Abstract

T cells recognizing myelin auto-antigens penetrate into the CNS to induce inflammatory autoimmune disease following complex sequential interactions with individual components of the vascular blood-brain barrier (BBB), particularly endothelial cells, and perivascular phagocytes. To determine the functional consequences of these processes, two-photon intravital imaging was performed to compare the behavior of three myelin-specific GFP-expressing T cell lines with different potentials for transferring Experimental Autoimmune Encephalomyelitis. Imaging documented that, irrespective of their pathogenic potential, all T cell lines reached the CNS and interacted with vascular endothelial cells indistinguishably, crawling on the luminal surface, preferably against blood flow, before crossing the vessel wall. In striking contrast, after extravasation the T cell motility and their interactions with perivascular antigen presenting cells (APCs) varied dramatically. While highly encephalitogenic T cells showed a low motility, made stable contacts with local APCs and became activated, the corresponding contacts of weakly encephalitogenic T cells remained short, their motility high and their activation marginal. Supplying auto-antigen, via either local injection or by transfer of antigen-pulsed meningeal APCs, lowered their motility and prolonged the contact duration of weakly encephalitogenic T cells to values characteristic for highly pathogenic ones. Only after exogenous antigen supply, the weakly encephalitogenic T cells became activated, infiltrated the CNS parenchyma, and triggered clinical EAE, suggesting that the strength of the antigen-dependent signals received by immigrating effector T cells from leptomeningeal APCs is crucial for their pathogenic effect within the target tissue. To directly correlate the activation of encephalitogenic T cells with their dynamic behavior in the CNS, a truncated fluorescent derivative of nuclear factor of activated T cells (NFAT) was introduced as a real-time activation indicator. Two-photon imaging documented the activation of the auto-reactive T cells extravasated into the perivascular space, but not within the vascular lumen. Activation correlated with reduced T cell motility, and it was related to contacts with the local APCs. However, it did not necessarily lead to a long-lasting arrest, as individual, activated T cells SUMMARY 2 were able to sequentially contact other APCs. A spontaneous cytosol-nuclear translocation of the marker was noted only in T cells with a high pathogenic potential. The translocation implied the presentation of an auto-antigen, as the weakly pathogenic T cells, which remained silent in the untreated hosts, were activated upon the instillation of exogenous auto-antigen. It is proposed here that the presentation of local auto-antigen by BBB-associated APCs provides stimuli that guide autoimmune T cells to the CNS destination and enable them to attack the target tissue. In addition, a theoretical, physicist approach was used for modeling T cell activation in the leptomeningeal space. Assuming that T cells have evolved to gain their activation signal in a way that is energetically optimal for them, two possible scenarios for T cell activation were compared. The first one assumes that, after finding an APC presenting the epitope of interest, the T cell will stop and interact with the APC until it becomes fully activated. The second model considers the possibility that a T cell can accumulate activation signals from different APCs while scanning them without stopping, until a certain threshold is exceeded and the T cell becomes activated. Using this approach, it is proposed that the T cells in EAE are more likely to become activated following the first scenario. However, in a more natural environment such as a lymph node, the second scenario could give them some advantages