Index | Search | Home | Table of Contents

Cragg, G.M., J.E. Simon, J.G. Jato, and K.M. Snader. 1996. Drug discovery and development at the National Cancer Institute: Potential for new pharmaceutical crops. p. 554-560. In: J. Janick (ed.), Progress in new crops. ASHS Press, Arlington, VA.

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

Gordon M. Cragg, James E. Simon, Johnson G. Jato, and Kenneth M. Snader

    1. Alternative Methods of Production
    2. The Large-Scale Production of Paclitaxel
    3. NCI Policies for Biomass Production
    4. Sustainable Harvesting of Calophyllum species. Production of (-)-Calanolide B
    5. Feasibility Studies of the Cultivation of Ancistrocladus korupensis
  5. Fig. 1

Plants have formed the basis for the treatment of diseases in traditional medicine systems for thousands of years, and continue to play a major role in the primary health care of about 80% of the world's inhabitants (World Health Organization statistic, Farnsworth et al. 1985). In the area of cancer treatment, many claims have been made for the beneficial effects of plants (Hartwell 1982), though many of these claims may be viewed with some skepticism since cancer, as a specific disease entity, is likely to be poorly defined in terms of folklore and traditional medicine. Nevertheless, the discovery and development of efficacious anticancer agents, such as vinblastine and vincristine isolated from the Madagascar periwinkle, Catharanthus roseus (L.) G. Don, provided convincing evidence that plants could be a source of novel cancer chemotherapeutic agents. While the natural product isolated as the active compound might not be suitable for development as an effective drug, it can provide a suitable lead for conversion into a clinically useful agent. This approach is well illustrated by the development of the anticancer drugs, etoposide, and teniposide, as semisynthetic derivatives of epipodophyllotoxin, isolated from Podophyllum peltatum L. and P. emodii Wall. (Cragg et al. 1993a).


In 1960, the National Cancer Institute (NCI) initiated a plant collection program in collaboration with the United States Department of Agriculture (USDA) (Perdue 1976). During the next twenty one years, over 35,000 plant samples representing some 12,000 to 13,000 species were collected by the USDA, mainly from temperate regions; over 114,000 extracts were tested for antitumor activity, primarily in the in vivo L1210 and P388 mouse leukemia systems. While many active agents belonging to a wide variety of chemical classes were isolated and characterized (Cragg et al. 1993a), few satisfied the stringent requirements for preclinical and clinical development. The major clinically active agents to emerge from this program were paclitaxel isolated from Taxus brevifolia Nutt. and other Taxus species (Suffness 1995), and hycamptamine (topotecan), CPT-11, and 9-aminocamptothecin, all semisynthetic derivatives of camptothecin isolated from Camptotheca acuminata Decne (Wall and Wani 1993). The structures of these compounds are shown in Fig. 1.

In 1986, NCI expanded its program to the collection of plants from tropical and subtropical regions. Contracts were awarded to Missouri Botanical Garden (Africa and Madagascar), New York Botanical Garden (Central and South America), and the Univ. of Illinois at Chicago, assisted by the Arnold Arboretum and the Bishop Museum in Honolulu (Southeast Asia), and were renewed, after open competition, in 1991. Collections have been performed in over 25 countries, with the contractors working in close collaboration with qualified organizations in each source country. The collaboration of the source country organizations and scientists has been indispensable in the procurement of the necessary collection and export permits, in the successful performance of in-field collecting activities and taxonomic identifications, as well as the provision of facilities for the preparation, packaging, and shipment of the samples to the NCI natural products repository in Frederick, Maryland. In turn, the NCI program has provided support for expanded research activities by source country scientists, and the expansion of source country holdings of their flora through deposition of a voucher specimen of each species collected in the national herbarium. Through its Letter of Collection (LOC), the NCI has committed itself to policies of collaboration with source countries in the drug discovery and development process, and fair and equitable compensation in the event of commercialization of a drug developed from a plant collected within their borders (Cragg et al. 1994). A key provision of the LOC is the commitment to utilize source country resources, either through sustainable harvest or cultivation, in the large-scale production of an agent for preclinical and clinical development. In the event that a drug is licensed to a pharmaceutical company for advanced development and commercial production, the successful licensee will be required to seek as its first source of supply the natural resources available from the source country, provided a mutually agreeable fair price can be determined.

