It is always a little difficult to know where to start and when to stop in time when discussing the historical influence of natural products upon drug discovery because, even today, materials that were identified as late as the 1970s (though used for many centuries as a mixture) are still influencing chemists and biological scientists to use the “native product” and/or a modification as either probes for specific targets or as a treatment in its own right. Later in the chapter, we will demonstrate how natural product structures are still valid models upon which to base 21st century drugs.

Throughout the ages, humans have relied on Nature to cater for their basic needs—not the least of which are medicines for the treatment of a wide spectrum of diseases. Plants, in particular, have formed the basis of sophisticated traditional medicine systems, with the earliest records, dating from around 2900–2600 BCE,¹  documenting the uses of approximately 1000 plant-derived substances in Mesopotamia²  and the active transportation of medicinal plants and oils around what is now known as Southwest Asia. These include oils of Cedrus species (cedar) and Cupressus sempervirens (cypress), Glycyrrhiza glabra (liquorice), Commiphora species (myrrh) and Papaver somniferum (poppy juice), all of which are still used today for the treatment of ailments ranging from coughs and colds to parasitic infections and inflammation. In addition to plants, around 120 mineral substances were also listed as “medicinal in nature” including materials now identified as arsenic sulfide, sulfur, lime, potassium permanganate and even rock salt. In most cases, the materials were delivered as infusions (teas), ointments, medicated wines, enemas and even by fumigation—methods still in use in pharmaceutical delivery systems even today. By approximately 700 BCE, the concept of “contagion” was developing, though it would be millennia before the relationship of microbes to plagues, etc., was formally established. Although what is interesting is a description of the use of “rotten grain” in treating wounds; it is tempting to speculate that this might have been a method of administering a crude antibiotic formulation to a patient.

Egyptian medicine dates from about 2900 BCE with the best known record being the “Ebers Papyrus” dating from 1500 BCE, documenting over 850 drugs, mostly of plant origin³  including opium, cannabis, linseed, aloe and garlic. At around the same time, the Chinese Materia Medica was being extensively documented, with the first record dating from about 1100 BCE (Wu Shi Er Bing Fang, containing 52 prescriptions), although records from the Pent'sao are reputed to date from ∼2700 BCE. These were followed by works such as the Shennong Herbal (∼100 BCE, 365 drugs) and the Tang Herbal (659 CE, 850 drugs).^4,5  Likewise, documentation of the Indian Ayurvedic system dates from before 1000 BCE (Charaka; Sushruta and Samhitas with 341 and 516 drugs respectively).^6,7 

The Greeks and Romans contributed substantially to the rational development of the use of herbal drugs in the ancient Western world with Hippocrates (∼460 to 377 BCE) being considered the father of medicine through an anonymous treatise known as Corpus Hippocraticum, which covered the usage of mainly plant-based mixtures but with an emphasis on the correct diet. Sources included extracts of poppy, henbane and mandrake, alongside juniper and saffron. Entertainingly, one might well argue that establishing a potential resistance to poisoning (assassination rather than happenstance) also contributed to the evolution of Greek pharmacy around 100 BCE, with the preparation of Mithridaticum, a combination of 54 ingredients, made for Mithridates, who was the King of Pontus at that time.⁸  The mixture was “improved” by Andromachus, Nero's physician to contain 70 ingredients and was still available under the name Theriac in various European pharmacopoeias until the 19th century. Dioscorides, a Greek physician (100 CE), accurately recorded the collection, storage and use of medicinal herbs during his travels with Roman armies throughout the then “known world”, publishing his famous five volume botanical work, De Materia Medica at that time; details are available⁹  from the US National Library of Medicine (NLM). The majority of his “drugs” (80%) were based on plant sources, with animal and mineral sources making up ∼10% each. Almost at the same time, Galen (130–200 CE.), a practitioner and teacher of pharmacy and medicine in Rome, was well-known for his complex prescriptions and formulae used in compounding drugs, with details given in his herbal, De Simplicibus, of 473 drug entities.

