《天然药物化学》课程参考文献（海洋天然产物）Marine Natural Products and Related Compounds in Clinical and Advanced Preclinical Trials.pdf_大学文库

Marine Natural Products and Related Compounds in Clinical and Advanced Preclinical Trials† David J. Newman* and Gordon M. Cragg Natural Products Branch, Developmental Therapeutics Program, NCI-Frederick, P.O. Box B, Frederick, Maryland 21702 Received February 10, 2004 The marine environment has proven to be a very rich source of extremely potent compounds that have demonstrated significant activities in antitumor, antiinflammatory, analgesia, immunomodulation, allergy, and anti-viral assays. Although the case can and has been made that the nucleosides such as Ara-A and Ara-C are derived from knowledge gained from investigations of bioactive marine nucleosides, no drug directly from marine sources (whether isolated or by total synthesis) has yet made it to the commercial sector in any disease. However, as shown in this review, there are now significant numbers of very interesting molecules that have come from marine sources, or have been synthesized as a result of knowledge gained from a prototypical compound, that are either in or approaching Phase II/III clinical trials in cancer, analgesia, allergy, and cognitive diseases. A substantial number of other potential agents are following in their wake in preclinical trials in these and in other diseases. Introduction The initial discoveries from the marine environment that led to the belief that true marine-derived drugs would not be overly long in reaching the market can be traced to the reports of Bergmann on the discovery and subsequent identification of spongothymidine and spongouridine in the early 1950s from the Caribbean sponge Tethya crypta.1-3 These reports actually led to a complete reversal of the then current dogma, which prior to these discoveries was “that for a nucleoside to have biological activity, it had to have ribose or deoxyribose as the sugar, but that the base could comprise a multiplicity of heterocycles and even carbocycles”. The subsequent explosion of compounds is described with the relevant citations by Suckling4 and Newman et al., 5 and these discoveries led to the identification of a close analogue, cytosine arabinoside, as a potent antileukemic agent; this compound (1) subsequently was commercialized by Upjohn (now Pharmacia) as Ara-C. Other closely related compounds such as adenine arabinoside (Ara-A) (2), an antiviral compound synthesized and commercialized by Burroughs Wellcome (now Glaxo SmithKline) and later found in the Mediterranean gorgonian Eunicella cavolini, and even azidothymidine (AZT) (3) can be traced back to this initial discovery of the “other than ribose-substituted bioactive nucleosides”. The advent of scuba techniques approximately 60 years ago and their subsequent utilization by natural products chemists and biologists working closely with them led to questions such as, Why are certain marine invertebrates not prey for organisms higher up the evolutionary tree? Why do fish not eat particular algae? Why do two sponges grow and expand until they touch, but do not grow over each other? One possibility was that the organisms have some form of chemical communication or defense that enables an individual organism to establish a particular niche and thrive there. One has to realize that these marine invertebrates and marine plants, with very few exceptions, are sessile and require a “foot-hold” on a nonmoving, fixed substrate (rock or coral) that permits them to feed by filtration of the seawater flowing in and around them. Initial attempts at determining the chemistries of marine organisms were simply extensions of tried and true phytochemical techniques. Thus, easily accessible organisms (generally sponges and encrusting organisms such as ascidians) were collected by hand using snorkel or simple scuba systems, and then their chemical components were extracted and identified. Any biological activity was found as an afterthought in these initial experiments (though as shown above, active compounds could be found by these techniques that would ultimately be useful as treatments for human diseases). A corollary to the more systematic searching for marinederived products was that very sensitive analytical tools had to be used, as in general, the amounts of bioactive materials that could be recovered were exceedingly small. There are examples given later in detail, but levels of 1 mg of compound per 3 kg of organism were not uncommon. Thus high-field NMR (originally 200 MHz and then up through 600-800 MHz), mass spectrometry that involved MS-MS techniques, and chromatographic methods of all types were used. It should be emphasized that HPLC, the use of which is effectively a sina qua non in modern isolation methods, was not generally available until the late 1970s, and thus isolations often required large amounts of materials due to the level of sophistication of the techniques available. In retrospect, this is one of the major reasons that the field evolved slowly. Discovery of a given compound was easy in relative terms, but development, which required large amounts, was, and still is, not simple, as will be shown in examples later in the review. Although Paul Scheuer at the University of Hawaii was the first (marine) natural products chemist to systematically explore the chemistry of marine invertebrates, from his original work in Hawaii in the 1950s until his death † Dedicated to the late Dr. D. John Faulkner (Scripps) and the late Dr. Paul J. Scheuer (Hawaii) for their pioneering work on bioactive marine natural products. * To whom correspondence should be addressed. Tel: (301) 846-5387. Fax: (301) 846-6178. E-mail: dn22a@nih.gov. 1216 J. Nat. Prod. 2004, 67, 1216-1238 10.1021/np040031y This article not subject to U.S. Copyright. Published 2004 by the Am. Chem. Soc. and the Am. Soc. of Pharmacogn. Published on Web 06/11/2004

early in 2003, initially investigating marine toxin structures, the work of Rinehart at the University of Illinois at Champaign-Urbana and of Pettit at Arizona State University led the way in the discovery of biologically active molecules (i.e., potential human use pharmaceutical agents) from the marine environment. Both of these research groups were funded by the U.S. government but in somewhat different ways in the beginning. Pettit was part of an antineoplastic drug discovery effort whereby organisms were collected and extracted by NCI-funded groups, and Table 1. Status of Marine-Derived Natural Products in Clinical and Preclinical Trials name source status (disease) comment didemnin B Trididemnum solidum Phase II (cancer) dropped middle 90s dolastatin 10 Dolabella auricularia (marine microbe derived; cyanophyte) Phase I/II (cancer) many derivatives made synthetically; no positive effects in Phase II trials; no further trials known girolline Pseudaxinyssa cantharella Phase I (cancer) discontinued due to hypertension bengamide derivative Jaspis sp. Phase I (cancer) licensed to Novartis, Met-AP1 inhibitor, withdrawn 2002 cryptophycins (also arenastatin A) Nostoc sp. & Dysidea arenaria Phase I (cancer) from a terrestrial cyanophyte, but also from a sponge as arenastatin A; synthetic derivative licensed to Lilly by Univ. Hawaii, but withdrawn 2002 bryostatin 1 Bugula neritina Phase II (cancer) now in combination therapy trials; licensed to GPC Biotech by Arizona State Univ.; may be produced by bacterial symbiont TZT-1027 synthetic dolastatin Phase II (cancer) also known as auristatin PE and soblidotin cematodin synthetic derivative of dolastatin 15 Phase I/II (cancer) some positive effects on melanoma pts in Phase II; dichotomy on fate ILX 651, synthatodin synthetic derivative of dolastatin 15 Phase I/II (cancer) in Phase II for melanoma, breast, NSCLC ecteinascidin 743 Ecteinascidia turbinata Phase II/III (cancer) in 2003 licensed to Ortho Biotech (J&J); produced by partial synthesis from microbial metabolite aplidine Aplidium albicans Phase II (cancer) dehydrodidemnin B, made by total synthesis E7389 Lissodendoryx sp Phase I (cancer) Eisai’s synthetic halichondrin B derivative discodermolide Discodermia dissoluta Phase I (cancer) licensed to Novartis by Harbor Branch Oceanographic Institution kahalalide F Eylsia rufescens/ Bryopsis sp. Phase II (cancer) licensed to PharmaMar by Univ. Hawaii; revision of structure ES-285 (spisulosine) Spisula polynyma Phase I (cancer) Rho-GTP inhibitor HTI-286 (hemiasterlin derivative) Cymbastella sp Phase II (cancer) synthetic derivative made by Univ. British Columbia; licensed to Wyeth KRN-7000 Agelas mauritianus Phase I (cancer) an agelasphin derivative squalamine Squalus acanthias Phase II (cancer) antiangiogenic activity as well Æ-941 (Neovastat) shark Phase II/III (cancer) defined mixture of <500 kDa from cartilage; antiangiogenic activity as well NVP-LAQ824 Synthetic Phase I (cancer) derived from psammaplin, trichostatin, and trapoxin structures Laulimalide Cacospongia mycofijiensis preclinical (cancer) synthesized by a