Lymphoproliferation in children after liver transplantation.

Abstract

Posttransplant lymphoproliferative diseases (PTLDs) are a group of complications that hamper results and outcome of both solid-organ and bone marrow transplantation. Most of these diseases are related closely to Epstein-Barr virus (EBV) infection in the context of induced immunosuppression (1–3). Described for more than 30 years, PTLDs remain a major concern despite continuous progress in the knowledge of their risk factors and management. However, recent insights into the pathogenesis and new diagnostic tools offer promising future strategies and stimulate this review. DEFINITIONS AND EPIDEMIOLOGY Posttransplant lymphoproliferative diseases are a very heterogeneous group of diseases characterized by lymphoproliferation after transplant. This lymphoproliferation does not have the same clinical and histologic pattern that lymphoma in nonimmunosuppressed patients does, and is most often driven by EBV infection. In this review, the term PTLDs includes all the EBV-related posttransplant lymphoproliferations. Some authors differentiate more specifically mononucleosis-like syndrome, polymorphic, or lymphomatous PTLD (3,4), which we include under the term PTLDs and differentiate when necessary. Posttransplant lymphoproliferations not related to EBV are beyond the scope of this review. Such PTLDs probably are a different entity, are rare in children, and have later onset with distinct pathogenesis and worse response to treatment (5). Posttransplant lymphoproliferations are by far the most common tumors found in children after organ transplantation (52% of all tumors found during long-term follow-up versus 15% in adult recipients) and comprise 73% of childhood malignancies in this population, illustrating the early onset of this disease (6). In pediatric liver graft recipients, PTLDs incidence ranges between 2.8% and 20% with up to 60% mortality. However, these figures have improved in more recent years thanks to increased awareness and better management of the disease (1–3,7–10). RISK FACTORS AND PATHOGENESIS Epstein-Barr virus infection, and especially primary infection, is a well-known risk factor for PTLDs (2,7), highlighting the increased incidence of PTLD in children as compared with adults (11). Because of their young age, 60% to 80% of pediatric liver transplant recipients are not yet infected by EBV. Primary EBV infection occurs in most children within the first 3 months after transplant, without symptoms or graft dysfunction. When present, symptoms relate closely to emerging PTLD, which mainly occur concomitantly to primary EBV infection (2,8). The high rate and the early onset of this EBV primary infection highlight that the virus probably is acquired rapidly through the graft or blood products. Epstein-Barr virus transmission by the EBV-infected lymphocytes of the donor, which are transplanted with the graft, can lead to PTLD of donor origin often localized to the liver, more specifically to the hilum. However, a recipient origin seems to be more frequent, as discussed below. The load of immunosuppression, probably more than its type, is a second trigger for PTLD. Use of high doses of tacrolimus and OKT3 or antithymocyte globulins all were shown to increase the risk for PTLD (2,10). This may explain variability in the incidence of PTLDs among patients with different transplanted organs (1,9). Third, in addition to the virus itself, the high EBV load in peripheral blood or plasma of transplant recipients also is predictive of PTLD (12–15). These three factors underline the hypothesized PTLD pathogenesis in which EBV is the cornerstone (Fig. 1). The first step of natural EBV infection is a lytic phase, mainly located in the epithelial cells of the oropharynx, leading to new virus production and death of the infected cells. After primary infection, the EBV lytic cycle is usually inactive in healthy carriers but may be reactivated in some conditions, such as immunosuppression. Concomitant with the productive phase, lifelong infection is promoted by latent infection in B lymphocytes, leading to EBV-infected B lymphoblasts in which EBV genome is integrated as an episome and multiplies with the host cell. Early expansion of the infected blasts may be important for viral survival, and high viral load probably favors reactivation of the lytic cycle (16). Latent EBV has a complex survival strategy involving 1) the production of viral proteins that can transform human cells or inhibit antigen processing, 2) the secretion of a cytokine and soluble receptor that act against the cytotoxic cellular immunity, and 3) the down-regulation of its gene expression (17). In this latent pathway, EBV is a well known and potentially transforming virus, and is clearly associated with a variety of human malignancies. However, EBV infection is generally silent thanks to a strong host cellular immune response directed at viral-specific antigens