The liver is a site at which apoptotic CD8+ cells accumulate during the clearance phase of peripheral immune responses. Normal mouse liver contains an unusual mixture of lymphocytes in which natural killer (NK) and NK-T cells are abundant and apoptotic T cells are present, and we interpret these cell populations as, respectively, agents and targets of an intrahepatic T-cell trapping and killing mechanism. In support of this idea, direct perfusion of activated lymphocyte populations through the normal liver results in the selective retention of activated CD8+ T cells. T cells trapped in this manner undergo apoptosis in the liver. This mechanism could explain the importance of the liver in oral tolerance, the phenomenon of tolerance induced by portal vein infusion of antigenic cells, the tolerance to allogeneic liver allografts, and the persistence of some liver pathogens including hepatitis C.

Elucidating cellular mechanisms that maintain the intrahepatic immune balance is crucial to our understanding of viral or autoimmune liver diseases and allograft acceptance. Liver sinusoidal endothelial cells (LSECs) play an important role in modifying local immune responses to tolerance in major histocompatibility complex (MHC) I-restricted models, whereas their contribution in the MHCII context is still controversial. In an MHCII chimeric mouse model that excludes MHCII-mediated antigen presentation by professional antigen-presenting cells, we demonstrated that LSECs prime CD4 ؉ T cells to a CD45RB low memory phenotype lacking marker cytokine production for effector cells that was stable in vivo following immunogenic antigen re-encounter. Although these cells, which we term T LSEC , had the capacity to enter lymph nodes and the liver, they did not function as effector cells either in a delayed-type hypersensitivity reaction or in a hepatitis model. T LSEC inhibited the proliferation of naïve CD4 ؉ T cells in vitro although being CD25 low and lacking expression of forkhead box protein (FoxP)3. Furthermore, these cells suppressed hepatic inflammation as monitored by alanine aminotransferase levels and cellular infiltrates in a T cell-mediated autoimmune hepatitis model in vivo. Conclusion: T LSEC first described here might belong to the expanding group of FoxP3 ؊ regulatory T cells. Our findings strengthen the previously discussed assumption that CD4 ؉ T cell priming by nonprofessional antigen-presenting cells induces anti-inflammatory rather than proinflammatory phenotypes. Because recruitment of CD4 ؉ T cells is increased upon hepatic inflammation, T LSEC might contribute to shifting antigen-dependent immune responses to tolerance toward exogenous antigens or toward endogenous self-antigens, especially under inflammatory conditions. (HEPATOLOGY 2009;50:1904-1913 U nder physiological conditions, the liver seems rather to support the induction of peripheral tolerance than to establish immunity. 1 In particular, liver sinusoidal endothelial cells (LSECs) are believed to shift the hepatic immune balance toward tolerance by presenting major histocompatibility complex (MHC) Irestricted antigens, thereby inducing anergy and tolerance. 2,3 LSECs constitutively express MHCII and costimulatory molecules 4 and take up exogenous antigens. 2 Naïve CD4 ϩ T cells migrate into the liver 5,6 and interact with LSECs. 7 This raises the question whether LSECs also prime CD4 ϩ T cells. Coculturing naïve CD4 ϩ T cells with LSECs resulted in stimulation of the T cells. 8 Another recent study using LSEC preparations depleted of cells of hematopoietic origin has created doubts about the priming capacity of LSECs for naïve CD4 ϩ T cells in the absence of professional antigenpresenting cells (APCs). 9 Priming of naïve CD4 ϩ T cells by other nonprofessional APCs such as aortic endothelial cells or naïve B cells induces distinct T cell populations lacking cytokine production and displaying suppressive capaci-

Activation and differentiation of lymphocytes have profound effects on their trafficking. Whereas naive T cells recirculate through lymphoid organs, activated cells localize predominantly in other compartments. Here, we report that changes in migratory properties of T cells occur immediately upon activation via the TCR. One hour stimulation is enough to target T cells into lung and liver following i.v. injection. The high localization within lung and liver and the lack of recirculation through lymphoid tissues are key features of activated lymphocytes. Regardless of the source, in vitro as well as in vivo activated lymphocytes show this behavior, which is not caused by increased cell size. Accumulation in the lung requires protein synthesis and is partly mediated by LFA‐1, in contrast to the acquisition of liver "homing" properties. Intravital microscopy reveals firm adhesion of activated cells within periportal sinusoids of the liver. Selective homing to other organs, such as skin or mucosa, was not observed, regardless of the cell's origin. These data indicate that activation quickly switches the trafficking program of lymphocytes from recirculation to sequestration; it is tempting to speculate that especially the induced trapping in the liver has a distinct role in limiting systemic T cell responses.