During the Dark and Middle Ages (5th to 12th centuries), the Arabs (covering Southwest and Central Asia) preserved much of the Greco-Roman expertise and expanded it to include the use of their own resources, together with Chinese and Indian herbs unknown to the Greco-Roman world. Thus Rhazes, a Persian physician in the early 900s, gave the first accurate descriptions of measles and smallpox and Avicenna, an Arab physician of the late 900–early 1000 era, codified the then current knowledge with his epic and encyclopaedic work, Canon Medicinae, a book that influenced the practice of medicine for the next 600 plus years.¹⁰  This was subsequently superseded by the work of Ibn al-Baitar (whose full name was Abu Muhammad Abdallah Ibn Ahmad Ibn al-Baitar Dhiya al-Din al-Malaqi) an Arab, born in Malaga towards the end of the 12th Century (died 1248 CE), but who travelled extensively over the Muslim world and produced two extremely well known treatises, one on botany, that described over 1400 plants, over 200 of which had never previously been recorded (Kitab al-Jami fi al-Adwiya al-Mufrada) and the other, a comprehensive compilation known as the Corpus of Simples in English (Kitab al-Mlughni fi al-Adwiya al-Mufrada in Arabic). Both books were translated into Western languages in later centuries. For those readers wanting to further investigate the influence of the Arabic schools, details can be found at the NLM.

From a Western perspective, following on from the ∼1500 CE time frame, the person whose ideas permeated the West for the next two hundred or so years was Paracelsus, whose real name was Theophrastus Phillipus Aureolus Bombastus von Hohenheim, born in Switzerland in 1493. He attempted to “modernise” the then existing works of his forerunners by perhaps the first use of alchemy to separate “good from bad” effects of treatments. Although he was probably responsible for the derivation of what later became known as the “doctrine of signatures”, he did place pharmacy on a relatively sound chemical footing and may best be known for the use of mercury as a treatment for syphilis and for the value of mineral waters, plus substituting more simple herbal remedies for some of the complex mixtures handed down from the time of Galen. However, pharmacy was still an empiric science, as shown by publication of the herbal The London Pharmacopoeia in 1618 and then in 1676, the book, Observationes Medicae, by the English physician Thomas Sydenham. This latter publication was used for close to two centuries as a standard textbook and contained such “observational remedies” as the use of laudanum (opium in alcohol), quinquina (Jesuit bark preparations for malaria) and iron for iron-deficiency anaemias.

The subtitle for this section could easily be “The experimental chemist discovers the active principles of major drug preparations”. Although it is not often realised, the initial discoveries that can be considered to have revolutionised drug discovery and development were made by European chemists (known as apothecaries at that time) in the 1803–1805 time frame, building upon the physico-chemical principles evolving in the recent past from the work of experimental and theoretical chemists such as Proust, Davy, Gay-Lussac, Berzelius, Dalton, etc. This body of theory and experiment led away from “polypharmacy” towards the “pharmacology” of single (pure) agents” which was probably first enunciated by Cadet de Gassicourt¹¹  in 1809.

An excellent example of this change would be the story of morphine 1. The initial report of the isolation of fractions from the opium poppy was reputedly made by Derosne¹²  in 1803 at the Institute of France and then published in 1814.¹³  However, this preparation had no narcotic properties whatsoever and was probably noscapine with a little meconic acid extracted by the ethanol–water system that he used. A controversy arose because the German pharmacist Seturner then published his work in 1805,¹⁴  claiming that he had commenced work before Derosne; however inspection of this title implies investigation of the acidic and not the basic fractions of opium, probably meconic acid, as shown in a paper the following year.¹⁵  In 1817, Seturner's use of a different extraction technique—hot water extraction followed by precipitation with ammonia—led to colourless crystals that had the narcotic properties of opium.¹⁶  What surprised the scientists reading this publication at the time was that the material obtained was alkaline, not acidic; thus this was the first non-acidic compound with biological properties purified from a plant.

