Thèse Astrocyte-T Cell Interactions In Neuroinflammation H/F - 33

Établissement : Université de Bordeaux
École doctorale : Sciences de la Vie et de la Santé
Laboratoire de recherche : Biologie des maladies cardiovasculaires
Direction de la thèse : Candice CHAPOULY ORCID 0000000191731615
Début de la thèse : 2026-11-01
Date limite de candidature : 2026-05-20T23:59:59

La sclérose en plaques (SEP), une maladie inflammatoire démyélinisante du système nerveux central (SNC), représente un problème majeur de santé publique, car les traitements actuels ne permettent pas de prévenir la neurodégénérescence. Dans la SEP, les cellules immunitaires envahissent de manière anormale le parenchyme à partir de la circulation sanguine après avoir franchi la barrière hémato-encéphalique (BHE). Plus précisément, des anomalies dans les fonctions des cellules T régulatrices (Treg), associées à une augmentation des cellules T effectrices (sous-ensembles de cellules T CD4 activées périphériquement, T-helper (Th) 1/17) et des cellules T CD8, entraînent la formation de lésions inflammatoires et la dégénérescence dans la SEP. Après avoir traversé la BHE, les cellules T sont piégées dans l'espace périvasculaire (PVS) où elles rencontrent les pieds astrocytaires. Des signaux spécifiques médiés par le contact entre ces derniers et les cellules T infiltrantes régulent probablement le trafic et les voies immunitaires essentielles à la progression de la maladie. Cependant, ces signaux restent à identifier. Le récepteur Notch apparaît comme un régulateur clé du développement des sous-ensembles de cellules T favorisant l'auto-immunité dans la SEP. Le blocage de Notch atténue la gravité de l'encéphalite auto-immune expérimentale (EAE) et fait basculer l'équilibre immunitaire d'une réponse médiée par Th1/Th17 vers une réponse médiée par Treg. De plus, nous avons montré que le ligand Notch Dll4 est exprimé par les astrocytes réactifs et soutient l'astrogliose pendant l'EAE.
Nous émettons l'hypothèse que la protéine astrocytaire Dll4 interagit avec ses récepteurs Notch1 et 2 sur les lymphocytes T CD8 et CD4 infiltrés, d'abord dans le PVS, puis dans le parenchyme, afin de stimuler la fonction des sous-ensembles de lymphocytes T au cours de la progression de la maladie.
Notre objectif est d'identifier comment la voie Dll4-Notch contrôle le comportement des lymphocytes T au cours de la neurodégénérescence.
Objectif 1 : Le Dll4 stimule-t-il différemment le phénotype des cellules T dans le PVS par rapport au parenchyme dans l'EAE ?
Nous induirons l'EAE chez des souris dont les astrocytes ont été rendus insensibles au Dll4. Nous analyserons la pathologie dans l'EAE aiguë et chronique et isolerons les unités neurovasculaires de la moelle épinière et la fraction parenchymateuse. Nous établirons ensuite le profil des cellules T (capacité d'activation/migration) dans les deux fractions.
Objectif 2 : Comment l'expression de Dll4 dans les astrocytes contrôle-t-elle le comportement des lymphocytes T infiltrés dans la SEP ?
Nous allons réguler à la baisse ou à la hausse l'expression de Dll4 dans les astrocytes humains (HA). Nous allons co-cultiver des HA avec des cellules T provenant de donneurs atteints de SEP ou sains (HD), dont le profil sera ensuite analysé par cytométrie en flux et RNAseq.
Objectif 3 : La signalisation Dll4-Notch entre les astrocytes et les cellules T est-elle une voie effectrice commune favorisant la neurodégénérescence dans d'autres maladies que la SEP ?
Nous analyserons le profil d'infiltration des cellules T et leur réponse au Dll4 astrocytaire chez des souris atteintes de démence vasculaire et d'accident vasculaire cérébral ischémique.