To date, over 45,000 plant samples have been collected by the NCI contractors, and over 40,000 have been extracted to yield more than 87,000 organic solvent and aqueous extracts. These extracts are tested in vitro for selective cytotoxicity against panels of human cancer cell lines representing major disease types, including leukemias, breast, central nervous system, colon, lung, ovarian, prostate, and renal cancers (Boyd and Paull 1995), as well as for anti-AIDS activity in a screen comprising human lymphoblastoid cells infected with the live AIDS virus (Bader 1992). Extracts showing significant activity in either screen are subjected to bioassay-guided fractionation aimed at the isolation of the pure, active agents. Of the more than 44,000 extracts tested so far in the in vitro human cancer cell line screen, less than 1% have shown some level of selective cytotoxicity. In some instances, the patterns of differential cytotoxicity have been associated with known classes of compounds such as cardenolides, cucurbitacins, lignans, and quassinoids, but others appear to be new leads which are being investigated further. Over 36,000 extracts have been tested in the anti-AIDS screen, and approximately 10% have exhibited some in vitro activity; however, most of the active extracts are aqueous and, in the majority of cases, the activity has been attributed to the presence of ubiquitous chemotypes, such as polysaccharides and tannins. Such compounds are not a current NCI focus for drug development and are typically eliminated early in the drug discovery process (Cardellina et al. 1993). A number of in vitro active anti-AIDS agents have been isolated and selected for preclinical development. Of these, michellamine B isolated from the leaves of the Cameroon liana, Ancistrocladus korupensis, and the calanolides isolated from Calophyllum species collected in Sarawak, Malaysia, are in advanced preclinical development, and their production is discussed later in this paper.


Alternative Methods of Production

The isolation of an active compound is the first stage in the development of a new agent which might be developed as a drug for advancement to clinical trials and possibly to commercial use (Grever et al. 1992). While the initial plant sample (0.3-1.0 kg) collected generally yields enough extract (10-40 g) to permit the isolation and structural elucidation of the pure, active constituent, subsequent secondary testing and preclinical development might require gram or even kilogram quantities. Approval of an agent for clinical development could require multi-kilogram quantities.

In order to isolate sufficient quantities of an active agent for preclinical development, re-collections of 5 to 200 kg of the dried plant material might be necessary, preferably from the original collection site. Such large re-collections necessitate surveys to determine the abundance and distribution of the plant, as well as the variation in drug content with the season of harvesting. The feasibility of propagation and the potential for mass cultivation of high-yielding phenotypes of the plant would also need to be assessed. If problems are encountered due to the scarcity of the wild plant or inability to adapt it to cultivation, alternative sources need to be sought. Other species of the same genus or closely related genera may be analyzed for drug content, and other biomass production techniques, such as plant tissue and cell culture can be investigated. Another potential route for bulk production of the active agent is total synthesis, but experience has shown that the complex structures of most bioactive natural products require the development of multi-step bench-scale syntheses which often are not readily adapted to economically feasible large-scale production. Thus, despite over 30 years of extensive research into the synthesis and tissue culture production of the commercial anticancer drugs, vinblastine and vincristine, isolation from the source plant, Catharanthus roseus, grown in various regions of the world, remains the most economically viable method of large-scale production.