Subsequent conversion into heroin 2 was first reported in 1874 by Wright in the UK as a result of boiling morphine acetate; the process was commercialised by Bayer AG in 1898. The subsequent use and abuse of these compounds is much too complex to discuss here, but one major discovery came in the early 1970s when Pert and Synder reported the identification of opioid receptors in brain tissue.¹⁷  This report was followed closely by the identification of “endogenous morphine-like substances” in 1975 by Kosterlitz and Hughes,¹⁸  which over the next few years led to the identification of enkephalins, endorphins and dynorphins—all of which had the common N-terminal sequence of Tyr-Gly-Gly-Phe-(Met/Leu), leading to the concept that morphine actually mimics this sequence.¹⁹ 

Irrespective of the exact timing of the isolation of morphine, alkaloids were discovered at an ever increasing rate from plant sources over the next 50 or so years, confirming the influence of chemistry on pharmacology and drug development in its simplest form from 1817 to roughly the middle of the 19th century. Thus, emetine 3 was probably the first alkaloid to be purified and reported by Pelletier and Magendie²⁰  in 1817 from Ipecacuanha, closely followed the same year by the isolation of strychnine 4 from Strychnos by Pelletier, now working with Caventou. Then in 1820, the same workers reported the isolation of quinine 5 from Cinchona species.²¹  They developed a commercial process for the preparation of quinine and, in addition to its subsequent use in the treatment of malaria, it was also used extensively as a “tonic” and an antifever drug. Though not documented specifically, the realisation of its use for malaria probably followed after the usage for other “illnesses” in northern Europe.

Over the next few years, a veritable “treasure trove” of potential drug structures was reported, though one must realise that the structures were not identified for many years if not many decades following their initial isolation. Thus in 1819, brucine 6 and caffeine 7 were purified, followed in 1820 by colchicine 8, in 1820, codeine 9,²²  in 1833, atropine 10²³  and papaverine 11²⁴  in 1848. During this time frame, the first plant-derived alkaloid to be purified, have its structure elucidated and finally synthesised was coniine 12. The compound was extracted in 1826, followed by determination of its structure in 1870 and then synthesised by Ladenberg²⁵  in 1881. Even today, over 180 years since it initial isolation, the compound is still a candidate for drug discovery, this time being a model for induction of apoptosis in trypanosomal infections.²⁶ 

No historical perspective of natural product derived drugs would be complete without a discussion of aspirin (acetylsalicylic acid)—probably the most widely utilised drug of all time when the numbers of tablets consumed worldwide on an annual basis are considered. Even today, where presumably the major pharmacological effect is modulation of the cyclooxygenase isoforms, its full activity is still not fully defined.

Salicin 13 was first introduced into medicinal use by Maclagan²⁷  in 1876 as the single agent, although as a part of “herbals” the use of extracts of Salix or Spiraea ulmaria (the source of “spirin” in “aspirin”) for treatments of fevers and pain dated from the days of Hippocrates. It is probable that the formal identification of salicin in more “modern” medicine would be the letter from the Reverend Edmund Stone to the President of the Royal Society in 1763 covering the use of the compound for treatment of “fever”. There are various “stories and/or anecdotes” over the transition from salicin to acetylsalicylic acid (aspirin), but the most probable steps were as follows.

Piria, working with Spiraea species, first isolated salicylaldehyde in 1839 while working in Dumas’ laboratory and then prepared salicylic acid.²⁸  This was followed by the first synthesis of acetylsalicylic acid in 1853 by Gerhardt²⁹  but without any pharmacological investigation. The story that Hoffmann at Bayer synthesised acetylsalicylic acid to overcome the taste problems with salicylic acid in order to help his father “take his medicine” has been revised relatively recently by Sneader.³⁰  It would appear, in spite of some discussion on Sneader's paper in later issues of the British Medical Journal, that acetylsalicylic acid was synthesised by Hoffmann, but “under the direction of ” Arthur Eichengrün at Bayer and that the compound was in fact under “impromptu clinical testing” before the 1898 time frame. Eichengrün left Bayer in 1908 and proceeded to develop materials and processes such as flame-retardant fibres and also injection moulding of plastics. However, when the Hoffmann story was published as a footnote in a book on chemical engineering in 1934, ³¹  Eichengrün, being of Jewish descent, could not comment due to the political climate of 1930s Germany and it was not until 1949 that he was able to publish his story of the discovery.³² 

In 1775, the English physician Withering reported his extensive work on a potential treatment for “dropsy” that he had developed as a result of studies on a plant decoction that the local inhabitants were using for their own treatment. He subsequently investigated the methods used and identified the foxglove, Digitalis purpurea, as the potential source and also demonstrated what today would be known as a “narrow therapeutic index” for the preparation. It was realised subsequently that purification would have to be performed, but in spite of significant efforts, it was not until 1867 that Nativelle³³  was able to produce an effective crystalline preparation that he named “crystallised digitalin”. A few years later in 1875, the German pharmacologist, Schmiedeberg was able to produce digitoxin 14.³⁴  Subsequently, other compounds with a similar pharmacology were discovered by a combination of what would now be known as ethnopharmacology/ethnobotany—yielding ouabain from Acocanthera bark and roots and strophantin from Strophantus species. Both of these agents were used as arrow poisons in Africa, albeit in very crude form.

