The study, led by Alexandra McCarron and colleagues at Adelaide University and the Women’s and Children’s Hospital in South Australia and published in Gene Therapy, set out to answer a question that has real clinical consequences. Cystic fibrosis is caused by mutations in the CFTR gene, which encodes a chloride channel on the surface of airway cells. Highly effective modulator drugs have transformed care for many patients, but roughly ten percent of people with cystic fibrosis carry nonsense variants that produce no CFTR protein at all, making pharmaceutical rescue impossible. Others cannot tolerate the drugs or derive no measurable benefit. For these patients, gene-addition therapy, delivering functional CFTR genetic material directly to the airways, remains one of the few credible paths to slowing lung disease.

Clinical translation, however, has repeatedly disappointed. Every cystic fibrosis gene therapy product tested in humans to date has stalled at Phase I or II trials, failing to meet predefined efficacy endpoints despite strong preclinical performance. The researchers argue that a key reason may be the inadequacy of the preclinical models themselves. Commonly used rodent models of cystic fibrosis do not develop the severe, muco-obstructive, infected lung disease seen in patients, so candidate therapies are routinely validated under conditions that bear little resemblance to the target organ. Infections are a particularly glaring omission, since chronic Pseudomonas aeruginosa colonization, embedded in biofilms within viscous sputum, defines advanced cystic fibrosis lung disease and is strongly linked to hospitalization and early death.

Earlier work supports the concern that infection matters. Adenoviral vector transfer was halved in Pseudomonas-infected mice, and adeno-associated virus transfer was impeded in mice infected with either Bordetella bronchiseptica or Pseudomonas. Yet the picture is not uniform: in one study, the liposomal reagent lipofectamine was completely blocked by infection while a VSV-G pseudotyped lentiviral vector and PEI/DNA polyplexes retained their efficacy. The lesson, the researchers note, is that every vector, pseudotype and serotype must be tested individually under infected-lung conditions.

To do this properly, the team exploited two tools they had developed previously. The first was a pair of cystic fibrosis rat models, generated by CRISPR/Cas9 gene editing, which unlike mice display many features of the human disease but, importantly, do not spontaneously develop lung infections. The second was a method for establishing a localized, chronic Pseudomonas infection by delivering bacteria embedded in agarose beads directly into the right main bronchus through a miniature bronchoscope. The beads immobilize the bacteria and prevent their clearance, producing a persistent infection with robust inflammation and tissue remodeling, while confining the disease to one lung to minimize mortality risk.

The experimental design was a 2 x 2 factorial study: wild-type or CFTR knockout rats, each receiving either Pseudomonas-laden or sterile beads, with ten animals per group. Seven days after infection was established, the researchers delivered a lentiviral vector pseudotyped with the vesicular stomatitis virus glycoprotein, or VSV-G, carrying a luciferase reporter gene under the EF1α promoter. This vector platform has appealing properties for lung therapy: it transduces dividing and non-dividing cells, integrates stably into the host genome for long-term expression, and elicits comparatively little immune response. Because VSV-G enters cells through the low-density lipoprotein receptor, which sits on the basolateral surface of airway epithelial cells behind tight junctions, the team first applied a small dose of lysophosphatidylcholine, a lipid that transiently opens those junctions and gives the vector access to the cell layer.

Bioluminescence imaging at seven days after vector delivery produced the study’s first key finding: there was no significant difference in gene expression between infected and non-infected animals, regardless of genotype. The inflamed, bacteria-colonized lung had not impeded the vector at all. Histology confirmed that the infection was doing real damage, with widespread alveolar consolidation, mucus plugging of airways, goblet cell hyperplasia, perivascular inflammatory cuffs and proliferation of bronchus-associated lymphoid tissue in both wild-type and CF rats, yet the vector still got through.

The second imaging time point delivered the surprise. By day 14, bioluminescence in the sterile-bead control animals had fallen 136-fold from its day 7 level, whereas in infected animals the decline was only about seven-fold. A statistically significant interaction between infection status and time confirmed that infection genuinely attenuated the loss of transgene expression. The researchers propose a plausible mechanism: infection injures the airway epithelium, further disrupting tight junctions beyond what the lysophosphatidylcholine achieves, and thereby grants the vector enhanced access to the basal stem cell niche. Because injured lungs actively repair themselves, gene-modified progenitor cells may then proliferate and differentiate to replace damaged cells, carrying the transgene into daughter cells more effectively than in the quiescent, intact airways of healthy controls. Notably, the magnitude of the decline did not correlate with bacterial load, suggesting the infected environment itself, rather than the number of bacteria, modifies expression persistence.

The findings also carry reassurance on safety. Critics have worried that opening tight junctions in an infected lung could let bacteria escape into the bloodstream, but spleen cultures showed no dissemination, no procedure-related deaths occurred, and the animals tolerated repeated bronchoscopic interventions well. This is also the first study to use a cystic fibrosis animal model to test how infection affects gene transfer. CF rats carried about forty times more bacteria than wild-type rats on average, an unsurprising observation given that compromised mucociliary clearance and defective CFTR-mediated bacterial uptake are hallmarks of the disease, though high variability kept the difference short of statistical significance.

The study has caveats the authors acknowledge candidly. The agarose bead model, while producing authentic inflammation and remodeling, does not replicate the polymicrobial, true-biofilm infections of human cystic fibrosis lungs, and it may be severe enough to mask genotype-dependent differences. Only one vector, one luciferase reporter and one time window were tested, and luciferase, an intracellular enzyme, is a simpler protein than CFTR, whose folding, trafficking and membrane insertion could be far more vulnerable to infection-induced cellular stress. Follow-up beyond day 14 would establish whether the infection-induced advantage endures. Even so, the message is consequential: chronic bacterial infection, long viewed as a fundamental barrier, does not impair lentiviral gene transfer and may even help expression persist. The real obstacles to clinical success in cystic fibrosis gene therapy may lie elsewhere, and each candidate vector, particularly for the modulator-ineligible patients most likely to have advanced lung disease, deserves evaluation under infected-lung conditions before it reaches the clinic.

Subject of Research: Effect of chronic Pseudomonas aeruginosa lung infection on lentiviral vector gene transfer in cystic fibrosis rat models

Article Title: Chronic lung infection does not impair lentiviral vector gene transfer in a cystic fibrosis rat model

Article References: McCarron, A., Reyne, N., Boog, B., Cmielewski, P., & Donnelley, M. (2026). Chronic lung infection does not impair lentiviral vector gene transfer in a cystic fibrosis rat model. Gene Therapy. https://doi.org/10.1038/s41434-026-00652-y

Image Credits: AI Generated

DOI: 10.1038/s41434-026-00652-y

Keywords: cystic fibrosis, gene therapy, lentiviral vector, Pseudomonas aeruginosa, CFTR, VSV-G, lung infection, bioluminescence imaging, airway epithelium, preclinical models, gene transfer, inflammation