Anatomic Considerations

Because the vascular supply of the liver derives principally from the portal venous system draining the gut, a heavy antigen load is delivered to the liver, particularly after meals. Portal venous blood flows slowly through the vast network of hepatic sinusoids, which are discontinuously lined by fenestrated endothelium lacking a basement membrane (Fig. 9.4; see also Chapter 6). The sluggish flow of blood allows for the efficient capture of antigens by leukocytes traveling in the blood within the sinusoids and by the endothelial cells lining the sinusoids. The microscopic anatomy of the liver also favors the ability of blood-borne leukocytes to interact with hepatic parenchymal cells and resident immune cells of the liver.

FIGURE 9.4 Microvascular hepatic anatomy. A, The liver is organized as lobules defined by their relation to portal vascular bundles and central veins. B, Liver sinusoidal endothelial cells (LSECs) line the hepatic sinusoids and have fenestrated membranes. A variety of immune cells exist within the sinusoids and are able to traverse the sinusoidal membrane to enter and exit the space of Disse, which is in contact with hepatocytes.

(From Crispe IN, 2003: Hepatic T cells and liver tolerance. Nat Rev Immunol 3:51-62.

Tolerance Versus Immunity

One of the most intriguing aspects of liver immunology is the propensity of antigens passing through the liver to produce tolerance rather than immunity. In experimental animal models and clinical studies of humans undergoing liver transplantation, a greater propensity for graft acceptance has been noted compared with transplantation of other solid organs. A liver transplant protects a kidney allograft transplanted simultaneously from the same donor (Creput et al, 2003). Trying to decipher the reasons for the liver’s inherent propensity for tolerance has been a particularly active area of research in the field of transplantation and cancer immunobiology.

Multiple competing theories to explain the cause of liver tolerance have been advanced. One postulate is that, as in central thymic tolerance, clonal deletion of antigen-specific T cells occurs in the liver. Another theory is that liver allografts release large amounts of soluble MHC class I molecules that potentially could function to neutralize donor-specific antibodies or cytolytic T cells. Kupffer cells (KCs) and natural killer (NK) T cells have been implicated as being important in the development of hepatic tolerance, because the depletion of either has been associated with a loss of oral tolerance. Liver DCs also have been suggested to play a role in liver tolerance, because as a group they have been shown to be less immunogenic than DCs from lymphoid organs such as the spleen (Bamboat et al, 2009). In addition, the numbers of several unusual cell types are elevated in the liver, and more recent studies have shown that multiple lineages of intrahepatic leukocytes are unique in function and phenotype compared with their counterparts from lymphoid organs. Controversy exists, however, as to whether these differences are due to de novo development, conditioning within the liver, or chemotactic signals in the liver that attract a unique population of cells. Regardless, the differences in composition and function are likely to play an important role in defining the nature of liver immunology. We devote a portion of this chapter to the major populations of liver immune cells and their role in various liver diseases.

Antigen-Presenting Cells

The experimental observation that antigens passing through the liver can lead to tolerance makes the understanding of antigen presentation in the liver particularly germane (Li et al, 2004). Antigen-presenting cells (APCs) play a crucial role in determining the nature of adaptive immune responses. The manner in which an APC presents any particular antigen can alter dramatically the response of antigen-specific T cells. Specifically, when antigen presentation occurs in conjunction with the appropriate costimulatory molecules, T cells proliferate and develop an immunogenic phenotypic and functional profile (Fig. 9.6). In contrast, antigens presented in the absence of costimulation lead to anergy or activation-induced T-cell death, two of the mechanisms of peripheral tolerance induction and maintenance.

FIGURE 9.6 Antigen-presenting cells (APCs) and T-cell activation. The three main APCs are dendritic cells (DCs), macrophages, and B cells. The immune response depends on the context in which antigens are presented to T cells. The type of APC and the presence or absence of costimulatory molecules are important in determining whether a T cell has no response (anergy) or is activated.

(From Abbas AK, 2003: Cellular and Molecular Immunology. Philadelphia, Saunders.)

