Abstract
Hepatocyte transplantation (HT) is emerging as a promising alternative to orthotopic liver transplantation (OLT) in patients with certain liver-based metabolic disease and acute liver failure. Hepatocytes are generally infused into the portal venous system, from which they migrate into the liver cell plates of the native organ. One of the major hurdles to the sustained success of this therapy is early cell loss, with up to 70% of hepatocytes lost immediately following infusion. This is largely thought to be due to the instant blood-mediated inflammatory reaction (IBMIR), resulting in the activation of complement and coagulation pathways. Transplanted hepatocytes produce and release tissue factor (TF), which activates the coagulation pathway, leading to the formation of thrombin and fibrin clots. Thrombin can further activate a number of complement proteins, leading to the activation of the membrane attack complex (MAC) and subsequent hepatocyte cell death. Inflammatory cells including granulocytes, monocytes, Kupffer cells, and natural killer (NK) cells have been shown to cluster around transplanted hepatocytes, leading to their rapid clearance shortly after transplantation. Current research aims to improve cell engraftment and prevent early cell loss. This has been proven successful in vitro using pharmacological interventions such as melagatran, low-molecular-weight dextran sulphate, and N-acetylcysteine (NAC). Effective inhibition of IBMIR would significantly improve hepatocyte engraftment, proliferation, and function, providing successful treatment for patients with liver-based metabolic diseases.
Introduction
Hepatocyte transplantation (HT) offers a promising alternative to liver transplantation for patients with certain liver-based metabolic diseases and acute liver failure. Human hepatocytes are isolated from whole donor livers or tissue obtained from split liver procedures using a collagenase perfusion technique and purified using centrifugation (Fig. 1)1–4. The number of cells transplanted aims at approximating 5% of the normal liver mass, and cells are usually administered under radiological guidance via the portal vein or spleen.
Isolation of human hepatocytes. Hepatocytes are isolated from donor liver tissue using a collagenase perfusion technique and purified using centrifugation. Cells are either transplanted directly into the portal vein or cryopreserved for emergency cases. Viability and metabolic activity is checked before administration via the portal vein.
HT has potential advantages over orthotopic liver transplantation (OLT) in certain cases because the procedure is less invasive with significantly less complications than organ transplantation. The technique can also provide temporary life-saving treatment until a patient can undergo whole-organ transplantation or, potentially, future gene therapy2,5. HT can improve the phenotype of a number of metabolic liver diseases. To date, HT in patients with Crigler–Najjar syndrome has been shown to decrease serum bilirubin levels by 30–60% and to maintain the expression of UDP glucuronosyltransferase activity for up to 9 months5–10. Furthermore, partial hepatectomy followed by HT increased concentrations of hepatocyte growth factor (HGF), leading to increased hepatocyte function 11 . Patients with factor VII deficiency have shown up to a 70% decrease in the requirement for recombinant factor VII (fVII) treatment 12 , and in familial hypercholesterolemia, there was up to a 70% decrease in low-density lipoprotein (LDL) production 13 . Phenotypic improvement has also been seen in glycogen storage disease (GSD) 1a and 1b, with normal glucose-6-phosphatase activity and increased triglyceride levels observed14,15. HT has also been shown to improve a number of urea cycle defects including ornithine transcarbamylase (OTC) deficiency, where patients showed a decrease in ammonia production and an increase in urea production16,17; argininosuccinate lysase (ASL) deficiency, in which there was sustained hepatocyte engraftment with increased ASL activity 18 ; and tyrosinemia type 1, which improved clotting factor levels and decreased bilirubin concentrations 19 . Such promising results suggest that HT has clinical potential in alleviating the severity of metabolic liver diseases, making day-to-day management easier and improving quality of life for patients. Although HT has great potential, there are still many bottlenecks that limit the efficacy of the technique. This includes scarcity of donor organs, decreased viability of cells due to cryopreservation, early cell loss, and lack of engraftment into the liver. So far, only rarely have patients with metabolic disease demonstrated a sustained long-term improvement in their condition following HT 18 .
