Can we stop antibiotic resistance in ruining promising drugs before they reach patients?

What if we could predict in early development, which antibiotics bacteria are most likely to become resistant to? In a review published in Nature Reviews Microbiology, researchers from the HUN-REN Biological Research Centre, Szeged (HUN-REN BRC Szeged) and the Hungarian Centre of Excellence for Molecular Medicine (HCEMM) in Szeged propose a five-step decision-making framework that could help identify resistance risks early, refine promising drug candidates and prevent costly failures.

A new antibiotic may perform exceptionally well in the laboratory, only to fail during clinical trials as bacteria evolve resistance to it.

“Drug candidates are assessed for efficacy and safety early in development, but how readily bacteria can evolve resistance is often evaluated much later, when there is nearly no room to make changes,” explains Petra Szili, a researcher at the HUN-REN Biological Research Centre, Szeged (HUN-REN BRC Szeged).

But how can we account for evolution?

The review's authors, Csaba Pál, Petra Szili and Szilvia Juhász, propose a five-step decision-making framework. Beyond assessing how readily resistance can emerge, it examines whether resistance persists in the absence of antibiotics, makes pathogens more dangerous, reduces the effectiveness of other drugs or involves resistance genes with the potential to spread. The framework groups antibiotic candidates into five main risk categories, helping developers decide which compounds to prioritise, redesign or pursue for more targeted applications.

Crucially, the aim is not simply to rule out high-risk candidates. By identifying the genetic changes that allow bacteria to escape a drug's effects, researchers could also inform the redesign of the molecule itself. The goal is to develop antibiotics against which resistance is more difficult to evolve.

Many of the necessary tools already exist, from laboratory evolution experiments and genetic analyses to methods for identifying resistance genes. In a 2025 study published in Nature Microbiology, researchers in Szeged investigated 13 recently approved or in-development antibiotics. Potential routes to resistance were identified for almost all of them, but these varied considerably between bacterial species. In other words, studying a single bacterial strain cannot tell us everything about an antibiotic's future.

The importance of identifying these risks early is illustrated by the case of GSK2251052. Despite its initial promise, resistance emerged during phase II clinical trials, prompting the discontinuation of its development for complicated urinary tract and systemic Gram-negative infections in 2012. The compound has since been investigated again against non-tuberculous mycobacteria. Resistance, therefore, does not necessarily spell the end for a molecule, but it may fundamentally change the direction of its development.

There is also a less widely recognised risk: resistance can sometimes make bacteria more dangerous rather than weaker. Certain changes to the bacterial cell surface, including those associated with resistance to colistin and other membrane-targeting agents, can also alter how pathogens interact with the immune system.

“We need to investigate not only whether bacteria survive an antibiotic, but also how resistance may change their ability to damage host tissues. Some resistance mechanisms can make pathogens more harmful to the host,” emphasises Szilvia Juhász, a researcher at HCEMM.

The Szeged researchers are now translating their earlier experimental findings into a framework for drug development. By bringing together laboratory evolution, genetic studies and techniques for investigating the spread of resistance genes, they propose a practical approach to identifying vulnerabilities in drug candidates before they enter costly clinical trials.

“Pharmaceutical companies invest years and substantial resources in developing new antibiotics. If research can reveal a molecule's evolutionary vulnerabilities early on, those findings deserve attention. Acting on them could save time and resources, while improving our chances of delivering treatments that remain effective in the long term,” says Csaba Pál, senior author of the review.

The authors hope that predicting resistance will eventually become as routine a part of antibiotic development as assessing efficacy and safety. The antibiotics of tomorrow must be designed not only to tackle the bacteria we face today, but also to anticipate how those bacteria may evolve.