Potential conflict of interest: Nothing to report.
Supported by National Institutes of Health grants R01DK71753 (to N.M.), P01AI81678 (to A.W.T., D.A.G., D.B.S., A.J.D., N.M.), ASTS Genentech Presidential Student Mentor Award (to M.Z.), and the Ruth L. Kirschstein National Research Service Award T35 DK065521 (to M.Z.).
Present address for Antonino Castellaneta: Gastroenterology Unit, University of Bari, Italy.
Dendritic cells (DCs) induce and regulate both innate and adaptive immune responses; however, their in vivo functional importance in hepatic ischemia/reperfusion (IR) injury is perplexing. We hypothesized that liver-resident DC and locally recruited blood-borne DC might have distinctive roles in hepatic IR injury. We tested this hypothesis by using DC-deficient, fms-like tyrosine kinase 3 ligand (Flt3L) knockout (KO) mice in hepatic warm (70% partial clamping for 60 minutes) and cold IR injury (liver transplant [LTx] with 24-hour cold storage). Flt3L KO liver and lymphoid organs contained virtually no CD11c+F4/80− DC. Hepatic warm IR injury was significantly lower in Flt3L KO than in wildtype (WT) mice with lower alanine aminotransferase (ALT) levels, reduced hepatic necrosis, and lower neutrophil infiltration. Hepatic messenger RNA (mRNA) and protein levels for inflammatory cytokines (tumor necrosis factor alpha [TNFα], interleukin [IL]-6) and chemokines (CCL2, CXCL2) were also significantly lower in Flt3L KO than in WT mice, indicating that lack of both liver-resident and blood-borne DC ameliorated hepatic warm IR injury. Adoptive transfer of splenic or hepatic WT DC into Flt3L KO or WT mice increased hepatic warm IR injury, suggesting injurious roles of DC infusion. When Flt3L KO liver was transplanted into WT mice, ALT levels were significantly higher than in WT to WT LTx, with enhanced hepatic necrosis and neutrophil infiltration, indicating a protective role of liver-resident DC. Conclusion: Using both warm and cold hepatic IR models, this study suggests differential roles of liver-resident versus blood-borne DC, and points to the importance of the local microenvironment in determining DC function during hepatic IR injury. (HEPATOLOGY 2013;57:1585–1596)
Liver ischemia reperfusion (IR) injury is a clinically relevant condition that occurs during liver transplantation (LTx), trauma, elective liver surgeries, and incidental events such as hypovolemic shock, and results in hepatic dysfunction or failure. In LTx, IR injury is the major contributor to posttransplant liver graft dysfunction and patient morbidity. The current shortage of liver grafts available for transplantation has led to the use of organs from extended criteria donors and increased the severity of IR injury and risks of liver graft dysfunction.1, 2 IR injury is a profound proinflammatory event initiated by the innate immune system, one that involves neutrophil and macrophage recruitment and inflammatory cytokine and chemokine production, resulting in hepatic parenchymal damage and liver dysfunction. Improved understanding of the mechanisms underlying hepatic IR may lead to more effective therapeutic intervention to minimize liver dysfunction due to IR injury.
