Skip to Main Content
Skip Nav Destination

Three broad chemical classes of bioactive macrocyclic natural products are discussed in this chapter: macrolidic antibiotics, macrolides that have antitumor or immunological effects and cyclic peptides that may or may not contain lactone (depsipeptide ) linkages. In a number of cases, particularly from marine sources, they have been identified from assessment of biosynthetic clusters discovered through analysis of the genomes of symbiotic microbes. Examples are given from each of these groups, including short introductions to ansamycin-type Hsp90 inhibitors and the myxobacterial metabolites, the epothilones. Due to the fact that a number of molecules isolated from one source maybe produced by another or even a consortium, the molecules are grouped for discussion according to their biological activities rather than their nominal source(s). Also briefly discussed are some synthetic studies on macrocycles, in some cases de novo synthetic and, in others, based upon natural product structures.

Natural products have been an important source of pharmacologically active molecules throughout the history of medicinal chemistry and a selection of these molecules has advanced to provide clinically validated, marketed therapeutics for numerous indications, most notably as antibiotics, immunosuppressives and anti-tumor agents. Among these, structurally diverse and complex naturally derived macrocycles have demonstrated an impressive record of efficacy as pharmaceutical agents, and are playing an increasingly important role in the treatment of a range of serious diseases. These macrocycles have received intense recent interest from the medicinal chemistry community driven in part by their activity in biological systems, such as those mediated by protein–protein interactions, that are difficult to prosecute with more typically drug-like small molecules. In addition, the selectivity afforded by their complexity and the remarkable ability of some of these macrocyclic natural products to provide significant systemic exposures on oral dosing despite physical properties that lie substantially outside of normal parameters for achieving oral bioavailability, make them an attractive chemical class. Thus, these macrocyclic natural products and related biosynthetically and semi-synthetically derived macrocycles are of great value not only for their own intrinsic pharmacological activity but also for their potential as tools used to understand how to design molecules with the properties necessary to produce highly effective therapeutics for the treatment of human disease.

Further testaments to the current level of focus on macrocycles in drug discovery are the excellent reviews published over the last five years on key aspects of the medicinal chemistry of this chemotype.1–4  The most recent of these reviews, by Giordanetto and Kihlberg,1  provides an analysis of the physical properties of macrocyclic drug molecules, including cLogP, polar surface area, hydrogen bond donor count and molecular weight, in relationship to their ability to be orally efficacious. In addition, in 2008, Driggers et al.2  provided an excellent overview of macrocycles as drug leads and candidates at that time. This was followed by a review in 2011 by Marsault and Peterson3  showing the use of macrocycles over a wide area of medicinal chemistry and, in 2012, a review from Mallinson and Collins focusing primarily on potential anticancer agents.4 

In their 2014 review1  Giordanetto and Kihlberg reported 68 registered macrocyclic drugs approved for human use (identified through mining of the GVK BIO online structure-activity relationship database (GOSTAR)), along with a set of 35 macrocyclic drug candidates in clinical development (identified using the Adis R&D Insight database). The latter set did not include compounds in early clinical trials whose structures were not in the public domain.1  Most of these drugs, which have macrocyclic rings comprised of 12 or more atoms, fall into three chemical classes, namely macrolidic antibiotics, macrolides that have antitumor or immunological effects, and cyclic peptides that may or may not contain lactone (depsipeptide) linkages.

Of the 68 identified macrocyclic registered drugs,1  34 are used for the treatment of infections, mainly of bacterial origin, while 10 are used for the treatment of a variety of cancers; the remaining 28 are applied in cardiovascular, gynecological and immunological therapeutic areas, as well as in a range of indications, including anesthesiology and pain (these numbers do not total 68 since some agents have two or more activities). The majority of these molecules are natural products or directly derived from natural products (48 and 18, respectively), while eribulin (vide infra) is totally synthetic but modeled on the marine natural product, halichondrin B, and Sugammadex is a modified γ-cyclodextrin. Nineteen of the 68 drugs are administered orally, with 15 of these belonging to the macrolide classes. The parenterally administered macrocyclic drugs include all of the cyclic peptides, with the exception of cyclosporin A which is orally delivered.

As with the registered drugs briefly discussed above, of the 35 macrocycles identified as agents in clinical development 14 are under investigation for the treatment of various cancers, 10 are in infectious disease trials, and the remaining 11 are under investigation for indications ranging from endocrinology to ophthalmology.1  While these agents are also predominantly natural products (17) or natural product-derived molecules (8), with the largest chemical class being cyclic peptides (11), 10 of the clinical candidates are of de novo design. A total of 43% of these clinical candidates are administered orally, a significant increase over the 28% of all registered macrocyclic drugs that are orally administered, and nine of the 10 de novo designed macrocycles in trials are orally bioavailable. The increasing numbers of orally active macrocyclic drug candidates indicates that organic and medicinal chemists are learning to apply the lessons provided by bioactive natural macrocyclic agents, such as those presented in this chapter, for the design of fully synthetic molecules having desirable pharmacological properties, including oral bioavailability.

In strongly endorsing the impressive therapeutic record of natural product-derived macrocycles reported in earlier reviews and commentaries, the following sections will expand on the discussion of the macrocyclic chemical classes mentioned above, while adding a number of other chemical compounds that fall under the general description of “macrocycles” that have originated from plant and marine sources. Some of these compounds have arisen from the assessment of biosynthetic clusters that has led to the identification of novel agents, frequently not associated with the “expected” macro-organism source.

Also included are two short sections covering natural product macrocycles, which are described in detail in two later chapters of this volume. In the case of the ansamycin Hsp90 inhibitors, some of the very early work has been referenced, as the authors of this chapter were involved in the initial production of geldanamycin for the NCI's work with 17-AAG and 17-DMAG. This is followed by some current examples, where the manipulation of biosynthetic clusters in the producing bacteria has led to ansamycin structures that are completely novel and Hsp90 active. Similarly, with the epothilones, which are covered extensively in Chapter 3, particularly from a synthetic chemistry aspect, it is shown how genetic manipulation of the base-producing cluster and expressing it in heterologous hosts permitted the production of quantities of the four basic epothilones, and also materially aided in the utilization of myxobacteria as sources of novel agents, mainly via genomic techniques.

It can be successfully argued that the discovery of the actinomycins by Waksman and Woodruff in 19405  led to at least two firsts: the first crystalline antibiotic and the first demonstration of anti-tumor activity (actinomycin C) in vitro.6  This was followed by a report later in 1952 by Schulte demonstrating the first clinical studies with these agents.7  Over the last 60 plus years, actinomycins, usually as actinomycin D (Figure 1.1, 1), have been used as treatments for a variety of tumor types. Currently actinomycin D is used primarily for the treatment of rhabdomyosarcoma and Wilms’ tumor in children and young adults, with a very recent example being its reported use in the treatment of a patient with uterine embryonal rhabdomyosarcoma,8  50 years after its formal launch in 1963.

