J. Nat. Prod. 2009, 72, 507-515 (cid:9)
`
`507
`
`Contributed Reviews
`
`Tubulin-Interactive Natural Products as Anticancer Agents'
`
`David G. I. Kingston*
`
`Department of Chemistry, M/C 0212, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061-0212
`
`Received September 10, 2008
`
`This review provides an overview of the discovery, structures, and biological activities of anticancer natural products
`that act by inhibiting or promoting the assembly of tubulin to microtubules. The emphasis is on providing recent
`information on those compounds in clinical use or in advanced clinical trials. The vinca alkaloids, the combretastatins,
`NPI-2358, the halichondrin B analogue eribulin, dolastatin 10, noscapine, hemiasterlin, and rhizoxin are discussed as
`tubulin polymerization inhibitors, while the taxanes and the epothilones are the major classes of tubulin polymerization
`promoters presented, with brief treatments of discodermolide, eleutherobin, and laulimalide. The challenges and future
`directions of tubulin-interactive natural products-based drug discovery programs are also discussed briefly.
`
`Introduction
`
`Natural products have proven to be the most reliable single source
`of new and effective anticancer agents. Thus Newman and Cragg
`have shown that 63% of anticancer drugs introduced over the last
`25 years are natural products or can be traced back to a natural
`products source,2 and similar observations have been made by many
`others.3-7 A recent review by Butler lists 79 natural products or
`natural product analogues that entered clinical trial as anticancer
`agents in the 2005-2007 time frame.8 Natural products have not
`only yielded new and effective drugs but have also provided insight
`into new mechanisms of action, and cancer treatment would be
`immeasurably poorer without the insights and the compounds
`provided from Nature. The reasons for the effectiveness of natural
`products are at least twofold. In the first place, there is a high
`correlation between the properties of drugs and those of natural
`products 9'10 Second, natural products usually have built-in chirality
`and are thus uniquely suited to bind to complex proteins and other
`three-dimensional biological receptors.
`Among the various mechanisms of action of natural products,
`that of interaction with the cellular protein tubulin is one of the
`most important, and over 25% of the new clinical candidates listed
`by Butler operate by this general mechanism." Two major classes
`of anticancer drugs owe their effectiveness to this mechanism; the
`first class is that of the tubulin polymerization inhibitors, and the
`second is that of tubulin polymerization promoters. This review
`covers natural products or modified natural products that interact
`with tubulin and that are in clinical use or are in advanced
`development toward clinical use.
`The cellular protein tubulin is a crucial protein for cellular
`replication. The cell cycle involves the replication of DNA and
`the packaging of the resulting replicated chromosomes into two
`daughter cells. The separation of the daughter chromosomes in
`mitosis is brought about by microtubules, which are formed by the
`polymerization of a- and f3-tubulin. The microtubules radiate in
`cells from centrosomes and from the poles of mitotic spindles and
`are formed by attachment of GTP-tubulin to the growing end of
`an existing microtubule. Microtubules undergo rapid assembly and
`disassembly in cells, and this property enables those associated with
`the mitotic spindle to generate a large collection of structures and
`thus to produce those structures that will interact constructively
`
`* Tel: (540) 231-6570. Fax: (540) 231-3255. E-mail: dkingston@vt.edu.
`
`with the centromeres of daughter chromosomes and generate the
`necessary aligned chromosomes at metaphase. The normal func-
`tioning of tubulin assembly and disassembly is thus crucial to cell
`division, and any interference with this process will disrupt cell
`division and cause cell death by apoptosis.