Dendritic cells (DCs) are potent antigen-presenting cells (APCs) that induce and regulate adaptive immune responses. In the field of transplantation their roles and functions have been studied intensively in the context of allograft rejection and tolerance.3-6 However, DCs are also known to play significant roles in innate immune reactions. Potent proinflammatory responses in IR injury are mediated by the innate immune system by way of recognition of endogenous danger signals, damage-associated molecular patterns (DAMPs), by pattern recognition receptors (PRRs). DCs express a large repertoire of PRRs and promptly respond to signals by profound phenotypic change and functional maturation. After direct activation by DAMPs, DCs secrete proinflammatory cytokines, chemokines, and other mediators. In addition, DCs have an additional specialized ability as APC to initiate adaptive immune responses; activated DCs express high levels of major histocompatibility complex (MHC) molecules bearing pathogen-derived peptides to potently activate antigen-specific T cells and mount adaptive immune responses.7, 8
Roles of DC in hepatic IR injury have been studied previously by manipulating DC populations in genetically modified mice; however, the findings have been rather ambivalent. In the hepatic partial warm IR model, DC depletion in CD11c-diphteria toxin receptor (DTR) transgenic mice resulted in increased liver injury, and adoptive transfer of conventional DC from wildtype (WT), but not Toll-like receptor (TLR)9 knockout (KO) or interleukin (IL)-10 KO mice, reduced injury, suggesting protective roles of DC in hepatic IR through TLR9-mediated IL-10 secretion.9 However, using the same hepatic warm IR model, increasing DC numbers with exogenous granulocyte-macrophage colony-stimulating factor (GM-CSF) has been shown to enhance liver injury by way of necrotic cell released high-mobility group box 1 (HMGB1) and TLR4 activation,10 indicating detrimental roles of DC. In view of these contradictory reports on the role of DC in early innate immune responses during hepatic IR injury, we hypothesized that liver-resident DC and recruited blood-borne DC might have distinct roles in hepatic IR injury. Tissue-resident DC display unique properties that allow them to maintain immune homeostasis despite sampling and sensing danger signals.11 In particular, liver-resident DC are known to maintain an immature status, despite constant exposure to bacterial products released from gut microbes.12, 13
Fms-like tyrosine kinase 3 ligand (Flt3L) is a potent endogenous hematopoietic growth factor, and Flt3L KO mice exhibit profound reductions in DC in the spleen, lymph nodes, and thymus.14 In addition, in vivo administration of Flt3L to mice and humans causes dramatic increases in various DC subsets in lymphoid and nonlymphoid organs, including the liver.15-17 Modification of DC populations in Flt3L KO mice or with Flt3L administration has been used previously to examine the functional roles of DC.18-21 Although the Flt3L KO mouse also shows mildly decreased leukocyte cellularity and reduced natural killer (NK) cells,14 they provide a useful model for examining the role of DC in innate immune responses. Accordingly, we tested our hypothesis by investigating the role of DC in both hepatic warm IR and LTx-induced cold IR models using mice lacking Flt3L. The results obtained suggest that liver-resident DC may play distinct roles, compared to inflammatory, recruited blood-borne DC or lymphoid tissue DC in acute hepatic IR injury.
Collagenase B was purchased from Boehringer (Ridgefield, CT). A trypsin inhibitor, bovine serum albumin, ethylene diamine tetraacetic acid, ethylene glycol tetraacetic acid, and histodenz were obtained from Sigma (St. Louis, MO). Ca+Mg+-free Hank's balanced salt solution (HBSS), 4-(2-hydroxyethyl)-1-piperazine ethanesulfonic acid (HEPES) buffer, Roswell Park Memorial Institute (RPMI)-1640 medium, L-glutamine, fetal bovine serum, and gentamicin were acquired from Life Technologies (Grand Island, NY). FITC-, PE-, PECy5-, PECy7-, or Pacific Blue-conjugated monoclonal antibodies (mAbs) directed against mouse CD3 (145-2C11), CD4 (H129.19), CD8 (53-6.7), CD11c (HL3), CD11b (M1/70), CD40 (3/23), CD45 (30-F11), CD45.1 (A20), CD80 (16-10A1), CD86 (GL1), programmed death ligand-1 (PD-L1) (MIH5), F4/80 (BM8), B220/CD45R (RA3-6B2), IAb β-chain (MHC class II, 25-9-17), or NK1.1 (pk136), as well as appropriate immunoglobulin isotype controls, were obtained from eBiosciences (San Diego, CA) or BD Bioscience (San Diego, CA). Immunomagnetic microbeads for mouse (m)PDCA-1+ and CD11c+ cell isolation were obtained from Miltenyi Biotec (Auburn, CA).
Flt3L KO mice (Taconic Farms, Germantown, NY) on a C57BL/6 background (B6) and WT B6 and B6 CD45.1 mice (Jackson Laboratory, Bar Harbor, ME) were housed under specific pathogen-free conditions and used between 8 and 14 weeks of age. All procedures described were performed according to the guidelines of the National Research Council's Guide for the Humane Care and Use of Laboratory Animals, and approved by the Institutional Animal Care and Use Committee at the University of Pittsburgh.