expressed both on the lytically infected cells and on the latently infected EBV lymphoblasts, and thanks to a constant balance between these two parameters (17–19). New gene mutations discovered in the X-linked lymphoproliferative disease and involving interactions between T cells and B-EBV–infected lymphocytes highlight the necessity of this immune surveillance (17). Some EBV-infected B lymphoblasts finally differentiate in memory B cells, presenting no more viral-specific antigens, and, therefore, are invisible to cellular immunity and represent the final EBV reservoir (18). In this healthy carrier status, EBV most often is undetectable in the blood, and the lifelong EBV carriage is generally asymptomatic. In immunosuppressed, healthy patients who are immunized against EBV before transplant, this mechanism is almost unchanged. The only difference is an increased virus burden in the blood, the result of an increased EBV-infected memory B-cell pool and not of circulating EBV-blasts, as previously supposed (20). This phenomenon probably is the result of a more frequent EBV-lytic cycle in immunosuppressed subjects, allowing apparition of more latently infected B-cell clones. However, PTLDs are characterized by EBV-lymphoblast proliferation. Inappropriate immunosuppression, leading to imbalance between host surveillance and viral infection, make possible this proliferation (17). These EBV-proliferating cells express specific viral antigens and remain possible targets for the immune system, explaining the well-described regression of PTLDs after immunosuppression tapering (21). Most often, this alleviated immunosuppression does not lead to immediate graft rejection, probably because of the particular antiinflammatory cytokine profile induced by EBV-proliferating blasts (17,21). If latent EBV infection is principally involved in PTLD pathogenesis, the presence of the lytic cycle may play a role, for example, by producing newly infected B lymphocytes and increasing the viral load. The observation that transcripts of an EBV-lytic gene increase in the blood of patients with PTLD sustains this theory (22). Although PTLD also can occur in patients with previous immunity against EBV before transplant or in long-term observation of patients who undergo uncomplicated primary EBV infection, these presentations are much less common in children than are early-onset (within 2 years) PTLDs related to primary infection. These late-onset PTLDs are similar to PTLDs observed in adult patients with a same incidence inferior at 2%. Their pathogenesis seems to be somewhat different and can involve both B-EBV lymphoblasts and B-infected memory B cells, more often displaying gene mutations. Therefore, complete response to withdrawal of immunosuppression is less common, and additional treatments may be required (see below), impairing the prognosis (11,23–25).FIG. 1.: Pathogenesis, follow-up, and treatment of Epstein-Barr virus (EBV)–related PTLD in children after liver transplantation. Sixty to 80% of patients do not have immunity against EBV before transplant and most of them have primary EBV infection in the 3 months after transplant; the virus probably is transmitted with the graft or the blood products. The majority of PTLD occur within 2 years after transplant as complications of this primary infection. PTLD occurrence in transplant recipients is favored by imbalance between the EBV-infected cells and the EBV specific T-cell immunity, which is impaired by immunosuppression. Quantitative determination of these two parameters allows earlier diagnosis, allows observation of response to treatment, and will probably lead to efficient preemptive strategies in the near future. Gold standard treatment of PTLD is decrease or withdrawal of immunosuppression. Additional or alternative treatments are summarized, aiming all to either increase the immune response or inhibit the EBV-infected cells. Treatment currently under investigation or not widely available are in [ ].Additionally, some interesting reports question a direct effect of immunosuppressive drugs such as cyclosporine or tacrolimus, which may enhance growth of different tumoral cell lines in vitro and in vivo, probably by inducing the synthesis of transforming growth factor-β (26,27). Concomitant viral infections such as cytomegalovirus or hepatitis C are other risk factors for PTLD that have been discussed (28,29). CLINICAL PRESENTATION AND DIAGNOSIS Posttransplant lymphoproliferative disease presentation is often aspecific, and it is important to keep it in mind because prognosis seems to be related to early diagnosis. Several clinical symptoms are similar to those described in the natural EBV infection. However, if mononucleosis symptoms are related to the host immune response, those of PTLDs are caused by the proliferation of EBV-infected cells. Fever and impaired general condition such as irritability, poor appetite, or weight loss are invariably reported. Lymph node