Figure 1.1

Structures 1 to 8; Actinomycin D; Erythromycin and Early Derivatives.

Figure 1.1

Structures 1 to 8; Actinomycin D; Erythromycin and Early Derivatives.

Close modal

Two major mechanisms of action, involving intercalation of DNA and stabilization of cleavable complexes of topoisomerases I and II with DNA, have been discussed in a review by Mauger and Lackner.9  In 2003, Gniazdowski et al. reported that actinomycin D interacted with downstream proteins associated with transcription, and this activity appeared related to its DNA-intercalating ability, but at levels well below those demonstrating anti-tumor activity.10  In contrast, in a recent review, Leung et al. did not confirm this earlier report but did show that another complex depsipetide, echinomycin demonstrated such activity at the transcription factor level (vide infra).11  Studies reported in 2009 by Kang and Park showed that actinomycin D binds to oncogenic promoter G-quadruplex DNA repressing gene expression.12  Formally, actinomycins can be considered as two separate depsipeptides (the macrocycles) linked via a phenoxazine nucleus with differing amino acids in the depsipeptides, depending upon the actinomycin variant in question. For a much fuller description of the history of actinomycins, the significant chemical synthetic and semi-synthetic programs since its introduction and their activities, the 2012 review by Mauger and Lackner should be consulted.9 

Although not the first antibiotic to go into general clinical use, erythromycin (Figure 1.1, 2), which was introduced in 1952, was certainly the first of the bioactive macrolidic agents to become an anti-infective drug. Even today, 60 years later, variations, usually with a change in the salt form, are still in clinical trials, and the parent molecule has been almost a “poster child” for what could be done to follow the biosynthesis, initially by using radio-labeled production of the base aglycone, the macrolide erythronolide B13  and then the work over many years by Abbott scientists and their successors in the production of variations of the base macrolide. These derivatives were prepared by biosynthetic manipulations as recombinant DNA technologies advanced, coupled to fundamental knowledge of how to manipulate gene clusters in antibiotic producing microbes. A large amount of the work using these techniques was performed by the now defunct Kosan, Inc., and modifications of the base structure using biosynthetic techniques to produce what might best be called, un-natural natural products are still being published,14  with the alkynyl substituted erythromycin (Figure 1.1, 3) being an excellent example.

Over the last 25 years, in addition to this type of biosynthetic process which has not yet led to an approved agent, a number of macrolide antibiotics have been launched based upon the erythromycin core structure. In each case, the compounds were optimized to overcome problems with the parent molecule from an antibiotic perspective. These approved agents included the first azalide azithromycin (Figure 1.1, 4) launched in 1988, a long-acting agent that is now generic in the USA. Other approved drugs with the base erythronolide structure include midecamycin (Figure 1.1, 5) launched in 1985, rokitamycin (Figure 1.1, 6) in 1986, roxithromycin (Figure 1.1, 7) in 1987, clarithromycin (Figure 1.1, 8) in 1990, flurithromycin (Figure 1.2, 9) in 1997, dirithromycin (Figure 1.2, 10) in 1993, telithromycin, the first ketolide, (Figure 1.2, 11) in 2001, and fidaxomycin (Figure 1.2, 12) in 2011. In addition to these, solithromycin (Figure 1.2, 13) is now in Phase III clinical trials for the treatment of Community-acquired Bacterial Pneumonia (CaBP).

Figure 1.2

Structures 9 to 13; Later Erythromycin-based Derivatives.

Figure 1.2

Structures 9 to 13; Later Erythromycin-based Derivatives.

Close modal

In the 1960s and later, synergistic antibiotics that were mixtures of two macrolides, known collectively as streptogramins, were in use as agents to alter food uptake and metabolism, predominantly in ruminants. Following the advent of Gram positive resistant organisms, in particular methicillin resistant Staphylococcus aureus (MRSA), and the lack of suitable treatments for these and other more resistant Gram positive microbes in humans, these old compounds were used as templates for chemical modification, leading to the approval and subsequent launch in 1999 of the quinuprisitin (Figure 1.3, 14)/dalfoprisitin (Figure 1.3, 15) 1:1 mole ratio defined mixture under the name Synercid® in the USA. At this moment, another similar mixture of related compounds is in Phase II clinical trials with Novexel, using linopristine (Figure 1.3, 16) and flopristine (Figure 1.3, 17) as the components of the mixture.

Figure 1.3

Structures 14 to 22; Synergistic Antibiotic Pairs & Rifamycin-based Molecules.

Figure 1.3

Structures 14 to 22; Synergistic Antibiotic Pairs & Rifamycin-based Molecules.

Close modal

The base structure of the rifamycin class of macrocyclic antibiotics can be best thought of as a single or fused ring (usually two) linked within a larger macrolidic ring (Figure 1.3, 18). The base molecule in this series was launched in the middle 1960s as an antimycobacterial (tuberculosis) agent, and in the intervening five decades, well over 300 variations on the structure have been reported as being in biological assessments ranging from in vitro testing, through clinical trials, to becoming approved drugs. A search of the Thomson–Reuters Integrity™ database in December 2013 showed 177 different compounds listed of similar structure, with 7 being shown as launched.

A relatively recent paper by Mariani and Maffioli15  gives an excellent comparison of the various rifamycins and also leads into discussions of other ansamycin molecules including the geldanamycins and ansamitocins, both of which will be mentioned later.

In addition to rifamycin, four other variations have been marketed since that approval: rifampicin in 1967 (Figure 1.3, 19), rifamixin in 1988 (Figure 1.3, 20), rifabutin (Figure 1.3, 21) in 1992 and rifapentine (Figure 1.3, 22) in 1998. As of December, 2013, there appear to be no rifamycin-like molecules in clinical trials for mycobacterial infections, though a recent publication in the infectious disease literature does imply that increased doses of these agents in conjunction with other anti-tuberculosis drugs are still viable treatments.16 

Of additional interest is the report in August 2012, that three rifamycins, rifampicin, rifamixin and rifabutin, have been found to be effective in preventing the growth and cellular respiration of multidrug resistant (MDR) Acinetobacter baumanii (MDRAb), which is an important pathogen associated with wound infections afflicting US military personnel.17 

Although not an erythromycin-based molecule, nor an ansamycin of the normal basic structure for these agents due to the linkages of the macrolide ring, a very recent paper from the Fenical group at the Scripps Institute of Oceanography reported the isolation and identification of the 14-ring macrolide known as anthracimycin (Figure 1.4, 23) from a streptomycete isolated from marine sediments.18  The name chosen was due to its activity against both Bacillus anthracis and methicillin-resistant Staphylococcus aureus and according to the authors only one other structure similar to this has been reported to date and that was from a myxobacterium in 2008.18  Thus, even today novel bioactive macrocyclic agents are still being found.

Figure 1.4

Structures 23 to 35; Non-Rifamycin Ansamycin Structures.