variety of investigators Curacin A Lyngbya majuscula preclinical (cancer) synthesized, more soluble combi-chem derivatives being evaluated vitilevuamide Didemnum cucliferum & Polysyncraton lithostrotum preclinical (cancer) diazonamide Diazona angulata preclinical (cancer) synthesized and new structure elucidated eleutherobin Eleutherobia sp. preclinical (cancer) synthesized and derivatives made by combichem; can be produced by aquaculture sarcodictyin Sarcodictyon roseum preclinical (cancer) (derivatives) combi-chem synthesis performed around structure peloruside A Mycale hentscheli preclinical (cancer) salicylihalimides A Haliclona sp. preclinical (cancer) first marine Vo-ATPase inhibitor; similar materials from microbes, synthesized thiocoraline Micromonospora marina preclinical (cancer) DNA polymerase R inhibitor ascididemnin preclinical (cancer) reductive DNA-cleaving agents variolins Kirkpatrickia variolosa preclinical (cancer) Cdk inhibitors dictyodendrins Dictyodendrilla verongiformis preclinical (cancer) telomerase inhibitors GTS-21 (aka DMBX) Phase I (Alzheimer’s) modification of a worm toxin; licensed to Taiho by Univ. Florida manoalide Luffariaella variabilis Phase II (antipsoriatic) discontinued due to formulation problems IPL-576,092 (aka HMR-4011A) Petrosia contignata Phase II (antiasthmatic) derivative of contignasterol; licensed to Aventis IPL-512,602 derivative of 576092 Phase II (antiasthmatic) with Aventis IPL-550,260 derivative of 576092 Phase I (antiasthmatic) with Aventis ziconotide (aka Prialt) Conus magus Phase III (neuropathic pain) licensed by Elan to Warner Lambert CGX-1160 Conus geographus Phase I (pain) contulakin G CGX-1007 Conus geographus Phase I (pain & epilepsy) conantokin G; discontinued AMM336 Conus catus preclinical (pain) ω-conotoxin CVID ø-conotoxin Conus sp. preclinical (pain) conotoxin MR1A/B CGX-1063 Thr10-contulakin G preclinical (pain) modified toxin ACV1 Conus victoriae preclinical (pain) R-conotoxin Vc1.1 Reviews Journal of Natural Products, 2004, Vol. 67, No. 8 1217

the extracts tested by NCI contractors for their ability to inhibit the growth of tumors in mice. The active principles were then isolated by NCI-funded groups by following the bioactivity in mice. This was probably the first large-scale application in the marine area of what has come to be known as “bioactivity-driven isolations”. Rinehart, however, was funded by a number of U.S. government agencies but initially used his MS-MS and NMR capabilities to determine the potential structures of the bioactive agents that he found in organisms collected predominately in the Caribbean during NSF-funded expeditions. There have been a number of recent reviews covering aspects of this area, either not in as much detail or from a clinical or preclinical aspect. The reader should consult them for comparative purposes;6-13 any reviews that are specific to a class of agents will be cited under the agents themselves. We will discuss agents by clinical activities rather than by source or chemical class, and in order to aid the reader, we have shown in Table 1 all of the sources, diseases, trial level achieved at date of review submission, and a short comment where necessary on the compounds that we discuss in the review. The order is by type of pharmacological activity and then clinical and/or preclinical results for each activity, with those that have been discontinued listed first in each disease. Introduction to Agents that Entered Antitumor Clinical Trials The significant number of compounds from marine sources that have been entered into antitumor preclinical and clinical trials since the early 1980s is due to two serendipitous findings. The first is that the agents elaborated by marine organisms must be affected by the dilution effects of seawater; thus any “chemical warfare” agent must be extremely potent, as it has to overcome dilution en route to its target. This process may be considered as analogous to the role of phytoalexins in the plant kingdom, or similar to the emission of pheromones by insects, though the purpose in the latter case is to attract rather than repel! The other is that the U.S. National Cancer Institute (NCI) has funded, either directly or indirectly, most of the search for agents active against cancer, irrespective of the source. Thus, one has the systems in place for collection, bioactivity determinations, and subsequent testing in animals and humans, with the aim of finding new and potent treatments for cancers. Because of the extremely long time frame involved in such processes (for example, paclitaxel (Taxol), took over 20 years from structural determination and reporting until FDA approval in the early 1990s), the compounds that will be discussed fall into two approximate time frames: those from the initial collection programs (which aided the didemnin B discovery vide infra) and those that are further back in the current system that have been discovered as a result of the modified NCI screens utilizing the 60 cell line (or functional equivalent) screen that has been in use from the early 1990s. Some of the agents whose mechanisms of action (MOA) were discovered as a result of the latter screening system are now either just entering or about to enter clinical trials. Agents Now Withdrawn from Antitumor Clinical Trials Didemnin B. This compound (4) was isolated by Rinehart’s group from extracts made of the tunicate Trididemnum solidum14 that demonstrated excellent antiviral activity and subsequent cytotoxic activity against P388 and L1210 murine leukemia cell lines. Didemnin B was advanced into preclinical and clinical trials (Phases I and II; see Table 3 in Nuijen et al. for a discussion of these trials13) under the auspices of the NCI in the very early 1980s as the first defined chemical compound directly from a marine source to go into clinical trials for any major human disease. Despite many different treatment protocols and testing against many types of cancer, the compound turned out to be too toxic for use, and trials were officially terminated in the middle 1990s by NCI. Even though this compound did not make it to Phase III trials and then to market, the experience gained from these efforts was immensely helpful in aiding the trials of other natural product-derived agents/compounds. Thus Rinehart’s group developed methods of large-scale isolation and purification and, as would become essential much later in time, total syntheses that permitted significant structureactivity relationships to be derived.15 This work permitted materials to be provided to others so that basic biochemical studies could be performed, leading to the identification of a potential MOA for this compound, with the binding to elongation factor 1-R (ef1-R) being reported by Crews et al. in the middle 1990s.16 Subsequent reports from Crews’ group showed that didemnin B binds noncompetitively to palmitoyl protein thioesterase,17 and the following year, Johnson and Lawen reported that rapamycin inhibited the didemnin-induced apoptosis of human HL-60 cells, perhaps by binding to the FK-506 binding protein(s).18 Inferentially, from this latter result, didemnin B might bind to or modulate the FK-binding proteins as part of its immunomodulatory process and thus lead to cell death via apoptosis. Table 2. Phase I and Phase II Combination Studies with Bryostatin 1 year phase schedule dose range tumor type(s) # pts CR PR SD side effects reference 2001 I 24 h infusion & bolus of vincristine, dose escalation of bryostatin, 1-5 cycles 12.5-62.5 µg‚M2 bryostatin; 1.4 mg‚M2 of vincristine B-cell cancer 25 1 2 4 myalgia; neuropathy Dowalti et al.324 2002 I 24 h infusion, days 1 & 11, AraC on days 2, 3, 9, 10, bryostatin dose escalation, fixed AraC, 1-6 cycles 12.5-50 µg‚M2 bryostatin; 1-3 gm‚M2 AraC leukemia 23 5 1a 0 myalgia; neutropenia Cragg et al.325 2002 I 24 h infusion, fludarabine for days 2-6, repeat at 28d, or reverse addition order, 6-9+ cycles 16-50 µg‚M2 bryostatin; 12.5-25 mg‚M2 FAra CLL; NHL 53 bb b neutropenia Roberts et al.326 2003 II 1 h infusion of paclitaxel on 1, 8, & 15d; 24 h infusion of bryostatin on 2, 9, 16d, repeated on 28d cycle, 1-4 cycles 40-50 µg‚M2 bryostatin; 90 mg‚M2 paclitaxel NSCLC 11 0 2a 5 myalgia Winegarden et al.327 a Some question as to response level. b 23 “nondefined objective responses”. 1218 Journal of Natural Products, 2004, Vol. 67, No. 8 Reviews

