Index | Search | Home | Table of Contents

Cragg, G.M., M.R. Boyd, J.H. Cardellina II, M.R. Grever, S. Schepartz, K.M. Snader, and M. Suffness. 1993. The search for new pharmaceutical crops: Drug discovery and development at the National Cancer Institute. p. 161-167. In: J. Janick and J.E. Simon (eds.), New crops. Wiley, New York.

The Search for New Pharmaceutical Crops: Drug Discovery and Development at the National Cancer Institute

Gordon M. Cragg, Michael R. Boyd, John H. Cardellina II, Michael R. Grever, Saul Schepartz, Kenneth M. Snader, and Matthew Suffness

    1. Early Development of Taxol
    2. Current Status of Taxol Development
    3. Alternative Sources of Taxol
  5. Fig. 1
  6. Fig. 2
  7. Fig. 3
  8. Fig. 4

In 1937, the United States National Cancer Institute (NCI) was established with its mission being "to provide for, foster and aid in coordinating research related to cancer." In 1955, NCI set up the Cancer Chemotherapy National Service Center (CCNSC) to promote a cancer chemotherapy program, involving the procurement, screening, preclinical development, and clinical evaluation of new agents. All aspects of drug discovery and preclinical development are now the responsibility of the Developmental Therapeutics Program (DTP), a major component of the Division of Cancer Treatment (DCT). During the past 35 years, over 400,000 chemicals, both synthetic and natural products, submitted by investigators and organizations worldwide, have been screened for antitumor activity, and NCI has played a major role in the discovery and development of many of the available commercial and investigational anticancer agents (Driscoll 1984). Prior to 1960, the screening of natural products was concerned mainly with the testing of fermentation products, but the establishment of an interagency agreement with the United States Department of Agriculture in 1960 for the collection and screening of plants marked the start of a systematic program for the discovery and development of anticancer agents from plant sources. Initially, collections were made in the United States and Mexico, but were later expanded to about 60 countries through both field collections by USDA personnel and procurements from contract suppliers. Between 1960 and the termination of this particular collection program in 1982, some 114,000 extracts of an estimated 35,000 plants were tested for antitumor activity, mainly using the in vivo L1210 and P388 mouse leukemias as the primary screening models. The collection strategy, and the scope and achievements of this program, have been reviewed in depth in earlier reports (Perdue 1976; Suffness and Douros 1979, 1982), and will not be discussed in this paper.

Although a number of plant-derived antitumor compounds, such as taxol (Fig. 1) and camptothecin (Fig. 2), were discovered in the 1970s and subsequently have become of great interest at NCI in the 1990s, the program was discontinued in the early 1980s, since it was perceived that few novel active leads were being isolated from natural sources. Indeed, there was concern that natural products were failing to yield novel agents possessing activity against the more resistant solid tumors. This apparent failure might, however, be attributed more to the nature of the primary screens rather than a deficiency of nature. In 1985, development of a new in vitro human cancer cell line screen, incorporating panels of cell lines representing various solid tumor disease-types (e.g. colon, lung, ovarian), was initiated. This revision of the antitumor screening strategy also was accompanied by the implementation of a new NCI natural products program involving new procurement, extraction, and isolation components. Additionally, the initiation in 1987 of a major new program within NCI for the discovery and development of anti-HIV agents provided yet further impetus and resources for the renewed focus upon natural products.


