Immunological Mechanisms of Rejection
Transplantation of an allograft induces a formidable immunological response by the recipient designed to destroy the foreign organ efficiently and quickly. Successfully controlling this complex multifaceted cascade is the critical element required for graft survival. It is beyond the scope of this chapter to explore fully our evolving understanding of the immunological mechanisms of rejection. The interested reader is referred to excellent reviews from leaders in the field. Presented here is a summary of the important principles needed by the transplantation hepatologist to understand the basis of the clinical manifestations and therapeutic management of rejection.
The stage is set for the recipient’s immune response to the allograft even before reperfusion of the graft. Inherent donor characteristics such as age, blood type, human leukocyte antigen (HLA) match, and graft type, are factors influencing several aspects of the recipient’s immune response. The process of organ procurement itself inevitably causes tissue damage to the graft—the most obvious example being ischemic reperfusion injury. The first responder to this early graft damage is the innate immune system, which drives a rapid, non-specific inflammatory response independent of immune recognition of foreign tissue antigens. Macrophages, neutrophils, and natural killer (NK) cells are activated by recognizing damage-associated molecular patterns resulting in the production of pro-inflammatory cytokines (interleukin 1 [IL1], IL6, tumor necrosis factor, interferons) and potent chemokines. The innate immune system can also actively complement cascades that create tissue injury independently of alloantibody binding and is able to activate NK cells capable of cell killing. Also, p-selectin, an important adhesion molecule, is upregulated on endothelial cells and mediates the characteristic ischemic-reperfusion vascular injury that alters the permeability and integrity of the endothelium. This in turn fuels further tissue injury and engages the systemic response of the host to the graft, promoting the engagement of the adaptive immune system.
The adaptive immune system mounts a specific response induced by recognition of the genetic disparity between the host and foreign tissue. In the case of transplantation, this response is initially driven predominantly by T cells responding to the mismatch of major histocompatibility complex (MHC) molecules between the donor and recipient. The HLAs associated with the MHC complex are classified as class I (HLA A, B, and C) or class II (HLA DQ, DR, and DP). In the normal liver, class I antigens are widely expressed on all cell types, whereas class II antigens are expressed mostly on antigen-presenting cells (APCs) and Kupffer cells. With rejection, class I and II expression are induced preferentially on endothelial cells and the biliary epithelium. Antibodies directed against class II fix complement more avidly as compared with class I antigens. Although MHC molecules are one of the most important targets, non-MHC targets, such as the angiotensin II1 receptor (AT1R), are increasingly recognized as important. The potency of the T cell response is enhanced by the wide repertoire of tissue antigens that T-cell precursors are primed to recognize. However, as shown in T-cell–deficient animal models, T-cell responses alone are not sufficient to reject even a fully mismatched graft. Also, despite the vigor of the innate immune response, it alone is seldom powerful enough to destroy the allograft. It is the dynamic interplay between both the innate and adaptive immune systems augmenting each other’s potency that ultimately leads to graft destruction.
How recipient T cells interact with their target antigens is important to understanding therapeutic options. Direct allorecognition occurs when the T-cell receptor engages intact MHC complexes presented by the donor’s own APCs. Indirect allorecognition differs in that the T cell engages donor-derived peptide presented on the recipient’s APC after degradation by the recipient’s antigen-processing pathways ( Fig. 17.1 ). The importance of having both pathways is that donor-derived APCs are a relatively small and finite population and can be quickly depleted, whereas the recipient’s APC population is renewable and able to maintain a long-term sustained response.
The next important step after recognition of foreign antigen is activation of the recipient T cells. The activation signal delivered to T cells through the T-cell receptor CD3 requires costimulation through separate pathways to activate T cells fully ( Fig. 17.2 ). Costimulatory pathways are complex; some cause upregulation and some cause downregulation of the activation response. CD28, expressed constitutively on T cells, is an important upregulator through its binding to members of the B7 family (CD80 and CD86), which are potent inducers of IL2 expression. In contrast, CD152 (CTLA-4), also expressed on activated T cells, competes with CD28 binding to its ligands, and is therefore an important downregulator of the activation process. The other important costimulatory pathway is the tumor necrosis factor (TNF) and TNF receptor family, characterized by the CD40 and CD40 ligand (CD154) interaction. Costimulatory blockade has already been exploited by newer therapeutic agents such as belatacept, a monoclonal antibody that binds to B7 receptors. Monoclonal antibody inhibition of the CD40-CD40L pathway is under clinical trial development.
