CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

p<.05; VCAM, p<.01; PECAM, p<.001) and post-synaptic neuronal integrity (PSD95, p<.0001; synaptophysin, ns) in cortical regions, but not in basal ganglia. Additionally, within HIV+ subjects: i) HO-1 RNA is higher (p<.05), and PSD95 is lower (p<.0001) in basal ganglia vs. cortical regions, consistent with our recent macaque findings; ii) endothelial adhesion molecules are higher in basal ganglia vs. cortical regions (ICAM-1 p<.01, VCAM p<.01, PECAM, ns); and iii) PSD95 is lower in basal ganglia, but not cortical regions, in HIV+ vs HIV- subjects (p<.01). Conclusion: We identified regional human brain variation in the anti-oxidant response (HO-1 RNA) that links it to post-synaptic neuronal injury and increased endothelial adhesion molecule expression. This suggests greater vulnerability of the basal ganglia (vs cortical regions) to HIV neuronal injury, possibly due to increased drive for transendothelial immune cell adhesion and migration. Such endothelial cell function may be regionally regulated by HO-1 expression. 432 PLATELET-ENDOTHELIAL INTERACTIONS MAY PROTECT AGAINST VIRAL ENTRY IN THE BRAIN Claire E. Lyons 1 , Hannah Schneider 1 , Elizabeth L. Engle 1 , Kevin M. Najarro 1 , Suzanne E. Queen 2 , Craig N. Morrell 3 , Joseph Mankowski 2 , Kelly A. Metcalf Pate 2 1 Johns Hopkins University, Baltimore, MD, USA, 2 Johns Hopkins University School of Medicine, Baltimore, MD, USA, 3 University of Rochester, Rochester, NY, USA Background: The brain is an important sanctuary site for HIV and these viral reservoirs are an important barrier to cure. In the SIV-infected macaque model, perivascular infiltrates of infected cells are characteristic of central nervous system (CNS) disease. Platelet decline occurs due to multiple mechanisms in SIV, and is associated with increased risk for the development of CNS disease. We sought to determine if activated platelet-endothelial interactions contribute to platelet decline and are associated with these infected perivascular infiltrates in the SIV-infected pigtailed macaque, and to define how these interactions affect the blood brain barrier. Methods: Platelet activation was monitored throughout infection for SIV- infected macaques and uninfected controls. Brains were evaluated to determine CNS disease status and for immunohistochemistry for platelet-endoethelial binding and perivascular macrophage cuffs. Confluent monolayers of brain microvascular endothelial cells (BMECs) were exposed to washed platelets or media in transwells and permeability quantified. Results: Platelets harvested from infected macaques that went on to develop CNS disease during terminal infection demonstrated less activation than macaques without CNS disease during acute (P = 0.04) and asymptomatic (P < 0.0001) infection. Brains from SIV-infected macaques were more likely than brains from uninfected controls to have platelets bound to vascular endothelium during acute (RR 4.0, P=0.03) and terminal (RR 3.6, P=0.04) infection. 6 of the 10 SIV+macaques had CNS disease during terminal infection, and resident Mac387+ (RR 3.4, P=0.0001) or CD163+macrophages (RR 1.44, P=0.0005) but not non-resident CD68+macrophages (RR 1.2, P=0.2) were observed with increased likelihood around platelet-lined vessels in these animals. SIV-infected macrophages were similarly observed with increased likelihood around platelet- lined vessels (RR 1.5, P=0.007). Permeability of BMECs decreased two-fold following incubation with platelets from SIV infected macaques compared with uninfected macaques (P=0.01). Conclusion: Activated platelet-endothelial interactions may represent a protective mechanism against development of infected macrophage infiltrates in CNS disease that is removed in the context of HIV-associated thrombocytopenia. 433 HIV-1 INDUCED NEUROPATHOLOGY OF A HUMANIZED MICROGLIAL MOUSE Saumi Mathews , Edward Makarov, Howard E. Gendelman, Larisa Y. Poluektova, Santhi Gorantla University of Nebraska Medical Center, Omaha, NE, USA Background: Studies on HIV central nervous system infection and brain viral reservoirs have been hampered with the dearth of small animal models. Immune deficient mice reconstituted with human immune system are susceptible to HIV infection and have proven to be potent tools to study HIV pathogenesis, prevention and therapeutic development. However, any reflection of CNS has been hampered by a lack of human glial cells in any currently available rodent models. Human perivascular macrophages, microglia and astrocyte need be present to mimic brain viral reservoirs and virus-induced neuropathogenesis. To this end, we developed a new immunodeficient

