New drugs from old bugs

Advances in the understanding of their biosynthesis are renewing interest in natural products - and in particular the polyketide metabolites - as valuable leads for drug discovery. Barrie Wilkinson, head of chemistry at Biotica, discusses the options.

The track record of natural products (NPs) as a source of valuable and innovative therapeutic agents is unrivalled. NPs and their derivatives (semi-synthetic derivatives and NP-inspired compounds) together represent some 40% of the top 100 selling prescription drugs, with estimated pharmaceutical sales of in excess of US$40bn during 2003. Viewed another way, of the 877 new small chemical entities introduced between 1981 and 2002, roughly half (49%) were NPs, their derivatives, or NP-inspired compounds.complex processes

Despite these successes NPs have fallen from favour within the pharmaceutical industry. Particular reasons include difficulties in synergising high-throughput screening methods with crude NP containing extracts, and the rise of combinatorial chemistry.

The complexity of NP structures (viewed by many as the reason for their historical success) has compounded these issues, as this complexity does not always sit comfortably within the lead optimisation process.

There are signs, however, that NPs are once again provoking interest as a source of lead compounds. Several screening and biotechnology companies have sprung up during the past decade to exploit advances in NP science, and governmental interest can be seen by the funding of several programmes: the DTI's 'LINK Post Genomics' initiative in the UK; and, in the US, the call for new applications to the National Institutes of Health for development of 'new methodologies for natural products chemistry', which falls within the Molecular Libraries & Imaging component of the Roadmap for Medical Research.

Of the US$40bn of NP-related sales for 2003, around half can be attributed to a single class of compounds - the polyketides. This group of natural products encompasses a bewildering array of structural space and biological activity. These diverse compounds, produced commonly by soil-dwelling bacteria, are linked by a common and straightforward biosynthetic origin: the condensation of simple carboxylic acids in a process akin to that of fatty acid biosynthesis.

Although quite different from the small, Lipinski-type molecules that form the basis of most screening campaigns, polyketides make excellent lead compounds as drug candidates for a number of reasons. Their typically large molecular weight translates to a significantly greater molecular surface area that is available for interacting with protein surfaces. This characteristic means that polyketides are excellent at mediating complex protein-protein interactions and, in particular, are able to modulate complex interactions between multiple binding partners - an arena in which small molecule collections are often found wanting. This offers great potential for affecting cancer-associated pathways, where many important targets participate in numerous complex interactions.

biosynthetic engineering

There are, of course, problems as well as advantages inherent in following polyketides as lead compounds. Their relatively high molecular weight, coupled with stereochemical complexity and richness of functional groups, often means that total, and even semi-synthesis, can be extremely difficult, an issue that can limit the medicinal chemist's ability to explore structural space around lead compounds.

In response to these limitations chemists and biologists have learned how to harness the natural mechanisms for generating these complex structures. The ability to rationally alter nature's biosynthetic machinery has been most effectively demonstrated for the polyketides, leading to the new avenue of research described as biosynthetic engineering.

The biosynthesis of polyketides is achieved by large 'megasynthases', called modular polyketide synthases (PKSs), which operate as 'production lines' for polyketide assembly. These systems contain repeating modules of activities, where each module contains all of the catalytic functionality required to catalyse one round of polyketide chain elongation and its modification. Inherent to this process is the selective utilisation of varying types of monomer; the selection of differing levels of reductive processing of introduced b-keto groups; and the controlled introduction of multiple stereochemical centres.

In addition, the PKS controls the number of chain elongation cycles performed, as well as the choice of chain termination strategy employed; this latter step may lead to the further introduction of functional groups and/or macrocyle formation, including varying chemo-types and ring sizes.

data sequences

These PKS products are further elaborated by the action of 'tailoring enzymes' to introduce oxygen or other heteroatoms, acyl and alkyl groups; and, very commonly, complex deoxysugar moieties. In this way another layer of complexity is added to poly-ketide structures.