In 2002, Vera and Joullie19 published an excellent review of didemnins as cell probes and targets for syntheses and also made some reasonable arguments that the dosing schedules used in the early clinical trials may well have been nonoptimal for demonstrating activity as a cytotoxin rather than as an immunosuppressive/modulator. It will be interesting to compare the dosing schedules and responses for didemnin B and aplidine (Aplidin; PharmaMar, vide infra) in man once the latter are fully reported in the literature. Although didemnin B was not successful, a very close chemical relative is currently in clinical trials (cf. aplidine below), and in 2000 Rinehart published an overview of these compounds as part of a discussion of antitumor compounds from tunicates, which the reader may consult for further details.20 Dolastatin 10. The dolastatins are a series of cytotoxic peptides that were originally isolated in very low yield from the Indian Ocean mollusk Dolabella auricularia by Pettit’s group as part of its work on marine invertebrates.21-25 Due to the potency and mechanism of action of dolastatin 10 (5), a linear depsipeptide that was shown to be a tubulin interactive agent binding close to the vinca domain at a site where other peptidic agents bound,26,27 the compound entered Phase I clinical trials in the 1990s under the auspices of the NCI. Since the natural abundance was so low, Pettit and others developed synthetic methods that provided enough material under current Good Manufacturing Practice (cGMP) conditions to allow clinical trials to commence.25 Dolastatin 10 progressed to Phase II trials as a single agent, but although tolerated at the doses used, which were high enough to give the expected levels in vivo to inhibit cell growth, it did not demonstrate significant antitumor activity in a Phase II trial against prostate cancer in man.28 Similarly, no significant activity was seen in a Phase II trial against metastatic melanoma, even though again, levels high enough to affect cells were demonstrated.29 There are other dolastatins and molecules related to them that are still in clinical and preclinical trials; they will be covered in later sections. Girolline (Girodazole). This very simple compound, a substituted imidazole (6), was reported from the sponge Pseudaxinyssa cantharella30 and was shown by workers at Rhone-Poulenc Rorer to be an inhibitor of protein synthesis, acting preferentially on the termination step in eukaryotic protein synthesis, in contrast to other known protein synthesis inhibitors such as emitine, homoharringtonine, anguidine, and bruceantin, which generally act at either the initiation or elongation steps.30 Girolline proceeded to Phase I clinical trials in man, but the trials were stopped due to significant hypertensive effects seen in treated patients. In 2002, Schiavi et al. 31 reported on the synthesis of one of the two possible thiazole derivatives of girolline, 5-deazathiogirolline (7), hoping that this simple substitution might alter the human toxicity characteristics. Although protected intermediates in the synthetic scheme were about 10% as active in girolline in comparable systems, the final deprotected product (the thiazole derivative) was effectively inert. Bengamide Derivatives. Bengamides A (8) and B (9) were first reported in 1986 as antihelminthic compounds (together with some antibiotic and cytotoxic activites) by Crews’ group at the University of California, Santa Cruz.32 The number of bengamide analogues isolable from the same sponge was extended to bengamide G, with details being reported on their isolation and absolute stereochemistry in two more papers from the same group.33,34 In a subsequent paper with workers from Novartis, the number of compounds in the group was extended, and their antitumor activities were reported.35 The bengamides were evaluated by Novartis (initially by Ciba-Geigy), as Ciba-Geigy was the then current industrial partner of the UCSC group in an NCI-funded National Cooperative Natural Products Drug Discovery Group (NCNPDDG). As a result of their intrinsic activities, a synthetic program was put in place that developed a derivative of bengamide A (10) as a clinical candidate. This derivative was shown to be an inhibitor of methionine aminopeptidases and subsequently entered Phase I clinical trials in 2000, but was withdrawn in the middle of 2002. Cryptophycins. These compounds were reported from two blue-green algae, initially by a group from Merck in 1990 using a Nostoc species (ATCC 53789) originally isolated from a lichen on a Scottish Island; they reported only the antifungal activity, finally deciding not to proceed with development, as it was too toxic. Moore’s group at the University of Hawaii then identified the same compound36 from a nonmarine cyanophyte, Nostoc sp. strain GSV-224, Reviews Journal of Natural Products, 2004, Vol. 67, No. 8 1219

and in addition, almost contemporaneously, a similar molecule was reported by Kobayashi et al. from an Okinawan sponge (see below). The University of Hawaii and Wayne State University licensed the natural cryptophycins and synthetic derivatives to the Lilly Company for advanced preclinical and clinical development. This led to the selection of cryptophycin 52 (LY355703) (11) as a Phase I clinical candidate in the middle 1990s, with a single publication37 in late 2002 giving the Phase I and pharmacological results from a variety of schedules, with an intermittent schedule being chosen for Phase II studies. The routes, both chemical and pharmacological, leading to the choice of this particular derivative were described by Shih and Teicher38 of the Lilly Research Laboratories. The compound progressed toward Phase II trials, but in 2002, cryptophycin 52 was withdrawn from trial (personal communication, Dr. R. Moore). Although the original cryptophycins came from terrestrial cyanophytes and the clinical candidate came from semisynthetic modifications of the natural product, in 1994 Kobayashi et al. reported that an acetone extract of the Okinawan sponge Dysidea arenaria had potent cytoxicity,39 and on purification, the compound arenastatin A (12) subsequently turned out to be identical to cryptophycin 24 (12) reported by Moore’s group in 1995.40,41 A later report from the Japanese group42,43 demonstrated that arenastatin A and synthetic analogues also are tubulin interactive agents similar in activity to the other cryptophycins reported by Moore et al. Agents Currently in Clinical Trials as Antitumor Agents Bryostatins. In 1968, NCI commissioned a large-scale (for those days) collection of the bryozoan Bugula neritina by Jack Rudloe of the Gulf Specimen Company off the west coast of Florida that was sent to Pettit’s group for chemical workup. The aqueous 2-propanol extract was subsequently tested by NCI for its intrinsic activity as an antitumor agent in the then current P388 and L1210 murine leukemia in vivo models. Subsequently, the extract was found to be inactive against L1210 but to give a 68% increase in life span using P388 at the same concentration.44 Following significant amounts of work by Pettit and his group, including more collections on a larger scale, significant problems with isolation as a result of dealing with vanishingly small quantities of a very potent agent, and problems related to assay reproducibility, the compound was purified and identified as bryostatin 3 (13), one of a series of closely related compounds that now number 20.44-49 Subsequent work by Pettit’s group identified two other geographic areas where significant (in relative terms) quantities of bryostatin 1 (14) could be isolated from B. neritina colonies. What is important, however, is that although a number of reports have been made about other taxa producing bryostatins, in almost all cases, on careful examination, the putative producing organism actually contains B. neritina. However, as a result of prodigious efforts on the part of Pettit and collaborators and workers at NCI-Frederick, by 1990 there was enough cGMP-grade material to commence systematic clinical trials, though prior to this time frame, small quantities of bryostatin 1 had been supplied to a variety of collaborators so that basic biochemical studies and initial clinical trials in the U.K. could be performed. From these studies, which are summarized in recent reviews by a number of authors,48-50 it was shown that bryostatins bind to the same receptors as the tumorpromoting phorbol esters, the protein kinase C (PKC) isozymes, but have little or no tumor promoter activity. A recent paper from Hale’s group51 where they made a modified analogue has shown that the binding site for this compound and, by inference, the bryostatins is almost certainly at the cysteine-rich domain 2 (CRD2) in human PKC-R. As a result of this binding, the PKC isozymes in various tumor cells are significantly down-regulated, leading to inhibition of growth, alteration of differentiation, and/or death. To date, bryostatin 1 has been in more than 80 human clinical trials, with more than 20 being completed at both the Phase I and Phase II levels. There have been some responses to the compound as a single agent with effects ranging from complete remission (CR), to partial remission (PR), to stable disease (SD). However, the use as a single agent is probably not the optimal usage for this compound. More detailed reports of the clinical development are given in the recent reviews by Pettit50 and by Clamp and Jayson.52 However, when bryostatin is combined with another cytotoxin, such as the vinca alkaloids or nucleosides, and the carcinomas are leukemic in nature, then the response rates, even in Phase I trials, begin to demonstrate that such mixed treatments may well be worth further investigation (cf. Table 2 for details/citations). Thus a combination with high levels of AraC and low levels of bryostatin in patients with leukemias, in a population that included patients who had failed high-dose AraC (HiDaC) therapy, five of 23 patients presented with complete responses in a recent Phase I trial. Similarly, patients with chronic lymphocytic leukemia (CLL) and nonHodgkins lymphoma (NHL) treated with fludarabine and bryostatin were reported to show close to 50% “objective responses” in the trial report. With non-small cell lung cancer (NSCLC) and paclitaxel/bryostatin, seven of 11 1220 Journal of Natural Products, 2004, Vol. 67, No. 8 Reviews

