High-throughput metabolomics are an essential tool in antibiotics discovery. The Netherlands Metabolomics Centre and the research group of Prof. Thomas Hankemeijer participate in CARES.

Microorganisms generally produce several different antibiotics. The “early” antibiotics, those dicovered in the 1940s, occur in 1-10% of all streptomycetes. Vancomycin, which was discovered in 1953, is found in 1 on 100,000 microorganisms and daptomycin, discovered in the early eighties, in only 1 in 10 million. This illustrates that the low-hanging fruit has been picked and that it is getting increasingly difficult to find novel compounds. Moreover, when you search for novel antibiotics, these more commonly occuring antibiotics tend to dominate the results. Therefore, in screens for novel antibiotics, primarily known compounds are found. This phenomenon, called replication, is one of the main reasons for pharmaceutical companies to pull out of antibiotic research. Advanced metabolomics strategies can be applied to circumvent this problem.

Metabolomics

To discriminate known from novel bioactive compounds, the molecules need to be identified. This requires new Mass Spectrometry- and NMR-based metabolomics technologies. Early identification allows for efficient prioritization of promising molecules: valuable time and resources can be spent on novel rather than known compounds.

High throughput technologies

The metabolomics workflow should be fully automated to allow the screening of larger number of positive samples from the initial screening. In addition, compounds identified should be validated on their antimicrobial activity in an automated and efficient manner. The Leiden Metabolomics Centre focusses on the development of metabolomics based technologies and instrumentation and provides access to their high throughput technology.

Despite the extensive screening efforts in the 20th century, there is still a huge repository of yet to be discovered antimicrobial compounds. A major challenge we now face is to access these hidden treasure troves and to exploit nature's biodiversity. This program aims at three major approaches.

(1) Utilizing known antibiotics-producers

Filamentous microorganisms (filamentous fungi and bacteria of the order of actinomycetales) are the major antibiotic producers. Despite 70 years of extensive exploitation of these microorganisms, sequencing of their genomes established the presence of silent (also referred to as sleeping or cryptic) antibiotic biosynthetic gene clusters, suggesting that the potential of these organisms for novel drug production is much larger than originally anticipated. Therefore, it is imperative that technologies are developed to activate the expression of silent antibiotics, or to express the gene clusters in an optimised “super”-host, and to discriminate them from known bioactive molecules.

(2) Accessing unculturable microorganisms

Moreover, many of the antibiotics that have not yet been identified are likely produced by microorganisms that cannot be genetically engineered or are unculturable. Metagenomics approaches, where all DNA from an environment is sequenced, revealed a huge number of uncultured microorganisms. Only a fraction (ca. 1%)  of the microorganisms was ever grown in a laboratory. Many microorganisms cannot be easily grown in the lab because e.g. they live in symbiosis with other (micro)organisms, or because they need specific compounds that are present in their natural environment but not in the lab. The 99% “unculturable majority” represents a huge unexplored potential in nature that can be unlocked.

(3) Exploiting compound classes

Finally, there is an enormous potential in making new-to-nature antimicrobial compounds. There are several compound classes that have raised attention for their possibilities:

• Posttranslationally modified peptides (e.g. lanthipeptides, sactipeptides and lassopeptides) are highly stable molecules and have great combinatorial possibilities by shifting microdomains and combining modification enzymes, and combinatorial gene libraries can comprise over 100.000 species.
• Non-ribosomally produced peptides (NRP’s) such as the glycopeptide antibiotics vancomycin and daptomycin, or the beta-lactam antibiotics (penicillins, carbapenems, cevalosporins) are formed by large protein complexes called non-ribosomal peptide synthetases (NRPS). NRPS can be readily engineered to obtain new-to-nature compounds with high specific activity.
• Other major classes of antibiotics whose potential is still far from exhausted include the polyketides (tetracyclins, erythromycin, rapamycin) and the aminoglycosides (kanamycin, streptomycin). However, resistance against these antibiotics is often readily obtained by target modification.
• An important category of molecules that so far has received relatively little attention and are therefore specifically mentioned, are molecules that target resistance. By far the best known example is clavulanic acid, which targets beta-lactamases. Clavulanic acid is added to the drug amoxicillin to allow treating infections with beta-lactamase-producing pathogens, and marketed as Augmentin©. The success of clavulanic acid is due to its pharmacokinetic properties which are very similar to those of the beta-lactam antibiotics themselves. Specific approaches to find molecules that tackle antibiotic resistance are called for.

Bioinformatics

Analysing the genomes of either known antibiotics-producers, or yet uncultured microorganisms from nature, requires bioinformatics to interpret the sequence data. By looking for sequences resembling those of genes known to be involved in antibiotics production, like polyketide synthases (PKS) or nonribosomal peptide synthetases (NRPS), novel gene clusters can be identified.