Figure 1.4

Structures 23 to 35; Non-Rifamycin Ansamycin Structures.

Close modal

Very recently, one of the first “plant-derived” tubulin interactive compounds that can be considered a bioactive macrocycle to enter clinical trials, maytansine from the Ethiopian tree Maytenus serrata, was effectively granted a new lease of life as a slightly modified “warhead” that could be conjugated to a monoclonal antibody to provide potent and selective anti-tumor agents. From the initial determination of its structure (Figure 1.4, 24) natural product chemists wondered if the compound was microbial in origin, due to its similarity to the “ansa” antibiotics such as the rifamycins. In 1977, scientists at Takeda Chemical Industries reported the structures of the bacterial products, the ansamitocins, which very closely resembled the maytansenoids. Later work on compounds isolated from the bacterium, subsequently renamed as Actinosynnema pretiosum, demonstrated that they were in fact identical to those isolated from other plant genera. The work leading up to this determination has been well covered in a review by Kirchning et al. in 200819  and/or the chapter by Yu et al. in 2012,20  as these cover the chemistry and biosynthesis of these microbial compounds.

As mentioned above, “precursors” of maytansine from microbial sources, specifically DM-1 (Figure 1.4, 25) and DM-4 (Figure 1.4, 26), with suitable chemical linkages, have been used as warheads linked to specific monoclonal antibodies directed against tumor-linked epitopes. The utility of such antibody–drug conjugates has been discussed by Senter in 200921  and amplified the following year by both Alley22  and Caravella and Lugovskoy.23  The DM1-linked conjugates are specifically described in a 2010 review by Lambert (from Immunogen) covering these constructs and their clinical efficacies.24  It is recommended that this article be read in conjunction with the 2011 paper by Kümler et al. covering the story of trastuzumab emtansine,25  the combination of Herceptin® with a specific linkage to DM1 that is cleaved by enzymes on uptake into the tumor. The combination was approved in 2012 and launched in 2013 in the USA. Recently, a clinical update was published by Barginear et al., which provides detailed clinical reports, though it covers the period before the approval.26 

Thus, a bioactive macrocycle that had failed clinical trials in the late 1970s, due to limited efficacy and substantial toxicity, has now become a potent and active treatment for specific breast cancers by repurposing it as a cytotoxic warhead on a targeted therapeutic.

This particular ansamycin (Figure 1.4, 27) has quite a chequered past. It went into clinical trials in Europe as a tubulin interactive agent27  but was discontinued as a result of lack of activity.28  It languished for many years until a report in 2005 identified the producing organism as an endophytic microbe, not the host fungus presumed to be the producer.29,30  The story of the genetic dissection and the biosynthesis of the rhizoxin complex, plus a full analysis of the symbiotic bacterium was recently published by the Hertweck group who were responsible for the whole discovery of this unique biological interaction.31  A total synthesis of another one of the rhizoxin analogues that also has bioactivity, WF-1360F (Figure 1.4, 28), has recently been reported.32 

In this section, apart from a very short historical introduction, the use of genetic modifications to biosynthetic pathways to biochemically produce and identify macrocyclic molecules that have Hsp90 binding activity, giving structures that have not yet been approached synthetically, will be presented. A much fuller story of the discoveries that led to the clinical development of a number of macrocyclic compounds as Hsp90 inhibitors using the basic geldanamycin skeleton will be presented in Chapter 2. It should be noted that there are also a wide variety of non-macrocyclic synthetic Hsp90 inhibitors, such as those from the Chiosis group at Memorial Sloan Kettering, which have been identified through intensive drug discovery programs, some of which have entered clinical trials for oncology.33 

The benzoquinone ansamycin antibiotic geldanamycin (Figure 1.4, 29) from Streptomyces hygroscopicus var geldanus was first reported by The Upjohn Company in 197034  and shown to have anti-parasitic activity. Subsequent studies demonstrated anti-tumor activity that was thought to be due to inhibition of the tyrosine specific kinase, v-Src, involved in regulating growth and cell proliferation as well as several signal transduction pathways.35,36  In 1994 however, the compound was shown to bind to the ATPase heat shock protein 90 (Hsp90) by Whitesell and coworkers.37  Then, three years later, Stebbins et al. reported that geldanamycin specifically bound to an ATP binding site at the N-terminus of Hsp90, altered its chaperone activity and indirectly led to cell death.38  The history, predominantly from an NCI perspective of the various modifications of geldanamycin leading up to the initial development of tanespimycin (17-AAG; Figure 1.4, 30), was described by Snader in 200539  with an interim report in 2005 on other details of the biological activity by Kingston and Newman,40  and then further updated in 2012 by Snader.41 

Using the knowledge derived from total genome sequences and the ability to “mix and match” genes within biosynthetic clusters and, in certain cases, to be able to add exogenous genes, potentially active structures produced from such efforts have been disclosed in the last few years. Thus analogues, such as thiazinogeldanamycin (Figure 1.4, 31) and 19-hydroxy-4,5-dihydro-geldanamycin (Figure 1.4, 32), have been reported from engineered strains of S. hygroscopicus JCM4427 together with other known derivatives,42  though their bioactivities have not been fully delineated.

What is also of interest is that a close relative of geldanamycin, macbecin (Figure 1.4, 33) was reported to be an Hsp90 inhibitor by researchers from Biotica in 2008. They demonstrated that this compound had both in vitro and in vivo activity in mice, and was more water soluble and less toxic than the geldanamycin derivatives in current clinical trials.43  In a later paper the same year,44  they reported the optimization and production of macbecin-based molecules that were derived by genetic modification of the macbecin biosynthetic complex in an Actinosynnema pretiosum subsp. pretiosum, nominally the same genus and species from which the ansamitocins were first isolated.

Since a number of reports attributed the “off-target” bioactivities of the base geldanamycin structure to the quinone moiety undergoing redox cycling, the Biotica team chose a molecule that had no quinonoid ring, as it was replaced by a phenol (Figure 1.4, 34). The production of this compound was then further optimized to >200 mg L−1 by genetic manipulation in the same microbe. Further investigation demonstrated that it had activity in vitro and in vivo similar to that shown by 17-AAG, but was a tighter binder to Hsp90 and active at a lower molar dose in both cellular and murine assays.44  Somewhat similar modifications, but this time using the geldanamycin producer S. hygroscopicus, were reported in 2011 by Wu et al., producing a molecule (Figure 1.4, 35) similar to that optimized by the Biotica group, effectively differing only in two substituents from the phenol-containing macbecin analogue.45  Thus, different modified ansamycin macrocycles with HSP90 activity are available for future screening.

The isolation of this 20-membered class of macrolidic cytotoxins from the fouling invertebrate Bugula neritina over 30 years ago has led to massive collections of the nominal producing organism and to very elegant syntheses of various components.

