Modern and Emerging Perspectives on Natural Product Utilisation

NATURAL PRODUCTS, DERIVATIVES AND MIMICS AS ANTITUMOUR AGENTS

David J. Newman and Gordon M. Cragg

Natural Products Branch, Developmental Therapeutics Program, NCI-Frederick, P. O. Box B, Frederick, MD, 21702, USA

(Note: The opinions expressed in this article are the opinions of the authors, not necessarily those of the US Government)

1 INTRODUCTION

From early in the 1940s, the quest for agents that may ameliorate the scourge of the manifold diseases clustered under the term cancer, has involved all aspects of chemistry and pharmacology and throughout all these years, compounds from natural sources, microbes, plants, and latterly, marine invertebrates, have played an extremely important part.

The roles played by these natural products have changed with time and the increase of scientific knowledge. Thus, their initial role was as the major source of drugs used for direct treatment. This was followed by use as active scaffolds upon which chemists would practice their skill and in current times, natural products and derivatives are acting as modulators of specific cellular pathways in the tumour cell.

Currently, the 164 (as of June 2009) small molecule compounds available to the physician (depending upon their individual country) as antitumour agents can be categorized as follows: N (natural products: 25; 17%), ND (natural product-derived, usually semi-synthetic modifications: 50; 31%), S (totally synthetic drugs: 42; 26%), S/NM (synthetic drugs/natural product mimics: 16; 10%), S* (synthetic drugs inspired by natural products: 20; 12%) and S*/NM (synthetic drugs inspired by natural products/natural product mimics: 7; 4%), using the nomenclature of Newman et al. Recently another author, using a somewhat different series of definitions where only "direct and slightly modified natural products are counted as such", has shown the influence of natural product scaffolds using a different schematic and readers should consult Bailly's article for further information.

As a further refinement, the influence of microbial secondary metabolites on approved agents in cancer treatment can be seen in Table 1. Of the 47 commercially available compounds listed (including everolimus, which was approved for anticancer use after being approved for immunosuppression), 26 (or 55%) are either directly from microbes or are derivatives of microbial secondary metabolites. Trabectedin 1 is probably from microbial commensals rather than from the marine tunicate, Ecteinascidia turbinata. The remaining 21 agents (45%) are nominally synthetic but the major proportion are either modified nucleosides whose chemical antecedents can be traced back to Bergmann's discoveries of marine-sourced nucleosides containing arabinose, or are isosteres of ATP.

Since the full history of such sources with their manifold digressions into chemical and pharmacological space would be a massive undertaking, we have elected to highlight the influence of certain natural product classes that have derivatives in or approaching clinical trials in order to demonstrate how even today, in the midst of massive efforts related to combinatorial chemical techniques coupled to rational drug design, natural products from many ecological niches are still a major source of novel scaffolds upon which to base potential antitumour agents. For a discussion of the earlier history and derivation of the major classes of compounds derived from nature that are in current use, the reader is directed to the 2005 compendium on Anticancer Agents from Natural Products and the articles therein.

Due to space constraints generally we will discuss only the agents that are currently in Phase II and Phase III clinical trials in general, though where a particular class of compounds spans the range from preclinical to Phase III, we will include most of the molecules in order to demonstrate the breadth of possibilities.

2 NON-MARINE MICROBIAL AGENTS CURRENTLY IN CLINICAL TRIALS

Currently there are at least 17 agents derived from identified microbial sources in various clinical trials against a wide variety of human tumours in both paediatric and adult patients. These encompass structures based on staurosporins, anthracyclines, and bicyclic thiobridged compounds, with mechanisms of action (MOA) including inhibitors of topoisomerase I, II and histone deacetylase (HDAC).

2.1 Staurosporin Based Molecules

Staurosporin 2 is a pan-protein kinase C (PKC) inhibitor and the role of the base structure is evident if one looks at the number of derivatives that are in clinical trials at this moment. Becatecarin 3, a rebeccamycin 4 analogue with a sugar moiety that has been extensively modified, is claimed not to be a PGP substrate, though it is transported by ABCG2 and can induce ABCG2 expression in cancer cells. It is in trials for leukaemia and other tumour types both with and without other cytotoxins. A recent report shows efficacy in a Phase II trial against advanced biliary cancers with an overall response rate (partial response in 5% and stable disease in 35%) of 40%, which has encouraged extension to a Phase III trial in due course. A current listing of trials with this agent can be found via the "clinicaltrials.gov" web site and data in early 2009 can also be found in the "Gateways to Clinical Trials" report by Tomillero and Moral.

