CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

We hypothesize that one-year survival after a diagnosis of DLBCL in ART-naïve HIV+ adults will be associated with superior immune reconstitution of the T-cell repertoire. Methods: For cohort 1, serial peripheral blood mononuclear cell (PBMC) samples and clinical data were collected from 30 HIV+ adult subjects enrolled on prospective studies conducted by the Center for AIDS Research (CFAR) at the University of Washington, Seattle, WA. 1-4 PBMC samples were collected pre-ART and 2-6 PBMC samples post-ART from each subject. PBMC samples were also collected from 16 HIV- adult control subjects. For cohort 2, serial PBMC samples and clinical data are being collected from HIV+ adult subjects presenting to the Uganda Cancer Institute – Fred Hutchinson Cancer Centre in Kampala, Uganda for treatment of pathologically confirmed DLBCL. To date, PBMC samples have been collected from 50 subjects. High-throughput T-cell receptor β chain (TRB) sequencing has been performed on all 168 samples from the HIV+ cohort; analysis of the HIV+ lymphoma cohort is ongoing. Results: The TRB repertoire in the HIV+ cohort prior to ART initiation was significantly more “clonal” (less diverse) than that observed in the HIV- controls. Following initiation of ART, an increase in repertoire diversity was observed, accompanied by a substantial improvement in CD4+ T-cell count. Increased repertoire diversity was associated with an increase in the number and frequency of “public” TRB sequences, many which are associated with CD8+ T-cell responses to HIV epitopes. Conclusion: Initiation of ART in HIV+ adults was associated with changes in the global and pathogen-specific T-cell repertoires. ART initiation was also associated with increases in the frequency of “public” TRB sequences associated with immunodominant CD8+ T-cell responses to HIV and other viral pathogens, suggesting that recovery of CD4+ T-cells may enable expansion of pathogen- specific and tumor-specific CD8+ T-cells. Boris P. Hejblum 1 , Daniel Commenges 1 , Chariff Alkhassim 1 , Raphael Gottardo 2 , François Caron 3 , Rodolphe Thiébaut 1 1 INSERM, Bordeaux, France, 2 Fred Hutchinson Cancer Research Center, Seattle, WA, USA, 3 University of Oxford, Oxford, UK Background: Flow and mass (CyTOF) cytometry are high-throughput technologies quantifying multiple surface and intracellular markers at the level of a single cell. Improvements of these technologies allow to describe millions of individual cells from a single blood sample according to several dozens of markers (up to 2^30 cell populations with 30 markers). This generate high-dimensional datasets, whose manual analysis, called manual gating, is highly time-consuming and poorly reproducible. We have developed 2 machine learning approaches to perform automatic gating without human intervention. Methods: The first method is a new Bayesian nonparametric approach (NPflow) with Dirichlet process mixture (DPM) of multivariate skew t-distributions to performmodel-based clustering of flow-cytometry data. DPMmodels directly estimate the number of cell populations from the data, avoiding model selection issues, and skew t-distributions provides robustness to outliers and non-elliptical shape of cell populations. To accommodate repeated flow-cytometry measurements, such as in a clinical trial, a sequential strategy relying on a parametric approximation of the posterior is also proposed (NPflow seq). The second one (cytometree) is based on the construction of a binary tree, whose nodes represents cellular sub-populations. At each node, a binary split between different cellular populations is done according to the normalized difference of Akaike Information Criteria (AIC) between the two normal mixture models considering either one or two possible sub-populations. Post-processing of the tree structure and derived populations allows us to automatically provide a complete annotation of the derived populations. Results: The good performance of the methods are shown on simulated data and on an experimental benchmark datasets (FLOWCAP1) as shown in Table 1, as well as in a real dataset from an HIV vaccine trial. Compared to other available approaches the newmethods, especially cytometree, performed at the top position with the shortest runtime. Also, the F-measures>0.90 demonstrate the validity of the newmethods in comparison with the gold standard (consensus of 8 experimentalists). Conclusion: The constant increase of the number of markers available to characterize cell populations leads to an untractable situation with manual gating. However, improvements of machine learning approaches allow for the automatic analysis of cytometry samples, yielding the relative count for all possible cell populations.