Pathophysiology of the Instant Blood-Mediated Inflammatory Reaction
Although immunosuppression is generally used in HT for liver-based metabolic conditions, early cell loss due to the innate immune response has not been successfully targeted. Transplantation of hepatocytes, islets of Langerhans, and therapies involving mesenchymal stem cells (MSCs) have all been associated with cell damage and significant cell loss very soon after administration, leading to poor engraftment and cell function 20 .
Early hepatocyte cell loss has been demonstrated in dipeptidyl peptidase IV-deficient F344 rats, which showed hepatocytes present in large numbers in portal areas shortly after transplantation 21 . There was significant clearance of hepatocytes from these portal areas and low engraftment into the liver sinusoids with an average 3.3-fold decline of the total number of transplanted hepatocytes between 2 and 24 h 21 . This significant cell clearance was attributed to the presence of granulocytes, phagocytes, and activated macrophages surrounding the hepatocytes in portal areas within 6 h posttransplantation 21 . This has been confirmed clinically with explanted livers from metabolic liver disease patients who had undergone HT followed by OLT. Patients were shown to have inflammatory thrombi containing lymphocytes, histocytes, and multinucleated macrophages after HT 22 .
This inflammatory reaction is now commonly referred to as the instant blood-mediated inflammatory reaction (IBMIR), in which cells are recognized by the innate immune system, leading to rapid activation of both complement and coagulation pathways 23 . Additionally, inflammatory cells including granulocytes, monocytes, Kupffer cells, and natural killer (NK) cells are activated, leading to rapid clearance and death of transplanted hepatocytes 24 . IBMIR-induced cell loss has been well described in islet transplantation in in vitro blood perfusion systems, in rodent and primate studies, and in patients undergoing clinical islet transplantation25–29.
The coagulation cascade is usually activated following endothelial cell surface damage, leading to the activation of fVIIa. Hepatocytes have been shown to trigger activation of the extrinsic coagulation cascade by producing tissue factor (TF), which binds to fVII and fVIIa, leading to the activation of fVIIa (Fig. 2). The TF–VIIa complex initiates the activation and cleavage of factor X (fX), producing fXa, which associates with cofactor Va to form the prothrombinase complex that converts prothrombin to thrombin 30 . Thrombin initiates the conversion of fibrinogen to fibrin and activates fXIII to promote fibrin cross-linking and the formation of fibrin clots. Thrombin also amplifies its own expression through the activation of cofactor VIII to VIIIa, which binds fIXa to further activate fX. Thrombin activates platelets by binding to receptors on their cell surface and is required for their aggregation and formation of clot structure.
Activation of IBMIR following HT. Transplanted hepatocytes produce and release TF, which binds to the coagulation fVIIa and initiates the activation and cleavage of fXa, which is responsible for the conversion of prothrombin to thrombin. This is followed by platelet binding to the cell surface and the conversion of fibrinogen to fibrin-forming fibrin clots. Thrombin has also been shown to activate complement proteins. C3 activates C3 convertase resulting in the cleavage of C3 to C3a and C3b, which activates C5 convertase, resulting in the cleavage of C5 to C5a and C5b-9. C5b is responsible for the production of the MAC, which binds to the membrane of target cells resulting in cell lysis.
The complement and coagulation cascades share similar characteristics, with both consisting of circulating zymogens that rely on a series of enzymatic reactions, producing an active protease and a nonenzymatic ligand31,32. There are over 30 complement proteins; however, the most important is complement component 3 (C3). The complement cascade consists of three distinct pathways including the alternative, classical, and lectin pathways; however, all lead to the cleavage of C3 into C3a and C3b by C3 convertase (Fig. 2). The alterative pathway is the first to act and is directly activated by binding of the C3b fragment to the foreign cell surface, targeting it for destruction by phagocytes, while the smaller C3a fragment acts as a chemoattractant for phagocytes and other inflammatory cells 32 . The lectin pathway is activated by mannose binding lectin (MBL), which circulates in plasma in its inactive form in complex with MBL-associated protease (MASP) 1 and 2. The MBL complex binds to mannose containing carbohydrates on the foreign surface 32 . Last, the classical pathway is activated when C-reactive protein or immunoglobulins such as IgM or IgG bind to specific antigens on a target surface. All three pathways lead to the covalent binding of C3b to the foreign surface. C3b binds to C3 convertase producing C5 convertase, which results in the cleavage of C5 into C5a and C5b. C5b triggers the terminal pathway and initiates the formation of the MAC, which creates pores in the target cell membrane 33 . The liver is a major source of complement proteins, primarily produced by inflammatory cells and hepatocytes. Interestingly, it has been shown that hepatocytes synthesize up to 15 times more C3 than macrophages 34 .
