|Home | About | Journals | Submit | Contact Us | Français|
For over 40 years, small organic molecules derived naturally from microbes and plants have provided a number of useful cancer chemotherapeutic drugs. The search for naturally occurring lead compounds of this type has continued in recent years, with the constituents of marine fauna and flora as well as those of terrestrial microorganisms and plants being investigated for their anti-cancer activities. In the present short review, selected new compounds or their derivatives are described that have obtained for the first time recently from organisms of diverse biological origin, with potential use as cancer chemotherapeutic agents. It may be seen that such promising lead compounds tend to rapidly generate considerable interest among scientists such as synthetic organic chemists and biologists. Consequently, the supply of a given precious natural product sample may be enhanced, and it may be possible to determine a preliminary notion of structure-activity relationships and of its potential mechanism of action.
The term “natural product” is generally regarded as being synonymous with “secondary metabolite”, and these organic substances are of relatively small molecular weight (<3,000 Daltons) and of considerable structural diversity. Such compounds tend to be in the correct chiral form to exhibit biological activity, and it has been postulated that this facilitates species survival by repelling or attracting other organisms . For over 40 years, natural products have played a very important role as established cancer chemotherapeutic agents, either in their unmodified (naturally occurring) or synthetically modified forms . For example, antitumor antibiotics from microbes include the anthracyclines (such as doxorubicin), bleomycin, dactinomycin (actinomycin), and mitomycin C. In turn, members of four classes of plant-derived compounds are used widely as antitumor agents, namely, the bisindole (vinca) alkaloids, the camptothecins, the epipodophyllotoxins, and the taxanes . In addition, there are several examples of promising natural product-derived antineoplastic agents currently in advanced clinical development or recently approved, not only from microbes (e.g., the epothilones and the enediynes) and plants (e.g., the combretastatin and homoharringtonine analogs), but also of marine origin (e.g., the bryostatins, ecteinascidin 743, kahalalide F) . Of a total of 155 a nticancer agents approved for use in Western medicine and Japan since the 1940s, 47% were classified as either natural products per se (14%), semi-synthetic derivatives of natural products (28%), or otherwise derived from natural products (5%) . The present clinically used anticancer agents of natural origin, as well as those compounds of this type now in advanced clinical trial [4–8], are known to exhibit considerable structural diversity.
Collections or “libraries” of purified natural products are regarded as possessing an inherently large degree of “drug likeness”, since it has been pointed out that nearly all such compounds have some receptor-binding capability . Natural product molecules tend to contain chiral centers, complex ring systems, and certain heteroatoms, and so it has been suggested that such compounds be included in combinatorial libraries in order to increase the resultant biological diversity . The development of computerized technologies and robotics has greatly increased the efficiency and speed of the natural product screening process affording a more rapid evaluation of drug candidates [10,11]. It is now possible using “hyphenated” chromatographic and spectroscopic techniques to determine the structures of natural products when these compounds are found in complex mixtures in crude extracts used for biological screening, even if present at very low concentrations. For example, the complete structure and absolute configuration of a component of interest [(R)-gossypol] in an acetone-water extract of the twigs of Thespesia danis Oliver were determined using HPLC-PDA-MS-SPE-NMR, combined with circular dichroism spectroscopy .
Among the largest groups of taxonomically identified classes of organisms that may be studied as sources of new anticancer drugs are arthropods, higher plants, and marine invertebrates . In addition, natural product researchers have examined other taxonomic classes of organisms found all over the world, including algae, bacteria, fungi, and even terrestrial vertebrates [2,13]. However, it must be pointed out that natural product drug discovery for anticancer agents requires special procedures involved with sample collection, inclusive of the development of “benefit-sharing” agreements with source countries, whether the samples are of marine or terrestrial origin . There is a tendency for natural product chemists to specialize on the types of organisms they work, such higher plants or marine fauna, due to the different methods of organism collection and work-up in the laboratory. However, there is increasing evidence that the same secondary metabolite of significance as a potential anticancer agent may be produced by more than one type of organism. To give but one recent example, the aryltetralin lignan podophyllotoxin, is a plant constituent from Podophyllum peltatum L. that is used as a precursor to three epipodophyllotoxin anticancer drugs. This compound was also reported to be produced by two strains of endophytic fungi on P. peltatum, and thus may be able to be produced in the future commercially by fermentation rather than by cultivation .
