|Home | About | Journals | Submit | Contact Us | Français|
Studies on liver interstitial dendritic cells (DC) indicate that the maturation and function of these important antigen-presenting cells may be downregulated by continual exposure to microbial products from the gut, in particular, bacterial lipopolysaccharide. New evidence is emerging for a role of specific intracellular regulators of signal transduction, and of cytokines in the hepatic microenvironment, that may contribute to a hyporesponsive state in liver DC. Analysis of signaling molecule expression within DC in liver transplant tissue is likely to uncover its relation to allograft outcome.
Interstitial dendritic cells (DC), including those in the liver, are uniquely well-equipped migratory antigen (Ag)-presenting cells (APC), derived from CD34+ stem cells. They induce and regulate immune reactivity (1, 2). Although viewed traditionally as instigators of organ allograft rejection, donor-derived or host DC have also been implicated in transplant tolerance in experimental models (3, 4). This functional dichotomy of DC is governed by various factors, the most important of which appears to be their stage of differentiation/activation/maturation. We and others have shown that immature or `semi-mature' conventional myeloid (m)DC, deficient in surface co-stimulatory molecules (CD80, CD86, and inducible costimulator ligand), can inhibit alloAg-specific T cell responses. Such DC can prolong transplant survival across major histocompatibility barriers,- in some studies indefinitely, in the absence of immunosuppressive therapy (5-7). In humans, in vivo administration of autologous, Ag-pulsed, immature mDC can induce peptide-specific T regulatory cells (Treg) (8), and inhibit effector T cell responses to model Ags (9). Other DC subsets also have tolerogenic properties. Thus, plasmacytoid DC (pDC), an important source of type-I interferons (IFNs), can regulate T cell responses, including induction of Treg (4, 10, 11), and promote transplant tolerance (4).
The comparative ease with which liver allografts are accepted is well-known, but the mechanisms that underlie `hepatic tolerogenicity' remain unclear. We and others have described several important functional differences between normal mouse liver DC and those isolated concomitantly from secondary lymphoid tissue (12, 13). When compared to splenic DC, freshly-isolated liver DC produce lower levels of IL-12, secrete comparatively high levels of IL-10, and are inferior stimulators of naive allogeneic T cells (12-16). Adoptive transfer of liver-derived mDC progenitors to allogeneic hosts elicits IL-10 secretion by recipient T cells (17) and promotes transplant survival (18). Specific immunomodulatory cytokines secreted by unique cells in the liver (Kupffer cells, stellate cells and hepatocytes) may regulate liver DC function and T cell stimulatory ability. In particular, the cytokines IL-10 and transforming growth factor (TGF)β, expressed constitutively by several liver cell types and upregulated by others in response to activation or stress (19), appear to play important roles. Signaling cascades initiated by LPS, IL-6, IL-10 and TGFβ may shape the function of intrahepatic APC, particularly DC, to favor tolerance within the liver, and systemically.
We have shown that, unlike secondary lymphoid tissue DC, freshly-isolated murine liver DC are refractory to the exogenous Toll-like receptor (TLR) 4 ligand, bacterial lipopolysaccharide (LPS; endotoxin) suggesting that, in the normal steady-state, their continued in situ exposure to microbe-derived factors from the gut may result in a state of `endotoxin tolerance.' Activation of nuclear factor (NF)κB,- a critical determinant of DC maturation, is impaired (Fig. 1). Moreover, stimulation with LPS confers resistance (`cross tolerance') to subsequent stimulation with other TLR ligands, e.g. the TLR9 ligand, CpG (13). We have postulated that constant exposure of DC to LPS within the hepatic microenvironment further attenuates TLR4 and TLR9 signaling, invoking a mechanism(s) that prevents chronic liver inflammation. Taken together, these phenomena may contribute to the altered function of liver DC and to inherent liver tolerogenicity.
Mechanisms underlying endotoxin tolerance that have been described in macrophages, include elevated expression of negative regulators of the TLR signaling cascade (20). However, mechanisms underlying endotoxin tolerance in DC in general, and in liver DC, in particular, are poorly defined. Our recent findings show that the signaling adaptor DNAX-activating protein of 12KDa (DAP12), that can mediate inhibition of TLR activation, is upregulated in normal liver DC (Sumpter TL et al, in preparation) and that intrahepatic IL-6/signal transducer and activator of transcription (STAT)3 signaling inhibits liver DC maturation and function (21). Consequently, we hypothesize that LPS in the liver may induce negative regulators of TLR signaling, resulting in endotoxin tolerance and impaired T cell allostimulatory capacity of liver DC. TGFβ, IL-6, and IL-10 may potentiate endotoxin tolerance through molecular `cross-talk' between their respective signaling pathways (Smad for TGFβ, and STAT3 for IL-6 and IL-10) and the TLR signaling pathway. We further suggest that the tolerogenic phenotype attributed to liver DC, and their resistance to maturation, may regulate alloreactive T cell responses in vivo and contribute to the induction/maintenance of allograft tolerance (Fig 2).