Intracellular signaling pathways are the final step in the cell activation sequence. Simplistically, phosphorylation of several different signaling molecules in the cytoplasm generates transcription factors that carry the activation message to the nucleus, initiating the transcription of many genes. Two of the most critical upregulate IL2 and the high-affinity IL2 receptor. IL2 is a critical growth factor, promoting T-cell cycle progression and clonal expansion. Central to the mechanism of tacrolimus and cyclosporine is the inhibition of calcineurin, an important intracellular activation messenger that inhibits IL2 production.
Following T-cell activation, complex processes direct T-cell differentiation into many different subsets of T cells with specialized functions. CD4 + class II–restricted T cells, and their several subsets, have a predominantly helper function. CD8 + class I–restricted T cells generally have a cytotoxic function. In the transplantation setting, many factors influence T-cell differentiation, including the degree of ischemic-reperfusion injury, donor and recipient MHC (and potentially non-MHC) mismatch, and the overall immune responsiveness of the donor, which may already have been modified by immunosuppressive strategies begun before or at the time of transplantation. Also, control of the T-cell response includes checks and balances mediated by T regulatory cells and characterized by expression of the transcription factor Foxp3p.
B cells also have an important function in the immune response as the producers of potentially destructive antibodies. In addition, B cells are able to express complement receptors and can function as an APC. The activation of B cells is generally dependent on T cell help and is therefore linked to the complex differentiation pathways of T cells. Antibody-mediated injury most often occurs as a result of the cytotoxic effect of the complement-binding antibody. Destructive antibodies associated with graft injury can be generated against mismatched HLA MHC molecules, as well as against non-HLA targets, such as blood group antigens, endothelin type A receptor, and the MHC class I chain-related molecule A. Autoantibodies directed against antigens expressed on endothelium, such as AT1R, can induce a severe local injury characterized by microvascular coagulation and complement activation causing multiple thromboses within the graft.
Two other important concepts in graft injury are the dynamics of cell trafficking that bring destructive cells into the graft and the upregulation of adhesion molecules that localize destructive cells to their targets. Chemokines secreted by activated leukocytes play a key role in recruiting cells activated in the secondary lymphoid tissues back into the graft. Adhesion molecules such as intercellular adhesion molecule-1 tether leukocytes to the endothelium, allowing extravasation into the graft. Cytotoxic T cells, once bound to their target cell, release perforin, which creates pores in the target cell membrane through which the protease granzyme B, also released by the T cell, enters the cell and triggers apoptosis (programmed cell death).
The ability of specialized T and B cells to retain long-term memory to previous activation by specific antigens is another important concept relevant to short- and long-term graft outcomes. Memory cell responses require less stimulation and have a reduced activation threshold as compared with naïve cells. As appreciated clinically, late or multiple episodes of T cell acute rejection are frequently more difficult to control compared with initial rejection episodes. Moreover, B cells previously exposed to tissue antigens by sensitizing events, such as blood transfusions, pregnancy, or prior transplantation, are likely to mount a more robust antibody response if the donor tissue presents these antigens again. This memory response likely augments the observed increased risk of early rejection and graft loss seen in some patients with preformed donor-specific antibody (DSA).
The Liver as a “Privileged” Organ
The liver has long been regarded as immunologically privileged. In early studies of large animal models, it was observed that pigs transplanted across fully mismatched MHC antigens could be maintained without immunosuppression. As clinical experience grew in solid organ transplantation, clinicians recognized that liver transplant (LT) recipients seemed to require less immunosuppression, have less rejection, a decreased risk of sustained graft damage or graft loss from rejection, and a reduced incidence of chronic rejection compared with recipients of other organs. It was in LT recipients that first reports emerged of successful weaning of all immunosuppression, and that immunosuppression could be stopped indefinitely in some children treated for post-transplantation lymphoproliferative disease.