In the following part of this review, a number of promising new lead compounds isolated recently from diverse organisms with potential for use as antineoplastic agents are described. The compounds chosen are exemplary of the considerable structural variety present in plant, marine, and terrestrial microbial organisms, and no attempt has been made to make the compound selections comprehensive in this short review. Typically, isolation chemists working in a natural product laboratory may isolate only a few milligrams of a promising lead compound of novel structure. This would then be tested against a number of human cancer cells, such as the U.S. National Cancer Institute (NCI) 60-cell line panel [15,16], the application of which may also yield valuable information on mechanism of action [e.g., 17]. Compounds may also be shown to be active in a mechanistic in vitro assay germane to cancer chemotherapy. Ideally, the in vitro activity of the new compound should be evaluated in an in vivo model such as a xenograft of a human tumor in an immunodeficient mouse, but this may require over 100 mg of pure compound. Lead compound scale-up usually may be performed by re-isolation using a further collection of a plant or a marine specimen, or by additional fermentation if the compound is of terrestrial microbial origin. Since these procedures may be quite time-consuming, steps can be taken to make a more rapid decision as to whether a given lead compound natural product has the potential for further development as an anticancer agent. For example, the hollow fiber assay was developed at the NCI [18–20] as a relatively rapid in vivo test for use prior to traditional xenograft assays. Human cancer cells are propagated within fibers that are implanted either subcutaneously (s.c.) or intraperitoneally (i.p.) in immunodeficient mice. In our hands, this method is rapid, economical, requires only about 25 mg of test compound, and is reasonably predictive of activity in a traditional xenograft assay [21,22]. In our experience, however, only a small percentage of compounds determined to be cytotoxic against human cancer cells ultimately prove to have in vivo activity on subsequent biological testing. Therefore, it is by no means certain that any of the compounds discussed in the next section will ever represent a new chemotype of anticancer compounds for human use. However, it is hoped that the selection of these natural products, which is necessarily subjective, will illustrate to the reader both the chemical diversity of the compounds shown and the biological diversity of the organisms of origin.
Bioactivity-directed fractionation of an extract of the leaves of Alvaradoa haitiensis Urb. (Picramniaceae), collected in the Dominican Republic, using human oral epidermoid carcinoma cells (designated KB), led to the isolation and structure elucidation of ten new anthracenone C-glycosides, alvaradoins E–N . Of these, (−)-alvaradoin E (1) was the most active and was further evaluated in vivo in the P388 assay where it showed modest antileukemic activity (125% T/C) at a dose of 0.2 mg/kg per i.p. injection . Using the in vivo hollow fiber model, alvaradoin E demonstrated significant growth inhibition at the i.p. site when tested with KB, LNCaP, and Col2 cells . A series of experiments was then conducted to demonstrate that the cytotoxicity of alvaradoin E (1) was effected through apoptosis. Treatment of prostate cancer cells (LNCaP) for 24 h or less at concentrations ranging from 70 nM to 1.12 µM caused early signs of apoptosis including chromatin condensation (as judged by DAPI analysis) and mitochondrial membrane depolarization (by DiOC6 uptake). Late events in apoptosis were also evaluated in HL-60 cells by flow cytometry using an annexin V-FITC assay, which assesses cell viability, and the TUNEL assay, which measures DNA cleavage. Alvaradoin E significantly decreased (8.6 fold) cell viability and increased (10 fold) DNA cleavage within 36 hours of exposure . The first four compounds in the “alvaradoin” series, alvaradoins A-D, were described in 1999, but they were not investigated for their potential anticancer activity at that time .