`Although the most dramatic effect of the tubulin-interactive drugs
`is that of changing the extent of microtubule polymer mass, either
`decreasing it for the tubulin polymerization inhibitors or increasing
`it for the tubulin polymerization promoters, cancer cell growth can
`be inhibited at concentrations significantly lower than those
`necessary to exert these macroscopic effects. This fact can be
`explained by the observation that cell growth inhibition at low
`concentrations is caused by the suppression of microtubule dynam-
`ics.'1
`The structure of the tubulin heterodimer has been solved by
`electron diffraction.' Both the vinca alkaloids and the taxane drugs
`bind to /3-tubulin, but at different locations on the protein; the vinca
`alkaloids bind to /3-tubulin between amino acids 175 and 213,13
`while paclitaxel binds both to an N-terminal unit of fi-tubulin14
`and to the region bounded by amino acids 217-231.'5 Colchicine,
`which is not a clinically used drug for cancer but which has been
`studied extensively, binds between the two subunits. The epothilones
`also bind at the paclitaxel site.16
`
`Inhibitors of Tubulin Polymerization
`
`The Vinca Alkaloids. The first natural products to enter clinical
`use were the bisindole alkaloids vinblastine (1) and vincristine (2).
`These complex compounds were isolated from the Madagascar
`periwinkle Catharanthus roseus (L.) G. Don (previously known
`as Vinca rosea L.) in the late 1950s and early 1960s by two
`independent groups. Their discovery makes an interesting story,
`because one of the groups working on them, that of Robert Noble
`and Charles Beer at the University of Western Ontario, was actually
`looking for substances that could affect blood glucose levels, and
`the discovery of the antileukemic activity of the extract was made
`after the serendipitous observation of its effects on white blood
`cell counts. These observations led to the isolation and structure
`elucidation of the bisindole alkaloid vincaleukoblastine (1); the
`name was later shortened to vinblastine." The related alkaloid
`leurocristine (later named vincristine, 2) was isolated by Gordon
`Svoboda and his colleagues at Eli Lilly." The alkaloids bind to
`fi-tubulin at a different site from the taxane drugs and colchicine
`
`10.1021/np800568j CCC: $40.75 © 2009 American Chemical Society and American Society of Pharmacognosy
`Published on Web 01/06/2009
`
`AVENTIS EXHIBIT 2039
`Mylan v. Aventis
`IPR2016-00627
`
`

`
`508 Journal of Natural Products, 2009, Vol. 72, No. 3 (cid:9)
`
`Reviews
`
`and act to prevent tubulin assembly.19 The compounds consist of
`two subunits: an upper catharanthine ring system linked to a lower
`vindoline ring system by a single bond.
`
`OH (cid:9)
`
`z (cid:9)
`
`OH
`
`N
`H
`Me0C0
`
`Me0
`
`N
`H
`Me0C0
`
`Me0
`
`N
`H
`Me0C0
`
`OCOMe
`COOMe
`
`Me0
`
`H
`Me
`
`1
`
`OCOMe
`HO COOMe
`CHO
`2
`
`OCOMe
`
`Me HO COOMe
`
`4
`
`N
`H
`Me0C0
`
`Me0
`
`OCOMe
`
`N
`Me HO COOMe
`
`6
`
`Vinblastine and vincristine have been used in clinical oncology
`for almost 50 years, and their use has been revieWed.29 Vincristine
`is used in combination chemotherapy of acute lymphoblastic
`leukemias and lymphomas, while vinblastine is used in combination
`chemotherapy to treat bladder and breast cancers. Perhaps the most
`significant impact of vinblastine has been as part of the curative
`regimen for Hodgkin's disease.
`Vindesine (3) was the first analogue of vinblastine to enter
`clinical use. It differs from vinblastine in having an amide function
`rather than a methyl ester on the vindoline ring and in lacking an
`acetyl group on this ring system. It has a somewhat higher
`hematological toxicity than vincristine, but it has been incorporated
`into several effective combination regimens for treatment of
`leukemia, lymphoma, and non-small cell lung cancer (NSCLC).21,22
`Vinorelbine (4) is a semisynthetic derivative of vinblastine in
`which the bridge linking the indole ring to the piperidine nitrogen
`has been shortened by one carbon, and water has been eliminated
`from the piperidine ring. It was launched in 1989 by Pierre Fabre
`for the treatment of nonmetastatic breast cancer and NSCLC, and
`it is available both in iv and oral formulations.'