patients in a Phase II trial demonstrated positive responses (PR/SD) but no CRs. Currently (01/2004), there are four Phase I and five Phase II trials underway (data from the NCI clinical trials web site http://clinicaltrials.gov), and in every case, these are combination studies with biologicals such as interleukin 2 or granulocyte macrophage-colony stimulating factor (GM-CSF), nucleoside derivatives such as gemcitabine, cladribine, or AraC, or other cytotoxic agents such as paclitaxel, vincristine, or cisplatin. These combinations are being tested against leukemias and lymphomas and ovarian and prostate carcinomas. Hopefully, results similar to those demonstrated in Table 2 will be reported in due course. In all of the clinical trials so far reported the major cause of dose-limiting toxicity (DLT) appears be myalgia, but in almost all cases reported this was treatable by standard supportive therapies and patients continued on trial. Details as to the protocols for all trials and the results reported are given in two articles currently in press.53,54 A very interesting “side effect” of the use of bryostatin as a clinical candidate was the early realization that wild collections would not suffice to produce enough of the material for use as a clinical entity. For example, to obtain enough material for the initial clinical trials under NCI auspices, it was necessary to begin with 13 metric tonnes55 of wild-collected B. neritina and then process the material using large-scale chromatographic techniques in order to produce 18 g of cGMP bryostatin 1. Subsequently, NCI funded in-sea and on-land aquaculture (total NCI expenses above $1M) in order to establish the parameters necessary to produce bryostatin 1 in sufficient quantities at a “reasonable” cost if it progressed through the development pipeline. The processes involved and the successful results have recently been reviewed by Mendola,56 and this review should be consulted for specific information as to methods, economics, etc. Since the publication of the first structure by Pettit in 1982, these molecules have been the target of many synthetic chemistry groups. Many partial syntheses have been published where specific portions of the molecule have been made, but to date, only three of the 20 reported bryostatins have been synthesized. The first was the enantioselective total synthesis of bryostatin 7 in 1990 by Masamune et al.,57 the second by Evans et al. on the enantiomeric total synthesis of bryostatin 2 in 1999,58 and the third, the synthesis of bryostatin 3, by the group of Nishiyama and Yamamura59 in 2000. In addition to these papers, three excellent review articles have been published covering information available through 2002, on the syntheses of these three and other partial bryostatin structures including bryostatin 1, and should be consulted for specific details of reaction schema and comparisons of routes.48,49,60 From inspection of the three reviews referred to above it can be stated that the total synthesis of bryostatin 1 is not the process that one would wish to utilize to produce this agent. However, if one could synthesize a simpler analogue with comparable activity, then chemical production of such an agent might well be a viable option. In 1986, Wender et al. analyzed the potential binding site of the phorbol esters on PKC as a guide to the design of simpler analogues of these agents.61 In 1988, this work was expanded by modeling bryostatin 1 onto the same binding site as a result of the initial results indicating that bryostatin 1 interacted with PKC.62 Subsequently, the modeling work was refined to produce three analogues that would maintain the putative binding sites at the oxygen atoms at C1 (ketone), C19 (hydroxyl), and C26 (hydroxyl) in the original molecule. These requirements gave rise to structures (15-17) that maintained the recognition features but removed a significant amount of the peripheral substituents. These molecules demonstrated nanomolar binding constants when measured in displacement assays of tritiated phorbol esters, with the figures being in the same general range as bryostatin 1, and two compounds (15, 16) had activities in in vitro cell line assays close to those demonstrated by bryostatin 1 itself.63-66 Following on from these examples, modifications were made to the base structure (15) to introduce a second lactone (18) that had an 8 nM binding affinity and also inhibited P388 with an ED50 of 113 nM.67 Concomitantly, modifications were made to the base analogue (15) where different fatty acid esters were made (structures not shown). These, too, exhibited binding affinities for PKC isozymes in the 7-232 nM range depending upon the fatty acid used.68 To show the versatility of the base structure, recently Wender published a simple modification where by removal of a methyl group in the C26 side chain in compound 15 to produce compound 19, the binding affinity to PKC was increased to the picomolar level,69 and the compound demonstrated greater potency than bryostatin 1 in in vitro cell line assays. Finally, at the end of 2003, he published improved syntheses of the molecules, which could permit further rapid improvements of the model compound(s) with the potential for much greater overall yields.70,71 One very interesting question arising from the search for bryostatin sources was, why is the nominal producing organism so ubiquitous, but the number of B. neritina colonies that actually produce detectable bryostatin 1-3 levels so low and geographically spread? One possible answer to this question came from the work of Haygood and her collaborators at the Scripps Institution of Oceanography. Haygood showed that the bryozoan is actually the host to a symbiotic microorganism that may well be the actual producer of the compound; in an elegant series of experiments, she and her colleagues demonstrated by use of molecular probes the presence of a putative type I polyketide synthase (PKS) gene fragment in the microbial Reviews Journal of Natural Products, 2004, Vol. 67, No. 8 1221

flora of colonies that produced bryostatin but that was absent in the corresponding flora of nonproducers.72 In addition, Davidson and Haygood demonstrated that there are subdivisions within B. neritina samples taken from the same sites but at different depths. Thus, at depths greater than 9 m (the D or deep type), bryostatins 1-3 and minor components are found (these are also known as producers of chemotype O for “octa-2,4-dienoatic chain”), whereas, at less than 9 m (S or shallow type), only the minor derivatives are seen (chemotype M). The symbiotic microbes (Candidatus Endobugula sertula) isolated from each type differ in their mitochondrial carboxylase I (CO I) sequences by 8%, giving rise to the possibility that the bryozoans are also different taxonomically.73 There were reports at the Society for Industrial Microbiology (SIM) meetings in 2002 and 2003 that demonstrated that Haygood and collaborators were pursuing the possibility of transferring this particular PKS fragment to other, more amenable microbes in order to further investigate the possibility of producing bryostatin by fermentation. At the recent 6th International Marine Biotechnology Conference (IMBC) in Chiba, Japan, Haygood74 reported on the current status of the PKS search, suggesting that this system resembled that reported by Piel75,76 for the Paederus beetle’s pseudomonal symbiont PKS that produces pederine, in that there are no acyl-transferase (AT) domains in the clusters, unlike the usual PKS system, but that an AT domain was found in another, more remote, area of the overall PKS system. Further work is ongoing utilizing such “remote” AT domains from another organism. It will be very interesting to follow these results if Haygood is successful, as cultivation of the organism, or a surrogate with the bryostatin PKS system expressed, would potentially solve any supply problems if bryostatin becomes a commercial drug. Dolastatin Derivative, TZT-1027 (Auristatin PE or Soblidotin). As a result of the synthetic processes alluded to earlier, many derivatives of the dolastatins have been synthesized with TZT-1027 (20), now in Phase I clinical trials in Europe, Japan, and the United States under