strain supplemented with human interleukin-34 (IL-34) transgene to support microglial development in humanized mice. These human microglial mice were used to study HIV-1 induced neuropathogenesis. Methods: Human CD34+ hematopoietic stem cells transplanted NOD.Cg- PrkdcscidIl2rgtm1SugTg(CMV-IL34)1/Jic (CD34-NOG-hIL-34)mice developed human “microglia like” in the presence of tissue specific ligand-IL-34. To identify relationships between HIV-1 infection of microglia and neuropathology, mice at 6 months of age were infected with HIV-1ADA and brain tissues were subjected for histopathologic (glial and neuronal) and transcriptomic (mouse and human) profiling by next generation sequencing. Results: CD34-NOG-hIL-34 mice showed sustained plasma viremia with CD4+T cell loss and productive human microglial infection. Reductions in neuronal and synaptic architectures was observed in brain subregions by reduced expression of synaptophysin, MAP2 and neurofilament-H. Astrogliosis and microgliosis were evident. Molecular profiling of these infected brains revealed a significant increase in human genes such as IFIT1-5, ISG15, MX2, OAS1 pertaining to interferon signaling. The other pathways which were upregulated included toll- like receptor and pattern-recognition receptor indicating activation of innate immune response and increased inflammation. Whilst analysis of mouse genes indicated that ERK, integrin, MAPK, apoptosis signaling etc. were differentially regulated in association with neurodegeneration. Conclusion: Human microglial mouse closely reflects the pathobiology of HIV-1 infection with astrogliosis, microgliosis, productive viral infection of microglia, synaptic alterations and inflammatory responses. This model will prove useful for studies of neural-glial cross talk and studies designed to locate and eliminate the viral reservoir. 434 MTDNA HAPLOGROUP B IS ASSOCIATED WITH EXECUTIVE FUNCTION AND RECALL IN HISPANIC PLWH Karen E. Volpe 1 , Todd Hulgan 1 , David C. Samuels 1 , Asha R. Kallianpur 2 , Ronald J. Ellis 3 , Donald Franklin 3 , Scott L. Letendre 3 , Robert K. Heaton 3 1 Vanderbilt University, Nashville, TN, USA, 2 Cleveland Clinic Lerner College of Medicine, Cleveland, OH, USA, 3 University of California San Diego, San Diego, CA, USA Background: Neurocognitive impairment (NCI) is common among persons living with HIV (PLWH) and impacts quality of life and functional status. Mitochondrial DNA (mtDNA) haplogroups are associated with HIV-related outcomes. NCI appears to be more frequent in persons of Hispanic ethnicity with HIV. We previously observed less NCI in Hispanic-ancestry PLWH having mtDNA haplogroup B. The current analysis focused on domains of NC performance in this population. Methods: CHARTER is a prospective observational study of neurologic outcomes of PLWH. Comprehensive neuromedical, neurocognitive (NC), and laboratory assessments were performed. Haplogroups were assigned using mtDNA sequence data. Outcomes were age-, sex-, and education-corrected domain deficit (DDS) and mean T scores (TS), and impairment status by clinical rating score for seven NC domains: motor, working memory, verbal, executive function, processing speed, learning and delayed recall. Race/ethnicity- stratified analyses of baseline domain scores/status included univariate Wilcoxon rank sum and Chi2 tests. Linear regression models adjusted for nadir CD4 cell count, neurocognitive comorbidity status, and current ART use. Results: Major haplogroups and NC domain data were available in 1016 persons (median age 43 years, 23% female, 478 [47%] non-Hispanic black, 440 [43%] non-Hispanic white, and 98 [10%] Hispanic). Major haplogroups were not independently associated with NC domains among non-Hispanic black or white PLWH. Among 98 Hispanic PLWH, mean executive function, learning, and recall TS were higher and DDS lower (less impaired) among 17 with haplogroup B compared to other haplogroups (Wilcoxon p<0.05 for all comparisons). Three Hispanic PLWH (18%) with haplogroup B had impaired delayed recall vs. 34 (42%) with other haplogroups (Chi2 p=0.06). With adjustment for the covariates above, haplogroup B was no longer associated with learning DDS and TS (p=0.1 and 0.2) or recall DDS (p=0.08), but remained significantly associated with executive function DDS and TS (p≤0.05 for both), and recall TS (p=0.03). Other domains did not differ by haplogroup B status. Conclusion: Previously identified NC differences in Hispanic CHARTER participants with mtDNA haplogroup B were greatest for the delayed recall and executive function domains. If validated in independent cohorts, this finding could inform neuroimaging and other assessments to define mechanisms by which mtDNA variation may influence NC performance in PLWH.

Poster Abstracts

CROI 2019 157