Reprinted from the Winter 1970 issue of t'he THE AMERICA HORTICULTURAL \t{AGAZIl\'E Copyright 1970 by The American Horticultural Society, Inc. Cephalotaxus­ Source of Harringtonine, A Promising New Anti..Cancer Alkaloid ROBERT E. PERDUE, JR.,l LLOYD A. SPETZMAN,l and RICHARD G. POWELL2 The plumyews (Cephalotaxus) are yew-like evergreen trees and shrubs. The genus includes seven species native to southeastern Asia from Japan and Korea to Taiwan and Hainan, and west through China to northeastern India. Two species are in cultivation in the United States, C. harringtoniaJ (Fig. 1 & 2) of which there are several varieties (one often listed as C. drupacea) , and C. fortunii (Fig. 3). The cultivars are shrubs up to about 20 feet in height; most have broad crowns. The linear and pointed leaves are spirally arranged or in two opposite ranks. The upper sur­ Fig. 1. Japanese plumyew (Cephalo­ face is dark shiny green with a conspicu­ taxus harringtonia var. drupacea), an ous mid-rib; the lower surface has a evergreen shrub about 6 ft. high, at broad silvery band on either side of the the USDA Plant Introduction Station, mid-rib. These bands are made up of Glenn Dale, Maryland. This photo­ conspicuous white stomata arranged in graph was made in 1955. The plant is numerous distinct lines. Leaf length is now about 7 ft. high, but the lower variable, from about one inch in varie­ branches have been severely pruned ties of C. harringtonia to three or four to provide material for chemical re­ inches in C. fortunii. The leaves are search. about one-eighth to one-sixth inch wide. The "fruit" is olive-like, about one have a deeply wrinkled kernel. The inch long, wi th a fleshy outer coat, an leaves are similar to those of Torreya inner woody shell, and a smooth surfaced and Taxus. Torreya leaves can be distin­ kernel with two cotyledons. guished by the narrower, brownish Cephalotaxus fruit superficially re­ stomatic bands, a mid-rib that is not semble those of Torreya but the latter distinct on the upper surface, and their spiny point. Taxus leaves can be distin­ 1 Botanists, Crops Research Division, Agricul­ guished by their pale green stomatic tural Research Service, U.S. Department of Ag­ bands. riculture, Beltsville, Maryland. Extracts of Cephalotaxus have demon­ 2 Chemist, Northern Utilization Research and strated significant activity against leuke­ Development Division, Agricultural Research Service, U.S. Department of Agriculture, Peoria, .mia -in -laboratory mice. One of the prin­ Illinois. ci pal active cbnstituents is a new alka- WI TER 1970 19 PHOTO, U.S. DEPARTMENT OF AGRICULTURE Fig. 2. Japanese plumyew composite loid, harringtonine. It was isolated and photo showing (left to right) termi­ has been partially characterized by nal branch with immature fruit and chemists at USDA's Northern Utiliza­ view of lower leaf surfaces; young tion Research and Development Divi­ seed; terminal branch with view of sion in Peoria, Illinois.s upper leaf surfaces; mature fruit (upper right) with seed exposed A sample of C. harringtonia seed was (Ix); and enlarged (12x) view of received in 1960 for testing in the USDA lower leaf surface (lower right) illus­ oilseed screening program, a cooperative trating white bands of stomata that botanical-chemical effort within the Ag­ are characteristic of the genus. (Ma­ ricultural Research Service to discover ture seed photographed in September seed oils that differ significantly from 1951, other material photographed the usual vegetable oils, such as cotton from fresh specimens collected in seed, peanut, and soybean. An aqueous early July 1969. extract of the seed was subsequently submitted for testing in the anti-tumor A seed sample of C. fortunii was screening program of the Cancer Che­ screened for anti-tumor activity in 1963 motherapy National Service Center of and was also active against L-1210. the ational Cancer Institute. This ex­ Cephalotaxus is the second genus of tract showed promising activity against conifers that produces substances which lymphoid leukemia L-1210, currently re­ show activity against L-1210. A different garded as the most reliable test system anti-leukemic alkaloid was isolated from for predicting similar clinical activity in stem bark of Pacific yew (Taxus brevi­ man. This test system is predictive not folia) and has since been identified in only for leukemia but for solid cancers roots, stems, and leaves of other species as well. of yew. Cephalotaxus appears more promising, however, because the active 3 R. G. Powell, D. Weisleder, C. R. Smith, Jr.. principle occurs in much higher amounts and 1. A. Wolff. Harringtonine, a tumor-inhibit­ ing alkaloid from Cephalotax1.1s harringtonia. and the plants appear to produce other mtracts of Papers, 158th National ACS Meet­ alkaloids with similar activity. ing, ew York, September 