No abstract

The pathogenesis of autoimmune liver diseases is poorly understood. Animal models are necessary to investigate antigen presentation and priming of T-cells in the context of autoimmunity in the liver. Transgenic mouse models were generated in which the model antigen ovalbumin is expressed in hepatocytes (TF-OVA) or cholangiocytes (ASBT-OVA). A utoimmune hepatitis and cholangitis are triggered by autoreactive T-cells. Animal models are needed to study the early events in their pathogenesis, namely, the priming of autoreactive T-cells. The requirements for an immune-based animal model are restriction of the immune response to the liver, antigen specificity, and the potential to study the priming of CD8 and CD4 T-cells. Several animal models of autoimmune hepatitis have been developed, 1 none of which fulfills these criteria. Transgenic expression of foreign major histocompatibility complex (MHC) class I molecules in the liver has been widely used. [2][3][4] However, this model has a significant disadvantage: unlike the pathophysiology of an immune or autoimmune reaction, CD8 T cells recognize their antigen on hepatocytes but the antigen is not presented by professional antigen-presenting cells (APCs) in this model. Injection of the synthetic peptide ova p257-264 (SIINFEKL) into mice followed by transfer of antigenspecific CD8 T-cells has also been used, 5 but the possibility of peptide binding to MHC-I molecules on T-cells themselves and their activation by each other 6 or by unrelated cells is a possible concern. Infection with virus expressing ovalbumin 7 results in temporary expression of ovalbumin but also activation of innate immune mechanisms. No animal model exists that allows the simulta-

The liver is tolerogenic in many situations, including as an allograft and during the response to allogeneic MHC expressed on hepatocytes. The majority of data that address this issue focus on endogenous Ags. Little is known about CD4+ T cells and their fate under tolerizing conditions, especially with respect to fully differentiated CD4+ effector T cells. In this study, we used the adoptive transfer of populations of TCR-transgenic CD4+ T cells, skewed toward the Th1 or Th2 phenotype, to test whether either apoptotic or immune deviation mechanisms apply to cytokine-producing CD4+ T cells that enter the liver. After transfer, Th1 and Th2 cells could be detected up to 25 days in lymphoid organs and the liver. Intravenous high dose Ag application resulted in accumulation, proliferation, and subsequent deletion of effector cells within the liver. Th1 cells lost their capacity to produce cytokines, whereas IL-4 expression was sustained within Th2 cells from the liver. However, there was no evidence for a deviation of Th1-programmed cells toward a Th2 (IL-4) or regulatory T cell (IL-10) pattern of cytokine expression. We used isolated populations of liver-derived APCs to test whether the liver had the capacity to impose a bias toward IL-4 expression in T cells. These experiments showed that liver sinusoidal endothelial cells selectively suppress the expansion of IFN-γ-producing cells, yet they promote the outgrowth of IL-4-expressing Th2 cells, creating an immune suppressive milieu within this organ. These data suggest that presentation of Ags in the liver leads to modulation of immune response in terms of quantity and quality.

The endothelium plays an important role in the exchange of molecules, but also of immune cells between blood and the underlying tissue. The endothelial molecule S-Endo 1 antigen (CD146) is preferentially located at endothelial junctions and has been claimed to support endothelial integrity. In this study we show that the monoclonal antibody ME-9F1 recognizes the extracellular portion of murine CD146. Making use of ME-9F1 we found CD146 highly expressed and widely spread on endothelial cells in the analyzed murine tissues. In contrast to humans that express CD146 also on T cells or follicular dendritic cells, murine CD146 albeit at low levels was only found on a subset of NK1.1 + cells. The antibody against murine CD146 is useful for immunomagnetic sorting of primary endothelial cells not only from the liver but from various other organs. In vitro, no evidence was seen that the formation and integrity of endothelial monolayers or the transendothelial migration of T cells was aVected by antibody binding to CD146 or by crosslinking of the antigen. This makes the antibody ME-9F1 an excellent tool especially for the ex vivo isolation of murine endothelial cells intended to be used in functional studies.

We have previously shown that naïve CD8 ؉ T cells recognizing their cognate antigen within the liver are retained and undergo activation in situ, independent from lymphoid tissues. Intrahepatic primary T cell activation results in apoptosis and may play a crucial role in the ability of the liver to induce tolerance. Although adhesion molecules required for intrahepatic retention of T cells that have undergone previous extra-hepatic activation have been characterized, adhesive interactions involved in selective antigen-dependent intrahepatic retention of naïve CD8 ؉ T cells have not been investigated. By adoptively transferring radiolabeled T cell receptor (TCR)-transgenic CD8 ؉ T cells into recipient animals ubiquitously expressing the relevant antigen, we show that 40% to 60 % of donor antigenspecific naïve CD8 ؉ T cells were retained in the liver within 1 hour after transfer, despite ubiquitous expression of the antigen. Intravital microscopy showed that most donor naïve T cells slowed down and were irreversibly retained intrahepatically within the first few minutes after adoptive transfer, strongly suggesting that they were directly activated by liver cells in situ. This process was largely dependent on LFA-1 and ICAM-1, but was independent of blocking with antibodies against VCAM-1, ␣4 integrin, P-selectin, VAP-1, and ␤1 integrin. ICAM-2 seemed to play only a minor role in this process.