The fight against multi-resistant germs

Artificial intelligence (AI) has greatly advanced antibiotic research in all areas. Today, AI can analyse millions of genetic sequences, chemical structures and laboratory findings in a matter of moments, helping to identify patterns. This makes it much easier to identify potentially effective substances in natural products — a significant advantage, given that it would otherwise take years to develop and test thousands of natural products. AI models can learn which molecular characteristics are associated with resistance or efficacy, enabling them to predict which substances are most likely to be effective against multi-resistant germs. AI therefore accelerates screening, reduces failures and helps resources to be used in a more targeted manner.

Müller and his research team are working with AI to identify natural substances that are effective against bacteria. These substances arise in environments where microorganisms compete with each other, such as in soil. There, many bacteria and fungi produce antibiotic substances in order to prevail over others. The MICROBELIX project provides a basis for discovering such new antibiotics. Initially, it is purely citizen science: volunteers can send in soil samples from as many natural habitats as possible. HIPS then examines these samples using metagenome analysis to discover new soil bacteria, with the aim of finding clues about soil biodiversity, as well as potential new active substances that could be used to treat infections. “We have already received well over 1,000 samples and, thanks to Microbelix, found some promising candidates whose effects we are now testing in preclinical research,” says Müller.

Of all the substances, corallopyronin A is closest to clinical approval. It acts as an inhibitor of bacterial DNA-dependent RNA polymerase (RNAP), which is the enzyme that enables the translation of DNA to RNA in bacteria. Müller and his team are developing corallopyronin A in collaboration with DZIF researchers in Bonn. “We are currently on the verge of entering the clinical research stage with corallopyronin A and are optimistic that our substance will overcome the hurdles,” says Müller.

A few years ago, a pharmaceutical company would have stepped in and taken over the further development steps based on the convincing preclinical study data. However, given the current stage of development of their substances, it is unlikely that the industry will become involved. Large pharmaceutical companies in particular are increasingly withdrawing from antibiotic research because it is no longer considered lucrative. There are several reasons for this, including the fact that prices achieved are considered far too low. Additionally, infections are typically cured within a few days, meaning only a limited amount of medication is required — in stark contrast to drugs for chronic diseases, such as insulin for diabetes, which must usually be administered lifelong.

This problem has now been recognised, and the regulatory authorities are attempting to counteract it. There are several government-led approaches to solving the problem. The so-called voucher model, for example, is currently being trialled. 'This means that when the industry develops an antibiotic, it can apply for faster approval of another drug,' explains Mark Brönstrup, Head of the Department of Chemical Biology at the HZI in Braunschweig. Another approach is the so-called 'Netflix model': companies receive a fixed amount for developing and making an antibiotic available, regardless of how much it is actually used. 'Many ideas are currently being discussed to encourage the pharmaceutical industry to invest in antibiotic research again, but a major breakthrough is still pending,' says Brönstrup.

At least smaller pharmaceutical and biotech companies are specialising in antibiotic research in response to these incentives. However, it will likely be some time before sustainable successes can be celebrated and the major pharmaceutical companies rejoin the field of antibiotic research. Until then, the pressure to find new substances with antibiotic properties will continue to weigh heavily on academic research.

Research programmes within the framework of public-private partnerships have proven successful in this regard. Once a promising candidate has been identified, public institutions or foundations provide funding to further advance development. For instance, the development of Corallopyronin A is being funded to the tune of €5.6 million by the Japanese Global Health Innovative Technology (GHIT) Fund, among others.

A few rooms away from Müller's working groups, Anna Hirsch's research team is taking a different approach to the problem. “Rather than taking a molecule and trying to figure out its mechanism of action, we start with a biological target structure. Using computer-aided and experimental methods, we design and produce a molecule that could be a good fit,” says Hirsch. She also has a number of substance classes in development. The most advanced of these is a substance class that acts against Pseudomonas aeruginosa, a widespread hospital germ that is becoming increasingly resistant to common antibiotics.

Hirsch and her team have taken an unusual approach to treating infections caused by Pseudomonas aeruginosa. They have developed a class of substances that does not kill the bacterium, but rather reduces its virulence — that is to say, its harmful effect on the human body. This substance class targets an enzyme in P. aeruginosa that enables the bacterium to break down human tissue and evade the immune response. This enzyme is called elastase and is often abbreviated to LasB in technical jargon. “By inhibiting this protein, we prevent Pseudomonas aeruginosa from spreading in humans and allow the immune system to intervene.”

This is, in a sense, a pacifist approach. 'We don't want to destroy the bacterium itself, but rather to disarm it and render it harmless so that it can be eliminated by the immune system,' explains Hirsch.

The big advantage of this method is that resistance is likely to develop much more slowly. This is because certain strains of bacteria are killed by an antibiotic, except for those that have mutated in such a way that the antibiotic is no longer harmful to them. However, if the bacterial strains are not eliminated anyway, the pressure to develop resistance decreases.

In addition to the search for natural substances and the design of new active ingredients, which are currently being pursued worldwide in a number of academic research locations, there are other smaller branches of antibiotic research that could also lead to new ways of combatting multi-resistant germs. One such branch is bacteriophage research, which is being conducted in Germany at the DZIF and the HZI, among others. Put simply, this involves identifying bacterial viruses that specifically target multi-resistant bacteria. Other research teams are attempting to modify existing antibiotics, such as penicillin, to make them effective against multi-resistant germs again.

The use of artificial intelligence has significantly accelerated research in all areas. "Thanks to AI, we can assess how promising a drug candidate is much earlier and with fewer experiments. This saves us resources, which we can then use specifically for the most promising substances,' says Müller.

Thanks to these developments, the fight against multi-resistant germs is slowly gaining momentum in favour of drugs. This is good news. However, as the new WHO report shows, this progress is urgently needed and comes at just the right time.

World Antimicrobial Resistance Awareness Week: The search for new weapons against bacterial infections