Intriguing hypotheses have been invoked to explain the tolerogenic effect of the liver. Two important anatomical features likely play a role—the unique structure and large surface area of the sinusoidal microvascular bed, and the portal vein circulation that brings a multitude of oral antigens and endotoxins into the liver. The liver has evolved to tolerate this antigen burden by means of a number of specialized APCs, including dendritic cells, Kupffer cells, hepatic stellate cells, and endothelial sinusoidal cells. These APCs are thought to present antigens to T cells in a manner that induces T-cell apoptosis, anergy, or differentiation into regulatory T cells. Other properties of the liver that may confer immune protection in the transplantation setting are the secretion of high levels of soluble HLA antigens and the ability of Kupffer cells to phagocytose immune complexes, platelet aggregates, and activated complement complexes. Perhaps most important is the regenerative properties of the liver itself.
Immunoresponsiveness of Children
Physicians experienced in the care of both adults and children after solid organ transplantation share a common impression that rejection in pediatric patients is more frequent and often more difficult to treat than rejection in adult patients. Higher CD4 counts in children characterize lymphocyte subset differences between healthy children and adults. Moreover, children show increased antibody formation after blood transfusions. Ettenger and colleagues noted that before renal transplantation, children 5 years of age or younger had an increased absolute number of T cells and an increased T-helper to T-cytotoxic suppressor ratio. The rate of spontaneous blastogenesis, a nonspecific measure of the alloreactivity of T cells, was also increased.
Although it is accepted that immune responsiveness is depressed at the extremes of age, it is still debated whether very young solid organ transplant recipients are relatively hyporesponsive in comparison with older children. One study reported that children transplanted when younger than 3 months still had a 42% incidence of rejection, comparable to that in older children, whereas another report found a lower incidence of rejection in younger children. In the Studies of the Pediatric Liver Transplantation (SPLIT) registry, the 12-month probability of rejection incrementally increased with age; it ranged from 43.8% in infants younger than 6 months up to 57.9% in children older than 13 years. Overall, the data suggest that the astute clinician taking care of children after LT will assume a robust immunological response to the graft, even in infancy.
Principles of the Diagnosis of Rejection
Seldom are there suggestive clinical clues that rejection is occurring. In early rejection episodes, patients may develop a fever, the liver graft may swell in response to the inflammatory response, and there may be poorly localized abdominal pain. Occasionally, there is right upper quadrant abdominal pain, back pain, or referred right shoulder pain. Abdominal distention, non-specific irritability, and loss of appetite can occur. These non-specific signs and symptoms force the clinician to rely heavily on more objective findings, usually the elevation of levels of the liver aminotransaminases (aspartate transaminase [AST] and alanine transaminase [ALT]), which generally increase before the serum bilirubin level. Levels of biliary canalicular-based enzymes, gamma glutamyl transpeptidase (GGT), and alkaline phosphatase may also increase. GGT is specifically associated with biliary epithelial injury and is more helpful than the much less specific unfractionated alkaline phosphatase. Leukocytosis, eosinophilia, and thrombocytopenia may occur. Because all these laboratory findings are not specific for rejection—and importantly can be caused by other serious vascular, biliary, and infectious complications—it is imperative that the diagnosis of rejection be quickly proven and not assumed. Not only is the early treatment of rejection important for reversal, but delaying the diagnosis of other treatable complications such as vascular or biliary compromise may have serious consequences. In particular, increasing immunosuppression to treat presumed rejection when the actual underlying problem is infection is a mistake that should be rigorously avoided.
The liver biopsy remains the gold standard for diagnosis. It is the only modality, sometimes in conjunction with other studies, that helps protect the clinician from missing another treatable diagnosis and confirms the need to accept the risk of increased immunosuppression. Particularly in the early post-transplantation period, the pattern of histological injury that characterizes acute rejection usually has enough distinct features to differentiate from biliary complications, vascular complications, and intrahepatic infection (e.g., cytomegalovirus [CMV], Epstein-Barr virus [EBV], cholangitis). Early and aggressive treatment of any of these other early post-transplantation complications is essential to protect graft function and early graft survival.
The value of the allograft biopsy has been enhanced by newer methods beyond histology that assess the tissue expression of micro-RNAs (miRNAs) signatures associated with rejection. Specifically, in LT recipients, the usefulness of miRNAs profiles in diagnosing rejection has been reported in both liver tissue and serum.