TLR (TLR1-13 in mice; TLR1-10 in humans) recognize conserved bacterial and viral motifs. The result of ligand binding and TLR activation is inflammation, mediated by the transcription factors, NFκB and interferon (IFN) regulatory factor (IRF) (22-24). TLR4, the most studied TLR, induces NFκB activation and secretion of pro-inflammatory cytokines, such as IL-12, in response to LPS. After primary TLR4 activation, the net response upon subsequent LPS stimulation is dampened by negative regulatory proteins to prevent chronic inflammation (endotoxin tolerance). Cross-talk through common adaptor and regulatory molecules can result in ensuing downregulation of signaling cascades initiated by other TLR ligands (cross tolerance). Impaired NFκB activation, such as that associated with endotoxin tolerance, may generate tolerance-promoting DC. Thus DC in which NFκB is rendered inactive using NFκB decoy oligodeoxynucleotides, can promote tolerance to mouse heart (25, 26) or rat partial liver allografts (27).
Numerous proteins have been identified in macrophages as negative regulators of TLR signaling involved in endotoxin tolerance (14, 28). DAP12 is a candidate adaptor molecule for controlling endotoxin tolerance in liver DC. Notably, DAP12-deficiency is associated with increased susceptibility to endotoxemia in mice (29). In pDC (relatively abundant in the liver), as well as in macrophages, DAP12 downregulates activation in response to TLR9 (30) and TLR4 ligation (29). In airway CD11c+ cells (presumptive DC), DAP12 blocks NFκB activation in response to bacterial infection (31). However, in functionally mature, murine CD8α+ DC, DAP12 blocks activity of indoleamine dioxygenase (IDO), an immunosuppressive enzyme (32). The discrepancy between the role of DAP12 in IDO regulation and TLR responses may reflect differences in DC subsets, maturation status, or interactions with other signaling pathways. IL-1R-activated kinase (IRAK)-M is an IL-1R family member that functions as an inducible regulator of TLR signaling in macrophages. It has been associated with endotoxin tolerance in an endotoxemia model (33). These, as well as other negative regulators of TLR signaling (20), may play a key role in endotoxin tolerance in liver DC subsets.
A prevailing paradigm is that DC maturation status controls the induction of T helper type-2 (Th2) cells or Treg and subsequent tolerance, with immature DC able to promote tolerance, and mature DC equipped for induction of immunity. The transcriptional program that controls DC maturation, and the subsequent induction of tolerance is multivariate, and includes but is not limited to members of the STAT family and the canonical pro-inflammatory transcription factor, NFκB.
STAT family members purported to play a role in DC maturation include STAT1 and 3. STAT1 transduces signals initiated by the IFNγ receptor, and amplifies IL-12p40 production induced by NFκB (34). STAT1-deficient mice have impaired Th1 responses dependent on DC but not T cells. Analysis of bone marrow-derived DC from STAT4-deficient mice has revealed impaired expression of MHC class II, CD40, CD80 and CD86 (35). STAT1 activation in DC is controlled by suppressor of cytokine signaling (SOCS)-1. DC lacking SOCS-1 not only have higher STAT1 activity, but are more potent inducers of Th1 responses (36).
STAT1 drives DC maturation, but also the induction of IDO (37). IDO, in turn, downregulates T cell proliferation and can drive activation of Treg (38). In the liver, IDO is expressed by APC during tolerance induction, but expression tapers when tolerance is achieved (39). Moreover, blocking IDO activity early after transplantation prevents long-term, rat liver allograft survival, but does not cause acute rejection (40). STAT1-driven IDO may be a crucial negative regulator of Th1 responses for maintaining liver tolerance.
STAT3 activation is driven by multiple cytokines, including IL-6 and IL-10, and prevents LPS-induced DC maturation (41). IL-10-driven STAT3 activation blocks NFκB access to the IL-12p40 promoter (42). IL-6-driven STAT3 activation blocks CD80 and CD86 expression on liver CD11c+ cells (21). Gut microbial products drive the IL-6/STAT3 axis in the liver, thereby blocking maturation of liver DC (21).
As described above, prior exposure to LPS induces a transient state of cellular hyporesponsiveness to subsequent LPS stimulation known as endotoxin tolerance, that represents a reprogramming of cells, possibly as a means of adaptation to bacterial infection. It is clear that in such cells tolerance does not represent global inhibition of endotoxin-driven functions. Thus, although LPS-induced production of diverse cytokines (e.g. TNFα, IL-12) is suppressed, LPS-induced production of anti-inflammatory mediators (eg. IL-1 receptor antagonist or IL-10) is unaltered (43, 44). Similar effects are seen after priming with other TLR ligands, or with mixtures of pro-inflammatory cytokines (45). The molecular mechanisms underlying endotoxin tolerance appear to be multifactorial. Changes in LPS receptor complex expression or function, alterations in TLR-driven signaling pathways or pathways of feedback inhibition, and/or primary effects on transcription, mRNA stability and translational efficiency have all been implicated (41, 45-48).