enlargement or hepatosplenomegaly often are present with frequent lymphoproliferative involvement of the graft. Tonsillitis is common in lymphoproliferation concomitant with primary EBV infection, although stridor and respiratory distress are usually later symptoms. Intestinal symptoms including vomiting, diarrhea, or gastrointestinal bleeding are sometimes the first signs of the disease. Sepsislike symptoms are also described. A major biologic preliminary sign of PTLDs is increased γ-globulin concentration, with monoclonal/oligoclonal production. High C-reactive protein level, hypoalbuminemia, neutropenia, and thrombopenia are also reported (1,2,7–10). All these symptoms or signs should trigger a search for EBV infection and lymphoproliferation. If no lymphoid mass can be detected while these symptoms are present and EBV infection is confirmed, the diagnosis of preliminary PTLD (pre-PTLD) or nonspecific viral syndrome (2,4) has to be considered and the patient must be observed carefully. When possible, any suspected lymphoid mass should be biopsied to confirm the precise diagnosis of PTLD. Biologic investigations include total blood cell counts, inflammatory parameters, liver function tests, protein electrophoresis, numeration and clonality of circulating lymphocytes, and extensive microbial and virologic screening. Ultrasound or computer tomography scan, including cerebral computer tomography scan are necessary to detect lymphoproliferations because central nervous system involvement can occur with minimal symptoms. Surgical biopsy is preferable to needle biopsy when possible (3). Bone marrow functional defects lead to perform bone marrow aspiration. Digestive endoscopy with biopsy is a useful tool in the presence of gastrointestinal symptoms (30). Tonsillectomy in characteristic clinical context allows exact diagnosis and is the first step of treatment (31). Histopathology gives the final and precise PTLD diagnosis according to the previously described classification (32). Histopathology should include EBV search by in situ hybridization for EBER-1 (33) and immunostaining for T- and B-lymphocyte surface proteins and for cellular immunoglobulins (34). Monomorphic and multifocal diseases seem to have unfavorable prognosis. Some authors distinguish the mononucleosis-like form, in which EBV-linked lymphoproliferation is found without tissue architecture abnormalities, as not “real” PTLD (3,4). However, this frequent form in children is part of the continuum of EBV disease and requires careful management, the reason we include it with PTLDs. Monoclonality of the tumor may be demonstrated by molecular genetic analysis of immunoglobulin gene rearrangement or of EBV-DNA clonality, even if its clinical significance is not yet clearly proven (34,35). Gene mutations are not the rule, and their prognosis is controversial even if some authors report the relevance of BCL-6 or c-myc abnormalities (23,35). In most series, the donor or host origin of the tumor is not detailed. The latter probably predominates, but tumors of donor origin seem to have better prognosis, probably related to the more localized pattern (36). EBV INFECTION FOLLOW-UP It is now clear that the crucial point in PTLD diagnosis and management is closely monitoring EBV infection. If specific IgM serology is not clearly delayed, regarding EBV primary infection diagnosis (8), major limits in its sensitivity and specificity in comparison with polymerase chain reaction lead to its progressively decreasing usefulness. Its only indication probably remains in pretransplant EBV status determination, even if passive immunization must be considered in a very young child with positive antibodies (4,37). The cornerstone of diagnosing EBV infection is the EBV viral load and the host EBV-specific cellular immune response determination (Fig. 1). An increased EBV viral load, determined by semiquantitative or quantitative polymerase chain reaction, is usually the rule in PTLDs. The predictive value of its decrease during PTLD treatment is also described (12,13,15,38,39). Because of the short EBV doubling time in PTLDs, weekly samplings probably are necessary in high-risk patients (15). However, the various ways to determine and express the results hamper data comparison, and the finding of an abnormal EBV viral load may occur because it may vary depending of these parameters. Viral load expression as EBV copies by number of blood cells or by blood quantity should be avoided because of the high variability of DNA extraction efficiency. New real-time, quantitative polymerase chain reaction with normalization of the EBV concentration to coamplified genomic DNA will probably offer more reproducible and standardized values in the future (40). Moreover, increased viral loads are reported in asymptomatic patients, especially in the pediatric population at the time of primary infection, and in patients recovering from PTLDs without relapse, and this limits the