Dendritic Cells

Dendritic cells (DCs) are rare, heterogeneous leukocytes that are primarily responsible for the capture of antigens in the periphery and subsequent presentation to immune effector cells. DCs are now recognized as the most potent APCs of the immune system and seem to arise from multiple hematopoietic cell lineages. Murine DCs are typically isolated based on their expression of the integrin CD11c. Immature DCs are specialized to capture antigens and then migrate to lymph nodes, where they can interact with T cells. After an encounter with pathogenic stimulus, such as bacterial lipopolysaccharide, DCs undergo phenotypic and functional changes, whereby their ability to capture antigens is diminished, but they increase their expression of MHC class II and T cell costimulatory molecules. Costimulatory molecule expression is essential in facilitating antigen presentation and efficient T cell activation. The ability of DCs to capture antigens in the periphery, translocate to areas rich in T cells, and then present antigens places DCs in a central role in immune regulation, bridging innate and adaptive immunity.

The body of literature studying the function of DCs grown in vitro from murine bone marrow progenitors or human peripheral blood mononuclear cells is vast. DCs isolated directly from the spleen, lymph nodes, or thymus also have been well studied. In contrast, because of the rarity of DCs in the liver, few studies have shown the phenotype or function of liver DCs. Initial studies focused on DC progenitors grown from liver nonparenchymal cells cultured with various cytokines, including granulocyte-macrophage colony-stimulating factor (GM-CSF). These early studies showed that liver DC progenitors are relatively immature cells with poor immunostimulatory ability that can promote immunologic tolerance and prolong solid organ transplants in animal models (Lu et al, 1994, 2001; Lau & Thomson, 2003).

To gain a better understanding of liver DCs as they function in situ, recent focus has been on studying freshly isolated liver DCs. In our initial studies, we wished to increase the number of liver DCs to facilitate their isolation. Because GM-CSF had been used to grow DC progenitors in culture from liver nonparenchymal cells, and because we had previously shown that in vivo GM-CSF overexpression increases the number of myeloid DCs in the spleen (Miller et al, 2002a), we suspected that it might expand liver DCs as well. We found that overexpression of GM-CSF using an adenoviral vector led to a dramatic increase in the number of highly immunostimulatory liver DCs (Pillarisetty et al, 2003). We also showed that a large proportion of the cells recruited to the liver by GM-CSF were DC precursor cells that develop into CD11c⁺ DCs in culture, which suggests that DC development can occur in the liver. We used immunohistochemistry to locate both types of recruited cells. The liver DCs were concentrated around the central veins, whereas the DC precursors were distributed diffusely throughout the liver parenchyma.

Our experience with expanding liver DCs sheds light on in situ liver DC development, but we wished to understand the phenotype and function of liver DCs in normal, untreated mice. When our ability to isolate liver DCs from normal mice improved, we began to study them in depth. We found that compared with the relatively well-studied DCs from the spleen, CD11c⁺ liver DCs were immature and only weakly immunostimulatory (Pillarisetty et al, 2004). We also noticed that, in contrast to spleen DCs, liver DCs are heterogeneous in their expression of MHC class II and costimulatory molecules. We further separated CD11c⁺ liver DCs based on their expression of the myeloid marker CD11b and the lymphoid marker CD8α⁺, which have been commonly used to define murine DC subtypes. Myeloid (CD11b⁺) and lymphoid (CD8α⁺) liver DCs, which each comprise approximately 10% of the total population of DCs in the liver, were as able to activate T cells as were their splenic counterparts. The bulk of the remaining cells, which had low to no expression of CD11b and CD8α, were only poorly immunostimulatory. This study showed that the presence of these atypical DCs was primarily responsible for the weakly activating nature of liver DCs on the whole. More recently, we discovered that despite having multiple APCs, the CD11c^hi subset of DCs within the liver are required for effective presentation of soluble protein (Plitas et al, 2008). Using a transgenic mouse in which CD11c^hi DCs can be depleted selectively, we found that activation of antigen-specific CD8⁺ T cells in the liver only occurred in the presence of CD11c^hi DCs.