Astrocyte-immune cell interactions in multiple sclerosis (MS)
MS, an inflammatory demyelinating disease of the central nervous system (CNS), represents a major public health issue as current therapies do not prevent neurodegeneration. In MS, immune cells improperly invade the parenchyma from the bloodstream after crossing the blood-brain barrier (BBB). Specifically, defects in regulatory T cell (Treg) functions combined with an increase in effector (peripherally-activated CD4 T cell subsets T-helper (Th) 1/17) and CD8 T cells drive inflammatory lesion formation and degeneration in MS. After crossing the BBB, T cells are trapped in the perivascular space (PVS) where they encounter astrocytic endfeet. At this checkpoint, specific contact-mediated signals between astrocytes and infiltrating immune cells likely regulate trafficking and immune pathways critical to CNS inflammatory disease activity. How astrocyte-immune cell interactions control chronic progressive disease after initial relapse is not known.
Astrocytes in MS, the doctor Jeckyll and Mr Hyde cells of the neurovascular unit
The concept of the BBB structure and integrity has recently been expanded to include contributions, among others, from endothelial cells, pericytes, microglia and astrocytes. However, how signals from these cells participate to BBB dysfunction and neuroinflammation remains unclear and is of considerable translational interest to the field of neuro-immunology. Specifically, while it is now established that BBB breakdown promotes immune cell infiltration into the PVS during neuropathology, astrocytes described as reactive, are emerging as ambivalent cells having complex roles in both recruiting and restricting neuroinflammatory infiltration. Reactive astrocytes produce, on one hand, pro-inflammatory and pro-permeability factors and, on the other hand, neuroprotective factors and their behavior is determined in a context-specific manner by signaling events that vary with the nature (infections, MS, Alzheimer's and Parkinson's disease) and severity of CNS insults.
Dll4-Notch signaling in astrocytes, a new player controlling inflammation in MS
Notch1 receptor is a central effector of astrogliosis in a wide range of neuropathological contexts notably MS and experimental autoimmune encephalomyelitis (EAE) where Notch1 signal transduction is activated in reactive astrocytes. Notch signaling is stimulated by the interaction of Notch receptors with their ligands, in particular Delta-like 4 (Dll4), leading to the Notch intracellular domain (NICD) release, which translocates to the nucleus to promote transcription.
Interestingly, activation of the Notch1-Stat3 axis in EAE has been shown to control the production of inflammatory cytokines by reactive astrocytes via the long non-coding RNA Gm13568, also implicated in MS pathogenesis. Moreover, the Notch1-Stat3 pathway has been identified as an effector of inflammation-induced differentiation of neurotoxic A1 astrocytes in a model of spinal cord injury. Surprisingly, Dll4 expression has only been reported once, anecdotally, in reactive astrocytes, following brain injury. We recently identified Dll4 as being highly expressed by reactive astrocytes by performing RNA sequencing on non-stimulated vs. reactive human astrocytes, suggesting a role, currently unexplored, for Dll4 in astrocytes during neuroinflammation.
Dll4-Notch signaling in T cells, other key players controlling inflammation in MS
Mature naive CD4 and CD8 T cells express Notch1 and Notch2 receptors, with upregulated expression following T cell receptor (TCR) stimulation, and Notch signaling participates in T cell differentiation. Indeed, Notch acts as an unbiased amplifier of T cell differentiation promoting either Th1 or Th2 responses under different polarizing conditions and it also contributes to the maintenance of the Th17 response. Moreover, in vivo work suggests that Notch signaling curtails Treg function. Finally, novel emerging properties of the Dll4-Notch signaling pathway have been identified in pathological condition, notably its role in contributing to the balance of the CD4/CD8, Th17/Treg ratio both in experimental autoimmune uveitis and EAE models.
We hypothesize that astrocytic Dll4 interacts with its receptors Notch1 and 2 on infiltrated CD8 and CD4 T cells, first in the PVS and second in the CNS parenchyma, to drive T cell subset function during disease progression.

Aim1: Does astrocytic Dll4 upregulation differentially drive T cell function via Notch signaling (1) in the PVS vs. parenchyma and (2) during acute vs. chronic EAE?