specificity of the method (13,38). Determining the plasma or serum EBV viral load to improve this specificity is still controversial (14,15). Anti-EBV–specific cellular immunity has been well described for about ten years (41). Several EBV-derived antigens from latent and lytic viral proteins are recognized by cytotoxic CD8 T cells in an HLA class I– restricted presentation (19). Although this CD8 T-cell response is the main cellular response, an additional role of the CD4 and NK cells is increasingly reported (42). Three new rapid and sensitive diagnostic tools are available to study this specific cellular immunity. Two detect, at a cellular level and after appropriate specific stimulation (by an EBV-immortalized autologous lymphoblastoid cell line or by EBV-derived antigenic peptides), the EBV-induced cytokine secretion (most often gamma interferon) of EBV-specific T cells, by enzyme linked immunospot assay (Elispot) or by intracellular staining and flow cytometry analysis (43,44). The third is based on the in vitro synthesis of a fluorescent tetrameric complex that mimics viral epitope on an HLA type I molecule. This complex has a very high affinity for the T cells specific for this viral epitope and stains them, allowing direct and highly sensitive visualization of these T cells by flow cytometry analysis, but without direct functional assessment (45). Based on one of these three techniques, some recent small series report the protective role of EBV-specific cellular immunity after bone marrow transplantation and its correlation to the viral load (46,47). In our recent experience using Elispot technique in pediatric liver graft recipients, we demonstrate deficiency of the anti-EBV cellular immunity in patients with PTLDs, the ratio between viral load and specific cellular response being predictive of the emergence of PTLD (48). Follow-up of this immunity also allows better management of immunosuppression during PTLD treatment, thanks to better determination of the timing for reintroducing these drugs (46,48). Finally, in our experience, high specific cellular immunity concentrations seem to be the rule in asymptomatic patients with high viral load, leading to equilibrium and explaining the lack of specificity of the viral load alone. PREVENTION AND TREATMENT The gold standard in treating PTLDs is modulating immunosuppression (21). Preemptively tapering cyclosporine or tacrolimus to reach low blood concentrations is proposed in high-risk previously EBV-negative patients with high viral load and in patients with preliminary PTLDs or nonspecific viral syndrome, to avoid emergence of PTLDs (2,4,37). Careful clinical, biologic, virologic, and radiologic monitoring is necessary. Additionally, future preemptive strategies probably will include follow-up of the specific cellular immunity (46). Efforts also will aim to decrease PTLD incidence by establishing more appropriate immunosuppressive therapies. In kidney transplant recipients, authors reported a decreased incidence of EBV-related complications with a steroid-free immunosuppressive protocol using acyclovir and mycophenolate mofetil (49). Some new immunosuppressive drugs, which inhibit EBV-transformed B-lymphocyte growth, are also under study for PTLD prevention (50). Classically, withdrawal of all immunosuppressive drugs is described after PTLD diagnosis, except for steroids, which are kept at a maintenance dose. Response to treatment is followed daily, clinically and biologically, with control of imagery and viral load every week. The timing for restarting immunosuppression is controversial and currently based on biologic or histologic signs of graft rejection (2,4,8,51). After a median delay of 4 weeks, 60% to 74% of patients with immunosuppression withdrawal experienced acute rejection, with a 6% to 17% incidence of chronic rejection and a graft loss rate of 3% to 17% (2,51,52). Most patients had no evidence of PTLD at the time of rejection, and the viral load was normalized concomitantly (38). In more recent data, we demonstrate that, in patients with PTLDs, deficient EBV-specific T-cell response rapidly increases after withdrawal of immunosuppressive drugs. Close monitoring of this specific cellular immunity allows early reintroduction of immunosuppressive drugs without PTLD relapse or graft rejection, making way for promising future strategies (48). Most often, antiviral drugs such as acyclovir or ganciclovir are given intravenously in addition to immunomodulation therapy, even if their efficiency on the latent cycle of EBV infection and related PTLDs remain largely unproven (4). Possible benefits of these agents could be limiting the EBV switch to the lytic cycle and consecutive apparition of new latently infected clones, or therapeutic action on concomitant cytomegalovirus disease. Some authors also report preemptive interest in a 100-day course of intravenous ganciclovir in high-risk patients (EBV-seronegative recipients with