the auspices of either Teikoku Hormone, the originator, or the licensee, Daiichi Pharmaceuticals. This compound is also known as Auristatin PE and Soblidotin, and an initial report on Phase I studies was given in abstract form77 at the American Society for Clinical Oncology (ASCO) in 2002. Recently, a further report from investigators at Teikoku Hormone indicated that in nude mice the transfected vascular endothelial growth factor (VEGF)-secreting human lung cell line SBC-3/VEGF and also the mock transfected cell line were effectively totally inhibited as either early or advanced stage xenografts at levels of 1 or 2 mg‚kg-1, conditions under which only vincristine was similarly active and combretastatin was not, even at 500 mg‚kg-1. What was of significant interest in addition to these results was that TZT-1027 also exhibited a potent antivascular effect at these levels, thus suggesting that a dual mechanism might well be possible with this agent.78 Very recently, a multinational group of investigators demonstrated the potential for a directed delivery of this compound to prostate cancer cells, by using the upregulation of the adhesion molecule, E-selectin, that is found in the epithelium of prostate carcinomas and demonstrated that a monoclonal antibody directed to this protein with auristatin linked via a cathepsin B-labile linker gave more than 85% inhibition of growth of prostate carcinoma cell lines in mouse models.79 Dolastatin Derivative, Cematodin (LU-103793). Another derivative of dolastatin 15 known as Cematodin (21) (and also as LU-103793) was placed into Phase I clinical trials by BASF Pharma under their Knoll division for treatment of breast and other cancers. The results from six trials have been reported at the Phase I level with doselimiting toxicities being neutropenia or cardiotoxicity. A number of these trials used a very rapid bolus (5 min iv), and others used a longer time frame, even up to 5 days of continuous infusion; from them, the investigators’ recommended ranges for Phase II studies were at the 2.5-10 mg‚ M2 dose levels.80-85 The compound progressed into Phase II studies against malignant melanoma, metastatic breast cancer, and non-small-cell lung cancer, and reports demonstrated no objective responses in any of the trials86-88 although stable disease was seen in both the melanoma and breast cancer trials and there was a subjective increase in a quality of life measure in the lung trial. Currently, there is some dichotomy in the literature as to whether work is actively continuing with this compound; thus Amador et al. report Phase II trials still ongoing as of June 2003 in breast, ovarian, lung, prostate, and colon carcinomas,12 whereas it is now currently (01/2004) listed as discontinued in the Prous Ensemble database. Dolastatin Derivative, ILX651 (Synthadotin). There have been six scientific reports in the last two years on the Phase I studies with this agent (22), all as presentations at ASCO meetings,89-92 the American Association for Cancer Research (AACR) meeting,93 or the joint USEuropean (AACR-NCI-EORTC) molecular targets meeting.94 ILX651 is an orally active third generation dolastatin 15 derivative that was licensed by Ilex from BASF Pharma, and two reports95,96 in 2003 indicated that Ilex Oncology is initiating Phase II studies in melanoma, breast, and nonsmall-cell lung cancers, as there were responses in these tumors in Phase I patients. From a nonclinical perspective, dolastatin 15 has proven to be a useful “bioprobe” in tubulin interaction studies. Thus, by using tritium-labeled dolastatin 15, Hamel’s group at NCI97 recently reported that the vinca domain in tubulin may well be composed of a series of overlapping domains rather than being a single entity, as different levels/types of competition were found when selected tubulin interactive agents were used to investigate the binding characteristics of the labeled probe. Source(s) of the Dolastatins. Similarly to the situation with the bryostatins, there was always a potential question 1222 Journal of Natural Products, 2004, Vol. 67, No. 8 Reviews

Reviews Journal of Natural Products.2004.Vol.67.No.8 1223 with the dolastatins as to whether they were microbial in with Et743.110 Subsequently,he improved the synthetic origin,as peptides with unusual amino acids have been schema and developed a refined process that produced both well documented in the literature as coming from the Et743 and phthalascidin at much higher yields.111 Other Cyanophyta.In the past few years,this supposition has synthetic chemistry groups have continued work on the been shown to be fact.Thus,in 1998,workers at the basic compound,but as yet,none of their compounds have Universities of Guam and Hawaii reported98 the isolation had any biological activity reported in the literature.2. and purification of simplostatin 1(23)from the marine The natural compound was licensed by the University cyanobacterium Simploca hynoides.This molecule differed of Illinois to the Spanish Company PharmaMar for subse- from dolastatin 10 by the addition of a methyl group on quent development.Following very large-scale wild collec- the first N-dimethylated amino acid.Subsequently,in 2001 tions and aquaculture on both land and in-sea in efforts to the same groups reported the direct isolation of dolastatin obtain enough source material for further preclinical and 10 from another marine cyanobacterium that was known clinical workup,PharmaMar chemists performed an el- to be grazed on by D.auricularia.99 Dolastatin 10 was in egant semisynthesis from the marine Pseudomonas fluo fact isolated from the nudibranch following feeding of the rescens metabolite cyanosafracin B that provided cGMP cyanophyte,thus confirming the original hypothesis (per- grade Et743 from a 21-step synthetic process on a scale sonal communication.Dr.V.J.Paul). large enough to provide enough material for clinical trials. Very recently,the MOA of symplostatin 1 was evaluated This was feasible despite a low overall vield of 1.4%because both in vitro and in vivo,and it was shown to be similar to the starting material could be obtained on a large scale by dolastatin 10 but to be somewhat more toxic to mice at fermentation.The original paper114 together with a rela- comparable doses.100 In addition,two further examples of tively recent review article,115 both from the PharmaMar dolastatin-like peptides isolated from different collections group,should be consulted for further details as to syn- of the ubiquitous cyanophyte Lyngbya majuscula have thetic strategies,etc.,employed for production of this recently been reported in the literature,viz.,dolastatin 16 compound. from a Madagascan collection by Nogle and Gerwick101 and A number of reports have been published in the litera- homodolastatin 16 from a Kenyan collection by Davies- Coleman et al.,102 further evidence for the microbial source ture over the past few years giving possibilities as to the MOA(s)of Et743 when tumor cells are treated in vitro.A of these peptidic cytotoxins. significant problem with some of the reports is that the concentration(s)used in the experiments are orders of magnitude greater than those that demonstrate activity in vitro.These levels are in the low nanomolar to high picomolar range,and thus care should be taken when evaluating published work on the MOA of this compound. At physiologically relevant concentrations the MOAs of Et743 have been shown to include the following:effects on the Transcription-coupled Nucleotide Excision Repair process(TC-NER)116.117 and interaction between the Et743 DNA adduct and DNA transcription factors,in particular the NF-Y factor.118 In the recent review on Et743 by a Dutch group,19 further details as to other possible mech- anisms are given in their Table 1;the references that they Ecteinascidin 743.Antitumor activity from the ascid cite should be consulted for in-depth information and ian Ecteinascidia turbinata had been reported as early as discussion for other potential MOAs ascribed to Et743.As 1969 by Sigel et al.,103 but it was not until 1990 that the addenda to the results given in the paper above,there were structures of the active components were published simul- two presentations at the AACR-NCI-EORTC molecular taneously by Rinehart et al.104 and Wright et al.105 targets meeting in November 2003 reporting gene expres- The structure of the most stable member of the series. sion profiles on sarcoma lines using the"Oncochip",a 6700 known as Et743 from the absorption maximum,is shown gene array of genes prevalent in cancer cell proliferation. (24).The base structure,without the exocyclic isoquinoline The first,using cells from treated sarcoma patients,120 group,is a well-known chemotype106 originally reported reported that when the ICso values for Et743 were ~250 sized by Coreyl09 in a chemical"tour de force",and as a nM)those that are physiologically relevant.106 result of his synthetic approach,his group also made a The compound was placed into human clinical trials version where the exocyclic ring was a phthalimido sub- while these mechanisms were being worked out,and by situtent.This compound,phthalascidin,demonstrated 2002 it had been in over a 1000 patients in Phase I and significant activity in the same test systems used initially Phase II trialss covering a variety of cancers.Results from