`Vinflunine (Javlor, 5) is a dihydrodifluoro derivative of vinorel-
`bine. It can be prepared by treatment of vinorelbine (4) with HF/
`SbF5 in CHC13; the proposed mechanism involves chlorination of
`the cation generated by protonation of the cyclohexenyl double bond
`and isomerization. The chloro compound then loses a hydride ion
`and is fluorinated to a chlorofluoro compound, which is finally
`converted to vinflunine.26 Vinflunine interacts with tubulin in a
`qualitatively similar way to vinblastine, but detailed studies indicate
`that it has quantitatively different properties from the classic vinca
`alkaloids.27 It is in phase III trials at Pierre Fabre for the treatment
`of bladder cancer and NSCLC, based on the observation of clinically
`significant activity in phase II studies for the treatment of bladder,
`non-small cell lung, and breast cancers,26.28 and it is also being
`evaluated for second-line chemotherapy in hormone refractory
`prostate cancer (HRPC) and for HER2-overexpressing metastatic
`breast cancer (with Trastuzumab),29 and for ovarian cancer.39
`Anhydrovinblastine ((Hydravin, KRX-0403, 6) is an analogue
`of vinblastine that differs from its parent by one molecule of water.
`It can also be considered a homologue of vinorelbine with an
`additional carbon in the indole-piperidine bridge. It entered phase
`I trial for the treatment of advanced solid tumors, including NSCLC,
`soft tissue sarcoma, and colorectal cancer. Stable disease was noted
`in one patient with metastatic sarcoma to the lungs and in three
`patients with metastatic NSCLC.31 Keryx discontinued development
`in 2005, but trials may still be ongoing by Prescient Neuropharma.8
`Combretastatins and Analogues. The vinca alkaloids are the
`only inhibitors of tubulin polymerization in clinical use, but several
`
`compounds with this mechanism of action are in advanced clinical
`trials and will be discussed here.
`Combretastatin A-4 (7, CA4) was originally isolated by Pettit et
`al. from the root bark of the Combretum caffrum tree, also known
`as the Cape Bushwillow.32 It has been shown to target the
`microtubule, inhibiting the polymerization of tubulin to microtu-
`bules. However, it is much more cytotoxic than its activity against
`tubulin would seem to warrant, and four explanations have been
`proposed to explain this discrepancy: (i) CA4 targets, in vivo, a
`subpopulation of tubulin; (ii) CA4, alongside tubulin, recognizes a
`yet unidentified relevant target; (iii) the interaction between tubulin
`and CA4 is qualitatively different; (iv) CA4, for as yet unknown
`reasons, rapidly and preferentially disrupts in vivo cellular processes
`in the endothelium that require tubulin assembly.33 Combretastatin
`A4 phosphate (CA4P, 8), is a simple derivative of the natural
`product that was prepared to increase water solubility.34 CA4P has
`also been shown to be a vascular disrupting agent, and data from
`phase I studies have established that it can selectively reduce tumor
`blood flow at well-tolerated doses.35 Its vascular effect can be
`explained by its disruption of the endothelial cytoskeleton.36 CA4P
`is in phase III clinical trials sponsored by Oxigene, Inc. for treatment
`of cervical, colorectal, NSC lung, prostate, ovarian, and thyroid
`cancers; reviews of the clinical results to date have appeared
`recently.37'38
`
`R
`
`OMe
`
`7 R = OH
`8 R = OPO3Na2
`11 R = NHCOCH(NH2)CH2OH
`
`9 R = H
`10 R = OPO3Na2
`
`.NHCOMe
`
`OPO3Na2
`
`12
`
`Combretastatin A-4 is accompanied in the Cape Bushwillow with
`numerous congeners, differing in the ring substitutions (CA series)
`and by reduction of the internal stilbene double bond to give the
`CB series.39 Combretastatin A-1 (9) is a simple hydroxyl derivative
`of combretastatin A-4, and it is also in phase I clinical development
`as its diphosphate derivative OXi4503 (10).8 A study of its effects
`on mice indicated that substantial microvascular damage to liver
`tumors and minimal normal liver injury occurred, and it was
`concluded that "a combination of OXi4503 with other chemothera-
`peutic modalities might achieve complete tumor eradication and
`improve long-term survival".49
`
`

`
`Reviews (cid:9)
`
`The relative simplicity of the combretastatin skeleton has
`spawned the synthesis of numerous analogues. The medicinal
`chemistry of these compounds is beyond the scope of this review,
`but it is covered in detail in recent reviews:41-43 Two of the many
`analogues that have been prepared are in advanced development.