The liver biopsy in trained hands carries a low risk of complications. Some centers use only ultrasound-guided biopsies, and others use them only for technical variant grafts (segmental deceased donor or living donor grafts) when the graft is more centrally located in the abdomen or if there are other complications present, such as ascites. In very small children, the safety of a percutaneous liver biopsy can be improved by gelatin sponge pledget tract embolization. Although the serious complication rate (e.g., bleeding requiring transfusion from hemoperitoneum, hemothorax, hepatic hematoma, or hemobilia) is low, ranging from 1% to 3%, less invasive substitutes for liver biopsy to diagnose rejection have been investigated but no reliable biomarker has emerged. Transient elastography, measuring liver stiffness, has been most useful in evaluating fibrosis, including in pediatric LT (PLT) recipients. The inflammatory response in acute rejection has been observed to increase liver stiffness, but other complications such as hepatic infection, portal tract edema associated with biliary obstruction, and hepatic vein outflow obstruction would also be expected to increase liver stiffness acutely. For most experienced PLT hepatologists, the risk of missing the diagnosis of rejection—or missing the diagnosis of other treatable complications—is higher than the risk of the biopsy itself.
Acute Cell-Mediated Rejection
Acute cell-mediated rejection (CMR) is the prototype of the most common form of early rejection in the liver allograft. Modern immunosuppressive strategies, as discussed in Chapter 18, have made an impact on the incidence of acute CMR. Interestingly, in the few randomized controlled trials of new immunosuppressants that enrolled children, most notably the comparison between cyclosporin and tacrolimus, the incidence and severity of rejection were reduced but there was no demonstrated improvement in patient or graft survival. As has been consistently described by single-center experiences, and perhaps best characterized by a comprehensive multivariate analysis of the SPLIT registry, the most powerful predictor of patient and graft loss was serious infection, not rejection. The immunosuppression needed to control rejection is the biggest driver of the risk of infection, and this fundamental dilemma has continued to motivate efforts to minimize immunosuppression safely and decrease the risk of infection without jeopardizing graft function. The ability to find measures that would allow the reliable and objective individualization of immunosuppression, which cannot only be expected to change over time but also change in response to other causes of immune activation, is one of the biggest unmet needs in transplantation. This is particularly relevant to the care of PLT recipients, most of whom are younger than 5 years of age at transplantation, and will require management of immunosuppression over decades of life to prevent and control the immunological response to the graft.
Acute CMR is common, and there is good evidence that an early episode of acute CMR, if diagnosed and treated promptly, has no short- or long-term impact on graft function. An analysis of rejection in 1902 recipients of a first PLT from the SPLIT registry showed that rejection occurred most often in the first 3 months. The cumulative rejection rate was 0.45 at 3 months and increased only modestly to 0.59 at 24 months. The median time to first rejection was 16 days, the average number of rejection episodes per patient per year was 0.51, and more than one rejection episode occurred in 18.5% of children; steroid-resistant rejection was relatively unusual and occurred in 11.2% of children. When Kaplan-Meier probabilities of rejection over time were examined for various factors, there was a trend toward less rejection in children younger than 6 months of age and in recipients of living donor grafts, findings also reported by other investigators. However, in the multivariate analysis, initiation of immunosuppression with tacrolimus as opposed to cyclosporine was the only factor that showed a significantly lower probability of rejection at 6 months, 51% versus 64% ( P = .01). No difference in patient or graft survival was noted in those who underwent tacrolimus or cyclosporine induction.
Somewhat surprisingly, the incidence of rejection in pediatric recipients of living related grafts has not been consistently shown to be statistically less frequent than in recipients of deceased donor grafts, despite better HLA compatability. Therefore in general, pediatric transplant programs do not alter their immunosuppressive protocols for recipients of living donor as compared with deceased donor grafts.
The diagnosis of late acute rejection is more diffcult and often delayed and carries a different prognosis compared with early rejection. It is frequently associated with low levels of immunosuppression, which may be physician prescribed or related to patient or parent nonadherence. The diagnosis may be delayed because of less frequent follow-up, and the liver biopsy specimen may have been more difficult to interpret because of features of hepatitis such as non-specific inflammation, centrilobular venulitis, and necrosis. In addition, a delayed response to therapy and evidence of bile duct loss are associated with an increased risk of progression to chronic rejection.