The Large-Scale Production of Paclitaxel

The development of paclitaxel as an effective drug for the treatment of breast and ovarian cancers illustrates how escalating demands necessitated the development of various methods of biomass production. While the original wild source, the bark of the Pacific yew (Taxus brevifolia), provided adequate supplies of the drug for preclinical and early clinical studies, it was soon apparent that the destructive method of harvesting the bark would not meet the large demands resulting from the observation of clinical efficacy against ovarian cancer. An extensive program was initiated to develop alternative, renewable resources , and the current needs are now being met through the harvesting of needles from wild and cultivated Taxus species; in addition, scale-up production through tissue culture has been developed by the company, Phyton Inc., but whether such in vitro systems will be commercially viable as a source for paclitaxel or as a source of high taxane-yielding plants is as yet unproven. The evolution and solution of the paclitaxel supply problems has been reviewed in detail (Cragg et al. 1993b).

NCI Policies for Biomass Production

In response to the experience gained in the production of paclitaxel, the NCI has implemented policies which permit the study of various methods of biomass production at an early stage of development of a new anticancer or anti-AIDS agent (Cragg et al. 1993b). Through a Master Agreement (MA) mechanism, pools of qualified organizations have been established with expertise in: the large-scale re-collection of source plant materials; the cultivation of source plants; and source plant tissue culture. Allowance has been made for two phases in the cultivation and tissue culture projects, one involving the initiation of pilot-scale studies aimed at exploring the feasibility of techniques for biomass production, and the second involving the application of methods developed in the feasibility studies to large-scale production. When a plant-derived agent is approved for preclinical development, a Master Agreement Order (MAO) Request for Proposals (RFP) for projects in one or more of the above areas may be issued to the relevant pools of MA Holders who then submit technical and cost proposals addressing the particular RFP specifications. An award is made to the MA Holder whose proposal is judged to be best suited to the Government requirements.

Sustainable Harvesting of Calophyllum species. Production of (-)-Calanolide B

Several calanolides have been isolated as potential anti-AIDS drugs from Calophyllum species collected in the rainforest regions of Sarawak, Malaysia. Calanolide A is a novel coumarin isolated from leaves and twigs of Calophyllum lanigerum Miq. var. austrocariaceum (T.C.Whitmore) P.F.Stevens, collected in 1987 as part of the NCI contract with the Univ. of Illinois at Chicago for collections in Southeast Asia (Kashman et al. 1992). Recollections of plant material identified as C. lanigerum from the same general location as the original collection failed to yield substantial quantities of calanolide A required for preclinical development. This isomer has not been detected in any related species analyzed thus far, but an extensive analytical survey has shown that the latex of C. teysmanii var. inophylloide P.F. Stevens collected in the same region yields a related compound, (-)-calanolide B (costatolide) which also shows significant anti-HIV activity (Fuller et al. 1994). Latex collections are made by making small slash wounds in the bark of mature trees. After scraping the latex from the trees, the wounds heal and new collections can be performed. Repeated collections have not affected the health of the trees, and 50 kg has been collected from several hundred trees for the production of sufficient (-)-calanolide B for possible preclinical and clinical development. The latex collections are being performed by the Sarawak State Department of Forests working in collaboration with the NCI collection contractor, the Univ. of Illinois at Chicago. In addition, these two organizations are studying the propagation and cultivation of the source species in Sarawak.

Feasibility Studies of the Cultivation of Ancistrocladus korupensis

In 1987, a sample of the leaves of a liana identified as an Ancistrocladus species was collected in the Korup region of southwestern Cameroon as part of the NCI contract with Missouri Botanical Garden for collections in Africa and Madagascar. Extracts of the leaves exhibited significant in vitro anti-HIV activity, and the dimeric naphthylisoquinoline alkaloid, michellamine B, was isolated as the active agent (Manfredi et al. 1991; Boyd et al. 1994). The plant was later identified as a new species and named Ancistrocladus korupensis D. Thomas & Gereau (Thomas and Gereau 1993). Michellamine B shows in vitro activity against both HIV-1 and HIV-2, and is in advanced preclinical development.