The initial discovery of bryostatin 3 (Figure 1.5, 38) was indirectly reported in 1970.46  Subsequent developments leading to the report of the isolation and X-ray structure of bryostatin 1 (Figure 1.5, 36) in 1982,47  and the multi-year program that culminated in the isolation and purification of 18 bryostatins (Figure 1.5, 36–53), have been well documented.48–54 

Figure 1.5

Structures 36 to 55; Bryostatins.

Figure 1.5

Structures 36 to 55; Bryostatins.

Close modal

All of the known bryostatins possess a 20-membered macrolactone ring with three remotely substituted pyran rings linked by a methylene bridge and an (E)-disubstituted alkene; all have geminal dimethyls at C8 and C18, and a four carbon sidechain (carbons 4–1) from the A ring to the lactone oxygen, with another four carbon chain (carbons 24–27) on the other side of the lactone oxygen to the C ring. Most have an exocyclic methyl enoate in their B and C rings, though bryostatin 3 (Figure 1.5, 38), in particular, has a butenolide rather than the C-ring methyl enoate, and bryostatins 16 and 17 (Figure 1.5, 51 and 52) have glycals in place of the regular C19 and C20 hydroxyl moieties.54,55 

Work reported from the Peoples’ Republic of China in 199856  and in 200457  gave the structure for bryostatin 19 (Figure 1.5, 54), purified from a South China Sea collection of Bugula neritina. Then, in the same year, this report was followed by the publication by Lopanik et al. reporting the isolation of bryostatin 20 (Figure 1.5, 55) from an Atlantic-sourced Bugula neritina.58  Comparison with the structures of the other 18 bryostatins shows that these are closely related to bryostatin 3 in terms of their basic ring components.

Bryostatin 1 has been through well over 80 Phase I and Phase II clinical trials, with or without the addition of a cytotoxic agent in the protocol. One Phase I trial using bryostatin and temsirolimus was “actively recruiting” as of the beginning of July 2013 (NCT00112476) in the ClinicalTrials.gov web site, but in December, 2013, the status had moved to completed, but without any details published as of that date. One more trial, a potentially interesting Phase II trial investigating its use in Alzheimer's disease, is listed as “unknown status”, with no information posted since 2008 (NCT00606164). The major dose-limiting toxicity for bryostatin appears to be very significant myalgia in patients. Whether this side effect can be ameliorated by alteration of dosing regimens is not clear.

Some excellent chemistry groups have synthesized a number of the bryostatins (regular structures are shown in Figure 1.5), and have devised simplified molecules based upon the basic skeleton that have much higher (orders of magnitude in some cases when compared to bryostatin 1) in vitro activities and can be produced by total synthesis (see discussion later in this section). An early synthetic example was the discussion of routes to bryostatin 1 by Masamune in 1988,59  but it should be pointed out that with the exception of the Trost synthesis referred to later, none of these syntheses was a substitute for the isolation and purification of bryostatin 1 from natural sources. However, the first total synthesis of any bryostatin, was the enantiomeric synthesis of bryostatin 7 in 1990 by Masamune's group,60  which was followed in 1998–99 by details of an enantiomeric synthesis of bryostatin 2 from the Evans’ group.61,62  The synthesis of bryostatin 3 was reported by Nishiyama and Yamamura in 200063  together with a fuller explanation of the strategies utilized by Ohmori in 2004.64  These earlier syntheses along with the reported partial syntheses of other bryostatins were reviewed in detail through 2002 by Hale et al.54  The total synthesis of bryostatin 7 using methodologies that would allow modifications to the base structure to be performed was reported in 2006 by the Hale group.65  Two years later, in 2008, Trost and Dong published their elegant synthesis of bryostatin 1666  involving some novel metal-linked catalysis steps67  that included a ruthenium tandem alkyne-enone coupling, and then a palladium catalysed alkyne-ynoate macrocyclization to give the cyclized precursor of bryostatin 16.

A truly excellent compendium and thorough discussion of the chemistry efforts around the synthesis of the bryostatins was published in 2010 by Hale and Manaviazar covering the published results up to then.55  It should be read by any chemists interested in the manifold methods and specific methodological differences that can be, and have been, used in both successful and unsuccessful syntheses of these agents. However, in spite of all of these methods, up to 2011, no de novo synthesis of bryostatin 1 had been published. Then, in 2011, the Keck group reported the first complete total synthesis of this agent.68  This report was rapidly followed by a paper from Manaviazar and Hale with details of a shorter route69  to the same compound. Later the same year, Trost et al. published on ring expanded versions of bryostatins obtained by total synthesis,70,71  so the synthetic story of this class of macrocycles has not yet finished.

A number of simplified bryostatin analogues (often called “bryologs”) have been synthesized using methods such as function-oriented synthesis. This technique was employed by Wender and other workers to develop simplified analogues with comparable or much improved activities, in some cases orders of magnitude in in vitro assays.72–75  Further information was given in a review by Newman76  and in a recent 2013 paper by the Keck group.77  What is also of significant import is the recent report by the Wender group of in vitro anti-HIV activity for some of their newer analogues. It will be interesting to see if these can be further developed to provide in vivo activity.78 

Thus, just as in the case of the erythromycin basic structure, the synthesis and biological activities of macrocyclic compounds first reported almost 45 years ago, are still being investigated, with perhaps more interesting biological discoveries to come, particularly as it is now almost certain that bryostatins are produced by an as yet uncultured microbe found initially in the larvae of the bryozoan. Current information can be obtained by inspection of the review by the Haygood group published in 2010.79 

The epothilones are bioactive macrocycles that have led to very significant numbers of analogues being made by a variety of methods, including what was possibly the first example of genetic manipulation in the Myxobacteriales to optimize production of a desired molecule. A thorough review of the drug discovery effort surrounding this family of macrocycles and their subsequent clinical development can be found in Chapter 3 of this volume.

The initial identification by Reichenbach and Höfle of the 16-membered macrolides epothilones A and B (Figure 1.6, 56 and 57) from Sorangium cellulosum So ce90, is covered in detail in reviews by Reichenbach and Höfle.80–82  These discoveries, coupled to the report of their activity as tubulin stabilizers in 1995 by workers at Merck,83  led to work in the USA and Germany on the production of these molecules and other variations from bioengineered organisms.

Figure 1.6

Structures 56 to 59; Epothilones and Rapamycin-based Macrolides.

Figure 1.6

Structures 56 to 59; Epothilones and Rapamycin-based Macrolides.