Inspection of any textbook of pharmacognosy, pharmacology or natural product chemistry in the middle of the 20th century would show a large number of very similar “cardiac glycosides” isolated from a multiplicity of plant sources and, even as the bufodienolides, from Amphibia. However, it was not for three-quarters of a century after the purification of digitalis that a mechanism of action of these agents was firmly established as inhibitors of the sodium/potassium ATPase pump in membranes. Even today, novel agents based upon variations of an unusual plant-sourced “cardiac glycoside” are in preclinical trials as antitumour drugs with an example being UBS-1450 15; oleandrin 16 is the subject of a 2008 PCT patent application for use as a “functional food” in cancer therapy by a Japanese company.

The breadth of disease areas that could be covered from a historical perspective starting in the mid-1920s or so is staggering. Due to space limitations and because specific areas are covered by other authors, we have limited our discussions to early antibiotics (antibacterial, fungal and viral) and cancer, even though the initial work in this “series of diseases” (cancer) was performed using materials from war gases (the nitrogen mustards/alkylating agents) and then, in the early 1960s, workers using both microbes and plant sources proceeded to report and then to use multiple agents from these sources as treatments for cancers. For all these diseases, we give examples in each case of how the early historical structures have led to novel agents based upon them still being synthesised (or even biosynthesised) in the 2008 time frame.

It is probably true that if one had to name the natural product that has saved the most lives, directly or indirectly since its original discovery, penicillin G 17 would be the molecule of choice. In this day and age, there are few people in the developed countries who can remember the pre-antibiotic age with any clarity. Some, over the age of 75, may have hazy memories of relatives dying at young ages due to bacterial infections, but that is not the norm. However, let us set the stage for the reader.

The first usage of natural products as true antibacterials rather than as surface sterilants (e.g. use of thymol and other essential oils) can be fixed in time as the later stage of World War II, with the use of microbial derived secondary metabolites such as penicillin and streptomycin being the exemplars known in the West. This occurred as a result of the recognition by Fleming in the late 1920s of the activity of penicillin (though there were anecdotal reports of scientists such as Tyndall, Roberts and Pasteur in the 1870s recognising antagonism between various bacteria), leading ultimately to the well-known and documented use of penicillins G and V³⁵  and streptomycin (discovered by Waksman³⁶ ) in the early 1940s. However, it now appears that the (then) USSR was using the antibiotic Gramicidin S (Soviet Gramicidin^37–39 ) as a treatment for war wounds at the end of World War II.

It is recognised, however, that the use of Prontosil® 18 pre-WWII led to the introduction of synthetic antibacterials, with the first clinical efficacy report in 1933 ultimately leading to the award of the Nobel Prize for Medicine in 1938 to Domagk. This could also be thought of as the first formal prodrug in the antibiotic field as the active principle, sulfanilamide 19 is a structural analogue of para-aminobenzoic acid (PABA) and an essential nutrient of many bacteria and in particular, the cocci. PABA competitively inhibits dihydropteroate synthase, thus leading to inhibition of folic acid and bacterial death. So although synthesised in the absence of such knowledge and for an entirely different purpose, it was in retrospect an isostere of a natural product.⁴⁰  Using the nomenclature of Newman et al.⁴¹  this would now be classified as an “S/NM” or “synthetic but natural product mimic”.

Although the aminoglycosides such as streptomycin, neomycin and the gentamicins have a long and storied history as treatments for antibacterial infections, particularly in the early days when streptomycin was a treatment for both infected wounds and also for tuberculosis, few modifications of the basic molecule(s) went into clinical use, mainly due to the complexity of chemical modification of saccharidic-based structures. Thus, we do not discuss this class further or molecules such as the rifamycins and their manifold derivatives. Instead, due to space constraints, we show how β-lactams, macrolides, tetracyclines, glycopeptides and pleuromutilins—all “ancient antibiotic structures”—are even today still being utilised as base scaffolds on which to build molecules.