An initial botanical survey indicated that the range of A. korupensis was limited to the Korup National Park and that the vines were found in limited abundance with several related species closely resembling A. korupensis. Fallen leaves collected from the forest floor were found to contain reasonable quantities of michellamine B, and collections of these leaves should provide sufficient quantities of the drug to complete advanced preclinical development. The collection of fallen leaves thus far, has obviated the necessity for large-scale harvest of fresh leaves, which would be most difficult to collect from a liana, and avoids the possible endangerment of a wild species found endemic primarily in a national park. As larger quantities of leaves would be required if any clinical efficacy is observed, the NCI utilized its Master Agreement mechanism to provide a contract with the Center for New Crops & Plant Products, Purdue Univ., to examine the feasibility of cultivating A. korupensis in Cameroon as a potential long-term source of michellamine B. By cultivating high michellamine B yielding plants, the limited wild stands could be preserved and the danger of encroachment into a national park avoided. Purdue Univ. established such a project, working in concert with American scientists such as Roy Gereau and Jim Miller from the Missouri Botanic Garden (the original contracting source which collected this plant), Duncan Thomas from Oregon State Univ., and Cameroon scientists and institutions including the Univ. of Yaounde, the Korup Project, and the World Wide Fund for Nature which is coordinating conservation projects in Korup National Park (Simon et al., 1995). The project is being performed entirely in Cameroon except for the analyses of michellamine B and the entire project staff on site are Cameroonians. This project has the full permission and cooperation of the Cameroon government and is overseen by an Intraministrial Committee for Research on A. korupensis. Working in collaboration with the government of the source country is valuable because at present few legal regulations address bioprospecting of plants for non-timber use in Cameroon (Jato et al. 1996).

This project has several objectives. First, to complete a botanical survey of the region so that the total number of wild vines could be determined as well as to identify high michellamine B yielding vines. To accomplish this objective a rapid and quantitative assay was needed. An NCI contractor, Science Applications International Corporation (formerly Program Resources, Inc.), Fort Detrick, MD, developed a very efficient quantitative assay for michellamine B based upon HPLC. Over 1,000 analyses have already been performed and this cooperation enabled the project team to identify high michellamine B yielding phenotypes (Simon et al. 1995).

More than 1,000 kg of dried leaves from the forest floor were also collected in the first year of this project. Leaf collections were made in different regions by a team of locally-hired and trained leaf collectors. Leaves were air-dried, bulked in sacks and stored. The alkaloids appear very stable and not subject to easy degradation. The bulk collection of leaves provides a buffer of raw material for future NCI preclinical studies. As fallen leaves do contain considerable amounts of michellamine B, this technique can be used on a recurrent basis as a sustainable method to collect leaves without any harvesting or cutting of the wild vines. Such a technique can be used for both native stands as well as for cultivated vines (Thomas et al., 1994).

The geographical areas of collection and the averaged michellamine B content from each of the areas was calculated based upon subsampling of the leaf collection bags. Results indicate that the Rengo Camp and the Ekundo-Kundu areas yielded the highest michellamine values averaging 5.5% and 4.5% (w/w basis), respectively. This was followed by Akpasang, Chimpanzee Camp, Mana River, and the lowest, Ikassa. The very high values from the Rengo Camp and the Ekundo-Kundu sites need to be reconfirmed, but suggest that wild vines from these two areas should also be relatively higher in michellamine B than vines from other areas. The completed botanical survey identified several new populations, though all were still in the Korup region. An estimated 10,000 vines are in the wild, and plants are found at an elevational range of 50-160 m in skeletal highly leached soils and with soil pH about 4.0.

An analysis of 791 samples indicated that the average michellamine B content in leaves was 2.1l% (dry wt.). This distribution reflects all manner of samples that we collected (single leaf, leaf clusters, and leaves of varying ages). The highest levels of michellamine B were found in mature leaves but young, fully expanded pale green leaves also contained michellamine B. Younger leaves and the older, fallen, brown leaves had significantly less alkaloid (Simon et al. 1995). More than 400 individual vines were sampled for michellamine B and variation between vines was significant. Despite problems in sampling procedures, high michellamine B vines were identified and are being vegetatively propagated. Of interest to note is that some samples appeared devoid of michellamine B.