with the dolastatins as to whether they were microbial in origin, as peptides with unusual amino acids have been well documented in the literature as coming from the Cyanophyta. In the past few years, this supposition has been shown to be fact. Thus, in 1998, workers at the Universities of Guam and Hawaii reported98 the isolation and purification of simplostatin 1 (23) from the marine cyanobacterium Simploca hynoides. This molecule differed from dolastatin 10 by the addition of a methyl group on the first N-dimethylated amino acid. Subsequently, in 2001, the same groups reported the direct isolation of dolastatin 10 from another marine cyanobacterium that was known to be grazed on by D. auricularia. 99 Dolastatin 10 was in fact isolated from the nudibranch following feeding of the cyanophyte, thus confirming the original hypothesis (personal communication, Dr. V. J. Paul). Very recently, the MOA of symplostatin 1 was evaluated both in vitro and in vivo, and it was shown to be similar to dolastatin 10 but to be somewhat more toxic to mice at comparable doses.100 In addition, two further examples of dolastatin-like peptides isolated from different collections of the ubiquitous cyanophyte Lyngbya majuscula have recently been reported in the literature, viz., dolastatin 16 from a Madagascan collection by Nogle and Gerwick101 and homodolastatin 16 from a Kenyan collection by DaviesColeman et al.,102 further evidence for the microbial source of these peptidic cytotoxins. Ecteinascidin 743. Antitumor activity from the ascidian Ecteinascidia turbinata had been reported as early as 1969 by Sigel et al.,103 but it was not until 1990 that the structures of the active components were published simultaneously by Rinehart et al.104 and Wright et al.105 The structure of the most stable member of the series, known as Et743 from the absorption maximum, is shown (24). The base structure, without the exocyclic isoquinoline group, is a well-known chemotype106 originally reported from microbes, where the compound classes are saframycins, naphthyridinomycins, safracins, and quinocarcins. Similar molecules were reported from marine mollusks, i.e., jorumycin from the nudibranch Jorunna funebris107 and from sponges, the renieramycins, with the latest variation, renieramycin J, being recently reported by Oku et al.108 However, with Et743, the exocyclic substituent was novel, as was the bridging sulfur. The yield from natural sources was very low, and in order to provide enough material to perform basic studies as to the MOA and in vitro and in vivo animal studies, significant amounts of the ascidian had to be collected from areas around the Caribbean. The compound was synthesized by Corey109 in a chemical “tour de force”, and as a result of his synthetic approach, his group also made a version where the exocyclic ring was a phthalimido subsitutent. This compound, phthalascidin, demonstrated significant activity in the same test systems used initially with Et743.110 Subsequently, he improved the synthetic schema and developed a refined process that produced both Et743 and phthalascidin at much higher yields.111 Other synthetic chemistry groups have continued work on the basic compound, but as yet, none of their compounds have had any biological activity reported in the literature.112,113 The natural compound was licensed by the University of Illinois to the Spanish Company PharmaMar for subsequent development. Following very large-scale wild collections and aquaculture on both land and in-sea in efforts to obtain enough source material for further preclinical and clinical workup, PharmaMar chemists performed an elegant semisynthesis from the marine Pseudomonas fluorescens metabolite cyanosafracin B that provided cGMP grade Et743 from a 21-step synthetic process on a scale large enough to provide enough material for clinical trials. This was feasible despite a low overall yield of 1.4% because the starting material could be obtained on a large scale by fermentation. The original paper114 together with a relatively recent review article,115 both from the PharmaMar group, should be consulted for further details as to synthetic strategies, etc., employed for production of this compound. A number of reports have been published in the literature over the past few years giving possibilities as to the MOA(s) of Et743 when tumor cells are treated in vitro. A significant problem with some of the reports is that the concentration(s) used in the experiments are orders of magnitude greater than those that demonstrate activity in vitro. These levels are in the low nanomolar to high picomolar range, and thus care should be taken when evaluating published work on the MOA of this compound. At physiologically relevant concentrations the MOAs of Et743 have been shown to include the following: effects on the Transcription-coupled Nucleotide Excision Repair process (TC-NER)116,117 and interaction between the Et743 DNA adduct and DNA transcription factors, in particular the NF-Y factor.118 In the recent review on Et743 by a Dutch group,119 further details as to other possible mechanisms are given in their Table 1; the references that they cite should be consulted for in-depth information and discussion for other potential MOAs ascribed to Et743. As addenda to the results given in the paper above, there were two presentations at the AACR-NCI-EORTC molecular targets meeting in November 2003 reporting gene expression profiles on sarcoma lines using the “Oncochip”, a 6700 gene array of genes prevalent in cancer cell proliferation. The first, using cells from treated sarcoma patients,120 reported that when the IC50 values for Et743 were ∼250 nM) those that are physiologically relevant.106 The compound was placed into human clinical trials while these mechanisms were being worked out, and by 2002 it had been in over a 1000 patients in Phase I and Phase II trials8 covering a variety of cancers. Results from Reviews Journal of Natural Products, 2004, Vol. 67, No. 8 1223

1224 Journal of Natural Products.2004.Vol.67.No.8 Reviews the European Phase I and pharmacokinetic trials were into Phase I clinical trials in 1999 under the auspices of recently reported by Twelves et al.,122 and details of the PharmaMar in Canada,Spain,France,and the U.K.for human pharmacokinetics(PK)and activities against bone treatment of both solid tumors and non-Hodgkin's lym- tumor cells in vitro were also published recently.123 In 2001. phoma.A summary of five of the trial results is given in Et743 was licensed to Johnson and Johnson (Ortho Bio- Table 2 of Amador et al..12 which should be consulted for tech)under the brand name Yondelis,with the generic specific dosage details,and the actual abstracts from the name of trabectedin.Two recent full reports on the Phase three ASCO meetings may be consulted for further II trials have been published 19.124 giving details of toxicities information.140-144 These were successfully completed with and response levels in sarcomas and other carcinomas with over 200 patients and demonstrated that a dosage of up to both pretreated and naive patients,and at the November 5 mg.M2 was well tolerated in either a 3 or 24 h infusion 2003 AACR-NCI-EORTC molecular targets symposium, every other week.135 The DLT was muscle pain that was there were a further series of reports showing objective responsive to either dose limitation or addition of carnitine. responses in long-term follow-up studies in sarcoma in Interestingly,in the presence of carnitine,the maximum Phase II studies,125 preliminary results from a combination tolerated dose could be increased by 40%to 7 mg-M2.Phase study of Et743 and doxorubicin in untreated sarcoma and II clinical trials are now underway in Europe comparing non-anthracycline-treated breast cancer patients where PR the two dosage regimens in renal and colon carcinomas, and SD were observed,126 and the potential for the use of together with an outpatient regimen,and very recently(07/ paclitaxel and ET-743 where a PR has been observed in a 2003),the European Commission's COMP/EMEA awarded Phase I study.127 The article by the Dutch group gives in- orphan drug status145 for acute lymphoblastic leukemia depth discussions of most of the so-far reported trial (ALL)to aplidin.Other Phase II trials are also ongoing in results,119 and for further information on other aspects,the Europe and Canada covering renal,head and neck,and review by D'Incalci and Jimeno should also be consulted.128 medullary thyroid,but no patient details have yet been As a result of these earlier trials,Et743 was preregis- published. tered in the EU and granted orphan drug status for sarcoma The precise MOA of this agent is not yet