`The serinamide derivative AVE8062A (11, AC-7700) is in clinical
`trials in Europe and the United States, and its primary therapeutic
`effect is to reduce blood flow to the tumor.36'45 The compound
`ZD-6126 (ANG453, 12) can be considered as an analogue of both
`combretastatin and colchicine, and it is rapidly converted in vivo
`into N-acetylcolchinol. It too is a vascular disrupting agent.46 Its
`phase II trials were halted by AstraZeneca due to problems with
`the method of administration, but Angiogene has solved this
`problem.8 The development of combretastatin A-4 and these
`analogues as anticancer agents has been reviewed.43 Although it is
`too early yet to tell which (if any) of these combretastatin analogues
`will enter clinical use, it is safe to predict that at least one and
`probably more than one drug based on the combretastatins will
`enter clinical use over the next few years.
`NPI-2358. Fungi have yielded relatively few tubulin inhibitors,
`but one such is the diketopiperazine halimide (13a). NPI-2358 (13b)
`is a simple analogue of halimide, which is active as a tubulin-
`depolymerizing agent." It is in phase I trials at Nereus.• 8
`
`0
`
`13a
`
`0
`
`O
`
`13b
`
`Halichondrin B and Eribulin. Complex marine natural
`products of the halichondrin class were isolated by Uemura and
`Hirata from the western Pacific sponge Halichondria okadai49 and
`later by Pettit et al. from an Axinella sp.5° Several members of the
`class showed strong cytotoxicity, with the most potent being
`homohalichondrin B and halichondrin B (14). These compounds
`were shown to bind to tubulin and to inhibit tubulin polymerization.
`Halichondrin B is a noncompetitive inhibitor of vinblastine binding
`to tubulin and has no effect on colchicine binding.51 It showed
`subnanomolar activity in the NCI 60-cell line pane152 and excellent
`activity in various animal models," and it was thus a clear candidate
`for clinical development. The major obstacle to clinical development
`was the issue of compound supply, since it was obtained only in
`miniscule amounts from its marine source.
`
`Journal of Natural Products, 2009, Vol. 72, No. 3 509
`
`at Eisai Research Institute for activity. This led to the realization
`that several truncated halichondrins had significant bioactivity, and
`so a research program was initiated to develop a simplified and
`thus synthetically accessible anticancer agent based on the hali-
`chondrin B skeleton. The truncated halichondrin B analogue eribulin
`(15, E7389) was discovered as a result of this program, and its
`synthesis was effected in a highly convergent manner, although
`this still required over 70 steps.54.55 The synthesis of this highly
`cytotoxic compound on an industrial scale had its own unique
`challenges, but these were successfully overcome so that eribulin
`mesylate could enter clinical trials. Like its parent compound,
`halichondrin B, eribulin acts as an inhibitor of tubulin polymeri-
`zation.56 Eribulin mesylate is in phase III trials for the treatment
`of prostate, sarcoma, breast, NSCL, bladder, head and neck, and
`ovarian cancers. Its discovery and development have been re-
`viewed,57 and its clinical potential has also been reviewed.58'59 An
`encouraging overall response rate of 15% was observed in a phase
`II trial in NSCLC patients, and the compound also showed
`promising results in treatment of breast cancer. One significant
`advantage over other tubulin-interactive agents is that eribulin
`appears to have a lower neurotoxicity than the other agents. It is
`concluded that "E7389 would probably be a very welcome addition
`to the available agents used to treat women with advanced breast
`cancer".59
`Dolastatins. Dolastatin 10 (16) was isolated from the sea hare