In September, 1986, three five-year contracts were awarded for plant collections in tropical and sub-tropical regions worldwide. In awarding the contracts, NCI specified that collections should encompass a broad taxonomic diversity, but that emphasis should be given to reputed medicinal plants when reliable information was available. Each contract called for the collection of 1,500 samples of 0.3 to 1.0 kg (dry weight) per year, with different plant parts constituting separate, discreet samples. Each sample is assigned a unique NCI collection number, expressed in the form of a barcode label, which is attached to the cloth sample bag in the field. Detailed documentation of each sample is required, including taxonomy, plant part, date and location of collection, habitat, hazards (e.g. thorny), and, when available, any information concerning medicinal uses and preparations. At least five voucher specimens of each plant species are collected; one is donated to the national herbarium in the country of collection, and one is deposited with the Botany Department of the Museum of Natural History, Smithsonian Institution. Dried plant samples are shipped by air freight to the NCI Natural Products Repository (NPR) in Frederick, Maryland. An import permit has been provided to the contractor maintaining the repository (Program Resources, Inc.: PRI) by the Animal and Plant Health Inspection Service (APHIS) of the United States Department of Agriculture, which has provided excellent support to NCI in facilitating the import of thousands of plant samples. On arrival at the NPR, the samples are stored in large -20°C walk-in freezers for at least 48 h; this freezing of samples is required by APHIS in order to minimize the survival of plant pests and pathogens.

The initial contracts were awarded to the Missouri Botanical Garden, the New York Botanical Garden, and the University of Illinois in Chicago (assisted by the Arnold Arboretum and the Bishop Museum) for collections in Africa and Madagascar, Central and South America, and Southeast Asia, respectively. These contracts were recompeted recently and awarded again to these same organizations for a further five years, beginning September, 1991. In investigating the natural resources of other countries, many of the latter in the developing world, NCI has recognized the need to develop substantive collaborations with the scientific community in countries participating in the collection programs. Measures aimed at promoting such collaborations, and establishing mechanisms for compensation of contributing organizations, groups, or individuals in these countries, have been formulated and incorporated in the NCI Letter of Intent. The Letter of Intent contains both short-term and long-term measures aimed at assuring countries participating in NCI-funded collections of NCI's good intentions (National Cancer Institute, unpublished). In the short-term, NCI periodically invites appropriate officials or scientists from local scientific organizations to visit the drug discovery facilities in Frederick, Maryland to discuss the goals of the drug discovery and development program. When laboratory space and resources permit, suitably-qualified scientists are invited to spend periods of up to a year working with scientists in NCI facilities on research topics of mutual interest to NCI and the collection country organizations. To date, scientists from 14 countries have visited NCI and United States Collection contractor facilities for periods of one to two weeks, while chemists from five countries are carrying out collaborative research projects with chemists in NCI facilities.

As screening test data become available, these are provided to collection contractors for dissemination to interested scientists in countries participating in their collection programs. Each country receives only data obtained from organisms collected within its own borders, and scientists are requested to keep data on active organisms confidential until NCI has had sufficient time to assess the potential for development of new agents from such organisms. The request for confidentiality is related to the possibility of obtaining patents on these new agents; in the event of a patent being licensed to a pharmaceutical company for development and eventual marketing of a drug, NCI will make its best effort to insure that a clause is included in the licensing agreement requiring the company to pay a percentage of the royalties, accruing from sale of the drug, to the country of origin of the organism yielding the drug. This form of compensation is regarded as a potential long-term benefit, since development of a drug to the stage of marketing can take 10 to 20 years from its time of discovery. Another potential benefit to the country of origin is the development of large-scale harvesting or cultivation programs to supply sufficient raw material for bulk production of the drug. In licensing a patent on a new drug to a pharmaceutical company, NCI will require the company to seek, as its first source of supply, the natural product produced in the country of origin. The Letter of Intent already has formed the basis for agreements between research organizations in Madagascar and Tanzania and NCI, and will also be used in establishing formal collaborations with research organizations in countries not formally participating in current NCI contract collections.


Plant samples are transferred from the Natural Products Repository to the Extraction and Grinding Laboratory (EGL; also operated by PRI). Ground material is sequentially extracted at room temperature with a 1:1 mixture of methanol and methylene chloride and water to give organic solvent and aqueous extracts. Five 100 milligram samples of each extract are weighed out into small vials to give aliquots suitable for screening, while the remaining materials are kept as bulk samples suitable for subsequent fractionation and isolation studies, if necessary. All extract samples are assigned discreet NCI sample numbers, and are returned to the NPR for storage at -20°C until requested for screening and further investigation.