7-12, 1959. Research is con tinuing at the orth- 20 AMERICA HORTICULTURAL SOCIETY approximately 0.10 to 0.70 percent. This compares with a yield of approximately 0.80 percent from the seed of C. harring­ tonia. Harringtonine makes up only 6 percent of the crude a'kaloid extract or 0.05 percent of the original fruit sample. While this yield seems small, it is rela­ tively high in comparison with that of many other anti-cancer agents of plant origin. Harringtonine recently passed its first important hurdle when it was cleared for preclinical pharmacological evalua­ tion. It will now be tested in dogs and other animals to determine possible ad­ verse side effects and appropriate doses [or administration to human patients. It now appears unlikely that harring­ tonine will be adaptable to chemical synthesis. The plant is likely to provide the only practical source. Fortunately, the alkaloid is obtainable in fairly good yield. Since it is related to cephalotax­ ine, which occurs in larger amounts, a partial synthesis from cephalotaxine is a possibility. An appreciable quanti ty of harring­ tonine will be required for preclinical pharmacology and for clinical trials that, hopefully, will follow. While seed are the best source of the drug, they are Fig. 3. Chinese plumyew (Cephalo­ difficult to obtain. C. harringtonia var. taxus fortunii) Western Szechwan drupacea is native to the mountain for. Province, China, photographed by ests of Japan, from central Honshu to E. H. Wilson during his 1907-1909 Shikoku and Kyushu. Commercial seed expedition. dealers in the United States formerly obtained seed from Japan but have been ern Laboratory to complete the deter­ unable to obtain supplies from that mination of the chemical structure of country for several years. Current sup­ harringtonine. It is now known that it plies of seed come from Europe, espe­ is related to cephalotaxine, an alkaloid cially from cu tivated plants in Italy, but isolated from Cephalotaxus in 1963 by the supply fluctuates from year to year chemists at Ohio University.4 and in some years there is Ii ttle if any Alkaloids occur in all parts of Cepha­ seed produced. This seems also to be the case in Japan where in 1968 the plants lotaxus but they are concentrated in the failed to produce any appreciable quan­ kernel of the fruit. Crude alkaloid ex­ tity of seed. We continue to try to de­ tracts were obtained from roots, large velop sources of seed, and even small stems, twigs, and leaves of several varie­ amounts are welcome, but it is evident ties of C. harringtonia and from twigs that we are likely to be dependent on and leaves of C. fortunii in amounts of vegetative material for future supplies of the drug. 4 W. W. Paulder, G. 1. Kerley, and J. McKay, We are now making a survey of nur­ The alkaloids of Cephalotaxus drupacea and Cephalotaxus fortunei. Journal of Organic series in the United States to locate Chemistry 28:2194-2197 (1963). major sources of Cephalotaxus in this WINTER 1970 21 country. The purpose of our survey is samples will be procured to assess the two-fold. First, we must develop an effect of soils and geography. immediate source of plant material to This assessment will be accomplished provide adequate amounts of the drug by bio-assay in mice rather than by for research to be conducted in the chemical techniques. While the latter are fmmediate future. Secondly, we need to more accurate, the former are less time­ identify the Cephalotaxus germ plasm consuming, and will permit assessment now available in the United States, and of a greater range of material. This to determine which domestic sources evaluation will follow the pattern of our will provide the greatest yield of the evaluation of Camptotheca acuminata,5 drug. source of another anti-leukemic alkaloid. The seven-fold fluctuation in yield of The Camptotheca evaluation demon­ crude alkaloids suggests that production strated that the principal active constitu­ of alkaloids by Cephalotaxus may vary ent was most abundant in the roots and from one species or variety to another. It wood, and least abundant in the twigs may also be dependent on other factors and leaves. This evaluation also indi­ such as soils or geography, age of plants, cated that age of plants, season of collec­ part of the plant tested, or season during tion, and location had little effect on which samples are collected. We hope to biological activity and, presumably, on locate large nursery stocks that are de­ yield of the active constituent. We welcome information from nur­ scended vegetatively from older plants serymen and other readers on sources of in the same area, then set up an evalua­ nursery stock or of older plants, espe­ tion program to determine the best sour­ cially those that produce good crops of ces of harringtonine. Plants will be pur­ fruits.