Aim1 focuses on (1) experimental observations of reactive astrocytes at the BBB and in the CNS parenchyma and (2) T cell infiltration behavior in response to Dll4 astrocytic upregulation during the acute and chronic phases of EAE. Data from C. Chapouly's lab show that Dll4 conditional deletion in astrocytes leads to milder EAE associated to markedly decreased CD4 T cell infiltration during the initial acute phase. We hypothesize that astrocytic Dll4 promotes CD8 and Th1/Th17 CD4 T cell-mediated immunity, acting on their activation, proliferation, differentiation and/or migration abilities, which worsen disease and allow for secondary progression.
Mouse model: Dll4flox/flox (fl/fl) mice, under breeding in our animal facility, will be administered intravenously with adenoviruses (5×1011 vg/mouse/virus) carrying the Cre recombinase under a GfaABC1D promoter (an empty vector will be used as a negative control) (plasmids from Addgene). This will allow the KO of Dll4 astrocytic expression at 3 distinct time points following EAE induction. Adenoviruses are currently being built by the Vectub platform (Bordeaux).
EAE will be induced in Dll4fl/fl mice (10wk-old females, 10/group/study) with MOG35-55 peptide emulsified in complete Freund's adjuvant. On the same day, and again the following day, mice will receive intraperitoneal injection of pertussis toxin. Animals will develop EAE (10-14 post immunization) and become paralyzed. Daily scoring of mice start on day 7 and continue until the end of the study when the animals are euthanized. Scoring is on the scale of 0 to 4 (1-limp tail, 2-partial hind limb paralysis, 3-complete hind limb paralysis, 4-complete hind and partial front leg paralysis). To measure stage-specific effects of astrocytic Dll4 deletion on clinical course of EAE, we will induce Dll4 KO at 3 different time points: 1) acute peak, 2) partial remission and 3) chronic progression of the disease (days 10, 30 and 45 post immunization).
Rational: Dll4 is upregulated in reactive astrocytes both in EAE and active lesions of patients with MS (pwMS) and astrocytic Dll4 sustains astrogliosis during EAE. Pathology readout: Clinical score of paralysis will be assessed daily from day 7 to day 50 post-immunization. Groups of mice will be sacrificed 5 days after inducing Dll4 astrocytic KO, at day 15, 35, and 50. Spinal cord tissues will be harvested at the 3 time points to correlate scoring to a histological analysis of pathology (demyelination, oligodendrocyte loss and axonal transection). Inflammation readout: Reactive astrocyte phenotypes (as described in M. Absinta et al. 23) (Vim, Lcn2, Serpina3, Gfap, C3, IL-6, IL-8) will be analyzed by immuno-histology. Then, to selectively study the phenotype of T cells infiltrated into the parenchyma at the 3 time points, we will pharmacologically deplete T cells at the periphery in mice induced with EAE. To selectively study the phenotype of perivascular infiltrated T cells at the 3 time points, neurovascular units from EAE induced mice will be isolated following a protocol routinely performed and already published in C. Chapouly's lab (Dextran gradient separation and MACS Miltenyi isolation kits). We will profile immune cells for their activation, proliferation, differentiation and migration ability as well as Notch signaling pathway activity (Hes1, Hes5, Hes7, Hey1, Hey2, Ccnd1 and c-Myc) at the 3 time points on isolated neurovascular units and parenchymal fractions as well as cervical lymph nodes as a reference. More specifically, we will perform flow cytometry analyses (expression of CD3, CD4, CD8, CD44, CD69, CD25, PD1, Lag3, TIM3, Ki67, CD62L, CXCR3, CCR5, CXCR6, CD49d, CD49a, P2RX7, CD103, T-Bet, Eomes, Roro, Foxp3 at the single cell level) and bulk RNA-seq of cell-sorted subsets, including antigen-experienced (CD44high) CD4 T cells, antigen-experienced CD8 T cells, and Foxp3 expressing Tregs. We will also investigate the functional properties of the T cells by their production of cytokines and cytotoxic molecules (Granzymes, Perforin) following PMA/ionomycin stimulation.