known,but it by the European Commission's Committee for Orphan appears to block VEGF secretion and blocks the corre- Medicinal Products(COMP)of the European Agency for sponding VEGF-VEGF-Receptor-1 (also known as flt-1) the Evaluation of Medicinal Products(EMEA).However, autocrine loop in leukemic cells.146 In addition,effects on in late July 2003,the EU's Committee for Proprietary Med- the kinasesINK,src,and p38-MAPK,possibly mediated icinal Products(CPMP)recommended,on a majority vote. via glutathione depletion,were recently reported,147 with that marketing authorization for advanced soft tissue the end result being induction of the apoptotic cascade in sarcoma not be granted for the EU.This decision was ap MDA-MB-231 breast cancer cells at levels of 5 nM.below pealed in September 2003 by PharmaMar,129 but in Decem- the blood levels achievable in man.In these experiments, ber 2003 the appeal was denied.130 The compound was general caspase inhibitors decreased apoptotic efficacy by granted orphan drug status for ovarian cancer by the CPMP ~50%,thus implicating at least two different mechanisms during the appeal process on sarcoma referred to above.131 of apoptosis,one via caspases,the other not involving Further evidence as to the possibilities of combination caspase activation.Of significant interest are the recent studies has been reported by D'Incalci et al.,132 where work reports by Straight et al.on the effects of aplidine on ARO- with mice demonstrated that there was synergy against 81 anaplastic thyroid cancer cells148 and of Bravo et al.on the Et743-resistant/cisplatin-partially resistant ovarian cell human thyrocytes from different pathologies.149 In the first line HOC 8 when cisplatin was added to the treatment case,induction of apoptosis was observed together with a protocol.Both drugs were used at their maximum tolerated reduced or absent expression of angiogenic genes,and in dose (MTD)levels,thus demonstrating that although the second case,a low but constant apoptotic rate was synergy occurred with activity,there was no cross/ established that caused over 90%reduction in cell numbers synergistic toxicity shown. within 72 h at 100 nM aplidine.Thus in these cell types, One of the predominant toxicities exhibited by Et743 in aplidine had both a cytotoxic and an antiangiogenic effect. preclinical studies was hepatotoxicity.particularly in the In leukemic cells obtained from pediatric patients,ap female rat,and similar effects had been seen in human lidine demonstrated little cross-resistance with other cy- patients but could be controlled by dose-reduction.How- totoxic drugs,and in particular,bone marrow cells from ever,in a recent publication,Donald et al.133 demonstrated normal patients were 2-7 times more resistant to aplidine that pretreatment with high-dose dexamethosone gave than the cells from leukemia patients,indicating that complete protection against hepatotoxicity in this animal. studies with other cytotoxins could be justified in cancer Thus such a treatment in humans may well be a method patients.The original paper should be consulted for specific of controlling this Et743-related toxic side effect. sets of drug combinations/level of interactions.150 Further Currently Et743 is in a variety of Phase II trials in the evidence for a lack of myelosuppression by aplidine,Et743, United States and Europe for the treatment of sar- and kahalide F,compounds currently in Phase II,II/III, coma134.135 and is listed as being in Phase III in Europe in and II,respectively,has been reported by the PharmaMar the Prous Ensemble database at time of writing. group using a murine competitive repopulating model as Aplidine.This compound,formally dehydrodidemnin B the test system,but these findings will have to be con- (25).was first reported in a patent application in 1989,with firmed in human patients/bone marrow cells as well.151 a U.K.patent issued136 in 1990 and then referred to in the What is very interesting.both chemically and pharma- paperl5 from Rinehart's group in 1996 on the structure- cologically,is that the removal of two hydrogen atoms,i.e. activity relationships among the didemnins.The antitumor conversion of the lactyl side chain to a pyruvyl side chain, potential was first reported by PharmaMar scientists137.138 appears to significantly alter the toxicity profile,as this is in 1996,and the total synthesis was reported in a patent the only formal change in the molecule when compared to application39 in 2000 and the patent was issued in 2002. didemnin B.although the comments on dosage regimens The compound,generic name "aplidine or dehydrodi- under didemnin B (vide supra)from Vera and Joullie!9 demnin B"and with a trade name of Aplidin,was placed should be taken into account when such comparisons are

the European Phase I and pharmacokinetic trials were recently reported by Twelves et al.,122 and details of the human pharmacokinetics (PK) and activities against bone tumor cells in vitro were also published recently.123 In 2001, Et743 was licensed to Johnson and Johnson (Ortho Biotech) under the brand name Yondelis, with the generic name of trabectedin. Two recent full reports on the Phase II trials have been published119,124 giving details of toxicities and response levels in sarcomas and other carcinomas with both pretreated and naive patients, and at the November 2003 AACR-NCI-EORTC molecular targets symposium, there were a further series of reports showing objective responses in long-term follow-up studies in sarcoma in Phase II studies,125 preliminary results from a combination study of Et743 and doxorubicin in untreated sarcoma and non-anthracycline-treated breast cancer patients where PR and SD were observed,126 and the potential for the use of paclitaxel and ET-743 where a PR has been observed in a Phase I study.127 The article by the Dutch group gives indepth discussions of most of the so-far reported trial results,119 and for further information on other aspects, the review by D’Incalci and Jimeno should also be consulted.128 As a result of these earlier trials, Et743 was preregistered in the EU and granted orphan drug status for sarcoma by the European Commission’s Committee for Orphan Medicinal Products (COMP) of the European Agency for the Evaluation of Medicinal Products (EMEA). However, in late July 2003, the EU’s Committee for Proprietary Medicinal Products (CPMP) recommended, on a majority vote, that marketing authorization for advanced soft tissue sarcoma not be granted for the EU. This decision was appealed in September 2003 by PharmaMar,129 but in December 2003 the appeal was denied.130 The compound was granted orphan drug status for ovarian cancer by the CPMP during the appeal process on sarcoma referred to above.131 Further evidence as to the possibilities of combination studies has been reported by D’Incalci et al.,132 where work with mice demonstrated that there was synergy against the Et743-resistant/cisplatin-partially resistant ovarian cell line HOC 8 when cisplatin was added to the treatment protocol. Both drugs were used at their maximum tolerated dose (MTD) levels, thus demonstrating that although synergy occurred with activity, there was no cross/ synergistic toxicity shown. One of the predominant toxicities exhibited by Et743 in preclinical studies was hepatotoxicity, particularly in the female rat, and similar effects had been seen in human patients but could be controlled by dose-reduction. However, in a recent publication, Donald et al.133 demonstrated that pretreatment with high-dose dexamethosone gave complete protection against hepatotoxicity in this animal. Thus such a treatment in humans may well be a method of controlling this Et743-related toxic side effect. Currently Et743 is in a variety of Phase II trials in the United States and Europe for the treatment of sarcoma134,135 and is listed as being in Phase III in Europe in the Prous Ensemble database at time of writing. Aplidine. This compound, formally dehydrodidemnin B (25), was first reported in a patent application in 1989, with a U.K. patent issued136 in 1990 and then referred to in the paper15 from Rinehart’s group in 1996 on the structureactivity relationships among the didemnins. The antitumor potential was first reported by PharmaMar scientists137,138 in 1996, and the total synthesis was reported in a patent application139 in 2000 and the patent was issued in 2002. The compound, generic name “aplidine or dehydrodidemnin B” and with a trade name of Aplidin, was placed into Phase I clinical trials in 1999 under the auspices of PharmaMar in Canada, Spain, France, and the U.K. for treatment of both solid tumors and non-Hodgkin’s lymphoma. A summary of five of the trial results is given in Table 2 of Amador et al.,12 which should be consulted for specific dosage details, and the actual abstracts from the three ASCO meetings may be consulted for further information.140-144 These were successfully completed with over 200 patients and demonstrated that a dosage of up to 5 mg‚M2 was well tolerated