EBV-seropositive donors). However, the necessity of prolonged hospitalization limits this (37). Treatment also may include drugs or cytokines that modulate immunity, such as intravenous immunoglobulins, alpha or gamma interferon, and antiinterleukin-6 antibodies, to enhance the anti-EBV immune response or to inhibit viral properties (4,53,54). Again, direct evidence of any advantage of these therapies has not been demonstrated in controlled comparative studies, and their use is based on case reports and on theoretical knowledge. In our experience, we observe an outcome comparable to that of others without using this kind of treatment (8). Improvement usually occurs after 2 to 4 weeks of treatment. If not, or in the case of worsening despite discontinuing immunosuppression therapy, lesions must be biopsied again and additional treatment considered (4,24). The use of anti–B-cell monoclonal antibodies (anti-CD21 or -CD24) has been described with success, but is no longer available (55). If tumor cells express the B-cell marker CD20 at histology, anti-CD20 monoclonal antibodies (rituximab, Roche, 4 weekly infusions of 375 mg/m2) has been proposed as a second step of PTLD treatment (56). In central nervous system involvement, these antibodies should be used intrathecally because they do not cross the blood–brain barrier (57). Preliminary data about rituximab therapy are promising, and controlled studies are currently under way. In adult patients immunized against EBV before transplantation and who have late-onset PTLDs, tumor progression is observed in three of four cases despite rituximab therapy, with rapid regression of the peripheral blood viral load, underlining the probability of a different pathogenesis for this type of PTLD compared with early-onset PTLD that occurs in patients previously negative for EBV (25). As mentioned above, for healthy immunosuppressed patients but in higher proportions, increased EBV peripheral blood load in adult transplant recipients with PTLD seems to be caused by an increased burden of nonmalignant EBV-infected memory B cells, probably resulting from increased production of EBV by tumoral lymphoblasts, leading to newly infected nonmalignant B cells. However, neoplastic cells are probably restricted to lymphoid tissues and do not appear in peripheral circulation. If the memory B cells respond well to rituximab, tumoral cells probably have variable sensitivity to this drug and disease may progress despite normalization of the blood viral load (25). Chemotherapy is indicated in lymphomatous PTLD, whether in primary presentation or in evolution from a polymorphic to a lymphomatous form. This evolution is not exceptional, highlighting the necessity for repeating histologic analysis in disease progression. Chemotherapies used are those classically described for the same tumors in nonimmunosuppressed patients, without apparent increased toxicity if immunosuppression therapy is withdrawn (4,24). Surgery and radiotherapy are discussed case by case. Reconstitution of the EBV-specific cellular immune response by infusing the patient's own anti-EBV cytotoxic T cells, grown in vitro with appropriate stimulation, is one promising, future preemptive or curative PTLD treatment (58). This interesting technique may not be restricted to patients immunized against EBV before transplantation as previously supposed, and de novo activation of this cellular immunity from the blood of EBV-seronegative patients with successful use in subsequent PTLD has been described (59). However, it remains a very heavy procedure, limiting routine application. Some authors also report duality in the CD8 T-cell response to EBV. Types 1 and 2 CD8+ T-cells are described in parallel to the classic subsets of CD4+ T-cells. EBV-specific type 2 CD8+ T-cells seem to be less cytotoxic than type 1 and can activate resting B cells, potentially reactivating latent EBV infection with possible consequence on PTLD (60). Infusion of donor bone marrow cells could be another way to decrease the incidence of PTLDs (61). Finally, a future anti-EBV vaccine offers great promise, and several reports describe the EBV lytic cycle and derived antigens as the target. The viral protein gp350, which EBV enters in the cell, is a potential candidate currently under study (62,63). CONCLUSION In conclusion, further knowledge of PTLD pathogenesis and management in children after liver transplant will allow earlier diagnosis and treatment and improve prognosis, despite no current decrease in incidence. Indeed, PTLD-linked mortality rates have decreased from 60% to around 10% to 20%, and even zero in more recent series (2,7,8,10,51,52). However, PTLD incidence remains unchanged, between 5% to 10%. Recent insights in EBV infection diagnosis, follow-up, and subsequent, new preemptive therapy will probably decrease the incidence and related morbidity (mainly graft rejection) in the future, improving results of liver transplantation in children.