(+)-Neopeltolide (2) was isolated from a taxonomically uncharacterized sponge of the family Neopeltidae, order “Lithistida” that was collected in deep water off the coast of Jamaica. The compound was assigned structurally as macrolide with an isoxazole ring in a side chain, and showed some similarities in structure to the previously known compound, leucascandrolide A. However, due to the paucity of material, the absolute configuration was not determined . Neopeltolide was found to be a cytotoxic agent for three cancer cell lines, with potency in the nM range. In a cell cycle analysis study, the compound caused blockage at the G1 stage, when evaluated at a dose of 100 nM in the A549 lung adenocarcinoma model . Shortly after being reported in the literature, neopeltolide was totally synthesized, and its structure has been revised, with the correct structure being a diastereomer with inverted configuration at C-11 and C-13 from that originally proposed [27–30]. Two follow-up studies on the mechanism of action of neopeltolide have indicated that it targets the cytochrome bc1 complex in yeast and mammalian cells [31,32].
Baker and colleagues have recently isolated a new enamide-bearing polyketide, (−)-palmerolide A (3), from Synoicum adareanum, a circumpolar tunicate that is common to the shallow waters around Anvers Island on the Antarctic Peninsula . Palmerolide A was found to be selectively active against melanoma cells (i.e., UACC-62, LC50 = 18 nM) in NCI’s 60 cell line panel. Interestingly, this compound was found to be approximately three orders of magnitude more active against melanoma cells than the other cell lines tested in the panel . While palmerolide A showed moderate activity against one colon cancer cell line (HCC-2998, LC50 = 6.5 µM) and one renal cancer cell line (RXF 393, LC50 = 6.5 µM), the compound was not toxic (LC50 > 10 µM) to the remaining cells, yielding a selectivity index of 103 for the most sensitive cells. Analysis by the COMPARE algorithm revealed that palmerolide A possesses an activity profile that correlated to several vacuolar ATPase inhibitors. Palmerolide A was thus found to potently inhibit V-ATPase (IC50 2 nM) and was also active in the NCI in vivo hollow fiber assay, although details of the cell types involved have not appeared in the literature . The Baker group addressed the relative configuration of palmerolide A in a further communication . Palmerolide A has proved to be an attractive target molecule for synthetic organic chemists, and it has now been established that the actual structure is the ent-(19-epi-20-epi)- derivative of that proposed at the time of its isolation [35,36]. It has proven possible to synthesize five stereoisomers of palmerolide, which will be invaluable for follow-up biological studies .
In 1986, Pettit and colleagues reported the isolation of gram quantities of the parent compound, (+)-pancratistatin, a phenanthridone alkaloid, from the bulbs of the plant Pancratium littorale Jacq. (Amaryllidaceae) . The compound was found to be potently cytotoxic for the P338 murine cancer cell line (ED50 0.01 µg/ml) and to demonstrate in vivo activity against in the M-5076 ovary sarcoma tumor in mice (53–84% life extension at 0.38–3.0 mg/kg) . More recently, pancratistatin has been synthesized [39,40], and in has been shown mechanistically to target mitochondria in human lymphoma cells, resulting in apoptosis by causing activation of caspase-3 and flipping phosphatidyl serine to the outer leaflet of the plasma membrane . Pancratistatin also caused caspase-3 activation in the Jurkat (human T-cell leukemia) cell line, and activated the Fas receptor within membranous lipid rafts . Despite considerable interest in pancratistatin as a potential anticancer and antiviral agent, preclinical drug formulation has been hindered by the peer water solubility and bioavailability of this compound. However, one approach towards enhancing the solubility of pancratistatin has been to produce its sodium pancratistatin 3,4-O-cyclic phosphate salt (4), as reported by Pettit and colleagues in 2004 . It was shown very recently by Shnyder and colleagues that like the parent compound, 4 is also potently cytotoxic for cancer cell lines. Moreover, 4 was demonstrated to have efficacy in a DLD-1 human colon tumor model, causing necrosis and decrease in functional vasculature at the maximum tolerated dose of 100 mg/kg . Despite the fact that pancratistatin is not a new natural product, this approach to rendering this alkaloid more water soluble seems highly effective, and should promote wide further interest among members of the biomedical community.