Orthotopic Liver Transplantation (LTx).
Under isoflurane anesthesia, orthotopic LTx without hepatic artery reconstruction was conducted as described.22 Liver grafts were perfused with 1.0 mL of University of Wisconsin (UW) solution by way of the portal vein, stored in UW solution for 24 hours at 4°C, and implanted orthotopically by anastomosing the suprahepatic vena cava with a running 10-0 suture and the portal vein and inferior vena cava with the cuff technique. The bile duct was connected by way of ligation over the stent. Recipients were euthanized at 3 and 12 hours after LTx (n = 5-9 for each group at each timepoint).
Warm IR Model.
Nonlethal segmental (70%) hepatic warm ischemia was performed for 60 minutes under ketamine/xylazine anesthesia.23 Animals in the sham-operated group (n = 3 for each) were given anesthesia and underwent exposure of the liver hilum without ischemia. Animals were euthanized at 3, 6, and 12 hours after reperfusion (n = 6-9 for each group at each timepoint).
Liver Damage Assessment.
Serum alanine aminotransferase (ALT) levels were measured at the clinical laboratory of the University of Pittsburgh Medical Center.
Isolation of Hepatocytes and Nonparenchymal Cells (NPCs).
Hepatocytes and hepatic NPCs were isolated from the liver by the collagenase digestion method.24 Briefly, each liver was perfused in situ by way of the infrahepatic inferior vena cava, initially with Ca+Mg+-free HBSS containing 5 mM ethylene glycol tetraacetic acid and 10 mM HEPES, and then with HBSS containing 0.025% collagenase B, 5 mM HEPES, 56 mg calcium dichloride, and 0.005% trypsin inhibitor. NPC and parenchymal cells were liberated from the excised liver, and the initial cell suspension filtered through a 70-μm nylon mesh. Hepatocytes and NPC were separated by low-speed centrifugation (5 × 45g for 5 minutes) and washed by high-speed centrifugation (150g for 10 minutes).
Isolation of Liver and Spleen DC Subsets and Adoptive Transfer.
CD11c+ cells were isolated from livers and spleens of mice given the endogenous DC poietin human Flt3L (10 μg/mouse/day intraperitoneally for 10 days) as described.18 DCs were enriched by density centrifugation using histodenz. Plasmacytoid DC (pDC; mPDCA-1+) were positively purified using mPDCA-1 immunomagnetic microbeads and a paramagnetic LS column. For conventional myeloid DC (mDC) purification, mPDCA1− cells collected in the negative fraction after pDC selection were first run through LD columns to remove residual mPDCA1+ cells, then incubated with CD11c microbeads. mDC were positively selected by passage through a paramagnetic LS column. Ten million (>95% pure) DC were adoptively transferred by way of lateral tail vein injection.
Flow Cytometric Analysis.
After single cell suspensions were treated with Fc-blocking rat antimouse CD16/32 mAb (2.4G2) to prevent nonspecific binding, five-color flow cytometry was conducted by way of 30-minute incubation of the cells with fluorochrome-conjugated mAbs. Flow analysis was performed with an LSR II flow cytometer (BD Biosciences).
Routine Pathology and Immunohistochemistry (IHC).
Liver samples were fixed in 10% formalin, embedded in paraffin, sectioned (4 μm), and stained with hematoxylin and eosin. The severity of hepatic necrosis was scored blindly by one of the authors (A.J.D.). Sections were also stained with naphthol AS-D chloroacetate esterase staining kit (Sigma Diagnostics) for polymorphonuclear leukocytes (PMN). Liver tissue was also embedded in an optimal cutting temperature compound, and 6 μm cryosections were stained with anti-CD11c mAb with nuclear Hoechst staining. Positively stained cells in five high-power fields (×200) per each section were counted in a blinded fashion.