Extracts are tested in vitro for selective cytotoxicity against panels of human cancer cell lines representing major disease-types, such as lung, colon, melanoma, ovarian, renal, brain, and leukemia (Boyd 1989). Anti-HIV activity is also determined in vitro by measuring the survival of virus-infected human lymphoblastoid cells in the presence or absence of the extracts (Weislow et al. 1989). Extracts found to exhibit anti-HIV activity or a cytotoxicity profile of interest in the human tumor cell line screen are subjected to bioassay-guided fractionation by a team of NCI chemists and biologists to isolate the pure compounds responsible for the observed activity. In bioassay-guided fractionation, all fractions produced at each stage of the separation are tested for activity in the relevant bioassay, and the subsequent fractionations are performed only on the active fraction(s). This process of fractionation and testing is continued until the pure active component(s) is isolated. The initial plant collection sample (0.3 to 1.0 kg) will generally yield enough extract (10 to 40 g) to permit isolation of the pure, active constituent in sufficient quantity for complete structural elucidation. Subsequent secondary testing and preclinical studies (pharmacology, formulation, toxicology), however, might require gram or even kilogram quantities, depending on the degree of activity and toxicity of the active agent (Driscoll 1984).

As of September, 1991, approximately 26,000 plant samples had been collected, and over 14,000 had been extracted to yield over 28,000 extracts. More than 14,000 of these extracts had been tested in the anti-HIV screen, and about 1,300 had shown some activity. Of these, roughly 1,000 are aqueous extracts, many of which are likely to contain ubiquitous compound-types, such as polysaccharides or tannins, as the active constituents. Such compounds currently are not considered a priority for drug development, and typically are eliminated early in the process of fractionation. Amongst the agents which have been isolated and found to show interesting in vitro anti-HIV activity are prostratin (Fig. 3) from the Samoan medicinal plant, Homalanthus acuminatus, collected by ethnobotanist, Paul Cox (Gustafson et al. 1991), and the michellamines (Fig. 4) from the Cameroon plant, Ancistrocladus abbreviatus (Manfredi et al. 1991).

In order to obtain sufficient quantities of an active agent for early preclinical development, recollections of 5 to 50 kg of the dried plant material are carried out, preferably from the original site of collection. Should the preclinical studies justify development of the agent towards clinical trials, considerably larger amounts of plant material would be required, and data on the distribution and abundance, as well as the drug content of the various plant parts, would need to be collected. In addition, the potential for mass cultivation of the plant would need to be assessed. If problems are encountered due to lack of abundance or inability to adapt the plant to cultivation, a search for alternative sources may be necessary. Other species of the same genus or of closely related genera can be analyzed for drug content, and techniques, such as plant tissue culture, can be investigated. While total synthesis must always be considered as a potential route for bulk production of the active agent, it should be noted that the structures of most bioactive natural products are extremely complex, and bench-scale syntheses generally are not readily adapted to economical large-scale production.

The large-scale production of taxol, which typifies the challenges and problems associated with this phase of drug development, is discussed in the next section.


Early Development of Taxol

Taxus brevifolia Nutt bark was first collected by the USDA in 1962, as part of the exploratory screening program of the NCI CCNSC. An extract was shown to be active in the KB cytotoxicity assay, and the isolation of taxol as the active constituent was reported by Wani et al. (1971). A survey of the various parts of T. brevifolia showed that extracts of the bark were consistently more active than those of other parts, and the bark was selected for further collections. Although taxol exhibited moderate in vivo activity against the P388 and L1210 murine leukemia models, there was little interest in developing it further. Observation of strong activity against the B16 melanoma system in 1974 to 1975, however, revived interest, and, in 1977, it was adopted by NCI as a candidate for preclinical development. In 1978, taxol was shown to exhibit significant activity against several human tumor xenograft systems, including the MX-1 mammary tumor, and, in 1979, Horwitz and Manfredi reported its unique mechanism of action in promoting tubulin polymerization and stabilizing microtubules against depolymerization (Manfredi and Horwitz 1984). Formulation studies were completed in 1980, and toxicology studies initiated. Following completion of toxicology in 1982, approval was granted for INDA filing, and Phase I trials were started in 1983.