To date, the number of penicillin and cephalosporin-based molecules produced by semi- and total synthesis is well in excess of 20 000. Most started with modification of the fermentation product, 6-amino-penicillanic acid 20 or the corresponding cephalosporin, 7-amino-cephalosporanic acid 21, both of which can be produced by simple chemical or biochemical deacylation from penicillin or cephalosporin C. The number above is only approximate as a significant proportion of structures from industry were never formally published, or were only mentioned in the patent literature—particularly if they had marginal or no significant activity levels over those which had been reported previously.

In 1948, the ring-expanded version of penicillin, cephalosporin C 22, was reported from Cephalosporium sp. by Brotzu; its structure was determined in 1961 by the Oxford group.^42,43  As with the penicillin nucleus, this ring-expanded molecule, 7-aminocephalosporanic acid 21, also served as the building block for many thousands of cephalosporins with the first orally active molecule, cephalexin 23 being introduced in 1970. Since that time, a multitude of cephalosporins have been synthesised with the aim of producing molecules that are more resistant to β-lactamases.

In order to give extra “medicinal life” to β-lactams that were no longer useful due to the presence of both constitutive and inducible β-lactamases, efforts were made in the late 1960s and early 1970s—particularly by Beecham (now part of GlaxoSmithKline) and Pfizer—to find molecules that would have similar pharmacokinetics to the β-lactams but would inhibit the “regular” β-lactamases that were part of the pathogenic microbe's defence systems. Beecham discovered the naturally occurring clavulanate family, with clavulanic acid 24 being incorporated into the combination known as Augmentin® (a 1 : 1 mixture of amoxicillin and clavulanic acid launched in 1981), thus extending the franchise of this particular β-lactam well beyond its original patent date. The Pfizer entrant was basically des-amino penicillanic acid 25 with a sulfoxide in place of the sulfur; in tazobactam 26, which was originally synthesised by Taiho and launched by Lederle (now Wyeth), one of the gem methyl groups was replaced by a 1,2,3-triazol-1-yl-methyl substituent. Even today, 17 years after the last introduction, no other β-lactamase inhibitors have made it to commercialisation. Currently, as mentioned earlier, amoxicillin is combined with clavulanate or ticarcillin, sulbactam with ampicillin, and tazobactam with piperacillin. All these inhibit only class A serine-based β-lactamases, leaving a vast number of other enzymes where inhibitors are required.⁴⁴ 

Along with the search for the β-lactamase inhibitors, efforts were underway to produce the simplest β-lactam, the monobactam. Following many years of unsuccessful research at major pharmaceutical houses, predominately in the synthetic chemistry areas, came reports from Imada et al. in 1981⁴⁵  and a Squibb group led by Sykes,⁴⁶  who both demonstrated the same basic monobactam nucleus 27. What is important to realise is that no molecules synthesised before the discoveries of these natural products had a sulfonyl group attached to the lactam nitrogen, which is an excellent method for stabilising the single four-membered ring.

Since that time, a significant number of variations on that theme have been placed into clinical trials and, in some cases, such as Aztreonam® 28, into commerce. Recently (late 2007, early 2008), this compound was submitted for approval in the EU and the USA as the lysinate salt for the inhalation treatment of Pseudomonas aeruginosa in cystic fibrosis under an Orphan drug category. As of late 2008, the Food and Drug Administration (FDA) was requiring further information and the status of the EU application was not yet known.

That these base structures and others discovered after the early 1940s are still valid as scaffolds upon which to base new drugs is shown by the following data. Since 2000, three penems (biapenem 29, ertapenem 30 and doripenem 31), which although produced synthetically were based upon the structure of thienamycin 32, and two cephalosporins—cefovecin 33 (a veterinary drug) and ceftobiprole medocaril 34—have been approved for marketing. Currently, there is one penem, tebipenem pivoxil 35, which has been pre-registered in Japan with the aim of approval in early 2009.