Our investigation into liver DCs has now been extended into humans. We found that freshly isolated DCs from human liver exhibit tolerogenic properties when compared with autologous blood DCs. Liver DCs are weaker stimulators of T cells and favor the production of antiinflammatory IL-10, which induces the differentiation of naïve CD4⁺ T cells into regulatory T cells with suppressive function (Fig. 9.7; Bamboat et al, 2009a).

FIGURE 9.7 Human liver dendritic cells (DCs) are tolerogenic. A, Freshly isolated human liver and autologous blood DCs were cultured in media alone or in the presence of lipopolysaccharide (LPS). Supernatant interleukin-10 (IL-10) was measured 24 hours later. B, Liver and blood DCs were cultured with allogeneic naïve CD4⁺ T cells in the presence or absence of an anti–IL-10 blocking antibody (αIL-10). After 5 days of culture, the T cells were restimulated using antibodies against CD3 and CD28. The percentages of CD4⁺FoxP3⁺ cells were determined 24 hours later using flow cytometry. C, Autologous antigen-specific CD4⁺ T cells recognizing a pool of human leukocyte antigen (HLA) class II–restricted peptides (peptides) were generated by two rounds of stimulation with either liver or autologous blood DCs. The frequencies of interferon (IFN)-γ–producing CD4⁺ T cells after 24 hours of culture with peptide-pulsed target cells (APCs + peptides + T) were then determined by enzyme-linked immunosorbent spot-forming analyses. Antigen-presenting cells (APCs) and T cells cultured alone made no detectable IFN-γ (not shown). *P < .05; **P < .01; ***P < .001.

Kupffer Cells

Liver macrophages, referred to as Kupffer cells (KCs), have long been believed to be the primary phagocytic cells of the liver. KCs represent the largest pool of macrophages in the body, derived from monocytic precursors in the blood. They are typically found in the hepatic sinusoids; however, they also can migrate through the space of Disse to interact with hepatocytes (see Fig. 9.4). KCs have been thought to play a major role in antigen presentation and have been implicated in portal venous tolerance. Evidence also suggests that KCs may regulate T-cell responses to antigens and induce immune tolerance in the setting of liver allografts (Sun et al, 2003).

Techniques that allow for the more precise delineation of cellular subtypes have called into question some of the concepts about KCs and many other traditionally defined immune cells. Specifically, the advent of flow cytometry has allowed us to define cells by their surface marker expression, rather than simply by size or other crude physical characteristics shared by many cell types. This point is illustrated by the fact that many liver DCs express the cell surface marker CD11b, which is typically expressed by cells of monocytic lineage and has been used to define KCs in the liver. Much of the antigen-presenting ability previously attributed to KCs may be a result of DC function. The lack of consistent definitions of cells by cell surface marker expression and the continued use of less precise methods of cell isolation have led to continued ambiguity regarding KCs. An alternative explanation for the overlap in functional definitions for DCs and KCs is that DCs represent a more differentiated state of the KC lineage.

Liver Sinusoidal Endothelial Cells

Liver sinusoidal endothelial cells (LSECs) are highly specialized cells that line the hepatic sinusoids. LSECs are distinguished by the presence of fenestrations, or windows, in their cellular membranes. These fenestrations are believed to facilitate the selective passage of antigens between the sinusoid and the hepatic parenchyma. This strategic placement puts LSECs in the ideal position to interact with antigens and immune cells passing between the liver and the portal venous system.

Several studies have shown that in addition to serving as the building blocks lining the sinusoids, LSECs are immune cells with the ability to capture and present antigen and activate T cells (Knolle & Limmer, 2001). As with KCs, considerable controversy surrounds the immunologic function of LSECs. In contrast to earlier work, we have shown more recently that although LSECs are highly capable of capturing various antigens in vivo and in vitro, they lack the ability to activate T cells in the absence of exogenous costimulation (Katz et al, 2004). The differences in results may derive from the use of more specific methods of cell isolation in the latter study. The finding that LSECs are not independently capable of triggering a T cell–mediated immune response does not, however, exclude the possibility that LSECs, in concert with DCs or KCs, play an important role in antigen presentation in the liver.

Effector Cells