Current Status of Taxol Development

Taxol has shown significant clinical activity against refractory ovarian cancer (Rowinsky et al. 1990), and recent clinical trial results indicate good activity against advanced breast cancer. Further trials against these disease-types are in progress using combinations of taxol with other agents, such as cisplatin, adriamycin, and granulocyte colony-stimulating factor (G-CSF). Other ongoing trials include small cell lung and non-small cell lung, colon, head and neck, prostate, and upper gastro-intestinal cancers.

Early preclinical and clinical development required only modest amounts of taxol, and, up to 1990, approximately 4 kg had been isolated. Discovery of its efficacy in the treatment of refractory ovarian cancer increased the demand to over 25 kg per year, which requires processing of 340,000 kg of Taxus brevifolia bark, equivalent to about 38,000 trees. In February, 1991, Bristo-Myers Squibb (BMS) entered into a Cooperative Research and Development Agreement (CRADA) with NCI after being selected in an open competition involving a number of pharmaceutical companies interested in collaborating with NCI in the development of taxol. Under the CRADA, BMS is responsible for the continued production of taxol. Bark collections are continuing in the Pacific Northwest, mainly under the direction of Hauser Northwest, a subsidiary company of Hauser Chemical Research of Boulder, Colorado. Harvesting is permitted only from areas designated for clear-cutting, and bark is not accepted from unauthorized sources. The projected harvest for 1991 is 340,000 kg, sufficient to meet the taxol needs for the treatment of 12,000 patients suffering from ovarian cancer. All the bark collected is designated for production of taxol for use in clinical trials, and is being processed by Hauser Chemical Research using procedures approved by the Food and Drug Administration (FDA).

The bulk production of taxol for clinical use will continue to rely on the bark source for the next 2 to 3 years and, in June, 1991, BMS entered into a cooperative agreement with the USDA Forest Service and the Bureau of Land Management whereby BMS will fund a comprehensive inventory of Taxus brevifolia on Government land. To date, no detailed inventory of this tree has been carried out, but estimates, based on stand information and satellite imagery, have indicated that up to 130 million yew trees occur on 720,000 ha of National Forest lands in Oregon and Washington.

Alternative Sources of Taxol

While the annual demand for taxol to treat patients with ovarian cancer exceeds 25 kg for the United States alone, the promise being shown in the treatment of breast cancer and other serious cancers could result in annual demands exceeding 200 to 300 kg. NCI, BMS and other organizations involved in taxol production fully realize that alternative sources of taxol need to be developed, and intensive studies are being directed to meet these needs. NCI and its Frederick-based Contractor, Program Resources, Inc. (PRI) are collaborating with various organizations in undertaking analytical surveys of Taxus species worldwide. Surveys of T. brevifolia have been completed, or are ongoing, in collaboration with Weyerhaeuser Company, USDAFS, and the Canadian Ministry of Forests. These surveys are identifying high-yielding trees which will be propagated in nurseries to use as seed stock for replanting and mass-cultivation programs.

Analyses of the needles of T. brevifolia have shown that the yield of taxol is generally 5 to 10 times lower than that from bark. Analyses of the needles of other Taxus species and cultivars, however, indicate that the taxol yield is often considerably higher than that from T. brevifolia bark, and, in addition, substantial amounts of key taxol precursors, such as 10-desacetylbaccatin III, are usually present. The ready conversion of these precursors to taxol or related active agents by relatively simple semi-synthetic methods has been achieved by several research groups (Denis et al. 1988; Holton 1990). The availability of millions of ornamental Taxus plants representing a variety of species and cultivars in United States nurseries, as well as the abundant supply of several other wild Taxus species in other countries, makes the isolation of taxol and key taxanes an attractive proposition for long-term bulk production from this renewable resource. The variation of content of these compounds in the needles of different commercial cultivars and wild species is being evaluated, as is the propagation of a number of high-yielding cultivars.

