Roche's new antibiotic takes on superbug resistant to nearly all drugs
The pharma giant is preparing to launch the final phase of trials for zosurabalpin against infections caused by carbapenem-resistant Acinetobacter baumannii. The arrival of this drug could close one of the most terrifying gaps in antibiotic resistance, where doctors have been forced to use toxic cocktails for years.
Doctors call it the "Iraq bacterium" — Acinetobacter baumannii returned from hot spots with wounded soldiers and firmly entrenched itself in ICUs worldwide. It is resistant to nearly everything in medicine's arsenal and kills 40 to 60 percent of patients with invasive infections. The last new class of antibiotics against gram-negative bacteria emerged more than half a century ago. Now Roche has announced: zosurabalpin is moving into the final, third phase of trials. The bet is the first fundamentally new weapon in 50 years against a superbug that the WHO and CDC have listed among the most dangerous threats to humanity.
An antibiotic found in a library of 45,000 molecules
The story of zosurabalpin began not with a stroke of genius but with methodical screening. Roche and Harvard scientists ran nearly 45,000 compounds through screening until they stumbled upon a tethered macrocyclic peptide — a molecule that hits where no one had aimed before.
Conventional antibiotics attack the bacterial cell wall, ribosomes, or DNA gyrase. Zosurabalpin does something fundamentally different: it blocks the transport of lipopolysaccharides from the inner membrane to the outer membrane of the bacterium. For Acinetobacter baumannii, this is a death sentence. Lipopolysaccharides, which are supposed to build the protective shell, accumulate inside the cell and become toxic. The bacterium dies not from destruction but from poisoning by its own building material.
Larry Tsai, Senior Vice President at Roche and Global Head of Immunology Product Development, explains the company's bet without mincing words: "Drug-resistant Acinetobacter are present in every country worldwide and disproportionately affect hospitalized patients, causing invasive infections such as pneumonia and sepsis."
From petri dish to ICU: what has already been proven
Preclinical data look promising. Zosurabalpin showed high efficacy against clinical isolates of CRAB in lab tests, as well as in mouse models with lung and thigh infections caused by pan-resistant strains. The first phase of human trials confirmed the drug's safety, tolerability, and pharmacokinetics — and, critically, existing resistance mechanisms do not affect it.
Now the most serious phase begins. In the third phase, Roche will enroll about 400 patients with invasive CRAB infections at over 100 centers in Europe, North and South America, and Asia. Half will receive zosurabalpin, half will receive standard therapy, which today consists of toxic cocktails of polymyxins, sulbactam, and tigecycline. The trial launch is scheduled for late 2025 or early 2026.
The context of this trial is grim. Mortality from invasive CRAB infections ranges from 40 to 60 percent. In intensive care units, where the bacterium feels at home, it causes ventilator-associated pneumonia and sepsis. A recent cohort study showed that mortality among patients with CRAB pneumonia was 53.8 percent versus 12.5 percent for those whose strains were still susceptible to carbapenems. The average hospital stay is 18 days, and the average bill is nearly $7,000. And that's in the best case. With total drug resistance, time is measured in days, and treatment options simply run out.
Narrow spectrum as an advantage
Zosurabalpin is a narrow-spectrum antibiotic. It targets Acinetobacter specifically and leaves other bacteria alone. In an era when doctors dream of "broad-spectrum super-antibiotics," this sounds like a drawback. But Roche presents it as an advantage.
The logic is ironclad. Broad-spectrum antibiotics decimate the microbiome, trigger Clostridium difficile, and breed new resistance mechanisms. A narrow spectrum means the drug hits only the intended target, leaving the rest of the flora undisturbed. But it also means a modest market. CRAB infections are not millions of patients with community-acquired pneumonia but tens of thousands of severe hospital cases. For Big Pharma, accustomed to blockbusters, this has always been an economic puzzle.
Who wins, who loses, and why Merck paid $9.5 billion
The antibiotic market is experiencing a strange renaissance. Not long ago, major players were leaving the sector en masse — development was too expensive, treatment courses too short, generics too cheap. Antibiotics didn't pay off. But something has changed.
In December, Merck bought Cubist for $9.5 billion. Roche, alongside zosurabalpin, struck a deal with Meiji Seika Pharma and Fedora to develop the beta-lactamase inhibitor OP0595 — the potential value of the agreement reaches $750 million. And Roche's portfolio also includes RG6436, a LepB inhibitor undergoing early clinical trials against carbapenem-resistant gram-negative infections.
The driver of this revival is not philanthropy but cold calculation. In the US alone, over two million cases of drug-resistant infections are reported annually. Direct healthcare costs reach $20 billion. By 2050, according to some forecasts, superbugs could kill up to 10 million people a year — as many as cancer does today. Governments have begun introducing incentives: accelerated approval, extended patent protection, purchase guarantees. The economics are slowly turning around.
The biggest winners will be ICUs. Doctors who for years have been forced to administer polymyxins — drugs with horrific nephrotoxicity, developed back in the 1940s and abandoned precisely because of toxicity — will have an alternative. Immunocompromised patients, whom Acinetobacter hunts most ruthlessly, will benefit. Roche will win by staking out a territory where competitors can be counted on one hand.
Those who still hope that old antibiotics will last another couple of decades lose. And insurance companies lose, as they will have to pay for innovative drugs instead of dirt-cheap colistin. But they seem to have no choice.
Concrete forecast: 2030 and beyond
Zosurabalpin will not hit the market tomorrow. The third phase of trials will start around 2025–2026 and take several years. Roche plans to file for approval before 2030. That means doctors will get the drug in about four to five years — if efficacy data hold up and regulators don't drag out the review.
In parallel, work is underway on derivatives of zosurabalpin. A study published in April 2025 showed that modified versions of the molecule could also work against Klebsiella pneumoniae and potentially Shigella flexneri. If Roche can broaden the spectrum while preserving the mechanism of action, this would be not just one drug but an entire platform — like penicillins or cephalosporins in their day.
A separate storyline is competition. While zosurabalpin moves toward the finish line, doctors are squeezing the most out of existing drug cocktails. A new study published in Open Forum Infectious Diseases in January 2026 showed that the combination of polymyxin B with meropenem and sulbactam produces a 100 percent synergistic effect in vitro. It's a temporary solution — but it helps bridge the gap until the new drug class arrives.
And the most important forecast: zosurabalpin is not the end of the war against superbugs but a reconnaissance in force. It proves that new targets can be found in the seemingly impenetrable armor of gram-negative bacteria. If this approach works, the next candidates will come faster — and perhaps in ten years, the phrase "drug-resistant infection" will no longer sound like a death sentence.
— Editorial Team