in either a 3 or 24 h infusion every other week.135 The DLT was muscle pain that was responsive to either dose limitation or addition of carnitine. Interestingly, in the presence of carnitine, the maximum tolerated dose could be increased by 40% to 7 mg‚M2. Phase II clinical trials are now underway in Europe comparing the two dosage regimens in renal and colon carcinomas, together with an outpatient regimen, and very recently (07/ 2003), the European Commission’s COMP/EMEA awarded orphan drug status145 for acute lymphoblastic leukemia (ALL) to aplidin. Other Phase II trials are also ongoing in Europe and Canada covering renal, head and neck, and medullary thyroid, but no patient details have yet been published. The precise MOA of this agent is not yet known, but it appears to block VEGF secretion and blocks the corresponding VEGF-VEGF-Receptor-1 (also known as flt-1) autocrine loop in leukemic cells.146 In addition, effects on the kinases JNK, src, and p38-MAPK, possibly mediated via glutathione depletion, were recently reported,147 with the end result being induction of the apoptotic cascade in MDA-MB-231 breast cancer cells at levels of 5 nM, below the blood levels achievable in man. In these experiments, general caspase inhibitors decreased apoptotic efficacy by ∼50%, thus implicating at least two different mechanisms of apoptosis, one via caspases, the other not involving caspase activation. Of significant interest are the recent reports by Straight et al. on the effects of aplidine on ARO- 81 anaplastic thyroid cancer cells148 and of Bravo et al. on human thyrocytes from different pathologies.149 In the first case, induction of apoptosis was observed together with a reduced or absent expression of angiogenic genes, and in the second case, a low but constant apoptotic rate was established that caused over 90% reduction in cell numbers within 72 h at 100 nM aplidine. Thus in these cell types, aplidine had both a cytotoxic and an antiangiogenic effect. In leukemic cells obtained from pediatric patients, aplidine demonstrated little cross-resistance with other cytotoxic drugs, and in particular, bone marrow cells from normal patients were 2-7 times more resistant to aplidine than the cells from leukemia patients, indicating that studies with other cytotoxins could be justified in cancer patients. The original paper should be consulted for specific sets of drug combinations/level of interactions.150 Further evidence for a lack of myelosuppression by aplidine, Et743, and kahalide F, compounds currently in Phase II, II/III, and II, respectively, has been reported by the PharmaMar group using a murine competitive repopulating model as the test system, but these findings will have to be confirmed in human patients/bone marrow cells as well.151 What is very interesting, both chemically and pharmacologically, is that the removal of two hydrogen atoms, i.e., conversion of the lactyl side chain to a pyruvyl side chain, appears to significantly alter the toxicity profile, as this is the only formal change in the molecule when compared to didemnin B, although the comments on dosage regimens under didemnin B (vide supra) from Vera and Joullie19 should be taken into account when such comparisons are 1224 Journal of Natural Products, 2004, Vol. 67, No. 8 Reviews

made in the future. Similarly, the resemblance to didemnin B is emphasized by the recent work of Cardenas et al., who reported152 that in DMSO solution aplidine, like didemnin B, does not exhibit a formal â-turn in its side chain in approximately 20% of its solution conformers, thus suggesting that the presence of such a turn is not required for biological activity. As the authors point out, there may well be other, as yet unrecognized minor conformers that are responsible for some/all of the biological activities demonstrated. Halichondrin B (and Derivatives). Halichondrin B (26) is one of a series of compounds originally isolated and reported153,154 by Uemura et al. in 1985 from the Japanese sponge Halichondria okadai. Subsequently, a number of sponges from other areas of the Pacific and Indian Oceans were reported to contain one or more of these macrolides, including Axinella sp. from the Western Pacific,155 Phakellia carteri from the Eastern Indian Ocean,156 and from a deep water Lissodendoryx sp. off the East Coast of South Island, New Zealand.157 Although there was enough halichondrin B available for some initial experiments and to determine that the possible mechanism of action was as an tubulin interactive agent, affecting tubulin depolymerization at a site close to, but distinct from, the vinca site,158-160 and to show initial in vivo activity,161 there was not enough material for further development work. In 1992, NCI issued a request for groups that could provide a variety of scarce natural products from natural sources, and a consortium from New Zealand composed of the University of Canterbury (who had discovered that a deep water Lissodendoryx sp. produced the halichondrins at approximately 1 mg‚kg-1 wet weight) and the National Institute for Water and Atmospheric Research (NIWA) was successful in convincing the NCI to fund a large-scale recovery and isolation program as a joint venture with them and the New Zealand Government. Following an environmental assessment of the potential collection area paid for by the Developmental Therapeutics Program (DTP) of the NCI, the NZ Government gave permission to collect 1 metric tonne from the Kaikoura shelf at a depth of 100 m and greater by trawling. Following extensive workup, these samples produced 300 mg of halichondrin B, but what was just as important, were the experiments conducted by NIWA scientists (also partially funded by DTP/NCI) that demonstrated that the deepwater Lissodendoryx could be successfully aquacultured in water as shallow as 10 m and still produce the halichondrin complex at levels roughly comparable with those found from wild collections. Concomitantly with the start of this large-scale wild collection program, Kishi’s group at Harvard, also funded by the DTP/NCI, reported that they had successfully synthesized both halichondrin B and norhalichondrin B.162 Working with the U.S. division of the Japanese pharmaceutical company Eisai, Kishi’s synthetic schemes were utilized to synthesize a large number of variants of halichondrin B, particularly smaller molecules that maintained the biological activity but were intrinsically more chemically stable, due to the substitution of a ketone for the ester linkage in the macrolide ring. Two of these agents were subsequently evaluated by DTP in conjunction with the Eisai Research Institute in the United States, and one of the compounds, originally ER-086526 (NSC 707389) and now renumbered E7389 (27), entered Phase I clinical trials in 2002 in conjunction with the NCI. At the 2003 ASCO meeting, there were two presentations on E7389, one showing pharmacokinetics of this agent in man163 in the current Phase I trial demonstrating that levels above those required for cytotoxicity in vitro were achievable for up to 72 h at doses below the DLT of 0.5 mg‚M2, and the other demonstrating that this agent exhibits p53-independent anticancer activity versus nonsmall cell lung cancer (NSCLC) in vitro at the 0.5 pM level,164 orders of magnitude below the 1500 pM levels achievable in man. Thus NSCLC may well be a worthwhile target for this agent. Details of the biology and chemistry of this compound and other compounds in the series have recently been published by both the Harvard165 and Eisai scientists,166,167 thus demonstrating the power of current synthetic chemistry when applied to a very potent marine-derived natural product. Using the synthetic techniques described, enough cGMP material, produced by total synthesis, was provided to the NCI for the initial clinical trials. Discodermolide. This polyhydroxylated lactone (28) was reported by the Harbor Branch group in 1990 following isolation from the Caribbean sponge Discodermia dissoluta, originally collected at a depth of 33 m off the Bahamas,168 with a revision to the stereochemistry being published the following year.169 Originally, the compound was judged to be a new immunosuppressive and an incidental cytotoxin.170-172 However, in 1996, it was reported that discodermolide bound to microtubules more potently than Taxol, a discovery that confirmed in silico studies at the University of Pittsburgh.173 Concomitantly with these reports, a variety of chemical synthetic groups had seen discodermolide as a good candidate for total synthesis. Thus the initial report from Harbor Branch (which as noted above was later corrected) led to synthesis of the (-)-isomer by Schreiber’s group174 and others, and then in the late 1990s-2003, Marshall and Johns,174 Halstead,175 Smith et al.,176 and Paterson et al.177 all reported syntheses that would produce varied isomers in good yield. Recently, Paterson and Florence have published an excellent reReviews Journal of Natural Products, 2004, Vol. 67, No. 8 1225