Analytical surveys of needles of a number of Taxus species are being undertaken by NCI and PRI in collaboration with various organizations (names in parentheses). They include T. baccata from the Black Sea-Caucasus region of the Soviet Union (T. Elias, Director, Rancho Santa Ana Botanic Garden), T. canadensis from the Gaspe Peninsula of Quebec (Canadian Ministry of Forests), T. globosa from Mexico (R. Nicholson, Arnold Arboretum), and various Taxus species from the USDA in Beltsville, Maryland (J. Duke). An investigation of T. wallichiana from Himalayan regions is in progress, while BMS is collaborating with Chinese groups in the study of T. chinensis and T. yunnanensis. The potential of nursery cultivars is being explored by Zelenka Nursery in collaboration with Ohio State University and the University of Mississippi; 16,000 kg of dried needles is to be supplied to NCI by this collaborative group through an interagency agreement between NCI and USDA signed in September, 1991. A research agreement was recently signed by BMS and Weyerhaeuser Company for the mass propagation of high-yielding Taxus cultivars, and Weyerhaeuser currently has 500,000 seedlings being grown in their nurseries as mother-stock for this project.

Thus, extensive efforts are well under way to develop the renewable needle resources, but various parameters first need to be established before this resource can totally replace the bark as a source of human-use taxol. Unlike the bark, the method and duration of drying of the needles appears to be critical to achieving optimum yields of taxol, and the drying and storage processes are being studied by a number of groups in collaboration with NCI and BMS. In addition, only taxol derived from the bark has the necessary approval by the Food and Drug Administration for human use. Use of a new source, such as needles, requires development of a new isolation procedure following good manufacturing practices (GMP) for FDA approval, and the new needle-derived taxol also requires FDA approval. While these development and approval processes are not viewed as major obstacles, they will nevertheless take some time to complete. Until then, reliance on the bark as the major source of clinical taxol will remain.

Another potential source receiving attention from research groups worldwide is plant cell culture. USDA has patented a process for production of taxol, and is collaborating with Phyton Catalytic in the development of this process. Success in this area has also been reported by Escagenetics Corporation, and both these efforts are receiving NCI grant support.

There is often the perception that, once the structure of a natural product has been determined, large-scale production will inevitably be achieved by total chemical synthesis. Natural product drugs are often very complex molecules with many chiral centers, and, as such, pose formidable synthetic challenges. Thus, the important plant-derived anticancer drugs, vinblastine and vincristine, are still isolated from the source plant, Catharanthus roseus, despite over 20 years of synthetic endeavor. Likewise, microbially-derived anticancer drugs, such as the bleomycins and daunorubicin, are still produced by fermentation rather than total synthesis. Taxol, with 11 chiral centers, and hence 2048 potential diastereomeric isomers, is no exception, and, while small-scale total syntheses will no doubt be achieved, the development of an economically viable process on the 100 kg scale is likely to present considerable problems. These synthetic endeavors could be productive in another sense, however, since some simpler intermediates might be discovered which are more amenable to large-scale synthesis and retain the desired anticancer properties. In addition, as noted earlier, semisynthetic procedures, involving plant-derived taxol precursors, have already contributed substantially to the ultimate long-term solution of the taxol and taxol analog supply problems.

NCI is playing a key role in all these research endeavors through grant support totaling over $2.5 million annually. In addition to studies of plant genetics and propagation, synthesis, and tissue culture, NCI is supporting research in the area of biosynthesis, formulation, metabolism, and tubulin-binding. BMS is also supporting research directed at improving the supply and production of taxol and key taxanes.

NCI has been criticized for not having addressed the taxol supply problem at a much earlier stage in its development. It should be noted, however, that it was only the observation of good activity against refractory ovarian cancer in 1989 that established confidence in the future of taxol as an efficacious clinical agent. With the future of taxol as a clinical agent now assured, NCI and its CRADA partner, BMS, are vigorously pursuing its development by all possible means.


Fig. 1. Taxol.

Fig. 2. Camptothecin.

Fig. 3. Prostratin.

Fig. 4. Michellamines.

Last update September 9, 1997 aw