Interactions between the complement and coagulation pathways are poorly understood (Fig. 2). Complement proteins may directly enhance coagulation through platelet activation and modification of the lipid membrane, resulting in activation of TF 31 . Furthermore, thrombin may also have a role in amplifying the complement cascade 35 . Krisinger et al. showed thrombin efficiently cleaved C5 at a newly identified site, R947, generating C5T and C5bT 35 . The cross talk between such related biological cascades has clinical importance with evidence suggesting an inhibitor of the coagulation cascade may also prevent activation of the complement pathway, reducing the effects of IBMIR with a single inhibitor 23 .
Activation of both complement and coagulation pathways leads to the stimulation of a number of inflammatory cells including neutrophils, monocytes, macrophages, and NK cells (Fig. 3). Macrophages or Kupffer cells in the liver are responsible for phagocytosis of foreign cells, and translocation of nuclear factor κB (NF-κB) to the macrophage nucleus initiates translocation of a number of inflammatory cytokines including interleukin-1 (IL-1), IL-2, IL-6, IL-12, and chemokine (C-X-C motif) ligand 8 (CXCL-8) 32 . These act as chemokines for a number of inflammatory cells such as neutrophils and NK cells. Neutrophils function in a similar way to macrophages, producing a number of degradative enzymes and granules to lyse the target cell. NK cells are lymphocytes that circulate in an active state and act through interferon-γ (IFN-γ) production and through receptor-mediated binding that allows distinction between foreign and innate cells.
Intravenously transplanted hepatocytes cause activation of inflammatory cells including neutrophils, monocytes, Kupffer cells, and NK cells. Such inflammatory cells recognize transplanted hepatocytes and activate mechanisms including receptor-mediated phagocytosis, cytokine and chemokine production, and production of cytotoxic granules such as fas and perforin. Activation of such inflammatory responses contributes to hepatocyte cell lysis and death.
Thus, activation of processes such as coagulation, complement, and inflammatory cell recruitment has the potential to cause significant harm to transplanted cells, preventing engraftment and the success of the transplant. Current research has aimed to discover the possibility and mechanisms of IBMIR-induced hepatocyte cell loss and investigate reliable and effective methods of preventing this from occurring.
Activation of Coagulation in HT
A single-stage clotting assay showed that hepatocytes in contact with human plasma resulted in procoagulant activity (PCA) that increased significantly with increasing concentrations of hepatocytes. PCA was absent in fVII-deficient plasma, further suggesting the role of TF in coagulation activation. A Chandler loop model was designed in which polyvinyl chloride (PVC) tubing was coated with heparin, incubated at 37°C, and placed on a rocker to generate a flow of 45 ml/min to mimic portal vein blood flow. Using this model, it was shown 5 × 10 5 hepatocytes induced clot formation in ABO-matched blood circulating for 30 min. This correlated with a significant drop in platelet count and increased D-dimer levels 36 . The same authors have shown PCA in human adult liver-derived mesenchymal progenitor cells (hALPCs). Thromboelastography revealed shorter clotting times in plasma and whole blood treated with hepatocytes, while the tubing loop model revealed a significant drop in platelet count and increased D-dimer levels 37 .
The source of TF in HT was directly tested in mouse models of acetaminophen (APAP) overdose 38 . Following HT, thrombin–antithrombin (TAT) expression in plasma was significantly higher in wild-type mice transplanted with control hepatocytes compared to wild-type mice receiving hepatocytes from genetically modified mice with selectively deleted TF expression 38 . This proves donor hepatocytes are the source of procoagulant TF, triggering IBMIR.