`Dolabella auricularia by Pettit as part of an extensive series of
`investigations; it is the most potent member of a fairly large class
`of related compounds.6 •
`4 It binds to tubulin at a distinct site for
`peptide antimitotic agents near the exchangeable nucleotide and
`vinca alkaloid sites, inhibiting tubulin polymerization.62 The relative
`simplicity of the structure and the difficulty of sourcing the natural
`product made chemical synthesis the preferred approach, and
`dolastatin 10 has been synthesized by the Pettit group63 and others.64
`Dolastatin 10 entered phase I clinical trials in the 1990s, with the
`finding that 40% of patients developed moderate peripheral neur-
`opathy.65 It then progressed to phase H trials under the auspices of
`the National Cancer Institute (NCI) for the treatment of several
`solid tumors, including liver, bile duct, gallbladder cancer, pan-
`creatic cancer, and advanced kidney cancer, but the results of these
`trials were not encouraging, as summarized in two recent reviews 37.66
`
`14
`
`15
`
`Fortunately for the future of this class of compounds, Kishi
`developed a total synthesis of halichondrin B.53 The synthesis was
`designed in a convergent fashion, and his intermediates were tested
`
`16
`
`17
`
`18
`
`The novel activity of the dolastatins spurred the development of
`several analogues, and the simplified analogue, TZT-1027 (auristatin
`PE, Soblidotin, 17), was among several analogues prepared for SAR
`studies.67 It was selected for further development based on its
`reduced toxicity as compared with dolastatin 10. Clinical results
`to date are mixed, with antitumor activity observed in some cases
`but not others.37
`
`

`
`510 Journal of Natural Products, 2009, Vol. 72, No. 3
`
`A second derivative in clinical development is the dolastatin 15
`analogue tasidotin (18). Tasidotin is also a tubulin-interactive drug,
`weakly inhibiting tubulin polymerization to microtubules but
`strongly suppressing the dynamic instability of microtubules.68 It
`has completd phase I trials69•7° and is currently in phase II trials
`under Genzyme.8 Further details of the discovery and development
`of the dolastatins are available in a comprehensive review.65
`Noscapine. Noscapine (19) is an old compound, being a
`commonly used antitussive agent without significant side effects,71
`but it has recently been found to bind to tubulin and alter its
`conformation, assembly properties, and microtubule dynamics.72
`It shows good oral bioavailability in mice73 and is active against
`H460 NSCLC cells in nude mice.74 The analogue EM105 (20) is
`more potent and regresses breast cancer xenografts in nude mice
`without significant toxicity.75 Noscapine and its analogues are thus
`interesting lead compounds, and noscapine is in phase I trials by
`Cougar Biotechnology.76
`
`CI
`
`Hemiasterlin. Hemiasterlin (21a) is a tripeptide that was first
`isolated by Kashman from the sponge Hemiasterella minor, with
`reported activity against the P388 murine leukemia cell line.77 It
`was later reisolated by Andersen, who reported that it had
`antitubulin activity, producing abnormal mitotic spindles at low
`concentrations and microtubule depolymerization at higher con-
`centrations.78 Synthetic studies identified the phenylalanine analogue
`HTI-286 (21b) as a more accessible and more potent analogue,79
`and both HTI-2868° and hemiasterlin8 are in clinical trials."
`
`0
`
`Rhizoxin. Rhizoxin (22) was isolated in 1984 together with
`several congeners from the plant pathogenic fungus Rhizopus
`chinensis.82 It is an inhibitor of tubulin polymerization83 and is
`more potent than maytansine against human and murine tumor
`cells.84 It has been synthesized85 and has been evaluated in clinical
`trials,86 but has not yet entered clinical use.