Synthetic biology

Once novel, promising gene clusters are identified, those candidate gene clusters should be expressed in an optimized host to verify the lead. Therefore, it will be crucial to develop an efficient pipeline for heterologous production of biosynthetic gene clusters. Large pieces of DNA (>50 kilobases) will have to be synthesized in an automated manner and cloned as well as modified, making use of innovative synthetic biology technologies such as yeast artificial genomes and the CRISPR/CAS system. In that way, the antibiotic can be produced in a tractable production host which is important for upscaling and for genetic manipulation. Moreover, synthetic biology is crucial in unlocking the potential of compound classes like NRP’s and posttranslationally modified peptides. For example, NRPS are modular enzymes, where the domains specify the catalytic steps and therefore determine which NRP is formed. By shuffling the domains, it is possible to create entirely new NRP’s that can then be screened for antimicrobial properties.

Natural producers of antibiotics often don't make high amounts of these compounds. Increasing the production of antibiotics is vital in both the research and the development phase.

Chemical Ecology

Low production of antibiotics is a general problem. To enable detection of novel antibiotics, they need to be present at sufficient amounts. However, the discovery of silent antibiotic biosynthetic gene clusters in known antibiotics-producing organisms suggested that there is still an unexplored potential for novel drug production in these organisms. To activate expression of these gene clusters, the ecology of the microorganisms can provide leads.
There are several successful examples of this concept: In the lab bacteria are grown in axenic cultures, i.e. as a single species. However, in their natural environment, bacterial communities consist of many different species producing all kinds of signals that trigger responses. Recent studies show that co-cultivation can be an effective strategy for finding novel molecules. Another approach is to look for molecular switches that trigger expression of silent antibiotics. From the ecology of actinomicetes it is known that many antibiotics are produced during the starvation phase, when the cells break down their mycelium to produce spores. It turns out that one of the break-down products of the cell wall, N-acetylglucosamine, functions as a trigger for the production of antibiotics, including antibiotics that are not commonly produced by this organism. Finally, novel culturing techniques can also help  to grow microorganisms that belong to the 99% unculturable majority, and have never been grown in a lab. The iChip from Kim Lewis’ lab is a nice example of this approach: by locking bacteria in a chip and placing the entire chip in the natural environment, it is possible to grow many more microorganisms than by trying to isolate bacteria the traditional way (streaking on a plate). Lewis and his team discovered a new antibiotic using this technique: teixobactin.
In these examples, fundamental knowledge from microbial ecology is successfully applied to increase antibiotics production. This indicates the importance of understanding the ecology of these microorganisms and the necessity to develop novel culturing technologies.

Bioprocess design

When promising lead compounds are identified, it is vital to increase production to enable thorough characterization of the chemical and pharmacochemical properties of the molecules. Also further into the R&D-pipeline, larger amounts of antibiotics are needed for preclinical and clinical testing.
To obtain significant amounts of a compound, the production needs to be scaled up to larger volumes: instead of using plates or shake flaks, bioreactors are needed. Moreover, the production processes should be optimised. Expertise in bioprocess design is required to accomplish this.

Once novel antibiotics have been discovered with specific activities, preferably small spectrum, against major MDR pathogens (ESKAPE pathogens, MDR-TB, CDI), they can be further developed by e.g. chemical modification, and thoroughly characterized before going into clinical development.

Medicinal chemistry

Promising compounds can be chemically modified by employing state-of-the-art organic chemistry, making use of approaches such as click chemistry or protective aptamers. The combined technologies offer an unprecedented combinatorial power to produce hundreds of thousands of different compounds that can be screened for desired antimicrobial activity.

Prioritizing drug leads and clinical development

The chance of success of drug leads needs to be assessed. All promising compounds will be produced in sufficient amounts and thoroughly characterized with respect to antimicrobial spectrum, structure, stability, pharmacokinetics, toxicity and route of administration. A vital issue is the rate of resistance development: to determine the natural reservoir of resistant microbes and the resistance mechanism. To prepare for stage I clinical trials, testing in animal models will be performed.

When molecules have successfully passed all initial tests and trials, high-yield production systems need to be developed, e.g. by safe and sustainable biotechnological production systems.

Translational research on the mode of action and the onset of resistance

Mode of action studies will be based on a combination of genomics studies to assess the cellular responses to the drugs, which gives insight into the likely cellular target (cell wall, DNA, RNA or protein synthesis, etc), suppressor selection and structural drug-target interaction studies as well as molecular modeling.

Based on the clinical observational studies and metabolomics, proteomics, RNA seq and NGS, existing and new models will be translated into antibiotic killing assays and models to study the development of resistance. A recent example is a metabolomics profile of TB in patients and zebrafish larvae, providing insights in underlying disease mechanisms and to identify and test novel treatment options, and to test identified compounds. Another option is the development of specific organ-on-a-chip models as in-vitro model to study antibiotics and resistance in a specific tissue environment of the host. In-vitro antimicrobial susceptibility testing is done with a well-defined set of reference microorganisms to determine the minimal inhibitory concentrations and minimal bactericidal concentrations.