Activation of coagulation following HT has been demonstrated clinically in a child with OTC deficiency 23 . An hour following intraportal infusion of 8.3 × 10 7 cryopreserved hepatocytes, there was a significant increase in TAT, fXI–antithrombin, fXII–antithrombin, and platelet depletion 23 . This suggests activation of IBMIR within 60 min of HT. Similarly, a 9-month-old girl with Crigler–Najjar syndrome received 2.6 billion cryopreserved hepatocytes over 2 weeks in 14 infusions. The patient showed increased D-dimer levels but normal fibrinogen, prothrombin time, and activated partial thromboplastin time 36 . The authors suggest cell-dependent coagulation may occur in the small liver sinusoids and therefore may not be detected in high peripheral and portal flow 36 . Evidence of thrombosis has also been demonstrated in explant livers of children with metabolic livers that received HT. Hepatocytes, identified by the bile salt export protein (BSEP), were found in the portal vein thrombi 22 .
Activation of Complement in HT
Complement activation of C3a and C5b-9 has been demonstrated in a tubing loop model containing ABO-matched human blood and 1 × 10 5 fresh hepatocytes 23 . The addition of melagatran, a direct thrombin inhibitor and low-molecular-weight dextran sulfate (LMW-DS), an antithrombotic agent, reduced both TAT expression and C3a expression, further suggesting the association between the complement and coagulation cascades. Melagatran, a thrombin inhibitor, reduces the generation of thrombin through the production of chondroitin sulphate and indirectly inhibits the complement system39,40.
Complement activation has been implicated in the acute death of allogenic rat models following intraportal infusion of rat hepatocytes 41 . Transplantation of WAG-Rij hepatocytes into Lewis rat participants resulted in high morbidity rates and large thrombi containing complement components including C3 and IgM. No deaths occurred when WAG-Rij hepatocytes were transplanted into WAG-Rij rats 41 . Lewis rat recipients pretreated with the complement inhibitor cobra venom factor had high survival, suggesting complement activation was a vital contributor to the higher morbidity rates 41 .
The role of complement activation in HT has been further investigated using transfected hepatocytes. Hammel et al. infected hepatocytes with an adenovirus containing the genes encoding the human cytomegalovirus (hCMV) promoter and the human CR1 receptor, which is a potent inhibitor of the C3 and C5 convertases, essential in the classical and alternative complement pathways 42 . Transplantation of hepatocytes expressing the CR1 adenovirus prevented immediate rejection in immunocompetent Nagase analbuminemic rats and increased albumin production. Such genetic modification of hepatocytes may be useful in reducing complement activation following transplantation into sensitized recipients 42 .
Infiltration of Inflammatory Cells
Once administered via the portal vein, transplanted hepatocytes become entrapped in the liver sinusoids, migrate across the endothelial barrier, and integrate into the parenchyma23,43. However, the presence of transplanted hepatocytes within the sinusoids temporarily occludes blood flow and activates ischemia–reperfusion events 44 . Ischemia–reperfusion leads to the activation of Kupffer cells, which are specialized macrophages residing in liver sinusoids, responsible for the clearance of transplanted hepatocytes and the 70% cell loss observed following transplantation 44 . Gadolinium chloride (GdCl3) has been used to deplete Kupffer cells and increase hepatocyte cell survival 44 . Control rats showed transplanted cells in 34% of liver radicles compared to 66% in GdCl3-treated rats 72 h following transplantation. Furthermore, approximately twofold more hepatocytes were observed in the periportal sinusoids and the liver parenchyma 1, 2, and 7 days following transplantation. The authors suggest inhibition of Kuffper cell activation offers a way to prevent early cell loss following transplantation and improve engraftment of transplanted hepatocytes.