`
`Promoters of Tubulin Polymerization
`
`The Taxanes. Paclitaxel (23) was isolated by Wall and Wani in
`the 1960s from bark of the Pacific yew, Taxus brevifolia, and given
`the name taxo1.87 Its initial discovery was greeted with underwhelming
`enthusiasm, because of the obvious problems of compound supply
`and solubility, and because it showed only relatively modest in vivo
`activity against the then current antileukemic models at the NCI.88
`Fortunately, the B16 melanoma solid tumor assay was introduced in
`the early 1970s, and paclitaxel showed excellent and reproducible
`activity against this solid tumor, with an increase in life span (ILS) of
`126%, 32%, and 86% in three separate experiments.88 Even with these
`data the problems of developing paclitaxel loomed large, but fortunately
`the late Matthew Suffness, who had joined the NCI in 1976, recognized
`
`paclitaxel's potential and was instrumental in presenting the case for
`its development to the NCI Decision Network Committee. This
`committee approved paclitaxel as a development candidate in 1977.88
`
`Reviews
`
`Ph NH 0
`
`Ph
`
`OH
`
`AcO U OH
`
`OS H _ 0
`HO AcO
`OCOPh
`
`23
`
`In vivo studies of paclitaxel in the then new human solid tumor
`xenograft assays in nude mice were carried out in 1978, and these
`were rewarded when it was found to show strong activity against
`the MX-1 breast xenograft; this discovery was important in
`maintaining interest in the compound.88 The final significant finding
`came with the discovery by Horwitz that paclitaxel promoted the
`assembly of microtubules;89 this new mechanism was a crucial
`added factor in raising interest in this compound. The solubility
`problem was overcome by the development of an emulsion
`formulation in Cremophor, a polyethoxylated castor oil, which
`unfortunately caused allergic reactions in some patients. These
`problems were overcome by premedication with antihistamines and
`the use of extended intravenous infusions, which were initially for
`24 h but were later reduced to 3 h.9°
`Paclitaxel, or taxol as it was then known, was found to have
`clinical activity against ovarian cancer in 198991 and against breast
`cancer in 1991.92 Further development was taken over by Bristol-
`Myers Squibb (BMS) in 1991 under a Cooperative Research and
`Development Agreement (CRADA) with the NCI. BMS was able
`to trademark the name Taxol for their formulation of the drug, and
`the generic name paclitaxel was applied to the chemical compound
`formerly known as taxol. Taxol was approved by the FDA in
`December 1992 for the treatment of refractory breast cancer and
`refractory ovarian cancer and was launched on the U.S. market by
`Bristol-Myers Squibb the following year. It was later approved for
`treatment of breast cancer after failure of combination chemotherapy
`for metastatic disease, for second-line treatment of AIDS-related
`Kaposi's sarcoma, and for NSCLC in combination with cisplatin.93
`Extensive continuing clinical trials are evaluating combinations of
`paclitaxel with other drugs for treatment of many other cancers 94.95
`The enormous demand for paclitaxel caused by its excellent
`activity created a significant supply crisis, which was initially solved
`by aggressive collection of T. brevifolia bark. A viable semisynthetic
`route was then developed. In this semisynthesis the more readily
`available 10-deacetylbaccatin III (24) is first converted to 7-trieth-
`ylsilylbaccatin III (25), and this is coupled with the protected
`fl-lactam 26 in the key step, yielding the protected paclitaxel 27.
`Final deprotection gives paclitaxel (23) in excellent overall yield
`(Scheme 1).96 This synthetic route essentially ended the supply
`crisis. Paclitaxel is now also produced commercially by plant tissue
`culture methods.97
`As noted above, paclitaxel acts by promoting the assembly of
`tubulin into microtubules. This results in the inhibition of the normal
`dynamic reorganization of the microtubule network that is essential
`for vital cellular functions and leads ultimately to cell death by
`apoptosis. The actual mechanism by which paclitaxel stabilizes the
`microtubule is still under investigation, but Xiao et al. speculate
`that "Taxol induces a loss of flexibility in the involved regions that
`prevents a "roll out" of lateral contacts in microtubules that would
`otherwise open up their wall".98
`The success of paclitaxel has spurred enormous interest in finding
`improved analogues as well as improved formulations designed to
`circumvent the problems associated with the Cremophor formula-
`tion, and several analogues are currently in clinical trials. The only
`analogue approved for clinical use to date in the United States is
`docetaxel (28), a semisynthetic analogue developed in France.99
`Docetaxel was launched in 1995 and is the drug of choice in treating
`advanced NSCLC that is refractory to primary therapy.'°° It is also
`
`(cid:9)
`

`
`Reviews (cid:9)
`
`Scheme 1. Semisynthesis of Paclitaxel
`
`Journal of Natural Products, 2009, Vol. 72, No. 3 511
`
`TIPSO, Ph
`
`Ph
`
`0
`
`O
`
`26
`
`R.1 0 0 OR2
`
`DMAP,
`pyridine
`96%
`4.°
`HO''sH _
`AcO
`HO
`OCOPh
`24 R1 = R2 = H
`25 R1 = Ac, R2 = SiEt3
`
`0
`Ri NH 0
`Ph
`0'
`OTIPS
`
`0.5% HCI
`23
`
`AcO 0 OSiEt3
`
`91%
`I,.