Although we mentioned earlier that only three β-lactamase inhibitors have been marketed, there is now a potential “cepahalosporinase inhibitor/cephalosporin combination” in trials. Forest Pharmaceuticals recently announced that the cephalosporin, ceftaroline fosamil acetate 36, which is currently in Phase III clinical trials, has been combined with Novexel's synthetic β-lactamase inhibitor,⁴⁷  NXL-104 (AVE-1330A) 37 and has entered Phase I trials.

Even though the base molecule or its better known chloro-derivative, aureomycin and later the dimethyl amino derivative, doxycycline, have been stalwart members of the physician's armamentarium for 40–50 years, in 2005, Wyeth had the glycyl derivative of a modified doxycycline molecule, tigecycline 38, approved for complicated skin and soft tissue infections. Tigecycline has broad-spectrum activity including both Gram-positive and Gram-negative bacteria and methicillin-resistant Staphylococcus aureus (MRSA). It shows that relatively simple chemical modifications can even give very old base structures a new lease on life and be effective against clinically important infections.

Vancomycin, a natural product that was first approved in 1955, is still the prototype for structural variations with the same mechanism of action: the binding to the terminal l-Lys-D-Ala-d-Ala tripeptide in Gram-positive cell wall biosynthesis. The compounds below are semi-synthetic modifications of the same basic structural class (glycopeptides) as the prototype vancomycin, thus following in the “chemical footsteps” of the β-lactams; currently, there are three semi-synthetic glycopeptides, oritavancin 39, telavancin 40 and dalbavancin 41, in late stage clinical development.

In all cases, their antibacterial mechanism involves inhibition of cell wall biosynthesis initially via the vancomycin target, although the exact mechanisms can vary with the individual agent. In the case of oritavancin, from very recent data it would appear that the agent is comparable with vancomycin in its inhibition of trans-glycosylation, but is more effective as a transpeptidation inhibitor.⁴⁸  As mentioned above, all are semi-synthetic derivatives of natural products, with oritavancin being a modified chloroeremomycin (a vancomycin analogue), dalbavancin being based on the teicoplanin relative, B0-A40926 and telavancin (TD-6424) being directly based on chemical modification of vancomycin.⁴⁹ 

That one may combine the characteristics of two separate agents working at different targets within the same basic biological area is shown by the work of Theravance (also the originator of telavancin), which has combined a cephalosporin with vancomycin itself to produce the hybrid TD-1792 42, which is currently in Phase II trials against complicated skin and soft tissue infections. Thus, two old antibiotic classes can produce novel agents—again underscoring the possibilities of reworking older structures if one understands their history.

Following on the track of novel modifications of old structures, since 2000 there have been four molecules formally based upon the erythromycin chemotype that have either been approved (telithromycin 43 in 2001) or entered clinical trials; cethromycin (ABT-773; Phase III; 44), EDP-420 (EP-013420, S-013420; 45) and the product of glyco-optimisation, CEM-101 (Phase I; 46). Cethromycin 44 is currently in Phase III trials for use against community acquired pneumonia (CAP) and is being evaluated as an anti-anthrax agent (and against other biodefence targets) by the National Institute of Allergy and Infectious Diseases (NIAID) and the US Army. The interesting modification of the base erythromycin structure, the “bicyclolide” EDP-420 (45) a novel, bridged bicyclic derivative originally designed by Enanta Pharmaceuticals,^50,51  is currently in Phase II trials for treatment of CAP by both Enanta and Shionogi. Interestingly, this molecule is also active in a murine model of Mycobacterium avium, a common infection in immunosuppressed patients,⁵²  which may well expand its usage in the future.

Demonstrating yet again that older antibiotic structures have significant validity for today's diseases, GlaxoSmithKline (GSK) received approval in 2007 for a modified pleuromutilin, retapamulin 47, for the treatment of impetigo in paediatric patients. The base structure, pleuromutilin 48, was first reported in 1951 from the basidiomycete Pleurotus mutilis (FR.) Sacc and Pleurotus passeckerianus Pilat.⁵³  In the mid-1970s, a significant amount of work was reported on the use of derivatives of pleuromutilin as veterinary antibiotics;⁵⁴  thus the subsequent utilisation of the base molecule as a source of human use antibiotics is very reminiscent of the work that led to the approval of Synercid® in the late 1990s, as the base molecules in that case were also used extensively in veterinary applications.