Inflammatory cells from the recipient liver have also been demonstrated to stimulate activation of a number of cytokine–chemokine genes within 3 weeks after HT 45 . This included chemokine ligands, chemokine receptors, and regulatory cytokines. In the DPPIV rat model, gene expression levels of cytokine–chemokine genes increased from 2- to over 120-fold. Depletion of Kupffer cells and neutrophils resulted in the normalization of 19 of 25 genes. This was associated with increased cell engraftment for up to 2 weeks 45 .
NK cells have also been shown to accumulate around transplanted hepatocytes 46 . Wesolowska et al. developed a combined protocol of nonlethal whole-body irradiation, administration of anti-asialo GM1 antiserum (to eliminate NK cells), and reconstitution with bone marrow cells to eliminate inflammatory cell activation following transplantation. Lewis rats received this combined treatment over 3 days before transplantation of syngenic hepatocytes into the spleen. Such treatment led to the formation of hepatocyte islands 14 days following transplantation, which increased in number at 30 days. Immunohistochemical staining revealed the absence of cytotoxic granulocytes, macrophages, and NK cells 24 . Combined treatment may be a potential way to eliminate inflammatory cells involved in the innate immune response and prevent early hepatocyte cell loss. However, clinically, the risks of such treatment may need to be considered.
It is also important to consider the potential benefit of the inflammatory response. To a certain extent, Kupffer cell production of cytokines such as ILs, tumor necrosis factor-α (TNF-α), and free radicals is important for endothelial disruption, which is necessary for transplanted hepatocytes to cross the endothelial barrier and enter the sinusoidal plates21,47. It may be important to consider the effects complete inhibition of the inflammatory response has on the hepatocyte translocation and engraftment. It may be more effective to dampen the response but not cause complete inhibition.
Current Therapeutic Strategies to Inhibit Ibmir
Inhibition of IBMIR has been well explored in the setting of islet transplantation using a variety of different methods (Table 1). This includes systemic anticoagulant drugs such melagatran, LMW-DS, and N-acetylcysteine (NAC). These have been shown to inhibit platelet consumption, decrease the expression of complement and coagulation proteins, and reduce the number of inflammatory cell subtypes in blood perfusion models48,49. Several anti-inflammatory proteins have also been investigated in vivo. This includes alpha-1 antitrypsin (AAT) and activated protein C, which have been shown to downregulate the expression of a number of inflammatory cytokines and improve islet survival and engraftment in mouse models of islet transplantation50–53. An anti-TF antibody decreased posttransplant markers of coagulation following islet transplantation in cynomolgus monkeys, resulting in higher C-peptide (CP) levels and prolonged graft function 29 . A novel approach has been to coat cell surfaces with macromolecular heparin complexes 28 . Tubing loops with heparin-coated islets resulted in no macroscopic clotting, and the generation of TAT and drop in platelet count were significantly attenuated compared to untreated loops 28 . Alternatively, it is possible to coat the cell surface of islets with poly(ethylene glycol) (PEG), which allows binding of a number of different anti-inflammatory proteins such as sCR154,55.
Inhibitors of IBMIR and Their Mechanisms of Action in Islet Transplantation
MSCs have also been shown to trigger IBMIR after exposure to human blood, with activation dependent on the cell passage number. The authors showed visible clot formation in the Chandler loop model, which was prevented by the addition of the anti-TF blocking agent FVIIai and by monoclonal anti-TF antibodies 56 .
Currently, only a limited number of anti-inflammatory drugs have been tested to inhibit IBMIR in the context of HT.
Therapeutic Strategies to Inhibit Ibmir in HT
Systemic administration
In 2011 Gustafson et al. studied a number of systematic coagulation inhibitors to minimize IBMIR in HT and investigated the effect of thrombin inhibition, a fVIIa inhibitor (iFVIIa), and an anti-TF monoclonal antibody. In a tubing loop model, both inhibitors prevented platelet consumption after 15 min, but this was not maintained after 30 min. iFVIIa did not significantly decrease the concentration of coagulation factors TAT, FXIIa-AT, FXIa-AT, and complement factors C3a and C5b-9G 23 . In contrast, the direct thrombin inhibitor melagatran (10 μM) and LMW-DS prevented platelet depletion and significantly decreased concentrations of TAT, FXIIa-AT, FXIa-AT, C3a, and C5b-9G after 60 min 23 .