Total RNA was extracted from liver tissue using the TRIzol reagent (Invitrogen, Carlsbad, CA), and RNA content measured using 260/280 UV spectrophotometry. Messenger RNA (mRNA) expression for IL-6, IL-10, TGF-β, TNF-α, CCL2, and CXCL2 was quantified by SYBR Green two-step, real-time RT-PCR using an ABI PRISM 7000 Sequence Detection System (PE Applied Biosystems).22 The expression of each gene was normalized to β-actin mRNA content and calculated using comparative Ct methods.25
Quantification of Cytokine Protein Levels.
Liver samples were homogenized in tissue extraction reagent and protein fractions obtained. Cytokines and chemokines in samples (100 mg protein) were assessed using the Luminex multiplexing platform (MiraiBio, Alameda, CA) with the BioSource 20-plex mouse cytokine bead set (BioSource-Invitrogen, San Diego, CA) as per the manufacturer's specifications.26
The data are presented as means and standard errors of the mean. Comparisons between the groups at different timepoints were performed with Student t test or by analysis of variance with StatView (Abacus Concepts, Berkeley, CA). Differences were considered significant at P < 0.05.
Flt3L KO Liver Shows Markedly Reduced Frequencies of DC.
Flt3L KO mice exhibit markedly reduced numbers of mDC and pDC in the lymphoid organs.14, 27 Nonlymphoid organs, hearts, but not epidermis (skin) from Flt3L KO mice have been reported to have profound depletion of interstitial CD11c+ MHC class II+ DC19; however, hepatic DC populations in Flt3L KO mice have not been studied. To determine whether DC were also depleted from the liver of Flt3L KO mice, we performed flow cytometric analysis of hepatic NPC, as well as IHC. Flt3L KO livers exhibited marked reductions in CD11c+ F4/80− DC compared to WT livers (Fig. 1A). Both B220+ CD11c+ pDC and B220− mDC were reduced in Flt3L KO livers. Using IHC, WT livers showed primary concentrations of CD11c+ DC in portal tract, but Flt3L KO livers had virtually no detectable CD11c+ portal tract DC (Fig. 1B). These results confirmed pronounced reductions of hepatic DC populations in Flt3L KO mice.
Hepatic Warm IR Injury Is Significantly Ameliorated in Flt3L KO Animals and Enhanced by Flt3L Administration.
To determine the roles of DC in hepatic IR injury, we first conducted the warm IR procedure in WT and Flt3L KO mice. Unlike in sham-operated controls, severities of liver IR injury, determined by serum ALT levels at 3, 6, and 12 hours of reperfusion, were significantly less in Flt3L KO than in WT animals (Fig. 2A). Significant hepatic necrosis, mainly in zones 2 and 3, was seen in WT livers. In contrast, Flt3L KO mice had significantly reduced hepatic necrosis (Fig. 2B). Reduced hepatic injury in Flt3L KO mice was associated with less PMN infiltration compared to WT livers at all timepoints examined and the differences were statistically significant at 6 and 12 hours (Fig. 2C).
Alternatively, hepatic warm IR was induced in animals with increased numbers of DC after treating WT B6 mice for 10 days with Flt3L.18 Flt3L-treated WT animals exhibited marked increases in ALT levels compared to phosphate-buffered saline (PBS)-treated controls (Fig. 2D). These results indicated that increased numbers of DC induced by Flt3L enhance the severity of hepatic warm IR.
Flt3L Deficiency Reduces Inflammatory Cytokine/Chemokine Expression in Hepatic Warm IR Injury.
The intensities of hepatic inflammatory responses were also examined by determining mRNA up-regulation for inflammatory cytokines and chemokines. In WT mice, hepatic mRNA levels for IL-6, TNF-α, CCL2 (MCP-1), IL-10 and CXCL2 (MIP2) promptly increased at 3 hours of warm IR and decreased gradually thereafter. mRNA levels for TNF-α CCL2, IL-10, and CXCL2 were down-regulated significantly in Flt3L KO compared with WT animals (Fig. 3A). TGF-β mRNA expression increased marginally in WT mice with IR and sham, whereas Flt3L KO animals showed significantly reduced TGF-β mRNA at 3 and 12 hours. IL-6 mRNA levels were similar in Flt3L KO and WT animals. In accordance with the mRNA levels, Flt3L KO animals consistently showed significantly lower hepatic protein levels of inflammatory cytokines at 12 hours (Fig. 3B).