(+)-Pericosine A (5) was isolated over 10 year ago as a relatively structurally simple C7 cyclohexenoid derivative from a strain OUPS-N133 of the fungus, Periconia byssoides, isolated from the gut of the sea hare, Aplysia kurodai . Pericosine A was more cytotoxic than an analog, pericosine B, when evaluated in vitro against the P388 lymphocytic leukemia cell line (ED50 0.12 vs. 4.0 µg/ml, respectively). Activity in vivo was suggested when mice were injected i.p. with P388 leukemia cells on day 0, and administered with 25 mg/kg of 5 on days 1 and 5. The median survival times for treated and non-treated mice were 13 and 10.,7 days, respectively . Full stereostructures were reported for pericosines C-E, three further analogs of periocosines A and B, and it was indicated that 5 inhibits protein kinase EGFR and topoisomerase II . Several synthetic procedures have been published for the pericosines A and B [e.g., 47–49]. In 2007, the absolute stereostructure of (+)-pericosine A (5) was revised following the total synthesis of both the (+)- and (−)- forms .
Rocaglamide or rocaglate derivatives are known to be toxic to insects and cancerous mammalian cells [50,51]. These cyclopenta[b]benzofurans can also inhibit protein synthesis and induce cell cycle arrest in the G2/M-phase [52–54] and are potent and specific inhibitors of tumor necrosis factor alpha (TNFα) and phorbol 12-myristate 13-acetate (PMA)-induced NF-κB activity in human T cells . Our laboratories have studied extracts of the fruits and twigs of Aglaia foveolata Pannell (Meliaceae; originally misidentified as Aglaia silvestris), which were collected in Kalimantan, Indonesia, as part of a natural products anticancer drug discovery program . (−)-Silvestrol (6), a rocaglate derivative with an unusual pendant dioxanyl ring, was isolated from both the fruits and twigs of A. foveolata by bioassay-guided fractionation, and its structure and absolute stereochemistry determined by spectroscopic data interpretation and by X-ray crystallography . Silvestrol was shown to be highly cytotoxic against a panel of human cell lines derived from breast, prostate and lung cancers. The potency of silvestrol (ED50 1.2 to 1.5 nM) was similar to that observed for paclitaxel (ED50 0.7 to 4.7 nM) and camptothecin (ED50 = 30 nM). These in vitro studies were followed by analysis of silvestrol in vivo . When tested at doses up to 5 mg/kg, silvestrol inhibited proliferation of all cell lines, particularly the human prostate cancer line designated LNCaP (up to 83% inhibition) with no detectable gross toxicity to the mice . Silvestrol was also tested in the P388 murine lymphocytic leukemia model and found to be active at its maximum tolerated dose of 2.5 mg/kg when administered as five daily i.p. injections. A maximum increase in lifespan corresponding to a treatment/control ratio of 150% was achieved . Silvestrol has been isolated from a second species, Aglaia leptantha Miq. by Meurer-Grimes et al., and found to inhibit the growth of PC3 human prostate cancer cells in a xenograft study . Very recently, (−)-silvestrol has been totally synthesized by two independent groups [58,59].
Studies have been conducted to elucidate the cellular mechanism of action of silvestrol (6) in LNCaP cells. Microarray analysis of the molecular signature induced in cultured LNCaP cells by silvestrol revealed that 20 apoptosis and cell cycle related genes were significantly altered . Genes that were up-regulated included p21 (a potent cyclin-dependent kinase inhibitor that is governed by p53), and p300 (a transcription factor co-regulator). Conversely, silvestrol down-regulated p53 at the RNA and protein level within 15 minutes of exposure. Within 6 hours of silvestrol exposure, no p53 could be detected by immunoblot. This effect was accompanied by down regulation of MDM2, the E3 ligase specific for p53, and was not prevented by lactacystin, suggesting that silvestrol-induced degradation of p53 is not mediated by the proteasome. Cell cycle analysis by flow cytometry demonstrated that silvestrol induces a block in the cycle at the G2/M checkpoint .
Cell cycle arrest induced by silvestrol (6) led to cell death by apoptosis . Studies on the mechanism of apoptosis induction revealed that silvestrol disrupted the mitochondrial trans-membrane potential and caused cytochrome c release into the cytoplasm. Furthermore, silvestrol produced an increase of Bcl-xl phosphorylation and increased Bak expression. While caspases-2, −9 and −10 were found to be involved in silvestrol-mediated apoptosis, caspases-3 and −7 were not. Experiments using various cell-permeable inhibitors showed that only the general caspase inhibitor Boc-D-Fmk completely inhibited the formation of apoptotic bodies. These studies suggest that silvestrol induces apoptosome/mitochondrial pathway to stimulate apoptosis in treated cells .