Close modal

In 2000, workers at Kosan reported the isolation of the epothilone producing gene cluster from Sporangium cellosum strain SMP44 and expressed it in Streptomyces coelicolor CH999 together with expression of EpoK, the cytochrome P450 that epoxidizes epothilones C and D to A and B, respectively, in E. coli.84  Contemporaneously, the original group at the GBH in Braunschweig also published their epothilone biosynthetic cluster from S. cellulosum So ce90.85  Two years later, in 2002, Julien and Shah from Kosan demonstrated that the comparatively low yields seen in the S. coelicolor construct were materially improved by use of Myxococcus xanthus TA as a host, deleting EpoK and then using a variety of fermentation “tricks” including the well-known industrial technique of adding an adsorbent resin to the fermentations.86  These and later reports on further fermentation optimization from the Kosan group should be consulted for specific information.87–90  In addition, the work from the GBH concentrating on Sorangium species as a model organism should also be consulted from the original report from Gerth et al. in 200391  to the excellent review by Wenzel and Müller in 2009 on the impact of genomics on myxobacterial metabolomes.92 

Rapamycin and its close chemical relatives can almost be called “a molecule for most diseases” since the rapamycins now cover molecules that have biological properties ranging from initial anti-fungal activities through immunomodulation to anti-tumor therapies, and even as a molecule to use in stents to avoid plaque formation in blood circulation.

In 1975, scientists at Ayerst laboratories reported the 31-membered macrocyclic antibiotic rapamycin (sirolimus; Figure 1.6, 58a) to be a potential anti-fungal agent that was produced by the fermentation of a strain of Streptomyces hygroscopicus isolated from soil samples in Rapa Nui (Easter Island).93–95  Rapamycin was unsuccessful as an anti-fungal agent due to its immunosuppressant effects, however its activity against syngeneic murine tumors was reported a few years later in 1984 by Sehgal and co-workers at Ayerst Research Laboratories.96  At this time, the initial anti-tumor activity of rapamycin was not further developed, but as will be shown below, this parent structure (Figure 1.6, 58a) has since led to the production of several molecules with a variety of different pharmacological activities including, as mentioned above, anti-tumor activity. In the early 1990s, the molecular target of rapamycin in yeast was identified as TOR or “target of rapamycin”,97  followed three years later by the identification of the mammalian homolog, mTOR,98  with these reports ultimately leading to the development of a wide variety of anti-cancer and other pharmacologic agents.

Initially, chemical modifications were made at the carbon atom at C43  on the rapamycin parent structure with numeration as in Zech et al.99  rather than the alternative numbering system of McAlpine et al.,100  which was based upon a comparison with FK506 (which would give a C40  substitution), ultimately leading to a total of four clinically approved drugs, rapamycin (sirolimus), everolimus, temsirolimus, and zotarolimus. In 1999, sirolimus (rapamycin) (Figure 1.6, 58a) was approved as an immunosuppressive agent and is currently in Phase I/II trials for the treatment of various cancers. In a similar manner, everolimus (Figure 1.6, 58b), was initially launched in 2004 as an immunosuppressive agent and it then was approved for the treatment of kidney, brain, pancreatic, and breast cancers in 2009, 2010, 2011, and 2012, respectively. Then, to add to the armamentarium of just this compound, in 2012 everolimus was released by Abbott to be used as a coating for stents in the treatment of coronary and peripheral arterial diseases. It is also currently in Phase III trials for treating diffuse large B-cell lymphoma (NCT00790036), liver (NCT01035229), and stomach (NCT00879333) cancers. Temsirolimus (CCI-779, Figure 1.6, 58c) was first approved as a treatment for renal carcinoma in 2007, then in 2010 was approved in Japan and, as with its close chemical relatives, is currently in Phase II trials for the treatment of various carcinomas, mainly under the support of the NCI. Zotarolimus (Figure 1.6, 58d) was launched in the USA in 2005 for the treatment of arterial restenosis, as a component of a stent, and recently the EU approved a stent containing novolimus, a metabolite of rapamycin that has a C7-hydroxy group in place of the methoxyl in the parent molecule (cfFigure 1.6, 58a).

Merck and Ariad Pharmaceuticals collaborated to develop another rapamycin derivative, ridaforolimus (AP-23573, Figure 1.6, 58e), which is in Phase III clinical trials for the treatment of soft tissue carcinoma (NCT00538239) and bone cancer (NCT00538239). In contrast, from a chemical perspective, Wyeth Pharmaceuticals developed a rather interesting derivative of rapamycin, ILS-920, with a modified ring structure (Figure 1.6, 59). By modification of the triene portion of the molecule mTOR binding would be disrupted. However, ILS-920 may have a different target as it is a non-immunosuppressive neurotrophic analogue reported to exhibit a binding affinity over 900-fold higher for FKBP52 than FKBP12. This analogue promotes neuronal survival and outgrowth in vitro, and binds to the β1 subunit of L-type calcium channels (CACNB1).101  ILS-920 was under development for the treatment of stroke102  with a Phase I clinical trial for the treatment of acute ischemic stroke completed (NCT00827190). Interestingly, FKBP52 inhibition is reported to affect tubulin interactions in cells103  and this activity was exploited to screen natural products that inhibit the formation of a complex between FKBP52 and androgen receptors. Since this interaction may play a role in the progression of prostate cancer,104  there is a possibility that ILS-920 may exhibit anti-tumor activity, although no reports of such activity have been published as yet. In addition to reports showing effects in the brain, there was a report in 2008 that demonstrated that rapamycin plus lithium may aid in the treatment of Huntingdon's disease, so this structure may advance into areas not envisioned in earlier days.105 

Two prodrugs of rapamycin, Abraxis’ ABI-009, a nanoparticle encapsulated formulation of rapamycin, and Isotechnika's TAFA-93 (structure not yet published), have also advanced into Phase I clinical trials. The structure of the latter molecule likely includes either the rapamycin core structure or may only have modifications at the C43  hydroxyl group, avoiding both the FKBP-12 and TOR binding sites, as modifications anywhere else are thought to negate the intrinsic biological activity of these derivatives.106,107 

Another area of the rapamycin story has to do with the use of bacterial “genetic engineering”. In these programs, rapamycin derivatives are produced that are composed of biosynthetic gene clusters that have been modified, or expressed in unusual environments, with the aim of producing analogues with different structures and perhaps also different biological activities. Following on from the pioneering work of the Demain group at MIT,108,109  one of the groups that spent many years studying the biosynthesis of rapamycin, and then developed methods of utilizing the information in order to produce novel compounds was the group led by Leadlay and Staunton at Cambridge University, which spawned the formation of the biotech company, Biotica.110 

The methodologies and compounds developed since the 2004 review by Demain109  have been presented by a number of authors in the last few years, commencing with a review by Graziani covering 2003 to 2008,111  which should be read in conjunction with the 2010 review by Park et al.112  These papers demonstrate the multiplicity of materials that can be produced by modification of biosynthetic units, and show the problems involved in the regulation of any biosynthetic process. Among the problems addressed is what is now realized to be the “Achilles Heel” of mutasynthetic processes designed to increase yields of desired molecules, namely the provision in the microbe used of sufficient precursors so as to be able to maintain growth and also increase production of the desired molecules. An excellent example of this, though not from rapamycin studies, is shown in the recent paper from Keasling's group discussing the production of artemisinic acid from a bio-engineered yeast strain as a precursor for the semi-synthesis of artemisinin.113 

