Le développement d'un traitement curatif du virus de l’immunodéficience humaine (VIH) est devenu le défi majeur pour le futur mais les réservoirs et une réponse immunitaire insuffisamment efficace constituent des barrières pour l’élimination du virus. La réponse immune contre l’infection VIH dépend en partie de la capacité des cellules hôtes à présenter correctement les épitopes viraux pour induire une réponse spécifique des lymphocytes T CD8+ cytotoxiques (CTL). Cette présentation des épitopes viraux qui pourrait être optimisée par vaccination, dépend du système HLA (Human Leukocyte Antigen) du patient présentant un déterminisme génétique individuel. Correctement pré-stimulés, les CD8 cibleraient et détruiraient efficacement les cellules productrices du virus. Les précédents essais vaccinaux stimulant la réponse CTL n’ont pas montré d’efficacité dans la réponse virologique, probablement parce qu’ils sont composés d’épitopes "génériques" et qu’ils ne prennent pas en compte la variabilité du VIH ni l’immunogénétique des patients.L’objectif de ce travail est d’identifier des épitopes archivés dans l’ADN proviral, susceptibles d’induire une réponse CTL en considérant la variabilité du VIH-1 et la variabilité immunogénétique des patients infectés par le VIH-1 en succès thérapeutique. L’idée, à terme, est d’utiliser les peptides correspondants comme base de vaccin thérapeutique et de permettre au système immunitaire de prendre le relai du traitement antirétroviral pour contrôler l’infection virale.Cent quarante patients infectés par le VIH-1, suivis au CHU de Bordeaux en succès thérapeutique depuis plus de 6 mois ont été inclus dans le projet Provir/Latitude 45 entre 2012 et 2017. Une cartographie de la répartition des sous-types viraux du VIH-1 en Aquitaine a d’abord été réalisée. L’analyse de plus de 3200 génotypes de résistance VIH-1 effectués entre 2012 et 2016 a permis de déterminer que le sous-type viral majoritaire infectant les patients vivants avec le VIH dans la région est le sous-type B, suivi du CRF02_AG qui est majoritaire parmi les sous-types viraux non B. Si l’on se focalise sur les patients inclus dans ce projet on retrouve des répartitions similaires. Suite à cette première analyse, un nouveau virus recombinant composé de CRF06_cpx et de sous-type B a pu être identifié. Il est référencé en tant que CRF98_cpx. Un des patients inclus au sein du projet est d’ailleurs infecté par ce virus. Afin d’identifier des épitopes candidats pouvant servir de base à un vaccin thérapeutique pour l’ensemble de la population ou pour un groupe de la population en fonction de leurs caractéristiques immunogénétiques, nous avons combiné les données des séquences virales avec le typage HLA des patients afin de prédire in silico l’affinité entre un HLA et un peptide via les algorithmes d’IEDB. Pour automatiser l’analyse des données, un logiciel, TutuGenetics, a été développé. Ce logiciel instrumentalise les algorithmes d’IEDB et permet d’étudier la variabilité de la présentation des épitopes par les différents HLA du patient grâce à un score MHC IC50 déterminé pour chaque couple HLA-séquence virale. Ce logiciel a été validé en comparant l’analyse des données issues du séquençage Next Generation Sequencing avec le séquençage Sanger. Finalement, pour les 140 patients inclus dans le projet Provir, TutuGenetics a effectué un découpage des séquences virales par pas de 8 à 10 acides aminés et les valeurs MHC IC50 ont été définies pour toutes les combinaisons HLA-épitope. Une analyse plus fine nous a ensuite permis de déterminer une liste de 15 épitopes avec une forte affinité in silico pour les HLA majoritaires finalement retenue pour une cocktail vaccinal.Ces données in silico vont dans une prochaine phase être confirmées in vitro via des tests d’immunologie fonctionnelle, puis in vivo chez le macaque. / HIV (Human Immunodeficiency Virus) cure is the major challenge of the future but latent reservoir and inefficient immune response do not allow virus elimination. The immune response against HIV infection depends on host cells ability to correctly present viral epitopes to induce specific cytotoxic CD8+ T lymphocytes (CTL) response. This presentation of viral epitopes which could be improved by vaccination depends on the HLA (Human Leukocyte Antigen) system of the patient which is extremely variable. Accurately pre-stimulated, CTL would target and destroy efficiently virus-producing cells. Previous vaccine trials stimulating CTL response haven’t shown efficient virological response, presumably because epitopes used are generic without taking into account HIV-1 or patient’ immunogenetic variability.The aim of this work is to identify epitopes archived in the proviral DNA, considered to induce CTL response according to the HIV-1 and immunogenetic variability of the patients at therapeutic success. The goal is to use these peptides for a therapeutic vaccine and educate the immune system to control the viral replication without any antiretroviral treatment.One hundred and forty patients infected with HIV-1, followed at the University Hospital of Bordeaux, at therapeutic success for more than 6 months have been included in the Provir/Latitude 45 project between 2012 and 2017. A mapping of the distribution of viral subtypes of HIV-1 in Aquitaine was first performed. Analysis of more than 3200 HIV-1 genotypes conducted from 2012 to 2016 determined that the major viral subtype infecting patients living with HIV in the region is subtype B, followed by CRF02_AG which is predominant among the non-B viral subtypes. Focusing on the patients included in this project led us to find similar distributions. Following this initial analysis, a new recombinant virus composed of CRF06_cpx and subtype B could be identified. It is now referenced as CRF98_cpx. One of the patients included in the project is infected with this virus. In order to identify candidate epitopes that can serve as a therapeutic vaccine for the entire population or for a population group based on their immunogenetic characteristics, we combined the viral sequence data with the HLA typing of patients to predict the affinity in silico between an HLA and a peptide via IEDB algorithms. To automate data analysis, a software package, TutuGenetics, has been developed. This software exploits IEDB algorithms and makes it possible to study the variability of the presentation of epitopes by the different HLAs of the patient thanks to an MHC IC50 score determined for each HLA-viral sequence pair. This software has been validated by comparing the analysis of data from Next Generation Sequencing (NGS) with Sanger sequencing. Finally, for 140 patients included in the Provir project, TutuGenetics sliced the viral sequences in steps of 8 to 10 amino acids and MHC IC50 values were defined for all HLA-epitope combinations. A finer analysis allowed us to determine a list of 15 epitopes with high in silico affinity for major HLAs. These epitopes were selected by applying different filters and only HLA-peptide couples with more than 10 patients sequenced per position and by HLA were kept in this "Optimal_Provir" list.These in silico data will in a next phase be confirmed in vitro via functional immunology tests, then in vivo in macaques.

