CROI 2018 Abstract eBook

Abstract eBook

Poster Abstracts

Results: When the last time point (pre-euthanasia) was compared to pre- inoculation baseline, [18F]-DPA714 total binding was decreased in 4 animals (Fig. 1A) and increased in one (Fig.1D). We found an inverse relationship between binding and CSF VL: binding decreased when VL increased and vice versa (Fig. 1B and E). Pathology results in 4 of the animals compared to 3 uninfected monkey brains showed unchanged or mildly decreased Iba1 staining in the background compared to controls (Fig.1C) but with increased diffuse CC3/PARP staining. There was however evidence of multiple microglial nodule with increased Iba1 and CC3/PARP staining. The only animal showing increased binding on PET at the last time point (Fig.1D and F) showed a combination of diffuse and focal microglial activation on IF staining. Conclusion: In SIV encephalitis macaques, we found an inverse relationship between CSF VL and microglial activation by imaging and IF staining with microglial loss/apoptosis associated with very high CSF VL and microglial activation associated with lower CSF VL. Apoptosis staining co-localized with microglial and neuronal stains suggesting cellular death/loss in association with high VL. Our results provide a new insight into the role/status of microglia/ macrophages in early stages of the disease when very high CSF viral loads are observed (Feibig stages II-IV).

from one HIV+ participant is shown (Figure). Cluster analysis (a) shows 5 distinct groups of cells, and heat map (b) shows differentially expressed genes that identify the clusters as CD8 T cells (34%), CD4 T cells (23%), NK cells (15%), and monocytes (15%). Conclusion: We successfully employ scRNA-seq to classify CSF immune cells. The depth and breadth of gene expression profiles obtained with single-cell resolution will allow us to identify unique CNS cell populations associated with persistent CNS immune activation in ART-suppressed HIV.

Poster Abstracts

443 HIV-RNA IN THE OLFACTORY MUCOSA OF HIV-POSITIVE PATIENTS Mattia Trunfio 1 , Tiziano Allice 1 , Luca Bertero 1 , Daniele Imperiale 2 , Sebastiano Catera 2 , Veronica Pirriatore 1 , Giovanni Di Perri 1 , Maria Enrica Amasio 2 , Paola Cassoni 1 , Valeria Ghisetti 3 , Stefano Bonora 1 , Andrea Calcagno 1 1 University of Torino, Torino, Italy, 2 Maria Vittoria Hospital, Torino, Italy, 3 Amedeo di Savoia Hospital, Torino, Italy Background: Central nervous system (CNS) HIV infection is assessed on cerebrospinal fluid (CSF) through lumbar puncture (LP), although CSF represents a proxy of brain tissue and LP is a non risk-free procedure. Since olfactory mucosa (OM) is the only CNS tissue that is easily accessible, nasal brushing (NB) may represent an alternative safe diagnostic technique to evaluate CNS viral compartmentalization Methods: Naive and on cART HIV-positive patients undergoing LP for clinical reasons were included. After informed consent, patients underwent (<72 hours from the LP) a NB; OM swabs were inserted in Copan UTM viral transport medium. CSF and plasma HIV-RNA (plVL) were quantified with CAP/CTM v.2.0 HIV-1 (Roche Molecular, USA, detection limit 20 copies/mL). OM HIV-RNA was measured with a modified CAP/CTM procedure. Immunovirological data and CSF biomarkers [ELISA: tau, ptau, neopterin, S100β; immunoturbidimetric method: Reiber diagrams] were recorded. Data are reported as medians (interquartile ranges) and analyzed through non-parametric tests Results: 47 patients were included (74.5%male, 85.1% Caucasian, median age 51 years [47-57]). 19 patients were naïve (plVL and CSF HIV-RNA 5.2 and 3.3 Log 10 copies/mL; current CD4 44 cell/uL), while 28 on cART (plVL and CSF HIV-RNA undetectable and 1.3 Log 10 copies/mL; current and nadir CD4 456 and 177 cell/uL). CSF escape (CSFE) was observed in 5 patients (10.6%). Mild discomfort, sneezing and lacrimation were the only reported side effects for NB. OM HIV-RNA was detectable in 23 patients (17 naïve: 3.1 Log 10 copies/mL [2.0- 3.7]; 6 on cART: 1.9 Log 10 copies/mL [1.6-2.1]) and correlated with current CD4 count (ρ -.509, p<.01), Log 10 plVL and CSF HIV-RNA (ρ .746, p<.01 and ρ .663, p<.01), PBMC HIV-DNA (ρ .582, p<.01) and CSF neopterin (ρ .671, p<.01). OM HIV-RNA was higher than plVL in 5 treated patients (4 aviremic including 2 CSFE) and than CSF HIV-RNA in 5 naïve and 4 treated patients (3 aviremic including 1 CSFE). An OM escape (OM HIV-RNA more than 1 Log 10 above plVL) was associated with higher risk of CSFE (OR 12.7, p=.01) Conclusion: Non-invasive monitoring of HIV tissue replication may have major implications; OM HIV-RNA has been safely measured in around half NB samples. The reported association with HIV-DNA and neopterin is promising as a marker of residual viral burden and immune activation. Further studies are required to establish whether OM HIV-RNA may be a reliable surrogate of CSF/brain tissue HIV-RNA

442 SINGLE-CELL RNA SEQUENCING TO CHARACTERIZE CSF IN VIROLOGICALLY SUPPRESSED HIV

Shelli Farhadian 1 , Chryssa Zografu 1 , Sameet Mehta 1 , Jennifer Chiarella 1 , Kevin Robertson 2 , Sarah Yosief 2 , Jenna Pappalardo 1 , David Hafler 1 , Serena S. Spudich 1 1 Yale University, New Haven, CT, USA, 2 University of North Carolina Chapel Hill, Chapel Hill, NC, USA Background: The cellular mechanisms underlying persistent central nervous system (CNS) immune activation in HIV during antiretroviral therapy (ART) are incompletely understood. Advances in single-cell RNA sequencing (scRNA- seq) permit simultaneous examination of thousands cellular transcriptomes, allowing for a previously unrealized level of cellular resolution. Here we develop the first successful application of massively parallel scRNA-seq to cerebrospinal fluid (CSF) to profile thousands of single cell transcripts from CNS tissue. Methods: CSF and blood sampling and neuropsychological testing (total Z scores) were performed in a pilot study of HIV+ (n=4) individuals on ART with plasma viral load <20 cps/mL for >1 yr, and HIV- controls (n=3). Fresh cells were isolated from 25cc CSF on the day of collection, then ~10,000 cells were applied to nanowell arrays designed for single isolation and preloaded with uniquely barcoded beads. Cells were lysed, mRNAs were reverse transcribed, and single cell libraries prepared with NexteraXT. ~5000 cells/sample were sequenced (Illumina; average depth 60,000 reads/cell). Single-cell transcripts were analyzed with Seurat R package for identification of highly variable genes, dimensionality reduction, and unsupervised clustering. Results: The mean age in HIV+ and controls was 49 years, with total Z scores -0.87 and -0.96, and CD4 594 cells/ul and 860 cells/ul, respectively. 2 participants had HIV detected in CSF (38 and 99 cps/mL). For all samples, we successfully produced single cell libraries with average cDNA size 400-700bp. 2,000 cells per sample were randomly selected for further computational analysis, and filtered for high quality cells (>200 and <2500 unique transcripts and <10%mitochondrial RNA). This yielded a mean 802 cells per sample for further analysis, and 3-5 clusters per sample. A representative analysis of CSF

CROI 2018 157