People with HIV who had cognitive impairment carried blood-protein sugar chains depleted of galactose and sialic acid, and blocking the enzymes that strip those sugars prevented inflammation and memory deficits in two mouse models, according to a study published on 5 June 2026 in Med, a Cell Press journal. The work, led by Leila B. Giron and Mohamed Abdel-Mohsen of Northwestern University Feinberg School of Medicine with the AIDS Clinical Trials Group (ACTG) and funded by the National Institutes of Health, was carried forward as a poster at AIDS 2026 in Rio de Janeiro on 29 July, according to ACTG's own announcement that day.
The human part of the study is observational and small. The team analysed IgG and plasma N-glycans from a longitudinal subset of 40 people on antiretroviral therapy in the ACTG HAILO cohort (A5322), sampled over eight years and matched by age, sex and ethnicity, then an independent cross-sectional subset of 80 people from the same cohort, 40 women and 40 men, half of each with cognitive impairment. Pro-inflammatory agalactosylated and hypo-sialylated structures, including the glycan known as G0F, were higher in those with impairment and correlated with lower composite NPZ4 neuropsychological scores. In the validation subset, the differences were driven by the women; they were not seen among the men. Association is not proof of cause, and the paper notes that cognitive impairment affects at least 24 per cent of people with HIV despite suppressive treatment.
The causal claims rest on animals. In bone marrow-liver-thymus humanised mice infected with a transmitted-founder HIV strain, a combination of oseltamivir (marketed as Tamiflu) and the experimental broad-spectrum inhibitor DANA, each at 12.5 mg/kg, reduced HIV-associated inflammation, immune activation and accelerated epigenetic aging. Because that model is poorly suited to behaviour, the group turned to C57BL/6 mice infected with EcoHIV, a chimeric virus, giving the same drug pair daily by the intranasal route from day 25 after infection and testing spatial learning and working memory in the radial arm water maze from day 28. Across four independent experiments (one in female animals, three in male), treatment prevented the infection-related deficits. In cell assays, neither drug affected HIV entry, and oseltamivir alone did not preserve sialic acid.
The authors are explicit that this is not a treatment recommendation. "We are not saying yet that people should take flu drugs to prevent cognitive decline," Abdel-Mohsen said in a Northwestern release issued the day of publication, adding that these drugs have not been tested for this purpose, dose or duration. No clinical trial of sialidase inhibitors for cognition in people with HIV has been registered on the strength of this work.