Of the 12 million people who inject drugs worldwide, 13% are chronically infected with Human Immunodeficiency Virus (HIV), i.e., they live with HIV. Chronic opioid use affects the host immune system and increases an individual’s susceptibility to HIV infection. However, it is unclear how opioid use changes the course of HIV pathogenesis. Particularly, there is a gap in understanding how opioids impact HIV latency. Latency results in a reservoir of infected quiescent cells that evade antiviral immune responses, are not targeted by antiretroviral therapy (ART), and allow HIV viremia to rebound upon treatment interruption. While in vitro studies show that opioids modulate the activity of transcription factors involved in T-cell activation and HIV transcription, few studies have investigated whether opioid use impacts HIV latency in vivo in HIV-infected people. In this research, peripheral blood mononuclear cells (PBMCs) were utilized from People with HIV (PWH) with or without recent opioid use or opioid use disorder (OUD) who were enrolled in the Linking Infection and Narcology Care-Part II (LINC-II) and Studying Partial Agonists for Ethanol and Tobacco Elimination in Russians with HIV (St PETER HIV ARCH) studies conducted in St. Petersburg, Russia. Intact proviral DNA digital droplet PCR (ddPCR) assays were performed on PBMCs from antiretroviral treated PWH, with (n=8) or without (n=11) current OUD, to quantify intact and defective proviral genomes. Samples from ART-treated PWH with OUD compared to those without OUD had similar levels of intact and defective proviruses. To evaluate latency reversal, PBMCs from ART-treated PWH with or without OUD, were activated with anti-CD3/28 beads and RT-ddPCR assays were performed to measure HIV LTR-gag RNA. A variable response in PWH without OUD was seen where half of the samples showed an increase in HIV RNA upon activation. Interestingly, only 1 of 8 samples from PWH with OUD showed an increase in HIV transcription. However, no suppression of HIV reactivation was found in vitro from latent cells generated using a primary CD4+ T-cell latency model in the presence or absence of morphine. Similarly, no differences in HIV integration and transcription in vitro were observed between morphine and control conditions. Additionally, expression of opioid receptors was not detected in primary PBMCs, CD4 T cells, or macrophages. These results show that PWH with OUD have a pool of persistent HIV proviruses that are refractive to reactivation, although opioids did not affect HIV replication and latency reactivation in vitro. The discrepancy in these in vitro and in vivo results and the lack of expression of opioid receptors in immune cells suggests that while opioids do not directly impact HIV replication, latency, and reactivation in target CD4+ cells, opioids could indirectly shape the HIV reservoir in vivo by modulating general immune functions, neuroderived factors or other cells that are responsive to opioids. Eradication of the latent HIV reservoir is necessary to achieve a cure for HIV/AIDS. One approach for latency eradication is the “shock and kill” approach that entails stimulating viral production with latency-reversing agents followed by the killing of cells actively producing the virus by immune clearance. However, this approach does not induce all intact proviruses, leaving a residual reservoir. An alternative approach is to permanently repress HIV expression precluding viral rebound after ART discontinuation. Here, a nuclease-deficient disabled Cas9 (dCas9) coupled with a transcriptional repressor domain derived from Kruppel-associated box (KRAB) was used to epigenetically silence the proviral DNA. I show that specific guide RNAs (gRNAs) and dCas9-KRAB repress HIV-1 transcription and reactivation of latent HIV-1 provirus. This repression is correlated with chromatin changes, including decreased H3 histone acetylation and increased histone H3 lysine 9 trimethylation, which are histone marks that are associated with transcriptional repression. dCas9-KRAB-mediated inhibition of HIV-1 transcription suggests that CRISPR can be engineered as a tool for block-and-lock strategies. The research presented here provides evidence of opioid-mediated modulation of HIV-1 latency reactivation in PWH with opioid dependency. Additionally, we show that HIV-1 reactivation can be suppressed by epigenetic remodeling of the HIV-1 promoter using a repurposed CRISPR/Cas9 system.