CROI 2024 Abstract eBook

Abstract eBook

Poster Abstracts

545

Predictors of CD4/CD8 T-Cell Inversion After 96 Weeks of ART Initiated During Acute HIV Robert Paul 1 , Kyu Cho 1 , Carlo P. Sacdalan 2 , Ferron F. Ocampo 2 , Phillip Chan 3 , Lydie Trautmann 4 , Julie Ake 5 , Somchai Sriplienchan 6 , Nathornsorn Poltubtim 6 , Jacob Bolzenius 1 , Sandhya Vasan 4 , Napapon Sailasuta 7 , Kilian Pohl 8 , Serena S. Spudich 3 , for the RV254/SEARCH 010 Study Team 1 University of Missouri St Louis, St Louis, MO, USA, 2 SEARCH Research Foundation, Bangkok, Thailand, 3 Yale University, New Haven, CT, USA, 4 US Military HIV Research Program, Silver Spring, MD, USA, 5 Walter Reed Army Institute of Research, Silver Spring, MD, USA, 6 SEARCH, Bangkok, Thailand, 7 University of Hawaii at Manoa, Honolulu, HI, USA, 8 Stanford University, Stanford, CA, USA Background: >50% of people with HIV (PWH) do not achieve CD4/CD8 T-cell ratio normalization (ratio>1.0) on antiretroviral therapy (ART). Recent findings from RV254/SEARCH 010, a longitudinal investigation of acute HIV infection (AHI) and long-term response to ART, identified a combination of immune and behavioral factors (e.g., mental health) that predicted CD4/CD8 T-cell ratio inversion at week 144 of ART. This study examined a larger array of potential predictors of CD4/CD8 T-cell inversion, including multimodal indices of brain structure/function quantified using 3T MRI before ART initiation. Methods: Archival data from RV254/SEARCH 010 were examined for individuals who completed neuroimaging (volumetrics, diffusion tensor imaging (DTI), and resting state connectivity) at enrollment followed by 144 weeks of sustained ART. Gradient boosted multivariate (GBM) regression with repeated cross validation was utilized to identify predictors for persistent CD4/CD8 T-cell ratio inversion at week 144 from baseline neuroimaging, demographic, immune, viral, cognitive, and mental/behavioral health indices. Results: 74 Thai males with an average duration of HIV infection of 19.5 (7.2) days were included in the analysis. Study participants were of 27.4 (6.0) years of age and had a median viral load (log 10 ) of 6.01 (IQR 5.43-6.79) copies/mL, CD4+ T-cell count of 330, CD8+ T-cell count of 526, and a CD4/CD8 T-cell ratio of 0.64. After 144 weeks of ART, all study participants were virally suppressed. 34 individuals had a CD4/CD8 T-cell ratio <1.0. The GBM analysis identified 10 baseline features that predicted CD4/CD8 T-cell ratio inversion at 144 weeks (average model performance of .72). The algorithm included larger volumes in 6 brain regions (medial superior frontal gyrus, superior temporal gyrus, rectus gyrus, nucleus accumbens, cerebellar lobes IV-V, pallidum), lower connectivity between the salience and visuospatial network, lower high-density lipoprotein (HDL) levels, lower radial diffusivity in the external capsule, and higher IL-1α. Conclusion: Perturbations in brain structure and function, possibly reflecting early inflammatory mechanisms, dyslipidemia, and inflammation prior to ART commenced in acute infection stratify individual risk for persistent CD4/ CD8 T-cell inversion following sustained use of ART. Further investigation of potential causal pathways is needed to establish mechanisms and therapeutic opportunities.