`0
`_ H
`HO Ac0
`OCOPh
`27
`
`used in the treatment of hormone refractory prostate cancer,101,102
`advanced breast cancer,103 and other cancers including head and
`neck, stomach, and ovarian cancers.1°4
`
`0
`
`~OA
`Ph
`
`0
`
`NNH
`
`OH
`
`HO O OH
`
`HO AcO
`OCOPh
`
`28
`
`In addition to paclitaxel and docetaxel, the albumin nanoparticle
`(nab) formulation of paclitaxel Abraxane was approved in 2005 in
`the United States for the treatment of advanced breast cancer. Since
`Abraxane consists only of albumin-bound paclitaxel nanoparticles,
`it is free of Cremophor and requires no premedication. A recent
`review concludes "these studies have demonstrated that nab
`technology has increased the therapeutic index of paclitaxel
`compared with the conventional, solvent-based formulation",105
`while another review states "Abraxane was safe, effective, induced
`higher response rate and longer time to progression compared with
`Taxol in patients with metastatic breast cancer".1°6
`Although paclitaxel and docetaxel are the only taxanes currently
`approved for clinical use, there are many analogues in clinical trials,
`and the most important of these are summarized briefly below.
`There are four taxanes in phase III clinical trials. Larotaxel
`dihydrate (29, Sanofi-Aventis)1°7 is an interesting derivative of
`docetaxel in which the methyl group at the C-8 position has formed
`a cyclopropyl ring; the paclitaxel analogue of 29, with a phenyl
`group replacing the tertiary-butyloxy group in the side chain of
`29, was formed on treatment of 7-epi-paclitaxel with DAST.1°8
`Larotaxel is in phase III trials for treatment of breast and pancreatic
`cancers, and a report from a phase II trial indicated that it has a
`favorable therapeutic index in women with taxane-pretreated
`metastatic breast cancer.I°9 Paclitaxel poliglumex (30, Xyotax) is
`a conjugate of paclitaxel with a biodegradable polyglutamic acid;
`this feature was designed to increase water solubility and improve
`its pharmacokinetic profile. It is in phase III trials by Cell
`Therapeutics for the treatment of NSCLC and ovarian cancer.'10.111
`Cabazitaxel (31, XRP-6258, TXD258, Sanofi-Aventis) is a dimethoxy
`derivative of docetaxel that is in phase III trials in combination
`with prednisone for treatment of hormone refractory metastatic
`prostate cancer.12 It has the advantages that it is not a substrate
`for P-glycoprotein13 and that it can cross the blood—brain barrier.