The 2013 publications and patent applications from the Biotica group demonstrate the potential of these processes in producing engineered strains for producing novel rapalogs114  and for utilizing the products for biosynthetic medicinal chemistry.115–117  Using a combination of biosynthetic engineering and molecular modeling, multiple structural modifications were made to the rapamycin (sirolimus) scaffold (Figure 1.6, 58a) to generate a series of rapalogs having various combinations of changes, including C6-demethylation, C7-O-demethylation, C14-carbonyl reduction to CH2, C29-O-demethylation or reduction to CH2, and C44-O-demethylation or reduction to CH2.116  The rapalog lacking the C6-methyl and 7-O methyl groups showed significantly enhanced cell line growth inhibition (mean IC70 = 7.4 µM) in a 37 cancer cell line screen, compared to rapamycin (mean IC70 = 16.3 µM).116 

These agents can be either cyclic peptides with “regular” amino acids comprising their backbone, or molecules with what, at times, were thought to be non-proteinogenic (non-ribosomal) amino acid backbones. The discoveries in the last few years, particularly by the groups of Ireland and Schmidt, (ref. 118 and publications cited therein), plus the recent review by Dunbar and Mitchell119  have now demonstrated that a number of these (discussed later) are ribosomally encoded but then modified by other processing enzymes.

Although penicillins V and G are often touted in the West as being the first antibiotics to go into general use, in 1944 Gause and Brazhnikova reported on a variation of gramicidin, named as gramicidin S (Soviet).120  Rather than being a linear peptide as were the earlier gramicidins A–C,121,122  which are 15-mer peptides with alternating D and L amino acids, gramicidin S was subsequently shown to be a macrocycle (Figure 1.7, 60). This initial report was followed the same year by a clinical paper showing efficacy in over 1500 patients in what was then the USSR.123  Thus, this membrane-active cyclic peptide can also lay claim to being one of the first antibiotics to go into general use, though due to its “country of origin” the discovery may not have been given the credit that it deserved, though the publications were in two UK journals during the height of World War II. However, just to demonstrate that even today, almost 70 years after the initial report of its activity, variations are still being investigated. Thus the paper by Kapoerchan et al. in 2012 shows what can occur in terms of increased bioactivity when “inverted” analogues of the base molecule are produced.124  For example, the analogue (Figure 1.7, 61) shows increased antibacterial and reduced hemolytic activity compared to gramicidin S.

Figure 1.7

Structures 60 to 65; Mono- and Tri-cyclic Peptide-based Macrocycles.

Figure 1.7

Structures 60 to 65; Mono- and Tri-cyclic Peptide-based Macrocycles.

Close modal

Ziconotide (Figure 1.7, 62), a peptidic cone snail toxin, has the distinction of being the first “direct from the sea” agent to be approved for any disease. In late December of 2004, the FDA approved this peptidic compound for treatment of intractable neuropathic pain, with “phantom limb pain” being a particular indication for this drug. Since the method of drug delivery is via an intra-thecal syringe from a reservoir in the peritoneum of the patient, the number of patients willing to tolerate the intricate delivery system is low, but there are studies in the literature demonstrating its value in very specific situations.125 

This compound is the exact equivalent of the 25-residue peptide isolated from the venom of the cone snail Conus magus under the name of ω-conopeptide MVIIA.126  Over 200 variations on the structure were made before the realization that the native peptide was the most effective. The Irish company Elan purchased the compound and rights and it was approved as stated earlier. In 2012 Olivera and collaborators published an excellent review in Biochemistry (Moscow)127  covering these peptidic cone snail agents, which should be read in conjunction with a review a year earlier by Daly and Craik from the University of Queensland.128 

This 2011 paper suggested that it might well be feasible to modify some of these toxins by linking their N and C termini together to obtain orally active molecules. In 2012 evidence for this methodology was published by the Queensland group, paving the way for agents in the future that may be orally dosed.129,130  Thus, the cyclic analogue of α-conotoxin MII, cMII-6 (Figure 1.7, 63), was synthesized and shown to have greatly enhanced proteolytic stability in human serum while retaining most of the activity of the parent peptide against the nicotinic acetylcholine receptor nAChRα6, a target of potential importance for treatment of Parkinson's disease.129  In addition, Craik et al. have reported the development of an orally active cyclic analogue of α-conotoxin Vc1.1 from the cone snail Conus victoriae, cVc1.1 (Figure 1.7, 64), via cyclization of the peptide backbone.131  The cyclic analogue was shown to be more stable in simulated intestinal fluid and human serum compared to linear Vc1.1, and it showed greater selectivity in the inhibition of calcium channels versus the α9α10nAChR. Testing in the rat CCI-model of neuropathic pain demonstrated that cVc1.1 produced analgesia when delivered orally, and was 120 times more potent than gabapentin, currently used for the treatment for neuropathic pain.131 

The patellamide cyclic peptides are part of a large group of peptidic cytotoxins that were originally thought to be the products of non-ribosomal peptide synthetases. However, in a recent review118  Schmidt et al., utilizing new generation sequencing of the as yet uncultured cyanophyte Prochloron isolated from the tunicate, have demonstrated that these cyclic peptides, an example being Patellamide D (Figure 1.7, 65), are the products of ribosomal peptide synthesis followed by “tailoring” to produce the oxazoles and thiazoles in their structures. These are formed by subsequent cyclization of the regular amino acids coded for in the initial sequences. The details of their discovery and the potential applications are given in detail in the review referenced above.

The initial compounds, both from natural sources and synthesis were relatively simple structures consisting of a zinc-binding warhead (a hydroxamate or epoxide), a 4-carbon linker (mimicking the ɛ-amino side chain of lysine) and an aromatic “end”. However, within the original natural product structures was a tetrapeptide known as trapoxin A (Figure 1.8, 66), first reported by Itazaki et al. in 1990,132  with recognition of the mechanism of action, irreversible inhibition of deacetylation of acetylated histone molecules, being published by Kijima et al. three years later.133 

Figure 1.8

Structures 66 to 79; Cyclic Histone Deacetylase Inhibitors.

Figure 1.8

Structures 66 to 79; Cyclic Histone Deacetylase Inhibitors.