A recent study by Pelletier and colleagues showed that silvestrol (6) can inhibit translation in eukaryotic cells  and enhance the sensitivity of lymphoma cells to doxorubicin. The eukaryotic translation initiation factor (eIF) 4F is a complex that stimulates ribosome recruitment to mRNA. The eIF4F complex is composed of three subunits including eIF4FE eIF4FG and eIF4A, which is a DEAD-box RNA helicase that unwinds the 5′ mRNA structure to permit association with the 43S preinitiation complex. Silvestrol targets eIF4A to prevent its association with eIF4FE and eIF4FG, thus preventing the formation of the eIF4F complex and subsequent translation. These findings suggest a novel mechanism for silvestrol’s anticancer activity.
Recent studies by Grever and colleagues [63,64] indicate that the cytotoxicity of silvestrol (6) is selective for B cells relative to T cells. Silvestrol was found initially to be selectively cytotoxic for chronic lymphocytic leukemia (CLL) cells relative to normal peripheral blood monocyte cells, and its cell killing was independent of two key resistance factors, p53 and Bcl-2 . The B-cell-specific activity was observed both ex vivo and in vivo using the Tcl-1 transgenic murine model of CLL. Silvestrol significantly decreased leukemic B cells in vivo, without detectable effects on T cell number. Silvestrol caused a rapid proteasome independent reduction in Mcl-1 protein, which is a member of the Bcl-2 family of proteins associated with the outer mitochondrial membrane . These promising results have recently prompted the Drug Development Group of NCI to select silvestrol for preclinical evaluation at the IIA level (including additional sourcing of the plant of origin and preliminary formulation, pharmacokinetics, and toxicology), potentially leading to clinical development. Recently, silvestrol was found to occur in the leaves of one of its plants of origin, A. foveolata, so that this plant part could be used as a renewable resource for the production of silvestrol, in the event this compound continues to show great promise .
In 2001, the structures of the bicyclic depsipeptides, (−)-spiruchostatin A (7) and spiruchostatin B were reported from a Pseudomonas extract . Spiruchostatin A (also known as YM753) closely resembles FK228, a depsipeptide in clinical trials for cancer as an inhibitor of histone deacetylase (HDAC) . Both spirochostatins A and B have been synthesized [e.g., 68–70]. Recently, spiruchostatin A was subjected to extensive biological testing, and shown to exhibit greater potency as a HDAC inhibitor than several other compounds, with an IC50 value of 2±0.2 nM (cf. 3.1±0.1 for FK228) . It also showed growth inhibitory effects in the range 1.6–16 nM (GI50) for 14 cancer cells representing a variety of cancers . Moreover, this compound was evaluated in mice bearing WiDr colon tumor xenografts (3 mg/kg, three time a week for two weeks), leading to significant tumor growth inhibition by day 14. It was concluded that 7 causes selective accumulation of acetylated histones in tumor tissues specifically .
Gerwick and colleagues have isolated a new depsipeptide named (−)-symplocamide A (8) containing two unusual amino acid units. This compound was isolated from a marine cyanobacterium collected in Papua New Guinea and identified only to the species (Symploca) level The complete stereostructure of symplocamide A was determined, and the compound was found to be an extremely potent cytotoxin, exhibiting IC50 values of 40 nM and 29 nM for H460 lung cancer and neuro-2A neuroblastoma cells, respectively . The compound was also found to show more than a 200-fold selectivity for the inhibition of the protease enzyme, chymotrypsin, relative to trypsin, and, as such, it was suggested that symplocamide A should be tested for its effects as a proteasome inhibitor .
From the examples shown in this review, it has been shown that biologically diverse organisms from the land and sea continue to afford new small-molecule organic compounds with potential anticancer activity. There is a tendency for these organisms to be sourced from ever more geographically remote regions of the earth. The need for isolation chemists, synthetic organic chemists, biochemists, and biologists to work together to accelerate the development of promising new molecules towards preclinical trials must be emphasized. Natural products representative of new chemotypes with novel biochemical mechanisms of action remain of much interest as compounds that might lead to the alleviation of the cancer scourge.
•• of outstanding interest
• of special interest