Adoptive Transfer of WT DC Restores Warm IR Injury.
Hepatic warm IR injury was significantly ameliorated in Flt3L KO animals with reduced numbers of DC. However, Flt3L KO mice are also known to exhibit reduced numbers of other leukocytes, such as NK cells. To determine whether the reduced hepatic warm IR injury in Flt3L KO mice was due solely to the deficiency in DC populations, we next conducted adoptive transfer experiments to restore blood-borne DC populations in Flt3L KO animals. Various types of DC (10 million DC per animal) were obtained from WT B6 mice and adoptively transferred into Flt3L KO mice 10 minutes before initiating warm IR injury. When CD11c+ splenic DC were transferred into Flt3L KO mice, ALT levels at 6 hours of reperfusion increased marginally compared to those in PBS-injected control Flt3L KO (Fig. 4A). Transfer of hepatic DC into Flt3L KO mice led to significantly greater increases in ALT levels compared to those in PBS-treated control Flt3L KO mice. The liver contains both mDC and pDC, and mDC are known to produce IL-12 and TNF-α. In contrast, pDC have distinct capacity to produce type-1 interferons (IFNs), key effectors of inflammatory activation.28 Therefore, we next adoptively transferred hepatic pDC or mDC into Flt3L KO mice. Interestingly, both hepatic pDC and mDC significantly increased ALT levels compared to PBS-infused control Flt3L KO mice (Fig. 4A). Although ALT levels were higher with pDC infusion than with mDC, no statistical differences were obtained between these two DC subsets.
As adoptively transferred DC might function differently in an environment lacking Flt3L, we also conducted the adoptive transfer experiment in WT animals undergoing hepatic warm IR injury. Both hepatic pDC and mDC increased ALT levels in WT animals compared to PBS-infused controls, although statistical significance was not achieved (Fig. 4B).
These results indicate that adoptive transfer of DC in hepatic warm IR is detrimental and increases liver injury. To determine whether augmented hepatic injury was associated with preferential migration of adoptively transferred DC into the liver, frequencies of CD11c+ cells were analyzed in IHC. At 6 hours after warm IR in Flt3L KO mice, adoptively transferred WT splenic or hepatic CD11c+ DC were found in multiple locations (e.g., spleen, lymph node, lung), without obvious differences depending on splenic or hepatic origin. CD11c+ cells were more easily detected in the liver with warm IR injury (Fig. 5) compared to other organs (data not shown). The numbers of adoptively transferred CD11c+ cells in the liver were higher with hepatic than with splenic DC infusion, and correlated with the severity of injury. It is unclear, however, whether hepatic migration of adoptively transferred DC is the cause or result of increased liver injury. DC might migrate preferentially to the liver and directly cause liver injury. Alternatively, infused DC could be recruited secondarily to the injured liver through their expression of appropriate chemokine receptors. In the steady-state (without injury), adoptively transferred WT splenic or liver DC migrated similarly to the normal liver of WT and Flt3L KO mice, suggesting that the lack of liver-resident DC does not affect adoptively transferred DC recruitment to the liver (Supporting Fig. 1).
Liver-Resident DC Promptly Up-regulate Costimulatory Molecules in Cold IR.
As adoptive transfer of DC increased liver IR injury, we next distinguished responses of liver-resident DC from infiltrating blood-borne DC using the hepatic ischemic injury model associated with LTx. WT B6 liver grafts were transplanted into CD45.1 B6 recipients with 24 hours cold preservation in UW to induce significant liver injury.22, 24 At 6 hours after LTx, liver NPCs were isolated and the activation status of donor and recipient CD11c+ DC was analyzed by determining the degree of costimulatory molecule expression. Donor liver-resident DC (CD45.1−) promptly up-regulated CD80 and PD-L1, compared to expression levels seen in pretransplant naïve liver DC (Fig. 6). CD40 and CD86 expression also increased, but statistical significances were not obtained (data not shown). Notably, donor (graft) DC showed higher coregulatory molecule expression than that seen on host (CD45.1+) infiltrating DC. Interestingly, MHC class II expression was significantly reduced in both graft resident and host infiltrating DC, and this was due to posttransplant loss of MHC class II+ DC that normally reside in naïve liver. Prompt up-regulation of costimulatory molecules in response to IR injury suggests an important role of liver-resident DC in hepatic IR injury.