This combination of a few simple biosynthetic steps, carried out upon relatively few simple precursors, is responsible for providing the great structural diversity of the polyketides.

Moreover, the molecular logic which provides for this assembly process is readily understood, is common to the biosynthesis of most polyketides, and is 'hard wired' into the DNA sequence encoding any given polyketide biosynthetic gene cluster. It is, in fact, possible to convert the data encoded within the DNA sequence directly into a predicted structure of the molecule, for which the genes encode the production.

This molecular logic also provides an opportunity for altering the structure of the particular polyketide produced: hybrid genes can be generated by utilising sequence data from multiple gene clusters and splicing together specific sequences in a controlled manner. The resulting hybrid genes give rise to hybrid proteins, which subsequently lead to the production of new compounds with rationally predetermined structural variation.

new weapon

It is the ability to harness these biosynthetic processes in a rational and controllable manner that provides a novel and valuable addition to the drug discovery armoury.

Biotica has developed a powerful range of biosynthetic engineering methods that are able to produce a number of polyketide analogues on a selected template by exploiting the molecular logic described above. Historically, the biosynthesis of novel antibiotic macrolides was the first area for the use of this technology, but recent developments in anticancer therapeutics have revealed new polyketide templates for which this technology is able to open up new chemical space.

A particularly illuminating example can be found for the immunosuppressive and potential antitumour agent rapamycin (Sirolimus): a macrolide polyketide produced by the soil bacterium Streptomyces hygroscopicus. Rapamycin and its analogues are the focus of much current clinical interest as the only validated inhibitors of the central kinase mTOR (Mammalian Target Of Rapamycin).

mTOR is a serine/threonine kinase (a member of the phosphoinositide-kinase-related-kinse family) and a central controller of eukaryotic cellular processes related to growth and proliferation, sensing nutrients and mitogen and allowing progression from G1 to S phase. As such mTOR represents a promising target for potential anticancer therapy.

novel space

Rapamycin mediates its action through complex protein-protein interactions. It first binds the cytosolic protein FKBP12, before the resulting heterodimeric complex binds mTOR. This interaction with mTOR is atypical for compounds acting upon kinases as it targets a site distal to the ATP binding site and exerts its inhibitory effect upon kinase activity through an allosteric interaction.

Biotica has applied its biosynthetic engineering platform to produce rapamycin analogues (named rapalogues), and augmented this with chemical synthesis to produce a library of rapalogues that occupies novel chemical and structural space broader than that addressed to date using synthetic chemistry alone; i.e. such as for the small number of rapamycin-based mTOR inhibitors in clinical trials (CCI-779 (Wyeth) and RAD-001 (Novartis)).

This was achieved by deleting multiple genes within the biosynthetic cluster and reintroducing them by combinatorial methods. This was further combined with orthogonal feeding of exogenous carboxylic acids and amino acids as the deleted strains lacked the ability to make certain natural precursors of this type.

platform technology

This library not only addresses issues of potency but contains rapalogues with very different physiochemical properties that address limitations of solubility and metabolism inherent to the parent compound and its simple analogues. This approach has led to the production of promising drug candidates with properties distinct from other rapamycin analogues.

Biotica is also applying its platform technology to develop polyketides as lead compounds against another promising cancer target, HSP90, and as inhibitors of angiogenesis.

In addition to the biosynthetic engineering methods highlighted here, a number of other 'new generation' approaches for mining the value of NPs have arisen.

These include new ways for finding the cryptic compounds encoded within the genomes of micro-organisms, and through the expression of environmental DNA in suitable hosts, thereby providing access to the genetic and biosynthetic potential of silent and unculturable organisms. It is believed that over 99% of microbial diversity has yet to be cultured and have its chemical potential investigated.

Through the successful application of these new approaches, in combination with the significant advances in separations and analytical science and structure elucidation capabilities made during the last decade, it appears that polyketides and other natural products may yet make a return as one of the major sources of lead compounds in drug discovery.