Pathology of Acute Cell-Mediated Rejection
An international group of experts in LT, including pathologists, immunologists, and transplant hepatologists and surgeons, comprise the Banff Working Group on Liver Graft Pathology. This group has been convening since the late 1990s, and their work has been instrumental in establishing criteria for the diagnosis of various types of rejection following LT. The Banff group recognized that internationally agreed-on criteria for rejection would be essential to compare patient outcomes and to allow for meaningful analyses of clinical trials.
The now well-accepted definition of acute rejection proposed in 1995 was “inflammation of the allograft, elicited by genetic disparity between the donor and recipient, primarily affecting interlobular bile ducts and vascular endothelia, including portal veins and hepatic venules and occasionally the hepatic artery and its branches.” The Banff group’s more detailed histopathological description (based on core biopsies) included the following: (1) a mixed but predominantly mononuclear portal infiltrate, often including eosinophils usually not progressing beyond the limiting plate; (2) bile duct inflammation and damage; and (3) subendothelial inflammation of portal vein branches or terminal hepatic arteries in the portal tracts ( Fig. 17.3 ). A minimum of two of these criteria was required to make the diagnosis of rejection. The Banff group went on to develop a grading system for acute liver rejection ( Table 17.1 ) and also proposed a rejection activity index based on a scoring system for each of the three general criteria. These grading and scoring systems have been particularly useful in multicenter trials assessing graft outcomes.
The Banff group also emphasized the importance of correlating the histopathological findings with clinical information, noting that biopsy findings consistent with mild rejection can occur with normal liver function tests and no evidence of graft dysfunction. In such cases, treatment may not be warranted. Difficulties in diagnosis secondary to the sometimes focal nature of the histological picture of rejection, inadequate tissue from core biopsies, and treatment before obtaining the biopsy can also confound diagnostic accuracy.
Treatment of Acute Cell-Mediated Rejection
The first treatment strategy for acute CMR is generally high-dose intravenous steroids, most often methylprednisolone. The dose and duration of the steroid pulse have never been standardized. Some centers use three consecutive or alternate day similar doses, and others initiate treatment with a 10- to 20-mg/kg/day bolus, followed by a tapering schedule. An example of guidelines for treating CMR is shown in Table 17.2 . In some select cases, if the rejection episode appears mild both clinically and histologically, increasing the target tacrolimus levels into a higher therapeutic range is sufficient to reverse the episode.
|Methylprednisolone (MP), |
Bolus and Taper
|Children ≤ 50 kg||Children ≥ 50 kg|
|Day 1||MP 20 mg/kg/d IV ÷ q6h||MP 1,000 mg/d IV ÷ q6h|
|Day 2||MP 10 mg/kg/d IV ÷ q6h||MP 200 mg/d IV ÷ q6h|
|Day 3||MP 8 mg/kg/d IV ÷ q6h||MP 180 mg/d IV ÷ q6h|
|Day 4||MP 6 mg/kg/d IV ÷ q6h||MP 80 mg/d IV ÷ q6h|
|Day 5||MP 4 mg/kg/d IV ÷ q6h||MP 40 mg/d IV ÷ q6h|
|Day 6||MP 2 mg/kg/d IV ÷ q12h||MP 20 mg/d IV ÷ q12h|
|Day 7||Oral prednisone, 0.3 mg/kg/d q day max||Oral prednisone, 20 mg q day|
|When on high-dose steroids|
|Treatment of steroid-resistant rejection|
Steroid-resistant rejection is less common with the widespread use of tacrolimus as compared with cyclosporine as the primary immunosuppression. The diagnosis of steroid-resistant rejection should be confirmed by an additional biopsy because the increased immunosuppression needed to treat this entity successfully carries considerable risk.