`Taxoprexin (32, DHA-paclitaxel, Luitpold Pharmaceuticals) is a
`2'-acyl paclitaxel and is in phase III trials for treatment of NSCLC
`and in phase II trials for several other cancers. Phase II studies on
`the compound have been reported.' 14.115
`Compounds in phase II clinical trials include five new compounds
`and two formulations of paclitaxel. NK-105 is a polymeric micellar
`macromolecule encapsulated formulation of paclitaxel in phase II
`clinical trials at Nippon Kayaku for the treatment of solid tumors;
`it is reported to be a more potent radiosensitizing agent than
`paclitaxe1.116 EndoTAG-1 is a formulation of paclitaxel encapsu-
`lated in a positively charged lipid-based complex and is in phase
`II clinical trials at MediGene for the treatment of advanced
`pancreatic cancer."' TPI-287 (33, NBT-287, Tapestry Pharma-
`ceuticals) is in phase II trials for treatment of advanced pancreatic
`cancer and hormone refractory prostate cancer. It is designed to
`overcome acquired resistance to taxane-based therapies by circum-
`
`AcO 0
`
`0
`
`But0A NH 0
`
`Ph _
`OH 2E1,0
`
`_ 0
`AcO
`OCOPh
`
`HO
`
`29
`
`CH3O 0 OCH3
`
`0
`
`ButOANH 0
`
`Ph _ 0'
`OH
`
`0
`
`Fl (cid:9)
`OH AcO
`OCOPh
`
`31
`
`0
`
`PhANH 0
`
`Ph
`
`0
`O o
`
`AcO 0 OH
`
`0.0
`
`HO O
`OCOPh
`
`0
`11n)c
`b n
`
`COOH
`
`COOH
`
`30 (a + b averages 10)
`
`AcO 0 OH
`
`O
`
`Ph'''''NH 0
`
`Ph _ 0'
`0 O
`
`H 0
`0 Ac0
`OCOPh
`
`32
`
`venting the P-glycoprotein drug efflux mechanism, and it is more
`potent than paclitaxel in paclitaxel-resistant tumors. Results of its
`phase I studies have been reported:18'119 Ortataxel (34) was
`originally developed by Bayer, but the clinical trials encountered
`severe neutropenia12° and were discontinued. The compound was
`then licensed exclusively to Spectrum Pharmaceuticals in 2007. It
`is currently in phase II clinical trials for the treatment of non-small
`cell lung cancer (NSCLC),121 and it has also shown in vivo activity
`in animal models against head and neck squamous cell carcinoma
`(HNSCC).122 Milataxel (35), which was developed by the new
`company Taxolog, is in phase II trials under Wyeth.118 One recent
`report of a phase II study in colorectal cancer indicated that it had
`the side effect of neutropenic sepsis, necessitating close surveillance
`if the drug is to be used at its MTD.123 Tesetaxel (36) was
`developed by Daiichi Sankyo Co. Ltd. for treatment of colorectal
`and gastric cancer. It was withdrawn from development in 2006
`because of its failure to show clear benefit over existing agents,'24
`but it was recently licensed by Genta, who hope to restart trials.8
`BMS-188797 (37) is a simple 4-carbonate derivative of paclitaxel
`that showed objective responses in four out of 16 patients in a phase
`I trial, including three complete remissions in ovarian and cervical
`cancer patients;125 a phase I study of the drug in combination with
`carboplatin has also been reported.126 Nab-docetaxel, the nanopar-
`ticle albumin bound docetaxel formulation related to Abraxane, is
`in early clinical studies at Abraxis BioScience for the treatment of
`hormone refractory prostate cancer and other solid tumors.127
`Four compounds that were in phase I trials, simotaxel (38,
`Wyeth) and TL-310 (39), from Taxolog, and the Bristol-Myers
`Squibb analogues BMS-275183 (a 4-carbonate derivative, 40) and
`BMS-184476 (41), are no longer in clinical trials. Phase II trials
`on BMS-275183 were halted by BMS, phase I trials on BMS-
`184476 have been completed,128 phase I trials on simotaxel were
`halted by Wyeth, and there is no information on TL-310 on
`clinicaltrials.gov.8
`The Epothilones. Epothilones A (42) and B (43) were discovered
`as antifungal agents in 1986 as metabolites of the myxobacterium
`Sorangium cellulosum.129 They attracted only moderate scientific
`interest until the report by Bollag et al. in 1995 that they had the same
`mechanism of action as paclitaxel, stabilizing the tubulin polymer and
`causing apoptotic cell death.13° This report triggered a large amount
`of research on the biology and chemistry of these compounds, which
`has been reviewed on several occasions.129"1-133 The progression
`
`(cid:9)
`(cid:9)
`

`
`512 Journal of Natural Products, 2009, Vol. 72, No. 3
`
`Reviews
`
`Sagopilone (ZK-EPO, 45) is a fully synthetic epothilone, in contrast
`to patupilone, which is a natural fermentation product, and ixabepilone,
`which is a modified fermentation product.142 It is in phase II clinical
`development at Bayer Schering Pharma for the treatment of recurrent
`ovarian cancer, metastatic breast cancer, small-cell lung cancer, prostate
`cancer, and other solid tumors.143 It has good activity in MDR models,
`probably because it evades the PgP