Close modal

Following these identifications/MOAs, both new and old tetrapeptides were reported to demonstrate similar activities, with apicidins (Figure 1.8, 68 and 69) being reported by the Merck group in 1996134,135  and the much older tetrapeptides, chalmydocin and microsporin A also being identified as HDAC inhibitors.136 

However, in 2006, a series of unusual tetrapeptides with three unnatural α-aminoacids were reported from the sponge Mycale izuensis and named as azumamides A–E.137  Two of these compounds (azumamides A and E; Figure 1.8, 69 and 70) were then synthesized together with a close analogue (Figure 1.8, 71) that exhibited higher potency as an HDAC inhibitor,138  a year after the report of the total synthesis of the A and E compounds by Izzo et al.139  Very recently, the Olsen group at the Technical University of Denmark reported on their synthetic work on the azumamide skeleton, which included the first synthesis of the B to D variants, confirming the original structural assignments, together with a full isotype profiling of this class of agents and identifying that the C and D analogues are potent and specific inhibitors of HDAC10 and 11.140 

Contemporaneously with the determination of the MOA of trapoxin, a novel cytotoxic microbial agent was reported by Ueda et al. in 1994 from the phytotoxic bacterium Chromobacterium violaceum,141  and its MOA was determined in 1998 by comparison to the activities of trichostatin A.142  This compound, known nowadays by a variety of names, FK-228, depsipeptide, romidepsin or Istodax® (Figure 1.8, 72), can be thought of as the prototype structure for the cyclic agents (prodrugs) that contain a dithio bridge that when opened, will complex the zinc atom in the active site of the HDAC enzyme.143  FK228 was recommended for FDA approval in September of 2009 as a treatment for cutaneous T-cell lymphoma and was launched in the USA in March, 2010.

However, FK228 was only the beginning of the isolation, identification and often, total synthesis of a number of depsipeptides all containing the dithio bridge and all acting as prodrugs for interaction with various isotypes of HDACs. In 2001, Masuoka et al. reported on the spiruchostatins A and B (Figure 1.8, 73 and 74) and in 2003, Chen et al.144  reported the synthesis of FR-901375 (Figure 1.8, 75) a molecule closely related to FK228 that had only been reported in a Japanese patent in 1991. Spiruchostatin A was synthesized by the Ganesan group with a report in early 2004145  and later, using the material synthesized by Ganesan, a group at the University of Southampton determined that this agent was a potent inhibitor of the Class 1 HDAC isotype.146  Very recently a Japanese group reported on the synergistic activity of spiruchostatin A (under its code number, OBP-801) with the PI3K inhibitor LY294002 in an in vitro assay of human renal carcinoma cell lines; thus this agent may have potential in a tumor system that is resistant to most current therapies.147 

In addition to the agents discussed above, a third set of microbial products have also been identified that have similar structures and activities. These are the thailandepsins, which were originally reported as a result of genomic analysis of a specific bacterium and whose original structures have undergone revision following total synthesis,148  demonstrating that the compound known as burkholdac B (Figure 1.8, 76) is identical to the revised structure for thailandepsin A and has an activity against the MCF7 breast cancer cell line with an IC50 value of 60 pM. A recent paper from Wilson et al., demonstrated activity against ovarian cell lines in the nanomolar range with thailandepsin A, thus extending the range of activities for this class of compounds.149 

These is another microbial product, this time however, from a marine cyanophyte,150  that also has a masked thiol group. However, rather than a dithio linkage, in this compound (largazole; Figure 1.8, 77), the thiol is masked as a thioester and requires hydrolysis to release the activated thiol for interaction with HDAC isotypes.151  Following the report of the structure and its synthesis, a number of other groups have synthesized not only the base structure, but modifications thereon, including Ghosh and Kulkarni,152  the Phillips group,153  Luesch with synthetic modifications,154  and Schreiber and Williams155  all within 4 months of the presentation of the structure at an ACS Meeting, thus proving that if there is a novel and active agent from Nature, synthetic chemists are very willing to synthesize it and produce analogues.

Since that time, a number of other groups have extended the syntheses and activities, and these reports should be consulted by readers interested in modification of natural product structures.156–159  In addition, expanded biological activities have also been reported for largazole and derivatives, including osteogenic activity,160  sensitization of lymphoma cells to nucleosides,161  proteosomal degradation,162  and inhibition of liver fibrosis.163 

Finally for this section there have been some very interesting reviews published in the last few years covering both syntheses from natural products or using combinatorial chemistry to provide different base structural molecules. Thus in 2009, Olsen and Ghadiri demonstrated how focused combinatorial libraries of cyclic α3β-tetrapeptides led to the identification of potent HDAC inhibitors, working from natural product scaffolds with modifications.164  This work was followed in 2010 by the Hanessian group's demonstration how non-natural macrocyclic inhibitors of HDACs could be synthesized.165  This later work gave synthetic molecules that demonstrated selectivity for the Class 6 HDAC isotype (Figure 1.8, 78 and 79) depending upon their configuration at a given center. Then in 2013, there were three reviews published with one covering case studies on the synthesis of bioactive cyclodepsipeptide natural products by Stolze and Kaiser166  and two that covered more of the biological aspects of these molecules.167,168 

One of the simplest macrolides, mycolactone A/B, is a 12 membered ring with a pendant unsaturated polyhydroxy fatty acid ester and an upper unsaturated alkyl side chain (Figure 1.9, 80) and it is one of the nastiest bioactive macrolides known from a physiological effect aspect.169  This is the compound that causes the chronic necrotizing skin ulcer known as “Buruli Ulcer”. The causative agent is a component of various strains of the mycobacterium, Mycobacterium ulcerans, and until the recent total synthesis by Altmann's group at the ETH, studies were limited by the lack of pure compound(s) to work with.170 

Figure 1.9

Structures 80 to 82; Mycolactones A/B, Halichondrin and Eribulin.

Figure 1.9

Structures 80 to 82; Mycolactones A/B, Halichondrin and Eribulin.

Close modal

Mycolactone A/B (both E and Z isomers coexist in the purified microbial isolate) was difficult to obtain in pure form and therefore trying to determine its mechanism of action and thus potential treatments was very difficult. Soon after the publication by Altmann et al., a group in France reported a diverted total synthesis of mycolactone A/B and analogues, thus providing the potential to expand the SAR, though they found that the natural product was still the most toxic agent.171 

In 2013, Altmann, working with a group of Swiss scientists including immunologists and tropical disease specialists, using mycolactones with differing lower side chains (the fatty acid esters) together with some changes on the upper side chain, have shown that the lactones have no antibiotic activities but are apoptotic, necrotic, and immunosuppressive in their interactions with mammalian cell types.172  A derivative in which the lower side chain of mycolactone A/B was truncated to an acetyl residue was unique in strongly inhibiting the metabolism and cell proliferation of a murine L929 fibroblast cell line, while being non-cytotoxic at the highest concentrations used. Thus, through the use of total synthesis and techniques that permit “relatively” simple variations to the side chains, a series of these very toxic compounds can be obtained for future use by biologists and physicians investigating this very unpleasant disease.

Due to its extraordinary anti-tumor activity, the antitubulin marine natural product halichondrin B (Figure 1.9, 81) was chosen for preclinical development in 1992 by the Developmental Therapeutics Program (DTP) at the NCI. However, clinical development was severely impeded due to the limited amounts of compound available from natural sources.