A reasonably close relative to staurosporin is the indolocarbazole K-252a 5, derivatives of which have been placed into clinical trials in neurodegenerative diseases as well as in cancer. For cancers, the simple derivative lestaurtinib 6 is currently in five trials, with four being at the Phase II level (two in myelofibrosis and one each in acute myelogenous leukaemia and polycythemia vera), and one at Phase I (paediatric neuroblastoma). The same basic structure is also in Phase II/III clinical trials as midostaurin 7 and the ring-opened version of staurosporin, known as enzastaurin 8, is in multiple Phase II trials with some recent positive effects 11 and has reportedly just entered Phase Ill trials in lymphoma. Enzastaurin also has potential in the treatment of human transitional cell carcinoma in a preclinical model with gemcitabine. Although not formally an antitumour compound, the modified maleimide, sotrastaurin 9 is also a PKC inhibitor and is in Phase II trials for psoriasis, which is a skin disease that antitumour agents may well help in controlling. This compound also has activity as an immunomodulator and is in Phase II clinical trials for this indication as well. Whether it will follow in the footsteps of the rapamycin analogues (see below) has yet to be determined.

2.2 Anthracycline Based Molecules

The anthracyclines, isolated from bacteria of the order Actinomycetales, are probably the most utilized microbial chemical class of antitumour agents, both directly from nature and as semi-synthetic modifications, with two of the most useful being daunorubicin 10 and its natural derivative doxorubicin (adriamycin) 11. Currently, doxorubicin is a major component of the treatment regimen for breast cancer. Although there have been many similar molecules isolated and described in the literature, it is doxorubicin and its more modem derivatives such as epirubicin 12, pirirubicin 13, idarubicin 14 and, more recently, valrubicin 15 and amrubicin 16 that have been approved for cancer treatment. The 2005 review by Arcamone should be consulted for details of their history.

Just to demonstrate that the base structure is still in contention as a drug in the 21st century, there are several liposomal, PEGylated or proprietary formulations of doxorubicin and daunorubicin being investigated in phase I/II/III clinical trials, including DOXO-EMCH 17, NK-91 1 (a pegylated doxorubicin), SP-1049C (doxorubicin in a Biotransport™ carrier), CPX-351 (liposomal mixture of daunorubicin and cytarabine) and Sarcodoxome (liposomal doxorubicin). The targeted version of doxorubicin (TAP-doxorubicin 18) where a four residue peptide aids in transportation and acts as an inhibitor until cleaved off by intracellular peptidases, is in phase II trials in Africa and Europe under the code number DTS-201 for the treatment of breast cancer and hormone-refractory prostate cancer respectively.

In addition to doxorubicin-based drug candidates, there are other modified anthracyclines at various stages of development. Thus in phase I/II trials, there is CNDO-101 19 in which a quinone moiety has been chemically modified and the cardiac toxicity is reported to be absent as a result. There are three more chemical modifications of the doxorubicin skeleton in phase II trials: annamycin 20, a liposomal variant of which is also in phase I/II trials for leukaemia; sabarubicin 21, in which the major structural modification is an extra sugar moiety; berubicin hydrochloride 22, which has a benzyl ether attached to the sugar moiety, reported to permit crossing of the blood-brainbarrier, was removed from Phase II clinical trials against glioma as a result of a business decision, but the structural features are still important.

2.3 Enediynes

One of the most important recently approved microbial compounds (as the warhead for a monoclonal antibody [Mab]-linked delivery system) is the enediyne, calicheamicin γ1I23. The reason for its importance is that although Calicheamicin γ1I has in vitro cytotoxic activity at the sub-picomolar level, it was not developed further as a single compound as it was just too toxic to pursue in spite of its exquisite activity. In addition to its very potent activity, it was also the first non-protein bound structure identified of a new chemical class, the enediynes; though the neocarzinostatin chromophore 24 was identified in 1985. These agents currently number 13 with the two basic structural types (the so-called 9- and 10 membered endiynes), differing in the number of carbon atoms in the endiyne system. Two of the 13, sporolides A and B (25, 26) and cyanosporasides A and B (27, 28) are rearrangement products of putative enediynes. These agents have been well reported by various groups, the Wyeth (originally Lederle) discoverers of the calicheamicins from Micromonospora echinospora ssp calichensis covering the ten-membered category, while the chemistry and biosynthesis of the molecules is reviewed by a number of investigators.

Following binding to DNA and subsequent activation, these agents undergo an unprecedented rearrangement which causes cleavage of both strands of DNA leading to the death of the cell. As mentioned earlier, Wyeth gained FDA approval for gemtuzumab ozogamicin (Mylotarg®), in 2000. This monoclonal antibody-warhead construct for use against chronic myologenous leukaemia, is possibly the most potent antitumour agent yet approved for clinical use. Due to the success of this type of construct, calicheamicin has been linked to a variety of other monoclonal antibodies which are in various clinical trials. CMD-193, where the antibody is directed against an anti-Lewis Y antigen has now been withdrawn from Phase I trials but CMC-544 or inotuzumab ozogamicin, which is a conjugate of calicheamicin and an anti-CD22 monoclonal antibody is in five Phase I/II/Ill trials against a variety of lymphomas. If further information is desired on such agents, the recent review article by Castillo et at gives details of the current impact of monoclonal antibodies, with or without microbial-sourced attachments, on cancer treatments.