Donor Liver Resident DC Deficiency Significantly Worsens Cold IR Injury Associated with LTx.
To determine if liver-resident DC have significant functional roles, we conducted LTx using Flt3L KO liver grafts. Flt3L KO to WT LTx with 24 hours cold storage in UW solution resulted in marked increases in serum ALT levels to >10,000 IU/L at 3 and 12 hours, compared to those in WT to WT LTx (<7,000 IU/L) (Fig. 7A). Increased IR injury in Flt3L KO to WT compared with WT to WT LTx was also evident in significantly greater hepatic necrosis (Fig. 7B). Frequencies of PMN were significantly higher in Flt3L KO than in WT liver grafts (Fig. 7C). Host DC migrated into Flt3L KO liver grafts and were detected by CD11c staining (Fig. 7B). Flt3L KO to WT LTx resulted in dramatically increased liver injury and inflammatory cell infiltration, suggesting a protective role of liver-resident DC.
Hepatic IR injury is a frequently encountered clinical problem. It is initiated by innate immune cells through the recognition of endogenous danger signals by PRRs.29, 30 The findings of this study support the notion that DCs play a role in innate immune responses during the early phase of hepatic IR injury; a profound lack of DC in Flt3L KO mice mitigated hepatic injury and diminished proinflammatory responses in the hepatic warm IR model. Further, the restoration of DC populations by adoptive cell transfer restored hepatic injury. These results indicate that blood-borne DC overall have proinflammatory and injurious roles in IR injury. However, the absence of DC solely in liver grafts in the Flt3L KO to WT LTx-induced cold IR model markedly increased hepatic injury. Thus, our findings also suggest that tissue-resident liver DC have a protective role in hepatic IR injury, in contrast to the proinflammatory and injurious role of blood-borne DC.
The in vivo roles of DC in innate immune responses have been studied previously by depleting DC using DT administration in CD11c-DTR mice. In hepatic warm IR9 and acetaminophen-induced hepatic injury,31 liver damage increases in DT-treated and thus DC-depleted CD11c-DTR mice. These studies have highlighted a protective role of DC against liver injury. The current study, on the other hand, used Flt3L KO mice and demonstrated amelioration of warm IR liver injury in a setting lacking DC. Thus, the proinflammatory role of DC during warm IR injury in this study contrasts with the report by Bamboat et al.9 demonstrating a protective role of DC by way of IL-10 production in the same liver warm IR model. Moreover, conflicting roles of DC have been reported recently in liver fibrosis studies in CD11c-DTR mice; Connolly et al.32 found that DC promoted fibrosis through the secretion of TNF-α, whereas Jiao et al.33 demonstrated that DC regulated the regression of liver fibrosis in a matrix metallopeptidase-9-dependent manner. In these studies, the precise definition of DC and macrophages, particularly in nonlymphoid tissues, has been a matter of considerable debate because accurate differentiation of these two cell populations can be difficult.34 Contradictory roles of DC in these studies might suggest that DC play a dual role in hepatic IR injury, exhibiting both proinflammatory and regulatory actions during progression of liver damage, depending on the characterization/definition of DC, their phenotype, or maturation status. Additionally, although these studies have focused on DC in the liver, the experimental models used have not distinguished liver-resident DC from recruited blood-borne DC during liver injury, and the experimental outcomes and conclusions are based on the depletion of both liver-resident and recruited blood-borne DC in CD11c-DTR or Flt3L KO mice.
