Target the Bug, Not the Whole Body: New Work Aims at Biofilms and Resistant Klebsiella

What the papers show

In a preprint, researchers describe nanoparticles that carry vancomycin directly to Staphylococcus aureus biofilms, using DNA aptamers specific to that bacterium to guide PLGA particles to the target. The particles reached 7.9% vancomycin loading, plus or minus 2.2%, with a biphasic release profile, and aptamer conjugation improved retention inside S. aureus biofilms in laboratory tests and in rat knee joint models compared with non-targeted controls. The point is not a stronger antibiotic. It is delivery that stays where the infection is, in a setting where biofilms on implants are hard to reach and hard to clear. m:…9855

A second preprint takes the same targeting logic to prevention. An mRNA-LNP vaccine aimed at OmpX, an outer membrane protein of Klebsiella pneumoniae, produced high levels of OmpX-specific antibodies and elevated interferon-gamma and interleukin-4 levels in mice, and serum from immunised mice promoted greater complement-mediated lysis of Klebsiella cells than control serum. Immunisation then gave 100% survival after a lethal Klebsiella challenge in the mouse model. This suggests a vaccine strategy that does not depend on matching one bacterial serotype after another, though the evidence so far is in mice, not in people. m:…9556

Why targeting matters now

A third preprint shows why a targeted map of what bacteria actually do during infection is useful. Using a metabolic labelling method called BONCAT together with conventional proteomics, researchers identified 152 infection-associated proteins in Acinetobacter baumannii during infection of human lung epithelial cells, with 117 of those proteins not detected in axenic reference proteomes. The infection-associated proteins were enriched for signal peptides and bacterial secretion system components. In other words, bacteria studied alone do not show the full repertoire they use when they meet host cells, so drug and diagnostic work based only on laboratory cultures may be aiming at the wrong list. m:…9040

Surveillance from a tertiary oncology centre in Kazakhstan explains the cost of getting this wrong. In that preprint, carbapenem-resistant Klebsiella pneumoniae was the dominant phenotype, peaking at 76.4% in 2024, and aggregate multidrug resistance rates of 76 to 87% were two to three times higher than the global pooled mean of 28.7% reported in the study. Amikacin resistance rose from 8.5% in 2022 to 38.1% in early 2026, with 60% resistance in intensive care units, leaving polymyxins and tigecycline as the only agents with retained activity at the centre as of 2026. That is the backdrop for targeted delivery and vaccination: when broad options narrow, specificity becomes the remaining lever. The same surveillance preprint also reported that Escherichia coli with extended-spectrum beta-lactamase and multidrug resistance emerged as the dominant pathogen in haematological and surgical departments by 2025 to 2026, which suggests the resistance problem is shifting across organisms as well as within Klebsiella. m:…1310

What to watch

None of this is a treatment yet. The nanoparticle work rests on laboratory and rat knee joint models, the vaccine work rests on a mouse model, and the Acinetobacter work maps proteins during cell infection rather than testing a therapy. The implication for readers who fund or run infection programmes is practical rather than clinical: prioritise platforms that can prove they reach the pathogen in its infection state, and surveillance that shows which resistant phenotypes dominate a specific centre, before assuming a broader antibiotic will cover the gap. The next informative step in each case is a test outside the original model, not a larger claim about the same model. m:…9855 m:…9556 m:…9040 m:…1310