It is quite possible that a number of antibiotics based upon this elderly scaffold will enter late stage human trials as currently there are two “mutulins” in Phase I clinical trials for use against Gram-positive infections. Although structures have not yet been released, they are BC-3205 and BC-7013 from Nabriva in Vienna, with the former for oral use and the latter for topical use.

Although a very considerable amount of time and effort was expended in the early days of antibiotic discovery (by this we mean the mid to late 1940s), only three general use antifungal agents entered clinical practice as a result. Perhaps the first clinically used antifungal natural product (our information on Russian efforts in this field under the old USSR is effectively nil), was griseofulvin 49, which although launched in 1958, was originally reported in 1939. Its non-polyene structure was defined in a series of papers in 1952 using classical techniques⁵⁵  and, even today, close to 70 years after it was first described, is still in clinical use against dermatophytes—the only class of fungi that it is active against; long-term treatment is necessary due to its insolubility.

Perhaps the best known clinical agent is the heptaene polyene, amphotericin B 50, isolated from Streptomyces nodosus and first reported in 1956. The full structure was not elucidated until 1970 when it was determined by X-ray crystallography,⁵⁶  closely followed by a description of the absolute configuration determined by utilising the iodo-derivative for X-ray and by mass spectroscopy.⁵⁷  Quite recently, 50 years after its initial discovery, a full review giving the highlights of the chemistry around the compound was published by Cereghetti and Carreira.⁵⁸ 

Although many polyenes with varying numbers of conjugated double bonds have been reported since those early days, only one other compound of this class (in fact the first identified in 1950 of this general structural class), the tetraene nystatin 51, has gone into general clinical use and like amphotericin B its primary indication is for candidiasis. It was first reported from Streptomyces noursei and, as with amphotericin, its structure was reported in the 1970 time frame by two groups, one using classical chemical degradation plus proton NMR⁵⁹  and the other via mass spectroscopy.⁶⁰  Confirmation of the proposed hemiketal structures of both amphotericin B and nystatin was published subsequently by the Rinehart laboratory in 1976.⁶¹ 

Since 2000, three natural product-derived antifungal drugs from the echinocandin/pneumocandin class of glucan synthase inhibitors have been approved for human use.^62,63  In order, these were caspofungin 52 (2001, Merck) which recently has been shown to function successfully in both invasive candidiasis and in candidaemia,⁶⁴  micafungin 53 (2002, Astellas, see Chapter 15), which is currently in clinical trials for paediatric disease⁶⁵  and anidulafungin 54 (2006, Pfizer).^66,67  Another modification of the basic echinocandin structure, aminocandin 55 (HMR-3270), a semi-synthetic derivative of deoxymulundocandin, is currently in Phase I clinical trials with Phase II studies reported as being scheduled.⁶⁸ 

It can be argued quite successfully (and has been a number of times) that the derivation of the nucleoside-based antiviral agents can be traced back to the time frame 1950 to 1956, when Bergmann et al. reported^69–71  on two compounds they had isolated from marine sponges, spongouridine 56 and spongothymidine 57. What was significant about these materials was that they demonstrated, for the first time, that naturally occurring nucleosides could be found using sugars; importantly other than ribose or deoxyribose and having biological activity, as it had been “then current dogma” that one could change the “nucleoside bases” but ribose or deoxyribose had to be the sugar to maintain biological activity. These two compounds can be thought of as the prototypes of all of the modified nucleoside analogues made by chemists that have crossed the antiviral and antitumour stages since then.

Once it was realised that biological systems would recognise the base and not pay too much attention to the sugar moiety, chemists began to substitute the “regular pentoses” with acyclic entities and with cyclic sugars with unusual substituents. These experiments led to a vast number of derivatives that were tested extensively as antiviral and antitumour agents over the next 30+ years. Suckling, in a 1991 review,⁷²  showed how such structures evolved in the (then) Wellcome laboratories, leading to AZT and, incidentally to Nobel Prizes for Hitchens and Elion, though no direct mention was made of the original arabinose-containing leads from natural sources.