Kishi's group reported the total synthesis for halichondrin B in 1992.173  In collaboration with scientists at the Eisai Research Institute (ERI) in Woburn, MA, Kishi demonstrated that the right hand macrolide half of the molecule (approx MW of 600) retained all or most of the potency of the much larger parent compound. Chemists at the ERI, working very closely with Kishi's group, then synthesized over 200 analogues.174,175 

In conjunction with the DTP, they demonstrated that the modified truncated macrocyclic ketone, eribulin (E7389; Figure 1.9, 82), where the lactone in the macrocyclic ring was converted to a ketone and the ring pyran adjacent to the ketone was converted to a furan with a simple 3-carbon side chain, gave a molecule that had greater in vivo stability, possessed comparable bioactivity to and lower toxicity in vivo than halichondrin B (obtained by DTP in conjunction with New Zealand scientists).

Subsequently, the Eisai group has demonstrated that relatively minor changes to the “tail” of the eribulin molecule results in much lower propensity for inducing P-glycoprotein susceptibility while retaining in vivo potency.176  Incorporation of a morpholine in the “tail” demonstrated oral activity in a subcutaneous LOX melanoma model,177  and modification by ring closure at the “tail” yielded a different morpholino derivative that demonstrated intravenous in vivo activity in an orthotopic murine model of a human glioblastoma.178 

The development of Eribulin (E7389), perhaps the most complex drug molecule yet produced by total synthesis, from a marine derived antitumor agent, halichondrin B, is a compelling example of the power of the DTS approach.

Although many more examples of bioactive macrocylic compounds could be given, the molecules discussed above cover over 70 years of compound discovery, and are all based upon natural products. As was discussed in the Introduction, data reported in a 2013 review of macrocyclic drugs and clinical candidates indicated that the 68 registered macrocyclic drugs in use at the time were almost exclusively of natural origin. In the case of the 35 macrocyclic compounds in clinical trials that were considered, 25 were of natural origin, with 10 being of de novo design. Of significance was the observation that 43% of the clinical development candidates were orally administered, as compared to only 28% of the registered macrocyclic drugs. Even more noteworthy was the fact that 9 of the 10 de novo designed clinical candidates were orally bioavailable. These figures reflect a clear trend towards the increased development of orally bioavailable macrocyclic therapeutic agents, and suggest that synthetic and medicinal chemists are being guided by structural cues provided by bioactive natural macrocyclic agents in the design of synthetic mimics having the desired pharmacological properties, including oral bioavailability.

Two brief examples illustrating the utilization of macrocyclization by drug discovery programs to provide fully synthetic bioactive molecules are shown below. The first demonstrates the use of data from X-ray crystallography of the peptide KPF-pY-VNV bound to the Grb2-SH2 domain, to afford a lead series based upon a peptidomimetic structure that was pre-organized to give a β-turn configuration. This basic structure was then cyclized and slightly modified, yielding a potent in vitro active material. Further work by the same group produced a malonate derivative that is still biologically active (Figure 1.10, 83).179,180  The second example is the work of Zapf et al. on the synthesis of a series of Hsp90 inhibitors via use of the Buchwald-Hartwig cyclization reaction. Thus, starting from an acyclic clinical candidate and leveraging a variety of synthetic methodologies to explore linkers with different functionalities, coupled to biological activity analyses at each stage, a potent orally dosed in vivo active Hsp90 inhibitor was identified (Figure 1.8, 84).181 

Figure 1.10

Structures 83 to 84; Syntheses of Totally Synthetic Bioactive Macrocycles.

Figure 1.10

Structures 83 to 84; Syntheses of Totally Synthetic Bioactive Macrocycles.

Close modal

Finally, it has been the intent of this chapter to demonstrate the impressive variety of structures of bioactive natural product-based macrocycles, and also to show, albeit only with a few examples, that synthetic or quasi-synthetic macrocyclic compounds, derived from a natural product pharmacophore or close spatial relative, can provide drug-like structures on which to base the search for novel drug entities.

Thus, in the field of macrocyclic chemistry it is apparent that there can be a definite continuum from natural product structures to totally synthetic molecules that perform the same pharmacological tasks, aside from perhaps antibiotics, though that area may yet be conquered as shown by the success of the bicyclic, but not yet macrocyclic, quinolone-based antibacterial agents colloquially known as “floxacins”.182 

The area of macrocyclic chemistry thus offers an exciting and relatively new approach to effective drug discovery and development, substantially different to the focus on small molecules which, to date, has been favored by most medicinal chemistry drug discovery programs. This and subsequent chapters present promising evidence of a sustained and increasing interest in this field of drug discovery and development that augurs well for the availability of novel and more effective treatments for many of the serious diseases afflicting the global population.

1

The opinions expressed in this article are those of the authors, not necessarily those of the US Government.

Figures & Tables

Figure 1.1

Structures 1 to 8; Actinomycin D; Erythromycin and Early Derivatives.

Figure 1.1

Structures 1 to 8; Actinomycin D; Erythromycin and Early Derivatives.

Close modal
Figure 1.2

Structures 9 to 13; Later Erythromycin-based Derivatives.

Figure 1.2

Structures 9 to 13; Later Erythromycin-based Derivatives.

Close modal
Figure 1.3

Structures 14 to 22; Synergistic Antibiotic Pairs & Rifamycin-based Molecules.

Figure 1.3

Structures 14 to 22; Synergistic Antibiotic Pairs & Rifamycin-based Molecules.

Close modal
Figure 1.4

Structures 23 to 35; Non-Rifamycin Ansamycin Structures.

Figure 1.4

Structures 23 to 35; Non-Rifamycin Ansamycin Structures.

Close modal
Figure 1.5

Structures 36 to 55; Bryostatins.

Figure 1.5

Structures 36 to 55; Bryostatins.

Close modal
Figure 1.6

Structures 56 to 59; Epothilones and Rapamycin-based Macrolides.

Figure 1.6

Structures 56 to 59; Epothilones and Rapamycin-based Macrolides.

Close modal
Figure 1.7

Structures 60 to 65; Mono- and Tri-cyclic Peptide-based Macrocycles.

Figure 1.7

Structures 60 to 65; Mono- and Tri-cyclic Peptide-based Macrocycles.

Close modal
Figure 1.8

Structures 66 to 79; Cyclic Histone Deacetylase Inhibitors.

Figure 1.8

Structures 66 to 79; Cyclic Histone Deacetylase Inhibitors.

Close modal
Figure 1.9

Structures 80 to 82; Mycolactones A/B, Halichondrin and Eribulin.

Figure 1.9

Structures 80 to 82; Mycolactones A/B, Halichondrin and Eribulin.

Close modal
Figure 1.10

Structures 83 to 84; Syntheses of Totally Synthetic Bioactive Macrocycles.

Figure 1.10

Structures 83 to 84; Syntheses of Totally Synthetic Bioactive Macrocycles.

Close modal

References

Close Modal

or Create an Account

Close Modal
Close Modal