relationship to cognitive function in people with HIV (PWH) enrolled in the ACTG aging cohort study, HAILO. Methods: Clinical data and plasma were obtained from 550 PWH aged ≥45y with ≥2 neuropsychological tests (NP) and HIV RNA<200 copies/ml. Four NP scores (Trail Making Test Part A and B, Digit Symbol, and Hopkins Verbal Learning Test-Revised) were standardized as z-scores. The average of these z-scores (NPZ4) at plasma collection or subsequent visits were provided by HAILO and used as primary outcomes; longitudinal scores were summarized by calculating the slope for each participant. We used a single molecule array platform to quantify NFL and GFAP. Associations between BBMs and baseline NPZ4 scores or slopes were assessed with linear regression. Longitudinal analyses were done for the total cohort and a subset with NPZ4 decline (slope <0). Models were adjusted for age, sex at birth, race, and years of education; regressions of slopes were also adjusted for baseline NPZ4 and weighted by number of visits. Results: Mean age was 53y (range 45-77y), and observation time was 6.2y, 19.6% were female, and 28.4% non-Hispanic Black. Median [IQR] baseline NFL, GFAP, and NPZ4 scores were 10.7 pg/ml [7.98, 14.62], 77.5 pg/ml [58.80, 107.86], and 0.11 [-0.48, 0.75], respectively; median [IQR] NPZ4 slope was 0 [-0.06, 0.06]. In adjusted models, higher NFL or GFAP was associated with worse baseline NPZ4 (-0.05 per 10pg/ml NFL; p=0.06; -0.14 per 100 pg/ml GFAP; p=0.04) and became significant after excluding outliers NFL≥25 (p=0.03) but not GFAP≥200 (p=0.2). While baseline NFL (p=0.11) and GFAP (p=0.053) showed a weak relationship with NPZ4 slope in the total cohort, NFL had a stronger association in the subset with cognitive decline (n=263; -0.02 annual decrease per 10pg/ml NFL, p<0.01; Figure 1). The variance explained was ≤10% in all models. Conclusion: In this cohort of PWH, with the majority in mid-life and cognitively stable, NFL and GFAP were associated with a minor decrement in cross-sectional NPZ4, and NFL was weakly associated with longitudinal NPZ4. NFL more strongly predicted NPZ4 slope in a prespecified subset with cognitive decline. Overall, results highlight the limitations of NFL and GFAP as predictive biomarkers of cognitive decline in this age range.

Poster Abstracts

547

CD4 Count Predicts Longitudinal Trajectories of Central Executive Network Connectivity in PLWH Joshua M Schrock , Yufen Chen, Casey Armstrong, Robin Nusslock, Ann Ragin Northwestern University, Chicago, IL, USA Background: The brain's central executive network (CEN) plays a key role in the pathogenesis of depression, problematic substance use, and cognitive impairment, all of which are common comorbidities of HIV. Previous cross sectional research has reported that resting-state functional connectivity (rsFC) in the CEN is lower in people living with HIV (PLWH) compared to participants without HIV, suggesting reduced integrity of the CEN in PLWH. But it is unclear when these alterations emerge and what clinical markers predict their emergence. Methods: Resting state functional magnetic resonance imaging scans were conducted at baseline and two-year follow-up in a cohort (median age=29) of PLWH in early HIV infection (n=49) and healthy controls (n=20). A sub-sample (n=50, six female, 44 male) completed scans at both timepoints. Blood samples were collected to measure clinical immune markers and plasma HIV RNA.

546

Assessing Blood-Based Biomarkers as Predictors of Cognitive Decline in PWH: ACTG A5322 (HAILO) Shibani S Mukerji 1 , Petra Bachanová 2 , Linzy V. Rosen 2 , Hemi Park 1 , Rommi Kashlan 1 , Pia Kivisäkk 1 , Felicia C. Chow 3 , Kunling Wu 4 , Raha M. Dastgheyb 5 , Leah H. Rubin 5 , Katherine Tassiopoulos 4 , Robert A. Parker 1 , Emily P. Hyle 1 1 Harvard Medical School, Boston, MA, USA, 2 Massachusetts General Hospital, Boston, MA, USA, 3 University of California San Francisco, San Francisco, CA, USA, 4 Harvard TH Chan School of Public Health, Boston, MA, USA, 5 Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA Background: Blood-based markers (BBMs) show promise in evaluations of Alzheimer's disease and other cognitive disorders. We examined plasma neurofilament light chain (NFL) and glial fibrillary acid protein (GFAP) and their

CROI 2024 147