Showing that “Mother Nature” may follow chemists rather than the reverse, or conversely that it was always there but the natural product chemists were “slow off the mark”, arabinosyladenine 58 (Ara-A or Vidarabine®) was synthesised in 1960 as a potential antitumour agent,⁷³  with its antiviral activities reported by Schabel⁷⁴  in 1968 with production via fermentation of Streptomyces antibioticus NRRL3238 being reported in a British patent in 1969⁷⁵  and isolated, together with spongouridine, from a Mediterranean gorgonian (Eunicella cavolini) in 1984.⁷⁶ 

Building on from these original discoveries, medicinal chemists over the next 40+ years made a very large number of “substituted nucleosides” varying the base and the sugar moieties (including molecules that were acyclic), leading to the very well-known antiviral agents, acyclovir 59 and its later prodrug derivatives and AZT 60.

Although a significant number of antiviral vaccines have either been approved or are in clinical trials for a variety of viral diseases, small molecules based upon “modified nucleosides” are still being approved by either the FDA or the European Medicines Agency (EMEA). As in earlier days, agents originally approved as antiviral agents may later be shown to have potential utility as antitumour agents.

Since 2000, seven such agents have been approved for antiviral treatments covering anti-HIV, hepatitis B and cytomegalovirus (CMV). Rather than give details of each, we discuss below the importance of just two compounds of this class that would not have been synthesised without the historical perspective.

In 2001, tenofovir disoproxil fumarate 61, a prodrug of tenofovir was approved for treatment of HIV, subsequently being preregistered in the USA for treatment of hepatitis B. Emtricitabine 62, a reverse transcriptase inhibitor, was approved in 2003 for HIV. What is of import is that these compounds are now part of fixed dose combination therapies for treatment of HIV, either two drug (tenofovir disoproxil fumarate/emtricitabine) or three drug Atripla® (tenofovir disoproxil fumarate/emtricitabine/efavirenz) formulations. Thus, even 50+ years after Bergmann's discovery of bioactive arabinose nucleosides, small molecules synthesised as result of his discoveries are still in clinical use and others are in clinical trials for treatment of viral diseases.

A recent book⁷⁷  details most of the agents from natural sources that have entered the oncologist's armamentarium over the last 50 or so years. However, it is indicative of the importance of natural product sources that 14 microbial-sourced pure compounds have been approved for use in various countries since 1954 (Table 1.1) and ten modified microbial-sourced natural products (Table 1.2) for a total of 24 compounds from this source.

In addition to these, there are 13 products from (nominal) plant sources that have entered clinical use (Table 1.3). Of these, only three are the natural products, the rest are derivatives. However, just as in the marine environment, where there is now significant evidence of the involvement of microbes/protists (single-celled organisms from all three domains of life) in the production of the secondary metabolites isolated from the host macroorganism, there have now been a significant number of recent publications^78–86  that demonstrate that endophytic fungi isolated from the plants thought to be the sources of the base compounds in Table 1.3 can produce the same compound—albeit in very low yield—on fermentation of the purified microbes under conditions where “carryover” is not feasible. An argument that was used against such reports was based on the very low yields seen. However, as shown by Keller's group, the genetic control of secondary metabolic clusters in fungi (Aspergillus nidulans) is extremely complex⁸⁷  and it is possible that the very low yields are due to a lack of information as to the control systems involved.

From the history and examples presented above, it can be seen that natural products in the developed world have led to many different drug entities. It should be emphasised that, in the other ∼80% of the world, combinations of plants (and their associated microflora) are still the major source of medications for the manifold illnesses that afflict mankind.

Lest one might assume that natural products have had their day, we finish with two pie charts (codes from Newman et al.⁴¹ ) that demonstrate the continued involvement of Mother Nature's chemistry late in 2008 covering the sources of small molecule drugs, January 1981 to October 2008 (Figure 1.1), and sources of small molecule antitumour agents the 1930s to October 2008 (Figure 1.2).

In closing, we suggest that interested readers should consult the following three recent review papers that demonstrate how, even today, almost 20 years into the combinatorial chemistry era, chemists and biologists are still “learning chemical history from Nature”—the review by Kaiser et al. on biology-inspired compound libraries,⁸⁸  and the two reviews on natural products in the modern age by Ganesan⁸⁹  and Butler.⁹⁰