The first approach is to bolster maintenance immunosuppression by the addition of drugs such as mycophenolate mofetil, azathioprine, or a mammalian target of rapamycin (mTOR) inhibitor (sirolimus or everolimus). Antilymphocyte biologics such as polyclonal antithymocyte globulin (ATG) or monoclonal anti-CD3 drugs are reserved for moderate or severe steroid-resistant rejection but are accompanied by important side effects. Success has also been reported using basiliximab, an IL2 receptor-blocking monoclonal antibody generally used in induction regimens, to treat steroid-resistant rejection. Basiliximab is a non-depleting antibody with a more favorable safety profile as compared with ATG. Monitoring for infectious complications with this additive immunosuppressive burden is essential, particularly for the emergence or reactivation of cytomegalovirus (CMV) or Epstein-Barr virus (EBV) disease. This includes the risk for EBV-related, post-transplantation lymphoproliferative disease. Using antiviral drugs such as valganciclovir, combined with increased monitoring for viremia (most easily accomplished by following serial measures of the polymerase chain reaction assay for CMV and EBV), is a common practice in managing children with rejection. The sites of action of immunosuppressive drugs on rejection pathways are shown in Fig. 17.4 .
Chronic rejection, typically defined histologically as vanishing bile duct syndrome ( Fig. 17.5 ), is an important cause of graft dysfunction and graft loss, but is now increasingly rare in the modern era of immunosuppression. Investigators attribute this decline in chronic rejection to the increasing use of tacrolimus in LT. One of the first clinical benefits ascribed to the use of tacrolimus was its ability to reverse early chronic rejection, a property not shared by other immunosuppressive drugs available at that time. In reviewing the extensive use of tacrolimus in Pittsburgh, Jain et al. reported that chronic rejection occurred in 3.1% of 1048 tacrolimus-treated patients and was virtually absent in the pediatric cohort. This was a sharp reduction from the reported incidence of 15% to 20% in the era of cyclosporine primary immunosuppression.
Under-immunosuppression, and in particular non-adherence, is a known risk factor for chronic rejection. A study of risk factors for chronic rejection in 385 pediatric liver recipients at the University of Chicago found that recipients of deceased donor organs, African Americans, patients with two or more episodes of rejection, patients with post-transplantation lymphoproliferative disease, CMV disease, and those with autoimmune hepatitis as the indication for transplantation had a significantly higher risk for chronic rejection. Chronic rejection may evolve in a continuum with multiple episodes of inadequately controlled acute rejection within the first year after transplantation but usually presents years after transplantation, generally in the context of inadequate immunosuppression.
Clinically, the first clue is often the unexplained onset of pruritis with elevated GGT levels, sometimes with only a modest elevation in transaminase levels. In the teen years, it is not uncommon that the diagnosis is made after the teenager arrives jaundiced in the emergency room, having failed several follow-up clinic visits.
Pathology of Chronic Rejection
Obliterative arteriopathy, manifested by foamy cells and intimal hyperplasia involving the medium-sized arterioles, is the underlying mechanism for the bile duct loss that characterizes chronic rejection ( Fig. 17.6 ). Interestingly, the ischemic injury caused by obliterative arteriopathy is the common pathway to chronic rejection in other organs causing bronchiolitis obliterans in the lung, coronary artery vasculopathy in the heart, and transplant glomerulopathy in the kidney. The immune attack on the graft arteriolar endothelium is the underlying mechanism of the so-called transplant vasculopathy seen in chronic rejection across different organs. Although the mechanism of injury to the graft endothelium is not well elucidated, injurious antibodies (as discussed later) are thought to play an important role.
The Banff Working Group has published useful criteria for standardizing the staging and reporting of chronic rejection of the liver. The essential diagnostic elements are as follows: (1) the presence of bile duct atrophy/pyknosis; (2) convincing foam cell obliterative arteriopathy; and (3) bile duct loss affecting more than 50% of the bile ducts. Unfortunately, because a needle biopsy is seldom deep enough to include the larger arterioles showing the pathogonomic obliterative arteriopathy, more reliance is placed on evaluating the bile duct injury. The Banff group further differentiated the features of early chronic rejection from late chronic rejection ( Table 17.3 ). In early chronic rejection, the parenchyma surrounding the terminal hepatic vein is infiltrated by mononuclear cells, mild fibrosis, and cell dropout. In the portal tracts there is bile duct damage but, less frequently, bile duct loss. At this stage, reversibility is more likely with immunosuppressive strategies. In comparision, in late chronic rejection, the bile duct loss is prominent, bridging fibrosis may be severe, and necrosis and inflammation may be less evident. This is seldom a reversible injury.