAS ingredients and is regulated by the Center for Food Safety and Applied Nutrition, a food division of the FDA. In contrast to dietary supplements, medical foods attracted relatively little attention until recently. According to the FDA's "Regulation of Medical Foods" (http://vm.cfsan.fda.gov/~dms/ds-medfd.html), "a medical food is used under care or supervision of the physician or prescribed by a physician when a patient has special nutrient needs in order to manage a disease or health condition. The label must clearly state that the product is intended to be used to manage a specific medical disorder or condition." MCBTs added to cosmetics (cosmeceuticals) fall outside most of the regulations and allowed claims. The cosmetic industry relies on consumer and selfpolicing, assuming some risk. More information about GRAS ingredients and other categories of botanical therapeutics are presented in Table 1; additional information on the most common US regulatory paths for MCBTs is summarized in Box 3 and Figure 4.
Figure 4: Flowchart depicting the development paths for different categories of MCBTs in the United States.
Time and cost of commercialization are estimates and will vary for individual products.
Full size image (104 KB)Challenges in bringing MCBTs to market
The relative decline in natural product–based drug discovery is often blamed on the advent of high-throughput screening12. Although well paired with combinatorial chemistry, high-throughput screening is not easily adaptable to complex mixtures produced from natural sources. This is mainly due to the high cost per sample, complexity of resupply, difficulty in isolation and characterization of actives, lack of reproducibility and interference from compounds in complex mixtures1, 4.
Nevertheless, comparative analysis of structural diversity in natural-product mixtures and combinatorial libraries suggests that nature still has an edge over synthetic chemistry, despite the fact that combinatorial libraries display more elemental diversity. Indeed, superior elemental diversity does not compensate for the overall molecular complexity, scaffold variety, stereochemical richness, ring-system diversity and carbohydrate constituents of natural-product libraries24. It is generally believed that the complexity, diversity and vast number of plant-produced secondary metabolites will continue to constitute a resource beyond the capacity of current synthetic chemistry24. However, the relatively low cost of combinatorial libraries, the simplicity and speed of isolating and characterizing bioactives from these libraries and their compatibility with high-throughput screening, will continue to represent a challenge to botanical therapeutics discovery, unless a strategy to adapt natural-product extracts to the high-throughput screening format is identified.
'Seed-to-pill' standardization is an even greater challenge for MCBTs that is less applicable to single-ingredient drugs25. Full standardization of MCBTs is possible only when the major bioactive components are known and the range of their allowable fluctuations in the final product is established. Similar characterization and standardization is necessary for potentially hazardous compounds in MCBTs. Unfortunately, this is often not the case for botanical dietary supplements currently on the market, many of which are not standardized at all. Others are standardized on the basis of easy-to-quantify 'marker' compounds, which may not be related to efficacy. Additional functional standardization, based on the direct measurements of biological activity and toxicity in different batches, may be warranted for pharmaceutical-grade products. Such standardization is possible only when the mechanism of action of MCBTs is known, and may involve in vitro enzymatic, binding or transcriptional assays.
Standardization of MCBTs should start with correct botanical identification and prevention of product adulteration by other plant species or chemicals. Consistent cultivation practices and controlled environmental conditions are important for the standardization of the final product because growing conditions can cause significant qualitative and quantitative variations in the composition of plant natural products. In addition to epigenetic factors imposed by the environment, the quality of many MCBTs can be improved by the use of a clonally propagated, disease-free, genetically uniform stock. Inputs for NCE pharmaceutical manufacturing are strictly controlled; in the same way controlling genetic and environmental inputs during the cultivation of green plants is essential for the quality and consistency of MCBTs derived from them.
How to move forward?
Synthetic and computational approaches combined with further advances in microbial natural products are a major focus of current drug discovery efforts. Yet, in many cases, combinatorial and other synthetic approaches may get a good head start if they use plant natural products as templates. Current technologies for NCE discovery from plants heavily favor constitutive compounds, which are present at high concentrations in the source tissues. Prefractionation of extracts before testing, using chromatographic or partitioning methodologies, may improve the chances of detecting minor pharmacologically active components.
An encouraging recent trend within the FDA is a movement to increase safety and efficacy by requiring more scientific justification for MCBTs. The agency believes that scientifically developed, tested and standardized products should enjoy a broader range of FDA allowable health. Unfortunately, the "Qualified Health Claims Guidance of 2004" designed to stimulate good science, was too confusing to be widely accepted. And although the trend for increased safety and scientifically justified claims is also observable in Europe, the European regulatory environment for MCBTs is complex and has not been fully standardized between different EU countries26.
