Impaired blood-brain barrier integrity is associated with intrathecal immunoactivation and signs of neuronal injury in untreated and treated HIV

Birgitta Anesten1, Lars Hagberg1, Henrik Zetterberg2, Staffan Nilsson3, Bruce Brew4, Dietmar Fuchs5, Richard W Price6, Magnus Gisslén1.
Affiliates: 1Dpt of Infectious Diseases, 2Dpt of Psychiatry and Neurochemistry, University of Gothenburg, 3Mathematical Sciences, Chalmers University of Technology, Gothenburg, Sweden, 4Dpt of Neurology, University of New South Wales, Sydney, Australia, 5Division of Biological Chemistry, Biocenter, Innsbruck Medical University, Innsbruck, Austria, 6Department of Neurology, University of California San Francisco, San Francisco, California, USA

View poster-presentation

Background
HIV infects the central nervous system and give rise to a chronic infection and immune activation in the brain. Blood-brain barrier (BBB) damage is prevalent, especially in patients with severe neurocognitive impairment and HIV-associated dementia. The aim of this study was to assess the prevalence of BBB disruption in different stages of HIV-infection and to study risk factors and potential association with neuronal injury.

Methods
Using a cross-sectional design and archived CSF samples from three cohorts (Gothenburg, Sweden; San Francisco, USA; Sydney, Australia), we compared BBB integrity as measured by CSF/plasma albumin ratio with cerebrospinal fluid (CSF) and blood concentrations of HIV RNA and neopterin, CSF WBC count, and CSF neurofilament light chain protein (NFL) in 7 HIV-infected groups (n =657 ): four groups of untreated ‘neuroasymtpomatic’ (NA) subjects without HIV-associated dementia (HAD) defined by blood CD4+ T cells of >350, 200-349, 50-199, and 6 months of viral suppression to <50 copies/mL of HIV RNA in blood. An HIV-uninfected control group was also examined (n=53). Continuous variables were log10 transformed where appropriate.

Results
The albumin ratio was significantly increased in HAD as compared to all other groups with highest levels in those with more severe dementia. 68 % of patients with HAD had increased albumin ratio compared to 16 % of NA subjects without treatment. No significant difference in albumin ratio was found between NA untreated subjects, irrespective of CD4 cell count strata, subjects on ART, and HIV-negative controls. Significant correlations between BBB integrity and CSF WBC count (p<0.01, rs=0.16), CSF HIV RNA levels (p<0.01, rs =0.14), serum (p<0.001, rs =0.20) and CSF (p<0.001, rs =0.25) neopterin, and age (p<0.001, rs=0.24) were found in NA untreated subjects. In a multiple linear regression analysis only age and CSF neopterin stood out as independent predictors to albumin ratio with adjusted estimates in the multivariate analysis. Furthermore, a significant correlation was found between albumin ratios and CSF NFL concentrations in untreated patients (p<0.001, rs=0.40) as wells as in patients on ART (p<0.001, rs =0.57). Albumin ratio was confirmed as an independent predictor of CSF NFL together with age, CD4 cell count and CSF WBC count.

Conclusion
BBB disruption is normally a late phenomenon and mainly found in patients with HAD. Neither untreated neuroasymtpomatic subjects, nor patients on antiretroviral treatment had albumin ratios that was significantly different from HIV-negative controls. 16 % of untreated asymptomatic HIV-infected subjects had signs of BBB dysfunction but without relationship to the CD4 cell count. Intrathecal immunoactivation as measured by CSF neopterin is a risk factor of BBB impairment and the association with CSF NFL suggest a pathogenic pathway between inflammation, BBB disruption and neural injury before development of dementia.