Aim2: How does Dll4 upregulation in reactive astrocytes control T cell behavior through Notch receptor activation in MS?
Aim2 focuses on deciphering Notch signaling response and underlying activated pathways involved in the reactive astrocyte-lymphocyte crosstalk in MS. Specifically, CD8 and CD4 T cell profiles (activation, proliferation, differentiation and adhesion pathways) in response to astrocytic Dll4 expression level will be unraveled. Preliminary data show that activated human T cells essentially express the receptors Notch1 and 2. Moreover, while Th1 and Th17 hardly express any Dll ligand, Dll3 (an antagonist of Notch) is weakly expressed by Tregs while Notch2NLC (a positive regulator of Notch) is downregulated when compared to Th1. We, therefore, hypothesize that expression of distinct combinations of Notch family members by the various T cell subsets leads to differential effects upon interaction with astrocytic Dll4.
Co-culture model: primary human astrocytes (HA, ScienCell) will be co-cultured with primary human CD8 T cells and CD4 Th1, Th17 and Treg isolated from PBMC (from healthy donors (HD) or pwMS).
Task a) T cell subset behavior analysis in response to astrocytic DLL4 expression level: Primary HA will be cultured in HA growing medium for 24 hours. Astrocyte DLL4 expression will either be compromised using DLL4 siRNA 24 or upregulated using DLL4-expressing lentivirus. HA treated with DLL4 siRNA vs. CTRL siRNA will then be serum starved and stimulated by IL-1 for 24h to induce reactivity.
In parallel, human primary CD8 T cells and CD4 Th1, Th17 and Treg will be isolated from PBMC of HD and expanded as routinely done in A. Astier's lab. In brief, Th1, Th17 and Tregs will be purified from PBMC by flow cytometry and expanded for 2 weeks before co-culture with the HA. Reactive HA knockout for DLL4 and non-reactive HA overexpressing DLL4 vs. respective controls will be co-cultured in T cell medium with CD8 T cells or the different expanded CD4 Th subsets in the presence or absence of CD3/CD28 beads (dynabeads). Kinetic experiments (day 2 and 5) will first be performed to assess Notch expression related to T cell activation. Cells will be examined, as routinely done, by using large flow cytometry panels (17 colors on the Symphony) aiming at following T cell activation, differentiation and migration (incl. CD3, CD4, CD8, Ki67 (proliferation) and Notch 1, Notch 2, Dll3, CD69, CD25, CD49b, LAG3, IL-10, IFNg, TNF, Grz A, GM-CSF, IL-17, c-MAF, t-BET, RORO, EOMES, CD161, CD11a, CD49d, PSGL1).
Furthermore, we will separate HA from T cells after 2 or 5 days of co-culture (MACS Miltenyi human ACSA-1, CD8 and CD4 T cell Microbeads isolation Kits). Purified CD8 T cells and CD4 Th1, Th17 and Treg lysates will then be harvested to perform bulk RNA-seq. In parallel, HA lysates will be harvested to verify DLL4 knockdown/upregulation efficiency and astrocyte reactivity (as described in aim1). RNA-seq databases will allow us to identify activated or repressed genes in T cell populations cultured with control vs. DLL4-knockdown reactive HA or DLL4-upregulated non-stimulated HA, allowing identification of putative biomarkers implicated in T cell subtype activation, proliferation, differentiation and adhesion. In parallel, Treg suppressive function will be assessed by co-culturing the Tregs with autologous CTV-prelabelled CD4 T cells depleted in Tregs.