In addition to the isolation, development and biosynthetic work mentioned above on the overall class, a significant number of papers covering aspects of synthesis have been published in the last few years. These include work that permitted reassessment of the original structures by direct comparison of the possible isomers obtained by synthesis with the data from the natural product.

Perhaps the best current example in this respect is the synthetic effort around uncialamycin. The original structure was determined on less than 500 micrograms of material by Davies et al. following isolation from an unidentified streptomycete extracted from the lichen Cladonia uncialis, that appeared to be related to Streptomyces cyanogenus. Due to the lack of material, the discoverers were unable positively to determine the stereochemistry at position 26 in the natural product. Nicolaou's group rapidly synthesized both of the enantiomers and demonstrated that the 26 (R) enantiomer 29 was the natural product and the 26 (S) or epi-enantiomer 30 the unnatural one. Further biological investigations have shown that both epimers were very potent antibiotics (confirming the original work of Davies et al), and also that they were potent cytotoxins in the NCI 60 cell line panel, with 5 to 10 fold less activity for the 26-epi enantiomer. This work demonstrated that even molecules as complex as these are amenable to current synthetic methods.

2.4 Rapamycins and Epothilones

Rapamycin 30 was originally reported in 1975 as a potential antifungal agent under the code number AY-22,989 from an Easter Island isolate identified as a Streptomyces hygroscopicus. Unfortunately, the antifungal activity was not adequate for further development, but in 1984 its potential as a possible antitumour agent was reported by workers at Ayerst Canada using syngeneic murine tumours together with a hint of oral activity.

The antitumour activity was not developed at that time but the rapamycin skeleton has now spawned a plethora of molecules with a variety of different pharmacologic activities including cancer. The first modifications were at one site (the carbon atom at C43) and led to four more clinical drugs in addition to rapamycin. The rapamycin base molecule was approved as sirolimus 31 in 1999, initially as an immune-suppressive agent and now the same molecule is in Phase I/II trials against various cancers.

Everolimus 32, the second variant, was launched in 2004 as an immunosuppressive agent and it is currently in further Phase II/III trials for various cancers in the EU, Japan and the USA having recently been approved for renal cancer in the EU and the USA under the name Afinitor®. The third variation temsirolimus (CCI-779) 33 was approved as a treatment for renal carcinoma in the USA in 2007 under the name Toricel and is in a number of Phase I/II/III trials against various carcinomas in the USA, with a number under the auspices of the National Cancer Institute. The fourth, zotarolimus 34 was launched in the USA in 2005 for treatment of restenosis as part of a drug-eluting stent and is still in a number of trials comparing its activity against various different drug stents. The fifth, biolimus A9 35 was launched in 2008 where the rapamycin derivative was linked to the stent via a biodegradable polymer.

Although there are five clinical agents as shown above, the skeleton is still under refinement with the following agents at various stages of development. The closest compound to possible approval at the present time is deforolimus (A23573) 36 which is in Phase III clinical trials against sarcoma. There are two prodrugs of rapamycin currently in phase I trials, Abraxis' ABI-009 (which is a nanoparticle encapsulated formulation of rapamycin) in cancer and lsotechnika's TAFA-93 (structure not yet published), which is an immunosuppressive. It should be noted that all of these (where structures are known) are either the base molecule or have been modified at only the one site, the C43 alcoholic hydroxyl group that avoids both the FKBP-12 and the Target of Rapamycin (TOR) binding sites. Modifications in other areas were thought to negate the basic biological activity of this molecule.

However, a rapamycin derivative ILS-920 37 with a modified macrolide ring structure is currently in Phase 0 (first in man) clinical trials and is headed for Phase I trials shortly. ILS-920 has a modification in the triene portion of the molecule, designed to disrupt mTOR binding and appears to have a different target as it is a nonimmunosuppressive neurotrophic rapamycin analogue and has demonstrated over a 200 fold higher binding affinity for FKBP52 over FKBP12. It has been reported to promote neuronal survival and outgrowth in vitro and to bind to the β1 subunit of L-type calcium channels (CACNB 1). Inhibition of FKBP 52 is reported to affect tubulin interactions in cells so there is an interesting possibility that this agent may also have antitumour activity, though no reports of such activity have yet been published. Finally, the patent literature reveals other variations that are in preclinical studies in a variety of pharmacologic areas.