The central nervous system (CNS) is an immunologically privileged site, where entry of circulating immune cells is tightly controlled by the blood-brain barrier (BBB) at the level of CNS microvessels and by the blood-cerebrospinal fluid barrier (BCSFB) at the level of the choroid plexus epithelial cells. We have hypothesized that the immune privilege of the CNS is established by its unique anatomy that resembles the architecture of a medieval castle being surrounded by two walls. Immune cell entry into the CNS parenchyma thus obeys special mechanisms as it requires the crossing of two barriers, namely the outer BBB and BCSFB followed by the glia limitans. Our research is devoted to understand the molecular mechanisms involved in the migration of immune cells in the CNS during immunosurveillance and neuroinflammation. Combining our expertise in vascular biology and neuroimmunology we have established in vitro and in vivo approaches to study immune cell interactions with the brain barriers mainly in the context of multiple sclerosis and stroke employing animal models and more recently also human in vitro brain barrier models and patient derived immune cells. Pioneering intravital microscopy of the cervical spinal cord of the mouse (Videos 1 and 2) and integrating sophisticated in vitro and in vivo live cell imaging technologies, transgenic mouse models as well as novel in vitro models of the mouse and human blood-brain barrier (BBB) and the blood-cerebrospinal fluid barrier (BCSFB) we have defined molecular mechanisms involved in the migration of different immune cell subsets across the BBB and BCSFB during immunosurveillance and neuroinflammation.
Defining the molecular mechanism mediating the interaction of different T cell subsets with the blood-brain barrier in vivo during initiation and ongoing autoimmune neuroinflammation
Employing experimental autoimmune encephalomyelitis (EAE) as an animal model for multiple sclerosis (MS) and more recently novel transgenic mouse models of CD8 T-cell mediated autoimmune neuroinflammation in combination with novel adhesion molecule-deficient mouse models and function blocking antibodies our work has identified distinct mechanisms for the recruitment of CD8 versus CD4 T cells across the BBB in vivo. In collaborations with Antal Rot (London, UK) and Roland Liblau (Toulouse, France) we found novel traffic signals, e.g. the atypical chemokine receptor 1 ACKR1/DARC (Minten et al, Brain 2014) or the junctional adhesion molecule B (JAM-B) (Martin-Blondel et al., EJI, 2015) to be critically involved in CD4 and CD8 T cell migration across the BBB, respectively. Building on our pioneering research on studying T cell migration across spinal cord microvessels in the cervical spinal cord microvasculature, which allowed us to precisely show the roles of selectins and a4b1-integrins in this process (Vajkoczy et al., JCI, 2001), we expanded the surgical window preparation allowing for twophoton imaging (Haghayegh Jahromi et al. 2017, Front. Immunol). We are developing tools to push forward the limits of bioimaging and have established a novel controllable synchronization tool allowing to use twophoton imaging to visualize tissue regions subjected to periodic movement due to animal breathing. The established image processing pipeline allows for further distortion correction in the acquired images, thus enabling precise measurement of cellular motility parameters. Combined with the use of suitable reporter mouse models this technology will now allow to study the cellular pathway of T-cell diapedesis across the BBB in vivo.
Figure: Inflammatory cuff in the cerebellum of a mouse with EAE.
Green = laminin, red = CD45+ inflammatory cells, blue = nuclei.
Figure: Encephalitogenic T cells (green) interact with inflamed spinal cord micorvessels (red)
The role of blood-brain barrier tight junctions in neuroinflammation
In an effort to delineate the role of BBB tight junctions (TJs) in mediating BBB permeability versus diapedesis of inflammatory cells during EAE we have shown that transgenic mice with inducible endothelial cell specific expression of claudin-1 showed reduced BBB leakiness and thus milder disease while recruitment of inflammatory cells was undisturbed (Pfeiffer et al., Acta Neuropathol, 2011). In collaboration with Alexandre Prat (Montréal, Canada) we defined a novel role for ALCAM in maintaining BBB integrity as it is associated with TJ molecule assembly at the BBB explaining the increased permeability of CNS blood vessels in ALCAM KO animals (Lécuyer et al.,. PNAS, 2017). Employing in vitro studies we could demonstrated that junctional integrity of the BBB is regulated independently of the cellular pathway of leukocyte diapedesis (Abadier et al., EJI 2015). Recently we showed that JAM-B is localized in BBB TJs and its absence ameliorates EAE due to the trapping of immune cells in perivascular and leptomeningeal spaces (Tietz et al., BBI, 2018). In collaboration with Mikio Furuse (Okazaki, Japan) and Christer Betzholtz (Uppsala, Sweden) we recently established a claudin-3 KO mouse and could clarify that while claudin-3 is expressed by choroid plexus epithelial cells it is absent in brain endothelial cells and thus does not contribute to BBB TJ integrity (Castro Dias et al., Sci Reports 2019).
Claudin-3 is localized to tight junctions of the BCSFB but not the BBB. Immunofluorescent staining for claudin-3, PECAM-1 and nuclei in brain parenchyma (left) and choroid plexus (right).
Role of the choroid plexus in immune cell entry into the CNS during experimental autoimmune encephalomyelitis
Based on our early observation that during EAE the adhesion molecules ICAM-1 and VCAM-1 are upregulated at the blood-cerebrospinal fluid barrier (BCSFB) in parallel to their upregulation at the BBB (Steffen et al., AJP 1994) we have speculated on a function of the choroid plexus in CNS immunity. We could show functional expression of ICAM-1 and VCAM-1 at the BCSFB and in collaboration with Federica Sallusto (Bellinzona, Zurich, Switzerland) have provided evidence that Th17 cells may enter the CNS via the choroid plexus in a CCR6/CCL20 dependent manner to initiate EAE. We have established in vitro models of the mouse (Lazarevic and Engelhardt, FBCNS, 2016) and human BCSFB allowing for in depth modeling of the immune function of the BCSFB.