Task b) Demonstration of a Notch-dependent reciprocal dialogue between unstimulated astrocytes and T cell subsets from PwMS or HD: Primary HA will be cultured in HA growing medium for 24 hours. In parallel, human primary CD8 T cells, CD4 T cells depleted of Tregs (StemCell kit) and Tregs will be isolated from 10 pwMS (relapsing-remitting form) treated with Natalizumab (BIONAT cohort, allowing to enrich cultures in potentially pathogenic/CNS invading T cells) and 10 HD, and the T cells will be assessed for the different Notch family members. Preliminary data suggest that expression of Notch1 is increased in activated MS CD4 T cells compared to HD. Non-stimulated HA will then be co-cultured in T cell medium with CD8 T cells or distinct CD4 Th subsets from pwMS or HD for 2 and 5 days to measure whether the lymphocytes can reciprocally impact HA reactivity through Notch signaling pathway. More specifically, following separation by MACS technology, HA (cultured with T cell subsets from pwMS or HD) lysates will be harvested and astrocyte reactivity as well as Notch signaling pathway activity (as described in aim1) will be assessed by RT-qPCR and/or western blot.
Task c) Identification of molecular actors associated to Notch signaling pathway involved in CD8+ T cells and CD4+ Th subset identity: The implication of specific Notch family members (Notch1/Notch2/Notch2NLC and Dll3) in primary T cells from both HD and pwMS will then be determined by genetic loss and upregulation approaches, and we will assess the impact on both T cell and HA phenotypes in co-cultures. Phenotype of the cells will be first examined by flow cytometry using the panels generated in task a). Furthermore, after separation by MACS technology, HA and T cell lysates (4-5 donors per condition) from the different conditions will be harvested for RNA-seq analyses. Modulated pathways will then be identified and further validated by using downregulation (Crispr/Cas9) or upregulation (agonists/lentivirus vectors) strategies in vitro.
Task d) Notch signaling analysis in acute and chronic active CNS lesions from pwMS: Finally, following identification of the T cell infiltrate profile and response to astrocytic Dll4 in the PVS vs. the CNS parenchyma under EAE condition (aim1), we intend to measure the spatial distribution of Dll4-Notch signaling activity in human tissues from pwMS using spatial transcriptomic (Visium technology, 10x Genomics, PUMA, Magendie neurocenter, Bordeaux), in 5 pwMS and 5 non-neurologic controls obtained from the NeuroCEB (www.neuroceb.org) biobank. We will compare the distribution and activity of the Notch pathway (Hes1, Hes5, Hes7, Hey1, Hey2, Ccnd1 and c-Myc transcription levels) in the astrocytes and different T cell subsets in acute vs. chronic active lesions in order to correlate Notch signaling activity with disease severity. Other pathways specifically enriched in the acute vs. chronic active lesions will be identified for future prospects.
Aim3: Is Dll4-Notch signaling between astrocytes and lymphocytes a common effector pathway to promote neurodegeneration in CNS pathologies other than MS? Conversion from acute inflammatory injury to a chronic neurodegenerative state is a clinical hallmark of MS that also occurs in ischemic stroke and dementia and the role of Notch signaling has been suggested in these pathologies. We hypothesize that upregulation of Dll4 in reactive astrocytes and subsequent Notch signaling activation in infiltrated immune cells is a shared hallmark for various diseases leading to neurodegeneration. Based on the literature and our expertise, we plan on studying this shared hallmark in vascular dementia caused by ameroid constrictor-induced brain hypoxia and in a transient ischemic stroke model (insertion of a monofilament through the common carotid artery to the origin of the middle cerebral artery). In each model, induced in C57BL/6 mice, Dll4 upregulation in reactive astrocytes will be assessed and correlated to neurodegeneration severity using immuno-histology. Additionally, the CNS-infiltrating T cells isolated from the PVS and parenchymal fractions will be profiled as described in aim1. In parallel, we will induce (1) vascular dementia and (2) transient ischemic stroke in Dll4fl/fl mice. We will measure effects of astrocytic Dll4 deletion, induced 5 days before cognitive testing or Doppler analysis, on the clinical course of dementia (Y-maze, water maze and nesting to measure cognitive decline) and stroke (Doppler to measure blood flow and confirm occlusion). Brain tissues will be harvested to perform histological analysis and to profile immune cells on isolated neurovascular units and parenchymal fractions as described in aim1.