The CD11c-DTR transgenic mouse has been developed to allow specific deletion of CD11c+ DC to investigate their role in various experimental settings. After injected DT enters the cells by way of DTR-mediated endocytosis, DT catalyzes ADP-ribosylation of elongation factors and inhibits protein synthesis, resulting in rapid apoptosis of both mitotic and terminally differentiated cells. CD11c-DTR mice carry a transgene encoding a simian DTR-GFP fusion protein under control of the murine CD11c promoter, and injection of DT leads to rapid depletion of CD11c+ DC, whereas other murine cells are not killed by DT due to low affinity of DT for the rodent DTR.35 Several recent studies, however, point to limitations of the model. First, deletion is not as specific for DC as believed originally; the CD11c promoter is active in several macrophage populations (e.g., splenic marginal zone and alveolar macrophages), and these cells are also depleted in CD11c-DTR mice with DT administration.36 Second, the transgene is apparently “leaky” and expressed by nonhematopoietic cells, and DT treatment can induce morbidity and mortality in CD11c-DTR, but not WT mice, unless bone marrow chimeras are generated.37 Third, significant CXCL2 production and neutrophilia have recently been associated with DC depletion in CD11c-DTR mice.38 Thus, although inducible ablation of DC in CD11c-DTR mice has been an attractive approach, liver injury in DT-treated CD11c-DTR mice could be enhanced by direct parenchymal injury caused by DT, depletion of other types of cells expressing CD11c, and/or significant neutrophilia.
By contrast, Flt3L KO mice do not require administration of any toxic substance; whereas the mice are deficient in both myeloid and plasmacytoid DC, NK cells are also reduced.14 NK cells are an important component of the innate immune system, and lack of NK cells in warm IR experiments might have contributed to liver warm IR injury amelioration in this study. However, NK cell deficiency in Flt3L KO liver grafts does not explain increased liver injury in our Flt3L KO to WT LTx experiments.
Furthermore, compared to blood-borne DC, tissue-resident DC are known to play unique roles in tissue immunity to maintain physiological function and homeostasis of each organ. Contradictory findings in previous works and the current study may reflect different degrees of liver-resident DC depletion, as well as distinct roles of liver-resident versus blood-borne DC in various experimental settings. The human body hosts complex microbial communities that outnumber our own cells (∼100 trillion microbial symbionts in the gut, skin, and oral cavity).39 Even in the steady state, the sterility of tissues varies, and organ-specific innate immune systems must be regulated to allow coexistence of commensal microbiota and the sensing of real danger differently.40 A sterile organ, such as the spleen, is required to activate innate immunity to maintain host sterility through the induction of proinflammatory responses. In contrast, the liver is constantly exposed to gut microbial products,41-44 and innate immunity must be tightly regulated to maintain a peaceful coexistence of the microbial products in the portal blood and hepatic sentinel cells. The liver employs various strategies to tolerate continuous exposure to pathogens and to avoid unnecessary innate activation and tissue damage. Gut-derived bacterial products maintain liver-resident DC in an immature status by way of IL-6/STAT3 pathway activation.12 Liver DC may also express negative regulators of PRR signaling (e.g., IRAK-M, Tollip, SOCS, SHIP) to prevent excessive and prolonged innate inflammatory responses.45-47 Further, we have shown previously that PD-L1 deficiency in liver grafts increases liver injury in PD-L1 KO to WT LTx.24 Liver-resident DC might control inflammatory responses by promptly up-regulating negative costimulatory molecules and regulating inflammatory cell infiltrates. Hence, liver DC and blood-borne DC might have distinct functions in responding differently to innate signals during hepatic IR injury.
In summary, using Flt3L KO mice, this study has demonstrated that DC play significant roles in acute hepatic IR injury. The virtual absence of DC in liver warm IR in Flt3L KO mice mitigated innate inflammatory responses and liver injury. Proinflammatory roles of blood-borne DC were also seen with adoptive transfer of WT DC. In contrast, lack of liver-resident DC in Flt3L KO to WT LTx-induced cold IR injury was augmented significantly. By using both warm and cold liver IR models, this study suggests distinct roles of liver-resident and recruited blood-borne DC in innate immune reactivity; liver-resident DC that are exposed continuously to gut microbial products function to control innate immune responses and injury, whereas blood-borne DC in a sterile microenvironment display proinflammatory activities.
The authors thank Carla Forsythe for editorial assistance, and Derek Barclay and Rita M. Sico for technical support.