319 MACHINE LEARNING REVEALS T-CELL ACTIVATION PATHWAYS INDUCED BY INFLUENZA VACCINE Megan Cole 1 , Zainab Saeed 1 , A. Torm Shaw 1 , Yanping Guo 1 , Katja Hoschler 2 , Alan Winston 1 , Graham S. Cooke 1 , Sarah Fidler 1 , Graham P. Taylor 1 , Katrina Pollock 1 1 Imperial College London, London, UK, 2 Public Health England, London, UK Background: Seasonal influenza vaccine responsiveness is dependent on a specialised subset of CD4+ T-cells that are susceptible to infection with HIV. The role of CXCR5+ circulating T-follicular helper cells (cTFH) in this response is unclear in individuals with treated suppressed HIV infection. Investigations of potential biomarkers for HIV integration have identified rare CD4+ T-cells highly expressing the FC gamma receptor CD32, but with unknown function. We hypothesised that CD32 is upregulated on cTFH in response to seasonal influenza vaccine and used unsupervised computer algorithms to explore the cellular relationships arising. Methods: 16 men with treated, suppressed HIV infection and 14 healthy control subjects receiving quadrivalent influenza vaccine (QIV) during the 2017-18 Northern Hemisphere influenza season were studied. Peripheral blood mononuclear cells (PBMCs) were collected prior to and after vaccination. Thawed PBMCs were stained with a pre-optimised cocktail of flurochrome- conjugated antibodies, before acquisition on a BD Fortessa flow cytometer. The data were analysed using T-stochastic neighbour embedding analysis (t-SNE) and Spanning-tree progression analysis of density-normalized events in FlowJo v10.4.2 and FCS express v6plus. Results: cTFH more frequently expressed CD32 at Day 7 post QIV (p=0.0009) and returned to baseline at Day 28 (p<0.0001) with no difference in those with and without HIV infection. t-SNE identified three populations (P1, P2 and P3) of CD4+ T-cells that were defined by their expression of CXCR5 and CD32. P1 (CXCR5hiCD32hi) and P3 (CXCR5midCD32lo/mid) frequency was constant but P2 (CXCR5lo/midCD32lo/mid) was more frequent at Day 7 (p=0.0261) and expressed the cTFH activation markers programmed cell death 1 (PD-1) and inducible T-cell co-stimulator (ICOS). SPADE indicated a branched hierarchy of clustered nodes corresponding to P1, 2 and 3. Consistently, a central memory CXCR5mid node gave rise to a CXCR5hiCD32hi node that was unaffected by QIV and two vaccine-inducible activated ICOS+PD- 1+CD38+CXCR3+CXCR5+CD32mid/hi nodes. Conclusion: Circulating CXCR5+CD4+ T-cells fall into three major related populations. A parent population of cTFH-like cells gives rise to a vaccine- responsive cTFH population that upregulates CD32 and a vaccine-unresponsive population persistently expressing CD32 and CXCR5. These relationships were present irrespective of HIV infection in individuals receiving QIV and could be used to inform vaccine design. 320 COMPARATIVE ANALYSIS OF THE MAGNITUDE AND QUALITY OF VACCINE- ELICITED T-CELL RESPONSE Bianca Schulte 1 , Franco Pissani 2 , Michael A. Eller 2 , Bruce T. Schultz 1 , Mary Marovich 2 , Prasert Thongcharoen 3 , Somchai Sriplienchan 4 , Supachai Rerks- Ngarm 5 , Punnee Pitisutthithum 3 , Stefan Esser 6 , Galit Alter 7 , Merlin L. Robb 2 , Jerome H. Kim 2 , Nelson L. Michael 2 , Hendrik Streeck 6 1 University Hospital of Duisburg-Essen, Essen, Germany, 2 US Military HIV Research Program, Silver Spring, MD, USA, 3 Mahidol University, Bangkok, Thailand, 4 Armed Forces Research Institute of Medical Sciences in Bangkok, Bangkok, Thailand, 5 Ministry of Public Health, Nonthaburi, Thailand, 6 University Hospital Essen, Essen, Germany, 7 Massachusetts General Hospital, Boston, MA, USA Background: CD4 T cell responses that provide efficient help for B cells to generate a long-lived, high-affinity antibody response are considered to be an important component of an HIV vaccine. However, it is unknown how to elicit effective HIV-specific CD4 T cell responses and which vaccination strategy induces optimal T follicular helper (Tfh) cell responses. Here we evaluated and

Poster Abstracts

318 FINDING THE CELLS AMIDST THE DATA

CROI 2019 116