The antioxidant NAC has been shown to inhibit the activity of thrombin and coagulation fVII and fX 57 . NAC has been shown to inhibit PCA in whole blood containing hepatocytes in a dose-dependent manner 36 . PCA was significantly lower in plasma containing hepatocytes and 10–25 mM/L NAC. Furthermore, platelet counts and D-dimer concentrations were maintained at a similar level to the control 36 .
In hALPCs, the combination of bivalirudin and heparin treatment inhibited reduced clotting time and improved both engraftment and cell survival 37 .
Encapsulation
An alternative approach to prevent activation of IBMIR response following HT is to encapsulate cells in microbeads made from purified alginate 58 . The semipermeable membrane within the microbeads allows glucose and oxygen to pass through, maintaining necessary metabolic function while also protecting against immunocompetent cells by preventing the entry of antibodies54,59. Encapsulating hepatocytes in alginate microbeads improved the survival of mice with acute liver failure by providing metabolic support, decreasing the number of cytokines and reducing the inflammatory stress on the liver 58 . However, the encapsulation of cells using alginate or agarose can produce large capsules that make transplantation into the liver through the portal vein difficult, as major blood vessels would be occluded 54 . Teramura et al. carried out microencapsulation of cells in an ultra-thin polymer membrane, which showed reduced platelet aggregation, decreased TAT levels, and lower levels of C3a and C5b compared to alginate-coated cells when exposed to human whole blood and serum 54 .
Cytotopic Therapeutic Agents
Currently, novel cytotopic therapeutic agents are being investigated as a novel way to target IBMIR in both whole-organ transplantation and cell transplantation. Such agents are designed to bind to the cell membrane of transplanted cells and provide local inhibition of complement and coagulation pathways without the need for systemic treatment, which may result in bleeding risk. Currently, cytotopic agents such as Mirococept are being developed and investigated for kidney, islet, and HT, as well as the prevention of myocardial and intestinal ischemia–reperfusion injury60–63. Mirococept is derived from the human complement regulatory protein CR1, a transmembrane protein found on the cell surface of inflammatory cells 64 . CR1 is a potent inhibitor of the C3 and C5 convertases 64 . In both rat kidney transplantation and human tumor nephrectomy specimens, immunohistochemical analysis showed that Mirococept localized to the glomerular and tubular structures of the kidney. After 24 h, neutrophil activity and deposition of C3a and C5-9 were significantly lower in Mirococept-treated grafts compared to controls 65 . In 2012, 76 patients had received Mirococept, including one liver patient. These initial results showed Mirococept had no adverse side effects, and the next step is a multicenter phase 2b clinical trial 64 .
Conclusions
Evidence of IBMIR has been demonstrated following HT in blood perfusion models, animal experiments, and patients. Transplanted hepatocytes produce and release TF, leading to the activation of coagulation and complement pathways and eventually cell lysis. Inflammatory cells are further activated, leading to the rapid clearance of hepatocytes shortly after transplantation. Such activation of the innate immune response following HT is a limiting factor to the success of this technique. Current research in HT aims to reduce this early cell loss and improve cell engraftment. This includes pharmacological interventions, cell encapsulation, and novel cytotopic therapeutic agents. Improved cell engraftment and survival will significantly increase the application of HT and provide treatment for patients who are unsuitable or unable to receive OLT.
Footnotes
Acknowledgments
The authors would like to acknowledge support from the Medical Research Council Centre for Transplantation at King's College London and the Department of Health via the National Institute for Health Research (NIHR) Comprehensive Biomedical Research Centre (BRC) award to Guy's and St Thomas’ National Health Service (NHS) Foundation Trust in partnership with King's College London and King's College Hospital NHS Foundation Trust. C. Lee was supported by a BRC Interdisciplinary Ph.D. studentship award. Open access for this article was funded by King's College London. The authors declare no conflicts of interest.