Despite the significant levels of LPS in normal liver, several groups have confirmed that freshly-isolated liver DC are phenotypically immature (12, 14, 28). Anti-inflammatory cytokines produced in the liver (constitutively, and in response to cell activation/stress), such as TGFβ and IL-10, may keep liver DC immature/confer maturation resistance and play a role in endotoxin tolerance. The TGFβ ligand-receptor interaction results in phosphorylation of the receptor-activated Smads (Smad 1,2,3,5 and 8), which translocate to the nucleus with Smad4 to induce transcription (49). Global loss of TGFβ signaling in mice is associated with endotoxin hypersensitivity (50), though the cell types involved are not known. TGFβ blocks NFκB activation following ligation of TLR2, 4 or 5, and downregulates the TLR adaptor protein MyD88, interfering with signaling initiated by TLR2, 4 or 5 (51). STAT3 directly antagonizes NFκB, and may also play a role in endotoxin tolerance (52). STAT3 is activated by a number of cytokines found in the liver, including IL-6 and IL-10. We have shown (21) that by activating the IL-6/STAT3 axis, gut microbial products inhibit liver DC activation/maturation and thereby elevate the threshold needed to translate triggers of innate immunity into adaptive immune responses. TGFβ and IL-10 act synergistically to induce endotoxin tolerance in monocytes, even in the absence of LPS (53). Constitutive TGFβ and IL-10 production, as seen in the liver, may play an important role in the induction of endotoxin tolerance in liver DC.
Very little information is available on DC phenotype and function in human liver grafts/biopsies, especially in relation to liver allograft outcome. We have recently begun applying advanced imaging techniques to liver biopsy tissue to gain important insights into the expression of factors by liver DC that may be key to regulation of alloimmunity and graft outcome. These novel methods have enabled us to examine liver DC maturation in human liver allografts, using archived formalin-fixed, paraffin-embedded (FFPE) tissue.
Conventional light microscopy-based immunostaining of tissue has a limited capacity for multiplex analysis, due to the difficulty in discriminating overlapping chromagens. Traditional immunofluorescent staining is better suited for multiplex immunostaining, however the use of FFPE liver tissue is hindered by endogenous autofluorescence, especially in liver tissue, that can mask the imaging of traditional probes, such as FITC, CY3, or Texas Red. We have employed new techniques to overcome the artifacts that limit immunofluorescence analysis in FFPE tissues. First, we have integrated fluorescent nanoparticle colloidal semi-conductor quantum dots (Qdots) into our established standard Ab-based immunostaining techniques. Qdots have significantly brighter fluorescent emission than traditional fluorescent probes and, therefore, can overcome the endogenous autofluorescence in FFPE tissues (54). Additionally, multiple Qdots are commercially available with narrow emission spectra from a single excitation source which permits simultaneous analysis of several Ags in a single tissue section. Second, image acquisition using an advanced liquid crystal tunable filter CCD camera system and supporting software (Nuance 2, CRI Inc, Woburn MA) to detect Qdot fluorescence helps facilitate simultaneous analysis of multiple Ags. This imaging system and supporting software rapidly captures images at 10nm intervals and can discriminate and subtract endogenous tissue autofluorescence from specific Qdot fluorescence signals.
We have now begun to use multiplex Qdot immunostaining to investigate intrahepatic DC, examining CD11c (mDC), CD83, HLA-DR and phosphorylated STAT3 (pSTAT3) in human liver allografts (Fig. 3). In these preliminary results we can identify CD11c+ HLA-DR+ DC expressing pSTAT3, but little CD83. Thus it is now possible to relate expression of regulators of DC maturation in human liver biopsies in relation to allograft outcome.
Freshly-isolated murine liver DC, before or after stimulation with TLR ligands, show impaired NFκβ phosphorylation and inferior ability (compared with secondary lymphoid tissue DC) to secrete bioactive IL-12 and to induce IFNγ production and proliferation in allogeneic T cells. This deficiency is associated with enhanced expression by liver DC, compared with spleen DC, of negative regulators of TLR signaling (DAP12 and IRAK-M), and with inhibitory IL-6/STAT3 activity. LPS-stimulated liver DC can induce alloAg-specific T cell anergy. Using advanced imaging technology, we can now visualize DC in greater detail and their expression of maturation markers in human liver transplant biopsies (including tissue from tolerant patients). We can also quantify the expression of transcription factors that regulate DC maturation (e.g. STAT3) within these important APC. It will now be important to ascertain the relation between expression of these molecules by liver DC in situ and liver transplant outcome.
Supported by National Institutes of Health (NIH) grant R01 AI60994 and the Roche Organ Transplantation Research Foundation (874279717). TLS is in receipt of a Basic Science Fellowship from the American Society of Transplantation and a Postdoctoral Research Fellowship from the American Liver Foundation, and JGL is supported by NIH T32 AI74490