Many A. korupensis seedlings were also collected from the forest by digging the plants up and replanting into polyethylene plastic bags. While lower levels of michellamine B were expected from these young plants, their relative concentrations could be effective markers for high michellamine B. Seedling plants which already exhibit higher michellamine B levels, given the similar age and sampling techniques, may represent a genetic source that will continue to exhibit higher levels of michellamine B over time. Variation in plant growth and in michellamine B content among seedlings was also observed. Plants with the highest michellamine B content (0.6%-0.8% dry wt.) were identified and are also being propagated.

In parallel to the large leaf collection, wild vines were assayed several times for michellamine B content. Samples indicating high values of this dimeric alkaloid were then re-sampled for verification. Vines containing michellamine B contents >3.5% were targeted for propagation. By the end of Year 1, many of the high michellamine B yielding plants were being vegetatively propagated. Methods to improve the vegetative propagation success rate are being investigated. As all work is taking place in the jungle, specialized facilities to propagate and maintain the propagules were needed. Therefore, a medicinal plant nursery was designed and constructed by Purdue and the Korup Project staff at the Korup National Park. This new facility now provides, for the first time, specialized propagation units and a nursery for the Ancistrocladus plants which now comprise our germplasm collection.

Field plots have been established including unreplicated demonstrations which provide initial insight into the growth patterns of the plants once introduced into fields of full-sun, shade, and into darkened forest. Seven field studies using more than 5,000 plants are ongoing--all under varying environmental and field conditions which include cleared areas, secondary growth areas, mature forest, and underneath mature palms. In additon, plant population and fertilizer studies in open cleared areas are underway. Plantings in the forest are situated under varying light conditions. Although under natural conditions the seedlings are found only in shaded conditions, plants growing under full-sun are growing rapidly. The plants appear to be responsive to both sunlight and fertilizer, such as nitrogen. Plants growing in full-sunlight look good and have exhibited robust and large leaves. As the plantings mature, trellis systems or trees which the A. korupensis plants can eventually use for structural support will be needed. This will provide needed background information and test various shading and trellising systems. Both low and high input production systems within agricultural and agroforestry systems are being explored.

The collection, germplasm preservation, and horticultural studies should permit us to develop strategies to collect the leaves and introduce this plant into cultivation, whether in an open field, through enrichment of the jungle with high yielding clones, or in fallow ground. All this is being done in a manner compatible with the needs of the Cameroon people, the Cameroon government and the preservation of the rainforest to which the plant is endemic.

These studies become integral to drug discovery programs for several reasons. First, we need to ensure the availability of the raw plant material. In general, such a discovery is often focused on a wild plant that has not previously been studied--chemically, botanically, or horticulturally. The procurement of the raw plant material is not always easy. In the case of A. korupensis, this plant was not used in local medicine and, thus, is not part of the regional ethnopharmacopia. This suggests that a random screening of flora was successful, since a study only of locally used medicinal plants would never have led to the discovery of this plant. Secondly, as the discovery of new drugs from plants should not infringe upon the natural and undisturbed forest areas to which they may be native, the establishment of partnerships between the countries to which a potential drug plant candidate is native and the country that seeks to develop a drug from the plant is critical. Most tropical countries lack the financial support to initiate and sustain such a preservation and development program.


Natural products and plant-derived products continue to be excellent sources of new drug candidates. A program such as the one that has evolved at the National Cancer Institute can achieve three of the most important goals. It can be successful at helping to protect the rights and benefits of the source country which provides the new bioresource. It can anticipate and reduce the problems of providing adequate biomass supplies for drug studies by encouraging early cultivation and plant tissue culture programs. Finally, it can achieve both of these goals in a manner compatible with the needs of the source country. and still be successful in our third and primary goal, that of discovering new pharmacologically useful drug candidates.


Fig. 1. Clinically active agents identified during plant collection programs conducted by the National Cancer Institute.

Last update August 24, 1997 aw