Human in vitro model of the BCSFB
HIBCPP are grown on the inverted filters before turning the filter to grow the BCSFB with a physiological polarization of the epithelial cells.
Modeling the multi-step T cell extravasation across the inflamed BBB under physiological flow in vitro
In tight in-house collaboration with Ruth Lyck, we have set up microfluidic models of the mouse and human BBB allowing to study the multi-step T cell migration across the BBB under conditions of physiological flow by live cell imaging (Coisne et al., FBCNS, 2013). We found barrier characteristics and presence of flow to critically influence the cellular pathway of T-cell diapedesis across the BBB. While endothelial VCAM-1 mediates CD4 T cell arrest, subsequent CD4 T cell polarization and crawling is mediated by endothelial ICAM-1 and ICAM-2 (Steiner et al., JI 2010). Interestingly, cell surface levels of endothelial ICAM-1 direct the diapedesis of CD4 T cells to transcellular versus paracellular sites (Abadier et al., EJI 2015). In addition to CD4 Th1 cells, CD4 Th17 cells contribute to neuroinflammation in MS. Therefore, we have started to side-by-side compare the multi-step interaction of Th1 (red) and Th17 cells (green) on primary mouse brain microvascular endothelial cells under physiological flow conditions. Identification of potential differences will allow to improve safety of current MS therapies or identify novel therapeutic targets for specifically inhibiting the trafficking of pathogenic T cells to the CNS.
To develop automated and unbiased analysis of the multi-step T cell migration across the BBB in our in vitro microfluidic assays we have designed an automated data processing pipeline for T-cell segmentation and tracking on in the phase contrast images acquired in these assays. It is based on specifically developed 3D Deep Convolutional Network (DNN).
Analysing T cell crawling on the BBB. Left: Crop from single raw phase contrast image from the 30-min time-lapse image sequence. Middle: T cells segmented using 3D DNN are highlighted in red. Right: image overlaid with reconstructed T-cell tracks.
To explore subcellular components involved in directing the cellular pathway of T-cell diapedesis across the BBB we have started to combine live cell imaging of T-cell migration across our mouse in vitro BBB in model established from primary mouse brain microvascular endothelial cells(pMBMECs) under physiological flow, with serial block face scanning electron microscopy (SBF-SEM) in collaboration with Benoit Zuber (Bern).
Ultrastructural analysis of T cell diapedesis across pMBMECs nder physiological flow in vitro. Representative image acquired with SBF-SEM of a CD4 T cell crossing the pMBMEC monolayer, obtained with the IMOD software. A significant part of the T cell nucleus has has squeezed below the thin endothelial monolayer. Yellow arrow indicates the direction of the flow, which goes from the viewer’s perspective to the next image acquired. Scale bar: 1µm.
In collaboration with Fabien Gosselet (Lens, France) and James McGrath (Rochester, USA) we have developed a breakthough microfluidic human in vitro cerebrovascular barrier (CVB) model featuring stem cell-derived brain-like endothelial cells (BLECs) and nanoporous silicon nitride (NPN) membranes (μSiMCVB, Mossu et al, JCBFM, 2018). The small size and superior imaging characteristics make the μSiMCVB highly suitable to study the interaction of rare patient-derived immune cells with the human BBB.
Figure: Interaction of human T cells (green) with the human blood-brain barrier in vitro. Red = endothelial junctions, blue = nuclei.
In a recent approach we have started in collaboration with Renaud Du Pasquier (Lausanne) and Eric Shusta (Madison-Wisocnsin, USA) to establish novel human in vitro neurovascular unit models from induced pluripotent stem cell (iPSCs) derived from multiple sclerosis (MS) patients and healthy controls. to investigate the pathogenesis of MS.
Role of neutrophils in reperfusion injury after ischemic stroke
Infiltration of neutrophils into the brain after ischemic stroke is considered central to reperfusion injury and thus increasing disability of patients suffering from ischemic stroke. Within a collaborative effort of the EU FP7 funded European Stoke network we have discovered that neutrophils do not migrate into the brain parenchyma after ischemic stroke but rather remain within the confines of the neurovascular unit (Enzmann et al., Acta Neuropathologica, 2013). In vitro observations on neutrophil interaction with the ischemic or cytokine stimulated BBB confirmed that ischemia does not trigger neutrophil extravasation across the BBB under flow (Gorina et al., JI 2014). In collaboration with Jan Klohs (Zurich, CH) we have developed near-infrared fluorescence (NIRF) imaging allowing for non-invasive monitoring of neutrophil accumulation in the brains of live mice after ischemic stroke and confirmed a role for a4-integrins in this process (Vaas et al, JCBFM, 2016). We furthermore showed that lack of ICAM-1 does not protect mice from experimental ischemic stroke (Enzmann et al., Transl Stroke Res, 2018). We propose that CNS immune privilege extends to sterile inflammation and thus does not lead to neutrophil extravasation into the CNS parenchyma.
CD45+ leukocytes are confined to the vascular compartment lined by pan-laminin stained basement membranes in the ischemic brain in the C57BL/6J mouse