Regulatory issues aside, the biggest bottleneck in the development of novel, safe and effective products for the MCBTs market is the lack of scientific approaches and methodologies to discover, characterize and standardize MCBTs. Also contributing to the bottleneck is the complexity of MCBTs combined with fact that most of the federal and private R&D funding is currently directed toward synthetic NCEs and biologics. However, synthetic NCE-drug discovery would greatly benefit from the development of innovative strategies designed to detect pharmacological activities of minor plant extract components and/or detect secondary metabolites from plants affected by biotic and abiotic stresses.
So which R&D strategies will be most useful for discovering novel MCBTs? Innovative technologies to discover MCBTs, such as high-content screening, are emerging. High-content screening is designed to assess the effect of a treatment on multiple pharmacological targets at cellular or organismic levels4, 27.
For example, rapidly emerging gene-expression RT-PCR arrays28 allow simultaneous detection of the effects of MCBTs on multiple genes involved in a particular disease-related pathway or pathways. These arrays provide a lower throughput, mixture-friendly format (acceptable for MCBTs) and are more predictive of pharmacological activity in an organism. Efforts to establish human disease models and high-content screening methods in Caenorhabditis elegans, Drosophila melanogaster or zebrafish may also lead to discovery of new MCBTs. Screening approaches in zebrafish are particularly promising29. Owing to their small size, abundance of embryos, transparent bodies, ease of compound applications and evolutionary closeness to humans, zebrafish are a particularly useful model for discovering MCBTs and studying their modes of action.
In addition to novel pharmacological screening, better approaches to characterization of molecular interactions in mixtures need to be developed. Recent advances in analytical and structural chemistry instrumentation allow for rapid isolation and structural characterization of bioactives from progressively smaller biological samples30. However, these advances are not very useful for deciphering complex molecular interactions, which significantly affect the efficacy and safety of MCBTs. Even more difficult to study are the interactions between MCBTs, food and drugs. The need to develop multiplex approaches capable of detecting beneficial interactions within mixtures, such as '-omics' and systems biology technologies, is only beginning to be realized and is not yet an established part of the MCBTs discovery process.
Traditional approaches such as bioprospecting, in which samples are collected from natural sources and tested (a 'grind and find' approach), or alternatively, indigenous knowledge helpful in studying and exploiting biochemical or genetic resources is collected, will continue to be important. Indeed, bioprospecting combined with the utilization of indigenous and traditional medical knowledge has been central to the history of the discovery of botanical therapeutics. The Convention on Biological Diversity (CBD) in 1992 (http://www.biodiv.org), which has been signed and approved by nearly every country in the world,with the notable exception of the United States, sought to provide guidelines for bioprospecting. In particular, the CBD asserts full national sovereignty over countries' biological resources and promotes their sustainable use. It states that bioprospecting activities based on random screening or traditional knowledge should provide for the informed consent and equitable benefit sharing with the local people. Unfortunately, the lack of acceptance or full understanding of the CBD has somewhat limited bioprospecting in recent years. However, with less than 1% of bacterial and 5% of fungal species currently known or even culturable, the potential for novel microbial sources of drugs is vast31, and suggests that bioprospecting will remain a significant source of potential new compounds. This potential may be particularly large in extreme, stressed and otherwise unique environments. These drug-discovery-rich hotspots range from geothermal vents to the intracellular spaces of higher plants, populated by endophytic fungi, some of which only make bioactive compounds when growing in association with a plant32. Bioprospecting in the future may reap benefits from increased participation in collection and screening by local scientists in biodiversity-rich regions, limiting the need for transfer of material.
Conclusions
It is not yet clear what role MCBTs will play in the future world of medicine. An important argument favoring MCBTs is that they provide combination therapies which can simultaneously target various elements of human diseases, providing efficacy and safety unmatched by NCEs. This multitargeted 'birdshot' approach may provide a viable alternative to the 'silver bullet' NCE approach better designed to work with one symptom or pathogen at a time. Other advantages of MCBTs may lie in their ability to deliver multiple analogs of a pharmacophore in a single dose, potentially improving overall pharmacokinetic properties and patient response.
We hope that future regulatory policies will favor scientifically designed, validated and standardized MCBTs. Botanical drugs, approved by the FDA, prescribed by physicians, reimbursed by insurers and marketed by pharmaceutical companies are well positioned to play a greater role in medicine. The fact that MCBTs, particularly botanical drugs, can be advanced to clinical testing much faster than their synthetic cousins should allow the pharmaceutical industry to save time and money getting a product into advanced clinical development. According to the FDA, over 200 investigational new drug (IND) applications covering botanical drugs are currently pending, although very few have made it to the NDA stage (S. Chen, FDA, personal communication).
The adage 'an apple a day keeps the doctor away' may stand the test of time. New methods for screening, validation and standardization of MCBTs, a favorable regulatory environment and ongoing improvements in analytical methods for isolation and characterization of actives should sustain the contribution of plants to the health and wellness of the world's population.
